All animal experiments were conducted in accordance with the Guiding Principles for the Care and Use of Laboratory Animals and were approved by the Institute of Hydrobiology, Chinese Academy of Sciences (Approval ID: IHB 2013724).

The wild-type AB/TU zebrafish were maintained in recirculating aquaculture system at 28.5°C under standard conditions according to zebrafish care and maintenance protocol. Epithelioma papulosum cyprini (EPC) cells were maintained at 28°C in 5% CO₂ in medium 199 (Hyclone) supplemented with 10% FBS, 100 U/ml penicillin and 100 μg/ml streptomycin. The wild-type Edwardsiella piscicida (PPD130/91 strain) were grown in tryptic soy broth (TSB, BD Biosciences) (19).

Based on zebrafish mRNA sequence (GenBank accession No: NM_194395), primers FLAG-BIRC2F/FLAG-BIRC2R were used for cloning the open reading frame (ORF) of zebrafish BIRC2, and inserted into the p3×FLAG-CMV ™-14 Expression Vector (Sigma-Aldrich). The primer sequences used for plasmid construction were listed in Table 1 .

Domain architecture analysis was performed using Batch CD-Search (NCBI), and search results were visualized through the TBtools (20). Construction of phylogenetic tree was performed using the Neighbor-Joining method with 10,000 bootstrap replications using MEGA X software. Accession numbers of BIRC2 proteins from different species used are as follows: zebrafish (Danio rerio) BIRC2, XP_005161213.1; common carp (Cyprinus carpio) BIRC2, XP_018950329; channel catfish (Ictalurus punctatus) BIRC2, NP_001187106.1; rainbow trout (Oncorhynchus mykiss) BIRC2, XP_036845322.1; Japanese medaka (Oryzias latipes) BIRC2, XP_023818594.1; Atlantic cod (Gadus morhua) BIRC2, XP_030216815.1; spotted gar (Lepisosteus oculatus) BIRC2, XP_006627852; human (Homo sapiens) BIRC2, NP_001157.1; house mouse (Mus musculus) BIRC2, NP_031491.2; chicken (Gallus gallus) BIRC2, NP_001007823.1; green anole (Anolis carolinensis) BIRC2, XP_016848156.1; and common frog (Rana temporaria) BIRC2, XP_040196143.1.

Zebrafish embryos and larvae including 6 h post-fertilization (hpf), 12 hpf, 1 days post-fertilization (dpf), 2 dpf, 3 dpf, 4 dpf, 5 dpf and 7 dpf were collected, and used for quantivative RT-PCR (qRT-PCR) to determine the constitutive expression of zebrafish BIRC2. To analyze the expression level of zebrafish BIRC2 in response to bacterial infection, zebrafish larvae at 4 dpf were infected with E. piscicida (2 × 10⁸ CFU/ml) at 28°C. Zebrafish larvae were collected at 6, 24 and 72 h post-infection (hpi), and used for qRT-PCR. The primer sequences used for qRT-PCR were listed in Table 1 .

To determine the effect of zebrafish BIRC2 on apoptosis, 1 × 10⁶ EPC cells were plated into 6-well plates overnight and then transiently transfected with 2 μg BIRC2-FLAG or p3×FLAG. After 24 h later, the cells were infected with E. piscicida at an MOI of 5, 10 or left untreated. At 6 hpi, these cells were washed with ice-cold PBS, and proceeded with the FITC Annexin V Staining using the FITC Annexin V Apoptosis Detection Kit (Vazyme, China) on a CytoFLEX LX Flow Cytometer (Beckman, USA) according to the manufacturer’s instructions. The data were analyzed using the software CytoExpert.

To determine the role of zebrafish BIRC2 in bacterial infection, 200 ng/μl p3×FLAG empty plasmid or BIRC2-FLAG was microinjected into zebrafish embryos at the one-cell stage. At 4 dpf, the hatched larvae were infected with 2 × 10⁸ CFU/ml E. piscicida. To determine whether the negative regulation of zebrafish BIRC2 on E. piscicida infection was related to the suppression of caspases and the accumulation of p53, the caspase activator PAC-1 (Selleck, #S2738) and p53 inactivators pifithrin-μ (Selleck, #S2930) and pifithrin-α (Selleck, #S2929) were used for treating zebrafish larvae microinjected with p3×FLAG or BIRC2-FLAG at 4 dpf. The final concentration of treatment is 10 μm for PAC-1, 50 μm for Pifithrin-μ, and 5 μm for Pifithrin-α. After 12 h treatment, these larvae were infected with 2 × 10⁸ CFU/ml E. piscicida.

At 24 and/or 48 hpi, 10 larvae per group were washed in PBS three times and then lysed in 1 ml of PBS via homogenization for plating serial dilutions on TSB agar. CFU of E. piscicida were enumerated after 24 h of incubation at 28°C. For larval survival analysis, exposures were performed with three repetitions for each group, and the numbers of each repetition were 20 larvae. The numbers of surviving larvae were counted daily for 4 or 5 d. GraphPad Prism 7 was used to generate survival curves. To determine the effect of zebrafish BIRC2 on the expression levels of apoptosis-related genes and NF-κB genes, 30 larvae per group were collected at 24 and 48 hpi, and used for qRT-PCR. To determine the effect of PAC-1 on the expression levels of apoptosis-related genes and NF-κB genes, 30 larvae per group were collected at 48 hpi, and used for qRT-PCR.

For qRT-PCR analysis, total RNA was extracted from those samples collected above by Trizol reagent (Invitrogen). RNase-free DNase I (Thermo) was used to remove genomic DNA contamination at 37°C for 30 min. The first-strand cDNA was generated from total RNA using RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific). qRT-PCR was performed on a BIO-RAD CFX96 Real-Time System using SYBR^® Green Master Mix (Bio-RAD) according to the procedure as follows: preincubation at 95°C for 3 min, then 45 cycles at 95°C for 10 s, 58°C for 20 s, and 72°C for 30 s. Each sample was tested in triplicate. The efficiency of primer pairs was calculated by making a dilution series. The Ct values were correlated with the DNA input over seven orders of magnitude (R² >0.999). The amplification efficiency of BIRC2, casp3a, casp3b, casp8, casp9, NF-κB1, NF-κB2, p53, and GAPDH primer pairs is 90.0, 93.5, 93.4, 94.4, 92.7, 93.7, 92.9, 94.3, and 92.5%, respectively. The housekeeping gene GAPDH was used for normalizing the Ct values of target genes. The fold changes relative to control group were calculated using the 2^−ΔΔCt method. All primers used for qRT-PCR were presented in Table 1 .

Seeded on coverslips were 2 × 10⁵ EPC cells inside 24-well plates that were co-transfected with FLAG plus GFP, FLAG plus p53-GFP, BIRC2-FLAG plus p53-GFP, and then infected with E. piscicida at an MOI of 1. At 6 hpi, the cells were washed with PBS for three times, fixed with 4% formaldehyde at 37°C for 15 min, permeabilized with PBS containing 0.1% Triton^® X-100 for 10 min, incubated with anti-FLAG antibody (1:5,000, Sigma-Aldrich, # F3165) overnight, and incubated with ReadyProbes™ Alexa Fluor^® 594 Goat Anti-Mouse IgG Antibody (Invitrogen, #R37121) for 30 min. After washing with PBST, the cells were stained with NucBlue Fixed (Invitrogen, #R37606) and observed on a confocal microscope (SP8; Lecia, Wetzlar, Germany).

To investigate the regulating effect of zebrafish BIRC2 on the p53 protein with or without bacterial infection, 1 × 10⁶ EPC cells were plated into 6-well plates, and transfected with indicated plasmids and concentration. After 48 h, these cells were infected with E. piscicida at an MOI of 1 or left untreated. At 6 hpi, these cells were used for protein extraction using RIPA buffer (Thermo Scientific™) containing Protease Inhibitor Cocktail. All samples were analyzed by Western blotting with anti-GAPDH (Proteintech, #10494-1-AP), anti-FLAG (Sigma-Aldrich, #F3165), and anti-pTurboGFP (Evrogen, #AB513) antibodies.

To investigate the exact effect of zebrafish BIRC2 on the protein regulation of cytoplasmic and nuclear p53 in the case of E. piscicida infection, 1 × 10⁶ EPC cells plated into 6-well plates were transfected with indicated plasmids and concentration. At 6 hpi, the cells were collected and used for subsequent nuclear and cytoplasmic protein extraction using Subcellular Protein Fractionation Kit (Thermo Scientific™, #78840) according to the manufacturer’s instructions. Extracts were analyzed by Western blotting with anti-HDAC1 (ab41407; Abcam), anti-tubulin (ab6046; Abcam), anti-FLAG (F3165; Sigma-Aldrich) and anti-pTurboGFP (AB513; Evrogen) antibodies.

To determine the interaction between zebrafish BIRC2 and p53, EPC cells were transfected with the indicated plasmids mixes (1,000 ng for each plasmid). After 48 h, the cells were collected and lysed in IP lysis buffer containing Protease Inhibitor. The Anti-FLAG M2-Agarose Affinity Gel (Sigma-Aldrich) was added to the cell lysates, incubated at 4 ˚C overnight with gentle mixing. The eluted proteins and input proteins were analyzed by Western blotting with anti-FLAG and anti-pTurboGFP antibodies.

All data were expressed as means and standard errors of means (SEM) of three independent experiments. A two-tailed Student’s T-test or a one-way ANOVA was used to determined significant differences for qRT-PCR (*p <0.05; ** p <0.01). The log-rank test was used to test significant differences for survival analysis.

The ORF length of zebrafish BIRC2 is 1,887 bp (GenBank accession No. NM_194395), encoding 628 amino acids. Sequence analysis by using NCBI CDD reveals that vertebrate BIRC2 proteins from zebrafish, common carp, channel catfish, rainbow trout, Japanese medaka, Atlantic cod, spotted gar, frog, green anole, chicken, mouse, and human have the same domain architecture, and all proteins contain three BIR domains, one UBA domain, one CARD domain, and one C-term RING domain ( Figure 1A ). A phylogenetic tree is constructed using the N-J method based on the amino acid sequences from these vertebrate BIRC2 proteins. The results show that zebrafish BIRC2 is grouped with common carp and channel catfish BIRC2 proteins, with bootstrap value 100% ( Figure 1B ).

The expression of zebrafish BIRC2 was examined in eight samples involved in early developmental stages and in zebrafish larvae with or without E. piscicida infection. The results from qRT-PCR showed that the transcript level of zebrafish BIRC2 sharply increased from 6 to 12 hpf, then decreased from 12 hpf to 2 dpf, next maintained a relatively stable level from 2 to 5 dpf, and finally increased at 7 dpf ( Figure 2A ). Compared with the control group without E. piscicida infection, the expression of BIRC2 in zebrafish larvae infected with E. piscicida had no significant change at 6 and 24 hpi during the early stage of infection, but significantly increased at 72 hpi during the late stage of infection ( Figure 2B ).

To test whether zebrafish BIRC2 has anti-apoptosis function, EPC cells transfected with p3×FLAG empty plasmid or BIRC2-FLAG were infected with E. piscicida, and next used for measuring the apoptosis rate by flow cytometry. At the early apoptosis stage, the apoptosis rates for cells transfected with BIRC2-FLAG had no significant change in the absence of infection, and were significantly lower than that transfected with p3×FLAG empty plasmid when the MOIs of E. piscicida were 5 and 10 ( Figure 3A ). At the late apoptosis stage, the apoptosis rates for cells transfected with BIRC2-FLAG had no significant change in the absence of infection and at an MOI of 5, but significantly lower than that transfected with p3×FLAG empty plasmid at an MOI of 10 ( Figure 3B ).

To determine the effect of zebrafish BIRC2 in bacterial infection, zebrafish larvae overexpressed with BIRC2-FLAG or the empty plasmid were infected with E. piscicida. Compared with the control group transfected with the empty plasmid, bacterial load in the zebrafish larvae overexpressed with BIRC2-FLAG were higher both at 24 and 48 hpi ( Figure 4A ). The survival rate of zebrafish larvae overexpressed with BIRC2-FLAG was reduced about 23.33% ( Figure 4B ).

To determine the possible correlation between zebrafish BIRC2 and caspases or p53, we infected zebrafish larvae overexpressed with BIRC2-FLAG or the FLAG empty plasmid with E. piscicida or left untreated, and examined the expression levels of casp3a, casp3b, casp8, casp9, and p53 along with the expression levels of NF-κB1 and NF-κB2. Compared with the control larvae microinjected with the empty plasmid, the overexpression of zebrafish BIRC2-FLAG significantly decreased the expression of casp8, and increased the expressions of p53, NF-κB1, and NF-κB2 at 24 hpi, especially for p53 with the 47.35-fold induction ( Figure 4C ). At 48 hpi, overexpression of zebrafish BIRC2-FLAG significantly decreased the expressions of casp3b, casp8, and casp9, increased the expressions of p53 and NF-κB1, and no significant changes for casp3a and NF-κB2 ( Figure 4D ).

Since overexpression of zebrafish BIRC2 inhibited the expression levels of caspases in the case of E. piscicida infection, we further investigated whether the inhibition of zebrafish BIRC2 on caspases were required for the negative regulation of zebrafish BIRC2 on the E. piscicida infection. The facts that overexpression of zebrafish BIRC2 promoted the proliferation of E. piscicida in zebrafish larvae were again confirmed ( Figures 5A, B ). However, treatment of zebrafish larvae microinjected with BIRC2 with the caspase activator PAC-1 significantly inhibited the zebrafish BIRC2-mediated bacterial proliferation at 24 hpi ( Figure 5A ), and completely blocked the zebrafish BIRC2-mediated bacterial proliferation at 48 hpi ( Figure 5B ). Compared with zebrafish larvae microinjected with the FLAG empty plasmid, the bacterial loads in zebrafish larvae microinjected with BIRC2 even were significantly decreased by the PAC-1 treatment ( Figure 5B ). Most of all, PAC-1 treatment tremendously improved larvae survival. All the larvae overexpressed with zebrafish BIRC2 without the PAC-1 treatment died at 4 dpi, however the survival rate of zebrafish larvae overexpressed with BIRC2-FLAG but with the PAC-1 treatment was 61.67% at 5 dpi. Compared with the control group microinjected with the FLAG empty plasmid in the absence of PAC-1 treatment, no significant difference was observed for the survival rate of zebrafish larvae overexpressed with BIRC2-FLAG and with the PAC-1 treatment ( Figure 5C ).

Next we examine whether the caspase activator PAC-1 treatment rescued the zebrafish BIRC2-mediated inhibition on caspases. In the zebrafish larvae microinjected with the FLAG empty plasmid, PAC-1 treatment induced the expressions of BIRC2, casp8, casp3a, casp3b, casp9, and NF-κB2, decreased the expression of NF-κB1, and no significant regulation on p53 ( Figures 5D–K ). In the zebrafish larvae microinjected with BIRC2-FLAG, PAC-1 treatment induced the expressions of casp8, casp3a, casp3b, casp9, and NF-κB2, and no significant regulation on the expressions of BIRC2, p53, and NF-κB1 ( Figures 5D–K ). Compared with the group microinjected with the FLAG empty plasmid but without the PAC-1 treatment, PAC-1 treatment had no significant regulation on casp8 and increased the expression of casp9 in zebrafish larvae microinjected with BIRC2-FLAG ( Figures 5E, H ).

Overexpression of zebrafish BIRC2 significantly increased the expressions of p53, especially at 24 hpi. We further investigated the effect of zebrafish BIRC2 on the protein levels of p53. In the absence of infection, the induced expressions of p53 by zebrafish BIRC2 were dependent of the dose variation ( Figure 6A ). Similar results were observed in the case of E. piscicida infection ( Figure 6B ). In mammals, the actions of p53 as well as its localization are not restricted to the nucleus. p53 has been found to induce apoptosis in the cytoplasm through the intrinsic mitochondria-mediated pathway (21). To examine the location of zebrafish BIRC2-mediated p53 accumulation, cytoplasmic and nuclear fractions were prepared from E. piscicida-infected EPC cells transfected with various indicated plasmids. p53-GFP protein was detected both in the cytoplasm and the nucleus, however BIRC2-FLAG was detected only in the cytoplasm. Furthermore, overexpression of zebrafish BIRC2 was found to accumulate p53 both in the cytoplasm and the nucleus ( Figures 6C, D ).

To further investigate the interactions between zebrafish BIRC2 and p53, we next conducted subcellular localization studies using transient-expression vectors to express p53-GFP and BIRC2-FLAG. Expression of the fusion proteins from these plasmids in transfected EPC cells has been confirmed by Western blotting ( Figure 6 ). In control experiments, pTurboGFP fluorescence was observed over the cytoplasm and the nucleus of infected cells. p53-GFP only had a weak signal observed in the cytoplasm and the nucleus of infected cells in the absence of zebrafish BIRC2, but the obviously enhanced fluorescence signal for p53-GFP was observed in the cytoplasm of infected cells in the case of zebrafish BIRC2. Fluorescence associated with BIRC2-FLAG (Red) was only detected in the cytoplasm of infected cells. Furthermore, the merged image confirmed that zebrafish BIRC2 colocalized with p53 in the cytoplasm of infected cells, which was indicated by yellow fluorescence ( Figure 7A ).

Co-IP assays were next performed to further confirm the interaction between zebrafish BIRC2 and p53. Anti-FLAG-conjugated agarose beads were used to precipitate FLAG-containing protein complex from EPC cells. No band of p53-GFP was observed in Lane 2 (using anti-GFP antibody for IP product), which suggested that FLAG protein failed to interact with p53. However, immunoblotting analysis of anti-FLAG immunoprecipitate with an anti-GFP antibody (Lane 3) showed that the association of zebrafish BIRC2 with p53 was readily detected ( Figure 7B ).

Since zebrafish BIRC2 was found to accumulate p53 both in the cytoplasm and the nucleus, we would like to investigate whether the induction of the cytoplasmic and nuclear p53 by zebrafish BIRC2 were required for the negative regulation of zebrafish BIRC2 on the E. piscicida infection. Pifithrin-μ is a cytoplasmic p53 inhibitor, which can inhibits p53 binding to mitochondria (22), however pifithrin-α is widely used as a specific inhibitor of p53 transcription activity by acting at a stage after p53 translocation to the nucleus (23, 24). Here, the p53 inactivators pifithrin-μ and pifithrin-α were used to inhibit the cytoplasmic and nuclear p53, respectively. Although treatment of pifithrin-μ inhibited the proliferation of E. piscicida in zebrafish larvae ( Figure 8A ), no significant differences were observed for the survival rates of zebrafish larvae overexpressed with the FLAG empty plasmid between these two groups without or with the pifithrin-μ treatment ( Figure 8B ). However, pifithrin-μ treatment significantly inhibited the proliferation of E. piscicida in zebrafish larvae, and completely blocked the zebrafish BIRC2-mediated bacterial proliferation at 24 hpi, when compared with zebrafish larvae overexpressed with the FLAG empty plasmid or BIRC2-FLAG in the absence of pifithrin-μ treatment ( Figure 8A ). Furthermore, pifithrin-μ treatment completely rescued larval survival for zebrafish microinjected with BIRC2 in the case of E. piscicida infection ( Figure 8B ). Similar results were observed for pifithrin-α treatment ( Figures 8C, D ).