Dynamic Docking of Macrocycles in Bound and Unbound Protein Structures with DynaDock

Area of Science:

• Computational chemistry
• Molecular modeling
• Drug discovery

Background:

• Macrocyclic compounds possess unique conformational properties due to their constrained yet flexible ring structures.
• Accurate structural description is crucial for computational modeling, especially molecular docking, which often struggles with macrocycle flexibility.
• Existing docking tools frequently use rigid receptor models and pre-generated ligand conformers, limiting accuracy for flexible molecules.

Purpose of the Study:

• To optimize a molecular dynamics-based sampling and docking pipeline for accurate prediction of macrocyclic compounds.
• To develop a dihedral classification for detailed conformational analysis of macrocyclic rings.
• To enable fully flexible docking of macrocycles against both bound and unbound protein structures.

Main Methods:

• Developed a dihedral classification procedure for conformational analysis of macrocyclic rings.
• Generated structural ensembles of macrocycles through molecular dynamics-based sampling.
• Performed molecular docking using a fully flexible approach against bound and unbound protein structures.

Main Results:

• Including ring conformers close to the bound state in the initial ensemble improved docking success rates.
• The fully flexible approach achieved high and decent accuracy in predicting bioactive conformations for bound and unbound protein structures, respectively.
• Docking failures were primarily attributed to flexible substituents rather than the macrocyclic ring itself.

Conclusions:

• Incorporating full molecular flexibility in docking pipelines enhances the prediction accuracy of macrocyclic compound conformations.
• Optimized sampling and flexible docking strategies are essential for accurate computational studies of macrocycles.
• Further improvements for challenging cases involving flexible substituents can be achieved with explicit molecular dynamics simulations.