Virus-like Particles of Nodavirus Displaying the Receptor Binding Domain of SARS-CoV-2 Spike Protein: A Potential VLP-Based COVID-19 Vaccine

Since the outbreak of the coronavirus disease 2019 (COVID-19), various vaccines have been developed for emergency use. The efficacy of the initial vaccines based on the ancestral strain of severe acute respiratory syndrome coronavirus type 2 (SARS-CoV-2) has become a point of contention due to the emergence of new variants of concern (VOCs). Therefore, continuous innovation of new vaccines is required to target upcoming VOCs. The receptor binding domain (RBD) of the virus spike (S) glycoprotein has been extensively used in vaccine development due to its role in host cell attachment and penetration. In this study, the RBDs of the Beta (β) and Delta (δ) variants were fused to the truncated Macrobrachium rosenbergii nodavirus capsid protein without the protruding domain (CΔ116-MrNV-CP). Immunization of BALB/c mice with the virus-like particles (VLPs) self-assembled from the recombinant CP showed that, with AddaVax as an adjuvant, a significantly high level of humoral response was elicited. Specifically, mice injected with equimolar of adjuvanted CΔ116-MrNV-CP fused with the RBD of the β- and δ-variants increased T helper (Th) cell production with a CD8+/CD4+ ratio of 0.42. This formulation also induced proliferation of macrophages and lymphocytes. Overall, this study demonstrated that the nodavirus truncated CP fused with the SARS-CoV-2 RBD has potential to be developed as a VLP-based COVID-19 vaccine.


Immunogenicity of the Chimeric VLPs
After one week of acclimatization, the mice were immunized subcutaneously at weeks 2 (first injection), 5 (first booster), and 8 (second booster) with adjuvanted or unadjuvanted (i) C∆116-MrNV-CP, (ii) C∆116-MrNV-CP β-RBD , (iii) C∆116-MrNV-CP δ-RBD , (iv) a mixture of VLPs (C∆116-MrNV-CP β-RBD and C∆116-MrNV-CP δ-RBD in equal concentrations; Mix-VLPs), and (v) HEPES buffer (negative control). Sera were collected from the mice prior to each injection, and ELISA was performed to determine the antibodies against the RBDs. Generally, anti-RBD antibodies were not detected in the sera collected before immunization with the chimeric VLPs ( Figure 4). Following the first injection with either the chimeric VLPs of C∆116-MrNV-CP β-RBD or C∆116-MrNV-CP δ-RBD or the Mix-VLPs, regardless of the presence or absence of the AddaVax adjuvant, antibodies against the β-( Figure 4b) and the δ-SARS-CoV-2 RBD (Figure 4c) were detected. The anti-β-and anti-δ-SARS-CoV-2 RBD antibodies increased by about two folds after the first booster. Subsequently, the second booster further increased the antibody levels.

Immunogenicity of the Chimeric VLPs
After one week of acclimatization, the mice were immunized subcutaneously at weeks 2 (first injection), 5 (first booster), and 8 (second booster) with adjuvanted or unadjuvanted (i) CΔ116-MrNV-CP, (ii) CΔ116-MrNV-CP β-RBD , (iii) CΔ116-MrNV-CP δ-RBD , (iv) a mixture of VLPs (CΔ116-MrNV-CP β-RBD and CΔ116-MrNV-CP δ-RBD in equal concentrations; Mix-VLPs), and (v) HEPES buffer (negative control). Sera were collected from the mice prior to each injection, and ELISA was performed to determine the antibodies against the RBDs. Generally, anti-RBD antibodies were not detected in the sera collected before immunization with the chimeric VLPs ( Figure 4). Following the first injection with either the chimeric VLPs of CΔ116-MrNV-CP β-RBD or CΔ116-MrNV-CP δ-RBD or the Mix-VLPs, regardless of the presence or absence of the AddaVax adjuvant, antibodies against the β-( Figure  4b) and the δ-SARS-CoV-2 RBD (Figure 4c) were detected. The anti-β-and anti-δ-SARS-CoV-2 RBD antibodies increased by about two folds after the first booster. Subsequently, the second booster further increased the antibody levels.  at weeks 2 (before injection), 5 (3 weeks after 1st injection), and 8 (3 weeks after 1st booster), and 9 (1 week after 2nd booster) were analyzed. The immunized sera were used to determine the presence of antibodies against the β, δ and Wu-SARS-CoV-2 RBDs coated on the microtiter plate wells. Statistical significance (p < 0.001) is denoted by letters and Roman numerals shown above the bars. Insignificant differences are indicated by the same letter. Asterisks (*) denote significant differences (p < 0.001) compared to the animal group receiving the Mix-VLPs (a mixture of C∆116-MrNV-CP β-RBD and C∆116-MrNV-CP δ-RBD VLPs) with or without adjuvant. The standard deviations of triplicate measurement are represented by error bars.
Interestingly, these antibodies also interacted with the RBD derived from the ancestral Wuhan strain (Figure 4a), although the level of antibodies was only one third of the total antibody captured by the RBDs derived from βand δ-SARS-CoV-2. With respect to the negative controls, anti-RBD antibodies were not detected in the sera collected from the mice injected with HEPES or VLPs of C∆116-MrNV-CP with or without the adjuvant. The test group inoculated with the Mix-VLPs produced higher levels of anti-β-and anti-δ-SARS-CoV-2 RBD antibodies when compared with the groups immunized separately with C∆116-MrNV-CP β-RBD and C∆116-MrNV-CP δ-RBD (Figure 4b,c).

Determination of Mouse Splenocytes by Immunophenotyping
Mouse spleens collected one week after the second booster (week 9) were used to analyze the presence of cytotoxic T lymphocyte (CTL; CD3 + and CD8 + ) and T helper (Th; CD3 + and CD4 + ) cells. The highest percentage of CD3 + CD4 + cell population was observed in the splenocytes isolated from the mice administered with the Mix-VLPs regardless of the presence of adjuvant, followed by those from the test groups inoculated with the VLPs of C∆116-MrNV-CP β-RBD and C∆116-MrNV-CP δ-RBD with or without the adjuvant (Table 1). The negative control groups inoculated with C∆116-MrNV-CP, regardless of the presence of adjuvant, also triggered proliferation of significantly high levels of CD3 + CD4 + and CD3 + CD8 + compared to the buffer-inoculated mice. Figure 5a shows that the CD8 + /CD4 + T-cell ratio was significantly altered in the test groups immunized with the VLPs of C∆116-MrNV-CP β-RBD or C∆116-MrNV-CP δ-RBD or the Mix-VLPs, with or without adjuvant, compared to the negative control groups. Additionally, high levels of macrophage population were also detected in animal groups immunized with the VLPs of C∆116-MrNV-CP β-RBD or C∆116-MrNV-CP δ-RBD or the Mix-VLPs, with or without the adjuvant, whereas such macrophage elevation was not observed in the negative control groups ( Figure 5b).

Quantification of Cytokines
Cytokine quantification with multiplex ELISA showed that the mice immunized with the chimeric proteins exhibited significant elevations in all the serum cytokines tested, except IL-6 ( Figure 6). The increments were more obvious in the animal groups immunized with the VLPs of CΔ116-MrNV-CP β-RBD or CΔ116-MrNV-CP δ-RBD or Mix-VLPs, with or without the adjuvant, than in the negative control groups inoculated with HEPES buffer or the VLPs of CΔ116-MrNV-CP, with or without the adjuvant. The result indicated that the presence of SARS-CoV-2 RBD on the surface of nodavirus VLPs, regardless of the variants, could elicit the production of IL-5, IL-12p70, IFN-γ, and TNFα. Likewise, mice that received the VLPs of CΔ116-MrNV-CP β-RBD or CΔ116-MrNV-CP δ-RBD , with or without adjuvant, also effectively induced the cytokines effectively compared to mice injected with the buffer and the VLPs formed by CΔ116-MrNV-CP, with or without adjuvant. The result also revealed that both pro-inflammatory Th1 (IFN-γ, IL-12p70, and TNF-α) and anti-inflammatory Th2 (IL-5) cytokines were elevated following administration of VLPs displaying the RBDs of the β-or δ-variants.

Quantification of Cytokines
Cytokine quantification with multiplex ELISA showed that the mice immunized with the chimeric proteins exhibited significant elevations in all the serum cytokines tested, except IL-6 ( Figure 6). The increments were more obvious in the animal groups immunized with the VLPs of C∆116-MrNV-CP β-RBD or C∆116-MrNV-CP δ-RBD or Mix-VLPs, with or without the adjuvant, than in the negative control groups inoculated with HEPES buffer or the VLPs of C∆116-MrNV-CP, with or without the adjuvant. The result indicated that the presence of SARS-CoV-2 RBD on the surface of nodavirus VLPs, regardless of the variants, could elicit the production of IL-5, IL-12p70, IFN-γ, and TNFα. Figure

Discussion
The COVID-19 pandemic has resulted in substantial morbidity and death across the globe. Multiple variants have emerged due to the high mutability and transmissibility of the virus. The original wild-type strain isolated in Wuhan, China, and its subsequent variants with higher transmissible rates have heightened the need for global vaccinations. Various types of vaccines have been speedily developed to prevent further spread of the virus and reduce the severity of the disease. Current SARS-CoV-2 vaccines available in the market include the mRNA-based vaccines (Comirnaty and Spikevax), viral-vectorbased vaccines (COVISHIELD and Ad26.COV2.S), and inactivated virus vaccines (Coro-naVac and Covaxin). The effectiveness of these vaccines was reported to be up to 95%

Discussion
The COVID-19 pandemic has resulted in substantial morbidity and death across the globe. Multiple variants have emerged due to the high mutability and transmissibility of the virus. The original wild-type strain isolated in Wuhan, China, and its subsequent variants with higher transmissible rates have heightened the need for global vaccinations. Various types of vaccines have been speedily developed to prevent further spread of the virus and reduce the severity of the disease. Current SARS-CoV-2 vaccines available in the market include the mRNA-based vaccines (Comirnaty and Spikevax), viral-vector-based vaccines (COVISHIELD and Ad26.COV2.S), and inactivated virus vaccines (CoronaVac and Covaxin). The effectiveness of these vaccines was reported to be up to 95% [1, 24,25]. However, there have been concerns around the world about the lower effectiveness against new variants, particularly VOCs [26]. The first VOC, namely the Alpha variant (α-SARS-CoV-2, known as B.1.1.7 variant), was identified in the UK. Alpha-SARS-CoV-2 contains an amino acid substitution of D614G, a point mutation on the viral spike glycoprotein [27]. This variant is more transmissible than the wild-type Wuhan strain, and it has caused a greater rate of mortality in the UK [27]. Besides the α-variant, the vaccine effectiveness of the ChAdOx1 nCoV-19 vaccine against the β-variant (lineage B.1.351 isolated in South Africa) dropped from 89.3% to 21.9%, and, in separate studies, collectively, the vaccine effectiveness of it dropped to 62% [11,28]. Another study showed that the vaccine effectiveness of Novavax vaccine (NVX-CoV2373) against the B.1.1.7 and B.1.351 variations dropped to 85.6% and 60%, respectively, in a phase III trial [29]. These data attest the decreasing efficacy of these licensed COVID-19 vaccines against the new variants.
In this study, the P-domain of the full-length of MrNV-CP was removed by deleting 116 amino acid residues from its C-terminus. Removal of the P-domain allows the fusion peptide to be efficiently displayed on the surface of the VLPs formed by the truncated capsid protein, C∆116-MrNV-CP. The RBDs of βand δ-VOCs were separately displayed on the outer surface of the chimeric VLPs assembled from C∆116-MrNV-CP β-RBD and C∆116-MrNV-CP δ-RBD , respectively. The results demonstrated that these chimeric proteins, with M r of~52 kDa, were successfully expressed and purified, as analyzed with SDS-PAGE. STEM analysis revealed that the diameters of the VLPs formed by these chimeric proteins were similar to those of C∆116-MrNV-CP. Several studies showed that short peptides, such as the IAV M2e, HBV 'a' determinant, and GE11 peptide that fused to the full-length MrNV-CP, had no effect on the diameter of the VLPs formed [20,30,31]. Our previous study showed that the fusion of the domain III of the Japanese encephalitis virus (JEV) envelope protein, consisting of 133 amino acid residues at the C-terminus of the full-length MrNV-CP, drastically reduced the diameter of the VLPs from 30 to 18 nm [21]. In the present study, the VLPs formed by C∆116-MrNV-CP β/δ-RBD also demonstrated a similar reduction in diameter to~18 nm. Nevertheless, it is still unsure how the foreign epitopes are displayed on the C∆116-MrNV-CP. A high-resolution three-dimensional structure of these chimeric VLPs would provide insights into this enigma.
In the immunogenicity assay, ELISA was used to determine the level of anti-RBD antibodies induced by the VLPs of C∆116-MrNV-CP β-RBD and C∆116-MrNV-CP δ-RBD and the Mix-VLPs in the presence or absence of the AddaVax adjuvant. The antisera of mice immunized with the chimeric VLPs were captured by the RBDs of the Wu-, β-, and δ-strain coated on the microtiter plate wells. The result showed that the antibody specific against the β-SARS-CoV-2 RBD was elicited in mice either immunized with the VLPs of C∆116-MrNV-CP β-RBD or the Mix-VLPs with or without the AddaVax adjuvant. Similarly, the antibody against the δ-SARS-CoV-2 RBD was also produced in mice inoculated with the VLPs of C∆116-MrNV-CP δ-RBD or the Mix-VLPs, regardless of the presence or absence of the AddaVax adjuvant. Interestingly, the Wu-RBD coated on the plate showed a lower affinity to antisera elicited by the VLPs displaying the β-RBD or δ-RBD and the Mix-VLPs with or without the adjuvant. This result indicated that the anti-RBD antibodies produced in mice immunized with the VLPs of C∆116-MrNV-CP β-RBD and C∆116-MrNV-CP δ-RBD demonstrated a lower affinity towards the Wu-SARS-CoV-2 RBD.
A good understanding of the T cell responses towards the chimeric VLPs is important to develop effective COVID-19 vaccines. In this study, the fusion of RBDs at the C-terminal end of C∆116-MrNV-CP enhanced the Th cell population. The Mix-VLPs in the presence of the AddaVax adjuvant showed the highest level of CD4 + Th cells. At the same time, the level of CTLs showed no significant difference compared with that of other study groups except for the negative control groups inoculated with the HEPES buffer or the VLPs of C∆116-MrNV-CP with or without the adjuvant. Most of the immunized groups showed activation of CD4 + Th activity. Previous studies using the VLPs formed by the full-length MrNV-CP as nano-carriers for various epitopes showed that the activation of CD4 + cells was common compared to CD8 + cells [20,30,45]. Similar observations were also reported in SARS-CoV-2 infection, which corresponded well with the ability of CD4 + T cell immune responses in suppressing initial SARS-CoV-2 infections [46][47][48]. In managing SARS-CoV infection in animal models, CD4 + T cells were noted to be more predominant than CD8 + T cells [49,50]. Compared to CD4 + and CD8 + T cells, SARS-CoV-2-specific CD4 + T cells showed a strong correlation with reduced COVID-19 severity [47]. In acute COVID-19 cases, rapid activation of SARS-CoV-2-specific CD4 + T cells was shown to cause milder symptoms with faster viral clearance [51]. On the other hand, the lack of CD4 + T cells specific to SARS-CoV-2 has been linked with more severe cases (>day-22 post-symptom onset in some patients) [47,51,52]. Due to their capacity to destroy virus-infected cells, CD8 + T cells are essential for viral clearance.
Macrophages are among the first immune cells to interact with viruses that invade the human body. In the SARS-CoV-2-induced inflammatory response, monocytes and macrophages play a critical role in clearing the virus by the innate immune response [53]. The present study showed that macrophage activation occurred in all the immunized mouse groups except the negative control group inoculated with the HEPES buffer. Theobald et al. [54] reported that the SARS-CoV-2 S protein triggers macrophage activation. Interestingly, the activation of macrophages depends on the types of VLPs. Han et al. [55] reported that norovirus VLPs could not efficiently increase dendritic cell (DC) maturation and antigen presentation, but they are capable of activating macrophages. Another study by Lenz et al. [56] demonstrated that the papillomavirus VLPs could induce activation of DCs. Overall, activation of macrophages or DCs is antigen-receptor-dependent, which facilitates the binding and uptake of VLPs [57].
Blanco-Melo et al. [58] demonstrated a different inflammatory response associated with SARS-CoV-2 infection in COVID-19 patients. Individuals with comorbidities are more likely to have an "inappropriate and inadequate immune response". This might encourage viral replication and exacerbate symptoms associated with illness severity [58]. IFN-γ is a type II interferon generated by lymphocytes such as CD4 + and CD8 + T cells, Treg cells, CD8 T cells, FoxP3 + , NK cells, and B cells [59]. Thus, IFN-γ is a central antiviral immune mediator. A study indicated that the blood IFN-γ level was greater in COVID-19 patients than that in healthy individuals. The increased levels of this and other cytokines could likely be due to the activation of Th1 and Th2 cells [60]. In this study, a high titer of IFN-γ was detected in the mice inoculated with the Mix-VLPs with AddaVax adjuvant, followed by those immunized with this formulation without the adjuvant. This is likely due to the presence of the two RBDs from βand δ-SARS-CoV-2. Sun et al. [61] showed that the IFN-armed RBD dimer enhanced the production of IFN-γ by the CD4 + T cells. In the same study, the mice immunized with monomeric RBD gave rise to lower IFN-γ compared to the dimeric RBD.
DCs and macrophages release IL-12p70 in response to microbial stimuli such as viral infections. In combination with IL-15, IL-18, and type I IFN, the cytokine increases NK cells' cytotoxicity and causes IFN secretion [62]. In this study, all the adjuvanted study groups except C∆116-MrNV-CP with adjuvant showed high levels of IL-12p70. Several studies showed that the adjuvanted vaccines enhanced the stimulation of IL-12p70 [7,63]. Moreover, IFN-γ, IL-12p70, and TNF-α were stimulated by most of the COVID-19 vaccines, indicating the activation of these cytokines by the viral proteins, particularly the S protein [64][65][66].
Inflammation is the initial innate immune response triggered by pro-inflammatory cytokines or chemokines. Therefore, excessive production of these pro-inflammatory molecules during an infection could lead to an overwhelmed immune response known as the cytokine storm, which may lead to acute respiratory distress syndrome (ARDS) or multiple organ failure. TNF-α and IL-6 are commonly involved in inflammation during SARS-CoV-2 infection [67]. The frequency of specific TNF-α induced by the Mix-VLPs with the adjuvant in the current study is similar to that reported in clinical trials of COVID-19 vaccines such as SARS-CoV-2 FINLAY-FR-1A dimeric-RBD recombinant vaccine and recombinant spike protein nanoparticle vaccine [68,69]. A mixture of VLPs with the adjuvant enhanced the population of TNF-α-producing effector T cells, indicating induction of cellular immunity. COVID-19 mitigation relies heavily on T cell responses. CD4 + T cells are effector cells that produce IFN-γ, TNF-α, and other cytokines, besides collaborating with B cells [70,71]. TNF-α released by effector T cells and innate immunity cells can destroy virus-infected cells. Multiplication of T cells, cytokine generation, and host survival are all aided by the cellular responses [68].
After immunization with either the VLPs of C∆116-MrNV-CP β-RBD or C∆116-MrNV-CP δ-RBD or the Mix-VLPs, the mice secreted more IFN-γ, IL-12p70, and TNF-α than IL-5 or IL-6, suggesting a Th1-dominant response. Although IL-5 and IL-6 were elicited, their titers were lower than the Th1 cytokines. Overall, immunization with the Mix-VLPs in the presence of the AddaVax adjuvant elicited a higher Th1 response than Th2 response in BALB/c mice.

Construction of Truncated MrNV-CP without the Protruding Domain (C∆116-MrNV-CP)
The coding fragment of MrNV-CP without the P-domain (C∆116-MrNV-CP) was amplified from the plasmid pTrcHis-TARNA2, containing the coding sequence of fulllength MrNV-CP [17]. The primers used in PCR are as follows: Forward primer: 5 -GGGTAAACCATGGCCCTTAACATCAAGATGGCTAGAGGTAAA-3 (underlined sequence indicates NcoI restriction site), and reverse primer: 5 -TTTTTGAATTCGCCCTTCCCTAACT GTGAAATTTCCACTGGTGT-3 (underlined sequence indicates EcoRI restriction site). The PCR reaction was carried out with the Velocity DNA polymerase (Bioline, London, UK). Initially, the DNA was denatured at 98 • C for 5 min, followed by DNA amplification with 35 cycles of denaturation, annealing, and extension at 98 • C for 30 s, 61 • C for 1 min, and 72 • C for 1 min, respectively, and completed with a final extension at 72 • C for 10 min. Both the purified PCR product and plasmid pTrcHis2 TOPO were digested with NcoI and EcoRI restriction enzymes and purified using the Gel Cleanup Kit (Qiagen, Hilden, Germany) based on the recommended protocol. Ligation of the digested pTrcHis2 TOPO plasmid and purified PCR product was performed with T4 DNA ligase (Promega, Madison, WI, USA) overnight at 4 • C. The ligated plasmid was introduced into E. coli TOP10 competent cells (Thermo Fisher Scientific, Waltham, MA, USA) using the heat-shock method.

Expression and Purification of Chimeric VLPs
The expression of the C∆116-MrNV-CP was adapted from a previous study [17]. E. coli containing the recombinant plasmid was grown in Luria Bertani (LB) broth containing ampicillin (100 µg/mL) overnight at 37 • C. The overnight culture (20 mL) was transferred into LB broth (1 L), and cultured at 37 • C at 200 rpm until the OD 600 of the culture reached 0.6. The culture's temperature was lowered to 25 • C before adding 1 mM IPTG, and the culture was incubated for 5 h. The bacterial cells were recovered by centrifugation at 8000× g for 10 min, and the protein was purified as described by Goh et al. [17].

Synthesis of the Coding Sequences of the Wuhan, Beta, and Delta Variants of the SARS-CoV-2 RBD
The DNA fragments containing the SARS-CoV-2 RBD coding sequences of the ancestral Wuhan strain and βand δ-VOCs were synthesized. The RBD coding sequence of the Wuhan strain was obtained from 2019-nCoV WHU01 (GenBank accession no: MN988668.1), whereas the coding sequences of the βand δ-variants were obtained from hCoV-19/Malaysia/IMR_WC75452/2021 (GISAID accession no: 1263540) and hCoV-19/Malaysia/IMR_035575/2021 (GISAID accession no: 2931924), respectively. To enable ligation of the synthesized Wuhan and β-variant RBD coding sequences to the plasmid containing the coding sequence of C∆116-MrNV-CP, EcoRI and HindIII restriction sites were included at the 5 -and 3 -ends, respectively, whereas the coding fragment of the δ-variant RBD was designed to contain HindIII and the SnaBI recognition sequence at the 5 -and 3 -ends, respectively. The synthesized DNA fragments were ligated into the pUCIDT vector and stored in TE buffer at a final concentration of 4 µg/µL. The plasmid was introduced into E. coli BL21 (DE3) (Thermo Fisher Scientific, Waltham, MA, USA) competent cells.

Ligation of the SARS-CoV-2 RBD Coding Fragments to the pTrcHis2-TOPO Vector
The coding fragments synthesized as mentioned above were ligated into the pTrcHis2 TOPO TA expression vector. Firstly, the coding fragments were amplified by PCR using the proofreading polymerase; Velocity DNA Polymerase (Meridian Bioline, London, UK). The PCR primers used to amplify these coding fragments are listed in Table 2. The PCR reactions were performed as described in Section 4.1 except the annealing temperature was changed to 59 • C. The PCR products were then purified with the QIAquick PCR Purification Kit (Qiagen, Hilden, Germany), ligated as described above before being introduced into E. coli BL21 (DE3) (Thermo Fisher Scientific, Waltham, MA, USA) competent cells for protein expression.

Expression and Purification of Beta, Delta, and Wuhan-SARS-CoV-2 RBDs
The positive clones harboring the plasmids encoding the β-, δ-, and Wu-SARS-CoV-2 RBDs were cultured in 50 mL of LB broth supplemented with ampicillin (100 µg/mL) overnight at 37 • C. An aliquot of the overnight culture (10 mL) was transferred to 1 L LB broth containing ampicillin (100 µg/mL) in a 2 L conical flask. The culture was incubated at 37 • C at 200 rpm until the OD 600 of the culture reached~0.6. Protein expression was induced by adding IPTG (1 mM) for 5 h at 37 • C. Purification of β-, δ-, and Wu-SARS-CoV-2 RBDs was performed as described by Kumar et al. [21].

Construction of Plasmids Encoding the Beta Variant of SARS-CoV-2 RBD
The positive transformants containing the plasmids encoding the β-SARS-CoV-2 RBD were cultured in LB broth supplemented with ampicillin (100 µg/mL) at 37 • C for 18 h. The plasmids were extracted using the QIAprep Spin Miniprep Kit (Qiagen, Hilden, Germany) according to the standard protocol. Plasmids encoding the C∆116-MrNV-CP and the DNA fragment of β-SARS-CoV-2 RBD were linearized with EcoRI and HindIII, purified, and ligated with T4 DNA ligase. The ligated plasmid was transformed into E. coli BL21 (DE3) (Thermo Fisher Scientific, Waltham, MA, USA) competent cells via the heat-shock transformation method.

Construction of Plasmid Encoding the Delta Variant of SARS-CoV-2 RBD
The positive transformants carrying the plasmid encoding the δ-SARS-CoV-2 RBD were grown in LB broth supplemented with ampicillin (100 µg/mL), and incubated at 37 • C for 18 h. The recombinant plasmid was isolated with the QIAprep Spin Miniprep Kit (Qiagen, Hilden, Germany) according to the standard protocol. The coding region of the δ-SARS-CoV-2-RBD-containing restriction enzyme recognition sites was amplified using the forward primer (5'-AGGGCGAATTCGATGGTGGCGGAAATATTACAAACT-3'; underlined sequence indicates EcoRI restriction site), and the reverse primer (5 TACGTAAGCTTCAAC AGTTGCTGGTGCATGTAGAAGTTCAAAA-3'; underlined sequence indicates HindIII restriction site). The PCR reaction contained Velocity DNA polymerase (0.5 U, 0.5 µL), dNTP mix (0.2 mM, 0.25 µL), 5× Hi-Fi reaction buffer (5 µL), forward and reverse primers (10 µM, 0.5 µL), and nuclease-free water was added to a final volume of 25 µL. PCR reactions were performed as described in Section 4.1 except that the annealing temperature was changed to 60 • C. The PCR product and plasmid pTrcHis2 TOPO encoding C∆116-MrNV-CP were digested with EcoRI and HindIII restriction enzymes, ligated, and introduced into E. coli BL21 (DE3) competent cells with the heat-shock transformation method.

Expression and Purification of Truncated MrNV-CP Fused with Beta or Delta SARS-CoV-2 RBDs
Escherichia coli cells carrying the recombinant plasmids encoding for C∆116-MrNV-CP β-RBD or C∆116-MrNV-CP δ-RBD were inoculated into 50 mL LB broth containing ampicillin (100 µg/mL) and cultured overnight at 37 • C. An aliquot (10 mL) of overnight culture was transferred to 1 L LB broth and incubated at 37 • C until the culture reached an OD 600 of~0.6. Protein expressions were induced by adding 1 mM IPTG into the culture, and it was incubated at 25 • C for 5 h. The cell pellets were harvested by centrifugation at 8000× g for 8 min. Purification of recombinant proteins was modified from a previous study by Kumar et al. [21], using a fast-protein liquid chromatography (FPLC) system (Äkta Purifier; GE Healthcare, Uppsala, Sweden). The HiTap TM SP HP 1 mL column (GE Healthcare, Buckinghamshire, UK) was washed with 10 column volume (CV) of binding buffer (50 mM HEPES, 100 mM NaCl, pH 7.4). The lysates were loaded onto the column, and bound proteins were eluted via a gradient of NaCl concentration (50 mM HEPES, 1 M NaCl, pH 7.4) at a flow rate of 1 mL/min and fractionated using a Frac-950 collector. The eluted proteins were then analyzed with SDS-PAGE and Western blotting.

SDS-Polyacrylamide Gel Electrophoresis and Western Blotting
Purified proteins were analyzed with SDS-PAGE and Western blotting. The expression of C∆116-MrNV-CP, C∆116-MrNV-CP β-RBD , and C∆116-MrNV-CP δ-RBD was confirmed with western blotting using the anti-His monoclonal antibody (1:5000 dilution in TBS; MERCK, Germany) as the primary antibody and the goat anti-mouse monoclonal antibody conjugated with alkaline phosphatase (1:5000 dilution in TBS; Santa Cruz Biotechnology, Dallas, TX, USA) as the secondary antibody.

Scanning Transmission Electron Microscopy (STEM)
Purified protein samples at an appropriate concentration (~0.4 mg/mL) were applied onto copper grids for 5 min before the grids were negatively stained with 2% (w/v) uranyl acetate for 6-8 min. The grids were air-dried completely before examining under a scanning transmission electron microscope (Cold FESEM, Regulus 8230, Hitachi Co., Tokyo, Japan).

Immunophenotyping of Mouse Splenocytes
Mouse spleens were harvested at week 9 and used for immunophenotyping. In brief, PBS was added to the spleens before the samples were meshed through a cell strainer (BD Biosciences, Franklin Lakes, San Jose, CA, USA). To obtain single-cell suspensions, the meshed samples were centrifuged at 300× g for 10 min before adding erythrocyte lysis (EL) buffer (155 mM NH 4 Cl, 10 mM KHCO 3 , 0.1 mM EDTA; pH 7.4) to resuspend the cell pellet for 10 min at 4 • C. The cells were then washed with PBS (10 mL) and resuspended in PBS supplemented with 1% (w/v) BSA. The supernatant was discarded, and EL buffer was added to the cell pellet until a whitish pellet was obtained. The pellet was then gently resuspended in ice-cold PBS-BSA (1 mL).

Immunogenicity of the Chimeric VLPs
The sera collected from the mice on the second, fifth, eight, and ninth weeks were analyzed using ELISA. The RBDs of the β, δ, and Wu variants of SARS-CoV-2 were expressed and purified as described above. The purified β-, δand Wu-RBDs of SARS-CoV-2 were immobilized on a 96-well microtiter plate overnight at 4 • C. Excess proteins were removed by washing three times with TBST [0.01% (v/v) tween-20, 50 mM Tris-HCl, 150 mM NaCl; pH 7.4] before blocking with 1x milk-diluent (KPL, Milford, MA, USA; 200 µL) for 1 h at 25 • C. The wells were then washed again as above before adding serum samples (1:5000 dilution in TBS; 100 µL) and incubated for 2 h at 25 • C. Next, the wells were washed again as above before incubating with the alkaline-phosphatase-conjugated anti-mouse antibody (1:5000 dilution in TBS; 100 µL) for 2 h at 25 • C. Lastly, the wells were washed again as above before adding p-nitrophenyl phosphate (100 µL per well), and incubated for 20 min in the dark for color development. The absorbance at a wavelength of 405 nm was measured with a microtiter plate reader (BioTek, Winooski, VT, USA).

Statistical Analysis
The variations of antibody titers, cytokine quantification, and immunophenotyping were statistically analyzed using a one-way ANOVA. Duncan's multiple-range method was applied to differentiate significant differences among different groups. p-values of less than 0.05 and 0.001 are considered significant and very significant, respectively. All data analysis was performed with SPSS statistics software (IBM Corporation, Armonk, NY, USA).

Conclusions
In conclusion, the recombinant plasmids harboring the coding sequences of C∆116-MrNV-CP β-RBD and C∆116-MrNV-CP δ-RBD were successfully constructed and introduced into BL21 (DE3) competent cells. DLS and STEM showed that the E. coli-expressed C∆116-MrNV-CP β-RBD and C∆116-MrNV-CP δ-RBD self-assembled into a homogeneous population of spherical VLPs. The immunogenicity of the VLPs of C∆116-MrNV-CP β-RBD and C∆116-MrNV-CP δ-RBD was studied in BALB/s mice, and the result demonstrated the potential of these VLPs as SARS-CoV-2 vaccine candidates due to their ability to elicit humoral and cellular immune responses. Although the immunological study suggests that the VLPs of C∆116-MrNV-CP β-RBD and C∆116-MrNV-CP δ-RBD are potential COVID-19 vaccines, a virus challenge study in animal models will provide insights into the protective efficacy of the chimeric VLPs against SARS-CoV-2 infection.