Abstract
AA amyloidosis is one of the most prevalent forms of systemic amyloidosis and affects both humans and other vertebrates. In this study, we compare MAS solid-state NMR data with a recent cryo-EM study of fibrils involving full-length murine SAA1.1. We address the question whether the specific requirements for the reconstitution of an amyloid fibril structure by cryo-EM can potentially yield a bias towards a particular fibril polymorph. We employ fibril seeds extracted from in to vivo material to imprint the fibril structure onto the biochemically produced protein. Sequential assignments yield the secondary structure elements in the fibril state. Long-range DARR and PAR experiments confirm largely the topology observed in the ex-vivo cryo-EM study. We find that the β-sheets identified in the NMR experiments are similar to the β-sheets found in the cryo-EM study, with the exception of amino acids 33-42. These residues cannot be assigned by solid-state NMR, while they adopt a stable β-sheet in the cryo-EM structure. We suggest that the differences between MAS solid-state NMR and cryo-EM data are a consequence of a second conformer involving residues 33-42. Moreover, we were able to characterize the dynamic C-terminal tail of SAA in the fibril state. The C-terminus is flexible, remains detached from the fibrils, and does not affect the SAA fibril structure as confirmed further by molecular dynamics simulations. As the C-terminus can potentially interact with other cellular components, binding to cellular targets can affect its accessibility for protease digestion.