Theoretical chemistry in drug discovery

Abstract

Ideally, theoretical chemistry would furnish predictions of AGbindi”g prior to synthesis of possible drugs. There are economic as well as scientific imperatives behind this. Pharmaceutical companies may make several thousand compounds for every one which ever gets as far as being tested in a human being, at a cost of hundreds of millions of dollars. Not surprisingly, it is a difficult problem for theoretical chemistry, but significant progress has been made. The techniques used include quantum mechanics, molecular mechanics and statistical mechanics. Quantum chemistry, with its firm physical basis, would be the method of choice but for the computational demands when large numbers of atoms are involved. Molecular mechanics, the use of purely empirical potential functions, does permit the incorporation of thousands of atoms into calculations, but the major recent advances have been achieved by using molecular mechanics potentials within statistical mechanical methodologies. Most notably this has permitted the inclusion of explicit solvent molecules and counterions into computations which do yield free energies, or at least differences in binding free energies, not merely treating isolated molecules [ 11. The way these theoretical tools are employed depends on how much information is available at the atomic level for the problem under investigation. We can list a hierarchy of situations starting with the most tractable. 1. The receptor structure is known in atomic detail from experimental crystallography or NMR spectroscopy. 2. The receptor structure is predicted by theoretical methods. 3. The receptor remains an unknown macromolecule. These categories of problem will now be considered separately.