Do crystal structures obviate the need for theoretical models of GPCRs for structure-based virtual screening?

Area of Science:

• Structural Biology
• Computational Chemistry
• Pharmacology

Background:

• Recent X-ray crystallography of beta-2 adrenoreceptor (β2AR) provided crucial insights into GPCR ligand discovery.
• This advancement raised questions about the utility of pre-existing 3D theoretical models of GPCRs.

Purpose of the Study:

• To compare the structural similarity and virtual screening effectiveness of historical β2AR models against newly available crystal structures.
• To evaluate the role of theoretical models in structure-based drug discovery for GPCRs.

Main Methods:

• Collected theoretical β2AR models (homology and de novo) from various research groups, all developed before crystal structure availability.
• Performed virtual screening using these models and X-ray crystal structures against large chemical libraries.
• Assessed model performance based on hit rates and structural quality, focusing on binding pocket residue similarity (RMSD) and interaction accuracy.

Main Results:

• Theoretical β2AR models showed significant differences from crystal structures in transmembrane helix definition and overall packing.
• Several theoretical models yielded higher hit rates in virtual screening for β2AR agonists and antagonists compared to X-ray structures.
• Screening performance correlated with local structural quality, particularly the accuracy of binding pocket residue conformations and protein/ligand interactions.

Conclusions:

• Well-constructed theoretical GPCR models can effectively capture essential binding pocket features.
• These models may offer practical advantages over X-ray structures for structure-based drug discovery, especially for identifying GPCR ligands.
• Local structural accuracy is key for the success of theoretical models in virtual screening applications.