Proteolysis-Targeting Chimeras (PROTACs) eliminate disease-relevant proteins by stabilizing short-lived, metastable protein–protein interactions (PPIs) between a target protein and an E3 ligase. These transient encounter complexes are central to PROTAC activity, yet they are difficult to predict and are often invisible to structure-based or AI-only modeling approaches that favor stable, native interactions. Here, we show that explicitly accounting for metastable PPIs provides an effective design principle for rational PROTAC development. Using large-scale molecular dynamics (MD) simulations combined with integrative generalized master equation (IGME) modeling, we mapped the ensemble of metastable PPIs between receptor-interacting serine/threonine-protein kinase 1 (RIPK1) and the von Hippel-Lindau (VHL) E3 ligase, and identified four distinct PPIs suitable for degrader engagement. We find that different PPIs preferentially accommodate different linker architectures, providing a mechanistic explanation for the counterintuitive experimental observation that PROTACs with very short and very long linkers can consistently yield potent degradation. Guided by this insight, we established a virtual screening workflow for linker design and identified the benzylic position on the VHL ligand as a viable linkage site. Rationally designed novel PROTACs featuring this site achieve complete degradation with subnanomolar to low-nanomolar potency. Our results demonstrate that incorporating protein dynamics into degrader design substantially expands the accessible PROTAC design space.
Wu et al. (2026) studied this question.