The Synergy of Synthesis, Computation and NMR Spectroscopy to Design Conformationally Controlled α-Helix Mimetics

Abstract

In the search for high affinity therapeutics, medicinal chemists are turning their attention to free ligand conformations as it is well known that if a large conformational reorganisation is required upon binding, then a weaker binding affinity will be observed. This has been highlighted as an important strategy in
the inhibition of protein–protein interactions (PPIs), due to the large, flat and featureless binding sites they present. This thesis discusses the design, synthesis and conformational analysis of a novel class of conformationally controlled α-helix mimetics, aimed at targeting aberrant PPIs. Their ability to target
the p53-Mdm2 PPI with reasonable affinity is demonstrated.

Molecular Mechanics (MM) conformational search calculations were used to design a conformationally controlled scaffold that mimicked the i, i + 3/4 and i + 7 positions of the α-helix. Following the design of a general scaffold, appropriate side chains were added to mimic those of p53, and protein-ligand molecular docking calculations were performed to assess the binding of the designed ligand to Mdm2.
Molecular docking was also used to design a library of p53 mimetics with binding affinities comparable to that of Nutlin-2, a highly successful p53-Mdm2 inhibitor.

With promising p53 mimetics designed using computation, their synthesis was explored using lithiation–borylation. Using lithiation–borylation a diverse set of side chains can be incorporated onto the scaffold by the insertion of the correct benzoate ester into the synthetic sequence. Thus, broad libraries of α-helix mimetics bearing different side chains can be synthesised using the same iterative methodology.

To confirm that the designed p53 mimetics exhibit the conformational bias that was predicted by MM calculations, a hybrid quantum mechanics (QM) and NMR spectroscopy approach was used to explore the conformation of one of the p53 mimetics. Experimentally measured scalar coupling constants, interproton distances derived from 1D-NOESY spectroscopy and both 1H and 13C chemical shifts were compared to the Boltzmann averaged values, calculated using DFT calculations. An excellent correlation was observed between the two data sets, confirming the predicted conformational bias. Finally, the binding of the p53 mimetics to Mdm2 was explored using 1H15N HSQC spectroscopy. The
chemical shift perturbations (CSPs) that were observed upon binding of the ligands were mapped to amino acids located primarily in, or on the periphery of, the p53 binding pocket on Mdm2. The CSPs were tracked with increasing concentrations of ligand to obtain dissociation constants (Kd). Two
different p53 mimetics were tested and binding affinities as low as 8.8 µM were obtained. Additionally, the binding of two control molecules was assessed using 1H15N HSQC spectroscopy, to confirm the importance of both the hot-spot groups and conformational control in obtaining high affinity therapeutics.

Date of Award	12 May 2020
Original language	English
Awarding Institution	The University of Bristol
Supervisor	Adrian J Mulholland (Supervisor), Craig P Butts (Supervisor) & Varinder K Aggarwal (Supervisor)