Improved Sampling of Adaptive Path Collective Variables by Stabilized Extended-System Dynamics

Area of Science:

• Computational Chemistry
• Biocatalysis
• Enzyme Mechanisms

Background:

• Biocatalytic reactions are often complex and multistep, making it challenging to define reaction coordinates beforehand.
• Accurately simulating these intricate molecular transitions requires advanced computational techniques.

Purpose of the Study:

• To develop and demonstrate an efficient enhanced sampling algorithm for exploring complex biocatalytic reaction pathways.
• To improve the simulation efficiency and convergence for studying enzyme mechanisms.

Main Methods:

• Utilized adaptive path collective variables (PCVs) to converge to the minimum free energy path (MFEP).
• Combined PCVs with the well-tempered metadynamics extended-system adaptive biasing force (WTM-eABF) hybrid algorithm.
• Implemented a novel stabilization algorithm for extended-system methods to handle PCV discontinuities.
• Employed the multistate Bennett's acceptance ratio (MBAR) estimator to accelerate simulation convergence.

Main Results:

• Demonstrated dramatically increased sampling efficiency through the fast adaptation of WTM-eABF to path updates.
• Successfully addressed discontinuities in PCVs using a new stabilization algorithm.
• Showcased accelerated simulation convergence by integrating the MBAR estimator.
• Applied the method to the initial step of pseudouridine synthases' enzymatic reaction, confirming its efficacy.

Conclusions:

• The developed path WTM-eABF method, enhanced with PCVs and MBAR, significantly improves the efficiency of exploring complex molecular transitions in biocatalysis.
• This approach provides a powerful tool for elucidating intricate enzymatic reaction mechanisms.
• The study highlights the potential of adaptive sampling techniques in computational biochemistry.