- Email: [email protected]
The C-terminal tail of aquaporin-2 determines apical trafficking
Kidney International, Vol. 68 (2005), pp. 1999–2009
The C-terminal tail of aquaporin-2 determines apical trafficking MICHIO KUWAHARA, TOMOKI ASAI, YOSHIO TERADA, and SEI SASAKI Department of Nephrology, Tokyo Medical and Dental University Graduate School, Tokyo, Japan
The C-terminal tail of aquaporin-2 determines apical trafficking. Background. Aquaporin-2 (AQP-2) proteins are mainly expressed at the apical region of the collecting duct cells. We previously reported three different mutations in the C-terminus of AQP-2 that all-cause autosomal-dominant nephrogenic diabetes insipidus. When one of these mutant AQP-2s was expressed in Madin-Darby canine kidney (MDCK) cells, it was mistargeted to the basolateral membrane, suggesting a critical role of the C-terminal tail in the apical trafficking of AQP-2. Methods. Portions of the AQP-2 C-terminal tail (residues 226271) were mutated by the polymerase chain reaction (PCR) technique and inserted into the pcDNA3.1 vector. Constructs were transfected into MDCK cells to examine the localization of mutated AQP-2 proteins by immunofluorescence microscopy. Cell surface expression was detected by biotinylation assay. Results. The wild-type AQP-2 was localized at the apical membrane, whereas mutants lacking residues 262-271 (the last 10 amino acids) were predominantly distributed in the endoplasmic reticulum. Deletion mutants of the initial (226-240del) and middle (241-252del) portions of the C-terminal tail were identified at the apical membrane, suggesting that residues 226252 have no involvement in apical targeting. An AQP-4–AQP-2 chimera in which a portion of the AQP-4 C-terminal tail was replaced by the corresponding site in AQP-2 (residues 256-271) was found at the apical membrane. The sequence of the last 4 amino acids of AQP-2 (G-T-K-A) corresponds to a PDZinteracting motif. Our investigations identified a mutant of this portion mostly localized to the subapical region. Further, apical expression was found to be significantly decreased in mutants lacking a consensus sequence for cyclic adenosine monophosphate (cAMP)–dependent phosphorylation (residues 253-256). Conclusion. The sequence at 256-271 is sufficient for apical trafficking in AQP-2. The putative PDZ-interacting motif (GT-K-A, residues 268-271) plays a key role in apical membrane expression. In addition, cAMP-dependent phosphorylation was found to be critical for apical targeting.
Water movement across the cell membrane is an essential process in the maintenance of cellular homeostasis. Water molecules move through the aquaporin (AQP) Key words: AQP-2, C-terminal tail, PDZ-interacting motif, signalinduced proliferation-associated gene-1. Received for publication October 5, 2004 and in revised form March 15, 2005, and May 13, 2005 Accepted for publication June 3, 2005 C
2005 by the International Society of Nephrology
water channels, a family of selective pores that belong to the major intrinsic protein (MIP) superfamily [1, 2]. Hydropathy analysis of the MIP superfamily predicts six transmembrane regions with N- and C-termini localized in the cytoplasm. AQPs are found virtually in all species, including more than ten kinds so far identified in mammals [1, 3, 4]. Aquaporin-2 (AQP-2) is a vasopressinregulated water channel that plays a central role for the urinary concentration from its sole site of expression in the principal cells of the connecting tubule and collecting duct cells [5, 6]. The binding of vasopressin to the V2 receptor on the basolateral membrane is generally thought to activate cyclic adenosine monophosphate (cAMP)– dependent protein kinase, a process that encourages cytoplasmic AQP-2–containing vesicles from the subapical region to infiltrate the apical membrane, and thereby increase its water permeability [7]. After the effect of vasopressin subsides, AQP-2 channels are retrieved from the apical membrane by endocytosis and the water permeability returns to a low level. In our previous report, we identified three mutations of the AQP2 gene confirmed to cause the autosomaldominant form of nephrogenic diabetes insipidus (NDI) [8]. All three of these mutations were localized to a single allele in exon 4 of the AQP2 gene at the C-terminal tail (721delG, 763-772del, and 812-818del). The wild-type AQP-2 is a 271 amino acid protein, whereas these mutant genes were predicted to encode 330-333 amino acid proteins due to a frameshift mutation of the stop codon. Our Xenopus oocyte expression study of these mutants revealed an impairment in the trafficking to the plasma membrane [8]. When one of these mutants (763-772del) was expressed in polarized epithelia of Madin-Darby canine kidney (MDCK) cells, the mutant AQP-2 was mistargeted to the basolateral membrane, suggesting that the AQP-2 C-terminus was responsible for apical targeting [9]. Misrouting of C-terminal mutant AQP-2 has been also reported in other cases of autosomal-dominant NDI [10, 11]. These observations suggest that the C-terminal tail of AQP-2 plays a critical role in apical trafficking. The C-terminal tail of AQP-2 is composed of 46 amino acid residues [5]. We hypothesized that a portions of the residues in this sequence interact with proteins responsible for apical targeting. The AQP-2 C-terminal tail 1999
2000
Kuwahara et al: Apical trafficking of AQP-2 and C-terminus
Table 1. C-terminal amino acid sequence of the wild-type and mutant AQP2s AA 230 WT-AQP2 262-271del 252-271del 242-271del 232-271del 226-240del 241-255del 241-252del 269-271del 256-268del 262-268del S256A
240
250
260
270
PAKSLSERLAVLKGLEPDTDWEEREVRRRQSVELHSPQSLPRGTKA PAKSLSERLAVLKGLEPDTDWEEREVRRRQSVELHS PAKSLSERLAVLKGLEPDTDWEEREV PAKSLSERLAVLKGLE PAKSLS –——————————EPDTDWEEREVRRRQSVELHSPQSLPRGTKA PAKSLSERLAVLKGL—–——————————SVELHSPQSLPRGTKA PAKSLSERLAVLKGL-–————————–RRQSVELHSPQSLPRGTKA PAKSLSERLAVLKGLEPDTDWEEREVRRRQSVELHSPQSLPRG PAKSLSERLAVLKGLEPDTDWEEREVRRRQ—————————TKA PAKSLSERLAVLKGLEPDTDWEEREVRRRQSVELHS—————TKA PAKSLSERLAVLKGLEPDTDWEEREVRRRQAVELHSPQSLPRGTKA
C-terminal tail of the wild-type (WT) AQP2 is composed of 46 amino acid (AA) residues, starting from P231 to A271.
contains several possible domains/motifs associated with protein trafficking. Residues 253-256 (R-R-G-S) make up a consensus sequence for cAMP-dependent protein kinase (R-R-X-[S/T]) [5]. Indeed, cAMP-dependent phosphorylation was observed in wild-type AQP-2, but not in S256A AQP-2 [12, 13]. Further, the plasma membrane expression of AQP-2 was found to be inhibited in S256A AQP-2 [13–16]. The final four amino acid sequence of AQP-2 (G-T-K-A) corresponds to a class I PDZ domain binding motif (X-[S/T]-X-). Similarly, residues 231-234 (S-E-R-L) comprise a class III PDZ-interacting motif (-X-[D/E/K/R]-X-). Given that PDZ proteins are known to tether membrane proteins [17, 18], these domains may play a role in the apical distribution of AQP-2. In the present study we transfected several AQP-2 mutants in MDCK cells and identified their intracellular localization to determine the role of the AQP-2 C-terminus in apical targeting. METHODS Preparation of cDNA and construction of mutants Wild-type cDNA of human AQP-2 with a Kozak consensus translation initiation site just upstream of the initial codon (ATG) was inserted into the pcDNA3.1 vector (Invitrogen, CH Groningen, The Netherlands) between the HindIII site and XbaI site [9]. To generate hemagglutinin (HA) epitope-tagged AQP-2, the nucleotide sequence encoding the HA tag derived from human hemagglutinin protein (YPYDVPDYA) was introduced just after the initiation codon of AQP-2 cDNA. A method of polymerase chain reaction (PCR)–based mutagenesis was used to construct a series of C-terminal mutants of HA-tagged AQP-2 (Table 1), as described previously [9, 19]. The PCR products were digested with HindIII and XbaI, and inserted into the pcDNA3.1 vector (Invitrogen). All of the mutations were confirmed by a fluorescence DNA sequencer. In a separate experiment, HA-tagged rat AQP-4 was cloned into the pcDNA3.1 vector. A part of the AQP-4 C-terminal
tail (residues 285-323) (SQVETEDLILKPGVVHVIDIDRGDEKKGKDSSGEVLSSV) was replaced by the corresponding site (residues 256-271) (SVELHSPQSLPRGTKA) from AQP-2 (AQP-4–AQP-2). A Cterminal deletion mutant of AQP-4 (285-323del) was also generated. Cell culture and stable transfection MDCK cells (Japanese Collection of Research Biologicals #9029, Osaka, Japan) were cultured and transfected stably with wild-type or mutant AQP-2 and AQP-4 using previously described methods [9]. Cells were grown at 37◦ C under a humidified atmosphere of 5% CO 2 . Ten micrograms of the expression constructs of AQP2 were transfected into MDCK cells grown subconfluently on dishes using the calcium-phosphate precipitation technique. After 24 hours, the cells transfected with the wild-type AQP-2 and mutant AQP-2 were trypsinized, split among six dishes, and cultured in Dulbecco’s modified Eagle’s medium (DMEM) containing 700 lg/mL of G418 (Sigma Chemical Co., St. Louis, MO, USA). Ten to 14 days after the transfection, individual colonies were selected using cloning rings and expanded. The clones were isolated after the selection. Confocal immunofluorescence microscopy Transfected MDCK cells were grown on a permeable supported membrane (Transwell-Clear) (Corning, Cambridge, MA, USA). After 3-day culture in DMEM, confluent cells were incubated for 30 minutes with DMEM containing 10−5 mol/L forskolin (Sigma Chemical Co.) to induce apical expression of AQP-2. In some protocols, cells were incubated for 30 minutes with 0.5 mmol/L 8-(4-chlorophenylthio)-cAMP (Sigma Chemical Co.). The cells were then fixed with 2% paraformaldehyde in serum-free DMEM, washed with phosphate-buffered saline (PBS), permeabilized with 0.2% Triton X-100 in PBS, blocked with 1% bovine serum albumin (BSA) in PBS at 4◦ C overnight, and
Kuwahara et al: Apical trafficking of AQP-2 and C-terminus
stained [9]. Antibodies against the following proteins were used to double stain the proteins in the transfected cells: wild-type AQP-2 (rabbit or rat, an antibody against 15 C-terminal amino acids generated by Sawady Technology) (Tokyo, Japan) (diluted 1:400); HA peptide (rat) (1:200) (Roche Diagnostics, Penzberg, Germany); Na-K-adenosine triphophatase (ATPase), a basolateral membrane marker protein (mouse) (1:400) (Upstate Biotechnology, Lake Placid, NY, USA); 78 kD glucose-regulated protein (GRP 78), an endoplasmic reticulum marker protein (rabbit) (1:200) (Stressgen, Victoria, Canada); TGN 38, a trans-Golgi marker protein (mouse) (1:200) (BD Biosciences, San Jose, CA, USA); lysosome-associated membrane protein 1 (Lamp1), a late endosomal/lysosomal marker protein (rabbit) (1:200) (kindly provided by Dr. E. Kominami, Department of Biochemistry, School of Medicine, Juntendo University); and vesicle-associated membrane protein 2 (VAMP2), a intracellular vesicle marker protein (rabbit) (1:200) (Calbiochem, La Jolla, CA, USA). After 1-hour incubation with a mixture of two antibodies in PBS containing 1% BSA, the cells were washed with PBS, further incubated with an antibody against rabbit or rat IgG (1:200) (Alexa Fluor 488) (Molecular Probes, Eugene, OR, USA) and rat or rabbit IgG (1:200) (Alexa Fluor 568) (Molecular Probes) for 1 hour, washed again with PBS and mounted in a mounting medium (Prolong Antifade Kit) (Molecular Probes). In some protocols the cells were pretreated with 100 mmol/L of trimethylamineN-oxide or 2% dimethyl sulfoxide (DMSO) as a chemical chaperone for 4 hours just before the experiment. Immunofluorescent images were obtained with an LSM510 laser scanning confocal microscope system (Carl Zeiss, Jena, Germany) using a 63× water-immersion objective. Immunoblot analysis The immunoblot analysis was performed as described earlier [9]. Initially, the transfected cells were homogenized for 30 minutes in radioimmunoprecipitation assay (RIPA) buffer [50 mmol/L Tris-HCl, 150 mmol/L NaCl, 1.0% NP-40, 0.5% deoxycholate, 0.1% sodium dodecyl sulfate (SDS), pH 8.0] containing a protease inhibitor cocktail (Boehringer Manheim, Manheim, Germany). Next, the samples were centrifuged, denatured and separated by SDS-polyacrylamide gel electrophoresis (PAGE), and transferred to Immobilon-P filters (Millipore, Bedford, MA, USA) using a semidry system. Ten micrograms of total membrane protein was loaded in each lane. The filters were examined by autoradiography after blocking, and two 1-hour incubations with antibody against HA peptide and 125 I-protein A solution, respectively. Mock-transfected cells were treated similarly to provide control. In one protocol, the samples were incubated with endoglycosidase H (Boehringer
2001
Manheim) according to the manufacturer’s instructions. The band intensities were calculated using NIH images to compare the levels of protein expression quantitatively. Four independent sets of experiments were performed in each sample. Results were expressed as means ± SE in arbitrary units and statistical significance was evaluated by analysis of variance (ANOVA). Cell surface biotinylation Cells were seeded at a density of 1.5 × 105 cells/cm2 on a 12-well membrane support (1.0 cm2 ) (TranswellClear) (Corning). After a 3-day culture in DMEM, confluent cells were incubated for 30 minutes with 10−5 mol/L forskolin (Sigma Chemical Co.). Confluent monolayer cells were washed twice with ice-cold PBS containing 0.1 mmol/L CaCl 2 + 1 mmol/L MgCl 2 (PBSCM). All subsequent steps were essentially performed at 4◦ C. Sulfo-NHS-LC-Biotin (2 mg/mL) (Pierce, Rockford, IL, USA) in biotinylation buffer (2 mmol/L CaCl 2 , 150 mmol/L NaCl, 10 mmol/L triethanol amine, pH 8.0) was added to either the apical or basolateral compartment of the filter chamber. Compartments not receiving Sulfo-NHS-LC-Biotin were filled with an equivalent volume of biotinylation buffer alone. After 20 minutes of incubation with gentle agitation, the cells were washed with biotinylation buffer and the labeling procedure repeated. Unreacted Sulfo-NHS-LC-Biotin was quenched by adding quenching buffer (100 mmol/L glycine in PBS-CM). The cells were washed twice with ice-cold PBS-CM and the filters were cut from the plastic support and incubated for 30 minutes at 37◦ C with 1 mL of lysis buffer [150 mmol/L NaCl, 50 mmol/L TrisHCl at pH 7.5, 5 mmol/L ethylenediaminetetraacetic acid (EDTA), 1% NP-40, 0.1% SDS, 0.5% deoxycholate) containing a protease inhibitor cocktail. The cells were scraped from the filters, transferred to Eppendorf tubes, and centrifuged at 14,000g for 5 minutes. Next, 40 lL of streptavidin-agarose (Sigma Chemical Co.), prepared by two washes with high salt buffer (500 mmol/L NaCl, 5 mmol/L EDTA, 50 mmol/L TrisHCl, pH 7.5, 0.1% Triton-X) and two washes with lysis buffer was added to the supernatant. After a 12-hour incubation at 4◦ C with end-over-end rotation, the beads were centrifuged at 14,000g for 5 minutes and washed twice with high-salt buffer, twice with lysis buffer, and once with 10 mmol/L Tris-HCl, pH 7.5. After removing the supernatant, the beads were resuspended in Laemmli sample buffer and denatured for 30 minutes at 37◦ C. Subsequent procedures were essentially same as those used for the immunoblots. Four to six independent experiments were done and the band intensities obtained by NIH images were expressed as means ± SE in arbitrary units. Statistical significance was evaluated by ANOVA or nonpaired t test.
2002
Kuwahara et al: Apical trafficking of AQP-2 and C-terminus
Fig. 1. Immunocytochemical localization of the wild-type and mutant aquaporin-2 (AQP2) transfected in Madin-Darby canine kidney (MDCK) cells. (A to C) Double staining of the wild-type AQP-2 and Na-K-adenosine triphosphatase (ATPase). The wild-type AQP-2 was detected by an antibody against 15 C-terminal amino acids of AQP-2 (A) and the HA-tagged wild-type AQP-2 was detected by an antibody against HA peptide after (B) and before (C) incubation with forskolin. (D to H) Double staining of the 262-271del and Na-K-ATPase (D), 78 kD glucose-related protein (GRP 78) after (E) and before (F) forskolin treatment, trans-Golgi marker protein (TGN 38) (G), and lysosome-associated membrane protein 1 (Lamp1) (H). (I and J) Double staining of the 262-271del and Na-K-ATPase (I) and GRP 78 (J) after pretreatment with a chemical chaperone, trimethylamine-N-oxide. (K) Double staining of Na-K-ATPase and the 226240del. (L to N) Double staining of the 241255del and Na-K-ATPase (L) and TGN 38 after (M) and before (N) the incubation with forskolin. (O and P) Double staining of NaK-ATPase and the 241-252del after (O) and before (P) the forskolin treatment. Cells were viewed with a confocal microscope at 630×. XY and YZ sections are shown in the upper and lower panels, respectively (white bars indicate 10 lm). HA, hemagglutinin.
RESULTS Figure 1A shows the double staining of the wild-type AQP-2 transfected in MDCK cells and a basolateral membrane marker, Na-K-ATPase. An antibody against the AQP-2 C-terminal detected AQP-2 proteins at the apical membrane. Similarly, an antibody against HA identified the apical distribution of the HA-tagged wild-type AQP-2 (Fig. 1B), suggesting that the insertion of the HA tag had no inhibitory effects on the apical expression of the wild-type AQP-2. Based on this finding, we used the anti-HA antibody to identify HA-tagged AQP2 mutants in subsequent protocols. Before treatment with forskolin, the wild-type AQP-2 appeared to be distributed at the subapical region besides the apical membrane (Fig. 1C). Thus forskolin further redistributed the wild-type AQP-2 to the apical membrane. Similar findings were observed when cells were incubated with 8(4-chlorophenylthio)-cAMP (not shown). These results suggested that the stimulation of cAMP-dependent protein kinase enhanced apical AQP-2 trafficking. In contrast to the wild-type AQP-2, the 262-271del, an AQP-2
mutant lacking the last 10 amino acid residues, was localized in cytoplasm with a granular pattern (Fig. 1D). Double staining of the 262-271del and 78 kD glucoseregulated protein, an endoplasmic reticulum marker protein, revealed a close colocalization irrespective of forskolin treatment (Fig. 1E and F). On the other hand, the distributions of the 262-271del proteins were not coincident with that of TGN 38, a trans-Golgi marker protein (Fig. 1G), and slightly overlapped with that of Lamp1, a late endosomal/lysosomal marker protein (Fig. 1H). The intracellular localizations of the 252-271del, 242271del, 232-271del were similar to that of the 262-271del (not shown). On the basis of these results, we speculated that AQP-2 mutants lacking the C-terminal 262-271 residues were basically retained in the endoplasmic reticulum through the action of the endoplasmic reticulum quality-control mechanisms. Preincubation with a chemical chaperone, trimethylamine-N-oxide, redistributed the 262-271del proteins mainly at the subapical area (Fig. 1I). A small portion of the 262-271del proteins was colocalized with TGN 38 in this condition (Fig. 1J). A similar
2003
Kuwahara et al: Apical trafficking of AQP-2 and C-terminus
Table 2. Apparent localization of the wild-type and mutant aquaporins (AQPs) Construct
Localization of AQP protein after [F(+)] and before [F(−)] forskolin treatment
Wild-type AQP-2 262-271del 252-271del 242-271del 232-271del 226-240del 241-255del 241-252del 269-271AAA 269-271del 256-268del 262-268del S256A AQP4-AQP2
Intracellular F(+) apical, F(−) apical, subapical F(+), F(−) subapical, apical F(+) apical, F(−) apical, subapical F(+), F(−) apical, subapical F(+), F(−) apical, subapical Intracellular Intracellular F(+), F(−) intracellular, apical F(+) apical, F(−) apical, subapical Intracellular Intracellular F(+), F(−) intracellular, apical F(+) apical, F(−) apical, subapical
Number of cell lines
Number of cells
3 4 4 4 3 3 3 3 4 3 3 3 4 3
35 30 18 14 15 28 31 28 25 31 17 16 30 27
Abbreviations are: F(+), apical; F(−), apical, subapical F(+); F(−), intracellular.
redistribution of the 262-271del was observed after incubation with DMSO, another chemical chaperone (not shown), suggesting that chemical chaperones may be able to partially correct the retention of the mutant AQP-2 in the endoplasmic reticulum, as previously reported [20, 21]. In contrast to the truncation mutants, the 226-240del, a mutant lacking the initial one third of the C-terminal tail, was confined to the apical membrane area (Fig. 1K). It thus appeared that the consensus sequences of the class III PDZ domain (residues 231-234) play no part in the apical targeting of AQP-2. The 226-240del was distributed apical and subapical region before incubation with forskolin (not shown). Staining of the 241-255del, the deletion of the middle portion of C-terminus was identified at the subapical cytoplasm as well as the apical membrane (Fig. 1L). Some of the 241-255del proteins were colocalized with TGN 38 (Fig. 1M), suggesting that the 241-255del was distributed at both the Golgi complex and apical membrane. Localization of the 241-255del seemed to be similar before and after incubation with forskolin (Fig. 1M and N). In contrast, apical localization was observed in the 241-252del (Fig. 1O), a mutant in which a consensus sequence for cAMP-dependent protein kinase is preserved. Forskolin as well as 8-(4chlorophenylthio)-cAMP apparently stimulated the apical delivery of the 241-252del to some extent (Fig. 1O and P). These findings had two implications: first, that the sequence between 226 and 252 plays no part in apical sorting, and second, that cAMP-dependent phosphorylation contributes to the translocation of AQP-2 from the subapical region to the apical membrane. Table 2 summarizes the apparent localization, the number of generated cell lines, and the number of cells examined in each AQP-2 construct. Next, we decided to perform a chimera study to make up for the shortfall of meaningful results in the truncation studies due to the endoplamic reticulum retention. AQP-4 proteins are known to exist at the basolateral
membrane of the principal cells in the collecting duct. As expected, the wild-type AQP-4 transfected in MDCK cell was expressed at the basolateral membrane (Fig. 2A). Deletion of the last 39 amino acid residues of AQP4 (285-323del) did not change the basolateral expression (Fig. 2B). In contrast, AQP-4–AQP-2, a chimera in which the last 16 amino acids of AQP-2 (residues 256271) were substituted for the corresponding site in the AQP-4 C-terminal tail (residues 285-323), was localized at the apical membrane (Fig. 2C). These results suggested that the residues at 256-271play a critical role in the apical trafficking. Forskolin appeared to stimulate the apical trafficking of AQP-4–AQP-2 slightly (Fig. 2C and D). 8(4-chlorophenylthio)-cAMP had a similar effect on the AQP-4–AQP-2 (not shown). Our next endeavor was to examine the role of amino acids at 256-271 in further detail. The last three amino acids of AQP-2 were replaced by alanines (269-271AAA) due to the compatibility of the sequence of the last four amino acids of (G-T-K-A) with a PDZ-interacting domain. The 269-271AAA was apparently distributed at the apical side (Fig. 3A), but not confined to the apical membrane in the manner seen in the wild-type AQP-2 (Fig. 1A and B). The localization of the 269-271AAA was similar even before the incubation with forskolin (Fig. 3B), suggesting that forskolin had little effect on the distribution of the 269-271AAA proteins. Double staining of the 269-271AAA and intracellular organelle markers showed that the 269-271AAA was rarely colocalized with 78 kD GRP, TGN 38, and Lamp1 (Fig. 3C to E), but closely colocalized with VAMP2 (Fig. 3F). The distribution of the mutant lacking the last three amino acids of AQP-2 (269-271del) was similar to that of the 269-271AAA (not shown). These observations suggested that the putative PDZ-binding motif (residues 268-271) plays a critical role in trafficking from the intracellular vesicles to the apical membrane. The other mutants tested, namely, the 256-268del and 262-268del were
2004
Kuwahara et al: Apical trafficking of AQP-2 and C-terminus
Fig. 2. Immunocytochemical localization of the wild-type and mutant aquaporin-4 (AQP4) doubly stained with Na-K-adenosine triphosphatase (ATPase). (A) The wild-type AQP-4. (B) 285-323del. The AQP-4 mutant that lacks the last 39 amino acid residues. (C and D) AQP-4–AQP-2. The AQP-4 chimera (residues 1-284 in AQP-4) with the C-terminal tail of AQP-2 (residues 256-271 in AQP-2) after (C) and before (D) the incubation with forskolin. Cells were viewed with a confocal microscope at 630×. XY and YZ sections are shown in the upper and lower panels, respectively (white bars indicate 10 lm).
Fig. 3. Immunocytochemistry of aquaporin-2 (AQP-2) and the basolateral membrane or intracellular organelle. (A to F) Double staining of the 269-271del and Na-K-adenosine triphophatase (ATPase) after (A) and before (B) the forskolin treatment, 78 kD glucose-related protein (GRP 78) (C), trans-Golgi marker protein (TGN 38) (D), lysosome-associated membrane protein 1 (Lamp1) (E), and vesicle-associated membrane protein 2 (VAMP2) (F). (G to J) Double staining of the S256A and Na-K-ATPase after (G) and before (H) the incubation with forskolin, TGN 38 (I) and VAMP2 (J). Cells were viewed with a confocal microscope at 630×. XY and YZ sections are shown in the upper and lower panels, respectively (white bars indicate 10 lm).
predominantly coincident with 78 kD GRP (not shown). Ser-256 is a site for cAMP-dependent phosphorylation that regulates the exocytosis of AQP-2 [13, 14]. S256A, the mutant of Ser-256, was apparently localized in both the cytoplasm and apical region regardless of folkloric treatment (Fig. 3G and H), as well as partially colocalized with both TGN 38 (Fig. 3I) and VAMP2 (Fig. 3J). We checked the synthesis of the AQP-2 proteins in MDCK cells by immunoblot analysis using an antibody against HA tag. The apparent molecular size of the wildtype AQP-2 was 29 kD, and those of the mutant AQP-2 proteins were all close to the expected sizes (Fig. 4A). An extra band was identified at a position approximately 3 kD higher than the original band in four AQP-2 mutants lacking the last ten amino acid residues (262-271del, 252271del, 242-271del, and 232-271del), the 256-268del, and the 262-268del. The extra band was digested after the incubation with endoglycosidase H (Fig. 4B). Endoglycosidase H–sensitive glycoproteins are known to localize
in the endoplasmic reticulum [22]. In contrast, an additional band was not visible in four AQP-2s possessing the final 16 amino acid tail (wild-type AQP-2, 226-240del, 241-255del, and S256A) and the 269-271del. To evaluate the protein expression quantitatively, we calculated the band intensities using NIH images. The band intensities were 100 ± 31 (means ± SE in arbitrary units) in the wild-type AQP-2 (N = 4), 117 ± 25 in the 226-240del (N = 4), 128 ± 34 in the 241-255del (N = 4), 91 ± 28 in the 269-271del (N = 4), and 82 ± 23 in the S256A (N = 4). These values were not significantly different, suggesting equivalent protein expression levels. In our last experiment, we performed biotinylation assay to examine the cell surface expression of AQP. Expression at the apical membrane was detected in the wildtype AQP-2, 226-240del, 241-255del, 269-271del, S256A (Fig. 5A), wild-type AQP-2, AQP-4–AQP-2, and 241252del (Fig. 5B). Basolateral expression was observed only in the wild-type AQP4 (Fig. 5A). In contrast, the
Kuwahara et al: Apical trafficking of AQP-2 and C-terminus
2005
C
W T
on tro l AQ P 26 2 227 1d el 25 227 1d el 24 227 23 1de 2l 27 1d el 22 624 24 0d e 125 l 5d 26 el 927 25 1de l 626 8d el S2 56 26 A 226 8d el
A
B Endo
(−)
(+)
262-271 del
(−)
(+)
252-271 del
(−)
(+)
256-268 del
262-271del, 252-271del, and 256-268del proteins showed no cell surface expression whatsoever (Fig. 5A). These results were consistent with the protein localization shown in Figures 1 to 3. In Figure 5A series, the band intensities of the wild-type AQP-2 (100 ± 27) (means ± SE in arbitrary units) (N = 4) and 226-240del (85 ± 21) (N = 4) were not significantly different (Fig. 5C). In contrast, the apical expression was significantly decreased in the 241255del (28 ± 11) (N = 4) (P < 0.05), 269-271del (38 ± 12) (N = 4) (P < 0.05), and S256A (13 ± 6) (N = 4) (P < 0.05). In Figure 5B series, AQP-4–AQP-2 (35 ± 14) (N = 4) (P < 0.05) was significantly lower than the wild-type AQP-2 (100 ± 22) (N = 4) and 241-252del (91 ± 17) (N = 4). We quantitatively determined the effect of forskolin on the apical surface delivery of AQP proteins in the last protocol (Fig. 5D). Forskolin significantly increased apical expression of the wild-type AQP-2 (100 ± 15 to 208 ± 21) (N = 6) (P < 0.01), 241-252del (87 ± 14, to 193 ± 20) (N = 6) (P < 0.01), 241-255del (39 ± 6 to 61 ± 8) (N = 6) (P < 0.05), and AQP-4–AQP-2 (48 ± 7 to 76 ± 9) (N = 6) (P < 0.05) (Fig. 5E). In contrast, the effect of forskolin was absent in the 269-271AAA (65 ± 11 to 71 ± 8) (N = 6). DISCUSSION Previous reports suggested that the phosphorylation of Ser-256 is a key signal for the plasma membrane targeting of AQP-2 [13–16], and the S256A proteins were distributed at the intracellular vesicles [15] or Golgi complex [16], but basically absent in the apical membrane. Similarly, our findings identified the partial expression of S256A at the Golgi complex and intracellular vesicles (Fig. 3). The apical expression of biotin-labeled S256A was indeed identified, but the level of expression reached only 13% of that of the wild-type AQP-2 (Fig. 5). A low level of S256A expression at the apical surface is consistent with recent evidence that another mechanism
Fig. 4. Immunoblot of membrane fractions of the Madin-Darby canine kidney (MDCK) cells transfected with the wildtype aquaporin-2 (AQP-2) and C-terminal AQP-2 mutants detected by an antibody against HA peptide. (A) Lane 1, control (mock-transfected cells); lane 2, the wild-type AQP-2; lane 3, 262-271del; lane 4, 252-271del; lane 5, 242-271del; lane 6, 232-271del; lane 7, 226-240del; lane 8, 241-255del; lane 9, 269271del; lane 10, 256-268del; lane 11, S261A; and lane 12, 262-268del. Ten micrograms of total membrane protein was loaded in each lane. (B) Samples of the 262-271del, 252-271del, and 256-268del before (−) and after (+) incubation with endoglycosidase H (Endo). HA, hemagglutinin.
independent of Ser-256 phosphorylation takes part in AQP-2 apical trafficking. Studies in transfected LLCPK 1 cells have suggested that AQP-2 is constitutively recycled between the plasma membrane and intracellular vesicles in a vasopressin-independent manner [23]. This phosphorylation-independent apical trafficking of AQP-2 has also been suggested by others [24–26]. We performed cell surface biotinylation assay to examine the effect of forskolin quantitatively. The apical expression of the wild-type AQP-2 was increased by 108% in response to forskolin (Fig. 5). The basal expression level before forskolin treatment and the magnitude of stimulation by forskolin were similar in the wild-type AQP-2 and 241252del, a mutant possessing the consensus sequence for cAMP-dependent protein kinase. On the other hand, the basal expression of the 241-255del and AQP-4–AQP-2 were significantly lower than that of the wild-type AQP2, reaching 39% and 48% of the latter level, respectively (Fig. 5). The lack of the consensus sequence for cAMPdependent protein kinase (residues 253-256) might be responsible for the decrease in apical trafficking of these proteins. Forskolin significantly increased apical delivery of the 241-255del and AQP-4–AQP-2 by 56% and 58%, respectively, suggesting the presence of the site for cAMP-dependent phosphorylation in these proteins (Fig. 5). However, the degree of forskolin-induced stimulation was lower than that of the wild-type AQP-2 in the 241-255del and AQP-4–AQP-2. We speculate from these observations that the serine residue corresponding to S256 in the wild-type AQP-2 is still cAMP-dependent phosphorylation site and important for apical trafficking in these proteins. Our results on the AQP-4–AQP-2 chimera revealed an important role of residues 256-271 in the apical expression of AQP-2 (Fig. 2). The basolateral trafficking in AQP-4 is determined by a tyrosine-based motif (residues 277-280) and dileucine-based motif (residues 288-294) at the C-terminal sequence [27]. A C-terminal deletion
2006
Kuwahara et al: Apical trafficking of AQP-2 and C-terminus
A
B
AP
C on tro W l T AQ AQ P2 P4 -A 24 QP 2 125 2d el
C
W T
on tro l AQ P2 W TA 26 QP 4 227 1d 25 el 227 22 1d el 624 0d 24 el 125 5d 26 el 927 1d 25 el 626 8d el S2 56 A
BL
C
100
50
*
*
*
D FSK (−)
(+)
(−)
(+)
(+)
(−)
(+)
el 2d 25 1-
24
AQ P4
S2
27 926
(−)
-A Q P2
56
A
el 1d
el 5d
24
1-
25
0d 24
22
6-
W T
AQ P2
el
*
(−)
(+)
(−)
(+)
26
9-
27
1A
AQ P4 -A Q
AA
P2
el 5d 25 24
124
W
1-
25
2d
el
AQ P2 T
C
on
tro
l
AP
E FSK (−)
(+)
(−)
(+)
(−)
(+)
(−)
(+)
(−)
(+)
** **
1A
AA
*
27
AQ P4 -A Q P2
el 24
1-
25
5d
el 2d 25 124
W
T
AQ P2
*
9-
100
26
200
Fig. 5. Biotinylation of cell surface proteins. Confluent monolayers of transfected Madin-Darby canine kidney (MDCK) cells were incubated with Sulfo-NHS-LC-Biotin and solubilized. Biotin-labeled proteins were separated from unlabeled proteins by adding streptavidin-agarose. Aquaporin (AQP) proteins present in the pellet were detected by immunoblot using an antibody against HA peptide. Biotin was added either to the apical (AP) or basolateral (BL) compartment of the filter chamber. (A) Lane 1, control (mock-transfected cells); lane 2, wild-type AQP-2; lane 3, wild-type
Kuwahara et al: Apical trafficking of AQP-2 and C-terminus
mutant of AQP-4 in this study (285-323del) was localized at the basolateral membrane, probably because it still possessed the tyrosine-based motif. Unexpectedly, however, the addition of the AQP-2 C-terminal sequence (residues 256-271) to 285-323del AQP-4 resulted in apical expression of this chimera. It thus appeared that the sequence at residues 256-271 in AQP-2 was sufficient for apical trafficking and dominant over the basolateral signal of the tyrosine-based motif in AQP-4. In contrast to our findings, a chimera protein of the transmembrane region of placental alkaline phosphatase with the AQP-2 C-terminal tail (residues 220-271) failed to express at the apical membrane in an earlier study, suggesting that the C-terminal sequence was not sufficient for apical expression [28]. Notwithstanding, this earlier chimeric protein study did not rule out nonspecific processing defect between the endoplasmic reticulum and cell surface. The chimera protein in our study was so carefully designed that the corresponding sites of the C-terminal tail were exchanged between AQP-2 and AQP-4. A recent study examined the role of AQP-2 C-terminus by generating chimeras of AQP-1 and AQP-2 [29]. AQP-1 with AQP-2 residues 220-271 showed apical localization, whereas AQP-1s with AQP-2 residues distal to 230 were distributed at apical and basolateral membranes. This implied that amino acids 220-229 are important for apical sorting in AQP-2. While the present study found that residues 226-229 were probably not responsible for apical expression (226-240del) (Fig. 1), it did not examine the significance of residues 220-225. AQP-2 possesses a potential class I PDZ domain binding motif (X-S/T-X-) at the C-terminus (G-T-K-A) (268-271). PDZ domains are ubiquitous protein-protein interacting modules that recognize specific sequence motifs of transmembrane receptors, channels, and other PDZ-containing proteins. Indeed, recent evidence suggests that PDZ-interacting motifs mediate apical sorting of numerous channels and transporters, including cystic fibrosis transmembrane conductance regulator (CFTR), Na/H exchanger regulatory factors (NHERFs), c-aminobutyric acid transporter 3 (GAT-3), multidrugresistance-associated protein 2 (MRP2), and Na/Pi cotransporter 1 and 2 (Npt1 and Npt2) [17, 18]. For example, the last three amino acids at the C-terminus of the CFTR (T-R-L) comprise a class I PDZ-interacting motif. The mistargeting of a mutant lacking this domain (S1455X) to the basolateral membrane was found to lead
2007
the pathogenesis of cystic fibrosis [30]. Interaction of NHERF and ezrin-radixin-moesin-binding phosphoprotein 50 (EBP50), another PDZ protein, appears to regulate NHE3 through the protein kinase A pathway and b-adrenergic receptor coupled pathway [31]. In the present study, the 269-271AAA, a deletion mutant of a putative AQP-2 PDZ ligand, was apparently distributed in the subapical area irrespective of incubation with forskolin (Fig. 3). Biotinylation analysis showed a decrease in the confirmable apical expression of the 269271AAA (34% of that of the wild-type AQP-2) (Fig. 5). The 269-271AAA was mostly colocalized with intracellular vesicles, with some additional staining in the Golgi complex, late endosome/lysosome, and endoplasmic reticulum (Fig. 3). The signal-induced proliferationassociated gene-1 (SPA-1) is a guanine triphosphatase (GTPase)-activating protein (GAP) for Rap1 [32]. Our group recently identified SPA-1 as a PDZ protein that directly binds to the potential PDZ-interacting motif of AQP-2 (residues 268-271) [33]. We showed that SPA-1 was colocalized with AQP-2 in the collecting duct and that the GAP activity of SPA-1 was at least partially necessary for the apical sorting of AQP-2. Now that the Rap1 GTPases may be critical regulators of vesicle trafficking [34, 35], we speculated that the SPA-1 binding to AQP-2 reduces the levels of Rap1GTP that trigger F-actin disassembly and thereby leads to an activation of AQP-2 exocytosis from the intracellular vesicles to the apical membrane [33]. The retention of most of the AQP-2– containing vesicles in the subapical compartment in the present study was presumably due to the deletion of the putative PDZ-interacting motif. The apical localization of AQP-2 observed in this study was consistent with previous transfection studies of AQP2 protein with MDCK cells [36, 37]. In contrast to MDCK cells, AQP-2 transfected into LLC-PK 1 cells was translocated to the basolateral membrane upon vasopressin stimulation [38]. It has been reported that LLC-PK 1 cells exhibit mislocalization to the apical membrane in several proteins such as the low-density lipoprotein receptor (LDLR), transferrin receptor (TfR), and H-K-APTase beta subunit [39, 40]. The misrouting of LDLR and TfR was due to the lack of the l1B subunit of AP-1B complex because transfection of cDNA l1B into LLC-PK 1 cells restored the basolateral distribution of these proteins [39]. However, introduction of l1B failed to correct the basolateral localization of AQP-2 and the apical localization
AQP-4; lane 4, 262-271del; lane 5, 252-271del; lane 6, 226-240del; lane 7, 241-255del; lane 8, 269-271del; lane 9, 256-268del; and lane 10, S256A. (B) Lane 1, control; lane 2, wild-type AQP-2; lane 3, AQP-4–AQP-2; lane 4, 241-252del. (C) Summary of apical surface expression. ∗ P < 0.05 compared with wild-type AQP-2. (D) Effect of forskolin (FSK) on apical expression of AQP proteins. Cell surface biotinylation was assessed before (−) and after (+) the incubation with forskolin. Lanes 1 and 2, control; lanes 3 and 4, wild-type AQP-2; lanes 5 and 6, 241-252del; lanes 7 and 8, 241-255del; lanes 9 and 10, AQP-4–AQP-2; lanes 11and 12, 269-271AAA. (E) Summary of forskolin action on apical surface expression. ∗ P < 0.05; ∗∗ P < 0.01 compared with before pretreatment with forskolin. HA, hemagglutinin.
2008
Kuwahara et al: Apical trafficking of AQP-2 and C-terminus
of H-K-APTase beta subunit in LLC-PK 1 cells [7, 40]. These findings suggested that the LLC-PK 1 cells contain another defect in sorting machinery besides the lack of the l1B subunit. Recent evidence suggests that AQP-2 is expressed at the basolateral membrane of connecting tubule (CNT) cells and inner medullary collecting duct (IMCD) principal cells [6, 41–43]. Acute administration of oxytocin redistributed cytoplasmic AQP-2 proteins to the apical as well as the basolateral membrane in CNT cells and IMCD principal cells of Sprague-Dawley rats [42]. Moreover, 2-hour treatment with desamino-Cys1 , D-Arg8 vasopressin (dDAVP) markedly increased apical expression of AQP-2 in CNT cells and IMCD principal cells of Brattleboro rats [43]. These results suggested that the stimulation of cAMP-dependent protein kinase accelerates apical delivery of AQP-2 even in these cells. The mechanisms underlying the basolateral targeting of AQP-2 is unknown at present, although oxytocin and hypertonicity were suggested to cause basolateral redistribution of AQP-2 [42, 44]. Sorting mechanisms in CNT cells and IMCD principal cells might be different from those in cortical and outer medullary collecting duct principal cells. ACKNOWLEDGMENTS We thank Dr. E. Kominami for providing us an antibody against Lamp1. This work was supported by a Grant-in-Aid for Scientific Research and Grant-in-Aid for Scientific Research on Priority Areas from Japan Society for the Promotion of Science (JSPS). Reprint requests to Michio Kuwahara, M.D., Department of Nephrology, Graduate School, Tokyo Medical and Dental University, Tokyo 1138519, Japan. E-mail: [email protected]
REFERENCES 1. AGRE P, KING LS, YASUI M, et al: Aquaporin water channels—From atomic structure to clinical medicine. J Physiol 542:3–16, 2002 2. ISHIBASHI K, KUWAHARA M, SASAKI S: Molecular biology of aquaporins. Rev Physiol Biochem Pharmacol 141:1–32, 2000 3. HATAKEYAMA S, YOSHIDA Y, TANI T, et al: Cloning of a new aquaporin (AQP10) abundantly expressed in duodenum and jejunum. Biochem Biophys Res Commun 287:814–819, 2001 4. ISHIBASHI K, MORINAGA T, KUWAHARA M, et al: Cloning and identification of a new member of water channel (AQP10) as an aquaglyceroporin. Biochim Biophys Acta 1576:335–340, 2002 5. FUSHIMI K, UCHIDA S, HARA Y, et al: Cloning and expression of apical membrane water channel of rat kidney collecting tubule. Nature 361:549–552, 1993 6. NIELSEN S, DIGIOVANNI SR, CHRISTENSEN EI, et al: Cellular and subcellular immunolocalization of vasopressin-regulated water channel in rat kidney. Proc Natl Acad Sci USA 90:11663–11667, 1993 7. BROWN D: The ins and outs of aquaporin-2 trafficking. Am J Physiol 284:F893–F901, 2003 8. KUWAHARA M, IWAI K, OOEDA T, et al: Three families with autosomal dominant nephrogenic diabetes insipidus caused by aquaporin-2 mutations in the C-terminus. Am J Hum Genet 69:738–748, 2001 9. ASAI T, KUWAHARA M, KURIHARA H, et al: Pathogenesis of nephrogenic diabetes insipidus by aquaporin-2 C-terminus mutations. Kidney Int 64:2–10, 2003 10. MARR N, BICHET DG, LONERGAN M, et al: Heteroligomerization of an aquaporin-2 mutant with wild-type aquaporin-2 and their mis-
11. 12.
13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.
24.
25. 26.
27.
28. 29. 30. 31. 32.
33.
routing to late endosomes/lysozomes explains dominant nephrogenic diabetes insipidus. Hum Mol Genet 11:779–789, 2002 KAMSTEEG EJ, BICHET DG, KONINGS IB, et al: Reversed polarized delivery of an aquaporin-2 mutant causes dominant nephrogenic diabetes insipidus. J Cell Biol 163:1099–1109, 2003 KUWAHARA M, FUSHIMI K, TERADA Y, et al: cAMP-dependent phosphorylation stimulates water permeability of aquaporin-collecting duct water channel protein expressed in Xenopus oocytes. J Biol Chem 270:10384–10387, 1995 FUSHIMI K, SASAKI S, MARUMO F: Phosphorylation of serine 256 is required for cAMP-dependent regulatory exocytosis of the aquaporin-2 water channel. J Biol Chem 272:14800–14804, 1997 KATSURA T, GUSTAFSON CE, AUSIELLO DA, et al: Protein kinase A phosphorylation is involved in regulated exocytosis of aquaporin-2 in transfected LLC-PK 1 cells. Am J Physiol 272:F817–F822, 1997 VAN BALKOM BW, SAVELKOUL PJ, MARKOVICH D, et al: The role of putative phosphorylation sites in the targeting and shuttling of the aquaporin-2 water channel. J Biol Chem 277:41473–41739, 2002 PROCINO G, CARMOSINO M, MARIN O, et al: Ser-256 phosphorylation dynamics of aquaporin 2 during maturation from the ER to the vesicular compartment in renal cells. FASEB J 17:1886–1888, 2003 HARRIS BZ, LIM WA: Mechanism and role of PDZ domains in signaling complex assembly. J Cell Sci 114:3219–3231, 2001 JELEN F, OLEKSY A, SMIETANA K, et al: PDZ domains—Common players in the cell signaling. Acta Biochim Pol 50:985–1017, 2003 LEFEBVRE B, FORMSTECHER P, LEFEBVRE P: Improvement of the gene splicing overlap (SOE) method. Biotechniques 19:186–188, 1995 TAMARAPPOO BK, VERKMAN AS: Defective aquaporin-2 trafficking in nephrogenic diabetes insipidus and correction by chemical chaperones. J Clin Invest 101:2257–2267, 1998 YANG B, MA T, XU Z, et al: cDNA and genomic cloning of mouse aquaporin-2: functional analysis of an orthologous mutant causing nephrogenic diabetes insipidus. Genomics 57:79–83, 1999 HALBAN PA, IRMINGER JC: Sorting and processing of secretory proteins. Biochem J 299:1–18, 1994 GUSTAFSON CE, KATSURA T, MCKEE M, et al: Recycling of AQP2 occurs through a temperature- and bafilomycin-sensitive trans-Golgiassociated compartment. Am J Physiol Renal Physiol 278:F317– F326, 2000 VALENTI G, PROCINO G, CARMOSINO M, et al: The phosphatase inhibitor okadaic acid induces AQP2 translocation independently from AQP2 phosphorylation in renal collecting duct cells. J Cell Sci 113:1985–1992, 2000 SUN TX, VAN HOEK A, HUANG Y, et al: Aquaporin-2 localization in clathrin-coated pits: Inhibition of endocytosis by dominant-negative dynamin. Am J Physiol Renal Physiol 282:F998–F1011, 2002 LU H, SUN TX, BOULEY R, et al: Inhibition of endocytosis causes phosphorylation (S256)-independent plasma membrane accumulation of AQP2. Am J Physiol Renal Physiol 286:F233–F243, 2004 MADRID R, LE MAOUT S, BARRAULT M-B, et al: Polarized trafficking and surface expression of the AQP4 water channel are coordinated by serial and regulated interactions with different clathrin-adaptor complexes. EMBO J 20:7008–7021, 2001 DEEN PMT, VAN BALKOM BWM, SAVELKOUL PJM, et al: Aquaporin2: COOH terminus is necessary but not sufficient for routing to the apical membrane. Am J Physiol 282:F330–F340, 2002 VAN BALKOM BWM, GRAAT MPJ, VAN RAAK M, et al: Role of cytoplasmic termini in sorting and shuttling of the aquaporin-2 water channel. Am J Physiol 286:C372–C379, 2004 MOYER BD, DENTON J, KARLSON KH, et al: A PDZ-interacting domain in CFTR is an apical membrane polarization signal. J Clin Invest 104:1353–1361, 1999 LAMPRECHT G, WEINMAN EJ, YUN CH: The role of NHERF and E3KARP in the cAMP-mediated inhibition of NHE3. J Biol Chem 273:29972–29978, 1998 HATTORI M, TSUKAMOTO N, NUR-E-KAMAL MS, et al: Molecular cloning of a novel mitogen-inducible nuclear protein with a Ran GTPase-activating domain that affects cell cycle progression. Mol Cell Biol 15:552–560, 1995 NODA T, HORIKAWA S, FURUKAWA T, et al: Aquaporin-2 trafficking is regulated by PDZ-domain containing protein SPA-1. FEBS Lett 18:139–145, 2004
Kuwahara et al: Apical trafficking of AQP-2 and C-terminus
34. MOSKALENKO S, HENRY DO, ROSSE C, et al: The exocyst is a Ral effector complex. Nat Cell Biol 4:66–72, 2002 35. ALTSCHULER Y, HODSON, MILGRAM SL: The apical compartment: Trafficking pathways, regulators and scaffolding proteins. Curr Opin Cell Biol 15:423–429, 2003 36. VALENTI G, FRIGERI A, RONCO PM, et al: Expression and functional analysis of water channels in a stably AQP2-transfected human collecting duct cell line. J Biol Chem 271:24365–24370, 1996 37. DEEN PM, RIJSS JP, MULDERS SM, et al: Aquaporin-2 transfection of Madin-Darby canine kidney cells reconstitutes vasopressinregulated transcellular osmotic water transport. J Am Soc Nephrol 8:1493–1501, 1997 38. KATSURA T, VERBAVATZ JM, FARINAS J, et al: Constitutive and regulated membrane expression of aquaporin 1 and aquaporin 2 water channels in stably transfected LLC-PK 1 epithelial cells. Proc Natl Acad Sci USA 92:7212–7216, 1995 39. FOLSCH H, OHNO H, BONIFACINO JS, et al: A novel clathrin adaptor
40. 41. 42. 43. 44.
2009
complex mediates basolateral targeting in polarized epithelial cells. Cell 99:189–198, 1999 DUFFIELD A, FOLSCH H, MELLMAN I, et al: Sorting of H,K-ATPase beta-subunit in MDCK and LLC-PK cells is independent of mu 1B adaptin expression. Traffic 5:449–461, 2004 COLEMAN RA, WU DC, LIU J, et al: Expression of aquaporins in the renal connecting tubule. Am J Physiol Renal Physiol 279:F874– F883, 2000 JEON US, JOO KW, NA KY, et al: Oxytocin induces apical and basolateral redistribution of aquaporin-2 in rat kidney. Nephron Exp Neohrol 93:e36–e45, 2003 CHRISTENSEN BM, WANG W, FROKIAER J, et al: Axial heterogeneity in basolateral AQP2 localization in rat kidney: effect of vasopressin. Am J Physiol Renal Physiol 284:F701–F717, 2003 VAN BALKOM BW, VAN RAAK M, BRETON S, et al: Hypertonicity is involved in redirecting the aquaporin-2 water channel into the basolateral, instead of the apical, plasma membrane of renal epithelial cells. J Biol Chem 278:1101–1107, 2003
The C-terminal tail of aquaporin-2 determines apical trafficking MICHIO KUWAHARA, TOMOKI ASAI, YOSHIO TERADA, and SEI SASAKI Department of Nephrology, Tokyo Medical and Dental University Graduate School, Tokyo, Japan
The C-terminal tail of aquaporin-2 determines apical trafficking. Background. Aquaporin-2 (AQP-2) proteins are mainly expressed at the apical region of the collecting duct cells. We previously reported three different mutations in the C-terminus of AQP-2 that all-cause autosomal-dominant nephrogenic diabetes insipidus. When one of these mutant AQP-2s was expressed in Madin-Darby canine kidney (MDCK) cells, it was mistargeted to the basolateral membrane, suggesting a critical role of the C-terminal tail in the apical trafficking of AQP-2. Methods. Portions of the AQP-2 C-terminal tail (residues 226271) were mutated by the polymerase chain reaction (PCR) technique and inserted into the pcDNA3.1 vector. Constructs were transfected into MDCK cells to examine the localization of mutated AQP-2 proteins by immunofluorescence microscopy. Cell surface expression was detected by biotinylation assay. Results. The wild-type AQP-2 was localized at the apical membrane, whereas mutants lacking residues 262-271 (the last 10 amino acids) were predominantly distributed in the endoplasmic reticulum. Deletion mutants of the initial (226-240del) and middle (241-252del) portions of the C-terminal tail were identified at the apical membrane, suggesting that residues 226252 have no involvement in apical targeting. An AQP-4–AQP-2 chimera in which a portion of the AQP-4 C-terminal tail was replaced by the corresponding site in AQP-2 (residues 256-271) was found at the apical membrane. The sequence of the last 4 amino acids of AQP-2 (G-T-K-A) corresponds to a PDZinteracting motif. Our investigations identified a mutant of this portion mostly localized to the subapical region. Further, apical expression was found to be significantly decreased in mutants lacking a consensus sequence for cyclic adenosine monophosphate (cAMP)–dependent phosphorylation (residues 253-256). Conclusion. The sequence at 256-271 is sufficient for apical trafficking in AQP-2. The putative PDZ-interacting motif (GT-K-A, residues 268-271) plays a key role in apical membrane expression. In addition, cAMP-dependent phosphorylation was found to be critical for apical targeting.
Water movement across the cell membrane is an essential process in the maintenance of cellular homeostasis. Water molecules move through the aquaporin (AQP) Key words: AQP-2, C-terminal tail, PDZ-interacting motif, signalinduced proliferation-associated gene-1. Received for publication October 5, 2004 and in revised form March 15, 2005, and May 13, 2005 Accepted for publication June 3, 2005 C
2005 by the International Society of Nephrology
water channels, a family of selective pores that belong to the major intrinsic protein (MIP) superfamily [1, 2]. Hydropathy analysis of the MIP superfamily predicts six transmembrane regions with N- and C-termini localized in the cytoplasm. AQPs are found virtually in all species, including more than ten kinds so far identified in mammals [1, 3, 4]. Aquaporin-2 (AQP-2) is a vasopressinregulated water channel that plays a central role for the urinary concentration from its sole site of expression in the principal cells of the connecting tubule and collecting duct cells [5, 6]. The binding of vasopressin to the V2 receptor on the basolateral membrane is generally thought to activate cyclic adenosine monophosphate (cAMP)– dependent protein kinase, a process that encourages cytoplasmic AQP-2–containing vesicles from the subapical region to infiltrate the apical membrane, and thereby increase its water permeability [7]. After the effect of vasopressin subsides, AQP-2 channels are retrieved from the apical membrane by endocytosis and the water permeability returns to a low level. In our previous report, we identified three mutations of the AQP2 gene confirmed to cause the autosomaldominant form of nephrogenic diabetes insipidus (NDI) [8]. All three of these mutations were localized to a single allele in exon 4 of the AQP2 gene at the C-terminal tail (721delG, 763-772del, and 812-818del). The wild-type AQP-2 is a 271 amino acid protein, whereas these mutant genes were predicted to encode 330-333 amino acid proteins due to a frameshift mutation of the stop codon. Our Xenopus oocyte expression study of these mutants revealed an impairment in the trafficking to the plasma membrane [8]. When one of these mutants (763-772del) was expressed in polarized epithelia of Madin-Darby canine kidney (MDCK) cells, the mutant AQP-2 was mistargeted to the basolateral membrane, suggesting that the AQP-2 C-terminus was responsible for apical targeting [9]. Misrouting of C-terminal mutant AQP-2 has been also reported in other cases of autosomal-dominant NDI [10, 11]. These observations suggest that the C-terminal tail of AQP-2 plays a critical role in apical trafficking. The C-terminal tail of AQP-2 is composed of 46 amino acid residues [5]. We hypothesized that a portions of the residues in this sequence interact with proteins responsible for apical targeting. The AQP-2 C-terminal tail 1999
2000
Kuwahara et al: Apical trafficking of AQP-2 and C-terminus
Table 1. C-terminal amino acid sequence of the wild-type and mutant AQP2s AA 230 WT-AQP2 262-271del 252-271del 242-271del 232-271del 226-240del 241-255del 241-252del 269-271del 256-268del 262-268del S256A
240
250
260
270
PAKSLSERLAVLKGLEPDTDWEEREVRRRQSVELHSPQSLPRGTKA PAKSLSERLAVLKGLEPDTDWEEREVRRRQSVELHS PAKSLSERLAVLKGLEPDTDWEEREV PAKSLSERLAVLKGLE PAKSLS –——————————EPDTDWEEREVRRRQSVELHSPQSLPRGTKA PAKSLSERLAVLKGL—–——————————SVELHSPQSLPRGTKA PAKSLSERLAVLKGL-–————————–RRQSVELHSPQSLPRGTKA PAKSLSERLAVLKGLEPDTDWEEREVRRRQSVELHSPQSLPRG PAKSLSERLAVLKGLEPDTDWEEREVRRRQ—————————TKA PAKSLSERLAVLKGLEPDTDWEEREVRRRQSVELHS—————TKA PAKSLSERLAVLKGLEPDTDWEEREVRRRQAVELHSPQSLPRGTKA
C-terminal tail of the wild-type (WT) AQP2 is composed of 46 amino acid (AA) residues, starting from P231 to A271.
contains several possible domains/motifs associated with protein trafficking. Residues 253-256 (R-R-G-S) make up a consensus sequence for cAMP-dependent protein kinase (R-R-X-[S/T]) [5]. Indeed, cAMP-dependent phosphorylation was observed in wild-type AQP-2, but not in S256A AQP-2 [12, 13]. Further, the plasma membrane expression of AQP-2 was found to be inhibited in S256A AQP-2 [13–16]. The final four amino acid sequence of AQP-2 (G-T-K-A) corresponds to a class I PDZ domain binding motif (X-[S/T]-X-). Similarly, residues 231-234 (S-E-R-L) comprise a class III PDZ-interacting motif (-X-[D/E/K/R]-X-). Given that PDZ proteins are known to tether membrane proteins [17, 18], these domains may play a role in the apical distribution of AQP-2. In the present study we transfected several AQP-2 mutants in MDCK cells and identified their intracellular localization to determine the role of the AQP-2 C-terminus in apical targeting. METHODS Preparation of cDNA and construction of mutants Wild-type cDNA of human AQP-2 with a Kozak consensus translation initiation site just upstream of the initial codon (ATG) was inserted into the pcDNA3.1 vector (Invitrogen, CH Groningen, The Netherlands) between the HindIII site and XbaI site [9]. To generate hemagglutinin (HA) epitope-tagged AQP-2, the nucleotide sequence encoding the HA tag derived from human hemagglutinin protein (YPYDVPDYA) was introduced just after the initiation codon of AQP-2 cDNA. A method of polymerase chain reaction (PCR)–based mutagenesis was used to construct a series of C-terminal mutants of HA-tagged AQP-2 (Table 1), as described previously [9, 19]. The PCR products were digested with HindIII and XbaI, and inserted into the pcDNA3.1 vector (Invitrogen). All of the mutations were confirmed by a fluorescence DNA sequencer. In a separate experiment, HA-tagged rat AQP-4 was cloned into the pcDNA3.1 vector. A part of the AQP-4 C-terminal
tail (residues 285-323) (SQVETEDLILKPGVVHVIDIDRGDEKKGKDSSGEVLSSV) was replaced by the corresponding site (residues 256-271) (SVELHSPQSLPRGTKA) from AQP-2 (AQP-4–AQP-2). A Cterminal deletion mutant of AQP-4 (285-323del) was also generated. Cell culture and stable transfection MDCK cells (Japanese Collection of Research Biologicals #9029, Osaka, Japan) were cultured and transfected stably with wild-type or mutant AQP-2 and AQP-4 using previously described methods [9]. Cells were grown at 37◦ C under a humidified atmosphere of 5% CO 2 . Ten micrograms of the expression constructs of AQP2 were transfected into MDCK cells grown subconfluently on dishes using the calcium-phosphate precipitation technique. After 24 hours, the cells transfected with the wild-type AQP-2 and mutant AQP-2 were trypsinized, split among six dishes, and cultured in Dulbecco’s modified Eagle’s medium (DMEM) containing 700 lg/mL of G418 (Sigma Chemical Co., St. Louis, MO, USA). Ten to 14 days after the transfection, individual colonies were selected using cloning rings and expanded. The clones were isolated after the selection. Confocal immunofluorescence microscopy Transfected MDCK cells were grown on a permeable supported membrane (Transwell-Clear) (Corning, Cambridge, MA, USA). After 3-day culture in DMEM, confluent cells were incubated for 30 minutes with DMEM containing 10−5 mol/L forskolin (Sigma Chemical Co.) to induce apical expression of AQP-2. In some protocols, cells were incubated for 30 minutes with 0.5 mmol/L 8-(4-chlorophenylthio)-cAMP (Sigma Chemical Co.). The cells were then fixed with 2% paraformaldehyde in serum-free DMEM, washed with phosphate-buffered saline (PBS), permeabilized with 0.2% Triton X-100 in PBS, blocked with 1% bovine serum albumin (BSA) in PBS at 4◦ C overnight, and
Kuwahara et al: Apical trafficking of AQP-2 and C-terminus
stained [9]. Antibodies against the following proteins were used to double stain the proteins in the transfected cells: wild-type AQP-2 (rabbit or rat, an antibody against 15 C-terminal amino acids generated by Sawady Technology) (Tokyo, Japan) (diluted 1:400); HA peptide (rat) (1:200) (Roche Diagnostics, Penzberg, Germany); Na-K-adenosine triphophatase (ATPase), a basolateral membrane marker protein (mouse) (1:400) (Upstate Biotechnology, Lake Placid, NY, USA); 78 kD glucose-regulated protein (GRP 78), an endoplasmic reticulum marker protein (rabbit) (1:200) (Stressgen, Victoria, Canada); TGN 38, a trans-Golgi marker protein (mouse) (1:200) (BD Biosciences, San Jose, CA, USA); lysosome-associated membrane protein 1 (Lamp1), a late endosomal/lysosomal marker protein (rabbit) (1:200) (kindly provided by Dr. E. Kominami, Department of Biochemistry, School of Medicine, Juntendo University); and vesicle-associated membrane protein 2 (VAMP2), a intracellular vesicle marker protein (rabbit) (1:200) (Calbiochem, La Jolla, CA, USA). After 1-hour incubation with a mixture of two antibodies in PBS containing 1% BSA, the cells were washed with PBS, further incubated with an antibody against rabbit or rat IgG (1:200) (Alexa Fluor 488) (Molecular Probes, Eugene, OR, USA) and rat or rabbit IgG (1:200) (Alexa Fluor 568) (Molecular Probes) for 1 hour, washed again with PBS and mounted in a mounting medium (Prolong Antifade Kit) (Molecular Probes). In some protocols the cells were pretreated with 100 mmol/L of trimethylamineN-oxide or 2% dimethyl sulfoxide (DMSO) as a chemical chaperone for 4 hours just before the experiment. Immunofluorescent images were obtained with an LSM510 laser scanning confocal microscope system (Carl Zeiss, Jena, Germany) using a 63× water-immersion objective. Immunoblot analysis The immunoblot analysis was performed as described earlier [9]. Initially, the transfected cells were homogenized for 30 minutes in radioimmunoprecipitation assay (RIPA) buffer [50 mmol/L Tris-HCl, 150 mmol/L NaCl, 1.0% NP-40, 0.5% deoxycholate, 0.1% sodium dodecyl sulfate (SDS), pH 8.0] containing a protease inhibitor cocktail (Boehringer Manheim, Manheim, Germany). Next, the samples were centrifuged, denatured and separated by SDS-polyacrylamide gel electrophoresis (PAGE), and transferred to Immobilon-P filters (Millipore, Bedford, MA, USA) using a semidry system. Ten micrograms of total membrane protein was loaded in each lane. The filters were examined by autoradiography after blocking, and two 1-hour incubations with antibody against HA peptide and 125 I-protein A solution, respectively. Mock-transfected cells were treated similarly to provide control. In one protocol, the samples were incubated with endoglycosidase H (Boehringer
2001
Manheim) according to the manufacturer’s instructions. The band intensities were calculated using NIH images to compare the levels of protein expression quantitatively. Four independent sets of experiments were performed in each sample. Results were expressed as means ± SE in arbitrary units and statistical significance was evaluated by analysis of variance (ANOVA). Cell surface biotinylation Cells were seeded at a density of 1.5 × 105 cells/cm2 on a 12-well membrane support (1.0 cm2 ) (TranswellClear) (Corning). After a 3-day culture in DMEM, confluent cells were incubated for 30 minutes with 10−5 mol/L forskolin (Sigma Chemical Co.). Confluent monolayer cells were washed twice with ice-cold PBS containing 0.1 mmol/L CaCl 2 + 1 mmol/L MgCl 2 (PBSCM). All subsequent steps were essentially performed at 4◦ C. Sulfo-NHS-LC-Biotin (2 mg/mL) (Pierce, Rockford, IL, USA) in biotinylation buffer (2 mmol/L CaCl 2 , 150 mmol/L NaCl, 10 mmol/L triethanol amine, pH 8.0) was added to either the apical or basolateral compartment of the filter chamber. Compartments not receiving Sulfo-NHS-LC-Biotin were filled with an equivalent volume of biotinylation buffer alone. After 20 minutes of incubation with gentle agitation, the cells were washed with biotinylation buffer and the labeling procedure repeated. Unreacted Sulfo-NHS-LC-Biotin was quenched by adding quenching buffer (100 mmol/L glycine in PBS-CM). The cells were washed twice with ice-cold PBS-CM and the filters were cut from the plastic support and incubated for 30 minutes at 37◦ C with 1 mL of lysis buffer [150 mmol/L NaCl, 50 mmol/L TrisHCl at pH 7.5, 5 mmol/L ethylenediaminetetraacetic acid (EDTA), 1% NP-40, 0.1% SDS, 0.5% deoxycholate) containing a protease inhibitor cocktail. The cells were scraped from the filters, transferred to Eppendorf tubes, and centrifuged at 14,000g for 5 minutes. Next, 40 lL of streptavidin-agarose (Sigma Chemical Co.), prepared by two washes with high salt buffer (500 mmol/L NaCl, 5 mmol/L EDTA, 50 mmol/L TrisHCl, pH 7.5, 0.1% Triton-X) and two washes with lysis buffer was added to the supernatant. After a 12-hour incubation at 4◦ C with end-over-end rotation, the beads were centrifuged at 14,000g for 5 minutes and washed twice with high-salt buffer, twice with lysis buffer, and once with 10 mmol/L Tris-HCl, pH 7.5. After removing the supernatant, the beads were resuspended in Laemmli sample buffer and denatured for 30 minutes at 37◦ C. Subsequent procedures were essentially same as those used for the immunoblots. Four to six independent experiments were done and the band intensities obtained by NIH images were expressed as means ± SE in arbitrary units. Statistical significance was evaluated by ANOVA or nonpaired t test.
2002
Kuwahara et al: Apical trafficking of AQP-2 and C-terminus
Fig. 1. Immunocytochemical localization of the wild-type and mutant aquaporin-2 (AQP2) transfected in Madin-Darby canine kidney (MDCK) cells. (A to C) Double staining of the wild-type AQP-2 and Na-K-adenosine triphosphatase (ATPase). The wild-type AQP-2 was detected by an antibody against 15 C-terminal amino acids of AQP-2 (A) and the HA-tagged wild-type AQP-2 was detected by an antibody against HA peptide after (B) and before (C) incubation with forskolin. (D to H) Double staining of the 262-271del and Na-K-ATPase (D), 78 kD glucose-related protein (GRP 78) after (E) and before (F) forskolin treatment, trans-Golgi marker protein (TGN 38) (G), and lysosome-associated membrane protein 1 (Lamp1) (H). (I and J) Double staining of the 262-271del and Na-K-ATPase (I) and GRP 78 (J) after pretreatment with a chemical chaperone, trimethylamine-N-oxide. (K) Double staining of Na-K-ATPase and the 226240del. (L to N) Double staining of the 241255del and Na-K-ATPase (L) and TGN 38 after (M) and before (N) the incubation with forskolin. (O and P) Double staining of NaK-ATPase and the 241-252del after (O) and before (P) the forskolin treatment. Cells were viewed with a confocal microscope at 630×. XY and YZ sections are shown in the upper and lower panels, respectively (white bars indicate 10 lm). HA, hemagglutinin.
RESULTS Figure 1A shows the double staining of the wild-type AQP-2 transfected in MDCK cells and a basolateral membrane marker, Na-K-ATPase. An antibody against the AQP-2 C-terminal detected AQP-2 proteins at the apical membrane. Similarly, an antibody against HA identified the apical distribution of the HA-tagged wild-type AQP-2 (Fig. 1B), suggesting that the insertion of the HA tag had no inhibitory effects on the apical expression of the wild-type AQP-2. Based on this finding, we used the anti-HA antibody to identify HA-tagged AQP2 mutants in subsequent protocols. Before treatment with forskolin, the wild-type AQP-2 appeared to be distributed at the subapical region besides the apical membrane (Fig. 1C). Thus forskolin further redistributed the wild-type AQP-2 to the apical membrane. Similar findings were observed when cells were incubated with 8(4-chlorophenylthio)-cAMP (not shown). These results suggested that the stimulation of cAMP-dependent protein kinase enhanced apical AQP-2 trafficking. In contrast to the wild-type AQP-2, the 262-271del, an AQP-2
mutant lacking the last 10 amino acid residues, was localized in cytoplasm with a granular pattern (Fig. 1D). Double staining of the 262-271del and 78 kD glucoseregulated protein, an endoplasmic reticulum marker protein, revealed a close colocalization irrespective of forskolin treatment (Fig. 1E and F). On the other hand, the distributions of the 262-271del proteins were not coincident with that of TGN 38, a trans-Golgi marker protein (Fig. 1G), and slightly overlapped with that of Lamp1, a late endosomal/lysosomal marker protein (Fig. 1H). The intracellular localizations of the 252-271del, 242271del, 232-271del were similar to that of the 262-271del (not shown). On the basis of these results, we speculated that AQP-2 mutants lacking the C-terminal 262-271 residues were basically retained in the endoplasmic reticulum through the action of the endoplasmic reticulum quality-control mechanisms. Preincubation with a chemical chaperone, trimethylamine-N-oxide, redistributed the 262-271del proteins mainly at the subapical area (Fig. 1I). A small portion of the 262-271del proteins was colocalized with TGN 38 in this condition (Fig. 1J). A similar
2003
Kuwahara et al: Apical trafficking of AQP-2 and C-terminus
Table 2. Apparent localization of the wild-type and mutant aquaporins (AQPs) Construct
Localization of AQP protein after [F(+)] and before [F(−)] forskolin treatment
Wild-type AQP-2 262-271del 252-271del 242-271del 232-271del 226-240del 241-255del 241-252del 269-271AAA 269-271del 256-268del 262-268del S256A AQP4-AQP2
Intracellular F(+) apical, F(−) apical, subapical F(+), F(−) subapical, apical F(+) apical, F(−) apical, subapical F(+), F(−) apical, subapical F(+), F(−) apical, subapical Intracellular Intracellular F(+), F(−) intracellular, apical F(+) apical, F(−) apical, subapical Intracellular Intracellular F(+), F(−) intracellular, apical F(+) apical, F(−) apical, subapical
Number of cell lines
Number of cells
3 4 4 4 3 3 3 3 4 3 3 3 4 3
35 30 18 14 15 28 31 28 25 31 17 16 30 27
Abbreviations are: F(+), apical; F(−), apical, subapical F(+); F(−), intracellular.
redistribution of the 262-271del was observed after incubation with DMSO, another chemical chaperone (not shown), suggesting that chemical chaperones may be able to partially correct the retention of the mutant AQP-2 in the endoplasmic reticulum, as previously reported [20, 21]. In contrast to the truncation mutants, the 226-240del, a mutant lacking the initial one third of the C-terminal tail, was confined to the apical membrane area (Fig. 1K). It thus appeared that the consensus sequences of the class III PDZ domain (residues 231-234) play no part in the apical targeting of AQP-2. The 226-240del was distributed apical and subapical region before incubation with forskolin (not shown). Staining of the 241-255del, the deletion of the middle portion of C-terminus was identified at the subapical cytoplasm as well as the apical membrane (Fig. 1L). Some of the 241-255del proteins were colocalized with TGN 38 (Fig. 1M), suggesting that the 241-255del was distributed at both the Golgi complex and apical membrane. Localization of the 241-255del seemed to be similar before and after incubation with forskolin (Fig. 1M and N). In contrast, apical localization was observed in the 241-252del (Fig. 1O), a mutant in which a consensus sequence for cAMP-dependent protein kinase is preserved. Forskolin as well as 8-(4chlorophenylthio)-cAMP apparently stimulated the apical delivery of the 241-252del to some extent (Fig. 1O and P). These findings had two implications: first, that the sequence between 226 and 252 plays no part in apical sorting, and second, that cAMP-dependent phosphorylation contributes to the translocation of AQP-2 from the subapical region to the apical membrane. Table 2 summarizes the apparent localization, the number of generated cell lines, and the number of cells examined in each AQP-2 construct. Next, we decided to perform a chimera study to make up for the shortfall of meaningful results in the truncation studies due to the endoplamic reticulum retention. AQP-4 proteins are known to exist at the basolateral
membrane of the principal cells in the collecting duct. As expected, the wild-type AQP-4 transfected in MDCK cell was expressed at the basolateral membrane (Fig. 2A). Deletion of the last 39 amino acid residues of AQP4 (285-323del) did not change the basolateral expression (Fig. 2B). In contrast, AQP-4–AQP-2, a chimera in which the last 16 amino acids of AQP-2 (residues 256271) were substituted for the corresponding site in the AQP-4 C-terminal tail (residues 285-323), was localized at the apical membrane (Fig. 2C). These results suggested that the residues at 256-271play a critical role in the apical trafficking. Forskolin appeared to stimulate the apical trafficking of AQP-4–AQP-2 slightly (Fig. 2C and D). 8(4-chlorophenylthio)-cAMP had a similar effect on the AQP-4–AQP-2 (not shown). Our next endeavor was to examine the role of amino acids at 256-271 in further detail. The last three amino acids of AQP-2 were replaced by alanines (269-271AAA) due to the compatibility of the sequence of the last four amino acids of (G-T-K-A) with a PDZ-interacting domain. The 269-271AAA was apparently distributed at the apical side (Fig. 3A), but not confined to the apical membrane in the manner seen in the wild-type AQP-2 (Fig. 1A and B). The localization of the 269-271AAA was similar even before the incubation with forskolin (Fig. 3B), suggesting that forskolin had little effect on the distribution of the 269-271AAA proteins. Double staining of the 269-271AAA and intracellular organelle markers showed that the 269-271AAA was rarely colocalized with 78 kD GRP, TGN 38, and Lamp1 (Fig. 3C to E), but closely colocalized with VAMP2 (Fig. 3F). The distribution of the mutant lacking the last three amino acids of AQP-2 (269-271del) was similar to that of the 269-271AAA (not shown). These observations suggested that the putative PDZ-binding motif (residues 268-271) plays a critical role in trafficking from the intracellular vesicles to the apical membrane. The other mutants tested, namely, the 256-268del and 262-268del were
2004
Kuwahara et al: Apical trafficking of AQP-2 and C-terminus
Fig. 2. Immunocytochemical localization of the wild-type and mutant aquaporin-4 (AQP4) doubly stained with Na-K-adenosine triphosphatase (ATPase). (A) The wild-type AQP-4. (B) 285-323del. The AQP-4 mutant that lacks the last 39 amino acid residues. (C and D) AQP-4–AQP-2. The AQP-4 chimera (residues 1-284 in AQP-4) with the C-terminal tail of AQP-2 (residues 256-271 in AQP-2) after (C) and before (D) the incubation with forskolin. Cells were viewed with a confocal microscope at 630×. XY and YZ sections are shown in the upper and lower panels, respectively (white bars indicate 10 lm).
Fig. 3. Immunocytochemistry of aquaporin-2 (AQP-2) and the basolateral membrane or intracellular organelle. (A to F) Double staining of the 269-271del and Na-K-adenosine triphophatase (ATPase) after (A) and before (B) the forskolin treatment, 78 kD glucose-related protein (GRP 78) (C), trans-Golgi marker protein (TGN 38) (D), lysosome-associated membrane protein 1 (Lamp1) (E), and vesicle-associated membrane protein 2 (VAMP2) (F). (G to J) Double staining of the S256A and Na-K-ATPase after (G) and before (H) the incubation with forskolin, TGN 38 (I) and VAMP2 (J). Cells were viewed with a confocal microscope at 630×. XY and YZ sections are shown in the upper and lower panels, respectively (white bars indicate 10 lm).
predominantly coincident with 78 kD GRP (not shown). Ser-256 is a site for cAMP-dependent phosphorylation that regulates the exocytosis of AQP-2 [13, 14]. S256A, the mutant of Ser-256, was apparently localized in both the cytoplasm and apical region regardless of folkloric treatment (Fig. 3G and H), as well as partially colocalized with both TGN 38 (Fig. 3I) and VAMP2 (Fig. 3J). We checked the synthesis of the AQP-2 proteins in MDCK cells by immunoblot analysis using an antibody against HA tag. The apparent molecular size of the wildtype AQP-2 was 29 kD, and those of the mutant AQP-2 proteins were all close to the expected sizes (Fig. 4A). An extra band was identified at a position approximately 3 kD higher than the original band in four AQP-2 mutants lacking the last ten amino acid residues (262-271del, 252271del, 242-271del, and 232-271del), the 256-268del, and the 262-268del. The extra band was digested after the incubation with endoglycosidase H (Fig. 4B). Endoglycosidase H–sensitive glycoproteins are known to localize
in the endoplasmic reticulum [22]. In contrast, an additional band was not visible in four AQP-2s possessing the final 16 amino acid tail (wild-type AQP-2, 226-240del, 241-255del, and S256A) and the 269-271del. To evaluate the protein expression quantitatively, we calculated the band intensities using NIH images. The band intensities were 100 ± 31 (means ± SE in arbitrary units) in the wild-type AQP-2 (N = 4), 117 ± 25 in the 226-240del (N = 4), 128 ± 34 in the 241-255del (N = 4), 91 ± 28 in the 269-271del (N = 4), and 82 ± 23 in the S256A (N = 4). These values were not significantly different, suggesting equivalent protein expression levels. In our last experiment, we performed biotinylation assay to examine the cell surface expression of AQP. Expression at the apical membrane was detected in the wildtype AQP-2, 226-240del, 241-255del, 269-271del, S256A (Fig. 5A), wild-type AQP-2, AQP-4–AQP-2, and 241252del (Fig. 5B). Basolateral expression was observed only in the wild-type AQP4 (Fig. 5A). In contrast, the
Kuwahara et al: Apical trafficking of AQP-2 and C-terminus
2005
C
W T
on tro l AQ P 26 2 227 1d el 25 227 1d el 24 227 23 1de 2l 27 1d el 22 624 24 0d e 125 l 5d 26 el 927 25 1de l 626 8d el S2 56 26 A 226 8d el
A
B Endo
(−)
(+)
262-271 del
(−)
(+)
252-271 del
(−)
(+)
256-268 del
262-271del, 252-271del, and 256-268del proteins showed no cell surface expression whatsoever (Fig. 5A). These results were consistent with the protein localization shown in Figures 1 to 3. In Figure 5A series, the band intensities of the wild-type AQP-2 (100 ± 27) (means ± SE in arbitrary units) (N = 4) and 226-240del (85 ± 21) (N = 4) were not significantly different (Fig. 5C). In contrast, the apical expression was significantly decreased in the 241255del (28 ± 11) (N = 4) (P < 0.05), 269-271del (38 ± 12) (N = 4) (P < 0.05), and S256A (13 ± 6) (N = 4) (P < 0.05). In Figure 5B series, AQP-4–AQP-2 (35 ± 14) (N = 4) (P < 0.05) was significantly lower than the wild-type AQP-2 (100 ± 22) (N = 4) and 241-252del (91 ± 17) (N = 4). We quantitatively determined the effect of forskolin on the apical surface delivery of AQP proteins in the last protocol (Fig. 5D). Forskolin significantly increased apical expression of the wild-type AQP-2 (100 ± 15 to 208 ± 21) (N = 6) (P < 0.01), 241-252del (87 ± 14, to 193 ± 20) (N = 6) (P < 0.01), 241-255del (39 ± 6 to 61 ± 8) (N = 6) (P < 0.05), and AQP-4–AQP-2 (48 ± 7 to 76 ± 9) (N = 6) (P < 0.05) (Fig. 5E). In contrast, the effect of forskolin was absent in the 269-271AAA (65 ± 11 to 71 ± 8) (N = 6). DISCUSSION Previous reports suggested that the phosphorylation of Ser-256 is a key signal for the plasma membrane targeting of AQP-2 [13–16], and the S256A proteins were distributed at the intracellular vesicles [15] or Golgi complex [16], but basically absent in the apical membrane. Similarly, our findings identified the partial expression of S256A at the Golgi complex and intracellular vesicles (Fig. 3). The apical expression of biotin-labeled S256A was indeed identified, but the level of expression reached only 13% of that of the wild-type AQP-2 (Fig. 5). A low level of S256A expression at the apical surface is consistent with recent evidence that another mechanism
Fig. 4. Immunoblot of membrane fractions of the Madin-Darby canine kidney (MDCK) cells transfected with the wildtype aquaporin-2 (AQP-2) and C-terminal AQP-2 mutants detected by an antibody against HA peptide. (A) Lane 1, control (mock-transfected cells); lane 2, the wild-type AQP-2; lane 3, 262-271del; lane 4, 252-271del; lane 5, 242-271del; lane 6, 232-271del; lane 7, 226-240del; lane 8, 241-255del; lane 9, 269271del; lane 10, 256-268del; lane 11, S261A; and lane 12, 262-268del. Ten micrograms of total membrane protein was loaded in each lane. (B) Samples of the 262-271del, 252-271del, and 256-268del before (−) and after (+) incubation with endoglycosidase H (Endo). HA, hemagglutinin.
independent of Ser-256 phosphorylation takes part in AQP-2 apical trafficking. Studies in transfected LLCPK 1 cells have suggested that AQP-2 is constitutively recycled between the plasma membrane and intracellular vesicles in a vasopressin-independent manner [23]. This phosphorylation-independent apical trafficking of AQP-2 has also been suggested by others [24–26]. We performed cell surface biotinylation assay to examine the effect of forskolin quantitatively. The apical expression of the wild-type AQP-2 was increased by 108% in response to forskolin (Fig. 5). The basal expression level before forskolin treatment and the magnitude of stimulation by forskolin were similar in the wild-type AQP-2 and 241252del, a mutant possessing the consensus sequence for cAMP-dependent protein kinase. On the other hand, the basal expression of the 241-255del and AQP-4–AQP-2 were significantly lower than that of the wild-type AQP2, reaching 39% and 48% of the latter level, respectively (Fig. 5). The lack of the consensus sequence for cAMPdependent protein kinase (residues 253-256) might be responsible for the decrease in apical trafficking of these proteins. Forskolin significantly increased apical delivery of the 241-255del and AQP-4–AQP-2 by 56% and 58%, respectively, suggesting the presence of the site for cAMP-dependent phosphorylation in these proteins (Fig. 5). However, the degree of forskolin-induced stimulation was lower than that of the wild-type AQP-2 in the 241-255del and AQP-4–AQP-2. We speculate from these observations that the serine residue corresponding to S256 in the wild-type AQP-2 is still cAMP-dependent phosphorylation site and important for apical trafficking in these proteins. Our results on the AQP-4–AQP-2 chimera revealed an important role of residues 256-271 in the apical expression of AQP-2 (Fig. 2). The basolateral trafficking in AQP-4 is determined by a tyrosine-based motif (residues 277-280) and dileucine-based motif (residues 288-294) at the C-terminal sequence [27]. A C-terminal deletion
2006
Kuwahara et al: Apical trafficking of AQP-2 and C-terminus
A
B
AP
C on tro W l T AQ AQ P2 P4 -A 24 QP 2 125 2d el
C
W T
on tro l AQ P2 W TA 26 QP 4 227 1d 25 el 227 22 1d el 624 0d 24 el 125 5d 26 el 927 1d 25 el 626 8d el S2 56 A
BL
C
100
50
*
*
*
D FSK (−)
(+)
(−)
(+)
(+)
(−)
(+)
el 2d 25 1-
24
AQ P4
S2
27 926
(−)
-A Q P2
56
A
el 1d
el 5d
24
1-
25
0d 24
22
6-
W T
AQ P2
el
*
(−)
(+)
(−)
(+)
26
9-
27
1A
AQ P4 -A Q
AA
P2
el 5d 25 24
124
W
1-
25
2d
el
AQ P2 T
C
on
tro
l
AP
E FSK (−)
(+)
(−)
(+)
(−)
(+)
(−)
(+)
(−)
(+)
** **
1A
AA
*
27
AQ P4 -A Q P2
el 24
1-
25
5d
el 2d 25 124
W
T
AQ P2
*
9-
100
26
200
Fig. 5. Biotinylation of cell surface proteins. Confluent monolayers of transfected Madin-Darby canine kidney (MDCK) cells were incubated with Sulfo-NHS-LC-Biotin and solubilized. Biotin-labeled proteins were separated from unlabeled proteins by adding streptavidin-agarose. Aquaporin (AQP) proteins present in the pellet were detected by immunoblot using an antibody against HA peptide. Biotin was added either to the apical (AP) or basolateral (BL) compartment of the filter chamber. (A) Lane 1, control (mock-transfected cells); lane 2, wild-type AQP-2; lane 3, wild-type
Kuwahara et al: Apical trafficking of AQP-2 and C-terminus
mutant of AQP-4 in this study (285-323del) was localized at the basolateral membrane, probably because it still possessed the tyrosine-based motif. Unexpectedly, however, the addition of the AQP-2 C-terminal sequence (residues 256-271) to 285-323del AQP-4 resulted in apical expression of this chimera. It thus appeared that the sequence at residues 256-271 in AQP-2 was sufficient for apical trafficking and dominant over the basolateral signal of the tyrosine-based motif in AQP-4. In contrast to our findings, a chimera protein of the transmembrane region of placental alkaline phosphatase with the AQP-2 C-terminal tail (residues 220-271) failed to express at the apical membrane in an earlier study, suggesting that the C-terminal sequence was not sufficient for apical expression [28]. Notwithstanding, this earlier chimeric protein study did not rule out nonspecific processing defect between the endoplasmic reticulum and cell surface. The chimera protein in our study was so carefully designed that the corresponding sites of the C-terminal tail were exchanged between AQP-2 and AQP-4. A recent study examined the role of AQP-2 C-terminus by generating chimeras of AQP-1 and AQP-2 [29]. AQP-1 with AQP-2 residues 220-271 showed apical localization, whereas AQP-1s with AQP-2 residues distal to 230 were distributed at apical and basolateral membranes. This implied that amino acids 220-229 are important for apical sorting in AQP-2. While the present study found that residues 226-229 were probably not responsible for apical expression (226-240del) (Fig. 1), it did not examine the significance of residues 220-225. AQP-2 possesses a potential class I PDZ domain binding motif (X-S/T-X-) at the C-terminus (G-T-K-A) (268-271). PDZ domains are ubiquitous protein-protein interacting modules that recognize specific sequence motifs of transmembrane receptors, channels, and other PDZ-containing proteins. Indeed, recent evidence suggests that PDZ-interacting motifs mediate apical sorting of numerous channels and transporters, including cystic fibrosis transmembrane conductance regulator (CFTR), Na/H exchanger regulatory factors (NHERFs), c-aminobutyric acid transporter 3 (GAT-3), multidrugresistance-associated protein 2 (MRP2), and Na/Pi cotransporter 1 and 2 (Npt1 and Npt2) [17, 18]. For example, the last three amino acids at the C-terminus of the CFTR (T-R-L) comprise a class I PDZ-interacting motif. The mistargeting of a mutant lacking this domain (S1455X) to the basolateral membrane was found to lead
2007
the pathogenesis of cystic fibrosis [30]. Interaction of NHERF and ezrin-radixin-moesin-binding phosphoprotein 50 (EBP50), another PDZ protein, appears to regulate NHE3 through the protein kinase A pathway and b-adrenergic receptor coupled pathway [31]. In the present study, the 269-271AAA, a deletion mutant of a putative AQP-2 PDZ ligand, was apparently distributed in the subapical area irrespective of incubation with forskolin (Fig. 3). Biotinylation analysis showed a decrease in the confirmable apical expression of the 269271AAA (34% of that of the wild-type AQP-2) (Fig. 5). The 269-271AAA was mostly colocalized with intracellular vesicles, with some additional staining in the Golgi complex, late endosome/lysosome, and endoplasmic reticulum (Fig. 3). The signal-induced proliferationassociated gene-1 (SPA-1) is a guanine triphosphatase (GTPase)-activating protein (GAP) for Rap1 [32]. Our group recently identified SPA-1 as a PDZ protein that directly binds to the potential PDZ-interacting motif of AQP-2 (residues 268-271) [33]. We showed that SPA-1 was colocalized with AQP-2 in the collecting duct and that the GAP activity of SPA-1 was at least partially necessary for the apical sorting of AQP-2. Now that the Rap1 GTPases may be critical regulators of vesicle trafficking [34, 35], we speculated that the SPA-1 binding to AQP-2 reduces the levels of Rap1GTP that trigger F-actin disassembly and thereby leads to an activation of AQP-2 exocytosis from the intracellular vesicles to the apical membrane [33]. The retention of most of the AQP-2– containing vesicles in the subapical compartment in the present study was presumably due to the deletion of the putative PDZ-interacting motif. The apical localization of AQP-2 observed in this study was consistent with previous transfection studies of AQP2 protein with MDCK cells [36, 37]. In contrast to MDCK cells, AQP-2 transfected into LLC-PK 1 cells was translocated to the basolateral membrane upon vasopressin stimulation [38]. It has been reported that LLC-PK 1 cells exhibit mislocalization to the apical membrane in several proteins such as the low-density lipoprotein receptor (LDLR), transferrin receptor (TfR), and H-K-APTase beta subunit [39, 40]. The misrouting of LDLR and TfR was due to the lack of the l1B subunit of AP-1B complex because transfection of cDNA l1B into LLC-PK 1 cells restored the basolateral distribution of these proteins [39]. However, introduction of l1B failed to correct the basolateral localization of AQP-2 and the apical localization
AQP-4; lane 4, 262-271del; lane 5, 252-271del; lane 6, 226-240del; lane 7, 241-255del; lane 8, 269-271del; lane 9, 256-268del; and lane 10, S256A. (B) Lane 1, control; lane 2, wild-type AQP-2; lane 3, AQP-4–AQP-2; lane 4, 241-252del. (C) Summary of apical surface expression. ∗ P < 0.05 compared with wild-type AQP-2. (D) Effect of forskolin (FSK) on apical expression of AQP proteins. Cell surface biotinylation was assessed before (−) and after (+) the incubation with forskolin. Lanes 1 and 2, control; lanes 3 and 4, wild-type AQP-2; lanes 5 and 6, 241-252del; lanes 7 and 8, 241-255del; lanes 9 and 10, AQP-4–AQP-2; lanes 11and 12, 269-271AAA. (E) Summary of forskolin action on apical surface expression. ∗ P < 0.05; ∗∗ P < 0.01 compared with before pretreatment with forskolin. HA, hemagglutinin.
2008
Kuwahara et al: Apical trafficking of AQP-2 and C-terminus
of H-K-APTase beta subunit in LLC-PK 1 cells [7, 40]. These findings suggested that the LLC-PK 1 cells contain another defect in sorting machinery besides the lack of the l1B subunit. Recent evidence suggests that AQP-2 is expressed at the basolateral membrane of connecting tubule (CNT) cells and inner medullary collecting duct (IMCD) principal cells [6, 41–43]. Acute administration of oxytocin redistributed cytoplasmic AQP-2 proteins to the apical as well as the basolateral membrane in CNT cells and IMCD principal cells of Sprague-Dawley rats [42]. Moreover, 2-hour treatment with desamino-Cys1 , D-Arg8 vasopressin (dDAVP) markedly increased apical expression of AQP-2 in CNT cells and IMCD principal cells of Brattleboro rats [43]. These results suggested that the stimulation of cAMP-dependent protein kinase accelerates apical delivery of AQP-2 even in these cells. The mechanisms underlying the basolateral targeting of AQP-2 is unknown at present, although oxytocin and hypertonicity were suggested to cause basolateral redistribution of AQP-2 [42, 44]. Sorting mechanisms in CNT cells and IMCD principal cells might be different from those in cortical and outer medullary collecting duct principal cells. ACKNOWLEDGMENTS We thank Dr. E. Kominami for providing us an antibody against Lamp1. This work was supported by a Grant-in-Aid for Scientific Research and Grant-in-Aid for Scientific Research on Priority Areas from Japan Society for the Promotion of Science (JSPS). Reprint requests to Michio Kuwahara, M.D., Department of Nephrology, Graduate School, Tokyo Medical and Dental University, Tokyo 1138519, Japan. E-mail: [email protected]
REFERENCES 1. AGRE P, KING LS, YASUI M, et al: Aquaporin water channels—From atomic structure to clinical medicine. J Physiol 542:3–16, 2002 2. ISHIBASHI K, KUWAHARA M, SASAKI S: Molecular biology of aquaporins. Rev Physiol Biochem Pharmacol 141:1–32, 2000 3. HATAKEYAMA S, YOSHIDA Y, TANI T, et al: Cloning of a new aquaporin (AQP10) abundantly expressed in duodenum and jejunum. Biochem Biophys Res Commun 287:814–819, 2001 4. ISHIBASHI K, MORINAGA T, KUWAHARA M, et al: Cloning and identification of a new member of water channel (AQP10) as an aquaglyceroporin. Biochim Biophys Acta 1576:335–340, 2002 5. FUSHIMI K, UCHIDA S, HARA Y, et al: Cloning and expression of apical membrane water channel of rat kidney collecting tubule. Nature 361:549–552, 1993 6. NIELSEN S, DIGIOVANNI SR, CHRISTENSEN EI, et al: Cellular and subcellular immunolocalization of vasopressin-regulated water channel in rat kidney. Proc Natl Acad Sci USA 90:11663–11667, 1993 7. BROWN D: The ins and outs of aquaporin-2 trafficking. Am J Physiol 284:F893–F901, 2003 8. KUWAHARA M, IWAI K, OOEDA T, et al: Three families with autosomal dominant nephrogenic diabetes insipidus caused by aquaporin-2 mutations in the C-terminus. Am J Hum Genet 69:738–748, 2001 9. ASAI T, KUWAHARA M, KURIHARA H, et al: Pathogenesis of nephrogenic diabetes insipidus by aquaporin-2 C-terminus mutations. Kidney Int 64:2–10, 2003 10. MARR N, BICHET DG, LONERGAN M, et al: Heteroligomerization of an aquaporin-2 mutant with wild-type aquaporin-2 and their mis-
11. 12.
13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.
24.
25. 26.
27.
28. 29. 30. 31. 32.
33.
routing to late endosomes/lysozomes explains dominant nephrogenic diabetes insipidus. Hum Mol Genet 11:779–789, 2002 KAMSTEEG EJ, BICHET DG, KONINGS IB, et al: Reversed polarized delivery of an aquaporin-2 mutant causes dominant nephrogenic diabetes insipidus. J Cell Biol 163:1099–1109, 2003 KUWAHARA M, FUSHIMI K, TERADA Y, et al: cAMP-dependent phosphorylation stimulates water permeability of aquaporin-collecting duct water channel protein expressed in Xenopus oocytes. J Biol Chem 270:10384–10387, 1995 FUSHIMI K, SASAKI S, MARUMO F: Phosphorylation of serine 256 is required for cAMP-dependent regulatory exocytosis of the aquaporin-2 water channel. J Biol Chem 272:14800–14804, 1997 KATSURA T, GUSTAFSON CE, AUSIELLO DA, et al: Protein kinase A phosphorylation is involved in regulated exocytosis of aquaporin-2 in transfected LLC-PK 1 cells. Am J Physiol 272:F817–F822, 1997 VAN BALKOM BW, SAVELKOUL PJ, MARKOVICH D, et al: The role of putative phosphorylation sites in the targeting and shuttling of the aquaporin-2 water channel. J Biol Chem 277:41473–41739, 2002 PROCINO G, CARMOSINO M, MARIN O, et al: Ser-256 phosphorylation dynamics of aquaporin 2 during maturation from the ER to the vesicular compartment in renal cells. FASEB J 17:1886–1888, 2003 HARRIS BZ, LIM WA: Mechanism and role of PDZ domains in signaling complex assembly. J Cell Sci 114:3219–3231, 2001 JELEN F, OLEKSY A, SMIETANA K, et al: PDZ domains—Common players in the cell signaling. Acta Biochim Pol 50:985–1017, 2003 LEFEBVRE B, FORMSTECHER P, LEFEBVRE P: Improvement of the gene splicing overlap (SOE) method. Biotechniques 19:186–188, 1995 TAMARAPPOO BK, VERKMAN AS: Defective aquaporin-2 trafficking in nephrogenic diabetes insipidus and correction by chemical chaperones. J Clin Invest 101:2257–2267, 1998 YANG B, MA T, XU Z, et al: cDNA and genomic cloning of mouse aquaporin-2: functional analysis of an orthologous mutant causing nephrogenic diabetes insipidus. Genomics 57:79–83, 1999 HALBAN PA, IRMINGER JC: Sorting and processing of secretory proteins. Biochem J 299:1–18, 1994 GUSTAFSON CE, KATSURA T, MCKEE M, et al: Recycling of AQP2 occurs through a temperature- and bafilomycin-sensitive trans-Golgiassociated compartment. Am J Physiol Renal Physiol 278:F317– F326, 2000 VALENTI G, PROCINO G, CARMOSINO M, et al: The phosphatase inhibitor okadaic acid induces AQP2 translocation independently from AQP2 phosphorylation in renal collecting duct cells. J Cell Sci 113:1985–1992, 2000 SUN TX, VAN HOEK A, HUANG Y, et al: Aquaporin-2 localization in clathrin-coated pits: Inhibition of endocytosis by dominant-negative dynamin. Am J Physiol Renal Physiol 282:F998–F1011, 2002 LU H, SUN TX, BOULEY R, et al: Inhibition of endocytosis causes phosphorylation (S256)-independent plasma membrane accumulation of AQP2. Am J Physiol Renal Physiol 286:F233–F243, 2004 MADRID R, LE MAOUT S, BARRAULT M-B, et al: Polarized trafficking and surface expression of the AQP4 water channel are coordinated by serial and regulated interactions with different clathrin-adaptor complexes. EMBO J 20:7008–7021, 2001 DEEN PMT, VAN BALKOM BWM, SAVELKOUL PJM, et al: Aquaporin2: COOH terminus is necessary but not sufficient for routing to the apical membrane. Am J Physiol 282:F330–F340, 2002 VAN BALKOM BWM, GRAAT MPJ, VAN RAAK M, et al: Role of cytoplasmic termini in sorting and shuttling of the aquaporin-2 water channel. Am J Physiol 286:C372–C379, 2004 MOYER BD, DENTON J, KARLSON KH, et al: A PDZ-interacting domain in CFTR is an apical membrane polarization signal. J Clin Invest 104:1353–1361, 1999 LAMPRECHT G, WEINMAN EJ, YUN CH: The role of NHERF and E3KARP in the cAMP-mediated inhibition of NHE3. J Biol Chem 273:29972–29978, 1998 HATTORI M, TSUKAMOTO N, NUR-E-KAMAL MS, et al: Molecular cloning of a novel mitogen-inducible nuclear protein with a Ran GTPase-activating domain that affects cell cycle progression. Mol Cell Biol 15:552–560, 1995 NODA T, HORIKAWA S, FURUKAWA T, et al: Aquaporin-2 trafficking is regulated by PDZ-domain containing protein SPA-1. FEBS Lett 18:139–145, 2004
Kuwahara et al: Apical trafficking of AQP-2 and C-terminus
34. MOSKALENKO S, HENRY DO, ROSSE C, et al: The exocyst is a Ral effector complex. Nat Cell Biol 4:66–72, 2002 35. ALTSCHULER Y, HODSON, MILGRAM SL: The apical compartment: Trafficking pathways, regulators and scaffolding proteins. Curr Opin Cell Biol 15:423–429, 2003 36. VALENTI G, FRIGERI A, RONCO PM, et al: Expression and functional analysis of water channels in a stably AQP2-transfected human collecting duct cell line. J Biol Chem 271:24365–24370, 1996 37. DEEN PM, RIJSS JP, MULDERS SM, et al: Aquaporin-2 transfection of Madin-Darby canine kidney cells reconstitutes vasopressinregulated transcellular osmotic water transport. J Am Soc Nephrol 8:1493–1501, 1997 38. KATSURA T, VERBAVATZ JM, FARINAS J, et al: Constitutive and regulated membrane expression of aquaporin 1 and aquaporin 2 water channels in stably transfected LLC-PK 1 epithelial cells. Proc Natl Acad Sci USA 92:7212–7216, 1995 39. FOLSCH H, OHNO H, BONIFACINO JS, et al: A novel clathrin adaptor
40. 41. 42. 43. 44.
2009
complex mediates basolateral targeting in polarized epithelial cells. Cell 99:189–198, 1999 DUFFIELD A, FOLSCH H, MELLMAN I, et al: Sorting of H,K-ATPase beta-subunit in MDCK and LLC-PK cells is independent of mu 1B adaptin expression. Traffic 5:449–461, 2004 COLEMAN RA, WU DC, LIU J, et al: Expression of aquaporins in the renal connecting tubule. Am J Physiol Renal Physiol 279:F874– F883, 2000 JEON US, JOO KW, NA KY, et al: Oxytocin induces apical and basolateral redistribution of aquaporin-2 in rat kidney. Nephron Exp Neohrol 93:e36–e45, 2003 CHRISTENSEN BM, WANG W, FROKIAER J, et al: Axial heterogeneity in basolateral AQP2 localization in rat kidney: effect of vasopressin. Am J Physiol Renal Physiol 284:F701–F717, 2003 VAN BALKOM BW, VAN RAAK M, BRETON S, et al: Hypertonicity is involved in redirecting the aquaporin-2 water channel into the basolateral, instead of the apical, plasma membrane of renal epithelial cells. J Biol Chem 278:1101–1107, 2003