Poliovirus-nonsusceptible Vero cell line for the World Health Organization global action plan

Polio or poliomyelitis is a disabling and life-threatening disease caused by poliovirus (PV). As a consequence of global polio vaccination efforts, wild PV serotypes 2 and 3 have been eradicated around the world, and wild PV serotype 1-transmitted cases have been largely eliminated except for limited regions. However, vaccine-derived PV, pathogenically reverted live PV vaccine strains, has become a serious issue. For the global eradication of polio, the World Health Organization is conducting the third edition of the Global Action Plan, which is requesting stringent control of potentially PV-infected materials. To facilitate the mission, we generated a PV-nonsusceptible Vero cell subline, which may serve as an ideal replacement of standard Vero cells to isolate emerging/re-emerging viruses without the risk of generating PV-infected materials.


Results and discussion
Genomic genes and cDNAs of the human PV receptor homologs in Vero cells. The cell receptor of PV (PVR; also known as CD155) is a glycosylated membrane protein with an N-terminal extracellular domain, single transmembrane domain, and C-terminal cytoplasmic domain (Fig. 1a) 1 . The human genome has only one PVR-encoding gene, which expresses four alternative splicing transcripts. Among the transcript variants, only two variants have the transmembrane domain and function as the cell receptor of PV 1 . The AGM genome contains two paralogs of the human PVR 20,21 , hereafter referred to as PVR1 and PVR2, which probably arise because of gene duplication. A previous study showed that both AGM PVRs have the potential to form functional PVRs 21 .
In both genes, the first exon encodes the translational initiation codon, the signal peptide domain, and a part of the extracellular immunoglobulin-like V-type domain, which are essential for functional PVR 1,21 . We amplified the first exon and flanking introns of the PVRs in the Vero JCRB9013 cell line (which is equivalent to the Vero ATCC CCL-81 cell line) by genomic polymerase chain reaction (PCR), and sequenced them to design guide RNA (gRNA) targeting this exon. The amplified region of the PVR2 gene of Vero cells was found to contain synonymous single nucleotide variations (SNVs) when compared with counterparts in AGM reference sequences (Fig. 1b,c). Additionally, we obtained several full-length cDNA clones of Vero cell PVR1 and PVR2  Table 1). (b) Nucleotide sequences of the initial ATG codon and downstream region in Vero-PVR1 and Vero-PVR2. Genomic DNA of Vero cells were extracted and the coding region located on exon 1 of PVR1 and PVR2 were sequenced. DNA sequences of Vero-PVR1 and Vero-PVR2 are shown in lowercase letters; the predicted amino acid sequences are shown in capital letters. Vero-PVR2 was found to have a single-nucleotide variation (SNV) ( †) in the open reading frame (ORF), probably due to heterozygosity of the gene. (c) DNA sequencing chromatogram of a Vero-PVR2 region. Sanger sequencing of Vero-PVR2 showed an SNV in the ORF of exon 1 (see also panel b). and sequenced them individually. Whereas only one type of sequence was obtained from the PVR1 cDNA set ( Supplementary Fig. 1a), two types were obtained from the PVR2 cDNA set ( Supplementary Fig. 1b), which indicates that different patterns of SNVs, including synonymous and nonsynonymous mutations, exist between the two alleles of PVR2 gene. Of note, Vero cells are pseudodiploid 7,8 . Disruption of PVR1, PVR2, or both in Vero cells. Following the scheme shown in Fig. 2a, we constructed PVR-disrupted Vero cell lines, obtaining two PVR1 single-knockouts (SKOs), two PVR2 SKOs, and six PVR1/PVR2 double-knockouts (DKOs) cell clones ( Supplementary Fig. 2). Among the PVR1/PVR2 DKO cell clones, two clones (ΔPVR1/2-1 and ΔPVR1/2-2) were derived from a PVR1 SKO clone (ΔPVR1-1) while four clones (ΔPVR1/2-3, ΔPVR1/2-4, ΔPVR1/2-5, and ΔPVR1/2-6) were derived from another PVR1 SKO clone (ΔPVR1-2; Supplementary Fig. 2). No obvious morphological changes were observed among these clones (Fig. 3a). Nevertheless, the ΔPVR1-2 cell clone had retarded entry into the logarithmic phase of cell growth after plating, when compared with the parental, ΔPVR1-1, ΔPVR2-1, and ΔPVR2-2 cell clones (Fig. 3b). Thus, we chose ΔPVR1-1, not ΔPVR1-2, as the Vero ΔPVR1 representative for depositing to a cell bank (see also below).
In western blot analysis with an anti-human PVR monoclonal antibody, which was expected to cross-react with AGM PVRs, parental Vero cells exhibited broad bands, with band mobility corresponding to ~ 60-40 kDa (Fig. 2b). As peptide-N-glycosidase F (PNGase F)-treatment sharpened the bands (Fig. 2c), the observed band broadness was largely due to the heterogeneity of protein glycosylation. Compared with the parental control, the band signal was greatly reduced with Vero PVR1 SKO cells, but not with Vero PVR2 SKO cells. The signal retained with Vero PVR1 SKO cells disappeared with Vero PVR1/2 DKO cells, which indicates that the band represented the PVR2 protein (Fig. 2b,c). Compared with parental Vero cells, all PVR KO clones retained high susceptibility to MV and RV, with similar time courses in virus production ( Supplementary Fig. 4, 5). All PVR KO clones also retained high susceptibility to JEV. However, the JEV titer levels of Vero, ΔPVR1-1, and ΔPVR1-1-derived PVR1/2 DKO clones, and PVR2 SKO clones peaked at three days post-infection ( Supplementary Fig. 6b,c) while those of ΔPVR1-2 and ΔPVR1-2-derived PVR1/2 DKO cell clones (except for ΔPVR1/2-6) peaked two days post-infection ( Supplementary  Fig. 6d). The cause of the different characteristics of Vero ΔPVR1-1 and ΔPVR1-2 clones is unknown.

PV-nonsusceptible
Among the Vero cell mutant clones established in this study, we chose Vero ΔPVR1-1, Vero ΔPVR2-1, and Vero ΔPVR1/2-1 clones as representative cell lines of Vero PVR1 SKO, Vero PVR2 SKO, and Vero PVR1/2 DKO cell clones, respectively (Supplementary Fig. 2). It should be noted that the phenotypes of the representative cell lines were not due to any clonal effects as other mutant cell lines established in this study broadly exhibited the same phenotypes (Figs. 2, 3, 4 and Supplementary Figs. [3][4][5][6]. For the global distribution, these representative cell lines have been deposited in the Japanese Collection of Research Bioresources (JCRB) Cell Bank, a non-profit public cell bank in Japan. Of note, the representative cell lines deposited to the cell bank were renamed Vero ΔPVR1, Vero ΔPVR2, and Vero ΔPVR1/2, respectively, for simplicity ( Supplementary Fig. 2).
Since GAPIII requires the stricter control of PV, we conducted infection experiments using vaccine PV strains with a lower biorisk than wild-type PV. We verified that these Vero cells are not susceptible to PV using attenuated strains (Fig. 4). Since all PV strains use the same PVR 22 , the loss of PVR may prevent viral receptor binding and the subsequent entry of virulent wild-type PV strains. When performing transfection experiments using RNA extracted from potentially PV-infectious materials, PVR-deficient Vero cells are still susceptible to the first replication cycle occurring in the cell immediately after viral RNA transfection. The duration for which a genomic deletion or insertion by genome editing technologies is stable or how many repeating cell passages are permissive for safety currently remain unknown. To ensure quality control, PVR deficiency may be confirmed in these cell lines by the DNA sequencing of a genomic PCR product (see Methods and Supplementary Fig. 2). Information on cell lines will be shared with the WHO Containment Advisory Group via Polio Global Specialized Laboratory in the National Institute of Infectious Diseases (NIID). The quality control of the cell line will also be conducted under the supervision of NIID on request.
In the present study, we focused on Vero JCRB9013 as a model of Vero cells. Although Vero JCRB9013 is the most widely used Vero cell line, it is not optimal for the propagation of some viruses. The establishment of PVR-deficient cells using other Vero cell sublines and/or the expression of specific virus receptors in PVRdeficient Vero cells may strengthen the GAPIII mission. Therefore, we are now attempting to establish stable transformants of the Vero ΔPVR1/2-1 and ΔPVR1/2-2 cell lines expressing the human signaling lymphocytic www.nature.com/scientificreports/ activation molecule in order to isolate the measles virus. When an outbreak of an emerging and re-emerging infectious disease occurs, the identification of the causative pathogen is crucial. Additionally, in vitro pathogen culture systems are invaluable for fully characterizing the pathogen and the development of anti-pathogenic drugs. The present study provides evidence that when Vero ΔPVR1/2 cells are employed as host cells for non-PV viruses, the resultant cell culture can be excluded from classification as "potentially PV-infected materials". Thus, Vero ΔPVR1/2 cells may serve as an ideal replacement of standard Vero cells to isolate emerging/reemerging viruses without the risk of generating potentially PV-infected materials, which is in alignment with the WHO GAPIII mission. A study on PVR-disrupted mutant cell lines of human cervical cancer-derived HeLa and rhabdomyosarcoma-derived RD-A cells with a similar aim, to be used mainly for non-polio enterovirus isolation, was recently published after the submission of this manuscript 23 .   www.nature.com/scientificreports/ Rv; 5′-CCT TGT GCC CTC TGT CTG TGG ATC CTG-3′ were used. For Vero-PVR2, two primer sets were designed because our genomic DNA sequence analysis showed that PVR2 has a nucleotide polymorphism at its 5′ terminus (Supplementary Fig. 1b). One of the primer sets for Vero-PVR2 was cloning primer 2G-Fw (5′-ATG GCC GCC GCA TGG CCT CCG CTG CTG-3′) and 1/2-Rv (5′-CCT TGT GCC CTC TGT CTG TGG ATC  CTG- Construction of a vector for the CRISPR/Cas9 system. Guide sequences containing the target sequences were cloned into the BsmBI site of the pSELECT-CRISPR cas9 plasmid as described previously 24,25 .

Methods
The target sequences specific for Vero-PVR1 and -PVR2 are shown in Fig. 1b. Vero cells were transfected with the constructed plasmids using the Lipofectamine 2000 reagent (Thermo Fisher Scientific, MA, USA) according to the manufacturer's protocol. Three days after transfection, cells were cultivated for three days in growth medium supplemented with 20 µg/ml of puromycin (Takara Bio Inc.) for PVR1 KO or 40 µg/ml for PVR2 KO. Puromycin-resistant colonies were further grown without puromycin, isolated using cloning cups, and subcultured.
Western blotting. All procedures were performed at 4 °C or on ice, unless otherwise noted. Cells were washed with phosphate buffered saline (PBS) three times, removed from the plates using cell scrapers (IWAKI, Tokyo, Japan) in 1 ml of PBS, and centrifuged at 10,000 × g for 3 min. The cell pellets were stored at -30 °C until analysis. Each frozen pellet was thawed on ice and lysed for 10 min with 200 µl of 100 mM Tris-HCl (pH 7.0) buffer containing 0.5% NP-40, 0.5% sodium deoxycholate, 10 mM ethylenediaminetetraacetic acid (EDTA), and 100 mM NaCl. After centrifugation at 1,000 × g for 3 min to remove insoluble debris, the supernatant fraction was collected as cell lysate. Protein concentrations of cell lysates were measured using a bicinchoninic acid protein assay kit (Takara Bio Inc.). Cell lysate containing the same amount of protein was mixed with 10 volumes of methanol and then centrifuged at 20,000 × g for 30 min to precipitate proteins. After drying by air, the precipitated proteins were dissolved and denatured in modified Laemmli sample buffer (Bio-Rad, CA, USA), containing100 mM dithiothreitol rather than 5% 2-mercaptoethanol, by boiling for 10 min. Fifteen micrograms of protein was electrophoresed on 15% e-PAGEL sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) gels (ATTO, Tokyo, Japan) and transferred to polyvinylidene difluoride membranes (ATTO). The anti-PVR/CD155 rabbit monoclonal antibody D3G7H (Cell Signaling Technology, MA, USA; 0.1 µg/ml) and anti-GAPDH rabbit monoclonal antibody 14C10 (Cell Signaling Technology; 0.1 µg/ml) were diluted with Can Get Signal Immunoreaction Enhancer Solution (TOYOBO Life Science, Osaka, Japan) and used as primary antibodies. A horseradish peroxidase (HRP)-labeled anti-rabbit IgG antibody (Jackson ImmunoResearch, PA, USA; 0.7 µg/ml) in Tris-buffered saline with 0.05% Tween 20 was used as a secondary antibody. Proteins were detected by Immobilon Western Chemiluminescent HRP Substrate (Merck Millipore, MA, USA) and a LAS-3000 mini chemiluminescence imaging system (Fujifilm, Tokyo, Japan) with ImageGauge software (Fujifilm).

Deglycosylation of proteins.
To remove N-linked glycans, 25 µg of protein was precipitated from cell lysate as described in Methods and treated with peptide-N-glycosidase F (PNGase F; New England Biolabs, MT, USA) according to the manufacturer's instructions.
PV infection and titration. Vero and PVR KO cells seeded in a 12-well plate were infected with PV Sabin 1 and Sabin 3 strains (National Institute for Biological Standards and Control, Hertfordshire, UK) at MOI of 10 or 0.01. After 1 h incubation at 35 °C, the cells were washed three times with PBS (−) and cultured at 35 °C in a 5% CO 2 humid atmosphere with MEM (Thermo Fisher Scientific) containing 2% FBS (Biowest, Nuaillé, France) and penicillin-streptomycin (Cosmo Bio Co., Tokyo, Japan). At appropriate intervals, cells were dissociated by pipetting and collected together with culture supernatants as the culture fluid. The culture fluid was subjected to three freeze-thaw cycles and then centrifuged at 1,000 × g for 5 min to remove debris. Virus titer was determined by 50% end-point dilution assay (50% tissue culture infectious dose, TCID 50 ) using HEp-2 cells (ATCC CCL-23), a HeLa sub-line (ATCC, VA, USA). Viral RNA copy numbers in the culture fluid was also measured by real-time reverse transcription-PCR 26,27 .