Alberto A Amarilla1, Julian D J Sng1, Rhys Parry1, Joshua M Deerain2, James R Potter1, Yin Xiang Setoh1,3, Daniel J Rawle4, Thuy T Le4, Naphak Modhiran1, Xiaohui Wang1, Nias Y G Peng1, Francisco J Torres1, Alyssa Pyke5, Jessica J Harrison1, Morgan E Freney1, Benjamin Liang1, Christopher L D McMillan1, Stacey T M Cheung1, Darwin J Da Costa Guevara1, Joshua M Hardy6, Mark Bettington7, David A Muller1, Fasséli Coulibaly6, Frederick Moore5, Roy A Hall1,8, Paul R Young1,8, Jason M Mackenzie9, Jody Hobson-Peters10,11, Andreas Suhrbier12,13, Daniel Watterson14,15, Alexander A Khromykh16,17. 1. School of Chemistry and Molecular Biosciences, University of Queensland, St Lucia, QLD, Australia. 2. Department of Microbiology and Immunology, Peter Doherty Institute for Infection and Immunity, University of Melbourne, Melbourne, VIC, Australia. 3. Microbiology and Molecular Epidemiology Division, Environmental Health Institute, National Environmental Agency, Singapore, Singapore. 4. QIMR Berghofer Medical Research Institute, Herston, QLD, Australia. 5. Queensland Health Forensic & Scientific Services, Queensland Department of Health, Coopers Plains, QLD, Australia. 6. Infection & Immunity Program, Biomedicine Discovery Institute and Department of Biochemistry and Molecular Biology, Monash University, Clayton, VIC, Australia. 7. School of Medicine, University of Queensland, Kelvin Grove, QLD, Australia. 8. Australian Infectious Diseases Research Centre, Global Virus Network Centre of Excellence, Brisbane, QLD, Australia. 9. Department of Microbiology and Immunology, Peter Doherty Institute for Infection and Immunity, University of Melbourne, Melbourne, VIC, Australia. jason.mackenzie@unimelb.edu.au. 10. School of Chemistry and Molecular Biosciences, University of Queensland, St Lucia, QLD, Australia. j.peters2@uq.edu.au. 11. Australian Infectious Diseases Research Centre, Global Virus Network Centre of Excellence, Brisbane, QLD, Australia. j.peters2@uq.edu.au. 12. QIMR Berghofer Medical Research Institute, Herston, QLD, Australia. andreas.suhrbier@qimrberghofer.edu.au. 13. Australian Infectious Diseases Research Centre, Global Virus Network Centre of Excellence, Brisbane, QLD, Australia. andreas.suhrbier@qimrberghofer.edu.au. 14. School of Chemistry and Molecular Biosciences, University of Queensland, St Lucia, QLD, Australia. d.watterson@uq.edu.au. 15. Australian Infectious Diseases Research Centre, Global Virus Network Centre of Excellence, Brisbane, QLD, Australia. d.watterson@uq.edu.au. 16. School of Chemistry and Molecular Biosciences, University of Queensland, St Lucia, QLD, Australia. a.khromykh@uq.edu.au. 17. Australian Infectious Diseases Research Centre, Global Virus Network Centre of Excellence, Brisbane, QLD, Australia. a.khromykh@uq.edu.au.
Abstract
The current COVID-19 pandemic is caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). We demonstrate that despite the large size of the viral RNA genome (~30 kb), infectious full-length cDNA is readily assembled in vitro by a circular polymerase extension reaction (CPER) methodology without the need for technically demanding intermediate steps. Overlapping cDNA fragments are generated from viral RNA and assembled together with a linker fragment containing CMV promoter into a circular full-length viral cDNA in a single reaction. Transfection of the circular cDNA into mammalian cells results in the recovery of infectious SARS-CoV-2 virus that exhibits properties comparable to the parental virus in vitro and in vivo. CPER is also used to generate insect-specific Casuarina virus with ~20 kb genome and the human pathogens Ross River virus (Alphavirus) and Norovirus (Calicivirus), with the latter from a clinical sample. Additionally, reporter and mutant viruses are generated and employed to study virus replication and virus-receptor interactions.
Thpan class="Gene">e currn>an class="Gene">ent COVID-19 pandemic is caused by thesevere acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Wedemonstrate that despite the large size of the viral RNA genome (~30 kb), infectious full-length cDNA is readily assembled in vitro by a circular polymeraseextension reaction (CPER) methodology without the need for technically demanding intermediate steps. Overlapping cDNA fragments are generated from viral RNA and assembled together with a linker fragment containing CMV promoter into a circular full-length viral cDNA in a single reaction. Transfection of the circular cDNA into mammalian cells results in the recovery of infectious SARS-CoV-2 virus that exhibits properties comparable to the parental virus in vitro and in vivo. CPER is also used to generate insect-specific Casuarina viruswith ~20 kb genome and thehuman pathogensRoss River virus (Alphavirus) and Norovirus (Calicivirus), with the latter from a clinical sample. Additionally, reporter and mutant viruses are generated andemployed to study virus replication and virus-receptor interactions.
pan class="Chemical">Positivn>an class="Gene">e-strand RNA viruses encompass a large number of viruses from 41 virus families that are assigned to four large orders[1] and include major human pathogens such as the flaviviruses, e.g., West Nile virus andZika virus (ZIKV); alphaviruses, e.g., chikungunya virus (CHIKV) andRoss River virus (RRV); picornaviruses, e.g., poliovirus and the caliciviruses, e.g., human norovirus (HuNoV); and the pathogenic coronaviruses, e.g., severe acute respiratory syndrome coronavirus (SARS-CoV), Middle East respiratory syndrome coronavirus (MERS-CoV), and the most recent SARS coronavirus 2 (SARS-CoV-2). Traditionally, reverse genetic systems for positive-strand RNA viruses are based on assembling full-length cDNA clones in a plasmid vector or, for larger viruses likecoronaviruses, in bacterial or yeast artificial chromosome vectors (BAC and YAC, respectively) and propagating them in bacteria or yeast[2,3]. Incorporating bacteriophage T7 or SP6 promoters upstream of the viral 5′UTR sequence allows cell-free generation of viral RNA with T7 or SP6 RNA polymerases. The incorporation of eukaryotic expression promoters instead of T7 or SP6 promoters enables the generation of viral RNA from transfectedDNA by host cell RNA polymerase II[2,4]. While these reverse genetics approaches have been extensively used in the RNA virology field, propagation of full-length cDNA clones, particularly for larger viruses, is problematic due to toxicity of some viral sequences for bacteria and/or yeast and the presence of cryptic transcription, splicing, and termination signals, which often leads to deleterious mutations and/or deletions[3]. Several approaches have been developed to overcome these issues, including (i) the use of very low copy number plasmids, (ii) mutating cryptic sites, (iii) generating full-length DNA templates for in vitro RNA transcription by in vitro ligation of DNA fragments, and (iv) cotransfecting a mixture of overlapping DNA fragments with the first fragment containing eukaryotic expression promoter upstream of the viral 5′UTR sequence[3,5]. While proven to be useful for some positive-strand RNA viruses, these approaches haveeither not been successful or have not been attempted for most RNA viruses. These approaches also often require bespoke optimization of conditions, e.g., use of specialized plasmid vectors and bacterial strains, limited choice of fragments due to specific locations of restriction sites, and the requirement for large sequence overlaps.
Hpan class="Gene">ern>an class="Gene">ein, we advanced the circular polymeraseextension reaction (CPER) methodology that we previously developed for flaviviruses[6-8] to allow the generation of RNA viruses that have large genomes and that contain polyA tails such as SARS-CoV-2 (~30 kb) and the insect-specific member of the same viral order, Casuarina virus (~20 kb). We also successfully applied this CPER method to polyA tail-containing representatives of other positive-strand RNA virus families, such as the arthritogenic alphavirus RRV and two caliciviruses, murine norovirus (MNV) and HuNoV, with the latter generated from a human fecal sample. We further demonstrate the utility of this CPER methodology by generating reporter and mutant viruses and illustrate their application in the studies of virus replication and virus–receptor interactions. The ability to rapidly generate and manipulate RNA viruses using this CPER approach provides a robust avenue to facilitate fundamental discoveries[9,10] and ultimately help develop new interventions[11].
Results
Developing CPER for SARS-CoV-2
pan class="Species">Coronaviruses arn>an class="Gene">e positive-strand RNA viruses with some of the largest RNA virus genomes ranging from 26.4 to 31.7 kb. They belong to theCoronaviridae virus family in the order of Nidovirales[12]. The virus family generally causes respiratory tract infections in humans and includes common coldcoronaviruses, MERS, SARS-CoV, andSARS-CoV-2. Thedisease caused by SARS-CoV-2 is calledCoronavirus disease 2019 (COVID-19), where symptoms can include the often fatal acute respiratory distress syndrome. As of March 2021, more than 116 million people have been infected, with more than 2.5 million deaths[13].
Thpan class="Gene">e n>an class="Species">SARS-CoV-2 genome is ~30 kb long, encodes as many as 13 ORFs processed into 26 viral proteins, has short 5′ and 3′ untranslated regions (UTR) and, importantly, has a polyA tail at theend of the 3′UTR (Fig. 1a)[14]. To generateSARS-CoV-2 using CPER, viral RNA from passage 4 (P4) of the Australian SARS-CoV-2 isolate QLD02 (GISAID accession EPI_ISL_407896) was used as a template to generate first-strand cDNA (Fig. 1a). The cDNA was then used to PCR amplify six fragments that collectively encompassed theSARS-CoV-2 genome using high-fidelity DNA polymerase (Fig. 1a, b). Each fragment contained overlapping sequences of only 20 nucleotides, although thesewere selected for high GC content to allow the use of more optimal annealing temperatures (Fig. 1a). To facilitateDNA circularization and viral RNA transcription in eukaryotic cells, the linker fragment contains the last 20 nucleotides of SARS-CoV-2 3′UTR (that overlap with fragment F6), a polyA tail containing 30 adenines and thehepatitis delta virus ribozyme (HDVr) to generate the authentic 3′end of viral RNA, an SV40 polyA signal for efficient transcription termination, a spacer sequence to separate the functional elements, a CMV promoter for in vivo transcription of viral RNA by the cell RNA polymerase II, and the first 37 nucleotides of SARS-CoV-2 5′UTR (that overlap with fragment F1) (Fig. 1a).
Fig. 1
Generation of SARS-CoV-2 by CPER and characterization of properties of recovered viruses in cells and mice.
a Schematics of SARS-CoV-2 genome and overlapping SARS-CoV-2 fragments amplified from SARS-CoV-2 cDNA and circularized with a linker fragment containing the last 20 nucleotides of SARS-CoV-2 3′UTR, 30As, hepatitis delta virus ribozyme (HDVr), SV40 pA signal for transcription termination, spacer sequence, CMV promoter, and first 37 nucleotides of SARS-CoV-2 5′UTR. The resultant SARS-CoV-2 CPER product was then directly transfected into HEK293T cells, then cocultured with Vero E6 cells for virus recovery. b Agarose gel electrophoresis of PCR-amplified SARS-CoV-2 fragments 1–6 and the linker fragment showing a representative image of three experimental repeats. c Representative plaque morphologies of the wild-type (WT) SARS-CoV-2QLD02 isolate and CPER-recovered viruses (CPER1 and CPER2) stained using crystal violet at 2 dpi (top) and Immuno-plaque assay (iPA) with anti-spike protein monoclonal antibody at 14 hpi (bottom). d Growth kinetics of WT (red) and CPER-generated viruses (green and light blue) over a 3-day time course in Vero E6 cells infected at a multiplicity of infection MOI = 0.01, n = 2 independent experiments with three replicates in each, statistical analysis was performed by two-way analysis of variance with Tukey’s multiple comparisons test against WT virus. Mean values for each virus at each time point are shown ± SD. e End-point virus titers of nasal turbinates and lung tissues from JAX K18-hACE2-transgenic mice infected intranasally with 8 × 104 FFU/mouse of WT SARS-CoV-2QLD02 isolate (red) and CPER-recovered viruses (blue). At 5 days post infection, mice were sacrificed, and virus titers were determined by TCID50 assay on Vero E6 cells. For statistical analysis between WT and CPER viruses, unpaired t-test with Welch’s correction was used, p values are two-sided. Mean values for each treatment are shown ± SD. For WT virus, three biological replicates were used, for CPER-generated virus six biological replicates were used, results are from one experiment. f Full lungs from mice infected with WT virus, representative CPER virus, or infected (mock), harvested at 5 days post infection and stained with hematoxylin and eosin (H&E). Scale bar is 5 mm. g Selected H&E stained images of lung sections from mice infected with WT or representative CPER virus showing sloughing of the bronchial epithelium as indicated by the arrowhead, bronchi occluded with edema (O) and red blood cells (R). h Additional features of SARS-CoV-2 infection in lungs of infected mice showing the collapse of alveolar spaces (A) and edema. The scale bar is 100 µm. Representative images from f–h are from three independently analyzed samples for each treatment. Source data for d and e are provided in the Source Data file.
Generation of SARS-CoV-2 by CPER and characterization of properties of recovered viruses in cells and mice.
a Schpan class="Gene">ematics of n>an class="Species">SARS-CoV-2 genome and overlapping SARS-CoV-2 fragments amplified from SARS-CoV-2 cDNA and circularizedwith a linker fragment containing the last 20 nucleotides of SARS-CoV-2 3′UTR, 30As, hepatitis delta virus ribozyme (HDVr), SV40 pA signal for transcription termination, spacer sequence, CMV promoter, and first 37 nucleotides of SARS-CoV-2 5′UTR. The resultant SARS-CoV-2CPER product was then directly transfected into HEK293T cells, then coculturedwith Vero E6 cells for virus recovery. b Agarose gel electrophoresis of PCR-amplifiedSARS-CoV-2 fragments 1–6 and the linker fragment showing a representative image of threeexperimental repeats. c Representative plaque morphologies of thewild-type (WT) SARS-CoV-2QLD02 isolate andCPER-recovered viruses (CPER1 andCPER2) stained using crystal violet at 2 dpi (top) and Immuno-plaque assay (iPA) with anti-spike protein monoclonal antibody at 14 hpi (bottom). d Growth kinetics of WT (red) andCPER-generated viruses (green and light blue) over a 3-day time course in Vero E6 cells infected at a multiplicity of infection MOI = 0.01, n = 2 independent experiments with three replicates in each, statistical analysis was performed by two-way analysis of variancewith Tukey’s multiple comparisons test against WT virus. Mean values for each virus at each time point are shown ± SD. eEnd-point virus titers of nasal turbinates and lung tissues from JAX K18-hACE2-transgenic miceinfected intranasally with 8 × 104 FFU/mouse of WT SARS-CoV-2QLD02 isolate (red) andCPER-recovered viruses (blue). At 5 days post infection, micewere sacrificed, and virus titers weredetermined by TCID50 assay on Vero E6 cells. For statistical analysis between WT andCPER viruses, unpaired t-test with Welch’s correction was used, p values are two-sided. Mean values for each treatment are shown ± SD. For WT virus, three biological replicates were used, for CPER-generated virus six biological replicates were used, results are from oneexperiment. f Full lungs from miceinfectedwith WT virus, representativeCPER virus, or infected (mock), harvested at 5 days post infection and stainedwith hematoxylin andeosin (H&E). Scale bar is 5 mm. g Selected H&E stained images of lung sections from miceinfectedwith WT or representativeCPER virus showing sloughing of the bronchial epithelium as indicated by the arrowhead, bronchi occludedwith edema (O) and red blood cells (R). h Additional features of SARS-CoV-2 infection in lungs of infectedmice showing the collapse of alveolar spaces (A) andedema. The scale bar is 100 µm. Representative images from f–h are from three independently analyzed samples for each treatment. Sourcedata for d ande are provided in the SourceData file.
Full-lpan class="Gene">ength n>an class="Species">SARS-CoV-2 cDNA was assembled from the viral cDNA fragments and the linker fragment into a circular DNA in a singleCPER reaction using a high-fidelity DNA polymerase. The process does not require any of the intermediate steps commonly used in other coronavirus reverse genetics systems[2,15-17]. TheCPER reaction mix (Supplementary Fig. 1), without any further purification, was then directly transfected into HEK293T cells. To recover the virus, transfectedHEK293T cells were coculturedwith the highly permissive Vero E6 cells. Two independent transfection experiments wereperformed (CPER1 andCPER2), with both yielding infectious virus. TheCPER viruses were amplified once in Vero E6 cells to generate viral stocks. CPER3 virus was generated by transfecting theCPER reaction mix into ACE2-expressing HEK293T cells[18] and amplifying the recovered virus in Vero E6 cells.
Nanoporpan class="Gene">e sn>an class="Gene">equencing of the viral cDNA fragments PCR-amplified from P4 viral cDNA showed they had the same swarm/quasi-species sequence variation as thewild-type (WT) P4 virus (Supplementary Table 1). Comparison of the consensus genome sequences between P4 viral cDNA andPCR-amplified fragments used in CPER assembly (deposited to GenBank, ID: MW772455) did not identify any changes. Thesedata illustrate that the high-fidelity PCR used to generate the cDNA fragments that were subsequently used for theCPER reactions had faithfully amplified the original viral sequence.
Culturpan class="Gene">e of n>an class="Species">SARS-CoV-2 viruses in Vero E6 cells is well known to select rapidly for mutations, often in or around thefurin cleavage site[19-22]. This remains true for a number of reverse genetics systems[16,21,23], although deep sequencing of recovered viruses is not always provided[15,17,24-26]. Such selection was also seen for CPER viruses recovered from Vero E6 cells (coculturedwith transfectedHEK293T) (Supplementary Table 1). The selection of certain amino acid changes was different for CPER1, 2, and 3 viruses (Supplementary Table 1), with such variability in selection also reported for other reverse genetics systems[16,21,27]. As noted previously[21,22], the choice of cell lines can affect these selection processes, with our recovery of CPER3 virus using ACE2-HEK293T cells showed a decrease in the number of affected sites from 4 to 1 (Supplementary Table 1). Given theSARS-CoV-2 genomeencodes >9700 amino acids, theCPER method using ACE2-HEK293T cell thus achieved nearly complete (99.99%) amino acid sequence fidelity.
To pan class="Gene">examinn>an class="Gene">e viral properties in cells, a standard plaque assay with crystal violet staining of infected cells at 2 days post infection (dpi) was performed and showed similar plaque sizes for CPER1 andCPER2 viruses and the parental WT QLD02 isolate (Fig. 1c, top row). Immuno-plaque staining (immuno-plaque assay, iPA) with anti-spike protein monoclonal antibody CR3022 showed clear viral foci (Fig. 1c, bottom row). This methodwas used to determine viral titers (foci forming units, FFU) as it allows detection of viral foci as early as 14 h post infection (hpi). For comparison of virus replication properties in vitro, Vero E6 cells wereinfectedwith CPER1, CPER2, andWT QLD02 viruses at MOI = 0.01, with virus titers determined by iPA on 1, 2, and 3 dpi. Both CPER viruses replicatedwith comparable kinetics to the parental WT QLD02 virus (Fig. 1d).
To comparpan class="Gene">e virus propn>n>an class="Gene">erties in vivo, the JAX K18-hACE2-transgenicmouse model was used. Thesemicedevelop a respiratory disease resembling severeCOVID-19[28,29], but also present with a fulminant brain infection that is associatedwith mortality[30]. Although the virus can be found in the brain of ~20% of COVID-19patients, neurological signs and symptoms are thought to arise from systemic reactions or complicationsrather than being associatedwith extensivebrain infection[31]. Groups of three JAX K18-hACE2-transgenic micewereinfected intranasally with 8 × 104 FFU/mouse of QLD02, CPER1, andCPER2 viruses. On day 5 after infection, micewere sacrificed, and virus titers in nasal turbinates and lungs weredetermined by TCID50 assays using Vero E6 cells. Viral titers were similar for all three viruses with no significant differences in lung titers or nasal turbinate titers between WT QLD02 (n = 3 replicates) andCPER1 andCPER2 viruses (combined n = 6 replicates) (p = 0.12 and 0.2, respectively, t-tests with Welch’s correction) (Fig. 1e). Lungs were harvested at 5 dpi and processed for histology, with hematoxylin andeosin (H&E) staining showing a series of profound pathological changes that were similar between CPER-derived andWT viruses (Fig. 1f). These included bronchioles occludedwith edema, red blood cells and sloughing of the bronchial epithelium (Fig. 1g), the collapse of alveolar spaces (Fig. 1h), predominantly mononuclear cell infiltrates (Supplementary Fig. 3a), thickening of alveolar septa (Supplementary Fig. 3b), smooth muscle hypertrophy/hyperplasia (Supplementary Fig. 3c), andedema in theextracellular matrix (Supplementary Fig. 3d). Automated quantitation of lung consolidation and cellular infiltration also showed no significant differences between CPER-derived andWT viruses (Supplementary Fig. 3e, f). While the number of animals was small, collectively thedata on viral titers in nasal turbinate and lungs, and pathological changes in lungs indicate that CPER-derived and parental WT QLD02 viruses behave similarly in vivo.
Utilizing CPER for the generation of SARS-CoV-2 D614G mutant and ZsGreen-expressing reporter virus
To furthpan class="Gene">er illustn>an class="Species">rate the utility of CPER as a SARS-CoV-2 reverse genetic system, we introduced an amino acid mutation, D614G, into thespike protein of the QLD02 isolate. SARS-CoV-2 variants carrying theD614G mutation have become thedominant circulating viruses worldwide[32]. SARS-CoV-2D614G isolates have recently been shown to produce higher infectious titers in the nasal washes andtrachea of infected hamsters, suggesting that theD614G mutation may enhance viral loads in the upper respiratory tract and subsequently could increase transmission[25,26]. To introduce theD614G mutation into the genetic background of QLD02, fragment F5 was split into two overlapping fragments, F5A and F5B, each containing GAT (D) to GGT (G) codon substitution in the overlapping region (Fig. 2a). These two fragments were amplified from QLD02 viral cDNA (Fig. 2b) and incorporated into theCPER assembly instead of fragment F5. TheCPER mix contained the remaining fragments F1–F4, F6, and the linker fragment (Fig. 2a, b) and used the same cycling conditions. The mutant virus was recovered by transfecting theCPER mix into HEK293T cells and coculturing them with Vero E6 cells. TheD614G virus formed similar size plaques to the parental QLD02 isolate, but both QLD02 andD614G viruses produced slightly larger size plaques than a recent QLD935 viral isolate (Fig. 2c). QLD935 isolate is a more recent viral isolate containing theD614G mutation and other amino acid changes (NSP3C1392F, NSP3T835I, NSP12 P323L) (GISAID accession EPI_ISL_436097). TheD614G mutation was verified by Sanger sequencing of the RT-PCR-amplified fragment F5 generated from viral RNA (Fig. 2d). Replication efficiencies of D614G and QLD02 in Vero E6 cells were similar, while QLD935 exhibited reduced replication on day 1, but not on day 3 (Fig. 2e). Hence, these results show that theD614G mutation alone is not responsible for the smaller plaque size anddelayed replication of the QLD935 isolate, with other changes in the genome responsible for this phenotype.
Fig. 2
Generation by CPER and characterization of SARS-CoV-2 D614G mutant virus and ZsGreen reporter virus.
a Schematics of D614G mutant virus genome and overlapping fragments used to introduce D614G mutation by CPER. Fragment 5 is split into subfragments 5A and 5B with their overlapping region incorporating D614G mutation. b Agarose gel electrophoresis of PCR-amplified SARS-CoV-2 fragments representative of at least three experiments. c Representative plaque morphologies of wild-type (WT) SARS-CoV-2QLD02 isolate, CPER-recovered D614G mutant virus, and SARS-CoV-2QLD935 isolate naturally harboring the D614G mutation. Infected cells were stained with crystal violet at 3 dpi. d Sanger sequencing of CPER-generated D614G mutant virus (top) and wild-type SARS-CoV-2QLD02 cDNA (bottom). e Growth kinetics of SARS-CoV-2QLD02 isolate (red circle), CPER-generated D614G mutant virus (black triangle), and SARS-CoV-2QLD935 (orange square) isolate over a 3-day time course in Vero E6 cells infected at MOI = 0.01, n = 2 independent experiments with three replicates in each. Statistical tests to examine differences in growth kinetics between virus isolates were analyzed using a two-way ANOVA. The differences between each time point were analyzed with Tukey’s multiple comparisons test. Adjustments to p values were not made for multiple comparisons. ****p ≤ 0.0001. Bar graph is shown as mean values ± SD. f Schematics of SARS-CoV-2ΔORF7a-ZsGreen reporter virus and overlapping fragments used to generate this virus. Fragment 6 is split into two subfragments, 6A and 6B to generate a 95 codon deletion in ORF7a (deleted codons 14-108), and ZsGreen gene is inserted in the place of this deletion. Images of ZsGreen fluorescence of Vero E6 cells infected with MOI = 0.1 of CPER-generated reporter virus taken at 40× magnification (g) and 100× magnification (h). Representative images from g and h are from three independently analyzed samples from each treatment and two independent experiments. i Representative plaque morphologies of wild-type (WT) SARS-CoV-2QLD02 isolate and CPER-generated ZsGreen reporter virus stained with crystal violet at 2 dpi. j Growth kinetics of SARS-CoV-2QLD02 (WT, blue) and CPER-generated reporter virus (ZsGreen, green) over a 3-day time course in Vero E6 cells infected at MOI = 0.01. n = 2 independent experiments with three replicates in each. Graph is shown as mean values ± SD. k Sanger sequencing showing positions 23603-23620 of the SARS-CoV-2 QLD02 isolate corresponding to the polybasic furin cleavage site (RRAR) in the spike protein. ZsGreen1, amplified in Vero E6 cells, shows ambiguous peak at 23616 (indicated with the arrowhead) corresponding to variable deep sequencing data at this position. ZsGreen2, amplified and passaged on Vero E6-TMPRSS2 cells, shows no change to the furin cleavage site over three passages. l RT-PCR with the CoV-6F/6R primer pair of fragment 6 containing ZsGreen insertion compared to the parental QLD02 isolate showing retention of insertion over three passages in Vero E6-TMPRSS2 cells. The ZsGreen2 virus amplifies a 4094 nt fragment, whereas the WT QLD02 virus amplifies a 3689 nt fragment. Gel electrophoresis is representative of at least three experiments. m Median fluorescence intensity (MFI) values from flow cytometry analysis (Supplementary Fig. 4) normalized to viral titers (MFI/log10(FFU/mL)) of ZsGreen2 virus over three passages in Vero E6-TMPRSS2 cells. Statistical analysis was an Unpaired t-test with Welch’s correction, p values are two-tailed and shown above comparison. Data shown as individual biological replicates from three independent passage experiments (n = 3) ± SD. Source data for e, j, and m are provided in the Source Data file. The scale bar for g and h is 100 and 10 µm, respectively.
Generation by CPER and characterization of SARS-CoV-2 D614G mutant virus and ZsGreen reporter virus.
a Schpan class="Gene">ematics of n>an class="Mutation">D614G mutant virus genome and overlapping fragments used to introduceD614G mutation by CPER. Fragment 5 is split into subfragments 5A and 5B with their overlapping region incorporating D614G mutation. b Agarose gel electrophoresis of PCR-amplifiedSARS-CoV-2 fragments representative of at least threeexperiments. c Representative plaque morphologies of wild-type (WT) SARS-CoV-2QLD02 isolate, CPER-recoveredD614G mutant virus, andSARS-CoV-2QLD935 isolatenaturally harboring theD614G mutation. Infected cells were stainedwith crystal violet at 3 dpi. d Sanger sequencing of CPER-generatedD614G mutant virus (top) andwild-typeSARS-CoV-2QLD02 cDNA (bottom). e Growth kinetics of SARS-CoV-2QLD02 isolate (red circle), CPER-generatedD614G mutant virus (black triangle), andSARS-CoV-2QLD935 (orange square) isolate over a 3-day time course in Vero E6 cells infected at MOI = 0.01, n = 2 independent experiments with three replicates in each. Statistical tests to examinedifferences in growth kinetics between virus isolates were analyzed using a two-way ANOVA. Thedifferences between each time point were analyzedwith Tukey’s multiple comparisons test. Adjustments to p values were not made for multiple comparisons. ****p ≤ 0.0001. Bar graph is shown as mean values ± SD. f Schematics of SARS-CoV-2ΔORF7a-ZsGreen reporter virus and overlapping fragments used to generate this virus. Fragment 6 is split into two subfragments, 6A and 6B to generate a 95 codon deletion in ORF7a (deleted codons 14-108), and ZsGreen gene is inserted in the place of this deletion. Images of ZsGreen fluorescence of Vero E6 cells infectedwith MOI = 0.1 of CPER-generated reporter virus taken at 40× magnification (g) and 100× magnification (h). Representative images from g and h are from three independently analyzed samples from each treatment and two independent experiments. i Representative plaque morphologies of wild-type (WT) SARS-CoV-2QLD02 isolate andCPER-generated ZsGreen reporter virus stainedwith crystal violet at 2 dpi. j Growth kinetics of SARS-CoV-2QLD02 (WT, blue) andCPER-generated reporter virus (ZsGreen, green) over a 3-day time course in Vero E6 cells infected at MOI = 0.01. n = 2 independent experiments with three replicates in each. Graph is shown as mean values ± SD. k Sanger sequencing showing positions 23603-23620 of theSARS-CoV-2 QLD02 isolate corresponding to the polybasic furin cleavage site (RRAR) in thespike protein. ZsGreen1, amplified in Vero E6 cells, shows ambiguous peak at 23616 (indicatedwith the arrowhead) corresponding to variabledeep sequencing data at this position. ZsGreen2, amplified and passaged on Vero E6-TMPRSS2 cells, shows no change to thefurin cleavage site over three passages. l RT-PCR with theCoV-6F/6R primer pair of fragment 6 containing ZsGreen insertion compared to the parental QLD02 isolate showing retention of insertion over three passages in Vero E6-TMPRSS2 cells. The ZsGreen2 virus amplifies a 4094 nt fragment, whereas theWT QLD02 virus amplifies a 3689 nt fragment. Gel electrophoresis is representative of at least threeexperiments. m Median fluorescence intensity (MFI) values from flow cytometry analysis (Supplementary Fig. 4) normalized to viral titers (MFI/log10(FFU/mL)) of ZsGreen2 virus over three passages in Vero E6-TMPRSS2 cells. Statistical analysis was an Unpaired t-test with Welch’s correction, p values are two-tailed and shown above comparison. Data shown as individual biological replicates from three independent passageexperiments (n = 3) ± SD. Sourcedata for e, j, and m are provided in the SourceData file. The scale bar for g and h is 100 and 10 µm, respectively.
pan class="Gene">CPER was also employed to generate a SARS-CoV-2 reporter virus expressing ZsGreen fluorescent protein. The reporter virus was constructed by replacing codons for amino acids 14-108 of ORF7awith ZsGreen sequence, similar to the reporter SARS-CoV-2 viruses generated by other reverse genetics systems[15-17,27]. TheCPER methodology for generating ZsGreen reporter virus was carried out using the linker fragment, SARS-CoV-2 fragments F1-F5, two subfragments of fragment F6 (F6A and F6B), and a fragment containing ZsGreen with overlapping ends for fragments F6A and F6B (Fig. 2f). Again, no intermediate cloning was required. The reporter virus ZsGreen1 was recovered using the same protocol as for theD614G mutant except that RNA from passage 3 of the QLD02 virus was used to generate viral cDNA for amplifying CPER fragments. In addition, TMPRSS2-expressing Vero E6 cells were generated (Supplementary Note 1) and used in a separateexperiment to coculturewith CPER-transfectedHEK293T cells to generate the ZsGreen2 virus (Supplementary Fig. 2). Nanopore sequencing of the reporter viruses showed significantly less variation in thefurin cleavage site compared to CPER1-3 viruses, with the least variation observed in ZsGreen2 virus generated in TMPRSS2-Vero E6 cells (Supplementary Table 2).
ZsGrpan class="Gene">en>an class="Gene">en expression was detected in Vero E6 cells infectedwith the reporter virus, but not theWT virus (Fig. 2g). ZsGreen showedER localization and formed visible intracellular foci (Fig. 2h), likely because amino acids 1–13 of the signal sequence (at the N-terminus) and theER retention sequence KRKTE (at the C-terminus) of ORF7awere retained. The reporter virus formed slightly smaller plaques at 2 dpi (Fig. 2i) and showed some reduction in replication in Vero E6 cells while still replicating to high titers of ~6 × 106 focus-forming units per milliliter (FFU/mL) (Fig. 2j). To assess reporter virus stability, ZsGreen2 was passaged three times on TMPRSS2-Vero E6 cells, and viral RNA from each passagewas subjected to RT-PCR for fragment F6 to assess ZsGreen insertion and for fragment F5 to sequence thefurin cleavage site. ZsGreen insertion was stably retained, andfurin cleavage site remained unchanged in the viral RNA at all passages (Fig. 2k, l). Flow cytometry of infected cells (Supplementary Fig. 4a, b) also showed that ZsGreen fluorescence remained stable over three viral passages (Fig. 2m). Hence, the ZsGreen-expressing SARS-CoV-2 reporter virus generated by CPERwas stable for at least three passages and could be used to facilitate fundamental studies requiring live virus detection in infected cells or animals and simplify screening of diagnostic tests, neutralizing antibodies, and antiviral drugs.
Generation of RRV by CPER in mammalian and mosquito cells
pan class="Species">RRV is a mosquito-tran>an class="Disease">nsmitted, positive-strand RNA virus that belongs to a globally distributed group of arthritogenic alphaviruses (family Togaviridae) that includes chikungunya (CHIKV), Mayaro, o’nyong nyong, and Sindbis viruses[33,34]. The acutedisease manifestations associatedwith these viruses primarily involves polyarthralgia/polyarthritis, fever, rash, and/or myalgia, with arthritic manifestations often persisting for months[34]. RRV causes ~4000 cases of RRV disease in Australia annually, with an epidemic involving more than 60,000 cases in thePacific Islands occurring between 1979 and 1980[33]. The 2004–2019 global CHIKVepidemic with greater than 10 million cases across more than 100 countries on four continents[33] further highlighted the ability of these alphaviruses to spread internationally[35]. The ability to easily manipulate alphavirus genomes should facilitate studies on inter alia virus–host interactions andevasion of antiviral response, thereby accelerating thedevelopment of new interventions.
Hpan class="Gene">ern>an class="Gene">ein, the prototypeRRV T48 isolatewas used[36]. To establish a RRVCPER, primersweredesigned to generate six fragments covering the full genome of RRV T48 based on the published sequence (GenBank accession number GQ433359)[37] (Fig. 3a). Viral RNA was purified from the supernatant of RRV-infected Aedes albopictus mosquito (C6/36) cells and used to generate the first-strand cDNA, which was then used to amplify six fragments by PCR with 20–30 nucleotides overlaps (Fig. 3a, b). In addition, alpha-UTR-linker fragments were generated containing either the CMV or OpIE2 promoter, the first 22 nucleotides of the 5′UTR, the last 47 nucleotides of the 3′UTR, the polyA tract with 63 adenines and the HDVr (Fig. 3a and Supplementary Note 3). Full-length cDNAs were assembled and circularized by CPER using six fragments covering the viral genome and the corresponding linker fragment, containing either the CMV promoter or the OpIE2 promoter. RRV viruses were recovered by transfecting the respectiveCPER output into HEK293T cells (CMV promoter) or C6/36 cells (OpIE2 promoter). RRV viruses recovered from HEK293T or C6/36 cells were passaged in C6/36 cells andwere compared to the parental RRV T48 WT virus. HEK293T-derived and C6/36-derivedCPER viruses produced similar mean plaques sizes in Vero cells to those produced by WT RRV viruses (Fig. 3c). In addition, similar replication kinetics of HEK293T-derivedCPER virus andWT RRV virus were compared in Vero cells andwere shown to be similar (Fig. 3d). Thus, CPERwas successfully used to generateRRV in mammalian and mosquito cells directly from the viral RNA without any intermediate cloning.
Fig. 3
Generation by CPER and characterization of wild-type and mutant Ross River viruses, and their application for studying virus–receptor interactions.
a Schematics of RRV T48 strain genome, overlapping fragments, and linker fragment used for CPER assembly. OpIE2—insect promoter, CMV—mammalian promoter. b Agarose gel electrophoresis of PCR-amplified RRV fragments and linker fragment. A representative of at least three experiments is shown. c Representative plaque morphologies of wild-type (WT) RRV viruses in Vero cells (top left) or A. albopictus C6/36 cells (top right) as well as of CPER-recovered RRV viruses generated using CMV promoter (bottom left) or OpIE2 promoter (bottom right). Infected cells were stained with crystal violet at 3 dpi. d Growth kinetics of WT (black square) and CPER-generated (white circle) RRV viruses in Vero 76 cells over a 3-day time course infected at MOI = 0.1, n = 3 independent experiments with two replicates in each, statistical analysis to compare each time point for each of the mutants against CPER WT virus was performed by two-way analysis of variance with Tukey’s multiple comparisons test. Adjustments were not made for multiple comparisons. Mean values of the three independent experiments ± SD of the mean are shown. e Schematics of overlapping RRV fragments amplified from RRVT48 cDNA and introduced furin site mutations at the E3/E2 cleavage junction. Modified furin cleavage sites are introduced from Influenza virus H5N1 (RRVH5) and Semliki forest virus (RRVSFV) through the overlapping regions in fragments 4 and 5. Location of furin cleavage is indicated by an arrowhead. f Representative foci and plaque morphologies of CPER-generated RRV furin site mutants as shown from immuno-plaque assay at 12 hpi (top) and crystal violet staining at 3 dpi (bottom). g Growth kinetics of CPER-generated RRV furin mutants (RRVSFV; white square, RRVH5; black square, RRVCPER; red circle), in mammalian (Vero) cells (left) and A. albopictus C6/36 cells (right), n = 3 independent experiments with two replicates in each. Statistical analysis was performed by two-way analysis of variance with Tukey’s multiple comparisons test. Adjustments were not made for multiple comparisons. ****p ≤ 0.0001. Mean values of the three independent experiments ± SD of the mean are shown. h Binding assay for E1/E2 RRV mutants using purified virus and recombinant hMXRA8 receptor and represented as equilibrium-binding affinity (Kd) at 37 °C. Data used to generate the nonlinear curve are available from Source Data file, data represent two independent experimental replicates (n = 2). Summary table presented here as bar graph. i Position of the RRV mutants shown in h on the reported CHIKV-hMXRA8 structure (PDB:6J08). The heterotrimers of E1 and E2 are shown in white and blue surface representation, respectively. hMXRA8 is shown in green surface representation in the left panel, and the binding footprint shown in orange on the right. Mutated residues are indicated in pink, with labels indicating the amino acid for CHIKV at each position and corresponding residue for RRV in brackets. Source data for d, g, and h are provided in the Source Data file.
Generation by CPER and characterization of wild-type and mutant Ross River viruses, and their application for studying virus–receptor interactions.
a Schpan class="Gene">ematics of n>an class="Species">RRV T48 strain genome, overlapping fragments, and linker fragment used for CPER assembly. OpIE2—insect promoter, CMV—mammalian promoter. b Agarose gel electrophoresis of PCR-amplifiedRRV fragments and linker fragment. A representative of at least threeexperiments is shown. c Representative plaque morphologies of wild-type (WT) RRV viruses in Vero cells (top left) or A. albopictus C6/36 cells (top right) as well as of CPER-recoveredRRV viruses generated using CMV promoter (bottom left) or OpIE2 promoter (bottom right). Infected cells were stainedwith crystal violet at 3 dpi. d Growth kinetics of WT (black square) andCPER-generated (white circle) RRV viruses in Vero 76 cells over a 3-day time courseinfected at MOI = 0.1, n = 3 independent experiments with two replicates in each, statistical analysis to compareeach time point for each of the mutants against CPERWT virus was performed by two-way analysis of variancewith Tukey’s multiple comparisons test. Adjustments were not made for multiple comparisons. Mean values of the three independent experiments ± SD of the mean are shown. e Schematics of overlapping RRV fragments amplified from RRVT48 cDNA and introducedfurin site mutations at theE3/E2 cleavage junction. Modifiedfurin cleavage sites are introduced from Influenza virusH5N1 (RRVH5) andSemliki forest virus (RRVSFV) through the overlapping regions in fragments 4 and 5. Location of furin cleavage is indicated by an arrowhead. f Representative foci and plaque morphologies of CPER-generatedRRVfurin site mutants as shown from immuno-plaque assay at 12 hpi (top) andcrystal violet staining at 3 dpi (bottom). g Growth kinetics of CPER-generatedRRVfurin mutants (RRVSFV; white square, RRVH5; black square, RRVCPER; red circle), in mammalian (Vero) cells (left) and A. albopictus C6/36 cells (right), n = 3 independent experiments with two replicates in each. Statistical analysis was performed by two-way analysis of variancewith Tukey’s multiple comparisons test. Adjustments were not made for multiple comparisons. ****p ≤ 0.0001. Mean values of the three independent experiments ± SD of the mean are shown. h Binding assay for E1/E2RRV mutants using purified virus and recombinant hMXRA8 receptor and represented as equilibrium-binding affinity (Kd) at 37 °C. Data used to generate the nonlinear curve are available from SourceData file, data represent two independent experimental replicates (n = 2). Summary table presented here as bar graph. i Position of theRRV mutants shown in h on the reportedCHIKV-hMXRA8 structure (PDB:6J08). The heterotrimers of E1 and E2 are shown in white and blue surface representation, respectively. hMXRA8 is shown in green surface representation in the left panel, and the binding footprint shown in orange on the right. Mutated residues are indicated in pink, with labels indicating the amino acid for CHIKV at each position and corresponding residue for RRV in brackets. Sourcedata for d, g, and h are provided in the SourceData file.
Utilizing CPER to investigate the role of E3/E2 furin cleavage site amino acids in RRV replication
To pan class="Chemical">dn>an class="Gene">emonstrate the utility of CPER methodology to modify theRRV genome, we generated two RRV mutants with changes in thefurin-recognition site (NRSRHRR↓SV) located between E3 andE2 genes (Fig. 3e). Semliki forest virus (SFV) (GTRHRR↓TV) andH5N1influenza virus hemagglutinin furin cleavage sites (ERRRKKR↓G) were incorporated instead of RRV cleavage site to generateRRV mutants RRV-SFV andRRV-H5, respectively (Fig. 3e). For each mutant, the 3′ and 5′ ends of theCPER fragments F4 and F5, respectively, were modified by PCR with primers incorporating thedesired mutations (Fig. 3e). These modified fragments were then mixedwith fragments F2, F3, and F6 and the alpha-UTR-linker fragment to generate a full-length circularizedRRV cDNA by CPER. Mutant viruses were recovered from CPER-transfectedHEK293T cells andwere passaged in C6/36 cells. The presence of thedesired mutationswas confirmed by Sanger sequencing. RRV-H5 produced similar size plaques on Vero cells to the control CPER-recoveredRRV (RRV-CPERWT), while theRRV-SFV formed smaller size plaques (Fig. 3f). RRV-H5 virus replicated significantly moreefficiently than RRV-CPERWT virus in Vero cells, whileRRV-SFV virus replicated significantly less efficiently early (1 and 2 dpi) in infection (Fig. 3g). Both mutants replicatedwith similar kinetics to theRRV-CPERWT virus in C6/36 cells (Fig. 3g). The results show that changes in thefurin cleavage site significantly affect the replication of RRV in mammalian cells but not in mosquito cells.
Utilizing CPER to investigate the interactions of RRV with the alphavirus receptor MXRA8
Thpan class="Gene">e cn>an class="Gene">ell adhesion moleculeMXRA8was recently identified as a receptor for multiple arthritogenic alphaviruses, including RRV andCHIKV[38]. Two subsequent papers explored the structural basis of the interaction between MXRA8 andCHIKV structural proteins. Cryo-electron microscopy and cocrystallography demonstrated that theMXRA8ectodomain adopts an inverted, curved conformation that engages with both viral E1 and E2 proteins[39,40]. CHIKVE2 receptor binding determinants have been investigated using a variety of techniques[38-41]. Here, we usedCPER to generateRRVs with site-specific mutations in E1 and E2 proteins to probe for receptor interactions (Fig. 3h, i). The presence of thedesired mutationswas confirmed by Sanger sequencing. Mutant viruses were grown at scale, gradient-purified, and purity confirmed by SDS-PAGE (Supplementary Fig. 5). A recombinant humanMXRA8-Fc fusion protein was used to analyzewhole virus–receptor interactions, with an increase in Kd (reduced binding affinity) observed for viruses with theE2 mutationsH64A, H64W, andA71D (Fig. 3h). Thus, amino acids at these positionsemerge to beessential for receptor interactions for both RRV andCHIKV[40]. However, theRRV-MXRA8 interaction did not tolerateW at position 64 (the residue present in CHIKV), indicating that the interactions between E2 residue 64 andMXRA8 residues D116/R196[40] may bedistinct for the two viruses. Of the threeE1 mutants, only N141Q showed a pronounceddecrease in receptor binding, with this residue, an N-linked glycosylation site, conserved across the arthritogenic alphaviruses. The result suggests that glycosylation at this residue may play a role in modulating interactionswith MXRA8. Theseexperiments exemplify the utility of this CPER approach to investigate the role of individual residues in key host–virus interactions using replication-competent viruses.
Generation of murine and HuNoVs by CPER
HuNoVs arpan class="Gene">e positivn>an class="Gene">e-strand RNA viruses belonging to the Norovirus genus in the Caliciviridae family. They are a major cause of acutegastroenteritis in developing anddeveloped countries, causing ~220,000 deaths annually[42]. Research endeavors seeking to develop interventions to control HuNoV outbreaks are severely hampered by the inability to cultivate HuNoVs in the laboratory. Recent studies have shown that HuNoV can replicate (i) in B-cell-like cell lines only when coculturedwith specific enteric bacteria[43], (ii) in enteric organoids[44], or (iii) in zebrafish larvae[45]. The closely related genogroup V MNV remains the most widely used virus for norovirus research[46]. Noroviruses have a ~7.5 kb genome that encodes for nine or ten proteins, which have various roles in replication of the viral genome, polyprotein cleavage, translation, host manipulation, and assembly of virus particles[47]. The RNA genome is covalently attached to the viral protein VPg at its 5′ end and is polyadenylated at the 3′end[47].
pan class="Gene">CPER for n>an class="Species">MNV was generated from fragments amplified from the plasmidDNA pSPORT-T7-MNV1 containing full-length MNV cDNA clone under control of T7 promoter[48]. Three fragments covering theentireMNV genome and containing 27–34 nucleotide overlaps werePCR-amplified, assembled into full-length cDNA, and circularizedwith a linker fragment containing polyA tail with 30 adenines (Fig. 4a, b and Supplementary Note 2) by CPER. As NIH3T3 cells are moreefficiently transfectedwith DNA than RAW264.7 cells, CPERwas first transfected into NIH3T3 cells (P0), generating ~2.5 × 102 PFU/mL by 3 dpt, and then amplified in RAW264.7 cells (P1) generating ~108 PFU/mL by 3 dpi (Fig. 4c). CPER-generated virus formed similar size plaques to the original MNV isolate in RAW264.7 cells (Fig. 4d). Western blot of CPER-generated virus with anti-MNV capsid (VP1) antibodies detected theexpected 57 kDa band representing full-length MNV capsid protein (Fig. 4e).
Fig. 4
Generation of mouse and human noroviruses and insect-specific mesonivirus CASV by CPER and their characterization.
a Schematics of human norovirus (HuNoV) and murine norovirus (MNV) genomes and of overlapping fragments and linker fragment used for CPER assembly. b Agarose gel electrophoresis of PCR-amplified MNV, HuNoV, and linker fragments showing a representative image of two experiments for HuNoV and at least three for MNV. c Virus titers quantitation of MNV CPER-transfected NIH3T3 cells at 3 days post transfection (MNVCPER P0) and of virus further amplified in RAW264.7 cells at 3 days post infection (MNVCPER P1), results indicate four independent transfections and subsequent infections. d Representative plaque morphologies of wild-type MNV (MNVWT) and CPER-recovered MNV (MNVCPER). e Immunoblot analysis of mock, MNVWT and MNVCPER infected NIH3T3 cell lysates probed with anti-MNV-VP1 (top) and anti-actin (bottom) antibodies blot are representative of two experiments. f HuNoV genomic RNA (gRNA) quantitation by RT-qPCR analysis of supernatants from CPER-transfected NIH3T3 cells at 3 and 72 h post transfection from two independent experiments (n = 2). g Schematics of Casuarina virus (CASV) genome and overlapping fragments as well as linker fragment used for CPER assembly. h Agarose gel electrophoresis of PCR-amplified CASV fragments and linker fragments showing a representative image of more than three independent experiments. i Immunofluorescence analysis by confocal microscopy of A. albopictus C6/36 cells transfected with CPER products generated by three different CPER cycling conditions (CPER1, CPER2, CPER3), showing representative images of one independent experiment. CPER3 has been reproduced in two separate independent experiments. The scale bar is 50 µm. Viral antigen is visualized by staining with anti-CASV ORF2a protein monoclonal antibody C.9D7 and cell nuclei stained with Hoechst 33342. j Growth kinetics of CPER-generated (blue square) and WT CASV (red circle) in C6/36 cells infected at MOI = 0.1, n = 2 independent experiments were conducted with three replicates. Statistical analysis was performed by two-way analysis of variance with Tukey’s multiple comparisons test. Graphs in c, f, and j are mean values ± SD. Source data for c, f, and j are provided in the Source Data file.
Generation of mouse and human noroviruses and insect-specific mesonivirus CASV by CPER and their characterization.
a Schpan class="Gene">ematics of n>an class="Species">human norovirus (HuNoV) andmurine norovirus (MNV) genomes and of overlapping fragments and linker fragment used for CPER assembly. b Agarose gel electrophoresis of PCR-amplifiedMNV, HuNoV, and linker fragments showing a representative image of two experiments for HuNoV and at least three for MNV. c Virus titers quantitation of MNVCPER-transfectedNIH3T3 cells at 3 days post transfection (MNVCPERP0) and of virus further amplified in RAW264.7 cells at 3 days post infection (MNVCPERP1), results indicate four independent transfections and subsequent infections. d Representative plaque morphologies of wild-typeMNV (MNVWT) andCPER-recoveredMNV (MNVCPER). e Immunoblot analysis of mock, MNVWT andMNVCPERinfectedNIH3T3 cell lysates probedwith anti-MNV-VP1 (top) and anti-actin (bottom) antibodies blot are representative of two experiments. f HuNoV genomic RNA (gRNA) quantitation by RT-qPCR analysis of supernatants from CPER-transfectedNIH3T3 cells at 3 and 72 h post transfection from two independent experiments (n = 2). g Schematics of Casuarina virus (CASV) genome and overlapping fragments as well as linker fragment used for CPER assembly. h Agarose gel electrophoresis of PCR-amplifiedCASV fragments and linker fragments showing a representative image of more than three independent experiments. i Immunofluorescence analysis by confocal microscopy of A. albopictus C6/36 cells transfectedwith CPER products generated by threedifferent CPER cycling conditions (CPER1, CPER2, CPER3), showing representative images of one independent experiment. CPER3 has been reproduced in two separate independent experiments. The scale bar is 50 µm. Viral antigen is visualized by staining with anti-CASVORF2a protein monoclonal antibody C.9D7 and cell nuclei stainedwith Hoechst 33342. j Growth kinetics of CPER-generated (blue square) andWT CASV (red circle) in C6/36 cells infected at MOI = 0.1, n = 2 independent experiments were conductedwith three replicates. Statistical analysis was performed by two-way analysis of variancewith Tukey’s multiple comparisons test. Graphs in c, f, and j are mean values ± SD. Sourcedata for c, f, and j are provided in the SourceData file.
To asspan class="Gene">ess n>an class="Chemical">whether CPER could be used to generate HuNoV from clinical samples, total RNA was purified from a deidentifiedpatient fecal sample, and cDNA was synthesized using an oligo (dT) primer. This cDNA was used to amplify three fragments covering the HuNoV genome (Fig. 4a, b). CPERwas conducted using the same protocol and cycling conditions for MNV, followed by transfection into NIH3T3 cells. Viral RNA was then purified from cell supernatants at 3 and 72 hours post transfection (hpt), reverse transcribed and quantified by real-timePCR. A clear increase in viral RNA was observed between 3 and 72 hpt (Fig. 4f), indicating virus assembly and release. No viral RNA was detected in the mock samples (Fig. 4f). Overall, the results show the successful generation of murine and HuNoVs by CPER and illustrate the utility of CPER as a reverse genetics system for generating replication-competent virus from clinical samples.
Generation of Casuarina virus by CPER
pan class="Species">Casuarina virus (n>an class="Species">CASV, Alphamesonivirus 4) is an insect-specific virus belonging to the Mesoniviridae family, order Nidovirales[49]. Mesoniviruses have a positive-sense single-stranded 20–21 kb RNA genome, with a comparable gene organization to thecoronaviruses[50], which also belong to the order Nidovirales. The virion architecture of mesoniviruses is similar to that of thecoronaviruses, with large globular heads attached to low-density stalks protruding from theenveloped virion[49,51]. In contrast, mesoniviruses replication is restricted to mosquitoes, with no ability to replicate in vertebrate cells[49,52]. The factors determining mosquito host specificity of CASV remain unknown. A simple andeasy-to-useCASV reverse genetic system is needed to gain insights into the role of these factors.
As for pan class="Species">SARS-CoV-2, thn>an class="Gene">e large size of theCASV genome (~20 kb) complicates conventional reverse genetics methods. We, therefore, developed theCPER methodology as a simple reverse genetics system for manipulating theCASV genome. Five fragments encompassing the completeCASV genomewith 30–43 nucleotides overlaps were amplified from viral cDNA (Fig. 4h). A linker fragment was also generated (Fig. 4g, h and Supplementary Note 4) that contained the OpIE2 promoter, the first 25 nucleotides of the 5′UTR, the last 40 nucleotides of the 3′UTR, a polyA tail with 53 adenines, and the HDVr. The full-length CASV cDNA was then assembled and circularized by CPER. Threedifferent CPER conditions that differed in the length of extension time and the number of cycles were tested, with all of them yielding infectious virus. The relativeefficiencies of different CPER conditions and the identity of CASV in transfected C6/36 cells were shown by immunofluorescence assay (IFA) using an anti-CASV monoclonal antibody (Fig. 4i). CPER-recoveredCASV replicatedwith comparableefficiency to theWT CASV in C6/36 cells (Fig. 4j). The results demonstrate the utility of CPER as a robust reverse genetic system for CASV.
Discussion
Hpan class="Gene">ern>an class="Gene">ein, weestablished a CPER-based methodology for the construction of SARS-CoV-2 andCASV anddemonstrated the utility of this method as a versatile reverse genetics platform for polyA tail-containing viruses with large genomes. Six or seven PCR-generated fragments, up to 5.5 kb in length with ~20–40 nucleotides overlap, are assembled in vitro into a full-length circular DNA with the aid of a linker fragment that containseukaryotic expression promoter upstream of the viral 5′end and polyA tail downstream of the viral 3′end followed by the HDV ribozyme. The incorporation of a polyA tail at theend of the viral 3′UTR and immediately prior to the HDV ribozyme allows to generate authentic viral 3′endwith polyA tail, while SV40 polyA signal downstream of HDVr ensures efficient transcription termination[8,53]. The 30 and 20 kb viral RNA genomes are then generated in the nucleus of cells transfectedwith CPER-generatedDNA andexported into the cytoplasm, where they initiate the generation of authentic viruses, i.e., viruses with replicative and pathogenic characteristics (for SARS-CoV-2) indistinguishable from WT viruses. Remarkably, theCPER methoddescribed herein reliably and rapidly produces authentic viruses despite (i) the large genome sizes that would likely exacerbate misassembly errors, (ii) the presence of nuclear RNA splicing activities[54], (iii) the multiple processes required for nuclear export and translation of RNA[55], and (iv) the presence of host cell enzymes that degrade viral RNA[56,57]. Thus, genome sizedoes not appear to be a limitation for theCPER methodology, which bodes well for future studies of RNA viruses with larger genomes, such as the recently discoveredplanarian secretory cell nidovirus that has a ~41 kb genome[58]. The ability to recover by CPER RNA viruses that have large genome and that contain polyA tail represent a significant advance from our previous CPERdevelopments for flaviviruses, which have relatively small (11 kb) genome anddo not contain polyA tail.
An altpan class="Gene">ern>an class="Gene">native reverse genetic system has been developed and used for generating pathogenic coronaviruses and for manipulating theSARS-CoV-2 genome, and is in someways similar to CPER[15,17,24,26]. The method involves an in vitro ligation for assembling full-length cDNA from a set of cloned cDNA fragments. RNA is then generated by in vitro transcription and is then transfected into cells to generate an infectious virus. The method involves cloning each viral cDNA fragment into a plasmid vector and propagation in bacteria, whereas CPER generates these fragments by RT-PCR. Some of theSARS-CoV-2 fragments in the in vitro ligation protocol were unstablewhen cloned and propagated in bacteria and therefore required the use of low copy number plasmid vectors, different Escherichia coli strains, and lower incubation temperature for more stable propagation[24]. Similar instability for some of theSARS-CoV-2 fragments was also encountered in YAC reverse genetics system[16]. The fragments for the in vitro ligation method are then liberated from the amplified plasmids using class IIS restriction enzymes, with unique recognition sites engineered at theend of each fragment. Fragments are then ligated using T4 DNA ligase by cohesiveends left after restriction enzymedigestion. In contrast, theCPER method anneals fragments by ~20–40 nucleotide overlaps, uses DNA polymerase to fill in the remaining fragment sequences, and generates circular dsDNA ready for transfection, all in one reaction. After transfection of the completedCPER reaction, the method utilizes cellular enzymes to repair the remaining “nicks” in thedsDNA and to transcribe the viral RNA[6-8]. Therefore, the primary advantages of theCPER approach are that the steps of engineering special restriction sites, cloning, bacterial propagation, in vitro ligation, and in vitro RNA transcription are not required.
Gpan class="Gene">enn>an class="Gene">eration of viral mutants by theCPER methodology described herein is a simple and rapid process and is exemplified for SARS-CoV-2 (Fig. 2a–e) andRRV (Fig. 3e–i). For example, to generate theD614G mutant of SARS-CoV-2, fragment F5 was replacedwith two subfragments F5A and F5B. F5A was generated using the same forward primer as for F5 and a reverse primer containing the nucleotide substitution coding for the new amino acid. F5B was generated by using the same reverse primer as for F5 and a forward primer that is complementary to the aforementioned primer with the substitution. The subfragments were then usedwith the rest of the fragments in theCPER assembly. Generating recombinant viruses with deletions and/or insertions can be achieved by generating subfragments with appropriate primers. For example, to develop theSARS-CoV-2-ΔORF7a-ZsGreen virus, the ZsGreen reporter genewas generated by replacing fragment F6 with two subfragments, F6A and F6B, flanking theORF7adeletion, and adding the ZsGreen fragment containing overlapping ends with F6A and F6B subfragments (Fig. 2f). Conceivably, multiple mutations, deletions and/or insertions can be simultaneously introduced by CPER by swapping multipleWT fragments with corresponding fragments containing thedesired mutations/deletions and/or by adding desired insertion fragments. However, we have not explored the upper limit of the number of cDNA fragments that can be successfully assembled using CPER, with each mutation/insertion requiring the addition of another fragment to theCPER reaction. If multiple mutations are needed in one fragment, this fragment could be generatedsynthetically and incorporated into theCPER reaction. We have shown previously that an infectious virus can be generatedentirely from synthetic DNA fragments using CPER[7]. If mutations are needed in multiple fragments, a mixture of synthetic and split fragments (e.g., F5a and F5b in Fig. 2a) might beenvisaged. Clearly, cDNA fragments derived from one virus isolate can also be used together with cDNA fragments from another virus isolate (and/or synthetic fragments) to generate viruses with a combination of desired mutations[8]. Synthetic fragments could also, for instance, be used to incorporate clusters of repeated sequences (e.g., microRNA target sites) as long as primers to amplify the fragments are located outside the repeat regions.
Thpan class="Gene">e vn>an class="Gene">ersatility of theCPER methodology was further illustrated herein by generating five polyA tail-containing positive-strand RNA viruses from four different virus families. In addition, the generation of viruses using insect cells instead of mammalian cells was illustrated for RRV (Fig. 3) andCASV (Fig. 4g–j) by incorporating an insect-specific promoter instead of a mammalianexpression promoter for transcription of viral RNA. CPERwas also applied to generate virus directly from RNA purified from a clinical sample, a process illustrated herein for HuNoV (Fig. 4a). This is particularly useful for viruses like HuNoV, which cannot be readily isolated by in vitro culture.
pan class="Chemical">Whiln>an class="Gene">e theCPER method is rapid and robust, it does have certain limitations. It relies on efficient DNA transfection; hence, the highly transfectableHEK293T cells provide an optimal choice. However, HEK293T cells do not support robust SARS-CoV-2 infection. Hence, the coculture of CPER-transfectedHEK293T cells with Vero E6was used to recover high virus titers. Propagation of SARS-CoV-2 in Vero E6 cells can lead to the selection of mutations[19,22], an issue seen in other reverse genetics systems[16,21,27] and observed herein for CPER-generated viruses. Our data support the view that choice of cell lines for SARS-CoV-2 recovery/amplification such as ACE2- and/or TMPRSS2-expressing cells[21,23] may reduce the selection of unintended mutations. In addition, using a lower passage virus with more homogeneous sequence as the source of cDNA to amplify fragments for CPER assembly could further alleviate the selection of minor variants. On the other hand, the ability of CPER to faithfully recapitulate viral quasi-species[6] may be beneficial for studies on virus adaptation andevolution.
In summary, pan class="Chemical">wn>an class="Gene">e haveestablished a simple, rapid, and versatileCPER-based reverse genetic platform for a range of positive-strand RNA viruses with polyA tails. We havedemonstrated its practical utility in various settings, particularly its application for viruses with large genomes likeSARS-CoV-2. As newSARS-CoV-2 variants continue to emerge[32,59,60], the ability to easily and rapidly manipulate such a large viral genome by CPERwill greatly expedite identifying newdeterminants of transmission and pathogenesis.
Methods
Cells
pan class="Species">Human n>an class="Disease">embryonic kidney 293T cells (HEK293T), African green monkey kidney cells Vero E6 and Vero 76 cells, murine macrophageRAW264.7, andmurine fibroblast NIH3T3 cells were maintained at 37 °C in Dulbecco modified Eagle medium (DMEM; Gibco) supplementedwith 10% fetal calf serum (FCS; Gibco), 1% GlutaMAX (Gibco), and 1% sodium pyruvate (100 mM; Gibco). A. albopictus C6/36 cells were grown at 28 °C in Royal Park Memorial Institute (RPMI) medium (Gibco, USA), 10% FCS (Gibco), and 1% GlutaMAX (200 mM; Gibco). Following viral infections, cells were maintained in corresponding media with 2% FCS and 10,000 U/mL of penicillin and 10,000 μg/mL of streptomycin (Gibco, USA).
Vpan class="Gene">ero n>an class="CellLine">E6-TMPRSS2 cells were generated by transduction with lentivirus containing puromycin-selectable codon-optimizedhumanTMPRSS2 construct and validatedwith anti-TMPRSS2 antibody (Abcam, ab109131) (full details are available in Supplementary Note 1). HEK293T-hACE2 cells were provided by Jesse Bloom (Fred Hutchinson Cancer Research Centre, Washington, USA)[18].
Viruses
Thpan class="Gene">e n>an class="Species">SARS-CoV-2 isolates were sequenced by Dr D. Warrilow (Queensland Health, Brisbane, Australia), hCoV-19/Australia/QLD02/2020 (QLD02) (GISAID Accession ID; EPI_ISL_407896) andhCoV-19/Australia/QLDID935/2020 (QLD935) (GISAID Accession ID: EPI_ISL_436097), named here as QLD02 and QLD935, respectively. The viruses were isolated from patient nasopharyngeal aspirates via inoculation of Vero E6 cells and virus stockswere then produced in Vero E6 cells. This work receivedethical clearance from the Queensland Health Forensic and Scientific Services HumanEthics Committee (EC00305), approval reference HEC 21-08. Passage 3 or 4 of QLD02 isolate on Vero E6 cells was used for purifying viral RNA and generating cDNA. All thework with infectious SARS-CoV-2 andCPER-generatedSARS-CoV-2 recombinants not involving animals was performed in a certifiedPC3 facility at The University of Queensland (UQ) and approved by the UQ Institutional Biosafety Committee (UQ IBC, approvals IBC/390B/SCMB2020 and IBC/1301/SCMB/2020). Work with RRV, CASV, and corresponding CPER-recovered viruses was performed in PC2 facilities at SCMB, UQ, and approved by UQ IBC (IBC/1205/SCMB/2018 and IBC/1289/SCMB/2020, respectively). Generation of MNV/HuNoV CPERwas approved by the University of Melbourne Institutional Biosafety Committee (approval IBC 2019.035). Ross River virus prototype strain T48 was initially isolated from Aedes vigilax mosquitoes collected near Ross River (Townsville, Australia) in 1959[61] and a stock produced in C6/36 cells. MNV1 was obtained from H. Virgin (Washington University School of Medicine) and propagated in RAW264.7 cells. Casuarina virus strain 0071 (GenBank accession number: NC_023986) was originally isolated from Coquillettidia xanthogaster mosquitoes collected in Darwin in 2010[49], and a virus was passaged 1–2 times in C6/36 cells to produce viral stock for RNA extraction.
Viral RNA extraction
For pan class="Species">SARS-CoV-2 ann>an class="Chemical">d RRV, 15 mL of supernatant from infected cells were concentrated to ~250 µL by using the Amicon 100 kDa column (Merck Millipore, USA). RNA from the concentrated virus was extracted using TRIzol LS reagent (Thermo Fisher Scientific, USA) as per the manufacturer’s protocol. For HuNoV, a deidentified HuNoV GII.4 positive fecal specimen (10% w/v in Hanks media) was provided by the Victorian Infectious Diseases Reference Laboratory under humanethics approval HESC 1749300.1. Viral RNA was extracted using TRIzol reagent (Invitrogen, USA) as per the manufacturer’s protocol. For CASV, viral RNA was extracted from the culture supernatant of infected C6/36 cells using a NucleoSpin RNA virus kit (Macherey-Nagel).
cDNA synthesis
For pan class="Species">SARS-CoV-2, viral RNA was used to prepare cDNA with Protoscript II first-strand cDNA synthesis kit and random primer mix containing a mixture of hexamers and an anchored-dT primer (dT23VN) as per manufacturer’s protocol (NewEngland Biolabs, USA). All primers used in this study are shown in Supplementary Table 3 with individual primer names referred to throughout the text. For RRV, viral RNA was used to prepare cDNA with an RRV-3′UTR- reverse primer and SuperScript III First-Strand Synthesis System (Thermo Fisher Scientific, USA). For HuNoV, viral RNA was used to prepare cDNA with Oligo(dT)23 primer (Sigma-Aldrich) and First-Strand Synthesis System (Thermo Fisher Scientific, USA). For CASV, viral RNA was used as a template for cDNA synthesis using SuperScript IV (Thermo Fisher Scientific) with priming by random hexamers (Promega, USA).
PCR amplification of DNA fragments for CPER assembly
SARS-CoV-2 fragments
Six pan class="Species">SARS-CoV-2 fragmn>an class="Gene">ents (Fig. 1a) were amplified from viral cDNA using high-fidelity PrimeStar GXL DNA polymerase (Takara Bio) and corresponding pairs of primers. For generating D614G virus, two subfragments of fragment 5, 5A, and 5B (Fig. 2a) were generated from viral cDNA with primers incorporating GAT to GGT change at codon 614 in thespike gene. For generating ZsGreen reporter virus, two subfragments of fragment F6 were generated from viral cDNA, subfragment F6A ending at codon 13 in the open reading frame 7 (ORF7a) and subfragment F6B starting at codon 109 of ORF7a (Fig. 2f). ZsGreen with the addition of overlapping regions for subfragments F6A and F6B (Fig. 2f) was generated by PCR from pCCI-SP6-ZIKV-ZsGreen, kindly provided by Dr A. Merits (Tartu University, Tartu, Estonia).
RRV fragments
Six ovpan class="Gene">erlapn>ping fragmn>an class="Gene">ents (Fig. 3a) were amplified from RRV viral cDNA using Q5 High-Fidelity DNA Polymerase (NewEngland Biolabs, USA). To generate thefurin site mutants (Fig. 3e), the 3′ and 5′ ends of theCPER fragments F4 and F5 were modified in two steps. Step 1, new fragments F4 and F5 that flank thefurin cleavage sites were amplified from cDNA. Step 2, fragments F4 and F5 from the first stepwere used as a template to create new amplicons. To create amino acid mutations in E2 and E1 genes (Fig. 3h), the newCPER fragments F4 and F5 for E2 mutations and F5 and F6 for E1 mutationswere generated from the viral cDNA by Q5 High-Fidelity DNA polymerase.
HuNoV and MNV fragments
Thrpan class="Gene">en>an class="Gene">e overlapping HuNoV fragments (Fig. 4a) were amplified from viral cDNA, while the three overlapping MNV fragments (Fig. 4a) were amplified from the full-length MNV clone pSPORT-T7-MNV1[48]. High-fidelity PrimeStar GXL DNA Polymerase (Takara Bio) was used for all PCR amplifications.
CASV fragments
Five overlapping fragments (Fig. 4g, h) were amplified from viral cDNA using high-fidelity PrimeStar GXL DNA polymerase (Takara Bio).
Linker fragments
Linker fragments for each viral CPER assembly were generated as described in Supplementary Notes.
CPER reactions, transfections, and virus recovery
SARS-CoV-2
pan class="Chemical">Purifin>an class="Gene">edSARS-CoV-2 cDNA fragments andSARS-CoV-2 linker fragment (Supplementary Note 2) were mixed in equimolar amounts (0.1 pM each) in a 50 µL reaction volume containing 200 µM of each dNTP, 1x PS GXL reaction buffer, and 2 μL of PrimeStar GXL DNA polymerase. The following cycling conditionswere used: initial denaturation at 98 °C for 30 s, followed by 12 cycles of denaturation at 98 °C for 10 s, annealing at 55 °C for 20 s, andextension at 68 °C for 10 min, followed by a final extension at 68 °C for 10 min. For the generation of theD614G mutant, subfragments F5A and F5B were used together with other SARS-CoV-2 fragments and the linker fragment in CPER reaction using the same cycling conditions. For generating ZsGreen reporter virus, subfragments F6A and F6B and ZsGreen fragment were used together with the other SARS-CoV-2 fragments and the linker fragment in theCPER reaction using the same cycling conditions. CPER reactionswere then used for transfectionswithout any purification. Two independent transfection experiments (CPER1 andCPER2) wereperformed by transfecting HEK293T cells with 50 or 25 µL of CPER reactions in a 6-well or a 12-well plate, respectively, using Lipofectamine LTX Plus reagent (Invitrogen, USA) as per manufacturer’s protocol. Six hours post transfection, cells were trypsinized and transferred to a confluent monolayer of Vero E6 cells in a six-well plate. Supernatants containing viruses were harvestedwhen cytopathic effect (CPE) was pronounced (6–9 days post transfection, reaching titers of up to ~106 FFU/mL as determined by iPA described below) and amplified once on Vero E6 cells to generate viral stockswith titers of ~5 × 106 − 107 FFU/mL. CPER3 virus was generated by transfecting HEK293T-ACE2 cells in a six-well platewith CPER mix using Lipofectamine LTX Plus reagent. This resulted in the recovery of 1.8 × 105 FFU/mL CPER3 virus in the culture fluid by day 8 after transfection (P0). The secreted virus was then amplified in T75 flask of Vero E6 cells for 3 days and reached the titer of 2.3 × 107 FFU/mL (P1).
For rpan class="Gene">en>an class="Species">covery of D614G mutant virus anddORF7a-ZsGreen (ZsGreen) reporter viruses, CPER fragments were amplified from passage 4 QLD02 cDNA (D614G) or passage 3 QLD02 cDNA (ZsGreen) and assembled by CPER as above. CPER reactions (50 µL) were then transfected into HEK293T cells in a six-well plate using Lipofectamine LTX Plus reagent, and at 6 hpt cells were trypsinized and transferred to a T75 flask of confluent Vero E6 cells (for D614G and ZsGreen1) or Vero E6-TMPRSS2 cells (for ZsGreen2). Viruses were harvested from supernatants when CPEwas pronounced (6–8 days post transfection), and virus titers were 1.1 × 107 FFU/mL for D614G mutant, 2.5 × 106 FFU/mL for ZsGreen1 virus, and 5 × 106 FFU/mL for ZsGreen2 virus. No additional amplification in Vero E6 cells was performed for these viruses.
RRV
pan class="Gene">Equimolar amounts (0.1 pM n>an class="Gene">each) of six RRV RT-PCR fragments andeither CMV promoter- or OpIE2 promoter-containing alpha-UTR-linker fragments (Supplementary Note 3) were assembled into a circular full-length cDNA using the Q5 High-Fidelity DNA Polymerase (NewEngland Bio Lab, USA) with the following cycling conditions: 98 °C for 2 min followed by 12 cycles of 98 °C for 10 s, 60 °C for 15 s, 68 °C for 10–13 min, and a final extension at 68 °C for 10–13 min. Subsequently, theCPER reactionswere transfecteddirectly without any purification into HEK293T cells (CMV promoter) or C6/36 cells (OpIE2 promoter). Transfection into HEK293T cells was performed using theLipofectamine LTX Plus reagent (Invitrogen, USA), while transfection into C6/36 cells was performed using TransIT-LT1 reagent (Mirus Bio, USA). Viruses were harvested from the supernatants of transfectedHEK293T or C6/36 cells and amplified on C6/36 cells to generate viral stocks. Virus titers weredetermined by iPA with mouse anti-E1 monoclonal antibody G8[62].
HuNoV and MNV
For both HuNoV anpan class="Chemical">d n>an class="Species">MNV, 0.1 pM of each of the three respective cDNA fragments and corresponding linker fragments (Supplementary Note 2) were combined in CPER reaction with PrimeSTAR GXL DNA polymerase (Takara Bio) and subjected to the following cycling conditions: 98 °C for 30 s, followed by 12 cycles of 98 °C for 10 s, 55 °C for 30 s, 68 °C for 10 min, and a final extension at 68 °C for 10 min. TheentireCPER reaction was transfectedwith Lipofectamine 3000 into NIH3T3 cells seeded in a well of a six-well plate. At 3 and 72 hpt, cell supernatant was collected, clarified from cell debris by centrifugation at 500 × g for 5 min and stored at −80 °C for passaging, plaque assays and RT-qPCR.
CASV
pan class="Gene">Equimolar amounts (0.1 pmol n>an class="Gene">each) of the fiveCASV fragments and theCASV OpIE2 linker fragment (Supplementary Note 4) were used in a CPERwith PrimeStar GXL DNA polymerase (Takara Bio) in a 50 µL reaction. Threedifferent cycling conditionswere used to produceCPER1, CPER2, andCPER3 reactions. TheCPER1 reaction was cycled at 98 °C for 2 min, followed by 20 cycles of 98 °C for 10 s, 55 °C for 15 s, 68 °C for 25 min, and a final extension of 68 °C for 25 min. TheCPER2 reaction was cycled at 98 °C for 2 min, followed by 35 cycles of 98 °C for 10 s and 68 °C for 15 min, and a final extension at 68 °C for 25 min. TheCPER3 reaction was cycled at 98 °C for 2 min, followed by 35 cycles of 98 °C for 10 s, 55 °C for 15 s, and 68 °C for 15 min, followed by a final extension at 68 °C for 15 min. Subsequently, 25 µL of the 50 µL CPER reaction was transfected into C6/36 cells in a six-well plate using TransIT-LT1 (Mirus Bio), according to the manufacturer’s instructions. At 7 days post transfection, the supernatant was harvested and passaged onto C6/36 cells. Immediately following the removal of the supernatant, transfected cells were assessed by IFA.
Nanopore sequencing and bioinformatic analysis
For pan class="Chemical">wholn>an class="Gene">e-genome sequencing of SARS-CoV-2 isolates, CPER-generated viruses, and zsGreen reporter viruses, thenCoV-2019 sequencing protocol v3 (Josh Quick, University of Birmingham) was usedwith minor modifications. Briefly, cDNA generated from RNA isolated from cell culture supernatant was amplified using with ARTIC network v2 primers using two-stepPCR amplification with Q5® High-Fidelity DNA Polymerase (NewEngland Biolabs, USA). For amplification of whole-genome zsGreen sequencing, the ORF7_zsGreen primer pair was spiked into PCR reactions (0.030 µM) to recover bridging amplicons up anddownstream of theORF7a replacement. For sequencing of CPERDNA, theCPER fragments were useddirectly without additional PCR amplification.
pan class="Chemical">PCR fragmn>an class="Gene">ents were purified using AMPure XP beads (Beckman Coulter, USA) and subjected to End Repair/dA-Tailing using the NEBNext® Ultra™ II Module (NewEngland Biolabs, USA) and multiplexed using theNative Barcoding Expansion kit (EXP-NBD104, Oxford Nanopore, UK) and Ligation Sequencing Kit (SQK-LSK109, Oxford Nanopore, UK). Prepared libraries were then quantified using the Qubit™ dsDNA HS Assay Kit (Thermo Fisher Scientific, USA) and loaded into equimolar concentrations totaling 20 fmol into a Flongle flow cell (FLO-FLGOP1, Oxford Nanopore, UK). Base-calling and adapter trimming of fast5 files wereperformed using the guppy_basecaller (Version: 3.3.0+ef22818). Clean fastq base-called files were subsequently mapped to theSARS-CoV-2 isolate QLD02 (GISAID accession EPI_ISL_407896) using the Burrows–Wheeler Aligner (Version: 0.7.13-r1126)[63] using theflags (mem -x ont2d). Depth of coverage of mapped alignment files was determined using samtools (v1.3) depth. Single-nucleotide variants of alignment files were identified using iVar (v1.2.2)[64] with a minimum quality score threshold of 20. Coverage and frequencies of variant positionswere visualized and calculated using Integrative Genomics Viewer (Version: 2.7.0)[65]. Base-called fastq data generated in this study have been deposited in the Sequence Read Archive hosted by theNational Center for Biotechnology Information with accession number PRJNA707404.
Standard plaque assays
pan class="Species">SARS-CoV-2 ann>an class="Chemical">d RRV plaque assays were conducted on Vero E6 and Vero 76 cells, respectively. The overlay medium for both contained 0.375% low-melting point agarose (Bio-Rad, USA). Cells were fixed at 2 or 3 dpi with 4% formaldehyde and stainedwith 0.2% crystal violet solution. RAW264.7 cells were seeded onto 12-well plates and incubated at 37 °C for 24 h until 70% confluent. Cell culture supernatants containing virus were tenfold serially diluted in DMEM, inoculated onto cell monolayer in duplicate, and incubated at 37 °C for 1 h. Following inoculation, overlay medium (70% DMEM, 2.5% [vol/vol] FCS, 13.3 mM NaHCO3, 22.4 mM HEPES, 200 mM GlutaMax, and 0.35% [wt/vol] low-melting-point agarose) was added to each well, solidified at 4 °C, and incubated at 37 °C. After 48 h, cells were fixed in 10% formalin for >1 h and overlay removed. Fixed-monolayers were stainedwith 0.2% crystal violet (80% PBS, 20% methanol) for 30 min and plaques enumerated[66].
Immuno-plaque assays
Viral titpan class="Gene">ers n>an class="Chemical">weredetermined using an optimized iPA for SARS-CoV-2[67]. Briefly, 4 × 104 cells per well of Vero E6were seeded in 96-well plates and incubated for 14 h to reach 100% confluence. Samples were tenfold serially diluted in DMEM supplementedwith 2% FCS and 10,000 U/mL of penicillin and 10,000 μg/mL of streptomycin (Gibco, USA) and 25 μL of each dilution incubatedwith the cells for 30 min at 37 °C. Then, 175 μL of overlay medium was added to each well without removing the inoculum. After 14 h post infection, the overlay medium was removed, and the monolayer was fixed by adding cold 80% acetone and kept for 30 min at −20 °C. Next, acetonewas removed, and the platewas fully dried for 2 h. Subsequently, viral foci in fixedinfected cells weredetected using themouse monoclonal antibody CR3022[67]. The plates were first blocked for 60 min at 37 °C by adding 150 μL of blocking solution Pierce Clear Milk Blocking buffer (Thermo Fisher Scientific Inc, USA, Cat. No. 37587) to 96-well plates. Following, fixed cells were probedwith primary mAb for 1 h at 37 °C, using 50 μL/well (50 μg/well). After that, plates werewashed five times with phosphate buffered saline (PBS) containing 0.05% Tween-20with 5 min incubating in between each wash. Subsequently, a fluorophore-conjugated goat anti-mouse IgG secondary antibody (IRDye® 800CW, LI-COR, USA, Cat. No. 926-32210) was added to each well (20 ng/well) and incubated for 1 h at 37 °C. Then, plates werewashed five times anddried before scanning. Finally, plates were scanned using the LI-COR Biosciences Odyssey Infrared Imaging System (Odyssey CLx, Li-COR, USA). Image acquisition of plates was undertaken using Image Studio Lite (v 5.2.5) software under the following settings: resolution 42 μm, quality medium and focus 3 mm. The foci were counted in each well and viral titers were calculated andexpressed as FFU/mL. RRV iPAs wereperformed at 12 hpi using the same method and anti-RRVE1 protein monoclonal antibody G8. Viral titers determined by iPAs wereexpressed in FFU/mL.
Viral growth kinetics
Growth kinetics for SARS-CoV-2wereperformed on Vero E6 cells and MOI = 0.01. Growth kinetics for RRVwereperformed on Vero 76 or C6/36 cells at MOI = 0.1. Growth kinetics for CASVwereperformed on C6/36 cells at MOI = 0.1.
ZsGreen live fluorescence and flow cytometry of SARS-CoV-2 ZsGreen reporter virus-infected cells
For livpan class="Gene">e fluorn>an class="Gene">escence, Vero E6 cells were seeded on coverslips in 24-well plates andinfectedwith MOI = 0.1 of SARS-COV-2 ZsGreen virus or WT QLD02 isolate. Coverslips with infected cells were fixed for 30 min at 4 °C with 4% paraformaldehyde (PFA) at 1 and 2 dpi andwashedwith PBS. The ZsGreen fluorescencewas ether visualizeddirectly on Nikon Eclipse Ts2 Epi-fluorescent Inverted Microscope at 40× magnification, and 488 nm filter (Fig. 2g), or coverslips were first mounted on glass slides in ProLong Diamond Antifade Mountant with DAPI (Invitrogen, USA) and then imaged on ZEISS LSM 710 laser scanning confocal microscope at 100× magnification with oil emersion and 488 nm filter for ZsGreen or 405 nm filter for DAPI (Fig. 2h).
For flopan class="Chemical">w cytomn>an class="Gene">etry, Vero E6-TMPRSS2 cells were seeded at 1 × 106 cells/well overnight andinfectedwith SARS-COV-2 ZsGreen2 virus at MOI 0.1. Two days post infection, the supernatant was used to passage the virus two more times on Vero E6-TMPRSS2 using MOI = 0.1. In each passage, supernatants and cells were collected. Supernatants were used for virus quantification by iPA, and cells were used to analyze ZsGreen expression by flow cytometry. The cells werewashedwith PBS, treatedwith 500 μL of trypsin for 5 min at 37 °C, spun at 500 g for 5 min, and suspended in 500 μL of DMEM (containing 2% FCS andP/S). Cells were fixedwith 4% PFA for 30 min. To remove thePFA, cells were spun down at 500 g for 5 min and suspended in 1 mL of the ice-coldPBS containing 5% heat-inactivated FCS once. Lastly, cells were spun down at 500 g for 5 min and then suspended in ice-coldPBS and passed through 70 μm mesh filter prior to BD AccuriTM C6 analysis. Data were then analyzed using Flowjo v10.
SARS-CoV-2 mouse infections and determination of viral titers in organs and tissues
pan class="Species">Mouse n>an class="Gene">experiments were approved by the QIMR Berghofer MRI Biosafety Committee and Animal Ethics Committee (project P3600) and conducted in accordancewith the “Australian Code for the care and use of animals for scientific purposes” as defined by theNational Health and Medical Research Council of Australia. Work was conducted in a biosafety level-3 (PC3) facility at the QIMR Berghofer MRI (Australian Department of Agriculture, Water and theEnvironment certification Q2326 and Office of the Gene Technology Regulator certification 3445). Heterozygous K18-hACE2-transgenic mice (B6.Cg-Tg(K18-ACE2)2Prlmn/J, The Jackson Laboratory, Bar Harbor, ME, USA) were bred in-house by crossing with C57BL/6J mice (Animal Resources Center, Canning Vale, WA, Australia). DNA from the tail was isolated using Extract-N-Amp TissuePCR kit (Sigma) andPCR genotyping undertaken as described (The Jackson Laboratory. Genotyping protocols database, Protocol 37654) except using primers forward—5′-CTTGGTGATATGTGGGGTAGA-3′, reverse 5′ CGCTTCATCTCCCACCACTT-3′ (recommended by NIOBIOHN, Osaka, Japan). Thermocycling conditionswere 94 °C 3 min, 35 cycles of 94 °C 30 s, 55.8 °C 30 s, 72 °C 1 min, and final extension of 72 °C 10 min. The conditions themicewere kept are as follows: light = 12:12 h dark/light cycle, 7:45 a.m. sunrise and 7:45 p.m. sunset, 15 min light dark anddark light ramping time. Enclosures: M.I.C.E cage (Animal Care Systems, Colorado, USA). Ventilation: 100% fresh air, eight complete air exchange/h/rooms. In-houseenrichment: paper cups (Impact-Australia); tissue paper, cardboard rolls. Bedding: PuraChips (Able scientific) (aspen fine). Food: Double bagged norco rat andmousepellet (AIRR, Darra, QLD). Water: deionizedwater acidifiedwith HCl (pH = 3.2). QLD02 andCPER-derived viruses were used to inoculate male 6–8-week oldK18-hACE2mice (n = 3 per group) intranasally with 8 × 104 FFU of each virus in 50 μL medium while under light anesthesia; 3% isoflurane (Piramal Enterprises Ltd., Andhra Pradesh, India) delivered using The Stinger, Rodent Anesthesia System (Advanced Anaesthesia Specialists/Darvall, Gladesville, NSW, Australia). Micewereeuthanized at day 5 post infection using CO2, and tissues were homogenized using four ceramic beads at 6000 rpm twice for 15 s (Precellys 24 Homogenizer, Bertin Instruments, Montigny-le-Bretonneux, France). After centrifugation for 10 min, 9400 × g at 4 °C, virus titers in supernatants weredetermined by CCID50 assays using Vero E6 cells.
Histology and scoring of mouse lung sections
Lungs pan class="Chemical">wn>an class="Gene">ere fixed in 10% formalin, embedded in paraffin, and sections stainedwith H&E (Sigma‐Aldrich, Darmstadt, Germany). Slides were scanned using Aperio AT Turbo (Aperio, Vista, CA, USA) and images extracted using Aperio ImageScope software v12.3.2.8013 (Leica Biosystems, Wetzlar, Germany). Automatic quantitation of nuclei count andwhitespace of H&E stained sectionswas undertaken using QuPath v0.2.3[68].
RRV ELISA-based MXRA8-Fc binding assays
pan class="Chemical">Purifin>an class="Gene">edRRV (50 µL, 10 μg/mL) was coated onto maxi-sorb ELISA plates (Thermo Fisher) overnight in PBS, pH 7.4. Plates were blockedwith PBS supplementedwith 1X KPL (SeraCare) for 1 h at room temperature. HumanMXRA8-Fc (residues 20-337, UniProtKB: Q9BRK3) and negativehumanPD1-Fc controls were serially diluted in PBS supplementedwith 1X KPL, 50 μL/well was added for 1 h at 37 °C. Plates werewashed four times with PBS containing 0.05% Tween-20. Plates were then incubatedwith 50 μL/well horseradishperoxide conjugated goat anti-human IgG (H+L) (Goat anti-Human IgG Secondary Antibody, HRP (Invitrogen, Thermo Fisher, Cat number: A18829, RRID: AB_2535606)) at 5 μg/mL in PBS supplementedwith 1X KPL for 1 h at 37 °C. After fivewashes with PBS containing 0.05% Tween-20, plates weredevelopedwith TMB (Life Technologies) and 2 M H2SO4. Absorbancewas read at 450 nm.
MNV western blot
Cpan class="Gene">ell lysatn>an class="Gene">es from Raw264.7 cells infectedwith CPER-derived or WT MNVs or Mock were harvested in KALB lysis buffer (150 mM NaCl, 50 mM Tris-HCl, pH 7.5, 1% (v/v) Triton X-100, 1 mM EDTA) supplementedwith protease inhibitor cocktail III (Austral Scientific). Cell lysates from Raw264.7 cells infectedwith CPER-derived or WT MNVs or Mock were harvested in KALB lysis buffer (150 mM NaCl, 50 mM Tris-HCl, pH 7.5, 1% (v/v) Triton X-100, 1 mM EDTA) supplementedwith protease inhibitor cocktail III (Austral Scientific). Samples were separated on a polyacrylamide gel and transferred to a polyvinylidene difluoride membrane. Membranes were blockedwith 5% BSA/TBS-T for 2 h at room temperature prior to probing with primary antibodies overnight at 4 °C. Primary antibodies, anti-MNV1 capsid antibody 5C4.10 (Cat: MABF2097 Merck Millipore) or anti-actin antibody (Cat: A2066 Sigma), wereeach diluted 1:1000 in 5% BSA/TBS-T. The following day, the membranewas washed three times with TBS-T and incubatedwith either anti-mouse HRP secondary antibody (Cat: G-21040, Thermo Fisher Scientific) or anti-rabbit HRP secondary antibody (Cat: A16035, Thermo Fisher Scientific). Secondary antibodies were prepared by 1:10,000 dilution in TBS-T. Probed membranes weredevelopedwith the Amersham ECL Western Blotting Detection Reagent and imaged on the GE Healthcare Life Sciences AI600 Imager.
HuNoV RT-qPCR
Viral RNA was n>an class="Gene">extracted from the supernatant of HuNoV CPER-transfectedNIH3T3 cells using the AccuPrep Viral RNA extraction Kit (Bioneer). cDNA was generated by reverse transcription using Sensifast RT (Bioline) at 25 °C for 10 min, 42 °C for 15 min, and 85 °C for 5 min and quantified using iTaq Universal SYBR Green Supermix (Bio-Rad) under the following cycling conditions: 50 °C for 8 min, 95 °C for 2 min, 40 cycles of 15 s at 95 °C, and 1 min of annealing/extension at 60 °C, followed by a final extension for 10 min. To quantify Viral RNA was extracted from the supernatant of HuNoV CPER-transfectedNIH3T3 cells using the AccuPrep Viral RNA extraction Kit (Bioneer). cDNA was generated by reverse transcription using Sensifast RT (Bioline) at 25 °C for 10 min, 42 °C for 15 min, and 85 °C for 5 min. cDNA was quantified using iTaq Universal SYBR Green Supermix (Bio-Rad) under the following cycling conditions: 50 °C for 8 min, 95 °C for 2 min, 40 cycles of 15 s at 95 °C, and 1 min of annealing/extension at 60 °C, followed by a final extension for 10 min. To quantify HuNoV gRNA copies a previously described protocol69 was used. In brief, a pGEM-T-easy plasmid containing theNS/VP1 overlap region (nucleotides 3521–6916) of HuNoV (GenBank ID: GU445325) gifted by Peter White (University of New South Wales, Australia) was used to generate in vitro transcribed RNA. The RNA was quantified using the Qubit RNA BR assay (Q10211 Thermo Fisher Scientific) and volume adjusted to 1010 genome copies/μL. A tenfold serial dilution was undertaken to generate a standard curvewhich was reverse transcribed along with CPER-derived viral RNA and qPCR performed using the NK2PF/NK2PR primer pair[69].
CASV IFA
Trapan class="Disease">nsfn>an class="Gene">ected cell monolayers were removed by adding trypsin and then seeded onto glass coverslips at a density of 1 × 105 with appropriate supplementation of media and FBS. Once adherent, the supernatant was removed, and the cell monolayers were fixed in 100% ice-coldacetone. Coverslips were blockedwith a blocking buffer (0.05 M Tris/HCl (pH 8.0), 1 mM EDTA, 0.15 M NaCl, 0.05% (v/v) Tween-20, 0.2% w/v casein) for 1 h at room temperature. Coverslips were then incubated for 1 h with the mAb 9D7 (anti-CASV) or naivemouse serum (1:1000 in blocking buffer). Following two washes with PBST (PBSwith 0.05% Tween-20), coverslips were incubated for 1 h with a goat anti-mouse IgG H+L AlexaFluor 488 (Invitrogen, A-11001) secondary antibody according to the manufacturer’s protocols. Immediately after removing the secondary antibody, Hoechst 33342 nuclear stain (Invitrogen, 62249) was used at 1:1000 in thedark for 10 min to stain nuclei. Lastly, thecoverslips were subjected to threewashes with PBST and then mounted onto glass microscope slides using ProLong Gold Anti-fade (Invitrogen, P36934). All coverslips were imagedwith the ZEISS LSM 510 META confocal microscope.
CASV TCID50 assay
For pan class="Species">CASV growth kinetics cell monolayers were inoculated in triplicatewith 500 µL of virus (growth medium for the negative control) at an MOI = 0.1. After incubating at room temperaturewith agitation for 1 h, the inoculum was removed, and the cells werewashed three times with PBS before being replenishedwith 1 mL of 2% FBS in RPMI 1640. Viral supernatant was harvested at 16, 36, and 48 h post infection, and virus titers weredetermined using TCID50 assay on C6/36 cells. For the TCID50 assay, supernatant from each time point (16, 36, and 48 h) was titrated onto C6/36 cells at tenfold serial dilutions starting at 10−1 andending at 10−16, with five replicates at each dilution increment. After 48 h, cells were fixed in 20% acetone, 0.02% bovineserum albumin (BSA) in PBS overnight. Titers were confirmed by ELISA with mAb 9D7 (murine anti-CASVspike, School of Chemistry and Molecular Biosciences, UQ)[70].
Statistical analysis
For gropan class="Chemical">wth kinn>an class="Gene">etics, a two-way ANOVA test with Tukey’s multiple comparisons test to compare treatments at each time point was used. For comparisons between nasal and lung titers of SARS-CoV-2 in mice unpaired t-test with Welch’s correction was used. All data were analyzed using GraphPadPrism software (v9.0.0). The level of statistical significancewas set at 95% (p ≤ 0.05).
Authors: Roy A Hall; Debra J Nisbet; Kim B Pham; Alyssa T Pyke; Greg A Smith; Alexander A Khromykh Journal: Proc Natl Acad Sci U S A Date: 2003-08-13 Impact factor: 11.205
Authors: Nidia M M Oliveira; Annette K Broom; John S Mackenzie; David W Smith; Michael D A Lindsay; Brian H Kay; Roy A Hall Journal: Clin Vaccine Immunol Date: 2006-07
Authors: James T Robinson; Helga Thorvaldsdóttir; Aaron M Wenger; Ahmet Zehir; Jill P Mesirov Journal: Cancer Res Date: 2017-11-01 Impact factor: 12.701
Authors: Xuping Xie; Kumari G Lokugamage; Xianwen Zhang; Michelle N Vu; Antonio E Muruato; Vineet D Menachery; Pei-Yong Shi Journal: Nat Protoc Date: 2021-01-29 Impact factor: 13.491
Authors: Chris Lauber; John Ziebuhr; Sandra Junglen; Christian Drosten; Florian Zirkel; Phan Thi Nga; Kouichi Morita; Eric J Snijder; Alexander E Gorbalenya Journal: Arch Virol Date: 2012-04-24 Impact factor: 2.574
Authors: Yixuan J Hou; Kenichi Okuda; Caitlin E Edwards; David R Martinez; Takanori Asakura; Kenneth H Dinnon; Takafumi Kato; Rhianna E Lee; Boyd L Yount; Teresa M Mascenik; Gang Chen; Kenneth N Olivier; Andrew Ghio; Longping V Tse; Sarah R Leist; Lisa E Gralinski; Alexandra Schäfer; Hong Dang; Rodney Gilmore; Satoko Nakano; Ling Sun; M Leslie Fulcher; Alessandra Livraghi-Butrico; Nathan I Nicely; Mark Cameron; Cheryl Cameron; David J Kelvin; Aravinda de Silva; David M Margolis; Alena Markmann; Luther Bartelt; Ross Zumwalt; Fernando J Martinez; Steven P Salvatore; Alain Borczuk; Purushothama R Tata; Vishwaraj Sontake; Adam Kimple; Ilona Jaspers; Wanda K O'Neal; Scott H Randell; Richard C Boucher; Ralph S Baric Journal: Cell Date: 2020-05-27 Impact factor: 41.582
Authors: Amir Saberi; Anastasia A Gulyaeva; John L Brubacher; Phillip A Newmark; Alexander E Gorbalenya Journal: PLoS Pathog Date: 2018-11-01 Impact factor: 6.823
Authors: Scott E Guimond; Courtney J Mycroft-West; Neha S Gandhi; Julia A Tree; Thuy T Le; C Mirella Spalluto; Maria V Humbert; Karen R Buttigieg; Naomi Coombes; Michael J Elmore; Matthew Wand; Kristina Nyström; Joanna Said; Yin Xiang Setoh; Alberto A Amarilla; Naphak Modhiran; Julian D J Sng; Mohit Chhabra; Paul R Young; Daniel J Rawle; Marcelo A Lima; Edwin A Yates; Richard Karlsson; Rebecca L Miller; Yen-Hsi Chen; Ieva Bagdonaite; Zhang Yang; James Stewart; Dung Nguyen; Stephen Laidlaw; Edward Hammond; Keith Dredge; Tom M A Wilkinson; Daniel Watterson; Alexander A Khromykh; Andreas Suhrbier; Miles W Carroll; Edward Trybala; Tomas Bergström; Vito Ferro; Mark A Skidmore; Jeremy E Turnbull Journal: ACS Cent Sci Date: 2022-03-29 Impact factor: 18.728
Authors: Seth T Kazmer; Gunter Hartel; Harley Robinson; Renee S Richards; Kexin Yan; Sebastiaan J van Hal; Raymond Chan; Andrew Hind; David Bradley; Fabian Zieschang; Daniel J Rawle; Thuy T Le; David W Reid; Andreas Suhrbier; Michelle M Hill Journal: Biomedicines Date: 2022-02-01
Authors: Farooq Rashid; Zhixun Xie; Muhammad Suleman; Abdullah Shah; Suliman Khan; Sisi Luo Journal: Front Immunol Date: 2022-08-08 Impact factor: 8.786