Development of a Cell-Based Luciferase Complementation Assay for Identification of SARS-CoV-2 3CLpro Inhibitors

Abstract

The 3C-like protease (3CL^pro) of SARS-CoV-2 is considered an excellent target for COVID-19 antiviral drug development because it is essential for viral replication and has a cleavage specificity distinct from human proteases. However, drug development for 3CL^pro has been hindered by a lack of cell-based reporter assays that can be performed in a BSL-2 setting. Current efforts to identify 3CL^pro inhibitors largely rely upon in vitro screening, which fails to account for cell permeability and cytotoxicity of compounds, or assays involving replication-competent virus, which must be performed in a BSL-3 facility. To address these limitations, we have developed a novel cell-based luciferase complementation reporter assay to identify inhibitors of SARS-CoV-2 3CL^pro in a BSL-2 setting. The assay is based on a lentiviral vector that co-expresses 3CL^pro and two luciferase fragments linked together by a 3CL^pro cleavage site. 3CL^pro-mediated cleavage results in a loss of complementation and low luciferase activity, whereas inhibition of 3CL^pro results in 10-fold higher levels of luciferase activity. The luciferase reporter assay can easily distinguish true 3CL^pro inhibition from cytotoxicity, a powerful feature that should reduce false positives during screening. Using the assay, we screened 32 small molecules for activity against SARS-CoV-2 3CL^pro, including HIV protease inhibitors, HCV protease inhibitors, and various other compounds that have been reported to inhibit SARS-CoV-2 3CL^pro. Of these, only five exhibited significant inhibition of 3CL^pro in cells: GC376, boceprevir, Z-FA-FMK, calpain inhibitor XII, and GRL-0496. This assay should greatly facilitate efforts to identify more potent inhibitors of SARS-CoV-2 3CL^pro.

Keywords: 3CLpro; COVID-19; SARS-CoV-2; antiviral; inhibitor; luciferase; protease.

Figure 1

Development of cell-based luciferase complementation reporters to detect inhibition of SARS-CoV-2 3CL^pro activity. (A) Strategy for detection of 3CL^pro inhibition using NanoBiT, a luciferase complementation reporter comprised of Large BiT (L) and Small BiT (S). When connected by a linker containing a 3CL^pro cleavage site, 3CL^pro cleavage should result in a loss of complementation and low luciferase activity. Conversely, 3CL^pro inhibitors should prevent cleavage, resulting in high luciferase activity. (B) Lentiviral vectors that express 3CL^pro and reporters. P2A, a self-cleaving peptide, was inserted between 3CL^pro and reporters. Reporters consisted of GFP linked to L and S in various orders. L and S were separated by the SARS-CoV-2 nsp4-nsp5 cleavage site. LTR, long terminal repeat; P_EF1α, EF1α promoter; IRES, internal ribosome entry site; puro, puromycin resistance gene; WPRE, woodchuck hepatitis virus post-transcriptional regulatory element. (C) Western blotting to examine reporter expression and cleavage. Lentiviral vectors expressing 3CL^pro WT or C145A (a catalytically inactive mutant) were transfected into 293T cells, and DMSO (control) or the 3CL^pro inhibitor GC376 (100 µM) was added. Cell lysates were collected 30 h post-transfection. 3CL^pro was detected using anti-SARS-CoV 3CL^pro or anti-P2A antibodies, whereas reporters were detected using anti-GFP or anti-L antibodies. Arrows denote cleavage products of expected sizes. For 3CL^pro C145A, two different plasmid amounts were transfected: 2 or 0.4 µg. HSP90 was included as a loading control. (D) NanoBiT luciferase activity. 293T cells were transfected (same sample order as in C), and NanoBiT activity (expressed as relative luciferase units, or RLU) was measured 30 h post-transfection. The data represent the mean ± standard deviation of four independent experiments; ns: not significant; * p < 0.05, ** p < 0.01, *** p < 0.001 (relative to 3CL^pro WT without drug; mixed effects analysis with Dunnett’s post-test).

Figure 2

The luciferase complementation assay accurately reflects concentration-dependent inhibition of SARS-CoV-2 3CL^pro-mediated cleavage. (A) Western blotting to examine reporter cleavage. 293T cells were transfected with the lentiviral vector containing 3CL^pro WT and the S-L-GFP reporter cassette, and DMSO (control) or GC376 (0.1, 0.4, 1.6, 6.3, 25, 100, or 200 µM) was added. Mock transfection (MT) and 3CL^pro C145A S-L-GFP were included as controls. On the GFP and LgBiT (L) panels, the bottom band corresponds to cleaved reporter (L-GFP, 46 kDa, indicated by arrows), while the top band corresponds to uncleaved reporter (S-L-GFP, 49 kDa). (B) Quantification of the percentage of uncleaved reporter from three independent western blots based on GFP (circle) or L (triangle) detection. (C) NanoBiT luciferase activity and cell viability. 293T cells were transfected with the S-L-GFP lentiviral vector, and DMSO (control) or GC376 (0.1, 0.4, 1.6, 6.3, 25, or 100 µM) was added. NanoBiT activity (circle, left axis) and cell viability (measured with the CellTiter-Glo 2.0 assay; triangle, right axis) were analyzed 30 h post-transfection and are expressed as relative luciferase units (RLU). Cell viability data were normalized to the DMSO sample, which was set to 100%. NanoBiT data were normalized to the DMSO and 100 µM GC376 samples, which were set to 0 and 100%, respectively. The data represent the mean ± standard deviation of four independent experiments; ns: not significant; * p < 0.05, ** p < 0.01, *** p < 0.001 (relative to no-drug; one-way ANOVA with Dunnett’s post-test).

Figure 3

The luciferase complementation assay can easily distinguish 3CL^pro inhibition from cytotoxicity. To determine if the luciferase complementation reporter assay can differentiate 3CL^pro inhibition from cytotoxicity, 293T cells were transfected with the S-L-GFP lentiviral vector, and DMSO (control) or the anti-cancer drug doxorubicin (0.1–100 µM) was added. NanoBiT activity (circle, left axis) and cell viability (measured with the CellTiter-Glo 2.0 assay; triangle, right axis) were analyzed 30 h post-transfection and are expressed as relative luciferase units (RLU). Cell viability data were normalized to the DMSO sample, which was set to 100%. NanoBiT data were normalized to the DMSO and 100 µM GC376 samples, which were set to 0 and 100%, respectively. The data represent the mean ± standard deviation of three independent experiments.

Figure 4

HIV protease inhibitors do not block SARS-CoV-2 3CL^pro activity in a cell-based luciferase complementation assay. To determine if HIV protease inhibitors are active against SARS-CoV-2 3CL^pro in cells, 293T cells were transfected with the S-L-GFP lentiviral vector, and DMSO (control) or inhibitors (0.1, 0.4, 1.6, 6.3, 25, or 100 µM) were added. NanoBiT activity (circle, left axis) and cell viability (measured with the CellTiter-Glo 2.0 assay; triangle, right axis) were analyzed 30 h post-transfection and are expressed as relative luciferase units (RLU). Cell viability data were normalized to the DMSO sample, which was set to 100%. NanoBiT data were normalized to the DMSO and 100 µM GC376 samples, which were set to 0 and 100%, respectively. The data represent the mean ± standard deviation of three independent experiments; * p < 0.05, ** p < 0.01, *** p < 0.001 (relative to no-drug; one-way ANOVA with Dunnett’s post-test).

Figure 5

Boceprevir, Z-FA-FMK, calpain inhibitor XII, and GRL-0496, inhibit SARS-CoV-2 3CL^pro in a cell-based luciferase complementation assay. To determine if other compounds are active against SARS-CoV-2 3CL^pro in cells, 293T cells were transfected with the S-L-GFP lentiviral vector, and DMSO (control) or compounds (0.1, 0.4, 1.6, 6.3, 25, or 100 µM) were added. NanoBiT activity (circle, left axis) and cell viability (measured with the CellTiter-Glo 2.0 assay; triangle, right axis) were analyzed 30 h post-transfection and are expressed as relative luciferase units (RLU). Cell viability data were normalized to the DMSO sample, which was set to 100%. NanoBiT data were normalized to the DMSO and 100 µM GC376 samples, which were set to 0 and 100%, respectively. The data represent the mean ± standard deviation of three independent experiments; ** p < 0.01, *** p < 0.001 (relative to no-drug; one-way ANOVA with Dunnett’s post-test).