Systematic multiscale parameterization of heterogeneous elastic network models of proteins

Area of Science:

• Computational biology
• Biophysics
• Structural biology

Background:

• Elastic Network Models (ENMs) are widely used for protein dynamics.
• Traditional ENMs often assume uniform properties, limiting their accuracy in complex biological systems.
• Atomistic simulations provide detailed but computationally expensive data on protein behavior.

Purpose of the Study:

• To present a novel method for parameterizing heterogeneous elastic network models (heteroENMs).
• To enable the reproduction of protein fluctuations observed in atomistic simulations using coarse-grained models.
• To develop adaptable protein models applicable to diverse conditions and environments.

Main Methods:

• Parameterization of heteroENMs based on atomistic simulation data.
• Application of the method across different levels of coarse-graining.
• Validation using three distinct biological systems: F-actin, carboxy myoglobin, and amphiphysin N-BAR domain.

Main Results:

• Accurate prediction of F-actin persistence length (6.1 ± 1.6 µm) compared to experimental values (9.0 ± 0.5 µm).
• HeteroENM demonstrated superior accuracy in predicting alpha-carbon atom fluctuations for carboxy myoglobin over uniform ENM and REACH models.
• Successfully captured distinct effective harmonic interactions for amphiphysin N-BAR domain in membrane-bound versus solution states.

Conclusions:

• The heteroENM parameterization method accurately reproduces atomistic simulation-derived fluctuations.
• This approach facilitates the development of versatile coarse-grained protein models for various biological contexts.
• The method offers a significant advancement in modeling protein dynamics and interactions with improved accuracy and applicability.