Na+,K+-ATPase: half-of-the-subunits cross-linking reactivity suggests an oligomeric structure containing a minimum of four catalytic subunits

Sensational site: the sodium pump ouabain-binding site and its ligands

Cardiotonic steroids (CTS), used by certain insects, toads, and rats for protection from predators, became, thanks to Withering's trailblazing 1785 monograph, the mainstay of heart failure (HF) therapy. In the 1950s and 1960s, we learned that the CTS receptor was part of the sodium pump (NKA) and that the Na+/Ca2+ exchanger was critical for the acute cardiotonic effect of digoxin- and ouabain-related CTS. This "settled" view was upended by seven revolutionary observations. First, subnanomolar ouabain sometimes stimulates NKA while higher concentrations are invariably inhibitory. Second, endogenous ouabain (EO) was discovered in the human circulation. Third, in the DIG clinical trial, digoxin only marginally improved outcomes in patients with HF. Fourth, cloning of NKA in 1985 revealed multiple NKA α and β subunit isoforms that, in the rodent, differ in their sensitivities to CTS. Fifth, the NKA is a cation pump and a hormone receptor/signal transducer. EO binding to NKA activates, in a ligand- and cell-specific manner, several protein kinase and Ca2+-dependent signaling cascades that have widespread physiological effects and can contribute to hypertension and HF pathogenesis. Sixth, all CTS are not equivalent, e.g., ouabain induces hypertension in rodents while digoxin is antihypertensinogenic ("biased signaling"). Seventh, most common rodent hypertension models require a highly ouabain-sensitive α2 NKA and the elevated blood pressure is alleviated by EO immunoneutralization. These numerous phenomena are enabled by NKA's intricate structure. We have just begun to understand the endocrine role of the endogenous ligands and the broad impact of the ouabain-binding site on physiology and pathophysiology.

Multimeric species in equilibrium in detergent-solubilized Na,K-ATPase

In this work, we find an equilibrium between different Na,K-ATPase (NKA) oligomeric species solubilized in a non-ionic detergent C12E8 by means of Dynamic Light Scattering (DLS), Analytical Ultracentrifugation (AUC), Small Angle X-ray Scattering (SAXS), Spectrophotometry (absorption at 280/350nm) and enzymatic activity assay. The NKA sample after chromatography purification presented seven different populations as identified by AUC, with monomers and tetramers amounting to ∼55% of the total protein mass in solution. These two species constituted less than 40% of the total protein mass after increasing the NKA concentration. Removal of higher-order oligomer/aggregate species from the NKA solution using 220nm-pore filter resulted in an increase of the specific enzymatic activity. Nevertheless, the enzyme forms new large aggregates over an elapsed time of 20h. The results thus point out that C12E8-solubilized NKA is in a dynamic equilibrium of monomers, tetramers and high-order oligomers/subunit aggregates. These latter have low or null activity. High amount of detergent leads to the dissociation of NKA into smaller aggregates with no enzymatic activity.

Oligomerization of the Na,K-ATPase in cell membranes

The higher order oligomeric state of the Na,K-ATPase alphabeta heterodimer in cell membranes is the subject of controversy. We have utilized the baculovirus-infected insect cell system to express Na,K-ATPase with alpha-subunits bearing either His(6) or FLAG epitopes at the carboxyl terminus. Each of these constructs produced functional Na,K-ATPase alphabeta heterodimers that were delivered to the plasma membrane (PM). Cells were simultaneously co-infected with viruses encoding alpha-His/beta and alpha-FLAG/beta Na,K-ATPases. Co-immunoprecipitation of the His-tagged alpha-subunit in the endoplasmic reticulum (ER) and PM fractions of co-infected cells by the anti-FLAG antibody demonstrates that protein-protein associations exist between these heterodimers. This suggests the Na,K-ATPase is present in cell membranes in an oligomeric state of at least (alphabeta)(2) composition. Deletion of 256 amino acid residues from the central cytoplasmic loop of the alpha-subunit results in the deletion alpha-4,5-loop-less (alpha-4,5LL), which associates with beta but is confined to the ER. Co-immunoprecipitation demonstrates that when this inactive alpha-4,5LL/beta heterodimer is co-expressed with wild-type alphabeta, oligomers of wild-type alphabeta and alpha-4,5LL/beta form in the ER, but the alpha-4,5LL mutant remains retained in the ER, and the wild-type protein is still delivered to the PM. We conclude that the Na,K-ATPase is present as oligomers of the monomeric alphabeta heterodimer in native cell membranes.

Role of the self-association of beta subunits in the oligomeric structure of Na+/K+-ATPase

The two subunits of Na(+)/K(+)-ATPase that are essential for function are alpha and beta. Previous cross-linking studies on the oligomeric structure of the membrane-bound enzyme identified alpha,beta and alpha,alpha associations, but only the former and not the latter could be detected after solubilization. To study the possibility of direct beta,beta association, the purified membrane enzyme and a trypsin-digested enzyme that occludes cations and contains an essentially intact beta and fragments of alpha were subjected to oxidative cross-linking in the presence of Cu(2+)-phenanthroline. Resolution of products on polyacrylamide gels, N-terminal analysis and reactivity with anti-beta antibody showed that, in addition to previously identified products (e.g. alpha,alpha and alpha,beta dimers), a beta,beta dimer, most likely linked through intramembrane Cys(44) residues of two chains, is also formed. This dimer was also noted when digitonin-solubilized intact enzyme, and the trypsin-digested enzyme solubilized with digitonin or polyoxyethylene 10-laurylether were subjected to cross-linking, indicating that the detected beta,beta association was not due to random collisions. In the digested enzyme, K(+) but not Na(+) enhanced beta,beta dimer formation. The alternative cross-linking of beta-Cys(44) to a Cys residue of a transmembrane alpha-helix was antagonized specifically by K(+) or Na(+). The findings (i) indicate the role of beta,beta association in maintaining the minimum oligomeric structure of (alpha,beta)(2), (ii) provide further support for conformation-dependent flexibilities of the spatial relations of the transmembrane helices of alpha and beta and (iii) suggest the possibility of significant differences between the quaternary structures of the P-type ATPases that do and do not contain a beta subunit.

Thermal denaturation of the Na,K-ATPase provides evidence for alpha-alpha oligomeric interaction and gamma subunit association with the C-terminal domain

Thermal denaturation can help elucidate protein domain substructure. We previously showed that the Na,K-ATPase partially unfolded when heated to 55 degrees C (Arystarkhova, E., Gibbons, D. L., and Sweadner, K. J. (1995) J. Biol. Chem. 270, 8785-8796). The beta subunit unfolded without leaving the membrane, but three transmembrane spans (M8-M10) and the C terminus of the alpha subunit were extruded, while the rest of alpha retained its normal topology with respect to the lipid bilayer. Here we investigated thermal denaturation further, with several salient results. First, trypsin sensitivity at both surfaces of alpha was increased, but not sensitivity to V8 protease, suggesting that the cytoplasmic domains and extruded domain were less tightly packed but still retained secondary structure. Second, thermal denaturation was accompanied by SDS-resistant aggregation of alpha subunits as dimers, trimers, and tetramers without beta or gamma subunits. This implies specific alpha-alpha contact. Third, the gamma subunit, like the C-terminal spans of alpha, was selectively lost from the membrane. This suggests its association with M8-M10 rather than the more firmly anchored transmembrane spans. The picture that emerges is of a Na,K-ATPase complex of alpha, beta, and gamma subunits in which alpha can associate in assemblies as large as tetramers via its cytoplasmic domain, while beta and gamma subunits associate with alpha primarily in its C-terminal portion, which has a unique structure and thermal instability.

Acid-labile ATP and/or ADP/P(i) binding to the tetraprotomeric form of Na/K-ATPase accompanying catalytic phosphorylation-dephosphorylation cycle

The Na/K-ATPase has been shown to bind 1 and 0.5 mol of (32)P/mol of alpha-chain in the presence [gamma-(32)P]ATP and [alpha-(32)P]ATP, respectively, accompanied by a maximum accumulation of 0.5 mol of ADP-sensitive phosphoenzyme (NaE1P) and potassium-sensitive phosphoenzyme (E2P). The former accumulation was followed by the slow constant liberation of P(i), but the latter was accompanied with a rapid approximately 0.25 mol of acid-labile P(i) burst. The rubidium (potassium congener)-occluded enzyme (approximately 1.7 mol of rubidium/mol of alpha-chain) completely lost rubidium on the addition of sodium + magnesium. Further addition of approximately 100 microM [gamma-(32)P]ATP and [alpha-(32)P]ATP, both induced 0.5 mol of (32)P-ATP binding to the enzyme and caused accumulation of approximately 1 mol of rubidium/mol of alpha-chain, accompanied by a rapid approximately 0.5 mol of P(i) burst with no detectable phosphoenzyme under steady state conditions. Electron microscopy of rotary-shadowed soluble and membrane-bound Na/K-ATPases and an antibody-Na/K-ATPase complex, indicated the presence of tetraprotomeric structures (alphabeta)(4). These and other data suggest that Na/K-ATP hydrolysis occurs via four parallel paths, the sequential appearance of (NaE1P:E.ATP)(2), (E2P:E.ATP:E2P:E. ADP/P(i)), and (KE2:E.ADP/P(i))(2), each of which has been previously referred to as NaE1P, E2P, and KE2, respectively.

Phosphorylation/dephosphorylation of reconstituted shark Na+,K(+)-ATPase: one phosphorylation site per alpha beta protomer

In the present investigation reconstitution of Na+,K(+)-ATPase increases the number of phosphorylation sites (EP) of solubilized enzyme from 4.2 +/- 0.3 nmol/mg to 6.9 +/- 0.6 nmol/mg. The latter figure corresponds to one phosphorylation site per alpha beta-promoter. A cholesterol content > 10 mol% in the liposome bilayer and a high extracellular [Na+] are necessary to obtain this high value. Spontaneous dephosphorylation after maximum phosphorylation in Na+ is biphasic both in solubilized enzyme and after reconstitution. The rate of dephosphorylation compares with the specific hydrolytic Na(+)-ATPase activity measured at exactly identical conditions for all three preparations assuming parallel dephosphorylation of at least two phosphointermediates. The distribution of EP-species is found to vary among the three enzyme preparation used, i.e., membrane bound, solubilized, and reconstituted Na+,K(+)-ATPase, however in all the equilibrium is strongly poised away from the E1P-form.

(Na+ + K+)-ATPase: on the number of the ATP sites of the functional unit

Questions concerning the number of the ATP sites of the functional unit of (Na+ + K+)-ATPase (i.e., the sodium pump) have been at the center of the controversies on the mechanisms of the catalytic and transport functions of the enzyme. When the available data pertaining to the number of these sites are examined without any assumptions regarding the reaction mechanism, it is evident that although some relevant observations may be explained either by a single site or by multiple ATP sites, the remaining data dictate the existence of multiple sites on the functional unit. Also, while from much of the data it is clear that the multiple sites of the unit enzyme represent the interacting catalytic sites of an oligomer, it is not possible to rule out the existence of a distinct regulatory site for ATP in addition to the interacting catalytic sites. Regardless of the ultimate fate of the regulatory site, any realistic approach to the resolution of the kinetic mechanism of the sodium pump should include the consideration of the established site-site interactions of the oligomer.

Cyclic AMP-dependent stimulation of Na,K-ATPase in shark rectal gland

Scatchard analysis of 3H ouabain bound to isolated rectal gland cells as a function of increasing ouabain concentrations produced a concave curvilinear plot that was resolved into two specific sites with either a high (I) or low (II) affinity for ouabain. Cyclic cAMP/theophylline (+/- furosemide, 10(-4) M) increased the amount of 3H ouabain bound to the high-affinity site I. Vanadate, a phosphate congener which promotes formation of the ouabain-binding state of the enzyme, mimicked the effects of cAMP/theophylline at low concentrations of ouabain, suggesting that cAMP/theophylline increases binding to site I by enhancing the rate of turnover of resident enzyme. Enhanced 86Rb uptake seen following cAMP/theophylline administration was primarily associated with increased flux through the high-affinity ouabain site, and this stimulation was not obliterated by the co-administration of furosemide. A model was presented which suggested the presence of two noninteracting pools of enzyme or isozymes which exhibit either a high or low affinity for ouabain. Cyclic AMP both stimulated turnover via site I, and modified the kinetics of binding of 3H ouabain to site II. The (ave) Kd of 3H ouabain for site II was increased from 3.6 microM (controls) to 0.5 microM (cAMP/theophylline) and the Hill coefficient was modified from 0.45 (controls) to 1.12 (cAMP/theophylline), suggesting a transition from a negative- to a noncooperative binding state. While furosemide reversed the effects of cAMP/theophylline on site II kinetics, it did not obliterate cAMP/theophylline effects on site I. This suggests that cAMP may alter the intrinsic turnover rate of this particular pool of Na,K-ATPase in shark rectal gland.

Modification of the conformational equilibria in the sodium and potassium dependent adenosinetriphosphatase with glutaraldehyde

Glutaraldehyde treatment of electroplax membrane preparations of Na,K-ATPase leads to irreversible changes in the enzymic behavior of the protein, which are not due to modification of the active site. When the glutaraldehyde treatment is carried out in a medium containing K+ and without Na+, the "K+-modified enzyme" so produced shows the following changes in enzymic properties: The steady-state phosphorylation by ATP and the rate of ATP-ADP exchange are decreased to approximately 40% of control, while Na,K-ATPase activity decreases to approximately 15% of control. Phosphatase activity is decreased very little, but the potassium activation parameters of the reaction are changed, from K0.5 approximately equal to 5 mM and nH = 1.9 in control to K0.5 approximately equal to 0.5 mM and nH = 1 in K+-modified enzyme. KI(app) for nucleotide inhibition of phosphatase activity is increased significantly. Changes in the cation dependence of the ATPase reaction are also observed. All of these effects can be explained by assuming that the cross-linking of surface groups in protein subunits when they are in conformation E2 shifts the intrinsic conformational equilibrium of the enzyme toward E2. We considered the simplest mathematical model for the coupling between K+ binding and the conformational equilibrium, with equivalent potassium sites that must be simultaneously in the same state. If one assumes that the potassium activation of phosphatase activity in the K+-modified enzyme reflects the affinity for K+ of E2, the behavior of the phosphatase activity in the native enzyme can be fit if there are only two potassium sites, whose affinity is 80-fold higher in E2 than in E1, and the equilibrium constant for E2 in equilibrium E1 is about 250. The same sites can explain the activation of dephosphorylation during ATP hydrolysis. Independent of the model chosen, potassium ions must be required for the catalytic action of form E2 and cannot be merely "allosteric activators". The enzyme modified with glutaraldehyde in a medium containing Na+ also has interesting properties, but their rationalization is less straightforward. The Na,K-ATPase activity is inhibited more than the "partial reactions", as in the K+-modified enzyme. We suggest that this is a generally expected result of modifications of the enzyme.

(Na/ + K+)ATPase has one functioning phosphorylation site per alpha subunit

(Na+ + K+)ATPase contains two different subunits, a catalytic subunit (alpha) and a subunit with uncertain function (beta). The enzyme binds ATP, ouabain and vanadate, and can be phosphorylated by ATP as well as by inorganic phosphate. From the previously reported maximal binding and phosphorylation capacities of 3.5--4.3 nmol P per mg protein (based on Lowry protein determination) and the earlier molecular weight value of approximately 250,000, a molar binding and phosphorylation capacity of 0.87--1.07 mol per mol enzyme was derived. As it is generally agreed that the enzyme molecule contains two alpha subunits or even a multiple of two, it has been suggested that the enzyme operates by means of a so-called "half-of-the-sites mechanism" whereby only of the two alpha subunits can be phosphorylated at any one time. We now present evidence that every alpha subunit can be phosphorylated simultaneously, which rules out the operation of such a mechanism.

Contact-site cross-linking agents

Contact-site cross-linking agents comprise a heterogeneous grouping of cross-linkers which share the common property of being able to cross-link only very closely juxtaposed residues in macromolecular complexes. We have defined contact-site cross-linking arbitrarily as the covalent joining of residues such that they are constrained to a distance which is equivalent to or less than their closest possible steric approach prior to becoming linked (1). We recognize two classes of contact-site cross-linkers, bridge type and zero-length type. The former, such as formaldehyde, become incorporated during cross-linking as one-atom bridges. The latter, such as the carbodiimides, operate as condensing agents with the result that the cross-linked residues become interjoined directly. Contact-site cross-linkers have been used in several ways as specific probes of both the static and dynamic aspects of macromolecular structure. They can yield precise structural information about macromolecular contacts when actual sites of cross-linking are determined by peptide or nucleotide mapping techniques. In this way exact contacts between histones in the nucleosome, between protein and RNA in the ribosome, and between RNA polymerase and DNA have been determined. Contact-site cross-linkers have also been used to probe the perturbation of contacts following macromolecular conformational changes. Certain histone-histone 'cross-linkable' sites are rendered unreactive after induction of chromatin conformational changes thus serving to localize sites of perturbation.