Background

Fish migration is the movement from one habitat or region to another typically for spawning [1, 2]. There are three styles among migratory fishes: oceanodromous fishes migrating exclusively within seawater, potamodromous fishes migrating exclusively within freshwater, and diadromous fishes migrating between seawater and freshwater [3, 4]. During migration, fish are confronted multiple challenges posed by environmental conditions including temperature, water flow, and salinity [5, 6]. Three hundred sixty-eight species of diadromous fishes are known (http://www.fishbase.cn), including salmonoids, osmeroids, galaxioids, lampreys, sturgeons, and sicydiine gobies [7,8,9]. However, the Far Eastern daces (Pseudaspius species) are among the only diadromous fishes in Cyprinidae, with limited migratory behavior observed in some other cyprinid species, including idle (Leuciscus idus), roach (Rutilus rutilus), and bream (Vimba vimba) [10,11,12].

Meanwhile, environmental salinity fluctuations exert significant influences on various aspects of fish biology such as osmotic regulation, hormone control, energy metabolism, and growth [13,14,15]. The Far Eastern daces, Pseudaspius species (Pisces, Cyprinidae), speciated and dispersed around the Sea of Japan [16]. The genus Pseudaspius encompasses two freshwater resident species (P. nakamurai and P. sachalinensis) and two diadromous species (P. hakonensis and P. brandtii), belonging to the subfamily Leuciscinae within the large family Cyprinidae [17]. As migratory species of cyprinid fishes, P. hakonensis and P. brandtii comprise a unique group with the freshwater-seawater transition [18], which has attracted much research interest. P. hakonensis and P. brandtii are commonly known as gold beachhead fish and black beachhead fish, respectively, and return from the Japan sea for spawning in Suifen River (Heilongjiang Province in China) from April to July each year [19]. Specifically, they migrate upstream into the estuary in batches after their gonadal maturation, and the parent fish immediately return to the offshore area after spawning; the juveniles grow in the river then they overwinter in the river depths before migrating to the sea [20, 21]. A previous study reported that P. brandtii evolved unique strategies and exhibited higher tolerance and adaptability to salinity and alkalinity than other cyprinids [22, 23]. Therefore, the P. hakonensis and P. brandtii genomes may represent a genetic resource for studies on seawater adaptation in migratory fishes.

Multiple studies have documented the genetic basis of seawater adaptation of salmonid fishes and disentangled osmoregulatory genes and functional groups [24,25,26]. Recently, genomic studies in salmon and trout strikingly revealed a small genomic region (including rock1 and greb1l genes) was strongly associated with migration behavior and timing [26,27,28,29]. Furthermore, simulated experiments conducted on fishes subjected to salinity treatment have revealed insights into the significance of the osmoregulatory gene family, associated metabolic pathways, and gene expression patterns in determining crucial adaptations to seawater [24, 30,31,32]. For example, the euryhaline Javafish medaka (Oryzias javanicus) genome sequencing provided insights into the molecular basis of osmotic regulation and hatching enzyme activity [32]. Nevertheless, the genetic mechanism of Pseudaspius species (P. brandtii and P. hakonensis) adapting to seawater is unclear.

In this study, we focused on two species of cyprinid fishes in seawater, P. hakonensis and P. brandtii, to unravel the genetic basis of their adaptation to seawater during the migratory process. To achieve this, we first generated high-quality, chromosome-level assemblies of the genomes of P. hakonensis and P. brandtii. By comparing these genomes with those of other teleosts, we identified genes that underwent positive selection and rapid evolution as well as several conserved non-coding elements (CNEs) that are likely responsible for their seawater adaptation. Importantly, we validated the Pseudaspius-specific genetic innovations using functional assays. The findings of this study shed light on the intricate genetic mechanisms underlying seawater adaptation in Pseudaspius species. These discoveries serve as a foundation for further investigations into the adaptation of migratory fishes and contribute to our understanding of their remarkable ability to thrive in osmotically contrasting.

Results

Genome assembly and genome annotation

Using Illumina, Nanopore, and Hi-C sequencing, we acquired a total of 236.91 Gb data for P. hakonensis and 231.44 Gb data for P. brandtii, respectively (Fig. 1A, B, and C, Additional file 1: Table S1). Subsequently, we generated chromosome-level genome assemblies for P. hakonensis and P. brandtii. The assemblies of P. hakonensis and P. brandtii genomes were 840.77 Mb and 848.49 Mb, with contig N50 length of 3.68 Mb and 3.63 Mb, GC content of 39% and 38%, respectively (Additional file 1: Table S2 and S3). To further refine the assemblies, we anchored the contigs onto 25 chromosomes with mounting rates of 96.54% for P. hakonensis and 97.20% for P. brandtii. The genome-wide Hi-C heatmaps showing chromosome crosstalk adhered to the interaction rule, with noticeably stronger signal strength around the diagonal compared to other positions (Additional file 2: Figure S1, Additional file 1: Table S4. This suggests the high quality and completeness of the genome assembly. The summary of genome characteristics of P. hakonensis and P. brandtii are shown in Fig. 1B, C. Evaluation of genome completeness based on BUSCO identified that 94.50% of the complete BUSCO in the P. hakonensis genome assembly including 88.00% of the complete and single-copy genes and 6.50% of the complete and duplicated genes, while 94.90% of the complete BUSCO in the P. brandtii genome assembly including 87.90% of the complete and single-copy genes and 7.00% of the complete and duplicated genes (Additional file 1: Table S5). These results confirm the high quality of our genome assemblies for both species, providing a solid foundation for our investigations into the genetic basis of seawater adaptation in P. hakonensis and P. brandtii.

Fig. 1
figure 1

Sampling sites and genome assemblies for P. hakonensis and P. brandtii. A A sampling site (red circle) in Suifen River (Helongjiang province, China) and a photo of the P. brandtii. B Circos plots showed the distributions of genomic components in P. hakonensis; the outer circle was the chromosome. C Circos plots showed the distributions of genomic components in P. brandtii; the outer circle represents the chromosome. D Chromosomal syntenic relationship of P. brandtii, P. hakonensis, and D. rerio. The number (1–25) represented the number of the chromosome

Full size image

We conducted an analysis of repetitive sequences and protein-coding genes in the genomes of P. hakonensis and P. brandtii. Approximately 36.93% of the bases in the P. hakonensis genome and 37.20% of the bases in the P. brandtii genome were classified as repetitive sequences (Additional file 1: Table S6). We also predicted a total of 23,839 and 24,012 protein-coding genes in P. hakonensis and P. brandtii genomes, respectively (Additional file 1: Table S7). Approximately 97.54% of P. hakonensis and 97.69% of P. brandtii predicted genes were successfully annotated using the Non-Redundant Protein Sequence Database (NR), InterProScan, Uniport (Universal Protein), Pfam, and eggNOG (evolutionary genealogy of genes: Non-supervised Orthologous Groups) databases (Additional file 1: Table S8). To gain further insights into the genomic relationships, we conducted a genome collinearity analysis between P. hakonensis, P. brandtii, and Danio rerio [33]. The results showed that these species exhibited good collinearity without any evidence of chromosomal fission or fusion events (Fig. 1D). This corroborates the high-quality and chromosome-level nature of our assembled genomes for P. hakonensis and P. brandtii, which provide a solid foundation for our investigations into the genetic basis of seawater adaptation in these two Far Eastern daces.

Phylogenetic relationships and population history

We conducted a phylogenetic analysis based on 1597 single-copy orthologous genes to explore the evolutionary relationships and divergence times of P. hakonensis, P. brandtii, and 16 other teleost fishes (Additional file 2: Figure S2). The time-calibrated phylogeny showed that P. hakonensis and P. brandtii formed a sister clade to grass carp (Ctenopharyngodon idella), bighead carp (Hypophthalmichthys nobilis), silver carp (H. molitrix), and blunt snout bream (Megalobrama amblycephala). The two clades shared a most recent common ancestor (MRCA) approximately 27.86 million years ago (Ma) during the Palaeogene period (Fig. 2A). The divergence time between P. hakonensis and P. brandtii was estimated to be 2.51 million years ago (Ma). These results shed light on the evolutionary history and genetic relationships of these two Far Eastern daces and their place within the broader teleost fish lineage.

Fig. 2
figure 2

Phylogenetic analysis and demographic histories of P. hakonensis and P. brandtii. A Phylogenetic tree and divergence times estimated for the P. brandtii, P. hakonensis, and other 16 teleosts. Error bars indicate 95% confidence levels. P. hakonensis and P. brandtii were highlighted in bold. B Demographic histories were reconstructed using PSMC. Blue lines represent P. hakonensis, and green lines represent P. brandtii. Dark grey box represents glacial periods, and French grey box represents interglacial periods

Full size image

The population history was reconstructed by calculating the heterozygosity of the genomes of two species using pairwise sequentially Markovian coalescent (PSMC) model [34]. The population demography analysis of the two species exhibited a similar starting point around 104 to 107 years ago, but thereafter, they followed divergent trends. These patterns could be correlated with significant climatic oscillations, including glacial-interglacial cycles and sea level changes. Such environmental factors might have played a crucial role in shaping the population dynamics and distribution of these Far Eastern daces over time (Fig. 2B). We observed that the first population contraction occurred around 6 million years ago (Ma), followed by a remarkable and massive expansion around 0.9 Ma, reaching its peak at approximately 0.6 Ma during the largest Quaternary glaciation (0.8–0.2 Ma). Subsequently, a second contraction phase began. These fluctuations in population size over geological timeframes shed light on the species’ responses to major climatic events and offer valuable insights into its evolutionary history and adaptive strategies. Similarly, the size of the P. brandtii population decreased around 9 Ma, followed by a population expansion at approximately 0.7 Ma and peaked around 0.4 Ma during the largest Quaternary glaciation (0.8–0.2 Ma). Our analysis suggests that both Pseudaspius species populations experienced multiple bottlenecks that correlated with glacial activities. Specifically, the population declines coincided with the onset of warm interglacial periods when deglaciation dramatically reduced the available habitat. Notably, throughout various periods of evolution, the effective population size of P. hakonensis remained relatively higher and more stable compared to P. brandtii. This observation may be attributed to P. hakonensis’ robust adaptability and resistance to changing environments. These population dynamics and differences in effective population size provide valuable insights into the responses and resilience of the two Pseudaspius species to historical climatic changes and the complex interplay between environmental conditions and their evolutionary strategies.

Gene family expansion and contraction

The expansion and contraction of gene families can have a significant impact on speciation and environmental adaptation [35]. The analysis of gene family evolution showed that 148 gene families were expanded, while 837 gene families were contracted in the common ancestor of P. hakonensis and P. brandtii, respectively (Additional file 2: Figure S3). Based on the results, further enrichment analysis was performed using the GO and KEGG databases. The expanded gene families were significantly enriched in 174 GO terms and 16 KEGG pathways (Additional file 2: Figure S4A and S4B), such as protein-DNA complex subunit organization (GO:0071824), nucleosome (GO:0000786), DNA packaging (GO:0006323), chromatin assembly (GO:0031497), DNA conformation change (GO:0071103), natural killer cell mediated cytotoxicity (ko04650), glycosylphosphatidylinositol (GPI)-anchored proteins (ko00537), and CD molecules (ko04090) that associated with cell growth and metabolism and cell damage repair.

Conversely, the contraction gene families were significantly enriched in 645 GO terms and 61 KEGG pathways (Additional file 2: Figure S4C and S4D), such as negative regulation of proteolysis (GO:0045861), activation of immune response (GO:0002253), cell adhesion molecules (ko04514), and phagosome (ko04145) that associated with immune response. These results suggested that these biological processes of gene family expansion and contraction in the two Pseudaspius species may be associated with their adaptation to the seawater.

Adaptation to osmoregulation

Migratory fish exhibit the capacity to persist in both freshwater and high salinity seawater environments by employing osmotic regulation mechanisms. These regulatory processes are essential for maintaining cellular homeostasis and achieving internal environmental balance [36]. The branch-site model was employed to identify the genes evolving under positive selection for seawater adaptation. Among the 7194 single-copy orthologous genes across seven species, a total of 355 positively selected genes (PSGs) were identified in the ancestor of P. hakonensis and P. brandtii genomes (Additional file 2: Figure S5). We found nine PSGs were strongly associated with osmoregulation, including cldn10, igf2, slc2a3, kcnn3, ocln, map2k7, aqp3, slc15a2, and cldn19 (Fig. 3A, Table 1, Additional file 1: Table S9), presumably reflecting the seawater adaptation of P. hakonensis and P. brandtii. Additionally, we detected 704 rapidly evolving genes (REGs) in the ancestor of P. hakonensis and P. brandtii genomes. Among them, seven genes related to osmoregulation were found, including ocln, slc2a3, prl, aqp11, slc6a8, slc12a1, and map3k3 (Fig. 3A, Table 2, Additional file 1: Table S9).

Fig. 3
figure 3

Genomic evidence of osmoregulation related genes in P. hakonensis and P. brandtii. A Genes and pathways involved in osmoregulation are shown in the colored circles and light blue box, respectively. PSGs are marked in pink circles, REGs are marked in purple circles. B Immunoblotting assay of prl in HEK293T cells. HEK293T cells were cultured in 6-well plates and transfected with indicated plasmids (3 μg each) for 24 h, and then the cells were harvested to perform immunoblotting. C prl-WT reduces cell crumbling and abscission after high osmotic pressure treatment. HEK293T cells were cultured in 6-well plates and transfected with indicated plasmids (3 μg each) for 8 h and then cultured under high osmolarity condition (600 mOsmol/kg) for 16 h and imaged under a microscope Nikon TE2000-U. D–E prl-WT reduces cell apoptosis in high osmolarity condition (600 mOsmol/kg) or low osmolarity condition (200 mOsmol/kg) detected by flow cytometry analysis (D), and statistical analyses were performed using the GraphPad Prism software (7.0) (E). HEK293T cells were cultured in 6-well plates and transfected with the indicated plasmids (3 μg each) for 8 h and then cultured under high osmolarity condition (600 mOsmol/kg) for 16 h (high group) or cultured under high osmolarity condition (600 mOsmol/kg) for 12 h followed by low osmolarity condition (200 mOsmol/kg) for 4 h (low group). Data show mean ± SD; Student’s two-tailed t test. *P < 0.05, **P < 0.01, ***P < 0.001, data from three independent experiments

Full size image
Table 1 Positively selected genes with osmoregulation of both P. hakonensis and P. brandtii
Full size table
Table 2 Rapid evolution genes with osmoregulation and opsin of both P. hakonensis and P. brandtii
Full size table

Among these important genes, ocln and slc2a3 genes had both rapid evolution and positive selection (Fig. 3A, Table 1 and Table 2). The gene slc2a3, namely the facilitative Na+-independent sugar transporters, which contributes to osmoregulatory activities in the fish kidney [37], showed positive selection and accelerated evolution. It was found with A42E and V147L specific amino acid site mutations in both P. hakonensis and P. brandtii, these mutations occurred in the MFS_1 domain and were likely to have an impact on the three-dimensional structure of the protein (Additional file 2: Figure S6A and S6B). Another gene, cldn10, which participates in the creation and maintenance for cation-selective pore junctions in salt-secreting tissues of teleost fishes [38], was found with Y79H and L92T specific amino acid site mutations which occurred in the Claudin_2 domain in both P. hakonensis and P. brandtii, and this was also likely to affect the three-dimensional structure (Additional file 2: Figure S6A and S6C).

The prl gene directly contributes to the survival and functionalization of branchial ion cells involved in osmoregulation (Additional file 2: Figure S6D) [39]. To further investigate whether rapid evolution of prl (prolactin) in Pseudaspius is involved in their salinity adaptation, we first cloned and tested the expression of prl wild-type (prl-WT) and mutant (prl-MUT) (Additional file 2: Figure S6D). Immunoblotting showed that both prl-WT and prl-MUT could be detected in transfected HEK293 cells (Fig. 3B). Given that hypertonicity induced by high bay salt, the main component of which is NaCl, decreases cell volume, increases cytosolic osmolality, and alters mitochondrial osmotic equilibrium (Michea et al., 2002), we analyzed whether prl gene prevents the corresponding hypertonic threat. As expected, cells treated with high osmolarity (600 mOsmol/kg) showed significant shrinkage and shedding, and then overexpression of prl, especially overexpression of prl-WT, significantly improved the cell status (Fig. 3C). Previous studies have shown that changes in osmolarity induce cells to enter the apoptotic program (Michea et al., 2002; Dmitrieva et al., 2000). To further elucidate the effect of prl on osmolarity-induced apoptosis of HEK293 cells, flow cytometry analysis was performed. As shown in Fig. 3D, E, apoptosis was significantly induced by high osmolarity (600 mOsmol/kg), in contrast to overexpression of prl-MUT, overexpression of prl-WT significantly inhibited osmolarity-induced cell apoptosis. Taken together, the results suggest that the Pseudaspius prl gene enhanced cell adaptation to altered osmolarity.

Vision adaptation

In addition to osmotic regulation, freshwater and marine transition also require a suite of adaptations in other physiological and life history traits, including diet, metabolism, defense from predators, reproductive strategies, and even vision (Van Nynatten et al., 2021). The spectra of available light differ significantly between marine and freshwater habitats. Light that penetrates the open ocean is known to shift towards the blue end of the spectrum with increasing depth (Jerlov, 1968). In contrast, the light spectrum in riverine environments attenuates more rapidly, is more spectrally diverse, but has been less explored concerning the visual systems of riverine animals (Levine and MacNichol, 1979; Costa et al., 2013). These distinctions in the aquatic light environment are believed to impact the visual systems of fishes, leading to freshwater fishes generally exhibiting more redshifted visual pigments (Rennison et al., 2016; Musilova et al., 2019).

Rhodopsin (rh1), the light-sensitive visual pigment expressed in rod photoreceptors, is specialized for vision in dim-light environments and has been confirmed to be associated with marine to freshwater croaker invasion (Van Nynatten et al., 2021) and was also reported to play an important role in migratory eels [40, 41]. Our results showed that the rh1 gene involved in phototransduction cascade exhibited a rapid evolutionary rate that may have implications in adapting to distinct light environments (Fig. 4A, Table 2). Subsequently, rh1 gene underwent 11 specific amino acid replacements which occurred in the Rhodopsin_N and 7TM_GPCR_Srsx domains in both P. hakonensis and P. brandtii, and these amino acid mutation sites also located in the key position of the functional domain (Fig. 4B, C). Therefore, the accelerated evolution of the rh1 gene is likely to be closely associated with the diverse light conditions encountered during the migration of P. hakonensis and P. brandtii.

Fig. 4
figure 4

Genomic evidence of visual opsin genes in P. hakonensis and P. brandtii. A Genes involved in phototransduction cascade are shown in the colored circles; a REG was marked in purple circle. B The three-dimensional structure of RH1 protein. C The rh1 gene had specific AA replacement in P. hakonensis and P. brandtii based on sequence alignments

Full size image

Conserved non-coding elements (CNEs) linked to osmoregulation

CNEs are usually distributed in the upstream and downstream regions of genes and served as the core components of gene regulatory networks to regulate gene expression [42]. To determine the extent of CNEs loss or gain in P. hakonensis and P. brandtii, we predicted genome wide CNEs with reference to the zebrafish genome. Consequently, we successfully identified 18,581 CNEs across 11 teleosts through syntenic alignments. We further analyzed the CNEs that were specifically deleted or inserted in P. hakonensis and P. brandtii and found that they were present in the neighborhood of 175 genes with 196 CNE sequences, including atg7 and usp45 genes related to osmoregulation. In particular, there was one CNE (~ 160 bp) near atg7 with an 8-bp specific deletion in the P. hakonensis and P. brandtii (Fig. 5A, C) and another CNE (~ 180 bp) near usp45 with a 7-bp specific insertion in the P. hakonensis and P. brandtii (Fig. 5B) using zebrafish annotations as reference, which may be associated with adaptation to seawater.

Fig. 5
figure 5

CNE sequence analysis that may be related to osmoregulation. A An 8-bp specific deletion of P. hakonensis and P. brandtii was located near target gene atg7. B A 7-bp specific insertion of P. hakonensis and P. brandtii was located near target gene usp45. C VISTA sequence conservation plot of the P. hakonensis and P. brandtii specific deletion CNE around atg7, using zebrafish as reference. Red arrows indicate missing bp. D, E Luciferase activity assays of the fragments of atg7 (D) and usp45 (E) promoter from Danio rerio, P. brandtii, and P. hakonensis under control (315 mOsmol/kg) or high osmolarity conditions (600 mOsmol/kg) in HEK293 cells. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

Full size image

To further explore whether the Pseudaspius-specific CNEs near atg7 and usp45 genes were involved in osmoregulation, we inserted these CNEs into the pGL3-Basic vector and co-transfected them into HEK293T cells, which were cultured under control (315 mOsmol/kg) or high osmolarity conditions (600 mOsmol/kg). Then, their promoter activity was determined using the dual-luciferase reporter assay system. The results of our reporter assay suggest that the upstream fragment of atg7 and usp45 from P. brandtii and P. hakonensis exhibit higher promoter activity compared to corresponding fragments from zebrafish (Fig. 5D, E). Notably, the promoter activity of P. brandtii and P. hakonensis atg7 was significantly induced at high osmolarity (Fig. 5D). Overall, these results demonstrated that the Pseudaspius-specific CNEs near atg7 and usp45 genes played a significant role in osmoregulation.

Discussion

Two of the few migratory species cyprinids, P. hakonensis and P. brandtii, are ideal fish species to investigate the genetic basis of adaptation to seawater [21, 43]. With advancements in sequencing technology, comparative genomics studies leverage available data to reveal the evolutionary patterns of adaptive traits in fishes in special environmental conditions [44,45,46]. Here, we generated chromosomal level genome assemblies of P. hakonensis and P. brandtii with 840.77 Mb and 848.49 Mb, respectively. Both of P. hakonensis and P. brandtii were anchored with 25 chromosomes and had good collinearity with zebrafish. Analysis of population history showed that two species formed sister groups with C. idella, H. nobilis, H. molitrix, and M. amblycephala in Cyprinidae. The genomic information of two species will contribute to a deeper understanding of the genetic mechanism of seawater adaptation.

Surviving in diverse aquatic environments requires ability to regulate their osmotic pressure balance, since salinity plays a crucial role in the ecological environment [47]. Previous studies investigated the biological process genes including metabolism, immune system, cell activity, and growth potentially associated with Nile tilapia adaptation to different salinities as well as phenotypic differentiation of morphology, physiology, and behavior between the migratory population and the freshwater resident population of Coilia nasus in the process of adaptive evolution [48, 49]. We hypothesized that these biological processes of gene family expansion and contraction in the two Pseudaspius species might be linked to the physiological and behavioral changes observed in the two Pseudaspius species as they adapted to the evolution of seawater.

In previous studies, hatching enzymes associated with the evolution of salinity adaptation has been identified in a whole genome of Oryzias javanicus, revealing the genetic mechanism of adaptation to different salinities [32]. Similarly, the Atlantic sturgeon has the physiological capability to move between salinity habitats with varying salinity. Tests on its growth and osmotic regulation have indicated its adaptability to these changing salinity conditions. [50]. In the current study, our analyses yielded a list of adaptively evolving genes, which comprised 355 PSGs and 704 REGs in the ancestor node of P. hakonensis and P. brandtii. Notably, several genes related to osmotic regulation were identified as PSGs and REGs. The candidate genes ocln, cldn10 and cldn19 encoded the components of a tight connection (Fig. 3A). Previous studies have provided evidence identifying the ocln gene as an osmoregulation-related gene in Acipenser baerii [51], while the cldn19 gene played significant roles in osmoregulatory physiology of Petromyzon marinus [52]. Moreover, the PSG cldn10, which predominantly expressed in osmoregulatory organs of teleost fishes [38, 53], was found have two specific AA replacements in the two Pseudaspius species (Additional file 2: Figure S6A). In addition, osmotic regulation affected the expression of many SLCs (solute carriers), such as slc15a2, slc6a8, slc12a1, and slc2a3 [24, 54,55,56]. For instance, slc2a3 gene was important in the fish kidney for osmoregulatory activities [56]. Here, slc15a2, slc6a8, slc12a1, and slc2a3 genes were involved in material (such as sodium/potassium/chloride, neurotransmitter, oligopeptide, and glucose) exchange (Fig. 3A). Importantly, slc2a3 is identified as a both positively selected and rapidly evolving gene that had two specific amino acid replacements in the MFS_1 domain (Additional file 2: Figure S6A and 6B). Meanwhile, several ion transporters and genes were also important in osmoregulation, such as kcnn3 that has been found to be upregulated in an euryhaline pupfish adaptation to high salinity conditions [57] (Fig. 3A).

Aquaporins (AQPs) are crucial for rapid osmoregulation by facilitating the diffusion of water and osmolytes across cellular membranes [58]. Recent studies suggested that the aqp3 gene potentially contributes to water metabolism and osmotic pressure regulation in Lateolabrax maculatus [59]. Moreover, hypertonic conditions have resulted in increased expression of aqp11 in Clarias magur hepatocytes [60, 61]. In particular, our genomic analyses showed that aqp3 and aqp11 genes may modulate the exchange of water molecules between the inside and outside of the cell membrane (Fig. 3A). We also found igf2 to be under selection, which influences osmoregulation by stimulating Na + /K + -ATPase activity and promoting chloride cell differentiation (Chandhini et al., 2021) (Fig. 3A). A prl gene has been showed to have a reduced expression level following seawater adaptation in Oncorhynchus rhodurus (Yada et al., 2010) (Fig. 3A). The prl and igf2 genes bind to receptors and transmit signals to PI3K, participating in the PI3K-Akt signaling pathway (Fig. 3A). It has been demonstrated that Mitogen-activated protein kinase (MAPK) cascades are major participants in osmotic stress response of fish, and are involved in regulating cell proliferation, growth, death, and other physiological processes [62,63,64]. The expression of the map2k7 gene has been showed to be upregulated in the muscle of Penaeus monodon during acute hyposaline stress, and map3k3 also contributes to the activation of the osmoprotective transcription factor nfat5 in response to high NaCl [65, 66]. The genes, in particular map2k7 and map3k3, part of MAPK signaling pathway, were identified as PSG and REG in the two Pseudaspius species, respectively (Fig. 3A). The role of PRL in osmoregulation has been discovered as early as 1950s by researching on killifish (Fundulus heteroclitus), with F. heteroclitus unable to survive in freshwater following hypophysectomy while PRL treatment allowed hypophysectomized killifish to survive [67]. Subsequently, several studies found a direct contribution of PRL to osmoregulation by knocking out this gene in zebrafish (Shu et al., 2016). However, it is still unclear whether the rapid evolving of prl gene in two Pseudaspius species contribute to their adaptation to seawater. In our study, by integrating comparative genomics and functional assays, we found that the rapid evolving of prl gene is indeed associated with their osmoregulation in the two Pseudaspius species. Osmoregulatory candidate genes are potentially important that contribute to the adaptation of Pseudaspius species to seawater.

Previous studies demonstrated that visual systems have undergone diversification in migratory fish [68]; freshwater eels changed their spectral sensitivity by switching chromophore types or expressing different opsin genes in response to changing light environments during migration from fresh water to deep sea environments [68, 69]. Moreover, cloning and expression analysis of freshwater and deep-sea opsin genes in the eel A. japonica have revealed that the rod-shaped visual pigment (rh1) changes from “freshwater” form to “deep-sea” form during maturation [70, 71]. The rh1 underwent positive selection during salmonid evolution, linked to divergent light environments [72]. Interestingly, our genomic analyses found that rh1 is a rapidly evolving gene with several amino acid replacements (Fig. 4). Therefore, the genetic innovations may be associated with the seawater light environment of P. hakonensis and P. brandtii.

Many CNEs function as cis-regulatory elements, such as enhancers, repressors, and insulators, and the presence or absence of CNEs plays a significant role in phenotypic differences and morphological innovation [73,74,75]. Notably, our genomic analyses found that there were CNEs loss near the atg7 and obtain close to usp45 (Fig. 5). Thus, we propose CNEs play pivotal roles in osmoregulation during seawater adaptation of P. hakonensis and P. brandtii. Previous reports have indicated adverse effects of hypotonic stress on various developmental stages of Macrobrachium rosenbergii larvae; the expression of the autophagy-related gene atg7 showed significant changes under hypotonic stress [76]. In salinity-induced transcriptome studies of marine and freshwater threespine stickleback, a decrease in usp45 expression was observed 6 h after the sudden exposure to seawater [77]. Notably, our genomic analyses found that there were CNEs loss near the atg7 and obtain close to usp45 (Fig. 5). Thus, we propose that this may play pivotal roles in osmoregulation and will yield insights into seawater adaptation of P. hakonensis and P. brandtii.

Conclusions

We acquired high-quality chromosome-level genome assemblies of P. hakonensis and P. brandtii. Genome evolution analysis provided insights into genetic changes related to osmoregulation and vision that may be critical to the seawater adaptation in P. hakonensis and P. brandtii. Our results provide opportunities to elucidate the molecular mechanisms underlying the evolutionary adaptations of migratory fish to seawater and present a valuable resource for future research on adaptive evolution in Pseudaspius species.

Methods

Sample DNA and RNA extraction, sequencing

P. hakonensis individual and P. brandtii individual were collected from Suifen River in Dongning City, Heilongjiang Province, China. Genomic DNA was extracted from the muscle using Puregene Tissue Core 81 Kit A (Qiagen, Maryland, USA). DNA integrity was checked using 1% agarose gel electrophoresis and Pultton DNA/Protein Analyzer (Plextech, Berkeley, USA). Total RNA was extracted from tissues of P. hakonensis and P. brandtii, including muscle, eye, heart, liver, kidney, gill, spleen, spermary, swim bladder, and brain, by using TRIzol reagent (Invitrogen, USA). RNA quality was checked with a NanoDrop ND-1000 spectrophotometer (Labtech, Palaiseau, France) and a 2100 Bioanalyzer (Agilent Technologies, Waldbronn, Germany).

A paired-end library with inset sizes of 150 bp was constructed according to the Illumina protocol and sequenced on the NovaSeq 6000 platform (Illumina) to assess the complexity of the genome, correction of the genome assembly, and assembly evaluation. Hi-C libraries were constructed and sequenced to obtain a chromosomal-level genome assembly. After reads with low-quality bases (reads with more than 10% N bases or low-quality bases ≤ 5), adapter sequences, and duplicated sequences were discarded, the clean reads were used for subsequent analysis. Nanopore libraries were generated according to the manufacturer’s instructions and sequenced on 13 flow cells using the GridION X5 DNA sequencer (Oxford Nanopore). For the Nanopore reads, the mean quality of each read was calculated, and any reads with mean quality > 7 were retained for subsequent analyses. An SMRTbell library with a fragment size of 20 kb was constructed using a SMRTBell template preparation kit 1.0 (PacBio) by following the manufacturer’s protocol, and the library was sequenced with the PacBio Sequel II system. Then, full-length refined consensus transcripts were generated using the PacBio Iso-Seq pipeline (https://github.com/PacifcBiosciences/IsoSeq).

Genome size estimation and genome assembly

All the filtered Illumina pair-end reads were used for k-mer frequency analysis [78]. The genomic size was estimated on the basis of the following formula: G = k-mer number/k-mer depth, where G represents the genome size, where k-mer number are the total number, and k-mer depth represents the average depth of 17-mers, respectively. After the Nanopore low-quality reads were filtered, the remaining high-quality reads were first calibrated using the CANU (version 1.5) [79] and then assembled using the genome assembler FLYE (version 2.6) [80] with default parameters. The original Nanopore reads were polished for three rounds using racon (version 1.2.1) [81] and further corrected for three rounds using Illumina reads with pilon (version 1.21) [82] for a consistent assembly. The filtered Hi-C reads were first aligned to the assembled genome using the BOWTIE (version 2.2.5) [83]; only the read pairs with both ends uniquely aligned to the genome were selected. Then, the corrected contigs and valid Hi-C reads were used to perform the chromosomal-level genome assembly by LACHESIS [84] with default parameters. In addition, a heatmap based on the interaction signals was constructed to evaluate the quality of the chromosomal-level genome assembly. Finally, the completeness of our final genome assembly was evaluated using BUSCO (version 3) [85].

Genome annotation

In order to predict the repeat elements in P. hakonensis and P. brandtii, REPEATMODELER (version 1.0.5) [86] was used to construct a de novo transposable element library, which was then used to predict repeats with REPEATMASKER (version 4.0.6) [86]. In the P. hakonensis and P. brandtii genomes, ab initio, homologues, and RNA-sequencing methods were used for predicting the protein-coding genes. The above three gene sets were integrated to yield a comprehensive and nonredundant gene set using EVIDENCEMODELER (EVM, version 1.1.1) [87]. Then, the integrated gene set was translated into amino acid sequences and performed gene functional annotations by searching against known databases, including NR, InterPro, Uniport, Pfam, and eggNOG, with a cutoff E value of 1e − 5.

Phylogenetic analysis

Protein sequences from a total of 18 species including P. hakonensis and P. brandtii (assembled in this study), C. Idella [88], H. nobilis, H. molitrix [89], M. amblycephala [90], D. rerio, Astyanax mexicanus [91], Ictalurus punctatus [92], Pelteobagrus fulvidraco [93], Gasterosteus aculeatus [94], Takifugu rubripes [95], Oreochromis niloticus [96], O. latipes [97], Xiphophorus maculatus [98], Hippocampus comes [99], Boleophthalmus pectinirostris [100], and Lepisosteus oculatus [101] were constructed by ORTHOMCL (version 2.0.9) with default settings, and 1597 single-copy orthologues were identified. Then, the protein sequences of single-copy orthologous genes were further aligned using MUSCLE [102] with default parameter settings. The maximum likelihood method was used to construct a phylogenetic tree using RAXML (version 8.1.19) [103] with 1000 bootstraps under the GTR + I model. MCMCTREE of the PAML package [104] was used to calculate the divergence times among P. hakonensis and P. brandtii and 16 other teleost fishes, and the fossil records were obtained from TIMETREE database (http://www.timetree.org/) for calibration; we set four calibration time points (C. Idella–L. oculatus: ~ 298.8–342.5 MYA; C. Idella–D. rerio: ~ 41.7–68.9 MYA; H. comes–G. aculeatus: ~ 99.5–116.7 MYA; O. niloticus–X. maculatus: ~ 81.2–91.6 MYA).

Demographic history

We concluded the demographic history for the P. hakonensis and P. brandtii by pairwise sequentially Markovian coalescent (PSMC) analysis [34]. Heterozygote sites across the genome were used to estimate changes in effective population size. To generate consensus diploid sequences for the two individuals separately by SAMtools (version 1.3.1) [105], the “fq2psmcfa” tool was used to create the input file for PSMC modeling, with the option -q20. Moreover, the effective population history was inferred using PSMC with 100 bootstraps and plotted using the psmc_plot.pl pipeline. The generation time for two species was set as 4 years, and the mutation rate was 1.3387e − 09 per year per nucleotide for P. hakonensis and 1.2125e − 09 per year per nucleotide for P. brandtii. In addition, we obtained the atmospheric surface air temperature (°C) and Eurasian ice volume (m sea level equivalent) data for the past 3 million years from the NCEI (http://www.ncdc.noaa.gov/).

Gene family expansion and contraction

To determine gene family expansion and contraction in the P. hakonensis and P. brandtii genomes, CAFÉ (version 3.1) [106] was used with default parameter settings. If the copy number of the P. hakonensis and P. brandtii were higher or lower than those found in their close ancestral clades, then we identified the gene family as a substantially expanded or contracted gene family. A conditional P-value for each gene family was calculated, and gene modules of the significantly expanded and contracted gene family were extracted and were subjected to GO and KEGG functional enrichment analyses with a false discovery rate (FDR) adjusted P-value < 0.05, respectively. During functional enrichment, the zebrafish genes were selected as background. GO terms or KEGG pathways with a P-value < 0.05 were considered significantly enriched.

Identification of positively selected genes and rapidly evolving genes

Single-copy orthologous genes were used in the detection of PSGs and REGs. The branch-site model incorporated in the PAML (version 4.8) package [107] was employed to detect PSGs in P. hakonensis and P. brandtii genomes, with the non-migrated fishes as the background. The null model used in the branch-site test assumed that the comparison of the substitution rates at nonsynonymous and synonymous sites (Ka/Ks ratio) for all codons in all branches must be ≤ 1, whereas the alternative model assumed that the foreground branch included codons evolving at Ka/Ks > 1. The P-values were computed based on the chi-square statistic adjusted by the false discovery rate (FDR) method, and genes with adjusted P-value < 0.05 were considered as PSGs. We identified REGs in the P. hakonensis and P. brandtii genomes using the branch model in the CODEML program in PAML. The likelihood ratio test (LRT) was employed to discriminate between alternative models for each orthologue with df (degree of freedom) = 1. We treated the orthologue as evolving with a significantly faster rate in foreground branch if the FDR adjusted P-value < 0. 05 and a higher ω (Ka/Ks) ratio in the foreground branch than in the background branches.

Identification of specific amino acid mutations and protein structure simulation

The homologous protein sequences of P. hakonensis and P. brandtii and other fishes were aligned and compared. A particular amino acid mutation was defined as the P. hakonensis and P. brandtii amino acid were different from acids of all other species including C. Idella [88], H. nobilis, H. molitrix [89], M. amblycephala [90], and D. rerio. Missing sites may be present in the protein sequences of species with poorly assembled genomes because of assembly quality among different genomes. Therefore, the specific amino acid mutation was considered only if at least 80% of species had valid amino acid information. Then, it was determined whether the mutated amino acids located in the protein domain of the P. hakonensis and P. brandtii using Pfam [108].

First, we preliminarily predicted the slc2a3, cldn10, and rh1 protein structures based on the amino acid sequence on the SWISS-MODEL online website and downloaded the PDB file (https://swissmodel.expasy.org/). Next, the predicted result with the maximal score was selected as the best structure and visualized by Pymol (http://www.PyMOL.org/).

Identification of conserved noncoding elements

To identify CNEs, using the D. rerio genome as a reference, whole genome alignments for 11 teleosts were generated, and the genome data were downloaded from ENSEMBL; they were P. hakonensis and P. brandtii (assembled in this study), C. Idella, H. nobilis, H. molitrix, M. amblycephala, D. rerio, I. punctatus, G. aculeatus, O. latipes, and L. oculatus. First, the results of the LAST whole genome alignment were submitted to the LAST subroutine “Maf-swap” [109], and the sequence in the maf format alignment was changed; then, the MULTIZ v3 subroutine “roast” [110] was used to extract sequence alignments at the same location on the chromosomes of different species. Next, further analyses were run with subprograms of PHAST [111]. “msa_view” was used to identify aligned 4D sites of alignments with parameters “–informat MAF –4d –features.” We estimated a neutral phylogenetic model using “phyloFit” [112]. We estimated rho (expected substitution rate of conserved elements relative to neutrality) using “phastCons” with “–estimate-rho –no-postprobs–most-conserved –score.” Conserved models for each window were combined by phyloBoot [112] and then used for initial conserved element prediction. The coding regions and non-coding regions of each gene based on the GFF3 file of the D. rerio genome were separated by using ANNOVAR [113] after conserved elements had been identified. Finally, a series of python scripts were used to detect the divergent CNEs and divergent coding genes. A final set of 18,581 CNEs with a minimum of 30-bp in length were identified. We mainly focused on CNEs which are located in the intergenic regions, in introns, and within the 10-kb upstream or downstream flanking regions of genes. Finally, the VISTA [114] was used to plot the alignments after determining the required CNE sequence.

Detection of apoptotic cells

HEK293T cells were cultured under different osmolarity conditions as indicated. HEK293T cells were cultured in 6-well plates and transfected with indicated plasmids (3 μg each) for 8 h and then cultured under high osmolarity condition (600 mOsmol/kg) for 16 h and imaged under a microscope Nikon TE2000-U. For flow cytometry analysis, cells were harvested and stained with FITC Annexin V and PI using the FITC Annexin V Apoptosis Detection Kit I (556,547, BD Pharmingen) according to the manufacturer’s instructions. The detection of apoptotic cells was performed using the Beckman CytoFLEXS, and the data were analyzed using the CytExpert software.

Dual-luciferase assay

The chemically synthesized fragments of specific CNEs near Danio rerio atg7 (NC: 007122.7_814119-815,755), P. brandti atg7 (Chromosome 21: 1,036,539–1038266) and P. hakonensis atg7 (Chromosome 21: 42,224–43,950), D. rerio usp45 (NC: 007127.7_32717090-32718855), P. brandti usp45 (Chromosome 10: 21,314,205–21315977), and P. hakonensis usp45 (Chromosome 10: 20,933,460–20935232) were inserted into the pGL3-Basic vector (Promega, USA) and were sequenced (Sangon Biotech, China). The HEK293T cells were seeded in 24-well plates for 24 h and then co-transfected with 200 ng constructed pGL3-Basic vectors and 3 ng pRL-CMV as an internal control using VigoFect (Vigorous Biotech). Sixteen hours after transfection, cells were cultured under control (315 mOsmol/kg) or high osmolarity conditions (600 mOsmol/kg) for 8 h and then harvested using passive lysis buffer. Luciferase activity in cell extracts was determined using a dual-luciferase reporter assay system (Promega) according to the manufacturer’s protocol. The light output of the transcription activity was divided by the output of the Renilla luciferase activity for sample normalization.

Statistical analysis

All statistical analyses were performed using the GraphPad Prism software (7.0). Results with error bars represent mean ± SD. Statistical analysis was performed using Student’s two-tailed t-test. A P-value less than 0.05 was considered significant. Statistical significance is indicated as follows: *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.