New insights into the Manila clam and PAMPs interaction based on RNA-seq analysis of clam through in vitro challenges with LPS, PGN, and poly(I:C)

Background Manila clam (Ruditapes philippinarum) is a worldwide commercially important marine bivalve species. In recent years, however, microbial diseases caused high economic losses and have received increasing attention. To understand the molecular basis of the immune response to pathogen-associated molecular patterns (PAMPs) in R. philippinarum, transcriptome libraries of clam hepatopancreas were constructed at 24 h post-injection with Lipopolysaccharide (LPS), peptidoglycan (PGN), and polyinosinic-polycytidylic acid (poly(I:C)) and phosphate-buffered saline (PBS) control by using RNA sequencing technology (RNA-seq). Results A total of 832, 839, and 188 differentially expressed genes (DEGs) were found in LPS, PGN, and poly(I:C) challenge group compared with PBS control, respectively. Several immune-related genes and pathways were activated in response to the different PAMPs, suggesting these genes and pathways might specifically participate in the immune response to pathogens. Besides, the analyses provided useful complementary data to compare different PAMPs challenges in vivo. Functional enrichment analysis of DEGs demonstrated that PAMPs responsive signal pathways were related to apoptosis, signal transduction, immune system, and signaling molecules and interaction. Several shared or specific DEGs response to different PAMPs were revealed in R. philippinarum, including pattern recognition receptors (PRRs), antimicrobial peptides (AMPs), interferon-induced proteins (IFI), and some other immune-related genes were found in the present work. Conclusions This is the first study employing high throughput transcriptomic sequencing to provide valuable genomic resources and investigate Manila clam response to different PAMPs through in vivo challenges with LPS, PGN, and poly(I:C). The results obtained here provide new insights to understanding the immune characteristics of R. philippinarum response to different PAMPs. This information is critical to elucidate the molecular basis of R. philippinarum response to different pathogens invasion, which potentially can be used to develop effective control strategies for different pathogens.


Background
Manila clam, Ruditapes philippinarum, is one of the most commercially important bivalves and reached over 4.2 million tons in 2017, which is widely distributed along the coast of China, Japan, and Korea [1]. R. philippinarum possesses many advantages as an aquaculture species, including wide salinity and temperature resistance, rapid growth, and pollution tolerance [2]. Nevertheless, R. philippinarum has been threatened with a huge challenge caused by pathogen invasion [3][4][5]. Pathogens can affect not only the development and survivorship of clams but also the quality and price of the product [6]. The majority of diseases in Manila clam are associated with Vibrio [7,8] and Perkinsus [9,10]. Diseases affecting R. philippinarum can result in mass mortality in aquaculture and cause large economic losses [11].
Although most bivalves lack a specific immune system, the innate response, which includes circulating hemocytes and multiple molecular effectors (PRRs, AMPs), appears to be an effective defense against external aggression [11]. Different receptors, regulators, and effectors, including pattern recognition receptors (PRRs), antimicrobial proteins (AMPs), and a variety of other molecules involved in agglutination, phagocytosis, and encapsulation, have been found in some Molluscs [12,13]. In recent years, more and more research has been focused on the immune system of Molluscs [7,14], and most of the research was focused on mussels, oyster, and scallop [15][16][17]. However, the available information on PRRs, AMPs, and immune-related signaling pathway of R. philippinarum is still limited.
In the past decades, the high-throughput RNA sequencing (RNA-seq) technique has been widely used to investigate molecular interactions between host and pathogen in Molluscs [6,11]. Some efforts have been made to enrich the clam gene database and to gain an in-depth understanding of the potential immune mechanism of R. philippinarum [11,18]. Recently, gene expression profiles of R. philippinarum hemocytes stimulated with Perkinsus olseni trophozoites, zoospores, and extracellular products under different experimental conditions were analyzed with RNA-seq on an Illumina platform [10]. The 454 pyrosequencing technology was used to obtain hemocytes transcriptome after in vitro immunestimulated in the Manila clam, and a large number of immune-related genes were found that play important roles in the defense mechanisms of R. philippinarum [11]. Besides, transcriptional study in Manila clam in response to brown ring disease revealed that most changes in response to brown ring disease were tissue-specific, and a lot of candidate genes involved in microbe recognition and killing were identified [6]. Recently, the whole-genome of the Manila clam was assembled and annotated in our previous study [19], and the molecular basis of its adaptation to hypoxia, parasites and aerial exposure stress were analyzed [19][20][21]. However, the transcriptome analysis of response and defense against different pathogens or pathogen-associated molecular patterns (PAMPs) in Manila clam was still a largely unexplored landscape.
PAMPs, a class of conserved small molecular motif in microorganism, could be recognized by the PRRs of multicellular organisms and then activate innate immune response [22,23]. Lipopolysaccharide (LPS), a well-characterized PAMP, is a component of the cell wall in Gram-negative bacteria [24]. It has been found that LPS could induce multiple innate immune responses in some Molluscs [25,26]. Peptidoglycan (PGN), a component of the bacterial cell wall, is extracted from both Gram-positive and Gram-negative bacteria [23]. In Crassostrea gigas, PGN was recognized by peptidoglycan recognition proteins (CgPGRPs) and CgTLR-6 [27,28]. Polyinosinic-polycytidylic acid (poly(I: C)) is a kind of synthetic double-stranded RNA associated with viral infection [29]. It has been demonstrated that poly(I:C) could significantly up-regulate the expression of CgIFNLP, CgIFNR-3, and CgCaspase8-2 in C. gigas [30,31]. In bivalves, pattern recognition molecules (PRMs) could recognize PAMPs and trigger the innate immune response [25,27,30,31]. Hence, elucidating the immune response patterns and defense mechanisms of R. philippinarum against different PAMPs has important biological significance in the interpretation of the immune function of Manila clam.
In this study, we employed high throughput transcriptomic sequencing to investigate Manila clam response to different PAMPs through in vivo challenges with LPS, PGN, and poly(I:C). The hepatopancreas transcriptome of R. philippinarum after stimulated with three different PAMPs (LPS, PGN, and poly(I:C)) were analyzed to reveal the immune response of the Manila clam against LPS, PGN, and poly(I:C), and to elucidate the shared and specific immune-related genes in immune signaling pathways of Manila clam facing different PAMPs stress. This work sheds light on the molecular basis of Manila clam response to different PAMPs, and provides new insights into the immune signaling and pathogen defense responses of R. philippinarum.

Genome-guided transcriptome assembly
A total of 533,660,306 raw reads were obtained, including 138,764,184 raw reads from the LPS treatment groups, 131,242,452 raw reads from the PGN treatment groups; 128,591,010 raw reads from the poly(I:C) treatment groups, and 135,062,660 raw reads from the PBS control groups corresponding to the constructed libraries. After low-quality reads (quality scores < 20), short reads (length < 60 bp), and ambiguous nucleotides were removed, a total of 266,830,153 clean reads were retained for further mapping and differential expression analysis. The clean reads were assembled into 35,919 unigenes with a mean length of 1681 bp, a minimum length of 122 bp, and a maximum length of 64,604 bp. Besides, 8266 novel genes were annotated in this study. Summarized trimming statistics and the number of sequenced reads per sample was shown in Additional file 1. The RNA sequencing data has been submitted to the NCBI SRA database (Accession number: PRJNA616201). Density distribution of expression level based on log10 (FPKM) in each library was exhibited in Additional file 2, which showed that LPS, PGN, and poly(I:C) groups were similar, while the PBS control group library was different with others three groups.

Detection of differentially expressed genes
A total of 1859 DEGs were identified from three PAMPs groups (LPS, PGN, and poly(I:C)) compared with the PBS group. The lowest number of DEGs (188) was found in the poly(I:C) group, while 832 and 839 DEGs were found in the LPS and PGN group, respectively. Of those DEGs, 635 up-regulated and 197 down-regulated in LPS group, 520 up-regulated and 319 down-regulated in PGN group and 86 up-regulated and 102 downregulated in poly(I:C) group were identified compared with PBS control group, respectively (Fig. 1). The number of up-regulated genes were significantly higher than the down-regulated genes in LPS and PGN group.

Comparison of transcriptome profiles elicited by different PAMPs
As showing in Fig. 3a and c, the DEGs in response to poly(I:C) were significantly less than that in response to LPS or PGN in R. philippinarum, while the number of DEGs in LPS and PGN groups were similar. A total of 255 genes were differentially expressed (≥ 2 fold change) in response to LPS challenge, which were also found in the DEGs of PGN challenge group, for example, Cathelicidin (CATH), scavenger receptor cysteine-rich (SRCR), IgGFc-binding protein (FCGBP), low affinity immunoglobulin epsilon Fc receptor (FCGR), Cell number    Table 1). KEGG pathway enrichment analysis of DEGs was conducted to detect significantly altered pathways in each group. A total of 28, 41, and 21 pathways were significantly enriched in LPS, PGN, and poly(I:C) group, respectively (P < 0.05) (Fig. 3b). These pathways were primarily involved in apoptosis, signal transduction, immune system, and signaling molecules and interaction (Table 2). Furthermore, some enriched immune-related pathways were shared by multiple PAMPs groups, such as Platelet activation (map 04611) and Focal adhesion (map 04510) (shared by LPS, PGN, and poly(I:C) group), Phagosome (map 04145) (shared by LPS and poly(I:C) group), ECM-receptor interaction (map 04512) and Complement and coagulation cascades (map 04610) (shared by PGN poly(I:C) group), Cell cycle (map 04110), p53 signaling pathway (map 04115), cGMP-PKG signaling pathway (map 04022), and Calcium signaling pathway (map 04020), (shared by LPS and PGN group) ( Table 2). Besides, some specific immunerelated pathways enriched by DEGs also were identified, such as Lysosome (map 04142) (  Validation of gene expression profiles using the quantitative real-time PCR

GO and KEGG enrichment analysis of DEGs
To validate the accuracy of RNA-seq results, we select both shared and specific immune-related genes that were differentially expressed in response to LPS, PGN, and poly(I:C) to perform the quantitative real-time PCR (qPCR) analysis. The specific primers of those genes were listed in Additional file 4. The fold change detected by qPCR was compared with that detected by RNA-Seq expression analysis (Fig. 7). As is shown in Fig. 7, nearly all of those DEGs shared the same trends in LPS, PGN, and poly(I:C) groups. In general, PAMPs responsive genes identified with quantitative real-time PCR experiments were consistent with the results of the Illumina sequencing analysis, indicating the accuracy of the RNAseq expression analysis.

Discussion
This study provides the first genome-based transcriptome analysis in the hepatopancreas of R. philippinarum under three different PAMPs challenge (LPS, PGN, and poly(I:C)). Due to the availability of R. philippinarum whole-genome sequence date [21], enable us could effectively extract reads that mapped to exons, avoid problems caused by intron-mapped reads [32], and found many novel genes which were not annotated in R. philippinarum genome. To better understand the innate immune system of R. philippinarum response to PAMPs challenge and to uncover the difference of Manila clam response to different PAMPs, the immune-related pathways, and genes that differentially expressed in different PAMPs challenge groups were analyzed. In the present study, several PRMs that are important in detecting PAMPs were identified, including the FREP, C1qL, FCN, TLR, PGRP, Perlucin, MMR1, C-type lectin (CTL), CL, LTL, SRCR, and sialic acid-binding lectin (Siglec) ( Table 1), most of which could activate innate immune response by recognizing specific structures that are exposed to pathogens [22]. Lysozymes is a kind of alkaline enzyme that lyse bacteria by hydrolyzing the b- (1,4) linkage between N-acetylmuramic acid and N-acetylglucosamine of the peptidoglycan in bacterial cell walls [33]. The role of lysosomes in inflammatory response has been reported in Molluscan species [33,34]. In bivalves, lysosomes were found within the granular hemocytes [35], and were released during degranulation of the granular haemocytes accompanies with phagocytosis [36]. The released enzymes then assist in the breakdown of foreign material [34]. It has been reported that a sub-lethal doses of the Listonella anguillarum (Gram-negative bacterium) tended to cause the destabilisation of the lysosomal membranes in Ostrea edulis, C. gigas, and Pecten maximus for a period of up to 48 h post-inoculation [35]. In the present study, the Lysosome pathway was conspicuously enriched between the LPS challenge group and the PBS control group (Fig. 5). In addition, five genes participating in the Lysosome pathways were significantly regulated in response to LPS challenge, indicating the Lysosome pathway plays a vital role in Gram-negative bacterium infection in R. philippinarum.    Intracellular PRMs, such as NOD-like receptor, could recognize virus-derived dsRNA and bacteria-derived peptidoglycan, and induce the inflammatory response [36]. More than 20 NOD-like receptors that play a key role in the recognition of intracellular ligands have been found in mammals [36]. Recently, it has been reported that Nod1 and Nod2 identify different sites of bacterial PGN and mediate innate immunity [37]. In Manila clam, the previous study reported that NOD-like receptor signaling pathway was enriched in the hepatopancreas transcriptome of R. philippinarum after challenged by Vibrio anguillarum [18]. In the current work, a total of 16 DEGs, such as caspase, CATH, DEFs, interferon-induced proteins (IFI), and Interferon-induced guanylate-binding proteins (GBP) in NOD-like receptor signaling pathway were highly regulated in response to PGN challenge in R. philippinarum (Fig. 6). Our finding indicated that NOD-like receptor signaling pathway might plays an important role in response to the Gram-positive bacterium in R. philippinarum.
Based upon the number of DEGs and pathways (Fig. 3a), R. philippinarum exhibited a stronger immune response to LPS and PGN challenge than to poly(I:C) challenge [38]. Interestingly, we found the transcriptomic responses of R. philippinarum to LPS and PGN shared more DEGs, including C1qL, CTL, SRCR, CATH, coagulation factor X (FA10), Leucinerich repeat-containing protein (LR74A), etc. This result might be due to both LPS and PGN are derived from bacteria (Gram-negative and Gram-positive, respectively), whereas poly(I:C) is a synthetic analog of dsRNA associated with viral infection [29].
Among those PRRs, C1qL is described remarkably expanded in R. philippinarum [42]. It has been reported that some C1qL proteins are the first PRMs of the complement system from the evolutionary perspective [43]. Complement C1q-like proteins contain the C1q domain and were grouped as the C1qDC [43]. A total of 168 different transcripts of C1qDC was found in Mytilus galloprovincialis, most of which show differential expression following challenge with Gram-positive or -negative bacteria [44,45], and the same expression pattern was also found in Mytilus edulis [46]. In this study, three C1qL genes were detected, and two C1qL genes showed up-regulated expression, while the other C1qL showed down-regulation. A similar result was found in R. philippinarum in response to brown ring disease [6] and other Molluscs [17,45]. Therefore, It is conceivable that some C1qL transcripts are up-regulated while the others are down-regulated providing a tailored response to pathogens in R. philippinarum [6]. In addition, C3, CTRP, FCGR, and FCGBP also were detected highly regulated in PAMPs treatment groups. C3 functions as the key molecular of complement system to distinguish and eliminate pathogens, and induce inflammatory responses [47]. The membrane-attached CTRP is the initial prototype of C1q and acting as immune PRM activating the proto-complement [43,44]. Besides, FCN, a derivative of the lectin pathway of complement activation [43], was found significantly up-regulated under LPS challenge in our study. To sum up, a rudimentary complement system with a group of expanded and diversified genes is suggested to exist in R. philippinarum and play a crucial role in the innate defense against pathogens [43].
PGRPs is considered to be a crucial immune molecule in Molluscs by detecting and eliminating invading bacteria [48][49][50][51][52]. The expression pattern of PGRP has been investigated in Chlamys farreri [49], Solen grandis [51], and Hyriopsis cumingi [52]. In this study, the upregulated expression of PGRPs was not only detected in PGN challenged clams but also in LPS and poly(I:C) challenged clams (Table. 1). Similar results were found in C. farreri and S. grandis [49,51]. In C. farreri, the CfPGRP-S1 was a constitutive and inducible acute-phase protein that was involved in the immune response against both the Gram-positive bacteria and Gramnegative bacteria infection [49]. The expression of SgPGRP-S2 was significantly up-regulated when S. grandis was stimulated by LPS, PGN, and b-1,3-glucan [51]. In Pacific oyster, a CgPGRP-S1S gene was found greatly contribute to efficient host defense systems, not only by direct interaction with bacteria, but also by triggering other defense pathways [50]. Therefore, we speculate PGRPs might not only serve as PRRs to recognize Gram-positive bacteria, but also participate in other defense pathways to respond to different pathogens invasion in R. philippinarum.
Immune effectors are usually induced by PRRs recognition and produced by epithelial cells from various organs, including AMPs, lysozymes, cytokines, antioxidant enzymes, and acute phase proteins [19]. In this study, two kind of AMPs (DEF, CATH) were annotated in DEGs in response to PAMPs. DEF is a large group of small antimicrobial peptides and involved in the host immune response against bacterial infection [53][54][55][56][57]. Various DEFs have been characterized in different bivalve species such as Venerupis philippinarum [53], C. gigas [54], Argopecten irradians [56], and R. philippinarum [55,57]. In C. gigas, Cg-def gene exhibits high activities against Gram-positive bacteria but low activity against Gram-negative bacteria and fungi [54]. It has been reported Rpdef showed the highest activity against Gram-positive bacteria and played an important role in the elimination of invading bacterium through membrane-disruptive activity in R. philippinarum [55,57]. In this study, two DEF (defensin and defensin-B) with six conserved cysteines were identified, which is consistent with previous reports of arthropod-like defensins in other Molluscs [54,56,58]. Both of the two DEF exhibited highly increasing expression in response to LPS and PGN challenge, especially the defensin gene increased 26.8-and 26.3-fold, respectively. The defensin-B showed a significantly up-regulated expression level in response to LPS, PGN, and poly(I:C). Our result indicates DEFs play vital roles in response to Gram-positive and -negative bacteria and virus invasion in R. philippinarum. Cathelicidins (CATHs) have broad antimicrobial activity against Gram-positive and Gram-negative bacteria in the insect [59]. However, little is known of the CATHs in bivalves, especially in R. philippinarum. In this study, the expression level of Cathelicidin-B1 in R. philippinarum after LPS and PGN challenge was upregulated (4.65-and 4.09-fold, respectively). We speculated that CATH is an essential molecule in R. philippinarum immunity system and may be utilized as executors for the incapacitation and elimination of Gram-positive and Gram-negative bacteria invasion.
Peroxidase-like protein (PO) belongs to the antioxidant enzymes [60]. Some reactive oxygen species (ROS) are highly harmful and toxic to organism and are significantly induced when attacked by invaders or stress [61,62]. Organisms have formed an antioxidant defense system that removes ROS to protect cells from damage caused by abundant ROS [12]. In the present study, the PO gene was found significantly down-regulated (− 4.32 fold change) in the LPS group, indicating the efficiency of ROS removing was decreased and the immune system was affected in R. philippinarum after LPS challenge. Heat shock proteins (HSPs) are evolutionarily ancient and highly conserved intracellular molecular chaperones [19]. The primary role of HSP is to function as molecular chaperones to modulate stress response [12]. When organisms are stressed by environmental conditions, the expression level of HSP would significantly increase, enabling the organism could resist the damage caused by adverse environment to maintain homeostasis and cell survival [63,64]. In this study, HSPs with up-regulated expression were found both in LPS, PGN, and poly(I:C) challenge group, indicating the activity of HSPs is closely linked to the innate immune system in R. philippinarum [65].
Another group of transcripts highly regulated in PAMPs challenged clams were interferon-induced proteins (IFI) and Interferon-induced guanylate-binding proteins (GBP). It has been reported that viral infection could trigger the interferon and interferon-induced genes highly upregulated in salmonids and rainbow trout [66,67]. In bivalves, CgIFNLP was found to increase significantly at 12 h (8.35-fold) and 24 h (4.95-fold) after poly(I:C) stimulation in C. gigas hemocytes [30]. GBP countered the antiviral effect by inhibition of its GTPase activity in Mammalian [68]. In this study, the expression of 5 IFI and 3 GBP genes was found highly regulated by poly(I:C) in R. philippinarum (IFI 3.46, 11.07, 5.10, -5.44, − 25.60, fold, and GBP -24.42, − 23.09, − 9.27 fold, respectively), while only 2 of IFI and 2 of GBP genes were regulated by LPS and PGN, and the expression level of those genes was similar. It is therefore plausible that some IFI were up-regulated, whereas others are switched down providing a tailored response to pathogens infection in R. philippinarum. Overall, these findings indicating IFI and GBP might play a crucial role in response to viral invasion in R. philippinarum.

Conclusions
The transcriptome comparison of the different PAMPs challenged Manila clam has provided new useful data to understand the molecular basis of the immune response to pathogens. The genome-based transcriptome analysis revealed LPS and PGN are more potent PAMPs in activating the immune response in R. philippinarum. LPS challenge group shared more immune-related DEGs and immune response pathways with PGN challenge group than poly(I:C) stimulation. Besides, some significantly enriched specific pathways directly related to immune response were found, such as the Lysosome pathway in the LPS challenge group, NOD-like receptor signaling pathway in the PGN challenge group. Moreover, some PRRs (FREP, C1qL, FCN, TLR, PGRP, Perlucin, MMR1, CTL, CL, SRCR, LTL, Siglec), AMPs (DEFs, CATH), interferoninduced proteins (IFI, GBP), HSPs, and PO were identified, which play pivotal roles in identification and clearance of invading pathogens in R. philippinarum. Our finding will aid understanding of R. philippinarum immune system and defense response to different pathogens invasion and provide new insights to develop effective control strategies for different pathogens.

Manila clam and PAMPs challenge
The wild adult Manila clams used in this study were collected from Jinshitan, Dalian, Liaoning Province, China.
The clams had an average shell length of 23.2 ± 1.0 mm, and an average weight of 5.4 ± 0.8 g. After being transported to the laboratory, the clams were cleaned to remove any fouling and were acclimated in aerated 20 L plastic tanks, containing water at 13.8 ± 0.6°C, pH 8.1 ± 0.1 with a salinity of 30 ppt. The clams were fed with Spirulina powder once a day for 2 weeks and the water was exchanged fully once per day to discharge waste products. Clams in each group were fasted at least 2 days before injection to avoid food contamination.

RNA extraction and library construction for Illumina sequencing
Total RNA was extracted from 30 mg hepatopancreas of each individual (each group n = 3) using RNAprep pure Tissue Kit (TianGene, Beijing, China), according to the manufacturer's protocol. The degradation and contamination of total RNA were monitored on 1% agarose gels. RNA purity and concentration were measured using the NanoPhotometer® spectrophotometer (IMPLEN, CA, USA) and Qubit® RNA Assay Kit in Qubit® 2.0 Fluorometer (Life Technologies, CA, USA), respectively.
A total amount of 3 μg RNA per sample was used as input material for the RNA sample preparations. Sequencing libraries were generated using NEBNext® UltraTM RNA Library Prep Kit for Illumina® (NEB, USA) following the manufacturer's recommendations and index codes were added to attribute sequences to each sample [70]. Fragmentation was carried out using divalent cations under elevated temperature in NEBNext First-strand Synthesis Reaction Buffer (5X) [70,71]. The first-strand cDNA was synthesized using random hexamer primer and M-MuLV Reverse Transcriptase (RNase H) [71]. Second strand cDNA synthesis was subsequently performed using DNA Polymerase I and RNase H. After adenylation of 3′ ends of DNA fragments, NEBNext Adaptor with hairpin loop structure was ligated to prepare for hybridization [71]. To select cDNA fragments of preferentially 250~300 bp in length, the library fragments were purified with AMPure XP system (Beckman Coulter, Beverly, USA) [72]. Then 3 μL USER enzyme (NEB, USA) was used with sizeselected, adaptor-ligated cDNA at 37°C for 15 min followed by 5 min at 95°C before PCR. PCR was performed with Phusion High-Fidelity DNA polymerase, Universal PCR primers, and Index (X) Primer [72]. At last, PCR products were purified (AMPure XP system) and library quality was assessed on the Agilent Bioanalyzer 2100 system [72].

Sequence filtering, mapping, and assembly
Raw reads of fastq format were firstly processed through in-house Perl scripts [73]. In this step, clean reads were obtained by removing reads containing adapter, reads containing ploy-N and low quality reads from raw data [73]. Clean reads were mapped to the reference genome of the R. philippinarum published in our previous study (https:// www.ncbi.nlm.nih.gov/genome/?term=txid129788[orgn]) [21]. Index of the reference genome was built using Hisat2 v2.0.5 and paired-end clean reads were aligned to the reference genome using Hisat2 v2.0.5 [73]. We selected Hisat2 as the mapping tool for that Hisat2 can generate a database of splice junctions based on the gene model annotation file and thus a better mapping result than other non-splice mapping tools [74,75]. Feature Counts v1.5.0-p3 was used to count the reads numbers mapped to each gene, and then FPKM of each gene was calculated based on the length of the gene and reads count mapped to this gene [76]. The P values were adjusted using the Benjamini & Hochberg method [77]. Next, the alignments are passed to StringTie (http://www.ccb. jhu.edu/ software/stringtie/) for transcript assembly. StringTie assembles the genes for each data set separately, estimating the expression levels of each gene and each isoform as it assembles them [78].

Differential expression analysis
Differential expression analysis of two groups was performed using the DESeq2 R package [74]. DESeq2 provides statistical routines for determining differential expression in digital gene expression data using a model based on the negative binomial distribution [74]. The resulting P-values were adjusted using the Benjamini and Hochberg's approach for controlling the false discovery rate [79]. Genes with an adjusted P-value < 0.05 found by DESeq2 were assigned as differentially expressed.

GO and KEGG enrichment analysis of differentially expressed genes
Using transcripts from the reference genome [21], we annotated gene functions using the Gene Ontology (GO), and Kyoto Encyclopedia of Genes and Genomes (KEGG) database. GO enrichment analysis of differentially expressed genes (DEGs) was implemented by the cluster Profiler R package, in which gene length bias was corrected [80]. GO terms with corrected P-value less than 0.05 were considered significantly enriched by differential expressed genes [74]. KEGG is a database resource for understanding high-level functions and utilities of the biological system, such as the cell, the organism and the ecosystem, from molecular-level information, especially large-scale molecular data sets generated by genome sequencing and other highthroughput experimental technologies (http://www.genome.jp/kegg/) [74]. We used the cluster Profiler R package to test the statistical enrichment of differential expression genes in KEGG pathways [70].

qPCR confirmation of Illumina sequencing data
To validate the Illumina sequencing data, twenty immune-related DEGs were chosen for quantitative realtime PCR (qPCR) analysis. The integrity and purity of RNA were determined by electrophoresis on a 1% agarose gel and a Nanodrop ND-2000 spectrophotometer (Thermo Electron Corp., Waltham, MA, USA), respectively. Total RNA was reverse-transcribed to cDNA with the PrimeScript RT reagent Kit (TaKaRa, Tokyo, Japan). The primers were designed with the Primer 5 software (Premier Biosoft International). The β-actin was selected as a reference gene for the qPCR analysis, due to its stably expressed characteristic [25,26]. The qPCR was performed with TB Green Premix ExTaqII (TaKaRa, Tokyo, Japan). The reactions were carried out in a total volume of 20 μL containing 2 μL of diluted cDNA (50 μg/μL), 1 μL of each primer, 10 μL of TB Green PCR Master Mix and 6 μL of H 2 O, with the following cycling profile: 94°C for 5 min, 40 cycles of 94°C for 30 s, 60°C for 30 s, and 72°C for 30 s. Each sample was processed in triplicate in the Roche LightCycler 480 Real-Time PCR System (Roche Diagnostics Burgess Hill, UK). The 2 -ΔΔCT method [81] was used to analyze the expression level.