SLC4A11 function: evidence for H⁺(OH⁻) and NH₃-H⁺ transport

Abstract

Whether SLC4A11 transports ammonia and its potential mode of ammonia transport (${\text{NH}}_4^{+}$, NH₃, or NH₃-2H⁺ transport have been proposed) are controversial. In the absence of ammonia, whether SLC4A11 mediates significant conductive H⁺(OH⁻) transport is also controversial. The present study was performed to determine the mechanism of human SLC4A11 ammonia transport and whether the transporter mediates conductive H⁺(OH⁻) transport in the absence of ammonia. We quantitated H⁺ flux by monitoring changes in intracellular pH (pH_i) and measured whole cell currents in patch-clamp studies of HEK293 cells expressing the transporter in the absence and presence of NH₄Cl. Our results demonstrate that SLC4A11 mediated conductive H⁺(OH⁻) transport that was stimulated by raising the extracellular pH (pH_e). Ammonia-induced HEK293 whole cell currents were also stimulated by an increase in pH_e. In studies using increasing NH₄Cl concentrations with equal ${\text{NH}}_4^{+}$ extracellular and intracellular concentrations, the shift in the reversal potential (E_rev) due to the addition of ammonia was compatible with NH₃-H⁺ transport competing with H⁺(OH⁻) rather than NH₃-nH⁺ (n ≥ 2) transport. The increase in equivalent H⁺(OH⁻) flux observed in the presence of a transcellular H⁺ gradient was also compatible with SLC4A11-mediated NH₃-H⁺ flux. The NH₃ versus E_rev data fit a theoretical model suggesting that NH₃-H⁺ and H⁺(OH⁻) competitively interact with the transporter. Studies of mutant SLC4A11 constructs in the putative SLC4A11 ion coordination site showed that both H⁺(OH⁻) transport and ammonia-induced whole cell currents were blocked suggesting that the H⁺(OH⁻) and NH₃-H⁺ transport processes share common features involving the SLC4A11 transport mechanism.

INTRODUCTION

SLC4A11 belongs to the SLC4 gene family of ${\text{HCO}}_3^{-}$ and ${\text{CO}}_3^{2-}$ transporters (42). The amino acid sequence of SLC4A11 is least homologous to other SLC4 transporters, and its function has remained controversial. It was originally thought that SLC4A11 functions as an electrogenic sodium borate cotransporter [Na⁺(n)-B(OH⁻)⁴⁻] or that it mediates cation (Na⁺ or H⁺) permeation (41). It was subsequently reported that SLC4A11 mediates Na⁺-OH⁻ cotransport (equivalent to Na⁺/H⁺ exchange) (20, 38), ${\text{NH}}_4^{+}$ transport (38), water flux (50, 54), H⁺(OH⁻) transport (22, 23, 37, 41), NH₃-2H⁺ cotransport (ammonia-driven H⁺ transport) (23, 30, 61), and NH₃ transport (33). SLC4A11 does not transport ${\text{HCO}}_3^{-}$ or ${\text{CO}}_3^{2-}$ unlike other SLC4 transporters (20, 33, 38).

SLC4A11 is expressed in multiple tissues including testis, thyroid gland, inner ear, mammary glands, esophagus, cerebellum, kidney, salivary glands, pancreas, spleen, liver, cornea, and trachea (8, 34, 41, 43). Naturally occurring mutations in SLC4A11 cause corneal disorders, including all cases of autosomal recessive congenital hereditary endothelial dystrophy (CHED) (2, 21, 51, 55), Harboyan syndrome (CHED with sensorineural hearing loss) (11, 49), and some cases of Fuchs’ endothelial corneal dystrophy (44, 50, 56). Patients with CHED have diffuse corneal edema shortly after birth with a diffusely edematous corneal stroma, and there is thickening of Descemet’s membrane and abnormal corneal endothelial morphology (12). In mice the loss of the Slc4a11 gene causes corneal endothelial dystrophy recapitulating the human disease (17), auditory sensorineural abnormalities (34), and renal ion and water transport abnormalities (15, 17).

The corneal endothelium utilizes glutamine to produce ATP for Na⁺-K⁺-ATPase activity to support physiological pump function (60). High levels of glutaminolysis lead to the production of ammonia as a byproduct. Slc4a11^−/− mice have disrupted expression of enzymes involved in glutamine metabolism with a predicted impairment in pump function and elevated levels of nitrotyrosine staining, a general marker of ammonia toxicity (60). Similar changes in glutaminolysis enzyme expression were detected in a conditionally immortal Slc4a11^−/− mouse corneal endothelial cell line (62). Given the potential importance of glutamine metabolism resulting in the generation of ammonia, there is a need for a cellular ammonia transport process to prevent its accumulation with resultant ammonia toxicity.

The role of SLC4A11 in mediating ammonia transport and the underlying transport mechanism is controversial. Ogando et al. (38) first suggested that human SLC4A11 expressed in PS120 cells transports ${\text{NH}}_4^{+}$. Zhang et al. (61) reported that the transporter mediates NH₃-2H⁺ cotransport in PS120 cells. Loganathan et al. (33) found that SLC4A11 expressed in oocytes transports NH₃ rather than NH₃-2H⁺. Myers et al. (37) concluded that mouse Slc4a11 does not transport ammonia per se. In the absence of ammonia, whether SLC4A11 mediates conductive H⁺(OH⁻) transport is also controversial. Kao et al. (22, 23) and Myers et al. (37) have provided evidence for SLC4A11-mediated conductive H⁺(OH⁻) transport whereas others have been unable to detect significant conductive H⁺(OH⁻) transport (33, 61). These discrepant results prompted our re-examination of the transport properties of SLC4A11, specifically whether it mediates H⁺(OH⁻) transport and the mechanism of ammonia transport. Our results indicate that SLC4A11 mediates conductive H⁺(OH⁻) transport in the absence of ammonia, and in the presence of ammonia our findings are compatible with an NH₃-H⁺ mode of transport. Furthermore, we have generated a theoretical model characterizing H⁺(OH⁻) and NH₃-H⁺ competition for transport through SLC4A11.

METHODS

HEK293 Cell Culture and Transfection

The cells were grown in DMEM supplemented with 10% FBS, l-glutamine (200 mg/L), and 1% penicillin-streptomycin at 37°C in a 5% CO₂ atmosphere and passaged onto 60-mm dishes (sulpho-NHS-SS-biotin studies) or onto PEI-coated coverslips (functional studies). The cells were transiently transfected using Lipofectamine 2000 (Thermo Fisher Scientific) according to the manufacturer's protocol except that the transfection solution was removed after a 2-h exposure and the cells were studied 24 h posttransfection. The cells were transfected with empty pTT vector (mock transfected) or various human SLC4A11-C constructs. Enhanced green fluorescent protein (eGFP)-coupled constructs were used to identify expressing cells in the patch-clamp experiments.

Measurement of Intracellular pH and H⁺ Flux

Following transient transfection with the various constructs, intracellular pH (pH_i) was measured in cells grown on coated coverslips after 24 h. The coverslips were placed in a custom-designed chamber on the stage of a microscope fluorometer (28) and loaded with the fluorescent pH_i probe BCECF using esterified BCECF-AM (Life Technologies) at room temperature for ∼20 min. The composition of the solution used for dye loading was (in mM) 140 KCl, 5 tetramethylammonium chloride (TMACl), and 5 HEPES. In each experiment, the fluorescence excitation ratio (500 nm/440 nm; 530-nm emission) obtained from ∼200 cells was averaged. The bathing solutions continuously perfused the coverslips at 2 ml/min (37°C). At the end of each experiment the intracellular fluorescence excitation ratio was calibrated with 5 μM valinomycin (Sigma-Aldrich) and 26 μM nigericin (Sigma-Aldrich) to equilibrate pH_i, extracellular pH (pH_e), and K⁺ (45). With the use of ${\text{NH}}_4^{+}$ addition/removal protocols, the intrinsic cell buffer capacity (β_i) was measured in HEPES-buffered solutions over a range of intracellular H⁺ (H_in⁺) values and calculated as Δ[${\text{NH}}_4^{+}$]/Δ[H_in⁺]. In the initial 10–15 s following a bath solution switch, the rate of change of pH_i (dpH_i/dt) was measured and converted to the rate of change of [H_in⁺] (d[H_in⁺]/dt). H⁺ flux (mM/s) was calculated as β_i × (d[H_in⁺]/dt). In NH₄Cl-containing solutions, the ammonia buffer capacity β_NH4 was calculated as 2.303 [${\text{NH}}_4^{+}$]_i and added to β_i to calculate the equivalent H⁺ flux. The flux of H⁺ versus OH⁻ flux cannot be distinguished thermodynamically and is referred to in the text as SLC4A11-mediated H⁺(OH⁻) flux for simplicity. The solutions used were nominally ${\text{HCO}}_3^{-}$-free, and since HEK293 cells have endogenous Na⁺-coupled ${\text{HCO}}_3^{-}$ transport activity (40) to prevent the generation of ${\text{HCO}}_3^{-}$ from ambient CO₂, the solutions were bubbled continuously with 100% O₂.

H⁺(OH⁻) Flux Following an Increase in pH_e

Transient H⁺(OH⁻) flux was induced following an increase in bath pH from 7.4 to 8.0 followed by a return to pH 7.4 under Na⁺-free conditions. The cells were initially bathed in a solution containing (in mM) 140 KCl, 5 TMACl, and 5 HEPES, pH 7.4, and after pH_i stabilized, the external solution pH was acutely increased from pH 7.4 to 8.0. H⁺(OH⁻) flux was measured following the acute increase in external pH to 8.0 and following the return to pH 7.4.

Equivalent H⁺(OH⁻) Flux Following an Increase in pH_e in the Presence of NH₄Cl

Transient H⁺ flux was induced following an increase in bath pH from 7.4 to 8.0 followed by a return to pH 7.4 under Na⁺-free conditions in the presence of 5 mM NH₄Cl. In these experiments the NH₃ concentration was kept constant at 0.153 mM at both pH values. The ${\text{NH}}_4^{+}$ concentration decreased from 4.85 to 1.22 mM at pH 8.0. The cells were initially bathed in solution that contained (in mM) 140 KCl, 5 NH₄Cl, and 5 HEPES, pH 7.4, and after pH_i stabilized, the external pH was acutely increased from pH 7.4 to 8.0 with a solution that contained (in mM) 140 KCl, 3.63 TMACl, and 1.37 NH₄Cl, and 5 HEPES, pH 8.0. Equivalent H⁺(OH⁻) flux was measured following the acute increase in external pH to 8.0 and following the return to pH 7.4.

Patch-Clamp Measurements of Electrogenic H⁺(OH⁻) and Ammonia-Induced Whole Cell Currents

The cells were replated (1:10 dilution) onto coated 35-mm tissue culture inserts (Bioptechs) 24 h following transfection. The tissue culture inserts were put onto a chamber on the stage of a fluorescent microscope (Axioskop 2 FS plus; Carl Zeiss). Cells with moderate eGFP fluorescence were chosen 2–3 h later for electrophysiological recording using a MultiClamp 700B patch amplifier (Molecular Devices). Borosilicate glass patch pipettes were prepared with a tip resistance 4–6.5 MΩ (tip diameter of 1–1.5 μm). A microagar salt bridge containing 2 M KCl was built into the electrode holder to ensure stable electrode potentials during recording (47). Using whole cell voltage clamp, steady-state currents were measured with a holding potential of −15 mV and a series of voltage pulses (400-ms) at 10-mV increments (−70 to +40 mV). Using the pClamp software (Clampex 10, Molecular Devices), the signal was low-pass filtered at 400 Hz and sampled at 2 kHz. Junction potentials calculated with Clampex 10 that were generated from different compositions of patch pipette and bath solutions were corrected for at all applied potentials. With the auto-whole cell capacitance and series resistance compensation, whole cell capacitance and series resistance were determined. Typically, the series resistance was compensated 80% (both correction and prediction). The cells were continuously perfused at a flow rate of ∼2 ml/min (22–24°C). Whole cell currents were measured using the solutions in the protocols described below.

H⁺(OH⁻) currents at different pH_e values.

The patch pipette solution contained (in mM) 45 cesium methanesulfonate, 1 CaCl₂, 10 tetraethylammonium chloride (TEA-Cl), 10 EGTA, and 100 HEPES, pH 7.4. For this and subsequent protocols, 100 mM HEPES were used to minimize any patch pipette pH changes due to pH_e changes or changes in extracellular NH₃; the bath solutions contained (in mM) 140 TMACl, 1.5 CaCl₂, 10 CsCl, and 10 HEPES. The bath pH was varied from 7.0 to 8.0. In this and subsequent protocols, cesium hydroxide (55 mM) was added to the patch pipette solution to pH the solution to 7.4. The final measured patch pipette osmolality was ~265 mosmol/kgH₂O with a bath osmolality of ~295 mosmol/kgH₂O.

Ammonia-induced currents at different pH_e values at constant ${NH}_{\mathit{4}}^{+}$ concentration and varying NH₃ concentrations.

The patch pipette solution contained (in mM) 45 cesium methanesulfonate, 2 NH₄Cl (0.023 NH₃; 1.977 ${\text{NH}}_4^{+}$), 1 CaCl₂, 8 TEA-Cl, 10 EGTA, and 100 HEPES, pH 7.4; the bath solution pH 7.0 contained (in mM) 138.014 TMACl, 1.986 NH₄Cl (0.009 NH₃; 1.977 ${\text{NH}}_4^{+}$), 1.5 CaCl₂, 10 CsCl, and 10 HEPES; pH 7.4 contained (in mM) 138 TMACl, 2 NH₄Cl (0.023 NH₃; 1.977 ${\text{NH}}_4^{+}$), 1.5 CaCl₂, 10 CsCl, and 10 HEPES; pH 7.6 contained (in mM) 137.987 TMACl, 2.013 NH₄Cl (0.036 NH₃; 1.977 ${\text{NH}}_4^{+}$), 1.5 CaCl₂, 10 CsCl, and 10 HEPES; pH 7.8 contained (in mM) 137.966 TMACl, 2.034 NH₄Cl (0.057 NH₃; 1.977 ${\text{NH}}_4^{+}$), 1.5 CaCl₂, 10 CsCl, and 10 HEPES; and pH 8.0 contained (in mM) 137.932 TMACl, 2.068 NH₄Cl (0.091 NH₃; 1.977 ${\text{NH}}_4^{+}$), 1.5 CaCl₂, 10 CsCl, and 10 HEPES.

Ammonia-induced currents following an increase in bath pH from 7.4 to pH 8.0 with an ${NH}_{\mathit{4}}^{+}$ concentration gradient at pH 8.0 and no NH₃ concentration gradient.

Protocols a–e were utilized. Protocol a used patch pipette and bath pH 7.4: NH₃ = 0.006 mM; ${\text{NH}}_4^{+}$ = 0.494 mM; bath 8.0: NH₃ = 0.006 mM; ${\text{NH}}_4^{+}$ = 0.124 mM. The patch pipette solution contained (in mM) 45 cesium methanesulfonate, 0.5 NH₄Cl, 1 CaCl₂, 9.5 TEA-Cl, 10 EGTA, and 100 HEPES, pH 7.4; the bath solution contained (in mM) 139.5 TMACl, 0.5 NH₄Cl,1.5 CaCl₂, 10 CsCl, and 10 HEPES, pH 7.4; 139.87 TMACl, 0.13 NH₄Cl, 1.5 CaCl₂, 10 CsCl, and 10 HEPES, pH 8.0. Protocol b used patch pipette and bath pH 7.4: NH₃ = 0.011 mM; ${\text{NH}}_4^{+}$ = 0.989 mM; bath 8.0: NH₃ = 0.011 mM; ${\text{NH}}_4^{+}$= 0.249 mM. The patch pipette solution contained (in mM) 45 cesium methanesulfonate, 1.0 NH₄Cl, 1 CaCl₂, 9.0 TEA-Cl, 10 EGTA, and 100 HEPES, pH 7.4; the bath solution contained (in mM) 139.0 TMACl, 1.0 NH₄Cl,1.5 CaCl₂, 10 CsCl, and 10 HEPES, pH 7.4; 139.74 TMACl, 0.26 NH₄Cl, 1.5 CaCl₂, 10 CsCl, and 10 HEPES, pH 8.0. Protocol c used patch pipette and bath pH 7.4: NH₃ = 0.023 mM; ${\text{NH}}_4^{+}$ = 1.977 mM; bath 8.0: NH₃ = 0.023 mM; ${\text{NH}}_4^{+}$ = 0.497 mM. The patch pipette solution contained (in mM) 45 cesium methanesulfonate, 2 NH₄Cl, 1 CaCl₂, 8 TEA-Cl, 10 EGTA, and 100 HEPES, pH 7.4; the bath solution contained (in mM) 138 TMACl, 2.0 NH₄Cl,1.5 CaCl₂, 10 CsCl, and 10 HEPES, pH 7.4; 139.48 TMACl, 0.52 NH₄Cl, 1.5 CaCl₂, 10 CsCl, and 10 HEPES, pH 8.0. Protocol d used patch pipette and bath pH 7.4: NH₃ = 0.034 mM; ${\text{NH}}_4^{+}$ = 2.966 mM; bath 8.0: NH₃ = 0.034 mM; ${\text{NH}}_4^{+}$ = 0.746 mM. The patch pipette solution contained (in mM) 45 cesium methanesulfonate, 3 NH₄Cl, 1 CaCl₂, 7 TEA-Cl, 10 EGTA, and 100 HEPES, pH 7.4; the bath solution contained (in mM) 137 TMACl, 3.0 NH₄Cl,1.5 CaCl₂, 10 CsCl, and 10 HEPES, pH 7.4; 139.22 TMACl, 0.78 NH₄Cl, 1.5 CaCl₂, 10 CsCl, and 10 HEPES, pH 8.0. Protocol e used patch pipette and bath pH 7.4: NH₃ = 0.057 mM; ${\text{NH}}_4^{+}$ = 4.943 mM; bath 8.0: NH₃ = 0.057 mM; ${\text{NH}}_4^{+}$ =1.243 mM. The patch pipette solution contained (in mM) 45 cesium methanesulfonate, 5 NH₄Cl, 1 CaCl₂, 5 TEA-Cl, 10 EGTA, and 100 HEPES, pH 7.4; the bath solution contained (in mM) 135 TMACl, 5.0 NH₄Cl,1.5 CaCl₂, 10 CsCl, and 10 HEPES, pH 7.4; 138.7 TMACl, 1.3 NH₄Cl, 1.5 CaCl₂, 10 CsCl, and 10 HEPES, pH 8.0.

Ammonia-induced currents following an increase in bath pH from 7.4 to pH 8.0 with an NH₃ concentration gradient at pH 8.0 and no ${NH}_{\mathit{4}}^{+}$ concentration gradient.

Protocols a-f were utilized. Protocol a used patch pipette and bath pH 7.4: NH₃ = 0.006 mM; ${\text{NH}}_4^{+}$ = 0.494 mM; bath 8.0: NH₃ = 0.023 mM; ${\text{NH}}_4^{+}$ = 0.494 mM. The patch pipette solution contained (in mM) 45 cesium methanesulfonate, 0.5 NH₄Cl, 1 CaCl₂, 9.0 TEA-Cl, 10 EGTA, and 100 HEPES, pH 7.4; the bath solution contained (in mM) 139.5 TMACl, 0.5 NH₄Cl,1.5 CaCl₂, 10 CsCl, and 10 HEPES, pH 7.4; 139.483 TMACl, 0.517 NH₄Cl, 1.5 CaCl₂, 10 CsCl, and 10 HEPES, pH 8.0. Protocol b used patch pipette and bath pH 7.4: NH₃ = 0.011 mM; ${\text{NH}}_4^{+}$ = 0.989 mM; bath 8.0: NH₃ = 0.045 mM; ${\text{NH}}_4^{+}$ 0.989 mM. The patch pipette solution contained (in mM) 45 cesium methanesulfonate, 1.0 NH₄Cl, 1 CaCl₂, 9.0 TEA-Cl, 10 EGTA, and 100 HEPES, pH 7.4; the bath solution contained (in mM) 139.0 TMACl, 1.0 NH₄Cl, 1.5 CaCl₂, 10 CsCl, and 10 HEPES, pH 7.4; 138.966 TMACl, 1.034 NH₄Cl, 1.5 CaCl₂, 10 CsCl, and 10 HEPES, pH 8.0. Protocol c used patch pipette and bath pH 7.4: NH₃ = 0.023 mM; ${\text{NH}}_4^{+}$ = 1.977 mM; bath 8.0: NH₃ = 0.091 mM; ${\text{NH}}_4^{+}$ = 1.977 mM. The patch pipette solution contained (in mM) 45 cesium methanesulfonate, 2 NH₄Cl, 1 CaCl₂, 8 TEA-Cl, 10 EGTA, and 100 HEPES, pH 7.4; the bath solution contained (in mM) 138 TMACl, 2.0 NH₄Cl,1.5 CaCl₂, 10 CsCl, and 10 HEPES, pH 7.4; 137.932 TMACl, 2.068 NH₄Cl, 1.5 CaCl₂, 10 CsCl, and 10 HEPES, pH 8.0. Protocol d used patch pipette and bath pH 7.4: NH₃ = 0.034 mM; ${\text{NH}}_4^{+}$ = 2.966 mM; bath 8.0: NH₃ = 0.136 mM; ${\text{NH}}_4^{+}$ = 2.966 mM. The patch pipette solution contained (in mM) 45 cesium methanesulfonate, 3 NH₄Cl, 1 CaCl₂, 7 TEA-Cl, 10 EGTA, and 100 HEPES, pH 7.4; the bath solution contained (in mM) 137 TMACl, 3.0 NH₄Cl,1.5 CaCl₂, 10 CsCl, and 10 HEPES, pH 7.4; 136.898 TMACl, 3.102 NH₄Cl, 1.5 CaCl₂, 10 CsCl, and 10 HEPES, pH 8.0. Protocol e used patch pipette and bath pH 7.4: NH₃ = 0.057 mM; ${\text{NH}}_4^{+}$ = 4.943 mM; bath 8.0: NH₃ = 0.227 mM; ${\text{NH}}_4^{+}$= 4.943 mM. The patch pipette solution contained (in mM) 45 cesium methanesulfonate, 5 NH₄Cl, 1 CaCl₂, 5 TEA-Cl, 10 EGTA, and 100 HEPES, pH 7.4; the bath solution contained (in mM) 135 TMACl, 5.0 NH₄Cl,1.5 CaCl₂, 10 CsCl, and 10 HEPES, pH 7.4; 134.83 TMACl, 5.17 NH₄Cl, 1.5 CaCl₂, 10 CsCl, and 10 HEPES, pH 8.0. Protocol f used patch pipette and bath pH 7.4: NH₃ = 0.114 mM; ${\text{NH}}_4^{+}$ = 9.886 mM; bath 8.0: NH₃ = 0.454 mM; ${\text{NH}}_4^{+}$ = 9.886 mM. The patch pipette solution contained (in mM) 45 cesium methanesulfonate, 10 NH₄Cl, 1 CaCl₂, 10 EGTA, and 100 HEPES, pH 7.4; the bath solution contained (in mM) 130 TMACl, 10 NH₄Cl, 1.5 CaCl₂, 10 CsCl, and 10 HEPES, pH 7.4; 129.66 TMACl, 10.34 NH₄Cl, 1.5 CaCl₂, 10 CsCl, and 10 HEPES, pH 8.0.

Mutant H⁺(OH⁻) currents and ammonia-induced currents at pH 8.0 with an NH₃ concentration gradient at pH 8.0 and no ${NH}_{\mathit{4}}^{+}$ concentration gradient.

In the absence of ammonia, H⁺(OH⁻) currents were measured at external pH 7.4 and 8.0. The patch pipette solution contained (in mM) 45 cesium methanesulfonate, 1 CaCl₂, 10 TEA-Cl, 10 EGTA, and 100 HEPES, pH 7.4; the bath solutions contained (in mM) 140 TMACl, 1.5 CaCl₂, 10 CsCl, and 10 HEPES. In the presence of ammonia, the patch pipette solution contained (in mM) 45 cesium methanesulfonate, 0.500 mM NH₄Cl (0.006 NH₃; 0.494 ${\text{NH}}_4^{+}$), 1 CaCl₂, 10 TEA-Cl, 10 EGTA, and 100 HEPES, pH 7.4; the bath solution pH 7.4 contained (in mM) 139.5 TMACl, 0.500 NH₄Cl (0.006 NH₃; 0.494 ${\text{NH}}_4^{+}$), 1.5 CaCl₂, 10 CsCl, and 10 HEPES; and at pH 8.0 contained (in mM) 140 TMACl, 0.517 NH₄Cl (0.023 NH₃; 0.494 ${\text{NH}}_4^{+}$), 1.5 CaCl₂, 10 CsCl, and 10 HEPES.

Patch-Clamp Experiments Monitoring of pH_i

EYFP is a pH-sensitive fluorophore with a pK of 7.1 (32). pH_i was monitored with EYFP in patch-clamped cells expressing a pcDNA3-EYFP-SLC4A11 construct. Fluorescent images of the cell expressing EYFP were acquired using a Mightex camera (CXE-B013-U; Mightexsystems) every 10 s, and the cell fluorescence was quantitated and subtracted from the background fluorescence. pH_i was monitored while the bath pH was changed from pH 7.4 to 8.0. The following solutions were used: the patch pipette solution contained (in mM) 45 cesium methanesulfonate, 1 CaCl₂, 10 TEA-Cl, 10 EGTA, and 100 HEPES, pH 7.4; the bath solutions contained (in mM) 140 TMACl, 1.5 CaCl₂, 10 CsCl, and 10 HEPES. In separate experiments, pH_i was monitored while the extracellular NH₃ concentration was increased from 0.114 to 0.454 mM at the same time the bath pH was increased from pH 7.4 to 8.0. The solutions used were as follows: the patch pipette solution contained (in mM) 45 cesium methanesulfonate, 10 NH₄Cl (0.114 NH₃; 9.886 ${\text{NH}}_4^{+}$), 1 CaCl₂, 10 EGTA, and 100 HEPES, pH 7.4; the bath solution pH 7.4: contained (in mM) 130 TMACl, 10 NH₄Cl (0.114 NH₃; 9.886 ${\text{NH}}_4^{+}$), 1.5 CaCl₂, 10 CsCl, and 10 HEPES; and the bath solution pH 8.0 contained (in mM) 129.66 TMACl, 10.34 NH₄Cl (0.454 NH₃; 9.886 ${\text{NH}}_4^{+}$), 1.5 CaCl₂, 10 CsCl, and 10 HEPES.

Sulfo-NHS-SS-Biotin Plasma Membrane Labeling

Plasma membrane proteins were labeled with sulfo-NHS-SS-biotin and using streptavidin-agarose resin were pulled down according to the following protocol: 24 h after transfection, the cells were washed repeatedly with ice-cold PBS (pH 8.0) and then incubated at pH 8.0 (4°C for 30 min) with 1.1 mM sulfo-NHS-SS-biotin (Thermo Fisher Scientific). The reaction was stopped with 50 mM Tris buffer (containing 140 mM NaCl, pH 8.0, 4°C). The cells were then collected, washed with PBS, and lysed on ice in a buffer containing 150 mM NaCl, 1% (vol/vol) Igepal (Sigma-Aldrich), 0.5% sodium deoxycholate (Thermo Fisher Scientific), 5 mM EDTA (Sigma-Aldrich), and 10 mM Tris·HCl, pH 7.5, with protease inhibitors (Roche Life Sciences). Following the pelleting of insoluble material [10-min centrifugation (20,000 g at 4°C)], the supernatant which contained >90% of the plasma membrane protein fraction was collected and incubated for 4 h (4°C) with 50 μl of streptavidin-agarose resin (Thermo Fisher Scientific) on a rotating shaker. Following brief centrifugation to pellet the resin and washing with the lysis buffer, bound proteins were eluted at 60°C for 5 min with 2× SDS buffer containing 2% β-mercaptoethanol (EMD Millipore, Billerica, MA). A previously characterized rabbit polyclonal antibody (1:3,000 dilution) against a carboxy terminus epitope in human SLC4A11 was used to probe immunoblots of plasma membrane pulled down proteins and whole cell lysates (23, 34).

Modeling SLC4A11 Ion Coordination Site

A homology model of the outward facing state of SLC4A11 was built with Modeller 9.18 (58), using the available crystal structure of human AE1 (4) and cryoEM structure of human NBCe1-A (19) as templates. The multiple sequence alignment of SLC4A11 and the two templates (AE1 and NBCe1) was done in advance with ClustalW (53). Five hundred three-dimensional (3D) models of SLC4A11 were then generated with the automodel algorithm implemented in Modeller 9.18, and their overall quality was assessed with DOPE and GA341 scores (36, 48). The best 10 models were analyzed for further quality assessment by manual comparison of several key structural elements identified previously from functional mutagenesis in the two templates. One final model, reproducing successfully the position and orientation of the functionally critical residues in the templates was selected. The SLC4a11 ion coordination site figure was prepared with PyMol (46).

Statistics

Student’s t test was used to compare group means. One-way ANOVA and Dunnett’s t test were used to compare multiple study group means with a control group. The results are depicted as means ± SE. P < 0.05 was considered statistically significant.

RESULTS

H⁺(OH⁻) Flux Following an Increase in pH_e in the Absence of Ammonia

pH_i transients were monitored following an increase in pH_e from 7.4 to 8.0 and return to pH 7.4. As shown in representative experiments in Fig. 1, the changes in pH_i induced by the bath pH changes were significantly greater in SLC4A11-expressing cells. The flux of H⁺(OH⁻) (pH 7.4 to 8.0) was 0.011 ± 0.002 mM/s (n = 8) in mock-transfected cells versus 0.053 ± 0.004 (n = 19) (P < 0.001) in SLC4A11-expressing cells and the flux of H⁺(OH⁻) (pH 8.0 to 7.4) was 0.009 ± 0.002 mM/s (n = 8) in mock-transfected cells versus 0.062 ± 0.005 mM/s (n = 19) (P < 0.001) in SLC4A11-expressing cells.

Fig. 1.

Changes in intracellular pH (pH_i) induced by extracellular pH (pH_e) changes (7.4–8.0–7.4). Transient H⁺(OH⁻) flux was induced following an increase in bath pH from 7.4 to 8.0 followed by a return to pH 7.4. A: mock-transfected cells. B: SLC4A11-transfected cells.

Whole Cell H⁺ Currents at Various pH_e Values

Whole cell currents were measured at pH_e values between 7.0 and 8.0 with the patch pipette at pH 7.4 in mock-transfected and SLC4A11-expressing cells. The whole cell currents are depicted in Fig. 2, A–C. Following mock subtraction, the current was measured at +40 mV and the conductance was measured at 0 mV at each pH value. The results were as follows: pH 7.0 (n = 6): 5.25 ± 0.588 pA; 0.292 ± 0.038 nS; pH 7.4 (n = 6): 5.36 ± 1.71 pA (n = 6); 0.304 ± 0.046 nS; pH 7.6 (n = 5): 32.1 ± 3.46 pA; 0.820 ± 0.066 nS; pH 7.8 (n = 7): 58.0 ± 3.56 pA; 1.04 ± 0.084 nS; pH 8.0 (n = 3): 92.7 ± 5.45 pA; 1.41 ± 0.078 nS. The reversal potential (E_rev) was measured at each pH value following mock subtraction. Figure 2D depicts a plot of E_rev versus pH_i-pH_o showing that the data fall on the predicted line representing the following equation: E = 58 × (pH_i-pH_o) demonstrating that the whole cell currents represent H⁺(OH⁻) currents through the transporter. In separate experiments, whole cell currents measured at the same H⁺ gradient and opposite direction from pH 7.4 are compared (Fig. 3); pH 7.15 versus 8.0 (H⁺ gradient ± 30 nM); pH 7.19 versus pH 7.8 (H⁺ gradient ± 24.2 nM); pH 7.26 versus pH 7.6 (H⁺ gradient ± 14.9 nM); pH 7.30 versus pH 7.52 (H⁺ gradient ± 10.1 nM). The currents measured following background subtraction at +40 mV at each pH value and conductance at 0 mV were as follows: pH 7.4: 5.36 ± 1.71 pA; 0.304 ± 0.049 nS; pH 7.15: 6.00 ± 1.14 pA; 0.330 ± 0.059 nS (n = 3) and pH 8.0: 94.1 ± 13.0 pA (n = 3, P < 0.01 versus pH 7.15); 1.43 ± 0.156 nS (n = 3, P < 0.05 versus pH 7.15); pH 7.19: 1.38 ± 0.275 pA; 0.382 ± 0.080 nS (n = 4) and pH 7.8: 64.3 ± 9.63 pA (n = 4, P < 0.005 versus pH 7.19); 1.24 ± 0.176 nS (n = 4, P < 0.05 versus pH 7.19); pH 7.26: 4.15 ± 1.96 pA; 0.337 ± 0.062 nS (n = 3) and pH 7.6: 34.1 ± 3.56 pA (n = 5, P < 0.05 versus pH 7.26); 0.779 ± 0.074 nS (n = 5, P < 0.05 versus pH 7.26); pH 7.30: 5.25 ± 1.21 pA; 0.365 ± 0.071 nS (n = 4) and pH 7.52: 22.6 ± 3.45 pA (n = 4, P < 0.05 versus pH 7.30); 0.662 ± 0.096 nS (n = 4, P < 0.05 versus pH 7.30). These findings show that raising the pH_e per se stimulates SLC4A11-mediated H⁺(OH⁻) transport.

Fig. 2.

Whole cell currents were measured at various extracellular pH (pH_e) values with an intracellular pH (pH_i) of 7.4 and current-voltage data are depicted. The bath pH was varied from 7.0 to 8.0. A: mock-transfected cells. B: SLC4A11-transfected cells. C: subtraction of the mock data from the SLC4A11 data. D: reversal potential (E_rev) measured at various pH_e values with an intracellular pH (pH_i) of 7.4.

Fig. 3.

Whole cell currents were measured at various extracellular pH (pH_e) values that differed from pH 7.4 with the same H⁺ gradient and opposite direction are compared and current-voltage data are depicted. A, D, G, and J: mock-transfected cells. B, E, H, and K: SLC4A11-transfected cells. C, F, I, and L: subtraction of the mock data from the SLC4A11 data.

Equivalent H⁺(OH⁻) Flux Following an Increase in pH_e From 7.4 to pH 8.0 at Constant NH₃ Concentration With a Decrease in the ${NH}_{\mathit{4}}^{+}$ Concentration at pH 8.0

Following an increase in pH_e from 7.4 to 8.0 and return to pH 7.4 in the presence of ammonia pH_i transients were monitored (Fig. 4). In these experiments the extracellular NH₃ concentration was 0.153 mM in both the pH 7.4 and 8.0 solutions and the ${\text{NH}}_4^{+}$ concentration was decreased from 4.85 mM at pH 7.4 to 1.22 mM at pH 8.0 creating a transcellular ${\text{NH}}_4^{+}$ and H⁺ gradient. As shown in representative experiments in Fig. 4, in mock-transfected cells equivalent H⁺(OH⁻) flux was greater in the presence of ammonia following generation of a transcellular H⁺ gradient: pH 7.4 to 8.0: 0.019 ± 0.003 mM/s (n = 7) versus 0.011 ± 0.002 mM/s (n = 8) in controls (P < 0.05) and pH 8.0 to 7.4: −0.025 ± 0.003 (n = 7) versus −0.009 ± 0.002 mM/s (n = 8) (P < 0.001) in controls. In SLC4A11-expressing cells the equivalent H⁺(OH⁻) flux was significantly greater in the presence of ammonia following generation of a transcellular H⁺ gradient; pH 7.4 to 8.0: 0.144 ± 0.009 mM/s (n = 8) versus 0.053 ± 0.004 mM/s (n = 19) in controls (P < 0.001) and pH 8.0 to 7.4: −0.112 ± 0.005 mM/s (n = 8) versus −0.062 ± 0.005 mM/s (n = 19) in controls (P < 0.001). The data are compatible with NH₃ coupled H⁺ transport (or ${\text{NH}}_4^{+}$ transport) contributing with H⁺(OH⁻) flux to the total equivalent H⁺(OH⁻) flux in SLC4A11-expressing cells in the presence of ammonia following an increase and decrease in pH_e.

Fig. 4.

Changes in intracellular pH (pH_i) induced by extracellular pH (pH_e) changes (7.4–8.0–7.4) in the presence of 5 mM NH₄Cl at a constant NH₃ concentration of 0.153 mM with a decrease in the ${\text{NH}}_4^{+}$ concentration from 4.85 to 1.22 mM at pH 8.0. A: mock-transfected cells. B: SLC4A11-transfected cells. C: summary of the pH 7.4–8.0 flux data in mock and SLC4A11-transfected cells. D: summary of the pH 8.0–7.4 flux data in mock and SLC4A11-transfected cells. *P < 0.05; #P < 0.001, presence versus absence of 5 mM NH₄Cl in the same cell type (Student’s t test).

Whole Cell Currents in the Presence of a Constant ${NH}_{\mathit{4}}^{+}$ Concentration and Varying Extracellular NH₃ Concentrations as a Function of pH_e

In the presence of 2 mM NH₄Cl, whole cell currents were measured at pH_e values between 7.0 and 8.0 with the patch pipette at 7.4 in mock-transfected and SLC4A11-expressing cells. The ${\text{NH}}_4^{+}$ concentration was kept constant, and there was no patch pipette to bath ${\text{NH}}_4^{+}$ gradient. The NH₃ patch pipette to bath gradient varied with the pH and was equal and opposite to the H⁺ gradient at each pH value. The ammonia-induced whole cell currents are summarized in Fig. 5, A–C. In the absence and presence of ammonia, the SLC4A11-mediated whole cell currents were subtracted from the mock induced currents. The results at +40 mV and conductances measured at 0 mV in the absence of ammonia are shown above (Whole Cell H⁺ Currents at Various pH_e Values and Fig. 2C). The results in the presence of ammonia were: pH 7.0: 8.45 ± 1.13 pA (n = 5), P < 0.05 vs. control, and 0.529 ± 0.030 nS (n = 5), P < 0.001 vs. control; pH 7.4: 18.7 ± 2.83 pA (n = 4), P < 0.001 vs. control, and 0.452 ± 0.020 nS (n = 4), P < 0.05 vs. control; pH 7.6: 47.6 ± 4.76 pA (n = 4), P < 0.05 vs. control, and 1.12 ± 0.104 nS (n = 4), P < 0.05 vs. control; pH 7.8: 89.7 ± 8.30 pA (n = 5), P < 0.01 vs. control, and 1.72 ± 0.180 nS (n = 5), P < 0.01 vs. control; and pH 8.0: 129.5 ± 7.77 pA (n = 4), P < 0.05 vs. control, and 2.08 ± 0.126 nS (n = 4), P < 0.01 vs. control. Figure 5, D and E, shows the comparison with the data obtained in the absence of ammonia. The results demonstrate that the ammonia-induced whole cell currents are stimulated by increasing the pH_e.

Fig. 5.

Whole cell currents were measured at various extracellular pH (pH_e) values with an intracellular pH (pH_i) of 7.4 in the presence of 2 mM NH₄Cl, and current-voltage data are depicted. The ${\text{NH}}_4^{+}$ concentration was kept constant and there was no patch pipette to bath ${\text{NH}}_4^{+}$ gradient. The NH₃ patch pipette to bath gradient varied with the pH and was equal and opposite to the H⁺ gradient at each pH value. A: mock-transfected cells. B: SLC4A11-transfected cells. C: subtraction of the mock data from the SLC4A11 data. D: summary of the subtraction (SLC4A11 minus mock) data of the effect of ammonia on whole cell currents (+40 mV) measured at various pH_e values. E: summary of the subtraction (SLC4A11 minus mock) data of the effect of ammonia on conductance (0 mV) measured at various pH_e values. *P < 0.05; ⁺P < 0.01; #P < 0.001, presence versus absence of 2 mM NH₄Cl at each specific pH (Student t test).

Ammonia-Induced Currents Following an Increase in pH_e From 7.4 to pH 8.0 With an ${NH}_{\mathit{4}}^{+}$ Concentration Gradient at pH 8.0 and No NH₃ Concentration Gradient

Whole cell currents were measured in mock-transfected and SLC4A11-expressing cells following an increase in bath pH from 7.4 to 8.0 in solutions containing various NH₃ and ${\text{NH}}_4^{+}$ concentrations without a patch pipette to bath NH₃ concentration gradient in the presence of patch pipette to bath 4:1 ${\text{NH}}_4^{+}$ gradient at pH 8.0. The E_rev value was measured in each solution. The mock-transfected data were subtracted from the SLC4A11 data. Figure 6A shows a plot of E_rev versus the extracellular NH₃ concentration. The theoretical E_rev based on the H⁺ gradient of 4:1 (intracellular to extracellular) is −34.8, which approximates the E_rev measured in the absence of ammonia (−30.4 mV). As shown in Fig. 6A, the E_rev measured at each NH₃ concentration remained unchanged in the ammonia containing solutions. The latter is what is theoretically expected for an NH₃-nH⁺ (n ≥ 1) or ${\text{NH}}_4^{+}$-nH⁺ (n ≥ 0) transporter (theoretical E_rev also −34.8 in the ammonia-containing solutions).

Fig. 6.

A: Whole cell currents were measured in mock-transfected and SLC4A11-expressing cells following an increase in bath pH from 7.4 to 8.0 in solutions containing various NH₃ and ${\text{NH}}_4^{+}$ concentrations in the presence of patch pipette to bath 4:1 ${\text{NH}}_4^{+}$ gradient (pH 8.0) without a patch pipette to bath NH₃ concentration gradient. Depicted is a plot of reversal potential (E_rev) versus the extracellular NH₃ concentration. B: following an increase in bath from 7.4 to 8.0 whole cell currents were measured in mock-transfected and SLC4A11-expressing cells in solutions containing various NH₃ and ${\text{NH}}_4^{+}$ concentrations in the presence of patch pipette to bath 1:4 NH₃ gradient (pH 8.0) that was equal and opposite to the H⁺ gradient without a patch pipette to bath ${\text{NH}}_4^{+}$ concentration gradient. Depicted is a plot of E_rev versus the extracellular NH₃ concentration.

Ammonia-Induced Currents Following an Increase in pH_e from 7.4 to pH 8.0 with an NH₃ Concentration Gradient at pH 8.0 and no ${NH}_{\mathit{4}}^{+}$ Concentration Gradient

Following an increase in bath from 7.4 to 8.0, whole cell currents were measured in mock-transfected and SLC4A11-expressing cells in solutions containing various NH₃ and ${\text{NH}}_4^{+}$ concentrations without a patch pipette to bath ${\text{NH}}_4^{+}$ concentration gradient in the presence of patch pipette to bath 1:4 NH₃ gradient that was equal and opposite to the H⁺ gradient. In each solution protocol, the E_rev was measured. The mock-transfected data were subtracted from the SLC4A11 data. A plot of E_rev versus the extracellular NH₃ concentration is shown in Fig. 6B. With an H⁺ gradient of 4:1 (intracellular to extracellular), the theoretical E_rev is −34.8, which approximates the E_rev measured in the absence of ammonia (−30.4 mV). In the presence of ammonia, the E_rev expected from NH₃-H⁺ transport is zero given equal and opposite H⁺ and NH₃ gradients. As shown in Fig. 6B as the NH₃ concentration increased, the measured E_rev shifted closer to zero reaching −9.6 mV at a patch pipette and bath ${\text{NH}}_4^{+}$ concentration of 9.89 mM and a patch pipette and bath NH₃ concentration of 0.114 and 0.454 mM, respectively. The data are compatible with NH₃-H⁺ transport (or ${\text{NH}}_4^{+}$ transport) competing with H⁺(OH⁻) for transport on SLC4A11 with increasing NH₃ concentration. The data are not compatible with NH₃-2H⁺-mediated SLC4A11 transport since E_rev at the higher ammonia concentrations should have approached −17.6 mV (the E_rev expected from an NH₃-2H⁺ transport mode). Other transport modes such as NH₃-3H⁺ (expected E_rev −23.4) or ${\text{NH}}_4^{+}$-H⁺ (expected E_rev −17.6 mV) are also excluded.

Monitoring pH_i in the Patch-Clamp Experiments Using EYFP

pH_i was monitored in patch-clamped cells following a change in pH_e from 7.4 to 8.0 (pH_i 7.4) and following a change to pH 8.0 in the presence ammonia with 4:1 bath to patch pipette NH₃ gradient (patch pipette and bath NH₃ concentration of 0.114 mM and 0.454 mM respectively). Figure 7 depicts representative experiments where the pH_i remained unchanged following an increase in pH_e to 8.0 in the absence (n = 7) and presence (n = 6) of 4:1 NH₃ gradient. These findings indicate that changes in pH_i were not a factor affecting the data.

Fig. 7.

Intracellular pH (pH_i) was monitored using EYFP in patch-clamped cells following a change in pH_e from 7.4 to 8.0 (pH_i 7.4) and following a change to pH 8.0 in the presence of ammonia with a patch pipette and bath NH₃ concentration of 0.114 and 0.454, mM, respectively (4:1 bath to patch pipette NH₃ gradient). Representative experiments are depicted following an increase in pH_e to 8.0 and in the absence and presence of 4:1 NH₃ gradient.

Effect of SLC4A11 Mutations on H⁺(OH⁻) Currents and Ammonia-Induced Currents

On the basis of homology with NBCe1-A and AE1, the predicted SLC4A11 ion coordination site includes residues adjacent to or within TM3, TM8, and TM10. We mutated residues in these locations as shown in Fig. 8 and measured H⁺(OH⁻) flux following an increase in bath pH from 7.4 to 8.0 and from pH 8.0 to 7.4. These results are shown in Fig. 8. We identified three residues E659, Y706, and H708 within or adjacent to TMs 8 and 10 that when mutated significantly decreased H⁺(OH⁻) flux. H⁺(OH⁻) currents and ammonia-induced currents were also measured in these mutant SLC4A11-expressing cells to determine whether H⁺(OH⁻) and NH₃-H⁺ interact (and compete) with the same SLC4A11 residues (Fig. 9). In SLC4A11-expressing cells whole cell currents were measured following an increase in bath pH from 7.4 to 8.0. In separate experiments whole cell currents were measured following an increase in pH_e from 7.4 to 8.0 in the presence of ammonia with the bath ${\text{NH}}_4^{+}$ concentration 0.494 mM at pH 7.4 and 8.0 while the bath NH₃ concentration increased from 0.006 mM at pH 7.4 to 0.023 mM at pH 8.0 (patch pipette to bath NH₃ gradient equal and opposite to the H⁺ gradient). As shown in Fig. 9, in all mutants there was a decrease in both H⁺(OH⁻) currents and ammonia-induced currents suggesting that these transport processes share common features involving the SLC4A11 transport mechanism. Figure 10 shows the results of sulpho-NHS-SS-biotin experiments showing the plasma membrane expression of these transporters in addition to P707A, which had a modest decrease in H⁺(OH⁻) flux (Fig. 8). The involved residues are depicted in the SLC4A11 predicted ion coordination site (Fig. 11).

Fig. 8.

Based on homology with NBCe1-A and AE1, specific residues adjacent to or within TM3, TM8, and TM10 were predicted to be in the SLC4A11 ion coordination site. Residues in these locations were mutated and extracellular pH (pH_e) was changed from 7.4 to 8.0–7.4 while intracellular pH (pH_i) was monitored to calculate the H⁺(OH⁻) flux. A: flux (pH 7.4–8.0). B: flux (pH 8.0–7.4). E659, Y706, and H708 when mutated significantly decreased H⁺(OH⁻) flux. *P < 0.05; #P < 0.001, comparison with wild-type SLC4A11 (one-way ANOVA and Dunnett’s t test).

Fig. 9.

Residues near the putative ion coordination site were mutated and H⁺(OH⁻) currents and ammonia-induced currents were measured to determine whether H⁺(OH⁻) and NH₃-H⁺ interact (and compete) with the same SLC4A11 residues. Whole cell currents were measured in mutant SLC4A11-expressing cells following an increase in bath pH from 7.4 to 8.0 in the absence of ammonia and in the presence of 0.5 mM NH₄Cl solutions, and current-voltage data are depicted. The bath ${\text{NH}}_4^{+}$ concentration was 0.494 mM at pH 7.4 and 8.0 while the bath NH₃ concentration increased from 0.006 mM at pH 7.4 to 0.023 mM at pH 8.0 (1:4 patch pipette to bath NH₃ gradient equal and opposite to the H⁺ gradient).

Fig. 10.

Sulpho-NHS-SS-biotin experiments showing the plasma membrane expression of the mutant SLC4A11 transporters. A: representative experiment showing immunoblot analysis of cell surface expression of wild-type and mutant SLC4A11 proteins. B: densitometric analysis of the ratio of SLC4A11 proteins to lysate intensity (n = 5).

Fig. 11.

Predicted SLC4A11 ion coordination site: A homology model of the outward facing state of SLC4A11 was built with Modeller 9.18, using the human AE1 crystal structure (4) and human NBCe1 cryo-EM structure (19) as templates. Multiple sequence alignment of SLC4A11 and the two templates (AE1 and NBCe1) was done in advance with ClustalW (53). Five hundred three-dimentional models of SLC4A11 were then generated with the automodel algorithm implemented in Modeller 9.18, and their overall quality was assessed with DOPE and GA341 scores (36, 48). The best 10 models were analyzed for further quality assessment and one final model, reproducing successfully the position and orientation of the functionally critical residues in the templates, was chosen.

A Model for SLC4A11 Ammonia Transport: NH₃-H⁺ and H⁺ Competitively Interact with the Transporter

For a model where the transport of NH₃-H⁺ and H⁺ occur independently and since NH₃-H⁺ and H⁺ are not coupled, the GHK equation (3) is applicable. For the E_rev calculation, if SLC4A11 transports NH₃-H⁺ and H⁺ independently in parallel, we have:

$$E_{\text{rev}}=\frac{RT}{F}\ln \frac{P_{\text{NH}3}\left({\left[{\text{NH}}_3\right]}_{\text{o}}\right)\left({\left[{\text{H}}^{+}\right]}_{\text{o}}\right)+P_{\text{H}}{\left[{\text{H}}^{+}\right]}_{\text{o}}}{P_{\text{NH}3}\left({\left[{\text{NH}}_3\right]}_{\text{i}}\right)\left({\left[{\text{H}}^{+}\right]}_{\text{i}}\right)+P_{\text{H}}{\left[{\text{H}}^{+}\right]}_{\text{i}}}$$ (1)

where P_NH3 is the affinity of NH₃-H⁺ and P_H is affinity of H⁺. If we assume that NH₃-H⁺ and H⁺ interact competitively with the transporter:

$$E_{\text{rev}}=\frac{RT}{F}\ln \frac{\frac{A_{\text{NH3-H}}}{A_{\text{NH3-H}}+A_{\text{H}}}\left({\left[{\text{NH}}_3\right]}_{\text{o}}\right)\left({\left[{\text{H}}^{\text{+}}\right]}_{\text{o}}\right)+\left(1-\frac{A_{\text{NH3-H}}}{A_{\text{NH3-H}}+A_H}\right){\left[{\text{H}}^{+}\right]}_{\text{o}}}{\frac{A_{\text{NH3-H}}}{A_{\text{NH3-H}}+A_{\text{H}}}\left({\left[{\text{NH}}_3\right]}_{\text{i}}\right)\left({\left[{\text{H}}^{+}\right]}_{\text{i}}\right)+\left(1-\frac{A_{\text{NH3-H}}}{A_{\text{NH3-H}}+A_{\text{H}}}\right){\left[{\text{H}}^{+}\right]}_{\text{i}}}$$ (2)

where A_NH3-H is the affinity of NH₃-H⁺ and A_H is the affinity of H⁺. We modeled the data from the protocol in Fig. 6B with Eq. 2 where ammonia-induced currents were measured following an increase in pH_e from 7.4 to pH 8.0 at various NH₄Cl concentrations without an ${\text{NH}}_4^{+}$ concentration gradient at pH 7.4 and 8.0. We fitted the NH₃ versus E_rev data with Eq. 2 (nonlinear least squares method). Figure 12 shows the best fit to the data with E_rev as a function of NH₃ concentration which gives an estimate of A_NH3-H+/(A_{NH3 +} A_H+) = 0.99995. The model described by Eq. 2 is consistent with the NH₃ versus E_rev data.¹ In addition, applying the model to the data in Fig. 5C, the predicted E_rev values at each pH were as follows: pH 7.0, +18.1 mV; pH 7.4, 0 mV; pH 7.6, −7.45 mV; pH 7.8, −13.6 mV; and pH 8.0, −18.4 mV. These values closely approximate the measured E_rev values: pH 7.0, +15.3 mV; pH 7.4, 0.248 mV; pH 7.6, −7.78 mV; pH 7.8, −12.3 mV; and pH 8.0, −20.0 mV (Fig. 13).

Fig. 12.

Model (NH₃-H⁺ versus H⁺) of the data in Fig. 6B depicting the NH₃ versus reversal potential (E_rev) data using a nonlinear least squares method (Eq. 2) where ammonia-induced currents were measured following an increase in extracellular pH (pH_e) from 7.4 to pH 8.0 at various NH₄Cl concentrations without a change in ${\text{NH}}_4^{+}$ concentration at pH 8.0. The data were fitted with Eq. 2 and depicted is the best fit to the data with E_rev as a function of NH₃ concentration. The model described by Eq. 2 is consistent with the NH₃ versus E_rev data.

Fig. 13.

The NH₃-H⁺ versus H⁺ model of the data in Fig. 6B was used to predict the pH-dependent reversal potential (E_rev) values in Fig. 5C. Plotted are the measured E_rev values, which closely approximate the predicted E_rev values.

DISCUSSION

In the present study H⁺(OH⁻) flux was quantitated by monitoring changes in pH_i and measuring whole cell currents in HEK293 cells expressing the transporter. Stimulation of whole cell currents by ammonia was measured under various conditions to determine the mode of transport. In the absence of ammonia, SLC4A11-mediated H⁺(OH⁻) transport that was stimulated by raising the pH_e. An increase in pH_e was also found to stimulate ammonia-induced whole cell currents. These effects are likely mediated by the deprotonization of key residues in the transporter involved in the ion transport mechanism. In experiments using varying NH₄Cl concentrations with equal extracellular and intracellular ${\text{NH}}_4^{+}$ concentrations and opposing H⁺ and NH₃ gradients (NH₃-H⁺ E_rev = 0), the E_rev shifted from −30.4 mV (absence of ammonia) to less negative values (−9.6 mV) compatible with NH₃-H⁺ transport [rather than NH₃-nH⁺ (n ≥ 2) transport], where NH₃-H⁺ competes with H⁺(OH⁻) for transport. In separate experiments, in the presence of ammonia and an acute transcellular H⁺ gradient there was an increase in the measured equivalent H⁺(OH⁻) flux as expected if NH₃-H⁺ flux was contributing to the observed pH changes. A theoretical model was created and used to fit the NH₃ versus E_rev data, which showed that NH₃-H⁺ and H⁺(OH⁻) competitively interact with the transporter with an affinity ratio A_NH3-H+/(A_NH3-H+ + A_H+) of 0.99995. Both H⁺(OH⁻) transport and ammonia-induced whole cell currents were decreased in several mutant SLC4A11 constructs involving the putative ion coordination site suggesting that H⁺(OH⁻) and NH₃-H⁺ transport processes share common features involving the SLC4A11 transport mechanism.

Ogando et al. (38) first reported that when PS120 fibroblasts expressing SLC4A11 were exposed to NH₄Cl, after the initial alkalinization phase due to cellular NH₃ entry, the acidification phase due to ${\text{NH}}_4^{+}$ entry was greater than mock-transfected cells. There was also a greater degree of acidification after NH₄Cl removal consistent with greater ${\text{NH}}_4^{+}$ influx. Similar findings were reported by Zhang et al. (61) who noted that NH₃-H⁺ transport would give similar results. In contrast, Loganathan et al. (33) reported that the alkalinization phase was greater in SLC4A11-expressing HEK293 cells that was attributed to SLC4A11-mediated NH₃ flux per se. Li et al. (30) also reported a greater alkalinization phase.

Zhang et al. (61) first reported NH₄Cl-induced changes in whole cell currents in PS120 fibroblasts expressing SLC4A11. Based on E_rev shifts in response to changes in external NH₃ and pH, it was suggested that SLC4A11 mediates NH₃-2H⁺ cotransport. In studies where the pH was changed to generate varying NH₃ and ${\text{NH}}_4^{+}$ concentrations, the authors concluded that NH₃-2H⁺ rather than ${\text{NH}}_4^{+}$ was the transport mode because NH₃ varied proportionally with the change in whole cell current. In addition, K⁺, which shares transport processes with ${\text{NH}}_4^{+}$, is not transported by SLC4A11 (16, 61). In keeping with these findings our data are compatible with NH₃-H⁺ cotransport. It should be noted that our data do not distinguish NH₃-H⁺ from ${\text{NH}}_4^{+}$ as the transported species or provide information as whether interconversion of these species within SLC4A11 occurs during the transport cycle as in the AmtB channel (25). However, in our model of the data, the affinity ratio A_NH4+/(A_NH4+ + A_H+) is close to zero suggesting that NH₃-H⁺ is being transported.

Loganathan et al. (33) studied SLC4A11-mediated ammonia transport in Xenopus oocytes. Exposure of control oocytes to NH₄Cl led to an intracellular acidification due to the preferential influx of ${\text{NH}}_4^{+}$ rather than NH₃. In SLC4A11-expressing oocytes, pH_i increased suggesting that NH₃ was the species being transported. The whole cell currents induced by ammonia exposure were attributed to an alkalinization-dependent opening of endogenous channels rather than NH₃-2H⁺ cotransport. Experiments measuring pH_i in HEK 293 cells following NH₄Cl addition were also compatible with SLC4A11-mediated NH₃ transport per se. SLC4A11-mediated conductive H⁺(OH⁻) transport in the absence of ammonia was also not detected in this study. Why these results differ from the present study is unclear.

Myers et al. (37) studied mouse Slc4a11 expressed in Xenopus oocytes. The authors concluded that the currents observed in the presence of ammonia do not involve Slc4a11-mediated NH₃ or ${\text{NH}}_4^{+}$ transport or an endogenous channel. Mouse Slc4a11 functioned as an H⁺(OH⁻) transporter that was stimulated by an increase in pH_e or pH_i. Several mechanisms for the increased currents observed were hypothesized including the following: 1) cell depolarization due to ${\text{NH}}_4^{+}$ influx, which would result in increased Slc4a11-mediated H⁺(OH⁻) flux; 2) the increase in pH_i due to enhanced H⁺(OH⁻) flux would have an independent effect on stimulating the transporter; and 3) pH_i may have been increased due to cellular NH₃ influx thereby stimulating the transporter.

Changes in patch pipette pH due to pH_e changes or changes in extracellular NH₃ were not a factor affecting the data. The patch pipette solution contained 100 mM HEPES to minimize any patch pipette solution pH changes. As shown in Fig. 7 using EYFP to monitor changes in pH_i in the patch-clamp experiments, following pH_e changes (experiments in Fig. 2), there were no changes in pH_i detected. In addition, increasing the extracellular NH₃ concentration and pH_e (experiments in Fig. 5) did not change pH_i. In experiments where the pH_e was changed (Fig. 2), the measured E_rev approximated the value predicted from the transcellular H⁺ gradient (bath to patch pipette). In the data in Fig. 6A, the E_rev remained unchanged approximating the theoretical value of −34.8 in all NH₄Cl containing solutions indicating that the H⁺ gradient between the patch pipette and bath remained at ~4:1 in all solutions. Ammonia-induced changes in membrane potential were also not a likely factor affecting the data. Voltage clamping the cells prevented ammonia-induced changes in cell membrane potential from independently driving H⁺(OH⁻) flux. In experiments where the cells were not voltage clamped (data in Fig. 4), a decrease in bath ${\text{NH}}_4^{+}$ and increase in bath pH in this protocol could potentially have increased the cell membrane potential (more negative), but this would have decreased equivalent H⁺(OH⁻) flux in contrast to the increase that was detected.

Our data demonstrate that in the absence of ammonia, SLC4A11 functions as an H⁺(OH⁻) conductive pathway. These findings complement previous studies measuring changes in pH_i and electrophysiological studies on mouse Slc4a11 (37). Previous studies by Zhang et al. (61) and Loganathan et al. (33) in Xenopus oocytes with measurements at −60- and −40-mV holding potential, respectively, concluded that SLC4A11-does not mediate significant conductive H⁺(OH⁻) transport. Our data demonstrate that conductive SLC4A11-mediated conductive H⁺(OH⁻) transport is more easily detected at membrane voltages that are less negative. Our findings are in agreement with Myers et al. (37) who concluded that mouse Slc4a11 mediates conductive H⁺(OH⁻) transport. Because of the activity of mouse Slc4a11 was independently stimulated by an in increase in pH_e and pH_i, it was suggested that OH⁻ is the most likely the transported species. Whether SLC4A11 transports H⁺ versus OH⁻ is difficult to address given that 1) the transport of each species affects pH identically; 2) H⁺ and OH⁻ transcellular concentration gradients are equal and opposite in direction; and 3) they have the same Nerst potentials. Of interest, there is evidence for OH⁻ transport in (C₆F₂)2Hg planar phospholipid membranes and freshwater algae Chara spp. Channels (1, 35).

Based on homology with the ion coordination sites in the SLC4 transporters NBCe1-A (19) and AE1 (4), we examined residues adjacent to or within TM3, TM8, and TM10 that could be involved in SLC4A11 ion coordination. Our results show that mutants E659A (TM8), Y706D (TM10), and H708K (TM10) (SLC4A11-C numbering) impair H⁺(OH⁻) flux and block both conductive H⁺(OH⁻) transport and ammonia-stimulated whole cell currents. E659A and H708D (SLC4A11-C numbering) are known mutations causing CHED (24, 27). We hypothesize that these residues are involved in ion coordination in the transporter and that H⁺(OH⁻) and NH₃-H⁺ transport processes share common features involving the SLC4A11 transport mechanism. The presence of E659 in the putative ion coordination site suggests that H⁺ rather than OH⁻ is the transported species. These findings complement our competitive binding model (Eq. 2) depicting competitive interaction of H⁺(OH⁻) and NH₃-H⁺ with the transporter. NBCe1-A and AE1 likely undergo conformational changes from outward facing to inward facing conformations during the transport cycle. By analogy, given its sequence homology, it might be predicted that SLC4A11 is a transporter that undergoes similar conformational changes. It is also possible however that SLC4A11 functions as an H⁺(OH⁻) and NH₃-H⁺ channel. H⁺ conduction in proteins is thought to occur either by a reversible protonation-deprotonation of specific residues or via a Grotthuss-type (water-hopping) mechanism (10). Interestingly, it has been suggested that a variant of the Grotthuss mechanism involves putative H⁺ hopping in channels that are occupied with ammonia appearing macroscopically as an increase in NH₃-H⁺ (or ${\text{NH}}_4^{+}$) permeability (52).

In addition to mediating ammonia and H⁺(OH⁻) transport, SLC4A11 transports water (54). Aquaporin proteins such AQP8 and AtTIP2;1 also transport ammonia and water via a mechanism involving ${\text{NH}}_4^{+}$ uptake and deprotonation (26). Aquaporins typically efficiently exclude protons from the pore via electrostatic interactions rather than proton wire interruptions (9, 13). Rat AQP1 mutated in the ar/R constriction region develops ammonia stimulated proton transport and a proton permeability (5). Artificial water channels can transport protons at a high rate (14, 31). The transport of water by SLC4A11 is analogous to the transport of water by other membrane transport proteins such as SGLT1, NIS, MCT1, NKCC1, and LeuT (29, 59, 63) that are thought to develop transient water-conducting (channel-like) states during the conformational changes that provide alternating access to substrates. Whether water permeates SLC4A11 through these transient water-conducting states or SLC4A11 functions as a water channel like aquaporin is currently unknown. In addition, whether there are structural overlapping features common to its water, ammonia, and H⁺(OH⁻) transport pathways is unknown. Of the known SLC4A11 mutations, the R125H CHED-causing mutation blocks all three modes of transport suggesting structurally (or structural interaction) shared features between these pathways (23).

In conclusion, our results suggest that SLC4A11 mediates conductive H⁺(OH⁻) transport and in the presence of ammonia the transporter also functions in the NH₃-H⁺ transport mode. An important question to be addressed is the role of SLC4A11-mediated transport in vivo. In corneal endothelial cells, SLC4A11 has been reported to be expressed on the basolateral membrane (54) and more recently in the inner mitochondrial membrane (39). To calculate the electrochemical gradient across the transporter, whereas the stromal pH is estimated to be 7.39–7.54 (depending on whether the eye is open or closed) (6) with a pH_i of ~7.37 (7), the stromal and intracellular ammonia concentrations have not been determined. There is also uncertainty regarding the V_m of endothelial cells (18, 57) making it difficult to predict whether the transporter acts as a cellular H⁺ and ammonia loader versus an H⁺ and ammonia efflux pathway. It is hypothesized that in mitochondria SLC4A11 acts as an H⁺ influx pathway resulting in mitochondrial uncoupling and depolarization (39). Further studies are needed to address the measurements of corneal stromal, endothelial, and mitochondrial ammonia concentrations in vivo to predict the direction of SLC4A11 transport.

GRANTS

I. Kurtz is supported by funds from the NIH (National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-077162), the Allan Smidt Charitable Fund, the Ralph Block Family foundation, and the Factor Family Foundation. Work in laboratory of S. Noskov was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC Grant RGPIN-315019). H. Zhekova was supported by the Alberta Innovates Health Solution Post-Doctoral Fellowship. Molecular dynamics simulations were performed on the Canada Foundation for Innovation-supported GladOS cluster at the University of Calgary and on the West-Grid/Compute Canada clusters under a Research Allocation Award (to S. Noskov).

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

X.M.S. and I.K. conceived and designed research; L.K., R.A., N.A., and D.N. performed experiments; L.K., X.M.S., and I.K. analyzed data; L.K., X.M.S., and I.K. interpreted results of experiments; L.K. and H.Z. prepared figures; H.Z. and I.K. drafted manuscript; L.K., S.N., A.P., and I.K. edited and revised manuscript; L.K., R.A., X.M.S., N.A., D.N., H.Z., S.N., A.P., and I.K. approved final version of manuscript.