Global Amyloid Mapping: Solving Amyloid Nucleation by Deep Mutagenesis

Introduction

Amyloid fibrils form and precipitate in more than 50 incurable human diseases, including Alzheimer’s and Parkinson’s disease. All proteins may be able to form amyloids, at least under certain circumstances. However, aggregation is actually rare as the process of amyloid formation is controlled by a high kinetic barrier: protein sequences have to cross a free energy barrier to nucleate transition states which then seed irreversible fibril formation.

Challenges in Studying Amyloid Formation

Short-lived transition states are extremely challenging to study using biophysical methods. We have recently developed a massively parallel genomics approach to quantify the rate of aggregation of thousands of protein sequences. We also have evidence that, by quantifying the interactions between mutations, we can capture the key interactions between residues in the transition state.

Project Aims

Here, we will unleash the potential of this approach by targeting the following aims:

1. Map the impact on amyloid nucleation of all possible mutations in >60 human and functional amyloids - generating reference atlases for clinical variant interpretation.
2. Build an energetic and structural model of the transition state of disease-associated amyloids.
3. Uncover sequences across the genome that nucleate amyloids in response to environmental stress.

Expected Outcomes

This project will uncover the rules required to understand, predict, and engineer amyloid formation. By identifying the mutations that do and do not lead to amyloid formation, we will reveal which variants accelerate aggregation and cause disease.

Significance of the Research

We will also address one of the most important questions in the field, i.e., identifying the interactions established in the transition states of amyloid nucleation, guiding the development of therapeutic approaches in amyloid diseases, including the worst kinds of dementia.

Future Applications

Finally, our results will feed a new generation of models and predictors of protein aggregation to employ for disease variant interpretation and the synthetic design of novel proteins.