Predicting Nirmatrelvir Resistance in SARS-CoV-2 M^pro Mutants with an Integrated Computational Framework

Area of Science:

• Biochemistry
• Computational Biology
• Drug Discovery

Background:

• Nirmatrelvir is a key inhibitor of SARS-CoV-2 main protease (M^pro).
• Amino acid substitutions in M^pro can lead to drug resistance, impacting treatment efficacy.
• Predicting the impact of mutations is crucial for developing next-generation antiviral drugs.

Purpose of the Study:

• To computationally map the resistance landscape of the M^pro-nirmatrelvir complex.
• To identify key residues and predict the effects of single-point mutations on nirmatrelvir binding affinity.
• To provide a tool for predicting drug resistance and designing improved M^pro inhibitors.

Main Methods:

• An integrated computational pipeline combining alanine-scanning based on generalized Born and interaction entropy (ASGBIE) and saturation mutagenesis.
• Alchemical free-energy methods including free-energy perturbation (FEP), thermodynamic integration (TI), and multistate Bennett acceptance ratio (BAR/MBAR) were employed.
• Analysis of 228 single-point mutations across 12 identified key residues.

Main Results:

• 12 key residues within 4 Å of nirmatrelvir were identified.
• M165 and L167 were found to be the least mutationally tolerant residues, with most substitutions reducing inhibitor affinity.
• The study identified known drug-resistant mutations and predicted novel high-risk mutations (e.g., M165E/K, L167Y/W).
• Structural analysis revealed that mutations disrupt hydrophobic packing and hydrogen-bonding networks, weakening binding.

Conclusions:

• The computational pipeline reliably predicts drug resistance pathways for M^pro inhibitors.
• Identified vulnerable residues and predicted mutations provide insights for designing more robust nirmatrelvir analogs.
• This work facilitates the rational design of next-generation inhibitors to overcome drug resistance.