Evaluation of SARS-CoV-2 Main Protease Inhibitors Using a Novel Cell-Based Assay.

As an essential enzyme of SARS-CoV-2, main protease (MPro) triggers acute toxicity to its human cell host, an effect that can be alleviated by an MPro inhibitor. Using this toxicity alleviation, we developed an effective method that allows a bulk analysis of the cellular potency of MPro inhibitors. This novel assay is advantageous over an antiviral assay in providing precise cellular MPro inhibition information to assess an MPro inhibitor. We used this assay to analyze 30 known MPro inhibitors. Contrary to their strong antiviral effects and up to 10 μM, 11a, calpain inhibitor II, calpain XII, ebselen, bepridil, chloroquine, and hydroxychloroquine showed relatively weak to undetectable cellular MPro inhibition potency implicating their roles in interfering with key steps other than just the MPro catalysis in the SARS-CoV-2 life cycle. Our results also revealed that MPI5, MPI6, MPI7, and MPI8 have high cellular and antiviral potency. As the one with the highest cellular and antiviral potency among all tested compounds, MPI8 has a remarkable cellular MPro inhibition IC50 value of 31 nM that matches closely to its strong antiviral effect with an EC50 value of 30 nM. Therefore, we cautiously suggest exploring MPI8 further for COVID-19 preclinical tests.

Introduction

COVID-19 has paralyzed much of the world. As of December 9, 2021, the total confirmed infections have reached above 267 million, and the total death toll has exceeded 5.2 million worldwide.[1] With vaccines available for COVID-19, many countries have been conducting immunization campaigns hoping that herd immunity will be achieved when the majority of the population is vaccinated.[2] Current COVID-19 vaccines are targeting the Spike protein of SARS-CoV-2, the pathogen of COVID-19.[3] Spike is a weakly conserved protein in a highly mutable RNA virus. Although SARS-CoV-2 shares overall 82% genome sequence identity with SARS-CoV, Spike has only 76% protein sequence identity shared between two origins.[4] The highly mutable nature of Spike has also been corroborated by the continuous identification of new SARS-CoV-2 variants with Spike mutations.[5] The most notable are Alpha, Beta, Delta, and Omicron variants. Accumulated evidence has shown an attenuated activity of approved vaccines against some new SARS-CoV-2 variants.[6] Booster vaccines might be developed for new virus variants. However, the situation will likely turn into an incessant race between viral mutation and vaccine development. The focus on vaccine development that is preventative toward COVID-19 has largely obscured the development of targeted therapeutics that are needed for treating patients with severe symptoms. By targeting a conserved gene in SARS-CoV-2, a small-molecule antiviral will likely be more successful than a vaccine in both prevention and treatment since it is generally easier to manufacture, store, deliver, and administer a small-molecule antiviral than a vaccine.

One demonstrated drug target in SARS-CoV-2 is its main protease (MPro).[7,8] Unlike Spike that is highly mutable, MPro is highly conserved. Its 96% protein sequence identity shared between SARS-CoV and SARS-CoV-2 is much higher than the overall 82% genome sequence identity shared between the two viruses.[3] Much work has also been done in the development of MPro inhibitors.[9,10] A strategy that most researchers have been following in the development of MPro inhibitors is to synthesize an inhibitor, test its enzymatic inhibition, and then carry out its structural and antiviral analysis to obtain information for next-round optimization. For most medicinal chemists, the bottleneck in this drug discovery process is the antiviral assay that requires the use of a BSL3 facility and is often not accessible. The antiviral assay itself may also lead to misleading results about the real mechanism of an inhibitor. The life cycle of SARS-CoV-2 (Figure A) requires a number of proteases that are from either the host or the virus itself. It has been shown that transmembrane protease serine 2 (TMPRSS2) can prime Spike for interactions with the human cell host receptor ACE2 during the virus entry process.[11] Cathepsin L (CtsL) also potentiates the membrane fusion between SARS-CoV-2 with the endosome in infected cells.[12] It has also been suggested that other cathepsins such as cathepsin B (CtsB) serve a role in the SARS-CoV-2 entry.[13] After the SARS-CoV-2 genomic RNA is released into the host cytosol, it is translated by the host ribosome to form two large polypeptides, pp1a and pp1ab. The processing of pp1a and pp1ab to 16 nonstructural proteins (nsps) requires proteolytic functions of two internal protease fragments, nsp3 and nsp5 that are also called papain-like protease (PLPro) and main protease (MPro), respectively. Some nsps package into an RNA replicase complex that replicates both genomic and subgenomic RNAs. The translation of subgenomic RNAs leads to essential structural proteins for packaging new virions. Furin is a host protease that can hydrolyze Spike to prime it for new virion packaging and release.[14] Based on our current understanding of SARS-CoV-2 pathogenesis and replication, there are at least three host and two viral proteases serving critical roles in the SARS-CoV-2 life cycle. Inhibition of any of these enzymes will potentially cause a strong antiviral effect. The catalytic similarity between these enzymes makes it possible that a developed inhibitor is unselective toward these enzymes. MPro, PLPro, CtsB, and CtsL are cysteine proteases with a similar catalytic mechanism. TMPRSS2 and furin are serine proteases. Although serine proteases are mechanistically different from cysteine proteases, many currently developed MPro inhibitors have covalent warheads such as aldehyde and ketone that are prone to form covalent adducts with TMPRSS2 and furin as well.[15,16] All of these proteases are also localized in different parts of the host cell. Their inhibition requires different characteristics in inhibitors such as cellular permeability and pH sensitivity. A simple antiviral assay for an inhibitor will likely lead to a positive result that is not from the inhibition of MPro and therefore causes a misunderstanding that can be detrimental to further rounds of lead optimization. Therefore, an assay system that directly reflects MPro inhibition in the host cell is critical for both the assessment and optimization of MPro inhibitors. In the current work, we describe such a system and its application in the evaluation of a number of developed and repurposed MPro inhibitors.

Figure 1

Life cycle of SARS-CoV-2 and two assays for MPro-targeting antivirals. (A) Cartoon diagram illustrating the life cycle of SARS-CoV-2. Seven sequential steps are labeled in blue. Proteins that are labeled in pink are targets for the development of antivirals. TMPRSS2, CtsL, and furin are three host proteases that prime Spike for viral entry and new virion packaging. ACE2, angiotensin-converting enzyme 2; RdRp, RNA-dependent RNA polymerase. (B) Antiviral assay based on the inhibition of virus-induced CPE and cell death. (C) Antiviral assay based on the inhibition of MPro-induced apoptosis in host cells and the fluorescence of the expressed MPro-eGFP fusion protein.

Results

The Rationale and the Establishment of a Cellular MPro Inhibition Assay for MPI8

A typical antiviral assay for SARS-CoV-2 is its induced strong cytopathogenic effect (CPE) in host cells that can be quantified by counting viral plaques (Figure B). An MPro inhibitor with high cellular potency will suppress this virus-induced CPE and therefore lead to host cell survival. A good cellular MPro inhibition assay will need to provide results similar to this CPE suppression process. Our original design for a cellular MPro inhibition assay was to express MPro in host cells that is fused with an N-terminal cyan fluorescent protein (CFP) and a C-terminal yellow fluorescent protein (YFP) and characterize the inhibition of autocleavage of this fusion protein in the presence of an inhibitor. MPro natively cuts off its fused protein at the C-terminus. We put an MPro digestion site between CFP and MPro as well. CFP and YFP form a Förster resonance energy transfer (FRET) pair.[17] Without an inhibitor, both CFP and YFP will be cleaved from the fusion protein in host cells, which leads to no FRET signal. In the presence of a potent inhibitor, the fusion protein will be intact and emit strong FRET signals. However, transfection of 293T cells with pECFP-MPro-EYFP (Figure S1), a plasmid containing a gene coding the CFP-MPro-YFP fusion protein led to the death of most transfected cells (Figure S2). Repeating this transfection process all led to the exact same result. It is evident that MPro can exert acute toxicity to its human cell host. The same observation has been made by others and used to develop assays as well.[18] MPI8 is an MPro inhibitor that our lab developed previously.[15] Antiviral analysis indicated that MPI8 has potency to totally suppress SARS-CoV-2-induced CPE in ACE2+ A549 cells at 0.2 μM. Given its approved antiviral potency, we used MPI8 as a positive control molecule for the analysis of cellular MPro inhibition. To alleviate the toxicity that was induced by the expression of CFP-MPro-YFP, we cultured 293T cells that were transfected with pECFP-MPro-EYFP in media containing 10 μM MPI8. The presence of MPI8 reduced death of transfected cells sharply. Interestingly, the overall expressed fusion protein was also significantly improved, showing much enhanced, directly detected yellow fluorescence from YFP (Figure S2). This positive correlation between the expression of CFP-MPro-YFP and the survival of transfected cells is likely due to the shutting-down of translation by active MPro. Since the measurement of a cell survival-correlated fluorescence increase in the presence of an inhibitor is much simpler than the FRET characterization, we decided to adopt this new way to analyze the cellular potency of MPro inhibitors. To explore whether MPro expression is correlated to the death of SARS-CoV-2-infected cells, we used SARS-CoV-2 WA1/2020 to infect both Vero E6 and ACE2+ A549 cells; a strong CPE was observed 12 h after infection, which correlated with strong MPro expression detected by Western blot (Figure S3).

Since a FRET system is not necessary, we modified our plasmid to express an MPro-eGFP fusion protein (Figure C) in host cells that can be easily analyzed using fluorescent flow cytometry. The expression of MPro-eGFP in host cells will trigger cell death that leads to weak fluorescence. This process can be reversed by adding a potent inhibitor with cellular activity. In order to use eGFP fluorescence to accurately represent expressed MPro, we introduced a Q306G mutation in MPro to abolish its cleavage of the C-terminal eGFP. MPro requires a free N-terminal serine for strong activity. To achieve this, we built two constructs as shown in Figure A and Figure S4. The first construct pLVX-MPro-eGFP-1 encodes MPro-eGFP with an N-terminal methionine that relies on host methionine aminopeptidases for its cleavage. The second construct pLVX-MPro-eGFP-2 encodes MPro-eGFP containing a short N-terminal peptide that has an MPro cleavage site at the end for its autocatalytic release. The transfection of 293T cells with two constructs showed that pLVX-MPro-eGFP-2 led to more potent toxicity to cells, and this toxicity was effectively suppressed when we provided 10 μM MPI8 in the growth media (Figure B). Therefore, we selected pLVX-MPro-eGFP-2 for all of our following studies. We have noticed that 72 h provided optimal fluorescence detection indicating a slow turnover of MPro-eGFP. To demonstrate that cellular fluorescence is positively correlated to the concentration of the provided MPI8, we transfected 293T cells with pLVX-MPro-eGFP-2, grew transfected cells in the presence of four MPI8 concentrations (0, 20, 40, and 160 nM) for 72 h, and then sorted cells using fluorescent flow cytometry (Figure C). Both the number and intensity of fluorescent cells (FL1-A signal, >1 × 106) were positively dependent on the provided MPI8 concentration, indicating the feasibility of using the system to characterize the cellular potency of an MPro inhibitor. To demonstrate this feasibility, we transiently transfected 293T cells with pLVX-MPro-eGFP-2 and grew transfected cells in the presence of a cascade of MPI8 concentrations between 0.001 and 10 μM. After 72 h, we sorted cells according to their eGFP fluorescence intensity. Cells with an FL1-A signal above 1 × 106 were analyzed. We built a MATLAB script to calculate the average eGFP fluorescence intensity of all analyzed cells and plotted the average eGFP fluorescence intensity against the MPI8 concentration as shown in Figure D. The data showed obvious MPI8-induced saturation of MPro-eGFP expression and fit nicely to a three-parameter dose-dependent inhibition mechanism in GraphPad Prism 9 for IC50 determination. The determined cellular IC50 value of MPI8 is 31 nM. To confirm that cell survival was from the direct inhibition of MPro protease activity by MPI8, we constructed pLVX-MPro(C145S)-eGFP whose encoded MPro has its active site cysteine mutated to serine. Transfecting 293T cells with this construct led to strong MPro(C145S)-eGFP expression that was detected by both a fluorescence measurement and Western blot and low cell death regardless of whether 1 μM MPI8 was present (Figures S5 and S6). Similarly, an MPro-targeting siRNA significantly reduced cellular apoptosis in 293T cells that transiently expressed MPro-eGFP (Figure S7). These results confirmed that host cell death was due to the protease activity of MPro. To confirm that MPI8 does not inhibit caspases that serve functions in apoptosis, we used antimycin A to induce apoptosis in 293T cells and cultured treated cells with or without 1 μM MPI8. In both conditions, the detected apoptosis levels were not significantly different (Figure S8). Collectively, these results confirm that MPI8 directly inhibits the protease activity of MPro in 293T cells to cause overall cell survival and overexpression of MPro-eGFP.

Figure 2

Validation of transiently expressed MPro and its cellular toxicity for the analysis of cellular potency of MPro inhibitors. (A) Design of two MPro-eGFP fusions. (B) 293T cells transiently transfected with pLVX-MPro-eGFP-2 and grown in the absence or presence of 10 μM MPI8. (C) 293T cells that were transiently transfected with pLVX-MPro-eGFP-2 expressed MPro-eGFP correlated with the concentration of MPI8 in the growth media. (D) Cellular IC50 determination of MPI8. 293T cells were transfected with pLVX-MPro-eGFP-2 and grown in the presence of different concentrations of MPI8 for 72 h before their sorting using flow cytometry. The average fluorescence intensity for cells with FL1-A signal higher than 2 × 106 was determined and used to plot against the MPI8 concentration. Data were fitted to the three-parameter dose-dependent inhibition mechanism to determine the cellular IC50 value.

Since MPI8 is highly effective in inhibiting MPro in cells, we used it in combination with pLVX-MPro-eGFP-2 to make stable 293T cells that continuously expressed MPro-eGFP. Using this stable cell line, we characterized MPro-induced apoptosis that was detected by antiannexin. After we withdrew MPI8 from the growth media, a strong apoptotic effect started to show after 24 h and continued to increase (Figure S9). Since MPI8 is a reversible covalent inhibitor, the relatively long incubation time for the observation of apoptosis is likely due to its slow release from the MPro active site. Due to concerns about residual MPI8 and its potential slow release from MPro in stable cells, we chose to perform a cellular potency characterization of all MPro inhibitors by performing a transient transfection of 293T cells and then growth in the presence of different inhibitor concentrations.

MPI1–7, MPI9, GC376, and 11a

MPI8 was one of 9 β-(S-2-oxopyrrolidin-3-yl)alaninal (Opal)-based, reversible covalent MPro inhibitors (MPI1–9) that we previously developed (Figure A).[15] GC376 is a prodrug that dissociates quickly in water to release its Opal component.[19] 11a is another Opal-based, reversible covalent MPro inhibitor that was developed in 2020.[9] All 11 compounds showed high potency in inhibiting MPro in an enzymatic assay.[15] Besides MPI8, we tested the cellular potency of all other 10 Opal inhibitors in their cellular inhibition of MPro by following the exact same procedure that we did for MPI8. As shown in Figure A, all tested Opal inhibitors promoted cell survival and the expression of MPro-eGFP significantly at 10 μM. However, data collected at different concentrations showed that only three inhibitors, MPI5, MPI6, and MPI7, induced the saturation of MPro-eGFP expression at or below 10 μM. Determined IC50 values for MPI5, MPI6, and MPI7 are 0.66, 0.12, and 0.19 μM, respectively (Table ). Based on the collected data, MPI2–4, MPI9, GC376, and 11a have IC50 values higher than 2 μM, and MPI1 has an IC50 value higher than 10 μM.

Figure 3

Structures of inhibitors that were evaluated in their cellular inhibition of MPro. (A) Reversible covalent inhibitors designed for MPro. (B) Investigational covalent inhibitors that were developed for other targets. (C) Inhibitors that were identified via high-throughput screening. (D) FDA-approved medications that have been explored as MPro inhibitors. (E) Diaryl esters that have high potency to inhibit MPro.

Figure 4

Cellular potency of literature-reported MPro inhibitors. K777 is included as a potential MPro inhibitor.

Table 1

Determined Enzymatic and Cellular IC50 Values in Inhibiting SARS-CoV-2 MPro for Different Inhibitors

compound ID	enzymatic IC50 (μM)	cellular IC50 (μM)	cellular IC50 (μM) with CP-100356	antiviral EC50 (μM)	compound ID	enzymatic IC50 (μM)	cellular IC50 (μM)	cellular IC50 (μM) with CP-100356	antiviral EC50 (μM)
MPI1[15]	0.100 ± 0.023	>10	>2	>5	MG-132	3.9 ± 1.0[16]; 3.0 ± 0.2d	n.d.c,e
MPI2[15]	0.103 ± 0.014	>2	>2	>5	calpain inhibitor II[16]	0.97 ± 0.27	>10		2.07 ± 0.76a
MPI3[15]	0.0085 ± 0.0015	>2	>2	>5	calpain inhibitor XII	0.45 ± 0.06;[16] 0.82 ± 0.08d	>10		0.49 ± 0.18a
MPI4[15]	0.015 ± 0.005	>2	1.8 ± 0.01	>5	K777[22]	>100	n.d.		0.62a
MPI5[15]	0.033 ± 0.002	0.66 ± 0.15	0.58 ± 0.06	0.073 ± 0.007	carmofur	1.35 ± 0.04;[7,23] 0.20 ± 0.01d	n.d.		>100b
MPI6[15]	0.060 ± 0.004	0.12 ± 0.03	0.075 ± 0.008	0.21 ± 0.02	tideglusib[7]	1.55 ± 0.30; 2.8 ± 0.1d	n.d.
MPI7[15]	0.047 ± 0.003	0.19 ± 0.03	0.075 ± 0.006	0.17 ± 0.02	ebselen[7]	0.67 ± 0.09; 0.98 ± 0.01d	n.d.		4.67 ± 0.80a
MPI8[15]	0.105 ± 0.022	0.031 ± 0.002	0.039 ± 0.007	0.030 ± 0.003	disulfiram[7]	9.35 ± 0.18; 2.2 ± 0.2d	n.d.
MPI9[15]	0.056 ± 0.014	>2	>2		PX-12[7]	21.4 ± 7.1	>10c
GC376	0.030 ± 0.0086[15]	>2	2.2 ± 0.2	3.37 ± 1.68a,[16]/0.70[20]	bepridil[24]	72 ± 3	n.d.		0.46a
11a	0.053 ± 0.005;[4] 0.031 ± 0.003[15]	>2	1.4 ± 0.1	0.53 ± 0.01a	chloroquine[25,31]	3.9 ± 0.2 > 10d	n.d.		5.47a
boceprevir	4.2 ± 0.6[16]/8.0 ± 1.5;[20] 7.3 ± 2.3d	≫10		1.31 ± 0.58a,[16]/15.57[20]	hydroxychloroquine[25,31]	2.9 ± 0.3 > 10d	n.d.		0.72a
telaprevir[21]	15.3	≫10			10-1	0.040 ± 0.004	>10	>10
calpeptin[16]	10.7 ± 2.8	n.d.			10-2	0.068 ± 0.005	>10	>10
MG-115	3.1 ± 1.0;[16] 2.7 ± 0.1d	n.d.c			10-3	5.72 ± 0.43	>10	>10

Primary CPE assay.

Genomic RNA quantification.

Toxic at 10 μM.

Determined separately by us.

n.d.: not detected.

Boceprevir, Telaprevir, Calpeptin, MG-132, MG-115, Calpain Inhibitor II, Calpain Inhibitor XII, and K777

Drug repurposing research has led to the identification of a number of both FDA-approved and investigational drugs as MPro inhibitors. These include boceprevir, telaprevir, and calpain inhibitor XII that have an α-ketoamide moiety; and calpeptin, MG-132, and calpain inhibitor II that have an aldehyde for covalent interactions with the MPro active site cysteine.[16,20,21] Some of these compounds display potency in inhibiting SARS-CoV-2 replication in host cells as well. We proceeded to characterize the cellular potency of these inhibitors using our developed cellular assay. K777 is a known CtsL inhibitor with high potency in inhibiting SARS-CoV-2 replication in a human cell host.[22] It has a vinylsulfonate moiety. Due to its propensity to form a permanent covalent adduct with the MPro active site cysteine, we tested its cellular potency in inhibiting MPro as well. As shown in Figure B, calpeptin, MG115, MG132, telaprevir, and K777 displayed a close to undetectable cellular inhibition of MPro up to 10 μM; boceprevir and calpain inhibitor II displayed a close to undetectable cellular inhibition of MPro up to 2 μM and a very weak cellular inhibition of MPro at 10 μM. Calpeptin XII exhibited the highest cellular inhibition of MPro among this group of inhibitors, but it has low potency with an estimated IC50 value higher than 10 μM.

Carmofur, Tideglusib, Ebselen, Disulfiram, and PX-12

Drug repurposing research has also shown that carmofur, tideglusib, ebselen, disulfiram, and PX-12 can potently inhibit MPro.[7] Carmofur is an antineoplastic agent that generates a permanent thiocarbamate covalent adduct with the MPro active site cysteine.[23] All of the other four compounds are redox active for covalent conjugation with the MPro active site cysteine. We applied our cellular assay to these drugs as well. As shown in Figure C, except for PX-12 that weakly inhibited MPro in cells at 10 μM, the other four compounds showed an undetectable cellular inhibition of MPro at all tested inhibitor concentrations. Since all repurposed drugs in this section and the previous section displayed low cellular potency, we recharacterized most of them in an enzyme inhibition assay to confirm their in vitro MPro inhibition potency. Our determined IC50 values shown in Table and Figure S10 are not significantly different from literature-reported results.

Bepridil, Chloroquine, and Hydroxychloroquine

Using computational docking analysis in combination with experimental examination to guide drug repurposing for COVID-19, we previously showed that bepridil, an antianginal drug, inhibited MPro and had high potency in inhibiting SARS-CoV-2 replication in host cells.[24] To provide a full picture to understand the mechanism of bepridil in inhibiting SARS-CoV-2, we applied our cellular assay to bepridil. As shown in Figure D, bepridil displayed a very weak inhibition of MPro in cells up to 10 μM. A previous publication reported that chloroquine and hydroxychloroquine are potent inhibitors of MPro.[25] We applied our cellular assay to these two drugs. At all tested concentrations, both drugs displayed a close to undetectable promotion of MPro-eGFP expression, which indicates very low MPro inhibition from both drugs in cells. Using both a commercial and homemade substrate, we recharacterized MPro enzymatic inhibition by chloroquine and hydroxychloroquine. Our data (Figure S11) showed that MPro retained 84% activity at 16 μM chloroquine, and hydroxychloroquine did not inhibit MPro up to 16 μM.

Diaryl Esters 10-1, 10-2, and 10-3

Benzotriazole esters that were contaminants in a peptide library were accidentally discovered as potent inhibitors of SARS-CoV MPro.[26] Based on their inhibition mechanism, a number of diaryl esters were developed later as potent SARS-CoV MPro inhibitors.[27] To show whether similar compounds will also inhibit MPro of SARS-CoV-2, we synthesized diaryl esters 10-1, 10-2, and 10-3 and characterized their enzymatic inhibition IC50 values as 0.067, 0.038, and 7.6 μM, respectively (Figure S12). Using our cellular assay, we characterized all three compounds as well. As shown in Figure D, all three compounds display observable potency in inhibiting MPro to promote MPro-eGFP expression at 2 and 10 μM. Their cellular MPro inhibition IC50 values are estimated above 10 μM.

Effect of CP-100356 on the Cellular Potency of Peptide-Based MPro Inhibitors

CP-100356 is a high-affinity inhibitor of multidrug resistance protein (Mdr-1/gp), a protypical ABC transport that exports toxic substances from the inside of cells. A previous report showed that CP-100356 enhanced the antiviral potency of MPro inhibitors significantly.[28] To investigate whether CP-100356 improves the cellular MPro inhibition potency of Opal inhibitors, we recharacterized MPI1–9, GC376, and 11a using our cellular assay in the presence of 0.5 μM CP-100356 (Figure ). Except for MPI8 that showed an inhibition curve in the presence of CP-100356 very similar to that in the absence of CP-100356 and had a determined IC50 value of 39 nM, all other Opal inhibitors displayed a better cellular MPro inhibition curve. MPI5 and MPI6 have IC50 values (580 and 75 nM, respectively) in the presence of CP-100356 that are slightly lower than those in the absence of CP-100356. The highest cellular potency improvement that we observed among all compounds was for MPI7. It displayed an IC50 value (75 nM) in the presence of CP-100356 that is 60% lower than that in the presence of CP-100356. The cellular potency improvement for MPI4, GC376, and 11a in the presence of CP-100356 also led to characterizable IC50 values of 1.8, 2.2, and 1.4 μM, respectively. We did a similar test for 10-1, 10-2, and 10-3. Providing CP-100356 did not significantly change the cellular MPro inhibition for all three compounds at all tested concentrations.

Figure 5

Cellular potency of selected compounds in their inhibition of MPro in the presence of 0.5 μM CP-100356.

Determination of Antiviral EC50 Values for MPI1–8

Our previous antiviral assay for Opal inhibitors was based on the on–off observation of a CPE in Vero E6 and ACE2+ A549 cells. To quantify antiviral EC50 values of MPI1–8, we conducted plaque reduction neutralization tests of SARS-CoV-2 in Vero E6 cells in the presence of MPI1–8. we infected Vero E6 cells with SARS-CoV-2, grew infected cells in the presence of different inhibitor concentrations for 3 days, and then quantified SARS-CoV-2 plaque reduction. Based on SARS-CoV-2 plaque reduction in the presence of MPI1–8, we determined antiviral EC50 values for MPI1–8 as shown in Figure and Table . MPI1–4 displayed a low antiviral potency with estimated EC50 values above 5 μM, and MPI5–8 have EC50 values determined to be 0.073, 0.21, 0.17, and 0.030 μM, respectively.

Figure 6

Plaque reduction neutralization tests (PRNTs) of MPI5–8 on their inhibition of SARS-CoV-2 in Vero E6 cells. DMSO was used as a negative control.

Discussion

MPI1–9 were previously developed as potent MPro inhibitors. All showed enzymatic IC50 values around or below 100 nM (Table ). Among them, MPI3 has the highest potency with an IC50 value of 8.5 nM. However, a CPE-based antiviral assay in Vero E6 cells showed that MPI3 weakly inhibited SARS-CoV-2.[15] On the contrary, MPI8 that has an enzymatic IC50 value of 105 nM displayed the highest potency in inhibiting SARS-CoV-2. A separate antiviral assay in ACE2+ A549 cells showed that MPI8 inhibited the SARS-CoV-2-induced CPE completely at 200 nM MPI8. Overall, the antiviral potency of MPI1–9 based on the on–off observation of the CPE correlates with their cellular MPro inhibition potency that we have detected using the new cellular assay. In order to confirm that the determined cellular potency results correlate closely with antiviral effects, we quantified antiviral EC50 values for MPI1–8 in Vero E6 cells. Overall, the general trends of determined potency for MPI1–8 from two assays correlate well with each other, indicating that the developed cellular assay is valid in assessing the antiviral potency of MPro inhibitors if these inhibitors act on MPro alone. CP-100356 improved the cellular potency for most Opal-based inhibitors, although this improvement is as dramatic as reported in Hoffman et al.[28] Therefore, the main reason for the low cellular and antiviral potency of MPI3 and other Opal-based inhibitors might not be their active exportation from cells. Possible reasons that may contribute to the low antiviral and cellular potency for these compounds include their potential low cell permeability and proneness to both extracellular and intracellular proteolysis. Although MPI8 is not the most potent Opal-based inhibitor according to its in vitro enzymatic inhibition potency, it has the best antiviral and cellular potency. Its determined cellular IC50 is 31 nM, which is less than a third of its in vitro enzymatic IC50 value. A likely reason is the possible accumulation of MPI8 in cells, which needs to be investigated. Other Opal-based inhibitors with high cellular potency are MPI5, MPI6, and MPI7. All display cellular IC50 values below 1 μM. Among all 30 inhibitors that we have tested, MPI5–8 show the highest potency, which warrants their further investigation for possible use in treating COVID-19. As far as we know, MPI8 is the compound with the highest cellular MPro inhibition potency and the highest SARS-CoV-2 antiviral potency in Vero E6 cells. We urge its preclinical investigation for treating COVID-19.

GC376 is an investigational drug for treating feline infectious peritonitis, a lethal coronavirus disease in cats.[19] Anivive Lifesciences Inc. did clinical investigations to repurpose GC376 for the treatment of COVID-19 patients. Although GC376 has high in vitro enzymatic MPro inhibition potency with an IC50 value of 30 nM, it shows relatively weak cellular potency (IC50 > 2 μM). This weak cellular potency correlates with its antiviral potency that was determined with an EC50 value of 3.37 or 0.7 μM from two separate studies.[16,20] In comparison to MPI8, GC376 is almost 2 orders of magnitude less potent in cellular and antiviral potency. A low cellular permeability and stability likely contribute to this low cellular and antiviral potency. 11a is an MPro inhibitor that has an antiviral EC50 value of 0.53 μM which is not significantly different from those of MPI6 and MPI7.[9] However, its cellular potency is much weaker compared to MPI6 and MPI7. Its estimated cellular IC50 value is higher than 2 μM. It is likely that 11a may interfere with other critical process(es) in the SARS-CoV-2 life cycle to exert a potent antiviral effect, which needs to be explored.

Boceprevir and telaprevir are two drugs approved for treating hepatitis C virus infection. Both have shown potency to inhibit MPro enzymatically, and boceprevir has also been characterized in an antiviral assay to show an EC50 value of 1.31 μM.[16] However, both drugs display very weak potency in their cellular MPro inhibition tests. Since we detected very weak cellular potency for boceprevir at 10 μM, boceprevir must hit on other key step(s) in the SARS-CoV-2 pathogenesis and replication pathway to convene its high antiviral effect. An investigation in this possibility will likely lead to the discovery of novel target(s) for COVID-19 drug development. Other aldehyde- and ketone-based inhibitors we have tested include calpeptin, MG-132, MG-115, calpain inhibitor II, and calpain inhibitor XII. Except for calpain inhibitor XII that showed a weak cellular inhibition of MPro with an estimated IC50 value higher than 10 μM, all others exhibited a close to undetectable MPro inhibition in cells up to 10 μM. Both calpain inhibitor II and XII have demonstrated antiviral potency toward SARS-CoV-2 with an EC50 value of 2.07 and 0.49 μM, respectively. Based on our cellular potency analysis of the two compounds, it is clear that their antiviral potency is not primarily from the inhibition of MPro. Wang et al. have explored compounds with dual functions to inhibit both MPro and host calpains/cathepsins as antivirals for SARS-CoV-2.[29] These compounds include calpain inhibitor II and XII. As such, they likely inhibit host proteases to cause potent antiviral effects. K777 weakly inhibited MPro in a kinetic assay but potently inhibited SARS-CoV-2 in an antiviral assay.[22] It showed undetectable potency in our cellular assay, which confirms that it must target other key process(es) in the SARS-CoV-2 life cycle.

Carmofur, tideglusib, ebselen, disulfiram, and PX-12 were discovered as MPro inhibitors from high-throughput screening. Although carmofur has an enzymatic IC50 value of 1.35 μM and generates a permanent covalent adduct with the MPro active site cysteine by forming a thiocarbamate, it showed undetectable cellular potency up to 10 μM in our cellular assay. This observation correlates well with its low antiviral potency.[23] The high chemical reactivity of carmofur likely contributes to its low cellular and antiviral potency. Tideglusib, ebselen, disulfiram, and PX-12 are redox-active compounds that can form covalent adducts with the MPro active site cysteine. Except for PX-12 that showed weak cellular potency at 10 μM, the other three drugs exhibited undetectable cellular potency up to 10 μM. Among the four compounds, only ebselen has been examined in an antiviral assay.[7] It has a determined EC50 value of 4.67 μM. Since ebselen showed undetectable cellular MPro inhibition up to 10 μM, its antiviral potency must be from its interference with other key process(es) in the SARS-CoV-2 life cycle. The revelation of the SARS-CoV-2 inhibition mechanism by ebselen will likely lead to the discovery of novel drug target(s) for COVID-19.

Bepridil is an MPro inhibitor with an enzymatic IC50 value of 72 μM but a much lower antiviral EC50 value of 0.46 μM in ACE2+ A549 cells. Bepridil is known to inhibit other human viral pathogens as well.[30] We detected a close to undetectable cellular MPro inhibition potency for bepridil up to 10 μM. This correlates with its relatively high enzymatic IC50 value. Therefore, bepridil must use a mechanism different from the inhibition of MPro to convene its high antiviral potency. This needs to be investigated. Chloroquine and hydroxychloroquine are two repurposed drugs for COVID-19 with demonstrated antiviral EC50 values of 5.47 and 0.72 μM, respectively.[31] Although TMPRSS2 was shown as a possible target of chloroquine and hydroxychloroquine,[32] a previous report showed that chloroquine and hydroxychloroquine potently inhibited MPro in an enzyme inhibition assay.[25] We tested both drugs using the new cellular assay but revealed close to undetectable cellular MPro inhibition up to 10 μM for both drugs. We recharacterized the enzymatic inhibition of MPro by both drugs as well. However, we were not able to detect any MPro inhibition by hydroxychloroquine up to 16 μM, and chloroquine exhibited very weak inhibition of MPro at 16 μM. Based on our cellular data, enzymatic inhibition data, and data from a separate study,[33] we are confident that both chloroquine and hydroxychloroquine do not potently inhibit MPro. Their antiviral activities are from different mechanism(s).

10-1, 10-2, and 10-3 are three diaryl esters in which 10-1 and 10-2 displayed high potency in inhibiting MPro enzymatically. All three compounds displayed a significant cellular MPro inhibition potency at 10 μM, but their potency is much lower than those of MPI5–8. Although 10-3 has a much weaker enzymatic inhibition potency than 10-1 and 10-2, its cellular potency is slightly better than those from 10-1 and 10-2. A likely explanation is that 10-3 is more stable than 10-1 and 10-2, which leads to a longer cellular time to convene its cellular MPro inhibition potency. Therefore, we recommend balancing cellular stability and enzymatic inhibition potency for the future development of diaryl esters as MPro inhibitors to achieve optimal antiviral effects.

As a protypical ABC transporter inhibitor, CP-100356 can potentially improve the intracellular accumulation of exogenous toxic molecules in cells. Providing CP-100356 improved the cellular activity for all Opal-based inhibitors except MPI8, albeit the improvement is not as great as what was reported for PF-00835231.[28] This is likely due to a low expression of Mdr-1/gp in 293T cells. Since CP-100356 is not an approved drug, its use in combination with an MPro inhibitor for COVID-19 treatment will face significant hurdles in clearing out toxicity and other clinical concerns. MPI8 showed a similar cellular potency in the presence and absence of CP-100356, suggesting MPI8’s high propensity to accumulate inside cells. This explains our observation that the determined cellular MPro inhibition IC50 value for MPI8 was 3-fold less than its determined enzymatic inhibition IC50 value. Data related to the use of CP-100356 support that MPI8 is optimal for cellular MPro inhibition. As the compound with the highest cellular and antiviral potency among all of the literature and new compounds that we have tested in the current study, we urge MPI8 for further investigations in treating COVID-19.

Conclusion

We have developed a cellular assay to determine the cellular potency of SARS-CoV-2 MPro inhibitors. Unlike an antiviral assay in which the interference of any key step in the SARS-CoV-2 life cycle may lead to a strong antiviral effect, this new cellular assay reveals only cellular MPro inhibition potency of a compound. It provides precise information that reflects real MPro inhibition in cells. Using this assay, we characterized 30 MPro inhibitors. Our data indicated that 11a, boceprevir, ebselen, calpain inhibitor II, calpain inhibitor XII, K777, and bepridil likely interfere with key processes other than the MPro catalysis in the SARS-CoV-2 pathogenesis and replication pathways to convene their strong antiviral effects. Our results also revealed that MPI8 has the highest cellular potency among all compounds that were tested. It has a cellular MPro inhibition IC50 value of 31 nM. MPI8 has been recently shown with dual inhibition effects against human cathepsin L but not other tested human proteases.[34] As the compound with the highest antiviral potency with an EC50 value of 30 nM, we cautiously advocate preclinical tests of MPI8 as a COVID-19 treatment.

Methods

Chemicals, Reagents, and Cell Lines from Commercial Providers

We purchased HEK293T/17 cells from ATCC; DMEM with high glucose with GlutaMAX supplement, fetal bovine serum, 0.25% trypsin-EDTA, phenol red, puromycin, Lipofectamine 3000, and dimethyl sulfoxide from Thermo Fisher Scientific; linear polyethylenimine MW 25000 from Polysciences; RealTime-Glo annexin V apoptosis and a necrosis assay kit from Promega; an EndoFree plasmid DNA midi kit from Omega Biotek; antimycin a from Sigma-Aldrich; GC376 from Selleck Chem; boceprevir, calpeptin, MG-132, telaprevir, and carmofur from MedChemExpress; ebselen from TCI; calpain inhibitors II and XII from Santa Cruz Biotechnology; MG-115 From Abcam; tideglusib, disulfiram, and PX-12 from Cayman Chemical; chloroquine diphosphate from Alfa Aesar; hydroxychloroquine sulfate from Acros Organics; and a fluorogenic MPro substrate DABCYL-Lys-Thr-Ser-Ala-Val-Leu-Gln-Ser-Gly-Phe-Arg-Lys-Met-Glu-EDANS termed as Sub3 from Bachem. K777 was a gift from Prof. Thomas Meek at Texas A&M University. The syntheses of MPI1–9 and 11a were shown in a previous publication.[15]

Plasmid Construction

We amplified MPro with an N-terminal KTSAVLQ sequence using primers FRET-MPro-for and FRET-MPro-rev (Table S1) and cloned it into the pECFP-18aa-EYFP plasmid (Addgene, 109330) between XhoI and HindIII restriction sites to afford pECFP-MPro-EYFP. To construct pLVX-MPro-eGFP-1, we amplified MPro with an N-terminal methionine using primers XbaI-MPro-f and MPro-HindIII-r (Table S1) and eGFP using primers HindIII-eGFP-f and eGFP-NotI-r. We digested the MPro fragment using XbaI and HindIII-HF restriction enzymes and the eGFP fragment using HindIII-HF and NotI restriction enzymes. We ligated the two digested fragments together with the pLVX-EF1a-IRES-Puro vector (Takara Bio 631988) that was digested at XbaI and NotI restriction sites. To facilitate the ligation of three fragments, we used a ratio of MPro, eGFP, and pLVX-EF1a-IRES-Puro digested products of 3:3:1. We constructed pLVX-MPro-eGFP-2 in the same way as pLVX-MPro-eGFP-1 except that we amplified the MPro fragment using primers XbaI-Cut-Mpro-f and MPro-HindIII-r (Table S1). XbaI-Cut-MPro-f encodes an MKTSAVLQ sequence for its integration to the MProN-terminus. To construct pLVX-MPro(C145S)-eGFP, two primers MPro-C145S-f and MPro-C145S-r (Table S1) were used to carry out a site-directed mutagenesis of pLVX-MPro-eGFP. All plasmids were sequence confirmed by Sanger sequencing.

Transfection and MPI8 Inhibition Tests Using pECFP-MPro-EYFP

We grew 293T cells to 60% confluency and then transfected them with pECFP-MPro-EYFP using Lipofectamine 3000. We added 10 μM MPI8 at the same time of transfection. After 72 h of incubation, cells were collected and analyzed by a flow cytometer as well as fluorescence microscopy. In order to obtain high-definition images, poly-d-lysine-coated glass bottom plates from Mattek were used for microimaging.

Transfection and Inhibition Tests Using pLVX-MPro-eGFP-1 and pLVX-MPro-eGFP-2

We grew 293T cells to 60% confluency and transfected them with pLVX-MPro-eGFP-1 or pLVX-MPro-eGFP-2 using Lipofectamine 3000. We added different concentrations of MPI8 from the nM to μM level at the same time of transfection. After 72 h of incubation, we analyzed the transfected 293T cells using flow cytometry to determine fluorescent cell numbers and the eGFP fluorescence intensity.

Establishment of 293T Cells Stably Expressing MPro-eGFP

To establish a 293T cell line that stably expresses MPro-eGFP, we packaged lentivirus particles using the pLVX-MPro-eGFP-2 plasmid. Briefly, we transfected 293T cells at 90% confluency with three plasmids including pLVX-MPro-eGFP-2, pMD2.G, and psPAX2 using 30 mg/mL polyethylenimine. We collected supernatants at 48 and 72 h after transfection separately. We concentrated and collected lentiviral particles from collected supernatant using ultracentrifugation. We then transduced fresh 293T cells using the collected lentivirus particles. After 48 h of transduction, we added puromycin to the culture media to a final concentration of 2 μg/mL. We gradually raised the puromycin concentration 10 μg/mL in 2 weeks. The final stable cells were maintained in media containing 10 μg/mL puromycin.

Apoptosis Analysis

We performed the apoptosis analysis of the MPro stable cells and cells transiently transfected with the pLVX-MPro-eGFP-2 plasmid using the RealTime-Glo annexin V apoptosis and necrosis assay kits from Promega. The cells were maintained in a high glucose DMEM medium supplemented with 10% fetal bovine serum (FBS), plated with a cell density of 5 × 105 cells/mL. We set up five groups of experiments including (1) HEK 293T/17, (2) HEK 293T/17 + MPI8 (1 μM), (3) HEK 293T/17 cells stably expressing MPro-eGFP, (4) HEK 293T/17 cells stably expressing MPro-eGFP + MPI8 (1 μM), and (5) HEK 293T/17 or HEK 293T/17 cells stably expressing MPro-eGFP + antimycin A (1 μM). Each experiment was repeated 5 times. The assay was performed according to the instructor’s protocol. Chemiluminescence was recorded at 12, 24, 36, 48, 60, and 72 h after plating the cells. The luminescence readings were normalized using HEK 293T/17 as a negative control.

Cellular MPro Inhibition Analysis for 29 Selected Compounds

We grew HEK 293T/17 cells in high-glucose DMEM with GlutaMAX supplement and 10% fetal bovine serum in 10 cm culture plates under 37 °C and 5% CO2 to ∼80–90% and then transfected cells with the pLVX-MPro-eGFP-2 plasmid. For each transfection, we used 30 mg/mL polyethylenimine and a total of 8 μg of the plasmid in 500 μL of the opti-MEM medium. We incubated cells with transfecting reagents overnight. On the second day, we removed the medium, washed cells with a PBS buffer, digested them with 0.05% trypsin-EDTA, resuspended the cells in the original growth media, adjusted the cell density to 5 × 105 cells/mL, provided 500 μL of suspended cells in the growth media to each well of a 48-well plate, and then added 100 μL of a drug solution to the growth media. These cells were then incubated under 37 °C and 5% CO2 for 72 h before their flow cytometry analysis.

Data Collection, Processing, and Analysis

The cell was incubated with various concentrations of drugs at 37 °C for 3 days. After 3 days of incubation, we removed the media and then washed cells with 500 μL of PBS to remove dead cells. Cells were then trypsinized and spun down at 800 rpm for 5 min. We removed the supernatant and suspended the cell pellets in 200 μL of PBS. The fluorescence of each cell sample was collected by a Cytoflex Beckman flow cytometer based on the size scatters (SSC-A and SSC-H) and forward scatters (FSC-A). We gated cells based on SSC-A and FSC-A and then with SSC-A and SSC-H. The eGFP fluorescence was excited by a blue laser (488 nm), and cells were collected at FITC-A (525 nm). After collecting the data, we analyzed and transferred the data to csv files containing information on each cell sample. We then analyzed these files using a self-written MATLAB program for massive data processing. We sorted the FITC-A column from smallest to largest. A 106 cutoff was set to separate the column into two groups, larger as positive and smaller as negative. We integrated the positive group and divided the total integrated fluorescence intensity by the total positive cell counts as Flu. Int. shown in all graphs. The standard deviation of positive fluorescence was also calculated. It was then plotted and fitted nonlinearly with an agonist curve (three parameters) against drug concentrations in the program Prism 9 from GraphPad for IC50 determination.

Kinetic Recharacterization of Chloroquine and Hydroxychloroquine

We prepared 10 mM stock solutions of hydroxychloroquine and chloroquine in a PBS buffer and carried out IC50 assays for both hydroxychloroquine and chloroquine by measuring activities of 50 nM MPro against a concentration range of 0–16 μM hydroxychloroquine and chloroquine. Serial dilutions of hydroxychloroquine and chloroquine were carried out in the assay buffer by keeping the PBS concentration the same. First, 100 nM MPro in the assay buffer (10 mM phosphate, 10 mM NaCl, 0.5 mM EDTA, pH 7.6) was treated with 2 times the working concentration of hydroxychloroquine and chloroquine at 37 °C for 30 min. Then, a 20 μM concentration of the fluorogenic MPro substrate Sub3 (prepared from a 1 mM stock solution of the dye in DMSO) in the assay buffer was added to the reaction mixture to a final concentration of 10 μM. Immediately after the addition of the substrate, we started to monitor the reaction in a BioTek Neo2 plate reader with an excitation wavelength at 336 nM and emission detection at 490 nM. Initial product formation slopes at the first 5 min were calculated by simple linear regression, and data were plotted in GraphPad Prism 9.

Synthesis of 5-Chloropyridin-3-yl 1H-Indole-7-carboxylate (10-1)

To a solution of 5-chloropyridin-3-ol (1 mmol, 130 mg) and 1H-indole-7-carboxylic acid in anhydrous dichloromethane (DCM), we added DMAP (0.1 mmol, 12 mg) and EDC (1.2 mmol, 230 mg). The resulting solution was stirred at room temperature overnight. Then, the reaction mixture was evaporated in vacuo, and the residue was purified with flash chromatography to afford 10-1 as a white solid (210 mg, 77%). 1H NMR (400 MHz, DMSO-d6): δ 11.34 (s, 1 H), 8.65 (dd, J = 10.8, 2.2 Hz, 2 H), 8.19 (t, J = 2.2 Hz, 1 H), 8.03–7.92 (m, 2 H), 7.47 (t, J = 2.9 Hz, 1 H), 7.23 (t, J = 7.7 Hz, 1 H), 6.65 (dd, J = 3.1, 1.9 Hz, 1 H). 13C NMR (101 MHz, DMSO-d6): δ 164.7, 147.9, 146.1, 143.0, 134.9, 131.2, 131.0, 130.2, 128.0, 127.9, 125.3, 119.2, 111.3, 102.5. ESI-HRMS: calcd for C14H10ClN2O2+, 273.0425; found, 273.0420.

Synthesis of 5-Chloropyridin-3-yl 1H-Indole-4-carboxylate (10-2)

To a solution of 5-chloropyridin-3-ol (1 mmol, 130 mg) and 1H-indole-4-carboxylic acid in anhydrous DCM, we added DMAP (0.1 mmol, 12 mg) and EDC (1.2 mmol, 230 mg). The resulting solution was stirred at room temperature overnight. Then, the reaction mixture was evaporated in vacuo, and the residue was purified with flash chromatography to afford 10-2 as a white solid (220 mg, 80%). 1H NMR (400 MHz, CDCl3): δ 8.72 (s, 1 H), 8.53 (dd, J = 7.2, 2.2 Hz, 2 H), 8.10 (dd, J = 7.5, 0.9 Hz, 1 H), 7.75 (t, J = 2.2 Hz, 1 H), 7.71 (dt, J = 8.1, 1.0 Hz, 1 H), 7.43 (t, J = 2.9 Hz, 1 H), 7.32 (t, J = 7.8 Hz, 1 H), 7.23 (ddd, J = 3.2, 2.1, 0.9 Hz, 1 H). 13C NMR (101 MHz, CDCl3): δ 165.0, 147.8, 145.7, 141.7, 136.7, 131.9, 130.0, 128.0, 127.2, 124.6, 121.3, 119.3, 117.4, 103.8. ESI-HRMS: calcd for C14H10ClN2O2+, 273.0425; found, 273.0420.

Synthesis of 5-Chloropyridin-3-yl 1H-Indole-3-carboxylate (10-3)

To a solution of 5-chloropyridin-3-ol (1 mmol, 130 mg) and 1H-indole-3-carboxylic acid in anhydrous DCM, we added DMAP (0.1 mmol, 12 mg) and EDC (1.2 mmol, 230 mg). The resulting solution was stirred at room temperature overnight. Then, the reaction mixture was evaporated in vacuo, and the residue was purified with flash chromatography to afford 10-3 as a white solid (190 mg, 69%). 1H NMR (400 MHz, DMSO-d6): δ 12.27 (s, 1 H), 8.58 (dd, J = 2.3, 1.0 Hz, 2 H), 8.40 (s, 1 H), 8.08 (t, J = 2.2 Hz, 1 H), 8.06–8.00 (m, 1 H), 7.60–7.51 (m, 1 H), 7.31–7.22 (m, 2 H). 13C NMR (101 MHz, DMSO-d6): δ 162.3, 148.0, 145.6, 142.8, 137.0, 135.1, 131.2, 130.8, 126.2, 123.4, 122.4, 120.8, 113.2, 104.8. ESI-HRMS: calcd for C14H10ClN2O2+, 273.0425; found, 273.0420.

Kinetic Characterization of 10-1, 10-2, and 10-3 in Inhibiting MPro

We performed MPro inhibition assays of 10-1, 10-2, and 10-3 according to the same procedure used for the kinetic characterization of hydroxychloroquine and chloroquine.

Characterization of the Cellular Potency of MPI1–9, GC376, 11a, 10-1, 10-2, and 10-3 in the presence of CP-100356

All cellular MPro inhibition assays for these 14 compounds were repeated with the addition of CP-100356 in DMSO to a final concentration of 0.5 μM. The overall assay process and analysis were identical to the assays without CP-100356.

Plaque Reduction Neutralization Tests of SARS-CoV-2 by MPI5–8

We seeded 18 × 103 Vero cells per well in flat-bottom 96-well plates in a total volume of 200 μL of a culturing medium (DMEM + 10% FBS + glutamine) and incubated cells overnight at 37 °C and under 5% CO2. The next day, we titrated compounds in separate round-bottom 96-well plates using the culturing medium. We then discarded the original medium used for cell culturing and replaced it with 50 μL of compound-containing media from round-bottom plates. We incubated cells for 2 h at 36 °C and under 5% CO2. After incubation, we added 1000 PFU/50 μL of SARS-COV-2 (USA-WA1/2020) to each well and incubated the plate at 36 °C and under 5% CO2 for 1 h. After incubation, we added 100 μL of overlay (1:1 of 2% methylcellulose and the culture medium) to each well. We incubated plates for 3 days at 36 °C and under 5% CO2. Staining was performed by discarding the supernatant, fixing the plates with 4% paraformaldehyde in the PBS buffer for 30 min, and staining with crystal violet. Plaques were then counted.

Safety Statement

No unexpected or unusually high safety hazards were encountered.