Introduction

Sheep colibacillosis is one of the most common bacterial diseases in large-scale sheep farms. The traditional method of controlling the bacterial disease is by antibiotic therapy, although this approach also has several disadvantages. The use of RNA-seq to screen circRNAs that antagonize sheep colibacillosis is the basis for analysing the molecular mechanism of disease resistance in sheep and thus for discovering candidate genes associated with disease resistance traits. Circular RNA (circRNA) is a type of special non-coding RNA (ncRNA), which is a new research hotspot, that is in a RNA family along with microRNA (miRNA) and long non-coding RNA (lncRNA)1. Using electron microscopy, Sanger et al.2 first found the closed circular single-stranded RNA molecules formed by covalent bonds in plant-infected viruses (viroids), which are highly thermostable. In 1990, researchers found that 20S RNA did not have free 5′ and 3′ ends in Saccharomyces cerevisiae, and it was determined by electron microscopy to be a circular RNA molecule3,4. Subsequently, circRNAs were found in the hepatitis D virus5, and circRNAs transcribed from the sex-determining region Y (Sry) were also found in the testes of mice6. It was also confirmed that circRNAs are present in human cells7.

Although circRNAs are widely found in a variety of cells over a long period of time, the study of circRNA has been slow over the past few decades, and the mechanism of gene expression and regulation is not fully understood8. For a long time, circRNAs were identified as by-products of alternative splicing during pre-mRNA processing. However, only a few circRNAs were found to be exons produced during alternative splicing9. Recent studies10 have found that circRNA is not a by-product of mRNA maturation and, like mRNA, is an important product of pre-mRNA processing; circRNA processing machinery competes with mRNA. At the same time, the classical clipping signal and clipping mechanism are also necessary for back-shear11. At present, studies on sheep disease resistance mainly focus on the prevention and treatment of disease12,13, but the important molecular mechanisms of disease resistance are seldom reported. In this study, we first screened for circRNAs that were differentially expressed in individuals that were antagonistic and sensitive to E. coli F17 fimbriae by RNA-seq. Miranda was used to analyse the circRNA-miRNA-mRNA interaction to find miRNA target genes and then verified by q-PCR. At the circRNA level, this study deepens our understanding of antagonizing E. coli F17 fimbriae in sheep and at the same time is expected to identify genes that can antagonize E. coli F17 fimbriae and solve key problems for Chinese local sheep breeding against E. coli disease. This study can be used as a foundation and can provide a theoretical basis for formulating a breeding strategy against E. coli in the future.

Results

Identification of transcripts in sheep spleens

After mapping the reference sequence, we identified 7,730 known circRNAs. The length of circRNAs primarily ranges from 200 bp to 900 bp, with an average length of 1943 bp and a GC content of approximately 43.5%. The number of exons of circRNA is mainly 2–4 (Fig. 1). The statistics of the variable shear signal (GT-AG) of the reverse cleavage site in the circRNA sequence were calculated and graphs were drawn (Fig. 2a). The statistics for the circRNA types are shown in Fig. 2b: overlapping circRNA accounts for 92.11%, exonic circRNA accounts for 3.27%, intergenic circRNA accounts for 3.18%, intronic circRNA accounts for 0.88% and antisense circRNA accounts for 0.56%. circRNAs were compared to the genomic elements to explore the distribution of circRNAs in the genome, to count the number of circRNAs predicted on each chromosome or scaffold, and to plot the results (Fig. 2c). It was found that circRNAs were primarily distributed on three chromosomes: NC_019458.2 (853), NC_019459.2 (772), NC_019460.2 (787).

Figure 1
figure 1

Summary of the length, GC content, and number of the exons of the predicted circRNAs. (a) Shows the length of circRNAs, which primarily ranges from 200 bp to 900 bp, with an average length of 1943 bp. (b) Shows a GC content of approximately 43.5%. (c) Shows the number of exons of circRNA, which is mainly 2–4.

Full size image
Figure 2
figure 2

circRNA cleavage site signal statistics, circRNA gene structure distribution and circRNA number distribution in each chromosome or the scaffold. (a) Shows the statistics of the variable shear signal (GT-AG) of the reverse cleavage site in the circRNA sequence. (b) Shows the statistics for the circRNA types, overlapping circRNA accounts for 92.11%, exonic circRNA accounts for 3.27%, intergenic circRNA accounts for 3.18%, intronic circRNA accounts for 0.88% and antisense circRNA accounts for 0.56%. (c) shows the number of circRNAs predicted on each chromosome or scaffold.

Full size image

Analysis and validation of DE transcripts

We used the RPM value to estimate the expression level of circRNA transcripts, and the expression level of circRNA transcripts was low (Fig. 3). We screened 31 up-regulated and 29 down-regulated DE circRNAs (Fig. 4). Differentially expressed circRNAs can be found on Table 1. To further validate the reliability of RNA-seq, 6 DE circRNAs were randomly selected, and their relative expression levels in antagonistic and sensitive lambs were confirmed by q-PCR and found to coincide with our RNA-seq results (Fig. 5), thus indicating that the RNA-seq data are reliable. Our analyses also show that high-throughput sequencing has the advantage of detecting genes with low expression levels.

Figure 3
figure 3

Expression patterns of circRNA transcripts, The Box-whisker Plot consists of five statistics: the minimum, the first quartile (25%), the median (50%), the third quartile (75%), and the maximum.

Full size image
Figure 4
figure 4

Differentially expressed circRNAs in antagonistic and sensitive lambs Note: Gray represents circRNAs that have no significant differences; Red represents significantly upregulated circRNAs; Green represents significantly downregulated circRNAs; Blue indicates that the difference multiple is more than 2 times, but the circRNA is not significant in the difference significance test. The horizontal axis is the display of log2 FoldChange, and the vertical axis is the display of log10 Pvalue.

Full size image
Table 1 Differentially expressed circRNA.
Full size table
Figure 5
figure 5

Relative expression levels of DE circRNAs between antagonistic and sensitive lambs Note: “**” means highly significant correlation; “*” means significant correlation; “ns” or “no SuperiorScript” means no significant correlation. The same as below.

Full size image

GO and KEGG pathway enrichment analyses of DE lncRNAs

A comparison of the DE circRNA and GO databases showed that a total of 60 circRNAs were annotated and classified into 297 functional subclasses. The results showed the oxidation-reduction process (GO: 0055114), transport (GO: 0006810), extracellular region (GO: 0005576), focal adhesion (GO: 0005925), extracellular exosome 0005615), zinc ion binding (GO: 0008270) and seven more subclasses of circRNA functions, while the remaining functional subclass circRNA was less distributed (Fig. 6a). A comparison of the DE circRNA and KEGG PATHWAY databases showed that a total of 60 circRNAs were annotated and grouped into 73 KEGG PATHWAYS. The results showed that there were more circRNAs in three KEGG pathways, including the estrogen signalling pathway (path: ko04915), protein processing in the endoplasmic reticulum (path: ko04141) and regulation of the actin cytoskeleton (path: ko04810). The remaining KEGG pathways have fewer circRNAs (Fig. 6b).

Figure 6
figure 6

Gene Ontology and KEGG pathway enrichment analyses of DE circRNAs.

Full size image

Prediction of target relationship of circRNA-miRNA-mRNA

We used Miranda to predict miRNA-bound circRNA and the target genes of the miRNA. The function of circRNAs was elucidated on the basis of the function of the target genes of miRNAs. The predicted results are shown in Table 2 below. To further validate the relative expression, 9 mRNAs were selected, and their relative expression levels in antagonistic and sensitive lambs were confirmed by q-PCR. Significant differences were found between the two groups (Fig. 7).

Table 2 Prediction of the target relationship of circRNA-miRNA-mRNA.
Full size table
Figure 7
figure 7

Relative expression level of the target genes of miRNAs between antagonistic and sensitive lambs.

Full size image

Discussion

Due to the limited efficiency of traditional molecular biology methods for the detection of circRNAs, circRNA has long been regarded as a product of the abnormal splicing of RNA7. In recent years, with the rapid development of bioinformatics and high-throughput sequencing technology, a large number of circRNAs have been identified in eukaryotes6,14,15, and they may play an important role in the regulation of gene expression9,16,17. It has been found that most circRNAs contain miRNA binding sites that act as efficient competitive endogenous RNAs and efficiently adsorb miRNAs to regulate the target genes of miRNA18,19, which have the following four biological functions: miRNA sponge effect17,18, protein translation template20,21, regulation of gene transcription22,23 and regulation of competition for linear RNA production8,10. However, investigations of lamb diarrhoea relative to circRNAs are limited. Hu sheep are a unique breed with high fecundity and a strong adaptability to warm-wet climates; they can be kept indoors all year round. This study provides the first overview of circRNAs in relation to diarrhoea in sheep, as well an investigation into their possible roles in disease resistance.

In the present study, we found that the expression level of circRNA was very low: 5 miRNAs binding to 6 circRNAs were identified by the Miranda software, and the adjacent genes of 4 circRNAs were identified as Btnl 1, GSTM1, NRAMP2 and B2M.

Btnl 1 is a key suppressor of T-cell activation and immune diseases24. The mechanism of action of Btnl 1 is different from those of Btnl 2 and BtnlA1, which directly inhibit T cell activation through anti-receptor binding24,25,26,27. The study found that the Btnl gene may be a new important local regulator of intestinal inflammation28. GSTM1 encodes the glutathione-S-transferase (GST) M1 enzyme, which is involved in the detoxification of various carcinogens of lung cancer29 and plays a key role in protecting cells from oxidative stress30. NRAMP2 is a metal transporter protein, and in the absence of manganese, NRAMP2 is involved in the regeneration of Mn in the Golgi and promotes plant root growth31. B2M encodes the beta chain of the major histocompatibility complex (MHC) class I molecule and is up-regulated in inflammatory and tumour cells32.

We also used Miranda software to predict the 3 miRNA target genes and verified that they were significantly differentially expressed in antagonistic and sensitive groups, namely, NEB, UBE3B, ADGRF2, LAMA1, LTF, MGAT5, TLN2 and SLC25A29.

NEB encodes nebulin, a large protein component of the cytoskeletal matrix that coexists with myofilaments in skeletal muscle. Mutations in the NEB gene are the most common causes of myotubes, accounting for approximately 50%33. UBE3B is a ubiquitin ligase (UBE3), and its unique combination of E2-binding enzymes provides high substrate specificity, which is required to target specific protein degradation34. ADGRF2 is a member of the adhesion G protein-coupled receptor family and plays an important role in adhesion in the cell-cell and cell-matrix35. LAMA1 mutations may be related to Poretti-Boltshauser syndrome, and studies have shown that LAMA1 deficiency can lead to cytoskeletal changes36. LTF is a member of the transferrin family, whose protein products initiate host defence against a wide range of microbial infections and antigenic activity37. The protein encoded by MGAT5 belongs to the glycosyltransferase family and is one of the most important enzymes involved in the regulation of the biosynthesis of glycoprotein oligosaccharides. Changes in oligosaccharides on cell surface glycoproteins cause significant changes in cell adhesion or migration behaviour; increased enzyme activity is associated with the development of invasive malignancies38. Talin is a large adapter protein that links the integrin family of adhesion molecules to F-actin; Talin 1 is required for integrin-mediated cell adhesion, and TLN2, like Talin 1, is considered to be unique. Transmembrane receptors bind to form new connections between the extracellular matrix and the actin cytoskeleton39. SLC25A29 encodes a nuclear-encoded mitochondrial protein that is a member of the large family of solute transporter family 25 (SLC25) mitochondria. The primary physiological role of SLC25A29 is to introduce basic amino acids into the mitochondria for mitochondrial protein synthesis and amino-acid degradation40.

GO is a bioinformatics tool that is widely used to study the functional relationship of genes. GO and KEGG pathway analyses of 60 DE circRNAs showed that the relevant circRNAs may potentially participate in the process of pili adhesion to intestinal mucosa. However, the role of these pathways in disease resistance remains largely unknown.

We found that a total of 60 circRNAs were significantly differentially expressed between antagonistic and sensitive groups, with 31 up-regulated and 29 down-regulated. In addition, we identified a total of 1,942 new circRNAs in both groups. To further verify the results of RNA-Seq, the expression levels of the six circRNAs were verified by q-PCR, and the results were consistent.

We studied the expression profiles of circRNAs in the spleens of antagonistic and sensitive sheep that developed diarrhoea to further understand their regulatory role in disease resistance in sheep. We found that circRNAs were differentially expressed in spleen tissues of antagonistic and sensitive lambs. Our research may help to determine how lambs resist the mechanism of diarrhoea. In addition, further studies of these circRNAs can provide a scientific basis for lambs to resist diarrhoea.

Methods

Ethics statement

The Institutional Animal Care and Use Committee (IACUC) of the government of Jiangsu Province (Permit Number 45) and Ministry of Agriculture of China (Permit Number 39) approved the animal study proposal. All experimental procedures were conducted in strict compliance with the recommendations of the Guide for the Care and Use of Laboratory Animals of Jiangsu Province and of the Animal Care and Use Committee of the Chinese Ministry of Agriculture. All efforts were made to minimize animal suffering.

Experimental design and sample collection

Experimental sheep were purchased from Jiangsu Xilaiyuan Ecological Agriculture Co., Ltd. in December 2016. A total of 18 three-day-old lambs showing normal growth and roughly similar weight were randomly selected, and all sheep were raised with segregation. To ensure their dietary requirements were met, all sheep were fed with 10% lamb milk powder prior to the experiment. Five-day-old lambs were fed 12.5% lamb milk powder and E. coli F17 bacteria liquid [4.6 × 108 colony-forming units (CFUs)·mL−1]41 and had ad libitum access to drinking water. The stool features42 of the experimental lambs were recorded daily. Lambs that exhibited diarrhoea for two days were classified as antagonistic and sensitive and then euthanized. The intestinal tissues were collected in 4% paraformaldehyde. The liver, spleen, duodenum, jejunum, and ileum of each lamb were collected and immediately frozen in liquid nitrogen until RNA extraction.

Library construction and sequencing

RNA was extracted from the spleen of three individuals per group. A NanoDrop 2000 Ultra Microscope and an Agilent 2100 Bioanalyser (Shanghai, China) were utilized for quality control of the extracted total RNAs (Annex 1). Ribosomal RNA was removed using a Ribo-Zero (TM) kit (Epicenter, Madison, WI, USA). Short fragments (approximately 200 bp in length) were obtained and used as templates for first-stand cDNA synthesis. Second-strand cDNA synthesis was performed using a buffer, dNTPs, RNase H, and DNA polymerase I. After PCR amplification and purification using the Qubit® dsDNA HS Assay Kit, the cDNA library was constructed using an NEBNext® Ultra™ RNA Library Preparation Kit. The cDNA library was sequenced on the Illumina HiSeq2500 platform by Shanghai OE Biomedical Technology Co. (sequencing read length: 150 bp).

Identification of circRNAs

The circBase43 database contains only circRNA sequences of human, mouse, nematode, lagomorph, and coelacanth. Since sheep are not included, we used CIRI44 de novo prediction of circRNA. According to the position of circRNA on the genome, circRNAs can be classified into the following five categories: exonic circRNA, intronic circRNA, antisense circRNA, sense overlapping circRNA, and intergenic circRNA.

Different expression analysis

After obtaining differentially expressed circRNAs, we performed Gene Ontology (GO) and KEGG pathway significance analyses of the source genes. DESeq45 is suitable for experiments with biological duplication. Differential expression analysis can be performed between sample groups to obtain the circRNA list for the difference between the two biological conditions. For experiments without biological duplication, edgeR46 differential expression analysis was used to obtain a list of circRNAs that were differentially expressed between the two samples.

GO and KEGG pathway analyses

After screening for differentially expressed transcripts, functional annotation was performed using GO enrichment analysis. Enrichment analysis involved counting the number of transcripts in each GO term, followed by Fisher’s exact test to assess statistical significance (p < 0.05). KEGG47 is the main public database used in pathway analysis, which was followed by Fisher’s exact test to assess statistical significance (p < 0.05).

circRNA-miRNA-mRNA interaction studies

As a miRNA target molecule, circRNAs are regulated by miRNAs. Because circRNAs contain multiple miRNA binding sites, analysis of circRNA-miRNA interactions can help elucidate the function and mechanism of circRNA acting as a sponge. For animals, we used Miranda48,49 to predict circRNAs that bind to miRNAs and the target genes of miRNA and elucidated the function of this portion of circRNAs based on the functional annotations of miRNA target genes.

Verification of the expression level of DE circRNAs

To verify whether the screened DE circRNAs play a role in the process of antagonism, q-PCR was used to detect the expression levels of DE lncRNAs and target genes of DE miRNAs in lamb spleens between the antagonistic and sensitive groups. The relative expression of each RNA was normalized to that of GAPDH using the 2−ΔΔCt method50, and the primers used in the amplification of the lncRNAs are shown in Table 3.

Table 3 Primers of GAPDH, DE circRNAs, and the target genes of DE miRNAs.
Full size table

Statistical analysis

All data were analysed by SPSS (version 22.0), and the relative expression levels of various transcripts were analysed by one-way ANOVA. Statistical significance was determined when p < 0.05. Each group contained three samples, and each experiment was repeated three times.