Identification of the core structure of lysozyme amyloid fibrils by proteolysis

Human lysozyme variants form amyloid fibrils in individuals suffering from a familial non-neuropathic systemic amyloidosis. In vitro, wild-type human and hen lysozyme, and the amyloidogenic mutants can be induced to form amyloid fibrils when incubated under appropriate conditions. In this study, fibrils of wild-type human lysozyme formed at low pH have been analyzed by a combination of limited proteolysis and Fourier-transform infrared (FTIR) spectroscopy, in order to map conformational features of the 130 residue chain of lysozyme when embedded in the amyloid aggregates. After digestion with pepsin at low pH, the lysozyme fibrils were found to be composed primarily of N and C-terminally truncated protein species encompassing residues 26-123 and 32-108, although a significant minority of molecules was found to be completely resistant to proteolysis under these conditions. FTIR spectra provide evidence that lysozyme fibrils contain extensive beta-sheet structure and a substantial element of non beta-sheet or random structure that is reduced significantly in the fibrils after digestion. The sequence 32-108 includes the beta-sheet and helix C of the native protein, previously found to be prone to unfold locally in human lysozyme and its pathogenic variants. Moreover, this core structure of the lysozyme fibrils encompasses the highly aggregation-prone region of the sequence recently identified in hen lysozyme. The present proteolytic data indicate that the region of the lysozyme molecule that unfolds and aggregates most readily corresponds to the most highly protease-resistant and thus highly structured region of the majority of mature amyloid fibrils. Overall, the data show that amyloid formation does not require the participation of the entire lysozyme chain. The majority of amyloid fibrils formed from lysozyme under the conditions used here contain a core structure involving some 50% of the polypeptide chain that is flanked by proteolytically accessible N and C-terminal regions