UC Berkeley

Once regarded as a precarious strategy, covalency is now a critical aspect of modern drug discovery. Covalent drugs, which possess a reactive functional group, bind to proteins by reacting with an orthogonal functional group present in a protein; covalent modification of a protein can perform better than their noncovalent binding in liganding intrinsically disordered proteins or those without deep binding pockets, such as needed in stabilizing molecular glue interactions for targeted protein degradation. Also, coupled with chemoproteomic platforms, covalent modification with a drug or probe can greatly simplify mechanism elucidation. This dissertation presents examples of applying covalent strategies to the discovery of molecular glue degraders for targeted protein degradation (TPD), and development and deployment of a class of photoaffinity labeling (PAL) probes for target engagement.In chapter II of this work, we sought to employ covalency to discover new molecular glue degraders which could stabilize weak, neomorphic interactions between a target protein and an E3 ubiquitin ligase. As proof of concept, we built an elaborated cysteine-reactive electrophile library based on JQ1, an inhibitor of the well-established cancer target BRD4, hypothesizing that these compounds could stabilize a PPI between BRD4 and an unknown E3 ligase. We screened these molecules in cells for ability to deplete BRD4 levels, and we identified a promising compound, HRG038, which induces proteosome-dependent degradation of BRD4. Using chemoproteomic and genetic approaches, we identified that this molecule covalently binds the E3 ubiquitin ligase CUL4DCAF16 through C173. We then transposed the electrophilic fragment, or “handle” from HRG038, onto other protein-targeting ligands, converting the ligands into degraders with limited success. Upon optimization of the handle, we achieved degradation of a widened scope of targets, and found that the new fragment was dependent on both C173 and C178, indicating a more plastic binding mode to accommodate more diverse ternary complexes.In chapter III of this work, we deployed an alternative PAL strategy, acyl silane-based probes, for use in live-cell target identification. Building on initial proof-of-concept work from the Toste and Nomura groups that acyl silanes labeled lysates, we demonstrated that acyl silane-based JQ1 probes can be used in live-cell chemoproteomic workflows to specifically label the target, BRD4. We compared our probes to the current PAL state of the art, minimalist diazirine probes, and found that acyl silanes were able to more specifically enrich BRD4 from cells. Also, we were able to tune the reactivity of acyl silanes by substitution about the silicon atom, improving the selectivity of accurate target labeling.