The European and Japanese eel NaCl cotransporters β exhibit chloride currents and are resistant to thiazide type diuretics

Abstract

The thiazide-sensitive Na⁺-Cl⁻ cotransporter (NCC) is the major pathway for salt reabsorption in the mammalian distal convoluted tubule, and the inhibition of its function with thiazides is widely used for the treatment of arterial hypertension. In mammals and teleosts, NCC is present as one ortholog that is mainly expressed in the kidney. One exception, however, is the eel, which has two genes encoding NCC. The eNCCα is located in the kidney and eNCCβ, which is present in the apical membrane of the rectum. Interestingly, the European eNCCβ functions as a Na⁺-Cl⁻ cotransporter that is nevertheless resistant to thiazides and is not activated by low-chloride hypotonic stress. However, in the Japanese eel rectal sac, a thiazide-sensitive NaCl transport mechanism has been described. The protein sequences between eNCCβ and jNCCβ are 98% identical. Here, by site-directed mutagenesis, we transformed eNCCβ into jNCCβ. Our data showed that jNCCβ, similar to eNCCβ, is resistant to thiazides. In addition, both NCCβ proteins have high transport capacity with respect to their renal NCC orthologs and, in contrast to known NCCs, exhibit electrogenic properties that are reduced when residue I172 is substituted by A, G, or M. This is considered a key residue for the chloride ion-binding sites of NKCC and KCC. We conclude that NCCβ proteins are not sensitive to thiazides and have electrogenic properties dependent on Cl⁻, and site I172 is important for the function of NCCβ.

INTRODUCTION

The thiazide-sensitive NaCl cotransporter (NCC) is the major NaCl reabsorption pathway in the distal convoluted tubule of the nephron and is the target of thiazide-type diuretics (1). Mutations in the SLC12A3 gene that encodes the human NCC reduce its activity and are the cause of salt- and potassium-wasting disease known as Gitelman syndrome (OMIM 263800) (2, 3). This disease features arterial hypotension accompanied by hypokalemic metabolic alkalosis and hypercalciuria. However, increased activity of NCC due to mutations in regulatory genes such as the WNK1 and WNK4 kinases or the ubiquitin ligase proteins KLHL3 and Cul3 produces a form of salt-sensitive hypertension characterized by hyperkalemia and metabolic acidosis known as familial hyperkalemic hypertension or Type II pseudohypoaldosteronism (OMIM 145260) (4–6).

We have extensively studied the physiological, biochemical, and pharmacological characteristics of NCC orthologs (7–12). NCCs consist of a twelve-segment transmembrane domain (TM), an extracellular domain (ECD) located between TM 7 and 8 with the presence of glycosylation sites, a N-terminal domain (NTD), and a C-terminal domain (CTD) located intracellularly (Fig. 1). We first observed clear differences in the ion transport and thiazide inhibition kinetics between the mammalian and flounder NCC orthologs (7–9). Then, by constructing chimeric proteins between rat NCC (rNCC) and flounder NCC (flNCC) and/or point mutations in the wild-type or chimeric transporters, we determined specific amino acids related to ion transport kinetics or diuretic binding. In particular, we observed that the difference in thiazide affinity between rNCC and flNCC is due to residues in the TM 8 and 12 (9) and established the particular importance of the amino acid residue cysteine in position 575 in the rNCC (serine in flNCC) located in the TM 11 region as responsible for the difference in affinity for thiazides between flounder and mammalian NCC (11). In addition, the characterization of a single-nucleotide polymorphism (SNP) in NCC (10) revealed that a glycine in transmembrane segment (TM) 4 plays an important role in chloride affinity. However, our knowledge of the structure-function relationship regarding ion cotransport or thiazide-binding sites has been limited by the lack of tertiary and quaternary structures of the NCC protein (13), and thus, specific sites for ion or thiazide binding have not yet been revealed. Recent studies in other members of the SLC12 family, such as the Na-K-2Cl cotransporter NKCC1 and the K-Cl cotransporters KCC1 and KCC2, suggested key residues constituting the ion-binding sites that need to be explored in NCC (14–16). As reported previously, NKCC1 presents two chloride binding sites. The first chloride-binding site (SCL1) is formed by the amino acids GVM located in sites 223–225 in TM1. The second chloride binding site (SCL2) includes the amino acids GIL located in sites 421–423 in TM2 and T611 in TM10 (17).

Figure 1.

Model of the secondary structure of NCC containing the 13 amino acid changes between eNCCβ and jNCCβ. A: the NCC secondary structure is depicted, with a central hydrophobic domain containing 12 putative transmembrane segments and an extracellular glycosylated loop facing the extracellular side of the cell located between transmembrane segments 7 and 8. The black dots depict the localization of the nonconserved amino acid residues between eNCCβ and jNCCβ. B: the 13 changes made and the amino acids that were substituted from eNCCβ to jNCCβ are shown. eNCCβ, European eel NCCβ; jNCCβ, Japanese eel NCCβ; NCC, Na⁺-Cl⁻ cotransporter.

In contrast to mammalians and the flounder, the European eel, Anguilla europea (eNCC), and the Japanese eel, Anguilla japonica (jNCC) eels (class Actinopterygii, order Anguilliformes, family Anguillidae), have two genes that encode NCC proteins (18, 19). The NCC expressed in the kidney (NCCα) has 65%–70% identity with mammalian and flounder NCCs, whereas a paralog gene, designated NCCβ, is expressed in the gut of eels and exhibits 55% identity with mammalian NCC genes. We recently performed functional characterization of the intestinal isoform of the European eel NCCβ (eNCCβ) (12). We observed that eNCCβ is indeed a Na-Cl cotransporter that is not sensitive to thiazides. However, Watanabe et al. (19) reported a hydrochlorothiazide-sensitive NaCl transport mechanism in rectal sac preparations of the Japanese eel, where jNCCβ is expected to be present. In this regard, eNCCβ and jNCCβ are 98% identical; the difference lies in 13 amino acids throughout their sequence. In our previous work (12), substitution of cysteine 379 with serine conferred partial sensitivity to thiazides in eNCCβ. Thus, we substituted all 13 residues to convert eNCCβ into jNCCβ to explore the sensitivity to the diuretic. In addition, during the functional characterization of eNCCβ, we noticed that in oocytes injected with eNCCβ cRNA and exposed to ²²Na⁺-free medium, ³⁶Cl⁻ uptake was reduced by 80%, leaving 20% activity in the absence of Na⁺, suggesting that eNCCβ could have electrogenic properties, which we have characterized in the present work.

EXPERIMENTAL METHODS

jNCCβ Cloning

We previously reported the cloning of eNCCβ (12) and analyzed the structural differences between eNCCβ and jNCCβ (19). The difference between these two orthologs is 13 amino acids located along the sequence. To obtain jNCCβ cDNA, we made an additional 13 point mutations in eNCCβ cDNA that we previously studied (12) using custom-made primers.

In Vitro cRNA Translation

To prepare cRNA for microinjection, rNCC, flNCC, sNCC, eNCCβ, jNCCβ, and rat NKCC2 cDNAs were digested at the 3'-end using Nhe I from Invitrogen (Carlsbad, CA), and cRNA was transcribed in vitro using the T7 RNA polymerase mMESSAGE mMACHINE (Ambion) transcription system. cRNA product integrity was confirmed on agarose gels, and concentration was determined by absorbance reading at 260 nm (DU 640; Beckman, Fullerton, CA). cRNA was stored frozen in aliquots at −80°C until use.

Xenopus Laevis Oocyte Preparation

Oocytes were harvested surgically from adult female Xenopus laevis frogs (Nasco) under 0.17% tricaine anesthesia and incubated in ND96 (in mM: 96 NaCl, 2 KCl, 1.8 CaCl₂, 1 MgCl, and 5 HEPES-Tris, pH 7.4) in the presence of collagenase B (2 mg/mL) for 1 h. After four washes in ND96, the oocytes were manually defolliculated and incubated at 16°C in ND96 supplemented with 2.5 mM sodium pyruvate and 5 mg/100 mL gentamicin overnight. The next day, stage V–VI oocytes were injected with 50 nL of water or 10–20 ng cRNA/oocyte. Then, the oocytes were incubated for 2 or 3 days in ND96 with sodium pyruvate and gentamicin, which were changed every 24 h (7–12) The protocol was approved by the Institutional Animal Care and Use Committee (IACUC).

Assessing ²²Na⁺ Tracer Uptake

A 30-min incubation in a Cl⁻-free ND96 medium containing 1 mM ouabain, 0.1 mM amiloride, and 0.1 mM bumetanide was followed by a 60-min uptake period in a K⁺-free, NaCl^hx00A0;− containing medium with ouabain, amiloride, bumetanide, and 1 μCi of ²²Na⁺/mL. The affinity for thiazide diuretics was assessed by exposing groups of cRNA-injected oocytes to concentrations of drugs from 10⁻⁸ to 10⁻³ M. rNCC and flNCC were used as control signals.

To determine the ion transport kinetics of jNCCβ, we performed experiments varying the concentrations of Na⁺ and Cl⁻. To maintain osmolarity and ionic strength, gluconate was used as a Cl⁻ substitute and N-methyl-d-glucamine as a Na⁺ substitute. jNCCβ was subjected to at least four different ion transport kinetic experiments with each set of solutions (7–10, 12).

Assessing ⁸⁶Rb⁺ Tracer Uptake

A 30-min incubation in K⁺- and Cl⁻-free medium with 1 mM ouabain was followed by a 60-min uptake period in the presence of Na⁺, K⁺, and Cl⁻ medium containing 1 mM ouabain and the absence or presence of furosemide 10⁻⁴ M. Rat NKCC2 was used as a control signal.

Assessing ³⁶Cl⁻ Tracer Uptake

A 30-min of incubation in Cl⁻-free ND96 medium containing 1 mM ouabain, 0.1 mM amiloride, and 0.1 mM bumetanide was followed by a 60-min uptake period in K⁺-free NaCl ⁻ containing medium with ouabain, amiloride, bumetanide, and 1 μCi of ³⁶Cl⁻/mL. To determine the Na⁺ dependency of the ³⁶Cl⁻ transport of jNCCβ, ³⁶Cl⁻ uptake was assessed in parallel groups using a K⁺- and Na⁺-free Cl ⁻ containing medium (12).

All uptakes were performed at 32°C. At the end of the uptake period, oocytes were washed five times in ice-cold uptake solution without isotopes to remove extracellular fluid tracers. After the oocytes were dissolved in 10% sodium dodecyl sulfate, tracer activity was determined for each oocyte by β-scintillation counting (12, 20)

Two-electrode Voltage Clamp

Oocyte membrane currents were recorded using an OC-720C voltage clamp (Warner Instruments, Hamden, CT) filtered at 2–5 kHz, digitized at 10 kHz, and recorded with PATCH MASTER software (HEKA, Germany); data were analyzed as previously described (21, 22). For periods when I–V protocols were not being run, the oocytes were clamped at a holding potential (V_h) of −50 mV, and the current was monitored and recorded. I–V protocols consisted of 100 ms at V_h followed by 500 ms of 20-mV steps from V_h to −200 mV and +80 mM ending with 100 ms at V_h. The I–V protocols were run in the presence of K⁺-free ND96 solution (mM: 96 NaCl, 1.8 CaCl, 1 mM MgCl, 5 mM HEPES, pH 7.5), Cl⁻-free ND96 (mM: 86 Na gluconate, 10 NaCl, 4.8 Ca gluconate, 1.0 Mg gluconate, 5 HEPES, pH 7.5), and Na⁺-free ND96 (mM: 96 choline chloride or 96 NMDG chloride or 96 LiCl, 1.8 CaCl, 1 mM MgCl, 5 HEPES, pH 7.5). All of the experiments were performed at room temperature. The oocytes were bathed in the test solution for 1–3 min before the I–V protocol was run.

Western Blotting

Western blot analysis was used to assess the protein expression of FLAG-jNCCβ-injected oocytes. Proteins extracted from 50 oocytes were quantified by Bradford’s technique, and 50 μg of each protein per lane was run using sample buffer containing 6% SDS, 15% glycerol, 0.3% bromophenol blue, 150 mM Tris, pH 7.6, and β-mercaptoethanol, resolved by Laemmli SDS-polyacrylamide (7.5%) gel electrophoresis, and transferred to a polyvinylidene difluoride membrane. Immunoblotting was performed using anti-FLAG monoclonal antibody (Sigma A8592 monoclonal ANTI-FLAG M2 peroxidase) (12). Membranes were exposed to anti-FLAG antibody overnight at 4°C and washed again. Bands were detected by using Immun-Star Chemiluminescent Protein Detection Systems (Bio-Rad).

Cell Surface Biotinylation of X. Laevis Oocytes

Oocytes were injected with cRNA encoding FLAG-jNCCβ and washed five times in ND-96 TEA buffer (in mM: 96 NaCl, 2 KCl, 1.8 CaCl₂, 1 MgCl₂, 5 HEPES, pH 8.8, and 10 TEA) and incubated for 30 min with 1.5 mg/mL Sulfo-NHS-LC-Biotin (Thermo; Pierce) in ice-cold ND-96-TEA. Oocytes were then washed five times in ND-96-TEA buffer and homogenized using a 25-gauge needle in a sucrose-based buffer (5 μL/oocyte) consisting of 250 mM sucrose, 0.5 mM EDTA, 5 mM Tris·HCl, pH 6.9, 1 mM phenylmethylsulfonyl fluoride (PMSF), and 10 μL/mL protease inhibitor cocktail (P8340; Sigma, St. Louis, MO). The samples were centrifuged for 7 min at 8,000 rpm, the supernatant was collected, and the protein concentration was assessed utilizing a Bradford assay (Bio-Rad). Streptavidin precipitation was carried out by adding 75 μL of streptavidin-agarose beads in 50% slurry (Upstate Cell Signaling Solutions, Merck) to 400 μg of biotinylated total protein diluted in 1 mL of Tris-buffered saline (100 mM NaCl, 50 mM Tris·HCl, pH 7.4). Samples were rolled overnight at 4°C. The beads were then washed one time with buffer A (in mM: 5 EDTA, 50 NaCl, 50 Tris·HCl, pH 7.4), twice with buffer B (500 mM NaCl, 20 mM Tris·HCl, pH 7.4), and once with buffer C (10 mM Tris·HCl, pH 7.4) with a 2-min spin at 5,000 g between each wash. After the last wash, buffer C was substituted with 30 μL of Laemmli sample buffer with 5% 2-mercaptoethanol (Sigma, Bio-Rad). Protein samples were heated to 65°C for 15 min before separation on a 7.5% polyacrylamide gel (23).

Statistical Analysis

The results are presented as the means ± SE, and significance was defined as two-tailed P < 0.05. The significance of the differences between two groups was tested by Student’s t test, and significance for three or more groups was tested by one-way ANOVA with multiple comparisons using Bonferroni correction.

RESULTS

Cloning of jNCCβ cDNA by Site-Directed Mutagenesis

Watanabe et al. (19) cloned and studied jNCCβ. In rectal sac preparations, they observed that hydrochlorothiazide at a concentration of 1 mM significantly reduced the absorption of Na⁺, Cl⁻, and water, which suggested that jNCCβ was sensitive to thiazide. We previously demonstrated that the presence of different thiazides at a concentration of 100 μM did not achieve an inhibitory effect on oocytes microinjected with eNCCβ cRNA, but under similar conditions, a reduction of ∼40% was observed in oocytes injected with eNCCβ-C379S-S194F-A482G cRNA (12) in which the residues C379, S194, and A482 were changed to serine, phenylalanine, and glycine, respectively, as occurs in jNCCβ. These findings led us to consider whether the rest of the differences between eNCCβ and jNCCβ that were not evaluated in our previous study might lend full thiazide sensitivity to eNCCβ, which had been converted to jNCCβ.

As shown in Fig. 1A, the structural differences between eNCCβ and jNCCβ are due to 13 amino acids along their sequence. Of these, six are located in the amino-terminal region, three in the TM regions, and four in the carboxyl-terminal region. In a previous study (12) by point mutagenesis, we made the abovementioned three amino acid changes that are different in the TM domains between eNCCβ and jNCCβ (eNCCβ-C379S-S194F-A482G). Following the same methodology, we introduced the 10 missing differences between the eNCCβ-C379S-S194F-A482G cDNA and the jNCCβ (Fig. 1B): a glycine at position 7 was substituted by a serine (G7S); a serine at position 32 by an asparagine (S32N); a leucine at position 50 by a proline (L50P); a valine at position 87 by a leucine (V87L); an alanine at position 131 by a threonine (A131T); a glutamine at position 140 by a histidine (Q140H); an alanine at position 651 by a serine (A651S); a threonine at position 720 by a lysine (T720K); a glutamine at position 821 by a leucine (Q821L); and finally an isoleucine at position 937 by a valine (I937V). The substitutions were made one by one until the final clone was obtained and sequenced to ensure that no unwanted mutations were introduced. The resulting jNCCβ sequence was identical to that reported by Watanabe et al. (19).

Functional Properties of Residues in eNCCβ That Generate jNCCβ

In Fig. 2, we compared the effect of hydrochlorothiazide in oocytes microinjected with eNCCβ, eNCCβ-C379S-S194F-A482G, and jNCCβ. We evaluated ²²Na⁺ influx in the presence or absence of extracellular Cl⁻ and in the presence of hydrochlorothiazide at a concentration of 100 μM. For these experiments, the data obtained from wild-type eNCCβ were set as 100%, and the other groups were normalized accordingly. As previously demonstrated (12), we observed that injection of oocytes with eNCCβ cRNA induced an increase in ²²Na⁺ influx that was Cl⁻ dependent but not sensitive to hydrochlorothiazide. eNCCβ-C379S-S194F-A482G was also chloride-dependent, but thiazide elicited ∼40% inhibition. Interestingly, this partial sensitivity was lost with the introduction of the remaining difference between eNCCβ and jNCCβ. As shown in Fig. 2, similar to eNCCβ, injection of ooctyes with jNCCβ induced a chloride-dependent Na⁺ uptake pathway that was not inhibited with hydrochlorothiazide. These results indicate that like the β-isoform of European eel, the β-isoform of Japanese eel is not sensitive to this thiazide.

Figure 2.

Effects of eNCCβ, eNCCβ-C379S-S194F-A482G, and jNCCβ on thiazide sensitivity. By point-directed mutagenesis, 10 amino acid substitutions were made on the cDNA of eNCCβ-C379S-S194F-A482G. Each substitution was made randomly and was added to the clone until jNCCβ was obtained. The changes that were made were A651S, V87L, S32N, Q140H, L50P, A131T, G7S, T720K, I937V, and Q821L. The mutations were confirmed by DNA sequencing. The oocytes were injected with cRNA from wild-type eNCCβ, from mutant clones, and from cRNA from jNCCβ. Three days later, the influx of Na⁺ was evaluated under control conditions (white columns), in the absence of Cl⁻ or in the presence of 100 µm hydrochlorothiazide (black columns). The data observed for each control group were taken as 100%, and the thiazide groups were normalized accordingly. The influx in the absence of thiazides for each group versus the influx under control conditions in the same group was similar except for eNCCβ-C379S-S194F-A482G (*P < 0.0001). eNCCβ, European eel NCCβ; jNCCβ, Japanese eel NCCβ; NCC, Na⁺-Cl⁻ cotransporter.

Expression and Functional Properties of jNCCβ

Xenopus laevis oocytes microinjected with 5 ng of FLAG-jNCCβ cRNA produced by in vitro transcription from jNCCβ cDNA expressed a protein of the expected molecular size (∼110 kDa) that was detected in the cRNA but not in the water-injected oocytes via anti-FLAG antibody (Fig. 3A). In addition, the plasma membrane expression of FLAG-jNCCβ protein was corroborated by its finding in the biotinylated membrane fraction.

Figure 3.

jNCCβ is a thiazide-resistant NaCl cotransporter. A: Western blotting analysis of total and membrane proteins (as stated) extracted from Xenopus oocytes injected with water or FLAG-jNCCβ cRNA. The membrane was blotted with anti-FLAG antibody or anti-actin antibody. B: ²²Na⁺ influx in oocytes injected with water or with jNCCβ cRNA in the presence of NaCl (white columns), the absence of extracellular Cl⁻ (gray columns) or the presence of 100 μm hydrochlorothiazide (black columns). C: ⁸⁶Rb⁺ influx in oocytes injected with water, NKCC2, or jNCCβ cRNA (as stated) in the presence (white columns) of extracellular chloride or 100 µm bumetanide (black columns). D: ³⁶Cl⁻ influx in oocytes injected with water or with jNCCβ cRNA (as stated) in the presence (white columns) or absence (gray columns) of extracellular sodium and the presence of 100 μm hydrochlorothiazide (black columns). *P < 0.0001 vs. the water-injected control or versus the absence of chloride or sodium of the same group. jNCCβ, Japanese eel NCCβ; MW, molecular weight; NCC, Na⁺-Cl⁻ cotransporter.

To study the functional properties of jNCCβ, ²²Na⁺ uptake was assessed in the presence or absence of extracellular Cl⁻ in Xenopus laevis oocytes that were microinjected with jNCCβ cRNA. As shown in Fig. 3B, oocytes injected with jNCCβ cRNA showed significantly increased ²²Na⁺ uptake (7.8 ± 1.4 nmol/h/oocyte), which decreased in the absence of extracellular Cl⁻ (1.1 ± 0.2 nmol/h/oocyte), and no sensitivity to hydrochlorothiazide was observed. Similarly, we observed that oocytes injected with jNCCβ cRNA showed significantly increased ³⁶Cl⁻ (15.9 ± 2.5 nmol/h/oocyte) uptake, which decreased in the absence of extracellular Na⁺ (2.6 ± 0.4 nmol/h/oocyte), and as expected, hydrochlorothiazide did not modify jNCCβ activity (Fig. 3C). These results demonstrated the interdependence of Na-Cl transport in jNCCβ, and because they were performed without K⁺ in the uptake medium, it is unlikely that jNCCβ also transported this ion. However, to completely discount this possibility, we determined ⁸⁶Rb⁺ uptake (as a K⁺ surrogate) in oocytes injected with jNCCβ cRNA. As shown in Fig. 3D, although injection of rat NKCC2 cRNA (rNKCC2) as a positive control significantly increased furosemide-sensitive ⁸⁶Rb⁺ uptake (3.9 ± 0.2 nmol/h/oocyte), no increase in ⁸⁶Rb⁺ uptake was observed in oocytes injected with jNCCβ (0.8 ± 0.1 nmol/h/oocyte) under the same conditions, confirming that jNCCβ is a K⁺-independent cotransporter. All these observations together revealed that jNCCβ and eNCCβ (12) exhibited similar behaviors.

Thiazide Inhibitory Kinetics and Profile in NCC and jNCCβ

The effects of different thiazides at a concentration of 100 μM were evaluated in oocytes injected with jNCCβ cRNA and flounder NCC cRNA (flNCC). As shown in Fig. 4A, we observed that all of the thiazides tested inhibited flNCC activity by 70%–90%, whereas none of the thiazides had any effect on oocytes injected with jNCCβ cRNA. This observation is similar to that reported for eNCCβ (12). In addition, the effect of increasing concentrations of metolazone on the functional expression of jNCCβ, eNCCβ, and rat NCC (rNCC) was analyzed. As shown in Fig. 4B, oocytes injected with rNCC cRNA were inhibited in a dose-dependent manner (7, 9), in contrast to oocytes injected with eNCCβ or jNCCβ cRNA, which were not inhibited by this thiazide.

Figure 4.

Thiazide inhibitory profiles in flNCC and jNCCβ. A: oocytes injected with flNCC or jNCCβ cRNA were exposed to tracer ²²Na⁺, and the influx was measured under control conditions (white columns), in the absence of extracellular Cl⁻ (gray columns), or in the presence of 100-μm metolazone (black columns), bendroflumethiazide (horizontally hatched columns), hydrochlorothiazide (squared columns), chlortalidone (left hatched columns), or trichlorothiazide (right hatched columns). The data observed for each control group were taken as 100%, and the thiazide groups were normalized accordingly. *P < 0.0001 vs. control conditions. B: dose-response curve for hydrochlorothiazide of rNCC, eNCCβ, and jNCCβ. Xenopus oocytes were injected with rNCC cRNA (circles), eNCCβ cRNA (boxes) or jNCCβ cRNA (triangle). Three days later, a dose-response curve was obtained by assessing the ²²Na⁺ influx in the absence of the diuretic or in its presence in concentrations ranging from 10⁻⁸ to 10⁻⁴ M. No effect was observed in any of the eNCCβ or jNCCβ cRNA groups, whereas an inhibitory effect of hydrochlorothiazide was observed for rat NCC, as described previously (7–12). *P < 0.05 vs. the control group. eNCCβ, European eel NCCβ; flNCC, flounder NCC; jNCCβ, Japanese eel NCCβ; NCC, Na⁺-Cl⁻ cotransporter; rNCC, rat NCC.

Ion Transport Kinetics

Previous works by our group showed that different NCC orthologs exhibit significant differences between their ion transport kinetics (7–9, 24). The affinity for both ions is higher in mammalian orthologs than in flNCC or eNCCβ. Moreover, Km values for Na⁺ and Cl⁻ are similar in mammalian NCCs, whereas in flNCC, Km values for Cl⁻ are lower than those observed for Na⁺ and similar to those reported for eNCCβ. Therefore, we evaluated the kinetics of ion transport in oocytes injected with jNCCβ cRNA.

As shown in Fig. 5A, uptake experiments were performed by fixing the Na⁺ concentration to 96 mM and changing the Cl⁻ concentrations from 0 to 100 mM. Uptake in water-injected oocytes (data not shown) was subtracted from each group of jNCCβ cRNA-injected oocytes. ²²Na⁺ uptake in jNCCβ showed saturation kinetics, with an estimated Km of 22 ± 9 mM and a V_max of 4.3 ± 0.5 nmol/oocyte/h. As shown in Fig. 5B, the Hill coefficient of the transport kinetics was 1.2, suggesting that at least one Cl⁻ ion was translocated.

Figure 5.

Chloride and sodium transport kinetic analysis in oocytes injected with jNCCβ cRNA. A: ²²Na⁺ influx in oocytes exposed to influx medium with increased concentrations of Cl⁻ from 0 to 96 mM. B: Hill plot analysis based on kinetic data obtained from A; the slope of the Hill plot is 1.1 ± 0.2. C: Na⁺ influx in oocytes exposed to influx medium with increased concentrations of Na⁺ from 2 to 80 mM in solutions with pH 7.4. A nonsaturation curve was observed. jNCCβ, Japanese eel NCCβ.

Surprisingly, by applying the Michaelis–Menten equation on eNCCβ to define the kinetics involved in Na⁺ transport, it was not possible to be determined because the movement of Na⁺ has a progressive increase in media containing 0–96 mM Na⁺ (Fig. 5C) without reaching saturation.

eNCCβ Is More Efficient in the Influx and Efflux of Ions than flNCC

Compared with its kidney orthologs, eNCCβ presents highly efficient transport activity. We evaluated the influx and efflux capacities of Na⁺ and Cl⁻ ions between flNCC and eNCCβ using oocytes injected with water, flNCC cRNA, and eNCCβ cRNA that were incubated in ND96 with 1 µCi/mL ²²Na⁺ or ³⁶Cl⁻ for 24 h to fill the cells with the respective radioactive under homeostatic conditions to assess the input of these isotopes and also to evaluate the tracer ion output in the presence or absence of extracellular Na⁺ and Cl⁻. The efflux of Na⁺ and Cl⁻ was assessed by incubating the oocytes in different solutions (ND96, without Cl⁻ or Na⁺) for 15 min; before this procedure, the oocytes were quickly washed three times with ND96 and one time with the solution in which they were incubated to avoid contamination from external radioactivity.

As expected from this family of secondary transporters, the movement of the ions follows their gradient, so under standard conditions, the entry of both Na⁺ and Cl⁻ into the cell is expected. However, it is interesting to observe (Fig. 6, A and C) that the entries of ²²Na⁺ and ³⁶Cl⁻ were significantly greater (P < 0.0001) in the group of oocytes injected with eNCCβ cRNA (²²Na⁺: 22.14 ± 1.28; ³⁶Cl⁻: 32.06 ± 1.57 nmol/h/oocyte) in contrast to the oocytes injected with flNCC cRNA (²²Na⁺: 11.10 ± 1.09; ³⁶Cl⁻: 14.12 ± 0.33 nmol/h/oocyte) or water (²²Na⁺: 4.47 ± 0.43; ³⁶Cl⁻: 5.79 ± 0.33 nmol/h/oocyte).

Figure 6.

The movement of Na⁺ and Cl⁻ is more efficient in eNCCβ than in flNCC. A: ²²Na⁺ influx in oocytes injected with water, with eNCCβ cRNA or flNCC cRNA for 24 h in ND96 medium (white columns). To assess the efflux of ²²Na⁺, these oocytes were incubated in ND96 or without Cl⁻ medium (gray bar) or without Na⁺ medium (black bar) for 15 min (B). C: ³⁶Cl⁻ influx in oocytes injected with water, with eNCCβ cRNA or flNCC for 24 h in ND96 medium (white columns). To assess the efflux of ³⁶Cl⁻, these oocytes were incubated in ND96 or without Cl⁻ medium (gray bar) or without Na⁺ medium (black bar) for 15 min (D). The ion influx and efflux were greater in oocytes injected with eNCCβ cRNA. *P < 0.01 vs. the water-injected control. eNCCβ, European eel NCCβ; flNCC, flounder NCC.

Similar to the influx, the ion efflux in oocytes injected with eel NCCβ cRNA when incubated in solution without Cl⁻ or Na⁺ was significantly higher than that in oocytes injected with flNCC cRNA or H₂O.

The efflux or output of the ions (Fig. 6, B and D) was calculated by subtracting the amount of ²²Na⁺ or ³⁶Cl⁻ that entered the oocytes in 24 h minus the concentration of ²²Na⁺ or ³⁶Cl⁻ that remained in the oocyte after being exposed for 15 min in ND96 or in solutions with or without Na⁺ or Cl ⁻ . The difference obtained is shown in Fig. 6, B and D. As seen in Fig. 6, B and D, the efflux levels of ions in oocytes injected with eel NCCβ cRNA when incubated in a solution without Cl⁻ (²²Na⁺: 4.5 ± 0.7; ³⁶Cl⁻: 7.5 ± 1.3 nmol/h/oocyte) or without Na⁺ (²²Na⁺: 4.3 ± 1.0; ³⁶Cl⁻: 8.2 ± 1.0 nmol/h/oocyte) were significantly higher than in flNCC cRNA-injected oocytes incubated without Cl⁻ (²²Na⁺: 1.8 ± 0.9; ³⁶Cl⁻: −1.6 ± 0.8 nmol/h/oocyte) or without Na⁺ (²²Na⁺: 1.1 ± 0.9; ³⁶Cl⁻: −0.9 ± 0.9 nmol/h/oocyte) or in oocytes injected with H₂O without Cl⁻ (²²Na⁺: 1.4 ± 0.3; ³⁶Cl⁻: 0 ± 0.3 nmol/h/oocyte) or without Na⁺ (²²Na⁺: 1.4 ± 0.3; ³⁶Cl⁻: 1.3 ± 0.3 nmol/h/oocyte).

It is important to observe in these figures that the greater efficiency of the efflux movement of ions in favor of their ion gradient is in oocytes injected with eel NCCβ cRNA compared with oocytes injected with flNCC cRNA. Oocytes injected with flNCC showed practically no ion efflux for 15 min.

Electrogenic Properties of the eNCCβ

The high efficiency of the transport of Na⁺ and Cl⁻ by eNCCβ, the significant ³⁶Cl⁻ uptake in the absence of Na⁺ (12), which we also observed in jNCCβ (Fig. 3C; P < 0.03; H₂O 1.33 ± 0.31 vs. eNCCβ 2.6 ± 0.42 nmol/h/oocyte), as well as the impossibility of defining Na⁺ transport kinetics with ²²Na⁺ similar to other NCCs led us to hypothesize that ell NCCβ proteins could exhibit electrogenic properties. We thus assessed the behavior of oocytes injected with water, rNCC, eNCCβ, or jNCCβ cRNA by two-voltage clamp analysis. Figure 7 shows the representative current sweeps in different voltages of oocytes injected with water and different NCC orthologs. As expected, the currents observed in the different voltages in the oocytes injected with rNCC (Fig. 7B) were similar to those in oocytes injected with water (Fig. 7A), demonstrating that rNCC is an electroneutral cotransporter producing no changes in the membrane potential. The observations were similar using 100 mM (Fig. 7, A and B) and 0 mM chloride (Fig. 7, E and F). In contrast, as depicted in Fig. 7, C and D, a clear and significant current was observed in oocytes injected with eNCCβ or jNCCβ cRNA. The current was chloride-dependent because it was considerably reduced in the same oocytes when switched to extracellular media without chloride (substituted with gluconate; Fig. 7, G and H). Figure 8 shows the corresponding I–V plots. Significant differences were observed between the current intensity in the presence 100 mM Cl⁻ versus the absence of this ion in the oocytes injected with the eNCCβ cRNA (Fig. 8C) or jNCCβ cRNA (Fig. 8D). No currents were observed in water- (Figs. 8A and 9A) or rNCC cRNA-injected oocytes (Figs. 8B and 9B) in ND96, with or without Cl⁻ or Na⁺. In contrast, as shown in Fig. 9C, the currents observed in eNCCβ cRNA-injected oocytes in the presence of ND96 with or without Na⁺ were identical. The observed current in eNCCβ cRNA-injected oocytes had a reversal potential of −20 mV. All these data suggested that the current observed in the oocytes injected with eNCCβ or jNCCβ cRNA is Cl⁻ and not Na⁺ dependent.

Figure 7.

Whole cell currents of two voltage clamp analyses of Xenopus oocytes injected with water, rNCC cRNA, or eNCCβ cRNA. Water-injected oocytes (A and E), rNCC cRNA-injected oocytes (B and F), and eNCCβ cRNA-injected oocytes (C and G). D and H: jNCCβ cRNA-injected oocytes were voltage clamped at −50 mV. The current was monitored and recorded with I–V protocols consisting of 100 ms at V_h followed by 20 mV steps of 500 ms from V_h to −180 mV and +80 mM ending with 100 ms at V_h. The oocytes were analyzed using 100 mM extracellular chloride. A–D: the same oocytes as above, but in the absence of extracellular chloride. E–H: no difference was observed in the currents of water- and rNCC cRNA-injected oocytes. In contrast, a higher current was present in both NCCβ cRNA-injected oocytes that was abolished in the absence of extracellular chloride. eNCCβ, European eel NCCβ; jNCCβ, Japanese eel NCCβ; NCC, Na⁺-Cl⁻ cotransporter; rNCC, rat NCC.

Figure 8.

NCCβ is dependent on extracellular chloride. Two voltage clamp analyses of Xenopus oocytes injected with water and different NCC cRNAs. Water-injected oocyte (A), rNCC cRNA-injected oocyte (B), eNCCβ cRNA-injected oocyte (C), and jNCCβ cRNA-injected oocyte (D) analysis using 100 mM of extracellular chloride or without extracellular chloride. No difference was observed in the currents of water- and rNCC cRNA-injected oocytes. A significant major current was present in both oocytes injected with eNCCβ cRNA that was abolished in the absence of extracellular chloride. eNCCβ, European eel NCCβ; jNCCβ, Japanese eel NCCβ; NCC, Na⁺-Cl⁻ cotransporter; rNCC, rat NCC.

Figure 9.

The current present in eNCCβ oocytes was not affected in the absence of extracellular sodium. I–V plots of corresponding Xenopus oocytes injected with water, rNCC cRNA, or eNCCβ cRNA. Water-injected oocytes (A), rNCC cRNA-injected oocytes (B), and eNCCβ cRNA-injected oocytes (C). The currents observed in water and rNCC-injected oocytes were similar. A major-intensity current was observed in eNCCβ oocytes in the presence or absence of extracellular sodium that reversed at −20 mV. eNCCβ, European eel NCCβ; NCC, Na⁺-Cl⁻ cotransporter; rNCC, rat NCC.

The Current in NCCβ Oocytes Was Reduced by Mutations in the Putative Chloride-Binding Site (I172)

An alignment of the sequences of different orthologs of NCC, NKCC, and KCC was performed at the sites previously reported as chlorine-binding sites (25). As shown in Fig. 10, the two sites reported for Cl⁻ binding in NKCC and KCC are present in the NCC. The first Cl⁻ site is highly conserved between NCC and NKCC (GIL, located in the TM6 region). According to what has been reported (25, 26), this site is coordinated with the binding and transport of sodium in NKCC, and the other Cl⁻ site is coordinated with the binding and movement of potassium in KCC and NKCC; certainly, the coordination between the Cl⁻ and Na⁺ ions by cryo-EM in NCC has not yet been defined. The amino acid sequence of the first Cl⁻ site for NKCC is different between KCC and NCC at amino acid 172, which changes a methionine in NKCC to an isoleucine in NCC and KCC, suggesting that the higher electronegativity of the R group (CH2-S-CH3) of methionine with respect to the aliphatic branch of isoleucine could have an important effect on eNCCβ function. To test this hypothesis, by site-directed mutagenesis, we changed the isoleucine to glycine (G), alanine (A), or methionine (M).

Figure 10.

Alignment analysis of chlorine-binding sites in NKCC1, KCC1, and NCC. Two Cl⁻-binding motifs have been described in NKCC1 and KCC1. We performed an alignment analysis of these sequences and added the sequences of eNCCβ, jNCCβ, flNCC, rNCCβ, and hNCC. Top: alignment analysis of the TM1 regions in different cotransporters where the Cl⁻₁-binding site (amino acid residues in GVI boxes) can be observed. This motif is highly conserved in NCC and KCC1, whereas in NKCC1, methionine is present instead of isoleucine. Bottom: alignment analysis of the TM6 and TM10 regions involved in forming the Cl⁻₂-binding site. At this site, the amino acid residues in the boxes are shown to be highly conserved in all the cotransporters analyzed. eNCCβ, European eel NCCβ; flNCC, flounder NCC; jNCCβ, Japanese eel NCCβ; NCC, Na⁺-Cl⁻ cotransporter; rNCC, rat NCC.

As shown in Fig. 11, compared with the eNCCβ-wild type, the activity of NCC among oocytes injected with eNCCβ-I172G (31 ± 2%, n = 40), eNCCβ-I172A (32 ± 2%, n = 40), or eNCCβ-I172M cRNAs (28 ± 3%, n = 50) with respect to oocytes injected with eNCCβ cRNA (100 ± 4% n = 50) was reduced but still Cl⁻ dependent and not sensitive to thiazides. These data demonstrated that the aliphatic branch of isoleucine is important for the function of the transporter.

Figure 11.

Substitution of the amino acid isoleucine of the GVI motif related to the Cl⁻₁-binding site reduces the activity of eNCCβ. Single-site directed mutagenesis was used to perform isoleucine 172 substitution with alanine, glycine, or methionine (I172A, I172G, or I172M) in the eNCCβ cDNA. Mutations were confirmed by DNA sequencing. Oocytes were injected with eNCCβ or mutant I172A, I172G, or I172M cRNA. Three days later, Na⁺ influx was assessed under control conditions (white columns) or in the absence of extracellular Cl⁻ (gray columns) or in the presence of 100 μm of hydrochlorothiazide (black columns). Thiazide-sensitive ²²Na⁺ uptake in eNCCβ was set to 100%, and values observed in mutant clones were normalized accordingly. At least four different experiments were performed with 10 oocytes. Normalization of the data was performed for each experiment before comparing the groups. The data shown represent the means ± SE of four different experiments. *P < 0.0001 vs. eNCCβ. eNCCβ, European eel NCCβ.

However, Fig. 12 shows that all mutations at I172 had a significant effect on the electrogenic property of eNCCβ. The I–V plot shows the current intensity averages of at least 6 oocytes (6–18 oocytes) per curve, with oocytes from four different frogs that were injected with water or different orthologs of wild-type or mutant NCC cRNAs (Fig. 12A). Oocytes injected with the cRNAs of electroneutral NCC orthologs (sNCC and flNCC) showed currents similar to that observed in oocytes injected with water (endogenous current). In contrast, oocytes injected with eNCCβ cRNA exhibited higher currents, which were significantly reduced in oocytes injected with eNCCβ cRNAs with mutations at the I172 site (eNCCβ-I172A, eNCCβ-I172G, and eNCCβ-I172M).

Figure 12.

The current in NCCβ oocytes was affected by mutations in the putative union chloride site (I172). A: I–V plots of corresponding Xenopus oocytes injected with water (open circles), rat NCC cRNA (black squares), squalus NCC cRNA (black triangles), eel NCCβ cRNA (black invert triangles), or mutant eel NCCβ cRNA (I172M gray diamonds, I172G gray hexagons, and I172A gray open circles). The oocytes were voltage clamped at −50 mV. The current was monitored and recorded with I–V protocols consisting of 100 ms at V_h followed by 20 mV steps of 500 ms from V_h to −150 mV and +70 mM ending with 100 ms at V_h. B: percentages of current observed at 150 mV in corresponding oocytes injected with water (black bar), rat NCC cRNA (gray bar), eel NCCβ cRNA (white bar) squalus NCC cRNA (clear gray bar), or mutant eel NCCβ cRNA (I172M slanted lines bar from left to right; I172G slanted lines bar from right to left; I172A horizontal line bar), considering that the function of eNCCβ is 100%. *P < 0.01 mutants vs. eNCCβ. eNCCβ, European eel NCCβ; NCC, Na⁺-Cl⁻ cotransporter.

Figure 12 shows the averages of the current intensities at 150 mV obtained from the different groups of oocytes injected with water, with rNCC, sNCC, eNCCβ, and their respective mutant cRNAs plotted as percentages. Interestingly, the eNCCβ-I172M (50 ± 4% n = 36, 18 oocytes) showed a significant and higher reduction than the other mutants (eNCCβ-I172A, 75 ± 8% n = 24, 12 oocytes, eNCCβ-I172G, 77 ± 7% n = 34, 17 oocytes) in the current observed in oocytes injected with the eNCCβ cRNA (100 ± 11% n = 30, 15 oocytes). The percentage of current from oocytes injected with eNCCβ-I172M was close to the percentages of current obtained from oocytes injected with sNCC (33 ± 5% n = 10, 6 oocytes), rNCC (42 ± 3% n = 14, 7 oocytes), or water (34 ± 3% n = 14, 7 oocytes). If the intensity current percentage of water oocytes was subtracted from that of each NCC oocyte group, the reduction in the intensity current of mutants over the eNCCβ intensity current was better evaluated (eNCCβ: 100 ± 17%; eNCCβ-172M: 22 ± 6%; eNCCβ-I172G: 64 ± 10%; eNCCβ-I172A: 62 ± 12%). These data suggest that the characteristic electrogenic property of eNCCβ is nullified by this amino acid change in the putative Cl⁻-binding site.

eNCCβ Is Not Regulated by the Kinases WNK1 and WNK3

An important characteristic of the members of the SLC12 family is that their activity is regulated by phosphorylation/dephosphorylation processes in serine/threonine residues located in the terminal amino and carboxy domains; SPAK/OSR1 kinases (oxidative stress response kinase 1/serine-proline-alanine-rich kinase STE-20) are involved in these signaling processes and are capable of binding to their target protein through the RFxV/I motif. The activity of SPAK and OSR1, in turn, is modulated by a family of lysine-free kinases (WNK) that is composed of four members, WNK1–WNK4 (23, 27). When cell volume decreases or when [Cl⁻]i levels are low, the SPAK/OSR1-WNK signaling cascade is responsible for activating NKCC and NCC and inhibiting KCC. In contrast, increased cell volume or elevated [Cl⁻]i stimulates KCC activity and inhibits Na-(K)-Cl cotransporters. We evaluated whether eNCCβ function is affected by the kinases WNK1 and WNK3, potent stimulators of NCC function. Oocytes were coinjected with eNCCβ cRNA and WNK1 or WNK3 cRNA, and oocytes were coinjected with rNCC and WNK1 or WNK3 cRNA; the latter were used as controls. Figure 13 shows that neither kinase increases the activity of eNCCβ, contrary to what is observed for the activity of rNCC, which increased the percentage of its function more than twofold in the presence of WNK1 and threefold in the presence of WNK3. What is striking is the great transport capacity observed in the eNCCβ since the percentage of its function increased sixfold with respect to that of the rNCC.

Figure 13.

WNK1 and WNK3 did not increase eNCCβ activity. The ²²Na⁺ uptake data were normalized to 100% of the rNCC activity with a previous subtraction of background (²²Na⁺ uptake of water oocytes group) and were considered the control group. The percentages of eNCCβ and rNCC activity in oocytes injected with eNCCβ or rNCC cRNA alone (stated with −) or coinjected WNK1 or WNK3 cRNA (stated with +) incubated in the absence Cl⁻ (black columns) or in the presence Cl⁻ (white columns) were assessed. *P < 0.0001 vs. the respective NCC group. eNCCβ, European eel NCCβ; rNCC, rat NCC.

DISCUSSION

We previously reported that the NCCβ isoform of the European eel (12) exhibits differences in functional properties with respect to renal NCCα orthologs: 1) NCCβ is thiazide-resistant, 2) is not activated by low-chloride hypotonic stress, and 3) is not regulated by the WNK-SPAK pathway. However, Watanabe et al. (19) cloned the cDNA of the intestine and rectum NCCβ from seawater-acclimated Japanese eel and showed that this isoform is located mainly in the rectum. Its expression is upregulated in eels adapted in fresh water, and in rectal sac preparations, Na⁺ and Cl⁻ transport that was sensitive to 1 mM hydrochlorothiazide (1 × 10⁻³ M) was observed, suggesting that the Japanese eel NCCβ is thiazide-sensitive. However, the functional properties of the Japanese eel NCCβ have not been explored. The structural differences between jNCCβ and eNCCβ lie in 13 amino acid residues, six of which are in the amino-terminal region, three in the TM regions, and four in the carboxy-terminal domain. We reported that substituting the three residues of eNCCβ for those present in jNCCβ (C379S-S194F-A482G) conferred ∼40% sensitivity to thiazides to eNCCβ. Thus, we explored whether changing all these residues conferred full sensitivity to thiazide in jNCCβ. However, the data show that jNCCβ exhibits the same behavior as eNCCβ; it is a Na⁺-Cl⁻ cotransporter that is nevertheless resistant to thiazides. Thus, although the substitution of three residues in the TM domains conferred certain sensitivity to thiazides in eNCCβ, the addition of the rest of the differences between both isoforms reversed that, and the NCCβ again became resistant to thiazides.

Since Watanabe et al. observed hydrochlorothiazide sensitivity using whole rectal sacs (19), we have no explanation for their findings. Clearing the functional expression system in Xenopus laevis oocytes using jNCCβ cRNA is more specific for defining the functional properties of the transporter. In addition, they used a high concentration of hydrochlorothiazide, which could potentially have other effects. A similar observation was made by Bazzini (28) on the effect of hydrochlorothiazide in rat small intestine and intestinal cells of the human colon (HT-29), in which they showed that 1 mM hydrochlorothiazide induced a significant depolarization in the intestinal epithelium, which was attributed to a possible functional interaction between the NCC and the epithelial Ca⁺ channels (ECaC1 and ECaC2), which increased their Ca⁺ absorption in the presence of hydrochlorothiazide. However, the assumption of NCC presence was based on RT-PCR and Western blot of the ileum and jejunum, not the colonic epithelium, and no clear functional evidence of an NCC sensitive to hydrochlorothiazide was observed in the ileum and jejunum. Until now, the NCCβ isoforms of the Japanese and European eel are the only NCCs from a tissue different from the kidney that have been characterized at the functional level, and neither of them were sensitive to thiazides. Therefore, the effect of hydrochlorothiazide on the movement of ions in the intestines is a fertile area of research for new protein targets and molecular interactions that explain its pharmacological effect.

The lack of sensitivity to thiazides is not the only characteristic that differentiates the NCCβ isoform from its renal orthologs. The ion transport capacity of NCCβ is clearly higher than that of flNCC under similar experimental conditions. This was observed for both influx and efflux. A possible explanation for this is that eNCCβ is not regulated by the WNK-SPAK pathway; therefore, the intracellular Cl⁻ concentration does not limit its function, as occurs in renal NCCs (29), allowing a full function of eNCCβ, regardless of the intracellular concentration of chloride. Supporting this, in our previous work, we observed that although rat NCC was activated by low chloride hypotonic stress, eNCCβ was not activated (12). Interestingly, eNCCβ shares the SPAK-binding site and the amino terminal phosphorylation sites in NCC required for regulation by WNKs-SPAK.

Another striking property of the NCCβ that we characterized in the present work is its electrogenicity. This is interesting since NCC is one of the electroneutral cation chloride cotransporters that transport cations and chloride in a balanced stoichiometry and thus produce no current. However, since we cloned the eNCCβ cDNA, we noticed that although most of the transport was Na⁺ and Cl⁻ interdependent with Hill coefficients of 1:1, the tracer ²²Na⁺ uptake induced by eNCCβ was completely chloride-dependent, whereas the tracer ³⁶Cl⁻ uptake was not completely Na⁺-dependent, suggesting a remaining Cl⁻ current. Supporting this, we observed that the entry of ³⁶Cl⁻ in eNCCβ cRNA-injected oocytes with respect to entry ²²Na⁺ was greater, whereas the influx of ³⁶Cl⁻ and ²²Na⁺ in flNCC cRNA oocytes was similar, suggesting a movement of extra Cl⁻ in NCCβ. A similar situation occurred when the ionic gradient was favored for the exit of the ions. We thus characterized the eNCCβ and jNCCβ currents using two voltage clamps.

We observed that the oocytes injected with the cRNA of NCCβ isoforms generated a current that was not present in control oocytes injected with water or rNCC. The current was reversed at −20 mV, which was not affected by the elimination of Na⁺ in the bath solution but disappeared completely in the absence of Cl⁻, strongly suggesting that it is a chloride-dependent current. Therefore, we propose that NCCβ presents an electrogenic property by an extra movement of Cl⁻ that is completely independent of Na⁺ and is the reason for functional differences with respect to renal NCC orthologs.

Within the families of carrier solutes (SLC), there are very homogeneous families, as is the case of the SLC12 family, which is made up of Cl⁻ cotransporters coupled to Na⁺ and/or K⁺ cations, with electroneutral characteristics and sensitivity to diuretics (30); however, the NCCβ isoform confers diversity to the family due to its lack of sensitivity to diuretics and its electrogenic property. This heterogeneity is not new within the SLC families, in which very heterogeneous families have been described, as is the case for the SLC5 family, which is made up of 12 members; 10 are Na⁺ cotransporters coupled to solutes such as glucose, myoinositol, or anions; 1 is a Na⁺:Cl⁻:choline cotransporter; and another is an ion channel activated by glucose (31), although most have electrogenic properties. Within its members, there are electroneutral transporters, as is the case of members 8 and 12, both of which are Na⁺:monocarboxylate transporters, but Slc5a8 has electrogenic properties (32), whereas Slc5a12 is electroneutral (21) due to the stoichiometry of its transported solutes; in addition, Slc5a8 has the property of independently moving a small proportion of Na⁺ to the monocarboxylates and is considered a Na⁺ leak (32).

We wanted to demonstrate that the current observed in oocytes injected with the RNA of the NCCβ isoforms was its own characteristic, and given the impossibility of having a specific inhibitor for the function of the NCCβ isoform and since it was known that the current is dependent on Cl⁻, it was decided to make point mutations in the possible sites of binding to Cl⁻ of the NCCβ, with the idea that if the binding of Cl⁻ was modified, the current intensity would also be modified. The identification of these sites was based on the high identity of amino acids that the members of the SLC12 family present in the transmembrane regions and the data reported by NKCC1 and KCC cryo-electron microscopy (cryo-EM) and molecular dynamics simulations, by which the tertiary and quaternary structures of NKCC1, KCC1, KCC2, and KCC4 were determined (13, 17, 33) as well as the protein zone and amino acids involved in the union of Na⁺, K⁺, Cl⁻ ions, which provided a new vision on the mechanisms of the ion movement and regulation of these transporters.

These works demonstrated a dimeric assembly of NKCC1 with a partially open inward conformation, and they observed a nonprotein zone of a partially spherical shape with a structure very similar to the typical substrate-binding pocket observed in APC transporters. At that site, they determined the positions of the Na⁺, K⁺, and 2Cl⁻ ions in the case of NKCC1, which were very similar to the positions of K⁺ and Cl⁻ for the KCCs; they also observed that parts of the carboxy-terminal domain interacted directly with the transmembrane domains, which would explain the importance of posttranslational changes in the carboxy-terminal zone on the regulation of ion movement by general conformational changes in the protein structure (17). It is important to clarify that cryo-electron microscopy (cryo-EM) studies and molecular dynamics simulations have not yet been reported for any NCC.

In Fig. 10, we present an alignment analysis of the TM1, TM6, and TM10 domains. Based on the reported GVI motifs in TM1, the GIL motif in TM6 and the tyrosine (Y) in the TM10 region have been proposed as two sites that bind and transport Cl⁻ in NKCC1, KCC1, and KCC2. For Cl⁻-binding site 1 (SCl1), it is proposed that Cl⁻ is coordinated with the backbone amide groups of Gly134, Val135, and Ile136, along with the hydroxyl group of Ser430, which is approximately at 4 Å from the Cl⁻ at the SCl1 site. In the case of Cl⁻-binding site 2 (SCl2), it is proposed that Cl⁻ interacts mainly with three backbone amide groups, Gly433, Ile434, and Met435, as well as the Tyr589 hydroxyl group of TM10 (Fig. 10). Although we know that NCC is an electroneutral Na-Cl cotransporter with a stoichiometry of 1:1, within its sequence, the two Cl⁻ sites reported for the other members of the SLC12 family are present, raising the question of whether they are functional in jNCCβ and eNCCβ.

In cryo-EM studies, SCl1 was proposed as the site that allows the transport of Cl⁻ for KCC1 and KCC3 because it is coordinated with K⁺, as this site sits deeply in the cytosolic cavity, whereas SCl2 is more exposed to the cytosolic solvent and its function would only be structurally stable. For NKCC1, molecular dynamics simulations revealed that Cl⁻ ions diffused first between TM1 and TM6 (SCl1), remaining transiently before moving deeper into the cavity toward SCl2. In some simulations, Cl⁻ diffused to a third site located between TM1 and TM6a; however, only two first sites were in the translocation pathway that resemble Cl⁻ coordination sites on other transporters.

Previously, our group reported that NKCC2 and NCC retain more than 85% identity in the TM 1, 2, 3, 6, 8, and 10 regions; the high percentage of identity between them led us to think that any difference in the amino acid sequence in these regions would reveal important functional data (34). In Fig. 10, we show the amino acids involved in the binding of Cl⁻ in different NCCs and in NKCC1, NKCC2, and KCC1. In the sequence, we observe that the methionine of the GVM motif of NKCC1 is replaced by an isoleucine in NCC and KCC1 (GVI). We thus evaluated the role that this isoleucine plays in the function of eNCCβ. By point mutagenesis, we substituted this methionine with isoleucine, alanine, or glycine to simulate a shortening of the isoleucine aliphatic chain and evaluated its effect on eNCCβ function. As shown in our results, the ²²Na⁺ uptake of the eNCCβ isoform was reduced by more than 60% (Fig. 11). Of note, the missense of isoleucine 154 for phenylalanine in humans (site homologous to the GVI motif under study) is associated with Gitelman Syndrome (30); therefore, the observed reduction in eNCCβ correlates with this Gitelman mutation in human NCC (hNCC) Notably, the reduction of the function of eNCCβ was not only observed when measuring ²²Na⁺ uptake but also in the Cl⁻ current intensity characteristic of this isoform, which was reduced in the three mutants of the isoleucine 172 site by M, A, or G. It is interesting to observe that the highest percentage of reduction was in the mutant I172M, very similar to that of background current observed in oocytes injected with water. These observations suggest that a Cl⁻-binding site is affected by the change in an amino acid with a negative polarity slightly greater than isoleucine, as is the case for methionine. The methionine R group includes a sulfur (S) that has two electrons in its last orbit, thus conferring an electronegative degree greater than that of isoleucine.

These results strongly suggest that the observed Cl⁻ current is indeed a functional property of the NCCβ isoform, supporting cryo-EM studies of NKCC1 and KCCs, in which it was proposed that in NKCC1, Na⁺ coordinates with Cl⁻, which binds to SCl1, and K⁺ coordinates with Cl⁻ in SCl2. In the case of KCCs (given that they do not transport Na⁺ and do not present the coordination site for this ion), it is proposed that the active site to move Cl⁻ is SCl1; however, the authors suggested that in SCl2, a third Cl⁻ site is necessary to stabilize the structure. Therefore, it is likely that in NCC, the Cl⁻-binding site in the SCl1 site is important for the structural stability of the protein.

The proposal that there are several Cl⁻-binding sites may explain why NCCβ moves an extra amount of Cl⁻, independent of Na⁺, possibly by moving a small proportion of the Cl⁻ from SCl1, given its electrogenic property, unlike the mammalian NCC, in which the SLC1 site stabilizes the structure.

In conclusion, our results show that the NCCβ isoforms of both European and Japanese eel are Na⁺:Cl⁻ cotransporters that, in contrast to the renal NCC ortholog, are not sensitive to thiazides, exhibit electrogenic properties, and are not regulated by intracellular chloride concentration and the WNK-SPAK pathway.

GRANTS

The work was supported by grants DK51496 from the National Institutes of Health (NIH) and A1-S-8290 from Conacyt Mexico (to G.G.).

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

G.G. conceived and designed research; E.M., C.P., N.V., D.M.O.-V., D.P.-A., L.R.-V., V.O.-S., and G.G. performed experiments; E.M., C.P., D.M.O.-V., D.P.-A., L.R.-V., and G.G. analyzed data; E.M., C.P., and G.G. interpreted results of experiments; E.M., C.P., D.M.O.-V., V.O.-S., and G.G. prepared figures; E.M., C.P., and G.G. drafted manuscript; E.M. and C.P. edited and revised manuscript; E.M. and C.P. approved final version of manuscript.