A proposed structural model for amyloid fibril elongation: Domain swapping forms an interdigitating β-structure polymer

We propose a model illustrating how proteins, which differ in their overall sequences and structures, can form the propagating, twisted β-sheet conformations, characteristic of amyloids. Some cases of amyloid formation can be explained through a 'domain swapping' event, where the swapped segment is either a β-hairpin or an unstable conformation which can partially unfold and assume a β-hairpin structure. As in domain swapping, here the swapped β-hairpin is at the edge of the structure, has few (if any) salt bridges and hydrogen bonds connecting it to the remainder of the structure and variable extents of buried non-polar surface areas. Additionally, in both cases the swapped piece constitutes a transient 'building block' of the structure, with a high population time. Whereas in domain swapping the swapped fragment has been shown to be an α-helix, loop, strand or an entire domain, but so far not a β-hairpin, despite the large number of cases in which it was already detected, here swapping may involve such a structural motif. We show how the swapping of β-hairpins would form an interdigitated, twisted β-sheet conformation, explaining the remarkable high stability of the protofibril in vitro. Such a swapping mechanism is attractive as it involves a universal mechanism in proteins, critical for their function, namely hinge-bending motions. Our proposal is consistent with structural superpositioning of mutational variants. While the overall r.m.s.d.s of the wild-type and mutants are small, the proposed hinge-bending region consistently shows larger deviations. These larger deviations illustrate that this region is more prone to respond to the mutational changes, regardless of their location in the sequence or in the structure. Nevertheless, above all, we stress that this proposition is hypothetical, since it is based on assumptions lacking definitive experimental support.