Building and refining protein models within cryo-electron microscopy density maps based on homology modeling and...

Area of Science:

• Structural Biology
• Computational Biology
• Biophysics

Background:

• Automatic modeling using cryoelectron microscopy (cryoEM) density maps is effective for individual proteins but limited for large assemblies due to computational constraints.
• Existing unsupervised methods struggle with the scale and complexity of macromolecular assemblies.
• There is a need for efficient and accurate methods to build atomic models from cryoEM data for large biological structures.

Purpose of the Study:

• To introduce EM-IMO (electron microscopy-iterative modular optimization), a novel supervised method for building, modifying, and refining protein models guided by cryoEM density maps.
• To develop and validate a multiscale refinement strategy combining EM-IMO and molecular dynamics for constructing atomic models of large macromolecular assemblies.
• To apply the developed method to build a complete backbone model of the grass carp reovirus virion.

Main Methods:

• Development of EM-IMO, a supervised refinement method allowing user-guided parameter input to accelerate model building.
• Benchmarking of EM-IMO using simulated density maps and homology models.
• Application of a multiscale refinement protocol integrating EM-IMO and molecular dynamics for cryoEM data of the grass carp reovirus.

Main Results:

• EM-IMO significantly speeds up model refinement by incorporating user-defined parameters.
• The multiscale refinement strategy successfully built backbone models for all seven conformers of the five capsid proteins of the grass carp reovirus.
• A complete backbone model of the grass carp reovirus capsid was reconstructed, providing structural and functional insights.

Conclusions:

• The EM-IMO method offers a practical and efficient approach for refining protein models using cryoEM density maps.
• The integrated use of homology modeling and a multiscale refinement protocol (supervised and automated) is a viable strategy for atomic model building from medium- to high-resolution cryoEM maps.
• This study demonstrates a powerful computational strategy for analyzing complex macromolecular structures like viral capsids.