The fish used for the experiment were the juvenile fish artificially hatched and cultured by our group. The fish were temporarily kept in a bucket of 1.5 cubic meters of water in a salinity of 28-30‰, a water temperature of 26-28 °C, and a DO of not less than 6 mg·L^-1. 180 fish of uniform size, healthy and vigorous with no damage on the body surface were selected for the salinity adaptation experiment after one week of temporary rearing, and the initial weight of the fish was 9.74 ± 0.85 g. The fish were divided into 10‰ salinity group, 30‰ salinity group and 35‰ salinity group with three biological replicates in each group. The fish were cultured in 9 buckets of 500 L size for 4 weeks, with 20 fish randomly placed in each bucket. The culture water salinity was adjusted downward by 4‰/d using fully aerated dechlorinated fresh water or upward by 4‰/d using sea crystals until the salinity of the experimental group reached the preset salinity and then the experiment was officially started. During the culture period, the water was fed twice daily with 6% body weight of commercial compound feed (46% crude protein and 8% crude lipid) without interruption of aeration and the water exchange rate was 30%.

Two sampling were conducted. In the first sampling, five R. canadum were randomly selected after seven days of culture, and eight tissues, including gill, intestine, body kidney, brain, stomach, muscle, spleen, and heart, were collected after anesthesia with eugenol (200 mg/L) for tissue distribution assay. In the second sampling, after four weeks of culture, five fish were randomly selected from each barrel, with a total of 15 fish. Among them,6 fish were anesthetized and three tissues of gill, intestine, and body kidney were taken for phenotypic analysis and qPCR detection, and 9 fish were mixed for transcriptome sequencing. All molecular samples were snap frozen in liquid nitrogen and stored at -80 °C after collection.

The fresh tissue was fixed using paraformaldehyde (4%) for 24 h. Afterwards, the tissue was orderly dehydrated using gradient alcohol, and the wax-soaked tissue was embedded in the embedding machine. And further the tissue was cut into slices with its thickness 4 μm, and the paraffin sections were dewaxed and further washed by distilled water. Lastly, the nucleus and cytoplasm were stained by hematoxylin and eosin, respectively.

For the complete identification of NHE and NKA gene family members in R. canadum, this study was based on the whole genome data of R. canadum (PRJNA634421) obtained in our laboratory and the NCBI public database, blast identification of NHE and NKA gene family members, recorded as the first round of screening results. The NHE and NKA gene family features were obtained from the Pfam database (http://pfam.xfam.org/) as PF00999 (Sodium/hydrogen exchanger family), PF00287 (Sodium/potassium ATPase beta chain), and PF00690 (Cation transporter/ATPase, N-terminus), respectively. The Hidden Markov Models (HMM) were used to obtain the features of the gene family. The HMMER 3.0 software was used to retrieve the whole-genome data, and the results of the second round of screening were tallied. Integrate the results, delete the mutilated or duplicate sequences and upload to SMART website and NCBI database for duplicate checks. The naming of NHE and NKA gene family members was based on reference comparisons and NCBI search results.

In order to further investigate the NHE and NKA gene family members, we conducted several analyses. Firstly, we determined the intron and exon length information, as well as the genomic localization, based on the genome annotation file gff. Additionally, we predicted the molecular weight (MW) and isoelectric point (PI) of the family members using the ExPasy website (http://web.expasy.org/). Moreover, we employed the MEME (Bailey et al., 2015) website (http://meme-suite.org/) to predict amino acid conserved motifs. Furthermore, we predicted the protein structural domains of NHE and NKA gene family members using the SMATR website. Finally, we utilized the TBtools software to map the NHE and NKA gene family structures and genomic localization.

The NHE and NKA family members from Homo sapiens, Mus musculus, zebrafish (Danio rerio), S. maximus, Rainbow Trout (Oncorhynchus mykiss), and L. japonicus were retrieved from the NCBI database. These sequences served as references for multiple amino acid sequence comparisons and homology analyses, which were performed using ClustalX1.83. The resulting phylogenetic tree was constructed by applying the neighbor-joining method (NJ) through MEGA-X (Kumar et al., 2016) software.

To investigate the impact of different salinity acclimation conditions on the expression patterns of NHE and NKA, RNA was extracted from the gills, intestine, and body kidney of R. canadum following 4 weeks of culture in salinities of 10‰, 30‰, and 35‰. RNA from nine fish in each salinity group was pooled to obtain one sample, and Illumina Hiseq™2000 was used to sequence the transcriptome. The raw mRNA sequencing data has been deposited in the NCBI Sequence Read Archive (SRA) under the accession number SRP202920 (published by our research groups) (Cao et al., 2020).

The raw mRNA sequencing data was processed using fastp (Chen et al., 2018) to remove low-quality data, and the remaining clean reads were mapped to the R. canadum genome (PRJNA634421) using HISAT2 software (Kim et al., 2015). StringTie software (Pertea et al., 2015) was then used to assemble the mapped reads. The expression of all genes in each sample (FPKM and reads count) was then calculated using RSEM (Li and Dewey, 2011), and the read count was normalized and analyzed for differentially expressed genes using edgeR (Robinson et al., 2010) (P<0.05 for significantly differentially expressed genes, FDR<0.05 and |log2FC|> 1 for highly significant differentially expressed genes).The resulting expression data (log₂ ^FPKM) were utilized to generate a gene expression heat map using TBtools, and correlation analysis was performed.

Gene-specific primers were designed based on the cDNA sequences for NHE and NKA gene family members, resulting in the amplification of fragments ranging from 100-230 bp ( Table 1 ). β-actin was chosen as the reference gene. qRT-PCR was performed using a Roche Light Cycler™ 96 real-time PCR machine and SYBR^®Select Master Mix. The expression levels of the three genes in 9 tissues and the expression levels of the genes in osmoregulatory organs such as gills, intestines, and kidneys after salinity adaptation were determined.

The amplification program consisted of an initial denaturation step at 95°C for 10 min followed by 40 cycles of denaturation at 95°C for 10 s, annealing at 60°C for 20 s, and extension at 72°C for 20 s. To minimize errors, three different R. canadum individuals were sampled for each salt treatment, and qPCR was repeated three times for each individual. The expression levels of NHE1, NHE2a, NHE2c, NHE5, NKAα1b, NKAα3a, NKAβ1b, and NKAβ3a genes were analyzed using the 2^-ΔΔCt method, and one-way ANOVA (LSD, Duncan) was performed using SPSS22.0 software.

After 30 days of domestication in low salinity water (10 ppt), the length (width) of gill filaments and gill lamellae of R. canadum increased significantly. The spacing between gill lamellae decreased, and the cells of gill lamellae were rounded and full. The number of chloride-secreting cells on gill filaments and gill lamellae decreased significantly. In the high salinity group (35 ppt), the number of chloride-secreting cells on gill filaments and gill lamellae increased slightly but not significantly. The width of gill filaments, gill lamellae, and cartilage tissues decreased significantly, and the spacing of gill lamellae increased ( Figure 1A ).

The microstructure of the intestine of juvenile R. canadum in the control group (30 ppt) showed that the single layer of columnar epithelium on the intestinal villi of juvenile R. canadum in the low salinity group became thicker, and the number of cupped cells decreased significantly. The size of the cupped cells did not change significantly. In the high-salinity group, the cytosol of the cup-shaped cells was enlarged, and the thickness of the unilamellar columnar epithelium and the number of cup-shaped cells on the intestinal villi did not change significantly ( Figure 1B ).

In the low salinity group, the tubular diameter of all levels of renal tubules of R. canadum increased, and the glomerulus was enlarged, full, and filled. The lumen of its capsule was small. In the high salinity group, the glomerulus atrophied, the lumen of the glomerular capsule increased, and the tubular diameter of all levels of renal tubules decreased slightly ( Figure 1C ).

In this study, the NHE and NKA families were characterized using genome−wide data and an HMM model. The analysis identified a total of 12 NHE family members and 12 NKA family members, including both single-copy and multi-copy genes. Specifically, NHE2 and NHE6 were found to be multi-copy genes, while NKA1, NKA3, and NKAβ1~3 were also identified as multi-copy genes. The coding sequence (CDS) of NHE genes ranged from 1818 to 2940 bp in length, with amino acid sizes ranging from 606 to 980 aa. The PI ranged from 5.5 to 9.45, and the Mw ranged from 67.04 kD to 107.53 kD. Similarly, the CDS of NKA genes ranged from 837 to 3099 bp in length, with amino acid sizes ranging from 279 to 1033 aa. The PI ranged from 5.01 to 8.09, and the Mw ranged from 32.67 kD to 113.47 kD ( Table 2 ).

The NHE family members of R. canadum had 2 to 16 Coding DNA Sequence (CDS), with NHE3, NHE7, NHE8, and NHE9 having 16 CDS, and NHEβ having the least number of CDS with only 2, which might be attributed to genome assembly issues. In contrast, NHE1 had the highest number of CDS with 15 ( Figure 2A ). The number of CDS of NKA family members ranged from 5 to 23, with NKAα members and NKAβ members showing polarized CDS numbers, where none of the former had less than 21 and all of the latter had less than 10 CDS, implying a correlation between CDS numbers and subtype classification ( Figure 2B ).

The present study aimed to analyze the motif composition of NHE and NKA genes in R. canadum using the MEME website. The results showed that both gene families contained 10 motifs arranged in an organized and regular manner. Most motifs of NHE genes were associated with Na⁺/H⁺ exchanger structural domains, except for motif4, while all motifs identified in NKA were associated with Cation transporter/ATPase and Hydrolase structural domains. Further analysis revealed that NHE motif1 and motif4 were mainly identified in NHE1~5, and motif 8 appeared only twice in NHE8 and once in all other members. In contrast, motif4 appeared in the anterior segment of the NHE9 sequence. Furthermore, NHEβ lacked motif7 and had more motif10 compared to NHE1( Figure 3A ). Concerning the NKA gene family, motif10 was found only in the NKAβ isoform, while the remaining nine motifs were ordered in the NKAα isoform ( Figure 3B ). The only difference was that motif4 was missing in NKAα1b.

The domain information of NHE and NKA genes in R. canadum was predicted using the SMART website, which revealed that NHE genes contained CPA1 and NHE structural domains ( Figure 3C ), and NKA genes contained NKA and CPA-N/C structural domains ( Figure 3D ).

Genomic localization shows that members of the NHE and NKA gene families of R. canadum localize to 10 and 9 superscaffolds, respectively ( Figure 4 ). Specifically, NHE2b and NHE6a were present simultaneously on superscaffold16 and superscaffold3, respectively, while the remaining eight members of NHE family were distributed randomly on a single superscaffold ( Figure 4A ). Similarly, 2-3 members of NKA family (NKAβ1b, NKAα1b, and NKAα1a; NKAα2 and NKAβ3) were simultaneously present on superscaffold24 and superscaffold14, respectively. In contrast, the remaining eight members of NKA family were randomly distributed on a single superscaffold ( Figure 4B ). Moreover, the multi-copy members of NHE family, NHE2a-c, were localized to the 9th, 3rd, and 24th superscaffold, respectively, whereas NHE6a and NHE6b were localized to the 3rd and 12th superscaffold, respectively ( Figure 4A ) multi-copy. Additionally, the multi-copy members of NKA family, NKAβ1-3 and NKAα3a and NKAα3b, were identified in the 7th, 24th, 23rd, 12th, 5th, and 14th superscaffold, respectively, and NKAα1a and NKAα1b were localized in both superscaffold 24 ( Figure 4B ).

The NHE and NKA gene family members of six species, including Homo sapiens, Mus musculus, D. rerio, S. maximus, O. mykiss, and Lateolabrax maculatus, were used as references to construct a phylogenetic evolutionary tree, which further verified the accuracy of the annotation of the NHE and NKA genes of R. canadum and revealed the variation of these genes during species evolution ( Figure 5 ). The analysis indicated that NHE1~9 of R. canadum were most closely related to teleost and furthest from mammals. NHE1 and NHEβ were clustered into a single clade with teleost such as D. rerio and S. maximus, respectively, before merging into one clade; NHE3 and NHE5 of R. canadum were merged into one clade, while H. sapiens and M. musculus were separate clades. The multi-copy genes NHE2a~c were clustered with other species and re-clustered with mammalian NHE4, respectively; NHE6a and NHE6b were alone, and their closest relatives were L. japonicus, S. maximus, D. rerio and O. mykiss ( Figure 5A ). Additionally, the NKAα subtype gene and NKAβ subtype gene of R. canadum were separately divided into the same branch with other species and were most closely located with teleost. Among them, NKAα4 of H. sapiens and M. musculus were separately merged with NKAβ1a of R. canadum. Moreover, NKAα1a~b and NKAα2 of R. canadum were closest together and merged into one branch, and NKAα1 and NKAα2 of H. sapiens and M. musculus were independently into one branch ( Figure 5B ).

To investigate the expression patterns of the NHE and NKA gene families in various tissues of R. canadum under normal seawater salinity, qRT-PCR was used to determine the gene expression abundance in nine different tissues. The results demonstrated that the NHE and NKA family members of R. canadum were widely expressed in all tissues, including the gill, brain, heart, intestine, kidney, liver, spleen, stomach and muscle( Figure 6 ). Specifically, the tissues with high expression of NHE were the gill, somatic kidney, and brain, while NKA was highly expressed in the gill, brain, and heart.

Of the single-copy genes, NHE1 and NHE3 showed similar expression patterns and were mainly concentrated in the gill and somatic kidney. On the other hand, NHE5 and NHE7 were highly expressed in the liver and brain, respectively. The highest expression signals of multi-copy genes NHE2a and NHE2b were detected in the brain and muscle, respectively, while the highest expression tissues were the gill and liver for NHE6a, and the somatic kidney and brain for NHE6b multi-copy. Furthermore, NHEβ and NHE1 expression patterns were inconsistent, with high expression detected only in the brain and trace expression in other tissues ( Figure 6A ).

The NKA family members NKAα1a, NKAβ1b, NKAβ2a, and NKAβ3a were highly expressed in osmolarity-regulating organs, such as the gill, intestine, and somatic kidney, respectively. Among them, NKAα1a was the most highly expressed in the gill. NKAα2 was expressed only in the brain and muscle, while NKAβ4 was highly expressed in the brain, with lower expression levels in other tissues. Most of the multi-copy genes (NKAα1b, NKAα3a, NKAα3b, NKAβ1a, NKAβ2a, NKAβ2b, NKAβ3a, and NKAβ3b) were highly expressed in brain tissues, with NKAα1b only detected as a fluorescent signal in brain tissues, not consistent with NKAα1a. NKAα3a and NKAα3b showed similar expression patterns, with high expression in the brain, heart, somatic kidney, and gill in descending order. NKAβ1a was highly expressed in several tissues, mainly in the brain, stomach, and muscle, while NKAβ1b was highly expressed mainly in the somatic kidney. NKAβ2a was highly expressed in the gill, intestine, brain, stomach, and muscle, while NKAβ2b was hardly expressed except in the brain and muscle. The expression pattern difference of NKAβ3a-NKAβ3b was similar to that of NKAβ2a-NKAβ2b. Additionally, fluorescent signals of other members were less frequently detected in the liver, spleen, and stomach, except for NKAα1a, NKAβ1a, and NKAβ2a, which showed expression ( Figure 6B ).

The effects of salinity acclimation on the expression patterns of NHE and NKA gene families in different tissues of R. canadum was investigated using a reference transcriptome sequencing approach and validated by qRT-PCR. The results demonstrated that the expression patterns of NHE and NKA genes were influenced by salinity and exhibited tissue-specific characteristics ( Figure 7 ). In gill tissues, the expression of NHE3 was significantly down-regulated under salinity 10 and 35 acclimation conditions compared to the control group acclimated at salinity 30. Additionally, NHE6b and NHE8 were significantly down-regulated under salinity 35 acclimation conditions, while NHE2a was significantly down-regulated under salinity 10 acclimation conditions. However, the experimental group exhibited significant upregulation in NHE2c expression, and NHE7 showed significant upregulation under salinity 35 acclimation conditions. In intestinal tissues, the expression levels of NHE genes were significantly increased under experimental conditions (salinity 10 and 35) compared to the salinity 30 control. Specifically, NHE7 was significantly upregulated under both low (10) and high (35) salinity acclimation conditions, while NHE1 and NHE6a-b were only significantly upregulated under salinity 35 acclimation conditions. NHE9 exhibited significant upregulation under salinity 10 acclimation conditions. In kidney tissues, NHE3 expression increased with higher salinity and showed highly significant upregulation at salinity 35, along with significant downregulation at salinity 10 compared to the salinity 30 control. NHE6a exhibited significant upregulation at salinity 35, whereas NHE2a showed significant downregulation at salinity 10. NHE5 and NHEβ expression levels remained relatively stable across all three salinity acclimation conditions in all tissues, while NHE2b and NHE5 expression levels were comparatively low under all conditions ( Figure 7A ).

In gill tissues, NKAα3a, NKAα3b, and NKAβ1a expression levels were significantly upregulated under salinity acclimation conditions of 10 and 35, with NKAα3a displaying particularly highly significant upregulation at salinity 10 compared to the control group at salinity 30. Conversely, NKAα1b expression was significantly downregulated at both salinities 10 and 35. In gill, intestinal, and kidney tissues, the expression levels of NKAα1a and NKAβ1b showed significant increases with increasing salinity. In intestinal tissues, NKAβ2a and NKAβ4 expression levels were significantly upregulated at salinity acclimation condition of 10, and exhibited highly significant upregulation at salinity acclimation condition of 35. Additionally, NKAα1b, NKAα3b, NKAβ1a, and NKAβ3a were significantly upregulated at salinity acclimation condition of 35 compared to the control. In kidney tissues, NKAβ1a expression was significantly downregulated at salinity 35, while NKAα3b showed significant upregulation at salinity 10 compared to the control group. However, the expression levels of NKAα2, NKAα3a, and NKAβ2b were relatively low under all conditions ( Figure 7B ). The qRT-PCR validation and RNA-seq were in general agreement in terms of the fold change in differential expression ( Figure 7C ). In the linear regression analysis of trend changes, R² = 0.8959 ( Figure 7D ). The results indicate that the gene expression analysis based on RNA-Seq data is reliable.

To investigate the differences in the expression patterns of some members of the NHE and NKA gene families of R. canadum under salinity acclimation, qPCR was performed on R. canadum acclimated to salinities of 10‰, 30‰ and 35‰ for 4 weeks in this study. The results showed that the relative expression of NHE1 was significantly up-regulated in the intestine and down-regulated in the kidney after increasing or decreasing salinity, while the expression was not significantly increased in the gills after decreasing salinity ( Figure 8A ). In addition, NHE2a expression in the gill did not change significantly in low-salt acclimation, while it was significantly down-regulated in high-salt acclimation ( Figure 8B ). NHE2c was significantly down-regulated in gill, intestine and somatic kidney with increasing salinity ( Figure 8C ), and NHE5 was also significantly down-regulated in somatic kidney. In addition, the expression pattern of NHE5 in both gill and intestine was significantly different in a “U” pattern ( Figure 8D ).

Following low-salt acclimation, significant down-regulation of NKAα1b and NKAβ3a was observed in the gills of R. canadum ( Figures 8E, H ), while NKAα3a and NKAβ1b did not exhibit significant changes ( Figures 8F, G ). In the intestine, NKAα1b and NKAβ1b did not exhibit significant changes ( Figures 8E, G ), whereas NKAα3a and NKAβ3a were significantly up-regulated and down-regulated, respectively ( Figures 8F, H ). In the somatic kidney, NKAα1b, NKAα3a, and NKAβ3a showed significant up-regulation ( Figures 8E, F, H ), while NKAβ1b did not change significantly ( Figure 8G ). Upon high-salt acclimation, the expression of NKAα1b and NKAβ3a in the gills remained consistent with low-salt acclimation ( Figures 8E, H ), while NKAα3a and NKAβ1b exhibited significant up-regulation ( Figures 8F, G ). In the intestine, the expression levels of NKAα1b, NKAα3a, and NKAβ1b were significantly increased ( Figures 8E, F, G ), while NKAβ3a expression levels were significantly decreased ( Figure 8H ). The expression pattern of NKAα3a and NKAβ3a in the somatic kidney was consistent with low-salt adaptation ( Figures 8F, H ), while NKAα1b and NKAβ1b expression were unaffected by salinity ( Figures 8E, G ).