Flounder (P. olivaceus) of both sexes, with an average weight of 50 ± 5 g and with a length of 12 ± 1.3 cm, were obtained from a farm in Qingdao, China. During the entire animal trial, flounder were acclimated in aquarium tanks using a recirculating seawater system at 20 ± 2°C and fed with commercial dry food pellets three times a day. All the experimental treatment were performed under natural 12-h light and dark cycle. After acclimation to the laboratory setting for 7 days, the healthy fish were used for subsequent experiments. New Zealand white rabbits (∼1 kg) and Balb/C(~30 g)mice were purchased from the Qingdao Animal Experimental Center (Shandong, China) and then used for antibody production. All animal work in this paper was conducted according to relevant national and international guidelines. Fish were anesthetized with tricaine methanesulfonate (MS222, Sigma, USA) before immunization, tissue collection, and challenge.

Polyclonal antibodies against flounder CD83, CD40, and MHCII, and monoclonal antibodies against flounder IgM and CD4 were previously produced in our laboratory (35–38), and all antibodies were diluted 1:1,000. The rabbit against OmpK antibodies were produced by Li et al. (39). The secondary antibodies included Dylight 488 (Abbkine; diluted 1:1,000) Rabbit or mouse IgG, Dylight 649 (Abbkine; diluted 1:1,000) Rabbit or mouse IgG, alkaline phosphatase (AP)-conjugated goat anti-rabbit IgG (H + L) (Abbkine; diluted 1:5,000), and HRP-conjugated goat anti-mouse IgG (Abbkine; diluted 1:5,000).

V. anguillarum (ATCC 43305/MH67) was isolated from diseased flounder (P. olivaceus) stored in the Laboratory of Pathology and Immunology of Aquatic Animals, Ocean University of China. The bacteria were grown in marine broth 2216E (5 g/L Tryptone, 1 g/L yeast extract, and 0.1 g/L C₆H₅Fe·5H₂O, pH 7.6) at 28°C to OD₆₀₀ = 0.2, and centrifuged at 2,000 g for 5 min to harvest the bacteria. Bacteria adjusted to different concentrations with PBS were used in the immunization experiment.

The CD80/86 sequence of P. fulvidraco, rainbow trout, O. niloticus, or zebrafish was used to search the flounder transcripts database published on National Center for Biotechnology Information (NCBI; http://www.ncbi.nlm.nih.gov) through BLASTn or BLASTp. Meanwhile, the published amino acid sequence of CD80/86 in teleost was compared to obtain the conserved sequence, and then the primers were designed to amplify the CD80/86 gene by using the head kidney cDNA of flounder. Based on the partial sequence searched, specific primers were designed by PrimerPremier 5.0 (Listed in Table 1 ) to extend 3’ and 5’ untranslated region (UTR) using cDNA from the spleen by rapid amplification of cDNA ends (RACE) method.

The full-length CD80/86 cDNA was assembled by DNAMAN. The potential open reading frame (ORF) was analyzed with the Finder program (https://www.ncbi.nlm.nih.gov/orffinder/). The signal peptide and the TM domain of the deduced protein sequences were predicted with the programs SignalP (http://www.cbs.dtu.dk/services/SignalP/) and TMHMM (http://www.cbs.dtu.dk/services/TMHMM/), respectively. Multiple sequence alignment was performed with the ClustalX program. Phylogenic trees were constructed using the neighbor-joining method with MEGA software (Version 6.0) and were bootstrapped 1,000 times. N-linked glycosylation sites were predicted with the NetGlycate Server (http://www.cbs.dtu.dk/services/). Secondary and 3D structures were analyzed using SMART, SWISS-MODEL, and I-TASSER (https://zhanglab.ccmb.med.umich.edu/I-TASSER/).

The RNA was extracted from seven tissues, namely, the intestine, liver, peripheral blood leukocytes (PBLs), muscle, head kidney, spleen, and gills using Trizol Reagent (Invitrogen, USA). The first-strand synthesis used the Primerscript™ First Stand cDNA Synthesis kit (Takara, China). RT-qPCR was performed using the LightCycler^® 480 II Real-Time System (Roche, Switzerland) with ChamQ Universal SYBR qPCR Master Mix (Vazyme, China). The primers and housekeeping genes are listed in Table 1 . Briefly, the cDNA was adjusted to 400 ng/ml. Each reaction consists of 10 μl of 2×ChamQ Universal SYBR qPCR Master Mix (Vazyme), 0.4 μl of forward and 0.4 μl of reverse primers (10 μM), 2 μl of cDNA, and 7.2 μl of ddH₂O. The thermal cycling profile consisted of an initial denaturation at 95°C for 30 s, followed by 40 cycles of denaturation at 95°C for 10 s, and extension at 60°C for 30 s. An additional temperature ramping step was utilized to produce melting curves of the reaction from 60°C to 95°C. The β-actin gene was used as an internal control. The 2^−ΔΔCT method was used to analyze the expression level of the CD80/86 gene.

The sequence encoding the extracellular domain except the signal peptide was amplified by polymerase chain reaction (PCR) and then embedded into pET-28a Vector (TaKaRa, Japan). After affinity purification of recombinant CD80/86 protein with a 6-histidine tag at the N-terminal end using His TrapTM HP Ni-Agarose (GE Healthcare China, Beijing, China), a total of 2.8 mg of recombinant protein was used to immunize New Zealand white rabbits (∼1 kg) and Balb/C mice (~30 g) (40). The specific procedure was followed as previously described. The antiserum was collected on the fifth day of the final immunization and purified by a protein A/protein G agarose column. The Abs titers and specificity were determined by ELISA, Western blot (WB), or mass spectrometry.

Western blotting was conducted as previously described (40). After SDS-PAGE, rCD80/86 and whole-cell protein extracts from PBLs were electrophoretically transferred onto a PVDF membrane (Merck Millipore). The membrane was blocked with 5% bovine serum albumin (BSA) in PBS for 1 h at 37°C, incubated with the rabbit anti-CD80/86 polyclonal antibodies or the unimmunized serum for 1 h at 37°C, and washed thrice with PBS-T, and then primary antibody was detected with AP-conjugated goat anti-rabbit IgG (H + L) or HRP-conjugated goat anti-mouse IgG (H + L) for 1 h at 37°C. After washing three times with PBST, the stained bands were captured by the Fusion FX Spectra (Vilber, France).

The leukocytes were isolated from peripheral blood by 34% (1.050 g/ml) to 51% (1.072 g/ml) v/v discontinuous Percoll density gradient. After being adjusted to 1 × 10⁶ cells/ml, 50 μl of cells was adhered to the slides for 2 h and then fixed with 4% paraformaldehyde for 20 min. The cells were blocked with 5% BSA at 37°C for 1 h. The CD80/86-specific antibodies mixed with mouse polyclonal antibodies MHCII (diluted 1:500), CD40, CD83, and monoclonal antibodies against flounder IgM, CD4-1, and CD4-2 were used as primary Abs at 37°C for 1.5 h. After three washes with PBS-T, the primary Abs were detected by Dylight 488 (Abbkine) rabbit or mouse IgG and Dylight 649 rabbit or mouse IgG. The nucleus was stained with 4,6-diamidino-2-phenylindole (DAPI) at room temperature for 20 min. All the Abs were diluted 1:1,000 and unimmunized serum was used as negative controls. Fluorescence images of the samples were obtained using a fluorescence microscope (Olympus DP70, Japan).

The eukaryotic expression plasmid pBudCE4.1 (Invitrogen, Carlsbad, CA, USA) contains two promoters, human cytomegalovirus (CMV) and human elongation factor 1α-subunit (EF-1α), which were used as a vector to express exogenous genes CD80/86 and OmpK. The CD80/86 from flounder was ligated into the HindIII site located downstream of CMV, and OmpK from V. anguillarum was ligated into the KpnI site located downstream of EF-1α (list in Table 2 ). The recombinant plasmids of pBudCE4.1, p-OmpK, p-CD80/86, and p-OmpK-CD80/86 were extracted using an EndoFree plasmid Kit (Tiangen, Beijing, China). The concentration of recombinant plasmid was adjusted to 200 μg/ml.

The healthy flounder were randomly divided into five groups (150 fish/group) and injected intramuscularly with 100 μl of pBudCE4.1, p-OmpK, p-CD80/86, and p-OmpK-CD80/86 recombinant plasmids, respectively. The equal volume of PBS was injected as a control treatment. To monitor changes in lymphocytes, the proportion of CD4 and IgM+ cells at different time points was examined by flow cytometry. The expression of recombinant plasmids in muscle was verified by indirect immunofluorescence. Variation of immune-related genes at the injection site was detected by qPCR.

Three fish were selected from each group on the 7th day after immunization, and muscle tissue (~0.5 cm³) was taken from the injection site. The tissues were fixed with 4% paraformaldehyde at room temperature for 2–10 min and then dehydrated with 30% sucrose solution. The tissues were infiltrated using an OCT mixture (optimal cutting temperature compound, SAKURA) and frozen using liquid nitrogen. Tissues were cut into ~7-μm-thick sections using a Leica CM 1900 slicer (Leica, German). After being blocked with 5% BSA, the sections were incubated with rabbit anti-CD80/86 or anti-OmpK, respectively. Then, the frozen sections were incubated with Dylight 649 rabbit IgG antibodies and DAPI in turn. Unimmunized rabbit serum was used as the negative control.

The lymphocytes in peripheral blood were isolated by discontinuous Percoll gradient density (1.020–1.070 g/ml) on the 1st, 7th, and 14th day, and at weeks 3, 5, and 7. The lymphocytes were adjusted to 1 × 10⁶ cells/ml and incubated with anti-CD4 monoclonal antibodies (mixed with CD4-1 and CD4-2, diluted 1:1,000) and anti-IgM monoclonal antibodies (diluted 1:1,000). Then, the cells were stained with Dylight 488 (Abbkine) mouse IgG antibody at 37°C for 1 h. Flow cytometry was performed using FACSCalibur (BD Biosciences) flow cytometers and analyzed by FlowJo 10.7 (TreeStar).

The muscle of the injection site was sampled 72 h and 168 h after vaccination. The RNA was adjusted to 1,000 ng/μl and reversed to cDNA. The expression profiles of immune-related genes including CD4-1, CD8, CD28, TNF-α, IFN-γ, IL-2, CD83, MHCI, and MHCII were investigated. The qPCR primers are listed in Table 3 and the amplification fragment size of the primers was between 150 and 250 bp and the amplification efficiency of the primers was within 90%–110%.

Sixty flounder were randomly selected from the five immunized groups at week 7. The fish were intraperitoneally injected with 1.0 × 10⁶ CFU/fish (10×LD50) V. anguillarum after the last sampling. Mortalities were recorded within 30 days after the challenge. The relative percent survival (RPS) was calculated using the following formula: RPS = {1 − (% mortality in immunized fish/% mortality in control fish)} × 100%. The log rank method was used to detect the significant difference in survival rate between the groups.

Statistical analyses were performed using Statistical Product and Service Solution (SPSS) 20.0 software (IBM, Armonk, NY, USA) with one-way analysis of variance (ANOVA) followed by t-test or Duncan’s multiple range test. All data are expressed as the mean ± standard deviation (SD). Statistical parameters and details of experiments are provided in the figure legends, and exact p-values are shown in the figures. GraphPad Prism software version 5.01 (GraphPad 265 Software, San Diego, CA, USA) was used for plotting graphs.

The full-length cDNA consists of a 35-bp 5’ untranslated region (UTR), a 906-bp ORF, and a 54-bp 3’ UTR (GenBank accession No. MT019837). The ORF sequence of CD80/86 encoded a protein of 301 amino acids with a putative molecular mass of 33.4 kDa with a predicted isoelectric point of 7.43. Seven N-linked glycosylated sites are distributed in the extracellular Ig-like structural domain ( Figure 1A ). The amino acid sequence alignment shows that CD80/86 shares overall 11.56%–48.03% amino acid identity with other teleost’ homologs, and CD80/86 in flounder is more similar to marine fish CD80/86 homologs than that in freshwater fish ( Figure 1B ). The N-J tree was constructed with bootstrap analysis (1,000 replicates) to explore the evolutionary relationship of CD80/86 between teleost and mammals, and the results showed that CD80/86 of flounder clustered into one group in teleost with high bootstrap support and demonstrated exclusivity with other B7 superfamily members, and CD80/86 in fish is closely related to B7-2 ( Figure 1C ).

The structure of CD80/86 was analyzed by SMART (http://smart.emblheidelberg.de/) and Expasy (https://www.expasy.org), as a typical type I transmembrane protein; CD80/86 in flounder is composed of a signal peptide (1–33 aa), two superfamilies’ Ig-like domain (36–130 and 132–224 aa) in the extracellular domain, a transmembrane domain (240–259 aa), and a cytoplasmic region (260–302 aa) ( Figure 1E ). Through homologous modeling using human programmed cell death 1 as templates (PDB Hit: 3FN3, C-score = −3.36, TM-score = 0.34 ± 0.11), a three-dimensional model was constructed ( Figure 1D ). The potential ligand binding site residues were predicted, and the residues 31, 62, 113, 114, 120, 121, 122, and 123 are involved in the binding of CD80/86 to its ligand. These results may indicate that ligand binding site residues mainly distributed in the IgV-like domain participate in ligand binding by folding to form steric structures. Overall, the CD80/86 in flounder was cloned and it shared similar structural features to CD80/86 homologs in other species.

The distribution profiles of CD28 at the transcriptional level were detected by RT-qPCR. Among the seven tissues, namely, the head kidney, spleen, PBLs, muscle, gill, liver, and intestine, the highest expression was detected in the head kidney, followed by the PBLs, spleen, and muscle ( Figure 2 ), while the expression of CD80/86 was lowest in the liver. To explore the distribution characteristics of CD80/86 in different cell types, the specific antibodies were produced and the average titers are higher than 1:100,000 (data not shown). The antibodies could react with the purified recombinant protein (~23 kDa) ( Figure 3A ) and a ~34-kDa band with the whole-cell protein extracts of PBLs ( Figure 3B ), which is consistent with the predicted molecular mass of cloned CD80/86. Moreover, the mass spectrometry result showed that 3 peptides were matched to the extracellular sequences ( Figure 3C ). These data suggest that the prepared polyclonal antibodies could recognize the natural CD80/86 in flounder. The isolated leukocytes in peripheral blood are shown in Supplementary Figure S1 . The distribution of CD80/86 molecules in cell subpopulations of flounder was detected by indirect immunofluorescence staining. Double-immunofluorescence staining results showed that CD80/86 was colocalized with CD83, MHCII, and CD40 on the membrane of PBLs, and not distributed on IgM+, CD3+, or CD4+ lymphocytes ( Figure 4 ). Those observations indicate that CD80/86 mainly distribute on the surface of APCs.

The bicistronic plasmid, pBudCE4.1, was used to co-express OmpK of V. anguillarum and CD80/86 ( Figure 5A ). The recombinant plasmids were amplified with the universal primers (listed in Table 2 ), and specific bands corresponding to the theoretical size bp can be observed in agarose gel ( Figure 5B ). Therefore, recombinant plasmids were successfully constructed and can be used as a vaccine to immunize. The expression of recombinant plasmids in vivo was verified by indirect immunofluorescence. After immunization, muscle tissue was sampled from the injection site. Notably, the positive signal of CD80/86 was not only observed in the p-CD80/86 and p-OmpK-CD80/86 groups but also in the pBudCE4.1-, p-OmpK-, and PBS-immunized groups on the 7th day, and more CD80/86 positive signals have been detected in the p-CD80/86 and p-OmpK-CD80/86 groups. After being injected with recombinant plasmids, the specific red fluorescent signal was only detected in the p-OmpK and p-OmpK-CD80/86 groups; no signal was observed in the pBudCE4.1, p-CD80/86, and PBS groups ( Figure 6 ). Those observations revealed that the vaccines can be effectively expressed in vivo, and the expressed proteins are recognized by prepared specific antibodies.

The expression profiles of CD4-1, CD8α, CD28, TNF-α, IFN-γ, IL-2, CD83, MHCIα, and MHCIIα in each experiment group were investigated. The expression of detected genes in the p-OmpK-CD80/86 group was higher than the groups immunized with p-OmpK alone (p < 0.05). The expression of CD4-1, CD8α, CD28, and TNF-α increased with time and IL-2 was also increased, induced by p-CD80/86 at 72 h. The higher expression of CD83, MHCIα, and MHCIIα, which are involved in antigen delivery, was detected at 72 h compared to 168 h. As shown in Figure 7 , the highest CD28 expression was monitored at 72 h in the pCD80/86 and p-OmpK-CD80/86 groups with no significant difference between the two groups (p >0.05) and then decreased at 168 h in the pCD80/86 group.

To further investigate the variations of CD4+ and IgM+ induced by the vaccines, the percentage of T/B lymphocytes was monitored at different time points ( Figure 8 and Figure S2 in the Supplementary Material ). Compared to the control group, the percentage of CD4+ cells of peripheral blood lymphocytes increased on the 7th day (p < 0.05) and reached the peak on the 14th day with 17.13 ± 1.16% and 18.46 ± 1.06% (p < 0.05) in the p-OmpK and p-OmpK-CD80/86 groups. Then, the percentage was slowly decreased on the 21st day, which is still higher than the percentage of CD4+ cells on the 7th day. The percentage of CD4+ cells did not change significantly (p > 0.05) among the immunized groups of pBudCE4.1 and p-CD80/86 on the 1st, 7th, and 21st day, while the percentage of p-CD80/86 significantly increased to 11.43 ± 0.70% (p < 0.05) on the 14th day compared to the groups immunized with PBS and pBudCE4.1 (7.5 ± 0.86% and 8.6 ± 1.07%). The variations of IgM+ cells showed a similar dynamic profile. In the p-OmpK and p-OmpK-CD80/86 groups, the levels of the IgM+ cell gradually increased, reaching the peak levels at week 5 (27.53 ± 0.98% and 30.36 ± 1.31%, p< 0.05), and then decreased slightly. Similarly, the groups immunized with PBS, pBudCE4.1, and p-CD80/86 plasmids did not show significant differences at weeks 1, 2, and 7 (p >0.05).

The fish from different immunized groups were challenged with live V. anguillarum at week 7. The death of flounder began on day 2, and rapid death occurred on days 5–8. The typical clinical features such as bleeding at the base of the fins and intestinal bleeding appeared on dead fish. After 20 days of challenge, the mortality rate stabilized in each group, and about 85%–90% of the flounder died in the PBS, pBudCE4.1, and p-CD80/86 groups with an RPS of 3.7% and 5.56%, respectively ( Table S1 in the Supplementary Material ). The vaccine of p-OmpK and p-OmpK-CD80/86 showed effective immune protection against V. anguillarum with an RPS of 46.30% and 61.11%, respectively ( Figure 9A ). The mortality rate of each group was analyzed by the log rank method, and the results showed that the mortality rate of the p-OmpK and p-OmpK-CD80/86 groups was significantly different from that of the PBS, pBudCE4.1, and p-CD80/86 groups ( Figure 9B ).