Quantitative feature extraction reveals the status quo of protein fibrillation in the cell nucleus

Stepwise fibrillation of otherwise soluble proteins to insoluble amyloid-like protein aggregates is a hallmark of neurodegenerative protein-misfolding diseases including Alzheimer's, polyglutamine diseases, and the prion encephalopathies. Investigation of protein aggregation mechanisms has considerably advanced in vitro due to recent technical innovation, whereas the development of analyses tools for intracellular protein fibrillation remains a major challenge. Here, we introduce a method that enables monitoring of the protein fibrillation status in the cell nucleus. We show that the amyloid indicator Congo red can be induced to bind to distinct nucleoplasmic microdomains that are describable by application of discrete mathematics on the image information. Since formation of Congo red-binding nuclear microdomains (CRBDs) correlates with increased amyloid formation and decreased solubility of endogenous proteins with homopolymeric polyglutamine (polyQ) stretches we introduce the idea that different protein fibrillation steps can be characterized intracellularly by graph theory-aided pattern recognition.