Expression and Functional Role of the γ Subunit of the Na,K-ATPase in Mammalian Cells*

The functional role of the γ subunit of the Na,K-ATPase was studied using rat γ cDNA-transfected HEK-293 cells and an antiserum (γC33) specific for γ. Although the sequence for γ was verified and shown to be larger (7237 Da) than first reported, it still comprises a single initiator methionine despite the expression of a γC33-reactive doublet on immunoblots. Kinetic analysis of the enzyme of transfected compared with control cells and of γC33-treated kidney pumps shows that γ regulates the apparent affinity for ATP. Thus, γ-transfected cells have a decreasedK′ATP as shown in measurements of (i)K′ATP of Na,K-ATPase activity and (ii) K+ inhibition of Na-ATPase at 1 μm ATP. Consistent with the behavior of γ-transfected cells, γC33 pretreatment increases K′ATP of the kidney enzyme and K+ inhibition (1 μm ATP) of both kidney and γ-transfected cells. These results are consistent with previous findings that an antiserum raised against the pig γ subunit stabilizes the E 2(K) form of the enzyme (Therien, A. G., Goldshleger, R., Karlish, S. J., and Blostein, R. (1997) J. Biol. Chem. 272, 32628–32634). Overall, our data demonstrate that γ is a tissue (kidney)-specific regulator of the Na,K-ATPase that can increase the apparent affinity of the enzyme for ATP in a manner that is reversible by anti-γ antiserum.

The functional role of the ␥ subunit of the Na,K-ATPase was studied using rat ␥ cDNA-transfected HEK-293 cells and an antiserum (␥C33) specific for ␥. Although the sequence for ␥ was verified and shown to be larger (7237 Da) than first reported, it still comprises a single initiator methionine despite the expression of a ␥C33-reactive doublet on immunoblots. Kinetic analysis of the enzyme of transfected compared with control cells and of ␥C33-treated kidney pumps shows that ␥ regulates the apparent affinity for ATP. Thus, ␥-transfected cells have a decreased K ATP as shown in measurements of (i) K ATP of Na,K-ATPase activity and (ii) K ؉ inhibition of Na-ATPase at 1 M ATP. Consistent with the behavior of ␥-transfected cells, ␥C33 pretreatment increases K ATP of the kidney enzyme and K ؉ inhibition (1 M ATP) of both kidney and ␥-transfected cells. These results are consistent with previous findings that an antiserum raised against the pig ␥ subunit stabilizes the E 2 (K) form of the enzyme (Therien, A. G., Goldshleger, R., Karlish, S. J., and Blostein, R. (1997) J. Biol. Chem. 272, 32628 -32634). Overall, our data demonstrate that ␥ is a tissue (kidney)-specific regulator of the Na,K-ATPase that can increase the apparent affinity of the enzyme for ATP in a manner that is reversible by anti-␥ antiserum.
The Na,K-ATPase is the sodium pump protein responsible for maintaining the electrochemical gradient present across the membranes of most animal cells (1). It consists of at least two subunits, ␣ and ␤, each of which exists as one of several isoforms (␣ 1 , ␣ 2 , ␣ 3 , and ␣ 4 and ␤ 1 , ␤ 2 , and ␤ 3 ; for review, see Ref. 2). The ␣ subunit, also known as the catalytic subunit, contains the binding sites for the enzyme's nucleotide and cation substrates, as well as the catalytic and regulatory (calcium-dependent and cAMP-dependent protein kinase C and A, respectively) phosphorylation sites. The role of the ␤ subunit is less clear, but it is required for normal processing and expression of the enzyme and may have a role in regulating the interaction of cations with the ␣ subunit (3). The different isoforms of the pump are expressed in a tissue-and develop-ment-specific fashion and are believed to be distinct in both function and modes of regulation (2).
A small single-transmembrane protein called the ␥ subunit was originally believed to be a third subunit of the pump. It was discovered by Forbush et al. (4) in 1978 and later cloned in rat, mouse, cow, sheep (5), human (6), and Xenopus laevis (7); it has sequence homology to a family of channel-inducing peptides (8 -10). Although its function has remained elusive, experiments in Xenopus oocytes have shown that the ␥ subunit alters the K ϩ affinity of the pump in a voltage-and Na ϩ -dependent fashion (7) and may induce cation channel activity (11). Our recent Western blot analysis using an anti-␥ antiserum indicated that ␥ protein is detected only in the kidney medulla but not in other tissues tested (red blood cells, heart, brain, and kidney glomerulus) including cultured cell lines derived from cells of the kidney tubule. We showed that the antibodies bound to the cytoplasmic tail of ␥ and stabilized the E 2 form of the enzyme, presumably by disrupting ␣-␥ interactions (12).
In this report we show that expression of ␥ in cells devoid of this protein results in a significant increase in apparent affinity for ATP and that the ␥-transfected cells resemble the ␣ 1 ␤ 1 ␥ kidney enzyme in that this effect is abrogated by antiserum raised against a 10-residue peptide of the C terminus of the ␥ subunit.

EXPERIMENTAL PROCEDURES
Antibodies-␥C33 is a rabbit polyclonal antiserum raised against a peptide representing the C-terminal 10 amino acids of the ␥ subunit. In the experiments reported herein, ␥C33 was used, and a control nonimmune serum was obtained from the same rabbit prior to immunization. The peptide, KHRQVNEDEL, was synthesized at the Alberta Peptide Institute, University of Alberta, and used either as the free peptide for competition studies or linked to keyhole limpet hemocyanin and emulsified with Freund's adjuvant before injection into rabbits. Antibody 6H is a mouse monoclonal antibody specific for the ␣ 1 isoform of the Na,K-ATPase, and was a generous gift from Dr. Michael Caplan, Yale University. Horseradish peroxidase-labeled secondary antibodies (donkey anti-rabbit) were purchased from BIO/CAN Scientific.
5Ј-RACE and pREP4-␥ Synthesis-5Ј-Rapid amplification of cDNA ends (5Ј-RACE) 1 was carried out using CLONTECH Marathon-ready cDNA from rat kidney following the manufacturer's instructions. Appropriate primers (see below) were synthesized, and the ␥ subunit gene sequence was amplified by polymerase chain reaction. The 5Ј-end primer contained a site for HindIII endonuclease (boldface), a Kozak sequence (underlined; see Ref. 13), and the first 24 bases of the ␥ subunit gene as determined by 5Ј-RACE (GGGGGGGAAGCTTGC-CGCCACCATGACAGAGCTGTCAGCTAACCAT). The 3Ј-end primer contained a BamHI endonuclease site (boldface) and bases complementary to the last 24 bases of the ␥ subunit gene as determined by Mercer et al. (5) (GGGGGGGATCCGTCACAGCTCATCTTCATTGACCT). The resulting DNA was then cleaved with these endonucleases and ligated into the corresponding sites of pREP4 vector (Invitrogen) to make pREP4-␥. Sequencing of the recombinant plasmid was carried out using a Pharmacia T7 sequencing kit. pREP4 and pREP4-␥ DNA used for * This work was supported by grants from the Medical Research Council of Canada (MT-3876 to R. B.), the Quebec Heart and Stroke Foundation (to R. B.), and the Weizmann Institute Renal Research Fund (to S. J. D. K.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The necleotide sequence reported in this paper has been submitted to the GenBank™/EBI Data Bank with accession number AF129400.
§ Recipient of a predoctoral scholarship from the Fonds pour la Formation de Chercheurs et l'Aide à la Recherche.
ʈ Transfections, Tissue Culture, and Membrane Preparations-HEK-293 cells at 50% confluency in a 14-cm culture plate were transfected with pREP4 or pREP4-␥ using FuGENE 6 reagent (Roche Molecular Biochemicals) and following the manufacturer's instructions. Cells were selected for 10 days, divided among 5 ϫ 14 cm plates, and allowed to grow to confluency (about 3 weeks) in Dulbecco's modified Eagle's medium containing 10% newborn calf serum and 400 g/ml hygromycin B. Cellular membranes from transfected cells and from rat kidney medulla were prepared by the procedure described elsewhere (12).
Western Blots-Western blot analysis and densitometry measurements were carried out as described previously (14) with the following modifications. 10% polyacrylamide gels were run on a Protean II gel electrophoresis apparatus (Bio-Rad), transferred to polyvinylidene difluoride membranes (Millipore), and blotted with 6H antibodies and ␥C33 antiserum, both at dilutions of 1:10,000.
Enzyme Assays-Na,K-ATPase and Na-ATPase assays were carried out at 37°C in a final volume of 100 l as described previously (12). For Na,K-ATPase assays, final concentrations of reactants were: 100 mM NaCl, 10 mM KCl, 40 mM choline chloride, 4 mM MgSO 4 , 1 mM EDTA, 30 mM Tris-HCl (pH 7.4), and varying concentrations of ATP as indicated. Na,K-ATPase activities shown represent the ATPase activities inhibited by 5 mM ouabain and ranged from 1500 to 4500, 130 to 180, and 110 to 130 nmol P i /mg/min for kidney, HEK-pREP4, and HEK-pREP4-␥ membranes, respectively. For Na-ATPase assays, final concentrations of reactants were: 20 mM NaCl, 20 mM choline chloride, 2 mM MgSO 4 , 1 mM EDTA, 5 mM EGTA, 20 mM histidine-Tris (pH 7.4), and 1 M ATP. To determine effects of K ϩ on Na-ATPase, choline chloride was replaced by the indicated concentrations of KCl. For assays of effects of anti-␥, membranes were preincubated for 1 h at 4°C in the presence of immune (␥C33) or nonimmune (preimmune) sera at a ratio of 1:100. For experiments of K inhibition of Na-ATPase, the sera were dialyzed for 48 h at 4°C against three changes of 1000 volumes of 5 mM imidazole (pH 7.4). When present, the 10-mer peptide was used in the antiserum preincubation at a concentration of 20 M. KЈ ATP values were calculated by analyzing ATP activation curves using the Michaelis-Menten formulation. All experiments shown are representative of at least three separate experiments, and each data point shown is the mean Ϯ S.E. of the difference between triplicate determinations carried out in the absence and presence of ouabain.

RESULTS
We showed previously that the ␥ subunit protein is expressed in a tissue-specific manner. Of the various rat tissues analyzed by Western blotting (kidney medulla, kidney glomerulus, red blood cells, heart, and axolemma), ␥ was detected only in the kidney medulla (12). In more recent experiments (not shown) this analysis has been extended to additional tissues of the rat, namely the lung, small intestine, stomach, and spleen. The ␥ protein could not be detected in these tissues except for a trace amount in the spleen (relative to ␣, amounting to Յ2% of that present in the kidney medulla). The kidney-specific presence of ␥ also holds true with mouse tissues (kidney, axolemma, and heart) analyzed similarly. 2 Expression of the ␥ Subunit in Mammalian Cells-Our earlier evidence of a modulatory role for the ␥ subunit on the conformational equilibrium of the Na,K-ATPase reaction was inferred from studies of the effects of an anti-␥ antiserum on enzymatic activity. To evaluate directly the functional role of ␥, it was essential to transfect cDNA encoding ␥ into mammalian cells devoid of ␥. An additional goal of such experiments was to establish the basis for the existence of ␥ as a doublet in the rat (5) as it is in the Xenopus kidney (7). Accordingly, we first used 5Ј-RACE to ascertain that the previously reported cDNA of the rat ␥ subunit comprised the full-length sequence and, if not, whether the doublet in Western blots is secondary to the presence of an additional start codon in the mRNA for the ␥ subunit as is the case for Xenopus kidney (7). The resulting sequence, shown in Fig. 1, confirmed the presence of a single initiator methionine. However, the ␥ cDNA thus obtained encodes a protein of 66 rather than 58 residues, as originally reported (5), and corresponds to the sequence subsequently revised by Minor et al. (11). The calculated molecular mass is 7237 Da. The dichotomy may be the result of either a cloning artifact or, possibly, an isoform variant. 3 Efforts to express ␥ in HeLa and HEK cells using a standard stable transfection system resulted in levels of expression that, compared with the kidney, were considered too low (␥:␣ Յ 0.1) given the relatively modest effects of anti-␥ on the kidney enzyme. In an effort to increase the level of expression, we used the plasmid pREP4 that combines the advantages of "classical" transient and stable expression systems. In addition to a hygromycin resistance gene, this plasmid contains an origin of replication that allows it to remain expressed episomally for several weeks in the nuclei of primate and canine cells. Thus, hygromycin can be used to select for cells that contain multiple copies of the gene (rather than just one). Accordingly, we subcloned the gene for ␥ (revised sequence shown in Fig. 1) in pREP4 and transfected HEK-293 cells with both recombinant and wild type plasmids. Membranes were made from the transfected HEK-pREP4-␥ and control HEK-pREP4 cells, and the amount of ␥ subunit protein relative to ␣ subunit protein was estimated by a comparison with kidney membranes using Western blot analyses of both the ␥ and ␣ subunits.
The blots shown in Fig. 2 indicate that the ␥ doublet is present in both kidney and HEK-pREP4-␥ membranes but not in control HEK-pREP4 membranes. The densities of the ␥ subunit doublet and ␣ subunit band of HEK-pREP4-␥ were compared with those of the kidney using several dilutions and varying times of exposure to film. We determined that pREP4-␥ membranes contain 34 Ϯ 12% (S.E.) of the amount of ␥ present in the kidney after normalizing for ␣ 1 densities. Assuming that 2 A. Therien the ␥:␣ ratio of kidney is 1:1 (7,15,16), this indicates that the stoichiometry of the ␥:␣ proteins in HEK-pREP4-␥ Ϸ 1:3. That this ratio reflects ␥ associated with ␣ was confirmed in Western blots of immunoprecipitates using the antibody 6H (not shown).
Functional Effects of ␥-We showed earlier that binding of antibodies raised to the ␥ polypeptide doublet associated with the pig kidney Na,K-ATPase binds to the cytoplasmic tail of the ␥ subunit (12). This binding was associated with partial inhibition of the Na,K-ATPase activity. Moreover, inhibition varied as a function of conditions that affect the rate-limiting step(s) during steady-state hydrolysis, for example varying pH. Thus, the inhibition (Ϸ30%) observed under conditions of optimal concentrations of substrates and at pH 7.4 decreased as the pH level increased and increased as pH decreased. We concluded that the antiserum caused a shift in the E 1 7 E 2 (K) equilibrium toward E 2 (K).
To maximize the inhibitory effect of the antiserum, particularly for tests of the effect of ␥ in the transfected cells in which the ␥:␣ ratio is lower than in the kidney medulla, we tested the prediction that inhibition would be greater at suboptimal ATP concentrations, under which conditions the E 2 (K) 3 E 1 sequence becomes even more rate-limiting (17). For these experiments, a 10-residue peptide representing the C terminus of the ␥ subunit was synthesized and used for the production of ␥C33 antisera, allowing confirmation of the specificity of the anti-␥ effects and providing free 10-mer peptide for competition studies. Fig. 3A shows a representative experiment on the effects of anti-␥ (serum ␥C33) on Na,K-ATPase activity of renal enzyme at near saturating (1 mM) and subsaturating (10 M) concentrations of ATP. As predicted, inhibition increases as the ATP concentration is lowered, from 36 Ϯ 4% inhibition at 1 mM ATP to 70 Ϯ 11% at 10 M ATP (averages of several experiments). In addition, the presence of excess amounts of free peptide corresponding to the C terminus of ␥ during the preincubation reversed completely the inhibition observed at both ATP concentrations; no effect on the activity of nonimmune serumtreated enzyme was observed. Fig. 3B is a Lineweaver-Burk plot of a representative experiment showing the effect of ATP concentration on activity. It shows that pretreatment of the enzyme with antiserum ␥C33 caused a 1.8-fold increase in KЈ ATP (for values, see inset in Fig. 4). V max for ␥C33-treated enzyme was 78 Ϯ 7% that for nonimmune serum-treated enzyme. The critical implication of this result is that anti-␥ reverses an increase in affinity effected by the ␥ subunit. This hypothesis was tested in HEK-pREP4-␥ cells and HEK-pREP4 cells.
We first compared the effect of ␥C33 on HEK-pREP4-␥, HEK pREP4 cells, and kidney enzymes, all assayed at 10 M ATP. The experiment (not shown) indicated that ␥C33 caused 33 Ϯ 2  3. Effect of ␥C33 antiserum and ␥C33-reactive peptide on ATP affinity of renal pumps. A, rat renal membranes were assayed for Na,K-ATPase activity at 100 M or 1 mM ATP after preincubation in the presence of ␥C33 antiserum or nonimmune rabbit serum and in the absence or presence of peptide representing the C-terminal 10 amino acids of the ␥ subunit (used to generate ␥C33). Differences between nonimmune and immune serum-treated enzyme are significant (p Ͻ 0.01 using Student's t test). o, nonimmune serum;^, nonimmune serum ϩ peptide; ٗ, ␥C33; , ␥C33 ϩ peptide. B, rat renal membranes were assayed for Na,K-ATPase activity (v) at different ATP concentrations after preincubation in the presence of ␥C33 (open circles) or nonimmune serum (filled circles). Lineweaver-Burk plots of a representative experiment are shown. and 82 Ϯ 15% inhibition of HEK-pREP4-␥ and kidney enzymes, respectively, and had no effect on the activity of HEK-pREP4 cells. This inhibition is consistent with the aforementioned relative amounts of ␥ in kidney versus HEK-pREP4-␥ cells. Experiments were then carried out to determine whether the ␥ subunit has any effect on KЈ ATP . The plots shown in Fig. 4 indicate that the HEK-pREP4-␥ enzyme has a significantly higher affinity for ATP compared with control HEK-pREP4 enzyme (for KЈ ATP values, see inset). The ␥-mediated-1.3-fold decrease in KЈ ATP in these cells, although modest, is in fact similar to the effect of ␥ in the kidney membranes, taking into account the lower ␣:␥ ratio in the transfected cells (approximately one-third that of kidney membranes). This being the case, we used a more sensitive assay of ATP affinity to magnify the effect of ␥ and to determine whether anti-␥ antiserum can reverse its effects. This assay takes advantage of the fact that K ϩ inhibits Na-ATPase activity at a very low (1 M) ATP concentration under which condition the (low affinity) ATPactivated K ϩ deocclusion reaction becomes rate-limiting. Accordingly, this inhibition decreases as the affinity for ATP at its low affinity binding site increases (18). As shown in Fig. 5A, K ϩ is less effective at inhibiting the Na-ATPase activity of pumps of ␥-transfected membranes than of control membranes. Experiments were then carried out to test and compare K ϩ inhibition, and the effect of anti-␥ thereupon, of the enzyme of the kidney medulla, HEK-pREP4-␥, and HEK-pREP4. Fig. 5B shows the percentage inhibition at 0.2 mM KCl of these pumps in the presence of nonimmune versus immune serum. Whereas preincubation of kidney and pREP4-␥ pumps with ␥C33 effected 2.1-and 1.5-fold increases in K ϩ inhibition, respectively, no ␥C33-mediated change was detected for HEK-pREP4 pumps.

DISCUSSION
The successful transfection of the ␥ subunit into mammalian cells with sodium pumps devoid of this subunit has enabled the direct analysis of the functional role of this Na,K-ATPaseassociated protein. Although the ␥ subunit does not appear to be necessary for normal Na,K-ATPase activity (7,15,19), its role as a modulator of function is consistent with its appearance in a tissue (kidney)-specific manner.
Recently, Béguin et al. (7) have shown that the rat ␥ subunit lowers the affinity of the pump for K ϩ in cRNA-injected Xenopus oocytes, at least in the absence of Na ϩ . A ␥-mediated decrease in KЈ ATP could explain this increase in KЈ 0.5 for K ϩ because, as a first approximation, ATP and K ϩ affinities are inversely related (20). However, that result may be confounded by the use of cRNA synthesized using the original sequence for rat ␥ (5). In a recent report, the human ␥ subunit was shown to induce cation channel activity in Xenopus oocytes (11), consistent with several reports of other channel-inducing membrane peptides (8,9,21). These proteins have homology with the ␥ subunit but are generally larger, and some contain possible protein kinases C and A phosphorylation sites at their Cterminal ends that are not present in the ␥ subunit (5, 8 -10). Although we have no information regarding such a role in our transfected cells, it should be pointed out that the sequence of the putative human ␥ subunit reported in the aforementioned study contains 30 extra amino acids at its N terminus (11) that are absent in rat ␥ (cf. Fig. 1). Whether the rat ␥ subunit also has a channel function and/or this extended N terminus confers a particular functional role in forming channels in Xenopus oocytes remains to be determined.
The N-terminal sequence of the rat ␥ subunit reported here and by Minor et al. (11) is different from the one originally reported (5). That it is the correct sequence is substantiated by the following observations. First, the ␥ subunit doublet present in membranes of transfected cells corresponds in size to that of kidney membranes (Fig. 2). Second, the presence of a lysine residue at position 13 (Fig. 1) where a glutamine was originally reported (5) is in accordance with the finding that the upper band of the rat ␥ subunit is cleaved by trypsin (treatment of intact right-side-out microsomes (12)). Third, preliminary results using matrix-assisted laser desorption ionization time-offlight (MALDI-TOF) mass spectroscopy indicate that the pig kidney ␥ subunit has a length of between 64 and 67 residues, 4 consistent with a length of 66 residues reported here and in Ref. 11.
The presence of two distinct bands of ␥ has been the subject of some controversy. Whereas Mercer et al. first showed that a single RNA species could yield two protein products evidenced on Western blots using an artificial translation system (5), Béguin et al. (7) showed that in X. laevis the two bands were secondary to the presence of two distinct start codons (7). Our results with 5Ј-RACE analysis preclude the presence of distinct ATG codons for the rat protein, indicating instead that posttranslational modifications are involved, because transfection of HEK-293 cells with a gene containing single start and stop codons yielded two bands of similar mobilities to those of the kidney ␥ subunit. In addition, preliminary mass spectroscopy results are consistent with the notion that the difference between the two bands is the result of post-translational modifications. 4 The differences in the ratio of the densities of the upper to the lower band between ␥ subunits of kidney and transfected HEK membranes (see Fig. 2) suggest tissue-specific variations in post-translational modifications. Whether each band has some distinct role remains to be determined.
Overall, our results suggest an interaction between the Na,K-ATPase and the C-terminal tail of the ␥ subunit that regulates ATP affinity and that is reversible upon binding of antibodies to ␥. The finding that ␥ increases the apparent affinity for ATP in ␥-transfected cells is completely concurrent with the effect of anti-␥ on the ␣␤␥ pump of the kidney tubule. Moreover, under conditions in which K ϩ sensitivity of Na-ATPase at low ATP concentration is used as a sensitive marker of differences in ATP affinity, the reversal of the ␥ effect by anti-␥ is similar with the enzyme of ␥-transfected cells and the kidney medulla. These similarities underscore our earlier interpretation of the effect of ␥ from analysis of the effects of the anti-␥ antiserum. Whether the increased apparent affinity for ATP is, in fact, a true increase in affinity or a reflection of an alteration in conformational equilibrium toward E 1 form(s) is unclear and requires further analysis, as does the question of whether the difference in ATP affinity can be evidenced in a change in apparent affinity for extracellular K ϩ . Whatever the case, it is the change in ATP affinity that is likely to be of major physiological relevance.
The increase in apparent affinity for ATP effected by ␥ is approximately 2-fold, as evidenced in either the effect of anti-␥ on the kidney enzyme or of ␥ transfected into HEK cells, extrapolating the ratio of ␥:␣ in HEK-pREP-␥ to that of the kidney. Such a change in apparent affinity may be of critical physiological importance. Although other physiological functions may be served by the ␥ subunit (as suggested recently by Jones et al. (22)), an almost 2-fold shift in ATP affinity is a potentially important regulatory mechanism. The ␥ subunit may serve to preserve the pumping activity in cells or conditions in which the ATP level falls suddenly. Relevant to this notion is the observation that the renal outer medulla is highly prone to anoxia because it works on the brink of anoxia even in normal circumstances (23,24). That the ␥ subunit effect is reversible upon addition of anti-␥ antibodies further underscores its physiological relevance. It may be hypothesized that, like the anti-␥ antibodies, some cytosolic factor binds to the ␥ subunit and disrupts its interactions with the enzyme. Mutational analysis of the C-terminal 10 amino acids that comprise the epitope reactive with anti-␥ may provide information on specific residues involved in ␣-␥ interactions.