Data-guided Multi-Map variables for ensemble refinement of molecular movies

Area of Science:

• Structural Biology
• Computational Biophysics
• Biochemistry

Background:

• Molecular dynamics simulations are crucial for understanding protein behavior.
• Structure-centric experimental data, like cryo-EM and X-ray crystallography, provide valuable insights into protein conformations.
• Existing methods often focus on fitting models to density maps, limiting dynamic and thermodynamic information recovery.

Purpose of the Study:

• To develop a data-guided molecular dynamics approach using 3D protein electron density to simultaneously determine free energy costs of conformational transitions and refine atomic structures.
• To leverage the Multi-Map methodology for monitoring concerted movements in various simulation types.
• To enable the estimation of average properties and real-space refinement of structures from simulation ensembles.

Main Methods:

• Utilized 3D electron density maps from cryo-EM or X-ray crystallography as collective variables for molecular dynamics simulations.
• Employed the Multi-Map methodology to track conformational changes during simulations.
• Constructed all-atom ensembles along Multi-Map variable values to analyze average properties and refine structures.
• Tested the approach on three proteins of varying sizes to validate its efficacy.

Main Results:

• Demonstrated the ability of density-guided simulations to capture conformational transitions between known intermediates in proteins.
• Showcased reversible simulated pathways with minimal hysteresis, requiring only low-resolution density information.
• Generated estimates for free energy differences directly comparable to experimental data.
• Achieved superior refined model quality compared to existing Protein Data Bank structures.
• Identified optimal conditions for quantitative agreement with experimental free energies using medium-resolution density and large structural transitions.

Conclusions:

• Data-guided molecular dynamics using electron density is a viable strategy for recovering thermodynamic information and refining protein structures.
• The Multi-Map methodology effectively monitors conformational dynamics driven by experimental data.
• This approach provides accurate free energy estimates and high-quality structural models, advancing the study of protein conformational landscapes.