Improved Estimates of Folding Stabilities and Kinetics with Multiensemble Markov Models

Area of Science:

• Computational Biology
• Biophysics
• Protein Dynamics

Background:

• Markov State Models (MSMs) are used to study protein folding dynamics from molecular simulations.
• Standard MSMs assume equilibrium sampling, which is often not met in ab initio folding studies, leading to inaccurate stability predictions.
• This limitation hinders the accurate estimation of protein folding thermodynamics and kinetics.

Purpose of the Study:

• To develop an enhanced-sampling protocol for improving the accuracy of MSMs in predicting protein folding stability.
• To enable reliable estimation of folding pathways, rates, and stabilities from non-equilibrium simulation data.
• To validate the protocol using a designed helical hairpin (Z34C) and compare with experimental data.

Main Methods:

• Performed unbiased folding simulations and used sparse time-lagged independent component analysis (tICA) for feature extraction.
• Employed umbrella sampling along the identified reaction coordinate to observe folding and unfolding transitions.
• Utilized multiensemble Markov models (MEMMs) for estimating folding thermodynamics and kinetics.

Main Results:

• The developed protocol accurately predicted folding pathways, rates, and stabilities for the Z34C helical hairpin.
• Results showed good agreement between simulation-based predictions and experimental measurements.
• Demonstrated the capability of the protocol to overcome limitations of traditional MSMs with non-equilibrium data.

Conclusions:

• Accurate simulation-based prediction of absolute protein folding stabilities is achievable with the proposed enhanced-sampling protocol.
• This methodology has significant implications for the computational design of stable miniproteins and peptidomimetics.
• The protocol offers a robust approach for studying protein folding mechanisms and thermodynamics.