Bryan Dickinson

Cellular regulation relies on precisely controlled interactions between proteins. A powerful strategy in chemical biology and drug discovery involves molecules that bring a target protein into close proximity with a regulator—an approach known as induced proximity. By driving the formation of specific protein complexes, induced proximity can trigger diverse outcomes, including targeted protein degradation, posttranslational modifications, altered localization, or immune recognition. There are two major classes of induced proximity molecules: bifunctional molecules and molecular glues. Bifunctional molecules, which independently bind two targets to bring them together, are relatively easy to design but are often limited by the “hook effect”, which reduces their effectiveness in cells. Molecular glues avoid this issue by binding two targets cooperatively, but their discovery is difficult because they depend on pre-existing structural complementarity or weak interactions between targets, restricting their broader use. As a result, the full potential of induced proximity remains constrained.

In this talk, I will describe our efforts to overcome these limitations by developing an evolution-based pipeline for discovering peptide-based molecular glues, built over the past decade. Our approach integrates two technologies: PANCS-binders (for rapid discovery of binary interactions) and PANCS-glues (for identifying ternary interactions). Using this pipeline, we discovered a model peptide glue, validated its unique cooperativity mechanism, and engineered it into a mammalian synthetic biology tool. We then applied PANCS-glues to create eight de novo glues, yielding new synthetic biology tools, protein degraders, and inhibitors. Importantly, this platform led us to discover a new class of “disordered glues,” which exploit disorder-to-order transitions to achieve cooperativity across two binding interfaces. This mechanism, combined with the PANCS-binder/PANCS-glues workflow, provides a versatile framework for designing cooperative molecular glues for virtually any protein pair of interest—while also demonstrating how protein disorder can be harnessed as a powerful design principle.