Porcine epidemic diarrhea virus suppresses DHX9-mediated antiviral transcription through its nucleocapsid protein

Abstract

Porcine epidemic diarrhea virus (PEDV) causes an enteric disease in piglets and leads to economic losses in the swine industry. Antiviral defense depends on interferon responses, yet how PEDV suppresses antiviral transcription remains incompletely understood. We show that DExH-box helicase 9 (DHX9) functions as a host restriction factor that is antagonized by the PEDV nucleocapsid (N) protein. PEDV infection decreases DHX9 protein abundance without affecting DHX9 transcript levels. DHX9 overexpression restricts PEDV replication, whereas DHX9 depletion enhances viral replication. Mechanistically, DHX9 forms a transcriptional complex with signal transducer and activator of transcription 1 (STAT1) and RNA polymerase II to promote interferon-stimulated gene (ISG) expression. The PEDV N protein binds the STAT1-interacting region of DHX9 and disrupts assembly of the DHX9–STAT1–RNA polymerase II complex, suppressing antiviral transcription. In parallel, the N protein promotes ubiquitination and proteasome-dependent degradation of DHX9 by enhancing its interaction with the E3 ubiquitin ligase MDM2, attenuating STAT1-associated antiviral signaling. Together, these findings identify DHX9 as a regulator of antiviral transcription during PEDV infection and reveal how PEDV suppresses antiviral responses through disruption and degradation of DHX9. This work provides insights into coronavirus immune evasion and suggests that the PEDV N–DHX9 axis may represent a target for antiviral strategies.

Porcine epidemic diarrhea virus suppresses DHX9-dependent antiviral transcription through its nucleocapsid protein by disrupting the DHX9–STAT1–RNA polymerase II complex and promoting MDM2-mediated DHX9 degradation to enhance viral replication.

Introduction

Porcine epidemic diarrhea (PED), caused by the porcine epidemic diarrhea virus (PEDV), is an acute, highly contagious intestinal disease¹. PED primarily affects piglets within 10 days of age, characterized by acute enteritis, severe diarrhea, vomiting, and dehydration, with mortality rates reaching up to 90% in infected neonates². This pathogen imposes substantial threats to global swine production systems, especially in intensive farming operations, leading to substantial economic repercussions³. PEDV is an enveloped, single-stranded positive-sense RNA virus with a genome of ~28 kb^4,5. The viral genome encodes two large polyproteins, pp1a and pp1b, from ORF1a and ORF1b, which are subsequently cleaved into 16 nonstructural proteins (nsp1–nsp16) involved in viral genome replication, transcription, and polyprotein processing⁶. The structural proteins—spike (S), envelope (E), membrane (M), and nucleocapsid (N)—are encoded by ORF2 and ORF4–6, respectively.⁷. In addition, the accessory protein ORF3, encoded by ORF3, is a critical virulence factor for PEDV⁸. To facilitate viral proliferation, PEDV has developed sophisticated strategies to subvert host antiviral defenses. Studies indicate that PEDV interacts with host factors to evade immune surveillance and achieve immune escape⁹. For instance, PEDV Nsp1 directly targets NOD-like receptor family CARD domain containing 5 (NLRC5), suppressing virus-mediated MHC-I expression and selectively blocking NLRC5 mRNA translation, thereby disguising infected cells as healthy ones to evade CD8⁺ T cell and NK cell cytotoxicity^10. Nsp2 exhibits dual regulatory effects on host antiviral immunity by targeting F-box/WD repeat-containing protein 7 (FBXW7)—enhancing the expression of retinoic acid-inducible gene I (RIG-I), TANK-binding kinase 1 (TBK1), and interferon-stimulated genes (ISGs) to promote immune responses while simultaneously inducing FBXW7 degradation to suppress immune activity¹¹. Moreover, the PEDV N protein participates in viral genome encapsidation while competitively associating with tripartite motif containing 25 (TRIM25), thereby disrupting the TRIM25-ZAP interaction and compromising E3 ubiquitin ligase activity, ultimately subverting ZAP-dependent antiviral mechanisms¹². The PEDV N protein also interacts with tripartite motif containing 28 (TRIM28) to inhibit the JAK-STAT1 signaling transduction, thereby facilitating PEDV replication¹³. Nevertheless, the mechanisms by which PEDV evades host innate immunity remain to be further elucidated. In particular, how PEDV interfaces with host nuclear transcriptional regulators to dampen interferon-stimulated gene (ISG) expression remains unclear.

To address this gap, we examined whether PEDV targets DExH-box helicase 9 (DHX9), a multifunctional DExD/H-box RNA helicase implicated in antiviral signaling and gene transcription. DHX9, a member of the DExD/H-box RNA helicase family, is a multifunctional enzyme involved in transcriptional regulation, RNA metabolism, splicing, ribosome biogenesis, RNA transport, miRNA processing, and translation¹⁴. Due to its versatility, DHX9 plays a dual role in viral infections: it can either be exploited by viruses to facilitate replication or contribute to host antiviral defense. For example, DHX9 enhances HIV-1 mRNA translation via its ATP-dependent helicase activity and promotes viral particle assembly to increase infectivity^15,16. Similarly, myxoma virus (MYXV) M029 protein interacts with protein kinase R (PKR) and DHX9 to support viral replication¹⁷, while classical swine fever virus recruits DHX9 through its 5′ and 3′ UTRs to enhance replication¹⁸. DHX9 also promotes the replication of influenza A virus, hepatitis C virus, hepatitis B virus, and foot-and-mouth disease virus^19–22. Conversely, DHX9 exhibits antiviral activity in certain contexts. During MYXV infection, cytoplasmic DHX9 forms antiviral granules that suppress late viral protein synthesis, restricting MYXV and another poxvirus replication²³. DHX9 plays a pivotal role in innate immunity. In myeloid dendritic cells, DHX9 acts as a mitochondrial antiviral-signaling protein (MAVS)-dependent sensor for viral double-stranded RNA (dsRNA). Through its dsRNA-binding motif, DHX9 interacts with the CARD domain of MAVS, activating nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) and interferon regulatory factor 3 (IRF3) to initiate antiviral responses²⁴. In plasmacytoid dendritic cells (pDCs), DHX9 serves as a CpG-B DNA sensor, triggering MyD88-dependent production of interferon-alpha (IFN-α) and tumor necrosis factor-alpha (TNF-α). DHX9 knockdown significantly reduces CpG-B-induced IFN-α levels and impairs pDC responses to DNA viruses²⁵. Additionally, DHX9 functions as a co-transcriptional activator by forming a multiprotein complex with NF-κB p65 and RNA polymerase II (RNA Pol II) via its ATPase/helicase domain, enhancing NF-κB-dependent transcription of antiviral genes²⁶. The pro- or antiviral role of DHX9 likely depends on its subcellular localization and interaction partners. For instance, nuclear DHX9 recruits RNA Pol II to amplify the transcription of ISGs, whereas cytoplasmic DHX9 may be co-opted by viruses to aid replication²⁷. To our knowledge, the role of DHX9 in PEDV infection is unknown. These observations prompted us to investigate whether DHX9 participates in PEDV immune evasion.

In this study, we demonstrate that DHX9 acts as a crucial mediator of innate immunity against PEDV, and PEDV infection downregulates DHX9 to impair STAT1-associated antiviral transcription and promote viral replication. Mechanistically, the PEDV N protein suppresses host antiviral responses not only by competitively inhibiting the DHX9-STAT1-Pol II complex formation but also by promoting MDM2-mediated degradation of DHX9. This study elucidates an immune evasion mechanism employed by PEDV, wherein the virus targets DHX9 to suppress antiviral responses and enhance viral replication.

Results

PEDV infection decreases the expression of DHX9 in cells

To explore the impact of PEDV infection on DHX9 expression, Vero and IPEC-J2 cells were infected with PEDV at an MOI of 1. The mRNA level of DHX9 was measured using RT-qPCR, and the protein level of DHX9 was assessed by western blotting. The results showed that DHX9 mRNA level remained unchanged, but DHX9 protein level decreased significantly in the cells at 24 h and 48 h post-infection (Fig. 1A, B). The findings indicate that PEDV infection results in a decrease in DHX9 protein levels in cells, which is not mediated at the transcriptional level.

Fig. 1. DHX9 protein level is decreased upon PEDV infection.

A, B The mRNA and protein levels of DHX9 were respectively detected by RT-qPCR and western blotting in Vero and IPEC-J2 cells infected with PEDV (MOI = 1) for the indicated time. The bottom panels show semi-quantitative analyses of the protein bands. Bar graphs represent the mean of three independent experiments (±SD). *P < 0.05, **P < 0.01, and ***P < 0.001.

DHX9 inhibits PEDV replication

Given PEDV-induced post-transcriptional reduction of DHX9, we next asked whether DHX9 influences PEDV replication. For this, a specific siRNA (siDHX9) was employed to knock down DHX9, while the pCAGGS-DHX9-HA plasmid was used to overexpress DHX9 in Vero and IPEC-J2 cells. Cells transfected with siNC (non-specific siRNA control) or pCAGGS-HA (empty vector) served as controls. RT-qPCR analysis revealed that siDHX9 reduced DHX9 mRNA levels by 55–60% in both cell lines, compared to the siNC group (Fig. 2A). Notably, DHX9 silencing did not significantly affect cell viability in either Vero or IPEC-J2 cells, indicating that the subsequent antiviral phenotypes were unlikely attributed to the cytotoxicity of DHX9 silencing (Fig. 2B). Subsequently, cells with DHX9 knockdown or overexpression were infected with PEDV at an MOI of 1, and the expression levels of PEDV S and N proteins were analyzed at 24 h and 48 h post-infection. Western blotting results demonstrated that DHX9 knockdown significantly increased the expression of PEDV S and N proteins, whereas DHX9 overexpression markedly reduced their expression (Fig. 2C, D). In line with these findings, viral titer assays showed that DHX9 knockdown significantly enhanced PEDV titers, while DHX9 overexpression had the opposite effect (Fig. 2E, F). Collectively, these results indicate that DHX9 plays a critical role in inhibiting PEDV replication in cells.

Fig. 2. DHX9 inhibits PEDV replication.

A Validation of siDHX9 by RT-qPCR in Vero and IPEC-J2 cells. B Cell viability of Vero and IPEC-J2 cells after siNC or siDHX9 transfection. C, D Vero and IPEC-J2 cells were transfected with siNC (control) or siDHX9 for 24 h (C), or with pCAGGS-HA or pCAGGS-DHX9-HA for 24 h (D), and then infected with PEDV at MOI = 1. At 24 or 48 h post-infection, the cells were harvested, and the whole cell lysates were subjected to western blotting with the indicated antibodies. Semi-quantitative analyses of the protein bands are shown in the right panels. E, F Released viral titers were detected using a TCID₅₀ assay. Bar graphs represent the mean of three independent experiments (±SD). *P < 0.05 and **P < 0.01.

DHX9 promotes ISG transcription to inhibit PEDV replication

Since DHX9 has been linked to interferon signaling and ISG transcription, we next examined whether DHX9 modulates the JAK-STAT1 pathway during PEDV infection. The RT-qPCR and western blotting results demonstrated that PEDV infection significantly suppressed the expression of multiple ISGs, including IFIT2, OASL, RSAD2, MX1, and ISG15, at both mRNA and protein levels. Notably, DHX9 knockdown enhanced this suppression, while DHX9 overexpression markedly alleviated the inhibitory effects of PEDV on these ISGs. However, DHX9 expression levels did not affect PEDV-mediated suppression of STAT1 and phosphorylated STAT1(pSTAT1) (Fig. 3A, B). Importantly, although DHX9 modulated ISG expression, DHX9 overexpression or silencing did not substantially alter STAT1 phosphorylation status, suggesting that DHX9 appears to function primarily at the transcriptional level downstream of STAT1 activation. Instead, these findings support a model in which DHX9 functions downstream of activated STAT1 to facilitate ISG-associated transcription during PEDV infection. To further confirm the role of DHX9 in regulating ISGs, we examined ISG expression in DHX9-knockdown and DHX9-overexpressing cells. DHX9 knockdown significantly reduced ISG mRNA levels and exacerbated the PEDV-induced downregulation, while DHX9 overexpression significantly increased their expression and reversed PEDV-mediated suppression (Fig. 3C, D), indicating that DHX9 promotes ISG transcription.

Fig. 3. DHX9 promotes ISG transcription to inhibit PEDV replication.

Vero cells and IPEC-J2 cells, transfected with siDHX9 or siNC (control) for 24 h, were pretreated with IFN-β for 2 h and infected with or without PEDV at MOI = 1 for 24 h, followed by western blotting with indicated antibodies (A) and RT-qPCR (C). The right panels show semi-quantitative analyses of the protein bands. Vero cells and IPEC-J2 cells, transfected with pCAGGS-DHX9-HA or pCAGGS-HA (control) for 24 h, were pretreated with IFN-β for 2 h and infected with or without PEDV at MOI = 1 for 24 h, followed by western blotting with indicated antibodies (B) and RT-qPCR (D). The right panels show semi-quantitative analyses of the protein bands. E Vero cells or IPEC-J2 cells, transfected with siISG15, siRSAD2, siOASL, or siNC (control) for 24 h, were pretreated with IFN-β for 2 h and infected with or without PEDV at MOI = 1 for 24 h, followed by western blotting with the indicated antibodies. The bottom panels show semi-quantitative analyses of the protein bands. F Released viral titers were detected using a TCID₅₀ assay. Bar graphs represent the mean of three independent experiments (± SD). ns, P > 0.05; *P < 0.05, **P < 0.01, and ***P < 0.001.

To further assess the role of DHX9 under basal conditions, we performed experiments in the absence of exogenous IFN stimulation. DHX9 overexpression alone exerted minimal effects on ISG expression in non-infected Vero cells, as shown by both western blotting and RT-qPCR analyses (Supplementary Fig. 1A, C). However, during PEDV infection, DHX9 overexpression markedly restored the expression of multiple ISGs, including IFIT2, OASL, RSAD2, MX1, and ISG15, which were otherwise suppressed by viral infection. To evaluate whether the downstream transcriptional machinery remains intact in this IFN-deficient context, we ectopically expressed STAT1 as a functional control. Notably, STAT1 overexpression increased the expression of several ISGs even in the absence of exogenous IFN stimulation, likely reflecting residual basal signaling activity sufficient to support downstream transcription. (Supplementary Fig. 1A–D). Importantly, a similar trend was observed in IPEC-J2 cells, where both DHX9 and STAT1 overexpression enhanced ISG expression and partially rescued the inhibitory effects of PEDV (Supplementary Fig. 1B, D).

To determine whether these ISGs contribute to PEDV replication, we silenced ISG15, OASL, and RSAD2 using siRNAs. After confirming the effective depletion of these proteins by western blotting (Supplementary Fig. 2), cells were infected with PEDV at an MOI of 1. At 48 h post-infection, western blotting revealed significantly increased PEDV S and N protein levels in ISG-silenced cells (Fig. 3E), accompanied by markedly elevated viral titers (Fig. 3F).

Collectively, these data demonstrate that although DHX9 alone is insufficient to drive ISG expression under basal conditions, it effectively restores ISG expression during PEDV infection. This effect occurs in a cellular context where the STAT1 signaling axis remains intact and responsive, highlighting that DHX9 likely functions by facilitating STAT1-dependent transcription rather than initiating ISG expression independently.

DHX9 interacts with PEDV N

To investigate how PEDV antagonizes DHX9, we next tested whether any PEDV structural proteins directly interact with DHX9. Vero cells were co-transfected with DHX9-HA and individual PEDV structural plasmids (pCAGGS-S-FLAG, pCAGGS-E-FLAG, pCAGGS-M-FLAG, or pCAGGS-N-FLAG) for 48 h, followed by co-IP. Western blotting showed that only N-FLAG was efficiently co-immunoprecipitated with HA-DHX9, whereas S-, E-, and M-FLAG were not (Fig. 4A). Next, we examined the subcellular distribution of this interaction. Nuclear and cytoplasmic fractionation indicated that DHX9 was predominantly nuclear, with minimal cytoplasmic presence, while PEDV N localized mainly in the cytoplasm but also exhibited a detectable nuclear pool (Supplementary Fig. 3A). Co-immunoprecipitation from fractionated lysates demonstrated that N robustly associated with DHX9 in the nucleus and was also detectable in the cytoplasm, suggesting that the interaction occurs in both compartments, with a stronger association in the nucleus (Supplementary Fig. 3B). To determine whether this interaction is conserved across species, N-FLAG was co-expressed with either African green monkey DHX9-HA or swine DHX9-HA in Vero cells or 293T cells. Both forms of DHX9 were robustly detected in N-FLAG immunocomplexes but not in FLAG controls, confirming that PEDV N binds DHX9 from both species (Fig. 4B). Moreover, PEDV N was also associated with endogenous DHX9 under more physiological conditions: N-FLAG overexpression in Vero and IPEC-J2 cells allowed anti-FLAG IP to pull down endogenous DHX9, and reciprocal IPs from PEDV-infected Vero cells with anti-DHX9 or anti-PEDV N further verified the interaction between native DHX9 and viral N protein (Fig. 4C). In PEDV-infected Vero cells, immunofluorescence staining revealed strong PEDV N expression accompanied by clear co-localization with endogenous DHX9 (Fig. 4D), supporting the interaction between DHX9 and PEDV N during viral infection. To facilitate interpretation of the line-scan plots in Fig. 4D, overlap of the red and green fluorescence peaks within the DAPI-positive region was interpreted as nuclear co-localization of the two proteins, whereas separation of the peaks or a shift relative to the DAPI signal was interpreted as cytoplasmic localization or reduced co-localization. These data indicate that PEDV N and DHX9 are spatially associated in infected cells in both nuclear and cytoplasmic compartments. Consistent with the infection-based observations, in the overexpression system, confocal microscopy showed that DHX9 was mainly distributed in the nucleus in mock-transfected cells, whereas PEDV N-GFP was detected in both the cytoplasm and nucleus. In co-transfected cells, DHX9 (red fluorescence) and N-GFP (green fluorescence) were well co-localized in the nucleus and partially co-localized in the cytoplasm, indicating a spatial association between these two proteins (Supplementary Fig. 4). As shown in Fig. 4E and Supplementary Table 1, swine DHX9 and PEDV N form multiple hydrogen bonds involving residues such as ILE411–PRO162, ASN408–TYR351, TYR405–ASN352, GLU634–LYS387, and GLU423–LYS388, indicating a stable protein–protein interface with a binding energy of −48.4 kcal/mol. In addition, we constructed a DHX9–N-CTD docking model based on the crystal structure of human DHX9 (PDB ID: 8SZP) and the crystal structure of the PEDV N C-terminal domain (N-CTD; PDB ID: 8WQK). This model revealed a well-defined binding interface between DHX9 and N-CTD, further supporting a direct interaction between these proteins (Supplementary Fig. 5 and Supplementary Table 1). Because human and porcine DHX9 are highly conserved, the difference in predicted hydrogen-bond numbers between the two docking models should be interpreted cautiously and may reflect local interface differences or docking geometry rather than a definitive difference in binding strength. To define the interaction interface and validate the binding sites, we constructed three truncated PEDV N plasmids: N-FLAG (Δ1–146 aa), N-FLAG (Δ147–265 aa), and N-FLAG (Δ266–442 aa) (Fig. 4F). By co-IP, DHX9 was found to interact with the CTD domain of the PEDV N protein (Fig. 4G).

Fig. 4. DHX9 interacts with PEDV N.

A Vero cells were co-transfected with pCAGGS-DHX9-HA and pCAGGS-S-FLAG, pCAGGS-E-FLAG, pCAGGS-M-FLAG, pCAGGS-N-FLAG, pCAGGS-HA, or pCAGGS-FLAG (control) for 48 h, respectively. Co-IP and western blotting were performed to detect the interaction between DHX9 and PEDV structural proteins with antibodies to FLAG and HA. B Vero cells or 293T cells were pretreated with MG132 for 4 h and then co-transfected with pCAGGS-DHX9-HA (left, African green monkey DHX9) or pCAGGS-sDHX9-HA (right, swine DHX9) together with pCAGGS-N-FLAG, or with the corresponding control plasmids pCAGGS-HA and pCAGGS-FLAG, for 48 h. Co-IP and western blotting were performed to detect the interaction between PEDV N and DHX9 using anti-FLAG and anti-HA antibodies. C Vero cells were transfected with pCAGGS-N-FLAG for 48 h (left panel) or infected with PEDV for 24 h (middle two panels), and IPEC-J2 cells were transfected with pCAGGS-N-FLAG for 48 h (right panel). Cell lysates were subjected to IP with rabbit anti-DHX9, mouse anti-FLAG, mouse anti-PEDV N, or the corresponding normal IgG, followed by western blotting with anti-FLAG, anti-DHX9, or anti-PEDV N antibodies as indicated. D Vero cells were infected with PEDV and subjected to immunofluorescence staining using anti-DHX9 (red) and anti-PEDV N (green) antibodies, with nuclei counterstained using DAPI (blue). Images were acquired using a laser confocal microscope. A white line was drawn across a representative cell in the merged image, and fluorescence intensities of DHX9 (red), PEDV N (green), and DAPI (blue) along this line were quantified using ImageJ and plotted as intensity–distance profiles to illustrate nuclear versus cytoplasmic co-localization. Scale bar = 20 μm. E Surface diagram of the docking model showing the binding interface between swine DHX9 (blue) and PEDV N (green). F Schematic representation of PEDV N-FLAG and its truncated derivatives. G Vero cells were pretreated with MG132 for 4 h and then harvested after co-transfection with DHX9-HA and N-FLAG or its truncated derivatives for 48 h. Co-IP and western blotting were performed to detect the interaction between DHX9 protein and N protein or its truncated derivatives with antibodies to FLAG and HA.

PEDV N can suppress the transcription of ISGs by disrupting the interactions between DHX9 and Pol II, as well as that between STAT1 and Pol II

To explore the mechanism by which PEDV suppresses the transcription of ISGs, we first co-transfected DHX9-HA and STAT1-FLAG plasmids into Vero cells for 48 h, followed by co-IP. By western blotting, STAT1-FLAG was found to interact with DHX9-HA (Fig. 5A). Similarly, endogenous STAT1 was also found to interact with DHX9 (Fig. 5B). As shown in Fig. 5C and Supplementary Table 1, DHX9 and STAT1 form multiple hydrogen bonds involving residues such as LYS606–ILE218, SER400–ASN233, GLU634–ARG304, GLU387–PRO490 and SER253–ARG494, indicating a stable DHX9–STAT1 complex. Furthermore, confocal microscopy showed co-localization of DHX9 (red fluorescence) and STAT1 (green fluorescence) in co-transfected cells (Fig. 5D), further confirming their interaction. To determine the subcellular compartment in which the DHX9–STAT1 complex is formed, we performed cytoplasmic/nuclear fractionation followed by co-IP in cells co-expressing DHX9-HA and STAT1-FLAG and stimulated with IFN-β. Western blot analysis of the fractionated samples showed that DHX9 was strongly enriched in the nuclear fraction, whereas STAT1 and pSTAT1 were also strongly enriched in the nuclear fraction and only weakly detectable in the cytoplasmic fraction (Supplementary Fig. 3A). When DHX9-HA was immunoprecipitated from each fraction using an anti-HA antibody, a robust co-precipitation of STAT1-FLAG was observed in the nuclear fraction, while only a weak signal was detected in the cytoplasmic fraction (Supplementary Fig. 3C). These results indicate that the DHX9–STAT1 complex is predominantly formed in the nucleus.

Fig. 5. DHX9 interacts with STAT1 and recruits RNA Pol II to promote the expression of ISGs.

A Vero cells were co-transfected with pCAGGS-DHX9-HA and pCAGGS-STAT1-FLAG, pCAGGS-HA, or pCAGGS-FLAG (control) for 48 h, respectively. Co-IP and western blotting were performed to detect the interaction between DHX9 and STAT1 with antibodies to FLAG and HA. B Vero cells were harvested after being pretreated with IFN-β for 2 h. The lysates were incubated with 1 μg rabbit anti-DHX9 or mouse anti-STAT1 antibody, or corresponding normal IgG. C Surface diagram of the docking model showing the binding interface between swine DHX9 (blue) and swine STAT1 (orange). D Vero cells were transfected with pCherry-DHX9 and/or pEGFP-STAT1 for 48 h and stimulated with IFN-β for the last 2 h. Co-localization of DHX9 and STAT1 was visualized under a laser confocal microscope. A white line was drawn across a representative cell in the merged image, and fluorescence intensities of DHX9-Cherry (red), STAT1-EGFP (green), and DAPI (blue) along this line were quantified using ImageJ and plotted as intensity–distance profiles on the right to illustrate their nuclear versus cytoplasmic co-localization. Scale bar = 20 μm. E–L Vero cells were pretreated with MG132 for 4 h and transfected with the indicated plasmids. At 24 h post-transfection, IFN-β was added during the final 2 h before harvesting. Cell lysates were incubated with 1 μg rabbit anti-DHX9 or mouse anti-STAT1 antibody, or the corresponding normal IgG control.

Studies have demonstrated that DHX9 interacts with STAT1 and recruits RNA polymerase II (Pol II) to participate in STAT1-mediated transcription, thereby promoting the expression of ISGs to combat viral infection²⁷. To investigate whether PEDV N protein affects the recruitment of Pol II through the interaction between DHX9 and STAT1, we examined the potential binding between DHX9 and Pol II. The results showed that IFN-β stimulation enhanced the interaction between DHX9 and Pol II (Fig. 5E), whereas PEDV infection or overexpression of PEDV N inhibited this interaction (Fig. 5F, G). To exclude the possibility that the PEDV N protein affects the binding of DHX9 to Pol II by degrading DHX9, we pretreated cells with MG132 and examined their binding status. The results showed that even when DHX9 degradation was inhibited, the PEDV N protein could still significantly suppress the binding of DHX9 to Pol II (Fig. 5H). This finding suggests that, in addition to inhibiting the binding of DHX9 to Pol II through DHX9 degradation, PEDV may also impair Pol II recruitment by interfering with the interaction between DHX9 and STAT1. Additionally, co-IP assays in IFN-β-stimulated Vero cells revealed that DHX9 knockdown significantly weakened the interaction between endogenous STAT1 and Pol II (Fig. 5I). Similarly, infection of PEDV or overexpression of PEDV N also inhibited the binding of STAT1 to Pol II (Fig. 5J, K). Subsequently, cells were pretreated with MG132 to eliminate the influence of PEDV N protein-mediated DHX9 degradation. The results showed that even when DHX9 degradation was inhibited, the PEDV N protein could still significantly suppress the binding of STAT1 to Pol II (Fig. 5L). These observations suggested that PEDV N might inhibit Pol II recruitment at least in part by destabilizing the DHX9-STAT1 complex.

PEDV N protein competitively disrupts the interaction between DHX9 and STAT1

Having noted that both DHX9–STAT1 and DHX9–PEDV N complexes predominantly form in the nucleus (Supplementary Fig. 3), we next examined whether PEDV N affects the interaction between DHX9 and STAT1. For this, we constructed three truncated DHX9 plasmids: DHX9-HA-D1 (1–252 aa), DHX9-HA-D2 (253–635 aa), and DHX9-HA-D3 (636–1276 aa) (Fig. 6A). These plasmids were co-transfected with either N-FLAG or STAT1-FLAG plasmid into cells, followed by co-IP. By western blotting, both PEDV N and STAT1 proteins were found to bind to the D2 region of DHX9 (Fig. 6B, C). Subsequently, we co-transfected DHX9-HA, STAT1-FLAG, and PEDV N plasmids into cells. The confocal microscopy demonstrated that while DHX9 (red fluorescence) and STAT1 (green fluorescence) well co-localized in co-transfected cells, the addition of PEDV N protein disrupted their co-localization (Fig. 6D). Similarly, co-IP results showed that PEDV infection or the overexpression of PEDV N protein inhibited the binding between DHX9 and STAT1 (Figs. 6E and Supplementary Fig. S6). To further substantiate that PEDV N affects the DHX9-STAT1 interaction, we co-transfected DHX9-HA and STAT1-EGFP plasmids with either N-FLAG or N-FLAG (Δ266–442 aa) plasmid into cells and performed co-IP assays. Western blotting showed that full-length PEDV N protein disrupted the DHX9-STAT1 interaction, while the N protein lacking the CTD domain did not affect this interaction (Fig. 6F). These findings demonstrate that PEDV N protein competitively disrupts the binding between DHX9 and STAT1.

Fig. 6. The PEDV N protein disrupts the interaction between DHX9 and STAT1.

A Schematic representation of DHX9-HA and its truncated derivatives. B Vero cells, co-transfected with N-FLAG and DHX9-HA or the indicated DHX9 truncation mutants (DHX9-D1-HA, DHX9-D2-HA, or DHX9-D3-HA) for 48 h, were subjected to Co-IP and western blotting to detect the interaction between PEDV N and DHX9 using anti-FLAG and anti-HA antibodies. C Vero cells, co-transfected with STAT1-FLAG and DHX9-HA or the indicated DHX9 truncation mutants (DHX9-D1-HA, DHX9-D2-HA, or DHX9-D3-HA) for 48 h, were pretreated with IFN-β for 2 h followed by Co-IP and western blotting to detect the interaction between STAT1 and DHX9 using anti-FLAG and anti-HA antibodies. D Vero cells, co-transfected with pCherry-DHX9, pEGFP-STAT1 and PEDV N for 48 h, were pretreated with IFN-β for 2 h and then imaged by laser confocal microscopy to visualize DHX9–STAT1 co-localization. A white line was drawn across a representative cell in the merged image, and fluorescence intensities of DHX9-Cherry (red), STAT1-EGFP (green), and DAPI (blue) along this line were quantified using ImageJ and plotted as intensity–distance profiles on the right to illustrate their nuclear versus cytoplasmic co-localization. Scale bar = 20 μm. E Vero cells, co-transfected with DHX9-HA, STAT1-FLAG and PEDV N for 48 h, were pretreated with IFN-β for 2 h followed by Co-IP and western blotting to detect the interaction between DHX9 and STAT1 using anti-FLAG and anti-HA antibodies. F Vero cells, co-transfected with DHX9-HA, STAT1-EGFP and N-FLAG or N-FLAG (Δ266–442 aa) for 48 h, were pretreated with IFN-β for 2 h followed by Co-IP and western blotting to detect the interaction between DHX9 and STAT1 using anti-FLAG and anti-HA antibodies.

PEDV promotes DHX9 degradation via ubiquitination, disrupting the DHX9-STAT1 interaction

To investigate how PEDV N downregulates DHX9, the pCAGGS-DHX9-HA plasmid was co-transfected with increasing doses (0, 1, 2, and 3 μg) of the pCAGGS-N-FLAG plasmid for 48 h. Western blotting revealed that DHX9 protein levels decreased in a dose-dependent manner with increasing PEDV N expression (Fig. 7A), suggesting that PEDV N negatively regulates DHX9 protein levels.

Fig. 7. PEDV N protein promotes viral replication by disrupting the DHX9-STAT1 interaction via ubiquitination.

A Vero cells were co-transfected with 0.5 μg of pCAGGS-DHX9-HA plasmid and different doses (0, 1, 2, and 3 μg) pCAGGS-N-FLAG plasmid for 48 h, followed by western blotting with the indicated antibodies. As a control, corresponding doses of the empty vector pCAGGS-FLAG (3, 2, 1, and 0 μg) were added. The right panel shows a semi-quantitative analysis of the protein bands. B Vero cells were infected with PEDV or transfected with pCAGGS-N-FLAG plasmid and then treated with CHX. Vero cells were harvested after different time points (0, 3, 6, and 12 h), followed by western blotting with the indicated antibodies. The lower panels show semi-quantitative analyses of the protein bands. C Vero cells were pretreated with 3-MA prior to either PEDV infection or transfection with pCAGGS-N-FLAG plasmid. After 24 h incubation, cells were subsequently harvested and analyzed by western blotting with the indicated antibodies. The lower panels show semi-quantitative analyses of the protein bands. D Vero cells were pretreated with MG132 and then infected with PEDV or transfected with pCAGGS-N-FLAG plasmid for 24 h. Cells were subsequently harvested and analyzed by western blotting with the indicated antibodies. The lower panels show semi-quantitative analyses of the protein bands. E Vero cells were harvested after infection with PEDV co-transfection with DHX9-HA, pHis-Ub, and pCAGGS-N-FLAG for 48 h. Co-IP and western blotting were performed to detect the ubiquitination of DHX9 protein with antibodies to FLAG, HA, and His. F Vero cells were harvested after co-transfection with DHX9-HA and MDM2-FLAG for 48 h. Co-IP and western blotting were performed to detect the interaction of DHX9 and MDM2 protein with antibodies to FLAG and HA. G Vero cells were harvested after transfection with pCAGGS-N-FLAG or pCAGGS-FLAG for 48 h. Co-IP and western blotting were performed to detect the interaction of DHX9 and MDM2 protein with the indicated antibody. H Validation of the siRNA targeting MDM2 by RT-qPCR in Vero cells. I Vero cells were co-transfected siMDM2 and pCAGGS-N-FLAG plasmid for 24 h, followed by western blotting with the indicated antibodies. For control, siNC or the empty vector pCAGGS-FLAG were added. The right panel shows a semi-quantitative analysis of the protein bands. J Vero cells were harvested after co-transfection with DHX9-HA, STAT1-FLAG, PEDV N, and siMDM2 for 24 h. Co-IP and western blotting were performed to detect the interaction of DHX9 and STAT1 with antibodies to FLAG and HA. Bar graphs represent the mean of three independent experiments (± SD). ns, P > 0.05; *P < 0.05, **P < 0.01, and ***P < 0.001.

Having observed that PEDV infection reduces the protein levels of DHX9 without altering its mRNA levels (Fig. 1A, B), we wondered whether PEDV infection downregulates DHX9 protein levels by promoting its degradation. To this end, Vero cells infected with PEDV or overexpressing PEDV N were treated with cycloheximide (CHX), which inhibits de novo protein synthesis by blocking the elongation of nascent peptide chains²⁸. Cells were harvested at 0, 3, 6, and 12 h post-treatment for western blotting. The results showed that DHX9 degradation was significantly accelerated in PEDV-infected or N-overexpressing cells compared to controls (Fig. 7B), indicating that PEDV downregulates DHX9 protein levels by promoting its protein degradation.

To further explore the mechanism of DHX9 degradation induced by PEDV, cells were treated with the autophagy inhibitor 3-MA or the proteasome inhibitor MG132. While 3-MA failed to prevent the decrease in DHX9 levels, MG132 was able to restore DHX9 protein levels in PEDV-infected or N-overexpressing cells (Fig. 7C, D). These findings suggest that PEDV promotes DHX9 degradation via the proteasome pathway.

Ubiquitination is a common mechanism for protein degradation in cells²⁹. To examine whether PEDV enhances DHX9 ubiquitination, cells infected with PEDV or overexpressing N were transfected with a HA-Ub plasmid and treated with MG132, followed by co-IP. The results demonstrated a marked increase in ubiquitination signals in PEDV-infected or N-overexpressing cells compared to controls (Fig. 7E). E3 ubiquitin ligases play a critical role in substrate recognition during ubiquitination³⁰. Previous studies suggest that MDM2, an E3 ubiquitin ligase, may regulate DHX9 ubiquitination³¹. To investigate this, DHX9-HA and MDM2-FLAG were co-expressed in Vero cells. The co-IP assays confirmed the interaction between DHX9 and MDM2 (Fig. 7F). Furthermore, PEDV infection or the overexpression of PEDV N enhanced the interaction between endogenous DHX9 and MDM2 (Fig. 7G and Supplementary Fig. 7). To validate the role of MDM2 in DHX9 ubiquitination, MDM2 was knocked down in Vero cells overexpressing PEDV N. RT-qPCR confirmed efficient MDM2 knockdown, and western blotting showed that MDM2 knockdown reversed the PEDV N-induced decrease in DHX9 levels (Fig. 7H, I), indicating that the PEDV N promotes DHX9 degradation via MDM2-mediated ubiquitination.

Finally, to determine whether PEDV N affects the DHX9-STAT1 interaction through ubiquitination, MDM2 was silenced or PEDV N was overexpressed in Vero cells co-transfected with DHX9-HA and STAT1-FLAG. Co-IP assays revealed that overexpression of PEDV N inhibited the DHX9-STAT1 interaction, which was reversed upon MDM2 knockdown (Fig. 7J). These results demonstrate that PEDV N disrupts the DHX9-STAT1 interaction by promoting MDM2-mediated ubiquitination and degradation of DHX9.

Discussion

The DExD/H-box helicases are broadly implicated in antiviral immunity. Prototypical members, including DDX3³², DHX15³³, and DHX33³⁴, primarily activate downstream antiviral signaling pathways through cytoplasmic RNA recognition or synergistic interactions with MAVS. Recent studies have revealed the nuclear immunoregulatory functions of RNA helicases. For instance, DDX46 recruits the RNA demethylase ALKBH5 via its DEAD-box helicase domain to selectively remove m6A methylation marks from antiviral transcripts during viral infection³⁵, while DDX3X is specifically recruited to enhancer-binding sites of the IFN-β promoter to directly regulate type I interferon transcription³⁶. DHX9 has been shown to positively regulate ISG expression downstream of type I interferon signaling²⁷. Here, we demonstrate that PEDV infection suppresses DHX9 expression and that DHX9 overexpression inhibits PEDV replication. Treatment of siDHX9- or pCAGGS-DHX9-HA-transfected Vero and IPEC-J2 cells with IFN-β revealed that DHX9 knockdown enhances PEDV-mediated suppression of ISG expression, while DHX9 overexpression reverses this inhibition. These findings suggest that PEDV suppresses DHX9-associated antiviral transcriptional responses during infection. Because Vero cells are deficient in endogenous type I interferon production, caution is needed to interpret canonical innate immune signaling in this model. In the present study, exogenous IFN-β stimulation was therefore used in several experiments to synchronize STAT1 activation and enhance the dynamic range of ISG-associated transcriptional responses. Importantly, our findings do not indicate that DHX9 independently activates canonical JAK–STAT signaling in the absence of cytokine stimulation. Rather, the data support a model in which DHX9 functions downstream of activated STAT1 to facilitate ISG-associated transcription. Consistent with this interpretation, DHX9 modulation had a  minimal effects on STAT1 phosphorylation itself but substantially affected downstream ISG expression. In addition, similar trends were observed in IFN-competent IPEC-J2 cells, supporting the broader relevance of the DHX9–STAT1 axis during PEDV infection.

DHX9 (also designated RNA helicase A or nuclear DNA helicase II) is a multifunctional ATP-dependent helicase that orchestrates diverse cellular processes, including DNA replication/repair, transcriptional regulation, post-transcriptional RNA modification, and mRNA translation³⁷. Regarding DHX9 as a key regulator of RNA splicing and translation, in vitro studies have demonstrated its dual functionality in either potentiating antiviral immune responses^38,39 or facilitating viral replication within host cells¹⁴. This functional versatility positions DHX9 as a critical host factor across viral life cycles, with proviral or antiviral activities that vary by virus and infection stage⁴⁰. During MYXV infection, DHX9 forms specialized cytoplasmic antiviral granules that sequester viral components to suppress late-stage protein synthesis and restrict viral proliferation in human cancer cells²³. Conversely, DHX9 can be co-opted by viruses like influenza A virus, where it interacts with NS1 to enhance viral transcription and replication¹⁹.

In our study, nuclear/cytoplasmic fractionation combined with co-immunoprecipitation revealed that both DHX9–STAT1 and DHX9–PEDV N complexes form in the nucleus (Supplementary Fig. 3). DHX9 was strongly enriched in the nuclear fraction, while STAT1 and pSTAT1 were also predominantly nuclear and only weakly detectable in the cytoplasm. Co-IP further confirmed that DHX9 interacts with STAT1 mainly in the nuclear fraction, consistent with the observation that DHX9–PEDV N complexes also preferentially form in the nucleus. These findings provide strong mechanistic support for the notion that PEDV N exerts its inhibitory effect primarily by interfering with the nuclear DHX9–STAT1 transcriptional complex, rather than simply through cytoplasmic interactions.

The nucleocapsid (N) protein plays crucial roles in modulating host antiviral responses, particularly through interferon signaling pathways^41,42. In IPEC-J2 cells, N protein specifically inhibits type III interferon-dependent NF-κB nuclear translocation without affecting poly(I:C)-induced type I/II interferon expression⁴³. In addition, N protein antagonizes IFN-β production and IFN-λ3 promoter activity by suppressing IRF-3 and NF-κB activation^44,45. Our study demonstrates that PEDV N protein impairs ISG transcription by disrupting DHX9-mediated STAT1-RNA polymerase II (Pol II) interactions, while establishing DHX9 as an important facilitator of STAT1-associated antiviral transcription. We selected IFIT2, MX1, ISG15, RSAD2, and OASL as representative ISGs because they were among the most consistently DHX9-responsive antiviral transcripts in our dataset and have been reported to restrict PEDV in previous studies⁴⁶. We agree that OASL is more likely to act indirectly by potentiating RNA-sensing/IFN signaling (e.g., RIG-I–dependent pathways) rather than functioning as a purely “direct-acting” antiviral effector. Accordingly, in the context of DHX9 overexpression and/or IFN stimulation, OASL is expected to contribute mainly by reinforcing and sustaining the broader ISG program. Notably, the increased PEDV replication observed upon OASL knockdown in our system supports that OASL still provides a measurable functional contribution even under an activated antiviral signaling background. Notably, other ISGs such as IFITM1 and IFITM3 have also been reported to be upregulated during PEDV infection and to facilitate viral entry and replication by interacting with the viral S1 protein and modulating endocytic trafficking⁴⁷, rather than by acting as classical STAT1–Pol II–dependent transcriptional effectors. In light of this distinct, entry-associated mechanism, IFITM family members were not included as readouts in our DHX9–STAT1–Pol II–focused assays. Nevertheless, it would be of interest in future studies to determine whether DHX9 indirectly influences IFITM expression or activity and whether PEDV N also modulates this proviral ISG program. These findings suggest that the function of DHX9 during viral infection is context-dependent, dictated by subcellular localization, post-translational modifications, and viral counterstrategies. We further identify two distinct mechanisms by which PEDV suppresses DHX9’s antiviral functions: (1) competitive disruption of the DHX9-STAT1-Pol II transcriptional complex by N protein, and (2) MDM2-mediated degradation of DHX9. Our data support an indirect model in which PEDV N interferes with the DHX9–STAT1 association, thereby potentially increasing the availability of DHX9 for MDM2-dependent ubiquitination. These two mechanisms may act in a coordinated manner, whereby disruption of the DHX9–STAT1 complex could enhance DHX9 accessibility to MDM2, ultimately promoting its ubiquitination and degradation. However, the precise molecular basis of this coordination remains to be determined. Another limitation of the current study is that several mechanistic conclusions were derived from transformed cell lines and partially from IFN-stimulated or overexpression-based systems, which may not fully recapitulate the physiological antiviral signaling environment in primary intestinal epithelial cells in vivo. Therefore, future studies in primary porcine cells, particularly primary intestinal epithelial models, will be important to further validate the physiological relevance of these observations.

Our findings suggest that PEDV suppresses DHX9-dependent antiviral transcription as part of its immune evasion strategy. By targeting DHX9 to suppress ISG expression, PEDV creates a cellular environment favorable for viral replication-a strategy analogous to that employed by HIV-1 and hepatitis B virus^15,21. However, our discovery of N protein-mediated DHX9 degradation represents an immune evasion mechanism among coronaviruses. These findings carry significant therapeutic implications. Restoring DHX9 function or disrupting PEDV N-DHX9 interactions may constitute viable antiviral strategies. Targeting the PEDV N–DHX9 axis or the associated ubiquitination pathway may provide potential directions for future antiviral development. Furthermore, given DHX9’s evolutionary conservation, our insights may inform therapeutic development against other coronaviruses.

In summary, this work demonstrates a mechanism by which PEDV suppresses DHX9-associated antiviral transcription, whereby PEDV downregulates DHX9 to subvert host defenses, supporting an important role for DHX9 in PEDV–host interactions. Future research is needed to explore whether DHX9-targeted interventions can enhance antiviral responses in vivo, which may inform future antiviral studies against PEDV and related coronaviruses.

Methods

Reagents

The chemical reagents used in this study included 3-methyladenine (3MA, Selleckchem, S2767), MG132 (Selleckchem, S2619), and cycloheximide (Selleckchem, S7418), which were dissolved in dimethyl sulfoxide (DMSO) and stored at −80 °C. IFN-β (MCE; HY-P7024). The primary antibodies used in this study were as follows: rabbit polyclonal anti-DHX9 (ABclonal, 1:1000 for WB, 4 μg for IP, 1:50 for IF, A17955), rabbit polyclonal anti-Lamin B1 (ABclonal, 1:1000 for WB, A11495), rabbit polyclonal anti-GFP (ABclonal, 1:1000 for WB, AE078), rabbit polyclonal anti-IFIT2 (ABclonal, 1:1000 for WB, A13760), rabbit polyclonal anti-OASL (ABclonal, 1:1000 for WB, A24529), rabbit polyclonal anti-RSAD2 (ABclonal, 1:1000 for WB, A8271), rabbit polyclonal anti-MX1 (ABclonal, 1:1000 for WB, A1780), rabbit polyclonal anti-ISG15 (ABclonal, 1:1000 for WB, A1182), rabbit polyclonal anti- POLR2A (ABclonal, 1:1000 for WB, A2107), rabbit polyclonal anti-MDM2 (ABclonal, 1:1000 for WB, A23388), rabbit polyclonal anti-pSTAT1 (Y701) (Abcam, 1:1000 for WB, ab30645), mouse monoclonal anti-GAPDH (Beyotime, 1:1000 for WB, AG019), mouse monoclonal anti-Actin (Beyotime, 1:1000 for WB, AA128), mouse monoclonal anti-STAT1 (Abcam, 1:1000 for WB, ab3987), mouse monoclonal anti-HA (Beyotime, 1:1000 for WB, 2 μg for IP, AH158), and mouse monoclonal anti-FLAG (Beyotime, 1:1000 for WB, 2 μg for IP, A519). The secondary antibodies used for immunoblotting were horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG (Beyotime, 1:1000 for WB, A0216) and HRP-conjugated goat anti-rabbit IgG (Beyotime, 1:1000 for WB, A0208). Mouse polyclonal antibodies to PEDV-N and PEDV-S were prepared in the laboratory (1:1000 for WB, 4 μg for IP, 1:50 for IF),

Cell culture

Vero E6 and 293T cells were purchased from American Type Culture Collection (Manassas, VA, USA) and cultured in Dulbecco’s modified Eagle’s medium (DMEM, Thermo Fisher Scientific, 11995500) supplemented with 10% fetal bovine serum (FBS) (Thermo Fisher Scientific, 10099). IPEC-J2 cells (porcine small intestinal epithelial cell line, DSMZ, Germany) were cultured in RPMI 1640 medium (Thermo Fisher Scientific, 11875119) supplemented with 10% FBS. Cells were cultured at 37 °C in a 5% CO₂ incubator.

Virus infection

The PEDV strain GDZH02 (GenBank accession number: KR153325) was propagated in Vero E6 or IPEC-J2 cells using the corresponding culture medium supplemented with 2 μg/mL trypsin (Thermo Fisher Scientific, 2500072). Viral titers were determined as 50% tissue culture infective dose (TCID₅₀)/0.1 mL using Vero cell assays. For comparative studies of viral replication, cells were infected at a multiplicity of infection (MOI) of 1. Viral stocks were prepared at an MOI of 1 for subsequent experiments. Vero or IPEC-J2 cells were infected with PEDV at various MOIs as required for specific experiments. Mock-infected controls were treated with phosphate-buffered saline (PBS, Thermo Fisher Scientific, 10010023). After 1.5 h of infection, the inoculum was removed, and cells were gently washed with PBS before being cultured in complete growth medium at 37 °C for the indicated time periods until harvest.

Cell viability assay (CCK-8)

Cells were seeded in 96-well plates and transfected with siDHX9 or negative control siRNA (siNC) using Lipofectamine 3000 reagent (Thermo Fisher, L3000015) according to the manufacturer’s instructions. At 24 h post-transfection, 10 μL of Cell Counting Kit-8 reagent (CCK-8, Genview, GK3607) was added to each well containing 100 μL culture medium and incubated for 1 h at 37 °C. Absorbance was measured at 450 nm using a microplate reader. Cell viability was calculated as (A_siRNA − A_blank)/(A_siNC − A_blank) × 100%.

Biochemical intervention

To inhibit the induction of autophagy, Vero cells grown to 80% confluence in 6-well cell culture plates were pretreated with 50 μM 3-MA for 4 h. To inhibit proteasome and certain hydrolase activities, Vero cells infected with PEDV or overexpressing the PEDV N protein were treated with 10 μM MG132. To block protein synthesis, Vero cells were treated with 50 μg/mL cycloheximide (CHX). Treatment with an equal amount of DMSO served as a control. For IFN-β stimulation, cells were treated with IFN-β at 1000 IU/mL for 2 h, unless otherwise indicated.

DNA constructs and RNA interference

pEGFP-STAT1, pEGFP-PEDV-N, pCherry-DHX9, pCAGGS-MDM2-FLAG and pCAGGS-STAT1-FLAG were deposited in our laboratory. The ubiquitin expression plasmid pcDNA3.1-Myc-Ub was purchased from Wuhan Miaoling Biotechnology Co., Ltd. (Wuhan, China), and the pcDNA3.1-His-Ub construct used in this study was generated from pcDNA3.1-Myc-Ub in our laboratory by replacing the Myc tag with an N-terminal His tag via subcloning using the following primers: forward primer, 5′-CACCATCACCATCACCATATGCAGATCTTTGTGAGAACCTTA-3′; reverse primer, 5′-CATGGTGGCGGTACCAAGCTTAAG-3′. The plasmids pN-FLAG (Δ1–146 aa), pN-FLAG (Δ147–265 aa), pN-FLAG (Δ266–442 aa), pDHX9-HA-D1 (1–252 aa), pDHX9-HA-D2 (253–635 aa), and pDHX9-HA-D3 (636–1276 aa) were synthesized by Tsingke Biotechnology (Shanghai, China). Full-length African green monkey DHX9 (Chlorocebus sabaeus; GenBank accession XM_007989248.3) was amplified by PCR from Vero cell cDNA and subcloned into the pCAGGS-HA vector using EcoRI (Thermo Fisher Scientific, ER0271) and KpnI (Thermo Fisher Scientific, ER0521) to generate the expression plasmid pCAGGS-HA-DHX9. Full-length porcine DHX9 (Sus scrofa; GenBank accession XP_020919437.1) was amplified by PCR from IPEC-J2 cell cDNA and cloned into pCAGGS-HA using EcoRI and KpnI to generate pCAGGS-HA-sDHX9. PEDV S, E, M, and N (GenBank, KR153325.1) genes were amplified via PCR and cloned into the pCAGGS-FLAG vector using EcoRI and XhoI to generate pCAGGS-S-FLAG, pCAGGS-E-FLAG, pCAGGS-M-FLAG, and pCAGGS-N-FLAG, respectively. The primers used are shown in Supplementary Table 2. Vero cells were grown to 60–80% confluence in 6-well cell culture plates and transfected with indicated plasmids or siRNAs using Lipofectamine 3000 reagent (Thermo Fisher, L3000015) according to the manufacturer’s instructions. Specific siRNAs targeting DHX9, ISG15, RSAD2, OASL, and MDM2 were synthesized by Sangon Biotech (Shanghai, China). Sequences of siRNAs are listed in Supplementary Table 3. Next, 2.5 μg of plasmid or 50 nM siRNA and 5 μL of P3000 were diluted in 125 μL of serum-free Opti-MEM (Thermo Fisher Scientific, 22600050); meanwhile, 6 μL of Lipofectamine 3000 was also diluted in 125 μL of serum-free Opti-MEM. The diluted solutions were thoroughly mixed and incubated at 37 °C for 15 min. The mixture was then pipetted into the cells, which were further cultured at 37 °C for 24 h. Following PEDV infection, the cells were incubated in fresh culture medium containing 8 μg/mL trypsin at 37 °C for 48 h. The levels of proteins targeted for knockdown were evaluated using western blotting.

Quantitative real-time RT-PCR (qRT-qPCR)

For gene expression analysis, total RNA in cells was isolated using Total RNA Kit I (Omega, R6834-02). Complementary DNA (cDNA) was synthesized using HiScript II Q RT SuperMix for qPCR (Vazyme, R223-01). Real-time qPCR was performed using ChamQ Universal SYBR qPCR Master Mix (Vazyme, Q711-02) and iQ5 iCycler Detection System (Bio-Rad, Hercules, CA, USA). Relative mRNA expression was assessed using the 2^−ΔΔCT method and normalized to the gene GAPDH. The primers used are shown in Supplementary Table 4.

Western blotting

After treatments, the cells were lysed with RIPA lysis buffer (Beyotime, P0013B) supplemented with a protease and phosphatase inhibitor cocktail (Beyotime, P1045) and phenylmethylsulfonyl fluoride (PMSF, 1 mM). The protein concentration was determined using a BCA protein assay kit (Thermo Fisher Scientific, 23225). Equal amounts of proteins (20 μg) were separated on 12.5% SDS-PAGE gels and transferred onto polyvinylidene difluoride (PVDF) membranes (Millipore, ISEQ00010). After blocking with 5% skimmed milk at 37 °C for 1 h, the membranes were incubated with primary antibodies at 4 °C overnight, and then exposed to the corresponding secondary antibodies conjugated to HRP at 37 °C for 1 h. The protein bands were detected using an ECL Plus kit (Beyotime, P0018) and imaged using a Tanon Fine Do X6 Chemiluminescent Imaging System (Bio-Equip, Shanghai, China). The protein bands of interest were semi-quantified using Image-Pro Plus 6.0 software. Densitometry results were reported as the ratio of the target proteins to GAPDH. *P < 0.05, **P < 0.01, and ***P < 0.001, ns> 0.05, n = 3. Band detection was within the linear range.

Co-immunoprecipitation

To detect the protein–protein interaction, co-immunoprecipitation (co-IP) was performed. Briefly, parental Vero cells or Vero cells co-transfected with indicated plasmids were incubated on ice with IP lysis buffer (Beyotime, P0013) containing 1 mM PMSF for 10 min. The resulting cell lysates were centrifuged at 14,000×g for 10 min at 4 °C. The supernatants were used for immunoprecipitation with appropriate antibodies and Pierce Protein A/G Magnetic Beads (88802, Thermo Fisher Scientific). Immunoprecipitated proteins were analyzed via western blotting using appropriate antibodies.

Molecular docking

Rigid-body protein–protein docking was performed using GRAMM-X (http://gramm.compbio.ku.edu/)⁴⁸. to evaluate potential interactions among DHX9, PEDV N, and STAT1. When available, experimentally solved structures were used, including the crystal structure of the PEDV N C-terminal domain (CTD; PDB ID: 8WQK) and the structure of human DHX9 (PDB ID: 8SZP). To capture porcine-specific interactions, three-dimensional models of porcine DHX9 and porcine STAT1 were predicted from their corresponding amino-acid sequences using AlphaFold 3, and additional docking analyses were performed between PEDV N and porcine DHX9, as well as between porcine DHX9 and porcine STAT1, as indicated. Docked complexes, protein–protein interfaces, and interacting residues were inspected and visualized using PyMOL, and interface parameters/residue contacts were further analyzed using PDBePISA (https://www.ebi.ac.uk/pdbe/pisa/)⁴⁹.

Confocal immunofluorescence microscopy

Vero cells or IPEC-J2 cells were grown in 35-mm glass-bottom Petri dishes. When required, the cells were transfected with the indicated plasmids prior to PEDV infection or co-transfected with the indicated plasmids for 48 h. After washing twice with PBS and fixation with 4% paraformaldehyde (Sigma-Aldrich, P6148) for 30 min, the cells were permeabilized with 0.2% Triton X-100 (Sigma-Aldrich, T8787) for 10 min. The cells were then blocked with 5% bovine serum albumin (BSA; Beyotime, ST023). Next, the cells were stained with the indicated primary antibodies and appropriate secondary antibodies for 1 h at 37 °C. Finally, the nuclei were stained with DAPI (Beyotime, C1002). Images were obtained using a STELLARIS 5 confocal microscope (Leica) with a 100× oil objective. Images were analyzed and quantified using Image-Pro Plus 6.0 software. Line-scan analysis was performed in ImageJ using the built-in Plot Profile function (no custom code or ImageJ macro was used). For representative merged confocal images, a straight line was drawn across the region of interest, and fluorescence intensity profiles were extracted along the same line for the corresponding channels (including DAPI and the indicated protein fluorescence channels). The DAPI signal was used to define the nuclear region along the line scan. Co-localization within the nucleus was inferred when the fluorescence peaks of the two protein channels overlapped within the DAPI-positive region, whereas peak separation or a shift relative to the DAPI signal indicated cytoplasmic localization or reduced co-localization.

Statistics and reproducibility

All quantitative data are presented as the mean ± standard deviation (SD). Western blot band intensities were quantified using Image-Pro Plus. Statistical analyses were performed using GraphPad Prism. Data normality was assessed using the Shapiro–Wilk test. For comparisons among multiple groups, one-way ANOVA was applied to normally distributed data, whereas the Kruskal–Wallis test was used for non-normally distributed data. For comparisons between two groups, two-tailed Student’s t test was used. A P value < 0.05 was considered statistically significant.

Unless otherwise indicated in the figure legends, all experiments were independently repeated at least three times with similar results. For cell-based assays, each biological replicate represented an independently cultured and treated cell sample. Technical replicates within the same experiment were averaged and considered as one biological replicate for statistical analysis. Representative images from immunoblotting, migration assays, and immunofluorescence experiments were selected from experiments with consistent results across independent replicates.

For western blot analyses, representative blots from at least three independent experiments are shown. Quantification of western blot band intensities was performed using independent biological replicates. The exact sample size (n), definition of replicates, and statistical tests used for each experiment are indicated in the corresponding figure legends.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Author contributions

X.L., S.H., J.L., and X.G. designed the work. X.L., G.C., Z.Y., G.L., Y.W., J.P., Y.W., and B.L. performed the experiments and analyzed the data. X.L., S.H., J.L., and X.G. wrote and edited the manuscript. All authors read and approved the final manuscript.

Peer review

Peer review information

Communications Biology thanks the anonymous reviewers for their contribution to the peer review of this work. Primary Handling Editors: Caroline Goujon and Kaliya Georgieva.

Funding

This work was supported by Lingnan Modern Agricultural Science and Technology Guangdong Provincial Laboratory Zhaoqing Branch’s 14th Five-Year Independent Project (grant number: P20211154-030303), and the Specific University Discipline Construction Project (no. 2023B10564003).

Data availability

The Supplementary Information file contains all Supplementary Figs. (Supplementary Figs. 1–7), the original uncropped western blots (Supplementary Fig. 8), and all Supplementary Tables (Supplementary Tables 1–4). Source data are provided in Supplementary Data 1. The newly generated plasmids in this study are available from the corresponding author upon reasonable request. All other data that support the findings of this study are also available from the corresponding author upon reasonable request.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s42003-026-10391-w.