key: cord-305496-t8ykkekl
authors: Stone, E. Taylor; Geerling, Elizabeth; Steffen, Tara L.; Hassert, Mariah; Dickson, Alexandria; Spencer, Jacqueline F.; Toth, Karoly; DiPaolo, Richard J.; Brien, James D.; Pinto, Amelia K.
title: Characterization of cells susceptible to SARS-COV-2 and methods for detection of neutralizing antibody by focus forming assay
date: 2020-08-21
journal: bioRxiv
DOI: 10.1101/2020.08.20.259838
sha: 
doc_id: 305496
cord_uid: t8ykkekl

The SARS-CoV-2 outbreak and subsequent COVID-19 pandemic have highlighted the urgent need to determine what cells are susceptible to infection and for assays to detect and quantify SARS-CoV-2. Furthermore, the ongoing efforts for vaccine development have necessitated the development of rapid, high-throughput methods of quantifying infectious SARS-CoV-2, as well as the ability to screen human polyclonal sera samples for neutralizing antibodies against SARS-CoV-2. To this end, our lab has adapted focus forming assays for SARS-CoV-2 using Vero CCL-81 cells, referred to in this text as Vero WHO. Using the focus forming assay as the basis for screening cell susceptibility and to develop a focus reduction neutralization test. We have shown that this assay is a sensitive tool for determining SARS-CoV-2 neutralizing antibody titer in human, non-human primate, and mouse polyclonal sera following SARS-CoV-2 exposure. Additionally, we describe the viral growth kinetics of SARS-CoV-2 in a variety of different immortalized cell lines and demonstrate via human ACE2 and viral spike protein expression that these cell lines can support viral entry and replication.

which human cell types may be more permissive to SARS-CoV-2 infection, with particular 72 uncertainty including if higher ACE2 expression coincides with increased risk for heightened 73 infection. Furthermore, it is unclear which cell types may be suitable for the successful evaluation 74 of antiviral and therapeutic drugs for in vitro screening before in vivo evaluation. Further 75 characterization of permissible cell types is necessary to improve our understanding of SARS-76

CoV-2 pathogenesis and to develop therapeutic strategies for the treatment of It has also been noted that following infection with SARS-CoV-2, people typically 78 seroconvert 20 days following symptom onset, and there is increasing evidence suggesting that 79 development of a neutralizing antibody response is a correlate of protection in patients recovered 80

from . The COVID-19 pandemic has highlighted the necessity for testing for 81 neutralizing antibodies. As SARS-CoV-2 infection can have a range of manifestations from 82 asymptomatic to fatal multiple organ failure [23] [24] [25] [26] [27] , antibody testing and serological surveys are 83 a critical tool for determining prior infection status and seroprevalence in a population . It is also 84 the goal of many candidate SARS-CoV-2 vaccines to induce neutralizing antibodies targeting the 85 viral spike protein, the major antigenic determinant of SARS-CoV-2 [28] . 86

For understanding the antibody response, assays that measure neutralizing antibody titer 87 are considered the gold standard. One such tool for evaluating neutralizing antibody response is a 88 plaque/focus neutralization reduction test (PRNT/FRNT), which evaluates the ability of polyclonal 89 sera samples to prevent or reduce infection of a cell monolayer in vitro. Previously, for SARS-90

CoV-2, only PRNT assays-which rely on the ability of virus to lyse infected cells and thus can 91 take 48-96 hours to develop-have been used in the assessment of the neutralizing antibody 92 response. It is not well known whether an FRNT-which uses an immunostaining protocol to 93 detect virus and does not depend on cell lysis, and thus is often more rapid-is amenable for 94 detection of SARS-CoV-2 neutralizing antibody. 95

In this study, we describe the growth kinetics of SARS-CoV-2 in multiple cell types and 96 the methods our laboratory has used to optimize a SARS-CoV-2 focus forming assay (FFA) to 97 improve sensitivity and specific detection. By characterizing the growth kinetics of SARS-CoV-2 98 on a variety of immortalized and primary cell lines, we have demonstrated which of these cell lines 99 is susceptible to infection by SARS-CoV-2.We also demonstrate that the FFA can be adapted to 100 measure the neutralization capacity of polyclonal sera in an FRNT. This high throughput FRNT 101 assay can be applied to sera from both animal models of SARS-CoV-2 infection, as well as human 102 SARS-CoV-2 infected patients, and can serve as a useful assay for describing the kinetics of the 103 neutralizing antibody response to SARS-CoV-2. Additionally, we have compared the expression 104 of ACE2 and SARS-CoV-2 spike on these cell lines to determine how spike expression correlates 105 with susceptibility. The tools developed in this study have practical applications in both the basic 106 science and translational approaches that will be critical in the ongoing effort to slow the COVID-107 19 pandemic. Based on our previous work optimizing the FFA for WNV, [29] we proceeded to adapt the 113 FFA for the detection of infectious SARS-CoV-2. Along with spot number, the spot size and 114 border definition provide valuable information on possible differences in viral strains. As we have 115 observed that the foci morphology, as well as spot number, can vary dramatically under different 116 growth conditions, we sought to test different growth conditions and cell lines to determine the 117 optimal conditions for SARS-CoV-2 viral titration. This goal was guided by previous studies that 118 have suggested that the use of Vero cells from varying origins can impact viral titer [30, 31] . Using 119 both Vero CCL-81 (ATCC® CCL-81™, referred to in this text as Vero WHO) and Vero E6 (Vero 120 1008, ATCC® CRL-1586™) cell lines, we determined if differences in foci number or size 121 occurred to decide if one cell line was superior for titration by FFA (Figure 1A-1C) . Although 122 many laboratories utilize Vero E6 cells for viral titer measurements of SARS-CoV-2 [32, 33] , in 123 our laboratory, Vero E6 cells typically resulted in about two-fold lower foci formation relative to Vero WHO cells (Figure 1A-C) . Figure 1A is a representative image of an FFA showing the 125 viral titration on both the Vero E6 and WHOs for both ~50 FFU and ~200 FFU when identical 126 numbers of cells are seeded per well. We noted that at identical higher dilutions of SARS-CoV-2 127 virus stocks, Vero WHO cells develop ~55 individual foci per well, whereas Vero E6 cells develop 128 ~27 foci per well ( Figure 1A) . The same pattern was observed at lower dilutions of virus, with 129 ~200 foci formation on Vero WHO cells yielding only ~100 foci on Vero E6 cells ( Figure 1A ) 130

The quantification of this difference in the Vero WHO and Vero E6 cell type is shown in Figure  131 1B. Interestingly, when we compared this observation to the genome copy number by 132 we noted that Vero E6 cells tended to produce significantly more virus across all 24, 48, and 72-133 hour timepoints (p = 0.0064) ( Figure 1C) . It is possible that the discrepancy between the FFA and 134

qRT-PCR data could be due to Vero WHO cells producing fewer genome copies yet more 135 infectious virus than E6 cells, while E6 cells sustain more viral replication but yield less infectious 136 virus. While we did not see any differences in foci morphology between the two Vero cell lines 137 we used, the significant difference in the number of foci observed between Vero E6 and Vero 138 WHO (p < 0.0001) suggests that Vero WHO cells record higher viral titers than the Vero E6 cells 139

for SARS-CoV-2 ( Figure 1B) . 140 141

In order for SARS-CoV-2 to form distinct foci, it is critical to plate the optimal cell density. 143

We examined the impact of cell density on foci formation for both Vero WHO and Vero E6 cells 144 by plating identical dilutions of SARS-CoV-2 virus stocks on 96-well plates seeded with differing 145 numbers of WHO or E6 cells (3 × 104, 1.5 × 104 or 3 × 104 cells/well) one day prior to infection 146 of the cell monolayer. At these concentrations, on the day of infection, 3×104, 1.5×104 and 3×104 147 cells/well resulted in monolayers that were 70, 80 and 90 percent confluent respectively for both 148 cell lines tested. Figure 1D is a representative image of the focus forming assay showing the foci 149 formations arising from different cell concentrations plated for both the Vero E6 and WHOs. The 150 quantification of the spot counts for Vero WHO cells is shown in Figure 1E . The quantification 151 of the spot counts for Vero E6 cells is shown in Figure 1F . For both the Vero WHOs and E6 cells, 152

we observed a significant increase in the number of foci formed when either 1.5 × 104 (E6 p = 153 0.0480, WHO p = 0.0094) or 3 × 104 (E6 p = 0.0057, WHO p = 0.0024) were seeded per well 154 compared to 3 × 103. However, there was no significant increase in foci formation when increasing the cell density from 1.5 × 104 cells/well to 3 × 104 cells/well. Thus, while the Vero E6 cells 156 fostered lower foci numbers at each confluency as compared to the WHO cells, viral titers were 157 not significantly different at 1.5 × 104 cells/well or 3 × 104 cells/well irrespective of whether the 158 Vero WHO (Figure 1E ) or Vero E6 ( Figure 1F ) cell line was seeded for the assay. We have 159 previously tested higher cell densities for FFAs and have noted that cell concentrations higher than 160 3 × 104 cells/well results in an overly confluent monolayer with more cells than can adhere to the 161 wells, leading to highly variable titer information (data not shown). The results of these studies 162 suggest that Vero WHO cells plated at either 1.5 × 104 or 3 × 104 cells/well was optimal for the 163 viral FFA. 164 165

Like plaque assays, the incubation time for the FFA is highly dependent on the viral 167 replication cycle within the cells and the time required for infectious progeny to be released and 168 spread to neighboring cells. Whileunlike traditional plaque assaysthe FFA is not dependent 169 on viral lysis of infected cells, the development of visible spots is dependent on the time it takes 170 for viral protein production to occur and for infectious virus to spread to neighboring susceptible 171 cells. To determine the optimal time frame for infection of SARS-CoV-2 on a Vero WHO cell 172 monolayer to form individual foci, we tested a variety of incubation times. In order to optimize 173 these conditions, identical dilutions of SARS-CoV-2 virus stocks were added to infect Vero WHO 174 cells seeded at a density of 1.5 × 104 cells/well, and incubated for 20, 24, 48, or 72 hours post 175 infection. Figure 1G shows representative images of identical dilutions of SARS-CoV-2 FFAs 176 developed after 24, 48, or 72 hour incubation times, respectively. The mean spot size of foci at 177 each timepoint is quantified in Figure 1H . The mean spot count per well at each time point is 178 quantified in Figure 1I . 179 Altering the incubation time had the most dramatic impact on mean foci size amid all other 180 parameters tested. While there were no significant differences in the size of foci formed between 181 20 hours post infection (HPI) and 24 HPI (p = 0.0632), we did observe the formation of 182 significantly larger foci between the 24 and 48 HPI timepoints (p = 0.0031) , as well as the 48 and 183 72 HPI timepoints (p = 0.0134) ( Figure 1H) . Interestingly, at the same virus dilution, we found 184 that there were fewer spots between 24 HPI (mean spot number of 10) relative to the 48 HPI time 185 point (p = 0.0369, mean spot number of 13.5) but not the 72 HPI time point (p = 0.1895, mean spot value 12.5). Similarly, we found that there were no significant differences in the number of 187 foci formed between 48 and 72 HPI (p = 0.4198) ( Figure 1I ). However, we noted that this 188 difference is within one standard deviation of the mean number of spots across all wells and was 189 insignificant when determining titers. From this result we concluded that incubation times greater 190 than 24 hours resulted in a slight but significant increase in spot number, while assays incubated 191 for up to 72 did not alter the spot number but did increase the spot size. This increase in spot size 192 but not number between 48 and 72 hours is a highly useful for the testing of anti-viral compounds 193 which may require longer incubation times. In addition, larger spot size makes this assay more 194 universally useful since laboratories without an automated machine can manually count spots, 195

where the large size will improve readability of the assay. 196

The FFA relies on an immunostaining protocol of an infected cell monolayer in order to 197 quantify infectious virus titer and is therefore dependent upon SARS-CoV-2-specific antibody 198

binding. For this purpose, polyclonal guinea pig sera (BEI: NR-0361) raised against SARS-CoV-199 2 produces reproducible staining with minimal background. However, we have also used human As efforts to understand replication and transmission of SARS-CoV-2 are underway and 208 vaccine development moves forward, more information regarding permissive cell types for SARS-209

CoV-2 infection and replication is needed. To determine the permissivity of several different cell 210 types to infection with SARS-CoV-2, we generated multistep growth curves for human, non-211 human primate, murine, hamster, and gastric adenocarcinoma cell lines. Each of these cell lines 212 was infected at a MOI = 0.05 and cells or cell supernatants were collected aseptically, and total 213 RNA was isolated. Cellular RNA was normalized to an internal RNaseP control for human and 214 non-human primate cells and GAPDH for murine cells. Genome equivalents were determined 215 using the Applied Biosystems TaqMan gene expression assay protocol for SARS-CoV-2 216 previously described [34] . 217

To identify susceptible cell lines, we first assessed genome copy number in non-human 218 primate African green monkey kidney epithelial cells (Vero WHO, Vero E6) as well as human 219 hepatocytes (Huh7.5, Huh7) and lung epithelial cells (A549, CALU-3). Figure 2A shows the 220 SARS-CoV-2 genome copy number for whole cells for human and non-human primate cell lines. 221 Figure 2B shows the SARS-CoV-2 genome copy number for cell culture supernatants for human 222 and non-human primate cell lines. In each cell line aside from Vero WHO and Huh7, genome 223 equivalents within the cell peaked at 24 HPI and remained relatively constant for the duration of 224 the experiment. SARS-CoV-2 genome equivalents in Vero WHO and Huh7 cells peaked at 48 HPI 225 and remained relatively constant for the duration of the experiment. The Vero E6 cell line reached 226 the highest titer at 6.44 × 108 copies/µL at 24 HPI, with the Huh7.5 and CALU-3 cell lines reaching 227 1.1 × 106 copies/µL and 5.0 × 105 copies/µL, at the same time point, respectively. Vero WHO and 228

Huh7 cell lines reached the highest titer, 1.3 × 106 copies/µL and 1.1 × 106 copies/µL, 229 respectively, at 48 HPI. The A549 cells, although susceptible to infection, appeared to support 230 little SARS-CoV-2 replication, reaching only 80 copies/µL at 24 HPI. 231

We also measured genome copy number in the cell culture supernatant for all of the cell 232 lines described ( Figure 2B ) For each cell line, viral RNA in the supernatant peaked at later 233 timepoints, either 72 HPI (Vero WHO, 4.7× 104 copies/µL; Vero E6, 1.4× 104 copies/µL; CALU-234 3, 1.3× 104 copies/µL) or 96 HPI (A549, 2.6 copies/µL; Huh7, 27 copies/µL; Huh7.5, 8.6 × 103 235 copies/µL). Of these cell lines, the Vero WHO cells had the highest titers in the supernatant, while 236

Vero E6, Huh7.5 and CALU-3 cells were comparable in terms of titer. As expected, A549 cells 237 that did not contain high titers of cell-associated virus also did not contain high titers in the 238 supernatant. Interestingly, while Huh7 cells support relatively high titers of cell-associated virus, 239 they do not appear to yield high titers in the supernatant. These results suggest that Vero E6 cells 240 are most permissible for SARS-CoV-2 replication among all tested cell types and would be the 241 ideal choice for propagation. Vero WHO, Huh7.5, and CALU-3 cells are also permissible cell 242 types for SARS-CoV-2 infection and replication, however A549 cells do not appear to be suitable 243 for high levels of SARS-CoV-2 replication. Our results also suggest that Huh7 cells appear to be 244 permissible for SARS-CoV-2 infection and replication, but do not appear to be suitable for egress 245 into the cell culture supernatant. 246

Due to ongoing SARS-CoV-2 vaccine development efforts, there is an urgent need to 247 develop and evaluate the susceptibility of small animal models to SARS-CoV-2 infection and COVID-19. Recent studies have suggested that rodents may be used for these purposes, as well as 249

to study the adaptive immune response to SARS-CoV-2 infection [34] [35] [36] [37] . To this end, we next 250 sought to determine permissivity of rodent cell lines to SARS-CoV-2 infection, namely 3T3 and 251 SHHC17 cell lines. Figure 2C shows the SARS-CoV-2 genome copy number for whole cells for 252 murine and hamster cell lines, with Vero WHO cells included for comparison. Figure 2D shows 253 the SARS-CoV-2 genome copy numbers for cell culture supernatants derived from murine and 254 hamster cell lines, with Vero WHO supernatant included for comparison. From total cellular RNA, 255

we detected only 42 copies/µL in 3T3 cells at 72 HPI, the time point at which the titer peaked. 256

Similarly, with SHHC17 cells, we detected only 50 copies/µL at 24 HPI, the time point at which 257 the titer peaked. In addition to total cellular RNA, we also examined supernatants from cell culture 258 and predictably found peak titers of only 3 copies/µL in 3T3 cells and just 33 copies/µL in 259 SHHC17 cells, both at 96 HPI. These results suggest that neither 3T3 nor SHHC17 cell lines are 260 suitable for supporting SARS-CoV-2 replication or egress without further experimental 261

Finally, because it is known that hACE2 is highly expressed by intestinal epithelial cells, 263

we sought to examine the permissivity of human gastric adenocarcinoma cell lines to SARS-CoV-264 2 infection [15] . For this purpose, we used AGS and MKN cell lines, and examined viral genome 265 copies associated with both total cellular RNA as well as the cell supernatant. Figure 2E shows 266 the SARS-CoV-2 genome copy numbers for whole cells from gastric adenocarcinoma lines, with 267

Vero WHO cells included for comparison. Figure 2F shows the SARS-CoV-2 genome copy 268 numbers for cell culture supernatants from gastric adenocarcinoma lines, with Vero WHO 269 supernatant included for comparison. We found that MKN cells yielded relatively high titers, with 270 cell-associated virus peaking at 1.0× 106 copies/µL at 24 HPI. Virus in the supernatant peaked at 271 5.0 × 104 copies/µL at 72 HPI. AGS cells yielded lower titers, with cell-associated virus peaking 272 at 4.6× 103 copies/µL at 24 HPI and virus in the supernatant peaking at 6.7 × 102 copies/µL 96 273 HPI. Interestingly, however, the titer in AGS cells appeared more variable compared to other time 274 points, increasing at 48 and 96 HPI and dropping at 24 and 72 HPI. These results suggest that these 275 gastric adenocarcinoma cell lines can support infection, replication and egress of SARS-CoV-2 as 276 well as, or in some cases better than, Vero cell lines. 277

Given that our studies conducted to quantify SARS-CoV-2 viral genome copies in 280 susceptible cell lines yielded results that highlighted highly permissive cell lines, like Vero WHO, 281 while also distinguishing less permissive cell lines, like A549, we sought to analyze spike and 282 hACE2 protein co-expression to determine if higher hACE2 expression correlated with higher 283 susceptibility to SARS-Cov-2 infection. 284

One facet of our understanding of the current SARS-CoV-2 outbreak that is rapidly 287 evolving is SARS-CoV-2 seroprevalence in the general population. At the same time, forming a 288 better understanding of and ability to assess the kinetics of the neutralizing antibody response to 289 SARS-CoV-2 could be essential in further vaccine and anti-viral development efforts. To this end, 290

we adapted the SARS-CoV-2 FFA for the quantification of neutralizing antibody (nAb) titers in 291 the form of an FRNT. This was accomplished by incubating serially diluted convalescent serum 292 from SARS-CoV-2 infected individuals with a known quantity of infectious SARS-CoV-2 (~60 293 FFU) and measuring foci formation. Infection was normalized to a PBS control to reflect the 294 percent neutralization of sera. 295

First, we sought to determine whether the FRNT could be used to detect a range of nAb 296 concentrations in human samples. Figure 4A shows the neutralization curves for human sera 297 samples, showing a decrease in virus neutralization as the serum is diluted out. These samples 298 were collected from 4 human subjects at the University of Puerto Rico following a positive qPCR 299 test for SARS-CoV-2. All subjects were in the convalescence period at the time of sample 300 collection. These assays were performed using deidentified residual sera samples. 301

Using the FRNT approach, we quantified the neutralizing antibody titer in the form of the 302 reciprocal serum dilution required to neutralize 50% of virus, or the FRNT50 value. This assay can 303 also be used to determine the FRNT90 value (i.e. required for 90% neutralization). The FRNT50 304

and FRNT90 values are reported in Figure 4B . The reciprocal serum dilutions required for 50% 305 neutralization for the human samples (HS_A, HS_C and HS_D) are 2.161, 3.183, and 2.002, 306

respectively. The reciprocal serum dilutions required for 90% neutralization for HS_A, HS_C and 307 HS_D are 1.377, 2.725, and 1.739, respectively. For HS_B, the reciprocal serum dilution required 308 to neutralize both 50% and 90% of the virus was below the lower limit of quantitation (LLOQ) for 309 the assay.

In order to increase confidence that these nAbs were the result of recent SARS-CoV-2 311 infection rather than cross-reactivity with the four circulating human common cold coronaviruses, 312

we performed an ELISA to examine binding of these sera samples to SARS-CoV-2 receptor-313 binding domain (RBD). Figure 4C shows the absorbance at 450 nm (A450) values indicating that 314 sera from these subjects contain antibodies that can bind specifically to the receptor binding 315 domain (RBD) of SARS-CoV-2. 316

Having confirmed that our assay can detect nAb to SARS-CoV-2 in human sera, we next 317 sought to demonstrate that this assay is applicable for numerous sera sources including non-human 318 primates and mice. To this end, we performed an FRNT with non-human primate (NHP) sera, 319 which consisted of pooled sera samples from a group of Rhesus macaques in the convalescent 320 phase following SARS-CoV-2 infection by multiple routes (BEI: NR-52401). Figure 4E shows 321 the neutralization curve for this pool of NHP sera. Low but detectable nAb titers were present in 322 this sample with an FRNT50 value of 2.161 and FRNT90 of 1.377, as depicted in Figure 4F . 323

Having shown that our assay can be utilized to quantify nAbs in NHP samples, we next 324 sought to demonstrate that this assay could also be used for quantifying nAbs in small animal 325 models such as mice. This also afforded us the opportunity to examine the nAb response both at 326 Having demonstrated that our FRNT assay can be used to quantify nAb titers for human 335 samples, non-human primates, and mice both in acute infection and memory responses, we next 336 sought to determine whether this assay could be used to quantify nAb resulting from a subunit or 337 DNA vaccine. To this end, we immunized C57BL/6 mice intramuscularly (i.m) with 50 µg of 338 DNA encoding the SARS-CoV-2 spike (MS_C). A subset of these mice was boosted 21 days later 339 (MS_B, MS_D) with 5 µg of DNA intramuscularly and sera collected 21 days following the boost. 340 Figure 4G shows the neutralization curves for these mice immunized with DNA encoding the immunization. From these data we were able to define the FRNT50 values for MS_B, MS_C, and 343 MS_D, which are shown in Figure 4H , and are 2.045, 1.584, 1.227, respectively. However, the 344 FRNT90 values were below the LLOQ for the assay, as well as the FRNT50 value for MS_A. These 345 results suggest that the FRNT can be used to detect nAbs in sera resulting from immunizations, in 346 addition to nAbs in human, NHP, and mouse sera resulting from SARS-CoV-2 infections. 347 348

To address the need for high-throughput, rapid quantification of infectious SARS-CoV-2, 350 our group has developed a focus forming assay (FFA) for SARS-CoV-2 using Vero WHO cells. 351

The strength of the FFA is the rapid visualization of individual foci forming from a single 352 infectious unit or focus forming unit (FFU). The FFA for SARS-CoV-2 can be developed in as 353 little as 24 hours, shorter relative to traditional plaque assays for human coronaviruses which can 354 take 2-5 days [32, 38, 39] . The focus forming assay is also amenable to a 96-well plate format, 355 allowing for assays to be scaled up or automated to handle large volumes of samples quickly 356 relative to assays requiring plates with 24 wells or fewer. Automating the quantification of foci 357 using equipment such as a CTL machine can also streamline the process of screening large 358 numbers of samples. One potential disadvantage of the focus forming assay is the requirement of 359 a SARS-CoV-2 specific antibody as the primary antibody for foci immunostaining. However, for 360 our assays we have found that polyclonal guinea pig serum provides reproducible staining with 361 minimal background when used at the appropriate concentrations, and numerous human 362 monoclonal antibodies are now commercially available and suitable for this purpose [40] [41] [42] . 363

In regard to the focus forming assay development, we initially hypothesized that the 364 absence of in Vero E6 cells would make them more susceptible to SARS-CoV-2 infection and 365 therefore a more sensitive choice of cell line for the focus forming assay. Surprisingly, we found 366 that Vero WHO cells were more suited to foci formation. It is worth noting that other labs have 367

shown that a higher titer and larger, clearer plaques result when Vero E6 cells are used in place of 368

Vero WHO cells when performing plaque-assay based titrations with SARS-CoV-2 [32]. This may 369 reflect differences between the Wuhan clinical isolate used in this study as opposed to the USA-370 WA1/2020 isolate or this may be an artifact of the focus forming assay. Because we find that by 371 qPCR, genome copy number is typically highest in Vero E6 cells, we hypothesize that more defective or non-infectious virus results from replication in Vero E6 cells. Additionally, high levels 373 of genome replication in Vero E6 may not correlate with ability to spread laterally in cell culture 374 and form foci. The discrepancy in SARS-CoV-2 replication in these two cell lines warrants further 375 study. 376

Our understanding of the impact of cell type on SARS-CoV-2 entry, replication, assembly, 377 and egress is in its infancy. These gaps in our knowledge were recently made evident by the use 378 of chloroquine and hydroxychloroquine-widely used anti-malarial drugs that create suboptimal To advance our understanding of the SARS-CoV-2 life cycle in susceptible cell types, we 390 generated multi-step growth curves for a variety of human, simian, and rodent cell types. In most 391 cases, viral replication peaked at 24 HPI in susceptible cell lines and this cell-associated virus was 392 maintained for the duration of the experiment. In many cases, however, the presence of virus in 393 the supernatant did not peak until 72-96 HPI. In the context of the viral replication cycle, our data 394 suggests that genome replication in vitro peaks after just 24 hours, however assembly and egress 395 from infected cells may take as long as 72-96 HPI. 396

While there are conflicting reports concerning the suitability of Huh7 cells for SARS-CoV-397 2 studies [32, 33] we observed a striking discrepancy was between cell-associated virus within the 398 total RNA and the virus detected within the cell supernatant. As much as 1.1 × 106 copies/µL of 399 cell-associated virus was detectable in RNA isolated from Huh7 cells, but virtually no detectable 400 virus was found in the cell supernatant. This suggested to us that viral entry-and hence the 401 production of cell-associated virus within the total RNA fraction-was independent of successful 402 viral egress. This trend did not hold for RIG-I-deficient Huh 7.5 cells [49] , suggesting that viral egress is interferon (IFN) sensitive. This observation is in accordance with previous studies 404 describing SARS-CoV-2 sensitivity to type I IFNs [50, 51] . Further studies are warranted to 405 determine what factors are necessary and sufficient for viral egress and could therefore serve as 406 potential therapeutic targets. 407

CoV-2 infection, pathogenesis, and possibly transmission [35, [52] [53] [54] , which may reflect differing 409 susceptibility of hamster cell types based on anatomical location of the isolated cells. 410

We did not observe a strong correlation between ACE2 and viral spike protein levels, nor 411 did we see a strong relationship between viral genome copy and ACE2 mRNA level. Our results 412

suggest that host cell susceptibility to SARS-CoV-2 infection is more complicated than ACE2 413 expression alone, thus warranting further investigation. 414

We have showed by ELISA that convalescent sera sourced from human, non-human 416 primate, and mice infected with SARS-CoV-2 can bind to the RBD of SARS-CoV-2. While the 417 S2 subunit of the SARS-CoV-2 spike protein is highly conserved among betacoronaviruses, 418

previous studies have showed that the RBD within the spike protein of SARS-CoV-2 is unique 419 [16, 55] . In serological studies of SARS-CoV-2, the presence of antibodies binding SARS-CoV-2 420 RBD is considered the most sensitive and specific indicator of previous SARS-CoV-2 exposure. 421

These results increase our confidence that within these polyclonal sera samples are neutralizing 422 antibodies that are specific to SARS-CoV-2, rather than cross-reactivity due prior coronavirus 423 exposure, as has been called into question by some [28, 55, 56] . 424

As other labs have noted [57] we observed that binding to SARS-CoV-2 RBD appears to 425 correlate with neutralization capacity, as human samples with a high AUC for RBD binding by 426 ELISA also had lower EC50 values, indicating that low concentrations of sera from these patients 427 were sufficient to neutralize 50% of a standardized amount of virus. We have showed that each of 428 these samples can effectively neutralize SARS-CoV-2 in vitro and that neutralization can be 429 with 100µL per well of diluted samples. This plate containing sample dilution on the cell 466 monolayer was placed in an incubator with 37°C, 5% CO2 for 1 hour. A solution of 2% 467 methylcellulose (Sigma-M0512-250G) in 1 × PBS was made in advance of the assay and stored at 468 4°C until ready to use. On the day of the assay and during the one-hour infection period, 2% 469 methylcellulose was diluted 1:1 in 5% DMEM and placed on a rocker to mix. The 1% 470 methylcellulose-media mixture (hereby referred to as overlay media) was stored at room 471 temperature until ready to use. After the one-hour infection period, the 96 well plate containing 472 sample dilution and cell monolayer was removed from incubator. Overlay was added to the plate 473 by adding 125 µL of overlay media to each well. This step reduces the uncontrolled spread of virus 474 throughout the monolayer on the well, making it difficult to distinguish individual foci. After the 475 addition of overlay media, the plate was returned to an incubator with 37°C, 5% CO2 for 24 hours. The plate was removed from the incubator and the media containing the overlay and sample was 481 aspirated off. One wash with 150µL of 1 × PBS per well was performed, taking care not to disrupt 482 the cell monolayer. 50µL per well of 5% PFA in PBS was added for the fixing step. With the 5% 483 PFA still in the plate, the plate was submerged in a bath of 5% Formalin buffered phosphate 484 (Fisher: SF100-4) in 1 × PBS for 15 minutes at room temperature. After 15 minutes, the plate was 485 removed from the formaldehyde bath and the 5% PFA removed from the monolayer. One wash 486 with 100µL of 1 × PBS (tissue culture grade) per well was performed. The plate was submerged 487 in a bath of 1 × PBS to rinse and removed from BSL-3 containment. Foci were visualized by an 488 immunostaining protocol. The 96-well plate was first washed twice with 150µL per well of FFA 489

Wash buffer (1 × PBS, 0.05% Triton X-100). The primary antibody consisted of polyclonal anti-490 SARS-CoV-2 guinea pig sera (BEI: NR-0361) and was diluted 1:15,000 with FFA Staining Buffer 491

(1 × PBS, 1mg/ml saponin (Sigma: 47036)).Then, 50µL per well of primary antibody was allowed 492 to incubate for 2 hours at room temperature or 4°C overnight. The 96-well plate was then washed 493 three times with 150µL per well of FFA Wash Buffer. The secondary antibody consisted of goat 494

anti-mouse conjugated horseradish peroxidase (Sigma: A-7289) diluted 1:5,000 in FFA Staining 495

Buffer. Similarly, 50µL per well of secondary antibody was allowed to incubate for 2 hours at room temperature or 4°C overnight. The plate was washed three times with 150uL per well of FFA 497 wash buffer. Finally, 50µL per well KPL Trueblue HRP substrate was added to each well and 498 allowed to develop in the dark for 10-15 minutes, or until blue foci are visible. The reaction was 499 then quenched by two washes with Millipore water and imaged immediately thereafter with a CTL 500 machine to quantify foci. 501 502

Briefly, sera samples were diluted 1:40 in 5% DMEM and added to the topmost row of a round 504 bottom 96-well plate. Sera samples were then serially diluted 1:3 down the remainder of the plate 505 in 5% DMEM. An equal volume of SARS-CoV-2 diluted to ~600 FFU/mL (~60 FFU/100µL) was 506 then added to the serially diluted sample, mixed thoroughly, and allowed to incubate at 37°C for 507 1 hour. Then 100µL of SARS-CoV-2+sera mixture was transferred to a Vero WHO cell monolayer 508 (as described in the focus forming assay). From this point, the assay was as described in the focus 509 forming assay section. 510 511

Binding of human polyclonal sera to recombinant SARS-CoV2 proteins was determined by 513 ELISA. A 1ug/mL mixture of 50µL per well containing recombinant protein in carbonate buffer 514 (0.1M Na2CO3 0.1M NaHCO3 pH 9.3) was used to coat MaxiSorp (ThermoFisher) 96-well plates 515 overnight at 4°C. Plates were blocked with blocking buffer (PBS + 5%BSA + 0.5% Tween) for 2 516 hours at room temperature the following day and washed four times with wash buffer. Polyclonal 517 sera was serially diluted in blocking buffer prior to plating. Sera was allowed to incubate for 1 518 hour at room temperature and washed four times with wash buffer. Following the one hour 519 incubation, goat-anti-human IgG HRP (Sigma) conjugated secondary (1:5000) was added and 520 allowed to incubate for 1 hour at room temperature. The plate was washed again four times with 521 wash buffer and the ELISA was visualized with 100µL per well of TMB enhanced substrate 522 (Neogen Diagnostics) and allowed to develop in the dark for 15 minutes. A solution of 1N HCl 523 was used to quench the reaction and the plate was read for an optical density of 450 nanometers 524 on a BioTek Epoch plate reader. The total peak area under the curve (AUC) was calculated using 525

GraphPad Prism 8. 526

Isolation of RNA from cell culture and culture supernatants 528 RNA was isolated from cell culture and supernatant using an Invitrogen Purelink RNA mini kit 529 according to the manufacturer's instructions. 530 531 RT-qPCR 532 hACE2 expression was measured by qRT-PCR using Taqman primer and probe sets from IDT 533 (assay ID Hs.PT.58.27645939). SARS-CoV-2 viral burden was measured by qRT-PCR using 534

Taqman primer and probe sets from IDT with the following sequences: were allowed to infect cell monolayer for 1 hour at 37°C, 5% CO2 before overlay with 2% 566 methylcellulose in 5% DMEM. Following a 24 hour incubation at 37°C, 5% CO2, cells were fixed 567 in a solution of 5% paraformaldehyde diluted in 1 × PBS. Foci were visualized via immunostaining 568 with polyclonal guinea pig anti-SARS-CoV-2 sera and goat anti-guinea pig conjugated HRP. KPL Neutralization capacity was measured by mixing serially diluted sera with a standardized amount 614 of virus (~60 FFU) and allowed to incubate for 1 hour at 37°C, 5% CO2 before allowed to infect a 615 cell monolayer as described in Figure 1 non-human primate, and mouse sera samples previously described. 

The Incubation Period of Coronavirus Disease

Publicly Reported Confirmed Cases: Estimation and Application. Annals of Internal 624 Medicine

Early transmission dynamics in Wuhan, China, of novel coronavirus-infected 626 pneumonia

Identification of a novel coronavirus causing severe pneumonia in 628 human: a descriptive study

A new coronavirus associated with human respiratory disease in China. 630 Nature

A pneumonia outbreak associated with a new coronavirus of probable bat 632 origin

The 634 species Severe acute respiratory syndrome-related coronavirus: classifying 2019-nCoV 635 and naming it SARS-CoV-2

The proximal origin of SARS-CoV-2

The molecular biology of coronaviruses

The M, E, and N Structural Proteins of the Severe Acute Respiratory 641

Syndrome Coronavirus Are Required for Efficient Assembly, Trafficking, and Release of 642 Virus-Like Particles

Structural basis for the recognition of SARS-CoV-2 by full-length human 644 ACE2

SARS-CoV-2 Cell Entry Depends on ACE2 and TMPRSS2 and Is 646 Blocked by a Clinically Proven Protease Inhibitor

Angiotensin-converting enzyme 2 is a functional receptor for the SARS 648 coronavirus

SARS-CoV-2 Reverse Genetics Reveals a Variable Infection Gradient in 650 the Respiratory Tract

SARS-CoV-2 Receptor ACE2 Is an Interferon-Stimulated Gene in 652

Human Airway Epithelial Cells and Is Detected in Specific Cell Subsets across Tissues

Specific ACE2 expression in small intestinal enterocytes may cause 655 gastrointestinal symptoms and injury after 2019-nCoV infection

Structure, Function, and Antigenicity of the SARS-CoV-2 Spike 659 Glycoprotein

TMPRSS2 and TMPRSS4 promote SARS-CoV-2 infection of human small 661 intestinal enterocytes

Syndrome Coronavirus Spike Protein for Membrane Fusion and Reduces Viral Control by 664 the Humoral Immune Response

Neutralizing antibodies correlate with protection from SARS-CoV-2 in 666 humans during a fishery vessel outbreak with high attack rate. medRxiv

Neutralizing antibody responses to SARS-CoV-2 in a COVID-19 recovered 669 patient cohort and their implications. medRxiv, 2020: p. 2020.03.30

Quantifying antibody kinetics and RNA shedding during early-phase 671 SARS-CoV-2 infection

Antibody responses to SARS-CoV-2 in patients with COVID-19

The epidemiology and clinical information about COVID-19. European 675 journal of clinical microbiology & infectious diseases : official publication of the European 676 Society of Clinical Microbiology

Clinical Characteristics of Coronavirus Disease 2019 in China

The epidemiology and pathogenesis of coronavirus 680 disease (COVID-19) outbreak

SARS-CoV-2 and viral sepsis: observations and hypotheses. The Lancet

Clinical findings in a group of patients infected with the 2019 novel 684 coronavirus (SARS-Cov-2) outside of Wuhan, China: retrospective case series

The receptor-binding domain of the viral spike protein is an 687 immunodominant and highly specific target of antibodies in SARS-CoV-2 patients

Propagation, Quantification, Detection, and 690 Storage of West Nile Virus

Development and optimization of a direct plaque assay for human and 693 avian metapneumoviruses

Enhanced isolation of SARS-CoV-2 by TMPRSS2-expressing cells

Isolation and characterization of SARS-CoV-2 from the first US 697 COVID-19 patient

SARS-coronavirus-2 replication in Vero E6 cells: replication kinetics, 699 rapid adaptation and cytopathology

mRNA induced expression of human angiotensin-converting enzyme 2 701 in mice for the study of the adaptive immune response to severe acute respiratory syndrome 702 coronavirus 2. bioRxiv

Syrian hamsters as a small animal model for SARS-CoV-2 infection and 704 countermeasure development

Mouse model of SARS-CoV-2 reveals inflammatory role of 707 type I interferon signaling. bioRxiv

A SARS-CoV-2 infection model in mice demonstrates protection by 709 neutralizing antibodies

Plaque assay for human coronavirus NL63 using 711 human colon carcinoma cells

Two Detailed Plaque Assay Protocols for the Quantification of 713 Infectious SARS-CoV-2. Current protocols in microbiology

Potently neutralizing and protective human antibodies against SARS-CoV-716 2

Potently neutralizing human antibodies that block SARS-CoV-2 receptor 718 binding and protect animals. bioRxiv, 2020: p. 2020.05.22.111005

Rapid isolation and profiling of a diverse panel of human monoclonal 720 antibodies targeting the SARS-CoV-2 spike protein

Hydroxychloroquine, a less toxic derivative of chloroquine, is effective in 722 inhibiting SARS-CoV-2 infection in vitro

Remdesivir and chloroquine effectively inhibit the recently emerged novel 724 coronavirus (2019-nCoV) in vitro

In Vitro Antiviral Activity and Projection of Optimized Dosing Design of 726

Hydroxychloroquine for the Treatment of Severe Acute Respiratory Syndrome Coronavirus 727 2 (SARS-CoV-2)

A Randomized Trial of Hydroxychloroquine as Postexposure 729

Prophylaxis for Covid-19

Chloroquine does not inhibit infection of human lung cells with SARS-731 CoV-2

Proteolytic processing of Middle East respiratory syndrome coronavirus 733 spikes expands virus tropism

Regulating intracellular antiviral defense and permissiveness to 736 hepatitis C virus RNA replication through a cellular RNA helicase, RIG-I

SARS-CoV-2 is sensitive to type I interferon pretreatment. 739 bioRxiv : the preprint server for biology

Antiviral activities of type I interferons to SARS-CoV-2 infection. 741 Antiviral research

Pathogenesis and transmission of SARS-CoV-2 in golden hamsters

Massive transient damage of the olfactory epithelium associated with 745 infection of sustentacular cells by SARS-CoV-2 in golden Syrian hamsters. Brain, 746 Behavior, and Immunity

Surgical mask partition reduces the risk of non-contact transmission 748 in a golden Syrian hamster model for Coronavirus Disease 2019 (COVID-19)

A serological assay to detect SARS-CoV-2 seroconversion in humans

Cross-reactivity towards SARS-CoV-2: the potential role of low-pathogenic 753 human coronaviruses. The Lancet Microbe

Human neutralizing antibodies elicited by SARS-CoV-2 infection