HCO secretion is the most important defense mechanism against acid injury in the duodenum. However, the identity of the transporter(s) mediating apical HCO secretion in the duodenum remains unknown. A family of anion exchangers, which include downregulated in adenoma (DRA or SLC26A3), pendrin (PDS or SLC26A4), and the putative anion transporter (PAT1 or SLC26A6) has recently been identified. DRA and pendrin mediate Cl−/base exchange; however, the functional identity and distribution of PAT1 (SLC26A6) is not known. In these studies, we investigated the functional identity, tissue distribution, and membrane localization of PAT1. Expression studies in Xenopus oocytes demonstrated that PAT1 functions in Cl−/HCO exchange mode. Tissue distribution studies indicated that the expression of PAT1 is highly abundant in the small intestine but is low in the colon, a pattern opposite that of DRA. PAT1 was also abundantly detected in stomach and heart. Immunoblot analysis studies identified PAT1 as a ∼90 kDa protein in the duodenum. Immunohistochemical studies localized PAT1 to the brush border membranes of the villus cells of the duodenum. We propose that PAT1 is an apical Cl−/HCO exchanger in the small intestine.
- bicarbonate secretion
- apical Cl−/HCO3− exchange
- putative anion transporter
- downregulated in adenoma
- chloride/formate exchanger
the duodenal epithelium secretes an alkaline, HCO -rich fluid (11). The secreted HCO is essential for protecting the duodenal mucosa against acid injury by neutralizing the acid delivered from the stomach (11, 15, 16). Functional studies have identified an apical Cl−/HCO exchanger as the main mechanism of HCO secretion in the small intestine (4,16, 18, 22, 27). This transporter mediates Cl−absorption in exchange for HCO secretion into the lumen. The apical Cl−/HCO exchanger functions in parallel with the Na+/H+ exchanger NHE3 and, as a result, is essential for the electroneutral absorption of Na+ and Cl− (16). The molecular identity of apical Cl−/HCO exchanger in the small intestine has remained unknown. Whereas DRA shows mild expression levels in the small intestine, it is highly abundant in the colon (14, 21, 25, 28) and is presumably involved in the absorption of Cl− and secretion of base (14, 21, 25).
Recent studies have identified a new class of anion exchangers, including downregulated in adenoma (DRA), pendrin (PDS), and putative anion transporter (PAT1) (10, 12, 14, 20, 28, 29, 32). None of these transporters are structurally related to the AE (AE-1, 3 and 3) family. Indeed, the homology at the amino acid level between DRA, PDS, or PAT1 and AE family members is <15% (the GenBank accession nos. NP-000333, NP-003031, and NP-005061 were used for AE-1, AE-2, and AE-3, respectively). DRA and PDS mediate Cl−/HCO exchange (21,33). DRA is abundantly expressed on the apical membranes of colonocytes, whereas PDS is expressed on the apical membranes of thyroid follicular cells and kidney cortical collecting ducts (14, 25, 26, 28, 33). Mutations in DRA cause congenital chloride diarrhea, which presents with severe diarrhea, volume depletion, and metabolic alkalosis (14). Mutations in PDS cause pendred syndrome, which is characterized by deafness, goiter, and impaired iodine organification, as evidenced by a positive perchlorate test (10). Both diseases are autosomal recessive hereditary disorders.
PAT1 was recently cloned from the pancreas based on homology to DRA and pendrin (20). PAT1 maps to chromosome 3 and encodes a 738 amino-acid protein (20). Immunohistochemical studies localized PAT1 to the apical membranes of the pancreatic duct cells and kidney tubules (20). A mouse ortholog of PAT1 was recently cloned and found to be expressed on the apical membranes of the kidney proximal tubule (19). Little is known about the functional identity of PAT1. Furthermore, the distribution and membrane localization of PAT1 in other epithelial tissues have not been studied. Accordingly, we examined the functional identity, tissue distribution, and membrane localization of PAT1 in mouse. The results indicate that PAT1 is a Cl−/HCO exchanger and is expressed on the apical domain of villous epithelial cells of small intestine.
PCR of mouse PAT1.
Mouse EST database was blast searched against the human PAT1 sequence (GenBank accession no. AF279265). An expressed sequence tag (EST) (GenBank accession no. AI747461), which is highly similar to human PAT1, was identified. On the basis of the cDNA sequence of the EST, the following oligonucleotide primers (5′-GGG AGA TTG AAG TGG AAG TGT ACA TC, and 5′-AAG GCC AGA CTG ACT GCA ATA C) were designed and used for RT-PCR on RNA isolated from mouse kidney cortex. An ∼200-bp PCR fragment was purified, which, on sequencing, was verified as mouse PAT1 and corresponded to the human nucleotides 2278–2468. The PCR product was used as a probe for Northern blot hybridization in mouse tissues.
RNA isolation and Northern blot hybridization.
Total cellular RNA was extracted from various mouse tissues including gastrointestinal tract segments, kidney, liver, heart, brain, and lung using TriReagent (6). Hybridization was performed according to Church and Gilbert (7). The membranes were washed, blotted dry, and exposed to a PhosphorImager screen (Molecular Dynamics, Sunnyvale, CA). 32P-labeled human and mouse probes were used for Northern experiments. For DRA hybridization, a 400-bp cDNA from the mouse DRA cDNA (EcoR I-EcoR I fragment) was used as a probe.
Enriched luminal membrane vesicle preparation.
For the duodenum, enriched luminal membrane vesicles were prepared by a modification of the divalent cation precipitation technique (9,17, 31). Briefly, mucosa was scraped with a glass slide from rinsed intestinal segments and homogenized in 50 mM mannitol, 2 mM Tris, pH 7.1, containing protease inhibitor cocktail (20 μg/ml leupeptin, 10 μg/ml pepstatin, 8.7 μg/ml phenylmethylsulfonyl fluoride, and 1.25 μg/ml aprotinin). CaCl2 (1 M) was added to a final concentration of 10 mM, and the homogenate was mixed for 15 min at 4°C, then spun at 2,000 g. The pellet was discarded, and the membrane vesicles present in the supernatant were pelleted at 17,500 g for 20 min, washed once in 100 mM NaCl, 50 mM Tris, pH 7.4 containing protease inhibitor cocktail, resuspended in the same buffer by passage through a 22-gauge needle, and frozen at −80°C. For isolation of membrane proteins from the proximal colon, mucosa was scraped with a glass slide from rinsed proximal colon and homogenized as above. Similar whole cell scrapings from the duodenum were used for comparison.
Cloning of human and mouse PAT1.
Full-length mouse PAT1 cDNA was cloned from the duodenum by RT-PCR. Briefly, total RNA was prepared from the duodenum, poly(A+)-selected using Oligotex latex beads (Qiagen), and then reverse transcribed at 42°C using SuperScript II RT (Life Technologies) and oligo(dT) primers. The following oligonucleotide primers were designed and used for RT-PCR: 5′-CGT CTG CAC TGC TCC CTC CTC CAT TG, and 5′-GAG TCC CAG GGC ATC CAT CCA TG (GenBank accession no. AY032863; Ref. 19). These primers encode nucleotides 45 to 2498 of mouse PAT1.
Full-length human PAT1 cDNA was cloned from cultured human pancreatic duct (CFPAC-1) cells by RT-PCR. Oligonucleotide primers (5′-ATG CCT TCA CTG TGT CTC TCT GGT CTT GCC and 5′-AAT ATG CAC CAG TTC CCT CCC TGT ACC GC) were designed based on human PAT1 sequence (GenBank accession no.AF279265).
Amplification of the human or mouse PAT1 cDNA by the PCR was performed according to Clontech Advantage 2 PCR kit protocal. Briefly, each PCR contained 5 μl cDNA, 5 μl 10× PCR buffer, 1 μl 10 mM dNTP's, 10 pmol of each primer, and 1 μl Advantage 2 polymerase mix in a final volume of 50 μl. Cycling parameters were: 95°C, 1 min; 95°C, 30 s; and 68°C, 4 min. After PCR, the product was gel purified (revealing a single band of ∼2.5 kb). Sequence analysis of the PCR products verified the sequence as human PAT1 or its mouse ortholog. The PCR products were ligated into pGEM-T easy vector for expression studies.
Synthesis of mouse or human PAT1 cRNA.
The capped PAT1 cRNAs were generated using mmessage mMACHINE kit (from Ambion) according to manufacturer's instruction. Briefly, the plasmids containing the full-length mouse or human PAT-1 cDNA were linearized, and the products were then in vitro transcribed to cRNAs, as described previously (34).
Expression of mouse or human PAT1 in Xenopus oocytes.
X. oocytes were injected with mouse or human PAT1 cRNA, as used before (34). Fifty nanoliters of cRNA (0.2–1.3 μg/μl) were injected with a Drummond 510 microdispenser via a sterile glass pipette. Intracellular pH (pHi) in oocytes was measured with the pH-sensitive fluorescent probe 2′, 7′-bis-(2-carboxyethyl)-5-(and-6)-carboxyfluorescein (BCECF) (Molecular Probes, Eugene, OR) as described previously (2, 5, 34). Oocytes were loaded with 10 μM BCECF-acetoxymethyl ester for 20 min at room temperature, transferred to a 1-ml perfusion chamber, and perfused at rate of 3 ml/min with the following solution (in mM): 63 NaCl, 33 NaHCO3, 2 KCl, 1.8 CaCl2, 1 MgCl2, and 5 HEPES. Ratiometric fluorescence measurements were performed using Attofluor digital imaging system (Attofluor, Rockwell, MD). Excitation wavelengths were 450 and 490 nm, and fluorescence emission intensity was recorded at 520 nm. Data analyses were performed using Attograph and Attoview software packages provided with the imaging system. The ratios were obtained from the submembrane region of the oocytes that were visualized with a ×10 objective. Measured excitation ratios were converted to pHi by using a calibration curve that was constructed with the high-K+/nigericin method at the end of each experiment.
To examine the Cl−/HCO exchanger activity, the chamber was first perfused with a Cl-containing solution that consisted of 63 mM NaCl and 33 mM NaHCO3 (see above). Oocytes were then switched to a Cl-free medium (63 mM Na-gluconate and 33 mM NaHCO ). All other Cl-containing chemicals (KCl, etc.) were replaced with gluconate salts. This maneuver results in cell alkalinization due to reversal of Cl−/HCO exchanger (33,35). On pHi stabilization in Cl-free medium, oocytes were switched back to the Cl-containing solution. This should result in cell acidification back to baseline due to activation of Cl−/HCO exchanger. The initial rate of cell pHi recovery was used as the rate of Cl−/HCO exchanger activity (33,35).
Immunoblotting and immunohistochemical staining of PAT1 in mouse duodenum.
Duodenum from normal mice was cut into slices and mounted on holders to form tissue blocks. The tissues were fixed in a solution containing 0.1% glutaraldehyde plus 2% paraformaldehyde in 0.1 M sodium cacodylate buffer, pH 7.20, and stored in 0.1 M cacodylate buffer, pH 7.20, at 4 °C. For immunohistochemistry, the tissue blocks were sectioned into 5 μm section, placed on slides (Fisher brand-Superfrost/Plus), and incubated at 75°C for 1 h. The immunohistochemistry studies were performed according to standard protocols. A PAT1 specific antibody, raised against the amino terminal amino acid sequence RRDYHMERPLLNQE of human PAT1 (Genebank accession no. AF279265), was applied to the slides in 1/40 dilution in PBS + 1% BSA and in the presence of saponin and incubated in a humidified chamber for 2 h at room temperature. The specificity of PAT1 antibody was demonstrated in enriched luminal membrane vesicles isolated from the duodenum, where the immunoblot analysis identified a ∼90-kDa band, which was blocked following preadsorption with the synthetic peptide (see results). The secondary antibody was applied to the slides in a dilution of 1:25. The peroxidase-antiperoxidase conjugate diluted in 1:100 in PBS + 1% BSA was applied to the slides. To develop a colored reaction product, the diaminobenzidine was used and the tissues were counterstained with Harris hematoxylin.
[32P]dCTP was purchased from New England Nuclear (Boston, MA). Nitrocellulose filters and other chemicals were purchased from Sigma (St. Louis, MO). RadPrime DNA labeling kit was purchased from GIBCO-BRL. BCECF was from Molecular Probes (Eugene, OR). mmessage mMACHINE kit was purchased from Ambion (Austin, Texas). The human multiple tissue blots were purchased from Clontech (Palo Alto, CA).
Values are expressed as means ± SE. The significance of difference between mean values were examined using ANOVA.P < 0.05 was considered statistically significant.
PAT1 mRNA expression.
To examine the distribution of PAT1 mRNA in mouse tissues, Northern blots were prepared using total RNA from several tissues and then probed with mouse PAT1 cDNA. The results were as shown in Fig.1 A: PAT1 mRNA levels are highly expressed in heart, kidney, and stomach but are low in colon and lung. A human multiple-tissue blot similarly indicates high expression levels of PAT1 in the stomach (data not shown).
The abundant expression levels in the stomach raised the possibility that PAT1 may also be expressed in the small intestine. Figure2 A shows a Northern blot prepared from RNA isolated from various segments of mouse small and large intestines and probed with PAT1 cDNA. As indicated, PAT1 mRNA levels are very abundant in small intestine (duodenum, jejunum, and ileum) but are low in the colon (cecum, proximal and distal colon). This pattern of expression is opposite that for DRA, which is predominantly expressed in colon and shows only a mild expression in the small intestine (Fig. 2 B).
Immunoblotting and immununohistochemical staining of PAT1 in the duodenum. In this series of experiments, we first examined the presence of PAT1 in luminal membrane vesicles isolated from duodenum using a specific PAT1 immune serum (see methods). As indicated in Fig. 3 A, the PAT1 immune serum labeled an ∼90-kDa band in luminal membrane vesicles isolated from the duodenum (Fig. 3 A, left). The detection of the 90-kDa band was specific, because the labeling was prevented by immune preadsorption (Fig. 3 A, right). To compare the expression of PAT1 in small and large intestines, whole cell scrapings from proximal colon or duodenum were subjected to SDS-PAGE and blotted against the PAT1 immune serum. As indicated, PAT1 is predominantly expressed in membrane proteins isolated from the duodenum (Fig. 3 B). However, the expression of PAT1 in membrane proteins from the proximal colon was faint (Fig. 3 B). The subcellular localization of PAT1 in the duodenum was examined by histochemistry. As shown in histochemical staining studies depicted in Fig. 3 C, PAT1 immune serum specifically labels the apical membrane domain of villous cells in the duodenum. Interestingly, the crypt cells were not labeled. Preimmune serum (Fig. 3 C) and preadsorbed serum (not demonstrated) did not demonstrate any labeling. These results are consistent with the localization of PAT1 in brush border membranes of the duodenum.
Functional identity of PAT1.
In these series of experiments, the functional identity of human (and mouse) PAT1 was examined using the oocyte expression system.
On the basis of structural similarity with DRA and pendrin (10,21, 29, 33), we speculated that PAT1 could function in Cl−/base exchange mode. In the first series of experiments, oocytes were injected with PAT1 cRNA, loaded with BCECF in the presence of a Cl− and HCO containing solution (see methods), and monitored for intracellular pH. The representative pHi tracings in Fig.4 demonstrate that switching to a Cl-free solution resulted in a rapid intracellular alkalinization in oocytes expressing PAT1. Switching back to the Cl-containing solution caused a gradual return of pHi to normal. Control oocytes did not demonstrate any pHi alteration in response to exposure to the Cl-free medium (Fig. 4). These results are consistent with PAT1 functioning as a Cl−/HCO exchanger. The averaged results of multiple experiments indicated that the rate of Cl−/HCO exchange activity was 0.10 ± 0.01 in oocytes expressing mouse PAT1 (n = 10) and was inhibited by ∼74% in the presence of 0.5 mM DIDS (n = 6). The Cl/base exchange activity in control oocytes was not different from zero.
In the absence of HCO , the rate of Cl−/base exchange activity (mediated via Cl−/OH− exchange) was only 0.025 ± 0.002 (n = 4) in oocytes injected with mouse PAT1 cRNA, indicating low affinity of PAT1 for OH−.
The results of the above experiments indicate that the expression of SLC26A6 (PAT1) is highly abundant in the small intestine but is low in the colon (Fig. 2 A). This pattern of expression is opposite that for DRA, which is predominantly expressed in colon but is minimally expressed in small intestine (Fig. 2 B). Immunohistochemical staining studies in the duodenum localized PAT1 to the brush border membrane domain of villus cells (Fig. 3). Expression studies in oocytes demonstrated that PAT1 functions in Cl−/HCO exchange mode (Fig. 4).
The upper gastrointestinal tract (and specifically duodenum) is exposed to an acidic chyme delivered from stomach that can achieve a pH as low as 1.5 (11, 13). One major defense mechanism for protecting the duodenal mucosa against acid injury is via secretion of bicarbonate (15, 16, 27). Several studies have examined HCO transport processes in the duodenal epithelium using isolated cells, membrane vesicles, and duodenal mucosa in vitro and in vivo. These studies have demonstrated the presence of three mechanisms for acid-base transport. An amiloride-sensitive Na+/H+ exchange (NHE1) and a DIDS-sensitive Na+/nHCO cotransporter (NBC) are located on the basolateral membrane, and a Cl−/HCO exchanger is present on the apical membrane of entrocyte (4, 8, 16, 18, 22, 23, 30) . The presence of these transporters has been verified, and their functional properties have been characterized.
The apical Cl−/HCO exchanger is the main mechanism for HCO secretion in the small intestine (11, 16, 22). This transporter absorbs luminal Cl− in exchange for HCO secretion. The apical Cl−/HCO exchanger functions in parallel with the Na+/H+ exchanger NHE3 and, as a result, is essential for the electroneutral absorption of Na+ and Cl− (11, 16). The molecular identity of apical Cl−/HCO exchanger in the small intestine has remained unknown. Whereas DRA shows mild expression levels in the small intestine, its major role is the absorption of Cl− and secretion of HCO in the colon (14, 21, 25, 28). In addition to the Cl−/HCO exchanger, several studies indicate the presence of an HCO conductive pathway that mediates bicarbonate secretion into the duodenum (4, 8, 30). Whether the bicarbonate conductive pathway is the same as cystic fibrosis transmembrane conductance regulator (CFTR) or is a distinct anion channel remains to be resolved. Both the Cl−/HCO exchanger and bicarbonate conductive pathway are downregulated in cystic fibrosis (8, 24, 30).
cDNA analysis indicates that PAT1 is closely related to a family of anion transport proteins (SLC26A) that includes the rat sulfate-anion transporter, the human diastrophic dysplasia sulfate transporter, the DRA gene, and pendrin (3, 10, 12, 14, 20, 28, 32). Sequence comparison revealed that PAT1 has ∼34% homology with DRA and pendrin, respectively, at the amino acid level (10, 20,28). Both DRA and pendrin mediate Cl−/HCO exchange (21, 33). DRA can function in more than one anion exchange mode. Indeed, several studies have suggested that DRA can function as a Cl−/HCO (21), a Cl−/OH− (25), and a sulfate/OH− exchanger (1).
PAT1 (or SLC26A6) is the first apical Cl−/HCO exchanger identified in the duodenum. This conclusion is based on functional expression experiments and immunolocalization studies (see results). Our studies further indicate that the expression pattern of PAT1 in the gastrointestinal tract is completely opposite that for DRA. Whereas DRA is predominantly expressed in the large intestine, the expression levels of PAT1 are very high in small intestine and stomach but are low in the colon. Both mouse and human PAT1 (SLC26A6) mediate Cl−/HCO exchange when expressed in oocytes (see Fig. 4 and results). Together, these results indicate that PAT1 is a major apical Cl−/base exchanger in the small intestine, whereas DRA is a major apical Cl−/base exchanger in the colon.
Whether PAT1 can function in other anion exchange modes is currently under investigation. A recent study indicates that a mouse ortholog of SLC26A6 or PAT1 is expressed in the kidney proximal tubule and can function as a Cl−/formate exchanger (19). For this reason, the mouse ortholog of SLC26A6 is named chloride/formate exchanger (CFEX) (19). With respect to carrying formate, PAT (or CFEX) is very similar to pendrin. Pendrin, which is another member of the SLC26A family (and is referred to as SLC26A4), can function in both Cl−/HCO and Cl−/formate exchange modes (33). It would be difficult to compare the rate of Cl−/HCO exchange activity with that of Cl−/formate exchange activity mediated via PAT1 (or pendrin) in oocytes or other expression systems by radioflux assay or other methods. Comparison of radiolabeled 36Cl flux in exchange for formate (Cl−/formate exchange) or bicarbonate (Cl−/HCO exchange) is fraught with complications, because the imposition of HCO gradient is only possible in the presence of a pH gradient that, if applied to formate flux experiment, would alter the formate gradient across the membrane (as formate will combine with H+ once it is transported across the oocyte membrane to form formic acid, which can then diffuse back across the membrane and change the pH gradient further). Whether the transport of formate via PAT1 in the small intestine is physiologically significant remains to be examined.
In conclusion, PAT1 or SLC26A6 (or CFEX) is expressed in the upper gastrointestinal tract, with mRNA expression predominantly limited to the small intestine and stomach. Immunohistochemical staining studies localized PAT1 to the apical domain of villus cells in the duodenum. Functional studies demonstrated that PAT1 functions in Cl−/HCO exchange mode. It is suggested that PAT1 is an apical Cl−/base exchanger in the small intestine and is mediating absorption of chloride and secretion of bicarbonate.
The authors acknowledge the excellent technical assistance of Tracy Greely and Ljiljiana Pavelic.
These studies were supported by National Institute of Diabetes and Digestive and Kidney Diseases Grants DK-54430 and DK-52821, a Merit review grant, a Cystic Fibrosis Foundation grant, and grants from Dialysis Clinic Incorporated (to M. Soleimani).
Address for reprint requests and other correspondence: M. Soleimani, Div. of Nephrology and Hypertension, Dept. of Medicine, Univ. of Cincinnati, 231 Bethesda Ave., MSB 5502, Cincinnati, OH 45267–0585 (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2002 the American Physiological Society