Mechanistic Basis for Kinetic Differences between the Rat (cid:1) 1, (cid:1) 2, and (cid:1) 3 Isoforms of the Na,K-ATPase*

Previous studies showed that the (cid:1) 1, (cid:1) 2, and (cid:1) 3 isoforms of the catalytic subunit of the Na,K-ATPase differ in their apparent affinities for the ligands ATP, Na (cid:2) , and K (cid:2) . For the rat isoforms transfected into HeLa cells, K (cid:1) ATP for ATP binding at its low affinity site is lower for (cid:1) 2 and (cid:1) 3 compared with (cid:1) 1; relative to (cid:1) 1 and (cid:1) 2, (cid:1) 3 has a higher K (cid:1) Na and lower K (cid:1) K (Jewell, E. A., and Lingrel, J. B (1991) J. Biol. Chem. 266, 16925–16930; J. S., Daly, S. E., Jewell-Motz, E. A., Lingrel, J. B, and Blostein, R. (1994) J. Biol. Chem . 269, 16668–16676). The experiments described in the present study provide insight into the mechanistic basis for these differences. The results show that (cid:1) 2 differs from (cid:1) 1 primarily by a shift in the E 1 º E 2 equilibrium in favor of E 1 form(s) as evidenced by (i) a (cid:1) 20-fold increase in IC 50 for vanadate, (ii) decreased catalytic turnover, and (iii) notable stabil-ity of Na,K-ATPase activity at

The Na,K-ATPase or sodium pump is an integral membrane protein complex that couples the exchange of three intracellular Na ϩ ions for two extracellular K ϩ ions to the hydrolysis of one molecule of ATP. The sodium pump is essential to the maintenance of the electrochemical gradients of Na ϩ and K ϩ across the plasma membrane of virtually all animal cells and consequently provides the driving force for the transport of nutrients such as glucose and amino acids into the cell. This enzyme is a P-type ATPase transporter and, as such, is directly phosphorylated on a conserved aspartate residue within its cytoplasmic domain. Both the phosphorylated and dephosphorylated forms of the enzyme exist in at least two states that undergo conformational transitions (E 1 P 3 E 2 P and E 2 3 E 1 ) that are coupled to the ion-translocating steps.
The Na,K-ATPase comprises two essential subunits: a large catalytic ␣ subunit (ϳ110 kDa), containing the phosphorylation site as well as the binding sites for Na ϩ , K ϩ , ATP and for the cardiac glycoside ouabain, and a smaller, highly glycosylated ␤ subunit (ϳ35-55 kDa) that ensures the proper folding and delivery of the ␣ subunit to the plasma membrane, possibly also modulating cation affinity (1,2). A third subunit, ␥ (ϳ7 kDa), was found in the kidney (3), where it functions as a regulator of the pump (for a review, see Ref. 4). Several isoforms of ␣ and ␤ have been described. In the case of ␣, there are four known isoforms that are expressed in a tissue-and developmentspecific manner. In the rat, they are distributed as follows. ␣1 is the ubiquitous, "housekeeping" isoform; ␣2 is expressed in muscle, nerves, and adult heart; ␣3 is present in nerves, brain, and fetal heart; and recently the protein for ␣ 4 was found in rat testis, where it may participate in spermatogenesis and sperm motility (5,6). For a comprehensive review on Na,K-ATPase isozyme structure and diversity in function, see Ref. 7.
Both our laboratory and others have compared the functional properties of the ␣1, ␣2*, and ␣3* isoforms of the rat Na,K-ATPase within the same mammalian membrane environment following transfection of their cDNAs into HeLa cells. (␣2* and ␣3* are the rat ␣2 and ␣3 isoforms rendered relatively resistant to ouabain, to permit their distinction from endogenous ouabain sensitive enzyme (8). ) The results showed that the isozymes exhibit differences in their apparent ligand affinities (see Table I). Both ␣2* and ␣3* have a 3-fold higher apparent affinity for ATP binding (low affinity site) compared with the ␣1 isoform (KЈ ATP values of 120 and 130 M, respectively, compared with 331 M) (9). In addition, while ␣1 and ␣2* show similar apparent affinities for Na ϩ and K ϩ , measurements both in broken membranes and in intact cells revealed a lower apparent affinity of ␣3* for intracellular Na ϩ and a higher apparent affinity for extracellular K ϩ (9, 10).
The present study investigates whether the aforementioned differences in apparent affinities for ligands are the consequence of primary differences in ligand interactions or whether and to what extent they reflect differences in the E 1 º E 2 conformational equilibrium. The results of the present study, together with our earlier work (11), show that the kinetic differences of the ␣2* isoform are a consequence of a shift in the poise of the conformational equilibrium toward the E 1 form of the enzyme. In the case of ␣3*, the mechanistic basis for the observed differences in ligand binding is less straightforward. The notable differences in apparent cation affinities are not explained by a shift in the poise of the E 1 º E 2 balance. Nevertheless, they do reflect, at least partly, a change in the rates of limiting steps of its reaction cycle.

EXPERIMENTAL PROCEDURES
Cell Culture and Membrane Preparation-HeLa cells expressing the rat ␣1, ␣2*, and ␣3* enzymes (a generous gift from Dr. J. B Lingrel) were grown in Dulbecco's modified Eagle's medium supplemented with 10% newborn calf serum, 100 units of penicillin G, 100 g/ml streptomycin, and 1 M ouabain as described previously (9).
NaI-treated microsomal membranes were prepared from the transfected HeLa cells as described elsewhere (8,12). Protein concentrations of the membrane preparations were determined using the Lowry assay as modified by Markwell et al. (13).
Enzyme Assays-Na,K-ATPase activity was measured as the release of 32 P i from [␥-32 P]ATP as previously described (14). Unless indicated otherwise, the membranes were preincubated for 10 min at 37°C with all reactants added except ATP. The reaction was initiated by the addition of [␥-32 P]ATP. Final concentrations for Na,K-ATPase activity measurements were 100 mM NaCl, 10 mM KCl, 3 mM MgSO 4 , 20 mM histidine (pH 7.4), 5 mM EGTA (pH 7.4), and 5 M ouabain (Sigma). To analyze kinetic behavior under various conditions and unless indicated otherwise, Na,K-ATPase measurements were carried out using 1 mM ATP as in our earlier studies (cf. Refs. 9 -11), which is reasonably close (Ն75%) to saturation. This was done in order to maximize the precision (maximal percentage of P i release due to Na,K-ATPase) of the kinetic studies of the relatively low activity transfected cells. A final concentration of 5 mM ouabain was used to determine base-line hydrolysis activity. Na-ATPase activity was measured at 1 M ATP as described previously (9), with varying amounts of added KCl and choline chloride to maintain constant chloride (40 mM) concentration. Where indicated, V max values shown were derived from activities measured at 1 mM ATP using values of KЈ ATP given in Table I.
For studies of vanadate sensitivity, inorganic orthovanadate (Fisher) solutions were prepared fresh prior to the experiment and added with the [␥-32 P]ATP solution to initiate the reaction at the indicated concentrations. Values of Na,K-ATPase obtained at various vanadate concentrations, expressed as percentages of that obtained in the absence of vanadate, were analyzed by fitting the data to a one-compartment model using a nonlinear least square analysis of a general logistic function as described elsewhere (15,16).
Effects of varying pH on Na-ATPase or Na,K-ATPase were determined as indicated in the above enzyme assays except using 30 mM MES 1 /Tris (pH values of 6.2 and 6.5) or 30 mM Tris/HCl (pH Ն7.0) instead of histidine.
Formation of K ϩ -occluded Enzyme-Assays for K ϩ dependence of K ϩ -occlusion (E 1 ϩ K ϩ 7 E 2 (K)), where E 1 is the K ϩ -free enzyme and E 2 (K) is the occluded enzyme, and the rate of deocclusion (E 2 (K) 3 E 1 ϩ K ϩ ) were measured indirectly as described elsewhere (11) with the modifications described by Therien and Blostein (17).
At least two different membrane preparations were assayed, and the data presented are representative of at least three independent exper-iments. Each value shown is the mean Ϯ S.D. of triplicate determinations.

RESULTS
In the experiments described below, the properties of ␣2* and ␣3* are described in terms of their similarities or differences relative to the ubiquitous ␣1 isoform.

Assessment of Differences in the E 1 /E 2 Conformational Equilibrium
Vanadate Sensitivity-Inorganic orthovanadate is a transition state analog of inorganic phosphate that binds to P-type ATPases in the form of the enzyme from which phosphate is released in the E 2 conformation (18). Accordingly, sensitivity of these enzymes to vanadate inhibition is a measure of the proportion of enzyme in the E 2 conformation during steady-state catalysis. Thus, for mutants of various P-type pumps, including plant and yeast proton pumps and the Na,K-ATPase, this criterion has provided insight into shifts in the E 1 /E 2 distribution (for recent examples, see Refs. 19 -21). From the present comparative analysis of the vanadate sensitivity of the highly homologous ␣1, ␣2*, and ␣3* Na,K-ATPase isoforms, it is evident that the Na,K-ATPase activity of ␣2* is ϳ20-fold less sensitive to vanadate inhibition than ␣1 (Fig. 1A), suggesting a shift in the E 1 /E 2 distribution toward E 1 . In contrast, no significant difference in vanadate sensitivity of ␣3* compared with ␣1 could be detected. Although the representative experiment shown in Fig. 1A was carried out at 1 mM ATP and 3 mM MgCl 2 , similar results (not shown) were obtained at 3 mM ATP, with 5 mM MgCl 2 added to maintain a 2 mM excess of Mg 2ϩ (IC 50 values were 0.80 Ϯ 0.15, 12.1 Ϯ 4.3, and 1.2 Ϯ 0.4 M for ␣1, ␣2*, and ␣3*, respectively). Fig. 1B shows that similar results were also obtained when pump turnover was measured in the absence of K ϩ (Na-ATPase activity), precluding effects of vanadate secondary to differences in interactions with K ϩ .
Catalytic Turnover-Catalytic turnover of the Na,K-ATPase is estimated as the ratio of V max to EP max , the latter measured at 0°C in the presence of 100 mM NaCl, with K ϩ omitted and oligomycin added to trap the enzyme in the (Na)E 1 P state (see "Experimental Procedures"). As shown in Table II, in the presence of K ϩ (Na,K-ATPase mode), the catalytic turnover of ␣2* is ϳ40% that of ␣1 as observed previously (9). The turnover of ␣3* is also lower (ϳ50% of ␣1). In contrast, when the turnover of Na-ATPase is estimated as the product of Na-ATPase/Na,K-ATPase and V max (Na,K-ATPase)/EP max , the turnover of ␣3* is similar (20 mM Na ϩ ; Table II) or even higher (100 mM Na ϩ ; not shown) than that of ␣1, the former probably reflecting the lower KЈ Na for ␣3*. Turnover of ␣2* remains reduced in both the presence and absence of K ϩ .
pH Sensitivity-The pH sensitivity profile of Na,K-ATPase and its relevance to the issue of rate-limiting reaction(s) were addressed in earlier studies of Forbush and Klodos (22). Their experiments showed that the pathway involving deocclusion of 1 The abbreviations used are: MES, 4-morpholineethanesulfonic acid; EP, the phosphoenzyme form of the Na,K-ATPase; Na ϩ cyt , intracellular Na ϩ ; K ϩ ext , extracellular K ϩ .
is partially rate-limiting at acidic pH, whereas the E 1 3 E 1 P and E 1 P 3 E 2 P processes are more rate-limiting at alkaline and neutral pH, respectively. In the present study, we compared the pH profiles of the three isoforms over the pH range of 6.0 -8.5. As shown in Fig. 2, the activity of all three isoforms declines as pH is decreased below or increased above the optimal. Compared with ␣1, the activity of ␣2* is diminished to a lesser extent at acidic pH and more at alkaline pH, consistent with relatively faster E 2 (K) 3 E 1 and slower E 1 33 E 2 P transitions and hence with a preponderance of enzyme in E 1 state(s). Although the pH profile of ␣3* resembles that of ␣1 at acidic pH, it decreases more significantly than ␣1 on the alkaline side of the optimum, the significance of which is discussed below.
Distinct Interactions with Ligands: Comparison of ␣3* with ␣1 and ␣2* Since the differences in ligand binding affinities of ␣3* and ␣1 hold true despite their similar sensitivities to vanadate, it may be inferred that ␣3* differs from ␣1 primarily in its interaction with alkali cation ligands. Nevertheless, certain aspects of ␣3* kinetic behavior have remained enigmatic and a further comparative analysis of ligand interactions was carried out to explore the mechanistic basis for ␣3*-specific properties.
Cytoplasmic Na ϩ Effects and the Role of Li ϩ as a Na ϩ Congener-To determine whether ␣3* has an intrinsically lower affinity for Na ϩ , ATP hydrolysis was measured in the absence of K ϩ and with the Na ϩ concentration varied up to 10 mM to favor interactions primarily with cytoplasmic Na ϩ activation sites. Comparisons of ␣3* and ␣1 were carried out at pH values of 6.2, 7.4, and 8.0. The results shown in Fig. 3 indicate that large differences in KЈ Na persist at all pH values tested, consistent with an intrinsic difference in apparent Na ϩ affinity between ␣1 and ␣3*. In the case of ␣1, Li ϩ can act as a congener of cytoplasmic Na ϩ as well as extracellular K ϩ (24, 25). The experiment shown in Fig. 4 was carried out to test whether this holds true for ␣2* and ␣3*. The results show that Li ϩ can replace Na ϩ and stim-

TABLE II
Catalytic turnover in absence and presence of K ϩ Na,K-ATPase turnover values were obtained from the ratio of Na,K-ATPase activity measured at 37°C in the presence of 100 mM NaCl, 10 mM KCl, and 1 mM ATP, to a maximal amount of EP measured at 0°C in the presence of 100 mM NaCl, 1 M ATP, and 20 g/ml oligomycin. S.D. values of the ratios were obtained from a Monte Carlo simulation of a joint probability distribution. Na-ATPase/Na,K-ATPase values are the averages of the ratios of Na-ATPase to Na,K-ATPase of three separate membrane preparations, where Na-ATPase was carried out at 20 mM NaCl as described under "Experimental Procedures." Na-ATPase turnover (a ϫ b) was estimated as the product of the Na,K-ATPase turnover (a) and the Na-ATPase/Na,K-ATPase ratio (b). Shown are the mean values Ϯ S.D. ulate ATP hydrolysis for both the ␣1 and ␣2* enzymes but not in the case of ␣3*. This finding underscores the large difference in cation selectivity of ␣3* compared with ␣1 or ␣2*.
Interactions with Extracellular K ϩ and Li ϩ -In an earlier study, we showed that K ϩ stimulates Na-ATPase activity of ␣2* at micromolar ATP concentration but inhibits both ␣1 and ␣3*. The explanation that E 2 (K) 3 E 1 of ␣2* is faster relative to other reactions of the cycle was substantiated by measurements of the E 2 (K) 3 E 1 rate of ␣1 and ␣2* (9). With Li ϩ replacing K ϩ , Na-ATPase activity of all three isoforms was stimulated in the order ␣2* Ϸ ␣3* Ͼ Ͼ ␣1, consistent with Li ϩ acting as a congener of K ϩ and with a known faster deocclusion compared with K ϩ (26). Nevertheless, the observation that K ϩ inhibits ␣1 and ␣3* to similar extents whereas Li ϩ stimulates ␣2* and ␣3* but not ␣1 suggests distinct properties of ␣3* with respect to interaction with extracellular alkali cations and/or the rate of another reaction step. Therefore, to gain insight into the basis for these isoform-distinct effects of Li ϩ and K ϩ , we compared ␣3* and ␣1 with respect to rates of K ϩ occlusion and deocclusion as described below.
K 0.5 for K ϩ Occlusion-The K ϩ dependence for formation of the K ϩ -occluded enzyme, E 1 ϩ K ϩ 7 E 2 (K), was measured indirectly as the decrease in phosphoenzyme (E 32 P formed by phosphorylation with [␥-32 P]ATP) following equilibration of the enzyme in the absence or presence of varying concentrations of K ϩ as described previously (9,27)). The decrease in E 32 P resulting from preincubation with K ϩ is a measure of the amount of E 2 (K ϩ ). Fig. 5 shows a representative K ϩ -occlusion profile of ␣3* and ␣1. Using the simple model B max [S]/(K 0.5 ϩ [S]) to describe K ϩ occlusion, where B max is the maximally bound enzyme (28), the values of the equilibrium constant for K ϩ occlusion, K 0.5 , were 0.031 Ϯ 0.007 and 0.046 Ϯ 0.008 mM for ␣1 and ␣3*, respectively, and, therefore, not significantly different (p Ͼ 0.5). (This contrasts with the Ն3-fold increase in K 0.5 for ␣2* compared with ␣1 (11).) Maximum formation of E 2 (K) with both ␣1 and ␣3* was observed at 1 mM KCl as shown previously for ␣1 (11). E 2 (K) 3 E 1 -The rate of the E 2 (K) 3 E 1 deocclusion process was estimated indirectly by measuring the rate of E 1 formation from E 2 (K) as described previously (11,27). The results of the representative experiment shown in Fig. 6 indicate that the K ϩ deocclusion rate for ␣3*, similar to that determined earlier for ␣2* (11), is faster than that of ␣1. The fold increase obtained in three similar independent experiments was 3.4 Ϯ 0.5. Since K 0.5 for K ϩ occlusion is similar for ␣1 and ␣3*, the rate of E 1 ϩ K ϩ 33 E 2 (K) must also be 3-fold faster for ␣3*. The implication of the faster deocclusion vis-à -vis the effects of K ϩ on Na-ATPase at micromolar ATP concentration is considered below.

DISCUSSION
In the present study, we have extended earlier comparisons of the rat ␣1, ␣2*, and ␣3* isoforms expressed in the same (HeLa) cell environment (8,10) to gain insight into the mechanistic basis for their distinct behavior. The following discus-  Fig. 1.   FIG. 3. Na ؉ activation kinetics of ␣1 and ␣3* at varying pH. Na-ATPase activity was measured in the presence of 1 M ATP with varying concentrations of NaCl as indicated. Data are presented as percentage of maximal Na-ATPase activity (control) measured at 10 mM NaCl. pH values shown were measured at room temperature. The data were fit to a noncooperative three-site model as described by Garay and Garrahan (23); i.e. v ϭ V max /(1 ϩ KЈ Na /[Na ϩ ]) 3 , where KЈ Na is the apparent affinity for intracellular Na ϩ . KЈ Na values were 0.125 Ϯ 0.036 and 0.552 Ϯ 0.018 mM at pH 6.2, 0.046 Ϯ 0.005 and 0.335 Ϯ 0.046 mM at pH 7.4, and 0.066 Ϯ 0.030 and 0.753 Ϯ 0.372 mM at pH 8.0 for ␣3* compared with ␣1 (p Ͻ 0.05 at all pH values). Symbols are as described in the legend to Fig. 1. sion focuses on the properties of ␣2* and ␣3* as they relate to those of the ubiquitous ␣1 isoform.
The ␣2 Isoform-From a comparison with the ubiquitous ␣1 isoform, it is now clear that the main distinguishing property of the ␣2* isoform is its poise in conformational equilibrium in favor of E 1 forms. There are several observations in support of this conclusion. Thus, earlier studies showed that the E 2 (K) 3 E 1 transition associated with K ϩ -deocclusion is faster for ␣2* than for ␣1 and that the apparent affinity for low affinity ATP binding is 2-3 times higher for ␣2* compared with ␣1 (8-10). Further supportive evidence is the present finding of a ϳ20fold increase in IC 50 for vanadate inhibition of ␣2* compared with either ␣1 or ␣3*, indicating a shift in equilibrium away from the E 2 state(s). Another point of evidence is the ␣2*distinct pH profile. As already discussed, the smaller decrease in activity of ␣2* at acidic pH and greater decrease at alkaline pH also implicate shifts of rates of partial reactions in favor of E 1 form(s).
It is noteworthy that the distinctive behavior of ␣2* vis-à -vis ␣1 is remarkably similar to the behavior of ␣1 mutants obtained by deletion of the first 32 residues from the amino terminus and by a Glu 233 3 Lys substitution in the cytoplasmic loop between M2 and M3. Generally similar differences from ␣1 were obtained in the aforementioned properties for these mutants, including faster E 2 (K) 3 E 1 , slower catalytic turnover, lower KЈ ATP for low affinity ATP binding, and decreased sensitivity of Na,K-ATPase activity to vanadate inhibition, all leading to the conclusion that these are mutants with the E 1 º E 2 equilibrium shifted toward E 1 (21,29). Current studies aimed to identify distinct cytoplasmic regions of ␣2* that underlie its altered conformational transitions are under way.
The ␣3 Isoform-A mechanistic explanation for the distinct kinetic properties of ␣3* is more complex. As shown previously, differences in K ϩ /Na ϩ antagonism at cytoplasmic Na ϩ activation sites underlie much of the variability in KЈ Na noted in studies with intact cells comprising high intracellular K ϩ versus studies with membrane fragments analyzed at relatively low K ϩ concentration (see Ref. 30 and Table I). When the complexity of competing effects of Na ϩ and K ϩ is eliminated by analyzing the reaction cycle in the absence of K ϩ , and under conditions of enzyme turnover with Na ϩ varied up to 10 mM to limit its interaction to primarily cytoplasm-facing sites, a markedly lower apparent affinity of ␣3* for Na ϩ is observed FIG. 4. Li ؉ -stimulated ATP hydrolysis. ATP hydrolysis at varying LiCl concentrations was determined at 1 M ATP as described under "Experimental Procedures." The data are expressed as percentage of maximal Li ϩ -ATPase activity. Symbols are as described in the legend to Fig. 1.   FIG. 5. Occlusion of K ؉ at 0°C. Formation of EP at various concentrations of KCl was determined as described under "Experimental Procedures." Data are presented as percentage of maximal E 32 P, which is the difference of E 32 P formed in the absence of K ϩ minus E 32 P formed in the presence of K ϩ . Values of maximal E 32 P were (pmol/mg) 18.66 Ϯ 0.76, and 14.47 Ϯ 0.34, for ␣1 and ␣3*, respectively. The results were fitted to a simple model of K ϩ binding (B max [K ϩ ]/(K 0.5 ϩ [K ϩ ])). Symbols are as described in the legend to Fig. 1.   FIG. 6. Rate of formation of E 1 from E 2 (K) at 10°C. Formation of E 1 was determined at the indicated times following occlusion of K ϩ as described under "Experimental Procedures." Data are presented as percentage of control E 32 P, which is the difference ((E 32 P formed in the absence of K ϩ at 0°C) minus (E 32 P formed in the presence of 4 mM K ϩ at 10°C))/((E 32 P formed in the absence of K ϩ at 0°C) minus (E 32 P formed in the presence of 4 mM K ϩ at 0°C). Symbols are as described in legend to Fig. 1.  (Fig. 3). This behavior supports the notion of ␣3-distinct cation ligation. It is noteworthy that an earlier analysis of a series of ␣1/␣3* chimeras expressed in HeLa cells failed to identify a region clearly responsible for the differences in apparent Na ϩ affinity between the two isoforms (31). This is not surprising in view of the present experiments, which indicate that a difference in cation binding or selectivity is not the sole determinant of ␣3versus ␣1-distinct apparent affinities for Na ϩ and K ϩ activation as discussed further below. The remarkable ability of ␣3* to utilize Na ϩ over Li ϩ focuses on its distinctive cation interactions with cytoplasmic Na ϩ and should be an invaluable criterion for identification of specific residues comprising either the selectivity filter, gate, or Na ϩ binding pocket of the catalytic subunit.
Despite the evidence of an intrinsic difference in KЈ Na of ␣3* compared with ␣1 and ␣2*, several observations suggest that the apparent Na ϩ and K ϩ affinities are affected, in an isoformspecific manner, by the rates of reactions associated with Na ϩ and K ϩ translocation. Thus, if one considers the sequence E 1 ϩ Na ϩ ϩ ATP 7 Na⅐E 1 P 7 E 2 P ϩ Na ϩ , a decrease in the rate of E 1 P formation from E 1 or a higher rate of the subsequent E 1 P 3 E 2 P reaction should effectively increase KЈ Na . The greater decline in activity of ␣3* at alkaline pH suggests greater rate limitation of the former reaction in the case of ␣3* (cf. Ref. 22). Furthermore, the conclusion that the phosphorylation reaction also contributes to rate limitation under V max conditions at physiological pH is supported by the lower turnover of ␣3* compared with ␣1 under near V max conditions in the presence of both Na ϩ and K ϩ (Na,K-ATPase activity).
On the other hand, compared with ␣1, the Na⅐E 1 P 3 E 2 P transition is probably intrinsically faster for ␣3* for the following reason. Inhibition of Na-ATPase by low concentrations of K ϩ is observed at micromolar ATP concentration with both the ␣1 and ␣3*, but not the ␣2*, isoform (9). For ␣1, this behavior is thought to be a consequence of the rate-limiting E 2 (K) 3 E 1 process, which at low ATP is slower than E 2 3 E 1 in the absence of K ϩ . This holds true despite the evidence for an intrinsically faster rate of formation of E 1 from E 2 (K) noted in this study.
The most straightforward explanation for the foregoing paradox is that reaction steps comprising the conformational transition of phosphoenzyme that lead to release of three Na ϩ ions to the extracellular milieu (Na 3 E 1 P 7 E 2 P ϩ 3Na ϩ ) as well as the K ϩ deocclusion process (E 2 (K) 3 E 1 ϩ K ϩ ) are proportionately faster for ␣3*, with the result that the K ϩ deocclusion process is similarly rate-limiting at low ATP concentration, for ␣3* as for ␣1. Relevant to the presumably faster Na 3 E 1 P º E 2 P ϩ 3Na ϩ for ␣3* is the observation that the human ␣3 enzyme, expressed in Xenopus oocytes, shows almost no voltage dependence of pump current over the range of Ϫ150 to ϩ50 mV in the presence of extracellular Na ϩ and K ϩ (32). (The greater voltage dependence of ␣2 observed in that study is consistent with more rate limitation of this step as well as with the preference of this isoform for the E 1 conformation.) This led these authors to suggest that there is an isoform-specific difference in the backward rate constant for the Na ϩ release step. Accordingly, less rate limitation of the Na 3 ⅐E 1 P 7 E 2 P ϩ 3Na ϩ process would be consistent with the loss of voltage dependence of the ␣3-catalyzed reaction in the presence of extracellular Na ϩ . Decreased voltage dependence may, in turn, reflect either the decreased affinity of ␣3 for allosteric Na ϩ inhibition of pump flux (33) or a decrease in Na ϩ antagonism of K ϩ ext binding (see Ref. 34). Thus, in either case, an increase in the relative rate of the forward E 1 P 7 E 2 P process of the Na ϩ and K ϩ reaction cycle is observed. Studies are currently under way to compare the voltage dependence of pump current in rat isoform-transfected HeLa cells.
With the human enzyme, little, if any difference in KЈ K of ␣3 was observed except under conditions of extreme hyperpolarization in presence of Na ϩ ext (32); in Sf-9 cells, the apparent affinity for K ϩ was somewhat lower for ␣3 compared with ␣1 (6). It is unlikely that the difference in KЈ K seen with the rat ␣3* enzyme is due to the mutation to ouabain resistance, since the difference persists when ouabain-resistant rat kidney ␣1 and rat axolemma (ϳ70% ␣3) enzymes are compared (30). More likely, the discordant findings suggest that the isoforms do not differ in their intrinsic K ϩ binding/occlusion but rather in the relative rates of reaction(s) associated with K ϩ -activated dephosphorylation of E 2 P. Accordingly, an increase in the rate of the proceeding K ϩ deocclusion process [E 2 (K) 3 E 1 ] or preceding E 1 P 3 E 2 P transition would effectively decrease and increase, respectively, the apparent affinity for K ϩ . Although the present study indicates that these two processes are faster for ␣3, the relative magnitudes of the increases probably differ when the enzyme is analyzed in both a different membrane environment and at a different temperature (Xenopus membranes assayed at ambient temperature (32) versus HeLa membranes assayed at 37°C).
In conclusion, the main distinguishing feature of ␣2* is its shift in conformational equilibrium toward E 1 . On the other hand, the ␣3* isoform resembles ␣1 in its conformational equilibrium. Its lower apparent affinity for Na ϩ cyt is due, in large part, to intrinsic differences in cation binding and, to some extent, to differences in rates of limiting forward and reverse steps in the catalytic cycle. The higher apparent affinity of ␣3 for K ϩ ext is probably secondary to differences in the rates of reactions preceding and subsequent to K ϩ -activated dephosphorylation.
Although the relationship between the kinetics of the individual isoforms and their distinct physiological roles in various tissues remains elusive, a few clues are notable. In neuronal tissue, the lower affinity of ␣3 for Na ϩ cyt suggests that ␣3 may only be activated when the Na ϩ cyt concentration reaches its maximum following repeated firing, possibly acting as a "spare pump" to restore membrane potential (8,10). It is also plausible that the higher ATP affinity of ␣3 compared with the ubiquitous ␣1 allows it to function at the low ATP concentrations found near the membrane (7). This ␣3-distinct property (␣2 as well) is reminiscent of the shift in ATP affinity of ␣1 conferred by interaction with the ␥ subunit in the renal outer medulla, which may allow the pump to function more efficiently under the near anoxic conditions found in this portion of the kidney (35). It is also plausible that the higher apparent ATP affinity of ␣2 and the maintenance of activity under acidic and metabolically depleted conditions are advantageous for skeletal muscle tissue during exercise. In fact, it has been shown that in both the rat (36) and human (37) exercise induces the translocation of ␣2␤1 pumps from an intracellular pool to the plasma membrane of skeletal muscle. That the presence of ␣3 in neonatal rat cardiac myocytes, with its low Na ϩ affinity, would be particularly effective in raising cytosolic Ca 2ϩ is supported by the coincidence of the switch from ␣3 to ␣2 to the shortening of the action potential duration in rat heart during ontogeny (38 -40). Although studies with cardiac myocytes in which ␣1 is down-regulated implicate ␣1 as having a major role in regulation of cardiac Ca 2ϩ (41), a specific role of ␣2 in Ca 2ϩ signaling during heart contraction was clearly evidenced in studies of genetically reduced levels of ␣2 in the heart (42). The information provided by the present study taken together with earlier studies of isoform-specific kinetics repre-sents an important step toward a better understanding of the basis for their distinct roles in native tissues.