An optimized circular polymerase extension reaction-based method for functional analysis of SARS-CoV-2

The CPER method builds principally on overlap extension PCR that fuses several double-stranded DNA (dsDNA) fragments containing 20–50-bp homologous ends into one large fragment [27]. Compared to the traditional overlap extension PCR, which uses a set of two distal primers to facilitate the generation of the combined fragment, CPER does not amplify fragments using such primers but instead utilizes an additional fragment that overlaps with the first and the last fragment to be joined, thereby circularizing the self-primed and extended dsDNA product. In CPER-based bacterial cloning, this additional fragment is typically a linearized plasmid vector generated by restriction digestion or PCR. As a result, the CPER product resembles a relaxed circular plasmid with staggered nicks which locate to the 5′ end of each strand of the individual fragment following the respective ‘round-the-horn’ amplification, as commonly seen in the QuikChange^® approach of site-directed mutagenesis [28].

Adaptation of the CPER approach to de novo assembly of infectious clones for positive-strand RNA viruses is primarily achieved by substituting a linker fragment for the linearized vector used in CPER-mediated plasmid cloning. The design of the linker fragment draws inspiration from plasmid-launched mRNA synthesis driven by the mammalian RNA polymerase II (Pol II) promoter, as the genomes of several positive-strand RNA viruses including flaviviruses and coronaviruses contain a 5′ cap structure like cellular mRNAs and undergo cap-dependent translation [29]. In addition to the Pol II promoter, the linker fragment also contains a polyadenylation signal for transcription termination and, importantly, a self-cleaving ribozyme sequence in front of the poly(A) signal to ensure accurate processing of the 3′ end of the RNA transcript to match the authentic viral genome sequence [30, 31]. Notably, while the linker fragment is usually cloned into a plasmid for long-term maintenance in E. coli, only the portion containing the mammalian transcription elements, but not the bacterial propagation cassettes, is amplified and used in CPER assembly.

Despite the successful adaptation of the CPER technology for the generation of infectious clones, CPER has a major intrinsic limitation, which is the presence of staggered nicks that impede efficient expression in mammalian cells. Whereas nicked plasmids are known to be seamlessly repaired upon transformation into E. coli, the precise fate of a circularized, nick-containing dsDNA inside a mammalian cell remains elusive. The presence of nicks in the template strand can cause Pol II pausing and likely also template misalignment, which may eventually lead to unwanted mutations [32]. In CPER-derived infectious clones, the circular template strand extended from each fragment contains a nick, which, depending on the genome segmentation scheme used for assembly, locates to different coding or noncoding regions of the viral genome. Although the sequence contexts in which the nicks situate may permit Pol II bypassing, how the template discontinuity affects the overall Pol II transcription efficiency, and hence the synthesis of full-length viral genomes, in mammalian cells remains unclear.

Another limitation of the current CPER approaches for SARS-CoV-2 rescue lies in the choice of cell lines for transfection of the CPER product [25, 26]. While HEK293-derived cell lines have been successfully used for reverse genetics systems for a variety of viruses from diverse families due to their robust transfectability, the use of HEK293 cells for SARS-CoV-2 rescue can be less efficient because of the unique cellular tropism of the virus and the critical host factors required for virus entry and replication. To date, three mammalian cell lines are commonly used for in vitro propagation of SARS-CoV-2 to high titers. These include Vero E6 (African green monkey kidney epithelial), Caco-2 (human colonic epithelial), and Calu-3 (human lung epithelial) cells. All three cell lines express the receptor for SARS-CoV-2, angiotensin-converting enzyme 2 (ACE2), while the latter two express also transmembrane serine protease 2 (TMPRSS2), a critical early entry cofactor [33]. In addition to priming direct cell membrane fusion, the presence of TMPRSS2 safeguards the integrity of the polybasic furin cleavage site in the viral spike gene, which is selectively deleted during serial passaging in Vero E6 cells due to viral host adaptation [34]. Therefore, Vero E6 cells stably expressing human TMPRSS2 (Vero E6-TMPRSS2) have been widely used for the propagation of ancestral and emerging SARS-CoV-2 strains including the variants of concern (VOCs). More importantly, given the nature that Vero cells lack interferon (IFN) production [35], it remains the first-line cell system for generating and propagating recombinant mutant viruses that are attenuated through selective ablation of viral gene functions that evade or antagonize IFN-mediated antiviral innate immunity (e.g. SARS-CoV-2 papain-like protease (PLpro) which is an IFN antagonist [36, 37]).

Taking these limitations into account, we rationally optimized CPER for SARS-CoV-2 by adding new steps to seal the nicks in the CPER product and by using a modified linker plasmid as well as a different cell line for transfection of the CPER product (Fig. 1A). Specifically, under the same genome segmentation scheme reported by Torii et al. [26], gel-purified viral cDNA fragments were phosphorylated at the 5′ end by using a T4 polynucleotide kinase. Equal molar amounts of the phosphorylated fragments were then subjected to CPER assembly using the cycling ‘condition 3’ as described previously [26]. Immediately before transfection, the nicks in the CPER product were sealed by using a high-fidelity and thermostable Taq DNA ligase that joins the extended 3′–OH terminus with its originating 5′–phosphorylated terminus, giving rise to a closed circular cDNA infectious clone. Then, the sealed CPER product was directly transfected into a monolayer of Vero E6-TMPRSS2 cells by using the TransIT-X2 dynamic delivery system (Fig. 1A). Furthermore, to ensure efficient Pol II termination and to prevent Pol II read-through in the linker region, which may confound ribozyme processing at the transcript 3′ end or interfere with new transcription initiation, we also replaced the ‘spacer’ sequence that is located between the poly(A) signal and CMV enhancer/promoter with a functional Pol II transcriptional pause signal from the human α2 globin gene known to minimize promoter crosstalk [38]. The resultant linker sequence was assembled with ampicillin resistance and origin of replication cassettes into a high-copy plasmid, named “pGL-CPERlinker” (Fig. 1B and Additional file 1: Figure S1).

Applying the optimized CPER workflow, infectious virus generated using as a template a BAC construct encoding a GFP reporter SARS-CoV-2 [19] could be rescued as early as day 3 post-transfection, as evidenced by the formation of GFP-positive syncytia (Fig. 1C). By day 5 post-transfection, massive cytopathic effects (CPE) could be observed, and the GFP signal declined afterwards due to significant cell death. In comparison, successful virus rescue using the ‘classical’ CPER approach was not observed until day 5 post-transfection (Fig. 1C), similar to previous reports [25, 26]. The peak titers of the passage 0 (P0) stocks obtained from either protocol were found to reach ~ 5 × 10⁵–10⁶ PFU/mL when harvested at full CPE, indicating that the two workflows differ primarily in the onset of productive viral replication launched by full-length viral genomes generated from the intracellularly delivered CPER products. Of note, we also used our new linker plasmid in the ‘classical’ CPER assembly, which therefore may not precisely represent the efficiency of virus recovery reported in the previous studies. Nonetheless, our results indicate that the optimized CPER workflow can accelerate SARS-CoV-2 rescue by at least 2 days.

We also applied the optimized CPER approach to rescue SARS-CoV-2 from purified viral genomic RNA [25]. Adopting again the 10-fragment scheme reported by Torii et al. [26], we successfully achieved specific amplification of all fragments from the first-strand cDNA that was synthesized from purified viral genomic RNAs of three different virus strains, including the ancestral strain WA1 and two VOCs (i.e. Beta and Omicron) (Fig. 2A). We also performed site-directed mutagenesis directly in the purified fragment #2 by overlap extension PCR using the pair of primers for fragment #2 amplification and a pair of mutagenesis primers, and could readily obtain the new mutant fragment #2 for all three viruses (Fig. 2B). Successful rescue of the WA1 and Beta viruses, as evidenced by CPE, was consistently observed between day 3 and day 4, and the passage 0 (P0) stocks were typically harvested on day 4 or day 5 when CPE was > 90%. The use of Vero E6-TMPRSS2 cells ensured the integrity of the furin cleavage site, as confirmed by sequencing of independently-rescued viruses (Fig. 2C). The CPER-derived recombinant viruses also displayed the same plaque morphology as their parental isolates (Fig. 2D), and the P0 virus titers consistently reached ~ 10⁶ PFU/mL in 4 independent rescue experiments (Fig. 2E).

SARS-CoV-2 genomic RNA extraction, cDNA synthesis, CPER transfection, and live virus experiments were all conducted in the BSL-3 facility of the Cleveland Clinic Florida Research and Innovation Center (CC-FRIC). Sterility-tested viral cDNA was handled in a BSL-2 laboratory following standard biosafety practices and procedures. All work was reviewed and approved by the CC-FRIC Institutional Biosafety Committee in accordance with the National Institutes of Health (NIH) Guidelines.

Safe handling of viral agents such as SARS-CoV-2 is of utmost importance. Work with SARS-CoV-2 including recombinant viruses engineered using CPER (or other reverse genetics) approaches requires adequate biosafety containment and is subject to institutional, local and/or federal regulations. Considering the ongoing debates about the dissemination of methods for reverse engineering of SARS-CoV-2 (see for example [42]), we consciously described in detail only the newly developed optimization steps of the CPER method, while mostly referring to published reports for the other steps of the CPER approach.

Vero E6 (#CRL-1586) and HEK293T (#CRL-3216) cells were purchased from the American Type Culture Collection (ATCC) and were maintained in Dulbecco’s modified Eagle’s medium (DMEM, Gibco) supplemented with 10% fetal bovine serum (FBS, Gibco), 2 mM L-Glutamine (Gibco), 1 mM sodium pyruvate (Gibco) and 100 U/mL of penicillin–streptomycin (Gibco). Vero E6 cells stably expressing human TMPRSS2 were generated by lentiviral transduction followed by selection with blasticidin (40 μg/mL; Invivogen) [37]. SARS-CoV-2 strains hCoV-19/USA-WA1/2020 (NR-52281), hCoV-19/USA/MD-HP01542/2021 (Lineage B.1.351; Beta variant) (NR-55282), and hCoV-19/USA/MD-HP20874/2021 (Lineage B.1.1.529; Omicron variant) (NR-56461) were obtained from BEI Resources, National Institute of Allergy and Infectious Diseases (NIAID), NIH.

Viral genomic RNA was purified from 280 μL virus-containing media (P0 pool upon harvest) using the QIAamp Viral RNA Mini Kit (Qiagen) as per the manufacturer’s instructions and eluted in 60 μL nuclease-free water. Reverse transcription for first-strand cDNA synthesis was performed by using the LunaScript RT SuperMix Kit (NEB) containing both oligo(dT) and random primers in a reaction consisting of 10 μL genomic RNA, 4 μL 5 × SuperMix and 6 μL nuclease-free water with the cycling condition as follows: 2 min at 25 °C, 20 min at 55 °C, and 1 min at 95 °C. One microliter of RNase H (5 U; Thermo Scientific) was subsequently added and the reaction mix was incubated at 37 °C for 20 min.

The bacterial artificial chromosome (BAC) encoding a GFP reporter SARS-CoV-2 in the background of hCoV-19/Germany/BY-pBSCoV2-K49/2020 (GISAID EPI_ISL_2732373) was kindly provided by Armin Ensser (Friedrich-Alexander University Erlangen-Nürnberg, Germany) and has been described previously [19]. pGL-CPERlinker was assembled from synthetic DNA oligonucleotides and fragments (IDT) as well as the ampicillin resistance cassette and the origin of replication derived from pUC19 (NEB).

To amplify the 10 viral cDNA fragments (either from BAC or the first-strand viral genomic cDNA), previously reported primer sets were used [26]. The primers for amplification of the linker fragment from pGL-CPERlinker are: GL-CPERlinkF (5′-CTTAGGAGAATGACAAAAAAAAAAAAAAAAAAAAAAAAAAAGGCCGGCATGGTCCCAGCC-3′) and GL-CPERlinkR (5′-GTTACCTGGGAAGGTATAAACCTTTAATACGGTTCACTAAACGAGCTCTGCTTATATAG-3′). Amplification of each fragment was carried out by using the PrimeSTAR Max DNA polymerase (Takara Bio) in a 50 μL PCR reaction containing 0.2 μM of each primer and 1 ng BAC or 2 μL viral cDNA as the template with the cycling condition as follows: 10 s at 98 °C; 35 cycles of 10 s at 98 °C, 5 s at 55 °C, 25 s at 72 °C; and 2 min at 72 °C. All PCR products were gel purified by using the Monarch DNA Gel Extraction Kit (NEB) and eluted in 20 μL nuclease-free water. The purified fragments were then 5′ phosphorylated in a 50 μL reaction containing 10 U of T4 polynucleotide kinase (NEB) and cleaned up through the Monarch PCR & DNA Cleanup spin columns (NEB). CPER assembly was performed as previously described by combining 0.05 pmol of each fragment in a 50 μL reaction containing 2.5 U PrimeSTAR GXL DNA polymerase (Takara Bio) and using the ‘condition 3’ cycling parameters [26]. Immediately before transfection, the CPER product was subjected to post-PCR nick sealing for 30 min at 50 °C and 30 min at 60 °C in a 25 μL reaction containing 1 mM β-nicotinamide adenine dinucleotide (NAD+) (NEB) and 0.5 μL HiFi Taq DNA ligase (NEB). The final CPER product was transfected into Vero E6-TMPRSS2 cells seeded into 6-well plates (~ 5 × 10⁵ cells per well) by using the TransIT-X2 Dynamic Delivery System (Mirus Bio) as per the manufacturer’s instructions. After 24 h, the culture media was replaced with DMEM containing 2% FBS, 2 mM L-Glutamine, 1 mM sodium pyruvate, 1 × non-essential amino acids (Gibco), 10 mM HEPES (Gibco), and 100 U/mL of penicillin–streptomycin. For the classical CPER method, the unsealed CPER product was first transfected into HEK293T cells (1.5 × 10⁶ cells seeded into 6-well plates) by using TransIT-LT1 (Mirus Bio), and the trypsinized cells were then overlaid onto Vero E6-TMPRSS2 cells at 6 h post-transfection, as previously described [25].

The titers of the P0 virus stocks were determined by plaque assay. Briefly, a monolayer-culture system of Vero E6-TMPRSS2 cells was incubated with ten-fold serially diluted virus-containing media. The inoculum was removed after 2 h, and the cell monolayers were washed twice with PBS and then overlaid with 1% colloidal microcrystalline cellulose (Sigma) in MEM containing 2% FBS, 2 mM L-Glutamine, 1 × non-essential amino acids, 10 mM HEPES, and 100 U/mL of penicillin–streptomycin. Plaques were visualized by Coomassie Blue staining on day 3. For P0 virus sequencing, viral genomic RNA was purified and the first-strand cDNA was synthesized as described above. Nine fragments encompassing the whole genome [43] were then amplified from the cDNA and subsequently subjected to Sanger (Azenta Life Sciences) or Nanopore sequencing (Plasmidsaurus).