Deep learning-guided ligand generation for the strigolactone receptor ShHTL7

Area of Science:

• Agricultural Science
• Computational Chemistry
• Molecular Biology

Background:

• Striga hermonthica is a major parasitic weed impacting crop yields in sub-Saharan Africa.
• The strigolactone receptor ShHTL7 is crucial for Striga seed germination and a potential target for weed control.

Purpose of the Study:

• To develop a computational workflow for discovering ShHTL7-targeted ligands.
• To identify potential selective chemical regulators for Striga hermonthica.

Main Methods:

• Integrated deep learning (REINVENT4) and transfer learning for ligand generation.
• Multi-parameter screening including physicochemical properties, ADMET, molecular docking, and simulations.
• Molecular mechanics/Poisson-Boltzmann surface area (MM/PBSA) analysis for binding energetics.

Main Results:

• A computational framework successfully generated and prioritized ShHTL7-targeted ligand candidates.
• Docking and molecular dynamics simulations indicated favorable interactions and stable complexes for selected ligands.
• The compound inh-117 showed promising binding energetics and residue contributions compared to a known ligand.

Conclusions:

• The study presents an effective computational strategy for prioritizing ShHTL7 ligands.
• This framework can guide experimental research towards developing selective Striga hermonthica regulators.
• The findings contribute to sustainable agriculture by offering new avenues for parasitic weed management.