Ensemble-Based Precision Refinement of All-Atom Nucleic Acid Force Fields Guided by NMR NOE Pair-Distance...

Area of Science:

• Computational chemistry
• Structural biology
• Biophysics

Background:

• All-atom simulations struggle to accurately model nucleic acid structures, particularly noncanonical motifs like loops and G-quadruplexes.
• Existing classical force fields often fail to reproduce experimental data for nucleic acid ensembles.

Purpose of the Study:

• To develop and validate a systematic refinement strategy for AMBER-based force fields.
• To improve the accuracy of molecular dynamics simulations for nucleic acid structures and dynamics.

Main Methods:

• Incorporation of nuclear Overhauser effect (NOE) distance data from NMR experiments.
• An ensemble-averaged optimization framework was used to refine force fields.
• Selective tuning of van der Waals interaction parameters.

Main Results:

• The refined force fields significantly reduce discrepancies between simulations and experimental data.
• Persistent simulation artifacts were removed, leading to improved free energy landscapes.
• The strategy demonstrated broad applicability across various DNA and RNA systems, including flexible loops and G-quadruplexes.

Conclusions:

• The developed transferable strategy substantially enhances the structural accuracy and predictive power of nucleic acid simulations.
• This approach enables more reliable modeling of complex nucleic acid conformational ensembles.
• Improved force fields are crucial for advancing the understanding of nucleic acid function and interactions.