Human SLC4A11 Is a Novel NH₃/H⁺ Co-transporter

Background: Mutations in SLC4A11 result in corneal endothelial dystrophies; however, the transport characteristics of this membrane protein remain unclear.

Results: SLC4A11 showed NH₄Cl-dependent currents indicative of a NH₃/H⁺ electrogenic co-transport mode.

Conclusion: SLC4A11 is a novel NH₃/H⁺ co-transporter, uncharacteristic of the SLC4 bicarbonate transporter family.

Significance: SLC4A11 ammonia transport capacity should be considered along with other NH₃ transporters/channels when examining tissue nitrogen homeostasis.

Abstract

SLC4A11 has been proposed to be an electrogenic membrane transporter, permeable to Na⁺, H⁺ (OH⁻), bicarbonate, borate, and NH₄⁺. Recent studies indicate, however, that neither bicarbonate or borate is a substrate. Here, we examined potential NH₄⁺, Na⁺, and H⁺ contributions to electrogenic ion transport through SLC4A11 stably expressed in Na⁺/H⁺ exchanger-deficient PS120 fibroblasts. Inward currents observed during exposure to NH₄Cl were determined by the [NH₃]_o, not [NH₄⁺]_o, and current amplitudes varied with the [H⁺] gradient. These currents were relatively unaffected by removal of Na⁺, K⁺, or Cl⁻ from the bath but could be reduced by inclusion of NH₄Cl in the pipette solution. Bath pH changes alone did not generate significant currents through SLC4A11, except immediately following exposure to NH₄Cl. Reversal potential shifts in response to changing [NH₃]_o and pH_o suggested an NH₃/H⁺-coupled transport mode for SLC4A11. Proton flux through SLC4A11 in the absence of ammonia was relatively small, suggesting that ammonia transport is of more physiological relevance. Methylammonia produced currents similar to NH₃ but with reduced amplitude. Estimated stoichiometry of SLC4A11 transport was 1:2 (NH₃/H⁺). NH₃-dependent currents were insensitive to 10 μm ethyl-isopropyl amiloride or 100 μm 4,4′- diisothiocyanatostilbene-2,2′-disulfonic acid. We propose that SLC4A11 is an NH₃/2H⁺ co-transporter exhibiting unique characteristics.

Introduction

SLC4A11 is a recently identified transporter ubiquitously expressed in somatic cells (1, 2), with significant expression in the corneal endothelium, cochlea, and kidney (3). Several SLC4A11 mutations have been implicated in the pathogenesis of Fuchs endothelial corneal dystrophy, congenital hereditary endothelial dystrophy, and Harboyan syndrome (corneal dystrophy with perceptive deafness) (4, 5). Three strains of SLC4A11 knock-out mice exhibited variable phenotypes that included corneal endothelial dysfunction, perceptive deafness, and polyuria or low urine osmolality (6–8). Although SLC4A11 is classified as a member of the SLC4 (solute-linked co-transporter 4) family of bicarbonate transporters, it shares rather low (14–20%) homology with other family members. Phylogenetic analysis of SLC4A11 indicates a separate branch of early evolutionary divergence in the SLC4 family, suggesting a divergent function (9).

Despite high physiological significance of SLC4A11, its transport characteristics remain unclear. Initial studies suggested that SLC4A11 is an electrogenic 2Na⁺-OH⁻ (or borate) co-transporter (10). However, more recent studies provided evidence that human SLC4A11 transports neither borate nor bicarbonate (9, 11). These studies suggested ethyl-isopropyl-amiloride (EIPA)³-sensitive Na⁺ and H⁺ (or OH⁻) permeability through SLC4A11; however, a precise stoichiometry could not be determined. A more recent report has suggested H⁺ permeability through SLC4A11 but independent of Na⁺ (12).

Interestingly, ammonium permeability (9) and water permeability (13) have also been suggested for SLC4A11, and the across-species highly conserved asparagine-proline-X (NPX) water channel motif has been identified in a transmembrane domain of SLC4A11 (N⁶³⁹PS) (13). Because several water channels, such as AQP3, AQP7, AQP8, AQP9, and AQP10, are ammonia-permeable (14), and a single-point mutation in the aromatic/arginine region (ar/R) can induce ammonia permeability in the ammonia-impermeable water channel AQP1 (15), we hypothesized that SLC4A11 may also exhibit ammonia permeability similar to AQPs. To test this hypothesis, we generated two Na⁺/H⁺ exchanger (NHE)-deficient Chinese hamster PS120 fibroblast cell lines stably expressing high or low levels of SLC4A11. Employing whole-cell patch clamp electrophysiology, we found that NH₄Cl generated significant EIPA- and 4,4′-diisothiocyanatostilbene-2,2′-disulfonic acid (DIDS)-insensitive currents in SLC4A11-expressing PS120 cells. These currents were sensitive to [NH₃] and [H⁺] gradients but were relatively unaffected by Na⁺, K⁺, or Cl⁻ at physiological pH. The displayed current-voltage relations are consistent with an electrogenic NH₃- and H⁺-coupled co-transport model.

Experimental Procedures

Transfection and Cell Culture

Human SLC4A11-HA-tagged or empty vector (EV) were stably transfected into NHE-deficient PS120 Chinese hamster CCL39 fibroblast cell line as described previously (9). The SLC4A11 (NP_001167561.1) cDNA was subcloned into a plasmid containing a neo^r gene for Geneticin G418 selection and a C-terminal HA tag (13). PS120 cells were cultured in DMEM (Gibco®, catalog no. 21063, Life Technologies, Inc.) supplemented with 5% fetal bovine serum (Gibco®, catalog no. 10082, Life Technologies), 1% antibiotic/antimycotic (100 units/ml penicillin and 0.25 μ/ml fungizone), and 1 mg/ml Geneticin® G418 (catalog no. 10131, Life Technologies). The cells were cultured at 37 °C in a 5% CO₂ water-jacketed incubator. Culture medium was changed every other day.

Western Blot

SLC4A11- or EV- PS120 cells were grown on 6-well plates for 3 days before protein extraction. Cells were rinsed with ice-cold PBS, scrapped off from the wells, and sonicated in ice-cold radioimmune precipitation assay buffer (50 mm Tris base, 150 mm NaCl, 0.5% deoxycholic acid-sodium salt, 2% SDS, and 1% Nonidet P-40, pH 7.5) supplemented with complete protease inhibitor mixture (Roche Diagnostics, Mannheim, Germany). The extract was centrifuged for 20 min at 12,000 rpm at 4 °C, and the supernatant was collected. Protein concentration was determined using the 280-nm absorbance assay with a NanoDrop spectrophotometer (Thermo Scientific, Rockford, IL). Protein (15 μg/well) in Protein Loading Buffer Blue (2×) (National Diagnostics, catalog no. EC-886, Atlanta, GA) was resolved on 9% SDS-polyacrylamide gels and wet-transferred to PVDF membranes (Bio-Rad). Membranes were blocked with 5% nonfat milk in TBST (25 mm Tris base, 137 mm NaCl, 0.1% Tween 20) and probed with primary antibodies in the same buffer overnight at 4 °C. The following primary antibodies were used: 1) rabbit polyclonal anti-SLC4A11 (1:1000; 16B12, Covance); 2) mouse monoclonal anti-HA (1:2000; Santa Cruz Biotechnology); or 3) mouse monoclonal α-tubulin antibody (1:5000; DM1A, Novus Biologicals). Next, membranes were probed with secondary antibody (goat anti-mouse IgG peroxidase-conjugated antibody, 1:5000, catalog no. A8924 (Sigma) or goat anti-rabbit IgG peroxidase-conjugated antibody, 1:5000, catalogue no. A0545 (Sigma)) for 1 h at room temperature. Bound secondary antibodies were detected using an enhanced chemiluminescence assay (Supersignal West Pico, Pierce). Band densities and background were quantified using UN-SCAN-IT gel 6.1 (Silk Scientific).

Intracellular pH Measurements

Intracellular pH measurements were performed as described previously (9). Briefly, PS120 cells were cultured on poly-l-lysine-coated 25-mm diameter glass coverslips (Neuvitro Corp., El Monte, CA) for 2–3 days. Before each experiment, cells were incubated with 10 μm pH-sensitive fluorescent dye 2′7′-bis(carboxyethy)-5(6)-carboxyfluorescein (BCECF)-acetoxymethyl ester (Molecular Probes, Inc., Eugene, OR) in Ringer's solution for 30 min at room temperature and washed in dye-free Ringer's solution for another 30 min. The Ringer's solution contained 143.5 mm Na⁺, 4 mm K⁺, 0.6 mm Mg²⁺, 1.4 mm Ca²⁺, 118 mm Cl⁻, 1 mm HPO₄²⁻, 10 mm HEPES, 30 mm gluconate⁻, and 5 mm glucose, pH 7.5. Na⁺-free Ringer's solution contains 143.5 mm NMDG⁺ instead of Na⁺. All of the experimental solutions were equilibrated with air and adjusted to pH 7.5 with 1 n NaOH at 37 °C. Osmolarity of all solutions was adjusted to 295–300 mosm with sucrose. Coverslips with subconfluent cells were mounted into a perfusion chamber. The chamber was then placed on a stage warmer (37 °C) of an inverted microscope (Eclipse TE200, Nikon, Japan). Solutions were kept at 37 °C in a warming box, and the flow of the perfusate (∼0.5 ml/min) was achieved by gravity. Cells were imaged with an oil immersion objective (×40; Nikon). BCECF fluorescence was excited alternately at 495 ± 10 and 440 ± 10 nm, and the emitted light was collected through a bandpass filter (520–550 nm). Fluorescence ratios (495 nm/440 nm) were obtained at 1 Hz and converted into pH_i using the high K⁺/nigericin calibration approach (16).

Electrophysiological Recordings of PS120 Cells

SLC4A11- and EV-PS120 cells were cultured on 25-mm glass coverslips as described above to ∼10–20% confluence. High expression Col4 SlC4A11-PS120 cells were used in all electrophysiology experiments. Whole-cell currents were amplified using an EPC7 amplifier controlled by pCLAMP 10 software (Molecular Devices, Sunnyvale, CA) in the episodic mode. The holding potential was set to −60 mV unless otherwise indicated. The currents were filtered at 3 kHz and sampled at 1 kHz. Each episode lasted 2 s and included a 100-ms voltage ramp from −100 to +100 mV. Series resistance compensation was set at 50%. CLAMPFIT 10 software (Molecular Devices, Sunnyvale, CA) was used for data analysis. Only recordings obtained with a series resistance of less than 10 megaohms were analyzed. Background currents were subtracted when analyzing current amplitude and consequent current density data. No background current subtraction was done when the reversal potential values were determined. The liquid junction potential was subtracted before reporting reversal potential values. An agar salt bridge was used to connect reference electrode and the bath to minimize possible Ag/AgCl reference electrode potential changes during solution exchanges. Complete solution exchange required about 10–15 s in the recording chamber used for electrophysiological experiments (22–23 °C).

Solutions for Electrophysiology

The compositions of extracellular (E) and pipette (intracellular; I) solutions are listed in Tables 1 and 2, respectively. Extracellular solutions were adjusted to desired pH (pH 6.5, 7.5, and 8.5) with 1 n HCl, 1 n NaOH, or 1 m NMDG base solution (for Na⁺-free E5). Intracellular solution pH was adjusted with tris(hydroxymethyl)aminomethane base to pH 7.2. Osmolality of all solutions was adjusted to 300–310 mosm with mannitol. 10 mm NH₄Cl, methylammonium, tetraethylammonium, or 5 mm (NH₄)₂SO₄ salt was added fresh at the day of the experiment. The (NH₄)₂SO₄ salt was used instead of NH₄Cl for Cl⁻-free E6 solution. Equal 0.17 mm NH₃, pH 6.5, pH 7.5, and pH 8.5 E2 solutions were made by equal molar substitution of NMDG-Cl in base E2 solution by 98.5 mm, 10 mm, and 1.16 mm NH₄Cl, respectively. Solution E8 was used to make 1, 3, 10, 30, and 100 mm NH₄Cl extracellular solutions by equal molar substitution of NMDG-Cl with NH₄Cl.

TABLE 1.

Composition of extracellular solutions

	E1	E2	E4 (0 K+)	E5 (0 Na+, 0 K+)	E6 (0 Cl−, 0 K+)	E8
	m	m	m	m	m	m
NaCl	143.5	45	143.5			80
NMDG-Cl		98.5		143.5		100
K2HPO4	1
KCl	3
Calcium gluconate	1.4	1.4	1.4	1.4	1.4	1.4
MgCl2	1.22	1.22	1.22	1.22	1.22	1.22
HEPES	10	10	10	10	10	10
Glucose	5.1	5.1	5.1	5.1	5.1	5.1
NaMeSO3					143.5

TABLE 2.

Composition of intracellular solutions

	I1	I1-NH4	I2
	mm	mm	mm
CsMeSO3	125	125	145
CaCl2	5	5
CsCl			10
MgCl2	2	2	2.2
HEPES	10	30	10
EGTA	10	10	0.5
NH4Cl		10
pH	7.2	7.2	7.2

Statistical Analysis

Statistical analysis was performed using SPSS Statistics version 21 (IBM Corp., Armonk, NY). Stoichiometry data were fitted to linear regression in SigmaPlot version 12 (Systat Software Inc., San Jose, CA). Throughout, data are presented as mean ± S.E., in which significance level is indicated as follows: p < 0.05 (*), p < 0.01 (**), and p < 0.001 (***).

Results

Ammonium Permeability through SLC4A11 Is Expression-dependent

Previously, using BCECF pH_i measurements, we observed increased Na⁺ and NH₄⁺ permeability and higher capacity to recover from acid loads in SLC4A11-expressing HEK293 or PS120 cells (9). To further clarify the underlying mechanisms, we generated two stably transfected PS120 cell lines (Na⁺/H⁺ exchanger-deficient), high expression colony 4 (Col4) and low expression colony 12 (Col12). We determined, by Western blot, that Col4 expressed 1.34 times more SLC4A11 than Col12 (anti-HA; Fig. 1A), whereas no endogenous SLC4A11 can be detected in EV-PS120 cells (anti-SLC4A11; Fig. 1A). Employing these two colonies of SLC4A11-PS120 cells, we examined the pH_i response to extracellular 10 mm NH₄Cl, relative to EV. Fig. 1B shows that NH₄Cl application caused a rapid pH_i rise in both SLC4A11-PS120 and EV-PS120 (labeled a) due to NH₃ gas diffusion and intracellular production of NH₄⁺ and OH⁻ (17). In phase b, pH_i plateaued in EV-PS120 cells (−0.01 ± 0.02 ΔpH/Δmin), whereas a significant acidification occurred in SLC4A11-PS120 cells. The acidification rate (ΔpH/Δmin) was 1.57 times higher in Col4 (−0.11 ± 0.02, n = 3) than in Col12 cells (−0.07 ± 0.01, n = 3) (Fig. 1C). Removing NH₄Cl (phase c) caused a prominent acidification: ΔpH_i −0.68 ± 0.10 in Col4 (n = 3) and 1.55 times less ΔpH_i −0.44 ± 0.03 in Col12 (n = 3), whereas essentially no acidification occurred (−0.08 ± 0.04, n = 5) in EV-PS120 cells (Fig. 1D). Phase b acidification is generally attributed to a finite permeability to NH₄⁺, and consequent phase c acidification is a H⁺ loading effect from the finite NH₄⁺ permeability in phase b (17). The extents of acidification during phases b and c were both positively associated with the SLC4A11 expression level (Fig. 1, C and D). Thus, we reasoned that SLC4A11 expression increased the apparent NH₄⁺ permeability in PS120 cells. This permeability could be solo NH₄⁺ transport or co-transport of NH₃-H⁺ (14), but it cannot be differentiated by BCECF pH_i measurements.

FIGURE 1.

NH₄Cl superfusion induced differential intracellular acidification in SLC4A11-PS120 relative to EV-PS120 cells. A, Western blot verification of SLC4A11 expression in EV (empty vector), Col4, and Col12 using anti-SLC4A11 and anti-HA antibodies. A doublet shows glycosylated and non-glycosylated forms of SLC4A11 protein (40). The bar graph on the right shows the density quantification of HA-SLC4A11 normalized to α-tubulin. B, pH_i response to 10 mm NH₄Cl pulse in EV and SLC4A11 PS120 cells. Each trace is the average of n = 5 EV, n = 3 SLC4A11 Col4, and n = 3 SLC4A11 Col12, respectively. The traces show the rapid alkalinization phase (a), slow acidification phase (b), and rapid acid loading phase (c). C, phase b apparent ammonium influx (ΔpH_i/Δmin) in SLC4A11 Col4 (n = 3, p = 0.001) and Col12 (n = 3, p = 0.01) is higher than in the EV control (n = 5) (one-way ANOVA with Tukey honest significant difference post hoc test, p = 0.001). D, phase c ammonium-induced peak acidification (ΔpH_i) during wash is higher in SLC4A11 Col4 (n = 3, p < 0.001) and Col12 (n = 3, p = 0.004) than in the EV control (n = 5) (one-way ANOVA with Tukey honest significant difference post hoc test, p < 0.001). E, left, EV, simple NH₃ diffusion equilibrates and raises pH_i (phase a in Fig. 1B), but no secondary acidification occurs due to lack of NH₄⁺ permeability (phase b). Right, SLC4A11, NH₃ diffusion caused similar pH_i rise as in EV (phase a); additional NH₄⁺ (or NH₃-H⁺) permeability from SLC4A11 adds net H⁺ influx and causes gradual acidification (phase b). Error bars, S.E. **, p < 0.01; ***, p < 0.001.

[NH₃]_o Not [NH₄⁺]_o Drives Extracellular NH₄Cl (eNH₄Cl)-dependent Inward Currents through SLC4A11

A previous study provided evidence that SLC4A11 may function as an electrogenic transporter in SLC4A11-HEK293 cells (10). Therefore, we next used the whole-cell patch clamp approach to determine whether the apparent NH₄⁺ (or NH₃-H⁺) permeability creates charge movements by applying 10 mm eNH₄Cl to SLC4A11-PS120 Col4 cells. NH₄Cl solutions were adjusted to three different extracellular pH levels (7.5, 6.5, and 8.5) to vary the proportion of NH₃ to NH₄⁺. No currents were observed in EV-PS120 during any of these NH₄Cl exposures (Fig. 2A, top). In contrast, in SLC4A11-PS120 shown in Fig. 2A (bottom), pH 7.5 eNH₄Cl application caused inward currents of −6.18 ± 1.09 pA/pF, (n = 7), whereas pH 6.5 eNH₄Cl induced significantly smaller inward currents (−0.22 ± 0.19 pA/pF, n = 4), and pH 8.5 eNH₄Cl elicited markedly larger inward currents (−18.43 ± 3.43 pA/pF, n = 4. Fig. 2B shows a summary bar graph of current density for EV- and SLC4A11-PS120 cells. In Fig. 2C, we plotted the current density against [NH₃]_o (left), [H⁺]_o (middle), and [NH₄⁺]_o (right), respectively. The eNH₄Cl-dependent current density varied directly with [NH₃]_o but inversely with [H⁺]_o or [NH₄⁺]_o. From pH 6.5 to 8.5, the 84-fold increase in current density was concomitant with a 75-fold increase in [NH₃]_o; conversely, there was essentially no change (17% decrease) in [NH₄⁺]_o and a 100-fold decrease in [H⁺]_o.

FIGURE 2.

eNH₄Cl-dependent inward currents through SLC4A11. A, sample trace of whole-cell current recording (−60 mV holding potential) in response to 10 mm eNH₄Cl at pH_o 6.5, 7.5, and 8.5 in EV- (top trace) and SLC4A11-PS120 cells (bottom trace). Bath pH changes themselves did not elicit significant currents (open arrows). A pH 8.5 wash after pH 8.5 NH₄Cl elicited outward current (star), and the pH 8.5 to 7.5 transition after NH₄Cl pulse elicited inward currents (filled arrow). Inset, schematic diagram showing currents carried by NH₃ and an unknown charge carrier by SLC4A11. B, bar graph summary of NH₄Cl-dependent current density in EV (n = 3) and SLC4A11 cells (n = 4) (repeated measure ANOVA, cell type, p = 0.005; pH, p = 0.009; Tukey honest significant difference post hoc test comparing currents at pH 6.5 versus pH 7.5, pH 7.5 versus pH 8.5, and pH 6.5 versus pH 8.5, p = 0.085, 0.001, and 0.017, respectively). C, SLC4A11 inward current density versus extracellular [NH₃], [H⁺], or [NH₄⁺] in 10 mm NH₄Cl at pH 6.5, 7.5, and 8.5. NH₄⁺ pK_a 9.25 was used for concentration calculation. Error bars, S.E. *, p < 0.05; ***, p < 0.001.

These results suggest that NH₃ is the actual transported species and that the [NH₃] gradient contributes to the driving force in SLC4A11 transport. To further test this hypothesis, we adjusted the NH₄Cl concentration to achieve equal [NH₃]_o = 0.17 mm in the three different pH solutions. This required 98.5 mm NH₄Cl in the pH 6.5 E2 solution, 10 mm NH₄Cl in pH 7.5 E2, and 1.16 mm NH₄Cl in pH8.5 E2, calculated from NH₄⁺ pK_a = 9.25 (equal molar substitution with NMDG-Cl). Fig. 3A shows a sample trace of current recording, and Fig. 3B shows the summary bar graph of current density. In this series, with [NH₃] being equal (0.17 mm), we observed similar amplitude in current density under pH 6.5 and 8.5 compared with pH 7.5. This result strongly supports the notion that [NH₃]_o rather than [NH₄⁺]_o drives the inward currents through SLC4A11. Interestingly, further analysis of current density revealed small differences between each pH condition (n = 3; repeated ANOVA, p = 0.001). It appears that [H⁺] gradients or extracellular pH affects SLC4A11 activity, whereas the inward current density is primarily determined by [NH₃]_o.

FIGURE 3.

eNH₄Cl-dependent inward currents when holding [NH₃]_o constant in SLC4A11-PS120. A, sample current trace in response to 0.17 mm NH₃ at pH_o 6.5, 7.5, and 8.5 in SLC4A11-PS120 cells. Current amplitude is similar across three pH solutions, with slightly larger current in pH 6.5 and slightly smaller current at pH 8.5. B, bar graph summary of NH₃/H⁺ maximum current density at pH 6.5, 7.5, and 8.5 after a 40-s exposure to NH₄Cl. Current density shows small differences under the three pH conditions, largest in pH 6.5 NH₃ and smallest in pH 8.5 (n = 3, repeated measure ANOVA, p = 0.001; pairwise comparison pH 6.5 versus pH 7.5, p = 0.022; pH 7.5 versus pH 8.5, p = 0.051; pH 6.5 versus pH 8.5, p = 0.012). Error bars, S.E. *, p < 0.05.

SLC4A11 eNH4Cl-dependent Current Is Independent of K⁺, Cl⁻, and Na⁺

Given that NH₃ is an uncharged molecule, we next tested whether K⁺, Cl⁻, or Na⁺ is the charge carrier of the eNH₄Cl-dependent currents. Refer to Table 2 for solution composition. Removal of K⁺ (E1 K⁺ versus E4 K⁺-free) had no effect on eNH₄Cl-dependent currents at pH 6.5, 7.5, and 8.5 (Fig. 4A; factorial ANOVA, p = 0.716). Similarly, removal of extracellular Cl⁻ (5 mm (NH₄)₂SO₄ was added in both Cl⁻-free E6 and Cl⁻-rich E4 solutions for comparison) did not affect the eNH₄Cl-dependent current at pH 7.5 (independent t test, p = 0.65; Fig. 4B). Removal of Na⁺ (E4 Na⁺-rich versus E5 Na⁺-free) did not significantly affect eNH₄Cl-dependent currents at pH 6.5 or 7.5 (Fig. 4C; factorial ANOVA Na⁺, p = 0.108; Na⁺/pH interaction, p = 0.058), but interestingly, the absence of Na⁺ enhanced eNH₄Cl-dependent inward currents by 38% at pH 8.5 (Fig. 4, C and D). At pH 8.5, eNH₄Cl-dependent current density in 143.5 mm Na⁺ E4 was −18.43 ± 3.43 pA/pF (n = 4), whereas it was −29.71 ± 4.62 pA/pF (n = 6) in Na⁺-free E5 (independent t test, p = 0.115; Fig. 4C). In summary, it is unlikely that K⁺, Cl⁻, or Na⁺ carries the eNH₄Cl-dependent currents. However, there is some interaction with Na⁺ at high pH_o.

FIGURE 4.

K⁺, Cl⁻, and Na⁺ are not substrates for SLC4A11 under physiological pH. A, the eNH₄Cl-dependent current is not affected by removal of extracellular K⁺ at pH 6.5, 7.5, and 8.5. Sample size (n) under each condition is as follows: K⁺-free, pH 6.5, n = 4; pH 7.5, n = 7; pH 8.5, n = 4; 5 mm K⁺, pH 6.5, n = 3; pH 7.5, n = 6; pH 8.5, n = 4 (factorial ANOVA, 5 mm K⁺ versus K⁺-free, p = 0.716; K⁺/pH interaction, p = 0.763). B, the eNH₄Cl-dependent current is not affected by removal of extracellular Cl⁻ at pH 7.5. (Cl⁻-free n = 5 versus Cl⁻-rich n = 6; independent t test, p = 0.645). C, bar graph shows that eNH₄Cl-dependent current is not affected by removal of extracellular Na⁺ at pH 6.5 and 7.5 but is enhanced at pH 8.5. Sample size (n) under each condition is as follows: Na⁺-free, pH 6.5 n = 4, pH 7.5 n = 6, and pH 8.5 n = 6; 145 mm Na⁺, pH 6.5 n = 4, pH 7.5 n = 7, and pH 8.5 n = 4 (factorial ANOVA, Na⁺ versus Na⁺-free, p = 0.108; Na⁺/pH interaction, p = 0.058). D, sample trace of whole-cell recording to test Na⁺ sensitivity. Removal of Na⁺ does not significantly change eNH₄Cl-dependent current and slightly reduces baseline current (triangles). Bath pH transition after pH 8.5 NH₄Cl elicited inward current (open arrow). Error bars, S.E.

SLC4A11 Co-transports NH₃ and H⁺

In Fig. 2A (bottom trace, solid arrow) and Fig. 4D (open arrow), we observed inward current induced by a bath pH decrease (8.5 → 7.5) after a pH 8.5 NH₄Cl pulse, suggesting that SLC4A11 can provide H⁺ conductance. Fig. 3 shows that [H⁺]_o affects SLC4A11 inward current density and kinetics in the presence of NH₃. Fig. 5A (and Figs. 2A and 4D), however, shows that simple bath pH changes do not induce noticeable current. These data suggest that H⁺ current is only measurable when there is NH₃ present. In Fig. 2A, a pH 8.5 wash after pH 8.5 NH₄Cl pulse induced an outward current (labeled with a star). We reasoned that this outward current is the efflux of NH₃ coupled to H⁺ resulting from outward NH₃ and H⁺ gradient during wash. To validate, we applied a pH 8.5 wash following a pH 7.5 NH₄Cl pulse, creating a greater outward [H⁺] gradient than a pH 7.5 wash. Fig. 5B shows smaller but measureable outward current (relative to the initial baseline) during this pH 8.5 wash, supporting the notion of NH₃-coupled H⁺ efflux. In contrast, no outward current is observed during a pH 7.5 wash after pH 7.5 eNH₄Cl pulse, indicating that an outward H⁺ gradient is needed. Fig. 5C (right trace) shows that later in the same cell, a pH 8.5 eNH₄Cl pulse was followed by a pH 8.5 wash. The outward current (point c) during the wash is consistent with outward gradients of both [NH₃] and [H⁺]. Interestingly, a return to pH 7.5 bath produced a significant inward current (point e), consistent with an increased [H⁺] inward gradient. These data show that significant H⁺ current through SLC4A11 can occur after an NH₄Cl pulse.

FIGURE 5.

Current-voltage relations of eNH₄Cl-dependent SLC4A11 currents. A, sample trace showing that bath pH transitions do not generate significant currents in the absence of NH₄Cl. B, sample trace showing that transient outward current was induced after pH 7.5 eNH₄Cl by a pH 8.5 wash, and transient inward currents were elicited during pH 8.5 → 7.5 bath transition. Light gray line, baseline current. C, same trace as in Fig. 2A, highlighting points for E_rev comparisons. D, corresponding I-V plot at the time points a, b, c, d, and e. E, schematic diagram showing current in SLC4A11 cells is carried by NH₃ and an unknown charge ion.

To provide further insight into this potential NH₃-H⁺ co-transport model, we examined the reversal potentials at key points in Fig. 5C. The liquid junction potential was estimated to be 8.9 mV in this E4-I1 solution pairing and was subtracted from raw E_rev values before analysis. Fig. 5D (a) shows the I-V relation at point a in Fig. 5C. The E_rev was −4.16 ± 1.47 mV (n = 10), inconsistent with solo NH₄⁺ conductance, which will otherwise result in large positive E_rev (14). Furthermore, we analyzed E_rev at time points b, c, d, and e (during and after pH 8.5 NH₄Cl pulse). Fig. 5D (b), during pH 8.5 NH₄Cl, E_rev was −21.56 ± 1.98 mV (n = 4) at the inward current peak. Ten seconds into the pH 8.5 wash (point c), E_rev shifted to −66.24 ± 2.03 mV (n = 4). Assuming that [H⁺]_o, [H⁺]_i, and [NH₃]_i changed very little from b to c in this short time frame, the decreasing [NH₃]_o should result in a left shift of E_rev to a more negative value and result in outward current. Indeed, we observed a negative shift in E_rev and NH₃-H⁺ outward currents, representing efflux of NH₃-H⁺ following NH₃ and H⁺ outward gradient. Returning to pH 7.5 bath, E_rev shifted from −66.37 ± 2.23 mV (n = 4, point d) to −37.60 ± 5.30 mV (n = 4, point e). This E_rev rightward shift was induced by increased [H⁺]inward gradient and resulted in inward currents. In summary, these data suggest that [NH₃] and [H⁺] changes synergistically affect current direction and apparent E_rev in SLC4A11-PS120 cells, supporting the hypothesis that NH₃ and H⁺ are transported in a coupled manner.

SLC4A11 H⁺ Permeability in the Absence of Ammonia Is Very Small

Fig. 1B suggests that the resting pH_i in SLC4A11-transfected cells is lower than EV. Analysis of data from more cells (Fig. 6A) shows that the resting pH_i was significantly lower in SLC4A11-PS120 Col4 (pH_i 6.98 ± 0.08, n = 20) relative to EV-PS120 (pH_i 7.32 ± 0.08, n = 23), suggesting that SLC4A11 provides some H⁺ permeability in the absence of NH₃, where a negative membrane potential could drive H⁺ influx, as proposed in a recent report (12). Figs. 2A and 5A, however, do not reveal any measureable H⁺ currents in the absence of NH₃. Nevertheless, increased H⁺ permeability is evident in SLC4A11-transfected cells during the acid recovery following NH₄Cl pulses in previous reports (9, 12). Fig. 6B shows that acid recovery rates after NH₄Cl pulse were significantly greater in Col4 and Col12 cells relative to EV. Next, to avoid cell exposure to NH₃, we used sodium acetate to acidify these cells and monitored acid recovery. Fig. 6, C and D, show a similarly enhanced acid recovery in SLC4A11 transfected cells during acidification with sodium acetate, indicating that this apparent H⁺ permeability is independent of NH₃. The H⁺ efflux during sodium acetate experiments was 4.7 μm/s (ΔpH/Δs × 14 mm/pH buffering capacity), similar to that reported recently for SLC4A11 H⁺ permeability (12). However, this efflux rate is quite small, only about 2.4% of Na⁺/H⁺ exchanger 1 (NHE1)-facilitated H⁺ flux (0.2 mm/s) (19), and ∼7.3 times smaller than H⁺ flux in NH₄Cl (see Fig. 1C). Moreover, the effects of this added H⁺ permeability on resting pH_i are not evident in bicarbonate-rich conditions (i.e. high buffering capacity) (9, 12) or NHE-expressing cells (9).

FIGURE 6.

H⁺ permeability through SLC4A11 in the absence of NH₃ is small. A, baseline pH_i in SLC4A11-PS120 cells (6.98 ± 0.37, n = 20) is significantly lower than in EV-PS120 cells (7.32 ± 0.41, n = 23) (independent t test, p = 0.008). B, acid recovery rates following NH₄Cl pulses (see Fig. 1) in Col4 (n = 3) and Col12 (n = 3) SLC4A11-expressing cells relative to EV (n = 5). (one-way ANOVA, p = 0.001; post hoc Bonferroni EV versus Col4, p = 0.001; EV versus Col12, p = 0.012). C, pH_i response to sodium acetate in EV and SLC4A11 cells. D, bar graph summary of acid load recovery rate in SLC4A11- (n = 4) or EV-PS120 (n = 4) cells during 40 mm sodium acetate pulse (independent t test, p = 0.011). Error bars, S.E. *, p < 0.05; **, p < 0.01; ***, p < 0.001.

Intracellular NH₃ Decreases SLC4A11 NH₃/H⁺ Inward Currents

Fig. 5C (d) demonstrates that an outward [NH₃] gradient can produce an outward current in SLC4A11-PS120 cells. Therefore, we next tested whether reversing or dissipating the transmembrane [NH₃] gradient by increasing intracellular [NH₃] would reverse any SLC4A11 electrogenic activity. Previously, we used a standard pipette solution containing 0 mm NH₄Cl (I1), whereas here we used 10 mm NH₄Cl pipette solution (I1-NH₄Cl). Liquid junction potential was estimated to be 8.5 mV in this E4-I1-NH₄Cl solution pairing and was subtracted from raw E_rev values before analysis.

We dialyzed I1-NH₄Cl into SLC4A11-PS120 cells reversing the transmembrane [NH₃] gradient, hoping this would be sufficient to stimulate outward current. However, when the cell was held at −60 mV, we observed no significant current at pH 7.5 or 8.5 (E4 bath solution) (Fig. 7A). During NH₄Cl application, inward current (Fig. 7A) was smaller than that with standard 0 mm [NH₃] intracellular solution. Fig. 7B shows current densities amounting to −0.63 ± 0.16 pA/pF (n = 3) at pH 7.5 NH₄Cl and −9.26 ± 1.73 pA/pF at pH 8.5 NH₄Cl (n = 3); both were significantly smaller than that in I1 dialysis (p = 0.03), consistent with a reduced driving force for NH₃-H⁺ inward currents because of reduced [NH₃]_o/[NH₃]_i gradient. Interestingly, the presence of intracellular NH₃ and/or the reduced NH₃ gradient also slows current development. Fig. 7C shows a significant difference in the time constants between I1-NH₄Cl, τ = 58.44 ± 17.58 s (n = 3), and I1, τ = 9.00 ± 0.88 s (n = 4). In addition, the outward current at the pH 8.5 wash was prolonged and more prominent with I1-NH₄Cl dialysis than I1 (Fig. 7C, ▵), indicating that increased [NH₃]_i can produce larger outward current.

FIGURE 7.

eNH₄Cl-dependent currents through SLC4A11 are sensitive to transmembrane [NH₃] gradient. A, sample trace of a whole-cell current recording in SLC4A11 cells with intracellular 10 mm NH₄Cl and 10 mm eNH₄Cl superfusion at −60 mV holding potential. B, bar graph summary of responses to eNH₄Cl in I1-NH₄Cl- (n = 3) versus I1-dialyzed cells (n = 4) (repeated measure ANOVA; intracellular NH₄Cl, p = 0.03; in NH₄Cl/pH interaction, p = 0.610). C, overlay of NH₃/H⁺ current development in I1-NH₄Cl and I1 pipette solution. Data were fitted to a standard exponential equation, τ = 58.44 ± 17.58 s for I1-NH₄Cl (n = 3) versus τ = 9.00 ± 0.88 s for I1 (n = 4) (independent t test, p = 0.02, equal variance not assumed). D, representative trace of whole-cell current recording in SLC4A11-PS120 cells with intracellular 10 mm NH₄Cl or N(CH₂CH₃)₄Cl at break-in (arrows) and during 10 mm eNH₄Cl superfusion at −10 mV holding potential. E, bar graph summary of break-in outward current density in cells with I1-NH₄Cl (n = 7), or I1-N(C₂H₅)₄Cl (n = 4) (independent t test, p = 0.009). Error bars, S.E. *, p < 0.05; ***, p < 0.001.

Whereas we could not see outward currents at baseline in I1-NH₄Cl dialysis with −60 mV holding potential, Fig. 7, D and E, shows that at −10 mV holding potential, outward currents can be seen on break-in (arrow). As a control, dialysis of I1-N(CH₂CH₃)₄Cl (tetraethylammonium chloride; Fig. 7D, bottom trace) did not activate any currents on break-in (arrow). (Extracellular N(CH₂CH₃)₄Cl superfusion does not generate currents in SLC4A11-PS120 cells (Fig. 8D).) Moreover, Fig. 7D shows that inward currents at pH 7.5 NH₄Cl were similarly reduced in an I1-NH₄Cl-dialyzed cell compared with an I1-N(CH₂CH₃)₄Cl-dialyzed cell. Also, Fig. 7D (top trace) shows outward currents at the beginning of the pH 7.5 wash in I1-NH₄Cl-dialyzed cells. This outward current, although small, is consistent with an increased [NH₃]_i.

FIGURE 8.

SLC4A11 can also transport methylammonia. A, sample trace of whole-cell current recording in response to extracellular 10 mm MA and change of extracellular pH in SLC4A11-PS120 cells. B, bar graph showing a comparison of MA- and NH₃-induced current density in SLC4A11 cells (factorial ANOVA, NH₃ versus MA, p = 0.048; pH, p < 0.001; substrate/pH interaction, p = 0.498; independent t test comparing NH₃ (n) and MA (n′) current density at pH 6.5 (n = 6 versus n′ = 3), pH 7.5 (n = 21 versus n′ = 6), and pH 8.5 (n = 10 versus n′ = 5), p = 0.293, <0.001, and 0.221, respectively). C, representative reversal potential shift after MA perfusion. D, sample trace of whole-cell recording showing that tetramethylammoniun N(C₂H₃)₄Cl does not generate currents in SLC4A11-PS120 cells. Error bars, S.E. ***, p < 0.001.

Methylammonia N(CH₃)H₂ Can Also Be Transported by SLC4A11

Because ammonia-permeable Rh and AQPs can also transport methylammonia (MA) (20), we next examined whether this ammonia derivative can be transported by SLC4A11. N(CH₃)H₃Cl application produced similar pH-sensitive inward currents, smaller at low pH and larger at high pH (Fig. 8A). Average current densities (pA/pF) at extracellular pH 6.5, 7.5, and 8.5 were −0.01 ± 0.03, −1.05 ± 0.10, and −11.60 ± 1.03, respectively, which were significantly smaller than NH₄Cl-dependent current density (p = 0.048). This is consistent with lower free [MA] due to the higher pK_a of MA⁺ (pK_a 10.64) than pK_a of NH₄⁺ (pK_a 9.25). Fig. 8, A and C, shows that the transient currents elicited at the pH 8.5 → 7.5 transition after pH 8.5 MA were similar to what was observed with NH₄Cl. The E_rev shift from −56.88 ± 3.37 mV (n = 4) at pH 8.5 to −24.39 ± 2.93 mV (n = 4) at pH 7.5 was also similar to that observed with NH₄Cl. As we indicated above, application of the bulkier tetraethylammonium N(CH₂CH₃)₄⁺ did not induce any current (Fig. 8D), indicating that there is a size limit for substrates of SLC4A11.

SLC4A11 Stoichiometry

We next set out to determine the stoichiometry of NH₃/H⁺ transport through SLC4A11 at physiological pH 7.5. According to the equation,

where m is the stoichiometry of NH₃, n is the stoichiometry of H⁺, Z_m is the charge of NH₃ (Z_m = 0), Z_n is the charge of H⁺ (Z_n = +1), R is the gas constant 8.314 J/mol*K, F is the Faraday constant 96,485 C/mol, and T is temperature in K (295 K), there is a linear relationship between E_rev and log[NH₃]_o with a slope of ln(10) × RT/F × m/n if we assume that [NH₃]_i, [H]_i, and [H]_o are constant (21). Using this equation, we first predicted the values of linear slopes at each possible m (NH₃)/n (H⁺) stoichiometry (at 22 °C): (i) 58.5 mV/log mm if NH₃/H⁺ stoichiometry is 1:1; (ii) 29.25 mV/log mm if NH₃/H⁺ stoichiometry is 1:2; and (iii) 19.50 mV/log if NH₃/H⁺ stoichiometry is 1:3 (dashed lines in Fig. 9B). We then experimentally determined the E_rev of SLC4A11 NH₃/H⁺ currents at different [NH₃]_o by sequentially superfusing SLC4A11-PS120 cells with 1, 3, 10, 30, and 100 mm NH₄Cl E8 solutions (I2 pipette solution; Fig. 8A). Liquid junction potential was estimated to be 13.8 mV and was subtracted from E_rev values before analysis. Using the adjusted E_rev at the current peak at each concentration, the plot was linear (r² = 0.840) with a slope of 28.30 ± 2.78 mV/log mm (p < 0.0001), closest to the 1:2 NH₃/H⁺ stoichiometry. As shown in Fig. 8A, we observed some current decay at higher NH₄Cl concentrations, which could be due to an increased [NH₃]_i leading to decreased [NH₃]_o/[NH₃]_i gradient and in turn diminishing the current and underestimating the actual E_rev. To examine this possibility, we also plotted E_rev obtained at the eNH₄Cl-current steady-state phase (plateaus). Indeed, we observed decreased E_rev values at high NH₄Cl conditions. Fig. 9B shows the plot of E_rev at plateau versus log[NH₃]_o. The slope is 25.24 ± 3.46 mV/log mm (r² = 0.699, p < 0.0001) still closest to the 1:2 NH₃/H⁺ stoichiometry slope.

FIGURE 9.

Stoichiometry of SLC4A11 estimated to be 1:2 NH₃/H⁺. A, sample trace of whole-cell current recording in response to sequentially increasing extracellular (pH_o 7.5) NH₄Cl concentration in SLC4A11-PS120 cell at −60 mV holding potential. E_rev data for maximum “peak” current was obtained at points indicated by dark black arrows, whereas E_rev for “plateau” currents were obtained at gray arrows. B, reversal potentials obtained at maximum current (“peak”) were plotted against extracellular log[NH₃]_o. The plot shows a linear relationship between E_rev and log[NH₃]_o with a slope of 28.30 mV/log mm (r² = 0.840, p < 0.0001). Dashed lines, predicted NH₃/H⁺ 1:1, 1:2, and 1:3 slope. Reversal potentials obtained at “plateau” currents were also plotted against extracellular log[NH₃]_o, with a linear slope of 25.24 mV/log mm (r² = 0.699, p < 0.0001). Error bars, S.E.

NH₃/H⁺ current is not sensitive to DIDS or EIPA

Previous reports indicated that EIPA inhibits SLC4A11 Na⁺-H⁺ permeability (9). In addition, it has been reported that DIDS has no effect on SLC4A11 Na⁺-H⁺ permeability (10), whereas another group reported that DIDS enhances SLC4A11 H⁺ permeability (12). Therefore, we tested whether these two compounds affect the SLC4A11 NH₃/H⁺ current in our model system. However, we found that neither 100 μm DIDS nor 10 μm EIPA affected NH₃/H⁺ inward currents in SLC4A11-PS120 cells (Fig. 10). It is well known that DIDS is a broad spectrum inhibitor of anion transporters, whereas EIPA is a potent Na⁺/H⁺ exchanger inhibitor. These data suggest unique pharmacological properties for SLC4A11 expressed in NHE-deficient PS120 cells and indicate that the previous studies may be complicated by the presence of endogenous transporters influenced by SLC4A11 activity.

FIGURE 10.

SLC4A11 eNH₄Cl-dependent currents are inhibited by neither 10 μm EIPA nor 100 μm DIDS. A, sample whole-cell current recording of eNH₄Cl-dependent current in SLC4A11 cells before (a), during (b), and after (c) 10 μm EIPA application. B, bar graph summary of eNH₄Cl-dependent current density before (a), during (b), and after (c) EIPA application (n = 6) (repeated measure ANOVA, p = 0.298). C, representative whole-cell current recording of eNH₄Cl-dependent current in SLC4A11 cells before (d), during (e), and after (f) 100 μm DIDS application. D, bar graph summary of eNH₄Cl-dependent current density before (d), during (e), and after (f) DIDS application (n = 7) (repeated measure ANOVA, p = 0.847). Error bars, S.E.

Discussion

In this study, we examined the potential mechanism for apparent NH₄⁺ permeability of SLC4A11 (9). Using the NHE-deficient PS120 fibroblast cell line, we found evidence indicating that human SLC4A11 is an electrogenic NH₃/H⁺ transporter. As such, SLC4A11 should be considered when examining ammonia: H⁺ balance along with other recently identified NH₃ channels/transporters, namely selected AQPs (AQP8, AQP3, AQP7, AQP9, and AQP10) (14), Amt proteins, methylamine permease (22), and rhesus protein (23). In addition, our findings may also provide insight into the pathophysiology of corneal endothelial dystrophy (24, 25) and an explanation for altered renal function in SLC4A11 knock-out mice (7, 8).

We initially observed apparent SLC4A11 NH₄⁺ permeability during studies of acid load recovery using an ammonium pulse approach. Significant acidification during the ammonium pulse was only seen in SLC4A11-PS120 cells. When Boron et al. (17) first studied NH₃-induced intracellular pH transients in the squid giant axon, it was calculated that the intracellular acidification during NH₄Cl pulse was due to a finite permeability to NH₄⁺, and the acidification rate is determined mainly by the absolute NH₄⁺ permeability. Also, the extent of the acid loading effect in the wash phase positively correlates with NH₄⁺ permeability (17). Our data show that SLC4A11 transfection increased the apparent NH₄⁺ permeability in PS120 cells relative to the low endogenous level in EV. SLC4A11 high expression Col4 and low expression Col12 showed higher and lower acidification rates, indicating that the apparent NH₄⁺ (or NH₃-H⁺ as equivalent) permeability is directly associated with SLC4A11 expression level.

The electrophysiological recordings revealed that SLC4A11-induced NH₄⁺ (or NH₃-H⁺) permeability results in NH₄Cl-dependent currents. The inward currents vary directly with solution [NH₃] but not [NH₄⁺] and were relatively unaffected by removal of Na⁺, K⁺, or Cl⁻. This is very similar to that observed in AQP8, which was proposed to be permeable to NH₃ rather than NH₄⁺ due to the hydrophobic pore inside the channel (14, 23). Interestingly, there is evidence that SLC4A11 has water channel properties and shares one NPX motif with AQPs (13). NH₃ and H₂O share similar size and dipole moment (1.5 versus 1.8 debye), and the aromatic/arginine regions in AQPs provide selectivity of NH₃ over H₂O (15). In the current study, we observed an 84-fold increase in current in response to a 75-fold increase in [NH₃]_o (Fig. 2), and when [NH₃] is equalized in different pH solutions by adding more NH₄Cl at lower pH and less at high pH (Fig. 3), we observed similar current amplitude, strongly supporting NH₃ as the actual species transported. Furthermore, the ammonia analog MA has a similar molecular radius as NH₃ (26) but a significantly higher pK_a. So there is less free uncharged [MA] than [NH₃] at the same pH, and as a result, MA generated less current in SLC4A11 than NH₃.

A distinctive characteristic of SLC4A11 that we observed here, which differentiates SLC4A11 from other ammonia channels/transporters, is that NH₃ is evidently co-transported with H⁺. Unlike the ammonia- and H⁺-permeable AQP1 Arg-195 mutant that shows H⁺ currents in the absence of NH₃ (15)_, SLC4A11 shows no currents in response to proton gradient changes alone (Fig. 5A). Proton gradient changes can induce currents only after NH₄Cl pulse (Fig. 2, filled arrow, and Fig. 5, B and D). When equalizing [NH₃]_o (Fig. 3), extracellular pH or [H⁺] gradient differences can affect apparent kinetics and current density of eNH₄Cl-dependent current. Reversal potential analysis of currents during and after NH₄Cl pulse is also consistent with an NH₃/H⁺ co-transport hypothesis. Although H⁺ is the putative charge carrier in the currents of all electrogenic NH₃ channels/transporters, we show here, for the first time, evidence that H⁺ is co-transported with NH₃ in SLC4A11. A coupled transport of H⁺ and NH₃ may be more beneficial for the cell because it would produce smaller cytosolic pH_i transients than if NH₃ were transported alone.

As an NH₃/2H⁺ co-transporter, the direction of SLC4A11 transport depends on the [NH₃] gradient, [H⁺] gradient, and cell membrane potential. In most of the experiments, the holding potential was −60 mV, and we introduced an inward [NH₃] gradient that resulted in inward currents. However, we observed outward currents on multiple occasions: 1) during pH 8.5 wash after pH 8.5 NH₄Cl (Fig. 5C, time c); 2) prominent and prolonged outward currents during wash periods in cells dialyzed with NH₄Cl (Fig. 7A); 3) smaller outward currents observed during pH 8.5 wash following pH 7.5 eNH₄Cl (Fig. 5B); 4) outward break-in currents observed with I1-NH₄Cl dialysis if the holding potential was reduced to −10 mV (Fig. 7, D and E); 5) at −10 mV with I1-NH₄Cl dialysis, outward currents were observed in pH 7.5 wash after pH 7.5 NH₄Cl (Fig. 7D). Given a 1:2 NH₃/H⁺ stoichiometry of SLC4A11 at pH_o 7.5, in a typical cell with pH_i 7.1 and membrane potential of −50 mV, an ∼15-fold [NH₃] outward gradient is needed to drive SLC4A11 transport in the outward direction. This could potentially be achieved, for example, in glutamine-metabolizing cells that generate significant amounts of NH₃ (27, 28). Although NH₃ can diffuse across most cell membranes, a NH₃/H⁺ co-transporter like SLC4A11 may offer an advantage of directional regulated ammonia transport, which could play an important role in the function of ammonia detoxification like other ammonia channels (e.g. AQP8 (27)).

Although the net movement of NH₃/H⁺ is physiologically equivalent to NH₄⁺, transport of NH₃ over NH₄⁺ offers the advantage of high selectively. K⁺ and NH₄⁺ ions are similar in several characteristics that are important for membrane transport: 1) they share an almost identical radius (NH₄⁺ 0.148 nm versus K⁺ 0.141 nm) (29); 2) they have similar charge (15); and 3) hydration energies are also similar (30). Therefore, it is not surprising that multiple K⁺ transporters/channels allow substitution of NH₄⁺ for K⁺ (e.g. in the Na-K-2Cl co-transporter NKCC2 and barium-sensitive K⁺ channels (31)). Reverse substitution could possibly occur in NH₄⁺-selective transporters to carry K⁺ nonspecifically (15). To avoid this, the Amt NH₄⁺ channel uses a two-step translocation process involving a hydrophobic pore to discriminate NH₄⁺ from K⁺ (32). For SLC4A11, transport of NH₃ instead of NH₄⁺ may also help resolve the selectivity problem. Because K⁺ free extracellular and intracellular solutions were used in most of the experiments presented here, K⁺ involvement in the recorded currents is unlikely. Indeed, K⁺ is not a substrate for SLC4A11 shown in Fig. 4A comparing current density in 5 mm K⁺ E1 versus K⁺ free E4. This suggests that SLC4A11 utilizes a mechanism ensuring its relative selectivity. However, greater understanding of the molecular mechanisms underlying the processes would require further investigation.

Park et al. (10) reported increased baseline pH_i in SLC4A11-transfected HEK293 cells, whereas our previous study reported no difference (9). Consistent with a more recent report (12), here we observed reduced baseline pH_i in SLC4A11-PS120 cells (Fig. 6A). Although the H⁺ permeability is evident, in the absence of NH₃, the H⁺ fluxes are very small, and no measureable H⁺ current can be recorded. Interestingly, the effect on baseline pH_i is made most evident in PS120 cells because of the lack of Na⁺/H⁺ exchangers. Previous studies show that when a more robust bicarbonate buffering system is present, baseline pH_i differences disappear (9, 10, 12). This suggests that SLC4A11 H⁺ permeability in the absence of ammonia is a relatively small contributor to pH_i regulation, whereas ammonia transport through SLC4A11 is of more physiological relevance.

Park at al. (10) also concluded that SLC4A11 is an Na⁺-coupled borate transporter (with a capability to transport hydroxyl ions (OH⁻) in the absence of borate). Interestingly, boric acid is a weak acid with a pK_a of 9.24, which is very close to that of NH₃ (pK_a = 9.25), making it largely exist as uncharged H₃BO₃ at physiological pH (33). The structure of boric acid is comparable with that of NH₃, with the central boron atom possessing three valence electrons, as nitrogen does in NH₃. The molecular radius of boric acid (2.573 Å) is comparable with that of urea (2.618 Å), which is also permeable through some ammonia-permeable AQPs (14). If a hydrophobic environment does exist in SLC4A11 supporting NH₃ transport, it may provide a path for the uncharged boric acid.

This study has several limitations. Because of the ability of Na⁺ to modulate SLC4A11 activity at pH 8.5, we only determined the apparent stoichiometry of NH₃/H⁺ transport at the physiological pH 7.5, under which Na⁺ does not appear to have any effect on eNH₄Cl-dependent currents. We are also aware of the limitation of sequential NH₄Cl superfusion during which the assumption of constant [NH₃]_i and [H⁺]_i is uncertain. Indeed, there was current inhibition at high NH₄Cl concentrations, possibly due to increased [NH₃]_i, which could lead to an underestimation of E_rev. Notably, plotting the plateau E_rev data versus log[NH₃]_o yielded a smaller slope, however still closest to the 1:2 NH₃/H⁺ model. Also, because NH₃ is diffusible through the PS120 cell membrane (shown in Fig. 1), we are not able to control [NH₃]_i and [H⁺]_i more tightly. In addition, the possibility of NH₄⁺-H⁺-coupled transport cannot be completely eliminated. Therefore, future investigations, such as SLC4A11 protein reconstitution into a planar lipid bilayer or SLC4A11 expression in an NH₃-impermeable cell line, would be useful to clarify the uncertainties with the 1:2 NH₃/H⁺ model.

In healthy individuals, urine ammonia concentration is around 30–50 mm, whereas plasma ammonia concentration ranges between 0.005 and 0.02 mm. Plasma ammonia can rise to 0.1 mm under pathophysiological conditions. The fact that we were able to record currents from 1 to 100 mm eNH₄Cl (0.0017–17 mm (NH₃)) with no saturation indicates that SLC4A11 can efficiently function in the physiological ammonia range. Ammonia channels/transporters play important roles in nitrogen homeostasis, not only in the major nitrogen handling organs like liver and kidney (27, 34), but also in lung (35), male reproductive system (36), and gastrointestinal tract (37). Directional ammonia transport capacity offered by SLC4A11 can play essential roles in ammonia detoxification in a wide range of tissues/cells. Especially in cells that actively metabolize glutamine and glutamate, such as neurons, skeletal muscle cells, and gastrointestinal tract epithelium (38, 39), ammonia is a major waste product that needs to be extruded. Indeed, SLC4A11 is not only associated with corneal endothelial dystrophy, but also may play roles in other physiological and pathophysiology processes. The ubiquitous expression of SLC4A11, in cornea endothelium, kidney, inner ear, salivary glands (submandibular and parotid), thyroid, mammary gland, testis, trachea, esophagus, pancreas, liver, spleen, and cerebellum (1, 6, 10) suggests a diverse physiological role. Because SLC4A11 is highly expressed in cornea endothelium, fibrocytes of the inner ear, and the loop of Henle, these three tissues could be the starting point for investigating SLC4A11 physiological function in regulating tissue nitrogen levels. It is well known that cornea and inner ear function depends on tightly controlled fluid and ion transport. At least 60 mutations have been identified as being associated with corneal endothelial dystrophy and/or perceptive deafness (40). A recent clinical study suggests that homozygous SLC4A11 mutated congenital hereditary endothelial dystrophy patients progress to Harboyan syndrome at a later age (perceptive deafness and corneal endothelial dystrophy (41)), whereas another study associated SLC4A11 mutants with the corneal ectasia disorder keratoconus (42). In the kidney, SLC4A11 is highly expressed in the thin descending limb of loop of Henle (7). Henle's loop epithelium has the capacity to reabsorb ammonia into interstitial spaces and, through the countercurrent multiplication mechanism, create a cortical-medullary increasing ammonium gradient to regulate urine ammonium excretion (18). In SLC4A11 knock-out mice, an impairment in the countercurrent multiplication process was suggested based on polyuria and urine hypo-osmolality (7). The ammonia transport property of SLC4A11 suggests that it can potentially play a role in medullary interstitial ammonia handling and ammonia countercurrent multiplication (18, 31), filling the missing piece of renal ammonia transport (31).

In summary, we have identified SLC4A11 as an additional member of the subfamily of ammonia channels and transporters and as an H⁺-coupled ammonia transporter with an apparent stoichiometry of 1:2 NH₃/H⁺. The SLC4A11 ammonia current is independent of K⁺, Na⁺, and Cl⁻ at physiological pH and is uniquely insensitive to DIDS and EIPA. These findings provide mechanistic insights into the physiological and pathophysiological processes that may occur in the cornea, kidney, and inner ear.