Jul 5, 1994 - Bernadette Olives$, Philippe NeauO, Pascal BaillyS, Matthias A. Hedigem, Germain RousseletO, ..... Church, G. M., and Gilbert, W. (1984) Proc.
Vol. 269, No. 50, Issue of December 16,pp. 31649-31652, 1994 Printed in U.S.A.
THE JOURNAL OF B I O L ~ I CCHEMISTRY AL 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc
Cloning and Functional Expressionof a Urea Transporterfrom Human Bone Marrow Cells* (Received for publication, July 5, 1994, and in revised form, August 25, 1994)
Bernadette Olives$, PhilippeNeauO, Pascal BaillyS, Matthias A. Hedigem, Germain RousseletO, Jean-Pierre CartronS, and Pierre RipocheOll From the VNSERM U76, Institut National de la Dunsfusion Sanguine, 6 rue Alexandre Cabanel, 75015 Paris, France, the $Ddpartement de Biologie Cellulaire et Mole‘culaire, CEA Saclay, 91191 Gif-sur-Yvette France, and theW e n a l Division, Department of Medicine, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts 02115
A rapid passive urea transport has been previously described in the mammalian renal inner medullary collecting duct epithelial cells and in mammalian erythrocytes. Recently,a vasopressin-regulated urea transporter(UT2) has been cloned fromarabbit kidney medullary cDNA library (You, G., Smith, C. P., Kanai, Y., Lee, W. S., Stelzner, M., and Hediger, M. A. (1993)Nature 365,844-847).We now report the cloning and characterization of a complementary DNA (HUTl1) encoding an urea transporter isolated from a human bone marrow library. Itencodes a 43,000-Dapolypeptide of 391 amino acids that exhibited 63%sequence identity with the rabbit urea transporter and a similar membrane topology. HUTll carries 2 putative glycosylation sites and 10 cysteines, of which only 7 are conserved at an equivalent position in UT2. HUT11transcripts have been identified in human erythroid and renal tissues. Expression studies inXenopus oocytes demonstrated that HUT11 mediates a facilitated urea transport that was inhibited, as described in mammalian erythrocytes, byvery low concentrations of phloretin, p-chloromercuribenzene sulfonate, and urea analogues. No unidirectional movements of charged molecules, glycerol, or waterwere associated with HUTll expression in oocytes. These findings suggest that HUTll is most likely responsible for the facilitated urea transportin human red blood cells.
under the control of hormonal regulation by antidiuretic hormone, the red blood cell urea transport appears to be constitutive, suggestingthe existence of two distinct urea transporters. The molecular characterization of these transporters has considerably progressed over the past 3 years. Recently, we have designed specific photoactivatable inhibitors, which allowed us to demonstrate the involvementof at least a 45-kDa protein in the red blood cell urea transport machinery (9). O n Martial et al.(10) and Zhang and Verkman (11) the other hand, havebeenabletoexpressfunctional urea transporters in Xenopus oocytes by micro-injectionof mRNA purified from amphibianbladder. Further experimentswithmRNApurified from rat kidney and liver (12) suggested the hypothesis of two distinct transporters, one able to respond to vasopressin and the other insensitive to this hormone. Finally, You et al. (13) reported the isolation from rabbit renal medulla of a complementary DNAcoding for a facilitated urea transporter, UT2. Its localization byin situhybridization in the inner medulla and in papilla suggests that this protein is the rabbit antidiuretic hormone-regulated urea transporter. We now report the isolation of a human bone marrow-derived a proteinshowingextensivesecDNA (HUT11)codingfor quence similarity with UT2. We also describe its functional characterization as t h e first cloned human urea transporter.
MATERIALSANDMETHODS DNA and Oligonucleotide Probes-ThecDNAprobe encoding the The rapid transport of urea through the membranesof the rabbit urea transporter(clone UT2)has been described previously(13). and antiinner medullary collecting duct epithelial cells is a critical phe- Sense primer (5’-CAGGTGATGTTTGTGAACAACCCC-3’) (5’-AGCGATGAGGAAAATGCC-3’) that correspond to sense primer of the mamnomenon in the urinary concentrating mechanism amino acids 67-74 (QVMFVNNP) and 242-247 (GIFLIA), respectively, malian kidney (1, 2). Macey a n d Yousef (3) proposed that t h e were synthesized from the sequence of the rabbit urea transporter. An existence of a rapid urea transport in red blood cells could internal oligonucleotide probe (5’-CAGAACCCATGGTGGGCC-3’) was permit their passage without damage through the renal medesigned from nucleotides encodingamino acids 87-92 (QNPWWA). RNAsfrom dulla, where theyare exposed to high urea concentrations. The PCR AmplificationandLibraryScreening-Poly(A+) urea permeability of these cell membranes is 100-1000 times spleen erythroblasts of an adult 6-thalassemic patient were extracted higher than the urea permeability of a lipid bilayer (4). Satu- by the guanidine isothiocyanate/cesiumchloride method(14) and purified on an oligo(dT)-cellulose column. First cDNA strands were syntheration of the transportat high urea concentrations (51, compe- sized using the first-strand cDNA synthesis kit (Pharmacia Biotech tition with urea analogues(6,7), and relatively specific inhibi- Inc.). The cDNA products were enzymatically amplified between the (7,s)strongly suggest sense and antisense primers (1 pg each) by using the Tu4 polymerase. tion by phloretin and mercurial reagents that a facilitated urea diffusion takes place in these tissues. The reaction was carried out in a Perkin-Elmer thermal DNA cycler comprising denaturation at 94 “C for is for30cycles,eachcycle Whereas the inner medullary collecting duct urea transport 1min, primer annealing at 45 “C for 1min, and chain extension at 72 “C * This investigation was supported by the Institut National de la for 1 min. Amplification products of the expected size (540 bp) were Sante et de la Recherche Medicale and by the Centre #Etudes de identified by hybridization with the internal probe, purified, and subSaclay. The costsof publication of this article were defrayed in part by cloned by blunt end ligation t o pUC18 vector. The 540-bpfragment was the payment of page charges. This article must therefore be hereby sequenced and used as a random primed [32PldCTP-labeled probe to screen about 1.6 x lo6 plaques of hgtll human bone marrow cDNA marked “advertisement” in accordance with 18 U.S.C.Section1734 library (Clantech). Positive recombinant clones were isolated and presolely to indicate this fact. 11 To whom correspondence and reprint requests should be addressed: pared with a QIAGEN lambda kit. EcoRI inserts were subcloned and Departement de Biologie Cellulaire et Moleculaire, Service deBiologie sequenced on both strands by the dideoxy chain termination method with Sequenase version 2.0 (U. S. Biochemical Carp.). Cellulaire, Centre #Etudes deSaclay,91191Gif-sur-Yvette,Cedex, France. Tel.: 33-1-69-08-43-21;Telex: ENERG 604641F; Fax: 33-1-69RNA Northern Blot Hybridization Analysis-Total RNAs extracted 08-80-46;E-mail: ripocheBdsvidf,cea.fr. from several tissues and cell lines were prepared as described (14),
3 1649
31650
Cloning and Characterization of a Human Urea Dansporter
resolved by electrophoresis on 6%(w/v) formaldehyde, 1%(w/v) agarose gel, and transferred tonylon filters (Zeta-probe GT, Bio-Rad). Hybridization with the32P-labeledHUTll cDNA probe and stringent washes were performed as described before (15). Expression in Xenopus laeuis Oocytes-The full-length cDNA insert encoding the human urea transporter (clone HUT111 was subcloned into the EcoRV-digested T7TS, a pGEM4ZT-derived vector containing the 5 ' - and 3'-untranslated sequences of the X. laeuis P-globin gene (kindly provided by P. Krieg, Austin,TX). Capped sense and antisense RNAs were synthesized using T7 and SP6 RNA polymerase, respectively, after linearizationof the vector (mCAPmRNA capping kit, Stratagene). Expressionof H U T l l cRNAinXenopus oocytes was studied as described before (10).Oocytes isolated from X. lueuis were dissected manually after a 2 mg/ml collagenase treatment a t 25 "C for 90 min to remove the follicular layerand kept overnightat 18 "C. Water (50 nl) or mRNA (50 nl: 0.4 pg.l.11") was injected into individual oocytes. They were then kept at 18 "C until flux measurements in a daily changed Barth'ssolution (88 mM NaCI, 1 mM KCl, 0.8 mM MgSO,, 0.4 m~ Ca(NO,),, 7.5 mM Tris, 2.4 mM NaHCO,, pH 7.6) containing 100 unitdml penicillin and 0.01 mg/ml streptomycin. To study the abilityof H U T l l to induce a water channel activity,oocytes were transferred tohypoosmotic conditions, and initial changes in theoocyte volume were measured at 22 "C using videomicroscopy (16). Oocyte Ion ?Funsport Actiuity-Electrical conductance of oocyte expressing HUTll was determined by voltage-clamp experiments as described by Maurel et al. (17). Oocytes injected with water were used as controls. Oocyte Urea fiunsport Actiuity-Oocyte urea transport activity was measured by a ['*C]urea uptake on the 3rd day after injection. In all 10 pCi/ml c3Hlrafthese experiments, the incubation medium contained finose (New England Nuclear, Dreieich, FRG) as a control of oocyte integrity.For the inhibitionexperiments,phloretin(Sigma) (0.10.6 mM) and pura-chloromercuribenzene sulfonate (pCMBS)' (Sigma) (0.5 mM) were added, 20 min and 1 h, respectively, before the assay to the oocyte medium and maintained during the urea uptake. Urea anaAldrich), logues suchas tMourea (Sigma), nitrophenylthiourea (NPTU, and (3,4-dichlorophenyl)-2-thiourea(DCPTU, Aldrich), which are inhibitors of urea fluxes, were used a t a concentration close to their respective K, calculated by Mayrand and Levitt(6). Photoufinity Labeling-The photoactivatable urea analogue 143azido-4-chlorophenyl)-3 methyl-2-thiourea (Me-ACPTU, synthesized by the Service des Molecules Marquees, CEA Saclay, France), a specific and reversible inhibitor (when used in the dark) of the humanred blood cell urea transporter, has been testedas described by Martial et al. (18). For photolabeling experiments, incubation with Me-ACPTU was performed for 30 min while oocytes were illuminated with a 12,000 lux beam of polychromatic light. RESULTS AND DISCUSSION
Cloning of the Human Urea l'kansporter-Spleen erythroblast messenger RNAs were used in a reverse transcribed polymerase chain reaction with two primers derived from the amino acid sequence of the UT2 clone (13). Sequence analysis indicated that theexpected 540-bp amplified fragment encoded a peptide sharing 71.4% amino acid sequence identity with the UT2 protein. The 540-bp fragment wasused as a probe to screen a human bone marrow cDNA library constructed in h g t l l . Four positive clones of differentsizes were isolated, among which the clone HUTll carried the largest insert. The coding sequence of HUTll hasbeen deposited to the GenBank data base under theaccession number L36121. This HUTll insert contains 1173 bases of open reading frame and 18 and 405 nucleotides of 5'- and 3'-untranslated sequences, respectively. It encodes a predicted 391-amino acid polypeptide sharing 62.4% sequence identity with the rabbit UT2 protein, which is composed of 397 residues (Fig. 1).Both HUTll and UT2 proteins are synthesized without a peptide signal. The predicted relative molecular mass of the human protein is 43,000 Da, but the protein also carries two potential
10
40
30
20
UT2
PKA PKC V V MEDS-SEIKV ETASSRTSWI QSSVAAGGKR ISRALGYITG
WTll
-sP-
UT2
EMKECAEGLK DKSPVFQFLD wVLRGT-
79
HUT11
% K ~ A N Q L K DKWVLQFID WILRGISQV~ FVNNP~SGIL
80
UT2
JYIGLE'VQNP P SQDRS€G&S.G
119
HUTll
IV'JGLLVQNP
120
UT2
DSPTMVXEN QVSPC@C!
39
FPKALGY~TG 4 0
WWALTGWLGT VVSTLMALLL
SQDRSLIASG
LLIAVFSDKG DYYWWLLLPV IVMSMSCPIL
159
HUT11
LYGYNATLVG VLMAVFSDKG DY~WWLLLPV CAMSM~CPI~ 160
UT2
W T I F S K W
HUT11
S S A L N S ~ S K WDLPVFTLPF NMALSMYLSA TGWYNPFF~A 200
UT2
D
P
H
Y
N
U
'
F
P
T
It
TLLQWSSVP N I
SEIQW ULRAIPVG- -1GQWGCDN
199
237
HUTll UT2 -
P
HUT11
PWTGGIFLGA
277 IUSSPLMCHAAIGSLLGI L A A G L S ~ S A ~ F280
UT2
TWQT H
m
317
HUT11
TWQT HLLALGCALF
320
UT2
WYVGAALTN VLSWGLPTC TWF'FCISALI FLUTTNNPA
357
HUT11
Tny4;VG"gLNFMAEVGLPAC
UT2
IYKLPLSKVT YPEANRWL TQERNRRSST ITKYQAYDVS
HUT11
IYKMPLSKVT
T U P F C ~ T L L F L I M T T ~ N S360 ~
PKC V
YPEENRIFYL-sP
L
397 391
FIG.1. Protein sequence comparisonof the human and rabbit urea transporter.The 391-aminoacid sequence of the humanprotein (HUT11,lower sequence) is aligned with the 397-amino acid sequence of the rabbit protein (UT2, upper sequence). Sequence differences between HUT11 and UT2 are shown with raised letters within the HUTll sequence. Potential protein kinase A(PKA)and protein kinase C (PKC) phosphorylation sites of UT2 are indicated. Potential N-glycosylation sites are boxed. Underlined amino acids within the UT2 sequence represent potential membrane-spanning regions.
N-linked glycosylation sites at Asn-211 and Asn-291. Sequence comparison suggested that potential protein kinase A and protein kinase C phosphorylation sites present in UT2 protein werenot conserved in HUT11. Of the 10cysteineresidues present in each protein, only 7 were conserved and aligned a t equivalent positions. Hydrophobicity profiles for the two proteins are very similar (not shown) and suggest 10 predicted membrane-traversing segments with the N and C terminus localized in the cytoplasm. Assuming that HUTll and UT2 have a similar topology, only the potentialN-glycosylation site at Asn-211 is predicted to be on the extracellular side of the membrane and is likely to carrya N-glycan chain, whereas the site at Asn-291 is located within a membrane-spanning region. The multimembrane-spanning domain structure of HUTll The abbreviations used are: pCMBS, pura-chloromercuribenzene andits sequence analogy t o therabbitureatransporter sulfonate; DCPTU, (3,4-dichlorophenyl)-2-thiourea; Me-ACPTU, 143strongly suggest that HUTll isa human analog of the rabbit azido-4-chlorophenyl)-3 methyl-2-thiourea; NPTU, nitrophenylthiourea transporter. This was supported by functional assays as urea; bp, base pair(s); kb, kilobase paids).
Cloning and Characterization of a Human Urea Dansporter
31651
HUT1 1
a al Y
3
30-
T !
0 L
aJ
a
-0
20v
Q Y
m
Y
a
V d 7
10
20
30
40
50
60
l0I
70 0
I
K '
Incubation time (min)
FIG.2. Time course of ['4Clurea uptake inX. Zaeuis oocytes. The assay was initiated by resuspending individual oocytes in 0.2 ml of Barth's solution containing 1 mM urea, 2 pCdm1 [l4C1urea (Amersham Corp.), and 10 $dm1 [3Hlrafinose. The urea uptake was stopped a€tm the indicated time by additionof 2 ml of ice-cold Barth's solution, followed by rapid filtration through glass fiber filters, which were washed with an additional 5 ml of the same solution.Oocytes were dissolved in glass vials by a 10% SDS treatment. Aftermixing with 5 ml of scintillation liquid (UltimaGold, Packard Instrument Co.), the oocyte-associated radioactivity was determined.
FIG.3. Inhibition ofHUT11-mediated urea transport by phloretin and pCMBS. Oocytes were preincubated 20 and 60 min, respectively. To test the reversibility of the pCMBS inhibition, subsequent incubation for 15 min in 5 mM 6-mercaptoethanol was realized before urea uptake measurements. A, phloretin inhibition: 0.1 nlM = 54 f 6%, 0.2 mM = 88 f 496,0.3 mM = 96 2 396, 0.6 mM = 98 2 2%; B , pCMBS inhibition: 0.5 mM = 65 f 6%.Data from at least three experiments are expressed as the means 2 S.E.
described below. Expression of HUT in Xenopus Oocytes-The time course of [14C]urea(1mM) uptake into HUTl1-cRNA-injected oocytes is shown in Fig. 2. After a 10-min incubation, the uptake of [14C]ureareached a maximumat 280 pmol. This value is in the range of the urea uptake in oocytes injected with UT2 (13). Control experiments performed with oocytes injected with the cRNAcoding for FA-CHIP (a frog-specificwater channel)(16) or water-injected oocytes exhibited a very slow urea uptake, and urea equilibrium was not reached after a 60-min incubation. On the other hand,a depleted urea permeability was measured on oocytes co-injected with sense and antisense HUT11-cRNA. These results suggest that HUTll mediates a facilitated urea transport. The urea permeability (P,,,,) of HUTl1-cRNA-injected oocytes calculated from initial rates of uptake gave a value of 2.4 x c d s , while the P,,,, of water-injected controls ranged between 1.2 and 2.0 x 1O"j c d s . Thus, HUTll increased to 20 times theoocyte plasma membrane urea Permeability. Assuming that HUTll and UT2 were present at the samecopy number in oocytes, the permeabilities of HUT11 and UT2 are comparable. Considering the number of urea transporters in red blood cells as 32,000 (19), the urea permeability at 25 "C as 2.7 x c d s (51, andthe surface areaas 1.4 x cm', the cm%. We individual carrierhas a permeability of 1.1 x then deduce that the number of functional protein copies per oocyte is close t o lo8 units. The total copy number is probably much higher, as shown for the GLUT1 and GLUT4 transporters where only a weak percentage of carrier was inserted in the plasma membrane (20). Urea uptakeof HUT11-cRNA-injectedoocytes was measured in thepresence of two known urea transport inhibitors, phloretin and pCMBS (Fig. 3). Phloretin (0.1 mM) inhibited P,,,, by 54%, while 0.2 mM phloretin reduced P,, to the basal urea permeability of oocyte plasma membrane. By comparison, P,,,,
ofUT2-injected oocytes was inhibitedby only 50%with 0.35 mM phloretin (13). This high efficiency of phloretin to inhibit the urea transport in oocytes injected with HUT11-cRNA can be related to the low inhibition constant (50%)of phloretin for red bloodcell urea transport inhibition.A1-hincubationwith 0.5 mM pCMBS reduced the urea transport by 65%. This mercurial inhibition was reversed by subsequent incubation in the reducing agent P-mercaptoethanol (5 mM).As in human red blood cells, P,,,, of HUT11-injected oocytes is strongly inhibited by a very low concentration of pCMBS (5). Urea analogues (50 mM thiourea, 0.1 mM NPTU, and 0.1 mM DCPTU) inhibited ureatransport by 88, 35, and 54%, respectively (Table I). This pattern of inhibition is similar to the one obtained in erythrocytes (6). Photoaffinity labeling with the photoactivatable urea analogue Me-ACPTU (0.1 mM) inhibited P,,,, by 73%. This inhibition was reversed by simple washout in the dark, but it was maintained when oocytes were illuminated for 30 rnin before washout of the inhibitor. Here again, the patternof inhibition and photolabeling obtained with Me-ACPTU on HUT11injected oocytes was similar to the pattern described with human red blood cells (9). The osmotic water permeabilities ofX. Zaeuis oocytes injected with water or with HUT11-cRNA did not change significantly and were, respectively, 6.5 2 1.6 x c d s (n = 5) and 7.4 2 0.9 x c d s (n = 5). Thus, we can exclude the possibility that HUT induces a water channel activity. On the other hand, no significant increase of the glycerol permeability (Pgly) was observed. With HUT11-cRNA-injected oocytes, P,, was 1.01 r 0.34 x c d s (n = 51, and the water-injected oocytes, P,, was c d s (n = 5). 1.15 r 0.36 x Electrical conductances of the plasma membrane of HUT11-cRNA-injected or water-injected oocytes are presented in Fig. 4. The similarity between the two curves supports the idea that no unidirectional movements of charged molecules
Cloning and Characterization
31652
of a Human Urea Dansporter
TABLEI Inhibition of HUTll-mediated urea transport by urea analogues Data (from at least three experiments) are expressed as the means2 S.E. Photoaffinity labeling experiments were realized with Me-ACPTU as described by Martial et al. (10). Inhibitor concentration Me-ACPTU Thiourea N€TU DCPTU 50 mM O.' mM O.' mM 0.1 m~ Washout Light + washout
% ofinhibition 8 8 + 6 3 5 . ~ 4 5 4 + 8 7 3 2 8 2 4 +
11
5 8 2 12
zocl-
-2 -
28SrRNA
+
l8SrRNA
+ L.,
va
FIG. 5. Northern blot analysis of HUTll transcript.Total RNAs (15 pgllane) fromHEL, B-lymphocyte (EBV+), HMy2.CIRmonocytic cell lines, or tumoral kidney or poly(A+)RNA (3 pg) from adult spleen erythroblasts and adult liver tissues were separated by agarose gel electrophoresis and hybridized with the 32P-labeledHUTll probe as described under "Materialsand Methods."Equal loading and absence of degradation were checked by staining with ethidium bromide. Autoradiography was for 72 h at -80 "C.
Water
1 oaI -
Y
c
E
L 3
The pattern of inhibition of the HUT11-mediated urea transport, the bone marrow origin of the HUTll cDNA, and Northern analysis support the view that HUTll could be responsible for the facilitated urea transport in human erythrocytes. Then, in contrast with UT2, HUTll could be a vasopressinindependent urea transporter. The availability of a cDNA coding for a human transporter will allow new experimental approaches for studying the facilitated urea transport in humans.
u
0
-1 00 7-1 50
-50
Membrane potential
50
(mv)
FIG.4. Membrane current-to-voltage relationship in oocytes injected with water or HUT11-cRNA. Membrane potential was held at -60 mV and stepped for 3 s to potentials from -120 to +80 mV with the membrane potential being returned to the holding potential after each step. Data are represented as the mean S.E. of currents measured on 12 cells.
are associated with HUTll expression in oocytes. In both types of injection, plasma membrane oocyte conductance was 0.51 2 0.11 microsiemens (n = 12). Northern blot analysis with the HUTll probe revealed the presence of 2.5- and 4.7-kb transcripts in human adultspleen erythroblasts and tumoral kidney. In the human erythroleukemic cell line HEL, 2.0- and 4.3-kb transcripts were detected, but there was no signal in RNA preparations from adult liver, monocyte(YMyB.CIR), and B-lymphocyte (EBV+) cell lines (Fig. 5). These findings suggest that the human urea transporter is present both in erythroid and renaltissues. Thisis in contrast with the rabbit UT2 transporter, which was absent from the spleen of anemic mice and rabbits (13).Whether the different transcripts identified in human tissues result from alternative splicing or poly(A+) choice of a single mRNA species is presently unknown. In addition, further investigations should determine the localization of HUTll transcripts along nephron segments and whether one or several species of urea transporter arepresent in the renal medulla.
Acknowledgments-We thank Frkdkrique Tacnet and Jacques Stinnakre for the voltage-clamp experiments, Veronique Berthonaud for oocyte injections, and P.-H. Romeo (INSERM U91,Crkteil, France) for the gift of f3-thalassemic erythroblasts. REFERENCES Sands, J. M., and Knepper, M. A. (1987)J. Clin. Inuest. 79, 138-147 Knepper, M. A., and Star, R. A. (1990)Arn. J. Physiol. 259, F393-F401 Macey, R. I., and Yousef, L. W. (1988)Am. J. Physiol. 254, C669-C674 Galluci, E., Micelli, S., and Lippe, C. (1975)Nature 255, 722-723 Brahm, J. (1983) J. Gen. Physiol. 82, 1-23 Mayrand, R. R., and Levitt, D. G. (1983)J. Gen. Physiol. 81,221-237 Chou, C. L., and Knepper, M. A. (1989)Am. J. Physiol. 257, F359-F365 Macey, R. I., and Farmer, R.E.L. (1970)Biochirn. Biophys. Acta 211,104-106 Neau, P., Degeilh, F., Lamotte,H., Rousseau, B., and Ripoche,P. (1993)Eur. J. Biochern. 218,447-455 10. Martial, S.,Ripoche, P., and Ibarra, C. (1991) Biochirn. Biophys. Acta 1090, 86-90 11. Zhang, R., and Verkman, A. S. (1991)Am. J. Physiol. 260, C 2 M 3 4 12. Hasegawa, H., and Verkman, A. S. (1993)Am. J . Physiol. 265, C5144520 13. You, G., Smith, C. P., Kanai, Y., Lee, W. S.,Stelzner, M., and Hediger, M. A. (1993)Nature 365, 844-847 14. Maniatis, T.,Fritsch, E. F., and Sambrook, J. (1982) Molecular Cloning: A Laboratow Manual, pp. . . HarborLaboratory,Cold .. 7.19-7.22, Cold Spring Spring Harbor, NY 15. Church, G. M., andGilbert, W. (1984) Proc. Natl.Acad.Sci. U.S . A . 81, 1991-1995 16. Abrami, L., Simon, M., Rousselet, G., Berthonaud,V.,Buhler, J. M., and Ripoche, P. (1994) Biochim. Biophys. Acta 1192,147-151 17. Maurel, C., Reizer, J., Schroeder, J. I., and Chrispeels, M. J. (1993)EMBO J. 6,2241-2247 18. Martial, S., Neau, P., Degeilh, F., Lamotte, H., Rousseau, B., and Ripoche, P. (1993) EW J. Phy~iol.423,51-58 19. Mannuzzu,L. M., Moronne, M. M., and Macey, R.I. (1993)J. Membr. Biol. 133, 85-97 20. Nishimura, H., Pallardo, F., Seidner, G. A,, Vannucci, S., Simpson, I. A., and Birnbaum, M. J. (1993) J. Biol. Chern. 268,8514-8520 1. 2. 3. 4. 5. 6. 7. 8. 9.
