All experiments were conducted in accordance with the guidelines and regulations outlined by the Management and Use of Laboratory Animals of Hubei Province and complied with China’s existing laws and regulations for biological research. The present study did not involve any endangered or protected species.

The experimental fish C. carpio (mean weight, 50 – 100 g) was obtained from the Guanqiao Experimental Station (Wuhan City, Hubei Province, China) and acclimatized at 25 ± 1°C. All the fish were fed twice daily with commercial pellets for 2 weeks before the experiments.

Naturally infected C. carpio was obtained from the temperature-controlled Fish Culture Center of the Institute of Hydrobiology, Chinese Academy of Sciences (Wuhan City, Hubei Province, China). The body surface of C. carpio was disinfected with 70% alcohol and placed in a biosafety container for dissection. A sterilized inoculum ring was inserted deep into the lesion of the fish, the needle was gently rotated for sampling, and streaked inoculation was performed on the surface of sterile brain heart infusion (BHI) broth, followed by incubation at a constant temperature of 28°C for 48 h. A single dominant colony in the plate was selected and inoculated on another BHI plate thrice, and the purified pathogenic strain was obtained after pure culture. Finally, one part of the purified strain was used for subsequent experiments and the other part was stored in 50% glycerol at −80°C for future use.

Bacteria were analyzed using bacterial biochemical microidentification tubes and test papers (Hangzhou Microbial Reagent. Co., Ltd. Hangzhou, China). A single colony was selected, inoculated in a microbiological reaction tube, and cultured at 28°C for 48 h. Following this, the biochemical indicators were tested (Table 1). The bacterial strains were preliminarily identified according to Bergey’s Manual of Systematic Bacteriology and Common Bacterial System Identification Manual (27).

The 16S rRNA of the pathogenic bacterial strain was amplified using the 16S rRNA primers 27F (AGTTTGATCMTGGCTCAG) and 1492R (GGTTACCTTGTTACGACTT). The PCR cycling conditions were as follows: initial denaturation at 95°C for 5 min; followed by 35 cycles at 95°C for 30 s, annealing at 58°C for 30 s, and extension at 72°C for 1 min; and a final extension at 72°C for 10 min. The two primers used for PCR amplification of the gyrB gene sequence were as follows: UP1 (forward): 5′-GAAGTCATCATGACCGTTCTGCAYGCNGGNGGNAARTTYGA-3′ and UP2r (reverse): 5′-AGCAGGGTACGGATGTGCGAGCCRTCNACRTCNGCRTCNGTCAT-3′. The PCR reaction conditions were as follows: pre-denaturation at 94°C for 5 min, denaturation at 94°C for 1 min, renaturation at 60°C for 1 min, and extension at 72°C for 1 min. After 30 cycles, incubation was performed at 72°C for 7 min. The reaction process was ended at 16°C, followed by preservation at 4°C. Amplification products were placed in 1% agarose gel for electrophoresis. Positive amplification products were recovered and sequenced by TSINGKE Biotechnology Co., Ltd (Wuhan, Hubei, China). The obtained 16S rRNA and gyrB gene sequences were analyzed by the NCBI BLAST retrieval system for sequence homology, and multiple alignments were made with the sequences of strains with high sequence similarity obtained from the GenBank database using ClustalX software. The neighbor-joining method was used to construct phylogenetic trees using Molecular Evolutionary Genetics Analysis (MEGA) 7.0 software. A bootstrap test was performed with 1000 repetitions.

The pathogenic bacterial strain was inoculated in BHI broth and cultured at 28°C under 200 rpm for 48 h. To determine the 50% lethal concentration of the pathogenic bacterial strain, four groups of C. carpio (n = 30) were intraperitoneally injected 100 μL of the bacterial strain (diluted in physiological saline) at concentrations of 1.0 × 10⁷ CFU/mL, 5.0 × 10⁷ CFU/mL, 1.0 × 10⁸ CFU/mL, and 5.0 ×10⁸ CFU/mL, while the control group was injected with sterile physiological saline (0.65% NaCl). Following this, mortality was monitored for the next 7 days.

After infecting the experimental fish (n = 100) with the measured half lethal concentration, the surviving fish were sampled from 0 to 7 days after the challenge. The blood, liver, head kidney, spleen, trunk kidney, and intestine (hindgut) of the challenged fish were randomly collected for the subsequent experiment (n = 5). Fish tissues injected with the same dose of saline were used as controls.

In total, 0.5 g of ground viscera (liver, head kidney, spleen, and intestine) was weighed and homogenized in BHI medium under sterile conditions using a Dounce tissue grinder (Sigma, USA). Serial dilutions (10⁻⁴, 10⁻⁵, and 10⁻⁶) were performed, 100 μL of the diluted sample was plated on BHI plates and incubated at 28°C for 48 h. Finally, the colony morphology was identified when the colonies were counted. CFU/g (log10) was calculated taking the dilution and total weight of the fish into consideration.

The whole target tissue was gathered and rinsed, and the immediate weight was recorded as the wet weight. Subsequently, the tissues were air dried for 2 days at 60°C, and their weight was recorded as the dry weight. The wet/dry weight ratio for each fish was calculated for evaluating tissue edema.

Tissue samples from the head kidney, spleen, liver, and intestine were dissected and immediately fixed in 10% neutral buffered formalin at 4°C, dehydrated in a graded ethanol series, leaned in xylene, embedded in paraffin, and cut into 4 μm sections using a rotary microtome (Leica, Germany). The sections were stained with HE staining for routine histological examination and observed using a light microscope (Olympus BX51, Japan).

Experimentally infected C. carpio was anesthetized with 3-aminobenzoic acid ethyl ester methanesulfonate (MS-222). Blood samples were collected from the caudal vein and placed for 1 h at room temperature. After centrifugation under 4500 rpm at 4°C for 15 min, the serum was collected and stored at −80°C. The serum biochemical indices of complement C3, superoxide dismutase (SOD), and lysozyme (LZM) were assessed using corresponding commercial kits (Nanjing Jiancheng Bioengineering Institute, Nanjing, China). The detection method refers to the kit instructions and previous reports (28).

Total RNA was extracted from six tissue samples (three from the control groups and three from the challenge groups, collected from the intestinal tissue after 3 days of P. shigelloides challenge) using TRIzol Reagent (Invitrogen, USA) and then treated with RNase-free DNase I (Thermo Scientific, USA) at 30°C for 30 min to remove genomic DNA contaminants. Subsequently, the purity, quantity, and integrity of the extracted total RNA samples were assessed using a NanoDropND-2000 spectrophotometer (Thermo, Waltham, USA), by 1.2% (w/v) agarose gel electrophoresis, and using an Agilent 2100 Bioanalyzer (Agilent Technologies, Richardson, USA), respectively. RNA samples with an RNA integrity number (RIN) > 8, 28S/18S > 0.7, and A260/280 of approximately 2.0 were used to construct the RNA-seq library. Poly (A) mRNA was isolated from total RNA using poly (dT) oligo-attached magnetic beads, and cDNA libraries were prepared using the TruSeq RNA Sample Preparation Kit (Illumina, USA). In total, six cDNA libraries were sequenced using the Illumina Nova-Seq (BGI) sequencing platform to generate 150-bp paired-end reads.

Read quality of the collected RNA-seq data was assessed using FastQC (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/) (29). The filtered clean data were subsequently mapped to the C. carpio genome (NCBI: ASM1834038v1) using HISAT2 v2.1.0 (30). All genes were annotated in the genome sequences available in six public databases, including NCBI’s non-redundant protein sequence (Nr) database, Swiss-Prot, Kyoto Encyclopedia of Genes and Genomes (KEGG) (31), Gene Ontology (GO) (32), evolutionary genealogy of genes: Non-supervised Orthologous Groups (eggNOG) (33), and Pfam (34).

DEGs between the control and challenge groups were identified using Baggerly’s test to calculate the fragments per kilobase of exon per million fragments mapped (FPKM) values. Benjamini–Hochberg correction was used to adjust the original P-values in Baggerly’s test in order to minimize the false discovery rate (FDR) (35). DEGs were identified if the associated PFDR was less than 0.05, and an absolute value of log2 (fold change) > 1 was regarded as the cut-off criterion. Clustering analysis was performed using the pheatmap package in R based on the FPKM values of DEGs. Subsequently, DEGs were enriched and subjected to GO term (36) and KEGG pathway analyses using clusterProfiler (37). A P-value < 0.05 indicated statistical significance, and the top 10 terms related to immunity were selected for visualization.

In total, six genes associated with immune-related gene testis development were randomly selected for qRT-PCR to validate the RNA-seq results. Reversed-transcribed cDNA obtained from total RNA used for transcriptome sequencing was synthesized using the PrimeScript™ RT Reagent Kit with gDNA Eraser (Takara, Shanghai, China), according to the manufacturer’s protocol. All cDNA samples were diluted to 5 ng/µL and stored at −80°C until use. Specific primers were designed on the basis of NCBI Primer-BLAST (NCBI, USA), as listed in Table S1. All qRT-PCR reactions were performed in triplicate, and target specificity was determined on the basis of dissociation curve analysis. β-actin was selected as the internal control to normalize the expression level of each gene. The relative expression level of the target gene versus the β-actin gene was calculated using the 2^−ΔΔCT method. The obtained data were statistically analyzed using GraphPad Prism 7.0 software.

The results have been reported as the means ± SEs after data preparation and statistical analysis using GraphPad Prism 7.0 software. Statistical significance of the findings in each experimental group relative to the control group was assessed using Student’s two-tailed t-test. Significance (P-value) has been indicated as follows: *(P < 0.05), **(P < 0.01), and ***(P < 0.001).

The symptoms of infected fish are shown in Figure 1; the symptoms mainly manifested as a mild floating head, outgroup swimming, slow swimming, and reduced feeding. In the middle stage of the disease, the fish stopped feeding completely, with slight eyeball protrusion, slight bleeding in the jaw and abdomen, and redness of the head (Figures 1A, B). In the later stage, death, abdominal bleeding, and swelling were noted (Figures 1C, D). After dissection, it was found that the dead fish had fluid accumulation (ascites) in the abdominal cavity. Moreover, the liver was swollen, with light yellow and irregular red spots distributed on it. Furthermore, the tissue felt slightly brittle, like bean paste (Figures 1E, F).

The dominant strain of pathogenic bacteria was isolated from the lesions of C. carpio (ulceration, ascites, hepatopancreas, and intestinal tract) and named Cc2021 (27). The bacterium grew well on BHI agar, and the colonies were round, beige opaque, smooth, moist, and slightly raised but with neat edges (Figure 2Aa). Gram staining results were microscopically observed. The bacterial cells stained red and appeared as short rods, which were arranged singly or in pairs, indicating that the strain was gram negative (Figure 2Ab). Negative stain electron microscopy revealed a straight rod with a terminal round shape, no spore or capsule, and two fascicular flagella at both ends. The size of the bacterium (short diameter × long diameter) was 0.8–1.0 μm × 3.0 μm (Figure 2Ac). These characteristics were consistent with the description provided in Bergey’s Manual of Systematic Bacteriology.

The physiological and biochemical characteristics of the pathogen are provided in Table 1. According to Bergey’s Manual of Determinative Bacteriology (9th ed.), the bacterium could hydrolyze arginine, D-fructose, fucose, glucose, and myo-inositol but not adipic acid, citric acid, D-mannitol, phenylalanine aminotransferase, phenylacetic acid, sucrose, salicin, sorbitol, and urease. The results of methyl red and malic acid tests were also positive.

To further identify the strain, its 16S rRNA and gyrB genes were sequenced. NCBI BLAST was used for homology analysis based on the 16S rRNA and gyrB gene sequences of the strain; the results revealed that the strain was 99% similar to P. shigelloides. Phylogenetic analysis was performed after screening sequences with high similarity in terms of sequence alignment; the results (Figures 2B, C) revealed that both 16S rRNA and gyrB gene sequences of P. shigelloides formed a single cluster. Therefore, based on its morphological, physiological, biochemical, and molecular characteristics, the strain was identified to be P. shigelloides.

To assess the pathogenicity of P. shigelloides in C. carpio, one group of C. carpio was injected with physiological saline (65% NaCl) (control group), while four groups were injected with four different concentrations (1.0 × 10⁷ CFU/mL, 5.0 × 10⁷ CFU/mL, 1.0 × 10⁸ CFU/mL, and 5.0 × 10⁸ CFU/mL) of the pathogenic bacterial strain (experimental groups). Subsequently, deaths were recorded (Table 2). Both high and low concentrations of the bacterial suspension resulted in typical symptoms, similar to those observed in naturally infected fish. After the injection of a high concentration (5.0 × 10⁸ CFU/mL) of the bacterial suspension, a large number of fish died within 2 days (73.3%), with the mortality reaching 100% within 7 days. The mortality rates of C. carpio were 80%, 53.3%, and 23.3% after the injection of 1.0 × 10⁸ CFU/mL, 5.0 × 10⁷ CFU/mL, and 1.0 × 10⁷ CFU/mL concentrations, respectively. On the other hand, no death was noted in the control group. These mortality results served as the basis for the follow-up regression challenge experiment. Tissue edema and bacterial loads in C. carpio were assessed after P. shigelloides challenge. The dry/wet weight ratios of the liver, spleen, head kidney, and body kidney of C. carpio showed an increasing trend with an increase in the challenge time. After 2 days post-challenge, the dry/wet weight ratios of the liver and head kidney gradually increased, while those of the spleen and head kidney tended to decrease (Figures 3A–D). The bacterial load in the tissues from the intestine, hepatopancreas, head kidney, body kidney, and spleen of C. carpio infected with P. shigelloides and the blood bacterial concentrations increased with an increase in the challenge time. The bacterial load in the intestinal tissues was the highest, followed by that in the tissues from the hepatopancreas, head kidney, and spleen. The lowest bacterial load was noted in the blood and body kidney tissues (Figure 3E).

To observe the histopathological state of P. shigelloides-infected C. carpio, HE staining was performed using tissues from the liver, head kidney, spleen, and intestine at each period. The typical pathological states are shown in Figure 4. In the control group, the hepatic cells were rich in cytoplasm, had an obvious nucleus and cytoplasm, had a clear cell boundary, and were evenly dispersed (Figure 4Aa). After the artificial challenge, the inflammatory cells in the liver of infected C. carpio showed focal aggregation, with an unclear structure and a fuzzy boundary. The hepatic cells were denatured and necrotic, and the nucleus and cytoplasm were concentrated. Some hepatic cells were vacuolated and had dissolved nuclei. The tissue was loose and had edema, and degeneration and necrosis were noted around the sinuses. Moreover, the tissue structure was destroyed and several hepatocytes were vacuolated (Figure 4Ab). The renal structure of infected C. carpio was disorganized, with swelling and degeneration of cells, a narrow renal cavity, and a transparent tube type. The renal tissue was filled with lymphocytes, accompanied by multiple bleeding spots and an obvious inflammatory reaction (Figure 4Bb). Moreover, ferriflavin deposition in the splenic tissue, atrophy of the white pulp, and a reduced area occupied by splenic pulp hyperemia were noted (Figure 4Cb). The intestinal structure of C. carpio in the control group was intact; the inner wall of the tissue was not damaged, intestinal epithelial villi were arranged tightly and in an orderly manner, and basal epithelial tissues were obviously stratified (Figure 4Da). Villus epithelial cells in the intestinal tissues of infected C. carpio fell off, with the infiltration of local lymphocytes and the swelling of and increase in goblet cells in the mucosal epithelium being noted (Figure 4Db).

To assess the degree of activation of antibacterial immune responses in the key immune organs and bloodstream of C. carpio during P. shigelloides challenge, C3, SOD, and LZM activities were assessed at different stages after the challenge. The expression level of C3 in the serum increased from 1 to 7 days and reached its maximum on the 5th day post-challenge, becoming nearly three times the normal expression level (Figure 5A). The expression levels of LZM in all the tissues of healthy common carp were similar. The LZM levels in the blood, head kidney, liver, and intestine significantly increased after the challenge. Moreover, the LZM levels in all the tissues continued to increase from 1 to 7 days post-challenge. LZM activity was the highest in the intestine, followed by the liver and blood, while it was the lowest in the head kidney after the challenge (Figure 5B). The SOD levels in each tissue of C. carpio showed a trend of first increasing and then decreasing after the challenge; the increase was most obvious on 1 and 3 days, followed by a decrease (Figure 5C). The highest SOD activity was noted in the intestinal tissue after the challenge; this may be related to enteritis caused by the bacterium.

All raw data generated in the present study have been uploaded to the NCBI Sequence Read Archive (SRA) under the registration number PRJNA793911. The original sequencing data included reads having low quality, joint contamination, and a high unknown base N content, which needed to be removed before data analysis in order to ensure the reliability of the results. After sequence filtration, a comparison of all genomes was performed, as shown in Figure S1A. Expression profiles of all the genes were obtained in both the control and challenge groups, with half reaching at least one FPKM (Figure S1B). Correlation analysis of the samples revealed that the correlation coefficients of Challenge 2 and Control 1 within the group were far less than 0.7; thus, these two samples were removed from subsequent analysis (Figure S1C). Analysis of FPKM expression levels for all the genes in the visualized heatmap revealed that the expression levels of Control 2 and Control 3 genes were consistent and those of Challenge 1 and Challenge 3 genes were consistent (Figure S1D).

To further assess the immune response mechanism of C. carpio during P. shigelloides challenge, comparative transcriptome analysis was performed for comparing the gene expression profiles in the intestinal tract of C. carpio between the control and challenge groups. For this analysis, three control group samples and three challenge group samples, collected on the 3rd day after the infection, were included. Based on sequence homology, GO enrichment analysis of DEGs in the intestinal tract was performed after P. shigelloides challenge, and the similarities and differences in these terms were analyzed. To further understand the functions of DEGs and the signaling pathways that they participate in, all DEGs were classified into three GO categories: biological process (BP; 20 subclasses), molecular function (MF; 20 subclasses), and cellular component (CC; 20 subclasses) (Figure S2A). KEGG pathway analysis of the intestinal tissue after P. shigelloides challenge revealed that there were 21 KEGG metabolic pathways involving different genes, mainly including the complement and coagulation cascade pathway, TNF signaling pathway, interleukin (IL)-17 signaling pathway, and toxoplasmosis and systemic lupus erythematosus pathway (Figure S2B).

The identification of DEGs can provide a deeper understanding of the changes in the immune mechanism of P. shigelloides-infected C. carpio. The heatmap in Figure 6A shows the expression of DEGs in each group. Intestinal gene expression was higher in the challenge group. Importantly, there was no significant difference in DEGs between the control and challenge groups. In total, 1704 genes were differentially expressed in the control and challenge groups, including 876 upregulated genes and 828 downregulated genes (Figure 6B). The upregulation level of genes was thus significantly higher than the downregulation level, which may be attributed to the cascade amplification of immune signal transmission.

To further assess the immune response after P. shigelloides challenge, we focused on immune-related DEGs classified into BP, MF, and CC categories, as shown in Figure 6C. The top 10 KEGG pathway terms are shown in Figure 6D. GO enrichment analysis of this module was dominated by functional categories related to immune system. In the BP category, acute inflammatory response, regulation of acute inflammatory response, and leukocyte chemotaxis were the top three terms. In the CC category, platelet alpha granule lumen, platelet alpha granule, and plasma lipoprotein particle were the top three GO terms. In the MF category, CXCR3 chemokine receptor binding, chemokine activity, and C3HC4-type RING finger domain binding were the top three terms. Furthermore, KEGG pathway enrichment of immune-related DEGs was analyzed. The immune-related signaling pathways of all enriched immune-related DEGs were selected as the top 10 ones, e.g., the TNF signaling pathway, IL-17 signaling pathway, antigen processing and presentation pathway, Chagas disease pathway, and Toll-like receptor (TLR) signaling pathway (Figure 6D). Some immune-related DEGs are shown in Table 3. These results suggested that both the innate and adaptive immune responses of C. carpio played vital roles in its response to P. shigelloides challenge. The discovery of the immune-related pathways and genes could provide a theoretical basis to understand the molecular mechanism of C. carpio infected with P. shigelloides or other bacteria.

To verify significant immune response candidates activated by P. shigelloides, qRT-PCR was performed on several candidates selected from the comparative transcriptome. The expression levels of IL-1β, Hsp70, TLR5a, MCH II, CD22, and IL-8 were verified in the blood, spleen, head kidney, and intestine at 0, 1, 3, 5, and 7 days after the challenge (Figure 7). The mRNA expression level of IL-1β dramatically increased and immediately reached the highest level on the 1st day after the challenge, following which it began to decrease gradually. The expression level of IL-1β was the highest in the intestine (Figure 7A). The mRNA expression level of Hsp70 gradually increased in all immune tissues from 1 to 7 days after the challenge; the expression level was higher in the head kidney and spleen (Figure 7B). The mRNA expression level of TLR5a was upregulated to a similar extent in all the tissues after the challenge, with the increase being the highest at 3 and 5 days (Figure 7C). Moreover, the mRNA expression level of MHC II was continuously upregulated after the challenge, with the most significant increase being noted in the intestine (Figure 7D). In contrast, the mRNA expression levels of CD22 and IL-8 were downregulated to varying degrees after the challenge in all the tissues (Figures 7E, F).