INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

THE INNER MEDULLARY collecting duct of the mammalian kidney responds to vasopressin with a marked increase in water permeability and in urea permeability (14, 17, 22). Both of these processes are critical to the overall regulation of water excretion. Both processes are mediated by cAMP (10, 27) and are initiated within 40 s of exposure of the collecting duct to increased levels of vasopressin (29). Studies of aquaporin-2 (AQP2), the vasopressin-regulated water channel, have demonstrated that the short-term effect of vasopressin to raise water permeability is a result of vasopressin-stimulated trafficking of AQP2-bearing vesicles to the apical plasma membrane (13, 15, 20, 30). It has been hypothesized that vasopressin may increase the urea permeability by a similar mechanism (18). However, this hypothesis has not yet been directly addressed.

Urea transport across the apical plasma membrane of the inner medullary collecting duct (IMCD) cells is facilitated by a specialized phloretin-sensitive urea transporter initially identified by physiological techniques (4, 21, 26) and eventually cloned by Shayakul et al. (25). The apical plasma membrane is the rate-limiting membrane for overall transcellular urea transport and is the site at which vasopressin regulates urea transport (4, 26). Therefore, the apical urea transporter is referred to as the vasopressin-regulated urea transporter (VRUT) of the IMCD (12). More recently, the cloned VRUT has been referred to as either "UT1" (25) or "UT-A1" (23). Immunocytochemical studies, using a polyclonal antibody raised to a synthetic peptide corresponding to the carboxy-terminal tail of the 929 amino acid protein predicted from the open reading frame of the UT1 cDNA clone, have confirmed apical localization of the urea transporter protein in the IMCD (18). The chief aim of the present study was to determine whether vasopressin activates VRUT by the same mechanism involved in activation of the AQP2 water channel, i.e., by stimulation of VRUT trafficking to the plasma membrane of IMCD cells.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Antibodies. Antibodies recognizing the carboxy-terminal region of AQP2 (L127) (5), the carboxy-terminal region of VRUT (L194) (18), and the middle loop of VRUT (L448) (28) have been previously characterized. They were affinity purified as described (5, 18, 28).

IMCD suspensions. IMCDs were prepared from rat whole inner medulla suspensions as described by Chou et al. (3). Inner medullas were dissected from rat kidneys and digested at 37°C with collagenase B (3 mg/ml; Boehringer-Mannheim, Indianapolis, IN) and hyaluronidase (600 U/ml; Worthington Biochemicals, Freehold, NJ) in HEPES-buffered solution (in mM: 118 NaCl, 5 KCl, 25 NaHCO₃, 4 Na₂HPO₄, 2 CaCl₂, 1.2 MgSO₄, 5 CH₃COONa, and 5.5 glucose) containing 0.1% (wt/vol) bovine serum albumin (ICN Biomedicals, Aurora, OH) under continuous supplement of 95% air-5% CO₂ until IMCDs were free from adherent thin limbs. DNase I (Boehringer-Mannheim) at a final concentration of 0.001% (wt/vol) was added to the digesting solution to reduce aggregation of separated tubule segments. After incubation for another 15 min in this solution, the suspension was separated into IMCD-enriched and IMCD-depleted fractions by low-speed centrifugation (50 g) (Sorvall RT-6000B). The pellet was resuspended in HEPES, and the 50 g centrifugation was repeated three times. The final pellet contained mostly IMCD fragments, whereas the supernatant contained thin limbs and vascular elements. Aliquots were taken for measurement of protein concentration (Pierce BCA protein assay reagent kit; Pierce Chemical, Rockford, IL).

Surface biotinylation of IMCD suspension. Biotinylation of IMCD suspensions was carried out as described by Gottardi et al. (9) with slight modification. Briefly, after prewarming suspensions in HEPES buffer solution containing 0.1% (wt/vol) BSA for 60 min at 37°C, suspensions were incubated in the same solution without vasopressin (AVP) or with 10¹⁰, 10⁸, and 10⁶ M AVP for 20 min. At the end of the incubation, suspensions were put into ice to inhibit endocytosis. After centrifugation at 50 g for 10 s at 4°C, the supernatant was discarded and the pellet was washed three times at 4°C with ice-cold biotinylation buffer solution (in mM: 215 NaCl, 4 KCl, 2.5 Na₂HPO₄, 2 CaCl₂, 1.2 MgSO₄, 5.5 glucose, and 10 triethanolamine, pH 7.4) without biotin. Final pellet was resuspended in 1 ml ice-cold biotinylation solution containing 3.0 mg/ml of NHS-LC-biotin (ImmunoPure no. 21335, Pierce) and incubated for 60 min at 4°C with gentle horizontal motion to ensure mixing. The suspensions were then washed three times with ice-cold biotin quenching solution (0.1 mM CaCl₂, 1 mM MgCl₂, and 260 mM glycine in PBS, pH 7.4) followed by gentle mixing for 20 min at 4°C with the same solution. After washing with ice-cold lysis buffer (150 mM NaCl, 5 mM EDTA, and 50 mM Tris, pH 7.4) containing protease inhibitors (1 µg/ml of leupeptin and 0.1 mg/ml of phenylmethylsulfonyl fluoride) without detergent three times, the suspension was solubilized in 1 ml of lysis buffer (which included 1% NP-40) for 60 min at 4°C with gentle motion. The samples were centrifuged at 14,000 g for 10 min at 4°C, and 900 µl of the supernatant was added to 200 µl of streptavidin beads (ImmunoPure no. 20349, Pierce). The bead suspension was incubated overnight (16 h) at 4°C with continuous agitation. After washing the beads three times with lysis buffer containing 1% NP-40, twice with high-salt wash buffer (500 mM NaCl, 5 mM EDTA, and 50 mM Tris, pH 7.5), and once with no-salt wash buffer (10 mM Tris, pH 7.5) at 4°C, biotinylated proteins were eluted from the beads with 100 µl of 1.5% SDS in Laemmli sample buffer incubated at 60°C for 15 min.

Experiments were done to test whether biotinylation is restricted to surface proteins on the plasma membranes. An IMCD tubule suspension pooled from five rats was divided into five tubes and biotinylated with 0, 0.02, 0.2, 2, and 10 mg of NHS-LC-biotin in 1 ml of biotinylation solution. After quenching of excess free biotin as above, the suspensions were pelleted by low-speed centrifugation and resuspended in ice-cold isolation solution [10 mM triethanolamine (pH 7.6), 250 mM sucrose]. The suspensions were homogenized in ice-cold isolation solution with protease inhibitors using a tissue homogenizer (Omni 1000 fitted with a micro-sawtooth generator). Differential centrifugation was carried out as described below, yielding a high-density membrane fraction and a low-density membrane fraction. Both fractions were solubilized in 1 ml of 1% NP-40 lysis buffer with protease inhibitors for 60 min at 4°C with gentle motion. These samples were incubated with streptavidin beads, which were subsequently eluted with Laemmli sample buffer to obtain biotinylated proteins as described above. The resulting samples were subjected to immunoblotting. The high-density fraction is enriched in plasma membranes and should contain biotinylated proteins. Conversely, since the low-density fraction is virtually devoid of plasma membranes (6, 13), it should not contain biotinylated proteins if biotinylation is limited to plasma membrane proteins.

Differential centrifugation. Differential centrifugation was carried out as previously described (6, 13). Inner medullary homogenates were initially centrifuged at 4,000 g for 10 min at 4°C (Tomy, MTX-150) to remove incompletely homogenized fragments and nuclei. The pellets were resuspended in ice-cold isolation solution with protease inhibitors and centrifuged again at 4,000 g for 10 min. The supernatants were collected and centrifuged at 17,000 g for 20 min (Sorvall RC2-B centrifuge with SS34 rotor). The pellets ("high-density membrane fraction") were retained, and the supernatants were then pelleted by centrifugation at 200,000 g for 1 h (Beckman ultracentrifuge with Ti-80 rotor) ("low-density membrane fraction"). As previously documented (6, 13), the low-density membrane fraction from this protocol is virtually devoid of plasma membranes.

Immunoisolation of AQP2-bearing intracellular vesicles. AQP2-bearing intracellular vesicles were immunoisolated using affinity-purified AQP2 polyclonal antibody covalently coupled to magnetic beads (Dynabeads no. M-280 with covalently attached sheep anti-rabbit IgG; Dynal, Lake Success, NY) according to the manufacturer's instructions. Either 5.1 µg of AQP2 antibody or 5.1 µg of preimmune IgG from the same rabbit was incubated with 1.7 mg magnetic beads in 100 µl of washing solution (PBS pH 7.4, 0.1% BSA) with 0.02% azide overnight at 4°C with gentle mixing. After washing four times in same solution with bidirectional mixing for 30 min each, the beads were resuspended in 1 ml of 0.2 M triethanolamine (pH 8.2) and were washed two times with the same solution. Then, 20 mM dimethylpimelimidate (DMP), a homobifunctional cross-linker, in 10 ml of 0.2 M triethanolamine solution was added to the beads and incubated for 45 min at room temperature with bidirectional mixing. After incubation in 10 ml of cross-linking solution without DMP for a further 2 h, beads were resuspended and incubated in 10 ml of 1% NP-40 containing 0.2 M triethanolamine for another 10 min to completely eliminate noncovalently bound IgG.

A low-density membrane fraction containing intracellular vesicles but not plasma membranes (6, 13) was prepared from inner medullas by differential centrifugation as described above, except that the low-density membrane fraction (200,000 g pellet) was resuspended in washing buffer. After washing the antibody-bearing beads in 1 ml of washing buffer three times, 166 µl of the low-density membrane fraction (~0.3 µg/µl total protein) was incubated with the beads for 5 h at 4°C with gentle mixing. After washing in 1 ml of washing buffer three times, the proteins were eluted by 100 µl of Laemmli sample buffer and were solubilized at 60°C for 10 min.

Electrophoresis and immunoblotting. Samples prewarmed at 37°C for 20 min were loaded on precast 12% SDS-PAGE minigels (Novex) and electrophoresed using an X-Cell II minicell (Novex, San Diego, CA). The proteins on the gel were transferred electrophoretically to nitrocellulose membranes using a Bio-Rad Mini Trans-Blot cell (Bio-Rad, Hercules, CA). After blocking with 5% (wt/vol) nonfat dry milk in blot wash buffer [150 mM NaCl, 50 mM NaH₂PO₄, and 0.05% (vol/vol) Tween 20, pH 7.5] for 30 min at room temperature, the membranes were incubated with either anti-VRUT antibody (L194) or anti-AQP2 antibody (L127) in antibody dilution buffer solution (blot wash buffer with 0.1% BSA and 0.02% NaN₃) overnight at 4°C. After washing with blot wash buffer, the nitrocellulose membranes were incubated with 0.16 µg/ml of donkey anti-rabbit IgG conjugated to horseradish peroxidase (Pierce no. 31458) in blot wash buffer containing 5% (wt/vol) nonfat dry milk at room temperature for 1 h. Sites of antibody-antigen reaction were visualized by chemiluminescence using SuperSignal Substrate (Pierce) and exposure to light-sensitive imaging film (Kodak no. 165-1579 Scientific Imaging Film).

Dot blotting. One microliter of each sample was applied directly onto nitrocellulose membranes. Subsequent procedure after air drying of membranes for 30 min at room temperature was the same as for immunoblotting as described above. The dot density was analyzed by scanning densitometry (Molecular Dynamics, Sunnyvale, CA) and normalized by the amount (in µg) applied onto the membrane. Data are presented as mean ± SE. Comparisons between groups were made by unpaired Student's t-test.

Experimental animals. For in vivo studies of short-term regulation of AQP2 and VRUT by vasopressin, we used pathogen-free male Brattleboro homozygous (di/di; Harlan Sprague Dawley, Indianapolis, IN) rats. All rats were fed a standard rat chow ad libitum containing 18% protein and 130 meq Na⁺ per kg (NIH-31 autoclavable rodent diet; Zeigler Bros., Gardner, PA) and had free access to water. The rats were maintained at all times under pathogen-free conditions.

Animal protocol: Differential centrifugation studies. Body weights were carefully matched (control, 240 ± 6 g; DDAVP, 253 ± 3 g). The rats were maintained in metabolic cages to allow urine collections. Urine osmolalities were measured using a vapor-pressure osmometer (model 5100C; Wescor, Logan, UT). Six rats received an intramuscular injection of 2 nmol DDAVP (Sigma Chemical, St. Louis, MO) dissolved in water, and six rats were injected with the same volume of 0.9% NaCl. One hour after injection, the rats were killed by decapitation. The kidneys were rapidly removed into ice-cold isolation buffer (composition described above), and the inner medullas were separated from the outer medulla and cortex using sharp, curved, iris scissors. Homogenization and differential centrifugation were carried out as described above to yield high-density and low-density membrane fractions.

Animal protocol: Immunocytochemical studies. Brattleboro rats, weighing 250-300 g, were anesthetized with intraperitoneal pentobarbital sodium (75 mg/kg) and injected via the femoral vein with DDAVP, a stable V₂-selective analog (1 ng/animal), or with vehicle. After 20 min, the kidneys were perfusion fixed by retrograde perfusion through the abdominal aorta with 8% paraformaldehyde in 0.1 M sodium cacodylate buffer, pH 7.2. Tissue blocks were postfixed in the same fixative for 2 h, infiltrated for 30 min with 2.3 M sucrose containing 2% paraformaldehyde, mounted on holders, and rapidly frozen in liquid nitrogen. Semi-thin cryosections (0.85 µm) were obtained with a cryoultramicrotome (Reichert Jung, Vienna, Austria), and the sections were placed on gelatin-coated glass slides. After preincubation with PBS containing 1% BSA and 0.05% M glycine, the sections were incubated with 1 µg IgG/ml of either affinity-purified anti-AQP2 (L127) or anti-VRUT antibody (L194) in PBS containing 1% BSA and 0.3% Triton X-100. The labeling was visualized using anti-rabbit IgG conjugated to horseradish peroxidase (P448, 1:100; DAKO, Glostrup, Denmark). Sections were counterstained with Meier counterstain. The following controls were performed: 1) incubation with protein A-purified rabbit IgG instead of primary antibody; 2) adsorption controls, made by preincubating the affinity-purified antibody with an excess of peptide conjugated to BSA; 3) and incubation without use of primary antibody or without primary and secondary antibody. All controls revealed absence of labeling.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Surface biotinylation of plasma membrane proteins in IMCD suspension. We used surface biotinylation to assay for protein trafficking to the plasma membrane in IMCD cells. This technique employs a membrane-impermeant reagent that reacts with primary amine groups to label all plasma membrane proteins with externally accessible lysines. Biotinylated proteins were collected by binding to streptavidin beads and were eluted from the beads with SDS-containing sample buffer for immunoblotting. A preliminary experiment to test whether biotinylation was restricted to plasma membrane proteins is summarized in Fig. 1. Here, high-density (17,000 g) and low-density (200,000 g pellet from 17,000 g supernatant) membrane fractions were prepared from IMCD suspensions after incubation with different amounts of NHS-LC-biotin as described in METHODS and probed with anti-AQP2 antibody. As previously demonstrated, the 17,000 g fraction from this differential centrifugation protocol is enriched in plasma membranes, whereas the 200,000 g fraction is virtually free of plasma membrane proteins from collecting ducts, with AQP2 being abundant in both fractions (6, 13). As shown in Fig. 1, differential centrifugation of the derivatized samples revealed that only the 17,000 g fraction contained biotinylated AQP2. AQP2 in the 200,000 g fraction, which contains diverse intracellular vesicles, was not biotinylated, supporting the idea that only the cell surface is biotinylated. Biotinylation of AQP2 was nearly maximal at 2 mg/ml NHS-LC-biotin. Based on this result, all subsequent biotinylation experiments were carried out using 3 mg/ml of NHS-LC-biotin for IMCD suspensions from one rat.

View larger version (81K): [in this window] [in a new window]   Fig. 1.   Surface biotinylation of rat inner medullary collecting duct (IMCD). IMCD suspensions were divided into equal aliquots and subjected to surface biotinylation with 0, 0.02, 0.2, 2, and 10 mg of NHS-LC-biotin. High-density membrane fraction (17,000 g) and low-density membrane fraction (200,000 g) were prepared by a series of centrifugations as described in METHODS. Immunoblot was probed with affinity-purified anti-aquaporin-2 (anti-AQP2) antibody.

Figure 2 shows a representative surface biotinylation experiment done to test whether the amount of AQP2 and VRUT in the plasma membrane is increased by incubation of the IMCD suspension with AVP. Figure 2A shows the results for AQP2 using immunoblotting (top) and dot blotting (bottom). As expected from previous studies demonstrating that vasopressin stimulates translocation of AQP2 to the plasma membrane of collecting duct cells, there was a marked dose-dependent increase in AQP2 labeling by surface biotinylation in response to AVP. These observations, therefore, provide an appropriate positive control for VRUT observations. Figure 2B shows the immunoblotting and dot blotting results for VRUT using the same samples. In contrast to AQP2, VRUT showed no perceptible change in labeling by surface biotinylation in response to increasing concentrations of AVP. As we have previously demonstrated (1), solubilization of membranes using a nonionic detergent results in stabilization of a VRUT complex (most likely a VRUT dimer) represented by a broad band centered at around 206 kDa.

View larger version (65K): [in this window] [in a new window]   Fig. 2.   Effect of arginine vasopressin (AVP) on extent of biotinylation of AQP2 and vasopressin-regulated urea transporter (VRUT) at IMCD cell surface. Immunoblots (top) and immuno-dot blots (bottom) of biotinylated proteins from IMCD suspension were probed with either affinity-purified anti-AQP2 (A) or anti-VRUT antibody (B). This is an example that is typical of 6 experiments. For immunoblots, 10 µl was loaded in each lane. For dot blots, 1 µl of each sample was directly applied to the nitrocellulose membrane.

Figure 3 summarizes quantification of six such studies using laser densitometry to assess dot density on AQP2 and VRUT dot blots. There was a significant increase in cell surface labeling of AQP2 to a maximum of 350% of the AVP-independent value, whereas there was no change in biotinylation of VRUT at the cell surface. Thus we find no evidence for vasopressin-stimulated trafficking of the urea transporter VRUT.

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Quantification of dot density in 6 biotinylation experiments. Solid bars, results from AQP2 dot blots; hatched bars, results from VRUT dot blots. Dot density was analyzed by scanning densitometry, and values were normalized to show percent change. * P < 0.05, significantly different based on ANOVA with Bonferroni's method employed for individual contrasts vs. the unstimulated control.

Immunocytochemical localization of AQP2 and VRUT: Effect of vasopressin. To address further whether vasopressin stimulates redistribution of VRUT in IMCD cells, immunocytochemical localization was carried out. Brattleboro rats were injected with either DDAVP (n = 3) or with vehicle (n = 3), and the kidneys were perfusion fixed with paraformaldehyde fixative as described in METHODS. Thin sections were prepared, and horseradish-peroxidase immunocytochemistry was carried out in the inner medulla using antibodies against AQP2 and VRUT. Figure 4, top, shows the localization of AQP2 in the IMCD after vehicle injection (Fig. 4A) and after DDAVP injection (Fig. 4B). In the vehicle-treated control IMCDs, AQP2 was distributed both to the apical region of the cell and throughout the cytoplasm (Fig. 4A). The AQP2 in the cytoplasm is known to be due to the presence of this water channel in intracytoplasmic vesicles (16). In contrast, after DDAVP, the labeling was almost exclusively in the extreme apical region of the IMCD cells (Fig. 4B), consistent with the conclusion that DDAVP stimulates translocation of AQP2 vesicles to the apical plasma membrane. Figure 4, bottom, shows VRUT labeling in the IMCD. In the IMCDs from vehicle-treated rats, the VRUT labeling of collecting duct cells was distributed both to the apex of the cells and intracytoplasmic vesicles (Fig. 4C), as previously demonstrated (16). In contrast to AQP2, VRUT labeling in the IMCD cells was not affected by DDAVP administration to the rats (Fig. 4D). Similar observations were made in two other pairs of rats (not shown). Labeling was absent in parallel sections when an IgG fraction of the preimmune serum was substituted for the primary antibody (not shown). These immunocytochemical results are consistent with the observations made in surface biotinylation experiments (above) and further support the conclusion that vasopressin does not stimulate trafficking of VRUT to the apical plasma membranes of IMCD cells.

View larger version (131K): [in this window] [in a new window]   Fig. 4.   Immunocytochemical localization of AQP2 and VRUT. Immunoperoxidase labeling of thin sections from inner medullas of Brattleboro rats treated with DDAVP (B and D) or vehicle (A and C). A and B were labeled with anti-AQP2 antibody; C and D were labeled with anti-VRUT antibody. Note that DDAVP stimulated a redistribution of AQP2, but not VRUT, toward the apical plasma membrane. Observations shown were confirmed in two other pairs of rats (not shown).

Differential centrifugation. Differential centrifugation techniques have previously been employed to study regulated trafficking of AQP2 (6, 13). We used the same protocol to investigate whether VRUT is regulated by trafficking in the IMCD. Inner medullas from Brattleboro rats injected intravenously with either 2 nmol of DDAVP (n = 4) or with vehicle (n = 4) were subjected to differential centrifugation as previously described (13). The DDAVP-injected rats showed an increase in urinary osmolality from 240 ± 9 to 682 ± 78 mosmol/kgH₂O 60 min after injection. The vehicle-injected control rats showed urinary osmolalities of 278 ± 26 mosmol/kgH₂O prior to injection and 356 ± 18 mosmol/kgH₂O 60 min after vehicle injection. Figure 5A shows immunoblots run with the low-density membrane fraction (pellet from 200,000 g spin using supernatant from 17,000 g spin), which contains predominantly intracellular vesicles. As shown, there was a marked reduction in AQP2 band density in the low-density membrane fraction from DDAVP-treated rats relative to vehicle-treated rats (Fig. 5A), indicating a depletion of AQP2 from intracellular vesicles. In contrast, when the same samples were run for VRUT, there was no evident decrease (Fig. 5B). Figure 6 shows a summary of quantification of the blots from the differential centrifugation experiments expressed as the ratio of band density of the high-density fraction (17,000 g) to the band density of the low-density fraction (200,000 g). As previously demonstrated (13), this ratio for the AQP2 blots was markedly increased by DDAVP administration, consistent with vasopressin-induced trafficking to the plasma membrane. However, the corresponding ratio for VRUT was unchanged. This combines with the evidence from the surface biotinylation studies and immunocytochemical studies (above) to show that although vasopressin stimulates trafficking of AQP2 in the IMCD, parallel trafficking of VRUT cannot be observed.

View larger version (68K): [in this window] [in a new window]   Fig. 5.   Differential centrifugation of rat inner medullary membranes showing effect of DDAVP on abundance of AQP2 and VRUT in low-density membrane fraction. Immunoblots were carried out using low-density membrane fraction from vehicle- or DDAVP-treated Brattleboro rats and were probed with either affinity-purified anti-AQP2 (A) or anti-VRUT antibody (B). Two micrograms of protein was loaded in each lane for AQP2 blots, and 15 µg protein per lane was loaded for VRUT blots.

View larger version (35K): [in this window] [in a new window]   Fig. 6.   Quantification of effect of DDAVP on the ratio of band density of the high-density fraction to the band density of the low-density fraction (HD:LD). Low-density membrane fraction and high-density membrane fraction from inner medullas of vehicle-treated (n = 4) or DDAVP-treated (n = 4) Brattleboro rats were prepared by a series of centrifugations as described in METHODS. Immunoblots were probed with either affinity-purified anti-AQP2 (A) or anti-VRUT antibody (B). Band density was analyzed by scanning densitometry. Ratio of band density of the high-density fraction (HD) to the band density of low-density fraction (LD) was significantly increased by DDAVP for AQP2 (P < 0.03, t-test), but there was no significant (NS) change in VRUT.

VRUT is not present in immunoisolated AQP2 vesicles. We immunoisolated IMCD vesicles using the anti-AQP2 antibody to test whether these vesicles contain VRUT. The starting material was a low-density membrane fraction (200,000 g pellet from the 17,000 g supernatant) from rat inner medullas that was previously demonstrated to be devoid of plasma membrane markers (6, 13). Typical results are shown in Fig. 7. In Fig. 7, left half, is an immunoblot probed with anti-AQP2. AQP2 normally runs as two bands at 29 kDa (nonglycosylated form) and 38 kDa (glycosylated form) (16). As can be seen, when vesicles were immunoisolated using anti-AQP2-bearing beads (left lane), there was a substantial yield of AQP2. In contrast, when purified preimmune IgG from the same rabbit was substituted on the beads (middle lane), there was little or no yield of AQP2, consistent with lack of substantial nonspecific binding of AQP2-bearing vesicles to the beads. The third lane shows the AQP2 present in the initial low-density membrane fraction. Figure 7, right half, shows an immunoblot run with the same samples, but probed instead with the anti-VRUT antibody. VRUT normally runs at 97 kDa when solubilized in SDS as done here and as seen in the right lane loaded with the initial low-density membrane fraction prior to the immunoisolation. As can be seen in Fig. 7, right half, little or no VRUT was detectable in the eluates from the anti-AQP2 beads (left lane) or the control beads (middle lane), despite the fact that much more of the eluates was loaded on this blot than for the blot probed with anti-AQP2. The bands seen just above the 48-kDa marker are due to the presence of small amounts of reduced IgG heavy chain that detached from the immunoisolation beads. From this experiment and three others of identical design (not shown), we conclude that AQP2-containing intracellular vesicles in the IMCD do not contain appreciable amounts of VRUT.

View larger version (40K): [in this window] [in a new window]   Fig. 7.   Absence of VRUT in immunoisolated AQP2 vesicles. Low-density membrane fraction from inner medulla was incubated with beads with covalently bound affinity-purified anti-AQP2 or IgG from preimmune serum purified by protein A column as described in METHODS. Proteins were eluted with Laemmli sample buffer. Immunoblots were probed with either affinity-purified anti-AQP2 (left) or anti-VRUT antibody (right). Two and 10 µl of eluate from beads were loaded for AQP2 and VRUT, respectively. Third lane of each panel was loaded with low-density membrane fraction used as starting material for the immunoisolation.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

It is well established that vasopressin dramatically increases the water and urea permeability of the IMCD epithelium (14, 17, 22) and does so via increases in intracellular cAMP (27). It is also well established that the rapid increase in water permeability seen with vasopressin stimulation (seen within minutes of vasopressin exposure) is a result of redistribution of AQP2-containing membrane domains within the collecting duct cell to increase the number of AQP2 water channels in the apical plasma membrane at the expense of AQP2 in intracellular vesicles (13, 15, 20, 30). The finding of a substantial degree of labeling of intracellular vesicles with a rabbit polyclonal antibody to VRUT (18) raised the possibility that vasopressin-stimulated increases in urea transport in the IMCD are due to stimulation of trafficking of VRUT to the plasma membrane, either in parallel to the trafficking of AQP2 vesicles or perhaps via translocation of vesicles containing both VRUT and AQP2. In this study, we have provided several types of evidence that refute this hypothesis. First, surface biotinylation studies, although demonstrating a marked increase in AQP2 at the surface of IMCD cells in response to vasopressin, failed to show any increase in surface labeling of VRUT. Second, immunocytochemical localization experiments revealed that, although vasopressin induces a major redistribution of AQP2 labeling in IMCD cells to the apex of the cells, there was no demonstrable redistribution of VRUT labeling. Third, differential centrifugation studies, although revealing a shift of AQP2 out of the low-density membrane fraction, demonstrated no such shift in the distribution of VRUT. Finally, VRUT was not present in intracellular vesicles immunoisolated from the IMCD using a polyclonal antibody to AQP2. Therefore, we conclude from these studies that the ability of vasopressin to activate VRUT does not depend on stimulation of trafficking of VRUT to the cell surface. This conclusion is in agreement with conclusions from physiological studies in isolated perfused IMCD segments demonstrating that the time course of urea permeability changes and water permeability changes in response to addition or washout of vasopressin can be dissociated under certain conditions, indicating that different physical processes account for short-term regulation of urea and water permeability (17). Similar conclusions could be drawn from the observed action of atrial natriuretic peptide in the IMCD, which decreases osmotic water permeability without affecting urea permeability (19). Furthermore, a similar dissociation of regulation of urea and water transport was previously demonstrated in the toad bladder (2), a vasopressin-responsive epithelium that has served as a reliable model of the mammalian collecting duct.

If the short-term regulation of VRUT by vasopressin does not depend on VRUT trafficking to the plasma membrane, then what is the mechanism of the response? One possibility is direct phosphorylation of the urea transporter which may alter its transport characteristics. The VRUT protein has been shown to possess several serines and threonines that are potential targets for phosphorylation by protein kinase A (25). Preliminary reports have provided evidence suggesting that VRUT is phosphorylated in response to physiological manipulations (11) and that serine-499 found in the middle loop of the 929 amino acid VRUT protein is crucial for activation of VRUT by cAMP (24). However, a functional role for such phosphorylation in the regulation of VRUT has not been reported. Another possibility is that regulatory proteins may exist that may exert vasopressin-mediated regulatory influences on VRUT, perhaps similar to regulatory proteins involved in cAMP-mediated regulation of the type 3 Na/H exchanger, NHE3 (31). Such a possibility has not been explored in detail for VRUT. Chemical cross-linking studies, however, have demonstrated that VRUT exists in the plasma membrane of IMCD cells as a protein complex of ~206 kDa (1). Although this complex is believed to be a VRUT homodimer, the possibility that additional proteins may exist in the complex has not been ruled out.

The present observations, although providing evidence that vasopressin-mediated regulation of VRUT does not occur via changes in VRUT trafficking to the plasma membrane, do not rule out the possibility that other modes of VRUT regulation involve regulated trafficking. For example, urea transport in isolated perfused IMCD segments is markedly increased by increases in the osmolality of the fluid bathing the tubule, a process at least in part independent of the vasopressin-mediated regulatory process (7, 8). Further studies are needed to address the possible role of VRUT trafficking in this response.