Prediction of protein aggregation and amyloid formation

Abstract

Protein aggregation accounts for the onset of more than 40 human disorders, including neurodegenerative diseases like Alzheimer's and Parkinson's but also non-neuropathic pathologies like Diabetes type II or some types of cancers. In all these diseases, the toxic effect is associated with the self-assembly of proteins into insoluble amyloid fibrils displaying a common regular cross-b structure. Surprisingly, cells also exploit the amyloid fold for important physiological processes, from structure scaffolding to heritable information transmission. In addition, protein aggregation often occurs during the recombinant production and downstream processing of therapeutic proteins, becoming the main bottleneck in the marketing of these drugs. In this context, approaches aiming to predict the aggregation and amyloid formation propensities of proteins are receiving increasing interest, both because they can lead us to the development of novel therapeutic strategies and because they are providing us with a global understanding of the role of protein aggregation in physiological and pathological processes. Here we illustrate how our present understanding of the physico-chemical and structural basis of protein aggregation has crystalized in the development of algorithms able to forecast the aggregation properties of proteins both from their primary and tertiary structures. A detailed description of these computational approaches and their application is provided.