Background

MicroRNAs (miRNAs) belong to a large family of endogenous noncoding RNAs (ncRNAs). MiRNAs may regulate the expression of target genes by binding to complementary regions in their 3′ untranslated regions (3’UTRs) [1, 2]. Many studies have shown that miRNAs play important regulatory roles in animal reproduction [3,4,5,6,7,8,9]. In recent years, miRNAs have also been found to be involved in the regulation of animal seasonal reproduction [10,11,12,2) were annotated and classified by alignment against ncRNAs. Among the diverse sequences of ovine pineal ncRNAs (including miRNAs, rRNAs, sRNAs, tRNAs and other RfamRNAs), the proportion of miRNAs was always the highest in each stage (Fig. 2b), and their values were also similar among different stages. However, the proportions of rRNAs, sRNA and other RfamRNA were relatively higher during anestrus than during the breeding season. In total, 427 miRNAs were identified in the ovine pineal gland, including 91 known miRNAs, 311 conserved miRNAs and 25 predicted novel miRNAs (Fig. 2c). Compared with the two stages (luteal and follicular phases) in the breeding season, expressed miRNAs (including known, conserved and novel miRNAs) were the least abundant in ovine anestrus (Fig. 2d).

Fig. 1
figure 1

Seasonal reproductive characteristics of Tan sheep (A) and ovarian sections of Tan sheep at different sampling stages in this study (B). (A) The anestrus season is usually from April to July for Tan sheep, and the other months are the breeding season. In the breeding season, every estrous cycle is approximately 17 days, including the luteal phase and follicular phase. (B) In ovarian sections from anestrous ewes, no large corpus luteum or follicles were observed (a). An obvious corpus luteum with a diameter of more than 5 mm was observed on the ovary surface at the luteal phase (b), and a large antral follicle was found in the follicular phase (c)

Full size image
Fig. 2
figure 2

Expression characteristics of pineal small RNAs in sheep at different reproductive stages. a Distribution of sequence lengths at different reproductive stages based on the abundance of clean reads. X axis: sequence lengths; Y axis: Percentage of reads number with each length. A: anestrus; L: luteal phase; F: follicular phase. b The composition of RNA classes at different reproductive stages. c The number of expressed miRNAs in the ovine pineal gland, including known, conserved and predicted novel miRNAs. d The number of expressed miRNAs at different reproductive stages

Full size image

Next, the functions of miRNAs that were specifically expressed in anestrus or the breeding season were predicted. KEGG pathway analysis (Additional file 3) showed that the target genes of miRNAs that are expressed specifically in anestrus were predominantly enriched in endocytosis, mTOR and MAPK signaling pathways. These pathways are mainly associated with hormone uptake, protein synthesis, and cell proliferation and differentiation. On the other hand, the target genes of miRNAs that were expressed specifically in the breeding season were predominantly involved in pathways such as the mTOR signaling pathway, apoptosis and axon guidance (Additional file 3). These pathways are mainly associated with protein synthesis, cell growth and death, and the formation of neuronal networks.

Meanwhile, the expression of miRNAs was also ranked in each reproductive stage, and the 20 most highly expressed miRNAs are displayed in Table 1. The results indicated that the top 20 miRNAs were similar between the two stages (luteal and follicular phases) in the breeding season; however, they were significantly different between the breeding season and anestrus. In anestrus, miR-142 (homology ID: aca-miR-5441) was the most abundant miRNA, accounting for 86% of the total expressed miRNA. KEGG analysis showed that the target genes of miR-142 were predominantly enriched in oxidative phosphorylation, glycerolipid metabolism and phosphatidylinositol signaling pathways. In addition to miR-142, high expression of miR-202 (homology ID: tae-miR-5086) and miR-2 (homology ID: cel-miR-51-5p) was also observed during anestrus. Oar-miR-181a, Oar-miR-26a and Oar-miR-143 showed the highest levels of expression in the breeding season. Additionally, Oar-let-7a was highly expressed in all reproductive stages.

Table 1 Top 20 miRNAs expressed in the sheep pineal gland at each reproductive stage
Full size table

Differentially expressed (DE) miRNAs among different ovine reproductive stages and their probable functions in the pineal gland

We determined the DE miRNAs among three reproductive stages (anestrus, follicular phase and luteal phase). The largest number of DE miRNAs was detected between anestrus and the follicular phase (Fig. 3a). Hierarchical clustering of miRNAs (Fig. 3b) also indicated that the differences in miRNA expression between anestrus and the follicular phase were greatest among the three stages analyzed. Furthermore, the majority of the DE miRNAs between anestrus and the two stages of the breeding season overlapped (Fig. 3c). Therefore, these overlap** miRNAs could be considered DE miRNAs between anestrus and the breeding season. To determine the probable biological functions of the overlap** miRNAs, we performed a pathway analysis of the target genes of these miRNAs. The analysis revealed that these miRNAs were mainly enriched in pathways related to protein biosynthesis, secretion and uptake (such as biosynthesis of amino acids, ribosome, cAMP signaling pathway, vascular smooth muscle contraction, axon guidance, dopaminergic synapses, and endocytosis pathway) and the phototransduction pathway (P < 0.01) (Fig. 3d, Additional file 4). Moreover, the results of the transcriptome analysis (Additional file 5) showed that the majority of the target genes in these pathways exhibited differential expression between the seasons. For example, RPLP1, RPLP2, RPL18A, RPL35, RPS5, RPS13 and RPSA in the ribosome pathway showed significantly lower expression levels in the breeding season than in anestrus (Additional file 6).

Fig. 3
figure 3

Outline of differentially expressed miRNAs among different reproductive stages.a Number of differentially expressed (DE) miRNAs detected among three stages. DE miRNAs were identified with the edgeR software package (version: 3.12). b Dendrogram of hierarchical clustering of expressed miRNAs among three reproductive stages. The clustering analysis was performed by pheatmap (v1.0.2). c Venn diagram showing the overlap of differentially expressed miRNAs among three comparisons (A vs. L; L vs. F; A vs. F). d Pathways in which the target genes of differentially expressed miRNAs between anestrus and the breeding season were mainly enriched. The color of the circle represents the P value at which a certain pathway is enriched. X axis: number of differentially expressed genes in the specific KEGG pathway. The KEGG pathways were analyzed by clusterProfiler package (v3.16.0)

Full size image

In addition, the overlap** differentially expressed genes between anestrus and the luteal phase and between anestrus and the follicular phase were also screened out to represent the expression differences in genes between nonbreeding and breeding seasons. The highly expressed genes during anestrus and the breeding season in the pineal gland of sheep are shown in Additional file 7. Some of the highly expressed genes in anestrus were related to protein synthesis and hormone secretion. Highly expressed genes in the breeding season were involved in the ribosome pathway, cAMP signaling pathway and other pathways.

Prediction of important miRNA–target gene pairs

The joint analysis of negatively correlated miRNA–mRNA pairs and miRNA target binding prediction has been demonstrated to be a feasible approach for predicting miRNA-target gene pairs [31, 32]. We therefore measured pineal mRNA profiles at different reproductive stages to examine miRNA–mRNA correlations at the whole-genome scale. Among the negatively correlated pairs, many miRNA-target gene pairs with binding sites were predicted. We first investigated the transcriptional regulatory role of miRNAs on key rate-limiting enzyme genes in melatonin synthesis. The expression of AANAT mRNA showed significant variation at different reproductive stages (Fig. 4). Therefore, the miRNAs that were significantly and negatively correlated with the AANAT expression pattern were predicted and summarized in Table 2. To validate the results from high-throughput sequencing, real-time quantitative PCR was performed for the five miRNAs in Table 2 and the AANAT gene. The results of quantitative PCR (Fig. 4) were consistent with the sequencing data, and these miRNAs exhibited an inverted expression to that of the AANAT gene. Additionally, for the expression of HIOMT mRNA, there was no significant difference among the reproductive stages. Therefore, miRNAs targeting the gene were not predicted in this study. In addition to AANAT, miRNAs potentially targeting several differentially expressed genes in the ribosome pathway were also predicted (Additional file 8).

Fig. 4
figure 4

Results of real-time quantitative PCR for the AANAT gene and five miRNAs. Data are the mean ± SEM from three ewes in each group

Full size image
Table 2 List of miRNAs that are negatively correlated with AANAT expression in the sheep pineal gland
Full size table

Validation of the target relationships between AANAT and candidate miRNAs

To further verify the target relationships between AANAT and candidate miRNAs, one miRNA (miR-89) that was predicted to show a target relationship (completely binding) and one (miR-201) that was predicted to show just partial binding but exhibited a high negative correlation coefficient were selected for validation in vitro. Among the miRNAs with a complete binding relationship, miR-89 was selected for validation because it was previously detected in sheep (defined as Oar-miR-214-3p) and suggested to be involved in the regulation of breeding activities [33]. Next, the 3′ UTR sequence of the AANAT gene in Tan sheep was obtained, and the sequence was consistent with the corresponding region of the reported AANAT mRNA sequence (NM_001009461.1) in a cross of the Dorset × Rambouillet sheep [34, 35]. The predicted target site of miR-89 in the 3’UTR of the AANAT gene is shown in Fig. 5a. The possible binding site of miR-201 with the 3’UTR of AANAT was also predicted by RNAhybrid software. The 2nd and 8th bases at the 5′ end of the miRNA showed semibinding states with the 3’UTR of AANAT (Fig. 5b). Then, the dual luciferase reporter system was used to further verify the targeting role of the miRNAs associated with the 3’UTR of AANAT in vitro. The miRNAs were successfully transfected into 293FT cells, and the efficiency of transfection is indicated in Fig. 6. As shown in Fig. 5c, the relative luciferase activity of luc2/hRluc in the group cotransfected with miR-89 and the wild-type 3′ UTR of AANAT was significantly lower (P < 0.01) than in the group transfected with the wild-type 3′ UTR of AANAT alone or the group cotransfected with miR-89 and the mutant-type 3′ UTR of AANAT. This result verified that miR-89 can target the 3′ UTR of AANAT and that the binding site is unique. Taken together, the results of validation of the target relationship and the observed expression of miRNA and AANAT mRNA during different stages implied that miR-89 participates in the negative regulation of AANAT mRNA expression by targeting its 3′ UTR. However, the data demonstrated that miR-201 had no apparent target effect on the 3′ UTR of AANAT (Fig. 5d). Therefore, the results of in vitro analysis demonstrated that the predicted miRNA–target mRNA pairs were accurate.

Fig. 5
figure 5

Results of prediction and validation of the target relationships between miRNAs and the 3’UTR of the AANAT gene. The target sites of miR-89 (a) and miR-201 (b) in the 3’UTR of the AANAT gene were predicted by RNAhybrid software. The results of validation of the target relationships of miR-89 and miR-201 through the dual luciferase reporter system are shown in (c) and (D). WT + miR-89: cells cotransfected with pmirGLO-AANAT wild-type 3’UTR (0.025 μg) and pcDNA6.2-GW/miRNA-89 (0.075 μg); WT + NE: cells cotransfected with pmirGLO-AANAT wild-type 3’UTR (0.025 μg) and pcDNA6.2-GW empty vector (0.075 μg); MUT + miR-89: cells cotransfected with pmirGLO-AANAT mutant-type 3’UTR (0.025 μg) and pcDNA6.2-GW/miRNA-89 (0.075 μg); MUT + NE: cells cotransfected with pmirGLO-AANAT mutant-type 3’UTR (0.025 μg) and pcDNA6.2-GW empty vector (0.075 μg); blank: cells without transfected vector. The test groups for miR-201 were similar to those for miR-89. **: P < 0.01

Full size image
Fig. 6
figure 6

Efficiency of miRNA transfection in cells. Expression of green fluorescent protein gene in the pcDNA6.2-GW/miRNA recombinant plasmid was detected after transfection for 24 h (magnification: 4×, scale bars = 1000 μm; a: fluorescent detection image, b: bright field image, c: merged fluorescent image). d. Compared with the empty plasmid group, the expression levels of miR-89 were significantly higher in cells with the pcDNA6.2-GW/miR-89 recombinant plasmid

Full size image

Discussion

Many recent studies have shown that miRNAs play important roles in the regulation of reproduction [3,4,5,6,7,8,9], including seasonal reproduction [10,11,12,10]. Genes with an FDR < 0.05 were defined as differentially expressed genes [64].

Based on the 3′-UTR sequences that were extracted according to gene annotation information, miRNAs potentially targeting candidate genes were simultaneously predicted using the miRanda and RNAhybrid algorithms. Pairs of DE miRNAs and genes showing negative correlations were identified using R software. The function of target genes was predicted by KEGG pathway analysis. The enrichment p-values were calculated using a hypergeometric distribution.

Validation of the expression of AANAT and its associated miRNAs

Real-time quantitative PCR (Q-PCR) was used to quantify the expression levels of AANAT mRNA and miRNAs that were randomly selected among those that were negatively correlated with AANAT in sequencing data. The stem-loop primers employed in RT-PCR of miRNAs and the primers employed in Q-PCR of AANAT and miRNAs are summarized in Additional file 9. Reverse transcription and Q-PCR were performed with the same method as in our previous study [10]. The data were analyzed with the 2–ΔΔCt method [65].

Validation of the target relationship between candidate miRNAs and AANAT

The 3′ UTR sequence of the AANAT gene of Tan sheep was obtained using the 3′-Full RACE Core Set with PrimeScript™ RTase kit (TaKaRa) and the specific primers GSP1 (5′-AGCGTCCACTGCCTGAAACC-3′) and GSP2 (5′-AGGTTTGGCTTCCATCCCG-3′). Following sequencing, the potential target site bound by candidate miRNAs was predicted using miRanda and RNAhybrid algorithms. For candidate miRNA-gene pairs, both a miRNA expression vector (pcDNA6.2-GW/EmGFP-miR) and a reporter vector (pmirGLO) in which the AANAT 3’UTR was fused to a firefly luciferase reporter were obtained to perform the dual luciferase reporter assay. The sequence that was inserted in the pcDNA6.2-GW/EmGFP-miR vector was designed based on the mature sequence of the miRNA (Additional file 10) and was synthesized by Invitrogen Trading (Shanghai) Company. The AANAT 3’UTR sequence that was inserted into the pmirGLO vector was synthesized by Shanghai Generay Biotech Company. The two vectors were cotransfected into 293FT cells using Lipofectamine 3000. The luciferase activity of luc2/hRluc in transfected cells will be significantly reduced when the miRNA targets the 3’UTR of the candidate gene. To determine whether there were other target sites besides the predicted one, a mutant-type 3’UTR of AANAT mutated at the predicted target site was also designed.