Approaching Elastic Network Models to Molecular Dynamics Flexibility.

Area of Science:

• Structural Biology
• Computational Biophysics
• Protein Dynamics

Background:

• Elastic network models (ENMs) are widely used for studying protein collective movements.
• Existing ENMs lack a consensus parametrization and often fail to accurately represent protein flexibility compared to molecular dynamics (MD).
• Standard ENMs tend to disperse crucial motions across multiple vibrational modes.

Purpose of the Study:

• To develop and validate a novel, improved elastic network model for protein dynamics.
• To systematically investigate the impact of residue connectivity and force field parameters on ENM accuracy.
• To create an ENM that better reproduces molecular dynamics flexibility and experimental conformational data.

Main Methods:

• Trained a new ENM against a comprehensive database of atomistic MD trajectories.
• Systematically explored residue connectivity, force constant analytical forms, and interaction thresholds.
• Developed a new potential function integrating sequential and spatial residue relationships.

Main Results:

• Identified contacts between the three nearest sequence neighbors as critical for fundamental protein motions.
• The new ENM demonstrates robustness across varying protein sizes and folds.
• Achieved systematic improvement over standard ENMs, matching MD results on long timescales.

Conclusions:

• The developed ENM accurately captures protein collective motions, including large conformational transitions and NMR ensemble diversity.
• This model offers a significant advancement for predicting protein dynamics and flexibility.
• The findings highlight the importance of incorporating both sequential and spatial residue information in coarse-grained models.