Fragment screening and biophysical method development for BET and non-BET bromodomain inhibitor discovery

Abstract

Fragment-based drug discovery has become a popular and successful method for identifying small molecule modulators of various protein targets. Despite drugs and chemical probes having significant 3D-character, traditional fragment libraries are highly 2D with few sp3 centers, stereocenters, and unsaturated ring systems. The use of 3D-enriched fragments is highly debated especially regarding targeting protein-protein interactions. The two main concerns with using 3D-fragments are that the increased complexity of the compounds will cause prohibitively low hit rates and that 3D-compounds are not well suited for targeting the highly 2D protein-protein interaction interfaces. However, it is hypothesized that using 3D-fragments will allow for a fuller exploration of the binding pockets resulting in selectivity amongst targets at the onset of drug discovery. The first part of this dissertation details my work using 3D-enriched fragments for targeting bromodomains, a class of epigenetic proteins that function through protein-protein interactions but have a well-defined binding site. Mis-regulation of bromodomains has been associated with many disorders and cancers making them an attractive therapeutic target. However, gaining selectivity amongst the 61 bromodomain isoforms has proven challenging as they all recognize the same substrate, N-ε-acetylated lysine, a post translation modification on histones. In one project I used a 3D-enriched fragment library to target the first bromodomain of BRD4. In a second project, I developed a screening workflow for screening cocktails of fragments against two proteins simultaneously. Taking advantage of the pros and cons of ligand- and protein-observed NMR binding assays, I used protein- and ligand-observed NMR, I used protein-observed fluorine (PrOF) NMR and 1H CPMG NMR in concert, to screen a 3D-enriched fragment library against the bromodomains of BPTF and PfGCN5. In a third project, I worked to increase the efficiency of 19F labeling bromodomains for their use in PrOF NMR. Here I used tyrosine phenol lyase to make fluorotyrosine from fluorophenol in cellulo for direct incorporation into proteins sans isolation. Lastly, in several other project I worked to crystalize and solve seven bromodomain holo and cocrystal structures to aid in inhibitor development.