Innate and early immune responses to Yersinia ruckeri in the spleen of rainbow trout (Oncorhynchus mykiss)

Background: Yersinia ruckeri is a pathogen that can cause enteric redmouth disease in salmonid species, damaging global production of economically important sh including rainbow trout (Oncorhynchus mykiss). Herein, we conducted the transcriptomic proling of spleen samples from rainbow trout at 24 h post-Y. ruckeri infection via RNA-seq in an effort to more fully understand their immunological responses. Results: We identied 2498 differentially expressed genes (DEGs), of which 2083 and 415 were up- and down-regulated, respectively. We then conducted a more in-depth assessment of 78 DEGs associated with the immune system including CCR9, CXCL11, IL-1β, CARD9, IFN, CASP8, NF-κB, NOD1, TLR8α2, HSP90, and MAPK11, revealing these genes to be associated with 20 different immunological KEGG pathways including the Cytokine-cytokine receptor interaction, Toll-like receptor signaling, RIG-I-like receptor signaling, NOD-like receptor signaling, and MAPK signaling pathways. Additionally, the differential expression of 8 of these DEGs was validated by a qPCR approach and their immunological importance was then discussed. Conclusions: Our ndings provide preliminary insight regarding the innate and early immune responses of rainbow trout following Y. ruckeri infection and the base for future studies of host-pathogen interactions in rainbow trout. Nucleotide-binding oligomerization domain-containing 1-like;


Background
Yersinia ruckeri is a pathogen that can cause enteric redmouth disease (ERM) or yersiniosis, resulting in signi cant mortality and economic losses associated with the global production of rainbow trout (Oncorhynchus mykiss). Rainbow trout are highly susceptible to ERM, although other species of sh can also be affected by this disease [1,2]. Multiple studies have sought to clarify the immunological responses of sh species to Y. ruckeri infection. In one study, Raida et al. determined that very susceptible trout species exhibited a robust and rapid-onset septicemic response to infection associated with the production of high levels of pro-in ammatory cytokines [3]. Similarly, these pro-in ammatory cytokines were also upregulated in the spleen of the vaccinated rainbow trout following Y. ruckeri challenge, albeit to a lesser extent than in naïve sh [4]. The spleen is a key secondary lymphoid organ that is thus closely associated with rainbow trout responses to Y. ruckeri infection, and signi cant changes in the expression of splenic immune-related genes have been detected following Y. ruckeri challenge [5,6]. However, no systematic analyses of patterns of rainbow trout splenic gene expression after Y. ruckeri infection have been conducted to date.
RNA sequencing (RNA-seq) is a high-throughput approach to analyzing transcriptomes that has frequently been employed in studies of sh species [7]. Several recent studies based on RNA-Seq analysis have explored rainbow trout responses to a range of pathogen types, such as splenic responses to Aeromonas salmonicida [8,9], infectious hematopoietic necrosis virus (IHNV) [10], and Ichthyophthirius multi liis [11]. Such transcriptomic analyses have offered new insights into the etiology of these diseases, and similar studies of Y. ruckeri infections may highlight viable approaches for treating or preventing yersiniosis in rainbow trout farming.
As such, we herein conducted a transcriptomic study assessing rainbow trout splenic immune responses to Y. ruckeri infection. After identifying infection-related differentially expressed genes (DEGs), we validated a subset of these genes via qPCR and conducted the functional annotation of immuneassociated DEGs. Together, our data offer a preliminary insight for future research regarding the immunological mechanisms involved in rainbow trout defensive response against Y. ruckeri.

RNA-sequencing and data processing
Genes associated with rainbow trout immune response to Y. ruckeri infection were identi ed by assessing spleen samples from YR-infected and control uninfected sh via RNA-sEq. In total, six cDNA libraries were prepared (from 3 per group), and raw data were generated (Table S1) and deposited in the NCBI Sequence Read Archive (SRA) under accession number PRJNA675742.
Following the completion of ltering, 44.07 G bp of clean data were extracted, with over 93.15-93.55% of the bases reads having a phred quality value ≥ 30 in the non-infected group compared to 92.87-93.43% in the YR-infected group. These quality scores were consistent with excellent quality data. Reads from these two groups exhibited GC contents of 49.14-49.64% and 49.00-49.18%, respectively (Table 1). The total number of expressed genes detected in samples from uninfected rainbow trout was slightly higher than that detected in YR-infected rainbow trout (Fig. 1

DEG Identi cation And Analysis
The Pearson's correlation coe cient values were used to assess relative gene expression in the uninfected and YR-infected groups (Fig. S1). A total of 2498 DEGs were identi ed by comparing these groups, of which 2083 (83.39%) were up-regulated and 415 (16.61%) were down-regulated, in YR-infected sh compared to uninfected sh (Table S3). Volcano and MA plots were also used to represent these gene expression trends (Fig. S2).
Of these DEGs, 2431 were classi ed successfully using the NR, Swiss-Prot, GO, COG, KOG, Pfam, and KEGG databases (Table 3). With respect to new genes, many DEGs were annotated using the NR and eggNOG databases, but few were annotated in the COG database.  (Fig. 2).
In addition, KEGG pathway enrichment analyses were performed to assess the functional roles of these DEGs during Y. ruckeri infection in rainbow trout. Assembled DEGs were analyzed with the KEGG database, leading to their classi cation into 6 categories ( Fig. S3). KEGG enrichment results, including the top 9 pathways enriched for > 50 genes (p < 0.05), are shown in Fig. 3. Four highly enriched pathways were detected through this KEGG analysis, including the NOD-like receptor signaling, cytokine-cytokine receptor interaction, Toll-like receptor signaling, and RIG-I-like receptor signaling pathways. The preferential enrichment of these pathways suggests that many of the genes differentially expressed between uninfected and YR-infected rainbow trout were related to the immune system.

Identi cation Of Immune-related DEGs
To better understand the intracellular signaling pathways during Y. ruckeri infection in rainbow trout, we therefore focused on 78 immune response-related DEGs identi ed in this study, including two new genes (Table S4). A heatmap was constructed based upon the fold-change expression values for these DEGs (Fig. 4), clearly demonstrating that almost all of these genes (74) were upregulated in spleen samples from YR-infected sh compared to spleen samples from uninfected sh, whereas only 4 genes were down-regulated after infection.
Further analysis of these immune-related DEGs revealed them to be primarily associated with 20 immunological KEGG pathways, including the MAPK signaling, Cytokine-cytokine receptor interaction, Toll-like receptor signaling, RIG-I-like receptor signaling, NOD-like receptor signaling, FoxO signaling, mTOR signaling, apoptosis, TGF-beta signaling, regulation of autophagy, ErbB signaling, cell adhesion molecule (CAM), intestinal immune network for IgA production, cytosolic DNA-sensing, phosphatidylinositol signaling system, and p53 signaling pathways ( Table 4). The top 3 pathways enriched in these genes included the NOD-like receptor signaling (31 genes), RIG-I-like signaling (35 genes), and Toll-like receptor signaling (51 genes) pathways (Fig. 5). As expected, all the eight immune-related DEGs exhibited similar expression trends when measured via both qPCR and RNA-Seq analysis, con rming the reliability of our analytical techniques (Fig. 6).

Discussion
ERM is a serious disease that impacts global salmonid populations [12]. While some studies have begun to characterize rainbow trout immune responses to Y. ruckeri infection [6,13], no systematic transcriptomic analyses of these responses have been conducted to date. The spleen plays central roles in orchestrating innate and adaptive immune responses in sh. Herein, we sequenced the spleen transcriptomes of rainbow trout infected with YR in comparison with those of control uninfected rainbow trout and we identi ed 2498 DEGs between these populations, of which 2083 were up-regulated whereas 415 were down-regulated in infected rainbow trout. Immune response-related DEGs were then assessed in additional detail in an effort to explore the basis of innate immune responses against Y. ruckeri infection in rainbow trout.
Cytokines are secreted by a range of cell types, and they act as immune response regulators that can be classi ed as interleukins (ILs), interferons (IFNs), tumor necrosis factors (TNFs), and chemokines [14]. Of the 78 immune-associated DEGs in the present study, 31 were classi ed into the cytokine-cytokine receptor interaction pathway, including chemokine (C-X-C motif) ligand (CXCL11), C-C motif chemokine receptor 9 (CCR9), caspase recruitment domain-containing protein (CARD9), IL-12, IL-1β, and IFN. Chemokines control the migration of particular immune cell subsets and coordinate both adaptive and innate immune responses to stressors [15]. The transcription of CXCd in rainbow trout has previously been shown to be induced in response to Y. ruckeri infection [16]. Herein, we observed the upregulation of both CXCL11 and CCR9 in the spleens of rainbow trout infected with this bacterium, consistent with the pathogen-induced chemokine regulation. CARD9, which is normally activated by CLRs [17], was 1.39-fold downregulated in response to Y. ruckeri.
Apoptosis is an important determinant of cellular survival in both physiological and pathological contexts, and can be triggered by factors such as hypoxia, chemical exposure, temperature stress, or immune responses to particular stimuli. Upon bacterial infection, a host's cells may undergo apoptotic death to mitigate the spread of the pathogen within host tissues [18]. Herein, we observed the upregulation of caspase 8 (CASP8), receptor-interacting serine/threonine-protein kinase 1-like (RIPK1), transcription factor p65-like (TF65), inhibitor of apoptosis protein-like protein (IAP), and NF-kappa-B inhibitor alpha-like (IκBα) following YR infection in rainbow trout. Caspases are proteases that serve as essential regulators of apoptotic cell death, with CASP8 having showed to be an upstream regulator of apoptotic cascades in sh [19]. Marked CASP8 upregulation has also previously been detected in headkidney and spleen leukocytes of Totoaba macdonaldi at 24 h post-infection with Vibrio parahaemolyticus and Aeromonas veronii [20]. NF-κB can control innate and adaptive immune-related gene expression, inducing apoptosis in response to numerous stimuli [21]. At the same time, NF-κB activation induces IκBα expression in rainbow trout, in turn resulting in the feedback inhibition of NF-κB [22]. Upregulation of IkBα, IAPs and RIPK1 detected in this study can suggest the compensatory activation of some inhibitors of apoptotic cell death, underscoring the complexities of cellular responses to Y. ruckeri in rainbow trout. Additional work must be done in order to understand in depth how the apoptotic processes.
Pattern recognition receptors (PRRs) serve as innate sensors that can rapidly detect and respond to conserved damage-and pathogen-associated molecular patterns (DAMPs and PAMPs, respectively), resulting in the induction of immune-related gene expression and anti-pathogen responses. PRRs detected in aquatic species to date include TLRs, NLRs, RLRs, and CLRs [23]. In the present study, we identi ed over 10 members of the TLR, NLR, and RLR gene families that were differentially expressed in the spleens of rainbow trout at 24 h post-Y. ruckeri infection, including heat shock protein 90 (HSP90), tumor necrosis factor alpha-induced protein 3-like (TNFAP3), nucleotide-binding oligomerization domaincontaining protein 1-like (NOD1), IL-1, IL-12, toll-like receptor 8α2 (TLR8α2), and interferon α (IFNα). HSPs are important regulators of sh immune responses [24,25], and HSP90 upregulation detected in the present research may be linked to the rainbow trout innate immune defenses to Y. ruckeri infection. NOD1 modulates the innate immune response of sh to bacterial peptidoglycan. Loss-and gain-of-function experiments have suggested that NOD1 can control rainbow trout pro-in ammatory cytokines in rainbow trout [26]. Palti et al. rst reported the presence of the TLR8α2 gene in rainbow trout, which they found to be somewhat downregulated in response to treatment with the human agonist of TLR7/8 known as R848 [27]. We found both NOD1 and TLR8α2 to have been downregulated in rainbow trout during the early stages of Y. ruckeri infection. Overall, these ndings suggest that the PRRs were differentially expressed in rainbow trout and may be important mediators of the initial induction of immunological responses to bacterial infection.
The MAPK signaling pathway is responsive to diverse extracellular stimuli and can modulate transcription factor expression and activation, controlling a range of biological processes including proliferation, apoptosis, and gene transcription. Recent evidence indicates that sh MAPKs can be induced by a range of stimuli. For example, agellin treatment is associated with MAPK11 upregulation in rock bream (Oplegnathus fasciatus) [28]. In contrast, in the present study we observed a 2.58-fold decrease of MAPK11 expression in the spleen of rainbow trout following Y. ruckeri infection, although additional validation of these results is warranted. We also observed a signi cant MAPK8 upregulation upon Y. ruckeri infection in rainbow trout. Moreover, a total of 22 DEGs involved in the MAPK signaling pathway seems to play key roles in the rainbow trout response to infection with this bacterium.

Conclusions
In summary, we conducted a transcriptomic analysis of spleen samples from rainbow trout infected with Y. ruckeri in an effort to better understand the immunological basis for responses to this pathogen, leading to the identi cation of several key immune-related DEGs. Overall, our results will provide a preliminary insight regarding the innate and early immune responses of rainbow trout following Y. ruckeri infection and the base for future studies of host-pathogen interactions in rainbow trout.

Experimental sh and bacteria
Healthy rainbow trout (~ 10 g) were obtained from Benxi Agrimarine Industries Inc. and housed in a 540 L berglass circulating water tank at a constant temperature of 14 ± 0.2℃ with a 12 h light/dark cycle and an 8.0 mg/L oxygen saturation. Fish were maintained under these conditions for 2 weeks and were fed commercial rainbow trout feed.
Y. ruckeri strain BH1206 was isolated from infected rainbow trout, con rmed to be pathogenic, and used for challenge experiments as previously published [29]. Bacteria were grown for 24 h in TSB medium (BD Difco, USA) and collected by spinning for 5 min at 6,000 xg prior to resuspension in sterile PBS (pH 7.2) at 6×10 7 CFU· mL − 1 .

Bacterial Challenge And Sampling
Prior to challenge test, healthy rainbow trout were kept under laboratory conditions in ow-through tanks at approximately 14°C with continuous aeration and fed twice a day at 1.2% of body weight with commercial sh feed. A subset of experimental sh was microscopically and bacteriologically examined to verify freedom of Y. ruckeri infection. Tricaine methanesulfonate (MS222) was used to anesthetize sh prior to the challenge or tissue sample collection. Experimental infection was induced by intraperitoneally (i.p.) injecting sh with 100 µL of BH1206 bacteria at 6×10 5 CFU per gram of sh body weight. An equivalent volume of PBS was injected into uninfected control sh. At 24 h post-infection, three sh per group were sacri ced by an overdose of anesthetic, and spleens were collected, washed to remove blood and fat, snap-frozen with liquid nitrogen, nally stored in liquid nitrogen tank. To con rm the presence of Y. ruckeri in experimental sh, the kidney was sampled to perform bacteriological examination.

RNA Isolation
Splenic RNA was isolated using Trizol (Invitrogen, USA), after which RNA integrity and purity were evaluated via 1% agarose gel electrophoresis and using an Agilent 2100 Bioanalyzer (Agilent Technologies, CA, USA), while a Qubit RNA Assay Kit and a Qubit 2.0 Fluorometer (Life Technologies, CA, USA) were utilized to measure RNA concentration. After RNA preparation, all downstream library preparation and sequencing were performed by Biomarker technologies CO., LTD (Beijing, China).

Library Construction And Sequencing
A total of 3 µg RNA per spleen sample was utilized for library construction using a NEBNext Ultra RNA Library Prep kit for Illumina (NEB, USA), with samples being a xed with appropriate barcodes. Following DNase I treatment, the remaining mRNA was puri ed and sheared into 200-250 bp fragments as discussed previously [30]. Library quality was assessed with an Agilent Bioanalyzer 2100 instrument, and a cBot Cluster Generation System with TruSeq PE Cluster Kit v4-cBot-HS (Illumina) was used to cluster barcoded samples. An Illumina Hiseq 2500 platform was then used for the paired-end sequencing of these prepared library samples.

Data Processing
Raw data were initially cleaned by removing reads that contained adapter sequences, poly-N sequences, and low-quality reads with the FastQC program (http://www.bioinforatics.babraham.ac.uk/projects/fastqc/), after which clean data Q30, GC-content, and sequence duplication levels were calculated. The Trinity software [31] was then used to assemble reads into EST clusters, followed by de novo assembly and alignment to the rainbow trout reference genome (http://www.genoscope.cns.fr/trout/data/) with TopHat (v.2.0.5). Functional annotation was performed by comparing unigenes to the following databases: Nr (NCBI non-redundant protein sequences); Nt (NCBI non-redundant nucleotide sequences);Pfam (Protein family); KOG/COG (Clusters of Orthologous Groups of proteins) [32]; Swiss-Prot (A manually annotated and reviewed protein sequence database); KO (KEGG Ortholog database) [33]; GO (Gene Ontology) [34].

DEGs Identi cation
The RSEM software was used to assess unigene expression based upon reads per kilobase of exon per million mapped reads (RPKM) [35]. The DESeq R package (1.10.1) was used to identify DEGs between infected and non-infected sh using a negative binomial distribution-based model, with P values being adjusted as indicated by the Benjamini and Hochberg approach to reduce the false discovery rate. DEGs were considered as those genes with an adjusted P-value < 0.05, and were represented with volcano and MA plots. The Kyoto Encyclopedia of Genes and Genomes (KEGG) database was used for functional enrichment analysis of DEGs, with pathways that had a Q-value ≤ 0.05 after correcting for multiple testing being considered signi cantly enriched.
Validation Of Immune-related DEGs By QRT-PCR To con rm the results of RNA-sequencing, eight immune-related DEGs (IL-1β, IL-8, TNF, NOD1, CARD9, TLR8α2, CCR9, and HSP90) were chosen for qRT-PCR-based validation using the same RNA samples prepared for RNA-seq using primers designed with the Premier primer 5 software (Table 5). EF-1α was used as a normalization control for these analyses. SYBR Green dye (Takara, China) and an ABI PRISM 7500 Fast Real-time PCR instrument were used for qRT-PCR based on provided protocols. All reactions were conducted in triplicate with the following thermocycler settings: 60 s at 95℃; 40 cycles of 15 s at 95℃, 45 s at 60℃. Melt curve analyses were conducted to con rm the speci city of ampli cation products. Relative gene expression was assessed via the 2 −△△CT approach [36].   Immune-related DEGs in the non-infected and YR-infected rainbow trout Comparison of DEG expression in qPCR and RNA-seq analyses. Relative gene expression levels were normalized to EF-1α

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. TableS1.xlsx TableS2.xlsx TableS3.xlsx