Design and validation of nucleosome binding peptides using BindCraft

Abstract

Chromatin is a sophisticated assembly of DNA, RNA and proteins in the nuclei of eukaryotic cells. Protein-protein interactions that govern chromatin function are key regulators of genome activity, making them prime targets for drug discovery. Engineering and modulating these contacts presents a viable pathway for innovative epigenetic therapies. Nucleosomes, the primary structural units of chromatin, feature various interaction interfaces, most notably the acidic patch. This region serves as a critical binding site for numerous chromatin-associated proteins, often via their intrinsically disordered regions. Advances in AI-driven protein design have recently opened new avenues for the rational development of molecules aimed at modulating these chromatin interactions. In this research, we implemented an integrated computational and experimental framework to design and validate peptides targeting the nucleosome acidic patch. We relied on the AI protein design approach implemented in BindCraft to design nucleosome binding peptides. To achieve precise quantitative analysis of these interactions, we refined a donor FRET quenching assay specifically for this application, enabling sensitive and consistent measurement of binding affinities for the designed sequences. We successfully identified peptides with high affinity for the acidic patch on canonical nucleosome core particles. These findings highlight the potential of AI-assisted design for chromatin-targeting agents and establish an effective workflow for their validation, laying the groundwork for synthetic epigenetic modulators and the broader use of protein design in managing 3D genome organization.