Abstract The discovery that disordered proteins are widespread in the human proteome has prompted the quest for methods to characterize the conformational properties that determine their functional and dysfunctional behaviour. It has become customary to describe these proteins in terms of structural ensembles and free energy landscapes, which offer conformational and thermodynamic insight. However, a current major challenge is to generalize this description to ‘kinetic ensembles’, thereby also providing information on transition rates between states. Approaches based on the theory of stochastic processes can be particularly suitable for this purpose. Here, we develop a Markov state model and illustrate its application by determining a kinetic ensemble of the 42-residue form of the amyloid-β peptide (Aβ42), whose aggregation is associated with Alzheimer’s disease. Using the Google Compute Engine, we generated 315 μs all-atom, explicit solvent molecular dynamics trajectories, validated with experimental data from nuclear magnetic resonance spectroscopy. Using a probabilistic-based definition of conformational states in a neural network approach, we found that Aβ42 is characterized by inter-state transitions no longer than the microsecond timescale, exhibiting only fully unfolded or short-lived, partially-folded states. We contextualize our findings by performing additional simulations of the oxidized form of Aβ42. Our results illustrate how the use of kinetic ensembles offers an effective means to provide information about the structure, thermodynamics, and kinetics of disordered proteins towards an understanding of these ubiquitous biomolecules.

A kinetic ensemble of the Alzheimer’s Aβ peptide