Formulation of 3D pharmacophore models for bradykinin B2-receptor antagonists

In order to make a further contribution to the elucidation of the essential structural features for bradykinin (BK) antagonism, we extracted 3D pharmacophore models from a training set consisting of nine relatively rigid, small organic, non-peptide molecules, reported to be more or less active, competitive BK B₂-receptor antagonists. This was accomplished by means of the expert system Catalyst. The information contained in one of these models was then used to identify relevant structural features and perform consensus molecular dynamics simulations including, in addition to the non-peptide antagonist set, four prototypical linear and cyclic peptide antagonists, and BK itself. This combined approach allowed us to identify regions of the conformational space shared by this series of compounds, which includes highly flexible, and, hence, otherwise almost inaccessible for conventional conformational sampling techniques, peptide molecules. As a result, we obtained a relatively small number of extended pharmacophore models, whose average, together with the original Catalyst model, could be used to search proprietary 3D databases for potential candidates. The results of such a search are also reported.