Development and optimization of a lean computational pipeline for prediction of PROTACs ternary complexes

PROTACs are heterobifunctional molecules able to induce the proteasomal degradation of a target protein by keeping it near to the cell ubiquitination machinery. In particular, they simultaneously bind the protein related to the diseases that has to be degraded (the Protein of Interest, POI) with a member of E3 ubiquitin ligase forming a ternary complex (TC). The Targeted Protein Degradation induced by PROTACs is one of the most exciting new modalities of the last 20 years, and they are gradually passing from fascinating molecular biology tools towards clinical settings. ARV-110, ARV-471 and KT-474 are the most relevant examples of PROTACs already in clinical trials. In principle, Targeted Protein Degradation can tackle a wide range of diseases, from oncology to psychiatric disorders passing through immune dysfunctionalities and rare diseases. The aim of this thesis is the development of a lean computational pipeline able to predict the geometry and the stability of the TC. The result was a reliable and intuitive method for modelling CT, suitable for inclusion in a drug discovery pipeline to prioritize compounds to be synthesized and subsequently submitted to physic-chemical and in-vitro ADME characterization. Firstly, the method was extensive validated against experimentally determined structures. Then it was successfully used to capture degradation cliffs, establishing intra-series scoring and susceptibility to protein mutations. Moreover, the implementation of explicit water molecules in ternary complex models has revealed unexplored facets of PROTACs ternary complexes. In conclusion the proposed method enables structure-based rational design of PROTACs providing a suitable platform for speeding up the discovery of novel active PROTACs.