Rapid, reliable, and reproducible cell fusion assay to quantify SARS-Cov-2 spike interaction with hACE2

COVID-19 is a global crisis of unimagined dimensions. Currently, Remedesivir is only fully licensed FDA therapeutic. A major target of the vaccine effort is the SARS-CoV-2 spike-hACE2 interaction, and assessment of efficacy relies on time consuming neutralization assay. Here, we developed a cell fusion assay based upon spike-hACE2 interaction. The system was tested by transient co-transfection of 293T cells, which demonstrated good correlation with standard spike pseudotyping for inhibition by sera and biologics. Then established stable cell lines were very well behaved and gave even better correlation with pseudotyping results, after a short, overnight co-incubation. Results with the stable cell fusion assay also correlated well with those of a live virus assay. In summary we have established a rapid, reliable, and reproducible cell fusion assay that will serve to complement the other neutralization assays currently in use, is easy to implement in most laboratories, and may serve as the basis for high throughput screens to identify inhibitors of SARS-CoV-2 virus-cell binding and entry.

SARS-CoV-2 belongs to the genus Coronavirus, in the family Coronaviridae. It is an enveloped, non-segmented, positive-sense single-stranded RNA virus [16,17]. Genomic sequences of SARS-CoV-2 and SARS-CoV show high similarity, with amino acid sequence identity being >76% [18]. The genome of SARS-CoV-2 is nearly 30 kb in length, including many open reading frames (ORFs) which express at least 27 proteins [17,19]. Among them, the surface spike glycoprotein (S) plays a key role in viral entry into target cells [20][21][22]. The receptor-binding subunit S1 attaches to the host cell via the cellular receptor human Angiotensin-Converting Enzyme 2 (hACE2), triggering proteolytic activation of S and subsequent conformational change of the S2 subunit, which facilitates the fusion of viral and cellular membranes [23][24][25].
An essential element of developing any prophylactic or therapeutic antiviral or vaccine is quantitative measurement of viral replication. The current gold standards for SARS-CoV-2 neutralization include pseudotyping using S and a suitable virus core encoding a reporter [26][27][28][29], or inhibition of live virus replication in vitro [30,31] or in animal models [32,33]. Pseudotyping requires production of vector supernatants at BSL2 or BSL2+ biocontainment that are then cryostored until use, with assay readout on susceptible cells after a few days; live virus requires BSL3 laboratory and readout by plaque reduction or similar assay 3-5 days after cell or animal infection.
To circumvent some of these issues we developed a cell fusion assay that utilizes S-expressing and hACE2-expressing cells that when mixed together provides a rapid and quantitative readout, based upon human immunodeficiency virus type 1 (HIV) Tat transactivation of an integrated HIV long terminal repeat (LTR)-firefly luciferase gene. Sera from COVID-19 + patients and anti-spike monoclonal antibodies inhibited cell fusion, which correlated highly with pseudotyping and use of live virus results. Soluble receptor binding domain (RBD) of S also inhibited cell fusion to a lesser extent, as did soluble hACE2. This assay is rapid and can be easily modified for high-throughput format, which will facilitate vaccine development, potent monoclonal antibodies screening, and in vitro drug testing against the virus.

Development of a quantitative assay for the measurement of S-hACE2-mediated cell fusion in transiently transfected cells
Spike or S protein mediates cell entry into susceptible target cells expressing hACE2 [24,28,[34][35][36]. We first tested whether co-expression of S and hACE2 could result in cell fusion. 293T cells were transiently co-transfected by calcium phosphate co-precipitation with CMV expression plasmids encoding S and hACE2. At 48 h cells were fixed and stained with crystal violet, and there was obvious multinucleate cell formation that was dependent upon both S and hACE2 (S1 Fig). To test whether this occurred when cells were individually transfected, plasmids encoding S and hACE2 were introduced separately into 293Ts along with CMV-Tat and HIV LTR-FFLUC, respectively, and cells mixed at 48 h. The next day cells were fixed, and cell syncytia were observed with higher amounts of plasmid transfected, up to 1 μg (S2A- S2E Fig). In parallel, cells were lysed 24 h after mixing, and we observed a marked increase in RLU (S2F Fig).
Because enumeration of multinucleate cells is at best semi-quantitative, we decided to focus on development of a quantitative cell fusion assay, based upon HIV LTR activation by HIV Tat. As an initial experiment 293T cells were transiently co-transfected with a plasmid encoding either empty vector (EV), S driven by CMV promoter with or without protease TMPRSS2, or VSV-G along with HIV long terminal repeat driving firefly luciferase (HIV LTR-FFLUC). After 48 hours, these cells were incubated with target 293T cells that had been co-transfected with a plasmid encoding EV or hACE2 with or without TMPRSS2, along with HIV Tat, each driven by the CMV promoter. After another 48 h cells were lysed and FFLUC activity quantified by luminometry. We observed a~100-fold increase in RLU when both S and hACE2 were each separately transfected, consistent with cell fusion (S3 Fig). As expected, when VSV-G was introduced the increase in RLU occurred independent of expression of hACE2. Interestingly, co-transfection of TRMPRSS2, the protease thought to activate S for cell fusion [34,37], did not further increase RLU activity, when introduced along with either S or hACE2 in the producers or targets, respectively (S3 Fig). This may be because 293Ts already express this protease or presence of this specific protease is not required for cell fusion in these cells.

Studies with soluble hACE2 and Spike ectodomain
We wished to test the validity of the cell fusion assay by using putative biological inhibitors. To characterize the cell fusion assay, increasing concentrations of soluble hACE2 protein were pre-incubated with S-expressing cells for 1 h prior to mixing with hACE2-expressing cells. Soluble hACE2 was able to inhibit cell fusion at a relatively high concentration, with calculated IC 50 of >3μM (Fig 1A). Entry of pseudotyped particles was also inhibited by soluble hACE2, with an IC 50 of 350 nM (Fig 1B).
We performed similar experiments with purified ectodomain of S protein (Fig 1C and 1D). Soluble ectodomain of S was pre-incubated with S-expressing cells prior to mixing with hACE2 expressing cells, with cell fusion measured the next day. Calculated IC 50 for soluble ectodomain of S was >679 nM (Fig 1C), whereas for pseudotyped particles IC 50 was~34 nM ( Fig 1D). This suggests that, compared to pseudotyping, cell fusion is more difficult to inhibit by at least 10-to 20-fold, using either of these biomolecules.

Serum from COVID-19+ patients inhibit cell fusion and syncytia formation
A convenience sampling of convalescent and acute illness sera from 12 COVID-19+ subjects, some with acute infection and some during recovery phase, were also tested in the cell fusion assay. Demographic and clinical characteristics of the subjects are shown in S1 Table. With a single exception, all tested sera inhibited cell fusion at varying titers (Fig 2A-2L). Setting aside outlier 027, IC 50 titers for the cell fusion assay varied between 13.75 and 353.5 ( Fig 2M). Serum was also tested in the pseudotyping assay (Fig 2N), and there was a high degree of correlation between cell fusion and pseudotyping IC 50 titers (Fig 2O). Similar to what we had observed with soluble hACE2, inhibition of cell fusion required more sera than inhibition of

PLOS PATHOGENS
Cell fusion assay to quantify SARS-Cov-2 spike interaction with hACE2 pseudotyping. This suggests that, depending on the biologic or drug being tested, compared to pseudotyping the cell fusion assay may be a more rigorous test of viral inhibition.

Development of stable cell lines to quantify S-hACE2 cell fusion
The data presented above demonstrate quantitation of cell fusion in transiently transfected cells. Because transient transfection is relatively complicated, occasionally unreliable, and not amenable to high throughput use, we decided to develop stable cell lines that could quantify cell fusion rapidly, reliably, and reproducibly. First, cell lines stably expressing S (TZMbl-Spike) or hACE2 (HOS-3734 or 3742) were generated, as described in Materials and Methods. TZMbls, based upon HeLa cells, have integrated HIV LTR-FFLUC and HIV LTR-LacZ cassettes that are both Tat-responsive; they have been widely used in the HIV field for titering virus stocks and performing pseudotyping assays, especially to measure neutralization by sera and cloned antibodies. The HOS cells used here had been transduced with a third-generation HIV vector in which tat remains intact. Importantly

Further characterization of the stable cell line cell fusion system
Anti-spike monoclonal antibodies, soluble hACE2 and spike RBD, and convalescent/acute illness human sera were tested in the stable cell line cell fusion system. Both commercial monoclonal antibody against Spike RBD mFc protein ( Fig 3A) and clone CR3022 ( Fig 3C) were able to inhibit cell fusion in the stable system, and the IC 50 values were 2.1 and 16.73 μg/mL, respectively, whereas in the pseudotyping assay IC 50 values were 15-20 fold less at 0.14 and 0.70 μg/ mL, respectively (Fig 3B and 3D). Peptide LCB1 was reported to efficiently neutralize pseudotyped Sars-Cov-2 virus entry [38]. When we tested its ability to inhibit cell fusion the IC 50 was 8.2 nM (Fig 3E), which is 12-fold greater than the IC 50 value for inhibiting pseudotyping (0.66 nM; Fig 3F). Furthermore, a time of addition experiment demonstrated that LCB1 could efficiently inhibit cell fusion at -1 and 0 h before target-producer cell co-incubation (S7 Fig). Soluble hACE2 inhibited cell fusion at higher concentrations, with an IC 50 of 1.39 μM (Fig 3G), whereas soluble spike RBD was virtually inactive (IC 50 >19.5 μM, Fig 3H).

PLOS PATHOGENS
Cell fusion assay to quantify SARS-Cov-2 spike interaction with hACE2 We randomly picked another 5 acute and convalescent sera such that 17 were tested using stable cell line system (HOS-3734). Clinical and demographic characteristics of these subjects are included in S1 Table. Again, 13 out of the 17 convalescent sera were able to inhibit cell fusion, with the exception of subjects 027, 247, 277, and 306 ( Furthermore, a time of addition experiment demonstrated that convalescent sera 045 could efficiently inhibit cell fusion at -1 and 0 h before target-producer cell co-incubation (Fig 6A  and 6C), and convalescent sera 054 could efficiently inhibit cell fusion at -1 h (Fig 6B and 6C). Beyond 2 h of co-incubation inhibition of cell fusion by sera 045 or 054 was minimal to nonexistent ( Fig 6C). As shown in Fig 6D, data from the cell fusion assay is largely consistent with that of pseudotyping, although it appears that even at time of addition = 2 h of both sera there is some degree of inhibition of pseudotyping.
In order to test the reproducibility of the cell fusion assay, LCB1 peptide and murine mAb against Spike protein (details of this mAb to be published separately) were each tested 5 times in 5 independent experiments performed at separate, discrete times over a 6 month period, with respective, aggregate IC 50 s of 0.096±0.04 μg/mL and 5.52±1.64 nM (see S11 Fig for data for LCB1 peptide). Thus, the data suggest that the stable cell lines are reliably well-behaved and may be a suitable platform for high-throughput screening of sera, monoclonal antibodies,biologics, and small molecules that inhibit SARS-CoV-2 virus entry into cells.

Discussion
Current prophylactic and therapeutic efforts to stem the global COVID-19 pandemic include vaccines, biologics, and small molecules [2,15,[39][40][41][42]. Virtually all of the vaccines target SARS-CoV-2 spike-hACE2 interaction, as do many of the other agents, in order to prevent virus binding and entry into target, receptor-bearing cells [15,43,44]. Current 'gold standards' to quantify the efficacy of such measures include inhibition of target cell infection by pseudotyped particles and fully replication-competent virus [6,28]. Both of these methods require consistent and reliable production and cryostorage of viral particles, use under BSL2+ or BSL3 biocontainment, and readout is typically performed several days later (although there are exceptions; see [45]). Herein we borrowed a page from the HIV playbook and report the development of stable cell lines in which quantification of cell fusion is available overnight after coincubating the cells.

PLOS PATHOGENS
Cell fusion assay to quantify SARS-Cov-2 spike interaction with hACE2

PLOS PATHOGENS
Cell fusion assay to quantify SARS-Cov-2 spike interaction with hACE2 We were impressed with the degree of syncytia formation after co-expression of S and hACE2 (S1 Fig), which also occurred when the protein products were separately expressed and the cells subsequently mixed (S2 Fig). Because counting multinucleate giant cells is at best semi-quantitative, we next used a reporter assay after transient transfection that relies upon HIV Tat trans-activating the HIV LTR [46]. Although transient transfection did give somewhat reliable and rapid results which correlated with S pseudotyping using HIV cores (Figs 1  and 2), it required repeated, fresh cell transfections of multiple plasmids, which is inconvenient, variable, and may not be reproducible if widely employed. The stable cell lines that we have developed, on the other hand, are facile to use and very reliable. They behave very consistently over many months, even in the absence of antibiotic selection. The actual RLU values obtained are quite high for cells plated in 96-well format and the dynamic range coupled with low standard deviations, based upon inhibition by sera and monoclonal antibodies tested, should be sufficient for a high throughput screen to identify inhibitors of S and hACE2 interaction.
At present it is unknown precisely how SARS-CoV-2 spreads and replicates in human tissues, including the respiratory tract [47]. Whether cell-free versus cell-to-cell transmission of

PLOS PATHOGENS
Cell fusion assay to quantify SARS-Cov-2 spike interaction with hACE2 this virus occurs in vivo is an open question; generally speaking cell-to-cell transmission of virus via a virological synapse is much more efficient than infection by cell free virus [48][49][50][51][52][53]. It is conceivable that SARS-CoV-2 causes cell fusion between S and hACE2-expressing cells in man [28,54]. There is increasing histopathological evidence of giant, multinucleate respiratory epithelial and intra-alveolar cells in COVID-19+ patients [55][56][57]. Although some of those cells had features suggestive of being virally-infected, definitive evidence is lacking. It should be made clear that our results here are not meant to address or answer that interesting question, but rather to demonstrate that a quantitative assay based upon cell fusion has been established for spike-hACE2 interaction.
That the cell fusion and pseudotyping IC 50 values for the convalescent and acute illness COVID-19+ sera (Figs 2 and 4 and 5 and S10) correlated highly suggests that the cell fusion assay is indeed measuring the ability of S and hACE2 to interact. Unsurprisingly, there was also significant correlation between the cell fusion, pseudotyping, and live virus assays in terms of IC 50 titers, although certainly with additional serum testing discrepancies may arise which might be worth scientific pursuit. It is also not surprising that more sera and antibody is needed to inhibit the cells from fusing. Although we have not attempted to quantify the numbers of spike and hACE2 proteins on the stable cell lines, typical expression levels would

PLOS PATHOGENS
be >10,000 molecules of each, whereas an intact SARS-CoV-2 virion may have just a few dozen spike trimers [58][59][60], depending upon its source and how it is made. Nor is it known how many S-hACE2 interactions are required to trigger irreversible pore formation and subsequent cell-cell or cell-virus fusion. Although we do not know the relative cell surface density of either S or receptor on either stable cell line, based upon stoichiometry alone it is quite likely that inhibition of cell-cell fusion is much more difficult than inhibition of virus-cell fusion, whether the latter is due to pseudotyped or replication-competent virus. Here, this is borne out in the much higher amounts of sera and antibody required to achieve 50% inhibition of fusion (Figs 2-5). More importantly, however, the IC 50 values for both assays are highly correlated (Figs 2 and 5 and S10), suggesting that the cell fusion assay has utility and in fact may be a more rigorous test for the inhibitory power of any antibody, serum, biologic, or even small molecule, when compared to other cell-based assays that rely on production of virus.
In addition to sera and monoclonal antibodies, we tested two other biologics-purified, soluble hACE2 and RBD of spike (Figs 1 and 3). Soluble hACE2, stabilized by an Fc domain for a longer half-life in plasma, has been proposed as a potential therapeutic agent, especially since the S-hACE2 interaction has a K D in the low nanomolar range [61,62]. hACE2 has now been subjected to saturation mutagenesis of the RBD-interacting residues, and several hACE2 variants have been identified that have even greater affinity for RBD (K D in the picomolar range) [63]. Here, we tested wt, soluble hACE2 (Figs 1A and 3G), and although it inhibited cell fusion the amounts required to do so were relatively high. Whether soluble or stabilized hACE2 variants are more potent against cell fusion will require further testing. At the moment soluble RBD and variants thereof have only been proposed and tested as potential vaccine candidates [64], not as therapeutics, so RBD testing here was purely academic. As anticipated, RBD did interfere with both pseudotyping and cell fusion, although for the latter the degree of inhibition was marginal at best. The fact that both soluble hACE2 and RBD inhibited cell fusion further corroborates the validity of the assay. A third biologic, peptide LCB1, also had low nM IC 50 inhibitory activity in the cell fusion assay, as would be predicted based upon pseudotyping results.
The cell fusion assay was highly reproducible in that repeated testing of both an inhibitory peptide and a murine mAb gave extremely consistent results over a period of many months, with at most a 2-fold variance in IC 50 values. This suggests that the cell fusion assay, when performed by other investigators throughout the world, should also behave similarly, since all that needs to be done is co-culture of two stable cell lines, without the need for making, storing, or testing any virus, live or pseudotyped. This will allow straightforward comparisons of the potency of various virus entry inhibitors between disparate laboratories.
In summary, we have developed a novel SARS-CoV-2 spike-hACE2 cell fusion assay that is rapid, reliable, and reproducible. Other than two stable, well-behaved cell lines it requires no specialized research reagents or laboratory equipment and should be easy to adapt for use in most investigative and clinical settings. It will allow for the testing of sera after vaccination or infection, to assess for level of immune protection, and it could be used for high throughput screening for compounds and biologics that interfere with virus-cell binding and entry.

Ethics statement
COVID-19+ convalescent serum was obtained from YNHH hospital (IMPACT research team). IMPACT study used the Yale university institutional review board (IRB) or Biomedical (HIC). And the approval was granted by the Yale HIC. All informed, written consent was obtained from all subjects.

Cell lines
Human embryonic kidney cell line 293T (#CRL-3216), human bone cell line HOS (#CRL-1543) were originally purchased from ATCC. Africa green monkey kidney cell line VeroE6 (#CRL-1586) were purchased from ATCC. TZMbl cells (#JC53BL-13) were obtained from the NIH AIDS Reagent Program. The HOS cells were stably transduced with a third generation HIV vector encoding tat, along with eGFP, mRFP, and bleomycin resistance gene; they were maintained in 200-400 μg/mL phleomycin (Invivogen) and were eGFP and mRFP-positive by flow cytometry. hACE2 was subsequently introduced by VSV G-mediated HIV-based transduction using pLV-EF1a-hACE2-cMYC-FLAG-IRES-Puro and pHIV-CMV-hACE2-IRE-S-Puro, respectively, to produce HOS-3734 and HOS-3742 cells, both cell lines maintained in selection using 10 μg/mL puromycin (Sigma-Aldrich). Those two vectors were also introduced into 293T cells to produce 293T-hACE2 cells for use with pseudotyped particles. Control HOS-2072 cells were created by transducing them with the empty vector HIV-CMV-IRESpuro and maintaining them in 10 μg/mL puromycin. TZMbl cells stably expressing S were created by co-transfecting TZMbl cells with pT-PB-SARS-CoV-2-Spike-IRES-Blasti along with pCMV-piggybac and resistant cells selected with 10 μg/mL blasticidin (Invivogen). The control stable cell line not expressing S was generated by co-transfecting pCMV-piggybac with pT-pB-IRES-Blasti and selecting for blasticidin-resistant cells.
Spike RBD domain plasmid was a gift of David Veesler (University of Washington) and the human ACE2 ectodomain plasmid was a gift of Jason McLellan (University of Texas at Austin). Both constructs have histidine tags to allow Ni column purification. Plasmids were transfected into Expi293F cells (ThermoFisher) using the manufacturer's protocol. Culture supernatant containing the secreted protein was harvested after 3-4 days, and dialyzed against its Ni-NTA binding buffer. Protein was then purified through Ni-NTA affinity chromatography (Qiagen), and size exclusion chromatography (SEC) using a Superdex 200 column (GE Life Sciences) equilibrated with its SEC buffer (listed below). SDS-PAGE was used to monitor purification steps and ensure protein homogeneity. Peak fractions were concentrated, flash frozen and stored at -80˚C for future use.

Cell fusion assay
For the transient transfection cell fusion assay producer 293T cells were co-transfected with pSV-TAT or pCMV-Tat and pcDNA-SARS-CoV-2-S, while target 293T cells were co-transfected with pLTR-LUC and pShuttle-hACE2 or hACE2 plasmids. At 48 h transfected cells were lifted, mixed 1:1, and after another 16-24 h cells were lysed and RLU measured by plate reader in 96-well format as described. Images of cell syncytia were captured with a Nikon TE2000 epifluorescence microscope running MetaMorph software. For the stable cell fusion assay HOS cell lines stably expressing HIV Tat and hACE2 (termed HOS-3734 and HOS-3742) were mixed 1:1 with TZMbl cells stably expressing S. After 16-24 h FFLUC activity was measured and syncytia images captured as described above. To observe LacZ activity, after cell fixation X-gal substrate was used as described. All experiments were performed with biological duplicates and repeated at least twice.

Cell fusion inhibition by serum or antibodies
Producer and target cells were generated as described above. With regards to the transient system, producer cells were lifted 48 h post transfection and 10 4 cells were resuspended in 100 μL medium per well in 96-well plates. Serial dilutions of antibody or serum were then added to producer cells and incubated at 37˚C for 1 hour. At that point 10 4 target cells (50 μL per well) that had been transfected 48 h prior were added to producer cells, and after another 24 h cells were lysed in 0.1 mL and RLU measured. With regards to the stable cell lines, 10 4 producer cells (TZMbl-Spike) in 100 μL of medium in the absence of blasticidin were seeded in 96 well plates. After 24 h, 70 μL of four-fold serially diluted antibody or serum was added into producer cells and incubated at 37˚C for 1 hour. Serum and antibody concentrations were the same as above. At that time 10 4 target cells (HOS-3734 or HOS-3742) in 50 μL medium were then added to the producer cells, and after another 24 h cells were lysed in 0.1 mL and RLU measured. Data were analyzed with non-linear regression using GraphPad Prism to determine the neutralization curve and the IC 50 values calculated.

Inhibition of cell fusion with soluble hACE2, Spike RBD and peptide LCB1
Seventy μL of four-fold serially diluted, purified soluble hACE2, spike RBD or peptide LCB1 were added to 96-well plates, which were seeded with 10 4 producer or target cells in 100 μL per well, respectively. hACE2 dilutions began at 10 μM, spike RBD at 43 μM, and peptide LCB1 at 9.176 μM. After 1 h 0.5×10 4 target or producer cells in 50 μL were added per well. After 16-24 h 100 μL of lysis buffer was added to each well and RLU measured. The assay was performed for both transient and stable cell lines.

Pseudotyped virus neutralization assay
Pseudotyped HIV-FFLUC was produced as previously described [65] but using pcDNA-SARS-CoV-2-S instead of VSV G or HIV Env plasmid. If necessary, pseudotyped particles were concentrated by ultracentrifugation. Serum from clinical samples, antibodies, or soluble proteins (hACE2 or Spike RBD) were serially diluted as indicated and pre-incubated with SARS-CoV-2 pseudotyped particles for 1 h at 37˚C before inoculation onto 293T-hACE2 target cells. After an overnight incubation fresh medium was added. After another 48 h cells were lysed and RLU measured. IC 50 values of sera, antibodies or soluble proteins were calculated using GraphPad Prism software. Non-linear regression with normalized response model was applied.

Replication-competent SARS-CoV-2 Nano luciferase neutralization assay
Experiments using infectious SARS-CoV-2 were performed in a Biosafety Level 3 facility, licensed by the State of Connecticut and Yale University. Nano luciferase expressing SARS-CoV-2 infectious clone ("icSARS-CoV-2-nLuc-GFP") was previously described and generously provided by Ralph Baric (UNC) [30]. P3 viral stock was generated in VeroE6 cells (cultured in DMEM containing 5%FBS, 1% sodium pyruvate, and 1% Penicillin-Streptomycin) by infecting at a MOI 0.01 for two-three days to generate a working stock. After incubation the supernatant was clarified by centrifugation (500g × 5min) and filtered through a 0.45-micron filter, and virus titer was determined by plaque assay on VeroE6 cells as previously described [66,67]. VeroE6 cells were plated at 3000 cells/well in 384 well clear bottom black cell culture plate (Greiner Bio-One). Twenty-four h post seeding, human sera, which were heat-inactivated at 56˚C for 30 min, were diluted 1:20 (starting dilution), then serially diluted 2-fold for 8 dilutions in DMEM (2% FBS, 1% sodium pyruvate, and 1% Penicillin-Streptomycin). 50 μL dilutions were added to 50 μL of icSARS-CoV-2-nLuc-GFP (MOI 0.01,~2PFU/ μL), and incubated for 1 hour at 37˚C. Media was then removed from cells and replaced with 20 μL of virus-sera mixture, and incubated for 48 h at 37˚C. 5 μL of Nano-Glo Luciferase substrate (Promega) was added to each well and luciferase signal was measured using Cytation 5 plate reader (BioTek). Values of serum-treated samples were normalized to non-serum controls. Assay was performed in triplicate and averages of these normalized values were plotted in Prism 9 (GraphPad). IC 50 values were calculated.

Western blotting
Expression of SARS-CoV-2 spike and hACE2 proteins in cells were verified by immunoblotting. In the transient system, cells transfected with plasmids encoding Spike or hACE2 were lysed with RIPA buffer 48 h post transfection. Stable cell lines expressing spike or hACE2 were similarly lysed. Samples were boiled for 10 min in the presence of SDS and DTT and size-separated on a pre-made SDS-PAGE gradient gel (Bio-Rad), which was then transferred onto PVDF filter membranes as described [65]. hACE2 and Spike proteins were detected by goat anti-hACE2 and anti-FLAG primary and rabbit anti-goat-HRP and rabbit anti-mouse-HRP secondary antibodies, respectively.