Toward structure prediction of cyclic peptides.

Area of Science:

• Computational chemistry
• Molecular modeling
• Biophysics

Background:

• Cyclic peptides show promise for modulating protein-protein interactions.
• Accurate structural prediction is key to advancing cyclic peptide drug development.

Purpose of the Study:

• To develop and validate a computational method for converged structural description of cyclic peptides.
• To assess the performance of standard protein force fields for cyclic peptide modeling.

Main Methods:

• Integration of bias-exchange metadynamics simulations.
• Application of Boltzmann reweighting and dihedral principal component analysis.
• Utilizing modified density peak-based cluster analysis for structural convergence.

Main Results:

• The developed computational method provides a converged structural description for cyclic peptides.
• Tested protein force fields exhibit over-stabilization of helical and extended conformations (α-helix, PPII/β regions).
• These regions are typically populated by linear peptides and proteins, not cyclic ones.

Conclusions:

• Existing protein force fields require re-parameterization to accurately model cyclic peptides.
• A force field that captures the full Ramachandran plot is essential for reliable cyclic peptide simulations.