Biomolecular recognition and protein dynamics in amyloid aggregation

The deposition of proteinaceous aggregates known as amyloid fibrils is related to the onset of numerous human pathologies termed protein folding diseases. Even though the amyloid fold is proposed to lie in an energy minimum, the process of amyloid aggregation is remarkably slow both in vivo and in vitro. This arises from the necessity to overcome high energy barriers that lead to the formation of intermolecular interactions that promote further polymerisation and/or the generation of aggregation-prone intermediate species. Structural characterisation of protein association in the early stages of amyloid assembly and most importantly, its link to protein dynamics and/or conformational changes remains poorly characterised due to the heterogeneity and the transient nature of the interactions involved. In this thesis, using NMR techniques sensitive enough to detect species that are lowly populated in solution, the interaction surfaces that lead to inhibition, promotion or nucleation of amyloid assembly by the protein β2-microglobulin (β2m) were mapped in atomic detail. As visualised by paramagnetic relaxation enhancement and analysis of chemical shift perturbation data using human β2m and its murine homologue, inhibition and promotion of assembly involve similar interfaces but the interactions differ in details sufficient to determine the fate of polymerisation. Inhibition occurs by the formation of specific, kinetically trapped off-pathway species, the accumulation of which results in prolongation of the onset of amyloid, while promotion involves more diffuse, weaker protein-protein interactions that cause destabilisation/conformational changes in both interacting partners. By contrast, nucleation progresses through a more complicated multiple-site binding, reminiscent of oligomer formation. Overall, the results reveal an atomic level description of the fine interplay between structure, affinity and dynamics that dictates the course of amyloid formation by proteins that are very similar both in terms of sequence and structure and provide a general framework of the mechanism by which β2m (hetero) polymerises.