Discovery of small molecules that modulate transient protein-protein interactions leading to amyloid formation

Protein-protein interactions control and coordinate a wide range of biological processes, but can also occur aberrantly and with toxic consequences. In amyloid diseases, exposure of new interaction interfaces on protein surfaces (e.g. as a consequence of anomalous post-translational modifications, mutations, or proteolytic cleavage) leads to aberrant self-assembly pathways which result in formation of fibrillar protein aggregates. Many of the interactions which drive self-assembly are transient and associated with highly heterogeneous populations of rapidly interconverting species, some of which can contribute towards toxicity by undergoing further aberrant interactions with other biological molecules. Amyloid disease is associated with over 50 distinct human proteins, spanning prevalent (e.g. Alzheimer’s disease, Parkinson’s disease) and rarer disorders (e.g. light chain amyloidosis and dialysis-related amyloidosis). However, despite the pervasiveness of some amyloid diseases, the molecular mechanisms of fibril assembly and the structures of assembly intermediates have remained largely elusive, as the heterogeneity and dynamics of these aggregating systems has hindered structural studies and drug development. Small molecules which are capable of perturbing protein-protein interactions in specific ways are valuable tools for interrogating transient association events. Developing such chemical tools for amyloid proteins could therefore provide powerful routes to facilitate the study of these challenging systems. In particular, the identification of molecules which stabilise specific oligomeric species offers opportunities for the structural and functional characterisation of key assembly intermediates. In this thesis, a chemical tool approach has been applied to study the aggregation of an amyloidogenic protein, β2-microglobulin (β2m), concentrating specifically on a truncation variant known as ΔN6, which lacks the N-terminal six amino acids and has increased amyloidogenicity. Due to the dynamic nature of this protein and its relatively featureless surface, focus was placed on the development of covalent ligands, to ensure that the resulting molecules would have the specificity and affinity required to be useful chemical tools. Based on existing structural models of ΔN6 dimers and higher order oligomers, and in combination with computational chemistry methods, two target sites were selected for ligand development. A rationally-designed library of 84 fragments was synthesised and screened against ΔN6 using a covalent, site-directed method called “tethering”. Screening revealed several promising leads, which were subsequently shown to be capable of perturbing the initial stages of ΔN6 self-assembly, driving formation of trapped tetrameric species with reduced amyloidogenicity. These molecules were used in combination with NMR to investigate the structure of tetrameric ΔN6. The results provide new insights into ΔN6 assembly mechanisms and, more generally, highlight the potential of this covalent screening approach to develop new chemical probes for the study of amyloid proteins.