Medicine, Baltimore, Maryland 21205 and the PCentre de Biochimie-Centre National de la .... cloning and sequencing of a third isoform in the Na+/H+.
THEJOURNAL OF BIOLOGICAL CHEMISTRY
Vol. 267,No. 13,Issue of May 5, pp. 9340-9346,1992 Printed in U.S.A.
0 1992 by The American Society for Biochemistry and Molecular Biology, Inc.
Cloning andSequencing of a Rabbit cDNA Encoding an Intestinal and Kidney-specific Na+/H+ Exchanger Isoform (NHE-3)" (Received for publication, December 31,1991)
Chung-Ming Tse, Steven R. Brant, M. Susan Walker, Jacques Pouyssegur$, and Mark Donowitz From the Departments of Physiology and Medicine, Division of Gastroenterology, The Johns Hopkins UniversitySchool of Medicine, Baltimore, Maryland21205 and the PCentre de Biochimie-Centre National dela Recherche Scientifique, Parc Valrose, 06034 Nice, France
We previously cloned, sequenced, and expressed two H+ exchangers include regulation of intracellular pH, in pardistinct mammalianNa+/H+exchanger isoforms(NHE- ticular recovery from an acid load (5), maintenance of cellular 1 and NHE-2). We report here the cloning of a compos- volume in response to anosmotic load (6), transepithelial Na+ ite cDNA which encodes a third mammalian isoform absorption in epithelial cells (41, and as a target for growth (NHE-3), which is expressed specifically in intestine factors (7). and kidney. The protein deduced from thelongest open Recently, Sardet and Pouyssegur (8, 9) cloned by genetic reading frame of this composite sequence has 832 complementation a cDNA, which encodes the human growth amino acids with a calculated M, of 92,747. The hy- factor activated Na+/H+ exchanger. This cDNA encodes a drophobicity plot of NHE-3 is very similar to that of NHE-1 and NHE-2. NHE-3 is also predicted to have protein of 815 amino acid and is now called NHE-1' (10). 10-12 membrane-spanning domains and a long cyto- Using a BarnHI-BarnHI fragment of human NHE-1 cDNA, plasmic domain which contains putative protein kinase we isolated a4-kb cDNA encoding a rabbit NHE-1. The phosphorylation motifs. NHE-3 exhibits overall 41% deduced protein has 816 amino acids and exhibits 95% idenamino acid identity with NHE-1. NHE-3 is likely a tity to itshuman homologue (10). Based on pharmacological, immunological, and kinetic glycoprotein as it has one potential N-linked glycosylation site, which is conserved in all NHEs identified. studies, it has been suggestedthat there aremultiple isoforms Northern blot analysis of poly(A+)RNA isolated from of Na+/H+ exchangers. Ileal villus epithelial cells carry out rabbit ileum using NHE-3cDNA as a probe hybridized intestinal Na+ absorption, part of which is carried out by a to a single 6.4-kilobasetranscript. More detailed tissue brush border Na+/H+ exchanger. Knickelbein et al. (11)demdistribution of message was performedby ribonuclease onstrated by plasma membrane vesicle studies that ileal villus protection assay. It was found that NHE-3 message is epithelial cells have Na+/H+ exchangers on both their apical only expressedin intestine and kidney, with the kidney and basolateral membranes, which differ in sensitivity to cortex having the most abundant message, followed by amiloride inhibition (amiloride Ki was 104 PM for the apical intestine and kidney medulla. In intestine, ileum and membrane and 11 PM for the basolateral membrane). Using ascending colon have the same amount of message, with antibodies against the cytoplasmic domain of NHE-1, we much lesser amounts in jejunum. The message is absent found that NHE-1 localizes to the basolateral membrane of from duodenum and descending colon, which lack the both ileal villus and crypt epithelial cells (10). Similarly, in neutral NaCl absorptive process. Thus, NHE-3 might the LLC-PK1/C14porcine kidney cell line, Haggerty et al. (12) be involved in Na+ absorption in intestinal and renal demonstrated that the apical membrane Na+/H' exchanger epithelial cells. is far more resistant to ethylisopropylamiloride (EIPA) inhibition than its basolateral counterpart (12). Using antibodies against an extracellular domain of the cloned porcine NHE1, it was found that porcine NHE-1 also localizes to the Na+/H+ exchangers are found in all mammalian cells (for basolateral membrane of LLC-PKJCl, cells (13). In addition, reviews, see Refs. 1 and 2). They catalyze the exchange of the cultured hippocampal neuron Na'/H+ exchanger is not extracellular Na+ for intracellular H+ in an electroneutral inhibited by either amiloride or its 5'-amino-substituted anmode with a stoichiometry of 1:l. Although Na' and H+ are alogues (14). Kinetically, most Na+/H+ exchangers are found the primary substrates, Li' is also transported, as is NH: in to have an internal H+-modifier site which functions in rerenal and intestinal brush border membranes (3, 4). Other covery from an acid load. The presence of an internal H+ monovalent cations, including K'Rb' , , Cs+, choline' and modifier site was first identified in the renal brush border tetramethylammonium+are poor substrates OP not transmembrane Na'/H' exchanger and subsequently has been ported at all, except that K' appears to be transported by the confirmed to be present in most other Na'/H+ exchangers in ileal brush border exchanger (4).Identified functions for Na'/ many other cell types (15, 16). Nevertheless, recently, it has been suggested that the rat distal colon apical Na+/H+ ex* This work wassupported by National Institutes of Health Grants changer and the rabbit ileal villus cell brush border and ROlDK26523 and R29DK43778, Training Grants T32DK07632 and basolateral Na+/H+exchangers lack this H+-modifier site (11, F32DK08672, and theMeyerhoff Digestive Disease Center. The costs of publication of this article were defrayed in part by the payment of 17).Although the above studies suggest that thereare multiple page charges. This article must therefore be hereby marked "adver- isoforms of Na+/H+exchanger, there is nothing known about tisement" in accordance with 18U.S.C.Section 1734 solely to indicate ~~~
~
this fact. The nucleotide sequencefs) reported in this paper has been submitted to theGenBankTM/EMBLDataBankwith accession numberfs) M87007.
'The abbreviations used are: NHE,Na+/H+ exchanger; EIPA, ethylisopropylamiloride;nt, nucleotide; bp, base pair; SDS, sodium dodecyl sulfate; kb, kilobase(s); PIPES, 1,4-piperazinediethanesulfonic acid.
9340
Na+/H+ Isoform Exchanger the differences among the isoforms at the molecular level. By genomic Southern Blotanalysis, we have demonstrated the existence of a Na+/H' exchanger gene family (10). We have recently reported the cloning and characterization of another serum activated, but low EIPA sensitivity Na+/H+ exchanger isoform (NHE-2), which is expressed in intestine, kidney, and adrenal gland (18).In thispaper, we present the cloning and sequencing of a third isoform in the Na+/H+ exchanger gene family (NHE-3), which is expressed only in intestinal and renal epithelial cells. EXPERIMENTAL PROCEDURES
Cloning of a Partial CDNA Encoding Human NHE-3-A human kidney cortex cDNA library in X g t l O was obtained as a generous gift from Dr. G. 1. Bell (The University of Chicago). Using a cDNA probe derived from the styl-sty1 fragment of rabbit NHE-2 (nt 126-1898 encoding amino acids 43-633) to screen this human kidney cortex cDNA library under low stringency conditions (hybridized at 37 "C in 40% formamide, 4 X SSC, 5 X Denhardt, 1%SDS andwashed at 42 "C in 1 X SSC and 0.1% SDS), we obtained four positive clones, which wereplaque-purified. Phage lysates were prepared by the plate lysate method and were further purified by DEAE-cellulose chromatography (19). Phage proteincoat was then solubilized with proteinase K and 0.1% SDS and the DNA obtained by ethanol precipitation. Restriction digestion of isolated phage DNAs with EcoRI released inserts from 0.5 to 3.2 kb, which were subcloned into pBluescript KS (Stratagene). Terminal sequencing showed that two of the four positive clones are NHE-1; one (called HKC-3), which is 1.4 kb, encodes part of human NHE-3 (Fig. l), and theremaining clone is unrelated. th Sequence of Rabbit NHE-3-We C ~ n ~ ofn gthe ~ ~ - & n gCoding used the above 1.4-kb human NHE-3 cDNA probe (HKC-3) to screen a rabbit kidney cortex library and a rabbit ileal villus cell cDNA library. Both rabbit libraries were in XZAP. The kidney cortex library was a generous gift from Drs. C. Montrose-Rafizadeh and W. Guggino (Physiology Department, The Johns Hopkins University, School of Medicine). The ileal villus cell library was prepared by us (10). Since wewere using a human cDNA to screen the rabbit library, low stringency conditions were used as described above. We obtained two positive clones, called RAKl and RAK3. These two clones were plaque-purified and inserts were rescued into pBluescript SK by in vivo excision by co-infecting the XL1-blue cells with the helper phage R408, as described previously (IO). Restriction digestion with EcoRI showed that RAKl is 0.72 kb and RAK3 is 1.89 kb (Fig. 2). Both clones were completely sequenced and found to be overlapping. RAKl was then used to rescreen the ileal villus cell cDNA library under high stringency conditions ( h y b r i d i z ~at 42 "C in 50% formamide, 4 X SSC, 5 X Denhardt and a final wash at 60 "C in 0.1 X SSC, 0.1% SDS). More than 15 independent but overlapping clones were obtained and two of these clones (RAI1 and RAI5) gave the complete coding sequence of rabbit NHE-3 (Fig. 2). RNA Isolation-Rabbit intestinal mucosa from duodenum, jejunum, ileum, ascending colon, and descending colon was obtained by light scraping of the corresponding tissues with a glass slide (20). Total RNA from various rabbit tissues, cultured cells (rabbit skin fibroblasts and a lymphoblast cell line, TP-3), and the above intestinal mucosa were isolated by a commercially available kit, RNAzoL, which utilizes a modification of the one-step procedure of Chomczynski and Sacchi (21). Poly(A+) RNA from rabbit ileal mucosa was purified from the corresponding total RNA bytwo passages of affinity chromatography on ~ligo(dT)-cellulose. N o r ~ ~Br bnt Analysis-Five pg of poly(A+)RNA from rabbit ileal villus mucosa was denatured with formaldehyde, size-fractionated by 1% agarose gel electrophoresis, and transferred to a nylon filter by capillary blotting (22). The membrane was prehybridized for 2 h under high stringency conditions, as described above for library screening. Hybridization was carried out for 20 h in the same solution containing lo8 cpm/ml of denatured 32P-labeled insert of RAI5 as described in Fig, 2. Insert of RAI5 waslabeled by the random-primed method of Feinberg and Vogelstein (23). Washing was also performed under high stringency conditions with a final wash at 62 "C, 0.1 x SSC, 0.1 X SDS. The blot was analyzed by autoradiography using Kodak XAR film. RibonucleaseProtection Assay-Tissue distribution of rabbit NHE3 message was performed by ribonuclease protection assay with the commercially available ribonuclease protection assay kit (Ambion).
NHE-3
9341
For preparation of [32P]cRNAprobes, we used an in vitro transcription kit ( M a x ~ r i p tAmbion). , Briefly, a cDNA fragment of NHE-3 was subcloned into pBluescript (Stratagene) and used as a template for production of antisense cRNA probe. Antisense NHE-3 ["PI cRNA probe (317 bp which corresponds to nt 1181-1498 and amino acid residues 395-500) was transcribed in vitro using T7 RNA polymerase and radioactive [32P]CTP(800 Ci/mmol), as described by Melton et al. (24). After the transcription reaction, DNA templates were removed by digestion with RQ1 RNase-free DNase (Promega). ["PI cRNA probes were purified by a centrifugation Sephadex column. For ribonuclease protection assay, labeled cRNA probes were hybridized overnight at 42 "C with 30 pg of total RNA which was isolated from multiple rabbit tissues. Hybridization solution contained 80% formamide, 40 mM PIPES, pH 6.4.400 mM NaOAC, pH 6.4, 1 mM EDTA. Following hybridization, hybridization mixtures were treated with 0.01 unit of RNase A and 20 of units RNase T1 a t 37 "C for 30 min to degrade single-str~dedhybridized probe. Labeled probes that hybridized to their complementary RNA, and thus "protected" from RNase digestion, were precipitated with ethanol and resuspended in DNA loading buffer. Half of each sample (which represents 15 pgof the total RNA initially added) was separatedon a 6% polyacrylamide sequencing gel and thegel was analyzed by autoradiography. cDNA Sequencing-Sequencing of cDNA clones were performed on both strands by Sanger's dideoxy termination procedure (25). Progressive unidirectional deletion clones were obtained by the method of EroIIIlmung bean nuclease digestion (26). Deletion plasmids were purified by the alkaline lysis method and used as doublestranded templatesfor sequencing. RESULTS
Cloni~ of a Partial cDNA ~ ~ ~ NHE-3-We ~ n g have previously isolated a cDNA encoding rabbit NHE-1(10). Using the PstI-AccI fragment of the rabbit NHE-1 cDNA to rescreen the rabbit ileal villus cell cDNA library, we obtained another cDNA encoding a second novel Na+/H+ exchanger isoform (NHE-2) which is 50-fold less sensitive to ethylisopropylamiloride inhibition than NHE-1 and is also activated by serum (18).In order to identify other possible members of the Na+/H+ exchanger gene family, we used the styl-sty1 fragment of the rabbit NHE-2cDNA to screen a human A g t l O kidney cortex library under low stringency conditions. We obtained four positive clones, one of whichshowed high homology with but is not identical to rabbit NHE-1 and NHE2. This clone, called HKC-3, was completely sequenced in both strands, and the amino acid sequence was deducedbased on its homology with NHE-1 and NHE-2. As was evident from the amino acid conservation of HKC-3 compared with NHE-1 andNHE-2, clone HKC-3 likely represented a partial sequence of a third isoform of Na+/H+ exchanger (NHE-3). As shown in Fig. 1,the amino acid sequence as deduced from HKC-3 (which encodes part of human NHE-3)was compared with the homologous area of rabbit NHE-3 that we subsequently cloned as described below. These two proteins are highly homologous and exhibit 94% amino acid identity (98% similarity) to each other. The putative potential N-linked glycosylation site is also conserved between these two proteins. At the nucleotide level, they have 91% identity (data not shown). Cloning and Sequencing of Rabbit NHE-3-HKC-3 cDNA was used as a probe to screen a rabbit kidney cortex XZAp cDNA library. Screening lo6 plaques yieldedtwo positive clones, called RAK1 and RAK3, as shown in Fig. 2. Both clones were completely sequenced and found to be overlapping. They have approximately 50% identity with rabbit NHE-1 and NHE-2 at the amino acid level and have 95% amino acid identity with HKC-3. This suggests that RAKl and RAK2 are partialcDNAs encoding rabbit NHE-3, a third distinct Na+/H+ exchanger isoform. RAKl was then used as a probe to rescreen both the rabbit kidney cortex and the
9342
Na+/H+Exchanger Isoform NHE-3 HUMAN RABBIT HUMAN
FIG. 1. Alignment of partial amino acid sequences of human NHE-3 as deduced from clone HKC3 (above, amino acid residues 1375) and the homologous sequence of rabbit NHE-3 (below, amino acid residues 131-505). Amino acids are indicated by their single-letter abbreviation. The partial amino acid sequences span eight membrane-spanning domains which are overlined (M5-Ml0, M5a, and M56). The conserved potential N-linked glycosylation site is indicated by +. These two sequences exhibit 95% amino acid identity. Amino acid numbers are shown on the right. “1” indicates identity and “:” and “.” indicate similarity.
M5 PNRLFFGNLGTILLYAVVGTWNAATTGLSLYGVFLSGLMGDLQIGLLDF 50
lIlIII:III.IIIIIIlIIIlIIIIlIlIIIIlIIII:Il:I.IIIIII
PNRLFFSNLGSILLYAVVGTVWNAATTGLSLYGVFLSLYGVFLSGIMGELKIGLLDF 180 M5a M5b LLFGSLMAAVDPVAVPAVFEQVHVNEVLFIIVFGESLLNDAVTVVLYNVF 100
IIIIII:IIIIIIII
IIll:IIIIllIIlIlIIIIIIIIIIIllIIIII
HUMAN
LLFGSLIAAVDPVAVLAVFEEVHVNEVLFIIVFGESLLNDAVTVVLYNVF 230 M6 ESFVALGGDNVTGVDCVKGIVSFFWSLGGTLVGWFAFLLSLVTRFTKH 1 5 0
RABBIT
QSFVTLGGDKVTGVDCVKGIVSFFVVSLGGTLVGWFAFLLSLVTRFTKH 2 8 0
HUMAN
VRIIEPGFVFIISYLSYLTSEMLSLSAILAITFCGICCQKYVKANISEQS 200
RABBIT HUMAN
VRVIEPGFVFIISYLSYLTSEMLSLSSILAITFCGICCQKYVKANISEQS 330 M9 M 8 ATTVRYTMKMLASSAETIIFMFLGISAVNPFIWTWNTAFVLLTLVFISVY 250
RABBIT
ATTVRYTMKMLASGAETIIFMFLGISAVDPLIWTWNTAFVLLTLLFVSVF
HUMAN
RAIGVVLQTWLLNRYRMVQLEPIDQVVLSYGGLRGAVAFALVVLLDGDKV 300
RABBIT HUMAN
RAIGVVLQTWLLNRYRMVQLELIDQVVMSYGGLRGAVAFALVALLDGNKV 430 M10 KEKNLFVSTTIIVVFFTVIFQGLTIKPLVQWLKVKRSEHREPRLNEKLHG 350
RABBIT
KEKNLFVSTTIIVVFFTVIFQGLTIKPLVQWLKVKRSEHREPKLNEKLHG
HUNAN
RAFDHILSAIEDISGQIGHNYLRDK 3 7 5
RABBIT
RAFDHILSAIEDISGQIGHNYLRDK 5 0 5
RABBIT
:IIl.IIII.IIlIIIIIIlIlIIIIIIIIIIlIIIIIIIlIllIlIIlI M l
+
II:IIIIIIlllIlIIIlIIIIIIII.IIIIllIIIIIIIIIIIIIIIII
IIIIIIIIIIIII:IIIIlIIIIIIIII:I:lIIIIIIIllIII:I:Il: IIIIIIJIIIIIIIIIIIIII
380
lIlII:IIlIIIllIIIIlI.IIII:lI
IIIIIIIIIIIIlIllIIIIlIIIIlIIIIIIIIIIlIIIII:IIIIIIl
480
IIIIIIIIIJIIIIIIIIIIIIIII
FIG.2. Schematic representation of rabbit NHE-3 cDNA clones. The protein coding region is represented by the hatched area and thenoncoding regions by open bars. The composite NHE-3 cDNA is 2656 bp long with an open reading frame of 2449 bp. Nucleotide numbers are indicated on the top of partial cDNA clones, the sizes of which are indicated at theright. rabbit ileal villus cell cDNA libraries under high stringency conditions. More than 15 totalpositive clones were obtained from both libraries. Characterizing the inserts derived from these positive clones by restriction mapping and terminal sequencing showed that they were overlapping. Two of these clones, RAIl and RAI5, were sequenced completely in both strands andgave the complete coding sequence of NHE-3 as shown in Fig. 2. Fig. 3 shows the composite nucleotide and amino acid sequences of rabbit NHE-3. We have sequenced 119 bp in the 5”untranslated region, 38 bp in the 3”untranslated region, and an open reading frame of 2499 bp. The initiation ATG triplet is in good agreement with Kozak’s consensus sequence for translation initiation (27). In the 5’-noncoding region, there is a seven amino acid mini-cistron upstream from the putative initiator methionine, beginning at -36 and ending at -12. The polyadenylation signal and the poly(A+) RNA tail have not yet been identified. The amino acid sequence deduced from the open reading frame of NHE-3 revealed a protein of 832 amino acid residues with a calculated M,of 92,747. It is probably a glycoprotein as three potential N -
linked glycosylation sites were identified at N325,N692,and NRI 1 Upon sequencing multiple overlapping cDNA clones, it was found that there are areas with a single base variation, that is T431+ C;C450 + T; C6’0 4T and + C. In three of these substitutions, therewas no change in thededuced amino acid because of codon degeneracy. However, in the T31+ C substitution, leucine 144 (CTC) was changed to proline (CCC). Fig. 4 compares the hydrophobicity plots of NHE-1 and NHE-3 using the Kyte-Doolittle algorithm (28). These two proteins have very similar hydrophobicity profiles. Each protein is predicted to have 12 membrane-spanning domains and a long cytoplasmic C terminus of similar length. It was proposed previously by Sardet and Pouyssegur (8)that theNa+/ H+ exchanger has 10 membrane-spanning domains. Therefore, in order to allow consistent nomenclature of membranespanning domains, we named the 12 membrane-spanning domains as 1-10 and 5a and 5b, the latterbeing two additional membrane-spanning domains between membrane-spanning domains 5 and 6. Fig. 5 compares the amino acid sequences of rabbit NHE-1 and rabbit NHE-3. Rabbit NHE-3 has41% amino acid identity with rabbit NHE-1. Membrane-spanning regions are more conserved than theextracellular and cytoplasmic loops; in particular, putative membrane-spanning domains 5a and 5b have amino acid identity of 95 and 85%, respectively. There is one potential glycosylation site which is conserved between the two proteins, that is N370in NHE-1 and N325in NHE-3. This asparagine is located on putative extracellular loop c (8). Both the Nterminus(the putative membrane-spanning domain 1and theextracellular loop a (8))and thecytoplasmic C terminus are areas which are very divergent between NHE1 and NHE-3. NHE-1 has a long putative extracellular loop a which has 73 amino acid residues. Extracellular loop a in NHE-3 has only 29 amino acids. An additional putative N linked glycosylation site is found in extracellular loop a in NHE-1 but this is absent in NHE-3. Although in the N terminal region, there is no significant conservation between the NHE-1 and NHE-3, in both putativemembrane-spanning
9343
Na+/H+Exchanger Isoform NHE-3
1
31
TGCCGCCTCACCCGCTCGCGCAAGGCTTC ACGCCGCCCACGCGCGTCCCGCGGCCTCGCGCAGAGGGACCCTCCCCACGCGACATGCGCGTCGGGCCCCGGCGCTGAGCGGCAGAGGCG ATGTCAGGGCGCCGGGGCTGCGGCCCTTGCTGGGGGTTGCTGCTGGCGCTGGTGCTGGCGCTGGGCGCGCTGCCCTGGACGCAGGGCGCC ~ ~ G U G G C G P C W G L L L A L V L A L G A L P W T Q G
-9 1 -1 90 A
GAGCAGGAGCATCACGATGAGATCCAGGGCTTCCAGATAGTCACGTTCAAATGGCACCACGTCCAGGACCCCTACATCATCGCGCTCTGG E Q E H H D E I Q G F Q I V T F K W H H V Q D P Y I I A L W
180
270
GTCCTCG,TGGCCAGCTTGGCTAAGATCGTCTTCCACCTGTCCCACAAAGTCACCAGTGTCGTCCCCGAGAGCGCGCTGCTCATCGTGCTA 61
V
L
V
A
S
L
A
K
I
V
~
H
L
S
H
K
V
T
S
V
V
P
E
S
A
L
L
I
V
L
121
GGCCTGGTGCTAGGCGGCATCGTCCTGGCCGCTGACCACATCGCCTCCTTCACGCTCACGCCCACCGTCTTCTTCTTCTACCTGCTGCCG G L V L G G I V L A A D H I A S F T L T P T V F F ~ Y L L P I 9 CCCATCGTGCTCGACGCCGGCTACTTCATGCCCAACCGGCTCTTCTTCAGCAACCTGGGCAGCATCCTGCTCTACGCCGTCGTGGGCACC P I V L D A G Y ~ ~ P N R L ~ ~ S N L G S I L L Y A V
540
151
GTGTGCAACGCGGCCACCACCGGGCTGTCGCTCTACGGAGTCTTCCTCAGCGGCATCATGGGCGMCTGAAGATCGGCCTCCTGGACTTC V W N A A T T C L S L Y G V F L S C I ~ G E L K I G L L D F CTGCTGTTCGGCAGCCTCATTGCTGCCGTGGACCCCGTGGCCGTGCTGGCCGTGTTCGAGGAGGTCCACGTCAACGAGGTTCTGTTCATC
630
91
181
211
L
L
F
G
S
L
I
A
A
V
D
P
V
A
V
L
A
V
F
360
450
V
E V H V N E V L F I t ATCGTTTTCGGGGAGTCGCTGCTAAACGACGCCGTCACGGTGGTGCTGTACAACGTGTTCCAATCTTTCGTGACACTGGGTGGTGACAAG I V F G E S L L A D A V T V V L Y N V F Q S F V T L G G D K V
T
G
V
D
C
V
K
G
I
V
S
F
F
V
V
S
L
120
G
G
T
L
V
G
V
V
F
A
F
810
L 900
CTGTCGCTGGTGACCCGCTTCACCAAGCACGTGCGCGTCATCGAGCCCGGCTTCGTCTTCATCATCTCCTACCTGAGCTACCTGACGTCC 211
L
GAGATGCTGTCCCTGTCGTCCATCCTGGCTATCACCTTCTGCGGCATCTGCTGTCAGAAGTATGTGAAAGCCAACATCTCGGAGCAGTCG E ~ L S L S S I L A I T F C G I C C Q K Y V K A N I S E Q S
990
301
GCCACCACGGTGCGCTACACCATGAAGATGCTGGCCAGCGGGGCTGAGACCATCATCTTCATGTTCCTGGGCATCTCGGCCGTGGACCCG
1080
331
S
A
L
T
V
T
T
V
R
R
F
Y
T
K
T
H
~
V
K
R
~
V
L
I
I
E
S
P
G
G
A
F
E
V
F
T
I
I
I
I
S
F
P
~
+
L
F
S
Y
L
t
G
T
I
S
S
A
V
L
I
W
T
U
N
T
A
F
V
L
L
T
L
L
F
V
S
V
F
R
A
I
G
V
V
L
Q
T
D
P
1110
CTCATCTGGACCTGGAACACGGCCTTCGTGCTGCTCACGCTGCTCTTCGTGTCCGTGTTCCGGGCCATCGGTGTCGTGCTGCAGACCTGG 361
W 1260
CTGCTGAACCGGTACCGCATGGTGCAGCTGGAGCTGATCGACCAGGTGGTCATGTCCTACGGCGGCCTGCGCGGGGCCGTGGCCTTCGCC 391
L
421
CTCGTCGCGCTGCTGGACGGGAACMGGTCMGGAGAAGAACTTGTTCGTCAGCACCACCATCATCGTCGTCTTCTTCACCGTCATCTTC 1 3 5 0 L V A L L D ~ N K V K E K ~ L F V S T T I I V V F ~ T V I F
451
Q
L
N
R
Y
R
~
V
Q
L
E
L
I
D
P
V
V
~
S
Y
G
G
L
R
G
A
V
A
F
L
T
I
K
P
L
V
Q
U
L
K
V
K
R
S
E
H
R
E
P
K
L
N
E
K
L
H
A
1440
CAGCGCCTGACCATCAAGCCGCTGGTGCAGTGGCTGAAGGTGAAGCGCAGCGAGCACCGGGAGCCCAAGCTGAACGAGAAGCTGCACGGC
G
T
E
GTGACAGGCGTGGACTCTGTGAAGGGCATAGTGTCCTTCTTCGTGGTGAGCCTGGGGGGCACGCTGGTGGGGGTGGTCTTCGCCTTCCTG 241
G
G
9
CGGGCTTTCGACCACATCCTCTCGGCCATCGAGGACATATCCGGCCAGATTG~~CACMTTATCTCAGGGACAAGTGGGCCMCTTCGAC 1 5 3 0 481
R
A
F
D
H
I
L
S
A
I
E
D
I
S
G
Q
1
G
"
R
N
Y
L
R
D
K
U
A
I
I
F
D
1620
CCGAGCTTCCTCAGCAAACTGCTCATGAGGCAGTCGGCCCAGAAGTCTCGGGACCGGATTCTGAACGTGTTCCACGAGCTGAACCTGAAG 511
R
R
F
L
S
~
L
L
~
R
Q
S
A
Q
K
S
R
D
R
I
L
N
V
F
H
E
L
N
L
A
541
D
1800
571
TTCAGCACGCCGCGCCCGTCCACCGTGGAGGCCTCCGTCTCCTACCTCCTGAGGGAGAGTGCCAGCGCCGTGTGCCTGGACATGCAGTCG F S T P R P S T V E A S V S Y L L R E S A S A V C L D M Q S CTGGAGCAGCGGCGGCGCAGCGTGCGCGACGCAGAGGACGTCATCACGCACCACACGCTGCAGCAGTACCTGTACAAGCCTCGGCAGGAG
1890
K
A
L
E
R
P
D
G
L
V
S
D
I
L
S
T
A
P
H
F
A
H
T
I
Y
L
R
G
L
Q
S
T
Q
P
L
Y
S
L
T
K
E
Y
D
K
~
R
P
~
E
R
L
E
Q
V
~
V
2160
G
D
R
AGAAACAGCAGTGTCCCCAACGGGAAGCTGCCCCTGGACAGCCCTGCCTACGGCCTCACCCTTAAGGAGAG~CAGCTGGMCTTCI\GAC
H
R
E
2010
P
V
G
CTGGAGTCCTTCAAGTCGGCCAAGCTGGGCCTCGGCCAGAGCAAGAAGGCGACCAAGCACAAGCGGGAGCGGGAGCGGGCGCAGAAGCGG L E S F K S A K L G L G Q S K K A T K H K R E R E R A Q K R
V
S
E
661
S
R
T
TACAAACACCTGTACAGCCGGCACGTGCTCAGCCCCAGCGAGGACGAGAAGCAGGACAAGGAGATCTTCCACCGCACCATGCGCAAGCGG 1980 Y K H L Y S R H V L S P S E D E K Q D K E I F H R T M R K R
S
R
V
631
N
R
Y
L
R
Q
S
601
691
E
I
K 1110
GATGCCATCAGCTACGTGACCGAGGGAGAGCGCCGAGGCTCCCTGGCCTTCATCCGCTCCCCGAGCACTGACAACATGGTCAATGTGGAC
L
E
S
D
CCCGAGGAGGCCCCCGACTACTACGAGGCTGAGAAGATGAGTGGGGGGATCGAGTTCCTGGCCAGCGTCACCAAGGACACGACCTCCGAC
2250
121
P
S
D
751
TCCCCTGCAGGAATTGACAACCCTGTGTTCTCCCCGGACGAGGACCTGGCCCCCAGCCTCCTGGCCAGGGTGCCACCCTGGCTGTCTCCC S P A G I D N P V F S P D E D L A P S L L A R V P P W L S
~
E
E
A
P
D
Y
Y
E
I
E
K
~
S
G
G
I
E
F
L
A
S
V
T
K
D
T
T
2340
GGGGAGGCCGTGGTGCCCTCCCAGAGGGCCCGCGTGCAGATCCCCTACTCCCCGGGCAACTTCCGCCGCCTGGCGCCCTTCCGGCTCAGC 781
G
E
A
V
V
P
S
Q
R
A
R
V
Q
I
P
Y
S
P
G
N
F
R
R
L
A
P
F
R
L
N
K
S
V
D
S
F
L
L
A
E
D
G
A
E
H
GACCGGACGCTCGTCCC
P
E
S
T
H
n
2430
S
AACAAGTCGGTGGACTCGTTCCTGC~GGCGGAGGCGGAGGACGGCGCTGAACACCCCGAGTCCACACACATGTGACACGGGCTCCGACACGCCGCT
811
D
2520
X 2531
FIG. 3. Nucleotide sequence of rabbit NHE-3and the deduced amino acid sequence of the protein. Nucleotides are numbered at the right of the sequence with respect to the putative translation initiation site. Amino acids are numbered at theleft of the sequence and are represented by their single-letter abbreviations. The 4 C, 4 T, C680 + in-frame upstream mini-cistron is underlined. "#" indicates base variations in cDNAs: T3' T and T1480 + C and "*n and "+" indicate the stop codon and theN-linked glycosylation site, respectively.
domains 1, there is a 7-9 amino acid stretch of hydrophobic residues of leucine, valine, and alanine. Furthermore, membrane-spanning domain 1 of both NHE-1 and NHE-3 contains a signal peptide consensus sequence, suggesting the possibility that thisregion might be cleaved off in themature protein (29). In theC terminus, the homology of the proteins decrease toward the 3'-end. Northern Blot Analysis and Tissue Distribution of NHE-3
Message-In order to determine the size of NHE-3 transcript, Northern blot analysis was performed using poly(A+) RNA isolated from rabbit ileal villus cells and RAI5 cDNA as the probe. As shown in Fig. 6a, RAI5 hybridized to a single transcript of 5.4 kb in theileal villus cell poly(A+) RNA. To define the magnitude of distribution of message among the tissues, ribonuclease protection assay was used. As shown in Fig. 66, NHE-3 message is only expressed in intestine and
Na+/H+Exchanger Isoform NHE-3
9344
(A)
M1 RNHE-1 MLLWSAVRGLSPPRIVPSLLVVVALAGLLPGLRSHGLQLSPTDSTTPDSg 50
HYDROPHOBIC I
3
I
I
s
9
..I1
NHE-I
I?
.
:.I
I :l:ll::l
I
RNHE-3 MSGRGGCGPCWGLLLALVLALGAL.P
.
I
:
..I.
..WTQGAEQEHHDEI... ..
38
RNHE-1 PSRERSIGDVTTAPPEVTPESRPVNRSVTEHGMKPRKAFPVLGIDYTHVR 100 RNHE-3
....................................
.:I.::.:.:
(1.
QGFQIVTFKWHHVQ 52 M2 M 3 RNHE-1 TPFEISLWILLACLMKIGFHVIPTISSIVPESCLLIVVGLLVGGLIKGVG 1 5 0
.I: l.ll:l:l:l
~ ~ : . l : l l l l . l l l l : l l : : l ~ ::.: :
11.11:
RNHE-3 D P Y I I A L W V L V A S L A K I V F H L S H K V T S V V P E S A L L I V L G L V L G G I V L ~ D102 M4 M5 199 RNHE-1 EKPPF.LQSEVFFLFLLPPIILDAGYFLPLRQFTENLGTIL1FAVVGTLW
. :.I
I I .lll.ll::lllll:l
I ~~lll::lllll:llllll:l
RNHE-3 HIASFTLTPTVFFFYLLPPIVLDAGYFMPNRLFFSNLGSILLYAVVGT~ 152 M5a RNHE-1 ~ F F L G G L M Y A V C L V G G E Q I N N I G L L D N L L F G S I I S A V D P V A V L A V F E E I 249
I
I1
.
I I
:l:l
.IIllI lllll:l~lllllllllllll:
RNHE-3 NAATTGLSLYGVFLSGIMGELKIGLLDFLLFGSLIAAVDPVAVLAVFEEV 202
MSb -2.01
I
HYDROPHILIC I
,
100
200
300
400
500
T
,
600
'1 I, 700
RNHE-1 HINELLHILVFGESLLNDAVTVVLYHLFEEFANY. . D H V G I V D I V L G z .
,
l:llllllllllllllll::l:~l~..
l:ll:l
1.1.
11.1
EO(
M6
RNHE-1 FFVVALGGVFVGVVYGVIAAFTSRFTAHIRVIEPLFVFLYSYMAYLS6L
AMINOACIDS
llll.lll~:llll::.:
347
I : I I l l l lll:.ll:.ll..l:
.:..Ill
RNHE-3 FFVVSLGGTLVGVVFAFLLSLVTRFTKHVRVIEPGFVFIISYLSYLTSEM 302 M
-
7
MI3
+
RNHE-1 FHLSGIMALIASGVVMRPYVEANISHKSHTTIKYFLKMWSSVSETLIFIF 397
1l:I:l:.
:
HYDROPHOBIC
3.0-
297
I::I
RNHE-3 HVNEVLFIIVFGESLLNDAVTVVLYNVFQSFVTLGGDKVTGVDCVKGIVS 252
:I:.
..)I.III(..I ll::l
: ~ ~ : . ~ . . ~ ~ : ~
RNHE-3 LSLSSILAITFCGICCQKYVKANISEQSATTVRYTMKMLASGAETIIFMF 352
7
NHE-3
M9
RNHE-1 LGVSTVAG.SHHWNWTTVISTLLFCLIARVLGVLGLTWFINKFR1VKLTP 446
II:I.I..
11
)II).
. I I :
:
l.:ll:
~ l : : ~ : : ~ : ~ . ~ .
RNHE-3 LGISAVDPLIWTWNTAFVLLTLLFVSVFRAIGVVLQTWLLNRYRMVQLEL 402 M10 RNHE-1 KDQFIIAYGGLRGAIAFSLGYLLDKKHFPNCDLFLTAIITVIFFTVFVQG 496
~ ~ . : : . ~1 1 1 ~. . .~. :~l l : ~ ... l ~. l : l ~l ~ l : :. ~ ~ ~ ~ . ~ .
RNHE-3 IDQVVMSYGGLRGAVAFALVALLDGNKVKEKNLFVSTTIIVVFFTVIFQG 452 RNHE-1 MTIRPLVDLLAVKKKQETKRSINEEIHTQFLDHLLTGIEDICGHYGHHHW 546
... : [ \ . : I . .
: ~ l : ~ l l :l l:: .~: .
:ll:~.:lll~:l:.~l:.:
RNHE-3 LTIKPLVQWLKVKRSEHREPKLNEKLHGRAFDHILSAIEDISGQIGHNYL 5 0 2 RNHE-1 K D K L N R F N K K Y V K K C L I . . A G E R S K E P Q L I A F Y H K M E M K Q A I E L V E S . . . 591
:I/:..I:::::.I
3.0
HYDROPHILIC
100
200
300
400
500
,
600
700
.:::I::
.::..:l.:::~:l~..l..
551
RNHE-1 .GGMGKIPSAVSTVSMQNIHPKALPAERILPALS................
624
I:::
1.
I:
RNHE-3 RDKWANFDRRFLSKLLMRQSAQKSRD.RILNVFHELNLKDA1SYVTEGER
1.1:
..
1 ) . I I:.
.:.
:
:.:I
RNHE-3 RGSLAFIRSP.STDNMVNVDFSTPRPSTVEASVSYLLRESASAVCLDMQS 600
EO(
AMINOACIDS
FIG.4. Hydrophobicity plots of deduced amino acid 8equences of rabbit NHE-1 ( A )and NHE-3 ( B ) .Average hydropathy indexes were calculated by the method of Kyte-Doolittle using a window of 15. Positive values represent hydrophobicityand negative values represent hydrophilicity. 12 predicted membrane-spanning domains are indicated from 1 to 10, 5a and 5b. The numbering of membrane-spanning domains is consistent with the 10 membranespanning domain model of Sardet and Pouyssegur (8) using the method of Engelman.
....,... KDKEEE1,RKILRNNLQKTRQRLRSYNRHTLVADPYEEAWNQ
665
RNHE-3 LEQRRRSVRDAEDVITHHTLQQYLYKPRQEYKHLYSRHVLSPSEDEKQDK
650
RNHE-1
:II:
I :..I..
I 1.11
. : . .: ::...
:.
:I
RNHE-1 MLLRRQKARQLEQ.KINNYLTVPAHKLDSPTMSRARIGSDPLAYEPKADL 714
:::I
. :.[I
I
... ....I ......I . ) .. . .
I.:.I
RNHE-3 EIFHRTMRKRLESFKSAKLGLGQSKKATKHKRERERAQKRRNSSVPNGKL 700 RNHE-1 PVITIDPASPQSPESVDLVNEELKGKVLGLSREPRVAEEAAEEDEDGGI. 763
:II. . . l l : : I : I ( .. .:: . . .I (. ) I .....DSPA....YGLTLKERELELSDPEEAPDYYEAEKMSGGIE
738
....VNRPKEPSSPCTDD..VFSPAPSDSPS.SQRMQRCLSDPGPHPEPG
806
I:
RNHE-3 PL... RNHE-1
I :.....)I:.
I
I l l ( . . . . ( I. I : . .
It..::
....
RNHE-3 FLASVTKDTTSDSPAGIDNPVFSPDEDLAPSLLARVPPWLSPGEAVVPSQ 788 RNHE-1 EGEPFIPKGQ 816 :
.
( I :.
RNHE-3 RARVQIPYSPGNFRRLAPFRLSNKSVDSFLLAEDGAEHPESTHM 832
kidney. It is expressed, in descending order, in kidney cortex, followed equally by ileum and ascending colon, then kidney medulla and least in jejunum. It is absent from duodenum and descending colon. It is also absent from other epithelial tissues including liver and trachea and is absent in all nonepithelial tissues, including brain, heart, skeletal muscle, adrenal gland, the cultured rabbit lymphoblast cell line TP-3, and cultured rabbit skin fibroblasts. DISCUSSION
We previously cloned a rabbit ileal villus cell basolateral membrane Na+/H+ exchanger (NHE-1)and provided evidence for the existence of a gene family of Na+/H+exchangers by genomic Southern blot analysis (10). The presence of such a gene family was confirmed by our cloning, sequencing, and expression of a second isoform NHE-2 (18).Although NHE2 is an EIPA-resistant Na+/H+exchanger isoform, its message expression is not restricted to kidney and intestine; itis found in a significant amount in adrenal gland and a small amount in skeletal muscle. This suggested to us that NHE-2 is not unique to epithelial cells and that there might be additional isoform Na+/H+ exchangers that are epithelial cell-specific. In the present study, we present the cloning and sequencing
FIG.5. Alignment of the amino acid sequences of rabbit NHE-1 (RNHE-I)and rabbit NHE-3 (RNHE-3).Rabbit NHE1 sequence was obtained as describedin Ref. 10. Amino acids are indicated by their single-letter abbreviation. Membrane-spanning domains are overlined (MI-MlO, M5a, and M56). The conserved Nlinked glycosylation site is indicated by "+". These two amino acid sequences exhibit an identity of 41%.Amino acid numbers areshown on the right. ' 1" indicates identity and ":"and I'." indicate similarity.
of a third member of the Na+/H+ exchanger gene family, NHE-3, which is expressed only in intestine and kidney. Using low stringency library screening conditions, we obtained a cDNA whichencodes part of the human NHE-3and also a composite cDNA which encodes the full-length coding sequence of rabbit NHE-3. Like human and rabbit NHE-1, which exhibited 95% amino acid identity (lo), human and rabbit NHE-3 are also conserved with amino acid identity of 94%, at least in the homologous region that we have cloned. Rabbit NHE-3 is also conserved with the rat NHE-3 with amino acid identity of 90% (data notshown).' Thus, the Na+/ 'By mutual agreement, prior to submission, we compared the rabbit NHE-3 amino acid sequence with that of the rat cloned by Drs. Orlowski and Shull.
Na+/H+Exchanger Isoform NHE-3
a
Standards -6.6 Kb
5.4 Kb-
-4.4 Kb
FIG. 6. Northern blot analysis and tissue distribution of NHE-3 message.A, Northern analysis: RA15 (see Fig. 2) was used
as a probe to determine the expression and the size of the NHE-3 transcriptby Northern analysis. 5 pg ofpoly(A+)RNA was hybridized to the probe under high stringency hybridizationconditions of 42 "C in 50% formamide, 4 X SSC, 4 X Denhardt's solution, 1%SDS, and high stringency washing conditions of 65 "C,0.1% SSC and 0.1% SDS. DNA size standards are shown in the right margin, and size of the message is shown in the left margin. B,ribonuclease protection assay: total RNA (30 pg), which was isolated from multiple rabbit tissues and rabbit cultured cells as shown on the top of the figure, was hybridized overnightat 42 "Cto a 317-bpantisense NHE-3 ["PI cRNA probe which corresponded to nt 1181-1498 under the conditions of 80% formamide, 40 mM PIPES, pH 6.4,400 mM NaOAc, pH 6.4, 1 mM EDTA. After hybridization, hybridization mixtures were treated with 0.01unit of RNase A and20 units of RNase T1 at 37 "C for 30min. Labeled probesthat had annealedto their complementary RNA were protected from RNase digestion and were ethanol-precipitated and resuspendedin DNA loading buffer. Half of each sample (representing 15 pg of the total RNA initially added) was analyzed by 6%polyacrylamide gel electrophoresis.The gel was then analyzed by autoradiography.
9345
These variations presumably arose from cDNAs synthesized from mRNA templates transcribed from different alleles of the NHE-3 gene. In the NHE-3 cDNAwe sequenced, we found that one out of four such variations resulted in an amino acid change: leucine 144to proline. This is a nonconservative substitution which is locatedin putative membranespanning domain 5. The significance of this substitution is unknown. On cloning and sequencing rabbit NHE-2 cDNA, it was also found that there were single base pair variations in the NHE-2 cDNA.3 In genomic Southern blot analysis of human NHE-3, we also found that there were EcoRI restriction fragment polymorphisms for human NHE-3.4 Thus, it appears that the Na+/H+exchanger genesexhibit allelic polymorphisms. All NHEs cloned to date have similar hydrophobicity plots (Fig. 4) (18), and analyses of these plots suggest that each protein has two distinct domains: an N-terminal membranespanning domain and a hydrophilic C terminus. Using the method of Engleman (30), Sardet and Pouyssegur (8) proposed that theNa+/H+exchanger has 10 membrane-spanning domains. However, as discussed previously (8),extracellular loop b is predicted not to be very hydrophilic, and two additional putative membrane-spanning domains are predicted in this region by using the Kyte-Doolittle (28) algorithm. After comparingthe amino acid sequencesof the three cloned NHEs (NHE-1, NHE-2, and NHE-3) and theirhydrophobicityplots using the method of Kyte-Doolittle, we favor the 12 membrane-spanning domain model. The two additional putative membrane-spanning domains, 5a and 5b, are hydrophobic, have higha! helical content using Chouand Fasman paradigm (31), and are found to be the most highly conserved regionin all cloned NHEs, suggesting that the putative membranespanning domains 5a and 5b might be important functional parts of the protein. Inclusion of these two additional membrane-spanning domains does not change the orientation of the N- and C-terminal parts of the protein. The C terminus has been shownto be cytoplasmic, and theputative extracellular loop a has been confirmedto be extra-cytoplasmicbased on antibody staining (9,10,13). As noted by Miller et al. (34), unlike other cloned membrane transport proteins (ie. the erythroid Cl-/HCO; exchanger(32) and a brain/erythroid glucose transporter protein (33)), theintron-exon boundaries of NHE-1 do not correspond with the membrane-spanning domain boundaries of NHE-1 when these boundaries were derived using Engelman plots.This remains true for NHE-1, when it is assigned the 12 membrane-spanning domains derived using the Kyte-Doolittle hydropathy paradigm. Both the N and the C terminus are very different among the members of the gene family. The C terminus is believed to be the regulatory domain and to contain sites for kinase phosphorylation. NHE-1 has both protein kinase C and calmodulin kinase I1 consensus sequences, but no CAMP-dependent protein kinase consensus sequences (10).NHE-3 has all three types of consensus sequences. Since each Na+/H+ exchanger is likely to be regulated differently by protein kinases, the divergence in the C terminus may represent the specificity of the protein for kinase regulation. NHE-1 has a very long extracellular loop a, but this loop is much shorter in NHE-3. This divergence in N terminus is found not only between each member of the gene family within one species but is also found among species, comparing human and rabbit NHE-1 (10) and rabbit and rat NHE-3 (data not shown)? Thus the N terminus might contain sequences which define
H+ exchanger gene family is not unique to rabbit; and the family members recognized to-date demonstrate little divergence among species, suggesting the conserved evolution of this gene family. We have sequenced2656 bp of rabbit NHE-3 cDNA. In the 5"noncoding region, there is an in-frame mini-cistron upstream of the putative start codon. Such a mini-cistron is also found in the 5'-untranslated region of NHE-1. However, the peptides encoded by these two mini-cistrons are unrelated. No such mini-cistron is present in the 112 bp of the 5'untranslated region of NHE-2 we have sequenced,and aninframe stop codon was found 57 bp upstream of the putative initiation site. Thus, it appears that not every memberof this gene family has such a mini-cistron. The function of these mini-cistrons is not known but they might be involved in posttranscriptional control. C.-M.Tse, J. Pouyssegur, andM.Donowitz, unpublishedresults. Base pair variations were found in certain areas in NHE-3 'S.R. Brant, E. Jabs, C.-M. Tse, and M.Donowitz, unpublished cDNAwhen we sequenced multiple NHE-3 cDNAclones. results.
Na+/H+Exchanger Isoform NHE-3
9346
the specificity of the expression of the protein, i.e. species-, tissue-, or localization-specific. Furthermore, the putative membrane-spanning domain 1 in NHE-1 and NHE-3 is predicted to contain asignal peptide sequence (29).This suggests the possibility that this region might be cleaved off in the mature protein and that the Na+/H+exchanger has 11 membrane-spanning domains in situ instead of 12 membranespanning domains as predicted. Kidney and intestine are known to be sites of Na+ absorption andNHE-3 message is restrictedto kidney and intestine, with most message in kidney cortex, followed in amount by intestine.Inintestine,NHE-3 is only expressed in ileum, ascending colon, and jejunum. The neutral NaCl absorptive process in intestine is believed to be composed of an apical Na+/H+ exchanger and an apical CI-/HCO; exchanger (35). Neutral NaCl absorption is found only in rabbit jejunum ileum, and ascending colon where NHE-3 message is localized. Consistent with the occurrence of the neutral NaCl absorptive process, NHE-3 message is notfound everywhere in intestine, being absent from duodenum and descending colon. Thus, the tissue distribution of NHE-3 message suggests that NHE-3 might be the apical Na+/H+ exchanger involved in neutral NaCl absorption. In conclusion, we have cloned a third rabbit Na+/H+ exchanger isoform which is only expressed in kidney and intestine. This 832-amino acid protein is likely to be a glycoprotein, is likely regulated by protein kinases, and is a candidate to be an epithelial cell apical Na+/H+ exchanger. Note Added inProof-After the submission of the paper, we stably transfected the full-length NHE-3 cDNA into PS120 cells, which are Chinese hamster lung fibroblasts deficient in all endogenous Na+/H+ exchangers. The transfected cells, after selection by G418 resistance in parallel with acid killing, as described in Ref. 10,demonstrated 1) amiloride sensitive Na+-dependent alkalinization as measured by fluorometry using the pH-sensitive dye BCECF and 2) H+-activated Na+ influx, as measured by *'Na+ uptake, which is also inhibited by amiloride. These resultsshow that NHE-3encodes a functional Na+/ H+ exchanger.
REFERENCES 1. Grinstein, S., Rutin, D., and Mason, M. J. (1989) Biochim. Bwphys. Acta 9 8 8 , 73-97 2. Aronson, P. S. (1985) Annu. Rev. Physiol. 47,545-560 3. Kinsella, J. L., and Aronson, P. S. (1981) Am. J. Physwl. 2 4 1 , C220-C226 4. Knickelbein, R. G., Aronson, P. S., Atherton, W., and Dobbins, J. W. (1983) Am. J. Physiol. 2 4 5 , G504-G510 5. Boron, W.F., and Boulpaep, E. L. (1983) J. Gen. Physiol. 8 1 , 29-52
6. Grinstein, S., Clarke, C.A., and Rothstein, A. (1983) J. Gen. Physiol. 82,619-638 7. Pouyssegur, J. (1985) Trends Biochem. Sci. 10,453-455 8. Sardet, C., Franchi, A., and Pouyssegur, J. (1989) Cell 5 6 , 271280 9. Sardet, C., Counillion, L., Franchi, A., and Pouyssegur, J. (1990) Science 247,723-726 10. Tse, C. M., Ma, A. I., Yang, V. W., Watson, A. J. M., Levine, S.,
Montrose, M. H., Potter, J., Sardet, C., Pouyssegur, J., and Donowitz, M. (1991) EMBO J. 10,1957-1967 11. Knickelbein, R.G., Aronson, P. S., and Dobbins, J. W. (1990) Am. J. Physiol. 259, G802-G806 12. Haggerty, C., Agarwal, N., Reilly, R.F.,Adelberg,E. A., and Slayman. C. W. (1988) . . Proc. Natl. Acad. Sci.U. S. A. 88.67976801 13. Igarashi, P., Reilly, R.F., Hildebrandt, F., Biemesderfer, D., Reboucas, N.A., Slayman, C.W., and Aronson, P. S. (1991) Kidney Int. 40, S84-S89 14. Raley-Susman, K. M., Cragoe, E. J., Jr., Sapolsky, R.M., and Kopito, R. R. (1990) J. Biol. Chem. 266, 2739-2745 15. Aronson, P. S., Nee, J., and Suhm, M. A. (1982) Nature 299, 161-163 16. Paris, S., and Pouyssegur, J. (1984) J. Biol. Chem. 2 5 9 , 1098910994 17. Rajendran, V., and Binder, H. J. (1990) J. Biol. Chem. 265,84088414 18. Tse, C.M., Watson, A. J. M., Ma, A. I., Pouyssegur, J., and Donowitz, M. (1991) Gastroenterology 100, A258 19. Helms, C., Graham, M.Y., Dutchik, J. E., and Olson, M.V. (1985) DNA (NY) 4,39-49 20. Weiser, M. M. (1973) J. Biol. Chem. 2 4 8 , 2536-2541 21. Chomcznski, P., and Sacchi, N. (1987) Anal. Biochem. 162,156159 22. Maniatis, T., Fritsch, E., and Sambrook, J. (1989) Molecular Cloning:A Ldoratory Manual, pp. 7.43-7.50. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY 23. Feinberg, A. P., and Vogelstein, B. (1983) Anal. Biochem. 1 3 2 , 6-13 24. Melton, D. A., Krieg, P. A., Rebagliati, M. R., Maniatis, T., Zinn, K., and Green, M. R. (1984) Nucleic Acids Res. 12, 7035-7056 25. Sanger, F., Nicklen, S., and Coulson, A. R. (1977) Proc.Natl. Acad. Sci. U. S. A . 7 4 , 5463-5467 26. Henikoff, S. (1984) Gene (Amst.)2 8 , 351-359 27. Kozak, M. (1987) Nucleic Acids Res. 15, 8128-8148 28. Kyte, J., and Doolittle, R. F. (1982) J. Mol. Biol. 1 5 7 , 105-132 29. Von Heijne, G. (1990) J. Membr. Biol. 115, 195-201 30. Engelman, D., Steitz, T., and Goldman, A. (1986) Annu. Reu. Biophys. Biophys. Chem. 1 5 , 321-353 31. Chou, P. Y., and Fasman, G. P. (1978) Adv. Enzymol. 47, 45147 32. Kopito, R. R., Anderson, M., and Lodish, H. F. (1987) J. Biol. Chem. 262,8035-8040 33. Williams, S. A., and Birnbaum, M. J. (1989) J. B i d . Chem. 2 6 3 , 19513-19518 34. Miller, R. T., Counillon, L., Pages, G., Lifton, R. P., Sardet, C., and Pouyssegur, J. (1991) J. Biol. Chem. 266, 10813-10819 35. Knickelbein, R. G., Aronson, P. S., Seifter, J., Schron, C. M., and Dobbins, J. W. (1985) Am. J. Physiol. 2 4 9 , G236-G245 '
