Introduction

Porcine deltacoronavirus (PDCoV) is a new porcine enteropathogenic virus. It causes watery diarrhea, dehydration, and high mortality in newborn piglets, which results in significant economic losses to the swine industry [1]. PDCoV is an enveloped RNA virus with a positive-stranded RNA genome that belongs to the genus Deltacoronavirus within the family Coronaviridae [2]. The full genome of PDCoV is approximately 25.4 kb in length and arranged in the following order: 5′ untranslated region (UTR), open reading frame 1a and b (ORF1ab), spike (S), envelope (E), membrane (M), non-structural gene 6 (NS6), nucleocapsid (N), non-structural gene 7 (NS7), and 3′ UTR [2]. PDCoV was initially reported in a molecular surveillance study in Hong Kong (China) in 2012, and then emerged in Ohio (USA) in 2014 [3, 4]. Subsequently, it spread to many countries including Thailand, South Korea, Canada, and China [5,6,7,22,23,24,25,26,43], and then further mapped to Sus scrofa 11.1. The known lncRNAs and mRNAs were obtained by comparing the assembled transcripts with annotated lncRNAs and mRNAs in the reference genome. Then, novel transcripts with length ≥ 200 bp and number of exons ≥2 were further retained from the transcripts that mapped to the reference genome and were located in intergenic regions. Then CNCI and CPC2 software were used to predict the protein-coding potential of novel transcripts. Novel lncRNAs were predicted by removing the novel transcripts that did not pass the protein-coding score test (CPC score < 0, CNCI score < 0), and the remaining transcripts were novel mRNAs. The expression of lncRNAs and mRNAs was calculated by fragments per kilobase of transcript per million mapped reads (FPKM) using RSEM software .

Analysis of small RNA-Seq data

High-quality clean reads were first screened from raw reads by removing the low-quality reads and adapter sequences. Then, the clean reads in a size range from 18 to 30 nt were further screened against the GenBank database (Release 209.0) [44] and Rfam (11.0) database [45] to remove small nucleolar RNA, small nuclear RNA, rRNA, tRNA, repeat sequences, exon and intron sequences. Eventually, the existed miRNAs were identified by searching all of the clean reads in the miRbase database [46] (Release 22). The known miRNAs that were still not included in the miRBase database were identified by aligning clean reads to those existed miRNAs of other species which included in miRBase. To distinguish with the existed miRNAs, x (−x) and y (−y) were used to represent the known miRNAs from the 5′ (−5p) and 3′ (−3p) arms of the miRNA precursor in this study. To identify the novel miRNAs, all of the unannotated clean reads were aligned with the swine reference genome (NCBI genome database: Sus scrofa 11.1) and the novel miRNAs were screened according to the results of hairpin structure prediction which performed by miReap software [47]. For gene identification, clean reads were screened from raw reads obtained from RNA sequencing by removing low-quality reads and reads with adapters or unknown nucleobases. Subsequently, the expressed genes and novel genes were identified by genome map**. Briefly, clean reads obtained from RNA sequencing were mapped to a swine reference genome (NCBI genome database: Sus scrofa 11.1), which was performed by HISAT software following the instructions of the software website (http://www.ccb.jhu.edu/software/hisat/index.shtml). The miRNA expression level was computed and normalized to transcripts per million (TPM) [48].

Screening of differentially expressed lncRNAs, mRNAs and miRNAs

The expression analysis of lncRNA, mRNA and miRNA in the two comparison groups were performed using DEGseq2 (http://www.bioconductor.org/packages/release/bioc/html/DESeq2.html) and edgeR (http://www.bioconductor.org/packages/release/bioc/html /edgeR.html). lncRNAs and mRNAs with a false discovery rate (FDR) value < 0.05 and a │log2 (fold change) │ > 1 and miRNAs with a P < 0.05 and a │log2 (fold change) │ > 1 were identified as differentially expressed ncRNAs or mRNAs.

Prediction of lncRNA and miRNA target genes

The potential target genes of DElncRNAs involved in cis- and trans-regulatory effects were predicted. We screened the mRNAs located 10 kb upstream and downstream of lncRNAs as cis target genes. The trans target genes were predicted by correlation analysis between lncRNAs and mRNAs. According to the FPKM values, the Pearson correlation coefficient between the lncRNAs and the mRNAs was calculated, and the threshold for positive correlation was set to Pearson correlation > 0.999.

Three software programs, TargetScan (Version 7.0), RNAhybrid (Version 2.1.2) and miRanda (Version 3.3a), were used to predict the targets of miRNAs. The genes that had reliable binding sites (minimum free energy (MFE) levels < − 10 kcal/mol, P values < 0.05) to miRNAs and commonly existed in 3 prediction results were considered to be credible targets of miRNAs. Targets involved in the innate immune response were annotated to InnateDB (https://www.innatedb.ca/).

RT-qRCR validation

Six miRNAs were randomly selected for RT-qPCR analysis using Bulge-loop™ miRNA RT-qPCR Primer Sets (RiboBio Inc., China) according to the manufacturer’s instructions. The primers for miRNAs and internal standard U6 were designed by RiboBio Inc. (China), and the sequences are covered by a patent. The RT-qPCR examination of lncRNAs and mRNAs was carried out using TB Green® Premix Ex Taq™ II (Tli RNaseH Plus) (Takara Biomedical Technology, China) following the manufacturer’s instructions. Primers for lncRNAs and mRNAs were designed and are listed in Additional file 10: Table S10. The RT-qPCRs were performed on an ABI7500 StepOnePlus Real-Time PCR System (Thermo Fisher Scientific, USA) and run in duplicate. U6 was used as the internal reference gene for miRNA relative expression validation, and GAPDH was used as the internal reference gene for lncRNA and mRNA relative expression validation. The relative expression level of each miRNA, lncRNA and mRNA was calculated using the 2−ΔΔCt method [49, 50].

Function and pathway enrichment analysis

To investigate the function of lncRNAs and miRNAs in IPEC-J2 cells, which were primarily influenced by PDCoV infection, GO functional enrichment analysis and KEGG pathway enrichment analysis of the predicted targets of DElncRNAs and DEmiRNAs were further performed by using of the GOseq R package and KOBAS software (3.0) [51]. The significantly enriched functional GO terms and pathways are represented in this study (P < 0.05).

Construction of competing endogenous RNA networks

Innate immunity-related lncRNAs, miRNAs and mRNAs were chosen for analysis. lncRNA-miRNA interactions were predicted by miRanda. Based on the results of the innate immunity-related target genes of DElncRNAs and DEmiRNAs as well as the lncRNA-miRNA interactions, visualization of the lncRNA-miRNA-mRNA interaction network was constructed using Cytoscape software (http://www.cytoscape.org/download.php).

Functional validation of miRNA for the regulation of PDCoV replication

Monolayer IPEC-J2 cells were transfected with miRNA mimic fluorescence controls under different concentrations using Lipofectamine 3000 (Thermo Fisher Scientific, USA). After 24 h, the fluorescence of the cells was observed by fluorescence microscopy (Olympus Corporation, Japan). Subsequently, monolayer IPEC-J2 cells were transfected with mimics and inhibitor as well as the controls of ssc-miR-30c-3p and ssc-miR-374b-3p, respectively, at a concentration of 150 nM. After 24 h, PDCoV at a final MOI of 1.0 was then used for adsorption in the transfected IPEC-J2 cells in an incubator at 37 °C in 5% CO2. The cells were respectively harvested at 24 h after transfection and at 12 hpi, and then the total RNA of each sample was extracted by TRIzol. miRNA RT product and cDNA was respectively obtained using the riboSCRIPT Reverse Transcription Kit (RiboBio Inc., China) and the PrimeScript™ RT reagent Kit (Perfect Real Time) (Takara Biomedical Technology, China) in an Applied Biosystems™ Veriti™ Dx 96-well Fast Thermal Cycler (Thermo Fisher Scientific, USA). The relative expression of ssc-miR-30c-3p and ssc-miR-374b-3p in the cells transfected with mimics or inhibitor as well as the controls of ssc-miR-30c-3p and ssc-miR-374b-3p were detected by RT-qPCR using Bulge-loop™ miRNA RT-qPCR Primer Sets (RiboBio Inc., China) according to the manufacturer’s instructions. The primers for miRNAs and internal standard U6 were designed by RiboBio Inc. (China), and the sequences are covered by a patent. The relative expression of the PDCoV M gene in different transfection groups was detected by RT-qPCR using TB Green® Premix Ex Taq™ II (Tli RNaseH Plus) (Takara Biomedical Technology, China). The RT-qPCRs were performed on an ABI7500 StepOnePlus Real-Time PCR System (Thermo Fisher Scientific, USA). Sequences of forward primer and reverse primer for RT-qPCR of PDCoV M gene are listed in Additional file 10: Table S10.