Three families with autosomal dominant nephrogenic diabetes insipidus caused by aquaporin-2 mutations in the C-terminus.

The vasopressin-regulated water channel aquaporin-2 (AQP2) is known to tetramerize in the apical membrane of the renal tubular cells and contributes to urine concentration. We identified three novel mutations, each in a single allele of exon 4 of the AQP2 gene, in three families showing autosomal dominant nephrogenic diabetes insipidus (NDI). These mutations were found in the C-terminus of AQP2: a deletion of G at nucleotide 721 (721 delG), a deletion of 10 nucleotides starting at nucleotide 763 (763-772del), and a deletion of 7 nucleotides starting at nucleotide 812 (812-818del). The wild-type AQP2 is predicted to be a 271-amino acid protein, whereas these mutant genes are predicted to encode proteins that are 330-333 amino acids in length, because of the frameshift mutations. Interestingly, these three mutant AQP2s shared the same C-terminal tail of 61 amino acids. In Xenopus oocytes injected with mutant AQP2 cRNAs, the osmotic water permeability (Pf) was much smaller than that of oocytes with the AQP2 wild-type (14%-17%). Immunoblot analysis of the lysates of the oocytes expressing the mutant AQP2s detected a band at 34 kD, whereas the immunoblot of the plasma-membrane fractions of the oocytes and immunocytochemistry failed to show a significant surface expression, suggesting a defect in trafficking of these mutant proteins. Furthermore, coinjection of wild-type cRNAs with mutant cRNAs markedly decreased the oocyte Pf in parallel with the surface expression of the wild-type AQP2. Immunoprecipitation with antibodies against wild-type and mutant AQP2 indicated the formation of mixed oligomers composed of wild-type and mutant AQP2 monomers. Our results suggest that the trafficking of mutant AQP2 is impaired because of elongation of the C-terminal tail, and the dominant-negative effect is attributed to oligomerization of the wild-type and mutant AQP2s. Segregation of the mutations in the C-terminus of AQP2 with dominant-type NDI underlies the importance of this domain in the intracellular trafficking of AQP2.

Functional characterization of a water channel of the nematode Caenorhabditis elegans.

A genome project for the species Caenorhabditis elegans has demonstrated the presence of eight cDNAs belonging to the major intrinsic protein (MIP) family. We previously characterized one of these cDNAs known as C01G6.1. C01G6.1 was confirmed to be a water channel and newly designated as AQP-CE1 [Am. J. Physiol. 275 (1998) C1459-C1464]. In this paper, we examined the function of another MIP protein encoded by F40F9.9. This cDNA encodes a 274-amino acid protein showing a high sequence identity with mammalian aquaporin-8 (AQP8) water channel (35%) and d-TIP (34%), an AQP of Arabidopsis. The expression of F40F9.9 in Xenopus oocytes increased the osmotic water permeability (P(f)) 10.4-fold, and the activation energy for P(f) from Arrhenius plot was 4.7 kcal/mol, suggesting that F40F9.9 is a water channel (AQP-CE2). AQP-CE2 was not permeable to glycerol or urea. Oocyte P(f) was reversibly inhibited by 58% after an incubation with 0.3 mM HgCl(2). To identify the mercury-sensitive site, four individual cysteine residues in AQP-CE2 (at positions 47, 132, 149, 259) were altered to serine by site-directed mutagenesis. Of these mutants, only C132S had a P(f) similar to that of the wild-type together with an acquired mercury resistance, suggesting that Cys-132 is the mercury-sensitive site. Similar results were obtained by the mutation of Cys-132 to alanine (C132A). Replacement of Cys-132 with tryptophan decreased P(f) by 64%, but P(f) was still 2.5 times higher than that of the control. Cys-132 is located in the transmembrane helix 3, close to the transition to the extracellular loop C. These results suggest that the transmembrane helix 3, including Cys-132, might participate in the aqueous pore formation, or, alternatively, that Cys-132 might contribute to the construction of the AQP protein.

Stimulation of Cl(-) secretion by L-alanine oligopeptides in the mammalian large intestine.

A variety of oligopeptides are probably released within the intestinal tissue under inflammatory conditions or during peptide absorption. To examine whether some of these peptides can affect intestinal transport functions, we determined the effects of L-alanine oligopeptide on short-circuit current (I(sc)) and transmucosal conductance (G(t)) in submucosa-mucosa preparations from the mouse cecum and guinea pig distal colon in vitro in Ussing chambers. L-Alanyl-L-alanine (Ala-Ala, 10 mM) added to the serosal side increased I(sc) and G(t), giving a peak followed by a sustained phase (the peak increase in I(sc) was 45 +/-6 microA/cm(2) and the increase in G(t) was 0.55+/-0.11 mS/cm(2)). The tripeptide, L-alanyl-L-alanyl-L-alanine (Ala-Ala-Ala, 10 mM), added to the serosal side also induced increases in I(sc) and G(t) by a similar degree. On the other hand, luminal Ala-Ala, and serosal L-alanine and L-alanine (10 mM) caused significantly smaller increases in I(sc) and G(t) ( approximately 15 microA/cm(2) and approximately 0. 15 mS/cm(2), respectively). The Ala-Ala induced increase in I(sc) was partially inhibited by serosal bumetanide (0.1 mM) and mucosal 5-nitro-2-(3-phenylpropylamino)benzoic acid (0.1 mM), and largely suppressed by removing Cl(-) from the bathing solution. The increase in I(sc) was largely suppressed by serosal low Ca(2+) and tetrodotoxin, but was not affected by indomethacin. In the guinea pig distal colon, serosal Ala-Ala (10 mM) evoked a transient increase in I(sc) by 23+/-7 microA/cm(2) and an increase in G(t) by 1.2+/-0.3 mS/cm(2). These results suggest that Ala-Ala, and probably also Ala-Ala-Ala, added to the serosal side stimulated electrogenic Cl(-) secretion mainly through the activation of submucosal secretomotor neurons in the mammalian large intestine.

Transmembrane helix 5 is critical for the high water permeability of aquaporin.

Aquaporin-2 (AQP2), a vasopressin-regulated water channel, plays a major role in urinary concentration. AQP2 and the major intrinsic protein (MIP) of lens fiber are highly homologous (58% amino acid identity) and share a topology of six transmembrane helices connected by five loops (loops A-E). Despite the similarities of these proteins, however, the water channel activity of AQP2 is much higher than that of MIP. To determine the site responsible for this gain of activity in AQP2, several parts of MIP were replaced with the corresponding parts of AQP2. When expressed in Xenopus oocytes, the osmotic water permeability (P(f)) of MIP and AQP2 was 48 and 245 x 10(-)(4) cm/s, respectively. Substitutions in loops B-D failed to increase P(f), whereas substitution of loop E significantly increased P(f) 1.5-fold. A similar increase in P(f) was observed with the substitution of the front half of loop E. P(f) measurements taken in a yeast vesicle expression system also confirmed that loop E had a complementary effect, whereas loops B-D did not. However, P(f) values of the loop E chimeras were only approximately 30% of that of AQP2. Simultaneous exchanges of loop E and a distal half of transmembrane helix 5 just proximal to loop E increased P(f) to the level of that of AQP2. Replacement of helix 5 alone stimulated P(f) 2.7-fold. Conversely, P(f) was decreased by 73% when helix 5 of AQP2 was replaced with that of MIP. Moreover, P(f) was stimulated 2.6- and 3.3-fold after helix 5 of AQP1 and AQP4 was spliced into MIP, respectively. Our findings suggested that the distal half of helix 5 is necessary for maximum water channel activity in AQP. We speculate that this portion contributes to the formation of the aqueous pore and the determination of the flux rate.

Functional analysis of aquaporin-2 mutants associated with nephrogenic diabetes insipidus by yeast expression.

Mutations of aquaporin-2 (AQP2) vasopressin water channel cause nephrogenic diabetes insipidus (NDI). It has been suggested that impaired routing of AQP2 mutants to the plasma membrane causes the disease; however, no determinations have been made of mutation-induced alterations of AQP2 channel water permeability. To address this issue, a series of AQP2 mutants were expressed in yeast, and the osmotic water permeability (P(f)) of the isolated vesicles was measured. Wild-type and mutant AQP2 were expressed equally well in vesicles. P(f) of the vesicles containing wild-type AQP2 was 22 times greater than that of the control, which was sensitive to mercury and weakly dependent on the temperature. P(f) measurements and mercury inhibition examinations suggested that mutants L22V and P262L are fully functional, whereas mutants N68S, R187C, and S216P are partially functional. In contrast, mutants N123D, T125M, T126M, A147T, and C181W had very low water permeability. Our results suggest that the structure between the third and fifth hydrophilic loops is critical for the functional integrity of the AQP2 water channel and that disruption of AQP2 water permeability by mutations may cause NDI.

Effects of missense mutations on rat aquaporin-2 in LLC-PK1 porcine kidney cells.

BACKGROUND: Mutations in the aquaporin-2 (AQP2) gene have been found in families with nephrogenic diabetes insipidus (NDI), but the pathophysiological mechanisms of how mutant AQP2 causes the disease are still not clear. METHODS: Wild-type (WT) AQP2 and four mutants-T126M, A147T, R187C, and S216P-were transiently expressed in LLC-PK1 cells. The osmotic water permeability of LLC-PK1 cells expressing AQP2 mutants was determined by stopped-flow light-scattering microphotometry. Cell surface expression, subcellular localization, and effects of vasopressin stimulation were examined by surface biotin labeling and confocal immunohistochemistry. RESULTS: The osmotic water permeability (Pf) of cells expressing WT increased significantly after vasopressin treatment, whereas the Pf of cells expressing T126M A147T, R187C, and S216P was not significantly different from that of the control even after vasopressin stimulation. Confocal immunohistochemistry demonstrated distribution of WT and A147T in early/recycling endosomal compartments and vasopressin-responsive translocation and surface expression. In contrast, stainings of T126M, R187C, and S216P were similar to that of Grp78, indicating that these mutants were misassembled and retarded in the endoplasmic reticulum. CONCLUSION: Our results indicated that the intracellular distribution and vasopressin-regulated trafficking of A147T is intact, in contrast to the other three mutants, of which both were impaired. Thus, it is conceivable that the disruption of the AQP2 channel function accounts for the pathogenesis of A147T NDI, whereas trafficking defects account for that of the other types, suggesting that the pathophysiology of AQP2-related NDI is heterogeneous.

[Myelodysplastic syndrome associated with gastric cancer and colon polyp].

A 62-year-old man was found to have myelodysplastic syndrome (MDS) and gastric cancer. Serum studies revealed hypogammaglobulinemia, and positive reactions to rheumatoid test and anti-nuclear antibody. Chromosomal analysis of the bone marrow revealed deletions of No. 5 and 7 chromosomes and that of the stomach showed complex abnormalities. It may be hypothesized that an initial event which selects a clone of stem cells could manifest with this sort of immunological abnormalities. Subsequent deranged immunosurveillance may be responsible for the increased risk of cancer. Alternatively an increased chromosomal instability which seems to be associated with immunodeficiencies might be responsible for cancer development.