Acute regulation of the SLC26A3 congenital chloride diarrhoea anion exchanger (DRA) expressed in Xenopus oocytes

Abstract

Mutations in the human SLC26A3 gene, also known as down-regulated in adenoma (hDRA), cause autosomal recessive congenital chloride-losing diarrhoea (CLD). hDRA expressed in Xenopus oocytes mediated bidirectional Cl⁻-Cl⁻ and Cl⁻-HCO₃⁻ exchange. In contrast, transport of oxalate was low, and transport of sulfate and of butyrate was undetectable. Two CLD missense disease mutants of hDRA were nonfunctional in oocytes. Truncation of up to 44 C-terminal amino acids from the putatively cytoplasmic C-terminal hydrophilic domain left transport function unimpaired, but deletion of the adjacent STAS (sulfate transporter anti-sigma factor antagonist) domain abolished function. hDRA-mediated Cl⁻ transport was insensitive to changing extracellular pH, but was inhibited by intracellular acidification and activated by NH₄⁺ at acidifying concentrations. These regulatory responses did not require the presence of either hDRA's N-terminal cytoplasmic tail or its 44 C-terminal amino acids, but they did require more proximate residues of the C-terminal cytoplasmic domain. Although only weakly sensitive to inhibition by stilbenes, hDRA was inhibited with two orders of magnitude greater potency by the anti-inflammatory drugs niflumate and tenidap. cAMP-insensitive Cl⁻-HCO₃⁻ exchange mediated by hDRA gained modest cAMP sensitivity when co-expressed with cystic fibrosis transmembrane conductance regulator (CFTR). Despite the absence of hDRA transcripts in human cell lines derived from CFTR patients, DRA mRNA was present at wild-type levels in proximal colon and nearly so in the distal ileum of CFTR(-/-) mice. Thus, pharmacological modulation of DRA might be a useful adjunct treatment of cystic fibrosis.

The SLC26 gene family (Kere et al. 1999; Everett & Green, 1999; Markovich, 2001; Alper, 2002) was first identified as a group of sulfate transporters in Neurospora. Expression cloning of the first mammalian sulfate transporter, SAT1 (SLC26A1), extended this family into vertebrates (Bissig et al. 1994; Markovich et al. 1994). DTD (SLC26A2), a sulfate-bicarbonate-oxalate exchanger of chondrocytes (Satoh et al. 1998), was identified by positional cloning of the diastrophic dysplasia gene via linkage disequilibrium mapping in the genetically isolated Finnish population (Hastbacke et al. 1994). Subsequent positional cloning of the genes underlying congenital chloride-losing diarrhoea (CLD) (Hoglund et al. 1996) and the Pendred syndrome of deafness and variably penetrant goitre (Everett et al. 1997) led to identification on chromosome 7q31 of the immediately adjacent genes SLC26A3 (down-regulated in adenoma, DRA) and SLC26A4 (pendrin) encoding polypeptides of 45 % amino acid identity. DRA (Schweinfest et al. 1993) is a Cl⁻-HCO₃⁻ exchanger of the enterocyte apical membrane (Kere et al. 1999). Pendrin (SLC26A4) is a chloride-bicarbonate-iodide-formate exchanger of the inner ear, thyroid, kidney, placenta and endometrium (Scott et al. 1999; Scott & Karniski, 2000; Royaux et al. 2001; Suzuki et al. 2002).

Additional SLC26 genes have been identified by homology searches of the human genome. Their products include the widely expressed SLC26A6 (PAT1/CFEX) Cl⁻-HCO₃⁻- formate- oxalate exchanger (Waldegger et al. 2000; Lohi et al. 2000; Knauf et al. 2001; Wang et al. 2001; Jiang et al. 2002), the SLC26A7 sulfate transporter of endothelial cells (Lohi et al. 2002a; Vincourt et al. 2002), the SLC26A8 anion transporter of testis cloned by protein interaction trap with a testis-specific RhoGAP (Toure et al. 2001; Lohi et al. 2002a), the SLC26A9 anion transporter of lung and the widely expressed, but little characterized, SLC26A11 (Lohi et al. 2002a). Prestin (SLC26A5) was cloned as one of several gene products specifically expressed in the cochlear outer hair cell, where prestin is likely to be the basolateral membrane mechanotransducer (Zheng et al. 2000). Prestin exhibits Cl⁻- and HCO₃⁻-dependent conformational change, but as yet no anion transport function (Oliver et al. 2001). SLC26 gene products are unrelated in amino acid sequence to the anion exchanger polypeptides of the SLC4 gene family.

In mammalian epithelia, SAT1/SLC26A1 is expressed basolaterally (Karniski et al. 1998), whereas DRA/SLC26A3 (Byeon et al. 1996; Greeley et al. 2001; Jacob et al. 2002), pendrin/SLC26A4 (Royaux et al. 2001) and PAT1/CFEX/SLC26A6 (Knauf et al. 2001; Wang et al. 2002) are located in lumenal membranes. The SLC26 gene products studied to date appear to function as Na⁺-independent anion exchangers with a variable range of anion selectivity. Their mechanisms of anion transport and modes of acute regulation remain little studied.

DRA/SLC26A3 is expressed in the apical membrane of surface enterocytes of proximal colon (Byeon et al. 1996) and duodenum (Jacob et al. 2002;), and at lower levels in ileum (Byeon et al. 1996). DRA is also expressed in seminal vesicles and sweat glands (Haila et al. 2000), and can be expressed in cultured cells derived from the pancreatic duct (Greeley et al. 2001) and trachea (Wheat et al. 2000). Loss-of-function mutations give rise to congenital CLD, in which daily intestinal fluid losses in stool can require daily therapeutic replacement of up to 30 l. Omeprazole, the only currently available drug treatment for the disease (Aichbichler et al. 1997), limits secretion of the gastric Cl⁻ which is delivered to distal sites of impaired reabsorption. However, the drug shows limited effectiveness.

Human DRA (hDRA) was originally identified as a transcript down-regulated in colonic adenomas and adenocarcinomas (Schweinfest et al. 1993). Levels of hDRA mRNA correlate inversely with colon tumour progression (Antalis et al. 1998). Conditional overexpression of hDRA in cultured mammalian cells has been correlated with reversible growth suppression. Constitutive overexpression of hDRA is not tolerated in most mammalian cultured cells (Chapman et al. 2002).

Although mouse DRA has been shown to function as a Cl⁻-HCO₃⁻ exchanger when overexpressed in HEK 293 cells (Melvin et al. 1999), functional analysis of hDRA as it pertains to congenital CLD has been relatively limited. In this study we document robust functional expression of hDRA in Xenopus oocytes, demonstrating by several methods that hDRA mediates bidirectional Cl⁻-Cl⁻ and Cl⁻-HCO₃⁻ exchange with a restricted anion selectivity. We define the first moderately potent inhibitors of hDRA, and we demonstrate independent regulation of hDRA by intracellular pH (pH_i) and by NH₄⁺, but not by extracellular pH (pH_o). We characterize portions of the carboxy- and amino-terminal cytoplasmic domains of hDRA which are required for basal and for regulated Cl⁻ transport activities. These data, along with demonstration of loss of function in two disease mutants, document the functional importance of the ‘sulfate transporter and anti-sigma factor antagonist’ (STAS) domain (Aravind & Koonin, 2000). We show that cystic fibrosis transmembrane conductance regulator (CFTR) co-expression with hDRA in Xenopus oocytes confers modest cAMP-stimulation upon hDRA-ClD-mediated Cl⁻-HCO₃⁻ exchange. Lastly, we show that hDRA mRNA expression persists in colon and ileum of a CFTR(-/-) mouse, allowing the suggestion of DRA modulation as an adjunct therapy for cystic fibrosis.

Methods

Materials

Na³⁶Cl, [³⁵S]sulfate and [α-³²P]UTP were purchased from ICN (Irvine, CA, USA). Other chemical reagents were of analytical grade and obtained from Sigma (St Louis, MO, USA), Calbiochem (San Diego, CA, USA) or Fluka (Milwaukee, WI, USA). Restriction enzymes and T4 DNA ligase were from New England BioLabs (Beverly, MA, USA). The EXPAND High-fidelity PCR system was obtained from Roche Diagnostics (Indianapolis, IN, USA). [¹⁴C]oxalate was the generous gift of C. Scheid and T. Honeyman (University of Massachusetts Medical Center) and [¹⁴C]butyrate was from NEN-Dupont (Boston, MA, USA). All other chemical reagents were from Sigma or Fluka and were of reagent grade.

Solutions

ND-96 solution (pH 7.40) consisted of (mm): 96 NaCl, 2 KCl, 1.8 CaCl₂, 1 MgCl₂, 5 Hepes, 2.5 sodium pyruvate, with 100 μg ml⁻¹ gentamicin. ND-96 flux medium and other flux media lacked sodium pyruvate and gentamicin. pH values of 7.4 and 8.0 in room air flux media were achieved with 5 mm sodium Hepes. 4-Morpholine-ethanesulfonic acid (Mes; 5 mm) was used for room air flux media of pH 5.0. In Cl⁻-free solutions, NaCl was replaced isosmotically with 96 mm sodium isethionate. The Cl⁻ salts of K⁺, Ca²⁺ and Mg²⁺ were substituted with the corresponding equimolar gluconate salts. Hepes-free CO₂-HCO₃⁻-buffered solutions of pH 7.4 were saturated with 5 % CO₂-95 % air at room temperature for ∼1 h, and differed from Cl⁻-free ND-96 in replacement of 24 mm sodium isethionate with 24 mm NaHCO₃⁻. The pH of CO₂-HCO₃⁻-buffered solutions was verified prior to each experiment. Addition to flux media of the weak acid salt sodium butyrate was in equimolar substitution for NaCl.

Mutagenesis

hDRA cDNA (Schweinfest et al. 1993) was subjected to four-primer PCR mutagenesis as described previously (Stewart et al. 2001, 2002). Oligonucleotide primers were obtained from Biosynthesis (Woodlands, TX, USA); sequences are available upon request. Mutagenized PCR products were ligated into appropriately engineered host plasmids to reconstruct the desired coding sequence. Plasmid DNA sequences of the PCR-amplified regions and their ligation junctions were confirmed on both strands. Reconstruction of full-length mutant hDRA was confirmed by restriction digest.

cRNA expression in xenopus oocytes

Mature female Xenopus (NASCO, Madison, WI, USA) were maintained and subjected to partial ovariectomy under tricaine-hypothermia anaesthesia as previously described (Humphreys et al. 1994), conforming to methods approved by the Institutional Animal Care and Use Committee (IACUC) of the Beth Israel Deaconess Medical Center. After several such procedures, frogs anaesthetized with tricaine were humanely killed. Stage V–VI oocytes were manually defolliculated following incubation of ovarian fragments with 2 mg ml⁻¹ collagenase A or collagenase B (Boehringer Mannheim, Indianapolis, IN, USA) for 60 min in ND-96. Oocytes were injected on the same day with cRNA or with 50 nl H₂O. Capped cRNA was transcribed from linearized cDNA templates with SP6 RNA polymerase (Ambion, Austin, TX, USA) and resuspended in diethylpyrocarbonate-treated water. RNA integrity was confirmed by agarose gel electrophoresis in formaldehyde. Injected oocytes were then maintained for 2–6 days at 19 °C.

Robust expression of hDRA cRNA in Xenopus oocytes required the presence of Xenopus β-globin untranslated regions flanking the hDRA coding region. Thus, ³⁶Cl⁻ influx into oocytes injected with hDRA cRNA transcribed from a pBF Xenopus oocyte expression vector template was tenfold higher than in oocytes injected with hDRA cRNA transcribed from a pBluescript template. All experiments shown used cRNA transcribed from pBF. Since addition of a 6 × His epitope tag to the N-terminus of human DRA reduced ³⁶Cl⁻ transport activity by 90–95 %, these epitope-tagged constructs were not used further.

Isotopic flux studies

Unidirectional ³⁶Cl⁻ influx studies were carried out in ND-96 for 15 or 30 min periods as previously described (Chernova et al. 1997). Total bath [Cl⁻] was 104 mm. Unidirectional ³⁶Cl⁻ efflux studies were as described by Chernova et al. (1997) and Stewart et al. (2001, 2002). Individual oocytes in Cl⁻-free ND-96 were injected with 50 nl of 130 mm Na³⁶Cl (10 000–12 000 c.p.m.). Following a 5–10 min recovery period, the efflux assay was initiated by transfer of individual oocytes to 6 ml borosilicate glass tubes, each containing 1 ml efflux solution. At intervals of 3 min, 0.95 ml of this efflux solution was removed for scintillation counting and replaced with an equal volume of fresh efflux solution. Following completion of the assay with a final efflux period in the absence of bath chloride (substituted by sodium isethionate and by gluconate salts of K⁺, Ca²⁺ and Mg²⁺), each oocyte was lysed in 100 μl of 2 % sodium dodecyl sulfate (SDS). Samples were counted for 3–5 min such that the magnitude of 2 s.d. was < 5 % of the sample mean.

Experimental data were plotted as ln(percentage c.p.m. remaining in the oocyte) vs. time. The ³⁶Cl⁻ efflux rate constants were measured from linear fits to data from the last three time points sampled for each experimental condition. Within each experiment, water-injected and DRA cRNA-injected oocytes from the same frog were subjected to parallel measurements. On each experimental day, activity of tested mutant DRA polypeptides was compared to wild-type (WT) DRA activity at bath pH 7.4. Each experimental condition and each DRA mutant was tested in oocytes from at least two frogs. Values for K_1/2 for extracellular Cl⁻ and for ID₅₀ (concentration of inhibitor at which the DRA Cl⁻ efflux rate constant was 50 % of the uninhibited value measured in the same oocyte within the same experiment) were calculated from curve fits to the Michaelis-Menten equation (Ultrafit 3.0, Elsevier, New York, USA).

To vary pH_i, oocytes were pre-exposed to 40 mm sodium butyrate (substituting for NaCl) prior to initiation of an efflux experiment to produce intracellular acidification. Upon its removal from the bath (with restoration of chloride) during the course of the efflux experiment, pH_i rapidly alkalinized while pH_o remained constant (Stewart et al. 2001). Other oocyte groups were exposed to 20 mm NH₄Cl during the course of efflux experiments. Drugs were added to the bath or were injected into oocytes either prior to or together with isotope as indicated.

The [³⁵S]sulfate influx experiments were performed in 150 μl influx medium containing 5 μCi carrier-free [³⁵S]sulfate (47 nm) in the absence or presence of 2 mm unlabelled sulfate. The [³⁵S]sulfate efflux experiments were carried out in ND-96 or in baths containing 64 mm sodium sulfate or 20 mm sodium sulfate plus 66 mm sodium isethionate (Chernova et al. 1997). The [¹⁴C]butyrate influx experiments were performed in 150 μl influx medium containing 167 μm (carrier-free) butyrate (3.33 μCi ml⁻¹) in the absence or presence of 40 mm added unlabelled butyrate, where sodium butyrate substituted for equimolar NaCl. The efflux of [¹⁴C]butyrate was measured into ND-96 bath solution. The [¹⁴C]oxalate influx experiments were performed in Ca²⁺- and Mg²⁺-free influx medium containing 1 mm oxalate (2.67 μCi ml⁻¹). Efflux of [¹⁴C]oxalate was measured into a bath containing ND-96.

Statistical analyses were carried out using paired or unpaired Student's two-tailed t tests or ANOVA for multiple group comparisons (Scheffé's or Fisher's test). Analyses were performed using Microsoft Excel.

Measurement of oocyte pH_i

Oocyte pH_i was monitored during bath superfusion using 2′,7′-bis(carboxyethyl)-5-(and −6)carboxyfluorescein (BCECF) fluorescence excitation ratio imaging, as previously described (Zhang et al. 1996; Goss et al. 2001). Cl⁻-HCO₃⁻ exchange was assayed by measurement of dpH_i/dt during substitution of bath Cl⁻ with isethionate followed by Cl⁻ restoration. Initial rates of intracellular alkalinization and acidification were computed by linear least squares fit to at least six ratio measurements. Cl⁻-HCO₃⁻ exchange activity in all groups of oocytes was indistinguishable in the presence and in the absence of bath Na⁺. Therefore, pooled data from both conditions is presented for each construct assayed. dpH_i/dt values were analysed by two-tailed t test or ANOVA.

Oocyte pH_i was also measured during bath superfusion using pH microelectrodes as described previously (Romero et al. 1997; Stewart et al. 2001, 2002). The electrodes were calibrated at pH 6.0, 7.0 and 8.0, and exhibited slopes of at least 55 mV (pH unit)⁻¹.

Immunoblots

Clarified oocyte lysates in 1 % Triton X-100 were prepared, separated by 8 % SDS-PAGE and transferred to nitrocellulose by semi-dry blotter. Immunoblots were developed with primary affinity-purified antibody to hDRA C-terminal peptide (Byeon et al. 1996) and secondary antibodies as described (Casula et al. 2001). The anti-hDRA antibody was not useful for confocal immunofluorescent detection of near-surface hDRA in fixed, permeabilized, intact oocytes prepared as described (Casula et al. 2001).

RNAse protection assay

CFTR(−/−) mice (Snouwaert et al. 1992) obtained from Jackson Laboratories (Bar Harbor, ME, USA) were bred at the Beth Israel Deaconess Medical Center and maintained after weaning on a Peptamen diet, as previously described (Freedman et al. 1999). A mouse DRA cDNA fragment encoding nt 1191–1478 (Genbank AF136751) was cloned by RT-PCR, sequenced and subcloned into pCR2. The probe encompassed 4733 nt of genomic sequence (Genbank supercontig NW-000050 nt 7095487–7100220) and bridged introns 9, 10 and 11 of the mouse DRA gene. ³²P-labelled and unlabelled antisense RNA probe was transcribed with SP6 RNA polymerase in the presence or absence of [α-³²P]UTP. Unlabelled sense RNA probe was transcribed with T7 polymerase (MaxiScript and Megascript in vitro transcription kits, Ambion). Proximal colon and distal ileum were harvested from two groups of mice with CFTR genotypes (−/−), (−/+) and (+/+). Mice in each group were 6–12 weeks of age. Total RNA was prepared using the RNEasy MiniKit (Qiagen, Valencia, CA, USA). The RNAse protection assay was performed with the Ambion RPA III kit according to manufacturer's instructions. Reaction products were electrophoresed on a 5 % urea-8 % polyacrylamide gel, dried on paper and subjected to autoradiography. Abundance of DRA mRNA in the ileum and colon of CFTR(−/+) mice relative to that in WT mice in each experiment was quantified from scanned autoradiographs using NIH Image version 1.62. Statistical analysis of raw density values was by two-tailed t test, applied to groups of samples run on the same gel within the same experiment. Normalized DRA mRNA levels in CFTR(−/−) mouse tissue were expressed as percentages of the mean WT level analysed on the same gel.

Results

Human DRA functions in xenopus oocytes as a Cl⁻-HCO₃⁻ exchanger with restricted anion selectivity

Human DRA (hDRA) expressed in Xenopus oocytes has been reported to transport ³⁵SO₄^2- and ³⁶Cl⁻ in unidirectional influx assays, but the rates of transport reported for both anions were picomoles per oocyte per hour (Silberg et al. 1995; Moseley et al. 1999), or three orders of magnitude slower than that normally exhibited by Cl⁻ transporters of the SLC4 or the SLC26 families. In contrast, mouse DRA expressed in HEK 293 cells mediated robust Cl⁻-HCO₃⁻ exchange (Melvin et al. 1999). We therefore studied hDRA function and anion selectivity in Xenopus oocytes.

hDRA-expressing oocytes displayed Cl⁻ uptake rates up to 20 nmol ³⁶Cl⁻ h⁻¹ (Fig. 1A), ∼50-fold higher than background and > 10³-fold higher than reported previously by Moseley et al. (1999).

Figure 1. Anion selectivity of human down-regulated in adenoma (hDRA)-mediated isotopic fluxes.

A, hDRA-mediated ³⁶Cl⁻ influx; representative of 6 similar experiments. B, [¹⁴C]butyrate influx into oocytes pre-injected with water or with hDRA cRNA; one of two similar experiments. C, hDRA-mediated [¹⁴C]oxalate influx; representative of three similar experiments (*P < 0.05). D, [³⁵S]sulfate influx into oocytes preinjected with water or with hDRA cRNA; one of two similar experiments (see opposite). The number of oocytes tested is given in parentheses. E, [³⁵S]sulfate efflux from DRA-expressing oocytes is not activated by acid bath pH. Efflux was measured from 5 individual oocytes each transferred from Cl⁻-free bath into pH 5 ND-96 and then into pH 7.4 ND-96. Inset: [³⁵S]sulfate efflux rate constants in each condition. F, ³⁶Cl⁻ efflux from two water-injected oocytes (upper traces) and from 5 DRA-expressing oocytes (lower traces) first in the absence of Cl⁻, then transferredinto 64 mm sulfate, then in ND-96 and finally returned to the Cl⁻-free bath. Inset: ³⁶Cl⁻ efflux rate constants for DRA-expressing oocytes.

(Moseley et al. (1999; see Figs 3 and 4) showed, as did Silberg et al. (1995), that hDRA mediated low picomole per hour rates of [³⁵S]sulfate uptake into Xenopus oocytes from an uptake medium containing 1 mm sulfate and 106 mm Cl⁻¹. However, Fig. 5 of Moseley et al. (1999) presented contradictory data. Whereas hDRA-mediated sulfate transport was equivalent in zero and 10 mm Cl⁻¹, sulfate uptake by hDRA was completely abolished when bath Cl⁻¹ was > 10 mm. Thus, these oocyte anion transport data are in question.)

Figure 3. Extracellular Cl⁻ dependence of hDRA-mediated ³⁶Cl⁻ efflux.

A, ³⁶Cl⁻ efflux into bath solutions containing increasing Cl⁻ concentrations. Uppermost trace represents a water-injected oocyte. Lower traces represent three hDRA-expressing oocytes. B, [Cl⁻] dependence of hDRA-mediated ³⁶Cl⁻ efflux from a single representative oocyte. The curve is the best Michaelis-Menten fit to the data. Mean K_1/2(Cl⁻_o) was 5.8 ± 1.5 mm for 13 hDRA-expressing oocytes studied with the same protocol.

Figure 4. CLD mutations in STAS domain exhibit loss of function.

Top: schematic diagram showing location of the chloride-losing diarrhoea (CLD) missense mutations I544N, and 2116delA which leads to termination at codon 711. Open bars indicate hydrophilic, putative cytoplasmic domains. The grey bar indicates the hydrophobic polytopic transmembrane domain. Hatch marks indicate that grey bar is shorter than if it was drawn to scale. Amino acid residue numbers under the bar mark approximate domain boundaries. Bottom: ³⁶Cl⁻ influx assay shows that both missense mutants exhibit loss of function; numbers of oocytes tested given in parentheses. Injected cRNA was 1 ng for wild-type (WT) hDRA and 10 ng for each mutant. Increasing cRNA to 50 ng did not lead to ³⁶Cl⁻ uptake by the mutants. Inset: hDRA immunoblot of whole oocyte lysates comparing WT hDRA with hDRA I544N polypeptide (10 ng injected cRNA for each). hDRA 2116delA, lacking the C-terminus, is undetectable by the anti-hDRA C-terminal peptide antibody.

In contrast, hDRA exhibited no uptake of ³⁵SO₄^2- above levels exhibited by water-injected oocytes, whether from baths containing 2 mm (Fig. 1D) or sub-micromolar (carrier-free) sulfate, and uptake was not increased at bath pH 5.0 (not shown). hDRA did not mediate detectable ³⁵SO₄^2- efflux into baths containing 104 mm Cl⁻ (ND-96) at either pH 5.0 or 7.4 (Fig. 1E). Neither did hDRA mediate detectable ³⁶Cl⁻ efflux into either 64 mm sulfate (Fig. 1F) or 20 mm sulfate (not shown).

(Data shown is for hDRA cRNA transcribed from template subcloned into the Xenopus expression vector pBF. Influx and efflux of [³⁵S]sulfate was also examined in oocytes expressing hDRA cRNA subcloned into pBluescript. Again, no sulfate transport was observed.)

Since DRA is expressed in the colon and colonic enterocytes (Chu & Montrose, 1996) and apical membrane vesicles exhibit butyrate-anion exchange (Rajendran & Binder, 1994), hDRA was tested for [¹⁴C]butyrate uptake at two bath butyrate concentrations. In neither condition did DRA exhibit butyrate uptake (Fig. 1B). hDRA did exhibit low level uptake (Fig. 1C) of [¹⁴C]oxalate at a bath concentration of 1.0 mm, as well as low rates of [¹⁴C]oxalate efflux into ND-96 (not shown), in contrast to the robust oxalate influx and efflux mediated by hPAT1/SLC26A6 (Jiang et al. 2002; M. N. Chernova & S. L. Alper, unpublished results).

Human DRA mediates both Cl⁻-Cl⁻ and Cl⁻-HCO₃⁻ exchange

As shown in Fig. 2A and B, DRA-mediated ³⁶Cl⁻ efflux required trans-anion. Extracellular media containing either 24 mm HCO₃⁻ or 104 mm Cl⁻ (ND-96) supported ³⁶Cl⁻ efflux. Figure 2C shows that, in a BCECF-AM-loaded, hDRA-expressing oocyte, bath Cl⁻ removal induced a strong intracellular alkalinization that was fully reversible by reintroduction of Cl⁻ to the bath, and was absent from a control water-injected oocyte. This Cl⁻-HCO₃⁻ exchange did not require bath Na⁺ (N-methyl-d-glucamine substitution, not shown), in agreement with results for mouse DRA (Melvin et al. 1999). At a mean resting pH_i of 7.03 ± 0.01 (mean ± s.e.m., n = 25), values for hDRA-mediated initial dpH_i/dt induced by Cl⁻ removal and restoration, respectively, were 0.022 ± 0.0007 and −0.022 ± 0.0008 u min⁻¹ (u = units; n = 14). The corresponding values for water-injected oocytes were 0.007 ± 0.0003 u min⁻¹ and −0.0064 ± 0.0005 u min⁻¹.

Figure 2. hDRA mediates Cl⁻-Cl⁻ and Cl⁻-HCO₃⁻ exchange.

A, ³⁶Cl⁻ efflux traces from water-injected oocytes (upper three traces) and from hDRA-expressing oocytes (lower four traces) during sequential incubations in baths without Cl⁻, baths containing 24 mm HCO₃⁻ and baths containing 96 mm Cl⁻ (ND-96). B, relative increase in stimulation of ³⁶Cl⁻ efflux upon changing from Cl⁻-free baths to baths containing the indicated anion (n = 4). C, pH_i of representative oocytes previously injected with water (□) or with hDRA cRNA (♦) during bath Cl⁻ removal and restoration. pH_i was measured by 2′,7′-bis(carboxyethyl)-5-(and −6)-carboxyfluorescein (BCECF) fluorescence ratio imaging. D, upper panel: pH_i of representative oocytes previously injected with water (lower trace) or with hDRA cRNA (upper trace) during bath Cl⁻ removal and restoration. pH_i was measured by pH-sensitive microelectrode. D, lower panel: membrane potential (V_m) traces recorded from the same oocytes previously injected with water (upper trace) or with hDRA cRNA (lower trace).

(This 0.1 unit difference in initial pH_i represents the difference between calculated values of intracellular HCO₃⁻ of 8.3 and 10 mm. Taken together with the oocyte's intrinsic buffer capacity of 18.9 mm (pH unit)⁻¹ (Stewart et al. 2002), the calculated proton equivalent flux in oocytes expressing DRA (0.87 mm min⁻¹) is not further stimulated by cAMP.)

Figure 2D confirms hDRA-mediated Cl⁻-HCO₃⁻ exchange as measured by pH-sensitive microelectrode (resting pH_i 6.92 ± 0.06 in 12 similar oocytes. Bath Cl⁻ removal led to dpH_i/dt 0.031 ± 0.004 u min⁻¹ in nine similar oocytes. The small changes in membrane potential of hDRA-expressing oocytes during restoration of bath Cl⁻ (−1 ± 1 mV, n = 9) were consistent with electroneutrality of hDRA-mediated Cl⁻-HCO₃⁻ exchange in these conditions, as previously suggested for mouse DRA (Melvin et al. 1999).

Figure 3 shows the extracellular Cl⁻ dependence of DRA-mediated ³⁶Cl⁻ efflux from oocytes. The K_1/2 value for extracellular Cl⁻ was 5.8 mm, similar to that of AE2 in Xenopus oocytes (Humphreys et al. 1994). K_1/2 values for extracellular Cl⁻ have not been reported for other SLC26 transporters.

Two recently described hDRA missense mutants that cause CLD are nonfunctional in Xenopus oocytes

CLD mutations have been found throughout the coding region of hDRA. Many are nonsense mutants or frameshift/termination mutants which truncate most of the protein and are expected to be nonfunctional. One hDRA disease mutation has been subjected to study of anion transport function, the major Finnish founder effect mutation, ΔV317 (Moseley, 1999). Mutations within the conserved sulfate transporter anti-sigma factor antagonist (STAS) domain of the cytoplasmic C-terminal tail (Aravind & Koonin, 2000) which leave the polytopic transmembrane domain intact have not been studied in this way. Therefore we tested the activity of two such mutations in the STAS domain. The disease-causing missense mutation I544N (Hoglund et al. 2001) is situated in the N-terminal portion of the STAS domain. It resides in the linker between β2 and α1, in an exposed site immediately adjacent to the highly conserved loop between β3 and α2 which has been implicated in nucleotide binding in the Bacillus subtilis antisigma-factor antagonist SPOIIAA. Residue 544 of DRA is conserved as a hydrophobic residue throughout the SLC26 gene family, and in related nonmammalian sequences (Aravind & Koonin, 2000). hDRA I544N is nonfunctional as assessed by ³⁶Cl⁻ influx into Xenopus oocytes (Fig. 4).

The disease-causing mutation 2116delA frameshifts the hDRA open reading frame at codon 706, terminating the DRA polypeptide at residue 711 in the C-terminal portion of the STAS domain (Hoglund et al. 1998). This frameshift deletes the two terminal α4 and α5 helices of the STAS domain, which exhibit little sequence conservation within a structure predicted to be conserved (Aravind & Koonin, 2000). This hDRA mutant was also nonfunctional in the ³⁶Cl⁻ influx assay. Consistent with the autosomal recessive inheritance pattern of congenital CLD, neither hDRA mutant I544N nor 2116delA exhibited a dominant negative phenotype when coexpressed with WT hDRA (not shown). I544N mutant polypeptide accumulated in oocytes to WT levels, but the degree of its surface expression remains unknown. Both WT DRA and the I544N mutant migrated with M_r similar to that of hDRA expressed in Sf9 insect cells (Byeon et al. 1998). While this paper was in revision, Ko et al. (2002) reported that the murine equivalents of the human DRA disease mutations L496R and I675–676ins expressed in HEK 293 cells were also functionally inactive.

Systematic analysis of hDRA cytoplasmic domain truncation mutants

The distal C-terminal cytoplasmic tail of hDRA harbours a Type I PDZ domain consensus recognition sequence, as do several other polypeptides of the SLC26 gene family. However, neither the mutant polypeptide hDRA T762V, which lacks the critical hydroxyl side chain in the −3 position, nor the hDRA truncation mutant 760X which lacks the entire consensus recognition sequence, impaired hDRA-mediated ³⁶Cl⁻ efflux activity (Fig. 5A). Neither mutation of a putative tyrosine phosphorylation site at Y756 in the mutant 755X, nor deletion of the adjacent cationic region in the mutant 751X diminished basal transport activity of hDRA (Fig. 5A). Even deletion of 44 C-terminal residues in the mutant 720X did not decrease hDRA activity in oocytes (Fig. 5B). In contrast, the mutant polypeptide Y756F exhibited a doubling of basal ³⁶Cl⁻ efflux activity (Fig. 5A). Removal of the entire N-terminal cytoplasmic domain (Δ_N52) or the entire C-terminal cytoplasmic domain (524X) each decreased but did not abolish DRA activity expressed in Xenopus oocytes (Fig. 5B; Δ_N52 differed from WT hDRA in a two-way comparison, but not by multiple comparison ANOVA). In contrast, internal deletion of the STAS domain (amino acids 560–720) with retention of the C-terminal 44 amino acids (‘-STAS’ in Fig. 5B) completely abolished anion transport activity, despite accumulation in oocytes of this mutant polypeptide to levels higher than WT (inset, Fig. 5B). Neither the reduced function mutant hDRA Δ_N52 nor the nonfunctional mutant hDRA with the internal deletion of the STAS domain exhibited a dominant negative phenotype when co-expressed with WT hDRA (not shown).

Figure 5. Effects of cytoplasmic domain truncations on hDRA basal activity.

A, left, schematic diagram showing C-terminal region of hDRA across which C-terminal deletions were generated. Italics indicate consensus binding site for PDZ domain. Asterisks indicate potential phosphoryation sites. A, right, ³⁶Cl⁻ efflux rate constants exhibited by WT hDRA and by the indicated hDRA mutants. One nanogram of cRNA was injected for all constructs. * Y756F differs significantly from WT DRA and all other groups by ANOVA. B, left, schematic diagram depicting the more extensive deletions generated. B, right, ³⁶Cl⁻ efflux rate constants exhibited by WT hDRA and by the indicated hDRA mutants. One nanogram of cRNA was injected for WT hDRA and 720X; 10 ng cRNA injected for the mutants (-STAS), 524X and Δ_N52. Number of oocytes tested given in parentheses. * Significant difference from WT DRA by ANOVA. Inset, hDRA STAS domain deletion polypeptide accumulates in oocytes to higher-than-WT levels. hDRA 524X cannot be detected by the anti-hDRA C-terminal peptide antibody.

hDRA is regulated by pH_i, but not by pH_o

The AE2 (SLC4A2) anion exchanger is acutely regulated by both pH_i and by pH_o, whereas the AE1 (SLC4A1) anion exchanger is comparatively insensitive to changing pH (Zhang et al. 1996; Stewart et al. 2001, 2002). hDRA-mediated ³⁶Cl⁻ efflux was unaffected by shifting pH_o from 5.0 to 7.4 (Fig. 6A and C). However, hDRA-mediated ³⁶Cl⁻ efflux was inhibited in the presence of 40 mm butyrate, and stimulated by its removal (Fig. 6B), suggesting inhibition by intracellular acidification, and removal of inhibition by intracellular alkalinization. Butyrate was neither a substrate for (Fig. 1B) nor an inhibitor of hDRA-mediated Cl⁻ transport (not shown). Removal of the entire C-terminal cytoplasmic domain (524X) abolished this regulation by pH_i. However, removal of the entire short N-terminal cytoplasmic domain (Δ_N52) or removal of the C-terminal 44 amino acids (720X) retained sensitivity to altered pH_i. Thus, the C-terminal cytoplasmic domain contains proximate regions absolutely required for regulation by pH_i, whereas the N-terminal cytoplasmic domain is dispensable for this regulation.

Figure 6. hDRA is inhibited by intracellular, but not by extracellular acidification.

A, ³⁶Cl⁻ efflux from a water-injected oocyte (uppermost trace) and several hDRA-expressing oocytes (lower traces) during transition from bath pH 5.0 to 7.4, followed by exposure to 2 mm DIDS at pH 7.4. B, ³⁶Cl⁻ efflux from a water-injected oocyte (uppermost trace) and several oocytes expressing either hDRA 524X (middle traces) or WT hDRA (lower traces) during exposure and removal of 40 mm butyrate, followed by switch to Cl⁻-free bath. C, summary of effect of pH_o on WT hDRA-mediated ³⁶Cl⁻ efflux. D, summary of relative increase in activation by butyrate removal of ³⁶Cl⁻ efflux from oocytes expressing WT hDRA or the indicated mutants. Number of oocytes tested given in parentheses.

hDRA is stimulated by NH₄⁺ but not by hypertonicity

NH₄⁺ and hypertonicity each stimulate AE2-mediated Cl⁻ transport, but AE1 activity is unaffected by either stimulus (Humphreys et al. 1994, 1997; Chernova, 2003). hDRA was also stimulated by NH₄⁺ by 2.6 ± 0.23-fold (Fig. 7). Removal of the entire C-terminal cytoplasmic tail from hDRA abolished sensitivity to NH₄⁺ in the mutant 524X, whereas removal from hDRA of the entire N-terminal cytoplasmic tail preserved (and perhaps enhanced) stimulation by NH₄⁺. Truncations across the C-terminal 44 residues of hDRA minimally reduced stimulation by NH₄⁺, as was also true for the candidate phosphorylation site mutants T762V and Y756F (Fig. 7). Hypertonicity (362 vs. 212 mosmol l⁻ isotonic medium) had virtually no effect on DRA-mediated ³⁶Cl⁻ influx or efflux (1.17 ± 0.12-fold stimulation, n = 18)

Figure 7. hDRA is stimulated by NH₄⁺.

A, ³⁶Cl⁻ efflux from a water-injected oocyte (uppermost trace) and several oocytes expressing either hDRA 524X (middle traces) or WT hDRA (lower traces) during exposure to NH₄⁺, followed by transition to Cl⁻-free bath. B, relative increase in stimulation by NH₄⁺ of hDRA-mediated ³⁶Cl⁻ efflux from oocytes expressing either WT hDRA or the indicated mutant polypeptides; *P < 0.05 by ANOVA. Number of oocytes tested given in parentheses.

Tests of other potential regulators

Although AE2 is sensitive to inhibition by the imidazole calmodulin antagonist, calmidazolium (Chernova et al. 2003), neither ³⁶Cl⁻ influx nor efflux mediated by WT hDRA or by the mutant 524X were inhibited by calmidazolium (10 μm in bath), whether in the absence or presence of NH₄⁺ (n = 4–8). The C-terminal cytoplasmic domain of mDRA has been reported to interact with the serine-threonine protein phosphatase regulatory subunit PP2A Bδ (Chang et al. 2001). However, injected serine-threonine phosphatase inhibitor calyculin A (100 nm estimated final intra-oocyte concentration) altered neither basal or NH₄⁺-stimulated Cl⁻ influx nor efflux activity of WT hDRA or the mutant 720X (n = 3–8, not shown). Since DRA is localized in the apical membrane above the terminal web of the actin cytoskeleton, the effects of actin cytoskeleton depolymerization and stabilization on hDRA were tested. However, neither the actin-depolymerizing drug cytochalasin D nor the polymeric actin-stabilizing drug NBD-phallicidin (10 μm in bath during 30 min preincubation and 30 min ³⁶Cl⁻ influx) altered WT hDRA activity (n = 7–9, not shown).

Inhibitor pharmacology of hDRA

The sensitivity of hDRA to inhibition by stilbene disulfonates remains controversial (Silberg et al. 1995; Byeon et al. 1998; Moseley et al. 1999; Melvin et al. 1999). We therefore screened several known anion transport inhibitors as hDRA blockers. As shown in Fig. 8, the nonsteroidal anti-inflammatory drug (NSAID) tenidap inhibited hDRA-mediated ³⁶Cl⁻ efflux with an ID₅₀ of 9.6 ± 1.9 μm. Tenidap's ability to acidify mammalian cells (McNiff et al. 1995) as well as Xenopus oocytes (Ducoudret et al. 2001) could explain this inhibition of hDRA. However, 100 μm tenidap inhibited the pH_i-insensitive hDRA mutant 524X by 93 ± 4 % (n = 4, not shown). Thus, tenidap blockade of hDRA function is not secondary to inhibition of hDRA by intracellular acidification.

Figure 8. hDRA is inhibited by tenidap much more potently than by stilbenes.

A, ³⁶Cl⁻ efflux into bath solutions containing increasing tenidap concentrations. Uppermost trace represents a water-injected oocyte. Lower traces represent three hDRA-expressing oocytes. B, tenidap concentration dependence of hDRA-mediated ³⁶Cl⁻ efflux from a single oocyte. Curve is the best fit to this data of the Michaelis-Menten equation. Mean ID₅₀ for tenidap for 6 similar hDRA-expressing oocytes was 9.6 ± 1.9 μm.

Table 1 summarizes the inhibitory properties of several NSAIDs and other drugs. Notable among them are the anti-inflammatory drugs niflumate (ID₅₀ 6.8 ± 2.5 μm), UK5099, an inhibitor of anion exchange in polymorphonuclear leukocytes (Simchowitz & Davis, 1990), and the non-stilbene blocker of red cell anion exchange and anion conductance NS1652 (Bennekou et al. 2000). DIDS was a weak inhibitor of hDRA in Xenopus oocytes. The inhibitor of cardiomyocyte Cl⁻-HCO₃⁻ exchange S20787 (Loh et al. 2001) stimulated hDRA activity by 57 ± 25 % when present at a concentration of 5 μm (n = 10, not shown).

Table 1.

Potency of inhibition of hDRA-mediated ³⁶Cl⁻ efflux into ND-96 solution

Inhibitory drug	ID50 (μM)	Inhibition	(n)
Niflumic acid	6.8±2.5	—	(6)
Tenidap	9.6±1.9	—	(6)
UK5099 (100 μM)	—	100%	(3)
NS1652 (20 μM)	—	69±5%	(3)
WW781 (20 μM)	—	39±14%	(3)
Hydroxycinnamate (1 mM)	—	34%	—
DIDS (0.5 mM)	—	26±5%	(3)
(2 mM)	—	66±1%	(3)

Values are means ± s.e.m. for (n) oocytes. Mean efflux rate constants (k) in the absence of drugs ranged between 0.014 and 0.035 min⁻¹ in these experiments.

CFTR confers modest cAMP-sensitivity upon hDRA in xenopus oocytes

Cl⁻-HCO₃⁻ exchange is activated and rendered cAMP-sensitive in cultured NIH-3T3 or HEK 293 cells which overexpress CFTR (Lee et al. 1999). In contrast, CFTR does not activate AE2 (SLC4A2) when coexpressed in Xenopus oocytes (L. Jiang, M. Chernova & S. Alper, unpublished observations), consistent with the basolateral localization of SLC4 polypeptides in most tissues. The colocalization of CFTR and DRA in intestine (Jacob et al. 2002) prompted a test of the effect of CFTR coexpression upon hDRA function in Xenopus oocytes (Fig. 9). hDRA exhibited cAMP-insensitive, bidirectional Cl⁻-HCO₃⁻ exchange activity (Fig. 9A and D). Bath Cl⁻ restoration returned pH_i to original values. Interestingly, CFTR alone exhibited a lower level of cAMP-insensitive, nominal Cl⁻-HCO₃⁻ exchange activity (Fig. 9B and D), in contrast to its cAMP-stimulated ³⁶Cl⁻ efflux and Cl⁻ conductance (data not shown). These anion exchange activities were not additive, as CFTR co-expression did not enhance basal hDRA-mediated Cl⁻-HCO₃⁻ exchange activity. However, addition of cAMP and 3-isobutyl-1-methylxanthine (IBMX) to oocytes co-expressing hDRA and CFTR acutely enhanced Cl⁻-HCO₃⁻ exchange activity (Fig. 9C and D). Nominal Cl⁻-HCO₃⁻ exchange activities in oocytes expressing hDRA and CFTR, individually and together, were all Na⁺-independent.

Figure 9. CFTR co-expression confers modest cAMP-sensitivity upon hDRA-mediated Cl⁻-HCO₃⁻ exchange.

A, hDRA-mediated Cl⁻-HCO₃⁻ exchange activity measured by BCECF fluorescence ratio imaging in representative oocytes subjected to sequential Cl⁻ removal and restoration in the absence (○) or presence (•) of 2 mm cAMP plus 1 mm IBMX. B, Cl⁻-HCO₃⁻ exchange activity in representative oocytes expressing CFTR and subjected to sequential Cl⁻ removal and restoration in the absence (▵) or presence (▴) of 2 mm cAMP plus 1 mm IBMX. C, Cl⁻-HCO₃⁻ exchange activity in representative oocytes co-expressing hDRA and CFTR, subjected to sequential Cl⁻ removal and restoration in the absence (▵) or presence (▴) of 2 mm cAMP plus 1 mm IBMX. D, summarized results of Cl⁻-HCO₃⁻ exchange activity for the number of oocytes given in parentheses. Resting pH_i was 7.03 ± 0.02 for ‘H₂O’ and ‘DRA’ (minus cAMP) groups, and 7.12 ± 0.02 for all other groups. Note that stimulation by cAMP and IBMX is observed only in oocytes co-expressing CFTR and hDRA³. *P < 0.001 compared to water-injected oocytes; **P < 0.001 compared to absence of cAMP/IBMX (adjacent bar).

Mouse DRA mRNA is expressed in ileum and colon of CFTR(−/−) mice

In CFPAC-1 human pancreatic duct-like carcinoma cells (Greeley et al. 2001) and in CFT-1 human tracheal epithelial cells (Wheat et al. 2000), each derived from a cystic fibrosis patient homozygous for the ΔF508 mutation, hDRA mRNA is undetectable in Northern blots loaded with 30 μg total RNA. However, these same cell lines stably overexpressing WT CFTR cDNA express easily detectable levels of hDRA mRNA. Since hDRA mRNA levels have not been reported in cystic fibrosis patient tissue samples, we examined levels of DRA mRNA in tissues from CFTR(−/−) mice (Snouwaert et al. 1992; Freedman et al. 1999). Figure 10 shows that DRA mRNA levels in proximal colon and distal ileum varied among individual mice with the same genotype, sometimes considerably. Variation in hDRA mRNA levels has also been reported among colonic mucosa samples from individual patients (Yang et al. 1998; Lohi et al. 2002b). The mean level of DRA mRNA in proximal colon of CFTR(−/−) mice (n = 10) was 139 ± 38 % (s.e.m.) of that in CFTR(+/+) mice (n = 8). Within each experiment, proximal colon DRA mRNA levels in the two groups were statistically indistinguishable (P = 0.07 in expt 1; P = 0.18 in expt 2). The mean level of DRA mRNA in distal ileum of CFTR(−/−) mice (n = 11) was 68 ± 14 % of that in CFTR(+/+) mice (n = 9). Within each experiment, ileal DRA mRNA levels in the two groups were statistically indistinguishable (P = 0.28 in expt 1; P = 0.16 in expt 2). DRA mRNA levels in two CFTR(−/+) mice were 120 % of the mean WT level. Thus, in a mouse model of cystic fibrosis maintained post-weaning on an elemental liquid diet, DRA mRNA expression in proximal colon and ileum was maintained near WT level in the absence of functional CFTR.

Figure 10. DRA mRNA levels in CFTR(−/−) and (+/+) mice.

RNAse protection assays of mouse DRA mRNA in whole proximal colon (caecum, A and C) and in distal ileum (B and D) from individual CFTR(+/+), (−/−) and (−/+) mice. Replica digestions of RNA from colon and ileum were tested for each numbered individual. Experiments 1 (A and B) and 2 (C and D) represent separate groups of age-matched individuals. Experiment 2 is representative of three independent digestions with the same RNA samples. The 18S and 28S RNAs were intact in all samples (not shown). ³²P-end-labelled size standards mark 300 and 400 nt. S, probe protected by hybridization with unlabelled sense DRA cRNA; A, probe unprotected by hybridization with unlabelled antisense DRA cRNA.

Discussion

Mutations in the human SLC26A3/DRA gene cause congenital CLD, secondary to absence of intestinal Cl⁻ reabsorption mediated by surface enterocyte apical membrane Cl⁻-HCO₃⁻ exchange. In the current work, we have characterized hDRA-mediated anion exchange in Xenopus oocytes, and presented initial observations on the structure-function relationships of acute regulation of hDRA function. hDRA mediated Cl⁻-HCO₃⁻ exchange as judged by ion flux studies and by pH measurements with pH-sensitive microelectrode measurements and with BCECF fluorescence excitation ratio imaging. The K_1/2 of extracellular Cl⁻ was ∼6 mm for nominal Cl⁻-Cl⁻ exchange. hDRA was insensitive to acidic pH_o but inhibited by acidic pH_i, yet activated by the oocyte acidifier, NH₄⁺. Truncations of the hDRA N-terminal and C-terminal cytoplasmic domain revealed that neither is absolutely required for transport activity, whereas deletion of the STAS domain with preservation of the C-terminal 44 amino acids abolished transport function in oocytes. Whereas the N-terminal cytoplasmic domain was not required for acute regulation by pH_i or by NH₄⁺, removal of the C-terminal cytoplasmic domain abolished regulation by both stimuli. Most potent among several newly described inhibitors of DRA were the NSAIDs niflumate (ID₅₀ 7 μm) and tenidap (ID₅₀ 10 μm). Co-expression of CFTR conferred modest cAMP-sensitive stimulation upon the normally cAMP-insensitive hDRA. Mouse DRA mRNA abundance in distal ileum and proximal colon of CFTR(−/−) mice was similar to that of age-matched WT and heterozygote mice.

Anion selectivity of hDRA-mediated anion transport

Previous studies of hDRA-mediated anion transport in Xenopus oocytes have reported equivalent magnitudes of sulfate and Cl⁻ flux (Silberg et al. 1995; Moseley et al. 1999). We have observed 1000-fold higher rates of hDRA-mediated Cl⁻ transport, but failed to detect hDRA-sulfate transport. hDRA exhibits high-affinity uptake of [³⁵S]sulfate from extracellular solutions containing 2 mm Cl⁻ when expressed in Sf9 cells (Byeon, 1998) and in HEK 293 cells (Chapman et al. 2002). The reason for low sulfate transport rates by hDRA in Xenopus oocytes, even in Cl⁻-free solutions, is not clear. However, it parallels the very low sulfate transport rates by the SLC4 anion exchanger, hAE1/SLC4A1, measured in oocytes (Chernova et al. 1997). Background oocyte sulfate uptake rates in the current study were comparable to those recently reported by others (Lohi et al. 2002a; Jiang et al. 2002). A K_1/2 value for DRA-mediated extracellular sulfate uptake has not been reported in Xenopus oocytes. Sulfate affinity of DRA (or related anion exchangers) may vary in different species (Jacob et al. 2002).

hDRA-mediated oxalate transport was of low magnitude and butyrate transport was undetectable. The human disease phenotype of congenital CLD (including high stool [Cl⁻], acidic stool pH and hypochloraemic metabolic alkalosis), together with the demonstration of robust Cl⁻-HCO₃⁻ exchange in hDRA-expressing oocytes (Figs 1 and 9) and mouse DRA-transfected HEK 293 cells (Melvin et al. 1999) together strongly suggest that Cl⁻-HCO₃⁻ exchange is the major physiological function of DRA. Our estimate of K_1/2 for extracellular Cl⁻, the first reported for an SLC26 family member, is sufficiently low to account for reabsorption of most of the lumenal Cl⁻ originating in gastric mucosal secretion. hDRA is probably not the only SLC26 Cl⁻ transporter in the human intestine (Wang et al. 2002). However, the clinical consequences of DRA deficiency clearly implicate it as the major distal Cl⁻ reuptake pathway in human distal ileum and colon.

Interestingly, the Na⁺-sulfate cotransporter SLC13A1 is not expressed in the human intestine, in contrast to prominent apical expression in rat and mouse (Markovich, 2001). Thus, luminal sulfate uptake into enterocytes of human ileum and jejunum may be mediated predominantly by other sulfate transport pathways. Although hDRA protein level is unchanged in severe ulcerative colitis (Lohi et al. 2002b), the disease-associated intestinal hyperabsorption of oxalate may be mediated by another transporter(s) (Knauf et al. 2001; Wang et al. 2002; Xie et al. 2002).

Regulation of hDRA by pH_i and NH₄⁺

If the role of hDRA is primarily that of Cl⁻ reabsorption rather than pH_i regulation, why should hDRA be sensitive to pH_i? The need for intracellular HCO₃⁻ to exchange with lumenal Cl⁻ suggests that a pH_i sensor in hDRA might sense adequacy of intracellular [HCO₃⁻] for this purpose. Such a sensor could serve to down-regulate hDRA activity before pH_i is acidified secondary to substrate exhaustion. Moreover, inhibition of hDRA by acidic pH_i should foster functional coupling to the NHE3 Na⁺-H⁺ exchanger in the apical membrane (Lamprecht et al. 2002). In contrast, hDRA stimulation by NH₄⁺, which is abundant in the colonic lumen, might serve to sustain hDRA-mediated Cl⁻ reabsorption in the colon over a range of pH_i values when called for.

In the basolateral membrane of ileal and colonic surface enterocytes, the AE2/SLC4A2 Cl⁻-HCO₃⁻ exchanger undergoes similar regulation by pH_i and by NH₄⁺. Unlike the regulation of AE2 by pH_o, hDRA is insensitive to changes in pH_o. Perhaps hDRA function is insulated in this way from lumenal pH values that vary over a wider range than do the interstitial pH values to which AE2 is exposed. Alternatively, if hDRA functions in tandem with NHE3, it might rely upon the latter to sense changes in pH_o. Interestingly, pendrin/SLC26A-mediated Cl⁻-Cl⁻ exchange in Xenopus oocytes is not sensitive to either pH_i, pH_o or NH₄⁺ (A. Stewart & S. Alper, unpublished results).

Role of cytoplasmic domains in hDRA function and regulation

The first 52 amino acids of the N-terminal cytoplasmic domain were not required for continued sensitivity to inhibition by acidic pH_i or for stimulation by NH₄⁺. This observation may be relevant to possible post-translational cleavage of N-terminal residues from hDRA (Byeon et al. 1996; Lohi et al. 2002b). Interestingly, removal of the far C-terminal amino acids, which include a Type I PDZ domain recognition site and a candidate tyrosine phosphorylation site, decreased neither basal transport activity nor the responses to pH_i and NH₄⁺. In contrast, the mutant Y756F exhibited increased basal activity, suggesting a possible role for phosphorylation of this residue in the tonic negative regulation of hDRA activity.

Further substantial deletion of the C-terminal cytoplasmic tail to remove the STAS domain reduced transport and abolished these regulatory sensitivities. Two disease mutations in the STAS domain exhibited complete loss of function in Xenopus oocytes. The STAS domain is present in all characterized members of the SLC26 gene family (Aravind & Koonin, 2000), but its function remains unknown. Chapman et al. (2002) have shown that hDRA with a truncated STAS domain lacks the growth inhibition phenotype of wild-type hDRA in transfected mammalian cells. Future experiments will test the possibility that the STAS domain has a nucleotide-binding or -sensing function.

Pharmacology of hDRA inhibitors

The stilbene sensitivity of DRA has been controversial. Our results show that hDRA-mediated ³⁶Cl⁻ influx exhibits only low sensitivity to DIDS, in agreement with the low affinity inhibition of mouse DRA reported by Melvin et al. (1999) even in the absence of extracellular Cl⁻. In contrast, Silberg et al. (1995) and Moseley et al. (1999) reported 90–100 % inhibition of hDRA-mediated transport of sulfate and Cl⁻ by 0.5 mm DIDS in the presence of 100 mm bath Cl⁻. hDRA-mediated sulfate uptake in the presence of 2 mm Cl⁻ was also sensitive to 0.5 mm DIDS in Sf9 cells (Byeon et al. 1998). The reasons underlying the differences among these results are currently not clear.

The demonstration of hDRA inhibition by NSAIDs is new. They are of considerably higher inhibitory potency than the stilbenes. Additional experiments will be required to test the kinetics of inhibition, as well as to test sensitivities to drugs with specificity towards cyclo-oxygenases 1 and 2.

Regulation of hDRA by CFTR

CFTR overexpression has been shown to induce cAMP-stimulated Cl⁻-HCO₃⁻ exchange in HEK 293 cells (Lee et al. 1999). Conversely, the cAMP-stimulated component of apical Cl⁻-HCO₃⁻ exchange is absent from human duodenum (Pratha et al. 2000) and from intestine of the CFTR (−/−) mice used in this study (Grubb & Boucher, 1999). The absence of CFTR expression in epithelial cells has been shown to be accompanied by low or absent levels of hDRA mRNA, whereas the same cells which overexpress CFTR also express hDRA mRNA. In the current work we show that CFTR co-expression with hDRA can confer modest sensitivity of hDRA-mediated Cl⁻-HCO₃⁻ exchange to acute stimulation by cAMP. This interaction may be mediated by a scaffolding protein which can interact with PDZ recognition motifs in both proteins (Lamprecht et al. 2002; Ko et al. 2002). Alternatively, the interaction may be direct, or may require diffusible intracellular messengers. It is unlikely to be secondary to ionic gradients, since cAMP-activation of CFTR lowers intracellular Cl⁻, and reduces the driving force for pH_i increase upon removal of bath Cl⁻. Similar cAMP-stimulated Cl⁻-HCO₃⁻ exchange activity has been noted in duodenal enterocytes (Jacob et al. 2002), T84 cells (Lee et al. 1999b), and epithelial cells from pancreatic duct (Wheat et al. 2000) and trachea (Greeley et al. 2001). However, the DIDS-sensitivity of most of the cAMP-stimulated Cl⁻-HCO₃⁻ exchange in all of these studies suggests that hDRA might not be responsible for that component. DIDS-sensitivity of Cl⁻-HCO₃⁻ exchange was a function of CFTR expression level in HEK 293 cells (Lee et al. 1999), suggesting graded induction of distinct endogenous activities or regulators of a single activity. The SLC26A6/PAT1 Cl⁻-HCO₃⁻ exchanger (Lohi et al. 2000; Waldegger et al. 2001) is a DIDS-sensitive candidate mediator of this activity (Wang et al. 2002; Xie et al. 2002; Ko et al. 2002).

Interestingly, CFTR expression alone also induced apparent Cl⁻-HCO₃⁻ exchange in Xenopus oocytes (Fig. 9). In some respects this observation resembles those of Lee et al. in HEK 293 cells (1999). Curiously, however, this activity in oocytes is itself not stimulated by cAMP, whereas Cl⁻ conductance measured by two-electrode voltage clamp and unidirectional ³⁶Cl⁻ efflux in room air are both greatly stimulated. This CFTR-associated Cl⁻-HCO₃⁻ exchange activity may represent regulation by CFTR of a latent endogenous Cl⁻-HCO₃⁻ exchanger of the oocyte, or it may be mediated by CFTR itself (Choi et al. 2001). Thus, Fig. 9 might also be explained (if less conventionally) as stimulation by hDRA of CFTR-mediated ‘Cl⁻-HCO₃⁻ exchange’.

While this paper was in editorial review, Ko et al. (2002), using HEK 293 cells, reported that CFTR coexpression activates and confers forskolin-sensitivity on Cl⁻-HCO₃⁻ and Cl⁻-OH⁻ exchange activities mediated by mouse DRA.

Modulation of hDRA as a potential adjunct therapy in cystic fibrosis

If in the intestine of cystic fibrosis patients with intestinal or pancreatic disease absorption of Cl⁻ continues in the absence of Cl⁻ secretion, then inhibition of the persistent Cl⁻ reabsorption may prove useful as an adjunct therapy to the widely desired induction of Cl⁻ secretion. Our newly described inhibitors include drugs in current clinical use for other indications. However, use of hDRA inhibitors would be futile if, as in cultured epithelial cells lacking CFTR, hDRA is not detectably expressed in cystic fibrosis patients. We therefore examined mouse DRA expression in distal ileum and proximal colon of the UNC CFTR(−/−) mouse. At least in this particular animal model of cystic fibrosis, DRA mRNA levels were minimally, if at all, reduced compared with wild-type and heterozygote tissues. This contrasts with increased mouse DRA mRNA levels in the NHE3(−/−) mouse (Melvin et al. 1999) and with substantially decreased hDRA mRNA levels in rat and mouse colitis models (Yang et al. 1998) and in human colonic specimens from patients with severe ulcerative colitis (Yang et al. 1998; Lohi et al. 2002b; note however that these studies differed in assessment of changes in hDRA polypeptide levels). The sustained expression of DRA mRNA in CFTR(−/−) mouse ileum and colon suggests that absence of hDRA mRNA in tracheal (Wheat et al. 2000) and pancreatic epithelial cell lines (Greeley et al. 2001) from cystic fibrosis patients may thus represent an adaptation required for continued growth in conditions of cell culture (Chapman et al. 2002), one which is overcome by CFTR overexpression.

If impaired HCO₃⁻ secretion is important in cystic fibrosis, then therapeutic enhancement of DRA expression might be beneficial. However, if impaired fluid (volume) secretion is a more important pathological problem in cystic fibrosis, then use of NSAIDs to inhibit hDRA-mediated Cl⁻ reabsorption may represent a plausible adjunct therapy for the intestinal (and perhaps also the pancreatic) pathologies of cystic fibrosis. High dose ibuprofen (20–30 mg kg⁻¹ day⁻¹) has been tested in the treatment of the inflammatory component of cystic fibrosis lung disease with modestly encouraging results (Chmiel et al. 2002). At still higher concentrations, however, ibuprofen exposure of T84 cells acutely inhibits CFTR (Devor & Schultz, 1998) and chronically lowers CFTR mRNA levels (Tondelier et al. 1999), perhaps by inhibition of NF-KB (Brouillard et al. 2001.) Niflumate and tenidap also elicit side-effects in adults. Thus, these anti-inflammatory drugs may serve only as lead compounds for the development of additional inhibitors or modulators of hDRA. These results nonetheless provide additional support for possible use of NSAIDs or their next-generation congeners in the pharmacotherapy of the intestinal disorders of cystic fibrosis. Pharmacological modulation of other members of the SLC26 gene family expressed in airway epithelia may similarly serve as adjunct therapy of cystic fibrosis lung disease.