Abstract

The renal urea transporter gene (UT-A) produces different transcripts in the inner medullary collecting ducts (UT-A1) and thin descending limbs of Henle's loop (UT-A2), coding for distinct proteins. Peptide-directed rabbit polyclonal antibodies were used to identify the UT-A2 protein in renal medulla of mouse and rat. In the inner stripe of outer medulla, an antibody directed to the COOH terminus of UT-A recognized a membrane protein of 55 kDa. The abundance of this 55-kDa protein was strongly increased in response to chronic infusion of the vasopressin analog 1-deamino-[8-d-arginine]vasopressin (DDAVP) in Brattleboro rats, consistent with previous evidence that UT-A2 mRNA abundance is markedly increased. Immunofluorescence labeling with the COOH-terminal antibody in Brattleboro rats revealed labeling in the lower portion of descending limbs from short-looped nephrons (in the aquaporin-1-negative portion of this segment). This UT-A labeling was increased in response to DDAVP. Increased labeling was also seen in descending limbs of long-looped nephrons in the base of the inner medulla. These results indicate that UT-A2 is expressed as a 55-kDa protein in portions of the thin descending limbs of Henle's loop and that the abundance of this protein is strongly upregulated by vasopressin.

  • loop of Henle
  • urinary concentrating mechanism
  • urea channel
  • aquaporin-1

urea transport in the renal medullary loop of Henle and collecting duct is critically important to the urine concentrating process and to the regulation of renal water excretion (11, 28). Our understanding of the physiological basis of urea transport in renal medullary epithelia has been greatly advanced by the identification of two distinct epithelial urea transporters via molecular cloning (26,27, 35). These two transporters are transcribed from the same gene from two distinct promoters. This gene has been referred to as “UT-A” to distinguish it from the gene coding for a urea transporter expressed in erythrocytes and endothelial cells, termed “UT-B” (24). Of these two UT-A isoforms (Fig. 1), the larger of the two, termed “UT-A1” (originally called “UT1”; Ref. 26), has an open reading frame of 929 amino acids and is expressed solely in the inner medullary collecting duct (25). A second isoform, termed “UT-A2” (originally called “UT2”; Refs. 27 and 35) corresponds to the COOH-terminal 397 amino acids of UT-A1. RT-PCR studies in microdissected renal tubules (25) found UT-A2 mRNA in the descending thin limbs of short loops of Henle in outer medulla and in descending limbs of long loops of Henle in inner medulla. UT-A2 expression appears to be absent in descending limbs of long loops in outer medulla. Northern blotting studies have revealed that the abundance of the UT-A2 transcript is increased in rats in response to water restriction (27) and by vasopressin infusion (21).1

Fig. 1.

Antibodies to urea transporters. Immunization sites for peptide-directed rabbit polyclonal urea transporter antibodies L446, L448, L403, and L194 relative to UT-A1 sequence (top) and UT-A2 sequence (bottom). UT-A1 is a 929-amino acid protein containing four 135 amino regions of exceptionally high hydrophobicity (labeled H1-H4). UT-A2 (bottom) is identical to the terminal 397 amino acids of UT-A1 (amino acid numbering is based on the UT-A1 sequence). L446 is directed to a portion of the amino-terminal tail of UT-A1 (amino acids 56–78). L448 is directed against a region in the middle cytoplasmic loop and recognizes UT-A1 but not UT-A2 (31). L403 (31) is directed against the COOH terminus and recognizes both UT-A1 and UT-A2. L194 is a different COOH-terminal antibody used in a previous study of urea transporter localization in the kidney (18). [Recently, two additional UT-A cDNAs have been cloned (10). One of them, termed “UT-A3,” comprises amino acids 1–460. The second cDNA, termed “UT-A4”, corresponds to amino acids 1–274 and amino acids 738–929 of UT-A1 spliced together as a single polypeptide.]

Urea transporter localization in the renal medulla has also been carried out by immunochemical approaches. Previous immunochemical studies of UT-A isoforms (18) utilized a peptide-directed antibody, termed L194, raised to the carboxy-terminal 19 amino acids of UT-A1, whose carboxy terminus is identical to that of UT-A2. This antibody labeled both the inner medullary collecting ducts and the thin descending limbs of Henle's loop in the inner stripe of the outer medulla, consistent with the expectation that both UT-A1 and UT-A2 would be recognized by the antibodies. The thin descending limbs labeled were identified as thin descending limbs of short-looped nephrons based on the demonstration that the labeled structures undergo a transition to thick ascending limbs in the outer medulla. Immunoblots run with membrane fractions from the inner medulla revealed a predominant band of 97 kDa, consistent with the expected molecular weight of UT-A1 (18). However, inexplicably, immunoblots run using membrane fractions from the outer medulla did not show a protein band of molecular weight consistent with that expected for UT-A2, despite extraordinarily strong immunocytochemical labeling of the descending limbs of short-looped nephrons. The inability to carry out immunoblotting studies of UT-A2 expression has hampered our ability to investigate how it is regulated. The present study describes approaches that allow recognition of UT-A2 on immunoblots and uses immunoblotting and immunocytochemical techniques to investigate UT-A2 localization and regulation in mouse and rat kidney.

METHODS

Experimental animals. All experiments were conducted in accord with animal protocols for rats and mice approved by the Animal Care and Use Committee of the National Heart, Lung, and Blood Institute. Pathogen-free male Sprague-Dawley rats (Taconic Farms, Germantown, NY) weighing 195–225 g and male Brattleboro (di/di) rats (Harlan Sprague Dawley, Indianapolis, IN) weighing 210–260 g were used. Pathogen-free male BALB/c mice (NCI/Frederick Cancer Research Facility, Frederick, MD, and Charles River) were also used. Rats and mice were kept in filter-top microisolators with autoclaved feed and bedding, to maintain a pathogen-free state.

Antibodies. To carry out immunoblotting and immunolocalization studies, we used peptide-directed polyclonal antibodies to UT-A sites depicted in Fig. 1. The antibodies (L403, L446, and L448) were affinity purified using columns on which the immunizing peptides were immobilized. For immunoblotting, the L403 antibody was used at an IgG concentration of 0.178 μg/ml. The L446 antibody was used at an IgG concentration of 0.383 μg/ml. An accounting of the UT-A isoforms expected to be recognized by these antibodies is given in Table1. All recognize UT-A1 in the inner medulla (Ref. 31 and unpublished data). To localize aquaporin-1, we used antibodies raised in chicken (LC18) to the same carboxy-terminal sequence used to prepare a rabbit polyclonal antibody against aquaporin-1 (30) (sequence: CGQVEEYDLDADDINSRVEMKPK; note addition of a cysteine residue at the amino terminal end of the peptide to facilitate conjugation). An antibody to von Willebrand factor raised in sheep was obtained from Cedarlane Labs (Ontario, Canada).

View this table:
Table 1.

Anti-UT-A antibodies used in this study

DDAVP infusion study. Under methoxyflurane anesthesia (Metofane; Pitman-Moore, Mundelein, IL), osmotic minipumps (model 2002; Alzet, Palo Alto, CA) were implanted subcutaneously in six Brattleboro rats to deliver 20 ng/h of 1-deamino-[8-d-arginine]vasopressin (DDAVP; Rhone-Poulenc Rorer, Collegeville, PA), a V2 receptor-selective agonist of vasopressin. Another six rats (controls) were implanted with minipumps containing vehicle (saline) alone. After 7 days of DDAVP or vehicle infusion, during which time they received ad libitum water and pelleted chow, the rats were killed and the kidneys were harvested for semiquantitative immunoblotting (described below).

Semiquantitative immunoblotting. Kidneys were dissected to obtain separate tissue samples for inner stripe of outer medulla (labeled “outer medulla”) and whole inner medulla (labeled “inner medulla”). The tissue was homogenized using a tissue homogenizer (Omni 1000 fitted with a micro-saw-toothed generator) in ice-cold isolation solution containing 250 mM sucrose/10 mM triethanolamine (Calbiochem, La Jolla, CA) with 1 μg/ml leupeptin (Bachem California, Torrance, CA) and 0.1 mg/ml phenylmethylsulfonyl fluoride (United States Biochemical, Toledo, OH). Total protein concentration was measured in the samples using a Pierce BCA protein assay reagent kit (Pierce, Rockford, IL). Total protein concentration was adjusted to approximately 1 μg/μl with isolation solution. Laemmli buffer, 5× (7.5% SDS, 30% glycerol, 1 M Tris, pH 6.8, bromophenol blue), with 30 mg/ml dithiothreitol was added to the samples in a volume ratio of 1:4 and heated to 60°C for 15 min to solubilize proteins. Equality of loading was assessed by running preliminary gels with all the samples (experimental vs. control; seeElectrophoresis and immunoblotting, below) and staining them with Coomassie blue dye. Densitometry of representative bands was used to compare actual loading among samples and to make minor corrections in loading prior to running SDS-PAGE for immunoblotting. To quantitate bands on immunoblots, we carried out densitometry using a laser densitometer (Molecular Dynamics, San Jose, CA) and ImageQuaNT software (Molecular Dynamics). To facilitate comparisons, we normalized the band density values by dividing by the mean value for the control group. Thus, the mean for the control group is defined as 100%. Normalized band densities for treated rats were compared with controls using an unpaired t-test when standard deviations were the same or by Welch t-test when standard deviations were significantly different (INSTAT; Graphpad Software, San Diego, CA). P < 0.05 taken as the criterion for statistical significance.

Differential centrifugation. Differential centrifugation was done in mouse outer medulla to determine the distribution of UT protein among the subcellular membrane fractions. Kidneys were dissected to obtain tissue samples from inner stripe of outer medulla. Tissue samples were homogenized using a tissue homogenizer (OMNI 1000 fitted with a micro-saw-toothed generator) in cold isolation solution. Homogenate samples were centrifuged sequentially at 1,000 g for 10 min, 4,000 g for 20 min, 17,000 g for 20 min, and then finally 200,000 g for 1 h. At each step, the pellet was saved and the supernatant was taken to the next step. After determination of total protein concentration, each of the pellets was solubilized with Laemmli sample buffer and was used for blots.

Electrophoresis and immunoblotting. SDS-PAGE was done on minigels of 12% polyacrylamide. The proteins were transferred from the gels electrophoretically to nitrocellulose membranes. After blocking with 5 g/dl nonfat dry milk, membranes were probed with the affinity-purified polyclonal antibody to the UT-A antibodies (at concentrations indicated above) in an antibody dilution buffer solution containing 150 mM NaCl, 50 mM sodium phosphate, 10 mg/dl sodium azide, 50 mg/dl Tween 20, and 1 g/dl BSA (pH 7.5). The secondary antibody was donkey anti-rabbit IgG conjugated to horseradish peroxidase (Pierce no. 31458) used at a concentration of 0.16 μg/ml. Sites of antibody-antigen reaction were visualized using luminol-based enhanced chemiluminescence (LumiGLO; Kirkegaard and Perry Laboratories, Gaithersburg, MD) before exposure to X-ray film (Kodak no. 165–1579 Scientific Imaging Film).

Immunocytochemistry. Tissue for immunocytochemistry was taken from 180- to 250-g male Sprague-Dawley rats, from 210- to 260-g male Brattleboro rats, or from pathogen-free male BALB/c mice. Kidneys were fixed with 2% paraformaldehyde (PF) by perfusion (2 min in PBS, 5 min in 2% PF, 2 min in a cryoprotectant of 10% EDTA in 0.1 M Tris). Fixed kidneys were sliced and further incubated in cryoprotectant for 60 min, wrapped in aluminum foil, and frozen on dry ice. Cryostat sections 12- to 15-μm thick were made and picked up on coverslips coated with HistoGrip (Zymed, San Francisco, CA). Sections were treated with 6 M guanidine for 10 min to uncover antigenic sites and washed three times with high-salt wash (50 ml PBS, 0.5 g BSA, and 1.13 g NaCl). A blocking agent for nonspecific binding sites (50 ml PBS, 0.5 g BSA, and 0.188 g glycine, pH 7.2) was then added to the sample for 20 min followed by the diluted primary antibody for overnight incubation at 4°C. Antibodies were diluted to 10 μg/ml with incubation medium (50 ml PBS, 0.05 g BSA, and 200 μl 5% NaN3). During the second day, these sections were rinsed three times for 5 min, one time for 15 min, and one time for 30 min with high-salt wash before incubation with the secondary antibody for 2 h at 4°C. Appropriate species-specific antibodies prepared specifically for multilabeling (Jackson Immunoresearch Labs, West Grove, PA) were diluted with incubation medium by 1:100. These samples were again washed five times with high-salt wash over the course of 1 h and then in PBS to remove the excess salt before mounting in Vectashield (Vector Labs, Burlingame, CA) and examination with a Zeiss LSM410 confocal microscope.

RESULTS

Identification of immunoreactive UT-A proteins in the mouse renal outer medulla. Because previous immunoblotting studies using outer medullary membrane fractions from rats failed to detect a protein with molecular mass compatible with UT-A2 (18), we prepared inner stripe outer medullary samples from another species, the mouse, in an attempt to detect the UT-A2 protein (Fig. 2). Figure 2 A shows results using whole homogenates from outer medullas of six different mice. In contrast with previous observations in rat, the mouse outer medullary blots demonstrated prominent labeling of a band at 55 kDa, with a second band at approximately 48 kDa. Both bands are in the approximate size range expected for UT-A2, a 397-amino acid protein, considering the fact that the molecular mass is likely to be augmented by glycosylation. As shown in Fig. 2 B, both bands were absent when the antibody was preincubated with an excess of the immunizing peptide (“specific peptide”) but not when it was preincubated with an unrelated peptide. In this deliberately overexposed immunoblot, weak peptide-ablatable bands were also seen at 97 kDa (presumably UT-A1) and at a lower apparent molecular mass (∼33 kDa), which may represent unglycosylated UT-A2 (seediscussion). As demonstrated in Fig. 2 C, when blots were probed with L446, a polyclonal antibody raised to the amino-terminal tail of UT-A1 (see Fig. 1), only the 48-kDa band was seen. This result suggests that the 55-kDa band corresponds to UT-A2, whereas the 48-kDa band corresponds to a distinct protein, possibly UT-A4 (see discussion). Differential centrifugation of a homogenate from the inner stripe of the outer medulla was able to produce some separation of the two bands, suggesting that they may occur in different structures (Fig. 2 D).

Fig. 2.

Immunoblots of mouse outer medullary proteins probed with antibodies to UT-A. A: outer medullary immunoblot showing labeling in mouse. Homogenates from inner stripes of outer medullas of each of 6 mice probed with antibody L403, directed to the carboxy-terminal tail of UT-A1 (Fig. 1). Each lane was loaded with 10 μg of outer medullary protein. There are prominent bands at 55 and 48 kDa. Note the absence of the 97-kDa band associated with UT-A1, as expected in outer medulla.B: preadsorption control. Immunoblots of the same sample of homogenate from inner stripe of mouse outer medulla (10 μg protein/lane) were probed with the L403 antibody preincubated with an unrelated peptide (sequence: CRIQRLEITGKKTTDEELEEMLES) (left) and with the specific peptide used for immunization of rabbits (right). Both the 55-kDa band and the 48- kDa band were ablated by preincubation with the specific peptide but not with the unrelated peptide. C: labeling with antibody directed to carboxy terminus of UT-A1 (L403) vs. antibody directed to amino terminus of UT-A1 (L446). Immunoblots of the same sample of homogenate from inner stripe of mouse outer medulla (10 μg protein/lane) were probed with L403 (left) or L446 (right). Only the 48-kDa band was recognized by the amino-terminal antibody. Failure of L446 to recognize the 55-kDa band suggests that this band corresponds to UT-A2 (cf. Fig.1). D: differential centrifugation of homogenate from inner stripe of mouse outer medulla: immunoblots probed with antibody L403. Most of the 48-kDa protein is found in the 1,000 g pellet (enriched in nuclei). Most of the 55-kDa protein is found in the 17,000g (microsomal) fraction, compatible with what is expected for a plasma membrane localization. Each lane was loaded with 10 μg of protein.

Immunofluorescence localization using the L403 COOH-terminal antibody in mouse outer medulla (Fig. 3) produced a labeling pattern similar to that previously observed in immunolabeling and in situ hybridization studies of rat outer medulla (18, 26, 27). Specifically, there is strong labeling in distinct longitudinal streaks in the inner stripe of the outer medulla, consistent with a urea transporter localization in the vascular bundle region. In rat, this pattern was found to be due to the presence of UT-A protein in descending limbs from short-loop nephrons (18), which are clustered around the vasa recta bundles. [The structure of the vascular bundles in mouse is similar, except that the descending limbs of short loops tend to be more integrated into the bundle (12, 13).]

Fig. 3.

Immunolocalization with antibody L403 in mouse. Inner stripe of the outer medulla is brightly labeled in a pattern corresponding to that of the vascular bundles where the thin descending limbs of short-looped nephrons are present, consistent with the localization seen in rat (18). Bar = 250 μm.

Immunofluorescence localization in rat outer medulla. The previous UT-A immunolocalization study showing labeling of thin descending limbs in the inner stripe of the outer medulla in rat was done using antibody L194 (see Fig. 1). To test whether the newer COOH-terminal antibody, L403, utilized in this study, also recognizes thin descending limbs in the inner stripe of rat outer medulla, we carried out immunofluorescence labeling experiments in rat outer medulla (Fig. 4). L403 labeling is shown in Fig. 4 A, which is compared with an adjacent section of rat outer medulla labeled with polyclonal antibody L448 (Fig. 4 B), which is directed to the middle loop of UT-A1 upstream from the start site of UT-A2 (see Fig. 1 and Table 1). In contrast to the very strong labeling by L403 of thin descending limbs in vascular bundles (Fig.4 A), L448, which is targeted to a site in UT-A1 upstream from the UT-A2 start site, does not label descending thin limbs of the loop of Henle significantly (Fig. 4 B).

Fig. 4.

Immunolocalization with UT-A antibodies in rat outer medulla: COOH-terminal antibody L403 (A) labels thin descending limbs of the loop of Henle associated with vascular bundles in the outer medulla as seen previously with antibody L194 (18). An antibody directed to a site upstream from the start site for UT-A2, L448 (B), fails to significantly label these structures in an adjacent section of the same region of the kidney. Bar = 250 μm.

To examine the specific vascular bundle structures labeled by urea transporter antibodies, we carried out triple labeling of oblique and transverse sections of the rat outer medulla (Fig.5). Figure 5 shows the structures labeled by COOH-terminal antibody to UT-A in green. Also shown are structures labeled with an antibody to aquaporin-1 (AQP1, red) and an antibody to von Willebrand factor (vWF, blue), an endothelial cell marker. This very low-magnification view shows a section that passes progressively from the inner medulla (IM) at the bottom through the inner stripe (IS) of the outer medulla in the middle until reaching the outer stripe (OS) of the outer medulla at the top of Fig. 5, where proximal tubules are being labeled by the antibody to aquaporin-1 (dense red labeling). Vascular elements shown in blue represent ascending vasa recta, which express von Willebrand factor but lack aquaporin-1. The pink structures are descending vasa recta, which have both von Willebrand factor and aquaporin-1. Thus the vascular bundles are readily recognized in the section. Two distinct vascular bundle regions (labeled VBa and VBb) can be noted on the basis of the presence or absence of urea transporter labeling (green). In vascular bundles of the lower portion of the inner stripe of the outer medulla (VBb), the blue-stained vasa recta are surrounded by structures containing abundant urea transporter protein (green) (see Fig. 5, inset) . These urea transporter-containing structures have been previously identified as descending limbs of short loops of Henle based on their continuity with thick ascending limbs in the outer medulla (18). Between the bundles are the thin descending limbs of long loops of Henle, which contain abundant aquaporin-1 (red) as previously demonstrated (16, 17). Thus, in the lower portion of the inner stripe of the outer medulla, the short-loop descending limbs express UT-A but not aquaporin-1, whereas the long-loop descending limbs express aquaporin-1 but not UT-A.

Fig. 5.

Immunolocalization of UT-A (COOH-terminal antibody) with respect to aquaporin-1 (AQP1) and von Willebrand factor (vWF). In this very low-magnification overview, a section from a control rat kidney is shown that passes from the inner medulla (IM) through the inner stripe (IS) of outer medulla to reach the outer stripe (OS) of outer medulla at top. Vascular elements (shown in blue) are labeled by an antibody to vWF, whereas structures labeled by an antibody to AQP1 are shown in red. Pink structures represent descending vasa recta, which have both vWF and AQP1. Labeling by COOH-terminal antibody to UT-A (in this case L194, but an identical pattern is seen with L403) is shown in green. The upper region of descending thin limbs (DTL) of short-looped nephrons (VBa) lacks UT-A labeling but do label with AQP1. Regions of short-looped nephron DTL close to the IM (VBb) are labeled by UT-A antibody, and these regions lack AQP1 labeling. Inset: cross section through a vascular bundle at the level where UT-A labeling is seen. Thin structures labeled by COOH-terminal UT-A antibodies (green) are arranged in the oval pattern around vasa recta (blue), with the aquaporin-1-labeled thin limbs in periphery. This distinctive anatomical arrangement in rat identifies the UT-A labeled structures as DTL of short-looped nephrons, whereas the aquaporin-1-labeled structures in the periphery are DTL of long-looped nephrons. Bar = 0.5 mm; inset bar = 50 μm.

In the upper portion of the inner stripe of the outer medulla, no structures express UT-A, as evidenced by the lack of green labeling (Fig. 5). The vascular bundles in this region (VBa) are surrounded by structures labeled with aquaporin-1 that are believed to include descending limbs of both short and long loops of Henle (16). Thus the descending limbs of short-looped nephrons are not homogeneous but vary with respect to UT-A and aquaporin-1 expression along their lengths.

Effect of DDAVP on abundance of 55-kDa isoform in rat. In preliminary experiments using homogenates of the inner medulla and from the inner stripe of outer medulla from rat prepared in the same manner as the mouse samples shown in Fig. 2 A, the COOH-terminal antibody L403 labeled a 55-kDa band similar to that seen in mouse (Fig.6). Note that, as in the mouse, there was a lower molecular mass band in the outer medulla of rat (approximately 46 kDa in rat). This band was nearly undetectable in the inner medulla. An additional band was seen at approximately 33 kDa, consistent with the apparent molecular mass reported for the nonglycosylated form of UT-A2 seen in in vitro translation experiments (10). Each of the three bands marked was ablated when the antibody is preincubated with the immunizing peptide (data not shown).

Fig. 6.

Immunoblot using polyclonal antibody L403 to probe whole homogenates of rat inner medulla (IM) and outer medulla (OM). Initial electrophoresis employed a 12% polyacrylamide gel, loading 22 μg of whole homogenate. After transfer to nitrocellulose, the top of the nitrocellulose membrane (above approximately 87 kDa) was cut off with a razor blade. This procedure allows long exposures of the blot to be made without interference from the strong UT-A1 band in inner medulla, which normally runs at around 97 kDa. See text for description.

Because UT-A2 mRNA abundance is markedly increased in the inner stripe of the outer medulla by DDAVP infusion in rats (21), we reasoned that if the 55-kDa band is the UT-A2 protein, then its density should be increased by DDAVP infusion. Results of an experiment to test this possibility are shown in Fig. 7 A. Homogenates were prepared from Brattleboro rats infused with DDAVP or vehicle by osmotic minipump for 7 days. Samples from six animals from each group were loaded for semiquantitative immunoblotting. Preliminary gels were identically loaded and stained with Coomassie blue dye to assure equal loading among lanes (not shown). In the inner stripe of the outer medulla (Fig. 7 A), the mean density of the 55-kDa band in DDAVP-infused Brattleboro rats was increased to 518 ± 57% of the mean value for vehicle-infused controls (P < 0.001). In contrast, the 46-kDa band was unchanged by hormone administration. The density of the 55-kDa band was also found to be increased in the inner medulla to 225 ± 35% of the value in vehicle-infused control rats (Fig. 7 B). The strong regulation of the abundance of the 55-kDa isoform, coupled with previous observation that the abundance of UT-A2 mRNA is markedly increased by DDAVP infusion (21), provides further support for the conclusion that the 55-kDa protein is UT-A2.

Fig. 7.

Semiquantitative immunoblots of renal medullary homogenates from 1-deamino-[8-d-arginine]vasopressin (DDAVP)-treated and vehicle-treated rats: each lane was loaded with a tissue homogenate sample from a different rat treated with either DDAVP (n = 6, left) or vehicle (n = 6, right) by osmotic minipump for 7 days. For the outer medulla (A), 26 μg of protein was loaded per lane; for inner medulla (B), 24 μg protein was loaded per lane. Preliminary polyacrylamide gels were loaded identically and stained with Coomassie blue dye to assure equality of loading among lanes. For inner medullary blot, the top of the nitrocellulose membrane (above approximately 87 kDa) was cut off with a razor blade, allowing a relatively long exposure of the blot to the film without interference from the strong 97-kDa UT-A1 band. Infusion of DDAVP for 7 days by osmotic minipump strongly increased the density of the 55-kDa band in the outer medulla (518 ± 57% increase,P < 0.001). Density of the band was also significantly increased in the inner medulla (225 ± 35% increase, P < 0.01).

Immunofluorescence localization in rat outer and inner medulla: Effect of DDAVP infusion. Figure 8shows representative immunofluorescence labeling of outer medulla from a DDAVP-infused Brattleboro rat (Fig. 8 A) and from a vehicle-treated control rat (Fig. 8 B) using the COOH-terminal antibody L403. There was consistently more intense labeling in DDAVP-infused animals (results typical of sections from 4 pairs of rats). In addition to a more intense labeling, the length of bundles labeled by the antibody was consistently greater in DDAVP-treated animals. This observation is in accord with the finding that the overall thickness of the inner stripe increases in response to DDAVP infusion in normal and Brattleboro rats (3, 5). In no case, however, did the thin limb labeling extend up to the end of the proximal straight tubule, as evidenced by comparison of fluorescence labeling and examination of the same sections by standard light microscopy (data not shown).

Fig. 8.

UT-A immunolabeling in the outer medulla of DDAVP-infused Brattleboro rat (A) and vehicle-infused rat (B) using polyclonal antibody L403: antibody concentrations, fixation conditions, labeling conditions, and microscope settings were the same for A and B. Labeling of descending limbs in DDAVP-treated animals (A) was greater than in vehicle-treated animals (B). Results are representative of those seen in multiple sections from four DDAVP-infused Brattleboro rats and four vehicle-infused control Brattleboro rats. Bar = 250 μm.

Figure 9 shows immunofluorescence labeling in cross sections from the base of the inner medulla in DDAVP-infused (Fig. 9, A and C) vs. vehicle-infused Brattleboro rats (Fig. 9, B and D). A comparison of A andB (single labeling with the L403 antibody) shows the striking increase in labeling of thin limbs in DDAVP-infused (Fig. 9 A) vs. vehicle-infused (Fig. 9 B) animals. (Note that antibody concentrations, fixation and labeling conditions, and confocal microscope settings were identical for A and B of Fig.9.) Figure 9, C and D, shows labeling in the same fields as Fig. 9, A and B, respectively, using an antibody to aquaporin-1. Aquaporin-1 is strongly expressed in thin descending limbs of long-looped nephrons in the inner medulla but not in thin ascending limbs (17). As can be seen, there is close overlap in UT-A and aquaporin-1 labeling after DDAVP infusion (Fig. 9 C), consistent with the localization of the induced UT-A protein in thin descending limbs of long-looped nephrons or vasa recta, the two sites of aquaporin-1 expression in the inner medulla (16, 17). Since vasa recta do not label with antibodies to UT-A (Fig. 5), we conclude that the UT-A-containing structures in inner medulla are thin descending limbs. Note that at this level of the inner medulla, the collecting ducts are not labeled with the UT-A antibody (Fig. 9, A andB), consistent with the low urea permeabilities measured in this portion of the inner medullary collecting ducts (22,23). Immunofluorescence labeling in deeper regions of the inner medulla did not reveal L403 labeling in thin limb segments either in control animals or in response to DDAVP (not shown).

Fig. 9.

UT-A immunolocalization in base of the inner medulla: effect of chronic DDAVP treatment. A and B: localization using the L403 antibody in DDAVP- and vehicle-infused Brattleboro rats, respectively.C and D: localization in the same section of LC18, an antipeptide antibody raised in chicken to AQP1, a marker for the thin descending limb. Labeling by L403 is nearly impossible to detect in the base of the inner medulla of vehicle-treated animals (B), although thin descending limbs of long-looped nephrons are present as shown by the AQP1 labeling in the same section (D). In DDAVP-treated animals, labeling by L403 is very strong (A) in thin descending limbs as identified by AQP1 colabeling (C). Note that at this level of the inner medulla, the collecting ducts are not labeled with the UT-A antibody, consistent with the low urea permeabilities measured in this segment (22, 23). Bar = 50 μm.

DISCUSSION

Based on the observations made in this study, we conclude: 1) that the UT-A2 urea transporter isoform is detectable by immunoblotting in mouse and rat outer medulla as a protein with an apparent molecular mass of 55 kDa; 2) that, in addition to its presence in thin descending limbs of short-loop nephrons, this protein is also present in the base of the inner medulla in the thin descending limbs of long-loop nephrons; and 3) that the abundance of this protein is markedly increased in response to infusion of the V2 receptor-selective vasopressin analog DDAVP in both the inner stripe of the outer medulla and in the base of the inner medulla. In the remainder of this discussion, we address the basis of these conclusions and the physiological significance of the findings.

Presence of 55-kDa UT-A isoform in mouse and rat outer medulla.In a previous study using an antibody to the COOH-terminal tail of UT-A1 (18), we could not detect the UT-A2 isoform in outer medulla by immunoblotting, despite the fact that the antibody strongly labeled the thin descending limbs of short-looped nephrons surrounding the vascular bundles of the outer medulla in tissue sections. A number of technical changes that we have introduced since that time have allowed greater sensitivity in immunoblotting, permitting us to identify a 55-kDa protein in the outer medullas of both mice and rats that has characteristics consistent with its identification as UT-A2 (see below). These technical changes include the introduction of a new COOH-terminal antibody L403, the use of whole homogenate samples rather than membrane fractions, and improvement of immunoblotting equipment and reagents available from commercial sources. The improvement afforded by the use of whole homogenate samples could stem from two factors: 1) The small type I thin descending limb cells that make up the epithelium of the descending limbs of short-loop nephrons may be more difficult to homogenize than are other epithelial cells of the outer medulla (e.g., thick ascending limb cells and collecting duct cells), which are much larger. 2) Preparation of membrane fractions in low ionic strength isolation solutions may encourage formation of insoluble complexes. We have not carried out an evaluation of the relative role of each of these factors in allowing us to now detect UT-A2 in blots.

Identification of the 55-kDa UT-A protein as UT-A2. The identification of the 55-kDa protein as UT-A2 is based on several criteria. 1) The size, 55 kDa, is consistent with the expected size of the 397-amino acid UT-A2 protein, if glycosylation is taken into account. 2) The 55-kDa protein was found in membrane fractions as would be expected for UT-A2, an integral membrane protein.3) Antibody L403 recognized it, but antibody L446 did not, consistent with the view that the 55-kDa protein has sequence common to the COOH-terminal (but not amino-terminal) portion of UT-A1, as would be expected for UT-A2. The L446 antibody is targeted to a portion of UT-A1 that is not contained in the UT-A2 polypeptide (Fig. 1).4) DDAVP administration strongly increased the abundance of the 55-kDa outer medullary protein in Brattleboro rats (Fig. 7). Previous studies have clearly demonstrated that DDAVP and arginine vasopressin markedly increase UT-A2 expression in the rat outer medulla at the mRNA level (21). The corresponding increase in the 55-kDa UT-A protein parallels the previously demonstrated increase in UT-A2 mRNA expression, indicating a likely relationship between the two responses. Although this evidence strongly points to identification of the 55-kDa protein as UT-A2, we cannot rule out that it corresponds to an as yet unidentified UT-A isoform.

If the 55-kDa protein is UT-A2, then what is the 48-kDa UT-A protein? Recently, two new UT-A cDNAs have been cloned, corresponding to additional splice variants derived from the UT-A gene (10). UT-A3 is identical to the first 460 amino acids of UT-A2, except for the substitution of D for G at position 460. UT-A4 comprises the first quarter and fourth quarter of UT-A1, spliced together, and is predicted to be 466 amino acids in length. The UT-A4 protein is predicted to be recognized by both antibodies L446 and L403 (Fig. 1) and to have a nonglycosylated molecular mass of 51 kDa. Hence, based on the evidence provided in the current study, we hypothesize that the 48-kDa protein is UT-A4, although more evidence would be required to draw this conclusion with certainty. Because the L446 antibody proved unsuitable for immunocytochemical localization studies, we have not yet been able to assess the cellular localization of the 48-kDa UT-A protein.

When the cloned cDNA for UT-A2 was translated in vitro, it produced a product with an apparent molecular mass of 33–35 kDa (10), which is smaller than 55 kDa seen in the present study, and indeed even smaller than the predicted size of the protein based on the open reading frame of the cDNA clone, viz. 43 kDa. This result suggests that the UT-A2 protein expressed in the rat and mouse outer medulla is modified in some way by posttranslational modification, most likely glycosylation. It is interesting that our outer medullary immunoblots with the COOH-terminal antibody did reveal a 33-kDa protein (Figs.2 B and 6) consistent with the size of the translation product, suggesting that this 33-kDa band may be due to the presence of a nonglycosylated form of the UT-A2 protein.

Heterogeneity of thin descending limbs in the inner medulla. Colabeling studies of UT-A with aquaporin-1 (Fig. 5) have demonstrated a separation of sites of water channel and urea transporter (UT-A) expression in thin descending limbs of the inner stripe of the outer medulla, consistent with previous studies of aquaporin-1 (16, 17) and UT-A (18) localization. In the lower portion of the inner stripe, the thin descending limbs of short-loop nephrons express UT-A but not aquaporin-1, whereas the thin descending limbs of the long-loop nephrons express aquaporin-1 and not UT-A. However, only the most distal portion of the descending limbs of short loops of Henle express UT-A, as previously demonstrated at the mRNA level by RT-PCR (25). Previously, it had been demonstrated that the upper part but not the lower part of the short-loop descending limb expresses aquaporin-1 (16). In this study, we have found that there is no detectable overlap between the aquaporin-1-expressing portion and the UT-A-expressing portion of the short-loop descending limb (Fig. 5).

Physiological significance of UT-A localization in thin descending limbs of Henle. Immunofluorescence localization of UT-A2 with the carboxy-terminal antibody L403 confirmed the previously observed strong expression in a population of structures adjacent to the vasa recta bundles (18). These structures were identified as thin descending limbs of short-loop nephrons based on their continuity with thick ascending limbs in the inner stripe of the outer medulla (18). These thin descending limbs are lined with so-called type I thin descending limb epithelium, described previously by Kriz and colleagues (1, 9, 12). Urea transporter expression at this site would be expected to facilitate the transfer of urea from the ascending vasa recta to the short loops of Henle in the outer medulla (Fig.10). The existence of this pathway is believed to enhance urinary concentrating ability by recycling urea originating from the inner medullary interstitium back to the nephron (2, 15, 33). Recent immunolocalization studies using antibodies to the UT-B gene product indicate that UT-B protein is expressed in the endothelial cells of the vasa recta at this level (34). Thus the combination of urea transporters in adjacent structures in the vasa recta bundles of the outer medulla may be important to the concentrating mechanism by assuring efficient recycling of urea and making urea available to distal sites of short-looped nephrons. The first evidence for this recycling pathway was obtained from the micropuncture studies of Lassiter et al. (14), which showed a large amount of net secretion of urea into the short loops of Henle, resulting in delivery of amounts of urea to the superficial distal tubule that can exceed the filtered load.

Fig. 10.

Schematic illustration of the distribution of UT-A2 and its possible physiological role in urea recycling. Structures are not drawn to scale. Shading indicates sites of UT-A expression. C, cortex; OS, outer stripe of outer medulla; IS, inner stripe of outer medulla; IM, inner medulla; DVR, descending vasa recta; and AVR, ascending vasa recta. Urea is reabsorbed from the inner medullary collecting duct, where it is trapped by countercurrent exchange processes (not depicted). Urea leaving the inner medulla via the AVR can exit via the fenestrated endothelium of the AVR. The propinquity of the AVR and short-loop descending limbs in the vascular bundles of the inner stripe permits transfer of the urea from the AVR into the descending limbs via UT-A2. UT-A2 in the inner medullary portion of the long-loop descending limb may recycle urea absorbed from the ascending thin limb of Henle (which is highly permeable to urea despite the absence of any known urea transporter) and from the AVR. See text for further explanation.

In addition to the localization in the short-loop descending limbs, the current studies demonstrate that a UT-A protein (presumably UT-A2) is expressed in thin descending limbs in the inner medulla, consistent with the results of RT-PCR studies in rats showing UT-A2 mRNA in inner medullary thin descending limbs (25) and in situ hybridization studies in rat showing UT-A2 mRNA in the base of the inner medulla (32). The labeled structures represent the inner medullary portion of thin descending limbs of long-looped nephrons but may not correspond strictly to the localization of the type III thin limb epithelium in rat. The type III epithelium begins a variable distance from the inner-outer medullary junction and continues deep into the inner medulla (12). In contrast, UT-A labeling in inner medulla generally began very close to the inner-outer junction and did not continue into the deep inner medulla. Localization at an electron microscopic level will be needed to clarify this issue. Isolated perfused tubule studies comparing urea permeability in descending limbs from the outer medullary portion of the long loops of Henle vs. descending limbs from the inner medullary portion of the long loops of Henle have demonstrated a substantially higher urea permeability in the inner medullary segments (6, 8). Thus, the measured urea permeabilities correlate with the relative expression levels of UT-A protein, supporting the view that the UT-A2 or a closely related UT-A isoform such as UT-A4 may function as a urea transporter in inner medullary descending limbs. A urea transporter at this site might function to promote recycling of urea reabsorbed from the thin ascending limb into the descending limb. In addition, urea that might otherwise exit the inner medulla via ascending vasa recta could recycle into thin descending limbs via this urea transporter (see Fig. 10). Entry of urea into the long-loop descending limbs at the base of the inner medulla provides a likely explanation for the extensive urea secretion deduced from micropuncture studies, which demonstrated that the rate of urea delivery in tubule fluid at bends of long loops of Henle is far in excess of the filtered load of urea (19, 20). Enhanced countercurrent exchange between ascending and descending structures would help maintain a high urea concentration in the inner medullary interstitium (4).

Regulation by vasopressin. Our immunoblotting showed that the abundance of the 55-kDa UT-A isoform, which we identify as UT-A2, is strongly induced by DDAVP infusion both in the inner stripe of the outer medulla as well as in the base of the inner medulla (Figs.7-9). Because this response can be elicited by DDAVP infusion, it seems likely that it is the result of V2 receptor activation. Although it is well established that the V2 vasopressin receptor is expressed in both thin ascending limbs and thick ascending limbs of Henle's loops, this receptor has not been reported to be expressed in the thin descending limb (7, 29), although these studies did not specifically examine descending thin limb from short-looped vs. long looped nephrons. Thus, a DDAVP-elicited increase in UT-A2 abundance in descending limbs from short-looped nephrons may possibly be due to an indirect response to DDAVP that is as a result of V2 receptor occupation at other renal tubule sites (21, 32). For example, the vasopressin-stimulated increase in UT-A2 expression could result from changes in medullary osmolality or release of diffusible autacoids from nearby cell types possessing V2 receptors, e.g., thick ascending limbs or collecting ducts.

DDAVP infusion not only increased the total abundance of UT-A2 in the outer medulla, but immunofluorescence labeling also demonstrated an upward spread of UT-A expression in the short-looped descending limbs in response to DDAVP infusion (Fig. 8). This upward spread was predictable from the observation by Promeneur et al. (21) that UT-A2 mRNA is upregulated in the upper half of the inner stripe of the outer medulla in response to DDAVP infusion.

Acknowledgments

We thank Donald Wade for preparation of Fig. 10 and Dr. Lise Bankir (Paris, France) for helpful discussions.

Footnotes

  • Address for reprint requests and other correspondence: J. B. Wade, Dept. of Physiology, 655 W. Baltimore St., Univ. of Maryland, Baltimore, MD 21201.

  • The work reported herein was supported by the intramural budget of the National, Heart, Lung, and Blood Institute (to M. A. Knepper, Project Z01-HL-01282-KE) and by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-32839 (to J. B. Wade). The Confocal Microscope Facility used for the immunolocalizations was funded by National Science Foundation Grant BIR9318061.

  • 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. §1734 solely to indicate this fact.

  • 1 In addition to UT-A1 and UT-A2, the UT-A gene appears to produce at least two other transcripts, “UT-A3” and “UT-A4,” whose cDNAs have recently been cloned (10). The putative structure of the coded proteins are summarized in the caption to Fig.1. The renal tubular localization and functional roles of these putative proteins remain undefined.

REFERENCES

View Abstract