The diversity of the glycan shield of sarbecoviruses related to SARS-CoV-2

Abstract

Animal reservoirs of sarbecoviruses represent a significant risk of emergent pandemics, as evidenced by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic. Vaccines remain successful at limiting severe disease and death, but the potential for further coronavirus zoonosis motivates the search for pan-coronavirus vaccines. This necessitates a better understanding of the glycan shields of coronaviruses, which can occlude potential antibody epitopes on spike glycoproteins. Here, we compare the structure of 12 sarbecovirus glycan shields. Of the 22 N-linked glycan attachment sites present on SARS-CoV-2, 15 are shared by all 12 sarbecoviruses. However, there are significant differences in the processing state at glycan sites in the N-terminal domain, such as N165. Conversely, glycosylation sites in the S2 domain are highly conserved and contain a low abundance of oligomannose-type glycans, suggesting a low glycan shield density. The S2 domain may therefore provide a more attractive target for immunogen design efforts aiming to generate a pan-coronavirus antibody response.

Keywords: CP: Immunology; CP: Microbiology; N-linked glycosylation; SARS-CoV-2; glycan shielding; pan-coronavirus.

Graphical abstract

Figure 1

Conservation of N-linked glycosylation sequons across a sample of sarbecovirus S proteins (A) Alignment of 78 sarbecovirus S protein sequences. The y axis represents the proportion of sarbecoviruses that possess an N-linked glycan attachment site, expressed as a percentage of the total sequences used. Peaks corresponding to glycan sites from SARS-CoV-2 are labeled in black with their position on SARS-CoV-2. N370 is colored red because it is highly conserved but not present in SARS-CoV-2. (B) Clustal Omega multiple sequence alignment of the sarbecoviruses analyzed in this study alongside other human CoVs. Each sarbecovirus is colored according to the clade, which has been classified previously.^, (C) Table of the sarbecoviruses analyzed in this study, displaying the name, the species from which it was isolated, and the region in which the isolate was discovered. (D) Reproduction of a model of the SARS-CoV-2 glycan shield from Allen et al. determined from aggregation of data from recombinant proteins from multiple sources. The protein is displayed in gray, and the glycans are colored according to the abundance of oligomannose-type glycans present at each site. (E) Bar chart depicting the number of sarbecoviruses containing an NxS/T motif within the subpanel selected for glycopeptide analysis. Each sarbecovirus was aligned to SARS-CoV-2, and the glycan sites are displayed relative to their position on SARS-CoV-2. See also Tables S1–S3.

Figure 2

Determination of site-specific glycosylation of sarbecoviruses by LC-MS (A) Sum of the oligomannose-type glycans located at each N-linked glycan site on the sarbecoviruses analyzed in this study. The sequences for all sarbecoviruses were aligned with the SARS-CoV-2 S protein, and the glycan sites are presented aligned to this protein. The oligomannose-type glycan content of previously published site-specific data for SARS-CoV-2 S protein is shown as red dots. The mean of all strains is displayed as a line, and the error bars represent ±SEM or, when only two datasets are present, the range of the two datasets. (B) The averaged glycan processing state of all sarbecoviruses aligned with the SARS-CoV-2 S protein. Glycans classified as oligomannose-type are colored green, and hybrid-type glycans are represented as a white bar with pink hatches. Complex-type glycans are colored pink and the proportion of unoccupied N-linked glycan sites is displayed in gray. See also Figure S1 and Tables S4–S7.

Figure 3

Clade-specific glycan processing of sarbecoviruses (A) Site-specific glycosylation of clade 1a sarbecoviruses, with the data displayed in a manner identical to Figure 2, with the symbols representing the oligomannose-type glycan content of individual strains and the bar graph representing the consensus glycosylation pattern at each site. (B) Site-specific glycosylation of clade 1b sarbecoviruses. (C) Site-specific glycosylation of clade 2 sarbecoviruses. (D) Site-specific glycosylation of clade 3 sarbecoviruses. Sites that are not present in a particular clade are labeled with an asterisk. Sites where the site-specific glycosylation could not be determined are labeled n.d. Error bars represent ± SEM. See also Tables S2–S7.

Figure 4

Modeling the glycan shield of sarbecoviruses with their site-specific glycosylation (A and B) 3D maps of the sarbecoviruses glycan shields are displayed top down (A) and side on (B). All models were constructed using SWISS-MODEL, GlycoShield, and the MS data displayed in Figure 3. Each model displays the protein sequence in gray. A representative Man₅GlcNAc₂ glycan was mapped onto each PNGS and is colored according to the oligomannose-type glycan content at each site, with 80% and above colored green, between 79% and 20% colored orange, and below 20% colored pink. The C-terminal region of the S protein was not resolved in the templates used to generate the models and therefore is not included.

Figure 5

Glycan shield map of RaTG13-CoV to investigate the distinct glycan processing observed in clade 1b sarbecoviruses (A) Reproduction of the RaTG13 model generated in Figure 4, with the glycans recolored according to the p.p. difference in oligomannose-type glycans between RaTG13 and pang17, with a positive number representing a higher abundance of oligomannose-type glycans on pang17 relative to RaTG13. The protein sequence is displayed as a cartoon, with discrepancies in the amino acid sequence between RaTG13 and pang17 represented as blue spheres. Sites displaying increased oligomannose-type glycans on pang17 are labeled, with (A) representing a side-on view and (B) a top-down view. (C) Comparing the site-specific oligomannose-type glycan content of clade 1b sarbecoviruses: SARS-CoV-2, pang17, and RaTG13. (D) P.p. change in oligomannose-type glycans between RaTG13 and pang17. A positive p.p. change represents a glycoform that was present in higher abundance on pang17 compared with RaTG13.