MARC-145 cells, which are kidney cells from an African green monkey, were grown in Dulbecco’s modified Eagle medium (DMEM; Hyclone, Logan, UT, USA) with 10% heat-inactivated fetal bovine serum (FBS; PAN-Biotech), and 100-times diluted penicillin and streptomycin (Hyclone). MARC-145 cells were used to validate the effect of miR-615 during PEDV infection. The IECs were cultivated in DMEM-F12 (Hyclone) supplemented with 10% FBS, penicillin, and streptomycin at the same concentrations as indicated above. All cells were kept in an incubator with 5% CO₂ at 37°C as previously described (Wang et al., 2014). PEDV strains CV777 (GenBank accession number KT323979.1) and NW-17 (GenBank accession number MF782686.1) were provided by Nuoweilihua Biotechnology Co., which have an S gene from the epidemic strain group II. Strain NW-17 could infect IECs without trypsin and was therefore selected as an optimal strain for IEC infection. Vero cells were used to prepared the PEDV stock by three cycles of freezing and thawing, and were stored at −80°C.

After 20 min of fixation at a 4:1 ratio of cold acetone and methanol, the cell samples were washed three times in phosphate-buffered saline (PBS). The cells were incubated with a monoclonal antibody (mAb) against the PEDV N or NF-κB p65 protein for 2 h. The cells were then rinsed three times with PBS and incubated for 1 h with a fluorescein isothiocyanate-conjugated AffiniPure Goat Anti-rabbit/Mouse secondary antibody (Sungene Biotech, Tianjin, China). Hoechst 33258 was utilized to stain the nucleus, and the cell samples were analyzed with a laser-scanning confocal microscope after three PBS rinses (Olympus). The experiment was performed at room temperature.

Deep sequencing was carried out at Novogene (Beijing, China). IECs were infected in triplicates for 24 h with CV777, NW-17, or a mock infection at a multiplicity of infection (MOI) of 1. The mock group was treated with PBS. Total RNA was isolated using Trizol reagent (Invitrogen). Nine small RNA libraries (triplicate samples of the control mock-, CV777-, and NW-17-infected groups) were generated for Illumina sequencing. The microarray assay was conducted as previously described (Hallman et al., 2013). The prediction of target genes of DEMiRs was performed using miRanda (John et al., 2004; version 3.3a), PITA¹, and RNAhybrid.² Gene Ontology enrichment analysis was performed to screen the potential functions of the significantly enriched target genes. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis was applied to identify the potential pathways associated with the target genes of the DEMiRs (Kanehisa et al., 2008). p-values were corrected using the Benjamini–Hochberg strategy (Benjamini and Hochberg, 1995). Statistical significance was defined as a corrected p-value <0.05.

Total RNA was extracted from IECs 24 h after infection at an MOI of 1. The Real-time Quantitative PCR Detection System and SYBR PrimeScript™ miRNA RT-PCR Kit were used to perform RT-qPCR. One microliter of each primer, 2.0 μl of diluted cDNA, and 12.5 μl of SYBR Green Premix Ex Taq II were included in each 25 μl reaction mixture. The thermocycling conditions comprised 95°C for 30 s, followed by 40 cycles of 95°C for 5 s and 60°C for 20 s. Each sample was processed in triplicates. Fifteen miRNAs were selected for testing. The expression of the U6 small RNA was used as the reference for data normalization. Table 1 contains the primer sequences for each miRNA and gene.

A miRNA mimics is a chemically synthesized double-stranded RNA identical to a mature miRNA sequence. A miRNA inhibitor is a chemically modified single-stranded RNA that is complementary to the mature miRNA sequenc. After 12 h of culture, MARC-145 cells and IECs were transfected for 24 h using Lipofectamine 3000 (Invitrogen) with mimic control (MC), miR-615 mimics, miR-615 inhibitor (miR-615 inhi; 100 nM), inhibitor control (IC; 100 nM), small interfering RNA (siRNAs; si-IRAK1), or siRNA control (SC; 50 nM), which were synthesized at RiboBio (Guangzhou, China). Subsequently, the cells were infected with PEDV at an MOI of 1.0. After 24 h of infection, the cells were examined for indirect immunofluorescent labeling, or collected for RNA quantification or western blotting.

pCDNA3.1+ plasmid was used to clone the Flag-IRAK1 plasmid at Genecreate using the BamHI and EcoRI restriction sites (Wuhan, China). The miRNA reporter plasmids, WT-pmirGLO-IRAK1 (wild-type, WT), and MuTpmirGLO-IRAK1 (mutant type; MuT) were subcloned into pmirGLO using the Nhel and SalI sites at Genecreate. pNiFty-luc plasmids is composed of a minimal Promoter, five NF-κB repeated transcription factor binding sites and a Luc (Luciferase) reporter gene of mammal.

The potential miR-615 target genes in the 3′-untranslated region (3′-UTR) of IRAK1 were identified using the luciferase vector pmirGLO (Promega, Madison, WI, USA). The wild-type (WT-pmirGLO-IRAK1) or mutant (MuTpmirGLO-IRAK1) 3′-UTR sections of IRAK1 were subcloned into the pmirGLO vector and co-transfected into 293 T cells with miR-615. NF-κB activity was detected using pNiFty-luc and pRL-TK plasmids with the Dual-Luciferase Reporter Assay System (Promega, Madison, WI, USA) according to the manufacturer instructions.

The cells were lysed using radioimmunoprecipitation assay buffer (Beyotime, Shanghai, China) and centrifuged at 4°C for 12,000 rpm. The concentration of the lysate was evaluated using a bicinchoninic acid (Thermo Scientific) protein assay kit. Each sample was diluted with 5Хloading buffer, boiled for 10 min, resolved on a sodium dodecyl sulfate-polyacrylamide gel electrophoresis gel, and then deposited onto polyvinylidene fluoride membranes at equal amounts. The membranes were blocked with 5% skim milk for 1 h and incubated overnight at 4°C with the primary antibodies against β-actin (1:1,000; 5,057; Cell Signaling Technology, Danvers, MA, USA), phospho-NF-κB p65 antibody (1:1,000; 3,033; Cell Signaling Technology), NF-κB p65 (1:1,000; 6,956; Cell Signaling Technology), and MyD88 (1:1,000; NB100-5698SS; RDSC). Proteins were detected using enhanced chemiluminescence detection reagents after being treated with secondary antibodies, HRP-conjugated goat anti-mouse IgG or goat anti-rabbit IgG (Beyotime). All samples were incubated along with β-actin as an internal standard. PEDV anti-nucleocapsid (N) protein antibody (1:1,000) was gifted by Prof. Tong Guangzhi, Shanghai Veterinary Research Institute.

All data were statistically analyzed using Student’s t-test in Graphpad Prism 9.3.1 for analysis of variance (ANOVA).

Villous epithelial cells are the primary target cells of PEDV infection; therefore, IECs are a suitable model for studying virus–cell interactions in the intestinal epithelia (Wang et al., 2014). IECs have been found to be susceptible to certain PEDV strains such as CV777, independent of high concentrations of trypsin (Cao et al., 2015b).

The IFA results of the IECs at 24 h post-inoculation (hpi) with CV777 and NW-17 (MOI = 1) suggested significantly higher fluorescence in the CV777- and NW-17-infected groups than that in the mock-infected group (Figure 1A). This observation confirmed that CV777 and NW-17 could infect IECs effectively.

To identify whether miRNAs play a role in virus–cell interactions, we performed high-throughput sequencing to obtain the miRNA profiles of mock-infected or PEDV (CV777 and NW-17)-infected IECs at an MOI of 1 at 24 hpi. In comparison to the mock-infected group, 102 known miRNAs were found to be differentially expressed in the CV777-infected group and 98 known miRNAs were differentially expressed in the NW-17-infected group. Among them, 78 miRNAs were commonly differentially expressed in both infection groups (Figure 1B).

A total of 55 high-abundance miRNAs were selected by setting the read count to >250 (p < 0.05; Figure 1C). Of these miRNAs, 28 were upregulated and 27 were downregulated. To validate the high-throughput results of the DEMiRs, RT-qPCR was used to analyze the expression of 13 DEMiRs that were common to both groups (Figure 1D).

Using the KEGG pathway database, the roles of the DEMiRs in response to the PEDV strains were predicted, showing enrichment in several pathways, including the NF-κB, apoptosis, MAPK, TNF, and TLR signaling pathways, which participate in antiviral activities (Figures 1E,F). The high enrichment scores for the NF-κB and TLR signaling pathways indicated that innate immunity plays an important role in PEDV infection.

The DEMiRs enriched in the NF-κB pathway, miR-615, miR-708, miR-221-5p, and miR-671-5p, were selected for further investigation. Three miRNAs (miR-708, miR-671-5p, and miR-221-5p) were found to exert inhibition on PEDV replication by RT-qPCR. In particular, miR-615 overexpression significantly boosted viral replication compared to that in the MC group (p < 0.01; Figure 1G).

To further evaluate the interaction between miR-615 and PEDV in IECs and MARC-145 cells during PEDV infection, PEDV genomic and miR-615 RNA were detected using RT-qPCR. In IECs, the expression of miR-615 showed a downward trend (p < 0.01; Figure 2A). In MARC-145 cells, miR-615 expression was up-regulated at 12 h and was down-regulated at 24 h (p < 0.001; Figure 2B). The overexpression of miR-615 promoted viral replication in IECs (p < 0.01; Figure 2C) and MARC-145 cells (p < 0.001; Figure 2E). Conversely, the downregulation of miR-615 expression inhibited viral replication compared to that in the IC group in the IECs (p < 0.001; Figure 2D) and MARC-145 cells (p < 0.001; Figure 2F).

Western blotting showed that the level of PEDV N protein expression increased in IECs (Figure 2G) and MARC-145 cells (Figure 2H) after transfection with the miR-615 mimic for 6 and 12 h.

Recent studies suggested that IFN-III has an important effect on the antiviral activity of small IECs (Zhang et al., 2018b). Consistently, we found that PEDV replication could be inhibited by IFN-λ3 (Figure 3A). Therefore, we hypothesized that miR-615 affects the expression of IFN-IIIs or IFN-Is, which in turn enhances viral replication. To test this hypothesis, we assessed the expression of IFN-IIIs in PEDV-infected IECs and of IFN-I in MARC-145 cells.

miR-615 mimics were transfected into the cells and/or poly (I:C) was used to stimulate the cells for 24 h. RT-qPCR was used to examine the expression of the two IFN-III subtypes, IFN-λ1 and-λ3. IFN-λ1 and λ3 expression was upregulated after stimulation with poly (I:C), which served as a positive control; however, the expression levels of IFN-λ1 and -λ3 significantly decreased following miR-615 transfection in IECs (p < 0.01; Figures 3B,C).

We also evaluated IFN-β and -α levels in miR-615-transfected MARC-145 cells. Poly (I:C) stimulation upregulated IFN-β and -α expression, whereas miR-615 transfection significantly downregulated their expression (p < 0.01; Figures 3D,E). The downregulation of IFN-Is and -III expression suggests that miR-615 may affect the IFN pathway in these two types of cells when infected with PEDV.

Based on our preliminary findings, we hypothesized that miR-615 restricts the NF-κB pathway from being activated. The expression of IFNs is induced when NF-κB binds to the positive regulatory domain (PRD) II in the nucleus after translocation. Thus, we further examined NF-κB activity to explore whether the pathway was inhibited by miR-615 or miR-615 inhibitors. The NF-κB reporter luciferase plasmid (pNiFty-luc; containing five PRDII sites), thymidine kinase promoter-Renilla luciferase reporter plasmid (pRL-TK; an internal control plasmid), and miR-615 mimics or inhibitors were co-transfected into IECs, followed by poly (I:C) stimulation; thereafter, cell lysates were collected to detect luciferase activity at 24 hpi. The luciferase reporter assays showed that miR-615 markedly inhibited poly (I:C)-induced PRDII activity in IECs (p < 0.01; Figure 4A). Conversely, miR-615 inhibitors increased PRDII activity (p < 0.01; Figure 4B). Comparable results were observed in MARC-145 cells (Figures 4C,D). These results verified our hypothesis that miR-615 inhibits NF-κB pathway activity in both IECs and MARC-145 cells.

Next, we explored the mechanism of NF-κB inhibition. The IFA results revealed a significant reduction in the nuclear abundance of p65 protein in the miR-615 transfection group as compared to that in the MC-transfected group in both IECs (Figure 5A) and MARC-145 cells (Figure 5C). The abundance of p-p65 was found to decrease following transfection with miR-615 mimics in both the poly (I:C)-induced and non-poly (I:C)-induced groups (Figure 5B). We further examined p-p65 proteins upstream of MyD88 to investigate the mechanism underlying NF-κB downregulation. The results showed that p-p65 expression was downregulated, whereas MyD88 expression was not changed with the transfection of miR-615 in IECs (Figure 5B) and MARC-145 cells (Figure 5D). These results suggested that miR-615 inhibits the NF-κB pathway by repressing p65 nuclear translocation.

To explore the mechanism whereby miR-615 inhibits NF-κB activation, target gene prediction was conducted using miRanda³, PITA⁴, and RNAhybrid⁵, which identified 167 target genes (Supplementary Table S1). The most interesting candidate target gene was IRAK1 as it had the highest prediction score (174) and was enriched in the NF-κB pathway. Therefore, we hypothesized that IRAK1 is a target gene of miR-615.

Two reporter gene plasmids were constructed containing WT and seed region-Mut target sites (MuT) with matching or mutated target gene seed sites of miR-615 in the 3′-UTR of IRAK1 (Figure 6A). miR-615 inhibited the luciferase activity of the WT reporter plasmid when compared with that in the MC-transfected group; however, miR-615 did not repress the activity of the mutated dual-luciferase reporter gene plasmid (Figure 6B). Moreover, transfection with miR-615 mimics downregulated IRAK1 protein expression, whereas transfection with miR-615 inhibitors upregulated its expression (Figure 6C).

To determine whether repression of the NF-κB pathway by miR-615 is dependent on the regulation of IRAK1 expression, we first examined the effect of IRAK1 on the NF-κB pathway using a dual-luciferase reporter assay with overexpression and knockdown of IRAK1 in PEDV-infected IECs stimulated with poly (I:C). The NF-κB pathway was activated by IRAK1 overexpression (Figure 7A) and was downregulated by knockdown of IRAK1 (Figure 7B). In addition, IRAK1 overexpression upregulated IFN-λ1 and -λ3 expression during viral infection (Figures 7C,D).

To further examine whether the effect of miR-615 on the NF-κB pathway was dependent on its regulation of IRAK1 expression, miR-615 and IRAK1 were overexpressed and the effects on the NF-κB pathway activity were examined. Porcine IECs were co-transfected with pNif-TK, pRL-TK, miR-615 mimics, or Flag-IRAK1, and poly (I:C) was used to induce NF-κB pathway activation. The rescue experiment showed that IRAK1 overexpression promoted NF-κB pathway activation and reversed the miR-615-dependent repression of the NF-κB pathway (Figure 7E). Collectively, these data indicate that miR-615 inhibits the NF-κB pathway by targeting IRAK1. The knockdown and overexpression of IRAK1 protein were further confirmed using western blotting (Figure 7F).