The role of N-glycosylation in spike antigenicity for the SARS-CoV-2 gamma variant

Abstract

The emergence of SARS-CoV-2 variants alters the efficacy of existing immunity towards the viral spike protein, whether acquired from infection or vaccination. Mutations that impact N-glycosylation of spike may be particularly important in influencing antigenicity, but their consequences are difficult to predict. Here, we compare the glycosylation profiles and antigenicity of recombinant viral spike of ancestral Wu-1 and the Gamma strain, which has two additional N-glycosylation sites due to amino acid substitutions in the N-terminal domain (NTD). We found that a mutation at residue 20 from threonine to asparagine within the NTD caused the loss of NTD-specific antibody COVA2-17 binding. Glycan site-occupancy analyses revealed that the mutation resulted in N-glycosylation switching to the new sequon at N20 from the native N17 site. Site-specific glycosylation profiles demonstrated distinct glycoform differences between Wu-1, Gamma, and selected NTD variant spike proteins, but these did not affect antibody binding. Finally, we evaluated the specificity of spike proteins against convalescent COVID-19 sera and found reduced cross-reactivity against some mutants, but not Gamma spike compared to Wuhan spike. Our results illustrate the impact of viral divergence on spike glycosylation and SARS-CoV-2 antibody binding profiles.

Keywords: antibodies; coronavirus; glycoprotein; glycoproteomics; mass spectrometry.

Fig. 1

Mutations of the spike protein of the SARS-CoV-2 gamma variant and evolution of T20N and R190S mutations. A) Schematic of the key domains of spike and locations of the 12 mutations present in the gamma variant. SS, signal sequence; HR, heptad repeat; CH, central helix; CD, connector domain; TM, transmembrane domain; CT, cytoplasmic domain. The spike construct used in this study comprised residues 1,204 of SARS-CoV-2 S with “GSG” substituted at the furin cleavage site and a molecular clamp substituted for the TM and CT domains. B) Early evolution and fixation of SARS-CoV-2 spike glycoforms in gamma (P.1 and descendent lineages). The left y-axis shows the proportion of sequences with complete collection dates on GISAID harboring T20N (red), R190S (blue) or both T20N/R190S (green). The right y-axis shows the cumulative number of P.1 sequences deposited on GISAID between 20 Nov 2020 to 1 Jul 2021.

Fig. 2

Wu-1 and gamma SARS-CoV-2 spike proteins exhibit different sensitivities to RBD and NTD specific mAbs. A) Indirect ELISA show binding curves of Wu-1 (red) and gamma (blue) spike to RBD- and NTD-specific antibodies as indicated. Kd shown for each curve, nM.

Fig. 3

Binding characterization of SARS-CoV-2 spike proteins of Wu-1, gamma and NTD mutants. A). Indirect ELISA show binding curves of RBD- and NTD-specific antibodies as indicated. B) Dissociation constant of mAbs against Wu-1 or gamma, mutated SARS-CoV-2 spike proteins.

Fig. 4

Site-specific N-linked glycosylation occupancy and structural heterogeneity of SARS-CoV-2 spike proteins of Wu-1, gamma and NTD mutants. Site-specific N-glycosylation occupancy at sites (A) N17, (B) N20 and (C) N188. D) Principal component analysis (PCA) of glycoform abundances excluding N188. Site-specific glycoform analysis at sites (E) N17 and N20 with the (F) log₂ fold change relative to Wu-1 (blue increased abundance, red decreased abundance, P < 0.05). For N17 in (E) the unmodified intensity for T20N, L18F/T20N, L18F/T20N/R190S and gamma were inferred from the occupancy analysis. (G) Site-specific glycoform analysis at site N188. HexNAc, N-acetylhexosamine; hex, hexose; Fuc, fucose; NeuAc, N-acetylneuraminic acid.

Fig. 5

SARS-CoV-2 spike proteins of Wu-1, gamma and NTD mutants have diverse glycosylation profiles. Clustered heatmap of the relative abundance of all identified glycoforms in all replicates, excluding N188. Both rows and columns were clustered using correlation distance and average linkage. The percentage of types of glycans observed in each selected cluster are represented in bar graphs (refer to the methods for the allocation of glycan-types). The full clustered heatmap with labelled rows (glycopeptide glycoforms) is shown in Supplementary Fig. 3A.

Fig. 6

The binding affinities of SARS-CoV-2 spike proteins of Wu-1, gamma and NTD mutants are not dependent on sialic acid content. Variant spike proteins were treated with neuraminidase for 1 h or left untreated. The proteins were then coated on ELISA plates and binding affinities determined for COVA2-17 or CB6 mAb.

Fig. 7

NTD-specific IgG titer of 30 COVID-19 patients (COVID-19 pos) and close contact (COVID-19 neg). Blood samples were obtained from 30 COVID-19 patients and close contact at 3 months after infection IgG titers against the recombinant NTD domains (^**P-value < 0.01, ^***P-value < 0.001).