Vasopressin-regulated urea transport in the renal inner medullary collecting duct (IMCD) is mediated by two urea channel proteins, UT-A1 and UT-A3, derived from the same gene (Slc14a2) by alternative splicing. The NH2-terminal 459 amino acids are the same in both proteins. To study UT-A1/3 phosphorylation, we made phospho-specific antibodies to UT-A sequences targeting phospho-serines at positions 84 and 486, sites identified previously by protein mass spectrometry. Both antibodies proved specific, recognizing only the phosphorylated forms of UT-A1 and -A3. Immunoblotting of rat IMCD suspensions or whole inner medullas showed that the V2R-selective vasopressin analog 1-deamino-8-d-arginine vasopressin (dDAVP) increases phosphorylation at Ser84 (in UT-A1 and UT-A3) and Ser486 (in UT-A1) by about eightfold. Time course studies in rat IMCD suspensions showed maximum phosphorylation within 1 min of dDAVP exposure, consistent with the time course of vasopressin-stimulated phosphorylation of the vasopressin-sensitive water channel aquaporin-2 at Ser256. Confocal immunofluorescence in Brattleboro rat medullary tissue showed labeling limited to the IMCD, which increased markedly in response to dDAVP. Immuno-electron microscopy studies showed that both phosphorylated forms were present mainly in intracellular compartments in the presence of vasopressin. These studies demonstrate regulated phosphorylation of both UT-A1 and UT-A3 in response to vasopressin in a manner consistent with coordinate regulation of UT-A and aquaporin-2 in the renal IMCD. The findings add to prior evidence for vasopressin-induced phosphorylation of UT-A1, providing evidence that UT-A3 may be regulated by phosphorylation as well.

  • phospho-specific antibody
  • confocal microscopy
  • immunofluorescence immunocytochemistry
  • immuno-electron microscopy

the renal inner medullary collecting duct (IMCD) is the sole site of expression of two urea transport proteins, UT-A1 and UT-A3 (Fig. 1), which are derived from a single gene (Slc14a2) by alternative splicing. Based on recent X-ray crystallography studies, these proteins probably function as urea channels (20), rather than carriers as previously believed. These urea channels mediate transepithelial urea transport, which allows rapid equilibration of urea between the lumen of the IMCD and the inner medullary interstitium. Thus, urea, which is generated as the chief nitrogen waste product from protein metabolism, can be excreted at high rates without impairing urine-concentrating ability (4, 12). Vasopressin strongly regulates the urea permeability of the IMCD (34). However, the mechanisms involved in this regulation remain unclear. [32P]-labeling experiments revealed that vasopressin signaling increases phosphorylation of UT-A1 (5, 44). Candidate phosphorylation sites (Ser486 and Ser499) in UT-A1 have been mutated and shown to be involved in urea transport regulation, although sites in UT-A3 have not been investigated in this manner (6). More recently, mass spectrometry-based phosphoproteomic studies identified several sites in addition to Ser486 and Ser499 (2, 16), namely Ser10, Ser62, Ser63, and Ser84 (2). Among these sites, three have been demonstrated by mass spectrometry to be regulated by vasopressin: Ser84, Ser486, and Ser499 (2, 6). Of these, Ser84 is present in both UT-A1 and UT-A3, whereas the other two sites are only present in UT-A1 (Fig. 1).

To investigate regulated phosphorylation at Ser84, we produced a phospho-specific antibody to this site and used it in immunoblotting and immunohistochemical experiments in rats. The results show that vasopressin stimulates phosphorylation of both UT-A1 and UT-A3 at Ser84 with a time course similar to that of vasopressin-induced phosphorylation of aquaporin-2 at Ser256. In addition, to confirm the role of the Ser486 site in UT-A1 and investigate its regulation by vasopressin, we produced a phospho-specific antibody to this site and confirmed that phosphorylation of this site is also increased by vasopressin.


Experimental animals.

Pathogen-free male Sprague-Dawley rats (Taconic Farms, Germantown, NY) or homozygous Brattleboro rats (Harland Sprague-Dawley, Indianapolis, IN) weighing 150–200 g were used in these studies. All experiments were conducted in accord with animal protocol H-0110R1 approved by the Animal Care and Use Committee of the National Heart, Lung, and Blood Institute or the boards of the Institute of Anatomy and Institute of Clinical Medicine, Aarhus University, according to the licenses for use of experimental animals issued by the Danish Ministry of Justice.

Phospho-specific UT-A antibodies.

Rabbit polyclonal phospho-specific UT-A1/3 antibodies were generated against 12-amino acid synthetic phosphopeptides surrounding rat Ser84 or Ser486 (see NP_062220.2 for sequence). For each antibody, two rabbits were immunized and affinity-purified (PhosphoSolutions, Aurora, CA), P7281 and P7282 for pSer84-UT-A1/3; P7283 and P7284 for pSer486-UT-A1. The specificities of these antibodies were tested by dot blotting. We also used a previously characterized rabbit antibody L446 targeted for the NH2-terminal tail of both UT-A1 and UT-A3 (7). IgG concentration of each antibody was quantified by a NanoDrop spectrophotometer (model ND-1000, NanoDrop Technologies, Wilmington, DE; P7281, 50 μg/ml; P7282, 60 μg/ml; P7283, 90 μg/ml; P7284, 70 μg/ml; L446, 140 μg/ml). Other antibodies used in this paper were rabbit anti-aquaporin-2 (AQP2; K5007) (14), rabbit anti-phospho-Ser256-AQP2 (14), and chicken anti-AQP2 (LL265) (3). All of these antibodies have been previously characterized.

IMCD suspensions.

IMCD suspensions were prepared as previously described with slight modification (9). After rats were euthanized, kidney inner medullas were dissected, minced, and digested into suspensions by incubation at 37°C for 70–90 min in digestion solution (250 mM sucrose, 10 mM triethanolamine, pH 7.6) containing collagenase B (3 mg/ml; Roche, Indianapolis, IN) and hyaluronidase (1,400 U/ml; Worthington, Lakewood, NJ). The resulting inner medullary suspension (whole IM) was subjected to three low-speed centrifugations (at 70 g, 30 s) to separate the IMCD-enriched fraction in the pellet from the non-IMCD fraction in the supernatant. When all three fractions were used, they were harvested by centrifugation at 1,000 g for 5 min and washed once. The pellets were resuspended in 1× Laemmli buffer (1.5% SDS, 10 mM Tris, pH 6.8) containing protease and phosphatase inhibitors and passed through a DNA shredder (Qiagen, Valencia, CA). After protein concentration determination by the BCA method (Thermo Scientific, Rockford, IL), samples were stored in the presence of 6% glycerol and 40 mM DTT after heating at 60°C for 15 min. When IMCD suspensions were used in physiological experiments, the IMCD-enriched pellet was gently washed and resuspended in bicarbonate-buffered solution containing (in mM) 118 NaCl, 4 Na2HPO4, 5 KCl, 25 NaHCO3, 2 CaCl2, 1.2 MgSO4, and 5.5 glucose (290 mosmol/kgH2O). IMCD suspensions were allowed to equilibrate with gentle stirring with an aid of a micro-stirring bar at 37°C with 95% air-5% CO2 supply for 10 min before use. Time course experiments were carried out by exposing IMCD suspensions to 1-deamino-8-d-arginine vasopressin (dDAVP) or vehicle for various lengths of time (0, 1, 5, 15, 30 min). Hormone incubation was terminated by spinning the suspensions at 14,000 rpm for 1 min to harvest the pellet containing IMCD segments. Samples were resuspended in 1× Laemmli buffer and treated as described above.

Semiquantitative immunoblotting.

Proteins were resolved by SDS-PAGE on polyacrylamide gels (Criterion, Bio-Rad, Hercules, CA) and transferred electrophoretically onto nitrocellulose membranes. Membranes were blocked for 30 min with Odyssey blocking buffer (Li-Cor, Lincoln, NE), rinsed, and probed with the respective affinity-purified antibodies at proper dilution (in Odyssey blocking buffer containing 0.1% Tween 20) overnight at 4°C. After 1-h incubation with secondary antibody (Alexa Fluor 680 goat anti-rabbit immunoglobulin G; Invitrogen, Carlsbad, CA) at 1:5,000 dilution, sites of antibody-antigen reaction were detected using an Odyssey infrared imager (Li-Cor).

Perfusion fixation of rat kidneys.

Rats under anesthesia were surgically prepared for retrograde perfusion of the kidneys via the abdominal aorta. The kidneys were first perfused with PBS for 10 s to wash out the blood, followed by ice-cold 4% paraformaldehyde for 5 min. The fixed kidneys were trimmed to expose all three major regions (cortex, outer medulla, and inner medulla), embedded in paraffin, and sectioned (4 μm) for immunofluorescence studies.

Immunofluorescence confocal microscopy.

Immunostaining was performed as previously described (43). In brief, paraffin-embedded whole kidney sections were dewaxed using xylene and rehydrated sequentially in 100, 95, 90, and 70% ethanol. Antigen retrieval was performed with microwave treatment for 15 min in TEG buffer (10 mM Tris and 0.5 mM EGTA, pH 9.0) followed by neutralization in 50 mM NH4Cl (in PBS). Blocking was performed using 1% BSA, 0.2% gelatin, and 0.05% saponin in PBS. Incubation with the primary antibody (diluted in 0.1% BSA and 0.3% Triton X-100 in PBS) was performed overnight (4°C). After being washed with 0.1% BSA, 0.2% gelatin, and 0.05% saponin in PBS, tissue sections were incubated for 1 h with secondary antibody (conjugated with either Alexa 488 or Alexa 568; Invitrogen) diluted in 0.1% BSA and 0.3% Triton X-100 in PBS. After subsequent washes with PBS, nuclei were counterstained with DAPI (4 μl of 0.2 mg/ml DAPI stock solution diluted in 10 ml PBS). The sections were then preserved in fluorescence mounting medium (S3023, Dako North America). For peptide blocking controls, antibody was preincubated with appropriate peptides at a 1:25 molar ratio for 2 h at 4°C before use. Sections were also incubated without primary antibody as a negative control. Confocal fluorescence images were acquired using a Zeiss LSM 510 META microscope and software (Carl Zeiss MicroImaging, Thornwood, NY).

Immunogold-electron microscopy.

We carried out immunogold labeling of rat renal inner medulla tissues following the procedure described by Moeller et al. (27), with rats treated with dDAVP for 60 min and control rats. Anti-pS84 (no. 7281, dilution 1:50), anti-pS486 (no. 7284, dilution 1:50), and anti-UT-A1/3 (L446, dilution, 1:300) were used.

Statistical analysis.

Data are presented as means ± SE. All statistical comparisons were made by t-tests. P < 0.05 was considered significant.


Specificities of phospho-specific UT-A1/3 antibodies.

Figure 1 shows the locations of phosphoserines targeted by the phospho-specific antibodies. Figure 2 shows the results of dot blotting testing the specificities of the antibodies. As shown, phospho-Ser84 UT-A1/3 antibody P7282 recognized only the phosphopeptide targeted to Ser84 of UT-A1/3 but not the corresponding nonphosphopeptide. Antibody P7281 was similar in specificity and produced only a very faint nonphosphopeptide signal that is barely visible in Fig. 2. Antibody P7282 was consequently used in immunoblotting. Antibody P7281, however, yielded better immunostaining in tissue sections than P7282, so it was used in immunofluorescence studies only with appropriate labeling controls. Similarly, one phospho-Ser486-UT-A1 antibody (P7284) showed specificity for its phosphopeptide. Antibody P7284 was used in both immunoblotting and immunofluorescence studies.

Fig. 1.

Phospho-Ser84 UT-A1/3 and phospho-Ser486 UT-A1 antibody recognition sites. The schematic drawing of UT-A1 and UT-A3 shows that the NH2-terminal 1–459 amino acids of UT-A1 are identical to UT-A3. The filled boxes represent hydrophobic regions that contain transmembrane domains of the protein. The residue numbers indicated here and elsewhere in this paper correspond to the rat sequence (RefSeq no. NP_062220.2). The antibodies recognizing phosphorylated Ser84 of both UT-A1 and UT-A3 (P7181 and P7282) were raised against a synthetic phosphopeptide targeting to Ser84. The antibody recognizing phosphorylated Ser486 of UT-A1 (P7284) was raised against a synthetic phosphopeptide targeting to Ser486. A previously characterized antibody recognizing total UT-A1 and UT-A3 (L446) was raised against amino acids 56–78 (7, 38).

Fig. 2.

Specificities of phospho-Ser84 UT-A1/3 antibody (P7281 and P7282) and phospho-Ser486 UT-A1 antibody (P7283 and P7284). In each dot blot, 0.01 μg of phosphopeptides was loaded to the upper end of the nitrocellulose membrane and corresponding nonphosphopeptides were loaded to the lower end. All 4 antibodies were used at a 1:1,000 dilution (see IgG concentrations in methods). Antibody P7282 and P7284 recognized phosphopeptides but not nonphosphopeptides.

Phospho-UT-A1 and UT-A3 localization in rat kidney.

Figure 3A shows immunoblots testing for the presence of phosphorylated UT-A1 and UT-A3 proteins in a whole inner medullary suspension and two subfractions, non-IMCD and IMCD-enriched (first 3 lanes of each blot). The results show that all three phosphorylation sites, pSer84-UT-A1, pSer84-UT-A3, and pSer486-UT-A1, are enriched in the IMCD fraction, consistent with the known exclusive distribution of UT-A1 and UT-A3 in IMCD (30, 38). Previous studies of UT-A1 demonstrated that it exists in the IMCD chiefly at two distinct molecular weights, 97 and 117 kDa (on 10% polyacrylamide gel), that represent two distinct levels of glycosylation (deglycosylated UT-A1 ran at 88 kDa.) (7). In Fig. 3A, these two previously identified bands are indicated by arrows labeled “UT-A1.” Also, previous studies of UT-A3 demonstrated that it is present chiefly at two distinct molecular weights, 67 and 44 kDa, that represent two distinct levels of glycosylation (deglycosylated UT-A3 ran at 40 kDa) (38). In Fig. 3A, these previously identified bands are indicated by arrows labeled “UT-A3.” Both phospho-specific antibodies P7282 and P7284 recognized pSer84-UT-A1 or pSer486-UT-A1 as a higher band at 117 kDa and a lower band at 97 kDa, consistent with previous observations showing that these bands represent two different glycosylation states (7). The phospho-specific antibody P7282 also recognized pSer84-UT-A3 at two molecular weights, ∼67 and ∼44 kDa, consistent with what was reported previously for UT-A3 (38) and confirmed here in an immunoblot using an antibody to total UT-A1/3 (antibody LL446; Fig. 3A, right). Furthermore, as shown in the last two lanes of each blot in Fig. 3A, V2 receptor-selective vasopressin analog dDAVP (1 nM, 30 min) strongly increased the abundance of pSer84-UT-A1, pSer84-UT-A3, and pSer486-UT-A1, but not total UT-A1/3, in IMCD, confirming that phosphorylation of Ser84 of UT-A1 and UT-A3 and Ser486 of UT-A1 is regulated by vasopressin. Figure 3B confirms that the tubule fractionation procedure successfully enriched the collecting duct-specific marker AQP2 and deenriched the loop of Henle marker AQP1 in the IMCD fractions. Preadsorption controls confirm the specificities of the phospho-specific antibodies (Fig. 3C), showing that the same bands described above were ablated by preincubation with the specific phospho-peptide. Note that in the pS486 blot, there is a pair of bands at ∼75 and ∼55 kDa (labeled with asterisks) that appear to be compatible with similar bands seen previously with the same antibody after limited chymotrypsin cleavage of UT-A1 (7). Thus, it appears possible that a small fraction of total UT-A1 may be subject to physiological proteolytic cleavage of the middle loop downstream from Ser486.

Fig. 3.

Immunoblotting using anti-phospho-Ser84 UT-A1 and UT-A3 and anti-phospho-Ser486 UT-A1 antibodies in rat inner medullary samples. A: in each blot, whole inner medullary suspension (whole IM) and 2 subfractions, non-inner medullary collecting duct (non-IMCD) and IMCD, were loaded (first 3 lanes) to test the presence of phosphorylated or nonphosphorylated UT-A1 and UT-A3 in IMCD. In addition, IMCD samples with and without treatment of 1-deamino-8-d-arginine vasopressin (dDAVP; 1 nM, 30 min) were loaded (last 2 lanes) to test the effect of vasopressin on phosphorylation of UT-A1 and UT-A3 on targeted sites. Samples were loaded on a 4–15% polyacrylamide gel. Loading was 30 μg per lane except for dDAVP-treated IMCD which was 20 μg. Antibodies P7282, P7284, and L446 labeled UT-A1 as 117- and 97-kDa bands (arrows) as previously found (7). Antibodies P7282 and L446 labeled UT-A3 as broad ladder-like 67- and 44-kDa bands (arrows), as prevously found (38). All antibodies, P7282, P7284, and L446, were used at a 1:1,000 dilution. B: an enrichment test for IMCD suspension preparation using collecting duct aquaporin (AQP2) as a marker for IMCD and thin descending limb aquaporin (AQP1) as a marker for non-IMCD. C: preadsorption control. Twenty micrograms IMCD samples treated with dDAVP were loaded per lane on a 12.5% polyacrylamide gel. Membranes were incubated with antibody with or without preadsorption with appropriate peptides for 4 h at room temperature before use. Peptide concentrations: 0.1 μg/ml. *Bands compatible with physiological proteolytic cleavage of UT-A1 (7).

Immunofluorescence labeling showed that both total and phosphorylated UT-A1/3 (both sites) are specifically localized to IMCD cells, with no labeling of other cell types (Fig. 4). Within the IMCD cell, phospho-UT-A1/3 were found in both the cytoplasm and apical region of the cell, a result consistent with previous immunolocalization studies of UT-A1 (30) or UT-A3 (38).

Fig. 4.

Immunofluorescence localization of phosphorylated UT-A1/3 in Sprague-Dawley rat kidney inner medulla. Phospho-Ser84 UT-A1/3 (left) and phospho-Ser486 UT-A1 proteins (middle), as well as total UT-A1/3 (right), are present only in collecting duct cells. P7281 and P7284 antibodies were used at a 1:200 dilution; L446 was used at a 1:100 dilution. (Preadsorption controls are shown below in Figs. 7 and 8.)

Time course of UT-A1 and UT-A3 phosphorylation.

Previously, using [32P] radioisotope labeling in IMCD suspensions, Zhang et al. (44) demonstrated that vasopressin could rapidly (within 2–5 min) increase phosphorylation of UT-A1. To test whether vasopressin stimulates phosphorylation of Ser84 or Ser486 of UT-A1 or UT-A3 in a similar fashion, IMCD suspensions were incubated with dDAVP for various lengths of time. As shown in Fig. 5A, there was a substantial increase in phosphorylation of UT-A1 (117- and 97-kDa bands) at Ser84 and Ser486 in response to dDAVP, beginning at the shortest incubation time (1 min) and persisting throughout the entire experimental period (30 min). (Note that the ratio of the upper UT-A1 band density to the lower UT-A1 band density appears to be greater for Ser486-phosphorylated UT-A1 than for Ser84-phosphorylated UT-A1.) Similarly, a rapid increase in phosphorylation can be seen at Ser84 in UT-A3. In contrast, there was no change in the abundance of total UT-A1 or UT-A3 in response to dDAVP. Phosphorylation of AQP2 at Ser256 was also increased within 1 min of dDAVP addition as previously demonstrated (8, 15). Figure 5B summarizes the time course data graphically, showing the magnitude of UT-A1 and UT-A3 phosphorylation induced by dDAVP as log2(dDAVP:control) as a function of time (n = 3).

Fig. 5.

Time course of UT-A1 and UT-A3 phosphorylation in response to dDAVP in rat IMCD suspensions. A: representative immunoblots for pSer84-UT-A1, pSer84-UT-A3, pSer486-UT-A1, and total UT-A1/3 in control and dDAVP-treated IMCD suspensions prepared from Sprague-Dawley rats. At each time point, IMCD treated with 1 nM dDAVP had a substantially higher UT-A1 and UT-A3 phosphorylation level compared with its control, whereas the total UT-A1 or UT-A3 did not significantly change. Phosphorylation of AQP2 at Ser256 was also measured as a positive control to demonstrate the efficacy of the dDAVP. Thirty micrograms of IMCD protein were loaded per lane on a 12.5% polyacrylamide gel. B: summary line graph depicts the changes in protein abundance of each protein, expressed as log2(dDAVP:control; n = 3). Band densities were quantified by drawing a single integration box around both bands for UT-A1 and both bands for UT-A3.

In vivo UT-A1 and UT-A3 phosphorylation in response to vasopressin.

We next tested whether these phosphorylation events can occur in the animal. The experiment was conducted on six Sprague-Dawley rats given 200 mM sucrose for drinking for 16 h before the experiment to lower the baseline level of phospho-UT-A. Thereafter, three rats received an intramuscular injection of 2 nmol dDAVP 1 h before euthanasia and the other three rats received saline as control. Rats receiving dDAVP treatment had a large increase in urine osmolality (before: 184 ± 13, after: 1,389 ± 65 mosmol/kgH2O, n = 3, P < 0.0001), whereas the control rats had only a small increase in urine osmolality over the 1-h waiting period (before: 155 ± 3, after: 297 ± 3 mosmol/kgH2O, n = 3, P < 0.05). Immunoblotting was performed on whole homogenates of the terminal 80% of kidney inner medulla (the papilla). The results show that inner medullas of dDAVP-treated rats have significantly higher abundances of all three phospho-UT-A proteins, pSer84-UT-A1, pSer84-UT-A3, and pSer486-UT-A1, than inner medullas from control rats (Fig. 6A). Figure 6B graphically depicts the magnitude of UT-A phosphorylation induced by dDAVP expressed as log2(dDAVP:control).

Fig. 6.

In vivo phosphorylation of UT-A1 and UT-A3 in response to dDAVP. Experiment was done on 6 water-loaded Sprague-Dawley rats. Three rats received intramuscular injection of 2 nmol dDAVP 1 h before euthanasia and 3 rats received saline only as control. Terminal inner medullas were dissected and homogenized for immunoblotting. A: immunoblots displaying phosphorylation changes in response to dDAVP in whole inner medulla homogenates. Thirty micrograms of sample protein were loaded per lane on a 12.5% polyacrylamide gel. Inner medullary whole homegenates from dDAVP-treated rats had significantly higher levels of pSer84-UT-A1, pSer84-UT-A3, and pSer486-UT-A1 than control rats. B: summary of quantification of immunoblots for phospho-UT-A1 and UT-A3 presented as log2(dDAVP:control). Band densities were quantified by drawing a single integration box around both bands for UT-A1 and both bands for UT-A3.

An in vivo dDAVP response experiment was also conducted in Brattleboro rats, a natural strain of rat that lacks endogenously circulating vasopressin, to examine the immunolocalization of phosphorylated UT-A1/3 in kidney inner medulla before and after vasopressin stimulation. Figure 7 shows a seemingly “all or none” pattern of pSer84-UT-A1/3 in cytoplasm and the apical region of Brattleboro rat IMCD after vasopressin stimulation (Fig. 7, A and B). Vasopressin-induced redistribution of AQP2 from intracellular domain to apical plasma membrane was also shown as a positive control (Fig. 7C). Preabsorption with the Ser84 phosphopeptide fully ablated the signal (Fig. 7A). Similar results were obtained for pSer486-UT-A1 as shown in Fig. 7, D–F. Again, preabsorption with the Ser486 phosphopeptide fully ablated the signal (Fig. 7D).

Fig. 7.

Immunofluorescence localization of pSer84-UT-A1/3 (A–C) and pS486-UT-A1 (D–F) in Brattleboro rat kidney inner medulla–effect of dDAVP. A: test of antibody (P7281) specificity using a peptide blocking assay (blue: nuclei, green: pSer84-UT-A1/3). Tissue sections were preincubated with antibody with or without preadsorption with Ser84 phosphopeptide for 2 h at 4°C before use. P7281 antibody was used at a 1:50 dilution. Peptide concentration: 0.5 μg/ml. Phosphorylation at Ser84 was detected in kidney sections obtained from dDAVP-treated rats but not vehicle-treated rats. Preadsorption with immunizing phosphopeptide ablated signal. B: immunolocalization of pSer84-UT-A1/3 and total UT-A1/3 in consecutive kidney sections. pSer84-UT-A1/3 and UT-A1/3 were detected in collecting ducts only. Both antibodies were used at a 1:50 dilution. C: double labeling for pSer84-UT-A1/3 (green) and AQP2 (red). After dDAVP treatment, AQP2 is mainly apical, whereas pSer84-UT-A1/3 is localized both apically and intracellularly. AQP2 antibody was used at a 1:100 dilution. A–C: identical confocal microscope settings were used to visualize vehicle and dDAVP-treated samples. Immunofluorescence using pS486-UT-A1 antibody was done in a manner similar to pS84-UT-A1/3 antibody as described above. D: test of antibody (P7284) specificity using a peptide-blocking assay (blue: nuclei, green: pSer486-UT-A1). Phosphorylation at Ser486 was detected in kidney sections obtained from dDAVP-treated rats but not vehicle-treated rats. Preadsorption with immunizing phosphopeptide ablated signal. E: immunolabeling of pSer486-UT-A1 and total UT-A1/3 in consecutive kidney sections. pSer486-UT-A1/3 and UT-A1/3 were detected in collecting ducts only. F: double labeling for pSer486-UT-A1 (green) and AQP2 (red). After dDAVP treatment, AQP2 is mainly apical, whereas pSer486-UT-A1 is localized both apically and intracellularly.

Immunogold-EM localization of phosphorylated urea channel proteins.

Immunogold-EM localization of pSer84- and pSer486-phosphorylated urea transporters in rat inner medulla showed no appreciable labeling in the absence of dDAVP pretreatment (not shown). After 60-min dDAVP (1 ng) treatment of rats, there was abundant labeling of IMCD cells, the majority of which was intracellular (Fig. 8). A large proportion of the gold particles was observed in the Golgi complexes and Golgi vesicles in the apical region of IMCD cells. Only a few gold particles were observed in direct association with the apical plasma membrane. Some pSer-84 gold particles were observed within the basolateral membrane infoldings (not shown). Using a UT-A antibody that recognizes both UT-A1 and UT-A3 total protein, abundant plasma membrane labeling, including both the apical plasma membrane and basolateral plasma membranes, was detected.

Fig. 8.

Immunogold electron microscopy demonstrates that pSer84 UT-A and pSer486 UT-A are predominantly associated with intracellular compartments. A: overview of an IMCD cell in midregion of inner medulla labeled with pSer84 UT-A. B: higher magnification of the area highlighted in A. Few gold particles are directly associated with the apical plasma membrane (apm). Abundant labeling is observed in the subapical Golgi apparatus (g) and intracellular vesicles. C: further example of the association of pSer84 UT-A with the subapical Golgi apparatus. D: overview of an IMCD cell in midregion of inner medulla labeled with pSer486 UT-A. E: higher magnification of the area highlighted in D. Few gold particles are directly associated with the apm. Abundant labeling is observed in subapical vesicles. F: pSer486 UT-A is also abundant in the subapical Golgi apparatus. G: overview of a principal cell in IMCD labeled with L446 (UT-A1/3 specific antibody). Inset: area highlighted in box 1 and demonstrates labeling of the subapical Golgi apparatus. H: area corresponds to box 2 in G and demonstrates abundant labeling of the apical plasma membrane. I: area corresponds to box 3 in G and demonstrates that some gold particles are also observed in the basolateral plasma membrane. In all images, gold particles are 10 nm in diameter.


In this paper, we used two new phospho-specific antibodies recognizing phosphorylated serines in the collecting duct urea transporters UT-A1 and UT-A3 to investigate vasopressin-mediated regulation of urea transporter phosphorylation. One antibody recognizes the phosphorylated Ser84 site present in both UT-A1 and UT-A3, which was demonstrated by LC-MS/MS analysis of rat collecting ducts by Bansal et al. (2). The second antibody recognizes the phosphorylated Ser486 site, present in UT-A1 but not UT-A3, which was demonstrated originally by Hoffert et al. (16) and has been extensively studied by Blount et al. (6). The latter study used mutational analysis to infer a role for the Ser486 phosphorylation in the regulation of UT-A1.

UT-A1 and UT-A3 are proteins of 929 and 460 amino acids (Fig. 1) that are derived from the same gene (Slc14a2) as a result of incorporation of alternative exons that contain either a stop codon in the case of the shorter UT-A3 (exon 12) or a full open reading frame that allows splicing to exon 14 in the case of the longer UT-A1 (23 exons) (28, 33, 35). Since exons 1–11 are the same in UT-A1 and UT-A3, the corresponding amino acid sequence is the same for the first 459 amino acids of both proteins. Our L446 antibody was made to recognize the NH2-terminal tail of both proteins allowing assessment of both UT-A1 and UT-A3 abundances on one immunoblot. The two are discriminated on the basis of size (Fig. 3). Figure 1 illustrates the relationship between the two proteins and labels the two phosphorylation sites under study in this paper, viz Ser84 and Ser486. It is at present unclear whether UT-A1 and UT-A3 function independently as separate transporters or are part of a more complex structure combining UT-A1 and UT-A3 in a single heteromeric transporter. However, both can function independently as homomeric transporters with heterologous expression in oocytes (6, 13, 24) or cultured cells (6, 36). UT-A1 and UT-A3 have been shown to be coregulated, at least in long-term regulation of transporter abundance. Chronic conditions associated with downregulation of both UT-A1 and UT-A3 include potassium depletion (19), ureteral obstruction (21), extracellular fluid volume expansion (41), and glucocorticoid excess (22). Also, Dahl salt-sensitive rats show increased levels of both UT-A1 and UT-A3 compared with control rats (11). However, studies of different hydration states in animals showed that UT-A1 and UT-A3 expression can be regulated differently (1, 23).

Since Ser84 is present in both UT-A1 and UT-A3, it was of particular interest to determine whether the increased abundance of phosphorylated Ser84 demonstrated by mass spectrometry (2) was due to vasopressin-induced phosphorylation of either UT-A1, UT-A3, or both. The results shown in Figs. 3 and 5 demonstrated that indeed phosphorylation of this site is markedly increased in both urea transporters in response to the V2 receptor vasopressin analog dDAVP. Furthermore, this result demonstrates regulated phosphorylation of UT-A3, which has not been known previously to be a target for regulated phosphorylation. The time course of phosphorylation of both Ser84 and Ser486 of UT-A1, and Ser84 of UT-A3, in IMCD suspensions indicates a very rapid response, within 1 min of vasopressin addition. This finding is coherent with time course studies of urea transport in isolated, perfused IMCD segments showing the initial rise in urea permeability ∼40 s after vasopressin exposure (40). It is also consistent with UT-A1 phosphorylation studies using [32P] incorporation to show increased phosphorylation in the same time frame (44). The time course was similar for both sites and for Ser84 in both UT-A1 and UT-A3, suggesting that the same signaling pathways were rate limiting in each case. In addition, the time course of AQP2 phosphorylation at Ser256 has been shown to be similar (8, 15). Previous studies suggested that this rate-limiting process for the vasopressin response is at or before the level of cAMP generation (29, 40).

We also confirmed the strong regulation of phosphorylation at Ser84 and Ser486 using immunohistochemistry. In general, with confocal immunofluorescence, only IMCD cells were labeled and labeling included both the cell periphery and intracellular regions. Phosphorylated UT-A1/3 is not restricted to the apical plasma membrane, contrary to what was seen previously in the case of phosphorylated AQP2 at Ser269 (14, 27). Thus, the results do not provide evidence pointing to a specific role of phosphorylation of UT-A1 or UT-A3 in plasma membrane retention or regulation of endocytosis as for Ser269 phosphorylation of AQP2 (26). Supporting this view were our findings from immunogold electron microscopy using the two antibodies that showed virtually exclusive presence of phosphorylated UT-A in intracellular compartments, with much of the intracellular labeling associated with the Golgi apparatus and Golgi vesicles in the apical region of the cells (Fig. 8). Our data highlight the need for high-resolution imaging techniques for confirmation of confocal microscopy studies, especially when drawing conclusions regarding subcellular distribution. It remains possible that the phosphorylation at either site plays a role in exocytosis of UT-A1 or UT-A3 as proposed by Blount et al. (6) and that the phosphorylation is rapidly removed on delivery of the urea channels to the plasma membrane. Another possibility is that some portion of plasma membrane UT channels is phosphorylated, but that the phosphorylation is masked by proteins that bind to the phosphorylation sites or due to conformational changes in channel structure.

The amino acid sequence surrounding the Ser84 phosphorylation site is SKRRESELPRRA (Ser84 is underlined). As can be seen, the Ser84 site is in a basic region of the protein (surrounded by arginine and lysine moieties up- and downstream). Thus, it is likely that Ser84 is phosphorylated by one or more “basophilic” kinases (25). The amino acid sequence surrounding the Ser486 phosphorylation site is FPRRKSVFHIE (Ser486 is underlined). Thus, Ser486 is also in a basic region with arginines and lysines upstream from the phosphorylation site. This again points to basophilic kinases as candidates for phosphorylation of Ser486. Vasopressin is known to signal in part through activation of the basophilic kinase protein kinase A (PKA). The phosphorylation motif signature for PKA is x-(R/K)-(R/K)-x-(S/T)-x (25). (This terminology gives the amino acid sequence surrounding the target serine or threonine where R/K means either arginine or lysine.) In collecting ducts, vasopressin has also been found to signal via pathways involving several additional kinases including myosin light chain kinase (MLCK) (8), Akt or protein kinase B (31), and ERK1/2 (31). Among these, only Akt has the phosphorylation motif signature to be a candidate for the two phosphorylation sites under study in this paper, i.e., Ser84 and Ser486 of urea transporters. The Akt phosphorylation target signature is R-x-R-x-x-(S/T)-ϕ, where ϕ is a hydrophobic amino acid (25). MLCK has only one known target, viz. myosin regulatory light chain A and B isoforms where Thr18 and Ser19 are phosphorylated (sequence PQRATSNVFA) (17, 18). The signature for Erk1/2-mediated phosphorylation has been reported to be x-P-x-(S/T)-P-x (proline moieties both up- and downstream from target Ser or Thr) (25). No ERK1/2 candidate sites have been identified in UT-A1 or UT-A3 by mass spectrometry as of this writing, although other critical sites may remain undiscovered (42). In addition, because inhibitors of calmodulin block some aspects of vasopressin signaling (10), it is worthwhile to consider CaM-kinase II which is strongly expressed in the IMCD (39) as a candidate for phosphorylation at Ser84 and Ser486 of UT-A1/3. The CaM-kinase II target signature has been reported to be x-R-x-x-(S/T)-x-x. Both Ser84 and Ser486 have a basic amino acid in the -3 position, making them compatible with CaM-kinase II targets. Furthermore, Rinschen et al. (32) recently reported that vasopressin activates CaM kinase II in collecting duct cells. Finally, Stewart et al. (37) proposed that urea transport by UT-A3 may be regulated via phosphorylation by an acidophilic kinase casein kinase II. Its target motif signature is x-x-(S/T)-x-x-(E/D)-x (25). The only demonstrated phosphorylation sites in UT-A1/3 that are compatible with this motif are Ser-62 and Ser-63 (DLRSSDEDS), although this site has not been found to show an increase in phosphorylation in response to vasopressin (2).

PKA is capable of phosphorylating both Ser84 and Ser486 in vitro (2) but such a demonstration does not imply that PKA is the only or even the chief kinase responsible for phosphorylation at these sites. However, Zhang et al. (44) demonstrated that the moderately specific PKA inhibitor H-89 blocked vasopressin-induced phosphorylation of UT-A1, pointing to a role for PKA in regulation of sites present in UT-A1, presumably Ser486 and/or Ser499.


This study was supported by the Intramural Budget of the NHLBI (Project Z01-HL-001285). S. Hwang was supported by an American Physiological Society Undergraduate Summer Research Fellowship. M. M. Rinschen was supported by the Biomedical Exchange Program, Hannover and the Braun Foundation, Melsungen. R. A. Fenton is funded by the Danish Medical Research Council, The Lundbeck Foundation, and the Novo Nordisk Foundation.


No conflicts of interest, financial or otherwise, are declared by the author(s).


We acknowledge the professional skills and advice of Dr. C. A. Combs [Light Microscopy Core Facility, National Heart, Lung, and Blood Institute (NHLBI)] regarding light microscopy-related data in this paper. E. Merete Locke is thanked for excellent technical assistance in immunogold electron microscopy.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
View Abstract