Identification of CD4 and H-2Kd-restricted cytotoxic T lymphocyte epitopes on the human herpesvirus 6B glycoprotein Q1 protein

The identification of Human herpesvirus 6B (HHV-6B) epitopes that are recognized by T-cells could contribute to the development of potential vaccines and immunotherapies. Here, we identified CD4+ and H-2Kd-restricted CD8+ T-cell epitopes on the glycoprotein Q1 of HHV-6B (BgQ1), which is a unique glycoprotein and essential for HHV-6B viral entry, by using in vivo electroporation with a plasmid DNA encoding BgQ1, overlapping peptides spanning the BgQ1 sequence, ELISPOT assay for quantification of gamma interferon (IFN-γ), and computer-based T-cell epitope prediction programs. The CD4+ and CD8+ T-cell epitopes identified in BALB/c mice in this study could be a good animal model system for use in the development of T-cell responses, inducing HHV-6B vaccines or immunotherapies.

HHV-6B. The T-cells responding to the identified HHV-6 antigens are present at low frequency in healthy adults and need to be expanded in vitro for use in autologous immunotherapy [16][17][18] . In addition, to design vaccines and immunotherapies for HHV-6B infection in humans, a good animal model will be needed to analyze T-cell responses against the HHV-6B antigen.
DNA vaccines can engender not only humoral but also cellular immune responses 19 . And they offer some advantages including easy construction, preparation, stability. The vaccine immunogenicity and efficacy are significantly enhanced by in vivo electroporation 20 .
In this report, we attempted to determine T-cell epitopes on BgQ1, which is a unique glycoprotein and essential for HHV-6B viral entry. Due to the similarities in the peptide-binding motifs between H-2K d and HLA-A24, BALB/c mice were chosen as the animal model hosts [21][22][23][24] . We identified an H-2K d -restricted CD8 + T-cell epitope in BALB/c mice by using in vivo electroporation with a plasmid DNA encoding BgQ1, overlapping peptides spanning the BgQ1 sequence, ELISPOT assay for quantification of gamma interferon (IFN-γ), and computer-based T-cell epitope prediction programs.

Results
Identification of CD4 + or CD8 + T-cell stimulating peptides by using the synthetic overlapping peptides from BgQ1. As a first preliminary ELISPOT assay, overlapping peptides from BgQ1 (P1-P48) were used for stimulation. Peptides were divided into two groups: the first group contained the first 24 peptides (P1-P24) and the second group contained the other 24 peptides (P25-P48). Splenocytes from BALB/c mice immunized with a plasmid DNA encoding BgQ1 were stimulated with the individual peptides for 40 h, and IFN-γ spots were counted. Independent experiments were carried out from independent mice. As shown in Fig. 1a,b, splenocytes in the presence of 12 peptides (P7, P8, P11, P16, P17, P18 from the first half, and P33, P36, P37, P38, P43, P44 from the second half) showed more than 50 SFC/1 × 10 6 splenocytes.
In a second preliminary ELISPOT assay, splenocytes from immunized mice were stimulated with the 12 peptides that showed more than 50 SFC/1 × 10 6 splenocytes in the first ELISPOT. As shown in Fig. 1c, substantial IFN-γ production was confirmed in splenocytes after stimulation with 5 peptides (P11, P16, P17, P18, and P43). The remaining 7 peptides were considered to have lower immunogenicity. These preliminary results were further confirmed by an independently performed three-color flow cytometric analysis. To reveal the responsive T-cell subset, the findings of intracellular IFN-γ staining after stimulation with these 5 peptides were examined. The results demonstrated that CD8 + T-cells mainly produced IFN-γ in response to P17 and P18 (Fig. 2). On the other hand, CD4 + T-cells mainly produced IFN-γ in the presence of P16 and P43. CD4 + T-cells also produced IFN-γ in the presence of P11, but the IFN-γ signals were lower than the other 4 peptides.
Identification of an MHC class Ia restriction molecule for P17 and P18 of BgQ1. Since P17 and P18 were found to stimulate CD8 + T-cells, we next tried to determine which MHC class Ia molecule was involved in the presentation of peptides to CD8 + T-cells. BW5147 (H2k) lymphoma cell lines expressing either H-2K d , H-2D d , H-2L d , or just the wild type molecule (H2k) were co-cultured with P17 or P18, and used for stimulation of splenocytes from BgQ1-immunized mice. IFN-γ production for both P17 and P18 was only observed in H-2K d expressing BW5147 (Fig. 3), indicating P17 and P18 contained the H-2K drestricted CD8 + T-cell epitope.
As shown in Fig. 4a,b, the 9-mer peptide (AFCPMTSKL) from the overlapping region of P17 and P18 induced the strongest intracellular IFN-γ signals in the CD8 + T-cells by flow cytometric analysis, followed by the 10-mer (IAFCPMTSKL) and 8-mer (FCPMTSKL) peptides, indicating that this 9mer peptide is the optimal H-2K d -restricted CD8 + T-cell epitope on BgQ1, in BALB/c mice. Another 9-mer peptide (KPLTAMTAI) from the front half of P17 induced only a background level of intracellular IFN-γ synthesis.

Discussion
In this study, we identified an H-2K d -restricted CD8 + T-cell epitope on the BgQ1 molecule. In addition, we found that P43 and P16 of the BgQ1 contain CD4 + T-cell epitopes in BALB/c mice.
In general, CTL plays a crucial role in protective immunity against infection with intracellular pathogens, such as certain bacteria and viruses 25,26 . Although many things are still unknown about the mechanisms of T-cell mediated protection against HHV-6B infection, the necessity of T-cells to control HHV-6B replication is suggested by the higher incidence of persistent HHV-6B viremia in patients without proliferative T-cell responses 27 . To better understand CD8 + T-cell responses against HHV-6B infection, identifying the epitopes which are recognized by CD8 + T-cells is important. However, the HHV-6B genome encodes hundreds of proteins, making the identification of immunodominant epitopes complicated. Recent reports which defined CD4 + and CD8 + T-cell epitopes by using expanded T-cells in vitro focused on the HHV-6 proteins present at high levels in virus preparations 18 , or on the HHV-6 homologies of antigens defined for HCMV 16,17 . In this study, we focused on the BgQ1 protein, which plays an essential role in HHV-6B virus entry and is a promising candidate for antiviral therapy 13 .
In the present mouse model, we identified the H-2K d -restricted CTL epitope on the BgQ1 molecule. This epitope is not conserved in HHV-6A gQ1 sequence (strain U1102; PubMed accession NC 001664), but similar www.nature.com/scientificreports www.nature.com/scientificreports/ sequence (RFCPMTTKL) which has same amino acids in the main anchor positions of nonameric K d epitope is present (position 2 and 9).
In this study, mice were immunized with a codon-optimized plasmid DNA expressing the BgQ1 by in vivo electroporation to identify the CTL epitopes. DNA vaccination is a powerful tool for identifying T-cell epitopes. However, optimization of codon usage is an important consideration in constructing DNA vaccines 26 . DNA immunization with an optimized codon has been reported to result in enhanced CTL reactivity by increasing the translational efficiency of plasmid DNA 28,29 . Moreover, an in vivo electroporation technique to induce transient and reversible permeabilization of the cell membrane improves the efficiency of plasmid DNA transfection 30 . Electroporation enhances the immunogenicity and effectiveness of DNA vaccines by increasing antigen delivery up to 1000-fold over naked DNA delivery alone 20 .
IFN-γ producing cells were detected by three-color flow cytometric analysis. It might have been better if we had used Live/Dead and other markers of CTL activity, because the percentages of IFN-γ-producing cells in this study were relatively small. These markers may influence our results. However, similar number of IFN-γ producing cells with similar system were reported in another study 31,32 and IFN-γ production was consistently detected over the background level.
We also used the computer-based algorithm programs BIMAS and SYFPEITHI to predict the epitopes [33][34][35] . The strategy including DNA vaccines, overlapping peptide and these computer-based programs is an effective methods for narrowing down the amino acid region of the T-cell epitope 31 . Prediction in these programs is restricted to for 8-11mer length peptides. It would be possible CTL can recognize much longer peptides up to 14mer 36 , however we did not pursue this long CTL peptide in this study. We also found that P43 and P16 of the Overlapping peptides (5 µg/ml for each peptide) from glycoprotein Q1 of human herpesvirus 6B (BgQ1) were used for stimulation. Peptides were divided into two groups: the first group contained the first 24 peptides (P1-P24) (a) and the second group contained the other 24 peptides (P25-P48) (b). (c) Splenocytes from immunized mice were stimulated with the selected 12 peptides (5 µg/ml for each peptide). The results were expressed as spot forming cells (SFC)/1 × 10 6 splenocytes and the wells of "not reliably countable high signals" were defined as 300 SFC/1 × 10 6 splenocytes. Pep(-) indicates medium alone. The data are the mean ± SD of duplicate (a,b) and triplicated (c) wells. And the data are representative of three independent experiments. Independent experiments were carried out from independent mice. www.nature.com/scientificreports www.nature.com/scientificreports/ BgQ1 contain CD4 + T-cell epitopes in BALB/c mice. Especially, P43 contains 15mer peptide which showed high score (VNNIFTVQARYSKQN, score 31) for binding to the H-2A d molecule by SYFPEITHI Epitope Prediction. But MHC class II prediction tools do not perform as well as class I predictions 37 , there is room for further research.
In conclusion, we identified an H-2K d -restricted CD8 + T-cell epitope on the BgQ1 molecule and two peptides which contain CD4 + T-cell epitopes in BALB/c mice. These results warrant further study to examine whether the epitopes are applicable to animal models or humans and could be utilized in vaccines or immunotherapies.

Materials and Methods
Mice. Six-week-old female BALB/c mice (16-20 g) were obtained from Japan SLC (Shizuoka, Japan). The mice were housed in-house under specific pathogen-free conditions maintained at 22 ± 2 °C and 55 ± 5% relative humidity, in a 12-hour light/dark cycle environment, and provided with food and water ad libitum. The health condition of the mice was monitored daily. A total of 25 mice were used including preliminary experiments. All  www.nature.com/scientificreports www.nature.com/scientificreports/ mice were immunized with plasmid DNA and each experiment was independently repeated three times. All of the animal experimental procedures were approved by the Kobe University Institutional Animal Care and Use Committee (Permission number: P131101-R1) and carried out according to the Animal Experimentation Regulations of Kobe University.

Construction of the plasmid DNAs, pCAAGGS-MCS BgQ1m.
To optimize the gene expression in mammalian cells, the BgQ1 original cDNA (GenBank accession no. MK388090) codon was converted to more common in human genes and cloned into the pMX plasmid by Invitrogen (Carlsbad, CA). The amplified BgQ1 DNA fragments were digested with SacI and KpnI and ligated into the digested pCAGGS-MCS vector, yielding the plasmid pCAGGS-MCS BgQ1m. The plasmid was amplified in DH5α Escherichia coli and purified using QIAGEN Plasmid Mega Kit (QIAGEN, Hilden, Germany) following the manufacturer's instructions.
Peptides. BgQ1 of the HHV-6B HST strain has 516-amino-acid (aa). Except for the signal sequence (aa 1 to 25), peptides spanning the total 491 aa BgQ1 sequence were prepared as 20-mers overlapping by 10 residues ( Supplementary Fig. S1). Forty-eight lyophilized peptides in total (P1-P48) were purchased from Eurofins Genomics (Tokyo, Japan). The purity of the peptides was 50% or more, confirmed by high-performance liquid chromatography (HPLC) and verified for correct sequence by mass spectroscopy (MS). All peptides were dissolved in dimethyl sulfoxide (DMSO) at a concentration of 10 mg/ml and stored at −80 °C until use.
Immunization of mice. Mice were immunized with a plasmid DNA expressing the BgQ1 by in vivo electroporation. After anesthesia by isoflurane, 100 µg of pCAGGS-MCS BgQ1m was injected twice into both femoral muscles of the mice, and an NEPA21 electroporator (NepaGene, Tokyo, Japan) was used for transfection. Immunizations were performed three times at 2-week intervals and the spleens were removed one week after the third immunization.
BW5147 cells retrovirally transduced with one of the genes encoding H-2K d , H-2D d , or H-2L d were made as previously described 32 , and used to determine the CD8 + T-cell epitope-presenting MHC Ia molecule. The cells were cultured in RPMI 1640 medium with 10% heat-inactivated fetal bovine serum (RPMI/10FBS) in an incubator with a humidified atmosphere containing 5% CO 2 . FBS, which originate from Canada (endotoxin level ≤ 50 EU/mL, hemoglobin level ≤ 25 mg/dL), was purchased from Gibco (cat. # 12483020, Thermo Fisher Scientific, Waltham, MA).
Preparation of splenocytes from an immunized mouse. Immunized mice were euthanized by an intraperitoneal injection of sodium pentobarbital and cervical dislocation. The spleens were aseptically removed and single cell suspensions were prepared after dissociation of the spleen through a cell strainer. Splenocytes were isolated by Ficoll density gradient centrifugation, suspended in RPMI/10FBS at a concentration of 1 × 10 7 cells/ml, and used for assay.
Quantification of IFN-γ by ELISPOT assay. The wells of a 96-well MultiScreen HA plate (Millipore, Billerica, MA) were precoated and blocked with RPMI/10FBS according to the manufacturer's instructions (Ready-Set-Go Mouse IFN-γ ELISPOT kit; eBioscience, San Diego, CA). After blocking, 50 µl of RPMI/10FBS was added to each well followed by 100 µl of splenocytes (1 × 10 6 cells) from immunized mice. All dissolved peptides in DMSO were diluted in RPMI/10FBS to a concentration of 20 µg/ml, and 50 µl of the resulting solution was added to each well. The total volume was 200 µl/well and the final concentration of each peptide was 5 µg/ml. Measurements were made in duplicate for each peptide in the first preliminary ELISPOT, and in triplicate in the second preliminary ELISPOT.
The plates were incubated for 40 h at 37 °C in a 5% CO 2 humidified incubator. IFN-γ spot-forming cells (SFC) were detected and developed according to the manufacturer's instructions (Ready-Set-Go Mouse IFN-γ ELISPOT kit; eBioscience). As a minor modification, 100 µl of 3,3′,5,5′-tetramethylbenzidine-H substrate (Moss, Pasadena, MD) was used for development of spots, and the spots were developed for 3 min at room temperature. After stopping the substrate reaction by washing wells with water, the plates were dried. The spots were then enumerated using the KS ELISPOT system (Carl Zeiss Microscopy GmbH, Jena, Germany) for automated spot counting. Because too many spots can cause signal overlapping and a reduction in counting accuracy, the wells of undercounting high signals were defined as 300 SFC per 1 × 10 6 splenocytes. Intracellular IFN-γ staining and flow cytometry. Splenocytes from the immunized mice were treated with ammonium chloride-tris (ACT) buffer for 3 min at room temperature to remove red blood cells. The cells were then washed twice with RPMI 1640 medium and resuspended in RPMI/10FBS at a concentration of 1 × 10 7 cells/ml. The cells (2 × 10 6 cells) were incubated for 4 h at 37 °C in the presence or absence of peptide at a concentration of 10 µg/ml with Golgistop stock solution (monensin solution; BD Biosciences, San Jose, CA) diluted 1:1,500. The cells were then washed twice with 2% bovine serum albumin (BSA) in phosphate buffered saline (PBS) followed by staining with fluorescein isothiocyanate (FITC)-conjugated anti-CD8 (clone KT15; MBL, Nagoya, Japan) and phycoerythrin-indotricarbocyanine (PE/Cy7)-conjugated anti-CD4 (clone GK1.5; BioLegend, San Diego, CA) monoclonal antibodies (mAbs) for 30 min at 4 °C. The cells were washed twice and then intracellular cytokine staining (ICS) was performed by using a BD Cytofix/Cytoperm kit (BD Biosciences) according to the manufacturer's protocol. ICS for IFN-γ was performed with PE-conjugated anti-IFN-γ (clone XMG1.2; BD Biosciences) mAbs for 30 min at 4 °C. The cells were washed twice, resuspended in PBS with 2% BSA, and then analyzed by flow cytometry with a SA3800 spectral analyzer (Sony Corporation, Tokyo, Japan).

Data Availability
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.