Protein dynamics investigated by inherent structure analysis.

Area of Science:

• Computational biology
• Biophysics
• Protein dynamics

Background:

• Molecular dynamics (MD) simulations are crucial for understanding protein thermodynamics and dynamics.
• Accurate free-energy surface construction from MD trajectories requires effective snapshot discretization.
• Previous clustering methods have limitations in analyzing complex protein conformational landscapes.

Purpose of the Study:

• To introduce and validate the inherent structures (IS) approach for discretizing MD trajectories.
• To demonstrate the utility of IS for constructing accurate free-energy surfaces and conformational networks.
• To compare the IS method with traditional root-mean-square deviation (rmsd) clustering.

Main Methods:

• Utilizing inherent structures (IS) to discretize molecular dynamics simulation snapshots.
• Constructing configuration space networks from transitions between inherent structures.
• Applying the mincut method to derive one-dimensional free-energy profiles.
• Analyzing a 10-residue peptide beta-hairpin and the PDZ2 signaling domain.

Main Results:

• The IS approach provides a robust discretization, avoiding issues with standard clustering algorithms.
• A 10-residue peptide simulation revealed a complex native state organization with free-energy barriers up to 3 kcal/mol.
• Coarse-grained network analysis identified multiple pathways between conformational valleys, hidden by rmsd clustering.
• The IS method proved effective for the biologically relevant PDZ2 signaling domain.

Conclusions:

• Inherent structures offer a superior method for analyzing protein dynamics and free-energy landscapes from MD simulations.
• The IS approach reveals intricate conformational dynamics and multiple transition pathways, crucial for understanding protein function.
• This method enhances the study of complex biological systems, including signaling domains like PDZ2.