This study was approved by the Animal Care and Use Committee of Jiangsu Ocean University (protocol no. 2020-37, approval date: September 1, 2019). All procedures involving animals were performed in accordance with the guidelines for the Care and Use of Laboratory Animals in China.

The pathogenic A. hydrophila strain isolated from moribund M. amblycephala was stored at -80°C in our laboratory, which had been used in our previous studies (21). The stored A. hydrophila strain was cultured on Luria-Bertani (LB) agar plates, and single colonies of A. hydrophila were randomly isolated for expanding culture at 28°C for 16 h, which were then identified by PCR amplification and 16S rRNA sequencing (MAP, Shanghai, China). Then, the correctly identified A. hydrophila was stored at 4°C for further use.

Healthy juvenile M. amblycephala (1.12 ± 0.22 g), obtained from a fish farm in Guangzhou, China, were confirmed free of infection by examining with standard bacteriological and parasitological examination (22). Experimental fish were fed with commercial feed for temporary rearing and taming for 2 weeks in an indoor freshwater recirculating system before the culture experiment.

900 fish were divided into 5 groups (180 fish per group with 3 tanks): (I) control group (PBS instead of vaccines); (II) inactivated vaccine L group (1×10⁸ CFU/mL); (III) inactivated vaccine H group (1×10⁹ CFU/mL); (IV) OmpA vaccine L group (0.5 μg/μL); and (V) OmpA vaccine H group (1.0 μg/μL). As shown in Table 1 , for the first immunization, the emulsions of mixed vaccines and adjuvant were intraperitoneally injected with 20 μL for each fish. Then, the booster immunization was conducted 2 weeks after the first vaccination, which was directly immunized with 20 μL PBS or vaccines without adjuvant.

The experimental fish were fed 4 times daily (8:00, 11:00, 14:00, and 17:00) to apparent satiation (approximately 3% of the body weight), and the water was renewed every day to maintain suitable water quality. The water temperature was maintained at 26-28°C with the dissolved oxygen greater than 6.0 mg/L, and the pH value was approximately 7.2, while ammonia, nitrogen, and nitrite were lower than 0.1 mg/L.

The bacterial challenge assay was performed 2 weeks after the booster immunization as previously described (21), and experimental fish from each tank (60 fish/tank, 3 tanks/group) were assigned to 2 categories for calculating mortality (30 fish/tank) and sampling (30 fish/tank), respectively. The fish for sampling (body weight of 2.08 ± 0.25 g) were intraperitoneally injected with 20 μL of 1×10⁷ CFU/mL A. hydrophila (LD50 dose). Then, 3 individuals from each tank were randomly dissected after anesthetized with MS-222, and the blood, gill, hepatopancreas and intestines were collected at 0, 1, 3, 7, 14 and 30 d post infection (dpi). The fish for calculating mortality (body weight of 2.08 ± 0.25 g) were intraperitoneally injected with 20 μL of 5×10⁷ CFU/mL A. hydrophila, and the relative percent survival (RPS) was calculated using the formula: RPS = (1-% mortality in vaccinated fish/% mortality in control fish) × 100. Clinical signs and mortality were recorded everyday post infection for 14 d.

The blood was stored at 4°C overnight and centrifuged at 3000 rpm to isolate serum, which was used for Enzyme-linked immunosorbent assay (ELISA) assay and detection of enzymes activities, respectively. The hepatopancreas was homogenized and centrifuged at 4000 rpm for 10 min at 4°C for enzymes activities analysis. Immunohistochemical samples (gills, hepatopancreas and intestines) were fixed in 4% paraformaldehyde for 24 h at 4°C. Samples (gills, hepatopancreas and intestines) for RNA extraction were stored in sample protector for RNA (TaKaRa, Dalian, China) for 24 h at 4°C and then stored at -80°C until used.

The recombinant IgM (AGR34023.1), IgD (AGR34025.1), IgZ (AGR34024.1) and CD8 (XP_048023282.1) protein were prepared in the same way as OmpA. After three immunizations with purified recombinant protein, rabbit antiserum against IgM, IgZ, IgD, and CD8 were obtained and purified using Protein A/G Agarose Kit (Beyotime, Shanghai, China). The total proteins of M. amblycephala tissues were extracted with RIPA lysate (Beyotime) and the concentration was determined with a BCA kit (Beyotime).

The specificity of prepared polyclonal antibodies was verified with purified recombinant proteins and total tissue proteins of M. amblycephala by western blotting (28). Specifically, protein samples were separated on a 12% SDS-PAGE gel and transferred onto polyvinylidene difluoride (PVDF) membranes using a Trans-Blot apparatus (BioRad, Berkeley, CA, USA). Non-specific reactivity was blocked with 5% (w/v) skim milk powder in TBS (150 mM NaCl, 20 mM Tris-base, pH 7.4). The PVDF membrane was incubated with polyclonal antibodies (1: 2000) and HRP-conjugated goat anti-rabbit IgG (H+L) (Beyotime; 1: 2000) for 1 h, respectively. Finally, a DAB Kit (Beyotime) was used for detection.

The titer of serum immunoglobulin M (IgM) specific to A. hydrophila was detected by ELISA (29). In short, 96 well plates were coated with 200 μL of 5 μg/mL A. hydrophila ultrasonic fragments (same strain as used for inactivated vaccine preparation and challenge experiment) overnight at 4°C, and then blocked with 1% BSA at 22°C for 2 h, which were further incubated with M. amblycephala serum samples (1: 50 dilution) 96 well plates at 4°C for 12 h. Followed by incubating the 96 well plate with rabbit anti-M. amblycephala IgM specific antibody (1: 3000 dilution) at 22°C for 1 h, and incubating with goat anti-rabbit IgG-HRP secondary antibody (Beyotime; 1:2000 dilution) at 22°C for 1 h. After washing, the TMB substrate was added to incubate at 22°C for 5 min, then the termination solution was added to stop color reaction, and measured the absorbance at 450 nm. Internal positive and negative control samples were included. The relative OD values were calculated by the formula: Relative OD_{450 nm} = (OD_sample - OD_negative) × (average of OD_positive of all plates)/OD_positive.

Three fixed tissues from each group were dehydrated with gradient ethanol, cleaned in xylene substitute, embedded in paraffin blocks, sectioned at 4 μm thickness, and stained with H&E (Sigma, St. Louis, Missouri, USA) (21). Three random fields of each slide were captured (400×) using a light microscope (Nikon, Tokyo, Japan). Histopathological change scores were obtained by evaluating the degree of tissue lesion. The score of 0 means asymptomatic, and the higher the score, the more serious the lesion.

The expression and distribution of immune-related proteins (CD8, IgM, IgZ, and IgD) were detected by immunohistochemistry assay as previously described (28). Briefly, after dehydration, transparency and penetration, samples were embedded in paraffin. Sections (4 μm) were prepared, deparaffinized and rehydrated, then endogenous peroxidase was removed by immersion in 3% hydrogen peroxide. Sections were incubated with primary antibody (1: 2000) overnight at 4°C and then incubated with goat anti-rabbit IgG-HRP secondary antibody at 37°C for 30 min. The immunoreaction products were visualized using a DAB kit, and sections were counterstained with hematoxylin and embedded in glycerol and then positive cells were counted in three random fields using a light microscope (Nikon) with magnifications (×400 for hepatopancreas and intestine). The density of positive cells per mm² was estimated as positive cells per mm² (total of positive cells/total area counted).

Total RNAs of collected samples were extracted using the RNA Easy Fast Tissue Kit (TIANGEN, Beijing, China) and the quality and concentration were determined by agarose gel electrophoresis and NanoDrop 2000 (Thermo Fisher Scientific, Wilmington, DE, USA), respectively. In addition, cDNA was synthesized using the PrimeScript^® RT reagent Kit with gDNA Eraser (TaKaRa) following the manufacturer’s protocol.

The expression patterns of M. amblycephala immune-related genes (TNFα, iNOS, IL-1β, IL-6, IL-10, C3, SAA, TCR α, MHC I, MHC II, CD4, CD8, IgM, IgZ and IgD) and A. hydrophila 16S rRNA gene were analyzed by qRT-PCR, as previously reported (28). Briefly, qRT-PCR was performed on the ABI StepOne Plus real-time PCR system (PerkinElmer Applied Biosystems, CA, USA) using the QuantiNova™ SYBR^® Green PCR Kit (TaKaRa). All the reactions were performed in triplicate with glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as the reference gene, and the primers are listed in Supplemental Table 1 . The relative expression levels of target genes were measured in terms of the threshold cycle (Ct) value using the 2^-ΔΔCt method (30). Gene expression levels of the control group were set as 1, and that of the vaccinated groups were presented as fold change. A. hydrophila abundance in different tissues of M. amblycephala from different groups was monitored by measuring A. hydrophila 16S rRNA transcripts levels, which was calculated according to the Ct values of qRT-PCR assay (29).

In this study, the activities of lysozyme (LZM), alkaline phosphatase (AKP), and acid phosphatase (ACP) in the serum and hepatopancreas were measured according to methods previously described (21). Briefly, according to the ratio of tissue weight (g) to phosphate buffer saline (PBS) volume (mL) of 1:9, hepatopancreas were homogenized. After centrifugation at 2500 rpm for 10 min, the supernatant was separated to determine the activities of antimicrobial enzymes. The activities of LZM, AKP and ACP were detected using the corresponding enzyme activity detection kit (Nanjing Jiancheng Bioengineering Institute, Nanjing, China).

In the present study, data were presented as mean ± standard error (SE). The statistical significance was assessed by one-way analysis of variance (ANOVA), and multiple comparisons were performed using the Tukey method of SPSS 25.0. The P-value of 0.05 was considered as statistically significant difference.

The recombinant OmpA protein was mainly expressed in the inclusion body with high purity ( Figures 1A, B ). The mortality rate in the control group was as high as 85.6% post A. hydrophila infection, while those of the inactivated vaccine L, inactivated vaccine H, OmpA vaccine L and OmpA vaccine H groups were 19.8%, 36%, 15% and 21%, respectively, which were significantly lower than that of the control group ( Figure 1C ). In addition, the RPS values of inactivated vaccine L, inactivated vaccine H, OmpA vaccine L and OmpA vaccine H groups were 76.9%, 58%, 82.3% and 75.5%, respectively. These results indicated that prepared vaccines showed excellent immunoprotective effects, especially OmpA vaccine L group.

Obvious clinical symptoms were observed in the control group at 1 dpi, including congestion of fish body, hemorrhage and swelling around anus and intestines, as well as severe hemorrhage and necrosis of gills. At 3 dpi, control fish presented obvious congestion of gills, and the intestines became thinned, brittle and were easy to break. In contrast, the symptoms of vaccinated fish mainly included slight congestion of gills at 1-3 dpi, and the intestines became thinned and brittle in the inactivated vaccine groups at 1 dpi ( Supplemental Figure 1 ). After 7 dpi, all the survived fish gradually recovered and the clinical symptoms disappeared. The record of clinical symptoms revealed the immunoprotective effect of A. hydrophila vaccine.

The specificity of prepared anti-IgM antibody and no cross reaction between IgD, IgZ and IgM was verified by western blotting ( Supplemental Figures 2A, E ). No significant difference in IgM levels between control and vaccinated groups before and at 7 dpi, with just slight fluctuations. At 14 dpi, the specific IgM levels in the vaccinated groups were up-regulated to the peak and higher than that of the control group, especially these of the two OmpA subunit vaccine groups were more significant ( Figure 3 ), which indicated that vaccination increased the production of specific antibodies and enhanced host resistance to A. hydrophila infection.

The 16S rRNA transcripts levels of A. hydrophila in different tissues were detected to compare the bacterial abundance in different groups at various time points post infection. In the intestines, 16S rRNA transcripts of most vaccinated groups were lower than that of the control group, especially these of the OmpA vaccine groups at all time points. The OmpA vaccine groups also showed lower levels of 16S rRNA transcripts in the hepatopancreas and gills. However, the up-regulated levels of 16S rRNA transcripts were observed in the gills at 3 and 7 dpi of most groups, which should be related to the method of intraperitoneal injection. Overall, these results revealed that vaccines immunization (especially OmpA vaccines) decreased the A. hydrophila abundance in the host, thereby protecting juvenile M. amblycephala against infection ( Figure 4 ).

The specificity of all prepared antibodies was validated by western blotting which showed excellent specificity ( Supplemental Figure 2 ). Four immune-related proteins (CD8, IgM, IgZ and IgD) showed strong positive immunoreactive signals in both tissues ( Table 3 ). In general, the number of immune positive cells was much more abundant in the hepatopancreas than in the intestines (except CD8⁺ cells), which also maintained abundant levels post infection, while that in the intestines just showed partial up-regulation post infection ( Figure 5 ).

In the intestines, the expression of adaptive immune-related genes (C3, TCR α, MHC I, MHC II, CD4, CD8, IgM, IgZ and IgD) was higher in immunized groups after 2 times vaccination, and most of them were also significantly up-regulated in the vaccinated groups post infection, except for CD4 and TCR α ( Figure 6A ; Supplemental Figure 6A ). In addition, proinflammatory cytokines (TNFα, iNOS, IL-1β, IL-6 and SAA) showed similar expression patterns to adaptive immune-related genes, especially the significant up-regulation of acute phase protein SAA, while the expression of anti-inflammatory cytokine IL-10 was mainly decreased post infection. These results indicated that vaccine immunization enhanced host innate and adaptive immunity in the intestines, especially the high dose vaccine groups.

In the gills, the expression of most adaptive immune-related genes was also higher in immunized groups after vaccination and post infection, except for MHC I ( Figure 6B ; Supplemental Figure 6B ). The expression of most proinflammatory cytokines was relatively low after vaccination, but up-regulated significantly at 3 and 7 dpi (SAA up-regulated significantly at 1 dpi). These results also revealed that vaccine immunization strengthened host innate and adaptive immunity in the gills.

In the hepatopancreas, similar expression patterns were found in adaptive and innate immune-related genes with low expression levels before 1 dpi, which started up-regulated from 3 dpi to 30 dpi (except for IgZ and MHC I). In addition, the vaccinated groups showed higher expression levels than those of the control group at most time points, indicating that vaccine immunization also enhanced host immunity ( Figure 6C ; Supplemental Figure 6C ).

The activities of LZM, ACP and AKP in the hepatopancreas and serum of juvenile M. amblycephala post infection were detected to evaluate effects of vaccination on host antibacterial ability. In the serum, the LZM activities of most groups reached the peak levels at 1 dpi, and immunized groups showed much higher activities compared with control group (P < 0.05). In the hepatopancreas, all vaccinated groups showed much higher LZM activities at 0, 1 and 7 dpi. Although LZM activity of the control group increased significantly at 3 and 14 dpi, LZM activity of the OmpA vaccine groups was dramatically higher (P < 0.05). In general, the LZM activities of all groups were up-regulated post infection in hepatopancreas and serum ( Figures 7A, D ), especially vaccinated groups, indicating that vaccine immunization enhanced the host antibacterial ability.

The activities of ACP showed different levels and trends in the hepatopancreas and serum, which maintained high and stable levels in the serum upon infection, and the control and vaccinated groups showed no significant difference with just a partial increase at 7 dpi ( Figures 7B, E ). In contrast, the hepatic ACP activities of inactivated vaccine H and OmpA vaccine H groups were higher than that of the control group at 0 dpi, which gradually decreased post infection. Thus, these results indicated that the bactericidal effect of serum was more excellent than that of hepatopancreas, and vaccination significantly enhanced the bactericidal effect of serum post infection, but not that of hepatopancreas.

The activities of AKP were fluctuant within limits in both serum and hepatopancreas, except for the extremely significant up-regulation of vaccinated groups in the hepatopancreas at 0 dpi, but the regulation of vaccination was also inconsistent at various time points post infection, indicating that the variable effects of vaccination on inflammatory response upon infection. In a word, vaccines immunization (especially OmpA vaccines) enhanced the antibacterial ability of juvenile M. amblycephala against A. hydrophila infection ( Figures 7C, F ).