Mouse Slc4a11 expressed in Xenopus oocytes is an ideally selective H⁺/OH⁻ conductance pathway that is stimulated by rises in intracellular and extracellular pH

Abstract

The SLC4A11 gene encodes the bicarbonate-transporter-related protein BTR1, which is mutated in syndromes characterized by vision and hearing loss. Signs of these diseases [congenital hereditary endothelial dystrophy (CHED) and Harboyan syndrome] are evident in mouse models of Slc4a11 disruption. However, the intrinsic activity of Slc4a11 remains controversial, complicating assignment of its (patho)physiological role. Most studies concur that Slc4a11 transports H⁺ (or the thermodynamically equivalent species OH⁻) rather than HCO₃⁻, but disparities have arisen as to whether the transport is coupled to another species such as Na⁺ or NH₃/NH₄⁺. Here for the first time, we examine the action of mouse Slc4a11 in Xenopus oocytes. We simultaneously monitor changes in intracellular pH, membrane potential, and conductance as we alter extracellular pH, revealing the electrical and chemical driving forces that underlie the observed ion fluxes. We find that mSlc4a11 is an ideally selective H⁺/OH⁻ conductive pathway, the action of which is uncoupled from the cotransport of any other ion. We also find that the activity of mSlc4a11 is independently enhanced by both extracellular and intracellular alkalinization, suggesting OH⁻ as the most likely substrate and providing a novel explanation for the apparent NH₃-dependence of Slc4a11-mediated currents reported by others. We suggest that the unique properties of Slc4a11 action underlie its value as a pH regulator in corneal endothelial cells.

one of the most tightly regulated parameters in the body is pH. Extracellular pH (pH_e) is predominantly maintained by the CO₂/HCO₃⁻ buffering system and is set at a value close to 7.4 by the balance between Pco₂ (controlled by the lungs) and [HCO₃⁻] (controlled by the kidneys) in blood plasma (5, 18). Intracellular pH (pH_i) is maintained at values slightly lower than pH_e, a value that is determined by the balanced actions of acid loaders [such as Cl⁻/HCO₃⁻ exchangers of the solute carrier 4 (Slc4) family] that counter rises in pH_i and acid extruders [such as Na⁺/H⁺ exchangers (Slc9 family) and Na⁺-HCO₃⁻ cotransporters (also Slc4 family)] that counter falls in pH_i (14, 48). Beyond regulation of pH_i, tightly coupled acid-extruding and acid-loading activities produce a net influx of Na⁺ and Cl⁻ that can contribute to cell volume homeostasis (16, 23, 29).

The extent of the Slc4 family of solute carriers was finalized in 2001 with the cloning of human SLC4A11 (46). Of the transporters encoded by the other nine genes in the family (Slc4a1–Slc4a10; there is no Slc4a6), three are Cl⁻/HCO₃⁻ exchangers (AE1, AE2, and AE3) and five are Na⁺-coupled HCO₃⁻ transporters (NBCe1, NBCe2, NBCn1, NBCn2/NCBE, and NDCBE). The molecular actions of the Slc4a9 and Slc4a11 gene products, however, remain controversial (44, 47). Slc4a11—aka the bicarbonate-transporter-related protein BTR1—is the most divergent of the 10 Slc4 members, sharing less than 20% sequence identity with other members at the amino-acid level (48). Moreover, unlike most other Slc4 proteins, Slc4a11 does not transport HCO₃⁻ (28, 34, 42). Nonetheless, Slc4a11 is a clinically important protein: individuals carrying SLC4A11 mutations exhibit congenital hereditary endothelial dystrophy (CHED) and Harboyan syndrome: both rare forms of corneal dystrophy, the latter also being associated with progressive deafness (12, 53). Signs of these diseases (corneal edema from birth followed by endothelial cell dysmorphia and loss with age as well as decreased endocochlear potential and auditory brainstem response) are recapitulated in strains of Slc4a11-null mice (19, 20, 35, 51), along with a mild urine-concentrating defect (19, 20) that has yet to be observed in human probands (33). Consistent with its pathological roles, Slc4a11 is expressed in corneal endothelial cells (19, 51), fibroblasts of the inner ear (35), and in renal epithelia (8, 19).¹ However, assignment of a precise physiological or pathophysiological role for Slc4a11 at any of these locations has been hindered by a lack of consensus regarding its molecular action and its roster of substrates.

Slc4-like genes that are predicted to encode proteins which are identifiably Slc4a11-like, as opposed to AE-like or NBC-like, can be found in most metazoan genomes, indicating a long-held specialization of role (44). Because mammalian Slc4a11 does not appear capable of supporting HCO₃⁻ transport, and because Slc4-like proteins from Arabidopsis (a plant) and Saccharomyces (a yeast) are boric-acid rather than bicarbonate transporters, the original assignment of human Slc4a11 as a Na⁺-coupled borate transporter seemed fitting (17). However, doubts were subsequently voiced in the literature about the physiological requirement for a borate transporter in mammalian cells and none of the four groups, including ourselves, that have subsequently investigated Slc4a11 action have been able to produce any evidence for borate transport by human Slc4a11 in cell lines or oocytes (6, 28, 30, 34, 42, 51).² Even the original study that favored borate transport stopped short of demonstrating the translocation of borate across the plasma membrane, reporting only that borate enhances Slc4a11-dependent pH_i changes (43). One element of the original paper that has persisted is the concept that, regardless of the presence of borate, Slc4a11 confers a H⁺/OH⁻ (the two are thermodynamically indistinguishable) permeability to its host cell that is under the influence of membrane potential (30, 43, 59). However, each of these subsequent studies has ultimately arrived at different conclusions regarding the molecular action of Slc4a11 and no consensus has yet emerged. Some find evidence that its H⁺/OH⁻ permeation is uncoupled from any other substrate (30). Some find evidence that its action is coupled to (or at least additionally permeable to) Na⁺ (42, 43) and yet others report that Slc4a11 is an electrogenic NH₃-2H⁺ cotransporter (59). The authors of a parallel line of investigation find Slc4a11 to be H₂O and NH₃ permeable (34, 51).

With studies of Slc4a11-null mice in mind, we focused our attention on discerning the molecular action of mouse Slc4a11 (mSlc4a11) in Xenopus oocytes. The observation of enhanced H⁺/OH⁻ permeability in mSlc4a11-expressing cells also forms the starting point of the present study. Here we test the current hypotheses regarding mammalian Slc4a11 action using a combination of two-electrode voltage-clamp and H⁺-selective microelectrodes. We examine the H⁺/OH⁻ fluxes associated with Slc4a11 action while considering the chemical and electrical driving forces acting upon the pathway. This approach allows us to demonstrate for the first time that mouse Slc4a11 is a unique entity in the realm of pH regulation: an uncoupled and ideally selective H⁺/OH⁻ conductive pathway, whose action is acutely stimulated by rises in extracellular and intracellular pH. We use our new insights to reconcile previous observations of Slc4a11-expressing cells and propose a single, unifying model of Slc4a11 action.³

MATERIALS AND METHODS

cDNA preparation.

Full-length mouse Slc4a11 cDNA (IMAGE clone no. 40060299) was purchased from Thermo Fisher Scientific (Grand Island, NY) and sent for automated DNA sequencing (Eurofins MWG Operon, Louisville, KY). The results revealed three silent and two nonsilent base mismatches compared to the reference mouse sequence (GenBank DNA accession no. BC111884). The two missense substitutions, predicted to result in the missense substitutions Ala151Thr and Ala850Ser, were corrected by QuikChange mutagenesis, performed according to the manufacturer's instructions using the complementary primer pairs no. 1/2 (T151A conversion) and no. 3/4 (S850A conversion) in Table 1. The corrected open-reading frame was amplified by PCR and subcloned into the pGH19 expression vector (38) between XmaI and HindIII restriction sites. The forward subcloning primer (no. 5 in Table 1) included a consensus Kozak sequence before the ATG and the reverse primer (no. 6 in Table 1) included the natural termination codon. The resulting clone was named mSlc4a11.pGH19. The equivalent of the human CHED-linked mutation R125H (R99H in the mouse sequence) was introduced by QuikChange mutagenesis using primers no. 7 and no. 8 in Table 1. The resulting clone was named R99H/mSlc4a11.pGH19.

Table 1.

Primers used for the subcloning of mSlc4a11 and R99H/mSlc4a11

Primer	Sequence (5′–3′)
1	GACCTGGACCTGCTCATGGCTAAGCTCTTTACAGATG
2	CATCTGTAAAGAGCTTAGCCATGAGCAGGTCCAGGTC
3	CTACCCCGAATCATTGAGGCCAAGTACTTGGATGTC
4	GACATCCAAGTACTTGGCCTCAATGATTCGGGGTAG
5	CGAAGCCCGGGCCACCATGTCACAGAATGAACAC
6	GCAACAAGCTTTCAGGGCCTATGCTCAGC
7	CTTTGAGGAAGAGGTCCATGCACACCGGGACCTGGATG
8	ACATCCAGGTCCCGGTGTGCATGGACCTCTTCCTCAAAG

cRNA synthesis.

pGH19 clones were linearized using NotI and the linearized cDNA was purified using a MinElute PCR Purification Kit (QIAGEN, Valencia, CA). Capped cRNA was generated using the T7 mMessage mMachine Transcription Kit (Thermo Fisher Scientific). The quantity and quality (A₂₆₀/A₂₈₀ ratio) of the cRNA was assessed using a Nanodrop 2000 Spectrophotometer (Thermo Fisher Scientific).

Xenopus oocytes.

The extraction of ovaries from Xenopus laevis frogs was performed in accordance with the rules and recommendations of the Institutional Animal Care and Use Committee (IACUC) at the University at Buffalo, which reviewed and approved the protocol. Oocytes were extracted from female Xenopus laevis frogs (Xenopus Express, Brooksville, FL) as described elsewhere (38). In brief, frogs were anesthetized by immersion in 0.2% tricaine solution for 15 min (or until unresponsive to toe pinch) and small sections of ovary were surgically extracted for washing in Ca²⁺-free “NRS” buffer (in mM: 82 NaCl, 2 KCl, 20 MgCl₂, 5 HEPES, adjusted to pH 7.45 with NaOH). Frogs were subsequently euthanized by exsanguination. Following the Ca²⁺-free washes, oocytes were liberated from the tissue sections by digestion in NRS buffer plus 2 mg/ml type-IA collagenase (C2674: Sigma-Aldrich, St. Louis, MO) for 15–35 min. Finally, liberated cells were washed in NRS buffer, followed by ND96 solution (aka pH_e = 7.5 solution: Table 2), followed by OR3 medium. OR3 medium contains 14 g/l of Leibovitz's L-15 medium (10-045-CV: Thermo Fisher Scientific), 100 U/ml penicillin, 100 μg/ml streptomycin, 5 mM HEPES, adjusted to pH 7.5 using NaOH. The osmolality of OR3, measured using a Vapro vapor pressure osmometer (Wescor, Logan, UT), was adjusted to 195 ± 5 mOsmol/kg osmolality with H₂O.

Table 2.

Composition (in mM) of solutions used in electrophysiological assays

	Solution
	Na+	NMDG+	K+	Ca2+	Mg2+	Cl−	Gluconate−	Buffer	pH
pHe = 6.5	96		2	1.8	1	103.1		5 (Bis Tris)	6.5
pHe = 7.5	96		2	1.8	1	103.1		5 (HEPES)	7.5
pHe = 8.5	96		2	1.8	1	103.1		5 (Bicine)	8.5
Na-free		96	2	1.8	1	103.1		5 (Bicine)	8.5
Cl-free	96		2	1.8	1		103.1	5 (Bicine)	8.5

Biotinylation.

Biotinylation of oocytes was performed with the Pierce Cell Surface Protein Isolation Kit (Thermo Fisher Scientific, Rockford, IL) using a version of the manufacturer's protocol adjusted for oocytes. Briefly, groups of fifteen oocytes were rinsed in ice-cold PBS (diluted to 200 mOsmol/kg to be isosmotic with oocyte cytosol) and each group was transferred into a separate well of six-well plate that contained the same PBS solution plus 0.24 mg/ml of Sulfo-NHS-SS-Biotin biotinylation reagent. After a 1 h incubation in the dark at 4°C with gentle agitation, the reactions were quenched by adding 10% vol/vol of the supplied quenching solution and the cells were washed in ice-cold TBS (pH 7.5). The cells were lysed in lysis buffer (Roche Diagnostics cOmplete, EDTA-free Protease Inhibitor cocktail tablet + 1% Triton X-100 in TBS pH 7.5) and insoluble debris was pelleted by centrifugation. The supernatant containing soluble material was incubated with neutravidin beads for 1 h using the supplied columns. Unbound protein was washed from the columns with lysis buffer, and the bound (i.e., biotinylated) protein was eluted from the beads using 1× SDS buffer containing 50 mM DTT. The eluted material was the “biotinylated fraction” sample. Samples were resolved by polyacrylamide gel electrophoresis using 3–8% Tris-acetate protein gels (Thermo Fisher Scientific) and transferred to PVDF membranes using an iBlot dry blotting system (Invitrogen). Membranes were probed for Slc4a11 protein using a custom anti-BTR1 antibody (Thermo Fisher Scientific) that was generated in rabbits against the extreme C-terminal epitope “LPRIIEAKYLDVMDAEHRP” in the cytosolic Ct of the protein. Immunoreactivity was visualized using an HRP-conjugated goat anti-rabbit secondary antibody (no. 1706515 Bio-Rad, Hercules, CA) and Pierce ECL 2 Western-blotting substrate (Thermo Fisher Scientific) in conjunction with a Pierce MyECL imager (Thermo Fisher Scientific).

Electrophysiological assays.

Xenopus oocytes were placed in a recording chamber (no. RC-3Z, Warner Instruments, Hamden, CT) and impaled with microelectrodes pulled from Clark borosilicate capillary glass (no. GC200TF-10, Warner Instruments) that were mounted in half-cell holders. Glass was pulled using a P-1000 micropipette puller (Sutter Instrument, Novato, CA) to achieve a tip resistance, when filled with saturated KCl of 0.5 MΩ. Voltage measuring and current-passing electrodes were used filled with saturated KCl (no. SP138-500, Thermo Fisher Scientific). pH-measuring electrodes were filled at the tip with a short (∼0.5 mm) column of H⁺-selective ionophore cocktail B (Sigma Aldrich) and backfilled with a solution that contained (in mM) 40 KH₂PO₄, 15 NaCl (pH adjusted to 7.0 with NaOH). Voltage-measuring and current-passing electrodes were connected to an OC-725C oocyte clamp (Warner Instruments) to measure membrane potential (V_m) and currents (I_m). The H⁺-selective microelectrode was connected to a Duo 773 electrometer, with pH being reported as the difference between that signal and the signal from the voltage-sensing electrode. Voltage-clamp protocols were set and run from the pClamp 10.4 software suite (Molecular Devices, Sunnyvale, CA): oocytes were clamped at the desired value of V (typically the spontaneous membrane potential) prior to initiation of a protocol in which the potential was stepped to +20 mV for 100 ms before returning to its initial value for 100–400 ms before being stepped to 0 mV and so on in −20 mV intervals, with the final step being to − 160 mV. Superfusing solutions were changed using a six-channel perfusion pinch-valve control system (no. VC-6, Warner Instruments). Detailed descriptions of solutions used in this study (e.g., osmolality and ion composition) are shown in Table 2.

Data analysis.

Current-voltage (I–V) relationships (generated using Clampfit 10) and pH_i/V_m traces (acquired using custom software written by Dale Huffman from the laboratory of Walter Boron, Case Western Reserve University, Cleveland, OH) were transferred to Excel for calculation of G_m or d[H⁺]_i/dt. Statistical analysis was performed by MiniTab17 using a general linear model with Tukey's post hoc comparison in which injected material (H₂O vs. cRNA) and extracellular solution were treated as two interacting, fixed factors. The cell number (e.g., 1 through 6 of n = 6) was considered as a random factor to allow for pairing of responses to serial solution changes by individual cells. The analysis output assigns a letter to each data group to denote its statistical difference from other data groups; we show these letters above each data point in our figures. Data points that share a letter are considered to be statistically indistinguishable (confidence interval 95%).

RESULTS

Raising extracellular pH (pH_e) increases the H⁺/OH⁻ permeability of mSlc4a11-expressing oocytes.

Figure 1A shows a representative experiment in which we simultaneously monitored intracellular pH (pH_i) and spontaneous membrane potential (V_m) of a H₂O-injected oocyte during exposure to pH_e = 6.5, pH_e = 7.5, and pH_e = 8.5 solutions. The average resting value of pH_i in pH_e = 7.5 solution was 7.15 ± 0.03 (n = 6) and the average V_m was −46 ± 3 mV (n = 6). During exposure to pH_e = 6.5 solution, pH_i of these cells tended to acidify and V_m tended to depolarize (Fig. 1A). During exposure to pH_e = 8.5 solution, the pH_i of H₂O-injected cells tended to alkalinize⁴ and V_m tended to hyperpolarize. The small changes in [H⁺]_i that we detected while changing pH_e did not achieve statistical significance in our sample (Fig. 1C, n = 6; groups that share the same letter were not significantly different from each other; 95% confidence; result of a general linear model analysis with Tukey's pairwise comparison). On the other hand, the more hyperpolarized V_m in pH_e = 8.5 solution was significantly different from the V_m in either pH_e = 6.5 or pH_e = 7.5 solution (gray data points in Fig. 1D, n = 6). The white data points in Fig. 1D represent the averaged calculated values of E_H (the equilibrium potential for H⁺/OH⁻) given the average measured pH_i of each cell (from data such as those in Fig. 1A) at each value of pH_e. The slope of the line drawn through these points has the expected Nernstian ideal value of −58 mV/decade. The slope of the dotted line that describes the actual relationship between V_m and pH_e in H₂O-injected cells, as pH_e increases from 6.5 to 7.5, is −5 mV/decade. The slope of the dotted line that describes the relationship between V_m and pH_e as pH_e increases from 7.5 to 8.5 is −8 mV/decade. Thus we observe no evidence of substantial H⁺/OH⁻ permeability in H₂O-injected oocytes.

Fig. 1.

Extracellular pH (pH_e)-dependence of membrane potential (V_m) and intracellular pH (pH_i) in H₂O-injected and mSlc4a11-expressing oocytes. A and B: pH_i and V_m of an oocyte that had been injected with H₂O (A) or mSlc4a11 cRNA (B) was monitored while the cell was exposed to extracellular solutions of pH 6.5, 7.5, or 8.5. The gray dashed boxes exemplify the periods over which the rates of [H⁺] change were calculated for C (i.e., omitting rapid changes close to the time of solution change that could be associated with artifacts or other non-Slc4a11-dependent phenomena). The arrow in B shows an enhanced depolarization of cells in pH_e = 6.5 solution following exposure to pH_e = 8.5 solution. C: the rate of change of intracellular [H⁺] during exposure to pH_e = 6.5, 7.5, or 8.5 solutions (n = 6). Rates were calculated from the change in [H⁺] that occurred during the periods such as enclosed by gray boxes in A and B. The letters above each bar denote statistically indistinguishable groups reported by general linear model with Tukey's post hoc analysis (confidence interval 95%). D and E: average V_m of H₂O-injected (D) or mSlc4a11-expressing (E) cells during exposure to pH_e = 6.5, 7.5, or 8.5 solutions (gray data points), with letters to denote statistically indistinguishable groups. V_m values were taken at the end of the exposure period to each solution. For each value of pH_e, each “n” represents an averaged value from the 2–3 exposure periods to which each cell was subjected. The white data points indicate the equilibrium potential for H⁺ determined at the average pH_i from A or B at each value of pH_e. Note that standard error bars are included but are eclipsed by the data point markers.

Figure 1B shows an equivalent experiment to that in Fig. 1A, performed on an mSlc4a11-expressing oocyte. The average pH_i of these cells was 6.95 ± 0.01 (n = 6) and the average resting V_m was −34 ± 3 mV (n = 6). Thus these cells are more acidic and more depolarized at rest than H₂O-injected oocytes (P < 0.05; n = 6; unpaired t-test for each parameter). During exposure to pH_e = 6.5 solution, the rate of acidification of mSlc4a11-expressing cells tended to be greater than, but was not significantly different from, that of H₂O-injected cells (Fig. 1C) even when assessed with the less conservative t-test (P = 0.10). We find that the relationship between V_m and pH_e in mSlc4a11-expressing cells vs. H₂O-injected cells is described by a similar slope (i.e., −5 mV/decade), as pH_e increases from 6.5 to 7.5 (gray data points in Fig. 1E). On the other hand, mSlc4a11-expressing cells behave very differently from H₂O-injected cells during exposure to pH_e = 8.5 solution. Firstly, mSlc4a11-expressing cells exhibit a significant alkalinization (Fig. 1, B and C). Secondly, the slope of the dotted line that describes the relationship between V_m and pH_e (as pH_e increases from 7.5 to 8.5) is −38 mV/decade (Fig. 1D), indicating substantial H⁺/OH⁻ permeability of the mSlc4a11-expressing plasma membrane. A final notable feature of the behavior of mSlc4a11-expressing cells is that the extent of depolarization upon exposure to pH_e = 6.5 solution is transiently exaggerated (in the direction of E_H) by ∼3 mV if the cell is bathed in pH_e = 8.5 (see arrow in lower panel of Fig. 1B) rather than pH_e = 7.5 solution immediately prior to the solution change. This indicates that the loss of H⁺/OH⁻ permeability upon extracellular acidification is not instantaneous.

Raising pH_e increases the membrane conductance of mSlc4a11-expressing oocytes.

To examine the membrane currents (I_m) responsible for the pH-dependent shifts in V_m observed in Fig. 1, we subjected oocytes to our voltage-clamping protocol. Figure 2A shows a set of current-voltage (I–V) relationships gathered from a H₂O-injected oocyte during exposure to pH_e = 7.5 and pH_e = 8.5 solutions. The conductance of the membrane (G_m) was unaltered by the rise in pH_e, as summarized for a larger number of cells in Fig. 2D (n = 6; groups that share the same letter are not significantly different from each other; 95% confidence; result of a general linear model analysis with Tukey's pairwise comparison). Figure 2B shows an equivalent set of I–V plots gathered from an mSlc4a11-expressing oocyte. In pH_e = 7.5 solution, the G_m of mSlc4a11-expressing cells is not significantly different from that of H₂O-injected cells (white bars in Fig. 2D). However, exposure of mSlc4a11-expressing oocytes to pH_e = 8.5 solution resulted in the expected hyperpolarization of V_m (evident in the negative shift of the zero-current value of V in Fig. 2B) as well as a gradual increase in the magnitude of G_m throughout the 20 min exposure period (gray vs. black data in Fig. 2, B and D). We will return to this phenomenon of rising G_m vs. time in the section entitled “Raising intracellular pH increases the H⁺/OH⁻ conductance of mSlc4a11-expressing oocytes.” An example of the current steps elicited by our voltage-clamp protocol after 20 min of exposure to pH_e = 8.5 solution is shown in the gray inset in Fig. 2B, and reveals no acute time- or voltage-dependence of I_m. A mutant mSlc4a11 [R99H/mSlc4a11: equivalent to the CHED-causing mutation R125H in human Slc4a11 (24, 52)] accumulated to a similar extent as wild-type mSlc4a11 in the plasma membrane of oocytes (P = 0.23, paired two-tailed t-test, n = 4: example of biotinylation results shown in Fig. 2D/gray inset). However, oocytes expressing the mutant did not exhibit any rise in G_m (Fig. 2, C and D) nor did their V_m hyperpolarize to any greater degree than of H₂O-injected cells in response to the application of pH_e = 8.5 solution (P = 0.27 vs. the ΔV_m of H₂O-injected cells, unpaired two-tailed t-test, data not shown).

Fig. 2.

pH_e-dependence of membrane conductance (G_m) in H₂O-injected and mSlc4a11-expressing oocytes. A–C: representative sets of current-voltage (I–V) relationships for an oocyte that had been injected with H₂O (A), mSlc4a11 cRNA (B), or cRNA encoding an “R99H” mutant of mSlc4a11 (C). A representative current ladder obtained for the Slc4a11-expressing oocyte bathed for 20 min in pH_e = 8.5 solution is shown in the inset in B. D: average membrane conductance calculated for six of each cell type from data such as that in A, B, and C. Letters above each bar denote statistically indistinguishable groups. The inset in D shows one of three Western blots of biotinylated oocyte protein, probed for mSlc4a11 immunoreactivity.

mSlc4a11 is ideally permeable to H⁺/OH⁻.

The equation below (Eq. 1) restates the Goldman-Hodgkin-Katz equation. Equation 1 predicts that, if mSlc4a11 is perfectly selective for H⁺ or OH⁻ (i.e., P_H = 1; P_K, P_Na, and P_Cl = 0), the slope of the relationship between the transmembrane H⁺ gradient and V_m will approach the Nernstian ideal of 58 mV/pH unit:

$$\begin{array}{l}{\mathit{V}}_{\text{m}}=58\ast \log \left(\frac{{\mathit{P}}_{\text{H}}{\left[{\text{H}}^{+}\right]}_{\text{e}}+{\mathit{P}}_{\text{K}}{\left[{\text{K}}^{+}\right]}_{\text{e}}+{\mathit{P}}_{\text{Na}}{\left[{\text{Na}}^{+}\right]}_{\text{e}}+{\mathit{P}}_{\text{Cl}}{\left[{\text{Cl}}^{-}\right]}_{\text{i}}}{{\mathit{P}}_{\text{H}}{\left[{\text{H}}^{+}\right]}_{\text{i}}+{\mathit{P}}_{\text{K}}{\left[{\text{K}}^{+}\right]}_{\text{i}}+{\mathit{P}}_{\text{Na}}{\left[{\text{Na}}^{+}\right]}_{\text{i}}+{\mathit{P}}_{\text{Cl}}{\left[{\text{Cl}}^{-}\right]}_{\text{e}}}\right)\\ {\mathit{V}}_{\text{m}}=58\ast \left({\text{pH}}_{\text{i}}-{\text{pH}}_{\text{e}}\right)\end{array}$$ (1)

Acidic shifts in pH_e prevent mSlc4a11 action from dominating V_m (Fig. 1E); therefore, instead of monitoring the response of pH_i and V_m to changes in pH_e, we opted to monitor the response of V_m at constant pH_e as we altered pH_i. Figure 3A shows a representative experiment in which we bathed an mSlc4a11-expressing oocyte in pH_e = 8.5 solution, allowed V_m to hyperpolarize (per Fig. 1B) and, when we observed no further change in V_m, applied a voltage-clamp to maintain V_m at its spontaneous (“zero-current” or “steady-state”) value. This steady state is shown in Period I of Fig. 3A and provides us with the first of two data points that we gathered from each cell that describes the relationship between the transmembrane pH gradient (pH_i - pH_e) and V_m, that are plotted for each cell in Fig. 3C (open circles). To obtain the second data point (filled circles in Fig. 3C), we caused pH_i to rise by taking advantage of the H⁺/OH⁻ permeability of mSlc4a11-expressing cells; clamping V to +20 mV for a short period of time (Period II in Fig. 3A) provides the necessary acid-extruding driving force. Once pH_i had risen by a substantial amount (on average 0.29 ± 0.04 units, n = 6, which approximates to a halving of [H⁺]_i), we interrupted the pH_i trajectory by clamping V to a new value that represented the zero-current value at the new value of pH_i (Period III in Fig. 3A). As seen in the traces in Fig. 3A, we made several fine adjustments to the value of V until we located its zero-current value. We held the cell in clamp at this value of V until we observed no further change in pH_i, and then noted our second “pH_i − pH_e” vs. V_m data point for the cell (filled circles in Fig. 3C). Figure 3B shows one of five equivalent experiments that we performed on H₂O-injected oocytes. Note that the pH_i in this cell does not change substantially in response to the voltage-clamp maneuvers. Figure 3C shows the accumulated data points gathered from experiments such as those in Fig. 3A, as well as the ideal Nernstian relationship between V_m and the transmembrane pH gradient (gray line). The average slope of the experimentally determined “pH_i − pH_e^” vs. V_m relationships is 64 ± 4 mV/pH unit (n = 6).

Fig. 3.

The relationship between pH_i and V_m in H₂O-injected and mSlc4a11-expressing oocytes. A and B: representative voltage, current, and pH_i traces recorded from an mSlc4a11-expressing (A) and a H₂O-injected oocyte (B) exposed to pH_e = 8.5 solution. During Period I, V_m was clamped at its spontaneous (zero-current) value. During Period II, V_m was clamped to +20 mV to drive electrogenic acid extrusion. During Period III, V_m was clamped at its new zero-current value, reflecting the increased pH_i. C: a plot of zero-current V_m vs. transmembrane pH gradient from six cells such as that represented in A. White data points were collected from each cell during Period I in A. Black data points were collected towards the end of Period III in A.

Raising intracellular pH increases the H⁺/OH⁻ conductance of mSlc4a11-expressing oocytes.

Because pH_i rises in mSlc4a11-expressing oocytes during exposure to pH_e = 8.5 solution, we hypothesized that the apparent time-dependent increase of G_m, represented in Fig. 2, could represent a dependence on pH_i. Figure 4A shows one of six experiments in which we monitored the pH_i of an mSlc4a11-expressing oocyte bathed in pH_e = 8.5 solution. At regular intervals throughout the recording (typically each minute, see deflections in V trace in Fig. 4A), we initiated a 5-s step-wise voltage-clamp protocol to monitor changes in G_m (bottom trace in Fig. 4A). During Period I of the experiment we allowed each cell to spontaneously hyperpolarize and alkalinize for 5 min in response to the application of pH_e = 8.5 solution. During Period II, we applied an electrical driving force to promote faster mSlc4a11-mediated acid-extrusion by voltage-clamping the membrane potential at −60 mV (E_H = −84 mV in the example shown). G_m increases nearly 100-fold over the 0.45 pH unit rise. The evolution of the current steps evoked by our voltage-clamp protocol over the course of the experiment is represented in the inset in Fig. 4A. The average relationship between pH_i and G_m for all six cells tested is plotted in Fig. 4B.

Fig. 4.

The relationship between pH_i and G_m in mSlc4a11-expressing oocytes. A: representative pH_i and voltage traces recorded from an mSlc4a11-expressing cell during exposure to pH_e = 8.5 solution prior to, and during, voltage-clamp at −60 mV to drive electrogenic acid-extrusion. The deflections in the voltage trace denote the points at which current-voltage relationships were obtained. The values of G calculated for the cell at each point are plotted in bottom panel; the gray data point was obtained prior to the application of the voltage-clamp while the cell was spontaneously hyperpolarizing towards initial E_H. The inset shows the current ladders at the measurement points labeled * and #. B: the relationship between pH_i (rounded to 2 decimal places) vs. G_m averaged from data gathered from six mSlc4a11-expressing cells. Low confidence observations are marked with crosses, higher confidence observations are marked with open or filled circles accompanied by standard error bars.

Given the apparent pH_i sensitivity of mSlc4a11, we hypothesized that the lack of mSlc4a11 conductivity at physiological pH (pH_e = 7.5 solution in Fig. 2) could be a function of pH_i rather than pH_e per se. In a set of experiments represented by the example shown in Fig. 5A, we compared G_m in pH_e = 8.5 vs. pH_e = 7.5 solutions while pH_i was held at a constant elevated value close to 7.5. We first obtained a value for G_m in an mSlc4a11-expressing oocyte at rest in pH_e = 7.5 solution (“Point 0” in Fig. 5, A and C). Next, we exposed the cell to pH_e = 8.5 solution and applied a voltage-clamp (V_h = 0 mV) to rapidly alkalinize pH_i. Once pH_i was close to 7.5, we voltage-clamped the cell to −58 mV (the predicted value of E_H for the condition pH_e = 8.5/pH_i = 7.5) to hold pH_i constant_. When pH_i was stable, we made three consecutive measurements of G_m, 1 min apart (“Group I” in Fig. 5A). Next, we exposed the oocyte to our pH_e = 7.5 solution, immediately switching the voltage-clamp to 0 mV (the predicted value of E_H for the new condition pH_e = 7.5/pH_i = 7.5) to prevent changes in pH_i and again measured G_m (“Group II” in Fig. 5A). Finally, we returned to our pH_e = 8.5 solution, reclamped the cell to −58 mV, and measured G_m again (“Group III” in Fig. 5A). Figure 5B reports the average pH_i at which each group of G_m values was obtained from six cells and shows that the pH_i was not different in pH_e = 7.5 solution (Group II) vs. either bracketing measurements in pH_e = 8.5 solution (Groups I and III). We note that pH_i was slightly more acidic than predicted if H⁺ was at electrochemical equilibrium under these conditions (see discussion). Figure 5C reports the average values of G_m determined at each value of pH_e at this constant, elevated value of pH_i. These data show two things:

Fig. 5.

pH_i-dependence vs. pH_e-dependence of membrane conductance (G_m) in mSlc4a11-expressing oocytes. A: representative pH_i and voltage traces recorded from an mSlc4a11-expressing oocyte during exposure to pH_e = 7.5 and pH_e = 8.5 solution under voltage-clamp. When the extracellular solution was pH 8.5, V_m was initially held at 0 mV to promote a mSlc4a11-mediated rise in pH_i to ∼7.5, at which point V_m was clamped to the predicted value of E_H (0 mV) to stabilize pH_i at ∼7.5. When the extracellular solution was pH 7.5, V_m was held at −58 mV to maintain pH_i at ∼7.5. Point 0 and Groups I–III denote regions of the trace where I–V relationships were measured to allow calculation of G_m measurements. There was no significant change in pH_i or G_m between the first and last measurement in Groups I, II, or III and thus we took the average of the three readings for each group to be a single replicate to be reported in B and C. B: average values (n = 7 for all) of pH_i measured at each of the points (or group, in which case the three values in each group were averaged) noted in A, with letters to indicate groups that are statistically indistinguishable (general linear model, confidence interval 95%). C: average values (n = 7 for all) of G_m measured at each of the points noted in A, with letters to indicate groups that are statistically indistinguishable.

1) G_m is significantly greater when pH_i ∼7.5 vs. ∼7.0 (Group II vs. Point 0); i.e., the conductivity of mSlc4a11 at pH_e = 7.5 increases with pH_i.

2) G_m is substantially greater when pH_e = 8.5 vs. 7.5 even at elevated values of pH_i (Groups I and III vs. Group II); i.e., the enhancement of G_m in Fig. 2 is not solely a reflection of increased pH_i; mSlc4a11 is truly and additionally pH_e sensitive.

Removing extracellular Na⁺ or Cl⁻ does not greatly influence the H⁺/OH⁻ conductance of mSlc4a11-expressing oocytes.

Results in Fig. 3 show that mSlc4a11 is ideally permeable to H⁺/OH⁻ and therefore not substantially permeable to Na⁺. However, these results do not rule out the possibility that mSlc4a11 H⁺/OH⁻ permeability is Na⁺ dependent. Figure 6A shows a representative experiment in which we exposed an mSlc4a11-expressing oocyte to pH_e = 8.5 solution, clamped V_m at −10 mV to promote acid-extrusion, and monitored G_m. Once G_m had reached its maximum value, as judged from 3 consecutive readings 1 min apart that showed no substantial change, we applied a bath solution identical to our pH_e = 8.5 solution except for the replacement of all Na⁺ with NMDG⁺. We monitored G_m for a further 3 min and, noting no obvious alteration in G_m, terminated the experiment. The G_m data points with (*) or without (#) Na⁺ in Fig. 6A were generated from the I–V plots shown in Fig. 6B, which in turn were generated from the current ladders shown in Fig. 6C. The averaged values of G_m with or without Na⁺ are shown for a larger number of cells in Fig. 6D. Although G_m was significantly greater in mSlc4a11-expressing cells in the absence of Na⁺ (result of a paired two-tailed t-test, n = 6), the change is insubstantial and the currents are qualitatively similar in the presence vs. the absence of Na⁺. Figure 7, A–D, shows a similar suite of experiments designed to probe the influence of extracellular Cl⁻ on mSlc4a11 action. Here we find that the complete replacement of Cl⁻ with gluconate⁻ causes a significant (result of a paired two-tailed t-test, n = 6) yet insubstantial fall in G_m with, again, no obvious qualitative change in the currents.

Fig. 6.

The influence of extracellular Na⁺ on G_m in mSlc4a11-expressing oocytes. A: representative experiment showing progressive changes in G_m in an mSlc4a11-expressing oocyte clamped at −10 mV in pH_e = 8.5 solution. Gray data points indicate measurements made in Na⁺-free solution. B and C: data points in A marked with * and # are generated from data similarly marked in B accompanying representative I–V plots and C accompanying current ladders. D: averaged data from six cells showing the influence of extracellular Na⁺ on G_m. Statistic is the result of a two-tailed, paired t-test. Data equivalent to that shown in A–D, but showing the influence of extracellular Na⁺.

Fig. 7.

The influence of extracellular Cl⁻ on G_m in mSlc4a11-expressing oocytes. A–D: data equivalent to that shown in Fig. 6, but showing the influence of extracellular Cl⁻.

NH₃/NH₄⁺ indirectly increases the H⁺/OH⁻ conductance of mSlc4a11-expressing oocytes by depolarizing V_m.

Others have presented data in support of the hypothesis that Slc4a11 is an NH₃-2H⁺ cotransporter (59). To reproduce these data, we obtained I–V relationships from H₂O-injected cells (example in Fig. 8A) and cells expressing mSlc4a11 (example in Fig. 8B) during exposure to our pH_e = 7.5 and pH_e = 8.5 solutions in the presence of 5 mM NH₄Cl. The initial response to NH₃/NH₄⁺ application at pH_e = 7.5 of both H₂O-injected and mSlc4a11-expressing cells was a rapid depolarization of V_m. The V_m of H₂O-injected oocytes depolarized from −51 ± 2 mV to −16 ± 2 mv (n = 6) and mSlc4a11-expressing oocytes depolarized from −38 ± 1 mV to −3 ± 1 mV (n = 6). Only in mSlc4a11-expressing cells was this depolarization accompanied by a rapid increase in G_m that became statistically significant when cells were subsequently exposed to NH₃/NH₄⁺ at pH_e = 8.5 (Fig. 8C). Note that the magnitude of the NH₃/NH₄⁺-induced G_m in pH_e = 8.5 solution is similar to that which can be achieved by depolarization induced alkalinization in the absence of NH₃/NH₄⁺ (e.g., Fig. 6D).

Fig. 8.

NH₃/NH₄⁺-dependence of G_m in mSlc4a11-expressing oocytes. A and B: representative sets of I–V relationships for an oocyte that had been injected with H₂O (A) or mSlc4a11 cRNA (B) and sequentially exposed to pH_e = 7.5 solution, pH_e = 7.5 solution containing 5 mM NH₄Cl, and pH_e = 8.5 solution containing 5 mM NH₄Cl. C: average membrane conductance calculated for six of each cell type from data such as that in A and B, with letters above each bar to denote statistically indistinguishable groups (general linear model, confidence interval 95%).

We hypothesized that, if this conductance does represent the action of an electrogenic, NH₃/NH₄⁺-coupled transport process (rather than an indirect stimulation of uncoupled H⁺/OH⁻ conductance that follows depolarization) the NH₃/NH₄⁺-induced rise in G_m ought to be reversed by removal of NH₃/NH₄⁺. Figure 9A shows a representative experiment in which we monitored the G_m of an mSlc4a11-expressing oocyte in pH_e = 8.5 solution during and after exposure to NH₃/NH₄⁺. The first I–V relationship was obtained at spontaneous V_m, 1 min into an exposure to pH_e = 8.5 solution; the calculated G_m is the gray data point in Fig. 9A. The cell was then exposed to pH_e = 8.5 solution containing 5 mM NH₄Cl, which caused V_m to substantially depolarize. G_m was determined in these conditions from further I–V relationships obtained 1 min apart from each other (filled symbols). Finally, the original pH_e = 8.5 solution (lacking NH₃/NH₄⁺) was restored to the bath while V_m was voltage-clamped at its new depolarized value (−5 mV in the example shown in Fig. 9A, to mimic the depolarizing effect of NH₃/NH₄⁺ presence) and five further I–V relationships were obtained at 1, 2, 3, 5, and 10 min following NH₃/NH₄⁺ removal (open symbols). The G_m data points with (*) or without (#) NH₃/NH₄⁺ in Fig. 9A were generated from the I–V plots shown in Fig. 9B, which in turn were generated from the current ladders shown in Fig. 9C. Figure 9D shows the average G_m values measured at each point and demonstrates that, while the removal of NH₃/NH₄⁺ causes a significant decrease in G_m, the majority of the conductance persists even 10 min after extracellular NH₃/NH₄⁺ removal.

Fig. 9.

NH₃/NH₄⁺-independence of G_m in mSlc4a11-expressing oocytes. A: representative experiment showing progressive changes in G_m in an mSlc4a11-expressing oocyte exposed to pH_e = 8.5 solution containing 5 mM NH₄Cl at spontaneous V_m (black data points), or subsequently to NH₃/NH₄⁺-free pH_e = 8.5 solution (white data points). Data points marked with * or # were generated from data similarly marked in B accompanying representative I–V plots and C accompanying current ladders. D: averaged data from six cells showing the influence of extracellular NH₃/NH₄⁺ removal on NH₃/NH₄⁺-induced G_m. Statistic is the result of a two-tailed, paired t-test.

DISCUSSION

Our conclusions regarding the molecular action of mSlc4a11.

Our data show that Slc4a11 is a pH_i and pH_e-stimulated H⁺/OH⁻ conductance. A major strength of our approach that allows us to extend this observation beyond previous studies with mammalian cells is that we simultaneously monitor and control V_m as we observe pH_i changes. Thus for the first time we report that mSlc4a11-expressing cells exhibit an exaggerated V_m response to pH_e changes. By matching our observations to the predicted V_m response of an ideally H⁺/OH⁻-permeable membrane, we make the novel observation that the apparent H⁺/OH⁻ permeability of mSlc4a11-expressing cells increases as pH_e rises (Fig. 1D). Furthermore, at pH_e = 8.5—the value in our assay at which H⁺/OH⁻ permeability appears to be greatest—the relationship between the transmembrane pH gradient and V_m is close to Nernstian (Fig. 3). This is a particularly powerful result when considering the relative scarcity of H⁺/OH⁻ compared to other ions in our solutions such as Na⁺. Equation 1 predicts that even if mSlc4a11 had a P_Na/P_H/OH permeability ratio of 10⁻⁶, V_m would tend towards E_Na [∼ +60 mV (57)] rather than E_H and would be essentially unresponsive to changes in the transmembrane pH gradient. The same holds true for H⁺/OH⁻ vs. all other components of our solutions that are present at mM concentration. Our data also rule out any obligate Na⁺, or Cl⁻, dependence of the mSlc4a11-mediated H⁺/OH⁻ current (Figs. 6 and 7).

Another unique finding of our study is that the mSlc4a11 conductance is acutely and intrinsically sensitive to changes in pH_e and pH_i, with rises in either parameter, independently resulting in an increase in G_m (Figs. 4 and 5). This observation suggests that OH⁻ is the charge carrier, but there is clearly an allosteric effect of pH_i and pH_e because under no circumstances did we observe a rectification in the I–V relationships that would be expected if the pH_e- or pH_i-dependent rises in G_m were simply caused by a rise in OH⁻ (or even HCO₃⁻)⁵ on one side of the membrane. The diminution of G_m caused by lowering pH_i explains one confounding observation in Fig. 2: the reason that mSlc4a11 does not appear to be substantially conductive at physiological pH_e. At rest, mSlc4a11-expressing oocytes are unusually acidified (pH_i = 6.95), far below the values of pH_i in our H₂O-injected oocytes (pH_i = 7.22) or the value expected in, for example, a corneal endothelial cell [pH_i ∼ 7.3 (4, 26)]. The lower steady-state pH_i in oocytes in culture at pH_e = 7.5 is consistent with the unopposed action of a H⁺/OH⁻ channel: the driving force acting on mSlc4a11 expressed in a cell with initial conditions pH_i = 7.22, and V_m = −50 mV (per our H₂O-injected cells) will promote acid loading. mSlc4a11-mediated acid-loading should continue until electrochemical equilibrium is reached, or until pH_i falls to a value at which mSlc4a11 is not substantially conductive. As V_m (−34 mV) is close to E_H (−32 mV) at rest in these cells, it is likely a combination of these factors apply. Thus the more acidified and depolarized state of mSlc4a11 expressing cells vs. H₂O-injected cells is a predictable consequence of the expression of a H⁺/OH⁻ conductance.

The predominant electrogenic permeability of the membrane of mSlc4a11-expressing cells is to H⁺/OH⁻. Accordingly, pH_i of these cells is almost exclusively under electrical control. However, under zero-current conditions we observe that pH_i is always slightly more acidic than predicted if H⁺ were at electrochemical equilibrium (e.g., non-zero x-axis intercepts extrapolated from data in Fig. 3C and the lower-than-7.5 value of pH_i in Fig. 5B). This phenomenon is consistent with one or a combination of the following factors:

1) The action of an electroneutral acid-loading activity such as an endogenous Cl⁻/HCO₃⁻ exchanger (under these particular assay conditions, pH_i is typically more alkaline than that of a H₂O-injected oocyte, which would tend to promote such endogenous acid loading mechanisms),

2) The pH_i sensed by mSlc4a11 at the plasma membrane following a period of robust acid-extrusion being more alkaline than the bulk pH_i sensed by our pH_i electrode, the tip of which is likely positioned tens of μm away from the inner surface of the membrane, or

3) The pH at the cell surface (pH_s) sensed by mSlc4a11 being more acidic than bulk pH_e; indeed even in H₂O-injected oocytes pH_s is slightly more acidic than bulk pH_e (15).

Importantly, none of our observations are consistent with mSlc4a11 mediating acid-extruding activities such as Na⁺/H⁺ exchange or Na⁺-OH⁻ cotransport. Thus we conclude that mSlc4a11 is an ideally selective H⁺/OH⁻ conductance. Whether this conductance represents the action of a facilitative carrier (or H⁺/OH⁻ exchanger) rather than a bona fide ion channel could be experimentally determined in future work by study of single-channel events in membrane patches expressing mSlc4a11.

To our knowledge, although many proteins are capable of conducting H⁺/OH⁻, the major passive H⁺-selective conductance in animals is the voltage-gated H⁺-channel H_v1, the time and voltage-dependence of which are not features of the currents associated with mSlc4a11 (9, 10, 50). In summary, we find mSlc4a11 to be a unique entity: an ideally H⁺/OH⁻-selective, voltage and time-independent conductance that is independently stimulated by rises in pH_e and pH_i.

Comparisons with the data of others regarding Slc4a11 action.

We must preface any discussion of previous studies with a disclaimer that ours uses mouse rather than human Slc4a11. The mouse and human orthologs of Slc4a11 are 85% identical at the amino -acid level, with a major disparity at the extreme N-terminus of the protein, which is 29–72 aa shorter in the mouse ortholog depending on which human variant is considered [SLC4A11-A, B, or C (30)]. The role, if any, of these N-terminal appendages is unknown, but comparison with other Slc4 proteins suggests that these are more likely to play a regulatory role than to determine intrinsic transport action (e.g., refs. 37, 45, 54). We must also point out that most previous studies draw on experiments that have been conducted using mammalian cells, whereas we are using oocytes. The use of oocytes provides the advantage of studying mSlc4a11 in the absence of complications from substantial endogenous pH-regulating mechanisms. The use of oocytes also requires us to note that we are studying mSlc4a11 action in the absence of any natural binding partners/accessory subunits; thus our study considers the intrinsic properties of the mSlc4a11 polypeptide.

Our data are in best agreement with those of Kao and coworkers (30), who studied human SLC4A11-C in HEK cells. As we find with oocytes, the pH_i of SLC4A11-expressing HEK cells is also more acidic at rest than that of mock-transfected cells, although not to the same extent as the oocytes in our study. This difference is explained by the acid-extruding action of an EIPA-sensitive NHE that is endogenous to HEK cells; in the presence of EIPA the pH_i of these cells falls to a value that is presumably close to E_H (30). The findings of Kao et al. are exemplified by the observation that lowering extracellular K⁺ (or Na⁺ in gramicidin-permeabilized cells), a maneuver that ought to hyperpolarize V_m, results in an acidification of Slc4a11-expressing cells. This response is consistent with expression of electrogenic H⁺/OH⁻ permeability. A potential disparity between the behaviors of human SLC4A11-C in HEK cells vs. mSlc4a11 in oocytes is that, unlike the conductance mediated by mSlc4a11, the H⁺/OH⁻ fluxes mediated by the human clone show no obvious sign of pH-dependence. However, methodological and technical differences between the two studies complicate a direct data comparison.

Other studies report that mSlc4a11-mediated H⁺/OH⁻ transport is coupled to the transport of another molecular species. Several groups have reported that mSlc4a11 interacts with Na⁺ (28, 42, 43). One group found that removing extracellular Na⁺ caused HEK cells to acidify. However, Na⁺ removal also caused a rapid depletion of Na⁺ in both mock- and mSlc4a11-transfected cells in that study (43); both were thus likely to be hyperpolarized by this maneuver. Therefore, with hindsight, the acidification could be explained as the predictable effect of hyperpolarization on Slc4a11-mediated H⁺/OH⁻ conductance rather than a direct Na⁺-dependence of Slc4a11 action. Consistent with this hypothesis, Kao et al. (30) only observe acidification upon Na⁺ removal in cells that have been permeabilized with gramicidin. Beyond the observation of apparent Na⁺-dependence, that same group also reported that intracellular [Na⁺] increased in mSlc4a11-expressing cells when they were exposed to a pH_e = 8.2 solution and decreased when exposed to a pH_e = 6.2 solution: an observation seemingly consistent with Na⁺/H⁺ exchange or Na⁺-OH⁻ cotransport. The caveat to this result is that these maneuvers elicit dramatic pH_i changes in these cells and the observed changes in the fluorescence of the SBFI Na⁺ indicator could be explained by the pH sensitivity of SBFI (13). The robustness of a second report of Slc4a11 Na⁺ permeability in mammalian cells (42) is unclear as that activity was unusually sensitive to EIPA and may represent the action of an endogenous protein (59). Finally, two groups do not find any evidence for the influence Na⁺ on Slc4a11 action (30, 34). In summary, taken together with our data, there is no robust evidence for the Na⁺ permeability of Slc4a11 or the obligate Na⁺-dependence of Slc4a11-mediated H⁺-OH⁻ transport.

Two recent studies report an interaction of Slc4a11 with NH₃ (34, 59) Both groups of researcher find, as we do (Fig. 8), that the application of NH₃/NH₄⁺ induces the appearance of currents and that the magnitude of these currents increase as pH_e rises. The authors of the first study (performed in PS120-NHE-deficient Chinese hamster ovary-cells) conclude that these currents represent the coupled cotransport of NH₃ and H⁺, (59) while the authors of the second study (performed in oocytes) find against either NH₃-coupled or uncoupled H⁺/OH⁻ conductance and suggest instead that these currents could represent an endogenous pH-sensitive NH₄⁺ channel (34) Because we only observe these currents in Slc4a11-expressing cells (Fig. 2 and Fig. 3), because we observe these currents in the absence of NH₃/NH₄⁺ (Figs. 2– 7), and because we find that the NH₃/NH₄⁺-induced currents persist long after the removal of NH₃/NH₄⁺ (Fig. 9), we conclude that the currents observed in the presence of NH₃/NH₄⁺ do not involve mSlc4a11-mediated NH₃/NH₄⁺ transport or the action of an endogenous NH₄⁺ channel. There are three mechanisms by which NH₃/NH₄⁺ application could indirectly enhance Slc4a11-mediated H⁺/OH⁻ conductance:

1) Independently of the presence of heterologously expressed Slc4a11, the application of NH₃/NH₄⁺ rapidly and robustly depolarizes both mammalian cells and oocytes (1, 7). This depolarization in itself is sufficient to increase mSlc4a11-mediated H⁺/OH⁻ currents by increasing the driving force for mSlc4a11-mediated acid-extrusion, as we observe in mSlc4a11-expressing oocytes that are voltage-clamped at values more positive than spontaneous V_m (e.g., Fig. 4A).

2) The rise in pH_i brought about by Slc4a11-mediated acid-extrusion further increases G_m.

3) The influx of NH₃ across the plasma membrane would exert an additional stimulatory effect by further alkalinizing pH_i (22, 27, 55)⁶.

Interpretation aside, the major disparity between our study and the findings of the two NH₃-related reports lies in their lack of evidence for H⁺/OH⁻ conductance in the absence of NH₃/NH₄⁺. Actually, the authors of the study in PS120 cells did detect evidence for uncoupled H⁺/OH⁻ conductance, but concluded that it was relatively small and therefore of less physiological significance than NH₃/H⁺ cotransport (59). However, the assay used is not conducive to the appearance of these currents: cells were voltage-clamped at − 60 mV while the cells were exposed to a pH_e = 8.5 solution. We would not expect to detect substantial currents during a brief exposure to these solutions, especially when voltage-clamped at a value that is presumably close to E_H. In fact we find that the H⁺/OH⁻ conductance can be similarly robust in the absence vs. the presence of NH₃/NH₄⁺ (Fig. 9). The authors of the oocyte study find that the magnitude of the change in zero-current potential in NH₃/NH₄⁺ solution as pH_e rises from 7.4 to 8.4 (expected to be −58 mV, if pH_i remains constant) is inconsistent with coupled NH₃-H⁺ cotransport or uncoupled H⁺/OH⁻ conduction. However, two factors should be taken into account when considering these data:

1) V_m is probably not dominated by Slc4a11 action in these cells as the elicited currents are 10 times smaller than those that we observe in the presence of NH₃/NH₄⁺ and are undetectable in the absence of NH₃/NH₄⁺ (i.e., V_m may not be the same as the reversal potential of Slc4a11).

2) In the presence of NH₃/NH₄⁺ when the currents are greatest in those experiments, V_m includes a contribution from the depolarizing influence of NH₃/NH₄⁺ (i.e., V_m may not be the same as the reversal potential of Slc4a11 even in the presence of NH₃/NH₄⁺). Furthermore, pH_i is not the same when pH_e = 7.4 vs. 8.4. In fact when maximally conductive, Slc4a11 is likely capable of rapidly dissipating transmembrane pH gradients in which case no change in E_rev would be expected as pH_e is changed.

In summary, we see no electrophysiological evidence in the work of others that cannot be explained by the action of a pH_i and pH_e-sensitive H⁺/OH⁻ conductance.

The pathophysiological role of mSlc4a11.

An interesting and parallel line of investigation finds Slc4a11 to be permeable to both H₂O (at pH_e = 7.4 in the absence of NH₃/NH₄⁺) and NH₃ (34, 51), like members of the aquaporin protein family (40). As many ion transporters exhibit H₂O-permeability (31, 58) and as it is theoretically possible for a membrane protein to conduct both water and H⁺ (32), these two hypotheses are not mutually exclusive. Indeed, many H₂O-carrying transport proteins such as Na⁺-glucose and H⁺-lactate cotransporters have been described to carry H₂O during transport cycles at rates not dissimilar to that reported for Slc4a11 (58). In the case of H₂O-carrying transporters, as opposed to H₂O channels, molecular dynamic simulations indicate that the large-scale conformational changes associated with alternating access transport mechanisms permit the transient formation of water-permeable pathways (31). As we do not yet know the mechanism by which H⁺/OH⁻ is carried by Slc4a11 (channel vs. transporter), it is premature to speculate how the H⁺/OH⁻ and H₂O conducting pathways are related. It is even possible that the two pathways are entirely discrete, e.g., one passes through each monomer and another passes through the interface between two monomers in a dimer. However, as both actions are blocked by mutations equivalent to the CHED-causing mutation R125H, it is entirely likely that both activities are related to the physiological function of Slc4a11.

A key question is how might a H₂O-permeable, pH-sensitive H⁺/OH⁻ conductance support corneal health? The role of the corneal endothelial cells in which Slc4a11 is expressed—and which are dysfunctional or lost in CHED—is to promote fluid re-secretion from the collagenous corneal stroma (49). The stroma has the potential to draw water from the aqueous humor due to the oncotic pressure exerted by its proteinaceous matrix: if unopposed, as in CHED, the fluid accumulation distorts the matrix causing loss of vision. The endothelial cells transport Na⁺, HCO₃⁻, and lactate⁻ (and thence water) from the stromal fluid and back into the aqueous humor to maintain a normal state of stromal hydration (Fig. 10A). Leading this action in the basolateral (stromal facing) membrane of endothelial cells are the Na⁺-2HCO₃⁻ cotransporter NBCe1 and the H⁺-coupled lactate transporter MCT1 (3). The basolateral membrane is also the location of Slc4a11. The role of H₂O permeability in this membrane is obvious and has been discussed previously by its original proponents (51), so we will confine our discussion to consideration of the value of a pH-sensitive H⁺/OH⁻ conductance.

Fig. 10.

Simplified model of ion transport across the corneal endothelium. A: Slc4a11 is expressed in the same membrane as the Na⁺-2HCO₃⁻ cotransporter (NBCe1) and H⁺-Lac⁻ cotransporter (MCT1). Although the actions of these two transporters are coupled, stromal [Lac⁻] can vary throughout the day. Imbalances in NBCe1/MCT1 action could disturb pH_i and transendothelial fluid movement. B and C: Slc4a11 action could respond to and compensate for such imbalances to maintain pH_i and endothelial function, preventing a rise in pH_i when stromal lactate is low and MCT1 action is suboptimal (B) and supporting MCT1 action when stromal lactate is elevated (C).

The acid-base status of the fluid compartments on either side of the basolateral membrane is under constant challenge from the transmembrane flux of H⁺ and HCO₃⁻: the pH in the immediate vicinity of a robust acid-base transport can be as much as 1 pH unit different from bulk pH (11, 36). Although the actions of NBCe1 and MCT1 are functionally coupled (2), they are presumably not always perfectly balanced as stromal [lactate] varies substantially throughout the day (21). Basolateral NHE1 and AE2 are situated to respond to and counter changes in pH_i in these cells so the uniqueness of the role of Slc4a11 presumably lies with its unique molecular action, i.e., the ability to respond to changes in pH_e/pH_s and V_m and, being uncoupled from other ion gradients, its ability to operate either as an acid loader or an acid extruder depending on prevailing conditions. Stromal pH is 7.5 and endothelial pH_i is 7.3, meaning that E_H ∼ −12 mV. Unfortunately, there is a great deal of uncertainty concerning the V_m of endothelial cells; values have been reported that are more negative than − 35 mV (56), in which case Slc4a11 would act as an acid loader unless pH_e was to rise above 7.9, and as depolarized as 0 mV (25), in which case Slc4a11 would act as an acid extruder. There could be numerous roles for such a protein; we suggest two scenarios in which a pH-sensitive H⁺/OH⁻ conductance could be useful:

1) As an electrogenic acid-loader, Slc4a11 would provide a constant and preemptive influx of H⁺ that, along with MCT1 action, would minimize pH changes caused by the constant electrogenic influx of HCO₃⁻, especially when stromal [lactate] is low (Fig. 10B). This would presumably be preferable to balancing the pH gradient across the basolateral membrane via the action of the Cl⁻/HCO₃⁻ exchanger AE2. pH balance via the functional coupling of AE2 to NBCe1 would result in futile cycles of HCO₃⁻/HCO₃⁻ exchange, thereby reducing the efficiency of transendothelial bicarbonate transport. Furthermore, the resulting net uptake of NaCl could disturb endothelial cell volume and transendothelial fluid transport.

2) MCT1 action requires support from an extracellular carbonic anhydrase to maintain an adequate supply of H⁺ to drive H⁺/Lac⁻ cotransport (41). During periods of elevated MCT1 activity, such as during sleep when eye closure promotes anaerobic metabolism and lactate production, Slc4a11 could provide additional support in response to the depletion of surface H⁺ (elevation in pH_s) to extrude and recycle H⁺ (Fig. 10C). This action would promote MCT1 activity, while preemptively countering intracellular acidification.

Such actions could contribute to pH_i and cell volume homeostasis, while optimizing the efficiency of endothelial fluid transport. Thus loss of Slc4a11 would be expected to have an immediate impact on endothelial function with longer-term detrimental consequences for endothelial health. This hypothesis could explain why both loss of endothelial function and subsequent loss of endothelial cells occurs in CHED.

GRANTS

This work was supported by startup funding from the Dean of the School of Medicine and Biomedical Sciences and the Department of Physiology and Biophysics at UB:SUNY (to M. D. Parker). Salary support was provided to E. J. Myers via a Carl W. Gottschalk Research Scholar Grant from the ASN Foundation for Kidney Research (to M. D. Parker). The subcloning of mSlc4a11.pGH19 and the generation of the anti-Slc4a11 antibody were supported by an R21 award (EY021646) from the National Institutes of Health (to M. L. Jennings and M. D. Parker).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

E.J.M., M.L.J., and M.D.P. conceived and designed the research; E.J.M., A.M., and M.D.P. performed experiments; E.J.M. and M.D.P. analyzed data; E.J.M. and M.D.P. interpreted results of experiments; E.J.M. and M.D.P. prepared figures; E.J.M. and M.D.P. drafted manuscript; E.J.M., M.L.J., and M.D.P. edited and revised manuscript; E.J.M., A.M., M.L.J., and M.D.P. approved final version of manuscript.