Computational assembly of polymorphic amyloid fibrils reveals stable aggregates.

摘要: Amyloid proteins aggregate into polymorphic fibrils that damage tissues of the brain, nerves, and heart. Experimental computational studies have examined structural basis nucleation short fibrils, but ability to predict precisely quantify stability larger aggregates has remained elusive. We established a complete classification fibril shapes developed tool called CreateFibril build such complex, polymorphic, modular structures automatically. applied landscapes, technique we reveal reliable parameters, assess stability. constructed HET-s, Aβ, amylin up 17 nm in length, utilized novel dipolar solvent model captured effect dipole-dipole interactions between water very large molecular systems their aqueous Our results validate experimental data for HET-s suggest (to our knowledge) findings amylin. In particular, predicted correct parameters (rotation angles, packing distances, hydrogen bond lengths, helical pitches) one three predominant protofilaments. structurally characterize all known Aβ including recently classified as wrapped fibrils. Finally, elucidate assert native is more stable than its amyloid form. database models tested, along with energy are available at http://amyloid.cs.mcgill.ca.