Regulation of Na/K-ATPase expression by cholesterol: isoform specificity and the molecular mechanism

Abstract

We have reported that the reduction in plasma membrane cholesterol could decrease cellular Na/K-ATPase α1-expression through a Src-dependent pathway. However, it is unclear whether cholesterol could regulate other Na/K-ATPase α-isoforms and the molecular mechanisms of this regulation are not fully understood. Here we used cells expressing different Na/K-ATPase α isoforms and found that membrane cholesterol reduction by U18666A decreased expression of the α1-isoform but not the α2- or α3-isoform. Imaging analyses showed the cellular redistribution of α1 and α3 but not α2. Moreover, U18666A led to redistribution of α1 to late endosomes/lysosomes, while the proteasome inhibitor blocked α1-reduction by U18666A. These results suggest that the regulation of the Na/K-ATPase α-subunit by cholesterol is isoform specific and α1 is unique in this regulation through the endocytosis-proteasome pathway. Mechanistically, loss-of-Src binding mutation of A425P in α1 lost its capacity for regulation by cholesterol. Meanwhile, gain-of-Src binding mutations in α2 partially restored the regulation. Furthermore, through studies in caveolin-1 knockdown cells, as well as subcellular distribution studies in cell lines with different α-isoforms, we found that Na/K-ATPase, Src, and caveolin-1 worked together for the cholesterol regulation. Taken together, these new findings reveal that the putative Src-binding domain and the intact Na/K-ATPase/Src/caveolin-1 complex are indispensable for the isoform-specific regulation of Na/K-ATPase by cholesterol.

INTRODUCTION

The Na/K-ATPase was discovered by Jens C. Skou in 1957 (45). It is an integral membrane protein responsible for translocating sodium and potassium ions across the cell membrane against electrochemical gradients by hydrolyzing ATP (28). The functional Na/K-ATPase consists of α-, β-, and γ-subunits. The α-subunit (around 112 kDa) is a catalytic subunit that contains binding sites for ATP, cation, and other ligands (26). So far, four distinct isoforms of the α-subunit have been identified with different tissue-specific expression patterns (42, 44). The α1-isoform is ubiquitously expressed. The α2-isoform is found in skeletal muscles, cardiac myocytes, adipocytes, and brain. The α3 is mainly expressed in the neuronal cells and α4 in testis (5, 19, 25, 33, 55). These isoforms differ in the structure, kinetic properties, and protein interactions (4, 43). The Na/K-ATPase is also known as a signal transducer that binds with stimuli such as ouabain to activate multiple protein kinase cascades involving Src kinase, epidermal growth factor receptors, and mitogen-activated protein kinase and facilitates reactive oxygen species production (51, 52). The signaling Na/K-ATPase is implicated in the cell functions under physiological and pathological processes (3, 24, 30, 37, 38, 50). Moreover, it has been demonstrated that Na/K-ATPase and Src form a functional receptor to regulate downstream signaling pathways (16, 23, 27, 47). To study the effects of different Na/K-ATPase α-isoforms, our laboratory developed α-isoform-specific expressing mammalian cell lines. The cell line that expresses only α1 is able to stimulate Src kinase-dependent signaling, but α2- or α3-cell lines fail to do so (27, 34, 49). This implies that the signaling mechanism of Na/K-ATPase is isoform specific and the molecular heterogeneity of Na/K-ATPase isoforms may be physiologically relevant.

Cholesterol accounts for 20% of the total lipids in mammalian cell membranes. Na/K-ATPase is also located at the plasma membrane and ubiquitously expressed in all mammalian cells. Our laboratory has recently revealed a novel reciprocal regulation between Na/K-ATPase and cholesterol. On the one hand, Na/K-ATPase α1 is involved in controlling cellular cholesterol distribution (7). Downregulation of Na/K-ATPase α1 redistributes cholesterol from the plasma membrane to cytosolic compartments in α1-knockdown cells (PY-17) and α1^+/−-mice and also alters cholesterol metabolism in vivo. On the other hand, cholesterol also plays an important role in the regulation of Na/K-ATPase. Reduction in plasma membrane cholesterol in pig kidney LLC-PK1 cells could activate Src kinase by phosphorylation at tyrosine 418 and stimulate the endocytosis and degradation of α1 through a Src-dependent pathway and consequently decrease Na/K-ATPase α1-expression. In the brains and livers of a mice model with intracellular cholesterol trafficking defects, the expression of α1 was similarly reduced (9). Caveolin-1 (Cav-1) is the molecular marker of caveolae, which is plasma membrane invagination concentrated with many signaling proteins including Na/K-ATPase (48). Interestingly, the NH₂-terminal cytosolic tail of Na/K-ATPase α1 also interacts with Cav-1 in caveolae and regulates Cav-1 trafficking (6). Furthermore, Src-mediated Cav-1 phosphorylation promotes swelling and release of caveolae (15, 56). It appears that Na/K-ATPase α1, cholesterol, Src, and Cav-1 work together as a signaling complex.

To further understand the mechanism of how different Na/K-ATPase isoforms are differentially regulated, we tested whether membrane cholesterol depletion affected the expression of α-isoforms in different isoform-expressing cells, Src-binding mutant cells, and Cav-1 knockdown cells. Our new findings indicate that the cholesterol-regulated expression of the Na/K-ATPase α-subunit is isoform specific through Src and Cav-1. The intact Na/K-ATPase/Src/Cav-1 complex and the NaKtide sequence [S415 to Q434, which is responsible for Src binding (26)] in the nucleotide-binding domain, especially A425, are indispensable for this regulation. These results indicate the uniqueness and importance of α1 to hold the signaling complex together and may advance our understanding of why cholesterol dysregulation could impose different effects on different organs or tissues.

MATERIALS AND METHODS

Materials.

Dulbecco’s modified Eagle’s medium (DMEM) with l-glutamine, 4.5 g/L glucose, and HEPES was purchased from Sigma-Aldrich (St. Louis, MO). Fetal bovine serum was purchased from Atlanta Biologicals and trypsin from Corning (Flowery Branch, GA). The mouse monoclonal anti-Na/K-ATPase α1-antibody for Western blot analysis was obtained from the Developmental Studies Hybridoma Bank at the University of Iowa (cat. no. α6F, RRID: AB_528092, dilution 1:1,000). The mouse monoclonal anti-Na/K-ATPase α1-antibody for immunocytochemistry was purchased from Upstate Biotechnology, Inc. (Lake Placid, NY, cat. no. 05-369, RRID: AB_309699, dilution 1:100). The rabbit polyclonal anti-Na/K-ATPase α2 antibody (cat. no. AB9094, RRID: AB_347648, dilution 1:1,000 for Western blot and 1:200 for immunocytochemistry) and the rabbit polyclonal anti-Na/K-ATPase α3-antibody (cat. no. 06-172, RRID: AB_310066, dilution 1:1,000 for Western blot and 1:200 for immunocytochemistry) were from Millipore (Billerica, MA). The rabbit monoclonal anti-caveolin antibody was obtained from Cell Signaling (Danvers, MA, cat. no. 3267, RRID: AB_2275453, dilution 1:1,000). Mouse anti-α-tubulin was from Sigma-Aldrich (cat. no. T5168, RRID: AB_477579, dilution 1:10,000). The goat anti-rabbit (cat. no. sc-2004, RRID: AB_631746, dilution 1:1,000) and goat anti-mouse (cat. no. sc-2005, RRID: AB_631736, dilution 1:1,000) secondary antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Alexa Fluor 555-conjugated anti-mouse IgG (cat. no. A-31570, RRID: AB_2536180, dilution 1:100) and anti-rabbit IgG (cat. no. A-21428, RRID: AB_2535849, dilution 1:100) were obtained from Life Technologies (Eugene, OR). The U18666A compound was from Cayman Chemical (Ann Arbor, MI). Filipin was purchased from Sigma-Aldrich. The Amplex Red cholesterol assay kit was from Molecular Probes, Inc. (Eugene, OR).

Cell lines and cell culture.

Different cell lines used in this study were generated as previously described and summarized in Table 1 (23, 27, 29, 34, 49, 53). Briefly, pig kidney epithelial cell line (LLC-PK1) was purchased from the American Type Culture Collection (Manassas, VA). PY-17 cells are Na/K-ATPase α1-knockdown cell line derived from LLC-PK1 by the siRNA method. The wild-type rat α1-rescued cells (AAC-19), α2-rescued cells (LX-α2-4), and α3-rescued cells (LM-α3-1) were generated from PY-17 cells transfected with rat α1-, α2- and α3-expressing vectors, respectively. The α1-mutant cells (A425P) were from PY-17 cells by substituting an amino acid at alanine 425 in rat α1. The α2-mutant cells (LY-a2) were from PY-17 cells transfected with the mutant α2-vector, which substituted four amino acids in rat α2- with the corresponding α1-amino acids. Cav-1 knockdown cells (C2-9) were derived from LLC-PK1 by utilizing Cav-1-specific siRNA. Mycoplasma-free cells (passages 8–13) were cultured in DMEM with 10% fetal bovine serum, 100 units/ml penicillin, and 100 μg/ml streptomycin in a 5% CO₂ humidified incubator. When cells reached ∼90% confluence, they were serum starved for 24 h before treatment.

Table 1.

List of cell lines included in this study and summary of their features

Cell Line	Na/K-ATPase Species and Isoform	Src Binding Capacity in Na/K-ATPase	Cav-1 Expression Maintenance
LLC-PK1	Pig α1	+	+
PY-17	Pig α1 (α1-knockdown)	−	−
AAC-19	Rat α1 (wild type)	+	+
LX-α2-4	Rat α2 (wild type)	−	−
LM-α3-1	Rat α3 (wild type)	−	+
LY-a2	Rat α2 (mutant)	+	+
A425P	Rat α1 (mutant)	−	−
C2-9	Pig α1 (Cav-1 knockdown)	−	−

Western blot analysis.

Cells were washed with phosphate-buffered saline (PBS) and solubilized in ice-cold modified radioimmune precipitation buffer (1% Nonidet P-40, 1% sodium deoxycholate, 150 mM NaCl, 1 mM EDTA, and 150 mM Tris·HCl, pH 7.4) with protease inhibitor cocktail and phosphatase inhibitors from Sigma-Aldrich. The obtained cell lysates were then centrifuged at 14,000 g for 15 min, and supernatants were collected for protein assay. Following denaturation by warming for 20 min or boiling for 5 min, cell lysates containing equal amounts of protein were loaded onto the SDS-PAGE gel for separation, then transferred to nitrocellulose blotting membrane (GE Healthcare Life Sciences, Germany), and probed with indicated antibodies. The signals were detected with chemiluminescent substrate from Thermo Fisher (Rockford, IL) and quantified by ImageJ software.

Immunofluorescence staining and confocal microscopy.

Cells were grown on the uncoated glass coverslips (Fisherbrand, Pittsburgh, PA) until 90% confluence, serum starved for 24 h, and then treated with U18666A compound (10 μg/ml) for 24 h or 48 h. The cells were fixed with ice-cold methanol for 20 min, blocked with 4% BSA from Sigma for 30 min, and then incubated with specific primary antibodies overnight at 4°C. After three washes, Alexa Fluor 555-conjugated anti-mouse/rabbit secondary antibodies were added and incubated for 1 h at room temperature. Cells on coverslips were washed three times, mounted with Vectashield mounting medium, and imaged using a Leica TCSII-SP5 laser scanning microscope (Leica, Mannheim, Germany). The images were then analyzed and quantified using the ImageJ software. Specifically, immunofluorescence from outside and within 1 μm of the cell surface is selected and considered as total cellular signals, and immunofluorescence internal and away from 1 μm of the cell surface is as intracellular region. The ratio between intracellular and total signals was calculated as an indicator of endocytosis as described previously (9).

Cell surface biotinylation.

Cells were cultured in 60-mm-treated polystyrene dishes (Denville Scientific, Metuchen, NJ) until 90% confluence and washed three times with ice-cold PBS containing 1 mM EDTA. Cells were then incubated with 2.5 ml biotinylation buffer (10 mM triethanolamine, pH 9.0 and 250 mM sucrose) containing 2.5 mg/ml sulfo-NHS-SS-biotin (Thermo Fisher) for 1 h on ice with shaking. To quench unreacted biotin, cells were rinsed three times with PBS-EDTA containing 100 mM glycine for 15 min on ice. Cells were then washed twice with PBS and solubilized in 200 μl lysis buffer (150 mM NaCl, 50 mM Tris·HCl, 5 mM EDTA, 1% Triton X-100, and protease inhibitor, pH 7.5). Cell lysates were rotated for 15 min and then centrifuged at 17,000 g for 10 min at 4°C. After determination of protein concentration, the solution containing 100 μg of protein and 35 μl of streptavidin agarose beads (Thermo Fisher, Rockford, IL) in lysis buffer were rotated overnight at 4°C. The supernatant was removed after centrifuge at 17,000 g for 1 min, and the beads were washed four times with PBS. The beads were incubated with 35 μl of Laemmli loading buffer for 30 min at 55°C to elute protein. The eluted protein was then subjected to SDS-PAGE gel, transferred to nitrocellulose membrane, and probed with indicated antibodies.

Cellular fractionation.

Subcellular fractions were obtained via sucrose gradient fractionation as previously described (7). Briefly, cells were washed twice with ice-cold PBS and collected in 2 ml 500 mM sodium carbonate (pH 11.0) solution. The cell lysates were homogenized using a Polytron tissue homogenizer (three 6-s bursts) and then sonicated (three 40-s bursts). Two milliliters of 90% sucrose in MBS (25 mM MES and 150 mM NaCl, pH 6.5) were mixed with cell homogenates to final 45% sucrose content. The suspension was placed in the bottom of ultracentrifuge tubes, loaded with 4 ml of 35% sucrose and then 4 ml of 5% sucrose (both in MBS with 250 mM sodium carbonate). The samples were centrifuged at 39,000 rpm for 17 h in a SW41 rotor (Beckman Instruments). Twelve 1-ml gradient fractions (numbered from top to bottom) were collected. Fractions 4 and 5 were combined and diluted with 4 ml of MBS and centrifuged at 40,000 rpm in type 65 rotor (Beckman Instruments) for 1 h. The pellets were resuspended in 250 μl of MBS and considered as the caveolae-enriched fraction.

Plasmid constructs and transfection.

RFP-Rab7 plasmid was purchased from Addgene (Cambridge, MA). When cells reached 80% confluence, they were transfected 6 h with the plasmid using Lipofectamine 2000 as described previously (9).

Free cholesterol staining assay.

The distribution of free cholesterol was performed by Filipin staining as described previously (9). Briefly, cells were fixed with 4% paraformaldehyde in PBS for 30 min, quenched with 50 mM NH₄Cl in PBS for 10 min, and then permeabilized with 0.1% saponin in PBS for 10 min. Cells were then blocked in PBS with 1 mg/ml filipin (Sigma-Aldrich), 0.2% BSA, and 0.2% fish skin gelatin (Sigma-Aldrich) for 30 min. After three washes with PBS, cells were mounted with Vectashield mounting medium and imaged with a Leica TCSII-SP5 laser scanning microscope (Leica, Mannheim, Germany).

Mice and diets.

Wild-type male 129 SvJ/Black Swiss mice (25–30 g) were used for all experiments, as previously described in detail (22a). Mice were fed a high-fat diet (HFD) (36% weight/weight-adjusted calories from fat, Bio-Serv S3282) for 20 wk while controls were maintained on a normal chow (NC) diet (Harlan Teklad, TD 2918). All studies were approved by the Cleveland Clinic Institutional Animal Care and Use Committee.

Histology and immunohistochemistry.

Kidneys were fixed in 4% formaldehyde (pH 7.2), paraffin embedded, and cut into 4-μm sections. The tissue sections were deparaffinized with xylene and rehydrated by sequential incubations in ethanol and water. A Vectastain Elite-ABC kit (Vector Laboratories) (Burlingame, California) was used following manufacturer's protocol. For Na/K-ATPase histology analysis, 10 images were randomly taken with a bright-field microscope with a ×20 lens and quantitative morphometric analysis was performed using automated and customized algorithms/scripts for batch analysis (ImageIQ, Inc., Cleveland, OH) written for Image Pro Plus 7.0 as described in detail (22b). For Na/K-ATPase count, 10 images were randomly taken with a bright-field microscope with a ×20 lens, distinct and diffused staining was counted and averaged for each kidney. For Na/K-ATPase area, we included both diffused and distinct staining of Na/K-ATPase in the kidney.

RNA isolation and reverse transcription-quantitative PCR.

RNA extraction, cDNA preparation, and reverse transcription-quantitative PCR were all performed utilizing the Qiagen (Germantown, MD) automated workflow system, which utilizes the QIAcube HT and QIAgility liquid handling robots. RNA from kidney tissue was isolated utilizing QIAzol/chloroform extraction methodology via automated liquid handling equipment (QIAcube HT). Approximately 500 ng of extracted RNA were used to synthesize cDNA (Qiagen, RT2 First Strand Kit, cat. no. 330404). RT-PCR was performed utilizing QIAGEN’s Rotor-Gene Q thermocycler. Calculation of gene expression was conducted by comparing the relative change in cycle threshold value (ΔCt). Fold change in expression was calculated using the 2−ΔΔCt equation as previously described (22a).

Statistical analysis.

Data are presented as means ± SE of at least three independent experiments. Statistical analysis was performed using the Student’s t test for comparison between two groups or one-way ANOVA for comparison among more than two groups. The significance was accepted at P < 0.05. All statistical analyses were performed using GraphPad Prism 6 software.

RESULTS

Regulation of cellular Na/K-ATPase α-subunit by membrane cholesterol is isoform specific.

Compound U18666A, an amphipathic steroid, is widely used to specifically block the intracellular trafficking of cholesterol. It inhibits the egress of free cholesterol from late endosomes and lysosomes by binding to a transport protein named Niemann-Pick C1 (NPC1) which helps cholesterol leave the lysosome (32). We previously reported that U18666A led to reduction in the plasma membrane cholesterol and consequently plasma membrane levels of α1-subunit of Na/K-ATPase (9). To confirm the regulation of Na/K-ATPase α1 by cholesterol reduction, we first tested the effect of U18666A on total cellular content (plasm membrane and intracellular compartments) of α1-subunit in LLC-PK1 and AAC-19 cells. LLC-PK1 cells are pig kidney epithelial cells that express only Na/K-ATPase α1. To reduce the interference from endogenous Na/K-ATPase α1, we rescued α1-knockdown PY-17 cells with rat α1 and generated stable cell lines (AAC-19) (Table 1) (27). AAC-19 cells expressed over 95% of exogenous α1 with no signal of other Na/K-ATPase isoforms (34, 49). As depicted in Fig. 1A, a 30% and 50% downregulation of Na/K-ATPase α1 was recorded in LLC-PK1 after U18666A treatment at 10 μg/ml for 24 h and 48 h, respectively, consistent with the previous report (9). In AAC-19, a time- and dose-dependent downregulation of α1 was also observed after U18666A treatment (Fig. 1, B and C). To investigate the regulation of other α-isoforms, we generated rat α2-rescued PY-17 cells (LX-α2-4) and rat α3-rescued PY-17 cells (LM-α3) with no detectable level of α1 expression (Table 1) (34, 49). As depicted in Fig. 1, D–F, U18666A failed to change the total expression of α2 or α3 after U18666A treatment. These results indicate that U18666A specifically regulates the total expression of Na/K-ATPase α1 but not α2 or α3.

Fig. 1.

U18666A decreases total Na/K-ATPase α1-expression but does not alter total α2- or α3-expression. A: LLC-PK1 cells were serum starved and treated with 10 µg/ml U18666A for 24 h and 48 h. Total cell lysates were prepared, and Na/K-ATPase α1-expression was analyzed by Western blot. B and C: AAC-19 cells were serum starved and treated with 10 µg/ml U18666A for 24 h and 48 h (B) and different doses for 24 h (C). Total cell lysates were prepared, and Na/K-ATPase α1 expression was analyzed by Western blot. D and E: LX-α2-4 cells were serum starved and treated with 10 µg/ml U18666A for 24 h and 48 h (D) and different doses for 24 h (E). Total cell lysates were prepared, Na/K-ATPase α2-expression was analyzed by Western blot. F: LM-α3 cells were serum starved and treated with 10 µg/ml U18666A for 24 h and 48 h. Total cell lysates were prepared, and Na/K-ATPase α3 expression was analyzed by Western blot. The signals were quantified using the ImageJ software and calculated. The results were normalized to α-tubulin. The quantitative data are presented as means ± SE and n = 3–4 in each group. **P < 0.01, ***P < 0.001, and ****P < 0.0001, compared with control based on the unpaired t test (A, B, D, and F) and one-way ANOVA with Dunnett’s multiple-comparisons test (C and E).

To check the distribution of cholesterol in different α-isoform-expressing cells, we detected free cholesterol in AAC-19, LX-α2-4 and LM-α3 cells after U18666A treatment using filipin, a fluorescent chemical compound that binds to unesterified cholesterol in cells. After exposure with U18666A for 24 h, the weaker free cholesterol signals in the plasma membrane and stronger signals in the perinuclear region were detected in all of the different isoform cells compared with the control (Fig. 2). These data indicate that free cholesterol redistributes from the plasma membrane to intracellular compartments after U18666A treatment. It further supports the notion that it is the reduction of plasma membrane cholesterol that downregulates the total expression of the Na/K-ATPase α1 as previously reported (9) but not α2 or α3. In other words, the regulation of cellular Na/K-ATPase α by plasma membrane cholesterol is isoform specific.

Fig. 2.

U18666A decreases plasma membrane cholesterol in AAC-19, LX-α2-4, and LM-α3 cells. AAC-19, LX-α2-4, and LM-α3 cells were serum starved, treated with 10 µg/ml U18666A for 24 h, and then fixed and stained with filipin to detect free cholesterol. Representative fluorescent images are shown. Scale bar = 10 µm.

To further determine the effect of cholesterol on the cellular localization of Na/K-ATPase α-isoforms, AAC-19, LX-α2-4 and LM-α3 cells were immunostained using the α-isoform-specific antibodies. While the majority of the Na/K-ATPase α1 resided in the plasma membrane in AAC-19 control cells, some α1-isoforms appeared in the intracellular compartments after U18666A exposure for 24 h (Fig. 3A). However, U18666A did not induce the intracellular distribution of α2-isoforms in LX-α2-4 cells as depicted in Fig. 3B. Interestingly, we also observed α3-isoforms redistributed into intracellular compartments in LM-α3 cells without change in total expression (Fig. 3C). Taken together, the confocal imaging results reveal that the reduction in the plasma membrane cholesterol may induce the endocytosis of Na/K-ATPase α1 and α3 but not α2.

Fig. 3.

Membrane cholesterol reduction decreases plasma membrane Na/K-ATPase α1 and α3 but does not change cellular localization of α2. AAC-19 (A), LX-α2-4 (B), and LM-α3 (C) cells were serum starved, treated with 10 µg/ml U18666A for 24 h, and then fixed for immunostaining of Na/K-ATPase isoforms. Representative fluorescent images for cellular distribution of Na/K-ATPase α-isoforms in the presence or absence of U18666A are shown. Scale bar = 10 µm. The signals were quantified using the ImageJ software, and the ratio between intracellular and total signals was calculated. Results are presented as box and whisker plots. The number of the cells used for quantification was 20–30 from 3 separate experiments. ****P < 0.0001, compared with control based on the unpaired t test.

To provide further evidence that U18666A stimulates endocytosis of α1, we measured the plasma membrane expression of α1 by cell surface biotinylation. The surface expression of α1 was reduced by ∼40% after U18666A treatment in AAC-19 cells (Fig. 4A). The similar change in total expression and distribution of α1 suggests that U18666A stimulates the endocytosis and then degradation of the Na/K-ATPase as we reported previously (9). In contrast, no change in surface expression of α2 suggests that U18666A did not induce the endocytosis of α2 (Fig. 4B). Interestingly, the surface expression of α3 also decreased in LM-α3 cells which is compatible with the immunostaining result (Figs. 3C and 4C). However, the total expression level of α3 was not altered (Fig. 1F). This may suggest that Na/K-ATPase α3 was endocytosed but not degraded.

Fig. 4.

Membrane cholesterol reduction decreases surface expression of Na/K-ATPase α1 and α3 but does not alter α2. AAC-19 (A), LX-α2-4 (B), and LM-α3 (C) cells were serum starved, treated with 10 µg/ml U18666A for 24 h, and then biotinylated as described in materials and methods. The same amount of bound fraction was analyzed by Western blot with anti-Na/K-ATPase α1-, α2-, α3-primary antibodies, respectively. Representative Western blots are shown for surface and total expression of Na/K-ATPase α. The signals were quantified using the ImageJ software and calculated. The quantitative data are presented as means ± SE and n = 4 in each group. ***P < 0.001, compared with control based on the unpaired t test.

To directly demonstrate that α1 traffics to the late endosome/lysosomes after U18666A treatment, we transfected AAC-19 cells with late endosome/lysosomes marker RFP-tagged Rab7. As depicted in Fig. 5A, α1 accumulated in intracellular compartments and was colocalized with the Rab7-positive vesicles after U18666A treatment, while the majority of α1 resided in the plasma membrane in control cells. To further confirm that α1 is degraded in the lysosomes, we added the ubiquitin-proteasome inhibitor (MG-132) into U18666A-treated AAC-19 cells. As depicted in Fig. 5B, MG-132 blocked U18666A-induced downregulation of α1. Similarly, to further test whether U18666A induces the accumulation of α3 in late endosome/lysosomes, we also transfected LM-α3 cells with Rab7. Although the majority of α3 accumulated in intracellular compartments, no colocalized signal between α3 and Rab7 was observed in U18666A-treated LM-α3 cells (Fig. 5C). Our data strongly suggest that U18666A stimulates the endocytosis of Na/K-ATPase α1 and α3 but then degradation of only α1, not α3. Taken together, the regulation of Na/K-ATPase α by plasma membrane cholesterol is isoform specific and α1 is unique in the regulation through endocytosis pathway.

Fig. 5.

Membrane cholesterol reduction leads to endocytosis of both Na/K-ATPase α1 and α3 but degradation of only Na/K-ATPase α1. AAC-19 (A) and LM-α3 (C) cells were transfected with RFP-Rab7 for 6 h and then treated with 10 µg/ml U18666A for 24 h. Afterward, cells were fixed and immunostained with anti-Na/K-ATPase α1- and α3-antibody, respectively. Representative fluorescent images are shown for Na/K-ATPase α1/α3-staining, RFP-Rab7, and merged figures. The arrows point to the colocalization of α1 and RFP-Rab7. Scale bar = 10 µm. B: AAC-19 cells were serum starved and treated with 10 µg/ml U18666A for 24 h in the presence or absence of 20 µM MG-132. Total cell lysates were prepared, and Na/K-ATPase α1-expression was analyzed by Western blot. The signals were quantified using the ImageJ software and calculated. The results were normalized to α-tubulin. The quantitative data are presented as means ± SE and n = 3 in each group. *P < 0.05, compared with control based on one-way ANOVA with Dunnett’s multiple-comparisons test.

To investigate the changes in Na/K-ATPase expression and distribution by dysregulation of cholesterol metabolism in vivo, we examined kidneys from normal chow (NC)-fed and high-fat diet-fed (HFD) mice. As depicted in Supplemental Fig. S1A (all Supplemental material is available at https://doi.org/10.6084/m9.figshare.11921085.v2) quantitative RT-PCR revealed no difference in Na/K-ATPase α1 mRNA levels in kidney sections between mice on NC and HFD. Immunohistochemistry studies showed no obvious difference in Na/K-ATPase α1-expression in distinct tubular basolateral membrane but displayed a more diffused α1-staining pattern in HFD fed mice (Supplemental Fig. S1B). Taken together, although we did not detect a difference in total cellular Na/K-ATPase α1, we found the redistribution of Na/K-ATPase α1 from the plasma membrane to cytoplasm in kidneys from HFD fed mice. It suggests that endocytosis of Na/K-ATPase α1 may be associated with abnormal cholesterol metabolism in vivo.

Regulation of cellular Na/K-ATPase α-subunit by membrane cholesterol requires its interaction with Src.

Our laboratory has identified NaKtide sequence (S415 to Q434) in the nucleotide-binding domain, which inhibits Src activity (26). We previously generated a stable cell line expressing A425P mutant in the NaKtide sequence of Na/K-ATPase α1 and determined that the mutant α1 retained normal pumping function but defective in interacting and regulating Src (Table 1) (23). To further probe the Src-related molecular mechanism of the regulation of Na/K-ATPase expression by cholesterol, we checked the expression of mutant α1 after U18666A treatment. As depicted in Fig. 6A, A425P total mutant α1 appeared to be resistant to U18666A, similar to α2. This finding indicates that the A425P mutant fully blocks the effect of cholesterol depletion on total Na/K-ATPase α1 expression. As depicted in Fig. 6B, there was also no change in surface expression of mutant A425P α1 after U18666A treatment. These results indicate that Src-binding capacity in α1 is indispensable for its regulation by membrane cholesterol.

Fig. 6.

Na/K-ATPase/Src interaction defect blocks the regulation of Na/K-ATPase α1 by membrane cholesterol. A425P cells were serum starved and treated with 10 µg/ml U18666A for 24 h and 48 h. A: total cell lysates were prepared and total Na/K-ATPase α1-expression was analyzed by Western blot. B: the biotinylated fractions were prepared, and surface expression of Na/K-ATPase α1 was analyzed by Western blot. Representative Western blots are shown for total and surface expression of Na/K-ATPase α1. The signals were quantified using the ImageJ software and calculated. The quantitative data are presented as means ± SE and n = 4 in each group. No significant difference was shown between groups based on the unpaired t test.

The wild-type α2-subunit does not contain putative Src-binding sequences. Nevertheless, we generated a mutant-α2-expressing stable cell line (LY-a2) containing the α1-like Src interacting sequence and partially restoring the α1-like signaling phenotype (Table 1) (53). To verify the importance of α1-Src interaction in cholesterol-mediated Na/K-ATPase regulation, we performed the following experiments with LY-a2 cells. We first measured the effect of cholesterol reduction on the total expression of mutant α2. As shown in Fig. 7A, no change was detected in 24 h, which was consistent with the prior observations in LX-α2-4 cells; but significant reduction (∼30%) was recorded in 48 h. Second, to verify the α1-like effect in mutant α2 cells, we performed immunostaining analysis. Concomitantly, a significant reduction in the plasma membrane levels of α2 and an increment in the vesicular accumulation of α2 were readily detectable in U18666A-treated cells (Fig. 7B). Third, because U18666A upregulated the accumulation of mutant α2 in intracellular vesicles, we assessed whether cholesterol regulated the endocytosis and degradation of existing α2. The surface expression of mutant α2 was tested by biotinylation assays. As expected, no significant change in 24 h, yet a 40% reduction in 48 h was observed (Fig. 7C). Moreover, we transfected LY-a2 cells with RFP-tagged Rab7, and found the intracellular accumulation of mutant α2 was clearly colocalized with the Rab7-positive vesicles (Fig. 7D). We further checked whether mutant α2 is degraded in the lysosomes by using MG-132 in LY-a2 cells. As depicted in Fig. 7E, MG-132 blocked U18666A-induced downregulation of mutant α2. This suggests U18666A stimulates the endocytosis and degradation of mutant α2 in LY-a2 cells. These findings indicate that the gain-of-Src binding mutations in α2 are capable of partially restoring the regulation of α2 by cholesterol depletion.

Fig. 7.

Mutant Na/K-ATPase α2 partially restores the regulation of Na/K-ATPase α by membrane cholesterol. A: LY-a2 cells were serum starved and treated with 10 µg/ml U18666A for 24 h and 48 h. Total expression of Na/K-ATPase α2 was analyzed by Western blot. Representative Western blots are shown for total expression of Na/K-ATPase α2. The signals were quantified using the ImageJ software and calculated. The quantitative data are presented as means ± SE and n = 4 in each group. ***P < 0.001, compared with control based on the unpaired t test. B: LY-a2 cells were treated with 10 µg/ml U18666A for 24 h and immunostained with anti-Na/K-ATPase α2 antibody. The signals were quantified using the ImageJ software, and the ratio between intracellular and total signals was calculated. Results are presented as box and whisker plots. The number of the cells used for quantification was 20–30 from 3 separate experiments. ****P < 0.0001, compared with control based on the unpaired t test. C: LY-a2 cells were treated with 10 µg/ml U18666A for 24 h and 48 h. The biotinylated fractions were prepared, and surface expression of Na/K-ATPase α2 was analyzed by Western blot. Representative Western blots are shown for surface expression of Na/K-ATPase α2. The signals were quantified using the ImageJ software and calculated. The quantitative data are presented as means ± SE and n = 4 in each group. **P < 0.01, compared with control based on the unpaired t test. D: LY-a2 cells were transfected with RFP-Rab7 for 6 h and then treated with 10 µg/ml U18666A for 24 h. Afterward, cells were fixed and immunostained with anti-Na/K-ATPase α2 antibody. Representative fluorescent images are shown for α2-staining, RFP-Rab7, and merged figures. The arrows point to the colocalization of α2 and RFP-Rab7. Scale bar = 10 µm. E: LY-a2 cells were where serum starved and treated with 10 µg/ml U18666A for 24 h in the presence or absence of 20 µM MG-132. Total expression of Na/K-ATPase α2 was analyzed by Western blot. The signals were quantified using the ImageJ software and calculated. The results were normalized to α-tubulin. The quantitative data are presented as means ± SE and n = 3 in each group. *P < 0.05, compared with control based on one-way ANOVA with Dunnett’s multiple-comparisons test.

Regulation of cellular Na/K-ATPase α by membrane cholesterol requires Cav-1.

Because the above data show that the regulation of Na/K-ATPase α2 is not fully restored in the gain-of-function α2 mutant cells (LY-a2), we further explored whether some other protein is also involved in the regulation of Na/K-ATPase expression by cholesterol. We previously reported that Cav-1, a required structural component of caveolae, interacted with Na/K-ATPase through its NH₂-terminal region and regulated the function of each other (6, 29, 48). According to prior studies, there is difference in Cav-1 expression among the α2-wild-type cells and mutant cells, as well as the Src-binding mutant cells as summarized in Table 1 (23, 53). Unlike wild-type α2, the mutant α2 restores the expression of Cav-1 and is capable of Src regulation as with the α1-isoform. The expression of A425P mutant could not restore either Src regulation or Cav-1 expression. In addition, cholesterol and Cav-1 also affect each other. Therefore, we reasoned that Cav-1 might be important for the regulation of cellular Na/K-ATPase by cholesterol depletion. We therefore tested α1-expression in the Cav-1 knockdown LLC-PK1 cell line (C2-9) after U18666A treatment. As reported (29), Cav-1 expression levels are undetectable in C2-9 cells compared with the control cells (Fig. 8A). As illustrated in Fig. 8B, Cav-1 knockdown rendered total α1 to be resistant to U18666A (Fig. 8B). The unaltered trafficking of α1 in C2-9 cells was further confirmed by immunostaining methods (Fig. 8C). The studies of Cav-1 knockdown cell lines demonstrate that Cav-1 is also required for regulation of Na/K-ATPase by cholesterol depletion.

Fig. 8.

Knockdown of Cav-1 inhibits the regulation of Na/K-ATPase α1 by membrane cholesterol. A: representative Western blots show the total amount of Cav-1 in LLC-PK1 and C2-9 cells. B: C2-9 cells were serum starved and treated with 10 µg/ml U18666A for 24 h. Total cell lysates were prepared, and Na/K-ATPase α1 expression was analyzed by Western blot. The signals were quantified using the ImageJ software and calculated. The results were normalized to α-tubulin. The quantitative data are presented as means ± SE and n = 3 in each group. C: C2-9 cells were treated with 10 µg/ml U18666A for 24 h and fixed for immunostaining of Na/K-ATPase α1. Representative fluorescent images for cellular distribution of Na/K-ATPase α1 in the presence or absence of U18666A are shown. Scale bar = 10 µm. The signals were quantified using the ImageJ software, and the ratio between intracellular and total signals was calculated. Results are presented as box and whisker plots. The number of the cells used for quantification was 20–30 from 3 separate experiments. No significant difference was shown between groups based on the unpaired t test.

Regulation of cellular Na/K-ATPase α by membrane cholesterol requires complex formation among Na/K-ATPase α1, Src, and Cav-1.

We previously found that Na/K-ATPase α1 knockdown redistributes Cav-1 and cholesterol from plasma membrane lipid domains to other cellular compartments (6, 7). Moreover, Src and Na/K-ATPase distribution could be partially restored by expression of mutant-α2 but not by wild-type α2 (Table 1) (53). To further investigate the distribution difference of Cav-1, Na/K-ATPase, and Src in isoform-specific cells, Src-binding mutant cells and Cav-1 knockdown cells, we performed the sucrose density gradient fractionation of these cells. The low-density fractions 4/5 were prepared from these cells and considered as caveolin-enriched caveolar fractions in the plasma membrane with around 60% of total Cav-1 present in these fractions (Fig. 9A). As presented in Fig. 9, B and C, when different cells were analyzed, we found that both Na/K-ATPase and Src were concentrated in fractions 4/5 in control AAC-19 cells, but redistributed to higher density gradients in LX-α2-4 cells, and partially restored by expression of mutant-α2 in LY-a2 cells, which is consistent with what has been reported (53). Moreover, in LM-α3 cells, Na/K-ATPase still resided in fraction 4/5, but the distribution of Src was altered. Interestingly, in Src-binding mutant cells, both Na/K-ATPase and Src were redistributed by A425P mutant. Cav-1 knockdown changed both Na/K-ATPase and Src distribution. The differences in the subcellular distribution of Cav-1, Na/K-ATPase, and Src in these different cell lines, especially as shown in AAC-19 and LY-a2 cells, imply that Na/K-ATPase α1, Src, and Cav-1 together as an intact complex with normal Src-binding function plays an important role in the regulation of cellular Na/K-ATPase by plasma membrane cholesterol.

Fig. 9.

Different isoforms of the Na/K-ATPase, α1-Src interaction mutation, and Cav-1 knockdown redistribute Na/K-ATPase and Src in caveolar fractions. Subcellular fractions from AAC-19, LX-α2-4, LY-a2, LM-α3, A425P, and C2-9 cells were collected and subjected to sucrose gradient fractionation as described in materials and methods. Equal volumes of caveolin-enriched fractions (4/5) and fractions 6–11 were taken for Western blots analysis of Cav-1 (A), Na/K-ATPase α (B), and Src (C). The signals were quantified using the ImageJ software. The quantitative data are presented as means ± SE and n = 3–4 in each group. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001, compared with control; ^#P < 0.05 and ^##P < 0.01, compared between LX-α2-4 and LY-a2 groups based on one-way ANOVA with Tukey’s multiple-comparisons test.

DISCUSSION

Several studies suggest that there is a Src-dependent interplay among Na/K-ATPase, cholesterol, and Cav-1. Na/K-ATPase is able to regulate the distribution of cholesterol and Cav-1 (6, 7). On the other hand, plasma membrane pools of cholesterol regulate the distribution and expression of Na/K-ATPase through Src-dependent pathways (9). The interaction between Na/K-ATPase and Cav-1 in caveolae regulates Cav-1 trafficking and Na/K-ATPase functions (6, 48). Cholesterol depletion increases the caveolae internalization and Cav-1 degradation (18, 36), while in vivo studies revealed a regulatory role of Cav-1 in cholesterol homeostasis (13, 35). Depletion of cholesterol or Cav-1 reduces ouabain-induced Na/K-ATPase endocytosis and Src signal transduction (29, 41). In this study, we demonstrate that the plasma membrane pool of cholesterol regulates the expression and distribution of Na/K-ATPase α1, which is isoform specific through Src and Cav-1 pathway. The Na/K-ATPase/Src/Cav-1 complex, as well as the NaKtide sequence (S415 to Q434) in the nucleotide-binding domain, especially A425, is important for this regulation.

It is important to mention that the expression and distribution of Na/K-ATPase are regulated by plasma membrane cholesterol pools but not U18666A itself considering the similar change in Na/K-ATPase by acute cholesterol depletion with cholesterol-extracting agent methyl-β-cyclodextrin (9). Interestingly, we failed to record significant changes in expression of Na/K-ATPase until longer treatment periods with U18666A in some mutant cells (e.g., LY-a2 cell in Fig. 7). The longer time required for a response may be due to the defect in downstream signaling, which leads to delayed redistribution of cholesterol and then Na/K-ATPase. Therefore, we hypothesize that U18666A regulates Na/K-ATPase in an indirect manner through cholesterol.

The findings presented in Fig. 9 provide evidence that the functional complex of Na/K-ATPase α, Src, and Cav-1 is important to the regulation of Na/K-ATPase by cholesterol. When both Na/K-ATPase and Src are concentrated in caveolar fractions, if Src-binding function is normal in α, U18666A can regulate the expression of Na/K-ATPase α such as in AAC-19 cells. When Na/K-ATPase and Src redistribute from caveolar fractions with loss-of-Src binding function in α, U18666A could not regulate Na/K-ATPase in either LX-α2-4 or A425P cells. When the distribution of Na/K-ATPase and Src is partially restored with gain-of-Src binding mutations in LY-a2 cells, regulation of Na/K-ATPase by U18666A is also partially restored compared with LX-α2-4 cells. With concentrated Na/K-ATPase in caveolar fractions but redistribution of Src and defective Src-binding function in LM-α3 cells, U18666A also fails to regulate the total expression of Na/K-ATPase. When Cav-1 was knocked down, Na/K-ATPase and Src moved to intracellular compartments, and then U18666A could not regulate Na/K-ATPase.

In view of the fact that phosphorylation of Cav-1 at Tyr-14 by Src promotes swelling and release of caveolae (15, 56), and cholesterol regulates endocytosis and degradation of Na/K-ATPase in a Src-dependent manner (9), we propose a molecular mechanism among the interplay of Na/K-ATPase α1, cholesterol, Src, and Cav-1 (Fig. 10). When cholesterol is concentrated in plasma membrane lipid domains (e.g., caveolae), Src binds with Na/K-ATPase, which keeps Src in an inactive state, and Cav-1 maintains caveolae integrity on cell membrane, allowing Na/K-ATPase, Src, and Cav-1 to work together as a functional complex. Depletion of plasma membrane cholesterol may change Na/K-ATPase conformation, release the Src kinase, and result in activation of Src kinase by autophosphorylation of the Y418 site. The activated Src triggers subsequent Cav-1 phosphorylation and caveolae endocytosis, which could cause Na/K-ATPase α1 endocytosis and then degradation. Disruption of Na/K-ATPase α1, Src, and Cav-1 complex or loss-of-Src binding mutation of A425P disrupts the regulation completely. This explains why the signaling complex mainly resides in the cholesterol-enriched caveolae and how cholesterol depletion actives Src and reduces Na/K-ATPase. Moreover, several reports show changes in the activity of Na/K-ATPase and downstream signaling ERK when cholesterol levels change (12, 14) and fit our proposed model. Interestingly, not all Na/K-ATPase isoforms can be regulated by cholesterol, and our proposed mechanism also explains the isoform-specific regulation of Na/K-ATPase by cholesterol. What remains unresolved is why U18666A stimulated α3-internalization but not α3-degradation. It suggests that the process after endocytosis is different between Na/K-ATPase α1 and α3. The α3 may recycle back to the membrane after endocytosis rather than being sorted to the late endosome/lysosome for degradation. Further research is required to reveal why α1 and α3 are sorted differently after membrane cholesterol depletion.

Fig. 10.

The proposed model of Na/K-ATPase/Src/Cav-1/cholesterol interplay. This model shows how plasma membrane pool of cholesterol regulates Na/K-ATPase α1 through the Src and Cav-1 pathway. A: as cholesterol concentrates in caveolae, this maintains Src in an inactive state binding with Na/K-ATPase and keeps Cav-1 in the cell membrane. The intact Na/K-ATPase/Src/caveolin-1 complex only applies to the Na/K-ATPase α1-isoform but not the α2- or α3-isoforms. The arrow points to the A425 site in Src binding sequence in N domain of Na/K-ATPase, which is essential to the regulation of Na/K-ATPase by cholesterol. B: as cholesterol reduces in plasma membrane, it changes Na/K-ATPase conformation to free the Src kinase domain and autophosphorylate Src leading to an open conformation of Src, and it results in activation of Src. Cav-1 is then phosphorylated by the activated Src, which allows for endocytosis of caveolae. C: this leads to Na/K-ATPase α1 endocytosis and then degradation, which consequently decrease plasma membrane Na/K-ATPase.

The different reactions of Na/K-ATPase α-isoforms to cholesterol is of physiological and pathological relevance. Expression of the different α-isoforms would confer cells with susceptibility to cellular cholesterol and adaptation to particular needs of each cell type. On the other hand, Na/K-ATPase may also contribute to diseases that are associated with abnormal cholesterol transport such as atherosclerosis, obesity, and diabetes (8, 20, 22, 46). Our in vivo study might suggest that dysregulation of cholesterol metabolism induced redistribution of Na/K-ATPase α1 in kidney from HFD-fed mice and imply that cholesterol regulated Na/K-ATPase α1 can be physiologically important.

Niemann-Pick type C disease is a neurodegenerative lysosomal cholesterol storage disorder, characterized by the accumulation of cholesterol and glycosphingolipids in various neurovisceral tissues. Nonfunctional NPC protein caused by gene mutations impairs intracellular trafficking of cholesterol in brain, liver, and spleen organs leading to associated dysfunction (1). However, it is still unclear how lipid redistribution leads to cell and tissue dysfunction. Several studies have shown that Na/K-ATPase is important to neuronal cell death and astrocyte glutamate uptake (21, 54). Moreover, our previous report showed downregulation of Na/K-ATPase α1 in brain and liver of NPC1 knockout mice with no change of α2 or α3 in brain, which is consistent with the observations in this study (9). Therefore, we hypothesize that the isoform-specific regulation of Na/K-ATPase by cholesterol contributes to the pathogenesis of NPC1 disease. There are some ongoing therapies in development including histone deacetylase inhibitors, which can correct cholesterol storage defects and increase the expression of mutant NPC1 proteins (39, 40). Considering the unclear pharmacology mechanism of the drugs, our proposed model may provide a potential explanation about how restoring cholesterol by the drugs could recover the function and viability of neuron cells.

Peripherally, our proposed model may also explain the pathogenesis and progression of nonalcoholic steatohepatitis (NASH), which is characterized by steatosis, necro-inflammatory changes, and various degrees of liver fibrosis (31). Although the precise pathophysiology is not fully understood, dysregulated cholesterol metabolism and transport have been described in NASH (2). Accumulation of intracellular cholesterol and reduction of plasma membrane cholesterol in the liver may activate Src and subsequently alter Na/K-ATPase α1-expression and distribution. Although it remains to be investigated if Na/K-ATPase is regulated in NASH liver, it would be of interest to do further studies.

Furthermore, considering no curative treatment for NPC1 disease until recently and new specific agents targeting cholesterol transport being required for NASH, disruption of the interplay among Na/K-ATPase α1, cholesterol, Src, and Cav-1, including the pharmacological blockade of the NaKtide sequence, may help to restore Na/K-ATPase α-expression and subsequently cholesterol trafficking. The future endeavors will provide the therapeutic potential for NPC1 and NASH disease.

GRANTS

This work was supported by the National Heart, Lung, and Blood Institute Grant HL137004 (to D.J.K.), American Heart Association Scientist Development Grant 14SDG18650010 (to D.J.K.), and American Heart Association Scientist Development Grant 17SDG33661117 (to Y.C.).

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

S.V.P., Z.X., and Y.C. conceived and designed research; J.Z., I.L., P.R.D., and D.J.K. performed experiments; J.Z., X.L., H.Y., and D.J.K. analyzed data; J.Z., D.J.K., and Y.C. interpreted results of experiments; J.Z., P.R.D., W.H.W.T. and K.W. prepared figures; J.Z. and Y.C. drafted manuscript; J.Z., I.L., Z.X., and Y.C. edited and revised manuscript; J.Z., X.L., H.Y., I.L., P.R.D., D.J.K., W.H.W.T., K.W., S.V.P., Z.X., and Y.C. approved final version of manuscript.