Regions of the Catalytic α Subunit of Na,K-ATPase Important for Functional Interactions with FXYD 2*

The γ modulator (FXYD 2) is a member of the FXYD family of single transmembrane proteins that modulate the kinetic behavior of Na,K-ATPase. This study concerns the identification of regions in the α subunit that are important for its functional interaction with γ. An important effect of γ is to increase K+ antagonism of cytoplasmic Na+ activation apparent as an increase in KNa′ at high [K+]. We show that although γ associates with α1, α2, and α3 isoforms, it increases the KNa′ of α1 and α3 but not α2. Accordingly, chimeras of α1 and α2 were used to identify regions of α critical for the increased KNa′. As with α1 and α2, all chimeras associate with γ. Kinetic analysis of α2front/α1back chimeras indicate that the C-terminal (Lys907-Tyr1018) region of α1, which includes transmembrane (TM)9 close to γ, is important for the increase in KNa′. However, similar experiments with α1front/α2back chimeras indicate a modulatory role of the loop between TMs 7 and 8. Thus, as long as the α1 L7/8 loop is present, replacement of TM9 of α1 with that of α2 does not abrogate the γ effect on KNa′. In contrast, as long as TM9 is that of α1, replacement of L7/8 of α1 with that of α2 does not abolish the effect. It is suggested that structural association of the TM regions of α and FXYD 2 is not the sole determinant of this effect of FXYD on KNa′ but is subject to long range modulation by the extramembranous L7/8 loop of α.

The Na,K-ATPase or sodium pump is an integral membrane protein found in the cells of virtually all higher eukaryotes. It couples the hydrolysis of one molecule of ATP to the electrogenic exchange of three intracellular Na ϩ for two extracellular K ϩ ions against their electrochemical gradients. The sodium pump is an oligomer of two subunits, a large catalytic ␣ subunit and a smaller, highly glycosylated ␤ subunit whose role is to ensure the proper folding, insertion, and maturation of ␣ in the plasma membrane. Four isoforms of ␣ and three of ␤ are expressed in a tissue-and development-specific manner. (for reviews, see Refs. 1 and 2).
The sodium pump is subject to diverse modes of regulation by substrates, membrane-associated components, hormones, and neurotransmitters (3). In recent years, considerable attention has been focused on tissue-specific regulation by small transmembrane proteins referred to as FXYD proteins (for reviews, see Refs. 4 -7). These proteins belong to a gene family of small, single transmembrane proteins, with the exception of MAT-8 which has two transmembranes (TM(s)). 3 There are at least seven members of which several (FXYDs 1-4 and 7) appear to regulate the kinetic behavior of the sodium pump in distinct and specific ways, in particular the apparent affinity of the enzyme for ligands (Na ϩ , K ϩ , and ATP). Members possess a 35-amino-acid signature sequence that includes 6 invariant amino acids before, in, and after the TM domain. The N-terminal FXYD (Phe-Xaa-Tyr-Asp) motif is invariant, along with two glycines in the TM domain, and a serine in the C-terminal domain. The first FXYD member identified and characterized was ␥ (FXYD 2), originally called the ␥ "subunit" and present predominantly in the kidney. Other members include phospholemman (PLM or FXYD 1) MAT-8 (FXYD 3), CHIF (FXYD 4), RIC or Dysadherin (FXYD 5), Phosphohippolin (FXYD 6) and FXYD 7. A phospholemman-like protein, PLMS, is present in the shark rectal gland.
The ␥ modulator exists as two main variants, ␥a and ␥b, with distinct and overlapping localization along the nephron (8,9). Mass spectrometry indicates that they differ only in the N terminus (10). In rat ␥a, TELSANH is replaced by Ac-MDRWYL in ␥b. Previous studies in our laboratory (8,(11)(12)(13)(14) using membrane fragments isolated from ␥-transfected rat ␣1-HeLa cells have shown that ␥ serves at least two distinct regulatory effects on pump kinetics and that these effects are the same for both variants. Thus, (i) ␥ increases the apparent affinity for ATP and (ii) increases K ϩ /Na ϩ competition at cytoplasmic Na ϩ activation sites, as seen by an increase in KЈ Na at high [K ϩ ] concentration. In addition, in intact cells, ␥ increases the apparent K ϩ affinity (15,16).
Although the functional effects of ␥ on Na,K-ATPase have been extensively characterized, the structural basis of these effects is largely unknown. Earlier cryoelectron microscopy of the purified renal (␣1␤1) pump (17) suggested that the ␥ subunit is located in a pocket comprising TM9, TM6, TM2, and possibly TM4. Furthermore, recent homology modeling of the Na,K-ATPase based on the high resolution structure of the Ca-ATPase as well as cross-linking experiments have shown further that ␥ makes contacts with TM2, TM6, and TM9 of the ␣ subunit (18,19).
An important issue concerns the region(s) of ␣ that interact(s) with the ␥ modulator, focusing on the interactions that are critical for producing the modulatory effects of ␥. Mutagenesis of the ␣1 subunit (Ala replacement of Phe 956 and Glu 960 ) indicate the importance of TM9 in affecting the apparent affinity for extracellular K ϩ (18). However, neither of these replacements abrogated the increase in KЈ Na effected by ␥.
The present study focuses on regions of the ␣ subunit with which ␥ interacts to effect the increase in KЈ Na at high [K ϩ ] (K ϩ /Na ϩ antagonism). Kinetic studies using ␥ mutants (14,20) and mimetic peptides comprising the transmembrane domain of ␥, ␥-TM (21), have clearly shown that the increase in K ϩ /Na ϩ antagonism is mediated by the TM region of ␥. Here we show that although ␥ interacts with ␣1, ␣2, and ␣3 and increases KЈ Na of ␣1 and ␣3, a significant change in the KЈ Na of ␣2 could not be detected. Accordingly, an ␣1/␣2 chimera approach was used to gain insight into regions of ␣ that are relevant to its functional interaction with ␥ resulting in increased K ϩ /Na ϩ antagonism.
Membrane Preparations-NaI-and non-NaI-treated microsomal membranes were prepared from the chimeric and WT cells as described earlier (23,24). The protein concentration was determined using a detergent-containing modification (25) of the Lowry method (26).
Polyacrylamide Gel Electrophoresis and Western Blotting-SDS-PAGE and Western blotting were carried out as described previously (8). Following transfer of the SDS-PAGE gels to polyvinylidene difluoride membranes, a section of the membrane comprising the lower molecular mass proteins (Յ28 kDa) was analyzed with a polyclonal ␥ antibody (␥C32) raised against the C-terminal KHRQVNEDEL peptide that is essentially the same as ␥C33 used previously (8). The remaining membrane was blotted with monoclonal antibodies to detect the ␣ subunit as described in the figure legends.
Co-immunoprecipitation-The method used was a modification of Garty et al. (27) as described elsewhere (21). Band densities were quantified using Molecular Dynamics ImageQuant software.
Enzyme Assays-Na ϩ -dependent activation of Na,K-ATPase activity was measured as described previously (14). Briefly, Na,K-ATPase activity was determined at 100 mM KCl and varying concentrations of NaCl and with 10 M ouabain present to inhibit endogenous pumps. Baseline ATPase activity was determined at 100 mM KCl and absence of NaCl and with 100 mM choline chloride added to maintain a constant (200 mM) chloride concentration. For determination of the effects of the ␥-TM peptide on ␣2, the permeabilized membranes were assayed as previously described (21), except that the peptide concentration in the final assay mixture was 1.67 M. All experiments were carried out at least three times on two clones each of mock-and ␥-transfected cells, with assays performed in triplicates.
Data Analysis-The data were analyzed using the Kaleidagraph computer program (Synergy Software) with the non-interactive model of cation binding described by Garay and Garrahan (28) where v represents the rate of the reaction, V max is the maximal rate, and KЈ Na is the apparent affinity for Na ϩ . Superscript 3 denotes the number of Na ϩ binding sites. Values of V max and KЈ Na were obtained from this fitting procedure.

RESULTS
Earlier studies with cRNA injected Xenopus oocytes showed that the ␥ subunit can associate with the three major isoforms of the ␣ subunit (15). In the present study, this finding was confirmed with mammalian cells. For these experiments rat ␣1-, ␣2-, and ␣3-HeLa cells were transfected with ␥ (␥b variant) and following isolation of stable transfectants, membranes were isolated, solubilized with detergent, and subjected to immunoprecipitation with anti-␥ antibodies using the conditions developed by Garty et al. (27) as outlined under "Experimental Procedures." These authors showed that co-immunoprecipitation is efficient only in conditions (solubilization in C 12 E 10 in the presence of either Rb ϩ ϩ ouabain as used here, or Na ϩ ϩ oligomycin) that preserve native pump structure (27). Samples taken before and after immunoprecipita- Membranes were solubilized and immunoprecipitated with anti-C-terminal antibody ␥C32 as described under "Experimental Procedures." Aliquots corresponding to 5 and 45 g of the original solubilized protein taken before (Ϫ) and after (ϩ) immunoprecipitation, respectively, were resolved by SDS-PAGE and analyzed by Western blotting using the following monoclonal antibodies: for the top section of the blot, A277 (Sigma) for ␣1, McB2 (gift of Dr. K. Sweadner) for ␣2 and A273 (Sigma) for ␣3; for the bottom section, ␥C32 to detect ␥. Efficiencies of co-immunoprecipitation as determined by densitometry of the bands (ratio of the ␣/␥ density postimmunoprecipitation/preimmunoprecipitation) were 6.2% for ␣1␥b, 5.9% for ␣2␥b, and 6.2% for ␣3␥b.

TABLE 1
Isoform specificity of the ␥-mediated modulation of K ؉ /Na ؉ antagonism KЈ Na was measured at either 100 mM K ϩ (␣1, ␣2) or 50 mM K ϩ (␣3) and represents the average of at least three experiments (average Ϯ S.E.).

Functional Interactions of FXYD 2 with Na,K-ATPase
tion were subjected to SDS-PAGE and Western blotting with anti-␣specific antibodies as indicated. The results (Fig. 1) show that ␣1, ␣2, and ␣3 associate with ␥ because all are immunoprecipitated by anti-␥ antibodies. As shown by Lindzen et al. (20) a relatively small fraction (Ϸ5%) of the total protein is precipitated under these same conditions. As indicated in the legend to Fig. 1, quantitation of the bands indicates that the efficiency of immunoprecipitation is similar (Ϸ6%) for all three isoforms.
Effects of ␥ on the Na ϩ -activation Kinetics of the ␣ Isoforms-Although ␥ associates with all three isoforms of ␣, the question remains whether the ␥-mediated increase in K ϩ /Na ϩ antagonism seen previously for ␣1 also holds true for ␣2 and ␣3. In the experiments described below, Na,K-ATPase activity for ␣1 and ␣2 was measured with NaCl varied and KCl kept constant at 100 mM. ␣3 was assayed at 50 mM KCl because KЈ Na of WT ␣3 is high (16.5 mM at 100 mM K ϩ , see Ref. 29) and even higher with ␣3-␥b membranes, precluding satisfactory extrapolation to V max . The results presented in Table 1 show that, whereas ␥ increases KЈ Na of ␣1 and ␣3 by almost Ϸ1.6-fold, a significant change in KЈ Na of ␣2 could not be detected. It is noteworthy that like the effect of full-length ␥, the transmembrane mimetic peptide of WT ␥ (␥-TM) increases KЈ Na of ␣1 but not ␣2.
The lack of KЈ Na modulation of ␣2 pumps by ␥, despite the high sequence identity between ␣1 and ␣2, provided a unique opportunity to use an ␣1/␣2 chimera approach to identify regions of ␣ interaction with ␥, which are functionally important for the KЈ Na effect. Fig.  2 depicts a linear representation of the ␣1-front half/␣2-back half (␣1 f /␣2 b ) and the reverse ␣2-front half/␣1-back half (␣2 f /␣1 b ) chimeras used for these experiments. The high degree of amino acid identity (86%) among the isoforms notwithstanding, the regions of most diversity between ␣1 and ␣2 are within residues 1-311 and 429 -565 (␣1 numbering), which span the cytoplasmic N terminus and part of the large cytoplasmic loop between TM4 and TM5 (see top of Fig. 2).
Association of ␥ with ␣1 f /␣2 b and ␣2 f /␣1 b Chimeras and Its Effects on Their Na ϩ -activation Kinetics-The co-immunoprecipitation experiment shown in Fig. 3 indicates that similar to its association with ␣1 and ␣2 (see Fig. 1), ␥b associates with all ␣1 f /␣2 b and ␣2 f /␣1 b chimeras.
In the experiments depicted in Fig. 4, the effects of ␥ on the apparent KЈ Na of ␣1 f /␣2 b and ␣2 f /␣1 b chimeras were measured at 100 mM K ϩ and compared with the effects of ␥ on WT ␣1 and ␣2 isoforms.  Table 2). Fig. 4 (top panel), Fig. 5, and Table 2 indicate that ␥ causes a significant increase in the KЈ Na of all chimeras in the ␣2 f /␣1 b configuration, similar to that reported for WT ␣1. It is notable that this holds true for chimera (␣2 (1-904) /␣1) comprising only TMs 8 -10 of ␣1 and thus encompassing TM9, which has been shown to be close to (17) and interact with ␥ (18,19).
As shown in Fig. 4 (bottom panel), Fig. 5, and Table 2, ␥ did not increase KЈ Na of chimeras ␣1 f /␣2 b up to ␣1 (1-875) /␣2, similar to its lack of effect on the WT ␣2 enzyme. However, a notable ␥-mediated increase in KЈ Na of ␣1 (1-907) /␣2 was observed. This latter effect of ␥ was not seen at low (10 mM) K ϩ concentration (experiment not shown) indicating that, like WT ␣1 (14), the ␥-mediated increase in KЈ Na of the ␣1 (1-907) /␣2 chimera reflects increased cytoplasmic K ϩ /Na ϩ antagonism. Taken together and as discussed below, these observations suggest that the  Although Table 2 shows that there are notable differences in KЈ Na values of mock-transfected ␣1, ␣2, and chimeras thereof, it is clear that the ␥-mediated increases in KЈ Na are independent of these intrinsic differences.

DISCUSSION
Earlier studies showed that the ␥ modulator has distinct kinetic effects on the Na,K-ATPase of cultured mammalian cells such as rat ␣1-transfected HeLa cells transfected with either ␥a or ␥b. One effect is a ␥-mediated increase in the apparent ATP affinity (11,12). Another effect is a ␥-mediated increase in KЈ Na at high K ϩ concentration (8,14) because of an increase in K ϩ antagonism of cytoplasmic Na ϩ activation, which can have a significant effect under phys-iological conditions in which the intracellular Na ϩ concentration limits the rate of pump activity. (For further discussion of the physiological importance of this ␥ effect, see Ref. 7). It is noteworthy that these differences in apparent affinities are consistent with the differences in affinities seen with ␣1␤1 pumps in ␥-rich kidney preparations compared with ␥-free tissues (13). An additional distinct effect, observed only in intact HeLa cells, is an increase in apparent affinity for extracellular K ϩ (16). This effect may be similar to a membrane potential-dependent decrease in K (0.5)K reported with Xenopus oocytes bathed in Na ϩ -containing medium (15,30).
Regions of ␥ Important for Modulation of Ligand Affinities-Distinct regions of ␥ are important for its aforementioned distinct effects. The cytoplasmic C terminus is responsible for the ␥-mediated decrease in KЈ ATP . This effect has been localized to its penultimate four residues (14), because their deletion abrogates the decrease in KЈ ATP . However, this ␥   Table 2 and are expressed as the percent increase in KЈ Na caused by ␥.

TABLE 2 Summary of the effects of ␥ on K Na
Na,K-ATPase assays were carried out as described in the legend to Fig. 4 effect can be modified by long range perturbations as seen in experiments in which the effect was abrogated by deletion of the first seven residues from the extramembranous N terminus (mutant ␥bN⌬7; Ref. 14). The ␥-mediated increase in KЈ Na has been attributed to the ␥-TM domain alone (21), because a ␥-TM peptide that is devoid of extramembranous regions elicits this property (also see Table 1) and a disruptive mutation in the TM helix (␥bG41R) abrogates the ␥ effect of both fulllength ␥ (14) as well as ␥-TM (21).
Regions of ␣ Involved in Functionally Important ␥-␣␤ Associations-Several earlier studies have provided insight into regions of ␣ with which ␥ associates. Thermal denaturation experiments suggested a region between TMs 8 -10 (31). Cryoelectron microscopy (17) as well as homology modeling (18) based on SERCA-phospholamban interaction (32) have pointed to the importance of TM2, TM6, and TM9.
Regions of ␥-␣␤ pump interaction relevant to the kinetic effects of ␥ are only beginning to be identified. Recent mutagenesis studies have indicated the importance of TM9 in abolishing the functional effect of ␥ on the voltage dependence of K ϩ affinity seen in ␥ cRNA-injected oocytes (18). The present study focuses on regions of the ␣ subunit that interact with ␥ to effect its modulation of KЈ Na . The results with ␣2 f /␣1 b chimeras are consistent with a key role of TM9 in mediating the K ϩ antagonism of cytoplasmic Na ϩ activation. Thus, the ␥-mediated increase in KЈ Na at high K ϩ concentration, characteristic of its effect on ␣1, is seen with all ␣2 f /␣1 b chimeras including chimera ␣2 (1-904) /␣1. In the C-terminal portion (residues 907-1018 of ␣1), only TM9 is proximal to ␥ (17)(18)(19). Even if one considers all ␣ TMs near ␥, namely TM2, TM6, and TM9 as in the model of Fuzesi et al. (19), and possibly TM4 seen by cryoelectron microscopy (17), it is notable that amino acid sequences of TM4 and TM6 of ␣1 and ␣2 are identical. In TM2, only one of the three conservative differences between ␣2 and ␣1 is also different between ␣2 and ␣3, yet ␥ increases KЈ Na of both ␣1 and ␣3, but not ␣2. In TM9, there is one non-conservative residue difference in ␣2, namely Leu 951 in ␣2 is Phe in both ␣1 and ␣3. This difference may be key to the functional importance of TM9. The putative importance of TM9 notwithstanding, the effects of ␥ on KЈ Na of the reverse ␣1 f /␣2 b suggest that its role is modified by extramembranous region(s) as discussed below.
Role of the L7/8 Loop-Similar to the lack of its effect on WT ␣2, ␥ does not increase KЈ Na of the ␣1 f /␣2 b chimeras up to ␣1 (1-875) /␣2, whereas like its effect on WT ␣1, ␥ increase KЈ Na of ␣1 (1-907) /␣2. Thus, replacement of residues 873 SRLLGIRLDWDDRTT 887 in the extracellular L7/8 loop of ␣2 with the analogous residues of ␣1 ( 876 FHLLGIRET-WDDRWI 890 ), 5 points to an important but complex role of the L7/8 loop. Thus, as long as the ␣1 L7/8 sequence is present, replacement of TM9 of ␣1 with that of ␣2 does not abrogate the ␥-dependent increase in KЈ Na . On the other hand, as long as TM9 is that of ␣1, replacement of the ␣1 L7/8 sequence with the analogous sequence of ␣2 does not abolish the KЈ Na effect. Interestingly, there are six non-conservative residue differences between ␣1 and ␣2, some of which presumably underlie the isoform-distinct effect of ␥ on KЈ Na .
The aforementioned role of the L7/8 loop may not be surprising. This loop appears to be the focal region for ␥-␣-␤ interactions. Accordingly, at the extramembranous extracellular region, the ␤ subunit interacts with L7/8 of ␣1 as shown in several earlier studies (19,(33)(34)(35)(36), and ␥ can also be cross-linked to ␤ (19), and as already suggested (19), the site of ␥-␤ interaction is probably close to the site of ␣-␤ interaction.
Although the ␣1 versus ␣2 difference in L7/8 is probably key to the loss of the ␥ effect in ␣2, the effect must be long range and quite likely affecting associations within the membranous regions because the isoform-distinct effects of ␥ are seen in the present study in which the addition of only the transmembrane region of ␥ increased KЈ Na of ␣1 but not ␣2 pumps (see Table 1). It is noteworthy that, like ␥, the functional effects of FXYD7 are also isoform-specific (37). Like ␥, FXYD 7 associates with ␣1, ␣2, and ␣3, but its effect on affinity for K (0.5)K (decrease in apparent affinity) is seen with ␣1 and ␣2, but not ␣3. Similarly to ␥, FXYD 7 resides in a groove made up of TMs 2, 6, and 9 of the ␣ subunit (18).
Conclusion-Our findings support the view that the structural associations of transmembrane regions of the catalytic ␣ subunit with FXYD proteins are not necessarily the sole determinants of the kinetic effects of these regulators on cation affinities. In the case of ␥, the extramembranous L7/8 loop of ␣ appears to modulate intramembranous ␣-␥ interactions to effect the ␥-mediated increase in K ϩ antagonism of cytoplasmic Na ϩ activation.