Computational approaches to identify small-molecule inhibitors of non-traditional drug targets

Abstract

Non-traditional targets for therapeutic intervention are those proteins that have not evolved to bind small molecules, but have instead evolved to bind other macromolecules. Such targets include protein–protein interaction sites, protein–RNA interaction sites and protein–DNA interaction sites. Modulating these biologically important targets will allow us as a community to develop novel therapeutics, but still remains a major challenge. In this thesis, I describe two different computational approaches that I have developed: one for identifying small-molecule inhibitors of protein–protein interactions, and the other for identifying small-molecule inhibitors of protein–RNA interactions. To specifically target protein interaction sites, I have developed a docking method called DARC (Docking Approach using Ray-Casting). This method quantitatively measures the complementarity between the protein surface and a ligand, by using ray-casting to map and compare their shapes. I have applied DARC to carry out a virtual screen against the protein interaction site of the protein Mcl 1, allowing us to identify 6 new inhibitors of this exciting target. To specifically target protein-RNA interactions, I have developed a mimicry-inspired strategy that extracts a “hotspot pharmacophore” from the structure of a protein-RNA complex, and then uses this as a template for ligand-based virtual screening. I have applied this strategy to screen for compounds that inhibit the Musashi-1 / NUMB mRNA interaction, allowing us to identify a new class of compounds that inhibit this interaction in both biochemical and cell-based assays. This thesis is outlined as follows. In the first chapter, I will compare the structural features of inhibitor-bound complexes of traditional versus non-traditional protein targets. In the second chapter, I will present the DARC method and its application to Mcl-1. In the third chapter, I will present various enhancements to DARC method that result in both speed and performance improvements. Finally, in the Fourth chapter I will present the “hotspot mimicry” approach for targeting protein-RNA interactions and application of this approach in identification of inhibitors for Musashi 1 / NUMB mRNA interaction.