AI-Accelerated Design of Targeted Covalent Inhibitors for SARS-CoV-2.

Area of Science:

• Drug Discovery
• Computational Chemistry
• Virology

Background:

• The COVID-19 pandemic necessitates novel antiviral therapies beyond vaccines.
• Emerging SARS-CoV-2 variants require rapid antiviral discovery workflows.
• Existing pipelines focus on noncovalent inhibitors; covalent inhibitors offer an alternative strategy.

Purpose of the Study:

• To develop an AI-driven pipeline for designing covalent antiviral candidates targeting SARS-CoV-2 main protease (Mpro).
• To integrate computational design with experimental validation for efficient lead discovery.
• To identify and characterize novel covalent inhibitors of Mpro.

Main Methods:

• A closed-loop artificial intelligence pipeline was developed for designing covalent inhibitors.
• Deep learning was used to incorporate linkers and electrophilic warheads.
• Experimental validation employed native mass spectrometry and FRET-based assays.
• Room-temperature X-ray crystallography and molecular dynamics simulations were used for structural analysis.

Main Results:

• Four chloroacetamide-based covalent inhibitors of Mpro were identified.
• Inhibitors demonstrated micromolar affinities with a K_I of 5.27 μM.
• Experimental binding modes were consistent with computational predictions.
• Molecular dynamics simulations suggested potential for improved selectivity and reduced toxicity.

Conclusions:

• The developed modular, data-driven AI pipeline is effective for discovering potent and selective covalent inhibitors.
• This approach provides a versatile platform for targeting emerging viral threats.
• Covalent inhibitors targeting Mpro show promise for COVID-19 treatment.