Predicting potential SARS-CoV-2 mutations of concern via full quantum mechanical modelling.

Area of Science:

• Computational chemistry
• Molecular modeling
• Structural biology

Background:

• Accurate prediction of intermolecular binding is crucial for understanding biological interactions.
• Simulating large molecular systems, like viral spike proteins and host receptors, presents significant computational challenges.
• Recent advancements in computational methods enable larger-scale quantum mechanical simulations.

Purpose of the Study:

• To predict and characterize the binding of SARS-CoV-2 spike variants (Wuhan, Omicron, and two Omicron-based) to the human ACE2 receptor.
• To assess the energetic contributions of individual amino acids to binding affinity.
• To predict the impact of single amino acid mutations on binding efficacy.

Main Methods:

• Utilized the quantum mechanics complexity reduction (QM-CR) approach for large-scale electronic structure simulations (approx. 13,000 atoms).
• Analyzed four SARS-CoV-2 spike variants: Wuhan, Omicron, and two Omicron-based variants.
• Performed experimental validation by comparing variant binding efficacy to cells expressing hACE2.

Main Results:

• The QM-CR model successfully predicted and characterized binding interactions between spike variants and hACE2.
• Identified specific amino acid contributions to binding energy.
• Predicted the beneficial effect of the A484K mutation on ACE2 binding, which was later observed in variant BA.2.86.

Conclusions:

• The QM-CR computational model is effective for identifying critical mutations influencing intermolecular interactions.
• This approach can inform the engineering of molecules with high specificity for target interactions.
• Computational modeling offers a powerful tool for predicting viral evolution and guiding therapeutic development.