Predicting Metal-binding Proteins and Structures Through Integration of Evolutionary-scale and Physics-based...

Area of Science:

• Biochemistry
• Structural Biology
• Bioinformatics

Background:

• Metals are crucial for protein function, involved in stabilization, catalysis, and regulation.
• Identifying metal-binding proteins and their structures is vital for understanding biological processes and diseases.
• Existing computational methods require enhancement for accurate prediction and structural modeling of protein-metal interactions.

Purpose of the Study:

• To develop and validate a novel computational workflow, ESMBind, for predicting metal-binding proteins and their complex structures.
• To improve the accuracy of residue-level metal-binding predictions and 3D structure generation.
• To apply the ESMBind workflow to identify novel metal-binding proteins in uncharacterized fungal pathogens.

Main Methods:

• Utilized evolutionary scale modeling (ESM-2 and ESM-IF) for residue-level metal-binding probability prediction.
• Developed a physics-based approach incorporating metal-placement and energy minimization for 3D structure generation.
• Combined evolutionary and physics-based methods into the integrated ESMBind workflow.

Main Results:

• The ESMBind workflow demonstrated superior performance in both residue-level and 3D-level predictions compared to existing models.
• Successfully applied ESMBind to 142 uncharacterized fungal pathogen proteins.
• Identified potential metal-binding proteins implicated in fungal infection and virulence.

Conclusions:

• ESMBind provides a powerful and accurate computational tool for identifying and characterizing metal-binding proteins.
• The workflow advances the study of metalloproteins, with significant implications for understanding fungal pathogenesis.
• This approach facilitates the discovery of novel therapeutic targets in fungal pathogens.