The cell-surface expression and function of multisubunit plasma membrane proteins are regulated via interactions between catalytic subunits and auxiliary subunits. Subunit assembly in the endoplasmic reticulum is required for the cell-surface expression of the enzyme, but little is known about subunit interactions once it reaches the plasma membrane. Here we performed highly quantitative analyses of the catalytic (α1) and auxiliary (β1 and β3) subunits of Na+/K+-ATPase in the HeLa cell plasma membrane using isoform-specific antibodies and a cell-surface protein labeling procedure. Our results indicate that although the β-subunit is required for the cell-surface expression of the α-subunit, the plasma membrane contains more α-subunits than β-subunits. Pulse-labeling and chasing of the cell-surface proteins revealed that degradation of the β-subunits was much faster than that of the α1-subunit. Ubiquitylation, as well as endocytosis, was involved in the fast degradation of the β1-subunit. Double knockdown of the β1- and β3-subunits by RNAi resulted in the disappearance of these β-subunits but not the α1-subunit in the plasma membrane. All these results indicate that the α- and β-subunits of Na+/K+-ATPase are assembled in the endoplasmic reticulum, but are disassembled in the plasma membrane and undergo different degradation processes.
The functional expressions of multisubunit membrane proteins in the plasma membrane require proper folding of the nascent polypeptide chain and assembly with other subunits in the endoplasmic reticulum (ER). The subunit assembly in the ER is, in some cases, a prerequisite for the functional protein complex to exit the ER. The proper assembly of the subunits allows the complex to exit the ER and appear on the cell surface after a series of modifications in the Golgi apparatus. The unassembled subunits are retained in the ER and, in some cases, degraded via a quality control system. This assembly-dependent expression mechanism plays a critical role in controlling the quantity of the enzyme in the plasma membrane.
There is an increasing number of reports on the structure and function of the auxiliary subunits of membrane proteins, including ion channels, receptors and ion-motive ATPases (Chow and Forte, 1995; Flucher et al., 2005; Geering, 2006; Hanlon and Wallace, 2002; Isom, 2001; Mayer, 2005). The structures of the auxiliary subunits are extremely diverse. The β-subunits of the voltage-gated Ca2+ channel (Catterall, 1995) and the K+v channel (Rettig et al., 1994) do not carry any membrane-spanning domains and exist in the cytoplasm, whereas the β-subunits of the voltage-gated Na+ channel and the KATP channel have 1 and 17 membrane-spanning regions, respectively. The functions of the auxiliary subunits are also diverse. The auxiliary subunit of the KATP channel (SUR) affects the biophysical properties of the channel in many ways and is regarded as an intricate component of the channel (Shi et al., 2005; Teramoto, 2006). However, the auxiliary subunits of Na+/K+-ATPase and Na+ and Ca2+ channels do not play a central role in the catalytic or ion-transport functions, but rather modulate the expression and/or the activity of the catalytic subunits in the plasma membrane (Chow and Forte, 1995; Hanlon and Wallace, 2002; Isom et al., 1992; Isom et al., 1995; Perez-Reyes et al., 1992; Shi et al., 1996).
Na+/K+-ATPase is an ion pump that translocates Na+ and K+ ions against the electrochemical gradients across the plasma membrane (Skou, 1957). The enzyme is composed of a large catalytic subunit (α-subunit), a membrane-spanning auxiliary subunit (β-subunit) and a small regulatory subunit (γ-subunit). The α-subunit contains the binding sites for ATP, a specific inhibitor (ouabain) and ions (Jorgensen and Andersen, 1988; Lingrel and Kuntzweiler, 1994; Skou and Esmann, 1992). The β-subunit is a single-membrane-spanning polypeptide with a small cytoplasmic domain and a large extracellular portion, usually containing three disulfide bridges and several sugar modifications (Chow and Forte, 1995; Fambrough et al., 1994; Okamura et al., 2003; Yu et al., 1994). In the current model, the functional expression of the enzyme in the plasma membrane requires assembly of the α- and β-subunits in the ER (Fambrough et al., 1994; Geering, 1991; Geering et al., 1989). The assembled enzyme can exit the ER, pass through the Golgi complex and finally reach the plasma membrane. Thus, the β-subunit is indispensable for the structural and functional maturation of the enzyme, as well as its intracellular transport to the plasma membrane.
Because the 1:1 assembly of the α- and β-subunits is necessary for the functional expression of Na+/K+-ATPase in the plasma membrane, it has been an a priori concept that the both subunits remain as a 1:1 complex in the plasma membrane. In this report,
however, we demonstrate, by utilizing a cell-surface protein labeling method and an antibody-based protein quantifying procedure, that the plasma membrane contains more α-subunits than β-subunits.
HeLa S3 cells express one catalytic subunit isoform and two auxiliary subunit isoforms
Four isoforms of each α- and β-subunit of Na+/K+-ATPase (α1-α4, official protein symbols ATP1A1-ATP1A4; β1-β4 or ATP1B1-ATP1B4) have been identified in humans (Okamura et al., 2003). RT-PCR using isoform-specific primers indicated that HeLa S3 cells used in this study expressed only one isoform of the α-subunit (α1), which is a ubiquitous α-isoform expressed in almost all tissues and cell types (Fig. 1A). The α2-isoform (found in brain and skeletal muscle), α3-isoform (found in brain) and α4-isoform (found in testis) could not be detected in HeLa cells, consistent with a previous report (Shamraj and Lingrel, 1994). However, two β-isoforms (β1 and β3) were found in HeLa cells, but the other two isoforms (β2, found in brain and testis, and β4, found in skeletal muscle) could not be detected (Fig. 1A).
Characterization of isoform-specific antibodies and establishment of an immunoblot-based protein quantification method
To examine the subunit stoichiometry in HeLa cells, an immunoblot-based subunit-quantification method was established by utilizing isoform-specific antibodies. First, isoform-specific antibodies for each isoform (α1, β1 and β3) were prepared and carefully characterized (see Materials and Methods). The band patterns from the anti-β1 and anti-β3 antibodies were apparently different, indicating that they do not crossreact. Furthermore, these antibodies detected both glycosylated and nonglycosylated forms of the β-subunits in the immunoblot analysis; treatment of the cell lysate with N-glycosidase F completely removed sugar moieties and produced a single band at ∼32 kDa for β1 and ∼28 kDa for β3. It should be noted that the total immunoreactive signal was not changed by the glycosidase treatment, indicating that immunodetection by these antibodies was not affected by the sugar modification of the β-subunits. An additional experiment examining crossreactivity using HA-tagged β1- and β3-subunits and immunoprecipitation, confirmed the specificity of these antibodies (Fig. 1C).
As each antibody has a different sensitivity in the immunoblot analysis, the intensity ratio of the immunoreactive bands does not necessarily reflect the molecular ratio of the protein. Therefore, the relative sensitivity of these isoform-specific antibodies was examined before they were used in the quantitative analysis. In brief, tandem HA-tagged α1-, β1- and β3-isoforms were expressed in HeLa cells and immunopurified using anti-HA polyclonal antibody. Each purified sample was then subjected to immunoblot analysis using both anti-HA antibody and isoform-specific antibody (Fig. 1C). The immunoreactive signals on the immunoblot were carefully quantified and the sensitivity ratios of the antibodies were obtained by dividing the signal from the isoform-specific antibody by that of the corresponding anti-HA antibody. Based on five independent experiments, the sensitivity ratio obtained was 1.0:9.6(±0.7):7.2(±1.1) (mean ± s.d., n=5) for anti-α1: anti-β1:anti-β3 antibodies, respectively, in our experimental condition. The sensitivity of the anti-α-subunit antibody was lower than that of the anti-β-subunit antibodies because the anti-α1 antibody was originally raised against the α-subunit of chicken Na+/K+-ATPase (Takeyasu et al., 1988).
The sensitivity ratio of the antibodies obtained above was then applied to examine the molecular ratio of α1-, β1- and β3-subunits in HeLa cells. The total HeLa cell lysate was subjected to immunoblot analysis using isoform-specific antibodies under the same conditions as described for Fig. 1C (Fig. 2A). The molecular ratio of these subunits was obtained by adjusting the intensities of these immunoreactive signals with the sensitivity ratio obtained in Fig. 1C. The careful quantification of immunoreactive signals revealed that the molecular ratio of the α1-, β1- and β3-subunits in the whole HeLa cell lysate was 1.0:0.21(±0.058):0.41(±0.065) (mean ± s.d., n=4), indicating that HeLa cells express more of the α-subunit than the total amount of the β-subunits.
HeLa cell plasma membrane shows unbalanced subunit stoichiometry
To distinguish the cell-surface protein fraction from the intracellular ones, the above-mentioned procedure was utilized with a membrane-impermeable biotinylation reagent to quantify the protein on the cell surface (see supplementary material Fig. S1 for details of the labeling method). HeLa cells were incubated with a membrane-impermeable biotinylation reagent, sulfo-NHS-SS-biotin, for 1 hour at 4°C. The membrane impermeability and cell-surface specificity of this biotin labeling were confirmed by both western blot analysis and immunofluorescence microscopy (see supplementary material Fig. S1). Biotinylation of the cell surface did not affect the affinity for ouabain, a specific inhibitor of Na+/K+-ATPase, and the number of ouabain-binding sites on the cell surface (data not shown). The biotinylated cell-surface proteins were purified with streptavidin-agarose and subjected to western blot analysis using isoform-specific antibodies. The efficiency of the labeling and affinity purification was optimized so that more than 95% of the cell-surface proteins could be recovered (see supplementary material Fig. S1).
The biotin-labeled cell-surface proteins were subjected to immunoblot analysis by the same procedure described in Fig. 1C. Quantification of the immunoreactive bands revealed that the molecular ratio of α1-, β1- and β3-subunits in the plasma membrane was measured as 1.0:0.030(±0.0021):0.22(±0.062) (mean ± s.d., n=4) (Fig. 2), indicating that the plasma membrane possesses an extremely small number of the β1-subunit compared with the α1-subunit. The amount of the β3-subunit was larger than that of the β1-subunit but still much smaller than that of the α1-subunit, demonstrating that the subunit stoichiometry of Na+/K+-ATPase in the plasma membrane is not 1:1.
A summary of the quantitative analyses is shown in Fig. 2B. The relative amounts of the α1-, β1- and β3-subunits in intracellular compartments (ER, Golgi complex, and other intracellular vesicles) were obtained by subtracting the cell-surface amount from the total amount of each subunit (see the figure legend for the detailed procedure). When the total amount of the α1-subunit was counted as 1, the molecular ratio of α1:β1:β3 was 0.74:0.20:0.35 in the intracellular organelles and 0.26:0.01:0.06 in the plasma membrane. The amount of the intracellular α-subunit (0.74) was slightly larger than that of the total β-subunits (0.20 + 0.35), suggesting the existence of α1-subunits free of the β-subunits in the ER. Approximately one quarter (0.26) of the total α1-subunits exist in the plasma membrane. This surface-intracellular ratio of the α1-subunit varies among different cell types: ∼50% in chick sensory neurons (Tamkun and Fambrough, 1986), 40% in chick myogenic culture (Wolitzky and Fambrough, 1986) and 50% in mouse L-cells (our unpublished results). Although the α- and β-subunits are assembled in the ER with a 1:1 ratio and then travel to the plasma membrane, the result here demonstrated that the amount of the β-subunits in the plasma membrane is much smaller (0.01+0.06) than that of the α1-subunit (0.26).
Different lifetimes of the α- and β-subunits in the plasma membrane
The results of the quantitative analyses of the cell-surface proteins suggest that the fates of the α- and β-subunits might be different in the plasma membrane. To examine this possibility, the lifetime of the α- and β-subunits in the plasma membrane were investigated by pulse-chase experiments using cell-surface biotin labeling. HeLa cells were pulse-labeled with sulfo-NHS-SS-biotin for 1 hour at 4°C. The labeled proteins were then chased for 1, 3 or 5 hours at 37°C in normal medium. At the end of the chase period, the cells were either directly lysed or lysed after treatment with glutathione (GSH), which would cleave off the biotin moieties only on the cell surface (supplementary material Fig. S1). Subsequent affinity purification with streptavidin-agarose and immunoblot analysis characterized the stability of the cell-surface proteins.
The amount of the Na+/K+-ATPase α1-subunit was fairly constant, and internalization of the subunit was not detected during the first 5-hour chase period (Fig. 3). Thus, the α1-subunit, once expressed on the cell surface, stably remains in the plasma membrane at least for 5 hours and slowly undergoes internalization and degradation processes, which is consistent with previous pulse-chase experiments using [35S]Met which showed that the half-life of the α-subunit in cultured neurons is about 40 hours (Tamkun and Fambrough, 1986). When cells were chased in the presence of 1 μM ouabain, the α1-subunit was swiftly internalized within 1 hour and remained inside the cell for the next few hours without degradation (data not shown). These data match well those of previous studies in which ouabain treatment induces the internalization of the enzyme (Cook et al., 1982; Griffiths et al., 1983; Lamb and Ogden, 1982; Nunez-Duran et al., 1996; Nunez-Duran et al., 1988).
In contrast to the α1-subunit, the β1-subunit was swiftly internalized and degraded within 5 hours (Fig. 3). After 5 hours, the total amount of the labeled β1-subunit was reduced to about 50% (Fig. 3B). There was also accumulation of the labeled β1-subunit in the intracellular compartment (Fig. 3, +GSH), indicating that the β-subunit is internalized before being degraded. The degradation of the β3-subunit was much slower than that of the β1-subunit, but still faster than the α1-subunit; approximately 80% of the β3-subunit remained on the cell surface after 5 hours (Fig. 3B). These results suggest that the degradation of each type of subunit is independently regulated in the plasma membrane.
The ubiquitin-proteasome pathway is involved in the quick degradation of the β-subunit
To examine the degradation mechanism of the Na+/K+-ATPase subunits, the biotin-labeled cell-surface proteins were chased in the presence of various inhibitors. Treatment of HeLa cells with chloroquine, an inhibitor of the endocytotic pathway, completely abolished the internalization and the degradation of both β1- and β3-subunits (Fig. 3B), indicating that endocytosis is indeed involved in the degradation of the β-subunits. Interestingly, in the presence of an inhibitor for proteasomal degradation (LLnL), the degradation of the β1-subunit was significantly blocked (Fig. 3B). A similar result was obtained with another proteasomal inhibitor, lactacystin (data not shown). The effect of the proteasome inhibitors on the β3 degradation was relatively small compared with that on β1 (Fig. 3B).
Since the proteasome-dependent protein degradation requires the ubiquitylation of the target protein, we examined whether the β-subunits are ubiquitylated. All ubiquitylated proteins were immunoprecipitated with anti-ubiquitin antibody and the existence of the α- and β-subunits was detected by immunoblotting. Under normal conditions, none of the subunit types could be detected in the immunoblot. However, when the cells were pretreated with the proteasome inhibitor lactacystin for 5 hours and all of the immunoprecipitation procedures were performed in the presence of the inhibitor, the anti-β1-antibody detected a smear signal over 50 kDa, suggesting polyubiquitylation on the β1-subunit (Fig. 4). However, the ubiquitylation of the α- and β3-subunits was hardly detected even in the presence of lactacystin, possibly because the amounts of the internalized proteins were too small. These results match the result in Fig. 3B well, indicating that the ubiquitin-proteasome pathway is involved in the fast degradation of the β1-subunit.
The α-subunit remains in the plasma membrane in the β-subunit double-knockdown cells
The fast degradation of the β-subunits in the plasma membrane suggests that a significant amount of the α-subunit in the plasma membrane is not being assembled with the β-subunit. To examine this possibility, RNA interference was employed to knock down the β-subunits in HeLa cells. Immunoblot analysis demonstrated that 24 hours after the introduction of siRNA (for β1 or β3), the amount of the target protein in the plasma membrane was reduced to less than 5% of that of the nontransfected cells (Fig. 5A). The amount of each subunit (α1, β1 and β3) in the total cell lysate fraction and in the plasma membrane fraction was examined by immunoblot analysis.
When the β1-isoform was knocked down, the amount of the β3-subunit in the plasma membrane was increased, whereas the total amount of the β3-isoform in the cell remained constant (Fig. 5A,B). Similarly, when the β3-isoform was knocked down, the amount of the β1-isoform in the plasma membrane was increased without changes in the total cellular amount of the β1-isoform. The amount of the α-subunit in the whole cell and in the plasma membrane was not changed (Fig. 5A,B).
When both β1- and β3-isoforms were knocked down simultaneously, the amount of the α-subunit in the whole cell and in the plasma membrane decreased (Fig. 5A,B), confirming that at least one β-isoform is necessary for the expression of the α-subunit in the plasma membrane. However, even when the amounts of both β1- and β3-subunits were reduced to less than 5% of the nontransfected cells (24 hours), a significant amount of the α1-subunit (∼80%) still remained in the plasma membrane. An examination of this situation by following the time course after transfection indicated that the amounts of β1- and β3-subunits in the plasma membrane were swiftly reduced after transfection, and reached the minimum level after 24 hours, whereas more than 80% of the α1-subunit still remained in the plasma membrane after 24 hours (Fig. 5C). It should be noted that the degradation rate of the α-subunit after the complete reduction of the β-subunits in the whole cell fraction was similar to the normal degradation rate of the α-subunit examined by a pulse-chase experiment using [35S]Met (∼30 hours, data not shown). These results indicate the existence of the α-subunit in the plasma membrane that is not assembled with the β-subunits.
To investigate whether disassembled α-subunits are functional, the activity of Na+/K+-ATPase was estimated by the cell viability. The β1 and β3 double-knockdown cells maintained normal viability up to 32 hours, judged from a conventional cell proliferation assay (data not shown). As a comparison, the treatment of HeLa cells with 100 μM ouabain (KI of Na+/K+-ATPase is ∼0.1 μM) started to abolish the cell viability after several hours and completely killed the cells after 24 hours. This result indicates that the disassembled α-subunits in the plasma membrane have a pump function, at least partially, if not completely.
The amount of β-subunits but not α-subunits in the plasma membrane is affected by cell density
The existence of disassembled α-subunits in the plasma membrane suggests a dynamic rearrangement of the subunit assembly state in the plasma membrane and the possibility of the involvement of the β-subunit in cellular processes other than the regulation of the pump activity. Because previous studies demonstrated that the rat β2-subunit is involved in cell-cell-adhesion in neuronal cells (Gloor et al., 1990; Muller-Husmann et al., 1993; Schmalzing et al., 1992), the possible involvement of the β-subunits in cell-cell adhesion is also most likely in HeLa cells.
HeLa cells cultured to different confluencies (20, 40, 60, 80 and 100%) were labeled with sulfo-NHS-SS-biotin and the amount of α1-, β1- and β3-subunits in the cell-surface fraction, as well as in the total cell lysate, was quantified by immunoblot analysis as described in Fig. 2 (Fig. 6). The amount of the α1-subunit both in the total cell lysate and in the plasma membrane fraction was almost constant at different cell densities. However, the amount of the cell-surface β-subunits (β1 and β3) increased as the cell confluency became higher. This result suggests that the β-subunits, but not the α-subunit, of Na+/K+-ATPase are likely to be involved in cell-cell adhesion.
In this study, we performed quantitative analyses of the turnover of the Na+/K+-ATPase subunits in HeLa cells using subunit- and isoform-specific antibodies combined with cell-surface protein labeling. Our quantitative analyses on the cell-surface amounts of the Na+/K+-ATPase subunits suggested that once expressed in the plasma membrane, the α- and β-subunits show different behavior. The examination of cell-surface proteins using a membrane-impermeable biotinylation reagent demonstrated that there is a large excess of the α-subunit relative to the amount of β-subunits in the plasma membrane. We also found that this unusual stoichiometry results from the faster degradation of the β-subunits and that the α-subunit can exist in the plasma membrane without the β-subunit. The different behavior of the catalytic and auxiliary subunits raises several critical issues regarding the dynamics of the subunit assembly and disassembly in the plasma membrane.
Disassembly of the auxiliary subunit from the catalytic subunit occurs in the plasma membrane
The Na+/K+-ATPase β-subunit ensures the correct folding of the α-subunit upon assembly in the ER. Only the α/β-assembled enzyme with the correct folding may exit the ER, follow a maturation process in the Golgi complex and reach the plasma membrane (Beguin et al., 1998; Fambrough et al., 1994; Geering, 1991). Previous experiments using Xenopus oocytes have demonstrated that injection of cRNA for the α-subunit does not increase the cell-surface amount of the enzyme, but the co-injection of β-subunit cRNA drastically increases the amount of cell-surface enzyme (Geering et al., 1989; Yoshimura et al., 1998; Zhao et al., 1997). Many types of cells express excess α-subunits (as seen in Fig. 2) and the biosynthesis of β-subunits is the rate-limiting step for the production of functional α/β-complex (Takeyasu et al., 1989; Taormino and Fambrough, 1990). Our results from RNAi experiments (Fig. 5) also demonstrated that the β1- and β3-isoforms are complementary in delivering the α-subunit to the plasma membrane and that at least one of the β-isoforms is required for the expression of the α-subunit in the plasma membrane. Therefore, it was thought that the α/β subunit stoichiometry in the plasma membrane would be 1:1.
However, our quantitative analysis using isoform-specific antibodies and the cell-surface labeling technique demonstrated that the amount of the α-subunit is larger than that of the β-subunits in the plasma membrane (Fig. 2). Since there has been no evidence of unassembled α-subunits being transported from the ER to the plasma membrane in mammalian cells, this unusual stoichiometry is simply due to the longer lifetime of the α-subunit in the plasma membrane than the β-subunit (Fig. 3). Our RNAi experiment clearly demonstrated that when almost all of the β-subunits (β1 and β3) were knocked down, a significant amount of the α-subunit still existed in the plasma membrane (Fig. 5), indicating that the α-subunit possesses a longer lifetime than the β-subunits, and that even after disassembly from the β-subunit, it can exist in the plasma membrane without the β-subunit.
The present result seems to be contrary to the previous notion that the α:β ratio is 1:1 in the plasma membrane, even considering possible oligomeric forms of the enzyme, such as α2β2 (Taniguchi et al., 2001). However, most previous biochemical assays often used crude membrane preparations that included not only the plasma membrane fraction but also significant amounts of other intracellular membrane organelles, especially ER and Golgi complex. In this study, we clearly separated the cell-surface fraction from the intracellular fraction and performed highly quantitative analyses for each subunit and isoform. Since the cell-surface fraction of the enzyme is much smaller than the intracellular fraction (Fig. 2), the crude membrane fraction seems to primarily contain the intracellular fraction of the enzyme.
The β-subunit-independent function of the α-subunit
Our results obtained from the RNAi experiment demonstrated that a significant number of the α-subunits are separated from the β-subunits on the cell surface while maintaining significant enzymatic functionality (Fig. 5). The functionality of the unassembled α-subunit was also demonstrated in previous studies. Blanco et al. have demonstrated that the α-subunit may not necessarily require the β-subunit for the basal ATPase activity (Blanco et al., 1994). Furthermore, the amino acid sequence analysis of the Na+/K+-ATPase catalytic subunits demonstrated that α-subunits lacking the assembly domain with the β-subunit (Lemas et al., 1994; Lemas et al., 1992) exist in some lower eukaryotes (Okamura et al., 2003). Thus, it is possible that the free α-subunit in the plasma membrane also has some pump activity. The reassembly of the α- and β-subunits (DeTomaso et al., 1994) may occasionally occur in the plasma membrane. Because the different β-isoforms confer different pump activities and ATPase activity to the enzymes (Blanco and Mercer, 1998; Sweadner, 1989), this disassembly-reassembly of the subunits may regulate the pump activity in the plasma membrane.
Na+/K+-ATPase reportedly functions as a signal transducer. Na+/K+-ATPase has been shown to interact with multiple signaling proteins to transmit ouabain signal to various intracellular compartments via multiple signaling pathways (Aydemir-Koksoy et al., 2001; Kometiani et al., 1998). Binding of ouabain to Na+/K+-ATPase induces the binding of the cytoplasmic tyrosine kinase Src, resulting in the formation of an active complex that phosphorylates other signaling molecules (Haas et al., 2002). Na+/K+-ATPase also directly interacts with ankyrin and adductin, as well as proteins in caveolae. Xie and colleagues have recently reported that cultured cells contain a significant amount of non-pumping Na+/K+-ATPase in the plasma membrane (Liang et al., 2007), indicating that there are functionally different pools of Na+/K+-ATPase. These `non-pumping' pumps may consist of some portion of the disassembled α-subunits and be mainly involved in signal transduction.
The auxiliary subunit undergoes a different degradation pathway from the catalytic subunit
The polyubiquitylation of the cytosolic proteins has been known to function as a signal for proteasome-dependent protein degradation (Laing et al., 1997; Levkowitz et al., 1998). The stability of the cell-surface receptors and channels are also regulated by ubiquitylation (Hicke, 1999; Kolling and Losko, 1997; Luscher and Keller, 2001; Rotin et al., 2001; Ward et al., 1995). For the membrane proteins, monoubiquitylation is implicated in the regulation of the endocytic pathway (Aguilar and Wendland, 2003; Haglund et al., 2003; Hicke, 2001; Rotin et al., 2000; Strous and Govers, 1999). In the case of the β1-subunit of Na+/K+-ATPase in HeLa cells, the polyubiquitylation seems to function as a signal for proteasomal degradation (Fig. 4).
Interestingly, although degradation of the β1-subunit was significantly blocked by LLnL, which is an inhibitor of proteasomal degradation (Fig. 3B), both the endocytosis and the degradation of the β-subunit were almost completely blocked by chloroquine, which is an inhibitor of endocytosis (Fig. 3B). These results indicate that endocytosis is a prerequisite for the ubiquitin-dependent degradation of the β-subunit and that this rapid endocytosis is independent of the α-subunit turnover. The entire life cycle of the Na+/K+-ATPase subunits are depicted in Fig. 7.
Possible involvement of the β-subunit in cell-cell adhesion
In addition to the involvement in the intracellular transport of the α-subunit and the regulation of the pump activity, the β-subunit has been suggested to play a role in other cellular functions. The β2-isoform of mouse Na+/K+-ATPase has been identified as a cell-adhesion molecule (Gloor et al., 1990; Muller-Husmann et al., 1993; Schmalzing et al., 1992). Our results in Fig. 6 show that the amounts of the cell surface β-subunits increase when the cell density becomes higher, whereas the amount of the α-subunit does not change significantly. These results indicate that the quantity of the β-subunit in the plasma membrane is regulated depending on the cell density, and suggest the general involvement of the β-subunits in cell-cell adhesion.
The amino acid sequence analysis of the β-subunits and other adhesion molecules also supports this notion. One of the structural characteristics of the cell adhesion molecules (e.g. cadherin) is the existence of at least one Ig-fold domains in the extracellular region (Nollet et al., 2000; Takeichi, 1995). The β-subunit of the Na+-channel has an Ig-fold domain in the extracellular region and was demonstrated to be involved in cell-cell adhesion via a direct interaction with the extracellular matrix (Srinivasan et al., 1998; Undrovinas et al., 1995; Xiao et al., 1999). Recent bioinformatic analysis and x-ray crystallography demonstrated that the Ig-fold structure can be found in various proteins (Halaby et al., 1999; Pearl et al., 2005). Each Ig fold is composed of sequential and antiparallel short β-sheets, some of which are relatively hydrophobic (Halaby et al., 1999). Although the three-dimensional structure of the Na+/K+-ATPase β-subunit has not been elucidated, the prediction of the secondary structure from the primary structure demonstrated that the human β-subunits possess sequential short β-sheets in the extracellular region, some of which are relatively hydrophobic (supplementary material Fig. S2). The extracellular domain of the β-subunit may possess the Ig-fold structure for cell-cell adhesion. The β-subunit can be found only in multi-cellular eukaryotes (Okamura et al., 2003). Furthermore, our BLAST and FASTA searches of unicellular eukaryotes and prokaryotes whose genome projects have been completed did not find any proteins homologous to the β-subunit (our unpublished data).
In this report, we demonstrated that the stoichiometry of the Na+/K+-ATPase α- and β-subunits is not necessarily 1:1 in HeLa cells. Our results obtained here, together with the previous findings, support the idea that the α/β complex is necessary for intracellular transport, but becomes highly mobile once it reaches the plasma membrane. The α- and β-subunits might independently be involved in cellular functions other than ion pumping. The evolutionary process of the P-type ATPases strongly supports this possibility.
Materials and Methods
HeLa S3 cells were cultured in Dulbecco's Modified Eagle's Medium (DMEM, Sigma) containing 10% fetal bovine serum (Hyclone) at 37°C with 5% CO2. All of the experiments described in this paper were performed using the cells at 80% confluency, unless otherwise indicated.
The antibody against the Na+/K+-ATPase α-subunit was previously raised against chicken Na+/K+-ATPase α-subunit (α5) (Takeyasu et al., 1988). This monoclonal antibody recognizes the human Na+/K+-ATPase α-subunit in western blot analysis. The human β1-subunit-specific antibody (clone M17-P5-F11) was purchased from Affinity Bioreagents. The epitope of this antibody (LETYP) exists only in the β1-subunit and not in the β3-subunit of human Na+/K+-ATPase. The rabbit polyclonal antibody specific to the β3-subunit was raised against the C-terminal region of the β3-subunit (SQDDRDKFLGRVMFKITARA). The amino acid sequence of the corresponding region of the β1-subunit is SEKDRFQGRFDVKIEVKS, which shows less than 40% amino acid identity. A rabbit was immunized with the synthetic polypeptide and the antibody was affinity-purified from the antiserum by protein-G-Sepharose beads. The final concentration of the anti-β3 antibody was 1 mg/ml. The specificities of these antibodies were confirmed by immunoprecipitation and western blot analysis. Anti-ubiquitin polyclonal antibody (rabbit) was a kind gift from Dr Yokota (Yamanashi University, Japan) and anti-ubiquitin monoclonal antibody was purchased from Nippon Bio-Test Laboratories. Anti-HA polyclonal (HA11) and monoclonal (16B12) antibodies were purchased from Babco. The antibody against lamin was a kind gift from Dr Yamaguchi (Nagoya University, Japan) and anti-BiP/GRP78 antibody was purchased from Stressgen.
cDNA fragments encoding the human Na+/K+-ATPase α1-, β1- and β3-subunits were amplified by PCR from a cDNA pool of HeLa cells (Clontech) and subcloned into pBluescript II SK– vector (Stratagene). The nucleotide sequences were confirmed by a genetic analyzer ABI310 (Applied Biosystems). A tandem HA tag (YPYDVPDYAAAYPYDVPDYA) was introduced at Ala14 in the α1-subunit and Met1 in the β1- and β3-subunits. These recombinant cDNA fragments were subcloned into the mammalian expression vector pCDNA3.1(+) (Invitrogen). The standard lipofection procedure was used to introduce these expression vectors into HeLa cells (Effectene, Qiagen).
The total RNA was purified from HeLa cells with a commercial total RNA isolation kit (Qiagen) following the manufacturer's protocol. The first strand was synthesized with reverse transcriptase (ReverTra-Plus™, TOYOBO). The complementary strand of the mRNA from human brain, skeletal muscle and testis were purchased from Clontech (Quick-clone). Amplification by PCR was performed using LA-taq DNA polymerase (Takara), single-stranded DNA template and isoform-specific primers (α1, α2, α3, α4, β1, β2, β3 and β4 of human Na+/K+-ATPase; the nucleotide sequences of these primers are listed in supplementary material Table S1), which amplify an ∼300bp DNA fragment.
Determining the sensitivity ratio of the isoform-specific antibodies
Each HAx2-tagged human α1-, β1- and β3-subunit was transiently expressed in HeLa cells in a 60 mm culture dish. The cells were lysed on ice with 500 μl RIPA buffer (50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1% Triton X-100, 0.5% sodium deoxycholate and 0.1% SDS) containing protease inhibitors and the nucleus was then removed by centrifugation (3300 g for 10 minutes at 4°C). The supernatant was incubated with anti-HA polyclonal antibody (HA11, Babco) and protein-A-Sepharose beads (25 μl bead volume) at 4°C for 4 hours. The beads were washed with RIPA buffer and finally resuspended with 50 μl PBS. 5 μl of each bead suspension was subjected to SDS-PAGE and subsequent immunoblot analysis using anti-HA monoclonal antibody (16B12, Babco). The band intensity was carefully measured with image analysis software (Scion Image) to obtain the intensity ratio (α1:β1:β3=xα1:xβ1:xβ3). Different amounts of the sample were applied to the gel to ensure that the amount of protein and the signal intensity were in a linear relationship. Equal amounts of each immunoprecipitated protein were also subjected to immunoblot analysis using isoform-specific antibody [anti-α1 (1:5000 dilution of ascitic fluid), anti-β1 (1:20,000 dilution of ascitic fluid), or anti-β3 (1:5000 dilution from the purified stock, 1 mg/ml)]. The band intensities were carefully measured as described above to obtain the intensity ratio (α1:β1:β3=yα1:yβ1:yβ3). The sensitivity ratio of these antibodies was obtained as anti-α1:anti-β1:anti-β3=(yα1/xα1):(yβ1/xβ1):(yβ3/xβ3). All the quantitative immunoblot analyses using these specific antibodies described in this study were performed with the same procedure and conditions described here.
Labeling of cell surface proteins by biotin
HeLa cells were cultured in a 60 mm culture dish. The cells were washed three times with ice-cold PBS (10 minutes each) and incubated with 1% sulfo-NHS-SS-biotin or sulfo-NHS-LC-biotin (Pierce) in PBS(2+) (PBS containing 0.9 mM CaCl2 and 0.5 mM MgCl2) at 4°C for 30 minutes with gentle agitation. The biotinylation was terminated by the addition of PBS containing 50 mM glycine and 5 mg/ml bovine serum albumin. After repeated washes with PBS, the cells were either harvested or incubated in DMEM with 10% FBS for indicated periods in the presence or absence of inhibitors (1 μM ouabain, 10 μg/ml LLnL, 100 μM chloroquine) at 37°C. After 1, 3 and 5 hours, cells were either directly harvested or treated with glutathione (GSH) to strip biotins on the cell surface. For GSH stripping, the cells were washed three times with PBS and subsequently incubated for 15 minutes on ice with a solution containing 50 mM glutathione, 75 mM NaCl, 10% FBS and 75 mM NaOH.
The harvested cells were washed with PBS three times and resuspended in a buffer containing 0.5% Triton X-100, 0.1% SDS and 0.1% deoxycholate (500 μl) and incubated for 10 minutes at 4°C. The nucleus and unsolubilized cells were removed by low-speed centrifugation (800 g, for 5 minutes). The supernatant was then incubated with streptavidin-agarose beads (Pierce) (25 μl bead volume) for 4 hours at 4°C with gentle agitation. The beads were washed with PBS containing 0.5% Triton X-100 four times and resuspended in SDS-PAGE sample buffer (total 50 μl). 5 μl of the bead suspension was analyzed by SDS-PAGE and immunoblotting using the antibody against the α1-, β1- or β3-subunit of Na+/K+-ATPase. The dilutions of the antibodies were always constant in all experiments: anti-α1, 1:5000; anti-β1, 1:20,000; anti-β3, 1:5000. The bands were detected by an ECL® system (GE healthcare), captured as a digital image and quantified by using image analysis software (Scion Image).
Immunoprecipitation of ubiquitylated proteins
HeLa cells were harvested and washed with PBS three times. The cells were lysed with RIPA buffer containing protease inhibitors on ice for 10 minutes and the solubilized fraction was collected by centrifugation (1000 g at 4°C for 5 minutes). The lysate was pre-cleared with Protein-A-Sepharose beads (GE healthcare) for 4 hours at 4°C with rotation. The resin was removed by brief centrifugation and the supernatant was then incubated with polyclonal antibody or monoclonal antibody against ubiquitin together with Protein-A-Sepharose beads at 4°C for 2 hours with rotation. The beads were washed with ice-cold RIPA buffer containing protease inhibitors four times. The bound proteins were eluted with 1% SDS and analyzed by SDS-PAGE and immunoblotting.
The siRNAs against human Na+/K+-ATPase β1- and β3-subunits were purchased from Invitrogen (Stealth siRNA, the nucleotide sequence of these siRNAs is described in supplementary material Table S2). Double-stranded siRNA was introduced into HeLa cells by the standard lipofection method by following the manufacturer's protocol (Lipofectamine2000, Invitrogen).
This work was supported by a Grant-in-aid for the basic science research (B) from the Japanese Ministry of Education, Culture, Sports, Science and Technology. We thank Dr Ohniwa for the analysis of the β-subunit structure and Dr Yokota for providing the polyclonal antibody for ubiquitin.
Supplementary material available online at http://jcs.biologists.org/cgi/content/full/121/13/2159/DC1
- Accepted April 14, 2008.
- © The Company of Biologists Limited 2008