INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

THE BIOSYNTHETIC ASSEMBLY of the Na-K-ATPase, the best-characterized member of the P-type ion transporters, has been extensively studied (see Refs. 8, 16, and 33 for review). The minimal functional unit of the Na-K-ATPase consists of a polytopic -subunit and a glycosylated, single membrane-spanning -subunit. Additional tools for the investigation and characterization of interactions between - and -subunits in ATPases have been provided by the identification and characterization of a -subunit from another member of the P-type ATPases, the gastric H-K-ATPase (5, 21, 36, 37, 41), and of homologs of the -subunit of the gastric H-K-ATPase expressed in colon (10, 11) and kidney cells (4, 24, 28).

The -subunit is typically cited as the catalytic subunit for the H-K-ATPase and Na-K-ATPase, but enzymatic function is dependent on the interaction between the respective - and -subunits (7, 15, 23). Therefore, identification of subunit interactive sites is important to understand the functional cycle of these cation-exchange ATPases. The association between subunits is stable to mild detergents, such as Triton X-100 or Nonidet P-40, whereas the forces of interaction are clearly disrupted by denaturing detergents, such as dodecyl trimethyl ammonium bromide (DTAB) or SDS. These detergent-sensitive properties were first used to demonstrate the existence of a -subunit for the H-K-ATPase (HK) (36) and have also been exploited to identify specific segments of the -subunit that interact with the -subunit (40). Recently, Wang et al. (43), using a heterologous yeast expression system to monitor assembly of transfected chimeric ATPase subunits, found a region in an extracellular domain of the -subunit of the H-K-ATPase (HK) spanning Gln-905 to Val-930 that interacts with the extracellular domain of HK. In addition, Melle-Milovanovic et al. (34) used the yeast two-hybrid system to probe for sites of interaction between - and -subunits. They reported two regions in the ectodomain of the -subunit, Gln-64 to Asn-130 and Ala-156 to Arg-188, as containing sites of interaction with the -subunit. Geering et al. (17) demonstrated that a sequence motif Y₂₄₂Y(or F)PYY in the ectodomain of the Na-K-ATPase -subunit (NaK) is conserved among all -subunits, and the presence of P₂₄₄ in this motif is essential for assembly of Na-K-ATPase - and -subunits. Moreover, consistent with the results of Wang et al., Melle-Milovanovic et al. showed that the region Arg-898 to Arg-922 in the -subunit interacts strongly with the extracytoplasmic domain of the -subunit. The ability to form hybrid - heterodimers between the Na-K-ATPase and the H-K-ATPase has provided significant insight into not only the regulation of assembly of native - heterodimers but also the role of the -subunit in the modulation of transport activities of the holoenzyme (15, 19, 22, 23, 30).

We previously reported on two monoclonal antibodies (MAbs) against HK: MAb 2G11, which bound to the native enzyme at a site within the cytoplasmic NH₂-terminal region, and MAb 2/2E6, which bound within the large extracytoplasmic segment, but only when the enzyme was denatured (7). An objective for the present work was to test the hypothesis that MAb 2/2E6 recognizes a specific site of interaction between - and -subunits and to identify the epitope within HK.

Another objective of this work was to use the MAbs against topologically distinct regions of HK to study the regulation of differential assembly of HK and its impact on plasma membrane delivery of P-type ATPases in epithelial cells, a topic that has received relatively little attention (6). Investigation of this phenomenon is critically important to understanding the physiology of transport across epithelia, inasmuch as the Na-K-ATPase is responsible for the maintenance of transepithelial ion gradients that, in turn, drive the transepithelial transport of other ions and nutrients. Moreover, in the gastric acid-secreting oxyntic cell, the mechanism by which the Na-K-ATPase and the H-K-ATPase are separately targeted to the basolateral or apical membranes, respectively, must be precise, notwithstanding the high degree of homology between the - and -subunits of the pump enzymes (6). A similar process may exist in colon and kidney. When expressed in Xenopus, HK has been shown to act as a surrogate for conveying Na⁺-pumping functions for the -subunit of the Na-K-ATPase (NaK) (23). However, the situation appeared to be quite different when Gottardi and Caplan (18) transfected HK in a mammalian epithelial cell line, LLC-PK₁ cells. These authors reported that the expressed HK was exclusively targeted to the apical plasma membrane and endogenous Na-K-ATPase was targeted to its standard basolateral location, with very poor efficiency of HK/NaK assembly or surrogation of activity. In the present work, HK was transfected into two polarized kidney cell lines, MDCK cells and LLC-PK₁ cells, and the biosynthesis, assembly, and cell surface expression of HK were investigated.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Materials. All chemicals were reagent grade. The anti-HK MAbs 2/2E6 and 2G11 have been previously described (7). The MAb against HK was the kind gift of Dr. Adam Smolka (Medical University of South Carolina, Charleston, SC). The polyclonal anti-Na-K-ATPase antibodies were a kind gift from Dr. W. James Nelson (Stanford University, Palo Alto, CA). The anti-NaK MAb 6H (18) was obtained as a hybridoma supernatant from Dr. Alicia McDonough (University of Southern California, Los Angeles, CA). Prestained molecular weight and secondary antibodies coupled to horseradish peroxidase (HRP) were purchased from Bio-Rad (Hercules, CA). Pro-mix L-[³⁵S]amino acids in vitro cell labeling mix, streptavidin coupled to HRP, enhanced chemiluminescence (ECL) kit, protein A-Sepharose, protein G-Sepharose, and Sepharose CL-2B were purchased from Amersham Pharmacia Biotech. Trans-³⁵S-label in vitro labeling mix was purchased from ICN (Irvine, CA). Sulfosuccinimidobiotin (sulfo-NHS-biotin) was purchased from Pierce (Rockford, IL). MEM was purchased from Mediatech (Washington, DC). Cys- and Met-free DMEM, MEM with Hanks' salts, clostripain, and streptavidin-agarose were purchased from Sigma-Aldrich. Penicillin, streptomycin, amphotericin B (Fungizone) cocktail, and Hanks' balanced salt solution (HBSS) were purchased from the Cell Culture Facility at the University of California, San Francisco, Bio-Whittaker (Washington, DC), or Mediatech. FBS was purchased from Hyclone Laboratories (Logan, UT). G418 and ¹⁴C-labeled molecular weight markers were purchased from Life Technologies (Bethesda, MD). Endoglycosidase H (Endo H) and peptide N-glycosidase F (PNGase F) were purchased from Boehringer Mannheim. Lysyl endopeptidase C (Lys C) was obtained from Wako Chemical (Osaka, Japan). 4-(2-Aminoethyl)benzenesulfonyl fluoride (AEBSF) · HCl was purchased from Calbiochem-Novachem (La Jolla, CA). Pepstatin, leupeptin, chymostatin, antipain, and benzamidine were purchased from Chemicon (Temecula, CA). Wheat germ agglutinin (WGA) coupled to agarose was purchased from E-Y Laboratories (San Mateo, CA).

Immunohistochemistry of gastric glands and cell cultures. Gastric glands were isolated as previously described (1). The glands were fixed in 3.7% formaldehyde in PBS for 30 min at room temperature, then they were permeabilized by 1% Triton X-100 in PBS for 15 min. The fixed, permeabilized glands were immunostained directly or subjected to further treatment with denaturants before they were immunostained. Treatment with urea or detergents was carried out at room temperature, and the treated glands were subsequently washed with PBS before incubation with antibodies. The treated cells were incubated for 1 h with MAbs against HK (MAb 2/2E6 or MAb 2G11), washed three times in PBS, and incubated with fluorescent-labeled anti-mouse IgG antibody for 1 h. Previous work has shown that the epitope on HK for MAb 2/2E6 is in the extracellular domain and that for MAb 2G11 is in the NH₂-terminal cytoplasmic tail (7).

Biochemical characterization of - and -subunit interaction. H-K-ATPase-enriched gastric microsomes were prepared as previously described (29). As a general protocol, the microsomes were solubilized for 1 h at room temperature in 1% Triton X-100 in suspending medium (SM) consisting of 300 mM sucrose, Tris · HCl (pH 7.4), and 0.5 mM EDTA. The solubilized materials were applied to different affinity matrices before or after the H-K-ATPase was subjected to several modifying procedures.

Phage display peptide library kit. The heptapeptide display library phage kit was purchased from New England Biolabs. Phage selection, amplification, and DNA extraction were carried out according to manufacturer's instructions. Using polystyrene petri dishes coated with 2/2E6 antibody as selection plates, we carried out four rounds of selection (biopanning) to choose a pool of phages. Twenty individual clones were isolated, amplified, and analyzed with Western blot for 2/2E6. Seven of them showed that their minor coat protein strongly reacted with 2/2E6. The corresponding DNA sequences were sequenced, and the amino acids were deduced. Synthetic peptides were synthesized by BioMed. Competitive Western blots were carried out with primary antibody preincubated with the synthetic peptide for 1 h.

Construction of HK clones and transfection. cDNA encoding rat HK (5) was excised from the pBluescript SK() vector (Stratagene) with Apa I and Avr II and subcloned into pCB6, a vector in which the expression of the insert cDNA is driven by a cytomegalovirus promoter (2). The Apa I-digested end was filled in with Klenow and blunt-end ligated into the unique Hind III site of pCB6; the Avr II-cut end was ligated into the Xba I site. MDCK strain II cells, which are of European Molecular Biology Lab parentage and are the clone that delivers newly synthesized Na-K-ATPase exclusively to the basolateral membrane, were transfected with this plasmid by the calcium phosphate method, as described previously (35). After selection in G418, four positive clones were identified, expanded, and maintained in MEM supplemented with 5% FBS and antibiotic-antimycotic. The polarity of these four clones was verified by assaying for polarized secretion of the endogenous MDCK glycoprotein gp80, as described previously (42). The apical-to-basal ratio of gp80 secretion in all the clones described here was 4:1. The LLC-PK cells used in this study, stably transfected with HK (18), were the kind gift of Dr. Michael Caplan (Yale University).

Cell surface biotinylation. Cells grown on Costar 24-mm Transwell polycarbonate filters were cooled to 4°C with several washes of cold HBSS supplemented with 20 mM sodium HEPES, pH 7.4. The apical or basolateral surface of transfected MDCK cells was biotinylated with 0.5 mg/ml sulfo-NHS-biotin in HBSS with two 15-min incubations at 4°C. After biotinylation, cells were washed quickly three times with cold HBSS and incubated for 15 min at 4°C with cold MEM supplemented with 0.6% BSA, 20 mM sodium HEPES, and 50 mM glycine to quench any remaining unreacted biotin. The cells were lysed under denaturing conditions (0.5% SDS, 100 mM NaCl, 50 mM triethanolamine · HCl, 5 mM EDTA, 0.1 mM AEBSF, and 0.2% NaN₃, pH 8.1) or nondenaturing conditions [2.5% (wt/vol) Triton X-100, 100 mM NaCl, 100 mM triethanolamine · HCl, 5 mM EDTA, 0.1 mM AEBSF, 0.2% NaN₃, 5 µg/ml each of pepstatin and leupeptin, 10 µg/ml each of chymostatin and antipain, and 0.5 mM benzamidine]. After cell lysates were precleared with Sepharose CL-2B, HK was immunoprecipitated with MAb 2/2E6-protein A-Sepharose (specific for the ectodomain of HK) or MAb 2G11-protein G-Sepharose (specific for the cytoplasmic domain of HK), and biotinylated proteins were isolated with streptavidin-agarose. Immunoprecipitates were run on SDS-PAGE, transferred to nitrocellulose, and probed with MAb 2/2E6 to detect HK, anti-NaK monoclonal or polyclonal antibodies to detect NaK, or streptavidin-HRP to detect biotinylated proteins. After incubation with appropriate secondary antibodies coupled to HRP (if necessary), the reactive proteins were visualized by ECL. Signals from ECL blots were quantitated on a Bio-Rad GS-670 imaging densitometer.

Pulse-chase analysis of HK. Transfected MDCK or LLC-PK₁ cells grown on 24-mm Transwell plates were washed with PBS containing Ca²⁺ and Mg²⁺ and starved for 15 min at 37°C with Cys- and Met-free DMEM supplemented with 5% dialyzed FBS, 20 mM sodium HEPES (pH 7.4), and antibiotic-antimycotic. Cells were pulse labeled with 50 µCi of Pro-mix (New England Nuclear) or Trans-³⁵S-label (ICN, Irvine, CA) L-amino acids from the basolateral surface and chased for various periods. Cells were lysed under denaturing or nondenaturing conditions, as described above. Lysates were processed for immunoprecipitation, as described above. Immunoprecipitates were run on SDS-PAGE and fluorographed. In some cases, immunoprecipitates of pulse-labeled HK were digested with Endo H before SDS-PAGE and fluorography. The immunoprecipitated material was resuspended by boiling in 0.1 M sodium citrate (pH 6.0) and 1% SDS; Endo H was added to 3 mU, and the material was incubated overnight at 37°C.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Immunostaining of parietal cells by 2/2E6 requires denaturation. When we first introduced MAb 2/2E6, we pointed out that immunostaining did not occur with conventional formaldehyde fixation and permeabilization by Triton X-100 but "required alcohol delipidation of the glands." Figure 1 shows that H-K-ATPase of Formalin-fixed parietal cells is immunostained by MAb 2/2E6 only when a denaturant is applied (e.g., alcohol, SDS, DTAB, alkali). A titration with urea as a denaturant is shown in Fig. 2. For gastric glands fixed by Formalin and permeabilized with Triton X-100, immunostaining by 2/2E6 was barely evident up to 4.5 M urea and became uniformly apparent at 5 M urea. These data demonstrate that the epitope for 2/2E6 is not simply related to delipidation but, rather, is buried and becomes exposed only after denaturation.

View larger version (133K): [in this window] [in a new window]   Fig. 1.   H-K-ATPase of Formalin-fixed parietal cells is immunostained by monoclonal antibody (MAb) 2/2E6 only when a denaturant is applied. Gastric glands were fixed with Formalin and then subjected to a series of permeabilization procedures as follows: 1% Triton X-100 for 15 min (TX-100), 1% Triton X-100 for 15 min followed by a brief exposure to methanol (TX-100 + MeOH), 0.5% dodecyl trimethyl ammonium bromide for 15 min (DTAB), and 0.5% SDS for 15 min (SDS). Glands were then probed with MAb 2/2E6 followed by secondary FITC-labeled anti-mouse antibody and examined by fluorescence (top) and differential interference (bottom) microscopy. All fluorescent micrographs were captured with identical times of exposure. Scale bar, 20 µm.

View larger version (33K): [in this window] [in a new window]   Fig. 2.   Immunostaining by MAb 2/2E6 occurs when H-K-ATPase is denatured by urea. Formalin-fixed gastric glands were permeabilized with 1% Triton X-100 and subjected to titration with 0-5.5 M urea before they were probed with MAb 2/2E6. All fluorescent micrographs were captured with identical times of exposure. Scale bar, 20 µm.

Dissociation of - and -subunits exposes the 2/2E6 epitope. We previously reported that - association was stable in nonionic detergents and disrupted by ionic detergents, such as DTAB or SDS (36). The unmasking of the 2/2E6 epitope by DTAB or SDS in fixed gastric glands further supported an epitopic site within the sphere of - interaction. To investigate the action of SDS on the - association and unmasking of 2/2E6 epitope, we treated purified H-K-ATPase-containing membranes with various amounts of SDS, then added Triton X-100 to mop up the excess SDS. The treated samples were passed through a 2/2E6-protein A affinity column. The bound materials were eluted with 2% SDS and analyzed by SDS-PAGE and Western blot, as shown Fig. 3. For samples where the membranes were treated with 0.3% SDS, HK bound to the 2/2E6 column, demonstrating that these levels of SDS were sufficient to expose the binding site.

View larger version (62K): [in this window] [in a new window]   Fig. 3.   SDS dissociates subunits of H-K-ATPase and exposes epitope for MAb 2/2E6 on -subunit. H-K-ATPase-rich vesicles were treated with 0-1% SDS. An excess of Triton X-100 (10-fold volume dilution in 2% Triton X-100) was added to all samples, which were then passed through a 2/2E6-protein A affinity column. Bound materials were eluted with 2% SDS, separated by SDS-PAGE, and developed by silver staining. A significant quantity of -subunit was adsorbed by 2/2E6 column only when original treatment with SDS was 0.3%. Right lane shows a sample of control, untreated microsomes run of same gel, where relative amounts of - and -subunits can be seen in starting membrane material. mw, Molecular mass (in kDa).

View larger version (81K): [in this window] [in a new window]   Fig. 4.   SDS separates - and -subunits and exposes 2/2E6 epitope. H-K-ATPase was solubilized in 1% Triton X-100 and immobilized on a series of wheat germ agglutinin (WGA) columns. Individual columns were exposed to graded amounts of SDS, and respective eluants were probed by Western blotting for -subunit (effluent HK). MAb 2/2E6 was subsequently passed through columns, and, after thorough washing, bound 2/2E6 was released by 2% SDS and probed with horseradish peroxidase (HRP)-labeled goat anti-mouse antibody (2/2E6 bound). Locations of molecular mass standards are indicated on right. Densitometric quantitation of blots (bottom) reveals that release of effluent HK by SDS was roughly parallel to exposure of 2/2E6 epitope on WGA-bound -subunit of H-K-ATPase (HK). Values are expressed relative to %bound at 1% SDS.

The 2/2E6 epitope is identified as S₂₂₆LHY₂₂₉. The heptapeptide phage display library was used as described in METHODS AND MATERIALS to screen the peptide sequence corresponding to the epitope for MAb 2/2E6 binding. Seven individual clones of the phage were selected for amplification and DNA sequencing. The corresponding peptide sequences are shown in Table 1. Inspection of these sequences indicates that the common sequence for the epitope is SXHY. The most likely site is SLHY, corresponding to amino acids 226-229 of the rabbit HK sequence.

View larger version (41K): [in this window] [in a new window]   Fig. 5.   The 16-mer peptide corresponding to amino acids 224-239 of rabbit gastric -subunit competes with MAb 2/2E6 in binding to HK. Culture medium containing 2/2E6 was mixed with 0-50 µg of peptide, and mixtures were used for Western blots of H-K-ATPase-rich microsomes. Broad band between 60 and 80 kDa is strongly visible in control with no peptide present, whereas signal was correspondingly decreased as peptide concentration was increased.

MAb 2G11 interacts with the NH₂-terminal eight amino acids of the -subunit. Earlier work showed that MAb 2G11 binds to the cytoplasmic domain of HK (7). Through the use of site-specific proteases, we now provide a more definitive binding site. Figure 6, A and B, shows that, after digestion of microsomal vesicles with Arg-specific clostripain and subsequent deglycosylation with PNGase F, distinctly different peptides are recognized by our two antibody probes: a peptide of ~30 kDa is recognized by MAb 2/2E6 and a peptide of ~2 kDa is recognized by MAb 2G11. The NH₂-terminal cytoplasmic domain of the rabbit HK has potential Arg cleavage sites at Arg-13 and Arg-18, which would be predicted to yield NH₂-terminal peptides of ~2.2 and 1.5 kDa, respectively, as indicated by the following sequence: MAALQEK₇K₈SCSQR₁₃MEEFR₁₈HYCWN PDT. . . . .

View larger version (41K): [in this window] [in a new window]   Fig. 6.   Epitope for MAb 2G11 is within 8 most NH2-terminal amino acids of HK. H-K-ATPase-rich microsomal vesicles (100 µl containing 0.2 mg of protein) were digested with 10 µg of clostripain (Arg-C) or lysyl endoproteinase C (Lys-C) for 16 h at 37°C. Digestions were terminated by boiling in 0.5% SDS for 5 min. Samples were diluted 10- to 20-fold, 5 U of peptide N-glycosidase F (PNGase F) were added to deglycosylate, and samples were incubated at room temperature overnight. Samples were separated on 10% or 12% tricine gels and blotted to nitrocellulose. A: digestion by Arg-specific Arg-C produced a large -subunit fragment of ~30 kDa recognized by MAb 2/2E6 and a small fragment of ~2 kDa recognized by MAb 2G11. Lys-C digestion also produced a similar ~30-kDa fragment recognized by MAb 2/2E6, but it completely eliminated any peptide recognition by MAb 2G11. B: 2nd experiment of Arg-C digestion, but because of incomplete deglycosylation by PNGase F, some faint staining with 2/2E6 (but not with 2G11) can be seen at 60-80 kDa for fully glycosylated -subunit.

Immunostaining of HK expressed in LLC-PK₁ cells. Gottardi and Caplan (18) stably transfected a kidney cell line, LLC-PK₁ cells, with a gene for HK and subsequently showed that HK is synthesized in the endoplasmic reticulum (ER) and sorted to the apical plasma membrane without an accompanying -subunit. When we employed this same stably transfected LLC-PK₁ cell line, we observed a strong reaction of MAb 2/2E6 with the apical plasma membrane, even when the antibody was added to live cells (Fig. 7A). That is, neither fixation nor permeabilization was required for surface staining of the 2/2E6 epitope, although these treatments, without denaturation, were effective in exposing additional -subunit sites within an intracellular membrane compartment(s) (Fig. 7B). On the other hand, the MAb 2G11 epitope is exposed after only permeabilization of transfected cells (Fig. 7, C and D). Thus, in the LLC-PK₁ expression system, reaction of HK with 2/2E6 did not require denaturation, and since HK was not being coexpressed, the results are consistent with the interpretation that the epitopic site for MAb 2/2E6 is in a region of - association.

View larger version (174K): [in this window] [in a new window]   Fig. 7.   HK on cell surface of LLC-PK1 cells is immunostained by MAb 2/2E6 without fixation or permeabilization. Cultures of LLC-PK1 cells, stably transfected with HK, were probed with MAb 2/2E6 (A and B) or MAb 2G11 (C and D). In A and C, live cells were probed, washed, and labeled with secondary antibody. MAb 2/2E6 (A) clearly has access to -subunit on plasma membrane; MAb 2G11 (C) does not. In B and D, cells were fixed with Formalin and permeabilized with Triton X-100 before they were probed. Labeling can be seen for both antibodies, on surface and within cells. Scale bar, 10 µm for all low-power panels and 5 µm for high-power insets.

Immunostaining of HK in transfected MDCK cells. A second cultured epithelial cell system in which to study the expression and trafficking of HK was generated by expressing HK in MDCK cells (MDCK-HK cells). MDCK-HK cells were fixed in formaldehyde and stained with MAb 2/2E6 without (Fig. 8A) or with (Fig. 8B) postfixation treatment with methanol. Surprisingly, unlike the case with transfected LLC-PK₁ cells (Fig. 7), but similar to that seen for isolated rabbit gastric glands (Fig. 1), methanol or high concentrations of urea are required after fixation to effect staining of HK in MDCK-HK cells with MAb 2/2E6. Also, in contrast to data generated from the LLC-PK₁ model in this study and from Gottardi and Caplan (18), attempts to surface immunolabel HK by prebinding MAb 2/2E6 were unsuccessful (data not shown), as were attempts to surface immunoprecipitate metabolically radiolabeled HK (data not shown). These results imply that, in transfected MDCK cells, access of MAb 2/2E6 to its extracellular epitope on HK is restricted unless HK is further denatured. These results parallel those obtained with native H-K-ATPase in glands or isolated gastric microsomes. In the case of oxyntic cell membranes, access may be limited by association of HK with HK; however, because HK alone was transfected into MDCK-HK cells, another mechanism by which the MAb 2/2E6-binding epitope is masked must exist in MDCK-HK cells.

View larger version (156K): [in this window] [in a new window]   Fig. 8.   Immunostaining of HK-transfected Madin-Darby canine kidney (MDCK) cells requires denaturation to expose epitope for MAb 2/2E6. Cultures of transfected MDCK cells were fixed in formaldehyde and permeabilized with 1% Triton X-100. After fixation and permeabilization, cells were directly incubated with MAb 2/2E6 and then with FITC-labeled secondary antibody (A) or briefly exposed to methanol before immunostaining with 2/2E6 (B). C: cells directly probed with MAb 2G11 after fixation and permeabilization. D: cells briefly exposed to methanol before they were probed with MAb 2G11. Scale bar, 10 µm.

HK is expressed at the cell surface of MDCK-HK cells. Analysis of the distribution of anti-HK immunofluorescence in MDCK-HK cells suggested that HK may be present at the cell surface. To assay for HK at the cell surface of transfected MDCK cells, the apical and basolateral plasma membranes of MDCK-HK cells were separately biotinylated. Cell lysis was performed under denaturing conditions, and lysates were incubated with streptavidin-agarose to precipitate biotinylated proteins. Western blots of biotinylated proteins were probed with MAb 2/2E6. As seen in Fig. 9A for two different clones of MDCK-HK cells, HK appears to be expressed at the cell membrane, and cell surface expression of HK is apparently predominantly apical. In addition, in SDS gels, biotinylated HK migrates as a broad band of apparent molecular mass of 60-80 kDa, indicating that its oligosaccharide chains have been modified from the high-mannose core oligosaccharides (18, 23, 30). Alternatively, cell surface HK was biotinylated, and cell lysis was performed under denaturing conditions. HK was immunoprecipitated with anti-HK cytoplasmic domain MAb 2G11, and the immunoprecipitate was probed on Western blots with streptavidin-HRP. By this approach, cell surface HK was also observed to be predominantly apical (data not shown).

View larger version (48K): [in this window] [in a new window]   Fig. 9.   Biotinylation of HK at cell surface of transfected MDCK cells. A: apical (a) or basolateral (b) membranes of 2 clones of transfected MDCK cells were biotinylated. Cell lysis was performed under denaturing conditions. Biotinylated HK was precipitated with streptavidin-agarose and detected on Western blots with MAb 2/2E6 and enhanced chemiluminescence (ECL). B: association of HK with an endogenous 94-kDa protein at cell surface. After cell surface biotinylation, cell lysates were prepared under nondenaturing conditions. HK was immunoprecipitated with MAb 2G11. Biotinylated HK and any noncovalently associated, biotinylated proteins were detected on Western blots with streptavidin-HRP and ECL. Positions of prestained molecular mass markers are as follows: ovalbumin (55 kDa), BSA (86 kDa), and -galactosidase (120 kDa).

HK is associated with an endogenous 94-kDa MDCK cell protein at the cell surface of MDCK-HK cells. To determine whether HK was expressed alone at the cell surface in MDCK-HK cells, as was observed in other transfected cells (12, 14, 18, 23), the cell surface was biotinylated and cell lysis was effected under nondenaturing conditions. HK was subsequently immunoprecipitated with the anticytoplasmic domain MAb 2G11, and the immunoprecipitates were analyzed on Western blots by probing with streptavidin-HRP. As shown in Fig. 9B, a biotinylated endogenous 94-kDa MDCK cell protein coimmunoprecipitated with HK, suggesting that at least a fraction of HK at the cell surface is associated with another protein. This 94-kDa protein did not appear in immunoprecipitates from cells lysed under denaturing conditions (Fig. 9A), suggesting that the 94-kDa protein and HK are noncovalently associated.

Noncovalent association of HK with the 94-kDa protein occurs early in the biosynthetic pathway. Pulse labeling of transfected MDCK-HK cells with radiolabeled Cys and Met for 15 min and subsequent immunoprecipitation of pulse-labeled HK under nondenaturing conditions with MAb 2G11 resulted in the coimmunoprecipitation of two radiolabeled proteins: a 94-kDa protein and a protein migrating at ~50-55 kDa (Fig. 10A). The protein migrating at ~50-55 kDa is presumably a core-glycosylated, high-mannose ("immature") form of HK. This core-glycosylated form of HK has been previously identified in other heterologous expression systems (19, 23, 25, 26, 30), and the oligosaccharide chains are sensitive to hydrolysis by Endo H (Fig. 10B). Therefore, the association between these two proteins apparently occurs early in the secretory pathway.

View larger version (59K): [in this window] [in a new window]   Fig. 10.   Immunoprecipitation of metabolically radiolabeled HK. A: coimmunoprecipitation of a 94-kDa protein with pulse-labeled HK. Transfected MDCK-HK cells were metabolically radiolabeled with [35S]Cys and [35S]Met for 15 min and lysed under nondenaturing conditions. Lysates were subsequently immunoprecipitated with MAb 2G11. Samples were run on SDS-PAGE, and radiolabeled proteins were visualized by fluorography. B: metabolically pulse-labeled HK immunoprecipitated by MAb 2G11 is core glycosylated. An immunoprecipitate of HK was obtained as described for A and incubated without (lane 1) or with (lane 2) endoglycosidase H (Endo H). Oligosaccharides on pulse-labeled HK immunoprecipitated by MAb 2G11 are sensitive to hydrolysis by Endo H. After prolonged digestion with Endo H, although 94-kDa protein is still slightly visible, a noticeable loss of 94-kDa protein often occurred. C: association between HK and 94-kDa protein is noncovalent. Lanes 1 and 2, cells were lysed under denaturing conditions (SDS) and immunoprecipitated with MAb 2/2E6 (lane 1) or 2G11 (lane 2). From cells lysed under denaturing conditions, either MAb immunoprecipitates only HK. Lanes 3 and 4, sequential immunoprecipitation (seq) of pulse-labeled HK with MAbs 2/2E6 and 2G11. Cells were pulse labeled and lysed under nondenaturing conditions. Lysates were immunoprecipitated first with MAb 2/2E6 (lane 3) and then with MAb 2G11 (lane 4). Under nondenaturing conditions, MAb 2/2E6 immunoprecipitates only HK, whereas MAb 2G11 immunoprecipitates a complex of HK and a 94-kDa protein. Migration of molecular mass standards is indicated.

View larger version (59K): [in this window] [in a new window]   Fig. 11.   Pulse-chase analysis of metabolically radiolabeled HK in transfected MDCK-HK cells. MDCK-HK cells were pulse labeled and chased for 0-23 h. Cells were lysed under nondenaturing conditions, and HK was immunoprecipitated with MAb 2G11. Immunoprecipitates were separated on SDS-PAGE and visualized by fluorography. Positions of molecular mass markers are indicated.

The 94-kDa protein associated with HK is endogenous NaK. The identity of the 94-kDa protein associated with HK at the cell surface was revealed when the coimmunoprecipitating proteins from MDCK-HK cells were probed with anti-NaK antibodies on Western blots. NaK clearly coimmunoprecipitates with HK under nondenaturing conditions (Fig. 12A). The coimmunoprecipitation does not occur from cell lysates harvested under denaturing conditions, confirming that the association of HK with NaK is noncovalent. Moreover, when the surface-biotinylated complex of HK and 94-kDa protein is probed on Western blots, the surface-biotinylated 94-kDa protein clearly reacts with the anti-NaK antibodies (Fig. 12B). Thus, in MDCK-HK cells, HK can apparently assemble with endogenous NaK, and a fraction of this complex can be targeted to the apical plasma membrane.

View larger version (66K): [in this window] [in a new window]   Fig. 12.   Endogenous -subunit of Na-K-ATPase (NaK) coimmunoprecipitates with HK under nondenaturing conditions from MDCK-HK cells (A and B) and transfected LLC-PK1 cells (C and D). A: cells were lysed under nondenaturing (lane 1) or denaturing conditions (lane 2), and HK was immunoprecipitated with MAb 2G11. Immunoprecipitates were run with separation by SDS-PAGE, transferred to nitrocellulose, and blotted with an MAb against NaK. B: 94-kDa surface-biotinylated protein associated with HK reacts with polyclonal anti-NaK antibodies. After cell surface biotinylation, cell lysates were prepared under nondenaturing conditions. HK was immunoprecipitated with MAb 2G11. Biotinylated HK and any noncovalently associated, biotinylated proteins were detected on Western blots with streptavidin-HRP (lane 1) or polyclonal anti-NaK antibodies (lane 2) and ECL. Sample in lane 1 is same sample in Fig. 10B and is included for comparison with sample in lane 2, which was produced in same experiment and run on same gel and Western blot. C: coimmunoprecipitation of HK and a 94-kDa protein from metabolically radiolabeled, transfected LLC-PK1 cells. Cells were metabolically radiolabeled, and HK was immunoprecipitated by MAb 2G11. D: NaK is detected in Western blots of immunoprecipitates of HK. Transfected MDCK (lane 1), untransfected MDCK (lane 2), or transfected LLC-PK1 (lane 3) cells were lysed under nondenaturing conditions. Lysate was immunoprecipitated with MAb 2G11 and probed on Western blots with a monoclonal anti-NaK antibody. Positions of prestained molecular mass markers are shown.

HK assembles with endogenous NaK in transfected LLC-PK₁ cells. With the observation that HK assembles with endogenous NaK in MDCK-HK cells, we sought to reevaluate the assembly competence of HK expressed in LLC-PK₁ cells. From LLC-PK₁ cells, metabolically radiolabeled HK could be coimmunoprecipitated with a 94-kDa protein with MAb 2G11 (Fig. 12C). In addition, immunoprecipitation of HK with MAb 2G11 from cells lysed under nondenaturing conditions clearly results in the coimmunoprecipitation of endogenous NaK, as shown in the Western blot in Fig. 12D. These results suggest that HK may be generally competent to assemble with NaK in renal epithelial cells.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Two anti-HK MAbs have been further characterized. The epitope to MAb 2/2E6 appears to include the amino acids S₂₂₆LHY₂₂₉ in the extracellular domain of HK. In addition to evidence presented here, we have observed that MAb 2/2E6 reacts with all known HK containing the SLHY motif, whereas there is no recognition of those -subunits, dog (SLRY) or chicken (NLHY), where substitutions have occurred (8). The SLHY epitope appears to become masked when assembled with HK or NaK. The masking of this epitope in assembled, but not unassembled, H-K-ATPase suggests that SLHY may be part of a site for - interaction. Given that this site appears to be masked in heterodimers of HK-HK as well as NaK-HK hybrid heterodimers, this epitope may be a region that is involved in - association in all heterodimeric P-type ATPases. Alternatively, it is possible that the loss of recognition by MAb 2/2E6 may result from the "burying" of its epitope as a consequence of the intrinsic folding of the extracellular domain of HK, rather than assembly with an -subunit. We do not favor this interpretation, since HK, presumably alone at the cell surface of LLC-PK₁ cells, can bind MAb 2/2E6 without requiring denaturation or reduction of disulfide bonds. In these cells, HK should have satisfied all the quality control mechanisms, including proper folding, as a prerequisite to its exit from the ER for its appearance at the cell surface in the apparent absence of an -subunit.

Additional evidence suggests that MAb 2/2E6 binds to an epitope that is formed at an interface between - and -subunits. Two amino acids in the 2/2E6-binding epitope, L₂₂₇ and Y₂₂₉, are identical in all three isoforms of the -subunit of the Na-K-ATPase (1, 2, and 3) as well as HK, and the adjacent region Y₂₂₉Y(or F)P₂₃₁YYG is conserved throughout all known -subunits (8, 31, 39). The Y₂₂₉Y(or F)P₂₃₁YYG sequence is absolutely conserved in all -subunits of heterodimeric P-type ATPases characterized to date and overlaps with the SLHY motif in HK. The importance of this region in - assembly/association is further underscored by the work of Geering et al. (17) showing that the ability of the -subunit to assemble with the -subunit is virtually abolished by mutating the proline residue of YY(or F)PYYG. Thus these two overlapping motifs may be one part of HK that is responsible for the association of HK with NaK.

The second anti-HK MAb, MAb 2G11, was previously shown to bind to the cytoplasmic domain of HK (7). In this study we further defined the epitope to reside near the eight most NH₂-terminal amino acids. This region must be at or near an important site for functional activity, and possibly a cytoplasmic K⁺ binding site, inasmuch as association with MAb 2G11 inhibits H-K-ATPase activity and alters the K_A for activation of p-nitrophenyl phosphatase activity by K⁺ (7). It is also noteworthy that all mammalian species where HK is recognized by MAb 2G11 (mouse, rat, human, rabbit, and dog) have the same eight NH₂-terminal amino acids, whereas those -subunits with different NH₂-terminal sequences, including chicken HK and all isoforms of the -subunit of the Na-K-ATPase, are not recognized by MAb 2G11 (8).

We also showed here the utility of MAb 2G11 for immunoprecipitating assembled as well as unassembled HK. We used MAb 2G11 in conjunction with MAb 2/2E6 to evaluate the assembly of HK in transfected MDCK-HK and LLC-PK₁ cells. The simplest interpretation of our results is that there are apparently two cellular pools of HK. One pool is unassembled HK. In LLC-PK₁ cells, unassembled HK appears to transit to the plasma membrane; in MDCK-HK cells, unassembled HK largely appears to remain in the ER. The other pool of HK assembles with endogenous NaK in MDCK and LLC-PK₁ cells. Although the assembly of NaK-HK has been convincingly demonstrated in expression systems such as Xenopus oocytes (23, 25) and yeast (14, 15, 43), our results represent the first demonstration of assembly of HK with NaK in epithelial cells, a significant result for several reasons. First, published studies on the trafficking of transfected HK in LLC-PK₁ and MDCK cells used the assembly-sensitive MAb 2/2E6 (18). Moreover, several studies have shown that HK appears to transit to the plasma membrane in the absence of an -subunit in other cell types (12, 18, 19, 38). All these studies used MAb 2/2E6; thus analysis of these trafficking data would be restricted to that of unassembled HK. For MDCK-HK cells, we showed that a significant fraction of HK appears to be assembled with NaK at steady state; thus a significant population of HK was likely overlooked with respect to the analysis of trafficking of HK in these earlier studies. Because HK also assembles with NaK in LLC-PK₁ cells, the conclusions regarding HK trafficking in transfected epithelial cell lines may suffer from the same limitations (18, 38). In both previously published studies, it was stated that HK assembly with NaK was not detected. These differences in results regarding HK assembly are not likely the result of clonal variations, since the LLC-PK₁ cell line transfected with HK is common to both studies. The differences, however, are likely due to the different MAbs used by each investigator. In the light of the present data, it will be important to reevaluate the trafficking of HK with respect to assembled and unassembled HK in these epithelial cell lines. In addition, our preliminary unpublished data suggest that the trafficking of HK in MDCK-HK cells is more complex than formerly believed and may be regulated not only by assembly of HK, but also by time of culture or after induction of polarity (unpublished observations). Similarly, the targeting of endogenous Na-K-ATPase in MDCK cells as a function of the establishment of polarity is well documented (13, 32).

Second, these two transfected epithelial cell lines appear to display different phenotypes relative to the relationship between assembly of HK and its appearance at the cell surface. In LLC-PK₁ cells, assembly of HK with an -subunit does not appear to be a prerequisite for its exit from the ER and appearance at the cell surface. The mechanism by which HK appears to be able to reach the plasma membrane alone remains to be determined; alternatively, it remains to be determined whether HK in LLC-PK₁ cells assembles with another protein that is not detectable by the assays used in this study and whether the 2/2E6 epitope remains exposed in such a protein complex. Assembly with such a surrogate may then facilitate exit of this complex from the ER for delivery to the plasma membrane. On the other hand, MDCK-HK cells characterized here appear to function in a manner that is more consistent with the present views of quality control mechanisms for membrane protein assembly in the ER. That is, HK does not appear to exit the ER without assembly with an -subunit. In this case, it is the -subunit of the endogenous Na-K-ATPase. Conversely, numerous studies have shown that the exit of nascent -subunits from the ER is dependent on the presence of a -subunit (for review see Refs. 16 and 33).

Third, isoforms of the H-K-ATPase and Na-K-ATPase are clearly coexpressed in cells other than the gastric oxyntic cell, such as colonic (4, 11, 39) and renal medullary cells (27). The data presented here suggest that the potential for the assembly of hybrid heterodimers in epithelial cells is significant. Thus, in some instances, cells must possess regulatory mechanisms to prevent hybrid heterodimer assembly, as in the gastric oxyntic cell. One possible mechanism is to regulate subunit expression, hence assembly, at the level of mRNA translation of ATPase subunits, as shown recently in MDCK cells (20). Alternatively, there may be other intrinsic posttranslational mechanisms in place to prevent such hybrid assem