3-DIMENSIONAL PHARMACOPHORES FROM BINDING DATA

The application of HASL (hypothetical active site lattice) methodology has been successfully extended to generate putative pharmacophoric patterns in three dimensions capable of quantitatively predicting binding activity. The transformation of a HASL model to a pharmacophore is illustrated using pK(i) values published for 84 HIV-1 protease inhibitors. Starting with a HASL model generated at 2.00 Angstrom and containing 899 lattice points, a selective trimming process was used to identify significant lattice points. In this manner, a set of 11 points was found which represents a potential pharmacophoric pattern and predicts the pK(i) activity of the 84-inhibitor set with a correlation (r(2)) of 0.827. Furthermore, the locations of these points were found to coincide with a number of strategic binding areas within the known active site structure HIV-1 protease, thus providing a physical confirmation of their relevancy.