Comprehensive analysis of genes contributing to euryhalinity in the bull shark, Carcharhinus leucas; Na⁺-Cl⁻ co-transporter is one of the key renal factors upregulated in acclimation to low-salinity environment

Most cartilaginous fishes are principally marine species, and only a limited number of species, including the bull shark, Carcharhinus leucas, have the capacity to inhabit both seawater (SW) and freshwater (FW) environments (see Compagno, 1984; Hazon et al., 2003; Pillans and Franklin, 2004). Bull sharks spend most of their lifetime in SW, but they utilize a wide range of salinities throughout their life cycle (Thorson, 1971; Montoya and Thorson, 1982; Simpfendorfer et al., 2005; Carlson et al., 2010). Female bull sharks are thought to give birth in estuaries or river mouths, and neonates and juveniles are thought to migrate upstream to occupy inshore rivers as nursery habitats (Ortega et al., 2009; Werry et al., 2012). The use of FW habitats by juvenile bull sharks has been reported in the southern USA (Simpfendorfer et al., 2005; Ortega et al., 2009; Matich and Heithaus, 2015), central America (Thorson, 1971; Montoya and Thorson, 1982), Australia (Pillans and Franklin, 2004; Werry et al., 2012) and South Africa (Cliff and Dudley, 1991; McCord and Lamberth, 2009). Once bull sharks reach sexual maturity [approximately 160–180 cm in total length (TL) in Florida], they move from riverine and estuarine environments into full-strength SW for further growth and breeding (Curtis et al., 2011).

It is well recognized that marine cartilaginous fishes, including SW-adapted bull sharks, conduct urea-based osmoregulation. Plasma ion levels are maintained at approximately half that of SW, whereas total plasma osmolality is slightly hyperosmotic to the surrounding SW environment through the retention of urea. As a result, cartilaginous fish do not suffer dehydration in a SW environment (Smith, 1936; Robertson, 1975; Hazon et al., 2003; Hyodo et al., 2004b). In FW environments, bull sharks and other euryhaline elasmobranch species maintain high internal NaCl and urea levels, resulting in a plasma osmolality of approximately 600 mOsm (Janech and Piermarini, 2002; Pillans and Franklin, 2004) that is almost twice that of FW teleosts (Evans, 1993). This strategy is in contrast with FW stingrays (Potamotrygonidae), which, like FW teleosts, do not accumulate urea (Wood et al., 2002). Therefore, it is likely that bull sharks face incipient water gain and solute loss in a FW environment by branchial diffusion and urinary loss. Recently, the contribution of Na⁺/K⁺-ATPase (NKA) and Na⁺/H⁺ exchanger 3 was reported for branchial Na⁺ uptake in bull sharks in a FW environment (Reilly et al., 2011). However, mechanisms enabling the euryhalinity in cartilaginous fish are still largely unknown.

It is generally accepted that multiple organs are involved in maintaining body–fluid homeostasis in vertebrates. Among them, the kidney is one of the most important osmoregulatory organs in vertebrates including cartilaginous fishes. In a SW environment, more than 90% of filtered urea is reabsorbed from the primary urine and returned to the blood system, thereby reducing the urinary loss of urea (Smith, 1936; Kempton, 1953; Boylan, 1967). Reflecting their crucial function, the renal tubules of marine cartilaginous fish are highly elaborate and show unique features compared with those of other vertebrates (Lacy and Reale, 1985; Hentschel et al., 1998; Hyodo et al., 2014). We have conducted mapping of various pumps, channels and transporters in the nephron segments, and proposed a model for urea reabsorption in the kidney of SW-adapted cartilaginous fishes (Hyodo et al., 2004a, 2014; Yamaguchi et al., 2009; Kakumura et al., 2015; Hasegawa et al., 2016). In FW environments, in contrast, it is assumed that the kidney of the bull shark must excrete excess water caused by the osmotic influx of water, while concomitantly retaining ions and urea. Therefore, it is highly probable that the expression of key proteins involved in osmoregulation is altered in the kidney when bull sharks move from SW to FW habitats. Pillans et al. (2005) reported that the expression level of NKA decreased in the kidney of bull sharks following transfer from FW to SW, supporting our notion that mechanisms for retaining ions and urea are upregulated in a FW environment.

In the present study, we performed a transfer experiment of captive bull sharks from SW to FW and conducted a comprehensive search for transcripts exhibiting significant increases in expression levels in FW-acclimated bull sharks by RNA-sequencing (RNA-seq). The differential gene expression profiling with RNA-seq, and subsequent quantitative PCR and in situ hybridization analyses revealed that the expression of Na⁺-Cl⁻ cotransporter (NCC) was substantially upregulated in the late distal tubule (the fourth loop) and the collecting tubule (the final segment) by transfer of bull sharks from SW to FW. In contrast, similar changes in distribution and expression level of NCC were not observed in the houndshark, Triakis scyllium, when it was acclimated to 30% SW; this species cannot survive in the FW environment. Our findings indicate that the activation of NaCl reabsorption associated with NCC is one of the crucial mechanisms that contributes to the euryhalinity of the bull shark.

The bull sharks, Carcharhinus leucas (Müller & Henle 1839), used in this study [n=7, TL=199.9±16.0 cm, body mass (BM)=94.1±20.4 kg] were caught by set nets or reared in the Okinawa Churaumi Aquarium. They were kept in holding tanks filled with SW (20–23°C) in the aquarium under a constant photoperiod (12 h:12 h light:dark) and fed chopped fish. During the experimental period, they were kept without feeding.

Japanese banded houndsharks, Triakis scyllium Müller & Henle 1839 (n=9, TL=68.1±1.0 cm, BM=1.0±0.05 kg), were collected in Koajiro Bay, near the Misaki Marine Biological Station of the University of Tokyo. They were transported to the Atmosphere and Ocean Research Institute at the University of Tokyo and kept in 2×10³ liter holding tanks (20–22°C) under a constant photoperiod (12 h:12 h light:dark). The houndsharks were fed on squid for at least 2 weeks before the experiments and were kept without feeding during the experimental period (see Yamaguchi et al., 2009).

All animal experiments were conducted according to the Guidelines for Care and Use of Animals approved by the committees of the University of Tokyo and the Okinawa Churaumi Aquarium.

In March of 2012, three bull sharks kept in holding tanks were moved to an experimental tank (15×10³ liters, filled with full-strength SW, 35 ppt) 3 days before the onset of salinity change. It was ascertained that they could swim comfortably in the new tank. On day 1 of the experiment, the salinity was adjusted by adding FW to achieve a salinity of 80% SW. On day 2, the same degree of salinity change was performed to produce 60% SW. On days 3–8, the salinity of the tank was reduced by 10% per day, and the salinity was deemed to have reached nearly FW (less than 2 ppt) on day 8 (Fig. S1). Bull sharks were killed on days 8–10. Two SW control sharks were maintained in the holding tank with running SW, and were killed on days 8 and 9, and two further SW sharks were killed in November 2011 and February 2013, respectively. For sampling, bull sharks were initially anesthetized with an intramuscular injection of midazolam (2.5 mg kg⁻¹ BM) and then euthanized with an intravenous injection of 2,6-diisoprophylphenol (0.6–1.6 mg kg⁻¹ BM). Blood samples were collected from the dorsal vasculature by a syringe and were centrifuged at 10,000 g at 4°C for 10 min to obtain the plasma. Urine samples were collected using a urethral catheter that was inserted under anesthesia (All Silicone Foley Balloon Catheter, Create Medic Co. Ltd, Kanagawa, Japan). Plasma and urine samples were stored at −80°C until analysis. Ion and urea concentrations and the osmolality of plasma and urine were measured by Rapid Labo 1265 (Siemens Healthineers, Erlangen, Germany). Kidneys were dissected out and separated into the left and right halves. One half was frozen quickly in liquid nitrogen, while the other half was trimmed transversely to 1 cm slices and fixed in Bouin's solution without acetic acid (saturated picric acid:formalin=3:1) or 4% paraformaldehyde in 0.1 mol l⁻¹ phosphate buffer (pH 7.4) containing 150 mmol l⁻¹ NaCl and 350 mmol l⁻¹ urea at 4°C for 2 days. Other tissues (hypothalamus, pituitary, gill, ventricle, atrium, liver, muscle, stomach, intestine, rectum and rectal gland) were also collected and were frozen or fixed as described above.

With regard to the houndshark, frozen and fixed tissues collected in the previously conducted transfer experiment in 2009 were used. In brief, houndsharks were kept in two tanks (1–2×10³ liters, filled with full-strength SW, 34 ppt). On day 1, one tank was adjusted to a salinity of 80% SW by adding FW. On days 2 and 3, the same degree of salinity change was performed to produce 40% SW. FW was added on day 4 to produce a final salinity of 30% SW (10 ppt, n=4). Control houndsharks (n=5) were kept in full-strength SW during the dilution protocol. Houndsharks were maintained for 1 week in each salinity and then euthanized and sampled. Sampling procedures were as described in detail previously (Yamaguchi et al., 2009).

Total RNA was extracted from the frozen kidneys of SW (sharks no. 2 and 3) and FW-acclimated (sharks no. 5 and 6) bull sharks using the guanidium thiocyanate–phenol–chloroform mixture (ISOGEN, Nippon Gene, Toyama, Japan). The concentration and quality of the extracted RNA were assessed using a Qubit 2.0 Fluorometer (Thermo Fisher Scientific, Waltham, MA, USA) and an Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA, USA), respectively. Non-stranded libraries were then constructed using a TruSeq total RNA sample preparation kit (Illumina, San Diego, CA, USA). Sequencing was performed on a HiSeq 1500 (Illumina) to obtain 127 bp pair-end reads, as documented previously (Tatsumi et al., 2015). The adaptor sequences and low-quality reads were discarded using Trim Galore! v0.3.1 (https://www.bioinformatics.babraham.ac.uk/projects/trim_galore/) and a fastq quality filter v0.10.0 (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/).

The trimmed reads were subjected to de novo assembly to obtain transcript contigs for each shark using Trinity version r20140717 (Grabherr et al., 2011). The contigs from each individual shark were then clustered using CD-HIT EST v4.6.1 (Fu et al., 2012) and TGICL v0.0.1 (Pertea et al., 2003) to produce the consensus contigs. The read trimming, transcriptome assembly and merging of contigs were performed according to a previous study (Hara et al., 2015). From these consensus contigs, open reading frames (ORFs) and the deduced amino acid sequences were predicted using Transdecorder version r20140704 (Haas et al., 2013). The contigs were annotated using Trinotate version r20150708 (Bryant et al., 2017), which provides the homology information of known proteins and functional domains, as well as ab initio prediction of transmembrane and signal peptides. The trimmed sequence reads were then mapped to the contigs using bowtie2 v2.1.0 (Langmead and Salzberg, 2012) and the expressions were quantified using eXpress v1.5.1 (Cappé and Moulines, 2009). The differential expression analysis between SW and FW individuals was performed using edgeR v3.0.0 (Robinson et al., 2010; McCarthy et al., 2012). The sequencing information and assembly statistics are shown in Table S1. The accession numbers for the short-read data of RNA-seq analysis are: DRX162067, DRX162068, DRX162069 and DRX162070.

Complementary DNA cloning was performed as previously described in detail (Hasegawa et al., 2016). Total RNA was extracted from the kidney with ISOGEN. Two micrograms of total RNA was treated with DNase using a TURBO DNA-free kit (Life Technologies, Carlsbad, CA, USA) and reverse-transcribed to first-strand cDNA using a high-capacity cDNA reverse transcription kit (Life Technologies), following the manufacturer's instructions. Primers for cDNA cloning were designed to amplify entire ORFs based on the contig sequence data from the RNA-seq (Table S2). The target cDNAs were amplified with Kapa Taq Extra DNA polymerase (Kapa Biosystems, Boston, MA, USA), electrophoresed on a 1.2% agarose gel, excised and purified using the UltraClean 15 DNA Purification Kit (MO BIO Laboratories, Carlsbad, CA, USA). The purified fragments were ligated into a pGEM T-easy plasmid (Promega, Madison, WI, USA), and the nucleotide sequence was determined using an automated DNA sequencer (3130xl Genetic Analyzer; Life Technologies). The accession numbers for the partial sequences of bull shark Na⁺/K⁺-ATPase alpha subunit 1 (NKAα1), Na⁺-K⁺-Cl⁻ cotransporter-2 (NKCC2), Na⁺-Cl⁻ cotransporter (NCC), urea transporter (UT) and elongation factor 1 alpha subunit 1 (EF1α1) are LC462277, LC462278, AB769491, LC462279 and LC462280, respectively. The accession numbers for the partial sequences of houndshark NKAα1, NKCC2, NCC, UT and EF1α1 mRNA are AB669491, AB769486, AB769487, AB094993 and LC462281, respectively.

The expression levels of mRNAs were examined by real-time quantitative PCR (qPCR) using a 7900HT Sequence Detection System (Applied Biosystems) with a KAPA SYBR Fast qPCR kit (Kapa Biosystems), as previously described in detail (Hasegawa et al., 2016). The plasmids containing cloned cDNAs were serially diluted and were used as the known amounts of standard cDNAs for absolute quantification in qPCR analyses. The copy numbers of the standard cDNAs were calculated using BioMath Calculators (https://www.promega.co.uk/resources/tools/biomath-calculators/). Primer sets for qPCR were designed using Primer Express software or PrimerQuest (https://sg.idtdna.com/pages) (Table S2). As an internal control, EF1α1 mRNA was used. First-strand cDNAs were prepared from 11 tissues (hypothalamus, pituitary, gill, liver, muscle, ventricle, atrium, stomach, intestine, rectum and rectal gland) as described above, and were used for a tissue distribution analysis.

Kidneys fixed in 4% paraformaldehyde solution were embedded in Paraplast (McCormick Scientific, Richmond, IL, USA). Serial sections were cut at 7 µm thickness and mounted onto MAS-coated slides (Matsunami Glass, Osaka, Japan). The cDNA fragments of bull shark NKAα1 (765 bp), NKCC2 (845 bp), NCC (1079 bp) and UT (664 bp), and houndshark NKAα1 (859 bp), NKCC2 (936 bp), NCC (1080 bp) and UT (972 bp) were amplified using gene-specific primers (Table S2), and then used to synthesize digoxigenin (DIG)-labeled antisense and sense cRNA probes (DIG RNA labeling kit; Roche Applied Science, Mannheim, Germany), following the manufacturer's protocols. Hybridization and washing were conducted using a previously described protocol (Takabe et al., 2012). Stained sections were counterstained with Nuclear Fast Red (Vector Laboratories, Burlingame, CA, USA). Micrographs were obtained using a virtual slide system (Toco; Claro, Aomori, Japan).

For morphological observations, kidney sections fixed in Bouin's solution were stained with periodic acid–Schiff (McManus, 1946). Briefly, deparaffinized sections were oxidized in 0.5% periodic acid solution (Wako, Osaka, Japan) for 5 min. After washing in tap water and distilled water, sections were placed in Schiff's reagent (Wako) for 15 min, and then rinsed in sulfurous acid three times. Sections were then counterstained with Mayer's hematoxylin (Wako).

A peptide C+GFEDEAIVKELRKD from bull shark NCC was synthesized and coupled via cysteine to keyhole limpet hemocyanin. The conjugated peptide was emulsified with complete Freund's adjuvant and injected into a rat for immunization (Eurofins Genomics, Tokyo, Japan).

Kidneys fixed with Bouin's solution without acetic acid were embedded in Paraplast. Serial sections were cut at 7 µm thickness and mounted onto MAS-coated slides (Matsunami Glass). Rehydrated tissue sections were treated with antigen activator (Histo VT One; Nacalai Tesque, Kyoto, Japan) for 20 min at 90°C. Immunohistochemical staining for NCC was performed with the avidin-biotin-peroxidase complex kit (Vector Laboratories), as described previously (Hasegawa et al., 2016). NCC antiserum was diluted with 2% normal goat serum in PBS (pH 7.4) (1:90,000). The specificity of immunoreactive signals for NCC was confirmed by preabsorption of antibody with the synthetic antigen for 24 h at 4°C prior to incubation. Stained sections were counterstained with Mayer's hematoxylin (Wako). Micrographs were obtained using a digital camera (DSRi1; Nikon, Tokyo, Japan). For double-labeling fluorescence immunohistochemistry, the NCC antiserum (1:30,000) and anti-NKAα subunit antibody (1:2000; immunized in a rabbit; a gift from Prof. Kaneko, University of Tokyo) were mixed and incubated with deparaffinized sections for 48 h at 4°C. Sections were washed in PBS, and then incubated with fluorescein-labeled anti-rat IgG antibody (Alexa Fluor 555) and anti-rabbit IgG antibody (Alexa Fluor 488) (Thermo Fisher Scientific) for 2 h at room temperature. Sections were mounted with ProLong Gold Antifade Reagent (Thermo Fisher Scientific). Micrographs were obtained using a fluorescence microscope (BX53; Olympus, Tokyo, Japan).

Values are presented as means±s.e.m. and were compared using Student's t-test or the Mann–Whitney U-test, with the assumption of normality checked by the Shapiro–Wilk test. P-values <0.05 were considered statistically significant. All analyses were performed using KyPlot 5.0 software (Kyenslab, Tokyo, Japan).

The plasma and urine measurement of bull sharks kept in SW and acclimated to FW are summarized in Table 1. Plasma osmolality, ion and urea concentrations were decreased after the transfer to FW, and the changes in osmolality, Cl⁻ and urea were statistically significant. However, as previously reported (Pillans and Franklin, 2004), bull sharks maintained plasma levels of ion and urea concentrations at approximately 600 mOsm kg⁻¹ in the FW environment. Although we could not investigate the time course changes in the plasma parameters during the transfer experiment, osmolality and Na⁺ concentration of environmental water reached levels below those of FW-acclimated bull shark plasma on days 2 and 4, respectively (Fig. S1). Therefore, bull sharks were exposed to hypo-osmotic environmental conditions for a period of 1 week before sampling. Urine parameters, except for urea, were also decreased in the FW environment (Table 1).

In the present study, we could not measure glomerular filtration rate (GFR) and urine flow rate (UFR) of SW- and FW-acclimated bull sharks. In the Atlantic stingray, the transfer from ambient SW to 50% SW for 2 h increased GFR and UFR 3-fold and 9-fold, respectively (Janech et al., 2006). Therefore, in Table 2, the values for GFR and UFR reported in the Atlantic stingray were used to calculate the hypothetical absolute amount of reabsorbed NaCl in the bull shark kidney. The estimated values of NaCl and urea reabsorption in SW- and FW-acclimated bull shark kidneys were calculated from the plasma and urine concentrations of NaCl and urea in bull sharks and from the stingray GFR and UFR values in ambient SW (used for SW-acclimated bull sharks) and 50% SW (used for FW-acclimated bull sharks). The estimated values in Table 2 indicate that the transfer of bull sharks from SW to FW caused an approximately 50% increase in reabsorption of NaCl from the glomerular filtrate.

The plasma parameters of houndsharks after transfer to 30% SW (10 ppt) were reported previously (Yamaguchi et al., 2009). Plasma osmolality and the concentration of urea and ions were decreased in the 30% SW group. Our preliminary experiment showed that Japanese banded houndshark cannot survive when environmental salinity was lowered to less than 25% SW (8 ppt) (see Yamaguchi et al., 2009).

Renal mRNA expression profiles were examined in two SW and two FW bull sharks by RNA-seq to search for candidate genes exhibiting a significant difference in expression levels following the FW acclimation (see Materials and Methods). The clustering of the overall expression profiles showed a clear distinction between the two SW individuals and the two FW-acclimated individuals (Fig. S2), indicating that acclimation to the FW environment triggered changes in renal function at the molecular level. The application of dual cut-offs [false discovery rate (FDR)<0.05 and log₂ fold change (logFC)>1.5 of the fragment per kilobase of exon per million (FPKM)] identified 138 genes upregulated in the FW-acclimated group. Among them, six genes encoding the solute carrier (Slc) superfamily proteins were found to be upregulated: Na⁺/K⁺/Ca²⁺ exchanger 5 (Slc24a5), carnitine transporter 2 (Slc22a5, OCNT2), Na⁺-coupled neutral amino acid transporter 8 (Slc38a8), Ca²⁺-binding mitochondrial carrier protein 2B (Slc25a25), Na⁺-Cl⁻ cotransporter (NCC, Slc12a3) and glucose transporter 12 (Slc2a12) (Table 3). The Slc superfamily comprises proteins that are responsible for osmolyte transport (Hediger et al., 2004). In our analysis, NCC stably showed the largest FPKM values in FW individuals (over 120 in both individuals) that were approximately 8-fold higher than that of SW individuals. Therefore, in the following investigations, we focused on NCC as a candidate gene important for FW acclimation in the bull shark.

Tissue distribution of NCC mRNA was examined with quantitative RT-PCR in four bull sharks kept in SW and three bull sharks acclimated in FW (Fig. S3). NCC mRNA was predominantly expressed in the kidney. Expression of NCC mRNA was significantly higher in the kidneys of FW-acclimated bull sharks than SW individuals (Fig. S3 and Fig. 1). Minute amounts of expression were detected in the hypothalamus, pituitary and gills, but no significant change was observed between bull sharks in SW and FW (Fig. S3).

To validate the results of the RNA-seq and examine the possible involvement of other genes, we performed quantitative RT-PCR for some selected genes. As mentioned above, the expression of NCC mRNA was 10-fold higher in the kidneys of bull sharks acclimated to FW compared with that in SW individuals (Fig. 1). In addition to NCC, we examined the mRNA levels of Na⁺/K⁺-ATPase alpha subunit 1 (NKAα1; Atp1a1; a driving force of NaCl reabsorption), Na⁺-K⁺-2Cl⁻ cotransporter 2 (NKCC2; Slc12a1; a protein important for NaCl reabsorption in the mammalian thick ascending limb) and urea transporter (UT; Slc14a2) in the kidneys of bull sharks, because of their involvement in osmoregulation. The expression of NKAα1 mRNA was significantly higher in the kidneys of FW-acclimated individuals compared with that of bull sharks kept in SW. In contrast, no significant change was observed in the expression levels of NKCC2 and UT mRNA following the transfer of bull sharks from SW to FW, although the mean value of UT mRNA levels in FW was 1.5 times greater than that in SW (Fig. 1).

In contrast, in the kidneys of houndsharks, no significant change was observed in the NKAα1, NKCC2, NCC and UT mRNA levels after transfer to diluted SW (30% SW) (Fig. 1). Furthermore, NCC mRNA levels (NCC mRNA/EF1α1 mRNA) in houndsharks in full-strength SW were approximately 40 times lower compared with those of bull sharks in SW. This resulted in approximately 500 times higher expression levels of the NCC mRNA/EF1α1 mRNA in FW-acclimated bull sharks compared with values in 30% SW-acclimated houndsharks.

Similar to the kidneys of other marine cartilaginous fishes (Lacy and Reale, 1985; Hentschel et al., 1998; Hyodo et al., 2004a; Kakumura et al., 2015), the bull shark kidney consists of multiple lobules. Each lobule is separated into two morphologically distinguishable regions, a sinus zone and a bundle zone (Fig. 2A). The sinus zone occupied a large area in lobules and was filled with blood sinuses. The sinus zone was composed of two types of tubules: the larger diameter tubules with apical brush border correspond to the proximal segment II of the second loop (open arrows in Fig. 2B), while the late distal tubules (LDT) of the fourth loop (filled arrows in Fig. 2B) were characterized by a relatively smaller diameter and lower epithelial cell height, and were without an apical brush-border membrane. In contrast, in the bundle zone, renal tubules were densely packed, and five tubular segments of a single nephron (the descending and ascending limbs of the first and the third loops, and the final segment) were enclosed in a peritubular sheath (Fig. 2C,D).

Fig. 3 shows mRNA signals of NKAα1, NKCC2, NCC and UT in the bull shark kidney. In SW individuals, signals for NKAα1 mRNA were observed in both the bundle and the sinus zones (Fig. 3A). In the magnified images, NKAα1 mRNA signals were localized in the smaller diameter LDT in the sinus zone (filled arrows in Fig. 3B) and in tubules with the largest diameter in the bundle zone (early distal tubule, EDT) (open arrowheads in Fig. 3C). Some tubules neighboring renal corpuscles also showed NKAα1 mRNA signals; these tubules correspond to the transitional segments from the EDT to the LDT (open arrow in Fig. 3B) and from the LDT to the collecting tubule (CT) (filled arrowhead in Fig. 3B). Weak signals for NKAα1 mRNA were observed in the CT in the bundle zone (filled arrowhead in Fig. 3C). Therefore, positive signals for NKAα1 mRNA were continuously observed from the EDT to the CT (Fig. 3i). A similar distribution of NKAα1 mRNA was observed in the kidneys of FW-acclimated bull sharks. However, the signals were stronger compared with those of SW individuals, especially in the LDT, and were also evident in the descending limb of the third loop (Fig. 3D–F,ii). Signals for NKCC2 mRNA were observed almost exclusively in tubules in the bundle zone of both SW- and FW-acclimated bull sharks (Fig. 3G–L). In the bundle zone, intense signals of NKCC2 mRNA were detected in the EDT (open arrowheads in Fig. 3I,L). In contrast, in the sinus zone, only a small number of tubules located in the vicinity of renal corpuscles were positive for NKCC2 mRNA (open arrows in Fig. 3H,K). Detailed observation using serial sections revealed that the NKCC2 mRNA-expressing tubules in the sinus zone are the transitional segment from the EDT to the LDT (Fig. 3iii,iv). No difference was observed in the distribution and signal intensity of NKCC2 mRNA between SW and FW sharks.

Signals for NCC mRNA were very weak in SW individuals (Fig. 3M–O). The signals were detected in the sinus zone, but not in the bundle zone. The morphological characteristics of the tubules showing positive signals revealed that these tubules are the LDT (filled arrows in Fig. 3N). Meanwhile, in FW-acclimated individuals, strong signals for NCC mRNA were observed in the LDT (Fig. 3P,Q). In addition, intense signals were also detected in the anterior half of the CT in the bundle zone (filled arrowheads in Fig. 3R, and schematically shown in Fig. 3vi). Signals for UT mRNA were detected almost exclusively in the CT in the bundle zone in both SW- and FW-acclimated bull sharks (Fig. 3S,V), and the signal intensity was higher in FW-acclimated bull sharks than in SW individuals (filled arrowheads in Fig. 3U,X). A small number of tubules located in the vicinity of renal corpuscles showed UT mRNA signals in the sinus zone, and these segments correspond to the transitional segment from the LDT to the CT (filled arrowheads in Fig. 3T,W). The intensity of UT mRNA signals in this segment was also higher in FW-acclimated bull sharks (Fig. 3vii,viii).

Co-expression of NKAα1, NKCC2, NCC and UT mRNA was examined using the serial sections of FW individuals (Fig. 4). In the sinus zone, NCC mRNA was co-expressed with NKAα1 mRNA in the LDT (filled arrows in Fig. 4A,E). The LDT did not express NKCC2 or UT mRNA (Fig. 4C,G). NKCC2 mRNA was moderately expressed in the transitional segment from the EDT to the LDT, where NKAα1 mRNA was also expressed (open arrows in Fig. 4A,C). In the bundle zone, NKCC2 mRNA was co-expressed with NKAα1 mRNA in the EDT (open arrowheads in Fig. 4B,D), while co-expression of NKAα1, NCC and UT mRNA was found in the CT (filled arrowheads in Fig. 4B,F,H).

In the kidneys of houndsharks kept in full-strength SW, intense positive signals for NKAα1 mRNA were observed in the EDT (open arrowheads in Fig. 5C). The LDT (fourth loop of nephron) also showed positive signals for NKAα1 mRNA (filled arrows in Fig. 5B). In the LDT, the transitional segment from the EDT to the LDT showed higher expression (open arrows in Fig. 5A,B), while signals in the remaining part of LDT were weak (Fig. 5i). No difference was observed in the localization and signal intensity of NKAα1 mRNA between SW individuals and 30% SW-acclimated individuals (Fig 5A–F,i,ii). The expression of NKCC2 mRNA was similar to that of NKAα1 mRNA (Fig. 5G–L). Intense signals for NKCC2 mRNA were detected in the EDT in the bundle zone (open arrowheads in Fig. 5I,L), while weak or moderate levels of NKCC2 mRNA expression were seen in the LDT (open and filled arrows in Fig. 5H,K). No difference was observed in localization and signal intensity of NKCC2 mRNA between 100% SW and 30% SW groups (Fig 5G–L,iii,iv).

NCC mRNA was expressed only in a limited region of the fourth loop, which was the transitional segment from the LDT to the CT (Fig. 5M and filled arrowheads in N, schematically shown in v). In 30% SW-acclimated houndsharks, signals for NCC mRNA were much weaker than that in SW individuals (filled arrowheads in Fig. 5Q). UT mRNA was expressed in the CT (Fig. 5S and filled arrowheads in U). In the sinus zone, intense signals for UT mRNA were detected in the transitional segment from the LDT to the CT (filled arrowheads in Fig. 5T), and the signal intensity of the transitional segment was decreased by the transfer of houndsharks from full-strength SW to 30% SW (filled arrowheads in Fig. 5W), similar to NCC mRNA signals.

Co-expression of the four mRNAs was examined using the serial sections of SW individuals (Fig. S4). NKCC2 mRNA was consistently expressed with NKAα1 mRNA in the EDT (open arrowheads in Fig. S4B,D) and the LDT (open arrows in Fig. S4A,C). In the sinus zone, NCC mRNA was co-expressed with UT mRNA in the transitional segment from the LDT to the CT, where NKAα1 mRNA was also expressed (filled arrows in Fig. S4A,E,G). In contrast to the results in bull sharks, only UT mRNA expression was found in the CT (filled arrowheads in Fig. S4B,D,F,H).

The intracellular localization of bull shark NCC was examined using the antiserum raised against a synthetic polypeptide of bull shark NCC. Consistent with the results of in situ hybridization, immunoreactive signals for NCC were observed in the LDT of the nephron in FW individuals (open arrows in Fig. 6A), but not in SW individuals (Fig. 6D). Immunoreaction for NCC was localized to the apical membrane. The use of preimmune serum (Fig. 6B) and preabsorption of the anti-NCC serum with the synthetic NCC polypeptide (Fig. 6C) resulted in the disappearance of the NCC immunoreactive signals in the LDT. Double-labeling fluorescence immunohistochemistry showed co-localization of apically located NCC and basolaterally located NKA in the tubular cells of the LDT (Fig. 6F).

It is well recognized that most cartilaginous fishes are principally marine species, and that the bull shark is a rare euryhaline species with the capacity to inhabit both SW and FW environments. To further our understanding of the underlying mechanisms of euryhalinity in bull sharks, we focused on kidney function in the present study, as the kidney is the only organ by which excess water in the body can be excreted in a hypo-osmotic environment. By means of RNA-seq analysis and the subsequent molecular and histochemical analyses, we found that NCC expressed in the LDT and CT is one of the key molecules contributing to the successful FW acclimation of the bull shark.

In the SW environment, bull sharks maintain their plasma slightly hyperosmotic to surrounding SW by accumulating urea in the body, which is similar to other marine cartilaginous fishes (see also Pillans and Franklin, 2004). Transfer to a low-salinity environment resulted in a decrease in plasma ion and urea levels. However, the FW-acclimated bull sharks still maintained high plasma Na⁺ (213.7±9.0 mmol l⁻¹) and Cl⁻ (177.0±3.5 mmol l⁻¹) concentrations, and the resulting plasma osmolality was 617.7±13.0 mOsm, as has been previously reported for this species (Pillans and Franklin, 2004; Pillans et al., 2006) and for other euryhaline cartilaginous fishes including the Atlantic stingray, Hypanus sabina (Piermarini and Evans, 1998). These osmolality data indicate that bull sharks exposed to severe hypo-osmotic conditions must be excreting large amounts of diluting urine to compensate for the osmotic influx of water from the environment. In accordance with this idea, a concomitant decrease in urine osmolality and NaCl concentrations was observed. In the present study, however, due to technical limitations, we could not measure the GFR and UFR of SW- and FW-acclimated bull sharks, and thus could not calculate exact values for water excretion and NaCl reabsorption in the kidney after FW transfer. We therefore used the values for GFR and UFR that were reported in Atlantic stingrays transferred from ambient SW to 50%-diluted SW (Janech et al., 2006) to calculate the hypothetical absolute amount of reabsorbed NaCl in the bull shark kidney (Table 2). The estimated values in Table 2 indicate that the transfer of bull sharks from SW to FW caused an approximately 50% increase in reabsorption of NaCl from the glomerular filtrate.

It should be noted that, because of the larger osmotic difference between internal and environmental osmolalities, bull sharks acclimated in FW are considered to receive a greater influx of water compared with Atlantic stingrays transferred to 50% SW. In other words, bull sharks in the FW environment likely filter more plasma and excrete more urine compared with stingrays. The calculated values in Table 2 indicate a 1.5 times increase in NaCl reabsorption, which is lower than the value in stingrays acutely transferred to 50% SW (2.5 times increase) (Janech et al., 2006), but a greater increase in NaCl reabsorption is assumed in FW-acclimated bull sharks. Therefore, we searched for upregulated mRNAs encoding solute carrier superfamily proteins involved in NaCl reabsorption.

We previously identified the expression of NKAα1 and NKCC2 in the kidney of another cartilaginous fish, Callorhinchus milli (Kakumura et al., 2015). NKCC2 (Slc12a1) is well known to be localized in the thick ascending limb of mammalian Henle's loop together with NKA, where NaCl is actively reabsorbed for dilution of primary urine (Fenton and Knepper, 2007). In the elephant fish kidney, NKAα1 and NKCC2 were colocalized in the EDT and the posterior half of the LDT (Kakumura et al., 2015), and the NaCl uptake by NKCC2 is considered to be important for the urea reabsorption process (Hyodo et al., 2014). We thus initially focused on NKCC2 as a candidate molecule important for NaCl reabsorption in FW-acclimated bull sharks. However, no significant change was observed in the distribution and expression levels of NKCC2 mRNA following the transfer of bull sharks to a FW environment. Therefore, we performed RNA-seq analysis to search for other candidate genes contributing to NaCl reabsorption, and upregulated by FW acclimation.

In the RNA-seq analysis, a remarkable increase was detected in the expression of NCC mRNA. Consistent with the results of qPCR, only a modest level of NCC mRNA signal was found in the LDT of bull shark kidney in SW, whereas the intensity of hybridization signals for NCC mRNA considerably increased in the LDT of FW-acclimated bull shark. NCC is known to be expressed in the apical membrane of the mammalian distal convoluted tubule (Fenton and Knepper, 2007) and in the distal segment of the teleost fish kidney (Kato et al., 2011; Teranishi et al., 2013). Localization on the apical membrane was confirmed also in the LDT of bull sharks. In the LDT, NCC is co-expressed with NKAα1, which is well recognized as the generator for the Na⁺ electrochemical gradient that is the driving force for NaCl reabsorption via NCC (Fenton and Knepper, 2007). The upregulation in the NKAα1 mRNA signal was also observed in the LDT of FW-acclimated bull sharks. These results strongly suggest that the LDT contributes significantly to NaCl retention in the FW environment, and that NCC is a key molecule for NaCl reabsorption in the LDT (Fig. 7).

In cartilaginous fish, we initially identified NCC mRNA expression in the gills and kidney of the Japanese houndshark, T. scyllium (Takabe et al., 2016). The present qPCR analysis revealed that, in the kidney of houndsharks, expression levels of NCC mRNA were considerably lower compared with those in the kidney of FW-acclimated bull sharks, and no significant change was observed in NCC mRNA levels between SW- and 30% SW-acclimated houndsharks. In situ hybridization further demonstrated that NCC mRNA was expressed only in the limited region of the LDT in the houndshark kidney. In contrast to the kidney, NCC mRNA levels and the number of NCC-expressing cells in the gill increased in houndsharks after the transfer from SW to 30% SW (Takabe et al., 2016). In houndsharks, NCC is more likely to be important in the gills for absorption of NaCl from the environment. Although the timelines of transfer experiments were not exactly the same between bull sharks and houndsharks, our results revealed that the significant upregulation of NCC mRNA expression in the LDT was unique to the bull shark, further implying that the NaCl reabsorption in the LDT is critical for the euryhalinity of this species.

In addition to the LDT, upregulation of NCC mRNA expression was also found in the CT. Expression of NKAα1 mRNA was also observed, implying that NaCl is reabsorbed in the CT of FW-acclimated bull sharks. In contrast, neither NKAα1 nor NCC mRNA signals were detected in the CT of houndshark kidney. Therefore, NaCl reabsorption in the CT is another feature of the euryhaline bull shark kidney. The expansion of tubular segments contributing to NaCl reabsorption, in comparison to the kidneys of SW-acclimated bull sharks and houndsharks in SW and 30% SW, is most likely important for highly effective NaCl retention in FW environments (Fig. 7).

Alternatively, NaCl reabsorption in the CT may contribute to another important function, that is, urea retention. Based on the exclusive localization of UT, the CT has been implicated as the urea-reabsorbing segment in elephant fish and houndsharks (Hyodo et al., 2004a, 2014; Yamaguchi et al., 2009; Kakumura et al., 2015). The limited expression of UT in the CT has also been demonstrated in the bull shark kidney, implying that urea reabsorption is a common function of the CT in both marine and euryhaline cartilaginous fishes. We have proposed a model for urea reabsorption in the tubular bundle, in which the first and the third loops and the CT are wrapped with an impermeable peritubular sheath (Hyodo et al., 2014). The first step of this model is massive reabsorption of NaCl. The resulting increase in osmolality causes reabsorption of water, which produces a low urea environment in the tubular bundle. Urea is then reabsorbed from the CT via UT using the concentration gradient of urea as a driving force (Hyodo et al., 2014). In this model, NKA and NKCC2 expressed in the third loop (the EDT) contribute to the NaCl reabsorption. Expression of NKAα1 and NCC mRNA in the CT of FW-acclimated bull shark most probably enhances NaCl uptake in the tubular bundle, which in turn enhances water uptake, and finally urea reabsorption. When we calculated the amount of reabsorbed urea in bull sharks using the GFR and UFR data of the Atlantic stingray, the amount in the FW-acclimated bull shark kidney was approximately two-thirds that of bull sharks in SW (Table 2). However, again, it would be reasonable to consider that bull sharks in FW have greater GFR and UFR values compared with Atlantic stingrays in 50% SW, in order to excrete excess water in the body. If this is the case, bull sharks in FW environments reabsorb more urea than in the SW environment in order to maintain plasma urea levels (Fig. 7).

Currently, the function of the LDT in SW environment remains to be clarified. In elephant fish and houndsharks, NKCC2 mRNA is expressed in the LDT (present study and Kakumura et al., 2015), and the LDT is considered to be important for concentrating urea for subsequent urea reabsorption in the CT (Hyodo et al., 2014). However, in the bull shark kidney in a SW environment, neither NKCC2 nor NCC was observed in the LDT (Fig. 7). Further studies are needed to understand the function of LDT in the kidney of bull shark in the SW environment by identifying transporters that are expressed in the LDT.

In summary, we show for the first time that NCC expressed in the LDT and CT of the kidney is a key molecule for NaCl retention required for euryhalinity of bull sharks. Our comprehensive approach with RNA-seq effectively pinpointed several candidate genes, of which we focused on NCC. The remaining genes with significant changes in expression profiles will be characterized in further studies to ascertain their role in renal function in the euryhaline bull shark. In addition, other organs including the gill (Reilly et al., 2011), rectal gland (Pillans et al., 2005) and liver (Anderson et al., 2005) are thought to be involved in euryhaline mechanisms of bull sharks; RNA-seq analysis of bull shark gills is now underway. Comprehensive studies on these osmoregulatory organs will uncover a more complete picture of the mechanisms for euryhalinity in this unique elasmobranch, the bull shark.

We thank Drs Y. Yamaguchi, K. Hasegawa, M. Inokuchi and K. Sato for their invaluable support and discussion, the rest of the Laboratory for Phyloinformatics, RIKEN BDR, for their support in RNA-seq data production, Mr. Nomura and Ms. Y. Nakamatsu for their support on the transfer experiment of bull shark, and Dr A. Kato of the Tokyo Institute of Technology for the use of a virtual slide system. We also thank Prof C. A. Loretz of State University of New York at Buffalo and Prof. J. A. Donald of Deakin University for critical comments on the manuscript. The anti-NKA antiserum was provided courtesy of Prof T. Kaneko of the University of Tokyo.