Ligandable pockets for different E3 ligases predicted by mixed solvent molecular dynamics

E3 ligases are crucial components of the ubiquitin-proteasome system (UPS) that confer substrate specificity for proteasomal degradation. E3 ligases have been linked to various pathogenic mechanisms, and for this reason, these protein complexes have emerged as attractive drug target candidates for precise therapeutic intervention. Although there is a lack of clear-cut design strategies, a variety of potential activities can be envisioned from directly pharmacologically modulating E3 ligases, such as promoting a higher rate of (neo)substrate degradation or, in contrast, inhibiting their degradation. However, despite the tremendous clinical significance of this large protein family, only a single E3 ligase has FDA approved drugs.

To expand the current approaches, and to increase the arsenal of targeted E3 ligases, we present an efficient E3 ligase-based computational approach, where we used the PDB structure of a diverse set of 23 E3 ligases to identify and characterize their pockets. We encourage the scientific community to use all the ligandable pockets identified in thes work, to facilitate the development of small molecules targeting new, uncharted, E3 ligases.

The prediction of pockets has been performed using mixed-solvent molecular dynamics, using the pyMDMix setup. We employed a mixture of ethanol:water at 1:4 concentration. These systems where simulated for 3 replicas of 50ns each in AMBER, and the joint occupancy of the ethanol converted into energy grids following the standard pyMDMix protocol. Clusterization of the energy grids led to the energy hotspots (saved as ETA_hotspots.pdb in the subdirectory of each E3). Grouping those hotspots by surface distance, coupled by volume, balance between hydrophobic and polar hotspots led to the prediction of ligandable pockets. All pockets, ligandable and not, are represented in the pockets.pdb files in each subdirectory corresponding to their E3 ligase.

A compiled information about the pockets (energy efficiency, location, volume), coupled with their ligandablility prediction is detailed in predicted_pockets_information.xlsx. Finally, a pre-generated visualization session is stored for the ease of analyzing the pockets.