What Does Evolution Tell Us About The Structure Of A Functional Amyloid Protein?
pengfei tian,wouter boomsma,yong wang,daniel e otzen,mogens h jensen,kresten lindorfflarsen
DOI: https://doi.org/10.1016/j.bpj.2014.11.1253
IF: 3.4
2015-01-01
Biophysical Journal
Abstract:Amyloids are insoluble fibrillar protein aggregates. While they are commonly found in human diseases, it is becoming increasingly clear that this type of structure is essential for a range of biological functions, with prominent examples in organisms ranging from bacteria to human. Functional amyloid and pathogenic amyloid share similar physical and chemical properties. Unlike pathological amyloids, however, the structures of functional amyloids are formed by polypeptide sequences whose amyloid structure has been under a positive evolutionary selection pressure. This important distinction provides us with an opportunity to obtain structural insights from an unexpected source: the covariation of amino acids among sequences within the same family of a functional amyloid protein. There is a long history for the idea of using coevolution for molecular structure prediction, but recent growth in sequence databases and new, efficient algorithms to disentangle indirect couplings in a network, have dramatically improved our ability to predict residue-residue contacts. We used recently developed sequence analysis methods (EVcoupling, PSICOV and GREMLIN) to extract distance restraints from a multiple sequence alignment of a functional amyloid protein. Together with an efficient force field, these restraints allow us to determine atomic resolution structural models. We find that the protein forms a beta-helical structure, where each turn corresponds to previously identified repeat sequences. The proposed structure is validated by previously published solid-state NMR, electron microscopy and X-ray diffraction data, and confirms an earlier proposed model derived by complementary means. To our knowledge, this is the first time the analysis of correlated mutations and computer simulations have been used together to study the structure of a functional amyloid. The current study therefore serves as a probe into the potential applicability of the approach in this domain.