Extracellular Cl⁻ regulates human SO₄²⁻/anion exchanger SLC26A1 by altering pH-sensitivity of anion transport

Abstract

Genetic deficiency of the SLC26A1 anion exchanger in mice is known to be associated with hyposulfatemia and hyperoxaluria with nephrolithiasis, but many aspects of human SLC26A1 function remain to be explored. We report here the functional characterization of human SLC26A1, a DIDS-sensitive, electroneutral sodium-independent anion exchanger transporting sulfate, oxalate, bicarbonate, thiosulfate and (with divergent properties) chloride. Human SLC26A1-mediated anion exchange differs from that of its rodent orthologs in its stimulation by alkaline pH_o and inhibition by acidic pH_o but not pH_i, and in its failure to transport glyoxylate. SLC26A1-mediated transport of sulfate and oxalate is highly dependent on allosteric activation by extracellular chloride or nonsubstrate anions. Extracellular chloride stimulates apparent V_max of human SLC26A1-mediated sulfate uptake by conferring a two-log decrease in sensitivity to inhibition by extracellular protons, without changing transporter affinity for extracellular sulfate. In contrast to SLC26A1-mediated sulfate transport, SLC26A1-associated chloride transport is activated by acid pH_o, shows reduced sensitivity to DIDS, and exhibits cation-dependence of its DIDS-insensitive component. Human SLC26A1 resembles SLC26 paralogs in its inhibition by phorbol ester activation of PKC, which differs in its undiminished polypeptide abundance at or near the oocyte surface. Mutation of SLC26A1 residues corresponding to candidate anion binding site-associated residues in avian SLC26A5/prestin altered anion transport in patterns resembling those of prestin. However, rare SLC26A1 polymorphic variants from a patient with renal Fanconi Syndrome and from a patient with nephrolithiasis/calcinosis exhibited no loss-of-function phenotypes consistent with disease pathogenesis.

Introduction

The highly conserved human SLC26 gene family includes 11 members encoding at least 10 polypeptide gene products that mediate electroneutral or electrogenic anion transport in many cell and tissue types [15]. The SLC26 proteins are N-glycosylated [44] homodimers [14] (or tetramers [5,78]) with short N-terminal cytoplasmic domains followed by 14 transmembrane spans and C-terminal cytoplasmic STAS domains of variable length [3,24]. Mammalian SLC26 proteins transport small anion substrates including Cl⁻, HCO₃⁻, OH⁻, SO₄²⁻, oxalate, I⁻, and formate [56], whereas SLC26-related SulP proteins transport dicarboxylates, as well [36,74], playing important roles in maintaining cellular anion homeostasis [15]. Human SLC26 mutations cause Mendelian diseases such as diastrophic (chondro)dysplasia (SLC26A2) [4], congenital chloride diarrhea (SLC26A3) [80], Pendred syndrome or isolated enlargement of the vestibular acqueduct (SLC26A4) [8,79], and impaired male fertility (SLC26A8) [21]. SLC26A9 is a risk modifier gene for cystic fibrosis [9,77], and SLC26A11 polypeptide has been implicated in the neuronal swelling of cerebral edema [66]. In mice, genetically engineered loss of function led, in addition, to outer hair cell deafness (Slc26a5) [45]; oxalate urolithiasis (Slc26a6, Slc26a1) [18,34], and distal renal tubular acidosis (Slc26a7, Slc26a9) [83,84].

SLC26A1/SAT1 cDNA was first cloned from rat liver as a Na⁺-independent SO₄²⁻ transporter [7]. Human SLC26A1 exhibits modest cross-species conservation of amino acid sequence, sharing 78% identity with rat SLC26A1 and 77% with mouse SLC26A1, but only 44% identity with paralogous human SO₄²⁻ transporter SLC26A2 [61]. Previous studies of human and rodent SLC26A1 have demonstrated electroneutral, pH-regulated exchange of anions including SO₄²⁻, oxalate, HCO₃⁻, thiosulfate, and glyoxylate, sensitive to inhibition by 4,4'-diisothiocyano-2,2'-stilbenedisulfonic acid (DIDS) [40-42,61,70,82]. However, the ability of SLC26A1 to transport Cl⁻ remains controversial, with reports both detecting [42,61] and failing to detect Cl⁻ transport [41,51,82]. An unusual property of rodent SLC26A1-mediated SO₄²⁻ uptake measured in Xenopus oocytes is cis-stimulation by extracellular Cl⁻[69], as well as by other non-substrate anions, including Br⁻, I⁻, formate, and lactate [82]. The mechanism of this cis-stimulation has remained unclear.

SO₄²⁻ is an essential nutrient for development [16], biosynthetic sulfation [28]and hepatic sulfate conjugation reactions [25]. The widely expressed SLC26A1 is particularly abundant in sinusoidal membranes of hepatocytes and in basolateral membranes of intestinal enterocytes and renal proximal tubular cells, with higher expression in male than female rats [12,37,42,60-62]. The liver is also the main site of oxalate uptake from dietary absorption, oxalate biosynthesis from ascorbate and amino acid metabolism, and oxalate catabolism [53,76]. Hepatic SLC26A1 is the likely sinusoidal exchanger of intracellular oxalate for extracellular SO₄²⁻, enabling SO₄²⁻ uptake and oxalate secretion for eventual excretion by intestine and kidney [11]. Renal glomeruli freely filter both SO₄2− and oxalate. Most filtered SO₄²⁻ is reabsorbed by renal proximal tubular cells, which also secrete oxalate [6,32,38,71]. SLC26A1 in proximal tubular cell basolateral membrane likely mediates basolateral SO₄²⁻ efflux into the pericapillary interstitial space, in exchange for uptake of extracellular oxalate, which can then be secreted across the apical membrane by SLC26A6 [48].

Slc26a1^−/− mice exhibited urolithiasis, hyperoxaluria, hyperoxalemia, hyposulfatemia, hypersulfaturia, and increased susceptibility to hepatotoxicity [18]. However, oxalate secretion by isolated Slc26a1^−/− duodenum was unchanged [39], and Slc26a1 mRNA and protein abundance was unaltered in mice with hyperoxaluria of other causes, questioning the physiological role of SLC26A1 in hyperoxaluria (at least in duodenum) [23]. Human diseases caused by SLC26A1 mutations have not been reported.

In the present study, we further delineate the function and regulation of human SLC26A1 expressed in Xenopus oocytes, and test potential pathogenicity of selected candidate disease-associated missense variants. We report that extracellular Cl⁻ allosterically activates SLC26A1 by increasing SLC26A1 V_max for SO₄²⁻ transport without change in K_1/2 for SO₄²⁻_[o]. The increased V_max reflects a 2 log decrease in sensitivity of SO₄²⁻/anion exchange to inhibition by extracellular protons.

Methods

Materials

Na36Cl and H36Cl were from ICN (Irvine, CA) and Na₂³⁵SO₄ from ARC (Saint Louis, MO). ¹⁴C-oxalate originally from NEN-DuPont (Boston, MA) was a gift from C. Scheid and T. Honeyman (Univ. Mass. Med. Ctr.). Restriction enzymes and T4 DNA ligase were from New England Biolabs (Beverly, MA). EXPAND High-fidelity PCR System was from Roche Diagnostics (Indianapolis, IN). QuikChange II XL site-directed mutagenesis kit was from Agilent Technologies (Lexington, MA). MEGAscript^® T7 Transcription Kit was from Thermo Fisher Scientific (Waltham, MA). 4,4'-Diisothiocyanostilbene-2,2'-disulfonic acid (DIDS) was from Calbiochem (La Jolla, CA). Phorbol-12-myristate-13-acetate (PMA) was from LC Laboratories (Woburn, MA). Niflumic acid (NFA), glyoxylate, dibutyryl cyclic AMP sodium salt (dB-cAMP), 3-Isobutyl-1-methylanxthine (IBMX), tenidap and 1,2-Bis(2-aminophenoxy)ethane -N,N,N’,N’-tetraacetic acid tetrakis (acetoxymethyl ester) (BAPTA-AM) were from Santa Cruz Biotechnology (Dallas, TX). Hydrochlorothiazide (HCTZ) and R-(+)-[(2-n-Butyl-6,7-dichloro-2-cyclopentyl-2,3-dihydro-1-oxo-1H-inden-5-yl)oxy]-acetic acid (DIOA) were from Sigma (St. Louis, MO). DMSO Stock solutions of these drugs were stored at −20 °C. All other chemical reagents were from Sigma or Fluka (Milwaukee, WI) and were of reagent grade.

Solutions

Modified Barth's saline (MBS) consisted of (in mM) 85 NaCl, 1 KCl, 2.4 NaHCO₃, 0.82 MgSO₄, 0.33 Ca(NO3)₂, 0.41 CaCl₂, and 10 HEPES (pH adjusted to 7.40 with NaOH). ND96(Cl⁻) (referred to in cation substitution experiments as ND96(Na+)) consisted of (in mM) 96 NaCl, 2 KCl, 1.8 CaCl₂, 1 MgCl₂, and 5 HEPES, adjusted to 7.40. In Cl⁻-free ND96 or partial Cl⁻ substitution solutions, NaCl was replaced mole-for-mole with other sodium salts (substituted anion in parentheses: e.g. ND96(gluconate)) or with mannitol. In all Cl⁻-free ND96 solutions, KCl, CaCl₂, and MgCl₂ were substituted with the corresponding gluconate salts. In Na⁺-free ND96, NaCl was replaced by mole-for-mole with KCl or N-methyl-D-glucamine (NMDG-Cl; described with the substituted cation in parentheses: e.g. ND96(K⁺)). Oxalate-containing bath solutions were nominally free of Ca²⁺ and Mg²⁺. Addition of sodium butyrate (40 mM) was in mole-for-mole substitution for NaCl. ND96 solutions at pH 5.0 and pH 6.0 were buffered with 5 mM MES. ND96 solutions at pH values of 8.0 through 9.5 were buffered with 5 mM tricine.

Construction and mutagenesis of cDNA expression plasmids

Human SLC26A1 (NM_022042) was subcloned into the Xenopus oocyte expression vector pXT7. pXT7-SLC26A1 with a fused N-terminal HA-tag was constructed by PCR mutagenesis and ligation. SLC26A1 mutations C41W, A56T, M132T, E295Q, A400C, K401E, and Q556R were generated by QuikChange II XL site-directed mutagenesis kit using specific mutagenic oligonucleotides. (See Table S1 for oligonucleotide primer sequences).

Expression of cRNA in Xenopus oocytes

SLC26A1 cRNA and Xenopus oocytes were prepared as previously described [31]. Oocytes injected with 50 nl cRNA (1-40 ng) and uninjected oocytes were maintained at 17.5°C in MBS containing gentamicin for 2-5 days before use. Control uninjected and control water-injected oocytes exhibited indistinguishable levels of influx of 36Cl⁻, of ³⁵SO₄³⁻, and of 14C-oxalate (data not shown).

Isotopic influx experiments

Oocytes were prewashed (<1 min) in radioisotope-free influx medium to avoid introduction of trace (Cl⁻-containing) MBS to the influx assay. ³⁵SO₄²⁻ influx assays were of 30 min duration (or as indicated) in ND96(Cl⁻) or in Cl⁻-free ND96 bath solutions containing 1 mM (2 μCi) ³⁵SO₄²⁻ (or as indicated) in 150 μl volumes in microtiter plate wells. ³⁶Cl⁻ influx studies of 30 min duration were carried out in ND96(Na⁺), ND96 (K⁺) or ND96(NMDG), or in other bath solutions as indicated. Total bath [Cl⁻] was 103.6 mM (0.25 μCi/well). K³⁶Cl was used in Na⁺-free baths. ¹⁴C-oxalate influx studies were carried out for 30 min periods in nominally Ca²⁺- and Mg²⁺-free ND96(Cl⁻) with 1.0 mM sodium ¹⁴C-oxalate (0.375 μCi/well; 150 μl). Chemicals and drugs were diluted to the indicated concentrations in influx bath solution before uptake experiments. Influx experiments were terminated with 3 washes in cold ND96(cyclamate), followed by oocyte lysis in 150 μl of 2% sodium dodecyl sulfate (SDS). Duplicate 10μl aliquots of influx solution (in 150 μl 2% SDS) were used to calculate specific activities of radiolabeled substrate anions. Oocyte anion uptake was calculated from oocyte-associated counts per minute (cpm) and bath specific activity.

Isotopic efflux experiments

For ³⁵SO₄²⁻ efflux studies, individual oocytes were injected with 50 nl Na₂³⁵SO₄ (5,000~8,000 cpm). After a 5- to 10-min recovery period in ND96 (gluconate), the ³⁵SO₄²⁻ efflux assay was initiated by transfer of individual oocytes to 6 ml borosilicate glass tubes, each containing 1 ml efflux solution. At intervals of 1 or 3 min, 0.95 ml of this efflux solution was removed for scintillation counting and replaced with an equal volume of fresh efflux solution. After completion of the assay with a final efflux period in the presence of the inhibitor 200 μM DIDS, each oocyte was lysed in 150 μl of 2% SDS for counts of residual oocyte-associated radioisotope.

For ³⁶Cl⁻ efflux studies, oocytes were injected with 50 nl of 40 mM Na³⁶Cl (5,000 −8,000 cpm). After 5 min recovery in ND96(gluconate), oocytes were transferred into fresh efflux solutions and the assay was carried out as above. For ¹⁴C-oxalate efflux assays, oocytes were injected with 50 nl of 100 mM Na¹⁴C-oxalate (8,000-10,000 cpm, with final estimated intracellular concentration 10 mM). After a recovery period of at least 15 min, efflux was measured as above.

To vary oocyte pH_i, oocytes were pre-exposed to 40 mM Na butyrate (substituting for NaCl) for 30 min before initiation of an efflux experiment to produce intracellular acidification of 0.5 pH units to pH_i ~6.7-6.8 [75]. Upon removal of bath butyrate (with substitution by NaCl) during the efflux experiment, pH_i alkalinized back toward initial pH_i while pH_o remained constant.

Efflux data were plotted and anion efflux rate constants were calculated as previously reported [31].

Confocal immunofluorescence microscopy

Three or four days after injection with cRNA encoding wild-type or mutant HA-SLC26A1, 10–12 oocytes were fixed with 1.5% paraformaldehyde (PFA) in phosphate-buffered saline (PBS) for 30 min at room temperature and washed three times with PBS containing 0.002% Na azide (PBS-azide). Oocytes were then placed in PBS containing 1% SDS for 15 min to further permeabilize the surface membrane and unmask epitopes. After three subsequent washes in PBS-azide, oocytes were blocked for 1h in PBS containing 1% bovine serum albumin (PBS-BSA), then washed three times with PBS-BSA.

Fixed, permeabilized, and blocked oocytes were incubated overnight at 4°C with rabbit monoclonal anti-HA antibody (Cell Signaling Technologies, Danvers, MA) diluted 1:1000 in PBS-BSA, then washed three times in cold PBS-BSA. Antibody-labeled oocytes were then incubated 2-3 h at room temperature with Cy3-conjugated secondary goat anti-rabbit Ig (diluted 1:500, Jackson ImmunoResearch, West Grove, PA), again thoroughly washed in PBS-BSA, and stored at 4°C until imaging. Cy3-labeled oocytes were imaged using previously reported confocal microscope settings and parameters [31]. Polypeptide abundance at or near the oocyte surface was estimated by Image J quantitation of fluorescence intensity as previously described [31].

Statistics

Data were reported as means ± SE. Flux data were compared by Student’s paired or unpaired two-tailed t tests, or by ANOVA with Tukey post-hoc analysis (SPSS). Some concentration-dependence datasets were fit to the Michaelis-Menten equation using GraphPad Prism.

Results

SLC26A1 mediates SO₄²⁻_[i]/SO₄²⁻_[o] exchange

SLC26A1-mediated ³⁵SO₄²⁻ uptake remained linear for >30 min (Fig. 1a inset) and was maximal in oocytes previously injected with ≥ 5 ng SLC26A1 cRNA (Fig. 1a). In these conditions, SLC26A1-mediated ³⁵SO₄²⁻ influx exhibited a K_1/2 for SO₄²⁻_[o] of 0.26±0.04 mM, with maximal influx at ≥ 1 mM extracellular SO₄²⁻ (SO₄²⁻_[o]) (Fig. 1b). Accordingly, in most of the following ³⁵SO₄²⁻ influx experiments oocytes were injected with 5 ng cRNA, influx bath solutions contained 1 mM SO₄²⁻, and influx time period was 30 min. Efflux of ³⁵SO₄²⁻ from SLC26A1 expressing oocytes increased ~4-fold upon addition of 1 mM SO₄²⁻ to a Cl⁻ bath, and this increment was abrogated by addition of 200 μM DIDS (Fig. 1c,d), consistent with SLC26A1-mediated ³⁵SO₄²⁻_[i]/SO₄²⁻_[o] exchange. Two electrode voltage clamp experiments performed in Cl⁻ bath and in cyclamate bath (Fig. S1) detected no evidence for SLC26A1-mediated electrogenic influx of SO₄²⁻, consistent with previous reports [40].

Fig 1. SLC26A1 mediates SO₄²⁻/SO₄²⁻ exchange.

a) SO₄²⁻ uptake by SLC26A1-expressing oocytes saturates at ≥ 5 ng SLC26A1 cRNA (± SE for n=10). (Inset: Time-dependence of SO₄²⁻ uptake by SLC26A1-expressing oocytes (previously injected with 5 ng cRNA) in ND96(Cl⁻) containing 1 mM SO₄²⁻ (± SE for n=10). b) Oocytes expressing SLC26A1 (previously injected with 5 ng cRNA) mediate SO₄²⁻ uptake (30 min) from ND96(Cl⁻) in a [SO₄²⁻]-dependent manner. c) ³⁵SO₄²⁻ efflux traces of representative individual oocytes previously uninjected (triangles) or injected with 5 ng SLC26A1 cRNA (circles) during sequential exposure to ND96(Cl⁻), to 1 mM SO₄²⁻ in ND96(Cl⁻), and with subsequent addition of 200 μM DIDS. d) Mean ³⁵SO₄²⁻ efflux rate constants (± SE for n oocytes) from experiments as in panel c (**, p<0.01)

Alkaline pH_o stimulates SLC26A1-mediated SO₄²⁻ flux

SLC26A1 was sensitive to pH_o. ³⁵SO₄²⁻ uptake by SLC26A1-expressing oocytes at pH_o 5.0 was lower than at pH_o 7.4, whereas ³⁵SO₄²⁻ uptake at pH_o 8.5 was higher than at pH_o 7.4 (Fig. 2a). Alkaline pH_o-stimulated, DIDS-sensitive ³⁵SO₄^2- efflux from SLC26A1-expressing oocytes was evident also in SO₄²⁻_[o]-free ND96(C^l-) and in ND96(gluconate) bath solutions at pH_o 8.5 (Fig. 2b-e). The Cl⁻_[o]-independence of ³⁵SO₄²⁻ efflux was consistent with exchange of intracellular SO₄²⁻ for extracellular OH⁻. The pH_o dependence of SLC26A1-associated SO₄²⁻_[i]/SO₄²⁻_[o] exchange (measured in Cl⁻ bath) was consistent with a single major protonation site regulating transport (Fig. 2f,g). pH_i did not regulate SLC26A1-mediated ³⁵SO₄²⁻_[i]/SO₄²⁻_[o] exchange in Cl⁻ bath, as judged by the lack of change in efflux during reversible intracellular acidification of ~0.5 units [75] elicited by exposure to and removal of sodium butyrate (Fig. 2h,i). These data together suggested that SLC26A1 can mediate alkaline pH_o-stimulated and acid pH_o−inhibited ³⁵SO₄²⁻/SO₄²⁻ exchange and likely also ³⁵SO₄²⁻[i]^/OH−[o] exchange.

Fig 2. SLC26A1-mediated SO₄²⁻/anion exchange is regulated by pH_o but not pH_i.

a) SLC26A1-mediated SO₄²⁻ uptake from ND96(Cl⁻) containing 1 mM SO₄²⁻ at the indicated bath pH values. b) ³⁵SO₄²⁻ efflux traces of representative individual oocytes previously uninjected (triangles) or injected with 40 ng SLC26A1 cRNA (circles) during sequential exposure to SO₄²⁻-free ND96(Cl⁻) at pH 5.0, pH 8.5 and with subsequent addition of 200 μM DIDS. c) Mean ³⁵SO₄²⁻ efflux rate constants from experiments as in panel b (± SE for n oocytes). d) ³⁵SO₄²⁻ efflux traces of representative individual oocytes previously uninjected (triangles) or injected with 40 ng SLC26A1 cRNA (circles) during sequential exposure to Cl⁻-free, SO₄²⁻-free ND96(gluconate) at pH 5.0, pH 8.5 and with subsequent addition of 200 μM DIDS. e) Mean ³⁵SO₄²⁻ efflux rate constants from experiments as in panel d (± SE for n oocytes). f)35SO₄²⁻ efflux traces of representative individual oocytes previously uninjected (triangles) or injected with 5 ng SLC26A1 cRNA (circles) during sequential exposure to ND96(Cl⁻) containing 1 mM SO₄²⁻ baths of the indicated pH values, with subsequent addition of 200 μM DIDS. g) Mean pH_o-dependent ³⁵SO₄²⁻_[i]/SO₄²⁻_[o] exchange rate constants from experiments as in panel f (± SE for n oocytes). h) ³⁵SO₄²⁻ efflux traces of representative individual oocytes previously uninjected (triangles) or injected with 5 ng SLC26A1 cRNA (circles), preincubated 30 min with 40 mM Na butyrate in ND56(Cl⁻), then exposed sequentially in the presence of 1 mM SO₄²⁻ to ND56(Cl⁻) + 40 mM butyrate, to ND96(Cl⁻) without butyrate, and with subsequent addition of 200 μM DIDS. i) Mean ³⁵SO₄²⁻ efflux rate constants (± SE for n oocytes) from experiments as in panel h (**, p<0.01; ***, p<0.001).

Cl⁻_[o] cis-stimulates SLC26A1-mediated SO₄²⁻ transport

Previous data have suggested cis-stimulation of SLC26A1-mediated ³⁵SO₄²⁻ uptake by halides and other anions [42,69,82], and more extensive evidence for Cl⁻ regulation of SLC26A2 activity has been presented [55,69]. We found that human SLC26A1-mediated ³⁵SO₄²⁻ uptake in Xenopus oocytes was nearly completely dependent upon extracellular anions, as previously noted for mouse [42] and eel SLC26A1 [51]. The [Cl⁻]_o-dependence of ³⁵SO₄²⁻ uptake (measured in the presence of isosmotic mannitol substitution) exhibited a K_1/2 for Cl⁻_[o] of 13.3 mM (R²=0.81; Fig. 3a). Measured in the presence of isosmotic gluconate substitution, ³⁵SO₄²⁻ uptake did not absolutely require Cl⁻, and the [Cl⁻]_o-dependence of the gluconate-stimulated incremental fraction of ³⁵SO₄²⁻ uptake displayed a K_1/2 of 8.33 mM (R²=0.4; Fig. 3b).

Fig 3. Cl⁻_[o] cis-stimulates SLC26A1-mediated SO₄²⁻ transport.

a) SLC26A1-mediated SO₄²⁻ uptake (5 ng cRNA) from bath solutions containing 1 mM SO₄²⁻ and increasing [Cl⁻], with mannitol to balance osmolarity (± SE, n=9-10). b) SLC26A1-mediated SO₄²⁻ uptake (5 ng cRNA) from bath solutions containing 1 mM SO₄²⁻ and increasing [Cl⁻], with gluconate to balance osmolarity (± SE, n=10-20). c) ³⁵SO₄²⁻ efflux traces from representative individual oocytes previously uninjected (triangles) or injected with 5 ng SLC26A1 cRNA (circles) during sequential exposure to baths containing 1 mM SO₄2 first in ND96(gluconate), then in ND96(Cl⁻), with subsequent addition of 200 μM DIDS. d) ³⁵SO₄²⁻ efflux rate constants from experiments as in panel c (mean ± SE for n oocytes). e) ³⁵SO₄²⁻ efflux traces from representative individual oocytes previously uninjected (triangles) or injected with 40 ng SLC26A1 cRNA (circles) during sequential exposure to SO₄²⁻-free baths of ND96(gluconate) and ND96(Cl⁻), with subsequent addition of 200 μM DIDS. f) ³⁵SO₄²⁻ efflux rate constants from experiments as in panel e (mean ± SE for n oocytes). g) SLC26A1-mediated SO₄²⁻ uptake (5 ng cRNA) from ND96(Cl⁻) and ND96(gluconate) solutions containing increasing [SO₄²⁻] as indicated (± SE, n=20). h) SLC26A1-mediated SO₄²⁻ uptake (5ng cRNA) in ND96(Cl⁻) and ND96(gluconate) solutions containing 1 mM SO₄²⁻ at the indicated pH values (± SE, n=15-16;. **, p<0.01).

As shown in Fig. 3c,d, the slow ³⁵SO₄²⁻ efflux from SLC26A1-expressing oocytes into ND96(gluconate) containing 1 mM SO₄²⁻ was greatly accelerated by bath change from gluconate to Cl⁻. In contrast, ³⁵SO₄²⁻ efflux from SLC26A1-expressing oocytes into baths of SO₄²⁻-free ND96(gluconate) or SO₄²⁻-free ND96(Cl⁻) (both pH 7.4) was minimal and DIDS-insensitive (Fig. 3e,f). These results demonstrate that SLC26A1 fails to mediate ³⁵SO₄²⁻_[i]/Cl⁻_[o] exchange, and that the stimulatory effect of Cl⁻_[o] on ³⁵SO₄²⁻_[i]/SO₄²⁻_[o] exchange likely represents allosteric regulation of SLC26A1.

The mechanism by which Cl⁻_[o] exerts cis-stimulation on SLC26A1-mediated ³⁵SO₄²⁻ influx was subjected to more detailed investigation. Since multiple anions can cis-stimulate SLC26A1-mediated SO₄²⁻ transport ([82]; see also Fig. 3b and Fig. 4b) we tested the hypothesis that Cl⁻_[o] might modify SLC26A1 affinity for SO₄²⁻_[o]. However, the SLC26A1 K_1/2 value for [SO₄²⁻] was 0.24 ± 0.029 mM (R²=0.78) as measured in ND96(Cl⁻) and 0.24 ± 0.035 mM (R²=0.68) as measured in ND96(gluconate) (Fig. 3g), demonstrating that Cl⁻_[o] does not regulate the apparent affinity of SLC26A1 for SO₄²⁻_[o]. However, as shown in Fig. 3h, extracellular anions regulated the sensitivity of SLC26A1-mediated ³⁵SO₄²⁻ uptake to extracellular pH. The apparent proton affinity of SLC26A1 (as measured by ³⁵SO₄²⁻ uptake) was left-shifted by approximately two orders of magnitude by ND96(Cl⁻) as compared to uptake in ND96(gluconate) (Fig. 3h). These experiments show that extracellular anions mediate allosteric regulation of the pH_o-sensitivity of SLC26A1-mediated ³⁵SO₄²⁻ influx (likely exchange), and that regulation likely explains the cis-stimulatory effect of Cl⁻_[o] on uptake of ³⁵SO₄²⁻ by SLC26A1.

Fig 4. Cyclamate cis-inhibits SLC26A1-mediated SO₄²⁻ influx.

a) Normalized SLC26A1-mediated SO₄²⁻ uptake (5 ng cRNA) from bath containing 1 mM SO₄²⁻ in the presence of either ND96(Cl⁻) (normalized to 1.0), ND96(gluconate), ND96(isethionate), or ND96(cyclamate). Black bars show uninjected oocytes in ND96(Cl⁻) plus 1 mM SO₄²⁻. b) SLC26A1-mediated SO₄²⁻ uptake (5 ng cRNA) from bath containing 1 mM SO₄²⁻ in ND96(Cl⁻) or in bath containing 56 mM Cl⁻ plus either 96 mM D-mannitol, 48 mM gluconate, or 48 mM cyclamate. Black bars show uninjected oocytes in ND96(Cl⁻) plus 1 mM SO₄²⁻. All fluxes in panels a and b were DIDS-sensitive (not shown). c) ³⁵SO₄²⁻ efflux traces from representative individual uninjected (triangles) and 5 ng SLC26A1 cRNA-injected oocytes (circles) during sequential exposure to bath solutions containing 1 mM SO₄²⁻ in the presence of Cl⁻-free ND96(gluconate), then ND96(cyclamate), with subsequent addition of 200 μM DIDS. d) Mean ³⁵SO₄²⁻ efflux rate constants (± SE for n oocytes) as in experiments of panel c. e) ³⁵SO₄²⁻ efflux traces from representative individual uninjected (triangles) and 5 ng SLC26A1 cRNA-injected oocytes (circles) during sequential exposure to bath solutions containing 1 mM SO₄²⁻ in the presence of ND96(Cl⁻), ND96(cyclamate), then again ND96(Cl⁻), with final addition of 200 μM DIDS. f) Mean ³⁵SO₄²⁻ efflux rate constants of oocytes (± SE for n oocytes) as in experiments of panel e (*, p<0.05; **, p<0.01; ***, p<0.001).

Cyclamate inhibition of SLC26A1-mediated SO₄²⁻ influx is Cl⁻-dependent

SO₄²⁻ uptake from ND96(Cl⁻) was compared to that from a range of Cl⁻-free bath solutions. As shown in Fig. 4a, ³⁵SO₄²⁻ uptake was slightly reduced in ND96(gluconate) and ND96(isethionate), and substantially reduced in ND96(cyclamate). In contrast (not shown), ³⁵SO₄²⁻ uptake in Cl⁻-free bath solutions of sulfamate, methanesulfonate, nitrate and phosphate was comparable in rate to that in ND96(Cl⁻). Moreover, multiple sulfonate ("Good") buffers [26], at 40 mM concentrations failed to reduce SLC26A1-mediated ³⁵SO₄²⁻ uptake (Fig. S2).

The inhibitory effect of cyclamate on SO₄²⁻ transport might reflect direct inhibition at the putative SO₄²⁻ transport site, or reduced cis-stimulation at the hypothetical "cis-stimulatory anion site". Therefore, ³⁵SO₄²⁻ uptake was measured in a 1 mM SO₄²⁻ bath containing constant 56 mM Cl⁻ (producing maximal cis-stimulation; see Fig. 3a) plus osmotically balancing concentrations of D-mannitol, gluconate, or cyclamate. ³⁵SO₄²⁻ uptake was minimally impacted by bath substitution of 48 mM Cl⁻ with 96 mM D-mannitol or 48 mM gluconate, whereas substitution with 48 mM cyclamate decreased ³⁵SO₄²⁻ uptake significantly (Fig. 4b). However, bath change from ND96(gluconate) to ND96(cyclamate) did not inhibit SLC26A1-mediated, DIDS-sensitive ³⁵SO₄²⁻_[i]/SO₄²⁻_[o] exchange in Cl--free bath (Fig. 4c,d). In contrast, ³⁵SO₄²⁻_[i]/SO₄²⁻_[o] exchange in ND96(Cl⁻) bath was strongly and reversibly inhibited by bath cyclamate substitution (Fig. 4e,f), suggesting preferential inhibition by cyclamate of Cl⁻ action at the cis-stimulatory anion site, in addition to weaker binding or reduced agonism of cyclamate compared to Cl⁻ at the putative regulatory site.

SLC26A1 mediates oxalate transport

Oocyte expression of SLC26A1 increased ¹⁴C-oxalate uptake by almost 20-fold compared with that by uninjected oocytes (Fig. 5a). To prevent bath oxalate chelation, the oxalate influx assay was performed in the nominal absence of both Ca²⁺ and Mg²⁺, a condition which did not detectably affect influx of ³⁵SO₄²⁻ from either ND96(Cl⁻) or ND96(gluconate) baths (Fig. S3). Bath addition of 5mM oxalate to ND96(Cl⁻) increased DIDS-sensitive ³⁵SO₄²⁻ efflux from SLC26A1-expressing oocytes 3-fold (Fig. 5b,c), demonstrating ³⁵SO₄²⁻_[i]/oxalate_[o] exchange mediated by SLC26A1. In contrast, however, oocytes previously injected with ¹⁴C-oxalate exhibited no detectable DIDS-sensitive oxalate_[i]/SO₄²⁻_[o] exchange or oxalate_[i]/Cl⁻_[o] exchange, and the component of ¹⁴C-oxalate efflux detected also appeared DIDS-insensitive (Fig. 5d,e).

Fig 5. SLC26A1 mediates oxalate transport.

a) SLC26A1-mediated ¹⁴C-oxalate uptake into oocytes from Ca²⁺- and Mg²⁺-free baths containing 1 mM ¹⁴C-oxalate in ND96(Cl⁻), ND96(gluconate), or ND96(cyclamate) (5 ng cRNA vs. uninjected; n=20, means ± SE). b) ³⁵SO₄²⁻ efflux traces from representative individual oocytes previously uninjected (triangles) or injected with 5 ng SLC26A1 cRNA (circles), during sequential exposure to Ca²⁺ and Mg²⁺-free ND96(Cl⁻), then to 90.5 mM Cl⁻ containing 5 mM oxalate, with subsequent addition of 200 μM DIDS. d) Mean ³⁵SO₄²⁻ efflux rate constants (± SE for n oocytes) from experiments as in panel c. e) ¹⁴C-oxalate efflux traces from representative individual oocytes previously uninjected (triangles) or injected with 40 ng SLC26A1 cRNA (circles) during sequential exposure to ND96(Cl⁻), 64mM ^Na2^SO4 and subsequent addition of 200 μM DIDS. f) Mean 14C-oxalate efflux rate constants (± SE for n oocytes) from experiments as shown in panel e (*, p<0.05; ***, p<0.001).

Oocyte-expressed SLC26A1 mediates SO₄²⁻_[i]^/HCO3⁻_[o] exchange and SO₄²⁻_[i]/ thiosulfate_[o] exchange, but not detectable SO₄²⁻_[i]/glyoxylate_[o] exchange

DIDS-sensitive ³⁵SO₄²⁻ efflux from SLC26A1-expressing oocytes was stimulated by bath addition of either 24mM HCO₃⁻ (Fig. 6a,b) or 5mM thiosulfate (Fig. 6c,d), demonstrating SLC26A1-mediated ³⁵SO₄²⁻_[i]/HCO₃⁻_[o] exchange and ³⁵SO₄²⁻_[i]/thiosulfate_[o] exchange. Glyoxylate was previously identified as a substrate of rat SLC26A1 expressed in Xenopus oocytes [70]. However, 10mM bath glyoxylate failed either to stimulate efflux of ³⁵SO₄²⁻ from SLC26A1-expressing oocytes (Fig. 6e,f) or to cis-inhibit SLC26A1-mediated ⁵³SO₄²⁻ uptake (Fig. 6g). These data suggest that under the conditions tested in the current study, glyoxylate is not a substrate of human SLC26A1.

Fig 6. SLC26A1 mediates ³⁵SO₄²⁻_[i]/HCO₃⁻_[o] exchange and ³⁵SO₄²⁻_[i]/thiosulfate_[o] exchange.

a) ³⁵SO₄²⁻ efflux traces from representative individual oocytes previously uninjected (triangles) or injected with 40 ng SLC26A1 cRNA (circles) during sequential exposure to ND96(Cl⁻), then 80 mM Cl⁻ plus 24 mM HCO₃−/(5% CO₂), with subsequent addition of 200 μM DIDS. b) Mean ³⁵SO₄²⁻ efflux rate constants (± SE for n oocytes) from experiments as in panel a. c) ³⁵SO₄²⁻ efflux traces from representative individual oocytes previously uninjected (triangles) or injected with 5 ng SLC26A1 cRNA (circles) during sequential exposure to ND96(Cl⁻), then 96 mM Cl⁻ plus 5 mM sodium thiosulfate, with subsequent addition of 200 μM DIDS. d) Mean ³⁵SO₄²⁻ efflux rate constants (± SE for n oocytes) from experiments as in panel c. e) ³⁵SO₄²⁻ efflux traces from representative individual oocytes previously uninjected (triangles) or injected with 40 ng SLC26A1 cRNA (circles) during sequential exposure to ND96(Cl⁻), 89 mM Cl⁻ containing 10 mM glyoxylate, and subsequent addition of 200 μM DIDS. f) Mean ³⁵SO₄²⁻ efflux rate constants (± SE for n oocytes) from experiments as in panel e. g) SLC26A1-mediated SO₄²⁻ uptake from baths containing 1 mM SO₄²⁻ in ND96(Cl⁻), in 89 mM Cl⁻ containing 10 mM glyoxylate, and in ND96(Cl⁻) containing 200 μM DIDS; Black column shows uninjected oocytes in ND96(Cl⁻) (± SE , n=10; **, p<0.01; ***, p<0.001).

Acid pH_o stimulates Cl⁻ flux in SLC26A1 expressing oocytes

Cl⁻ has been reported to be a substrate of both mouse and human SLC26A1 polypeptides. We therefore compared pH sensitivity of SLC26A1-associated 36Cl⁻ flux with that of SLC26A1-mediated ³⁵SO₄²⁻ flux. Since oocytes previously injected with 5 ng SLC26A1 cRNA exhibited low Cl⁻ transport rates, these Cl⁻ transport experiments were performed with oocytes previously injected with 40 ng cRNA, and exhibited substantial Cl⁻ influx. However, in contrast to the pH-dependence of SLC26A1-mediated ³⁵SO₄²⁻ influx, 36Cl⁻ influx associated with SLC26A1 was significantly higher at pH_o 5.0 than at more alkaline pH_o values (Fig. 7a). SLC26A1-associated efflux of ³⁶Cl⁻ _[i] was similarly activated by acidic pH_o and comparatively inhibited by more alkaline bath pH_o values, whether in the presence of ND96(Cl⁻) or ND96(gluconate) (Fig. 7b-e), and independent of the order of exposure to acid and alkaline pH_o conditions (Fig. S4). Although the pH_o-dependence of ³⁶Cl⁻ influx is compatible with Cl⁻ _[o]/OH⁻_[i] exchange, the similar pH_o-dependence of Cl⁻ efflux is not compatible with this transport mechanism. Thus, the stimulatory effects of protons on SLC26A1-associated ³⁶Cl⁻ transport may reflect an allosteric regulation by pH that differs from pH_o-dependent regulation of SO₄²⁻ transport.

Fig 7. Effects of pH_o on Cl⁻ flux in SLC26A1-expressing oocytes.

a) Cl⁻ uptake from ND96(Cl⁻) at the indicated bath pH_o values into oocytes (± SE, n=10) previously uninjected (black bars) or injected with 40 ng SLC26A1 cRNA (white bars). b) ³⁶Cl⁻ efflux traces from representative individual oocytes previously uninjected (triangles) or injected with 40 ng SLC26A1 cRNA (circles) during sequential bath exposure to ND96(gluconate) at pH 5.0, then at pH 8.5, and with subsequent addition of 200 μM DIDS. c) Mean ³⁶Cl⁻ efflux rate constants (± SE for n oocytes) from experiments as in panel b. d) ³⁶Cl⁻ efflux traces from representative individual oocytes previously uninjected (triangles) or injected with 40 ng SLC26A1 cRNA (circles) during sequential bath exposure to ND96(Cl⁻) at pH 5.0, then at pH 8.5, and with subsequent addition of 200 μM DIDS. e) Mean ³⁶Cl⁻ efflux rate constants (± SE for n oocytes) from experiments as in panel d (**, p<0.01; ***, p<0.001).

Cl⁻_[o] stimulates Cl⁻_[i] efflux in SLC26A1-expressing oocytes

Our detection of ³⁶Cl⁻ efflux by SLC26A1-expressing oocytes (Fig. 7) prompted further investigation of this Cl⁻ transport process. ³⁶Cl⁻ efflux from these SLC26A1-expressing oocytes (40 ng cRNA) into SO₄²⁻-free ND96(gluconate) was accelerated upon bath change to ND96(Cl⁻) (Fig. 8a,b), consistent with Cl⁻_[i]/Cl⁻_[o] exchange (or with regulatory stimulation by Cl⁻_o of non-exchange Cl⁻ efflux). However, bath change from SO₄²⁻-free ND96(Cl⁻) to nominally Cl⁻-free 64 mM SO₄²⁻ retarded but did not abolish efflux of 36Cl⁻ (Fig. 8c,d), consistent with 36Cl⁻_[i]/SO₄²⁻_[o] exchange at low rates and with reduced DIDS-sensitivity (Figs 7, 8).

Fig 8. SLC26A1 mediates a small component of Cl⁻_[i]/Cl⁻_[o] self-exchange.

a) ³⁶Cl⁻ efflux traces from representative individual oocytes previously uninjected (triangles) or injected with 40 ng SLC26A1 cRNA (circles) during sequential exposure to ND96(gluconate) and ND96(Cl⁻), with subsequent addition of 200 μM DIDS. b) Mean ³⁶Cl⁻ efflux rate constants (± SE for n oocytes) from experiments as in panel a. c) ³⁶Cl⁻ efflux traces from representative individual oocytes previously uninjected (triangles) or injected with 40 ng SLC26A1 cRNA (circles) during sequential exposure to ^{ND96(Cl−), then to 64 mM Na}2^SO4^{, with subsequent addition of 200} μM DIDS. d) Mean ³⁶Cl⁻ efflux rate constants (± SE for n oocytes) from experiments as in panel c (*, p<0.05; ***, p<0.001).

SLC26A1-mediated transport of SO₄²⁻ and Cl⁻ differ in Na+_[o] dependence

Although SLC26A1 mediated substantial rates of Cl⁻ influx (Fig. 7a), neither SO₄²⁻_[i]/Cl⁻_[o] exchange nor Cl⁻_[i]/SO₄²⁻_[o] exchange mediated by SLC26A1 were detected (Fig. 3 and 8). SLC26A1-mediated ³⁵SO₄²⁻ uptake was Na⁺-independent (Fig. 9a), as previously reported [7,69,82]. In contrast, SLC26A1-associated Cl⁻ influx exhibited a complex cation dependence (Fig. 9b). ³⁶Cl⁻ uptake by SLC26A1-expressing oocytes was lower in ND96(NMDG) than in either ND96(Na⁺) or in ND96(K⁺). Moreover, whereas the Na⁺-dependent component of ³⁶Cl⁻ uptake (~45%) was DIDS-sensitive, the remaining, DIDS-insensitive component of uptake was comparable in magnitude to ³⁶Cl⁻ uptake in the absence of Na⁺ (Fig. 9b). This contrasts to the abolition of SLC26A1-mediated SO₄²⁻ uptake by 200 μM DIDS (Fig. 9a), even in oocytes previously injected with 40 ng cRNA (data not shown). SLC26A1-associated ³⁶Cl⁻ influx from ND96(Na⁺) was completely insensitive to 100 μl HCTZ and to 500 μM DIOA, suggesting that neither activation of endogenous oocyte Na⁺-Cl⁻ cotransport nor of oocyte K⁺-Cl⁻cotransport accounted for the observed SLC26A1-associated Na⁺-dependent Cl⁻ flux. SLC26A1-associated ³⁶Cl⁻ flux was also insensitive to 200 μM niflumic acid (NFA, Fig. 9c), in contrast to its partial sensitivity to 200 μM DIDS, and arguing against endogenous NFA-sensitive anion channels as mediators of SLC26A1-associated Cl⁻ transport.

Fig 9. SLC26A1-mediated uptakes of Cl⁻ and SO₄²⁻ differ in cis-Na+-dependence and apparent DIDS-sensitivity.

a) SO₄²⁻ uptake by oocytes (n=10) previously uninjected (black bars) or injected with 5 ng SLC26A1 cRNA (white bars), from baths containing 1 mM SO₄²⁻ in the presence of ND96(Na⁺) or ND96(NMDG). b) Cl⁻ uptake by oocytes (n=20) previously uninjected (black bars) or injected with 40 ng SLC26A1 cRNA (white bars), from baths of ND96(Na⁺), ND96(K⁺), or ND96 (NMDG), in the absence or presence (cross-hatched bars) of 200 μM DIDS. c) Normalized SLC26A1-associated Cl⁻ uptake (40ng cRNA) from ND96 bath in the absence or presence of 500 μM hydrochlorothiazide (HCTZ), 100 μM dihydroindenyloxyalkanoic acid (DIOA), 200 μM niflumic acid (NFA), or 200 μM DIDS (Mean values ± SE for n oocytes are normalized to absence of inhibitor) (*, p<0.05; **, p<0.01; ***, p<0.001).

SLC26A1 is inhibited by PKC

Several SLC26A1 family members are sensitive to PKC [31,63], and SLC26A1 has several potential PKC phosphorylation sites. As shown in Fig. 10a,b, ³⁵SO₄²⁻_[i]/SO₄²⁻_[o] exchange mediated by either SLC26A1 or HA-tagged SLC26A1 was strongly inhibited after oocyte exposure to the classical PKC activator, PMA, whereas DMSO exposure had no effects (data not shown). PKC inhibition of SLC26A1 did not reflect decreased expression at or near the oocyte surface (Fig. 10c), in contrast to PMA-mediated reduction in oocyte surface expression accompanying PMA-medicated inhibition of SLC26A2 [31].

Fig 10. Inhibition of SLC26A1 by phorbol 12-myristate 13-acetate (PMA).

a) ³⁵SO₄²⁻ efflux traces from representative individual oocytes uninjected (triangles) or injected with 5 ng SLC26A1 cRNA (circles) or HA-SLC26A1 cRNA (squares) during sequential exposure to ND96(Cl⁻) containing 1 mM SO₄²⁻, then for 30 min to SO₄²⁻-free ND96(Cl⁻) containing 200 nM PMA, followed by readdition of 1 mM SO₄²⁻ to the same bath, and with subsequent addition of 200 μM DIDS. b) Mean ³⁵SO₄²⁻ efflux rate constants (± SE for n oocytes) from experiments as in panel a. c) Confocal immunofluorescence micrographs of representative oocytes previously uninjected or injected with 5 ng HA-SLC26A1 cRNA, then incubated in 200nM PMA or DMSO vehicle (CTL) as indicated; lower right: normalized HA-SLC26A1 fluorescence intensity (mean ± SE, n=20) at periphery of oocytes treated with PMA or DMSO (*, p<0.05; ***, p<0.001).

Manipulations of cAMP and intracellular Ca²⁺ did not alter SLC26A1 activity

SLC26A1-mediated ³⁵SO₄²⁻ uptake from ND96(Cl⁻) or from ND96(gluconate) was not altered by oocyte treatment with dibutyryl-cAMP(dB-cAMP) and/or isobutylmethylxanthin (IBMX, Fig. S5a). Intracellular Ca²⁺ chelation by prolonged preincubation in BAPTA-AM or by preinjection with Na-EGTA similarly failed to alter SL26A1-mediated ³⁵SO₄²⁻ uptake from either ND96(Cl⁻) or ND96(gluconate) (Fig. S5b).

Mutations of highly conserved amino acid residues reveal functional similarities and differences of SLC26A1

The congenital chloride-losing diarrhea mutation SLC26A3^E293Q and the paralogous engineered mutations SLC26A4^E303Q, SLC26A2^E336Q, and SLC26A6^E298Q exhibited loss-of-function in Xenopus oocytes without decrease in apparent surface expression [31,63]. The corresponding SLC26A1 mutant SLC26A1E295Q also exhibited 50% reduction of ³⁵SO₄²⁻ influx (Fig. 11a), but no reduction of ³⁶Cl⁻ influx (Fig. 11b) or of apparent SLC26A1surface expression.

Fig 11. Functional tests of select mutations in highly conserved amino acid residues of SLC26A1.

a) SO₄²⁻ uptake from ND96(Cl⁻) and from ND96(gluconate) baths containing 1 mM SO₄²⁻ by oocytes (n=10) previously uninjected (black bars) or injected with 5 ng cRNA encoding SLC26A1^WT or the indicated mutants (white bars). b) Normalized Cl⁻ influx by oocytes (n=17-30) previously uninjected (black bar) or injected with 5 ng cRNA encoding SLC26A1^WT or the indicated mutants (white bars). c) Confocal immunofluorescence micrographs of representative oocytes, uninjected or previously injected with 5ng cRNA encoding HA-SLC26A1^WT or the indicated HA-tagged mutant cRNAs. Lower right: normalized HA-SLC26A1 fluorescence intensity at the oocyte periphery (mean ± SE, n=20; ***, p<0.001).

In chicken SLC26A5/prestin, residues S398 and R399 have been proposed as substrate-binding sites, since mutation of these residues strongly disrupted prestin transport function [27]. Expression in oocytes of the corresponding mutant SLC26A1A400C reduced SLC26A1-mediated ³⁵SO₄²⁻ uptake by only 25% in ND96(Cl⁻), and not at all in ND96(gluconate). Remarkably, ³⁶Cl⁻ influx into oocytes expressing mutant SLC26A1^A400C was unaffected. However, the adjacent corresponding mutant SLC26A1^K401E (corresponding to cPrestin R399) exhibited complete loss-of-function in uptake assays of ³⁵SO₄²⁻ and of ³⁶Cl⁻ (Fig. 11a,b). Neither putative anion substrate binding site mutant of human SLC26A1 exhibited reduced polypeptide abundance at or near the oocyte surface (Fig. 11c), suggesting preservation of overall transporter structure despite loss-of-function. HA-tagged SLC26A1 wildtype and mutant polypeptides showed similar results (data not shown).

SLC26A1 variants in the exome of a patient with recessive renal Fanconi syndrome do not confer loss-of-function

Inherited compound heterozygous variants SLC26A1^C41W and SLC26A1^A56T were found in the exome of a single Mexican child with recessive proximal tubular (PT) Fanconi Syndrome (~30x at ~80% coverage) in which mutations in Fanconi Syndrome gene NaPiIIa/SLC34A1 [47], and Fanconi-Bickel syndrome gene GLUT2/SLC5A2 [68] were not found. Both SLC26A1 variants were rare, with dbSNP-reported minor allele frequencies for C41W (rs43974931) of <0.5% and for A56T (rs142573758) of 0.02% . The C41W variant is predicted as "damaging" (SIFT) or "probably damaging" (Polyphen2). The A56T variant is predicted as "damaging" (SIFT) or "benign" (Polyphen2). The above characteristics and the shared basolateral proximal tubular localization of SLC26A1 with GLUT2 led to evaluation of the patient's SLC26A1 variants as candidate Fanconi Syndrome genes. Both ³⁵SO₄²⁻ uptake and ¹⁴C-oxalate uptake mediated by SLC26A1^C41W in ND96(Cl⁻) were normal or slightly elevated (Fig. 12a,b). ³⁵SO₄²⁻ uptake by SLC26A1A56T was also normal, and ¹⁴C-oxalate uptake was minimally decreased (Fig. 12a,b). Interestingly, ³⁶Cl⁻ uptake associated with expression of either variant was substantially increased (Fig. 12c), as was (nominal) surface abundance of variant polypeptide (Fig. 12d,e). However, such a gain-of-function disease mutation would likely show a dominant inheritance pattern, incompatible with the reported good health of the heterozygous parent. In addition, the hypothetical contribution of SLC26A1 A56T to proximal tubular basolateral membrane Cl⁻ permeability [2,59,81] (if indeed electroneutral) should not decrease transepithelial reabsorption of HCO₃⁻, glucose, or phosphate. Thus, the available functional data from expression in Xenopus oocytes do not support SLC26A1 as the renal Fanconi Syndrome disease gene in this family.

Fig 12. Functional tests of candidate pathological missense variants of SLC26A1.

Uptake of SO₄²⁻ (a),14C-oxalate (b) and Cl⁻ (c) by oocytes (n=10-30) previously uninjected (black bars) or injected with 5 ng cRNA (white bars) encoding SLC26A1^WT or compound heterozygous variants SLC26A1^C41W or SLC26A1^A56T found in a renal Fanconi Syndrome exome. d) Confocal immunofluorescence micrographs of representative oocytes previously uninjected or injected with 5ng cRNA encoding HA-SLC26A1^WT, HA-SLC26A1^C41W or HA-SLC26A1^A56T. e) Normalized HA-SLC26A1 fluorescence intensity at oocyte periphery (mean ± SE, n=20; *, p<0.05; **, p<0.01; ***, p<0.001).

SLC26A1 variants reported in a patient with nephrocalcinosis do not exhibit altered function

The rare heterozygous SLC26A1 variant SLC26A1^M132T was found in trans with the homozygous common variant Q566R in a patient with severe nephrocalcinosis requiring nephrectomy [19]. Since residue M132 is highly conserved across species orthologs, and the M132T-encoding variant is absent from dbSNP and predicted to be pathogenic, we subjected SLC26A1^M132T to functional testing in Xenopus oocytes, in isolation and in cis with common variant Q556R. We also coexpressed cRNAs encoding the double mutant with that encoding Q566R alone, to mimic the reported patient genotype. However, neither ³⁵SO₄²⁻ uptake from ND96(Cl⁻) or from ND96(gluconate) nor ¹⁴C-oxalate uptake from ND96(Cl⁻) was reduced in oocytes expressing SLC26A1^M132T in any of the combinations tested (Fig. 13), suggesting that variant SLC26A1^M132T did not cause nephrolithiasis in this patient.

Fig. 13. Functional tests of nephrocalcinosis-associated missense variants of SLC26A1.

a) SO₄²⁻ uptake from ND96(Cl⁻) or ND96(gluconate) by oocytes previously uninjected (black bar) or injected with 5ng cRNA (white bars) encoding SLC26A1^WT or rare variant SLC26A1^M132T, common variant SLC26A1^Q556R, nephrolithiasis-associated double variant SLC26A1^M132T/Q556R, or co-injected with cRNAs (2.5 ng each) encoding SLC26A1^M132T/Q556R + SLC26A1^Q556R (n=10). An independent experiment comparing only SLC26A1^M132T and SLC26A1^WT yielded indistinguishable results (n=20, not shown). b) ¹⁴C-oxalate uptake from ND96(Cl⁻) by oocytes previously uninjected (black bar) or injected (white bars) with 5 ng cRNAs (or 2.5 ng each for two cRNAs) encoding the indicated SLC26A1 polypeptide (n=10).

Discussion

Prompted by our preliminary identification of SLC26A1 as a candidate gene in a case of recessive renal Fanconi Syndrome, we conducted a comprehensive functional characterization of recombinant human SLC26A1 using the Xenopus oocyte heterologous expression system. SLC26A1-mediated ³⁵SO₄²⁻/SO₄²⁻ exchange was DIDS-sensitive, apparently electroneutral, stimulated by alkaline pH_o and inhibited by acid pH_o, but was not regulated by acid pH_i over the pH range tested. SLC26A1-mediated ³⁵SO₄²⁻ uptake was almost completely dependent upon extracellular Cl⁻, with K_1/2 for Cl⁻ cis-stimulation of 8-13 mM. The Cl⁻-stimulatory effect was mediated entirely by increasing V_max for SO₄²⁻ without altering K_1/2. The increased V_max reflected a dramatic, Cl⁻_[o]-dependent decrease in sensitivity of SLC26A1-mediated SO₄²⁻ transport to inhibition by acid pH_o. Cl⁻_[o] was the most effective among other extracellular anions that also conferred cis-stimulation of SLC26A1-mediated SO₄²⁻/SO₄²⁻ exchange.. SLC26A1-associated ³⁶Cl⁻ transport, unlike ³⁵SO₄²⁻ transport, was activated by acid pH_o and partially cation-dependent. The Na⁺-dependent component of Cl⁻ transport was DIDS-sensitive, but insensitive to inhibitors of cation chloride cotransport. Table 1 summarizes the major differences observed between human SLC26A1-mediated ³⁵SO₄²⁻ transport and SLC26A1-associated Cl⁻ transport.

Table 1.

Differences between SLC26A1-mediated SO4²⁻ transport and SLC26A1-associated Cl⁻ transport.

	SO42− transport	Cl− transport
pHo effects	alkaline pHo stimulates SO42− flux	acidic pHo stimulates Cl- flux
OH-/anion exchange	possible SO42−[i]/OH−[o] exchange	no detectable Cl−[i]/OH−[o] exchange
Cl−-SO42− exchange	no detectable SO42−[i]/Cl−[o] exchange	possible Cl-[i]/SO42−[o] exchange
Na+[o] effects	Na+[o]-independent SO42− influx	Na+[o]-dependent component of Cl− influx
DIDS-sensitivity	DIDS-sensitive	Na+[o]-sensitive component of Cl− flux is DIDS-sensitive

SLC26A1 also mediated 14C-oxalate uptake and SO₄²⁻_[i] /oxalate_[o] exchange, but oxalate efflux in exchange for bath SO₄²⁻ (14C-oxalate_[i]/SO42-_[o] exchange) was not detectable. SLC26A1 mediated SO₄²⁻ efflux in exchange for bath HCO₃-(35SO₄²⁻_[i]/HCO₃-_[o] exchange) and for bath thiosulfate (³⁵SO₄²⁻ _[i]/thiosulfate_[o] exchange), but not in exchange for bath glyoxylate.

Mutation of residues corresponding to those crucial for SO₄²⁻ transport by chicken prestin/SLC26A5 altered SLC26A1-mediated ³⁵SO₄²⁻ transport in ways similar, but not identical to those in prestin. However, neither SLC26A1 missense variants found in a complete Fanconi syndrome patient nor different variants previously associated with nephrolithiasis/nephrocalcinosis exhibited loss-of-function phenotypes consistent with pathogenicity.

SLC26A1-mediated SO₄²⁻ transport

Our determination of 0.26 mM for K_1/2 of human SLC26A1-mediated SO₄²⁻ uptake (Fig. 1) is similar to those previously reported for rat (0.14 mM, [7]), and reported as data not shown for mouse (0.31 mM, [42]) and human SLC26A1 (0.19 mM, [61]). The human K_1/2 value is appropriate for a baso-lateral epithelial transporter in the presence of serum SO₄²⁻ concentrations of ~0.25-0.41 mM [50] that can double in late gestation [17] and triple during renal insufficiency [52]. Krick et al [40] previously characterized K_1/2 values for cis-inhibition of rat SLC26A1-mediated SO₄²⁻ uptake by extracellular HCO₃⁻, thiosulfate, sulfite, and oxalate. As we observed for human SLC26A1 (Fig. S1), rat SLC26A1 function appeared electroneutral in Cl⁻ bath in both absence and presence of added SO₄²⁻ [40]. The related SO₄²⁻ transporter SLC26A2 was also shown to mediate electroneutral anion exchange [55].

pH-sensitivity of SLC26A1-mediated SO₄²⁻ transport

Acid pH_o inhibited and alkaline pH_o stimulated both influx and efflux of SO₄²⁻ by human SLC26A1 (Fig. 2). In marked contrast, SO₄²⁻ uptake by mouse SLC26A1 was stimulated by acid pH_o [82], and rat SLC26A1-mediated SO₄²⁻ uptake was inhibited by alkaline pH_o [40]. Human SLC26A1-mediated SO₄²⁻_[i]^/OH-_[o] exchange might account for part of the ³⁵SO₄²⁻ efflux stimulated by alkaline pH_o, but alkaline pH_o−stimulated SO₄²⁻ influx is less easily explained by SO₄²⁻[o]/^OH-_[i] exchange. The qualitatively identical regulatory effects of pH_o on both influx and efflux of SO₄²⁻ together suggest an allosteric regulatory effect of SO₄²⁻ transport. In the absence of pH_i-sensitive SO₄²⁻ transport by human SLC26A1, it is not straightforward to link the transporter's pH_o sensitivity to alkaline pH-stimulated proteoglycan sulfation [20] or to alkaline pH-stimulated metabolic N-sulfation of drugs [33]. We suggest that aa sequence differences in ecto-loops or elsewhere may explain the opposite pH-sensitivites of human and rodent SLC26A1.

Cl⁻_[o] regulates SLC26A1-mediated SO₄²⁻ transport by allosteric control of pH_o sensitivity

Human SLC26A1-mediated influx and efflux of SO₄²⁻ were stimulated greatly by Cl⁻_[o], and to comparable or lesser degrees by other extracellular anions tested (Fig. 4a,b; Supp. Fig. 4). Cl⁻_[o] exhibited cis-stimulation of ³⁵SO₄²⁻ uptake with Cl⁻ K_1/2 of ~8-13 mM (Fig. 3). Cl⁻_[o]-mediated cis-stimulation of SO₄²⁻ uptake has also been noted as a property of SLC26A1 from mouse [82], eel [51], and rat [69]. However, Cl⁻_[o] did not stimulate human SLC26A1-mediated ³⁵SO₄²⁻_[i]/Cl⁻_[o] exchange in the complete absence of SO₄²⁻_[o] (Fig. 3e,f).

Bath Cl⁻ cis-stimulated ³⁵SO₄²⁻ uptake via increased Vmax without effect on K_1/2 for extracellular anion. The cis-stimulation reflected not a true V_max increase but rather a ~2 log shift of the pH_o vs. SO₄²⁻ uptake curve towards acid pH_o by bath Cl⁻ as compared to bath gluconate, and likely represents allosteric regulation by bath anion. The related SO₄²⁻ transporter SLC26A2 is also allosterically regulated by Cl⁻_[o], but by a mechanism (or utilizing a regulatory site of different anion affinity relative to the transport site) that co-exists with (the more frequently observed) cis-inhibition of SO₄²⁻ uptake by Cl⁻_[o]. Stimulation by Cl⁻_[o] of mouse SLC26A2-mediated SO₄²⁻_[i]/OH-_[o] exchange (monitored by intracellular pH electrode) exhibited Cl⁻_[o] ^K1/2 of ~4-8 mM [55], values similar to those reported here for human SLC26A1.

The aa residues constituting the regulatory binding site for Cl⁻ and other anions in SLC26A1 and SLC26A2 remain to be defined. Fluoride ion was reported (as data not shown) to inhibit the allosteric regulatory effect of Cl⁻_[o] on SLC26A2 [55]. Cl⁻_[o] has also been reported to regulate Cl⁻/HCO₃-selectivity of CFTR in Xenopus oocytes [72], either directly via pore-widening [35] or via WNK kinase regulation [58]. The possible influences of these or related pathways on human SLC26A1 are unknown. Intracellular Cl⁻ regulatory sites with the GXXXP motif have been defined in SLC4A4/NBCe1-B [73], but a role for Cl⁻_[i] in regulation of SO₄²⁻ transport by SLC26A1 remains to undefined.

Regulation by Cl⁻_[o] with K_1/2 ~10 mM can be physiologically meaningful for an apical anion exchanger (such as SLC26A2) facing a low-Cl⁻ lumen [29]. However, understanding of such regulation is more challenging for a basolateral transporter such as SLC26A1 in proximal tubule and enterocytes exposed to interstitial fluid, for which a yet undefined anion, active in the presence of saturating Cl⁻, may be a more potent physiological ligand of the regulatory ecto-site. Recently, SLC26A1 has been localized to the apical membrane in maturation phase ameloblasts in developing teeth of rat [85], a context in which physiological regulation by Cl⁻_[o] might occur. Alternatively, the regulatory ecto-binding site might be an evolutionary legacy from ancestral SLC26 genes or the conditions in which they functioned.

Transport of oxalate and other anions

SLC26A1 from mouse [82] and rat [40] was previously shown to take up oxalate. Ethylene glycol-induced hyperoxaluria in female (but not male) rats was accompanied by elevated hepatic and renal levels of SLC26A1 polypeptide, but not mRNA [10]. Human SLC26A1 displayed oxalate influx higher in Cl⁻_[o] than in baths of gluconate or cyclamate. Human SLC26A1 mediated DIDS-sensitive ³⁵SO₄²⁻_[i]/oxalate_[o] exchange, but neither 14C-oxalate_[i]/Cl⁻_[o] exchange nor 14C-oxalate_[i]/SO₄²⁻_[o] exchange was detected (Fig. 5). This preferentially vectorial oxalate transport is consistent with the hypothesis that basolateral oxalate uptake by enterocytes is mediated by SLC26A1 as part of a transepithelial pathway that includes apical oxalate secretion into the gut lumen by SLC26A6 [18,49]. The macroscopically unidirectional transport may reflect a very low affinity of SLC26A1 for intracellular oxalate, or much higher affinities for other coextant substrate anions at the intracellular substrate binding site. However, such a system is not active throughout the gut, as evidenced by normal ex vivo oxalate secretion rates by the isolated SLC26A1^−/− mouse duodenum [39].

Human SLC26A1 exchange of SO₄²⁻_[i] for bath HCO₃- and bath thiosulfate (Fig. 6a-d) resembled that of rat SLC26A1 [40]. Thus, SLC26A1 may play a role in the action of exogenous thiosulfate in treatment or prophylaxis of calciphylaxis [54,57] and possibly also in the metabolism of H₂S, both a precursor of endogenous thiosulfate and a product of exogenous thiosulfate [43]. Unlike rat SLC26A1 [70], recombinant human SLC26A1 did not exhibit glyoxylate transport (Fig. 6e-g), despite the ability of chronic glyoxylate exposure to increase SLC26A1 mRNA and SO₄²⁻ uptake in human HepG2 cells [70]. Thus, our Xenopus oocyte data do not support the otherwise attractive model of hepatic SLC26A1-mediated glyoxylate/oxalate exchange proposed by Schnedler et al. [70].

Transport of Cl⁻ associated with SLC26A1 expression in Xenopus oocytes

Human SLC26A1 mediated robust uptake of Cl⁻ which is increased further by acid pH_o (Fig. 7a), in contrast to inhibition of human SLC26A1-mediated SO₄²⁻ transport by acid pH_o (Fig. 2). SLC26A1-mediated Cl⁻ efflux into Cl⁻ bath or gluconate bath was also increased by acid pH_o, results not consistent with SLC26A1-mediated Cl⁻_[i]/^OH-_[o] exchange. These results also differed from alkaline pH_o stimulation of SO₄²⁻ efflux, consistent with SLC26A1-mediated ³⁵SO₄²⁻_[i]/OH-_[o] exchange. SLC26A1-associated 36Cl⁻ efflux into baths of Cl⁻ or gluconate were of lower rate and lower DIDS-sensitivity than was the case for SLC26A1-mediated SO₄²⁻ transport (Fig. 8). Among other SLC26 polypeptides, acid pH_o also increased SLC26A9-mediated Cl⁻ currents [22,65].

Unlike the Na+-independence of SLC26A1-mediated SO₄²⁻ transport, SLC26A1-associated Cl⁻ flux displayed partial Na⁺-dependence (Fig. 9). NMDG⁺ substitution of bath Na⁺ abolished the DIDS-sensitive component of Cl⁻ flux. SLC26A1-mediated Cl⁻ uptake was insensitive to inhibition by cation-Cl⁻ cotransporter blockers and to niflumate. Although K⁺ substitution of bath Na⁺ did not reduce Cl⁻ uptake, its DIDS-sensitive component was abolished, suggesting that high bath K⁺ activated a DIDS-insensitive, SLC26A1-dependent Cl⁻ uptake. The reported Na⁺-dependent component of SLC26A9-mediated Cl⁻ transport was DIDS-sensitive [46,67] and represented (at least in part) coupled transport of Na⁺ with Cl⁻ [13], albeit insensitive to inhibitors of cation chloride cotransport. The previously reported lack of Cl⁻ transport by mouse SLC26A1 [82] was based on experiments carried out in Na⁺-free, NMDG⁺ solutions. In contrast, the previous report of Cl⁻ transport associated with human SLC26A1 expression [61] reflected experiments carried out in Na gluconate bath.

Protein Kinase C (PKC) regulation of SLC26A1

PKC activation by PMA was first shown to downregulate SLC26A6 activity through decreased cell surface expression, and pharmacological approaches suggested regulation by PKCδ [30]. Documentation of inhibition by PMA-mediated activation of PKC was extended to SLC26A2 [31] and SLC26A3, but SLC26A4 was unaffected by PMA. Activated PKCδ was more definitively implicated as inhibitor of SLC26A2, -A3, and -A6 through attenuation of inhibition by co-expressed kinase-dead PKCδ [64]. Here we have demonstrated PMA-mediated inhibition of SLC26A1 by a mechanism other than regulation of surface expression (Fig. 10). This PMA response may reflect involvement of a different PKC isoform. Further work will determine whether SLC26A1 inhibition by PMA is mediated via direct phosphorylation or requires intermediate signaling steps.

Divergent effects of SLC26A1 mutations on transport of SO₄²⁻ and Cl⁻

The structural fold of paralogous SLC26 proteins accommodates anion exchangers of varied selectivity and mechanism, anion channels, and proteins with apparent plasticity of anion selectivity and mechanism [1,74]. Therefore, the functional consequences of mutation in selected residues previously shown crucial for anion transport activity in other SLC26 family members were tested in SLC26A1. Human SLC26A3 disease mutant E293Q exhibited loss of ³⁶Cl⁻/anion exchange function without decrease in surface expression [64]. The corresponding human SLC26A1mutant E295Q showed partial reduction of ³⁵SO₄²⁻ transport without reduction in either 36Cl⁻ transport or polypeptide surface expression, highlighting differences between SO₄²⁻ and Cl⁻ transport by SLC26A1.

Chicken prestin and human SLC26A1 both mediate SO₄²⁻/oxalate exchange, but prestin can also mediate electrogenic Cl⁻/oxalate and Cl⁻/SO₄²⁻ exchange. Mutations in candidate anion substrate binding site residue R405 in the inner anion translocation pathway of chicken prestin/SLC26A5 completely abolished anion transport, and the corresponding mutation R399C in rat prestin abolished anion-dependent non-linear capacitance [27]. The corresponding SLC26A1 mutant K401E completely abolished transport of ³⁵SO₄²⁻ in the presence of either bath Cl⁻ or gluconate, and abolished 36Cl⁻ transport, as well. The adjacent SLC26A1 mutation A400C partially reduced ³⁵SO₄²⁻ transport only in the presence of bath Cl⁻, but not gluconate, and was without effect on SLC26A1-mediated ³⁶Cl⁻ uptake. Thus residues defined by mutagenesis as important for anion exchange by chicken prestin/SLC26A5 are also important for anion exchange by SLC26A1, but the mutations studied differentially affect SLC26A1-mediated transport of SO₄²⁻ and of Cl⁻. This difference may reflect, at least in part, a contribution of endogenous oocyte Cl⁻ transport pathways activated by recombinant expression of SLC26A1.

Xenopus oocyte functional expression data suggest that the SLC26A1 variants identified in a patient with renal Fanconi Syndrome and a patient with nephrolithiasis-calcinosis do not contribute to pathogenesis of these conditions

SLC26A1 variants found in a patient with Fanconi Syndrome exhibited no decrease in transport of either SO₄²⁻, oxalate, or Cl⁻ (Fig. 12). Indeed, the C41W variant increased transport of both Cl⁻ and SO₄²⁻ without change in oocyte surface expression. In contrast, variant A56T increased Cl⁻ transport by 4-fold but slightly decreased oxalate transport, both in parallel with 4-fold 65% increased polypeptide expression at or near the oocyte surface. These functional properties do not suggest a causative role for these SLC26A1 mutants in renal Fanconi syndrome.

The rare heterozygous SLC26A1 variant M132T and the common homozygous variant Q556R, found together in a patient with recurrent nephrolithiasis and nephrocalcinosis [19], did not alter transport of SO₄²⁻ or oxalate, whether expressed individually or co-expressed (Fig. 13). Thus, the functional properties of these variants expressed in Xenopus oocytes do not suffice to explain nephrolithiasis or nephrocalcinosis in this patient. We cannot rule out the possibility that detection of impaired trafficking, or transport phenotypes of either the Fanconi variants or the nephrolithiasis/ calcinosis variants of SLC26A1 might require expression inpolarized mammalian epithelial cells that express yet unidentified interaction partners of SLC26A1. Nonetheless, as a basolateral oxalate uptake mechanism of intestinal and proximal tubular epithelial cells, inclusion of the SLC26A1 gene in genetic variant screens of nephrolithiasis/calcinosis patients is reasonable.

Conclusion

The functional properties of human SLC26A1 emphasize the context-dependent functional complexity among not only paralogs but even species orthologs of SLC26 anion transporter polypeptides. Although human SLC26A1 shared with its paralogs the ability to mediate apparently electroneutral, cis-Cl⁻-stimulated exchange of sulfate and oxalate, and transport of bicarbonate and thiosulfate, it differed from those paralogs in several notable ways. Human SLC26A1 exhibited pH-dependent sulfate exchange that was opposite to that of rodent SLC26A1, as well as opposite to that exhibited by human SLC26A1-associated Cl⁻ transport. Human SLC26A1 differed from rat SLC26A1 in its inability to mediate oxalate/glyoxalate exchange. Human SLC26A1 exhibited partial and limited cation-dependence of its substantial associated Cl⁻ uptake. In contrast to other human SLC26 anion exchangers, human SLC26A1 inhibition by PKC was unassociated with decreased polypeptide expression at or near the oocyte surface.

The major mechanistic finding presented here is the demonstration that the cis-stimulation of human SLC26A1-mediated anion exchange by Cl⁻_[o] is a V_max effect mediated by allosteric regulation of the transporter's pH_o-sensitivity. The recently published crystal structure of the TMD of bacterial SulP/SLC26 multi-anion transporter SLC26Dg [24], and the confirmation of its previously suspected contextual modulation of anion selectivity [3,24,74] will aid in the structural analysis of allosteric regulatory interplay extracellular protons and anions in the control of human SLC26A1 activity.

Supplementary Material

424_2016_1823_MOESM1_ESM

Fig S1. SLC26A1 mediates electroneutral SO₄²⁻ uptake. Oocytes previously uninjected (a,c) or injected with 5 ng cRNA encoding SLC26A1 (b,d) were subjected to two electrode-voltage clamp measurements. Currents were recorded at the indicated clamp potentials during sequential exposure to baths of ND96(cyclamate), then of ND96(cyclamate) with added 5 mM SO₄²⁻, and finally of ND96(cyclamate) containing 5 mM SO₄²⁻ pulus 100 ⍰M DIDS (a,b) Alternatively, currents were recorded during sequential exposure to baths of ND96(cyclamate) followed by ND64(sulfate) (c,d).

424_2016_1823_MOESM2_ESM

Fig S2. Effects of sulfonate ("Good") buffers on SLC26A1. SLC26A1-mediated SO₄²⁻ uptake (5 ng cRNA) from baths containing 1 mM SO₄²⁻ in ND96(Cl⁻) (104 mM Cl⁻ plus 5 mM HEPES), or in baths containing 69 mM Cl⁻ plus 40 mM of the indicated buffers. SO₄²⁻ uptake by uninjected oocytes was in ND96(Cl⁻) containing 1 mM SO₄²⁻ (mean ± SE for (n) oocytes).

424_2016_1823_MOESM3_ESM

Fig S3. Effects of extracellular Ca²⁺ and Mg²⁺ on SLC26A1. SLC26A1-mediated SO₄²⁻ uptake (5 ng cRNA) from baths containing 1 mM SO₄²⁻ in either ND96(Cl⁻), ND96(gluconate), or their nominally Ca²⁺- and Mg²⁺-free versions, as indicated (mean ± SE, n=10).

424_2016_1823_MOESM4_ESM

Fig S4. pH_o regulates SLC26A1-associated Cl⁻ flux independent of order of exposure. a)³⁶Cl⁻ efflux traces from representative individual oocytes previously uninjected (triangles) or injected with 40 ng SLC26A1 cRNA (circles) during sequential exposure to ND96(Cl⁻) bath first at pH 8.5, then at pH 5.0, with subsequent addition of 200 μM DIDS. Note that this order of pH exposure is reversed from that shown in Fig 7b,d. b) Mean ³⁶Cl⁻ efflux rate constants (+ SE for (n) oocytes) from experiments as in panel a (***, p < 0.001).

424_2016_1823_MOESM5_ESM

Fig S5. Effects of cAMP and Ca²⁺ on SLC26A1.

a) SLC26A1-expressing oocytes (5 ng cRNA) were pre-incubated 30 min in baths of ND96(Cl⁻) or ND96(gluconate) in the absence or presence of 1 mM dibutyryl cAMP (db-cAMP) and/or 0.5 mM Isobutylmethylxanthine (IBMX), followed by a 30 min uptake assay in the same baths supplemented by 1 mM SO₄²⁻. b) SLC26A1-expressing oocytes (5 ng cRNA) were preincubated 4 h in ND96(Cl⁻) or ND96(gluconate) containing 100 μM BAPTA-AM or 0.2% DMSO (vehicle), then subjected to 30 min SO₄²⁻ uptake assays in the same baths supplemented by 1 mM SO₄²⁻. Additional SLC26A1-expressing oocytes were injected with 50 nl disodium EGTA (50 mM, pH 7.4) or NaCl (50 mM), incubated 30 min in ND96(Cl⁻) or ND96(gluconate), and subjected to 30 min SO₄²⁻ uptake assays in the same baths supplemented by 1 mM SO₄²⁻ (Mean ± SE, n=10).