High-Level ab Initio Calculations To Improve Protein Backbone Dihedral Parameters

Area of Science:

• Computational Chemistry
• Molecular Modeling
• Biophysics

Background:

• Accurate molecular mechanical (MM) force fields are crucial for simulating protein structure and dynamics.
• Existing MM dihedral parameters for protein backbones, such as those in AMBER, may not fully capture the nuances of Ramachandran angles (ϕ and ψ).

Purpose of the Study:

• To derive novel, high-accuracy molecular mechanical dihedral parameters for the protein backbone Ramachandran angles (ϕ and ψ).
• To improve the fidelity of molecular mechanics simulations by better representing protein conformational preferences.

Main Methods:

• High-level ab initio molecular orbital calculations were performed on model dipeptides (glycine and alanine).
• Fully relaxed (ϕ, ψ) contour maps were generated at the MP2/6-31G(d) level.
• Torsional energy profiles were computed using a higher level of theory (DF-LCCSD(T0)/Aug-cc-pVTZ//DF-LMP2/Aug-cc-pVTZ).
• New dihedral parameters were derived to fit these ab initio profiles, addressing discrepancies with existing AMBER force field variants.

Main Results:

• Ab initio calculations provided detailed torsional energy profiles for protein backbone dihedral angles.
• Significant differences were observed between ab initio profiles and those generated by standard AMBER force field parameters.
• New molecular mechanical dihedral parameters were successfully derived, demonstrating improved agreement with the high-level computational data.

Conclusions:

• The developed dihedral parameters offer a more accurate representation of protein backbone Ramachandran angles (ϕ and ψ).
• These new parameters are expected to enhance the accuracy of molecular mechanics simulations in predicting protein structures and dynamics.
• This work provides a foundation for refining force fields used in computational biology and drug discovery.