RNF216 Inhibits the Replication of H5N1 Avian Influenza Virus and Regulates the RIG-I Signaling Pathway in Ducks

Waterfowl (e.g., duck, goose) are the natural reservoir for avian influenza virus (AIV). Ducks are infected with some highly pathogenic AIVs without showing clinical symptoms. However, the mechanism of resistance to AIV infection in ducks is still unknown (1, 2). The genome of AIV comprises eight negative-sense, single-stranded RNA segments, which encode ten essential proteins and over eight accessory proteins (3, 4). Its polymerase genes (PB2, PB1, PA) are key regulators of AIV replication (5, 6). Moreover, these viral proteins interact with host proteins to facilitate viral replication or regulate host innate immune responses (7, 8).

When host cells are invaded by viruses, intrinsic pattern-recognition receptors sense their pathogen-associated molecular patterns to trigger innate immune responses. Based on the homology of protein structural domains, pattern-recognition receptors can be classified as TLRs, nucleotide-binding oligomerization domain-containing proteins, retinoic acid-inducible gene-I–like receptors (RLRs), and cytosolic DNA sensors (9, 10). The RLRs contain retinoic acid-inducible gene I (RIG-I), melanoma differentiation-associated protein 5 (MDA5), and laboratory of genetics and physiology 2 (LGP2). RIG-I and MDA5 recognize cytoplasmic viral RNA, and then their caspase activation and recruitment domains (CARDs) bind to the CARDs of mitochondrial antiviral signaling protein (MAVS) to activate MAVS (11). Subsequently, MAVS recruits TNF receptor-associated factor 3 (TRAF3) (12) via its TRAF-interacting motifs (TIMs) (13, 14). TRAF3 activates downstream TANK-binding kinase 1 (TBK1)/IκB kinase ε (IKKε), which in turn phosphorylates the transcription factors IFN regulatory factor (IRF) 3/7 to induce type I IFN (IFN-I) and the proinflammatory cytokines, ultimately resulting in the antiviral innate immune response (15). However, IFN-I must be strictly regulated to avoid its overproduction causing harm to the host (16).

The RIG-I signaling pathway is modulated by a variety of post-translational modifications, such as ubiquitination, phosphorylation, methylation, acetylation, and SUMOylation (17). The RING finger (RNF) family, a group of E3 ubiquitin ligases containing RING domains, catalyzes the ubiquitination of substrate proteins in mammals (8). Notably, the RNF family is extensively involved in regulating innate immune responses by ubiquitination (8, 18). RNF114 has been shown to positively and negatively regulate the IFN-I pathway (19–22). RNF114 promotes the K48-linked polyubiquitination and proteasomal degradation of TRAF3 to negatively regulate the RIG-I pathway (22). RNF116 catalyzes polyubiquitination of TRAF3 to positively regulate the RIG-I pathway (23). RNF216 promotes the K48-linked polyubiquitination and proteasomal degradation of TRAF3 to negatively regulate the RIG-I pathway (24). However, whether duck RNF216 (duRNF216) also regulates the IFN-I pathway in ducks remains unclear.

Host RIG-I recognizes viral RNA and then activates the RIG-I pathway, ultimately inducing IFN-I expression. Nevertheless, several viral proteins can interact with the key proteins of the RIG-I pathway, resulting in the regulation of IFN-I production. The VP2 protein of human bocavirus physically associates with RNF125 to abrogate RNF125-mediated RIG-I degradation, promoting IFN-I production (25). The NS3-4A protease of hepatitis C virus specifically targets RNF135 to block RIG-I activation, eventually suppressing IFN-I production (26). The cysteine protease 3Cpro of Coxsackievirus B3 cleaves MAVS to disrupt the RIG-I pathway (27). However, whether avian influenza virus proteins interact with duRNF216 to regulate the RIG-I pathway in ducks remains unknown.

In this article, we describe the function of the duRNF216 protein in ducks. Our data reveal that duRNF216 is an E3 ubiquitin ligase containing two RING domains. duRNF216 interacted with duTRAF3 but not with duRIG-I, duMDA5, duMAVS, duSTING, duTBK1, or duIRF7 in the duck RIG-I pathway. duRNF216 promotes the K48-linked polyubiquitination and proteasomal degradation of duTRAF3 to inhibit the RIG-I pathway in ducks. However, when host cells are infected by avian influenza virus, duRNF216 interacts with the AIV PB1 protein and damages polymerase activity to inhibit viral replication in the nucleus. Moreover, duRNF216 interacts with the AIV PB1 protein to weaken the downregulation of the RIG-I pathway by attenuating duTRAF3 polyubiquitination in ducks.

HEK293T, HeLa, and primary duck embryonic fibroblasts (DEFs) produced from 9-d-old Muscovy duck embryos were cultured in DMEM (Life Technologies) containing 10% FBS, 100 U/ml penicillin, and 100 g/ml streptomycin at 37°C and 5% CO₂. A/duck/Guangdong/212/2004 (H5N1) virus (DK212) is maintained in our laboratory. All virus experiments were performed in biosafety level 3 (BSL-3) facilities in compliance with approved protocols by the Biosafety Committee of South China Agricultural University.

Total RNA was isolated from DEFs using Vazyme Total RNA extraction reagent, and the RNA was then reverse-transcribed using Moloney murine leukemia virus reverse transcriptase (Promega, USA) to generate duck cDNA. Primers were created and used to amplify the full-length duRNF216 gene from cDNA in accordance with the predicted sequence of duck-RNF216 (duRNF216) from the National Center for Biotechnology Information (accession number XM 027469163.1, https://www.ncbi.nlm.nih.gov/nuccore/XM_027469163.1/). The GeneJET gel extraction kit (Thermo Scientific, USA) was used to purify the PCR products, which were then sent to be sequenced (Tsingke, China). The nucleotide sequence of duRNF216 is available from NCBI GenBank under the accession number OP598816 (https://www.ncbi.nlm.nih.gov/nuccore/OP598816). The different species RNF216 accession numbers are human (NP_001364085.1, https://www.ncbi.nlm.nih.gov/protein/NP_001364085.1), mouse (NP_542128.2, https://www.ncbi.nlm.nih.gov/protein/NP_542128.2), pig (XP_020941710.1, https://www.ncbi.nlm.nih.gov/protein/XP_020941710.1), cattle (NP_001180187.1, https://www.ncbi.nlm.nih.gov/protein/NP_001180187.1), chicken (XP_015149855.2, https://www.ncbi.nlm.nih.gov/protein/XP_015149855.2), zebrafish (XP_687871.3, https://www.ncbi.nlm.nih.gov/protein/XP_687871.3), and grouper (XP_033500116.1, https://www.ncbi.nlm.nih.gov/protein/XP_033500116.1). The RNF216 sequences of different species were aligned using SnapGene 4.2.4 (Insightful Science, USA) and Clustal Ω multiple alignment. MEGA 11 software (Sinauer Associates, USA) was used to conduct phylogenetic analysis of RNF216 genes with the neighbor-joining method. Similarity analysis of RNF216 was performed using Lasergene 7.1 (DNASTAR, USA). The functional domains of duRNF216 were predicted using the Simple Modular Architecture Research Tool (SMART, http://smart.embl.de/).

The coding segments of duck RIG-I (duRIG-I), duck MDA5 (duMDA5), and duck IRF7 (duIRF7) were inserted into the EcoRI/SacI sites of pCAGGS with a C-terminal V5 tag. The eukaryotic expression plasmids pCAGGS-duMAVS-V5, pCAGGS-duSTING-V5, pCAGGS-duTRAF3-V5, pCAGGS-duTBK1-V5, pCAGGS-duIRF7-V5, pCAGGS-DK212PB2-HA, pCAGGS-DK212PB1-HA, pCAGGS-DK212PA-HA, pCAGGS-DK212HA-HA, pCAGGS-DK212NP-HA, pCAGGS-DK212NA-HA, pCAGGS-DK212M1-HA, pCAGGS-DK212M2-V5, pCAGGS-DK212NS1-HA, and pCAGGS-DK212PB1-F2-HA were maintained in our laboratory. pCAGGS-duRIG-I(N)-V5 contains the N-terminal 187 amino acids (aa) of duRIG-I. The duRNF216 gene and its truncations were cloned into the pCAGGS-Flag-tagged vector by the ClonExpress II one-step cloning kit (Vazyme, China) (duRNF216-Flag, duRNF216(N)-Flag: 1–574 aa, and duRNF216(C)-Flag: 575–929 aa). The N-terminal (1–418 aa) and C-terminal (419–567 aa) fragments of duTRAF3 were inserted into the EcoRI/SacI sites of pCAGGS with the C-terminal V5 tag and named duTRAF3(N)-V5 and duTRAF3(C)-V5, respectively. At the TIM motif of duRNF216, S442D mutations were introduced by overlap extension PCR, and Y440A Q442A mutations were introduced into duTRAF3-V5. The mutated DNA fragments were cloned into the pCAGGS-Flag/V5-tagged vector and named duRNF216(S326D)-Flag and duTRAF3(Y440A Q442A)-V5. For construction of the fluorescence fusion protein, eGFP and AsRed were fused by a linker (GGGGSGGGGSGGGGS) to duTRAF3 (duTRAF3-eGFP) and duRNF216 (duRNF216-AsRed), respectively. pRK5-HA-tagged ubiquitin plasmids (wild type, KO, K6, K11, K27, K29, K33, K48, K63, K6R, K11R, K27R, K29R, K33R, K48R and K63R) were purchased from Addgene.

The human polymerase (pol) I promoter (225 bp) and the murine terminator (33 bp) sequences were cloned from the plasmid pHW2000 (28–30). The human RNA polymerase I promoter, the 3′-untranslated region of the influenza A virus NP gene, the firefly luciferase gene, the 5′-untranslated region of the influenza A virus NP gene, and the murine terminator were sequentially cloned into the pUC vector to construct the human pol I vNP-firefly luciferase reporter plasmid. The chicken pol I vNP-firefly luciferase reporter plasmid was constructed by replacing the human pol I promoter with the chicken RNA pol I promoter (GenBank accession number DQ112354). The report plasmid contains a polymerase I promoter to drive transcription of a viral-like RNA. Following the transfection of the plasmid encoding viral-like RNA, subsequent transfection introduces the viral polymerase in trans, initiating the amplification and expression of the minigenome. Measurement of the reporter gene product, such as luciferase, provides an estimate of polymerase activity (31–34).

The Abs were mainly purchased from Invitrogen, Sigma, Sino Biological, Abcam, and LC-COR. Major Abs used included HA tag mAb (26183, Invitrogen), HA tag polyclonal Ab (71–5500, Invitrogen), monoclonal anti-Myc Ab (MA1-980, Invitrogen), monoclonal anti-Flag Ab (F1804, Σ), anti-V5 tag Ab (ab27671, Abcam), V5 tag rabbit mAb (13202S, CST), GAPDH mAb (MA5-15738, Invitrogen), monoclonal anti-β-actin Ab (A5316, Σ), influenza A PB2 polyclonal Ab (PA5-32221, Invitrogen), influenza A PB1 polyclonal Ab (PA5-34914, Invitrogen), influenza A PA polyclonal Ab (PA5-32223, Invitrogen), influenza A nucleoprotein/NP Ab (Sino Biological, 11675-MM03T), IRDye 800CW goat anti-rabbit IgG (92632211, LC-COR), and IRDye 800CW goat anti-mouse IgG (92632210, LC-COR).

Avian IFN-β and IRF7 promoter plasmids (pGL3-IFN-β-Luc, pGL3-IRF7-Luc) and pRL-TK (Promega, E2241) reference plasmid were maintained in our laboratory (35). To determine whether duRNF216 affected IFN-β/IRF7 promoter activity in ducks, DEFs were transfected with the pRL-TK reference plasmid and pGL3-IFN-β/IRF7-Luc promoter plasmids, pCAGGS (empty vector [EV]) or expression plasmids encoding duRNF216. After 16 h, the DEFs were transfected with the polyinosinic–polycytidylic acid [poly(I:C)] (5 μg/ml) or 5′ppp-dsRNA (500 ng/ml). After 24 h, luciferase assays were performed.

To determine whether duRNF216 affected IFN-β promoter activity in ducks, DEFs were transfected with the pRL-TK reference plasmid and pGL3-IFN-β-Luc promoter plasmids, pCAGGS (EV) or expression plasmids encoding duRIG-I(N), duMAVS, duTBK1, duIRF7, and duRNF216. Luciferase assays were performed 24 h post-transfection.

To determine whether duRNF216 affected IRF7 promoter activity in ducks, DEFs were cotransfected with the pRL-TK reference plasmid, pGL3-IRF7-Luc promoter plasmids, duRIG-I(N) and duRNF216, or pCAGGS (EV). Luciferase assays were performed 24 h post-transfection.

To determine whether duRNF216 affected AIV polymerase activity, the pRL-TK (Promega, E2241) reference plasmid (10 ng), human pol I vNP-firefly luciferase reporter plasmid (100 ng), pCAGGS-PA (100 ng), pCAGGS-PB1 (100 ng), pCAGGS-PB2 (100 ng), pCAGGS-NP (200 ng), and pCAGGS-duRNF216-Flag, or pCAGGS (EV) were cotransfected into HEK293T in a 24-well plate. The pRL-TK reference plasmid (10 ng), chicken pol I vNP-firefly luciferase reporter plasmid (100 ng), pCAGGS-PA (100 ng), pCAGGS-PB1 (100 ng), pCAGGS-PB2 (100 ng), pCAGGS-NP (200 ng), and pCAGGS-duRNF216-Flag, or pCAGGS (EV) were cotransfected into DEFs in a 24-well plate. Luciferase assays were performed 24 h post-transfection. The luciferase activity was assayed using a dual luciferase assay instrument (Promega, USA) according to the manufacturer’s operating instructions. The results are presented as the normalized luciferase activity (the ratios of firefly/Renilla luminescences). Influenza A virus polymerase activity luciferase reporter assays were performed according the previous studies (31–34).

IFN-β and IRF7 luciferase reporter assays were performed using previously described methods (35). The luciferase activity was assayed using a dual luciferase assay instrument (Promega, USA) according to the manufacturer’s operating instructions.

To determine whether overexpression of duRNF216 altered the mRNA levels of IFN-β, IFN-α, Mx, OAS, and PKR, DEFs were transfected with pCAGGS (EV) or expression plasmids encoding duRIG-I(N), duMAVS, duTBK1, and duIRF7, as well as the duRNF216 expression plasmid as indicated. The relative mRNA levels of IFN-β were measured by qPCR at 24 h post-transfection. Duck β-actin was used to normalize qPCR in the tested group. qPCR procedures and data processing were performed according to our previous study (36).

To determine whether duRNF216 affected AIV RNA synthesis in infected cells, DEFs were transfected with pCAGGS (EV). After 24 h of transfection, the DEFs were infected with a dose (MOI = 1) of the H5N1 virus DK212. Total RNA was extracted from infected DEFs and reverse transcribed with specific reverse transcription primers (Table I), and qPCR was used to measure the RNA levels of DK212 NP segment mRNA, cRNA, and vRNA after 12 h of infection (37, 38). Duck IFN-β expression levels in DEF culture supernatants were determined using kits from Mlbio (Shanghai, China) following the manufacturer’s instructions. Three biological replicates were set for each group in each experiment. All experiments were repeated at least three times with consistent results.

In our previous study, a coimmunoprecipitation assay was performed in HEK293T cell systems (39). HEK293T cells growing in 100-mm dishes were cotransfected with duRNF216-Flag and duTRAF3-V5, duRNF216-Flag and EV, or EV and duTRAF3-V5. After 24 h of transfection, the cells were lysed at 2–8°C and centrifuged at 10,000–13,000g for 10 min. The lysate supernatants were incubated with anti-Flag M2 agarose beads (A2220, Σ), anti-HA agarose beads (26182, Thermo), or anti-V5 agarose beads (A7345, Σ) for 4–8 h at 4°C. To determine whether duTRAF3 ubiquitination could be detected, HEK293T cells were cotransfected with duTRAF3-V5, Ub-HA, and duRNF216-Flag. Cell lysates were incubated with anti-V5 agarose beads. The supernatant was discarded after centrifugation, and the beads were washed three to five times with lysis buffer before being resuspended in SDS loading buffer. Elution of coimmunoprecipitates was bathed in boiling water for 5 min, and the samples were analyzed by Western blotting. All experiments were repeated at least three times with consistent results.

HeLa cells were grown on cover slips in 24-well plates. The cells were fixed in 4% paraformaldehyde at 25°C for 30 min after experimental treatment. Then, the cell membrane structure was permeabilized with 0.3% Triton X-100 in PBS at 25°C for 30 min. The cell nuclei were visualized with 4,6-diamidino-2-phenylindole (DAPI; Dingguo) and visualized. The fluorescence intensity profile was measured along the line drawn by Olympus, with a scale bar of 20 μm.

For knockdown of duRNF216 genes in DEFs, three shRNAs targeting duRNF216 were synthesized (Genepharma, China), and the target sequences were as follows: shduRNF216-1, GGGCACTACATGAACTCAAGG; shduRNF216-2, GGAGCAGTATCAGAAGGATGG; and shduRNF216-3, GCAAGGAGTGCCTAATTAAAT. The vectors containing shduRNF216 or shNC (negative control) were transfected into DEFs with Lipofectamine 2000. After 24 h, Western blot analysis and qPCR assays were used to measure the knockdown efficiency of duRNF216.

To determine whether knockdown of duRNF216 genes leads to a decrease in the mRNA expression level of duRNF216 in ducks, shduRNF216-3 was transfected into DEFs. After 24 h of transfection, qPCR was conducted to detect the transcription levels of duRNF216 compared with the shNC group in DEFs.

To determine whether duRNF216 deficiency altered the mRNA production of IFN-β in ducks, the shduRNF216-3 plasmid was transfected into DEFs along with duRIG-I(N). After 24 h of transfection, qPCR was used to measure the transcription levels of IFN-β.

In the antiviral experiments, cells were grown on 12-well plates until they reached 85% confluence, and expression plasmid (duRNF216-Flag) or EVs were transfected into DEFs. After 24 h post-transfection, DEFs were infected with H5N1 AIV (DK212) at a dose of MOI = 0.001. Cell culture supernatants and cell lysates were collected 12, 24, 36, and 48 h postinfection, a 50% tissue culture-infective dose (TCID₅₀) assay was used to determine the virus titer in the supernatant, and Western blot analysis was used to determine the level of viral NP protein expression in the cell lysates. Knockdown of duRNF216 for determination of viral growth curves and polymerase activity was based on overexpression of duRNF216 in DEFs.

The data are presented as the means ± SD, and multiple comparisons were conducted using an unpaired t test and ANOVA in GraphPad Prism 9 software. *p < 0.05; **p < 0.01; ***p < 0.001; and ****p < 0.0001 were significant. “ns” indicates no significant difference.

The duRNF216 gene CDS is 2,790 bp, encodes 929 amino acids, and contains a TIM domain [PxQx(T/S)], two RING domains and an “in-between RING” domain (Fig. 1A). The amino acid sequence of duRNF216 shared 90.0% similarity with those of chicken RNF216, 72.5% with human RNF216, 72.6% with mouse RNF216, 46.3% with grouper RNF216, and 49.9% with zebrafish (Fig. 1B). Phylogenetic analysis showed that duRNF216 was clustered into the bird clade (Fig. 1C).

RIG-I senses intracellular nonself RNA to undergo configurational change, and then its CARD domain binds to the CARD domain of MAVS to activate MAVS on the mitochondrial membrane. Activated MAVS recruits TRAF3 to form the RIG-I–MAVS–TRAF3 complex, and then TRAF3 induces kinase complexes, including TBK1 and IKKε. Through several phosphorylation steps, kinases ultimately induce IFN-I expression (40, 41). In mammals, RNF216 targets TRAF3 and inhibits the activation of the RIG-I pathway (24). To investigate whether duRNF216 affected the duck RIG-I pathway, duRNF216 was transfected with IFN-β or IRF7 promoter reporter plasmids into DEFs. Dual-luciferase assay results showed that duRNF216 obviously inhibited IFN-β and IRF7 promoter activity, which was triggered by the duRIG-I CARD domain [aa 1–187, duRIG-I(N)], 5′ppp-dsRNA, and poly(I:C), respectively (p < 0.01), but duRNF216 cannot completely eliminate the upregulation of IFN-β and IRF7 promoter activity induced by duRIG-I(N), 5′ppp-dsRNA, and poly(I:C) (Fig. 2A–F). In addition, qPCR assays showed that duRNF216 also significantly downregulated the mRNA levels of IFN-β, IFN-α, Mx, OAS, and PKR induced by duRIG-I(N) (p < 0.0001), but duRNF216 cannot completely eliminate the upregulation of the IFN-I and ISG genes expression induced by duRIG-I(N) (Fig. 2G–K).

To further confirm whether duRNF216 influences the production of IFN-β, three shRNA plasmids (shduRNF216-1, shduRNF216-2 and shduRNF216-3) were transfected into cells to knock down duRNF216. Western blot results indicated that shduRNF216-3 targeting duRNF216 obviously downregulated the protein expression level of duRNF216-Flag compared with shduRNF216-1 or shduRNF216-2 in HEK293T cells (Fig. 3A). In addition, three shRNAs were transfected into DEFs. qPCR assays indicated that shduRNF216-3 significantly decreased the mRNA expression level of duRNF216 compared with that in the shNC group in DEFs (Fig. 3B). Furthermore, the shduRNF216-3 plasmid was transfected into DEFs along with duRIG-I(N). Knockdown of duRNF216 obviously promoted the expression of IFN-β in DEFs (p < 0.05) (Fig. 3C). Therefore, our results reveal that duRNF216 inhibits the RIG-I pathway in ducks.

To examine whether duRNF216 regulates the RIG-I pathway, duMAVS, duTBK1, and duIRF7 were coexpressed with duRNF216 in DEFs together with an IFN-β promoter plasmid. The results showed that duRNF216 obviously inhibited IFN-β promoter activity induced by duMAVS (p < 0.001) but not by duTBK1 or duIRF7 (Fig. 4A). Meanwhile, to determine whether duRNF216 altered the production of IFN-β, duMAVS, duTBK1, and duIRF7 were coexpressed with duRNF216 in DEFs. qPCR assays showed that duRNF216 significantly downregulated the mRNA expression of IFN-β induced by duMAVS (p < 0.05) but not that induced by duTBK1 and duIRF7 (Fig. 4B). To further identify the molecules of the duck RIG-I signaling pathway that are targeted by duRNF216, duRNF216 was cotransfected with duRIG-I, duMDA5, duMAVS, duSTING, duTRAF3, duTBK1, and duIRF7 in HEK293T cells. Coimmunoprecipitation (co-IP) results showed that duRNF216 interacted with duTRAF3 but not with duRIG-I, duMDA5, duMAVS, duSTING, duTBK1, or duIRF7 (Fig. 4C, 4D). Meanwhile, to study the localization of duRNF216 and duTRAF3, we transfected duRNF216-AsRed and duTRAF3-eGFP into HeLa cells. We observed that duRNF216 was colocalized with duTRAF3 in the cytoplasm by confocal microscopy (Fig. 4E). These results showed that duRNF216 targeted duTRAF3. Additionally, MAVS recruits TRAF3 proteins via its TIMs and initiates signal transduction in mammals (13, 14). To determine whether duRNF216 inhibits duMAVS to recruit duTRAF3, duTRAF3 and duMAVS were transfected with duRNF216 in HEK293T cells. The results demonstrated that duRNF216 inhibited duMAVS to recruit duTRAF3 in a dose-dependent manner (Fig. 4F). Therefore, duRNF216 targets duTRAF3 and inhibits duMAVS to recruit duTRAF3 to negatively regulate the duRIG-I pathway.

The TRAF family performs different functions by recognizing and binding to various structural motifs in mammals. One of the conserved TIMs is PxQx(T/S) (13). We identified an N-terminal motif of duRNF216, PAQGS, which is one conserved motif of TIM (Fig. 5A). To detect the interaction region of duRNF216 and duTRAF3, duTRAF3 was transfected with duRNF216, duRNF216(N) (1–574 aa), and duRNF216(C) (575–929 aa) in HEK293T cells. We observed that duTRAF3 interacted with duRNF216 and duRNF216(N) but not with duRNF216(C) (Fig. 5B). Meanwhile, duRNF216 was transfected with duTRAF3, duTRAF3(N) (1–418 aa), and duTRAF3(C) (419–567 aa) in HEK293T cells. We found that duRNF216 interacted with duTRAF3 and duTRAF3 (C) but not with duTRAF3 (N) (Fig. 5C). These results showed that duRNF216 interacted with duTRAF3 via its TIM domain and TRAF domain of duTRAF3.

Previous studies showed that human RNF216 S263 and TRAF3 Y441/Q443 were the binding sites between RNF216 and TRAF3. Sequence alignment showed that duRNF216 322-PAQGS-326 and duTRAF3 439-LYSQP-443 were homologous to human RNF216 259-PMQES-263 and TRAF3 440-LYSQP-444, respectively. To examine whether duRNF216 S326D and duTRAF3 (Y440A/Q442A) affected the interaction between duRNF216 and duTRAF3, duTRAF3 was transfected with duRNF216 and duRNF216 (S326D) into HEK293T cells. We found that duTRAF3 bound duRNF216 but not duRNF216 (S326D) (Fig. 5D). In addition, duRNF216 was transfected with duTRAF3 and duTRAF3 (Y440A/Q442A) in HEK293T cells. Our results illustrated that duRNF216 interacted with duTRAF3 but not with duTRAF3 (Y440A/Q442A) (Fig. 5E). Thus, we demonstrate that duck RNF216 S326 and TRAF3 Y440/Q442 are the binding sites between duRNF216 and duTRAF3.

The RNF proteins are a family of E3 ubiquitin ligases that catalyze the ubiquitination of a wide range of target proteins (8, 42). Considering that RNF216, as an RNF protein, targets TRAF3 in mammals, we investigated whether duRNF216 catalyzes duTRAF3 ubiquitination. We transfected duRNF216, duTRAF3, and wild-type or KO, K6, K11, K27, K29, K33, K48, and K63 Ub plasmids into HEK293T cells to perform ubiquitination assays. The co-IP results showed that duRNF216 mediated duTRAF3 ubiquitination (Fig. 6A, lane 4). Furthermore, duRNF216 promoted K48-linked polyubiquitination of duTRAF3 (Fig. 6B, lane 4; Fig. 6D, lane 8), whereas duRNF216 was unable to ubiquitinate duTRAF3 in the presence of K48R and KO Ub mutants (Fig. 6E, lanes 2, 8). In contrast, duRNF216 did not affect the K63-linked polyubiquitination of duTRAF3 (Fig. 6C, lane 4).

Previous studies found that human RNF216 polyubiquitinated TRAF3, leading to its proteasomal degradation (24). To examine whether duRNF216 promotes the ubiquitination and degradation of duTRAF3, duTRAF3 was transfected with duRNF216 in DEFs. The results indicated that duRNF216 promoted the degradation of duTRAF3, but duTRAF3 degradation mediated by duRNF216 was obviously inhibited in cells treated with the proteasomal inhibitor MG132 and bortezomib (PS-341). In addition, duRNF216-mediated degradation of duTRAF3 was not obviously inhibited by the autophagic–lysosomal pathway inhibitors 3-methyladenine, chloroquine, and bafilomycin A1 (Fig. 6F). Therefore, our results reveal that duRNF216 catalyzes K48-linked polyubiquitination and degrades duTRAF3 by the proteasome.

To determine whether duRNF216 affects the replication of H5N1 avian influenza viruses, H5N1 virus DK212–infected DEFs overexpressing duRNF216 were used. The results showed that the virus titer at all time points was significantly lower in DEFs overexpressing duRNF216 than in the control group (p < 0.01) (Fig. 7A). Meanwhile, in contrast to the control group, overexpression of duRNF216 decreased the expression of NP protein, as shown by Western blot analysis. Moreover, the antiviral activity of duRNF216 increased in a dose-dependent manner (p < 0.01) (Fig. 7B). To investigate whether duRNF216 inhibits the polymerase function of H5N1 subtype AIV to affect viral replication, duRNF216-Flag, PB2, PB1, PA, and NP plasmids of H5N1 virus DK212, polymerase reporter plasmid (hu/chpolI-NP-Luc), and internal reference plasmid (pRL-TK) were transfected into HEK293T (or DEF) cells. A dual-luciferase reporter assay indicated that overexpression of duRNF216 in HEK293T or DEFs apparently suppressed the H5N1 AIV polymerase activity (p < 0.05) (Fig. 7C, 7D). To examine whether duRNF216 affects the transcription and synthesis of viral RNAs, DEFs transfected with duRNF216-Flag were infected with the H5N1 virus DK212. We found that overexpression of duRNF216 in DEFs suppressed AIV vRNA, mRNA, and cRNA synthesis (Fig. 7E). In addition, knockdown of the duRNF216 gene resulted in an increase in polymerase activity (p < 0.05) in DEFs (Fig. 7H). Thus, duRNF216 protein strongly restricts viral replication by inhibiting viral polymerase activity and viral RNA synthesis.

We found that duRNF216 to some extent inhibited IFN-β production, but duRNF216 cannot completely eliminate the upregulation of the IFN-β induced by the activators (Fig. 2). In this article, we further revealed that the overexpression of duRNF216 led to a reduction in the expression of IFN-β in AIV-infected DEFs, as detected by ELISA and qRT-PCR (p < 0.01). However, the expression level of IFN-β in AIV-infected DEFs overexpressing duRNF216 was still higher than that in uninfected control cells, which would also hinder the viral replication (Fig. 7F, 7G).

To determine whether knockdown of duRNF216 affects the replication of H5N1 avian influenza viruses, H5N1 virus DK212–infected DEFs knocking down duRNF216. As depicted in Fig. 7I, using shRNA (shRNF216-3) to knock down duRNF216 resulted in an increase in virus titer at 24, 36, and 48 hours postinfection (hpi) compared with the control group (shNC) in DEFs (p < 0.01). Further investigation revealed that knockdown of duRNF216 led to an increase in the expression of IFN-β in AIV-infected DEFs, as detected by ELISA and qRT-PCR at 12, 24, 36, and 48 hpi (p < 0.01) (Fig. 7J, 7K). Notably, the maximum decrease of viral titer in DEFs was ∼5-fold lgTCID₅₀ by overexpressing duRNF216 (Fig. 7A), whereas the maximum increase of viral titer was only ∼2-fold lgTCID₅₀ by knocking down duRNF216 (Fig. 7I). The possible reason was that when DEFs were infected with avian influenza virus, viral RNA activated the duRIG-I pathway to significantly upregulate IFN-β expression (Fig. 7F, 7G, 7J, 7K). Moreover, although duRNF216 downregulated the IFN-β expression during virus infection, the expression level of IFN-β in AIV-infected DEFs overexpressing duRNF216 was still higher than that in uninfected control cells, which would also hinder the viral replication (Fig. 7F, 7G, 7J, 7K).

Therefore, our results reveal that duRNF216 significantly inhibits the replication of H5N1 avian influenza virus. However, the expression level of IFN-β in AIV-infected DEFs overexpressing duRNF216 was still higher than that in uninfected control cells, which would also hinder the viral replication.

To further investigate whether the RNF216 protein targets viral proteins to inhibit viral replication, duRNF216 was cotransfected with the expression plasmids of H5N1 AIV proteins (PB2, PB1, PA, NP, HA, NA, M1, M2, NS1, and PB1-F2) in HEK293T cells. We found that duRNF216 interacted with the PB1 protein but not with other viral proteins (Fig. 8A–J). The duRNF216-Flag was transfected into DEFs; 12 h post-transfection, H5N1 virus DK212 (MOI = 0.001) infected the DEFs. After 24 h of infection, the lysed DEFs were incubated with anti-Flag agarose beads. Western blot was then used to examine the coprecipitation of PB1 protein and duRNF216 protein. We found that the duRNF216 protein interacted with the H5N1 virus PB1 protein in the infected DEFs (Fig. 9A). To investigate the interaction region between duRNF216 and the H5N1 virus PB1, we constructed expression plasmids of PB1 protein fragments [pCAGGS-PB1(1–266)-V5, pCAGGS-PB1(267–492)-V5, and pCAGGS-PB1(493–757)-V5]. The expression plasmids of PB1 protein fragments were transfected with the duRNF216-Flag plasmid into HEK293T cells. The results showed that the H5N1 virus PB1(1–266) interacted with duRNF216 (Fig. 9B). Further results showed that duRNF216 interacted with the H5N1 virus PB1 protein, and duRNF216 colocalized with PB1 in the nucleus (Fig. 9C). duRNF216 affected the interaction of PB1 and PA, which suggested that duRNF216 influenced the assembly of AIV polymerase (Fig. 9D). However, there was no ubiquitination or degradation of the PB1 protein mediated by RNF216 (Fig. 9E–G). Thus, duRNF216 interacts with the viral PB1 protein to inhibit the replication of avian influenza virus.

Furthermore, we analyzed whether the PB1 protein of H5N1 virus DK212 affected the regulation of the duRIG-I pathway mediated by duRNF216. The dual-luciferase assay results showed that duRNF216 inhibited IFN-β and IRF7 promoter activity, which was triggered by duRIG-I(N), 5′ppp-dsRNA, and poly(I:C), but the H5N1 AIV PB1 protein slightly attenuated the inhibitory effect. Notably, the promoter activity induced by the activators alone was higher than that induced by cotransfection with activator and duRNF216 or activator, duRNF216, and PB1 (Fig. 10A–F). The qPCR results showed that duRNF216 inhibited IFN‐I and ISGs production, which was triggered by duRIG-I(N), but the H5N1 virus PB1 protein slightly weakened the inhibitory effect. Notably, the expression levels of IFN-I, Mx, OAS, and PKR induced by duRIG-I(N) alone were higher than those induced by cotransfection with duRIG-I(N) and duRNF216 or duRIG-I(N), duRNF216, and PB1 (Fig. 10G–K). Further studies showed that the H5N1 virus PB1 protein competed with duTRAF3 for binding to duRNF216, thereby hindering the ubiquitination of duTRAF3 mediated by duRNF216 (Fig. 11A). To test whether the PB1 protein affected duTRAF3 degradation mediated by duRNF216, PB1, duRNF216, and duTRAF3 were transfected into DEFs. The results indicated that duRNF216 promoted the degradation of duTRAF3, but the duTRAF3 degradation was to a certain extent inhibited by PB1 (Fig. 11B). Consequently, H5N1 virus PB1 protein competes with duTRAF3 for binding to duRNF216, thereby slightly attenuating the negative regulation of the innate immunity pathway mediated by duRNF216.

The RNF family is present in various organisms from animals to plants. The RNF family, a group of E3 ubiquitin ligases containing one or two RING finger domains, plays multiple roles in innate immunity (8, 18). Based on phylogenetic analysis, the RNF family is grouped into five different subfamilies, including tripartite motif-containing, PA-TMRING, RING between RING, membrane-associated RING-CH, and RING-Ub interacting motif families (8). The results from mammalian studies show that certain RNF family proteins of E3 ubiquitin ligases interact with key proteins in the RIG-I pathway to regulate the production of IFN-I. It is still unclear whether duck RNF proteins regulate the production of IFN-I in ducks. We determined the full-length duRNF216 gene from Muscovy ducks. Phylogenetic analysis revealed that duRNF216 was clustered into the bird clade and shared 90.0% similarity in amino acid sequence with chicken RNF216 but shared only 71.4–73.2% similarity with mammalian RNF216. In addition, duRNF216, similar to human RNF216, contains a TIM domain (PxQxT/S), two RING domains, and an “in-between RING” domain, which suggests that duRNF216 and mammalian RNF216 may possess similar functions.

Certain host proteins interact with key proteins of the RIG-I pathway and modify or degrade these proteins to regulate this pathway, ultimately regulating the expression of IFN-I. The RNF family, as a group of E3 ubiquitin ligases, catalyzes the ubiquitination of immune-related proteins to regulate innate immunity in mammals (8). RNF135 catalyzes K63-linked polyubiquitination of RIG-I to effectively evoke innate immune responses against RNA virus infection (43). RNF166 binds to TRAF3 and TRAF6 and catalyzes their ubiquitination, thus promoting IFN-β production (23). RNF128 targets TBK1 and catalyzes K63-linked polyubiquitination of TBK1, which promotes TBK1 kinase activity to induce IFN-β production (44). RNF125 promotes K48-linked polyubiquitination and proteasomal degradation of RIG-I to inhibit the RIG-I signaling pathway (45). RNF39 cripples RLR signaling by catalyzing K48-linked polyubiquitination and proteasomal degradation of DDX3X (46). RNF5 catalyzes the K48-linked polyubiquitination of MAVS at K362 and K461 and promotes the degradation of MAVS in proteasome-dependent pathways (47). Thus, some RNF proteins positively and negatively regulate the IFN-I signaling pathway by targeting critical proteins in mammals. In this study, we found that duRNF216 repressed the expression of IFN-I in the duRIG-I pathway. duRNF216 did not interact with duRIG-I, duMDA5, duMAVS, duSTING, duTBK1, or duIRF7 in the duck RIG-I pathway. duRNF216 targeted duTRAF3 and inhibited duMAVS to recruit duTRAF3 in a dose-dependent manner. Furthermore, duRNF216 promoted K48-linked polyubiquitination and proteasomal degradation of duTRAF3 to negatively regulate the expression of IFN-I. Notably, when DEFs were infected with avian influenza virus, viral RNA activated the duRIG-I pathway to significantly upregulate IFN-I expression (Fig. 7F, 7G, 7J, 7K). Moreover, although duRNF216 downregulated the IFN-I expression during virus infection, the expression level of IFN-β in AIV-infected DEFs overexpressing duRNF216 was still higher than that in uninfected cells, which would hinder the viral replication (Fig. 7F, 7G). Additionally, AIV PB1 protein competed with duTRAF3 for binding to duRNF216 to reduce degradation of TRAF3 by proteasomes in the cytoplasm, thereby slightly weakening duRNF216-mediated downregulation of IFN-I, which would slightly upregulate the expression of IFN-I in the PB1 and duRNF26-cotransfected DEFs compared with those in the duRNF216-transfected cells (Fig. 10, 11).

Many studies have shown that certain RIG-I pathway regulatory proteins directly interact with viral proteins to inhibit viral replication. For instance, the negative regulatory protein RNF5 of the RIG-I pathway restricts SARS-CoV-2 replication by targeting its envelope (E) protein for degradation (47, 48). RNF178 (also known as MARCH8) negatively regulates innate immunity and directly degrades the S and M proteins of SARS-CoV-2 and the M2 protein of influenza viruses to impair viral infectivity (49–52). In addition, studies have found that MARCH8 has a wide range of antiviral activity by specifically inactivating different viral fusion proteins, including Ebola virus, HIV-1, and H5N1 avian influenza virus (53). RNF153 (also known as MARCH5) negatively regulates the RIG-I pathway and directly degrades the HBx protein of HBV to alleviate HBV-mediated liver disease (54, 55). The negative regulatory factor RNF125 of the RIG-I pathway reduces viral titer by degrading host factors involved in HIV-1 transcription (45, 56). RNF114 acts as an inhibitor of Porcine Reproductive and Respiratory Syndrome Virus replication by degrading virus Nsp12 (57). Our results showed that duRNF216 protein interacted with the H5N1 AIV protein PB1, but not PB2, PA, NP, HA, NA, M1, M2, NS1, or PB1-F2 (Fig. 8). During AIV infection, duRNF216 protein targeted the core protein PB1 of viral polymerase to hinder viral polymerase assembly and viral RNA synthesis, which strongly decreased viral polymerase activity in the nucleus, ultimately restricting viral replication. In contrast, although duRNF216 downregulated the IFN-β expression during virus infection, the expression level of IFN-β in AIV-infected DEFs overexpressing duRNF216 was still higher than that in uninfected cells, which would also hinder the viral replication (Fig. 7F, 7G). Overall, duRNF216 strongly inhibits the replication of avian influenza virus in DEFs.

In summary, transfection with duRIG-I(N), 5′ppp-dsRNA, and poly(I:C) activated duRIG-I in DEF cells, leading to a strong upregulation of IFN-β expression. Simultaneously, duRNF216 targeted duTRAF3 and dose-dependently inhibited duMAVS recruitment to duTRAF3. Notably, duRNF216 induced partial degradation of duTRAF3, exerting a negative regulatory effect on the duRIG-I pathway. However, it was observed that duRNF216 could not completely eliminate the upregulation of IFN-I induced by these activators. Upon AIV infection in DEF cells, viral RNA activated duRIG-I, resulting in a robust upregulation of IFN-β expression. Moreover, duRNF216 targeted duTRAF3, causing partial degradation of duTRAF3 and a slightly negative regulation of the duRIG-I pathway. Despite duRNF216 downregulating IFN-β expression during virus infection, the IFN-β level in AIV-infected DEFs overexpressing duRNF216 remained higher than that in uninfected cells, contributing to the hindrance of viral replication. Furthermore, the AIV PB1 protein engaged in competition with duTRAF3 for binding to duRNF216, leading to a reduction in TRAF3 degradation by proteasomes. Importantly, during AIV infection, duRNF216 directly targeted the AIV PB1 protein to strongly decrease viral polymerase activity in the nucleus, thereby restricting viral replication in DEFs. Thus, these findings expand our knowledge of the mechanisms by which duRNF216 influences AIV replication in ducks (as shown in the Graphical Abstract).

The authors have no financial conflicts of interest.

The authors thank the Key Laboratory of Animal Vaccine Development, Ministry of Agriculture and Rural Affairs.