Protection of the receptor binding domain (RBD) dimer against SARS-CoV-2 and its variants

ABSTRACT The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) recombinant protein vaccines have been widely used in the real world and shown good protective effects. A vaccine prepared from the ancestral SARS-CoV-2 receptor-binding domain (RBD) homodimer was previously made a candidate in view of its effectiveness in rodents and nonhuman primates. Here, we report that the RBD homodimers of ancestral SARS-CoV-2 as well as the variant RBD dimers from the Beta, Delta, Lambda, Omicron, and Omicron sublineages, which were rapidly prepared using our universal dimeric protein platform, elicit both strong immunogenicity and good protection in vivo. The ancestral RBD vaccine was verified to provide cross-protection against the SARS-CoV-2 Delta variant from lethal challenge. A heterogeneous booster with Omicron BA.1 dimeric RBD vaccine based on a two-dose ancestral vaccine prime reduced the viral loads in Omicron BA.1 virus-challenged animals. In addition, vaccines prepared from dimeric Omicron XBB.1.5 RBD completely protected the mice from lethal challenge by Omicron XBB.1.16 and reduced the viral infection in the respiratory tract of Syrian hamsters. Thus, RBD homodimer vaccines can confer good protection against SARS-CoV-2 and its variants when used in homogeneous or heterogeneous boosting schemes. IMPORTANCE Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) variants achieved immune escape and became less virulent and easily transmissible through rapid mutation in the spike protein, thus the efficacy of vaccines on the market or in development continues to be challenged. Updating the vaccine, exploring compromise vaccination strategies, and evaluating the efficacy of candidate vaccines for the emerging variants in a timely manner are important to combat complex and volatile SARS-CoV-2. This study reports that vaccines prepared from the dimeric receptor-binding domain (RBD) recombinant protein, which can be quickly produced using a mature and stable process platform, had both good immunogenicity and protection in vivo and could completely protect rodents from lethal challenge by SARS-CoV-2 and its variants, including the emerging Omicron XBB.1.16, highlighting the value of dimeric recombinant vaccines in the post-COVID-19 era.

activities/tracking-SARS-CoV-2-variants).Due to the interaction between viral evolu tion, such as viral variation and combination, and human interventions, including human activities, physical quarantine, usage of personal protective equipment, vaccine inoculation, and therapeutic use, mutations frequently occurred in the SARS-CoV-2 spike (S) protein and especially in the receptor-binding domain (RBD) (3).Mutations that influence transmissibility, virulence, or immunity are of great concern to both scientists and the public (4).For instance, D614G in the S protein enhanced the infectivity, competitive fitness, and transmission of the SARS-CoV-2 Alpha variant (5-7); K417N, E484K, and N501Y in the S protein of the Beta variant increased the binding affinity for the ACE2 receptor, thus increasing the risk of transmission and reducing neutralization (8)(9)(10); and both K417N and E484A were predicted to have an overwhelmingly disruptive effect, making the Omicron variant more likely to cause breakthrough infections (11,12).
To address this unprecedented, complex, and volatile epidemic, researchers, as well as manufacturers, spared no effort in developing vaccines to contain the spread of the epidemic as quickly as possible (13,14).Fortunately, dozens of COVID-19 vaccines, including inactivated vaccines, recombinant protein vaccines, mRNA vaccines, and viral vector vaccines were developed and proved to effectively reduce the fatality rate or severe disease rate in clinical trials (15).Nevertheless, whether the existing vaccines on the market or in development can provide cross-protection against the circulating and emerging SARS-CoV-2 variants is unclear (13).Because vaccine development usually lags behind the viral emergence and the pace of vaccine development is often slower than the pace of viral evolution, new strategies to design broad-spectrum vaccines that perhaps have some foresight for emerging SARS-CoV-2 variants are emerging (16,17).In addition, some compromise solutions with feasibility and practicality, such as new immunogen designations, immunogen combinations, and homogeneous or heteroge neous boosting, which aim to achieve high neutralizing titers or provide diverse epitopes to enable cross-protection and reduce the breakthrough infections, matter equally (18)(19)(20)(21).
In our previous study, a homodimer protein strategy based on our dimeric protein platform was adopted to develop a SARS-CoV-2 vaccine that was characterized by high quality, low cost, strong immunogenicity, and good protection (22,23).Using an Fc tag at the C-terminus, ancestral SARS-CoV-2 RBD homodimers were able to form through a pair of disulfide bonds, and then the protein was converted into a recombinant protein without the tag by utilizing a thrombin cleavage site at the C-terminus of the RBD, which is a completely different method from the tandem strategy employed by ZF2001 (24).Using this dimeric protein platform, the RBD homodimer vaccine was rapidly prepared and entered a phase II clinical trial.Here, we prepared several variant RBD homodim ers from SARS-CoV-2 Beta, Delta, Lambda, and Omicron variants as candidate vaccines using the same strategy and evaluated the efficacy of the updated vaccines against an emerging Omicron sublineage.This study provides a universal strategy for dimeric RBD vaccine development and highlights the reliability and feasibility of dimeric RBD vaccines against SARS-CoV-2.

Preparation and characterization of RBD homodimers
The genetic sequence of the SARS-CoV-2 S protein has undergone many mutations and deletions during the evolution from the ancestral virus to the Omicron variant.The RBD, which is the active center of the S protein, mediates the binding of viruses to host receptors and has the highest mutation frequency.The Beta, Delta, and Lambda RBDs had two or three mutations in comparison to the ancestral RBD, while 15 mutations occurred in the Omicron RBD (Fig. 1A).Using our universal dimeric protein platform (22), RBD proteins of ancestral SARS-CoV-2 and its variants including Beta, Delta, Lambda, and Omicron BA.1 were prepared as no-tag homodimers.In brief, the introduction of an Fc tag to the C-terminus of the RBD promoted dimer formation.Then, the tag was removed by thrombin digestion, and naked RBD homodimers linked by disulfide bonds were obtained.
As shown in Fig. 1B, the sizes of the reduced and nonreduced RBD dimers were approximately 30 and 60 kDa, respectively, which further confirmed the type of connection mode inside the dimer.And the small differences in size between the RBDs RBD dimers to hACE2-his was analyzed and is listed in the table.The corresponding K D s were calculated using BIAcore T200 Evaluation 3.0 software with the "1:1 binding" model as the curve fitting method.

Full-Length Text
Journal of Virology were due to the glycosylation modification, which was demonstrated by PNGase F assay in Fig. S1.Using size exclusion chromatography (SEC), we confirmed that the purity of the five RBD dimers exceeded 97% (Fig. 1C).Furthermore, we verified the exposure of major antigenic sites of the five RBD dimers by surface plasmon resonance (SPR) assays on the human ACE2 receptor protein.As expected, the five RBD dimers all showed high affinity for ACE2, and the K D ranged from 1.2 to 4.4 nM (Fig. 1D and E).These data indicated that the ancestral and variant RBD homodimers exhibited good properties, which enabled the preparation of vaccines.

Immunogenicity of RBD dimers in BALB/c mice
RBD dimers were prepared into ancestral, Beta, Delta, Lambda, and Omicron BA.1 vaccines using a formulation that included an aluminum hydroxide (AL) adjuvant, and the vaccines were tested in BALB/c mice according to a three-shot regimen with 2-week intervals as the schedule presented in Fig. 2A.As detected by enzyme-linked immuno sorbent assay (ELISA), the five vaccines all elicited high specific binding antibody titers to the RBDs, especially after the second and third shots, which led to titers of up to ~1:10 6 and ~1:10 7 , respectively (Fig. 2B).Sera collected from the ancestral, Delta, and Omicron BA.1 vaccine groups after the third shot were used for the detection of cross-binding antibodies against the five RBDs.As shown in Fig. 2C, the binding of serum antibodies from the ancestral or Delta vaccine group to the Omicron BA.1 RBD was obviously weaker than the binding to the other three RBDs.Moreover, the binding of serum antibodies from the Omicron BA.1 vaccine group to ancestral, Beta, Delta, and Lambda RBDs was weaker than the binding to the Omicron BA.1 RBD itself.This finding indicates that a narrow spectrum of cross-reactive antibodies is induced by the Omicron BA.1 RBD.Furthermore, sera collected from the five groups after the third shot were used for cross-neutralizing antibody detection against ancestral SARS-CoV-2 and its variants through the authentic virus plaque reduction neutralization test (PRNT).As shown in Fig. 2D, the PRNT 50 s of serum antibodies collected from the ancestral, Beta, Delta, and Omicron BA.1 vaccine group against the ancestral viruses varied, exhibiting a trend toward lower levels of antibodies from the Omicron BA.1 vaccine in particular, which was consistent with the previous reports (25,26).Moreover, the PRNT 50 s of serum antibodies from the ancestral, Beta, and Delta vaccine groups against the Omicron BA.1 virus decreased to varying degrees.Notably, serum antibodies from the Omicron BA.1 vaccine group only showed prominent neutralization of the Omicron BA.1 virus itself and exhibited low cross-neutralizing activity against the ancestral, Beta, and Delta viruses.
In addition, splenocytes from ancestral vaccine-vaccinated mice were stimulated with either ancestral or Omicron BA.1 RBD to observe the CD4 + T-cell response detected by enzyme-linked immunospot assay (ELISPOT).In Fig. 2E, in comparison to the ancestral RBD mRNA vaccine, our ancestral RBD dimer vaccine principally triggered a Th2-biased immune response, which was reflected by the significantly increased release of IL-4 and IL-10 rather than IFN-γ or IL-2.Moreover, the release of cytokines stimulated by Omicron BA.1 RBD was weaker than that stimulated by ancestral RBD, which may imply the presence of a small number of conserved epitopes between ancestral and Omicron BA.1 RBD.These results collectively indicated that the RBD dimers perform well in triggering an antibody response, despite the cross-reactive antibody being weak between Omicron and the previous variants as well as the ancestral RBD.

The ancestral RBD vaccine provided cross-protection against the SARS-CoV-2 Delta variant in a lethal challenge model
Human ACE2-transgenic C57BL/6J mice were vaccinated with the ancestral RBD vaccine following a regular procedure and challenged with the ancestral SARS-CoV-2 or Delta variant virus as described in Fig. 3A.As in previous experiments, sera were collected after each vaccination to measure the levels of specific or cross-binding antibodies by ELISA and to detect neutralizing antibodies by PRNT.Mice were all challenged with 1 × 10 3 plaque-forming units (PFUs) of either the ancestral or Delta virus after the third Consequently, the ancestral RBD vaccine could elicit cross-binding and cross-neutral izing antibodies in C57BL/6J mice as well as in BALB/c mice after the second and third vaccinations.Specifically, the titer of binding antibodies targeting the ancestral RBD after the third vaccination was ~1:10 7 and that targeting the Delta RBD dimer was ~1:10 6 (Fig. 3B).Correspondingly, the titer of neutralizing antibodies targeting the ancestral virus produced by the ancestral vaccine was as high as 1:7,403 after the third vaccination, while that targeting Delta was 1:1,053 (Fig. 3C).These data showed that the levels of cross-binding and neutralizing antibodies against Delta variant declined while remaining high.
After the challenge, body weight loss was sustained until death occurred at 4-8 dpi in the vehicle groups; in contrast, mice in the vaccinated groups all survived within 14 days after the challenge with either the ancestral SARS-CoV-2 or Delta variant (Fig. 3D and E).By analyzing the tissues of mice at the endpoint, we found that the viral loads in the lung and brain as well as nasal turbinate were all under the detection limit, indicating the complete clearance of viruses at 14 dpi (Fig. 3F, G, and H).In line with the viral loads, the lesions in the lung and brain were significantly alleviated and only symptoms that were difficult to eliminate in a short time remained (Fig. 3I, J, and K).These results demonstrated that the ancestral RBD vaccine can provide cross-protection from SARS-CoV-2 Delta variant infection.

Heterogeneous booster with Omicron BA.1 RBD vaccine can defend against SARS-CoV-2 Omicron BA.1 infection
Considering that the ancestral RBD vaccine has difficulty in providing sufficient crossprotection for the Omicron variant, we further investigated the efficacy of an alternative vaccination regimen for defending against Omicron, as presented in Fig. 4A.Based on a two-shot ancestral vaccine prime, a heterogeneous booster with the Omicron BA.1 vaccine could slightly improve the titer of binding antibodies to Omicron BA.1 RBD, while the Delta vaccine could not (Fig. 4B and C).Meanwhile, the booster with the Delta vaccine cloud did not markedly elevate the levels of neutralizing antibodies against the Omicron BA.1 virus, whereas the booster with Omicron BA.1 vaccine cloud did (Fig. 4D).
Upon finishing the vaccination procedure, mice were challenged with 5 × 10 4 PFUs of Omicron BA.1 virus, and viral load measurement and pathological assessment of the lung were performed.As a result, there were no substantial changes in weight loss within 2 days after the Omicron BA.1 virus challenge (Fig. S2A).The viral load in the lung was approximately 10 5 PFUs per gram at 2 dpi, and this was reduced by boosting with the Omicron BA.1 vaccine (Fig. 4E).In addition, the lesions in the lung were also obviously relieved by boosting with the Omicron BA.1 vaccine (Fig. 4F and G; Fig. S2B).These results suggested that a heterogeneous booster with Omicron BA.1 RBD vaccine could combat Omicron BA.1 infection in vivo.

The Omicron XBB.1.5 RBD vaccine reduces the infection of Omicron XBB.1.16 in the respiratory tract of Syrian hamsters
The Omicron XBB.1.5RBD dimer was prepared in the same way as the previous dimers on our dimeric protein platform (Fig. S3).Formulated with the AS03 adjuvant (27), the Omicron XBB.1.5vaccine was administrated via intramuscular (IM) injection with a regular three-shot regimen in a Syrian hamster model (Fig. 5A).Specific antibodies from the Omicron XBB.1.5vaccine group for Omicron XBB.1.5 and XBB.1.16RBD were detected by ELISA after each vaccination, and 1:10 4 was achieved after the third shot.In contrast, the titers of specific antibodies from the ancestral vaccine group were 1:10 3 (Fig. 5B and  C).In line with that, the titer of neutralizing antibodies against Omicron XBB.1.16virus from the Omicron XBB.1.5vaccine group was 1:145 after the third shot, while sera from the ancestral vaccine group had almost no neutralizing effects on Omicron XBB.1.16virus (Fig. 5D).
After the challenge with Omicron XBB.1.16virus, slight body weight changes were observed in Syrian hamsters (Fig. S4).At 3 dpi, the viral load in the nasal turbinate and lung peaked at 10 5 PFUs per milliliter, which was reduced to 10 3 PFUs per milliliter by the Omicron XBB.1.5vaccine; moreover, the ancestral vaccine could not reduce the viral loads in either the nasal turbinate or lung and the same results were observed for viral copies (Fig. 5E and F).Since the tissue injury of Syrian hamsters is generally mild, we mainly analyzed the viral infection in the trachea and nasal turbinate.Consistent with the viral loads, the Omicron XBB.1.16virus infection in the trachea and nasal turbinate was obviously lessened by the Omicron XBB.1.5vaccine, while this did not occur with the ancestral vaccine (Fig. 5G and H).
From the neutralizing test, the neutralizing titer of the Omicron XBB.1.5vaccine group was evaluated with the vaccination times and reached at 1:336 after the second shot, while the sera from the Omicron XBB.1.5vaccine group had no neutralizing effect on the ancestral virus (Fig. 6C).
After the challenge by ancestral or Omicron XBB.1.16virus, mice continued to lose weight and died at 4-8 dpi in the vehicle groups, while all mice survived in the Omicron XBB.1.5vaccine-vaccinated group challenged by Omicron XBB.1.16virus; in contrast, none survived in the ancestral virus-challenged group (one reached the end of mercy) (Fig. 6D and E).From the viral load results, we observed that live virus in the lung, brain, and nasal turbinate was cleared by the Omicron XBB.1.5vaccine but could not be cleared by the ancestral vaccine, and the viral copies showed similar results (Fig. 6F, G, H, and I).Tissues such as lung, brain, and nasal turbinate at the endpoint were taken for analysis, and the Omicron XBB.1.5vaccine relieved the lesions in both the lung and the brain (Fig. 7A, C, and D), and significantly reduced the Omicron XBB.1.16virus infection in the nasal turbinate (Fig. 7B).The results highlight the cross-protection of the Omicron XBB.1.5RBD vaccine against the emerging SARS-CoV-2 variant, Omicron XBB.1.16.

DISCUSSION
The ongoing COVID-19 pandemic caused by SARS-CoV-2 has taken a substantial toll and profoundly impacted the human society.As the epidemic spread, the SARS-CoV-2 Omicron variant became the predominant circulating variant since it was first discovered in South Africa in November 2021 (29).In particular, the Omicron variant contains more than 60 mutations in its genome and more than 30 mutations in the S protein (30), implying its strong immune escape, thus highlighting the need for updated vaccines.
Our study evaluated the immunogenicity and protection of the dimeric RBD vaccines from variants such as Beta, Delta, Lambda, Omicron BA.1, and XBB.1.In addition, it is worth mentioning that the ancestral RBD vaccine and Omicron XBB.1.5vaccine can hardly induce cross-reactivity and especially cross-protection between each other.In our study, ancestral RBD-based vaccine induced limited crossreactive and -neutralizing antibodies against XBB variants, and four out of five ancestral RBD vaccine-vaccinated mice naturally died after XBB.1.16challenge, only one reached the end of mercy and was put to death, this means that the rare effective epitopes provided by ancestral RBD, which work in either humoral immune response or cellular immune response, thus cannot provide reliable protection.It is clear that the neutralizing antibody level is a basic and major indicator for evaluating the protection potency (31), rare evidence supports that only cellular immune responses can provide complete protection.However, it is different from that making targets other than the highly diverse RBD, such as the conserved nucleoprotein, as antigen for inducing a cellular immune response, which does provide some protective effect (32).Anyhow, the limited cross-protection between the ancestral and Omicron variants emphasizes the specificity of the emerging variants, which needs ongoing attention, and specific vaccine updates are necessary.
Besides, we demonstrated that heterogeneous boosters may be a practical vaccina tion regimen, considering that almost all the people have been vaccinated or infected.Since COVID-19 vaccines developed based on ancestral SARS-CoV-2 have a reduced ability or even no ability to neutralize the Omicron variant (33)(34)(35), some findings suggest that the vaccines are still able to elicit robust T-cell responses against the Omicron variant, and a booster dose of vaccine could recall and expand the preexisting memory immune response against SARS-CoV-2, as well as the de novo induction of immune responses, resulting in protection against Omicron infection (36,37).Therefore, a homogenous or heterogeneous boosting strategy may be able to combat emerging variants such as Omicron, with the advantage that it can improve the titers of neutraliz ing antibodies as well as provide diverse epitopes that play a key role in both B and T-cell responses (38).These findings may also explain why moderate neutralizing antibodies could still provide some in vivo cross-protection using the boosting strategy in our study and clinical trials reported elsewhere (39,40).
In the past 3 years, numerous SARS-CoV-2 vaccines have been applied to populations around the world and great positive results have been achieved with joint efforts.They are mainly prepared by recombinant protein, mRNA, adenovirus-vectored and virus-inac tivated technologies, construction of diverse antigen ingredients and adjuvants, and work by what looks like different immune responses (41).Among them, the recombinant protein vaccines were well-researched both in preclinical and clinical trials, which were based on tandem RBD dimer (42), S timer (43), IFN-α-fused RBD dimer (44), and some other reported antigen designation.The major advantage of the recombinant protein vaccine, including our vaccines reported here, is that they are safe and technically mature, but the strong immunogenicity often depends on suitable adjuvants.Three adjuvants, including aluminum, AS03, and c-GAMP, were employed in our study, despite that they all worked well with the RBD dimers in triggering antibodies and providing protection, a comprehensive comparison of their immune mechanism is worth further study.
The adenovirus-vector vaccine and nucleic acid vaccine are more effective in stimulating the cellular immune response (45) reflected by Th1-type immune cytokines release, while the protein subunit vaccine including the one reported here and the inactivated-virus vaccine mainly induced Th2-type immune responses.Th2-type immune response is helpful for antibody production, while its excessive and persistent activation may lead to tissue damage.In view of that vaccines generally work through short-term local immune stimulation to form long-term immune memory, so this has not occurred with SARS-CoV-2 vaccines including our reported one in the clinical or preclinical trials so far.Definitely, the risk should be monitored along with the virus variation to ensure adequate safety.
After all, this study demonstrated the effectiveness of the RBD dimer as a valid antigen for the SARS-CoV-2 vaccine and the versatility of the dimeric protein platform, while there are small limitations, such as a lack of consistency in comparisons of different adjuvants; thus, which combination of adjuvant and RBD dimer is optimal cannot be determined.Additionally, the BALB/c or C57BL/6J mice seem to respond better to RBD dimer vaccines than Syrian hamsters, which is reflected by both the antibody response and in vivo protection; therefore, a suitable model is critical for vaccine evaluation.In addition, the detailed mechanism of different vaccination routes in different antigenadjuvant combinations is also worthy of further study in the future.
Vero E6 cells (donated by Prof. Zhengli Shi from the Wuhan Institute of Virology) were maintained in Dulbecco's modified Eagle's medium (DMEM, Gibco, NY, USA) supplemen ted with 10% fetal bovine serum (FBS, Gibco, NY, USA) and cultured at 37°C with 5% CO 2 for SARS-CoV-2 passage, titration, and neutralization.The ExpiCHO-S cells were cultured in a chemically defined serum-freeCHO medium and ExpiCHO Expression Medium (Gibco, NY, USA) for RBD protein expression.

Surface plasmon resonance
SPR assays were carried out using a BIAcore T200 (GE Healthcare).The ancestral, Beta, Delta, Lambda, and Omicron BA.1 RBD dimers were immobilized on CM5 chips (GE Healthcare).Human ACE2-his was diluted to 20, 10, 5, 2.5, and 1.25 nM to bind to RBD dimers.The interaction was assayed using a flow rate of 30 µL/min, an association time of 120 seconds, and a dissociation time of 300 seconds.The chip was regenerated with pH 1.5 glycine solution.The respective K D values were calculated using BIAcore T200 Evaluation 3.0 (software) with "1:1 binding" as the curve fitting method.

Vaccination and challenge
BALB/c mice (n = 8 or n = 6) were vaccinated with AL-adjuvanted RBD dimers (10 µg per mouse) according to a homogeneous prime-boost-boost protocol with 2-week intervals.For the ELISPOT assay, the ancestral RBD-LNP mRNA vaccine (2 µg per mouse) was made as a control, which was applied according to a conventional prime-boost protocol with a 3-week interval.Blood samples were collected from the ophthalmic vein 7 or 14 days after each vaccination.Specifically, 200 µL of formulated RBD vaccine (ancestral, Beta, Delta, Lambda, or Omicron BA.1) was injected into the hind leg muscle (IM) of each mouse.Whole blood was collected from mice and stored at 4°C, and sera were separated after centrifugation at 4°C.Syrian hamsters (n = 4) were vaccinated with Omicron XBB.1.5RBD dimers (10 µg per hamster) formulated with AS03 via IM injection according to a homogeneous three-dose prime-boost-boost protocol with 2-week intervals.Syrian hamsters that underwent the vaccination program were challenged with 1 × 10 4 PFUs of Omicron XBB.1.16virus, and lung, nasal turbinate tissue, and trachea were obtained at 3 dpi for viral detection.
Adult hACE2-transgenic C57BL/6J mice (n = 6, n = 3, or n = 5) were vaccinated with RBD dimers (10 µg per mouse) formulated with AL via IM injection according to a three-dose homogeneous prime-boost-boost protocol or a two-dose homogeneous prime accompanied by one-dose heterogeneous booster protocol with 2-week intervals, or vaccinated with Omicron XBB.1.5RBD dimers formulated with 2′, 3′-cGAMP intra nasally according to a homogeneous two-dose prime-boost protocol with a 3-week interval.Serum samples were collected every 2 weeks or 7 or 10 days after the final vaccination.
Human ACE2-transgenic C57BL/6J mice that underwent the vaccination program were transferred to the A3 lab for challenge.A total of 5 × 10 4 PFUs of Omicron BA.1 virus were diluted with DMEM and used to inoculate each animal via the nasal cavity after isoflurane anesthesia via the airway in nonlethal models, and three mice were executed for viral detection at 2 dpi.For the lethal challenge, 1 × 10 3 PFUs of the ancestral and Delta virus, or 1 × 10 4 PFUs of Omicron XBB.1.16virus were used.Bodyweight and behavior were monitored during the whole experimental period.After the challenge, lung, brain, or turbinate tissues were dissected for virus detection or histopathology examination at the endpoint.

Enzyme-linked immunosorbent assay
A 96-well polystyrene high-binding flat-bottom plate (Greiner, Germany) was coated with ancestral, Beta, Delta, Lambda, Omicron BA.1, Omicron XBB.1.5,or XBB.1.16RBD at 1 µg/mL with a volume of 100 µL/well overnight at 4°C.The plates were blocked with 2% nonfat dried milk at room temperature for 1 hour after washing with PBST (phosphate buffer solution with 0.1% Tween-20) three times.Then, 10-fold gradient-diluted sera were added to each well, and incubation was performed at room temperature for 1 hour.Then, the plates were incubated with mouse or hamster horseradish peroxidase-conju gated secondary antibodies at room temperature for another hour after washing five times with PBST.Then, 100 µL/well of tetramethylbenzidine solution (Proteintech, China) was added after washing five times as above.Ten minutes later, 100 µL of 2 M HCl was added to each well, and the absorbance at 450 nm was measured with a Synergy H1 microplate reader (BioTek, USA).Samples with values greater than twice those in the controls were considered positive.

Enzyme-linked immunospot assay
BALB/c mice (female, 6-8-week-old, n = 6) were vaccinated with AL adjuvant, ancestral RBD dimer vaccine, or ancestral RBD mRNA vaccine.The AL adjuvant and ancestral RBD dimer vaccine (10 µg/mouse) were administrated three times at 2-week intervals, and the RBD mRNA vaccine was administrated twice at a 3-week interval, as described in a previous report (46).Seven days after the last vaccination, splenocytes (5 × 10 5 /well) from each mouse were added to the IFN-γ/IL-2/IL-4/IL-10 antibody precoated plate kit (MabTech, USA) following the manufacturer's instruction.Then, ancestral or Omicron BA.1 RBD protein was added to the wells.After incubation for 36 hours, the cells were removed, and the plates were processed in turn with a biotinylated detection antibody, followed by incubation with streptavidin and chromogenic substrate.Afterward, the liquid was poured, and the plate was washed with water to stop the color development process.The ELISPOT plate was placed on the holder, and the parameters for spot counting were adjusted.Images were captured and processed with an ImmunoSpot S6 reader (Cellular Technology Limited, USA).

Plaque reduction neutralization test
Vero E6 cells seeded in a 24-well plate grown to monolayer confluence were used for plaque formation.Specifically, gradient-diluted sera in 200 µL DMEM were incubated with 400 PFUs of ancestral, Beta, Delta, or Omicron BA.1 and Omicron XBB.1.16SARS-CoV-2 in 200 µL DMEM at 37°C for 1 hour.Then, 150 µL of the mixture was used to inoculate each well of a 24-well plate with two replicates, and incubation was performed for 1 hour at 37°C.Then, the infectious mixture was completely removed from the cells, and 1 mL of 0.9% methylcellulose-2% FBS-DMEM was added to each well.Four or five days later, the plate was soaked with 8% formaldehyde overnight and stained with 0.1% crystal violet.The plaques were manually counted.

Viral load measurement in the tissues
Viral loads in the lung, brain, or nasal turbinate tissues were either measured by quantitative RT-PCR (qRT-PCR) for viral copies or plaque formation to measure the viral titers.Using a common protocol, the right lung or brain, the whole or half nasal turbinate was homogenized in 1 mL DMEM, 200 µL was used for viral RNA isolation according to the protocol of the RNeasy Mini kit (Qiagen, Germany), and the total RNA was eluted with 30 µL RNase-free water.Seven microliters of eluted product was transcribed in a 20 µL reaction system, and then 1 µL cDNA was used as a template for qRT-PCR with a HiScript II Kit (Vazyme, China).Viral copies were quantified using 1 µL of cDNA by a standard curve method on QuantStudio I (ABI, USA) with a pair of primers targeting the S gene.The standard curve was set from seven points in a 20 µL reaction system (2.35 × 10 8 copies, 2.35 × 10 7 copies, 2.35 × 10 6 copies, 2.35 × 10 5 copies, 2.35 × 10 4 copies, 2.35 × 10 3 copies, and 2.35 × 10 2 copies).Samples with <2.35 × 10 2 copies were defined as negative.The number of copies in positive samples was converted with the following equation: sample well copies × 20 × 30/7/0.2mL/weight in grams for the lung or brain in 1 mL, and this normalized value is presented in the figure with limit of detection (LOD, 1 × 10 6 copies/gram or mL).
For the detection of live virus in tissue, homogenized fluid was directly used for infection or 10-fold gradient dilution (LOD, 670 PFUs/gram for lung, 335 PFUs/gram for brain, and 67 PFUs/mL for nasal turbinate), and 150 µL of diluted fluid was used to inoculate Vero E6 monolayer cells seeded in 24-well plates with two replicates.The remaining steps were the same as the description of the PRNT protocol.

FIG 1
FIG 1 Preparation and characterization of RBD dimers.(A) Mutations in the ancestral, Beta, Delta, Lambda, and Omicron BA.1 RBDs (amino acids 319-541) and a flow chart of the preparation of RBD dimers.(B) Analysis of the five RBD dimers by reduced or nonreduced SDS-PAGE and Coomassie brilliant blue staining.(C) Analysis of the five RBD dimers by size-exclusion chromatography; the corresponding purities are indicated in the graph.(D and E) The binding affinity of five

FIG 2
FIG 2 Cross-reactive immune response induced by the five RBD dimer vaccines.(A) Vaccination scheme.Six-to eight-week-old female BALB/c mice (n = 8 or 6) were vaccinated with ancestral, Beta, Delta, Lambda, and Omicron BA.1 RBD dimers formulated with AL at a dose of 10 µg per mouse on days 0, 14, and 28.Serum samples were obtained from orbital veins 7 or 14 days after each vaccination.(B) The titers of specific binding antibodies targeting the corresponding immunogens in the five vaccinated groups were measured by ELISA.(C) Cross-binding antibodies in the ancestral, Delta, and Omicron BA.1 vaccine-vaccinated groups targeting the five immunogens were detected by ELISA, and statistical analysis was referred to ancestral, Delta, or Omicron BA.1 antigen, respectively.(D) The titers of neutralizing antibodies against the authentic ancestral, Beta, Delta, and Omicron BA.1 viruses in the five vaccinated groups were measured (Continued on next page)

FIG 2 ( 6 FIG 3
FIG 2 (Continued) by PRNT.Geometric mean titers calculated from PRNT 50 s are presented on the top of each column.(E) Splenocytes from ancestral vaccine-vaccinated BALB/c mice were stimulated with either ancestral or Omicron BA.1 RBD, and Th1/2 cytokines, such as IFN-γ, IL-2, IL-4, and IL-10, were detected by ELISPOT, and AL and ancestral mRNA vaccine were used as controls.

FIG 4
FIG 4 Heterogeneous booster with the Omicron BA.1 vaccine reduced the viral load and pathological injury in SARS-CoV-2 Omicron BA.1-challenged mice.(A) Vaccination and challenge scheme.Six-to eight-week-old male K18-hACE2-transgenic C57BL/6J mice (n = 6) were vaccinated with 10 µg RBD vaccine on days 0, 14, and 28.Serum samples were obtained from orbital veins 14 days after each vaccination.SARS-CoV-2 Omicron BA.1 viruses (5 × 10 4 PFUs) were used to inoculate mice via the nasal cavity.(B and C) Binding antibodies to the Omicron BA.1 RBD proteins in the Delta or Omicron BA.1 vaccine-boosted groups were detected by ELISA.(D) The levels of neutralizing antibodies targeting the Omicron BA.1 virus in the Delta and Omicron vaccine-boosted groups were measured by PRNT.Geometric mean titers are presented in the column indicating the third shot booster.(E) Three mice were executed for viral titer detection in the lung at 2 dpi.(F) Lung tissues from mice dissected at 2 dpi were analyzed by H&E staining.Representative images from each group are displayed.The arrow indicates an area with hemorrhage, alveolar wall thickening, and inflammatory cell infiltration.

FIG 5
FIG 5 The Omicron XBB.1.5RBD vaccine reduced Omicron XBB.1.16infection in Syrian hamsters.(A) Vaccination and challenge scheme.Six-to eight-week-old female Syrian hamsters (n = 4) were vaccinated with 10 µg Omicron XBB.1.5RBD dimers formulated with AS03 via IM injection according to a prime-boost-boost protocol with 2-week intervals.The ancestral RBD vaccines formulated with AL and AL only were made as controls.Sera were collected on days 0, 14, 28, and 35.SARS-CoV-2 Omicron XBB.1.16virus (1 × 10 4 PFUs) was used to inoculate mice via the nasal cavity on day 35.(B) and (C) Titers of specific antibodies targeting Omicron XBB.1.5 and XBB.1.16RBD were measured after each vaccination.(D) Titers of neutralizing antibodies against Omicron XBB.1.16virus were measured, and geometric mean titers are presented on the top of the columns.(E and F) The viral titers and copies in the nasal turbinate and lung were measured at dpi 3. (G and H) Immunofluorescence analysis of viral antigens in the trachea and nasal turbinate.Sections were stained with antibodies targeting SARS-CoV-2 NP (red), and DAPI was used to stain the nucleus (blue).

FIG 6
FIG 6 The Omicron XBB.1.5RBD vaccine protected mice from lethal challenge by Omicron XBB.1.16.(A) Vaccination and challenge scheme.Six-to eight-weekold K18-hACE2-transgenic C57BL/6J mice (n = 5) were vaccinated intranasally with 10 µg Omicron XBB.1.5RBD dimers formulated with 2′, 3′-cGAMP on days 0 and 21.Serum samples were obtained from orbital veins on days 14 and 28.SARS-CoV-2 (Omicron XBB.1.16, 1 × 10 4 PFUs) was used to inoculate mice via the nasal cavity.(B) The levels of cross-binding antibodies targeting the ancestral or Omicron XBB.1.5RBD were measured by ELISA.(C) The levels of cross-neutralizing antibodies targeting the ancestral or Omicron XBB.1.16virus were measured by PRNT.(D and E) Body weight and survival rate were monitored throughout the whole experimental period.(F and G) The viral copies and titers in the lung, brain, and nasal turbinate from ancestral virus-challenged mice.(H and I) The viral copies and titers in the lung, brain, and nasal turbinate from Omicron XBB.1.16virus-challenged mice.

FIG 7
FIG 7 Reduction of pathological damage and infection in Omicron XBB.1.5RBD-vaccinated mice.(A) Lung and brain tissues at the endpoint were analyzed by H&E staining.Representative images from each group are displayed.Arrows in the lung images are used to indicate areas that exhibit fibrin exudation, hemorrhage, alveolar wall thickening, and inflammatory cell infiltration, and arrows in the brain images are used to indicate areas with cellular edema and neuronal degeneration necrosis in the hippocampus.(B) Immunofluorescence analysis of Omicron XBB.1.16infection in the trachea of XBB.1.5RBD-vaccinated mice.(C and D) Accumulative pathological scores from different pathological indicators in the lung and brain were calculated for each mouse.

Full
5, and confirmed their efficacy in protecting against the emerging Omicron variant in both a lethal challenge model and a mild model.In particular, the AL-adjuvanted ancestral RBD IM vaccine completely protected mice from lethal challenges from the ancestral and Delta viruses; the AS03-adjuvanted Omicron XBB.1.5RBD IM vaccine significantly reduced Omicron XBB.1.16virus infection in the respiratory tract; and the 2′, 3′-cGAMP-adjuvan ted Omicron XBB.1.5RBD IN vaccine completely protected mice from lethal challenge from the Omicron XBB.1.16virus.These results together suggest that the RBD dimer is a good antigen for the SARS-CoV-2 vaccine, regardless of the adjuvants and vaccination routes.