Lipid Rafts Are Required for Effective Renal D₁ Dopamine Receptor Function

Abstract

Effective receptor signaling is anchored on the preferential localization of the receptor in lipid rafts, which are plasma membrane platforms replete with cholesterol and sphingolipids. We hypothesized that the dopamine D₁ receptor (D₁R) contains structural features that allow it to reside in lipid rafts for its activity. Mutation of C347 palmitoylation site and Y218 of a newly identified Cholesterol Recognition Amino Acid Consensus motif resulted in the exclusion of D₁R from lipid rafts, blunted cAMP response, impaired sodium transport, and increased oxidative stress in renal proximal tubule cells. Kidney-restricted silencing of Drd1 in C57BL/6J mice increased blood pressure (BP) that was normalized by renal tubule-restricted rescue with D₁R-wild-type but not the mutant D₁R-347A that lacks a palmitoylation site. Kidney-restricted disruption of lipid rafts by β-MCD jettisoned the D₁R from the brush border, decreased sodium excretion, and increased oxidative stress and BP in C57BL/6J mice. Deletion of the PX domain of the novel D₁R-binding partner sorting nexin 19 (SNX19) resulted in D₁R partitioning solely to non-raft domains, while silencing of SNX19 impaired D₁R function in renal proximal tubule cells. Kidney-restricted silencing of Snx19 resulted in hypertension in C57BL/6J mice. Our results highlight the essential role of lipid rafts for effective D₁R signaling.

Introduction

Dopamine produced by renal proximal tubules, independent of renal nerves, engenders natriuresis through its cognate receptors by inhibiting the sodium pump and sodium transporters/exchangers (1–5). The dopamine receptors are classified as D₁-like (D₁R and D₅R) or D₂-like (D₃R, D₄R, and D₅R), depending on their coupling with stimulatory or inhibitory G proteins, Gα_s or Gα_i, respectively. G protein-coupled receptor kinases (GRKs), especially GRK4, in the case of D₁R, are activated when dopamine receptors are stimulated leading to receptor phosphorylation, internalization, and desensitization (1–5). Dopamine accounts for more than 60% of renal Na⁺ excretion during moderate volume expansion (3,4). Germline deletion of the Drd1 gene in mice (Drd1^−/−) results in decreased sodium excretion and increased blood pressure (6).

The D₁R is expressed in both apical and basolateral membranes of the renal proximal tubule (RPT), especially in lipid raft microdomains and caveolae (7,8). Caveolae, a subset of lipid rafts, are small (60–80 nm) invaginations formed by the polymerization of caveolins (CAVs) with cholesterol (9). Lipid rafts are tightly packed, highly organized plasma membrane microdomains that are enriched in phospholipids, glycosphingolipids, and cholesterol (10,11). These abound in the apical and basolateral plasma membranes of polarized epithelial cells, such as renal proximal tubule cells. Lipid rafts form a semi-rigid structure composed of different proteins for effective GPCR signaling. Other functions of lipid rafts include the formation of platforms for extracellular matrix adhesion and intracellular cytoskeletal tethering to the plasma membrane. Caveolae are localized at the inner leaflet of the plasma membrane (9). Caveolae are found mostly at the basolateral membrane of renal epithelial cells that faces the blood supply; the basolateral membrane is most active during signal transduction (10–14). The renal D₁R partitions to the lipid raft/caveolae in human RPT cells (hRPTCs) and interact with the lipid raft protein marker CAV-1 (8,15) and CAV-2 (16), which are crucial in regulating D₁R turnover and signaling.

In this study, we tested the hypothesis that the renal D₁R must reside in specific microdomains, i.e., in lipid rafts, for its full activity. We demonstrate that the D₁R has critical structural determinants, such as the palmitoylation sites and the CRAC domain (17), which are essential for its targeting to and activity in lipid rafts. Kidney-restricted silencing of Drd1 increases blood pressure in C57BL/6J mice that is normalized by renal tubule-restricted rescue with D₁R-wild-type but not the mutant D₁R-347A that lacks a D₁R palmitoylation site. Kidney-restricted disruption of lipid rafts by β-MCD jettisoned the D₁R from the brush border, decreased sodium excretion, and increased oxidative stress and blood pressure in C57BL/6J mice. The adaptor protein sorting nexin 19 (SNX19) is important in the palmitoylation, lipid raft microdomain localization, and function of renal D₁R. Silencing of SNX19 inhibits D₁-like receptor-mediated inhibition of renal sodium transport and stimulation of cAMP production in human renal proximal tubule but not distal convoluted tubule cells. Kidney-restricted, siRNA-mediated, silencing of Snx19 results in oxidative stress and elevated blood pressure in C57BL/6J mice.

Materials and Methods

Plasmid DNA Constructs and AAV Vectors

Human wild-type D₁R-WT (Accession: NM_000794, Clone ID: OHu23409C), ΔhD₁R-C347A, ΔhD₁R-C351A, and ΔhD₁R-Y218S and wild-type SNX19-WT (Accession: NM_014758) and ΔhSNX19-(-PX) were purchased from GenScript (Piscataway, NJ, USA). An upstream KpnI restriction enzyme digestion site and a downstream BamHI site were introduced into the respective genes by PCR. The PCR fragments were subcloned into the pcDNA3.1 vector with N-Myc (D₁R) or FLAG (SNX19) tag and sequenced which showed 100% fidelity. E. coli TOP10 strain was used for the DNA subcloning and expression of the plasmids.

The AAV vectors harboring hD₁R-WT (AAV-D₁R-WT) or the mutant ΔhD₁R-C347A (AAV-ΔhD₁R-C347A), with the expression controlled by a CMV promoter were constructed using the plasmid pAV-FH (Vigene Bioscience Inc., Rockville, MD, USA) as previously described (16,18,19).

Cell Culture and Transfection

Immortalized hRPTCs were generated from the normal portions of renal surgical specimens from normotensive Caucasian male patients who had renal cell carcinoma (8,15,20–22). Human embryonic kidney (HEK) 293 cells (ATCC, Manassas, VA, USA) and hRPTCs were maintained at 37°C in 95% air/5% CO₂ in DMEM/F-12 medium (Gibco, Grand Island, NY, USA) supplemented with 5% penicillin-streptomycin-amphotericin (Corning Life Sciences, Manassas, VA) and 10% FBS (Gibco, Grand Island, NY, USA). When 90% confluent, the cells were subcultured for use in experimental protocols using trypsin-EDTA (ThermoFisher; Rockford, IL, USA). The cells tested negative for Mycoplasma infection. Utilization of cells was limited to <25 passages to avoid the effects of cellular senescence. The cells were transfected with the plasmids using FuGENE® 6 (Cat# E2691; Promega, Madison, WI, USA) with a 3:1 reagent to plasmid DNA ratio. Stable transfection was done through treatment with 500 μg/μl of G418 (Invivogen, San Diego, CA, USA).

Cell Surface Biotinylation and Isolation

hRPTCs stably transfected with hD₁R-WT, ΔhD₁R-C347A, ΔhD₁R-C351A, ΔhD₁R-Y218S, and ΔhD₁R-LL were cultured in 10-cm dishes. The cells were washed with ice-cold PBS then incubated with 2 ml (500 μg/ml PBS) EZ-Link™ Sulfo-NHS-LC-LC-Biotin (Thermo Fisher Scientific, Rockford, IL, USA) at room temperature for 30 min. The cells were washed 3x with PBS and then 100 mM glycine. The cells were lysed with NP-40 Lysis Buffer (Amresco, Solon, OH, USA), containing Halt™ Protease and Phosphatase Inhibitor Single-Use Cocktail 1X (Thermo Fisher Scientific, Rockford, IL, USA). The protein concentration was measured using Pierce BCA Protein Assay Kit (Thermo Fisher Scientific, Rockford, IL, USA) to ensure uniform protein concentrations prior to immobilization with streptavidin. Washed Dynabeads™ MyOne™ Streptavidin T1 (Thermo Fisher Scientific, Rockford, IL, USA) magnetic beads were incubated with the biotinylated proteins in PBS for 30 min at room temperature, using gentle rotation. The beads were separated with a magnet for 2–3 min then washed 4–5x with PBS containing 0.1% BSA. The proteins were eluted with the Lane Marker Reducing Sample Buffer (5X) (Thermo Fisher Scientific, Rockford, IL, USA). The samples were heated at 60°C for 15 min, subjected to SDS-PAGE, and then immunoblotted with anti-N-Myc and anti-GGT antibody.

Lipid Raft Isolation and Sucrose Gradient Ultracentrifugation

The cells grown to confluence in 15-cm dishes were washed with PBS, scraped, pelleted, and centrifuged at ~800 x g for 1 min at 4°C. After discarding the supernatant, SPE Buffer I from the FOCUS™ SubCell kit (G-Biosciences, St Louis, MO, USA) was added to the pellet. The samples were vortexed and incubated on ice for 10 min. Then, the cells were lysed using a Dounce homogenizer with 20 strokes per sample. To pellet the nuclei, the samples were centrifuged at 700 x g for 10 min at 4°C. The supernatant was transferred to a new tube and subsequently centrifuged at 12,000 x g for 15 min at 4°C to pellet the mitochondria. The remaining supernatant was transferred to a new tube and subsequently centrifuged at 100,000 x g for 60 min at 4°C in a SW50.1 swinging bucket rotor (Beckman Coulter™, Palo Alto, CA, USA), to pellet the membrane fraction. The plasma membrane-enriched pellet was subjected to detergent-free sucrose gradient ultracentrifugation (16,23–26). Twelve fractions were collected and immunoblotted for N-Myc-tagged D₁R and CAV-1.

Western Blotting

Cell lysates were prepared on ice using Laemmli sample buffer (Bio-Rad; Hercules, CA) with a cocktail of Halt™ Protease and Phosphatase Inhibitors (Thermo Fisher Scientific, Rockford, IL, USA). The samples were heated at 60°C for 15 min. Proteins were resolved in 10% Criterion™ TGX™ Precast Midi Protein Gel (Bio-Rad; Hercules, CA, USA). The primary antibodies against N-Myc (Proteintech™, Rosemont, IL, USA), CAV-1 (Abcam, Cambridge, MA, USA), gamma-glutamyl transferase (GGT) (Genetex; Irvine, CA, USA), GODZ (Santa Cruz Biotechnology, Dallas, TX, USA), SERZ-β (Bioss Abs, Woburn, MA, USA), PPT-1 (Novus Biologicals, Centennial, CO), PTT-2 (Thermo Fisher Scientific, Rockford, IL, USA), GRK4 (Santa Cruz Biotechnology, Dallas, TX, USA), and SNX19 (ProteinTech, Rosemont, IL, USA) were prepared in an antibody diluent (Thermo Fisher Scientific, Rockford, IL, USA). The secondary antibodies, donkey anti-goat IgG-HRP (Santa Cruz Biotechnology, Dallas, TX, USA) and donkey anti-rabbit IgG-HRP (Santa Cruz Biotechnology, Dallas, TX, USA), were diluted with 5% non-fat dried milk. Hyperfilm ECL (GE Healthcare, Buckinghmanshire, UK) was used for western blotting detection. The bands were analyzed by ImageJ.

Co-Immunoprecipitation

Cell lysates were prepared on ice, using NP-40 lysis buffer (Thermo Fisher Scientific, Rockford, IL, USA) with a cocktail of Halt™ Protease and Phosphatase Inhibitors (Thermo Fisher Scientific, Rockford, IL, USA). Co-immunoprecipitation was carried out using DynaBeads Protein G (Thermo Fisher Scientific, Rockford, IL, USA), following the manufacturer’s protocol. Normal rabbit or mouse IgG (Santa Cruz Biotechnology, Dallas, TX, USA) was used as negative control.

cAMP Measurement

The cells were cultured in 6-well plates in complete culture medium until they reached 90% confluence. The cells were serum-starved and pre-treated with the phosphodiesterase inhibitor IBMX (Millipore Aldrich, St. Louis, MO, USA) for 2 hr prior to treatment with fenoldopam monohydrobromide, 1 μM for 15 min at 37°C (Sigma Aldrich, St. Louis, MO, USA). Intracellular cAMP was measured using the Cyclic AMP Direct EIA Kit (Arbor Assays, Ann Arbor, MI, USA). The protein concentration of the samples was measured utilizing Pierce BCA Protein Assay Kit (Thermo Fisher Scientific, Rockford, IL, USA) to normalize cAMP levels. Data are expressed as mean ± standard error of the mean (pmol/mg protein/min) which was subsequently converted to % change of control.

Sodium Transport Studies

The cells were grown to confluence under polarized conditions on Polyester (PET) Membrane Transwell-Clear Inserts (Corning, Lowell, MA, USA) in 12-well plates. Transepithelial electrical resistance (TEER) using the Epithelial Voltohmmeter (EVOM) was used as a parameter to monitor the confluency of the cells. The cells were serum-starved for 2 hr and treated with vehicle, ouabain octahydrate 50 μM (Sigma Aldrich, St. Louis, MO, USA), fenoldopam monohydrobromide 1 μM, or the combination of fenoldopam and ouabain for 1 hr at the basolateral side. After washing the cells with PBS, 5 μl of DMSO and pluronic acid were mixed with Sodium Green™ Tetraacetate (Molecular Probes, Eugene, OR, USA). The cells were washed gently with PBS and the fluorescence emission was read in Victor II (excitation 485 nm and emission 535 nm). Protein concentration of the samples was also determined utilizing Pierce BCA Protein Assay Kit (Thermo Fisher Scientific, Rockford, IL, USA) to normalize the sodium levels. Results were reported as fluorescence units (reading from the Victor II)/mg protein and from these numbers, % of inhibition/activation was calculated.

Laser Scanning Confocal Microscopy

The cells were grown on poly-D-lysine-coated coverslips to 50% confluence. These were fixed for 20 min with 4% paraformaldehyde, permeabilized for 5 min with 0.05% Triton X-100, and immunostained using N-Myc and CAV-1 for 1 hr at room temperature. The cells were then treated with the appropriate fluorophore-conjugated secondary antibodies (Molecular Probes, Grand Island, NY, USA). To study the effect of inhibitors of palmitoylation on D₁R targeting to lipid rafts, the hRPTCs were pre-treated with 2-fluoropalmitic acid (Santa Cruz Biotechnology, Dallas, Texas, USA), 2-bromopalmitate (Millipore Aldrich, Louis, MO, USA), or cerulenin (Millipore Aldrich, Louis, MO, USA) at 37°C for 1 hr.

Immunohistochemistry of the mouse kidney samples was performed following standard protocol. The slides were deparaffinized in xylene at 60°C for 1 hr and then washed twice for 3 min in 100%, 95%, and 75% ethanol. The slides were boiled in sodium citrate buffer (pH 6.0) for 3 min under pressure for antigen retrieval, cooled to room temperature, and then blocked in 1% BSA for 30 min at room temperature. The sections were immunostained for Cholera toxin subunit B (CTxB), SNX19, GRK4, and D₁R, for 24 hr at 4°C. Lotus tetragonolobus lectin (LTL) conjugated with Alexa fluor® 647 was used to target the lectin-rich brush border and plasma membranes of the renal proximal tubules. DAPI was used to visualize the nuclei.

Confocal images of hRPTCs and kidney sections were obtained sequentially in separate channels to avoid bleed-through using a Carl Zeiss LSM 510 META with ×63/1.4 NA oil-immersion objective (Carl Zeiss, Thornwood, NY, USA) and processed using Zeiss 510 META with Physiology 3.5 and Multiple Time Series 3.5 software.

Intracellular ROS Detection

Intracellular ROS was quantified using the ROSstar™ 550 (Li-cor, Lincoln, NE, USA), which is a cell-permeable hydrocyanine probe that is initially non-fluorescent but becomes fluorescent after oxidation by ROS. It is specific for oxygen radicals, in particular for superoxide and hydroxyl radicals. Cultured RPTCs were treated with 50 μM of ROSstar 550. After incubation for 30 min at 37°C, the cells were washed 2X times with PBS and the fluorescence signal was quantified immediately by a plate reader at 540 nm excitation and 560 nm emission. The data were normalized by the protein concentration of each well.

Animal Care

Adult (8 wk-old) male C57BL/6J mice were obtained from Jackson Laboratory (Bar Harbor, ME, USA). The studies were conducted in accordance with U.S. National Institutes of Health guidelines for the ethical treatment and handling of animals in research and approved by the George Washington University Institutional Animal Care and Use Committee (IACUC). The mice were housed in a temperature-controlled facility with a 12:12-hr light-dark cycle and fed with mouse chow and water ad libitum for at least 2 wk before any studies were performed.

In Vivo β-MCD Infusion

Uninephrectomized adult male C57BL/6J mice on normal salt diet were given a 7-day renal subcapsular infusion of β-MCD (80 μg/kg/min) through an osmotic minipump. α-MCD or vehicle (PBS) was also infused in other mice as controls. On the day of the experiment, the mice were anesthetized with pentobarbital (50 mg/kg body weight, intraperitoneal) before cannulation of the femoral artery for blood pressure monitoring. A PE-90 catheter was also inserted through a cystotomy and exteriorized to collect urine for one hour. The remnant kidney was then perfused with 4% paraformaldehyde and collected for immunohistochemistry.

Telemetry

The mice were anesthetized with pentobarbital sodium (50 mg/kg, intraperitoneal). A transmitter (cat# TA11PA-C10, Data Sciences International) for mice was turned on magnetically 24 hr before and soaked in warm saline solution 10 min prior to the surgery (27). The catheter end was inserted into the temporarily occluded left carotid artery and the main body was secured in a skin pocket made by blunt dissection of the right flank of the animal. The incision was closed using 4–0 Ethilon®. Buprenorphine (0.05–0.10 mg/kg, intraperitoneal) was given to alleviate post-surgical pain. The mice were then closely observed and kept warm until ambulatory. The animals were then returned to the Animal Research Facility (ARF) telemetry room. Data were continuously collected and recorded through a dedicated computer running Dataquest. The mice were allowed to recover for 5–7 days to equilibrate the system in order to obtain accurate cardiovascular phenotyping. The mice were maintained on normal salt diet and were provided with unlimited access to drinking water.

Renal DRD1 and SNX19 Gene Silencing and Renal DRD1 Rescue

Uninephrectomized adult male C57BL/6J mice were anesthetized with pentobarbital (50 mg/kg, intraperitoneal) and placed on a heated board to maintain rectal temperature at 37°C. A polyethylene tube connected to an osmotic minipump (ALZET® Cupertino, CA, USA) was inserted underneath the renal capsule to deliver continuously the Drd1-specific siRNA or Snx19-specific siRNA (3 μg/day; Qiagen, Germantown, MD, USA) for 7 days. The siRNA was prepared in an in vivo transfection reagent (Mirus Bio, Madison, WI, USA) under sterile conditions. The tube was secured on top of the renal capsule using a drop of surgical glue, while the minipump was secured onto the abdominal wall by a suture. Infusion of vehicle or non-silencing mock siRNA (Qiagen, Germantown, MD, USA) served as controls.

On the 7^th day of siRNA infusion, blood pressure was measured under pentobarbital anesthesia using Cardiomax II to confirm the presence of elevated blood pressure. D₁R-WT and ΔD₁R-347A packaged in AAV9 vectors were infused into the renal tubules via the retrograde ureteral route, as described previously (18,19). Briefly, the distal portion of the ureter closest to the bladder and the renal artery supplying the target kidney were clamped off with micro-venous clips in pentobarbital-anesthetized mice. The ureter was then exposed and punctured with a 35-gauge needle to aspirate the urine. Using another syringe, 100 μL of the AAV vector (10¹¹ viral genome particles) was injected retrogradely into the renal tubules via the ureter. The needle was withdrawn and the injection site on the ureter was clipped to prevent leakage. The arterial and the ureteral clips were kept in place for 30 min for maximum exposure to the infusate. The arterial and ureteral clips were then removed, and the abdominal contents were replaced in the proper order. The incision site was closed using a double layer of 6–0 ethilon absorbable sutures for the muscle and 6–0 silk sutures for the skin. The animals were allowed to recover and then returned to the ARF. On the 14^th day post-rescue, the mice were placed in metabolic cages for 24 hr to monitor for water and food consumption, as well as urine and fecal production. Urine Na⁺ was measured using Easylyte® Analyser (Medica Corporation, Bedford, MA, USA). UNaV was calculated as urine volume × Na⁺ (mEq/liter). The next day, the animals were anesthetized, and blood pressure was measured using Cardiomax II. The mice were sacrificed, and the blood, organs, and tissue samples were collected and stored accordingly for further experimentation.

Statistics

Numerical data are expressed as mean ± standard error of the mean. Significant difference between 2 groups was determined by Student’s t-test, while that among ≥3 groups was determined by one-way ANOVA followed by Holm-Sidak post-hoc test. Values of P<0.05 were considered significant. Statistical analysis was performed using SigmaStat 3.5 (Richmond, CA, USA).

Results

Structural Determinants of D₁R

Numerous post-translational modifications of proteins, such as palmitoylation of amino acid residues, serve as membrane anchors that promote lipid raft partitioning of several GPCRs (28). Palmitoylation is a process that involves the formation of thioesters between palmitate and cysteine residues, catalyzed by a family of palmitoyl acyltransferases to enhance the hydrophobicity of GPCRs (29). The C-terminus of D₁R possesses cysteine residues at positions C347 and C351, but not those at C297, C306, and C385, that undergo palmitoylation (30) (Figure 1).

Figure 1. Partial primary sequence (171–387 a.a. residues) of the human D₁R.

The palmitoylation sites, Cholesterol Recognition Amino Acid Consensus (CRAC) domain, and boundaries of the transmembrane domains (TM, underlined), extracellular (EC) and intracellular (IC) domains, and the C-terminal tail are indicated.

We used the bioinformatics software CLC Main Workbench 7.7.2 (Qiagen) to identify a putative Cholesterol Recognition Amino Acid Consensus or CRAC motif, which is an ubiquitous and loosely defined sequence ([LV]-x(1,5)-Y-x(1,5)-[KR]) that has been shown to interact with cholesterol (17) and implicated in lipid raft partitioning (31). CRAC motifs are present in GPCRs, such as the rhodopsin receptor, β₂-adrenergic receptor, and serotonin_1A receptor. The central tyrosine residue of the CRAC motif is critical for the cholesterol-binding property of the peripheral benzodiazepine and sigma-1 receptors. Substitution of tyrosine to serine impairs the receptors^’ ability to bind to cholesterol and partition to lipid rafts (17,32).

To determine if the D₁R palmitoylation sites and CRAC domain are critical for receptor function, we initially used PolyPhen-2 (http://genetics.bwh.harvard.edu/pph2/) to perform ab initio prediction of the possible effects of an amino acid substitution on the structure and function of a human protein, which predicted the 347A, 351A, and 218S mutations (CRAC motif Y>S) to be “probably damaging” (Figure 1).

D₁R Palmitoylation and SNX19

Palmitoylation is carried out at the Golgi by a DHHC (Asp-His-His-Cys) family of palmitoyl acyltransferases, such as the Golgi-specific DHHC zinc finger protein (GODZ) and, to a lesser extent, its paralog, the Sertoli cell gene with a zinc finger domain-β (SERZ-β) (33). We found that GODZ (55 kDa), but not SERZ-β, co-immunoprecipitated with the D₁R and GRK4 (Figure 2A). A higher band, conceivably representing the less enzymatically active GODZ dimers (34), co-immunoprecipitated with D₁R, GRK4, and sorting nexin 19 (SNX19). We also found that palmitoyl-protein thioesterase 1 (PPT-1), but not PPT-2, co-immunoprecipitated with the D₁R and GRK4 and may possibly be responsible for the removal of the palmitate from D₁R and GRK4 at the lysosomes (Figure 2A). GRK4 undergoes palmitoylation at C561 and C578 (35), which are important for its plasma membrane localization (36). SNX19 is a member of the sorting nexin family which plays a pivotal role in the molecular sorting of GPCRs, including the SNX5 for D₁R (37) and SNX1 for D₅R (38). We evaluated the subcellular distribution and colocalization of D₁R, GODZ, and SNX19 in hRPTCs via confocal microcopy and observed that in the basal state, they all colocalized at the perinuclear area, where the Golgi is located (Figure 2B). These results suggest the presence of an organized complex of enzyme (GODZ) and targeted substrates (D₁R and GRK4) that are conceivably held in place by a scaffold or adaptor protein (SNX19).

Figure 2. Interaction of endogenous D₁R, SNX19, GRK4, GODZ, and caveolin-1 (CAV-1).

A. Co-immunoprecipitation (IP) of D₁R, SNX19, and GRK4 with GODZ and palmitoyl-protein thioesterase 1 (PPT-1) and PPT-2 at the basal state. Normal rabbit IgG was used as negative control and regular immunoblot (IB) of the cell lysate was used as positive control. Results of 1 of 3 experiments are shown. B: Basal distribution D₁R (pseudocolored green), SNX19 (pseudocolored magenta), and GODZ (pseudocolored red) at the perinuclear area in serum-starved hRPTCs. Colocalization of all three is denoted by the discrete white areas. DAPI was used to visualize the nucleus (pseudocolored blue). 630x magnification, scale bar = 10 μm. C: Serum-starved hRPTCs were pre-treated with palmitoylation inhibitors 2-fluoropalmitate (2-FP), 2-bromopalmitate (2-BP), or cerulenin for 2 hr, or vehicle as control. Basal colocalization of D₁R (pseudocolored green) and CAV-1 (pseudocolored red) is indicated by the discrete yellow areas in the D₁R + CAV-1 image of vehicle-treated cells. Wheat germ agglutinin and DAPI were used to visualize the plasma membrane and nucleus (pseudocolored magenta and blue), respectively. 630x magnification, scale bar = 20 μm. Representative images are shown from 1 of 3 experiments. About 20–30 cells were scored per treatment.

To determine if palmitoylation is crucial for the lipid raft targeting of the D₁R, we evaluated the colocalization of D₁R and the lipid raft marker CAV-1 in the presence or absence of pharmacological inhibitors of palmitoylation (Figure 2C). In the basal state, the D₁R was distributed in the cytoplasm and plasma membrane, where it colocalized with CAV-1. However, pre-treatment with the inhibitors of palmitoylation, i.e., 2-fluoropalmitate, 2-bromopalmitate (39), and cerulenin (40), limited the plasma membrane distribution of D₁R and CAV-1 and promoted their accumulation at the juxtanuclear area. The C-terminus of CAV-1 contains palmitoylation sites at C133, C144, and C156 (41) that are equally susceptible to the effects of the palmitoylation inhibitors.

D₁R mutants mistarget to the non-raft domains

Site-directed mutagenesis at ΔD₁R-347S did not interfere with the plasma membrane targeting of the receptor, as previously reported (42). We confirmed this observation by heterologously expressing the D₁R-wild-type (WT) and mutants (ΔD₁R-347A, ΔD₁R-351A, and ΔD₁R-218S) in hRPTCs and evaluating their distribution at the plasma membrane. The D₁R-WT and the D₁R mutants were found at the plasma membrane-enriched fraction, denoted by the plasma membrane marker gamma-glutamyl transferase (GGT) (Figure 3A). These observations were corroborated via confocal microscopy which showed the distribution and colocalization of D₁R-WT and the above mutants with the plasma membrane (Figure 3B). By contrast, the ΔD₁R-LL mutant remained mostly cytoplasmic in distribution, as previously reported (33).

Figure 3. D₁R mutants are targeted to the plasma membrane but not the lipid raft.

hRPTCs were transfected with the N-Myc-tagged D₁R-WT or mutants (347A, 351A, 218S, and LL). The ΔD₁R-LL is a dileucine mutant that does not target the plasma membrane (42) and was used as an additional control to non-transfected hRPTCs. A: Total cell lysates were collected, and plasma membrane proteins were extracted via biotinylation and isolation. The heterologously expressed D₁R was immunoblotted, using anti-Myc, and γ-glutamyl-transferase (GGT) as loading control. B: hRPTCs were grown on cover slips, transfected with the N-Myc-tagged D₁R-WT or mutants, and grown to 50% confluence prior to fixation, double immunostaining, mounting, and confocal microscopy. An anti-Myc antibody was used to visualize the N-Myc-tagged D₁R (pseudocolored green). Wheat germ agglutinin (WGA) was used to visualize the plasma membrane (pseudocolored red), and DAPI to visualize the nuclei (pseudocolored blue). Basal colocalization of the D₁R with the plasma membrane is denoted by the discrete yellow areas (yellow arrows) in the merged images. 630x magnification, scale bar = 10 μm

We next evaluated the ability of D₁R-WT and mutants to segregate into the lipid raft microdomains. The receptors were heterologously expressed in hRPTCs and subjected to sucrose gradient ultracentrifugation to separate the more buoyant lipid rafts (fractions 1–6) from the less buoyant non-raft domains (fractions 7–12). The heterologously expressed D₁R-WT was distributed in both lipid raft and non-raft microdomains and to a greater extent in the latter than in the former region (Figure 4). By contrast, the D₁R mutants exclusively segregated to the non-raft microdomains.

Figure 4. Lipid raft distribution of D₁R-WT and mutants.

A: Plasma membrane-enriched fractions were prepared and subjected to detergent-free sucrose gradient ultracentrifugation and the 12 fractions were immunoblotted for N-Myc which was used to tag the D₁Rs. Caveolin-1 (CAV-1) was used as the lipid raft marker. Immunoblots from 1 of 3 experiments is shown.

D₁R mutants are dysfunctional

To determine the functional repercussions of the D₁R mutants which do not reside in lipid rafts, intracellular cAMP levels were measured in non-transfected and transfected cells that were treated with the D₁R/D₅R agonist fenoldopam or vehicle as control. There are no agonists that can distinguish the D₁R from the D₅R; D₁R and D₅R interact for a full response of either receptor, especially the D₁R-mediated stimulation of cAMP production (20). Agonist stimulation resulted in the expected increase in cAMP production in non-transfected (NT) and D₁R-WT cells, but not in cells expressing the D₁R mutants (Figure 5A). The overexpression of the mutant receptors somehow over-ride the response of the endogenous D₁R. The mutant receptors partition to non-raft domains leading to impaired function (cAMP production). Furthermore, this may be related, in part, to the increased production of reactive oxygen species in 218S and 347A (see below), when the D₁R function is impaired; the D₁R can inhibit the production of reactive oxygen species (3,5,25). Slightly higher basal cAMP levels were observed in cells expressing the ΔD₁R-347A and ΔD₁R-218S. Disruption of lipid rafts with methyl-β-cyclodextrin (β-MCD) increased the basal adenylyl cyclase activity but prevented the agonist-stimulated effect in hRPTCs that may be related to the presence of the adenylyl cyclase isoform targets (adenylyl cyclase 5 and 6) of D₁R in lipid rafts (23). We have suggested that lipid rafts keep adenylyl cyclase activity in a less active state in hRPTCs (24).

Figure 5. D₁R mutants have impaired activity.

A: The cells were serum-starved and pre-treated with IBMX for 2 hr before fenoldopam (1 μM) or vehicle treatment for 30 min to determine the cAMP response. The data were normalized for protein concentration. B: The hRPTCs were grown under polarized conditions on Transwells® to 100% confluence. The cells were serum-starved and pre-treated with the Na⁺-K⁺/ATPase (NKA) inhibitor ouabain for 60 min prior to 30-min fenoldopam treatment. Both treatments were given at the basolateral compartment. Intracellular sodium was measured using sodium green assay and the data were normalized for protein concentration. NT= non-transfected hRPTCs; WT=D₁R wild-type transfected hRPTCs; 347A, 351A, and 218S are hRPTCs transfected with the corresponding D₁R mutants. *P<0.05, vs. vehicle-treated D₁R-WT (Figures 5A and 5B) or NT hRPTCs (Figure 5A), one-way ANOVA and Holm-Sidak post-hoc test, n=4–5/group.

We next evaluated the effect of these receptors on sodium transport. In D₁R-WT cells, treatment with fenoldopam, which inhibits Na⁺-K⁺/ATPase (NKA) and sodium transporters/exchangers (3–6,8), or ouabain (a specific NKA inhibitor), or both at the basolateral side of polarized hRPTCs grown in Transwells, resulted in the expected increase in intracellular sodium because of inhibition of sodium egress from inside the cell to outside the cell via the basolateral membrane (Figure 5B). However, there was no increase in intracellular sodium concentration when cells expressing the mutant D₁R were stimulated with fenoldopam only, demonstrating the failure of the D₁R mutant receptors to inhibit the sodium pump. The inability of fenoldopam to inhibit sodium transport in cells expressing the mutant D₁R is probably related to its failure to produce signal transducers, such as cAMP and protein kinase A (3, 20). Ouabain alone or ouabain with fenoldopam was able to inhibit NKA in cells expressing the mutant D₁R, indicating the presence of a normally functional NKA in the cells with mutant D₁R.

Lipid raft disruption in C57BL/6J mouse kidney leads to hypertension and oxidative stress

To determine if our in vitro finding of impaired D₁R function due to mistargeting of D₁R mutants to the non-raft domains occurs in vivo, we infused β-MCD subcapsularly into the remnant kidney of uninephrectomized adult male C57BL/6J mice fed normal salt diet for 7 days. We observed an increase in the systolic blood pressure (SBP) in mice treated with β-MCD, but not in the control (vehicle- and α-MCD-treated) mice (Figure 6A). Sodium excretion (UNaV) was decreased in mice that received β-MCD, but not in control (vehicle- and α-MCD-treated) mice (Figure 6B).

Figure 6. Disruption of lipid rafts only in the kidneys of C57BL/6J mice results in hypertension and impaired sodium excretion.

Adult (8–10 wk) male C57BL/6J mice on normal salt (0.8% NaCl) diet were uninephrectomized prior to a 7-day renal subcapsular minipump infusion of the cholesterol depletor β-MCD to disrupt the lipid raft. α-MCD and vehicle were used as negative controls. A & B: The mice were housed individually in metabolic cages to collect 24-hr urine samples on the 6^th day after the start of the drug/vehicle treatment. Blood pressure was measured under pentobarbital anesthesia on the 7^th day. C: The mice were sacrificed, and the kidneys were flash frozen in isopentane, fixed with 4% paraformaldehyde, sectioned, and immunostained. The brush border (using Lotus tetragonolobus Lectin, LTL), lipid raft (using cholera toxin subunit B, CTxB, green arrows), D₁R (yellow arrows), SNX19 (yellow arrows), GRK4 (yellow arrows), and nuclei (using DAPI) were visualized via confocal microscopy. Three to 4 mice per group and 630x magnification, scale bar = 20 μm

We next evaluated the morphological changes that lipid raft disruption brought about to the renal parenchyma. While there were no gross changes in the structure of the renal proximal tubule, there was an almost complete disappearance of lipid rafts (denoted by the absence of Cholera toxin B [CTxB] staining) in the brush borders of renal proximal tubules (Figure 6C). Moreover, we also observed the redistribution of D₁R and SNX19 and, to a lesser extent, GRK4, from the brush borders to the cytoplasm, which may help explain the decreased sodium excretion and increased blood pressure in the β-MCD-treated mice.

We next evaluated the presence of oxidative stress by measuring the levels of urinary isoprostanes (43) in β-MCD-treated mice and found that the urinary levels were increased compared with control (vehicle- and α-MCD-treated) mice (Figure 7A). Disruption of lipid rafts and lipid raft clustering is associated with oxidative stress in glomerular epithelial cells and hepatocytes (44–46) but not breast cancer cells. β-MCD facilitates the hepatoxicity of ochratoxin (46). Therefore, we further explored the role of lipid raft residency of D₁R on oxidative stress in mouse RPTCs and found that β-MCD treatment doubled the production of ROS in the treated cells, indicating the presence of heightened oxidative stress when the lipid rafts were perturbed (Figures 7B and 7C). Pre-treatment of the cells with diphenyleneiodonium (DPI, an inhibitor of NADPH oxidase or NOX) (36) abrogated the increase in ROS production in mouse RPTCs with disrupted lipid rafts, indicating the involvement of NOX in the oxidative stress (Figure 7C). To determine the involvement of D₁R and SNX19 in the development of oxidative stress, we silenced the endogenous expression of these proteins in mouse and human RPTCs and observed an increase in ROS production in both cell lines (Figures 7D and 7E, respectively), indicating the translational potential of our work. To ascribe directly the observed increase in oxidative stress to the D₁R mutants, we heterologously expressed the D₁R-WT and mutants in HEK-293 cells and observed a basal increase in ROS levels in cells expressing ΔD₁R-347A and ΔD₁R-218S, but not ΔD₁R-351A or D₁R-WT (Figure 7F). HEK-293 cells were used because these cells express negligible endogenous D₁R and D₅R (16,47).

Figure 7. D₁R mutants cause oxidative stress.

A: Urine isoprostanes were measured and normalized for urine creatinine to determine the presence of oxidative stress in uninephrectomized adult male C57BL/6J mice on normal salt (0.8% NaCl) diet that received a 7-day renal-restricted lipid raft disruption via a renal subcapsular minipump infusion of the lipid raft disruptor β-MCD. α-MCD and vehicle were used as controls. (+) = 1% H₂O₂, (–) = without ROSstart. B and C: ROS production in mouse RPTCs treated with β-MCD, or α-MCD and vehicle as controls, for 2 hr. ROS was measured using ROSstart™ 650. In another set of experiments, diphenyleneiodonium (DPI; a NOX inhibitor) was used to determine if the increased ROS production was due to enhanced NOX activity. *P<0.05, vs. vehicle, α-MCD, or with DPI, one-way ANOVA and Holm-Sidak post-hoc test, n=3/group. (+, positive control, 200–250 ± 10% vs. vehicle) = 1% H₂O₂, (-, negative control) = without ROSstart. D and E: ROS production in mouse (D) and human (E) RPTCs in which DRD1 or SNX19 was silenced using siRNA. Non-silencing “Mock” siRNA was used as control. ROS was measured via ROSstart™ 650. *P<0.05, vs. Mock, one-way ANOVA and Holm-Sidak post-hoc test, n=3/group. (+, positive control, 200–250 ± 10% vs. vehicle) = 1% H₂O₂, (-, negative control) = without ROSstart. F: ROS production in HEK-293 cells transfected with human D₁R-WT or mutants for 72 hrs. ROS was measured with ROSstart™ 650. *P<0.05, vs. untransfected HEK293 (HEK), human D₁R-WT (WT), or ΔD₁R-351A (351A), one-way ANOVA and Holm-Sidak post-hoc test, n=3/group. (+, positive control, 200–250 ± 10% vs. vehicle) = 1% H₂O₂, (-, negative control) = without ROSstart

Gene rescue with D₁R-WT normalized the hypertension in mice

Moreover, we performed renal tubular-restricted gene rescue in adult male C57BL/6J mice on normal salt diet that developed high blood pressure following renal-restricted silencing of renal Drd1. The increased blood pressure in renal D₁R-depleted mice was normalized by the renal tubule-restricted rescue with D₁R-WT but not ΔD₁R-347A mutant. The increase in blood pressure in the ΔD₁R-347A, relative to D₁R-WT, was observed at nighttime (Figure 8A, bottom figure) when the mice, being nocturnal animals, were mostly active. There was no difference in the expression levels of total D₁R or endogenous D₅R in the mice that were rescued with either D₁R-WT or D₁R-347A (Figure 8B).

Figure 8. Renal tubule-restricted DRD1 gene rescue using D₁R-WT, but not 347A mutant, normalizes the hypertension in renal-restricted D₁R-depleted C57BL/6J mice.

A: A 7-day (D7) renal subcapsular infusion of Drd1 siRNA (red), but not of non-silencing “Mock” siRNA (blue), to the kidney of C57BL/6J increased the systolic blood pressure (SBP) measured under anesthesia (top figure). A renal tubule-restricted rescue using D₁R-WT but not D₁R 347A mutant, subcloned in AAV vectors normalized the blood pressure at day 21 (D21) in the mice with renal-restricted Drd1 silencing (red). Baseline, before Drd1 silencing = day 0 (D0), *P<0.05, Drd1 siRNA vs. mock or WT vs. 347A post-rescue, Student’s t-test. Blood pressure was also monitored via telemetry in conscious mice (bottom figure). *Rescue of with wild-type D₁R (Drd1 siRNA + DRD1-WT [AAV]) blunted the increase in nighttime BP induced by renal-restricted D₁R silencing (Drd1 siRNA). Shown are the values for the last week of siRNA treatment and the last week of Drd1 siRNA + CAAV (control) and Drd1 siRNA + DRD1-WT (AAV). *P<0.05, vs. mice treated with Drd1 siRNA + CAAV. B. Immunoblotting of renal D₁R, D₅R, and GAPDH. M = ΔD₁R-347A, WT = D₁R-WT.

Lipid raft disruption may or may not affect blood pressure and lipid raft disruption has effects other than that related to D₁R in the kidney (48–49). The role of CAV-1 in the regulation of blood pressure in mice has been inconsistent; germline deletion of Cav-1 in mice has been shown to be associated with low, normal, or increased blood pressure (50–52). Therefore, we hypothesized that the D₁R is directed to lipid raft microdomains where it can interact with proteins important in D₁R signal transduction and function. SNX19 is a member of the sorting nexin family which plays a pivotal role in the molecular sorting of GPCRs, including the SNX5 for D₁R (37) and SNX1 for D₅R (38). As shown in Figure 2B, SNX19 and D₁R colocalize at the perinuclear area; they also colocalize in the plasma membrane, specifically in lipid raft, that is increased by fenoldopam treatment (Figure 9A). In another study, we found that treatment with fenoldopam beyond 5 minutes causes the internalization of D₁R and SNX19 (Figure 9B). In the basal state in the mouse kidney, the D₁R and SNX19 colocalize in the plasma membrane of mouse renal proximal tubules; the intravenous infusion of a dose of fenoldopam (2 μg/kg/min/5 min) does not affect blood pressure but increases the co-localization of D₁R and SNX19 in the cytosol of mouse renal proximal tubules (Figure 9C). Stimulated emission depletion (STED) microscopy confirmed the colocalization of D₁R and SNX19 in mouse renal proximal tubules (Figure 9D).

Figure 9. D₁R-GFP and SNX19 are expressed in lipid raft and non-raft domains.

A: The colocalization of D₁R-GFP with CAV-1 (lipid raft marker, arrows) and SNX19 with CAV-1 and, to a lesser degree, to CD71 (non-raft domain marker, arrows) is increased by fenoldopam (1 μM, 1 & 3 min). Images were obtained via laser confocal microscopy. 630x magnification, bar scale = 10 μm. B: Fenoldopam (D₁-like receptor agonist,1 μM) internalized and increased the colocalization of D₁R and SNX19 at 5 min & returned to 0 min position after 60 min in RPTCs. PM=plasma membrane. C: D₁R and SNX19 basally co-localized in RPTs of untreated mouse kidney (“vehicle”). Fenoldopam (2 μg/kg/min/5 min) increased the co-localization of D₁R and SNX19 in the cytosol of mouse RPTs. WGA= wheat germ agglutinin, plasma membrane marker. DAPI=nucleus. D: Colocalization of D₁R (pseudocolored green) and SNX19 (pseudocolored magenta) at the plasma membrane of renal proximal tubule of C57BL/6J mouse on normal salt diet (0.8% NaCl, 7 days) is denoted by the white circumscribed areas. The composite image was obtained via STED microscopy

There is specificity of the D₁R and SNX19 interaction because D₁R but not D₅R as the immunoprecipitant, co-immunoprecipitated with SNX19 in mouse renal proximal tubule cells (Figure 10A). Reverse co-immunoprecipitation shows that SNX19, not SNX13 or SNX25, co-immunoprecipitated with D₁R (Figure 10B).

Figure 10. D₁R, not D₅R, co-immunoprecipitates (IP) with SNX19 (Figure A).

Reverse co-IP shows that SNX19, but not SNX13 or SNX25, co-IPs with D₁R (Figure 10B). Negative control = Co-IP with IgG from host that generated the antibody. IB=regular immunoblot.

We next determined if SNX19 is required for the lipid raft targeting of D₁R by expressing a mutant SNX19 [ΔSNX19-(-PX)] in which the PX domain has been deleted and thus, cannot target the plasma membrane. The PX domain is a phosphoinositide-binding domain that is involved in targeting a protein to the lipid rafts of the plasma membrane (53). In the presence of the SNX19-WT, the endogenous D₁R partitioned mainly to the lipid raft (Figure 11). However, the expression of ΔSNX19-(-PX) resulted in the mistargeting of decreased levels of endogenous D₁R to the non-raft fractions. Endogenous SNX19-WT is mainly found in lipid rafts. These results underscore the importance of the D₁R palmitoylation sites for lipid raft targeting that may be mediated by SNX19.

Figure 11. SNX19 mutants prevent D₁R partitioning to the lipid raft.

hRPTCs were transfected with the FLAG-tagged wild-type SNX19 (SNX19-WT) or mutant SNX19-(-PX). Similar to the D₁R studies, sucrose gradient ultracentrifugation was performed, and the 12 fractions were immunoblotted for endogenous D₁R. Endogenous SNX19 was also prepared from non-transfected hRPTCs.

To corroborate the crucial role of SNX19 on the lipid raft localization and activity of D₁R in hRPTCs, we studied the effect of the D₁-like receptor agonist, fenoldopam (1 μM/15 min) on cAMP production in hRPTCs and human distal convoluted tubule cells and on sodium transport in hRPTCs in which SNX19 is silenced using siRNA (Figure 12). Silencing SNX19 decreased the fenoldopam-mediated increase in cAMP production in hRPTCs but not in human distal convoluted tubule cells. The impairment in fenoldopam-mediated increase in cAMP production in hRPTCs was accompanied by the abrogation of the ability of fenoldopam to inhibit basolateral sodium transport.

Figure 12. Effects of SNX19 silencing on cAMP production and Na⁺ transport in human renal proximal and renal distal tubule cells.

SNX19 siRNA treatment blunted cAMP production (Figure 12A) and intracellular Na⁺ (Na⁺ green) accumulation (Figure 12B) in human RPTCs treated with the D₁-like receptor agonist fenoldopam (F; 1 μM/15 min). ^#P<0.05 vs. vehicle (V), SNX19 siRNA, *P<0.05 vs. vehicle non-silencing “Mock” or SNX19 siRNA, one-way ANOVA, Holm-Sidak post-hoc test, n=4–5/group. Silencing of SNX19 in human distal convoluted tubule cells (RDTCs) (Figure 12C) did not affect the ability of the D₁-like receptor agonist fenoldopam (Fen, 1 μM/15 min) to stimulate cAMP production. n=6/group

As further evidence for the importance of SNX19 on the homeostatic regulation of blood pressure by renal D₁R, we also performed renal-restricted silencing of Snx19, similar to that performed for D₁R (Figure 8). siRNA-mediated knockdown of renal Snx19 increased the systolic blood pressure of C57BL/6J mice on normal salt diet (0.8% NaCl) (Figure 13). This increase in blood pressure was accompanied by a decrease in the expression of renal D₁R.

Figure 13. Effects of renal-restricted Snx19 silencing on blood pressure and renal D₁R expression in C57BL/6J mice.

siRNA-mediated silencing of Snx19 in the kidney of C57BL/6J mice on normal salt diet (0.8% NaCl) increased blood pressure (Figure 13A) and decreased renal Drd1 mRNA (Figure 13B) and protein (Figure 13C) expressions. *P<0.05, vs. non-silencing “Mock” siRNA, Student’s t-test, n=3–4/group. GAPDH was used as housekeeping gene for gene expression.

Discussion

Our current study shows that the D₁R has several intrinsic features that allow it to reside in lipid rafts. The D₁R has two palmitoylation sites at C347 and C351 (30) and a single putative CRAC motif (current study), which when mutated (but not 351A) results in D₁R dysfunction, leading to blunted cAMP response, impaired sodium excretion, and increased oxidative stress. Additionally, the D₁R contains a C-terminal di-Leu (L344+L345) motif that is important for its plasma membrane trafficking (42); however, its role in lipid raft partitioning has not been explored. The D₁R also contains a “CAV binding motif” in transmembrane domain 7 (7). Decreased CAV-1 expression in hRPTCs leads to the uncoupling of D₁R from adenylyl cyclase, reduced association of NKA with adaptor protein-2, internalization of NKA (8), and reduced D₁-like receptor-mediated inhibition of Na⁺ transport (11). Renal-restricted silencing of Cav-1 expression with siRNA or renal-restricted disruption of lipid rafts with β-MCD decreases sodium excretion and produces hypertension in sodium-loaded rats (11) and mice on a normal salt diet (current study). By contrast, global Cav-1 knockout mice have either low (50) or normal (51) blood pressure. However, the low or normal blood pressure in mice with germline (global) deletion of Cav-1 has been related to heart failure and increased generation of nitric oxide from eNOS (51–55). The germline deletion of Cav-1 in mice fed a high salt diet slightly but non-significantly increased blood pressure that was aggravated by nitric oxide inhibition with N^G-nitro-L-arginine methyl ester to a greater extent in Cav-1^−/− than Cav-1^+/+ mice (54). The difference in the blood pressure between renal-restricted silencing of Cav-1 and germline deletion of Cav-1 (51,52,54,55) or vascular smooth muscle-restricted deletion (53), could be due to organ/tissue/cell specificity of the effects of lipid rafts, or genes for that matter.

Other intrinsic properties of the D₁R may favor its partitioning to the lipid rafts. Overall, the hydrophobicity score calculated via GPMAWlite (https://www.alphalyse.com/customer-support/gpmaw-lite-bioinformatics-tool/start-gpmaw-lite) for D₁R is +0.33 compared with those of the interleukin 4 receptor α (−0.33) and CD71 (−0.23), which mainly reside in non-raft domains (56,57). The more positive the value, the more hydrophobic are the component amino acids, and the more miscible the protein or transmembrane is to the hydrophobic regions of the plasma membrane. As was suspected, the hydrophobicity scores for the 7 transmembrane domains of D₁R are +2.23, +1.89, +1.61, +1.84, +1.20, +2.02, and +0.66. The transmembrane length may also favor lipid raft partitioning. The D₁R transmembrane domains are 24-amino acid residues long, which is the average length for long transmembranes that prefer the plasma raft membrane. Reducing the number from 24 to 18 amino acid residues markedly decreases the raft partition coefficient (58).

The importance of both the palmitoylation process and the presence of the CRAC motif on the dopamine receptors are underscored by our studies and those by others. Jensen et al. showed that C347, but not C351, is critical in human D₁R function and plasma membrane expression in COS-7 cells (59). However, Kong et al. showed that the palmitoylation of D₁R is not essential for the localization of D₁R in caveolae (60). There are three major differences in the former and our studies. First, Kong et al. used double-mutant D₁R (347A+351A), while we used single-mutant D₁R (347A or 351A). Second, their experiments were performed in SV40-transformed COS-7 cells, which are fibroblast-like cells from the African green monkey kidney, in contrast to our studies which were performed using fully differentiated, hTERT-immortalized RPTCs from humans. Third, they performed detergent-free sucrose gradient fractionation on whole cell lysates, which include intracellular membranes, while we performed detergent-free sucrose gradient centrifugation on plasma membrane-enriched fractions (13). In a subsequent report, Kong et al reported that “constitutive palmitoylation may serve to stabilize the D₁R during agonist-dependent caveolar internalization” (60).

Mutation of the palmitoylation sites or the CRAC motif resulted in D₁R dysfunction. A higher basal cAMP production was observed in the presence of the D₁R mutants, similar to the results observed upon the disruption of lipid rafts (24). The exclusion of these mutants from the lipid rafts may have prevented the inhibitory effect of the lipid raft protein CAV-1 on Gα_s and/or facilitated GTP binding (61). Alternatively, these mutations may have changed the basal conformation of the receptor into a constitutively active version (62) by locking the receptor in either a high affinity “agonist conformation” or low affinity “antagonist conformation” (63). In contrast to our current and published (40) results which showed an impairment of cAMP production in response to the D₁R/D₅R-specific agonist fenoldopam among the mutant D₁Rs, Jin et al. reported that D₁R palmitoylation mutants had normal adenylyl cyclase activity in response to dopamine (61). The difference in our results may result from their use of: intact cells (vs. isolated plasma membranes in our studies) to measure adenylyl cyclase activity (vs. intracellular cAMP levels); dopamine (the natural ligand for all of the dopamine receptor isoforms, vs. fenoldopam, which is more specific for D₁R/D₅R (64)); and baby hamster kidney cells vs. adult hRPTCs.

Our study demonstrates that disruption of the lipid raft via cholesterol depletion, siRNA-mediated silencing of Drd1 or Snx19 genes, or Drd1 C347A and Y218S mutations result in increased oxidative stress in RPTCs or C57BL/6J mice. Lipid raft bestows protection against oxidative stress by maintaining NOX in the inactive state in hRPTCs (24), the opposite of what is observed in normotensive rat RPTCs (65). However, RPTCs from hypertensive rats have greater basal oxidase activity conceivably due to greater NOX2 and Rac1 abundance in their lipid rafts (65). Agonist stimulation of D₁R may directly regulate the level of ROS through its interaction with NOX and paraoxonase 2 (PON2) (25) and nuclear factor E2-related factor 2 (66). The absence of increased ROS levels and basal cAMP production when ΔD₁R-351A is expressed is reminiscent of Jensen et al’s findings that it is mainly the ΔD₁R-C347 mutant that is impaired functionally (59). The ability of fenoldopam, through the D₁R/D₅R, to inhibit NOX activity is, in part, due to its ability to increase cAMP production, via a PKC and PKA crosstalk (67). However, while ΔD₁R-351A was able to prevent the ability of fenoldopam to increase cAMP and inhibit basolateral sodium transport, its ability to decrease ROS production persisted. This could be due to the ability of D₁R to stimulate antioxidant enzymes such as superoxide dismutase, glutathione peroxidase, glutamyl cysteine transferase, and heme-oxygenase 1, independent of cAMP. Indeed, the ability of D₁R and D₅R to stimulate PON2 activity is related, in part, to its translocation to non-raft domains (25).

We have demonstrated that disruption of lipid rafts via β-MCD impairs the localization of renal D₁R in the brush borders of proximal tubules, resulting in decreased sodium excretion and high blood pressure in mice on normal salt diet, similar to the results by Gildea et al in rats (11). In our study, the concomitant decrease in brush border distribution of SNX19 and GRK4 may conceivably contribute to the D₁R dysfunction. The activity of the sodium transporters and NKA may also be directly affected by raft disruption since they are mostly lipid raft-bound (68). The crucial role of the C347 palmitoylation site was corroborated by our gene rescue studies; renal tubule-restricted rescue with DRD1-WT but not DRD1-347A mutant, subcloned in AAV vectors, normalized the blood pressure at day 21 in the mice with renal-restricted silencing of Drd1. The siRNA-mediated renal-restricted Drd1 knockdown mouse model was used in lieu of the Drd1 knockout mice because we wanted to eliminate non-renal factors in the process.

SNX19 is normally exclusively found in the lipid raft and may be an additional, extrinsic factor that ensures the subsequent lipid raft partitioning of D₁R. It may serve as a hub that provides an entropic advantage and promotes the effective interaction between D₁R and the GODZ enzyme, resulting in D₁R palmitoylation and lipid raft distribution (15–17,23). Unlike SNX1 and SNX5, SNX19 only possesses the canonical PX domain but not the Bin/Amphiphysin/Rvs (BAR) domain, which is a dimerization and membrane-curvature sensing module (69). It is conceivable that the ability of SNX19 to foster the partitioning of D₁R to the lipid raft may reside in the PX domain, in the absence of the BAR domain, since the PX domain is important in protein localization in lipid rafts (70). Lipid raft targeting is not specific to D₁R but probably needs SNX19 to target it to lipid rafts, ensuring D₁R specificity. Deletion of the PX domain of SNX19 prevents the targeting of D₁R to lipid rafts and renal-restricted silencing of Snx19 in mice increases blood pressure. Moreover, SNX19 regulates the renal expression of D₁R but not D₅R, the other D₁-like dopamine receptor.

In summary, renal D₁R requires lipid raft residency for its full functionality (signal transduction), which includes increase in cAMP production, inhibition of sodium transport, inhibition of ROS production, and blood pressure homeostasis (Figure 14). SNX19 acts as an adaptor or scaffold protein for the palmitoylation of renal D₁R required for lipid raft targeting.

Figure 14. Palmitoylated D₁R in lipid raft microdomains.

Nascent D₁R undergoes post-translational modification at the Golgi apparatus. This includes palmitoylation of the cysteine residues at positions 347 and 351 found at the C-terminus through the GODZ enzyme, a process that may be facilitated by the adaptor protein SNX19. Palmitoylation promotes the partitioning of the receptor into the lipid raft where other components of the signal transduction pathway are localized, including the GRK4, G proteins, β-arrestins, adenylyl cyclases, other receptors, and effector proteins, such as the NOX and NKA, for normal signal transduction and appropriate cellular response. Negligible signaling may occur in the non-raft domains where some of the signaling molecules are missing in the case of the D₁R.

Perspectives

Disruption of lipid rafts is implicated in renal diseases such as glomerulosclerosis, lupus nephritis, and unilateral ureteropelvic junction obstruction (12,71,72) and other cardiovascular diseases such as heart failure, myocardial infarction, and pulmonary hypertension (73). Lipid rafts are important in dopamine function via the D₁R. Abnormal D₁R signaling is involved in the pathogenesis of renal disease and hypertension. Thus, it is important to evaluate the dynamics among the various components of the D₁R signaling clusters in the lipid rafts of renal tubule cells. Moreover, other determinants of lipid raft partitioning such as CARC motif, GXXG motif, tilted domain, CAV-binding motif, and N-linked myristoylation of glycine residues in the D₁R should be studied to determine their biological relevance to D₁R-mediated regulation of renal function and blood pressure.

Abbreviations:

AAV

adeno-associated virus

ANOVA

analysis of variance

ARF

animal research facility

BCA

bicinchoninic acid

β-MCD

methyl-beta-cyclodextrin

cAMP

cyclic adenosine monophosphate

CAV

caveolin

CRAC motif

cholesterol recognition/interaction amino acid consensus

co-IP

co-immunoprecipitation

CTxB

Cholera Toxin subunit B

DAPI

diamino phenylindole

D₁R

dopamine 1 receptor

GPCR

G protein-coupled receptor

DMEM/F-12

Dulbecco’s Modified Eagle Medium/Nutrient Mixture F-12

GGT

gamma-glutamyl transferase

GAPDH

glyceraldehyde 3-phosphate dehydrogenase

hTERT

human telomerase reverse transcriptase

GODZ

Golgi-associated DHHC-type zinc finger protein

GRK4

GPCR kinase 4

LTL

Lotus tretragonolobus lectin

NKA

sodium/potassium-ATPase

NOX

NADPH oxidase

PPT1/2

palmitoyl-protein thioesterase 1/2

PON2

paraoxonase 2

RPTCs

renal proximal tubule cells

ROS

reactive oxygen species

siRNA

short, interfering RNA

SNX19

sorting nexin 19

SERZ-β

sertoli cell gene with a zinc finger domain-β

STED

Stimulated Emission Depletion Microscopy

UNaV

urinary sodium excretion

WT

wild-type