SARS-CoV-2 RdRp Inhibitors Selected from a Cell-Based SARS-CoV-2 RdRp Activity Assay System

The COVID-19 pandemic, caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), urgently needs effective prophylactic and therapeutic drugs. RNA-dependent RNA polymerase (RdRp), essential for replicating and transcribing a viral RNA genome, is highly conserved in coronaviruses; thus, it is a potential target for inhibiting coronavirus infection. In this study, we generated the cell-based SARS-CoV-2 RdRp activity assay system by modifying a previously reported cell-based MERS-CoV RdRp activity assay system to screen for SARS-CoV-2 RdRp inhibitors. The assay system consisted of an expression plasmid encoding SARS-CoV-2 RdRp and an RdRp activity reporter plasmid. RdRp activity in the cells could be conveniently detected by luminescence after transfection. We confirmed that SARS-CoV-2 RdRp replicated double-stranded RNA using immunofluorescence staining and the inhibition of RdRp activity by remdesivir and lycorine using this system. Moreover, the Z-factor of this system was calculated to be 0.798, suggesting the reproducibility and reliability of the high-throughput screening system. Finally, we screened nucleoside and nucleotide analogs and identified adefovir dipivoxil, emtricitabine, telbivudine, entecavir hydrate, moroxydine and rifampin as novel SARS-CoV-2 RdRp inhibitors and therapeutic candidates for COVID-19 This system provides an effective high-throughput screening system platform for developing potential prophylactic and therapeutic drugs for COVID-19 and emerging coronavirus infections.


Introduction
The coronavirus disease 2019 (COVID-19) pandemic, caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), broke out in December 2019. As of July 2021, COVID-19 has resulted in approximately 190 million confirmed cases with a 2% death rate, according to the World Health Organization [1].
Bonferroni's multiple comparison test. The non-linear regression analysis of IC50 was conducted using GraphPad Prism ® 9.1.2 (GraphPad Software Inc., San Diego, CA, USA).
Then, the replicated sense-oriented NLuc RNA was translated. Therefore, NLuc activity indicated the activity of SARS-CoV-2 RdRp, whereas FLuc activity acted as an internal control.

Western Blot Assay
HEK293T cells were transfected with the indicated plasmids for 24 h and lysed in the Glo Lysis Buffer (Promega Corporation). The lysates were separated on a gradient gel (Bio-Rad Laboratories, Hercules, CA, USA) and transferred to a nitrocellulose membrane (Bio-Rad Laboratories). The membrane was incubated with antibodies against Flag (Cat no. ab125243, Lot no. GR3348594-1, Abcam, Cambridge, UK) or β-actin (Cat no. #3700, Lot no. 5, Cell Signaling Technology, Danvers, MA, USA), then with an HRP-conjugated secondary antibody (Cat no. ab6728, Lot no. GR3200472-2, Abcam) and detected with chemiluminescence substrates (Thermo Fisher Scientific, Waltham, MA, USA) using the ChemiDoc TM Touch Imaging System (Bio-Rad).

Immunofluorescence Staining Assay
HEK293T cells were fixed with 4% paraformaldehyde and permeabilized with 0.2% Triton X-100 in PBS. After blocking with 3% bovine serum albumin, the cells were incubated with anti-double-stranded RNA (dsRNA) antibody K1 (Cat. No. 10020200, Batch no. K1-1715, Scicons, Susteren, the Netherlands) and then with the AlexaFluor555 conjugated anti-mouse immunoglobulin G (Thermo Fisher). The labeled cells were mounted on slides with the SlowFade Gold anti-fade reagent with DAPI (Invitrogen) and visualized by fluorescence microscopy (Olympus Corporation, Tokyo, Japan) and the CellSense program (Olympus).

Statistical Analysis
The data were presented as mean ± standard error of the mean. Statistical comparisons were conducted using one-or two-way analysis of variance (ANOVA) followed by Bonferroni's multiple comparison test. The non-linear regression analysis of IC 50 was conducted using GraphPad Prism ® 9.1.2 (GraphPad Software Inc., San Diego, CA, USA).
In addition, we constructed a plasmid encoding the SARS-CoV-2 nsp12/RdRp human codon-optimized sequence and tagged an N-terminal or C-terminal Flag to compare whether the N-term or C-term Flag tag could interrupt the RNA polymerase activity. At first, we confirmed the production of RdRp protein by the transfection of pCI-SARS2-nsp12-N-term Flag (pCI-SARS2 nsp12N) and pCI-SARS2-nsp12-C-term Flag (pCI-SARS2 nsp12C) using western blotting with an anti-Flag antibody ( Figure 2A). pCI-SARS2-nsp12N or pCI-SARS2-nsp12C were transfected dose-dependently with the p(+)FLuc-(−)UTR-NLuc reporter plasmid to test the functional activity of C-term or N-term Flag-tagged SARS-CoV-2 RdRp. We found that the NLuc values were dose-dependently enhanced by the increasing concentrations of pCI-SARS2-nsp12N or pCI-SARS2-nsp12C. The relative NLuc values of the cells expressing SARS nsp12N and nsp12C were comparable, suggesting that the Flag tag did not disrupt the SARS-CoV-2 RdRp activity ( Figure 2B). The relative NLuc value of the cells expressing SARS2-nsp12N was 1.1-fold higher than that of the cells expressing SARS2-nsp12C, at 2.7 and 2.4-fold at 120 ng of plasmid, respectively. So, we selected pCI-SARS2 nsp12N for further study.

Effect of Accessory Proteins nsp7 and nsp8 SARS-CoV-2 RdRp Activity
Among the SARS-CoV viral proteins, nsp7 and nsp8 have been reported as co-factors of RdRp [15][16][17][18]. Therefore, we tested the effect of nsp7 and nsp8 proteins on RdRp activity in this cell-based assay system. We generated the following three types of plasmids expressing the nsp7 and nsp8 genes: plasmids with a C-term Flag tag (pCI-SARS2-nsp7C and pCI-SARS2-nsp8C) or without Flag tag (pCI-SARS2-nsp7 and pCI-SARS2-nsp8) and a plasmid containing the internal ribosome entry site (IRES) between nsp8 and nsp7 (pCI-SARS2-nsp8-IRES-nsp7). The encoded proteins, C-term Flag tag nsp7 protein at 10 kDa, C-term Flag tag nsp8 protein at 22 kDa and N-term Flag tag nsp12 at 102 kDa, were detected in the HEK293T cells after transfection with pCI-SARS2-nsp7C, pCI-SARS2-nsp8C and pCI-SARS-nsp12N using western blotting ( Figure 3A).

Inhibition of SARS-CoV-2 RdRp Activity by Remdesivir and Lycorine
We further verified the cell-based system using remdesivir. Remdesivir is an adenosine analog that inhibits RdRp activity and coronavirus infection; it is FDA-approved for treating COVID-19 patients. Thus, it can be used as a positive control of the inhibition of RdRp activity [19,20]. When we treated the cells transfected with pCI-SARS2-nsp12N and the p(+)FLuc-(−)UTR-NLuc reporter plasmid with remdesivir at the indicated concentra-tions, the activity of SARS-CoV-2 RdRp, indicated by the relative Nano-luciferase activity, was decreased in a dose-dependent manner. Meanwhile, the activity of FLuc, acting as an internal control, was maintained. Thus, the IC 50 of remdesivir was calculated to be 2.585 ± 0.273 µM ( Figure 4A and Table 1).  The IC50 values were calculated using non-linear regression analysis (right graph). The data were representative of at least three independent experiments and presented as mean ± standard error of the mean.

Inhibition of SARS-CoV RdRp Activity by Nucleos(t)ide
We screened nucleos(t)ide analogs using the cell-based SARS-CoV-2 RdRp activity assay system to identify inhibitors of SARS-CoV-2 RdRp activity. Adefovir dipivoxil is a nucleoside analog inhibiting the reverse transcriptase activity of HBV; it is FDA-approved for chronic hepatitis B [25]. When we applied adefovir dipivoxil to our cell-based system, it dose-dependently reduced SARS-CoV-2 RdRp activity with an IC50 value of 3.785 ± 0.866 μM ( Figure 5A). Its inhibitory effect on SARS-CoV RdRp activity was comparable to that of remdesivir. NLuc activities were measured and NLuc/FLuc ratios were determined. Statistical comparisons were conducted using one-way analysis of variance (ANOVA) followed by Bonferroni's multiple comparison test. * p < 0.05; *** p < 0.001; **** p < 0.0001 vs. 0 µM (left graph). The IC 50 values were calculated using non-linear regression analysis (right graph). The data were representative of at least three independent experiments and presented as mean ± standard error of the mean. Recently, we reported a natural alkaloid, lycorine, as a broad-spectrum inhibitor of coronavirus infections and a MERS-CoV RdRp inhibitor, using a cell-based MERS-CoV RdRp assay system [21]. We examined whether lycorine could inhibit SARS-CoV-2 RdRp activity by treating the cells transfected with p(+)FLuc-(−)UTR-NLuc and pCI-SARS2-nsp12N for 15 h with a lycorine and measured the SARS-CoV-2 RdRp activity. Lycorine dose-dependently reduced the NLuc activity, whereas the FLuc value remained unchanged. Lycorine completely inhibited SARS-CoV-2 RdRp activity at 4.4 µM and the IC 50 was calculated to be 1.465 ± 0.033 µM, suggesting that lycorine inhibited SARS-CoV-2 RdRp activity more effectively than remdesivir ( Figure 4B).
We tested another natural compound, cepharanthine, which was also reported to inhibit HCoV-OC43 and SARS-CoV-2 coronavirus infections [22,23]. Cepharanthine was recently suggested by virtual screening to bind at the interface active pockets of SARS-CoV-2 RdRp, nsp7 and nsp8 [24]. So, we examined the effects of cepharanthine on SARS-CoV-2 RdRp activity with or without nsp7 and nsp8 using this system. Cepharanthine treatment did not affect the activity of SARS-CoV-2 RdRp in this cell-based system ( Figure 4C and Supplementary Figure S1). Therefore, we confirmed the inhibitory effects of remdesivir and lycorine, but not cepharanthine, on SARS-CoV-2 RdRp activity by the cell-based SARS-CoV-2 RdRp activity assay system.

Inhibition of SARS-CoV RdRp Activity by Nucleos(t)ide
We screened nucleos(t)ide analogs using the cell-based SARS-CoV-2 RdRp activity assay system to identify inhibitors of SARS-CoV-2 RdRp activity. Adefovir dipivoxil is a nucleoside analog inhibiting the reverse transcriptase activity of HBV; it is FDAapproved for chronic hepatitis B [25]. When we applied adefovir dipivoxil to our cellbased system, it dose-dependently reduced SARS-CoV-2 RdRp activity with an IC 50 value of 3.785 ± 0.866 µM ( Figure 5A). Its inhibitory effect on SARS-CoV RdRp activity was comparable to that of remdesivir. SARS-CoV-2 RdRp with an IC50 of 48.929 ± 14.370 μM ( Figure 5E). In addition, rifampin, also known as rifampicin and an antibiotic for tuberculosis [29], displayed an inhibitory effect on SARS-CoV-2 RdRp activity with an IC50 of 49.434 ± 4.020 μM ( Figure 5F). Therefore, our cell-based SARS-CoV-2 RdRp activity assay system has identified these nucleos(t)ide analogs as inhibitors of SARS-CoV-2 RdRp activity, among them, adefovir dipivoxil was the most effective inhibitor. NLuc activities were measured to determine the NLuc/FLuc ratios. Statistical comparisons were conducted using one-way analysis of variance (ANOVA) followed by Bonferroni's multiple comparison test. * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001 vs. 0 µM (upper graph). The IC 50 values were calculated using non-linear regression analysis (lower graph). The data were representative of at least three independent experiments and presented as mean ± standard error of the mean.
Emtricitabine is known to inhibit the activity of human immunodeficiency virus reverse transcriptase via its incorporation into the DNA, terminating the DNA chain elongation [26]. We found that emtricitabine effectively inhibited SARS-CoV-2 RdRp activity with the IC 50 value of 15.375 ± 3.602 µM in this assay system ( Figure 5B).

Discussion
We have created a cell-based SARS-CoV-2 RdRp activity assay system by altering the cell-based MERS-CoV RdRp activity assay system [12]. The system consists of a bicistronic p(+)FLuc-(−)UTR-NLuc reporter plasmid and the pCI-SARS2-nsp12N plasmid. The NLuc activity of the cells transfected with the plasmids represented the RdRp activity and FLuc activity was used as the internal control. We have used this system to screen and discover inhibitors of SARS-CoV-2 RdRp.
We examined the effect of the Flag tag on RdRp activity by comparing its activity with that of the C-term Flag-tagged RdRp and N-term Flag-tagged RdRp. The N-terminal region of RdRp is important for protein folding [30,31]. However, the activities of these tagged proteins were comparable. Therefore, we chose the N-term Flag-tagged RdRp for this study due to the 10% higher activity. These results were consistent with the observation that the activity of N-term Flag-tagged MERS RdRp was higher than C-term Flag-tagged RdRp in the cell-based MERS-CoV RdRp assay system [12].
In addition, we tested the p(+)FLuc-(−)UTR-NLuc reporter plasmid with polyA 33 or without polyA 33 under the 3 -UTR of SARS-CoV. Although PolyA 33 stabilizes RNA, stimulates translation and is used as a template for generating negative-sense RNA [32], we unexpectedly found that RdRp activity was decreased in the cells transfected with p(+)FLuc-(−)UTR-NLuc with polyA 33 compared to p(+)FLuc-(−)UTR-NLuc without polyA33. Therefore, we used the reporter plasmid without polyA 33 for a cell-based SARS-CoV-2 RdRp activity assay system. Two nsp8, one nsp7 and one nsp12 protein of SARS-CoV-2 form the active RdRp complex with template-primer RNA and co-factors nsp8 and nsp7 proteins were reported to confer the processivity of RdRp [9,15,17,18]. We tested the effect of nsp7 and nsp8 on RdRp activity in our cell-based system. In this system, the RdRp activity without nsp7 and nsp8 proteins was already detectable and the nsp7 and nsp8 proteins only increased RdRp activity slightly. Thus, we conducted the cell-based SARS-CoV-2 RdRp activity assay system without using the plasmids encoding nsp7 and nsp8 to screen the inhibitors by targeting the nsp12 function only. Then, we should define the effect of inhibitors on the interaction and efficacy of RdRp complex with nsp7 and nsp8 proteins in more detail.
Single-strand RNA viruses replicate their RNA in the cytoplasm of the infected cells and the corresponding dsRNA foci have been detected at the peri-nuclear region by immunofluorescence staining [33]. Here, we also visualized the peri-nuclear foci in our cell-based SARS-CoV-2 RdRp activity assay system, confirming that RdRp in the system could generate p(+)FLuc-(−)UTR-NLuc reporter plasmid-originated dsRNA replicates. Moreover, the high Z-factor of this assay system confirmed its reliability and reproducibility for the SARS-CoV-2 RdRp inhibitor screening HTS system.
We validated our cell-based SARS-CoV-2 RdRp activity assay system by testing various drugs that have been effective against coronaviruses. Remdesivir is an FDA-approved firstclass drug for COVID-19 and an adenosine analog inhibitor of SARS-CoV-2 RdRp [9,10]. In our system, remdesivir dose-dependently inhibited the RdRp activity. In addition, we tested lycorine, a natural alkaloid and non-nucleoside inhibitor of MERS-CoV RdRp and in silico inhibitor of SARS-CoV RdRp [21]. In our system, lycorine inhibited the activity of SARS-CoV RdRp dose-dependently, suggesting that lycorine was a more effective SARS-CoV RdRp inhibitor than remdesivir. These data were consistent with the findings that lycorine more effectively inhibited SARS-CoV-2 infection at the IC 50 value of 0.878 ± 0.022 µM compared with remdesivir at the IC 50 of 6.499 ± 0.256 µM. Moreover, the binding affinity of lycorine to SARS-CoV-2 RdRp at −6.2 kcal/mol is stronger than that of remdesivir at −4.7 kcal/mol [21]. Cepharanthine inhibits coronavirus infections by blocking the Ca 2+ -permeable channels during virus entry [22,23,34]. Recently, it was also suggested that cepharanthine may also bind to the interface active pockets of the SARS-CoV-2 nsp12-nsp7 and nsp12-nsp8 by virtual screening [24]. We tested cepharanthine in our assay system, but it did not display any inhibitory effect on SARS-CoV-2 RdRp activity with or without nsp7 and nsp8.
Finally, we tested the nucleos(t)ide analogs using our system to discover other inhibitors of SARS-CoV-2 RdRp. We found that the adefovir dipivoxil for treatment of HBV infection [25] effectively inhibited SARS-CoV-2 RdRp activity at a level comparable to remdesivir. The other HBV inhibitors, telbivudine [35] and entecavir hydrate [27,36], inhibited SARS-CoV-2 RdRp activity, although at a lower inhibitory ability level than adefovir dipivoxil. A nucleoside reverse transcriptase inhibitor of HIV, emtricitabine [37], also inhibited SARS-CoV-2 RdRp activity. In addition, we found that moroxydine, a broad antiviral agent against DNA and RNA viruses [28] and rifampin, a macrocyclic antibiotic for tuberculosis [38], repressed SARS-CoV-2 RdRp activity moderately. Thus, we have identified six nucleos(t)ide/ analogs as SARS-CoV-2 RdRp inhibitors using the cell-based SARS-CoV-2 RdRp activity assay system. Among them, adefovir dipivoxil was likely the strongest inhibitor comparable to the already-reported RdRp inhibitors, remdesivir and lycorine.
In summary, we have established a cell-based SARS-CoV-2 RdRp activity assay system to screen the inhibitor of SARS-CoV-2 RdRp. We confirmed the inhibitory activity of remdesivir and lycorine on SARS-CoV-2 RdRp using this system. In addition, we screened the nucleos(t)ide analogs and identified six nucleos(t)ide analogs as novel SARS-CoV-2 RdRp inhibitors and therapeutic candidates for the COVID-19. Thus, our assay system can provide an effective HTS platform for developing prophylactic and therapeutic drugs for COVID-19 and other emerging coronavirus infections.
Author Contributions: J.S.M. contributed methodology, investigation, data curation, visualization and writing-original draft; S.K. contributed formal analysis, investigation, data curation, visualization, writing-review and editing, validation, supervision, project administration, resources and funding acquisition; Y.-H.J. contributed conceptualization, methodology, formal analysis, investigation, data curation, visualization, writing-original draft, writing-review and editing, validation, supervision and project administration. All authors have read and agreed to the published version of the manuscript.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.