Comparative molecular modeling analysis of­5-amidinoindole and benzamidine binding to thrombin and trypsin: specific H-bond formation contributes to high 5-amidinoindole potency and selectivity for thrombin and factor Xa

Abstract

The coagulation cascade enzymes thrombin and factor Xa are known to have specificity pockets very similar to those of trypsin and plasmin. However, comparative molecular modeling analysis of the crystal structures of benzamidine-thrombin and benzamidine-trypsin, in conjunction with a docking analysis of 5-amidinoindole and related inhibitors in both enzymes reveals subtle differences between the specificity sites of the two types of enzymes. Specifically, thrombin and factor Xa, which have an alanine residue at position 190, exhibit increased activities for the rigid and more bulky bicyclic derivatives of benzamidine (e.g. amidinobenzofuran, amidinothiophene and amidinoindole), because of additional hydrophobic and H-bond interactions between the inhibitors and the specificity sites, whereas enzymes with a serine residue at position 190, like trypsin and plasmin, exhibit little difference in activity among the same set of compounds because of the orientational restriction imposed on the inhibitors by Ser190, which forms an additional H-bond with the amidino group of the inhibitors. Enzymes of both groups show similar responses to the flexible aminobenzamidine since the smaller size and the rotatable anilino group of the inhibitor would allow the inhibitor to achieve favorable electrostatic interactions with both groups of enzymes. 5-amidinoindole is the most dramatic example of the rigid bicyclic type inhibitor. Based on our docking analysis, we propose that a selective H-bond with the hydroxyl group of the catalytic Ser195 and the subtle differences in steric fit imposed by Ala/Ser at position 190 explain the high potency and selectivity of 5-amidinoindole for thrombin and factor Xa.

Optimizing structural modeling for a specific protein scaffold: knottins or inhibitor cystine knots

“…To this end, if the structures of all homologs of a given protein target are available, the differential analysis of local environments in different model subgroups can help to design highly selective molecules interacting with one subfamily but not with the remaining proteins of the concerned super-family. Homology models can also be useful for the prediction of ligand binding sites [16], for functional annotations [17], or as starting folds for experimental structure determinations [18]. Of course, the best achievable structural model accuracy is critical to extract reliable information from predicted protein folds and give precise answers to the above issues.…”