Human Tissue Factor: a Molecular Dynamics Study

Kristin Patterson1, Coray M. Colina1, Lalith Perera2, and Lee G. Pedersen3. (1) Department of Materials Science and Engineering, Pennsylvania State University, 121 Steidle Building, University Park, PA 16802, (2) Laboratory of Quantitative and Computational Biology, National Institute of Environmental Health Science, Research Triangle Park, Raleigh, NC 27709, (3) University of North Carolina, Campus Box 3290, Chapel Hill, NC 27599

Most descriptions of the coagulation process state that factor VII (FVII) binds to tissue factor (TF) when an injury occurs and blood leaves the vasculature and encounters TF in the extravascular compartment. Assembly of the FVII(a)/TF complex then leads to initiation of coagulation. The Tissue Factor (TF) pathway (or the extrinsic pathway) is thus believed to play a primary role in initiating blood coagulation during normal hemostasis as well as during many pathologic situations, including arterosclerosis and septicemia [1, 2]. However, there is indirect evidence that this scenario is not accurate [3, 4].

Several studies have also been performed to elucidate the importance of the so-called allosteric disulphide bonds. Human TF consists of 263 residues, the first 219 of which comprise the extracellular region [5]. Extracellular TF (219) consists of two immunoglobulin-like domains connected by a single polypeptide linker (PRO 102 to ASN 107). Each domain is formed by two antiparallel b-sheets. TF has two disulphide bridges at positions 49-57 and 186-209. It has been suggested that the Cys-186-Cys209 disulfide repositions the adjacent F and G beta strands that reorient nearby residues and enable productive binding of the TF substrates.

In this work, we have used Molecular Dynamics (MD) simulations to predict the solution

structure of free TF (fTF) and TF bound to factor VIIa (bTF). A complete refined solution equilibrated model for TF based on available crystal structures (1BOY, 2HFT, 1DAN) using MD simulations in aqueous medium for over 15 ns is presented.

The X-ray crystal structures (1BOY, 2HFT, 1DAN) were used initially as starting points in modeling the unbound and bound structures. The complete models (after missing regions, water molecules, and ions have been added) were subjected to aqueous-phase molecular dynamics (MD) simulations, where unconstrained dynamics were performed using the Cornell et al. force-field (using AMBER9/PMEMD9 programs [6]). The Particle Mesh Ewald (PME) method was used in all aqueous simulations to account for long-range interactions during dynamics. NPT simulations at 300 K, SHAKE algorithm for bond constraints and a 1.0 fs time step were used. The solute, ions/counter-ions and crystal water molecules, were immersed in rectilinear periodic boxes of at least 20Å each side. Standards steps of energy minimization and dynamics were carried out before beginning the “production run”. Finally, the post-equilibration NPT dynamics at 300 K, or production run, were completed for at least 15 ns depending of the study case.

We observe significant inter-domain motion of fTF in the present simulations, where the polypeptide linker has the major contribution to the overall motion. From our preliminary molecular dynamic simulations however, we observe only small changes relative to the disulphide bond presence (or not). We have performed these simulations to compare and complement mutational studies. From our long molecular dynamic simulations, the Cys186-Cys209 disulfide bond of TF (the so-called allosteric disulfide), does not appears to play a structural role on FVIIa activation or reorientation of nearby residues in TF. Although these preliminary results seems reasonable based on the antiparallel beta sheets structure of TF, where each beta sheet contain 7 strands, and where the presence of the disulfide bond near the membrane surface does not appear to provide extra strength to the structure, a much more thorough study is needed.

Finally, these equilibrated solution structures for tissue factor should provide insight into the details of how tissue factor enhances the action of FVIIa. The possibility to predict these dynamic conformations could represent a crucial impact for the areas of

protein and drug design. It is also important to keep in mind that the propose carefully constructed models will, since they intrinsically incorporate the experimental data, provide a complete, dynamical molecular understanding of structures and, as such, will be a natural starting point for suggesting mutational experiments, protein engineering and therapeutic design leads, in addition to provide a uniform foundation with which to incorporate new structural information.


1.Bajaj, S. P., Schmidt, A. E., Agah, S., Bajaj, M. S. and K. Padmanabhan, J. Biol. Chem. 281, 24873-24888 (2006)

2.Davie, E. W., Fujikawa, K., and W.Ksie, Biochemistry, 30, 10363-10370 (1991)

3.Hoffman, M., Colina, C. M., Harger, A. G., Arepally, G., Pedersen, L. and D. M. Monroe, J. Thromb. Haesm.

4.Bach, R. R. Arterioscler Thromb, Vasc. Biol. 26, 456-461 (2006).

5.Daubie, V., Pochet, R., Houard, S. and P. Philippart, J. Tissue Eng. Reg. Med. 1, 161-169 (2007)

6.Case, D.A. , Darden, T.A. Cheatham, III, T.E. Simmerling, C.L. Wang, J. Duke, R.E. Luo, R. Merz, K.M. Pearlman, D.A. Crowley, M. Walker, R.C. Zhang, W. Wang, B. Hayik, S. Roitberg, A.,G. Seabra, K.F. Wong, F. Paesani, X. Wu, S. Brozell, V. Tsui, H. Gohlke, L. Yang, C. Tan, J. Mongan, V. Hornak, G. Cui, P. Beroza, D.H. Mathews, C. Schafmeister, W.S. Ross, and P.A. Kollman (2006), AMBER 9, University of California, San Francisco.