Predicting Protein Dynamics and Allostery Using Multi-Protein Atomic Distance Constraints.

Area of Science:

• Computational Biology
• Structural Biology
• Biophysics

Background:

• Studying protein dynamics, conformational ensembles, and allostery is challenging for molecular dynamics (MD) due to timescale limitations.
• Existing coarse-grained methods struggle to capture large conformational changes in proteins.

Purpose of the Study:

• To present ExProSE (Exploration of Protein Structural Ensembles), a novel distance geometry-based method for generating protein structural ensembles.
• To provide a fast and accessible framework for exploring protein structure and dynamics.
• To assess ExProSE's ability to predict allosteric sites.

Main Methods:

• ExProSE generates an ensemble of protein structures from two input structures using a distance geometry approach.
• The method was evaluated using apo/holo protein pairs and compared against existing coarse-grained methods (tCONCOORD, NMSim) and targeted MD.
• Additional constraints representing potential modulators were incorporated to predict allosteric sites.

Main Results:

• ExProSE successfully generates protein ensembles, outperforming tCONCOORD and NMSim for T4-lysozyme and achieving comparable results to targeted MD.
• ExProSE demonstrates capability in predicting allosteric sites, ranking the correct pocket first or second for 27 out of 58 tested allosteric proteins.
• The performance in allosteric site prediction is comparable and complementary to existing computational methods.

Conclusions:

• ExProSE offers a unified, fast, and accessible framework for exploring protein structure and dynamics, addressing limitations of traditional MD.
• The method effectively generates native-like conformational ensembles and shows promise in identifying allosteric sites.
• ExProSE is freely available, facilitating broader research in protein structure-function relationships.