A computationally designed antigen eliciting broad humoral responses against SARS-CoV-2 and related sarbecoviruses

ACE2-binding sarbecoviruses within the Betacoronavirus genus pose a high pandemic risk. SARS-CoV (2002–2003) and SARS-CoV-2 (current pandemic) exemplify zoonotic spillovers from bat reservoirs that harbor diverse ACE2-binding sarbecoviruses. Ongoing circulation has led to continual mutation, especially in the spike RBD, generating variants of concern (VOCs) with enhanced transmission and immune escape (for example, N501Y, L452R, T478K; multiple mutations in Omicron lineages). Because many RBD mutations cluster in receptor-binding motifs targeted by potent neutralizing antibodies, there is an urgent need for next-generation vaccines that provide broad protection across ACE2-binding sarbecoviruses and emerging VOCs. This study aimed to design and assess a computationally derived, immune-optimized single RBD-based antigen capable of eliciting cross-sarbecovirus humoral responses and serving as a booster on backgrounds of existing Wuhan-based spike immunity.

The field has explored several strategies for broad coronavirus vaccines, including chimeric spike mRNAs and mosaic/cocktail nanoparticle vaccines displaying RBDs from multiple coronaviruses, which have shown promise for pan-sarbecovirus and pan-betacoronavirus immunity but complicate manufacturing due to multiple constructs. Antibody responses to the RBD have been structurally classified into classes 1–4; classes 1–3 are potent neutralizers. Glycan masking of immunodominant, variable epitopes has been used by viruses (hepatitis C, Lassa, influenza) to evade immunity and can be leveraged to refocus responses to conserved epitopes. The rapid antigenic evolution of SARS-CoV-2, particularly in Omicron lineages, underscores the importance of variant-resilient designs and strategies to overcome original antigenic sin from Wuhan-based spike vaccines.

In silico antigen design: Sarbecovirus spike protein sequences (NCBI Virus, June 2020) were aligned (MUSCLE), pruned to RBD, filtered at 95% identity, and used to reconstruct a phylogeny (IQ-TREE). A phylogenetically optimized core RBD (T2_13) was generated (HyPhy) to balance conserved and distinct epitopes across sarbecoviruses. Epitope-informed variants were created by matching conserved epitopes for cross-reactive mAbs S309 (class 3) and CR3022 (class 4) to SARS-CoV (T2_14, T2_15) and by inserting a glycosylation sequon (N-X-T/S) to mask the divergent class-1 B38 epitope (T2_17 from T2_14; T2_18 from T2_15), selected via FoldX energy scoring. Soluble and transmembrane-anchored forms (T2_13_TM, T2_17_TM) were modeled (MODELLER, SCWRL, GROMACS) and stability-checked (FoldX). Genes were codon-optimized for human expression (GeneOptimizer) and cloned into pEVAC. In vivo screening in BALB/c mice: Female BALB/c mice (n=6/group) received DNA immunizations (50 µg, s.c.) of designs or SARS-CoV-2 RBD at 30-day intervals, with bleeds every 15 days. Cross-reactive binding to cell-surface spike proteins (SARS-CoV, SARS-CoV-2, WIV16, RaTG13) was assessed by FACS. ELISAs quantified RBD-specific binding. T2_17 was selected as lead based on broad binding. Outbred species and neutralization: Guinea pigs (n=8/group) received intradermal DNA via PharmaJet Tropis at 28-day intervals (3 doses) with control SARS2_RBD_P521N. Binding ELISA and pseudovirus neutralization (lentiviral pseudotypes bearing spikes of SARS-CoV, SARS-CoV-2, WIV16, RaTG13 and VOCs) generated IC50 values. ACE2-competition ELISA evaluated receptor-blocking activity. Rabbits (GMP pEVAC_T2_17 DNA, intradermal Tropis) were immunized at 14-day intervals (4 doses) with serial bleeds, assayed for binding and pseudovirus neutralization against SARS-CoV-2 VOCs. K18-hACE2 mouse prime-boost and challenge: Homozygous K18-hACE2 mice were primed i.m. with AZD1222 (1.4×10^6 vp) and boosted 4 weeks later with AZD1222, T2_17 DNA, or MVA-T2_17; controls received PBS. Eight weeks post-boost, mice were challenged intranasally with SARS-CoV-2 Victoria (B lineage) or Delta; weight and survival were monitored. A longitudinal cohort was primed with AZD1222 and boosted at 20 weeks with AZD1222, T2_17 DNA, or MVA-T2_17 (or PBS); neutralization over time and peptide microarray mapping (15-mer, 14-aa overlap) of RBD-specific antibodies were performed. mRNA immunogens: Chemically modified mRNAs encoding T2_17 (soluble) and T2_17_TM (membrane-anchored) were formulated in LNPs and tested in BALB/c mice (5 or 10 µg; 2 doses, 21–28 days apart). Controls included full-length prefusion-stabilized spike mRNA and BNT162b2. Guinea pigs received T2_17_TM or full-length spike mRNA (3.15 µg or 15 µg; two doses, 3-week interval). Binding ELISAs and pseudovirus neutralization against sarbecoviruses and VOCs (including Omicron BA.1 and XBB.1.5) were performed. Assays and analyses: FACS-based binding, ELISA AUCs, pseudotype microneutralization (IC50), ACE2 competition ELISA, and nonparametric statistics (two-tailed Mann–Whitney U) were used. Sample sizes: typically n=6–12 per group as indicated by figure legends.

- In BALB/c mice, multiple DIOSynVax-designed RBD antigens elicited cross-reactive binding to SARS-CoV, SARS-CoV-2, WIV16, and RaTG13 spikes; T2_17 showed the best or second-best median binding across all four, leading to its selection. ELISAs confirmed robust binding to SARS-CoV and SARS-CoV-2 after two DNA immunizations; T2_17 induced stronger binding to SARS-CoV than SARS-CoV-2, comparable to SARS-CoV-2 RBD against SARS-CoV-2. - Guinea pigs (DNA): T2_17 induced higher binding to SARS-CoV than the control SARS2_RBD_P521N after one and two immunizations, with comparable SARS-CoV-2 binding; after three doses, SARS2_RBD_P521N was higher for SARS-CoV-2, while T2_17 remained higher for SARS-CoV. Neutralization: SARS-CoV-2 neutralizing antibodies were detected after the first dose; potent SARS-CoV neutralization developed after two doses, stronger for T2_17 than control. At bleed 6 (28 days post third dose), T2_17 yielded significantly higher neutralizing titres against SARS-CoV, WIV16 and RaTG13 than SARS2_RBD_P521N. ACE2 competition showed both antisera abrogated RBD–ACE2 binding comparably to WHO convalescent standard. - Rabbits (DNA): By 2 weeks after the third/fourth immunization, sera exhibited robust neutralization across SARS-CoV, SARS-CoV-2, and VOCs Beta, Gamma, Delta and Omicron BA.1, as well as WIV16 and RaTG13. - K18-hACE2 mice (AZD1222 prime then boost): Boosting with T2_17 (DNA or MVA) increased binding titres to SARS-CoV and SARS-CoV-2. Two weeks post-boost, T2_17 DNA and MVA groups neutralized Delta significantly better than AZD1222-boosted sera. SARS-CoV neutralization was detected only in the T2_17(MVA) group pre-challenge. All vaccinated groups (not PBS) were protected from weight loss and survived challenge with Victoria or Delta. - Longitudinal K18-hACE2: Boost at 20 weeks showed significantly higher SARS-CoV-2 neutralization at 4 weeks post-boost in the T2_17(MVA) group vs AZD1222. Peptide microarrays revealed more RBD peptide hits in AZD1222/T2_17(MVA) compared with AZD1222/AZD1222, indicating broader RBD-focused humoral responses after T2_17 boosting. - mRNA studies: In BALB/c mice, T2_17_TM elicited higher binding titres than soluble T2_17 at 5 µg; both were comparable at 10 µg. In guinea pigs, T2_17_TM (3.15 µg) induced broader neutralization than full-length spike (15 µg), with significantly higher titres against WIV16, SARS-CoV, and Omicron BA.1 at 6 weeks post-boost; full-length spike showed negligible BA.1 neutralization. Neither vaccine neutralized XBB.1.5 at low dose, but at 15 µg, T2_17_TM sera neutralized XBB.1.5 whereas full-length spike did not. Overall, T2_17/T2_17_TM elicited broad cross-sarbecovirus neutralization across platforms (DNA, MVA, mRNA).

The study demonstrates that a single computationally designed, immune-refocused RBD antigen (T2_17) can induce broad humoral responses across ACE2-binding sarbecoviruses and multiple SARS-CoV-2 VOCs. By masking a divergent, immunodominant class-1 epitope (B38 site) with an engineered glycan while preserving conserved epitopes (for example, S309, CR3022), the design appears to bias responses toward conserved, variant-resilient sites, yielding cross-neutralization of SARS-CoV, SARS-CoV-2, WIV16 and RaTG13. As a heterologous booster on an AZD1222 prime background, T2_17—particularly as an MVA vector—enhanced neutralization against the Delta variant and uniquely elicited SARS-CoV neutralization in mice, without compromising protection upon live virus challenge. Across modalities (DNA, MVA, mRNA), T2_17/T2_17_TM consistently broadened serologic coverage, outperforming homologous Wuhan full-length spike in neutralization breadth against divergent sarbecoviruses and Omicron BA.1, and achieving XBB.1.5 neutralization at higher mRNA dose. These results address the need for next-generation vaccines with breadth against zoonotic sarbecoviruses and evolving variants, offering a manufacturable single-antigen alternative to multi-component chimeric or mosaic approaches. T2_17 may also help mitigate immune imprinting from Wuhan-based spike vaccines by focusing responses on RBD and conserved epitopes.

T2_17, a DIOSynVax-designed RBD antigen with engineered epitope presentation, elicits broad, cross-sarbecovirus humoral immunity in mice, guinea pigs and rabbits and enhances protection and neutralization breadth as a booster in AZD1222-primed K18-hACE2 mice. The antigen is effective across vaccine platforms (DNA, MVA, mRNA), neutralizing SARS-CoV, SARS-CoV-2 (including VOCs Alpha, Beta, Gamma, Delta, Omicron BA.1) and, at higher mRNA dose, XBB.1.5. Designed prior to variant emergence, T2_17’s performance validates the DIOSynVax computational and structural engineering approach. Future work should update designs to incorporate VOC sequence diversity, consider combining conserved structural/non-structural antigens, and clinically evaluate T2_17 as a booster to overcome immune imprinting; phase 1 trials have been initiated.

- Neutralization titres against recent VOCs (for example, XBB.1.5) were lower than against the Wuhan strain and required higher mRNA dose for detectable neutralization, indicating potential need for updated designs. - Differences in vector platforms influenced breadth: SARS-CoV neutralization was observed in MVA-boosted but not DNA-boosted mice after AZD1222 prime; vector effects were not fully dissected. - Protection and immunogenicity were evaluated in animal models; human immune landscapes are more complex due to prior infections and vaccinations, necessitating clinical validation. - Emphasis was on humoral responses; cellular immunity was not detailed and could impact breadth and durability. - The study design predated many later VOCs; inclusion of newer sequences may further optimize breadth.