| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 279, Issue 35, 36339-36348, August 27, 2004
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
From the Department of Biochemistry and Molecular Genetics, University of Illinois at Chicago, Illinois 60607-7170
Received for publication, March 11, 2004 , and in revised form, June 4, 2004.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
heterodimer in cell membranes is the subject of controversy. We have utilized the baculovirus-infected insect cell system to express Na,K-ATPase with -subunits bearing either His6 or FLAG epitopes at the carboxyl terminus. Each of these constructs produced functional Na,K-ATPase heterodimers that were delivered to the plasma membrane (PM). Cells were simultaneously co-infected with viruses encoding -His/ and -FLAG/ Na,K-ATPases. Co-immunoprecipitation of the His-tagged -subunit in the endoplasmic reticulum (ER) and PM fractions of co-infected cells by the anti-FLAG antibody demonstrates that protein-protein associations exist between these heterodimers. This suggests the Na,K-ATPase is present in cell membranes in an oligomeric state of at least ()2 composition. Deletion of 256 amino acid residues from the central cytoplasmic loop of the -subunit results in the deletion -4,5-loop-less (-4,5LL), which associates with but is confined to the ER. Co-immunoprecipitation demonstrates that when this inactive -4,5LL/ heterodimer is co-expressed with wild-type , oligomers of wild-type and -4,5LL/ form in the ER, but the -4,5LL mutant remains retained in the ER, and the wild-type protein is still delivered to the PM. We conclude that the Na,K-ATPase is present as oligomers of the monomeric heterodimer in native cell membranes. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
-subunit in addition to the catalytic -subunit. The -subunit is composed of about 1000 amino acids with 10 transmembrane segments and encodes all of the essential domains for ATP hydrolysis and ion transport. The -subunit is a glycoprotein of around 300 amino acids with a single transmembrane segment and an intracellular amino terminus. Recent data from a number of systems has established that the -subunit probably plays an influential role in the catalytic activity of the system (2, 3) in addition to its essential role in the delivery of the heterodimers to the PM1 from the site of synthesis in the ER (4–6). The subunit assembly takes place in the ER, and assembled heterodimers are delivered to the PM (for review, see Ref. 6).
Although the heterodimeric state of the Na,K-ATPase is well established, the organization and possible higher order oligomerization of heterodimers in native membranes remains a controversial subject (1, 7, 8). It has been claimed that the Na,K-ATPase heterodimer in native membranes exists as an oligomer of multiple heterodimer units (7, 9, 10). The existence of these oligomers has been postulated to account for kinetic observations that are difficult to explain with a simple monomeric unit (11–13). Evidence for the existence of such multimers or for associations among monomers has derived from cross-linking experiments (10, 14–16), single particle images (17), thermal inactivation studies (18), and association among different isoforms (9, 19, 20). Other recently published work demonstrates that the major cytoplasmic loop between the M4 and M5 membrane segments associates in an ATP-dependent way when expressed in isolation (21). Conversely, it has been claimed that the monomeric solubilized enzyme is capable of normal enzymatic activity (22). Furthermore, cross-linking between -subunits was not observed in cell membranes containing a low density of Na,K-ATPase molecules (23) or after detergent solubilization (24), leading to the suggestion that higher order associations may only be a result of unusually high Na,K-ATPase densities in some enriched preparations.
In the present work we have utilized the expression of recombinant Na,K-ATPase molecules in baculovirus-infected insect cells, where viruses contain the cDNA that encodes for -subunits bearing either His6 or FLAG epitopes. Expressed in this system, the level of Na,K-ATPase has been estimated to reach up to 1–2% of the total membrane protein (5). We show that each of these epitope-tagged -subunits expressed with wild-type -subunits results in the synthesis and delivery of functional Na,K-ATPase heterodimers to the PM of the insect cells. We show that co-infection of cells with both viruses and subsequent immunoprecipitation of either ER or PM fractions from infected insect cells by an anti-FLAG antibody results in co-immunoprecipitation of the -subunit bearing the His6 epitope. We also show that co-expression of wild-type Na,K-ATPase heterodimers with non-functional heterodimers containing a deleted -subunit, lacking most of the nucleotide binding and phosphorylation domains, produces oligomeric molecules. These data provide strong evidence for association of native heterodimers in cell membranes.
|
EXPERIMENTAL PROCEDURES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
1 cDNA as a template. The His6 (HHHHHH) or FLAG (MDYKDDDD) epitope tag was inserted on the carboxyl terminus of the 1-subunit to create -His and -FLAG, respectively. The -4,5-loop-less (-4,5LL) mutant is a deletion (
Asp443-Val707) of the sheep 1 M4M5 loop. As a result of mutagenesis, a Glu residue was inserted at the junction of the amino- and carboxyl-terminal regions such that the M4M5 loop in the -4,5LL sequence is M4-G442ET708-M5.
Mutated 1 cDNA was sub-cloned in the pOCUS-2 vector (Novagen) and subsequently ligated into the pFBD (Invitrogen) vector in the multicloning site 2. The wild-type sheep 1-subunit cDNA was placed at multicloning site 1 in the same pFBD. The Bac to Bac system (Invitrogen) was used create recombinant baculovirus particles, each of which contain cDNA for - and -subunits under separate promoters (5). Mutations in the 1-subunits were confirmed by sequencing of DNA isolated from recombinant baculovirus using Easy-DNA kit (Invitrogen). Protein expression of the -4,5LL mutant was also confirmed by its mobility in SDS-PAGE, visualized via Western blot analysis using the anti-KETYY antibody.
Protein Expression and Isolation—Log-phase high viability (
98% viability, determined by trypan blue exclusion) High Five cells were infected with recombinant baculovirus in the presence of 1% ethanol (v/v). The progress of each infection was monitored by recording the cell density and viability. Infections were harvested by centrifugation (500 x g, 10 min) between 3 and 5 days post-viral infection, when cells retained 60–75% viability. The harvested High Five cells were homogenized, whole cells were removed, and lysate was either ultracentrifuged to collect total membranes or separated on a five-step sucrose gradient into fractions enriched in ER, Golgi apparatus (G), or PM as previously described (5). Membrane pellets were resuspended in HB (250 mM sucrose, 2 mM EDTA, 10 mM Tris, pH 7.4) and stored on ice or at –20 °C for long term storage. Protein concentration was determined by the Lowry assay using bovine serum albumin as a standard (25).
ATPase Activity—ATPase activity assay of isolated membranes was performed in triplicate for 30 min at 37 °C with 16 µg of membrane protein as previously described (25). Na,K-ATPase activity was determined as the difference in the amount of phosphate liberated in the presence and absence of 0.3 mM ouabain, expressed as µmol of Pi/mg of protein/h.
Western Blot Analysis—Equal protein concentrations of ER-, G and PM-enriched fractions were subjected to SDS-PAGE in 7.5 or 10% acrylamide/bisacrylamide gels in the presence of 0.75% -mercaptoethanol, and transferred to nitrocellulose membranes. Western blots were blocked and probed with anti-1, anti-1, anti-FLAG, anti-KETTY, anti-His6-HRP, or anti-hCTR1 antibody as previously described (4). Goat anti-mouse HRP (Pierce #31430) or goat anti-rabbit HRP (Pierce #31460) secondary antibodies were used when appropriate. Chemiluminescent reagents (Pierce) were used for signal detection.
Antibodies—The anti-KETYY antibody, which was raised in rabbit against the sheep 1 carboxyl-terminal amino acid sequence 1111KETYY1116 (a gift from Dr. Jack Kyte, University of California, San Diego, CA), was used for both Western blot and co-immunoprecipitation. Likewise, the anti-1 antibody (Affinity Bioreagents #MA3–929), which recognizes an epitope in the region of amino acid residues 496–506 in the M4M5 loop of the 1-subunit was used for both Western blot and co-immunoprecipitation experiments. The anti-1 antibody (Affinity Bioreagents #MA3–930) was used to detect the 1-subunit on Western blots. The anti-FLAG antibody (Sigma #F7425) was used for both Western and co-immunoprecipitation. The anti-His6 antibody (Affinity Bioreagents #PA1–983) was used for co-immunoprecipitation and anti-His6-HRP conjugate antibody (Qiagen #34460) was used for Western blot detection of the His6 tag. A rabbit polyclonal antibody raised against an epitope in the carboxyl-terminal region of hCTR1 was used for Western blot detection of hCTR1.
Protein Solubilization—Membrane preparations (500–1000 µg) were pelleted in a TLA 45 rotor at 45,000 rpm for 30 min. Pellets were resuspended in 200 µl of immunoprecipitation buffer (IPB) (150 mM NaCl, 10 mM KCl, 2.5 mM MgCl2, 25 mM Hepes, pH 7.4, 1 mg/ml leupeptin, 2 mg/ml antipain, 1 mg/ml pepstatin, 100 mg/ml L-1-tosylamido-2-phenylethyl chloromethyl ketone, and 100 mg/ml phenylmethylsulfonyl fluoride) supplemented with detergent. For subsequent immunoprecipitation of carboxyl-terminal-tagged -subunits, 2% n-dodecyl--D-maltoside (DDM) was used as the solubilizing detergent, as previously described (4). Solubilization of the -4,5LL construct required 1% Anagrade Fos-Choline-12 (Anatrace #F308). Initial suspensions were passed through a 28-gauge x 0.5-inch needle until homogenous and incubated at room temperature (
25 °C) for 30 min. Suspensions were centrifuged at 14,000 x g for 5 min to remove insoluble material. In the absence of detergent, no - or -subunit was detected by Western blot in the supernatant after this centrifugation step. In the presence of detergent, a significant portion of each subunit was detectable in the supernatant. After solubilization, the supernatant was removed, and an aliquot (1 to 5%) used for Western blot analysis (S sample). The remaining soluble material (supernatant) was used for immunoprecipitation. The insoluble pellet was resuspended in 200 µlof IPB using a 28-gauge x 0.5-inch needle until homogenized, and an equal volume (1–5%) was subjected to Western blot analysis (IS sample).
Co-immunoprecipitations—Supernatant from solubilization (200 µl: see above) was transferred to a fresh tube containing IPB and antibody (anti-FLAG, anti-His6, anti-1, or anti-KETYY) for a total volume of 1 ml. Two sets of controls were done in parallel. Only soluble protein (no antibody) was added to one control, and only antibody (no protein) was included in the other control. The samples were incubated at 4 °C with end-over-end rotation overnight. Protein G-Sepharose beads (50–100 µl at 1:1 IPB) were added to all samples and incubated at 4 °C with rotation for 4 h. Beads were pelleted at 500 x g for 5 min, and supernatant was removed by aspiration. The beads were washed 3–5 times for 5 min with rotation at 4 °C in 1 ml of IPB followed by 500 x g for 5 min and the aspiration of supernatant. Precipitated protein was removed from the beads by incubation for 30 min to overnight at room temperature with 30 µlof4x SDS sample buffer (100 mM Tris-Cl, pH 6.8, 10% glycerol, 8% SDS, 0.75% -mercaptoethanol, 0.0001% bromphenol blue). Beads were removed by centrifugation at 21,000 x g for 5 min, then the supernatant was resolved on a 7.5% SDS-polyacrylamide gel, and a Western blot analysis was performed.
Trypsin Digestion—Membrane preparations in HB (see above) were supplemented with 20 mM NaCl, 20 mM KCl, 2 mM ATP-Tris or no ligand. Volumes were adjusted such that the final protein concentration was 4 mg/ml after the addition of trypsin at 1 mg/ml. Digestion was stopped by the addition of denaturing sample buffer after 2 h at room temperature, and proteolyzed fragments were analyzed by Western blot analysis.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
1-subunit and sheep 1-subunit bearing either the His6 or FLAG epitope tag at the carboxyl terminus (-His/ or -FLAG/). Cells were harvested and Dounce-homogenized, and ER-, G-, and PM-enriched fractions were separated on a five-step sucrose gradient, as previously described (5) Western blot analysis of the membrane fractions by the anti-1 antibody confirms that, like the wild-type -subunit, both tagged -subunit constructs are delivered to the PM (Fig. 1A, top panel). Only the appropriately tagged constructs are detectable with the anti-FLAG and anti-His6-HRP antibodies, and there is no cross-reactivity (Fig. 1A, bottom panels). Fig. 1A also demonstrates the expression and trafficking of the -subunit. Previous work in the baculovirus expression system has established that delivery of the -subunit to the PM is dependent upon the heterodimeric assembly with the -subunit (4, 5). The appearance of the -subunit in the PM fraction (Fig. 1A) demonstrates that heterodimers containing these tagged -subunits are sorted like wild-type protein and delivered to the PM of the cells. Fig. 1B shows the ouabain-sensitive ATPase activities in each of the membrane fractions of the tagged -subunit construct expressions. These ATPase activities are comparable to the ATPase activities of wild-type protein with equivalent protein expression levels (data not shown). Likewise, the distribution of the Na,K-ATPase activity between the membrane fractions is the same as we have previously reported with the wild-type protein (4), in which the PM fraction contains the greater proportion of Na,K-ATPase activity. The data presented in Fig. 1 establish that attachment of the epitopes to the carboxyl terminus of the -subunit has no detrimental effect on the normal heterodimer synthesis, assembly, delivery to the PM, or activity of the Na,K-ATPase.
|
-FLAG/ and -His/ baculoviruses. Each virus was at a concentration 1 order of magnitude higher than the cell density of the suspension. This level of each virus is sufficient to independently infect the entire population of cells. Co-infection for the purpose of simultaneously expressing Na,K-ATPase subunits in insect cells has been utilized previously (26). The top two panels of Fig. 2A show Western blot reactivity by both the anti-His6-HRP and anti-FLAG antibodies, confirming that both tagged -subunits are present in the indicated membrane fractions. Likewise, the -subunit is present in all fractions (Fig. 2B, bottom panel). Co-infection with the -FLAG/ and -His/ baculoviruses does not alter the level or cellular distribution of Na,K-ATPase activity (compare Fig. 1B and Fig. 2B).
|
-subunit in 2% DDM before co-immunoprecipitation. Note that a significant portion of the -subunit is present in the supernatant. In the absence of detergent, the -subunit is not observed in the supernatant fraction, indicating that the presence of detergent is necessary to separate the -subunit from the membrane pellet (i.e. solubilized the -subunit) under these conditions (data not shown). The lane in Fig. 2C marked S represents 5% of the solubilized fraction subsequently used as protein input for co-immunoprecipitation. Since the -subunit expresses at a higher level than the -subunit, a significant portion of the -subunits are not assembled in heterodimers with -subunits (4, 5). Therefore, only a portion of this solubilized -subunit is available for co-immunoprecipitation by antibodies against the -subunit. Fig. 2D shows co-immunoprecipitation of the wild-type -subunit by either the anti-His6 (fifth lane) or the anti-FLAG antibodies (fourth lane). It should be noted that in control immunoprecipitations lacking either precipitating antibody (first lane) or protein input (second and third lanes) we were unable to reproduce this immunoprecipitation, demonstrating the specificity of the co-immunoprecipitation. These results confirm that each of the tagged -subunits assemble with the -subunit as well as demonstrating that the anti-FLAG and anti-His6 antibodies are capable of immunoprecipitating the FLAG and His6-tagged -subunits, respectively. Fig. 2D shows that the -subunits which are co-immunoprecipitated the most efficiently with the -subunit tend to be on the higher side of the molecular weight range for the -subunit, indicating higher glycosylation and a more mature -subunit.
Since the immunoprecipitation efficiency of the anti-FLAG antibody appeared slightly higher than that of the anti-His6 antibody (compare Fig. 2D, fourth and fifth lanes), the ER and PM fractions of co-infected cells were subjected to immunoprecipitation with anti-FLAG antibody. The immunoprecipitates were then probed with the anti-His6-HRP antibody to investigate the presence of oligomers containing both of the epitope tagged -subunits. Before immunoprecipitation, membranes were solubilized in 2% DDM. Fig. 3A shows the Western blot analysis detecting the His6-tagged -subunit from the ER and PM fractions before solubilization and in the soluble and insoluble fractions. Interestingly, Fig. 3A reveals that more of the Na,K-ATPase -His6 subunit is solubilized from the PM fraction than from the ER fraction under these solubilization conditions. The soluble ER and PM fractions were used as input for the co-immunoprecipitation experiment in Fig. 3B. The anti-FLAG antibody was able to precipitate the -His6 subunit from ER and PM membrane fractions, as indicated by Western blot detection of the immunoprecipitate with the anti-His6-HRP antibody (Fig. 3B, fourth and fifth lanes). Control precipitations lacking the anti-FLAG antibody confirm that the positive coimmunoprecipitation result is not due to nonspecific binding to beads (Fig. 3B, first and second lanes). Because there is no cross-reactivity between the epitope antibodies (Fig. 1A), this result indicates the -FLAG subunits must be in stable association with -His subunits in the DDM-solubilized ER and PM fractions.
|
-tagged Na,K-ATPase from co-localization in membrane patches. If DDM solubilization does not separate the heterodimers from the membrane, the differently tagged subunits might be expected to co-immunoprecipitate whether or not protein-protein associations occurred due to their co-localization in membrane patches. To test the ability of DDM-solubilized membrane proteins to sufficiently disrupt non-associated membrane proteins, we performed an additional control experiment. Baculovirus-mediated co-expressions in High Five cells were carried out with the tagged -subunits and the unrelated membrane protein, hCTR1. hCTR1 is the human copper transporter that has been functionally expressed recently in insect cells using the baculovirus system (27). The hCTR1 protein does not specifically associate with the Na,K-ATPase, but it is delivered to the PM, like the Na,K-ATPase. Fig. 3C demonstrates the cellular localization by Western blot analysis of fractionated membranes co-expressing hCTR1 and either -His/ or -FLAG/. Like wild-type Na,K-ATPase, hCTR1 is delivered to the PM. The complex pattern of bands seen with hCTR1 is due to glycosylation and the presence of some degradation products (18 kDa and below) of this protein (27). Total membrane preparations of cells co-expressing hCTR1 with -His/ or -FLAG/ were subjected to the same 2% DDM solubilization protocol as used in Fig. 3, A and B. The first lane of each blot in Fig. 3D shows 2% of the soluble protein input for immunoprecipitation experiments. Due to preferential solubilization of higher glycosylated hCTR1, slightly altered gel conditions, and reduced levels of secondary antibody for Western blot detection, the hCTR1 signal in Fig. 3D appears tighter and of only the higher molecular species observed in Fig. 3C. Immunoprecipitation experiments in Fig. 3D (second lane of each blot) shows that neither the anti-FLAG nor the anti-His6 antibodies co-immunoprecipitated hCTR1 with the tagged Na,K-ATPase. Note that there is a protein G band present in all immunoprecipitation and control lanes that runs at a higher molecular mass (about 31 kDa) than hCTR1 (about 26 kDa). The control experiment presented in Fig. 3D demonstrates that solubilization in 2% DDM is sufficient to separate non-associated but co-localized membrane proteins. In summary, the data presented in Fig. 3 demonstrate the oligomeric association of the Na,K-ATPase heterodimers after solubilization from membranes and the lack of association of the subunit with hCTR1.
Recently published work has shown that the large cytoplasmic loop between the M4 and M5 transmembrane segments, when expressed in isolation, form oligomers in an ATP-dependent fashion (21). This loop contains the phosphorylation and nucleotide binding domains (28) for which significant structural information is provided by crystal structures of sarco(endo)plasmic reticulum calcium ATPase and EM structures of the Na,K-ATPase (29–31). Furthermore, cross-linking experiments (16) and chimera studies (9, 20) using the intact protein suggest that regions of association between oligomers occur in the M4M5 loop. We investigated the potential oligomerization of the Na,K-ATPase wild-type heterodimer with a mutant of the -subunit lacking a significant portion of the M4M5 loop. This anomalous Na,K-ATPase, termed the -4,5LL (for M4M5-loop-less) is shown in Fig. 4. The M4M5 loop, which is normally composed of the stretch of amino acid residues from about 354 to 774, lacks amino acid residues 443–708 in the -4,5LL mutant. This eliminates about 65% of the loop, with regions missing from both the phosphorylation (P) and nucleotide binding (N) domains (see Fig. 4). However, the -4,5LL retains the essential site of enzymatic phosphorylation, Asp-369, and the ATP binding motif GCGXNDXP.
|
-4,5LL/ baculovirus encoded expression of the -4,5LL mutant in High Five cells. For comparison, a wild-type expression is shown in adjacent lanes. Western blot detection with the KETYY antibody, which recognizes the carboxyl terminus of the -subunit, shows that the -4,5LL mutant migrates at the predicted mass of about 73 kDa (top panel of Fig. 5A). Due to the elimination of the epitope within the M4M5 loop, the -4,5LL mutant is not detected by the 1 antibody (Fig. 5A, bottom panel). Fig. 5A also shows that although the -subunit expressed from the -4,5LL/ virus is targeted to the PM (Fig. 5A, middle panel), the -4,5LL mutant is predominantly retained in the ER, unlike the wild-type heterodimers. ER retention of the -subunit can occur for a number of reasons including the loss of a PM targeting signal, recognition of misfolded protein by the ER quality control machinery, or failure to associate with the -subunit. It is interesting that we do not observe extensive degradation of the -4,5LL mutation, unlike some other -subunit mutations we have characterized previously (32). Although it would be undetectable by these methods if the KETYY moiety were removed first, the absence of extensive degradation suggests that the -4,5LL is not greatly denatured and is integrated appropriately into the membrane.
|
-4,5LL mutant might be retained in the ER due to the presentation of misfolded protein, we looked at ligand-dependent digestion of the -subunit. It has been well established that the Na,K-ATPase assumes different conformational states when bound to different ligands. Therefore, we tested the trypsin sensitivity of the wild-type and the -4,5LL mutant in the presence of Na, K, ATP, or no ligand. Mild trypsin treatment reveals increased trypsin sensitivity of the -4,5LL (Fig. 5B). The carboxyl-terminal -4,5LL tryptic product migrating at 32 kDa is predicted to result from a cleavage within the M4M5 loop. Fig. 5B shows that the presence of ATP provides protection against tryptic cleavage at this site. Although the -4,5LL deletion fails to produce ouabain-sensitive Na,K-ATPase activity (data not shown), Fig. 5B demonstrates that the -4,5LL mutant achieves a reasonably folded state, although it is retained in the ER. Furthermore, the -4,5LL does not show increased tryptic sensitivity at further carboxyl-terminal sites, suggesting that the carboxyl-terminal of the -4,5LL mutant has a fairly normal fold.
Co-infection with the wild-type and the -4,5LL/ baculoviruses does not alter the subcellular localization of the wild-type or mutant -subunits, as demonstrated by Western blot analysis of fractionated membranes in Fig. 6A. Likewise, the presence of the -4,5,LL subunit did not greatly inhibit or modify the distribution of the Na,K-ATPase activity of the wild-type protein (Fig. 6B). We have shown previously that -subunits not associated with the -subunit are retained in the ER (5). Therefore, if the -45LL mutant were not able to assemble with the -subunit, the ER retention of the -45LL would be explained. Our initial attempts at immunoprecipitation of the -4,5LL subunit revealed that, unlike the wild-type protein, this mutant is not soluble in 2% DDM. Therefore, we tried solubilizing the co-expressed wild-type and -4,5LL mutant in a number of detergents that have been previously used for the solubilization of the Na,K-ATPase and/or other membrane proteins. As Fig. 6C shows that at least 50% of the wild-type -subunit was solubilized by all of the detergents tested. From the six detergents shown in Fig. 6C, only SDS was able to solubilize the -4,5LL subunit. Unfortunately, the presence of SDS during immunoprecipitation caused nonspecific binding of the Na,K-ATPase to the protein G-Sepharose beads in the absence of antibody. Therefore, we turned to less widely used detergents. We found that the -4,5LL subunit could be solubilized using 1% Fos-choline (see Fig. 8A).
|
|
-subunit co-immunoprecipitation experiments after solubilization in 1% Fos-choline. Fig. 7A demonstrates the relative amount of the -subunit solubilized from ER membranes containing either wild-type - and -subunits or the -4,5LL mutant and wild-type -subunits. Only the ER fractions were used because the -4,5LL is retained in the ER (Fig. 6A). The soluble protein fractions shown in Fig. 7A were used as input for immunoprecipitation by the anti-1(top panel, Fig. 7B) or anti-KETYY (bottom panel, Fig. 7B) antibodies, followed by Western blot detection for the -subunit. The anti-1 antibody was able to co-immunoprecipitate the -subunit from wild-type (Fig. 7B, top panel, fifth lane)-containing membranes but not from the -4,5LL/ expression (Fig. 7B, top panel, fourth lane). This was expected because the -4,5LL does not contain the epitope recognized by the anti-1 antibody. Note that antibody bands are also detected on the Western blot due to reactivity of the anti-mouse-HRP secondary antibody with the anti-1 antibody (Fig. 7B, top panel, first, fourth, and fifth lanes). Negative controls lacking antibody or protein input confirm that the association is specific (Fig. 7B, top panel, second and third lanes). The lower panel of Fig. 7B shows that the anti-KETYY antibody, which recognizes the carboxyl terminus of the -subunit, was able to co-immunoprecipitate the -subunit from both expressions (Fig. 7B, bottom panel, fourth and fifth lanes). Again, negative control lacking antibody or protein input produced no signal (Fig. 7B, bottom panel, first through third lanes). These results demonstrate that although the -4,5LL lacks a major part of the cytoplasmic loop, it is still able to associate with the -subunit. The failure of the -4,5LL subunit to exit the ER (Fig. 5A) is not then due to a lack of association with the -subunit. Heterodimerization with the -subunit further suggests the correct membrane insertion of the -4,5LL mutant because the region of the -subunit that has been implicated in assembly with the -subunit (the M7M8 loop) is carboxyl-terminal to the deleted M4M5 loop.
|
-4,5LL/ protein (Fig. 6A). Fig. 8A shows solubilization of two different ER preparations in 1% Fos-Choline. Whereas nearly all of the wild-type -subunit was present in the soluble fraction, less than half of the -4,5LL mutant subunits were solubilized. Each soluble lane in Fig. 8A shows 5% of the protein input, which was subject to immunoprecipitation. Western blot detection by the anti-KETYY antibody after precipitation with the anti-1 antibody detects both the wild-type -subunit and the -4,5LL mutant in the precipitated material (Fig. 8B, fourth and fifth lanes), indicating that the wild-type -subunit and -4,5LL mutant are associated in the Fos-Choline-solubilized preparation. Control experiments utilizing co-infected hCTR to test for protein solubilization were not performed because we found hCTR solubilized in Fos-Choline nonspecifically associates with protein G-Sepharose beads. However, control precipitations of the wild-type and -4,5LL mutant lacking antibody or protein input confirm that the co-immunoprecipitation result in Fig. 8B, fourth and fifth lane are not due to nonspecific binding to beads or antibody cross-reactivity (Fig. 8B, first through third lanes). |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
subunits, exists in a higher oligomeric state of at least two -subunits. We demonstrate the existence of stable protein-protein associations between heterodimers by co-immunoprecipitation with anti-FLAG antibody of -subunits that bear the His6 epitope along with -subunits that bear the FLAG epitope. Control experiments established that this result is not accounted for by cross-reactivity of the antibodies or due to opportunistic co-localization of the tagged -subunits within unsolubilized membrane patches. We also demonstrate that such epitope-tagged -subunits also associate with -subunits. Our findings thereby imply the existence of a higher oligomeric state of at least ()2 in native membranes. These oligomeric heterodimers are present in enriched fractions of the ER and PM of High Five cells, so it is likely that the oligomeric unit is formed after synthesis in the ER and then delivered (via the Golgi network) to the PM. Our results do not address the issue of whether or not such associations among heterodimers have significant modifying effects on the enzyme activity.
The oligomeric state of the Na,K-ATPase has been the subject of ongoing controversy within the field for more than 20 years. In 1983, Brotherus et al. (33) demonstrated by analytical ultracentrifugation that the C12E8-solubilized heterodimer was a single unit that could be reconstituted in phospholipids to regain Na,K-ATPase activity. The same group later showed the soluble heterodimer retained the ion occlusion properties of the membrane bound Na,K-ATPase (22). This work established that the heterodimer forms the membrane ion binding sites and suggested that oligomerization of multiple units is not necessary to transport ions. This early work, however, could not exclude the possibility that Na,K-ATPase may spontaneously form higher order associations in vivo or reform them in vitro. Subsequently, a significant body of evidence has accumulated which suggests that the Na,K-ATPase can form higher oligomers. Furthermore, this association of subunits has been claimed to influence Na,K-ATPase functioning (for review, see Ref. 7). It should be noted that subsequent work with C12E8-solubilized membranes has demonstrated the existence of ()2 and higher order oligomers in addition to heterodimers, the relative amounts of which were dependent on buffer ionic composition (34). More recently, work with a Na,K-ATPase preparation from salt-adapted duck nasal gland has supplied evidence for the monomeric heterodimer as the functional unit in the membrane. This preparation, which can be fully phosphorylated, contains primarily monomeric heterodimers (35, 36). However, in light of mounting evidence from many methods and systems demonstrating the oligomerization of the Na,K-ATPase, including our current work, it seems likely that in a native membrane environment, the Na,K-ATPase can and does form higher order oligomers.
Numerous other studies have been presented in the literature that support the oligomerization of the Na,K-ATPase. For more extensive reviews, see Refs. 7 and 8. Briefly, physical interactions between , , and have been implicated from cross-linking studies (14, 16, 24, 37–39), co-immunoprecipitations have demonstrated the interaction of different -subunit isoforms (19), thermal inactivation studies suggest contact (18), fluorescence resonance energy transfer provides distance measurements consistent with ()2 (40), and low angle laser light scattering photometry coupled with high performance gel chromatography (41) as well as single molecule detection (17) indicate the presence of , ()2, and higher order oligomers. Numerous studies in which half-site phosphorylation and half- or quarter-site ATP-analog binding stoichiometries have been interpreted in a model that proposes that the Na,K-ATPase co-exits simultaneously in the E1ATP, E2ATP, or EPATP forms (42). The existence of such oligomerization of the Na,K-ATPase in native membranes provides a mechanism that can explain the recent observation of dominant-negative -mutants in Drosophila (43).
In an earlier study using baculovirus-infected insect cells and immunoprecipitation using isoform-specific antibodies, Blanco et al. (19) provide evidence of stable associations of the -heterodimers in unfractionated cell membranes (19). They also showed that such oligomerization did not take place between -subunits from the Na,K-ATPase and the H,K-ATPase. The present work confirms and extends those findings using epitope tags on the wild-type 11 heterodimers and demonstrates that the oligomeric forms occur in membranes from both the ER and PM. It should be borne in mind that we estimate the heterologously expressed Na,K-ATPase represents only 1–2% of the total membrane protein in the insect cells (5), so it is unlikely that the observed associations are produced by over-enrichment or artificially induced close interactions of the heterodimeric units.
The observation of Na,K-ATPase oligomerization naturally leads to the question of how heterodimeric units interact to form higher order complexes. Work utilizing Na,K-ATPase and H,K-ATPase chimeras suggests that the cytoplasmic loop of the Na,K-ATPase is important for oligomerization (9). These data are in agreement with the observation that the bacterially expressed major cytoplasmic loop between membrane segments M4 and M5 in isolation can dimerize in an ATP-dependent way (21). However, if the M4M5 loop is exclusively responsible for oligomerization, it is surprising that, in the work presented here, the -4,5LL mutant, from which 65% of the M4M5 cytoplasmic loop had been removed, is able to associate with the wild-type protein. The formation of oligomers between the wild-type and the -4,5LL mutant demonstrates that the full M4-M5 loop is not necessary for oligomeric protein-protein interactions and suggests that an association interface may extend beyond the cytoplasmic loop, possibly including other regions of the -subunit or the -subunit. It is interesting that although stable associations are formed in the ER between wild-type and -4,5LL -subunits, this does not enable these (') assemblies to leave the ER in the same way as ()2.As an extension of our current findings, we are initiating studies on the effects of Na,K-ATPase ligands on the oligomerization.
Our findings that the -4,5LL mutant is retained in the ER raises some interesting questions. Because the deleted region of the -subunit in the -4,5LL mutant is on the cytoplasmic side of the membrane, retention by ER luminal quality control apparatus does not seem to be a likely explanation for the ER retention unless some segments of the -4,5LL mutant are incorrectly inserted in the membrane. However, our observations that 1) the -4,5LL is relatively stable, i.e. not readily degraded in the ER, and 2) the -subunit assembles with the -4,5LL mutant support the idea that the -4,5LL achieves a fairly normal membrane insertion. ER retention of the -4,5LL subunit even though it associates with the -subunit also supports our previous conclusion that assembly, although necessary, is not sufficient for exit from the ER (4). It should be noted that our present result implies that structural elements of the -subunit may also be important to accomplish ER exit.
We do not know yet whether the formation of oligomeric subunit association is modulated via direct interactions alone or if other components such as the cytoskeleton play a role. Based on primary structure, two putative ankyrin binding domains have been identified in the Na,K-ATPase -subunit and claim to be important for Na,K-ATPase trafficking (44, 45). These sites are located in the M2M3 and the M4M5 cytoplasmic loops. The putative ankyrin binding site in the M4M5 loop, which is deleted in the -4,5LL mutant, is predicted to be buried based on alignment of the Na,K-ATPase in the sarco-(endo)plasmic reticulum calcium ATPase crystal structure (46). Because the -4,5LL is able to associate with the wild-type Na,K-ATPase, oligomerization cannot occur only through cytoskeletal associations with the putative ankyrin binding site in the M4M5 loop. Studies are currently under way to investigate the role of the other putative ankyrin binding site in the M2M3 loop.
In summary, we have shown that the Na,K-ATPase heterodimers, when heterologously expressed in insect cells, form stable intermolecular associations leading to at least bis-heterodimers in both the ER and PM compartments. Our work does not address the issue of whether or not such associations among heterodimers have significant modulating effects on the enzyme activity, trafficking, or cellular stability of the Na,K-ATPase. It is likely that these intermolecular interactions among Na,K-ATPase heterodimers may modulate interactions between the Na,K-ATPase and other cellular proteins. The components and/or interactions that mediate the oligomerization of the Na,K-ATPase heterodimers and the potential interaction of the cytoskeleton remain to be determined.
|
FOOTNOTES |
|---|
To whom correspondence should be addressed: Dept. of Biochemistry and Molecular Genetics, University of Illinois at Chicago, Molecular Biology Research Bldg., 900 S. Ashland Ave., Chicago, IL 60607-7170. Tel.: 312-355-2732; Fax: 312-355-1765.
1 The abbreviations used are: PM, plasma membrane; ER, endoplasmic reticulum; -4,5LL, -4,5-loop-less; DDM, n-dodecyl--D-maltoside; G, Golgi apparatus; HRP, horseradish peroxidase; IPB, immunoprecipitation buffer; Chaps, 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonic acid.
|
ACKNOWLEDGMENTS |
|---|
-subunit mutant, technical guidance, and helpful discussions. |
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
