The spike protein is the multimeric engine of coronavirus entry. Its larger S1 subunit carries the receptor-binding machinery that latches onto ACE2 on the surface of human cells, while the smaller S2 subunit is the fusogenic core that, once triggered, refolds to drive merger of the viral envelope with the host cell membrane. Between these two subunits sits a polybasic cleavage motif, the RRAR sequence that distinguishes SARS-CoV-2 from many related coronaviruses and that is recognised and cut by furin and other proprotein convertases within the producing cell. This cleavage event, termed priming, is not decorative: it pre-activates the spike so that the virus is poised for rapid membrane fusion upon receptor engagement, a property widely credited with contributing to efficient human-to-human transmission.

What makes the region around the cleavage site chemically complex is that the spike is one of the most heavily glycosylated proteins known in virology, carrying 22 N-linked glycans per protomer. These carbohydrates serve the canonical functions of shielding immunodominant epitopes and escorting the protein through the folding and quality-control machinery of the endoplasmic reticulum and Golgi apparatus. But several of the glycosylation sequons sit remarkably close to the S1/S2 junction, raising a long-standing question: do these sugars physically modulate access of furin to the polybasic motif, and if so, does removing them cripple the entry functions of the spike?

To answer this, Muriel Lavie, Emma Louvet, Jules Patron, Xavier Hanoulle and Sandrine Belouzard systematically mutated the asparagine residues at positions 603, 616 and 657, individually and in combination, thereby ablating the corresponding N-glycans without otherwise altering the spike sequence. They then measured three tightly linked readouts in cell systems: the efficiency with which the spike was cleaved into its S1 and S2 subunits, the capacity of the processed spike to mediate cell–cell fusion, and its ability to support entry of pseudotyped particles into target cells expressing ACE2. The comparative dimension of the design is what gives the work its evolutionary punch, because the same mutant panel was assessed on two distinct genetic backgrounds: the D614 variant and its G614 descendant that rapidly became globally dominant in 2020, and the XBB.1.5 Omicron variant that circulated widely in 2023.

The first major finding concerns a compensatory basic residue. When the three glycans were removed simultaneously, cleavage of the spike did not simply collapse; instead, the authors found that processing of the glycan-deficient spike became dependent on arginine at position 634, a basic residue lying within the cleavage-proximal region. In other words, the loss of the sugar shield exposed a functional redundancy in the local sequence: with the glycans gone, the enzymatic machinery that processes the spike leans on the R634 residue to achieve productive cutting. This interplay between carbohydrate occupancy and a specific cationic side chain underscores how finely balanced the protease-recognition landscape around the S1/S2 junction really is, and it suggests that the local electrostatic and steric environment — not merely the canonical RRAR motif — determines whether priming occurs efficiently.

The second finding is that on the D614 and G614 backgrounds, the glycans at 603, 616 and 657 and the R634 residue are each required for full fusion and entry competence. Mutating these positions on those variants produced measurable deficits in both cell–cell fusion assays and pseudotype entry, indicating that the glycan cluster and the compensatory basic residue act in concert to maintain the entry functions that follow successful cleavage. The consistency across D614 and G614 is notable because the G614 substitution itself is known to stabilise the spike and increase its incorporation into virions; the new data indicate that even on that stabilised backbone, the cleavage-site glycans remain non-negotiable structural elements.

The picture changes, however, when the same experiments are repeated on XBB.1.5. In the Omicron background, mutations of the three glycosylation sites and of R634 had a milder impact on spike processing and entry functions than they did in the ancestral variants. The authors conclude from this pattern that the functional roles of N603, N616, N657 and R634 have evolved alongside the circulating variants. This is a subtle but important point for evolutionary virology: Omicron lineages accumulated dozens of changes across the spike, including extensive remodeling of the receptor-binding domain and adjustments near the S1/S2 region, and the new results show that these changes have partially uncoupled entry competence from the glycan-and-arginine dependency that governed earlier forms of the virus.

Mechanistically, the results fit a model in which the upstream glycans act as steric modulators of furin accessibility. Bulky N-linked carbohydrates projecting from the peptide backbone around residues 603 to 657 can either occlude or funnel the protease toward the polybasic motif, depending on their precise position and occupancy. Removing them appears to perturb the local conformation in a way that shifts the burden of recognition onto nearby basic residues such as R634. At the same time, the glycans’ contribution to fusion and entry in D614 and G614 suggests they do more than merely regulate cleavage access; they may help stabilise the metastable pre-fusion architecture that the S2 subunit must maintain until the moment of triggering. Their reduced importance in XBB.1.5 implies that other sequence features of the Omicron spike have taken over some of that stabilising or presenting role.

The work also carries implications that extend beyond SARS-CoV-2 itself. The dependence of proteolytic priming on glycan context echoes findings in other enveloped viruses, where carbohydrates adjacent to cleavage sites modulate activation of fusion proteins. For coronavirus surveillance, the message is that mutations at N-glycosylation sequons near the S1/S2 boundary should not be dismissed as neutral surface noise; depending on the genetic background in which they appear, they can reshape how efficiently the spike is primed and how readily the virus enters cells. Conversely, the variant-dependent behaviour means that any prediction of a glycan mutation’s effect must account for the rest of the spike sequence — a caution against interpreting such changes in isolation.

For therapeutic and vaccine design, the study adds a further layer of nuance. Glycan-shielded regions of the spike are notoriously difficult targets for antibody-based countermeasures, but understanding which sugars are functionally load-bearing, and how their roles shift between variants, can inform the design of stabilized immunogens and small molecules that target the cleavage-proximal region. The Lille team’s findings, supported by the region Hauts-de-France through the REACT’EU programme and conducted with the imaging facilities of the BioImaging Center Lille Nord-de-France, reinforce a theme that has run through coronavirus structural biology since the pandemic began: the spike is not a static target but a dynamically glycosylated, protease-regulated machine, and its evolution must be read through that lens. As SARS-CoV-2 continues to diversify, the sugar chains flanking its activation switch may yet prove to be among the quiet arbiters of which variants rise and which fade.

Subject of Research: Role of N-glycans near the SARS-CoV-2 spike furin cleavage site in spike processing and viral entry

Article Title: Glycans located upstream of the furin cleavage site in SARS-CoV-2 Spike protein influence its processing and functions

Article References: Lavie, M., Louvet, E., Patron, J., Hanoulle, X., & Belouzard, S. (2026). Glycans located upstream of the furin cleavage site in SARS-CoV-2 Spike protein influence its processing and functions. Scientific Reports. https://doi.org/10.1038/s41598-026-70382-z

Image Credits: AI Generated

DOI: 10.1038/s41598-026-70382-z

Keywords: SARS-CoV-2, spike glycoprotein, furin cleavage, N-glycosylation, S1/S2 processing, membrane fusion, viral entry, D614G, Omicron XBB.1.5, glycan shield, R634, Scientific Reports