Generating Ensembles of Dynamic Misfolding Proteins

Area of Science:

• Biochemistry
• Structural Biology
• Computational Biology

Background:

• Protein misfolding and aggregation are central to neurodegenerative diseases.
• Early stages involve dynamic, disordered protein conformers.
• Visualizing these transient species at atomic resolution is challenging.

Purpose of the Study:

• To review experimental and computational methods for studying early protein misfolding dynamics.
• To discuss challenges in characterizing rapidly interconverting molecular species.
• To explore machine learning applications for identifying aggregation-relevant protein conformers.

Main Methods:

• Review of experimental techniques (e.g., NMR, cryo-EM) for capturing protein dynamics.
• Computational approaches including molecular dynamics simulations.
• Application of machine learning algorithms to analyze heterogeneous protein ensembles.

Main Results:

• Dynamic disorder and large-scale motions characterize early misfolding species.
• Experimental restraints often average over millisecond-timescale motions, complicating analysis.
• Machine learning can extract specific, aggregation-prone sub-ensembles from complex data.

Conclusions:

• Understanding dynamic protein conformers is crucial for disease mechanisms.
• Machine learning offers a promising avenue for dissecting aggregation pathways.
• Specific interactions within dynamic ensembles of alpha-synuclein and beta2-microglobulin were investigated using this approach.