Molecular modelling methods for prediction of sequence-selectivity in DNA recognition.

Area of Science:

• Computational Chemistry
• Molecular Biophysics
• Drug Discovery

Background:

• Predicting ligand-DNA interactions is crucial for drug development.
• Molecular modeling offers a computational approach to study these interactions.
• The generalized Born (GB) method combined with molecular mechanics (MM) is a popular technique.

Purpose of the Study:

• To evaluate the utility of the molecular mechanics/generalized Born (MM/GB) approach for predicting relative binding affinities of DNA minor groove binding ligands.
• To analyze the energetics of binding for specific ligands to different DNA sequences.
• To provide insights into setting up and analyzing molecular dynamics (MD) simulations for drug binding studies.

Main Methods:

• Application of molecular mechanics/generalized Born (MM/GB) methods.
• Utilizing molecular dynamics (MD) simulations to model ligand-DNA interactions.
• Analysis of binding energetics, including DNA perturbation energy and configurational entropy changes.

Main Results:

• The MM/GB approach, enhanced with specific calculations like DNA perturbation energy, shows improved agreement with experimental data for relative binding affinity predictions.
• Certain methodological elaborations, such as calculating configurational entropy changes, require careful interpretation.
• The study successfully ranked the binding affinity of Hoechst 33258 to various A/T-rich DNA duplexes.

Conclusions:

• Molecular modeling with MM/GB can reliably rank the relative binding affinities of ligands to DNA minor grooves.
• Calculating absolute binding affinities using current MM/GB methodologies remains challenging and less reliable.
• The study highlights the strengths and limitations of computational approaches in predicting ligand-DNA interactions.