Recent Calculations on Human Thrombin Inhibitors

It is noteworthy that the MD simulations, even without reparametrization of the LIE equation, gives a good ranking of the rather diverse set of thrombin inhibitors. The calculations thus reproduce, e.g., the large affinity difference between the S-form of Tla, which is a potent inhibitor in the nM 

It is also encouraging that the calculations reproduce the stereospecificity of the two enantiomers of Tla. Here the predicted binding free energy difference is 2.1 kcal/mol in favor of the 5-enantiomer while the experimental difference is roughly 1.4 kcal/mol. The structures of these 

The results for thrombin show that our previous parametrization of the LIE coefficients holds rather well in this case, provided that a constant term of -2.9 kcal/mol is added. At present it is not clear to us why thrombin would require such a constant term while, e.g., trypsin does not, but this issue is currently under investigation (see also Ref. 47 for a discussion of thrombin versus trypsin). Furthermore, one should note that with our computational procedures and the Gromos87 force field the results for thrombin inhibitors differ from those of Ref. 35 as well as Ref. 43. That is to say, three independent studies involving thrombin inhibitors have arrived at significantly different parametrizations of the LIE equation, that in all cases reproduce the experimental data well. It therefore seems clear that the differences in the computational procedures have a definite effect on the parameters of the binding energy approximation. 

---