Gene duplication leads to subfunctionalization of paralogs. In mammals, IFN-γ is the sole member of the type II IFN family and binds to a receptor complex consisting of IFN-γR1 and IFN-γR2. In teleost fish, IFN-γ and its receptors have been duplicated due to the teleost-specific whole-genome duplication event. In this study, the functions of an IFN-γ–related (IFN-γrel) cytokine were found to be partially retained relative to IFN-γ in grass carp (Ctenopharyngodon idella [CiIFN-γrel]). CiIFN-γrel upregulated the expression of proinflammatory genes but had lost the ability to activate genes involved in Th1 response. The results suggest that CiIFN-γrel could have been subfunctionalized from CiIFN-γ. Moreover, CiIFN-γrel induced STAT1 phosphorylation via interaction with duplicated homologs of IFN-γR1 (cytokine receptor family B [CRFB] 17 and CRFB13). Strikingly, CiIFN-γrel did not bind to the IFN-γR2 homolog (CRFB6). To gain insight into the subfunctionalization, the crystal structure of CiIFN-γrel was solved at 2.26 Å, revealing that it forms a homodimer that is connected by two pairs of disulfide bonds. Due to the spatial positions of helix A, loop AB, and helix B, CiIFN-γrel displays a unique topology that requires elements from two identical monomers to form a unit that is similar to IFN-γ. Further, mutagenesis analyses identified key residues interacting with CiIFN-γrel receptors and those required for the biological functions. Our study can help understand the subfunctionalization of duplicated IFN-γ paralogs in fish.

IFN is a secretory protein with a broad spectrum of antiviral, antitumor, and immunomodulatory functions. In tetrapods, IFNs can be divided into type I, II, and III families based on the receptors they interact with and biological functions (13). Type I and type III IFNs have multiple members and activate distinct receptors to induce the expression of antiviral effectors. Type I IFNs bind to IFNAR1/IFNAR2, whereas type III IFNs interact with IFN-λ receptor 1/IL-10 receptor 2 (4, 5). The type II IFN family consists of a single member and exerts weak antiviral functions through IFN-γ and IFN-γR2 (6, 7). It is mainly produced by CD4+ Th1 lymphocytes, innate lymphoid cells, and NK cells and plays a central role in the regulation of Th1 response and activation of macrophages (6, 8, 9). Teleost fish have two type II IFN members, namely IFN-γ and IFN-γ–related (IFN-γrel) molecule, which have multiple copies of paralogs and share low sequence homology (e.g., 17.2–25.5% aa identities between IFN-γrel and fish IFN-γ molecules) (10). It is believed that the genes coding for IFN-γ and IFN-γrel had diverged during the teleost specific genome duplication event occurred in the ancestor of teleosts. The Ifn-γ and Ifn-γrel genes are tandemly linked in the genome and have a four-exon/three-intron genomic organization (1113). Despite low sequence homology (16.9–26.9% aa identities) with their mammalian counterparts, fish IFN-γ exhibit conserved functions such as the enhancement of phagocytic activity and antigen (Ag) presentation by macrophages and the induction of the production of nitric oxide (NO) (1315). However, information on teleost IFN-γrels remains limited (16, 17). Studies in common carp have shown that IFN-γrel is mainly produced by IgM+ cells and is involved in regulating humoral immunity (18). In goldfish, IFN-γrel induces the expression of genes encoding IFN receptors and NO synthase in leukocytes (16).

In mammals, IFN-γ binds to a tetrameric receptor complex and triggers a cascade of signals in target cells via the JAK/STAT pathway (19, 20). The receptor complex consists of two receptors: IFN-γR1, primarily for ligand binding, and IFN-γR2, for signal transduction. Both receptors are single transmembrane proteins and belong to the class II cytokine receptor family. Upon activation by IFN-γ, IFN-γR1 and IFN-γR2 oligomerize to recruit JAK1 and JAK2, leading to recruitment, phosphorylation, and formation of STAT1 homodimer, which subsequently translocates to the nucleus. Phosphorylated STAT1 homodimer binds to the IFN-γ activation site in the promoter of target genes, thereby initiating gene transcription (21, 22). Although the signaling pathway of IFN-γ in inducing phosphorylation of STAT1 and gene transcription is conserved in fish (2325), whether it can be activated by IFN-γrel is unclear.

Previous studies have shown that fish IFN-γ and IFN-γrel activate distinct receptors to elicit functions. IFN-γ interacts with the cytokine receptor family B (CRFB) 6 and CRFB13, whereas IFN-γrel binds to CRFB6 and CRFB17 (26, 27). Fish CRFB13 and CRFB17 are homologs of mammalian IFN-γR1, whereas CRFB6 is equivalent to IFN-γR2. A recent report has demonstrated that Fugu (Takifugu rubripes) IFN-γ displays weaker binding affinity with CRFB13 than with CRFB17 (28). CRFB6, which is shared by IFN-γ and IFN-γrel, has low binding affinity with its ligands and is primarily responsible for initiation of cellular signaling. Crfb6 is expressed at relatively low levels in cells and is inducible. It is generally perceived that the CRFB6 transcript level may play a role in the regulation of cellular responses mediated by the ligands (29).

The crystal structure of IFN-γ has been solved (30). Human (Homo sapiens) IFN-γ forms a globular homodimer with overall dimensions of ∼60 × 40 × 30 Å (30). The IFN-γ monomer consists of six tightly associated α-helices, each harboring 9 to 21 residues. The 12 α-helices of the IFN-γ dimer are arranged in parallel to form a 2-fold axis, in which the four middle α-helices from each peptide chain are intertwined to stabilize the overall structure, whereas the N- and C- terminal helices are organized in a domain-swapped manner. No apparent antiparallel four-helix domains are present. The stability of the IFN-γ dimer relies on the hydrophobic interaction between the helices, but mostly between C and D helices. Helix C is the most hydrophobic helix buried in the core of the dimer (30, 31). The ternary complex structure of human IFN-γ/IFN-γR1/IFN-γR2 shows that helices A, B, and F, loop AB, and C terminal region are involved in the interaction with IFN-γR1, whereas helix D primarily interacts with IFN-γR2 (32). Recently, the crystal structure of flounder (Paralichthys olivaceus) IFN-γ has been solved and shows a seven α-helical topology, contrasting with the six α-helical structure seen in mammalian IFN-γ molecules. One notable feature of the flounder IFN-γ dimer is that the C terminus is extended due to the additional helix G (33), suggesting that the conformation changes may impact on the interaction with its receptors.

Gene duplication and loss lead to neo- and subfunctionalization of genes. In this article, we found that as a duplicate of IFN-γ, carp IFN-γrel bound to CRFB13 and CRFB17 and shared part, but not all, of the functional properties of IFN-γ. To gain insight into its subfunctionalization, we solved the crystal structure of the IFN-γrel molecule and investigated the interactions with the receptors.

Healthy grass carp (Ctenopharyngodon idella; ∼100 g) were obtained from a local fish farm in the suburb of Shanghai, China. Fish were kept in an indoor aquarium, with circulating water at 25 ± 2°C for 2 weeks before experiments. Fish were anesthetized with MS-222 (100 mg/l; Sigma-Aldrich), and the head kidney cells were isolated and extracted. All experiments were conducted under the guidelines of Shanghai Ocean University on the use of animals for research (SHOU-DW-2019-003).

Human embryonic kidney (HEK) 293 cells were cultured at 37°C in DMEM (Life Technologies) supplemented with 10% FBS (Life Technologies) and 1% antibiotics (10,000 U/ml penicillin and streptomycin; Life Technologies) in an incubator with 5% CO2. The epithelioma papulosum cyprini (EPC) cell line was cultured in M199 medium (Life Technologies) containing 10% FBS and 1% antibiotics at 28°C in a 5% CO2 incubator. The Chinese hamster ovary (CHO)-S cells and HEK293F cells were cultured in ExpiCHO expression medium (Life Technologies) and Expi293 expression medium (Life Technologies), respectively, in a rotating incubator (90 rpm) supplemented with 8% CO2 at 37°C.

The mature peptide of grass carp IFN-γrel (GenBank accession number: ACN56578.1, starting from F25) was amplified by PCR using primers listed in Supplemental Table I. The C. idella IFN-γrel (CiIFN-γrel) cytokine fragment was ligated into pET-21d at the restriction sites of NcoI and BamHI, giving rise to pET-21d-CiIFN-γrel. The pET-21d-CiIFN-γrel plasmid was transformed into the Escherichia coli (E. coli) Rosetta (DE3) cells for the production of native recombinant protein and into the methionine-auxotrophic E. coli B834 (DE3) cells for production of selenomethionine (SeMet)–substituted protein. After induction with isopropyl β-d-thiogalactoside (1 mM), the protein was expressed as inclusion bodies, which were dissolved in a denaturing buffer containing 6 M guanidine hydrochloride (VWR International), 10% glycerol, 50 mM Tris-HCl (pH 8), 100 mM NaCl, and 10 mM EDTA at a final concentration of 30 mg/ml. Ten-milliliter protein solution was then dropped into 1 l refolding buffer (100 mM Tris-HCl, 2 mM EDTA, 400 mM l-arginine–HCl, 0.5 mM oxidized glutathione, and 5 mM reduced glutathione [pH 8]) at 4°C and stirred for 48 h to allow protein to be refolded. The refolded protein was concentrated to 30 ml using a 10-kDa cutoff filter (Millipore), mixed with 120 ml equilibration buffer (20 mM Tris-HCl [pH 8] and 300 mM NaCl), and concentrated again into 5 ml. The protein was purified by size-exclusion chromatography using a Superdex 75 16/60 column (GE Healthcare) and checked by gel electrophoresis under reducing and nonreducing conditions. The concentration was determined by the bicinchoninic acid method. The protein was stored at −80°C before use.

Hanging drop vapor diffusion was used to crystallize the protein. Sparse matrix screen kits such as Crystal Screen Cryo I/II, Crystal Screen I/II, Index, Peg/Ion Screen (Hampton Research), and Wizard Classic 1–4 (Rigaku) were used to screen for the crystallization conditions. A total of 160 µl of mother liquor was added to the pool of the 48-well seat drop plates. The purified native and SeMet CiIFN-γrel proteins were concentrated to 5 mg/ml and 10 mg/ml, respectively, and mixed with reservoir buffer at a 1:1 ratio (v/v). Native and SeMet CiIFN-γrel crystals appeared after 2 d using the Wizard Classic 2 Kit (no. 12) and Index Kit (no. 75), respectively. Nylon loops (0.2 mm) were used to take the protein crystals for transferring to the reservoir solution containing 17% glycerin and flash-frozen in liquid nitrogen. Diffraction data of the native and SeMet CiIFN-γrel crystals were collected on the beamline BL18U at a wavelength of 0.97923 Å with an ADSC 315 CCD detector at the Shanghai Synchrotron Radiation Facility. The crystal intensities were indexed, integrated, corrected for absorption, scaled, and merged using the HKL-3000 software package. The SeMet CiIFN-γrel data were used as the search model, and the structure of CiIFN-γrel was solved by molecular replacement with Phaser in the CCP4 package (34, 35). The restrained refinement was performed using REFMAC5 (36), and refinement rounds were implemented using phenix as previously described (37). The CiIFN-γrel structure has been deposited in the Protein Data Bank (PDB) under accession number 7X45 (https://doi.org/10.2210/pdb7X45/pdb). The structural analyses were completed using the PyMOL (DeLano Scientific) program. The buried surface area (BSA) was analyzed using the online tool PDBePISA (https://www.ebi.ac.uk/msd-srv/prot_int/pistart.html).

The coding regions of CiIFN-γ (GenBank accession number AGQ16236.1) and CiIFN-γrel were cloned into the expression plasmid pcDNA3.4 (GENEWIZ). Both proteins were tagged with a six-histidine tag at the C terminus. Plasmid DNA was extracted using an E.Z.N.A. Endo-Free Plasmid Midi Kit (Omega Bio-tek) for transfection. The CHO-S and HEK293F cells were cultured in a 25-cm2 flask to reach 5 × 106 cells/ml and transfected with 25 μg pcDNA3.4-CiIFN-γ or pcDNA3.4-CiIFN-γrel using the ExpiFectamine CHO Transfection Kit (for CHO-S cells) (Life Technologies) and the ExpiFectamine 293 Transfection Kit (for HEK293F cells) (Life Technologies), respectively. Eighteen hours after transfection, the expression enhancement reagent was added to the cells. The CiIFN-γrel protein expressed in the HEK293F cells was collected at day 4 and the CiIFN-γ protein expressed in the CHO-S cells collected at day 8. After validation by Western blotting using an anti-6x His Ab (HUABIO), the proteins were purified using His Trap HP affinity columns (Cytiva) and confirmed by mass spectrometry (Sangon Biotech).

Nanoscale differential scanning fluorometry (nanoDSF) was performed using Prometheus NT.48 (NanoTemper Technologies). The rCiIFN-γrel and CiIFN-γ proteins were dissolved in a buffer containing 20 mM Tris-HCl (pH 8) and 300 mM NaCl at a concentration of 1 µg/µl. Thermal unfolding experiments were performed in the NT.48 instrument by loading 10 µl of protein samples into nanoDSF-grade high-sensitivity capillaries (PR-C006; NanoTemper Technologies) and exposed under thermal stress from 20°C to 95°C at a thermal ramping rate of 1°C/min. Protein unfolding was detected by the change of tryptophan fluorescence at 330 and 350 nm with a dual-UV detector. The onset temperature and melting temperature (Tm) were calculated by detecting the first derivative of the fluorescence ratio at both wavelengths (F350/F330) using PR.ThermControl software (NanoTemper Technologies).

Leukocytes were isolated from grass carp head kidney using a Percoll gradient approach and adjusted to 5 × 106 cells/ml as previously described (38). The head kidney leukocytes (HKLs) were stimulated with the CiIFN-γ or CiIFN-γrel protein purified from CHO-S or HEK293F cells (10, 50, and 250 ng/ml) at 28°C for 6 h and 12 h. The control cells were treated with PBS. Total RNA was extracted from cells using TRIzol reagent (Sigma-Aldrich), and cDNA was synthesized using a 2× Hifair II SuperMix Plus Kit (Yeasen) according to the manufacturer’s instructions. The cDNA samples were stored at −20°C.

Quantitative real-time PCR (qPCR) was carried out using the iTaq Universal SYBR Green Supermix (Bio-Rad Laboratories) and the LightCycler 480 II real-time PCR system (Roche). The PCR product of the target gene was purified and diluted in a series of 10-fold dilutions to obtain a standard curve for calculation of gene expression levels. The expression levels were calculated as arbitrary units normalized to that of elongation factor 1α (internal reference). Fold changes of expression were calculated by comparing the average expression levels of the experimental group with that of the corresponding control group. Gene primers for qPCR analysis are listed in Supplemental Table I.

Based on the modeled ligand/receptor complex, key residues of CiIFN-γrel potentially involved in the interaction with receptors were identified for mutagenesis. The mutated residues included His20, Gln23, Arg29, Thr31, Gln45, Glu76, Lys92, Lys93, Lys115, and Asp128. Primers used for mutations are described in Supplemental Table I. The plasmids expressing CiIFN-γrel mutants were extracted using an E.Z.N.A. Endo-Free Plasmid Midi Kit (Omega Bio-tek). Mutant plasmids were transfected into HEK293F cells and the culture media containing recombinant proteins collected for purification. Proteins were purified using Ni-NTA affinity columns and verified by Western blotting.

The pcDNA3.4-CRFB6-Flag (GenBank accession number AMT92201.1), pcDNA3.4-CRFB13-Ha (GenBank accession number AMT92202.1), and pcDNA3.4-CRFB17-Myc (GenBank accession number AMT92203.1) were constructed (GENEWIZ). For the coimmunoprecipitation (Co-IP) experiments, HEK293 cells were cultured in 25-cm2 culture flasks and cotransfected with 2.5 μg IFN-γrel-His plus 2.5 μg CRFB6-Flag, CRFB13-Ha, CRFB17-Myc, or pcDNA3.4 (vector) using jetOPTIMUS Transfection Reagent (Polyplus-transfection). After 24 h, cells were carefully washed with ice-cold PBS and lysed in RIPA lysis buffer (500 μl/flask; Beyotime) containing protease inhibitor mixture (CWBIO) followed by centrifugation at 13,000 × g at 4°C for 20 min. The cell supernatants were then incubated with 3 µl Rabbit/Mouse anti-6x His Tag Ab (HUABIO) at 4°C overnight and with 100 µl protein A/G agarose resin (Yeasen) at room temperature for 1 to 2 h. The resin was centrifuged at 1000 × g for 3 min, washed with PBS five times, and resuspended with 80 µl PBS. The cell lysate and the recovered resin were analyzed by Western blotting (39). The primary Abs included rabbit anti-6x His Tag Ab, rabbit anti-hemagglutinin (Ha) Tag Ab, mouse anti-Flag Tag Ab, and mouse anti-Myc Tag Ab (HUABIO).

To compare the binding affinity of wild-type (WT) and mutant CiIFN-γrel proteins with receptors, HEK293 cells were cotransfected with CiIFN-γrel WT or mutant plasmids plus pcDNA3.4-CRFB6-Flag, pcDNA3.4-CRFB13-Flag, or pcDNA3.4-CRFB17-Flag. The control group was transfected with pcDNA3.4 plasmid. Co-IP assay was performed as described above.

HKLs were isolated and stimulated with 50 or 250 ng/ml WT and mutant CiIFN-γrel proteins at 28°C for 6 h. The cells were then harvested for analysis of gene expression by real-time PCR. To date, there are no mAbs available for detecting STAT1 phosphorylation in fish. Therefore, we constructed an expression plasmid containing the full-length open reading frame of zebrafish STAT1a (GenBank accession number NP_571555), in which Thr679-Glu706 was replaced by the responding region of human STAT1 (SRPKEAPEPMELDGPKGTGYIKTELISVSEVHPSRLQTTD), which harbors human Tyr701 phosphorylation The EPC cells were seeded in 12-well plates, cultured overnight, and transfected with 1 μg STAT1a-Flag plasmid using jetOPTIMUS Transfection Reagent (Polyplus-transfection). After 24 h, the cells were stimulated with 10, 50, and 250 ng/ml WT CiIFN-γrel or 50 ng/ml mutant proteins for 15 min. The control group was treated with PBS buffer. The cells were lysed in RIPA lysis buffer (100 μl/well) containing PMSF (Beyotime) followed by centrifugation at 13,000 × g at 4°C for 20 min. The cell lysate was analyzed by Western blotting using the rabbit anti–phospho-Stat1 (Y701) (8D6) mAb (Cell Signaling Technology), mouse anti-Flag Tag Ab, or rabbit anti–β-actin Ab (HUABIO).

The qPCR data were analyzed using a one-way ANOVA and a least significant difference post hoc test (SPSS, Inc., Chicago, IL), with p < 0.05 and p < 0.01 considered significant.

IFN-γ is known to induce the expression of genes involved in innate and adaptive immune responses. In teleost fish, IFN-γ has been duplicated into multiple homologs of IFN-γ and IFN-γrel genes (Fig. 1). To investigate whether IFN-γ and IFN-γrel differentially modulate the expression of immune genes, the CiIFN-γ and CiIFN-γrel proteins were expressed in the CHO-S and HEK293F cells and purified (Supplemental Fig. 1A, 1B). Two protein bands of ∼17 kDa and 23 kDa were detected for CiIFN-γrel by SDS-PAGE analysis (Supplemental Fig. 1B) and Western blotting using an anti-6x His Ab (Supplemental Fig. 1D). Both bands were validated to be CiIFN-γrel by mass spectrometry (Supplemental Table II). For the CiIFN-γ, a single protein of ∼19.6 kDa was purified (Supplemental Fig. 1A) and verified by Western blotting (Supplemental Fig. 1C). Gel filtration chromatography and nonreducing gel electrophoresis indicate that native CiIFN-γrel is likely to form a dimer (Supplemental Fig. 1G–I). The recombinant proteins were used to stimulate the primary HKLs for 6 h and 12 h. It was found that Il-1β, Il-6 and Cxcl8-l1 were upregulated by both CiIFN-γ and CiIFN-γrel at the doses of 50 ng/ml and 250 ng/ml (Fig. 2A–C). Cxcl11.1b, a chemokine that primarily directs T cell migration, was also markedly induced by CiIFN-γ, but curiously not by CiIFN-γrel (Fig. 2D). Consistent with previous reports (13), MhcII (involved in Ag presentation) and T-bet (a transcription factor of CD4 Th1 cells) were increased after stimulation with CiIFN-γ (Fig. 2E, 2F). Conversely, we did not detect any alteration of expression of MhcII and T-bet after stimulation with CiIFN-γrel. Moreover, CiIFN-γ had stimulatory effects on the expression of Mx1 (except for the dose of 10 ng/ml at 6 h) and Stat1a at 50 ng/ml at 12 h (Fig. 2G, 2H). Lastly, the expression of genes mediating Ag processing, including Tap-1, Ciita, Ctsd, and Ctsl, were examined. It was found that Tap-1 and Ciita were induced by both CiIFN-γrel and CiIFN-γ, whereas Ctsd and Ctsl was activated by CiIFN-γ but not by CiIFN-γrel (Fig. 2I–L). A heat map summarizing gene expression is shown in (Fig. 2M, indicating that CiIFN-γrel displays a narrower spectrum of biological activities than CiIFN-γ.

FIGURE 1.

Gene copy numbers of type II IFNs and receptors in jawed vertebrates.

FIGURE 1.

Gene copy numbers of type II IFNs and receptors in jawed vertebrates.

Close modal
FIGURE 2.

IFN-γ and IFN-γrel differentially regulate the expression of immune genes. HKLs were stimulated with IFN-γ or IFN-γrel proteins for 6 h and 12 h, respectively. (AL) Total RNA was extracted for qPCR analysis of expressions of Il-1β, Il-6, Cxcl8-l1, Cxcl11.b, Mhc II, T-bet, Mx1, and Stat1a. (M) Heat map showing modulation of gene expression. Dark gray and light gray represent upregulation and no change of gene expression, respectively. Data are presented as mean + SEM (n = 4). *p < 0.05, **p < 0.01, ***p < 0.001 are considered significant.

FIGURE 2.

IFN-γ and IFN-γrel differentially regulate the expression of immune genes. HKLs were stimulated with IFN-γ or IFN-γrel proteins for 6 h and 12 h, respectively. (AL) Total RNA was extracted for qPCR analysis of expressions of Il-1β, Il-6, Cxcl8-l1, Cxcl11.b, Mhc II, T-bet, Mx1, and Stat1a. (M) Heat map showing modulation of gene expression. Dark gray and light gray represent upregulation and no change of gene expression, respectively. Data are presented as mean + SEM (n = 4). *p < 0.05, **p < 0.01, ***p < 0.001 are considered significant.

Close modal

Two homologs of IFN-γR1 (CRFB13 and CRFB17) are found in teleosts, whereas IFN-γR2 exists as a single copy (Fig. 1). To investigate the interaction of IFN-γrel with the receptors, HEK293 cells were cotransfected with IFN-γrel-His and CRFB6-Flag, CRFB13-Ha, CRFB17-Myc, or pcDNA3.4 for an immunoprecipitation assay. The results showed that IFN-γrel was coimmunoprecipitated with CRFB13 and CRFB17 but had relatively stronger binding affinity with CRFB17 than CRFB13. However, no binding of IFN-γrel with CRFB6 was detected (Fig. 3A). Moreover, we observed that CRFB6 could directly bind to CRFB13 and CRFB17 (Fig. 3B). The pairing relationship of receptors is summarized in (Fig. 3C. Taken together, these results suggest that IFN-γrel may activate two distinct receptor complexes consisting of CRFB13/CRFB6 and CRFB17/CRFB6 to trigger cellular responses.

FIGURE 3.

Interaction of IFN-γrel with receptors. (A) Co-IP analysis of IFN-γrel with CRFB6, CRFB13, and CRFB17. Five micrograms of IFN-γrel-His plus CRFB6-Flag, CRFB13-Ha, CRFB17-Myc, or pcDNA3.4 was transfected into the HEK293 cells and cultured for 24 h. (B) Co-IP analysis of CRFB6 with CRFB13 and CRFB17. CRFB6 plus CRFB13, CRFB17, or pcDNA3.4 were transfected into the HEK293 cells and cultured for 24 h. Abs against Ha tag, Flag tag, and Myc tag were used for immunoblotting. (C) The binding relationship between IFN-γrel and receptors.

FIGURE 3.

Interaction of IFN-γrel with receptors. (A) Co-IP analysis of IFN-γrel with CRFB6, CRFB13, and CRFB17. Five micrograms of IFN-γrel-His plus CRFB6-Flag, CRFB13-Ha, CRFB17-Myc, or pcDNA3.4 was transfected into the HEK293 cells and cultured for 24 h. (B) Co-IP analysis of CRFB6 with CRFB13 and CRFB17. CRFB6 plus CRFB13, CRFB17, or pcDNA3.4 were transfected into the HEK293 cells and cultured for 24 h. Abs against Ha tag, Flag tag, and Myc tag were used for immunoblotting. (C) The binding relationship between IFN-γrel and receptors.

Close modal

To understand the interaction of CiIFN-γrel with the receptors, we solved the crystal structure of CiIFN-γrel using single-wavelength anomalous dispersion (Fig. 4A, Supplemental Fig. 1E, 1F). Due to the high Rfree value of 0.367, the space group symmetry of the SeMet protein structure was reduced to P1. Using I/sigI of 2, equal to CC1/2 of 0.7, for the high-resolution cutoff, the CiIFN-γrel structure was resolved at a resolution of 2.26 Å with space group P3112 and refined to a final R work value of 19% and an Rfree value of 23% (Fig. 4A, Table I). Details of data collection, phasing, and refinement are given in Table I. The CiIFN-γrel is a homodimer with identical peptide chains related by a dyad axis. Each peptide chain is composed of six α helices: helix A, Glu11-His25; helix B, Val40-Leu46; helix C, Cys54-Ser75; helix D, Lys82-Tyr98; helix E, Lys99-Ser113; and helix F, Gly119-Lys140 (Fig. 4B). Among the six helices, both helices C and F contain 22 aa residues, whereas helix B is the shortest, containing 7 residues. Notably, helix C contains 10 relatively well-conserved hydrophobic residues, making it highly hydrophobic. Helix F is bent at Asp128 with an angle of 136°. Helices D and E are connected by a single residue (Lys99). Loop AB (Tyr26-Phe36) is the longest loop, in which residue Thr28-Trp33 of loop AB is not solved due to the poor electron density. His77-Gly80 of loop CD is also not solved (Fig. 4A, 4B).

FIGURE 4.

The structure of CiIFN-γrel. The two subunits of CiIFN-γrel are colored in magenta and lime green, respectively. Helices are indicated by letters. (A) The dimeric structure of CiIFN-γrel. (B) The monomeric structure of CiIFN-γrel. (C) A schematic diagram showing the organization of the interdigitating conformation of CiIFN-γrel. Helix B, loop BC, and helix C interlock with helix E, loop EF, and helix F. (D) Human IFN-γ monomeric structure. (E) A schematic diagram showing the organization of helix B-loop BC-helix C and segment helix E-loop EF-helix F. Dashed lines indicate unsolved loops.

FIGURE 4.

The structure of CiIFN-γrel. The two subunits of CiIFN-γrel are colored in magenta and lime green, respectively. Helices are indicated by letters. (A) The dimeric structure of CiIFN-γrel. (B) The monomeric structure of CiIFN-γrel. (C) A schematic diagram showing the organization of the interdigitating conformation of CiIFN-γrel. Helix B, loop BC, and helix C interlock with helix E, loop EF, and helix F. (D) Human IFN-γ monomeric structure. (E) A schematic diagram showing the organization of helix B-loop BC-helix C and segment helix E-loop EF-helix F. Dashed lines indicate unsolved loops.

Close modal
Table I.

Data collection and refinement statistics

Native CiIFN-γrelSeMet-CiIFN-γrel
Data collection   
 Space group P3112 P1 
 Cell dimensions   
  a, b, c (Å) 48.58, 48.58, 208.66 51.29, 51.29, 205.465 
  α, β, γ (°) 90, 90, 120 90, 90, 120 
  Resolution (Å) 42.07-2.33 (2.26–2.26) 44.42-2.44 (2.38–2.38) 
  Rsym or Rmerge (%)a 1.146 (0.112) 0.498 (0.074) 
  II 2.0 (15.4) 2.1 (9.3) 
  Completeness (%) 97.6 (99.8) 93.5 (90.06) 
  Redundancy 13.3 (17.8) 3.8 (3.6) 
Refinement   
 Resolution (Å) 42.072-2.26 44.418-2.380 
 No. of reflections 12,936 62,415 
Rwork/Rfree (%)b 0.1900/0.2366 0.2347/0.2740 
 No. of atoms   
  Protein 2042  
 Ligand/ion   
  Water 36  
 β-Factors   
  Protein 65.37  
 Ligand/ion   
  Water 49.067  
 RMSD   
  Bond lengths (Å) 0.007  
  Bond angles (°) 1.4475  
Ramachandran statistics   
 In favored regions (%) 95.69  
 Outliers (%) 0.00  
Native CiIFN-γrelSeMet-CiIFN-γrel
Data collection   
 Space group P3112 P1 
 Cell dimensions   
  a, b, c (Å) 48.58, 48.58, 208.66 51.29, 51.29, 205.465 
  α, β, γ (°) 90, 90, 120 90, 90, 120 
  Resolution (Å) 42.07-2.33 (2.26–2.26) 44.42-2.44 (2.38–2.38) 
  Rsym or Rmerge (%)a 1.146 (0.112) 0.498 (0.074) 
  II 2.0 (15.4) 2.1 (9.3) 
  Completeness (%) 97.6 (99.8) 93.5 (90.06) 
  Redundancy 13.3 (17.8) 3.8 (3.6) 
Refinement   
 Resolution (Å) 42.072-2.26 44.418-2.380 
 No. of reflections 12,936 62,415 
Rwork/Rfree (%)b 0.1900/0.2366 0.2347/0.2740 
 No. of atoms   
  Protein 2042  
 Ligand/ion   
  Water 36  
 β-Factors   
  Protein 65.37  
 Ligand/ion   
  Water 49.067  
 RMSD   
  Bond lengths (Å) 0.007  
  Bond angles (°) 1.4475  
Ramachandran statistics   
 In favored regions (%) 95.69  
 Outliers (%) 0.00  

Values in parentheses represent the highest-resolution shell.

a

Rmergei Σ hkl | I (hkl) − < I (hkl) > |/Σ hkl Σi Ii (hkl), where Ii (hkl) is the observed intensity, and < I (hkl) > is the average intensity from multiple measurements.

b

R factor = Σ(Fobs − Fcalc)/ΣFobs; Rfree is the R factor for a subset (5%) of reflections that were selected prior to the refinement calculations but not included in the refinement.

Comparative analysis shows that the IFN-γrel and IFN-γ monomers differ considerably in conformation (Fig. 4B–E). The CiIFN-γrel monomer seems tightly organized. The segment comprising helix A, helix B, and loop AB is located in the exterior surface and forms a concave downward bend to accommodate helix F (Fig. 4B). Furthermore, the segment containing helix B, loop BC, and helix C forms an interdigitating conformation with the segment containing helix E, loop EF, and helix F, which differs with that of IFN-γ (Fig. 4B, 4C). In the IFN-γ monomers, helices A, B, and F are located at opposite sides that are separated by elements consisting of helices C, D, and E (Fig. 4D, 4E). The contact interface involving helix A, loop AB, helix B and helix F is relatively big and has a BSA of 872 Å2, which is much higher than that in the human IFN-γ monomer (104 Å2).

Structural analyses have shown that no disulphide bond is present in the IFN-γ dimers of human, mouse, and flounder. Strikingly, the CiIFN-γrel dimer is connected by two pairs of intermolecular disulphide bonds (Fig. 5A). Cys52 of loop BC of one chain in the CiIFN-γrel is linked to Cys54 of helix C’ of another chain to form a disulfide bond and vice versa. Elements containing helix B-loop BC-helix C and helix B’-loop B’C’-helix C’ are directly linked by the two intermolecular disulfide bonds, which provide a strong interaction force to stabilize the spatial location of element helix A-loop AB-helix B.

FIGURE 5.

The unique conformation of CiIFN-γrel dimeric structure. Ribbon diagrams of CiIFN-γrel (A) and human IFN-γ (B) dimeric structures. The electron density map around the connecting region between helices B and C is shown (A, bottom). Cysteine residues are indicated and disulfide bonds shown as sticks. Cartoon and surface representation of CiIFN-γrel (C) and human IFN-γ (D) structures. Schematic diagrams depicting the connectivity of α-helices of CiIFN-γrel (E) and human IFN-γ (F) dimers. Elements above and below the plane of the figure are indicated by solid and dashed arrows, respectively.

FIGURE 5.

The unique conformation of CiIFN-γrel dimeric structure. Ribbon diagrams of CiIFN-γrel (A) and human IFN-γ (B) dimeric structures. The electron density map around the connecting region between helices B and C is shown (A, bottom). Cysteine residues are indicated and disulfide bonds shown as sticks. Cartoon and surface representation of CiIFN-γrel (C) and human IFN-γ (D) structures. Schematic diagrams depicting the connectivity of α-helices of CiIFN-γrel (E) and human IFN-γ (F) dimers. Elements above and below the plane of the figure are indicated by solid and dashed arrows, respectively.

Close modal

Both CiIFN-γrel and IFN-γ dimers are organized into two structural domains, each consisting of six α-helices. Four α-helical segments within each domain are derived from one chain, whereas the remaining two α-helical segments are contributed by the other chain (Fig. 5A, 5B). Helices A, B, C, and D of chain A and helices E’ and F’ of human IFN-γ-chain B form one domain, where helices A, B, C, and D constitute a cleft intercalated by helix F’ (Fig. 5B). However, the unusual spatial locations of helices A and B in CiIFN-γrel indicate that all of the elements of each chain are displayed on opposite side, thus altering the helical components. For instance, one domain is composed of helices A, B, E, and F of chain A and helices C’ and D’ of chain B (Fig. 5A, 5C). Unlike the intertwined subunits of IFN-γ, CiIFN-γrel has a concave surface but is devoid of a cleft (Fig. 5C, 5D). The connectivity of α-helices for the dimer formation of CiIFN-γrel is unique and has not been reported for any known cytokines (Fig. 5E, 5F).

One interesting finding is that the core element consisting of helices C, D, E, and F of CiIFN-γrel and IFN-γ monomeric structures can be superposed well; however, helix A, loop AB, and helix B do not match (Fig. 6A–C). The spatial positions of helix A, loop AB, and helix B are clearly different. However, when helix A, loop AB, and helix B are swapped between chain A and chain B, the resultant CiIFN-γrel monomer can be superposed with a human IFN-γ monomer (Fig. 6A, 6C). This topology is interesting because it requires elements from two identical monomers to form a unit that is comparable to IFN-γ. Taken together, these features may imply that IFN-γrel and IFN-γ share some, but not all, of the binding sites with the receptor complex.

FIGURE 6.

Structural superposition of CiIFN-γrel with IFN-γ molecules. Superposition of CiIFN-γrel monomer (magenta) with flounder (green-cyan, PDB: 6f1e) (A), bovine (yellow, PDB: 1d9c) (B), and human IFN-γ (blue, PDB: 1fg9) (C). Superposition of CiIFN-γrel dimer with flounder, bovine and human IFN-γ (D).

FIGURE 6.

Structural superposition of CiIFN-γrel with IFN-γ molecules. Superposition of CiIFN-γrel monomer (magenta) with flounder (green-cyan, PDB: 6f1e) (A), bovine (yellow, PDB: 1d9c) (B), and human IFN-γ (blue, PDB: 1fg9) (C). Superposition of CiIFN-γrel dimer with flounder, bovine and human IFN-γ (D).

Close modal

Helices D and E are tightly linked in CiIFN-γrel and flounder IFN-γ, and the angle between the two helices is similar in the two molecules (Fig. 6A). The angle is at least 20° wider than that in bovine and human IFN-γ, resulting in the extension of the CiIFN-γrel structure (Fig. 6A–D). The CiIFN-γrel structure is highly similar to that of flounder IFN-γ with a root mean square deviation (RMSD) value of 2.5 Å (Fig. 6A), but could not be superposed with that of mammalian IFN-γ structures due to the high RMSD values. For example, the RMSD values for CiIFN-γrel/bovine IFN-γ and CiIFN-γrel/human IFN-γ are 7.0 Å and 6.8 Å, respectively (Fig. 6B, 6C).

The CiIFN-γrel dimer has a prolate oval shape. Contrasting with the tightly intertwined IFN-γ dimer, the interaction between the two peptide chains of CiIFN-γrel seems relatively flexible. The two symmetric C helices in the CiIFN-γrel dimer are packed against each other, forming the core interface. All of the elements of each chain of CiIFN-γrel are placed on opposite sides and well apart. The two chains of CiIFN-γrel are linked via 24 hydrogen bonds, similar to that seen in human IFN-γ. However, those residues involved in the formation of hydrogen bonds are not conserved between CiIFN-γrel and IFN-γ. The BSA of CiIFN-γrel is 3293 Å2, which is comparable to that of flounder (4031 Å2), bovine (3809 Å2), and human (3646 Å2) IFN-γ, indicating that CiIFN-γrel is also a tightly interconnected dimer (Table II). The interface of the CiIFN-γrel dimer involves 70 residues (53% of total residues), of which 35 are hydrophobic residues. These residues contribute to a BSA of 1724 Å2, accounting for 52% of total BSA, which is similar to the BSA percentages of hydrophobic residues in the IFN-γ dimer of flounder (2407 Å2; 59%), bovine (2002 Å2; 55%), and human (2103 Å2; 55%) IFN-γ (Table II). The results indicate that dimerization of IFN-γrel and IFN-γ mainly relies on the hydrophobic interaction. It is worth noting that the CiIFN-γrel dimer has three distinct hydrophobic cores in the interface (Fig. 7A), in stark contrast with the IFN-γ dimers of human, bovine, and flounder, which have no apparent hydrophobic cores (Fig. 7B). The first hydrophobic core of CiIFN-γrel is composed of Phe39, Ala125, Phe129, Val132, and Phe133, the second involves Leu57, Leu58, Trp105, and Leu108, and the third involves Met62, Leu63, Ile65, Leu69, Phe70, and Val91 (Fig. 7A, 7C). Interestingly, these 15 residues are conserved in all known IFN-γrel molecules (Fig. 7C). In addition to the hydrophobic cores, the two pairs of disulphide bonds could be important in maintaining the stability of the CiIFN-γrel dimer. Interestingly, the cysteine residues involved in the formation of disulphide bonds are conserved in IFN-γrel molecules (Fig. 7C). Furthermore, we performed nanoDSF analysis to determine the melting temperatures of the CiIFN-γ and CiIFN-γrel proteins and observed that CiIFN-γ had a Tm value of 63.4°C, whereas the CiIFN-γrel protein had Tm values of 56.6°C and 64.3°C (Supplemental Fig. 1J–M).

FIGURE 7.

The hydrophobic cores of CiIFN-γrel. The CiIFN-γrel (A) and human IFN-γ (B) monomeric structures are described as surface. Conserved residues on the surface are colored blue. (C) Sequence alignment shows the conservation of hydrophobic residues in the IFN-γrel (magenta) and IFN-γ (blue and green) molecules. Cysteine residues are boxed.

FIGURE 7.

The hydrophobic cores of CiIFN-γrel. The CiIFN-γrel (A) and human IFN-γ (B) monomeric structures are described as surface. Conserved residues on the surface are colored blue. (C) Sequence alignment shows the conservation of hydrophobic residues in the IFN-γrel (magenta) and IFN-γ (blue and green) molecules. Cysteine residues are boxed.

Close modal
Table II.

The BSA values of CiIFN-γrel and IFN-γ dimers

CiIFN-γrelFlounder IFN-γBovine IFN-γHuman IFN-γ
Helix A 306.87 270.79 249.73 263.27 
Helix B 87.53 — 309.23 241.06 
Helix C 1074.29 838.94 677.67 913.27 
Helix D 469.02 170.89 191.06 160.92 
Helix E 345.36 354.28 371.66 437.57 
Helix F 564.58 1460.42 1257.4 1068.42 
Total BSA (Å) 3293.25 4031 3646.4 3809.4 
CiIFN-γrelFlounder IFN-γBovine IFN-γHuman IFN-γ
Helix A 306.87 270.79 249.73 263.27 
Helix B 87.53 — 309.23 241.06 
Helix C 1074.29 838.94 677.67 913.27 
Helix D 469.02 170.89 191.06 160.92 
Helix E 345.36 354.28 371.66 437.57 
Helix F 564.58 1460.42 1257.4 1068.42 
Total BSA (Å) 3293.25 4031 3646.4 3809.4 

The helices that provide the largest embedding area are shown in boldface.

In human IFN-γ, helix A, helix B, helix F, loop AB, and C terminal region are involved in the interaction with IFN-γR1, whereas helix D primarily interacts with IFN-γR2 (Fig. 8A–D) (32, 40). CiIFN-γrel has low sequence identity with mammalian (12.2–22.2%) and fish (17.2–25.5%) IFN-γ. Superposition analysis reveals that residues interacting with receptors are poorly conserved (Fig. 8C, 8D). Human IFN-γ interacts with IFN-γR1 mainly through 11 hydrogen bonds (Fig. 8A, 8C). In contrast, IFN-γ interacts with IFN-γR2 principally through a small pocket consisting of eight residues of helices A and D (Lys68, Ser69, Glu71, Thr72, Lys74, Glu75, Asp76, and Val79) (Fig. 8B, 8D). Moreover, Thr13, His20, Gln23, Asn127, Asp128, Met131, Asp134, and His141 of CiIFN-γrel could be involved in the interaction with IFN-γR1 (Fig. 8C). Due to the lack of a density map, only five residues, including Glu76, Asp89, Lys92, Lys93, and Asn97, have been identified to potentially bind to IFN-γR2 (Fig. 8D).

FIGURE 8.

Modeling and validation of key residues of CiIFN-γrel in the interaction with receptors. (A) Binary structure of human IFN-γ and IFN-γR1 complex (PDB: 1fg9). Residues of IFN-γ on the contact interface are indicated, and the hydrogen bonds are shown as red dashed lines. (B) Binary structure of human IFN-γ and IFN-γR2 (PDB: 6e3l). (C) Prediction of CiIFN-γrel residues involved in the interaction between ligand and receptor by superposition. Human IFN-γ and CiIFN-γrel are shown as a cartoon and colored gray, respectively. Residues of human IFN-γ (lime green) and CiIFN-γrel (salmon pink) are shown. (D) Prediction of residues of CiIFN-γrel involved in the interaction with CRFB6/IFNγR2 by superposition. Residues on the contact interface of human IFN-γ with IFN-γR2 complex are shown in cyan. (E) Residues selected for mutation. CiIFN-γrel is shown as surface and colored magenta. The transparency is set to 0.2. Residues (yellow) interacting with IFN-γR1 and IFN-γR2 (cyan) are shown as sticks. Co-IP analysis of IFN-γrel mutants with CRFB13 (F) and CRFB17 (G). HEK293 cells were transfected with the IFN-γrel WT or mutant plasmids plus CRFB13 or CRFB17 plasmids. After 24 h, immunoprecipitation assay was performed using rabbit anti-Ha Tag or mouse anti-Myc Tag Ab beads. Abs against 6x His tag, Ha tag, and Myc tag were used for immunoblotting.

FIGURE 8.

Modeling and validation of key residues of CiIFN-γrel in the interaction with receptors. (A) Binary structure of human IFN-γ and IFN-γR1 complex (PDB: 1fg9). Residues of IFN-γ on the contact interface are indicated, and the hydrogen bonds are shown as red dashed lines. (B) Binary structure of human IFN-γ and IFN-γR2 (PDB: 6e3l). (C) Prediction of CiIFN-γrel residues involved in the interaction between ligand and receptor by superposition. Human IFN-γ and CiIFN-γrel are shown as a cartoon and colored gray, respectively. Residues of human IFN-γ (lime green) and CiIFN-γrel (salmon pink) are shown. (D) Prediction of residues of CiIFN-γrel involved in the interaction with CRFB6/IFNγR2 by superposition. Residues on the contact interface of human IFN-γ with IFN-γR2 complex are shown in cyan. (E) Residues selected for mutation. CiIFN-γrel is shown as surface and colored magenta. The transparency is set to 0.2. Residues (yellow) interacting with IFN-γR1 and IFN-γR2 (cyan) are shown as sticks. Co-IP analysis of IFN-γrel mutants with CRFB13 (F) and CRFB17 (G). HEK293 cells were transfected with the IFN-γrel WT or mutant plasmids plus CRFB13 or CRFB17 plasmids. After 24 h, immunoprecipitation assay was performed using rabbit anti-Ha Tag or mouse anti-Myc Tag Ab beads. Abs against 6x His tag, Ha tag, and Myc tag were used for immunoblotting.

Close modal

Based on the above analysis and sequence alignment, we selected 10 key residues for functional analysis. These include His20, Gln23, Arg29, Thr31, Gln45, Glu76, Lys92, Lys93, Lys115, and Asp128 (Fig. 8E). Mutant plasmids were constructed and transfected into HEK293 cells with CRFB13 or CRFB17 plasmids for immunoprecipitation assay. The expression of mutants K92A, K102A, and H20A could not be detected by Western blotting and hence were not subjected to further analysis (Supplemental Fig. 2). Compared to the WT IFN-γrel, mutations of Q23A, T31A, E76A, and D128A resulted in weakened binding with CRFB13, whereas mutations of Q45A and K93A had the opposite effect. It must be noted that mutant R29A had comparable binding affinity with the WT IFN-γrel (Fig. 8F). As for the interaction of IFN-γrel with CRFB17, D128A significantly decreased the binding affinity with CRFB17, whereas other residues had no effect (Fig. 8G).

Having identified the residues of CiIFN-γrel interacting with the receptors, we investigated their effects on cellular signaling. To determine whether WT CiIFN-γrel induces STAT1a phosphorylation, we transfected STAT1a-Flag plasmid in the EPC cells. After 24 h, the cells were stimulated with 10, 50, or 250 ng/ml CiIFN-γrel purified from the HEK293F cells for 15 min. (Fig. 9A shows that phosphorylation of STAT1a increased in the cells stimulated with CiIFN-γrel in a dose-dependent manner. We then chose the dose of 50 ng/ml for evaluating the effect of IFN-γrel mutants on STAT1a phosphorylation. We observed that the STAT1a phosphorylation levels were decreased in the cells treated with R29A, T31A, Q45A, L93A, or D128A mutants compared with WT CiIFN-γrel, with marked inhibition detected for Q45A. In contrast, Q23A and E76A did not alter the STAT1a phosphorylation level (Fig. 9B).

FIGURE 9.

Lys93 and Asp128 of CiIFN-γrel are required for STAT1a phosphorylation and induction of proinflammatory cytokines. (A and B) The EPC cells were transfected with 1 µg STAT1a-Flag plasmid. After 24 h, the cells were incubated with PBS, WT, or mutant IFN-γrel proteins for 15 min and harvested for analysis of STAT1a phosphorylation by Western blotting. (C and D) HKLs were stimulated with 50 and 250 ng/ml of WT or mutant proteins. The expression of Il-1β and Cxcl8-l1 was analyzed by qPCR. Data are presented as mean +SEM (n = 4). *p < 0.05, **p < 0.01 are considered significant.

FIGURE 9.

Lys93 and Asp128 of CiIFN-γrel are required for STAT1a phosphorylation and induction of proinflammatory cytokines. (A and B) The EPC cells were transfected with 1 µg STAT1a-Flag plasmid. After 24 h, the cells were incubated with PBS, WT, or mutant IFN-γrel proteins for 15 min and harvested for analysis of STAT1a phosphorylation by Western blotting. (C and D) HKLs were stimulated with 50 and 250 ng/ml of WT or mutant proteins. The expression of Il-1β and Cxcl8-l1 was analyzed by qPCR. Data are presented as mean +SEM (n = 4). *p < 0.05, **p < 0.01 are considered significant.

Close modal

Next, we sought to evaluate whether the mutants affected the expression of target genes. Previously, Il-1β and Cxcl8-l1 were shown to be upregulated in the HKLs by WT CiIFN-γrel (Fig. 9C, 9D). Therefore, we stimulated the HKLs with CiIFN-γrel mutants for 6 h and evaluated their expression by qPCR. We found that K93A and D128A diminished the ability of CiIFN-γrel to induce the expression of Il-1β and Cxcl8-l1. E76A failed to induce Il-1β expression at the dose of 250 ng/ml. Intriguingly, R29A (250 ng/ml) and T31A (50 ng/ml) resulted in increased stimulatory effects on the expression of Cxcl8-l1 but not of Il-1β. The receptor interactions and functions of CiIFN-γ and CiIFN-γrel are summarized in (Fig. 10).

FIGURE 10.

Proposed model for the receptor interaction and functions of IFN-γ and IFN-γrel.

FIGURE 10.

Proposed model for the receptor interaction and functions of IFN-γ and IFN-γrel.

Close modal

Teleost fish possess duplicated type II IFN genes and receptors. In this study, we demonstrate that CiIFN-γrel upregulates genes that promote inflammation but not those involved in Th1 response cell response and antiviral defense (Fig. 2). This contrasts with the broad functions of IFN-γ (6, 8, 9) and suggests that IFN-γrel may have been subfunctionalized in teleost fish. Previous studies have shown that fish IFN-γ and IFN-γrels exhibit overlapping and distinct functions (16, 24). In common carp (Cyprinus carpio L.), IFN-γrel is expressed in LPS-activated IgM+ B cells and has been shown to be associated with B lymphocyte response (41). Furthermore, fish IFN-γ induce long-lasting stimulatory effects on NO production in monocytes, whereas IFN-γrel’s action appears to be transient (16). Our data support the notion that both IFN-γ and IFN-γrel are able to upregulate the expression of proinflammatory genes, including Il-1β and Cxcl8 (16) and that IFN-γrel’s action is short-lived. In pufferfish and Mandarin fish, IFN-γ weakly activates the expression of antiviral factors and enhances cell resistance to viral infection, whereas IFN-γrels have the opposite effect (42, 43). Consistent with these observations, grass carp IFN-γ had stimulatory effects on the expression of genes involved in the T cell immune response (Cxcl11.1b and T-bet), Ag presentation (e.g., MhcII, Ciita, Tap-1, Ctsd, and Ctsl), and antiviral defense (e.g., Mx1 and Stat1). Our data suggest that as a duplicate cytokine of IFN-γ, IFN-γrel has a narrower spectrum of the functional properties than IFN-γ.

In vertebrates, class II cytokine receptors are divided into two groups, with members of one group having longer extracellular domains for ligand binding and members of the other group having shorter intracellular domains for signal transduction (6, 7). IFN-γR1 belongs to the former subfamily, whereas IFN-γR2 is mainly responsible for activation of cellular signaling. In teleosts, CRFB13 and CRFB17 are equivalent to IFN-γR1, whereas CRFB6 is an ortholog of IFN-γR2 (14). Previous studies have shown that CRFB13 and CRFB17 are activated by IFN-γ and IFN-γrel, respectively (42, 44). In this study, we show that IFN-γrel is able to bind to both CRFB13 and CRFB17; however, it exhibits a weaker binding affinity with CRFB13, indicating that CRFB17 may act as the prime binding receptor for IFN-γrel (43). To our surprise, we did not detect the interaction of IFN-γrel with CRFB6. This may be due to the low binding affinity between IFN-γrel and CRFB6, which could not be detected by Western blotting. It is worth noting that both CRFB13 and CRFB17 bound to CRFB6 in the absence of a ligand. Therefore, we are tempted to propose a model in which IFN-γrel could bind to a preformed ligand/receptor complex able to facilitate STAT1 phosphorylation and transduce cellular signals (Fig. 10).

The finding that IFN-γrel and IFN-γ activate the same receptor chains may imply a novel mechanism to fine-tune the functions of IFN-γ. We propose that IFN-γrel may serve as a mediator to balance the actions of IFN-γ by competing for its receptors, for instance, to divert the roles of IFN-γ to the subfunctions when needed. Binding affinities of ligands with the receptors may also play vital roles in controlling this process. In addition, the conformation changes of the receptors following binding with IFN-γ and IFN-γrel may impact on the activation of downstream signaling. As shown in the current study that the architectures of the IFN-γ and IFN-γrel differ significantly, therefore, solving the binary and ternary structures of IFN-γrel and its receptors will shed light on the mechanism on the receptor sharing by IFN-γ and IFN-γrel and their signaling pathways.

Although we were unable to obtain the crystal structure of ligand and receptor complex, we solved the crystal structure of IFN-γrel. IFN-γrel forms a dimer that has unique topology for dimer formation. Known cytokines forming functional dimers include IFN-γ and IL-22, all of which are members of IL-10 family. The dimeric structure of CiIFN-γrel is similar to that of IFN-γ. However, the topology of their monomers differs significantly due to the spatial location of helices A and B. Of note, helix A, AB loop, and helix B cannot be superposed. The CiIFN-γrel dimer requires helices A and B of one peptide chain to engage helices C’, D’, E’, and F’ of the other peptide chain to form a dimeric structure similar to the IFN-γ dimer, suggesting that IFN-γrel and IFN-γ interact with the receptors differently. It is also possible that subfunctionalization of IFN-γrel may be regulated at the receptor level, because it could activate three putative receptor complex (Fig. 10). Although we established the interaction relationships of IFN-γrel with individual receptor chains, the exact receptor complex warrants further investigation in the future. It cannot be ruled out that IFN-γrel is responsible for other functions that are distinct from that of IFN-γ.

Several putative residues of FN-γrel were predicted to be involved in the interaction with receptors. Mutagenesis analysis revealed that Asp128 is critical for the binding of IFN-γrel with CRFB13 and CRFB17 and is indispensable for its biological activity. This aspartic residue is located in the bent helix F. In human IFN-γ, the corresponding residue (Glu112) is negatively charged and has been shown to be critical for maintaining the structure of helix F and its interaction with IFN-γR1 (45). It is worth noting that the aspartic acid and glutamic acid are conserved in the IFN-γrels (except for Electrophorus electricus) and IFN-γ (Fig. 7), suggesting that it is essential for the interaction of IFN-γrels and IFN-γ with IFN-γR1 in jawed vertebrates and their shared functions. However, in fish, IFN-γR1 has been duplicated into CRFB13 and CRFB17, both of which can be activated by IFN-γrel (28). It is expected that the key residues in the contact interface of CiIFN-γrel with CRFB13 and CRFB17 will be different. We also observed that Gln23, Thr31, and Glu76 were involved in the interaction with CRFB13 but not with CRFB17. Of note, E76A resulted in partial loss of biological activity. Taken together, we propose that subfunctionalization of IFN-γrel from IFN-γ is closely associated with the structural adaption between duplicated ligands and receptors.

In summary, we found that CiIFN-γrel has been subfunctionalized from IFN-γ. Both CiIFN-γ and CiIFN-γrel were able to induce proinflammatory genes but unlike CiIFN-γ, CiIFN-γrel has lost the capability to activate genes involved in the Th1 response. CiIFN-γrel was shown to interact with CRFB17 and weakly with CRFB13, but not with CRFB6. The crystal structure of CiIFN-γrel showed a novel topology for cytokine dimer formation. Functional analysis reveals that Lys93 and Asp128 are required for STAT1 phosphorylation and biological activities.

We thank Dr. Mingxian Chang, Institute of Hydrobiology, Chinese Academy of Sciences, for providing the EPC cell line and the Shanghai Synchrotron Radiation Facility of China for protein crystallization service.

This work was supported by the National Natural Science Foundation of China (32030112 and U21A20268) and the Ministry of Science and Technology of the People’s Republic of China (2018YFD0900302).

The CiIFN-γrel structure presented in this article has been submitted to the Protein Data Bank (https://doi.org/10.2210/pdb7X45/pdb) under accession number 7X45.

The online version of this article contains supplemental material.

Abbreviations used in this article:

     
  • Ag

    antigen

  •  
  • BSA

    buried surface area

  •  
  • CHO

    Chinese hamster ovary

  •  
  • CiIFN-γrel

    Ctenopharyngodon idella IFN-γ–related

  •  
  • Co-IP

    coimmunoprecipitation

  •  
  • CRFB

    cytokine receptor family B

  •  
  • EPC

    epithelioma papulosum cyprini

  •  
  • Ha

    hemagglutinin

  •  
  • HEK

    human embryonic kidney

  •  
  • HKL

    head kidney leukocyte

  •  
  • IFN-γrel

    IFN-γ–related

  •  
  • nanoDSF

    nanoscale differential scanning fluorometry

  •  
  • NO

    nitric oxide

  •  
  • PDB

    Protein Data Bank

  •  
  • qPCR

    quantitative real-time PCR

  •  
  • RMSD

    root mean square deviation

  •  
  • SeMet

    selenomethionine

  •  
  • Tm

    melting temperature

  •  
  • WT

    wild-type

1.
Pestka
S.
2007
.
The interferons: 50 years after their discovery, there is much more to learn.
J. Biol. Chem.
282
:
20047
20051
.
2.
Savan
R.
,
S.
Ravichandran
,
J. R.
Collins
,
M.
Sakai
,
H. A.
Young
.
2009
.
Structural conservation of interferon gamma among vertebrates.
Cytokine Growth Factor Rev.
20
:
115
124
.
3.
Fu
J. P.
,
S. N.
Chen
,
P. F.
Zou
,
B.
Huang
,
Z.
Guo
,
L. B.
Zeng
,
Q. W.
Qin
,
P.
Nie
.
2014
.
IFN-γ in turtle: conservation in sequence and signalling and role in inhibiting iridovirus replication in Chinese soft-shelled turtle Pelodiscus sinensis.
Dev. Comp. Immunol.
43
:
87
95
.
4.
Schreiber
G.
2017
.
The molecular basis for differential type I interferon signaling.
J. Biol. Chem.
292
:
7285
7294
.
5.
Kotenko
S. V.
,
G.
Gallagher
,
V. V.
Baurin
,
A.
Lewis-Antes
,
M.
Shen
,
N. K.
Shah
,
J. A.
Langer
,
F.
Sheikh
,
H.
Dickensheets
,
R. P.
Donnelly
.
2003
.
IFN-lambdas mediate antiviral protection through a distinct class II cytokine receptor complex.
Nat. Immunol.
4
:
69
77
.
6.
Boehm
U.
,
T.
Klamp
,
M.
Groot
,
J. C.
Howard
.
1997
.
Cellular responses to interferon-gamma.
Annu. Rev. Immunol.
15
:
749
795
.
7.
Pestka
S.
,
S. V.
Kotenko
,
G.
Muthukumaran
,
L. S.
Izotova
,
J. R.
Cook
,
G.
Garotta
.
1997
.
The interferon gamma (IFN-gamma) receptor: a paradigm for the multichain cytokine receptor.
Cytokine Growth Factor Rev.
8
:
189
206
.
8.
Schroder
K.
,
P. J.
Hertzog
,
T.
Ravasi
,
D. A.
Hume
.
2004
.
Interferon-gamma: an overview of signals, mechanisms and functions.
J. Leukoc. Biol.
75
:
163
189
.
9.
Heim
M. H.
2000
.
Intracellular signalling and antiviral effects of interferons.
Dig. Liver Dis.
32
:
257
263
.
10.
Zou
J.
,
C. J.
Secombes
.
2011
.
Teleost fish interferons and their role in immunity.
Dev. Comp. Immunol.
35
:
1376
1387
.
11.
Krause
C. D.
2016
.
Intron loss in interferon genes follows a distinct set of stages, and may confer an evolutionary advantage.
Cytokine
83
:
193
205
.
12.
Robertsen
B.
,
V.
Bergan
,
T.
Røkenes
,
R.
Larsen
,
A.
Albuquerque
.
2003
.
Atlantic salmon interferon genes: cloning, sequence analysis, expression, and biological activity.
J. Interferon Cytokine Res.
23
:
601
612
.
13.
Zou
J.
,
A.
Carrington
,
B.
Collet
,
J. M.
Dijkstra
,
Y.
Yoshiura
,
N.
Bols
,
C.
Secombes
.
2005
.
Identification and bioactivities of IFN-gamma in rainbow trout Oncorhynchus mykiss: the first Th1-type cytokine characterized functionally in fish.
J. Immunol.
175
:
2484
2494
.
14.
Grayfer
L.
,
M.
Belosevic
.
2009
.
Molecular characterization, expression and functional analysis of goldfish (Carassius aurutus L.) interferon gamma.
Dev. Comp. Immunol.
33
:
235
246
.
15.
Yang
K.
,
S.
Zhang
,
D.
Chen
,
A.
Zhang
,
X.
Wang
,
H.
Zhou
.
2013
.
IFN-γ-activated lymphocytes boost nitric oxide production in grass carp monocytes/macrophages.
Fish Shellfish Immunol.
35
:
1635
1641
.
16.
Grayfer
L.
,
E. G.
Garcia
,
M.
Belosevic
.
2010
.
Comparison of macrophage antimicrobial responses induced by type II interferons of the goldfish (Carassius auratus L.).
J. Biol. Chem.
285
:
23537
23547
.
17.
Liao
Z.
,
Q.
Wan
,
J.
Su
.
2016
.
Bioinformatics analysis of organizational and expressional characterizations of the IFNs, IRFs and CRFBs in grass carp Ctenopharyngodon idella.
Dev. Comp. Immunol.
61
:
97
106
.
18.
Arts
J. A.
,
E. J.
Tijhaar
,
M.
Chadzinska
,
H. F.
Savelkoul
,
B. M.
Verburg-van Kemenade
.
2010
.
Functional analysis of carp interferon-gamma: evolutionary conservation of classical phagocyte activation.
Fish Shellfish Immunol.
29
:
793
802
.
19.
Platanias
L. C.
2005
.
Mechanisms of type-I- and type-II-interferon-mediated signalling.
Nat. Rev. Immunol.
5
:
375
386
.
20.
Bach
E. A.
,
M.
Aguet
,
R. D.
Schreiber
.
1997
.
The IFN gamma receptor: a paradigm for cytokine receptor signaling.
Annu. Rev. Immunol.
15
:
563
591
.
21.
Samuel
C. E.
2001
.
Antiviral actions of interferons.
Clin. Microbiol. Rev.
14
:
778
809
.
22.
Darnell
J. E.
 Jr.
,
I. M.
Kerr
,
G. R.
Stark
.
1994
.
Jak-STAT pathways and transcriptional activation in response to IFNs and other extracellular signaling proteins.
Science
264
:
1415
1421
.
23.
Jung
C. Y.
,
J.
Hikima
,
M.
Ohtani
,
H. B.
Jang
,
C. S.
del Castillo
,
S. W.
Nho
,
I. S.
Cha
,
S. B.
Park
,
T.
Aoki
,
T. S.
Jung
.
2012
.
Recombinant interferon-γ activates immune responses against Edwardsiella tarda infection in the olive flounder, Paralichthys olivaceus.
Fish Shellfish Immunol.
33
:
197
203
.
24.
Ruan
B. Y.
,
S. N.
Chen
,
J.
Hou
,
B.
Huang
,
Z. A.
Laghari
,
L.
Li
,
P.
Nie
.
2017
.
Two type II IFN members, IFN-γ and IFN-γ related (rel), regulate differentially IRF1 and IRF11 in zebrafish.
Fish Shellfish Immunol.
65
:
103
110
.
25.
Dehler
C. E.
,
K.
Lester
,
G.
Della Pelle
,
L.
Jouneau
,
A.
Houel
,
C.
Collins
,
T.
Dovgan
,
R.
Machat
,
J.
Zou
,
P.
Boudinot
, et al
2019
.
Viral resistance and IFN signaling in STAT2 knockout fish cells.
J. Immunol.
203
:
465
475
.
26.
Aggad
D.
,
C.
Stein
,
D.
Sieger
,
M.
Mazel
,
P.
Boudinot
,
P.
Herbomel
,
J. P.
Levraud
,
G.
Lutfalla
,
M.
Leptin
.
2010
.
In vivo analysis of Ifn-γ1 and Ifn-γ2 signaling in zebrafish.
J. Immunol.
185
:
6774
6782
.
27.
Luo
K.
,
S.
Zhang
,
D.
Tang
,
L.
Xia
,
W.
Gao
,
G.
Tian
,
Z.
Qi
,
Q.
Xu
,
W.
Zhang
.
2018
.
Analysis of the expression patterns of the cytokine receptor family B (CRFB) and interferon gamma receptor (IFNGR) in Dabry’s sturgeon (Acipenser dabryanus).
Dev. Comp. Immunol.
84
:
420
426
.
28.
Lu
D. Q.
,
T. T.
Leng
,
X.
Ding
,
W.
Peng
,
M.
Yao
,
S. S.
Li
,
H. R.
Lin
,
Y.
Zhang
.
2014
.
Two IFNGR1 homologues in Tetraodon nigroviridis: origin, expression analysis and ligand-binding preference.
Dev. Comp. Immunol.
44
:
270
279
.
29.
Gao
Q.
,
P.
Nie
,
K. D.
Thompson
,
A.
Adams
,
T.
Wang
,
C. J.
Secombes
,
J.
Zou
.
2009
.
The search for the IFN-gamma receptor in fish: functional and expression analysis of putative binding and signalling chains in rainbow trout Oncorhynchus mykiss.
Dev. Comp. Immunol.
33
:
920
931
.
30.
Ealick
S. E.
,
W. J.
Cook
,
S.
Vijay-Kumar
,
M.
Carson
,
T. L.
Nagabhushan
,
P. P.
Trotta
,
C. E.
Bugg
.
1991
.
Three-dimensional structure of recombinant human interferon-gamma.
Science
252
:
698
702
.
31.
Bennett
M. J.
,
M. P.
Schlunegger
,
D.
Eisenberg
.
1995
.
3D domain swapping: a mechanism for oligomer assembly.
Protein Sci.
4
:
2455
2468
.
32.
Mendoza
J. L.
,
N. K.
Escalante
,
K. M.
Jude
,
J.
Sotolongo Bellon
,
L.
Su
,
T. M.
Horton
,
N.
Tsutsumi
,
S. J.
Berardinelli
,
R. S.
Haltiwanger
,
J.
Piehler
, et al
2019
.
Structure of the IFNγ receptor complex guides design of biased agonists.
Nature
567
:
56
60
.
33.
Zahradník
J.
,
L.
Kolářová
,
H.
Pařízková
,
P.
Kolenko
,
B.
Schneider
.
2018
.
Interferons type II and their receptors R1 and R2 in fish species: evolution, structure, and function.
Fish Shellfish Immunol.
79
:
140
152
.
34.
Collaborative Computational Project, Number 4
.
1994
.
The CCP4 suite: programs for protein crystallography.
Acta Crystallogr. D Biol. Crystallogr.
50
:
760
763
.
35.
Wang
J.
,
W.
Wang
,
J.
Xu
,
Z.
Jia
,
Q.
Liu
,
X.
Zhu
,
C.
Xia
,
J.
Zou
.
2021
.
Structural insights into the co-evolution of IL-2 and its private receptor in fish.
Dev. Comp. Immunol.
115
:
103895
.
36.
Murshudov
G. N.
,
A. A.
Vagin
,
E. J.
Dodson
.
1997
.
Refinement of macromolecular structures by the maximum-likelihood method.
Acta Crystallogr. D Biol. Crystallogr.
53
:
240
255
.
37.
Adams
P. D.
,
R. W.
Grosse-Kunstleve
,
L. W.
Hung
,
T. R.
Ioerger
,
A. J.
McCoy
,
N. W.
Moriarty
,
R. J.
Read
,
J. C.
Sacchettini
,
N. K.
Sauter
,
T. C.
Terwilliger
.
2002
.
PHENIX: building new software for automated crystallographic structure determination.
Acta Crystallogr. D Biol. Crystallogr.
58
:
1948
1954
.
38.
Wang
W.
,
J.
Wang
,
L.
Lei
,
J.
Xu
,
Y.
Qin
,
Q.
Gao
,
J.
Zou
.
2020
.
Characterisation of IL-15 and IL-2Rβ in grass carp: IL-15 upregulates cytokines and transcription factors of type 1 immune response and NK cell activation.
Fish Shellfish Immunol.
107
(
Pt A
):
104
117
.
39.
Qin
Y.
,
Z.
Sun
,
W.
Wang
,
J.
Xu
,
B.
Wang
,
Z.
Jia
,
X.
Li
,
J.
Wang
,
Q.
Gao
,
X.
Chen
,
J.
Zou
.
2021
.
Characterization of CD3γ/δ+ cells in grass carp (Ctenopharyngodon idella).
Dev. Comp. Immunol.
114
:
103791
.
40.
Thiel
D. J.
,
M. H.
le Du
,
R. L.
Walter
,
A.
D’Arcy
,
C.
Chène
,
M.
Fountoulakis
,
G.
Garotta
,
F. K.
Winkler
,
S. E.
Ealick
.
2000
.
Observation of an unexpected third receptor molecule in the crystal structure of human interferon-gamma receptor complex.
Structure
8
:
927
936
.
41.
Stolte
E. H.
,
H. F.
Savelkoul
,
G.
Wiegertjes
,
G.
Flik
,
B. M.
Lidy Verburg-van Kemenade
.
2008
.
Differential expression of two interferon-gamma genes in common carp (Cyprinus carpio L.).
Dev. Comp. Immunol.
32
:
1467
1481
.
42.
Li
L.
,
S. N.
Chen
,
Z. A.
Laghari
,
B.
Huang
,
H. J.
Huo
,
N.
Li
,
P.
Nie
.
2019
.
Receptor complex and signalling pathway of the two type II IFNs, IFN-γ and IFN-γrel in mandarin fish or the so-called Chinese perch Siniperca chuatsi.
Dev. Comp. Immunol.
97
:
98
112
.
43.
Peng
W.
,
D. Q.
Lu
,
G. F.
Li
,
X.
Zhang
,
M.
Yao
,
Y.
Zhang
,
H. R.
Lin
.
2016
.
Two distinct interferon-γ genes in Tetraodon nigroviridis: functional analysis during Vibrio parahaemolyticus infection.
Mol. Immunol.
70
:
34
46
.
44.
Grayfer
L.
,
M.
Belosevic
.
2009
.
Molecular characterization of novel interferon gamma receptor 1 isoforms in zebrafish (Danio rerio) and goldfish (Carassius auratus L.).
Mol. Immunol.
46
:
3050
3059
.
45.
Walter
M. R.
,
W. T.
Windsor
,
T. L.
Nagabhushan
,
D. J.
Lundell
,
C. A.
Lunn
,
P. J.
Zauodny
,
S. K.
Narula
.
1995
.
Crystal structure of a complex between interferon-gamma and its soluble high-affinity receptor.
Nature
376
:
230
235
.

The authors have no financial conflicts of interest.

Supplementary data