Cell culture systems for isolation of SARS-CoV-2 clinical isolates and generation of recombinant virus

Summary

A simple and robust cell culture system is essential for generating authentic SARS-CoV-2 stocks for evaluation of viral pathogenicity, screening of antiviral compounds, and preparation of inactivated vaccines. Evidence suggests that Vero E6, a cell line commonly used in the field to grow SARS-CoV-2, does not support efficient propagation of new viral variants and triggers rapid cell culture adaptation of the virus. We generated a panel of 17 human cell lines overexpressing SARS-CoV-2 entry factors and tested their ability to support viral infection. Two cell lines, Caco-2/AT and HuH-6/AT, demonstrated exceptional susceptibility, yielding highly concentrated virus stocks. Notably, these cell lines were more sensitive than Vero E6 cells in recovering SARS-CoV-2 from clinical specimens. Further, Caco-2/AT cells provided a robust platform for producing genetically reliable recombinant SARS-CoV-2 through a reverse genetics system. These cellular models are a valuable tool for the study of SARS-CoV-2 and its continuously emerging variants.

Graphical abstract

Highlights

• •Human cell lines overexpressing ACE2 and TMPRSS2 were tested for SARS-CoV-2 infection
• •Two cell lines, Caco-2 and HuH-6, exhibited remarkable susceptibility to infection
• •Caco-2 and HuH-6 cells allowed isolation of SARS-CoV-2 from clinical specimens
• •Caco-2 cells supported production of genetically stable recombinant SARS-CoV-2

Methodology in biological sciences; Virology

Introduction

Since entering the human population in late 2019, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has infected more than 674 million people worldwide and claimed over 6.8 million lives.^1,2 The unprecedented speed and scale at which the virus has spread across the globe and its continuous circulation in most parts of the world, has led to the emergence of new viral variants, some of which seem to be more infectious than their ancestral counterparts.^3,4,5,6,7 As scientists worldwide scramble to understand the biology of these newly emerging variants and find ways to control their spread, it is imperative to develop cell culture systems that allow isolation and propagation of a broad variety of SARS-CoV-2 variants in a robust and reliable manner.

Since the beginning of SARS-CoV-2, Vero and its derivative cell lines, such as Vero E6 (a.k.a. Vero C1008), have been used as the workhorse of virologic studies.⁸ These cells were adopted for SARS-CoV-2 research due to their history with other coronaviruses, such as SARS-CoV⁹ and Middle East respiratory syndrome coronavirus (MERS-CoV).¹⁰ However, Vero cells have several limitations. First, they are of simian origin, diminishing their utility for studying human host responses. Second, Vero cells are inefficient in direct isolation of SARS-CoV-2 from clinical specimens^11,12 and generation of recombinant virus through reverse genetics systems.^13,14 Finally, serial passaging of SARS-CoV-2 in Vero cells introduces mutations and deletions in the viral genome,¹⁵ chief among them are mutations that disrupt the furin cleavage site present in the spike protein.^16,17,18,19 These genotypically altered viruses replicate better in Vero cells but demonstrate attenuated fitness in human respiratory cells and in vivo models of virus infection.^15,17,20,21 In addition, SARS-CoV-2 particles lacking the furin cleavage site are poorly neutralized by antibodies,²¹ inadequately inhibited by IFN-stimulated genes, such as IFITM2,²² and elicit differential inflammatory responses in mouse lungs.²¹ This limits the generalizability of in vitro and in vivo studies carried out using Vero-expanded viral stocks.²³

The reliance on Vero cells for SARS-CoV-2 propagation partly stems from the lack of well-characterized human cell culture systems for virus isolation. Although a handful of human cell lines, either wild-type or overexpressing the SARS-CoV-2 receptor ACE2, have been proposed as more suitable platforms for SARS-CoV-2 studies,^8,24,25,26 their ability to produce high-titer virus stocks has not been explored. Also, it is unknown if these cells allow viral propagation without cell culture adaptation. A recent study reported a relatively stable propagation of SARS-CoV-2 in the human lung Calu-3 cells.²⁷ However, these cells are difficult to work with, owing to their exceedingly slow proliferation rate (doubling time of around 3–4 days²⁸), which limits their use for routine preparation of large-scale virus stocks. Another human cell line, HT1299, engineered to express ACE2, was employed to isolate the SARS-CoV-2 Omicron variant from clinical samples.¹¹ Yet, its usage has not gained a widespread acceptance in the field.

To fill these deficiencies, we aimed to establish a facile and robust, human-derived cell culture system for isolation and propagation of authentic SARS-CoV-2 with minimal genomic drift. We engineered a panel of 17 human cell lines, derived from SARS-CoV-2-relevant organs (e.g., lung, heart, kidney, brain, intestine), with two SARS-CoV-2 entry factors, ACE2 and TMPRSS2. When infected with the prototype SARS-CoV-2 WA-1 isolate, two cell lines, Caco-2/AT and HuH-6/AT, produced highly concentrated virus stocks, comparable in titer to those produced in Vero E6 cells. However, compared to Vero E6 cells, these cell lines exhibited increased sensitivity in isolating SARS-CoV-2 clinical isolates from nasal swabs. Also, these cells were remarkably more sensitive than Vero E6 cells in rescuing recombinant SARS-CoV-2 (rSARS-CoV-2) from the transfected viral genome. Finally, these cells allowed production of genetically reliable viral stocks. The human cell lines reported in this study are a valuable resource for isolation, propagation, and investigation of SARS-CoV-2.

Results

Most human cell lines are resistant to SARS-CoV-2 infection

Seventeen human cell lines (Figure 1A) were infected with the Vero-expanded SARS-CoV-2 WA-1 isolate at a low multiplicity of infection (MOI) of 0.1 and monitored for their ability to support viral propagation. Immunofluorescence (IF) analysis using an antibody directed against the viral nucleocapsid (N) protein showed that 10 of the 17 cell lines were resistant to infection (Figure 1B). The remaining seven had 4-48% infected cells at 72 h post-infection (hpi), indicating some level of permissivity. Within this group, six cell lines (HK-2, Calu-3, HuH-6, HuH-7, Huh-7.5, and Caco-2) supported viral spread, as evidenced by a significant increase in the number of infected cells from 24 to 72 hpi. The replication of WA-1 was slower in all cell types compared to Vero E6 cells (Figure 1B). These data indicate that only a limited number of human cell lines support SARS-CoV-2 infection and that this phenotype varies in ways that are unrelated to tissues of origin.

Figure 1.

Most human cell lines supported no or minimal SARS-CoV-2 infection

(A) The cell lines tested and their tissues of origin. The detailed description of each cell line is provided in the table.

(B) The indicated cells were infected with SARS-CoV-2 at an MOI of 0.1 and stained with the viral nucleocapsid (N) protein (red) at 24 and 72 hpi. The nuclei were counterstained with DAPI. The mean percentage of positive cells ±standard deviation of three biological replicates is shown. The IF images were captured with a 10× objective lens. The cell names are colored differently according to their tissues of origin. The experiment was repeated twice.

(C) Total abundance of ACE2 in cells and its cell surface-associated fraction was measured by Western blot and flow cytometry, respectively. The shades of blue indicate the intensity of the ACE2 signal, whereas the shades of red reflect the SARS-CoV-2 infection efficiency in each cell line. The data of Western blot and flow cytometry experiments are presented in Figure S1.

To examine whether susceptibility to SARS-CoV-2 infection was explained by the presence of the viral entry receptor ACE2, we determined ACE2 expression in all cell lines by Western blot (Figure S1A) and flow cytometry (Figure S1B). SARS-CoV-2 susceptible cell lines HK-2, Calu-3, HepG2, Huh-7.5, and Caco-2 showed detectable levels of ACE2 in lysates and on the cell surface, although these levels were substantially lower than those observed in the highly susceptible Vero E6 cells. Despite efforts, we failed to detect the expression of TMPRSS2, another SARS-CoV-2 entry factor, in any of the cell lines using commercially available antibodies (data not shown). In all, our results indicate that the resistance of cell lines to SARS-CoV-2 infection can largely be attributed to the lack or lower expression of viral entry factors (Figure 1C).

ACE2 and TMPRSS2 expression allows efficient SARS-CoV-2 infection of human cells

We next investigated whether exogenous expression of ACE2 and/or TMPRSS2 in human cells improves their susceptibility to SARS-CoV-2 infection. For this, we generated stable cell lines expressing human ACE2, TMPRSS2, or their combination (cells expressing the combination of human ACE2 and TMPRSS2 are hereinafter referred to as AT cells). Western blot and flow cytometric analysis of select AT cell lines confirmed the expression of ACE2 in cell lysates and on the cell surface, respectively (Figure S2). Unfortunately, none of the antibodies tested could detect the expression of TMPRSS2, either by Western blot or flow cytometry (data not shown).

First, we assessed the ability of the stable cell lines to support entry of SARS-CoV-2 pseudoparticles. Expression of ACE2 allowed varying levels of infection in different cell types. 293T, Calu-6, HuH-6, and Caco-2 cells, which displayed low infection rates in the absence of ACE2 (<8% infection), achieved high infection levels of 55%, 52%, 75%, and 93%, respectively, following ACE2 expression (Figures 2A and S3). Huh-7.5 and AC-16 cells were moderately affected, yielding 28%, 27%, and 25% infection, respectively. ACE2 expression had minimal effect on other cell types, reflected by infection efficiency of less than 10%. Overexpressing TMPRSS2 in the absence of ACE2 had mostly no or subtle effect on viral entry, with the exception of HuH-7 cells, where TMPRSS2 expression led to around 60% infected cells compared to 18% for wild-type cells.

Figure 2.

Expression of ACE2 and TMPRSS2 synergistically enhanced SARS-CoV-2 infection of multiple cell lines

(A) The cells were infected with SARS-CoV-2 pseudoparticles containing a GFP reporter for 24h, and the infection efficiency was monitored by flow cytometry. The mean ± standard deviation of triplicate samples is plotted. ND, not determined. The experiment was repeated three times. The data used to generate the heatmap are presented as a bar graph in Figure S3.

(B) Cells were infected for 24h with SARS-CoV-2 at an MOI of 0.01 followed by IF. Red, viral N-protein; Blue, cell nuclei. The IF images were captured with a 10× objective lens. The mean percentage of positive cells ±standard deviation of triplicate samples is shown. The experiment was repeated twice.

When ACE2 and TMPRSS2 were co-expressed, most of the cell lines became highly susceptible to infection. Eight cell lines, 293T/AT, HK-2/AT, A549/AT, Calu-6/AT, HuH-6/AT, HuH-7/AT, Huh-7.5/AT, and Caco-2/AT, had infection efficiencies either equal to or higher than that of Vero E6/AT cells. Some cell lines, such as MRC5, HT29, SK-N-SH, and RD, remained refractory to infection, even after ACE2/TMPRSS2 expression.

Since the pseudoparticle assay only allows single-cycle infection and does not provide information about post-entry steps of the virus life cycle, we confirmed our results with authentic SARS-CoV-2. The cells were infected with the SARS-CoV-2 WA-1 isolate at a low MOI of 0.01 to account for both viral replication and spread, and the infection was assessed by IF at 24 hpi. Under these conditions, no infection was observed in wild-type cells (Figure 2B). Consistent with the pseudoparticle assay, ACE2 overexpression alone led to increased infection in several cell lines. HuH-6 and Caco-2 cells exhibited the highest infection rates of 87% and 61%, respectively (Figure 2B). Some other cell lines, such as MRC5, HuH-7, Huh-7.5, AC-16, SK-N-SH, and RD, achieved infection rates of around 10–30%. When ACE2 and TMPRSS2 were expressed together, nine of the 17 human cell lines became highly susceptible to SARS-CoV-2 infection, with over 70% cells positive for the viral N protein. HK-2/AT, HuH-6/AT, HuH-7/AT, Huh-7.5/AT, Caco-2/AT, and AC-16/AT had over 90% infected cells, numbers comparable to Vero E6 cells (92% infected cells). Interestingly, in congruence with the pseudoparticle assay, HT29 and HUVEC cells remained resistant to infection even after ACE2/TMPRSS2 co-expression (Figure 2B). Overall, these results indicate that exogenous expression of ACE2 and TMPRSS2 renders most of the human cell lines susceptible to SARS-CoV-2 infection.

Time course analysis of SARS-CoV-2 replication in human cell lines

Next, we compared SARS-CoV-2 replication kinetics in human cells lines versus Vero E6 cells. For this, we chose five human cell lines, A549/AT, Caco-2/AT, HuH-6/AT, HuH-7/AT, and Huh-7.5/AT, which demonstrated high infection efficiencies in the experiment above (Figure 2B). HK-2/AT and AC-16/AT cells were excluded from this analysis, because ACE2/TMPRSS2 overexpression led to growth arrest in these cells (data not shown). We infected cells with SARS-CoV-2 at an MOI of 1 and carefully monitored virus replication at 2, 4, 8, 12, and 24 hpi through multiple orthogonal assays, such as IF, quantitative real-time PCR (RT-qPCR), RNA Scope, and plaque assay. The viral N protein was first detected at 4 hpi (Figure 3A), indicating that the viral RNA replication started at some point between 2 and 4 hpi. Consistently, the viral negative-sense RNA, which is the replicative intermediate and an authentic marker of viral RNA replication only became detectable at 4 hpi (Figure 3B). The number of N-positive cells increased with time, reaching around 70% for HuH-6/AT, HuH-7/AT, Huh-7.5/AT, and Caco-2/AT cells by 12 hpi, compared to 50% for Vero E6 cells, indicating relatively accelerated viral spread in human cells (Figure 3A). Over 90% cells were infected for all cell lines by 24 hpi, except for A549 cells where the percentage of positive cells was around 80%.

Figure 3.

Time course analysis of SARS-CoV-2 infection

(A) Cells were infected with SARS-CoV-2 (Washington isolate) at an MOI of 1 on ice for 1h followed by 3X washing with 1 mL of ice-cold 1X PBS. After incubation at 37°C for the indicated times, the cells were stained with an anti-N antibody. The IF images were captured with a 10× objective lens. Representative images from one of the two experimental repeats are shown. The values represent the mean percentage of positive cells ±standard error of the mean of two experiments. AT, cells expressing ACE2 and TMPRSS2.

(B) SARS-CoV-2 negative-sense RNA was detected by RNA scope. The cell nuclei were stained with DAPI. The images were captured with a 20× objective lens.

(C and D) The culture medium of cells described in A was subjected to RT-qPCR (C) and plaque assay (D). The bar graphs represent mean ± standard error of the mean of two experiments.

RT-qPCR analysis of the culture medium revealed ∼100 to 360-fold increase in viral RNA copy number at 24 hpi (A549/AT, 107.3-fold; Caco-2/AT, 201.6-fold; HuH-6/AT, 359.8-fold; HuH-7/AT, 133.3-fold, Huh-7.5/AT, 133.5-fold; Vero E6, 240.1-fold) (Figure 3C). A similar pattern was observed with the plaque assay, where at 24 hpi, the highest viral titers were obtained for Caco-2/AT (6.8 × 10⁵ PFU/mL), HuH-6/AT (5.0 × 10⁵ PFU/mL), and Vero E6 cells (8.5 × 10⁵ PFU/mL). A549/AT and HuH-7/AT cells yielded the lowest titer of 2.3 × 10³ and 2.4 × 10³ PFU/mL, respectively (Figures 3D and S4). Together, these results indicate that Caco-2/AT and HuH-6/AT cells are comparable to Vero E6 cells in supporting infectious SARS-CoV-2 production.

Caco-2/AT and HuH-6/AT cells allow efficient isolation of SARS-CoV-2 from clinical specimens

Earlier reports demonstrated rapid mutation of SARS-CoV-2 upon growth in Vero E6 cells that can revert upon re-culture in human cells.²⁷ This suggested that clinical isolates of SARS-CoV-2 experience a barrier to efficient replication in Vero E6 cells, perhaps decreasing the overall sensitivity of these cells for isolating virus from clinical specimens. We reasoned that our SARS-CoV-2 infection-optimized cells would show an increased ability to support the replication of viral clinical isolates compared to Vero E6 cells. To test this idea, we directly infected Caco-2/AT, HuH-6/AT, and Vero E6 cells with 20 clinical specimens, most of which had been determined to be SARS-CoV-2-positive by RT-qPCR, with CT values between 17 and 25. All samples had known genetic sequences and were classified as variants of concern (VOC) (Figure S5A).

The cells were inoculated with 5-fold serial dilutions of viral transport medium (VTM) containing nasopharyngeal (NP) swabs of infected individuals, and virus growth was assessed by IF at 72 hpi. HuH-6/AT and Caco-2/AT cells showed remarkable sensitivity, facilitating virus propagation from samples diluted as high as 1: 3125 in most cases (Figures 4A and 4B). In contrast, we did not detect virus replication in Vero E6 cells inoculated with samples diluted beyond 1:25, and in some cases, the cells remained uninfected even when inoculated with undiluted VTM (Figure 4A). Consistent with the published literature,^11,12 Vero E6 cells did not support Omicron isolation, whereas Caco-2/AT and HuH-6/AT allowed highly efficient detection of these isolates (Figures S5B and S5C). Surprisingly, none of the three Omicron isolates tested grew in Vero E6/AT cells, suggesting that the inability of Vero E6 cells to support Omicron replication is not entirely due to low levels of entry factors on the surface of these cells (Figure S5C). Further, HuH-6/AT and Caco-2/AT cells largely yielded all-or-none infection pattern; either all cells in a well were infected or there was no infection, indicating a quick and rampant viral spread across cells (Figure S5D). The plaque assay showed that Caco-2/AT and HuH-6/AT cells released high amounts of infectious virus in the culture medium, with the mean virus titer of 1.6 × 10⁶ PFU/mL (range: 3.6 × 10⁴–1.4 × 10⁷ PFU/mL) (Figure 4C). Overall, these findings suggest that Caco-2/AT and HuH-6/AT cells offer a superior alternative to Vero E6 cells for isolation of infectious SARS-CoV-2 from clinical specimens.

Figure 4.

Caco-2/AT and HuH-6/AT cells efficiently isolated virus from clinical specimens

(A) Vero E6, Caco-2/AT, and HuH-6/AT cells were infected for 72h with 5-fold serial dilutions of nasal material collected from infected individuals and stained with an anti-N antibody. Pie charts indicate the percentage of infected cells as measured by IF. Some representative images are shown in Figure S4C. Infectious virus was not recovered from sample J, although viral RNA was detected by RT-qPCR (Figure S5A), likely due to virus inactivation during sample collection or transportation.

(B) The highest dilution of clinical samples that produced cell infection. ∗p = 0.01 and ∗∗p = 0.009, as calculated by a two-tailed, unpaired t test with Welch’s correction.

(C) Virus titer in the culture medium of cells infected with undiluted nasal material. The bar graph represents mean ± SD of three technical replicates.

Caco-2/AT cells allow genetically stable propagation of SARS-CoV-2 over multiple passages

We hypothesized that the increased permissivity of modified human cell lines to SARS-CoV-2 was due to decreased evolutionary pressure the virus encounters in these cells compared to Vero E6 cells. To test this, we sequenced a SARS-CoV-2 clinical isolate grown in parallel in Vero E6, Vero E6/AT and Caco-2/AT cells over five passages. The cells were inoculated, in duplicate, with the clinical isolate representing the SARS-CoV-2 B.1 lineage. Every three days, when the cytopathic effect of infection became visible, the virus was passaged on to fresh cells at an MOI of 0.1 (virus titration was carried out on the same cell line in which the virus was propagated) (Figure 5A). Whole genome sequencing at the end of five passages showed that the virus grown in Caco-2/AT cells was genetically identical to the parental virus (Figure 5B). In contrast, the virus propagated in Vero E6 and Vero E6/AT cells accumulated 2–4 amino acid changes, some of which were localized in the spike region (Figure 5C). Although our samples size is too small to draw broad conclusions, the results indicate that Caco-2/AT cells may represent a suitable cell type for relatively stable propagation of SARS-CoV-2 clinical isolates in cell culture.

Figure 5.

Caco-2/AT cells supported genetically stable propagation of a SARS-CoV-2 clinical isolate

(A) Schematics of the passage experiment. The P0 virus stock, recovered from a clinical specimen in Vero E6 cells, was passaged five times on Vero E6, Vero E6/AT, and Caco-2/AT cells at an MOI of 0.1.

(B) Frequency of single-nucleotide variations in P0 and P5 virus. The peaks in the right panel indicate mutations relative to the P0 virus.

(C) Amino acid changes identified in the P5 virus.

Caco-2/AT cells provide a robust platform for production of recombinant SARS-CoV-2

The existing reverse genetics systems to produce rSARS-CoV-2 are inefficient, requiring several days to recover virus and often producing low-titer stocks, which then need to be amplified by additional passages in cell culture,^{13,14,29,30,31} running the risk of cell culture adaptation. The majority of these systems use Vero E6 as virus-producer cells. Since Caco-2/AT cells demonstrated a remarkable susceptibility to SARS-CoV-2, we tested the utility of these cells for producing rSARS-CoV-2 following transfection of the viral genetic material.

We employed the recently described circular polymerase extension reaction (CPER) method to generate a SARS-CoV-2 infectious clone.^13,14 Nine overlapping cDNA fragments covering the entire genome of the SARS-CoV-2 isolate (SARS-CoV-2/Hu/DP/Kng/19–020), as well as a linker fragment, were pieced together by CPER (Figure 6A). Two changes were made to the viral sequence; a D614G substitution was introduced into the spike protein, and the ORF7a gene was replaced with the mCherry fluorescent protein to enable convenient tracking of viral infection. The CPER product was transfected into Caco-2/AT, Vero E6, and Vero E6/AT cells, and the rescue of rSARS-CoV-2 was monitored by fluorescence analysis of transfected cells and titration of the recovered virus.

Figure 6.

Caco-2/AT cells allowed highly efficient production of recombinant SARS-CoV-2

(A) Schematic representation of the CPER system.

(B) Cells transfected with a CPER product generated from D614G-containing SARS-CoV-2/mCherry were imaged by fluorescent microscope at indicated days post-transfection. The images were captured with a 10× objective lens.

(C) Viral titer in the culture medium of cells in B was measured by the plaque assay on Caco-2/AT cells. The graph represents mean ± SD of two independent transfection experiments.

(D and E) 293T cells transfected with the CPER product of a non-reporter D614G-containing virus (D) or Omicron BA.1 virus (E) were co-cultured, 24h after transfection, with Caco-2/AT, Vero E6, and Vero E6/AT cells. The culture medium of co-cultured cells was subjected to the plaque assay using Caco-2/AT cells. The graphs show mean ± SD of triplicate samples. The results were confirmed in two independent transfection experiments.

Red fluorescent signal appeared as early as two days post-transfection (dpt) in Caco-2/AT cells and rapidly spread across the culture in the next 24 h, leading to extensive cell death by 4 dpt (Figure 6B). The detection of red signal coincided with the production of infectious virus, which achieved the highest titer of 2.4 × 10⁷ PFU/ml at 3 dpt (Figure 6C). The recovered virus had identical sequence to the transfected DNA (data not shown), indicating genetically stable propagation of virus. In contrast to Caco-2/AT cells, no red signal or viral recovery was seen for Vero E6 cells during the experimental duration of 7 days. Virus particles could be recovered from Vero E6/AT cells at 7 dpt, but the titer was only 6 × 10² PFU/ml. The lack of virus production from Vero E6 cells was not due to inefficient DNA transfection, as all three of the cell lines yielded a similar signal when transfected with a mammalian expression plasmid carrying the green fluorescent protein (Figure S6).

In a recent report, transfection of CPER product into HEK293T cells followed by co-culture with Vero E6 cells successfully generated rSARS-CoV-2, although it took 6–9 days to rescue virus and required further amplification of the recovered virus in Vero E6 cells to obtain the highest titer of 5 × 10⁶-10⁷ PFU/ml¹⁴ We tested this approach with Caco-2/AT cells. When co-cultured with HEK293T cells being transfected with the CPER product of a none-reporter D614G-containing virus, Caco-2/AT cells yielded 1.58 × 10⁶ PFU/mL within 24 h, and the titer reached the highest value of 8 × 10⁷ PFU/ml at day 2 post-coculture (Figure 6D). In contrast, Vero E6 and Vero E6/AT cells produced only 2 × 10³ and 2 × 10⁴ PFU/mL at day 1 and 6.4 × 10⁴ and 7.8 × 10⁵ PFU/ml at day 2 post-coculture, respectively. Importantly, the virus recovered from Caco-2/AT cells on day 2 post-coculture was genetically identical to the transfected DNA. We observed similar viral growth patterns with Omicron (BA.1 lineage). While Caco-2/AT cells produced the highest viral titer of 5.1 × 10⁶ PFU/ml within two days of co-culture, Vero E6 and Vero E6/AT cells took three and four days to produce the highest viral titer of 1.8 × 10⁵ and 5.6 × 10⁴ PFU/ml, respectively (Figure 6E). Together, these results indicate that Caco-2/AT cells provide an excellent platform for production of rSARS-CoV-2 with high titers.

Lastly, we passaged the Caco-2/AT-generated recombinant D614G-containing virus in Vero E6, Vero E6/AT, and Caco-2/AT cells three times and performed whole genome sequencing of the P3 stock to determine the number of mutations accumulated by these viruses during passaging in distinct cellular backgrounds. Each sample was run in duplicate. Genomic RNA was isolated from the P3 stock and subjected to amplification using the Artic 4.1 primer scheme followed by amplicon sequencing using Illumina MiSeq.³² Interestingly, when mutations were analyzed at the cutoff threshold of 0.5 (mutations present in more than half of the sequencing reads), no changes were seen, except for the virus recovered from Vero E6 cells (three changes in one of the two replicates) (Table S1). To further evaluate mutations present at a low frequency, we analyzed our sequencing data at the cutoff threshold of 0.1 (mutations present in as low as 10% of reads). Under these settings, while the number of mutations in Vero E6-recovered virus increased, leading to a number of amino acid changes, only one amino acid change was seen in Caco-2/AT-recovered virus. These results indicate that Caco-2/AT cells may allow relatively stable propagation of SARS-CoV-2 over multiple passages.

Discussion

A suitable cell culture system for producing authentic SARS-CoV-2 stocks should possess at minimum four key characteristics: 1) it should allow easy growth, maintenance, and scalability of cells; 2) it should enable robust and highly sensitive isolation of infectious virus from patient samples; 3) it should allow production of high-titer virus stocks; 4) it should support virus propagation with minimal cell culture adaptation. The Caco-2/AT cells meet these criteria and offer an optimal platform for isolation and growth of SARS-CoV-2 clinical isolates.

The cell lines described in this study have numerous advantages over previously reported cell systems for growing SARS-CoV-2. Unlike Calu-3 cells and human airway organoid-based cultures that have also been shown to support SARS-CoV-2 growth without significant genomic changes,^27,33 Caco-2/AT cells are easy to engineer, grow, and expand for generating large-scale virus stocks. Further, these cells allow growth of clinical isolates for almost all known viral variants and represent a robust platform for rapid generation of rSARS-CoV-2, including Omicron that has shown to be less replicative in Vero E6 cells.^11,12 It should however be noted that Caco-2/AT cells do not necessarily represent an all-encompassing cell line for answering diverse questions about SARS-CoV-2 biology and pathogenesis. In our previous study, we observed that some genes found to be downregulated upon SARS-CoV-2 infection are not expressed in Caco-2 cells.³⁴ Further, these cells are of non-lung origin, diminishing their utility for investigating lung-specific effects of SARS-CoV-2. Therefore, while Caco-2/AT cells constitute an invaluable resource for isolating SARS-CoV-2 from clinical specimens and generating high-titer recombinant virus stocks, the choice of cell types for downstream investigations depends on the question(s) being asked.

Published evidence suggests that the route of virus entry might be central to genetically stable growth of SARS-CoV-2 in cell culture. The spike mutations in Vero-expanded virus have been shown to provide selective advantage to the virus for cathepsin-mediated entry, the major mode of SARS-CoV-2 internalization in Vero E6 cells. Passaging the Vero-grown virus in cells where SARS-CoV-2 entry is mediated through TMPRSS2 leads to reversal of spike mutations.²⁷ Similarly, forced evolution of SARS-CoV-2 toward TMPRSS2 entry pathway, such as by inhibiting cathepsins or by overexpressing TMPRSS2, arrests cell culture adaptation of virus.²⁷ At least two lines of evidence suggest that SARS-CoV-2 enters Caco-2 cells mainly through the TMPRSS2 pathway. First, in line with the published literature,³⁵ we found that the viral entry into Caco-2 cells is more strongly inhibited by camostat mesylate, an inhibitor of TMPRSS2, as compared to E64D that inhibits cathepsins (Figure S7). This contrasts with Vero E6 cells where E64D, but not camostat mesylate, causes strong inhibition of SARS-CoV-2 infection. Second, co-expression of TMPRSS2 and ACE2 in Caco-2 cells does not boost SARS-CoV-2 pseudoparticle entry beyond what is achieved with ACE2-only expression (Figures 2A and S3). These findings suggest that the cells that support TMPRSS2-mediated SARS-CoV-2 entry might have increased propensity to allow genetically stable viral propagation.

Interestingly, while several cell lines became highly susceptible to SARS-CoV-2 infection following ACE2 and TMPRSS2 overexpression, many others remained refractory to infection (HT-29, HUVEC) or were only modestly infected (HeLa, RD) (Figure 2B). This can be explained by at least two possibilities. First, these cells have a post-entry block due to the lack of essential host dependency factors for SARS-CoV-2 replication. This scenario however is less likely, as these cells also exhibited poor susceptibility to SARS-CoV-2 pseudoparticles (Figure 2A). The second possibility is that ACE2 and/or TMPRSS2 are aberrantly processed in these cells, leading to their poor cell surface localization or inability to interact with the SARS-CoV-2 spike protein. This notion is supported by a recent paper showing that endogenous ACE2 is differentially processed in various cell lines, reflected by its changed electrophoretic mobility in western blot.²⁶ A detailed investigation is warranted to illuminate the details of how post-translational modifications influence the ability of ACE2 to mediate SARS-CoV-2 entry.

In summary, our study reports simple and robust cell culture systems that allow isolation of infectious SARS-CoV-2 from clinical specimens, recovery of recombinant virus by reverse genetics, and production of high-titer virus stocks without significant genomic drift. This progress is vital for the field as it will help generate SARS-CoV-2 preparations that can be employed to obtain reliable information about in vitro and in vivo characteristics of existing and future SARS-CoV-2 variants.

Limitations of the study

To obtain human cells highly susceptible to SARS-CoV-2 infection, we began with a panel of 17 cell lines engineered to express ACE2 and TMPRSS2 transgenes. Seven of these cell lines demonstrated a poor susceptibility to SARS-CoV-2 pseudoparticles, as reflected by less than 20% infection. Some of these cell lines were also resistant to authentic SARS-CoV-2 infection. An in-depth investigation will be needed to elucidate the nature and mechanism of block in these cells. It is possible that the ACE2 and/or TMPRSS2 transgenes are poorly expressed, or the encoded proteins are differentially processed in these cells, leading to their impaired trafficking to the cell surface or poor interaction with the SARS-CoV-2 spike protein. Alternatively, these cells may lack essential viral entry factors. All of these possibilities should be explored in the future studies.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Sars Nucleocapsid Protein Antibody	Rockland	Cat#200-401-A50 RRID:AB_828403
Anti-ACE2 antibody [EPR4435(2)]	Abcam	Cat#ab108252 RRID:AB_10864415
Human/Mouse/Rat/Hamster ACE-2 Antibody	R&D Systems	Cat#AF933 RRID:AB_355722
Goat IgG Isotype Control	Thermo Fisher Scientific	Cat#02–6202 RRID:AB_2532946
Anti-VSV-G [8G5F11] Antibody	Kerafast	Cat#EB0010 RRID:AB_2811223
beta Actin Monoclonal Antibody (AC-15)	Thermo Fisher Scientific	Cat#AM4302 RRID:AB_2536382
Goat anti-Rabbit IgG (H+L) Cross-Adsorbed Secondary Antibody, Alexa Fluor™ 488	Thermo Fisher Scientific	Cat#A-11008 RRID:AB_143165
Donkey anti-Goat IgG (H+L) Cross-Adsorbed Secondary Antibody, Alexa Fluor™ 488	Thermo Fisher Scientific	Cat#A-11055 RRID:AB_2534102
IRDye 680RD Donkey anti-Rabbit IgG	Li-COR	Cat#926–68073 RRID:AB_10954442
IRDye 800CW Donkey anti-Mouse IgG	Li-COR	Cat#926–32212 RRID:AB_621847
Bacterial and virus strains
MC1061 Competent E. coli	MCLAB	Cat#MC1061-100
SARS-CoV-2/human/USA/WA-CDC-02982586-001/2020	BEI Resources	Cat#NR-52281; NCBI accession #MN985325
Chemicals, peptides, and recombinant proteins
cOmplete™, Mini, EDTA-free Protease Inhibitor Cocktail	Roche	Cat#11836170001
PhosSTOP	Roche	Cat#4906845001
E64D	TOCRIS	Cat#4545
DMEM, high glucose, pyruvate	Gibco	Cat#11995065
DMEM/F-12, HEPES	Gibco	Cat#11330032
MEM	Gibco	Cat#11095080
RPMI 1640 Medium	Gibco	Cat#11875093
Opti-MEM™ I Reduced Serum Medium	Gibco	Cat#31985070
Puromycin dihydrochloride from Streptomyces alboniger	Millipore Sigma	Cat#P8833
Blasticidine S hydrochloride	Millipore Sigma	Cat#3513-03-9
X-tremeGENE™9 DNA Transfection Reagent	Millipore Sigma	Cat#XTG9-RO
TransIT-X2® Dynamic Delivery System	Mirus	Cat#MIR6000
DPBS, no calcium, no magnesium	Gibco	Cat#14190144
Human TruStain FcX™ (Fc Receptor Blocking Solution)	BioLegend	Cat#422302
Paraformaldehyde, 8% w/v aq. soln., methanol free	Thermo Scientific	Cat#047347.9M
DAPI (4',6-diamidino-2-phenylindole, dihydrochloride)	Thermo Scientific	Cat#62247
Critical commercial assays
Quick-RNA Viral Kit	ZYMO	Cat#R1035
qScript® 1-Step qRT-PCR Kit ROX	Quantabio	Cat#95058-200
mMESSAGE mMACHINE™ SP6 Transcription Kit	Invitrogen	Cat#AM1340
RNAscope®Multiplex Fluorescent Reagent Kit V2	ACD	Cat#323100
Pierce™ BCA Protein Assay Kit	Thermo Scientific	Cat#23225
Experimental models: Cell lines
HEK293T	ATCC	Cat#CRL-3216; RRID:CVCL_0063
HeLa	ATCC	Cat#CCL-2; RRID:CVCL_0030
HT-29	ATCC	Cat#HTB-38; RRID:CVCL_0320
HuH-6 Clone 5	JCBR Cell Bank	Cat#JCRB0401; RRID:CVCL_1296
HuH-7	JCBR Cell Bank	Cat#JCRN0403; RRID:CVCL_0336
HuH-7.5	Derivative of HuH-7	RRID:CVCL_7927
Hep-G2	ATCC	Cat#HB-8065; RRID:CVCL_0027
Calu-3	ATCC	Cat#HTB-55; RRID:CVCL_0609
Calu-6	ATCC	Cat#HTB-56; RRID:CVCL_0236
A549	ATCC	Cat#CCL-185; RRID:CVCL_0023
MRC-5	ATCC	Cat#CCL-171; RRID:CVCL_0440
HK-2	ATCC	Cat#CRL-2190; RRID:CVCL_0302
SK-N-SH	ATCC	Cat#HTB-11; RRID:CVCL_0531
VERO E6	ATCC	Cat#CRL-1586; RRID:CVCL_0574
RD	ATCC	Cat#CCL-136; RRID:CVCL_1649
Caco-2	ATCC	Cat#HTB-37; RRID:CVCL_0025
AC16 Human Cardiomyocyte Cell Line	Millipore Sigma	Cat#SCC-109; RRID:CVCL_4U18
HUVEC	ATCC	Cat#CRL-1730; RRID:CVCL_2959
Recombinant DNA
SARS-CoV-2, pCG1_SARS-2_S	Stefan Pohlmann35
pLOC_hACE2_PuroR	This study
pLOC_hTMPRSS2_BlastR	This study
pCSII-sars-cov-2 F1	Yoshiraru Matsuura13
pCSII-sars-cov-2 F2	Yoshiraru Matsuura13
pCSII-sars-cov-2 F3	Yoshiraru Matsuura13
pCSII-sars-cov-2 F4	Yoshiraru Matsuura13
pMW118-sars-cov2 F5	Yoshiraru Matsuura13
pcDNA3.1.-sars-cov2 F6	Yoshiraru Matsuura13
pCSII-sars-cov-2 F7	Yoshiraru Matsuura13
pCSII-sars-cov-2 F8	Yoshiraru Matsuura13
pcDNA3.1.-sars-cov2 F9-10	Yoshiraru Matsuura13
pMW118 CoV2-CMVlinker	Yoshiraru Matsuura13
SARS-CoV-2 E gene onto pIDTBlue	This study
pcDNA3.1-eGFP	Our previous study36
pCSII-sars-cov-2 Omicron F1	Our previous study36
pCSII-sars-cov-2 Omicron F3	Our previous study36
pCSII-sars-cov-2 Omicron F4	Our previous study36
pcDNA3.1.-sars-cov2 Omicron F6	Our previous study36
pCSII-sars-cov-2 Omicron F8	Our previous study36
pcDNA3.1.-sars-cov2 Omicron F9-10	Our previous study36
Software and algorithms
FlowJo	FlowJo	https://www.flowjo.com/
Prism	Graphpad by Dotmatics	https://www.graphpad.com/features
ImageJ	NIH and LOCI, University of Wisconsin	https://imagej.nih.gov/ij/download.html
MetaXpress	Molecular Devices	https://www.moleculardevices.com/products/cellular-imaging-systems

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to the lead contact and corresponding author, Mohsan Saeed (msaeed1@bu.edu).

Materials availability

All reagents generated in this study are available from the corresponding author with a completed Materials Transfer Agreement.

Experimental model and subject details

Biocontainment

All experiments involving authentic SARS-CoV-2 were performed in a state-of-the-art biosafety level 3 (BSL3) facility at the National Emerging Infectious Diseases Laboratories (NEIDL) of Boston University using biosafety protocols approved by the Institutional Biosafety Committee (IBC), comprising scientists, biosafety and compliance experts as well as local community members. The biosafety protocols were further approved by the Boston Public Health Commission. All personnel received rigorous biosafety, biosecurity, and BSL3 training before participating in experiments. Special personal protective equipment, including scrubs, disposable overalls, shoe covers, double-layered gloves, and powered air-purifying respirators, was used during BSL3 work.

Cells

All cell lines were incubated at 37°C and 5% CO₂ in a humidified incubator. Human embryonic kidney HEK293T cells (ATCC; CRL-3216), human cervical carcinoma HeLa cells (ATCC; CCL-2), human colorectal adenocarcinoma HT-29 cells (ATCC; HTB-38), human hepatoblastoma HuH-6 cells (JCRB-0401), human hepatocellular carcinoma HuH-7 cells (JCRB-0403) and its derivative Huh-7.5 cells, human hepatocellular carcinoma HepG2 cells (ATCC; HB-8065), human lung anaplastic carcinoma Calu-6 cells (ATCC; HTB-56), human lung adenocarcinoma A549 cells (ATCC; CCL-185), human normal lung fibroblast MRC-5 cells (ATCC; CCL-171), human kidney papilloma HK-2 cells (ATCC; CRL-2190), human neuroblastoma SK-N-SH cells (ATCC; HTB-11), African green monkey kidney Vero E6 cells, and human muscle rhabdomyosarcoma RD cells (ATCC; CCL-136) were maintained in DMEM (Gibco; #11995–065) containing 10% FBS and 1X non-essential amino acids, human colorectal adenocarcinoma Caco-2 cells (ATCC; HTB-37) were maintained in the same medium but containing 20% FBS, human cardiomyocyte AC16 cells (Millipore; SCC109) in DMEM/F12 (Gibco; #11330–032) containing 12.5% FBS, human lung adenocarcinoma Calu-3 cells (ATCC; HTB-55) in MEM (Gibco; #11095–080) containing 10% FBS, and human umbilical vein endothelial HUVEC cells (ATCC; CRL-1730) in RPMI 1640 medium (Gibco; #11875–093) containing 10% FBS. Mycoplasma negative status of all cell lines was confirmed. Detailed information regarding all the cell lines used was provided in the key resources table.

Antibodies and chemicals

Anti-SARS-CoV nucleocapsid (N) protein antibody (Rockland; #200-401-A50) was used for detection of SARS-CoV-2 N protein by IF. Anti-hACE2 antibodies included rabbit monoclonal antibody EPR4435(2) (Abcam; #ab108252) for Western blot and goat polyclonal antibody (R&D Systems; #AF933) for flow cytometry. Goat IgG isotype control antibody was from Invitrogen (#02–6202), Anti-VSV-G antibody from Kerafast (#EB0010), and ß-actin antibody from Invitrogen (#AM4302).

Plasmids

Expression plasmid encoding the spike protein of SARS-CoV-2, pCG1_SARS-2_S, has recently been described and was a kind gift from Stefan Pohlmann.³⁵ The plasmids, pLOC_hACE2_PuroR and pLOC_hTMPRSS2_BlastR, containing human ACE2 and TMPRSS2, respectively, have previously been described.³⁴ Plasmids encoding various fragments of the SARS-CoV-2 genome (Hu/DP/Kng/19–020 strain) and a linker plasmid encoding the last 43 nucleotides of the SARS-CoV-2 genome, a hepatitis delta virus ribozyme (HDVr), the bovine growth hormone (BGH) poly(A) signal, a cytomegalovirus (CMV) promoter, and the first 25 nucleotides of the SARS-CoV-2 genome were a kind gift from Yoshiraru Matsuura.¹³

Method details

Generation and characterization of stable cell lines

Cells stably expressing human ACE2, human TMPRSS2 or the combination of both, were generated by lentiviral transduction with pLOC_hACE2_PuroR, pLOC_hTMPRSS2_BlastR, or the combination of both plasmids. The ACE2-expressing cells were then selected with 1 μg/ml puromycin, the TMPRSS2-expressing cells with 1 μg/ml blasticidin, and the combination (AT) cells with 1μg/ml concentration of each of puromycin and blasticidin. The expression of ACE2 in select cell lines was confirmed by Western blot and flow cytometry.

VSV pseudoparticle production and infection

VSV pseudoparticles carrying the SARS-CoV-2 S protein were generated as previously described.³⁵ Briefly, 293T cells were seeded into a 6-well plate at the density of 5 × 10⁵ cells per well. The next day, the cells were transfected with 6 μg of pCG1_SARS-2_S using X-tremeGENE 9 DNA transfection reagent (Sigma-Aldrich). Twenty hours later, the cells were inoculated with a replication-competent vesicular stomatitis virus (VSV) modified to contain an expression cassette for GFP in place of the VSV-G open reading frame, VSV∗ΔG-GFP (kindly provided by Stefan Pohlmann of the Leibniz Institute for Primate Research (DPZ), Gottingen, Germany). Following 2h incubation of cells with VSV∗ΔG-GFP at 37°C, the inoculum was removed and the cells were washed three times with FBS-free DMEM. Two ml of DMEM/10%FBS supplemented with anti VSV-G antibody was then added to each well to neutralize any residual input VSV. Pseudoparticles were harvested 20h later, passed through 0.45 μ filter, and stored at -80°C.

For infection experiments, target cells grown in 24-well plates to the confluency of 70–80% were infected with pseudoparticles. Twenty-four hours later, the cells, including those floating in the culture medium due to cytopathic effects of VSV replication, were collected and fixed in 4% PFA. Flow cytometry was performed to determine the percentage of GFP positive cells.

SARS-CoV-2 stock preparation and titration

SARS-CoV-2 stock was prepared as previously reported.³⁴ Briefly, the SARS-CoV-2 Washington isolate (NCBI accession number: MN985325), obtained from the Centers for Disease Control and Prevention and BEI Resources, was passaged twice on Vero E6 cells to obtain the P2 working stock. For measurement of virus titer, we seeded Vero E6 cells into a 12-well plate at a density of 2.5 × 10⁵ cells per well. The next day, the cells were incubated with serial 10-fold dilutions of the virus stock (200 μl volume per well) for 1h at 37°C, followed by the addition of 1 ml per well of the overlay medium made of 1:1 mixture of 2X DMEM/4% FBS and 1.2% Avicel (DuPont; RC-581). After three days of incubation at 37°C, the overlay medium was removed and the cell monolayer was washed with 1X PBS and fixed with 4% paraformaldehyde. The fixed cells were then stained with 0.1% crystal violet for 1h at room temperature and rinsed with tap water. The numbers of plaques were counted and the virus titer was calculated.

Generation of recombinant SARS-CoV-2 by CPER

SARS-CoV-2 recombinant virus was generated by using a recently published CPER protocol.¹³ Plasmids encoding nine overlapping fragments (P1-P9) covering the entire SARS-CoV-2 genome (Hu/DP/Kng/19–020 strain) and a linker plasmid (described under plasmids) were used as templates. To obtain the recombinant D614G-containing virus, we introduced a D614G substitution into the spike protein in the P8 plasmid and, for some experiments, replaced the ORF7a open reading frame in the P9 plasmid with the mCherry fluorescent protein. As previously described, we obtained the recombinant Omicron by introducing the non-synonymous mutations found in an Omicron BA.1 lineage virus (USA-lh01/2021) into respective SARS-CoV-2 plasmids.³⁶ All fragments were then amplified by PCR using the previously described primers and protocol and connected by CPER to obtain the full-length SARS-CoV-2 cDNA clone under the CMV promoter.¹³

We used reverse transfection to introduce the assembled SARS-CoV-2 cDNA clone into cells. The transfection mix was prepared by mixing the CPER product with 500 μl of Opti-MEM (Thermo Fisher Scientific; #31985070) and 12 μl of TransIT-X2 Dynamic Delivery System (Mirus Bio; #MIR 6000). Following incubation at room temperature for 20 min, the transfection mix was added to a cell suspension containing a total of 1 × 10⁶ cells in 4 ml of DMEM/10% FBS medium. The cells were seeded into two wells of a 6-well plate to obtain the final density of 5 × 10⁵ cells/well. The next day, the culture medium was replaced with DMEM/2% FBS medium, and the cells were monitored for the next six days for appearance of red fluorescent signal and production of infectious virus.

For co-culture, we transfected 293T cells with the CPER product by reverse transfection and seeded them into a 6-well plate at a density of 5 × 10⁵ cells per well for 24h. The next day, we harvested the cells, split them into three portions, and seeded each portion on Caco-2/AT, Vero E6, and Vero E6/AT cells seeded the day before in 12-well plates at a density of 2 × 10⁵ cells per well. Recovery of infectious virus was monitored by the plaque assay at different days of overlay.

SARS-CoV-2 passaging in cells

Clinical isolate

For the first passage, cells seeded into 12-well plates at a density of 2 × 10⁵ cells per well were inoculated with 1/8 dilution of VTM in which the nasal swabs were immersed. The dilution was prepared in Opti-MEM containing 50 μg/ml of Gentamycin and 0.5 μg/ml of Amphotericin B. Following 1h incubation at 37°C for 1h, 1 ml of 1X DMEM supplemented with 10% FBS, 50 μg/ml of Gentamycin, and 0.5 μg/ml of Amphotericin B was added to each well. Three days later, the culture medium was collected, passed through 0.45μ filters, and stored at −80°C as a P0 stock. This virus was then passaged into Vero E6, Vero E6/AT or Caco-2/AT cells five times. The passaging was done by mixing 25 μl of culture medium from preceding passage with 175 μl of Opti-MEM and inoculating the cells seeded the day before in 12-well plates at a density of 2 × 10⁵ cells per well. This roughly yielded the MOI of 0.1 PFU/cell. Following 1h incubation at 37°C, the inoculum was removed and the cell monolayer was washed three times with 1X PBS. One ml of fresh medium containing 10% FBS was added to each well, and the cells were incubated at 37°C for three days before the culture medium was passaged again essentially as described above. After five passages, RNA was extracted from the culture medium and subjected to next generation sequencing.

Recombinant SARS-CoV-2

The D614G-containing virus, generated by CPER in Caco-2/AT cells, was passaged in Vero E6, Vero E6/AT or Caco-2/AT cells three times essentially as described above.

SARS-CoV-2 whole genome sequencing

The whole genome sequencing of the recombinant virus was performed using the Artic V4.1 primer scheme as previously described.³² Briefly, RNA was isolated from the supernatant of virus-infected cells using the Quick-RNA Viral Kit (Zymo, #1035), quantitated using the CDC N1 assay, and RNA diluted to Ct above 18. Amplicons generated were fragmented and sequenced using Illumina MiSeq. Reads were subjected to quality trimming and mapped to the DNA template used for CPER. Primers were trimmed and reads deduplicated prior to final mutation calling at 0.5 and 0.1 read frequencies.

Quantitative real-time PCR (RT-qPCR)

We isolated RNA from the cell culture supernatant of SARS-CoV-2-infected cells by using the Quick-RNA Viral Kit (Zymo; #R1035) according to the manufacturer’s instructions. We then quantified the viral RNA by single-step RT-qPCR using Quanta qScript One-Step RT-qPCR Kit (Quantbio; #95058) with primers and Taqman probes targeting the SARS-CoV-2 E gene as previously described.³⁷ Data were acquired using a Quantstudio3 Real-Time PCR System (Applied Biosystems) using the following conditions: 55°C for 10 min, denaturation at 94°C for 3 min, 45 cycles of denaturation at 94°C for 15 sec, and annealing at 58°C for 30 sec. The primers and probe used were as follow: E_Sarbeco_Forward: ACAGGTACGTTAATAGTTAATAGCGT, E_Sarbeco_Probe: FAM-ACACTAGCCATCCTTACTGCGCTTCG-BBQ, and E_Sarbeco_R: ATATTGCAGCAGTACGCACACA. For absolute quantification of the viral RNA, we generated the RT-qPCR standard by cloning a 389-bp fragment from the SARS-CoV-2 E gene onto pIDTBlue plasmid under an SP6 promoter and subjecting this fragment to in vitro transcription using the mMessage mMachine SP6 transcription kit (ThermoFisher; #AM1340).

Immunofluorescence

Virus-infected cells were fixed in 4% paraformaldehyde for 30 minutes. The fixative was removed, and the cell monolayer washed twice with 1X PBS. The cells were then permeabilized and incubated overnight at 4°C with anti-SARS-CoV Nucleocapsid antibody (1:2,000 dilution). The cells were then washed 5 times with 1X PBS and stained with Alexa Fluor 568-conjugated goat anti-rabbit secondary antibody (1:1000 dilution) (Invitrogen; #A11008) in the dark at room temperature for 1h and counterstained with DAPI. The cells were imaged on an IXM-C high content imager (Molecular Devices) with a 4x S Fluor objective lens at a resolution of 1.7 microns/pixel in the DAPI (excitation: 400 nm/40 nm, emission: 447 nm/60 nm) and TexasRed (excitation: 570 nm/80 nm, emission: 624 nm/40 nm) channels. Both channels were used to establish their respective laser autofocus offsets.

For quantitative analysis of images, cells were segmented in MetaXpress using the CellScoring module. First, objects between 7 and 20 microns in diameter and greater than 1800 gray level units in intensity were identified and classified as nuclei. Positive cells were taken as nuclei having TexasRed signal of 1500 gray level units or above within 10 to 20 microns of each nucleus. The remaining objects were set to negative cells. From these objects, the following readouts were measured and used for downstream analysis: Total number of positive and negative cells, total area of positive cells, and integrated intensity in the TexasRed channel for positive cells.

Flow cytometry

For cell surface analysis of ACE2, we harvested cells and washed them in FACS Buffer (2% FBS in 1X PBS). Cells were resuspended in 1:50 dilution of human F_C blocking solution (BioLegend; #422302) and incubated on ice for 10 min. Human ACE2 antibody or goat IgG isotype control was then added to the cells to obtain the final concentration of 5 μg/mL followed by 1h incubation on ice. The cells were washed with FACS buffer and incubated for 30 minutes on ice in the dark with 1:400 dilution of Alexa Fluor 488 donkey anti-goat secondary antibody (Invitrogen; #A11055). The cells were washed and resuspended in FACS buffer. Data were collected using a BD LSR II flow cytometer and analyzed with FlowJo software (version 10).

RNA scope

SARS-CoV-2 RNA in infected cells was detected by using the RNAscope® Multiplex Fluorescent Kit v2 (ACDBio; #323100). Briefly, the cells were fixed in 4% paraformaldehyde in PBS for 30 min at room temperature and incubated with 50 mM NaOH to denature RNA. The cells were then dehydrated by treatment with increasing concentrations of EtOH (50%, 70%, and 100%) and stored at -20°C until further processing. On the day of processing, the cells were rehydrated by using decreasing EtOH concentrations (100%, 70%, and 50%) followed by a 10-min incubation with 1X PBS. The cells were then blocked with H₂O₂ for 10 min at room temperature and digested with the kit-provided proteinase for 30 min at 40°C, followed by incubation with a forty ZZ ISH probe designed and synthesized by Advanced Cell Diagnostics (ACDBio). The probe (#845701) targeted the 21631-23303 nucleotide region (spike protein) of the viral negative-sense (replicative intermediate) RNA. Probe hybridization was carried out at 40°C for 2h in HybEZ™ II Oven (ACDBio), and the signal was amplified using kit-provided Pre-amplifier and Amplifiers. The last set of Amplifiers was conjugated to HRP-C2. Finally, the cells were incubated with the fluorophore, and the nuclei were stained with DAPI. The images were recorded using EVOS XL Core fluorescent microscope (ThermoFisher Scientific).

Western blotting

Cell lysates were prepared in 1x RIPA buffer containing 1x complete-mini protease inhibitor (Roche; #11836170001) and 1x phosphatase inhibitor cocktail (Roche; #04906837001). The lysates were then incubated on ice for 30 min and centrifuged at 12,000 xg for 20 min at 4°C. The supernatants were transferred to new ice-cold Eppendorf tubes and protein concentration was measured by the BCA assay using Pierce BCA Protein Assay kit (ThermoFisher Scientific; #23225). Equal amounts of protein were loaded on 4–12% SDS-PAGE gel and transferred onto nitrocellulose membrane. Following staining with primary and secondary (LiCor) antibodies, the bands were visualized by scanning the membrane with the LiCor CLx infrared scanner. The intensity of protein bands was measured in the open-source package, ImageJ.

Quantification and statistical analysis

Statistical analyses were performed on individual experiments, as indicated, with GraphPad Prism 9 Software.

Published: April 10, 2023

Data and code availability

• •The published article and supplemental information include all data generated and analyzed during this study.
• •This paper does not report original code.
• •Any additional information required to reanalyze the data reported in this paper is available from the corresponding author upon request.