Polymorphism and assembly dynamics of amylin fibrils

Abstract

Type-2 diabetes is associated with amyloid deposits of fibrillar amylin in the pancreatic islets of Langerhans. Fibrillogenesis co-occurs with death of the islet β-cells and may be a causative factor. In order to gain a deeper understanding of amylin fibril formation an in vitro model system was developed. Synthetic human amylin (hA) was used to investigate the structure, polymorphism and assembly dynamics of these fibrils. Synthetic hA spontaneously assembled into 5-nm wide protofibrils which further associated into higher order fibrils either by lateral (side-by-side) association or by coiling around each other. Circular dichroism spectroscopy showed that monomeric hA existed primarily in random-coil conformation. Upon aggregation into fibrils, the peptide conformation changed to a mixture of β-sheet and α-helical structure. In a new application of time-lapse atomic force microscopy, the growth dynamics of individual fibrils was investigated. Growing fibrils adsorbed to a mica surface were visualised at high resolution over time. This established that hA protofibrils grew bi-directionally in a non-polar fashion, consistent with current models of amyloid formation which postulate protofibrils as continuous β-sheets. Bai (1999) established that in contrast to full length hA which is cytotoxic to islet β-cells in vitro, its truncated counterpart hA(8-37) is inert. Therefore, a detailed comparison of fibril formation of truncated versus full length peptide was carried out. For full-length hA, the predominant fibril type formed in vitro was composed of three protofibrils. ln the case of truncated hA(8-37), the major fibril species was composed of two protofibrils. The kinetics of fibril growth from both peptides were monitored using the thioflavin-T binding assay, and results were strongly indicative for distinct mechanisms of fibrillogenesis. However, the observation that hA(8-37) formed fibrils which were similar to full length hA also indicated that amyloid-peptide induced cell death is more complex than simply "fibril-mediated". The model system developed in this thesis has the potential for further investigation of these complexities, thus bringing us closer to the goal of being able to control the growth and nature of amyloid deposits in vivo and improved ways of disease management.