The Emergence of the Spike Furin Cleavage Site in SARS-CoV-2

Abstract

Compared with other SARS-related coronaviruses (SARSr-CoVs), SARS-CoV-2 possesses a unique furin cleavage site (FCS) in its spike. This has stimulated discussion pertaining to the origin of SARS-CoV-2 because the FCS has been observed to be under strong selective pressure in humans and confers the enhanced ability to infect some cell types and induce cell-cell fusion. Furthermore, scientists have demonstrated interest in studying novel cleavage sites by introducing them into SARSr-CoVs. We review what is known about the SARS-CoV-2 FCS in the context of its pathogenesis, origin, and how future wildlife coronavirus sampling may alter the interpretation of existing data.

Keywords: COVID-19; coronavirus; furin cleavage site; virology.

Fig. 1.

Phylogenetic tree of the spike gene (A) and alignments of the S1/S2 region of the FCS by codon sequences (B). A codon alignment of the spike sequences was generated using PRANK version 170427 (Löytynoja 2021) and was then translated into amino acid residues for visualization. A phylogenetic tree was estimated from the codon alignment using IQTree version 1.6.12 with the options “-bb 1000 -alrt 1000” (Minh et al. 2020). The consensus tree from IQTree was rooted at midpoint and visualized using FigTree version 1.4.4 (http://tree.bio.ed.ac.uk/software/figtree/). Ultra-bootstrap values less than 100 are labeled in the tree. In Clade 1 and Clade 2, we collapsed the entries, which contained identical amino acid sequences in the 1,100–1,150 region of the full amino acid alignment, and arbitrarily selected SZ3 and WIV1 as the representatives of Clade 1 and Clade 2, respectively. (C) To focus on the entries that led to the indels in the S1/S2 region in the full alignment, we collapsed similar entries and subsequently used this visualization to facilitate the comparison of alignments in figure 2. The alignments were visualized using AliView version 1.26 (Larsson 2014). The last arginine (R) residue in the PRRAR motif from Wuhan-Hu-1 is indicated by the red arrows above each alignment. The FCS motif is also indicated by a box and label.

Fig. 2.

Amino acid alignments of subsets of SARSr-CoV genomes. The subsets were created by excluding select entries to illustrate how sensitive alignments in the spike S1/S2 region can be to lineage sampling. We examined three subsets: (A) without Wuhan-Hu-1, (B) without Wuhan-Hu-1 and RacCS203, and (C) without Wuhan-Hu-1, RacCS203, RaTG15, BM48-31, and RmYN02. The codon sequences from the subsets were aligned using PRANK, translated into amino acid, and then visualized using AliView. A visual comparison between each of the subsets and the full sequence alignment (at the top of each panel; taken from fig. 1C) shows that the alignment in the S1/S2 region is sensitive to including one or a few samples with a different amino acid sequence in the region.