Mouse Ae1 E699Q mediates SO₄²⁻_i/anion_o exchange with [SO₄²⁻]_i-dependent reversal of wild-type pH_o sensitivity

Abstract

The SLC4A1/AE1 gene encodes the electroneutral Cl⁻/HCO₃⁻ exchanger of erythrocytes and renal type A intercalated cells. AE1 mutations cause familial spherocytic and stomatocytic anemias, ovalocytosis, and distal renal tubular acidosis. The mutant mouse Ae1 polypeptide E699Q expressed in Xenopus oocytes cannot mediate Cl⁻/HCO₃⁻ exchange or ³⁶Cl⁻ efflux but exhibits enhanced dual sulfate efflux mechanisms: electroneutral exchange of intracellular sulfate for extracellular sulfate (SO₄²⁻_i/SO₄²⁻_o exchange), and electrogenic exchange of intracellular sulfate for extracellular chloride (SO₄²⁻_i/Cl⁻_o exchange). Whereas wild-type AE1 mediates 1:1 H⁺/SO₄²⁻ cotransport in exchange for either Cl⁻ or for the H⁺/SO₄²⁻ ion pair, mutant Ae1 E699Q transports sulfate without cotransport of protons, similar to human erythrocyte AE1 in which the corresponding E681 carboxylate has been chemically converted to the alcohol (hAE1 E681OH). We now show that in contrast to the normal cis-stimulation by protons of wild-type AE1-mediated SO₄²⁻ transport, both SO₄²⁻_i/Cl⁻_o exchange and SO₄²⁻_i/SO₄²⁻_o exchange mediated by mutant Ae1 E699Q are inhibited by acidic pH_o and activated by alkaline pH_o. hAE1 E681OH displays a similarly altered pH_o dependence of SO₄²⁻_i/Cl⁻_o exchange. Elevated [SO₄²⁻]_i increases the K_1/2 of Ae1 E699Q for both extracellular Cl⁻ and SO₄²⁻, while reducing inhibition of both exchange mechanisms by acid pH_o. The E699Q mutation also leads to increased potency of self-inhibition by extracellular SO₄²⁻. Study of the Ae1 E699Q mutation has revealed the existence of a novel pH-regulatory site of the Ae1 polypeptide and should continue to provide valuable paths toward understanding substrate selectivity and self-inhibition in SLC4 anion transporters.

the slc4 bicarbonate transporter superfamily encodes Na⁺-dependent and Na⁺-independent anion carriers (3, 30, 32, 44). SLC4A1/AE1/band 3 is the most thoroughly studied among the Na⁺-independent, electroneutral Cl⁻/HCO₃⁻ exchangers. AE1 is the major intrinsic protein of the erythrocyte membrane, where it serves to increase the total CO₂-carrying capacity of the blood (48). The kidney variant of AE1 is expressed in the basolateral membrane of the renal type A intercalated cell. AE1 mutations cause dominant hereditary spherocytosis, stomatocytosis, and ovalocytosis, and a distinct set of mutations cause dominant and recessive forms of distal renal tubular acidosis (2, 24).

In addition to 1:1 monovalent anion exchange, AE1 can mediate cotransport of H⁺/SO₄²⁻ in exchange for Cl⁻ or as a self-exchange reaction (27, 28). The binding site for the cotransported proton during H⁺/SO₄²⁻ cotransport is believed to be E681 of human AE1 (19). Treatment of intact human erythrocytes with Woodward's reagent K (WRK) followed by borohydride (BH₄) reduction converts AE1 E681 to the corresponding alcohol (hAE1 E681OH). This modification produces multiple alterations in anion transport. These changes include severely reduced rates of electroneutral Cl⁻/Cl⁻ exchange rates and dramatically increased H⁺-independent rates of electrogenic SO₄²⁻/Cl⁻ exchange and electroneutral SO₄²⁻/SO₄²⁻ exchange (15, 17). Similar changes were observed in the function of recombinant mouse AE1 (mAe1) mutated at the corresponding E699 to the amide Gln. In particular, the high rate of Cl⁻/Cl⁻ exchange characteristic of wild-type Ae1 became undetectable. In addition, the low wild-type rates of H⁺/SO₄²⁻ cotransport were converted to greatly accelerated rates of H⁺-independent, electroneutral SO₄²⁻/SO₄²⁻ exchange and electrogenic SO₄²⁻/Cl⁻ exchange (7).

In the course of investigating the altered properties of mAe1 E699Q-mediated SO₄²⁻ transport, we observed a novel pH dependence of SO₄²⁻ transport. We show in the present study that, in contrast to activation of wild-type AE1-mediated SO₄²⁻ transport by H⁺ consistent with H⁺/SO₄²⁻ cotransport, SO₄²⁻ transport by the mAE1 mutant E699Q was inhibited by acidic extracellular pH (pH_o) and activated by alkaline pH_o. This pattern of regulation, diametrically opposed to that of wild-type AE1, was shared by SO₄²⁻_i/Cl⁻_o and SO₄²⁻/SO₄²⁻ exchanges mediated by mAE1 E699Q, and by SO₄²⁻_i/Cl⁻_o exchange mediated by hAE1 E681OH. The extracellular pH_o(50) for ³⁵SO₄²⁻ efflux (the pH_o at which efflux was half-maximal) was not significantly alkaline-shifted in the presence of acidic intracellular pH (pH_i). Moreover, changing pH_i at constant pH_o had no effect on either SO₄²⁻_i/Cl⁻_o or SO₄²⁻/SO₄²⁻ exchange. Elevation of intracellular [SO₄²⁻] ([SO₄²⁻]_i) greatly attenuated inhibition of ³⁵SO₄²⁻ efflux by acidic pH_o. Elevated [SO₄²⁻]_i also increased K_1/2 for extracellular SO₄²⁻ ([SO₄²⁻]_o) in SO₄²⁻/SO₄²⁻ exchange and greatly increased K_1/2 for extracellular Cl⁻ ([Cl⁻]_o) in SO₄²⁻_i/Cl⁻_o exchange. mAE1 E699Q mediated exchange of SO₄²⁻_i with a wide range of extracellular oxyanions at rates comparable to those measured for SO₄²⁻_i/Cl⁻_o exchange and for SO₄²⁻/SO₄²⁻ exchange.

METHODS

Reagents.

Na³⁵SO₄²⁻ was purchased from ICN Biomedicals (Costa Mesa, CA) or from MP Biomedicals (Irvine, CA). Other chemical reagents were of analytical grade and were obtained from Sigma-Aldrich (St. Louis, MO), Fluka (St. Louis, MO), or Calbiochem (San Diego, CA). Transcription plasmid cDNA encoding mouse AE1 mutant E699Q was described previously (7).

Solutions for Xenopus laevis oocytes.

ND-96 medium consisted of (in mM) 96 NaCl, 2 KCl, 1.8 CaCl₂, 1 MgCl₂, 5 HEPES, and 2.5 sodium pyruvate, pH 7.40. Flux media lacked sodium pyruvate. pH values of 7.0, 8.0, and 8.5 in flux media were achieved with 5 mM HEPES. 4-Morpholino-ethanesulfonic acid (MES; 5 mM) was used for flux media of pH values 5.0 and 6.0. In Cl⁻-free solutions, NaCl was replaced isosmotically with 96 mM sodium isethionate, and equimolar K, Ca, and Mg gluconate substituted for the corresponding Cl⁻ salts. Addition to flux media of the weak acid salt sodium butyrate or of NH₄Cl was in equimolar substitution for NaCl.

Expression of cRNA in Xenopus oocytes.

Transcription template was generated by linearizing plasmid cDNA with HindIII. cRNA transcription with T7 RNA polymerase was performed with Ambion's Megascript Kit (Austin, TX). Female Xenopus laevis were purchased from NASCO (Madison, WI). Manually defolliculated oocytes were prepared as previously described (13), in compliance with a protocol approved by the Beth Israel Deaconess Medical Center Institutional Animal Care and Use Committee, were microinjected with 10–20 ng cRNA encoding mAe1 E699Q, and were then incubated in ND-96 supplemented with gentamicin at 19°C for 2–7 days before use in efflux experiments (7).

³⁵SO₄²⁻ efflux experiments in Xenopus oocytes.

Oocytes were injected with 50 nl of a solution containing Na₂³⁵SO₄ (0.25–0.5 μCi, 3–6 μM) in 130 mM of either HEPES, pH 7.4, MES, pH 5.0, or occasionally, Bis-Tris, pH 5.0. When indicated, the injectate additionally contained 130 mM cold Na₂SO₄, yielding a final estimated [SO₄²⁻]_i of ∼14 mM [including the endogenous [SO₄²⁻] of ∼1 mM (7)]. Following a 10-min recovery period in SO₄²⁻-free, Cl⁻-free medium containing Na isethionate, ³⁵SO₄²⁻ efflux was initiated by transfer of individual oocytes into 1 ml efflux medium containing 96 mM NaCl for assay of SO_4i/Cl⁻_o exchange, or containing 20 or 64 mM Na₂SO₄ for assay of SO₄²⁻_i/SO₄²⁻_o exchange. Experiments in which extracellular sulfate or chloride concentrations were varied used isethionate as substituting anion to maintain nominal isosmolarity. At regular intervals of 3 min, 950 μl of this medium was removed for scintillation counting and replaced with fresh medium. Experiments ended with a final efflux period in the presence of the AE1 inhibitor, 4,4′-diisothiocyanostilbene-2,2′-disulfonic acid (DIDS, 100 or 200 μM), before solubilization of the washed oocyte in 100 μl of 1% sodium dodecyl sulfate (SDS). All ³⁵SO₄²⁻ efflux experiments were conducted at room temperature.

Efflux cpm values for water-injected or cRNA-injected oocytes in the presence of DIDS were less than threefold above machine background values (∼20 cpm). Within each experiment, water-injected and cRNA-injected oocytes from the same frog were subjected to parallel measurements. With the exception of the SO₄²⁻/SO₄²⁻ exchange experiment presented in Fig. 8B, each experiment was performed with oocytes obtained from at least two frogs.

Fig. 8.

Elevated [SO₄²⁻]_i attenuates inhibition by acidic pH_o of mAe1 E699Q-mediated SO₄²⁻_i/Cl⁻_o exchange. A: ³⁵SO₄²⁻ efflux time courses of representative individual oocytes (estimated [SO₄²⁻]_i 14 mM, injected with HEPES, pH 7.4) previously injected with water (top trace) or with mAe1 E699Q cRNA (bottom traces), during sequential exposure to pH_o 5.0 and 7.4, then to 100 μM DIDS. B: mean normalized efflux rate constants for (n) oocytes at pH_o 5.0 and pH_o 7.4 measured as depicted in A. C: ³⁵SO₄²⁻ efflux time courses of representative individual oocytes (estimated [SO₄²⁻]_i 14 mM, injected with MES, pH 5.0) previously injected with water (top trace) or with mAe1 E699Q cRNA (bottom traces), during sequential exposure to pH_o 5.0 and 7.4, then to 100 μM DIDS. D: mean normalized efflux rate constants for (n) oocytes at pH_o 5.0 and 7.4 measured as depicted in C.

Efflux time course data were plotted as ln (% cpm remaining in oocyte) vs. time. Efflux rate constants were measured from linear least squares regressions calculated from the final three time points of each experimental condition and were corrected by subtraction of the efflux rate constant exhibited by water-injected oocytes studied in the same experiment. Samples were counted for 3–5 min such that the magnitude of 2 SD was <5% of the sample mean; >98% of the originally injected cpm was accounted for in the sum of the combined efflux fractions and the cpm remaining in the oocyte at the end of the experiment. Oocyte water space was assumed to 450 nl (500 nl acutely following injection of the 50-nl volume) such that final estimated oocyte [SO₄²⁻]_i was either ∼1 mM or ∼14 mM (7).

To describe the dependence of SO₄²⁻/anion exchange on pH_o, efflux rate constants measured at each pH_o value were fit (Ultrafit 3.0, Biosoft, Ferguson, MO; or Sigmaplot 8.0; Systat, San Jose, CA) by the following first-order logistic sigmoid equation:

(1)

where k_eff is the AE1 E699Q-mediated ³⁵SO₄²⁻ efflux rate constant (in units of min⁻¹), k_max is the maximal AE1 E699Q-mediated ³⁵SO₄²⁻ efflux rate constant, pH_o is the extracellular pH at which the rate constant was measured, pH_o(50) is the pH_o at which k_eff is half-maximal, and d = 0 or a nonzero value, as indicated in the text.

To describe the Cl⁻_o dependence of SO₄²⁻_i/Cl⁻_o exchange and the SO₄²⁻_o-dependence of SO₄²⁻/SO₄²⁻ exchange, ³⁵SO₄²⁻ efflux rate constants plotted as functions of [anion]_o were fit by the Michaelis-Menten equation modified to include an added constant, y₀:

(2)

where k and k_max are as above, K_1/2 is the apparent K_m for extracellular permeant substrate, S is the extracellular concentration of permeant anion substrate, and y₀ is the efflux rate constant at S = 0 (in the presence of the nominally nonpermeant anion, isethionate). Fits were slightly inferior without addition of the y₀ offset, but mean K_1/2 values calculated with and without the y₀ term did not significantly differ. The magnitude of the ratio y₀/k_max was larger for SO₄²⁻/SO₄²⁻ exchange experiments than for SO₄²⁻_i/Cl⁻_o exchange. The r² values of the nine fits to Eq. 2 of SO₄²⁻/SO₄²⁻ exchange at [SO₄²⁻]_i ∼ 1 mM were 0.99 for seven individual oocytes and 0.96 and 0.92 for two additional oocytes (Fig. 7B).

Fig. 7.

[SO₄²⁻]_i regulates K_1/2 for SO₄²⁻_o in mAe1 E699Q-mediated SO₄²⁻_i/SO₄²⁻_o exchange. A: ³⁵SO₄²⁻ efflux time courses of representative individual oocytes (estimated [SO₄²⁻]_i = 1 mM) previously injected with water (top trace) or with mAe1 E699Q cRNA (bottom traces) during sequential exposures to progressively increased [SO₄²⁻]_i, followed by 100 μM DIDS. B: plot of efflux rate constant vs. [SO₄²⁻]_o up to 20 mM for an individual oocyte measured as depicted in A, with Michaelis-Menten fit to data. Inset: data for same oocyte including measurement at 64 mM SO₄²⁻_o. C: ³⁵SO₄²⁻ efflux time courses of representative individual oocytes (estimated [SO₄²⁻]_i = 14 mM) previously injected with water (top trace) or with mAe1 E699Q cRNA (bottom traces) during sequential exposures to progressively increased [SO₄²⁻]_o. D: plot of efflux rate constant vs. [SO₄²⁻]_o up to 20 mM for an individual oocyte containing 14 mM [SO₄²⁻]_i measured as depicted in C, with Michaelis-Menten fit to data.

Measurement of oocyte pH_i.

Oocytes loaded 30 min at room temperature with 5 μM BCECF-AM were set into a Leiden closed perfusion chamber of 1 ml internal volume and mounted on the stage of an Olympus BH-2 epifluorescence microscope. BCECF fluorescence emission at 530 nm was recorded at alternating excitation wavelengths of 440 and 490 nm by an Image-1 digital ratio imaging system (Universal Imaging, West Chester, PA). Fluorescence ratios in a crescent-shaped region of interest just inside the oocyte edge were measured before and 10 min after injection of 50 nl of 130 mM MES, pH 5.0. In situ calibration of the BCECF fluorescence ratio to pH was as previously described (13, 35, 47).

Treatment of human erythrocytes with Woodward's reagent K.

Red blood cells from healthy adults were obtained by a venipuncture protocol approved by the University of Arkansas for Medical Sciences Institutional Review Board, washed, and modified with Woodward's reagent K (WRK) and NaBH₄ as described previously (15). Cells were washed three times and suspended at 5% hematocrit in 150 mM KCl and 10 mM MOPS, pH 7.0. After the suspension was chilled on ice for at least 20 min, solid WRK (Sigma-Aldrich, St. Louis, MO) was added to a final concentration of 2 mM. The suspension was mixed gently and incubated 10 min further on ice. NaBH₄ (Sigma-Aldrich) was added from a freshly prepared 1 M stock (in 0.1 N NaOH) to a final concentration of 2 mM. After 5 min on ice, an additional 2 mM NaBH₄ was added, and the suspension was incubated another 5 min. Between the NaBH₄ additions, 10 mM MOPS acid was added to the suspension to keep the pH from rising above 7. This procedure converts ∼80% of the copies of wild-type hAE1 to hAE1 E681OH (15).

³⁵SO₄²⁻ efflux from erythrocytes.

After treatment with WRK and NaBH₄, cells were washed three times in at least 20 volumes of 80 mM K₂SO₄ and 10 mM HEPES, pH 7.45, with a 10-min incubation at 37°C to allow SO₄²⁻ to replace cellular Cl⁻. Cells were then loaded with ³⁵SO₄²⁻ by incubating 1 h, 37°C in the same medium containing 1 μCi Na³⁵SO₄/ml. Red blood cell [SO₄²⁻]_i after this protocol of sulfate loading and labeling was ≥40 mM. To measure ³⁵SO₄²⁻/Cl⁻ exchange, loaded and labeled cells were washed twice at 0°C in K₂SO₄/HEPES medium containing no radioactivity and resuspended in media consisting of 120 mM KCl, or 60 mM KCl, 120 mM sucrose, in each case buffered with one of the following: 20 mM Na-HEPES (pH 7.5), 20 mM Na-MES, pH 6.4, 10 mM Na-MES/10 mM Na-glutamate, pH 5.7, or 20 mM Na-glutamate, pH 4.4. The suspension was incubated at 20°C, and aliquots were centrifuged at various times (4 time points per efflux) for determination of extracellular radioactivity. The rate constants for efflux were determined as previously described (15). Data plotting efflux rate constant vs. pH_o data were fit using equation 1, with d = 0.

Statistical analysis.

Data are reported as means ± SE. Statistical significance of pairwise differences was assessed by Student's paired and unpaired t-tests. pH_o(50) values were compared by Dunnett's two-way t-test. The level of significance was defined as P < 0.05.

RESULTS

Acidic pH_o inhibits mAe1 E699Q-mediated SO₄²⁻/Cl⁻ exchange.

mAe1 E699Q-expressing oocytes injected with HEPES, pH 7.4 (final estimated intracellular concentration 13 mM, to minimize changes in pH_i) showed a 2.7 ± 0.2-fold stimulation in the rate of ³⁵SO₄²⁻_i/Cl⁻_o exchange upon transition from pH_o 5.0 to pH_o 7.4 (Fig. 1, A and B). In a separate set of oocytes not injected with buffer, transition from pH_o 5.0 to 8.0 stimulated the rate of ³⁵SO₄²⁻_i/Cl⁻_o exchange by 7.5-fold (n = 15, not shown). This inhibition by acidic pH_o is diametrically opposed to the acceleration of human red blood cell AE1-mediated ³⁵SO₄²⁻_i/Cl⁻_o exchange by acidic pH_o (27, 28). E699Q-mediated ³⁵SO₄²⁻_i/Cl⁻_o exchange was nearly completely inhibited by 100 μM DIDS (Fig. 1, A and C) and so did not represent a nonspecific divalent anion leak. ³⁵SO₄²⁻_i/Cl⁻_o exchange was also inhibited 35% in a rapidly reversible manner by 1 mM Zn²⁺ (n = 8, not shown). More detailed analysis of the pH_o dependence of mAe1 E699Q-mediated ³⁵SO₄²⁻_i/Cl⁻_o exchange (Fig. 1C) revealed a pH_o(50) value (Fig. 1D) of 6.20 ± 0.18 (n = 7, r² = 0.97) calculated with parameter d of Eq. 1 as a variable (solid line), and 5.68 ± 0.14 (r² = 0.92) for d = 0 (dashed line).

Fig. 1.

Mouse Ae1 (mAe1) E699Q-mediated SO₄²⁻_i/Cl⁻_o exchange is inhibited by acidic extracellular pH (pH_o). A: ³⁵SO₄²⁻ efflux time courses of representative individual oocytes previously injected with water (top trace) or with cRNA encoding mAe1 E699Q (bottom traces). After oocytes were injected with 50 nl 130 mM HEPES pH 7.4 containing carrier-free ³⁵SO₄²⁻ [estimated oocyte [SO₄²⁻]_i 1 mM (7)], they were exposed sequentially to Cl⁻-containing solutions of pH_o 5.0 and 7.4, and then to the inhibitor, DIDS (100 μM). B: mean normalized efflux rate constants for (n) oocytes studied as depicted in A. C: ³⁵SO₄²⁻ efflux time courses of representative individual oocytes previously injected with water (top trace) or with cRNA encoding mAe1 E699Q (bottom traces), injected with ³⁵SO₄²⁻ as in A, and exposed sequentially to solutions of the indicated pH values, followed by DIDS (100 μM). D: plot of efflux rate constants vs. pH_o (n = 7) fit to Eq. 1 with parameter d = 0 (dashed line) or treated as a variable (solid line).

Acidic pH_i does not significantly shift pH_o sensitivity of mAe1 E699Q-mediated SO₄²⁻/Cl⁻ exchange.

Since altering pH_o modestly changes pH_i in Xenopus oocytes (41), we investigated directly the effects of acidic pH_i on the pH_o sensitivity of mAe1 E699Q-mediated ³⁵SO₄²⁻_i/Cl⁻_o exchange. Figure 2, A and B, shows that oocytes injected with MES, pH 5.0 (final estimated intracellular concentration 13 mM) exhibited a 3.8 ± 0.3-fold acceleration of ³⁵SO₄²⁻_i/Cl⁻_o exchange (n = 10) upon transition from pH_o 5.0 to 7.4. This concentration of MES reduced oocyte pH_i (as measured at pH_o 7.4 by BCECF ratio fluorimetry) by 0.44 + 0.13 units (n = 6). More detailed analysis of the pH_o dependence of mAe1 E699Q-mediated ³⁵SO₄²⁻_i/Cl⁻_o exchange in the setting of acidic pH_i (Fig. 2C) revealed a pH_o(50) value (Fig. 2D) of 6.53 ± 0.13 (n = 10, r² = 0.99) with parameter d of Eq. 1 as a variable, and 6.43 ± 0.09 (r² = 0.96) with parameter d = 0. The alkaline-shift in pH_o(50) value associated with the more acidic pH_i of MES-injected oocytes (Fig. 2D) was significant compared with that in HEPES-injected oocytes only when modeled with parameter d = 0 (Fig. 1D, dashed line, P < 0.0005), but not when d was allowed to vary (the better fit solid line of Fig. 1D, P = 0.15). Thus, the present data suggest that decreasing pH_i ∼0.45 units does not significantly alter pH_o dependence of mAe1 E699Q-mediated ³⁵SO₄²⁻_i/Cl⁻_o exchange.

Fig. 2.

Inhibition of mAe1 E699Q-mediated SO₄²⁻_i/Cl⁻_o exchange by acidic pH_o seems insensitive to modulation by intracellular pH (pH_i). A: ³⁵SO₄²⁻ efflux time courses of representative individual oocytes previously injected with water (top trace) or with cRNA encoding mAe1 E699Q (lower traces). After oocytes were injected with 50 nl 130 mM MES pH 5.0 containing carrier-free ³⁵SO₄²⁻, they were exposed sequentially to Cl⁻-containing solutions of pH_o 5.0 and 7.4, and then to 100 μM DIDS. B: mean normalized efflux rate constants for (n) oocytes studied as depicted in A. C: ³⁵SO₄²⁻ efflux time courses of representative individual oocytes previously injected with water (top trace) or with cRNA encoding mAe1 E699Q (bottom traces), injected with ³⁵SO₄²⁻ as in A, and exposed sequentially to solutions of the indicated pH values, followed by 100 μM DIDS. D: plot of efflux rate constant vs. pH_o (n = 10) fit to Eq. 1 with parameter d = 0 (dashed line) or treated as a variable (solid line).

Acidic pH_o inhibits mAe1 E699Q-mediated SO₄²⁻/SO₄²⁻ exchange.

Since mAe1 E699Q-mediated SO₄²⁻_i/Cl⁻_o exchange is electrogenic but SO₄²⁻/SO₄²⁻ exchange is electroneutral (7, 15), we investigated the effects of changing pH_o and pH_i on SO₄²⁻/SO₄²⁻ exchange rates. As shown in Fig. 3, A and B, upon injection of HEPES, pH 7.4, transition from pH_o 5.0 to 8.0 led to 2.4 ± 0.1-fold stimulation of SO₄²⁻/SO₄²⁻ exchange. This stimulation was completely reversible upon changing pH_o from 8.0 to 5.0 (n = 6, not shown). SO₄²⁻/SO₄²⁻ exchange by oocytes injected with MES, pH 5.0 (estimated intracellular concentration 13 mM) was also stimulated upon pH_o alkalinization from 5.0 to 8.0 (Fig. 3, C and D). Oocytes injected with unbuffered carrier-free ³⁵SO₄²⁻ similarly exhibited an eightfold stimulation of SO₄²⁻/SO₄²⁻ exchange upon pH_o transition from 5.0 to 8.0 (n = 9, not shown).

Fig. 3.

mAe1 E699Q-mediated SO₄²⁻_i/SO₄²⁻_o exchange is inhibited by acidic pH_o. A: ³⁵SO₄²⁻ efflux time courses of representative individual oocytes previously injected with water (top trace) or with cRNA encoding mAe1 E699Q (bottom traces). After oocytes were injected with 50 nl 130 mM HEPES pH 7.4 containing carrier-free ³⁵SO₄²⁻, each was exposed sequentially to Cl⁻-free solutions containing 20 mM SO₄²⁻ at pH_o 5.0 and 8.0, followed by DIDS (100 μM). B: mean normalized efflux rate constants for the experiments in A. C: ³⁵SO₄²⁻ efflux time courses of representative individual oocytes previously injected with water (top trace) or with cRNA encoding mAe1 E699Q (bottom traces). After oocytes were injected with 50 nl 130 mM MES pH 5.0 containing carrier-free ³⁵SO₄²⁻, each was exposed sequentially to Cl⁻-free solutions containing 20 mM SO₄²⁻ at pH_o 5.0 and 8.0, followed by 100 μM DIDS. D: mean normalized efflux rate constants for the experiments in C.

As shown in Fig. 4 in oocytes containing only endogenous sulfate (∼1 mM), intracellular acidification by ∼0.5 pH units at constant pH_o 7.4 did not alter rates of mAe1 E699Q-mediated SO₄²⁻/SO₄²⁻ exchange or of SO₄²⁻_i/Cl⁻_o exchange, whether acidification was imposed by exposure to 40 mM sodium butyrate or to 20 mM NH₄Cl (12, 40). Exposure to 20 mM KCl to assess the effect of moderate membrane depolarization without acidification also had no effect (n = 4, not shown).

Fig. 4.

At constant pH_o 7.4, mAe1 E699Q-mediated SO₄²⁻_i/Cl⁻_o exchange and SO₄²⁻_i/SO₄²⁻_o exchange are not inhibited by acidic pH_i. A: ³⁵SO₄²⁻ efflux time courses of representative individual oocytes previously injected with water (○, Cl⁻ bath) or with mAe1 E699Q cRNA (bottom two traces). Oocytes preequilibrated with 40 mM butyrate in Cl⁻-free conditions were exposed to baths containing either Cl⁻ (ND-96, ▿) or SO₄²⁻ (20 mM, •), and subjected to subsequent butyrate removal and 200 μM DIDS exposure. B: mean efflux rate constants for SO₄²⁻/Cl⁻ exchange and SO₄²⁻/SO₄²⁻ exchange in the presence (black bars) and absence (gray bars) of butyrate, measured as depicted in A. C: ³⁵SO₄²⁻ efflux time course of a representative individual mAe1 E699Q-expressing oocyte exposed to the absence and subsequent presence of NH₄Cl, followed by 200 μM DIDS. D: mean efflux rate constants for SO₄²⁻/Cl⁻ exchange in the absence (black bars) and presence (gray bars) of NH₄Cl, measured as depicted in C.

Thus, we have shown that both SO₄²⁻/SO₄²⁻ exchange and SO₄²⁻_i/Cl⁻_o exchange mediated by mAe1 E699Q were acutely regulated by changes in pH_o at near constant pH_i, but not by changes in pH_i at constant pH_o.

hAE1 E681OH-mediated SO₄²⁻/Cl⁻ exchange is also inhibited by acidic pH_o.

To test the possibility that the reversed pH dependence of mAe1 E699Q-mediated SO₄²⁻ transport might be a cell type-specific property of the Xenopus oocyte expression system, we tested pH_o dependence of hAE1 E681OH-mediated ³⁵SO₄²⁻ efflux from intact erythrocytes. As shown in Fig. 5, acidic pH_o inhibited ³⁵SO₄²⁻ efflux from WRK-BH₄-treated human red blood cells, with pH_o(50) value ∼5.6 tested at [Cl⁻]_o of 60 and 120 mM. In these conditions, red blood cell pH_i is expected to be close to neutral at neutral pH_o. Thus hAE1 E681OH-mediated SO₄²⁻_i/Cl⁻_o exchange is also inhibited by extracellular H⁺, a pattern opposite to that of wild-type hAE1 in erythrocytes (27).

Fig. 5.

Inhibition of hAE1 E681OH-mediated SO₄²⁻_i/Cl⁻_o exchange by acidic pH_o in human erythrocytes. Human erythrocytes treated with Woodward's reagent K (WRK) and BH₄⁻ undergo ∼80% conversion of hAE1 E681 (equivalent to mouse AE1 E699) to the corresponding alcohol (AE1 E681OH). ³⁵SO₄²⁻ efflux rate constants from WRK/BH₄-treated human erythrocytes were measured in media buffered at the indicated pH_o containing 120 mM KCl (•) or 60 mM KCl, 120 mM sucrose (▴). Indicated pH_o values were as measured in the efflux supernatants immediately after the final efflux time point. The curves represent fits with Eq. 1 with d = 0.

Elevated [SO₄²⁻]_i greatly increases the K_1/2 for Cl⁻_o of mAe1 E699Q-mediated SO₄²⁻_i/Cl⁻_o exchange.

Some aspects of hAE1 E681OH kinetics do not conform to the predictions of ping-pong kinetics (14). We therefore tested the effect of [SO₄²⁻]_i on the apparent K_1/2 for the extracellular substrate Cl⁻ of mAe1 E699Q. At 1 mM nominal [SO₄²⁻]_i (Fig. 6, A and B), ³⁵SO₄²⁻ efflux from oocytes injected with HEPES pH 7.4 exhibited a hyperbolic Cl⁻_o dependence of SO₄²⁻_i/Cl⁻_o exchange, with a K_1/2 for Cl⁻_o of 3.3 ± 0.1 mM (n = 5). Oocytes with nominal [SO₄²⁻]_i of 14 mM (Fig. 6, C and D) exhibited a similarly hyperbolic Cl⁻_o dependence of SO₄²⁻_i/Cl⁻ exchange, but with a ∼10-fold higher K_1/2 for Cl⁻_o of 40 ± 10 mM (n = 5). This increase of extracellular K_1/2 in response to elevation of intracellular substrate concentration is consistent with a ping-pong mechanism of SO₄²⁻_i/Cl⁻_o exchange (11).

Fig. 6.

[SO₄²⁻]_i regulates K_1/2 for Cl⁻_o in mAe1 E699Q-mediated SO₄²⁻_i/Cl⁻_o exchange. A: ³⁵SO₄²⁻ efflux time courses of representative individual oocytes (estimated [SO₄²⁻]_i 1 mM) previously injected with water (top trace) or with mAe1 E699Q cRNA (bottom traces) during sequential exposures to progressively increased [Cl⁻]_o, followed by 100 μM DIDS. B: plot of efflux rate constant vs. [Cl⁻]_o for an individual oocyte measured as depicted in A, with Michaelis-Menten fit to data. C: ³⁵SO₄²⁻ efflux time courses of representative individual oocytes (estimated [SO₄²⁻]_i = 14 mM) previously injected with water (top trace) or with mAe1 E699Q cRNA (bottom traces) during sequential exposure to progressively increased [Cl⁻]_o. D: plot of efflux rate constant vs. [Cl⁻]_o for an individual oocyte measured as depicted in C, with Michaelis-Menten fit to data.

Elevated [SO₄²⁻]_i increases the K_1/2 for SO₄²⁻_o of mAe1 E699Q-mediated SO₄²⁻/SO₄²⁻ exchange, but it does not greatly modify the enhanced affinity of the autoinhibitory “site” for SO₄²⁻_o.

The extent of similarity between mechanisms of electrogenic SO₄²⁻_i/Cl⁻_o exchange and electroneutral SO₄²⁻/SO₄²⁻ exchange mediated by mAe1 E699Q remains unknown. We therefore tested the effect of [SO₄²⁻]_i on the apparent K_1/2 for extracellular SO₄²⁻_o. At 1 mM nominal [SO₄²⁻]_i (Fig. 7, A and B), ³⁵SO₄²⁻ efflux from oocytes injected with HEPES pH 7.4 exhibited a hyperbolic [SO₄²⁻]₀ dependence out to a concentration of 20 mM, but it was inhibited by further elevation of [SO₄²⁻]_o to 64 mM (Fig. 7B, inset). A hyperbolic fit excluding the 64 mM efflux data indicated a SO₄²⁻/SO₄²⁻ exchange K_1/2 of 1.29 ± 0.33 mM for SO₄²⁻_o (n = 9). This value is somewhat larger than that of 0.3 mM observed for SO₄²⁻/SO₄²⁻ exchange in ghosts from E681OH human red blood cells (14). The SO₄²⁻/SO₄²⁻ exchange rate measured at 64 mM [SO₄²⁻]_o was 44 ± 7% (at [SO₄²⁻]_i ∼1 mM, n = 4) lower than the rates at 20 mM [SO₄²⁻]_o. Such apparent autoinhibition was not evident for SO₄²⁻_i/Cl⁻_o exchange for [Cl⁻]_o as high as 103 mM.

Oocytes with nominal [SO₄²⁻]_i of 14 mM (Fig. 7, C and D) exhibited a similarly hyperbolic Cl⁻_o dependence of SO₄²⁻/SO₄²⁻ exchange, but with a ∼2.7-fold higher K_1/2 for SO₄²⁻_o of 3.4 ± 0.68 mM (n = 8). Thus elevation of [SO₄²⁻]_i increases the K_1/2 of mAe1 E699Q for extracellular SO₄²⁻ (P < 0.02), but to a lesser degree than for extracellular Cl⁻. But the increases in both values are qualitatively consistent with a ping-pong mechanism for both electrogenic and electroneutral anion exchange by mAE1 E699Q. Measurements of SO₄²⁻/SO₄²⁻ exchange from similar oocytes with nominal [SO₄²⁻]_i of 14 mM during the transition from 20 to 64 mM [SO₄²⁻]_o showed apparent autoinhibition of 39 ± 7% (n = 12). Thus, changes in [SO₄²⁻]_i did not detectably alter the apparent affinity of the autoinhibitory site for SO₄²⁻_o.

Elevated [SO₄²⁻]_i severely attenuates inhibition by pH_o of both SO₄²⁻_i/Cl⁻_o exchange and SO₄²⁻/SO₄²⁻ exchange mediated by mAe1 E699Q.

In mAe1 E699Q-expressing oocytes containing 14 mM [SO₄²⁻]_i and 13 mM HEPES, pH 7.4, stimulation of SO₄²⁻_i/Cl⁻_o exchange by the pH_o transition from 5.0 to 7.4 (Fig. 8A) was reduced to 1.7 ± 0.3-fold (n = 17; Fig. 8B). In mAe1 E699Q-expressing oocytes containing 14 mM [SO₄²⁻]_i and 13 mM MES, pH 5.0, stimulation of SO₄²⁻_i/Cl⁻_o exchange by the pH_o 5.0–7.4 transition (Fig. 8C) was reduced further to 1.2 ± 0.2-fold (n = 9, Fig. 8D). In mAe1 E699Q-expressing oocytes containing 1 mM [SO₄²⁻]_i without injected buffer, SO₄²⁻/SO₄²⁻ exchange was stimulated eightfold by the pH_o 5.0-to-8.0 transition, but only 1.2 ± 0.4-fold in oocytes containing 14 mM [SO₄²⁻]_i (Fig. 9). Thus, elevated [SO₄²⁻]_i attenuated the pH_o sensitivity of both SO₄²⁻/SO₄²⁻ exchange and SO₄²⁻_i/Cl⁻_o exchange. In separate experiments, elevation of oocyte [SO₄²⁻]_i to ∼26 mM similarly abolished inhibition of both SO₄²⁻/SO₄²⁻ and SO₄²⁻_i/Cl⁻_o exchange by acidic pH_o (n = 5–7, not shown).

Fig. 9.

Elevated [SO₄²⁻]_i attenuates inhibition by acidic pH_o of mAe1 E699Q-mediated SO₄²⁻/SO₄²⁻ exchange. A: ³⁵SO₄²⁻ efflux time courses of SO₄²⁻/SO₄²⁻ exchange in representative individual oocytes previously injected with water (□, top trace) or with mAe1 E699Q cRNA [estimated [SO₄²⁻]_i = 1 mM (•) or 14 mM (○)], during sequential exposure to baths of pH_o 5.0 and 8.0, then to 200 μM DIDS. B: mean normalized efflux rate constants at pH_o 5.0 (black bars) and 8.0 (gray bars) for oocytes with estimated [SO₄²⁻]_i of 1 or 14 mM.

Extracellular oxyanion selectivity of mAe1 E699Q.

Human red blood cell anion exchange is characterized by more rapid transport of planar than of nonplanar anions. Red blood cell Cl⁻/anion exchange rates correlate with the vertical atomic dimensions of trigonal and pyramidal anions (8). Because oxyanion selectivity can reveal information about transporter binding pockets, and E699 is believed to constitute the H⁺ binding site for wild-type H⁺/SO₄²⁻ cotransport, we assessed oxyanion selectivity of mAe1 E699Q in the presence of 1 mM SO₄²⁻_i. Figure 10A shows that, as true for wild-type mAE1, sulfamate and methanesulfonate are transported at very slow rates, such as to be operationally impermeant in Xenopus oocyte assays. Rates of SO₄²⁻/SO₄²⁻ and SO₄²⁻_i/SO₃²⁻ _o exchange were nearly equivalent at ∼65–70% the rate of SO₄²⁻_i/Cl⁻_o exchange. SO₄²⁻_i/SeO_o was 65% the rate of SO₄²⁻_i/Cl⁻_o exchange, but SO₄²⁻_i/SeO_o exchange was roughly comparable in rate to SO₄²⁻_i/Cl⁻_o exchange. As true for wild-type hAE1 in red blood cells (8), SO₄²⁻_i/PO_o exchange was much slower than SO₄²⁻_i/PO_o exchange. SO₄²⁻ exchange rates for extracellular NO₃⁻ or NO₂⁻ were nearly equivalent to those for SO₄²⁻_i/Cl⁻_o exchange.

Fig. 10.

Oxyanion specificity of mAe1 E699Q. A: relative ³⁵SO₄²⁻ efflux rate constants (normalized to SO₄²⁻_i/Cl⁻_o exchange) into baths containing the listed anions at pH_o 7.4 or (as indicated) pH_o 8.2. *P < 0.05 compared with Cl⁻ bath at same pH_o (Student's paired two-tailed t-test; gray bars). B: pH_o dependence of ³⁵SO₄²⁻_i/PO_o exchange in oocytes containing ∼1 mM [SO₄²⁻]_i injected with HEPES, pH 7.4, or with MES, pH 5.0.

SO₄²⁻_i/PO_o exchange was activated approximately twofold by extracellular alkalinization of oocytes injected with HEPES, pH 7.4, and nearly threefold in oocytes injected with MES, pH 5.0 (Fig. 10B). Thus, acute inhibition of mAe1 E699Q-mediated ³⁵SO₄²⁻ efflux by pH_o is also a property of exchange with other oxyanions recognized by the extracellular substrate site of this mutant polypeptide.

DISCUSSION

Mouse Ae1 (mAe1) mutant E699Q expressed in Xenopus oocytes exhibits undetectable ³⁶Cl⁻ efflux and greatly accelerated ³⁵SO₄²⁻ efflux, mediated either by electrogenic SO₄²⁻_i/Cl⁻_o exchange or electroneutral SO₄²⁻/SO₄²⁻ exchange (7). These properties resemble characteristics of human red blood cell hAE1 E681OH, produced by chemical modification with WRK and BH₄ reduction (15). Both mAe1 E699Q and hAE1 E681OH mediate transport of SO₄²⁻ without cotransport of H⁺, in contrast to the H⁺/SO₄²⁻ cotransport characteristic of wild-type AE1. Here we report that the AE1 E699Q mutation of mouse AE1 reverses the wild-type pH_o dependence of SO₄²⁻ transport. In contrast to acid pH_o-activated SO₄²⁻ transport by wild-type AE1, SO₄²⁻ efflux from AE1 E699Q-expressing oocytes is inhibited by acid pH_o and stimulated by alkaline pH_o in a manner consistent with a single titratable “site.” This novel pattern of pH_o sensitivity is exhibited by mAe1 E699Q-mediated electrogenic SO₄²⁻_i/Cl⁻_o exchange as well as by electroneutral SO₄²⁻/SO₄²⁻ exchange and is also a property of hAE1 E681OH in chemically modified human red blood cells. Inhibition of SO₄²⁻ efflux from Xenopus oocytes by acid pH_o is severely attenuated by increased [SO₄²⁻]_i.

Thus, the E699Q mutation unmasks a novel pH_o-regulatory site on mAE1 that can be transregulated by intracellular substrate SO₄²⁻. We also show for mAe1 E699Q that elevated [SO₄²⁻]_i increases apparent K_1/2 ∼10-fold for Cl⁻_o and ∼2.5-fold for SO₄²⁻_o, consistent with sequential (ping-pong) mechanisms for both SO₄²⁻_i/Cl⁻_o exchange and for SO₄²⁻/SO₄²⁻ exchange. In addition, we document in mAe1 E699Q enhanced SO₄²⁻ affinity of an autoinhibitory site for SO₄²⁻.

Opposite pH sensitivities of SO₄²⁻/anion exchange by wild-type AE1 and mAE1 E699Q or hAE1 E681OH.

Cotransport of H⁺ and SO₄²⁻_o in exchange for Cl⁻_i by wild-type hAE1 in human red blood cells is activated by pH, with pK_a values of 5.5 (27, 28). However, whereas SO₄²⁻/SO₄²⁻ exchange in human (27, 36) and in mouse red blood cells (31) peaked at pH_o values between 5.9 and 6.5, with progressive inhibition upon further extracellular acidification, Cl⁻_i/SO₄²⁻_o exchange continued to increase upon further acidification until peaking at pH 4.5–4.0, with transport rates 20-fold higher than peak rates of SO₄²⁻/SO₄²⁻ exchange (26). In contrast, both SO₄²⁻_i/Cl⁻_o exchange and SO₄²⁻/SO₄²⁻ exchange mediated by mAe1 E699Q are inhibited by extracellular H⁺, with peak transport rates observed at values ∼≥pH_o 8. Sulfate transport by hAE1 E681OH in WRK/BH₄-treated human red blood cells exhibited similar inhibition by extracellular protons, with pH_o(50) ∼5.6 for SO₄²⁻_i/Cl⁻ exchange (Fig. 5) similar to the pH_o(50) value for mouse AE1 E699Q-mediated SO₄²⁻_i/Cl⁻_o exchange (Fig. 1D) as fit by Eq. 1 with parameter d = 0, but lower than the pH_o(50) value of 6.2 predicted from the better curve fit (Fig. 1D). Thus, the reversed pH dependence of mAe1 E699Q is not a function of the expression system or of the E-to-Q mutation, but rather results from replacement of the catalytically important carboxylate of E699/E681 with polar amide or alcohol residues,

The H⁺ binding site for H⁺/SO₄²⁻ cotransport in wild-type AE1 is likely E681 in the human protein (15) and E699 in the mouse protein (7). The identity of the protonation site mediating inhibition of SO₄²⁻ transport by mAe1 E699Q and by hAE1 E681OH is unknown, but it may be the inhibitory protonation site for wild-type AE1 SO₄²⁻/SO₄²⁻ exchange, apparently inactive or inaccessible to protonation in the presence of Cl⁻_o, and unmasked by mutation of E699/E681. Existence of the site was predicted from reversal of carrier asymmetry in hAE1 E681OH (15) and may relate to the inhibitory titratable side chain carboxylate modifiable by carbodiimides (5, 6). Similarly uncertain is a possible relationship to any among the conserved amino acid residues of mAe2 whose titration contribute to its H⁺-sensitive inhibition of monovalent anion transport (40–43, 45, 46). Intramolecular cross-linking of hAE1 with the cross-linker bis(sulfosuccinimidyl)-suberate (BS3) applied to intact erythrocytes also reveals an inhibitory protonation site coextant with the wild-type activating protonation site. The inhibitory site may reflect BS3's acid-shifting the pH_o dependence of monovalent anion exchange by 5 pH units such that Cl⁻/Br⁻ exchange, very slow at pH_o >8, is activated to maximal rates at pH 5–6 (18).

The ability of ∼14 mM SO₄²⁻_i to attenuate or prevent inhibition by acidic pH_o of mouse AE1 E699Q-mediated exchange of SO₄²⁻_i for either Cl⁻_o or SO₄²⁻_o might represent a transmembrane conformational effect of either an internal sulfate modifier site or, alternatively, of saturation of the intracellular substrate binding site. In either case, the effect must reflect the increased outward asymmetry of the carrier induced by elevated intracellular substrate. The ability of elevated [SO₄²⁻]_i to attenuate inhibition by acidic pH_o of hAE1 E681OH-mediated SO₄²⁻_i/Cl⁻_o exchange may be less than for mouse AE1 E699Q in oocytes, since red blood cell [SO₄²⁻]_i at the onset of the efflux period approximated 40 mM. This apparent difference between AE1 E681OH and mouse AE1 E699Q may reflect in part the more extensive acid pH_o range tested for WRK-BH₄ red blood cells (Fig. 5), likely accompanied by much larger pH_i excursions in red blood cells than in Xenopus oocytes (13, 41). Alternatively, the difference may reflect distinct SO₄²⁻_i affinities at the intracellular substrate site and/or the putative internal modifier site (21) in the mouse and human proteins.

Proton-mediated inhibition of SO₄²⁻/anion exchange among SLC4 polypeptides.

The Glu residue corresponding to mAe1 E699 and hAE1 E681 is conserved in the Na⁺-independent anion exchangers SLC4A2/AE2 and SLC4A3/AE3, and in the borate transporter, SLC4A11/BTR1. Preliminary results indicate that mutation to Gln of the corresponding mAe2 residue E1007 similarly abolishes Cl⁻/Cl⁻ and Cl⁻_i/HCO₃⁻_o exchange activities, and greatly accelerates activities of SO₄²⁻_i/Cl⁻_o and SO₄²⁻/SO₄²⁻ exchange. mAe2 E1007Q-mediated SO₄²⁻/SO₄²⁻ exchange is also inhibited by acidic pH_o, exhibiting eightfold activation upon pH_o shift from 5.0 to 8.0 (n = 4; Stewart AK and Alper SL, unpublished observations). These results differ from an earlier report that ³⁵SO₄²⁻ uptake by proteoliposomes reconstituted from microsomes of HEK-293 cells transiently transfected with mouse AE2 E1007Q was no slower at pH_o 5.5 than at pH_o 7.5 (38).

A conservative Asp substitution is present in all other SLC4 polypeptides, but sequence polymorphisms at this codon of SLC4 genes have not yet been described in the Single Nucleotide Polymorphism database. Mutation of the corresponding D754 to Cys in rat NBCe1/Slc4a4 abolished Na/HCO₃⁻ cotransport current despite normal oocyte surface expression (25). Mutation of the same Asp residue in human NBCe1 to Glu, Asn, and Arg resulted in Na⁺-dependent base flux at 23%, 9%, and 0% of wild-type rates (1). Possible alteration in substrate selectivity was not evaluated in either study.

Evidence favoring a ping-pong transport mechanism for mAe1 E699Q.

Many characteristics of anion exchange by wild-type human red blood cell AE1 and by hAE1 E681OH can be described by simple ping-pong kinetics with translocation as the rate-limiting step (15, 22). Preservation of ping-pong kinetics in settings of mAe1 E699Q's altered substrate selectivities and substrate-conditional electrogenic anion exchange had not been tested. The 10-fold increase in mAe1 E699Q K_1/2 for Cl⁻_o (Fig. 6) and the smaller increase in K_1/2 for SO₄²⁻_o produced by elevation of [SO₄²⁻]_i from ∼1 mM to ∼14 mM (Fig. 7) are both compatible with a substantial degree of intracellular substrate-driven polypeptide reorientation from inward to outward facing conformations of substrate binding sites. The smaller increase in K_1/2 for SO₄²⁻_o than for Cl⁻_o may reflect the predominance of outward facing sites in SO₄²⁻ medium, the large difference in magnitude of charge transfer during the inward anion translocation step (15), or intrinsic changes in carrier conformation resulting from occupancy of the outward facing substrate site by SO₄²⁻_o or Cl⁻_o. This proposed reorientation from inward to outward conformation was not very much influenced by pH_i (Fig. 9). Unlike hAE1 E681OH red blood cells, the [Cl⁻]_o dependence of ³⁵SO₄²⁻ efflux from oocytes by mAe1 E699Q was not sigmoidal (15) but hyperbolic. This difference may reflect different intrinsic consequences of mutation of the critical Glu residue to the alcohol as opposed to the amide. Alternatively, the difference may relate to the larger fractional contribution of electrogenic SO₄²⁻_i/Cl⁻_o exchange by hAE1 E681OH to total red blood cell conductance than that of mAe1 E699Q to total oocyte conductance.

Substrate selectivity of mAe1 E699Q.

The oxyanion substrate selectivity of mAe1 E699Q shows substantial differences from that of wild-type hAE1. Thus, whereas wild-type hAE1 exchanges Cl⁻_i for SO₄²⁻_o 100-fold more slowly than SO_o at neutral pH_o (8), mAe1 E699Q exchanges SO₄²⁻_i for SO₄²⁻_o and for SO_o at equivalent rates (Fig. 10). Whereas wild-type hAE1 exchanges Cl⁻_i for SeO_o 100-fold more slowly than for SeO_o (8), mAe1 E699Q exchanges SO₄²⁻_i for SeO_o only 35% more slowly than SeO_o. Wild-type hAE1 exchanges Cl⁻_i for PO_o 100-fold more slowly than for PO_o, mAe1 exchanges SO₄²⁻_i for PO_o only 10-fold more slowly than for PO_o. These differences suggest that the transition state in the E699Q mutant can accommodate certain oxyanion substrates larger in the z dimension (8) than tolerated by the wild-type protein, and that this flexibility can be modified by pH_o. The smaller monovalent anions NO₂⁻ and NO₃⁻ are transported by wild-type and mutant proteins at similar rates (50–70% those of Cl⁻_o), whereas the bulky monovalent anions sulfamate and methanesulfonate are transported at negligible rates by both proteins (Fig. 10).

Wild-type SLC4 polypeptides also transport CO₂ equivalents as the divalent oxyanion, CO. Wild-type AE1 in the human red blood cell mediates slow exchange of Cl⁻_i for the extracellular ion pairs lithium carbonate (Li:CO₃)⁻ and lithium sulfite (Li:SO₃)⁻. The even slower exchange of Cl⁻_i for the extracellular Li/oxalate ion pair is more rapid at pH_o 7.5 than at pH_o 6.0 (4), likely reflecting competition for H⁺/oxalate cotransport, with its opposite pH_o dependence (16). Electrogenic Na⁺/HCO₃⁻ cotransporter activity measured in basolateral membrane vesicles from rabbit kidney proximal tubule was postulated to cotransport Na⁺, HCO₃⁻, and CO functioning with 3:1 charge stoichiometry as in the native proximal tubule. SO was found to stimulate HCO₃⁻-dependent Na⁺ uptake, with characteristics suggesting occupancy of a proposed CO site distinct from the HCO₃⁻ site (39). However, recombinant rat kidney NBCe1 expressed in Xenopus oocytes, with its 2:1 charge stoichiometry attributed at first to Na⁺/2HCO₃⁻ cotransport (37) and later reinterpreted as Na/CO cotransport (10), revealed no evidence for either binding or transport of SO(9). Gln substitution of hNBCe1, the residue corresponding to hAE1 E681/mAe1 E699, has not been reported (1, 25), and the hNBCe1 mutant D754N was not tested for upregulation of SO transport (1).

Extracellular modifier site of mAe1 E699Q.

The extracellular modifier site of erythroid hAE1 is believed not to block access to the substrate site, but rather to slow the rate of the anion transporting conformational change. The modifier site likely has no physiological significance in the red blood cell, since Cl⁻ and HCO₃⁻ at their physiological systemic plasma concentrations interact with it only weakly. Model-dependent estimates of modifier site K_1/2 for Cl⁻_o range from 270 to 375 mM (20). However, the modifier site on kidney AE1 of the type A intercalated cell might gain importance under antidiuretic conditions, in which renal medullary interstitial [Cl⁻] can be sustained at 300 mM in the human and at 600–1,200 mM or more in rodents. The modifier site on erythroid AE1 might thus exert a significant effect also on red blood cells retarded in their passage through the antidiuretic renal medulla. An internal inhibitory modifier site requiring high concentrations of Cl⁻_i (21) may also become meaningful in these high ionic strength conditions.

The wild-type hAE1 extracellular modifier site exhibits highest affinity for I⁻ [although K_1/2 estimates range between 8 and 120 mM (18, 23)], with lesser affinity for Br⁻. The binding sites of the noncompetitive anion transport inhibitors NAP-taurine and NIP-taurine overlap the extracellular modifier site (23), and BS3 treatment attenuates or abolishes binding of I⁻ or NAP-taurine to the modifier site (18). SO₄²⁻ interaction with the modifier site has been little studied. However, no self-inhibition by SO₄²⁻ of human red blood cell anion exchange has been noted at concentrations up to 100 mM. Thus, the ∼40% inhibition of mAe1 E699Q-mediated SO₄²⁻/SO₄²⁻ exchange upon increasing [SO₄²⁻]_o from 20 to 64 mM represents a change in both the anion selectivity as well as apparent affinity of the external modifier site, if not the emergence of a novel modifier site. The existence of a similar modifier site in mAe1 E681OH, preferring occupancy with SO₄²⁻ over Cl⁻, was proposed to explain the transacceleration of SO₄²⁻_i efflux by Cl⁻_o to a degree higher than explicable by simple ping-pong kinetics (15). The interaction of SO₄²⁻ with a similar modifier site in hAE1 E681OH may mediate the cis-stimulation of SO₄²⁻_o uptake into WRK-BH₄-modified human red blood cells by 10–20 mM Cl⁻_o, possibly via displacement of SO₄²⁻_o from the modifier site by noninhibitory Cl⁻_o, but more likely via Cl⁻_o-accelerated post-outward translocation release of SO₄²⁻ from hAE1 E681OH (14).

The relationship of the inhibitory substrate modifier site of wild-type hAE1 to the inhibitory protonation site of mAe1 E699Q is not clear. Similarly unknown is the modifier site's relationship with either the second Cl⁻ binding site of hAE1 E681OH (14, 34), the second stilbene binding site of hAE1 E681OH (33), or the distinct binding sites for HCO₃⁻ and CO postulated in some models of NBCe1 activity (10, 29, 39). Nonetheless, the presumed structural similarities among homologous SLC4 polypeptides, most prominent in their transmembrane domains, encourages consideration and testing of these possibilities.

GRANTS

This work was supported by NIH Grants DK-43495 (to S. L. Alper) and GM-026861 (to M. L. Jennings).