Prioritizing Neuroactive Ligands Using Motif-Guided Virtual Discovery and Zebrafish Profiling

Area of Science:

• Computational chemistry
• Drug discovery
• Structural biology

Background:

• Virtual screening identifies many drug candidates, but experimental validation is challenging.
• In silico predictions often fail to correlate with in vivo efficacy or predict off-target effects.
• Prioritizing candidates for preclinical testing remains a significant bottleneck in drug discovery.

Purpose of the Study:

• To introduce Rosetta Engine for Anchoring Ligands with a Motif (REAL-M), a novel computational screening algorithm.
• To leverage structural interaction data from the Protein Data Bank (PDB) for improved ligand placement and selection.
• To address the challenge of prioritizing drug candidates for experimental validation.

Main Methods:

• Developed and applied the REAL-M algorithm, utilizing PDB structural data.
• Tested REAL-M using the hypocretin receptor as a case study.
• Validated predicted antagonists through cell-based assays (PRESTO-Tango) and larval zebrafish behavioral studies.

Main Results:

• 28 out of 30 REAL-M predicted hypocretin receptor antagonists significantly blocked agonist binding in cell assays.
• Six chemically diverse antagonists showed efficacy comparable to existing drugs.
• Three compounds mitigated hypocretin-induced hyperactivity in zebrafish, with specificity confirmed using a knockout model.

Conclusions:

• The REAL-M pipeline effectively prioritizes potent and specific drug candidates.
• This method demonstrates high predictive accuracy, reducing the need for extensive experimental screening.
• The REAL-M approach is adaptable to other protein targets with conserved binding pockets, accelerating drug discovery.