Ae4 (Slc4a9) is an electroneutral monovalent cation-dependent Cl⁻/HCO₃⁻ exchanger

Bicarbonate (HCO₃⁻) transporters play important roles in many physiological processes such as intracellular pH (pH_i) regulation (Russell and Boron, 1976), acid–base transport (Gawenis et al., 2007), and Cl⁻ (Leviel et al., 2010) and HCO₃⁻ secretion (Ishiguro et al., 1996), as well as epithelial fluid secretion (Walker et al., 2002). SLC4-like genes encode for a family of HCO₃⁻-transporting proteins in bacteria, plants, and animals (Parker and Boron, 2013). The 10 mammalian SLC4 members (SLC4A1–5 and SLC4A7–11) mediate diverse HCO₃⁻-transporting activities, including Cl⁻/HCO₃⁻ exchange, Na⁺-HCO₃⁻ cotransport, and Na⁺-driven Cl⁻/HCO₃⁻ exchange. It is well accepted that SLC4A1–3 (AE1–3, respectively) are Na⁺-independent Cl⁻/HCO₃⁻ exchangers, whereas SLC4A4, -5, and -7 (NBCe1, NBCe2, and NBCn1, respectively) are Na⁺-HCO₃⁻ cotransporters (Parker and Boron, 2013; Romero et al., 2013). Conversely, the functional properties of SLC4A8–11 have not been as well defined, but it appears that they may act as either Na⁺-independent (SLC4A9/AE4; Tsuganezawa et al., 2001; Ko et al., 2002) or Na⁺-driven Cl⁻/HCO₃⁻ exchangers (SLC4A8, -9, and -10; Wang et al., 2000; Leviel et al., 2010; Peña-Münzenmayer et al., 2015), as well as Na⁺-HCO₃⁻ cotransporters (SLC4A8, -9, and -10; Grichtchenko et al., 2001; Wang et al., 2001; Parker et al., 2008; Chambrey et al., 2013). In contrast, it has been suggested that SLC4A11 (BTR1) serves as a borate transporter (Park et al., 2004), Na⁺-dependent OH⁻ transporter (Ogando et al., 2013), NH₃/H⁺ cotransporter (Zhang et al., 2015), or DIDS-stimulated H⁺ (OH⁻) transporter (Kao et al., 2015).

The current evidence suggests that SLC4 family members play an important role in exocrine gland fluid secretion by promoting intracellular Cl⁻ accumulation in secretory acinar cells (Nguyen et al., 2004; Melvin et al., 2005). Indeed, we recently found that Ae2 (Slc4a2) and Ae4 (Slc4a9) are functionally expressed in salivary gland acinar cells, but only Ae4 appears to play an important role in fluid secretion (Peña-Münzenmayer et al., 2015). There is no agreement regarding the Ae4 ion–transporting mechanism, but it has been suggested that Ae4 may be a Na⁺-independent Cl⁻/HCO₃⁻ exchanger (Tsuganezawa et al., 2001; Ko et al., 2002) or a Na⁺-HCO₃⁻ cotransporter (Chambrey et al., 2013). In contrast, we found that Ae4 acts as a Na⁺-dependent Cl⁻/HCO₃⁻ exchanger (Peña-Münzenmayer et al., 2015) in salivary gland acinar cells, but the nature of the Na⁺ dependency is unknown.

The main goal of this study was to gain insight into the Na⁺ dependence of the ion-transporting mechanism used by Ae4, which may provide a better understanding of how Ae4 supports fluid secretion by the exocrine salivary glands. Mouse Ae4 (Slc4a9)-mediated transport was associated with HCO₃⁻, Cl⁻, Na⁺, and K⁺ fluxes, with no change in membrane potential, revealing that Ae4 is an electroneutral, nonselective cation-dependent Cl⁻/HCO₃⁻ exchanger. We propose that the lower intracellular [Cl⁻] observed in the salivary gland secretory cells of Ae4^−/− mice is the consequence of the loss of Ae4-mediated intracellular Cl⁻ accumulation in exchange for KHCO₃ efflux.

Mice were housed in micro-isolator cages with ad libitum access to laboratory chow and water with a 12-h light/dark cycle. Experiments were performed on 2–4-mo-old female and male animals. All animal procedures were approved by the National Institute of Dental and Craniofacial Research, National Institutes of Health, Animal Care and Use Committee (ASP 13-686). Acinar-specific Ae2^−/− mice were generated by crossing conditional knockout Ae2 mice (Ae2^flox/flox; Coury et al., 2013) with knock-in mice expressing Cre recombinase under the Aqp5 promoter (ACID mice; Flodby et al., 2010) as previously described (Peña-Münzenmayer et al., 2015). Mice lacking Ae2 and Ae4 in salivary gland acinar cells (Ae2^−/−-Ae4^−/− mice) were obtained by making the proper crosses between Ae2^flox/flox, ACID, and Ae4^−/− (Simpson et al., 2007) mice.

Plasmids encoding Mus musculus Ae4 (Slc4a9) transcript variant 3 (GenBank accession no. NM_172830) and Homo sapiens AE4 (SLC4A9) transcript variant 2 (GenBank accession no. NM_031467) were obtained from OriGene.

Mice were euthanized by CO₂ asphyxiation followed by cervical dislocation. SMGs from Ae2^−/− and Ae2^−/−-Ae4^−/− mice were surgically removed and finely minced with scissors and then digested by two consecutive incubation steps of 25 min each with 0.5 mg/ml collagenase 2 (type 2; Worthington Biochemical Corporation) in basal medium eagle (BME; Sigma-Aldrich). All solutions were gassed continuously with 95% O₂/5% CO₂ and maintained at 37°C.

CHO-K1 cells (Sigma-Aldrich) were grown in Petri dishes containing F12 medium (Invitrogen) supplemented with 10% (vol/vol) fetal bovine serum (Invitrogen) and penicillin-streptomycin-glutamine (100 U/ml, 100 µg/ml, 292 µg/ml) and maintained in a humidified 5% CO₂/95% air atmosphere. Cells were electroporated (Nucleofector II; Amaxa) with the plasmids mentioned above (4 µg DNA per reaction) using the Nucleofector kit V (Lonza) according to the manufacturer’s instructions and seeded onto 5-mm-diameter coverslips (Warner Instruments). It is important to note that at least 4 µg DNA was required per electroporation to measure a robust Ae4 activity in CHO-K1 cells 18–20 h after electroporation.

The fluorescent ion-sensitive dyes BCECF-AM (2′,7′-bis-(2-carboxyethyl)-5-(and-6)-carboxyfluorescein, acetoxymethyl ester; Rink et al., 1982), SPQ (6-methoxy-N-[3-sulfopropyl] quinolinium; Illsley and Verkman, 1987), SBFI-AM (1,3-benzenedicarboxylic acid, 4,4′-[1,4,10-trioxa-7,13-diazacyclopentadecane-7,13-diylbis(5-methoxy-6,12-benzofurandiyl)]bis-, tetrakis[(acetyloxy)methyl] ester), and PBFI-AM (1,3-benzenedicarboxylic acid, 4,4′-[1,4,10,13-tetraoxa-7,16-diazacyclooctadecane-7,16-diylbis(5-methoxy-6,2-benzofurandiyl)]bis-, tetrakis[(acetyloxy)methyl] ester; Minta and Tsien, 1989) were purchased from Invitrogen and used to measure pH_i, [Cl⁻], [Na⁺], and [K⁺], respectively. CHO-K1 cells were loaded with the different ion-sensitive dyes as follows: 2 µM BCECF-AM for 15 min at 37°C (95% O₂/5% CO₂), 10 µM SBFI-AM or 10 µM PBFI-AM for 1.5 h at room temperature (95% O₂/5% CO₂). For intracellular [Cl⁻] measurements, CHO-K1 cells were exposed to a hypotonic loading solution containing 10 mM SPQ as previously described (Chao et al., 1989). Dispersed acinar cells from SMGs were loaded with 2 µM BCECF-AM (15 min at 37°C and continuously gassed with 95% O₂/5% CO₂).

Imaging experiments were performed using an inverted microscope (IX71; Olympus) equipped with a Polychrome IV Imaging System coupled to a high-speed digital camera (Till Photonics). Images were acquired by alternate excitation at 490 and 440 nm (BCECF), 340 and 380 nm (SBFI and PBFI), or 340 nm (SPQ). Emissions were captured at 530 nm (BCECF) or 510 nm (SPQ, SBFI and PBFI) using Imaging WorkBench 6.0 software (INDEC BioSystems). The temperature of the solutions (listed in Table 1) was kept at 37°C during the experiments using a CL-100 bipolar temperature controller (Warner Instruments).

Simultaneous whole cell recordings (voltage and current clamp modes) and intracellular [Cl⁻] or pH_i measurements were performed at room temperature (22°C) on CHO-K1 cells using an Axopatch 200B amplifier (Molecular Devices) and a Polychrome IV monochromator (Till Photonics). The membrane voltage and the macroscopic currents were acquired to a computer hard drive using Clampex 10 software (Molecular Devices) via a Digidata 1440 Interface (Molecular Devices). BCECF- and SPQ-associated fluorescent images were acquired as described in the previous section using pipette (intracellular) solutions supplemented with 100 µM BCECF (free acid) or 1 mM SPQ, respectively. Membrane potential measurements were corrected by the changes in liquid junction potential (Barry, 1994).

Glass pipettes (Warner Instruments) were pulled using a P97 puller (Sutter Instrument) to give a resistance of 2–3 MΩ using the solutions listed in Table 2. The seal resistance was continuously calculated according to Ohm’s law from the difference in the membrane potential in response to −30-pA (current clamp mode), 60-mV (voltage clamp experiments at −100 mV), or −20-mV (voltage clamp experiments at 0 mV) square pulses of 250 ms every 5 s. Experiments in which the seal resistance was >500 MΩ throughout the experiment were considered for analysis.

The composition of external solutions used for single pH_i, Cl⁻, Na⁺, and K⁺ measurements are given in Table 1. The composition of pipette (internal) and external solutions used in simultaneous whole cell recordings and intracellular Cl⁻ or pH measurements are given in Table 2. The composition of the external solutions used to evaluate the selectivity of Ae4 for inorganic monovalent cations is shown in Table 3.

The osmolality of the external and pipette solutions were measured by a freezing-point depression osmometer (Advanced Instruments) and adjusted to 300 mOsm/kg with sucrose. All of the reagents used for the solutions were from Sigma-Aldrich unless otherwise indicated.

The changes in free Gibbs energy associated with the Ae4-mediated transport were calculated (at 37°C) using the following formula:

where R is the ideal gas constant 8.31*10⁻³ kJ*mol⁻¹*K⁻¹, T is the temperature in Kelvin (310.15 K = 37°C), and the other parameters are the intracellular and extracellular Cl⁻, HCO₃⁻, and Na⁺ concentrations.

The dependence of Ae4 for Na⁺ and K⁺ ions was obtained by fitting the Ae4 activities measured at different [Na⁺]_o and [K⁺]_o to the following Hill function using Origin 8.0 software (OriginLab):

where R_Cation is the alkalinization rate in response to a reduction in external [Cl⁻] at different [Cation] and R_Min and R_Max correspond to the minimal and maximal alkalinization rates. The EC₅₀ is the concentration required to achieve 50% of the maximal effect, and nH is the Hill number.

Results are presented as the mean ± SEM. Statistical significance was determined using the Student’s t test. Multiple-sample comparisons were performed using one-way ANOVA followed by Bonferroni’s post hoc test. A p-value <0.05 was considered statistically significant. Origin 8.0 software was used for statistical calculations. All experiments were performed for each condition using preparations from three or more separate mice or cell electroporations.

The Cl⁻/HCO₃⁻ exchanger activity in mouse salivary gland acinar cells is dependent on the expression of the Ae2 and Ae4 genes (Peña-Münzenmayer et al., 2015). Consequently, the acinar cells from Ae2^−/− mice serve as a physiological model to isolate native Ae4-mediated Cl⁻/HCO₃⁻ exchanger activity. As seen in Fig. 1 A (black circles), an extensive intracellular alkalinization consistent with Cl⁻/HCO₃⁻ exchange was observed when most of the extracellular Cl⁻ was isotonically replaced by gluconate (Cl⁻ concentration was reduced from 128.3 mM to 4 mM; see ion composition of solutions A and B in Table 1). The Cl⁻/HCO₃⁻ exchange activity present in Ae2^−/− acinar cells was dramatically reduced when external Na⁺ was isotonically replaced with the large organic cation NMDG (Fig. 1 A, green circles), demonstrating that Ae4 activity is Na⁺ dependent. Furthermore, there was nearly a total absence of an intracellular alkalinization in acinar cells from mice lacking Ae2 and Ae4 (Ae2^−/−-Ae4^−/−) in the presence of external Na⁺, in agreement with Ae4 being the major anion exchanger present in acinar cells from Ae2^−/− mice (Fig. 1 A, blue circles). A summary of the alkalinization rates calculated from experiments like those shown in Fig. 1 A (red lines) are presented in Fig. 1 B.

To further characterize the Na⁺ dependence of Ae4, we measured its Cl⁻/HCO₃⁻ exchanger activity by heterologous expression of mouse Ae4 cDNA in CHO-K1 cells. As seen in Fig. 2 A, an intracellular alkalinization comparable with that seen in native Ae2^−/− acinar cells (Fig. 1) was observed in Ae4-transfected CHO-K1 cells (black circles) when most of the extracellular Cl⁻ was replaced by gluconate. Like Ae2^−/− acinar cells, the Cl⁻/HCO₃⁻ exchanger activity in Ae4-transfected cells was Na⁺ dependent, as indicated by the lack of an intracellular alkalinization when external Na⁺ was replaced by NMDG (Fig. 2 A, blue circles). Furthermore, the Na⁺-dependent, Ae4-mediated alkalinization was absent under extracellular HCO₃⁻-free conditions (Fig. 2 A, green circles), consistent with Cl⁻/HCO₃⁻ exchange.

We also evaluated the Na⁺ dependence of Ae4-mediated Cl⁻ fluxes in CHO-K1 cells using the Cl⁻-sensitive dye SPQ. As shown in Fig. 2 B, a marked reduction in the intracellular Cl⁻ concentration was observed in Ae4-transfected cells when the external [Cl⁻] was reduced (black circles). In contrast, no reduction in [Cl⁻]_i was observed upon external Cl⁻ reduction when external Na⁺ was replaced by NMDG (Fig. 2 B, blue circles). Finally, as expected for a Cl⁻/HCO₃⁻ exchange process, there was also no change in the [Cl⁻]_i in Ae4-transfected cells under HCO₃⁻-free conditions (Fig. 2 B, green circles).

Together, Fig. 2 (A and B) shows that both HCO₃⁻ uptake (alkalinization) and Cl⁻ loss (reduction in [Cl⁻]_i) were prevented when external Na⁺ was replaced by NMDG or in the absence of extracellular HCO₃⁻, confirming that both HCO₃⁻ and Cl⁻ fluxes are tightly coupled to external Na⁺. Similarly, Fig. 2 C shows that CHO-K1 cells expressing the human AE4 protein displayed Cl⁻/HCO₃⁻ exchanger activity that was strongly dependent on extracellular Na⁺, demonstrating that the Na⁺ dependence displayed by mouse Ae4 is conserved in the human AE4 orthologue. A summary of the alkalinization rates calculated from experiments like those shown in Fig. 2 C (red lines) are presented in Fig. 2 D.

One possible explanation for the Na⁺ dependence of Ae4-mediated anion transport is an allosteric modulation of Cl⁻/HCO₃⁻ exchange by Na⁺. Alternatively, Na⁺ ions might be part of the Ae4-mediated transport cycle. To differentiate between these two possibilities, we measured changes in [Na⁺]_i using the Na⁺ indicator SBFI. As seen in Fig. 3, an increase in [Na⁺]_i was observed when Cl⁻/HCO₃⁻ exchange was activated by replacement of extracellular Cl⁻ with gluconate in Ae4-transfected CHO-K1 cells (black circles), but not in nontransfected cells (green circles), indicating that Ae4 mediates Na⁺ transport. The increase in the [Na⁺]_i induced by extracellular Cl⁻ removal recovered to the initial [Na⁺]_i when extracellular Cl⁻ was restored, consistent with Na⁺ moving in the opposite direction relative to Cl⁻ movement, although this could also be caused by Na⁺/K⁺-ATPase activity. Note that addition of the Na⁺-permeable ionophore gramicidin A caused a rapid increase in the [Na⁺]_i in both Ae4-transfected and nontransfected CHO-K1 cells.

The Na⁺ movement that accompanies Cl⁻/HCO₃⁻ exchange could be an electroneutral or electrogenic transport process, depending on the stoichiometry of ions transported by Ae4. Simultaneous whole cell recordings (current clamp mode) and intracellular [Cl⁻] measurements were performed to determine whether Ae4 activity is associated with changes in the membrane potential using pipette solution L (Table 2) supplemented with 1 mM of the Cl⁻ indicator SPQ. Under these experimental conditions, Ae4 expression in CHO-K1 cells caused a Na⁺-dependent decrease in [Cl⁻]_i upon [Cl⁻]_o reduction (like Fig. 2 B results) that was essentially absent in nontransfected cells (Fig. 4 A). A summary of the Cl⁻ efflux rates calculated from experiments like those shown in Fig. 4 A (red lines) are presented in Fig. 4 B. In contrast, the whole cell current clamp measurements from the same cells shown in Fig. 4 A demonstrated that the activity of Ae4-mediated Na⁺-dependent Cl⁻/HCO₃⁻ exchange is an electroneutral process as little, if any, change in the membrane potential was associated with ion fluxes (Fig. 4 C); i.e., the change in the membrane potential was <2 mV when a low Cl⁻, Na⁺-free external solution (solution M in Table 2) was switched to a low Cl⁻, Na⁺-containing solution (solution N in Table 2). A summary of the change in the membrane potential calculated from experiments like those shown in Fig. 4 C are presented in Fig. 4 D. Additionally, we also evaluated the voltage dependence of Ae4-mediated transport in CHO-K1 cells using the voltage clamp whole cell configuration. By supplementing the pipette solution L (Table 2) with 100 µM of the pH indicator BCECF (free acid), we compared the Na⁺-dependent alkalinization rates at −100 mV and 0 mV, respectively. As shown in Fig. 4 (E and F), an intracellular alkalinization was not observed in response to the addition of external Na⁺ in nontransfected cells at −100 mV (black line, n = 5) or 0 mV (green line, n = 5). In contrast, Ae4-transfected cells displayed robust alkalinization rates at −100 mV (Fig. 4 E, black circles) and 0 mV (Fig. 4 E, green circles) that were essentially identical, supporting our findings obtained by current clamp whole cell recordings that showed the ion-transporting mechanism of Ae4 is electroneutral rather than electrogenic. Additionally, no differences in the macroscopic currents (voltage clamp experiments) were found in Ae4-transfected versus nontransfected cells. The difference in current before and after Na⁺ addition to the external bath was −1.68 ± 0.44 pA/pF (n = 5) versus −1.27 ± 0.46 pA/pF (n = 5) at −100 mV (Ae4-transfected vs. nontransfected cells, P = 0.54, Student’s t test) and −0.28 ± 0.09 pA/pF (n = 5) versus −0.21 ± 0.14 pA/pF (n = 5) at 0 mV, respectively (Ae4-transfected vs. nontransfected cells, P = 0.69, Student’s t test).

The aforementioned experiments indicate that Ae4 acts as an electroneutral, Na⁺-dependent Cl⁻/HCO₃⁻ exchanger, although others have suggested that Ae4 may also support Na⁺-HCO₃⁻ cotransporter (NBC) activity (Chambrey et al., 2013). Consequently, we also tested whether Ae4 can mediate Na⁺-HCO₃⁻ cotransporter activity in the absence of extracellular Cl⁻ (solutions G, H, and I in Table 1, containing 10 µM EIPA to inhibit the activity of endogenous Na⁺/H⁺ exchangers) to isolate NBC-like activity (Burnham et al., 1997). Fig. 5 shows that a rapid intracellular acidification was induced by perfusion with a HCO₃⁻/CO₂-containing solution. In Ae4-expressing cells (Fig. 5, black circles), this acidification was dramatically blunted and then followed by a slower alkalinization that was absent in nontransfected cells (Fig. 5, green circles). Moreover, the Ae4-mediated intracellular alkalinization was tightly coupled to Na⁺ transport as removal of external Na⁺ prevented further alkalinization and caused an intracellular acidification.

The net Na⁺-HCO₃⁻ cotransporter–like activity observed in Ae4-transfected cells under Cl⁻-free conditions (Fig. 5) could be the result of bona fide Na⁺-HCO₃⁻ cotransport, Na⁺-driven Cl⁻/HCO₃⁻ exchange caused by incomplete intracellular Cl⁻ depletion or, alternatively, Na⁺-driven HCO₃⁻/HCO₃⁻ exchange where HCO₃⁻ substitutes for Cl⁻ in its absence. To differentiate between these possibilities, we used the whole cell (current clamp) configuration to control the intracellular ion composition while measuring the pH_i (low [HEPES], 1 mM, was included to permit the pH to change because of HCO₃⁻ influx). Ae4-mediated HCO₃⁻ influx was initiated by switching from a HCO₃⁻/CO₂-free to a HCO₃⁻/CO₂-containing external solution (conditions i and ii in Fig. 6 A, respectively) to induce an intracellular acidification, and then we compared the intracellular alkalinization rates after acidification. The external and pipette solutions were designed to impose a large inward Na⁺ gradient to promote HCO₃⁻ influx.

To differentiate between Na⁺-HCO₃⁻ cotransport and Na⁺-driven HCO₃⁻/HCO₃⁻ exchange using the protocol described in the previous paragraph, we then tested whether the Ae4-mediated alkalinization required intracellular HCO₃⁻ by comparing pipette solutions containing low Cl⁻ concentration with and without HCO₃⁻. Fig. 6 A depicts two possible scenarios: Ae4-promoted HCO₃⁻ influx by a mechanism independent (cotransport) of or dependent (exchange) on intracellular HCO₃⁻. Fig. 6 B shows that little if any intracellular alkalinization occurred in response to the HCO₃⁻/CO₂-induced acidification in Ae4-transfected cells dialyzed with a HCO₃⁻-free pipette solution (black circles; solution P in Table 2). Conversely, a distinct alkalinization was seen with a pipette solution containing 15 mM HCO₃⁻ (Fig. 6 B, green circles; solution Q in Table 2). Furthermore, there were no differences in the membrane potential changes between nontransfected and Ae4-transfected cells under these experimental conditions. Table 4 summarizes results obtained in experiments like those shown in Fig. 6 and also includes the membrane potential measurements recorded from nontransfected cells using the same pipette solutions. Additionally, we evaluated the voltage dependence of the intracellular alkalinization observed when a pipette solution containing 15 mM HCO₃⁻ supplemented with BCECF (free acid) was used. As shown in Fig. 6 C, voltage clamp whole cell recordings showed that the alkalinization rates were essentially identical in Ae4-expressing cells held at −100 mV and 0 mV, respectively. Additionally, the difference in the macroscopic currents (normalized by capacitance) before and after switching from HEPES-buffered to a CO₂-HCO₃⁻–buffered solution did not differ between Ae4-transfected versus nontransfected cells; the values were −2.60 ± 1.39 pA/pF (n = 5) versus −0.90 ± 0.64 pA/pF (n = 4) at −100 mV (Ae4-transfected vs. nontransfected, P = 0.34, Student’s t test) and 0.31 ± 0.17 pA/pF (n = 4) vs. 0.41 ± 0.07 pA/pF (n = 4) at 0 mV (Ae4-transfected vs. nontransfected, P = 0.58, Student’s t test). Collectively, these results indicate that the Ae4-mediated alkalinization can be driven by intracellular HCO₃⁻, which is not consistent with Na⁺-HCO₃⁻ cotransport activity, but with an electroneutral, Na⁺-driven Cl⁻/HCO₃⁻ exchange where HCO₃⁻ substitutes for Cl⁻ in its absence (HCO₃⁻/HCO₃⁻ exchange). In fact, given that Ae4 transport is electroneutral, the efficiency of intracellular HCO₃⁻ to promote Na⁺-2 HCO₃⁻ influx is similar to that displayed by intracellular Cl⁻; i.e., the alkalinization rates were not significantly different between pipette solutions containing either 15 mM HCO₃⁻ + 4 mM Cl⁻ (solution Q in Table 2) or a HCO₃⁻-free pipette solution containing 15 mM Cl⁻ (solution O in Table 2; results obtained under this condition are included in Table 4).

Salivary gland acinar cells that lack Ae4 (Ae4^−/−) display a lower resting state [Cl⁻]_i and secrete less agonist-induced saliva compared with wild-type control glands (Peña-Münzenmayer et al., 2015). These observations suggest that Ae4 acts as a Cl⁻ concentrative pathway that is important for stimulated Cl⁻-dependent fluid secretion and raise questions about the physiological transport mechanism by which Ae4 promotes Cl⁻ influx and NaHCO₃ efflux. Given that Ae4-mediated transport is an electroneutral process (see Fig. 4) and assuming the following fixed steady-state ion concentrations for SMG acinar cells (mM): [Cl⁻]_o = 124.6 (Mangos et al., 1973), [Cl⁻]_i = 50.1 (Peña-Münzenmayer et al., 2015), [HCO₃⁻]_o = 24.7 (serum; Mangos et al., 1973), [HCO₃⁻]_i = 19 (calculated; Patterson et al., 2012), [Na⁺]_o = 151.6 (Mangos et al., 1973), and [Na⁺]_i = 15.5 (calculated from rat sublingual acini; Zhang et al., 1993), we calculated the Gibbs free energy (ΔG) values that are associated with Ae4-mediated Cl⁻ influx by independently varying the intracellular concentrations of Cl⁻, Na⁺, and HCO₃⁻, respectively. ΔG values were obtained assuming the following potential stoichiometries that could produce electroneutral Ae4-mediated transport (Cl⁻:Na⁺:HCO₃⁻): 1:1:2, 1:2:3, and 2:1:3 (Eqs. 1, 2, and 3 in Materials and methods). Fig. 7 shows that intracellular Cl⁻ concentrations <7 mM at 1:1:2 (top left graph), 0.5 mM at 1:2:3 (middle left graph), and 26 mM at 2:1:3 (bottom left graph) stoichiometries, respectively, are required to reach the negative ΔG values that thermodynamically favor an Ae4-mediated Cl⁻ influx. The calculated intracellular Na⁺ concentrations (middle row) required to reach negative ΔG values and favor Ae4-mediated Cl⁻ influx are predicted to be >104 mM at 1:1:2 (top), 143 mM at 1:2:3 (middle), and 54 mM at 2:1:3 (bottom). The right panels show that [HCO₃⁻]_i >49 mM (1:1:2; top right), 84 mM (1:2:3, middle right), and 29 mM (2:1:3, bottom right) are required to reach negative ΔG values to promote Ae4-mediated Cl⁻ influx.

Because the intracellular Cl⁻, Na⁺, and HCO₃⁻ concentrations required to achieve negative ΔG values are quite far from the intracellular Cl⁻, Na⁺, and HCO₃⁻ concentrations reported for salivary gland acinar cells at rest and even during sustained stimulation, it seems unlikely that Na⁺ is the physiologically relevant cation. Thus, we hypothesized that K⁺, the main intracellular monovalent cation, supports Ae4-mediated Cl⁻/HCO₃⁻ exchange. Fig. 8 A shows that a substantial intracellular alkalinization was evoked in Ae4-transfected cells in response to a reduction in [Cl⁻]_o in both Na⁺ (black circles)- and K⁺ (green circles)-containing external solutions. In contrast, an intracellular alkalinization was not induced by low [Cl⁻]_o in Ae4-transfected cells when NMDG was the main external cation (Fig. 8 A, blue circles). Moreover, Fig. 8 B shows that the selectivity of the cation-dependent transport of Ae4 poorly distinguishes between the monovalent inorganic cations Na⁺, K⁺, Li⁺, Rb⁺, and Cs⁺; all tested cations evoked a robust Cl⁻/HCO₃⁻ exchange in Ae4-transfected cells.

The aforementioned results strongly suggest that like Na⁺, K⁺ and other monovalent cations may be transported by Ae4. To directly test whether K⁺ is transported by Ae4, the K⁺-selective probe PBFI was used to measure [K⁺]_i during stimulated Ae4 activity. Given that PBFI is only ∼1.5-times more selective for K⁺ than Na⁺ ions, we performed these experiments under Na⁺-free conditions to eliminate Na⁺ contamination of the K⁺ signal (Minta and Tsien, 1989). The experimental design is shown in Fig. 8 C. In brief, cells were loaded with PBFI and then perfused with a low [Cl⁻], Na⁺-free/high [K⁺] external solution to deplete the cells of Cl⁻ (Fig. 8, C and D, condition i). If Ae4 transports K⁺, then it would be expected that a combination of inwardly directed Cl⁻ and outwardly directed K⁺ gradients (by increasing [Cl⁻]_o and decreasing [K⁺]_o) would promote Cl⁻ influx in exchange for K⁺ (and HCO₃⁻) efflux, respectively (Fig. 8, C and D, condition ii). Fig. 8 D shows that a sustained reduction in the [K⁺]_i was observed when inward-directed Cl⁻ and outward-directed K⁺ gradients were imposed to Ae4-transfected cells, but little, if any, K⁺ efflux was observed in nontransfected cells, strongly suggesting that Ae4 indeed transports K⁺ ions. Dose–response experiments revealed that the affinities of Ae4 for Na⁺ and K⁺ ions were comparable (Fig. 9), where the EC₅₀ values for Na⁺ and K⁺ were 49 mM and 62 mM, respectively. The R_Min, R_Max, and nH obtained from Hill’s function fittings (see Eq. 4 in Materials and methods) to the data shown in Fig. 9 were 0.3, 1.5, and 2.0 for Na⁺ and 0.4, 1.6, and 1.8 for K⁺, respectively.

Assuming that physiological [Na⁺]_i + [K⁺]_i ≈ [Na⁺]_o + [K⁺]_o and that Ae4 nearly equally transports Na⁺ and K⁺ (Fig. 9), we calculated ΔG associated with a stoichiometry of 1:1:2 [Cl⁻:K⁺(Na⁺):HCO₃⁻]. Considering the aforementioned premises, it can be inferred from Eq. 1 that the Na⁺ and K⁺ gradients do not influence the direction of the Ae4-mediated transport. Therefore, the direction of the ion transport mediated by Ae4 will be governed by the Cl⁻ and HCO₃⁻ gradients. Using the same ion concentrations for the ΔG values shown in Fig. 7, the changes in ΔG value at different [Cl⁻]_i and [HCO₃⁻]_i were calculated. As shown in Fig. 10 A, a [Cl⁻]_i <73 mM (left graph) and a [HCO₃⁻]_i >16 mM (right graph) are sufficient to support Cl⁻ uptake by Ae4, ion concentrations that are in agreement with the physiological concentration range estimated for Cl⁻ and HCO₃⁻ in secretory acinar cells, 51.1 and 19 mM, respectively. Finally, a working model for Ae4 mediating Cl⁻ influx in exchange for K⁺(Na⁺)HCO₃ (efflux) in acinar secretory cells is shown in Fig. 10 B.

To better understand the physiological role of Ae4 in the kidney, exocrine glands and other tissues, it is essential to determine the ion transport mechanism underlying Ae4 function. The first cloning studies reported that rabbit (kidney β-intercalated cells; Tsuganezawa et al., 2001) and rat (kidney cDNA library; Ko et al., 2002) Ae4 orthologues are Na⁺-independent Cl⁻/HCO₃⁻ exchangers. In contrast, Chambrey et al. (2013) suggested that Ae4 is a Na⁺-HCO₃⁻ cotransporter in mouse kidney β-intercalated cells, whereas the Ae4 expressed in mouse salivary gland acinar cells appears to act as a Na⁺-dependent Cl⁻/HCO₃⁻ exchanger; however, detailed characterization of the Ae4-mediated ion transport mechanism was not performed (Peña-Münzenmayer et al., 2015). Thus, the aim of the present study was to gain further insight into the Ae4-mediated ion transport mechanism. We took advantage of Ae2^−/− mice to study the native Ae4 Cl⁻/HCO₃⁻ exchanger activity expressed in salivary gland acinar cells. A large Ae4-mediated intracellular alkalinization in response to a reduction in external [Cl⁻] that was dependent on external Na⁺ was observed in acinar cells. This response was not present in acinar cells from glands lacking both Ae2 and Ae4, confirming that the remaining activity in Ae2^−/− acinar cells corresponds to Ae4-mediated Na⁺-dependent Cl⁻/HCO₃⁻ exchange. Moreover, CHO-K1 cells expressing mouse Ae4 and human AE4 cDNAs mimicked the results obtained in native acinar cells. Collectively, our results demonstrate that both native and recombinant Ae4 act as Na⁺-dependent Cl⁻/HCO₃⁻ exchangers, although Na⁺-dependent Cl⁻/CO₃²⁻ exchange activity cannot be ruled out.

We next explored the nature of the Na⁺ dependence of Ae4-mediated Cl⁻/HCO₃⁻ exchange by addressing two basic questions: is the Na⁺ dependence caused by an allosteric modulation of Cl⁻/HCO₃⁻ exchange or is Ae4 able to transport Na⁺ ions? Consistent with the hypothesis that Ae4 activity is associated with Na⁺ transport, we found that Ae4-mediated HCO₃⁻ influx was accompanied by an increase in the intracellular [Na⁺]. Moreover, Ae4 activity did not generate a change in the membrane potential, demonstrating that Ae4 mediates an electroneutral, Na⁺ transport-dependent exchange of Cl⁻ and HCO₃⁻.

We also found that Ae4 can produce net Na⁺-HCO₃⁻ cotransport under nonphysiological Cl⁻-free conditions. Simultaneous whole cell recordings and pH_i measurements revealed that Na⁺-coupled HCO₃⁻ influx occurred with either Cl⁻ or HCO₃⁻ as the primary intracellular permeable anion, whereas little or no HCO₃⁻ influx was observed in a HCO₃⁻-free, low Cl⁻–containing pipette solution (Table 4). Moreover, there was no difference in the change of the membrane potential between Ae4-tranfected and nontransfected cells in response to HCO₃⁻/CO₂-induced acidification. Voltage clamp experiments comparing Ae4 activity at −100 and 0 mV showed no voltage dependence of Ae4-mediated ion transport, consistent with Ae4 mediating both electroneutral Cl⁻/NaHCO₃ or HCO₃⁻/NaHCO₃ exchange; however, we cannot rule out that OH⁻ ions can substitute for Cl⁻ ions and mediate OH⁻/NaHCO₃ exchange. Although our findings differ from those reported by Chambrey et al. (2013) regarding the activity encoded by Ae4 (i.e., Cl⁻/NaHCO₃ exchange vs. NaHCO₃ cotransport), both studies are in agreement with Na⁺ ions being transported by Ae4 in the same direction as HCO₃⁻ ions. A detailed characterization of the Cl⁻ dependence of Ae4-mediated activity in β-intercalated cells is required to evaluate whether the ion-transporting mechanism by Ae4 differs or is conserved in salivary glands and kidney, respectively.

Epithelial fluid secretion is a Cl⁻-dependent process in which basolateral Cl⁻ uptake and apical Cl⁻ efflux pathways are required to promote vectorial Cl⁻ secretion into the luminal space (Silva et al., 1977; Greger, 2000). The electroneutral basolateral Na⁺-K⁺-2Cl⁻ cotransporter (NKCC1) is the major Cl⁻ concentrative pathway in several tissues (Kunzelmann and Mall, 2002; Melvin et al., 2005), but a HCO₃⁻-dependent Cl⁻ concentrative mechanism has also proven to be important for secretion in many epithelia, including the duodenum and salivary glands (Walker et al., 2002; Peña-Münzenmayer et al., 2015). In fact, we found that salivary glands lacking the HCO₃⁻ transporter Ae4 (Ae4^−/−) display impaired fluid secretion as well as a lower [Cl⁻]_i in their acinar cells (Peña-Münzenmayer et al., 2015), suggesting that Ae4 is involved in a Cl⁻ concentrative pathway. If Ae4 indeed mediates Cl⁻ influx in acinar cells by promoting Cl⁻/NaHCO₃ exchange with a stoichiometry of 1:1:2 (Cl⁻, Na⁺, and HCO₃⁻), ΔG calculations predict that Ae4 mediates Cl⁻ efflux rather than Cl⁻ influx under physiological conditions. One possible explanation for the inconsistency between experimental data obtained from the Ae4^−/− mouse and the thermodynamic calculations could be off-target effects on the expression of another Cl⁻-uptake pathway. However, this seems unlikely because the Nkcc1 Na⁺-K⁺-2Cl⁻ cotransporter activity was not affected by Ae4 gene disruption (Peña-Münzenmayer et al., 2015). Alternatively, Ae4 could promote Cl⁻ influx and HCO₃⁻ efflux by another ion transport mechanism. To test this possibility, we considered whether the Na⁺ transport pathway of Ae4 might display cation promiscuity and, therefore, K⁺ might substitute for Na⁺ to promote Cl⁻/HCO₃⁻ exchange. We found that Ae4 expression promotes K⁺-dependent Cl⁻/HCO₃⁻ exchange and K⁺ efflux when an inward directed Cl⁻ gradient is imposed. Together, these results strongly suggest that Ae4 is able to transport both K⁺ and Na⁺ ions, and in fact, the cation transport pathway of Ae4 appears be nonselective as Li⁺, Rb⁺, and Cs⁺ ions also drive Ae4-mediated Cl⁻/HCO₃⁻ exchange.

Dose–response experiments for Na⁺ and K⁺ ions found that these two cations share almost identical EC₅₀, maximal alkalinization rates (R_Max), and Hill (nH) values obtained from Hill plots. The nH values obtained for Na⁺ and K⁺ (2.0 and 1.8, respectively) show that Na⁺ and K⁺ ions exert a cooperative effect on Ae4 activity. This finding suggests that (a) more than one Na⁺ or K⁺ ion is transported per cycle by Ae4 where the binding of the first cation facilitates the binding of another cation or (b) in addition to Ae4-mediated cation transport, Na⁺ or K⁺ ions may allosterically regulate Ae4 activity. More experiments in the future are required to address this open question.

Given that the sum of the external Na⁺ and K⁺ concentrations is nearly identical to the sum of intracellular Na⁺ and K⁺ concentrations, it can be inferred that the direction of the Cl⁻ and HCO₃⁻ transport by Ae4 is governed by the Cl⁻ and HCO₃⁻ chemical gradients; that is, the sum of the inward and outward cation driving forces cancel out. Further ΔG calculations assuming a stoichiometry of Cl⁻:K⁺(Na⁺):HCO₃⁻ of 1:1:2 (when [Cl⁻]_o = 124.6 mM, [Cl⁻]_i = 50.1 mM, [HCO₃⁻]_o = 24.7 mM, and [HCO₃⁻]_i = 19 mM, both [Na⁺]_o + [K⁺]_o and [Na⁺]_i + [K⁺]_i = 155 mM) confirm that the ΔG values are negative for Ae4-mediated Cl⁻ influx at the Cl⁻ and HCO₃⁻ concentrations reported for acinar secretory cells at rest, supporting a critical role of Ae4 in accumulating Cl⁻ above its equilibrium potential in exchange for HCO₃⁻ efflux in acinar cells. In addition to basolateral K⁺ channels, the K⁺ permeability displayed by Ae4 might also be important as a K⁺ recycling pathway to support Na⁺/K⁺ ATPase activity under some physiological conditions.

In summary, the present study provides evidence for the first time of a nonselective cationic transport pathway in a member of the Slc4a family of HCO₃⁻ transporters. Importantly, the cationic nonselectivity displayed by Ae4 might be shared with other Na⁺-dependent members of the Slc4a family, and thus, these transporters may play physiological roles different than those previously predicted. For instance, the electroneutral K⁺-HCO₃⁻ cotransporter activity described in the kidney (Leviel et al., 1992; Blanchard et al., 1996) as well as in other species (Zhao et al., 1995; Hogan et al., 1995a,b) might be encoded by a member of the Slc4a family.

We thank Drs. Z. Borok, E.D. Crandall, and U.P. Flodby for kindly providing the Aqp5-Cre (ACID) mice.

This study was supported by the Intramural Research Program of the National Institute of Dental and Craniofacial Research, National Institutes of Health (NIH; to J.E. Melvin) and NIH grant DK050594 (to G.E. Shull).

The authors declare no competing financial interests.

Merritt Maduke served as editor.