Cloning and Expression of a Chloride-dependent Na+-H+ Exchanger*

Electroneutral Na+-H+ exchange is present in virtually all cells, mediating the exchange of extracellular Na+ for intracellular H+ and, thus, plays an important role in the regulation of intracellular pH, cell volume, and transepithelial Na+ absorption. Recent transport studies demonstrated the presence of a novel chloride-dependent Na+-H+ exchange in the apical membrane of crypt cells of rat distal colon. We describe the cloning of a 2.5-kb full-length cDNA from rat distal colon that encodes 438 amino acids and has six putative transmembrane spanning domains. Of the 438 amino acids 375 amino acids at the N-terminal region are identical to Na+-H+ exchange (NHE)-1 isoform with the remaining 63 amino acids comprising a completely novel C terminus.In situ hybridization revealed that this transcript is expressed in colonic crypt cells, whereas Northern blot analysis established the presence of its 2.5-kb mRNA in multiple tissues. Despite its much smaller size compared with all other known Na+-H+ exchange isoforms, NHE-deficient PS120 fibroblasts stably transfected with this cDNA exhibited Na+-dependent intracellular pH recovery to an acid load that was chloride-dependent and inhibited both by 5-ethylisopropylamiloride, an amiloride analogue, and by 5′-nitro-2-(3-phenylproplyamino)benzoic acid, a Cl−channel blocker, but only minimally affected by 25 μm3-methylsulfonyl-4piperidonbenzoylguanidine, an NHE-1 and NHE-2 isoform inhibitor. In contrast to other Na+-H+ exchange isoforms in colonic epithelial cells, chloride-dependent Na+-H+ exchange mRNA abundance was increased by dietary sodium depletion. Based on these results we predict that chloride-dependent Na+-H+ exchange represents a new class of Na+-H+ exchangers that may regulate ion transport in several organs.

. At least seven NHE isoforms have been cloned to date and functionally expressed with a length ranging from 669 to 898 amino acids with 10 -12 putative transmembrane domains (1)(2)(3)(4)(5). NHE-1 isoform is ubiquitous and linked to intracellular pH (pH i ) and cell volume regulation, whereas other isoforms (e.g. NHE-2 and NHE-3) are present solely in epithelial cells (6,7). The NHE-1 isoform has been identified on the basolateral membrane of epithelial cells and on the plasma membrane of non-polar cells functioning as a "housekeeper," whereas the NHE-3 isoform is expressed predominantly on the apical membrane of polarized epithelial cells and has been linked to transepithelial sodium-dependent fluid absorption and pH i regulation (6,7). Consequently, loss of NHE1 or NHE3 function has a severe impact on cellular and organ functions (8,9).
Intestinal electroneutral Na ϩ absorption is the result of an apical membrane NHE and has been linked to the NHE-3 isoform whose message and protein are present in surface but not crypt cells (7). The long-standing model of fluid and electrolyte movement in the large and small intestine is that fluid absorptive processes are present in surface/villous cells, whereas secretory ones are localized to crypt cells (10). To study crypt cell function directly in the rat distal colon, we recently established methods to perform microperfusion of colonic crypts adapting methods used previously with renal tubules and to prepare apical membrane vesicles (AMV) from isolated crypt cells (11,12). In microperfusion studies of isolated rat colonic crypts, we demonstrated sodium-dependent fluid absorption (11) and therefore sought the identity of this unexpected phenomenon. By using both crypt AMV and microperfusion, we demonstrated that both [H ϩ ] gradient-driven 22 Na ϩ uptake and sodium-dependent recovery of pH i from an acid load, respectively, had an absolute requirement for chloride (12)(13)(14). Additional studies established that the chloride dependence of chloride-dependent Na ϩ -H ϩ exchange (Cl-NHE) most likely involves one or more Cl Ϫ channels, including cystic fibrosis transmembrane regulator, and not a Cl Ϫ -HCO 3 Ϫ exchange, as follows: 1) Cl-NHE activity in AMV was inhibited by both Cl Ϫ channel blockers and partially by a polyclonal antibody to cystic fibrosis transmembrane regulator (13); 2) Cl Ϫ -HCO 3 Ϫ exchangers are not present in crypt AMV (15); and 3) Cl-NHE activity was not inhibited by 100 M 4,4Ј-diisothiocyanostilbene-2,2Ј-disulfonic acid, a concentration that only inhibits chloride anion exchanges but not Cl Ϫ channels (13).
As the characteristics and the kinetic properties of the colonic Cl-NHE differed from those of other known NHE isoforms (14) and NHE-2 but not the NHE-3 isoform has been localized to the apical membrane of colonic crypt cells, it was likely that chloride-dependent Na ϩ -H ϩ exchange transport protein was mediated by a previously unidentified NHE isoform. These present studies report the following: 1) the isolation of a 2,498-bp full-length cDNA from colonic crypt cells that consists of a 5Ј end that is identical to the NHE-1 isoform and a completely novel 3Ј end segment; 2) the stable expression of this cDNA in NHE-deficient PS120 cells that manifested sodiumdependent pH i recovery that was chloride-dependent and inhibited by both 5-ethylisopropylamiloride (EIPA) and 5Ј-nitro-2-(3-phenylproplyamino)benzoic acid (NPPB); 3) the upregulation of Cl-NHE mRNA by dietary sodium depletion; and 4) the localization of its mRNA in colonic crypt cells and its expression in multiple tissues. The identification of this unique protein with wide distribution in several tissues may provide an explanation for Na ϩ -and Cl Ϫ -coupled fluid and electrolyte movement that has not been adequately elucidated by existing transport proteins. Thus, these studies indicate the identification of a new class of NHE proteins that are widely expressed and raise the possibility of their importance in cellular homeostasis in several organs.

EXPERIMENTAL PROCEDURES
Male Sprague-Dawley rats (200 -210 g body wt, Charles River Laboratories, Wilmington, MA) were used in these experiments and were fed a commercial rat diet. The rats were allowed free access to water. One group of rats was fed a sodium-free diet for 7 days, as described previously (16), to produce a secondary increase in serum aldosterone levels (17).
Preparation of Distal Crypt Cells-Crypt cells of the rat distal colon were prepared, as described previously (11,12). Briefly, colonic crypts were isolated by a modified rapid calcium chelation method (12). Isolated crypt and surface cells were isolated following incubation of everted colonic segments, as described previously (12). To establish the relative purity of surface and crypt cells, ouabain-sensitive and ouabain-insensitive H,K-ATPase activities were determined (18). Enrichment of ouabain-sensitive H,K-ATPase activity with a relative absence of ouabain-insensitive H,K-ATPase activity provided evidence of a predominant crypt cell preparation with a relative absence of surface cells. This method resulted in crypt cell preparations that were 10% contaminated by surface cells (18).
Cl-NHE Cloning Strategy-Total RNA and mRNA from crypt cells were prepared by standard methods. SuperScript preamplification system (Invitrogen) was strictly followed to prepare first strand cDNA using 5 g of mRNA, oligo(dT), and random hexamer primers in separate reactions. Negative control reactions were also performed without reverse transcriptase. Five microliters of first strand cDNA products were used as a template for the PCR in a 50-l reaction volume using sense and antisense primers designed from highly conserved F and J membrane spanning domains of NHE1-NHE4 (19) (sense primer, GATCTCAGCTGTGGACCCTGTGCT, and antisense primer, GCCCAT-GAAGATGAAGATGAGGGT). The following PCR program was used: 30 cycles of 94°C for 30 s, 50°C for 45 s, 72°C for 1 min, and a final extension of 5 min at 72°C using PCR system 2400 (PerkinElmer Life Sciences). The expected 500-bp reverse transcriptase-PCR products were obtained, subcloned into PCRII vector (CLONTECH), and sequenced in the Yale Sequencing Facility. Sequencing information of one clone of the reverse transcriptase-PCR products revealed significant homology with rat NHE-1 isoform that was chosen to screen a rat colonic cDNA library that yielded a single positive clone. This clone did not have a start codon but had a stop codon within 200 bp of the novel segment that extended from NHE-1 isoform sequence. A 5Ј-rapid amplification of cDNA ends (RACE) procedure was performed to isolate the 5Ј end of this clone, according to the manufacturer's recommendations (CLONTECH). A gene-specific antisense primer was used for the RACE procedure that was designed based on information obtained from the screening of the 5Ј end of the positive clone. The resulting RACE products were subcloned into PCRII vector (CLONTECH) and sequenced. The sequence revealed a continuous clone that extended toward the 5Ј end with both a start codon and a 5Ј-noncoding region. Full-length cDNA (GenBank TM accession number AF462063) was constructed by performing PCR using the primers designed and templates from both the screening and RACE products.
Northern blot hybridization was performed, as previously described (16), using mRNA prepared from different rat tissues with the 32 Plabeled full-length coding region or the 589-bp novel portion of the putative chloride-dependent Na ϩ -H ϩ exchange (Cl-NHE) cDNA as a probe. Colonic mRNA was isolated from a purified colonic epithelial cell preparation, as described previously (16). Northern blot hybridization was also performed using mRNA isolated from colonic epithelial cells of normal mice and from a segment of sigmoid colon obtained at operation for diverticulitis with the 32 P-labeled 189-bp novel portion of the open reading frame of Cl-NHE cDNA as a probe.
In Situ Hybridization-The 589-bp fragment of the novel NHE was subcloned into the PCR II vector (CLONTECH) and linearized, and antisense and sense cRNA probes were prepared using [ 33 P]UTP, T7 RNA polymerase, and SP6 RNA polymerase, respectively, using an in vitro transcription kit (Promega), according to the manufacturer's recommendations. In situ hybridization was performed using the singlestranded uniformly 33 P-labeled RNA probe on a cryostat section of rat distal colon according to the method described by Hogan et al. (20).
Cell Culture and Preparation of Stable Cell Lines-The complete coding region of the novel NHE cDNA was subcloned into the pIRES-EGFP vector (CLONTECH). PS120 cells lacking endogenous NHE (21) (kindly provided by Dr. Mark Donowitz) were transfected with pIRES-EGFP-NHE plasmid using Superfect reagent (Qiagen); a stable cell line was prepared by selecting and growing the cells in a growth medium containing G418 at a concentration of 700 g/ml for 4 weeks. Successful transfection and selection were confirmed by the presence of green fluorescence in all the cells under confocal microscopy.
pH i Measurements-Cl-NHE activity was measured in mock-transfected and stably transfected PS120 fibroblasts 3-4 days after seeding on coverslips. Cells were loaded with the pH-sensitive dye 2Ј,7Јbis(carboxyethyl)-5,6-carboxyfluorescein (10 M, 20 min at 37°C) and mounted in a chamber for superfusion (3 ml/min flow rate, 37°C). Intracellular pH (pH i ) was continuously monitored, and all cells were calibrated using a modified high K ϩ /nigericin technique, as described previously (21). All experiments were performed in the nominal absence of bicarbonate. The initial solution was a HEPES-buffered Ringer solution as follows(in mM): 125 NaCl, 3 KCl, 1 CaCl 2 , 1.2 MgSO 4 , 2 KH 2 PO 4 , 32.2 HEPES, pH 7.4. Cells were acidified using the NH 4 Cl (20 mM) prepulse technique and washed into a Na ϩ -or Na ϩ /Cl Ϫ -free solution (Na ϩ was replaced by N-methyl-D-glucamine, whereas Cl Ϫ was replaced by gluconate). The rate of Cl-NHE activity was determined from the initial steep pH i alkalinization rate upon readdition of Na ϩ or Na ϩ and Cl Ϫ in the absence or presence of inhibitors. Rates were calculated at the same initial pH i for all cells studied. NPPB and 3-methylsulfonyl-4-piperidonbenzoylguanidine (HOE694) were kindly provided by Aventis, Frankfurt, Germany.
All data were tested for significance using the unpaired Student's t test, and only results with p Ͻ 0.05 were considered as statistically significant.

RESULTS
The cloning strategy to identify a putative Cl-NHE cDNA was based on performing reverse transcriptase-PCR using primers that were designed on the basis of the J and F membrane spanning domains that are conserved in the NHE-1-4 isoforms (19), and mRNA was isolated from crypt cells of normal rat distal colon. The PCR products were subcloned and sequenced; the sequence of one of these PCR products had significant homology to the rat NHE-1 sequence. This cDNA was used to screen a rat colon cDNA library that yielded a single positive clone that represented a 1.9-kb cDNA consisting of a 700-bp fragment identical to the NHE-1 and 1.2-kb novel fragment with a stop codon within the first 200 bp. 5Ј-RACE was performed to clone the 5Ј end and obtained a cDNA fragment with a start codon. A 2,498-bp full-length cDNA with a 1,314-bp open reading frame, a 5Ј-noncoding sequence of 360 bp, and a 3Ј-noncoding sequence of 824 bp were constructed by PCR using the 1.9-kb cDNA and the 5Ј-RACE product. The completely novel fragment is 1,013 bp representing the 189-bp open reading frame and 3Ј-non-coding sequence of 824 bp. The fragment that was homologous to a portion of NHE-1 isoform is 1,485 bp representing a 1,125-bp coding region and a 360-bp 5Ј non-coding sequence. The nucleotide sequence of the putative Cl-NHE cDNA predicts a protein of 438 amino acids with a calculated molecular mass of ϳ50 kDa. Fig. 1 provides the deduced amino acid sequence of the putative Cl-NHE with a comparison to the other NHE isoforms known to date. The N-terminal portion of the coding region has 375 amino acids that are identical to that of NHE-1 isoform, whereas the C-terminal segment of 63 amino acids is completely novel. The novel 63-amino acid region has three poten- tial phosphorylation sites at Ser-437, Thr-407, and Thr-426; there is also a potential glycosylation site at Thr-428. Hydropathy plot analysis suggested six putative transmembrane domains (Fig. 2a), whereas phylogenetic tree analysis demonstrates that this novel NHE is closely related to the NHE-1 isoform (Fig. 2b).
To confirm that this cDNA encodes Cl-NHE whose functional activity has been identified in crypt but not in surface cells (12), in situ hybridization studies were performed using a 589-bp novel portion of the cDNA that consists of 189-bp open reading frame plus 400 bp of 3Ј-noncoding region. These studies demonstrated that this transcript is expressed predominantly in crypt cells of the rat distal colon (Fig. 3). The control study with the sense probe did not reveal evidence of expression in the crypt cells.
Northern blot analyses were performed with both the complete coding region and the 589-bp novel fragment as probes. When the complete coding region was used as a probe, hybridization was observed with both 4.8-(NHE-1 isoform) and 2.5-kb (the novel NHE) transcripts in mRNA isolated from epithelial cells of proximal and distal colon (Fig. 4B), an observation that is consistent with the presence of both a NHE-1 isoform segment and a novel segment in the coding region of this putative Cl-NHE cDNA. A Northern blot analysis that used the identical 589-bp novel fragment as a probe identified only the 2.5-kb transcript (Fig. 4A). This 2.5-kb transcript was present in several tissues besides distal colon including proximal colon, lung, liver, kidney, and heart indicating that the putative Cl-NHE mRNA is widely distributed. To assess whether Cl-NHE mRNA was present in other species Northern blot analyses were performed with mRNA prepared from normal mice and human sigmoid colonic mucosa using the novel C-terminal open reading frame (189 bp) fragment as a probe. Fig. 4C reveals that a 2.5-kb transcript was present in both mice and humans.
Prior studies (22) have demonstrated that dietary sodium depletion and aldosterone markedly reduces both colonic apical membrane NHE activity and NHE isoform mRNA abundance in colonic epithelial cells. Therefore, the effect of dietary sodium depletion on Cl-NHE mRNA abundance was determined in Northern blot analyses using the 589-bp probe and mRNA prepared from colonic epithelial cells from normal and dietary sodium-depleted rats. In these studies Cl-NHE mRNA abundance from dietary sodium-depleted rats was substantially increased (Fig. 5). Prior studies (14) had shown that Cl-NHE activity (as evidenced by [H ϩ ] gradient-stimulated 22 Na ϩ uptake by crypt AMV) was significantly increased by dietary sodium depletion.
In view of the unusually short sequence and its partial homology to NHE-1, it was critical to determine whether this small cDNA encoded for a protein with functional NHE activity. Therefore, PS120 fibroblasts lacking endogenous NHE function (20) were stably transfected with the putative Cl-NHE cDNA. pH i recovery after an intracellular acid load that had been induced by a NH 3 /NH 4 Cl prepulse was measured in the nominal absence of bicarbonate using 2Ј,7Ј-bis(carboxyethyl)-5,6-carboxyfluorescein (23). Untransfected PS120 fibroblasts showed no pH i recovery both in the presence or absence of sodium (0.001 Ϯ 0.005 units pH/min) (Fig. 6a) confirming the absence of NHE activity in these cells. In transfected cells in the absence of Cl Ϫ , only minimal sodium-dependent recovery of pH i was seen (Fig. 6b). However, when both Na ϩ and Cl Ϫ were present, there was a significant increase in pH i following an acid load (Figs. 6b and 7). Thus, sodium-dependent pH i recovery to an acid load in these cells transfected with the putative Cl-NHE cDNA has a requirement for Cl Ϫ .
The effect of both amiloride and its analogues and the nonspecific Cl Ϫ channel blocker, NPPB, on the expressed sodium/ chloride-dependent pH i recovery in these PS120 cells was also FIG. 2. a, hydropathy plot analysis. Hydropathy profile of Cl-NHE was determined according to the algorithm of Kyte and Doolittle using windows of 11 amino acids. Positive values correspond to hydrophobic segments, and negative values are indicative of hydrophilic segments. This analysis demonstrates this cDNA has six transmembrane domains. b, phylogenetic tree analysis. Phylogenetic tree analysis was generated by amino acid comparison of NHE-1 to NHE-5 isoforms and indicates that the Cl-NHE isoform is closely related to NHE-1 isoform. NHE-3 and NHE-5 isoforms are the least related isoforms to Cl-NHE.
examined. The sodium/chloride-dependent pH i recovery to an acid load was almost completely blocked by a low concentration of the amiloride derivative EIPA (Ϫ95 Ϯ 2%) (Fig. 8). Similar to the results in colonic crypt AMV, sodium/chloride-dependent recovery of pH i was relatively insensitive to amiloride (10 M, 114 Ϯ 7%, and 100 M, Ϫ77 Ϯ 3% of control). Twenty five M HOE694, at which concentration HOE694 is a specific inhibitor of the NHE-1 isoform, did not inhibit sodium/chloride-dependent pH i recovery to an acid load.
Because Cl-NHE activity in both colonic AMV and during crypt microperfusion studies is inhibited by NPPB, a nonspecific Cl Ϫ channel blocker, an additional study was performed with 500 M NPPB that significantly reduced sodium/chloridedependent pH i recovery to an acid load by Ϫ59 Ϯ 3%. Thus, the pharmacological profile distinguishes the novel Cl-NHE from all other known NHE isoforms and particularly from the closely related NHE-1 isoform (3,14). More importantly, the functional and pharmacological properties of this novel NHE isoform closely resemble those of the recently described Cl-dependent Na ϩ -H ϩ exchange in apical membranes of colon crypts strongly suggesting that this newly identified cDNA encodes the transport protein expressing Cl-NHE function in apical membranes of colonic crypts (12)(13)(14). DISCUSSION The classic model of fluid movement in the small and large intestine is that absorptive processes are present in surface/ villous cells and secretory processes in crypt cells (10). In addition, NHE-3 isoform has been linked to transepithelial electroneutral sodium absorption and fluid absorption following the demonstration that 1) glucocorticosteroids increase both NHE-3 isoform message and protein in rabbit ileum; 2) neither Na ϩ -H ϩ exchange nor NHE-3 isoform message and protein are present in the rabbit distal colon (6, 7); and 3) NHE-3 deficient, Poly(A) ϩ RNA was prepared from several adult rat tissues, and equal amounts (3 g) of mRNA were analyzed by Northern blot hybridization using a 589-bp (189-bp fragment of the novel portion of the coding region and 400-bp fragment of the 3Ј-non-coding region) cDNA as a probe. Cl-NHE mRNA was expressed in several tissues including distal colon, proximal colon, stomach, small intestine, liver, lung, heart, kidney, brain, testes, and spleen. B, Northern blot analyses with distal and proximal colonic mRNA. Poly(A) ϩ RNA was prepared from epithelial cells of distal and proximal colon of adult rats, and equal amounts (3 g) of mRNA were analyzed by Northern blot hybridization using as a probe the cDNA corresponding to the entire coding region of the putative Cl-NHE cDNA. There was hybridization to both 2.5-(Cl-NHE mRNA) and 4.8-kb (NHE-1 mRNA) transcripts. C, Northern blot analyses with mRNA from mice (lane 1) and human (lane 2) colonic mucosa. Poly(A) ϩ RNA was prepared from epithelial cells and mucosal scrapings, respectively, and equal amounts (3 g) of mRNA were analyzed by Northern blot hybridization using as a probe the 189-bp fragment of the novel portion of the coding region. Cl-NHE was expressed in both mice and human distal colon.
FIG . 5. Northern blot analysis of mRNA from normal (lanes 1  and 2) and dietary sodium-depleted (lanes 3 and 4) rats. mRNA was prepared from crypt epithelial cells of distal colon. Equal amounts (3 g) of mRNA were analyzed. Northern blot analysis was performed with the 589-bp cDNA fragment as a probe. Cl-NHE mRNA abundance was substantially increased by dietary sodium depletion.
but not NHE-2-deficient mice have diarrhea (8). Consistent with the previously postulated absence of fluid absorption in crypts (10), NHE-3 message and protein have been localized to surface but not crypt cells (7). However, recent studies from our laboratories (11) demonstrated sodium-dependent fluid absorption in rat colonic crypts. Thus, the demonstration of sodiumdependent fluid absorption was unexpected and led to studies to explore for the presence of other sodium transport processes in colonic crypt apical membranes. Subsequent studies (24) revealed that sodium-dependent fluid absorption in crypt microperfusion studies was chloride-dependent and inhibited both by EIPA and by NPPB, a nonspecific Cl Ϫ channel blocker. In addition, the inhibitory action of both EIPA and NPPB was observed only in the presence of Na ϩ and Cl Ϫ , respectively. As these characteristics of sodium-dependent fluid absorption parallel those of Cl-NHE activity when determined by 22 Na uptake by crypt apical membrane vesicles or by sodium-dependent pH i recovery during crypt microperfusion studies (12,13), it is likely that Cl-NHE is the transport process responsible for constitutive fluid absorption in colonic crypts. This Cl-NHE isoform does not represent a phenomenon restricted to rodents because Cl-NHE mRNA was identified in both colonic mucosal cells from normal mice and humans (Fig. 4C). These latter observations were paralleled by the demonstration of Cl-NHE activity in the distal colon crypt of mice and humans by the demonstration of sodium/chloride-dependent recovery of pH i to an acid load. 2 These present studies establish that Cl-NHE is a novel membrane transport protein that resembles only a portion of NHE-1 isoform. Of the 438 amino acids of the Cl-NHE open reading frame the N-terminal 375 amino acids are identical to that of NHE-1 isoform, whereas the 63-amino acid Cterminal segment is completely unique. This Cl-NHE protein has only six putative transmembrane spanning domains compared with 10 -12 putative transmembrane domains of other known NHE isoforms. Because structure-function studies of the NHE-1 isoform indicate that NHE-1 isoform is not functional when the C-terminal end is truncated to 515 amino acids or less (25), it is unlikely that NHE activity would be present if the N-terminal 375 amino acids were expressed in PS120 cells. In contrast, the demonstration that the Cl-NHE protein is functional when expressed in PS120 cells (see Figs. 6 -8) suggests that the 63-amino acid C-terminal fragment is essential for the observed Cl-NHE activity in the present study. Thus, this 63-amino acid peptide appears to be critical for both sodium and proton transport as well as for the observed chloride dependence. Earlier studies (13) have suggested the chloride dependence of Cl-NHE activity may involve a Cl Ϫ channel. Whether the chloride dependence is an 2 J. Geibel, unpublished observations. In untransfected fibroblasts no sodium-dependent pH i recovery was observed after intracellular acidification with an NH 3 /NH 4 Cl prepulse, whereas an increase in pH i was observed in Cl-NHE-transfected fibroblasts following the re-addition of Na ϩ and Cl Ϫ to the bath solution.
FIG. 7. Summary of sodium-dependent pH i recovery to an acid load. pH i recovery (mean Ϯ S.E.) in untransfected (n ϭ 90) and Cl-NHE transfected fibroblasts (n ϭ 258) in the absence or presence of Na ϩ or Cl Ϫ , respectively, demonstrates that both Na ϩ and Cl Ϫ are required for the function of Cl-NHE.
FIG. 8. Effect of amiloride and amiloride analogues on Na ؉ / Cl ؊ -dependent pH i recovery to an acid load in Cl-NHE-transfected PS120 fibroblasts. The protocol used in these studies is identical to that presented in Fig. 6. Fibroblasts were incubated in a Na-Ringer solution, exposed to a NH 3 /NH 4 Cl prepulse, and then washed into a Na-Ringer solution containing one of the following inhibitors that was added to the lumen side. The inhibitors used in these studies included the following: (a), 10 M amiloride, a specific inhibitor of ENaC; (b), 100 M amiloride, an inhibitor of most NHE isoforms; (c), 10 M EIPA, an amiloride analogue, that also inhibits all NHE isoforms; and (d), 25 M HOE694, an NHE-1 isoform-specific inhibitor (n ϭ 80 -100/data points).
intrinsic property of Cl-NHE or requires the presence of associated Cl Ϫ channels that could be linked to Cl-NHE via the novel C terminus requires clarification. In addition, future studies are required to establish whether Cl-NHE is a spliced variant of previously identified NHE isoforms (e.g. NHE1 isoform) or is a new NHE isoform.
The novel properties of Cl-NHE compared with other NHE isoforms previously identified in colonic and non-colonic epithelial cells are the absolute dependence of chloride for activity and the inhibition of Cl-NHE activity by NPPB. In native colonic tissue Cl-NHE activity was manifested by [H ϩ ]-labeled gradient stimulation of 22 Na ϩ uptake by AMV, sodium-dependent recovery of pH i to an acid load, and sodium-dependent fluid absorption (12,13,24). NPPB inhibited these three parameters of Cl-NHE activity at concentrations between 10 and 500 M.
In the present study sodium-dependent recovery of pH i from an acid load was almost completely chloride-dependent and was inhibited by 60% by 500 M NPPB. This relatively reduced sensitivity of the expressed Cl-NHE in PS120 cells may represent different properties of intrinsic Cl Ϫ channels in native colonic tissue and PS120 cells.
These present studies also provide evidence that the regulation of Cl-NHE by dietary sodium depletion differs from its regulation of other NHE isoforms present in colonic epithelial cells. Increased levels of aldosterone as a result of either subcutaneous aldosterone infusion or dietary sodium depletion both induce electrogenic sodium absorption via the epithelial Na ϩ channel ENaC and inhibit electroneutral Na ϩ -Cl Ϫ absorption in rat distal colon as a result of down-regulating both Na ϩ -H ϩ and Cl Ϫ -HCO 3 Ϫ exchanges (15,17). Subsequent studies revealed that aldosterone inhibited the activity, message, and protein of both NHE-2 and NHE-3 isoforms without altering NHE-1 function in colonic epithelial cells (22). In contrast, in experiments that characterized Cl-NHE activity in apical membranes prepared from crypt cells, Cl-NHE activity was increased by 60% by dietary sodium depletion (14). The present study establishes that this increase in Cl-NHE activity is paralleled by an up-regulation of Cl-NHE mRNA abundance (Fig.  5) suggesting that aldosterone regulates Cl-NHE at a transcriptional level.
There have been a few other reports of chloride dependence of NHE function but are dissimilar to the present series of observations (26,27). In rat mesangial cells Miyata et al. (26) demonstrated that the NHE response to hyperosmolar contraction required chloride, whereas in contrast, sodium-dependent recovery of pH i to an acid load did not require chloride. Studies of Cl-NHE in colonic crypts have demonstrated chloride dependence of sodium-dependent recovery of pH i to an acid load but have not as yet examined the role of Cl-NHE in cell volume regulation. In AP-1 cells transfected with NHE-1, -2, and -3 isoforms, Aharonovitz et al. (27) examined sodium-dependent pH i recovery to an acid load and observed a varying degree of chloride dependence. However, these investigators used NO 3 and SCN as their chloride substitute and did not adequately exclude an inhibitory effect on NHE isoform activity by NO 3 or by SCN. Some reports (28,29) had provided evidence of a minor role for chloride in NHE function in erythrocytes.
Several observations suggest that Cl-NHE may also be expressed in several other tissues and might provide a mechanism for Na ϩ and Cl Ϫ transport processes that have not been adequately explained by previously identified transport proteins. First, Northern blot analyses (shown in Fig. 4A) using the Cl-NHE-specific probe revealed expression in several epithelial and non-epithelial tissues including proximal colon, heart, lung, kidney, and liver. Second, perfusion studies of proximal renal tubule in NHE-2, NHE-3, and NHE-2/ NHE-3 knockout mice that were performed in the presence of luminal chloride demonstrated sodium-dependent proton secretion that was inhibited by low concentrations of EIPA (30). These findings of an unidentified NHE representing ϳ50% of total NHE activity in proximal tubule are consistent with the potential presence of Cl-NHE. Third, similar observations were reported from perfusion studies of the pancreatic duct in which sodium-dependent bicarbonate absorption (i.e. NHE) was also studied in NHE-2, NHE-3, and NHE-2/NHE3 knockout mice (31). These studies concluded that 55% of total sodium-dependent bicarbonate absorption was inhibited by 50 M HOE694 but was not due to either NHE-2 or NHE-3 isoforms. As these studies were performed in the presence of luminal Cl, Cl-NHE could account for these observations. This unexplained sodium-dependent transport process in the pancreatic duct was inhibited by cAMP. Although the effect of cyclic AMP on Cl-NHE isoform is not known as yet, cyclic AMP inhibits NHE-2 and NHE-3 isoforms but does not affect NHE-1 isoform (32).
In conclusion, these studies suggest that Cl-NHE may represent the molecular basis of both chloride-dependent Na ϩ -H ϩ exchange and sodium-dependent fluid absorption in colonic crypts. In addition, recent observations indicate that Cl-NHE is widely distributed in multiple organs (Fig. 4A) and is present in the colon of at least three species (i.e. rats, mice, and humans) (Fig. 4C). Cl-NHE has a wide tissue distribution and may be the transport mechanism responsible for one or more Na ϩ and Cl Ϫ transport processes that have not been adequately explained by existing transport proteins and, thus, may be important for cell and whole body electrolyte and volume homeostasis.