Chang Jun Lee

Proposed structural models of human factor Va and prothrombinase

Summary.  Background: The prothrombinase complex consists of factor Xa, FVa, calcium ions, and phospholipid membrane. The prothrombinase complex plays a key role in the blood coagulation process.

A revisit to the one form kinetic model of prothrombinase

Thrombin is generated enzymatically from prothrombin by two pathways with the intermediates of meizothrombin and prethrombin-2. Experimental concentration profiles from two independent groups for these two pathways have been re-analyzed. By rationally combining the independent data sets, a simple mechanism can be established and rate constants determined. A structural model is consistent with the data-derived finding that mechanisms that feature channeling or ratcheting are not necessary to describe thrombin production.

A computational modeling and molecular dynamics study of the Michaelis complex of human protein Z-dependent protease inhibitor (ZPI) and factor Xa (FXa)

AbstractProtein Z-dependent protease inhibitor (ZPI) and antithrombin III (AT3) are members of the serpin superfamily of protease inhibitors that inhibit factor Xa (FXa) and other proteases in the coagulation pathway. While experimental structural information is available for the interaction of AT3 with FXa, at present there is no structural data regarding the interaction of ZPI with FXa, and the precise role of this interaction in the blood coagulation pathway is poorly understood. In an effort to gain a structural understanding of this system, we have built a solvent equilibrated three-dimensional structural model of the Michaelis complex of human ZPI/FXa using homology modeling, protein–protein docking and molecular dynamics simulation methods. Preliminary analysis of interactions at the complex interface from our simulations suggests that the interactions of the reactive center loop (RCL) and the exosite surface of ZPI with FXa are similar to those observed from X-ray crystal structure-based simulations of AT3/FXa. However, detailed comparison of our modeled structure of ZPI/FXa with that of AT3/FXa points to differences in interaction specificity at the reactive center and in the stability of the inhibitory complex, due to the presence of a tyrosine residue at the P1 position in ZPI, instead of the P1 arginine residue in AT3. The modeled structure also shows specific structural differences between AT3 and ZPI in the heparin-binding and flexible N-terminal tail regions. Our structural model of ZPI/FXa is also compatible with available experimental information regarding the importance for the inhibitory action of certain basic residues in FXa. FigureSolvent equilibrated models for protein z-dependent protease inhibitor and its initial reactive complex with coagulation factor Xa (show here) are developed.

A proposed ternary complex model of prothrombinase with prothrombin: protein-protein docking and molecular dynamics simulations

The activation of prothrombin (II) to thrombin (IIa) is a key step during the blood coagulation process. Under physiological conditions, the most relevant enzyme for generating thrombin is the prothrombinase complex composed of factor (F)Xa and FVa assembled on a negatively charged membrane surface in the presence of calcium ions [1]. We present here a solutionequilibrated model for the ternary complex of FXa/FVa/II, the zymogen prothrombin docked onto our earlier prothrombinase model [2]. For this purpose, we used several computational methods: homology modeling, loop building and optimization, protein–protein docking and molecular dynamics (MD). Prothrombin is the zymogen of the enzyme thrombin which is generated by two cleavages at Arg320 and Arg271 by FXa in prothrombinase, which leads to distinctive intermediates: meizothrombin (Arg320 cleavage alone) and prethrombin-2 (Arg271 cleavage alone) with F1.2, respectively. Prothrombin consists of a Gla domain, two kringle domains (K1 and K2) and a serine protease (SP) domain composed of the A and B chains. Our prothrombin model is based on the recent X-ray crystal structure of human prethrombin-1 (residues 156–579) [3]. While the X-ray crystal structure of human prethrombin-1 shows distinctive structural features, several regions (residues 156–168, 255–273, 472–475) are missing. We constructed K2SP (residues 169–579) and fragment 1 (F1, residues 1–144) models of II separately and then docked each model onto our published FXa/FVa model [2]. The missing linker region (residues 145–168) in II was built using loop and optimization modules of the Modeller9v1 program [4]. We first completed residues 472–475 in the prethrombin-1 crystal structure using loop modules in the Modeller9v1 program, and then constructed an energy-optimized ensemble of 100 models for 169–579 segment of II including the missing loop (255–273). An optimization method involving conjugate gradients and molecular dynamics with simulated annealing available in Modeller was employed. Thirty representative models with distinctive conformation were selected considering the objective function of Modeller and stereochemical values from PROCHECK [5]. These 30 models were docked onto our prothrombinase structure using a parallelized in-house version of the FTDock program [6]. We then applied experimentally known binding data for the II interaction with FXa or FVa as biological filters. All the filtered docking conformations further ranked by residue level pair potential scores (RPScores) in FTDock were visually inspected. The experimental (Table S1) filters (cutoff distance 6.5 Å) used in docking K2-SP of II onto FXa/FVa were as follows: (i) 205–220 of II binding with FXa [7], (ii) 473–487 of II binding with FVa [8], (iii) 557–571 of II binding with FXa [9] and (iv) 241cn–252cn in the SP domain of FXa (cn = chymotrypsin numbering) binding with II [9]. We also restricted the distances between the Oc atom of the active site Ser195cn in the SP domain of FXawith C atoms of the two cleavage sites, Arg320 and Arg271, in II to be within 35 Å. The detailed parameter set for searching the conformational space in the docking process is given in [2]. After visual inspection on the docking candidates ranked by RPScores, we selected a model for which the distances between the active site Oc@Ser195cn in FXa and the two cleavage sites C@Arg320 and C@Arg271 are 30.9 and 20.8 Å, respectively. For F1 (residues 1–155), we used our prior solutionequilibrated model (residues 1–144) [10]. This model was docked onto the selectedmodel for K2-SP of II with FXa/FVa. Both Gla [11] and K1 [12] domains in II are known to bind with FVa during the activation of prothrombin to thrombin. Along with the biological filters used for docking, we required the interaction of thex-loop in the Gla domain with a putative membrane surface and chose a plausible distance between the C-terminus (Gly144) of K1 and theN-terminus (Gln169) of K2 for the docking to be within 65 Å. After visual inspection for the docking candidates, we selected a final docking model and completed the missing linker (145–168) of II in the docking model using Modeller. The final docking model included unwanted clashes. To refine the docking model and obtain a solution-equilibrated model, we performedMDsimulations. The basic features of the simulation protocol are essentially the same as in our previous modeling work [2]. The docking complex was immersed into TIP3P water boxes of 12.5 Å and then neutralized by three Correspondence: Lee G. Pedersen, Department of Chemistry, University of North Carolina, Chapel Hill, NC 27599-3290, USA. Tel.: +1 919 962 1578; fax: +1 919 962 2388. E-mail: lee_pedersen@unc.edu

Recent Estimates of the Structure of the Factor VIIa (FVIIa)/Tissue Factor (TF) and Factor Xa (FXa) Ternary Complex

The putative structure of the Tissue Factor/Factor VIIa/Factor Xa (TF/FVIIa/FXa) ternary complex is reconsidered. Two independently derived docking models proposed in 2003 (one for our laboratory: CHeA and one from the Scripps laboratory: Ss) are dynamically equilibrated for over 10 ns in an electrically neutral solution using all-atom molecular dynamics. Although the dynamical models (CHeB and Se) differ in atomic detail, there are similarities in that TF is found to interact with the gamma-carboxyglutamic acid (Gla) and Epidermal Growth Factor-like 1 (EGF-1) domains of FXa, and FVIIa is found to interact with the Gla, EGF-2 and serine protease (SP) domains of FXa in both models. FVIIa does not interact with the FXa EGF-1 domain in Se and the EGF domains of FVIIa do not interact with FXa in the CHeB. Both models are consistent with experimentally suggested contacts between the SP domain of FVIIa with the EGF-2 and SP domains of FXa.

A proposed structural model of human protein Z

Analysis of structures and sequence similarities of the vitamin K-dependent (VKD) proteins in the blood coagulation pathway has shown that these proteins have evolved through a series of gene duplications and diversification, acquiring a degree of functional diversity in the process [1–3]. Protein Z (PZ) is a VKD plasma glycoprotein that is highly conserved across different species. It is homologous to the blood coagulation factors VII (FVII), FIX, FX, and protein C (PC). However, PZ differs from other coagulation proteins in that it lacks the critical histidine and serine residues of the catalytic triad, and is therefore not a zymogen of a serine protease (SP) [4,5]. Human PZ was first isolated in 1984 [6], and its gene was characterized in 1998 [7]. PZ is relatively abundant in humans, with a wide plasma concentration range, and circulates as a complex with PZ-dependent protease inhibitor (ZPI) [8–10]. It has been shown to function as a cofactor in the inhibition of activated FX (FXa) by ZPI, causing a thousandfold increase in inhibition in a Ca-dependent manner [11]. The relative abundance in plasma and sequence similarity with other VKD proteins serve to pique our interest in the PZ system. In order to gain a better understanding of the function and evolutionary significance of PZ, it is important to start from structural information at an atomic level. Although PZ was isolated over 20 years ago, its physiologic role in relation to its molecular structure is still not clear. In this study, we propose a solvent-equilibrated structural model of human PZ using homology modeling and molecular dynamics (MD) simulation. The possibility of using this approach is supported by recent findings [12]. Human PZ is a single-chain molecule of 360 residues with four domains: a Gla domain (residues 1–46) with 13 ccarboxyglutamic acid (Gla) residues (residues 7, 8, 11, 15, 17, 20, 21, 26, 27, 30, 33, 35 and 40), two epidermal growth factor (EGF)-like domains – an EGF1 domain (residues 47–83) with a b-hydroxyaspartate (Bha) residue at 64, and an EGF2 domain (residues 85–126) – and an SP-like domain (residues 135–360) with high homology to the catalytic domain of other VKD SPs [4,5]. We have employed an iterative process of homology modeling and structure evaluation followed byMD simulation to construct a solvent-equilibrated structural model of PZ. Two different schemes were adopted to construct the homology model for PZ: one usingmultiple sequence alignment of human FVIIa [13], FIXa [14], FXa [15] and PC [16] sequences, and the other using FVIIa as a single template (see supplementary Fig. S1 for the sequence alignment). A multiple sequence alignment created with CLUSTALW [17] was used as an input to construct 30 homology models using MODELLER 8v2 [18] (see supplementary Fig. S1 for a description of MODELLER parameters). We chose the best model on the basis of the stereochemical values from PROCHECK [19] and objective function values fromMODELLER. This model had nine disulfide bonds (residues 18–23, 51–62, 56–71, 73–82, 89–101, 97–110, 112–125, 163–179 and 287–301). To obtain a refined solvent-equilibrated model and to remove bad contacts in the homology model, we performed an MD simulation using PMEMD9 in the AMBER9 [20] suite with the ff99SB forcefield. The total system was composed of 99 459 atoms, including 12 calcium ions, four sodium counterions for neutralizing the system, and 31 307 TIP3P water molecules. The seven conserved calcium ions in the Gla domain were placed on the basis of FVIIa, and the additional calcium ions were placed with malonate coordination (both carboxylates of Gla involved). Constraints were applied on the backbone during the early stages of minimization and the NPT (fixed number of molecules, pressure and temperature to simulate benchtop conditions) equilibration phase. The final 9-ns unconstrained trajectories (1.5-ps time interval in the NPT ensemble at 1 atm and 300.0 K) were analyzed. For the homologymodel that employed FVIIa as a single template, the refinement protocols above were followed. The backbone root mean square deviation (RMSD) plots of the simulation vs. the starting structures show that the domains are solution equilibrated by 9 ns (Fig. 1A). The overall structure is similar to that of other VKD proteins, especially FVIIa (Fig. 1B). The close contacts in the homology model are removed, and the stereochemical values are also well maintained throughout the simulation. The main findings of the simulation can be summarized as follows: (i) the secondary structural motif of each domain is well maintained and stable during the MD simulation; (ii) the individual domains show relatively smaller RMSDs and fluctuations than the overall structure – the larger RMSD for Correspondence: L. G. Pedersen, Department of Chemistry, UNCCH, Chapel Hill, NC, USA. Tel.: +1 919 962 1578; fax: +1 919 962 2388; e-mail: lee_pedersen@ unc.edu

A model for the unique role of factor Va A2 domain extension in the human ternary thrombin-generating complex

An all-atom human ternary model for the prothrombinase-prothrombin complex, including metal ions and post-translationally modified residues, was constructed from existing X-ray crystal structures. The factor Xa-prothrombin interface was taken from an existing ternary model, which locates the active site of factor Xa in the vicinity of prothrombin cleavage positions. The three sulfotyrosine residues at the C-terminal sequence of factor Va A2 domain are accommodated by modelling rational interactions with positively charged patches on the surface of prothrombin. The entire model is then solvent-equilibrated with molecular dynamics. This ternary model for the thrombin-generating complex provides an estimate as to the role of the C-terminus of the factor Va A2 domain: to establish an interface between FXa and prothrombin and to stabilize the orientation of this interface.

Molecular dynamic simulations of the binary complex of human tissue factor (TF(1-242) ) and factor VIIa (TF(1-242) /FVIIa) on a 4:1 POPC/POPS lipid bilayer.

Human tissue factor (TF, residues 1−263) is a cell-surface receptor consisting of three domains: an extracellular domain (residues 1−219=soluble form TF=sTF) composed of two C2-type immunoglobulin-like modules (N-module with residues 1−106 and C-module with residues 107−219), a transmembrane (TM) domain (residues 220−242), and a cytoplasmic domain (residues 245−263) [1]. After vessel injury TF activates the blood coagulation cascade by binding a serine protease factor VIIa (fVIIa). Full-length TF acts as a cofactor to substantially increase the activity of factor VII (fVIIa) for a substrate factor X (fX) in the presence of an anionic phospholipid membrane and Ca2+. sTF shows 4 % of the activity of full-length TF [2]. The mechanism of enhancement of activity by TF, whether soluble or full-length, is still elusive at the molecular level. Recent pioneering MD simulations by Ohkubo and workers [3] provided a new understanding of the dynamic features, structure, and atomic interaction of the proteins (sTF1–213 and fVIIa) and their binary complex (sTF1–213/fVIIa) on a lipid bilayer composed of 100 % DOPS (1,2-dioleoyl-sn-glycero-3-[phosphor-L-serine]). In this letter, we have modeled a transmembrane (TM) domain of TF to a linker region which connects the truncated sTF210 in the X-ray crystal structure (PDB code: 1DAN) [4] with the TM domain (see Supporting Information, SI). The newly modeled TF1–242/ fVIIa is incorporated onto a 4:1 POPC/POPS bilayer model with additional Ca2+ ions present beyond those bound to fVIIa. In building the initial structures, we adopted four criteria: 1) that the TM domain of TF penetrates the lipid bilayer, 2) that three hydrophobic residues in the gamma-carboxyglutamate-rich (GLA) domain of fVIIa (Phe4, Leu5, and Leu8) are located in the hydrophobic region immediately beneath the phosphate groups in POPC/POPS, 3) that the three charged residues in the linker of the tissue factor (Lys214, Glu216, and Glu219) are placed on the charged surface of the lipid membrane, 4) that the initial orientation of TF1–242/fVIIa on the lipid membrane is determined from a putative ternary complex of sTF/fVIIa/fXa [5]. For the initial structure based on these four criteria, we estimated the distance between Cα of Ser195 in active site of fVIIa and the nearest P atom to be ~80 A. We performed two independent simulations employing the Amber11 [6] and NAMD2.8 [7] packages with independent force-fields. The details of the molecular dynamics (MD) simulations are given in more detail in SI. Even though the four criteria for the initial structure set-up were applied to the both simulations, the MD simulation environments differed somewhat (see Table S1). Interestingly, the separate RMSD profiles of TF1–242/fVIIa, fVIIa, sTF, and GLA domain obtained by the independent methods are similar (see Fig. S1). The last snapshot (115 ns) of the Amber simulation is shown in Figure 1. We averaged the shortest distance between any phosphate atom and Cα of Ser195 for 100 ps snapshots during the last 10 ns of simulations by the Amber and NAMD packages and find these distances to be 77.96 A (±1.58) for Amber and 76.82 A (±2.24) for NAMD (Fig. S2). We thus obtained consistent values for the active site height irrespective of the method, force-field, and the locally different simulation environments. These consensus values of the height of Ser195 from the lipid membrane estimated by the two distinct simulations are comparable to the experimental estimates made using TM-containing TF (74 ± 2 A [8], 75.0 ± 1.8 A [9], 76 ± 3 A [10]). Ohkubo et al. [3] defined the reference point of the lipid surface as the average z-coordinate of all carboxy oxygens in the head groups of DOPS of proximal lipid leaflet, but found significantly larger distances of the active site of fVIIa above the surface for sTF1–213/fVIIa. These may reflect the differences in lipid composition, concentration of ions (Ca2+Na+Cl−) or the effect of the TM domain addition. We also investigated the dynamic local interaction between the GLA domain of fVIIa, the TF linker region and sTF1–213 (Table S2) with the bilayer. From the last 10 ns trajectories, we observed significant interactions involving residues Ser162 and Lys181 of TF in both the Amber and NAMD simulations. The involvement of these residues was also reported in Ohkubo et al.’s pure DOPS bilayer simulation [3]. Both Amber and NAMD simulations show new interactions between the tissue factor transmembrane helix and linker region with the bilayer that cannot be present for simulations with sTF (See Table S2). These new interactions may contribute to the smaller active site height that we observe. This concept is consistent with the desGLA-fVIIa/TF, fVIIa, fVIIa/TF comparative experiments [9], which suggest that it is TF, not fVIIa, that plays a dominant role in orienting fVIIa for the optimal position of active site on the phospholipid; while interactions involving the fVIIa GLA domain, the fX GLA domain, and phospholipid are critical for efficient proteolysis of factor X on a membrane surface [11]. The fVIIa active site location estimated in the current simulations may provide an optimal height and orientation for the interaction with a substrate fX, and therefore be a significant factor in the enhancement of the fVIIa catalysis provided by TF. In a docking of the final snapshots from the Amber and NAMD simulations with a ternary model of sTF/fVIIa/fXa [5], both show reasonable ternary structures on the lipid bilayer by minimal adjustment of linker region between EGF-1 and EGF-2 domains of fXa (See S4). Finally, Ca(POPS)2 structures, similar to those proposed from site-resolved multidimensional solid-state NMR (SSNMR) experiments, are dynamically generated in both simulations [14]. The final snapshot structures of the simulations are available on request from the authors. Figure 1 The orientation of TF/fVIIa on a bilayer of 4:1 POPC:POPS from the final snapshot (115 ns) of the Amber simulation (A) and its rotated view by 90° (B)

Analysis on long-range residue–residue communication using molecular dynamics

We investigated the possibility of inter‐residue communication of side chains in barstar, an 89 residue protein, using mutual information theory. The normalized mutual information (NMI) of the dihedral angles of the side chains was obtained from all‐atom molecular dynamics simulations. The accumulated NMI from an explicit solvent equilibrated trajectory (600 ns) with free backbone exhibits a parabola‐shaped distribution over the inter‐residue distances (0–36 Å): smaller at the end regimes but larger in the middle regime. This analysis, plus several other measures, does not find unusual long‐range communication for free backbone in explicit solvent simulations. Proteins 2014; 82:2896–2901.

Computational study of the putative active form of protein Z (PZa): Sequence design and structural modeling

Although protein Z (PZ) has a domain arrangement similar to the essential coagulation proteins FVII, FIX, FX, and protein C, its serine protease (SP)‐like domain is incomplete and does not exhibit proteolytic activity. We have generated a trial sequence of putative activated protein Z (PZa) by identifying amino acid mutations in the SP‐like domain that might reasonably resurrect the serine protease catalytic activity of PZ. The structure of the activated form was then modeled based on the proposed sequence using homology modeling and solvent‐equilibrated molecular dynamics simulations. In silico docking of inhibitors of FVIIa and FXa to the putative active site of equilibrated PZa, along with structural comparison with its homologous proteins, suggest that the designed PZa can possibly act as a serine protease.