Animals

Female red stingrays, Hemitrygon akajei (Müller & Henle, 1841) [n = 6, average disc width = 58 ± 3.2 cm, average body weight (BW) = 7.8 ± 1.2 kg] were caught in February 2020 in a bay at Ushimado and transported to Ushimado Marine Institute, Okayama University for plasma and urine collection. They were kept in holding tanks filled with running natural SW (34‰) under natural conditions, and were fed with shrimp at least three times a week. For tissue sampling, three more stingrays [average disc width = 40 ± 6.2 cm, average BW = 2.8 ± 1.2 kg] were caught in January and May 2018 in the same location and transported to the Atmosphere and Ocean Research Institute (AORI), The University of Tokyo: two females for RNAseq and one male for three-dimensional reconstruction. They were kept in a 500 L holding tank at 20°C under a constant photoperiod (12 L:12 D) for 3 weeks without feeding.

A female Japanese banded houndshark, Triakis scyllium (Müller & Henle, 1839) [fork length (FL) = 66.6 cm, BW = 2.3 kg], was collected in Koajiro Bay, near the Misaki Marine Biological Station of The University of Tokyo. It was transported to AORI and kept in a 2,000 L holding tank at 21°C under a constant photoperiod (12 L:12 D). Female cloudy catsharks, Scyliorhinus torazame (Tanaka, 1908) [n = 6, average FL = 39.7 ± 0.7 cm, average BW = 398.0 ± 21.2 g] were transported from the Ibaraki Prefectural Oarai Aquarium to AORI and were kept in a 1000 L or 3000 L holding tank filled with natural SW (37‰) at 16°C under a constant photoperiod (12 L:12 D). Catshark rather than houndshark was used for urine collection because of the smaller body size and less active character. The houndshark and catsharks were fed with diced squid and sardine twice a week. 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 Okayama University.

Sampling of tissue, plasma, and urine

For tissue sampling, the fish were anesthetized with 0.02% (w/v) ethyl 3-aminobenzoate methanesulfonate (Sigma-Aldrich, MO, USA) buffered with an equal amount of sodium bicarbonate (Sigma-Aldrich). After the whole (left and right halves) kidney was dissected out, one half was immediately frozen in liquid nitrogen and kept at –80°C until RNA extraction, while the other half was fixed in Bouin's solution without acetic acid (saturated picric acid: formalin = 3:1) for 2 days and then stored in 70% ethanol at 4°C until use.

For plasma and urine sampling, the stingrays and catsharks were anesthetized as described above. Blood samples were collected from the dorsal vasculature by a syringe with 0.1% (w/v) EDTA-2K as an anticoagulant and were centrifuged at 10,000 × g at 4°C for 10 min to obtain the plasma. Only females were used for collecting urine, since Kempton (1953) pointed out the possibility of contaminations of urine by seminal fluid in male individuals. Sexually mature stingray and catshark possess a recognizable opening of the urinary tract, which protrudes to the outside. By using a polyethylene cannula (SP-45, Natsume Seisakusyo, Tokyo, Japan) attached to a collecting device ( Supplementary Figure S1 (zs200038.s1.pdf)), samples of excreted urine were collected from free-swimming animals without contamination by SW or any other body fluids. Sodium and chloride ion concentrations, and osmolality were measured with an atomic absorption spectrophotometer (Z5300, Hitachi, Tokyo, Japan), a digital chloridometer (C-50AP, Jokoh, Japan), and a vapor pressure osmometer (5520/5600, Wescor, UT, USA), respectively. Urea concentration was measured by the colorimetric assay according to Rahmatullah and Boyde (1980).

Library preparation for RNA-seq and sequencing

Total RNA extracted from the frozen kidney of stingray with ISOGEN (Nippon Gene, Toyama, Japan) was treated with DNase I (Thermo Fisher Scientific, MA, USA) to digest genomic DNA and was purified using the Zymo RNA Clean & Concentrator column (Zymo Research, CA, USA). Concentration and length distribution of the RNA was checked with the Qubit RNA HS Assay Kit on a Qubit 2.0 Fluorometer (Thermo Fisher Scientific) and the RNA 6000 Nano Kit on a 2100 Bioanalyzer (Agilent Technologies, CA, USA). Sequencing libraries were prepared using 1 µg of the DNase I-treated total RNA with the TruSeq Stranded mRNA Sample Prep Kit (Illumina, CA, USA) and the TruSeq single-index adaptor (Illumina). RNA fragmentation was performed at 94°C for a shortened duration of 4 minutes, compared to the standard condition (94°C for 8 minutes) to generate libraries with longer insert size distribution (Hara et al., 2015). The optimal numbers of PCR cycles for the libraries were pre-determined to be 7 cycles using the KAPA Real-Time Library Amplification Kit (Roche Applied Science, Mannheim, Germany) as described previously (Kadota et al., 2017). Libraries were sequenced with the Rapid Run mode of HiSeq1500 (Illumina) using the HiSeq PE Rapid Cluster Kit ver2 (Illumina) and the HiSeq Rapid SBS Kit v2-HS (Illumina), and paired-end reads of 127 bases were produced. Base-calling was performed with RTA ver1.18.64, and the fastq files were generated with bcl2fastq ver1.8.4 (Illumina). The raw sequenced reads were deposited in the DNA Data Bank of Japan (DDBJ) under the accession number DRA009804.

cDNA cloning and RNA probe synthesis

Total RNA was extracted from the frozen kidney of stingray with ISOGEN (Nippon Gene). Two micrograms of total RNA were treated with DNase using a TURBO DNA-free kit (Thermo Fisher Scientific) and reverse-transcribed to first-strand cDNA using a high-capacity cDNA reverse transcription kit (Thermo Fisher Scientific), according to the manufacturer's protocols. Primer sets were designed to amplify cDNAs encoding NKAα1 (837 bp; GenBank accession no. LC530055), NKCC2 (768 bp; LC530056), and UT (1004 bp; LC530057) based on the contig sequence data from the in-house transcriptome database as mentioned above ( Supplementary Table S1 (zs200038.s1.pdf)). PCR was performed with KAPA Taq Extra (Kapa Biosystems, MA, USA) by using the kidney cDNA as template. The amplified products were electrophoresed on agarose gels, excised and ligated into pGEM-T Easy Vector (Promega, WI, USA). The nucleotide sequence was confirmed by an automated DNA sequencer (ABI PRISM 3100, Applied Biosystems, CA, USA). For digoxigenin (DIG)labeled RNA probe synthesis, a DNA fragment was amplified again from the sequenced plasmid by PCR with Primestar GXL (Takara Bio, Shiga, Japan) using the M13 Forward and Reverse primers and subsequently purified with Wizard SV Gel and PCR Clean-Up System (Promega). Antisense and sense RNA probes were then synthesized from the purified template using a DIG RNA labeling kit (Roche Applied Science), following the manufacturer's protocols.

In situ hybridization and morphological observation

The fixed kidney was embedded in Paraplast (McCormick Scientific, IL, USA). Serial sections were cut at 7 µm thickness and mounted onto MAS-GP-coated slides (Matsunami Glass, Osaka, Japan). Hybridization and washing were conducted according to a protocol previously described by Takabe et al. (2012). Stained sections were counterstained with Kernechtrot Stain Solution (MUTO PURE CHEMICALS, Tokyo, Japan). For morphological observations, kidney sections were stained with periodic acid–Schiff (PAS; McManus, 1946). Briefly, deparaffinized sections were oxidized in 0.5% periodic acid solution (Wako) 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 with wash solution (10% sulfurous acid:1N hydrochloric acid:distilled water = 24 mL:20 mL:400 mL) three times. Sections were then counterstained with Mayer's hematoxylin (Wako). Stained sections were mounted with Permount (Thermo Fisher Scientific). Micrographs were obtained using a virtual slide system (BZ-X800 and accompanying software, Keyence, Osaka, Japan) and a digital camera (DXM1200; Nikon, Tokyo, Japan).

Three-dimensional reconstruction of stained serial sections

To reconstruct the nephrons of stingray, serial paraffin sections (480 sections in total) were divided into four series (A to D). As illustrated in  Supplementary Figure S2 (zs200038.s1.pdf), the first two sections were mounted on a slide A1, and the next two sections were mounted on a slide B1. Finally, six sections were mounted on a slide, and 80 slides (20 slides for each of series A, B, C and D) were prepared in total. Slides of the series A were treated with PAS-hematoxylin staining, and slides of the series B, C and D were stained with in situ hybridization of NKAα1, NKCC2 and UT mRNAs, respectively. This enabled us to conduct reconstruction of nephron structure and mapping of transporter expressions simultaneously. For the houndshark nephron, serial sections (300 sections in total) were mounted onto one series of slides and stained by PAS-hematoxylin only, since the mapping of NKAα1, NKCC2 and UT was previously described (Hyodo et al., 2004; Imaseki et al., 2019). To obtain the three-dimensional (3D) images of a single nephron, each segment was highlighted by different colors using GIMP 2.8. The highlighted regions were then aligned and stacked by using StackReg and TurboReg plugins of ImageJ (Abràmoff et al., 2004). Finally, 3D images of the stacked sections were generated using the 3D Viewer plugin of ImageJ.

Statistical analysis

Values are presented as means ± s.e.m. Plasma parameters of stingray were largely different from those of catshark, most likely due to the difference in the baselines of the surrounding SW. Therefore, in this study, the original values of each parameter were not compared statistically. Comparisons of urine/plasma (U/P) ratio of each parameter between the stingray and catshark were performed with the Mann-Whitney U test. All P values were two-tailed, and P < 0.05 was considered statistically significant. Statistical analysis was performed by using KyPlot 6.0 software (Kyenslab, Tokyo, Japan).

Urea and ionic composition of plasma and urine in two elasmobranchs

Table 1 shows the concentrations of Na⁺ and Cl^–, and urea, and osmolality in the plasma and urine of stingray and catshark. The salinity of the holding water in the experiment with stingray was lower (osmolality, 946.5 mOsm/kg; Na⁺, 488.5 mM; Cl^–, 456.4 mM) than that with catshark (osmolality, 1063.5 mOsm/kg; Na⁺, 608.1 mM; Cl^–, 578.5 mM) as the sources of natural SW were different. Likewise, the plasma sodium, chloride and urea concentrations and osmolality were lower in stingray (Table 1). Because of this intrinsic difference, we compared the urine/plasma (U/P) ratio of each parameter between the stingray and catshark (Fig. 1). The U/P ratios of all the parameters were significantly lower in the stingray (osmolality, 0.79; Na⁺, 0.82; Cl^–, 0.94; urea, 0.32) than in the catshark (osmolality, 1.03; Na⁺, 1.26; Cl^–, 1.21; urea, 0.60).

Fig. 1.

The urine/plasma (U/P) ratio of each parameter shown in Table 1. Statistical analysis was performed by using the Mann-Whitney U test. Statistically significant differences between stingray and catshark are shown with asterisks (**P < 0.01).

Table 1.

Plasma and urine parameters of stingray and catshark (n = 6 for each).

Fig. 2.

Stingray and houndshark kidneys. (A, C) Low magnification cross-sectional overviews of the halves of the stingray and houndshark kidney, stained with PAS-hematoxylin. The areas surrounded by broken lines represent the bundle zones. Numbered boxes represent the nephrons examined in this study. The 3D images of corresponding nephrons are shown in Fig. 3. BZ, bundle zone; SZ, sinus zone. (B, D) Magnified views of the boundary between the BZ and SZ in each animal. Lines represent the boundaries. The tubular bundles containing the five renal tubules of the nephron (circled with broken lines and indicated by black arrowheads) are rarely found in the stingray (B) but frequently found in the houndshark (D). RC, renal corpuscle. (E) Schematic drawing of the segmentation of the stingray nephron. CV, central vessel; CD, collecting duct. PIa (p), descending proximal I (PI); PIb (p'), ascending PI; PII, proximal II; I (i), intermediate; EDT (e), early distal; LDT, late distal; CT(c), collecting. (F) Six types of the renal tubules identified in a single bundle of the stingray nephron (neck segment is not shown). Bars, 1 mm (A, C); 100 µm (B, D, F).

Histological observations of the stingray and banded houndshark nephrons

The stingray and houndshark kidneys were composed of multiple lobules, and each lobule was further divided into the bundle and sinus zones, where a line of renal corpuscle was found at the boundary between the two zones (Fig. 2A, B). In the bundle zone of houndshark kidney, we frequently observed cross-sectional views of a typical five-tubular bundle, which is composed of the descending and ascending segments of the first and third loops, and the descending collecting tubule wrapped together by a peritubular sheath (Fig. 2D). In contrast, such an apparent bundle unit was rarely observed in the stingray kidney (Fig. 2B), suggesting that the composition of the tubular bundle is somewhat different in stingray (see 3D nephron result below).

We selected three representative nephrons from the kidneys of stingray and houndshark (Fig. 2A, C), for examination of the histological and morphological characteristics of nephron segments. Based on the histological criteria in PAS-hematoxylin stained serial sections (Kakumura et al., 2015), seven major segments were identified in stingray: 1) neck; 2) proximal I (PI); 3) proximal II (PII); 4) intermediate; 5) early distal (EDT); 6) late distal (LDT); and 7) collecting (CT) (Fig. 2E, F). These classifications were later confirmed by the 3D reconstructed images. The proximal segments (PI and PII) were characterized by apical brush-border membrane and were distributed in the bundle and sinus zones, respectively (Fig. 2E, F). The PI was further divided into ascending PIa and descending PIb, composed of larger cells. The EDT possessed large columnar cells with apical membrane that was modestly PAS-positive (Fig. 2F). The intermediate segment crossing the bundle and sinus zones was a junctional tubule between the PII and EDT (Fig. 2E, F). In the sinus zone, the LDT was a small tubule composed of thin squamous cells, while the PII had a large diameter tubule with high columnar cells. The collecting tubule was composed of cuboidal cells, and the lumen was PAS-negative (Fig. 2F).

Fig. 3.

(A, B) Three-dimensional (3D) images reconstructed by using serial sections of the stingray (A) and houndshark (B) nephrons. Each component of the nephrons colored with dark blue, blue, yellow, white, orange, purple and green represent the renal corpuscle, PI, PII, intermediate, EDT, LDT and CT, respectively. For nephrons of (A-1) and (B-1), the images of whole nephron including the tubules both in the bundle zone and sinus zone are shown. For nephrons of (A-2), (A-3), (B-2), and (B-3), only the images of bundle were represented by omitting the tubules in the sinus zone. (C, D) Schematic drawings of the tubular configuration in the bundle zone of representative nephron in the stingray and houndshark, respectively. See Fig. 2 for the abbreviations.

Three-dimensional images of the stingray and houndshark nephrons

By assembling the serial section views, we generated 3D images of the nephrons of stingray and houndshark to investigate the presumed structural differences. As shown in Fig. 3A and B, a distinct difference was observed in the bundle zone; the tubular bundles of the stingray nephron were more compact than those of houndshark. In houndshark, a nephron in the bundle area could be further separated typically into two regions: a straight portion and a convoluted portion, as seen in Fig. 3B-2. The straight portion composed of the five tubular segments started in close association with the renal corpuscle and extended into the bundle zone, while the convoluted region was located at the periphery of the bundle zone and was composed mainly of EDT (orange tubules in Fig. 3B, D). In stingray, all of the bundles examined were dense and compact (Fig. 3A). The straight portion was absent or confined to a short segment near the renal corpuscle. The PI (descending and ascending segments of the first loop) and collecting tubule in the bundle were coiled around by the EDT to form a compact convoluted structure (Fig. 3A, C). All the structural characteristics of tubular bundles were consistent among all the nephrons examined so far in both species (Fig. 3).

Distributions of membrane transporters in nephron segments

In addition to the 3D reconstructions of nephrons, we conducted in situ hybridization to investigate the distribution of three major membrane transporters (NKAα1, NKCC2, and UT), which have been considered to be important for urea reabsorption in the tubular bundle. The localization of these transporters was already established in houndshark, as well as in elephant fish and bull shark (Hyodo et al., 2004; Kakumura et al., 2015; Imaseki et al., 2019). In the houndshark nephron, NKAα1 and NKCC2 mRNAs were co-expressed in the EDT and LDT, while UT mRNA was localized in the collecting tubule (Imaseki et al., 2019). In the stingray nephron, intense NKCC2 mRNA signals were found in the EDT and the first-half of the LDT (Fig. 4A–C). The expression of NKCC2 mRNA was mostly co-localized with the signals of NKAα1 mRNA (Fig. 4A–F). UT mRNA expression was found in the collecting tubule and the preceding transitional region where the LDT entered the bundle and became the collecting tubule (Fig. 4G–I). In addition to its location in the EDT and LDT, NKAα1 mRNA expression was also observed in the collecting tubule and the ascending segment of the first loop (PIb) (Fig. 4D–F).

Fig. 4.

(A, D, G) Localization of the stingray NKCC2, NKAα1, and UT mRNAs in the bundle zone detected by in situ hybridization. (B, E, H) With the serial sections that were used for 3D reconstruction, the tubules that originated from a single renal corpuscle were identified. The tubules colored blue, gray, white, orange, and green represent the PIa, PIb (p'), intermediate (i), EDT (e) and CT (c), respectively, as shown in the left panels (A, D, G). Each micrograph is identical to the left panels (A, D, G). (C, F, I) Schematic illustrations of the distributions for NKCC2, NKAα1, and UT mRNAs. (A) The intense signals of NKCC2 mRNA were found in the EDT of the bundle (e, indicated by closed arrowheads) and also found in the ascending portion of the LDT (open arrowhead). (D) In the bundle zone, the NKAα1 signals were widely found in the PIb (p'), EDT (e), and CT (c) (closed arrowheads). In the sinus zone, ascending portion of the LDT (open arrowheads) and the transitional portion (arrow) where the LDT enters the bundle and becomes CT express NKAα1. (G) The UT expression was found in the CT of the bundle (closed arrowheads) and the transitional portion of the LDT (arrow). The inset of (G) is high magnification showing the UT signal in the CT (c). See Figs 2 and 3 for the abbreviations. Bars, 50 µm.