key: cord-1022951-bp5mjxoa
authors: Lamers, Mart M.; Mykytyn, Anna Z.; Breugem, Tim I.; Wang, Yiquan; Wu, Douglas C.; Riesebosch, Samra; van den Doel, Petra B.; Schipper, Debby; Bestebroer, Theo; Wu, Nicholas C.; Haagmans, Bart L.
title: Human airway cells prevent SARS-CoV-2 multibasic cleavage site cell culture adaptation
date: 2021-01-22
journal: bioRxiv
DOI: 10.1101/2021.01.22.427802
sha: 894addc932fe4a218ef3aea7e33ca3427e3a021a
doc_id: 1022951
cord_uid: bp5mjxoa

Virus propagation methods generally use transformed cell lines to grow viruses from clinical specimens, which may force viruses to rapidly adapt to cell culture conditions, a process facilitated by high viral mutation rates. Upon propagation in VeroE6 cells, SARS-CoV-2 may mutate or delete the multibasic cleavage site (MBCS) in the spike protein that facilitates serine protease-mediated entry into human airway cells. We report that propagating SARS-CoV-2 on the human airway cell line Calu-3 - that expresses serine proteases - prevents MBCS mutations. Similar results were obtained using a human airway organoid-based culture system for SARS-CoV-2 propagation. Thus, in-depth knowledge on the biology of a virus can be used to establish methods to prevent cell culture adaptation.

Pseudoviruses containing MBCS mutations infected human airway organoids poorly 210 (Mykytyn et al., 2021) , indicating that these mutations could be prevented by virus 211 propagation in these cells. To produce stocks in human airway organoids, we differentiated 212 the organoids at 2D in transwell inserts at air-liquid interface for twelve weeks as described 213 before ( Fig 7A) (Mykytyn et al., 2021) . Apical cells, including ciliated cells, in these cultures 214 expressed TMPRSS2 as shown by immunohistochemistry (Fig 7B) . To produce viral stocks, 215 stock ( Fig 1A) . After a two-hour incubation, cells were washed three times to remove 217 unbound particles. On day two to five post-infection, apical washes were collected and 218 stored at 4°C. During virus collections, bound virus particles were released from cells by 219 pipetting directly on the cell layer. Virus collections from day two and day three (d2+3), and 220 day four and day five (d4+5) were pooled, centrifuged, and filtered to remove debris, dead 221 cells and mucus. In these cultures, ciliated cells were infected, as shown by confocal imaging 222 at day three post infection ( Fig 7C) . At day five, cultures exhibited widespread infection (Fig  223   7D ) and significant cytopathic effects including loss of ciliated cells (Fig 7D-E) and syncytium 224 formation ( Fig 7E) . To remove cytokines that could interfere in downstream experiments 225 (such as interferons), we exchanged the medium in the filtered virus collections three times 226 using an Amicon Ultra-15 column (100 kDa cutoff). The resulting titers from the d2+3 and 227 d4+5 stocks were 5.64 x 10 5 and 1.00 x 10 7 TCID50/ml, respectively, indicating that high titer 228 virus stocks can be made in human airway organoids. Sequencing demonstrated that the 229 high titer organoid stock (d4+5) had a 98.9% WT spike sequence, without multibasic 230 cleavage site mutations and the S686G mutation at only 1.1% (Fig 8A-B ). In accordance, the 231 Organoid P3 virus produced small plaques ( Fig 8C) . No major variants were detected in the 232 rest of the genome ( Fig 8D) . Next, we investigated S1/S2 cleavage of the VeroE6 P2, P3, 233

Calu-3 P3 and Organoid P3 virus stocks by immunoblot (Fig 8E) . The non-adapted Calu-3 and 234 organoid stocks were >85% cleaved, while the VeroE6 P2 and P3 stocks were 71.2% and 33% 235 cleaved, respectively ( Figure 8F ). The findings support that the Calu-3 and organoid stocks 236 are non-adapted and indicate that in vivo the S1/S2 cleavage takes place in the producing 237 cell. The rapid loss of the SARS-CoV-2 MBCS in cell culture has underlined that some in vitro 242 propagation systems may fail to model key aspects of the viral life cycle. As these mutations 243 directly affect the relevance and translatability of all laboratory SARS-CoV-2 experiments, it 244 is pivotal to sort out exactly why these occur in order to prevent them. We and others have 245 previously reported that the SARS-CoV-2 MBCS enhances serine protease-mediated entry, 246 the dominant entry pathway in human airway cells (Hoffmann et al., 2020; Mykytyn et al., 247 2021 which naturally express serine proteases, also prevented cell culture adaptation. Similar 253 results were obtained using a human airway organoid-based culture system for SARS-CoV-2 254

propagation 255

Our study shows that SARS-CoV-2 rapidly adapts to VeroE6 cell culture. Therefore, deep-257 sequencing of viral stocks, which offers a thorough analysis beyond the consensus 258 sequence, is essential. As none of the MBCS mutations co-occurred, consensus sequence 259 logos of culture adapted stocks were often WT, while actually only 10-20% of viral reads 260 contained the WT sequence. Therefore, besides reporting the consensus sequence SARS-261

CoV-2 studies should preferably also report the percentage of WT reads in the MBCS. The 262 first adaptation to occur in our stocks was the S686G mutation, which lies directly adjacent 263 to the MBCS and decreased Calu-3 infectivity, fusogenicity and S1/S2 cleavage, but not as severely as MBCS mutations. Interestingly, this mutation is rapidly positively selected in 265 ferrets (Richard et al., 2020), and also transmitted, suggesting that there are key differences 266 in transmission between humans and ferrets. Alternatively, it is possible that S686G 267 optimizes cleavage by a specific ferret protease. This indicates that the MBCS is an adaptation to serine proteases and that the serine 289 protease-mediated entry pathway is used for entry in vivo. This is in agreement with our 290 earlier observations that SARS-CoV-2 enters using serine proteases on airway organoids 291 (Mykytyn et al., 2021) and that MBCS mutant pseudoviruses could not efficiently infect 292 these cells. Low infectivity of MBCS mutants on the airway cell line Calu-3 was also noted by As new SARS-CoV-2 strains are emerging now and will continue to emerge for as long as 302 SARS-CoV-2 circulates in humans, there is a need to develop propagation systems that will 303 preserve genetic stability for any given SARS-CoV-2 mutant originating from a human 304 respiratory sample. The closer a culture system mimics the human respiratory tract the less 305 likely it is that a SARS-CoV-2 isolate will adapt. Therefore, we developed a human airway 306 organoid model for SARS-CoV-2 propagation (Figure supplement 8) . This model allows high 307 titer SARS-CoV-2 production and was most successful in removing MBCS mutations. In the 308 future we expect that organoid-based systems are likely to replace transformed cell lines 309

when producing viral stocks. The self-renewing capacity of organoids allows labs to share 310 organoid lines, allowing a level of reproducibility similar to that of transformed cell lines.

Organoids can be grown from a wide range of organs and species to best model the in vivo 312 environment of a particular virus. 313

In conclusion, this study shows that SARS-CoV-2 rapidly adapts to VeroE6 cell culture 315

propagation and that this can be prevented by using cell lines with an active serine 316 protease-mediated entry pathway (VeroE6-TMPRSS2 or Calu-3). Alternatively, a 2D airway anonymized and non-traceable to the donor. In this study we used organoids from one 375 donor, from which bronchial and bronchiolar organoids were grown. Differentiation of 376 human airway organoids at air-liquid interface was performed as described before (Lamers, 377 Beumer, et al., 2020). Cultures were differentiated for 8-12 weeks at air-liquid interface. 378

To produce stocks in human airway organoids, we differentiated the bronchial organoids in 381 transwell inserts at air-liquid interface for twelve weeks. A total of twelve 12 mm transwell 382 inserts were washed three times in AdDF+++ before inoculation at the apical side at a MOI {snp_file}. Sequence logo were generated with logomaker (PMC7141850) using a custom 528 python script. Plotting of mutation frequencies was done using R and ggplot2 (Hadley, 529 2016 Quantification of S1 cleavage from four independent pseudovirus productions. (C) Analysis 566 of S1/S2 cleavage by multiplex S1 (red) and S2 (green) immunoblot of SARS-CoV-2 S (WT) 567 and S686G mutant pseudoviruses. S0 indicates uncleaved spike; S1 indicates the S1 domain 568 of cleaved spike; VSV-N indicates VSV nucleoprotein (production control). 

A Single Amino Acid Change in the Marburg Virus Glycoprotein 681 Arises during Serial Cell Culture Passages and Attenuates the Virus in a Macaque 682 Model of Disease. mSphere

A Crisp(r) New Perspective on SARS-CoV-2 Biology

The 686 spike glycoprotein of the new coronavirus 2019-nCoV contains a furin-like cleavage 687 site absent in CoV of the same clade

Identification of Required Host Factors for SARS-CoV-2 Infection in 691 Human Cells

Characterisation of the transcriptome and proteome of 694 SARS-CoV-2 reveals a cell passage induced in-frame deletion of the furin-like 695 cleavage site from the spike glycoprotein

Alternatives to the use of fetal bovine serum: Serum-free cell 698 culture

Ggplot2

A Multibasic Cleavage Site in the 702

Spike Protein of SARS-CoV-2 Is Essential for Infection of Human Lung Cells

Coronavirus Spike Protein and Tropism 705 Changes

Furin Cleavage Site Is Key to SARS-CoV-2 Pathogenesis

Human Lung Stem Cell-Based Alveolospheres Provide Insights into SARS-CoV-711 2-Mediated Interferon Responses and Pneumocyte Dysfunction

SARS-CoV-2 growth, furin-cleavage-site adaptation and 715 neutralization using serum from acutely infected hospitalized COVID-19 patients

SARS-CoV-2 productively infects human gut enterocytes

An organoid-derived bronchioalveolar model for SARS-CoV-722 2 infection of human alveolar type II-like cells

Attenuated SARS-CoV-2 variants with deletions at the S1/S2 junction

Identification of Common 728 Deletions in the Spike Protein of Severe Acute Respiratory Syndrome Coronavirus 2

Host cell proteases: Critical determinants of 731 coronavirus tropism and pathogenesis

SARS-CoV-2 entry into human airway organoids is serine 735 protease-mediated and facilitated by the multibasic cleavage site

Human embryonic lung epithelial tips are multipotent progenitors that can 739 be expanded in vitro as long-term self-renewing organoids

SARS-coronavirus-2 replication in Vero E6 cells: replication 743 kinetics, rapid adaptation and cytopathology

The genome landscape of the african green monkey kidney-derived vero cell line

Rapid SARS-CoV-2 whole-genome 749 sequencing and analysis for informed public health decision-making in the 750 Netherlands

Comparative pathogenesis of COVID-19, MERS, and SARS in 753 a nonhuman primate model

Long-term expanding human airway organoids for disease 756 modeling

Progenitor identification and SARS-CoV-2 infection in human distal lung 759 organoids

Single Lgr5 stem cells build crypt-villus structures in vitro without 762 a mesenchymal niche

A novel protease obtained 764 from FBS-containing culture supernatant, that processes single chain form 765 hepatocyte growth factor to two chain form in serum-free culture

Nonreplicating vaccinia vector efficiently expresses 768 recombinant genes

Cell culture-selected substitutions in influenza A(H3N2) neuraminidase 772 affect drug susceptibility assessment

Deep-sequencing 775 of Marburg virus genome during sequential mouse passaging and cell-culture 776 adaptation reveals extensive changes over time

Genome-wide CRISPR screen reveals host genes that regulate 780 SARS-CoV-2 infection

Natural transmission of bat-like SARS-CoV-2PRRA variants in COVID-19 patients

Three-785 Dimensional Human Alveolar Stem Cell Culture Models Reveal Infection Response to 786 SARS-CoV-2

Infection of bat 788 and human intestinal organoids by SARS-CoV-2

A Novel 791 Coronavirus from Patients with Pneumonia in China

Figure supplement 8. Schematic workflow for the production of SARS-CoV-2 stocks on 2D air-liquid interface differentiated airway organoids

Step 1. 3D self-renewing airway organoids are grown from human lung tissue. Next, these are dissociated to single cells and differentiated at air-liquid interface for 4-12 weeks

Differentiated cultures are infected at a multiplicity of infection of 0.05 and washed daily for 5 days. The washes from day 2-5 are collected and stored at 4°C

Virus collections are cleared by centrifugation and filtered to remove debris larger than 0.45 μm. Next, the medium is exchanged three times using Amicon columns to remove cytokines and debris smaller than 100 kDa

Stocks can be characterized using plaque assays, Sanger sequencing and deep-sequencing