From May to November 2021, we collected 15 nasal swab samples and six saliva samples from 16 pediatric patients who exhibited clinical symptoms and had high SARS-CoV-2 viral titers as detected in our previous study (Kim et al., 2024) on the development of a new diagnostic method using the Proteinase K method for SARS-CoV-2 detection. The samples were collected at Inha University Hospital (Incheon, Korea) under protocol number 2021-07-008. These collected samples (Table 1), which were stored at −80 °C, were sent to the College of Medicine at Yonsei University. The handling and processing of these samples were conducted at Biosafety Level 3 (BSL3) in the Avison Biomedical Research Center, with the approval of the Institutional Biosafety Committee (IBC 2022-0320).

Attempts were made to isolate SARS-CoV-2 from pediatric samples using Vero cell lines (ATCC No. CCL-81). An overnight monolayer of Vero cells (80% confluence) was washed twice with 1 × phosphate-buffered saline (PBS) before inoculating the nasal and saliva samples, which were filtered through a 0.2 μm filter, at a diluted to 5% (w/v) in Dulbecco’s Modified Eagle Medium (DMEM) containing antibiotics prior to inoculation. After 1 h of adsorption at 37 °C with 5% CO₂, a maintenance medium was added. This medium consisted of Dulbecco’s Modified Eagle Medium (DMEM) supplemented with trypsin (10 μg/mL), yeast extract (0.04%), tryptose phosphate broth (0.6%), and Antibiotic-Antimycotic 100 × (4 μL/mL; Gibco, Thermo Fisher Scientific, Grand Island, NY, United States) at a ratio of 1:10. The inoculated cells were cultured for 3–4 days at 37 °C in a 5% CO₂ atmosphere and were blindly passaged five times. Among the pediatric nasal swab and saliva samples of SARS-CoV-2, one strain, named N15, was successfully isolated in Vero cells. All fifth blind passage samples were titrated for SARS-CoV-2 using digital RT-PCR (dRT-PCR).

Total RNA was extracted from the fifth passage using the RNeasy Mini Kit (Qiagen, Valencia, CA, United States), following the manufacturer’s instructions. The reaction mixture of 30 μl included 5 μl of primer–probe mixture (Supplementary Table 1; 20 pmol forward primer, 20 pmol reverse primer, and 10 pmol FAM-BHQ1 probe per reaction), 15 μl of 2X Dr. PCR OneStep dRT-PCR Mixture (Optolane, Lot 4M009B-2201), 5 μl of total RNA, and nuclease-free water up to 30 μl. The reaction mixtures were loaded into wells of LOAA Dr. Digital PCR cartridges (Optolane, Seongnam-si, Korea). The cartridges were placed into the POSTMAN equipment (Optolane, Seongnam-si, Korea) for uniform application, and then mounted on the LOAA equipment (Optolane, Seongnam-si, Korea). The thermal profile began with an incubation at 50 °C for 30 min to convert RNA into cDNA, followed by denaturation at 95 °C for 15 min. The subsequent cycling conditions for digital PCR included 40 cycles of 95 °C for 10 s and 60 °C for 15 s. Each sample produced 16,800 to 19,200 valid wells. Digital PCR results were analyzed using the “Optolane OnPoint Pro” software (Optolane, Korea).

The passage five of the N15 strain was inoculated into fresh Vero cells at a multiplicity of infection (MOI). After an absorption period of 1 h at 37 °C, the cells were washed twice with 1 × PBS and then cultured in the maintenance medium for 24 h (subsection 2.2). Following the 24 h incubation, the Vero cells were fixed with 4% formaldehyde, permeabilized with 0.1% Triton X-100 in PBS, and blocked with 5% bovine serum albumin in PBS for 1 h. Primary rabbit monoclonal antibodies against SARS-CoV-2 nucleocapsid protein (Abcam, cat. no. ab271180) were added to the cells and incubated at room temperature for 1 h. After washing, a fluorescently conjugated secondary antibody, Alexa Fluor 647-labeled goat anti-rabbit IgG (red), was added and incubated at room temperature for 2 h. The cell nuclei were stained with one drop of VECTASHIELD Antifade Mounting Medium with DAPI (Vector Laboratories, Burlingame, CA, United States) and allowed to dry. The images were analyzed using a FluoView FV1000 confocal microscope (Olympus Life Science, Tokyo, Japan).

Viral supernatant from passage level 5 was inoculated at a 0.1 MOI into confluent Vero cells in a 25T flask and incubated for 72 h. To observe the virus using a transmission electron microscope (TEM), the viral-inoculated Vero cells from the 25T flask were harvested at 2,500 rpm at 4 °C for 10 min. Subsequently, the cells were washed and dehydrated using a graded ethanol series (50%, 60%, 70%, 80%, 90%, 100%). The dehydrated cells were infiltrated with an Embed-812 embedding kit (Electron Microscopy Sciences, United States). The embedding was polymerized in an oven at 60 °C for 48 h. Finally, the samples were negatively stained and examined using an HT7800 transmission electron microscope (Hitachi High-Tech Corporation, Japan).

The supernatants of passage 5 and 8 of the N15 strain were harvested and subjected to 10-fold serial dilutions (10^–2 to 10^–6). Each dilution (0.25 mL) was then added in duplicate to wells of confluent Vero cells. After 1 h absorption period, the inoculum was discarded, and each well was rinsed with PBS. The cells were then overlaid with 1% low melting temperature agarose containing DMEM, 2% fetal bovine serum, and 1% penicillin/streptomycin. After 48 h of incubation at 37 °C, the plaques were stained using a Neutral Red Solution (0.33% w/v in DPBS; Sigma, Cat. No. N2889). After an additional 24 h of incubation, the number of plaques (N) was counted. The following formula was applied to determine the titer of viral stock: PFU/mL = N/(D × V), where D is the dilution factor, and V is the volume of diluted virus added per well.

The replication kinetics of the N15 strain were assessed and compared with those of two other SARS-CoV-2 lineages obtained from the Korean National Culture Collection for Pathogens: B.1 (NCCP-43344) and B.1.1.529 (NCCP-43408). The viruses were inoculated onto Vero cell monolayers at a multiplicity of infection (MOI) of 0.1. Virus adsorption was carried out for 1 h at 37 °C. Following adsorption, the cells were washed twice with phosphate-buffered saline (PBS) to remove unbound virus particles and then incubated in maintenance medium for up to 120 h. The maintenance medium consisted of Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 0.3% tryptose phosphate broth, 0.02% yeast extract, and 1 μg/mL trypsin. Supernatants were collected at 1-, 24-, 48-, 72-, 96-, and 120 h post-inoculation (hpi) to monitor viral replication. Viral RNA was extracted and quantified using dRT-PCR targeting the nucleocapsid (N) gene (Supplementary Table 1). All samples were analyzed in triplicate (see section “2.3 Digital RT-PCR based for SARS-CoV-2 confirmation”).

Total RNA of the N15 strain was extracted using the RNeasy mini kit (Qiagen, Valencia, CA, United States) and subsequently sent for next-generation sequencing at Macrogen (Korea). The library was constructed using the TruSeq Stranded Total RNA Library Prep Gold Kit, and sequencing was performed on the Illumina NovaSeq 6000 platform. A total of 49,285,634 reads were produced for the N15 sample, with a GC content of 39.8% and a Q30 score of 94.0%. For bioinformatic analysis, the reads were processed using the Snakelines (v.1.18) framework (Goga et al., 2021). The complete genomes of the N15 strain have been deposited into the NCBI GenBank with accession number PP915627. Genomic annotation of the N15 strain was performed using GATU (Tcherepanov et al., 2006) with the reference genome of the Wuhan-Hu-1 isolate (NC_045512.2).

For genetic classification, various levels of analysis were applied in this study. The first was based on genome sequences using Nextclade (Aksamentov et al., 2021) and Pangolin tool (O’toole et al., 2021). The second one based on the spike sequence using Hedgehog tool (O’toole et al., 2022). To determine the fluctuation over time of different lineages of SARS-CoV-2 in South Korea, complete genomic sequences of South Korea strains with collection dates (n = 192,501) were retrieved from GISAID (Supplementary Table 2). Spike sequences were extracted using the tool available at https://mafft.cbrc.jp/alignment/server/specificregion-last.html (last accessed September 2024). All 192,501 Korean spike sequences were then classified using the Hedgehog tool, resulting in 5,687 sequences belonging to lineage A_1 and 247 sequences belonging to lineage A_2 (Supplementary Table 2). The chronological frequency plot of lineages A_1 and A_2 of SARS-CoV-2 was generated using Microsoft Excel.

The isolated N15 strain was subjected to an antiviral assay against three drugs known to inhibit coronaviruses (Tiseo et al., 2023): Remdesivir (Cat. no. HY-104077), Molnupiravir (Cat. no. HY-135853), and Nirmatrelvir (Cat. no. HY-138687), all of which were purchased from MedChemExpress.¹ Vero and Calu-3 cells (6.25 × 10⁴ cells) were plated in a 48-well plate, cultured overnight, and pre-treated for 2 h with serial 2-fold dilutions of each drug at varying concentrations in triplicate experiments. The concentration range (μM) for Remdesivir were 10–0.156, Molnupiravir were 100–0.781, and Nirmatrelvir were 10–0.078. The cells were then inoculated with passage 5 of the N15 strain at a MOI of 0.01 for 1 h. After 1 h of inoculation, the virus was removed, the cells were washed once with PBS, and fresh medium containing dilutions of each drug was added to the cells. At 24 and 48 hpi, culture supernatants were harvested, and viral RNA was extracted for quantification of viral genomic copies by real-time RT-PCR. The half-maximal inhibitory concentration (IC₅₀) values were calculated from the normalized activity dataset-fitted curves using GraphPad Prism v.10.2.3.

The N15 strain of SARS-CoV-2 was successfully isolated in Vero cells after five blind passages. In Vero cells, cytopathic effects (CPE) were observed within 48 hpi, characterized by the cells beginning to detach from the monolayer and cluster into grape-like aggregates (blue arrows, Figure 1B). Besides of CPE, the replication of virus was confirmed through IFA and TEM. The IFA results demonstrated that the N15 strain at passage 5 exhibited strong staining for the nucleocapsid protein, which localized in the cytoplasm of Vero cells (bright red signal, inserted Figure 1B). These findings suggested that Vero cells had been productively infected by the N15 strain. The TEM results, shown in Figure 1C, reveal clusters of viral particles observed within membrane-bound vesicles in the cytoplasm. Figure 1D clearly illustrates the characteristic spherical shape and surface features of individual viral particles, measuring approximately 75 nm in size.

The N15 strain from both passage 5 and passage 8 produced large, light pink- colored circles on a red monolayer of cells (arrows, Figures 2A, B). The plaque assays revealed a significant increase in viral titers between passage 5 and passage 8. Specifically, the mean PFU/mL for passage 5 was 2.1 × 10⁵, whereas for passage 8, the mean PFU/mL increased to 1.7 × 10⁶. This indicates a considerable enhancement in viral replication or infectivity across the passages. The viral load in Vero cells increased significantly from passage 5 (mean 273,319 copies/ml) to passage 8 (mean 2,350,706 copies/ml) (Figure 2C). The higher expression levels of the nucleocapsid gene at passage 8 align with the increased viral titers observed in the plaque assay results.

Because SARS-CoV-2 does not induce CPE in certain cell lines, the replication of three viral strains (B.1, N15, and B.1.1.529) was assessed in Vero cells by quantifying nucleocapsid gene expression using dRT-PCR. As shown in Figure 3, all three strains exhibited a marked increase in viral RNA levels over time. Viral loads increased rapidly from 1 to 48 h post-infection (hpi), reaching approximately 6 log₁₀ copies/mL. After 48 hpi, the viral RNA levels plateaued, with only minor differences observed among the strains. Notably, all three strains reached similar nucleocapsid gene expression levels by 72 h post-infection (hpi) and remained stable up to 120 hpi. These results indicate that all tested strains replicate efficiently in Vero cells with comparable kinetics.

The genome of the N15 strain contained 29,893 nucleotides and exhibited a high similarity of 99.9% to the canonical SARS-CoV-2 reference genome (NC_045512.2). According to this reference, the N15 strain was predicted to encode 12 open reading frames (ORFs). Eight non-synonymous substitutions were observed in five out of 12 ORFs (Table 2). Among these mutations of N709S in the spike protein and T11M in the E protein were considered rare (based on information from CoVsurver²). Of the 23 sites in the nsp5 and 11 sites in the nsp12 protein associated with antiviral drug resistance (European Centre for Disease Prevention and Control, 2023; Flynn et al., 2023), no mutations were observed in the N15 strain, as well as in the other Korean strains of Pango lineage A found in different years from 2020 to 2023 (not shown).

Based on genome classification, the N15 strain belongs to Pango lineage A and Nextstrain clade 19B (Supplementary Figure 1). Based on the spike protein sequence, the N15 was classified as A_2 lineage. Unlike most globally circulating SARS-CoV-2 variants that possess the D614G mutation (O’toole et al., 2022), the N15 spike retains the ancestral D (aspartic acid) at position 614. The absence of the D614G mutation further supports its classification within the early A_2 lineage. In the epidemiological context, the unique situation of lineage A in South Korea is illustrated in the accompanying Figure 4. From January 2020 to July 2023, it was observed that lineage A predominantly circulated during the initial phase of the pandemic. After July 2021, sequences of lineage A were detected only sporadically; however, this lineage continued to persist at least until July 2023. The N15 strain was collected in September 2021, following a decline in the epidemic wave of lineage A.

In Figure 5, the antiviral efficacy of three drugs (Remdesivir, Molnupiravir, and Nirmatrelvir) were assessed against the N15 strain in Vero and Calu-3 cells. The inhibition percentage was evaluated at two different time points: 24 and 48 hpi. Remdesivir showed potent antiviral activity against the N15 strain in both Vero and Calu-3 cells with lower IC₅₀ values compared to Molnupiravir and Nirmatrelvir. Molnupiravir demonstrated moderate activity, while Nirmatrelvir exhibited strong inhibition but with varying IC₅₀ values across different cell lines and time points. These results suggest that Remdesivir may be the most effective among the tested compounds, especially in Vero cells, followed by Nirmatrelvir and Molnupiravir.