Re-evaluation of low-resolution crystal structures via interactive molecular-dynamics flexible fitting (iMDFF): a...

Area of Science:

• Structural Biology
• Computational Biology
• Biophysics

Background:

• High-resolution protein structures are abundant in the Protein Data Bank, but many structures are built manually using low-resolution or weak electron density data.
• Current model-building tools struggle with accuracy, leading to residual errors like atomic clashes and unfavorable conformations, which can mask larger protein fold errors.
• Even structures with good Rfree and MolProbity scores at resolutions up to 3.5 Å can contain significant errors, as seen in recent complement C4 structures.

Purpose of the Study:

• To introduce and evaluate the interactive molecular-dynamics flexible fitting (iMDFF) approach for improving manual model building in low-resolution cryo-EM density maps.
• To demonstrate iMDFF's effectiveness in detecting and correcting register errors in protein structures, particularly those derived from challenging datasets.
• To show that iMDFF can yield stereochemically sound protein models comparable to atomic resolution structures, even from low-resolution data.

Main Methods:

• Utilized a molecular-dynamics force field that explicitly models interatomic forces to reduce the search space for manual rebuilding.
• Applied the iMDFF approach for user-guided model building in electron density maps with resolutions lower than 3.5 Å.
• Corrected and re-refined three complement C4 structures (PDB entries 4fxg, 4fxk, 4xam) using iMDFF, achieving MolProbity scores below 1.7.

Main Results:

• iMDFF significantly reduced the complexity of conformational space, improving the detection and correction of register errors in six β-strands of the complement C4 structures.
• Manual rebuilding guided by iMDFF converged to solutions with stereochemical quality comparable to atomic resolution structures, even from maps below 3.5 Å resolution.
• The corrected and refined models led to an extension of the resolution for complement C4b from 4.2 Å to 3.5 Å, validated by paired refinement.

Conclusions:

• iMDFF is a powerful tool for improving the accuracy and stereochemical quality of protein models built from low-resolution and weak electron density data.
• The approach facilitates the detection and correction of subtle errors, such as register errors in β-strands, which are difficult to identify with conventional methods.
• iMDFF enables the refinement of protein structures to higher resolutions and improved quality, advancing the field of structural biology and protein structure determination.