Broadly neutralizing antibody induction by non-stabilized SARS-CoV-2 Spike mRNA vaccination in nonhuman primates

Immunization with mRNA or viral vectors encoding spike with diproline substitutions (S-2P) has provided protective immunity against severe COVID-19 disease. How immunization with Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) spike elicits neutralizing antibodies (nAbs) against difficult-to-neutralize variants of concern (VOCs) remains an area of great interest. Here, we compare immunization of macaques with mRNA vaccines expressing ancestral spike either including or lacking diproline substitutions, and show the diproline substitutions were not required for protection against SARS-CoV-2 challenge or induction of broadly neutralizing B cell lineages. One group of nAbs elicited by the ancestral spike lacking diproline substitutions targeted the outer face of the receptor binding domain (RBD), neutralized all tested SARS-CoV-2 VOCs including Omicron XBB.1.5, but lacked cross-Sarbecovirus neutralization. Structural analysis showed that the macaque broad SARS-CoV-2 VOC nAbs bound to the same epitope as a human broad SARS-CoV-2 VOC nAb, DH1193. Vaccine-induced antibodies that targeted the RBD inner face neutralized multiple Sarbecoviruses, protected mice from bat CoV RsSHC014 challenge, but lacked Omicron variant neutralization. Thus, ancestral SARS-CoV-2 spike lacking proline substitutions encoded by nucleoside-modified mRNA can induce B cell lineages binding to distinct RBD sites that either broadly neutralize animal and human Sarbecoviruses or recent Omicron VOCs.


Figure S2 .
Figure S1.Vaccine-induced SARS-CoV-2 specific IgG binding titers.Total serum IgG binding magnitude to Spike 2P (S-2P), anti-Spike S2 ECD (ectodomain) and anti-NTD IgG were tested by ELISA and shown as log area under the curve (logAUC).Symbols indicate the group mean value ± SEM of three replicates.

Figure
Figure S3.mRNA-LNP vaccination prevents virus replication in the lower respiratory tract after intranasal and intratracheal SARS-CoV-2 challenge in macaques.(A) SARS-CoV-2 envelope gene (E gene) sgRNA and nucleocapsid gene (N gene) sgRNA in bronchoalveolar lavage (BAL) and nasal swab samples were quantitated pre-challenge, and on Day 2, 4, and 7. (B) A representative image of nucleocapsid antigen staining from each group of mRNA-LNP vaccinated macaques is shown.All images are shown at 10x magnification.Scale bars, 100 μm.

Figure S4 .
Figure S4.Flow cytometric gating strategies for antigen-specific B cell isolation.The gating strategy to identify SARS-CoV-2 full length spike and RBD-specific memory B cells.

Figure S7 .
Figure S7.Monoclonal antibodies isolated from wildtype S-tm mRNA-LNP immunized rhesus macaques bind SARS-CoV-2 variants and SARS-related betacoronaviruses and possess different epitope specificities.Antibodies labeled Spike-reactive bound to various domains of Spike in initial binding screens and confirmatory ELISA.

Figure S9 .
Figure S9.Cryo-EM data processing for Antibody DH1338 in complex with SARS-CoV-2 S protein, Related to Figure Cryo-EM-1.(A) Representative micrograph.(B) Representative CTF fit.(C) Representative 2D class averages from cryo-EM dataset.Box size = 324 Å. (D) Refined 3D map segmented and colored by component, with the SARS-CoV-2 S-HexaPro protein colored in grey and DH1338 colored green.(E) Refined map shown in D colored by local resolution.(F) Fourier Shell Correlation (FSC) curves of the 3D reconstruction shown in D with horizontal blue line indicating FSC 0.143.(G) Local refinement of RBD-DH1338 Fv region.Left Blue mesh shows the mask that was used for local refinement.Right.Local Refined 3D density map colored by local resolution.(H) FSC curves of local refined map shown in G. (I) DH1338 HC binding interface with RBD, shown in sticks with electron density shown in blue mesh.(J) DH1338 LC binding interface with RBD.PDBID: 8DPZ.(K) ACE2 (yellow surface representation; PDB 6M0J) binding to RBD (Red cartoon with RBM shown in purple; PDB 6M0J) is sterically hindered by DH1338 (Blue, HC; light blue, LC; cartoon representation; PDB 8DPZ).

Figure S10 .
Figure S10.Cryo-EM data processing for Antibody DH1047 in complex with SARS-CoV-2 S protein.(A) Refined map of the SARS-CoV-2 2P spike bound to three DH1047 Fabs, highlighted in violet.(B) The from A. colored by local resolution.(C) Fourier Shell Correlation for the refinement of the full spike-Fab complex.(D) Local refinement region covered by the mask shown in green mesh and resulting map from local refinement, colored by local resolution.The resulting map included the spike RBD and Fv region of DH1047 (E) Fourier Shell Correlation of the locally refined map.(F-I) Representative views of the fit of the model to the local refinement map.(F) CDR-L1 to RBD contact.(G) CDR-L3 and H2 to RBD contact.(H) CDR-H2 to RBD contact.(I) CDR-H2 and H3 to RBD contact.

High-throughput ELISA binding screen of antibodies recovered from wildtype SARS- CoV-2 spike mRNA-LNP vaccinated rhesus macaques.
Antibodies were tested in unpurified cell culture supernatant at a single dilution.Yellow highlighted values are above 0.1 OD450nm and are considered positive.

Table XA .
Detailed interactions of DH1338 LC with RBD from Pisa Server.