Human Molecular Genetics, 2002, Vol. 11, No. 7 779-789
© 2002 Oxford University Press
Heteroligomerization of an Aquaporin-2 mutant with wild-type Aquaporin-2 and their misrouting to late endosomes/lysosomes explains dominant nephrogenic diabetes insipidus
160 Department of Cell Physiology, Nijmegen Center of Molecular Life Sciences, University Medical Center St Radboud, PO Box 9101, 6500 HB Nijmegen, The Netherlands, 1Department of Medicine, University of Montreal and Centre de Recherches, Hôpital du Sacre-Coeur de Montreal, Montreal, Quebec, Canada, 2Universitätskinderklinik, 35033 Marburg, Germany, 3Forschungsinstitut für Molekulare Pharmakologie, 13125 Berlin, Germany and 4Institut für Pharmakologie, 14195 Berlin, Germany
Received November 24, 2001; Revised and Accepted January 28, 2002.
DDBJ/EMBL/GenBank accession no. AF147093.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Autosomal nephrogenic diabetes insipidus (NDI), a disease in which the kidney is unable to concentrate urine in response to vasopressin, is caused by mutations in the Aquaporin-2 (AQP2) gene. Analysis of a new family with dominant NDI revealed a single nucleotide deletion (727
G) in one AQP2 allele, which encoded an AQP2 mutant with an altered and extended C-terminal tail. When expressed in oocytes, the tetrameric AQP2–727G was retained within the cell. When co-expressed, AQP2–727G, but not a mutant in recessive NDI (AQP2–R187C), formed hetero-oligomers with wild-type (wt) AQP2 and reduced the water permeability of these oocytes, because of a reduced plasma membrane expression of wt-AQP2. Expressed in renal epithelial cells, AQP2–727G predominantly localized to the basolateral membrane and late endosomes/lysosomes, whereas wt-AQP2 was expressed in the apical membrane. Upon co-expressing in these cells, wt-AQP2 and AQP2–727G mainly co-localized to late endosomes/lysosomes. In conclusion, hetero-oligomerization of AQP2–727G with wt-AQP2 and consequent mistargeting of this complex to late endosomes/lysosomes results in absence of AQP2 in the apical membrane, which can explain dominant NDI in this family. Together with other mutants in dominant NDI, our data reveal that a misrouting, instead of a lack of function, is a general mechanism for the ‘loss of function’ phenotype in dominant NDI and visualizes for the first time a mislocalization of a wild-type protein to late endosomes/lysosomes in polarized cells after oligomerization with a mutant protein. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
The solute-independent reabsorption of water to regulate whole-body water homeostasis is one of the key functions of the kidney. This process mainly occurs through renal collecting duct cells and is under control of the anti-diuretic hormone arginine-vasopressin (AVP). In conditions of hypernatremia or hypovolemia, AVP is released from the pituitary. Binding of AVP to the vasopressin type-2 receptor (V2R) initiates a cAMP signalling cascade, resulting in activation of protein kinase A (PKA). This protein phosphorylates the Aquaporin-2 (AQP2) water channel and presumably other proteins, leading to a redistribution of AQP2 from intracellular vesicles to the apical membrane. Using a sodium osmotic gradient as a driving force, water then enters the cell via AQP2 and crosses into the interstitial fluid through AQP3 and AQP4, which are located in the basolateral membrane. The reduction in urine concentrating ability in mice and man lacking functional AQP2–4 shows their essential roles in this process (1–3). After dissociation of AVP from its receptor, this process is reversed (4,5).
Nephrogenic diabetes insipidus (NDI) is a disease in which the kidney is unable to concentrate urine in response to AVP. Polyuria and polydipsia are present at birth and need to be recognized early to avoid dehydration, which can result in mental retardation. The defect in urine concentration is accompanied by other early symptoms including vomiting, anorexia, failure to thrive, fever and constipation (6). NDI is most commonly acquired (7), but occurs at a frequency of approximately 10 per million as an inherited disorder (8). Three patterns of inheritance have been described. The X-linked form of NDI is caused by mutations in the V2R gene, whereas the autosomal recessive and dominant forms of NDI are consequences of mutations in the AQP2 gene. Most mutations identified in the AQP2 gene result in the recessive form of NDI (1,9). Expression in oocytes of missense AQP2 mutants in recessive NDI revealed that all were impaired in their export from the endoplasmic reticulum (ER), presumably caused by misfolding of the mutant proteins (10,11). This ER retention provides the molecular basis for recessive NDI caused by these mutants, although some of these mutants are functional (10–13).
Until recently, one AQP2 mutant (AQP2–E258K) had been reported as causing dominant NDI. Expression studies in oocytes showed that AQP2–E258K was a properly folded functional water channel, which was retained in the Golgi region (14). In co-expression studies with wild-type (wt)-AQP2, a dominant-negative effect was observed, caused by an impaired routing of wt-AQP2 to the plasma membrane after hetero-oligomerization with AQP2–E258K (15). During the course of this study, a similar inpairment of wt-AQP2 in the routing to the plasma membrane of oocytes has been reported for three other AQP2 mutants in dominant NDI (721delG, 763–772del, 812–818del) (16).
We recently identified a new family in which NDI seemed to be inherited as an autosomal dominant trait. To identify the cause and the cell biological mechanism underlying NDI in this family, the genomic DNAs of the family members were screened for mutations and the mutant AQP2 protein encoded by genomic DNA of an affected individual was subjected to detailed analysis upon expression in oocytes and epithelial cells.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Identification of an AQP2 mutation in a family with dominant NDI
Recently, a NDI family was presented in which the female proband and her father were affected, while the mother and brother were healthy (Fig. 1). Since mutations in the sex-linked V2R gene are the most common cause of congenital NDI and could, based on the pedigree, not be excluded, its sequence was determined. However, no mutation in the V2R coding sequence was found. Subsequent haplotype analysis and sequence analysis of the AQP2 gene revealed that in this family, NDI segregated with a guanosine deletion at position 727 of the AQP2 gene in one allele, leading to a C-terminal frame shift, while the other allele of the affected family members encoded a wt-AQP2 protein (Fig. 1). This 727
G mutation was not observed in more than 60 AQP2 alleles of healthy individuals.
|
Urine osmolality after 1-deamino-8-D-arginine-vasopressin (dDAVP) infusion in the patients
Administration of the synthetic vasopressin analogue, dDAVP, to humans induces redistribution of AQP2 to the apical membrane of renal collecting duct cells, thereby initiating the concentration of urine. Upon admission to the hospital, the urine osmolarities were 116, 685 and 381 mmol/kg for the proband and her healthy mother and father, respectively (the father underwent a kidney transplantation). To test whether dDAVP would induce a temporal increase in the urinary concentrating ability in the NDI patient carrying the AQP2–727
G mutation, the urine osmolarity was measured in time following an intravenous dDAVP administration for 30 min. In contrast to other patients with dominant NDI and healthy individuals (14,16), however, no increase in urine osmolality was observed in the patient (124, 108, 104 and 118 mOsm/kg at 0.3, 1, 2 and 4 h, respectively).
In oocytes, AQP2–727G is impaired in its routing to the plasma membrane and confers a dominant effect on wild-type AQP2
To determine whether and, if so, how the identified mutation causes dominant NDI in this family, expression of the mutant protein was essential. However, because the 727G mutation changed the reading frame, the normal AQP2 gene stop codon was out of frame and the AQP2–727G protein coding sequence extended into the gene segment that normally represents the 3'-untranslated region (3'-UTR). Therefore, to obtain the exact coding sequence of the C-tail of AQP2–727G, the sequences of the 3'-UTRs of the AQP2 genes of the proband and of a healthy individual were determined. Sequence analysis of six cloned PCR fragments, obtained from three independent amplification reactions revealed that the 3'-UTR of the patient’s mutant AQP2 gene was identical to that of the healthy individual, except for a 
836A
C (3'-UTR) mutation (Fig. 2). Sequence analysis of genomic DNAs of 53 independent human subjects revealed that this nucleotide transition was present in 21 individuals, indicating that this mutation is a widely occurring polymorphism. Based on the obtained 3'-UTR sequence of the proband, the mutant AQP2 gene encodes a mutant AQP2 protein of 333 amino acids (35 kDa), in which the original C-terminal 29 amino acids of wt-AQP2 are altered, subsequently followed by 62 new amino acids (Fig. 2).
|
To determine the molecular mechanism underlying dominant NDI in this family, the 727
G mutation and the proband’s 3'-UTR were introduced into the wt-AQP2 cDNA. N-terminal tagging of AQP2 proteins has been shown not to interfere with AQP2 routing nor with its function in oocytes (15). Therefore, both cDNAs were also cloned downstream of a vesicular stomatitis virus (VSV) tag coding region to allow comparison of wt-AQP2 and AQP2–727G expression levels. Since low expression levels have been shown to be essential to reveal the dominant effect of AQP2–E258K (17), oocytes were subsequently injected with 0.3 ng cRNA encoding V-AQP2–727G, V-AQP2 or AQP2–R187C only or in combination (i.e. in these latter cases a total of 0.6 ng cRNA was injected). Copy RNA encoding AQP2–R187C, which is an ER-retained AQP2 mutant encoded in patients with recessive NDI (10), was taken along as a control. Two days after cRNA injection, osmotic swelling assays revealed that the water permeabilities (Pf ± SEM in µm/s) of oocytes injected with cRNAs encoding AQP2–R187C or V-AQP2–727G were not different from those of non-injected controls (24 ± 16), whereas oocytes injected with V-AQP2 cRNAs revealed a Pf of 118 ± 27 (Fig. 3A). In addition, the Pf of oocytes co-expressing V-AQP2–727G and V-AQP2 was significantly reduced compared to Pfs of oocytes expressing only V-AQP2 or those co-expressing V-AQP2 and AQP2–R187C (Fig. 3A). Since a decrease in Pf of co-injected oocytes could also be caused by a reduced expression of wt-AQP2, as a consequence of co-expression of the mutant protein (17), the expression levels of the individual AQP2 proteins in these oocytes was determined. Immunoblot analysis of total membranes of these oocytes, however, revealed similar expression levels throughout (Fig. 3B, TM), indicating that the low Pf of oocytes expressing V-AQP2–727G and V-AQP2 was not due to reduced levels of V-AQP2. The blots also show that V-AQP2–727G migrates as a protein of 
36 kDa, which is in line with the combined mass of the VSV tag (1 kDa) and that of AQP2–727G as deduced from its coding sequence (35 kDa). At a high level of V-AQP2–727G expression (Fig. 3B), obtained with injection of 10 ng cRNA, a Pf similar to that of oocytes injected with 0.3 ng V-AQP2 cRNA was obtained (Fig. 3A), indicating that V-AQP2–727G is a functional water channel, but that the conferred water permeability is strongly reduced compared to that of V-AQP2.
|
12) of these oocytes was measured. Non-injected oocytes were taken as a control. An asterisk indicates a significant decrease in Pf compared to oocytes expressing V-AQP2 (P < 0.01). (B) Expression in oocytes. From oocytes injected as described for (A), total membranes (TM) or plasma membranes (PM) were isolated (n Immunoblotting of isolated plasma membranes of injected oocytes revealed that, in contrast to V-AQP2, neither AQP2–R187C nor V-AQP2–727
G was detectable in the plasma membrane (Fig. 3B, PM), which clearly indicated that both mutants were impaired in their routing to the plasma membrane. Also, the plasma membrane/total membrane ratio of expression of V-AQP2 in oocytes co-expressing V-AQP2–727G (0.7 ± 0.1; n = 3) was reduced compared to the V-AQP2 ratios obtained for oocytes expressing V-AQP2 only (1.4 ± 0.1; n = 3) or together with AQP2–R187C (1.5 ± 0.2; n = 3; Fig. 3B). These results clearly indicated that V-AQP2 was retained inside the cell when co-expressed with V-AQP2–727G, but not with AQP2–R187C.
These data were corroborated by immunocytochemical analysis of oocytes, which revealed that V-AQP2–727G (data not shown) and untagged AQP2–727G (Fig. 4A) were mainly retained within the cell, whereas V-AQP2 (data not shown) and untagged wt-AQP2 (Fig. 4C) were solely present in the plasma membrane. Note that for immunocytochemistry, the oocytes expressing AQP2–727G were injected with 10 ng cRNA, which was needed to detect the protein. Consistent with the immunoblot data (Fig. 3B), these oocytes showed some AQP2–727G expression in the plasma membrane. Incubation of non-injected oocytes with affinity-purified antibodies specifically recognizing AQP2–727G (Fig. 4B) or wt-AQP2 (Fig. 4D) did not show any staining, illustrating the specificity of the antibodies used.
|
AQP2–727
G and wt-AQP2 form hetero-oligomersAQP2 is known to form homotetramers both in vivo and in vitro and the first mutant in dominant NDI, AQP2–E258K, was shown to exert its dominant effect by retaining wt-AQP2 in the Golgi-complex region after the formation of hetero-oligomers (15). To determine the oligomerization state of AQP2–727
G, membranes of oocytes expressing AQP2–727G, AQP-R187C or wt-AQP2 were isolated, dissolved in desoxycholate and sedimented on a sucrose gradient. Immunoblotting of fractions taken from these gradients revealed that AQP2–727G sedimented as a protein of 120–140 kDa (Fig. 5A). Since AQP2–727G has a deduced molecular mass of 35 kDa, the sedimentation as a protein complex of 120–140 kDa indicated that AQP2–727G is expressed as a homotetramer. As reported previously (15), the sedimentation of wt-AQP2 and AQP2–R187C as 97–130 kDa and 28–67 kDa complexes is consistent with a tetrameric and monomeric structure, respectively.
|
To test whether the formation of hetero-oligomers between AQP2–727
G and wt-AQP2 could explain its dominant effect, the wt-AQP2 cDNA was cloned downstream of a FLAG tag coding region (F-AQP2) to allow efficient protein-specific immunoprecipitations. Two days after cRNA injections, membranes of oocytes expressing F-AQP2, V-AQP2–727G or V-AQP2–R187C only or combined were solubilized and subjected to immunoprecipitations using anti-FLAG antibodies. Whereas immunoblotting of total membranes using VSV antibodies revealed a clear expression of both V-tagged AQP2 mutants (Fig. 5B, top), immunoblot analysis of the immunoprecipitates using VSV antibodies revealed that V-AQP2–727G, but not V-AQP2–R187C, co-precipitated with F-AQP2 (Fig. 5B, middle). Immunoblotting of the precipitates using wt-AQP2 specific antibodies revealed the presence of F-AQP2 in precipitates from all oocytes (co-)injected with F-AQP2 cRNAs (Fig. 5B, bottom). Therefore, the failure to detect V-AQP2–R187C in the immunoprecipitates was not caused by a lack of expression of F-AQP2 in oocytes injected with F-AQP2 and V-AQP2–R187C cRNAs.
In Madin–Darby canine kidney (MDCK) cells, AQP2–727G mistargets wt-AQP2 to late endosomes/lysosomes
In collecting duct cells and in MDCK cells stably expressing AQP2 (wt10 cells), AQP2 mainly localizes to intracellular vesicles and is redistributed to the apical membrane upon treatment with the adenylate cyclase activator, forskolin (18). Since the subcellular localization of AQP2–727G in oocytes was not clear and oocytes are, in contrast to renal collecting duct cells, not polarized, an AQP2–727G expression construct was transfected into MDCK cells. Several clones expressing AQP2–727G were grown to confluence, treated with forskolin, and subjected to immunocytochemistry using antibodies directed against AQP2–727G and various marker proteins for subcellular organelles. Wt10 cells were taken as a control. Confocal laser scanning microscopy revealed that AQP2–727G predominantly localized to the basolateral membrane and intracellular vesicles (Fig. 6C and D). As anticipated, wt-AQP2 in wt10 cells was localized in the apical membrane (Fig. 6A and B). In addition, vesicular AQP2–727G showed a clear co-localization with the lysosome-associated membrane protein 1 (Lamp1; Fig. 6C and D), which is a marker protein for late endosomes/lysosomes. This was not observed for wt-AQP2 (Fig. 6A and B). No co-localization of AQP2–727G was observed with markers for ER (PDI), cis-median Golgi (GOS28), trans-Golgi (giantin) and VAMP2-positive recycling vesicles. For AQP2–727G, the distribution was not changed when the forskolin treatment was omitted from the cells (data not shown).
|
To test whether AQP2–727
G could also exert a dominant-negative effect on wt-AQP2 in polarized MDCK cells, the AQP2–727G expression construct was also transfected into wt10 cells. Selection of clonal cell lines stably expressing both proteins was not successful, because immunocytochemical analysis of clonal cell lines that expressed both proteins (as revealed from immunoblots) always revealed expression of wt-AQP2 or AQP2–727G, but never of both proteins, in the same cell. Therefore, after 2 weeks of selection, pooled colonies of G418-resistent wt10 cells transfected with AQP2–727G DNA were grown to confluence, treated with forskolin and subjected to immunocytochemistry using antibodies specifically recognizing AQP2–727G, wt-AQP2 and Lamp1. Confocal laser scanning microscopy revealed that in cells that expressed both AQP2–727G and wt-AQP2, predominant co-localization of mutant and wt-AQP2 was observed with Lamp1 (Fig. 7, indicated by arrows). In contrast, in cells in which AQP2–727G expression was absent, wt-AQP2 was mainly localized in the apical membrane (Fig. 7). Although in pooled colonies of three independent transfection experiments only a few cells expressed all three proteins, they consistently showed a strong co-localization. These data, therefore, indicated that in polarized epithelial cells, AQP2–727G interacts with wt-AQP2 and redirects wt-AQP2 mainly to late endosomes/lysosomes.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
A common mechanism in dominant NDI
In this study, haplotype and sequence analysis of the V2R and AQP2 genes of a family with dominant NDI revealed a novel mutation in the AQP2 gene, consisting of a deletion of a guanosine at position 727 (Fig. 1), resulting in a protein of 35 kDa in which the last 29 amino acids of wt-AQP2 were exchanged, followed by an extension of 62 amino acids (Fig. 2). This protein sequence was based on the 3'-UTR of the AQP2 allele of the proband harbouring the 727
G mutation. Except for the727G mutation and the 836AC (3'-UTR) polymorphism, the obtained sequence differed at one position (994–995insG) from the recently submitted 3'-UTR region of the AQP2 gene (GenBank accession no. AF147093), which appeared to be a mistake in the submitted sequence and has been corrected. These sequences, however, differed at several positions from those reported previously (19). Recently, three other deletions (721delG, 763–772del, 812–818del) in exon 4 of the AQP2 gene were reported to cause dominant NDI (16). Except for the location where the out-of-frame C-terminal tails start (following Leu240, Arg254 or Lys270 and Pro242 for the three mutants and AQP2–727G, respectively), the extended tails of these mutants could be identical to that of AQP2–727G. Unfortunately, however, the origin of the 3'-UTRs of these AQP2 mutants was not described and, therefore, it is unclear whether the C-tails of these mutants are the same as that of AQP2–727G.
As found for many mutant proteins in recessive diseases (20,21), AQP2 mutants in recessive NDI are mostly non-functional and are retained in the ER, as evidenced by their expression as high-mannose glycosylated proteins of 32 kDa, besides non-glycosylated AQP2 of 29 kDa. In addition, they are expressed as monomers and are unable to oligomerize with wt-AQP2 (10–12,15,17). Although clearly different from AQP2 mutants in recessive NDI, our analyses on AQP2–727G and those performed on four other mutants in dominant NDI (14–16) reveal that also the mechanism underlying dominant NDI seems quite general. All studied AQP2 mutants in dominant NDI appear to be functional water channels (Fig. 3A for AQP2–727G) and are not retained in the ER as shown by the absence of high-mannose glycosylated bands on immunoblots (Figs 3B and 5 for AQP2–727G) and their subcellular localization (Figs 6 and 7 for AQP2–727G). Besides, they are expressed as tetramers (Fig. 5A for AQP2–727G) and are able to form hetero-oligomers with wt-AQP2 (Fig. 5B for AQP2–727G) Although the steady state localization of these hetero-oligomeric complexes might depend on the introduced mutation, the inability of these complexes to be transported to the apical membrane is likely to be fundamental to dominant NDI.
Interestingly, nearly all mutations in recessive NDI are found between the first and sixth transmembrane domain of the protein, whereas all mutations identified in dominant NDI are located in the C-terminal tail of AQP2. Since it is generally believed that recognition of misfolding of the mutant protein by the ER quality control system underlies ER retention (20), the functional and cell biological data obtained for AQP2 mutants in dominant NDI seem to indicate that, in contrast to AQP2 mutants in recessive NDI, mutations in the C-terminal tail of AQP2 do not affect the fundamental three-dimensional structure of AQP2. This is in line with the current knowledge that the highly conserved segment between the first and sixth transmembrane domain forms the pore of the water channel (22,23) whereas the less conserved C-terminal tail is important for its routing (24–26).
Misrouting of AQP2 mutants in dominant NDI
In the trans-Golgi network (TGN), proteins are sorted into different domains and vesicles to continue their route to different cellular destinations, a process which is directed by a multitude of proteins and lipids interacting with linear or non-linear targeting motifs along the peptide chain (27,28). Here, all mutants in dominant NDI appear to divert from wt-AQP2 and, possibly, each other (14–16). While wt-AQP2 is expressed in the plasma membrane of oocytes, all mutants in dominant NDI were impaired in their routing to the cell surface (Figs 3B and 4 for AQP2–727G and wt-AQP2). In contrast to AQP2–E258K, however, which was retained in the Golgi complex region of Xenopus oocytes located just below the cell surface (14), AQP2–727G was found in intracellular vesicles localized throughout the oocyte (Fig. 4). Expression of AQP2–721G, AQP2 763–772del and AQP2 812–818del revealed a similar pattern to that of AQP2–727G, although the specificity of their antibodies against the AQP2 mutants was not checked on control oocytes (16). Together, these results suggest that in oocytes, the steady state subcellular localization of all four mutants with nucleotide deletions is different from that of AQP2–E258K.
Since subcellular organelles of oocytes are poorly identified and oocytes are non-polarized cells, AQP2–727G was also studied in polarized MDCK cells. Co-localization studies in transfected MDCK cells revealed that, with or without forskolin stimulation, AQP2–727G mainly co-localized with Lamp1 (Fig. 6), whereas wt-AQP2 was localized in the apical membrane after stimulation and hardly co-localized with Lamp1 (Fig. 6). This indicated that AQP2–727G was mistargeted to late endosomes/lysosomes. Similarly, in forskolin-treated MDCK cells co-expressing AQP2–727G and wt-AQP2, both proteins predominantly co-localized with each other and Lamp1 (Fig. 7), which indicated that AQP2–727G is also able to hetero-oligomerize (and presumably heterotetramerize) with wt-AQP2 in MDCK cells, resulting in a misrouting of the complex to late endosomes/lysosomes. Since AQP2 is expressed as tetramers in MDCK cells and in vivo (15) and the AVP-regulated shuttling of wt-AQP2 in MDCK cells is similar to that described for renal collecting duct cells (18), our data indicate that the retention of AQP2/AQP2–727G hetero-oligomers in late endosomes/lysosomes is fundamental to dominant NDI in this particular family. If the three AQP2 gene mutations identified by Kuwahara et al. (16) result in identical C-terminal tails to that of AQP2–727G, it is likely that late endosomal/lysosomal mistargeting of AQP2 also explains dominant NDI in these families.
Localization of AQP2–727G to late endosomes/lysosomes
Expressed in MDCK cells, AQP2–727G mainly localized in the basolateral membrane and in late endosomes/lysosomes (Fig. 7). Other and weaker signals for AQP2–727G did not localize to the ER, to the cis-median or trans-Golgi complex or to VAMP2 positive vesicles, but may be undefined vesicles en route to late endosomes/lysosomes. Unfortunately, electron microscopic analysis of these cells could not address this issue, because no signal was obtained, which might be due to a low avidity of our AQP2–727G antibodies for their epitope. Upon co-expression with wt-AQP2, the basolateral localization was much less pronounced, which might be due to competition between targeting signals in the C-tails of wt-AQP2 and AQP2–727G monomers, because the steady state localization of a protein complex is determined by the hierarchy of sorting information contained within such a complex (29). At present, however, it is unknown whether the predominant steady-state expression of AQP2–727G (complexed or not with wt-AQP2) in late endosomes/lysosomes is a reflection of a defective delivery to the plasma membrane, altered endocytosis from the plasma membrane or an altered recycling. To resolve this issue, the targeting of AQP2–727G needs to be determined.
Although it is at present unclear which determinants in AQP2–727G cause its main retention in late endosomes/lysosomes, it is conceivable that this is caused by (i) the absence of motifs present in the wt-AQP2 tail, (ii) determinants present in the changed and/or extended tail of the mutant or (iii) a conformational change in the mutant protein resulting in the lysosomal retention as part of an intracellular protein quality control system.
(i) Of the wt-AQP2 segment missing in AQP2–727G, only Ser256, which is phosphorylated by PKA, has been shown to be essential for AQP2 redistribution from vesicles to the plasma membrane (24,25). Consistent with the absence of S256, the subcellular localization of AQP2–727G was not affected upon incubation with forskolin. In contrast to AQP2–727G, however, non-phosphorylated AQP2 is thought to reside in endosomal recycling vesicles rather than late endosomes/lysosomes (24,30) and does not co-localize with Lamp1 in unstimulated wt10 cells (P.Savelkoul and P.M.T.D., unpublished data). In this respect, another AQP2 mutant in dominant NDI (AQP2 812–818del) might be informative, because in this mutant the C-terminal tail as found in AQP2–727G only starts right after the last amino acid of AQP2 (Ala271) and therefore contains the entire wt-AQP2 segment missing in AQP2–727G (16). Similar to AQP2–727G, AQP2 812–818del was retained in oocytes, whereas wt-AQP2 missing Ala271 was only expressed in the plasma membrane (16). Although the subcellular localization of AQP2 812–818del in polarized cells remains to be determined, its retention in oocytes seems to indicate that the parts of wt-AQP2 that are missing in AQP2–727G do not cause its retention in late endosomes/lysosomes.
(ii) Alternatively, the added C-tail in AQP2–727G might be fundamental to its localization in late endosomes/lysosomes. In MDCK cells, several late endosomal/lysosomal proteins, such as Lamp1 and lysosomal acid phosphatase I (LAPI), are partly targeted from the TGN to the basolateral membrane, endocytosed and subsequently routed to late endosomes and lysosomes (31–33). Based on its steady-state subcellular localization (Fig. 6), this pathway might also be applicable to AQP2–727G. Besides these proteins, many agonist-induced G-protein coupled receptors are sorted to lysosomes after which they are degraded (34,35). Distinct lysosomal targeting motifs have been identified in the C-terminal tail of many of these proteins (32,34–38). In addition, the mannose-6 phosphate signal is a well-known targeting motif for lysosomal enzymes (39). Careful examination of the C-tails of AQP2–727G and the other nucleotide deletion mutants in dominant NDI (16), however, did not reveal any of these specific targeting motifs. Moreover, the inability of a co-expressed mutant C-terminal tail to restore the routing of AQP2–721delG to the plasma membrane of oocytes (16) also renders this explanation unlikely.
(iii) In addition to AQP2–727G, mistargeting to lysosomes in diseases has also been reported for a mutant lipoprotein lipase (LPL) in hyperlipoproteinemia and connexin 32 gap junctional protein in X-linked Charcot–Marie–Tooth disease (40,41). Since the mutations did not introduce a known lysosomal targeting motif in these proteins, it was speculated that the lysosomal targeting of the LPL mutant might be a consequence of an intracellular protein quality control system that would ensure the viability of the cell (40). Indeed, progressive aggregation of the luminal portion of furin in the TGN appeared to promote its targeting to lysosomes instead of secretory granules, which was suggested to be triggered by a conformational change (42).
Although the last explanation might be most likely, it remains to be determined which, if any, of these mechanisms is responsible for the late endosomal/lysosomal expression of AQP2–727G.
In conclusion, a new family with dominant NDI was presented in which the disease was caused by a novel mutation leading to an AQP2 mutant with an altered and extended intracellular C-terminal tail. This AQP2 mutant forms heterooligomers with wt-AQP2 and interferes the routing of wt-AQP2 to the apical plasma membrane by its mistargeting to late endosomes/lysosomes, thereby providing an explanation for NDI in this particular family. This study visualizes for the first time a mis-localization of a wt-protein to late endosomes/lysosomes in polarized cells after oligomerization with a mutant protein. Together with functional analysis of other AQP2 mutants in dominant NDI, our analyses furthermore reveal that an impaired routing, instead of function, is a general mechanism for the ‘loss of function’ phenotype in dominant NDI. Other plasma membrane channels, which are involved in recessive forms of diseases, have also been shown to be expressed as homotetramers (e.g. ROMK1, KIR6.2) (43–45). It is likely that also dominant forms of inheritance for these diseases will be identified, in which the mutant protein impairs the routing of the wild-type protein.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Patients
The female proband (852) is 10 years of age and has a documented lifelong history of polydipsia associated with hyposthenuric polyuria and elevated plasma concentrations of active renin and AVP. She is dystrophic with respect to weight [standard deviation score (SDS) of –1.95] and height (SDS of –2.76). Under current treatment with hydrochlorothiazide (0.5 mg/kg/day), daily urinary output is
3 l (6 ml/kg/h) with a creatinin clearance of 37 ml/min/1.73 m2. The urine and serum osmolalities are 112 mOsm/kg and 311 mOsm/kg, respectively. Further pathological laboratory signs are a hypokalemic alkalosis, moderate hypernatremia, hypercalcaemia and elevated levels of uric acid. The patient’s father suffered from severe NDI (fluid intake up to 12 l/day) leading to megacystis and bilateral mega-ureters. He finally developed end stage renal failure and underwent renal transplantation with a normal allograft function to date. The father’s mother was also reported to suffer from polyuria. The patient’s mother and brother are healthy. Infusion of dDAVP was performed as described previously (46) (0.3 µg/kg of body weight infused in 30 min).
Genetic analysis of the NDI family and cloning of AQP2–727G
The V2R genes of the patient and her father were sequenced as described previously (47). The AQP2 gene, which consists of four exons, of all family members was amplified using a forward primer 5' to the coding region of exon 1 (5'-GCG AGA GCG AGT GCC CG-3') and a reverse primer 3' to the coding region of exon 4 (5'-GCG GCC CTCA GGC CT-3'). With the generated AQP2 gene PCR fragment, the coding regions of exons 1–4 were amplified using flanking primers and manually sequenced (48). Haplotype analysis was carried out using AQP2 flanking markers (49). To obtain the proper coding region 3' of the mutation, the 3'-UTR of the AQP2 gene of the patient and of a non-affected person was amplified in three independent PCR reactions using the forward primer (5'-GTG GCC GAC GAG ATA CTT AGG-3') and the reverse primer (5'-GGT GCA GGC TCT CCA GAG-3'), which were designed according to the published human AQP2 sequence (19). Products from each PCR reaction were cloned into the EcoRV site of pBluescript-KS and transformed to bacteria. From each transformation, two separate clones were subjected to sequence analysis. For amplification and sequencing of the 3'-UTR of independent individuals, sense primer Aq11 (5'-C CTC TAC AAC TAC GTG CTG TTT CCG C-3') and antisense primer Aq13 (5'-TGCTCACAGCACTGGCCTG-3') were used.
To generate a construct driving the expression of AQP2–727G in oocytes, a NarI/SpeI fragment from an amplified AQP2 exon 4 segment from the patient, containing the 727G mutation and the 3'-UTR including the new stop codon, was cloned into the corresponding sites of the wt-AQP2 expression construct, pT7TsAQP2 (1), and named pT7TsAQP2–727G. To tag the AQP2 mutant, a DNA fragment coding for the vesicular stomatitis virus G protein (VSV-G) tag (EIYTDIEMNRLGK) was cloned into pT7TsAQP2–727G as described for pT7TsVSV-AQP2 (15). With this cloning step, a fusion protein was encoded in which the VSV tag was connected via a threonine residue to the N-terminus of AQP2–727G. For clarity, the N-terminal tags VSV and FLAG will be abbreviated to V- and F-, respectively.
In vitro transcription
The oocyte expression constructs encoding N-terminally tagged or untagged wt-AQP2 (pT7TsAQP2, pT7TsV-AQP2, pT7TsF-AQP2) or a mutant AQP2 in recessive NDI (pT7TsAQP2–R187C pT7TsV-AQP2–R187C) were as described previously (1,10,15). For transcription, these constructs, pT7TsAQP2–727G and pT7TsV-AQP2–727G, were linearized with SalI. Synthesis and purification of G-capped cRNAs and the subsequent analyses of their integrity and concentration were performed as described previously (15).
Water permeability measurements and membrane isolations
Isolation and de-folliculation of Xenopus oocytes was performed as described previously (15). The oocytes were injected with 10 ng cRNA encoding V-AQP2–727G or 0.3 ng cRNA encoding V-AQP2, V-AQP2–727G or AQP2–R187C only or in combination. Thus, with co-injections a total of 0.6 ng cRNA was injected. Two days after injection, vitellin membranes were removed and the water permeabilities were measured using a standard swelling assay (1). Statistical significance was determined using the Student’s t-test.
Isolation of total membranes (15) or plasma membranes (50) were performed as described previously. For each sample, 12 oocytes were taken. Digestion of proteins with N-glycosidase F (Boehringer Mannheim, Mannheim, Germany) was performed according to the manufacturer’s protocol.
Antibodies
To generate AQP2–727G specific antibodies, a 350 bp blunted KpnI/EcoRI fragment of pT7TsAQP2–727G encoding the last 62 amino acids of AQP2–727G, was cloned in frame into the blunted BamHI/EcoRI sites of pGEX-1X, downstream of the glutathione S-transferase coding sequence, and transformed into Escherichia coli BL-21. After IPTG induction, high amounts of the encoded fusion protein were obtained and isolated. Guinea pigs were primed with 200 µg of the fusion protein in complete Freund’s adjuvant. After 3 weeks, the animals were boosted with 100 µg of fusion protein in incomplete adjuvant, which was repeated until the titre of serum reached a plateau phase as verified by ELISA assays. To obtain affinity-purified antibodies, the serum was first passed over a gluthathione S-transferase coupled Affigel 15 column (Pharmacia Biotech, Uppsala, Sweden), followed by a column on which the fusion protein was immobilized. Antibodies were eluted with 0.1 M glycine (pH 2.8) and immediately neutralized.
Determination of the oligomerization state of AQP2–727G
Sixty oocytes were each injected with 10 ng cRNA coding for AQP2–727G, or 1 ng cRNA coding for wt-AQP2 or AQP2–R187C. After 2 days, total membranes were isolated and solubilized as described previously (15). Subsequently, the membrane proteins were subjected to sucrose gradient centrifugation; fractions were taken and then analysed by immunoblotting as reported previously (15). As sedimentation markers, a mixture of trypsine (28 kDa), BSA (67 kDa), phosphorylase B (97 kDa) and yeast alcohol dehydrogenase (150 kDa) were used. With these assays, the molecular mass of the studied protein is indicated by its peak fraction.
Co-immunoprecipitation
Binding of monoclonal FLAG antibodies (M2; Sigma, St Louis, MO) to Protein G-agarose beads (Pharmacia Biotech) was performed as described previously (15). Then, of 30 oocytes expressing F-AQP2, V-AQP2–727G or V-AQP2–R187C only or in combination, total membranes were isolated and solubilized as described previously (15). Following ultracentrifugation and thus removal of undissolved membranes, 300 µl of supernatant was diluted to 1200 µl with IPP100 (15) and rotated overnight with the protein G beads coupled to 
-FLAG antibodies. After five washes, proteins bound to the beads were dissolved in Laemmli buffer and subjected to immunoblotting.
Immunoblotting
Protein denaturation, SDS–PAGE and blotting were performed as described previously (15). The obtained blots were subsequently incubated with 1:3000 diluted affinity-purified AQP2: 257–271 rabbit antibodies, raised against the 15 C-terminal amino acids of rat AQP2 (-AQP2) (10), 1:1000 diluted guinea pigs -AQP2–727G or with 1:2000 diluted -VSV antibody (P5D4-ascitus) (51), in TBST-buffer (20 mM Tris, 140 mM NaCl, 0.1% Tween pH 7.6) supplemented with 1% non-fat dried milk. As secondary antibodies, a 1:5000 dilution of goat anti-rabbit IgG (Sigma), a 1:2500 dilution of goat anti-guinea pigs IgG (Jackson Immunoresearch, West Grove, PA) or 1:2000 sheep anti-mouse IgG (Sigma), all coupled to horseradish peroxidase were used. AQP2 proteins were visualized using enhanced chemiluminescence (Pierce, Rockford, IL).
Transfection of MDCK cells
To study AQP2–727G in polarized renal epithelial cells, the AQP2–727G-encoding cDNA fragment was cut from pT7Ts-AQP2–727G using BglII and SpeI, and cloned into the BglII and XbaI sites of the eukaryotic expression vector pCB6. After verifying the correct orientation, 25 µg of the purified expression construct was transfected into native MDCK type I cells (52) or MDCK cells stably expressing human AQP2 (wt10) (18) using the calcium-phosphate precipitation technique as described previously (53). Selection of G418-resistant clones was performed as described (26).
Immunocytochemistry
Two days after injection of oocytes with cRNA coding for AQP2–727G or wt-AQP2, vitellin membranes were removed, after which the oocytes were fixed, paraffin-embedded and sectioned as described previously (15). Stably MDCK cells and non-transfected control cells were grown to confluence on filters for 3 days and treated with 1 x 10–5 M forskolin for 45 min to induce apical localization of wt-AQP2. Next, the cells were fixed and prepared for immunocytochemistry as described previously (26). Filters with MDCK cells and oocyte sections were blocked with normal goat serum dilution buffer (GSDB) (26), washed and incubated overnight in GSDB containing 1:25 diluted guinea pig -AQP2–727G antibodies, 1:100 diluted monoclonal antibodies recognizing the late endosomal/lysosomal marker protein Lamp1 (-Lamp1; kindly provided by A.le Bivic) (31) or 1:100 diluted rabbit -AQP2 antibodies only or combined. After three washing steps of 10 min with permeabilization buffer (PB) (26), the cells were incubated for 1 h in GSDB containing 1:200 diluted Alexa 594-conjugated -guinea pig, 1:200 diluted Alexa 488-conjugated -mouse or 1:100 diluted Alexa 594-conjugated -rabbit antibodies (all from Molecular Probes, Pitchford, Eugene, OR) only or combined. For triple labelling, 1:200 diluted Cy5-conjugated -mouse antibodies (Jackson Immunoresearch), 1:100 diluted FITC-conjugated -rabbit antibodies (1:100; Sigma) and 1:200 Alexa 594-conjugated -guinea pig antibodies were used. Subsequently, the sections were washed three times for 10 min in PB and mounted in Vectashield (Vector Laboratories, Berlingame, CA). The subcellular localization of the different proteins was analysed by confocal laser scanning microscopy (Bio-Rad MRC-1000) using a 60x oil-immersion objective, a 32 Kalman collection filter, an aperture diaphragm of 2.8 and an axial resolution of 0.4 µm per pixel.
|
ACKNOWLEDGEMENTS |
|---|
This study was supported by grants from the Dutch Kidney Foundation (C95.5001) and European Community TMR (FMRX-CT97-0128) (P.M.T.D., C.H.v.O.), European Community Biotech (BFO4-CT98-0024) (P.M.T.D.), the Verbund Klinische Pharmakologie von Berlin-Brandenburg and the Chemischen Industrie (both to W.R.), the Canadian Institutes of Health Research (MOP-8126) and the kidney foundation of Canada (both D.G.B.). P.M.T.D. is an investigator of the Royal Netherlands Academy of Arts and Sciences.
|
FOOTNOTES |
|---|
+ To whom correspondence should be addressed. Tel: +31 24 361 7347; Fax: +31 24 361 6413; Email: peterd@sci.kun.nl
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
1 Deen,P.M.T., Verdijk,M.A.J., Knoers,N.V.A.M., Wieringa,B., Monnens,L.A.H., van Os,C.H. and van Oost,B.A. (1994) Requirement of human renal water channel aquaporin-2 for vasopressin-dependent concentration of urine. Science, 264, 92–95.
2
Ma,T., Song,Y., Yang,B., Gillespie,A., Carlson,E.J., Epstein,C.J. and Verkman,A.S. (2000) Nephrogenic diabetes insipidus in mice lacking aquaporin-3 water channels. Proc. Natl Acad. Sci. USA, 97, 4386–4391.
3 Chou,C.L., Ma,T.H., Yang,B.X., Knepper,M.A. and Verkman,S. (1998) Fourfold reduction of water permeability in inner medullary collecting duct of aquaporin-4 knockout mice. Am. J. Physiol. Renal Physiol., 43, C549–C554.
4 van Os,C.H. and Deen,P.M.T. (1998) Aquaporin-2 water channel mutations causing nephrogenic diabetes insipidus. Proc. Assoc. Am. Physicians, 110, 395–400.[ISI][Medline]
5
Nielsen,S., Chou,C.L., Marples,D., Christensen,E.I., Kishore,B.K. and Knepper,M.A. (1995) Vasopressin increases water permeability of kidney collecting duct by inducing translocation of aquaporin-CD water channels to plasma membrane. Proc. Natl Acad. Sci. USA, 92, 1013–1017.
6
van Lieburg,A.F., Knoers,N.V. and Monnens,L.A. (1999) Clinical presentation and follow-up of 30 patients with congenital nephrogenic diabetes insipidus. J. Am. Soc. Nephrol., 10, 1958–1964.
7 Nielsen,S., Frokiaer,J. and Knepper,M.A. (1998) Renal aquaporins: key roles in water balance and water balance disorders. Curr. Opin. Nephrol. Hypertens., 7, 509–516.[ISI][Medline]
8
Arthus,M.F., Lonergan,M., Crumley,M.J., Naumova,A.K., Morin,D., De Marco,L.A., Kaplan,B.S., Robertson,G.L., Sasaki,S., Morgan,K., Bichet,D.G. and Fujiwara,T.M. (2000) Report of 33 novel AVPR2 mutations and analysis of 117 families with X-linked nephrogenic diabetes insipidus. J. Am. Soc. Nephrol., 11, 1044–1054.
9 Van Lieburg,A.F., Verdijk,M.A.J., Knoers,N.V.A.M., van Essen,A.J., Proesmans,W., Mallmann,R., Monnens,L.A.H., van Oost,B.A., van Os,C.H. and Deen,P.M.T. (1994) Patients with autosomal nephrogenic diabetes insipidus homozygous for mutations in the aquaporin 2 water-channel gene. Am. J. Hum. Genet., 55, 648–652.[ISI][Medline]
10 Deen,P.M.T., Croes,H., van Aubel,R.A., Ginsel,L.A. and van Os,C.H. (1995) Water channels encoded by mutant aquaporin-2 genes in nephrogenic diabetes insipidus are impaired in their cellular routing. J. Clin. Invest., 95, 2291–2296.
11 Mulders,S.M., Knoers,N.V.A.M., van Lieburg,A.F., Monnens,L.A.H., Leumann,E., Wuhl,E., Schober,E., Rijss,J.P.L., van Os,C.H. and Deen,P.M.T. (1997) New mutations in the AQP2 gene in nephrogenic diabetes insipidus resulting in functional but misrouted water channels. J. Am. Soc. Nephrol., 8, 242–248.[Abstract]
12 Tamarappoo,B.K. and Verkman,A.S. (1998) Defective aquaporin-2 trafficking in nephrogenic diabetes insipidus and correction by chemical chaperones. J. Clin. Invest., 101, 2257–2267.[ISI][Medline]
13 Marr,N., Kamsteeg,E.J., Van Raak,M., van Os,C.H. and Deen,P.M. (2001) Functionality of aquaporin-2 missense mutants in recessive nephrogenic diabetes insipidus. Pflugers Arch., 442, 73–77.[ISI][Medline]
14 Mulders,S.M., Bichet,D.G., Rijss,J.P.L., Kamsteeg,E.J., Arthus,M.F., Lonergan,M., Fujiwara,M., Morgan,K., Leijendekker,R., van der Sluijs,P., van Os,C.H. and Deen,P.M.T. (1998) An aquaporin-2 water channel mutant which causes autosomal dominant nephrogenic diabetes insipidus is retained in the Golgi complex. J. Clin. Invest., 102, 57–66.[ISI][Medline]
15 Kamsteeg,E.J., Wormhoudt,T.A., Rijss,J.P.L., van Os,C.H. and Deen,P.M.T. (1999) An impaired routing of wild-type aquaporin-2 after tetramerization with an aquaporin-2 mutant explains dominant nephrogenic diabetes insipidus. EMBO J., 18, 2394–2400.[ISI][Medline]
16 Kuwahara,M., Iwai,K., Ooeda,T., Igarashi,T., Ogawa,E., Katsushima,Y., Shinbo,I., Uchida,S., Terada,Y., Arthus,M.F. et al. (2001) Three families with autosomal dominant nephrogenic diabetes insipidus caused by aquaporin-2 mutations in the C-terminus. Am. J. Hum.Genet., 69, 738–748.[ISI][Medline]
17
Kamsteeg,E.J. and Deen,P.M.T. (2000) Importance of aquaporin-2 expression levels in genotype-phenotype studies in nephrogenic diabetes insipidus. Am. J. Physiol. Renal Physiol., 279, F778–F784.
18 Deen,P.M.T., Rijss,J.P.L., Mulders,S.M., Errington,R.J., van Baal,J. and van Os,C.H. (1997) Aquaporin-2 transfection of Madin-Darby canine kidney cells reconstitutes vasopressin-regulated transcellular osmotic water transport. J. Am. Soc. Nephrol., 8, 1493–1501.[Abstract]
19
Uchida,S., Sasaki,S., Fushimi,K. and Marumo,F. (1994) Isolation of human aquaporin-CD gene. J. Biol. Chem., 269, 23451–23455.
20
Kim,P.S. and Arvan,P. (1998) Endocrinopathies in the family of endoplasmic reticulum (er) storage diseases: disorders of protein trafficking and the role of er molecular chaperones. Endocr. Rev., 19, 173–202.
21 Hobbs,H.H., Brown,M.S. and Goldstein,J.L. (1992) Molecular genetics of the LDL receptor gene in familial hypercholesterolemia. Hum. Mutat., 1, 445–466.[Medline]
22 Murata,K., Mitsuoka,K., Hirai,T., Walz,T., Agre,P., Heymann,J.B., Engel,A. and Fujiyoshi,Y. (2000) Structural determinants of water permeation through aquaporin-1. Nature, 407, 599–605.[Medline]
23
Fu,D., Libson,A., Miercke,L.J., Weitzman,C., Nollert,P., Krucinski,J. and Stroud,R.M. (2000) Structure of a glycerol-conducting channel and the basis for its selectivity. Science, 290, 481–486.
24 Katsura,T., Gustafson,C.E., Ausiello,D.A. and Brown,D. (1997) Protein kinase A phosphorylation is involved in regulated exocytosis of aquaporin-2 in transfected LLC-PK1 cells. Am. J. Physiol., 41, F816–F822.
25
Kamsteeg,E.J., Heijnen,I., van Os,C.H. and Deen,P.M.T. (2000) The subcellular localization of an squaporin-2 tetramer depends on the stoichiometry of phosphorylated and nonphosphorylated monomers. J. Cell Biol., 151, 919–930.
26
Deen,P.M.T., Van Balkom,B., Savelkoul,P.J.M., Kamsteeg,E.J., Van Raak,M., Jennings,M.L., Muth,T.R., Rajendran,V. and Caplan,M.J. (2002) Aquaporin-2: COOH terminus is necessary but not sufficient for routing to the apical membrane. Am. J. Physiol. Renal Physiol., 282, F330–F340
27 Ikonen,E. and Simons,K. (1998) Protein and lipid sorting from the trans-Golgi network to the plasma membrane in polarized cells. Semin. Cell Dev. Biol., 9, 503–509.[ISI][Medline]
28 Brown,D. and Breton,S. (2000) Sorting proteins to their target membranes. Kidney Int., 57, 816–824.[ISI][Medline]
29 Lisanti,M.P. and Rodriguez-Boulan,E. (1990) Glycophospholipid membrane anchoring provides clues to the mechanism of protein sorting in polarized epithelial cells. Trends Biochem. Sci., 15, 113–118.[ISI][Medline]
30
Christensen,B.M., Zelenina,M., Aperia,A. and Nielsen,S. (2000) Localization and regulation of PKA-phosphorylated AQP2 in response to V(2)-receptor agonist/antagonist treatment. Am. J. Physiol. Renal Physiol., 278, F29–F42.
31
Nabi,I.R., Le Bivic,A., Fambrough,D. and Rodriguez-Boulan,E. (1991) An endogenous MDCK lysosomal membrane glycoprotein is targeted basolaterally before delivery to lysosomes. J. Cell Biol., 115, 1573–1584.
32
Honing,S. and Hunziker,W. (1995) Cytoplasmic determinants involved in direct lysosomal sorting, endocytosis and basolateral targeting of rat lgp120 (lamp-I) in MDCK cells. J. Cell Biol., 128, 321–332.
33 Bermingham,N.A., Martin,J.E. and Fisher,E.M. (1996) The mouse lysosomal membrane protein 1 gene as a candidate for the motorneuron degeneration (mnd) locus. Genomics, 32, 266–271.[ISI][Medline]
34 Moore,R.H., Tuffaha,A., Millman,E.E., Dai,W., Hall,H.S., Dickey,B.F. and Knoll,B.J. (1999) Agonist-induced sorting of human ß2-adrenergic receptors to lysosomes during downregulation. J. Cell Sci., 112, 329–338.[Abstract]
35
Trejo,J. and Coughlin,S.R. (1999) The cytoplasmic tails of protease-activated receptor-1 and substance P receptor specify sorting to lysosomes versus recycling. J. Biol. Chem., 274, 2216–2224.
36
Haft,C.R., de la Luz,S.M., Hamer,I., Carpentier,J.L. and Taylor,S.I. (1998) Analysis of the juxtamembrane dileucine motif in the insulin receptor. Endocrinology, 139, 1618–1629.
37 Prill,V., Lehmann,L., von Figura,K. and Peters,C. (1993) The cytoplasmic tail of lysosomal acid phosphatase contains overlapping but distinct signals for basolateral sorting and rapid internalization in polarized MDCK cells. EMBO J., 12, 2181–2193.[ISI]