Homology model-based virtual screening for GPCR ligands using docking and target-biased scoring

Area of Science:

• Computational chemistry
• Structural biology
• Pharmacology

Background:

• G-protein coupled receptors (GPCRs) are crucial drug targets.
• Virtual screening aids in identifying novel GPCR ligands.
• Homology modeling is a key technique for structure-based drug design.

Purpose of the Study:

• To improve homology model-based virtual screening for GPCR ligands.
• To combine ligand-supported homology modeling with interaction fingerprint similarity (IFS) for enhanced scoring.
• To evaluate the developed approach for identifying antagonists of metabotropic glutamate receptor subtype 5 (mGluR5).

Main Methods:

• Generated GPCR homology models using ligand-supported modeling.
• Applied ligand-receptor interaction fingerprint-based similarity (IFS) for scoring and ranking compounds.
• Conducted retrospective virtual screening experiments for mGluR5 antagonists.
• Compared IFS performance against conventional scoring functions (Dock-Score, PMF-Score, Gold-Score, ChemScore, FlexX-Score).

Main Results:

• The IFS approach demonstrated significantly higher enrichment rates compared to conventional scoring functions.
• The method achieved good results without being biased towards the chemical classes of reference structures.
• Successfully identified potential antagonists for mGluR5 in retrospective screening.

Conclusions:

• The combined approach of ligand-supported modeling and IFS is effective for GPCR virtual screening.
• This method offers a robust strategy for structure-based GPCR drug discovery.
• The approach shows potential as a general framework for GPCR virtual screening.