Innovations in Targeting Dynamic Proteins With Small Molecule Modulators

Abstract

For decades, small molecule modulators of protein function have been vital for advancing both understanding of basic biology and therapeutic discovery; however, there are classes of proteins for which small molecule discovery has lagged. The majority of the proteins studied in this work are transcriptional coactivators, proteins that have great potential as therapeutic targets, yet are often classified as “undruggable” because of their atypical structure and mechanism of action. Coactivators are highly malleable and their interactions typically occur over broad surfaces areas with only moderate affinity, making them difficult to target. Recent advances in protein-protein interaction (PPI) inhibitor discovery, exploiting allostery and dynamic substructures within target proteins, has led to some success, but selectivity remains a challenge. Selectivity is pivotal for candidate probe molecules due to the extensive interaction network of these dynamic hubs. The following dissertation addresses the challenges of selectivity and presents small molecules that target coactivator interactions. A large portion of this work is dedicated to targeting the activator interaction domain (AcID) of the coactivator Med25. This domain contains a unique protein fold and is not required for basal transcription, but interacts with activators to regulate genes implicated in disease. Previous work from the Mapp lab led to identification of promising AcID PPI inhibitors. Here, this work was continued with the compound norstictic acid (NA). NA demonstrated high selectivity for AcID interactions and, using biochemical techniques in combination with molecular dynamics simulations, the mechanism of inhibition was elucidated: covalent modification of lysine residues within a dynamic loop leads to both orthosteric and allosteric inhibition of activator binding. NA proved useful for studying Med25 in a cellular context, engaging with full length protein from cell extracts and inhibiting the interaction between endogenous Med25 and transcriptional coactivators. Ultimately NA was used to probe the interaction of Med25 and ETV5, a transcription factor linked to metastasis, in a metastatic breast cancer cell line. Dosing with NA was able to decrease expression of the Med25•ETV5 regulated gene MMP-2, similarly to what was observed with KO of Med25. Inspired by the iterative screening approach used to discover NA, a new small molecule screening method for PPIs was developed, the first method for coactivator targets that directly incorporates selectivity at the primary level. A fluorescence polarization assay that simultaneously monitors multiple activator-coactivator interactions is presented. This method enables assessment of both selectivity and potency of candidate inhibitors in a single screen. A duplex assay containing AcID and CBP KIX has been optimized and a pilot screen has been conducted. The pilot screen was able to categorize compounds as Med25 AcID-selective, CBP KIX-selective, or dual inhibitors. Representative compounds from each subset were evaluated in secondary screening, leading to identification of novel inhibitors of both proteins. Ultimately, this approach is applicable to other coactivator-activator complexes. The final focus of this dissertation is development of a biochemical screen for the human serine protease TMPRSS2, which plays a key role in SARS-CoV-2 viral infection. TMPRSS2 is a difficult protein target because expression and purification is challenging, thus complicating inhibitor screening and in vitro studies. An expression and purification procedure to isolate active TMPRSS2 protease domain from E. Coli was developed. Biochemical methods were used to characterize the mechanism of known inhibitors, and, through integration with computational methodologies, novel TMPRSS2 targeting chemical scaffolds were identified.