Furin cleavage site in the SARS-CoV-2 coronavirus glycoprotein

coronavirus SpikeThe spike glycoprotein of the newly emerged SARS-CoV-2 contains a potential cleavage site for furin proteases. This observation has implications for the zoonotic origin of the virus and its epidemic spread in China.

The membrane of coronaviruses harbors a trimeric transmembrane spike (S) glycoprotein (pictured) which is essential for entry of virus particles into the cell. The S protein contains two functional domains: a receptor binding domain, and a second domain which contains sequences that mediate fusion of the viral and cell membranes. The S glycoprotein must be cleaved by cell proteases to enable exposure of the fusion sequences and hence is needed for cell entry.

The nature of the cell protease that cleaves the S glycoprotein varies according to the coronavirus. The MERS-CoV S glycoprotein contains a furin cleavage site and is probably processed by these intracellular proteases during exit from the cell. The virus particles are therefore ready for entry into the next cell. In contrast, the SARS-CoV S glycoprotein is uncleaved upon virus release from cells; it is likely cleaved during virus entry into a cell.

Proteolytic cleavage of the S glycoprotein can determine whether the virus can cross species, e.g. from bats to humans. For example, the S glycoprotein from a MERS-like CoV from Ugandan bats can bind to human cells but cannot mediate virus entry. However, if the protease trypsin is included during infection, the S glycoprotein is cleaved and virus entry takes place. This observation demonstrates that cleavage of the S glycoprotein is a barrier to zoonotic coronavirus transmission.

Examination of the protein sequence of the S glycoprotein of SARS-CoV-2 reveals the presence of a furin cleavage sequence (PRRARS|V). The CoV with the highest nucleotide sequence homology, isolated from a bat in Yunnan in 2013 (RaTG-13), does not have the furin cleavage sequence. Because furin proteases are abundant in the respiratory tract, it is possible that SARS-CoV-2 S glycoprotein is cleaved upon exit from epithelial cells and consequently can efficiently infect other cells. In contrast, the highly related bat CoV RaTG-13 does not have the furin cleavage site.

Whether or not the furin cleavage site within the S glycoprotein of SARS-CoV-2 is actually cleaved remains to be determined. Meanwhile, it is possible that the insertion of a furin cleavage site allowed a bat CoV to gain the ability to infect humans. The furin cleavage site might have been acquired by recombination with another virus possessing that site. This event could have happened thousands of years ago, or weeks ago. Upon introduction into a human – likely in an outdoor meat market – the virus began its epidemic spread.

Furins are also known to control infection by avian influenza A viruses, in which cleavage of the HA glycoprotein is needed for entry into the cell. Low-pathogenic avian influenza viruses contain a single basic amino acid at the cleavage site in the HA protein which is cleaved by proteases that are restricted to the respiratory tract. Insertion of a furin cleavage site in the HA of highly pathogenic avian H5N1 influenza viruses leads to replication in multiple tissues and higher pathogenicity, due to the distribution of furins in multiple tissues.

Acquisition of the furin cleavage site might be viewed as a ‘gain of function’ that enabled a bat CoV to jump into humans and begin its current epidemic spread.

{ 15 comments… add one }
  • Ben 14 February 2020, 11:08 am

    This is fascinating, in particular the difference between SARS (2003) and SARS-CoV-2 (2019). I have a couple questions, but I’ll summarize what I understood first, in case I’m not following.

    They explain that there are two necessary cleavages that must occur for infection. First, the S protein must be split into S1/S2, because those two sub-units mediate the distinct tasks of attachment and entry, respectively. Then the S2 must be cleaved at the “S2-prime” site, splitting the fusion peptide (FP) from the “internal fusion peptide” (IFP), because “it is likely both… participate in the viral entry process.” About the S2 cleaving, they say: “The furin-like S2′ cleavage site … is identical between the 2019-nCoV and SARS-CoV”.

    About the S1/S2 cleaving they say that it “exhibits different motifs” among CoVs, and point to a “site1 & site2” on SARS-CoV-2. If I understand correctly, they say that site 2, which also appears on both SARS-CoV-2 and SARS-CoV, is known to be left uncleaved on viral egress and therefore must somehow be cleaved prior to viral attachment in the absence of another S1/S2 cleavage mechanism. “Site 1” is the new furin-like cleavage site, and it does not appear in SARS-CoV (they do say that other CoV, e.g., MERS-CoV) “harbor[] furin-like [S1/S2] cleavage sites”). This region on SARS-CoV-2 “contains 12 additional nucleotides upstream” to site 1, “which corresponds to a canonical furin-like cleavage site”.

    If site 1 is cleaved on SARS-CoV-2 S1/S2 on viral exit, then there is no need for further cleaving on entry, so this is the potential “gain-of-function” for the new virus. This potential for an infectivity gain — due to the cleaving of S1/S2 on exit, as compared to the SARS necessity for cleaving on entry — seems to be the most important idea here. Obviously there are lots of factors that go into “infectivity” (and maybe I’m improperly associating within-host infectivity of cells with the observed epidemiological-scale infectivity), but if this effect is important, then have two questions:

    (1) If MERS-CoV also had this “site 1” furin-like cleavage site, why did it have a similarly low infectivity to SARS?

    (2) The “site 1” aa sequence in SARS-CoV-2019 is different than those of the other CoV possessing a furin-like cleavage “site 1”. Is it possible to find this specific sequence in another viruses, which could implicate them as a co-infection partner that led to this mutation?

  • Wang 15 February 2020, 10:23 pm

    The putative furin cleavage site as listed in the blog exists in both COVID-19 and bat strain RaTG13.

  • Etienne Decroly 16 February 2020, 1:10 pm

    No this site is absent in the RaTG13 and in the pangolin sequence.
    Our observations is also present in the excellent paper on the Spike cryo-EM from MCLeLLan lab.
    “Cryo-EM Structure of the 2019-nCoV Spike in the Prefusion Conformation
    Daniel Wrapp, Nianshuang Wang, Kizzmekia S Corbett, Jory A Goldsmith, Ching-Lin Hsieh, Olubukola Abiona, Barney S Graham, View ORCID ProfileJason S McLellan
    doi: https://doi.org/10.1101/2020.02.11.944462

  • Dr.Boudemagh 16 February 2020, 2:02 pm

    Ballistol is a powerfull anti viral product:
    https://algerietouteheure.com/2020/02/12/coronavirus-ballistol /

  • Virus Watcher 16 February 2020, 2:34 pm

    One well known human coronavirus, HCoV-NL63 also uses ACE2 for cell entry.

    Clinical spectrum looks like COVID19 :

    “mild to moderate upper respiratory tract infections, severe lower respiratory tract infection.”

    So its not necessarily ACE2-affinity and the SARS-spike protein that makes the virus “new and dangerous” ?
    Maybe COVID19 could (hopefully) be not that “SARS-like” sever after all ?
    Maybe there is 10x more, but completely unreported very mild “common cold” type of COVID19 illness prevalence in China ?

    Whats you virologists take on this ?

  • Wang 16 February 2020, 3:27 pm

    Sorry for the mistake I made in my alignment. PRRA is indeed unique to COVID-19. My question is how important the acquired cleavage site is for viral infection. If it is important for viral exit, then the released viral particles would have the truncated spike protein missing S1. If it is important for viral entry, how do you reconcile with the fact that other coronaviruses such as SARS would be able to infect without this site?

  • Z. Galen Wo 18 February 2020, 8:07 pm

    “Sequence Analysis Indicates that 2019-nCoV Virus Contains a Putative Furin Cleavage Site at the Boundary of S1 and S2 Domains of Spike Protein”

  • Minae 19 February 2020, 12:11 am

    “a furin cleavage sequence (PRRARS|V)” should be PRRAR|SV. i believe that this is just a typo.

  • Carsten 20 February 2020, 12:29 pm

    “Upon introduction into a human – likely in an outdoor meat market – the virus began its epidemic spread.”
    Please stop spreading unfounded rumors. (This goes both ways!) There is no evidence that the meat market was the origin of the outbreak. Neither were the first cases related to the market (Huang, Lancet, 2020) nor were any animals found that could have been the source.

  • John Harris 21 February 2020, 10:24 pm

    E. Braun, D. Sauter. Furin-mediated protein processing in infectious diseases and cancer, August, 2019.


    Current review of therapeutic inhibition of furin.

  • John Harris 22 February 2020, 1:37 pm

    A search turned up four furin inhibitors that have been identified and extracted from traditional medicinal plants in China and India. Some have been tested for antiviral activity against Dengue. All four are flavonoids.

    The appeal of plants and extracts with a history of use in traditional medicines is that there is experience – sometimes long experience — in observing their impacts on human patients. Few of the perhaps 150 agents that have been identified as furin inhibitors have ever been in clinical trials.

    In 2010 at the University of Ottawa, three flavonoid furin inhibitors were extracted from the traditional medicinal plant, Oroxylum Indicum: baicalein, chrysin and oroxylin. Oroxylin was the most potent inhibitor of furin with a k sub i of 5 µM. All three inhibit other PCs in addition to furin.


    Luteolin is another flavonoid with furin inhibitory activity. It is found in many types of plants including medicinal herbs. It was tested as a furin inhibitor for its antiviral activity against Dengue. The K sub i was in the high fifties, as recounted here:


  • Mario Garrett 22 February 2020, 2:19 pm

    Reply to Carston. The Huang et al (2020) paper in the Lancet makes EXPLICIT reference to the seafood market.

    “We report the epidemiological, clinical, laboratory, and radiological characteristics, treatment, and clinical outcomes of 41 laboratory-confirmed cases infected with 2019-nCoV. 27 (66%) of 41 patients had a history of direct exposure to the Huanan seafood market.”

  • Jon Perry 23 February 2020, 5:22 pm

    I’m doing an animation that will show how we currently think cell entry works for COVID-19. Is it safe to say we currently think endocytosis is needed (as was reported for SARS), or might this furin site allow direct fusion at the cell membrane?

    Here’s the paper claiming endocytosis was needed for SARS: https://www.nature.com/articles/cr200815

Leave a Comment