Both mammals and birds can concentrate urine hyperosmotic to plasma via a countercurrent multiplier mechanism, although evolutionary lines leading to mammals and birds diverged at an early stage of tetrapod evolution. We reported earlier (Nishimura H, Koseki C, and Patel TB. Am J Physiol Regul Integr Comp Physiol 271: R1535–R1543, 1996) that arginine vasotocin (AVT; avian antidiuretic hormone) increases diffusional water permeability in the isolated, perfused medullary collecting duct (CD) of the quail kidney. In the present study, we have identified an aquaporin (AQP) 2 homolog water channel in the medullary cones of Japanese quail, Coturnix coturnix (qAQP2), by RT-PCR-based cloning techniques. A full-length cDNA contains an 822-bp open reading frame that encodes a 274-amino acid sequence with 75.5% identity to rat AQP2. The qAQP2 has six transmembrane domains, two asparagine-proline-alanine (NPA) sequences, and putative N-glycosylation (asparagine-124) and phosphorylation sites (serine-257) for cAMP-dependent protein kinase. qAQP2 is expressed in the membrane of Xenopus laevis oocytes and significantly increased its osmotic water permeability (Pf), inhibitable (P < 0.01) by mercury chloride. qAQP2 mRNA (RT-PCR) was detected in the kidney; medullary mRNA levels were higher than cortical levels. qAQP2 protein that binds to rabbit anti-rat AQP2 antibody is present in the apical/subapical regions of both cortical and medullary CDs from normally hydrated quail, and the intensity of staining increased only in the medullary CDs after water deprivation or AVT treatment. The relative density of the ∼29-kDa protein band detected by immunoblot from the medullary cones was modestly higher in water-deprived/AVT-treated quail. The results suggest that 1) medullary CDs of quail kidneys express a mercury-sensitive functioning qAQP2 water channel, and 2) qAQP2 is at least partly regulated by an AVT-dependent mechanism. This is the first clear identification of AQP2 homolog in nonmammalian vertebrates.
- urine concentration
- water permeability
- arginine vasotocin
- avian aquaporin 2
- Coturnix coturnix
- medullary cone
- collecting duct
birds are the only nonmammalian vertebrates that can produce hyperosmotic urine by a countercurrent urine concentration mechanism. Avian kidneys contain loopless and looped nephrons. Loopless nephrons lack the Henle's loop, and their collecting tubules open into the collecting duct (CD) at a right angle. Loopless nephrons thus do not contribute directly to the formation of hyperosmotic urine. Looped nephrons have Henle's loop and run parallel to the CD (for review, see Refs. 5 and 8). Looped nephrons can produce concentrated urine by a countercurrent multiplier mechanism, utilizing recycling of a single solute (NaCl) (29). In general, osmolal urine-to-plasma (U/P) ratios are lower in birds than in mammals, and the maximum osmolal U/P ratio is 1.5–2.0 after dehydration (4). We previously reported that the medullary CD of Japanese quail, Coturnix coturnix, isolated and perfused in vitro, shows considerable basal diffusional water permeability that is increased only modestly by arginine vasotocin (AVT; avian antidiuretic hormone, ADH) (30). This relatively small effect of AVT is not due to limited availability of adenylate cyclase/cAMP because forskolin increases diffusional water permeability nearly 20-fold (30).
In mammalian kidneys, the aquaporin 2 (AQP2) water channel is selectively expressed in apical/subapical membranes of connecting tubules and cortical and medullary CDs (13, 42). Arginine vasopressin, via the cAMP/protein kinase A signal pathway, regulates trafficking of AQP2 and also markedly stimulates the production of AQP2 mRNA and protein (21, 33). We therefore hypothesize that bird kidneys are likely to have AVT-dependent and AVT-independent water channels in the medullary cones. We found previously that an immunoreactive AQP2 homolog (qAQP2) protein exists in the apical and subapical regions of medullary CDs from Coturnix quail (45). The molecular and functional properties of qAQP2, however, have not been identified. The purposes of this study, therefore, are 1) to characterize the molecular and functional properties of qAQP2, and 2) to investigate whether qAQP2 is regulated by AVT.
MATERIALS AND METHODS
Animals and maintenance.
Japanese quail, C. coturnix, were hatched in our laboratory (29, 30). Eggs were purchased from GQF Manufacturing (Savannah, GA) and incubated for 17 days. Hatched birds were fed Game Bird Startena (Purina Mills, St. Louis, MO; protein content, 30%; NaCl, 0.3–0.6%) for 2 wk, and then Chick Start and Grow (Purina; protein content, 17%; NaCl, 0.3–0.6%). Drinking water containing multiple vitamins (GQF Manufacturing) was given for the first 3 wk, followed by tap water ad libitum. Experiments were performed with quail of both sexes, 7–35 wk (all sexually mature) of age. The photoperiod (10:14-h light-dark cycle) was controlled. Adult female Xenopus laevis were purchased (NASCO; Ft. Atkinson, WI) and kept in an aquarium (filtered fresh water) in a cold room (15°C) and fed commercial pellets (NASCO).
Adult quail (35 wk of age) of both sexes were water deprived for 48 h to presumably stimulate production of AQP2 mRNA. The kidneys were quickly removed after decapitation of the quail and placed in chilled avian Ringer solution (in mM: 115 NaCl, 5.0 KCl, 25 NaHCO3, 0.5 NaH2PO4, 2.0 Na2HPO4, 10 Na acetate, 1.0 MgCl2, 1.9 CaCl2, 8.3 d-glucose, 5.0 l-alanine; pH 7.4; 312 mosmol/kgH2O). Medullary cones were isolated under a dissecting microscope and homogenized (Polytron, Brinkman; Westbury, NY). The total RNA was extracted using a Trizol reagent (GIBCO Life Technologies; Grand Island, NY) containing a monophasic solution of phenol and guanidine thiocyanate, followed by extraction with chloroform and isopropanol. The concentration of RNA was determined by measuring the absorbance at 260 nm. RNA quality was monitored by an A260:A280 ratio of 1.8-2.0 and by RNA electrophoresis on either 1.2% agarose (FMC Bioproducts; Rockland, ME)/2.2 M formaldehyde gel or 1.2% agarose/Tris-acetate/EDTA gel, scanning through the integrity of 28S and 18S ribosomal RNA bands.
First-strand cDNA was synthesized from total RNA with either random or oligo(dT) primers, and AQP-related sequences were amplified with an AdvanTag PCR Kit (Clontech Lab; Palo Alto, CA) with a pair of degenerate primers selected from two conserved regions of the MIP family [MIP-forward 5′-(T/C)TIAA(T/C)CCIGCIGTIAC-3′; MIP-reverse 5′-AAI(G/C)(T/A)IC(G/T)IGCIGG(A/G)T-3′] (13). These reactions yielded a fragment of the expected size (∼370 bp), which was subcloned into a pCR 2.1-vector (Invitrogen Life Tech; Carlsbad, CA) for sequencing. Individual clones of each PCR fragment were further amplified by PCR with T7 primer and M13 reverse primer. Clones yielding a PCR fragment of ∼570 bp (370 + 140 bp linker arms of the vector) were selected for sequencing. After identification of the cDNA segment according to known AQP sequences, the sequence was extended in the 3′ and 5′ directions with 5′-CATCCTACACGAGATCACCCCAGC-3′ and 5′-TGATCTCGTGTAGGATGGCAGCCC-3′, respectively, using rapid amplification of cDNA ends (RACE) kit (Smart RACE; Clontech). The RACE products were cloned into the pCR4-TOPO vector with a T/A cloning kit (Invitrogen) for sequencing. The full coding sequence was obtained from oligo(dT)-primed cDNA, using specific sense primer 5′-GGCAGGTCGACTCCTTGCAGCCTCCATG-3′ (engineered HincII restriction site is underlined) and antisense primer 5′-GAACAGCGGCCGCTCTCTCAGCTCCTCTCC-3′ (engineered NotII restriction site is underlined). Full cDNA was cloned into the pBluescript II KS(−) vector (Stratagene; La Jolla, CA).
Expression of qAQP2 in Xenopus oocytes and measurement of water permeability.
The quail AQP2 (qAQP2) cDNA was subcloned into the pBluescript II KS (−) vector (Stratagene; La Jolla, CA) at the NotI and HincII multicloning site. A Xenopus beta-globin 5′-untranslated region was substituted for the qAQP2 ATG initiation codon (32). The plasmid was linearized with NotI and transcribed/capped using T3 polymerase (mMESSAGE mMACHINE, Ambio, Austin, TX). After proteinase K treatment, phenol-chloroform extraction, and ethanol precipitation (−20°C, twice), the cRNA was dissolved in RNase-free water for oocyte injection. Human AQP1 (hAQP1) was used as a positive control; the hAQP1 was either a gift from Dr. Peter Agre's laboratory (Johns Hopkins School of Medicine; Baltimore, MD) or a Xenopus vector containing hAQP1 purchased from American Type Culture Collection (ATCC; Rockville, MD).
Stage V and VI oocytes from adult female Xenopus laevis (NASCO) were isolated and defolliculated by incubation with 2 mg/ml collagenase (Type IA, Sigma; St. Louis, MO) in Ca2+-free buffer for 60 min at room temperature (46) followed by 16 h incubation in 200 mosmol/kgH2O modified Barth's buffer at 18°C. Oocytes were injected with 50 nl water or 15 ng (in 50 nl) of cRNA and incubated at 18°C for 24, 48, 72, or 96 h. The 72-h incubation time yielded the maximum expression of AQP cRNA and was used for the subsequent experiments. Oocytes were transferred to a superfusion chamber (2 ml/min, 22°C); after equilibration, the perfusate was changed to 70 mosmol/kgH2O Barth's buffer (diluted), and the time-course increase in volume (area) was examined (every 30 s for 5 min; N = 7–16/group). To determine whether qAQP2 is susceptible to mercury, oocytes were exposed to HgCl2 (0.3 mM) for 5 min before initiation of the time-course study. Oocyte volumes (V), measured at time t on a high-performance CCD camera (Cofu; San Diego, CA) and an NIH Image analyzer, were expressed relative to the volume at time zero (V0) as V/V0 = (A/A0)3/2, where A is the area of oocyte at time t and A represents the area at time zero (A0). Osmotic water permeability (Pf) was determined from the initial slope of the time course of V/V0[d(V/V0)/dt], initial oocyte volume (V0), initial oocyte surface area (S), the molar volume of water (Vw, 18 cm/mol), and the osmolarity inside (Osmin) and outside (Osmout) the cell, using the following equation (46):
AQP protein expression on Xenopus oocyte membranes after injection of qAQP2 cRNA or hAQp1 cRNA was examined using immunofluorescence analysis. The oocytes were fixed in a 3.7% (vol/vol) formaldehyde solution containing 80 mM PIPES (pH 6.8), 5 mM Na EGTA, 1 mM MgCl2, and 0.2% (vol/vol) Triton X-100 for 4 h at room temperature (18). After postfixation in absolute methanol at −20°C overnight, the oocytes were rehydrated in Na phosphate buffer (PBS, pH 7.2), incubated for 16 h at room temperature in PBS containing 100 mM NaBH4, and were used for immunofluorescence labeling. The oocyte was incubated with rabbit anti-rat AQP2 (1:500 dilution) (gift from Dr. S. Sasaki, Tokyo Medical and Dental University, Tokyo; Ref. 13) or anti-rat AQP1 (1:500 dilution) (Chemicon International, Temecula, CA) in 2% (wt/vol) bovine serum albumin for 24 h (4°C) and then for 24 h in fluorescein-5′-isothiocyanate-conjugated goat anti-rabbit IgG (1:100). The bisected oocytes (cut in half) were mounted in 0.5-mm well slides, and fluorescence labeling was examined with a confocal scanning laser microscope (Bio-Red MRC-2400, Hemel Hemstead, UK) connected to a Nikon Optiphot.
Measurement of qAQP2 mRNA by RT-PCR.
The kidneys, brain, liver, heart, adrenal glands, and intestine were quickly removed from normally hydrated Coturnix quail (25 wk of age, N = 5), freed from surrounding connective tissues and washed in chilled aerated avian Ringer solution. Tissues were snap-frozen in liquid nitrogen and placed on dry ice. Frozen tissues were homogenized by a Polytron homogenizer (Brinkman) on ice in 4 M guanidine thiocyanate (Fluka Chemical; St. Louis, MO) containing 25 mM Na citrate (pH 7), 0.5% Na lauryl sarcosinate, and 0.72% (vol/vol) β-mercaptoethanol (Sigma); total RNA was extracted by the method of Chomczynski and Sacchi (7) using phenol-chloroform and precipitated in isopropanol. The concentration and quality of RNA were determined as above.
The cDNA was reverse-transcribed from total RNA (designated amount) with 100 pmol of random hexadeoxynucleotide primer (Pharmacia; Peapack, NJ) in 20 μl of a mixture containing 20 U of RNase inhibitor (Promega; Madison, WI) and 200 U of Moloney Murine Leukemia Virus Reverse Transcriptase (GIBCO Life Sciences) in the presence of 50 mM Tris·HCl (pH 8.3), 75 mM KCl, 5.0 mM MgCl2, 1.25 mM dNTP, and 5 mM dithiothreitol (16). After 90 min incubation at 37°C, the reaction was stopped at 65°C for 5 min, and the yields were stored at −80°C until use. We examined the possible contamination of sample RNA with genomic DNA by incubating the sample RNA without reverse transcriptase. Both series were simultaneously processed for PCR amplification.
Double-stranded cDNAs were synthesized and amplified, after initial denaturing (94°C for 5 min) by incubating (total volume, 25 μl in duplicate) the reverse transcription (RT) reaction product (3 μl) with 1 U of Taq polymerase (Roche; Mannheim, Germany) and 0.1 μM each of the 5′- and 3′-primer pairs (sense strand, 5′-CATCCTACACGAGATCACCCCAGC-3′, residue 329–353; antisense strand, 5′-CTCAACCGACTGCCGCCGCCGTGC-3, residue 797–774), in the presence of 10 mM Tris·HCl buffer (pH 8.3), 50 mM KCl, 5 mM MgCl2, 0.35 mM each dNTP, for 29 cycles at 94° (denaturation), 60° (primer annealing), and 72°C (extension/synthesis), respectively, for 30, 30, and 60 s. PCR using the above primers yielded a 469-bp product.
To conduct RT-PCR within the exponential phase of the reaction, the number of cycles, the primer-annealing step, and the polymerization step were optimized (16). The yield of PCR products from total RNA (0.2–2.0 μg) and the effect of the number of PCR amplification cycles (25–31 cycles) were examined, and those representing the midlinear part of the log dose-response curves were selected (0.8 μg and 29 cycles). PCR incubations with the RT reaction product, from which reverse transcriptase was deleted and to which the pCR4-TOPO plasmid (∼30 pg) including the full-length sequence of qAQP2 was added, are designated, respectively, negative and positive control. The glyceraldehyde-3-phosphate dehydrogenase (GAPDH) of quail mRNA (accession no. Z19086) obligatorily expressed in the same RT reaction was similarly determined. PCR amplification of GAPDH was performed for 24 cycles with a pair of primers (sense: 5′-TCAAGGGCACTGTCAAGGCTGAG-3′; antisense: 5′-CTGGTTTCTCCAGACGGCAGGTC-3′). Twenty microliters of all PCR products were electrophoresed on 1.5% agarose (Invitrogen, Carlsbad. CA) in Tris-borate-EDTA buffer.
SDS polyacrylamide gel electrophoresis and Western immunoblot analysis.
Coturnix quail (7–10 wk of age, N = 15) were divided into two groups: 1) normally hydrated (N = 7) and 2) water deprived (48 h) (presumably to enhance the endogenous AVT level and stimulate qAQP mRNA/protein synthesis) plus 50 ng/kg of AVT injected 1 h before death (presumably to optimize the trafficking of qAQP2 in the apical membrane region). Medullary cones devoid of superficial cortical regions were dissected in chilled avian Ringer solution and homogenized on ice by Polytron (10 s, 5 times) in homogenizing buffer [10 mM Tris, 250 mM sucrose, 0.1 mM EDTA, 0.1 mM phenylmethylsulfonyl fluoride (PMSF), 1 μg/ml leupeptin, and 1 μg/ml pepstatin A, pH 7.2]. To obtain membrane/submembrane fractions, the homogenate was centrifuged first at 5,000 g for 15 min (twice) and then (supernatant) at 100,000 g for 30 min (15). The pellets were resuspended in Tris·HCl buffer (50 mM Tris, 100 mM NaCl, 10 mM MgCl2, 0.1 mM PMSF, 1 μg/ml leupeptin, and 1 μg/m pepstatin A; pH 7.2). The protein concentration was assayed using the methods of Lowry et al. (22).
Protein samples solubilized in a loading buffer (Laemmli sample buffer; Bio-Rad; Hercules, CA) containing 2% SDS were electrophoresed on 12% SDS-PAGE minigel, using the Bio-Rad Mini-PROTEAN II Electrophoresis system. Proteins were transferred to nitrocellulose membranes (Bio-Rad Mini-Trans-Blot Electrophoretic Transfer system). The blots were blocked overnight with 3% BSA in TBS buffer (20 mM Tris, 500 mM NaCl, 0.05% Tween-20; pH 7.5), washed, and incubated for 1.5 h in rabbit anti-rat AQP2 (diluted 1:500 in TBS containing 0.5% BSA) or nonimmune rabbit serum. The blot was incubated 1 h in horseradish peroxidase-conjugated goat anti-rabbit IgG (Sigma; diluted 1:10,000 in Tris-buffered saline-Tween 20 (0.05% vol/vol) containing 0.5% BSA) and then incubated with the enhanced chemiluminescence substrate and exposed to film (Kodak) to visualize the immunoreactive bands. The relative optical density of bands was measured by the NIH Image program (version 1.62; NIH; Bethesda, MD) for semiquantitative analysis.
Localization of qAQP2 and enhancement of immunoreactive qAQP2 expression after dehydration or AVT treatment were examined in three groups (20 wk of age): 1) normally hydrated/0.2 ml saline-injected (sc) (N = 3), 2) water deprived for 48 h (N = 4), and 3) AVT (50 ng·kg−1·day−1 sc, tap water ad libitum) for 3 days (killed 1∼2 h after the last injection, N = 4). Quail were anesthetized with pentobarbital sodium (Nembutal, 50 mg/kg ip; Abbott Laboratories; North Chicago, IL), and the left ischiadic artery was cannulated with polyethylene tubing (Clay Adams; Parsippany, NJ). While the bifurcation of the abdominal aorta and the right ischiadic artery were manually occluded, 5 ml of heparin solution (50 U/ml PBS, pH 7.4) was quickly infused, followed by a fixative containing a periodate-lysine-paraformaldehyde mixture (PLP) for 3 min at physiological pressure for the bird. Then, either the kidneys were cut into slices (2–3 mm thick), or medullary cones were dissected and immersed in PLP for an additional 1–4 h at room temperature. After equilibration with graded concentrations of sucrose, the tissue specimens were embedded in the OCT compound and were kept at −70°C until being sectioned into 10-μm-thick slices by a cryostat. Localization of qAQP2 was examined by incubating the sliced tissues with the above-mentioned rabbit anti-rat AQP2 antibody (1:1,000 dilution) followed by treatment with biotinylated goat anti-rabbit IgG (Vector Laboratories; Burlingame, CA; 1:200 dilution) and peroxidase-antiperoxidase complex (Vectastain ABC kit) for 30 min. The peroxidase was visualized by reaction with diaminobenzidine and hydrogen peroxide (Fast DAB tablets, Sigma). The tissues were counterstained by hematoxylin.
Intensities of immunoreactive qAQP2 staining of individual CDs were graded according to two major factors: 1) approximate percentage of positively labeled cells within one CD, and 2) area and intensity of staining within one epithelial cell. Labeling intensities were scored on a range of 1.0–5.0, as follows: grade 1, similar to background showing no specific staining in epithelial cells; grade 2, diffuse and weak staining is seen at the apical or subapical area in a small number of epithelial cells; grade 3, approximately one-third of epithelial cells express positive immunoreactive qAQP2 staining on the apical membrane, and the rest of epithelial cells also show diffuse weak staining at subapical regions; grade 4, approximately two-thirds of epithelial cells show clear staining on the apical membrane, and the rest of the cells exhibit diffuse staining at the subapical regions; and grade 5, the majority of epithelial cells express strong staining at both the apical and subapical regions. Fifteen tubules from the proximal tubule, thick ascending limb, and cortical (superficial) CD per bird were randomly selected and graded. All medullary CDs visualized in cross sections of medullary cone slices were examined and graded for each bird. Occurrence (% of positive tubules of 15 tubules or all tubules examined) and average intensity scores (mean of 2 slices/bird, and then mean ± SE for the group) were calculated. Histological images were examined by Olympus BX50 microscope and photographed by a digital camera (Olympus, C-2020zoom, Tokyo).
All the data are shown as means ± SE. For statistical analysis, a single- or two-factor ANOVA was used, followed by Tukey's honestly significantly different unbalanced test when applicable. The difference between control and experimental groups in immunoblot analysis was determined by Student's t-test. The difference was considered significant at a P value of <0.05.
Cloning of cDNA and analysis of amino acid sequence of qAQP2.
We initially obtained five partial cDNA that have residues 90% identical to the corresponding region of hAQP2 and 82% identical to the rat sequence. The fragment was extended in the 3′ and 5′ termini using a RACE technique in combination with use of the specific primers (see materials and methods). The full length of cDNA shows an 822-bp open reading frame containing an ATG codon and encoding 274 amino acids (Fig. 1). Sequence alignments with rat AQP2 and human AQP2 revealed, respectively, ∼76% and ∼77% overall identity. Hydropathy analysis of the deducted amino acid sequence indicates that qAQP2 has six potential transmembrane domains and five connecting loops (loops A–E) (Fig. 1B). Loops B and E contain the asparagine-proline-alanine (NPA) motif conserved in the aquaporin family. Furthermore, qAQP2 contains residues equivalent to those of the putative N-glycosylation site (asparagine-124), putative mercury-sensitive cysteine-182 site, and putative phosphorylation site for cAMP-dependent protein kinase (serine-257) in the cytoplasmic carboxyl terminus. A phylogenetic comparison between qAQP2 and reported mammalian and nonmammalian AQPs revealed that qAQP2 diverged from progenitor AQPs earlier than rat and sheep AQP2 but that the horizontal distance between qAQP2 and mammalian AQP2 is small (Fig. 2).
Osmotic water permeability of oocytes injected with qAQP2.
The time-dependent changes in volume and osmotic water permeability (Pf) of the oocytes after exposure to hyposmotic (70 mosmol/kgH2O) Barth's solution are shown in Fig. 3. The oocytes injected with qAQP2 and hAQP1 (positive control) cRNA swelled profoundly within 3 min, whereas the oocytes injected with water showed no swelling (Fig. 3A). The calculated Pfs (10−4 cm/s) of qAQP2-injected (15 ng) oocytes (107.7 ± 38.4) and hAQP1-injected (15 ng) oocytes (115.8 ± 38.4) are significantly (P < 0.01) higher than that of water-injected oocytes (4.93 ± 0.92) (Fig. 3B). Preincubation with 0.3 mM HgCl2 for 5 min inhibited swelling of the qAQP2-injected oocytes in hyposmotic media by 81.3%.
Using immunofluorescence analysis, we investigated whether qAQP2 protein is expressed on the membrane of the qAQP2 cRNA-injected oocyte. Rabbit anti-rat AQP2 antibody, raised against a synthetic peptide consisting of 17 amino acids identical to those in the rat AQP2 COOH terminus (Fig. 1), was used to recognize qAQP2. The oocytes injected with qAQP2 cRNA (Fig. 4B) or human AQP1 cRNA (Fig. 4C) revealed intense fluorescent labeling of the plasma membrane, whereas the oocytes receiving a water injection showed no staining in the membrane (Fig. 4A).
Tissue-specific expression of qAQP2 mRNA.
The tissue distribution of qAQP2 mRNA determined by RT-PCR with specific primers is shown in Fig. 5. The RT-PCR product from total RNA from the kidney, but not from the brain, heart, liver, adrenal glands, or intestines, exhibited a single 469-bp band when analyzed by agarose gel electrophoresis (Fig. 5A). This fragment matched the expected size of qAQP2. The level of qAQP2 mRNA (intensity of PCR product band) was higher in medullary cones than in the cortex (Fig. 5B). All RNA preparations were free from DNA contamination (no RT-PCR product band from the mixture from which reverse transcriptase was deleted). The GAPDH mRNA was ubiquitously expressed in all organs examined (Fig. 5B).
Effect of water deprivation and vasotocin on qAQP2 protein expression.
The above-indicated rAQP2 antibody recognized proteins of ∼29 kDa (predicted molecular size) and ∼38 kDa extracted from medullary cones of quail kidneys (Fig. 6). The relative optical densities of the bands at ∼29 kDa derived from the water-deprived plus AVT-treated birds were more distinct (Fig. 6, top) and significantly higher (Fig. 6, bottom) than those from normally hydrated controls. There was no difference in density of the ∼38-kDa band between these two groups. We also observed a less intense band at a slightly higher region (∼45 kDa). Figure 6 shows a part of this band.
Immunolocalization of qAPQ2 and the effect of water dehydration and vasotocin.
Thin frozen sections of the cortical area and medullary cones incubated with anti-rat AQP2 antisera are shown in Fig. 7. In normally hydrated quail, relatively weak positive staining is noted at the apical and/or subapical regions of the CD in the superficial area (presumably from a loopless nephron) (Fig. 7, A and B). Medullary CDs have larger diameters and show more intense apical/subapical staining (Fig. 7C). The intensity and area (apical membrane and subapical regions immediately below the apical membrane) increased after 48 h water deprivation (Fig. 7D) or treatment with AVT (50 ng·kg−1·day−1 ip, 3 days) (Fig. 7E). Positive staining was not seen in the basolateral area of CDs, CDs incubated with control rabbit serum (Fig. 7F), or in renal tubules from other than CDs. The percentage of positively stained tubules and the intensity scores of randomly selected proximal tubules, thick limbs, superficial CDs, and medullary CDs are summarized in Table 1 (see materials and methods for grading criteria). Only a negligible background level of staining was noted in proximal tubules and thick limbs. In normally hydrated birds, >90% of the CDs from superficial (cortical) regions and virtually all medullary CDs contain some positively stained epithelial cells. Water-deprived (48 h) quail or those that received repeated AVT injection showed significantly higher intensity of immunoreactive qAQP2 in medullary CD cells, whereas no significant difference was seen among the three groups in superficial CDs.
The AQPs are a family of small, hydrophobic, major intrinsic proteins (MIP), originally cloned in mammalian lens as MIP26 (for review, see Refs. 1, 26, 40). AQPs consist of homotetramers in which each monomer contains six membrane-spanning α-helical segments forming five loops with the amino and carboxyl termini in the intracellular milieu. The transmembrane domains reveal two internal repeats; each contains a short-sequence NPA motif highly conserved among AQPs; mutation of this region alters water permeability (18). In mammalian kidneys, AQP2 provides the water transport pathway across the apical plasma membrane of the principal cells of the CD (13, 25), and, in the inactive state, AQP2 resides in the subapical vesicles.
Using degenerate primers constructed based on the conserved AQP sequence and RT-PCR cloning techniques, we successfully identified an AQP2-homolog sequence from the medullary cones of the Coturnix quail kidney, indicating that the AQP molecule is evolutionarily stable. We noted that qAQP2 has six potential transmembrane domains and contains the NPA structure between domains 2 and 3 (loop B) and domains 5 and 6 (loop E). Loops B and E are significantly hydrophobic in mammalian AQPs (17). The putative N-glycosylation site (asparagine-124) and putative phosphorylation site for cAMP-dependent protein kinase (serine-257) in the cytoplasmic carboxyl terminus are equivalent to those seen in the rat AQP2 (11).
We demonstrated that Xenopus laevis oocytes injected with qAQP2 transcript express immunoreactive AQP2 protein that evokes profound Pf in hyposmotic media, equivalent in magnitude to the Pf obtained by human AQP1, suggesting that qAQP2 acts as a water channel. Inhibition of Pf by HgCl2 agrees with the presence of the putative mercury-sensitive site (cysteine-182) in a location equivalent to that in rat AQP2 (2, 13); this indicates that the mercury-sensitive property of the AQP water channel is phylogenetically conserved.
In Coturnix quail, the CDs comprise intercalated cells and mucus-secreting cells, presumably equivalent to the principal cells of mammalian kidneys that possess V2 receptors for ADH (30). We found that immunoreactive qAQP2 is expressed on the apical side of the CD cells in both superficial regions and medullary cones from normally hydrated quail, suggesting that at a normally hydrated state with a low level of circulating AVT (37), water is transported via qAQP2 water channels. While the majority of epithelial cells of the CD are positively labeled, a few cells, possibly intercalated cells, exhibit no staining. We noted that qAQP2 mRNA expression was detected in the kidneys but not in other organs that contain high levels of total RNA and that the mRNA level was higher in medullary cones than in cortical superficial regions. This intrarenal heterogeneity of qAQP2 mRNA levels agrees with that of the immunohistochemical localization of immunoreactive qAQP2 protein. It remains to be determined in microdissected tubules, however, whether CDs that originate in loopless reptilian-type nephrons express lower levels of qAQP2 mRNA and protein than CDs from looped nephrons. Immunoblots of the membrane protein fractions from medullary cones, after separation by SDS-PAGE, revealed a specific band with an ∼29-kDa protein; the protein size agrees with that of rat AQP2 (12). The relative density of the 29-kDa band was significantly higher in the birds treated with water deprivation plus AVT, whereas the density of the band of higher molecular mass (∼38 kDa) did not change. It remains to be examined whether the latter band represents glycosylated protein (35) or may merely be a nonspecific labeling to heterologous antibody.
In birds, the diffusional water permeability of the medullary CD is considerable and is only modestly increased by AVT (30); the AVT-induced cAMP production in medullary cones/CDs is also less than that evoked by ADH in mammalian CDs (14, 30). The present study also indicates that water deprivation and/or treatment with AVT increases immunoreactive qAQP2 levels, but that AVT's effect is smaller than that of either exogenously administered or endogenously elevated ADH in rats on AQP2 mRNA (33) or on protein (10, 15, 25). In normally hydrated quail kidneys, the level of immunoreactive qAQP2 (%positively stained and intensity) in CDs from the cortical area tends to be slightly lower than that from medullary cones, agreeing with the weaker expression of qAQP2 mRNA in the former. In addition, immunoreactive qAQP2 levels significantly increased only in medullary CDs after water deprivation/AVT treatment, suggesting that heterogeneity exists between superficial (presumably loopless nephrons) and medullary (looped mammalian-type) nephrons in qAQP2 expression and its responses to AVT. The mechanism of this heterogeneity is unclear at present. Because both AVT and forskolin stimulate cAMP production in the cortical and medullary areas, it is possible that this heterogeneity may be attributed to an effect downstream of the signal pathway. Further investigation is needed on the effect of AVT on qAQP2 mRNA and protein using isolated CDs derived from cortical loopless and medullary looped nephrons. Interestingly, the enhancement of qAQP2 levels after 48-h water deprivation was not significantly different from that after 3 consecutive days of injection of AVT at a considerable dose, suggesting that as in fowl (37), dehydration makes endogenous AVT sufficiently high. Due to the limited availability of anti-rat AQP2 antibody, the combination of water deprivation plus AVT was not examined in the present study. In a separate study, we noted that medullary CDs from Coturnix quail that received water deprivation plus AVT (single injection 1–2 h before death) exhibited clear concentration of FITC-conjugated rat AQP2 antibody labeling at the apical membrane (28).
In mammalian kidneys, vasopressin regulates AQP2 expression in two stages: first, short-term (min) regulation of CD water transport via enhancement of AQP trafficking and, second, longer-term (hours-days) modulation of CD water permeability via upregulation of AQP2 mRNA and protein (15, 21, 39). AQP2 is primarily regulated by a vasopressin/cAMP-dependent vesicular trafficking mechanism via V2 receptors (for review, see Refs. 36, 40). Using primary culture of CD cells from quail kidney, we examined whether AVT stimulates trafficking of AQP-containing vesicles. Although condensation of FITC-conjugated immunoreactive particles was seen at cell membrane regions 15–20 min after application of 10−5 M of AVT, the results were not conclusive. This may partly be due to rapid phenotypic modulation of CD cells, similar to the observation reported for rats (23); this agrees with the finding that AVT increased the water permeability of isolated and perfused quail CDs (30).
Stoichiometric studies indicate that in principal cells of CDs, three of four monomers of AQP2 must be phosphorylated before initiation of trafficking from the subapical endosomes to the apical membrane (19) and that serine-256 of the AQP2 is critical for triggering vasopressin/cAMP-induced translocation (11). Furthermore, a series of amino acid mutations of the sixth transmembrane domain in the rat revealed that the dileucine motif (217–218, 222–223) plays an important role in the vasopressin-induced translocation of AQP2 (43). The current study indicates that qAQP2 also possesses an equivalent serine residue and di-isoleucine (217–218) and isoleucine-leucine motifs (222–223). Although the effect of this substitution of leucine for isoleucine in AVT-induced vesicle trafficking of qAQP2 remains to be examined, it has been reported that replacement of leucine with isoleucine residues results in a decreased efficiency of targeting of lysosomal membrane protein to lysosomes (34), which may partly explain the only modest effect of AVT on qAQP2 expression/water permeability.
Previously we proposed that urine concentration occurs in avian kidneys by the recycling of a single solute (NaCl) and that the Na+-K+-2Cl− cotransporter function, which provides an energy source for the countercurrent multiplier mechanism, may be regulated by local factors such as tubular urine flow and NaCl concentration (29, 31). The present study suggests that, similar to the mammalian mechanism, AVT likely enhances urine concentration in avian kidneys via the qAQP2 water channel. In addition, AVT reduces the single-nephron glomerular filtration rate by constricting afferent arterioles of cortical loopless nephrons and thus reduces the volume flow passing through the medullary CD, leading to the increase in tubular urine concentration via prolonged contact time (8). This vascular effect may supplement AVT's relatively weak direct effect on CD water permeability. In humans, hereditary nephrogenic diabetes insipidus (NDI) occurs by mutations of the genes encoding either the vasopressin V2 receptor or AQP2 (1, 20), including premature termination of translation that leads to autosomal recessive NDI (9, 24). A possibly nephrogenic DI chicken strain exists that produces abundant dilute urine with normal or higher plasma AVT levels (6).
AQPs are phylogenetically old molecules and AQP and aquaglyceroporin homologs are present in plants, bacteria, and invertebrates (for review, see Ref. 3). Although it is anticipated that all vertebrate kidneys possess AQP water channels, AQP isoforms have been cloned from tissues of only a few species of nonmammalian vertebrates, including gill and rectal gland of teleost fish, toad bladder (for review, see Refs. 27, 28), frog skin (38), and Xenopus laevis oocytes (41). Most nonmammalian AQPs exhibit homology to AQP1 or AQP3, and the present study is the first clear demonstration of an AQP2 homolog that appears to be regulated by AVT. From a taxonomical view point, the reptilian evolutionary line leading to the mammals (Synapsida) departed from the reptilian-avian line (Archosauria) at a very early stage of tetrapod evolution (44). It is interesting to consider whether the prototypes of AQP2 and the urine concentrating mechanism evolved in ancestral tetrapods and are conserved in birds and mammals or whether water conservation via AVT-regulated AQP isoforms evolved independently during diversion of birds and mammals.
In summary, we have cloned and characterized qAQP2 from Coturnix quail medullary cones that shows high homology to mammalian AQP2 and has conserved amino acid residues common to mammalian AQP2. The qAQP2 acts as a mercury-sensitive water channel in Xenopus laevis oocytes and is expressed in apical/subapical regions of CDs both in the superficial cortical region and, more markedly, in medullary cones. The lack of enhancement of qAQP2 protein levels in cortical CDs and the only modest stimulation in medullary CDs evoked by water deprivation/AVT treatments suggest that a population of qAQP2 may not be linked to the regulatory pathway by AVT.
We are grateful for support by National Science Foundation Grant IBN-9986633 and National Institutes of Health (NIH) Grant HL-52881 (Principal Investigator: H. Nishimura) and by NIH Grant MG-61943 (Principal Investigator: Z. Fan).
We thank P. Agre, Johns Hopkins Univ., for the generous gift of AQP plasmid, and Dr. J. Kyle, Univ. of Chicago, for kindly helping us with production of AQP cRNA. We also thank P. Jiao and J. Bolen for excellent technical assistance.
Present address for L. Zhang: Center for Research on Reproduction and Women's Health, and Abramson Family Cancer Research Institute, Univ. of Pennsylvania School of Medicine, Philadelphia, PA 19104.
Preliminary studies were presented at the annual meetings of the Federation of American Societies for Experimental Biology, 2001 and 2003, and at the American Physiological Society Conference on Comparative Physiology, 2003. qAQP2 cDNA sequence has been published in GenBank (accession no. AY430098).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2004 the American Physiological Society