Amyloid fibrils are filamentous aggregates of misfolded proteins that, despite their diversity in sequence, embrace a characteristic cross-β sheet architecture1, 2, 3, 4, 5, 6. Amyloidosis is the clinical manifestation of the superficial/intracellular deposition of these amyloid fibrils, vaguely linked to life-threatening neurodegenerative disorders such as Alzheimer’s and Parkinson’s7. Challenges in crystallization of amyloid fibrils served as a bottleneck for structural studies throughout the twentieth century. However, large leaps of progress in various facets of cryo-electron microscopy- such as the development of direct electron detectors and image processing packages for helical reconstruction, now provide an unprecedented possibility of gaining atomic-level structural insights into these enigmatic helical filaments8, 9, 10, 11, 12, 13, 14, 15, 16, 17. High-throughput structure determination with cryo-EM does not only promise to unravel the etiology of amyloid-linked diseases, but also hold the possibility to lead the development of novel therapeutic strategies that leverage the knowledge of structural diversity of amyloid fibrils.

Cryo-EM single particle analysis (SPA) permits three-dimensional reconstruction of proteins and their complexes in a near-native environment18, 19, 20. The protein sample is first plunge-frozen in liquid ethane, suspending the protein in a thin amorphous layer of ice. Generally, thousands of images with many protein particles captured in different orientations are acquired. These images are processed iteratively until cryo-EM maps with complete angular distribution and resolution sufficient for atomic model building are obtained. Helical reconstruction using single particle cryo-EM takes advantage of the helical symmetry present in the protein sample such as amyloid filaments or bacteriophage tails etc., though a careful analysis is required regarding the correct helical parameters and handedness of the cryo-EM map9, 15, 21, 22. Currently, cryo-EM SPA is the main method employed for high resolution helical reconstruction of amyloids. However, the possibility cannot be excluded that isolated/recombinantly purified proteins might not always reflect the native fold or protein interactions. Cryo-electron tomography (cryo-ET) complements cryo-EM SPA by permitting direct imaging of sufficiently thin tissue slices of interest with different tilting angles, thus studying proteins in their native environment23, 24.

This review aims to illuminate the recent advances in our structural understanding of amyloid filaments implicated in neurodegenerative diseases. Amyloid filaments are often polymorphic, i.e., the same primary sequence of amino acids folds into a myriad of higher secondary and tertiary structures, thus changing the entire architecture of helical assembly. The polymorphism of amyloid fibrils is a sophisticated subject posing several critical questions that continue to be explored through numerous recent cryo-EM studies. What factors drive this polymorphism? Do different neurodegenerative diseases have distinct amyloid folds? Is a particular amyloid fold consistent across different samples and regions of the brain for a particular disease? Are the current extraction methods of brain-derived fibrils biased towards certain amyloid folds? Seeded aggregation, which basically is the in vitro amplification of ex vivo amyloid seeds- are these amplified assemblies structurally faithful to the seed structures? Can the disease-associated folds be reproduced through in vitro studies? Furthermore, can cryo-EM be employed to understand and develop amyloid interactions with potential drug candidates or positron emission tomography (PET) ligands for selective labelling in vivo? The scope of this review covers recent advancements towards understanding in part these previously unexplored aspects of pathological amyloids.