Renal intercalated cells are rather energized by a proton than a sodium pump

H⁺ V-ATPase Controls Steady-State Cell Volume and Resting Membrane Potential of Renal β-ICs.

The ICs of the kidney are located in a distal segment, the collecting duct, which also contains principal cells (13). ICs are specialized for H⁺ and HCO₃⁻ transport. They are enriched with mitochondria and have a high cytoplasmic content of carbonic anhydrase II (14). Two functionally distinct subtypes of ICs have been identified in the cortical collecting duct (CCD): the β-ICs secrete HCO₃⁻, whereas the α-ICs secrete H⁺. An apical Cl⁻/HCO₃⁻ exchanger and a basolateral H⁺ V-ATPase mediate secretion of base by the β-ICs, whereas α-ICs secrete acid by an apical H⁺ V-ATPase and a basolateral Cl⁻/HCO₃⁻ exchanger. In both cell types, it is the same H⁺ V-ATPase that is located either in the apical membrane of the α-ICs or in the basolateral membrane of the β-ICs (15). However, there is now general agreement that the apical Cl⁻/HCO₃⁻ exchanger of the β-IC is pendrin (16), whereas the basolateral exchanger of the α-ICs is a variant of the red cell anion exchanger 1 (Ae1) (17).

All living cells contain impermeable anionic colloids, which are mostly made up of proteins and organic phosphates. As a result of this, there is a high concentration of nondiffusible anions across the cell membrane, thus generating significant osmotic force between extracellular and intracellular compartments, also known as the Donnan effect. In typical animal cells, the Na⁺/K⁺ P-ATPase creates steep gradients for Na⁺ and K⁺ across the plasma membrane. Because the plasma membrane has a higher permeability for K⁺ over Na⁺ and anions, an outwardly directed K⁺ leak creates an inside-negative membrane potential. In cells that use a proton pump as bioenergizer, like plant and fungi cells, active expulsion of protons by the pump directly generate an inside-negative membrane potential. In both cases, the inside-negative membrane potential is critical to maintain steady-state cell volume because it drives a steady flow of inorganic anions, mostly Cl⁻, out of the cell, thereby counteracting Donnan’s effect. Thus, if the H⁺ V-ATPase instead of the Na⁺/K⁺ P-ATPase plays the role of plasma membrane bioenergizer in renal ICs, then dissipation of the resting membrane potential and cell swelling are expected to occur after inhibition of the H⁺ V-ATPase but not of the Na⁺/K⁺ P-ATPase. Therefore, we first tested the impact of blockade of either the H⁺ V-ATPase or the Na⁺/K⁺ P-ATPase on the volume of β-ICs and principal cells in CCDs isolated from mouse kidney and superfused in vitro (18). We subsequently applied either 40 nM bafilomycinA1, a highly specific blocker of the H⁺ V-ATPase (19), or 100 µM ouabain to block the Na⁺/K⁺ P-ATPase, and changes in cell volume were measured by monitoring the quenching of the fluorescent probe calcein using real-time two-photon imaging as described previously (20) (Movie S1). Application of bafilomycin A1 led to a significant increase in IC volume (Δ = +42 ± 4%, n = 12), as evidenced by the quenching of calcein fluorescence. In line with our hypothesis, principal cell volume measured in the same tubules was unaffected by bafilomycin A1. Conversely, ouabain induced significant cell swelling of principal cells (Δ = +38 ± 4%, n = 9), but not of ICs (Fig. 1 A–C).

We next assessed the impact of the H⁺ V-ATPase on the resting potential of both intercalated and principal cells. Changes in membrane resting potential were monitored by measuring the quenching of fluorescence of the voltage-sensitive dye ANNINE-6, as previously described (21). Application of bafilomycin A1, as shown in Fig. 1 D and E, led to a significant depolarization of ICs (Δ = +34 ± 2%, n = 8), indicating that the resting membrane potential in these cells critically depends on this pump. In contrast, bafilomycin A1 had no effect on the resting membrane potential of principal cells. Importantly, Muto et al. (22) have reported previously that blockade of the Na⁺/K⁺ P-ATPase by ouabain led to a marked depolarization of principal cells, but not of ICs. Taken together, these results indicate that the H⁺ V-ATPase acts as a bioenergizer of IC’s plasma membrane, whereas the Na⁺/K⁺ P-ATPase appears to be dispensable in this cell type.

NaCl Transepithelial Absorption by Renal ICs Is Energized by the H⁺ V-ATPase but Not the Na⁺/K⁺ P-ATPase.

One of the most prominent features of renal epithelial cells is their ability to mediate vectorial transepithelial NaCl transport. This process is dependent upon the activity of the Na⁺/K⁺ P-ATPase that converts the energy derived from metabolism into a steep inwardly directed sodium gradient. This sodium gradient energizes in turn numerous secondary or tertiary active transport systems. We recently examined transport properties of renal ICs on isolated renal tubules and identified an electroneutral thiazide-sensitive transport system in ICs (6). In these cells, NaCl absorption results from the functional coupling of the sodium-independent anion exchanger pendrin (Pds/Slc26a4) and of the sodium-dependent chloride/bicarbonate exchanger (Ndcbe) (Slc4a8).

The luminal bicarbonate concentration in nephron segments expressing pendrin is expected to be very low due to avid reabsorption of bicarbonate in the proximal tubule and the loop of Henle. Hence, we assume that the bicarbonate required for sustaining NaCl absorption via ICs comes from active bicarbonate secretion by pendrin. Moreover, pendrin accumulates of chloride into the cells, which is expected to favor sodium and bicarbonate uptake via Ndcbe. Pendrin has been shown to be energized by an outwardly directed bicarbonate gradient, which results from primary active proton extrusion by the H⁺ V-ATPase (23). Thus, we tested the dependence of transepithelial NaCl absorption on either the Na⁺/K⁺ P-ATPase or the H⁺ V-ATPase. As indicated above, two distinct transport pathways account for Na⁺ transepithelial absorption in the collecting duct: the first depends upon the epithelial sodium channel (ENaC), is electrogenic, amiloride-sensitive, and thiazide-resistant, and is located in the principal cells where it drives K⁺ secretion (24); the second depends upon the parallel action of pendrin and the Na⁺-driven Cl⁻/HCO₃⁻ exchanger Ndcbe, is electroneutral, thiazide-sensitive, and amiloride-resistant, and is restricted to ICs (6). Inhibition of the Na⁺/K⁺ P-ATPase by 10⁻⁴ M ouabain abolished transepithelial voltage (V_te) and K⁺ secretion in isolated microperfused mouse collecting ducts (Fig. 2 A and B). Similar effects were obtained by blocking the epithelial sodium channel ENaC of principal cells with 10⁻⁵ M amiloride. Application of both amiloride and ouabain (even at a concentration 10-fold higher) had no additional effects on V_te and K⁺ secretion. Because V_te and K⁺ secretion depend on the activity of the amiloride-sensitive Na⁺ channel ENaC in principal cells (Fig. S1) and because ouabain had the same effects than 10⁻⁵ M amiloride, these experiments showed that 10⁻⁴ M ouabain is able to block ENaC-dependent Na⁺ absorption completely. We next assessed the effects of the same inhibitors on Na⁺ and Cl⁻ absorption. Na⁺ flux was only partially inhibited by either 10⁻⁵ M amiloride or 10⁻⁴ M ouabain (Fig. 2C). Again, the simultaneous application of both blockers did not lead to significant additive effects, demonstrating that ouabain alone is sufficient to block the amiloride-sensitive Na⁺ absorption but does not affect amiloride-resistant Na⁺ transport. In contrast, Cl⁻ transport was neither affected by application of amiloride, ouabain, nor simultaneous application of both compounds (Fig. 2C). We next tested the effects of 5 × 10⁻⁸ M bafilomycin A1 (19). Basolateral application of bafilomycin A1 fully inhibited the amiloride-resistant component of Na⁺ and Cl⁻ absorption (Fig. 2D). These experiments demonstrate that sodium absorption by principal cells is primarily energized by the Na⁺/K⁺ P-ATPase, whereas sodium absorption by ICs might be exclusively energized by the H⁺ V-ATPase.

Basolateral Na⁺ Exit in β-ICs Occurs Through Ae4-Mediated Na⁺-HCO₃⁺ Cotransport.

In epithelial cells, the Na⁺/K⁺ P-ATPase also provides a basolateral exit pathway for sodium. In the absence of the Na⁺/K⁺ P-ATPase, the parallel action of pendrin and Ndcbe energized by the H⁺ V-ATPase is predicted to lead to net accumulation of Na⁺ and HCO₃⁻ into the cell. Thus, we hypothesized that, in the nominal absence of the Na⁺/K⁺ P-ATPase, Na⁺ transport across the basolateral membrane of ICs might be mediated by a bicarbonate-dependent sodium transporter energized by the H⁺ V-ATPase.

We have previously reported that Ae4/Slc4a9 is specifically expressed in β-ICs (25). We also detected Slc4a9 transcript by RT-PCR in cDNA of CCDs isolated from mouse kidney (Fig. S2). The localization and transport characteristics of Ae4/Slc4a9 are to some extent controversial. Concerning the different reported sites of Ae4 localization, previous studies lacked validation of the specificity of the Ae4 antibodies used on knockout tissue (26, 27). Even though Ae4 shares more similarities with Na⁺-HCO₃⁻ cotransporters than with Cl⁻/HCO₃⁻ exchangers of the SLC4 superfamily (28, 29), it has initially been cloned as a 4,4′-diisothiocyanatostilbene-2,2′-disulfonic acid (DIDS)-insensitive Na⁺-independent Cl⁻/HCO₃⁻ exchanger (27). Subsequently, Ae4 was reported to be rather DIDS sensitive (26). Finally, others reported that Ae4 might mediate Cl⁻-independent Na⁺-HCO₃⁻ cotransport rather than Cl⁻/HCO₃⁻ exchange (29, 30). To study the function of Ae4 function in vivo and to assess its potential role in Na⁺ extrusion across the basolateral membrane of ICs, we disrupted Slc4a9 in mice (Fig. S3). Slc4a9^−/− mice from heterozygous matings followed Mendelian ratios and had no obvious phenotypical abnormalities. Anti-mouse Slc4a9 antibodies detected the presence of Ae4 in cells of the collecting duct from Slc4a9^+/+ mice by immunohistochemistry (Fig. 3 A and B). The staining was completely abolished in kidney sections from Slc4a9^−/− mice demonstrating the specificity of the antisera generated (Fig. 3C). Ae4 labeling was exclusively detected at the basolateral membrane of renal epithelial cells that were identified as β-ICs because of apical expression of pendrin (Fig. 3 D and E). In contrast, α-ICs, which exhibit basolateral Ae1/Slc4a1 staining, were devoid of Ae4 labeling (Fig. 3F). The basolateral localization of Ae4 was further demonstrated by immunogold EM experiments (Fig. 3 G and H).

We also used a commercially available rabbit anti-human AE4 antibody (Alpha Diagnostics; catalog no. AE41-A) to examine Ae4 localization in CCDs isolated from rabbit kidney because apical localization of Ae4 has been previously reported in this species using an anti-rat Ae4 antibody (27). Fig. 4 I–K shows Ae4 labeling in a subpopulation of cells also stained with fluorescent peanut lectin, a marker of β-ICs in rabbit (31, 32). Moreover, colocalization studies of Ae4 with zona occludens protein 1 (ZO-1), a marker of the tight junction that delimits apical vs. lateral domains of epithelial cells, confirmed that, similar to mouse, Ae4 staining is restricted to the basolateral membrane of the cells (Fig. 3 L and M).

To study the role of Ae4 for NaCl reabsorption in its normal cellular context, CCDs dissected from the kidney of either Slc4a9^+/+ or Slc4a9^−/− were microperfused in vitro as described (33). Subsequently, the effects of peritubular Na⁺ removal or addition on IC’s intracellular pH were measured. In CCDs isolated from Slc4a9^+/+ mice, removal of Na⁺ from the peritubular solutions led to a marked intracellular acidification that was reversible upon restitution of Na⁺ to the bath (Fig. 4 A–C). These Na⁺-dependent changes of pH_i were not detected when tubules were perfused in CO₂/HCO₃⁻-free solutions, indicating that they reflect Na⁺-coupled HCO₃⁻ fluxes and not Na⁺/H⁺ exchange. In CCDs isolated from Slc4a9^−/− mice, Na⁺-dependent pH_i changes were dramatically decreased independently of the presence of or absence of CO₂/HCO₃⁻, which is in accordance with Na⁺-dependent HCO₃⁻ fluxes mediated by Ae4. However, Slc4a9 deletion did not fully abolish the Na⁺-dependent HCO₃⁻ fluxes, suggesting that another Na⁺-HCO₃⁻ cotransporter coexists in the basolateral membrane of β-ICs (Fig. 4 A–C). We next analyzed whether Ae4 disruption blocks NaCl absorption by these cells. We have shown previously that NaCl absorption by ICs or principal cells can easily be distinguished since the former is amiloride resistant, whereas the latter is amiloride sensitive (6). CCDs isolated from Slc4a9^+/+ mice exhibited significant amiloride-resistant Na⁺ and Cl⁻ absorption that was abolished in CCDs isolated from Slc4a9^−/− mice (Fig. 4D). The effect in Slc4a9^−/− mice was similar to that observed in Slc4a8^−/− (Ndcbe) mice. Slc4a8 has previously been demonstrated to mediate the apical entry pathway for Na⁺ in these cells (6) (Fig. 4E). Taken together, these experiments demonstrate that Ae4 mediates basolateral Na⁺-HCO₃⁻ cotransport when expressed in its normal environment and that it mediates sodium extrusion from renal β-ICs.

Ethic Statements.

All animal protocols conformed to the “Protocol of Animal Welfare” (Amsterdam Treaty; www.eurocbc.org/page673.html) and were approved by the Institutional Animal Care and Use review board of Université Pierre et Marie Curie or of the University of Jena.

Microperfusion of Isolated CCDs.

Experiments were performed as described previously in detail (6). Changes in pH_i were monitored using the pH-sensitive dye 2′,7′-bis-(2-carboxyethyl)-5-(and-6)-carboxyfluorescein (BCECF). ICs were distinguished from principal cells by their virtue of binding fluorescein-labeled peanut lectin [peanut agglutinin (PNA)] (Vector Labs) as described previously (6). For each tubule, three to four ICs were analyzed, and the mean gray level was measured with the Andor IQ software (Andor Technology). Ion fluxes on isolated CCDs were measured as previously described (6).

Generation of Slc4a9 Knockout Mice.

A fragment comprising exons 1–5 of the Slc4a9 gene was isolated from a 129/SvJ mouse genomic λ library (Stratagene) to generate the targeting vector for homologous recombination. Genotyping was performed by analyzing genomic DNA from tail biopsies. Mice were genotyped either by Southern blot or by PCR.

Northern Analysis.

Total RNA was isolated from various tissues of an adult C57BL/6 mouse using the RNeasy kit (Qiagen) according to the manufacturer’s instructions. Ten micrograms of total RNA were separated by electrophoresis and blotted following standard protocols.

Slc4a9/Ae4-Antibody Generation.

The Ae4 antisera were raised in rabbits against murine AE4 (GenBank accession no. NM_172830.2) with an N-terminal epitope KLPGQGDFESSDAHE(C) and a C-terminal epitope (C)PEEEETIPENRSEPE. The peptides were coupled via an N- or a C-terminal cysteine (in brackets) to KLH carrier. After immunopurification both antibodies gave consistent results in immunohistochemistry.

Immunostaining and EM of Kidney Sections or Isolated Tubules.

Cryosections of 4% (mass/vol) paraformaldehyde-fixed tissues were labeled with antibodies directed against Slc4a1/Ae1 (1:1,000 from guinea pig; a gift from Carsten Wagner, Zurich), Slc26a4/pendrin (1:2,000 from guinea pig; a gift from Carsten Wagner), and Slc4a9/Ae4 (1:500 for immunohistochemistry and 1:1,000 for EM; see above) using standard protocols.

CCDs were microdissected from normal rabbit kidney and fixed in 1:4 dilution of Prefer concentrate (glyoxal fixative) in Dulbecco’s PBS for 15 min (32). CCDs were then stained as previously described (32) with antibodies directed against Ae4 (rabbit anti-human AE4; Alpha Diagnostic International) or AE1:IV12 provided as a kind gift from M. L. Jennings (49) (University of Arkansas, Little Rock, AR). Colabeling was also performed with PNA-FITC (Vector Laboratories) or ZO-1, mouse monoclonal antibody-Alexa Fluor⁴⁸⁸ (catalog no. 339188; Invitrogen) was accomplished in a separate tertiary incubation. Three-dimensional reconstruction of individual ICs using the Fluoview software was performed to visualize the distribution of Ae4 in β-ICs and Ae1 in α-ICs.

Two-Photon Imaging and Semiquantitative Measurements of Changes in Cell Volume and Membrane Voltage in ICs.

CCDs were isolated and perfused as described before (50). Cell volume was monitored using the cell volume marker calcein (Invitrogen) and cell voltage using ANNINE-6, a newly synthesized voltage-sensitive dye designed for ultrafast (1-ms) neural signal detection as described before (21). Preparations were visualized using a two-photon excitation laser scanning confocal fluorescence microscope (TCS SP2 AOBS MP confocal microscope system; Leica Microsystems). Images were collected in time series at 1 Hz and analyzed with the Leica LCS imaging software (LCS 2.61.1537) Quantification Tools. In some experiments, CCDs were preincubated with bafilomycin (40 nM) or ouabain (100 µM; both from Sigma). Each perfused CCD was dissected from a different animal.