A virtual screening approach applied to the search for trypanothione reductase inhibitors.

A prediction algorithm of the binding affinity of ligands to trypanothione reductase (TR), the enzyme replacing glutathione reductase in the metabolism of trypanosomatidae, has been used for the "virtual screening" of a data base of 2500 molecular sketches and has detected several structures of putative TR ligands. Most of these compounds turned out to be micromolar inhibitors of TR, as predicted by the algorithm. While their inhibitory potencies are lower than those of previously reported compounds, one of the molecules reported here could represent the lead toward a structurally different class of TR inhibitors. The fully automated prediction algorithm converts the 2D molecular sketches into 3D ligand structures, explores the conformational space of the latter, and performs a grid-based, rigid-body docking of the resulting family of ligand conformations into the TR site, calculating enthalpic and entropic binding indexes and predicting the binding affinity. The docking model has also been used to obtain hints about the binding modes of TR ligands.