key: cord-0989479-yk60ol95
authors: Ogando, Natacha S.; Dalebout, Tim J.; Zevenhoven-Dobbe, Jessika C.; Limpens, Ronald W.; van der Meer, Yvonne; Caly, Leon; Druce, Julian; de Vries, Jutte J. C.; Kikkert, Marjolein; Bárcena, Montserrat; Sidorov, Igor; Snijder, Eric J.
title: SARS-coronavirus-2 replication in Vero E6 cells: replication kinetics, rapid adaptation and cytopathology
date: 2020-04-20
journal: bioRxiv
DOI: 10.1101/2020.04.20.049924
sha: 1b3f2d41d8b3ee09175d8e4fd6dcf7660bc9efed
doc_id: 989479
cord_uid: yk60ol95

The sudden emergence of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) at the end of 2019 from the Chinese province of Hubei and its subsequent pandemic spread highlight the importance of understanding the full molecular details of coronavirus infection and pathogenesis. Here, we compared a variety of replication features of SARS-CoV-2 and SARS-CoV and analysed the cytopathology caused by the two closely related viruses in the commonly used Vero E6 cell line. Compared to SARS-CoV, SARS-CoV-2 generated higher levels of intracellular viral RNA, but strikingly about 50-fold less infectious viral progeny was recovered from the culture medium. Immunofluorescence microscopy of SARS-CoV-2-infected cells established extensive cross-reactivity of antisera previously raised against a variety of nonstructural proteins, membrane and nucleocapsid protein of SARS-CoV. Electron microscopy revealed that the ultrastructural changes induced by the two SARS viruses are very similar and occur within comparable time frames after infection. Furthermore, we determined that the sensitivity of the two viruses to three established inhibitors of coronavirus replication (Remdesivir, Alisporivir and chloroquine) is very similar, but that SARS-CoV-2 infection was substantially more sensitive to pre-treatment of cells with pegylated interferon alpha. An important difference between the two viruses is the fact that - upon passaging in Vero E6 cells - SARS-CoV-2 apparently is under strong selection pressure to acquire adaptive mutations in its spike protein gene. These mutations change or delete a putative ‘furin-like cleavage site’ in the region connecting the S1 and S2 domains and result in a very prominent phenotypic change in plaque assays.

Corresponding author: Eric J. Snijder (e.j.snijder@lumc.nl) 23 24

Keywords: plaque phenotype, evolution, RNA synthesis, antisera, furin cleavage site, 25 antiviral drugs 26 27

Abbreviations: SARS-CoV, severe acute respiratory syndrome coronavirus; CoV, 28

Coronavirus; CPE, cytopathic effect; HCoV, human coronavirus; MERS-CoV, Middle East 29 respiratory syndrome coronavirus;; nsp, non-structural protein; S protein, spike protein; ACE2, 30

angiotensin-converting enzyme 2; NGS, next-generation sequencing; RO, replication 31 organelle; DMV, Double-membrane vesicle; PEG-IFN-α, pegylated interferon alpha; UTR, 32 untranslated region. 33 34 very similar and occur within comparable time frames after infection. Furthermore, we 48 determined that the sensitivity of the two viruses to three established inhibitors of coronavirus 49 replication (Remdesivir, Alisporivir and chloroquine) is very similar, but that SARS-CoV-2 50 infection was substantially more sensitive to pre-treatment of cells with pegylated interferon 51

alpha. An important difference between the two viruses is the fact that -upon passaging in 52

Vero E6 cells -SARS-CoV-2 apparently is under strong selection pressure to acquire adaptive 53 mutations in its spike protein gene. These mutations change or delete a putative 'furin-like 54 cleavage site' in the region connecting the S1 and S2 domains and result in a very prominent 55 phenotypic change in plaque assays. 56 57 INTRODUCTION set of smaller nsps (nsp7-10) that mainly appear to serve as cofactors/modulators of other 112

The newly emerged SARS-CoV-2 was rapidly identified as a CoV that is relatively closely 114 BetaCoV/Australia/VIC01/2020; (34), which will be used throughout this study. Until now, this 139 isolate has been provided to 17 other laboratories worldwide to promote the rapid 140 characterization of SARS-CoV-2, in this critical time of lockdowns and other preventive 141 measures to avoid a collapse of public health systems. 142

In this report, we describe a comparative study of the basic replication features of SARS-CoV 143 and SARS-CoV-2 in Vero E6 cells, including growth kinetics, virus titres, plaque phenotype 144 and an analysis of intracellular viral RNA and protein synthesis. Additionally, we analysed 145 infected cells by light and electron microscopy, and demonstrated cross-reactivity of 13 146 available SARS-CoV-specific antisera (recognising 10 different viral proteins) with their SARS-147

CoV-2 counterparts. Finally, we established the conditions for a medium-throughput assay to 148 evaluate basic antiviral activity and assessed the impact of some known CoV inhibitors on 149 SARS-CoV-2 replication. In addition to many anticipated similarities, our results also 150 established some remarkable differences between the two viruses that warrant further 151 investigation. One of them is the rapid evolution -during virus passaging in Vero cells -of a 152 specific region of the SARS-CoV-2 S protein that contains the so-called 'furin-like cleavage 153 Dried agarose gels were used for direct detection of viral mRNAs by hybridization with a 32 P-186 labeled oligonucleotide probe (5'-CACATGGGGATAGCACTAC-3') that is complementary to 187 a fully conserved sequence located 30 nucleotides upstream of the 3' end of the genome and 188 all subgenomic mRNAs produced by SARS-CoV-2 and SARS-CoV. After hybridization, RNA 189 bands were visualised and quantified by phosphorimaging using a Typhoon-9410 variable 190 mode scanner (GE Healthcare) and ImageQuant TL software (GE Healthcare). In order to 191 verify the amount of RNA loaded, a second hybridization was performed using a 32 P-labeled 192 oligonucleotide probe recognizing 18S ribosomal RNA (5'-GATCCGAGGGCCTCACTAAAC- Leiden, the Netherlands) while including appropriate quality controls after each step of the 210 procedure. Sequencing was performed using a NovaSeq 6000 Sequencing System (Illumina). 211

Subsequently, sequencing reads were screened for the presence of human (GRCh37.75), 212 mouse (GRCm38.p4), E. coli MG1655 (EMBL U00096.2), phiX (RefSeq NC_001422.1) and 213 common vector sequences (UniVec and ChlSab1.1). Prior to alignment, reads were trimmed 214 to remove adapter sequences and filtered for sequence quality. The remaining reads were 215 mapped to the SARS-CoV-2 GenBank reference sequence (NC_045512.2; (38)). Data 216 analysis was performed using Bowtie 2 (39). Raw NGS data sets for each virus sample 217 analysed in this study are deposited in NCBI Bioproject and available under the following 218 links: ---. Only SARS-CoV-2-specific reads were included in these data files. 219

To study evolution/adaptation of the S protein gene, we performed an in-depth analysis of 220 reads covering the S1/S2 region of the S protein gene. This was done for the p2 stock and for 221 the four virus samples of the plaque picking experiment shown in Fig. 1a . First, all reads 222 spanning nt 23,576 to 23,665 of the SARS-CoV genome were selected. Next, reads 223 constituting less than 1% of the total number of selected reads were excluded for further 224 analysis. The remaining number of reads were 3,860 (p2 stock), 1,924 (S5p1), 2,263 (S5p2), 225 4,049 (S5p3) and 3,323 (L8p1). These reads were translated in the S protein open reading 226 frame and the resulting amino acid sequences were aligned, grouped on the basis of 227 containing the same mutations/deletions in the S1/S2 region and ranked by frequency of 228 occurrence (Fig. 1b) . 229 230

The SARS-CoV-specific rabbit or mouse antisera/antibodies used in this study are listed in 232 Table 1 . Most antisera were described previously (see references in Table 1) , with the 233 exception of three rabbit antisera recognizing SARS-CoV nsps 8, 9 and 15. These were raised 234 using full-length (His)6-tagged bacterial expression products (nsp8 and nsp15) or a synthetic 235 peptide (nsp9, aa 4209-4230 of SARS-CoV pp1a), which were used to immunize New Zealand 236 white rabbits as described previously (40, 41). Cross-reactivity of antisera to SARS-CoV-2 237 targets was evaluated microscopically by immunofluorescence assay (IFA) and for some 238 antisera (nsp3 and N protein) also by Western blot analysis. Double-stranded RNA was 239 detected using mouse monoclonal antibody J2 from Scicons (42). 240

Cells were grown on glass coverslips and infected as described above (43). At 12, 24, 48 or 241 72 h p.i., cells were fixed overnight at 4°C using 3% paraformaldehyde in PBS (pH 7.4). Cells 242 were washed with PBS containing 10 mM glycine and permeabilized with 0.1% Triton X-100 243 in PBS. Cells were incubated with antisera diluted in PBS containing 5% FCS. Secondary 244 antibodies used were an Alexa488-conjugated goat anti-rabbit IgG antibody (Invitrogen), a 245

Cy3-conjugated donkey anti-rabbit IgG antibody (Jackson ImmunoResearch Laboratories) 246

and an Alexa488-conjugated goat anti-mouse IgG antibody (Invitrogen). Nuclei were stained 247 with 1 µg/ml Hoechst 33258 (ThermoFischer). Samples were embedded using Prolong Gold 248 (Life Technologies) and analysed with a Leica DM6B fluorescence microscope using LASX 249 software. 

Vero E6 cells 294 SARS-CoV-2 isolate BetaCoV/Australia/VIC01/2020 was received as a stock derived from two 295 consecutive passages in Vero/hSLAM cells (34). The virus was then propagated two more 296 times at low MOI in Vero E6 cells, in which it caused a severe cytopathic effect (CPE). We 297 also attempted propagation in HuH7 cells, using the same amount of virus or a ten-fold larger 298

inoculum, but did not observe any cytopathology after 72 h (data not shown). At 24 h p.i., 299 immunofluorescence microscopy (see below) revealed infection of only a small percentage of 300 the HuH7 cells, without any clear spread to other cells occurring in the next 48 h. We therefore 301 conclude that infection of HuH7 cells does not lead to a productive SARS-CoV-2 infection and 302 deemed this cell line unsuitable for further SARS-CoV-2 studies. 303

The infectivity titre of the Leiden-p2 stock grown in Vero E6 cells was analysed by plaque 304 assay, after which we noticed a mixed plaque phenotype (~1:3 ratio of small versus large 305 (plaques; data not shown) while a virus titre of 7 x 10 6 PFU/ml was calculated. To verify the 306 identity and genome sequence of the SARS-CoV-2/p2 virus stock, we isolated genomic RNA 307 from culture supernatant and applied next-generation sequencing (NGS; see methods for 308 details). The resulting consensus sequence was found to be identical to the sequence 309 previously deposited in GenBank (accession number MT007544.1) (34), with one exception 310 (see below). Compared to the SARS-CoV-2 GenBank reference sequence (NC_045512.3) 311

(38) and other field isolates (29), isolate BetaCoV/Australia/VIC01/2020 exhibits >99.9% 312 sequence identity. In addition to synonymous mutations in the nsp14-coding sequence 313

(U19065 to C) and S protein gene (U22303 to G), ORF3a contains a single non-synonymous 314 mutation (G26144 to U). Strikingly, the 3' untranslated region (UTR) contains a 10-nt deletion 315 (nt 29750-29759; CGAUCGAGUG) located 120 nt upstream of the genomic 3' end , which is 316 not present in other SARS-CoV-2 isolates thus far (>670 SARS-CoV2 sequences present in 317

GenBank on April 17, 2020). 318

In about 71% of the 95,173 p2 NGS reads covering this position, we noticed a G23607 to A 319 mutation encoding an Arg682 to Gln substitution near the so-called S1/S2 cleavage site of the 320 viral S protein (see Discussion), with the other 29% of the reads being wild-type sequence. As 321 this ratio approximated the observed ratio between large and small plaques, we performed a 322 plaque assay on the p1 virus stock (Fig. 1a , leftmost well) and picked multiple plaques of each 323 size, which were passaged three times in Vero E6 cells while monitoring their plaque 324 phenotype. Interestingly, for several of the small-plaque virus clones (like S5; Fig. 1a) we 325 observed rapid conversion to a mixed or large-plaque phenotype during these three passages, 326

while large-plaque virus clones (like L8) stably retained their plaque phenotype (Fig. 1a) . NGS 327 analysis of the genome of a large-plaque p1 virus (L8p1) revealed that >99% of the reads in 328 the S1/S2 cleavage site region contained the G23607 to A mutation described above. No other 329 mutations were detected in the genome, thus clearly linking the Arg682 to Gln substitution in 330 the S protein to the large-plaque phenotype observed for the L8p1 virus. 331

Next, we also analysed the genomes of the p1, p2 and p3 viruses derived from a small-plaque 332 (S5) that was picked. This virus clone retained its small-plaque phenotype during the first 333 passage ( Fig. 1a; S5p1 ), but began to yield an increasing proportion of large(r) plaques during 334 subsequent passages. Sequencing of S5p2 (Fig 1b) revealed a variety of low-frequency reads 335 with mutations near the S1/S2 cleavage site motif (aa 681-687; PRRAR↓SV), with G23607 to 336 A (specifying the Arg682 to Gln substitution) again being the dominant one (in ~0.9% of the 337 reads covering nt 23,576 to 23,665 of the genome). At lower frequencies single-nucleotide 338 changes specifying Arg682 to Trp and Arg683 to Leu substitutions were also detected. 339

Furthermore, a 10-aa deletion (residues 679-688) that erases the S1/S2 cleavage site region 340 was discovered, as well as a 5-aa deletion (residues 675-679) immediately preceding that 341 region. The amount of large plaques increased substantially upon the next passage, with NGS 342 revealing the prominent emergence of the mutants containing the 10-aa deletion or the Arg682 343 to Gln point mutation (~22% and ~12% of the reads, respectively), and yet other minor variants 344 with mutations in the PRRAR↓SV sequence being discovered. Taken together these data 345 clearly link the large-plaque phenotype of SARS-CoV-2 to the acquisition of mutations in this 346 particular region of the S protein, which apparently provides a strong selective advantage 347 during passaging in Vero E6 cells. 348 349

To our knowledge, a detailed comparison of SARS-CoV-2 and SARS-CoV replication kinetics 351 in cell culture has not been reported so far. Therefore, we infected Vero E6 cells with the 352 SASR-CoV-2/p2 virus stock at high m.o.i. to analyse viral RNA synthesis, protein expression 353 and the release of infectious viral progeny (Fig. 2a) . This experiment was performed using 4 354 We also monitored viral protein production by Western blot analysis using antisera targeting 377 a non-structural (nsp3) and structural (N) protein. As expected from the RNA analysis, the 378 accumulation of both viral proteins increased with time, and was detected somewhat earlier 379 for SARS-CoV-2 than for SARS-CoV (data not shown). Overall, we conclude that in Vero E6 380 cells, SARS-CoV-2 produces levels of intracellular RNA and proteins that are at least 381 comparable to those of SARS-CoV, although this does not translate into the release of equal 382 amounts of infectious viral progeny (Fig. 2a) . 383 384

To be able to follow virus replication in SARS-CoV-2-infected cells more closely, we explored 386 cross-reactivity of a variety of antisera previously raised against SARS-CoV targets, in 387 particular a variety of nsps. In an earlier study, many of those were found to cross-react also 388

with the corresponding MERS-CoV targets (35), despite the relatively large evolutionary 389 distance between MERS-CoV and SARS-CoV. Based on the much closer relationship with 390 SARS-CoV-2, similar or better cross-reactivity of these SARS-CoV reagents was expected, 391 which was explored using immunofluorescence microscopy. 392

Indeed, most antisera recognizing SARS-CoV nsps that were tested (nsp3, nsp4, nsp5, nsp8, 393 nsp9, nsp13, nsp15) strongly cross-reacted with the corresponding SARS-CoV-2 target (Fig.  394   4 and Table 1 ), the exception being a polyclonal nsp6 rabbit antiserum. Likewise, both a 395 polyclonal rabbit antiserum and mouse monoclonal antibody recognizing the N protein cross-396 reacted strongly (Fig. 4b and Table 1 ). The same was true for a rabbit antiserum raised against 397 a C-terminal peptide of the SARS-CoV M protein (Fig 4e) . Labelling patterns were essentially 398 the elaborate membrane structures of the viral ROs are formed (Fig. 4a, c, d) . Punctate 401 structures in the same area of the cell were labelled using an antibody recognizing double-402 stranded RNA (dsRNA), which presumably recognizes replicative intermediates of viral RNA 403 synthesis (45, 46). The N protein signal was diffusely cytosolic (Fig. 4b) , whereas the M protein 404 labelling predominantly showed the expected localization to the Golgi complex (Fig. 4e) , where 405 the protein is known to accumulate (47). 406 407 408

We next used electron microscopy to investigate the ultrastructural changes that SARS-CoV-410 2 induces in infected cells, and focused on the membranous replication organelles (ROs) that 411 supports viral RNA synthesis and on the assembly and release of new virions (Fig. 5) . 412

Compared to mock-infected control cells (Fig. 5a-b) , various distinct membrane alterations 413 were observed in cells infected with either SARS-CoV or SARS-CoV-2 ( Fig. 5c-j) . At 6 h p.i., 414 larger regions with membrane alterations were found particularly in cells infected with SARS-415

CoV-2 (data not shown), which may align with the somewhat faster onset of intracellular RNA 416 synthesis in SARS-CoV2-infected Vero E6 cells (Fig. 3a) . and 5h-i, asterisks). In addition, convoluted membranes (45) were readily detected in SARS-420

CoV-infected cells, while zippered ER (25, 48, 49) appeared to be the predominant structure 421 in SARS-CoV-2-infected cells ( Fig. 5e and 5i , white arrowheads). As previously described for 422 SARS-CoV (45), also SARS-CoV-2-induced DMV appeared to fuse through their outer 423 membrane, giving rise to vesicles packets that increased in numbers as infection progressed 424 (Fig 5f and 5k , white asterisks). Virus budding near the Golgi apparatus, presumably into 425 smooth membranes of the ER-Golgi intermediate compartment (ERGIC) (50-52), was 426 frequently observed at 8 h p.i. (Fig. 5k-l and 5o-p) . This step is followed by transport to the 427 plasma membrane and release of virus particles into extracellular space. By 10 h p.i., released 428 progeny virions were abundantly detected around all infected cells (Fig. 5m-n and 5q-r) . contribute to the lower progeny titres obtained for this virus (Fig. 2a) . 434 435 436

In order to establish and validate a CPE-based assay to identify potential inhibitors of SARS-438

CoV-2 replication, we selected four previously identified inhibitors of CoV replication: 439

Remdesivir (53, 54), chloroquine (55, 56), Alisporivir (57, 58) and pegylated interferon alpha 440 (PEG-IFN-α) (35, 59). Cells were infected at low MOI to allow for multiple cycles of replication. 441

After three days, a colorimetric cell viability assay (60) was used to measure drug toxicity and In this report, we describe a comparative analysis of the replication features of SARS-CoV-2 464 and SARS-CoV in Vero E6 cells, one of the most commonly used cell lines for studying these 465 two viruses. However, in contrast to the stable phenotype exhibited by SARS-CoV during our 466 17 years of working with this virus in these cells, SARS-CoV-2 began to exhibit remarkable 467 phenotypic variation in plaque assays within a few passages after its isolation from clinical 468 samples (Fig. 1a) . In addition to the BetaCoV/Australia/VIC01/2020 isolate used in this study, nucleotide substitution in the S protein gene, clearly established that a single S protein 474 mutation (Arg682 to Gln) was responsible for the observed plaque size difference. This 475 mutation is localized near the so-called 'furin-like' S1/S2 cleavage site (Fig. 1b) Possibly, this pathway is also employed by our Vero E6-cell adapted SARS-CoV-2 mutants 488 that have lost the furin-like cleavage site, like clone L8p1 and multiple variants encountered in 489 S5p3 (Fig. 1a) . These variants contain either single point mutations or deletions of 5 to 10 aa 490 (Fig. 1b) , resembling variants recently reported by other laboratories (30, 70, 71) . Interestingly 491 similar changes were also observed in some clinical SARS-CoV-2 isolates that had not been 492 passaged in cell culture (70). It is currently being investigated why mutations that inactivate 493 the furin-like cleavage site provide such a major selective advantage during SARS-CoV-2 494 passaging in Vero E6 cells and how this translates into the striking large-plaque phenotype 495 documented in this paper. 496

An additional remarkable feature confirmed by our re-sequencing of the 497 BetaCoV/Australia/VIC01/2020 isolate of SARS-CoV-2 is the presence of a 10-nt deletion in 498 the 3' UTR of the genome (34). Screening of other available SARS-CoV-2 genome sequences 499 indicated that the presence of this deletion apparently is unique for this particular isolate, and 500 likely represents an additional adaptation acquired during cell culture passaging. This deletion 501 maps to a previously described "hypervariable region" in the otherwise conserved 3' UTR, and 502 in particular to the so-called s2m motif (72) that is conserved among CoVs and also found in 503 several other virus groups (73, 74). The s2m element has been implicated in the binding of 504 host factors to viral RNAs, but its exact function has remained enigmatic thus far. Strikingly, 505

for the mouse hepatitis coronavirus the entire hypervariable region (including s2m) was found 506 to be dispensable for replication in cell culture, but highly relevant for viral pathogenesis in 507 mice (72). Although the impact of this deletion for SARS-CoV-2 remains to be studied in more 508 detail, these previous data suggest that this mutation need not have a major impact on SARS-509

CoV-2 replication in Vero E6 cells. This notion is also supported by the fact that the results of 510 our antiviral screening assays (Fig. 6) correlate well with similar studies performed with other 511 SARS-CoV-2 isolates (54, 75, 76). Clearly, this could be different for in vivo studies, for which 512 it would probably be better to rely on SARS-CoV-2 isolates not carrying this deletion in their 513

Vero E6 cells are commonly used to isolate, propagate, and study SARS-CoV-like viruses as 515 they support viral replication to high titres (77-81). This may be due to a high expression level 516 of the ACE-2 receptor (82) that is used by both SARS-CoV-2 and SARS-CoV (9) and/or the 517 fact that they lack the ability to produce interferon (83, 84). It will be interesting to evaluate 518 whether there is a similarly strong selection pressure to adapt the S1/S2 region of the S protein 519 when SARS-CoV-2 is passaged in other cell types. Such studies are currently in progress in 520 our laboratory and already established that HuH7 cells may be a poor choice, despite the fact 521 that they were used for virus propagation (9, 85) and antiviral screening in other studies (54, 522 86). Immunolabelling of infected HuH7 cells (data not shown) revealed non-productive 523 infection of only a small fraction of the cells and a general lack of cytopathology. While other 524 cell lines are being evaluated, as illustrated above, the monitoring of the plaque phenotype 525 (plaque size and homogeneity) may provide a quick and convenient method to assess the 526 composition of SARS-CoV-2 stocks propagated in Vero E6 cells, at least where it concerns 527 the evolution of the S1/S2 region of the S protein. 528

Given the ongoing SARS-CoV-2 pandemic, the detailed characterization of its replication cycle 529

is an important step in understanding the molecular biology of the virus and defining potential 530 targets for inhibitors of replication. The cross-reacting antisera described in this study (Table  531 1) will be a useful tool during such studies. In general, the subcellular localization of viral nsps 532 and structural proteins (Fig. 4) and the ultrastructural changes associated with RO formation 533 ( Fig. 5) were very similar for the two viruses. We also observed comparable replication kinetics 534 for SARS-CoV-2 and SARS-CoV in Vero E6 cells, although clearly lower final infectivity titres 535

were measured for SARS-CoV-2 (~50-fold lower; Fig. 2 ). Nevertheless, RNA synthesis could 536 be detected somewhat earlier for SARS-CoV-2 and the overall amount of viral RNA produced 537 exceeded that produced by SARS-CoV (Fig. 3) . This may be indicative of certain assembly or 538 maturation problems or of virus-host interactions that are different in the case SARS-CoV-2. 539

These possibilities merit further investigation, in particular since our preliminary EM studies 540 suggested intriguing differences with SARS-CoV where it concerns the presence of spikes on 541 the surface of freshly released SARS-CoV-2 particles ( Fig. 5n and 5r) . 542

Our analysis of SARS-CoV-2 subgenomic mRNA synthesis revealed the increased relative 543 abundance of mRNAs 7 and 8 (~4-and ~2-fold, respectively) when SARS-CoV-2 was 544 compared to SARS-CoV. Mechanistically, these differences do not appear to be caused by 545 extended base pairing possibilities of the transcription regulatory sequences that direct the 546 synthesis of these two mRNAs (24). As in SARS-CoV, mRNA7 of SARS-CoV-2 encodes for 547 two proteins, the ORF7a and ORF7b proteins, with the latter presumably being expressed 548 following leaky ribosomal scanning (32). Upon its ectopic expression, the ORF7a protein has 549 been reported to induce apoptosis via a caspase-dependent pathway (87) and/or to be 550 involved in cell cycle arrest (88). The ORF7b product is a poorly studied integral membrane 551 protein that has (also) been detected in virions (89). When ORF7a/b or ORF7a were deleted 552 from the SARS-CoV genome, there was a minimal impact on the kinetics of virus replication 553 in vitro in different cell lines, including Vero cells, and in vivo using mice. In another study, 554 however, partial deletion of SARS-CoV ORF7b was reported to provide a replicative 555 advantage in CaCo-2 and HuH7 cells, but not in Vero cells (90). 556

The SARS-CoV ORF8 protein is membrane-associated and able to induce endoplasmic 557 reticulum stress (91, 92), although it has not been characterised in great detail in the context 558 of viral infection. Soon after the emergence of SARS-CoV in 2003, a conspicuous 29-nt (out-559 of-frame) deletion in ORF8 was noticed in late(r) human isolates, but not in early human 560 isolates and SARS-like viruses obtained from animal sources (93-95). Consequently, loss of 561 ORF8 function was postulated to reflect an adaptation to the human host. The re-engineering 562 of an intact ORF8, using a reverse genetics system for the SARS-CoV Frankfurt-1 isolate, 563 yielded a virus with strikingly enhanced (up to 23-fold) replication properties in multiple 564 systems (96) . Clearly, it remains to be established that the increased synthesis of mRNAs 7 565 and 8 is a general feature of SARS-CoV-2 isolates, and that this indeed also translates into 566 higher expression levels of the accessory proteins encoded by ORFs 7a, 7b and 8. If 567 confirmed, these differences definitely warrant an in-depth follow-up analysis as CoV 568 accessory proteins in general have been shown to be important determinants of virulence. 569

They may thus be relevant for our understanding of the wide spectrum of respiratory disease 570 symptoms observed in COVID-19 patients (97). 571

Based on the close ancestral relationship between SARS-CoV-2 and SARS-CoV (98), one 572 might expect that the patterns and modes of interaction with host antiviral defence 573 mechanisms would be similar. However, our experiments with type I interferon treatment of 574 Vero E6 cells (Fig. 6 ) revealed a clear difference, with SARS-CoV-2 being considerably more 575 sensitive than SARS-CoV, as also observed by other laboratories (76). Essentially, SARS-576

CoV-2 replication could be inhibited by similarly low concentrations of PEG-IFN-alpha-2a that 577

inhibit MERS-CoV replication in cell culture (35). Taken together, our data suggest that SARS-578 CoV-2 is less able to counter a primed type I IFN response than SARS-CoV (76, 99). 579

Previously identified inhibitors of CoV replication were used to further validate our cell-based 580 assay for SARS-CoV-2 inhibitor screening. These compounds inhibited replication at similar 581 low-micromolar concentrations and in a similar dose-dependent manner as observed for 582 SARS-CoV (Fig. 6 ). Remdesivir is a prodrug of an adenosine analogue developed by Gilead 583

Sciences. It was demonstrated to target the CoV RNA polymerase and act as a chain 584 terminator (100-102). The clinical efficacy of Remdesivir is still being evaluated and, after 585 some first encouraging results (103), worldwide compassionate use trials are now being 586 conducted. Likewise, hydroxychloroquine and chloroquine have been labelled as potential 587 "game changers" and are being evaluated for treatment of severe COVID-19 patients (104). 588

Both compounds have been used to treat malaria and amebiasis (105), until drug-resistant 589

Plasmodium strains emerged (106). These compounds can be incorporated into endosomes 590 and lysosomes, raising the pH inside these intracellular compartments, which in turn may lead 591 to defects in protein degradation and intracellular trafficking (68, 107). An alternative 592 hypothesis to explain their anti-SARS-CoV activity is based on their impact on glycosylation 593 of the ACE2 receptor that is used by SARS-CoV (56). Finally, as expected, the non- Vero E6 cells, separated in an agarose gel and probed with a radiolabelled oligonucleotide 961 recognizing the genome and subgenomic mRNAs of both viruses. Subsequently, the gel was 962 re-hybridized to a probe specific for 18S ribosomal RNA, which was used as a loading control. 963 

Coronaviridae Study Group of the International Committee on Taxonomy of V. The 624 species Severe acute respiratory syndrome-related coronavirus: classifying 2019-nCoV 625 and naming it SARS-CoV-2

Bat origin of a new human coronavirus: there and back 627 again

Coronavirus 629 infection in acute lower respiratory tract disease of infants

Identification of new human coronaviruses

A novel 634 coronavirus associated with severe acute respiratory syndrome

Coronavirus as a possible 637 cause of severe acute respiratory syndrome

Isolation of a 639 novel coronavirus from a man with pneumonia in Saudi Arabia

Genomic characterization of a newly discovered coronavirus associated with acute 643 respiratory distress syndrome in humans

A pneumonia outbreak 645 associated with a new coronavirus of probable bat origin

A Novel Coronavirus 648 Emerging in China -Key Questions for Impact Assessment

Evolutionary history, potential 651 intermediate animal host, and cross-species analyses of SARS-CoV-2

Identifying SARS-CoV-2 654 related coronaviruses in Malayan pangolins

The proximal origin of 656 SARS-CoV-2

Cross-host evolution of 658 severe acute respiratory syndrome coronavirus in palm civet and human

Middle 661 East respiratory syndrome coronavirus neutralising serum antibodies in dromedary 662 camels: a comparative serological study

A SARS-like cluster of circulating bat coronaviruses shows potential for human 668 emergence

Discovery of a rich gene pool of 670 bat SARS-related coronaviruses provides new insights into the origin of SARS 671 coronavirus

Origin and evolution of pathogenic coronaviruses

Evidence 675 for an Ancestral Association of Human Coronavirus 229E with Bats

Broad receptor 678 engagement of an emerging global coronavirus may potentiate its diverse cross-species 679 transmissibility

Fatal swine acute diarrhoea 681 syndrome caused by an HKU2-related coronavirus of bat origin

Discovery of 684 the first insect nidovirus, a missing evolutionary link in the emergence of the largest RNA 685 virus genomes

The Nonstructural Proteins Directing Coronavirus RNA 687 Synthesis and Processing

A unifying structural and functional model of the coronavirus replication 690 organelle: tracking down RNA synthesis

SARS coronavirus accessory proteins

Therapeutic options for the 2019 novel coronavirus (2019-nCoV)

The Curious 696 Case of the Nidovirus Exoribonuclease: Its Role in RNA Synthesis and Replication 697 Fidelity

Genomic characterisation and 699 epidemiology of 2019 novel coronavirus: implications for virus origins and receptor 700 binding

A new coronavirus associated 702 with human respiratory disease in China

The novel 704 coronavirus 2019 (2019-nCoV) uses the SARS-coronavirus receptor ACE2 and the 705 cellular protease TMPRSS2 for entry into target cells

Unique and 707 conserved features of genome and proteome of SARS-coronavirus, an early split-off 708 from the coronavirus group 2 lineage

Mechanisms and 710 enzymes involved in SARS coronavirus genome expression

Isolation 714 and rapid sharing of the 2019 novel coronavirus (SAR-CoV-2) from the first diagnosis of 715 COVID-19 in Australia. the medical journal of australia

MERS-coronavirus replication induces severe in vitro cytopathology and is strongly 718 inhibited by cyclosporin A or interferon-alpha treatment

721 Identification of a novel coronavirus in patients with severe acute respiratory syndrome. 722

724 Reverse genetics of SARS-related coronavirus using vaccinia virus-based 725 recombination

Complete genome 727 characterisation of a novel coronavirus associated with severe human respiratory 728 disease in Wuhan

Fast gapped-read alignment with Bowtie 2

Proteolytic processing of the replicase ORF1a 732 protein of equine arteritis virus

Production and characterization 734 of monoclonal antibodies against the nucleocapsid protein of SARS-CoV

Monoclonal 737 antibodies to double-stranded RNA as probes of RNA structure in crude nucleic acid 738 extracts

Localization of mouse hepatitis virus nonstructural proteins and RNA synthesis indicates 741 a role for late endosomes in viral replication

Debio 743 025, a cyclophilin binding molecule, is highly efficient in clearing hepatitis C virus (HCV) 744 replicon-containing cells when used alone or in combination with specifically targeted 745 antiviral therapy for HCV (STAT-C) inhibitors

SARS-coronavirus replication is supported by a reticulovesicular network of modified 749 endoplasmic reticulum

Double-stranded RNA 751 is produced by positive-strand RNA viruses and DNA viruses but not in detectable 752 amounts by negative-strand RNA viruses

Ultrastructure and origin of membrane vesicles associated with the 755 severe acute respiratory syndrome coronavirus replication complex

Infectious 758 bronchitis virus generates spherules from zippered endoplasmic reticulum membranes. 759 mBio

The Porcine Deltacoronavirus

Replication Organelle Comprises Double-Membrane Vesicles and Zippered 762

Endoplasmic Reticulum with Double-Membrane Spherules

Replication of coronavirus MHV-A59 in sac-cells: 764 determination of the first site of budding of progeny virions

Ultrastructural characterization of SARS coronavirus

The 770 intracellular sites of early replication and budding of SARS-coronavirus

Coronavirus 773 Susceptibility to the Antiviral Remdesivir (GS-5734) Is Mediated by the Viral Polymerase 774 and the Proofreading Exoribonuclease. mBio

Remdesivir and chloroquine 776 effectively inhibit the recently emerged novel coronavirus (2019-nCoV) in vitro

In vitro inhibition of severe acute 779 respiratory syndrome coronavirus by chloroquine

Chloroquine is a potent inhibitor of SARS coronavirus infection and spread

Human coronavirus NL63 replication is cyclophilin A-dependent and inhibited by non-786 immunosuppressive cyclosporine A-derivatives including Alisporivir

Alisporivir inhibits MERS-and SARS-coronavirus replication in cell culture, but not 790 SARS-coronavirus infection in a mouse model

Pegylated interferon-alpha protects type 1 pneumocytes against SARS 793 coronavirus infection in macaques

Assay Guidance Manual. Bethesda (MD)

The spike 798 glycoprotein of the new coronavirus 2019-nCoV contains a furin-like cleavage site 799 absent in CoV of the same clade

The Proteolytic Regulation of Virus Cell Entry by Furin and Other Proprotein 801 Convertases

Discovery of a novel 803 coronavirus associated with the recent pneumonia outbreak in humans and its potential 804 bat origin

Host cell entry of Middle East respiratory syndrome coronavirus 806 after two-step, furin-mediated activation of the spike protein

Furin cleavage of the SARS coronavirus spike 809 glycoprotein enhances cell-cell fusion but does not affect virion entry

Cathepsin L functionally cleaves the severe acute 814 respiratory syndrome coronavirus class I fusion protein upstream of rather than adjacent 815 to the fusion peptide

Coronavirus cell entry occurs through the endo-/lysosomal pathway in a 818 proteolysis-dependent manner

SARS coronavirus, but not 820 human coronavirus NL63, utilizes cathepsin L to infect ACE2-expressing cells

Identification of a common deletion 823 in the spike protein of SARS-CoV-2. bioRxiv

825 Characterisation of the transcriptome and proteome of SARS-CoV-2 using direct RNA 826 sequencing and tandem mass spectrometry reveals evidence for a cell passage induced 827 in-frame deletion in the spike glycoprotein that removes the furin-like cleavage site

A hypervariable region 830 within the 3' cis-acting element of the murine coronavirus genome is nonessential for 831 RNA synthesis but affects pathogenesis

A conserved RNA pseudoknot in a putative 833 molecular switch domain of the 3'-untranslated region of coronaviruses is only marginally 834 stable

RNA genome conservation and secondary structure in 836 SARS-CoV-2 and SARS-related viruses

lopinavir, emetine, and homoharringtonine inhibit SARS-CoV-2 replication in vitro

SARS-CoV-2 sensitive to type I 841 interferon pretreatment

Isolation, 843 sequence, infectivity and replication kinetics of SARS-CoV-2. bioRxiv

Enhanced 845 isolation of SARS-CoV-2 by TMPRSS2-expressing cells

Apical entry and release 848 of severe acute respiratory syndrome-associated coronavirus in polarized Calu-3 lung 849 epithelial cells

Exogenous ACE2 851 expression allows refractory cell lines to support severe acute respiratory syndrome 852 coronavirus replication

SARS-associated coronavirus replication in cell lines

Discovery 856 of novel human and animal cells infected by the severe acute respiratory syndrome 857 coronavirus by replication-specific multiplex reverse transcription-PCR

Studies on the mechanism of the priming 860 effect of interferon on interferon production by cell cultures exposed to poly(rI)-poly(rC)

Regulation of the interferon system: evidence that Vero cells 863 have a genetic defect in interferon production

A Novel Coronavirus from Patients 865 with Pneumonia in China

Overexpression of 7a, a 870 protein specifically encoded by the severe acute respiratory syndrome coronavirus, 871 induces apoptosis via a caspase-dependent pathway

SARS coronavirus 7a protein 873 blocks cell cycle progression at G0/G1 phase via the cyclin D3/pRb pathway

The ORF7b protein of severe acute 876 respiratory syndrome coronavirus (SARS-CoV) is expressed in virus-infected cells and 877 incorporated into SARS-CoV particles

The 879 SARS-coronavirus-host interactome: identification of cyclophilins as target for pan-880 coronavirus inhibitors

The 8ab protein of SARS-CoV is a 882 luminal ER membrane-associated protein and induces the activation of ATF6

SARS-Coronavirus Open Reading Frame-8b 885 triggers intracellular stress pathways and activates NLRP3 inflammasomes

Molecular evolution of the SARS coronavirus during the course of the 888 SARS epidemic in China

Isolation and 890 characterization of viruses related to the SARS coronavirus from animals in southern 891 China

Severe Acute Respiratory 893 Syndrome (SARS) Coronavirus ORF8 Protein Is Acquired from SARS-Related 894 Coronavirus from Greater Horseshoe Bats through Recombination

Attenuation of 897 replication by a 29 nucleotide deletion in SARS-coronavirus acquired during the early 898 stages of human-to-human transmission

Understanding SARS-CoV-2-Mediated Inflammatory Responses: 900 From Mechanisms to Potential Therapeutic Tools

Genomic characterization of the 902 2019 novel human-pathogenic coronavirus isolated from a patient with atypical 903 pneumonia after visiting Wuhan

Inhibition of 905 novel beta coronavirus replication by a combination of interferon-alpha2b and ribavirin

The antiviral compound 908 remdesivir potently inhibits RNA-dependent RNA polymerase from Middle East 909 respiratory syndrome coronavirus

Structural Basis for the Inhibition 911 of the RNA-Dependent RNA Polymerase from SARS-CoV-2 by Remdesivir. bioRxiv

-2: structural requirements at both nsp12 RdRp and nsp14 915 Exonuclease active-sites

First Case of 917 2019 Novel Coronavirus in the United States

A 919 Rush to Judgment? Rapid Reporting and Dissemination of Results and Its 920 Consequences Regarding the Use of Hydroxychloroquine for COVID-19

In Vitro Antiviral Activity and 923 Projection of Optimized Dosing Design of Hydroxychloroquine for the

Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2)

Chloroquine-resistant malaria

Chloroquine analogues in drug discovery: new directions of uses, 928 mechanisms of actions and toxic manifestations from malaria to multifarious diseases

Snijder 933 EJ. SARS-coronavirus replication/transcription complexes are membrane-protected and 934 need a host factor for activity in vitro

Images from a mock-infected cell are included for comparison

These virus-induced structures 984 accumulated in large clusters in the perinuclear region by 8 h p.i

These regions primarily contained DMVs

Additionally, virus-induced convoluted membranes

At 10 h p.i., vesicle packets (f, j, white asterisks), 989 which seem to arise by fusion of two or more DMVs through their outer membrane, became 990 abundant in the RO regions. (k-r) Examples of virion assembly and release in infected cells

Virus particles budding into membranes of the ERGIC (k-l, o-p, arrowheads). The black 992 arrowheads in the boxed areas highlight captured budding events

Subsequently, virus particles are transported to the plasma membrane which, at 10 h p.i., is 994 surrounded by a large number of released virions (m, q, boxed areas enlarged in n and r

N, nucleus; m, mitochondria; G, Golgi apparatus. Scale bars: 1 µm (a

Assay to screen for compounds that inhibit SARS-CoV-2 replication

Inhibition of SARS-CoV-2 replication (coloured bars) was tested in Vero E6 cells by developing 1000

CPE-reduction assay and evaluating several previously identified inhibitors of SARS-CoV, 1001 which was included for comparison (grey bars)

and (d) pegylated interferon alpha-2. Cell viability was assayed using the CellTiter 96®

Aqueous One Solution cell proliferation assay (MTS assay)

was evaluated in parallel using mock-infected, compound-treated cells. The graphs show the 1006 results of 3 independent experiments, each performed using quadruplicate samples (mean ± 1007 SD are shown)