G Protein-coupled Receptor Kinase 4 (GRK4) Regulates the Phosphorylation and Function of the Dopamine D₃ Receptor

Abstract

During conditions of moderate sodium excess, the dopaminergic system regulates blood pressure and water and electrolyte balance by engendering natriuresis. Dopamine exerts its effects on dopamine receptors, including the dopamine D₃ receptor. G protein-coupled receptor kinase 4 (GRK4), whose gene locus (4p16.3) is linked to essential hypertension, desensitizes the D₁ receptor, another dopamine receptor. This study evaluated the role of GRK4 on D₃ receptor function in human proximal tubule cells. D₃ receptor co-segregated in lipid rafts and co-immunoprecipitated and co-localized in human proximal tubule cells and in proximal and distal tubules and glomeruli of kidneys of Wistar Kyoto rats. Bimolecular fluorescence complementation and confocal microscopy revealed that agonist activation of the receptor initiated the interaction between D₃ receptor and GRK4 at the cell membrane and promoted it intracellularly, presumably en route to endosomal trafficking. Of the four GRK4 splice variants, GRK4-γ and GRK4-α mediated a 3- and 2-fold increase in the phosphorylation of agonist-activated D₃ receptor, respectively. Inhibition of GRK activity with heparin or knockdown of GRK4 expression via RNA interference completely abolished p44/42 phosphorylation and mitogenesis induced by D₃ receptor stimulation. These data demonstrate that GRK4, specifically the GRK4-γ and GRK4-α isoforms, phosphorylates the D₃ receptor and is crucial for its signaling in human proximal tubule cells.

During conditions of moderate sodium excess, the dopaminergic system sits at the fulcrum of homeostatic control of water and electrolyte balance and blood pressure (1, 2). Dopamine promotes natriuresis by inhibiting sodium chloride reabsorption in specific segments of the nephron. Dopamine exerts its action on dopamine receptors, which belong to the family of G protein-coupled receptors (GPCRs).² The dopamine receptors are classified into two subtypes based on their ability to increase cAMP levels, sequence similarity, G protein coupling, and pharmacological profiles (3, 4). The D₁-like dopamine receptors activate adenylyl cyclase by coupling to stimulatory Gα_s/Gα_olf and include the D₁ (D₁R) and D₅ receptors (D₅R). The D₂-like dopamine receptors inhibit adenylyl cyclase by coupling to Gα_i/Gα_o and consist of the D₂ (D₂R), D₃ (D₃R), and D₄ (D₄R) receptors. The D₃R has also been shown to couple to Gα_o, Gβγ, and to the stimulatory Gα_s (5, 6).

The signal transduction that follows ligand occupation of a GPCR is tightly regulated to limit the specificity and extent of cellular response. GPCR-mediated signal transduction is rapidly dampened via receptor desensitization or the waning of the responsiveness of the receptor to agonist with time. Desensitization involves receptor phosphorylation and is carried out by either GPCR kinases (GRKs) or second messenger-activated kinases such as protein kinase A and protein kinase C. Homologous desensitization involves GRKs that selectively phosphorylate only agonist-activated receptors, whereas heterologous desensitization is carried out by second messenger-dependent kinases that indiscriminately phosphorylate agonist-activated receptors and those that have not been exposed to the agonist (7).

The GRKs are serine/threonine protein kinases comprising seven isoforms that are grouped into three subfamilies. GRK1 and GRK7 belong to the rhodopsin kinase subfamily and are expressed exclusively in the retina (8–10). GRK2 and GRK3 phosphorylate the β-adrenergic receptor and belong to the β-adrenergic receptor kinase subfamily (11), and GRK4, GRK5, and GRK6 belong to the GRK4 subfamily. GRK4 is highly enriched in the testis and, to a lesser degree, in the kidneys (12, 13). Four splice variants of human GRK4 result from the alternative splicing of exons 2 and 15 (11). GRK4-α is considered the full-length version, whereas GRK4-β, -γ, and -δ are shortened versions of GRK4-α (14). The coding region of the GRK4 gene, whose 4p16.3 locus has been linked to essential hypertension (15, 16), contains several single nucleotide polymorphisms, including R65L, A142V, and A486V, which have been linked to essential hypertension and/or salt sensitivity in various ethnic groups (17).

The D₃R gene is found at 3q13.3 (18), a locus that is also linked to essential hypertension (19, 20). Sequence analysis of the D₃R gene shows the presence of several single nucleotide polymorphisms, which do not correlate with either essential hypertension among Japanese (21) or with blood pressure levels and diabetic nephropathy among Finns (22). However, D₃R knock-out mice develop a renin-dependent form of hypertension and fail to excrete a sodium load (23).

The D₃R has a long third intracellular loop that contains several putative GRK phosphorylation sites (24). A previous study evaluated the ability of GRK2 and GRK3 to phosphorylate D₃R and showed that co-transfection of GRK3, but not GRK2, resulted in a weak phosphorylation of the heterologously expressed, dopamine-stimulated D₃R in HEK-293 (25), a human embryonic kidney cell line. We tested the hypothesis that GRK4 is required in D₃R signaling in terminally differentiated human renal proximal tubule cells (hPTCs) by determining the spatiotemporal dynamics of the interaction of D₃R and GRK4 through their subfractionation in membrane microdomains and subcellular co-localization via confocal microscopy and bimolecular fluorescence complementation assay (BiFC). We also identified which of the GRK4 splice variants are involved in D₃R phosphorylation and evaluated the physiological roles of GRK4 in D₃R signaling in the hPTCs. We now report that D₃R and GRK4 co-fractionate in lipid rafts and co-localize in both hPTCs and WYK kidneys. Moreover, D₃R is phosphorylated by GRK4-γ and GRK4-α isoforms, and the absence of GRK4 impairs D₃R-mediated mitogenesis and activation of p44/42 in hPTCs.

EXPERIMENTAL PROCEDURES

Constructs

The human wild-type GRK4 (α, β, γ, and δ) splice variants were derived from in-house cDNA clones. Each cDNA was subcloned into the inducible mammalian expression vector, pcDNA4/TO (Invitrogen) for tetracycline-regulated expression in Chinese hamster ovary cells (T-REx CHO) (Invitrogen). The cloned inserts were completely sequenced. The expression of each GRK4 isoform was initially verified in HEK-293 cells via Western blot.

pBiFC-YN155-bJun and pBiFC-YC155-bFos vectors (26, 27) were kind gifts of Dr. Tom K. Kerppola. The b-Jun and b-Fos sequences were removed from the vectors, and new multiple cloning sequences were inserted into both vectors to make new pBiFC vectors, designated as pBiFC-YN155 and pBiFC-YC155. The YN vector contains the N-terminal portion of EYFP (residues 1–155) and the YC vector contains the C-terminal portion (residues 156–241) of EYFP, both fused to the C terminus of the protein of interest. D₃R cDNA was subcloned into the pBiFC-YN vector. GRK4γ was PCR-amplified and inserted into the pBiFC-YC vector, followed by oligonucleotide-directed mutagenesis to change the in-frame stop codon within the PmeI TAA to AAA to create GRK4γ-YC. The fidelity of the cloned inserts was verified by DNA sequencing, and their expression was evaluated by reverse transcription-PCR and immunoblotting.

Cell Lines and Cell Culture

Immortalized hPTCs were obtained from normotensive Caucasian males. The cells were maintained in DMEM/F-12 (Invitrogen) with 10% fetal bovine serum, epidermal growth factor (10 ng/ml), insulin, transferrin and selenium mixture (5 μg/ml each), and dexamethasone (4 ng/μl) at 37 °C in an incubator with humidified 5% CO₂ and 95% air. Cells with low passage numbers (<18) were used to avoid the confounding effects of cellular senescence.

T-REx CHO cells (Invitrogen) were grown in DMEM/F-12 supplemented with 10% tetracycline-free fetal bovine serum and blasticidin (10 μg/ml; Invitrogen) to select for cells harboring the pcDNA6/TR plasmid. The cells were stably transfected with the various GRK4 isoforms using Lipofectamine^TM 2000 (Invitrogen), and double transfectants were selected with the addition of both Zeocin (100 μg/ml, Invitrogen) and blasticidin to the cell culture medium. The cell lines tested negative for Mycoplasma infection. There are no reports on the endogenous GRK4 expression in CHO cell lines; previous experiments on D₁R desensitization in this cell line required the stable transfection of GRK4 (13).

Sucrose Gradient Fractionation

To prepare lipid raft and non-lipid raft fractions, sucrose gradient ultracentrifugation was performed using a detergent-free protocol (28) with modifications (29). hPTCs grown to 95% confluence were harvested and lysed in 500 mmol/liter sodium carbonate (pH 11) and homogenized by sonication. One ml of the homogenate was diluted with 2 ml of 80% sucrose and overlaid with 6 ml of 35% sucrose and 3 ml of 5% sucrose, and spun at 160,000 × g in a Beckman SW40 rotor at 4 °C for 18 h. The sucrose solutions were prepared in MBS solution (25 mm MES, pH 6.7; 150 mm NaCl). After centrifugation, 12 1-ml fractions were collected and labeled 1 to 12 from top to bottom. Aliquots of each fraction were mixed with Laemmli buffer, boiled, and immunoblotted. Treatment with methyl-β-cyclodextrin (MCD, a cholesterol depletor) for 30 min prior to the collection of the cells served as control.

Immunoblotting

Total cell lysates were prepared using lysis buffer supplemented with protease inhibitors. Uniform amounts of protein (measured using the BCA kit, Pierce) were prepared using Laemmli loading buffer, resolved in 10% SDS-polyacrylamide gel, and transferred onto a nitrocellulose membrane. The membrane was blocked and sequentially probed with the primary antibody and the corresponding secondary antibody. Chemiluminescence detection was carried out using SuperSignal West Dura Substrate (Pierce) followed by autoradiography.

Confocal Microscopy

hPTCs were grown on poly-d-lysine-coated coverslips to 50% confluence and treated with the D₃R agonist PD128907 (1 μm) at the indicated time points. The cells were washed three times with ice-cold PBS, and the cell membrane was labeled with the cell-impermeant EZ-link sulfo-NHS-SS-Biotin (Pierce) for 30 min on ice and then washed with ice-cold PBS supplemented with 10 mm glycine (pH 7.0) to remove the excess biotin. The cells were fixed with 4% paraformaldehyde, permeabilized with 0.05% Triton X-100 for 10 min, double-immunostained for D₃R and GRK4 using monoclonal anti-D₃R antibody (Zymed Laboratories Inc.) and rabbit polyclonal anti-GRK4 antibody (Santa Cruz Biotechnology) for 1 h, and re-probed with goat anti-mouse (H+L)-Alexa 555 and goat anti-rabbit (H+L)-Alexa 633 (Molecular Probes) secondary antibodies. The membranes were re-probed with Cy3-conjugated avidin for 10 min. The coverslips were mounted on glass slides using hard-set Vectashield mounting medium.

Four-μm-thick sections of formalin-fixed, paraffin-embedded WKY rat kidneys were prepared and mounted on slides. After deparaffination, rehydration, permeabilization, and antigen retrieval, the sections were double-immunostained with monoclonal anti-D₃R antibody (Zymed Laboratories Inc.) and rabbit polyclonal anti-GRK4 antibody (Santa Cruz Biotechnology). The tissues were probed with goat anti-mouse (H+L)-Alexa 488 and goat anti-rabbit (H+L)-Alexa 568 (Molecular Probes) secondary antibodies and mounted using Vectashield mounting medium. Negative controls were likewise prepared by omitting the primary antibodies. Co-localization of endogenous D₃R and GRK4 in both hPTCs and WKY rat kidney sections was evaluated by laser scanning confocal microscopy using an Olympus IX-70 inverted microscope equipped with a 60X/1.4 NA oil immersion objective at the Microscopy and Imaging Shared Resource of the Lombardi Cancer Center, Georgetown University.

Co-immunoprecipitation

Serum-starved hPTCs were treated with PD128907 or PBS as control, and the cell lysates were prepared using RIPA lysis buffer. Equal amounts of cell lysates (500 μg of protein) were mixed with monoclonal anti-D₃R (Zymed Laboratories Inc.), or normal mouse IgG (Santa Cruz Biotechnology) as negative control, or polyclonal anti-GRK4 (Abgent) as positive control. The immune complexes were pelleted out, and the bound proteins were eluted using 30 μl of Laemmli buffer. The samples were subjected to immunoblotting and probed for GRK4 using a polyclonal anti-GRK4 antibody (Abgent). Only a quarter of the eluate for the positive control was run alongside the test samples to avoid oversaturation of the band during x-ray radiography. The ability of D₃R to co-immunoprecipitate with the two other members of the GRK4 subfamily, i.e. GRK5 and GRK6, was also evaluated by probing for these proteins using specific anti-GRK5 and anti-GRK6 polyclonal antibodies (Santa Cruz Biotechnology).

Bimolecular Fluorescence Complementation

pBiFC-GRK4γ-YC and pBiFC-D₃R-YN were co-transfected into hPTCs grown on coverslips using FuGENE 6 transfection reagent (Roche Applied Science). Untransfected cells were also prepared as negative controls. The cells were incubated at 37 °C for 24 h and at 32 °C for 16 h to allow the maturation of the fluorophore. The cells were serum-starved prior to D₃R stimulation with PD128907 (1 μm) at the indicated time points. The cell membrane was labeled with the cell-impermeant EZ-link sulfo-NHS-SS-Biotin for 30 min on ice and then washed with ice-cold 10 mm glycine (pH 7.0) in PBS. The cells were fixed with 4% paraformaldehyde and permeabilized with 0.05% Triton X-100 in PBS. The cells were probed with Cy3-conjugated avidin and mounted on glass slides using Vectashield mounting medium containing the nuclear stain DAPI. Confocal images were obtained with an Olympus IX-70 laser scanning confocal microscope with a Plan Apo 60X/1.4 NA oil immersion objective. To determine the extent of protein-protein interaction, i.e. BiFC formation in the hPTCs, the mean fluorescence intensity was quantified using the Scion Image software (National Institutes of Health, Bethesda).

Whole Cell Phosphorylation Assay

T-REx CHO cells stably transfected with the various GRK4 isoforms were grown to 60% confluence and then transiently transfected with full-length human D₃R subcloned into pcDNA4/myc,His using FuGENE 6 transfection reagent to express His-tagged D₃R. The cells were treated with the tetracycline analog, doxycycline, for 24 h to induce the expression of the GRK4 splice variants. The culture medium was replaced with phosphate-free DMEM (Invitrogen) containing doxycycline for 1 h and replaced anew with fresh phosphate-free DMEM with doxycycline and 106 μCi of [³²P]H₃PO₄, and the cells were returned to the incubator for another 90 min. The cells were stimulated with PD128907 (1 μm, 30 min) or with PBS as control. Thereafter, the cell lysates were prepared using RIPA buffer supplemented with protease and phosphatase inhibitors (Sigma). His-tagged D₃R was pulled down from lysates with uniform amounts of total protein (500 μg) using the MagneHis^TM protein purification system (Promega). In a separate set of experiments, aliquots of the lysates were resolved by SDS-PAGE and probed for glyceraldehyde-3-phosphate dehydrogenase to confirm the uniformity of protein concentration. Copious washing was done to minimize residual nonspecific ³²P radioactivity. His pulldown assay was used in lieu of immunoprecipitation with a D₃R-specific antibody to avoid the possible confounding effect of the heavy chain of the immunoprecipitant, which has approximately the same molecular size as D₃R (∼50 kDa). The eluates were mixed with Laemmli buffer, resolved in 10% SDS-PAGE, and subjected to autoradiography for 24–48 h. T-REx CHO cells stably transfected with the pcDNA4/TO vector without insert and transiently transfected with pcDNA4/myc,His-hD₃R plasmid were used as negative control. The proteins were transferred onto a nitrocellulose membrane after 48 h and then probed for D₃R-His using an anti-His monoclonal antibody.

Mitogenesis Assay

D₃R-mediated mitogenesis of hPTCs was assessed using the Delfia cell cytotoxicity kit (PerkinElmer Life Sciences). Approximately 1 × 10⁴ hPTCs were seeded into a 96-well Isoplate^TM (PerkinElmer Life Sciences) and grown in complete media. Bromodeoxyuridine (BrdUrd, 10 μm), a thymidine analog, was added to the medium 24 h prior to the end of the assay. To determine the extent of mitogenesis after 18 h of D₃R stimulation, the cells were fixed, the DNA was denatured, and the amount of incorporated BrdUrd was determined using a europium-labeled monoclonal anti-BrdUrd antibody. Unbound antibody was washed off, and an inducer buffer was added into each well to dissociate the europium from the antibody, resulting in highly fluorescent chelates. The europium fluorescent signal was then determined by time-resolved fluorometry (with a 400-μs window time and delay) using a Victor³ multilabel reader. The same number of hPTCs cells was seeded on a separate 96-well Isoplate^TM and grown overnight before fixation with 4% paraformaldehyde and nuclear staining with DAPI. The DAPI fluorescence was detected using a Victor³ multilabel reader. The results are reported as europium fluorescence normalized to that of DAPI.

Inhibition of GRK4 Function and Expression

GRK function was inhibited by transfecting heparin (1 μm/24 h; Sigma), a nonspecific GRK inhibitor (30), into the hPTCs using FuGENE 6 transfection reagent. In additional studies, GRK4 expression was silenced via GRK4-specific siRNA (5 nm/72 h; Qiagen) transfected into the cells using Hyperfect transfection reagent (Qiagen). Nonsilencing “mock” siRNA (Qiagen) was also transfected into cells as negative control. Untransfected cells served as an additional control. The extent of inhibition of GRK4 expression in the cells was evaluated by reverse transcription-PCR and immunoblotting. The GRK4 mRNA expression was evaluated using GGAGGAGGTCGATCAAAGAA (sense) and AGACACACCCGGTAGCAAAC (antisense), resulting in a 343-bp product. For normalization, β-actin mRNA expression was also evaluated using AGAAAATCTGGCACCACACC (sense) and CTCCTTAATGTCACGCACGA (antisense), resulting in a 388-bp product.

MAP Kinase Phosphorylation

Serum-starved hPTCs grown to 90% confluence were treated with PD128907 (1 μm) at the indicated time points to determine the phosphorylation state of the p44/42 MAP kinases. Total cell lysates were prepared in the presence of a mixture of protease and phosphatase inhibitors (Sigma) and then immunoblotted for phosphorylated p44/42 (Cell Signaling) and for a panel of phosphorylated proteins (Cell Signaling) thereafter.

Statistical Analysis

Numerical data are expressed as means ± S.E. Significant difference between two groups was determined by Student's t test, whereas that among three or more groups was determined by one-way ANOVA followed by Holm-Sidak or Dunnett's post hoc test, as indicated. p < 0.05 was considered significant. Statistical analysis was performed using SigmaStat 3.5 (SPSS Inc., Chicago).

RESULTS

Distribution of D₃R and GRK4 in Membrane Microdomains

In the plasma membrane, distinct islands of lipids and proteins collectively known as lipid rafts exist to spatially concentrate functionally related sets of proteins to facilitate and augment their interaction, thereby enhancing the efficiency and specificity of their signal transduction. To evaluate the subfractionation of endogenous D₃R and GRK4 in lipid rafts, sucrose gradient ultracentrifugation was performed. Both D₃R and GRK4 were found in non-lipid raft fractions (fractions 7–12) and in lipid raft fractions (fractions 1–6), as denoted by the abundance of the lipid raft marker, caveolin-1 (Fig. 1). Estimation of total proteins found in these microdomains revealed that >90% are found in non-lipid rafts, and <10% are found in lipid rafts (31). Treatment with MCD, a cholesterol-depleting reagent, resulted in the disruption of the cholesterol-enriched lipid rafts and the re-distribution of caveolin-1, D₃R, and GRK4 to the non-lipid rafts.

FIGURE 1.

Distribution of endogenous D ₃R and GRK4 in membrane microdomains of hPTCs. Lipid and non-lipid membrane fractions of hPTCs were prepared by sucrose gradient centrifugation to determine the basal membrane distribution of both D₃R and GRK4. Twelve fractions were obtained (fractions 1–6 correspond to lipid rafts and fractions 7–12 to non-lipid rafts) and immunoblotted for D₃R, GRK4, and caveolin-1, a commonly used marker of lipid rafts. Cells pretreated with MCD, a cholesterol-depleting and lipid membrane-disrupting agent, were used as control. n = 3 independent experiments.

Co-localization between D₃R and GRK4

In hPTCs, endogenously expressed D₃R was both membrane-bound and scattered in the cytoplasm under basal conditions, suggesting basal turnover of the receptors (Fig. 2A). D₃R stimulation resulted in the rapid internalization and sequestration of the receptors at the perinuclear area (5 and 15 min), followed by the dispersal of the receptors to the membrane (30 min). GRK4 was distributed both at the surface membrane and dispersed in the cytoplasm under basal conditions. D₃R stimulation resulted in the internalization of GRK4 to the perinuclear area, where co-localization with D₃R was observed strongly at 5 and 15 min of agonist treatment.

FIGURE 2.

Co-localization of D ₃R and GRK4 in hPTCs and kidney sections from Wistar Kyoto rats. A, hPTCs grown on poly-d-lysine-coated coverslips were serum-starved for 1 h and treated with the D₃R agonist PD128907 (1 μm) at the indicated duration of treatment. The cell membrane was labeled with a membrane-impermeant biotin, after which the cells were fixed with 4% paraformaldehyde and permeabilized with 0.05% Triton X-100 in PBS. The cells were double-immunostained for D₃R (pseudocolored red) and GRK4 (pseudocolored green). The membrane (pseudocolored blue) was probed with Cy3-conjugated avidin. The distribution and co-localization of D₃R and GRK4 (shown as discrete yellow areas in merge and inset images) and of D₃R or GRK4 and the cell membrane (magenta and cyan in inset images, respectively) were evaluated by laser scanning confocal microscopy. B, formalin-fixed, paraffin-embedded kidney sections of WKY rats were prepared to determine in vivo co-localization of D₃R (pseudocolored red) and GRK4 (pseudocolored green) by confocal microscopy. Differential interference contrast (DIC) images were also obtained to show the cellular and histological boundaries. G indicates glomerulus; yellow arrow indicates proximal tubule (S₁ segment); red arrow indicates proximal tubule (S₃ segment), and white arrow indicates distal tubule. Scale bar, 10 μm, ×600 magnification, n = 3–5 independent experiments.

The in vivo correlation of the D₃R and GRK4 co-localization was evaluated in paraffin sections of WKY rat kidney (Fig. 2B). GRK4 was strongly expressed in the proximal tubules (S₁ and S₃ segments), the distal convoluted tubules, and the glomeruli. On the other hand, D₃R was predominantly expressed in the proximal tubules (S₁ and S₃ segments) and distal convoluted tubules, in agreement with previous reports (32–34), and to a lesser extent in the glomeruli. The segments that showed the strongest co-localization signal were the proximal and distal tubules.

Direct Interaction between D₃R and GRK4

To determine the physical interaction between the two proteins, total cell lysates were immunoprecipitated using a monoclonal anti-D₃R antibody, and the eluate was immunoblotted for GRK4. An ∼55-kDa band was visualized (Fig. 3A), which corresponded to GRK4 because a similar band was obtained when the lysate was immunoprecipitated with a polyclonal anti-GRK4 antibody. Minimal co-immunoprecipitation was observed when a pool of normal mouse IgGs was used as immunoprecipitant. There was no difference in the extent of co-immunoprecipitation between D₃R and GRK4 under basal conditions and after treatment with a D₃R agonist. D₃R failed to immunoprecipitate both GRK5 and GRK6 (supplemental Fig. S1A).

FIGURE 3.

Interaction of D ₃R and GRK4. A, capacity for D₃R and GRK4 to physically interact was determined in hPTCs treated with the D₃R agonist PD128907 (1 μm) at the indicated duration of treatment. Total cell lysates were prepared, and a monoclonal anti-D₃R Ab, or normal mouse IgG as negative control, or polyclonal anti-GRK4 Ab as positive control was used as immunoprecipitant. The immune complexes were precipitated using agarose-A/G beads and immunoblotted for GRK4. A ∼55-kDa band corresponding to GRK4 was visualized. B, hPTCs were double-transfected with D₃R and GRK4-tagged with the C and N termini of the fluorescent protein EYFP, respectively. The cells were grown for 48 h post-transfection, serum-starved for 2 h, stimulated with the D₃R agonist PD128907 (1 μm) at the indicated time points, and then prepared for confocal microscopy. The cell membrane (CM) was biotinylated with a membrane-impermeant biotin to allow visualization (pseudocolored red) and co-localization with the BiFC signal (pseudocolored green). Co-localization of the BiFC signal with the CM is indicated by yellow punctate areas in the merged images. An overlay of the BiFC signal and the nucleus (pseudocolored blue) is shown to indicate intracellular distribution. Untransfected cells treated with the D₃R agonist for 5 min were used as negative control (control). Scale bar, 10 μm, ×600 magnification, n = 3–4 independent experiments.

The BiFC assay was performed to determine the spatial and temporal aspects of the D₃R and GRK4 interaction in their natural cellular environment. This technique has the advantage of allowing the visualization of protein-protein interaction in cells with neither the need to disrupt subcellular compartmentalization nor the use of exogenous fluorophore-labeled antibodies (35). GRK4 tagged with the N terminus of EYFP and D₃R tagged with the C terminus of the same fluorophore were co-expressed in hPTCs. The physical interaction between D₃R and GRK4 in situ was indicated by a fluorescent signal visualized by confocal microscopy (Fig. 3B). In untransfected control cells, no fluorescence was observed, even after 5 min of agonist treatment. However, fluorescence signal was observed (4.16 ± 0.42 a.u. (arbitrary units) versus 1.06 ± 0.12 a.u. for untransfected control cells) in untreated, co-transfected cells, suggesting basal interaction between the two proteins. Treatment with the D₃R agonist markedly increased the fluorescence at 5 and 15 min of treatment (18.1 ± 1.02 a.u. and 17.0 ± 0.95 a.u., respectively), indicating that enhanced D₃R and GRK4 interaction is agonist-driven in hPTCs. Roughly 30% of scored cells (30 cells/experiment) in three separate experiments showed the fluorescent signal. To determine the spatial aspects of D₃R and GRK4 interaction, cell membrane proteins were labeled with a membrane-impermeant biotin to allow the visualization of the surface membrane, whereas DAPI was used to visualize the nucleus. The BiFC signal was observed both on the surface membrane and at the cytoplasm at 5 and 15 min of agonist treatment.

GRK4-mediated D₃R Phosphorylation

The functional significance of the interaction between D₃R and GRK4 was initially assessed by determining the ability of GRK4 to phosphorylate D₃R and by identifying which of the four GRK4 splice variants phosphorylate the receptor. The capacity of the GRK4 splice variants to phosphorylate D₃R was assessed in T-REx CHO cells, which uniformly express the GRK4 splice variants (supplemental Fig. S1B), by the amount of ³²P that was transferred to the receptor upon agonist activation. Agonist treatment did not change the extent of D₃R phosphorylation in the empty vector-transfected control cells (Fig. 4). Under basal conditions, GRK4-γ and GRK4-α mediated an ∼50% increase (48 ± 0.1 and 43 ± 0.04% increase, respectively) in phosphorylated D₃R compared with untreated control cells. Treatment with the D₃R agonist resulted in a 3-fold and 2-fold increase in D₃R phosphorylation in cells expressing GRK4-γ (298 ± 0.9%) and GRK4-α (195 ± 0.9%), respectively, compared with untreated control cells. Basal and agonist-mediated D₃R phosphorylation by GRK4-β and GRK4-δ tended to increase compared with those of control cells but did not reach statistical significance.

FIGURE 4.

GRK4-mediated p hosphorylation of D ₃R. T-REx CHO cells stably transfected with the tetracycline-inducible GRK4 splice variants, or empty vector as control, were transfected with His-tagged D₃R. The expression of the transgenes was induced by the addition of doxycycline, a tetracycline analog, 24 h prior to D₃R activation. The transfected cells were metabolically labeled with [³²P]H₃PO₄ before stimulation with the D₃R agonist PD128907 (1 μm). The heterologously expressed D₃R was pulled down from uniform amounts of protein (500 μg, confirmed by immunoblotting for glyceraldehyde-3-phosphate dehydrogenase (GAPDH)) using a His pulldown kit and resolved in 10% SDS-PAGE. Thereafter, the gel was subjected to autoradiography for 24–48 h. Blots showing the amounts of phosphorylated D₃R (phospho-D₃R) and total His-tagged D₃R are shown in the upper panel. Band densities were quantified via Scion densitometry software. Data are expressed as mean ± S.E. , p < 0.05, versus untreated cells transfected with the same GRK4 isoform, t test. *, p < 0.05, versus untreated empty vector transfected control, one-way ANOVA followed by Holm-Sidak post hoc test. n = 3 independent experiments.

D₃R-mediated Mitogenesis

Stimulation of D₃R promotes mitogenesis, or cell proliferation, a function that is shared with the other D₂-like dopamine receptors (36). D₃R-mediated mitogenesis was evaluated in hPTCs using the amount of BrdUrd that was incorporated into the DNA of dividing cells (Fig. 5A). Agonist-stimulated D₃R resulted in increased mitogenesis compared with vehicle-treated controls (135.0 ± 5.4 versus 100 ± 5.0%). The observed increase in cell proliferation was specifically elicited by D₃R stimulation because this response was abolished by co-treatment with the D₃R antagonist GR103691 (103.72 ± 5.0%), which had no effect on mitogenesis by itself (94.3 ± 4.9%). To ascertain if GRKs are required for the D₃R-mediated mitogenesis in hPTCs, GRK function was inhibited through an overnight pretreatment with heparin, a nonselective GRK inhibitor (30). As was previously observed, D₃R activation resulted in increased mitogenesis compared with vehicle-treated, non-heparin-transfected control (121.9 ± 4.9 versus 100 ± 2.0%) (Fig. 5B). However, agonist treatment failed to promote mitogenesis in cells pretreated with heparin (95.1 ± 3.3%), indicating that one or several GRKs are required for D₃R-mediated mitogenesis in hPTCs. Pretreatment with heparin had no effect on mitogenesis (97.3 ± 2.2%).

FIGURE 5.

D ₃R-mediated m itogenesis. A, hPTCs were grown in 24-well plates and incubated with BrdUrd, a thymidine analog, before treatment with the D₃R agonist PD128907 (PD, 1 μm) and/or the D₃R antagonist GR103691 (GR, 1 μm) or vehicle as control (C). B and C, function or expression of GRK4 was inhibited by transfecting the hPTCs with heparin (Hep) or GRK4-specific silencing siRNA (siRNA), respectively, prior to BrdUrd incubation and treatment with the D₃R agonist PD128907 or vehicle (V) as control. Cells transfected with vehicle (control) or nonsilencing siRNA (mock) served as controls. Thereafter, the cells were fixed and the extent of cell proliferation was determined by the amount of BrdUrd that was incorporated using an anti-BrdUrd Ab conjugated with europium. Europium fluorescence was determined by time-resolved fluorometry using a Victor³ multilabel reader and normalized for that of the nuclear stain DAPI. Data are expressed as mean ± S.E. , p < 0.05, versus vehicle-treated pair, t test. *, p < 0.05, versus vehicle-treated (untransfected) control, one-way ANOVA, and Holm-Sidak post hoc test, n = 3 independent experiments performed in duplicates.

To prove that GRK4 is involved in the D₃R-induced mitogenic response in hPTCs, the expression of GRK4 was knocked down using GRK4-specific silencing siRNA. Previous experiments showed that siRNA treatment down-regulated the GRK4 mRNA expression by 65% and depleted the endogenous GRK4 by at least 70% (supplemental Fig. S2A). In control cells (untransfected cells and cells transfected with nonsilencing mock siRNA), D₃R stimulation resulted in increased mitogenesis compared with vehicle (127 ± 2.9 versus 100 ± 2.1% and 125.7 ± 4.8 versus 102.9 ± 1.6% in untransfected and mock siRNA-transfected cells, respectively) (Fig. 5C). In contrast, no mitogenic response was elicited in GRK4-depleted cells, even with D₃R stimulation (102.7 ± 7.7 versus 109 ± 5.4%), indicating that GRK4 is crucial to D₃R signaling in hPTCs.

To determine whether the MAP kinase pathway is involved in GRK4-mediated mitogenesis upon D₃R stimulation, hPTCs were treated with PD128907 at various time points, and the extent of MAP kinase phosphorylation was evaluated by immunoblotting. D₃R stimulation increased the amount of phosphorylated p44/42 MAP kinases as early as 5 min after agonist treatment, which leveled off at 30 min (supplemental Fig. S2B). In another set of experiments, PD128907 treatment for 1 h resulted in a 3-fold increase in phosphorylated p44/42 in nontransfected and mock siRNA-transfected control cells (100 ± 12.5 versus 277.8 ± 33.3% and 98.2 ± 13.1 versus 316.5 ± 35.0%, respectively, for phospho-p44; and 100 ± 10.5 versus 301.7 ± 30.4% and 110 ± 7.7 versus 333.6 ± 31.6%, respectively, for phospho-p42) and about a 2-fold increase in a downstream target of p44/42, p90RSK (100 ± 2.2 versus 173 ± 10.3% and 101 ± 3.5 versus 240.7 ± 11.7%) (Fig. 6). However, depletion of GRK4 resulted in the failure of p44/42 and p90RSK to be phosphorylated (88.74 ± 11.7 versus 94.7 ± 11.3%, 95.8 ± 22.3 versus 107.7 ± 16.5%, and 111.7 ± 5.9 versus 110 ± 4.8%, respectively). AKT and S6 ribosomal protein remained unphosphorylated upon D₃R activation.

FIGURE 6.

GRK4-dependent MAP k inase p hosphorylation. hPTCs transfected with GRK4-specific siRNA or nonsilencing siRNA (mock) were treated with the D₃R agonist PD128907 (1 μm, 1 h), and the total cell lysates were immunoblotted using a mixture of Abs that detects the phospho-p44/42 (Erk1/2) MAP kinases, phospho-p90RSK (a downstream target of p44/42), phospho-Akt (which is involved in cell survival and apoptosis), and phospho-S6 ribosomal protein (which is phosphorylated by p70 S6 kinase in response to growth factors and mitogens). elF4E detects the total target protein and is used for normalization. Nontransfected (control) cells served as an additional negative control. Nonspecific bands are seen above phospho-S6 and phospho-p44/42 and below phospho-Akt. Band densities were quantified by Scion densitometry software. Data are expressed as mean ± S.E. *, p < 0.05, versus others, one-way ANOVA, and Holm-Sidak post hoc test, n = 3 independent experiments.

DISCUSSION

Views on how cell membranes are organized and how they facilitate GPCR signal transduction initiated by an extracellular stimulus have radically changed in recent years. A plethora of biological and biophysical studies has shown the presence of cholesterol- and sphingolipid-enriched lipid rafts in the outer leaflet of the lipid bilayer that compose the plasma membrane. This unique lipid composition bestows greater order and less fluidity on the lipid rafts, making these rafts ideal for the establishment of high fidelity GPCR signal transduction networks. GPCR signaling requires at least three classes of proteins that must interact to allow the propagation of the signal and ensure that the signal leads to appropriate biological responses, i.e. the GPCR, the cognate G protein, and effectors/second messengers. Considering that GPCRs and effector proteins are expressed at relatively low concentrations in the mammalian cell (<10,000 for GPCRs and adenylyl cyclase) (37, 38), it is crucial that these proteins are found at the same microdomain to achieve effective concentrations for interaction (39). We now report that both D₃R and GRK4 co-localize in lipid rafts of human renal proximal tubule cells. Interestingly, both Gα_s and Gα_i, but not Gα_q (40), as well as several isoforms of adenylyl cyclase are targeted to the lipid raft in HEK-293 cells (29, 41). Thus, it appears that most of the proteins involved in D₃R signaling, e.g. D₃R, Gα_i, and Gα_s, adenylyl cyclase, and GRK4, co-segregate in lipid rafts, conceivably to form a cohesive network for selective D₃R activation and efficient propagation of the signal.

In the basal state, we found the D₃R both in the cytoplasm and at the plasma membrane of hPTCs. These findings are consistent with previous studies that showed endogenous D₃R at the subapical region and in the cytoplasm of proximal tubule cells, with the latter representing the recycling pool of D₃R (34), and heterologously expressed D₃R at the cell membrane in HEK-293 cells (42). We also report that in the rat kidney D₃R is predominantly expressed in proximal and distal tubules and, to a much lesser extent, in the glomeruli, in agreement with previous reports (32, 34). GRK4 is also distributed in both the cytoplasm and the cell membrane of hPTCs. GRK4 has been previously reported to associate with the cell membrane through palmitoylation of the cysteine residue at its C terminus (11). Agonist stimulation of D₃R results in the rapid internalization of both the receptor and GRK4 and their accumulation at the perinuclear area.

In the rat kidney, we found a similar pattern of distribution of both GRK4 and the D₃R, i.e. in proximal and distal tubules and the glomeruli, and hence co-localization is mostly observed in these structures. Although GRK4 expression was first thought to be confined mainly to the testis and the brain (11, 43), later reports showed that GRK4 isoforms, especially GRK4γ, are expressed in the human kidney (12, 44), especially in the proximal tubules (13), and in other tissues, such as the human myometrium (45).

Our data show that D₃R and GRK4 do not only co-localize but also physically interact in hPTCs at basal conditions and upon D₃R stimulation, as demonstrated by the co-immunoprecipitation assay. The lack of noticeable increase in the extent of co-immunoprecipitation upon agonist treatment may reflect the nature of the assay, which essentially removes subcellular compartmentalization, thus allowing the proteins to mix and interact freely. This interaction was confirmed by BiFC, a method that is based on the in situ formation of a fluorescent complex when two nonfluorescent fragments of a fluorophore are brought together by the interaction between the proteins tagged with the fragments (26, 27, 35). A faint BiFC signal is observed at the basal state, which markedly increased at 5 and 15 min after agonist activation in cells co-transfected with D₃R and GRK4. Moreover, the interaction between agonist-activated D₃R and GRK4 is initiated at the surface membrane and continues into the cytoplasm, specifically at the juxtanuclear area.

Several studies have evaluated the role of GRKs on D₃R function and signaling in the central nervous system. In particular, GRK2 and GRK3 have been reported to regulate the constitutive signaling activity of D₃R by controlling the stability of the filamin A-anchored signaling complex formed around D₃R (46). An increase in basal GRK2/3 activity reduces the interaction between D₃R, filamin A, and β-arrestin, which eventually leads to a dramatic reduction in D₃R-mediated G protein signaling and results in a fast feedback of short term fluctuations in dopamine concentrations in the central nervous system to the activity of D₃R autoreceptors (46). In HEK-293 cells, GRK3, but not GRK2, weakly phosphorylates the heterologously expressed, dopamine-stimulated D₃R (25).

We now show that D₃R is phosphorylated by GRK4 in proximal tubule cells. Of the four GRK4 splice variants described in humans (11), only GRK4-α, the full-length version, and GRK4-γ, which lacks the sequence encoded by exon 15 resulting in a 46-residue deletion near the C terminus, are able to phosphorylate D₃R upon agonist activation. GRK4-γ results in a 3-fold increase in D₃R phosphorylation, whereas GRK4-α leads to a 2-fold increase, indicating that both variants are important in D₃R signaling in hPTCs where all four GRK4 splice variants are expressed (44). The extent of GRK4-mediated D₃R phosphorylation is similar to that observed in agonist-activated D₁R by GRK4-α in HEK-293 cells (47).

The demonstration that GRK4 mediates the phosphorylation of agonist-activated D₃R in hPTCs offers insights into the critical role that GRK4 plays in renal dopamine signaling. GRK4 phosphorylates both D₁R and D₃R, the main renal receptors of the D₁-like and D₂-like dopamine receptor subfamilies. It is conceivable that the ability of GRK4 to target both receptors might help explain the observed functional synergism between activated D₁R and D₃R in both in vitro and in vivo studies. Co-expression of these receptors in HEK-293 cells increases the affinity of dopamine to D₁R compared with cells expressing the D₁R alone and promotes greater adenylyl cyclase stimulation (48). Unlike the other D₂-like receptors, activated D₃R weakly inhibits cAMP production, although it robustly inhibits adenylyl cyclase 5 (49), an isoform that is not expressed in the renal proximal tubule (50). In this nephron segment, D₃R stimulation may even result in increased cAMP production because D₃R may weakly activate adenylyl cyclase 2 (49), which is expressed in the renal proximal tubule (50). The stimulatory Gα_s, which is normally linked to D₁R and can decrease NHE3 activity by itself (51), and Gα_q/11, which is involved in the D₁-like receptor inhibition of Na⁺/K⁺-ATPase (52), can also couple to D₃R (5, 6, 53). D₁R and D₃R have been also shown to co-localize and co-immunoprecipitate in WKY rat renal proximal tubule cells (54). In the rat, the natriuretic response to pramipexole, a D₃R agonist, is attenuated by a D₁-like receptor agonist (55). Both D₁R and D₃R down-regulate the expression of the angiotensin II type 1 receptor (54, 56), which promotes salt retention.

Dopamine-mediated mitogenesis results from the activation of the D₂-like dopamine receptors. The mitogenic effect of activated D₃R has been demonstrated in neuroblastoma-glioma hybrid NG108-15 (57) and CHO cells (58) but not in opossum kidney cells (59); however, D₃R stimulation activates the ERK pathway, which is required for mitogenesis, in HEK-293, a human kidney cell line (60). In this study, we found that agonist activation of D₃R in hPTCs results in increased mitogenesis that was dependent upon GRK4 activity. The mitogenic response was specific to D₃R stimulation because co-treatment with a full D₃R antagonist abrogated the proliferative response.

The ability of GRKs to mediate D₃R signaling in hPTCs is confirmed by the use of a GRK inhibitor. Heparin has been shown to be a potent inhibitor of all seven GRK isoforms through competitive inhibition with the kinase and noncompetitive inhibition with ATP (30). Pretreatment with heparin results in the failure to elicit mitogenesis upon D₃R activation in hPTCs, indicating that GRK activity is important in promoting mitogenesis in these cells. To ascribe the mitogenic response to GRK4, GRK4-specific siRNA was used to knock down the expression of endogenous GRK4 in hPTCs. D₃R stimulation fails to promote mitogenesis in GRK4-depleted hPTCs, but not in mock siRNA-transfected and untransfected control cells, thus confirming the role of GRK4 in mediating D₃R-induced mitogenesis in these cells.

Classical G protein effectors such as phospholipase C and protein kinase C, which mediate heterologous desensitization, are not involved in the early signaling events that trigger the D₃R mitogenic response (61). D₃R-regulated mitogenesis is mediated by the activation of MAP kinases involved in cell proliferation (36). The stimulation of MAP kinases, e.g. ERK (62), by D₃R is mediated by Gα_i/Gα_o, phosphatidylinositol 3-kinase, atypical protein kinase C, and other signaling components like MEK, Ras, and Raf-1 but does not require phospholipase C, β-arrestin, or receptor sequestration (58). Gβγ has been shown to be involved in D₃R-mediated MAP kinase activation in HEK-293 cells (60); however, activation of this receptor does not result in a surfeit of Gβγ subunits in these cells (63), suggesting that this may not be the major pathway by which D₃R promotes cell proliferation. Interestingly, dopamine, which targets all of the dopamine receptor subtypes, has been shown to increase the phosphorylation of the p44/42 MAP kinases in primary renal proximal tubule cells (64). We now show that PD128907, a highly selective D₃R agonist, increases the phosphorylation of p44/42 MAP kinases, as well as their downstream target p90RSK in a time-dependent manner, and that siRNA-mediated depletion of GRK4 abrogates this response in hPTCs. Taken together, our data indicate that GRK4 is required for D₃R-mediated mitogenic response, which is signaled through the MAP kinase pathway.

An alternative pathway by which D₂-like receptors elicit mitogenesis is through the transactivation of a receptor tyrosine kinase (RTK) (36), which effectively recruits the RTK cascade in response to dopamine. In HEK-293 and CHO cells, agonist-activated D₃R transactivates the epidermal growth factor receptor (60), an RTK. This transactivation may be dependent upon a direct interaction between the GPCR and RTK (65–67). We hypothesize that the GPCR-RTK interaction may be impaired in the absence of GPCR phosphorylation, e.g. D₃R phosphorylation by GRK4γ and GRK4α in hPTCs. Previous studies on angiotensin II type 1 receptor and V2 vasopressin receptors, members of the GPCR family, showed the requirement for receptor phosphorylation by GRK5 and GRK6, members of the GRK4 subfamily, to promote β-arrestin-mediated ERK activation (68, 69). Overexpression of GRK5/6 increased the β-arrestin-mediated ERK activation (69). The authors have speculated that GRK5/6-mediated phosphorylation of the receptor might induce the receptor to adopt a conformation that favors ERK activation via β-arrestin (69). Interestingly, we have reported in a preliminary study that GRK4γ is needed for the proper orientation of D₁R in plasma membrane microdomains (70) and may be responsible for the recruitment of cytoplasmic D₁R to the plasma membrane in the early phase of D₁R stimulation (71, 72). This may also be true for D₃R. The requirement for GRK4, in its capacity to phosphorylate and/or properly position the receptor in the membrane, may explain the inhibitory effect of heparin or of decreased GRK4 expression on D₃R-induced mitogenesis and also opens new views into the role of receptor phosphorylation on signal transduction. However, persistently increased constitutive GRK4 activity results in desensitization of D₁R, seen in human essential hypertension (1, 13, 17). Whether a similar state occurs with the D₃R remains to be determined.

In summary, D₃R and GRK4 are found in both surface membrane and the cytoplasm of hPTCs under basal conditions. In surface membranes, both proteins co-segregate in non-lipid and, more importantly from a functional standpoint, in lipid rafts where they are presumably in close association with other proteins involved in D₃R signaling. D₃R and GRK4 interact in vitro (hPTCs) and in vivo (proximal and distal tubules and the glomerulus of WKY rat kidney). The enhanced interaction following agonist activation of the receptor is initiated at the surface membrane and continues into the cytoplasm. GRK4-γ and GRK4-α promote a 3- and 2-fold increase in agonist-induced D₃R phosphorylation, respectively. Finally, GRK4 is required for D₃R-mediated mitogenesis and activation of the MAP kinase pathway in hPTCs.