Background

Toxocara canis is one of the most common causative agents of toxocariasis that is a neglected zoonosis with a global distribution, causing ocular larva migrans (OLM), visceral larva migrans (VLM) or neurotoxocariasis (NT) in humans and animals [1]. The global prevalence of Toxocara infection in humans equals 19%, extrapolating to the global population, approximately 1.4 billion people worldwide were estimated to be infected with or threatened by Toxocara, with 8%, 11%, 13%, 24%, 28%, 34% and 38% in the Eastern Mediterranean region, the Europe, the North Americas, the Western Pacific regions, the South Americas, the South-East Asia and African, respectively [2]. Multiple factors may affect the prevalence of Toxocara infection in different regions, such as socio-ecological, environmental and genetic effects [3]. The global prevalence of T. canis infection in dogs equals 11.1%, and dogs infected with T. canis can continuously excrete eggs into the environment, contaminating water sources and vegetation [4]. When dogs are infected with infectious eggs of T. canis, the larvae can be hatched in the duodenum within 2–4 h. Then the larvae penetrate the intestinal wall and migrate to reach the liver at approximately 24 hpi; subsequently, the larvae continue migration with blood flow, passing the heart and reaching the lung at 96 hpi; the larvae then return to the digestive tract through the trachea and throat; after 36 dpi, the larvae develop into adults in the intestinal tract [5]. T. canis has a complex life cycle in definitive hosts and paratenic hosts, which ensures that T. canis persists in hosts [5].

Bone marrow, an important hematopoietic and immune organ in the body, can produce red blood cells (RBCs), leukomonocytes and various white blood cells, which play indispensable roles in the host’s defense against parasite infection [6]. For example, bone marrow-derived eosinophils with cytotoxic functions can damage the worms that invade the host [7]. Genomics, transcriptomics, proteomics and metabolomics are widely used to explore the biological functions of genes, proteins and small metabolites of T. canis as well as the T. canis-host interactions [8,9,10,11].

miRNAs are a class of short, single-stranded, non-coding RNAs of about 22 nucleotides (nt) in length and can affect the gene expression post-transcriptionally [12]. miRNAs not only can regulate the development, differentiation and apoptosis in cells, but also participate in the development process of many diseases [13]. Over the past decade, thousands of miRNAs have been identified in helminths, many of them were associated with parasite development and interactions of parasites-hosts [14]. For example, some miRNAs of T. canis were predicted to be associated with the development, reproduction, host-parasite interactions or drug resistance [15]. Altered host miRNAs induced by T. canis infection play important roles in the pathogenesis of parasite migration in the liver [16]. Competing endogenous RNA (ceRNA) network analysis of lncRNA-miRNA-mRNA in puppy lung after T. canis infection improved the understanding of the miRNA regulatory mechanisms [17]. However, the effect of T. canis infection on the expression of miRNAs in dog bone marrow remains unclear.

In this study, the expression trajectories of miRNAs in the bone marrow of Beagle dogs were profiled by RNA-seq technology at different time points after T. canis infection. Through the bioinformatics analysis, we identified several differentially expressed (DE) miRNAs (DEmiRNAs) that could play regulatory roles in immune response in infected dogs, which may contribute to understanding the pathogenesis of toxocariasis.

Results

Characterization of the RNA sequencing data

A total of 272,834,730 raw reads and 264,942,218 clean reads were obtained by RNA-seq. The average raw data of each sample was 0.75 Gb. A total of 585 miRNAs were identified in this study, including 321 known miRNAs and 264 novel miRNAs. The Q20 and Q30 of the raw data were both > 95%, and the percentage of GC was approximately 50% (Table 1), suggesting that these RNA-seq data were of good quality and can be used for subsequent bioinformatics analysis.

Table 1 Quality control results of sequencing data
Full size table

Differential expression of miRNAs and qRT-PCR validation of the RNA-seq results

A total of 39 DEmiRNAs were identified in this study. Among them, four DEmiRNAs were identified at 24 hpi and all were up-regulated; 8 DEmiRNAs were identified at 96 hpi, including two up-regulated miRNAs and six down-regulated miRNAs; 27 DEmiRNAs were identified at 36 dpi, including 13 up-regulated miRNAs and 14 down-regulated miRNAs (Fig. 1a and Table S1). However, no overlapping DEmiRNAs between either of the three infection stages were identified (see the Venn diagram in Fig. 1b). The qRT-PCR validation results are depicted in Fig. 2, which showed that the qRT-PCR results were consistent with the overall trend of RNA-seq expression levels, further demonstrating that the RNA-seq results are reliable.

Fig. 1
figure 1

Comparison of differentially expressed (DE) miRNAs (DEmiRNAs) in the bone marrow of Beagle dogs infected by Toxocara canis at 24 h post-infection (hpi), 96 hpi, and 36 days post-infection (dpi). a The number of DEmiRNAs at 24 hpi, 96 hpi and 36 dpi. Green represents the total number of DEmiRNAs, red represents the number of up-regulated DEmiRNAs, and blue represents the number of down-regulated DEmiRNAs. b Venn diagram showing the numbers of common and unique DEmiRNAs at the three infection stages

Full size image
Fig. 2
figure 2

Validation of differentially expressed (DE) miRNAs (DEmiRNAs) using qRT-PCR. The error bar represents the standard deviation based on three repetitions. The X-axis shows DEmiRNAs, and the Y-axis shows the relative expression level of DEmiRNAs

Full size image

Target gene prediction and GO annotation analysis of DEmiRNAs

A total of 48, 523, and 1,167 potential target genes of DEmiRNAs were predicted at 24 hpi, 96 hpi and 36 dpi, respectively (Table S2). GO annotation analysis revealed that a total of 31 potential target genes of two DEmiRNAs were significantly enriched in 384 GO terms at 24 hpi, with one immune-related GO term such as immune system development (GO: 0002520); a total of 476 potential target genes of 7 DEmiRNAs were significantly enriched in 836 GO terms at 96 hpi, with 26 immune-related GO term, such as cytokine production involved in immune response (GO:0002367), regulation of lymphocyte mediated immunity (GO:0002706), cell activation involved in immune response (GO:0002263); a total of 1,028 potential target genes of 17 DEmiRNAs were significantly enriched in 857 GO terms at 36 dpi, with 22 immune-related GO term, such as T-helper 17 type immune response (GO:0072538), neutrophil mediated immunity (GO:0002446), humoral immune response mediated by circulating immunoglobulin (GO:0002455) (Table S3). The top 10 significantly enriched biological process (BP), cellular component (CC), and molecular function (MF) terms of the three stages after T. canis infection are shown in Fig. 3.

Fig. 3
figure 3

Scatter plots of the top 30 enriched Gene Ontology (GO) terms of differentially expressed (DE) miRNAs (DEmiRNAs) at (a) 24 h post-infection (hpi), (b) 96 hpi, and (c) 36 days post-infection (dpi). The X-axis represents the rich factor and the Y-axis shows the GO terms. The greater the rich factor, the greater the degree of term enrichment. The color of the dots indicates the enrichment score [-log10(P value)], where red indicates high enrichment, while green indicates low enrichment. The dot size indicates the number of the potential target gene of DEmiRNAs in the corresponding GO term. Larger dots indicate the greater number of potential target genes enriched in this GO term

Full size image

KEGG pathway enrichment analysis

KEGG pathway analysis revealed that seven potential target genes of two DEmiRNAs were significantly enriched in five signaling pathways at 24 hpi, such as Hematopoietic cell lineage (cfa04640), Cell adhesion molecules (CAMs) (cfa04514), Jak-STAT signaling pathway (cfa04630), Hippo signaling pathway (cfa04390), and Vitamin digestion and absorption (cfa04977); 35 potential target genes of 6 DEmiRNAs were significantly enriched in seven signaling pathways at 96 hpi, such as Focal adhesion (cfa04510), Colorectal cancer (cfa05210), ECM-receptor interaction (cfa04512), Chagas disease (American trypanosomiasis) (cfa05142), Thyroid hormone signaling pathway (cfa04919); a total of 83 potential target genes of 10 DEmiRNAs were significantly enriched in eight signaling pathways at 36 dpi, such as Neurotrophin signaling pathway (cfa04722), Adrenergic signaling in cardiomyocytes (cfa04261), Dopaminergic synapse (cfa04728), GnRH signaling pathway (cfa04912), Long-term potentiation (cfa04720) (Table S4). The top 20 representative pathways at the three infection stages are shown in Fig. 4.

Fig. 4
figure 4

Scatter plots of the top 20 enriched Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway of differentially expressed (DE) miRNAs (DEmiRNAs) at (a) 24 h post-infection (hpi), (b) 96 hpi, and (c) 36 days post-infection (dpi). The X-axis represents the rich factor; the Y-axis shows the KEGG pathways. The greater the rich factor, the greater the degree of pathways enrichment. The color of the dots indicates the enrichment score [-log10(P value)], where red indicates high enrichment, while green indicates low enrichment. The dot size indicates the number of the potential target gene of DEmiRNAs in the corresponding pathways. Larger dots indicate the greater number of potential target genes enriched in this pathway

Full size image

Discussion

Previous studies have shown that miRNAs can play critical roles in the pathogenesis of T. canis during the liver phase and lung phase of parasite development [16, 17]. However, the effect of T. canis infection on miRNA expression of dog bone marrow remains unclear. In this study, RNA-seq was used to examine the expression levels of miRNAs in bone marrow of Beagle dogs at different stages after T. canis infection, which provided useful resources for further understanding the pathogenesis of T. canis infection. 39 DEmiRNAs were identified during the three stages after T. canis infection, and most of them had multiple target genes, suggesting that these DEmiRNAs may exert a broader range of functions during T. canis infection in dogs.

We identified the least number of DEmiRNAs and potential target genes at the initial time point of infection and showed an increasing trend with the time of infection. At the same time, all DEmiRNAs identified at the early stage of infection were up-regulated, while a large proportion of DEmiRNAs was down-regulated at the later stage of infection. This is consistent with our anticipation of the stages of the host immune system response to infection. In the early stage of infection, a small number of DEmiRNAs in the host activates the innate immune response and prevents further invasion of T. canis through the inflammatory response. As the infection proceeds, at 96 hpi, the host activates the adaptive immune response through regulation of a portion of the dysregulated DEmiRNAs. At the late stage of infection, the host promotes the activation and differentiation of immune cells through many dysregulated DEmiRNAs. We speculate that this is closely related to the increasing complexity of parasite-host interactions.

Target gene prediction and DEmiRNAs analysis

At 24 hpi, four DEmiRNAs, such as cfa-miR-122 and cfa-miR-193b, were predicted to be able to regulate 48 potential target genes. CD22 is a specific adhesion molecule on the surface of B lymphocytes, which can activate and regulate the function of B lymphocytes through ligand-dependent and non-dependent mechanisms, and can respond to exogenous antigen or self-antigen [18]. Recognition of antigens by the B-cell antigen receptor (BCR) is the main process in the B-cell immune response [19], and Toll-like receptors (TLRs) are the pathogen recognition receptor that initiates the innate immune response [20]. CD22 has been shown to inhibit BCR and TLR signaling in B cells [19, 21]. At 24 hpi, T. canis larvae hatch in the small intestine of the infected dogs, then reach the liver via the intestinal wall, lymphatic vessels and capillaries; and during this process, excretory-secretory products (ESPs) of T. canis larvae will induce the immune response of dogs [5]. cd22 was predicted as a potential target gene for cfa-miR-193b in this study (Table S2). We speculate that up-regulated cfa-miR-193b could enhance the host immune response to T. canis infection, which could facilitate host resistance against T. canis infection. Erythropoietin (EPO) plays important role in the survival of erythroid progenitors by binding to its cell surface receptor, promoting cell survival, proliferation and differentiation in the production of mature RBCs [22]. In this study, epo could be down-regulated by up-regulated cfa-miR-122, which could affect the maturation of RBCs, and lead to anemia symptoms in puppies. The migration of T. canis larvae can cause damage to the body, resulting in a series of inflammatory reactions [5]. A previous study found that miR-382 contributes to promote the inflammatory responses in the host [23]. The up-regulated cfa-miR-382 induced by T. canis infection may be more specific on the expected effects of identified target genes in the pro-inflammatory response at the early infection stage in the infected dogs. In conclusion, the target genes predicted at 24 hpi can serve as host innate immune response components, promoting host resistance to T. canis infection, and also serving as a bridge that links innate and adaptive immunity.

At 96 hpi, eight DEmiRNAs, such as novel_328, were predicted to be able to regulate 523 potential target genes. The novel_328 was down-regulated at 96 hpi and was predicted to have multiple potential target genes in this study, such as transforming growth factor beta 1 (tgfb1) and thymocyte expressed, positive selection associated 1 (tespa1) (Table S2). TGFB1 is a pleiotropic and multifunctional growth factor that is mainly expressed in the immune system. It has potent immunomodulatory properties and can exert a potent anti-inflammatory function [24]. The down-regulation of novel_328 could contribute to the immune regulation and anti-inflammatory effect of the host by regulating the expression of tgfb1. TESPA1 is one of the members of TCR signaling, which is essential for T cell selection and maturation through regulating TCR signaling during T cell development [25]. Tespa1-deficient mice exhibit impaired thymocyte development, and the decrease of CD4+ and CD8+ T cells in the matured thymus [25]. The down-regulation of novel_328 in the puppy bone marrow may promote the puppy immune response to resist T. canis infection at 96 hpi by affecting the expression of tespa1. However, the biological function of tgfb1 and tespa1 in T. canis infection requires further experimental investigation. Therefore, we speculate that at 96 hpi, the activation of host adaptive immune response is closely associated with the potential target genes (tgfb1 and tespa1), and the biological functions of tgfb1 and tespa1 in T. canis infection should be deeply investigated in subsequent studies.

At 36 dpi, 27 DEmiRNAs, such as cfa-miR-331 and novel_129, were predicted to be able to regulate 1,167 potential target genes. cfa-miR-331 and novel_129 were significantly down-regulated at 36 dpi. Semaphorin 7A (SEMA7A), acting as an effective immunomodulator, can be expressed in activated lymphocytes and myeloid cells, which can promote cell differentiation of the immune system [26], and can significantly increase Th1/Th17 cytokine secretion [27]. Sema7a was predicted to be a potential target gene of cfa-miR-331 in this study (Table S2). The down-regulation of cfa-miR-331 may promote host TH1 immune response in T. canis infection at 36 dpi by affecting the expression of sema7a. As a member of the IgSF, CD96 is involved in both immune and developmental processes [28]. A previous study showed that CD96 may contribute to the differentiation and activation of T cells, such as enhancing the activation and effector responses of CD8+ T cells [29]. cd96 was predicted to be a potential target gene for novel_129 in this study (Table S2). The abnormal expression of novel_129 may be a self-protective manifestation in the host against T. canis infection through regulating the expression of cd96. In the late stages of T. canis infection in dogs, the predicted potential target genes act as immunomodulators, which may contribute to host self-protection and organismal stability.

GO annotation analysis of DEmiRNAs

At 24 hpi, GO annotation analysis showed that 384 GO terms were significantly enriched; among which, 351, 24 and 9 GO terms were included in the three categories of BP, MF and CC, respectively (Table S3). EPO, a glycoprotein hormone, is a key cytokine regulator of erythropoiesis in the bone marrow and plays an important role in erythropoiesis by Epo/Epo receptor (EpoR) signaling pathways [30]. At 24 hpi, GO terms such as response to erythropoietin, cellular response to erythropoietin, erythropoietin-mediated signaling pathway and erythrocyte maturation were significantly enriched in the bone marrow of infected puppies,, which may be a compensatory response to visceral injury and anemia of the puppies caused by T. canis infection at the early stage. These results indicated that puppies may mobilize the production of RBCs at this time to promote the recovery of body damage.

At 96 hpi, 836 GO terms were significantly enriched; among which, 674, 95 and 67 GO terms were included in the three categories of BP, MF and CC, respectively (Table S3). ESPs of T. canis larvae can induce complex host immune responses during infection [8]. At 96 hpi, 476 potential target genes of seven DEmiRNAs were mainly associated with GO terms of BP; among which, many potential target genes were associated with immune reaction, such as T cell mediated immunity, regulation of T cell mediated immunity, regulation of humoral immune response, regulation of adaptive immune response and regulation of B cell mediated immunity. The significant enrichment of these immune response-related GO terms indicated that T. canis infection has caused a strong immune response in the infected puppies.

At 36 dpi, 857 GO terms were significantly enriched; among which 692, 112 and 53 GO terms were included in the three categories of BP, MF and CC, respectively (Table S3). MAP kinase, located in the protein kinase cascade, is the main component that controls embryogenesis, cell differentiation, cell proliferation and cell death pathways [31]. Mitogen-activated protein kinases (MAPKs) pathway can be activated by various pro-inflammatory factors or inflammatory reactions, which plays an important role in mediating effective inflammatory response and cytokine production [32, 33]. In this study, some potential target genes of DEmiRNAs were significantly enriched in positive regulation of MAPK cascade, regulation of MAP kinase activity, MAP kinase activity, stress-activated MAPK cascade and activation of MAPKKK activity. When T. canis returns to the intestine of the puppies, it will continue to cause inflammatory response, which may not be limited to the intestine and also occur in the bone marrow. Biological processes related to MAPKs could play important roles in the host at the late stage of T. canis infection.

KEGG pathway enrichment analysis

At 24 hpi, Cell adhesion molecules (CAMs), Hematopoietic cell lineage, Vitamin digestion and absorption, Hippo signaling pathway, and Jak-STAT signaling pathway were significantly enriched (Fig. 4a). The CAMs of IgSF regulate some vital processes in the life course of cells, including cell proliferation and cell differentiation in time and space [34]. In the process of inflammation, CAMs can mediate the adhesion between leukocytes and endothelial cells, so that leukocytes can migrate to the inflammatory site [35]. A previous study found that CAMs play important roles in goats resistance to gastrointestinal nematodes (GINs) infection [36]. The CAMs may stimulate host immune function in the early stage of T. canis infection, which maintains the stability of the body and plays an important role in host defense against T. canis invasion.

At 96 hpi, the potential target genes of DEmiRNAs were significantly enriched in Focal adhesion and ECM-receptor interaction (Fig. 4b). ECM-receptor interaction affects cellular activities, including cell proliferation, differentiation, migration and survival [37]. ECM-receptor interaction as an immune-related pathway helps parasites adapt to the host [38]. We speculate that ECM-receptor interaction may contribute to the migration of T. canis larvae to suitable survival sites and avoid the clearance of the host immune system. The Focal adhesion pathway can perform a variety of biological functions, including the immune response and the barrier function, as well as promote the migration of neutrophils from blood vessels to infected sites by participating in endothelial conformational changes during the inflammatory process [39, 40]. A previous study has shown that gene expression related to focal adhesion may change in the host infected with Trichinella spiralis, leading to focal adhesion changes and the formation of epithelial monolayer tight junctions, which may be beneficial to host defense against H. spiralis infection [39]. Puppies could resist T. canis larvae infection or migration to other tissues by focal adhesion pathway at 96 hpi.

At 36 dpi, the target genes of DEmiRNAs were enriched in the Neurotrophin signaling pathway, Long-term potentiation, Axon guidance and Protein processing in endoplasmic reticulum (Fig. 4c). Neurotrophins and their receptors in neurotrophin signaling pathway are widely expressed in the immune system [41]. Neurotrophins can regulate the functions of lymphocytes and eosinophils, and promote B cell maturation, while also acting on hematopoietic lineage cells to regulate their hematopoietic function [42]. Neurotrophins are increased during the inflammatory response, parasitic infection, etc. [42]. We speculate that the Neurotrophin signaling pathway may play a critical role in host inflammation and tissue repair at 36 dpi by activating immune inflammatory cells through this pathway.

Conclusions

The present study examined the miRNA expression profiles in the bone marrow of Beagle puppies during the different stages of T. canis infection. The findings reveal dynamic changes in expression profiles of miRNAs, reflecting the complexity of the T. canis-host interaction. cfa-miR-193b, novel_328, and cfa-miR-331 are associated with the host immune response. novel_129 could regulate host self-protection mechanisms. cfa-miR-382 may contribute to promote host inflammatory responses. GO and KEGG pathway analysis revealed that these DEmiRNAs were involved in many critical biological processes and the regulation of diseases and immune signaling pathways. These findings also provided valuable information for further understanding the pathogenesis of T. canis infection.

Materials and methods

The collection of bone marrow tissue samples

A total of 18 Beagle dogs from 3 litters, 6–7 weeks old, were purchased from and housed at the National Canine Laboratory Animal Resource Center (Guangzhou, China). All puppies were examined using the sugar flotation method to ensure that there was no intestinal parasite infection. Moreover, each puppy’s serum was tested by indirect ELISA using larval excretory-secretory antigen to eliminate the T. canis infection [10]. According to the migration time points of the T. canis in definitive dog host, these puppies were divided into three time point groups: 24 h post-infection (hpi), 96 hpi, and 36 days post-infection (dpi). Each time point group included infected dogs and uninfected control dogs. Puppies from the same litter are equally distributed to reduce the influence of individual factors. Infected puppies orally received 1 mL of normal saline solution containing 300 infectious T. canis eggs, while uninfected control puppies received 1 mL of normal saline orally. At each specified time point post-infection as described above, puppies were anesthetized using Zoletil 50™ (Virbac, Nice, France), and then were euthanized through injecting KCl solution into the heart [10]. Then, the tibial bone marrow was rapidly separated aseptically and was stored in liquid nitrogen for RNA extraction. The potential infectious waste materials, such as puppy feces, infectious eggs and discarded tissues, were autoclaved before disposal.

RNA isolation

Total RNA was extracted from bone marrow samples with Trizol (Life Technologies, Carlsbad, USA). The genomic DNA in extracted RNA was removed with DNase I (NEB, Ipswich, USA). The purity and concentration of above RNA were determined with a Qubit® 2.0 Flurometer (Life Technologies, Carlsbad, USA). The integrity of the isolated RNA was verified using an Agilent Bioanalyzer 2100 system (Agilent Technologies, SantaClara, USA). The high-quality RNAs with RNA integrity number (RIN) value > 8.0 were used to construct the small RNA-seq libraries.

Library preparation and small RNA sequencing

Three μg total RNA per sample of puppy bone marrow was used as input material to generate the small RNA-seq library by using NEBNext® Multiplex Small RNA Library Prep Set for Illumina® (NEB, Ipswich, USA), and the index codes of 3' SR Adaptor (NEB, Ipswich, USA) were added to attribute sequences to each puppy bone marrow sample, following the manufacturer’s recommendations. Then, first-strand cDNA was synthesized, and the cDNA was amplified using LongAmp Taq 2X Master Mix, SR Primer for Illumina and index (X) primer. The PCR products were electrophoresed by 8% PAGE (100 V, 80 min), and the DNA fragments (including aptamers) of approximately 140–160 bp were recovered and purified to assess the quality on the Agilent Bioanalyzer 2100 system using DNAHigh Sensitivity Chips. The clustering of the index-coded samples was performed on a cBot Cluster Generation System using TruSeq SR Cluster Kit v3-cBot-HS (Illumia, San Diego, USA) according to the manufacturer’s instructions. After cluster generation, the libraries were sequenced on an Illumina Hiseq 2500 platform and 50 bp long single-end reads were generated.

Identification of miRNAs

Raw data (raw reads) that containing poly-N, with 5' adapter contaminants, without 3' adapter or the insert tag and low-quality reads were firstly filtered by custom perl and python scripts to obtain clean data (clean reads). Then, the Q20, Q30 and GC-content of raw data were calculated. The small RNAs of clean reads ranging from 18–30 nt were mapped to reference sequences by Bowtie to search known miRNAs [43]. Using the known miRNAs of Canis familiaris in MiRbase 22.0 as a reference, the improved software miRdeep 2 and srna-tools-cli were used to identify potential miRNAs and draw the stem-loop structure of the identified miRNAs [44, 45]. According to the characteristics of the hairpin structure of miRNA precursor, miREvo and miRdeep 2 were used to predict the novel miRNAs, and the secondary structure and minimum free energy of the small RNA tags that were unannotated in the preceding steps were analyzed [44, 46]. To ensure that each unique small RNA was mapped to only one annotation, the following priority rule was set up: known miRNA > rRNA > tRNA > snRNA > snoRNA > NAT-siRNA > novel miRNA > ta-siRNA.

The prediction of miRNA target genes and the enrichment analysis of GO and KEGG

The prediction of miRNA target genes was conducted by miRanda [47]. The expression level of miRNA was estimated by transcript per million (TPM) [48]. The transcriptional level of miRNA was assessed by the DESeq R package, and the DEmiRNA was identified by the thresholds of P values < 0.05 between infected dogs and uninfected control dogs [49]. GO annotation and KEGG pathway analysis of the target genes of DEmiRNAs were performed by GOseq R package and KOBAS 3.0, respectively [50,51,52]. P values < 0.05 were considered significantly enriched for GO and KEGG terms.

The qRT-PCR validation of RNA-Seq results

To validate the accuracy of the RNA-seq data, seven DEmiRNAs identified at 36 dpi were selected for verification by quantitative real-time PCR (qRT-PCR). The qRT-PCR quantitative experiment was performed on LightCycler480 (Roche, Basle, Switzerland). The miRNA reverse transcription step, qRT-PCR amplification procedures and melting curve analysis were performed strictly as previously described [17]. The selected DEmiRNAs and their PCR primers are summarized in Table S5. The verification results were calculated by the 2−ΔΔCT method, with U6 small ribonucleic acid as the internal reference for normalizing the level of miRNAs.