ARTICLE

Top Abstract Article References

THE DISCOVERY THAT MANY ENZYMES in the cell are expressed in multiple molecular forms has become one of the most fascinating and intriguing phenomena in biology. In 1959, Markert and Moller (134) first used the term "isozymes" to describe the separate proteins that catalyze the same biochemical reaction. Since then, researchers have looked for structural variants of enzymes to understand their physiological function. In some cases, duplication of genes or alternative posttranscriptional processing of the gene message results in isozymes with unique biological properties. The development of multiple isozymes often provides the functional versatility cells need to fulfill their physiological requirements. This review covers the present status in the continuously growing field of the isozymes of the Na-K-ATPase, emphasizing the functional properties and regulatory mechanisms of the individual Na pump polypeptides.

The Na-K-ATPase, or Na pump, is a membrane-bound protein that establishes and maintains the high internal K⁺ and low internal Na⁺ concentrations typical of most animal cells. By using the energy from the hydrolysis of one molecule of ATP, it transports three Na⁺ out in exchange for two K⁺ that are taken in. The electrochemical gradient the Na-K-ATPase generates is critical in maintaining the osmotic balance of the cell, the resting membrane potential of most tissues, and the excitable properties of muscle and nerve cells. In addition, the Na⁺ gradient provides the energy that fuels the Na-coupled transporters. These secondary transport systems mediate the translocation of ions (H⁺, Ca²⁺, Cl, PO³₄, SO²₄), substrates (glucose and amino acids), and neurotransmitters across the plasma membrane (83, 84, 109, 190, 191). In the kidney, the Na-K-ATPase plays a primary role in driving the reabsorption of Na⁺ and water. Thus the enzyme is essential in the maintenance of body fluid and electrolyte homeostasis (83, 109).

The Na-K-ATPase or Na pump belongs to a widely distributed class of P-type ATPases that are responsible for the active transport of a variety of cations across cell membranes (127, 171). P-type ATPases are found in both prokaryotic and eukaryotic cells, and are responsible for transporting H⁺, Na⁺, Mg²⁺, K⁺, Ca²⁺, Cu²⁺, and Cd²⁺ (127). All these enzymes use the hydrolysis of ATP to drive the transport of cations against an electrochemical potential. The P-type designation refers to the unique characteristic of these enzymes in forming a transient, phosphorylated aspartyl residue during the catalytic cycle. Accompanying the phosphorylation-dephosphorylation process, the P-type ATPases bind, occlude, and transport ions by cycling between two different cation-dependent conformations, called E1 and E2 (167, 171, 191). The precise molecular mechanisms that couple the hydrolysis of ATP to the conformational changes and the translocation of ions remain unknown.

In addition to having a common reaction mechanism, P-type ATPases also display comparable tertiary structures, equivalent membrane topological organization, and several highly conserved protein domains (127). The region around the phosphorylated aspartate, the TGES/A (threonine, glycine glutamine, serine/alanine) motif between transmembrane domains two and three, and several regions involved in ATP binding exhibit a high degree of amino acid homology. In addition to implying a common evolutionary ancestor, the similarities in the structure and reaction mechanism of these diverse proteins often make conclusions drawn from one transporter relevant to the others (46).

The Na-K-ATPase is an oligomer composed of stoichiometric amounts of two major polypeptides, the - and the -subunits. The primary structures and membrane organization of the Na-K-ATPase - and -subunits are depicted in Fig. 1. The -subunit is a multispanning membrane protein with a molecular mass of ~112,000 Da that is responsible for the catalytic and transport properties of the enzyme. The -subunit contains the binding sites for the cations, ATP, and the inhibitor, ouabain (reviewed in Refs. 120, 147, 160, 165). The -subunit is a polypeptide that crosses the membrane once and, depending on the degree of glycosylation in different tissues, has a molecular weight between 40,000 and 60,000 Da. The -subunit is essential for the normal activity of the enzyme (31, 52, 139), and it appears to be involved in the occlusion of K⁺ (126) and the modulation of the K⁺ and Na⁺ affinity of the enzyme (33, 35, 62, 64, 102). In addition, in vertebrate cells, the -subunit may act as a chaperone, stabilizing the correct folding of the -polypeptide to facilitate its delivery to the plasma membrane (reviewed in Refs. 52, 139).

View larger version (37K): [in this window] [in a new window]   Fig. 1.   Scheme of the membrane topology of the - and -isoforms of the Na-K-ATPase. Sequences of rat 1- and 1-isoforms are shown. Residues are colored to indicate the amino acid homology among the different -isoforms (1, 2, 3, and 4) or -isoforms (1, 2, and 3).

A third protein, termed the -subunit, has also been identified in purified preparations of the enzyme (166). The -subunit is a small, hydrophobic polypeptide of 8-14 kDa that was considered to be a contaminant of purification, until it was shown that it could be covalently labeled by photoaffinity derivatives of ouabain (78, 123, 169). Other evidence that the -subunit is a component of the Na-K-ATPase is that the subunit colocalizes with the -subunit in nephron segments and coimmunoprecipitates with -complexes (148). Also, the high degree of homology among -subunits from several species suggests that the subunit may be important in Na-K-ATPase function.

Expression studies have shown that the -subunit is not required for normal Na-K-ATPase activity (59, 91, 176). However, recently it was shown that the -subunit can modify the voltage dependence of K⁺ activation of the 1 1-isozyme when expressed in Xenopus oocytes (20). In addition, it appears that the -subunit can stabilize the E1 conformation of the enzyme (204) and may be required for cavitation in mouse embryos (108). Interestingly, the -subunit belongs to a family of small membrane proteins involved in the passage of ions across the plasma membrane. This family of proteins, which includes phospholemman (158), channel inducing factor (CHIF; Ref. 12), and Mat-8 (mammary tumor, 8 kDa; Ref. 152), induces ion channel activity when expressed in Xenopus oocytes. Consistent with these findings, the human -subunit induces cation-selective channels when expressed in Xenopus oocytes (150). The physiological significance of this activity and whether it requires the other Na-K-ATPase subunits is unknown. Although there is increasing evidence that the -subunit can modify Na-K-ATPase function, the exact role of the subunit in Na-K-ATPase function awaits further experimentation.

Na Pump Isoform Diversity

7

4

The -isoforms of the Na-K-ATPase have been cloned from several mammalian species. An analysis of the phylogenetic distribution of the -proteins using antibodies directed to conserved and specific regions of the isoforms indicates that isoforms appear to exist in all mammals, including both placental and marsupial species (164). The identifications of Na-K-ATPase -isoforms in the chicken (70) and teleost fishes (164) imply their broad presence among vertebrates (201). Moreover, the finding of -isoforms in crustaceans (14) and possibly in hydra (45) and platyhelminths (159) indicates that the divergence of Na-K-ATPase genes occurred early in evolution.

Further studies have shown that the Na pump molecular diversity also extends to its -subunit. At present, three different Na-K-ATPase -isoforms have been identified. Two of the isoforms, 1 and 2, have been found in different tissues of mammals (82, 136) and birds (115), whereas 3 has been detected in amphibians (85), mouse (27), rat, and human (82, 136). Moreover, in the zebra fish, an Na-K-ATPase -subunit highly homologous to 3 has been isolated (7). In this species, however, a detailed characterization of other -isoforms has not been accomplished, making it difficult to establish a correlation with the subunits of other species.

Both the - and -isoforms of the Na pump exhibit a tissue-specific pattern of expression. The 1-isoform in association with the 1-subunit is found in nearly every tissue (reviewed in Refs. 69, 117, 119, 121, 196). In addition, 1 1 is the principal isozyme of the kidney, a tissue that has classically been used as the source of Na-K-ATPase (reviewed in Refs. 83, 84, 109, 191). The presence of other Na-K-ATPase isoforms in the kidney has been a matter of debate (2, 8, 51, 53, 71, 73, 88, 104, 140, 206). The 2- and 3-isoforms have been detected in renal cortex, medulla, and papilla using both reverse-transcription followed by polymerase chain reaction amplification and [³H]ouabain titration analysis (125). Measurement of mRNA and protein levels indicates that the 2- and 3-isoforms correspond to only 0.1% of the 1 1 enzyme of the kidney (125). In contrast to the broad tissue distribution of 1 and 1, the other - and -polypeptides are more restricted in their expression. The pattern of Na-K-ATPase isoform expression, at both the mRNA and protein level, has been studied with most detail in the rat. The 4-isoform is a testis-specific isoform (185). The 2-isoform predominates in adipocytes (130), muscle (99, 114), heart (215), and brain (143), and 3 is abundant in nervous tissues (95, 143). In the nervous system, the expression of Na-K-ATPase isoforms reaches its greatest intricacy. Neuronal cells may express a unique isozyme or multiple isozymes consisting of the various -heterodimer combinations (117, 119, 121, 196). Whereas neurons are the principal source of the 3 polypeptide (reviewed in Refs. 117, 198), glial cells preferentially express 2 (40, 43, 76). Also, the -isoforms are distributed in a tissue-dependent manner. The 2-isoform is found in skeletal muscle (114), pineal gland (189), and nervous tissues (162), whereas 3 is present in testis, retina, liver, and lung (11, 133). In addition, the pattern of expression of the Na-K-ATPase isoforms is subjected to developmental as well as hormonal regulation and can be altered during disease (39, 49, 50, 68, 217).

Additional Mechanisms Involved in the Production of Na-K-ATPase Isozymes

Other short variants of the 1 polypeptide have been found in Xenopus laevis embryos. In Xenopus, deletions at the NH₂ terminus of the protein result from the alternative initiation of translation at codons encoding either Met³⁴ or Met⁴³. Sequential mutation of each of these residues and expression of the corresponding cRNAs in Xenopus oocytes shows that both truncated subunits are catalytically competent when coexpressed with the -isoform. Moreover, the -subunit with Met⁴³ at the NH₂ terminus exhibits a lower affinity for K⁺ than the Met³⁴ truncated or full-length polypeptides (42). These results support the hypothesis that the NH₂ terminus of the 1-subunit, which is rich in lysine and contains an excess of positively charged amino acids, forms a cation-selective gate (187, 210).

Another mechanism that contributes to the heterogeneity of the Na pump is the association of the - and -isoforms into different -heterodimers. The possibility that combinations of different heterodimers could be formed was suggested from the expression of multiple isoforms in several tissues and cells (43, 117, 162, 181, 182, 221). For example, in the pineal gland, both the 1 2 and 3 2 Na-K-ATPase isozymes are present (189). The 3 2-isoform combination is also found in retinal photoreceptor cells (181, 182), whereas the cells of the choroid plexus express the 1 1- and 2 1-isozymes (221). Specific cell populations within the central nervous system (117, 198), the ciliary epithelial cells of the eye (81), the sensory organ of the ear (202), and the choroid plexus (221) express several isoforms with the potential of multiple  and  combinations. The promiscuous association between - and -isoforms was confirmed in insect cells expressing different combinations of the rat Na-K-ATPase polypeptides. In this expression system, each -isoform could properly assemble with either the 1- or the 2-subunit into catalytically active complexes (30, 33, 35, 36). Similarly, nonspecific association was found for the Na-K-ATPase chicken isoforms exogenously expressed in mammalian cells (116). In this manner, multiple Na-K-ATPase isozymes (1 1, 1 2, 2 1, 2 2, 3 1, and 3 2) can result from the pairing of the various - and -polypeptides (33, 35).

The complexity of the Na-K-ATPase may also extend to the oligomeric structure of the enzyme. We have recently demonstrated that the different -isoforms are able to specifically and stably associate into oligomeric complexes (32). The existence of multiple sets of oligomers raises the possibility of an even greater degree of intricacy in the molecular heterogeneity of the Na pump. However, whether the quaternary structure of the Na pump influences the functional properties of the enzyme is still unclear.

Structure of - and -Isoforms

Across species the degree of identity for the 1- and 2-isoforms is ~92% and is over 96% for 3. There is also a high degree of identity (87%) among the 1-, 2-, and 3-isoforms. In contrast, 4 is the most divergent, sharing a 78% identity with the 1-isoform. As shown in Fig. 1, the highest structural variability among the isoforms occurs at the NH₂ terminus, the extracellular ouabain binding site between transmembrane segments 1 and 2, and the cytoplasmic region between amino acids 403 and 503. In contrast, the greatest similarities occur in the cytoplasmic middle region where the ATP binding and phosphorylation sites are located, the transmembrane hydrophobic regions and the COOH-terminal region (reviewed in Refs. 117, 120, 147).

The -subunit undergoes translational or posttranslational modifications, some of which appear to be isoform specific. In the mature 1- and 2-polypeptides, the first 5 amino acids are cleaved (129). It is not known whether the 3-isoform is processed in a similar fashion. A tissue-specific, posttranslational modification has been suggested for the 3-isoform. An isoform-specific antibody that reacts with a region near the ATP binding site identifies 3 in the brain but does not react to the same isoform from heart. This lack of reactivity has been interpreted as a blockage of the antibody epitope by a modification of the protein during synthesis (10). The nature of this modification, as well as its occurrence in other tissues and -isoforms, is unknown.

The amino acid sequence of the -isoforms has been deduced from the rat, human, chicken, Xenopus, and mouse cDNAs (82, 115, 133, 136). In the rat, the 1-isoform has 304, the 2-isoform has 290, and the 3-isoform has 279 amino acids. All -isoforms share a common basic structure (Fig. 1). The -isoforms consist of a short NH₂-terminal cytoplasmic region, a transmembrane spanning segment, and a large extracellular domain. The homology of the 1- and 2-isoforms across mammalian species is ~95%. This value drops to 60% when nonmammalian species are included in the comparison. The homology among different -isoforms is lower than that found for the catalytic subunit (Fig. 1). Compared with 1-, the 2-polypeptide exhibits 58% similarity (34% identity, 24% favored substitutions), whereas the 3-subunit is 68% homologous, with an identity of 39%. The similarity between 2 and 3 reaches 61%, with 49% of the residues conserved (133). Interestingly, the 2-subunit primary structure is more closely related to the H-K-ATPase -isoform than to the 1-isoform, suggesting that the 2- and H-K-ATPase -isoform genes diverged more recently than 1 and 2 (188). The transmembrane domain of the -subunit is the most highly conserved region both among isoforms and species.

All -isoforms are heavily glycosylated. The 1-isoform from mammals has three N-linked glycosylation sites. The putative N-linked glycosylation sites for the 2-isoform vary depending on the species. Thus the chicken subunit has four potential glycosylation sites, the rat has seven, the human has eight, and the mouse 2-polypeptide has nine (reviewed in Ref. 52). It is not known whether all the sites are used. Inhibition of glycosylation of the Na-K-ATPase 1-subunit with tunicamycin results in catalytically competent Na pumps with normal affinity for ouabain (200, 218, 219). Expression of an enzyme in which all -subunit N-linked glycosylation sites are mutated also renders an active enzyme with conserved K⁺ and ouabain affinities. However, the reduction in the ability of the nonglycosylated -subunit to assemble with the -subunit and the higher sensitivity of the enzyme to proteolysis suggest that glycosylation may play a role in protein folding (16).

Another important feature in the structure of the 1-subunit is the presence of three disulfide bridges, which in the rat polypeptide occur between Cys¹²⁵Cys¹⁴⁸, Cys¹⁵⁸Cys¹⁷⁴, and Cys²¹²Cys²⁷⁵. All the cysteines, but not their relative positions within the sequences, are conserved in the 2- and 3-isoforms, suggesting similarities in the tertiary structure of the polypeptides. Treatment of the Na-K-ATPase with reducing agents results in inactivation of Na-K-ATPase, suggesting that the disulfide bonds are required for enzyme function. The removal of just one of the disulfide bonds by site-directed mutagenesis of the involved cysteines has been shown to be sufficient to abolish proper assembly of the -subunits (16). The location and role of sulfhydryl bridges in the other Na pump -isoforms are unknown.

Enzymatic Properties of Na-K-ATPase Isozymes

Although several studies suggested differences in the enzymatic behavior of the Na pump isozymes, discrepancies in the reported kinetic parameters led to uncertainties in the functional characteristics of the isozymes. In contrast, the reactivity of the Na pump isoforms toward ouabain is less controversial. Because of their broad sensitivities to cardiotonic steroids, this is particularly evident for the rodent Na-K-ATPase isozymes. In the rat, for example, 1 was reported to be 100-fold more resistant to ouabain than + (2 + 3) (67, 100, 135, 194, 195). On the basis of the differential sensitivities to trypsin, early experimental evidence also suggested that the 3-isoform has a high affinity for ouabain (209). Working with partially purified Na-K-ATPase from rat brain stem, researchers identified three populations of Na-K-ATPase with different sensitivities to ouabain, possibly corresponding to 1, 2, and 3 (22, 29). In rabbit, pig, dog, and human, the differences in the cardiotonic steroid affinity may not be as marked, since the 1-isoform in these species is much more sensitive to ouabain (23, 88, 90, 92, 104, 131, 179).

With the introduction of molecular biology techniques to the Na pump field, a better understanding of the function of each isoform became possible. The availability of the cDNAs for the different Na pump isoforms allowed studies in heterologous expression systems. Consequently, several investigators have used mammalian cells to separately express each Na pump isoform. Thus, when expressed in cultured rodent cells, the rat 2- and 3-isoforms have been shown to display K_i values for ouabain of ~40 and 80 nM, respectively (89, 113). Through study of the molecular basis for the difference between primate and rodent cells in their sensitivities to the cardiac glycosides, it was shown that the rat 1-isoform could confer ouabain resistance to ouabain-sensitive CV-1 cells (65, 113). Using the differences in ouabain sensitivities between rodents and other mammals, Jewell and Lingrel (105) studied the kinetic properties of the rat 1-, 2-, and 3-subunits. By changing two residues in the extracellular loop between the first and second transmembrane segments, they converted the rat 2 and 3 into ouabain-resistant forms. The mutated isoforms were stably transfected into HeLa cells by their ability to support growth in 1 mM ouabain (105). Through this method, the expressed isoforms could be functionally separated from the endogenous ouabain-sensitive Na-K-ATPase. Analysis of the individual isozymes demonstrated that the 1 1 and 2 1 display similar affinities for Na⁺, K⁺, and ATP, whereas the 3 1-isozyme exhibits a lower affinity for Na⁺ compared with 1 1 and 2 1. Subsequently, the highly ouabain-sensitive Na-K-ATPase from rat brain or dog heart, ascribed to 31, was also reported to have a lower apparent affinity for Na⁺ than the intermediate and resistant components (1 and 2) from the same tissues (23, 80). To identify the structural basis for the difference in the Na⁺ requirements between the isoforms, a series of chimeric 1/3-subunits were expressed in HeLa cells. Analysis of the Na⁺ affinities of each chimera did not reveal a region clearly responsible for the differences in Na⁺ dependence between the isoforms, indicating that multiple residues spanning the -polypeptide may be cooperating in the binding and transport of the cation (106).

Expression of hybrid Na-K-ATPase molecules between - and -isoforms from different species in Xenopus oocytes has also helped in elucidating the characteristics of the Na pump isozymes. Through this system, the 1 1 Na-K-ATPase has been shown to exhibit a higher activation by Rb⁺ than the 1 2- and 1 3-isozymes (102, 177). Utilizing heterologous expression in yeast, which lack endogenous Na-K-ATPase activity, Farley and colleagues (97, 98) successfully obtained catalytically competent Na pump molecules by expressing the 1-subunit from sheep and the -subunit from dog. Later, they determined that the ouabain binding constant of the sheep 1 and rat 3 is between 5 and 10 nM (63). In addition, coexpression of the 1- or 3-isoforms with chimeric molecules between the Na-K- and H-K-ATPase -subunits indicated that the -polypeptide modulates the K⁺ and Na⁺ dependence of the enzyme (62, 64).

Our approach to the study of the Na-K-ATPase isozymes has included the use of the baculovirus expression system. This system employs the Autographica californica virus to express foreign genes in insect cells. By using recombinant baculoviruses coding for different - and -isoforms we have been able to successfully obtain functional Na pump molecules in cultured Sf-9 insect cells, a cell line derived from the ovary of the fall armyworm, Spodoptera frugiperda. Sf-9 cells are able to produce high amounts of the virally directed Na-K-ATPase polypeptides. Moreover, Sf-9 cells contain very low levels of endogenous Na-K-ATPase, allowing the analysis of the expressed enzymes in an environment relatively free from contaminating Na-K-ATPase activity (36, 59). In addition, the recombinant viruses can be used independently and in combination to study assembly and activity of the various Na-K-ATPase isoforms.

The insect cells are able to correctly assemble the various - and -isoforms into functional enzyme (30, 33, 35, 36, 59). Infecting cells with the various recombinant baculoviruses has demonstrated that all -isoforms can stably assemble with either the 1- or the 2-polypeptides. This is consistent with the results of other investigators (1, 116, 177). Although all possible  and  pairs result in active enzyme, some combinations, such as 1 2, exhibit lower activity, suggesting that not all associations are equally favored. Evidence for a selective interaction of certain - and -subunits has been reported using Xenopus oocytes (178). The importance of this preferential assembly in directing isozyme expression is unknown.

By analyzing the dose-response curves toward Na⁺, K⁺, ATP, and the inhibitor ouabain, the enzymatic properties of the various Na pump isozymes were determined. Table 1 summarizes the kinetic parameters for the rat Na-K-ATPase isozymes expressed in the insect cells. As shown, the apparent affinity for Na⁺ varies with a rank of order 2 2 > 2 1 > 1 1 = 3 2 > 3 1. Also, the apparent affinity for K⁺ differs among the isozymes, following the sequence 1 1 > 2 1 = 2 2 > 3 1 = 3 2. For the activation by ATP, the enzymes composed of the 2- and 3-isoforms display equivalent K_m values, which are approximately four times lower than that of the 1 1. Altogether, these results indicate that the major kinetic differences occur among Na-K-ATPases that differ in the -subunit composition. These observations are in reasonably good agreement with the kinetic properties reported for the rat 1-, 2-, and 3-isoforms expressed in HeLa cells (57, 58, 105, 154, 203). The only difference between our results and those obtained in HeLa cells is the K⁺ affinity of the 3-isozymes, which we found to be lower than that of the 1 or 2. The differences in the membrane environment of mammalian and invertebrate cells, the analysis of hybrid Na pump molecules in the HeLa cells, or the alteration of the isoforms by mutation of their ouabain binding site may be responsible for the disparity in results.

As mentioned before, the most conspicuous kinetic difference among the isozymes corresponds to the reactivity toward ouabain, with 3 1 and 3 2 displaying a high, the 2 1 and 2 2 an intermediate, and 1 1 a low sensitivity to the cardiotonic steroid (36, 155). Interestingly, the rat 1 1 enzyme expressed in insect cells is over 2-fold more sensitive and the 3 1 approximately 20-fold less sensitive to ouabain than the native enzyme. This difference in sensitivity between the native and expressed Na-K-ATPase may be a result of the different lipid environments of the enzyme. Alternatively, the sensitivity to ouabain may be influenced by the -subunit or by the oligomeric structure of the -subunit.

At present, the enzymatic properties of Na-K-ATPases composed of the 4-isoform have not been analyzed. To determine the ouabain sensitivity of the 4-isoform, we performed dose-response curves for the ouabain inhibition of Na-K-ATPase activity in membrane preparations from rat testis, the only tissue known to contain this isoform. As shown in Fig. 2A, the experimental data is best fitted assuming the existence of two isozymes with different sensitivities to the cardiotonic steroid. One of the components corresponding to ~45% of the total Na-K-ATPase has a calculated K_i of 1.4 ± 0.3 × 10⁴ M and very likely represents the 1-isozyme. The second component represents ~55% of the total activity and exhibits a high affinity for the inhibitor, with a K_i of 1.8 ± 1.0 × 10⁹ M. As previously shown (185), immunoblot analysis of the testis preparation showed that the tissue does not express the 2- or 3-isoforms (Fig. 2B). A complete characterization of the -subunits in the rat testis has not been performed. However, 1 and 3 mRNAs have been detected in the gonad (133). The possibility that 1 3 represents the component of high affinity for ouabain is unlikely, since the other -subunits do not influence the kinetics of the enzyme toward the cardiotonic steroids. Therefore, the highly ouabain-sensitive Na-K-ATPase encountered probably corresponds to 4. It is unknown which -isoform naturally associates with 4. Consequently, the highly ouabain-inhibitable ATPase activity could correspond to 4 1 or 4 3. Undoubtedly, a complete determination of the kinetic characteristics of this isoform will provide insight into its role in Na-K-ATPase function.

View larger version (18K): [in this window] [in a new window]   Fig. 2.   A: dose-response curves for the ouabain inhibition of Na-K-ATPase from rat testis membranes. Na-K-ATPase activity was determined after preincubation of the samples for 30 min at 37°C in the reaction mixture containing 120 mM NaCl, 30 mM KCl, 3 mM MgCl2, 0.2 mM EGTA, 30 mM Tris · HCl (pH 7.4), and the indicated ouabain concentrations. The reaction was started by the addition of [-32P]ATP. Values are expressed as percentage of maximal activity in the absence of the inhibitor. Curves represent the best fit of the data assuming the existence of two enzyme populations with different ouabain affinities. Each value is the mean, and error bars represent the standard errors of the mean of 4 experiments performed in triplicate. B: immunoblot analysis of the Na-K-ATPase -subunit expression in rat testis. Testis proteins (30 µg) were separated by SDS-PAGE (7.5% gel) and transferred to nitrocellulose. Na-K-ATPase -polypeptides were detected using isoform-specific antibodies as before (32). A membrane preparation from rat brain (20 µg) is shown as control.

As shown in Table 1, the -isoform composition does not influence the kinetic properties of the Na-K-ATPase isozymes as much as the -polypeptides do. However, the -subunit is able to influence the Na⁺ requirement of the enzyme. Thus, for the 2- and 3-isoforms, the apparent affinity for Na⁺ increases when associated with 2. A modulatory effect of the -isoform on the catalytic properties of the -subunit has also been shown using Xenopus oocytes. Hybrid 1 1- and 1 2-isozymes consisting of a Torpedo - and a rat -subunit, have different affinities for Rb⁺ (102, 103, 177). Moreover, comparison of results from transfected HeLa cells (57, 105) with native enzyme from pineal gland (189) suggests an effect of the -isoform on the Na⁺ dependence of the rat 3-isoform. At present, the role the -subunit plays in the reaction cycle of the Na-K-ATPase is uncertain. Characterization of the kinetics of hybrid Na-K-ATPases composed of the 1- or 3-isoforms and chimeric -subunits containing the NH₂ terminus of 1 and the COOH terminus of the H-K -subunit suggests that the -subunit influences the interaction of the enzyme with Na⁺ and reduces the ability of Na⁺ in forming the phosphoenzyme complex from ATP (64). It also appears that the -subunit can influence the K⁺ affinity of the Na-K-ATPase (62). With use of the Xenopus oocytes as an expression system, the 1 2 enzyme exhibits a lower affinity for K⁺ than 1 1 (102, 103, 177). Similarly, this trend was maintained in Sf-9 cells where the 2- and 3-isoforms in combination with 2 exhibit a slightly lower affinity for K⁺ than the 1 counterparts (33, 35). In addition, the -subunit has been shown to be actively involved in forming or stabilizing the K⁺-occluding complex of the enzyme (126). Because in insect cells the -isoforms are mainly expressed in their core-glycosylated forms (36, 59), the difference in Na⁺ affinity between isozymes containing 1 or 2 most likely depends on the dissimilar amino acid structure of the -isoforms rather than the sugar composition. Additional evidence of the importance of the -subunit in modulating the activity of the Na-K-ATPase has been suggested from the kinetic analysis of the human 1 3 Na-K-ATPase expressed in insect cells. The 3-isoform slightly influences the Na⁺, K⁺, and ouabain affinities of the enzyme (214).

A physiological role for the difference between Na-K-ATPases composed of 1 and 2 is difficult to infer. Although 2 confers 2 and 3 with a similar apparent affinity for Na⁺ as the 1 1 enzyme, 2 2 and 3 2 still retain the particular K⁺, ATP, and ouabain reactivities of their counterpart 1-containing isozymes. The importance of the 2-polypeptide in Na-K-ATPase function may be related to other properties intrinsic to this subunit. For example, an unexpected functional property of the 2-isoform is that, in addition to its role as a constituent of the Na pump, it serves as a cell adhesion molecule that mediates Ca²⁺ independent neuron-astrocyte interactions (82, 177). In addition, 2 has been shown to be involved in neuronal migration and neurite outgrowth (5, 82, 153). Coinciding with its role in cell-cell interaction, the decrease in 2 mRNA levels in human kidney, lung, and liver carcinomas suggests that this subunit may be involved in tumorigenesis (3). Although the link between cell recognition and ion transport is unknown, it is conceivable that 2 might modulate Na pump function to provide the ionic environment required in such specific cellular processes. The importance of the 2 has recently been demonstrated in mice carrying a targeted deletion of the gene for this isoform (132). The 2-deficient animals exhibit motor incoordination at 15 days of age, develop tremors and paralysis of extremities, and subsequently die at 17-18 days after birth. The observation that at different areas of the nervous system there is swelling and vacuolization of astrocyte processes suggested the requirement of 2 in maintaining ionic homeostasis. However, the cause of death in these rodents is unclear, since 2 is not necessary for the long-term survival of cultured telencephalic grafts obtained from the transgenic mice (101).

Physiological Relevance of Na Pump Isozymes

A differential function for the Na-K-ATPase isozymes is suggested by the stringent regulation of the expression of the - and -polypeptides under various physiological conditions. During development, a change in the relative amount of Na pump isoforms occurs in several tissues (157). A remarkable example of regulation of Na-K-ATPase isoform expression is found in the developing rat heart. In this tissue, between the second and third week of life, there is a switch from the 3-isoform to 2-isoform, which becomes the predominant isoform of the adult myocardium (124). Interestingly, this event coincides with important changes in the electrophysiological properties of the myocardium (121, 124). The Na pump isozymes may also be differentially regulated by hormones (reviewed in Ref. 68). Hormones can elicit their action by modulating the expression of a particular isoform or by directly affecting the activity of an individual Na-K-ATPase (68). For example, in rat skeletal muscle, insulin is able to produce a rapid translocation of preexisting 2 Na-K-ATPase from intracellular stores to the plasma membrane (99). This results in the recruitment of additional functional Na pumps to the cell surface and increased Na-K-ATPase activity. In addition, the normal expression of Na-K-ATPase isoforms can be altered by pathological conditions. For instance, in several cardiac diseases, the Na-K-ATPase isoform composition of the heart is modified. These changes may reflect the cellular attempt to reestablish the altered homeostasis associated with the disease. Consequently, it seems conceivable that the preferential regulation or modification of a specific isoform may be important in finely adjusting cellular ionic homeostasis.

The difference among rodent Na-K-ATPase isozymes in ouabain affinity is intriguing, although its physiological relevance remains unknown. It is possible that differences in ouabain binding may have a role in the regulation of Na-K-ATPase activity. Only rodents are known to have Na-K-ATPases that vastly differ in their ouabain affinities. In other species, the 1 1-isozyme is much more sensitive to the cardiotonic steroids, making the isoform differences in ouabain sensitivity less conspicuous. A detailed study of the ouabain binding constants for species other than rodents has not been accomplished. Nevertheless, more than one ouabain affinity site has been reported in non-rodent species (66, 67, 88, 90, 104, 131). In humans, both single as well as complex ouabain binding have been found in the heart (66, 67, 184), which is known to express the 1-, 2-, and 3-isoforms of the Na pump (216). In the heart, the existence of multiple isoforms is undoubtedly important in mediating the action of the cardiotonic steroids. It is clear that cardiotonic steroids act through their ability to inhibit the Na-K-ATPase. This inhibition results in a transient increase in intracellular Na⁺, which induces through the Na/Ca exchange system an increase in intracellular Ca²⁺. The rise in internal Ca²⁺ is taken up by the SR via an SR Ca-ATPase. Because of the elevated Ca²⁺ content in the SR, more Ca²⁺ can be released when the cardiocyte is stimulated, leading to an augmented force of contraction and an increased cardiac output (37, 48, 118, 141, 145, 205). The relevance of the cardiotonic glycosides as Na-K-ATPase modulators is supported by the finding of endogenous, digitalis-like compounds in mammals (reviewed in Refs. 37, 38, 60, 61, 111, 168). An isomer of ouabain has been found to be produced in the bovine hypothalamus (93, 94, 220), and ouabain itself is present in the adrenal glands of several species including humans (37, 38, 86, 87). Thus endogenous ouabain may regulate cardiovascular tonicity by inhibiting a discrete number of Na-K-ATPase pumps to modulate cellular excitability and heart and vascular muscle contractility.

Studying the effect of Ca²⁺ on the activity of the different rat Na-K-ATPase isozymes expressed in Sf-9 cells, we found that the cation differentially affects the function of each isoform (34). As has been shown before, the 1 1-isozyme is quite resistant to Ca²⁺ (15). In contrast, the 2 1- and 3 1-isozymes display K_i values 10- and 100-fold lower than that of the 1 1, respectively (Table 1). Similarly, in rat axolemma, myometrium, and skeletal muscle, the Na-K-ATPase activity that has a high sensitivity to ouabain is inhibited by physiological concentrations of Ca²⁺ (142, 207). This suggests that in excitable cells after depolarization, when the intracellular Ca²⁺ concentration rises to 5-10 µM, the 1 1-isozyme remains active, while the 2 1- and 3 1-isozymes are functioning at approximately one-half of their maximal capability. This may be particularly relevant in the heart, where the rise in intracellular Ca²⁺ elicited by the cardiotonic steroids may be enhanced by further inhibition of the ouabain-sensitive 2- and 3-isoforms in a positive feedback mechanism. The differential affinity to Ca²⁺ also ensures that only a fraction of the total enzyme is inhibited, the same fraction that is sensitive to the cardiotonic steroids. Interestingly, in contrast to the 1-isoform, 2-isoform in astrocytes and 3-isoform in neurons and myocytes exhibit a reticular distribution on the plasma membrane which parallels the underlying endoplasmic reticulum or SR (110). This distribution is identical to that of the Na/Ca exchanger. Thus, while 1 is regulating bulk cytosolic Na⁺, 2 and 3 may adjust Na⁺ concentrations in the restricted cytosolic space between the plasma membrane and reticulum to indirectly control Ca²⁺ (110).

Regulation of Na-K-ATPase Isozymes

Recent evidence suggests that intracellular messengers can also affect Na-K-ATPase activity. Depending on the tissue, activation of protein kinases can induce an increase or decrease in Na pump activity (26, 54, 68, 138, 149, 172). Agents that increase cellular cAMP, as well as exogenous derivatives of cAMP, lead to Na-K-ATPase inhibition in the medullary thick ascending limb of the loop of Henle and the cortical collecting duct (18, 26, 96, 149, 172, 174) and to Na-K-ATPase stimulation in the proximal convoluted tubules (47, 74). This response seems to be mediated, at least in part, by protein kinase A (PKA)-directed phosphorylation of Ser⁹⁴³ of the Na pump -subunit (77). Phosphorylation of the Na pump -subunit is reversible, as demonstrated by the decrease in Na-K-ATPase activity after activation of a dopamine- and cAMP- regulated phosphoprotein (DARPP-32), an endogenous inhibitor of protein phosphatase 1 (PP1) (6). This suggests that a phosphorylation/dephosphorylation event may dynamically regulate the activity of the Na pump. Moreover, in the proximal nephron and in cultured canine kidney cells, the Na-K-ATPase is inhibited by phorbol esters or diacylglycerol analogs, in a process that involves protein kinase C (PKC) activation (79, 172, 183) and, possibly, the phosphorylation of the Na pump -subunit at Ser¹⁶ (17). In addition, several other mechanisms have been postulated concerning the way by which protein kinases exert their effects on the Na pump. For example, in amphibian nephron cells, PKC produces Na-K-ATPase inhibition by increasing cell endocytosis and internalization of Na pump molecules (21), and in isolated nephron segments, PKA may stimulate phospholipase A₂ (PLA₂) and the production of arachidonic acid and its metabolites (173). Moreover, in the kidney another kinase, protein kinase G (PKG), and the production of cGMP appear to mediate the loss of salt and water produced by acetylcholine, bradykinin, and atrial natriuretic peptide (144, 175). In contrast, in the duck salt gland cGMP stimulates ouabain-sensitive p-nitrophenylphosphatase hydrolysis by the Na pump (192).

Although the effect of intracellular messengers on the function of the 1 1-isozyme has been well documented (26, 68, 138), the regulation of the other Na-K-ATPase isozymes by protein kinases is poorly understood. We studied the effect of different protein kinases on the rat Na-K-ATPase isoforms expressed in Sf-9 cells. Similar to mammalian cells, insect cells glycosylate, phosphorylate, acylate, and perform most other posttranslational modifications dictated by the primary amino acid sequence. In addition, they display a complex signal transduction machinery comparable to that of their mammalian counterparts (156). As shown in Fig. 3, activation of PKA, PKC, and PKG, using dibutyryl-cAMP, phorbol 12-myristate 13 acetate (PMA), or dibutyryl-cGMP, respectively, is able to differentially modify the function of the Na-K-ATPase isozymes. Specificity of the response is demonstrated by the reversibility of the effect by the respective protein kinase competitive inhibitors. As shown, PKC activation leads to inhibition of all isozymes, PKA activation stimulates the activity of the Na-K-ATPase 3 1 and decreases that of the