No abstract

Previous structure-activity studies of captopril and related active angiotensin-converting enzyme (ACE) inhibitors have led to the conclusion that the basic structural requirements for inhibition of ACE involve (a) a terminal carboxyl group; (b) an amido carbonyl group; and (c) different types of effective zinc (Zn) ligand functional groups. Such structural requirements common to a set of compounds acting at the same receptor have been used to define a pharmacophoric pattern of atoms or groups of atoms mutually oriented in space that is necessary for ACE inhibition from a stereochemical point of view. A unique pharmacophore model (within the resolution of approximately 0.15 A) was observed using a method for systematic search of the conformational hyperspace available to the 28 structurally different molecules under study. The method does not assume a common molecular framework, and, therefore, allows comparison of different compounds that is independent of their absolute orientation. Consequently, by placing the carboxyl binding group, the binding site for amido carbonyl, and the Zn atom site in positions determined by ideal binding geometry with the inhibitors' functional groups, it was possible to clearly specify a geometry for the active site of ACE.

A conceptually and computationally integrated approach to pharmacophore model validation is discussed at length. It allows one, for the first time, to address the question of pharmacophore existence and assessment of its uniqueness in a rigorous and quantitative manner. The approach has been effectively applied to the analysis of seventeen structurally diverse potent inhibitors of angiotensin converting enzyme and the resultant pharmacophore greatly refines the previously suggested models.

A molecular mechanics program (Maximin) is described with emphasis on its distinctive features. With Maximin, energy can be minimized while maintaining specified geometric relationships within and/or among given sets of atoms. Additionally, two alternative methods for conformational comparisons are supported: a set of flexible molecules can be mapped onto a rigid reference structure, or treating all molecules as flexible entities, one can minimize the conformational variance of the set. The latter method is described here for the first time. Calculations of methotrexate‐dihydrofolate reductase interaction energy, and the analysis of a series of structurally diverse inhibitors of angiotensin converting enzyme are reported. Algorithmic description of the original features of Maximin is also provided.