renal potassium secretion and reabsorption is mediated by various ion transport mechanisms and plays a major role in maintaining the systemic K⁺ equilibrium. To carry out these processes, ion-transporting proteins must be differentially expressed along the segments of the nephron, and their subcellular distributions in either the apical or basolateral membranes of renal epithelial cells must be rigorously controlled. Active secretion of K⁺ into the lumen of the renal collecting tubule requires the participation of the basolateral Na⁺-K⁺-ATPase and apically polarized K⁺ transporting ion channels. Previous studies revealed that under conditions of low dietary intake, luminal potassium secretion is reduced and is instead superseded by active K⁺reabsorption (46). A large body of evidence indicates that active K⁺ reabsorption in the renal collecting tubule is mainly attributable to the activity of several H⁺-K⁺-ATPases. However, the molecular mechanisms that regulate these pump activity, expression and subcellular distributions have yet to be identified.

The H⁺-K⁺-ATPases belong to the family of P-type ion-transporting ATPases and are structurally related to the Na⁺-K⁺-ATPases (32). These heterodimeric proteins are composed of a catalytic α- and an associated β-subunit (26). The polytopic α-subunit spans the membrane 10 times, is not glycosylated, and includes the enzyme's sites for ATP binding and ion translocation. Assembly with the single membrane spanning and highly glycosylated β-subunit is a prerequisite for posttranslational processing and transport of the newly synthesized ion pump from the endoplasmic reticulum (ER) to the plasma membrane (16). Although members of the H⁺-K⁺-ATPase family show strong structural homology to one another, their unique functional characteristics clearly differentiate two subgroups. One subgroup is defined by the gastric H⁺-K⁺-ATPase, which is also responsible for acid secretion in the parietal cells of the stomach. The other class is composed of the nongastric H⁺-K⁺-ATPases. Both types are expressed along different nephron segments and show characteristic pharmacological sensitivities (13, 46). Recent studies suggest that nongastric H⁺-K⁺-ATPases can transport protons in exchange for potassium but act primarily as Na⁺-K⁺-ATPases under physiological conditions (10, 21, 22).

Interactions with different β-subunits are known to modify the biochemical and cell biological properties of P-type ATPases (7, 17). It was recently shown that the rat colonic nongastric H⁺-K⁺-ATPase can assemble with the β-subunit of the Na⁺-K⁺-ATPase, with the β-subunit of the gastric H⁺-K⁺-ATPase and with a newly identified H⁺-K⁺ β-subunit from colon (8, 9, 28). The human nongastric H⁺-K⁺-ATPase, ATP1AL1, shows different β-subunit affinities (21, 35). The ATP1AL1 α-subunit will interact with the gastric H⁺-K⁺-ATPase β-subunit but not with the Na⁺-K⁺- ATPase β-subunit isoforms when expressed by transfection in HEK-293 cells. Gastric H⁺-K⁺-ATPase β-subunit expression has been detected in epithelial cells of the renal collecting tubule (5).

Both functional and structural localization studies have detected the gastric H⁺-K⁺-ATPase in the luminal membranes of stimulated gastric parietal cells (25) and renal tubule epithelial cells (46). In gastric parietal cells under resting conditions, the H⁺-K⁺-ATPase is stored in the membranes of tubulovesicular elements (TVEs). After stimulation of the cells, the TVEs fuse with the apical membrane and expose the H⁺-K⁺-ATPase to the gastric lumen. Ion pump inactivation and restoration of the TVEs is regulated by endocytosis (11). The requisite endocytosis signal is localized on the short N-terminus of the gastric H⁺-K⁺-ATPase β-subunit. Interestingly, it was recently shown that similar mechanisms also influence the H⁺-K⁺-ATPase-mediated K⁺reabsorption in renal cells (46). Although the majority of functional and pharmacological studies predict an apical distribution for the nongastric H⁺-K⁺-ATPases, their actual cellular localizations and modes of regulation are not clear. Except for the rat colonic nongastric H⁺-K⁺-ATPase subtype, which was immunohistochemically localized in the apical membrane of rat colon and renal principal cells (PC) (29, 38), none of the various nongastric H⁺-K⁺-ATPase subtypes has been structurally localized in renal epithelial cells. Moreover, it is conceivable that nongastric H⁺-K⁺-ATPases can also be differentially localized in epithelial cells.

Sorting information that directs the gastric H⁺-K⁺-ATPase to the apical membrane has been identified in the fourth transmembrane domain of the α-subunit and is restricted to a sequence of 8 amino acids (14). In the nongastric H⁺-K⁺-ATPases, the sequence in this region is essentially identical to that present in the α-subunit of the basolateral sodium pump. Thus either the sorting behaviors or the sorting signals manifest by the nongastric H⁺-K⁺-ATPases must differ substantially from those employed by their gastric counterpart. To elucidate the sorting properties of nongastric H⁺-K⁺-ATPases, we investigated the spatial distribution of the human nongastric H⁺-K⁺-ATPase, ATP1AL1, expressed by transfection in polarized renal epithelial cells.

EXPERIMENTAL PROCEDURES

Tissue culture.

MDCK cells were maintained in minimal essential medium supplemented with Earle's salts (EMEM) and LLC-PK₁ cells were maintained in Dulbecco's modified Eagle's medium (DMEM). Both media contained 10% fetal bovine serum (Sigma), 2 mMl-glutamine, and 50 U/ml penicillin and streptomycin (Life Technologies, Grand Island, NY). Stably transfected cells were grown in the presence of 0.9 g/l Geneticin (Life Technologies). Cells were cultured under standard conditions (37°C, 5% CO₂) and passaged twice a week with 0.05% trypsin and 0.5 mM EDTA (Life Technologies).

Stable transfection.

ATP1AL1 α-subunit cDNA and H⁺-K⁺-ATPase β-subunit cDNA were subcloned in the mammalian expression vector pJB20 and pCB6, respectively, as previously described (21). Single transfections of MDCK and LLC-PK₁cells with ATP1AL1 α-subunit and cotransfections with the H⁺-K⁺-ATPase β-subunit were performed with the PerFect transfection kit from Invitrogen (San Diego, CA) and DOTAP liposomal transfection reagent (Roche, Mannheim, Germany) according to the manufacturer protocols. Both cell types do not express ATP1AL1 or the gastric H⁺-K⁺-ATPase β-subunit endogenously. After Geneticin selection for 3 wk, ATP1AL1 α-subunit and H⁺-K⁺-ATPase β-subunit cotransfected cells were cultured in ouabain (0.5–1 μM) containing medium for 3 days. Only cells functionally expressing the H⁺-K⁺-ATPase, ATP1AL1, survived this additional selection. Afterwards, transfected cells were cloned by single cell dilution.

Cell surface biotinylation.

Transfected cells were grown on 24-mm polycarbonate Transwell filter inserts (0.4 μm pore, Costar) for 10 days. Medium was replaced daily. Before biotinylation, expression of the transfected cDNAs was enhanced by incubation for 12 h with media containing 10 mM sodium butyrate. Cell surface biotinylation was performed as previously described (19) using NHS-SS-Biotin (Pierce, Rockford, IL). Apical and basolateral biotinylation was performed for 40 min, at pH 9.0, and biotinylated proteins were separated with streptavidin-agarose beads (Pierce). Protein concentration was determined (12), and similar amounts (0.2–1 μg) of biotinylated proteins were analyzed by SDS-polyacrylamide gel electrophoresis and Western blotting using a rabbit polyclonal, affinity-purified anti-ATP1AL1 (α_XG) antibody (1:1,000; Ref. 21), a monoclonal mouse anti H⁺-K⁺-ATPase β-subunit antibody (1:250; kindly provided by J. Forte and D. Chow, University of California at Berkeley, Berkeley CA), and a mouse monoclonal anti Na⁺-K⁺-ATPase β-subunit antibody (Upstate Biotechnology). Detection was performed using either goat anti-mouse or goat anti-rabbit antibodies (1:1,000) conjugated to horseradish peroxidase (Sigma, St. Louis, MO) and developed by the enhanced chemiluminescence technique (ECL; Amersham, Arlington Heights, IL).

Immunofluorescence.

Transfected MDCK and LLC-PK₁ cells were seeded at low densities (1 × 10⁵ cells/24-mm filter) on polycarbonate Transwell filter inserts (Corning Costar) and grown for at least 7 days. Medium was changed daily, and 12 h before fixation supplemented with 10 mM sodium butyrate. Cells were washed with PBS⁺ (150 mM NaCl, 10 mM NaH₂PO₄, 0.1 mM CaCl₂, and 1 mM MgCl₂; pH 7.4) and fixed with ice-cold methanol for 7 min. Blocking (30 min) and antibody dilution was performed with GSDB [16% goat serum (Sigma), 0.3% Triton X-100, 0.1% bovine serum albumin (Sigma), 0.45 M NaCl, and 20 mM NaH₂PO₄, pH 7.4 (6)]. Diluted primary antibodies [α_XG, 1:200; H⁺-K⁺-ATPase β-subunit, 1:50; anti-Na⁺-K⁺-ATPase β-subunit, 1:100; PDI (Dianova)] were incubated for 1 h at room temperature. After several washing steps using PBS⁺, secondary anti-mouse FITC- or Rhodamine Red-labeled and anti-rabbit TRITC- or fluorescein-labeled antibodies (1:100) were incubated with the samples for 1–6 h at room temperature. After additional washing steps, filters were mounted on coverslips with Vectashield (Vector Laboratories, Burlingame, CA). Confocal images were generated on a Zeiss model LSM 410 and on an Olympus Fluoview laser-scanning microscope. Images are the product of eightfold line averaging, andxz cross sections were generated with a 0.2-μm motor step.

⁸⁶Rb⁺ uptake measurements.

⁸⁶Rb⁺ uptake measurements with H⁺-K⁺-ATPase α- and gastric H⁺-K⁺-ATPase β-subunit transfected and untransfected MDCK wild-type cells were performed as previously described for transfected HEK-293 cells (21,22). ⁸⁶Rb⁺ uptake was determined in the presence of an extracellular Na⁺ concentration of 10 mM and in the presence of ouabain concentrations ranging from 0.1 nM and 1 mM. Data are reported as means ± SE.

RESULTS

The subcellular steady-state localization of the human ATP1AL1-ATPase was assessed in stably transfected MDCK and transfected LLC-PK₁ wild-type cells. We examined the assembly and plasma membrane expression of the transfected ATP1AL1 α-subunit with the endogenous Na⁺-K⁺-ATPase β-subunit by confocal immunofluorescence microscopy and surface biotinylation. Previous studies in our lab showed that the strong endogenous expression of Na⁺-K⁺-ATPase in MDCK and LLC-PK₁ cells provides sufficient pools of sodium pump β-subunit for assembly and targeting of transfected α-subunits (14, 18). Therefore, the epithelial cells were stably transfected only with the ATP1AL1 α-subunit, which is not endogenously expressed. To enhance the expression of the transfected α-subunit, clonal filter-grown cell lines were treated with sodium butyrate (10 mM) for 12 h. Colocalization of both proteins was monitored using an affinity-purified polyclonal ATP1AL1 antibody, a monoclonal Na⁺-K⁺- ATPase β-subunit antibody, and TRITC- and FITC-labeled secondary antibodies, respectively. Our confocal immunofluorescence results revealed no plasma membrane colocalization for these two subunits.

The en face (xy) view in Fig.1A revealed predominantly perinuclear ATP1AL1 staining (arrow in Fig. 1A), consistent with ATP1AL1 accumulation in the ER. Cross sections in thexz direction confirmed the intracellular ATP1AL1 distribution and the absence of colocalizations with the laterally concentrated Na⁺- K⁺-ATPase β-subunit (Fig.1B). The intracellular accumulation of ATP1AL1 was confirmed by colocalization analysis with an antibody against the protein disulfide isomerase (PDI), a marker of the ER. The transfected ion pump (Fig. 1C; fluorescein-labeled) and the ER marker (Fig.1D; Rhodamine Red-labeled) were localized in the perinuclear region. Merging a representative detail of both stainings (yellow-orange color) revealed partial colocalization in the ER and ER-derived vesicle structures (Fig. 1E). The same results were obtained by analyzing ATP1AL1 transfected LLC-PK₁cells (data not shown). Because the strong intracellular ATP1AL1 α-subunit accumulation could possibly mask a small degree of surface colocalization with the Na⁺-K⁺-ATPase β-subunit, we performed surface biotinylation experiments with ATP1AL1 transfected cells. We used only transfected MDCK cells, because LLC-PK₁ cells have been shown to be unsuitable for quantitative surface biotinylation analysis (19). Similar amounts of apically or basolaterally biotinylated proteins (Fig.2; lanes 1 and 2, respectively) were separated on polyacrylamide gels, and the corresponding Western blots were probed with ATP1AL1 and Na⁺-K⁺-ATPase β-subunit-specific antibodies. Although the ATP1AL1 α-subunit is highly enriched in intracellular compartments (as seen in Fig. 1), the α-subunit could not be found in either the apical or in the basolateral biotinylated membrane protein fraction. Only the Na⁺-K⁺-ATPase β-subunit (∼50–60 kDa) could be detected in the basolateral membrane (Fig.2, lane 2), which reflects its colocalization with the endogenous Na⁺-K⁺-ATPase α-subunit. Therefore ATP1AL1 does not reach the plasma membrane together with the β-subunit of the related Na⁺-K⁺- ATPase in MDCK and LLC-PK₁ cells when expressed by itself.

Based on previous studies, which showed interaction between ATP1AL1 and the β-subunit of the gastric H⁺-K⁺-ATPase (21, 22) in unpolarized HEK-293 cells, assembly and targeting of both subunits was tested in cotransfected MDCK cells. Double-transfected cells were cloned by functional selection. In contrast to untransfected or to monotransfected cells, in cells expressing ATP1AL1 and the gastric H⁺-K⁺-ATPase β-subunit the activity of the heteromeric ion pump enables the cells to grow in media containing up to 1 μM ouabain (21). Although the proliferation rate seems to be reduced during ouabain selection, the transfected ATPase is able to compensate for the inhibited endogenous Na⁺-K⁺-ATPase (K_i for ouabain ∼10⁻⁷ M).

To further document the presence of pump activity in cells expressing ATP1AL1 in association with the gastric H⁺-K⁺-ATPase β-subunit, we measured the ouabain sensitivity of ⁸⁶Rb⁺ uptake in cotransfected MDCK cells. ⁸⁶Rb⁺ uptake measurements (Fig. 3) were performed as previously described (21). In these studies it was shown that single transfection of the gastric H⁺-K⁺-ATPase β-subunit in HEK-293 cells does not alter the ouabain sensitivity and the ⁸⁶Rb⁺uptake activity of the endogenously expressed Na⁺-K⁺- ATPase. By extrapolating the data from HEK-293 cells to the MDCK system, we compared⁸⁶Rb⁺ uptake activity of three different ATP1AL1/gastric H⁺-K⁺- ATPase β-subunit cotransfected MDCK cell lines with untransfected MDCK wild-type cells. Cotransfected MDCK cells are characterized by threefold higher uptake in the presence of a ouabain concentration (1 nM) that does not influence the endogenous influx. Ouabain concentrations up to 1 μM inhibit the endogenous Na⁺-K⁺-ATPase activity in wild-type cells but have only a marginal effect on the activity of the ATP1AL1-ATPase, which is fully blocked only at ouabain concentrations exceeding 1 mM. The viability of ATP1AL1/gastric H⁺-K⁺-ATPase β-subunit transfected MDCK cells in the presence of micromolar ouabain concentrations and the ouabain sensitivity profile are in line with previous findings from transfected HEK-293 cells (21, 22). These data confirm functional ATP1AL1- ATPase expression in the MDCK renal epithelial cell line.

Confocal immunofluorescence analysis of the cotransfected MDCK cells reveals that the ATP1AL1 α-subunit accumulates exclusively at the apical plasma membrane, which is shown in the xz cross section of Fig. 4A. There is no intracellular ATP1AL1 α-subunit accumulation as found in single transfected cells (xz cross section in Fig. 1B). In parallel, the cotransfected gastric H⁺-K⁺-ATPase β-subunit was predominantly apical but also laterally localized (Fig. 4B andxz cross section). The subcellular localization of both subunits was independent of their expression level, as found by analyzing more than 10 independent, stable cotransfected and clonal MDCK cell lines. Merging both staining patterns revealed clear apical colocalization of ATP1AL1 and the H⁺-K⁺-ATPase β-subunit (Fig. 5, top). The additional lateral β-subunit staining (Fig. 5, top; FITC) is due to the high β-subunit expression level and reflects the distribution of unassembled H⁺-K⁺-ATPase β-subunit protein. This result is in line with previous findings that documented gastric H⁺-K⁺-ATPase β-subunit targeting to the basolateral membrane of singly transfected MDCK cells (37). In contrast, in the same ATP1AL1/gastric H⁺-K⁺- ATPase β-subunit doubly transfected cells, the endogenous Na⁺-K⁺-ATPase β-subunit (FITC) was entirely basolateral and showed no overlapping distribution with the apically sorted ATP1AL1 protein (TRITC), confirming the lack of plasma membrane colocalization of these two polypeptides (Fig. 5,bottom). It is interesting to note that, despite the fact that we generated clonal cell lines, the expression level of ATP1AL1 and the gastric H⁺-K⁺-ATPase β-subunit varied considerably among neighboring cells (xy section in Fig. 4,A and B). Similar results have been obtained in transfections of MDCK cells with other transport proteins (T. R. Muth, personal communication). We assume that the heterogenous expression levels are due to differences in the transcription and/or translation efficiencies of both proteins. Nevertheless, the steady-state polarization of the transfected α- and β-subunits were shown to be independent of their protein expression level as found by analyzing more than 10 different clonal cell lines.

Further confirmation of our immunofluorescence data was obtained through surface biotinylation of ATP1AL1 α-subunit and gastric H⁺-K⁺-ATPase β-subunit transfected MDCK cells (Fig. 6). The ATP1AL1 α-subunit (∼110 kDa) was present in the apical membrane (Fig. 6A), whereas the Na⁺-K⁺-ATPase β-subunit was not found in the apically biotinylated plasma membrane protein fraction (Fig.6B). This results were reproducible with four different cotransfected MDCK cell lines. Western blots incorporating serial dilutions of biotinylated proteins for purposes of quantitation indicated that the steady-state distribution of the ATP1AL1 protein was ∼80% apical and 20% basolateral (not shown). This ratio is in line with the polarity ratios determined by other groups for apically polarized marker proteins in MDCK cells, such as the influenza virus hemagglutinin (31). Contrary to the predominantly apical ATP1AL1 α-subunit expression, the gastric H⁺-K⁺-ATPase β-subunit seems to be equally distributed in both plasma membrane domains (Fig. 6A). Taken together, the biotinylation and immunofluorescence data clearly show that ATP1AL1 is predominantly polarized to the apical plasma membrane of transfected MDCK cells.

Surprisingly, cotransfected LLC-PK₁ cells showed a completely different behavior. Both the ATP1AL1 subunit (Fig.7, A and B) and the gastric H⁺-K⁺-ATPase β-subunit (Fig.7C) accumulated in the lateral plasma membrane. Although technical limitations precluded the acquisition of quantitative biotinylation data, the analysis of six different transfected LLC-PK₁ cell clones strongly supports this result. Previous studies on LLC-PK₁ cells in our lab showed that the gastric H⁺-K⁺-ATPase β-subunit is differentially polarized to the apical membranes of LLC-PK₁ cells and to the basolateral membranes of monotransfected MDCK cells (37). Based on these results, the basolateral localization of the gastric H⁺-K⁺-ATPase β-subunit in ATP1AL1-expressing LLC-PK₁ cells and its apical localization in MDCK cells provides strong evidence for sorting signals residing on the ATP1AL1 α-subunit playing a dominant role in the sorting of the complex.

DISCUSSION

The human P-type ATPase, ATP1AL1, belongs to the subgroup of nongastric H⁺-K⁺-ATPases and is involved in K⁺ reabsorption in various tissues, including colon and kidney. Functional characterizations reveal that ATP1AL1 and the related rat colonic nongastric H⁺-K⁺-ATPase mediate K⁺ reabsorption in exchange for either sodium ions or protons. In addition to their functional properties, the establishment and regulation of the spatial distributions of nongastric H⁺-K⁺-ATPases in polarized renal epithelial cells must play a critical role in determining their physiological activity. Little is known of the cell biologic properties of this pump subfamily. It is known that the nongastric H⁺-K⁺-ATPases undergo upregulation after K⁺ depletion (for review see Ref. 27) and that their functional plasma membrane expression is modulated by endocytosis (46). However, the molecular signals and pathways responsible for this behavior in renal epithelial cells remain to be determined.

Although five functionally distinguishable K⁺-dependent K⁺-ATPase activities are detected in renal epithelial cells (for review see Refs. 27 and 46), the cellular distributions of only two, the gastric H⁺-K⁺-ATPase and the nongastric colonic H⁺-K⁺-ATPase, have been investigated. The gastric subtype was identified in the apical membranes of α-type intercalated cells (IC), in PC, and in the basolateral membranes of rat β-type ICs (2,46), whereas the nongastric colonic α-subunit protein was apically localized in renal PCs and in the colonic epithelium (29, 38). Studies of the distributions of H⁺-K⁺-ATPase isoforms have been complicated by the possible existence of multiple species-specific subtypes as well as by low expression levels of the nongastric ATPases under normal circumstances. Previous studies showed that the mRNA of the human ATP1AL1 ion pump, for example, was only detectable in sensitive RT-PCR experiments (20, 34). Furthermore, although this pump may be upregulated in response to K⁺ deprivation, this effect cannot be practically exploited to examine its distribution in its native human renal tissue. To examine the subcellular targeting and functional activity of ATP1AL1, we expressed this ATPase by stable transfection in MDCK and LLC-PK₁ cells. These cell lines manifest characteristics of cells from the distal nephron and the proximal nephron segments, respectively, and thus are suitable expression systems for H⁺-K⁺-ATPase characterization.

We have previously shown that the strong expression of the endogenous Na⁺-K⁺-ATPase in these cells provides sufficient Na⁺-K⁺-ATPase β-subunit to support the assembly, proper folding and targeting of exogenous sodium pump and sodium pump chimeras expressed by transfection (14,18). Nevertheless, our immunofluorescence and biotinylation data demonstrate that in both transfected MDCK and LLC-PK₁ cells, the ATP1AL1 α-subunit expressed by itself does not reach the plasma membrane. This result corresponds to previous findings that documented poor assembly between ATP1AL1 and the Na⁺-K⁺-ATPase β-subunit in in vitro expression studies (21, 35). The present data further demonstrate, therefore, that the Na⁺ pump β1-subunit is not an effective partner for ATP1AL1. Instead, ATP1AL1 once again exhibits a strong preference for functional assembly with the gastric H⁺-K⁺-ATPase β-subunit (21). In this context it is interesting to note that the related rat colonic H⁺-K⁺-ATPase protein and its guinea pig ortholog appear to associate with the Na⁺pump β1-subunit and the gastric H⁺-K⁺-ATPase β-subunit, both in vivo and in Xenopus oocyte expression studies (1, 9, 28). Previous elegant studies identified in an extracellular loop of the Na⁺ pump, of the gastric H⁺-K⁺-ATPase, and in the colonic H⁺-K⁺-ATPase a stretch of 26 amino acids that is highly conserved and plays a major role in β-subunit specificity (30, 33). Amino acid residues that were shown to prefer assembly between the colonic H⁺-K⁺-ATPase and the gastric H⁺-K⁺-ATPase β-subunit (46) are identical in the human ATP1AL1 α-subunit. Whether the shown β-subunit specificity is a property unique among the nongastric H⁺-K⁺- ATPases to the ATP1AL1 subtype or whether it is due to species-specific preferences remains to be determined.

We find that ATP1AL1 is differentially sorted to opposite membrane domains when expressed in transfected MDCK or LLC-PK₁cells. The apical polarization of ATP1AL1 in MDCK cells corresponds to predictions based on physiological and pharmacological studies, which reveal the presence of functionally similar ATPase activities in the luminal membranes of outer medulla and cortical collecting duct cells after K⁺ deprivation (4). The basolateral distribution observed in LLC-PK₁ cells lacks similar functional correlations. It should be noted, however, that at least the related gastric H⁺-K⁺-ATPase α-subunit protein has also been detected basolaterally in situ (2). Thus it is possible that these pumps are differentially sorted by the multiple epithelial cell types that line the nephron.

The distinct ATP1AL1 localization patterns raise the question as to which molecular signals are responsible for the disparate steady-state distributions of this protein in MDCK and LLC-PK₁ cells. Since the gastric H⁺-K⁺-ATPase β-subunit accumulates basolaterally in monotransfected MDCK cells (37) but is colocalized to the apical membrane after cotransfection with ATP1AL1 and vice versa in LLC-PK₁cells, the β-subunit cannot contain the dominant sorting signals responsible for the distribution of the holoenzyme complex. It has been shown that N-linked glycosylation can play a functional role in apical targeting of transmembrane proteins (24, 46). It is conceivable, therefore, that cell-type-specific glycosylation of the gastric H⁺-K⁺-ATPase β-subunit could influence the cellular distribution of the complex in different epithelial cell types. To test this possibility, we blocked N-linked glycosylation in transfected MDCK cells using tunicamycin (20 μM for 12 h). Unglycosylated β-subunit and ATP1AL1 were still strictly polarized to the apical membrane, excluding a functional role for N-linked glycosylation in apical sorting of ATP1AL1 (data not shown). Furthermore, association of ATP1AL1 with apically sorted, detergent-insoluble glycosphingolipid-rich membrane domains (for review see Ref. 3) appears not to occur in either cell type based on detergent solubilization and density gradient centrifugation experiments (data not shown).

Based on these observations, we suggest that the ATP1AL1 α-subunit embodies plasma membrane sorting information that is differentially recognized by cell type-specific mechanisms. Similar conclusions have also been drawn in studies of the gastric H⁺-K⁺- ATPase β-subunit (37) and the human dopamine transporter expressed in MDCK and LLC-PK₁ cells (23). Very recently it was shown that LLC-PK₁ cells lack the epithelial cell-specific clathrin adaptor subunit μ1B, which can recognize basolateral sorting signals (15). In LLC-PK₁ cells, the absence of μ1B leads to apical sorting of some basolateral membrane proteins. In context with the basolateral ATP1AL1 α-subunit/gastric H⁺-K⁺-ATPase β-subunit localization in LLC-PK₁ cells and the apical distribution of H⁺-K⁺-ATPase β-subunit in single transfected LLC-PK₁ cells (37), μ1B cannot be involved in the basolateral steady-state localization at least of the functional holoenzyme complex.

Comparison of the sequences of the apical gastric H⁺-K⁺-ATPase α-subunit, the basolateral Na⁺-K⁺- ATPase α-subunit, and ATP1AL1 may shed some light on the molecular nature of the sorting information. A signal sufficient to ensure the apical sorting of the gastric H⁺-K⁺-ATPase resides within the fourth transmembrane domain (TM4) of this pump's α-subunit (14). The sequence of the gastric H⁺-K⁺-ATPase and Na⁺-K⁺-ATPase α-subunit TM4s differ by only 8 amino acids. The residues associated with apical pump sorting are not conserved in ATP1AL1 and, instead, are highly homologous to the corresponding sequence of the basolaterally sorted Na⁺-K⁺-ATPase. Presumably, the Na⁺-K⁺-ATPase-like TM4 sequence of ATP1AL1 is recognized by cell type-specific sorting mechanisms leading to this protein's basolateral delivery in LLC-PK₁ cells. According to this model, additional or distinct information must be responsible for this protein's apical distribution in MDCK cells. The preparation and expression of chimeric pump constructs will be necessary to test this hypothesis.

FOOTNOTES

• These studies were supported by the National Institutes of Health Grants DK-17433 and GM-42136. J. Reinhardt was supported by a postdoctoral fellowship of the Deutsche Forschungsgemeinschaft (Re1284).

• Address for reprint requests and other correspondence: M. Caplan, Dept. of Cellular and Molecular Physiology, Yale Univ., School of Medicine, New Haven, Connecticut 06520 (E-mail:michael.[email protected]edu).

• 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. §1734 solely to indicate this fact.