An engineered bio-nanocapsule (BNC) comprising modified hepatitis B surface antigen L protein was used as a physical scaffold for envelope protein domain III (D3) of Japanese encephalitis virus (JEV). At the N terminus, the BNC contained a two-tandem repeat of the Z domain (ZZ) derived from Staphylococcus aureus protein A (ZZ-BNC). The Lys-rich ZZ moiety exposed on the surface of ZZ-BNC was used for chemical conjugation with the JEV D3 antigen, which had been expressed and purified from Escherichia coli. Immunization of mice with D3 loaded on the surface of ZZ-BNC (ZZ-BNC:D3) augmented serum IgG response against JEV and increased protection against lethal JEV infection. The present study suggests that innocuous recombinant antigens, when loaded on the surface of ZZ-BNC, can be transformed to immunogenic antigens.
B cell receptor
envelope protein domain III
- E. coli
hepatitis B surface antigen
hepatitis B virus
human papilloma virus
Japanese encephalitis virus
- S. aureus
- S. cerevisiae
N-succinimidyl 3-(2-pyridyldithio) propionate
sulfosuccinimidyl 6-(3′-[2-pyridyldithio]-propionamido) hexanoate
a two tandem repeat of the Z domain derived from S. aureus protein A
BNC containing ZZ on the surface
D3 loaded on the surface of ZZ-BNC
Virus-like particles, being strongly immunogenic, have the potential to be effective vaccines [1-3]. In the past, vaccines such as those for HBV  or HPV [5, 6] have been developed using recombinant proteins that contain VLPs. Additionally, RTS,S, the most advanced malaria vaccine candidate to date, was developed using a VLP consisting of the HBsAg fused to the repeat domain of the circumsporozoite protein . The strong immunogenicity of VLPs is derived from their structure. This has a highly organized surface displaying a repetitive array of protein epitopes that efficiently crosslink surface Igs (BCRs) on B cells [8-11], that mediate strong signal transduction. B cells present protein epitopes through major histocompatibility complex class II–T cell receptor interactions, thus enabling effective antibody responses [12-14]. Furthermore, substances similar in size to VLPs remain stable in vivo and are transported to lymph nodes . Therefore, research on VLPs has focused on the development of various subunit vaccines . It seems logical that vaccine development using VLP technology is possible. However, the vast majority of proteins derived from pathogens are unable to form VLPs and vaccine development for such pathogens is urgently required. VLP technology could also be applicable to antigens derived from viruses that are unable the potential to form VLPs. It is possible to construct a chimeric VLP by integrating a recombinant foreign antigen into, or onto, the surface of a heterologous VLP [2, 3].
An N terminus-modified BNC with a two-tandem repeat of the Z domain (ZZ-BNC), a derivative of the B domain of protein A from S. aureus, has previously been developed [16, 17]. Recombinant ZZ-BNC was constituted with modified HBsAg L protein (ZZ-L protein), and produced in S. cerevisiae [16, 17]. ZZ-BNC has a Lys-rich domain (ZZ) on its surface. The amino-group of the Lys residue is often used as a target site for constructing protein conjugates by chemical reactions. We hypothesized that the ZZ region would be a scaffold moiety for a recombinant antigen: that it might be possible to render such a recombinant antigen more immunogenic by presenting it to B cells as a repetitive and organized array on ZZ-BNC.
In this study, we selected the E protein D3 antigen of JEV for evaluating ZZ-BNC as an antigen carrier VLP. The D3 protein cannot form a VLP by itself; therefore, its immunogenicity is low. However, the D3 antigen has been demonstrated as a potential vaccine candidate against JE [18-20]. We previously reported that D3 protein is readily expressed in E. coli . As an approach to enhancing the immunogenicity of D3 protein, we produced a chimeric BNC using the D3 protein and ZZ-BNC. We evaluated whether the BNC generated by using chemical coupling to load the D3 antigen onto the surface of ZZ-BNC, could induce protective immunity against lethal JEV challenge in mice.
1 MATERIALS AND METHODS
1.1 Expression and purification of recombinant proteins
To construct D3 protein expression plasmid, the D3-coding gene was amplified by PCR from a plasmid containing the D3 gene . Sense and antisense oligonucleotide primers containing NcoI or XhoI restriction enzyme recognition sequences (sense primer: 5′-CGC CCA TGG ACA AAC TGG CTC TGA AAG GCA CA-3′ and anti-sense primer: 5′-GCG CTC GAG CGT GCT TCC AGC CTT GTA CC-3′) were used in the PCR. The amplified fragment was digested with NcoI and XhoI and the digested fragment subcloned between the NcoI-XhoI sites of the pET-21d plasmid vector (Merck KGaA, Darmstadt, Germany) (pET-21d-D3). The D3 expression plasmid was transformed into E. coli BL21(DE3) (BioDynamics Laboratory, Tokyo, Japan). Transformants were cultured in Luria-Bertani medium containing ampicillin (50 µg/mL); protein expression was induced over 4 hrs using 0.5 mM isopropyl β-D-1-thiogalactopyranoside. Cells were harvested when the OD of the culture at 600 nm (OD600) had reached 0.5–0.6 and resuspended in sonication buffer. Ice-cold PBS (pH 7.0) was used as a sonication buffer, the cells being disrupted by a Tomy UD-201 sonicator (20 mins × 3, output 2, duty 30; Tomy, Tokyo, Japan), then centrifuged (10,000 g, 20 mins, 4°C) to prepare inclusion bodies. The inclusion bodies were suspended in a small volume of sonication buffer supplemented with 8 M urea. The solubilized inclusion bodies were centrifuged as before to remove insoluble debris. The protein solution was diluted twice with PBS containing 4 M urea and then mixed with nickel-affinity resins (GE Healthcare, Little Chalfont, UK) overnight at 4°C. Columns were filled with the resins and nickel-affinity chromatography conducted for purification. Urea was removed by gradually decreasing its concentration in the wash solution (4, 1, 0.25, 0.063, 0.016, and 0 M) during the washing process. The eluted D3 protein was concentrated using a size-exclusion membrane filter (Amicon Ultra15, MWCO 10,000; Millipore, Billerica, MA, USA) and dialyzed (Spectra/Por CE Dialysis Membrane 8–10 kDa, Spectrum Labs, Rancho Dominguez, CA, USA) against PBS 6 times at 4°C for over 3 hrs. In this study, ZZ-BNC was produced in S. cerevisiae and analyzed as previously described [17, 22, 23].
1.2 Chemical conjugation
To construct a chimeric BNC, in which the D3 antigen is loaded on the surface of ZZ-BNC, a chemical conjugation method was conducted. Briefly, 1 mg of ZZ-BNC (0.4 mg/mL) was treated with sulfo-LC-SPDP (Thermo Scientific, Rockford, IL, USA) (final concentration 20 mM) for 1 hr at room temperature to generate pyridyldithiol-activated ZZ-BNC. Separately, 2 mg of D3 (0.125 mg/mL) was treated with SPDP (Thermo Scientific) (final concentration 20 mM) for 30 mins at room temperature before treatment with DTT (final concentration 50 mM) for 30 mins at room temperature to generate sulfhydryl-exposed D3. Finally, chemically modified ZZ-BNC (1 mg) and D3 (1 mg) were mixed in PBS and incubated overnight at room temperature for conjugation (Conjugation scheme, Fig. 1a right panel). The generated fusion complex was purified using a desalting column (Zeba Spin Desalting Column; Thermo Scientific) to remove excess crosslinker and by-products and for buffer-exchange with PBS. Generation of the conjugation complex was confirmed by human IgG–ELISA using mouse ZZ-BNC or D3 antisera.
1.3 Human IgG–enzyme-linked immunosorbent assay
A human IgG–ELISA was performed to analyze the chimeric BNC as previously described . Briefly, 5 µg/mL human IgG (Sigma-Aldrich, St. Louis, MO, USA) in bicarbonate buffer (15 mM Na2CO3, 35 mM NaHCO3, pH 9.6, 50 µL/well) was used as a capture molecule for the chimeric BNC and applied to a microtiter plate (Sumilon; Sumitomo Bakelite, Tokyo, Japan). The reaction was performed at 4°C overnight. Following this, 10% skim milk in PBS was used as a blocking solution (37°C, 1 hr). Samples (500 ng of total protein/well) were added to the plate and incubated at 37°C for 1 hr. Then, mouse anti-ZZ-BNC antisera (1:2000) or mouse anti-D3 antisera (1:2000) was added and incubated at 37°C for 1 hr. Finally, anti-mouse IgG conjugated to alkaline phosphatase (1:4000; Sigma-Aldrich) and its substrate, p-nitrophenylphosphate (Bio-Rad Laboratories, Redmond, WA, USA) were added to plates. The OD415 was determined using a microplate reader (Bio-Rad Laboratories).
1.4 Immunization of mice
Seven-week-old female BALB/c mice (Japan SLC, Shizuoka, Japan) (9–12 per group) were immunized s.c. with 30 µg of recombinant D3 protein or ZZ-BNC:D3 (60 μg: 30 μg of D3 and 30 microgram of ZZ-BNC) at weeks 0, 2 and 4, with or without 200 μL of alum (Imject Alum; Thermo Scientific). According to the manufacturer's immunization protocol, 100 µL of a mouse brain-derived formalin-inactivated JE vaccine (Beijing-1 strain, Chemo-Sero-Therapeutic Research Institute, Kumamoto, Japan; a positive control) was administered i.p. to 7-week-old female BALB/c mice (12 per group) twice at 3 day intervals during week 6. Animal experimental protocols were approved by the institutional Animal Care and Use Committee, and animal experiments were conducted according to the institutional ethical guidelines for animal experiments.
1.5 Measurement of virus-specific immune response in mice
Serum IgG titers were determined by indirect ELISA using antisera induced and collected at week 6. Briefly, 96-well Sumilon plates were coated with 100 µL mouse brain-derived formalin-inactivated JE vaccine per well. The blocking solution was 10% skim milk in PBS supplemented with Tween 80. Following blocking, serially diluted antisera and AP-conjugated goat anti-mouse IgG (1:4000, Sigma-Aldrich) were added to the plate, followed by addition of substrate. After reacting for 20 mins at room temperature, the OD415 was measured using a microplate reader (Bio-Rad Laboratories). The serum IgG titer was defined as the serum dilution that resulted in an OD415 of 0.1, or the serum dilution at which a one-magnitude-higher dilution resulted in an OD415 < 0.1. The virus neutralization assay and challenge experiments were performed as previously described .
1.6 Statistical analyses
The Wilcoxon–Mann–Whitney test was used to compare antibody titers between the PBS control group and individual immunization groups or between two different immunization groups. Kaplan–Meier analyses with the log-rank test were conducted to compare survival rates of JEV-infected mice between the PBS control group and indicated immunization groups. All statistical analyses were conducted using JMP software version 8.0 (SAS Institute, Cary, NC, USA).
2.1 Protein expression and purification
The D3 protein was expressed by E. coli in inclusion bodies. These were dissolved in 8 M urea, and protein refolding was performed using a nickel-affinity resin column. ZZ-BNC was produced in S. cerevisiae, which was purified and analyzed as previously described [17, 22, 23]. These prepared D3 antigen and ZZ-BNC (Fig. 1a, left panel) were used for chemical conjugation. Endotoxin concentrations were less than 50 pg/µg of protein (Pyrogent Single Test Vials; Cambrex, East Rutherford, NJ, USA).
2.2 Construction of chimeric bio-nanocapsule
To construct a chimeric BNC (ZZ-BNC:D3), the D3 antigen was coupled with ZZ-BNC by chemical conjugation (Fig. 1a, right panel). The generated complex was evaluated by human IgG–ELISA (Fig. 1b). It was evident that ZZ-BNC:D3 reacted more strongly with both anti-D3 and anti-ZZ-BNC mouse serum than did the D3 antigen or a mixture of ZZ-BNC and D3 (Fig. 1b). Because human IgG–ELISA can detect only the molecule that included the Z domain, this finding suggests that the D3 antigen was loaded on ZZ-BNC. The Z-averages (diameter) and ζ-potentials of ZZ-BNC:D3 were measured with the dynamic light-scattering model Zetasizer Nano ZS , and found to be 48.1 nm (polydispersity index, 0.193) and −6.88 mV, respectively. This finding suggests that ZZ-BNC:D3 retained a particle structure. The generated chimeric BNC was used for further immunological experiments.
2.3 Immune response induced by chimeric bio-nanocapusle in mice
BALB/c mice were immunized s.c. with the D3 antigen or ZZ-BNC:D3, with or without alum adjuvant. Serum IgG titers were examined to determine the reactivity to the formalin-inactivated JE vaccine (Fig. 2). We found that immunization with ZZ-BNC:D3 induced higher IgG response than did the D3 antigen alone.
The ability of the induced mouse antisera to prevent viral infection in vitro was examined by virus neutralization assays . The average titers of the induced neutralizing antibodies ranged from 50 to 150 (Fig. 3a). In particular, the titers of the ZZ-BNC:D3 with alum adjuvant were significantly higher than those of the D3 antigen with alum adjuvant (P < 0.01, Fig. 3a).
2.4 Protective efficacy of chimeric bio-nanocapsule against viral infection
Whether ZZ-BNC:D3 immunization could protect BALB/c mice against JEV challenge was evaluated using the JEV JaGAr01 strain (50 × LD50) (Fig. 3b). Mice immunized with formalin-inactivated JE vaccine (positive control) were completely protected against lethal challenge, whereas those immunized with PBS (negative control) were completely unprotected. In the absence of adjuvant, the mice immunized with ZZ-BNC:D3 were moderately protected (50%), which was better than the protection provided by the D3 antigen alone (10%). In the presence of alum adjuvant, ZZ-BNC:D3 provided better protection (70%) than that provided by the D3 antigen alone (44.4%). The survival rate for the ZZ-BNC:D3 immunization group, regardless of the presence or absence of alum adjuvant was significantly higher than that for the PBS control group (P = 0.002) (Fig. 3b).
Vaccination is one of the most effective strategies for protection against infectious diseases. A large number of vaccines in current use contain either attenuated or inactivated forms of the pathogen they are protecting against. Although attenuated and inactivated vaccines are much safer than the pathogen itself and are commonly excellent inducers of immune responses, there is always a slight risk to regain to a virulent phenotype in vivo, which would be detrimental to the host. Therefore, a focus of modern vaccinology has been to develop vaccines based on non-infectious subunits, such as recombinant proteins or epitopes . However, these recombinant proteins are intrinsically weak antigens, necessitating an efficient strategy for enhancing immune responses. To overcome this, various approaches such as VLP technology are being developed to increase the immunogenicity of proteins [2, 3]. If the target antigen can form a VLP, that VLP can be used as an antigen. However, if the target antigen cannot form a VLP structure, the target antigen can be fused to a heterologous VLP to form a chimeric VLP.
In this study, we examined the possibility of constructing a chimeric VLP based on ZZ-BNC as a vaccine platform. There are two approaches to construction of a chimeric VLP. The first is the genetic fusion method, where a foreign antigen is inserted at the N- or C- terminal, or internal region of a subunit protein of a parental VLP. Small and non-conformational epitopes are often used as foreign antigens for generation of chimeric VLPs because the size or conformation of the inserted antigen influences VLP assembly . The alternative approach is chemical conjugation, where it is necessary to prepare two materials, a parental VLP and foreign antigen. This approach is advantageous as it can display comparatively large antigens, conformational epitopes or non-proteinaceous antigens.
ZZ-BNC is derived from HBV VLP . VLP has strong immunogenicity and assembly ability. Because ZZ-L protein fused with a foreign protein can form a nanocapsule , the genetic fusion method can be used to generate a chimeric BNC. In addition, because the ZZ-moiety in the N terminus of the ZZ-L protein is exposed to the periphery of ZZ-BNC, this moiety is ideal site for chemical coupling to a foreign antigen. Furthermore, many Z moieties (approximately 120 ZZ molecules) exist as conjugate sites on the surface of BNC . This abundance of conjugate sites permits generation of an effective chimeric BNC with high load-carrying capacity.
Analysis of neutralization antibody responses (Fig. 3a) suggests that ZZ-BNC:D3 provides a suitable structure for the display of native conformational D3 epitopes, thereby enhancing production of neutralizing antibodies and protecting against JEV challenge (Fig. 3b). We predict at least two reasons for the protective effects of ZZ-BNC:D3.
First, the generated chimeric BNC has a high-molecular weight, which can enhance the immunogenicity of protein antigens [26, 28]. Previously, we investigated enhancement of immunogenicity of a high-molecular weight soluble D3 protein aggregate generated by self-crosslinking and compared it to the monomeric D3 antigen (unpublished data). The high-molecular weight D3 soluble aggregate provided moderate protection (approximately 50–60%) against JEV challenge, whereas the monomeric D3 antigen provided low protection (approximately 20–30%). These results suggest the importance of high-molecular weight proteins in the enhancement of immunogenicity. Loading the D3 antigen on ZZ-BNC resulted in the creation of a high-molecular weight D3 protein moiety. We estimated that the protective effect conferred by ZZ-BNC:D3 is partially attributable to its high-molecular weight. In addition, the protective effect of protein aggregation by chemical self-crosslinking was lower than that of ZZ-BNC:D3. Because the scales of molecular weight differ between soluble protein aggregates alone and BNCs, the immunogenicity of ZZ-BNC:D3 may be stronger than that of the soluble D3 aggregate.
Second, the protective effect of chimeric BNC may be attributable to its highly repetitive surface structure and organized display of D3 antigens: these features permit efficient cross-linking BCRs and a strong constitutive activation signal [29, 30]. The optimal distance between epitopes for B cell activation is approximately a minimum of 20–25 epitopes arranged at 5–10 nm intervals . The diameter of ZZ-BNC is approximately 30 nm and ZZ-BNC is constituted with approximately 120 ZZ-L protein molecules. Thus, the distance between each ZZ-L protein molecule would be approximately 4–5 nm. Because we loaded the D3 antigen on the ZZ moiety, ZZ-BNC:D3 may have a suitable epitope arrangement with regard to distance. Additionally, ZZ-BNC contains phospholipids such as phosphatidylcholine and phosphatidylethanolamine [22, 23] that stimulate innate immunity similarly to TLR ligands. These properties may account for the strong immunogenicity of the chimeric BNC.
In this study, we achieved a degree of success using the D3 antigen to coat BNC. However, because this strategy was not fully effective, a more effective approach is required. Recently, we reported that a targeting molecule displayed by a ZZ-BNC and LP combination strategy increases immunogenicity of the D3 antigen . LPs allow the packaging of pattern recognition receptor molecules such as TLR ligands, which can enhance innate immunity. Thus, in the future, a combination of BNC and a targeting function or LP utilization may be more successful for the development of a BNC-based vaccine.
This work was supported by the Program for Promotion of Basic Research Activities for Innovative Biosciences of the Bio-oriented Technology Research Advancement Institution.
The authors have no conflicting financial interests.