DF-1 cells were purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA). DF-1 cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) (Gibco, USA) supplemented with 10% FBS and maintained at 37°C and 5% CO₂.

The ALV-J strain, SCAU-HN06, was a kind gift from Professor Ming Liao and Professor Weisheng Cao (College of Veterinary Medicine, South China Agricultural University). Briefly, DF-1 cells were infected with ALV-J train SCAU-HN06 (10⁵ TCID₅₀/0.1mL). After 2 h incubation, washed by 1×PBS, the media was replaced with DMEM and added 1% FBS. For certain experiments, cells were selected with a suitable inducer concentration for 2 h prior to ALV-J virus infection.

Flag-tagged CH25H plasmid was stored in our laboratory (17). The CHMP4B, TSG101, RPL23, PCBP2 were cloned into a PCAGGS vector with a C-terminal myc tag. PCR amplification primer sequences are as follows: chicken CHMP4B forward primer, 5’-ATGTCGGGGATCCTGGG-3’ and reverse primer, 5’-TTACATGTTTCCTGCCCAAG-3’; and chicken RPL23 forward primer, 5’-CGCCTCCGTCTCTTCCG-3’ and reverse primer 5’-TGCACAAAGATGAGCACGTT-3’. The vector of chicken TSG101 and chicken PCBP2 was synthesized by Wuhan Genecreate Biological Engineering Co., Ltd (Wuhan, China). Small interfering RNA (siRNA) targeting Chicken TSG101, CHMP4B, RPL23, PCBP2, and normal control (NC) siRNA (siN0000001-1-5) were synthesized by RiboBio (RiboBio Inc, China). The siRNA sequences are as follows:

DF-1 cells were transfected with these plasmids or siRNAs using Lipofectamine 3000 reagent, according to the manufacturer’s instructions.

25HC was purchased from Sigma-Aldrich. Caspase-1 activity assay kit was purchased from Beyotime Institute of Biotechnology. Lipofectamine 3000 was purchased from Invitrogen. Annexin V-FITC and PI apoptosis detection kits were purchased from Keygen Biotech Company (Nanjing, China). Antibodies used in the present study were as follows: ALV-J envelope protein JE9-specific mouse monoclonal antibody (provided by Professor Qin Aijian of Yangzhou University), rabbit anti-LC3II antibody (NB100-2220SS, Novus Biologicals), rabbit anti-ATG5 antibody (12994, Cellsignal), mouse monoclonal anti-flag antibody (AP007M, Bioworld), rabbit anti- SQSTM1/p62 antibody (P0067, Sigma-Aldrich), and anti-myc antibody (A02060, Abbkine), rabbit anti-β-actin antibody (AF5003, Beyotime), goat anti-rabbit IgG (H+L) antibody (A21020, Abbkine), Goat anti-mouse IgG/HRP antibody (A25112, Abbkine).

Total RNA was extracted from DF-1 cells at the indicated time points using TRIzol reagent (TaKaRa, Kusatsu, Japan) and reversely transcribed to cDNA using a PrimeScript RT Reagent kit (TaKaRa, Kusatsu, Japan). qPCR analysis was performed on a CFX96 PCR system (Bio-Rad) using SYBR green fast mixture (Bio-Rad, Hercules, CA, USA). Expression levels were quantified using the 2 ^−ΔΔC(T) method and normalized to GAPDH. Primers used for RT-qPCR are listed in Supplementary Table 1.

Caspase-1 activity assay kits (C1102, Beyotime Biotechnology, Shanghai, China) were used to analyze Caspase-1 activities, following the manufacturer’s instructions. According to a previous study, the expression levels of ALV group-specific antigen (p27) were analyzed (17).

Apoptosis were analysed by double staining with Annexin V-FITC (fluorescein isothiocyanate) and propidium iodide (PI). Annexin V binds to phosphatidylserine on the outer plasma membrane of apoptotic cells, while PI labels necrotic cells with membrane damage. The prepared cells were gently washed three times with pre-chilled PBS. Subsequently, 500 μL Annexin V binding buffer was added, and the cells were mixed gently. Cell suspensions were then incubated with 5 μL FITC Annexin V and 5 μL propidium iodide (PI) for 15 min at room temperature in the dark. Then, the percentage of viable cells (Annexin V-/PI-), early apoptotic cells (Annexin V+/PI-), and late apoptotic or necrotic cells (Annexin V+/PI+) were analyzed by flow cytometry (BD AccuriC6, Biosciences, USA), and data was processed using FlowJo7.6 software.

DF-1 cells were cultured in 24-well PET cell culture inserts. Flag-tagged CH25H plasmid and control plasmid were transfected into DF-1 cells. After 24 h, DF-1 cells were fixed with 4% paraformaldehyde, permeabilized with 2% Triton-X-100, blocked with blocking solution (P0252, Beyotime) for 20 min, and incubated with primary antibody anti-Flag overnight at 4°C. Secondary antibody FITC-labeled Goat Anti-Mouse (A0568, Beyotime) was added, and the nuclei were stained with DAPI (P0131, Beyotime). Imaging was performed with a Leica TCS-SP5 confocal microscope with a ×40 objective.

The DF-1 cells in a six-well culture plate were divided into five treatment groups: DF-1 cells without any treatment (NC group), DF-1 cells infected with ALV-J train SCAU-HN06 (10⁵ TCID₅₀/0.1mL) (ALV-J group), DF-1 cells transfected with the plasmid pflag-EGFP following by ALV-J (10⁵ TCID₅₀/0.1mL) infection (EGFP group), DF-1 cells transfected with the plasmid pflag-CH25H following by ALV-J (10⁵ TCID₅₀/0.1mL) infection (CH25H group), and DF-1 cells pretreated with 1 μM 25HC for 12 h (17) following by ALV-J (10⁵ TCID₅₀/0.1mL) infection (25HC group). Cells were harvested at 48 h after ALV-J infection and were fixed in 2.5% glutaraldehyde (Solarbio, China) at 4°C for 6 h. Resin impregnation, adding neat resin and polymerizing at 65°C. The intracellular structure of the stained cells was observed with a transmission electron microscope (TEM) (Hitachi HT7800).

After treatment, total protein was extracted from DF-1 cells and quantified by the Bradford BSA Protein Assay Kit (Beijing Dingguo Changsheng Biotechnology, China). Proteins were separated by SDS-polyacrylamide gel electrophoresis (4-15% SDS-PAGE) and were wet-transferred onto PVDF membrane (Bio-Rad) for immunoblotting. Membranes were incubated with the primary antibody (1:1000) at 4°C overnight. Protein bands were visualized with an Odyssey FC infrared imaging system and analyzed with Image Studio (LICOR Biosciences, NE, USA).

Transfected DF-1 cells were washed three times with pre-cooled PBS, lysed with IP lysate buffer, and extracted with total protein buffer (Thermo Fisher Scientific), supplemented with 2% PMSF (ST507, Beyotime), then centrifuge at 13,000 rpm for 10 minutes at 4°C. 50 μL Protein G magnetic beads (Thermo Fisher Scientific) were conjugated with 5 µg anti-Flag and were gently mixed on a rotator mixer at 4°C for 2 h. Subsequently, the pretreated supernatants were further incubated with anti-Flag antibody-conjugated magnetic beads on a rotator mixer overnight at 4°C. Finally, the bound proteins were eluted for SDS‐PAGE. The proteins pulled‐down by IP were subjected to mass spectrometry analysis performed by Shenzhen Wininnovate Bio Co., Ltd (Shenzhen, China).

According to the signal peptide and transmembrane analysis, we constructed pBT3-STE-CH25H as a bait vector and pPR3-N-CHMP4B as a prey vector, while combinations of pTSU2-APP and pNubG-Fe65 were used as positive controls, combinations of pTSU2-APP and pPR3-N were used as negative controls. Combinations of pBT3-STE-CH25H and pPR3-N, pBT3-STE-CH25H and pPR3-N-CHMP4B, pBT3-STE-CH25H and POST-NubaI, pTSU2-APP and pNubG-Fe65, pTSU2-APP and PPR3-N were co-transformed into the yeast strain NMY51. Different co-transformed yeast cells were carried out ten-fold serial dilution, and 4 ten-fold series dilutions of each co-transformant were dropped synthetic dropout (SD) medium lacking Leu and Trp and SD medium lacking Ade, His, Leu, and Trp with 30 mm 3AT. Photographs were taken after 5 d of incubation.

All experiments were independently repeated at least twice, and each experiment included three replicates. Statistical analysis and graph generation were performed using GraphPad Prism 5 (GraphPad Software Inc., USA). All statistical data are presented as the mean ± SEM, and significance was assessed using two-way ANOVA. Differences with P-values < 0.05 (*), 0.01 (**), and 0.001 (***) were statistically significant.

Apoptosis is the most common form of programmed cell death (20). Apoptosis was measured in DF-1 cells infected with ALV-J to identify whether ALV-J infection induces activation of apoptosis. Flow cytometry results showed that DF-1 cells undergo apoptosis at 24 h or 48 h of ALV-J infection (Figures 1A-C). Moreover, compared to normal control (NC) cells, the mRNA expression of apoptosis-related genes including Bax, Cyt c, Caspase-3, and Caspase-9 were significantly upregulated in DF-1 cells infected with ALV-J, while Bak, and Bcl-2 were not significantly different (Figure 1D). These results suggested that ALV-J infection induced DF-1 cells apoptosis.

Programmed cell death (PCD) or regulated cell death (RCD) plays an important role in cancers as well as in diseases (21). Our previous study had proven that expression levels of the CH25H could be significantly induced in the immune system organs of clinical chickens and ALV-J infected cells (17, 22), which suggested that the host may exerts an antiviral function by inducing the expression of CH25H. To further explore the ability of CH25H to resist ALV-J virus, DF-1 cells were transfected with pflag-CH25H for 24 h prior to being infected with ALV-J. The result of qRT-PCR (Supplementary Figure 1A), ELISA (Supplementary Figure 1B) and immunoblot (Supplementary Figure 1C) analyses showed that overexpression of CH25H can inhibit the viral titer of 10⁵ TCID₅₀/0.1mL ALV-J. While CH25H exerts its anti-ALV-J function, it also induces host cell death (Supplementary Figure 1D), which suggested that CH25H may be closely related to cell apoptosis. Thus, we hypothesized that CH25H restricts ALV-J infection by regulating host apoptosis.

DF-1 cells overexpressing CH25H were detected by flow cytometry to test this hypothesis. Flow cytometry results showed that overexpression of CH25H significantly inhibited early apoptosis, but the percentage of late apoptotic or necrotic cells were significantly increased (Figures 2A-C). And,we found that overexpressing CH25H inhibited early apoptosis response in ALV-J infected DF-1 cells, while late apoptotic or necrotic cells were significantly increased (Figures 2D-F). Next, key apoptosis-related genes including Bax, Cyt c, Caspase-3, Caspase-9, Bak, and Bcl-2 were examined. Notably, the genes encoding the proapoptotic factors, Bax, Bak, Cyt c, Caspase-3, and Caspase-9, were not significantly different, while the anti-apoptotic marker Bcl-2 was significantly up-regulated in DF-1 cells overexpressing CH25H, as compared to vector control (Figure 2G). Besides, DF-1 cells were transfected with CH25H expression plasmid followed by ALV-J infection, and those encoding the proapoptotic factors, Bak, Cyt c, Caspase-3, and Caspase-9, were not significantly different, while the anti-apoptotic marker Bcl-2 was significantly up-regulated, as compared to ALV-J infected vector control (Figure 2H). Therefore, overexpression of CH25H inhibited apoptosis and increased late apoptosis or necrosis, and CH25H did not exert its anti-ALV-J effect through apoptosis.

Pyroptosis is an inflammatory cell death dependent upon the inflammatory caspases that plays a crucial role in defense against microbial infections (21, 23). In order to explore whether CH25H restricts ALV-J infection by regulating pyroptosis, we then examined the expression levels of caspase-1 and inflammatory factors IL-1β, IL-18, NLRP3. As a result, we found that CH25H overexpression elevates levels of IL-1β, IL-18 and NLRP3 but not caspase-1 (Supplementary Figure 2), suggesting that no pyroptosis occurred. These results suggested that overexpression of CH25H does not induce pyroptosis.

Autophagy is considered a mechanism of programmed cell death and is an evolutionarily conserved self-eating phenomenon and has multiple effects on antiviral innate immunity (24–26). A previous study showed that ALV-J inhibits autophagy and promotes viral self-replication (12). We therefore further hypothesized that CH25H inhibits ALV-J by promoting autophagy. We overexpressed CH25H in DF-1 cells to verify this hypothesis, and then autophagy-related proteins were analyzed by immunoblot. According to the fact that rapamycin (Rapa) is an autophagy inducer, a concentration of 1 μM Rapa (13) was used to analyze the expression levels of microtubule-associated protein 1 light chain 3 II (LC3II), autophagy-related gene 5(ATG5), and SQSTM1/p62 proteins (p62). Detection of LC3 protein levels by immunoblotting is one of the most common methods to measure autophagy, and comparing LC3II levels between samples is an appropriate method to determine autophagy (27). ATG5, an essential factor for proper autophagosome formation and maturation (28), was significantly increased. p62, which is a substrate of autophagy, was significantly downregulated.

As expected, CH25H overexpression significantly increased the LC3II level compared with the empty vector (Figures 3A, B). In addition, immunoblot analysis showed that the levels of LC3II and ATG5 were significantly promoted, while p62 was significantly decreased, at 48 h post-infection (hpi), with overexpressing CH25H before ALV-J virus infection (Figures 3C, D).

DF-1 cells were pretreated with 25HC to explore whether 25HC induces autophagy, and the immunoblot assays for LC3II, ATG5, and p62 were performed to detect autophagy. The results showed that the expression levels of LC3II and ATG5 were significantly promoted, while p62 were significantly decreased, compared with normal control (NC) cells (Figures 3E, F). Meanwhile, in ALV-J-infected DF-1cells, 25HC pretreatment significantly increased the expression levels of LC3II and ATG5, but significantly inhibited the expression level of p62 (Figures 3G, H). Altogether, these results indicated that both the overexpression of CH25H and the treatment with 25HC promoted autophagy.

Based on the above results, we speculated that CH25H-induced autophagy acts as a defense mechanism to inhibit ALV-J replication. To test this hypothesis, DF-1 cells were pretreated with Rapa for 12 h prior to being infected with ALV-J for 6 and 24 h. qRT-PCR and ELISA analyses showed that the expression level of ALV-J gp85 (Figure 4A) and ALV-J p27 (Figure 4B) were significantly decreased in Rapa-treated cells compared to Rapa-untreated (mock) cells. At the same time, autophagy is also detected by immunoblot analysis with an anti-LC3B antibody. As expected, ALV-J infection inhibits autophagy (Figures 4C, D). Next, autophagosomes were further determined by TEM. Consistently, autophagosomes were present in both CH25H transfected and 25HC treated cells, while a comparative few autophagosomes were observed for ALV-J-infected and control vector-transfected cells (Figure 4E). Altogether, these data suggested that CH25H-induced autophagy is a host defense mechanism to inhibit ALV-J viruses.

A separate study proposed that CH25H likely acts through cholesterol-independent mechanisms that inhibit viral replication (19). Thus, we hypothesized that CH25H inhibits ALV-J replication through interacting with proteins in the autophagy pathway. To prove this hypothesis, we performed IP-MS analysis and a two-hybrid assay to identify CH25H-interacting proteins. The flag-tagged CH25H expression plasmid and the control plasmid were transfected into DF-1 cells for 24 h and infected with the ALV-J virus. Confocal microscopy analysis showed that exogenous CH25H proteins were located in the cytoplasm in DF-1 cells (Figure 5A), which prompted us to consider that CH25H-interacting proteins are localized in the cytoplasm. Subsequently, we screened for possible CH25H target proteins by Co-IP (Figure 5B) combined with IP-MS experiments. About 65 unique proteins were identified in CH25H overexpression cells followed by ALV-J infection, compared with the EGFP overexpression control group (Supplementary Table 2), in which 10 interacting proteins were localized to the cytoplasm (Supplementary Figure 3A). Referring to the candidate proteins function annotation, candidate proteins ontology classification, candidate proteins biological process analysis, molecular functions analysis, and KEGG pathway analysis shown in Supplementary Figures 3B-D and Supplementary Table 2, 4 proteins (Figure 5C) were expected as candidate CH25H-interacting proteins.

The overexpression vectors of selected candidate proteins were overexpressed in DF-1 cells followed by ALV-J infection. The overexpression efficiencies were shown as Supplementary Figures 4A-D. qRT-PCR analyses results showed that overexpression of TSG101 had not significant different effect on ALV-J gp85 expression (Figure 5D); however, ALV-J gp85 was significantly decreased in cells overexpressing CHMP4B at 48 h (Figure 5E) but were not significantly different in cells overexpressing of RPL23 (Figure 5F) and PCBP2 (Figure 5G), compared to empty vector control. Conversely, silencing of TSG101 inhibits the expression of ALV-J gp85 at 48 hpi (Figure 5H), and silencing of CHMP4B promotes the expression of ALV-J gp85 at 48 hpi (Figure 5I); silencing of RPL23 (Figure 5J) and PCBP2 (Figure 5K) had not significant different effect on ALV-J gp85 expression. The interference efficiencies were shown as Supplementary (Figure 4E–H). Therefore, we speculated that CH25H might interact with CHMP4B and inhibited ALVJ replication.

Next, the interaction between CH25H and CHMP4B was verified by yeast two-hybrid assays. As expected, yeast two-hybrid assay results showed that CH25H interacted with CHMP4B (Figure 6A). Moreover, co-transfection of CH25H and CHMP4B significantly decreased the protein expression level of ALV-J env (Figures 6B, C), and significantly reduced the mRNA level of ALV-J gp85 (Figure 6D), at 48 hpi. Finally, whether co-overexpressing of CH25H and CHMP4B affected autophagy was investigated. The results clearly demonstrated that LC3II and ATG5 protein levels were significantly increased, whereas p62 expression levels were significantly decreased, in CH25H and CHMP4B co-overexpressing DF-1 cells infected with ALV-J. Altogether, the above data support that CH25H binds to CHMP4B and inhibits ALV-J replication by promoting autophagy.