Evaluating molecular mechanical potentials for helical peptides and proteins.

Area of Science:

• Computational biophysics
• Molecular dynamics simulations
• Protein structure prediction

Background:

• Accurate molecular mechanics force fields are essential for simulating protein dynamics.
• Previous AMBER force field variants showed limitations in characterizing helix-coil equilibria and protein stability.
• Evaluating force field performance requires rigorous testing with peptides and larger protein systems.

Purpose of the Study:

• To quantitatively evaluate multiple AMBER all-atom force field variants for helix-coil equilibria in explicit solvent.
• To assess the performance of AMBER force fields using large ensembles of helical peptides and a flexible five-helix-bundle protein (apolipophorin-III).
• To identify the most accurate AMBER force field for predicting protein structural properties and dynamics.

Main Methods:

• Utilized a global distributed computing network for large-scale molecular dynamics simulations.
• Achieved absolute conformational convergence for capped A(21) and F(s) helical peptides.
• Simulated the 164-residue apolipophorin-III protein over 100 nanoseconds.

Main Results:

• AMBER-99SB exhibited helix-destabilizing tendencies and caused apolipophorin-III unfolding.
• AMBER-03 showed issues with specific amino acid substitutions, lacked 3(10) helical content, and deviated from NMR data.
• AMBER-99phi accurately stabilized apolipophorin-III and matched experimental NMR ensemble data for gyration radius and solvent-exposed surface area.

Conclusions:

• AMBER-99phi demonstrates superior performance in modeling helix-coil equilibria and protein structure compared to AMBER-99SB and AMBER-03.
• The findings highlight the importance of force field selection for reliable computational studies of protein structure and dynamics.
• AMBER-99phi is recommended for future simulations requiring accurate characterization of protein conformational ensembles.