Abstract
Molecular glues are small molecules that stabilize protein-protein interactions, enabling new molecular pharmacologies, such as targeted protein degradation. They offer advantages over proteolysis targeting chimeras (PROTACs), which present challenges associated with the size and properties of heterobifunctional constructions, but glues lack the rational design principles analogous to PROTACs. One notable exception is the ability to alter the structure of Cereblon (CRBN)-based molecular glues and redirect their activity toward new neo-substrate proteins. We took a focused approach toward modifying the CRBN ligand, 5'-amino lenalidomide, to alter its neo-substrate specificity using high-throughput chemical diversification by parallelized sulfur(VI)-fluoride exchange (SuFEx) transformations. We synthesized over 3,000 analogs of 5'-amino lenalidomide using this approach and screened the crude products using a phenotypic screen for cell viability, identifying dozens of analogs with differentiated activity. We characterized four compounds that degrade G-to-S phase transition 1 (GSPT1) protein, providing a proof-of-concept model for SuFEx-based discovery of CRBN molecular glues.