Fibril modelling by sequence and structure conservation analysis combined with protein docking techniques: β2-microglobulin amyloidosis

Obtaining atomic resolution structural models of amyloid fibrils is currently impossible, yet crucial for our understanding of the amyloid mechanism. Different pathways in the transformation of a native globular domain to an amyloid fibril invariably involve domain destabilization. Hence, locating the unstable segments of a domain is important for understanding its amyloidogenic transformation and possibly control it. Since relative conservation is suggested to relate to local stability [H. Benyamini, K. Gunasekaran, H. Wolfson, R. Nussinov, Conservation and amyloid formation: a study of the gelsolin-like family, Proteins 51 (2003) 266-282. [24]], we performed an extensive, sequence and structure conservation analysis of the β2-microglobulin (β2-m) domain. Our dataset include 51 high resolution structures belonging to the “C1 set domain” family and 132 clustered PSI-BLAST search results. Segments of the β2-m domain corresponding to strands A (residues 12-18), D (45-55) and G (91-95) were found to be less conserved and stable, while the central strands B (residues 22-28), C (36-41), E (62-70) and F (78-83) were found conserved and stable. Our findings are supported by accumulating observations from various experimental methods, including urea denaturation, limited proteolysis, H/D exchange and structure determination by both NMR and X-ray crystallography. We used our conservation findings together with experimental literature information to suggest a structural model for the polymerized unit of β2-m. Pairwise protein docking and subsequent monomer stacking in the same manner suggest a fibril model consistent with the cross-β structure.