We queried amino acid sequences of spike protein from National Center for Biotechnology Information (NCBI). Alphacoronavirus (HCoV-229E, HCoV-NL63, BatAlpha-CoV and CCoV-HuPn2018), Betacoronavirus (HCoV-OC43, HCoV-HKU1, SARS-CoV-1, SARS-CoV-2, MERS-CoV and Bat-CoV-RaTG13), Gammacoronavirus (IBV, Canada goose-CoV and Duck-CoV), Deltacoronavirus (PD-CoV, PCoV-HKU15 and White eye-CoV HKU16).

We focused on the spike (S) protein of various coronaviruses, including SARS-CoV-1, MERS-CoV, and SARS-CoV-2, as the antigenic target for vaccine development. The S proteins of coronaviruses were inserted into plasmid vectors to formulate individual DNA vaccines for each type of coronavirus. Analogously, the S protein of coronavirus was combined into the Vaccina virus TianTan (VTT) constructing VTT vaccines. Based on the differences in vector types and immunogen categories adopted by the vaccines, they are designated as DNA-SARS1, DNA-SARS2, DNA-MERS, VTT-SARS1, VTT-SARS2 and VTT-MERS respectively. The constructions of DNA and VTT vector vaccines were described in patents (CN101020055A and CN114032217A) and article (20). DNA vector vaccines containing spike proteins of SARS-CoV-1, SARS-CoV-2, and MERS-CoV coronaviruses were mixed at a 1:1:1 ratio to prepare a trivalent DNA vector vaccine; similarly, VTT vector vaccines containing spike proteins of SARS-CoV-1, SARS-CoV-2, and MERS-CoV coronaviruses were mixed at a 1:1:1 ratio to prepare a trivalent VTT vector vaccine.

The immunogenicity assessment of the vaccines was conducted on 6 to 8 weeks old female BALB/c mice, which were maintained under specific pathogen-free (SPF) conditions. Furthermore, the efficacy of these vaccines in conferring immune protection was validated through the utilization of 6-week-old Lvov’s Golden Syrian hamsters (LVG golden hamsters). BALB/c mice and LVG golden hamsters were purchased from Beijing Vital River Laboratory (Beijing, China).

To verify the immunogenicity and immune protection effect of the vaccines, we have designed two immunization protocols.

Firstly, BALB/c mice in the vaccine groups (n=5) were intramuscularly (i.m.) immunized with coronavirus DNA vaccines via the electroporation method (electroporation parameters: 100 V, 60 ms, interval 999 ms, 10 times; WJ-2005-Intelligent live cell gene delivery instrument, NINGBO SCIENTZ BIOTECHNOLOGY CO.LTD) at days 0 and 14, with a dose of 50 μg per mouse per immunization, administered to the tibialis anterior muscles; Subsequently, the mice were intramuscularly (i.m.) boosted with recombinant VTT vaccines at day 70, with a dose of 5×10⁶ plaque-forming units (pfu) per mouse. However, the mice in control group (n=3) were immunized empty DNA vector and VTT. On the 14th day after the end of immunization, the blood and spleens of the BALB/c mice were collected to detect humoral and cellular immune responses respectively.

The second experimental protocol was designed to validate the protective effect of the trivalent vaccine in LVG golden hamsters, with twelve animals per group (6 females and 6 males). The immune doses of DNA and recombinant VTT vaccines are 50 μg and 1×10⁷ pfu per mouse each time respectively. 14 days after immunization, the golden hamsters were challenged with the XXB strain (SARS-CoV-2 XBB RY166, 6.125lgCCID50/ml) via the intranasal. Following viral challenge, the body weight of hamsters in the vaccine groups and the PBS control group (n=6 per group; 3 females and 3 males) was recorded daily, and nasopharyngeal swabs were collected from these hamsters for viral load detection. On the other hand, after the XBB challenge, on days 3, 5, and 7, two LVG golden hamsters per group (1 female and 1 male) were sacrificed, and their bronchial tissues and lung tissues were collected for viral load detection.

The specific binding antibodies were detected using enzyme-linked immunosorbent assay (ELISA). The high affinity plates (Corning) were coated with coronavirus receptor binding domain (RBD) protein 100 μL per well (1.5μg/mL) overnight at 4°C and then blocked with 5% skim milk-2% bovine serum albumin (23). The immunized sera were diluted in series and then added into the coated plates with 100 μL per well. HRP-conjugated goat anti-mouse ((Southern Biotech) was diluted at 1:2500 to be used in ELISA experiments. Color development was achieved using TMB as the substrate (Beijing Kinghawk Pharmaceutical), and the reaction was terminated after 10 minutes. Optical density was detected using dual wavelengths of 450nm and 460nm. Endpoint titers with an absorbance over two times that of the control group were considered positive.

The neutralizing antibody levels were determined by pseudovirus neutralization assay. The construction of vesicular stomatitis virus (VSV)-based pseudoviruses for coronaviruses was based on previous literature (24). The inactivated immune sera were serially diluted and then incubated with pseudoviruses at 37°C for 1 hour. Subsequently, 100 μL of Huh7 cells (2×10⁵ cells/well) were added to each well and cultured under 37°C, 5% CO₂ culture conditions for 24 hours. The neutralizing titers at 50% (NT50) were detected using Ultra-High Sensitivity luciferase reagents (PerkinElmer, MA) by a chemiluminescence detector (VICTOR3, PerkinElmer). The values of NT 50 were calculated by the Reed-Muench method.

The specific T cell responses were accessed using the mouse IFN-γ ELISpot assay set (BD Biosciences). The ELISpot plates were coated with purified anti-mouse IFN-γ (5μg/mL) overnight at 4°C and then blocked for 2 hours at room temperature (RT). Fresh mouse splenocytes were seeded in the blocked ELISpot plates, and the splenocytes from each group of mice were stimulated with three coronavirus peptide pools, respectively (20, 22). The positive control was stimulated with cell stimulation cocktail (Invitrogen), and the negative control group was stimulated with dimethyl sulfoxide (DMSO) (Sigma). The splenocytes were cultured with the peptide pools for 18 hours, washed, and then incubated with biotinylated anti-mouse IFN-γ (2μg/mL) for 2 hours at RT according to the instruction manual. Subsequently, the streptavidin-HRP (1:100) were added into the plates and incubated an hour. Finally, the 3-Amino-9-ethylcarbazole (AEC) substrate set (BD Biosciences) was used for chromogen in ELISpot assay. The results were counted using the ELISpot reader instrument (AID GmbH).

To assess protective efficacy of the vaccines on golden hamsters, we measured the levels of viral load in pharyngeal swabs, lung tissues (part of the right lung), and tracheas of golden hamsters after virus challenge. The tissue samples underwent homogenization, followed by centrifugation to separate the supernatant. Subsequently, the total RNA was extracted from the tissue supernatant. The total RNA was extracted from tissues using the Fastpure Cell/Tissue Total RNA Isolation Kit V2 (Vazyme). The TaqPath™ 1-Step RT-qPCR Master Mix, CG (Thermo Fisher Scientific Inc.) kit was employed to detect the levels of viral load. The system and procedure for PCR reactions followed the instruction manual. Viral load was detected using the LightCycler 480 II Real-Time Fluorescent Quantitative PCR instrument (Roche). Primers and probes were synthesized by Beijing Tsingke Biotechnology Co., Ltd. (prime-F: TTACAAACATTGGCCGCAAA, prime-R: GCGCGACATTCCGAAGAA and probe: FAM-ACAATTTGCCCCCAGCGCTTCAG-BHQ1).

The antibody levels, neutralizing antibody titers, and viral load levels were described using geometric means, whereas the results of cellular immune responses were presented using mean values. The statistical analyses were performed using the Kruskal-Wallis test followed by the Dunn’s multiple comparisons test. The GraphPad Prism 9.5.1 software was utilized for plotting, and statistical significance was indicated by p values (*p < 0.05, **p < 0.01, or ***p < 0.001).

To assess the immunogenicity of the vaccine, BALB/c mice were primed with DNA vaccines and then boosted with VTT vaccines (DNA-prime/VTT-boost strategy). Subsequently, the humoral and cellular immune responses in the mice were evaluated 14 days post the third vaccination. The DNA vaccination was administered with electroporation (electroporation parameters: 100V, 60ms, with an interval of 999ms, repeated 10 times) (Figure 1A).

Humoral immune responses were assessed by measuring levels of binding antibodies and neutralizing antibodies. We detected the IgG antibody titers of immunized sera against SARS-CoV-1-RBD, MERS-CoV-RBD, and SARS-CoV-2-RBD via ELISA. The trivalent-DNA/VTT group, SARS1-DNA/VTT group, and SARS2-DNA/VTT group demonstrated specific binding antibodies against SARS-CoV-1-RBD, with geometric mean titers (GMTs) of 271,529, 76,195, and 20,159, respectively. In contrast, the SARS-CoV-1-RBD-specific binding antibody titer in the MERS-DNA/VTT group was only 126. The GMT of SARS-CoV-1-RBD-specific binding antibodies was highest in the trivalent vaccine group, significantly higher than that in the MERS-DNA/VTT group (P = 0.0057) (Figure 1B). Trivalent-DNA/VTT and MERS-DNA/VTT groups exhibited robust specific binding antibody responses against MERS-CoV-RBD protein, with geometric mean titers (GMTs) of 912,280 and 724,077, respectively. In contrast, the SARS1-DNA/VTT and SARS2-DNA/VTT groups elicited significantly lower specific binding antibody responses against MERS-CoV-RBD protein, with GMTs of 178 and 112, respectively. Statistical analysis revealed that the GMT of specific binding antibodies against MERS-CoV-RBD protein in the Trivalent-DNA/VTT group was significantly higher than that in the SARS1-DNA/VTT group (P = 0.0236) (Figure 1C). The results indicated that both the Trivalent-DNA/VTT and SARS2-DNA/VTT groups exhibited high titers of SARS-CoV-2-RBD-specific binding antibodies, with geometric mean titers (GMTs) reaching 1,024,000 and 1,149,401 respectively. In contrast, the SARS1-DNA/VTT group displayed moderate levels of SARS-CoV-2-RBD-specific binding antibody, with a GMT of 11,404. Notably, the MERS-DNA/VTT group demonstrated the lowest level of SARS-CoV-2-RBD-specific binding antibodies, exhibiting a GMT of merely 126, which was significantly lower than that observed in the trivalent-DNA/VTT and SARS2-DNA/VTT groups (P = 0.0151 and P = 0.0445, respectively) (Figure 1D). The results showed that the trivalent vaccine group developed strong and broad-spectrum binding antibody activities, while the monovalent vaccine groups only elicited a high level of binding antibody response against themselves.

To further investigate the humoral immune responses induced by the vaccines, we also assessed the levels of neutralizing antibodies in the mice. The trends of neutralizing antibodies in the vaccine groups were similar to those of binding antibodies (Figures 1E–G). The trivalent-DNA/VTT group exhibited robust neutralizing responses against SARS-CoV-1, SARS-CoV-2, and MERS-CoV, with neutralizing antibody GMTs of 3256, 4720, and 15963, respectively. In contrast, the SARS1-DNA/VTT group elicited low titers of neutralizing antibodies against SARS-CoV-2 and MERS-CoV, with GMTs of only 20 and 71, respectively. The SARS2-DNA/VTT group, while effective in generating neutralizing antibody against SARS-CoV-2, failed to neutralize MERS-CoV and displayed limited neutralizing capacity against SARS-CoV-1(GMT = 135). Conversely, the MERS-DNA/VTT induced robust neutralizing responses against MERS-CoV (GMT = 28,452) but exhibited virtually no neutralizing activity against SARS-CoV-2 and SARS-CoV-1. Statistical analysis revealed that the GMTs of neutralizing antibodies against SARS-CoV-1 and SARS-CoV-2 in the Trivalent-DNA/VTT group were significantly higher than those in the MERS-DNA/VTT group (P = 0.0033, P = 0.0469, respectively) (Figures 1E-G). Furthermore, the neutralizing responses against MERS-CoV in MERS-DNA/VTT group were markedly higher compared to those from the SARS2-DNA/VTT group (P = 0.0042). Notably, the neutralizing reactions against SARS-CoV-2 in the SARS2-DNA/VTT group were significantly higher than those in the MERS-DNA/VTT group (P = 0.0405). The trivalent vaccine demonstrated robust neutralizing capability against three types pseudoviruses of SARS-CoV-1, SARS-CoV-2, and MERS-CoV, whereas the neutralizing antibodies elicited by monovalent vaccines exhibited comparable neutralizing potency to homologous pseudoviruses as those by the trivalent vaccine, yet significantly weaker neutralizing capacity against heterologous pseudoviruses.

The specific T cell responses were measured using an ELISpot assay. The trivalent-DNA/VTT group exhibited moderate levels of IFN-γ⁺ T cells (653 spot forming unit/10⁶ splenic lymphocytes) against the SARS-CoV-1 S peptide pool, which were 2.32-fold (282 SFU/10⁶ splenic lymphocytes) and 2.39-fold (273 SFU/10⁶ splenic lymphocytes) higher than those induced by the SARS1-DNA/VTT and SARS2-DNA/VTT groups, respectively. Notably, the specific T cell response against the SARS-CoV-1 S peptide pool in the trivalent-DNA/VTT group was significantly stronger than that of MERS-DNA/VTT group (P = 0.0152), which had only 76 SFU/10⁶ splenic lymphocytes (Figure 2A). Both the trivalent and monovalent MERS vaccines immunized mice elicited robust MERS-CoV S-specific IFN-γ⁺ T cell responses in immunized mice. In contrast, SARS1-DNA/VTT and SARS2-DNA/VTT groups barely induced MERS-CoV S -specific IFN-γ⁺ T cell responses in mice. Specifically, the MERS-CoV S-specific IFN-γ⁺ T cell response induced by the MERS-DNA/VTT group was statistically different from that of the SARS2-DNA/VTT groups (P = 0.0376, respectively) (Figure 2B). The trivalent-DNA/VTT group generated strongly specific IFN-γ⁺ T cell responses targeting the S protein of SARS-CoV-2, which were significantly higher than that observed in the MERS-DNA/VTT group (P = 0.0093). While the SARS1-DNA/VTT group induced a lower level of specific IFN-γ⁺ T cell response against the SARS-CoV-2 S peptide pool, amounting to 221 SFU/10⁶ splenic lymphocytes. Notably, mice immunized with the MERS-CoV vaccines elicited barely any specific IFN-γ⁺ T cell response against the SARS-CoV-2 S peptide pool (Figure 2C). Compared to monovalent vaccines, the trivalent vaccine elicited strong specific IFN-γ⁺ T cell responses targeting the S proteins of SARS-CoV-1, MERS-CoV, and SARS-CoV-2.

To analyze suitable coronavirus S proteins as immunogens, we retrieved S protein amino acid sequences from representative viruses of the four coronavirus genera (Alpha, Beta, Gamma, Delta) in the NCBI database and conducted phylogenetic analysis. Nine coronaviruses known to cause human respiratory diseases belong to the Alpha (light green in the phylogenetic tree), Beta (light red) and Delta (purple) genera. Notably, the viruses - SARS-CoV-1, MERS-CoV, and SARS-CoV-2 are the Beta genus and exhibit close genetic relatedness (Figure 3A). Accordingly, we have constructed a trivalent vaccine targeting these three viruses for preparedness against potential future human coronavirus outbreaks. Trivalent coronavirus vaccines immunize mice using a Prime/Boost strategy, which not only elicits immune responses against a single viral strain but also covers different subtypes within the same genus (Figure 3B). This provides crucial evidence for the development of broad-spectrum vaccines and holds significant application value, particularly in addressing viral variations and emerging coronavirus threats.

The golden hamsters were vaccinated with DNA vaccines at week 0 and week 2, and then boosted with VVT vector vaccines at week 6. Following challenge with the SARS-CoV-2 XBB (RY166) at week 8, Figure 4A. Starting from the second day after challenged with the SARS-CoV-2 XBB (RY166), the body weight of the hamsters began to decline. In the vaccine groups, the hamsters’ body weight ceased to decline on the fifth day and started to increase on the sixth day. In contrast, the hamsters in the PBS control group experienced a persistent decline in body weight, initiating from the second day and continuing until the sixth day, with a weight reduction nearly 10% (Figure 4B). The results indicated that, starting from the fourth day post-challenge, the viral load in the pharyngeal swabs of hamsters began to decline. Compared to the PBS control group, a more reduction in viral load was observed in the pharyngeal swabs of hamsters in the vaccine group (Figure 4C). On the eighth day, the geometric mean titer (GMT) of viral load in the trivalent vaccine group dropped to 91 copies/mL. The viral load GMT in the PBS group was approximately 19-fold higher than that in the trivalent vaccine group (P = 0.0283) and was more than 8-fold higher than that in the COVID-19 vaccine group. After challenge with SARS-CoV-2 XBB (RY166), the vaccine group of hamsters exhibited a fast decline in virus load in the trachea and lung tissues compared to the control group (Figures 4D, E).

After challenge with the SARS-CoV-2 XBB variant, golden hamsters exhibited the maximum body weight loss on days 5–6 post-infection, therefore, lung tissue samples on day 5 were used for pathological section examination. Overall, among the three experimental groups, the PBS control group exhibited severe pulmonary pathological lesions, characterized by marked thickening of the alveolar septa and a substantial infiltration of monocytes and lymphocytes (Figure 5A). In contrast, the SARS2-DNA/VTT group exhibited the mildest pathological damage: the epithelial structures of the bronchi and bronchioles remained largely intact, with only mild inflammatory manifestations such as focal infiltration of monocytes and lymphocytes observed in the alveolar septa (Figure 5B). As for the Trivalent-DNA/VTT group, multiple small patchy pulmonary consolidations were detected, and the extent of inflammatory injury was slightly more severe than that in the SARS2-DNA/VTT group (Figure 5C). Collectively, these results demonstrate that vaccination can effectively mitigate lung injury induced by the SARS-CoV-2 XBB variant in golden hamsters (Figure 5D).