Variations within the Glycan Shield of SARS-CoV-2 Impact Viral Spike Dynamics

Abstract

The emergence of SARS-CoV-2 variants alters the efficacy of existing immunity, whether arisen naturally or through vaccination. Understanding the structure of the viral spike assists in determining the impact of mutations on the antigenic surface. One class of mutation impacts glycosylation attachment sites, which have the capacity to influence the antigenic structure beyond the immediate site of attachment. Here, we compare the site-specific glycosylation of recombinant viral spike mimetics of B.1.351 (Beta), P.1 (Gamma), B.1.617.2 (Delta), B.1.1.529 (Omicron). The P.1 strain exhibits two additional N-linked glycan sites compared to the other variants analyzed and we investigate the impact of these glycans by molecular dynamics. The acquired N188 site is shown to exhibit very limited glycan maturation, consistent with limited enzyme accessibility. Structural modeling and molecular dynamics reveal that N188 is located within a cavity by the receptor binding domain, which influences the dynamics of these attachment domains. These observations suggest a mechanism whereby mutations affecting viral glycosylation sites have a structural impact across the protein surface.

Keywords: SARS-CoV-2; glycosylation; mass spectrometry; molecular dynamics; variant of concern.

Graphical abstract

Figure 1

Three-dimensional representation of the SARS-CoV-2 S protein, showing regions of the spike that differ from the Wuhan strain in VOCs Beta, Gamma, Delta and Omicron. Each protomer is colored in different shades of blue and N-linked glycans colored in gray. Amino acid mutations are color coordinated into lime green, Beta; yellow, Gamma; pink, Delta; purple, Omicron; red, ≥2 VOCs. Model of the Wuhan-hu-1 S glycan shield by Zuzic et al. 2022.

Figure 2

Schematic of soluble recombinant SARS-CoV-2 variant S glycoproteins. Domains are independently colored corresponding to each variant including: NTD, N-terminal domain; RBD, receptor binding domain; FP, fusion peptide; HR1, heptad repeat 1; CH, central helix region; CD, connector domain; HR2, heptad repeat 2; GSAS, mutated furin cleavage site. Glycans are depicted as black forks and stabilizing proline mutations are colored orange. Mutations (compared with engineered Wuhan-hu-1 strain) are color matched to the domains of each variant, with conserved mutations between variants labelled in red. Deleterious mutations are highlighted using dashed lines.

Figure 3

Site-specific glycan analysis of recombinant Wuhan-hu-1, B.1.351, P.1, and B.1.1.529 S protein using LC-MS. (A) Glycan compositions are grouped into their corresponding categories, with complex-type glycans displayed in pink, hybrid as hatched pink and white, oligomannose in green and unoccupied in grey. Glycan sites not present are denoted as “N.P.” and glycan sites that could not be resolved are denoted as “N.D.”.

Figure 4

Map of the site-specific glycosylation of the SARS-CoV-2 S protein. Using a previously generated model of the Wuhan-hu-1 glycan shield by Zuzic et al. 2022, the site-specific glycosylation determined in Figure 3 was used to color the map. If the oligomannose-type glycan content was 80% or more then the site is colored green. Between 79% and 20% correspond to mixed sites and are colored orange, and sites containing less than 20% oligomannose-type glycans correspond to complex-type glycan sites and are colored pink. Each variant is based on the Wuhan-hu-1 model; however, the additional glycan sites present in the Gamma variant have been added and are labelled appropriately. The dashed lines depict the position at which the recombinant S proteins are truncated to remove the transmembrane domain.

Figure 5

(A) Representative snapshot (793 ns) from 1.050 μs MD production trajectory of the reconstructed SARS-CoV-2 P.1 S ectodomain from PDB 7sbs. The three protomers are color-coded according to the legend. All glycan structures are shown in blue with sticks, except for N331-FA2G2 (Chain B) and N188-Man5 (Chain C) highlighted in purple and green, respectively. Bar chart depicts the degree of oligomannose-type glycan trimming (Man₉GlcNAc₂-Man₅GlcNAc₂ denoted by Man9-Man5) at sites N188 and N234 in P.1 with oligomannose glycoforms displayed using increasing luminance of green. (B) Close-up view of the NTD of Chain C with the N188-Man5 embedded in the cavity and interacting with N331-FA2G2 from the adjacent, closed RBD (Chain B). (C) Close-up view of the NTD of Chain C rendered with cartoons (cyan), with the adjacent RBD (Chain B) as a white solvent-accessible surface. The residues that the N188-Man5 makes stable contact through the trajectory (as seen for all protomers) are highlighted with sticks (yellow) and labelled based on the PDB 7sbs numbering. The loop from residue 171 to 183 framing one side of the cavity is also highlighted in yellow. (D) Structure of the NTD (Chain C) represented by its electrostatic potential energy surface calculated with the Adaptive Poisson-Boltzmann Solver (APBS) plugin in Pymol (www.pymol.org). The location of the cavity occupied by the N188-Man5 is indicated with an arrow. (E) Kernel-density estimate (KDE) plots of the molecular volumes (in ų) of the NTD cavity in all different protomers calculated from 300 ns to 1.05 μs of the MD production trajectories for the N188-glycosylated model gamma S (cyan) and for the N188-non-glycosylated model (grey). The data from 0 to 300 ns suggested a significant readjustment of the structure and was considered conformational equilibration. Median values for all distributions are indicated below the plots with corresponding standard deviations in brackets. Molecular volumes calculated with PyVol, plots with seaborn (https://seaborn.pydata.org/), all molecular rendering, except for the electrostatic potential surface in panel d) with VMD (https://www.ks.uiuc.edu/Research/vmd/).