I initially developed a large library of small molecules extracted from catalogues of commercially available compounds. Our library design is unique because it is developed with a focus on proteolysis-targeting chimeras. Starting from over 74 million compounds, I prepared a carefully chosen, diverse selection of ~500,000 compounds for computational screening, and we also set up a local library of ~1,000 molecules for use within our research group and the Drug Discovery Unit in Dundee. These fragment libraries were extensively used within and beyond the project. I first screened for compounds binding to the PHD domain of BAZ2A and BAZ2B proteins and identified the first small molecules binding to these domains (Amato et al., ACS Chem. Biol., 2018). Additionally, I carried out modelling studies to unravel the recognition mechanism of the histone tail, where DNA wraps within the nuclei, by the PHD domains at the molecular level (Bortoluzzi et al., Biochem. J., 2017).
Another aim of my project was to implement and apply computational tools to identify binding sites in protein surfaces, and I applied them to discover binding sites in the surface of VHL, an E3 ubiquitin ligase involved in the physiological response to low-oxygen conditions and a common target hijacked by proteolysis-targeting chimeras (Lucas, van Molle, and Ciulli, J. Med. Chem., 2018). The structure-based optimisation of fragments binding to these cavities are an ongoing project within the research group. During the development of the pocket discovery campaign, I also found inspiring evidence of an additional binding site in the E3 ligase. Therefore, I embarked on a structure-based endeavour to design, synthesise, and test cyclic peptides that could interact with this region of the protein. This represented a unique opportunity from a career development perspective, since it provided hands-on experience in many experimental techniques beyond the computational field.
I also performed molecular modelling studies to deepen our understanding of PROTAC design and mode of action. We perturbed binding of a VHL-hijacking PROTAC by substituting specific atoms in its VHL-binding warhead. A first investigation following this strategy revealed the crucial role of a specific amino acid in VHL in ligand recognition (Soares et al., Bioorg. Med. Chem, 2018). Second, we added a fluorine atom to the VHL ligand, and investigated its impact in the behaviour of the small molecule. This is revolutionary, because the project involved creating a novel artificial amino acid that has potential applications in several research areas including protein engineering. In this project, I provided molecular modelling expertise that enabled interpretation of experimental data. Conversion of this fluorinated VHL ligand into a PROTAC generated a very potent degrader of a protein of interest (Testa et al., J. Am. Chem. Soc., 2018).
Finally, during my Action our group also obtained the first crystal structure of a PROTAC bound simultaneously to its two target proteins: the protein of interest and the E3 ubiquitin ligase. I performed extensive molecular dynamics simulations to study and manipulate the behaviour of the ternary complex in solution (partly published in Gadd, Testa, Lucas et al., Nat. Chem. Biol., 2017). The crystal structure further enabled the structure-based design of more selective PROTACs, which will be published soon. I further co-authored a review on E3 ubiquitin ligases and selectivity of substrate recognition (Lucas and Ciulli, Curr. Opin. Struct. Biol., 2017).
