Small molecule induced protein dimerization is a long sought-after goal in chemical biology and drug discovery. Recent progress, especially in the development of PROTACs, demonstrates that functional ligand induced dimerization of two proteins can be achieved de novo. PROTACs are heterobifunctional small molecules, which recruit an E3 ligase and target it to a protein of interest. In turn, the target is subject to proximity mediated ubiquitination and subsequent degradation. PROTACs offer great promise for otherwise undruggable therapeutic targets, such as transcription factors, or multi-protein complexes. We and others have shown that PROTACs can achieve sub-nanomolar efficacy in vitro and in vivo, however, the underlying molecular mechanism remains poorly understood.
Here we utilize comprehensive characterization of the ligand dependent interaction between the CUL4-RBX1-DDB1-CRBN (CRL4CRBN) ubiquitin ligase, and the BET family protein BRD4. We demonstrate that binding between these proteins is unexpectedly plastic. Through multiple X-ray crystal structures in conjunction with biochemical and biophysical characterization, we show that plastic protein-protein contacts between ligase and substrate result in several distinct low energy binding conformations, which are selectively bound by ligands. We demonstrate that computational protein-protein docking can predict the energy landscape of these plastic inter-protein contacts, and inform the design of heterobifunctional degraders. Utilizing this computational approach, we synthesize the first BRD4 selective degrader that can discriminate between highly homologous BET bromodomains. We provide structural data to demonstrate that selectivity is conferred by a single amino acid difference between the bromodomains of BRD4 and BRD2/3/T. Our findings that plastic inter-protein contacts confer selectivity for ligand-induced protein dimerization provide a conceptual framework for the development of heterobifunctional ligands, which will likely apply to systems beyond protein degradation such as allosteric regulators or protein-dimerization, and our computational methods will inform future design of small molecule dimerizers.