A plant-produced SARS-CoV-2 spike protein elicits heterologous immunity in hamsters

Abstract

Molecular farming of vaccines has been heralded as a cheap, safe and scalable production platform. In reality, however, differences in the plant biosynthetic machinery, compared to mammalian cells, can complicate the production of viral glycoproteins. Remodelling the secretory pathway presents an opportunity to support key post-translational modifications, and to tailor aspects of glycosylation and glycosylation-directed folding. In this study, we applied an integrated host and glyco-engineering approach, NXS/T Generation™, to produce a SARS-CoV-2 prefusion spike trimer in Nicotiana benthamiana as a model antigen from an emerging virus. The size exclusion-purified protein exhibited a characteristic prefusion structure when viewed by transmission electron microscopy, and this was indistinguishable from the equivalent mammalian cell-produced antigen. The plant-produced protein was decorated with under-processed oligomannose N-glycans and exhibited a site occupancy that was comparable to the equivalent protein produced in mammalian cell culture. Complex-type glycans were almost entirely absent from the plant-derived material, which contrasted against the predominantly mature, complex glycans that were observed on the mammalian cell culture-derived protein. The plant-derived antigen elicited neutralizing antibodies against both the matched Wuhan and heterologous Delta SARS-CoV-2 variants in immunized hamsters, although titres were lower than those induced by the comparator mammalian antigen. Animals vaccinated with the plant-derived antigen exhibited reduced viral loads following challenge, as well as significant protection from SARS-CoV-2 disease as evidenced by reduced lung pathology, lower viral loads and protection from weight loss. Nonetheless, animals immunized with the mammalian cell-culture-derived protein were better protected in this challenge model suggesting that more faithfully reproducing the native glycoprotein structure and associated glycosylation of the antigen may be desirable.

Keywords: SARS-CoV-2; challenge; glycoprotein; glycosylation; immunogenicity; vaccine.

Figure 1

Production of a soluble, stabilized spike in plants. (A) Gel filtration profile of affinity-captured spike. Peak 1, fractions 31-35 = aggregates; Peak 2, fractions 36-40 = trimers. (B) Negative stain electron microscopy of fractions 31-35 derived from peak 1 (aggregates) in A. Large circles indicate large, clumps of aggregated spike protein. (C) Negative stain electron microscopy of fractions 36-40 derived from peak 2 (trimers) in A. Small circles indicate spike trimers. (D) Coomassie-stained BN-PAGE of fractions 31-35, peak 1 (aggregates) in A. (E) Coomassie-stained BN-PAGE of fractions 36-40, peak 2 (trimers) from A. (F) Western blotting of fractions 31-35 (lane 1) and 36-40 (lane 2) from peaks 1 and 2 respectively in A using polyclonal mouse anti-His antibody [Serotech, MCA1396]. (G) Three-dimensional reconstruction and class averages of purified trimeric spike protein (fractions 36-40) from electron micrographs shown in (C).

Figure 2

Site-specific N-glycosylation of plant and HEK293F-produced spike. (A) Fully glycosylated model of the SARS-CoV-2 spike glycoprotein in a native-like prefusion conformation, generated previously by (Zuzic et al., 2022). The individual N-glycan sites were colored according to the site-specific abundance of oligomannose-type glycans determined by LC-MS for spike protein produced using glycan enhanced N.benthamiana cells. (B) Model of the SARS-CoV-2 S protein produced in an identical manner to panel A except colored according to LC-MS analysis of spike protein produced by transient transfection of HEK293F cells.

Figure 3

The change in site-specific N-glycan occupancy of plant-produced and mammalian cell-derived spike. (A) Model of the SARS-CoV-2 S glycoprotein in the native-like prefusion conformation generated by Zuzic et al., (2022). Individual N-glycan sites are colored according to the percentage point change in glycan occupancy of plant-derived protein relative to the comparator HEK293F-produced protein and therefore positive and negative values represent a relative increase or decrease in the glycan occupancy each sequon. (B) Bar graph depicting the percentage point changes depicted in panel A. A negative value indicates lower glycan occupancy in the plant-produced spike protein. Sites that could not be resolved are denoted with an asterisk.

Figure 4

Immunogenicity of spike trimers in hamsters when produced in plants and mammalian cells. (A) Immunization schematic depicting the timing of immunizations and sampling during the experiment. (B) Neutralizing antibody titres (ID₅₀) against the matched wildtype (Wuhan) viral isolate. (C) Neutralizing antibody titres (ID₅₀) against heterologous Delta virus. Pre-bleed= day -2, first bleed = day 14, final bleed = day 46.

Figure 5

Vaccine-mediated protection against heterologous challenge. (A) Change in weight following experimental challenge. (B) Grading of lung pathology following challenge. (C) Representative images of histopathology findings in lung sections. (D) Viral load following challenge.