Lipid rafts are cholesterol- and shingolipid-enriched membrane microdomains implicated in membrane signaling and trafficking. To assess renal epithelial raft functions through the characterization of their associated membrane proteins, we have isolated lipid rafts from rat kidney by sucrose gradient fractionation after detergent treatment. The low-density fraction was enriched in cholesterol, sphingolipid, and flotillin-1 known as lipid raft markers. Based on proteomic analysis of the low-density fraction, the protein with the highest significance score was the α-subunit of Na+-K+-ATPase (NKA), whose raft association was validated by simultaneous immunoblotting. The β-subunit of NKA was copurified from the low-density fraction. To test the role of lipid rafts in sorting and membrane delivery of renal-transporting epithelia, we have chosen to study thick ascending limb (TAL) epithelium for its high NKA activity and the property to be stimulated by antidiuretic hormone (ADH). Cultured rabbit TAL cells were studied. Cholesterol depletion and detergent extraction at warmth caused a shift of NKA to the higher-density fractions. Comparative preparations from blood monocytes revealed the absence of NKA from rafts in these nonpolarized cells. Short-term exposure of rabbit TAL cells to ADH (1 h) caused translocation and enhanced raft association of NKA via cAMP activation. Preceding cholesterol depletion prevented this effect. TAL-specific, glycosylphosphatidylinositol-anchored Tamm Horsfall protein was copurified with NKA in the same raft fraction, suggesting functional interference between these products. These results may have functional implications regarding the turnover, trafficking, and regulated surface expression of NKA as the major basolateral ion transporter of TAL.
- protein sorting
- antidiuretic hormone
- detergent resistant microdomains
- sodium pump
the kidney performs its essential tasks in the maintenance of body fluid homeostasis by segment-specific transepithelial transport within the nephron. Ion transporters and channels allow the selective passage of solutes and electrical charge across the plasma membrane to permit vectorial transepithelial exchange or transport among intracellular compartments (for a review, see Ref. 18). Their physiological function is determined by their subcellular localization, the pathways of their apical or basolateral translocation upon activating or deactivating stimuli, and by the general cellular mechanisms, including intracellular protein networks and sorting signals involved (for a review, see Ref. 29). There is increasing knowledge on the detail of their transport activity, which is determined by the shuttling between vesicular storage pools, the plasma membrane, and an endosomal recycling machinery, involving endocytosis or transcytosis (3).
It has recently been shown for various membrane proteins involved in renal transepithelial transport that their site of action may depend on the local lipid composition. The lipid bilayer of the plasma membrane of all cells is organized into microdomains rich in glycosphingolipids and cholesterol, which are insoluble after treatment with nonionic detergents such as Triton X-100 at low temperature. These microdomains of a size of 50 to 200 nm are commonly referred to as lipid rafts (for a review, see Refs. 10 and 36). Proteins may enter lipid rafts before they reach the cell surface. Lipid rafts preferentially move to the apical membrane but may also be directed basolaterally (31). They may occur as caveolar invaginations of the plasma membrane, carrying isoforms of caveolin (27). Caveolae have earlier been localized also to the basolateral aspects of the collecting duct cells (6). The presence of caveolins, however, is not obligatory for the distribution of transporters into lipid rafts (24). The Na+/H+ exchanger NHE3 has been shown to distribute in lipid rafts, and its activities and trafficking were suggested to be lipid raft-dependent based on cholesterol depletion (CD) studies (28). The epithelial sodium channel, ENaC, has as well been localized to lipid rafts (17). Similarly, proteins such as the thick ascending limb (TAL) glycoprotein Tamm Horsfall protein (THP), putatively associated with TAL transport functions (4), may be integrated within rafts via their transmembrane glycosylphosphatidylinositol (GPI) anchor (8). For the majority of transporters and auxiliary proteins, however, structural or functional association with rafts has not been examined. Having adapted the technology for lipid raft biochemistry in renal cells, we were interested in gaining more insight into their role in nephron function. We therefore used tandem mass spectrometry [liquid chromatography-electrospray ionization quadrupole time-of-flight mass spectrometry (LC-ESI-Q-TOF-MS)] to screen isolated lipid raft fractions from rat kidney for the presence of epithelial transporters. One major finding from this proteomic approach was the presence of significant amounts of the α-subunit of Na+-K+-ATPase (NKA) in lipid rafts. This interesting finding was unexpected in the light of the precedent literature reporting absence of NKA from lipid rafts in MDCK cells (5) and enterocytes (15). On the other hand, more recent data contrastingly were suggestive of a presence of NKA within rafts, and signaling function of the enzyme in cardiac tissue was related with raft integration and the presence of caveolin (22, 23); in fish gills, NKA was further reported to be activated by the glycosphingolipid, sulfogalactosylceramide, which can be enriched in rafts (21). We attempted to resolve this controversial issue with respect to renal NKA. Therefore, we have studied the potential integration of NKA in renal cortical epithelia by means of biochemical and functional evidence. We suggest that the basolateral sodium “pump” in its segment of highest expression, the TAL, may in fact functionally depend on its association with basolaterally oriented lipid rafts.
MATERIALS AND METHODS
Adult male Sprague-Dawley rats (from the local animal facility; 250 g body wt) were anesthetized by intraperitoneal injection of nembutal (40 mg/kg body wt), followed by opening of the abdominal cavity. The experimental protocol was approved by an independent state government review committee under registration no. G0062/05 (Landesamt Berlin). Animals were perfusion-fixed for histochemical analysis of the kidneys as described previously (7). In brief, animals underwent cannulation of the abdominal aorta and perfusion with sucrose/PBS solution (330 mosmol/kgH2O, pH 7.3) for 15 s at a pressure of 220 mmHg, immediately followed by perfusion with paraformaldehyde (3% in PBS, pH 7.3) for another 5 min. Kidneys were then removed and dissected for further preparations. For immunohistochemical analysis, tissues were immersed in 800 mosmol sucrose/PBS solution for 12 h, shock-frozen in liquid nitrogen-cooled isopentane, and stored at −70°C. Alternatively, native kidneys were removed for biochemical analysis.
Protein identification by tandem mass spectrometry.
The low-density fraction from rat kidney tissue homogenates (postnuclear supernatant) was prepared for proteomic characterization of lipid raft proteins obtained from discontinuous floating assay as detailed below. The bands of proteins with molecular mass between 50 and 150 kDa, as separated by SDS gel electrophoresis, were excised, washed with 50% (vol/vol) acetonitrile in 25 mM ammonium bicarbonate, dehydrated in acetonitrile, and dried in a vacuum centrifuge. Gel pieces were incubated in 10 μl of 5 mM ammonium bicarbonate containing 60 ng trypsin (sequencing grade, modified; Promega). After 15 min, 10 μl ammonium bicarbonate was added to keep the gel pieces wet during enzymatic cleavage (12–16 h). To extract peptides, 20 μl of 0.5% (vol/vol) trifluoroacetic acid in acetonitrile was added, and samples were sonicated for 3 min, vacuum-dried, and reconstituted in 6 μl of 0.1% (vol/vol) trifluoroacetic acid (TFA) and 10% (vol/vol) acetonitrile in water.
Tandem mass spectrometry experiments were performed in an ESI-Q-TOF mass spectrometer (Ultima; Micromass) equipped with a Z-spray nanoelectrospray source and a CapLC liquid chromatography system. Liquid chromatography separations were performed on a capillary column (PepMap C18, 3 μm, 100 Å, 150 mm × 75 μm ID; Dionex) at an eluent flow rate of 200 nl/min using a linear gradient of 3–64% acetonitrile/0.3% TFA in water during 60 min. The mass spectrometer was operated in the positive ion mode using PicoTip spray capillaries (New Objective; Woburn). Argon was used as collision gas (pressure 6.0 × 10−5 mbar). The processed MS/MS spectra (MassLynx version 4.0 software) and the MASCOT program were used to search in the NCBI nonredundant protein database. A score level of >43 was considered significant. The major results are listed in Table 1, stressing the role of NKAα.
SV40-transformed rabbit TAL cells obtained from rabbit kidney medulla were cultured as described (37). We have selected these cells for their fine structure and polar organization criteria by electron microscopy analysis. Cells were seeded on culture dishes or glass cover slips (Sigma) and cultured in DMEM/Ham's F-12 medium containing l-glutamine, 15 mM HEPES, 1% penicillin/streptomycin, 5% FCS, 1% l-glutamine, and 1% nonessential amino acids (7.5% CO2, 37°C; GIBCO). Medium was changed every 2 days. Cells were utilized up to the 10th passage. Monocytes were prepared from human peripheral blood of healthy donors by differential centrifugation on Ficoll-Hypaque (GIBCO), maintained in polystyrene culture flasks for adherence (2 h), and subsequently washed with RPMI medium containing 10% FCS (Seromed) to remove nonadherent cells (14). The adherent cells comprising ∼80% monocytes, and the nonadherent cells dominated by lymphocytes, were harvested and homogenized for biochemical analysis. Cell viability was tested by trypan blue staining (Sigma) before homogenization.
The following antibodies were used: mouse monoclonal antibody against rat α1-subunit of NKA (1:1,000; Upstate), mouse monoclonal antibody against rat β1-subunit of NKA (1:200; Upstate), rabbit polyclonal antibodies against rat tubulin (1:500; Sigma), rat clathrin (1:200; Progen), rat caveolin-1 (1:200; Santa Cruz), rat flotillin-1 (1:200; Beckton-Dickinson), and rat THP (1:500; kindly provided by J. R. Hoyer, Philadelphia, PA). For tissue localization, 4-μm-thick paraffin sections were deparaffinized, rinsed in PBS, and incubated with blocking medium followed by the antibodies. Cultured cells grown on cover slips for 3–7 days were fixed with cold methanol, rinsed, and incubated with antibody. Bound primary antibodies were detected with Cy-2-, Cy-3 (both 1:200)-, or horseradish peroxidase-conjugated secondary anti-mouse or anti-rabbit antibodies (1:2,000; Dianova). Combined application of anti-NKAα and anti-THP antibodies was performed to identify colocalization. 4,6-Diamidino-2-phenylindole (Abcam) was used for nuclear counterstain. Image acquisition was performed as previously described (3) using a Leica DMRB microscope (Leica). Images were taken with a digital camera (Spot 32; Diagnostic Instruments).
Cholera toxin and NKAα double staining.
rbTAL cells grown on cover slips were fixed with ice-cold acetone (−20°C) and incubated with fluorescent marked cholera toxin (CT), a ligand for the lipid raft marker ganglioside GM1; the conjugate was cross-linked with anti-CT (Vybrant lipid raft labeling kit; Sigma) and double stained with anti-NKAα followed by incubation with Cy-3-coupled secondary antibody. Cells were viewed in a confocal laser-scanning microscope with a multilaser system (Ar laser 458–514 nm, HeNe laser 543 nm, HeNe laser 633 nm) under standard digital scanning conditions (Leica). Multichannel detection was checked by sequential scanning analysis to avoid overlapping fluorescence emission signal to be recorded in double-immunostained cells.
Kidneys were homogenized using a tissue homogenizer (5 min) and cells by ultrasonication (25 s) in sucrose buffer containing 10 mM triethanolamine, pH 7.5, 250 mM sucrose, and protease inhibitors (Complete; Roche Diagnostics), and the homogenates were centrifuged at 300 g (30 min) to remove whole cells, nuclei, and mitochondria. The supernatant (S1) was then centrifuged at 17,000 g (30 min) for preparation of the membrane fraction (P1), and the supernatant (S2) was centrifuged at 120,000 g (1 h) for vesicle preparation (P2). Alternatively, the supernatant S1 was centrifuged at 120,000 g (1 h) to obtain a pellet fraction containing both vesicles and membranes (P3). Protein concentration was measured using a BCA Protein Assay reagent kit (Pierce); 20 μg protein/lane were run on 8 or 10% SDS-polyacrylamide minigels; gels were silver stained or blotted on nitrocellulose membranes, incubated with anti-NKAα (recognizing the 98-kDa band), anti-NKAβ (48 kDa), anti-tubulin (50 kDa), anti-clathrin (180 kDa), anti-THP (80 kDa), and anti-CD14 antibody (55 kDa; antibody from Cymbus Biotechnology). Antibodies were visualized by chemiluminiscence (ECL kit; Amersham).
Triton X-100 solubility assay.
S1 supernatants of tissue and cell homogenates were incubated on ice for 1 h in sucrose buffer containing 1% Triton X-100 and centrifuged at 120,000 g for 1 h. The Triton X-100-insoluble pellet fraction was resuspended in sucrose buffer without Triton X-100. Supernatants and pellets were analyzed by Western blotting as described above.
Sucrose gradient fractionation (floating assay).
Vesicles and membrane fractions (P3) from cells or kidney homogenates were resuspended in 250 mM sucrose buffer containing 1% Triton X-100. Alternatively, 10 μM cytochalasin B (30 min; Sigma) was used before Triton X-100 extraction (35). The insoluble fraction was centrifuged (120,000 g), resuspended in 1 ml buffer containing 40% sucrose, overlayed with 4 ml of a continuous gradient (30–5% sucrose; continuous floating assay, lipid raft proteins ranging in fractions near the 20% sucrose level), or alternatively with 2 ml 30% sucrose plus 2 ml 5% sucrose (discontinuous floating assay, lipid raft proteins ranging in the interface between 5 and 35% sucrose levels), and centrifuged at 200,000 g for 16 h. Fractions were collected from the top of the gradients and analyzed by gel electrophoresis and Western blotting. For disintegration of lipid rafts, P3 was incubated with Triton-X-100 at 37°C for 1 h (33).
Analysis of lipid composition.
Samples (10 μl) of fractions from floating assays were diluted with 100 μl 50 mM HEPES, pH 8.0, and mixed with 200 μl methanol/chloroform. Two-phase extraction lipids were separated by TLC using a mixture of chloroform-methanol-H2O (65:25:1; see Ref. 35). Lipids were visualized by 20% H2SO4 at 150°C. Cholesterol, phosphatidylcholine, and sphingomyelin (Sigma) were used as standards.
Depletion of membrane cholesterol.
For CD, the rbTAL cell cultures were incubated for 1 h, at 37°C, with 10 mM methyl-β-cyclodextrin (MβCD; Sigma; see Ref. 35) and homogenized, and P3 was analyzed by floating assays and Western blots as described above. In addition, cells were cultured in the presence or absence of 4 μM lovastatin and 0.25 μM mevalonate (Sigma; see Ref. 20) for 48 h and subsequently for 2 h with 10 mM MβCD, followed by immunochemical analysis. CD was controlled in cell culture by staining with filipin. The cells were fixed for 30 min on ice with 4% paraformaldehyde and then stained for 15 min at room temperature with filipin in 15% glycerol. Images were taken with a digital camera (Spot 32; Diagnostic Instruments).
Antidiuretic hormone treatment.
Cells were incubated in culture flasks or on cover slips with 1 × 10−7 mol/l antidiuretic hormone (ADH; Sigma) in culture medium for 10, 30, 60, 120, or 180 min or 24 h as established earlier (38), harvested and homogenized, or stained using immunochemical methods. Alternatively, cells were concomitantly exposed to ADH and the inhibitory cAMP analog, Rp-cAMPS (10 μM; Sigma) for 1 h.
The magnetic cell separation system (MACS; Miltenyi Biotec) with protein G microbeads was used for immunoprecipitation. S1 from tissue and cell homogenates were incubated on ice for 1 h in sucrose buffer containing 1% Triton X-100 and centrifuged at 120,000 g for 1 h. The Triton X-100-insoluble pellet fraction was resuspended in sucrose buffer and incubated with the specific antibodies and protein G microbeads. The magnetically labeled immune complex is bound to columns placed in a magnetic field, whereas other proteins were washed away (nonprecipitated fraction). Immunoprecipitated protein was eluted from the columns outside of the magnetic field. Eluates of precipitated and nonprecipitated fractions were analyzed by Western blots as described above. As control, THP antibodies were preincubated with the antigen THP overnight at 4°C and prepared as described above.
Analysis of fractions from floating assays.
Samples were prepared from rat kidney homogenates by Triton X-100 extraction on ice and SDS gel electrophoresis performed on fractions from floating assays on discontinuous gradients (5, 30, and 40% sucrose), availing buoyancy of lipid raft proteins to low-density fractions on the top of the gradient. Silver gel staining showed various proteins present in the low-density fractions (Fig. 1A; 5% sucrose) with different prominent bands. ESI-Q-TOF analysis of the low-density fraction was performed after polyacrylamide SDS gel electrophoresis and staining. Protein bands of molecular masses between 50 and 150 kDa were excised for analysis. A number of proteins, each with a significant score >43, were identified with NKAα revealing highest significance (score of 628; Table 1 and Fig. 1B). In addition, the raft protein alkaline phosphatase (30) was detected in parallel, thus validating the assay (Fig. 1C). Based on these findings, the association of NKAα with lipid rafts was subjected to further study.
Antibodies suitable for biochemical analysis were tested by immunohistochemistry first (Fig. 2, A and B). Rat kidney typically showed expected basolateral NKAα staining of all epithelia with preferential signal intensity in the distal tubule (Fig. 2A). Cultured rbTAL cells also expressed NKAα basolaterally (Fig. 2A). As a positive reference protein to be recovered regularly in the floating fractions (32), tubulin was used; its immunoreactivity was broadly distributed throughout the cytoplasm of distal epithelial and rbTAL cells. Clathrin was used as a negative reference for its reported absence from lipid rafts (26); its localization is shown in the apical cytoplasm of epithelial cells, whereas in rbTAL cells, it showed a diffuse to perinuclear distribution (Fig. 2A). Double staining with TAL-specific THP demonstrated the strongest NKAα immunoreactivity in the TAL (Fig. 2B). By contrast, the widely distributed lipid raft marker caveolin-1 is not expressed in TAL (Fig. 2B). Instead, we therefore used antibody to flotillin-1 as a raft marker protein (Fig. 2A).
Triton X-100 solubility.
Extracted proteins were subjected to single-step high-speed separation by centrifugation in cold Triton X-100. The resulting detergent-soluble and detergent-resistant fractions were analyzed by immunoblotting. Both the soluble and insoluble fractions from kidney revealed the presence of NKAα. Tubulin showed a strong band in the detergent-resistant fractions and a weaker band in the soluble fraction. Signal for clathrin was unexpectedly dominant in the insoluble and only weak in the detergent-soluble fraction (Fig. 3A).
Sucrose gradient analysis (floating assay).
We next fractionated vesicle and membrane preparations from kidney and cell homogenates by subjecting them to high-speed centrifugation in cold detergent and loaded the detergent-resistant fraction on a discontinuous sucrose gradient. This step permitted the separation of membranes according to their different buoyant density. NKAα partitioned into all of the three density fractions in approximately equal amounts, whereas the signals for the β-subunit were low in the 5 and 30% fractions but strong in the 40% fraction, respectively (Fig. 3B). Tubulin signal showed varying intensities in all three density fractions, whereas clathrin was present only in the high-density fractions. To compare for polarity criteria, human blood monocytes were analyzed in parallel (Fig. 3C). On discontinuous sucrose gradient, only the high-density band showed NKAα signal in the immunoblot, demonstrating absence of the enzyme from the lipid raft fraction. Used as a positive control, the lipopolysaccharide receptor CD14 was substantially detectable in the lipid raft-containing low-sucrose fraction (34). Tubulin showed weak signals in the low-density fraction at 5% and strong signals at 30 and 40% sucrose (Fig. 3, B and C). In the analysis of lymphocytes, NKAα was absent from the lipid raft fractions as well (data not shown). To further confirm the floating behavior of kidney and rbTAL cell membrane proteins with gradient separation, we prepared continuous sucrose gradients loaded with the detergent-resistant fractions (Fig. 3D). Immunoblotting revealed that the NKAα signal peaked near the 20% fraction, as expected for lipid raft proteins. With the use of the reference proteins THP and flotillin, immunoblot signal with these proteins was concentrated as well in the 20% fraction range. Similar results were obtained with rbTAL cell extracts subjected to continuous sucrose gradient assays (data not shown). Our findings thus demonstrate that NKAα is included in the fractions typically enriched in lipid rafts.
Lipid analysis of floating fractions.
Lipid composition analysis revealed that cholesterol and sphingomyelin were enriched in the low-density fraction from discontinuous floating assay (Fig. 4A), or in the 20% range of the continuous gradient (data not shown). Phosphatidylcholine as a membrane component not specific for rafts was found in all fractions.
Influence of temperature and CD on floating ability of NKA.
Treatment of rat kidney homogenates with Triton X-100 at 37°C, a temperature that leads to raft disruption, resulted in the near absence of NKAα and the lipid raft marker flotillin from low-density fractions when subjected to discontinuous sucrose gradient fractionation (Fig. 4B). Lipid composition analysis of the same samples detected cholesterol and sphingomyelin in the detergent-soluble fraction, which demonstrates the disruption of lipid rafts after incubation at 37°C (Fig. 4C). In an attempt to disrupt association of lipid raft proteins with cytoskeletal components, we treated rbTAL cells with cytochalasin B before Triton X-100 extraction. The floatation of the cytoskeletal protein tubulin in the low-density fraction was inhibited after cytochalasin treatment, whereas floatation of NKAα was moderately increased (data not shown). Treatment of rbTAL cells with MβCD to disrupt the lipid rafts induced a shift in the distribution of NKAα and tubulin from low- to high-density fractions (Fig. 4D). The detection of clathrin in the high-density fraction was not influenced.
Involvement of lipid rafts in the trafficking of NKA.
To demonstrate the involvement of cholesterol-rich microdomains in the trafficking of NKAα protein in the cell, rbTAL cells were cultured on cover slips in cell culture medium with vehicle or the cholesterol synthesis blocker lovastatin combined with mevalonate and MβCD for CD. Cells were then treated with ADH or vehicle. Both CD and ADH treatment had minimal effects on cell viability. Double staining of the lipid raft marker ganglioside GM1, using the FITC-labeled β-subunit of cholera toxin, and of NKAα demonstrated partial colocalization of both markers (Fig. 5A). Treatment with ADH for 1 h increased the amount of fluorescent spots, revealing dual staining for NKAα and GM1. Filipin immunostaining indicated successful CD by its reduction in staining intensity (Fig. 5B). NKAα immunostaining was reduced by CD irrespective of ADH treatment. Treatment with the inhibitor of ADH-mediated cAMP activation, Rp-cAMPS, prevented the stimulation by ADH.
We next treated rbTAL cells in culture flasks with ADH sequentially for different time intervals and analyzed plasma membrane and vesicle fractions by Western blotting (Fig. 6, A and B). NKAα was increased in the plasma membrane fraction within 10 min, and plateaued after stimulation for 1 h, whereas the signal in vesicles moderately decreased during the first 10 min, suggesting translocation of NKAα in the cell membrane (Fig. 6, A and B). Discontinuous floating assays demonstrated that a higher proportion of NKAα was detected in the lipid raft containing low-density fractions after treatment with ADH (Fig. 6C). Disintegration of lipid rafts by incubation at 37°C markedly reduces the presence of NKAα in extracts from both ADH-treated and untreated cells.
Immunoprecipitation of NKAα with proteins expressed in TAL.
In TAL, NKA is expressed together with THP, and, like NKA, THP immunoreactivity is located to a substantial proportion at the basolateral cell membrane (data not shown). In addition, both proteins occur in lipid rafts (Fig. 3D). We therefore concluded that both proteins are integrated in the same lipid rafts. Rat kidney or rbTAL homogenates treated with cold Triton X-100 were immunoprecipitated with an antibody against THP (Fig. 6D). Western blot using an antibody for NKAα was performed on precipitated (Fig. 6D, lanes 1, 2, and 6) and nonprecipitated (Fig. 6D, lanes 3, 4, and 7) fractions. Approximately one-third of the NKAα-immunoreactive signal was detected in the immunoprecipitated fraction, whereas two-thirds were in the nonimmunoprecipitated fraction from kidney homogenates. We used rbTAL cell homogenates and found that the signals had approximately the same intensity in the two fractions. These results suggest that NKAα and THP are partly colocalized within the same lipid rafts. Negative controls were performed by preincubating THP antibody with THP antigen, which led to a marked reduction of the Western blot signal immunoprecipitating NKAα with THP antibody. (Fig. 6D, lane 5).
The present study combines a proteomic approach with biochemical and cell biological analysis for protein identification in glycosphingolipid- and cholesterol-enriched membrane microdomains or lipid rafts. One-dimensional SDS-electrophoresis separation was combined with mass spectrometry-based proteomics (ESI-Q-TOF) for global protein analysis of the lipid raft fractions from rat kidney homogenates. As a major result, the main catalytic subunit of NKA, NKAα, had reached by far the highest probability score among the identified proteins listed in Table 1. Another eight of these were raft associated according to published data. Among them, alkaline phosphatase has been particularly well documented with respect to its lipid raft properties (30). The remaining proteins were not necessarily intrinsic components of lipid rafts because of a potentially loose association with rafts; their lower scores may reflect this (20). Using biochemical and cell biological assays, we have confirmed the integration of NKAα in lipid rafts from kidney homogenate and rbTAL cells. We have further shown that, in TAL cells, the proportion of NKAα associated with rafts may be increased upon a locally activating hormonal stimulus (ADH), suggesting an interaction of rafts with polar sorting and surface expression of the enzyme.
Technically, to characterize our lipid raft preparations, we have applied established detergent insolubility criteria. We have used Triton X-100 extraction of the tissue and cell homogenates at cold to obtain the lipid raft fraction, including integrated proteins or protein complexes, based on its low buoyant density and relative insolubility in the nonionic detergent (for a review, see Ref. 36). We have further analyzed the floating properties of the raft populations using sucrose gradient fractionation (floating assays). NKAα was found to be present in the detergent-insoluble fraction and in the low-density fractions from the discontinuous and continuous floating assays; approximately one-third of the immunoreactive NKAα was detected in the latter, suggesting that a minor but significant proportion of NKA partitions into lipid rafts at baseline condition. To verify temperature dependence, control assays were performed at 37°C, resulting in the near absence of NKAα from the low-density fractions. The nature of the lipid rafts, as assessed by TLC, was confirmed by significant enrichment of sphingolipid and cholesterol selectively in the low-density fractions, as shown elsewhere (35). Because lipid raft association of a basolateral membrane protein was an uncommon, although not unprecedented, finding (1, 26, 31), further controls were included testing copurification of positive and negative reference proteins. Comparative analysis of NKA in nonepithelial blood cells was additionally performed. Among the positive reference proteins, flotillin-1 as a well-established lipid raft marker (17, 33), THP as a GPI-anchored raft- and TAL-specific product (8), and tubulin, usually included in assemblies of lipid rafts with cytoskeletal proteins (20), were applied. The coated pit-related protein clathrin was used as a negative reference based on its clear-cut absence from the low-density sucrose gradient fractions throughout (26). Its dominance in the insoluble fraction from detergent solubility assay is probably related to its property to form high-order oligomers that are largely unrelated to lipid rafts. A minor proportion of the cellular clathrin pool may nevertheless be lipid raft associated, although potential methodological and cell type-related discrepancies must be considered (22). Using CD with MβCD in rbTAL cells, we clearly produced a shift of NKAα to the high-density fractions, and this was also observed with tubulin, underlining that the presence of NKAα in the low-density floating fractions was lipid raft dependent, which agrees with analogous observations on other raft-associated products (17, 28, 35).
Because both α- and β-subunits are required for regular activity of NKA (16), and because the proteomic analysis showed the presence of the NKAβ precursor, we also tested for copurification of NKAβ in the lipid raft preparations. The resulting weak signal in the raft fractions obtained from tissue and cultured cells suggests that NKAβ, to some extent as well, is associated with lipid rafts. Because both subunits may be assembled at the plasma membrane (11), their common integration within rafts may thus be functionally relevant. Establishing the absence of an association of NKA with lipid rafts in nonpolarized blood monocytes (34), we could further provide indirect support that raft-related NKA in polarized cells may have a distinct role serving vectorial transepithelial transport.
Our methodological assays thus contribute to the currently increasing evidence that, in addition to GPI-anchored or acylated proteins originally associated with lipid rafts, ion channels and transporters with multiple membrane-spanning domains, such as NKA, may partition into rafts for potential interaction (17, 21, 26, 28). The renal epithelial presence of NKA in lipid rafts, given the precedents in literature, was denied or questioned by some (5, 15, 34), whereas more recent, contrasting findings may be found as well (21, 26, 39). In the intestinal mucosa, serving vectorial transport like renal epithelia, Hansen et al. (15) used floating assay and immunoelectron microscopy to demonstrate that basolateral sorting of NKA was unaffected by CD, whereas raft association and luminal insertion of aminopeptidase N, a brush-border enzyme, was profoundly disturbed; they concluded that rafts serve as lateral sorting platforms chiefly for apically destined membrane traffic. Others have used endogenous NKA of MDCK cells as a negative control to demonstrate lipid raft association of a transfected potassium channel together with caveolin, but not of NKA (5). By contrast, non-transport-related signaling of a caveolar pool of NKA detected in the light fraction purified from cardiac myocyte and proximal kidney cell line extracts was reported to function in a ouabain-sensitive manner, tethering the inositol trisphosphate receptor and phospholipase C into a calcium-regulatory complex (22, 23, 39). Ouabain-induced signal propagation was further related to endocytosis of NKA “signalplexes” in a caveolae-/lipid raft-dependent manner (22). An analogy with the present results, however, is limited since we have applied contrasting, primarily lipid raft-oriented purification protocols. We and others (6) have furthermore shown that the major form of caveolin, caveolin-1 (27), was mainly present in the basolateral membranes of connecting tubule and collecting duct epithelia but absent from TAL. Although total kidney homogenates had been analyzed, our protocols may therefore have included a proportion of the signaling type of NKA. Other more comparable supportive results came from the analysis of lipid rafts determining NKA activation in fish gills upon sea water adaptation compared with a baseline fresh water condition in which NKA was absent from rafts; specifically, raft-typic sulfogalactosylceramide enrichment was proposed to confer increased activity to the enzyme since this glycosphingolipid may function as a cofactor for NKA (21). Interestingly, also in rat liver cells, a basolateral association of NKA and lipid rafts has recently been shown (26).
A role for lipid rafts in polar sorting and functional activation of membrane proteins has been shown for other raft-associated transporters and channels (17, 26, 28). Exposure to short-term ADH, a potent stimulus not only for water channel and NKA activation in collecting duct (12), but also for V2-receptor-mediated, cAMP-dependent transport activation in TAL (9, 13, 19), was therefore selected to test for lipid raft dependence in presumed translocation events of NKA in this epithelium. A hierarchy of mechanisms governing polar insertion of NKA has earlier been identified in MDCK cells (25). Interestingly, in that study, application of fumonisin B1, which inhibits sphingolipid synthesis, had caused a disorder in the polar sorting of glycosylphosphatidylinositol-anchored glycoprotein (GP-2), related to THP, as well as NKA, suggesting lipid raft-mediated influences in polar sorting of these products. Comparingly, we have interpreted the present immunohistochemical approach with antibody staining of NKAα and concomitant lipid raft labeling by cholera toxin-B as a strong indication for rafts mediating sorting of NKA; this was based on the significant shift of NKAα-immunoreactive sites into raft sites upon 1 h exposure with ADH acting via cAMP in the rbTAL cells. These results were cytochemically corroborated by CD and controlled by cellular filipin staining, which led to diminished fluorescent NKAα signal on one hand and to a reduction in short-term ADH-induced membrane insertion on the other. Although the former effect may be related with profound redistribution of newly synthesized NKA in cells lacking rafts, the latter may also demonstrate the need for rafts in reactive translocation to the membrane, as demonstrated elsewhere for luminal transporters (28). Further confirmation came from biochemical assays demonstrating a shift of NKA from a vesicular pool to the plasma membrane upon short- and long-term exposure to ADH and from floating assays showing a shift of immunoreactive NKAα in the lipid rafts fraction. Comparable results have been obtained studying aldosterone-induced partitioning of ENaC into lipid rafts (17).
Finally, we discovered copurification and joint localization of the TAL major glycoprotein THP and NKA within lipid rafts. This may be functionally relevant regarding the mentioned relation between NKA and GP-2 for polar trafficking in MDCK cells (25). Like GP-2, THP may be sorted apically and basolaterally, and, because its GPI anchor may generally influence sorting events, coinsertion of THP and NKA in the same rafts may be related with sorting events.
In summary, proteomic screening of the lipid raft fraction from rat kidney revealed the highest probability score for NKA. Biochemical and cell biological analysis confirmed the integration of NKA in lipid rafts of the TAL, the site of highest NKA activity. The proportion of NKA associated with rafts of TAL may be increased under stimulation by ADH, suggesting a role for rafts in polar sorting and surface expression of the enzyme.
We thank Rolf K. H. Kinne for the gift of rbTAL cells. We are also indebted to Frauke Serowka for skilful technical assistance.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society