Intracellular sites of AQP2 S256 phosphorylation identified using inhibitors of the AQP2 recycling itinerary

Abstract

Vasopressin (VP)-regulated aquaporin-2 (AQP2) trafficking between cytoplasmic vesicles and the plasma membrane of kidney principal cells is essential for water homeostasis. VP affects AQP2 phosphorylation at several serine residues in the COOH-terminus; among them, serine 256 (S256) appears to be a major regulator of AQP2 trafficking. Mutation of this serine to aspartic acid, which mimics phosphorylation, induces constitutive membrane expression of AQP2. However, the intracellular location(s) at which S256 phosphorylation occurs remains elusive. Here, we used strategies to block AQP2 trafficking at different cellular locations in LLC-PK1 cells and monitored VP-stimulated phosphorylation of S256 at these sites by immunofluorescence and Western blot analysis with phospho-specific antibodies. Using methyl-β-cyclodextrin, cold block or bafilomycin, and taxol, we blocked AQP2 at the plasma membrane, in the perinuclear trans-Golgi network, and in scattered cytoplasmic vesicles, respectively. Regardless of its cellular location, VP induced a significant increase in S256 phosphorylation, and this effect was not dependent on a functional microtubule cytoskeleton. To further investigate whether protein kinase A (PKA) was responsible for S256 phosphorylation in these cellular compartments, we created PKA-null cells and blocked AQP2 trafficking using the same procedures. We found that S256 phosphorylation was no longer increased compared with baseline, regardless of AQP2 localization. Taken together, our data indicate that AQP2 S256 phosphorylation can occur at the plasma membrane, in the trans-Golgi network, or in cytoplasmic vesicles and that this event is dependent on the expression of PKA in these cells.

NEW & NOTEWORTHY Phosphorylation of aquaporin-2 by PKA at serine 256 (S256) occurs in various subcellular locations during its recycling itinerary, suggesting that the protein complex necessary for AQP2 S256 phosphorylation is present in these different recycling stations. Furthermore, we showed, using PKA-null cells, that PKA activity is required for vasopressin-induced AQP2 phosphorylation. Our data reveal a complex spatial pattern of intracellular AQP2 phosphorylation at S256, shedding new light on the role of phosphorylation in AQP2 membrane accumulation.

INTRODUCTION

Protein phosphorylation is a crucial posttranslational modification that can affect the trafficking of proteins between subcellular compartments as well as to and from the plasma membrane. Phosphorylation of the vasopressin (VP)-sensitive water channel aquaporin-2 (AQP2) plays a central role in water reabsorption across kidney principal cells (1–3). In the absence of phosphorylation at residue serine (S)256, AQP2 constitutively recycles between intracellular compartments and the plasma membrane (4, 5). Upon S256 phosphorylation, which inhibits AQP2 internalization, there is increased accumulation of AQP2 in the plasma membrane, resulting in an increase in collecting duct water permeability (1, 6). Dysregulation of this mechanism leads to disorders of water balance (7–12).

Several studies have shown that AQP2 can be phosphorylated on its COOH-terminus at S256, S261, S264, and S269 (1, 2, 13–15); their phosphorylation state serves specific roles in AQP2 regulation and trafficking. In this study, we focused on S256, a major regulator of AQP2 membrane trafficking. When VP stimulates VP receptor type 2 (V2R) in principal cells, adenylyl cyclase activity increases, leading to increased activity of cAMP-sensitive protein kinase A (PKA) (16). PKA subsequently phosphorylates S256, which leads to AQP2 membrane accumulation and water reabsorption. Although PKA is well accepted as the major kinase phosphorylating AQP2 S256, it has been suggested that other kinases may also phosphorylate AQP2 at this S256 residue (17, 18). Mutation of S256 to alanine (A) so that the site could not be phosphorylated caused AQP2 to be mainly cytoplasmic and no longer respond to VP. Conversely, mutating this serine to aspartic acid (D), which mimics phosphorylation, shifts AQP2 expression to the plasma membrane by inhibiting AQP2 internalization (endocytosis) from the cell surface (4, 15, 19, 20).

In addition to AQP2 phosphorylation leading to membrane trafficking, the cytoskeleton has been shown to play an essential role in the regulation of AQP2 intracellular trafficking and its membrane accumulation (21–25). The degree of AQP2 membrane expression is determined by a balance between its exocytosis and subsequent endocytosis. After insertion into the plasma membrane, AQP2 is retrieved from the membrane via a clathrin-coated pit mechanism (5, 26), a process that also requires actin polymerization (23, 27). Endocytosed AQP2-positive vesicles then interact with microtubules leading to the accumulation of AQP2 at the periphery of the Golgi where many microtubules terminate (24, 28–30). Some AQP2 from the perinuclear region near the Golgi apparatus is recycled and redirected back to the plasma membrane (28), but to accumulate in the plasma membrane, AQP2 endocytosis must be reduced by actin depolymerization (25, 31–36).

Many studies have separately described the role of the cytoskeleton or phosphorylation in AQP2 trafficking, but less is known about the direct relationship between the intact cytoskeleton, vesicle trafficking, and AQP2 phosphorylation. Do microtubules need to be intact for AQP2 to be phosphorylated, and does phosphorylation occur only in a particular cellular location? Here, we used various strategies, including microtubule disruption, to interrupt the trafficking and recycling of AQP2 in distinct cellular locations and determine the ability of VP to induce phosphorylation of AQP2 at S256 at these intracellular sites.

MATERIALS AND METHODS

Antibodies and Reagents

All chemical reagents were purchased from Sigma-Aldrich (St. Louis, MO) with the exceptions below. Forskolin (Fk) was from Tocris (Minneapolis, MN). Rabbit anti-phosphorylated (p)S256-AQP2 (ab111346) was purchased from Abcam (Cambridge, MA), and rabbit anti-AQP2 antibody was from Alomone Labs (Jerusalem, Israel). EEA1 (No. 610456), GM130 (No. 610822), and clathrin (heavy chain; No. 6104999) antibodies were purchased from BD Biosciences (East Rutherford, NJ). Mouse anti-PKA α catalytic subunit (sc-28315) was purchased from Santa Cruz Biotechnology (Dallas, TX). Mouse anti-c-myc was secreted into the medium of a hybridoma cell culture [1–910.2 (9E10), CRL-1729, American Type Culture Collection] (37). All secondary antibodies and horseradish peroxidase (HRP)-conjugated antibody were purchased from Jackson Immunoresearch (West Grove, PA). Cell culture reagents were obtained from ThermoFisher Scientific (Waltham, MA), and PBS was from Boston Bioproducts (Ashland, MA). FBS was from Atlanta Biologicals (Flowery Branch, GA). Paraformaldehyde (PFA) was purchased from Electron Microscopy Sciences (Hatfield, PA). The protease inhibitor complete Protease Inhibitor Cocktail was from Roche Diagnostics (Mannheim, Germany).

Cell Culture

LLC-AQP2 and LLC-S256A AQP2 cells expressing wild-type (WT) AQP2 or AQP2 with a serine-to-alanine point mutation at amino acid residue 256, respectively, and nontransfected LLC-PK1 cells were grown in DMEM with 10% FBS and additional l-glutamine (2 mM) at 37°C in a 5% CO₂ atmosphere as previously described (38, 39). Cells were split 1:10 every 3 days. DAPI staining tests were performed to confirm that cells were free of mycoplasma contamination. Rab11-green fluorescent protein (GFP) cells were generated by transfecting cells with GFP-Rab11 plasmid (a generous gift from Dr. Vladimir Marshansky) using 1 µg DNA in 3 µL Lipofectamine 2000 (Invitrogen, Waltham, MA) and grown under the same conditions as detailed above for 24 h before being subjected to cold block treatment. A PKA-null LLC-AQP2 cell line lacking both α and β PKA catalytic subunits was generated using CRISPR/Cas9. These cells [LLC-AQP2, PKA double knockout (dKO)] were grown under similar conditions as described above.

Generation of LLC-AQP2, PKA dKO Cells Using CRISPR/Cas9

LLC-AQP2 cells were transfected with pCMV-cas9-orange fluorescent protein (OFP) plasmids containing a guide RNA (gRNA) sequence specific for the β catalytic subunit of PKA (Prkacb). Plasmids were constructed using the Geneart kit (Invitrogen) and guide RNA sequence 5′- CGTTCATACCAAAGTTCAG-3′ in the Prkacb gene. Plasmids were transfected with Lipofectamine 3000 (Invitrogen) according to the manufacturer’s instructions. OFP-expressing cells were sorted into 96-well plates (∼1 cell per well) using a FACS Aria II cell sorting machine at the CRM Flow Cytometry Core Facility at Massachusetts General Hospital. All clones were grown under conditions similar to LLC-AQP2 cells. Sanger sequencing verified a 5-bp deletion in the Prkacb CRISPR target site of 5′-TTCAG-3′ at the end of the gRNA sequence. To generate dKO of α and β subunits, a clone with correct deletion of the β subunit was selected and transfected with plasmids targeting the α catalytic subunit of PKA (Prkaca). The methods were the same as described above and used the gRNA sequence 5′- ACACCTGCGGAGGATCGGG-3′ for Prkaca. These cells were again sorted by OFP expression. Lack of PKA α catalytic subunit expression was confirmed by Western blot analysis. Total PKA activity of the dKO PKA catalytic subunit cells was tested with the PKA Kinase Activity Kit (Enzo, Farmingdale, NY).

PKA Assay

PKA activity was analyzed using the PKA Kinase Activity Kit (Enzo Life Sciences) as previously described (17). Briefly, LLC-AQP2 cells and LLC-AQP2, PKA dKO cells were starved for 2 h before treatment. Cells were incubated at 37°C with and without VP (10 nM, 15 min). After incubation, cells were lysed in buffer according to the manufacturer’s protocol. The protein concentration of the lysates was determined with a BCA protein assay (ThermoFisher Scientific). Samples were diluted (0.1–0.3 µg) in the provided buffer to optimize the measurement of absorbance. The absorbance of each sample was measured using the DTX880 Multimode Plate Reader (Beckman Coulter, Pasadena, CA). PKA activity is expressed as the absorbance normalized to protein loaded per well (absorbance per µg protein). Each condition was performed in triplicate, and the results were averaged from three independent experiments.

Immunocytochemistry

Cells were grown on coverslips in 24-well plates. Cells were seeded on 10-mm coverslips (Ted Pella, Redding, CA) at 38,000 cells per well. Experiments were performed when cells reached 70% confluency. Before any experiment was performed, the medium was replaced with DMEM without FBS for 2 h. After this “starvation” period, cells were pretreated with colchicine (10 mM, 2 h) or methyl-β-cyclodextrin (MBCD; 10 mM,15 min). After pretreatment incubation, VP (10 nM) and Fk (1 mM) (VP/Fk) were added to both untreated and treated cells for 15 min. Cells were then fixed in 4% PFA in PBS and 5% sucrose for 20 min. Fixed coverslips were permeabilized with 0.1% Triton X-100 in PBS (5 min), washed three times in PBS, and blocked in 1% BSA for 10 min. The coverslips were then incubated with the primary antibody anti-pS256-AQP2 (10 µg/mL) for 1 h at room temperature. Under these conditions, anti-pS256-AQP2 specifically recognized pAQP2 at the plasma membrane and on intracellular vesicles (Supplemental Fig. S1). After three washes in PBS, coverslips were incubated with Alexa 488-conjugated donkey anti-rabbit antibody (DAR-Alexa 488; 5 µg/mL) for 1 h at room temperature. After three washes (5 min) in PBS, coverslips were exposed for 1 h to the cell supernatant containing c-myc antibody. After three washes, cells were incubated in the secondary antibody Alexa 568-conjugated donkey anti-mouse (DAM-Alexa 568; 10 µg/mL, Abcam). After incubation, coverslips were incubated 8 min with Alexa 647-conjugated wheat germ agglutinin (Alexa 647 WGA; 0.5 µg/mL, ThermoFisher Scientific). After the membranes were stained, coverslips were washed three times in PBS and mounted on glass slides using Vectashield with DAPI (Vector Laboratories, Burlingame, CA). Images were taken using a Zeiss LMS 800 confocal system equipped with a Plan-apochromat ×63/1.4 oil DIC UV-VIS-IR lens and operated using Zen 2.6 blue edition software (Carl Zeiss, Peabody, MA). The set of images including LLC-AQP2, PKA dKO cells was acquired using a Nikon 90i epifluorescence microscope system equipped with a Plan apo ×40/1.0 oil DIC H lens. Images were acquired using NIS-Elements AR 4.6 software (Nikon Instruments, Edgewood, NY). All acquired images were incorporated into the figures shown in this report using Photoshop CS5 software (Adobe, San Jose, CA). AQP2 quantification at the plasma membrane was performed using Volocity software v. 6.3 (Perkin-Elmer, Waltham, MA) as previously described (17, 39). Briefly, we highlighted regions of interest (ROIs) of each cell: the nucleus, the cytoplasm, and finally the plasma membrane. Staining of the plasma membrane by Alexa 647 WGA was used to distinguish the ROIs occupied by the plasma membrane and cytoplasm (an example is provided in Supplemental Fig. S2). DAPI fluorescence intensity was used to determine the ROI occupied by the nucleus. The relevant channel fluorescence intensity found in the nucleus was used as the background value. The mean fluorescence intensity of the plasma membrane was determined using at least 20 nonadjacent cells from 3 independent experiments.

Colocalization of AQP2 and Intracellular Compartment Markers

Intracellular compartments in cold block-treated cells were identified using antibodies against the early endosome marker (EEA1; 1.25 µg/mL), trans-Golgi network (TGN; clathrin, 0.5 µg/mL), or cis-Golgi marker (GM130, 2.5 µg/mL). Colocalization between recycling endosomes and AQP2 was analyzed in cells transfected with GFP-tagged Rab11. The GFP-Rab11 plasmid was transfected (1 µg DNA in 3 µL Lipofectamine 2000) as suggested by the company. After 24 h, cells were subjected to the cold block treatment.

The cold block to prevent vesicle trafficking out of the TGN was induced by incubating cells at 20°C for 2 h. During the last 15 min of the incubation, cells were treated with VP/Fk. Cells were fixed in 4% PFA and 5% sucrose at 20°C for 20 min and then at room temperature for 10 min. Next, cells were permeabilized with 0.25% SDS in PBS for 4 min and blocked with 1% BSA in PBS for 20 min. Total AQP2 and anti-pS256-AQP2 antibodies were applied as described above. Cells were incubated with EEA1, clathrin, and GM130 antibodies overnight at 4°C. After three washes in PBS, cells were incubated with secondary anti-rabbit IgG conjugated to Alexa 488 (6 μg/mL, Jackson ImmunoResearch Laboratories) before being washed and mounted in Vectashield with DAPI.

Images were acquired on a Zeiss LSM 800 confocal microscope equipped with an Airyscan detector (lens: Apochromat ×63, numerical aperture: 1.4). Three-dimensional images of each cell were taken. Each z-stack of six images (650 nm) was processed using Zen software (ZEN Digital Imaging for Light Microscopy, RRID:SCR 013672). The three-dimensional Airyscan images were processed, deconvoluted, and then analyzed using Zen software. All final images were assembled using Adobe Photoshop CS5 software.

Western Blot Analysis

LLC-AQP2 and LLC-AQP2, PKA dKO cells were plated at 150,000 cells/60-mm Petri dish. When cells reached confluency, they were starved and then treated with drugs as described above. After treatment, cells were lysed in 300-µL cold RIPA buffer (Boston Bioproducts) supplemented with a protease inhibitor cocktail (Complete Mini, Roche, Basel, Switzerland), EDTA (1 mM), and phosphatase inhibitors NaF (1 mM) and sodium orthovanadate (1 mM). The lysates were incubated at 4°C for 15 min and centrifuged for 10 min at 17,000 g. The protein concentration was determined using a BCA protein assay kit (ThermoFisher Scientific). The concentration of protein per well used in this experiment was predetermined to allow linear dynamic phosphorylation level detection (40, 41). Band intensities using anti-pS256-AQP2 antibody increased linearly from 5 to 20 µg protein per sample. Therefore, we subsequently used 10 μg of solubilized protein diluted with RIPA buffer, NuPage SDS sample buffer, and NuPage sample reducing agent to make up the final loading sample. Proteins were denatured at 70°C for 10 min. After denaturation, samples were centrifuged for 10 min. The samples and 10 µL of SeeBluePlus2 protein ladder (NuPage System, Invitrogen) were loaded into each well of NuPage 4–12% bis-Tris gels (ThermoFisher Scientific). After electroporation, samples were transferred onto Immun-blot PVDF membranes (Bio-Rad, Hercules, CA). Membranes were blocked with PBS-0.05% Tween 20 (PBS-T) containing 5% nonfat milk for 1 h at room temperature. Membranes were subsequently washed three times with PBS-T and incubated in primary antibody (anti-pS256-AQP2, 2 μg/mL) for 2 h at room temperature or at 4°C overnight. Anti-pS256-AQP2 detected two bands: the expected 28-kDa pS256 AQP2 band and a nonspecific band (at 95 kDa) in samples treated with VP/Fk. Membranes were then washed five times (PBS-T) and incubated with secondary antibody (DAR-HRP, 0.16 μg/mL) in 5% milk for 1 h at room temperature. After three washes with PBS-T, membranes were incubated in enhanced chemiluminescence substrate (ECL, Western lightning ECL, Perkin-Elmer). The intensity signal on the membranes was acquired using the G-box mini imaging system and GeneSys image acquisition software (Syngene, Frederick, MD). After the ECL reaction, membranes were washed five times with PBS-T, stripped by incubating in stripping solution (ThermoFisher Scientific) for 15 min, rinsed three times with PBS-T, and then blocked again with 5% milk and PBS-T. The efficacy of the stripping procedure was tested by exposing the membrane to DAR-HRP alone. After confirmation that the stripping was successful, membranes were rinsed with PBS-T and reprobed with anti-AQP2 or total AQP2 (0.15 μg/mL) antibodies as described above. Expression of PKA α catalytic subunits in LLC-AQP2, PKA dKO cells was compared with WT cells (LLC-AQP2) by Western blot analysis using similar steps to those described above. LLC-AQP2 cells (10 μg) and LLC-AQP2, PKA dKO (5−10 μg) lysed protein was loaded in wells. After transfer, the membrane was stained following the Pierce Reversible Protein Stain kit (ThermoFisher Scientific) protocol. After being stripped, the membrane was blocked with 5% milk and PBS-T and then incubated with anti-PKA α catalytic subunit antibody (4 µg/mL). After incubation, the membrane was washed three times in PBS-T and then incubated with secondary antibody (DAM-HRP, 0.16 μg/mL) in 3% milk for 1 h at room temperature. After three washes with PBS-T, the membrane signal was detected with ECL.

cAMP Activity Assay

Intracellular cAMP levels were analyzed using the Cyclic AMP direct Enzyme Immunoassay kit (Arbor Assays, Ann Arbor, MI). LLC-AQP2 cells (10,000 cells per well) were plated in 96-well plates until they reached confluency. Cells were pretreated with or without colchicine (10 µM) for 2 h. Thirty minutes before the end of the treatment, we added IBMX (1 mM) for 15 min and then VP/Fk (10 nM/1 mM) or PBS alone. Cells were solubilized, and cAMP levels were quantified according to the manufacturer’s instructions. Results were measured using a DTX880 Multimode plate reader (Beckman Coulter). The assay was performed three times independently, and each treatment was done in triplicate.

Mouse Kidney Tissue Preparation and Immunohistochemistry

Animal experiments were approved by the Institutional Committee on Research Animal Care, following the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Adult C57BL/6J mice received a single intraperitoneal injection of colchicine (0.5 mg/100 g body wt) or saline. After 2 h, mice were anesthetized by 2% isoflurane inhalation. The kidneys were perfused with PBS (37°C) for 1 min via the left cardiac ventricle and then perfused 3 min with PFA (4%)-lysine-periodate fixative (PLP). After the kidneys were harvested, they were incubated in PLP solution overnight (38, 39, 42). After five washes in PBS, tissues were cryoprotected in PBS and 30% sucrose overnight before being embedded in Tissue-Tek O.C.T. compound 4583 (Sakura Finetek, Torrance, CA). Cryosections (5 μm) were cut using a Leica CM3050S microtome (Buffalo Grove, IL) and placed onto positively charged slides (Denville Scientific, Holliston, MA). For immunostaining, sections were rehydrated in PBS for 20 min and then permeabilized with 0.1% Triton X-100 for 5 min. After three washes, nonspecific binding was blocked with 1% BSA in PBS for 10 min. The kidneys were incubated with anti-pS256-AQP2 (10 µg/mL) overnight at 4°C. After three washes in PBS, coverslips were incubated with goat anti-AQP2 antibody (C17; 0.4 µg/mL, Santa Cruz Biotechnology) at room temperature. After three washes in PBS, tissue samples were incubated with Alexa 488-conjugated donkey anti-rabbit antibody (2.4 µg/mL) and CY3-conjugated donkey anti-goat antibody (1.9 µg/mL) for 1 h at room temperature. After three washes in PBS, the slide was incubated 8 min with Alexa 647 WGA (0.5 µg/mL). After three rinses in PBS, sections were mounted using Vectashield with DAPI (Vector Laboratories). Images were taken using a Zeiss LSM 800 confocal microscope.

Statistical Analyses

Most of the data are expressed as means ± SD. Statistical analyses were performed using one-way ANOVA with a post hoc Tukey’s test. A two-tailed Student’s t test was also used when appropriate. For the experiments on the mouse kidney sections, a representative group of cells from images of several tissues was chosen for the AQP2 distribution analyses. To quantify the cellular distribution of AQP2, fluorescence intensity was measured along axes cross-sectioning the cells. Axes were chosen such that they passed adjacent to the nucleus without intersecting it. Fluorescence distributions for each cell were area normalized and aligned at the apical membrane on a common axis. To provide a visual representation of the distributions, the fluorescence intensity at each position was averaged and plotted with the error of the mean. The centroids of the fluorescence distributions were calculated. An equal variance two-tailed Student’s t test was used to study the change in centroid position due to colchicine treatment.

RESULTS

S256 Phosphorylation of AQP2 Can Occur at the Plasma Membrane

Under baseline conditions, AQP2 is located mainly in the cytoplasm on vesicles (Fig. 1A), and a similar distribution of pS256 was seen, although the staining was much less intense (Fig. 1B). After VP/Fk treatment, AQP2 distribution was shifted to the plasma membrane (Fig. 1D), as previously shown in our LLC-PK1 cells (38, 42), and there was also an increase in the intensity of pS256 AQP2 staining, detected now mostly at or near the plasma membrane (Fig. 1E). Quantification of total and pS256 AQP2 at the membrane showed a more than twofold increase in the presence of VP/Fk compared with basal conditions (Fig. 1, G and H). After VP/Fk treatment, we noted upper bands around 95 kDa in AQP2 pS256 blotted membranes but determined that it was irrelevant to S256 phosphorylation because we saw the same band in our mutated S256A cells treated with VP/Fk. These mutated S256A cells have serine residues replaced by alanine and are unable to be phosphorylated at this site (Supplemental Fig. S3). We, therefore, removed the upper, nonspecific 95 kDa band from subsequent figures. Determining its identity is the subject of ongoing work.

Figure 1.

Vasopressin (VP)/forskolin (Fk) causes phosphorylated serine 256 (pS256) aquaporin-2 (AQP2) membrane accumulation. Immunodetection using c-myc and specific pS256 AQP2 antibody was performed in LLC-AQP2 cells stably expressing c-myc-tagged AQP2 to detect total AQP2 (tAQP2; A and D) or phosphorylated AQP2 (pS256 AQP2; B and E), respectively. Colocalization of both tAQP2 (red) and pS256 AQP2 (green) is shown in the merged column (C and F). This image is representative of three independent experiments. Scale bars = 10 µm. The distribution of pS256 AQP2 and tAQP2 at the plasma membrane was quantified. The results are expressed as the ratio of tAQP2 or pS256 AQP2 at the plasma membrane to tAQP2 in the cell (G and H, respectively). These results are the average of three independent experiments. More than 40 cells were quantified for each condition. Values are means ± SD; n = 3. ****P < 0.0001.

When cells were treated with MBCD, a nonspecific endocytosis blocker, AQP2 accumulated on the plasma membrane, as previously shown (Fig. 2A) (5, 43). However, most pS256 AQP2 labeling was still seen in the cytoplasm (Fig. 2, B and C), similar to CT cells (Fig. 1B). Its accumulation in the plasma membrane, induced by MBCD, did not increase S256 phosphorylation of AQP2, confirming our prior findings that AQP2 membrane accumulation can be independent of S256 phosphorylation when endocytosis is inhibited (5). However, upon exposure of the MBCD-treated cells to VP/Fk, there was a striking increase of pS256 AQP2 detected at the plasma membrane (Fig. 2, D–F) and a significant increase in total cellular pS256 AQP2 detected by Western blot analysis that was of similar magnitude to that seen in cells not exposed to MBCD (Fig. 2, G and H). These data indicate that AQP2 phosphorylation can occur even when AQP2 is already inserted into the plasma membrane. This finding prompted us to use other modulators of AQP2 trafficking and cytoskeletal disruptors to further test our hypothesis that phosphorylation of AQP2 can occur at various cellular sites.

Figure 2.

Methyl-β-cyclodextrin (MBCD), a nonspecific endocytosis inhibitor, causes aquaporin-2 (AQP2) membrane accumulation without phosphorylation. Immunodetection using c-myc and specific phosphorylated serine 256 (pS256) AQP2 antibodies was performed in LLC-AQP2 cells stably expressing c-myc-tagged AQP2 to detect total AQP2 (tAQP2; A and D) or phosphorylated AQP2 (pS256 AQP2; B and E), respectively. MBCD treatment resulted in the accumulation of tAQP2 in the plasma membrane, but S256 AQP2 staining was weak in control cells (B). However, after vasopressin (VP)/forskolin (Fk) treatment, a strong AQP2 S256 signal was detected at the plasma membrane (E). Colocalization of both tAQP2 (red) and pS256 AQP2 (green) is shown in the merged column (C and F). This image is representative of three independent experiments. Scale bars = 10 µm. Western blot analysis of LLC-AQP2 cells was performed (G). Cells were pretreated 30 min with MBCD before VP/Fk was applied. pS256 was detected using specific pS256 AQP2 antibody (G, top). After membrane detection, the membrane was stripped and reprobed with specific antibody against AQP2 (G, bottom). Band intensities were analyzed (H). The results are the average of six independent Western blot analyses performed in duplicate. Values are means ± SD; n = 6. ****P < 0.0001 and ***P = 0.0004.

AQP2 Phosphorylation Is Induced by VP/Fk on Vesicles Scattered Throughout the Cytoplasm After Microtubule Disruption

To investigate the role of the microtubule cytoskeleton on AQP2 phosphorylation, we treated cells with colchicine, a disruptor of microtubules. Following colchicine treatment, AQP2 was localized on vesicles distributed throughout the cytoplasm and VP/Fk no longer produced the significant membrane accumulation of AQP2 seen in cells not exposed to colchicine (Fig. 3, D–G). Immunofluorescence imaging produced a noticeable increase in the staining intensity of the scattered vesicles with pS256 AQP2 antibody (Fig. 3, B and E), and Western blot analysis provided more quantitative evidence that VP/Fk induces an increase of S256 phosphorylation after colchicine treatment, similar to that found in control cells (Fig. 3H). The canonical signaling pathway appeared to be intact, because colchicine treatment did not affect the VP/Fk-induced increase in intracellular cAMP as quantified by ELISA (Fig. 3I).

Figure 3.

The increase of phosphorylated serine 256 (pS256) aquaporin-2 (AQP2) at the plasma membrane treated with vasopressin (VP)/forskolin (Fk) is abolished in the presence of colchicine (Col). Cells were pretreated 30 min with colchicine and then incubated in the presence or absence of VP/Fk. The c-myc and pS256 AQP2 antibodies detected total AQP2 (tAQP2; A and D) and pS256 AQP2 (B and E), respectively. Colocalization of both tAQP2 (green) and pS256 AQP2 (red) is shown in the merged column (C and F). Colchicine treatment resulted in a scattering of AQP2-containing vesicles throughout the cytoplasm (A and D), and the intensity of staining with S256 AQP2 was markedly increased on VP/Fk treatment (B vs. E). This image is representative of three independent experiments. Scale bars = 10 µm. The distribution of tAQP2 and pS256 AQP2 at the plasma membrane was quantified (G) and showed that VP did not result in an increase in plasma membrane AQP2 (either total or S256 phosphorylated) after microtubules were disrupted by colchicine and, in fact, there was a significant decrease in pS256 AQP2 at the membrane (G). However, VP/Fk treatment did result in a significant increase in the level of phosphorylated AQP2 in the cell as detected by Western blot analysis (H). The results are the ratios of tAQP2 or pS256 AQP2 at the membrane to all AQP2 in the cell. These results are the average of three independent experiments. More than 20 cells were quantified for each condition in each experiment. Values are means ± SD; n = 3. ***P = 0.0002. Western blot analysis of LLC-AQP2 cells was performed (H). pS256 and tAQP2 were also studied by Western blot analysis on cells treated as described above. After the detection of pS256 AQP2 using specific pS256 AQP2 antibody (H, top), the membranes were stripped and reprobed with specific antibody against tAQP2 (H, bottom). Band intensities were analyzed. The result is the average of eight independent Western blot analyses. Values are means ± SD; n = 8. *P = 0.0196 and **P = 0.0028. Intracellular cAMP levels in LLC-AQP2 cells pretreated with colchicine followed by VP/Fk or not were measured (I). This result is the average of three independent experiments (n = 3, *between CT and VP indicates P = 0.0409 and *between Col and Col + VP/FK indicates P = 0.0134). Values are means ± SD. NS, not significant.

We continued our study by using another drug that affects microtubules. Taxol, a microtubule stabilizer, prevents microtubule assembly but does not cause disassembly. In our cells, this treatment also caused scattering of AQP2-containing vesicles throughout the cytoplasm similar to colchicine, and AQP2 again no longer shifted its location to the plasma membrane upon VP/Fk treatment (Fig. 4, A and C). Similar to colchicine-treated cells, despite AQP2 being primarily localized in scattered cytoplasmic vesicles after taxol treatment, S256 continued to be phosphorylated after VP/Fk treatment. To further study whether AQP2 can be phosphorylated in other subcellular compartments, we used bafilomycin for our next experiment. Bafilomycin inhibits V-ATPase activity and indirectly affects endosomal trafficking along microtubules. We have previously shown that after bafilomycin treatment, AQP2 accumulated in perinuclear patches that partially overlap with the TGN (28, 44), and AQP2 no longer accumulates at the plasma membrane with VP/Fk treatment (Fig. 4, D–F). pS256 also concentrated in the TGN, as expected (Fig. 4E). Here again, we found that after bafilomycin treatment of cells, AQP2 S256 phosphorylation continued to occur upon incubation with VP/Fk (Fig. 4G). These data confirm earlier observations that functional microtubules are necessary for VP-induced plasma membrane accumulation of AQP2 to occur (25, 27, 29, 45), and our data imply that the S256 phosphorylation machinery can still interact with AQP2 on scattered cytoplasmic vesicles after inhibition of microtubule function whether by depolymerization with colchicine or stabilization with taxol or when endosomal trafficking along the microtubules is inhibited.

Figure 4.

The increase of phosphorylated serine 256 (pS256) aquaporin-2 (AQP2) in LLC-AQP2 cells in the presence of vasopressin (VP)/forskolin (Fk) is not affected by bafilomycin (Baf) or taxol. Cells were pretreated 30 min with taxol (A–C) or bafilomycin (D–F) and then incubated with VP/Fk. The c-myc and pS256 AQP2 antibodies detected total AQP2 (tAQP2; A and D) and pS256 AQP2 (B and E), respectively, in the scattered vesicles (after taxol; A–C) or in a perinuclear patch (after bafilomycin; D–F). The location of both tAQP2 (green) and pS256 AQP2 (red) shown in the merged column (C and F) showed a mixture of vesicles either labeled predominantly with one antibody or the other or costained with both antibodies. This image is representative of three independent experiments. Scale bars = 10 µm. Western blot analysis was performed on LLC-AQP2 cells incubated with or without VP/Fk (G). Cells were pretreated for 30 min with either bafilomycin or taxol before VP/Fk stimulation. pS256 was detected using specific anti-pS256 AQP2 antibody (G, top). After being stripped, the membrane was reprobed with specific antibody against AQP2 (G, bottom). Band intensities were analyzed (H). The result is the average of eight independent Western blot analyses. Values are means ± SD; n = 8. ****P < 0.0001, **P = 0.0011, and *P = 0.0196.

AQP2 Is Phosphorylated in Subcellular Compartments That Accumulate After Blockade of Recycling and Exocytosis With Cold (20°C) Treatment

Similar to bafilomycin, we used a 20°C cold block to induce AQP2 patch formation in the TGN (arrows in Fig. 5, A–C). Under cold block conditions, AQP2 exocytosis from the TGN was inhibited, whereas AQP2 endocytosis continued to occur, eventually forming a patch in the TGN, as shown in Fig. 5A. Most of the pS256 was also localized in the TGN under these conditions (Fig. 5B). After VP/Fk treatment, AQP2 phosphorylation at S256 increased significantly compared with controls not exposed to VP/Fk (Fig. 5D).

Figure 5.

The increase of phosphorylated serine 256 (pS256) aquaporin-2 (AQP2) in LLC-AQP2 cells treated with vasopressin (VP)/forskolin (Fk) is not affected by cold block inhibition of trafficking. Immunodetection of total AQP2 (tAQP2) and pS256 AQP2 in LLC-AQP2 cells was performed (A–C) in cold-blocked cells. Cells were exposed for 2 h to a 20°C temperature block to induce formation of a perinuclear patch by blocking the exocytotic pathway at the level of exit from the trans-Golgi network (TGN). The c-myc and anti-pS256 AQP2 antibodies detected tAQP2 (A) and pS256 AQP2 (B), respectively. Colocalization of both tAQP2 (red) and pS256 AQP2 (green) is shown in the merged column (C). This image is representative of three independent experiments. Scale bar = 10 µm. pS256 and tAQP2 were studied by Western blot analysis on cells treated as described above (D). After the detection of pS256 AQP2 using specific anti-pS256 AQP2 antibody (D, top left), the membranes were stripped and reprobed with specific antibody against tAQP2 (D, bottom left). Band intensities were analyzed (D, right). The result is the average of eight independent western blot analyses. Values are means ± SD; n = 8. **P = 0.0012 and *P = 0.0255.

Although cells were incubated at 20°C to inhibit exocytosis, we simultaneously disrupted microtubule function with colchicine and taxol and again found that AQP2 S256 was increased by VP/Fk (Fig. 6, A and B). We also used MBCD in cold-blocked cells and detected an increase in S256 phosphorylation with VP/Fk treatment (Fig. 6A) that was similar to that seen when cells were incubated at 37°C (Fig. 2G). Combining these results, we conclude that AQP2 can be phosphorylated independently of its microtubule cytoskeletal association and that phosphorylation occurs within the endo- and exocytosis trafficking and recycling pathways, including in the TGN domain, as well as at the plasma membrane.

Figure 6.

The phosphorylation of serine 256 (S256) aquaporin-2 (AQP2) in LLC-AQP2 cells due to vasopressin (VP)/forskolin (Fk) under cold block conditions is not affected by methyl-β-cyclodextrin (MBCD), colchicine, or taxol. LLC-AQP2 cells were incubated at 20°C for 2 h, and MBDC, colchicine (A) or taxol (B) was added 30 min before further incubation with or without VP/Fk. Western blot analysis of LLC-AQP2 cells was performed and showed that none of the drugs affected VP/Fk-induced S256 phosphorylation under these conditions. Phosphorylated S256 (pS256) AQP2 was detected using a specific antibody (A and B, top left). After being stripped, the membranes were reprobed with specific antibody against AQP2 (A and B, bottom left). Band intensities were analyzed (A and B, right). The result is the average of eight independent Western blot analyses for A (n = 8, ****P < 0.0001) and six independent Western blot analyses for B (n = 6, ***P = 0.002 and **P = 0.0095). Values are means ± SD.

PKA Is the Major Kinase Phosphorylating AQP2 at the S256 Residue Upon VP Treatment

PKA is well accepted to be the major kinase activated by VP that induces phosphorylation of AQP2 at S256 (1, 2, 20, 46). However, while VP signaling is predominantly PKA dependent, VP can induce activation of other downstream kinases, and AQP2 S256 can be phosphorylated and regulated in response to VP even in the absence of PKA (17, 18). To evaluate whether other kinases might phosphorylate AQP2 S256 in different subcellular compartments under our experimental conditions, we generated a stable LLC-AQP2 cell line without PKA α and β catalytic subunits (LLC-AQP2, dKO). Using CRISPR, we first generated cells with the PKA β catalytic subunit knocked out and confirmed a 5-bp deletion of the target gene (Prkacb) with Sanger sequencing (as there is no available commercial antibody against the PKA catalytic β subunit in Sus scrofa). We then transfected the CRISPR plasmid targeting the PKA α catalytic subunit into these cells. We confirmed the dKO again with Sanger sequencing, following by blotting the cell lysates using an antibody against the PKA α catalytic subunit, to confirm dKO of the catalytic α and β subunits (LLC-AQP2, dKO; Fig. 7A). We then confirmed the lack of PKA activity in our dKO cells (Fig. 7B). After VP treatment, the dKO cells no longer shifted AQP2 localization to the membrane as seen in LLC-AQP2 WT cells (Fig. 7D, inset); it remained mainly cytoplasmic. To confirm that the lack of response to VP is not because AQP2 no longer constitutively traffic to and from the plasma membrane, we used MBCD to inhibit endocytosis in these cells. We found that MBCD continued to cause membrane expression in our dKO cells (Fig. 7E) similar to its effect on LLC-AQP2 WT cells (Fig. 7E, inset), implying intact mechanisms of AQP2 endocytosis and exocytosis in PKA dKO cells.

Figure 7.

Protein kinase A (PKA) is the major kinase that phosphorylates aquaporin-2 (AQP2) serine 256 (S256). Deletion of the PKA α catalytic subunit was confirmed using specific anti-PKA α subunit antibody (A, top) compared with LLC-AQP2 wild-type cells using Coomassie blue as a marker for protein loading (A, bottom). We then detected PKA activity in wild-type and LLC-AQP2, double knockout (dKO) cells using a commercially available PKA activity assay (B), and this result is an average of eight independent experiments (n = 8, **P = 0.0185). ns, not significant. We treated cells with DMSO (CT; C), vasopressin (VP; D), and methyl-β-cyclodextrin (MBCD; E). Using c-myc antibody, we localized AQP2 in our LLC-AQP2, dKO cells and compared them with wild-type cells shown in the smaller insets (C–E). These images are representative of three independent experiments. Using the phosphoserine antibody against S256 AQP2, we detected phosphorylation of AQP2 in our wild-type (F, top) and LLC-AQP2, dKO cells (F, bottom) with or without VP treatment. We then separately pretreated cells with MBCD, colchicine, taxol, or bafilomycin followed by VP or subjected cells to cold temperature treatment for 2 h followed by VP in the last 15 min. Phosphorylation of S256 was detected as described above and compared with total AQP2 (G). The results are representative of at least three independent experiments. Values are means ± SD. ns, not significant. *P < 0.05, with actual P values as shown.

We then examined whether phosphorylation of S256 occurred in our LLC-AQP dKO cells. We found that VP no longer induced S256 phosphorylation (Fig. 7F). We next shifted AQP2 localization to various cellular compartments in our dKO cells as described above, including the plasma membrane, cytoplasm, and TGN using MBCD, colchicine, taxol, bafilomycin, and cold block and found that S256 phosphorylation no longer occurred under any of these conditions (Fig. 7G). This confirmed that PKA is the major kinase phosphorylating S256 under our experimental conditions, although the role of other PKA-dependent kinase in this process cannot be ruled out. Although other kinases may be involved in phosphorylating S256, as previously shown in similar PKA-null cells (18), the slight increases in phosphorylation that we observed (Fig. 7G) did not reach the level of significance compared with LLC-AQP2 WT cells. Of note, here we performed all experiments with VP, rather than VP/Fk, and we found similar results as in cells treated with VP/Fk in terms of AQP2 phosphorylation and AQP2 localization under various drug treatments.

Perinuclear Phosphorylation of S256 Induced by VP/Fk Occurs in the cis-Golgi

To further identify more precisely in which juxtanuclear intracellular vesicles AQP2 S256 phosphorylation occurs, we used high-resolution confocal imaging to semiquantify the degree of colocalization of pS256 with other populations of AQP2 using various organelle markers. We used cold-blocked cells to reduce exocytosis and constitutive recycling to better capture the potential site(s) of S256 phosphorylation. Using three-dimensional quantification of AQP2 colocalization with various organelle markers, we found that VP/Fk treatment induced a significant increase of S256 phosphorylation in GM130-positive vesicles (Fig. 8, E and F), indicating possible phosphorylation events occurring in the cis-Golgi and possibly the TGN, since GM130 labeling partially overlaps with the TGN (47). Surprisingly, we did not observe a change of colocalization in pS256 AQP2 and EEA-1 (Fig. 8, A and B) and clathrin-positive vesicles after VP/Fk treatment (Fig. 8, C and D), and there was, interestingly, a decrease of colocalization in Rab11-positive vesicles (Fig. 8, G and H) under these conditions. These results suggest that S256 phosphorylation can occur upon recycling back to the Golgi apparatus and TGN and possibly during transit through the cis-Golgi. Although pAQP2 was clearly detected in other juxtanuclear vesicles under cold block conditions, we detected no significant increase of total AQP2 in EEA1-positive endosomes, clathrin-positive vesicles that are involved in protein exit from the TGN, or Rab11-positive recycling vesicles after VP/Fk treatment.

Figure 8.

Differential changes in the ratio of phosphorylated serine 256 (pS256) aquaporin-2 (AQP2) vs. total AQP2 (tAQP2) in various intracellular compartments after vasopressin (VP)/forskolin (Fk) treatment under cold block conditions. Coimmunodetection of intracellular markers (green; A, C, E, and G, left) and both tAQP2 or pS256 AQP2 in LLC-AQP2 cells with and without VP/Fk for 15 min was performed. Here, we show only confocal images of pS256 AQP2 without VP/Fk for simplicity (A, C, E, and G, middle). Cells were incubated 2 h at 20°C and then incubated in the presence or absence of VP/Fk for 15 min. Colocalization of pS256 AQP2 (red) and intracellular compartment makers (green) is shown in the merged column (A, C, E, and G, right). This image is representative of three independent experiments. Scale bars = 1.5 µm. The colocalization of pS256 AQP2 or tAQP2 with EEA1, clathrin, GM130, or Rab11-green fluorescent protein (GFP) in the presence or absence of VP/Fk was quantified (B, D, F, and H). The results are expressed as the ratio of tAQP2 or pS256AQP2 colocalized with the different compartment markers in the cell. Changes in the ratio of tAQP2 and p256 AQP2 in the different compartments after VP/Fk stimulation were quantified in the confocal images. These results are the average of three independent experiments. More than 20 cells were quantified for each condition in B, D, and F, and 10 cells were quantified for the pS256 AQP2 experiments shown in H. Values are means ± SD; n = 3. ***P = 0.0003 and **P = 0.0052. ns, not significant.

pS256 AQP2 Is Located Throughout the Cytoplasm of Kidney Principal Cells in Colchicine-Treated Mice

In Fig. 3, we show in cells that disruption of the microtubule network by colchicine affects AQP2 membrane accumulation but not its phosphorylation. Here, we sought to confirm this observation using a modified in vivo model (29) to examine the change of AQP2 distribution in kidneys of colchicine-treated mice (Fig. 9A). The plasma membrane of the epithelium was marked with WGA. Under control conditions, total AQP2 and pS256 AQP2 accumulated close to the apical pole, but after colchicine treatment, total AQP2 and pS256 AQP2 were diffusely localized throughout the cytoplasm, similar to our findings in cells in vitro and as previously reported for total AQP2 in the rat kidney (Fig. 3A) (29). Fluorescence intensity distributions, spanning the width of a cell from the lumen to the basolateral membrane, widened with colchicine treatment, consistent with redistribution of AQP2 from the apical pole to a more diffuse cellular distribution (Fig. 9B). The center of these distributions also shifted away from the apical membrane with colchicine treatment.

Figure 9.

Colchicine redistributes aquaporin-2 (AQP2) in the mouse kidney cytoplasm while maintaining serine 256 (S256) phosphorylation of AQP2. Mice were injected intraperitoneally with 0.5 mg/100 g colchicine or saline alone for 6 h. Kidney sections were stained for total AQP2 and phosphorylated S256 (pS256) AQP2. Alexa 647-conjugated wheat germ agglutinin (WGA) was used to stain membranes. Total AQP2 and pS256 AQP2 were both mainly localized at the apical pole of principal cells under control conditions but were found diffusely scattered throughout the cytoplasm after colchicine treatment (A). To quantify the cellular distribution of AQP2, fluorescence intensity was measured along axes cross-sectioning the cells. For each condition, these distributions were averaged and plotted with the error of the mean. Colchicine treatment led to a widening of the fluorescence intensity distributions across each cell. In the total AQP2 population and the pS256 AQP2 subset, the centroids of the distributions of the colchicine-treated group were shifted intracellularly compared with the centroids of distributions of the control group (B). These data are a representation of an average of three independent experiments of three controls and three colchicine-treated animals (n = 3).

DISCUSSION

In this study, we show that the S256 residue on the COOH-terminus of AQP2 can be phosphorylated by VP/Fk in various subcellular compartments, including at the plasma membrane, in the cis-Golgi/TGN, and on vesicles scattered throughout the cytosol after microtubule disruption. PKA is the main kinase involved in phosphorylation of S256 AQP2 in these various cellular locations.

It is now well established that VP induces S256 AQP2 phosphorylation and causes parallel cytoskeletal changes that facilitate an increase in exocytosis and a decrease of endocytosis, resulting in AQP2 membrane accumulation (23, 27, 33, 48). Even today, new reports often incorrectly state that AQP2 phosphorylation “results in” vesicle exocytosis that increases water permeability, even though the phosphorylation state of AQP2 has little, if anything, to do with the actual membrane insertion (exocytosis) of AQP2 (5, 39). Our present data support this concept by showing that that AQP2 can be phosphorylated in several locations within epithelial cells, including when it is already present on the plasma membrane.

The first clear indication that AQP2 membrane exocytosis does not depend on S256 phosphorylation was our finding that MBCD causes AQP2 membrane accumulation of nonphosphorylated mutant S256A AQP2 in cultured epithelial cells (5). In the present study, we show by immunostaining that although MBCD causes AQP2 membrane accumulation, most of this AQP2 is not recognized by the S256 phospho-specific antibody. MBCD does not, therefore, cause an increase of AQP2 S256 phosphorylation despite the strong accumulation of AQP2 in the plasma membrane. Importantly, subsequent incubation of the MBCD pretreated cells with VP/Fk resulted in a significant increase in AQP2 S256 phosphorylation, much of which was already located at the cell surface in the plasma membrane. Thus, this provides evidence for a dissociation of S256 AQP2 phosphorylation and its cellular location. The location of AQP2 S269 phosphorylation seems to be somewhat more specific. Wong et al. (49), using knockdown of individual Rab proteins, showed that S269 is dephosphorylated during Rab5-mediated AQP2 endocytosis before AQP2 joins the recycling endosome and S269 can be phosphorylated in the recycling endosomes before apical membrane trafficking.

We then investigated the importance of the microtubule network and S256 phosphorylation of AQP2. Microtubules play important roles in intracellular trafficking and assist in the cellular compartmentalization of AQP2 (24, 27, 30). After endocytosis, AQP2 appears to reside in an apical perinuclear region that is initially distinct from the Golgi and TGN (28, 50, 51). Several microtubule-associated proteins have been detected in AQP2 vesicles (52, 53), and both dynein and myosin Vb motors interact with AQP2 facilitating its migration along the microtubules toward the intracellular compartment or the plasma membrane (45, 54). Similar to previous studies, we showed that AQP2 is scattered throughout the cytoplasm in both cells and tissues when microtubules are disrupted (24, 29, 30, 55). In addition, there is a significant reduction of AQP2 membrane accumulation upon VP treatment in microtubule-disrupted cells. Microtubule disrupters also reduce the effect of VP on water permeability in kidney tubules (56–58), and there is an increase in diuresis in colchicine-treated rats (59). Earlier studies also showed a decrease in VP-stimulated epithelial water permeability in the toad urinary bladder experimental model (60–63). In addition to a microtubule disrupter, we also used taxol, which is a microtubule stabilizer that prevents dynamic microtubule assembly. We found that it produced a similar phenotype to colchicine after immunocytochemistry, with AQP2 located on vesicles scattered in the cytoplasm. Neither colchicine nor taxol affected VP/Fk stimulation of AQP2 S256 phosphorylation. These data suggest that AQP2 phosphorylation occurs independently from changes in microtubule organization, but that membrane accumulation of AQP2 requires an intact cytoskeleton and dynamic microtubule assembly.

Under basal conditions, pS256 is distributed in several intracellular compartments, and we found that there is an increase of pS256 at the plasma membrane with VP treatment. Using high-resolution Airyscan and three-dimensional semiquantification of pS256 and various intracellular markers, we found a shift of pS256 to the GM130-positive cis-Golgi/TGN compartment with VP treatment but not to early endosomes positive for EEA-1, clathrin-positive vesicles, or Rab11 vesicles. This may indicate a role for phosphorylation, or the presence of phosphorylation machinery, in vesicles that are located in the cis-Golgi compartment, phosphorylating either newly synthesized AQP2 or AQP2 that recycles back to the cis-Golgi but not in the later post-Golgi compartments in the exocytotic pathway. Indeed, a role for AQP2 phosphorylation, as well as glycosylation, in rough endoplasmic reticulum-Golgi trafficking and processing was suggested previously (64, 65).

Perspectives and Significance

Our study determined whether AQP2 phosphorylation depends on its intracellular location and on a functional microtubular cytoskeleton. The data show that AQP2 can be phosphorylated by PKA at S256 while it resides in multiple intracellular compartments along the recycling itinerary. These results indicate that once internalized from the cell surface, subsequent S256 phosphorylation of AQP2 is unlikely to be involved in its recycling back to the plasma membrane. This new information increases our understanding of how phosphorylation contributes to cell surface accumulation and retrieval of AQP2 and will inform the search for novel therapeutic strategies to alleviate conditions of defective water homeostasis such as nephrogenic diabetes insipidus and syndrome of inappropriate antidiuretic hormone secretion.

DATA AVAILABILITY

Data will be made available upon reasonable request.

SUPPLEMENTAL DATA

Supplemental Fig. S1: https://doi.org/10.6084/m9.figshare.21202229.v1.

Supplemental Fig. S2: https://doi.org/10.6084/m9.figshare.21202232.v2.

Supplemental Fig. S3: https://doi.org/10.6084/m9.figshare.21202235.v2.

GRANTS

This work was supported by National Institutes of Health (NIH) Grant DK096586 (to D.B.). P.W.C. was supported by NIH Grant DK115901 and a generous philanthropic gift from Donald Glazer. C-C.S.L. was supported by NIH Training Grant 5T32DK007540A. The Nikon A1R and Zeiss LSM 800 with Airyscan confocal in the Program in Membrane Biology Microscopy Core were purchased using NIH Shared Instrumentation Grants 1S10RR031563-01 and 1S10OD021577-01 (to D.B.), respectively. Additional support for the Program in Membrane Biology Microscopy Core came from the Boston Area Diabetes and Endocrinology Research Center (NIH Grant DK057521) and the Massachusetts General Hospital Center for the Study of Inflammatory Bowel Disease (NIH Grant DK043351).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

P.W.C., M.B., S.M., A.K., T.C.P., C-C.S.L., A.V.N., R.B., and D.B. conceived and designed research; P.W.C., M.B., R.S.E.B., S.M., A.K., T.C.P., N.B., C-C.S.L., R.B., and D.B. performed experiments; P.W.C., M.B., S.M., A.K., T.C.P., N.B., C-C.S.L., A.V.N., R.B., and D.B. analyzed data; P.W.C., M.B., S.M., A.K., T.C.P., C-C.S.L., A.V.N., R.B., and D.B. interpreted results of experiments; P.W.C., M.B., R.S.E.B., S.M., A.K., T.C.P., N.B., C-C.S.L., R.B., and D.B. prepared figures; P.W.C., M.B., R.S.E.B. A.K., and R.B. drafted manuscript; P.W.C., M.B., A.K., T.C.P., C-C.S.L., A.V.N., R.B., and D.B. edited and revised manuscript; P.W.C., M.B., S.M., A.K., T.C.P., C-C.S.L., A.V.N., R.B., and D.B. approved final version of manuscript.