|Home | About | Journals | Submit | Contact Us | Français|
In moderate sodium-replete states, dopamine 1–like receptors (D1R/D5R) are responsible for regulating >50% of renal sodium excretion. This is partly mediated by internalization and inactivation of NaKATPase, when associated with adapter protein 2. We used dopaminergic stimulation via fenoldopam (D1-like receptor agonist) to study the interaction among D1-like receptors, caveolin-1 (CAV1), and the G protein– coupled receptor kinase type 4 in cultured human renal proximal tubule cells (RPTCs). We compared 2 groups of RPTCs, 1 of cell lines that were isolated from normal subjects (nRPTCs) and a second group of cell lines that have D1-like receptors that are uncoupled (uncoupled RPTCs) from adenylyl cyclase second messengers. In nRPTCs, fenoldopam increased the plasma membrane expression of D1R (10.0-fold) and CAV1 (1.3-fold) and markedly decreased G protein– coupled receptor kinase type 4 by 94±8%; no effects were seen in uncoupled RPTCs. Fenoldopam also increased the association of adapter protein 2 and NaKATPase by 53±9% in nRPTCs but not in uncoupled RPTCs. When CAV1 expression was reduced by 86.0±8.5% using small interfering RNA, restimulation of the D1-like receptors with fenoldopam in nRPTCs resulted in only a 7±9% increase in association between adapter protein 2 and NaKATPase. Basal CAV1 expression and association with G protein– coupled receptor kinase type 4 was decreased in uncoupled RPTCs (58±5% decrease in association) relative to nRPTCs. We conclude that the scaffolding protein CAV1 is necessary for the association of D1-like receptors with G protein– coupled receptor kinase type 4 and the adapter protein 2-associated reduction in plasma membrane NaKATPase.
The mechanisms by which renal dopamine, in concert with other sodium regulator pathways, regulates >50% of renal sodium excretion are not completely understood. Intrarenal administration of a dopamine 1 (D1)-like receptor antagonist (SCH23390) decreased sodium excretion by 57% in conscious uninephrectomized dogs1,2 and in anesthetized volume-loaded rats.3,4 Disruption of D1-like receptor (D1R or D5R) genes in mice increases salt-sensitive blood pressure.5 Dopamine and D1-like receptor agonists are natriuretic in experimental animals6 and humans,7,8 and ecopipam, a selective D1-like receptor antagonist, increases blood pressure in humans9 (presumably by preventing sodium excretion).
Dopamine inhibits sodium transport in several segments of the nephron, causing increased sodium excretion.5 In the renal proximal tubule (RPT), dopamine inhibits the sodium/hydrogen exchanger 3 (NHE3)10,11 and NaKATPase.12,13 The decrease in sodium/hydrogen exchanger 310 and NaKATPase14 from the plasma membrane involves binding to adapter protein 2 (AP2), followed by endocytosis. Dopamine stimulation and/or sodium concentration associate AP2 with NaKATPase, cause NaKATPase translocation from the plasma membrane into early and late endosomes, and decrease rubidium uptake, an index of NaKATPase activity in opossum and rodent cells.15-18 However, the effect of dopamine on NaKATPase internalization, activity, and association with AP2 has not been shown in human RPT cells (RPTCs). There may be species differences in the regulation of sodium transport. For example, in humans, the α subunit of NaKATPase expressed in the kidney is more sensitive to the inhibition by ouabain compared with rodents.19,20
Caveolin, especially caveolin-1 (CAV1), has been shown to be important in the NaKATPase internalization process.21,22 Caveolins, localized in lipid rafts of plasma membranes, tether and regulate signaling complexes into functional units (eg, G protein– coupled receptors).23,24 We have reported that fenoldopam (FEN; D1-like receptor agonist) stimulates caveolin 2 (CAV2) and D1R association in human embryonic kidney 293 cells, which do not express CAV1.25 Furthermore, FEN-stimulated membrane D1R recruitment and cAMP accumulation were greater in membranes with CAV2-containing lipid rafts than in those depleted of CAV2 with antisense oligonucleotides or in nonlipid rafts without CAV2.26
G protein– coupled receptor kinase 4 (GRK4) is 1 of 7 members of the GRK family and has been shown to be specifically located in RPTCs, where it constitutively phosphorylates and desensitizes D1R in specific plasma membrane microdomains.27,28 However, the role of the interaction among GRK4, CAV1, and D1-like receptors and NaKATPase internalization in human RPTCs has not been described. We tested the hypothesis that CAV1, interacting with GRK4, is necessary for D1R-mediated inhibition of NaKATPase activity in human RPTCs through binding with AP2. Because D1R signaling is impaired in hypertensive humans and rats,6,8,23,29,30 this interaction may be aberrant in essential hypertension.
We have generated RPTCs from institutional review board–approved normal human subjects (nRPTCs) and RPTCs from subjects in which D1R is uncoupled from adenylyl cyclase stimulation (uRPTCs)28,31-33 (see details in the online Data Supplement Methods section), as described previously.27,28,31,34-36
cAMP was measured both by a commercial ELISA (Cayman Chemical) and an intracellular real-time kinetic assay, which involves transfecting RPTCs with a plasmid containing a novel fluorescence resonance energy transfer (FRET) cAMP sensor (ICUE3) according to the method of Violin et al.37 Details of both methods are outlined in the online Data Supplement.
Detergent-free plasma membrane sheet isolation was performed using sulfo-NHS-SS-biotin, as reported previously.35 Total cell lysates (20 μg; 4°C) were loaded per lane for electrophoresis and immunoblotting.35 The proteins of interest were detected using rabbit polyclonal antibodies to CAV1 (1:500 dilution; BD Biosciences), D1R,26 and GRK4 (1:200 dilution; Santa Cruz; sc-13079), followed by a goat antirabbit infrared dye (IR Dye 800; Li-Cor) secondary antibody and imaged on an Odyssey infrared imaging system (Li-Cor).
GRK4 and CAV1 association was measured by coimmunoprecipitation of 1 mg of cellular protein in 1 mL of lysis buffer (M-PER; Pierce) and 2 μg of CAV1 monoclonal antibody (BD Biosciences). Detection used rabbit polyclonal anti-GRK4 (1:200 dilution; Santa Cruz; sc-13079), followed by goat antirabbit IR Dye 800 secondary antibody. Details are provided in the online Data Supplement.
CAV1 small interfering RNA (siRNA; target sequence: 5′CCGCATCAACTTGCAGAAA3′ and scrambled control: 5′CCGAACTGTTCGACACAAA3′)38 was designed and ordered prehybridized (Sigma-Genosys). The most effective GRK4 siRNA (target sequence: 5′AATACAAAGAGAAAGTCAA3′ and scrambled control: 5′AGAAGATAAGAACAATAAC3′) was chosen among 10 candidate target sequences by Western blotting. Details of transfection are provided in the online Data Supplement.
Measurement of NaKATPase internalization with AP2 was adapted from previously published techniques,15,13 which showed that the internalization of and decrease in NaKATPase activity are associated with AP2. Monensin was shown to be necessary to measure the effect of dopamine.17,18 Details are provided in the online Data Supplement.
Plasma membrane–localized NaKATPase expression was performed on RPTCs using similar treatments as for the NaKATPase experiments, with subsequent fixation and staining with a monoclonal antibody to NaKATPase. Details are outlined in the online Data Supplement.
We have established a sodium efflux assay that measures NaKATPase activity in cultured cells as ouabain-sensitive reduction in sodium export. Details are included in the online Data Supplement.
The data are expressed as mean±SE. Comparisons within and among groups were made by repeated-measures or factorial ANOVA, respectively, followed by Holm-Sidak or Duncan test. A t test was used for 2-group comparisons. A P<0.05 was considered significant.
Our cell lines were well characterized for their RPT origin.27,28 We measured cAMP in 2 nRPTC and 2 uRPTC cell lines, determined by the response to fenoldopam (FEN; 1 μmol/L; 30 minutes) or dimethyl sulfoxide (DMSO) vehicle control. We compared our ELISA with a novel FRET–based method (Figure 1). Figure 1A shows the >2-fold increase in cAMP accumulation (ELISA) induced by FEN in nRPTCs (but not in uRPTCs) that was blocked by the D1-like receptor antagonist, LE300 (10 μmol/L), indicating that the stimulatory effect was via D1-like receptors. A cAMP FRET biosensor, ICUE3,37 showed that D1-like stimulation with FEN caused a significant rise in the intracellular cAMP level in nRPTCs, which reached a plateau by 7 minutes (Figure 1B). Minimal effect was seen in uRPTCs, confirming our previous reports with older methods.28,32
We next studied the plasma membrane expression of D1R, GRK4, and CAV1 in nRPTCs and uRPTCs using Western blotting techniques (Figure 2). Figure 2A shows that D1R plasma membrane recruitment was increased (10.73±1.70-fold; P<0.05; n=4) over DMSO vehicle (VEH) control after FEN (1 μmol/L; 30 minutes) stimulation in nRPTCs. In uRPTCs, there was no significant increase in FEN-induced D1R recruitment to the plasma membrane. Basal levels of plasma membrane GRK4 expression were lower in uRPTCs than in nRPTCs (58.4 8% decrease; P<0.05; n=4). Figure 2B shows that FEN stimulation decreased plasma membrane GRK4 abundance in nRPTCs (94±8.0; P<0.05; n=4) but had no effect in uRPTCs. CAV1 basal plasma membrane expression was 40.0% (±7.0%; P<0.05; n=4) lower in uRPTCs than in nRPTCs and increased (1.3±0.09-fold; P<0.05; n=4) after FEN stimulation only in nRPTCs (Figure 2C). Figure 2D shows β-tubulin and Ponceau S loading controls.
To prove that the reduced recruitment of D1R to the plasma membrane in uRPTCs with FEN stimulation was because of GRK4, we silenced GRK4 gene expression with GRK4 siRNA (supplemental Figure S1, available in the online Data Supplement). The addition of GRK4-specific siRNA reduces GRK4 expression by 70.6±5.6 (P<0.05; n=3) by Western blotting (Figure S1A). We have reported that, in uRPTCs, GRK4 is constitutively active, and the D1R is phosphorylated, desensitized, and internalized.31,34 When siRNA to GRK4 was added to uRPTCs and compared with scrambled siRNA (SCR) and DMSO VEH controls, the FEN-mediated (1 μmol/L for 30 minutes) D1R plasma membrane expression was partially restored (P<0.05; n=6; Figure S1B).
The total cellular CAV1 expression level was increased by 47±16% (P<0.05; n=8) with FEN stimulation (1 μmol/L; 4 hours) when compared with DMSO VEH (Figure 3). Angiotensin (ANG) II stimulation (10 nM; 4 hours) reduced CAV1 expression by 67±8% (P<0.05; n=8).
Because CAV1 is a known negative regulator of GRKs,39 we studied the association of GRK4 and CAV1 in nRPTCs and uRPTCs with coimmunoprecipitation experiments. CAV1 was immunoprecipitated using a monoclonal anti-CAV1 antibody and immunoblotted with GRK4 rabbit polyclonal antibody (Figure 4). A 62-kDa immunoreactive band was observed corresponding with the predicted size of GRK4. CAV1/GRK4 coimmunoprecipitation was 58±5% less in uRPTCs compared with nRPTCs (P<0.05, t test; n=4 per group).
The inverse of this experiment is shown in Figure S2, where we immunoprecipitated GRK4 and detected CAV1 and GRK4. There was a 55.6±14.1% decrease in uRPTCs compared with nRPTCs (P<0.05; n=4), further validating the difference in GRK4/CAV1 association between the cell groups.
Treatment with β-methyl cyclodextrin (βMCD; 2 mmol/L; 1 hour), an agent that reduces cholesterol and CAV1 in the plasma membrane, decreased CAV1 membrane expression in nRPTCs (52±4%; P<0.05 versus PBS VEH, nRPTCs; ANOVA, Holm-Sidak test, n=8) but not in uRPTCs (Figure 5). The depletion of plasma membrane CAV1 by βMCD reduced CAV1 expression in nRPTCs to the basal levels seen in uRPTCs. Basal plasma membrane expression of CAV1 was lower in uRPTCs by 56±11% (P<0.05; n=8) relative to nRPTCs.
We used another approach to reduce CAV1 expression. siRNA to CAV1 reduced total cellular CAV1 expression by 86±8.5% (P<0.05, t test; n=8) in nRPTCs and 89±4% (P<0.05, t test; n=8) in uRPTCs compared with SCRtransfected control cells (Figure S3). Total cellular expression of CAV1 was once again lower (25.2±3.0%; P<0.05; n=8) in uRPTCs than in nRPTCs.
Internalization of NaKATPase was measured by immunoprecipitating AP2 with a rabbit polyclonal antibody and Western blotting with NaKATPase and AP2 monoclonal antibodies, producing single bands of the correct molecular weight for both AP2 and NaKATPase (Figure S4). Because there was only a single band for both proteins, further analysis was continued by immuno-dot blot (Figure 6). FEN stimulated the AP2-mediated NaKATPase internalization in nRPTCs (□), a response that was not seen in uRPTCs (). In nRPTCs, there was a 53 9% (P<0.05; n=6) increase in association between AP2 and NaKATPase with FEN (1 μmol/L for 30 minutes). This effect was completely blocked by βMCD or siRNA to CAV1. Control SCR had no effect on FEN-stimulated internalization of NaKATPase in nRPTCs. In uRPTCs, FEN had no effect on AP2-NaKATPase coimmunoprecipitation and was not affected by CAV1 siRNA or CAV1 SCR. Other controls used were βMCD alone, CAV1 siRNA alone, CAV SCR alone, or treatment without monensin; these controls had no effect on the basal association of AP2 and NaKATPase (Figure S5). Controls without monensin showed no change in cell-surface NaKATPase expression after FEN stimulation in nRPTCs in Figure S6. Thus, an increase in intracellular sodium concentration is necessary for NaKATPase internalization.
Confocal images of NaKATPase displayed a predominantly plasma membrane expression pattern in control (VEH) cells, which was reduced after FEN stimulation when an increase in cytoplasmic accumulation was seen (P<0.05, t test; n=3; Figure S7). uRPTCs displayed an identical expression pattern as control nRPTCs but displayed no measurable internalization on FEN stimulation (data not shown).
We measured NaKATPase activity as ouabain-inhibitable (P<0.01; n=3 versus VEH) sodium efflux (Figure S8). FEN reduced the sodium efflux in nRPTCs by 50.2±4.6% versus VEH (P<0.01; n=3), which was CAV1 dependent, because it was blocked by βMCD (P<0.01; n=3). FEN had no effect in uRPTCs.
These current studies demonstrate a CAV1 dependence on the D1R-like receptor coupling to NaKATPase in human RPTCs. The current study, along with our previous report,26 indicates that regulation of the D1R by CAV1 and/or CAV2 may depend on the surrounding microenvironment (lipids and associating receptors) in human RPTCs versus human embryonic kidney 293 cells.
CAV1 can also directly interact with GRKs 1, 2, and 5 in A431, NIH-3T3, and COS-1 cells,39 but a CAV1 interaction with GRK4 has not been reported. In human embryonic kidney 293 cells, GRK2 physically interacts with CAV226,39-41; CAV1 is not expressed in human embryonic kidney 293 cells.22 Thus, our demonstration of GRK4 coimmunoprecipitation with CAV1 is also novel, although not surprising. The lower GRK4/CAV1 association in uRPTCs is in keeping with the negative modulatory role of CAV1 on GRK4; GRK4 is constitutively active in uRPTCs.27
Our experiments demonstrate that CAV1 is also involved with GRK4 in the proper recycling, resensitization, and desensitization of the D1R.42 In contrast, in uRPTCs, basal membrane expression of CAV1 is decreased compared with nRPTCs. In uRPTCs, short-term FEN stimulation does not increase CAV1 plasma membrane protein expression, GRK4 plasma membrane protein does not change, and the D1R is not recruited to the plasma membrane. These studies suggest that genetic defects in GRK4 that interrupt its association with CAV1 lead to desensitization and impaired function of the D1R. The role of ANG II in this process is currently unknown, although it is possible that the decreased expression of CAV1 may be attributed to the lack of negative regulation of the uncoupled D1R on the ANG II type 1 receptor.43 The expression of the ANG II type 1 receptor is not different in the nRPTC and uRPTC cell lines examined (data not shown). However, we have demonstrated that ANG II type 1 receptor expression and function are increased in mice overexpressing human GRK4 142V and 486V genes, relative to mice overexpressing human GRK4 wild-type genes.44,45
In genetic hypertension, the ability of dopamine and D1-like receptor agonists to increase sodium excretion is impaired. RPTs from hypertensive rats have an impaired ability to decrease sodium transport (eg, sodium/hydrogen exchanger 3,46-49 NaHCO3 cotransporter,50 Cl/HCO3 exchanger,51 and NaKATPase6,52-56). Dopamine inhibits sodium transporter activity, in part by increasing its internalization, an action that is also impaired in hypertension.16,57 An increased activity of NaKATPase in RPTs has been shown in several rodent genetic models of hypertension.56,58 Our studies presented here in human RPTCs demonstrate that the scaffolding protein CAV1 is a mediator in the association of the D1R with its kinase, GRK4, and its role in AP2-associated reduction in plasma membrane NaKATPase. Ultimately, the transregulation of CAV1, D1R, and GRK4 will influence sodium transport. We are studying NaKATPase and sodium/ hydrogen exchanger 3–mediated Na+ flux, which will be reported in a future article.
There is a growing appreciation for the role of the dopaminergic system in regulating sodium balance and, hence, blood pressure. Genetic variants in a protein, GRK4, have been linked to essential hypertension and salt sensitivity in both humans and rodent models for these conditions. In these studies, we showed that CAV1, a principal membrane scaffolding protein, is involved in the organization of the D1R-like receptor signaling pathway and intracellular sodium. These studies further decipher the mechanisms of how dopamine, GRK4, and CAV1 control the sodium pump. The ubiquitous expression of CAV1 may limit its usefulness as a potential therapeutic target. However, selective targeting of GRK4, which regulates only a few G protein– coupled receptors in a limited number of organs, may be a new useful antihypertensive therapeutic approach.
We thank Helen E. McGrath for editorial assistance with this article.
Sources of Funding
This work was supported by National Institutes of Health grants HL074940, HL23081, DK39308, HL68686, and HL092196.
R.A.F. and P.A.J. were awarded a US Patent (No. 6 660 474) on “GRK Mutants in Essential Hypertension,” which has been assigned to Hypogen, Inc.