key: cord-0960156-uyrfqlfo
authors: Ju, Xiaohui; Zhu, Yunkai; Wang, Yuyan; Li, Jingrui; Zhang, Jiaxing; Gong, Mingli; Ren, Wenlin; Li, Sai; Zhong, Jin; Zhang, Qiangfeng Cliff; Zhang, Rong; Ding, Qiang
title: A novel cell culture system modeling the SARS-CoV-2 life cycle
date: 2020-12-13
journal: bioRxiv
DOI: 10.1101/2020.12.13.422469
sha: b426f8a31030d0248e918a5ba02e1c037710601b
doc_id: 960156
cord_uid: uyrfqlfo

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) causes the global pandemic of COVID-19, and no effective antiviral agents and vaccines are available. SARS-CoV-2 is classified as a biosafety level-3 (BLS-3) agent, impeding the basic research into its biology and the development of effective antivirals. Here, we developed a biosafety level-2 (BSL-2) cell culture system for production of transcription and replication-competent SARS-CoV-2 virus-like-particles (trVLP). This trVLP expresses a reporter gene (GFP) replacing viral nucleocapsid gene (N), which is required for viral genome packaging and virion assembly (SARS-CoV-2-GFP/ΔN trVLP). The complete viral life cycle can be achieved and exclusively confined in the cells ectopically expressing SARS-CoV or SARS-CoV-2 N proteins, but not MERS-CoV N. Genetic recombination of N supplied in trans into viral genome was not detected, as evidenced by sequence analysis after one-month serial passages in the N-expressing cells. Moreover, intein-mediated protein trans-splicing approach was utilized to split the viral N gene into two independent vectors, and the ligated viral N protein could function in trans to recapitulate entire viral life cycle, further securing the biosafety of this cell culture model. Based on this BSL-2 SARS-CoV-2 cell culture model, we developed a 96-well format high throughput screening for antivirals discovery. We identified salinomycin, tubeimoside I, monensin sodium, lycorine chloride and nigericin sodium as potent antivirals against SARS-CoV-2 infection. Collectively, we developed a convenient and efficient SARS-CoV-2 reverse genetics tool to dissect the virus life cycle under a BSL-2 condition. This powerful tool should accelerate our understanding of SARS-CoV-2 biology and its antiviral development.

length cDNA clone has been established using the in vitro ligation of cDNA 77 fragments 11,12 . This system has been shown to be efficient for the recovery of infectious 78 virus, and a reporter gene can be inserted into the viral genome to monitor virus 79 replication, providing a good tool for high-throughput antiviral screening. However, 80 experimentations involving live virus are restricted to BSL-3 laboratories, which 81 hinders the study of SARS-CoV-2 and development of countermeasures. Therefore, it 82 is urgent to develop an efficient non-BSL-3 experimental system for SARS-CoV-2. 83 Herein, we developed an N-based genetic complementation system to produce (Fig. 1) . 102 To assemble the molecular clone of SARS-CoV-2-GFP/N genome, we utilized 103 an in vitro ligation approach, which has been used for constructing the infectious clone the Caco-2-N cells but not in Caco-2 cells ( Fig. 2C and D) . RT-PCR analysis using a 139 primer set outside the N-encoding region confirmed that the N gene was indeed 140 replaced by GFP in the recombinant trVLP viral genome (Fig. 2E) . 141 To characterize the SARS-CoV-2-GFP/N trVLP infection, two spike-specific 142 mAbs (1F11 and 2F6) 13 were tested for their ability to neutralize infection of Caco-2-143 N cells. A neutralizing mAb specific for HIV gp120 (VRC-01) was also included as the showed that 1F11 and 2F6 inhibited trVLP infection in a dose-dependent manner; in 148 contrast, VRC01 had no effect on the trVLP infection (Fig. 2F) . 149 Soluble recombinant forms of the human ACE2 are able to bind SARS-CoV-2 150 spike protein and inhibit its interaction with cellular ACE2 15,16 . We therefore tested the 3'UTR that covers the region of the inserted GFP reporter gene. (Fig.3A) . RT-PCR 168 products of 1.5-Kbp and 1-Kbp were expected for WT genome and SARS-CoV-2-169 GFP/N genome, respectively. SARS-CoV-2-GFP/N trVLP was considerably stable 170 for at least 3 serial passages since the 1-Kb RT-PCR products were detected at P3 trVLP 171 (Fig. 3B) . The loss of GFP reporter gene was detected in the P4 trVLP as indicated by 172 amplicon of < 1 Kb size (Fig. 3B, Fig S3A) . No PCR product of greater than 1 Kb was 173 detected in the all samples, suggesting that no heterologous RNA inserted into the 174 SARS-CoV-2-GFP/N genome, at least in the GFP report region.

To characterize the trVLP sequence variations in an unbiased manner, we 176 performed deep sequencing analysis on the P1 and P10 trVLP genome. The deep 177 sequencing analysis provides deep coverage, on the order of 30 million reads per 178 sequencing sample (Fig. S3B) . Sequences of P1 or P10 were mapped to the SARS-8 CoV-2 and SARS-CoV-2-GFP/N trVLP genomes, respectively ( Fig. 3C and D) and 180 relative abundances of these sequences between P1 and P10, were also compared ( Fig   181   3E, Fig S3C) . The deep sequencing analysis could not detect N sequences in the both 182 P0 and P10 genome (Fig. 3C) , and GFP sequences were readily detected in the P1 183 genome with high abundance, however, it was rarely detected in the P10 genome ( Fig.   184 3D), due to GFP sequences deletion (Fig. 3B, Fig S3A) . Additionally, we found that 185 the subgenomic RNAs of ORF6, ORF7 and ORF8 were dramatically decreased in the 186 P10-trVLP infected cells compared with that of P1 VLP (Fig. 3E) , indicating that ORF6, To locate the splice sites according to these requirements we chose three splice sites in 205 the N protein, to have the N-intein A152C, S176C and G212C (Fig. 4A) . As for each 206 of N-intein above, we constructed two lentivirus vectors encoding either the N-or the 9 having N N -Int N and Int C -N C , respectively (Fig. 4A) . Each lentivirus vector included 209 appropriate regulatory elements (promoter and a polyadenylation signal) and a Flag tag 210 to allow detection of the full-length reconstituted N protein (Fig. 4A) . We then 211 transduced N N -Int N and Int C -N C either individually or together in Caco-2 cells, and the 212 full-length N protein reconstitution was assessed by Western blotting assay. We could occur; interestingly, N-intein (S176C) could ligate a full-length N (S176C), but fails to 223 support virus infection, suggesting that the S176C mutation probably impairs N protein 224 function ( Fig. 4C and D) . Consistent with the GFP expression, the subgenomic RNA 

Coronavirus N protein is an extensively phosphorylated, highly basic, vital 234 structural protein the primary function of which is to form a helical ribonucleoprotein 235 complex with viral RNA (RNP) as core structure of the virion. A variety of other 236 functions have been ascribed, such as viral genome transcription and replication, or 237 evasion of antiviral immunity. SARS-CoV-2 N protein is highly homologous to the N 238 protein of SARS-CoV, with 91% identity, while exhibited 48% identity with that of 239 MERS-CoV (Fig. 5A) . Several proteomics profiling analyses have been performed and 240 reveals that N protein of SARS-CoV-2 is extensively phosphorylated at multiple sites 241 ( Fig. 5A and Fig. S4) . However, the roles of N protein phosphorylation remain unclear. As SARS-CoV-2 N is heavily phosphorylated at multiple sites especially within 259 the central Ser-Arg (SR)-rich motif, we are interested in the roles of phosphorylation in 260 N function. For this purpose, we mutated S176, S413, S176/413, S105, S183, S188, 261 S206, S188/206 as the conservation of these residues with that of SARS-CoV into 262 alanine to specifically dissect their function. Notably, GSK-3 is the kinase responsible 263 for the phosphorylation of this SR-rich motif in SARS-CoV N protein, which are 264 primed by the phosphorylation of Ser-189 and Ser-207 (Ser-188 and Ser-206 in SARS-transduced with the N variants as indicated. As shown in Western blotting assay, the 267 mutations did not alter the protein expression and stability in the Caco-2 cell (Fig. 5F) , 268 We noted that N with the S188A/S206A double mutations migrated slightly faster than 269 WT and other mutants, probably because blockade of the initial priming Interestingly, most of the phosphorylation null mutants were able to assemble virus-like 276 particles with comparable or slightly reduced efficiencies than WT. However, 277 S188A/S206A double mutations completely abolish N function ( Fig. 5G-H) , 278 highlighting the critical role of S188 and S206 for N function. observed 60% reduction of the GFP fluorescence (Fig. 6A) . This is consistent with 299 recent reports that SARS-CoV-2 is sensitive to type I interferon treatment 11,25-27 . 300 Remdesivir and GC376, which targets virus RNA dependent RNA polymerase (RdRp) 301 and 3CLpro respectively, have been reported to be potent antivirals against SARS-CoV- To provide proof-of-concept that our system could be utilized in high-throughput 322 screening, we performed HTS of Topscience natural product library containing 377 323 drugs (Fig. 7A ) and the potential hit compounds were further assessed using authentic 324 SARS-CoV-2 to confirm the antiviral activities in vitro. DMSO or remdesivir were 325 included as the negative or positive control.

Among the 377 compounds of the compound library, 10 hit molecules showed 327 equal or higher inhibition with an inhibitory efficiency ≥ 60% (Fig. 7A) . In addition, 328 we excluded five hits due to the visible cytotoxicity. This criterion allowed the selection 329 of five hits as the highest confident hits: salinomycin, tubeimoside I, monensin sodium, 330 lycorine chloride and nigericin sodium (Fig. 7A) . Among these five compounds, detected and immunoprecipitated by an anti-FLAG antibody (Fig. S4A) . Multiple 377 amino acids in N protein can be phosphorylated, but our data demonstrated that most 378 of these phosphorylation may not be required for N function at least in vitro. Meanwhile, 379 we also identified numerous host factors associated with N protein (Fig. S4A ; Table   380 S2), notably, we also found that N protein could interact with G3BP1 and G3BP2, the 381 stress granule assembly proteins, which was in line with previous studies 39, 40 . Recent 382 studies found that N protein could impair the stress granule assembly to escape the 383 antiviral effect 40,41 . Thus, the trVLP system provides a new tool to study host factors 384 and viral proteins that may interact with N during SARS-CoV-2 infection. Table 1 . gene were calculated. Heatmaps were drawn by using R package "pheatmap" 499 (https://www.r-project.org). To quantify the junction-reads from subgenomic RNAs, 500 the STAR2.7.5c was used for reads mapping. The junction-reads was defined and 501 collected as described in Kim, Cell, 2020. A Sankey diagram was drawn by using R 502 packages named "networkD3" and "dplyr". 

to infect the Caco-2-N cells, and GFP expression analyzed by flow 661 cytometry/microscopy or viral subgenomic RNA abundance were determined by RT-662 qPCR. (C) Western blotting assay was performed to detect the N proteins expression in 663

Caco-2 cells transduced with distinct N genes from SARS-CoV-2, SARS-CoV or 664

D-E) The cell culture medium was collected from SARS-CoV-2 GFP/N 665 trVLP infected Caco-2 cells expressing N from SARS-CoV-2

CoV to infect the naïve Caco-2-N cells. GFP were observed using microscopy and 667 cellular RNA was extracted for RT-qPCR analysis to determine viral subgenomic RNA 668 levels. (F) Western blotting assay detected the expression of SARS-CoV-2 N WT or 669 mutants in Caco-2 cells

CoV-2 GFP/N trVLP infected Caco-2 cells expressing SARS-CoV-2 N mutants to 671 infect the naïve Caco-2-N cells. GFP were observed using microscopy and cellular 672 RNA was extracted for RT-qPCR analysis to determine viral subgenomic RNA levels

I) GSK-3 inhibitors LiCl or SB216763 treated Caco-2-N cells inoculated with SARS-674

CoV-2 GFP/N trVLP, the cell culture medium was then inoculated with Caco-2-N 675 cells. RNA was extracted for RT-qPCR analysis to determine viral subgenomic RNA 676 levels. Cell viability was evaluated by CellTiter-Glo assay. Error bars (E, H and I) 677 represent the standard deviations from one of three independent experiments performed 678 in triplicate

Significance 679 assessed by one-way ANOVA

Dose response 711 curves were generated using GraphPad Prism software version 7.0. IC50 values were 712 calculated using Prism software and is representative of one of three independent 713 experiments. Error bars represent the standard deviations from one of three independent 714 experiments performed in triplicate. (G) Comparison of antiviral activity and 715 cytotoxicity of each compound

Supplemental Figure 1. Generation of Caco-2 cell expressing SARS-CoV-2 N by 723 lentiviral transduction. (A) Scheme depicting the bicistronic lentiviral constructs for 724 expressing SARS-CoV-2 N protein with C-terminal Flag tag. (B) Representative flow 725 cytometry plots demonstrating efficient lentivirus transduction. Caco-2 cells were 726 transduced with pLVX-N-Flag-IRES-mCherry or not transduced

Supplemental Figure 2. GFP expression in Caco-2-N cells electroporated with 734

A) GFP expression in Caco-2-N cells electroporated 735 with SARS-CoV-2 GFP/N RNA. Caco-2-N cells were electroporated with 20 μg of 736 SARS-CoV-2 GFP/N RNA. From 21h-96h p.t., GFP expression in the cells was 737 observed with microscopy. (B) GFP expression was quantified by flowcytometry at 96h 738 post transfection of the RNA

Supplemental Figure 3. Characterization of the genetic stability of SARS-CoV-2

A) RT-PCR products from P10 virus infected cell passage were cloned 745 into pEASY-Blunt vector, and 12 colonies were randomly chosen for DNA sequences 746 analysis. Multiple deletions were detected in the amplicon. (B) Categories of mapped 747 reads from P1 and P10 virus infected Caco-2-N cells. (C) Canonical discontinuous 748 transcription (top) that is mediated by TRS-L (TRS in the leader) and TRS-B (TRS in 749 the body)

Identification of host factors associated with N protein and 755 phosphorylation on N protein by mass spectrometry. (A) Flag tagged N protein was 756 immunoprecipitated from Caco-2-N cells infected with recombinant SARS

GFP/N trVLP using Flag antibody, and the proteins were analyzed on SDS-PAGE gel

The proteins were visualized by Coomassie blue staining. N, G3BP1 and G3BP2 were 759 labelled. (B) Phosphorylated peptides of N protein derived from Caco-2-N cells

-N cells in which N was C-terminal Flag-tagged were collected and cell lysates were 761 immunoprecipitated with anti-Flag coupled beads

A novel coronavirus outbreak of global 768 health concern

Drug Evaluation during the Covid-19 Pandemic

Human Coronavirus: Host-Pathogen Interaction

Biochemical aspects of coronavirus 774 replication and virus-host interaction

Coronavirus: organization, replication and expression of genome

Structure, Function, and Evolution of Coronavirus Spike Proteins

Coronavirus nucleocapsid protein facilitates template switching and is 781 required for efficient transcription

Coronavirus gene 7 counteracts host defenses and modulates virus 783 virulence

Accessory proteins of SARS-785 CoV and other coronaviruses

Coronavirus reverse genetic systems: infectious clones and replicons

An Infectious cDNA Clone of SARS-CoV-2

SARS-CoV-2 Reverse Genetics Reveals a Variable Infection Gradient in the 792

Human neutralizing antibodies elicited by SARS-CoV-2 infection

Structural basis for broad and potent neutralization of HIV-1 by antibody 796

VRC01

Engineering human ACE2 to optimize binding to the spike protein of 798 SARS coronavirus 2

Potential host range of multiple SARS-like coronaviruses and an improved 800

ACE2-Fc variant that is potent against both SARS-CoV-2 and SARS-CoV-1

Characterization of SARS-CoV-2 ORF6 deletion variants detected in 803 a nosocomial cluster during routine genomic surveillance

Identification of multiple large deletions in ORF7a resulting in in-frame 805 gene fusions in clinical SARS-CoV-2 isolates

Discovery and Genomic Characterization of a 382-Nucleotide Deletion in 811

ORF7b and ORF8 during the Early Evolution of SARS-CoV-2

Semisynthesis of proteins by expressed protein ligation

The naturally split Npu DnaE intein exhibits an 816 extraordinarily high rate in the protein trans-splicing reaction

Glycogen synthase kinase-3 regulates the phosphorylation of severe acute 819 respiratory syndrome coronavirus nucleocapsid protein and viral replication

Nucleocapsid phosphorylation and RNA helicase DDX1 822 recruitment enables coronavirus transition from discontinuous to continuous transcription

Type I interferon susceptibility distinguishes SARS-CoV-2 from 825 SARS-CoV

Type I and Type III Interferons Restrict SARS-CoV-2 Infection of 827

Human Airway Epithelial Cultures

Type I and Type III Interferons -Induction, Signaling, Evasion, and 829 Application to Combat COVID-19

Remdesivir and chloroquine effectively inhibit the recently emerged novel 832

2019-nCoV) in vitro

Remdesivir Inhibits SARS-CoV-2 in Human Lung Cells and Chimeric 835

Expressing the SARS-CoV-2 RNA Polymerase in Mice

Discovery and Synthesis of a Phosphoramidate Prodrug of a Pyrrolo

Adenine C-Nucleoside (GS-5734) for the Treatment of Ebola

Emerging Viruses

Broad-spectrum antiviral GS-5734 inhibits both epidemic and 842 zoonotic coronaviruses

Therapeutic efficacy of the small molecule GS-5734 against Ebola virus 844 in rhesus monkeys

A Trial of Lopinavir-Ritonavir in Adults Hospitalized with Severe Covid-19

morphogenesis of avian coronavirus in Vero cells and their inhibition by monensin

Potential Antiviral Options against SARS-CoV-2 Infection

Emerging Pandemic Diseases: How We Got to COVID-19

A SARS-CoV-2 protein interaction map reveals targets for drug 859 repurposing

Nucleocapsid 861 protein of SARS-CoV-2 phase separates into RNA-rich polymerase-containing 862 condensates

A proposed role for the SARS-CoV-2 nucleocapsid protein 865 in the formation and regulation of biomolecular condensates

Identification of functional cis-acting RNA elements in the hepatitis E virus 868 genome required for viral replication