Transcriptomics and proteomics provide insights into the adaptative strategies of Tibetan naked carps (Gymnocypris przewalskii) to saline-alkaline variations

Abstract

Gymnocypris przewalskii is an exclusively cyprinid fish that inhabits Lake Qinghai, which is characterized by high salinity and alkalinity. To elucidate the molecular basis of the adaptation of G. przewalskii to a wide range of salinity‒alkalinity conditions, we performed morphological, biochemical, transcriptomic and proteomic analyses of the major osmoregulatory organs of the gills and kidney. Morphological examination revealed that mitochondria-rich cells were replaced by mucus cells in the gills during the transition of G. przewalskii from freshwater to lake water. In the kidney, the tight junction formed dense structure in the renal tubules under lake water condition compared with the loose structure in freshwater. The results of the biochemical assays revealed an increased content of total amino acids, indicating their potential roles as osmolytes and energy supplies in freshwater. The decreased urea concentration suggested that urea synthesis might not be involved in the detoxicity of ammonia. The transcriptomic and proteomic data revealed that genes involved in ion absorption and ammonia excretion were activated in freshwater and that genes involved in cell junction and glutamine synthesis were induced in lake water, which was consistent with the morphological and biochemical observations. Together with the higher levels of glutamine and glutamate, we proposed that G. przewalskii alleviated the toxic effect of ammonia direct excretion through gills under freshwater and the activation of the conversion of glutamate to glutamine under high saline-alkaline condition. Our results revealed different expression profiles of genes involved in metabolic pathways, including the upregulation of genes involved in energy production in freshwater and the induction of genes involved in the synthesis of acetylneuramic acid and sphingolipid in soda lake water. In conclusion, the appearance of mitochondria-rich cells and increased energy production might contribute to ion absorption in G. przewalskii to maintain ion and solute homeostasis in freshwater. The existence of mucus cells and dense junctions, which are associated with increased gene expression, might be related to the adaptation of G. przewalskii to high salinity-alkalinity.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12864-025-11336-z.

Introduction

Salinity-alkalinity is a critical environmental stressor that poses critical challenges to freshwater fish populations [1]. The adaptability of fish to saline-alkaline water has long been of great interest in biology and ecology. Euryhaline fishes have evolved independently in different lineages of teleosts, suggesting that diverse and complex osmo- and iono-regulatory strategies might be adopted in response to osmotic fluctuations [2]. Hence, these fish represent ideal models for investigating the molecular and evolutionary basis underlying the adaptive responses of fish to saline-alkaline variations. For example, physiological and morphological observations have revealed alterations in gill cell composition and junction, plasma endocrine levels and metabolites in muscle in response to saline-alkaline variation in many euryhaline fishes [3, 4]. Genomic analysis revealed that genes that function in cell junction, ion transport and transcriptional regulation were under positive selection and rapid evolution in the adaptation of cyprinid fishes to seawater and high salinity-alkalinity condition [5, 6]. Transcriptomic analyses revealed the transcriptional variations of genes involved in ion homeostasis, protein folding processing, endocrine regulation, and metabolism in diverse fish species, suggesting that the orchestration of cellular processes and tissue functions in osmoregulation [7–10]. These studies indicate that osmoregulation is an essential adaptive change for fish to address salinity‒alkalinity fluctuations; however, how euryhaline fishes maintain osmotic homeostasis at the molecular level are not fully understood [11].

Gills and kidneys are the primary sites for osmo- and iono-regulation in the response of fish to saline-alkaline alteration [12–14]. To counter the diffusive loss of ions and influx of water under hypotonic conditions, the gills and kidneys activate ion absorption, limit the inflow of water, and produce diluted urine while excreting excessive ions in the high saline-alkaline water [15–17]. These functional shifts involve morphological remodeling and cellular rearrangement, accompanied by gene expression regulation [18]. The appearance of mitochondrial-rich cells (MRCs) is coupled with variations in the expression of genes involved in transport, cell junctions and metabolism in the branchial epithelium of fish under hypertonic conditions [19–25]. Ion transporters and channel proteins, such as Na⁺-K⁺ ATPase (NKA) and Na⁺/H⁺ exchanger (NHE), as well as cell junction proteins for paracellular transport, such as the claudin, aquaporin and cadherin gene families, exhibit salinity-dependent expression patterns and facilitate the tolerance of fish to alterations in salinity-alkalinity [23–25]. To maintain the intracellular pH, carbonic anhydrases (CAs) and HCO₃⁻ transporter genes, such as solute carrier family 4 (SLC4a) and solute carrier family 9 (SLC9a), are differentially expressed in the gills and kidneys of fish during their transfer from freshwater to saline-alkaline condition [10, 26–31].

Considering the high-energy requirements for osmoregulation in teleosts, saline-alkaline shifts impact the metabolism of carbohydrates, lipids, proteins and amino acids [27, 32]. Salinity-alkalinity changes lead to variations in the expression of genes in gills, which are involved mainly in the pathways of glycolysis/gluconeogenesis, amino acid metabolism and steroid biosynthesis, and differentially expressed genes in the liver are enriched in lipid metabolism, protein processing, and N-glycan biosynthesis, suggesting tissue-specific responses to salinity-alkalinity changes [32]. Moreover, saline-alkaline-adapted fish have been reported to synthesize less toxic urea and amino acids to avoid ammonia toxicity under high salinity-alkalinity conditions [33, 34]. These energy-rich compounds, such as amino acids and amino sugars, serve not only as energy resources but also as important organic osmolytes to maintain cell volume and osmolality [35]. Therefore, alterations in salinity-alkalinity may reshape the metabolic signatures for osmoregulation and energy supplies in fish.

Gymnocypris przewalskii (subfamily Schizothoracinae, genus Gymnocypris) is an exclusive cyprinid fish that dwells in Soda Lake Qinghai (salinity 12–14 ppt, pH 9.0–9.2, altitude ~ 3200 m) on the Qinghai Tibetan Plateau (QTP) [36, 37]. It migrates to freshwater rivers (salinity ~ 0.02 ppt, pH 8.3–8.6) for spawning annually, demonstrating the ability to cope with a wide range of salinity-alkalinity [38]. Several studies have characterized the genomic and transcriptomic bases of the adaptation of G. przewalskii to high salinity and alkalinity, however, the molecular regulation underlying its tolerance to salinity-alkalinity variation remains elusive [39–41].

With advances in high-throughput technologies, multiomics analyses have been adopted to reveal the molecular mechanism involved in the adaptation of fish to environmental variations [42, 43]. In the present study, we conducted morphological, biochemical, transcriptomic and proteomic analyses of the gills and kidneys of G. przewalskii under freshwater (FW) and saline-alkaline lake water (LW), to identify the key genes and regulatory signatures involved in the adaptation to salinity-alkalinity alterations. Our results illustrate the regulatory basis of the adaptation of G. przewalskii to salinity-alkalinity fluctuations and shed light on the molecular mechanism of euryhalinity in fish.

Materials and methods

Experimental treatment and sample collection

All the animal experiments were conducted according to the guidelines described in the “Guidelines for Animal Care and Use” manual (NWIPB-202206) approved by the Animal Care and Use Committee, Northwest Institute of Plateau Biology, Chinese Academy of Sciences.

For experimental treatment, G. przewalskii was artificially reproduced from a wild population and reared under freshwater conditions in the rescue center of G. przewalskii. In total, 28 fish were randomly selected and assigned to two groups: the freshwater group (FW, body weight = 14.8 ± 3.21 g, n = 14) and the saline-alkaline lake water group (LW, body weight = 15.6 ± 3.57 g, n = 14), which were raised in two plastic pools filled with 1500 L of water. The FW group was reared in circulating water throughout the experiment. The lake water was obtained from the nearshore region of Lake Qinghai, where the salinity was 12 ppt and the pH was 9.0 and was transported to the experimental facility. In the LW group, the fish were initially freshwater and then transferred to 1:4, 1:2 and 1:1 dilutions of lake water by adding lake water according to volume. The fish were maintained for 48 h at each level before moving to the next level according to a previous report [21]. Finally, the fish in the LW group were reared in lake water with a final salinity of 12 ppt and a pH of 9.0 for two weeks. During the experiment, both groups were fed twice a day (9:00 and 16:00) with a commercial diet containing 43% protein (Hanye Biotechnology Co., Ltd, China) [44]. The dissolved oxygen (DO) was supplied and monitored during the experiment using YSI ProPlus (YSI Inc., USA), with values of 7.58 ± 0.07 mg/L and 7.41 ± 0.12 mg/L for the FW and LW groups, respectively. The water temperature was maintained at 10–12 °C using cooling system (HAILI, China), and half of the water was replaced each day. No fish died during the experiment.

After treatment, G. przewalskii were euthanized with 200 mg/L MS222 (Sigma, USA) to collect kidney and gill samples. For each treatment, blood samples were collected from the caudal vein of all individuals via a syringe with 1.5 mg/mL EDTA anticoagulant, after which the serum was separated, and the ammonia level was measured. Four gills and three kidneys from each treatment were collected and immediately stored in liquid nitrogen for RNA and protein purification. Four gill samples (the first gill arch from the right side) and kidney (right side) samples from each treatment group were collected for histological examination and transmission electron microscopy (TEM). Gills and kidneys from 10 samples were collected for the examination of urea and amino acid contents.

Transmission electron microscopy (TEM)

The tissue samples for transmission electron microscopy (TEM) were prepared as previously described [21]. Fresh kidneys and gills were fixed in 2.5% glutaraldehyde fixative (Servicebio Inc., China) for 2–4 h and then washed with PBS. The sample was dehydrated in a series of ethanol solutions, embedded in SPURR resin blocks and polymerized at 37 °C overnight. The sample was placed in a 60 °C oven for 48 h and then on a copper grid and subjected to uranium‒lead double staining (2% uranium acetate saturated alcohol solution, lead citrate, each for 15 min). After drying overnight at room temperature, the sample slices were observed under a transmission electron microscope (Hitachi HT7700).

Biochemical examination

A blood sample of 1 ml was centrifuged at 3000 × g for 15 min at 4 °C to obtain serum and plasma for the measurement of ammonia. The level of ammonia was examined using biochemical analyzer (Servicebio, China). The plasma was mixed with 5% trichloroacetic acid and homogenized. The kidney and gill tissues were homogenized, and the pH was adjusted to 1.80–2.00 with LiOH, which was used to determine the amino acid content of each sample. The contents of 19 free amino acids, including aspartate (Asp), asparagine (Asn), glutamine (Gln), glutamate (Glu), threonine (THR), serine (Ser), glycine (Gly), alanine (Ala), valine (Val), cystine (Cys), methionine (Met), isoleucine (Ile), leucine (Leu), tyrosine (Tyr), phenylalanine (Phe), histidine (His), lysine (Lys), arginine (Arg), proline (Pro), and histidine (His), were determined in the amino acid analyzer (S-433D amino acid analyzer, SYKAM, Germany) with standards (SYKAM, Germany).

Transcriptomic sequencing and data analysis

RNA-seq library construction

Four gill samples and three kidney samples from each of the FW and LW groups were used for RNA purification and library construction. The RNA samples were extracted using MiniBEST Universal RNA Extraction Kit (TaKaRa, China), and the quality and quantity of the RNA samples were assessed using 1% gel electrophoresis, a Qubit® 2.0 fluorometer (Life Technologies, USA), and Agilent Bioanalyzer 2100 system (Agilent Technologies, USA). The RNA-seq library was constructed according to the manufacturer’s instructions. Briefly, mRNA was captured from 1.5 μg of RNA, which was used for the synthesis of both first- and second-strand cDNA. After fragmentation, purification and PCR amplification, the final cDNA library was obtained, which was sequenced on the Illumina NovaSeq 6000 platform (Novogene Inc., China).

Identification of differentially expressed genes

Clean data were obtained by removing adapters, poly-N-containing reads (N% > 10%) and low-quality reads containing more than 50% low-quality bases (Q value ≤ 20) from the raw reads via fastp (version 0.18.0) [45] and then mapped to the G. przewalskii genome using HISAT (v2.2.4) [46]. Gene transcription levels were calculated for each sample using Cufflinks (v2.2.1) [47] to obtain FPKM values. Differential expression analysis was performed using the DESeq2 package [48] implemented in R. Genes with p values less than 0.05 and fold changes greater than 1.5 were defined as differentially expressed genes (DEGs). Based on DEGs, gene ontology (GO) enrichment was conducted using clusterProfiler in R [49], and significantly enriched pathways were identified according to p value less than 0.05.

Reverse transcription quantitative PCR (RT‒qPCR)

We performed RT‒qPCR experiments to validate the identified DEGs in RNA-seq analysis. The cDNA from four tissue samples in each group was synthesized using PrimeScrip™ RT reagent Kit with gDNA Eraser (TaKaRa, China), with the RT negative control using PCR water instead of the RNA sample. The PCR primers used are listed in Supplementary Table 1. PCR was performed with a cDNA template (1:5 dilution), Aptamer™ qPCR SYBR Green Master Mixture (Novogene, China), and forward and reverse primers (Table S1) in a final volume of 20 μl. The PCR conditions were as follows: 95 °C for 5 min; 40 cycles at 95 °C for 5 s and an annealing temperature of 55–60 °C for 30 s; 95 °C for 10 s; 60 °C for 30 s; and 95 °C for 15 s. We included two types of controls, the RT control and the PCR control, and no amplicon was detected. The housekeeping gene β-actin was used as the internal control to normalize the data [21, 50], and gene expression was measured using ^−ΔΔCt method [51]. Fold changes (FCs) were calculated between FW and LW, and Pearson’s correlation was used to evaluate the consistency between RNA-seq and RT‒qPCR.

Proteomics

Sample preparation

The same samples used for RNA-seq were used for tandem mass tag (TMT)-based protein proteomics. The tissue samples were homogenized in liquid nitrogen with lysis buffer and then ultrasonicated in ice water for 5 min. The samples were subsequently centrifuged at 12,000 × g at 4 °C for 15 min, after which the supernatants were collected and mixed with 10 mM DTT at 56 °C for 1 h. Iodoacetamide was added, and the mixture was incubated for 1 h at room temperature in the dark. The sample was then completely mixed with 4 volumes of acetone and incubated at -20 °C for 2 h. The precipitate was obtained by centrifugation at 12,000 × g at 4 °C for 15 min, after which it was quantified via a BSA assay (Bradford protein assay kit, Sangon Biotech) according to the manufacturer’s instructions.

Identification of differentially expressed proteins

Briefly, ~ 120 μg of protein was digested with trypsin at 37 °C overnight and then desalted using a C18 desalting column. The TMT labeling reagent was mixed with the washed supernatant and incubated at room temperature for 2 h. The sample was dissolved in mobile phase A solution (2% acetonitrile, pH 10.0) and fractionated using a C18 column (Waters BEH C18 4.6 × 250 mm, 5 μm) on a Rigol L3000 HPLC system. The separated peptides were analyzed via an EASY-nLCTM 1200 UHPLC system (Thermo Fisher) coupled with a Q Exactive HF-X mass spectrometer (Thermo Fisher), which generated the spectra for each fraction. The raw data were imported into Proteome Discoverer (v2.2) for peptide and protein identification and quantification. Two steps were applied to obtain high-quality proteomic results. First, peptide spectrum matches (PSMs) with confidence greater than 99% and proteins containing at least one unique peptide were maintained. Second, a false discovery rate (FDR) test was employed to remove peptides and proteins with FDRs greater than 1%, and the remaining peptides and proteins with high confidence were used for quantification and functional annotation. The t-test was conducted, and differentially abundant proteins (DAPs) were identified based on p values less than 0.05 and fold changes (FCs) greater than 1.2 [52].

Integrative analysis of the transcriptome and proteome

The correlation coefficients between the transcriptome and proteome were calculated based on the FC of the expressed transcripts and proteins between the FW and LW groups using cor package in R. The protein–protein interaction (PPI) network was constructed with genes whose transcription and protein expression were positively correlated. The genes were retrieved from the STRING database (http://string-db.org/), and the protein‒protein interactions were analyzed via Metascape and visualized via Cytoscape [53].

Immunohistochemical examination

The tissue samples were washed with PBS twice and dehydrated in a series of 30–100% ethanol. Each sample was treated with xylene two times, embedded in paraffin, and sliced into 5 μm sections. After dewaxing, endogenous peroxidase activity was blocked with serum. The sections were incubated with a primary antibody (1:500 dilution) against anti-Junctional Adhesion Molecule 1/JAM-A rabbit pAb (GB111265, Servicebio, China) and a secondary antibody (1:200 dilution) against Cy3-conjugated goat anti-rabbit IgG (GB21303, Servicebio, China), and the nuclei were stained with DAPI blue (Servicebio, China). Finally, the sections were observed and photographed under a fluorescence microscope (Zeiss, AxioImage A2).

Statistical analysis

The t-test was conducted to detect significant differences in biochemical parameters between the FW and LW groups. The significant difference was defined as p value less than 0.05, and the extremely significant difference was defined as p value less than 0.01. Before the t-test, we performed the Shapiro–Wilk normality test and F test to examine the normality and variance homogeneity of the data. The Pearson correlation coefficient (r²) was calculated between the FC of RT-qPCR and that of RNA-seq via GraphPad Systems (Prism 9).

Results

Morphological variations in the gills and kidneys of G. przewalskii

Mitochondrial-rich cells (MRCs) and red blood cells (RBCs) appeared in the lamella of the gill filament under FW conditions (Fig. 1a-b), and pavement cells (PVCs), pillar cells (PCs) and mucus cells (MCs) with mucus inside existed in the lamella of G. przewalskii in the LW group (Fig. 1c-d). In the kidney, TEM showed the variations in the junction structure in the renal tubules, with a dense form in the LW condition and a slightly leaky form in the FW condition (Fig. 2).

Fig. 1.

Transmission electron microscopy of G. przewalskii gills. a and (b) Microstructure of gill lamella under freshwater condition (FW). c and (d) Microstructure of gill lamella under lake water (LW). The scales represent 20 µm in (a) and (c) and 5 µm in (b) and (d). MRCs: mitochondria-rich cells. RBC: Red blood cells. MCs: mucus cells. PC: Pillar cells. PVC: Pavement cells

Fig. 2.

Transmission electron microscopy of G. przewalskii kidney. a and (b) Microstructure of renal tubules under freshwater condition (FW). c and (d) Microstructure of renal tubules in high-salinity and alkaline lake water (LW) of high salinity and alkalinity condition. The scales represent 5 µm in (a) and (c) and 2 µm in (b) and (d). The junction structure is labeled with white squares

Biochemical responses to saline-alkaline variation

The content of ammonia in the blood was significantly greater in the LW group than in the FW group (p < 0.05) (Fig. 3a). The urea level was significantly lower in the kidney of the LW group (p < 0.01) and did not significantly change in the gills (Fig. 3b). The total free amino acid content significantly decreased in both the gills and kidneys of the LW group (Fig. 3c). The contents of essential amino acids significantly decreased in the gills, the levels of Val, Ile, Leu, His and Lys significantly decreased, and Phe and Arg increased in the kidney (Fig. 3d-f). Among the nonessential amino acids (NEAAs), Asn, Glu and Gln were more abundant in both the gills and kidneys of the LW group than in those of the FW group (Fig. 3e-g). In addition, the remaining NEAAs were lower in the LW group than in the FW group.

Fig. 3.

Biochemical indices of G. przewalskii. a Ammonia content in the blood. b Urea content in the gills and kidneys. c The content of total free amino acids in the gills and kidneys. d The contents of essential amino acids in the gills. e The contents of nonessential amino acids in the gills. f The contents of essential amino acids in the kidneys. g The contents of nonessential amino acids in the kidneys. The blue bars indicate freshwater condition, and the orange bars indicate saline-alkaline condition of the lake water. FW, freshwater. LW, lake water. TFAAs, total free amino acids; EAAs, essential amino acids; NEAAs, nonessential amino acids. “*” indicates a significant difference with p value less than 0.05, and “**” represents an extremely significant difference with p value less than 0.01

Transcriptomic variations under different saline‒alkaline conditions

Overviews of the RNA-seq results

The transcriptomic patterns of the gills and kidneys of G. przewalskii were compared between freshwater (FW, 0.22 ppt, pH 7.8) and high saline-alkaline lake water (LW, 12 ppt, pH 9.0). We generated 54,601,180 to 76,292,610 clean reads for each sample, with Q30% greater than 92% (Table S2). The clean reads were aligned to the G. przewalskii genome, and the mapping ratios ranged from 76.54% to 83.51%. These data indicate the high quality of the RNA-seq data, which ensures the identification of differentially expressed genes (DEGs).

Identification of differentially expressed genes

We identified 1,424 differentially expressed genes (DEGs) in the gills of G. przewalskii, including 831 upregulated genes and 593 downregulated genes, under LW condition (Table 1 and Table S3). The transcriptomic analysis of the kidney revealed that 1,554 genes were significantly differentially expressed, including 726 upregulated genes and 828 downregulated genes, under LW treatment (Table 1 and Table S4). There were 240 common DEGs in both tissues, and their expression clearly separated the FW and LW groups into two tissues, suggesting that the transcriptional variations in the kidneys and gills were induced by salinity-alkalinity changes (Fig. 4a-b). We selected 10 genes to verify the RNA-seq results via RT‒qPCR. In terms of the fold changes, the Pearson correlation coefficient (r²) between RT‒qPCR and RNA‒seq was 0.84, suggesting the accuracy and reliability of the RNA‒seq results (Fig. 4c).

Table 1.

Differentially expressed genes and differentially abundant proteins

Tissue		Total	Up-regulation	Down-regulation
Gills	Transcripts	1,424	831	593
	Proteins	155	55	100
Kidney	Transcripts	1,554	726	828
	Proteins	200	92	108

Up- and down-regulation: the transcription expression and protein abundance in lake water condition versus those in freshwater

Fig. 4.

Differentially expressed genes in the gills and kidneys of G. przewalskii under freshwater and lake water conditions. a Venn diagram of DEGs between the gills and kidneys. b Hierarchical cluster analysis of DEGs in the gills and kidney. c Validation of the RNA-seq results via RT‒qPCR. DEG: differentially expressed gene. G, gills. K, kidney. FW, freshwater. LW, lake water. R², Pearson correlation coefficient

Functional enrichment analyses of DEGs

To understand the potential functional impact, we performed GO and KEGG enrichment analyses on the DEGs. GO analysis suggested that DEGs in the gills and kidney were overrepresented in functions related to ion transport, acid–base regulation, and immunity (Table S5 and Table S6). Moreover, genes related to ion transport and acid–base regulation, such as GO terms related to the regulation of pH (GO: 0006885), chloride transport (GO: 0006821) and antiporter activity (GO: 0015297), were downregulated in the LW group (Fig. 5a-b). The KEGG analysis revealed that 20 and 16 pathways were enriched with DEGs in the gills and kidney, which included 9 and 6 metabolic pathways, such as pyruvate metabolism (ko00620), glycolysis/gluconeogenesis (ko00010) and arginine and proline metabolism (ko00330) (Fig. 5c-d and Table S7-8).

Fig. 5.

Functional enrichment of differentially expressed genes (DEGs) in the gills and kidneys of G. przewalskii under freshwater and lake water conditions. GO enrichment of DEGs in the gills (a) and kidneys (b) of G. przewalskii. KEGG enrichment of DEGs in the (c) gills and (d) kidneys of G. przewalskii. For GO enrichment, yellow, pink and purple represent biological process (BP), molecular function (MF) and cellular component (CC), respectively. For KEGG analysis, yellow, pink, purple, blue and green represent metabolism, genetic information processing, environmental information processing, organismal systems and human diseases, respectively. The annotation of circles (from the outer to the inner side): the 1st circle indicates the GO term or KEGG pathway ID, and the scales denote the number of genes. The 2nd circle indicates the number of genes in the GO terms or KEGG pathway, and the color indicates the range of significant levels, with darker colors indicating higher significance levels. In the 3rd circle, blue and purple bars indicate the proportion of up- and downregulated genes, and the number of up- and downregulated genes is labeled below. The 4.^th circle indicates the rich factor (number of DEGs/number of genes in the GO terms or KEGG pathway)

Protein expression variations under different saline‒alkaline conditions

TMT-based proteomics generated 416,785 and 433,744 spectra in the gills and kidney, respectively, including 108,375 and 107,239 matched spectra, which resulted in the identification of 53,607 and 48,675 peptides. Overall, 8,364 and 8,306 proteins were identified in the gills and kidneys of G. przewalskii, among which 8,300 were expressed in all samples from both tissues.

In the gills, we identified 155 differentially abundant proteins (DAPs), including 55 activated and 100 repressed DAPs, under LW condition (Table 1 and Table S9). These DAPs of gills were enriched in the GO terms of tolerance induction (GO: 0002507), ATP hydrolysis coupled anion transmembrane transport (GO: 0099133) and response to L-glutamate (GO: 1,902,065) (Fig. 6a and Table S10). In the kidney, we observed the upregulation of 92 proteins and the downregulation of 108 proteins in the LW condition (Table 1 and Table S11). DAPs of the kidney were overrepresented in the GO terms of regulation of pH (GO:0006885), proton transmembrane transport (GO:1,902,600) and cellular response to osmotic stress (GO:0071470) (Fig. 6b and Table S12). These results indicate that both transport and metabolism are involved in the response of G. przewalskii to salinity-alkalinity changes. The Venn diagram revealed 8 common DAPs in both tissues, which clearly differed in expression between the two treatments (Fig. 6c-d). Among these common DAPs, the abundance of junctional adhesion molecule A (JAM-A, GPPG00046776) was significantly lower in the FW group than in the LW group in both tissues, which was consistent with the immunostaining results (Fig. 7).

Fig. 6.

Differentially abundant proteins in the gills and kidneys of G. przewalskii in freshwater and lake water. a GO enrichment of DAPs in gills. b GO enrichment of DAPs in the kidney. c Venn diagram of DAPs between the gills and kidney. d Hierarchical cluster analysis of DAPs in the gills and kidney. DAP: differentially abundant protein. G: gills. K: kidney. FW: freshwater. LW: lake water

Fig. 7.

Immunofluorescence of JAM-A in the gills and kidneys of G. przewalskii in freshwater and lake water. JAM-A: junctional adhesion molecule A. Red indicates JAM-A staining. The blue signals of DAPI indicate DNA staining. G: gills. K: kidney. FW: freshwater. LW: lake water. G, gill. K, kidney

Integrative proteomic and transcriptomic analyses

In total, 201 and 370 genes in the gills and kidney were positively correlated at the transcription and protein expression levels, with Pearson correlation coefficients (r²) of 0.1305 and 0.1511 for the gills and kidney, respectively (Fig. 8a-b). The positively correlated genes were enriched in GO functions related to the increased expression of GLUL (glutamate-ammonia ligase, GPPG00030516), cell adhesion, transport activities and metabolism, suggesting that they were influenced by variations in salinity-alkalinity through the gene-protein network in the gills and kidneys of G. przewalskii (Fig. 8c-d).

Fig. 8.

Integrative analysis of the transcriptome and proteome of the gills and kidneys of G. przewalskii. Nine-quadrant map of transcriptomic and proteomic correlations in the gills (a) and kidney (b). GO enrichment of positively correlated genes in the gills (c) and kidney (d). The size of the circles indicates the number of genes associated with the GO terms. The lines among circles indicate the interactions among genes obtained from the STRING database

Molecular network for osmoregulation in G. przewalskii

Among these positively correlated genes, genes related to acid‒-base homeostasis and transport in the gills and kidney under freshwater condition, including CA2 (carbonic anhydrase 2, GPPG00087498), GLS (glutaminase kidney isoform, GPPG00014160), SLC9a3 (Na⁺/H⁺ exchanger 3, NHE3, GPPG00000195 and GPPG00001973), SLC4a2 (Na⁺/HCO₃⁻ exchanger, AE2, GPPG00004442), SLC4a4 (Na⁺/HCO₃⁻ cotransporter, NBCe1, GPPG00045161), Na⁺-K⁺ ATPase subunits ATP1a1 (GPPG00073989) and ATP1b1 (GPPG00045014), were highly expressed. In addition, the expression of genes in the renin-angiotensin system, including PRCP (prolycarboxypeptidase, GPPG00017413), ENPEP (glutamyl aminopeptidase, GPPG00024682 and GPPG00014783) and REN (renin, GPPG00006133), was increased in the FW group, which was related to the reabsorption of Na⁺ in freshwater condition (Fig. 9a-b). Under LW condition, genes involved in the regulation of cell adhesion, including GEF-H1 (Rho/Rac guanine nucleotide exchange factor 2, GPPG00000179), cortactin (GPPG00080785), CDH5 (cadherin-5, GPPG00065431) and SYNPO (synaptopodin, GPPG00004218), were highly expressed in the gills and kidney (Fig. 9a-b). Increased expression of GLUL (glutamate-ammonia ligase, GPPG00030516) was observed in the LW group, which might play a role in reducing NH3 levels via the synthesis of glutamine.

Fig. 9.

Molecular network for osmoregulation in G. przewalskii under LW and FW conditions. a Simplified models of genes involved in ion transport, acid–base regulation, ammonia excretion and cell junctions. b Downregulation of genes involved in the biosynthesis of N-acetylneuramic acid. c Heatmaps of genes with positive correlations between the transcriptome and proteome. The gene symbol in red indicates upregulation in the FW. The gene symbol in blue indicates downregulation in FW, LW, and lake water. FW, freshwater. G, gill. K, kidney

Additionally, genes involved in metabolism presented varied expression levels in the kidneys and gills of G. przewalskii in the comparison of the FW and LW groups. For example, MDH1 (GPPG00008265, malate dehydrogenase) in the TCA cycle and CKB (creatine kinase B, GPPG00024448, GPPG00034640) and CKM (creatine kinase M, GPPG00030543) for energy homeostasis presented higher expression under FW condition. The higher expression of CPB1 (carboxypeptidase B, GPPG00062078), SLC1A1 (GPPG00032761) and SLC7A8 (GPPG00056567) indicated the potentially elevated function of protein digestion and absorption, which could explain the increased contents of amino acids under FW conditions. In the LW groups, genes involved in lipid and fatty acid metabolism, including Apoeb (apolipoprotein Eb, GPPG00079040 and GPPG00001858), NEU3 (GPPG00029676, sialidase-3), SMPD4 (GPPG00012835, sphingomyelin phosphodiesterase 4), CERS1 (ceramide synthase, GPPG00007981) and ALDH3A2 (fatty aldehyde dehydrogenase, GPPG00011297), presented relatively high expression under LW condition (Fig. 9b-c). We also found that genes for N-acetylneuraminic acid biosynthesis, including GFPT2 (glutamine: fructose-6-phosphate aminotransferase, GPPG00034273), GPI (glucose phosphate isomerase, GPPG00045054), and GNE (UDP-N-acetylglucosamine 2-epimerase/N-acetylmannosamine kinase, GPPG00007434) in gills and HK2 (hexokinase 2, GPPG00069499) and NANP (N-acylneuraminate-9-phosphatase, GPPG00038766) in the kidney, were upregulated under LW condition in G. przewalskii (Fig. 9b-c).

Discussions

G. przewalskii dwells in Lake Qinghai with high saline-alkaline conditions and migrates to freshwater during annual spawning, which demonstrated the strong capacity for osmoregulation [41, 54]. The current study combined morphological, biochemical, transcriptomic, and proteomic analyses to explore regulatory pathways and identify the crucial molecules underlying the adaptation of G. przewalskii to salinity‒alkalinity changes. In the present study, the correlation coefficients between transcriptomics and proteomics were 0.13 and 0.15 in the gills and kidneys, respectively, which are similar to observations in other fish [55]. The discrepancy between the transcriptome and proteome could be explained by posttranscriptional regulation, translation efficiency, and degradation rates of proteins and mRNAs [56, 57]. Moreover, miRNA‒mRNA interactions have been observed in the gill-derived cell line of G. przewalskii under salinity stress, suggesting the influence of posttranscriptional regulation on the correlation between transcriptomic and proteomic data [58].

Activation of ion transport and acid‒base regulation

During the transition from high to low salinity, the ability of fish to maintain ion homeostasis is highly challenging. Gill remodeling has been observed in the response of fish to saline-alkaline changes, particularly with the appearance of MRCs under hypotonic condition [21, 22, 33]. MRCs are mainly in charge of ion and solute transport, where transporters and exchangers, such as Na⁺-K⁺ ATPase (NKA), Na⁺-H⁺ exchangers (NHEs) and Na⁺-HCO₃⁻ cotransporters (NBCs), are primarily expressed for transepithelial transport [12, 22]. ATP1a1 and ATP1b1, consisting of the NKA subunit, are expressed in a salinity-dependent manner, and SLC9a3 (NHE3) has been recognized as an important regulator of Na⁺/H⁺ exchange [23]. In the present study, we detected MRCs in the gills of G. przewalskii under FW condition, which was coincident with the upregulation of DEGs and DAPs involved in ion transport. The increased expression of SLC9a3, ATP1a1 and ATP1b1 suggested the activation of ion uptake in G. przewalskii under FW condition (Fig. 9c), which is consistent with the expression patterns of these genes during the transition of several fish to soft, low-salinity water, such as rainbow trout [59], chum salmon [23], spotted sea bass [60], and marine medaka [11]. It has been reported that the blood osmolality of G. przewalskii is close to that in Lake Qinghai, which was reduced in the migration to the river [61–63].Therefore, the activation of ion absorption genes would contribute to maintain osmotic homeostasis in G. przewalskii under freshwater condition. Similarly, differential expression of genes involved in ion transport has been identified by comparing freshwater with high saline-alkaline water populations of Leuciscus waleckii, the cyprinid fish dwelling in the soda lake, Lake Dali Nur [6]. Interestingly, genes of the SLC9a and SLC2a gene families and NKA subunits showed higher expression in L. waleckii in high saline-alkaline lake water [6, 10]. Moreover, it has been reported that genes involved in ion uptake, including SLC9a3, ATPa1a and ATPa1b, are more highly expressed in freshwater than in seawater in fish acclimated to tidally changing salinities [64]. These phenomena suggest the involvement of ion absorption in maintaining the osmotic homeostasis of fish experiencing saline-alkaline fluctuations.

In addition to ion exchangers and channels, our results suggest the involvement of the renin‒angiotensin system (RAS) in the prevention of water and ion loss in G. przewalskii in FW. The RAS is responsible for the regulation of sodium and water homeostasis. PRCP, ENPEP and REN encode lysosomal pro-X prolylcarboxypeptidase, glutamyl aminopeptidase and renin, respectively, and are critical in the RAS. Renin is the rate-limiting enzyme in the RAS and is synthesized mainly in the kidney. Renin-null mice and zebrafish showed defects in renal development, and required extra saline injection for survival, indicating that renin is critical for ion reabsorption in the kidney [65, 66]. Renin initiates the RAS cascade to generate Ang II via the process of angiotensinogen by ACE and aminopeptidase, including ENPEP. PRCP converts Ang II to Ang (1–7), which could trigger the expression of downstream genes, such as ATPase and NKA, to strengthen the water and sodium reabsorption in proximal tubules [67]. The increased expression of REN, ENPEP and PRCP suggested that the activation of the RAS, together with increased levels of ion transepithelial transporter genes, might contribute to maintaining ion and water homeostasis in G. przewalskii under FW.

Acid‒base regulation

The high bicarbonate alkalinity of Lake Qinghai requires G. przewalskii to develop an adaptive response to maintain the acid‒base balance during migration from the soda lake to the freshwater river [54]. The genes associated with proximal tubule bicarbonate reclamation were activated in the gills and kidneys under saline-alkaline condition, which is consistent with observations in rainbow trout [59] and Marine Medaka [11]. Carbonic anhydrases (CAs) catalyze the interconversion of carbon dioxide (CO₂) and bicarbonate (HCO₃⁻) and regulate acid–base homeostasis by secreting H⁺ and absorbing HCO₃⁻ in couple with SLC9a3 (NHE3) and SLC4a4 (NBCe1) [10, 26]. The HCO₃⁻ transporters SLC4a2 (AE2) and SLC4a4 (NBCe1) and the Na⁺/H⁺ exchanger 3 SLC9a3 (NHE3) play crucial roles in the maintenance of intracellular ion and pH homeostasis [68, 69]. Physiological assays have demonstrated that G. przewalskii experienced transient metabolic acidosis during its migration from the lake to freshwater rivers [62]. The activation of CA2, SLC9a3, SLC4a2 and SLC4a4 would facilitate the reabsorption of HCO₃⁻ and Cl⁻ as well as the excretion of H⁺, which contributes to acid–base regulation and ion homeostasis in G. przewalskii in response to saline-alkaline changes.

Nitrogen waste excretion

Ammonia (NH₃) is a major form of nitrogenous waste in fish and is secreted into water through gill filaments through transporters [14, 70]. Rh glycoproteins, including Rhag, Rhbg and Rhcg, mediate ammonia excretion in fish [71, 72]. The greater transcription of Rhag (GPPG00059145) and Rhcg (GPPG00004431) in the gills of G. przewalskii under FW was consistent with the observation of the upregulation of Rhag glycoproteins under low pH, which is responsible for ${\text{NH}}_4^{+}$ exclusion through the gills of freshwater teleost fish [73, 74]. Additionally, we detected the activation of GLS in G. przewalskii, which could also promote ${\text{NH}}_4^{+}$ and energy production in FW by catalyzing the conversion of glutamine to glutamate [75]. Therefore, G. przewalskii eliminated ${\text{NH}}_4^{+}$ directly in gills, and the upregulation of GLS could increase to meet the energy demands of osmoregulation in G. przewalskii under FW.

On the other hand, high alkalinity impedes the conversion of NH₃ to ${\text{NH}}_4^{+}$ by reducing H⁺, which leads to its accumulation and toxic effects on fish [76]. Therefore, saline-alkaline-adapted fish have developed possible mechanisms to convert NH₃ to less toxic forms of nitrogenous products, such as amino acids and urea [77]. In line with previous reports, urea did significantly differ between the LW and FW groups of G. przewalskii [61]. Although the contents of most amino acids decreased, the levels of glutamine and glutamate increased with increasing salinity-alkalinity, indicating their importance in the adaptation of G. przewalskii to salinity-alkalinity. GLUL belongs to the glutamine synthetase family and catalyzes the conversion of glutamate and NH₃ to glutamine [77]. The activation of GLUL suggested that G. przewalskii reduced the toxicity of NH₃ through the generation of glutamine in LW, which was supported by the high levels of glutamate and glutamine. Therefore, G. przewalskii has adopted different methods for NH₃ detoxification, including the excretion of ${\text{NH}}_4^{+}$ in FW and the synthesis of glutamine in LW.

Transcellular and paracellular permeability

To avoid the passive invasion of ions and water loss, fish tend to adjust their paracellular permeability through the expression of genes in cell‒cell junctions [4, 78]. JAM-A is specifically localized at tight junctions and controls junction integrity and paracellular permeability [79]. Compared with the “leaky” junction in the FW, the upregulation of JAM-A suggested strengthened tight junctions in the kidney, which might explain the appearance of dense and intact barriers in the renal tubules. These morphological and expression alterations contribute to the protection of G. przewalskii from the loss of solutes in high-salinity environments. SYNPO and GEF-H1 reportedly localize at tight junctions and are critical for junction integrity and protective effects [80, 81]. It has been reported that CDH5 (cadherin 5, VE-cadherin) and cortactin form a complex and play a role in adherens junction assembly and stability [82]. The upregulation of SYNPO and GEF-H1 in the kidney and the increased expression of CDH5 and cortactin suggested the involvement of diverse cell junctions in the response of G. przewalskii to increased salinity and alkalinity, which was consistent with observations in many euryhaline fishes [59, 78, 83]. Moreover, MCs with mucus appeared in G. przewalskii gills in the LW. Mucus acts as a “waterproof” barrier and protects cells from hypertonic stress in fish [84–86]. N-acetylneuramic acid, a member of the sialic acid family, has been identified as a component of gill mucus [87]. Coincidently, the activation of a series of genes in the N-acetylneuramic acid biosynthesis pathway might be related to the induction of key components of mucus, which would limit the transcellular transport of water and ions through G. przewalskii gills under LW conditions [85]. Therefore, increased cell junctions and mucus production might block the paracellular and transcellular transport of solutes and contribute to the adaptation of G. przewalskii to high salinity-alkalinity.

The metabolic responses

Our analysis revealed that salinity-alkalinity variation resulted in transcriptomic and proteomic changes in genes involved in lipid, protein, amino acid and purine metabolism in G. przewalskii. MDH1 is a metabolic enzyme that fuels the TCA cycle [88, 89]. CK consists of subunits B and M and plays a central role in cellular energy metabolism [90]. The activation of MDH1, PC and CKs might promote the TCA cycle and energy homeostasis to meet the energy demand for ion uptake in G. przewalskii in FW. Genes involved in protein digestion and absorption, such as CPB1 [91], SLC1A1 and SLC7A8 [92], were upregulated, which is consistent with the increased contents of amino acids in the FW group. Amino acids are considered energy supplies for osmoregulation and organic osmolytes to maintain cell volume, and their contents are altered in fish tissues under different salinities [35]. Therefore, the increased amino acid contents might provide energy for ion uptake for osmoregulation under freshwater condition. Our results suggested that genes involved in sphingolipid metabolism, including NEU3 [93], CERS1 [94], and SMPD4 [95], were highly expressed under saline-alkaline condition. Sphingolipids, such as sphingomyelin and its intermediate product ceramide, are important components of lipid rafts in the cell membrane and are involved in the regulation of ion channels in tilapia and milkfish [78, 96]. The expression profiles indicated that the synthesis of sphingolipids might contribute to the high salinity‒alkalinity adaptation of G. przewalskii through potential changes in membrane structure.

Conclusions

By applying histological and biochemical analyses, transcriptomics and proteomics, the current study provides new insights into the physiological and molecular regulatory networks in the gills and kidneys of G. przewalskii to adapt to a wide range of salinities and alkalinities.

Under freshwater condition, G. przewalskii increased the expression of genes involved in ion absorption and acid–base regulation to maintain osmolality and the intracellular pH. Under high salinity-alkalinity lake water condition, G. przewalskii presumably produces glutamine for NH₃ detoxification via the upregulation of GLUL and increases the expression of genes in cell junctions and mucus to regulate transcellular and paracellular permeability. Additionally, the increased expression of genes involved in sphingolipid metabolism might lead to potential changes in membrane structure and components, and its role in the adaptation of G. przewalskii to high salinity-alkalinity needs further investigation. In conclusion, the present study investigated the morphological, biochemical and molecular responses of G. przewalskii to alterations in salinity-alkalinity, which provides helpful information for understanding the adaptative mechanism of the response of fish to salinity-alkalinity changes at the molecular level.

Abbreviations

DEG

Differentially expressed gene

DAP

Differentially abundant protein;

iTRAQ

Isobaric tags for relative and absolute quantitation

FC

Fold change

GO

Gene Ontology

RT‒qPCR

Real-time quantitative PCR

LW

Lake water

FW

Fresh water

K

Kidney

G

Gills

Authors’ contributions

Conceptualization: FT; Sampling: SF, YL, HQ; Investigation: BZ, RS, LY, XL; Bioinformatic analysis: RS, SL, and DQ; Writing-original draft: FT, BZ, RS, SL; Writing-review & editing: FT, BZ, SL, KZ, DQ, XL, HQ, SF, YL, LY; Supervision: FT and SL; Project administration: FT and KZ; Funding acquisition: KZ, SL, FT.

Funding

This work was supported by the National Science Foundation of China (32071489 and 32401305) and the Joint Foundation from the Chinese Academy of Sciences -People's Government of Qinghai Province on Sanjiangyuan National Park (LHZX-2021–03). Drs. Kai Zhao and Sijia Liu were supported by the "Kunlun Talent" Project of Qinghai Province. The field work was supported by the Sino BON-Inland Water Fish Diversity Observation Network and the 2023 award fund of Qinghai Provincial Key Laboratory of Animal Ecological Genomics.

Data availability

The raw transcriptomic data were deposited in Genome Sequencing Archive (GSA) under accession number of PRJCA034357 (https://ngdc.cncb.ac.cn/bioproject/browse/PRJCA034357). The raw data for proteomics were deposited in Integrated Proteome Resources (iPROX) under accession number IPX0005368000.

Declarations

Ethics approval and consent to participate

All the animal experiments were conducted according to the guidelines described in the “Guidelines for Animal Care and Use” manual (NWIPB-202206) approved by the Animal Care and Use Committee, Northwest Institute of Plateau Biology, Chinese Academy of Sciences.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.