OLIVES: A Go̅-like Model for Stabilizing Protein Structure via Hydrogen Bonding Native Contacts in the Martini 3...

Area of Science:

• Computational biology
• Biophysics
• Molecular dynamics simulations

Background:

• Coarse-grained molecular dynamics (CGMD) simulations are crucial for modeling complex biological systems.
• The Martini 3 force field is promising but requires bias potentials for folded protein stability.
• Existing bias potentials like elastic networks or Go̅-like models have limitations.

Purpose of the Study:

• To develop a novel algorithm, OLIVES, for coarse-grained protein simulations.
• To address the instability of folded proteins in the Martini 3 force field.
• To investigate the role of hydrogen bonds in protein structure stability within CGMD.

Main Methods:

• Developed the OLIVES algorithm to identify native contacts with hydrogen bond capabilities.
• Implemented a novel Go̅-like model for Martini 3 using OLIVES.
• Incorporated realistic hydrogen bond energies from ab initio calculations.
• Validated the model against known protein complexes.

Main Results:

• Protein structure instability in Martini 3 is partly due to the lack of hydrogen bond energy.
• Reintroducing a coarse-grained hydrogen bond network recovers protein stability.
• OLIVES eliminates the need for secondary structure restraints.
• OLIVES reduces bias terms and speeds up Martini 3 simulations by up to 30% on GPUs.

Conclusions:

• The OLIVES algorithm and its associated Go̅-like model enhance protein stability in Martini 3 simulations.
• Realistic hydrogen bond modeling is essential for accurate CGMD of proteins.
• OLIVES offers a more efficient alternative to existing bias methods for protein simulations.