Reversible Oxidative Modification
A Key Mechanism of Na+-K+ Pump Regulation
Angiotensin II (Ang II) inhibits the cardiac sarcolemmal Na+-K+ pump via protein kinase (PK)C-dependent activation of NADPH oxidase. We examined whether this is mediated by oxidative modification of the pump subunits. We detected glutathionylation of β1, but not α1, subunits in rabbit ventricular myocytes at baseline. β1 Subunit glutathionylation was increased by peroxynitrite (ONOO−), paraquat, or activation of NADPH oxidase by Ang II. Increased glutathionylation was associated with decreased α1/β1 subunit coimmunoprecipitation. Glutathionylation was reversed after addition of superoxide dismutase. Glutaredoxin 1, which catalyzes deglutathionylation, coimmunoprecipitated with β1 subunit and, when included in patch pipette solutions, abolished paraquat-induced inhibition of myocyte Na+-K+ pump current (Ip). Cysteine (Cys46) of the β1 subunit was the likely candidate for glutathionylation. We expressed Na+-K+ pump α1 subunits with wild-type or Cys46-mutated β1 subunits in Xenopus oocytes. ONOO− induced glutathionylation of β1 subunit and a decrease in Na+-K+ pump turnover number. This was eliminated by mutation of Cys46. ONOO− also induced glutathionylation of the Na+-K+ ATPase β1 subunit from pig kidney. This was associated with a ≈2-fold decrease in the rate-limiting E2→E1 conformational change of the pump, as determined by RH421 fluorescence. We propose that kinase-dependent regulation of the Na+-K+ pump occurs via glutathionylation of its β1 subunit at Cys46. These findings have implications for pathophysiological conditions characterized by neurohormonal dysregulation, myocardial oxidative stress and raised myocyte Na+ levels.
The ATP-dependent Na+-K+ pump transports Na+ out of cells in exchange for extracellular K+, against their respective electrochemical gradients. The Na+ and K+ gradients that it generates maintain the membrane potential essential for cellular electric excitability. However, the gradients also have an important broader role because they serve in coupled co- and countertransport processes for other ions and organic compounds. Regulation of the Na+-K+ pump, usually attributed to phosphorylation, is poorly understood. We have found that the ε isoform of protein kinase C (εPKC) mediates angiotensin II (Ang II)–induced inhibition of the pump in rabbit ventricular myocytes.1 However, inhibition could not be attributed to phosphorylation of the pump molecule or of the closely associated “FXYD protein” phospholemman that has been implicated in PKC-dependent pump regulation in cardiac myocytes.2 Instead, the dependence of Ang II–induced pump inhibition on NADPH oxidase and superoxide (O2·−)1 strongly suggests a role for oxidative signaling.
Oxidative signaling can occur via modifications of susceptible protein sulfhydryl groups (protein–SH). Their reactivities are too low to make them realistic targets for oxidative modification induced by O2·−,3 but O2·− can combine with nitric oxide (NO) to form ONOO− or be dismutated to H2O2, either of which can oxidize protein–SH. Of the possible oxidative modifications, formation of a mixed disulfide with the tripeptide glutathione (GSH) is a good candidate for mediating signaling: glutathionylation is facilitated by an abundance of GSH with a high negative redox potential,4 results in the addition of 305 Da negatively charged adduct with the potential for steric effects akin to those of phosphorylation,3,5 and is readily reversible. Glutathionylation has recently been shown to mediate regulation of important cell proteins including SERCA6 and actin.7
Glutathionylation of cysteines with typical redox potentials is not thermodynamically favored by an intracellular environment that is reducing8 but can be facilitated by subcellular compartmentalization.9 Coimmunoprecipitation of the Na+-K+ pump with NADPH oxidase subunits in cardiac myocytes1 suggests that the pump is in a microdomain that supports oxidative signaling. We examined whether any of the sulfhydryl groups in its α/β heterodimer are “reactive,” ie, susceptible to S-glutathionylation3 and found that a specific cysteine in the β1 subunit was glutathionylated by oxidant stimuli, including Ang II–induced activation of NADPH oxidase. We propose reversible glutathionylation offers an explanation for Na+-K+ pump regulation that does not involve phosphorylation of the pump molecule or associated FXYD proteins. However, because NADPH oxidase activation is phosphorylation-dependent, the scheme retains the firmly established role of protein kinases. Such oxidative regulation of the Na+-K+ pump has important physiological and pathophysiological implications.
Materials and Methods
We measured Na+-K+ pump current (Ip) as a ouabain-induced shift in holding current of voltage-clamped isolated rabbit cardiac myocytes.1 To detect S-glutathionylation of pump subunits, myocytes were loaded with biotinylated GSH. After lysis the biotin-tagged glutathionylated subfraction was precipitated using streptavidin-sepharose beads6 and immunoblotted for α1 and β1 Na+-K+ pump subunits. Separate experiments immunoblotted β1 subunit immunoprecipitate with an antibody against glutathionylated protein (anti-GSH antibody).
Xenopus oocytes were injected with Xenopus α1 and wild-type β1 or β1C46W mutant10 cRNAs or human α1 and β1 or β2 or β3 cRNAs. They were loaded with biotinylated GSH and glutathionylated protein precipitated with streptavidin-sepharose beads and immunoblotted for β1 subunit. Measurements of Na+-K+ pump current were performed by the 2-electrode voltage clamp technique.
Pig Kidney Na+-K+ ATPase
The hydrolytic activity of pig kidney Na+-K+ ATPase was measured under varying conditions of Na+, K+, and ATP, and enzyme conformational changes were detected with RH421 fluorescence.
An expanded Materials and Methods section is available in the Online Data Supplement at http://circres.ahajournals.org.
Glutathionylation of the Na+-K+ Pump
To determine whether the α1 and β1 subunits of the Na+-K+ pump have reactive cysteine residues susceptible to glutathionylation, myocytes were loaded with biotin-GSH. They were lysed and biotin-tagged glutathionylated proteins were precipitated using streptavidin beads. We performed immunoblotting on this glutathionylated subfraction for the most abundant isoforms of the α and β subunits of the pump, α1 and β1. Both α1 and β1 were easily detectable in total cell lysate and there was a strong signal for β1 in the biotin-tagged glutathionylated subfraction (Figure 1A). However, there was no evidence of glutathionylated α1 subunits. The biotin-tagged glutathionylated β1 subunit was not detected when the lysate was incubated with 1 mmol/L dithiothreitol (DTT) before precipitation by streptavidin. This sensitivity to 1 mmol/L DTT is supportive of a mixed disulfide bond between the β1 subunit and GSH.
We next examined the effect of a chemical oxidant on β1 subunit glutathionylation. Myocytes loaded with biotin-GSH were exposed to solutions containing 100 μmol/L ONOO− for 10 minutes. ONOO− increased β1 subunit glutathionylation detected by immunoblotting the glutathionyated protein subfraction (GSS-protein) precipitated by streptavidin as shown in Figure 1B. The α1 subunit was not detected in the GSS-protein subfraction, even after ONOO− exposure (data not shown). To obtain independent support for glutathionylation of β1 subunit, we immunoprecipitated β1 subunit, and immunoblotted the precipitate with antibody to GSS-protein. As seen with the biotin-GSH technique, a moderate degree of β1 glutathionylation was present at baseline, and exposure of the myocytes to ONOO− increased glutathionylation as shown in Figure 1C. The reverse immunoprecipitation experiment was performed, as shown in Figure 1D. Cell lysate was precipitated with the GSH antibody and immunoblotted for either α1 or β1 subunit. ONOO− increased glutathionylation of the β1 subunit, whereas α1 was not detected in the protein subfraction immunoprecipitated by the GSH antibody.
The effect of additional oxidant stresses on β1 subunit glutathionylation was investigated. Myocytes loaded with biotin-GSH were exposed to H2O2 for 15 minutes. This increased β1 subunit glutathionylation as shown in Figure 2A. Exposure of myocytes to 100 μmol/L paraquat, which increases O2·−-sensitive fluorescence11 also increased β1 subunit glutathionylation (Figure 2B). Because membrane-permeable, pegylated superoxide dismutase (SOD) abolishes the paraquat-induced increase in O2·−-sensitive fluorescence,11 we examined the effect of SOD on glutathionylation of the β1 subunit. Incubation of myocytes with 200 IU/mL pegylated SOD starting 5 minutes before exposure to paraquat abolished the increase in β1 subunit glutathionylation (Figure 2B). Paraquat-induced glutathionylation may partially be attributable to a direct effect of paraquat as a “redox-cycler,” producing O2·− anions. Alternatively, paraquat uncouples NO synthase (NOS), resulting in O2·− synthesis.12 We examined whether there is a physical association of endothelial (e)NOS and the pump. Cell lysate was immunoprecipitated with monoclonal antibodies to the α1 and β1 subunit of the pump and then immunoblotted for eNOS. Figure 2C shows that eNOS coimmunoprecipitated with the α1 and β1 subunits.
To examine whether pathophysiological conditions known to be associated with increased oxidative stress might induce β1 Na+-K+ pump subunit glutathionylation, we examined myocardium from a sheep model of infarction. As shown in Figure 2D, β1 subunit from myocardium immediately adjacent to an infarct zone showed increased glutathionylation compared with that in normal myocardium.
Receptor-Coupled Oxidant Signaling and Glutathionylation of the β1 Subunit
Because Ang II inhibits the cardiac Na+-K+ pump via PKC-dependent activation of NADPH oxidase, and NADPH oxidase subunits are physically associated with the pump,1 we examined the effect of Ang II on glutathionylation of the β1 pump subunit. We exposed myocytes to control solution or solution containing 100 nmol/L Ang II for 10 minutes before lysis. Figure 3A shows that Ang II increased glutathionylation of the β1 subunit as detected by the biotin-GSH technique. Results shown in Figure 3B confirm this with the independent GSH antibody technique.
Ang II–induced pump inhibition is abolished by SOD.1 Figure 3C shows that Ang II–induced β1 subunit glutathionylation is also abolished by preincubation of myocytes with 200 IU/mL SOD for 15 minutes. The Ang II–induced β1 subunit glutathionylation persisted in myocytes preincubated with 100 IU/mL catalase to eliminate H2O2. However, it was abolished by preexposing myocytes to 100 μmol/L ebselen to scavenge ONOO− or 1 mmol/L NG-nitro-l-arginine methyl ester (L-NAME) to inhibit NOS. L-NAME also reduced baseline β1 subunit glutathionylation.
Reversibility of Na+-K+ Pump β1 Subunit Glutathionylation
Figure 4A shows the increase in β1 subunit glutathionylation with either 10 minutes or 15 minutes of exposure of myocytes to paraquat. We examined whether the increase in glutathionylation is reversible by adding pegylated SOD to a final concentration of 200 IU/mL after 10 minutes of exposure to paraquat. The myocytes were then lysed after an additional 5 minutes. Scavenging of O2·− with SOD significantly reduced glutathionylation of β1 subunit compared to the glutathionylation in myocytes exposed to paraquat alone for either 10 or 15 minutes (Figure 4A).
Deglutathionylation is catalyzed by glutaredoxin 1 (GRx1). We examined whether this oxidoreductase is physically associated with the β1 subunit of the Na+-K+ pump. Cell lysate was immunoprecipitated with antibody to the β1 subunit and immunoblotted with monoclonal antibody to GRx1. Figure 4B shows that GRx1 coimmunoprecipitates with the β1 pump subunit. We used the whole-cell patch-clamp technique to examine whether GRx1 has functional effects on Na+-K+ pump current, Ip. The patch pipette solution included 100 μmol/L paraquat or was paraquat-free. Paraquat induced a large decrease in Ip. We included 10 μmol/L recombinant GRx1 in the pipette solution, and, to facilitate the conversion of GRx1 from its oxidized to its reduced state, 1 mmol/L GSH. GSH alone had no significant effect. However, GRx1 abolished the paraquat-induced Na+-K+ pump inhibition. The results of the patch clamp experiments are summarized in Figure 4C. Representative traces are shown in Online Figure III.
Oxidative Modification of the β1 Subunit and α/β Subunit Coimmunoprecipitation
Because a heterodimer of the catalytic α subunit and the regulatory β subunit is essential for Na+-K+ pump function, oxidation might inhibit pump activity by altering subunit interaction. Myocytes were exposed to control solutions or solutions that included 200 or 500 μmol/L ONOO− for 10 minutes. ONOO− had no effect on detection of α1 or β1 subunits in cell lysate by standard Western blot techniques (Online Figure I). Myocyte lysate was immunoprecipitated with α1 subunit antibody and immunoblotted for α1 or β1 subunit. The total α1 subunit immunoprecipitated by α1 antibody was similar in cells exposed to ONOO− or control solutions. However, ONOO− reduced the β1 subunit coimmunoprecipitated with α1 subunit as shown in Figure 5A. Reverse coimmunoprecipitation experiments were also performed. Myocyte lysate was immunoprecipitated using the β1 antibody and immunoblotted for α1 subunit. As shown in Figure 5B, ONOO− induced a reduction in α1 subunit detected in the β1 immunoprecipitate. A receptor-mediated oxidant signal induced by exposing myocytes to Ang II also markedly decreased α/β subunit coimmunoprecipitation, both when myocyte lysate was immunoprecipitated with α1 and immunoblotted with β1 (Figure 5C), and vice versa (Figure 5D).
Candidate β1 Subunit Cysteine
The β1 subunit has 7 cysteine residues, but 6 are linked in 3 disulfide bridges in the ectodomain leaving one, Cys4610 (numbering of amino acids corresponds to Xenopus β1 subunits), free as the likely candidate for oxidant modification. To investigate whether Cys46 is reactive, and whether it mediates oxidative regulation of the Na+-K+ pump, we expressed Xenopus Na+-K+ pump α1 subunits with wild-type or Cys46Trp mutant Xenopus β1 subunits in Xenopus oocytes. We have previously shown that this C46W substitution has no significant effect on the structural and functional maturation of the Na+-K+ pump.10 Figure 6A shows that most β1 subunits identified by Western blot analysis were core-glycosylated after 2 days of expression (lanes 7 to 10), reflecting the continuous synthesis from injected cRNA. However, a population of fully glycosylated subunits also appeared, (lanes 7 to 10). No β1 subunit could be detected in noninjected oocytes (lanes 11 and 12), in agreement with the low expression of endogenous, oocyte subunits.13 Oocytes were injected with biotin-GSH, and we examined for glutathionylation using streptavidin pull-down and immunoblotting. ONOO− induced glutathionylation of the core-glycosylated wild-type β1 subunit (lane 2). In control experiments, injection of inactivated ONOO− did not induce glutathionylation (data not shown). In contrast to wild-type β1 subunit, C46W mutant β1 subunit was not glutathionylated (lane 4).
To examine whether glutathionylation of Cys46 had an effect on Na+-K+ pump function, we measured Na+-K+ pump current in Na+-loaded oocytes (Imax) expressing wild-type or C46W β1 subunits. Oocytes showed a 5 to 6 fold increase in Imax compared with Imax measured in oocytes not injected with cRNA (Figure 6B). Injection with ONOO− significantly decreased Imax in oocytes expressing wild-type, but not mutant C46W β1 subunits (Figure 6B). This decrease was observed without injection of GSH as expected from its endogenous abundance with an estimated level of ≈2.5 mmol/L.14 There was no consistent effect detected of ONOO− on Imax of Na+-K+ pumps in oocytes expressing only native pumps, reflecting their low endogenous expression.
To test whether the ONOO−-induced decrease in Imax of oocytes expressing wild-type β1 subunits is attributable to a reduction in Na+-K+ pump expression at the cell surface or a decrease in turnover number, we performed ouabain binding studies on intact oocytes. Ouabain binding was similar with or without ONOO− treatment of oocytes expressing wild-type β1 subunits or β1 C46W mutants (Figure 6C). Na+-K+ pump turnover number, derived from Imax and the number of ouabain binding sites, is shown in Figure 6D. ONOO− decreased the turnover number of Na+-K+ pumps containing wild-type β1 subunits but not β1 C46W mutant.
A 99% cross-species homology of the transmembrane domain of the β1 subunit15 suggests the transmembrane domain of the β1 subunit is important for Na+-K+ pump function. We examined whether oxidant pump regulation also occurs for the human Na+-K+ pump. We expressed human wild-type α1 and β1 pump subunits in Xenopus oocytes and measured Na+-K+ pump current. Injection with ONOO− significantly decreased Imax as shown in Figure 7. The limited homology between the transmembrane domains of the β1 subunit and the β2 and β3 subunits of the Na+-K+ pump (57% to 61%)15 includes the key difference that β2 and β3 subunits have no free cysteine residues. Figure 7 shows that ONOO− has no effect on Imax in oocytes expressing β2 or β3 subunits, as was the case for the cysteine-free β1 Xenopus mutant.
Oxidative Modification of the β1 Subunit and Na+-K+ ATPase Kinetics
The effect of oxidative modification on Na+-K+ pump kinetics was examined in Na+-K+ ATPase-enriched membranes from pig kidney. β1 Subunit glutathionylation was detected at baseline using the GSH antibody technique without supplemental GSH. This was substantially increased by ONOO−, implying the presence of GSH in the preparation (Online Figure II, A). A GSH assay indicated that the concentration in the preparation was approximately 75 μmol/L. ONOO−- induced glutathionylation was associated with decreased α1/β1 subunit coimmunoprecipitation (Online Figure II, B).
We investigated effects of ONOO−/GSH on Na+-K+ ATPase steady state kinetics. Experimental conditions, Hill equations and the fitting parameters are shown in the Online Data Supplement. Activation of hydrolytic activity of control enzyme was compared with activation of enzyme incubated with 0.5 mmol/L ONOO− for 5 minutes. The enzyme was then incubated with 0.5 mmol/L GSH for 15 minutes to ensure GSH was not rate-limiting.
The effect of ONOO−/GSH on activation of Na-K ATPase by Na+ is shown in Figure 8A. Exposure to ONOO−/GSH inhibited maximal enzyme activity (Vmax), whereas the apparent Na+ affinity (K′Na) was unchanged (Online Table I). GSH itself did not affect either Vmax or K′Na (data not shown). The activation of hydrolytic activity by K+ is shown in Figure 8B. Again, ONOO−/GSH significantly inhibited maximal enzyme activity by ≈25%, whereas the apparent K+ affinity (K′K) was unchanged (Online Table II). Activation by ATP is shown in Figure 8C. Exposure to ONOO−/GSH significantly decreased both Vmax and the apparent ATP affinity constant (K′ATP) (Online Table III), whereas GSH alone affected neither Vmax nor K′ATP (data not shown). Exposure of the enzyme to 2 mmol/L DTT significantly reversed the inactivation induced by ONOO−/GSH from ≈30% to 10% (data not shown). This is very unlikely to be attributable to an effect on protein disulfide bonds, for example in the ectodomain of the β1 subunit, because much higher concentrations of DTT are required to disrupt these.16
We measured the vanadate sensitivity to test whether the decrease in K′ATP was secondary to a ONOO−/GSH-induced shift in the E1/E2 conformational equilibrium toward E1. Figure 8D shows that ONOO−/GSH right-shifted the vanadate inhibition curve. There was a significant increase in the inhibition constant, K′I, from 0.62±0.03 to 4.22±0.01 μmol/L (Online Table IV). This indicates stabilization of the E1 conformation with the lower vanadate affinity.
We followed the ATP phosphorylation reaction by stopped-flow fluorescence of the membrane probe RH42117 starting either from an E1 or an E2 conformation of the enzyme. Because the E2→E1 transition is the main rate-determining step of the overall reaction cycle,17 it rate limits the phosphorylation reaction that follows. Thus, the rate of the E2→E1 reaction can be measured as the rate of the phosphorylation reaction when the enzyme is initially stabilized in the E2 conformation. Two independent sets of experiments with 5 stopped-flow measurements each were performed under the different conditions. Representative traces are shown in Online Figure IV. Exposure of the enzyme to ONOO−/GSH significantly decreased the rate constant of the E1→E2P reaction from 45.2±3.0 to 26.9±2.8 sec−1 and for the E2→E1→E2P reaction from 31.1±2.8 to 16.8±2.5 sec−1.
In some experiments, the enzyme was exposed to ONOO− without supplemental GSH. Consistent with the presence of GSH in the enzyme preparation and the β1 subunit glutathionylation induced by ONOO− without supplemental GSH, supplemental GSH had no significant effect on Na+-K+ pump kinetics (Online Tables I through IV).
We have shown that oxidation induces glutathionylation of the β1 subunit of the Na+-K+ pump in rabbit cardiac myocytes, Xenopus oocytes, and pig kidney. In all models, glutathionylation was associated with inhibition of Na+-K+ pump activity, specifically a decrease in IP in cardiac myocytes, a decrease in Imax and turnover number in Xenopus oocytes, and a decrease in the steady-state hydrolytic activity and rate constant for the E2→E1 conformational change of the pump in pig kidney membrane. Mutational studies indicated that this oxidative regulation was mediated by glutathionylation of Cys46 of the β1 subunit of the pump.
Glutathionylation of the β1 subunit of the Na+-K+ pump is reversible in intact cardiac myocytes (Figure 4A). The rate of spontaneous breakage of disulfide bonds in the process of deglutathionylation is very low and would not allow efficient signaling.9 However, the process is catalyzed by a family of intracellular oxidoreductases. Of these, GRx1 has exclusive selectivity for GSS-protein versus other protein-mixed disulfides.8 We show that GRx1 coimmunoprecipitates with the β1 subunit (Figure 4B) and, when added to patch pipette solutions perfusing cardiac myocytes, abolishes paraquat-induced pump inhibition. These results showing enzymatically mediated reversibility support a functional role for glutathionylation of the β1 subunit in regulation of Na+-K+ pump activity.
The crystal structure of the Na+-K+ pump indicates that the reactive Cys46 of the β1 subunit is located in the bulk lipid of the membrane18 and therefore is expected to be poorly accessible to the cytosolic, hydrophilic GSH. However, the crystal structure was determined with the pump fixed in a state analogous to the E2·2K+·Pi configuration, and large-scale structural changes that include the β1 subunit may occur during the E2→E1 conformational change.18 Such structural change may make Cys46 accessible to GSH. In support of this, Cys46 is susceptible to oxidation induced by the water-soluble Cu-phenanthroline.19 The accessibility of Cys46 to the aqueous cytosol, at least after it has been glutathionylated, is also supported by the role of GRx1 because GRx1 is restricted to the cytosol.8
As expected from the lack of transmembrane cysteine residue in the β2 or β3 isoforms, heterodimer pumps, which include these subunits, were not regulated by an oxidant signal. They may serve in a “housekeeping” role, or they may be regulated by an alternative mechanism. These differences suggest that the expression of specific β subunit isoforms in particular cell types or subcellular domains contribute to differential regulation of the Na+-K+ pump.
Redox signaling may be mediated by direct modification of cysteine residues at the active sites of target proteins. Alternatively, oxidative modification may alter the tertiary structure of the protein and its interaction with other proteins.3 Regulation of the Na+-K+ pump by redox signaling may arise from a change in the interaction of the β1 subunit with the catalytic α1subunit. This is supported by the decrease in coimmunoprecipitation of α1/β1 subunits associated with glutathionylation. This decreased coimmunoprecipitation likely reflects a change in the physical interaction of the subunits revealed with exposure to detergents,19 although not necessarily that glutathionylation completely disrupts the heterodimer in vivo. Nevertheless, a very distinct effect of glutathionylation on α1/β1 interaction is suggested by the complete absence of α1 subunit in the glutathionylated protein subfraction precipitated with streptavidin. If the subunits remained closely associated, the α1 subunit would be expected to be indirectly precipitated by streptavidin via its association with the glutathionylated β1 subunit.
The critical role of the β subunit for Na+-K+ ATPase function and ion-binding was recently demonstrated in a high-resolution structure of the shark Na+-K+ ATPase.18 It may be of particular importance that Cys46 is only separated from one of the hydrogen bonds linking αM7 with the β1 subunit by 1 amino acid. The 2 units of the α1/β1 heterodimer move relative to each other during the transition from the E2 to E1 conformational state described by the Post–Albers scheme, and charged residues on the β1 subunit may affect this movement and hence functional properties of the holoenzyme.20
We used the pig kidney Na+-K+ ATPase preparation to examine effects of glutathionylation on the kinetic properties of the pump because purity, specific activity, and properties of this widely used preparation are well documented and widely accepted. Exposure of the preparation to ONOO− inhibited enzymatic activity, and RH421 fluorescence studies indicated that ONOO− decreased the rate constant for the E2→E1 conformational change. The shift in the conformational poise toward E1 may explain the ONOO− induced increase in apparent affinity to ATP. The E2→E1 reaction is a good candidate for physiologically significant regulation because, together with regulation by changes in the intracellular K+ and Na+ concentrations, it is the main rate-limiting step in the overall pump cycle.21
Kinetic studies on isolated Na+-K+ ATPase do not necessarily directly predict functional properties of the enzyme in the complex environment of the intact cell. This reservation is particularly important for interpretation of the data from patch clamped myocytes in this study. We included ATP in patch pipette solutions perfusing the intracellular compartment, but the diffusion coefficient for ATP is many orders of magnitude lower in the subsarcolemmal space than in the compartment perfused by the pipette solutions. Because ATP is continuously consumed by membrane ATPases, large concentration gradients within myocytes and a low concentration at the membrane can occur.22 This has important effects on the E2→E1 conformational change we identified as susceptible to oxidative signaling because the E2→E1 conformational change becomes sensitive to K+ at low ATP levels.23 The overall forward reaction is therefore inhibited by K+. The degree of inhibition is tissue-dependent, particularly evident in cardiac Na+-K+ ATPase,24 and a high intracellular K+ near physiological levels may enhance effects of oxidative signaling on the Na+-K+ pump in cardiac myocytes.1
It should be noted that there was baseline glutathionylation of the β1 subunit of the pump in both heart and kidney as also reported for some other proteins,4 allowing for the possibility that deglutathionylation from baseline may cause Na+-K+ pump stimulation. Inhibition could be induced by Ang II receptor–coupled activation of oxidative signaling emphasizing its physiological relevance. Oxidative regulation of pump function may also be of pathophysiological importance, in particular in heart failure. An increase in cardiac myocyte oxidant stress25 and raised levels of intracellular Na+26 are believed to contribute to its clinical manifestations. Oxidative modification of the key export route for Na+, the Na+-K+ pump, suggests that oxidant stress and raised myocyte Na+ are interrelated. Intuitively, one might therefore expect that treatment with antioxidants would be beneficial. However, clinical trials do not support such treatment. Mudd and Kass25 proposed that the effects of reactive oxygen species are dependent on the site of their synthesis in the myocyte and that a targeted inhibition is likely to be more successful than broad-acting antioxidants. Our findings that Ang II mediates Na+-K+ pump inhibition via activation of colocalized NADPH oxidase1 are in good agreement with this proposal and may partially explain the well-established clinical efficacy of angiotensin-converting enzyme inhibitors.
We thank Prof Stephen Hunyor and colleagues for kindly providing sheep myocardial samples.
Sources of Funding
Supported by grants from the North Shore Heart Research Foundation and the National Health & Medical Research Council (Australia). G.A.F. was supported by a Royal Australasian College of Physicians/High Blood Pressure Research Foundation Fellowship and the Medical Foundation, University of Sydney. K.G. was supported by Swiss National Science Foundation grant 3100A0-107513; and F.C. was supported by The Danish Medical Research Council.
↵*Both authors contributed equally to this work.
Original received October 31, 2008; resubmission received April 21, 2009; revised resubmission received May 21, 2009; accepted June 10, 2009.
White CN, Figtree GA, Liu CC, Garcia A, Hamilton EJ, Chia KK, Rasmussen HH. Angiotensin II inhibits the Na+-K+ pump via PKC dependent activation of NADPH oxidase. Am J Physiol Cell Physiol. 2009; 296: C693–C700.
Han F, Bossuyt J, Despa S, Tucker AL, Bers DM. Phospholemman phosphorylation mediates the protein kinase C-dependent effects on Na+/K+ pump function in cardiac myocytes. Circ Res. 2006; 99: 1376–1383.
Martinez-Ruiz A, Lamas S. Signalling by NO-induced protein S-nitrosylation and S-glutathionylation: convergences and divergences. Cardiovasc Res. 2007; 75: 220–228.
Hasler U, Crambert G, Horisberger JD, Geering K. Structural and functional features of the transmembrane domain of the Na,K-ATPase beta subunit revealed by tryptophan scanning. J Biol Chem. 2001; 276: 16356–16364.
White CN, Hamilton EJ, Garcia A, Wang D, Chia KKM, Figtree GA, Rasmussen HH. Opposing effects of coupled and uncoupled NOS activity on the Na+-K+ pump in cardiac myocytes. Am J Physiol Cell Physiol. 2008; 294: C572–C578.
Lupfert C, Grell E, Pintschovius V, Apell HJ, Cornelius F, Clarke RJ. Rate limitation of the Na+,K+-ATPase pump cycle. Biophys J. 2001; 8: 2069–2081.
Dempski RE, Hartung K, Friedrich T, Bamberg E. Fluorometric measurements of intermolecular distances between the alpha- and beta-subunits of the Na+/K+-ATPase. J Biol Chem. 2006; 281: 36338–36346.
Abraham MR, Selivanov VA, Hodgson DM, Pucar D, Zingman LV, Wieringa B, Dzeja PP, Alekseev AE, Terzic A. Coupling of cell energetics with membrane metabolic sensing. Integrative signaling through creatine kinase phosphotransfer disrupted by M-CK gene knock-out. J Biol Chem. 2002; 277: 24427–24434.
Czerwinski A, Gitelman HJ, Welt LG. A new member of the ATPase family. Am J Physiol. 1967; 213: 786–792.
Pieske B, Houser SR. [Na+]i handling in the failing human heart. Cardiovasc Res. 2003; 57: 874–886.