COVID-19: Avoiding culture shock with the SARS-CoV-2 spike protein

When culturing SARS-CoV-2 in the laboratory it is vital to avoid deletions in the gene for the spike protein that could affect the interpretation of experiments.
  1. Benjamin G Hale  Is a corresponding author
  1. Institute of Medical Virology, University of Zurich, Switzerland

Researchers across the globe are frantically trying to understand the biology of different SARS-CoV-2 variants to develop new therapeutics and bring the COVID-19 pandemic to an end. The process begins in the laboratory, with painstaking experiments to judge whether a drug can stop SARS-CoV-2 from replicating in cells in a dish, or whether a vaccine can stop an animal from getting sick. But such experiments require copious amounts of SARS-CoV-2, and the authenticity of that virus stock (which is usually cultured in animal cell lines) is paramount to ensure the validity of test results.

The standard 'go-to' cells for producing stocks of SARS-CoV-2 are derived from the Vero lineage (which are kidney cells isolated from an African green monkey almost 60 years ago). These cells are highly susceptible to viruses due to the absence of type I interferon cytokines, an important group of signaling proteins released by cells in the presence of viruses. Vero cells are a popular choice for isolating and propagating many viruses because they are well characterized and easy to maintain, they adhere to laboratory dishes, and they show visible structural changes when infected. However, like viruses circulating naturally in human populations, viruses grown in the laboratory also have a tendency to change and adapt to the environment that they are in.

An early study, since replicated by others, found that SARS-CoV-2 stocks cultured in Vero-derived cells often harbor mutations or deletions in the spike gene (Lau et al., 2020). The deletions remove an important region on the spike protein, called the multibasic cleavage site, which affects the ability of the virus to infect human airway cells (Hoffmann et al., 2020; Mykytyn et al., 2021; Pohl et al., 2021). Such viruses therefore do not behave like authentic SARS-CoV-2 in several aspects: they are less pathogenic (Johnson et al., 2021; Lau et al., 2020); they do not transmit (Peacock et al., 2021); and they exhibit altered sensitivities to inhibition by antiviral interferon-stimulated genes and patient antibodies (Johnson et al., 2021; Winstone et al., 2021).

These characteristics can make it difficult to interpret laboratory experiments that use Vero-produced viruses to assess the effectiveness of experimental drugs or vaccines. The ability of inactivated SARS-CoV-2 vaccines made in Vero cells to stimulate correct antibody responses might also be partially compromised, although this has not been formally assessed (Gao et al., 2020; Wang et al., 2020). Now, in eLife, Bart Haagmans and colleagues of the Erasmus Medical Center in Rotterdam and the University of Illinois at Urbana-Champaign – including Mart Lamers as first author – report simple methods for producing SARS-CoV-2 stocks in human cells that prevent such mutations and deletions in the gene for the spike protein (Figure 1; Lamers et al., 2021).

Culturing SARS-CoV-2 in Vero cells and human cells.

Schematic representation of a human SARS-CoV-2 patient sample passaged using the Vero cell line (which was isolated from an African green monkey almost 60 years ago; top), or the Calu-3 cell line (which is a human cell line; bottom). Vero cells are deficient in a serine protease called TMPRSS2 that is needed for the virus to enter cells at the plasma membrane; however, certain SARS-CoV-2 variants with spike deletions (shown in red) can enter Vero cells via a different pathway, so these variants are artificially selected for and come to dominate the virus population. The pathogenicity, transmission properties, and sensitivity to antiviral drugs and antibodies of the variants are different to those of the wild-type virus. Calu-3 cells are not deficient in TMPRSS2, so the authenticity of the spike gene is maintained, and studies with such virus stocks more faithfully recapitulate human virus biology.

Figure created with

First, Lamers et al. used a method called deep sequencing to confirm that repeated culture of SARS-CoV-2 in Vero cells leads to an increase in viral genomes with mutations or deletions in the important region of the spike gene. Less sensitive sequencing methods, or a reliance on 'consensus sequences', can leave researchers with the false impression that such deletions are absent.

Lamers et al. then pinpointed how these spike deletions give certain SARS-CoV-2 variants a replicative advantage in Vero cells, which enables them to dominate the virus population. Vero cells are deficient in a serine protease that is needed for SARS-CoV-2 to enter human airway cells at the plasma membrane: however, SARS-CoV-2 variants with the spike gene deletions exploit a different pathway (called endocytosis) to enter Vero cells. It seems the ability of the variants to take advantage of this second entry route allows them to dominate when Vero cells are used.

Lamers et al. next asked whether a human-airway cell line, Calu-3, might be a better alternative to culture SARS-CoV-2, given that this cell line possesses the necessary protease (Mykytyn et al., 2021). Indeed, they found that repeated culture of SARS-CoV-2 in Calu-3 cells (or Vero cells engineered to express the serine protease) prevented the accumulation of mutations and deletions in the SARS-CoV-2 spike. Moreover, Calu-3 cells were as good as Vero cells at producing the large amounts of virus necessary for further experiments. Thus, the authentic nature of SARS-CoV-2 stocks made in this way, as assessed by deep sequencing, will give confidence to the interpretation of subsequent experiments.

The results from this study make a compelling case for SARS-CoV-2 researchers to thoroughly characterize the genomic sequences of the virus stocks they produce (and to avoid consensus sequencing). Furthermore, researchers should consider the cells, the growth media and the additives used for virus production to prevent artefactual SARS-CoV-2 culture adaptations that might impact the evaluation of drug or vaccine effectiveness.

Cell culture systems to propagate viruses are the linchpin of virology, yet they have not really changed since the beginnings of the field. The application of modern technological advances, such as rational gene editing of cells or the use of in vivo-like organoid tissue models, promises to transform this critical aspect of virology and should allow researchers to update their methods to maintain experimental authenticity.


Article and author information

Author details

  1. Benjamin G Hale

    Benjamin G Hale is in the Institute of Medical Virology, University of Zurich, Zurich, Switzerland

    For correspondence
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-3891-9480

Publication history

  1. Version of Record published: May 18, 2021 (version 1)


© 2021, Hale

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.


  • 1,951
    Page views
  • 146
  • 3

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Benjamin G Hale
COVID-19: Avoiding culture shock with the SARS-CoV-2 spike protein
eLife 10:e69496.
  1. Further reading

Further reading

    1. Microbiology and Infectious Disease
    Xiao Lei, Daniel Gonçalves-Carneiro ... Paul D Bieniasz
    Research Article Updated

    The encapsidation of HIV-1 gRNA into virions is enabled by the binding of the nucleocapsid (NC) domain of the HIV-1 Gag polyprotein to the structured viral RNA packaging signal (Ψ) at the 5’ end of the viral genome. However, the subcellular location and oligomeric status of Gag during the initial Gag-Ψ encounter remain uncertain. Domains other than NC, such as capsid (CA), may therefore indirectly affect RNA recognition. To investigate the contribution of Gag domains to Ψ recognition in a cellular environment, we performed protein-protein crosslinking and protein-RNA crosslinking immunoprecipitation coupled with sequencing (CLIP-seq) experiments. We demonstrate that NC alone does not bind specifically to Ψ in living cells, whereas full-length Gag and a CANC subdomain bind to Ψ with high specificity. Perturbation of the Ψ RNA structure or NC zinc fingers affected CANC:Ψ binding specificity. Notably, CANC variants with substitutions that disrupt CA:CA dimer, trimer, or hexamer interfaces in the immature Gag lattice also affected RNA binding, and mutants that were unable to assemble a nascent Gag lattice were unable to specifically bind to Ψ. Artificially multimerized NC domains did not specifically bind Ψ. CA variants with substitutions in inositol phosphate coordinating residues that prevent CA hexamerization were also deficient in Ψ binding and second-site revertant mutants that restored CA assembly also restored specific binding to Ψ. Overall, these data indicate that the correct assembly of a nascent immature CA lattice is required for the specific interaction between Gag and Ψ in cells.

    1. Immunology and Inflammation
    2. Microbiology and Infectious Disease
    Sindhu Vangeti, Sara Falck-Jones ... Anna Smed-Sörensen
    Research Article

    During respiratory viral infections, the precise roles of monocytes and dendritic cells (DCs) in the nasopharynx in limiting infection and influencing disease severity are incompletely described. We studied circulating and nasopharyngeal monocytes and DCs in healthy controls (HC) and in patients with mild to moderate infections (primarily influenza A virus, IAV). As compared to HCs, patients with acute IAV infection displayed reduced DC but increased intermediate monocytes frequencies in blood, and an accumulation of most monocyte and DC subsets in the nasopharynx. IAV patients had more mature monocytes and DCs in the nasopharynx, and higher levels of TNFα, IL-6 and IFNα in plasma and the nasopharynx than HCs. In blood, monocytes were the most frequent cellular source of TNFα during IAV infection and remained responsive to additional stimulation with TLR7/8L. Immune responses in older patients skewed towards increased monocyte frequencies rather than DCs, suggesting a contributory role for monocytes in disease severity. In patients with other respiratory virus infections, we observed changes in monocyte and DC frequencies in the nasopharynx distinct from IAV patients, while differences in blood were more similar across infection groups. Using SomaScan, a high-throughput aptamer-based assay to study proteomic changes between patients and HCs, we found differential expression of innate immunity-related proteins in plasma and nasopharyngeal secretions of IAV and SARS-CoV-2 patients. Together, our findings demonstrate tissue-specific and pathogen-specific patterns of monocyte and DC function during human respiratory viral infections and highlight the importance of comparative investigations in blood and the nasopharynx.