Optimized Pseudotyping Conditions for the SARS-COV-2 Spike Glycoprotein

A codon-optimized, full-length Spike gene was synthesized and introduced in place of the green fluorescent protein (GFP) gene in plasmid pEGFP-N1. Because truncation of the cytoplasmic tail of SARS Spike was shown to enhance production of viral pseudotypes (15), we also generated a SARS-COV-2 Spike subclone with the last 19 amino acids deleted (Fig. 1). Pseudotyped particles were generated with the two Spike proteins as well as VSV-G with glycoprotein-defective MLV, HIV-1, and VSV particles (Fig. 1). To avoid potential false-positive signals, a reporter system was utilized where the Cre recombinase gene is expressed by HIV-1 and MLV particles. When these particles transduce the Cre reporter cell line, the protein causes a recombination in an engineered reporter element that results in GFP expression. Because the virus-producing cells do not express GFP, they are eliminated as a source of false-positive signals. This 293FT-based reporter cell line was further engineered to express human ACE2. Infectious particle production with all three types of particles was very low with the full-length Spike protein, but the Δ19 Spike produced viral titers approaching 10⁴ infectious particles/ml with both HIV-1 and VSV particles. The Δ19 truncation also enhanced infectivity with MLV particles, although the viral titer remained near background. Infectious particle production remained over 1,000-fold lower than that with the control glycoprotein, VSV-G.

To allow rapid quantitation with minimum background, we proceeded with an Env-defective HIV-1 provirus containing a Gaussia luciferase (Gluc) gene in the reverse orientation containing a forward intron (Fig. 2A) (21). This viral construct produces very low background because it is not capable of producing a luciferase signal unless the gene is reverse transcribed in the target cell. Infectivity with SARS-COV-2 Spike requires expression of the host receptor ACE2. Our starting cell line was generated using a retroviral transfer vector containing ACE2 and a puromycin selection cassette. To determine if introduction of an additional stable copy of ACE2 would enhance susceptibility to infection, we generated a second retroviral transfer vector with a blasticidin resistance cassette. Retroviral particles were generated with this vector and used to stably transduce 293FT or 293FT/ACE2 (Puro) cells. Each of the cell lines was transduced with HIV-1-Gluc particles pseudotyped with SARS-COV-2 Δ19 Spike (Fig. 2B). As expected, 293FT cells were not susceptible to infection with the HIV-1/SARS-COV-2 Δ19 Spike. However, both cell lines containing a single introduction of ACE2, and the cell line containing two introductions, had approximately 1,000-fold increases in the luciferase signal, but the surface ACE2 expression and viral susceptibility were roughly equivalent among the three cell lines (Fig. 2B and C).

The SARS-COV-2 Spike requires proteolytic priming during infection by either cysteine proteases CatB/L or serine protease TMPRSS2, produced in the target cell (3, 6). To determine whether introduction of TMPRSS2 would enhance susceptibility to infection, we synthesized a codon-optimized TMPRSS2 gene, introduced this gene into a retroviral transfer vector, and stably transduced 293FT or 293FT/ACE2 cells. The stable introduction of TMPRSS2 into 293FT cells did not impart sensitivity to transduction with HIV-1 particles pseudotyped with SARS-COV-2 Δ19 Spike (Fig. 3). However, stable introduction of TMPRSS2 into 293FT/ACE2 cells increased the Gluc signal from target cells by 5- to 10-fold. Neither ACE2 expression nor TMPRSS2 expression affected the Gluc signal from particles pseudotyped with VSV-G.

To explore additional genetic modifications in SARS-COV-2 that might enhance infectious particle production, we replaced the endogenous signal peptide with a well-characterized strong signal peptide (22), and separately, we introduced the modifications D614G and R682Q, alone or in combination, into SARS-COV-2 Δ19 Spike. Both of these modifications have been reported previously to increase SARS-COV-2 Spike efficiency (Fig. 4A) (8, 12–14). Infectious particle production was assessed using the Cre reporter system so that a viral titer could be determined, and transductions were performed in cell lines expressing ACE2 with or without TMPRSS2. Replacement of the signal peptide did not increase or decrease infectious particle production (Fig. 4B). However, the D614G and R682Q modifications both significantly enhanced infectious particle production under all conditions. The D614G modification enhanced infectious particle production by 2- or 1.5-fold in cells lacking or containing TMPRSS2, respectively. The R682Q modification enhanced infectious particle production by 5- or 3.5-fold in cells lacking or containing TMPRSS2, respectively. The D614G R682G double modification consistently produced the highest titer of infectious particles, although the difference between D614G R682G and R682G alone was not statistically significant. The highest average titer of 9 × 10⁵ was achieved with the D614G R682G Δ19 Spike in cells expressing both ACE2 and TMPRSS2. To compare this engineered cell to a cell line that is naturally susceptible to SARS-COV-2 infection, we did parallel infections of 293FT (ACE2⁺ TMPRSS2⁺) cells and Vero E6 cells. Because Vero E6 cells are resistant to HIV-1 infection, pseudotypes with MLV particles containing a GFP reporter were used. Infectivity in 293FT cells (ACE2⁺ TMPRSS2⁺) was consistently higher than that in VeroE6 cells, although the difference was not statistically significant.

To test if the pseudoparticles produced were suitable for use in a neutralization assay, we incubated HIV-1 particles pseudotyped with SARS-COV-2 Δ19 Spike, with or without the D614G R682Q modifications, with serial dilutions of control plasma and plasma from COVID-19 patients before adding 293FT/ACE2/TMPRSS2 cells. Robust and quantifiable neutralization was detected in plasma from COVID-19 patients, but not in control plasma (Fig. 5). The D614G R682Q modification did change the overall sensitivity to neutralizing antibodies.

The gammaretroviral transfer plasmids pQCXIP (puromycin resistance) and pQCXIH (hygromycin resistance) were obtained from Clontech. The blasticidin gammaretroviral transfer vector (pQCXIB) was generated by replacing the puromycin resistance cassette from pQCXIP with a blasticidin resistance cassette. The gammaretroviral vector for generating the Cre sensor cell line (MLV lox-mTomato-lox-GFP-Blast) was engineered into pQCXIB and contained the monomeric Tomato gene flanked by Lox sequences (ATAACTTCGTATAGCATACATTATACGAAGTTAT), which were followed immediately by enhanced GFP (eGFP). The ACE2 transfer vectors were generated by engineering the human ACE2 gene (NCBI reference sequence NM_021804.3) into pQCXIP and pQCXIH. The TMPRSS2 transfer vector was generated by synthesizing a codon-optimized version of the human TMPRSS2 gene (NCBI reference sequence NP_001128571.1) and engineering it into the pQCXIH vector. The HIV-1-CMV-Cre vector was an NL4-3-derived provirus that is defective in Vif, Vpr, and Env and has the Nef gene replaced with cytomegalovirus (CMV) Cre. The plasmid was generated by replacing the GFP gene from the previously described plasmid HIV-CMV-GFP+Vpu (26) with the Cre gene from plasmid MLV-Cre (27), provided by Alan Rein (NCI at Frederick). The MLV+ CMV Cre vector was generated by replacing the GFP gene from MLV-GFP (provided by Shan-Lu Liu, The Ohio State University) with the Cre gene from MLV-Cre (27), provided by Alan Rein (NCI at Frederick). The MLV GagPol expression construct was provided by Walther Mothes (Yale University). The HIV-1-Gluc vector was described previously (28). The VSV-G expression construct was obtained from the NIH AIDS Reagent Program (29). The SARS-COV-2 expression vector was generated by synthesizing a codon-optimized version of the SARS-COV-2 Spike (GenBank accession number MN985325.1). The Δ19 Spike and the Δ19 R682Q Spike were generated by PCR mutagenesis. The Spike subclone with the alternative signal peptide contained the previously described H7 peptide (22) and was generated by synthesizing the 5′ end of the Spike gene with the alternative signal peptide and introducing it into the SARS-COV-2 Δ19 Spike clone.

All transfections were performed in 6-well plates. 293FT cells were transfected with a total of 1 μg of plasmid and 4 μg of polyethyleneimine (PEI) (30). For HIV-1-CMV-Cre particles, cells were transfected with 900 ng of provirus and 100 ng of glycoprotein expression vector (Fig. 1) or with 800 ng of provirus and 200 ng of glycoprotein expression vector (subsequent figures). For HIV-1-Gluc particles, cells were transfected with 800 ng of provirus and 200 ng of glycoprotein expression vector. For MLV particles, cells were transfected with 500 ng of the MLV GagPol expression vector, 400 ng of MLV-CMV-Cre, and 100 ng of glycoprotein expression vector. For VSV particles, cells were transfected with 1 μg of glycoprotein expression vector and were infected 2 days posttransfection with >10⁷ infectious units/well of VSVΔG-GFP (Kerafast) (31). Cells were rinsed with phosphate-buffered saline (PBS) 1 h after infection, and the medium was replaced with complete medium supplemented with 2 μl of mouse hybridoma supernatant containing anti-VSV-G antibody I1 (Kerafast) to neutralize input virus. Neutralizing antibody was excluded from samples pseudotyped with VSV-G. VSV pseudoparticles were collected 24 h later.

Supernatants containing virus were frozen at –80°C for at least 4 h, thawed, and spun at 3,200 × g for 5 min, and the same volume of medium was added to target cells with 20 μg of hexadimethrine bromide per ml (H9268; Sigma). For assays with a fluorescent readout, infected cells were collected at about 2 to 3 days postinfection, fixed with 4% paraformaldehyde (PFA), washed with PBS, and analyzed on an Accuri C6 flow cytometer. For infections with a Gluc readout, transductions were allowed to proceed for 2 to 3 days, and 20 μl of supernatant from each well was transferred to a black 96-well plate for measuring Gaussia luciferase activity with 50 μl of 10 μm coelenterazine in 0.1 M Tris (pH 7.4) and 0.3 M sodium ascorbate (NanoLight Technology). Luminescence, representing infectivity, was measured from the supernatant using a PerkinElmer EnSpire 2300 multilabel reader.

The 293FT cell line was obtained from Invitrogen. All cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum, 2 mM l-glutamine, 1 mM sodium pyruvate, 10 mM nonessential amino acids, and 1% minimal essential medium (MEM) vitamins.

Cells were washed with PBS and treated with 10 mM EDTA. Cells were then collected with PBS, centrifuged at 300 × g for 5 min, and supernatant removed. Cells were blocked with 5% chicken serum in PBS for 30 min on ice. Cells were centrifuged at 300 × g for 5 min and supernatant removed. A goat anti-hACE2 ectodomain antibody (R&D Systems) was applied to cells at 1:100 in 1% chicken serum in PBS for 1 h. Cells were then washed 3 times with PBS, suspended in an Alexa Fluor 488-conjugated chicken anti-goat antibody solution (Thermo Fisher), and diluted 1:500 in 400 μl 1% chicken serum in PBS for 1 h. Cells were then washed 3 times with PBS and suspended in 100 μl of 4% PFA. After incubation for 20 min on ice, cells were centrifuged at 300 × g for 5 min, and the supernatant was removed. Cells were resuspended in 200 μl of PBS and analyzed via flow cytometry (Accuri).

The Cre sensor cell line was generated by transfecting 293FT cells with 500 ng MLV GagPol expression vector, 400 ng of retroviral transfer vector MLV lox-mTomato-lox-GFP-Blast, and 100 ng of VSV-G expression vector. Viral medium was used to transduce 293FT cells, and cells were selected with blasticidin (5 μg/ml) beginning 2 days postransduction and were maintained until control treated cells were all eliminated. A clonal isolate from these cells was selected that expressed mTomato but no GFP. The ACE2 cell lines were generated by transfecting 293FT cells with 500 ng MLV GagPol expression vector, 400 ng of retroviral transfer vector pQCXIP-ACE2 or pQCXIH-ACE2, and 100 ng of VSV-G expression vector. Viral medium was used to transduce 293FT cells or the 293FT sensor cell line, and cells were selected with puromycin (1 μg/ml) or hygromycin (200 μg/ml) beginning 2 days postransduction and were maintained until control treated cells were all eliminated. The TMPRSS2 cell line was generated by transfecting 293FT cells with 500 ng MLV GagPol expression vector, 400 ng of retroviral transfer vector pQCXIH-TMPRSS2, and 100 ng of VSV-G expression vector. Viral medium was used to transduce 293FT, 293FT/ACE2, or 293/Cre-sensor/ACE2 cells, which were selected with hygromycin (200 μg/ml) beginning 2 days postransduction and were maintained until control treated cells were all eliminated.

Deidentified patient plasma samples were obtained from the MU Health Care clinical laboratory. Plasma was heat inactivated for 30 min at 58°C. Twofold serially diluted plasma was incubated with HIV-1/SARS-COV-2 Spike pseudotypes for 1 h at 37°C. The mixture was subsequently incubated with 293/ACE2/TMPRSS2 cells, approximately 20,000 cells per well in a 96-well plate. Gluc measurements were taken 3 days postransduction. Neutralization assays were performed in triplicate, and titers were calculated using regression analyses to correspond to 50% inhibition.