Type III Interferon Restriction by Porcine Epidemic Diarrhea Virus and the Role of Viral Protein nsp1 in IRF1 Signaling

ABSTRACT

Type III interferons (IFNs) play a vital role in maintaining the antiviral state of the mucosal epithelial surface in the gut, and in turn, enteric viruses may have evolved to evade the type III IFN responses during infection. To study the possible immune evasion of the type III IFN response by porcine epidemic diarrhea virus (PEDV), a line of porcine intestinal epithelial cells was developed as a cell model for PEDV replication. IFN-λ1 and IFN-λ3 inhibited PEDV replication, indicating the anti-PEDV activity of type III IFNs. Of the 21 PEDV proteins, nsp1, nsp3, nsp5, nsp8, nsp14, nsp15, nsp16, open reading frame 3 (ORF3), E, M, and N were found to suppress type III IFN activities, and IRF1 (interferon regulatory factor 1) signaling mediated the suppression. PEDV specifically inhibited IRF1 nuclear translocation. The peroxisome is the innate antiviral signaling platform for the activation of IRF1-mediated IFN-λ production, and the numbers of peroxisomes were found to be decreased in PEDV-infected cells. PEDV nsp1 blocked the nuclear translocation of IRF1 and reduced the number of peroxisomes to suppress IRF1-mediated type III IFNs. Mutational studies showed that the conserved residues of nsp1 were crucial for IRF1-mediated IFN-λ suppression. Our study for the first time provides evidence that the porcine enteric virus PEDV downregulates and evades IRF1-mediated type III IFN responses by reducing the number of peroxisomes.

IMPORTANCE Porcine epidemic diarrhea virus (PEDV) is a highly contagious enteric coronavirus that emerged in swine in the United States and has caused severe economic losses. PEDV targets intestinal epithelial cells in the gut, and intestinal epithelial cells selectively induce and respond to the production of type III interferons (IFNs). However, little is known about the modulation of the type III IFN response by PEDV in intestinal epithelial cells. In this study, we established a porcine intestinal epithelial cell model for PEDV replication. We found that PEDV inhibited IRF1-mediated type III IFN production by decreasing the number of peroxisomes in porcine intestinal epithelial cells. We also demonstrated that the conserved residues in the PEDV nsp1 protein were crucial for IFN suppression. This study for the first time shows PEDV evasion of the type III IFN response in intestinal epithelial cells, and it provides valuable information on host cell-virus interactions not only for PEDV but also for other enteric viral infections in swine.

INTRODUCTION

A classic antiviral response in mammals is mediated by type I interferons (IFNs). Upon infection, host cells react quickly by producing IFN-α/βs, which trigger the production of hundreds of interferon-stimulating genes (ISGs) through the JAK-STAT pathway (1). Thus, type I IFNs and ISGs are the major components for the establishment of a host antiviral state and provide the first line of defense against viral infections. Recently, type III IFNs (IFN-λ) have been identified as a distinct class of antiviral cytokines and inducers of the expression of ISGs. Type III IFNs in human constitute IFN-λ1 (interleukin 29 [IL-29]), IFN-λ2 (IL-28A), IFN-λ3 (IL-28B), and IFN-λ4 (2–5). For swine, IFN-λ1 and IFN-λ3 have been described, but whether IFN-λ2 and IFN-λ4 are expressed or deficient in swine is unclear (6). The induction processes and mechanisms of action of type I IFNs and type III IFNs are similar (7–9). While the induction of type I IFNs requires all components of the IFN enhanceosome, type III IFNs can be produced by independent actions of IRF (interferon regulatory factor) and NF-κB (10). Similar to type I IFNs, IRF3, IRF7, and NF-κB are essential components of type III IFNs, but IRF1 plays a particular role (11). The mitochondrial antiviral signaling (MAVS) protein localizes to mitochondria and peroxisomes and participates in activating an antiviral response. While mitochondrial MAVS is responsible for the activation of type I IFNs, peroxisomal MAVS induces a type III IFN-dependent antiviral response (12, 13). Peroxisomal MAVS in intestinal epithelial cells (IECs) is particularly crucial for the IRF1-dependent activation of the type III IFN response (11).

A main difference between the type I and type III IFN systems is that receptors for type I IFNs are seemingly ubiquitous, whereas receptors for type III IFNs are confined to the mucosal epithelium (9). Thus, the type III IFN response is limited mostly to epithelial cells due to the highly restricted expression of specific receptors. Recent studies show that IECs express extremely low levels of type I IFN receptors and thus respond poorly to type I IFNs (14). IECs produce type III IFNs abundantly but produce type I IFNs poorly and thus selectively respond to type III IFNs. The robust response of type III IFNs to invading pathogens is due to the differentiation of IECs and the upregulation of the biogenesis of peroxisomes (11). For antiviral defense, IECs in the gut mucosa respond mainly to type III IFNs after viral infection, whereas other cell types in the gut rely on type I IFNs (14). In turn, viruses have evolved to evade the type III IFN response. Hepatitis C virus (HCV) NS3-4A cleaves peroxisomal MAVS to inhibit the peroxisome-dependent antiviral cellular response (15), and flavivirus also impairs peroxisomal biogenesis for the suppression of type III IFNs (16).

Porcine epidemic diarrhea (PED) is an acute and highly contagious enteric viral disease characterized by watery diarrhea, vomiting, and dehydration, resulting in high mortality rates in neonatal piglets (17–19). PED was first recognized in England in 1971 but was limited mostly to some parts of Europe until the 1990s, when it appeared in Asia. In 2010, highly virulent PED outbreaks occurred in China (20, 21). In the United States, PED was recognized for the first time in April 2013, and in the following 8 months, it caused more than 8 million deaths of pigs, resulting in substantial economic losses (19, 22–24). The causative agent is PED virus (PEDV), which is an enveloped, positive-sense, single-strand RNA virus grouped in the genus Alphacoronavirus of the family Coronaviridae in the order Nidovirales (http://ictvonline.org/virustaxonomy.asp). The PEDV genome is 28 kb long and codes for two large polyproteins, pp1a and pp1ab; an accessory protein, open reading frame 3 (ORF3); and four structural proteins, spike (S), membrane (M), envelope (E), and nucleocapsid (N) (25). The two large polyproteins are further processed to 16 nonstructural proteins, nsp1 through nsp16, by the proteinase activities of nsp3 and nsp5.

We have previously shown that PEDV inhibits the type I IFN response in cells and that nsp1 is the major type I IFN antagonist (26). The primary target cells for PEDV, however, are intestinal villous epithelial cells of swine and, to some extent, macrophages infiltrating the lamina propria (17, 27, 28). No information is available regarding whether and how PEDV modulates the type III IFN response in intestinal epithelial cells of swine. In the present study, we developed PEDV-susceptible porcine intestinal epithelial cells and showed that in these cells, PEDV suppressed type III IFN production. We further demonstrated that PEDV inhibited IRF1 activation and reduced the number of peroxisomes. Our findings indicate that PEDV evades the IRF1-mediated type III IFN response in porcine intestinal epithelial cells by reducing the number of peroxisomes, which is likely the key mechanism to thwart early antiviral signaling that emanates from these organelles.

RESULTS

Establishment of porcine intestinal epithelial cells (IPEC-DQ) susceptible to PEDV.

Vero cells are widely used for PEDV isolation and propagation (29, 30), and ST (swine testicle) and PK-15 (porcine kidney) cells also support PEDV replication (31). We previously described MARC-145 (a subline of MA-104 African green monkey kidney cells) and LLC-PK1 (porcine kidney) as alternative cell lines that are permissive to PEDV infection (26). Due to the different degrees of cytotoxicity with transfection and variable efficiencies of infection by PEDV, different cell types were used for different experiments. However, the primary target cells for PEDV in pigs are villous epithelial cells of the intestinal tract (17, 27, 28). Thus, Vero, MARC-145, LLC-PK1, and ST cells may be less suitable for studies of cell-virus interactions, especially innate immunity in the intestinal epithelia, and need to be validated in a porcine intestinal epithelial cell model. IPEC-J2 is a line of porcine intestinal epithelial cells established from the jejunum of a colostrum-deprived neonatal pig at 12 h of age. Some researchers reported that IPEC-J2 cells were susceptible to PEDV (32, 33). However, we and others found that these cells were not susceptible to PEDV (34). IPEC-J2 cells were nonhomogeneous, and as shown by immunofluorescence assays (IFAs), only a few cells appeared to be PEDV positive. We subcloned IPEC-J2 cells by limited serial dilutions and were able to obtain a homogeneous cell population, designated IPEC-DQ. These cells were first examined for their characteristics. Cytokeratin is the intracytoplasmic cytoskeleton protein of epithelial tissues, and pancytokeratin monoclonal antibody (MAb) staining showed the abundant expression of cytokeratins in IPEC-DQ cells, as in LLC-PK1 and IPEC-J2 cells (Fig. 1A). Sucrase-isomaltase is a transmembrane glycoprotein located in the brush border of the small intestine, and IPEC-DQ cells, as well as IPEC-J2 cells, showed expression of sucrase-isomaltase, whereas it was absent in LLC-PK1 kidney epithelial cells. The growth kinetics of IPEC-DQ cells was similar to that of IPEC-J2 cells, with a similar doubling time of approximately 53 h (data not shown). Based on these data, we concluded that IPEC-DQ retained the porcine intestinal epithelial cell phenotype. For examination of the susceptibility of IPEC-DQ cells to PEDV, cells were infected, and at 24 h postinfection (hpi), IFAs and Western blot (WB) analyses were conducted by using a virus-specific antibody (Ab). Specific fluorescence was observed in PEDV-infected IPEC-DQ cells, indicating the efficient replication of PEDV in these cells. In contrast, IPEC-J2 cells did not show any PEDV-specific fluorescence even at a multiplicity of infection (MOI) of 5 (Fig. 1B). Western blot analysis further confirmed the efficient and productive infection of IPEC-DQ cells by PEDV (Fig. 1C).

FIG 1.

Efficient replication of PEDV in IPEC-DQ porcine intestinal epithelial cells. (A) IPEC-DQ cells retain the intestinal epithelial phenotype. LLC-PK1, IPEC-J2, and IPEC-DQ cells were grown on coverslips and fixed with 4% paraformaldehyde in PBS at 4°C overnight. The cells were then permeabilized with cold methanol for 15 min at −20°C. Immunostaining was conducted to detect the presence of pancytokeratin (epithelial cell marker) and sucrase-isomaltase (intestinal cell differentiation marker). (B) Expression of the viral nucleocapsid (N) protein in PEDV-infected IPEC-DQ cells. IPEC-J2 and IPEC-DQ cells were infected with PEDV at MOIs of 5 and 1, respectively. Cells were then fixed at 24 hpi with 4% paraformaldehyde and stained with mouse anti-N MAb. (C) Expression of N protein in PEDV-infected IPEC-DQ cells detected by Western blotting. IPEC-J2 and IPEC-DQ cells were infected with PEDV at MOIs of 5 and 1, respectively, and Western blotting was conducted by using anti-N MAb at 24 hpi. (D) Induction of IFN in IPEC-DQ cells. Cells were stimulated with poly(I·C) for 12 h and harvested to determine the activation of type I and type III IFNs by RT-qPCR. The relative IFN induction levels were normalized to those of IFN-α.

Recent studies suggest that IFN induction and signaling in intestinal epithelial cells are unique, and type III IFNs play a key role in maintaining the antiviral state in the gut (14, 35, 36). Intestinal epithelial cells in mice have been shown to express a high level of IFN-λ, but not IFN-α/β, after reovirus infection (14). Thus, to examine whether IPEC-DQ cells also produced IFN-λ selectively, the induction levels of different types of IFNs were determined. After stimulation with poly(I·C), a >1,700-fold increase in the level of IFN-λ3 was observed in IPEC-DQ cells compared to IFN-α (Fig. 1D), suggesting that IPEC-DQ cells produced type III IFNs efficiently. Our results showed that IPEC-DQ cells were susceptible to PEDV infection and express type III IFNs, demonstrating that IPEC-DQ is a suitable cell model for the study of innate immunity possibly modulated by PEDV in the gut.

Antiviral activities of type III IFNs against PEDV.

For examination of the antiviral activities of type III IFNs against PEDV, cells were primed for 12 h with either IFN-λ1 or IFN-λ3 and inoculated with PEDV. Cells were washed once and replenished with fresh medium containing the respective IFN-λs. Cell culture supernatants were collected at 12, 18, 24, 36, and 48 h postinfection, and viral titers were determined for each collection. The PEDV titers were reduced by IFN-λ1 and IFN-λ3 over time, indicating that these IFNs inhibited PEDV production (Fig. 2A). IFN-λ3 had stronger antiviral activity against PEDV than did IFN-λ1. For evaluation of whether this IFN-λ-mediated antiviral activity was dose dependent, cells were pretreated with 10, 100, and 1,000 ng/ml of IFN-λ1 or IFN-λ3 for 12 h and infected with PEDV for 24 h in the presence of the respective IFNs. The treatment of cells with IFN-λ inhibited PEDV infection in a dose-dependent manner: the higher the concentration of IFN-λ, the stronger the inhibition of PEDV replication (Fig. 2B). The immunofluorescence signal for the PEDV N protein was decreased relative to the increased levels of IFN-λ1 and IFN-λ3 (Fig. 2C), further confirming that type III IFNs restricted PEDV infection in a dose-dependent manner. The anti-PEDV activity of type III IFNs was examined under three different treatment conditions. Cells were treated with IFN-λ1 or IFN-λ3 at 100 ng/ml 12 h before, at the time of, or after PEDV inoculation, and the culture supernatants were collected at 24 h postinfection for virus titration. Both IFN-λ1 and IFN-λ3 inhibited PEDV replication at comparable levels regardless of the different treatment conditions, with the highest level of inhibition being found for treatment before infection (Fig. 2D). Reverse transcription-quantitative PCR (RT-qPCR) was conducted for analysis of PEDV N gene expression in PEDV-infected cells, and both IFN-λ1 and IFN-λ3 reduced the amount of viral N mRNA in a dose-dependent manner (Fig. 2E) and under all three conditions (Fig. 2F). These data confirm that type III IFNs possessed anti-PEDV activities.

FIG 2.

Type III IFNs exert antiviral activity against PEDV. (A) Suppression of type III IFNs during PEDV infection. MARC-145 cells were seeded into 12-well plates and treated with 100 ng/ml of IFN-λ1 or IFN-λ3 for 12 h. The cells were then infected with PEDV at an MOI of 1 for 1 h, washed, and replenished with fresh infection medium containing the indicated subtypes of IFN-λ. Cell culture supernatants were collected at 12, 18, 24, 36, and 48 h postinfection and titrated for TCID₅₀. (B and C) Restriction of PEDV infection by type III IFNs is dose dependent. MARC-145 cells were seeded into 12-well plates and treated with different amounts of IFN-λ1 or IFN-λ3 (10 ng/ml, 100 ng/ml, and 1,000 ng/ml) for 12 h. The cells were then infected with PEDV at an MOI of 1 for 1 h, washed, and replenished with fresh infection medium containing the indicated IFNs. (B) Cell culture supernatants were collected at 24 h postinfection and titrated for TCID₅₀. (C) Cells were fixed for immunostaining with anti-N MAb and visualized under a fluorescence microscope. (D) Inhibition of PEDV infection by type III IFNs under three different treatment conditions. MARC-145 cells were pretreated with IFN-λ1 or IFN-λ3 (100 ng/ml) for 12 h and then inoculated with PEDV at an MOI of 1, washed, and replenished with fresh infection medium without IFN (Before). MARC-145 cells were coincubated with PEDV and IFN-λ1 or IFN-λ3 (100 ng/ml) for 1 h, washed, and replenished with fresh infection medium without IFN (During). Alternatively, cells were first inoculated with PEDV for 1 h and treated with IFN-λ1 or IFN-λ3 (100 ng/ml) during incubation for 12 h (After). (E and F) Relative expression levels of the nucleocapsid (N) mRNA in PEDV-infected cells determined by RT-qPCR. The experiments for panels B and D form a pair, and the cells were harvested for RT-qPCR to determine PEDV N mRNA transcript levels. β-Actin was used as an internal control for normalization. The mean values with standard deviations were obtained from three independent experiments, each in triplicate. Asterisks indicate statistical significance relative to the nontreatment group. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Inhibition of type III IFN production by PEDV.

Recent studies have shown that type III IFNs play a critical role in maintaining the antiviral state of the epithelium in the gut (14, 35, 36). We also showed that they inhibited PEDV infection (Fig. 2). Thus, we hypothesized that PEDV might have evolved to modulate the type III IFN response of the host. To investigate this possibility, we first assessed the expression of type III IFNs in PEDV-infected IPEC-DQ cells. Cells were inoculated with PEDV at an MOI of 5, and cell lysates were prepared at various times postinfection for RT-qPCR. Transfection with poly(I·C) as a positive control induced the strong induction of all three subtypes of IFN-λ (IFN-λ1, IFN-λ3, and IFN-λ4), with the highest level of production being found for IFN-λ3 (Fig. 3A to C). We were not able to detect IFN-λ2 under these conditions, suggesting the possibility of an IFN-λ2 deficiency in swine. Infection of IPEC-DQ cells with PEDV caused a slight induction of all subtypes of IFN-λ at 3 hpi and quickly downregulated their induction at 9 hpi and 12 hpi. We also evaluated type III IFN induction in LLC-PK1 cells. At an MOI of 0.01, no apparent induction was noticed in these cells throughout the observation period. Transcript levels of each subtype of IFN-λ were also negligible (data not shown). At a high MOI of 5, type III IFNs were still not inducible in LLC-PK1 cells, and the levels of the respective mRNAs were also negligible (data not shown), suggesting that the suppression of type III IFN responses by PEDV was not limited to a particular cell type. For examination of whether PEDV could induce the activation of type III IFN at any time of infection, cells were transfected with the IFN-λ1 luciferase reporter and inoculated with PEDV to determine virus regulation of IFN-λ1 activity. IFN-λ1 promoter-driven luciferase activity was barely detectable in PEDV-infected cells, compared to the intense activity in poly(I·C)-stimulated cells (Fig. 3D), suggesting that PEDV infection suppressed IFN-λ1 expression.

FIG 3.

Inhibition of type III IFN production by PEDV. (A to C) PEDV does not induce type III IFN production in intestinal epithelial cells. IPEC-DQ cells were infected with PEDV at an MOI of 5, and total cellular RNA was prepared at the indicated times postinfection to determine the type III IFN mRNA level by RT-qPCR. Uninfected cells stimulated with poly(I·C) for 12 h were used as a positive control. (D) PEDV does not activate the IFN-λ1 promoter. MARC-145 cells were grown in 12-well plates and cotransfected with p-55λ1-(−225/−36)-Luc and pRL-TK for 6 h, followed by PEDV infection at an MOI of 1. Cells were harvested at 3, 6, 9, and 12 h postinfection for dual-luciferase reporter assays. Mock-infected cells transfected with poly(I·C) for 12 h were used as a positive control. (E) PEDV inhibits the IFN-λ1 promoter. MARC-145 cells were grown in 12-well plates and transfected with p-55λ1-(−225/−36)-Luc for 6 h, followed by PEDV infection. The gene-transfected and PEDV-inoculated cells were then stimulated with poly(I·C) for 12 h, and cell lysates were prepared for dual-luciferase reporter assays. (F to H) PEDV suppresses the induction of type III IFN mRNA as determined by RT-qPCR. LLC-PK1 cells were grown in 12-well plates for infection for 12 h and then stimulated with poly(I·C) for 12 h. Total cellular RNA was prepared to determine type III IFN mRNA levels by RT-qPCR. Results were obtained from three independent experiments, each in triplicate, and are presented as mean values with standard deviations. Asterisks indicate statistical significance. ***, P < 0.001.

Since PEDV did not induce a type III IFN response during infection, IFN-λ1 reporter assays were performed with virus-infected, poly(I·C)-stimulated cells. Cells were transfected with the luciferase reporter and inoculated with PEDV, followed by stimulation with poly(I·C). IFN-λ1 activity in mock-infected cells was upregulated up to 125-fold after poly(I·C) stimulation. In sharp contrast, IFN-λ1 activity was significantly downregulated in virus-infected cells even after poly(I·C) stimulation (Fig. 3E), indicating virus-mediated suppression of the type III IFN response. For confirmation of IFN-λ suppression by PEDV, RT-qPCR was conducted for IFN-λ transcripts in PEDV-infected, poly(I·C)-stimulated cells. Poly(I·C) stimulation caused the upregulation of IFN-λ1, IFN-λ3, and IFN-λ4 expressions in mock-infected cells, as anticipated; however, PEDV infection significantly downregulated their expressions (Fig. 3F to H).

Identification of viral antagonists of type III IFN.

For the identification of viral proteins antagonizing the type III IFN response, individual PEDV proteins were examined for IFN-λ suppression by reporter assays. The expression levels of individually cloned viral proteins were first examined by Western blotting (Fig. 4A) using anti-FLAG antibody, as described previously (26). All genes were expressed well, as anticipated. Poly(I·C) stimulation upregulated IFN-λ1 transcription in glutathione S-transferase (GST)-expressing or empty-vector-transfected cells, as expected, and showed that cells recognized double-stranded RNA (dsRNA) and activated the IFN-λ1 promoter. Of the viral nonstructural proteins, nsp1, nsp3, nsp5, nsp8, nsp14, nsp15, and nsp16 were shown to downregulate IFN-λ1 activity (Fig. 4B), and the accessory protein ORF3 also inhibited IFN-λ1 activity. For the structural proteins, E, M, and N were identified as type III IFN antagonists (Fig. 4C).

FIG 4.

Identification of PEDV nsp1 as a potent type III IFN antagonist. (A) Expression of individually cloned PEDV proteins in gene-transfected cells. Each gene was transfected into HeLa cells, and the protein expression level was determined by Western blotting using anti-FLAG antibody. Transfection was conducted as described previously (26). Numbers indicate the respective nonstructural proteins representing nsp1 through nsp16. (B and C) Suppression of IFN-λ1 promoter activity by individual PEDV proteins. HeLa cells were grown in 24-well plates and cotransfected with p-55λ1-(−225/−36)-Luc and different PEDV genes along with pRL-TK at a ratio of 1:1:0.1. At 12 h posttransfection, cells were stimulated with poly(I·C) (0.5 μg/ml) for 12 h, and the luciferase activities were measured. (D) Dose-dependent inhibition of IFN-λ1 promoter activity by PEDV nsp1. HeLa cells were grown in 24-well plates and cotransfected with p-55λ1-(−225/−36)-Luc along with pRL-TK at a ratio of 1:0.1 with increasing amounts of the PEDV nsp1 gene. At 12 h posttransfection, cells were stimulated with poly(I·C) (0.5 μg/ml) for 12 h, and the luciferase activities were measured. (E to G) Suppression of type III IFN production by PEDV nsp1. LLC-PK1 cells were grown in 12-well plates and transfected with the empty vector pXJ41 or the PEDV nsp1 gene for 12 h. Cells were then stimulated with poly(I·C) for 12 h. Total cellular RNA was prepared for RT-qPCR to determine mRNA levels of IFN-λ1 (E), IFN-λ3 (F), and IFN-λ4 (G). Results were normalized by using β-actin mRNA and are expressed as relative changes in mRNA levels compared to those in cells without stimulation. Data were obtained from three independent experiments, each in triplicate. Asterisks indicate statistical significance. ***, P < 0.001.

Among the viral antagonists of type III IFN, nsp1 appeared to be one of the most potent suppressors (Fig. 4B). We examined whether the suppression of IFN-λ1 by nsp1 was dose dependent. Cells were cotransfected with p-55λ1-(−225/−36)-Luc and increasing amounts of the nsp1 gene, and luciferase assays were performed. The luciferase activities decreased as the amounts of nsp1 increased (Fig. 4D), indicating dose-dependent inhibition of IFN-λ1. To confirm the suppression of type III IFN by nsp1, RT-qPCR was conducted. While poly(I·C) alone upregulated IFN-λ1, IFN-λ3, and IFN-λ4 expressions, nsp1 inhibited their expressions significantly (Fig. 4E to G). Taken together, these data led to the conclusion that PEDV encoded multiple antagonists for the suppression of type III IFN, and nsp1 is the most potent viral antagonist.

Suppression of type III IFN is mediated by interference of IRF and NF-κB.

The expression of IFN-λ is dependent on the activation of the transcriptional factor IRF1, IRF3, IRF7, or NF-κB. A cluster of the IRF-binding and NF-κB-binding sites has been identified as the cis-acting elements for IFN-λ expression (8). For the determination of the transcription factors required for PEDV-mediated IFN-λ1 inhibition, a series of promoter mutants was used for reporter assays. The wild-type reporter p-55λ1-(−225/−36)-Luc contains the binding sites for both IRF and NF-κB for IFN-λ1 expression. The reporter mutants p-55λ1mut.IRF-Luc, p-55λ1mut.NF-κB-Luc, and p-55λ1mut.IRF/mut.NF-κB-Luc contain the binding-site mutations for IRF, NF-κB, and both IRF and NF-κB, respectively. Cells were transfected with individual reporter mutants and infected with PEDV, followed by stimulation with poly(I·C) and the determination of luciferase activities. Although the mutation of the IRF-binding or NF-κB-binding site abrogated IFN-λ1 activation compared to the activity of the wild-type promoter (Fig. 3E), the mutants still retained the ability to activate IFN-λ1 when stimulated with poly(I·C) (Fig. 5A). PEDV infection, however, suppressed the luciferase activity of the mutants. These results indicate that both IRF and NF-κB are involved in PEDV-mediated IFN-λ suppression. The expression of type III IFNs in intestinal epithelial cells is dependent mainly on IRF1 activation, whereas IRF3 and IRF7 play a minimal role (11). We thus examined whether PEDV suppression of the type III IFN response was due to IRF1 interference. Poly(I·C) stimulation of cells upregulated IRF1 activity, as anticipated. In virus-infected cells, however, IRF1 activity was significantly inhibited (Fig. 5B).

FIG 5.

Suppression of IFN-λ1 is mediated via IRF and NF-κB interference. (A) Suppression of IRF- and NF-κB-mediated IFN-λ1 activity by PEDV. MARC-145 cells were transfected with p-55λ1mut.IRF-Luc, p-55λ1mut.NF-κB-Luc, or p-55λ1mut.IRF/mut.NF-κB-Luc for 6 h and infected with PEDV at an MOI of 1 for 12 h. Cells were then stimulated with poly(I·C) for 12 h, and cell lysates were prepared for dual-luciferase reporter assays. (B) PEDV blocks IRF1 activation. MARC-145 cells were transfected with pIRF1-Luc for 6 h and infected with PEDV at an MOI of 1 for 12 h. Cells were then stimulated with poly(I·C), and cell lysates were prepared for dual-luciferase reporter assays. (C) Identification of IRF1 antagonists. HeLa cells were grown in 24-well plates and cotransfected with pIRF1-Luc, along with individual PEDV genes and pRL-TK at a ratio of 1:1:0.1. At 12 h posttransfection, cells were stimulated with poly(I·C), and the luciferase activities were measured. (D) Inhibition of IFN-λ1 activity is mediated by IRF and NF-κB interference. HeLa cells were grown in 24-well plates and cotransfected with p-55λ1-(−225/−36)-Luc or its mutant (p-55λ1mut.IRF-Luc, p-55λ1mut.NF-κB-Luc, or p-55λ1mut.IRF/mut.NF-κB-Luc), along with the PEDV nsp1 gene. At 12 h posttransfection, cells were stimulated with poly(I·C), and luciferase activities were measured. The reporter experiments were repeated three times, each time in triplicate. Asterisks indicate statistical significance using the empty vector pXJ41 as a control. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

To determine if other viral antagonists for IFN-λ also suppressed IRF1 activity, reporter assays were conducted. Strong luciferase activity was observed in cells transfected with the empty vector or in GST-expressing cells, as anticipated, when stimulated with poly(I·C) (Fig. 5C). In contrast, IRF1 activity was suppressed by nsp3, nsp5, nsp14, nsp15, nsp16, ORF3, E, and M in addition to nsp1 even after poly(I·C) stimulation. Of the total of 11 viral IFN-λ antagonists, nsp8 and N did not inhibit IRF1 activity, suggesting that these two proteins may function through pathways other than the IRF1-mediated pathway. nsp1 was also examined for its ability to suppress IRF-mediated and NF-κB-mediated IFN-λ1 responses. Indeed, PEDV nsp1 suppressed the activities of the IFN-λ1 promoter mutants (Fig. 5D), indicating that nsp1 inhibited type III IFNs by suppressing both IRF1 and NF-κB.

Inhibition of RIG-I/MDA5 and suppression of IRF1.

Since PEDV nsp1 appeared to inhibit IRF1 activity, we first examined IRF3 signaling. IRF3-mediated type III IFN signaling was activated when cells were transfected with a constitutively activated adaptor protein. Cells were cotransfected with p-55λ1-(−225/−36)-Luc, the nsp1 gene, and a plasmid expressing activated RIG-I, MDA5, MAVS, TANK-binding kinase 1 (TBK1), IκB kinase ε (IKKε), or IRF3 along with the pRL-TK internal control, and their reporter activities were determined. The transfection of such active molecules upregulated IFN-λ1 promoter activities (Fig. 6A to E). In contrast, these activities were downregulated when PEDV nsp1 was expressed. These results indicated that PEDV nsp1 possessed the ability to suppress IRF3-dependent type III IFN expression. PEDV nsp1 also inhibited the IFN-λ1 promoter activated by IRF3 (Fig. 6F), which suggests that IRF3-mediated type III IFN suppression might be a nuclear event.

FIG 6.

Inhibition of RIG-I/MDA5-mediated IFN-λ1 and IRF1 activities by nsp1. (A to F) Inhibition of RIG-I/MDA5-mediated IFN-λ1 activation by nsp1. HeLa cells were cotransfected with an active form of RIG-I (A), MDA5 (B), MAVS (C), TBK1 (D), IKKε (E), or IRF3 (F), along with the nsp1 gene and the p-55λ1-(−225/−36)-Luc reporter at a ratio of 1:1:0.1 for 24 h. Cell lysates were prepared to measure luciferase activities. (G to I) PEDV nsp1 inhibits RIG-I/MDA5-mediated IRF1 activation. HeLa cells were cotransfected with an active form of RIG-I (G), MDA5 (H), or MAVS (I), along with the nsp1 gene and the IRF1-Luc reporter at a ratio of 1:1:0.1 for 24 h for luciferase assays. Data were obtained from three independent experiments, each in triplicate, and are presented as the mean values with standard deviations. Asterisks indicate statistical significance. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

PEDV nsp1 degrades CREB-binding protein (CBP) in the nucleus, thus resulting in the suppression of type I IFN production (26). IRF3-mediated type III IFN production also relies on enhanceosome formation (11), and thus, CBP degradation by nsp1 in the nucleus may also affect type III IFN suppression. However, IRF1-mediated type III IFN production can be independent from enhanceosome assembly (10). We thus investigated whether nsp1 inhibited RIG-I/MDA5-mediated IRF1 activation. Transfection with an activated form of RIG-I or MDA5 upregulated IRF1 activity, and viral nsp1 counteracted this activity (Fig. 6G and H). The overexpression of activated MAVS also upregulated IRF1 activity, and nsp1 suppressed IRF1 activity (Fig. 6I). Since MAVS in peroxisomes is crucial for the activation of IRF1-mediated IFN-λ production (11), our findings suggest that the suppression of IRF1-mediated type III IFN production occurred downstream of MAVS activation.

Inhibition of IRF1 nuclear translocation by PEDV and nsp1.

To investigate the basis of the inhibition of the IRF1-mediated type III IFN response by PEDV, the expression kinetics of IRF1 was examined in virus-infected cells. IPEC-DQ cells were infected with PEDV at an MOI of 5, and Western blotting was conducted to investigate IRF1 expression at different times. The expression of the PEDV N protein indicated productive infection. The expression level of IRF1 was not altered by PEDV infection (Fig. 7A). The IRF1 protein contains a functional nuclear localization signal (NLS) located immediately downstream of the DNA-binding domain and is usually hidden, which results in IRF1 being distributed diffusely in the cytoplasm. When stimulated, IRF1 is instantly activated and translocates to the nucleus (37). To examine whether PEDV blocked IRF1 nuclear translocation, IPEC-DQ cells were infected with PEDV for 12 h and stimulated with poly(I·C), followed by antibody staining for IRF1. IRF1 became translocated to the nucleus when stimulated by poly(I·C) (Fig. 7B, second panel, yellow arrow). Similarly, in PEDV-infected cells, IRF1 remained in the cytoplasm (Fig. 7B, third panel), indicating that PEDV infection failed to activate IRF1 nuclear translocation. Furthermore, IRF1 in virus-infected cells remained in the cytoplasm even after poly(I·C) stimulation (Fig. 7B, bottom panel, white arrow), indicating that IRF suppression by PEDV was a cytoplasmic event. In IPEC-DQ cells, IRF1 was typically diffused throughout the cell, but poly(I·C) stimulation rendered IRF1 in the nucleus (Fig. 7C, second panel, yellow arrow). The expression of IRF1 and its cellular distribution, however, were not altered in nsp1-expressing cells (Fig. 7C, third panel), and when nsp1-expressing cells were stimulated, the amount of IRF1 in the nucleus was significantly reduced (Fig. 7C, bottom panel, white arrow). These data suggest that IRF1-mediated type III IFN antagonism by nsp1 occurred upstream of IRF1 activation.

FIG 7.

Inhibition of IRF1 nuclear translocation by PEDV and nsp1. (A) IPEC-DQ cells grown in 12-well plates were infected with PEDV at an MOI of 5 and lysed at the indicated times for detection of IRF1 and the PEDV N protein by Western blotting. β-Actin was used as a loading control. (B) PEDV blocks IRF1 nuclear translocation stimulated by poly(I·C). IPEC-DQ cells were grown on coverslips and infected with PEDV for 12 h, followed by stimulation with poly(I·C) for 12 h. Cells were then fixed and stained with anti-N antibody and anti-IRF1 antibody for 1 h. Alexa Fluor 594-conjugated goat anti-rabbit antibody and Alexa Fluor 488-conjugated goat anti-mouse secondary antibody were used to visualize IRF1 and N, respectively. Nuclei were stained with DAPI. Yellow arrows indicate IRF1 in the nucleus, and white arrows indicate IRF1 in the cytoplasm. (C) Inhibition of IRF1 nuclear translocation by PEDV nsp1. IPEC-DQ cells grown on coverslips were transfected with the PEDV nsp1 gene for 24 h and then stimulated with poly(I·C) for 12 h. Cells were then fixed and stained with anti-FLAG antibody and anti-IRF1 antibody for 1 h. Alexa Fluor 594-conjugated goat anti-rabbit antibody and Alexa Fluor 488-conjugated goat anti-mouse secondary antibody were used to visualize IRF1 and nsp1, respectively. Yellow arrows indicate IRF1 in the nucleus after stimulation in the absence of nsp1 expression. White arrows indicate IRF1 in the cytoplasm of nsp1-expressing cells.

Reduction of the number of peroxisomes in PEDV-infected cells.

MAVS localizes to both mitochondria and peroxisomes for the coordinated activation of an effective innate response (12, 13). The mitochondrial and peroxisomal MAVS pathways result in different but complementing responses: mitochondrial MAVS is associated with the activation of a stable response with delayed kinetics (type I IFN dependent), whereas peroxisomal MAVS is responsible for a rapid but short-lived antiviral response (type III IFN dependent) (12). Peroxisomal MAVS in intestinal epithelial cells preferentially triggers the IRF1-mediated expression of IFN-λ genes (11). The differentiation of intestinal epithelial cells upregulates the biogenesis of peroxisomes and promotes a robust type III IFN response to invading pathogens, and thus, adequate amounts of peroxisome-associated MAVS are essential for the type III IFN response. For examination of the peroxisomes in PEDV-infected cells, IPEC-DQ, LLC-PK1, and MARC-145 cells were infected with the virus, and the numbers and morphologies of peroxisomes were examined. Peroxisomal membrane protein 70 (PMP70) is one of the major components of the peroxisomal membrane during proliferation (38), and thus, it was used as a peroxisomal marker. In mock-infected cells, peroxisomes were abundantly distributed throughout the cytoplasm (Fig. 8A). In contrast, much smaller numbers of peroxisomes were observed in PEDV-infected MARC-145, LLC-PK1, and IPEC-DQ cells. One hundred fifty PEDV-infected cells were randomly chosen, and the average number of peroxisomes per cell was calculated. The results showed 30 to 50% reductions in the numbers of peroxisomes in PEDV-infected cells (Fig. 8B). This indicates that PEDV reduces the number of peroxisomes in PEDV-infected cells.

FIG 8.

Numbers of peroxisomes are decreased in PEDV-infected cells. (A) Reduction in the number of peroxisomes in PEDV-infected cells. MARC-145, LLC-PK1, and IPEC-DQ cells grown on coverslips were infected with PEDV for 24 h and incubated with anti-PMP70 (peroxisome marker) antibody and anti-N antibody for immunostaining. (B) Relative numbers of peroxisomes in PEDV-infected cells. The number of peroxisomes in PEDV-infected cells was compared to that in noninfected cells. One hundred fifty PEDV-infected cells were randomly chosen, and the number of peroxisomes per cell was quantified. Relative numbers are shown as relative percentages from three independent experiments. Asterisks indicate statistical significance. ***, P < 0.001.

Conserved residues of nsp1 are crucial for reduction of the number of peroxisomes and IRF1-mediated IFN-λ suppression.

Since PEDV nsp1 inhibited IRF1 by activated MAVS (Fig. 6I), nsp1 might be the protein that reduces the number of peroxisomes. To investigate this, IPEC-DQ cells were transfected with the nsp1 gene or its mutants and stained for PMP70. The peroxisomes were abundant and normally distributed throughout the cytoplasm in control cells (Fig. 9A, top panel), whereas in nsp1-expressing cells, the number of peroxisomes was decreased (Fig. 9A, second panel). Previously, we determined the predicted tertiary structure of nsp1 and reported that the conserved residues were critical for NF-κB regulation (39). The cellular distribution of nsp1 mutants in IPEC-DQ cells was consistent with that in HeLa cells (Fig. 9A). The K70A mutant was normally distributed and retained the ability to reduce the number of peroxisomes (Fig. 9A, third panel). The Δ37–51 deletion mutant was localized in the cytoplasm and also reduced the number of peroxisomes (Fig. 9A, fifth panel). The perinuclear N93A/N95A mutant (Fig. 9A, fourth panel) and other mutants, G38A/F39A, F44A, G87A, L98A/E99A/E100A, and Δ37–75, showed punctate patterns in the nucleus and did not interfere with the number of peroxisomes (Fig. 9A, 6th to 10th panels). The numbers of peroxisomes in nsp1- and mutant nsp1-expressing cells were quantified, and the data indicate that the conserved residues of nsp1 were crucial for the reduction of the number of peroxisomes (Fig. 9B).

FIG 9.

Conserved residues of nsp1 are crucial for reductions in the numbers of peroxisomes and IFN-λ suppression. (A) Conserved residues of nsp1 are critical for reductions in the numbers of peroxisomes. IPEC-DQ cells were grown on coverslips and transfected with the PEDV nsp1 gene or one of its mutants for 24 h. Cells were then stained with anti-PMP70 (peroxisome marker) and anti-FLAG antibodies for 1 h. Alexa Fluor 594-conjugated goat anti-rabbit antibody and Alexa Fluor 488-conjugated goat anti-mouse secondary antibody were used to visualize peroxisomes and nsp1, respectively. Nuclei were stained with DAPI. (B) Relative numbers (percentages) of peroxisomes in nsp1-expressing cells. The number of peroxisomes in nsp1-expressing cells was compared to that in pXJ41-transfected cells. One hundred fifty nsp1-expressing cells were randomly chosen, and numbers of peroxisomes were quantified. (C) Expression levels of wild-type PEDV nsp1 and its mutants in gene-transfected cells. HeLa cells were cotransfected with either p-55λ1-(−225/−36)-Luc or IRF1-Luc with each mutant of nsp1. After stimulation with poly(I·C), cell lysates were prepared for Western blotting with anti-FLAG antibody for individual mutants of nsp1. Lanes 1 through 13 represent the empty vector pXJ41; wild-type nsp1; and the G38A/F39A, F44A, T68A, K70A, T68A/K70A, G87A, N93A/N95A, L98A/E99A/E100A, L101A, Δ37–75, and Δ37–51 mutants, respectively. β-Actin was used as a loading control. (D and E) Conserved residues in nsp1 are crucial for IFN-λ suppression. HeLa cells were cotransfected with p-55λ1-(−225/−36)-Luc (D) or IRF1-Luc (E), along with an nsp1 mutant. Cells were then stimulated with poly(I·C), and cell lysates were prepared to measure the luciferase activities. Data were obtained from three independent experiments and are presented as the mean values with standard deviations. Asterisks indicate statistical significance compared with the relative luciferase value of nsp1. **, P < 0.01; ***, P < 0.001.

To further identify the residues for nsp1-mediated IFN-λ suppression, reporter assays were conducted by using IFN-λ1 and IRF1-luciferase constructs. Cells were cotransfected with individual nsp1 constructs along with p-55λ1-(−225/−36)-Luc or pIRF1-Luc and then stimulated with poly(I·C) for reporter assays. nsp1 and its mutants exhibited comparable expression levels (Fig. 9C). Among the nsp1 mutants, T68A, K70A, T68A/K70A, and L101A appeared to inhibit IFN-λ1 expression (Fig. 9D). The Δ37–51 deletion mutant localized in the cytoplasm but retained the IFN-λ1 suppression activity, which confirms that the suppression of type III IFNs by nsp1 is a cytosolic event. The cellular distribution of N93A/N95A was mostly perinuclear and reverted the suppression of type III IFNs. The cellular distributions of the G38A/F39A, F44A, G87A, L98A/E99A/E100A, and Δ37–75 mutants were punctate in the nucleus, and their activity for the suppression of type III IFNs was significantly reverted (Fig. 9D), suggesting that the highly conserved residues of nsp1 were crucial for IFN-λ suppression. Furthermore, the G38A/F39A, F44A, G87A, N93A/N95A, L98A/E99A/E100A, and Δ37–75 mutants reverted their IRF1 suppression activities (Fig. 9E), suggesting that the mutated amino acid residues were crucial for the IRF1-mediated suppression of type III IFN production.

DISCUSSION

Antiviral responses are mediated primarily by type I IFNs (IFN-α/β), leading to the expression of hundreds of ISGs for the establishment of an antiviral state. Recent evidence, however, shows that type III IFNs play a critical role in innate antiviral immunity in intestinal epithelial cells in the gut. Peroxisomes are the signaling platforms for innate antiviral activity, and peroxisomal MAVS triggers type III IFN production for rapid ISG expression independent of type I IFNs (12). Peroxisomal MAVS is essential for the IRF1-mediated type III IFN response (11), which plays a pivotal role in restricting viral infections in epithelial cells at mucosal surfaces (40). PEDV is an enteric virus causing severe diarrhea in pigs, but virtually no information is available as to whether and how the virus modulates the type III IFN response in intestinal epithelial cells. We first developed a subline of porcine intestinal epithelial cells (IPEC-DQ). These cells confer productive infection by PEDV and serve as an excellent cell model for porcine enteric viruses. The IFN-λ proteins can inhibit PEDV replication in a dose-dependent manner, and thus, they are potent antiviral cytokines. In turn, PEDV inhibits type III IFN production, and this inhibition is mediated by interference with IRF and NF-κB. PEDV reduces the number of peroxisomes for the suppression of IRF1-mediated type III IFN production. The conserved residues of PEDV nsp1 appear to be crucial for the reduction in the number of peroxisomes and, thus, the suppression of IRF1-mediated type III IFNs.

For humans, type III IFNs constitute four subtypes: IFN-λ1 (IL-29), IFN-λ2 (IL-28A), IFN-λ3 (IL-28B), and IFN-λ4 (2–5). For swine, IFN-λ1 and IFN-λ3 were described previously (6), and in the present study, we have identified IFN-λ4 gene expression in porcine intestinal cells by RT-qPCR (Fig. 3), implicating the importance of the type III IFN system for swine as well. We have not been able to identify IFN-λ2 in IPEC-DQ cells. The antiviral function of type III IFNs was initially thought to be analogous to that of type I IFNs when they were discovered. It appears, however, that type III IFNs exert antiviral activity mainly in epithelial cells at the mucosal surfaces via the induction of ISGs by the activation of the JAK-STAT signaling pathway (40). To activate the JAK-STAT pathway, type I IFNs signal through a heterodimeric IFNAR receptor consisting of IFNAR1 and IFNAR2, whereas type III IFNs signal through a distinct heterodimer of the IFNLR receptor consisting of IFNLR1 and IL-10R2. IL-10R2 is widely distributed in different cell types, whereas IFNLR1 is mostly restricted to epithelial cells (9), which renders type III IFNs that selectively exert their antiviral functions on epithelial cells. The receptors for type I IFNs are expressed in all nucleated cells. However, their expression levels are minimal in IECs, and thus, type I IFNs fail to protect these cells from enteric viral infections (14). The gut mucosa is equipped with the compartmentalized IFN system in which intestinal epithelial cells respond mainly to type III IFNs that are produced after viral infection, whereas other cells rely mostly on type I IFNs for antiviral defense (14). The cell tropism of enteric viruses determines the potency of the antiviral activities of the different types of IFNs in the gut. Rotavirus infects mainly IECs (41), and thus, infection is controlled exclusively by type III IFNs (36, 42). Reovirus replicates in both epithelial and nonepithelial cells (43), and thus, type I and type III IFNs cooperatively control reovirus infection such that type III IFNs restrict infection in epithelial cells, whereas type I IFNs exert antiviral activity on nonepithelial cells (14). Recombinant swine type III IFNs have been shown to restrict PEDV infection in both Vero and IPEC-J2 cells (44, 45). We have also shown that type III IFNs restrict PEDV infection in a dose-dependent manner (Fig. 2), which further confirms the strong anti-PEDV activities of type III IFNs. The swine type III IFNs also have greater antiviral activities against PEDV than does IFN-α, which is probably because type III IFNs can induce differential and more potent ISG expression than IFN-α in intestinal epithelial cells (44). In our study, IFN-λ3 restricts PEDV infection more efficiently than does IFN-λ1 (Fig. 2), which is consistent with the previously reported finding that IFN-λ3 exhibits more potent antiviral activity than does IFN-λ1 (44, 46). The potent antiviral activities of type III IFNs indicate therapeutic applications of these cytokines to enteric viral infections.

Type III IFNs exhibit potent antiviral activities against many viruses, particularly viruses that infect epithelial cells of the respiratory tract, blood-brain barrier, liver, urogenital tract, and gastrointestinal tract (40). Given the potent antiviral activity of type III IFNs, viruses have evolved to encode viral antagonists for the evasion of type III IFN responses. Treatment with recombinant IFN-λ1 inhibits the replication of foot-and-mouth disease virus (FMDV), and in turn, FMDV has evolved to interrupt the antiviral function of IFN-λ1 (47). FMDV infection does not induce type III IFNs, and several viral antagonists have been identified to inhibit IFN-λ1 expression (47). Among them, the leader proteinase L^pro is the most potent viral antagonist and disrupts the activation of IRF and NF-κB. The catalytic activities of the proteinase and the SAP (SAF-A/B, Acinus, and PIAS) domain are crucial for L^pro-mediated IFN-λ1 suppression (47). Yaba-like disease virus (YLDV) is a primate poxvirus that exclusively infects the skin. The secreted glycoprotein Y136 of YLDV binds to IFN-α/β for the inhibition of IFNAR signaling and also binds to type III IFNs to inhibit the antiviral response (48, 49). Similarly, we have reported that PEDV suppresses the type III IFN response, suggesting a potential therapeutic application of type III IFNs to the treatment and prevention of PED. PEDV codes for many antagonists of type I IFNs (26). Due to the similar induction and signaling pathways between the type I and type III IFNs, PEDV may use similar evasion strategies. Indeed, many of the type III IFN antagonists also suppress type I IFN production (Fig. 4), suggesting common mechanisms between IFN-α/β and IFN-λs. The type I IFN antagonists of PEDV inhibit the IRF3-mediated and NF-κB-mediated pathways (26, 50, 51), and we have shown here that PEDV also inhibits IFN-λ expression by interfering with IRF1 and NF-κB (Fig. 5). PEDV nsp1 suppresses IRF3-mediated type I IFN production by degrading CREB-binding protein in the nucleus (26) and also inhibits IRF3-mediated IFN-λ1 activation (Fig. 6). These findings suggest that nsp1-mediated CBP degradation in the nucleus may also result in the suppression of the type III IFN response.

Similar to type I IFNs, type III IFNs also require IRF3, IRF7, and NF-κB for induction (8, 9). However, the induction of type III IFNs does not need all the components of the enhanceosome (10). Epithelial cells preferentially induce the production of IFN-λ over IFN-α/β. Viruses may thus selectively cause type III IFN production. Hepatitis B virus (HBV) infection of hepatocytes induces the production of IFN-λ but not IFN-α/β (52). HCV infection also preferentially induces the production of IFN-λ over IFN-α/β (53). We have found that the induction of type III IFNs in IPEC-DQ cells is much more potent than that of type I IFNs (Fig. 1), which confirms that intestinal epithelial cells abundantly induce type III IFNs upon stimulation. This may be due to the relative abundance of peroxisomes in intestinal epithelial cells. The activation of mitochondrial MAVS induces type I IFN production, whereas the activation of peroxisomal MAVS predominantly induces IFN-λ production (11). IRF1 is the first identified IRF but does not induce the production of type I IFNs (54). IRF1 is also an ISG that is a prominent antiviral effector against many viruses (55). IRF1 can be induced by viral infections and translocates to the nucleus to activate the antiviral response (56). Porcine IRF1 does not induce type I IFN production but inhibits the replication of swine viruses, including swine influenza virus, pseudorabies virus, and transmissible gastroenteritis virus (TGEV) (56). IRF1 is dispensable for type I IFN production but plays a crucial role in regulating the MAVS-dependent signaling of peroxisomes for type III IFNs (11, 12). Peroxisomes are the signaling platforms for IRF1-dependent IFN-λ production. The differentiation process of IECs upregulates the biogenesis of peroxisomes for the robust response of type III IFNs (11), which highlights the importance of type III IFNs for the innate antiviral state in intestinal epithelial cells. Viruses may target IRF1 activation, peroxisomal MAVS activation, or peroxisomal biogenesis for the suppression of type III IFNs. HCV NS3-4A can traffic to and localize in peroxisomes and specifically cleaves peroxisomal MAVS and releases it into the cytosol (15). This action is intense in inhibiting the peroxisome-dependent antiviral response and dampens type III IFN responses. In flavivirus-infected cells, the peroxisomes are reduced in number and redistributed, resulting in the suppression of type III IFNs (16). Capsid proteins of West Nile virus (WNV) and dengue virus (DENV) bind to the Pex19 peroxisome biogenesis factor and impair the biogenesis of peroxisomes. The level of the peroxisomal matrix enzyme is also reduced in infected cells (16). PEDV also decreases the number of peroxisomes for the evasion of the IRF1-mediated type III IFN response in intestinal epithelial cells (Fig. 8). This finding suggests that the targeting of peroxisomes is likely a key strategy for virus evasion of type III IFN responses. PEDV nsp1 blocks IRF1 nuclear localization (Fig. 7C) and reduces the number of peroxisomes (Fig. 9A), indicating that targeting of peroxisomes is crucial for the nsp1-mediated type III IFN response. The conserved residues are essential for the nsp1-mediated suppression of NF-κB (39), IFN-λ (Fig. 9D), and IRF (Fig. 9E), which suggests that the conserved structures are crucial for nsp1-mediated innate immune evasion by PEDV.

Peroxisomes are intracellular organelles that play a central role in regulating various metabolic activities in mammalian cells. They act in concert with mitochondria to control the metabolism of lipids and reactive oxygen species. Most RNA viruses replicate in cellular membrane organelles or rely on them for entry and assembly. Coronaviruses and arteriviruses replicate in double-membrane vesicles (DMVs) and are assembled and released through membrane trafficking, which is associated with the endoplasmic reticulum (ER) and the Golgi complex. Despite the absence of evidence indicating that mammalian viruses replicate in peroxisomes, some viruses interact with peroxisomes. The influenza virus NS1 protein has been shown to bind to the peroxisomal enzyme 17-β-hydroxysteroid dehydrogenase (57), which may alter peroxisome-specific lipid metabolism for viral replication (58). The rotavirus spike protein VP4 contains peroxisome-targeting sequences, which may also be associated with lipid metabolism or cholesterol formation for viral infectivity (59). The HIV Nef protein interacts with thioesterase, which is the peroxisomal matrix protein (60, 61). This interaction is associated with the downregulation of CD4 in HIV-infected cells (62, 63). The interactions between viruses and peroxisomes may dampen the innate antiviral signaling of peroxisomes. In our study, PEDV and nsp1 reduce the number of peroxisomes for the suppression of type III IFN production. PEDV replication and the nsp1 protein may impair the biogenesis of peroxisomes by a direct interaction with essential peroxisome biogenesis factors, such as Pex19. It is also possible that PEDV and the nsp1 protein trigger the disassembly of peroxisomes by disrupting the peroxisomal structure. This disassembly can be visualized by electron microscopy, and it will be of interest to examine structural disassembly. The identification of cellular proteins that interact with nsp1 in peroxisomes and other viral factors that affect antiviral signaling from peroxisomes will also be interesting.

Coronavirus nsp1 is the first N-terminal cleavage product of the pp1a and pp1a/b polyproteins (64). It is one of the most divergent viral proteins in four different genera and does not harbor any known cellular functional motifs (65). Only alphacoronaviruses and betacoronaviruses code for nsp1, whereas gammacoronaviruses and deltacoronaviruses do not contain the nsp1 gene (64, 66–68). For betacoronavirus nsp1, the regulation of host cell and viral gene expressions has been documented, and nsp1 is also a potent IFN antagonist. Severe acute respiratory syndrome coronavirus (SARS-CoV) nsp1 inhibits reporter gene expression under constitutive and IFN-β promoters (69) and employs a two-pronged strategy to inhibit host gene expression. It tightly associates with the 40S ribosomal subunit for the inhibition of mRNA translation and the induction of endonucleolytic RNA cleavage of cellular mRNAs (70, 71). Murine hepatitis coronavirus (MHV) nsp1 affects cellular gene expression by suppressing IFN-β, IFN-stimulated response element (ISRE), and the simian virus 40 (SV40) promoter (72). Middle East respiratory syndrome coronavirus (MERS-CoV) nsp1 inhibits host gene expression by selectively targeting mRNAs transcribed in the nucleus for translation inhibition and mRNA degradation but spares exogenous mRNAs introduced directly into the cytoplasm or virus-like mRNAs that originate in the cytoplasm (73). However, nsp1 proteins in alphacoronaviruses and betacoronaviruses lack overall sequence similarity, and neither conserved motifs nor conserved domains are found in viruses of the Alphacoronavirus genus. Interestingly, the domains in SARS-CoV nsp1 responsible for the suppression of host gene expression and type I IFN production are absent in PEDV nsp1 (70, 74). Thus, it is plausible that nsp1 of alphacoronaviruses may have distinct functions in regulating the host innate immune response and gene expression. TGEV, a porcine alphacoronavirus, strongly inhibits host protein synthesis without binding to the 40S ribosomal subunit in a cell-free in vitro translation system (75). However, this translational suppression does not contribute to IFN suppression because TGEV induces the rapid and massive production of type I IFN both in vitro and in vivo (76, 77). SARS-CoV nsp1 potently inhibits host gene expression universally, including the endogenous protein β-actin (74). For PEDV, however, the suppression of protein translation by nsp1 may not be a universal event since the expression level of endogenous β-actin is unchanged after virus infection and nsp1 gene transfection (26, 39). Thus, it remains to be examined whether PEDV nsp1 also has functions similar to those in other coronaviruses for shutting off host gene translation and whether these biological activities contribute to the regulation of the IFN response.

In summary, we show that PEDV suppresses IRF1-mediated type III IFN production and that this suppression is correlated with decreased numbers of peroxisomes in swine intestinal epithelial cells. PEDV codes for many antagonists of type III IFNs and IRF1. PEDV nsp1 inhibits the RIG-I/MDA5-mediated activation of type III IFNs and IRF1 activities. Our data provide novel information on innate immune evasion by PEDV in intestinal epithelial cells, which may form the foundation for the future development of effective therapies and the control of PED.

MATERIALS AND METHODS

Cell culture.

Porcine kidney epithelial cells (LLC-PK1), derived from a healthy male pig 3 to 4 weeks of age (78), were obtained from K. Chang (Kansas State University, Manhattan, KS). LLC-PK1 cells were maintained in modified Eagle's medium (MEM; Corning Cellgro) with 5% heat-inactivated fetal bovine serum (FBS; Gibco) at 37°C in a humidified atmosphere of 5% CO₂. African green monkey kidney epithelial cells (MARC-145) (79) and Vero cells (ATCC CCL-81) were grown in high-glucose Dulbecco's modified Eagle's medium (DMEM; Corning Cellgro) supplemented with 10% heat-inactivated FBS. IPEC-J2 cells were obtained from A. Blikslager (North Carolina State University, Raleigh, NC). IPEC-J2 is a continuous line of epithelial cells derived from the jejunum of a 12-h-old, colostrum-unsuckled, mixed-breed piglet (80). These cells were maintained in DMEM–F-12 medium with 5% FBS supplemented with 5 mg/ml of insulin-selenium-transferrin (Life Technologies) and 5 ng/ml of epidermal growth factor (Life Technologies). IPEC-DQ cells, a subline of IPEC-J2 cells, were maintained in RPMI 1640 supplemented with 10% FBS.

Virus stocks, titration, and infection of cells.

The Colorado strain of PEDV (USA/Colorado/2013; GenBank accession no. KF272920) was obtained from the Agricultural Research Service, U.S. Department of Agriculture (Ames, IA). PEDV was propagated in FBS-free DMEM supplemented with 0.3% tryptose phosphate broth (Sigma, St. Louis, MO) and 0.02% yeast extract (Teknova, Hollister, CA) with different concentrations of trypsin 250 (Sigma-Aldrich, St. Louis, MO). LLC-PK1 and Vero cells were used for PEDV propagation and titration. The optimal trypsin concentrations for Vero, IPEC-J2, MARC-145, LLC-PK1, and IPEC-DQ cells were 5 μg/ml, 5 μg/ml, 2 μg/ml, 1 μg/ml, and 1 μg/ml, respectively. For the anti-PEDV activity of IFN-λ1 and IFN-λ3, cell culture supernatants were collected at the indicated times postinfection and titrated according to the 50% tissue culture infective dose (TCID₅₀) protocol. The viral titers were calculated by using the Spearman-Karber equation.

Antibodies and chemicals.

The following antibodies were used for IFAs and Western blot (WB) analyses: mouse anti-PEDV N MAb (catalog no. SD-1-5; Medgene, Brookings, SD) (1:1,000 dilution for WB and 1:200 dilution for IFA), mouse anti-β-actin MAb (C-4) (catalog no. sc-47778; Santa Cruz) (1:2,000 for WB), rat anti-FLAG Ab (catalog no. 200474; Agilent Technologies) (1:2,000 for WB and 1:200 for IFA), rabbit anti-IRF1 polyclonal antibody (PAb) (C-20) (catalog no. sc-497; Santa Cruz) (1:1,000 for WB and 1:200 for IFA), rabbit anti-PMP70 PAb as a peroxisomal membrane marker (catalog no. NBP187258; Novus Biologicals) (1:200 for IFA), mouse antipancytokeratin MAb (C-11) (catalog no. MA5-12231; ThermoFisher Scientific) (1:50 for IFA), mouse anti-sucrase-isomaltase MAb (A-12) (catalog no. sc-393424; Santa Cruz) (1:50 for IFA), and Alexa Fluor 594-conjugated (goat anti-rabbit) and Alexa Fluor 488-conjugated (goat anti-mouse) secondary antibodies (Thermo Scientific) (1:200 for IFA).

Recombinant IL-29/IFN-λ1 protein (catalog no. 1598IL025) and recombinant IL-28B/IFN-λ3 protein (catalog no. 5259IL025) were purchased from Bio-Techne (Minneapolis, MN). Poly(I·C) and DAPI (4′,6-diamidino-2-phenylindole) were purchased from Sigma (St. Louis, MO). A QIAamp Viral RNA minikit and an RNeasy minikit were purchased from Qiagen. Power SYBR green PCR master mix was purchased from Life Technologies (Carlsbad, CA). The Pierce ECL Western blotting substrate, methanol, and Triton X-100 were purchased from Thermo Scientific (Waltham, MA). Fluoromount-G mounting medium was purchased from Southern Biotech (Birmingham, AL).

Plasmids and transfections.

Plasmids expressing individual proteins of PEDV were constructed by cloning genes as a fusion with a FLAG tag under the control of the cytomegalovirus (CMV) promoter as described previously (26). The PEDV nsp1 mutants were made by PCR-based site-directed mutagenesis and described previously (39). The expression of each viral protein was confirmed by immunofluorescence and Western blot analyses. Firefly luciferase-based reporter assays were described previously (26). The wild-type IFN-λ1 luciferase reporter construct [p-55λ1-(−225/−36)-Luc] and its mutants (p-55λ1mut.IRF-Luc, p-55λ1mut.NF-κB-Luc, and p-55λ1mut.IRF/mut.NF-κB-Luc) were described previously (8). Plasmid pIRF1-Luc contained multiple copies of the IRF1 cis-acting enhancer element for measurement of IRF1 transcriptional activity and was purchased from Thermo Scientific (Waltham, MA). Renilla luciferase plasmid pRL-TK (Promega) contained the herpes simplex virus thymidine kinase (HSV-tk) promoter and was included to serve as an internal control. The activated stimulator pMAVS was obtained from J. Shisler (University of Illinois, Urbana, IL), and pMDA5 and pIKKε were kindly provided by B. Gotoh (Shiga University of Medical Science, Shiga, Japan). Activated pIRF3 was obtained from Y. Fang (Kansas State University, Manhattan, KS). Constitutively activated pRIG-I and pTBK1 were provided by S. Xiao (Huazhong Agricultural University, Wuhan, China). Transfections were performed by using Lipofectamine 2000 according to the manufacturer's instructions (Invitrogen, Carlsbad, CA).

RNA isolation and quantitative PCR.

Cells were washed once with phosphate-buffered saline (PBS) and lysed directly in cell culture dishes. Total RNA was isolated by using the RNeasy minikit according to the manufacturer's instructions (Qiagen). Possible contamination of cellular DNA was removed by treatment with DNase I. One microgram of total RNA was transcribed to cDNA by using Moloney murine leukemia virus (M-MLV) reverse transcriptase and random primers (Promega). The synthesized cDNA was subjected to quantitative PCR using SYBR green PCR mix according to the manufacturer's instructions (Life Technologies) with the ABI 7500 real-time PCR system. The RT-qPCR primers are listed in Table 1. The β-actin gene was used as an internal control for each experiment. Specificity was confirmed by sequencing of PCR products and melting-curve analysis for qPCR. The threshold cycle (C_T) values for target genes and the differences in their C_T values (ΔC_T) were determined. Relative transcription levels of target genes were presented as fold changes relative to the respective controls by using the 2^−ΔΔC_T threshold method (81).

TABLE 1.

Real-time PCR primer sets for cytokine genes used in this study

Gene	Forward primer (5′–3′)	Reverse primer (5′–3′)	Reference or GenBank accession no.
IFN-λ1	GGTGCTGGCGACTGTGATG	GATTGGAACTGGCCCATGTG	47
IFN-λ3	ACTTGGCCCAGTTCAAGTCT	CATCCTTGGCCCTCTTGA	FJ853389.1
IFN-λ4	GCTATGGGACTGTGGGTCTT	AGGGAGCGGTAGTGAGAGAG	NC_010456.4
PEDV N	GCAGCAAGTGTCCTAAAGAAGCA	GCTTGGGTTCTGCACAGATCT	KF272920
β-Actin	ATCGTGCGTGACATTAAG	ATTGCCAATGGTGATGAC	26

Dual-luciferase reporter assay.

To examine the activation of the IFN-λ1 promoter and its mutants during PEDV infection, MARC-145 cells were grown in 24-well plates to 80% confluence and transfected with the luciferase reporter and pRL-TK at a ratio of 1:0.1 for 6 h. Cells were then infected with PEDV at an MOI of 1 for 12 h, followed by stimulation with poly(I·C) for 12 h. For screening of viral antagonists of IFN-λ1, HeLa cells were grown in 48-well plates to 80% confluence and transfected with a luciferase reporter, the individual viral gene, and pRL-TK at a ratio of 1:1:0.1 for 12 h by using Lipofectamine 2000. To ensure similar protein expression levels of all the viral proteins, transfection was conducted as described previously (26), and the protein expression level was validated for each luciferase assay. Of note, three times more plasmids were transfected for the nsp3 and nsp16 genes to ensure comparable levels of protein expression. Cells were stimulated with 0.5 μg/ml of poly(I·C) for 12 h. For examination of IFN-λ singling by nsp1, cells were transfected for 24 h with luciferase reporters, the individual viral gene, the activator, and pRL-TK at a ratio of 1:1:1:0.1. Transfection was performed by using Lipofectamine 2000 according to the manufacturer's instructions (Invitrogen). For dual-luciferase assays, cells were lysed in 50 μl of passive lysis buffer for 20 min at room temperature (RT) with gentle shaking and placed on ice. Twenty microliters of the cell lysates was used for dual-luciferase assays according to the manufacturer's instructions (Promega). Luciferase values were normalized by using Renilla as the internal control. Results were presented as mean fold change values with standard deviations. Data were obtained from three independent experiments, each in triplicate.

Indirect-immunofluorescence assay and confocal microscopy.

Cells were grown on coverslips placed in 12-well plates to 80% confluence for PEDV infection. For staining of pancytokeratin and sucrase-isomaltase, cells were fixed with 4% paraformaldehyde in PBS overnight at 4°C and permeabilized by using 100% methanol at −20°C for 20 min. For transfection, cells were seeded onto coverslips and grown to 80% confluence. A transfection mix was added to cells for 4 h, followed by replacement with fresh medium, and cells were incubated for 12 h to allow gene expression. Cells were then transfected with poly(I·C) for 12 h and further incubated with fresh infection medium for an additional 12 h. Cells were fixed with 4% paraformaldehyde in PBS overnight at 4°C and permeabilized by using 0.1% Triton X-100 for 15 min at RT. Cells were incubated with 1% bovine serum albumin (BSA) in PBS for 30 min at RT and then incubated with the primary antibody for 1 h. After three washes with PBS, cells were incubated with a fluorochrome-conjugated secondary antibody in the dark for 1 h. The cell nuclei were stained with DAPI for 10 min. The coverslips were mounted onto microscope slides by using Fluoromount-G mounting medium. Fluorescence was visualized by using a Nikon A1R fluorescence microscope. Confocal microscope images were processed by using NIS-Elements analysis software (Nikon).

SDS-PAGE and Western blot analysis.

Cells were lysed at the indicated times in radioimmunoprecipitation assay (RIPA) buffer (20 mM Tris [pH 7.5], 150 mM NaCl, 1 mM EDTA, 1 mM phenylmethane sulfonyl fluoride [PMSF], 0.1% SDS, 0.5% sodium deoxycholate, 1% NP-40) containing a proteinase inhibitor cocktail (Promega) on ice for 30 min. The cell lysates were then sonicated and centrifuged at 4°C at 12,000 rpm (catalog no. 5415R; Eppendorf) for 10 min to remove insoluble components. Protein concentrations were determined by using the Pierce bicinchoninic acid (BCA) protein assay kit (Thermo Scientific). Proteins were resolved by SDS-PAGE and transferred onto an Immobilon-P membrane (Millipore). The membranes were blocked with 5% nonfat dry milk in Tris-buffered saline–Tween (TBST) for 1 h and incubated with the primary antibody at 4°C overnight. The membranes were then incubated with a horseradish peroxidase (HRP)-conjugated secondary antibody for 1 h at RT. Proteins were visualized by using enhanced chemiluminescence detection reagents (Thermo), and images were obtained by using the FluorChem R system according to the manufacturer's instructions (ProteinSimple).

Statistical analysis.

Statistical analyses were conducted by using GraphPad Prism 6 for analysis of variance (ANOVA). Asterisks in figures indicate statistical significance.