Atlantic salmon, 40-50g, were kept at the Aquaculture Research Station (Tromsø, Norway) in flat-bottomed circular 200 L tanks at an ambient temperature of approximately 12°C with 12/12 h illumination and fed commercial pelleted diets. Prior to treatment, the fish were anesthetized in Metacaine (50 mg L^-1, Norsk Medisinaldepot) and sacrificed using 100 mg L^-1 Metacaine before collecting the different tissues. The production of anti-rSsGATA-3 antibodies in rats were approved by the Instructional Animal Care and Use Committee of the Ocean University of China. All possible endeavors were made to minimize suffering and maintain animal welfare.

For the construction of the pSsGATA-3-RFP plasmid, the open reading frames of SsGATA-3 (GenBank No: EU418015) were retrieved from GenBank and cloned as previously described (Kumari et al., 2009). From a spleen cDNA library, the gene was subcloned into a pTagRFP-N vector (Evrogen, Moscow, Russia) by PCR using primer SsGATA3F/SsGATA3R ( Table 1 ) to generate pSsGATA-3-RFP plasmids.

For the construction of pETSsGATA-3a and pETSsGATA-3b, which expresses His-tagged recombinant SsGATA-3a (rSsGATA-3a, amino acid residues 1-260) and SsGATA-3b (rSsGATA-3b, amino acid residues 261-441), the coding sequences of SsGATA-3a and SsGATA-3b were amplified by PCR with primers SsGATA3aF/SsGATA3aR and SsGATA3bF/SsGATA3bR ( Table 1 ). The PCR products were ligated with the TA cloning vector T-Simple (TransGen, Beijing, China). The recombinant plasmids were digested with EcoRV to retrieve fragments containing SsGATA-3a and SsGATA-3b, which were inserted into pET259 at the EcoRV site to obtain pETSsGATA-3a and pETSsGATA-3b.

For the construction of SsIL-4/13a reporter gene plasmids, the 5’-flanking region sequence of the SsIL-4/13a (GenBank No: NM_001204895.1) was obtained from Genome (GenBank NO: NC_059454.1) according to the BLASTN programs. The Genomic DNA was isolated from spleen with the TIANNamp Marine Animals DNA kit (TransGen, Beijing, China). PCR amplified approximately 600 bp of 5’-flanking regions by primer SsIL-4/13aproF/SsIL-4/13aproR ( Table 1 ). For sequencing, the PCR products were cloned in TA cloning vector T-Simple (TransGen, Beijing, China). The Identification of transcription factor binding motifs was predicted with TRANSFAC^® (Biobase International) (Heinemeyer et al., 1998) and MatInspector version 6.2 (Cartharius et al., 2005). Sequences with successive removal of the 5’-flanking regions were amplified by PCR using forward primers SsIL-4/13aF1, SsIL-4/13aF2, SsIL-4/13aF3, SsIL-4/13aF4 and reverse primer SsIL-4/13aR ( Table 1 ) to generate deletion constructs p(-577/+20) Luc, p(-317/+20) Luc, p(-302/+20) Luc and p(-264/+20) Luc. The PCR products were inserted into reporter vector pMetLuc-2 (Clontech, Mountain View, CA, USA). All plasmids were transformed and grown in One Shot TOPO 10 Escherichia coli (Invitrogen, Carlsbad, CA, USA) and then isolated using Endo-free Plasmid Mini Kit (QIAgen, Hilden, Germany), verified by restriction map analysis and DNA sequencing.

The rSsGATA-3a and rSsGATA-3b were expressed and purified as described previously (Wang and Sun, 2016). Briefly, Escherichia coli BL21 (DE3) cells (TransGen, Beijing, China) were transformed separately with pETSsGATA-3a and pETSsGATA-3b. Then, the transformants were grown to the mid-logarithmic phase in LB medium at 37 °C, and isopropyl-β-dithiogalactopyranoside was added to the cultures to a final concentration of 1 mM. After growing at 30°C for an additional 6 hours, the cells were harvested by centrifugation (3000 g). His-tagged proteins were purified using Ni-NTA Agarose (QIAgen, Hilden, Germany) as recommended by the manufacturer. The proteins were concentrated with PEG20000 (Solarbio, Beijing, China), and endotoxins were removed as reported previously (Zhang and Sun, 2019). The concentrated proteins were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), and their concentrations were determined using the Bradford method with bovine serum albumin as a standard (Bradford, 1976).

The rat anti-rSsGATA-3a and rat anti-rSsGATA-3b serum were prepared following a protocol described previously (Liu et al., 2010). The antibodies were purified using Protein G-based Chromatography Media (Sigma, St. Louis, MO, USA) in line with the introduction of manufacture. The concentrations of purified antibodies were determined using the Bradford method with bovine serum albumin as a standard and adjusted to 1 mg ml^-1 (Bradford, 1976). The specificity of antibodies was detected by Western blotting and immunofluorescence.

Tissue and recombinant proteins were electro-transformed (30 V for 1.5 h) from a 12.5% SDS-PAGE onto a polyvinylidene fluoride (PVDF) membrane. Then the membrane was blocked with 5% skimmed milk for 1 hour at 37°C and incubated with rat anti-rSsGATA-3a (1:1000) or rat anti-rSsGATA-3b (1:1000) overnight at 4°C, respectively. After being washed thrice with PBST, the membrane was incubated with goat-anti-rat Ig-HRP (1:5000) (Santa Cruz Biotechnology, Santa Cruz, CA, USA) for 1 hour at 37°C. After washing three times with PBST, the immunocomplex was detected using Pierce™ ECL Western Blotting Substrate (Pierce, Rockford, IL, USA).

Chinook salmon embryonic (CHSE-214) cells were seeded in the standard flask (Nunc™) with L-15 medium (Invitrogen, Carlsbad, CA, USA) containing penicillin (60 μg ml⁻¹), streptomycin (100 μg ml⁻¹), 1% non-essential amino acids and 8% fetal bovine serum (FBS). The CHSE-214 cells were incubated at 20°C for one week. Cells were washed twice with 10 ml PBS before adding 1.5 ml trypsin (Sigma, St. Louis, MO, USA). Loosened cells were re-suspended in L-15 medium (8% FBS, 1% NEAA, without antibiotics), counted in Nucleocounter YC-100 (Invitrogen, Carlsbad, CA, USA), and seeded in separate wells (1 x 10⁵ cells well^-1). The pSsGATA-3-RFP and pTagRFP-N were transfected to cells with Lipofectamine LTX (Invitrogen, Carlsbad, CA, USA) according to supplier protocol. Mock-transfected cells were considered as control. After 48 h, the cells were fixed with 4% formaldehyde (w/v) (Invitrogen, Carlsbad, CA, USA) for 30 minutes. After washing thrice, the cells were incubated with 200 μl 5% skimmed milk for 2 h at room temperature. The cells were subsequently incubated with 200 μl rat anti-rSsGATA-3a (1:200), rat anti-rSsGATA-3b (1:200) or rat anti-Trx (1:200) as control for 1 h at 37°C with occasional shaking, then washed thrice with PBS and incubated for another 1 h at 37°C with goat-anti-rat Ig-FITC (1:500) (Abcam, Boston, MA, USA). After washing thrice, DAPI (Sigma, St. Louis, MO, USA) was used for nucleic acid staining. Micrographs were obtained by an inverted fluorescence microscope equipped with DAPI-365 and Texas Red 530-585 filters (Carl Zeiss Imager A2, Jena, Germany).

For reporter activity assay, the CHSE-214 cells were re-suspended in L-15 medium and seeded in 24 well culture plate (2 x 10⁵ cells well^-1). The reporter vectors, pSsGATA-3 and pTagRFP-N were transfected to the cells with Lipofectamine LTX according to the manufacturer’s instructions. 48 hours after transfection, the culture mediums of transfected cells were analyzed for luciferase activity and SEAP activity using the luciferase assay kit and great scape ™ SEAP chemiluminescent detection kit (Clontech, Mountain View, CA, USA), respectively.

Four-µm-thick formalin-fixed, paraffin-embedded (FFPE) tissue sections were mounted on StarFrost Advanced Adhesive slides (Engelbrecht, Kassel, Germany), followed by drying at 60°C for 2 h. After deparaffinization, slides were pretreated at 120°C in Citric acid buffer (pH=6.0) for 10 min to retrieve antigens. Sections were incubated for 20 min at room temperature in a humid chamber with 1% H₂O₂ and 100% methanol to quench endogenous peroxidase. After washing thrice, the sections were incubated with 5% skimmed milk for 2 h at room temperature. Then the sections were incubated with rat anti-rSsGATA-3b (1:200) or rat anti-Trx (1:200) overnight at 4°C, then washed thrice with PBS, and incubated for another 1 h at 37°C with goat-anti-rat Ig-HRP (1:500) (Abcam, Eugene, USA). After washing thrice, the tissue sections were immersed in a freshly prepared 3, 3’-diaminobenzidine (DAB) substrate reagent solution (Invitrogen, Carlsbad, CA, USA) for 10 minutes. Micrographs were obtained by microscope (Carl Zeiss Imager A2, Jena, Germany).

Statistical analysis was performed using one-way ANOVA and LSD multiple comparisons in SPSS 20.0 software (SPSS Inc., Chicago, IL, USA). Data are expressed as mean ± SD, and statistical significance was defined as P < 0.01.

The rSsGATA-3a and rSsGATA-3b proteins were expressed in and purified from transfected E. coli as His-tagged proteins, respectively ( Figure 1A ), and the anti-rSsGATA-3a and anti-rSsGATA-3b antibodies were prepared from rat. Western blotting showed that the rat anti-rSsGATA-3b antibodies could specifically recognize rSsGATA-3b ( Figure 1B ). To further examine the specificity of the rat anti-rSsGATA-3b antibodies, CHSE-214 cells were transfected with pSsGATA-3-RFP or pTagRFP-N. Fluorescence microscopy results revealed that SsGATA-3-RFP (red) in cells with pSsGATA-3-RFP transfection were identified in or close to the nuclei (blue) and could be specifically detected by rat anti-rSsGATA-3b antibodies (green) ( Figure 1C ). In contrast, red fluorescence protein in the pTagRFP-N transfection group was uniformly expressed in cells, but no green color was observed ( Figure 1C ). However, the anti-SsGATA-3a antibodies did not seem to bind GATA-3 (result not shown).

Immunohistochemistry and Western blotting were conducted to detect tissue distribution of SsGATA-3 in healthy Atlantic salmon tissues. The gray color corresponding to the immunocomplex was observed in the gills, gill associated lymphoid tissue (GIALT), intestine, spleen, liver, and head kidney ( Figures 2A-F ). In gills and GIALT, the SsGATA-3 proteins were expressed in pillar cells, lymphocyte-like cells, epithelial cells, chondrocytes, perichondrial cells, and some undifferentiated basal cells. In the spleen, intestine, liver, and head kidney, the SsGATA-3 positive cells were mainly found in lymphocyte-like cells. The results of Western blot as shown in Figure 2G , the SsGATA-3 protein was found in most of the analyzed tissues. The highest contents of SsGATA-3 proteins were revealed in the gills, followed by spleen, intestine, liver, head kidney, and liver in healthy salmon. The level of SsGATA-3 protein in muscle was low.

In 577 bp 5’-flanking regions (5’-FRs) of SsIL-4/13a, which exhibit four GATA-3 binding sites located in -445/-339, -310/-303, -268/-261, and -2/+7 ( Figure 3A ). In this study, we examined the potential effect of SsGATA-3 on the activity of the SsIL-4/13a promoter. For this purpose, four promoter reporter plasmids contain 577bp, 317bp, 302 bp or 264 bp 5’-FRs of SsIL-4/13a were created, in which the promoter activities were reflected by the activities of the luciferase reporter. CHSE-214 cells were transfected with pSsGATA-3 plus four pLucSsIL-4/13a reporter vectors respectively, and the luciferase activities were determined. The results showed that luciferase activities were significantly increased in the presence of pSsGATA-3 compared to the cells with pTagRFP-N p (-577/+20) luc and p (-317/+20) luc transfectants ( Figure 3B ). However, in p (-302/+20) luc and p (-264/+20) luc transfectants, luciferase activities were not significantly different between the groups in the presence of pSsGATA-3 and pTagRFP-N ( Figure 3B ).