The basolateral expression of mUT-A3 in the mouse kidney

【摘要】  Facilitative UT-A urea transporters play a central role in the urinary concentrating mechanism. There are three major UT-A isoforms found in the mouse kidney: mUT-A1, mUT-A2, and mUT-A3. The major aim of this study was to identify the location and function of mUT-A3. UT-A proteins were investigated using three novel mouse UT-A-targeted antibodies: ML446, MQ2, and ML194. ML446 detected mUT-A1 and mUT-A3. ML194 detected mUT-A1 and mUT-A2. Importantly, MQ2 was found to be selective for mUT-A3. MQ2 detected a 45- to 65-kDa signal in the mouse kidney inner medulla, which was deglycosylated to a 40-kDa protein band. Immunolocalization studies showed that mUT-A3 was strongly detected in the papillary tip, mainly in the basolateral regions of inner medullary collecting duct (IMCD) cells. Immunoblotting of subcellular fractions of inner medullary protein suggested that in mouse kidney mUT-A3 was present in plasma membranes. Consistent with this, immunoelectron microscopy demonstrated that mUT-A3 was predominantly localized at the basal plasma membrane domains of the IMCD cells in mouse kidney. Heterologous expression of mUT-A3-enhanced green fluorescent protein in Madin-Darby canine kidney cells showed that the protein localized to the basolateral membrane. In conclusion, our study indicates that mUT-A3 is a basolateral membrane transporter expressed in IMCD cells.

【关键词】  inner medullary collecting duct cells fluid homeostasis

FACILITATIVE UREA TRANSPORTERS regulate the movement of urea across biological membranes and are derived from the UT-A ( Slc14a2 ) and UT-B ( Slc14a1 ) genes ( 1, 12, 13, 16, 24 ). These specialized proteins play a central role in the urinary concentrating mechanism and are hence vital in fluid homeostasis ( 19 ). The UT-B gene gives rise to only two transcripts that encode the same protein ( 4, 12, 30 ). In contrast, there are five known splice variants of the UT-A gene ( 24 ). Four of these have been localized predominantly to the kidney, UT-A1 ( 5, 21 ), UT-A2 ( 5, 23, 31 ), UT-A3, and UT-A4 ( 7, 10, 22 ), whereas a fifth, UT-A5, is expressed in the testis ( 7 ).

Although the potential role of UT-A urea transporters in the urinary concentrating mechanism is established, the precise function of some of the renal UT-A isoforms remains to be clarified. In the rat, UT-A1 is present exclusively in the inner medullary collecting duct (IMCD) ( 14, 20 ). It is located in the apical membrane and subapical intracellular vesicles ( 14 ) and is stimulated in the short term by PKA agonists ( 5, 21 ). In contrast, UT-A2 is detected in the thin descending limbs of both the inner stripe of the outer medulla (ISOM) and the inner medulla ( 20, 29 ) and is not affected by short-term exposure to PKA agonists ( 5 ). Previously, UT-A3 has been located in the intracellular membranes and apical region of rat IMCD cells ( 28 ) and shown to be stimulated by PKA agonists ( 5, 10, 22 ). However, an almost identical protein (termed UrT1-C), which was unaffected by PKA agonists, has been suggested to function at the basolateral membrane of the same IMCD cells ( 22 ). At present, little is known about the location and regulation of the UT-A4 isoform that has only been detected in rat kidney using RT-PCR ( 10 ). We have previously characterized three UT-A isoforms isolated from mouse kidney, mUT-A1, mUT-A2, and mUT-A3, and shown that mUT-A1 and mUT-A3 are sensitive to PKA agonists, whereas mUT-A2 is not ( 5 ). In this study, to clarify the renal role of UT-A3, we determined the cellular location of UT-A3 in mouse kidney using three specific mUT-A antibodies and confirmed this location by heterologous expression of enhanced green fluorescent protein (EGFP) tagged mUT-A3. Our results show that mUT-A3 is predominantly expressed in basolateral membranes of IMCD cells, where we suggest it functions as a cAMP-regulated urea transporter.

METHODS

Antisera. To study the distribution of UT-A proteins in mouse kidney, three polyclonal antibodies were raised in rabbits ( Fig. 1; Eurogentec). Antiserum ML446 was raised against an immunizing peptide that corresponded to amino acids 57 to 75 (H 2 N-EEKDLRSSDEDSHIVKIEK-CONH 2 ) of mUT-A1 (GenBank accession number, AAM00357 . Antiserum MQ2 was raised against an immunizing peptide that corresponded to amino acids 444 to 461 (H 2 N-IYFLTVRRSEEEKSPNGD-CONH 2 ) at the COOH terminus of mUT-A3 (GenBank accession number, NP_109608 ). Finally, antiserum ML194 was raised against an immunizing peptide that corresponded to amino acids 912 to 930 (H 2 N-QEKNRRASTITKYQAYDVS-COOH) at the COOH terminus of mUT-A1. The antisera were affinity-purified using Affigel support columns (Bio-Rad) containing immobilized immunizing peptides, as previously described ( 8 ).

Fig. 1. Schematic representation of renal UT-A isoforms. UT-A1 (GenBank accession number, AAM00357 is shown at the top with the 4 hydrophobic domains represented as shaded boxes. Horizontal black lines below represent individual primary sequences of urea transporters compared with UT-A1. Antiserum ML446 was raised against amino acids 57 to 75 (H 2 N-EEKDLRSSDEDSHIVKIEK-CONH 2 ) of mUT-A1. Antiserum MQ2 was raised against amino acids 444 to 461 (H 2 N-IYFLTVRRSEEEKSPNGD-CONH 2 ) at the COOH terminus of mUT-A3 (GenBank accession number NP_109608 ,). Antiserum ML194 was raised against an immunizing peptide that corresponded to amino acids 912 to 930 (H 2 N-QEKNRRASTITKYQAYDVS-COOH) at the COOH terminus of mUT-A1. Diagonally shaded vertical lines represent epitopes of isoform-selective UT-A antisera. Intersection of horizontal and vertical lines indicates antiserum predicted to recognize protein.

Immunoblotting. Tissues were obtained from male adult NMR1 mice and homogenized in ice-cold buffer with a handheld dounce homogenizer. The homogenization buffer (pH 7.6) contained 12 mM HEPES, 300 mM mannitol, and several peptidase inhibitors added immediately before use: 1 µg/ml Pepstatin, 2 µg/ml Leupeptin, and 1 µg/ml phenylmethylsulfonyl fluoride (Sigma). For standard immunoblots, homogenates were initially centrifuged at 2,500 g for 15 min at 4° C. The resulting supernatant was centrifuged at 200,000 g for another 30 min at 4° C. These plasma membrane-enriched pellets were retained and resuspended in homogenization buffer. For subcellular fractionation experiments, homogenates of whole inner medulla underwent a series of differential centrifugation steps at 4° C: 1 ) 1,000 g for 10 min (pellet contains unhomogenized tissue); 2 ) 4,000 g for 20 min (pellet contains nuclei, mitochondria, and some plasma membranes); 3 ) 17,000 g for 20 min (pellet contains plasma membranes); and 4 ) 200,000 g for 1 h (pellet contains intracellular vesicles, while supernatant contains cytosolic proteins) ( 27 ). Total protein concentrations were determined using a Bio-Rad Protein Assay Reagent Kit (Bio-Rad). Laemmli sample buffer [5% SDS, 25% glycerol, 0.32 M Tris(hydroxymethyl)aminomethane, pH 6.8, bromophenol blue, 5% -mercaptoethanol] was added to protein samples in a ratio of 1:4, which were then heated at 60° C for 15 min. SDS-PAGE was performed on minigels of 8% polyacrylamide by loading 5 µg/lane of protein. Proteins were then transferred electrophoretically to nitrocellulose membranes (Gelman Sciences). After being blocked with 5% nonfat dry milk in washing buffer (15 mM Tris·HCl, ph 8.0, 150 mM NaCl, 0.01% Tween 20) for 1 h, the membranes were probed with affinity-purified antisera (ML446, MQ2, or ML194) for 16 h at 4° C. The membranes were rinsed in washing buffer for 3 x 10 min and then probed with goat anti-rabbit horseradish peroxidase (HRP)-linked secondary antiserum (Dako) at 1:5,000 dilution in 5% nonfat milk in washing buffer for 1 h. After another 3 x 10-min rinse in washing buffer, detection of protein was performed using the ECL Western Blotting Detection Reagents and ECL film (Amersham Pharmacia). The commercial Na-K-ATPase 1 -subunit antibody (05-369, Upstate Biotechnology) was used in conjunction with a 1:10,000 dilution of anti-mouse HRP-linked secondary antibody (NA931V, Amersham Pharmacia).

Immunolocalization. Male adult NMR1 and C57/BL6J mice were anesthetized with Inactin (100 mg/kg ip), perfused via the descending aorta with PBS, and tissues were perfusion-fixed with 4% paraformaldehyde/PBS. Kidneys were removed, dehydrated in an ascending series of ethanol concentrations (50-100%), and embedded in paraffin wax. Tissue was sectioned at 5 µm, mounted on Superfrost Plus slides (BDH), and allowed to dry overnight at 37° C. After xylene treatment and rehydration in a descending series of ethanol concentrations (100-70%), endogenous peroxidase was blocked by incubating sections for 30 min in 3% hydrogen peroxide in methanol. Antigen retrieval was performed by boiling sections for 10 min in a solution containing 25 mM Tris·HCl (pH 8.0), 10 mM EDTA, and 50 mM glucose before overnight incubation at 4° C with affinity-purified antiserum (ML446, MQ2, or ML194) diluted in 0.1% BSA, 0.3% Triton X-100 in PBS. Labeling was visualized with a 1:200 dilution of P448 goat anti-rabbit HRP-conjugated secondary antibody (Dako), followed by incubation with diaminobenzidine.

Immunoelectron microscopy. For immunoelectron microscopy, the frozen inner medulla of NMR1 mouse kidneys was freeze-substituted in a Reichert AFS freeze substitution unit. In brief, the samples were sequentially equilibrated over 3 days in methanol containing 0.5% uranyl acetate at temperatures gradually raised from -80 to -70° C and then rinsed in pure methanol for 24 h while increasing the temperature from -70 to -45° C and infiltrated with Lowicryl HM20 and methanol 1:1, 2:1, and, finally, pure Lowicryl HM20 before UV-polymerization for 2 days at -45° C and 2 days at 0° C. Immunolabeling was performed on ultrathin Lowicryl HM20 sections (60- to 80-nm thickness). Sections were pretreated with the saturated solution of NaOH in absolute ethanol (2-3 s), rinsed, and preincubated for 10 min with 0.1% sodium borohydride and 50 mM glycine in 0.05 M Tris, pH 7.4, supplemented with 0.1% Triton X-100. Sections were rinsed and incubated overnight at 4EC with anti-mUT-A3 antibody (MQ2) diluted in 0.05 M Tris, pH 7.4, supple-mented with 0.1% Triton X-100 and 0.2% milk. After rinse, sections were incubated for 1 h at room temperature with goat anti-rabbit IgG conjugated to 10-nm colloidal gold particles (GAR.EM10, BioCell Research Laboratories, Cardiff, UK). The sections were stained with uranyl acetate and lead citrate before examination in a Philips CM100 electron microscope.

Heterologous expression of mUT-A3-EGFP in Madin-Darby canine kidney II cells. To determine membrane targeting of mUT-A3, we transiently expressed a mUT-A3-EGFP fusion protein in the polarized renal epithelial cell line Madin-Darby canine kidney (MDCK) II and compared the distribution of expression with Kir1.1b (ROMK2), the secretory K + channel of the distal nephron that is targeted to the apical membrane of MDCK cells ( 15 ). The coding region of mUT-A3 was amplified by PCR and subcloned in frame into the Eco R1 site of the EGF protein expression vector pEGFP-C2 (Clontech, Oxford, UK). MDCK II cells (a kind gift from Prof. N. L. Simmons, University of Newcastle-upon-Tyne, Newcastle, UK) were routinely cultured on plastic in a humidified, 95% air-5% CO 2 atmosphere at 37° C in minimum essential medium (MEM) supplemented with 10% fetal bovine serum, 2.0 mM L -glutamine, 10,000 U/ml penicillin, and 10 mg/ml streptomycin (Sigma, Poole, UK). Before transfection, cells were passaged onto permeable filter inserts (6.5-mm-diameter Corning-Costar Transwells; 0.4-µm pore size; VWR International, Poole, UK) at a density of 3 x 10 5 cells per insert and grown to confluency (2-3 days). Confluent monolayers were transfected from the apical compartment using a Lipofectamine 2000 (Life Technologies, Paisley, UK) according to the manufacturer's instructions. Briefly, monolayers were incubated with 1 µg of cDNA and 6 µl of lipofectamine 2000 in serum-free OptiMEM-1 per filter for6hat37° C before reintroduction of serum-containing MEM. Cells were transfected either with the mUT-A3-EGFP fusion protein construct, p-EGFP-ROMK2 ( 15 ) or with p-EGFP-C2 alone. Membrane localization of mUT-A3-EGF protein was determined 48 h after transfection as follows. Monolayers were washed three times in sterile PBS, fixed for 10 min in a mixture of methanol and acetone (7:3) preequilibrated at -20° C, and then dried for 10 min before rehydrating in PBS. Monolayers were then washed with PBS and incubated in 1 mg/ml ribonuclease A (Sigma) in PBS at pH 7.0 for 10 min and then washed again in PBS. Nuclei were stained by incubation with 5 µg/ml propidium iodide (Sigma) in PBS for 10 min, rinsed in PBS, and allowed to dry at ambient temperature. The filters were removed from their inserts using a clean scalpel blade and mounted in Vectashield (Vector Laboratories, Peterborough, UK) on 1-mm-thick Super Premium microscope slides (VWR). Slides were stored at 4° C in the dark. Confocal images were obtained using a Zeiss LSM 510 META upright confocal microscope with a Plan-Apochromat x 63/1.4 oil immersion objective set for dual wavelength ( Ex = 488- and 543-nm) excitation and image acquisition.

RESULTS

Characterization of affinity-purified mouse UT-A-targeted antisera. Based on knowledge of the characterized mouse UT-A isoforms expressed in the inner medulla and the targeted epitopes, we predicted that ML446 and MQ2 would both detect mUT-A1 and mUT-A3, whereas ML194 would detect mUT-A1 ( Fig. 1 ). Immunoblots of papillary tip (PT) protein homogenates were probed with all three UT-A antibodies ( Fig. 2 A ). ML446 detected bands at 89 and 119 kDa and a 45- to 65-kDa signal. These were interpreted as representing UT-A1 in two different glycosylation states ( 2, 5 ) and UT-A3 ( 6, 28 ), respectively. In contrast, although MQ2 strongly detected the 45- to 65-kDa mUT-A3 signal, it did not detect mUT-A1 (i.e., was selective for mUT-A3). Furthermore, after total inner medulla protein was incubated at 37° C for 1 h both with and without PNGaseF enzyme ( Fig. 2 B ), the resulting immunoblots showed that the original 45- to 65-kDa smear detected by MQ2 was reduced by deglycosylation to a basic 40-kDa protein. This showed, as previously reported in the rat, that the 45- to 65-kDa band represented differentially glycosylated UT-A3 ( 6, 28 ). ML194 only detected bands at 89 and 119 kDa corresponding to mUT-A1. As predicted, this antiserum also detected 43-55 kDa corresponding to mUT-A2 in the ISOM ( Fig. 3 C ). All positive signals were ablated with preincubation of the antibody in the original immunizing peptide, indicating that these signals were specific for the targeted epitope.

Fig. 2. Characterization of affinity-purified mouse UT-A-targeted antisera. A : Western analysis of papillary tip protein showing representative immunoblots probed with ML446, MQ2, or ML194. ML446 detected bands at 89 and 119 kDa and a 45- to 65-kDa signal. These were interpreted as representing UT-A1 in 2 different glycosylation states and mUT-A3, respectively. MQ2 strongly detected the 45- to 65-kDa signal corresponding to mUT-A3, importantly it did not detect mUT-A1. ML194 detected bands at 89 and 119 kDa corresponding to mUT-A1. No signals were observed following preabsorption of antisera with immunizing peptides (+peptide), using a concentration of 1 mg of peptide per 10 ml. B : deglycosylation of mUT-A3 protein. Immunoblot of NMR1 whole inner medulla protein (5 µg/lane), treated at 37° C for 1 h with (+) or without (-) PNGaseF, probed with MQ2. The control mUT-A3 signal of 45-65 kDa was reduced by deglycosylation to a basic 40-kDa protein.

Fig. 3. Immunoblotting of dissected NMR1 mouse kidney (5 µg/lane) divided into papillary tip (PT), middle inner medulla (MIM), base inner medulla (BIM), inner stripe of outer medulla (ISOM), outer stripe of outer medulla (OSOM), cortex (C), and superficial cortex (SC). A : ML446 detects mUT-A1 (89 and 119 kDa) and mUT-A3 (45-65 kDa) throughout the PT, MIM, and BIM regions. B : MQ2 detects mUT-A3 in the PT, MIM, and BIM regions. C : ML194 detects mUT-A1 in the PT, MIM, and BIM regions and also mUT-A2 (43-55 kDa) in the ISOM.

Immunoblotting. To study the general pattern of expression of UT-A proteins in the kidney, we performed semiquantitative Western blot analysis on microdissected NMR1 mouse kidney. Kidneys were divided into seven different regions: PT, middle inner medulla (MIM), base inner medulla (BIM), ISOM, outer stripe of outer medulla, cortex, and superficial cortex. Immu-noblots of all kidney regions showed ML446 detected mUT-A1 and mUT-A3 in PT, MIM, and BIM lanes ( Fig. 3 A ). The strongest signals were in the PT, and these decreased toward the BIM. ML194 detected a similar pattern of bands that were of comparable intensity to the mUT-A1 signal detected by ML446. ML194 did not detect mUT-A3 in any region analyzed but detected a 43- to 55-kDa signal representing UT-A2 in the ISOM ( 29 ). The other two antibodies did not detect this signal. MQ2 recognized a 45- to 65-kDa mUT-A3 band in the same kidney sections as ML446 but did not detect mUT-A1 or mUT-A2 in any kidney sample. In summary, immunoblot analysis showed that 1 ) ML446 detected mUT-A1 and mUT-A3, 2 ) MQ2 was selective for mUT-A3, and 3 ) ML194 detected mUT-A1 and mUT-A2. It is also worthy of note that neither ML194 nor ML446, which are predicted to recognize UT-A4, detected a 50- to 55-kDa band, the predicted size of UT-A4. This suggests that UT-A4 may not be expressed in mouse kidney or, as is thought to be the case in rat, the level of expression may be very low and is below the detection limits of Western blot analysis.

Immunolocalization. ML446 and MQ2 stained only collecting ducts in the inner medulla, with no staining present in any other kidney region ( Fig. 4 ). ML446 staining was strong in PT and MIM and only partially stained BIM. MQ2 staining was strongest in PT, decreased in MIM, and was very weak in BIM. Similar to ML446, ML194 strongly stained collecting ducts in PT and MIM and only the lower part of BIM. In contrast to the other antisera, ML194 also stained structures in the ISOM ( Fig. 4 A, iv ). Based on the selectivity of our antibodies detailed above, our previous studies and those of others in the rat, these results were as predicted and confirm UT-A1 ( 14, 20, 22 ) and UT-A3 ( 22, 28 ) are present in IMCD cells, and UT-A2 is present in the ISOM ( 20, 29 ). High-magnification images of ISOM showed that UT-A2 was expressed on apical and basolateral membranes thin descending limbs ( Fig. 4 B, iv ).

Fig. 4. Immunolocalization of UT-A proteins in NMR1 mouse kidney. Immunoperoxidase labeling was performed on paraffin-embedded NMR1 mouse kidney sections (5 µM) stained with ML446, MQ2, or ML194. A : low magnification ( x 50). (i) : Secondary antibody only control. (ii) : ML446 stained PT, MIM, and lower part of BIM. (iii) : MQ2 stained PT and MIM and only weakly stained BIM. (iv) : ML194 stained PT, MIM, and lower part of BIM and also stained the ISOM. B : high magnification ( x 630) of ISOM ( i - iv ) and PT-MIM junction ( v - viii ). (i) : Secondary antibody only control. (ii) : ML446 and MQ2 (iii) did not stain structures in the ISOM, whereas ML194 (iv) stained thin descending limbs of ISOM, both apical membranes (open arrowheads) and basolateral membranes (filled arrowheads). (v) : Secondary antibody only control. (vi) : ML446 strongly stained basolateral membranes (filled arrowheads) and weakly stained in the apical membrane region (open arrowheads) and cell interior of IMCD cells. (vii) : MQ2 strongly stained basolateral membranes and the cell interior of IMCD cells. Because MQ2 is selective for UT-A3, these data suggest that mUT-A3 is expressed in the basolateral plasma membrane of IMCD cells. (viii) : ML194 stained basolateral membranes and showed weak cellular and apical membrane staining.

The subcellular patterns of staining in the IMCD were complex ( Fig. 4 B, v - viii ). In NMR1 mice, ML446 showed cellular staining and also strongly stained basolateral membranes. In some IMCD cells, weak apical membrane staining was also evident. MQ2 weakly stained cellular components and strongly stained basolateral membranes but importantly did not appear to stain apical membranes. Because MQ2 is selective for UT-A3, these data suggest that mUT-A3 is expressed in the basolateral plasma membrane of IMCD cells. ML194 stained basolateral membranes as well as apical membrane staining and some cellular staining. We conclude that the apical staining revealed by ML194 and ML446 most likely corresponds to mUT-A1. The finding that ML194 also stains basolateral membranes indicates that a protein containing the UT-A1 COOH-terminal epitope is also expressed in the basolateral membrane. At this stage, we cannot discern whether this signal represents UT-A1 or UT-A4, but our Western blot analysis suggests that UT-A1 is the predominant isoform in this region. In addition, using Western blot analysis, we were unable to detect any protein approximating to the molecular mass predicted for UT-A4 (50-55 kDa).

To further confirm our findings, we carried out a series of experiments in C57/BL6J mice to investigate whether the pattern of staining was altered in a different mouse strain. Interestingly, the pattern of IMCD staining in C57/BL6J mice ( Fig. 5 ) was basically very similar to that which we observed in NMR1 mice. MQ2 again strongly stained basolateral membranes, confirming the basolateral expression of mUT-A3. In addition, similar to NMR1 results, the basolateral membrane was also stained by ML446 and, although only weakly, ML194. MQ2 also stained the apical membrane and subapical region in some IMCD cells. This apparent greater apical staining was observed with ML446 and ML194 antisera as well and maybe reflects a greater expression of urea transporters overall in the apical IMCD membranes of C57/BL6J mice.

Fig. 5. Immunolocalization of UT-A proteins in C57/BL6J mouse kidney inner medullary collecting ducts (IMCDs). Immunoperoxidase labeling was performed on paraffin-embedded C57/BL6J mouse kidney sections (5 µM) stained with ML446, MQ2, or ML194. High magnification ( x 630; i ) ML446 stained both apical membranes (open arrowheads) and basolateral membranes (filled arrowheads). (ii) : MQ2 predominantly stained basolateral membranes, in a similar pattern to that observed in NMR1 IMCD cells. MQ2 also stained the apical region in a minority of IMCD cells. (iii) : ML194 staining was predominantly located on the apical membranes of IMCD cells.

Subcellular fractionation experiments. We confirmed that the 17,000- g fraction contained plasma membranes by performing Western blot analysis using an established Na-K-ATPase 1 -subunit antibody as a basolateral membrane marker. This analysis showed a 100-kDa Na-K-ATPase signal in the 17,000- g plasma membrane fraction ( Fig. 6 ). This was not detected in the 200,000- g pellet containing intracellular vesicles or the supernatant fraction containing cytosolic proteins.

Fig. 6. Immunoblotting of NMR1 whole inner medulla protein (5 µg/lane) after serial centrifugation. Each sample contains different subcellular fractions. 17,000 g : Plasma membrane enriched. 200,000 g : Intracellular vesicles. Supernatant: cytosolic proteins. This classification was confirmed using a monoclonal antibody raised against the plasma membrane-bound protein Na-K-ATPase. As expected, ML446 detects mUT-A1 and mUT-A3, MQ2 only detects mUT-A3, and ML194 detects mUT-A1 alone. In all cases, signals were strongest in the 17,000- g samples, weaker in the 200,000- g samples, and completely absent from the supernatant samples.

Using ML446, immunoblot analysis of serially centrifuged NMR1 whole inner medulla samples showed that both mUT-A1 and mUT-A3 were found predominantly in the 17,000- g plasma membrane-enriched fraction ( Fig. 6 ). Although signals were also apparent in the 200,000- g intracellular vesicle fraction, none were present in the supernatant containing cytosolic proteins. These patterns of expression were confirmed for both mUT-A1 and mUT-A3, using ML194 and MQ2, respectively. These results suggest that UT-A1 and UT-A3 in mouse IMCD cells are predominantly located in plasma membranes.

Immunoelectron microscopy of mUT-A3 in IMCD cells in mouse kidney. Immunoelectron microscopy was conducted using immunogold labeling of sections prepared from NMR1 kidney tissues embedded in Lowicryl HM20 by cryosubstitution. In sections from the middle part of inner medulla, immunogold labeling of mUT-A3 was observed in the basal part of the IMCD cells. As shown in Fig. 7 B, IMCD cells exhibited immunogold labeling of the basal plasma membrane (arrows) and intracellular vesicles in the basal part of the cell (arrowheads). In contrast, Fig. 7 C shows that although intracellular vesicles in the apical part of the cell were labeled (arrowheads), the apical plasma membrane was completely unlabeled. Immunolabeling controls using antibody preabsorbed with excess immunizing peptide produced no labeling (not shown).

Fig. 7. Immunoelectron microscopy of IMCD cells in mouse kidney. A : survey view of section of an IMCD; dashed and solid rectangles indicate the areas present at higher magnification in B and C, respectively. B : immunogold labeling of mUT-A3 is seen in the basal plasma membrane (arrows) and intracellular vesicles in the basal part of the cell (arrowheads). C : immunogold labeling of mUT-A3 observed in intracellular vesicles in the apical part of the cell (arrowheads), but none is seen in the apical plasma membrane. BM, basement membrane. Magnification: A : x 8,200. B : x 53,000. C : x 27,000.

Heterologous expression of mUT-A3-EGFP in MDCK II cells. To verify that mUT-A3 is targeted to basolateral membranes, we heterologously expressed mUT-A3 as an endogenously fluorescent EGF fusion protein in the polarized MDCK II epithelial cell line. Transient expression of pEGF protein alone resulted in fluorescence distributed evenly throughout the cell with no distinct plasma membrane labeling (not shown). Transfection with p-EGFP-ROMK2, the secretory potassium channel ROMK2 used here as an apical membrane targeting control, showed apical fluorescence ( Fig. 8 ). This finding was identical to that reported previously by us for the high-resistance strain MDCK I ( 15 ). In contrast, transient expression of pEGFP-mUT-A3 gave fluorescence that was localized to the lateral and basal cellular domains ( Fig. 8 ). These data show that mUT-A3 heterologously expressed in MDCK II cells is targeted to the basolateral membrane. This targeting is analogous to the immunostaining we observed using mUT-A3-selective antisera.

Fig. 8. Representative confocal images of Madin-Darby canine kidney II cells grown on filters transiently expressing enhanced green fluorescent (EGF) fusion proteins. For each monolayer of cells, single sections taken at the top ( i ), middle ( ii ), and bottom ( iii ) of the monolayer. Also shown are the corresponding XZ projections ( iv ), with the basolateral membrane orientated toward the bottom. Left : mUT-A3-EGFP. Right : ROMK2-EGFP. Cell nuclei were counter-stained with propidium iodide (red). For all panels, the scale bar where shown is 5 µm. Basolateral localization of mUT-A3 contrasts with both ROMK2, which is expressed at the apical cell pole, and EGFP which is not preferentially sorted (data not shown).

DISCUSSION

In the IMCD, AVP, acting through V 2 receptors, causes intracellular cAMP levels to increase and this stimulates an increase in IMCD urea permeability ( 17, 25 ). Consequently, urea is transported from the nephron lumen, across the IMCD epithelium, and into the deep inner medulla. This process, along with NaCl emanating from the action of the apical bumetanide-sensitive Na-K-Cl cotransporter in the thick ascending limb, creates a hypertonic medullary compartment that provides the driving force for water reabsorption from collecting ducts through aquaporin water channels. Urea transporters present in the thin descending limbs of the loop of Henle and descending vasa recta function to maintain a high medullary urea concentration and to recycle urea back into the nephron ( 11, 18, 26 ), thus playing an essential role in the counter-current multiplication system.

Two UT-A isoforms, UT-A1 and UT-A3, are strongly expressed in the IMCD. In the rat, UT-A1 has been immunolocalized to the apical membrane and intracellular compartment of IMCD cells ( 14 ). Initially, because of the relatively high cellular expression of UT-A1, it was thought that this protein might be trafficked to the apical IMCD membrane, bringing about the large increase in urea permeability observed in isolated, perfused IMCDs after administration of AVP ( 17 ). However, unlike aquaporin-2, Inoue et al. ( 9 ) found no evidence of UT-A1 trafficking to the apical membrane in rat IMCD following stimulation with AVP for 20 min.

Based on the current literature, it seems plausible that an increase in intracellular cAMP triggers direct phosphorylation of UT-A1, possibly via the classical cAMP target protein kinase A, which in turn activates UT-A1 ( 32 ) and leads to the observed increase in transepithelial urea flux.

The basolateral membrane urea pathway, although unresolved, has been assumed to be a UT-A protein because in vitro perfusion studies on rat tubules showed that addition of phloretin to the basolateral side of isolated IMCDs blocked transepithelial urea flux ( 3 ). UT-A3 is also highly expressed in the IMCD and like UT-A1 is also acutely regulated by PKA agonists ( 5, 21 ). However, its precise role and regulation have been the subject of contrasting reports ( 22, 28 ). Therefore, the aim of this study was to determine the location and possible role of UT-A3 in the mouse.

To ascertain this, we raised and characterized three antisera targeted to mouse UT-A proteins. As predicted, the antiserum raised to NH 2 terminus of UT-A1 recognized UT-A1 and UT-A3, and the antiserum targeting the COOH termini of UT-A1 recognized UT-A1 and UT-A2. Although both NH 2 - and COOH-termini antisera would be predicted to recognize UT-A4, throughout our studies we saw no conclusive evidence of UT-A4 protein expression in mouse kidney, suggesting that its expression level may be below the detection limits of our immunoblot analysis.

In contrast to ML194 and ML446, MQ2 was shown to be selective for mUT-A3. This was confirmed by the finding that the diffuse 45- to 65-kDa band was reduced by deglycosylation to a 40-kDa protein, as has previously been reported for rat UT-A3 ( 28 ). The reason MQ2 did not detect mUT-A1 remains unclear despite the fact that, with the exception of the terminal aspartic acid residue of UT-A3, both UT-A1 and UT-A3 contain the targeted epitope. Terris and colleagues ( 28 ) using a shorter immunizing peptide targeting 447-460 at the COOH terminus of rat UT-A3 reported very similar results that corroborate our findings. It is possible that the extra amino acid at the COOH terminus of UT-A3 present in both rat and mouse may be responsible for the inability of MQ2 and its rat analog to bind to UT-A1, or perhaps the folding of native mUT-A1 protein may potentially "mask" the MQ2 epitope.

Western blot analysis with both MQ2 and ML446 revealed that mUT-A3 expression was most abundant in the PT, decreased in the MIM, and was weakest in the base of the inner medulla. This pattern of expression was paralleled by the pattern of UT-A1 expression. Immunolocalization studies agreed with these data, although the MQ2 mUT-A3 signal was weaker than expected in the middle IMCD and undetectable in the base IMCD. This finding most likely reflects the relatively poor efficacy of MQ2 on immunolocalization slides, compared with immunblots, and the resulting reduction of all signals, rather than any differences in the pattern of mUT-A3 expression.

The overall renal distribution for mUT-A1, mUT-A2, and mUT-A3, detailed in this study with both immunoblotting and immunolocalization, strongly agrees with that previously reported in the rat ( 28 ) and also with a report by Shayakul et al. ( 22 ), in which in situ hybridization studies showed a strong rat UT-A3 (referred to as UrT1-C) mRNA signal in the PT. The location of UT-A1 and UT-A3 within the IMCD suggests they may play roles in a common pathway.

Data from light microscopy in two separate strains of mice (NMR1 and C57/BL6J), electron microscopy, and heterologous expression studies show that mUT-A3 is predominantly expressed in basolateral membranes of IMCD cells. Staining with the UT-A3-selective antisera clearly showed basolateral staining in both strains of mice. However, some evidence indicated slight differences between strains in that MQ2 infrequently stained apical membranes in C57/BL6J mice but did not stain apical membranes in NMR1 mice. This may indicate that in C57/BL6J, UT-A3 can also traffic to the apical membrane. The observed weak apical staining in NMR1 mice we saw with ML446 and ML194 we interpreted as representing UT-A1. This confirmed previous studies in rat showing UT-A1 to be expressed apically ( 14 ). We propose that UT-A1 and UT-A3 together represent a phloretin-inhibitable urea pathway that exists across the IMCD epithelia. It is possible that these proteins are expressed in series with UT-A1 predominantly expressed on the apical membrane and UT-A3 expressed predominantly on the basolateral membrane.

This urea pathway has been extensively studied in rat IMCD where transepithelial urea permeability is blocked by phloretin applied to either the apical or basolateral membranes ( 3 ). Whether, as these data suggest, a UT-A transporter is present in the basolateral IMCD membrane of the rat requires confirmation. In an attempt to address this question, Terris et al. ( 28 ) used an antiserum targeted to the same region of rat UT-A3 as our MQ2 antiserum, reported only cellular staining, and did not detect protein on either the apical or basolateral membranes. The disparity between our findings in the mouse and those of Terris and colleagues in the rat may reflect real differences between these two species. However, using immunoblot analysis Terris et al. detected strong expression of UT-A3 in plasma membrane-enriched IMCD fractions, indicating that in rat UT-A3 is likely to be expressed in a plasma membrane. It is unclear from these data whether this represents apical or basolateral expression ( 28 ). In light of these data, our favored alternative explanation for the disparity is that the antiserum used by Terris and colleagues did not detect plasma membrane-bound protein when used for immunocytochemistry.

Our data also suggest that another UT-A isoform may be expressed in the basolateral membrane. This suggestion is based on the finding that ML194 stains basolateral membranes in NMR1 and C57/BL6J. In addition, using an analogous antiserum to ML194, which targeted the COOH terminus of rat UT-A1, we previously reported a very similar pattern of staining in MF1 mice ( 5 ). Therefore, in three different mouse strains using antisera confirmed to target UT-A1, we observed basolateral membrane staining. Western blot analysis shows that only the two glycosylated forms of UT-A1 are expressed in the region corresponding to this staining; hence it seems likely that in mouse UT-A1 is also expressed in the basolateral membrane.

We previously showed that mUT-A3 is acutely responsive to cAMP ( 5 ). This, and the observation reported in the current study that mUT-A3 is expressed in the basolateral membrane of IMCD cells, has important implications for control of IMCD permeability in response to AVP. Our data suggest that the urea permeability of the basolateral membrane is acutely sensitive to cAMP. This suggestion is both novel and interesting because it implies that AVP causes an increase in both apical membrane urea permeability, mediated via UT-A1, and basolateral urea permeability, mediated by UT-A3 and possibly UT-A1.

We conclude that mUT-A3 is strongly expressed in basolateral membranes of mouse kidney IMCD cells. This basolateral location for mUT-A3 suggests an important role in the transcellular transport of urea across IMCD epithelia. This isoform would therefore play an integral part in the urinary concentration mechanism.

ACKNOWLEDGMENTS

The authors thank Z. Nikrozi for expert technical assistance and Dr. G. Howell of the Leeds Bioimaging Facility (supported by The Wellcome Trust) for advice on use of the confocal microscope.

GRANTS

We thank The Royal Society, the Biotechnology and Biological Sciences Research Council, and National Kidney Research Fund (UK) for funding this work. The Water and Salt Research Center at the University of Aarhus is established and supported by the Danish National Research Foundation (Danmarks Grundforskningsfond). Further support for this study was provided by the Karen Elise Jensen Foundation, the Human Frontier Science Program, the European Commission (KA 3.1.2 and KA 3.1.3 programmes), the Novo Nordic Foundation, the Danish Medical Research Council, the University of Aarhus Research Foundation, and the University of Aarhus. V. M. Collins is supported by a "White-Rose" collaborative studentship between the Universities of Leeds and Sheffield.

【参考文献】
  Bagnasco SM, Peng T, Janech MG, Karakashian A, and Sands JM. Cloning and characterization of the human urea transporter UT-A1 and mapping of the human Slc14a2 gene. Am J Physiol Renal Physiol 281: F400-F406, 2001.

Bradford AD, Terris JM, Ecelbarger CA, Klein JD, Sands JM, Chou CL, and Knepper MA. 97- And 117-kDa forms of collecting duct urea transporter UT-A1 are due to different states of glycosylation. Am J Physiol Renal Physiol 281: F133-F143, 2001.

Chou CL and Knepper MA. Inhibition of urea transport in inner medullary collecting duct by phloretin and urea analogues. Am J Physiol Renal Fluid Electrolyte Physiol 257: F359-F365, 1989.

Couriaud C, Ripoche P, and Rousselet G. Cloning and functional characterization of a rat urea transporter: expression in the brain. Biochim Biophys Acta 1309: 197-199, 1996.

Fenton R, Stewart G, Carpenter B, Howorth A, Potter E, Cooper G, and Smith C. Characterization of mouse urea transporters UT-A1 and UT-A2. Am J Physiol Renal Physiol 283: F817-F825, 2002.

Fenton RA, Cooper GJ, Morris ID, and Smith CP. Coordinated expression of UT-A and UT-B urea transporters in rat testis. Am J Physiol Cell Physiol 282: C1492-C1501, 2002.

Fenton RA, Howorth A, Cooper GJ, Meccariello R, Morris ID, and Smith CP. Molecular characterization of a novel UT-A urea transporter isoform (UT-A5) in testis. Am J Physiol Cell Physiol 279: C1425-C1431, 2000.

Ferguson CJ, Wareing M, Ward DT, Green R, Smith CP, and Riccardi D. Cellular localization of divalent metal transporter DMT-1 in rat kidney. Am J Physiol Renal Physiol 280: F803-F814, 2001.

Inoue T, Terris J, Ecelbarger CA, Chou CL, Nielsen S, and Knepper MA. Vasopressin regulates apical targeting of aquaporin-2 but not of UT1 urea transporter in renal collecting duct. Am J Physiol Renal Physiol 276: F559-F566, 1999.

Karakashian A, Timmer RT, Klein JD, Gunn RB, Sands JM, and Bagnasco SM. Cloning and characterization of two new isoforms of the rat kidney urea transporter: UT-A3 and UT-A4. J Am Soc Nephrol 10: 230-237, 1999.

Kokko J and Rector F. Countercurrent multiplication system without active transport in inner medulla. Kidney Int 2: 214-223, 1972.

Lucien N, Sidoux-Walter F, Olives B, Moulds J, Le Pennec PY, Cartron JP, and Bailly P. Characterization of the gene encoding the human Kidd blood group/urea transporter protein. Evidence for splice site mutations in Jknull individuals. J Biol Chem 273: 12973-12980, 1998.

Nakayama Y, Naruse M, Karakashian A, Peng T, Sands JM, and Bagnasco SM. Cloning of the rat Slc14a2 gene and genomic organization of the UT-A urea transporter. Biochim Biophys Acta 1518: 19-26, 2001.

Nielsen S, Terris J, Smith CP, Hediger MA, Ecelbarger CA, and Knepper MA. Cellular and subcellular localization of the vasopressin-regulated urea transporter in rat kidney. Proc Natl Acad Sci USA 93: 5495-5500, 1996.

Ortega B, Millar I, Beesley A, Robson L, and White S. Stable, polarised, functional expression of Kir1.1b channel protein in Madin-Darby canine kidney cell line. J Physiol 528: 5-13, 2000.

Sands J. Mammalian urea transporters. Annu Rev Physiol 65: 543-566, 2003.

Sands J and Knepper MA. Urea permeability of mammalian inner medullary collecting duct system and the papillary surface epithelium. J Clin Invest 79: 138-147, 1987.

Sands JM and Kokko JP. Current concepts of the countercurrent multiplication system. Kidney Int Suppl 57: S93-S99, 1996.

Sands JM, Timmer RT, and Gunn RB. Urea transporters in kidney and erythrocytes. Am J Physiol Renal Physiol 273: F321-F339, 1997.

Shayakul C, Knepper MA, Smith CP, DiGiovanni SR, and Hediger MA. Segmental localization of urea transporter mRNAs in rat kidney. Am J Physiol Renal Physiol 272: F654-F660, 1997.

Shayakul C, Steel A, and Hediger MA. Molecular cloning and characterization of the vasopressin-regulated urea transporter of rat kidney collecting ducts. J Clin Invest 98: 2580-2587, 1996.

Shayakul C, Tsukaguchi H, Berger UV, and Hediger MA. Molecular characterization of a novel urea transporter from kidney inner medullary collecting ducts. Am J Physiol Renal Physiol 280: F487-F494, 2001.

Smith CP, Lee WS, Martial S, Knepper MA, You G, Sands JM, and Hediger MA. Cloning and regulation of expression of the rat kidney urea transporter (rUT2). J Clin Invest 96: 1556-1563, 1995.

Smith CP and Rousselet G. Facilitative urea transporters. J Membr Biol 183: 1-14, 2001.

Star R, Nonoguchi H, Balaban R, and Knepper M. Calcium and cyclic adenosine monophosphate as second messangers for vasopressin in the rat inner medullary collecting duct. J Clin Invest 81: 1879-1888, 1988.

Stephenson J. Concentration of urine in a central core model of renal counterflow system. Kidney Int 2: 85-94, 1972.

Terris J, Ecelbarger CA, Sands JM, and Knepper MA. Long-term regulation of renal urea transporter protein expression in rat. J Am Soc Nephrol 9: 729-736, 1998.

Terris JM, Knepper MA, and Wade JB. UT-A3: localization and characterization of an additional urea transporter isoform in the IMCD. Am J Physiol Renal Physiol 280: F325-F332, 2001.

Wade JB, Lee AJ, Liu J, Ecelbarger CA, Mitchell C, Bradford AD, Terris J, Kim GH, and Knepper MA. UT-A2: a 55-kDa urea transporter in thin descending limb whose abundance is regulated by vasopressin. Am J Physiol Renal Physiol 278: F52-F62, 2000.

Yang B, Bankir L, Gillespie A, Epstein CJ, and Verkman AS. Urea-selective concentrating defect in transgenic mice lacking urea transporter UT-B. J Biol Chem 277: 10633-10637, 2002.

You G, Smith CP, Kanai Y, Lee WS, Stelzner M, and Hediger MA. Cloning and characterization of the vasopressin-regulated urea transporter. Nature 365: 844-847, 1993.

Zhang C, Sands JM, and Klein JD. Vasopressin rapidly increases phosphorylation of UT-A1 urea transporter in rat IMCDs through PKA. Am J Physiol Renal Physiol 282: F85-F90, 2002.

作者单位：1 School of Biological Sciences, University of Manchester, Manchester M13 9PT; 4 School of Biomedical Sciences, University of Leeds, Leeds LS2 9JT; 5 Department of Biomedical Sciences, University of Sheffield, Sheffield S10 2TN, United Kingdom; 2 Laboratory of Kidney and Electrolyte Metabolism, Nati