Discovery of Tetrahydroisoquinoline-Based SARS-CoV-2 Helicase Inhibitors with Iterative, Deep Learning-Enhanced...

Area of Science:

• Drug discovery and development
• Computational chemistry
• Virology

Background:

• The SARS-CoV-2 helicase (Nsp13) is a critical target for antiviral drug development.
• Previous structure-based drug discovery efforts require efficient screening methods.

Purpose of the Study:

• To identify novel SARS-CoV-2 helicase inhibitors using AI-enhanced virtual screening.
• To explore the antiviral potential of 3-phenyl-1,2,3,4-tetrahydroisoquinoline (THIQ) derivatives.

Main Methods:

• Three rounds of structure-based virtual screening (VS) incorporating Artificial Intelligence (AI).
• A deep neural network (DNN) was used in the third VS round to prioritize molecules.
• Molecular dynamics simulations were employed to elucidate the binding mechanism.

Main Results:

• AI-enhanced VS improved hit-identification efficiency by 21%.
• Six promising hits with selectivity indexes > 3 were identified.
• MWAC-3429, a THIQ derivative, showed potent cell-based antiviral activity (EC50 = 5.4 μM) without cytotoxicity.
• A novel allosteric inhibition mechanism at the Nsp13 helicase was proposed for the THIQ chemotype.

Conclusions:

• AI-driven VS is effective for identifying SARS-CoV-2 helicase inhibitors.
• THIQ derivatives represent a promising class of antivirals targeting SARS-CoV-2.
• Understanding the allosteric inhibition mechanism can guide future drug design.