WD40 repeats (or WDR domains) are some of the most abundant protein interaction domains in the human proteome, representing essential subunits of multiprotein complexes involved in a wide variety of cellular processes, ranging from cell signaling, growth, and division to protein degradation. The domain has a beta-propeller structure with an overall doughnut shape. The central pore of the domain is typically involved in interactions with peptide regions of key interaction partner proteins. While these pores of some WDR domains could be druggable pockets, often they can be challenging to target with small drug-like molecules due to their relatively shallow and solvent exposed nature. Here, using the AtomNet® model, a deep convolutional neural network for structure-based drug discovery, we screened billions of compounds against WDR domain-containing proteins, implicated in various illnesses such as heart disease and cancer. We identified several small drug-like molecules exhibiting micromolar binding affinities validated by differential scanning fluorimetry and surface plasmon resonance experiments. This effort will support SGC’s Target 2035 initiative to discover molecular probes for nearly all human proteins.

Discovery of inhibitors of WDR domain-containing proteins using a deep convolutional neural network

ACS Spring Meeting 2021

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.