Polypeptide folding using Monte Carlo sampling, concerted rotation, and continuum solvation.

Area of Science:

• Computational chemistry
• Statistical mechanics
• Biophysics

Background:

• Predicting protein structure is crucial for understanding biological function.
• Accurate computational methods are needed to explore complex conformational landscapes.
• Existing simulation methods face challenges in efficiently sampling relevant protein conformations.

Purpose of the Study:

• To develop and apply an efficient concerted rotation algorithm for Monte Carlo simulations.
• To accurately predict the native folds of polypeptides, including beta-hairpins and alpha-helices.
• To validate the algorithm's ability to sample relevant conformational space and reproduce experimental structures.

Main Methods:

• Utilized an efficient concerted rotation algorithm within Monte Carlo simulations.
• Incorporated flexible bond and dihedral angles with Gaussian bias for optimized sampling.
• Employed the generalized Born surface area (GBSA) model for solvation in water.
• Used the OPLS-AA force field for molecular mechanics calculations.

Main Results:

• Successfully folded three polypeptides: U(1-17)T9D, alpha(1), and trpzip2.
• Computed lowest-energy structures for two beta-hairpins closely matched NMR data.
• For alpha(1) peptide, predicted a preference for random coil structures, aligning with low-concentration NMR experiments.
• Demonstrated effective sampling of conformational space, locating native states from extended conformations.

Conclusions:

• The developed algorithm efficiently samples conformational space for polypeptide folding simulations.
• The methodology accurately predicts native structures for beta-hairpins and conformational preferences for alpha-helices.
• The combination of the OPLS-AA force field and GBSA solvent model performs well for beta-turn forming peptides.