A yeast-based system to study SARS-CoV-2 Mpro structure and to identify nirmatrelvir resistant mutations

Abstract

The SARS-CoV-2 main protease (Mpro) is a major therapeutic target. The Mpro inhibitor, nirmatrelvir, is the antiviral component of Paxlovid, an orally available treatment for COVID-19. As Mpro inhibitor use increases, drug resistant mutations will likely emerge. We have established a non-pathogenic system, in which yeast growth serves as an approximation for Mpro activity, enabling rapid identification of mutants with altered enzymatic activity and drug sensitivity. The E166 residue is known to be a potential hot spot for drug resistance and yeast assays identified substitutions which conferred strong nirmatrelvir resistance and others that compromised activity. On the other hand, N142A and the P132H mutation, carried by the Omicron variant, caused little to no change in drug response and activity. Standard enzymatic assays confirmed the yeast results. In turn, we solved the structures of Mpro E166R, and Mpro E166N, providing insights into how arginine may drive drug resistance while asparagine leads to reduced activity. The work presented here will help characterize novel resistant variants of Mpro that may arise as Mpro antivirals become more widely used.

Fig 1. M^pro confers a significant reduction in growth in yeast caused by decreases in a variety of cellular proteins.

A) The indicated SARS-CoV-2 genes regulated by a galactose inducible promoter were expressed in yeast and conferred growth defects compared to empty vector (EV). B) Bar graph shows the total growth of cultures after 72 hours normalized to EV. C) Galactose-induced (ga) expression of the catalytically inactive M^pro C145A mutant (C145A ga) does not confer a growth reduction compared to WT (M^pro ga). When grown in glucose (gl) all three strains grew equally well. D) Protein levels of the M^pro C145A mutant and wild-type M^pro (WT) are comparable. Shown are two biological replicates for each form of M^pro. Ratios of Flag:GAPDH signals relative to M^pro WT is shown at the bottom of each lane. E) Total protein lysates made from yeast expressing the wild-type M^pro (WT) or M^pro C145A mutant (MUT) were subjected to mass spectrometric analyses revealing 153 proteins were higher in abundance in the mutant relative to the wild-type. F) Gene Ontology (GO) analyses indicates an enrichment of proteins with functions in translation that are significantly reduced in the presence of M^pro versus M^pro C145A. Plots in A, B, C show averages from three biological replicates and error bars are standard deviations. (***) indicates differences (p<0.001)between EV and tested genes.

Fig 2. Yeast growth assays identify nirmatrelvir resistant M^pro mutants.

A) Total growth of cultures after 72 hours expressing M^pro in the presence of increasing doses of nirmatrelvir normalized to growth of yeast carrying empty vector (EV) are plotted. Growth is restored by nirmatrelvir in a dose dependent manner. B) No growth effects are observed in cells treated up to 200μM nirmatrelvir. C) Yeast expressing substitutions E166D and E166N grow as well as EV but E166R, P132H, and N142A results in significant growth reduction comparable to wild-type M^pro. D) Western analysis shows that mutants and wild-type M^pro are expressed at comparable levels. Ratios of Flag:GAPDH signals relative to M^pro WT is shown at the bottom of each lane. E—G) Cells expressing P132H and N142A remain sensitive to nirmatrelvir, indicated by growth recovery, but E166R appears to be resistant as there is a lack of growth even when treated with 200μM of nirmatrelvir. H) RC₅₀ measurements of each mutant in response to nirmatrelvir treatment. For all experiments, at least three biological and three technical replicates were performed. Error bars represent standard deviations. (*, p<0.01; ***, p<0.001) indicates differences compared to EV.

Fig 3. Enzymatic assays demonstrate that E166N and E166D have severe defects in catalytic activity and M^pro E166R is highly resistant to nirmatrelvir.

A) Michaelis–Menten plot of M^pro and its mutants with various concentrations of FRET substrate. The K_m, V_max, k_cat, and k_cat/K_m values are shown in the table. B) The IC₅₀ plots of nirmatrelvir, PF-00835231. and GC-376 against M^pro WT, N142A, E166R, E166N, and E166D. C) K_i plots of nirmatrelvir, GC-376, and PF-00835231 against M^pro, M^pro E166R, and M^pro N142A.

Fig 4. Crystal structures reveals structural basis for E166R resistance and E166N inactivity.

A) Apo M^pro WT (white, PDB 7JP1) aligned with apo M^pro E166N (green, PDB 8DDI). B) M^pro WT GC-376 complex (white, PDB 6WTT) aligned with M^pro E166R GC-376 complex (magenta, PDB 8DDM). WT hydrogen bonds are shown as black dashes, and mutant hydrogen bonds are shown as red dashes. GC-376 is shown in white for the WT structure and cyan for the mutant structure. Mutations are indicated with red text. Ser1 from an adjacent protomer is indicated with orange text.