Portal de Periódicos da CAPES

Primary Structure, Functional Expression, and Chromosomal Localization of the Bumetanide-sensitive Na-K-Cl Cotransporter in Human Colon

1995; Elsevier BV; Volume: 270; Issue: 30 Linguagem: Inglês

10.1074/jbc.270.30.17977

1083-351X

John A. Payne, Jianchao Xu, Melanie Haas, Christian Lytle, David C. Ward, Bliss Forbush,

Amino Acid Enzymes and Metabolism

By moving chloride into epithelial cells, the Na-K-Cl cotransporter aids transcellular movement of chloride across both secretory and absorptive epithelia. Using cDNA probes from the recently identified elasmobranch secretory Na-K-Cl cotransporter (sNKCC1) (Xu, J. C., Lytle, C. Zhu, T. T., Payne, J. A., Benz, E., and Forbush, B., III(1994) Proc. Natl. Acad. Sci. 91, 2201-2205), we have identified the human homologue. By screening cDNA libraries of a human colonic carcinoma line, T84 cell, we identified a sequence of 4115 bases from overlapping clones. The deduced protein is 1212 amino acids in length, and analysis of the primary structure indicates 12 transmembrane segments. The primary structure is 74% identical to sNKCC1, 91% identical to a mouse Na-K-Cl cotransporter (mNKCC1), 58% identical to rabbit and rat renal Na-K-Cl cotransporters (NKCC2), and 43% identical to the thiazide-sensitive Na-Cl cotransporters from flounder urinary bladder and rat kidney. Similar to sNKCC1 and mNKCC1, the 5′-end of the human colonic cotransporter is rich in G+C content. Interestingly, a triple repeat (GCG)7 occurs within the 5′-coding region and contributes to a large alanine repeat (Ala15). Two sites for N-linked glycosylation are predicted on an extracellular loop between putative transmembrane segments 7 and 8. A single potential site for phosphorylation by protein kinase A is present in the predicted cytoplasmic C-terminal domain. Northern blot analysis revealed a 7.4-7.5-kilobase transcript in T84 cells and shark rectal gland and a ~7.2-kilobase transcript in mammalian colon, kidney, lung, and stomach. Metaphase spreads from lymphocytes were probed with biotin-labeled cDNA and avidin fluorescein (the cotransporter gene was localized to human chromosome 5 at position 5q23.3). Human embryonic kidney cells stably transfected with the full-length cDNA expressed a ~170-kDa protein recognized by anti-cotransporter antibodies. Following treatment with N-glycosidase F, the molecular mass of the expressed protein was similar to that predicted for the core protein from the cDNA sequence (132-kDa) and identical to that of deglycosylated T84 cotransporter (~135-kDa). The stably transfected cells exhibited a ~15-fold greater bumetanide-sensitive 86Rb influx than control cells, and this flux required external sodium and chloride. Flux kinetics were consistent with an electroneutral cotransport of 1Na:1K:2Cl. Preincubation in chloride-free media was necessary to activate fully the expressed cotransporter, suggesting a [Cl]-dependent regulatory mechanism. By moving chloride into epithelial cells, the Na-K-Cl cotransporter aids transcellular movement of chloride across both secretory and absorptive epithelia. Using cDNA probes from the recently identified elasmobranch secretory Na-K-Cl cotransporter (sNKCC1) (Xu, J. C., Lytle, C. Zhu, T. T., Payne, J. A., Benz, E., and Forbush, B., III(1994) Proc. Natl. Acad. Sci. 91, 2201-2205), we have identified the human homologue. By screening cDNA libraries of a human colonic carcinoma line, T84 cell, we identified a sequence of 4115 bases from overlapping clones. The deduced protein is 1212 amino acids in length, and analysis of the primary structure indicates 12 transmembrane segments. The primary structure is 74% identical to sNKCC1, 91% identical to a mouse Na-K-Cl cotransporter (mNKCC1), 58% identical to rabbit and rat renal Na-K-Cl cotransporters (NKCC2), and 43% identical to the thiazide-sensitive Na-Cl cotransporters from flounder urinary bladder and rat kidney. Similar to sNKCC1 and mNKCC1, the 5′-end of the human colonic cotransporter is rich in G+C content. Interestingly, a triple repeat (GCG)7 occurs within the 5′-coding region and contributes to a large alanine repeat (Ala15). Two sites for N-linked glycosylation are predicted on an extracellular loop between putative transmembrane segments 7 and 8. A single potential site for phosphorylation by protein kinase A is present in the predicted cytoplasmic C-terminal domain. Northern blot analysis revealed a 7.4-7.5-kilobase transcript in T84 cells and shark rectal gland and a ~7.2-kilobase transcript in mammalian colon, kidney, lung, and stomach. Metaphase spreads from lymphocytes were probed with biotin-labeled cDNA and avidin fluorescein (the cotransporter gene was localized to human chromosome 5 at position 5q23.3). Human embryonic kidney cells stably transfected with the full-length cDNA expressed a ~170-kDa protein recognized by anti-cotransporter antibodies. Following treatment with N-glycosidase F, the molecular mass of the expressed protein was similar to that predicted for the core protein from the cDNA sequence (132-kDa) and identical to that of deglycosylated T84 cotransporter (~135-kDa). The stably transfected cells exhibited a ~15-fold greater bumetanide-sensitive 86Rb influx than control cells, and this flux required external sodium and chloride. Flux kinetics were consistent with an electroneutral cotransport of 1Na:1K:2Cl. Preincubation in chloride-free media was necessary to activate fully the expressed cotransporter, suggesting a [Cl]-dependent regulatory mechanism. The vectorial transport of chloride across epithelia is a prominent mechanism in the maintenance of water and electrolyte homeostasis. Chloride transport is involved in reabsorption of NaCl in the thick ascending limb of the loop of Henle in mammalian kidney (1Greger R. Schlatter E. Pflgers Arch. 1981; 392: 92-94Google Scholar) and in secretion of NaCl in a diverse array of secretory epithelia, including the intestine, trachea, and parotid and the avian and elasmobranch salt glands(2Dharmsathaphorn K. Mandel K.G. Masui H. McRoberts J.A. J. Clin. Invest. 1985; 75: 462-471Google Scholar, 3Fong P. Chao A.C. Widdicombe J.H. Am. J. Physiol. 1991; 261: L290-L295Google Scholar, 4Turner R.J. George J.N. Baum B.J. J. Membr. Biol. 1986; 94: 143-152Google Scholar, 5Torchia J. Lytle C. Pon D.J. Forbush III, B. Sen A.K. J. Biol. Chem. 1992; 267: 25444-25450Google Scholar, 6Hannafin J. Kinne-Saffran E. Friedman D. Kinne R. J. Membr. Biol. 1983; 75: 73-83Google Scholar). In all of these tissues, the chloride entry into the epithelia cell is mediated by a Na-K-Cl cotransport protein, which couples the electroneutral movement of sodium, potassium, and chloride ions. In order to carry out net salt transport, the Na-K-Cl cotransporter functions in concert with three other membrane proteins: chloride and potassium channels and the sodium pump. The importance of the proper functioning of these ion transport mechanisms in chloride secretory epithelia is exemplified by the disease states of cystic fibrosis and secretory diarrhea, where there are defects in the regulation of ion transport(7Halm D.R. Frizzell R.A. Lebenthal E. Duffy M. Textbook of Secretory Diarrhea. Raven Press Ltd., New York1990: 47-58Google Scholar). The Na-K-Cl cotransporter is characterized by a sensitivity to the sulfamoylbenzoic acid "loop" diuretics, furosemide, and bumetanide and a strict ion dependence of transport (for review, see (8Haas M. Am. J. Physiol. 1994; 267: C869-C885Google Scholar)). Among the Na-K-Cl cotransporters that have been described in various cells and tissues, there is substantial variation in both molecular weight and "loop" diuretic sensitivity(9Haas M. Annu. Rev. Physiol. 1989; 51: 443-457Google Scholar), suggesting the presence of different isoforms. For example, studies using the photosensitive bumetanide analog, 4-[3H]benzoyl-5-sulfamoyl-3-(3-thenyloxy) benzoic acid, have demonstrated the specific labeling of a ~150-kDa protein in membranes from dog kidney(10Haas M. Forbush III, B. Am. J. Physiol. 1987; 253: C243-C250Google Scholar), mouse kidney(11Haas M. Dunham P.B. Forbush III, B. Am. J. Physiol. 1991; 260: C791-C804Google Scholar), and duck red blood cells(12Haas M. Forbush III, B. Biochim. Biophys. Acta. 1988; 939: 131-144Google Scholar), whereas in shark rectal gland membranes, a ~195-kDa protein is observed(13Lytle C. Xu J.-C. Biemesderfer D. Haas M. Forbush III, B. J. Biol. Chem. 1992; 267: 25428-25437Google Scholar). Secretory epithelia, such as the rectal gland and parotid gland typically show a 10-fold lower affinity for bumetanide than absorptive epithelia, such as the thick ascending limb of the loop of Henle of mammalian kidney and flounder intestine(8Haas M. Am. J. Physiol. 1994; 267: C869-C885Google Scholar). Additionally, in absorptive and secretory epithelia, the Na-K-Cl cotransporter displays distinct differences in its polarized membrane distribution. In secretory epithelia, the Na-K-Cl cotransporter is an exclusively basolateral membrane protein, whereas in absorptive epithelia it is localized to the apical membrane(14O'Grady S.M. Palfrey C. Field M. Am. J. Physiol. 1987; 253: C177-C192Google Scholar). Recently, we reported the cloning, sequencing, and expression of a cDNA encoding the basolateral Na-K-Cl cotransporter from the shark rectal gland (sNKCC1)1 1The abbreviations used are: sNKCC1basolateral Na-K-Cl cotransporter from the shark rectal glandNKCCNa-K-Cl cotransporterkbkilobase(s)bpbase pair(s)TMtransmembrane segmentHEKhuman embryonic kidneyCHAPS3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonic acid. (15Xu J.C. Lytle C. Zhu T. Payne J.A. Benz E. Forbush III, B. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 2201-2205Google Scholar). In addition, we (16Payne J.A. Forbush III, B. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 4544-4548Google Scholar) and others (17Gamba G. Miyanoshita A. Lombardi M. Lytton J. Lee W.-S. Hediger M. Hebert S.C. J. Biol. Chem. 1994; 269: 17713-17722Google Scholar) have identified the apical Na-K-Cl cotransporter from mammalian kidney (NKCC2). We proposed that NKCC1 and NKCC2 represented distinct isoforms, since they display only 61% amino acid identity and are encoded by mRNAs of different sizes and tissue specificities(15Xu J.C. Lytle C. Zhu T. Payne J.A. Benz E. Forbush III, B. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 2201-2205Google Scholar, 16Payne J.A. Forbush III, B. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 4544-4548Google Scholar). In the present report, we have utilized homology to sNKCC1 to obtain the cDNA encoding the basolateral Na-K-Cl cotransporter (hNKCC1) from a human intestinal cell line, the T84 cell. The human Na-K-Cl cotransporter displays remarkable identity to sNKCC1 within the predicted transmembrane segments (88%), yet when examined in the same expression system, hNKCC1 exhibits considerably different ion affinities compared with sNKCC1. basolateral Na-K-Cl cotransporter from the shark rectal gland Na-K-Cl cotransporter kilobase(s) base pair(s) transmembrane segment human embryonic kidney 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonic acid. A human colonic (T84 epithelial cell line) cDNA library in Uni-Zap XR was obtained from Stratagene (La Jolla, CA). This library was screened with two 32P-labeled cDNA probes derived from the shark rectal gland Na-K-Cl cotransporter (nucleotides −298 to 1235 and 1236 to 3411; see (15Xu J.C. Lytle C. Zhu T. Payne J.A. Benz E. Forbush III, B. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 2201-2205Google Scholar)). Using low stringency hybridization (34°C in 50% formamide, 5 × SSPE, 5 × Denhardt's solution, 0.1% SDS, and 100 μg/ml fish sperm DNA) and high stringency wash conditions (55°C in 0.5 × SSC and 0.1% SDS), a single positive clone (TEF 1-1) was identified, carried through two rounds of plaque purification, and successfully excised into Bluescript SK- with the helper phage R408. A second T84 cDNA library in λZAP obtained from Dr. John R. Riordan (Hospital for Sick Children; Toronto, Ontario, Canada) was screened with a 5′-end 0.7-kb XbaI-XbaI fragment of TEF 1-1. Using similar screening procedures as described above, 23 positive cDNAs were obtained in Bluescript SK- (see Fig. 1). All cDNAs were characterized by Southern blot analysis and partial sequencing. The full-length cDNA was sequenced bidirectionally by the dideoxy chain termination method (18Sanger F. Nicklen S. Coulson A.R. Proc. Natl. Acad. Sci. U. S. A. 1977; 74: 5463-5467Google Scholar) using a combination of manual sequencing with Sequenase II (U. S. Biochemical Corp.) and automated sequencing (Applied Biosystems Inc.) with synthetic oligonucleotide primers and fluorescent dideoxy terminators. Analysis of the nucleotide sequence and deduced amino acid sequence were performed with programs from the Genetics Computer Group software. The program TBLASTN (19Altschul S.F. Gish W. Miller W. Myers E.W. Lipman D.J. J. Mol. Biol. 1990; 215: 403-441Google Scholar) was used to search GenBank. Identities with other sequences are given as the fraction of residues in hNKCC1, which are found to be identical in the matched sequence. T84 cells were obtained from Dr. James L. Madara (Brigham and Women's Hospital, Boston, MA) and cultured as described previously(20Dharmsathaphorn K. McRoberts J.A. Mandel K.G. Tisdale L.D. Masui H. Am. J. Physiol. 1984; 246: G204-G208Google Scholar). Total RNA was isolated from fresh rabbit tissues, spiny dogfish (Squalus acanthias) rectal gland, and T84 cells by the guanidine thiocyanate method(21Chomczynski P. Sachi N. Anal. Biochem. 1987; 162: 156-159Google Scholar). Poly(A)+ RNA was purified from total RNA using magnetic beads (PolyATtract, Promega Corp.). Samples of rabbit tissue mRNA were denatured by heating to 65°C in formamide and formaldehyde and size-fractionated on a 1% agarose gel. Fractionated mRNA was transferred to a nylon membrane by semidry blotting. The rabbit tissue Northern blot used in this study was the one used in (16Payne J.A. Forbush III, B. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 4544-4548Google Scholar). The membrane was completely stripped of previously hybridized probe by incubating in 0.1 × SSC and 0.5% SDS at 75°C, and confirmation of probe removal was determined by film exposure (48 h at −70°C). A blot of human tissue mRNA was obtained commercially (Clonetech, human MTN blot, 7760-1, lot 52805). The blots were prehybridized 4 h at 45°C in 50% formamide, 2 × SSPE, 2 × Denhardt's, 1% SDS, 200 μg/ml yeast RNA, 100 μg/ml fish sperm DNA and then hybridized for 24 h in fresh hybridization solution containing 106 cpm/ml 32P-labeled cRNA probe. Antisense cRNA probes were produced as run-off transcripts (MAXIscript, Ambion Inc) (nucleotides 650-1102 for Fig. 5a and nucleotides 796-1926 for Fig. 5b) from cDNA clones. The blots were subjected to a final wash of 20 min at 65°C in 0.1 × SSC and 0.5% SDS and analyzed by autoradiography at −70°C with an intensifying screen. Fluorescence in situ hybridization was performed as described previously(22Haas M. Ward D. Lee J. Roses A.D. Clarke V. E'Eustachio P. Lau D. Vega Saen de Miera E. Rudy B. Mammalian Genome. 1993; 4: 711-715Google Scholar). Metaphase spreads were prepared from human peripheral blood lymphocytes after methotrexate synchronization and bromodeoxyuridine incorporation. Probes were assigned to R bands generated by bromodeoxyuridine incorporation and subsequent photolysis(23Arnold N. Bhatt M. Ried T. Ward D.C. Wienberg J. Kessler C. Techniques and Methods in Molecular Biology: Nonradioactive Labelling and Detection of Biomolecules. SpringerVerlag, New York1993: 324-334Google Scholar). Briefly, TEF 1-1 cDNA was labeled with 11-biotin dUTP by nick translation. The probe DNA (60 ng) was coprecipitated with 3 μg of human Cot 1 DNA plus 6 μg of salmon sperm DNA. After resuspension in 12 μl of hybridization solution (50% formamide, 2 × SSC, 10% dextran sulfate) the DNA was denatured at 75°C for 10 min. Denatured DNA was applied directly to denatured metaphase spreads (50% formamide and 2 × SSC for 2 min at 70°C). The slides were incubated for 12 h at 37°C, washed 3 times in 42% formamide and 2 × SSC at 42°C, washed 3 additional times at low stringency (1 × SSC, 40°C), blocked with bovine serum albumin (3% in 4 × SSC for 30 min at 37°C), and stained with Hoechst 33258 (2.5 mg/ml) for 15 min. After washing briefly with distilled water, they were placed under a UV lamp at a distance of 10 cm for 1 h; they were subsequently incubated for an additional hour in 2 × SSC at 42°C. Avidin-DCS-fluorescein isothiocynate (Vector Laboratories) in 1% bovine serum albumin and 4 × SSC was used to detect probe sequences in which 11-biotin dUTP had been incorporated, and, after incubation for 30 min at 37°C, the slides were washed 3 times in 4 × SSC and 1% Tween 20 for 5 min at 42°C. Propidium iodide (200 ng/ml) in antifade (Dabco) was used to counterstain chromosomes. A CCD camera (PM512, Photometrics, Tucson, AZ) was used to visualize fluorescent signals; gray scale images were obtained sequentially for fluorescein and propidium iodide with precision filter sets (C. Zeiss), and the images were pseudocolored and merged. A full-length construct of the sequence for the human colonic Na-K-Cl cotransporter was prepared from the two cDNAs, TEF 11a and TEF 1-1 (see Fig. 1). Much of the 5′-untranslated region and all of the intronic DNA were removed from TEF 11a by subcloning a PstI-PstI fragment into Bluescript KS- (nucleotides −24 to 831). The small subcloned fragment of TEF 11a and the cDNA of TEF 1-1 were ligated at a common NcoI site and subcloned into the mammalian expression vector pJB20 at EcoRI and KpnI restriction sites of the polylinker(24Beck P.J. Orlean P. Albright C. Robbins P.W. Gething M.-J. Sambrook J.F. Mol. Cell. Biol. 1990; 10: 4612-4622Google Scholar). Transcription of the insert is under the control of the cytomegalovirus promoter, and the vector contains a neomycin (G418) resistance gene. The human embryonic kidney cell line (HEK-293) was maintained in Dulbecco's modified Eagle's medium (Life Technologies, Inc.) supplemented with 10% fetal bovine serum, penicillin (50 units/ml), and streptomycin (50 μg/ml) in a humidified incubator (5% CO2 at 37°C). The full-length construct was transfected into HEK-293 cells by calcium phosphate precipitation(25Puddington L. Woodgett C. Rose J.K. Proc. Natl. Acad. Sci. U. S. A. 1987; 84: 2756-2760Google Scholar). Nearly confluent cells were split 1:6 the day before transfection and were 50-70% confluent in a 10-cm dish immediately prior to transfection with 20 μg of DNA in 140 mM NaCl, 25 mM HEPES, 0.75 mM Na2HPO4, 125 mM CaCl2, pH 7.05. 48 h after transfection, cells were selected for neomycin resistance by growth in G418 (900 μg/ml), and after 3 weeks, single resistant colonies were amplified and screened by 86Rb influx measurements and Western blotting. CHAPS-solubilized membrane protein (50 μg) was deglycosylated by incubation (4 h at 37°C) with 20 units/ml of N-glycosidase F (Boehringer Mannheim) in a 0.05-ml solution containing 4% CHAPS, 0.15 M sodium phosphate buffer, pH 7.8, 2.5 mM EDTA, and protease inhibitors. Enzymatic treatment was terminated by the addition of electrophoresis sample buffer. HEK-293 cells were subcultured into 96-well plates (1:5 split) and grown to confluency (2-3 days). Prior to the flux assay, the cells were preincubated for 1 h in a low chloride hypotonic medium (1:2 dilution of isotonic low chloride medium but lacking ouabain) and then briefly washed in a low chloride isotonic medium containing 135 mM sodium gluconate, 5 mM potassium gluconate, 1 mM CaCl2, 1 mM MgCl2, 1 mM NaHPO4, 2 mM Na2SO4, 15 mM HEPES, pH 7.4, and 0.1 mM ouabain. The initial rate of 86Rb influx was determined in quadruplicate wells in the presence or absence of 200 μM bumetanide. Uptake was determined in isotonic medium in which 86Rb (1 μCi/ml) replaced potassium. Influx was terminated after 0.5 min by the addition of an ice-cold, high potassium saline followed by five rinses in phosphate-buffered saline. Cellular extracts (80 μl of 2% SDS) were assayed for 86Rb by Cherenkov radiation, for protein with the Micro BCA protein kit (Pierce) and for Na-K-Cl cotransporter production by Western blot analysis(13Lytle C. Xu J.-C. Biemesderfer D. Haas M. Forbush III, B. J. Biol. Chem. 1992; 267: 25428-25437Google Scholar). Reported Km and Ki values were obtained from nonlinear least squares fits to the data using one- or two-site (for chloride) models. Western blot analysis made use of a monoclonal antibody (T10) produced against a glutathione S-transferase fusion protein containing the hydrophilic C terminus of hNKCC1 (amino acids 902-1212).2 2C. Y. Lytle, J.-C. Xu, D. Biemesderfer, and B. Forbush, manuscript in preparation. Clones encoding the bumetanide-sensitive Na-K-Cl cotransporter in human colon (the T84 cell) were initially obtained by screening a cDNA library under low stringency conditions with nonoverlapping fragments of the shark rectal gland Na-K-Cl cotransporter (see "Materials and Methods" and (15Xu J.C. Lytle C. Zhu T. Payne J.A. Benz E. Forbush III, B. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 2201-2205Google Scholar)). The first library screening produced a cDNA (TEF 1-1; Fig. 1) with homology to the rectal gland Na-K-Cl cotransporter with a large open reading frame and poly(A) tail but that lacked the 5′-end of the coding region. A second library was screened with a 5′-end fragment of TEF 1-1 (0.7-kb; see "Materials and Methods"). 23 cDNAs were plaque-purified and excised into Bluescript SK-. Two independent cDNAs (TEF 11a and 20a) extended the 5′-end of TEF 1-1, providing the complete coding region of the human colonic Na-K-Cl cotransporter (Fig. 1). An overlapping nucleotide sequence of 4115-bp was identified, hNKCC1. The 5′-end of hNKCC1 is exceedingly rich in guanine and cytosine nucleotides with the first 900 bases containing 74% (G+C) content. We have reported previously the presence of a (G+C)-rich region at the 5′-end of sNKCC1(15Xu J.C. Lytle C. Zhu T. Payne J.A. Benz E. Forbush III, B. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 2201-2205Google Scholar). Interestingly, within this (G+C)-rich region of hNKCC1 is a triple repeat (GCG)7, which occurs within the coding region and contributes to a large alanine repeat (Ala93- Ala107; Fig. 2). This same triple repeat is observed in the putative basolateral Na-K-Cl cotransporter from mouse inner medullary collecting duct cells, mNKCC1(26Delpire E. Rauchman M.I. Beier D.R. Hebert S.C. Gullans S.R. J. Biol. Chem. 1994; 269: 25677-25683Google Scholar). 3 3The secretory and absorptive isoforms of the Na-K-Cl cotransporter are designated NKCC1 (15) and NKCC2 (16), respectively. The alternate designation BSC2 and BSC1 have also been employed (17, 26). Since alanine is assigned a positive hydropathic index (i.e. hydrophobic residue) by Kyte and Doolittle(42Kyte J. Doolittle R.F. J. Mol. Biol. 1982; 157: 105-132Google Scholar), the large alanine repeat encoded by the GCG trinucleotide repeat is responsible for the very hydrophobic region (residues color-coded red), which appears within the otherwise hydrophilic N-terminal domain of Fig. 3A.Figure 3:Hypothetical model of the human colonic Na-K-Cl cotransporter. The amino acid residues are color-coded by hydropathic index (A) as determined by the Kyte-Doolittle algorithm (42Kyte J. Doolittle R.F. J. Mol. Biol. 1982; 157: 105-132Google Scholar) or by the fractional identity of hNKCC1 to sNKCC1 (B). Both the hydropathic index and identity are averaged over a running window of 15 amino acids. Potential glycosylation sites between putative transmembrane segments 7 and 8 are indicated by branched lines. Secondary structural elements predicted by the PHD program (45Rost B. Sander C. Proteins. 1994; 19: 55-77Google Scholar, 46Rost B. Casadio R. Fariselli P. Sander C. Protein Science. 1995; (in press)Google Scholar) are shown (helices, α; wavy lines, β). No attempt has been made to pair β structural elements.View Large Image Figure ViewerDownload (PPT) The hNKCC1 cDNA includes a full-length open reading frame encoding 1212 amino acids, beginning with the first ATG (GCTATGG) downstream of a stop codon. The predicted molecular mass is 132 kDa, which is very similar to the core polypeptide identified in N-glycosidase-F-treated membranes from T84 cells with monoclonal antibodies developed against the shark rectal gland Na-K-Cl cotransporter (~135-kDa)(27Lytle C. Torchia J. Forbush III, B. J. Gen. Physiol. 1992; 100 (abstr.): 39Google Scholar). The full-length amino acid sequence of hNKCC1 is 93% identical to mNKCC1 and 73% identical to sNKCC1 (Fig. 2). The most divergent region is the N terminus, which includes a number of insertions and deletions when compared with mNKCC1 and sNKCC1. Outside of this area, hNKCC1 aligns perfectly with mNKCC1. We noted that a number of the clones isolated from one of the libraries contained intronic sequences with consensus splice sites for intron-exon boundaries ((28Mount S.M. Nucleic Acids Res. 1982; 10: 459-472Google Scholar), Fig. 1). Many of these clones were undoubtedly produced from incompletely processed mRNA during library construction (see (29Riordan J.R. Rommens J.M. Kerem B.S. Alon N. Rozmahel R. Grzelczak Z. Zielenski J. Lok S. Plavsic N. Chou J.L. Drumm M.L. Iannuzzi M.C. Collins F.S. Trsui L.C. Science. 1989; 245: 1066-1073Google Scholar)). This finding allowed us to identify some of the exons in the hNKCC1 gene (Fig. 1). Interestingly, one cDNA contained consensus splice sites and intronic DNA on either side of a 96-bp exon (Fig. 1, exon B), which corresponds exactly to the position of an alternatively spliced cassette exon in the rabbit kidney Na-K-Cl cotransporter(16Payne J.A. Forbush III, B. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 4544-4548Google Scholar). Similar to previously identified cation-chloride cotransport proteins(15Xu J.C. Lytle C. Zhu T. Payne J.A. Benz E. Forbush III, B. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 2201-2205Google Scholar, 16Payne J.A. Forbush III, B. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 4544-4548Google Scholar, 17Gamba G. Miyanoshita A. Lombardi M. Lytton J. Lee W.-S. Hediger M. Hebert S.C. J. Biol. Chem. 1994; 269: 17713-17722Google Scholar, 26Delpire E. Rauchman M.I. Beier D.R. Hebert S.C. Gullans S.R. J. Biol. Chem. 1994; 269: 25677-25683Google Scholar, 30Gamba G. Saltzberg S.N. Lambardi M. Miyanoshita A. Lytton J. Hediger M.A. Brenner B.M. Hebert S.C. Proc. Natl. Acad. Sci. U. S. A. 1993; 90: 2749-2753Google Scholar), hydropathy analysis of hNKCC1 predicts a large central hydrophobic region bounded by N- and C-terminal hydrophilic domains (Fig. 3A). As displayed in Fig. 3, 12 membrane-spanning α helices are predicted in our model of hNKCC1. Based on homology with sNKCC1, we predict that both terminal domains reside within the cytoplasm. The polypeptide sequence of hNKCC1 has five potential N-linked glycosylation sites. Two of these sites are located in a predicted extracellular hydrophilic region of the protein between putative transmembrane segments 7 and 8 (Figure 2:, Figure 3:). In comparing the primary structure of hNKCC1 to related proteins, we note that among the loops connecting predicted transmembrane domains, those that are predicted to be intracellular are considerably better conserved than are those that are predicted to be extracellular (e.g. sNKCC1, Fig. 3B). The Na-K-Cl cotransporter in the T84 cell is known to be stimulated by agents such as vasoactive intestinal peptide that activate adenylate cyclase(2Dharmsathaphorn K. Mandel K.G. Masui H. McRoberts J.A. J. Clin. Invest. 1985; 75: 462-471Google Scholar). Unlike sNKCC1, which has no consensus sites for phosphorylation by cAMP-dependent protein kinase, there is a single potential cAMP-dependent protein kinase site in hNKCC1 located within the predicted cytoplasmic C-terminal domain (Ser995; Fig. 2). This residue corresponds to one of three potential cAMP-dependent protein kinase phosphorylation sites in the rabbit kidney Na-K-Cl cotransporter (rNKCC2, Ser878; (16Payne J.A. Forbush III, B. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 4544-4548Google Scholar)). In addition, hNKCC1 contains two threonines (Thr217 and Thr1135; Fig. 2), which correspond to known phosphoacceptors in sNKCC1 (Thr189 and Thr1114; (15Xu J.C. Lytle C. Zhu T. Payne J.A. Benz E. Forbush III, B. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 2201-2205Google Scholar)). The fact that the region around both of these threonine residues in hNKCC and sNKCC1 is very well conserved suggests that these residues are also phosphacceptors in hNKCC1 (Fig. 3B). Using the TEF 1-1 cDNA, we localized the hNKCC1 gene to human chromosome 5 at position 5q23.3 by fluorescence in situ hybridization on R-banded chromosomes (Fig. 4). Five separate metaphase spreads were investigated. Two had distinct signals on both homologues of chromosome 5, with signals on all four chromatids. Three others had distinct signals on only one homologue with discrete signals on each chromatid. Evidence of hybridization at other chromosomal regions with TEF 1-1 was not observed. 26Delpire E. Rauchman M.I. Beier D.R. Hebert S.C. Gullans S.R. J. Biol. Chem. 1994; 269: 25677-25683Google Scholar have recently mapped mNKCC1 to mouse chromosome 18 in a region that they reported to be syntenic with human chromosome 5q31-33. Interestingly, the murine renal absorptive isoform, mNKCC2, has recently been cloned, and the gene localized to mouse chromosome 2(43Quaggin S.E. Payne J.A. Forbush III, B. Igarashi P. Mamm. Genome. 1995; (in press)Google Scholar). These studies clearly distinguish NKCC1 and NKCC2 as separate gene products. The level of expression of hNKCC1 and related transcripts in T84 cells, shark rectal gland, and several rabbit and human tissues was examined by Northern blot analysis (Fig. 5). Using antisense cRNA probes, a ~7.2-7.5-kb mRNA was expressed at high levels in T84 cells, rectal gland, rabbit stomach, rabbit kidney, and rabbit large intestine (Fig. 5A) and in all of the human tissues (Fig. 5B). A prominent message is also found at 6.3 kb in human heart and skeletal muscle, and a signal is detect