Phosphorylation of the Catalyic α-Subunit Constitutes a Triggering Signal for Na+,K+-ATPase Endocytosis*

Inhibition of Na+,K+-ATPase activity by dopamine is an important mechanism by which renal tubules modulate urine sodium excretion during a high salt diet. However, the molecular mechanisms of this regulation are not clearly understood. Inhibition of Na+,K+-ATPase activity in response to dopamine is associated with endocytosis of its α- and β-subunits, an effect that is protein kinase C-dependent. In this study we used isolated proximal tubule cells and a cell line derived from opossum kidney and demonstrate that dopamine-induced endocytosis of Na+,K+-ATPase and inhibition of its activity were accompanied by phosphorylation of the α-subunit. Inhibition of both the enzyme activity and its phosphorylation were blocked by the protein kinase C inhibitor bisindolylmaleimide. The early time dependence of these processes suggests a causal link between phosphorylation and inhibition of enzyme activity. However, after 10 min of dopamine incubation, the α-subunit was no longer phosphorylated, whereas enzyme activity remained inhibited due to its removal from the plasma membrane. Dephosphorylation occurred in the late endosomal compartment. To further examine whether phosphorylation was a prerequisite for subunit endocytosis, we used the opossum kidney cell line transfected with the rodent α-subunit cDNA. Treatment of this cell line with dopamine resulted in phosphorylation and endocytosis of the α-subunit with a concomitant decrease in Na+,K+-ATPase activity. In contrast, none of these effects were observed in cells transfected with the rodent α-subunit that lacks the putative protein kinase C-phosphorylation sites (Ser11 and Ser18). Our results support the hypothesis that protein kinase C-dependent phosphorylation of the α-subunit is essential for Na+,K+-ATPase endocytosis and that both events are responsible for the decreased enzyme activity in response to dopamine.

The natriuretic effect of dopamine (DA) 1 depends on its abil-ity to increase the glomerular filtration rate and/or to modulate directly tubular sodium reabsorption (1)(2)(3). Changes in vectorial transport of sodium induced by DA in renal tubules are largely mediated by inhibition of Na ϩ ,K ϩ -ATPase (4,5) and Na ϩ /H ϩ -exchanger activity (6). At the cellular level, DA triggers a specific signaling cascade that ultimately activates protein kinase C (PKC) (7), a process postulated to be responsible for the decreased Na ϩ ,K ϩ -ATPase activity.
Activators of PKC, such as phorbol esters and diacylglycerol analogs, decrease Na ϩ ,K ϩ -ATPase activity in isolated rat renal PCT segments (7,8) as well as the vectorial transport of sodium by isolated perfused PCTs (9). In isolated renal PCT cells, another activator of PKC, L-1-oleoyl-2-acetoyl-sn,n-acetoylglycerol, decreased Na ϩ ,K ϩ -ATPase activity determined as the rate of ouabain-sensitive oxygen consumption (10). In a renal cell line derived from opossum kidney (OK cells), but not from pig kidney (LLC-PK 1 cells), incubation with phorbol esters resulted in phosphorylation of the Na ϩ ,K ϩ -ATPase ␣-subunit and inhibition of its activity (11). However, stimulation of Na ϩ ,K ϩ -ATPase by phorbol esters has also been reported (12,13). Although phosphorylation of the ␣-subunit by PKC in a cell-free preparation was associated with a decrease in enzymatic activity (14 -16), it is not clear whether this effect occurs in intact cells in response to phorbol esters (11,17).
DA is produced locally in renal PCT cells (18 -20) where its synthesis is regulated physiologically during ingestion of a high salt diet (21). Contrary to the diverse effects of PKC stimulation by phorbol esters and diacylglycerols on Na ϩ ,K ϩ -ATPase activity, there is a consensus on the inhibitory action of DA on the enzyme. Moreover, we have recently demonstrated that inhibition of PCT Na ϩ ,K ϩ -ATPase activity by DA is associated with endocytosis of its ␣and ␤-subunits into early-(EE) and late (LE) endosomes via a clathrin-coated vesicle (CCV)-dependent mechanism (22). Nevertheless, despite the information gained during the last few years on the regulation of Na ϩ ,K ϩ -ATPase activity, it is not known whether inhibition of enzyme activity in intact cells depends on the phosphorylation of the catalytic subunit, or whether such phosphorylation is necessary for subunit endocytosis in response to a physiologic agonist such as DA.
In the present study, using intact renal PCT cells metabolically labeled with [ 32 P]orthophosphate, we have examined whether dopamine phosphorylates the Na ϩ ,K ϩ -ATPase ␣-subunit and whether this effect is responsible for the decreased enzymatic activity and subunit endocytosis.

EXPERIMENTAL PROCEDURES
Materials-The cAMP analog Rp-cAMPS was obtained from BioLog, Bremen, Germany. Bisindolylmaleimide was purchased from Calbiochem, San Diego, CA. All other chemicals were from Sigma. A monoclonal antibody, kindly provided by Dr. M. Caplan (Yale University), was used against the Na ϩ ,K ϩ -ATPase ␣-subunit (antibody A, which recognizes only the N-terminal first five residues of the ␣-subunit). Immunoprecipitation of the Na ϩ ,K ϩ -ATPase in the phosphorylation experiments and Western blots were performed using a polyclonal antibody (B) raised against the rat Na ϩ ,K ϩ -ATPase ␣-subunit (23). The identity of EE was determined with a polyclonal antibody raised against a rab5 synthetic peptide (Santa Cruz Biotechnology Inc., Santa Cruz, CA). The late endosome fraction was identified with a mannose-6-phosphate receptor antibody (courtesy of Dr. B. Hoflack, EMBL, Heidelberg, Germany).
Preparation of PCT Cells-PCT cells were prepared as described before (10,24). Briefly, male Sprague-Dawley rats (BK Universal, Sollentuna, Sweden) weighing between 150 -200 g were used. After the kidneys were removed and the cortex isolated, the tissue was minced on ice to a paste-like consistency. The cortical minceate was incubated with 0.7 mg/ml collagenase (Type I, Sigma) in 50 ml of Hanks' medium (Life Technologies,Inc., Gaithersburg, MD). The incubation was carried out at 37°C for 60 min, where the solution was continuously exposed to 95% O 2 , 5% CO 2 and was terminated by placing the tissue on ice and pouring through graded sieves (180 -75-53-38 M pore size) to obtain a cell suspension. It has been reported that phorbol esters regulate Na,K-ATPase differently depending on whether the tissue has been continously oxygenated during its preparation and incubation with the PKC activator. Although this effect was not reported to be involved in the DA response, we had taken the precaution to incubate cells in oxygenated solutions in all steps until the tissue was disrupted to immunoprecipitate the ␣-subunit or for preparation of BLM, EE, and LE.
Cell Culture and Transfection-The expression vector pCMV containing the rodent Na ϩ ,K ϩ -ATPase ␣ 1 -subunit cDNA was obtained from PharMingen (San Diego, CA). Preparation of the expression vector (myc/1.32) that encodes a shortened mutant of the ␣ 1 -subunit was as described by Shanbaky and Pressley (25). This vector expresses a rodent ␣-subunit in which the first 31 amino acids of the nascent polypeptide are replaced by an initiation methionine and a sequence of 10 amino acids (EQKLISEEDL) from the human c-myc oncogene product. Transfection of OK cells and selection of ouabain-resistant colonies were performed as described previously (26).
Determination of Na ϩ ,K ϩ -ATPase Activity-Cells were incubated in modified Hanks' medium in the presence or absence of 1 M DA at room temperature for different periods of time. The incubation was terminated by placing the samples on ice. Cell aliquots (approximately 10 -20 g of protein) were transferred to the Na ϩ ,K ϩ -ATPase assay medium (final volume of 100 l) containing in mM NaCl, 50; KCl, 5; MgCl 2 , 10; EGTA, 1; Tris-HCl, 50; Na 2 ATP, 7 (Calbiochem, La Jolla, CA); and [␥-32 P]ATP (NEN Life Science Products, specific activity 3000 Ci/mmol) in tracer amounts (3.3 nCi/l). Na ϩ ,K ϩ -ATPase activity was determined in permeabilized cells as described before (27,28).
Preparation of Endosomes-Cells were labeled with 32 P as detailed above. Cells in suspension (1.5 mg of protein/ml) were incubated under different protocols at room temperature. Incubation was terminated by transferring the samples to ice and adding cold homogenization buffer containing 250 mM sucrose and 3 mM imidazole, 2 mM EGTA, 10 mM NaF, 30 mM Na 4 O 7 P 2 , 1 mM Na 3 VO 4 , 1 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, 4 g/ml aprotinin, pH 7.4. Cells were gently homogenized (15-20 strokes) to minimize damage of the endosomes, using a Dounce homogenizer, and the samples were subjected to a brief (5 min) centrifugation (4°C, 3,000 ϫ g). Endosomes were fractionated on a flotation gradient as described (22), using essentially the technique of Gorvel et al. (30).
Preparation of Basolateral Plasma Membranes-After separation of EE and LE, another fraction (500 l) was collected at the 16 and 42% sucrose interface corresponding to cell ghosts, mitochondria, and plasma membranes. Basolateral membranes (BLM) were further purified according to Hammond et al. (31), using a Percoll gradient. Briefly, the collected material was diluted by adding 500 l of imidazole (3 mM, pH 7.4) buffer containing protease inhibitors (final sucrose concentration 25-26% w/w), and spun at 20,000 ϫ g for 20 min. The yellow layer was resuspended again in the supernatant (carefully removed from the brown pellet containing mitochondria and cell ghosts) and centrifuged at 48,000 ϫ g for 30 min. The supernatant was discarded, and the pellet was resuspended in 1 ml of buffer (300 mM mannitol and 12 mM HEPES, pH 7.6, adjusted with Tris) by gentle pipetting. To form a Percoll gradient, 0.19 g of undiluted Percoll (Pharmacia Biotech Inc.) was added to a 1-ml suspension (0.2-1 mg of protein). The suspension was gently mixed and centrifuged at 48,000 ϫ g for 30 min, and the ring of BLM was collected.
Miscellaneous-Protein content was determined according to Bradford (32). Western blots were developed with an ECL (Amersham, UK) detection kit. Scans were performed using a ScanJet IIc scanner (Hewlett Packard, Palo Alto, CA). Quantitation of the phosphorylated Na ϩ ,K ϩ -ATPase ␣-subunit was performed using a Fuji Bas 1000 Bioimaging analyzer (Fuji, Japan), and the data (arbitrary units) were analyzed using Tina 2.07 ray test software (Isotopenmessyerä te GmbH, Staulenhardt, Germany).
Statistics-Comparison between two experimental groups were made by the unpaired Student's t test. For multiple comparisons, oneway ANOVA with Sheffe's correction was used. p Ͻ 0.05 was considered significant.

RESULTS
In this study we sought to determine whether inhibition of Na ϩ ,K ϩ -ATPase activity and endocytosis was associated with phosphorylation of the ␣-subunit. In isolated renal PCT cells, incubation with DA decreased Na ϩ ,K ϩ -ATPase activity (nmol P i /mg prot/min, vehicle: 112 Ϯ 8 versus DA, 1 M: 60 Ϯ 2, n ϭ 4, p Ͻ 0.05), and this effect was blocked by PKC inhibitors (7,8). Intact renal PCT cells were metabolically labeled with 32 P and thereafter incubated for 2.5 min at room temperature with or without DA (Fig. 1A). The ␣-subunit was immunoprecipitated, separated by SDS-PAGE, and transferred to polyvinylidene difluoride membranes. In every experiment the amount of radioactivity (autoradiography or phosphoimager) incorporated into the ␣-subunit was corrected for the amount of protein present (Western blot), and the quantitative data are shown as percent of control at the bottom of each panel. DA increased the state of phosphorylation (to ϳ165% of control) of the ␣-subunit, as illustrated in Fig. 1A. This increased phosphorylation was inhibited by bisindolylmaleimide, a specific PKC inhibitor, but not by a cAMP-dependent protein kinase (PKA) inhibitor, suggesting that phosphorylation of the ␣-subunit induced by DA is mediated by PKC. Neither inhibitor affected the state of ␣-subunit phosphorylation in non-stimulated PCT cells.
Phosphorylation of the Na ϩ ,K ϩ -ATPase ␣-subunit by DA was time-dependent (Fig. 1B). It increased significantly after 1 min and was maximal at 2.5 min (178% of control), whereas it was no longer evident at 10 min. However, while the initial (1.0 and 2.5 min) increase in ␣-subunit phosphorylation corresponded to the decrease in enzyme activity, this correlation was no longer present at 10 min, i.e. enzyme activity remained inhibited (percent of control, 60 Ϯ 3, p Ͻ 0.05), whereas phosphorylation was similar to that of control cells.
To determine whether the Na ϩ ,K ϩ -ATPase has been dephosphorylated, we examined the effect of DA in the presence of a phosphatase inhibitor, 1 M okadaic acid (OKD) (Fig. 1C). Basal phosphorylation (resting condition ϭ control, C) was moderately higher (ϳ1.5-fold) in the OKD-treated cells. As hypothesized, in OKD-treated cells, DA (10 min) did increase the state of ␣-subunit phosphorylation, suggesting that at this time period it had been dephosphorylated by the action of protein phosphatases.
Because after 10 min the ␣-subunit has been dephosphorylated yet the decreased enzymatic activity persisted, it is possible that the dephosphorylated ␣-subunits no longer reside in the plasma membrane. To test this hypothesis, we evaluated the state of phosphorylation of the ␣-subunit in BLM and in LE. In BLM prepared from cells that have been preincubated with DA for 10 min, the state of phosphorylation of the immunoprecipitated ␣-subunit remained unchanged regardless of whether the PCT cells were previously treated with 1 M OKD or not, whereas it increased significantly in LE. This was evident, however, only if PCT cells had been preincubated with OKD (phosphorylation (percent of control): 104 Ϯ 6 without OKD versus 126 Ϯ 5 OKD-treated cells, n ϭ 3), supporting the notion that the ␣-subunit is dephosphorylated in LE.
In BLM prepared from cells that have been preincubated with DA for 2.5 min ( Fig. 2A, left panel) the state of phosphorylation of the immunoprecipitated ␣-subunit was significantly increased, regardless of whether the PCT cells were previously treated with 1 M OKD or not (phosphorylation (percent of control): 130 Ϯ 2.0 without OKD, 131 Ϯ 1.2 with OKD). These results suggest that the phosphorylated subunits (2.5 min) in the BLM are not affected by protein phosphatases Although the results described above support the concept that in response to DA the Na ϩ ,K ϩ -ATPase ␣-subunits are phosphorylated in the plasma membrane and then internalized and dephosphorylated in LE, the link between these two processes (i.e. whether phosphorylation is a requisite for endocytosis) is not clear. Therefore, we next used an epithelial cell line from OK transfected with the rat Na ϩ ,K ϩ -ATPase ␣-subunit cDNA carrying a deletion in the nascent 28 amino acids in which Ser 11 and Ser 18 , the putative phosphorylation sites for PKC (33,34), are absent. OK cells (non-transfected) behaved similarly to native PCT cells in their response to DA: DA decreased the Na ϩ ,K ϩ -ATPase activity, and this inhibition was associated with endocytosis of the ␣-subunit. Thus, they constitute a useful model to study the mechanisms of action of DA.
The relative expression of Na ϩ ,K ϩ -ATPase ␣-subunits in transfected OK cells was determined using antibody A raised against the first five amino acids of the ␣-subunit (which should not recognize the truncated form, OK␣ rat-t ) and compared with antibody B, raised against the holoenzyme. While antibody B recognized the Na ϩ ,K ϩ -ATPase ␣-subunit from both the full-length (OK␣ rat ) and OK␣ rat-t cells, antibody A detected only a slight presence of Na ϩ ,K ϩ -ATPase ␣-subunits in OK␣ rat-t (Fig. 3A), indicating that in OK␣ rat-t most of the ␣-subunits correspond to the truncated form.
We next evaluated the Na ϩ ,K ϩ -ATPase activity and its re- sponse to DA in OK␣ rat and OK␣ rat-t (Fig. 3B). While basal Na ϩ ,K ϩ -ATPase activity was similar in both groups of cells and comparable with that in earlier reports (13,26), incubation with DA resulted in a significant decrease in Na ϩ ,K ϩ -ATPase activity from OK␣ rat (p Ͻ 0.01), but not from OK␣ rat-t (p ϭ 0.567). The inhibitory effect of DA in OK␣ rat was abolished by coincubation with a PKC inhibitor, bisindolylmaleimide (percent of control: 99.3 Ϯ 7, n ϭ 3). We further examined whether this inhibition was associated with phosphorylation of the ␣-subunit (Fig. 3C). 1 M DA (3 min; room temperature) increased the state of phosphorylation of the ␣-subunit in OK␣ rat but not in OK␣ rat-t cells.
Last, to determine whether phosphorylation of the ␣-subunit was necessary for endocytosis, early and late endosomes were prepared from OK␣ rat and OK␣ rat-t cells incubated with DA (Fig. 4). DA stimulated the incorporation of ␣-subunits into EE and LE from OK␣ rat , and this effect was blocked by bisindolylmaleimide (percent of control, EE: 105 Ϯ 12, n ϭ 3; and LE: 96 Ϯ 17, n ϭ 3) or calphostin C (percent of control, EE: 98 Ϯ 13, n ϭ 3; and LE: 86 Ϯ 11, n ϭ 3). However, DA did not increase the incorporation of ␣-subunits from OK␣ rat-t . DISCUSSION In this report we have demonstrated that DA treatment of both isolated proximal tubule cells and OK cells transfected with the rodent ␣-subunit leads to inhibition of Na ϩ ,K ϩ -ATPase activity and phosphorylation and endocytosis of the ␣-subunit. In contrast, when the DA effect was examined in OK cells expressing the Na ϩ ,K ϩ -ATPase ␣-subunit isoform in which the putative PKC-phosphorylation sites were removed, DA-treatment neither inhibited the enzyme activity nor induced any significant phosphorylation or endocytosis of the ␣-subunit. These observations strongly suggest a causal link between PKC-dependent phosphorylation of amino acids at the ␣-subunit N terminus and Na ϩ ,K ϩ -ATPase inhibition and en-docytosis in response to a physiological agonist.
Inhibition of Na ϩ ,K ϩ -ATPase activity by DA in renal PCT involves the sequential activation of arachidonic acid, 20-HETE, and PKC (35). Although cAMP stimulation has been suggested to contribute to the action of DA (36,37), it is unlikely that it would be directly involved in Na ϩ ,K ϩ -ATPase regulation (phosphorylation of the ␣-subunit in renal PCT cells) because increased cAMP in this segment does not inhibit (7) but is rather associated with stimulation of Na ϩ ,K ϩ -ATPase activity (38). Accordingly, in this study phosphorylation of Na ϩ ,K ϩ -ATPase ␣-subunits was blocked by PKC-, but not cAMP-K, inhibition. Our observation differs from that reported by Beguin et al. (39), perhaps reflecting differences in the preparations used. We examined isolated PCT cells, where DA is synthetized and physiologically regulates Na ϩ ,K ϩ -ATPase activity, whereas Beguin et al. (39) used a reconstituted system in which the receptor (human dopaminergic DA 1A ) and the target (Bufo marinus Na ϩ ,K ϩ -ATPase ␣-subunit) were expressed in a cell line (COS-7) that normally does not express this regulatory system.
The present results suggest that inhibition of total cell Na ϩ ,K ϩ -ATPase activity is initially accomplished by phosphorylation of the ␣-subunit and that the activity remains decreased because the inhibited units no longer reside in the plasma membrane. Once the ␣-subunits become phosphoryl- ated, they are internalized by sequential translocation into CCV, EE, and finally LE, where they may be dephosphorylated. Because in CCV and EE the increased Na ϩ ,K ϩ -ATPase ␣-subunit abundance is not associated with increased enzymatic activity (22), it is unlikely that it could have been dephosphorylated in these compartments.
Endocytosis of the ␣-subunit requires phosphorylation by PKC because mutants lacking the PKC phosphorylation sites do not internalize in response to dopamine. The mechanisms by which membrane proteins are internalized have been studied extensively, and the consensus sequences of interaction with adaptins have been described. The amino acids that were deleted by the truncation do not bear any homology with known endocytic sequences. Thus, it appears that it is the lack of phosphorylation rather than impairment of a putative Na ϩ ,K ϩ -ATPase ␣-subunit-adaptin interaction that precludes the dopamine-dependent endocytosis in cells transfected with the truncated isoform. However, Beron et al. (40), using phorbol esters to stimulate PKC, have recently postulated that PKC activation is not necessary for Na ϩ ,K ϩ -ATPase ␣-subunit endocytosis in A6 cells. These observations, however, are not comparable with the effect of DA in PCT cells. Phorbol esters increased fluid phase endocytosis, and the signal could thus have occurred at any other target in the plasma membrane. In PCT cells, by contrast, DA selectively internalized the Na ϩ ,K ϩ -ATPase ␣/␤-subunits while the distribution of other basolateral membrane markers such as the glucose transporter GLUT-2 and the mannose 6-phosphate receptor remained unchanged (22). Incubation of PCT cells with phorbol esters, on the other hand, resulted in internalization of the GLUT-2 transporter as well as the Na ϩ ,K ϩ -ATPase ␣-subunit and also induced a significant change in the actin cytoskeleton organization. 2 Thus, endocytosis and phosphorylation of the Na ϩ ,K ϩ -ATPase ␣-subunit, as well as inhibition of its activity in response to DA (Ref. 22 and present study), require activation of PKC.
It has also been reported that phorbol esters stimulate Na ϩ ,K ϩ -ATPase activity (12,13,26) and that this effect is accompanied by phosphorylation of the ␣-subunit (12). Thus, although both effects (that of DA and of phorbol esters) share a common target, PKC, they are clearly different. For example, stimulation by phorbol esters of Na ϩ ,K ϩ -ATPase activity and phosphorylation of the ␣-subunit were significant after 15 min of incubation (12). In contrast, the effect of DA occurs already at 1 min, and after 10 min, the ␣-subunits are no longer phosphorylated and, in addition, they no longer reside in the plasma membrane. Finally, another reason why the effects of phorbol esters and DA are different in nature may be that the effect of DA on Na ϩ ,K ϩ -ATPase activity is mediated via a PKC isoform that can be activated by arachidonic acid metabolism and generation (in the PCT) of the cytochrome P-450 metabolite, 20-HETE, an eicosanoid that activates PKC (41). The action of DA might therefore involve an atypical PKC isoform that is not responsive to phorbol esters (42) but whose activation is rather dependent on membrane lipids.
In conclusion, while in intact cells the use of phorbol esters has not been proved to be an efficient probe to demostrate the relationship between phosphorylation of the Na ϩ ,K ϩ -ATPase ␣-subunit and inhibition of its activity (17), by using an agonist such as dopamine in cells where it is produced and exerts its physiologic action it was possible to demonstrate that phosphorylation of the ␣-subunit is associated with inhibition of Na ϩ ,K ϩ -ATPase activity and that this step is required for subunit endocytosis.