Sulfate transporters involved in sulfate secretion in the kidney are localized in the renal proximal tubule II of the elephant fish (Callorhinchus milii)

Abstract

Most vertebrates, including cartilaginous fishes, maintain their plasma SO₄²⁻ concentration ([SO₄²⁻]) within a narrow range of 0.2–1 mM. As seawater has a [SO₄²⁻] about 40 times higher than that of the plasma, SO₄²⁻ excretion is the major role of kidneys in marine teleost fishes. It has been suggested that cartilaginous fishes also excrete excess SO₄²⁻ via the kidney. However, little is known about the underlying mechanisms for SO₄²⁻ transport in cartilaginous fish, largely due to the extraordinarily elaborate four-loop configuration of the nephron, which consists of at least 10 morphologically distinguishable segments. In the present study, we determined cDNA sequences from the kidney of holocephalan elephant fish (Callorhinchus milii) that encoded solute carrier family 26 member 1 (Slc26a1) and member 6 (Slc26a6), which are SO₄²⁻ transporters that are expressed in mammalian and teleost kidneys. Elephant fish Slc26a1 (cmSlc26a1) and cmSlc26a6 mRNAs were coexpressed in the proximal II (PII) segment of the nephron, which comprises the second loop in the sinus zone. Functional analyses using Xenopus oocytes and the results of immunohistochemistry revealed that cmSlc26a1 is a basolaterally located electroneutral SO₄²⁻ transporter, while cmSlc26a6 is an apically located, electrogenic Cl⁻/SO₄²⁻ exchanger. In addition, we found that both cmSlc26a1 and cmSlc26a6 were abundantly expressed in the kidney of embryos; SO₄²⁻ was concentrated in a bladder-like structure of elephant fish embryos. Our results demonstrated that the PII segment of the nephron contributes to the secretion of excess SO₄²⁻ by the kidney of elephant fish. Possible mechanisms for SO₄²⁻ secretion in the PII segment are discussed.

sulfate (SO₄²⁻) is essential for various biological processes, including the production of highly sulfated proteoglycans by chondrocytes, detoxification of many endogenous and exogenous compounds, and maintenance of cell membranes (36). Accordingly, most vertebrates maintain the plasma SO₄²⁻ concentration ([SO₄²⁻]) within a narrow range (0.2–1.0 mM), and the kidney plays a pivotal role in maintaining SO₄²⁻ homeostasis. Terrestrial and freshwater (FW) habitats are SO₄²⁻-deficient environments, therefore, the kidneys of mammals and FW-teleost fishes reabsorb SO₄²⁻ from the filtrate to retain SO₄²⁻ in the body (35, 39). In contrast, the [SO₄²⁻] in seawater (SW) is >40-fold higher than that in the plasma of marine teleost fishes, which drink SW and excrete NaCl via the gills to avoid dehydration, and excrete SO₄²⁻ via the kidney to prevent hypersulfatemia (25, 41, 47). In stenohaline marine teleosts, the kidney lacks the distal tubule, and SO₄²⁻ is secreted from the proximal tubule (2).

Most cartilaginous fishes (sharks, skates, rays, and chimeras) inhabit a marine environment, and comparison of the composition of plasma and urine has revealed that the concentration of divalent ions in the urine (26–182 mM) is 10 to 100 times higher than that in the plasma (0.5–8 mM) (4, 45, 52, 53). This suggests that divalent ions, including SO₄²⁻, are secreted by the cartilaginous fish kidney. Indeed, a micropuncture study implied that there was a secretory segment for divalent ions in the nephron of the little skate (54). However, the nephron configuration of marine cartilaginous fish is considerably more complex than the simple nephron of marine teleost fish, as each nephron has an elaborate four-loop configuration and can be divided at least into 10 morphologically distinguishable segments (Fig. 1) (15, 22, 32). Although the micropuncture study suggested the second loop of the nephron as a secretory segment for divalent ions (54), the functional characteristics of the diverse nephron segments and the mechanisms of SO₄²⁻ secretion in the nephron are still largely unknown because of the complexity of the cartilaginous fish kidney.

Fig. 1.

Schematic drawing showing the four-loop nephron of cartilaginous fish (based on and modified from Ref. 22). The encircled numbers represent the number of loops. The representative segments that are common to both elasmobranchs and elephant fish nephrons are shown: PIa, proximal tubule Ia; PII, proximal tubule II; EDT, early distal tubule; LDT, late distal tubule; CT, collecting tubule; CV, central vessel; PS, peritubular sheath; RC, renal corpuscle.

There is an increasing amount of information on the distribution patterns of membrane transporters in osmoregulatory organs of cartilaginous fishes, which provides valuable information on the mechanisms that permit transepithelial movement of small molecules such as ions and urea (6, 7, 17, 18, 22, 43, 55). For example, we have conducted mapping of membrane transporters in the renal tubule of cartilaginous fishes and revealed the importance of the distal tubules and the collecting tubule for urea reabsorption (17, 18, 22, 64). In the present study, we cloned putative SO₄²⁻ transporters [solute carrier family 26 member 1 (Slc26a1) and member 6 (Slc26a6)] from the cartilaginous fish kidney and examined SO₄²⁻-transporting activities in vitro using Xenopus oocytes, to clarify the nephron segment and the mechanisms involved in the excretion of SO₄²⁻. Slc26a1 and Slc26a6 belong to the Slc26 gene family, which are anion transporters, and 10 Slc26 genes have been identified in mammals (1). The Slc26a1 and Slc26a6 proteins have been found to be localized in the proximal tubules of mammalian and marine teleost kidneys, where they play a role in SO₄²⁻ transport and homeostasis (24, 25–27, 39, 42, 62). For this study, we chose the holocephalan elephant fish (also called elephant shark, Callorhinchus milii) as a model cartilaginous fish, as elephant fish is one of the few species of cartilaginous fish for which the genome is being sequenced (59, 60). Therefore, this species is an excellent experimental model to examine molecular mechanisms involved in renal function in cartilaginous fish (17). We found that the proximal II segment of the second loop is most probably the nephron segment contributing to the SO₄²⁻ secretion in the elaborate cartilaginous fish nephron. Our findings provide crucial information to improve our understanding of cartilaginous fish kidney function.

MATERIALS AND METHODS

Adult fish and embryos.

Elephant fish (Callorhinchus milii; Bory de Saint-Vincent, 1823) of both sexes were collected in Western Port Bay, Victoria, Australia, using recreational fishing equipment, and transported to Primary Industries Research Victoria, Queenscliff, using a fish transporter. The fish for tissue collection were kept in a 10,000-l round tank with running SW under a natural photoperiod for at least 3 days before sampling (16). To obtain developing embryos, mature female fish were kept in 10,000-l round tanks under a natural photoperiod for approximately 2 mo from March to May and fed daily with pilchard. During that period, newly laid eggs were collected from the tanks and were maintained in 1,000-l tanks with running SW. Embryos at stage 36 of development were sampled at the end of September (57). The developmental stages were identified by using an established staging scheme (9).

For sampling, fish were anesthetized in 0.1% (wt/vol) 3-amino benzoic acid ethyl ester (Sigma-Aldrich, St. Louis, MO). Blood samples were obtained from the caudal vasculature with a syringe (adult fish) or a heparin-coated hematocrit capillary (embryos) (Terumo, Tokyo, Japan) and were centrifuged at 2,250 g for 10 min to obtain plasma. Urine that was present in the bladder-like structure of stage 36 embryos was collected with a syringe and fine-gauge needle. The urine and plasma were stored at −20°C. After decapitation of fish, tissues were dissected out and quickly frozen in liquid nitrogen, and kept at −80°C until required. For mRNA and protein localization, tissues were fixed in 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4) containing 350 mM urea, or Bouin's solution without acetic acid at 4°C for 2 days, and then washed in 70% ethanol, and stored at 4°C. All animal experiments were conducted according to the “Guidelines for the Care and Use of Animals” and were approved by the Animal Care and Use Committees of the University of Tokyo and Deakin University.

cDNA cloning.

Total RNA was extracted from the kidney with Isogen (Nippon Gene, Toyama, Japan). Two micrograms of total RNA was treated using a TURBO DNA-free kit (Life Technologies, Carlsbad, CA) and reverse-transcribed to first-strand cDNA using a high-capacity cDNA reverse transcription kit (Life Technologies), following the manufacturer's instructions. The amino acid sequences of Slc26a1 and Slc26a6 from clawed frog (Xenopus laevis) were used as BLAST queries to find candidate genes in the Elephant Shark Genome Database (http://esharkgenome.imcb.a-star.edu.sg/). The cDNA sequence encoding the entire open reading frame (ORF) of elephant fish (cm) Slc26a1 (cmSlc26a1) and Slc26a6 (cmSlc26a6) were determined by 3′- and 5′-RACE methods using the SMARTer RACE cDNA amplification kit (Clontech, Mountain View, CA), following the manufacturer's instructions. cDNAs were amplified with high-fidelity KOD-Plus DNA polymerase (Toyobo, Osaka, Japan), ligated into pGEM T-easy (Promega, Madison, WI), and the nucleotide sequence was determined by an automated DNA sequencer (3130xl Genetic Analyzer; Life Technologies) and BigDye Terminator v3.1 cycle sequencing kit (Applied Biosystems, Foster City, CA). All primers used are listed in Table 1. Putative transmembrane domains were predicted by a Kyte-Doolittle hydropathy plot constructed using the GENETYX software (GENETYX, Tokyo, Japan). The possible functional domains of cmSlc26a1 and cmSlc26a6 were predicted by online InterPro Scan software (21).

Table 1.

Primer sets used in the present study

Primers for 5′-RACE
Gene Name	Primer Sequence 5′ to 3′
cmSlc26a1	AS	CAAGGCAACTGCAATACTAGC
cmSlc26a6	AS	CCCGAATGTAAAGTAAATCAGC
Primers for 3′-RACE
cmSlc26a1	S	ACAATGAAGTCAGTCCTGAAGG
cmSlc26a6	S	AGATGTCCTCCTCAATTATTGG
Primer sets for cloning and functional assays
cmSlc26a1	S	CACTCTACAGTAACATCTGACAGC
	AS	GGTCTGAGAAATCTTAAATGGAAG
cmSlc26a6	S	ATCGCCAGCAAATCAAAAC
	AS	TTTTCTCCAGAGTTTGAGGATC
Primer sets for real-time qPCR assay
cmSlc26a1	S	GGAGAAATCCCCACAGGATTC
	AS	CAGCATCTATGGCAACTCTGAAAA
cmSlc26a6	S	CTCCACTGCCTGGCGAAA
	AS	AGTGAAGGGTGTCCATTCATGAG
Primer sets for in situ hybridization
cmSlc26a1	S	GAGAAGTGGAACGCCAGTGT
	AS	GTTTTCTTGCATGTTGCCACT
cmSlc26a6	S	(same with the 3′-RACE primer)
	AS	(same with the cloning primer)

S and AS mean sense and antisense, respectively. qPCR, quantitative PCR.

Tissue distribution of cmSlc26a1 and cmSlc26a6 mRNAs.

The tissue distribution of cmSlc26a1 and cmSlc26a6 mRNAs was examined by quantitative real-time PCR (qPCR) using a 7900HT Sequence Detection System (Life Technologies) and KAPA SYBR FAST ABI Prism qPCR kit (Kapa Biosystems, Boston, MA), as previously described in detail (64). Total RNA was extracted from 12 tissues (brain, gill, heart, liver, spleen, pancreas, spiral intestine, kidney, rectal gland, skeletal muscle, gonad, and uterus) of four adult fish (two males and two females), and 10 tissues (brain, gill, heart, liver, spleen, spiral intestine, kidney, rectal gland, urinary bladder, and skeletal muscle) of six stage 36 embryos using Isogen. To generate a standard curve for mRNA quantification, partial cDNA fragments were obtained with specific primer sets, which were purified using a QIAquick PCR purification kit (Qiagen, Hilden, Germany). The cDNAs were then serially diluted and were then used as the known amount of standard cDNAs for absolute quantification in qPCR analyses. The copy numbers of the standard cDNAs were calculated with BioMath Calculators (http://www.promega.com/techserv/tools/biomath/calc01.htm). As internal controls, elephant fish β-actin (cmACTB, NM_001292743) mRNA was used for adult fish, while elephant fish elongation factor 1α (cmEF1α, AB622989) was used for embryos. Previous studies have shown that, compared with ef1α gene, the expression level of the β-actin gene is unstable during development in zebrafish and elephant fish, while the β-actin gene is more suitable for adult fishes (38, 56). Primer sets for qPCR were designed using Primer Express software, and their sequences are shown in Table 1.

Molecular phylogenetic analysis.

The translated amino acid sequences of cmSlc26a1 and cmSlc26a6 were aligned with those of other vertebrate Slc26 proteins using Clustal W 2.1 program (58). The sequences of Slc26 family proteins were obtained from the DDBJ and Ensembl databases. Molecular phylogenetic trees were constructed by a Bayesian Metropolis-coupled Markov chain Monte Carlo method in the MrBayes 3.1.2 program (http://mrbayes.sourceforge.net/). We ran four separate Marcov chains for 1,000,000 generations and sampled them every 100 generations to create a posterior probability distribution of 10,000 trees. The first 2,500 trees were discarded as burn-in before stabilization, and then a 50% majority-rule tree was constructed from the subsequent trees. The reliability of the generated tree was shown by posterior probabilities in the Bayesian analysis.

In situ hybridization.

Kidneys that had been fixed in Bouin's solution without acetic acid were embedded in Palaplast (McCormick Scientific, Richmond, IL). Serial sections were cut at 6-μm thickness and mounted onto MAS-coated slides (Matsunami Glass, Osaka, Japan). The cDNA fragments (cmSlc26a1, 1361 bp; cmSlc26a6, 1383 bp; Na⁺,K⁺,2Cl⁻ cotransporter-2 [cmNKCC2; AB769493], 1337 bp) were amplified using gene-specific primers listed in Table 1, 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. Deparaffinized sections were treated with 5 mg/ml proteinase K (Sigma-Aldrich) in Tris-EDTA buffer [100 mM tris(hydroxymethyl)aminomethane and 50 mM ethylenediaminetetraacetic acid, pH 7.5] and then hybridized with 1 μg/ml DIG-labeled cRNA probes in hybridization buffer [50% formamide, 5 × SSC Buffer, 40 μg/ml bovine calf thymus DNA] at 58°C for 40 h. After hybridization, sections were washed in 2 × SSC (0.3 M NaCl and 33.3 mM sodium citrate, pH 7.0) for 30 min at room temperature, followed by 1 × SSC for 1 h at 65°C, and finally in 0.1 × SSC for 1 h at 65°C. Following immunohistochemical reaction with alkaline phosphatase-conjugated anti-DIG antibody (Roche Applied Science), hybridization signals were visualized with 4-nitro blue tetrazolium chloride (450 mg/ml) and X-phosphate/5-bromo-4-chloro-3-indolyl-phosphate (175 mg/ml) for 16–48 h at room temperature. Stained sections were mounted with Aquamount (BDH Laboratory Supplies, Poole, England). Micrographs were obtained using a digital camera (DXM1200; Nikon, Tokyo, Japan).

Immunohistochemistry.

The peptides IYIERTEREKPKVK+C from cmSlc26a1 and C+MEQPGSGPEKHTLE from cmSlc26a6, corresponding to the NH₂-terminal and COOH-terminal cytoplasmic domain, respectively, were synthesized and coupled via cysteine to keyhole limpet hemocyanin. These conjugated peptides were emulsified with complete Freund's adjuvant and injected into Japanese white rabbits for immunization (Eurofins Genomics, Tokyo, Japan).

Kidneys fixed with 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4) containing 350 mM urea were embedded in Paraplast. Serial sections were cut at 6-μm thickness and mounted onto MAS-coated slides. Immunohistochemical staining for cmSlc26a1 and cmSlc26a6 was performed with the avidin-biotin-peroxidase complex kit (Vector Laboratories, Burlingame, CA) or Elite avidin-biotin-peroxidase complex kit (Vector Laboratories). After rehydration, tissue sections were incubated sequentially with 1) 2% normal goat serum in PBS (pH 7.4; PBS-NGS) for 30 min at room temperature, 2) cmSlc26a1 (1:2,000) or cmSlc26a6 (1:20,000) antiserum diluted with PBS-NGS for 48 h at 4°C, 3) biotinylated, goat anti-rabbit IgG for 30 min at room temperature, 4) avidin-biotin-peroxidase complex for 45 min or 30 min at room temperature, and 5) 0.05% diaminobentizine tetrahydrochloride (Sigma-Aldrich) and 0.01% hydrogen peroxide in 50 mM Tris buffer (pH 7.2) for 10 min at room temperature. The specificity of immunoreactive signals for cmSlc26a1 and cmSlc26a6 was confirmed by preabsorption of antibodies with the synthetic antigens (2.5 × 10⁻⁶ M) for 24 h at 4°C prior to incubation. Stained sections were counterstained with hematoxylin. Micrographs were obtained using a digital camera (DXM1200; Nikon, Tokyo, Japan).

Measurement of osmolality and ion concentrations.

Plasma and urine osmolality were measured using a vapor pressure osmometer (Wescor 5520; Logan, UT). Anions (Cl⁻ and SO₄²⁻) in the plasma and urine were measured using ion chromatography (Shimadzu AV10; Kyoto, Japan), while cations (Na⁺, Mg²⁺, and Ca²⁺) were measured by an atomic absorption spectrophotometer (Hitachi 180-50; Tokyo, Japan). Urea concentration was measured using a Wako Urea NB test (Wako Pure Chemical Industries, Osaka, Japan).

Characterization of SO₄²⁻-transporting activities of cmSlc26a1 and cmSlc26a6 using [³⁵S]SO₄²⁻.

The entire coding regions of cmSlc26a1 and cmSlc26a6 cDNAs were inserted into the pGEMHE X. laevis expression vector. The constructs were linearized with NheI, and cRNAs were transcribed in vitro using the T7 mMESSAGE mMACHINE kit (Ambion, Austin, TX). Oocytes were surgically collected from X. laevis and incubated for 20 min with shaking in modified Barth solution (MBS) without calcium [MBS(-); in mM: 88 NaCl, 1 KCl, 2.4 NaHCO₃, 0.82 MgSO₄, 10 HEPES, pH 7.4] and were manually defolliculated after the treatment with 1 mg/ml collagenase (Sigma-Aldrich). Defolliculated oocytes were incubated in MBS with calcium [MBS(+); in mM: 88 NaCl, 1 KCl, 2.4 NaHCO₃, 0.3 Ca(NO₃)₂, 0.41 CaCl₂, 0.82 MgSO₄] for 24 h and then injected with 50 nl of water or 1 μg/μl of the cRNA (50 ng/oocyte). The injected oocytes were incubated at 18°C in MBS(+) supplemented with 100 units/ml penicillin and 10 mg/ml streptomycin sulfate for 48–72 h. The incubation medium was changed every 24 h, including the day of the uptake experiment.

On the day of assay, the injected oocytes were incubated in a preincubation buffer (in mM: 94 NaCl, 4.5 KCl, 1.8 CaCl₂, 1 MgCl₂, 10 HEPES-Tris, and 1 Na₂SO₄, pH 7.5) for 30 min at 18°C. Oocytes were then placed in one of the following four uptake buffers (Na⁺ buffer, Cl⁻-free Na⁺ buffer, K⁺ buffer, and Cl⁻-free K⁺ buffer) containing 10 μCi/ml [³⁵S]Na₂SO₄ (NEX041H; PerkinElmer, Waltham, MA) for 1 h at 18°C. The composition of each buffer is as follows: Na⁺ buffer (in mM: 94 NaCl, 4.5 KCl, 1.8 CaCl₂, 1 MgCl₂, 10 HEPES-Tris, and 1 Na₂SO₄, pH 7.5), Cl⁻-free Na⁺ buffer (in mM: 94 Na-gluconate, 4.5 K-gluconate, 1.8 Ca-gluconate, 1 Mg-gluconate, 10 HEPES-Tris, and 1 Na₂SO₄, pH 7.5), K⁺ buffer (in mM: 98.5 KCl, 1.8 CaCl₂, 1 MgCl₂, 10 HEPES-Tris, and 1 Na₂SO₄, pH 7.5), and Cl⁻-free K⁺ buffer (in mM: 98.5 K-gluconate, 1.8 Ca-gluconate, 1 Mg-gluconate, 10 HEPES-Tris, and 1 Na₂SO₄, pH 7.5). One mM of 4,4′-diisothiocyanatostilbene-2,2′-disulfonic acid (DIDS; Sigma-Aldrich) was used for inhibiting SO₄²⁻ transport. The uptake was terminated by washing oocytes in ice-cold buffer that they had been incubated in without [³⁵S]Na₂SO₄. Oocytes were dissolved in 10% SDS, and the radioactivity was measured using a liquid scintillation counter (Tri-carb 3100 TR; PerkinElmer).

Characterization of cmSlc26a1 and cmSlc26a6 by electrophysiology using Cl⁻-selective microelectrodes.

The cmSlc26a1 and cmSlc26a6 cDNAs were prepared as described above. Defolliculated X. laevis oocytes were injected with 50 nl of water or 0.5 μg/μl of the cRNA (25 ng/oocyte). Oocytes were incubated at 16°C in oocyte recipe 3 (OR3) media (48), and examined 3–6 days after injection.

Intracellular free Cl⁻ concentration was measured as intracellular Cl⁻ activity (aCl_i) using Cl⁻-selective microelectrodes prepared with Cl⁻-ionophore I, cocktail A (Fluka Chemical, Ronkonkoma, NY), as described previously (49). aCl_i was measured from the difference in potential between the Cl⁻ electrode and a KCl reference electrode impaled into the oocyte, and membrane potential (V_m) was measured as the difference between the KCl microelectrode and an extracellular calomel. Cl⁻ electrodes were calibrated using 10 and 100 mM NaCl, followed by a test of the specificity by using 100 mM NaHCO₃ and a point calibration in ND96 [in mM: 96 NaCl, 2 KCl, 1 MgCl₂, 1.8 CaCl₂, and 5 HEPES (pH 7.5)]. For Cl⁻-free buffer (0Cl-ND96), Cl⁻ was replaced with gluconate. For solutions containing SO₄²⁻, 10 mM NaCl or Na-gluconate was replaced with 5 mM Na₂SO₄ and 2.5 mM choline chloride or choline gluconate, pH 7.5 (25).

All X. laevis experiments were conducted according to the “Guidlines for the Care and Use of Animals” and were approved by the Animal Care and Use Committees of the University of Tokyo and Deakin University.

Statistical analysis.

Data are presented as means ± SE. Data from the radioisotope assay were analyzed by one-way ANOVA followed by Tukey-Kramer's multiple-comparison test. Unpaired t-tests were used for all other data analysis. P values less than 0.05 were considered statistically significant.

RESULTS

Identification of SO₄²⁻ transporters Slc26a1 and Slc26a6 in elephant fish.

Using the amino acid sequences of Slc26a1 and Slc26a6 proteins from frog (X. laevis) as queries, we found a number of fragments sharing high homology to Slc26a1 and Slc26a6 in the elephant fish genome database. We then obtained cDNA sequences encoding the entire ORF of Slc26a1 and Slc26a6. The elephant fish (cm) Slc26a1 (cmSlc26a1) and Slc26a6 (cmSlc26a6) mRNAs encoded a putative ORF of 720 and 769 amino acid residues, respectively. Molecular phylogenetic analysis with vertebrate Slc26 proteins confirmed that cmSlc26a1 (accession: LC089740) and cmSlc26a6 (accession: LC089741) belonged to the Slc26a1 and Slc26a6 subfamilies, respectively (Fig. 2). In teleosts, three Slc26a6 genes (slc26a6a, slc26a6b, and slc26a6c) have been identified and, accordingly, the three Slc26a6 subtypes grouped together in the teleost lineage of the Slc26a6 clade (Fig. 2). Both cmSlc26a1 and cmSlc26a6 contained 12 putative membrane-spanning domains with intracellular amino and carboxyl termini, as have been shown in mammals (34, 46) and teleosts (31, 39) (black horizontal bars in Fig. 3). Recently, the rat Slc26a5 was reported to have 14 transmembrane domains (13); possible additional transmembrane domains are indicated by open horizontal bars in Fig. 3. A Slc26/sulfate permease (SulP) transporter domain (PF00916) and a sulfate transporter and Ant-Sigma factor antagonist (STAS) domain (PF01740), which are features of the Slc26 transporter family (1), were found in both cmSlc26a1 and cmSlc26a6 (Fig. 3). A glutamate residue, which is considered to be important for the Slc26 transporter properties (40), was conserved in both cmSlc26a1 and cmSlc26a6 (Fig. 3). On the other hand, the three amino acids at the carboxyl termini of Slc26a3 and Slc26a6 were reported to interact with the PDZ (postsynaptic density 95/discs large/zona occludens 1) domain (10). The consensus sequence (TRL/TKF) was found in cmSlc26a6, but not in cmSlc26a1 (Fig. 3)

Fig. 2.

Molecular phylogeny of vertebrate Slc26 family proteins. The elephant fish (holocephalan) sequences identified in this study are shown in bold letters with selected vertebrates (mammals and teleosts) indicated in normal font. The accession numbers of genes and mRNAs used in the analysis are listed in Table 2. Numbers at branch nodes represent Bayesian posterior probabilities.

Fig. 3.

Primary structure of elephant fish (cm) Slc26a1 and Slc26a6 cloned from the kidney. The deduced amino acid sequences are aligned with those of human (hs) Slc26a1 and Slc26a6, respectively. The accession numbers of the genes are listed in Table 2. For simplicity and to reduce alignment gaps, other Slc26 sequences, such as teleost Slc26a1 and Slc26a6, have not been included in the alignment; these show considerable similarity to mammalian Slc26a1 and Slc26a6, respectively (25, 26, 31, 39). Gray shading indicates identical amino acid residues between cmSlc26a1 (cm1) and hsSlc26a1 (hs1) and between cmSlc26a6 (cm6) and hsSlc26a6 (hs6), respectively. Identical amino acid residues between cmSlc26a1 and cmSlc26a6 are indicated by asterisks. Hyphens denote gaps introduced for the alignment. The hash mark (#) represents the glutamate residue that is important for the Slc26 transporter properties (40). Wavy lines indicate the three amino acid residues that interact with the PDZ domain. TM, transmembrane domain; STAS, a sulfate transporter and Ant-sigma factor antagonist domain. Note that the Slc26 family has been considered to have 12 TMs (indicated by horizontal solid bars), while the rat Slc26a5 was recently reported to have 14 TMs; additional putative TMs are indicated by open horizontal bars.

Table 2.

The accession numbers of genes used in the analysis

Gene Name	Accession Number
cmSlc26a1	LC089740
cmSlc26a6	LC089741
Eel Slc26a1	AB111927
Eel Slc26a3	AB111930
Eel Slc26a6A	AB084425
Eel Slc26a6B	AB111928
Eel Slc26a6C	AB111929
Zebrafish Slc26a1	NM_001080667
Zebrafish Slc26a2	XM_680022
Zebrafish Slc26a3	FJ170816
Zebrafish Slc26a4	NM_001165915
Zebrafish Slc26a5	NM_201473
Zebrafish Slc26a6A	ENSDART00000149230
Zebrafish Slc26a6B	ENSDART00000045602
Zebrafish Slc26a6C	FJ170818
Zebrafish Slc26a11	NM_199767
Human Slc26A1	AF297659
Human Slc26a2	U14528
Human Slc26a3	NM_000111
Human Slc26a4	NM_000441
Human Slc26a5	AF523354
Human Slc26a6	AF279265
Human Slc26a7	AF331521
Human Slc26a8	AF331522
Human Slc26a9	AF314958
Human Slc26a11	AF345195
Mouse Slc26a2	NM_007885
Mouse Slc26a4	NM_011867
Mouse Slc26a5	NM_030727
Mouse Slc26a6	NM_134420
Mouse Slc26a7	NM_145947
Mouse Slc26a8	AF403499
Mouse Slc26a9	NM_177243
Mouse Slc26a11	NM_178743
Fugu Slc26a5	AB200327
Fugu Slc26a6A	AB200328
Fugu Slc26a6B	AB200329
Fugu Slc26a6C	AB200330
Fugu Slc26a11	AB200331
Tetraodon Slc26a1	ENSTNIT00000016422

Tissue distribution of cmSlc26a1 and cmSlc26a6 mRNAs in adult fish.

The tissue distribution of cmSlc26a1 and cmSlc26a6 mRNAs was quantitatively examined in four adult elephant fish (Fig. 4). The cmSlc26a1 mRNA was almost exclusively expressed in the kidney. The expression of cmSlc26a6 mRNA was also highest in the kidney, but moderate levels of expression were found in the brain, intestine, and gonad.

Fig. 4.

Tissue distribution of mRNAs encoding cmSlc26a1 (A) and cmSlc26a6 (B) in adult elephant fish. Data are presented as means ± SE. The values were normalized against those of cmACTB mRNA. n = 4 (two males and two females) for each tissue analysis, except for uterus which was n = 2.

Functional properties of cmSlc26a1 and cmSlc26a6: radioisotope assay.

The SO₄²⁻-transporting activities of cmSlc26a1 and cmSlc26a6 were examined using X. laevis oocytes and [³⁵S]-labeled SO₄²⁻ (Fig. 5). In response to exposure to Cl⁻-free Na⁺ buffer (0 mM Cl⁻, open bars in Fig. 5) containing SO₄²⁻, the oocytes expressing cmSlc26a6 showed significantly higher SO₄²⁻ uptake (405.8 ± 39.1 pmol/oocyte) compared with the uptake in water-injected control oocytes (8.2 ± 3.2 pmol/oocyte) (Fig. 5A). Meanwhile, exposure to Na⁺ buffer (100 mM Cl⁻, black bars in Fig. 5) did not induce SO₄²⁻ uptake in the oocytes expressing cmSlc26a6 (1.5 ± 0.1 pmol/oocyte in cmSlc26a6 cRNA-injected oocytes and 1.7 ± 0.3 pmol/oocyte in water-injected oocytes). The observed SO₄²⁻ uptake in the Cl⁻-free Na⁺ buffer was abolished (2.1 ± 0.4 pmol/oocyte) by the addition of 1 mM DIDS, an inhibitor of anion transport (Fig. 5A). Depolarization of oocytes by replacing Na⁺ with K⁺ in the Cl⁻-free assay medium resulted in a twofold increase in SO₄²⁻ uptake (Cl⁻-free K⁺ buffer; 845 ± 131.2 pmol/oocyte) (Fig. 5A). K⁺ depolarization of oocytes moderately increased SO₄²⁻ uptake in the K⁺ buffer containing 100 mM Cl⁻, but the increase was not statistically significant.

Fig. 5.

Functional characterization of cmSlc26a1 and cmSlc26a6 using [³⁵S]SO₄²⁻. A: [³⁵S]SO₄²⁻ uptake mediated by water-injected or cmSlc26a6 cRNA-injected oocytes was examined in Na⁺ buffer or K⁺ buffer in the presence (solid bars) or absence (open bars) of 100 mM Cl⁻. DIDS was used as a general anion exchanger inhibitor. ^a,b,cValues not sharing an identical letter are significantly different (P < 0.05). B: [³⁵S]SO₄²⁻ uptake mediated by water-injected or cmSlc26a1 cRNA-injected X. laevis oocytes. ^a,bValues not sharing the identical letter are significantly different (P < 0.05).

Injection of cmSlc26a1 cRNA into oocytes also induced SO₄²⁻ uptake (Fig. 5B). In contrast to the results in cmSlc26a6-expressing oocytes, significantly higher SO₄²⁻ transport was observed when 100 mM Cl⁻ (black bars in Fig. 5) was added in the Na⁺ buffer (260.8 ± 34.0 pmol/oocyte) (Fig. 5B). Exposure to Cl⁻-free (open bars in Fig. 5) Na⁺ buffer moderately increased SO₄²⁻ uptake (34.9 ± 2.2 pmol/oocyte); however, the level was 7 times lower than that after exposure to Na⁺ buffer. The observed SO₄²⁻ uptake in both assay media was abolished by the addition of 1 mM DIDS. No significant difference was observed in SO₄²⁻ uptake between exposure to Na⁺ buffer and K⁺ buffer (Fig. 5B).

Functional properties of cmSlc26a1 and cmSlc26a6: electrophysiological assay.

The Cl⁻/SO₄²⁻ exchanging activity was examined by monitoring the intracellular Cl⁻ activity (aCl_i) and membrane potential (V_m) of oocytes in response to exposure to Cl⁻-free medium containing SO₄²⁻ (Fig. 6). Initial aCl_i values were 47.3 ± 3.2 mM for cmSlc26a6-injected oocytes (n = 10), 45.1 ± 2.7 mM for cmSlc26a1-injectied oocytes (n = 3) and 46.3 ± 4.6 mM for water-injected oocytes (n = 5). Meanwhile, initial V_m values were −30.5 ± 2.6 mV for oocytes expressing cmSlc26a6 (n = 14), −39.1 ± 4.9 mV for oocytes expressing cmSlc26a1 (n = 3), and −50.8 ± 7.9 mV for water-injected oocytes (n = 9). Exposure to 5 mM SO₄²⁻ elicited a moderate hyperpolarization (−18.2 ± 3.7 mV; n = 10, P < 0.01) in cmSlc26a6-injected oocytes. In oocytes expressing cmSlc26a6, Cl⁻ removal from the assay medium caused marked reduction of aCl_i (−1.03 ± 0.23 mM/min, n = 10, P < 0.01, for cmSlc26a6; 0.03 ± 0.03 mM/min, n = 5, for control) and a marked hyperpolarization (−58.4 ± 8.0 mV, n = 10, P < 0.01, for cmSlc26a6; −4.1 ± 0.8 mV, n = 5, for control). Readdition of Cl⁻ elicited depolarization and recovery of aCl_i to the initial level. Control oocytes did not show these responses caused by sequential removal and supplementation of Cl⁻.

Fig. 6.

Functional characterization of cmSlc26a1 and cmSlc26a6 using Cl⁻-selective microelectrodes. Representative traces are shown for intracellular Cl⁻ activity (aCl_i) and membrane potential (V_m) of oocytes expressing either cmSlc26a1 and cmSlc26a6, and control oocytes. In the continuous presence of 5 mM SO₄²⁻, Cl⁻/SO₄²⁻ exchanging activity was monitored as changes in aCl_i and V_m after extracellular Cl⁻ was removed (0 Cl⁻) and replaced by gluconate.

For cmSlc26a1 cRNA-injected oocytes, there was no observed change in aCl_i or V_m in response to exposure to SO₄²⁻-containing medium or Cl⁻-free medium (Fig. 6).

In situ hybridization.

As described in other marine elasmobranchs, the elephant fish kidney consists of multiple, irregular lobules, and each lobule is separated into two zones, a sinus zone and a bundle zone (Fig. 7A). A single nephron makes four loops within the lobule. Beginning at the renal corpuscle, the first and third loops are situated in the bundle zone, and the second and fourth loops are in the sinus zone (Fig. 7M). In situ hybridization revealed positive signals for cmSlc26a1 and cmSlc26a6 mRNAs in the sinus zone but not in the bundle zone (Fig. 7, D and G). In the sinus zone, two major nephron segments are identifiable: a proximal II (PII) segment and a late distal tubule (LDT). PII is characterized by the largest tubular diameter with an extensive brush border on the apical membrane (white arrows in Fig. 7B), while the LDT has a relatively small diameter and thin epithelial cells without a brush border on the apical membrane (white arrowheads in Fig. 7B). The hybridization signals for both cmSlc26a1 and cmSlc26a6 mRNAs were detected in the PII segment (black arrows in Fig. 7, E and H), but not in the LDT (black arrowheads in Fig. 7, E and H). The signal intensity of cmSlc26a1 mRNA was considerably higher than that of cmSlc26a6 mRNA. This difference is consistent with the results of the qPCR experiment, in which the expression level of cmSlc26a1 mRNA in the kidney was about 10 times higher than that of cmSlc26a6 mRNA (Fig. 4). For comparison, the expression of Na⁺,K⁺,2Cl⁻ cotransporter-2 (cmNKCC2) mRNA was detected in the LDT (black arrowheads in Fig. 7K) and the early distal tubule (EDT) in the bundle zone (Fig. 7L), as previously described (17). No colocalization was observed between cmNKCC2 mRNA (black arrowheads) and the hybridization signals for cmSlc26a1 and cmSlc26a6 mRNAs (black arrows). No signal was observed in the negative controls in which sections were incubated with sense probes (data not shown).

Fig. 7.

Kidney sections subjected to either hematoxylin-and-eosin (HE) staining (A–C) or in situ hybridization with cRNA probes for cmSlc26a1 (D–F), cmSlc26a6 (G–I), or cmNKCC2 (J–L). A, D, G, and J are low-power micrographs. The kidney lobule is separated into two zones, a sinus zone (SZ) and a bundle zone (BZ). Signals for cmSlc26a1 and cmSlc26a6 mRNAs were detected only in the SZ, while cmNKCC2 mRNA was expressed in tubules in both the sinus and bundle zones. B, E, H, and K are magnified views of the SZ. In the SZ, two major nephron segments are identifiable: a proximal II (PII; arrows) segment and a late distal tubule (LDT; arrowheads). The signals for cmSlc26a1 and cmSlc26a6 mRNAs were colocalized in the PII segments, while cmNKCC2 mRNA was expressed in the LDT. C, F, I, and L are magnified views of the BZ. In the BZ, a cross-sectional view of the five tubular segments can be identified (C, inset). cmNKCC2 mRNA was expressed in the early distal tubule, while cmSlc26a1 and cmSlc26a6 mRNAs were not expressed in the bundle zone (F, I). M: schematic representation of the elephant fish nephron showing the localization of cmNKCC2 mRNA (modified on the basis of Ref. 22). The encircled numbers represent the number of loops. CV, central vessel; RC, renal corpuscle; PS, peritubular sheath. Scale bars: 500 μm (A, D, G, J) and 100 μm (B, C, E, F, H, I, K, L).

Immunohistochemical localization.

The intracellular localization of cmSlc26a1 and cmSlc26a6 was examined using antisera raised against synthetic polypeptides of each transporter, respectively. Consistent with the results of the in situ hybridization experiments, immunoreactive signals for cmSlc26a1 and cmSlc26a6 were observed in the PII segments of the nephron (arrows in Fig. 8). Immunoreaction for cmSlc26a1 was localized to the basolateral membrane (Fig. 8A), while cmSlc26a6 immunoreactivity was detected on the apical, brush border membrane (Fig. 8C). Preabsorption of the anti-cmSlc26a1 (Fig. 8B) and anti-cmSlc26a6 (Fig. 8D) sera with the synthetic cmSlc26a1 and cmSlc26a6 polypeptides, respectively, resulted in the disappearance of the immunoreactive signals on the membranes. The treatment of sections with preimmune sera showed no specific signals (data not shown).

Fig. 8.

Immunohistochemical localization of the cmSlc26a1 and cmSlc26a6 in the sinus zone of elephant fish kidney. Immunoreactive signals to cmSlc26a1 were detected in the basolateral membrane of PII cells (arrows), while signals to cmSlc26a6 were detected in the apical membrane of PII cells (C). Preabsorption of antibodies with the antigen peptides resulted in disappearance of the immunoreactive signals on the membranes (B, D). Sections are counterstained with hematoxylin. Scale bars: 50 μm.

Expression in embryos.

We discovered a urinary bladder-like structure in the perihatching embryos (stage 36 and hatched embryos) of elephant fish, and we measured ion concentrations in the plasma and urine of stage 36 embryos. The plasma concentration of Na⁺ in embryos was significantly lower than that in adult fish, while plasma urea concentration was higher in embryos compared with adult fish (Table 3). Other plasma parameters in embryos showed similar levels to those of adult fish. The urine osmolality of embryos was nearly equal to plasma osmolality. Meanwhile, divalent ions, such as Mg²⁺ and SO₄²⁻, were highly concentrated in the embryonic urine; concentrations of Mg²⁺ and SO₄²⁻ in the embryonic urine were >100 times higher than those in plasma. Although the Ca²⁺ concentration in the embryonic urine was significantly higher than that in the plasma, the concentration factor of Ca²⁺ (cf = 3.3) was considerably lower than that of Mg²⁺ (cf = 228) and SO₄²⁻ (cf = 194). The concentrations of monovalent ions Na⁺ and Cl⁻ in the urine of embryos were significantly lower than those in plasma.

Table 3.

Plasma and urine parameters in elephant fish embryos and adult

		Ions, mmol/l
Stage		Na+	Cl−	Mg2+	SO42−	Ca2+	Urea, mmol/l	Osmolality, mOsmol/kg
st.36	plasma	298.7 ± 4.2 (12)*	314.7 ± 8.5 (12)	1.3 ± 0.1 (12)	1.6 ± 0.1 (15)	3.9 ± 0.1 (12)	502.9 ± 7.5 (12)	1048.9 ± 3.1 (12)*
	urine	258.9 ± 43.0 (12)†	123.2 ± 22.9 (15)†	296.4 ± 17.7 (12)†	310.3 ± 9.3 (15)†	12.8 ± 1.9 (12)†	218.6 ± 15.2 (12)†	1055.8 ± 2.3 (12)
adult	plasma	394.0 ± 10.0 (5)‡	308.2 ± 4.3 (5)	2.3 ± 0.4 (5)	2.5 ± 0.4 (5)	3.6 ± 0.1 (5)	448.7 ± 8.8 (5)‡	1073.6 ± 3.4 (5)‡

Values are expressed as means ± SE.

^*Data are from Ref. 57.

^†A significant difference was observed between embryo plasma and urine (P < 0.05).

^‡A significant difference was observed between plasma of embryos and adult fish (P < 0.05).

Quantitative real-time PCR showed that cmSlc26a1 and cmSlc26a6 mRNAs were abundantly expressed in the kidney of embryos (Fig. 9). Furthermore, the tissue distribution pattern of cmSlc26a1 and cmSlc26a6 mRNAs in embryos (Fig. 9) was similar to that in adult fish (Fig. 4).

Fig. 9.

Embryonic tissue distribution of mRNAs. RNAs encoding cmSlc26a1 (A) and cmSlc26a6 (B) were detected in stage 36 embryos. The mRNA values were normalized against those of cmEF1α mRNA. Data are presented as means ± SE; n = 6.

DISCUSSION

Since seawater is a high, SO₄²⁻ environment, marine fishes need to excrete SO₄²⁻ from their kidneys to avoid hypersulfatemia. In the present study, we determined cDNA sequences from the kidney of the holocephalan elephant fish (Callorhinchus milii) that encoded the entire ORF of SO₄²⁻ transporters Slc26a1 and Slc26a6. The molecular phylogenetic and motif analyses revealed that cmSlc26a1 and cmSlc26a6 showed high similarity to Slc26a1 and Slc26a6, respectively, of other vertebrates. Functional and immunohistochemical investigations revealed that cmSlc26a1 is a basolateral SO₄²⁻ transporter, while cmSlc26a6 is an electrogenic Cl⁻/SO₄²⁻ exchanger on the apical membrane. Both cmSlc26a1 and cmSlc26a6 were colocalized in the PII segment of the nephron, suggesting that this segment is the site for the secretion of excess SO₄²⁻ in the highly elaborate nephron of cartilaginous fish.

cmSlc26a6 is an electrogenic Cl⁻/SO₄²⁻ exchanger.

In teleosts, three types of Slc26a6 (Slc26a6A, Slc26a6B, and Slc26a6C) have been found (25, 39). The Slc26a6A protein is expressed only in SW, suggesting that Slc26a6A is a candidate for the major apical SO₄²⁻ transporter that mediates SO₄²⁻ secretion in the kidneys of marine teleosts (25, 62). In the eel, Slc26a6A and Slc26a6B were expressed in the separate segments of the proximal tubule (62). In the elephant fish genome database, however, we found only a single gene encoding Slc26a6. Molecular phylogenetic analysis revealed that the three Slc26a6 genes in teleosts arose from duplication in the teleost lineage, supporting our finding that only single Slc26a6 gene exists in elephant fish.

The functional features of cmSlc26a6 showed high similarity to those of teleost and tetrapod Slc26a6 proteins, respectively. In mammals and teleosts, Slc26a6 mediates electrogenic exchange of various anions, including Cl⁻/SO₄²⁻ (25, 63), Cl⁻/oxalate, Cl⁻-OH⁻/HCO₃⁻, and electroneutral Cl⁻/formate (1). In our in vitro assay with radiolabeled SO₄²⁻, SO₄²⁻ uptake by the oocytes expressing cmSlc26a6 was induced only when Cl⁻ was removed from the assay medium. The SO₄²⁻ uptake was abolished by the use of Cl⁻-containing assay buffer or the addition of DIDS, an inhibitor of anion transport. These results imply that cmSlc26a6 acts as a Cl⁻/SO₄²⁻ exchanger using the Cl⁻ concentration gradient as a driving force. Furthermore, depolarization of oocytes by replacing Na⁺ with K⁺ significantly facilitated SO₄²⁻ uptake, suggesting that cmSlc26a6 is an electrogenic Cl⁻/SO₄²⁻ exchanger. These conclusions were further supported by the electrophysiological data. Specifically, the presence and absence of extracellular SO₄²⁻ and Cl⁻, respectively, caused a reduction in intracellular Cl⁻ activity and hyperpolarization. Depolarization and recovery of intracellular Cl⁻ activity were abruptly elicited by readdition of Cl⁻, implying the occurrence of Cl⁻ influx and SO₄²⁻ efflux. The observed characteristics of cmSlc26a6 are appropriate for SO₄²⁻ efflux across the plasma membrane, as extracellular Cl⁻ and negative membrane potential act as driving forces for the SO₄²⁻ secretion, as previously reported for marine teleosts (25).

cmSlc26a1 is an electroneutral SO₄²⁻ transporter.

The ion transport function of cmSlc26a1 is distinct from that of cmSlc26a6 as a significant SO₄²⁻ uptake activity was only detected in the presence of Cl⁻ in the assay buffer. When Cl⁻ was removed from the assay medium, only a minor SO₄²⁻ uptake was observed. These results imply that cmSlc26a1 is a SO₄²⁻ transporter but not a Cl⁻/SO₄²⁻ exchanger, and that Cl⁻ is required to exert SO₄²⁻ transporting activity. These conclusions were supported by the results of the electrophysiology experiments in which no change was observed in intracellular Cl⁻ activity in response to exposure to the medium containing SO₄²⁻, irrespective of Cl⁻ concentration in the assay medium. A similar effect of extracellular Cl⁻ on SO₄²⁻ transport was reported for the Slc26a1 of mammals and teleosts (33, 39, 46, 50, 63). For the rat transporter, SO₄²⁻ uptake by oocytes expressing Slc26a1 is increased only in the presence of extracellular Cl⁻ (50), while for the mouse and eel, only moderate levels of SO₄²⁻ transport are induced by Slc26a1, even in the absence of extracellular Cl⁻ (39, 63). Other monovalent anions, such as other halides, formate, and lactate, have also been reported to enhance SO₄²⁻ transport via Slc26a1 (63). However, a mechanism of SO₄²⁻ uptake activated by monovalent anions remains unknown.

Depolarization of oocytes (K⁺ medium) had no effect on SO₄²⁻ uptake in the oocytes expressing cmSlc26a1, suggesting that cmSlc26a1 is an electroneutral SO₄²⁻ transporter. In the electrophysiological assay, we detected no change in the membrane potential of oocytes expressing cmSlc26a1 in the presence of extracellular SO₄²⁻, supporting this conclusion. Currently, without knowing the exchanged anion or cotransported cation, the mode of SO₄²⁻ transport by cmSlc26a1 remains to be clarified. In the mammalian nephron, previous studies have demonstrated that SO₄²⁻ transport in basolateral membrane vesicles occurs via a SO₄²⁻/HCO₃⁻ exchanger (30, 44). Furthermore, rat Slc26a1 was shown to be localized on the basolateral membrane, where it functioned as an electroneutral SO₄²⁻/2HCO₃⁻ or oxalate²⁻/2HCO₃⁻ exchanger (24, 29). Further investigation is needed to clarify the ion(s) exchanged by cmSlc26a1 as a counterpart of SO₄²⁻.

Identification of the SO₄²⁻ secretory segment in the elephant fish nephron.

Elasmobranch kidneys are unusually complicated structures compared with other vertebrate kidneys, as they comprise multiple lobules, and each lobule is divided into two zones: a sinus zone and a bundle zone. Each nephron has an elaborate four-loop configuration (Fig. 1) and traverses repeatedly between the two zones (15, 22, 32). The in situ hybridization experiments of the present study revealed that the expression of both cmSlc26a1 and cmSlc26a6 mRNAs was in the sinus zone, but not in the bundle zone. The sinus zone of the elephant fish kidney mostly comprises two nephron segments: the PII (the 2nd loop of nephron) and the LDT (the 4th loop of nephron) (22). The PII and LDT are easily distinguishable by their morphological characteristics; the PII segment has the largest tubular diameter with an extensive brush border on the apical membrane, while the LDT shows a relatively small diameter and thin epithelial cells without a brush-bordered apical membrane (22, 32). The morphological features of the tubule that showed both cmSlc26a1 and cmSlc26a6 mRNA expression were those of the PII segment and not the LDT. This assessment is further confirmed by the comparison with cmNKCC2 mRNA signals, as our previous research has shown that cmNKCC2 mRNA is expressed in two separate diluting segments, namely, EDT in the bundle zone and the posterior half of the LDT in the sinus zone (22). A comparison using serial sections clearly showed that the cmSlc26a1 and cmSlc26a6 mRNA hybridization signals did not overlap with the cmNKCC2 signal. These observations strongly imply that in the elephant fish kidney, the secretory segment for SO₄²⁻ is the PII segment and not any other segment. The present result is consistent with a previous micropuncture study, in which the existence of a secretory segment for divalent ions was suggested to be the second loop of the little skate nephron (54).

In Fig. 10, we propose a hypothetical model for the secretion of SO₄²⁻ in the elephant fish kidney. In the PII segment, cmSlc26a1 (an electroneutral SO₄²⁻ transporter) and cmSlc26a6 (an electrogenic Cl⁻/SO₄²⁻ exchanger) were localized on the basolateral and apical membranes, respectively (Fig. 8, A and B). It is highly probable that the negative membrane potential drives SO₄²⁻ efflux via apically located cmSlc26a6 into the filtrate. Consistent with our notion, intense Na⁺-K⁺-ATPase (NKA) signals have been detected in the elephant fish kidney in several nephron segments, including the PII segment (22). The intracellular SO₄²⁻ is supplied from the blood sinus via a basolaterally located cmSlc26a1. Further investigations on ionic concentrations in the urine and the PII cells are necessary to confirm our model, but a similar system for SO₄²⁻ secretion has been proposed in the proximal tubule of marine teleost fishes (25, 62).

Fig. 10.

Schematic diagrams showing the elephant fish nephron and PII sulfate secretion. Localization of Slc26a1, Slc26a6, NKA, and NKCC2 is indicated in the diagram of elephant fish nephron. The upper left corner provides a hypothetical model for epithelial SO₄²⁻ secretion in the PII segment. cmSlc26a6 is an electrogenic Cl⁻/SO₄²⁻ exchanger localized on the apical membrane. The negative membrane potential and high concentration of Cl⁻ in the filtered urine most probably drives SO₄²⁻ secretion into the filtrate via Slc26a6. The intracellular SO₄²⁻ is supplied via basolaterally located cmSlc26a1 (electroneutral SO₄²⁻ transporter) from the blood sinus. The dashed line means that we have no direct evidence for SO₄²⁻/anion exchange or SO₄²⁻-cation cotransport for cmSlc26a1. The encircled numbers represent the number of loops. AM, apical membrane; BLM, basolateral membrane; CV, central vessel; NKA, Na⁺-K⁺-ATPase; RC, renal corpuscle; PS, peritubular sheath, Slc, solute carrier family.

In addition to the negative membrane potential, the concentration gradient for Cl⁻ is considered to be an important driving force for SO₄²⁻ secretion via cmSlc26a6. For secretion of SO₄²⁻ into the filtrate via the apically located cmSlc26a6 to occur, a high concentration of Cl⁻ in the filtrate is required to facilitate Cl⁻/SO₄²⁻ exchange. In this regard, the PII segment is well located to contribute to SO₄²⁻ secretion for the following reason. In the mammalian kidney, it is well recognized that the apical NKCC2 and the basolateral NKA contribute to active reabsorption of Na⁺ and Cl⁻ in the thick ascending limb of the loop of Henle (12, 14). In elephant fish, we found coexpression of cmNKA α1 and cmNKCC2 mRNAs in the EDT and posterior half of the LDT, implying that these distal segments contribute to Na⁺ and Cl⁻ reabsorption and the resulting dilution of the filtrate (22). Since the filtrate passes through the PII segment of the nephron prior to the diluting segments, it is highly probable that the concentration of Cl⁻ is still high in the PII segment, which would facilitate SO₄²⁻ secretion.

In Table 4, we estimated a driving force for SO₄²⁻ efflux via cmSlc26a6 as the ratio [SO₄²⁻]_out/[SO₄²⁻]_in, following the calculation by Kato et al. (25) in SW mefugu. The calculation was performed using the following equation (Table 4): Δμ_Slc26a6 = Δμ_Cl-SO4²⁻ = Δμ_Cl − Δμ_SO4²⁻ = RT × ln([Cl⁻]_in/[Cl⁻]_out) + (−1) × FV_m − {RT × ln([SO₄²⁻]_in/[SO₄²⁻]_out) + (−2) × FV_m}, where R is the gas constant, T is the absolute temperature, F is the Faraday constant, ln is the natural log, V_m is the membrane potential, and Δμ_ion is the electrochemical potential difference (J/mol). At the point of equilibrium, the calculated [SO₄²⁻]_out/[SO₄²⁻]_in values are 366.3, 246.5, and 165.9 when V_m is −80, −70, and −60 mV, respectively, under the following assumptions: 1) the stoichiometry of Cl⁻/SO₄²⁻ exchange is 1:1; 2) [Cl⁻]_out is 308.2 mM, which is the plasma [Cl⁻] of adult elephant fish; 3) [Cl⁻]_in is 20 mM, which is similar to the cytosolic [Cl⁻] of mammalian proximal tubular cells (5, 19, 28); and 4) the temperature is 293 K (20°C). These values were approximately two times higher than those calculated for SW mefugu (25). The ratio was >100, even when we assumed a higher cytosolic [Cl⁻]. Although it will be necessary to measure the ion concentration profiles of the intracellular fluid of the PII cells and of the filtrate flowing into the PII segment, the estimated values of the driving force suggest that cmSlc26a6 has the ability to secrete SO₄²⁻ under conditions that are relevant to the predicted composition of the fluid in the PII segment.

Table 4.

Calculations of [SO₄²⁻]_out/[SO₄²⁻]_in under various conditions of [Cl⁻]_in, [Cl⁻]_out, and V_m at the equilibria (Δμ_cmSlc26a6 = 0)

Vm, mV	[Cl−]in, mM	[Cl−]out, mM	[SO42−]out/[SO42−]in
−50	20	308.2	111.6
−60	20	308.2	165.9
−70	20	308.2	246.5
−80	20	308.2	366.3
−90	20	308.2	544.3
−70	20	308.2	246.5
−70	30	308.2	164.3
−70	40	308.2	123.3

As mentioned above, the localization of cmSlc26a1 and cmSlc26a6 in the PII segment was similar to that of Slc26a1 and Slc26a6A in the proximal tubule of marine teleost fishes, implying that the mechanisms by which elephant fish secrete SO₄²⁻ are similar to teleost fishes. Basolateral Slc26a1 and apical Slc26a6 have also been found in the mammalian proximal tubule (35), but in the mammalian kidney, net absorption of SO₄²⁻ occurs. An investigation using Slc26a6-null mice revealed that Slc26a6 is important for Cl⁻/base exchange in the apical membrane of the proximal tubule (61). In the mammalian kidney, as well as the FW teleost kidney (25, 39, 62), another apically located sulfate transporter, Slc13a1, is thought to contribute to SO₄²⁻ reabsorption via a Na⁺ coupled mechanism (37).

Environmental adaptation during development is an important aspect of research in animal homeostasis (23). In oviparous cartilaginous fishes, including the elephant fish, the embryos are enclosed in a tough and fibrous egg capsule in SW for the entire developmental period (∼6 mo). Although the egg capsule is important for protection from predation, it does not isolate the intracapsular ionic environment from that of the external SW. Previously, we demonstrated that the osmolality and ionic composition of the egg capsule fluid are similar to those of SW throughout the development of elephant fish (57). Despite these findings, this study found that the concentration of plasma Mg²⁺ and SO₄²⁻ were maintained at levels 10 times lower than those of SW (Mg²⁺, 50 mM; SO₄²⁻, 30 mM), implying that embryos, like adult fish, must regulate plasma Mg²⁺ and SO₄²⁻. In addition, we discovered a urinary bladder-like structure in the peri-hatching embryos (this study), which contained highly concentrated MgSO₄. We found that the kidney of elephant fish embryos expressed cmSlc26a1 and cmSlc26a6 mRNA, suggesting that the tubular secretion of SO₄²⁻ may already function in the kidney of elephant fish embryos. Surprisingly, the concentrations of Mg²⁺ and SO₄²⁻ in the embryonic urine were considerably higher than those reported in urine of adult elasmobranchs and holocephalans (4, 45, 52). Further studies are necessary to compare urine composition between adults and embryos of elephant fish, and to determine mechanisms by which elephant fish embryos can concentrate Mg²⁺ and SO₄²⁻ in their urine.

Perspectives and Significance

In the 1960s, it was revealed that the cartilaginous fish nephron had an extraordinarily, elaborate four-loop configuration (3). To understand the function of each nephron segment, renal micropuncture and microperfusion studies have been performed (8, 11, 51, 54). In addition to these physiological investigations, mapping of the membrane transporters in the nephron has revealed the possible function of each nephron segment, including urea reabsorption, NaCl reabsorption, glucose reabsorption, and urea synthesis (17, 22). The present study provides further information on nephron function in cartilaginous fish, as we have shown that the proximal segment II in the sinus zone of the elephant fish kidney is the site of SO₄²⁻ secretion. Future research will ascertain whether the PII segment is also important in the secretion of other divalent ions such as Mg²⁺ (20).

In euryhaline teleost fishes, the kidneys switch their role in SO₄²⁻ homeostasis from SO₄²⁻ absorption to secretion, and vice versa, depending on the salinity of their habitat (25, 62). Although most cartilaginous fishes inhabit marine environments, some cartilaginous fishes, e.g., bull shark and Atlantic stingray, are euryhaline. Thus, it would be interesting in future investigations to consider the role of the kidney in divalent ion homeostasis in euryhaline cartilaginous fish species.

GRANTS

This study was supported by Grant-in-Aid for Scientific Research (B) (Grant 26291065) and for Challenging Exploratory Research (Grant 26650110), and a Japan-Australia Research Cooperative Program (References 10034011-000062 and 13039901-000237) from the Japan Society for the Promotion of Science (JSPS) to S. Hyodo, Grant-in-Aid for Scientific Research (B) (Grant 26292113) from JSPS to A. Kato, and the National Institutes of Health (Grant DK-092408) to M. F. Romero.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

K.H., T.W., and S.H. conception and design of research; K.H., A.K., T.W., W.T., M.F.R., J.D.B., T.T., J.A.D., and S.H. performed experiments; K.H., A.K., T.W., M.F.R., and S.H. analyzed data; K.H., A.K., T.W., M.F.R., and S.H. interpreted results of experiments; K.H. and A.K. prepared figures; K.H. drafted manuscript; K.H., A.K., T.W., W.T., M.F.R., T.T., J.A.D., and S.H. edited and revised manuscript; K.H., A.K., T.W., W.T., M.F.R., J.D.B., T.T., J.A.D., and S.H. approved final version of manuscript.