Epithelioma papulosum cyprini (EPC) cells for western blotting and subcellular localization, Human embryonic kidney (HEK) 293T cells for coimmunoprecipitation (Co-IP) and dephosphorylation assays were cultured as described previously (54). Gibel carp brain (GiCB) cells for quantitative real-time PCR (qPCR) and viral infection were kindly provided by Prof. Zeng (Yangtze River Fisheries Research Institute, Chinese Academy of Fishery Sciences) (55). SVCV, a negative sense single-stranded RNA virus in the family Rhabdoviridae that could infect crucian carp and gibel carp (56) was propagated in GiCB cells until cytopathic effects (CPE) were observed, and then the culture media with cells were harvested and stored at -80°C until needed.

According to the genome sequences of gibel carp clone F, Cgptpn6-A and Cgptpn6-B cDNAs were amplified from gibel carp head kidney cDNA library by Rapid Amplification of cDNA Ends Polymerase Chain Reaction (RACE-PCR). PCR products amplified by a high-fidelity polymerase (TransGen Biotech) were purified and cloned into Trans5α Chemically Competent Cells. About 30 clones of each sample were sequenced and classified according to the specific SNPs among the sequences. The complete cDNA sequences of six Cgptpn6 transcripts were deposited in GenBank (accession numbers from OK142786-OK142791). Amino acid sequences and domains were predicted by open reading frame (ORF) Finder (https://www.ncbi.nlm.nih.gov/orffinder/) and SMART (http://smart.embl-heidelberg.de/), multiple amino acid sequence alignment was performed by DNAman version 7.0 software. Phylogenetic tree was constructed by bootstrap analysis (1000 replicates) using the neighbor-joining method (NJ) in MEGA 7.0 software (57).

All the amino acid sequences used in this study were obtained from GenBank (http://www.ncbi.nlm.nih.gov/) and Ensembl (http://www.ensembl.org). The accession numbers are as following: Homo sapiens SHP1, NP_002822.2; Mus musculus SHP1, NP_038573.2; Gallus gallus SHP1, NP_001026655.1; Lepisosteus oculatus SHP1, ENSLOCT00000009309.1; Danio rerio SHP1, NP_956254.1; Carassius auratus SHP1-A, XP_026109501.1; Carassius auratus SHP1-B, XP_026139710.1. The exon-intron structure was determined by aligning cDNA and genomic sequences. Syntenic analyses were conducted by comparing the chromosomic regions around ptpn6 genes in gibel carp chromosomes (CgA16 and CgB16) and crucian carp chromosomes (CaA16 and CaB16) with corresponding regions in H. sapiens chromosome 12, M. musculus chromosome 6, G. gallus chromosome 1, L. oculatus chromosome LG26, D. rerio chromosome 16. The genome information was obtained from the Ensembl genome database.

Chromosome preparation was performed as described previously (58). Five individuals of gibel carp clone F were injected phytohemagglutinin (PHA) (15-20 μg/g) in vivo and the head kidney cells were harvested by conventional hypotonic and fixation treatments. Briefly, the cells were exposed to a hypotonic solution for 30 min at room temperature and fixed for 30 min (with replacement of the fixative every 10 min without resuspension) in 3 ml of a 3:1 mixture of methanol and acetic acid. Finally, the cells were resuspended in 0.5 ml of fresh fixative and were spread on clean slides. The slides were prepared by the air-drying technique and storied at -20°C for FISH.

The bacterial artificial chromosome (BAC) clones containing Cgptpn6-A and Cgptpn6-B were screened by PCR. Then, Cgptpn6-A-BAC-DNA and Cgptpn6-B-BAC-DNA labeled by DIG-Nick Translation Mix and Biotin-Nick Translation Mix (Roche) respectively were used to perform FISH as described previously (35, 47). 4’, 6-diamidino-2-phenylindole (DAPI) was used to counterstain metaphase chromosomes. The results were acquired by Carl Zeiss upright fluorescence microscope Axio imager M2 (Analytical & Testing Center, IHB, CAS).

Total RNAs from 12 adult tissues, including brain, kidney, intestine, skin, gill, heart, liver, muscle, spleen, thymus, ovary and head kidney, and GiCB cells were extracted by Trizol reagent (Invitrogen). RNase-free DNase was used to purify RNA by removing all contaminating genomic DNA. The first-strand cDNA was synthesized by using a GoScript Reverse Transcription System (Promega) according to the manufacturer’s instructions. qPCR was performed with Fast SYBR Green master mix (BioRad) on a CFX96 Real-Time System (BioRad). PCR conditions were as follows: 95°C for 5 min, then 40 cycles of 95°C for 20 s, 60°C for 20 s, 72°C for 20 s. Eukaryotic translation elongation factor 1 alpha 1, like 1 (eef1a1l1) (M value = 0.74 < 1.5) was selected as the optimal reference gene for qPCR analysis according to the previous study (52). The primers of other IFN-related genes were also synthetized for qPCR ( Supplementary Table 1 ). The specificity of the PCR amplification for all primer pairs was verified from the dissociation curves. The relative gene expression levels were calculated with 2^-△△CT method. All the samples were analyzed in triplicates.

For Coimmunoprecipitation assay (Co-IP) and Western blotting, the ORFs of CgSHP1-A and CgSHP1-B were cloned into pCMV-Myc, pCMV-HA (Clontech) and pcDNA3.1(+), respectively. For subcellular localization, the ORFs of CgSHP1-A and CgSHP1-B were inserted into pEGFP-N3 (Clontech) vector. Owing the extremely high amino acid sequence identities (98.46%-100.00%) of the IFN-related genes and autophagy-related genes between gibel carp and crucian carp (C. auratus) used in this study, we chose the corresponding plasmids from crucian carp constructed previously. The ORF of C. auratus mediator of IFN regulatory factor 3 (IRF3) activation (CaMITA) [also called stimulator of interferon genes (STING)] (Gene accession number: MZ172421) and kinase TANK-binding kinase 1 (CaTBK1) (Gene accession number: MZ172419) were inserted into pCS2-mCherry vector (Clontech Laboratories). The ORFs of mitochondrial antiviral signaling protein (CaMAVS) (Gene accession number: MZ170793), CaMITA, CaTBK1, CaIRF3 (Gene accession number: MZ172420), microtubule-associated Protein 1A/1B-Light Chain 3 (CaLC3) (Gene accession number: XM_026238864.1), CaBeclin1 (Gene accession number: XM_026249455.1), and autophagy-related gene 14 (CaATG14) (Gene accession number: XM_026286484.1) were cloned into pCMV-HA, pCMV-Myc and pCMV-Tag2c vector. Compared to the crucian carp genome, CaMAVS, CaMITA and CaIRF3 localize in A subgenome, while CaTBK1 and CaATG14 belong to B subgenome. The plasmids containing CaIFN-luc and ISRE-Luc in pGL3-Basic luciferase reporter vectors were constructed as described previously (36). The primers including the restriction enzyme cutting sites used for plasmid construction were also listed in Supplementary Table 1 . These primers were designed with Oligo Calc (Oligonucleotide Properties Calculator) (http://biotools.nubic.northwestern.edu/OligoCalc.html).

Transient transfections were performed in EPC and GiCB cells seeded in 6-well or 24-well plates by using FishTrans Transfection Reagent (MeiSenTe Biotechnology) according to the manufacturer’s protocol [Total plasmid dosage (μg) and FishTrans (μl) dosage is at the ratio of 1:2]. For subcellular localization, EPC cells were plated onto coverslips in 6-well plates and transfected with indicated plasmids for 24 h. Following this, the cells were washed twice with phosphate-buffered saline (PBS) and fixed with 4% paraformaldehyde (PFA) for 1 h. After draining the fixative, the cells were stained with DAPI (1 μg/ml; Beyotime) for 5 min in a dark at room temperature. Finally, the coverslips were washed and observed with a Leica confocal microscope under a × 63 oil immersion objective (SP8; Leica Microsystems). Fluorescence intensity was analyzed with Image J.

For the antiviral assay, GiCB cells were seeded to 24-well plates and were transfected with 0.5 μg CgSHP1-A and CgSHP1-B or pcDNA3.1(+) vector, separately. At 24 h post-transfection, the GiCB cells were infected with SVCV at a multiplicity of infection (MOI = 0.01) and incubated at 28°C. At 48 h post-infection, the cell monolayers were washed with PBS, fixed with 4% PFA for 1 h, and stained with 0.05% crystal violet overnight to observe the CPE. The supernatants were subjected to 10-fold serial dilutions and then added (100 μl) onto a monolayer of GiCB cells cultured in a 96-well plate. After 48 h, the medium was removed and the cells were washed with PBS, fixed by 4% PFA and stained with 1% crystal violet. The virus titer was expressed as 50% tissue culture infective dose (TCID₅₀/ml). Results are the representative of three independent experiments.

EPC cells were seeded in 24-well plates, and 24 h later co-transfected with 250 ng luciferase reporter plasmid (CaIFNpro-luc or ISRE-Luc), 250 ng CgSHP1-A, CgSHP1-B or pcDNA3.1(+), and 50 ng Renilla luciferase internal control vector (pRL-TK, Promega). Then, the cells were transfected again with a mimic of viral dsRNA poly I:C or the double-stranded DNA mimetic poly dA:dT at 24 h post-transfection, and the cells were infected by SVCV. To further explore the role of SHP1 in the RLR-induced interferon reaction, EPC cells were co-transfected with plasmids as described above, then transfected 250 ng CaMAVS-, CaMITA-, CaTBK1- or CaIRF3-Myc expressing plasmid or pCMV-Myc empty vector respectively. At 24 h post-transfection or infected, the cells were washed in PBS and lysed for measuring luciferase activity by Dual-Luciferase Reporter Assay System, according to the manufacturer’s instructions (Promega). Fireflyluciferase activities were normalized on the basis of Renilla luciferaseactivity. The results were the representative of more than three independent experiments, each performed in triplicate. Luciferase and qPCR assay data are expressed as the mean ± standard error of the mean (SEM). Error bars indicate the SEM (n = 3, biologically independent samples). Data were analyzed using a Student’s unpaired t-tests. A probability (p) < 0.05 was considered statistically significant (∗), and p < 0.01 was considered extremely significant (∗∗).

Transfected HEK 293T cells were lysed in radioimmuno-precipitation (RIPA) lysis buffer [1% NP-40, 50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM EDTA, 1 mM NaF, 1 mM sodium orthovanadate,1 mM phenyl-methylsulfonyl fluoride, and 0.25% sodium deoxycholate] without phosphatase inhibitors. Protein dephosphorylation was carried out in 100 μl reaction mixtures consisting of 100 μg of cell protein and 10 U of calf intestinal phosphatase (CIP) (Sigma-Aldrich) (59). The reaction mixtures were incubated at 37°C for 1 h, then separated by 10% SDS-PAGE and transferred to polyvinylidene difluoride (PVDF) membrane (Bio-Rad). The membranes were blocked and incubated with indicated primary antibodies (Abs) at an appropriate dilution overnight at 4°C, washed three times with TBST buffer [25 mM Tris-HCl, 150 mM NaCl, 0.1% Tween 20 (pH 7.5)] and then incubated with secondary Abs. After additional three washes with TBST, the membranes were stained with Immobilon TM Western Chemiluminescent HRP Substrate (Millipore) and detected using an Image Quant LAS4000 system (GE Healthcare). Abs were diluted as follows: anti-β-actin (Cell Signaling Technology) at 1:3,000, anti-Flag (Sigma-Aldrich) at 1:3,000, anti-Myc (Santa Cruz Biotechnology) at 1:3,000, HRP-conjugated anti-mouse IgG or anti-rabbit IgG (Thermo Scientific) at 1:5,000. The results were the representative of three independent experiments.

HEK 293T cells seeded in 10 cm² dishes overnight were transfected with a total of 10 µg of the plasmids. At 24 h post-transfection, medium was removed carefully and cell monolayer was washed twice with 10 ml ice-cold PBS. Then the cells were lysed in 1 ml of RIPA lysis buffer containing protease inhibitor mixture (Sigma-Aldrich) at 4°C for 1 h on a rocker platform. The cellular debris was removed by centrifugation at 12,000 × g for 15 min at 4°C. The 100 μl supernatant was transferred to a fresh tube and the rest was incubated with 30 μl of anti-Flag or anti-Myc affinity gel (Sigma-Aldrich) overnight at 4°C with constant agitation. Immunoprecipitated proteins were collected by centrifugation at 5,000 × g for 1 min at 4°C, washed three times with lysis buffer, and resuspended in 100 μl SDS sample buffer (59). The immunoprecipitates and whole cell lysates were analyzed by western blotting with the indicated Abs.

Six Cgptpn6 transcripts cloned from gibel carp head kidney were clearly clustered into two homeologs (Cgptpn6-A and Cgptpn6-B), and each of them clearly possessed three alleles ( Supplementary Figure 1 ). The average identities among the three alleles (99.87% ± 0.09% for Cgptpn6-A and 99.73% ± 0.05% for Cgptpn6-B) were higher than that between Cgptpn6-As and Cgptpn6-Bs (89.28% ± 0.07%). The major differences between the Cgptpn6-As and Cgptpn6-Bs homeologs were observed in the 3′ untranslated region (UTR), where the sequence identity was only 73.10%. The ORFs of the three Cgptpn6-A alleles were all 1761 bp, encoding two CgSHP1-A proteins [CgSHP1-A1 and CgSHP1-A2/A3, 586 amino acids (aa)] with one aa difference at the 29th. The ORFs of the three Cgptpn6-B genes were also 1761 bp, encoding the same CgSHP1-B protein (586 aa, CgSHP1-B1/B2/B3). All gibel carp SHP1 proteins possessed three conserved domains (two SH2-domains [N-SH2, C-SH2] and a PTPc domain). The PTP signature motif (I/VHCSAGIGRTG) is identical among mammals and fish SHP1 ( Supplementary Figure 2A ). Multiple sequence alignments and phylogenetic analysis showed that almost identical (99.08% and 100.00%) gibel carp and crucian carp SHP1-A and SHP1-B were grouped into two separate branches, implying the duplication of the ptpn6 gene in the common ancestor of gibel carp and crucian carp, which then clustered with zebrafish SHP1 and spotted gar SHP1 ( Supplementary Figure 2B ).

Subsequently, the genomic structure and syntenic alignment of gibel carp ptpn6-A and ptpn6-B and other vertebrates were identified. Except chicken Ptpn6 (15 exons), both Cgptpn6-A and Cgptpn6-B, as well as other vertebrate ptpn6, were composed of 16 exons ( Figure 1A ). Cgptpn6-A and Cgptpn6-B possess almost identical lengths of exons except the first and last exons, which are the loci for transcribing the 5′ and 3′ UTR. However, the lengths of their introns are varied and their identity is only 52.70%. The identities between the corresponding introns of Cgptpn6-A and Cgptpn6-B ranged from 19.82% (11th intron) to 90.74% (9th intron). Similarities in the genomic structure between crucian carp ptpn6-A and ptpn6-B were also observed. In addition, only the lengths of the 1st, 7th, and 16th exons of Cgptpn6-A and Cgptpn6-B are different from zebrafish ptpn6, implying a highly conserved genomic structure in Cyprinidae fish. Human PTPN6 and mouse Ptpn6 have similar exon lengths, which are different from chicken and fish ptpn6 genes.

Each Cgptpn6 homeolog with three alleles was confirmed by FISH. Consistent with our previous studies (35, 36), three green Cgptpn6-A signals and three red Cgptpn6-B signals were located on the three different chromosomes when simultaneously using Cgptpn6-A-BAC-DNA and Cgptpn6-B-BAC-DNA as probes respectively ( Figure 1B ). Syntenic alignment showed that gibel carp chromosome CgA16 and CgB16 both retained approximately 60% of the analyzed homologous genes in zebrafish chromosome 16 and had a conserved gene block (styk1-phc1-atn1-mlf2-ptpn6). One homeolog of the other duplicated genes in CgA16 and CgB16 seemed to be deleted, and became singletons. For example, p3h3, pex5, clstns, lpcat3, and nod1 were mapped only in CgB16, while foxj2 and gstk1 ware located only in CgA16 ( Figure 1C ).

The distributions of Cgptpn6-A and Cgptpn6-B in 12 adult tissues of gibel carp were analyzed by qPCR. Two specific pairs of primers were designed to amplify Cgptpn6-A and Cgptpn6-B, respectively. Cgptpn6-A and Cgptpn6-B were abundantly expressed in the immune-related tissues, such as spleen, head kidney, kidney, and thymus ( Figure 2 ). Cgptpn6-A expression was remarkably higher (20-1726 folds) than that of Cgptpn6-B in all tissues, suggesting that Cgptpn6-A may play a dominant role in immune regulation.

Subsequently, the dynamic expression changes of Cgptpn6-A and Cgptpn6-B were investigated after stimulation with poly I:C, poly dA:dT, and SVCV. Cgptpn6-A and Cgptpn6-B expression increased up to 9.5-fold and 3-fold, respectively, at 24 h after poly I:C treatment ( Figure 3A ). Poly dA:dT showed a weaker stimulation (4.7-fold at 72 h) for the upregulation of Cgptpn6-A expression and showed no effect on the expression of Cgptpn6-B ( Figure 3D ). In contrast, Cgptpn6-A and Cgptpn6-B expression were remarkably up-regulated (2293- and 68-fold respectively) at 48 h after SVCV infection ( Figure 3G ). These results indicate that Cgptpn6-A and Cgptpn6-B may both participate in the host immune response, with Cgptpn6-A playing a dominant role. Similar to Cgptpn6, the other IFN-related genes, such as Cgifn, Cgirf3, Cgrig-i, and Cgviperin, all have two homeologs (52). Cgifn-A and Cgifn-B, as well as Cgirf3-A and Cgirf3-B, showed similar dynamic expression changes ( Figures 3B, C, E, F, H, I ) as Cgptpn6-A and Cgptpn6-B, implying potential association between Cgptpn6 and IFN response.

To explore the association between Cgptpn6 and innate immunity, the effects of Cgptpn6-A and Cgptpn6-B on IFN regulation were examined. The overexpression of CgSHP1-A and CgSHP1-B both remarkably inhibited CaIFN promoters and ISRE activities induced by poly I:C, poly dA:dT, and SVCV ( Figures 4A–C ). In comparison with CgSHP1-B, only a slightly stronger inhibition of CgSHP1-A was observed. Previous studies showed that fish IFN response could be triggered through the RLR signaling pathway (60). As shown in Figure 4D , the co-transfection of CgSHP1-A or CgSHP1-B can obviously inhibit the activities of CaIFNpro and ISRE promoted by CaMAVS, CaMITA, and CaTBK1. The repression effects seemed to be stronger by the co-transfection of CgSHP1-A relative to CgSHP1-B. In addition, both CgSHP1-A and CgSHP1-B had no significant effect on the activities of CaIFNpro and ISRE induced by CaIRF3.

These results were further supported by qPCR findings. The upregulated expression of RLR molecules (Cgrig-i-A and Cgrig-i-B), ifn (Cgifn-A and Cgifn-B), and ISGs (Cgviperin-A and Cgviperin-B) induced by poly I:C or poly dA:dT were remarkably reduced by CgSHP1-A or CgSHP1-B overexpression. Similarly, the inhibitory effect of CgSHP1-A was more significant than that of CgSHP1-B ( Figures 5A, B, E, F ). Similar to the results of the luciferase activity assay, the increased expression of Cgirf3-A and Cgirf3-B were rarely influenced by the CgSHP1-A or CgSHP1-B overexpression induced by poly I:C ( Figures 5C, D ), while the upregulated expression of Cgirf3-A and Cgirf3-B were decreased by the CgSHP1-A or CgSHP1-B overexpression induced by poly dA:dT ( Figures 5G, H ). These data demonstrate that CgSHP1-A and CgSHP1-B may negatively regulate IFN response through the RLR signaling pathway.

To further decipher the relationship between Cgptpn6 and the RLR signaling pathway, Co-IP experiments were performed using CgSHP1-A-Flag and RLR cascades with the Myc tag (CaMAVS-Myc, CaMITA-Myc, CaTBK1-Myc and CaIRF3-Myc). The results clearly showed that CgSHP1-A was efficiently associated with CaMITA and CaTBK1, not with CaMAVS and CaIRF3 ( Figure 6A ). However, CgSHP1-B-Myc could only be efficiently pulled down by CaTBK1-Flag ( Figure 6B ). The interactions between CgSHP1-A and CgSHP1-B with CaTBK1 were confirmed by the reverse assays ( Figures 6C, D ). The subcellular locations of CgSHP1s and CaMITA or CaTBK1 were also monitored in EPC cells. Consistent with the findings of a previous report (61), CgSHP1-A and CgSHP1-B both localized in the cytosol, implying that SHP1 is an intracellular cytoplasmic signaling enzyme ( Figures 6E, F ). The colocalization results showed that the green fluorescence signals of CgSHP1-A or CgSHP1-B were uniformly overlapped with the red signals of CaTBK1, and partly overlapped with the CaMITA signals in the cytosol ( Figures 6E, F ).

Next, we investigated the protein changes to determine the effect of CgSHP1-A and CgSHP1-B on the CaTBK1. Overexpression of CgSHP1-A or CgSHP1-B exerted little influence on the expression of CaTBK1 ( Figure 6G ). Since SHP1 is known as protein tyrosine phosphatase, we speculated that CgSHP1-A and CgSHP1-B may affect the post-translational status of some downstream molecules phosphorylated by TBK1. We first confirmed that CaMITA was indeed phosphorylated by CaTBK1. When CaMITA was co-transfected with CaTBK1, weakly shifted bands with higher molecular weights were detected. As expected, these bands disappeared after treatment with CIP ( Figure 6H ). We subsequently investigated the role of CgSHP1-A and CgSHP1-B in CaTBK1 kinase activity. The phosphorylated CaMITA was reduced with overexpression of CgSHP1-A or CgSHP1-B. Interestingly, CgSHP1-A degrades unphosphorylated CaMITA ( Figure 6I ). Taken together, these data demonstrate that both CgSHP1-A and CgSHP1-B inhibit CaTBK1-induced phosphorylation of CaMITA.

The interaction between CgSHP1-A and CaMITA was further confirmed by the reverse assay ( Figure 7A ). To determine the effect of CgSHP1-A on CaMITA, CgSHP1-A was cotransfected with CaMITA. Overexpression of CgSHP1-A caused a significant reduction of CaMITA in a dose-dependent manner ( Figures 7B, C ). Consistent with the no or very weak interaction between CgSHP1-B and CaMITA ( Figure 6B ), the overexpression of CgSHP1-B did not reduce CaMITA level ( Figure 7B ). A proteasome inhibitor (MG132) and an autophagy–lysosomal pathway inhibitor 3-methyladenine (3-MA) were used to examine the process underlying the CgSHP1-A-mediated CaMITA degradation. In comparison with the control (DMSO treatment) and MG132 groups, 3-MA could effectively block the degradation of CaMITA induced by CgSHP1-A in a dose-dependent manner ( Figures 7D, E ), implying that CgSHP1-A can degrade CaMITA via an autophagy-lysosomal pathway. To test this speculation, we preliminarily evaluated several autophagic components to identify which one could interact with CgSHP1-A. Co-IP assays showed that CgSHP1-A interacted with CaATG14 ( Figure 7F ). Similarly, the interaction between CaMITA and CaATG14 was also confirmed ( Figure 7G ). These data demonstrate that CgSHP1-A can degrade CaMITA probably through CaATG14-mediated autophagy signaling pathway.

Since CgSHP1-A and CgSHP1-B negatively regulate the IFN response, the modulation of CgSHP1 to the antiviral innate immune response was evaluated. In comparison with empty vector control infected with SVCV (MOI = 0.01), the overexpression of CgSHP1-A or CgSHP1-B in GiCB cells both resulted in an enhanced CPE ( Figure 8A ) and the viral titers increased about 10^4.5- and 10^3.94-fold respectively at 2 days post-infection ( Figure 8B ). In addition, the upregulated expression of Cgifn-A, Cgifn-B, Cgviperin-A and Cgviperin-B induced by SVCV were remarkably repressed by overexpression of CgSHP1-A or CgSHP1-B ( Figure 8C ). The inhibition effects of CgSHP1-A transfection were stronger than those of CgSHP1-B transfection. Meanwhile, more abundant transcripts of SVCV genes, n, p, m and g were detected in the CgSHP1-A overexpressed group ( Figure 8D ). These data indicate that both CgSHP1-A and CgSHP1-B negatively regulate the cellular antiviral response, in which CgSHP1-A is potentially dominant.