Timothy Cardozo
NYU Grossman School of Medicine
Detecting Weak Basal Target–E3 Ligase Interactions for Molecular Glue Discovery
Molecular glues (MGs) bring a target protein into contact with an E3 ubiquitin ligase substrate receptor, allowing ubiquitination and degradation of the target. MG discovery may be enhanced by discovering which of the hundreds of human E3 substrate receptors are capable of binding a given target. We tested whether computational docking can detect weak basal interactions between targets and substrate binding domains (SBDs), which could subsequently be stabilized by MGs. Six high resolution E3–target complexes with known or proposed MG mechanisms were analyzed. The glue molecule was removed and the target was docked back to the SBD. Complexes that produced near native poses in self-docking were then evaluated by cross-docking against a panel of 47 unrelated SBDs to assess specificity. Three of the six complexes generated near native docking poses: DCAF15–RBM39, CRBN–SALL4, and FBW1A(Beta-TrCP)–Beta-catenin, and these three showed better energy scores than the decoys, suggesting that interactions between the ligase receptor and target with weak affinity might be distinguished from noise by docking. Complexes involving disordered peptides or interfaces likely requiring induced conformational changes performed poorly however. These results indicate that docking can detect some weak target–ligase compatibility for certain structural classes and may provide a computational filter to narrow the ligase search space before experimental screening in molecular glue discovery.
Molecular glues (MGs) bring a target protein into contact with an E3 ubiquitin ligase substrate receptor, allowing ubiquitination and degradation of the target. MG discovery may be enhanced by discovering which of the hundreds of human E3 substrate receptors are capable of binding a given target. We tested whether computational docking can detect weak basal interactions between targets and substrate binding domains (SBDs), which could subsequently be stabilized by MGs. Six high resolution E3–target complexes with known or proposed MG mechanisms were analyzed. The glue molecule was removed and the target was docked back to the SBD. Complexes that produced near native poses in self-docking were then evaluated by cross-docking against a panel of 47 unrelated SBDs to assess specificity. Three of the six complexes generated near native docking poses: DCAF15–RBM39, CRBN–SALL4, and FBW1A(Beta-TrCP)–Beta-catenin, and these three showed better energy scores than the decoys, suggesting that interactions between the ligase receptor and target with weak affinity might be distinguished from noise by docking. Complexes involving disordered peptides or interfaces likely requiring induced conformational changes performed poorly however. These results indicate that docking can detect some weak target–ligase compatibility for certain structural classes and may provide a computational filter to narrow the ligase search space before experimental screening in molecular glue discovery.
