OXIDATIVE INHIBITION OF THE VASCULAR Na⁺-K⁺ PUMP VIA NADPH OXIDASE-DEPENDENT β₁ SUBUNIT GLUTATHIONYLATION: IMPLICATIONS FOR ANGIOTENSIN II-INDUCED VASCULAR DYSFUNCTION

Abstract

Background

Glutathionylation of the Na⁺-K⁺ pump’s β₁ subunit is a key molecular mechanism of physiological and pathophysiological pump inhibition in cardiac myocytes. Its contribution to Na⁺-K⁺ pump regulation in other tissues is unknown, and cannot be assumed given the dependence on specific β subunit isoform expression and receptor-coupled pathways. As Na⁺-K⁺ pump activity is an important determinant of vascular tone through effects on [Ca²⁺]_i, we have examined the role of oxidative regulation of the Na⁺-K⁺ pump in mediating Angiotensin II (Ang II)-induced increase in vascular reactivity.

Methods/results

β₁ subunit glutathione adducts were present at baseline and increased by exposure to Ang II in rabbit aortic rings, primary rabbit aortic vascular smooth muscle cells (VSMCs) and human arterial segments. In VSMCs, Ang II-induced glutathionylation was associated with marked reduction in Na⁺-K⁺ATPase activity, an effect that was abolished by the NADPH oxidase inhibitory peptide, tat-gp91ds. In aortic segments, Ang II-induced glutathionylation was associated with decreased K⁺-induced vasorelaxation, a validated index of pump activity. Ang II-induced oxidative inhibition of Na⁺-K⁺ ATPase and decrease in K⁺-induced relaxation were reversed by pre-incubation of VSMCs and rings with recombinant FXYD3 protein that is known to facilitate deglutathionylation of β₁ subunit. Knock-out of FXYD1 dramatically decreased K⁺-induced relaxation in a mouse model. Attenuation of Ang II signaling in vivo by captopril (8mg/kg/day for 7 days) decreased superoxide-sensitive DHE levels in the media of rabbit aorta, decreased β₁ subunit glutathionylation, and enhanced K⁺-induced vasorelaxation.

Conclusion

Ang II inhibits the Na⁺-K⁺ pump in VSMCs via NADPH oxidase-dependent glutathionylation of the pump’s β₁ subunit, and this newly identified signaling pathway may contribute to altered vascular tone. FXYD proteins reduce oxidative inhibition of the Na⁺-K⁺ pump and may have an important protective role in the vasculature under conditions of oxidative stress.

Graphical Abstract

INTRODUCTION

The ATP-dependent Na⁺-K⁺ pump transports Na⁺ out of cells in exchange for extracellular K⁺ against their respective electrochemical gradients. The gradients generated by the pump are essential for establishment of membrane potential in excitable cells, and are also vital to many other cellular functions including regulation of intracellular ion concentrations and transport of glucose and amino acids.

In the vasculature, it is well established that the Na⁺-K⁺ pump participates in the modulation of vascular smooth muscle contractility and hence vascular tone [1–4]. This is attributed predominantly to its coupling with the Na⁺/Ca²⁺ exchanger, with inhibition of the Na⁺-K⁺ pump increasing intracellular ([Na⁺]_i). This, in turn, activates the reverse mode of the Na⁺/Ca²⁺ exchanger, increasing intracellular Ca²⁺ ([Ca²⁺]_i) and, subsequently, vascular smooth muscle contractility [5–8]. Despite the physiologically important role of the Na⁺-K⁺ pump in vascular function, mechanisms for regulation of its activity in health and disease are poorly understood.

Numerous disease processes, including hypertension, atherosclerosis and diabetes are associated with abnormal vascular function and tone [9–11]. A common factor driving these conditions is the excess generation of oxygen-derived free radicals [12]. As well as quenching nitric oxide (NO) [12], superoxide anions and related reactive oxygen species (ROS) can alter the function of important cellular proteins through oxidative post-translational modification [13–15] and mediate inflammation, cell proliferation, migration, fibrosis and atherosclerosis [11, 16]. Neurohormonal abnormalities, particularly the increased activation of renin-angiotensin system (RAS), are characteristic of many vascular disease states [17]. The potent ability of Angiotensin II (Ang II) to activate NADPH oxidase in the cardiovascular system, augmenting production of ROS (reviewed in [10, 18, 19]) is a major contributor to the pathogenesis of vascular disease.

We have recently identified glutathionylation of the β₁ subunit as a key molecular mechanism of Na⁺-K⁺ pump inhibition in cardiac myocytes [20–22]. This oxidative modification involves the formation of a stable, but reversible mixed disulfide bond between the abundant intracellular tripeptide glutathione and reactive cysteine thiols. The resulting 305 Da negatively charged adduct confers effects on tertiary structure and protein function in a manner similar to phosphorylation and has an increasingly appreciated role in cell signaling [15, 23, 24]. Mutational studies performed confirmed the causal relationship between glutathionylation of Cys46 of the β₁ subunit and oxidant-induced pump inhibition [20], and kinetic studies showed an associated decrease in the rate-limiting E₁ → E₂ conformational change of the pump [20]. Interestingly, susceptibility of β₁ subunit to glutathionylation and thus the pump to oxidative inhibition is dependent on the conformation of the enzyme, being much greater in E1Na(3) than in the E2 state [25]. The Na⁺-K⁺ pump, previously known to associate with caveolin in caveolar microdomains [20, 26], co-immunoprecipitated with NADPH oxidase subunits in cardiac myocytes [27]. Ang II increased co-immunoprecipitation between the activated NADPH oxidase p47^phox subunit and the α₁ Na⁺-K⁺ pump subunit in isolated ventricular myocytes, and resulted in increased β₁ subunit glutathionylation and inhibition of the pump in an NADPH oxidase-dependent manner [20, 28].

Na⁺-K⁺ pump function is also modulated by the FXYD family of proteins. These membrane proteins, named after the FXYD signature sequence in their extracellular domain, co-localize with the Na⁺-K⁺ pump and play a role in kinase-dependent pump regulation [29]. We have recently demonstrated that a highly conserved cysteine residue in the FXYD protein family, provided that it is flanked by basic amino acids (as is the case for all but FXYD2), is susceptible to glutathionylation and is critical for a dynamic role in regulation of the Na⁺-K⁺ pump. FXYD proteins possessing the reactive cysteine residue facilitate deglutathionylation of the β₁ subunit, thus protecting against oxidative inhibition of the pump [30]. The role of FXYD proteins in the vasculature, particularly under conditions of high oxidative stress, is not known.

In the current study we have examined whether glutathionylation of the Na⁺-K⁺ pump’s β₁ subunit and oxidative Na⁺-K⁺ pump inhibition occurs in vascular smooth muscle cells (VSMCs), and whether this contributes to altered vascular function in response to Ang II and NADPH oxidase activation. The impact of FXYD proteins on oxidative Na⁺-K⁺ pump regulation in the vasculature is also investigated. Oxidative pump regulation cannot be assumed to occur or be relevant in vascular smooth muscle given tissue-specific expression of pump subunits, differential susceptibility of pump subunits to glutathionylation, and possible differences in the signaling pathways coupled to the Na⁺-K⁺ pump in vascular smooth muscle cells. The findings may have important implications for the regulation of vasomotor function in both health and disease.

METHODS

Animals, tissues and cells

Experiments were performed on aortas isolated from male New Zealand White rabbits weighing 3 to 3.5 kg. A group of rabbits received the angiotensin converting enzyme (ACE) inhibitor captopril (8mg/kg/day) in their drinking water for 7 days. The experimental protocol was approved by the local animal ethics committee. Anesthesia was achieved using subcutaneous injection of ketamine hydrochloride (50 mg/kg) and xylazine hydrochloride (50 mg/kg). Immediately after cessation of respiration, the chest was opened and the thoracic aorta was excised from the aortic valve to the diaphragmatic hiatus. After the vessels were taken out, they were bathed in Krebs-Henseleit buffer (KHB) containing (in mmol/L) NaCl 118, KCl 4.6, NaHCO₃ 27.2, MgSO₄ 1.2, CaCl₂ 2.5, KH₂PO₄ 1.2, and glucose 11.1 (pH 7.4), aerated with a 95% O₂/5% CO₂ gas mixture, and maintained at a constant temperature of 37°C. Connective and other adhesive tissue was completely removed. The segments were then cut at 5 mm intervals.

VSMCs were isolated from rabbit aortas and maintained in primary culture as previously described [31] in Medium 199 supplemented with 10% fetal calf serum, 2 mmol/L L-glutamine, 50 U/mL penicillin and 50 µg/mL streptomycin in 5% CO₂/95% room air at 37°C.

Segments of human left internal mammary or radial artery from patients undergoing coronary artery bypass graft surgery in our institution that were harvested as potential conduits but remained redundant at the end of surgery were obtained for in vitro experimentation. The use of human tissue was approved by the local research ethics committee, and informed consent was obtained from the tissue donors.

Aortas were also obtained from FXYD1^−/− mice, generated as previously described [32]. Briefly, a FXYD1 knock out (KO) mouse cell line was created by replacing the FXYD1 gene (exons 3–5) with an insert containing lacZ. Heterozygous breeding pairs, backcrossed onto C57Bl/6, were used to generate FXYD1 KO and wild-type (WT) littermates. Mice were deeply anesthetized with isoflurane and euthanized by cervical dislocation prior to excision of aorta for vascular studies.

Immunoblot analysis and detection of glutathionylated proteins

Glutathionylation was detected in VSMCs loaded with biotinylated glutathione ethyl ester (BioGEE) for 1 hour prior to treatment [20]. The biotin-tagged glutathionylated subfraction in cell lysate was precipitated using streptavidin-Sepharose beads and immunoblotted for α₁ and β₁ subunits and FXYD proteins. With an alternative technique, an antibody against a glutathionylated cysteine epitope (anti-GSH antibody) was used in co-immunoprecipitation protocols. Protein A/G plus agarose beads were used to immunprecipitate the glutathionylated proteins as previously described [20]. This technique was also used for detection of glutathionylation in rabbit aorta and human internal mammary artery samples, and for co-immunoprecipitation of p47^phox with α₁ pump subunit. Samples were subjected to SDS-PAGE and probed with appropriate antibodies using standard Western blot techniques.

Recombinant FXYD proteins

To examine the effect of FXYD proteins on glutathionylation of Na⁺-K⁺ pump β₁ subunit in VSMCs, and on the activity of pump in VSMCs and aortic segments, we exposed cells or aortic rings to recombinant FXYD proteins, synthesized as described previously [33]. The feasibility of achieving spontaneous membrane insertion was suggested by the hydrophobic nature of the FXYD proteins and by previously published findings that: 1) FXYD proteins associate with Na⁺,K⁺ pump subunits [34]; 2) exogenous, purified FXYD10 partitions into membrane fragments and associates with the Na⁺-K⁺ pump within 5 min at 23 °C as indicated by its co-immunoprecipitation and cysteine cross-linking with the α subunit [35, 36]; and 3) incubation of isolated cardiac myocytes in solution containing recombinant FXYD3 (rFXYD3) proteins (not natively expressed in the heart) results in association of the rFXYD3 protein with the native Na⁺-K⁺ pump [30].

Since immunodetection cannot distinguish between recombinant FXYD1 and FXYD1 native to the VSMCs, we used FXYD3. We also used a derivative with all 4 cysteines mutated to serine as a control, referred to as “Cys-less” FXYD3. Recombinant FXYD3 proteins were dissolved in dimethyl sulfoxide (DMSO) and then diluted in experimental solutions. The final solutions contained DMSO in a concentration of ∼0.05%. DMSO in concentrations up to ten-fold higher has no functional effects on the Na⁺-K⁺ pump in cardiac myocytes [37].

Na⁺-K⁺ ATPase activity measurement

VSMCs were washed three times with phosphate buffered saline (PBS), scraped in 1 mL of homogenization buffer (5 mmol/L histamine-imidazole, 2 mmol/L EDTA, 1 mmol/L EGTA, 1 mmol/L phenylmethylsulfonyl fluoride, protease inhibitor cocktail tablet (EDTA free; Roche Applied Science) and rinsed with an additional 1 mL of the same buffer. The homogenate was centrifuged (1,000 g for 5 min at 4°C) to remove unbroken cells and debris. The supernatant was then centrifuged at 60,000 g at 4°C for 1 h. The supernatant was the cytosolic fraction. The pellet (i.e. the membrane fraction) was resuspended in 500 µL homogenization buffer. Protein concentrations were determined by BCA protein assay kit (Pierce) and Na⁺-K⁺ ATPase activity was determined in membrane fractions of VSMCs. The total activity reaction mixture contained 130 mmol/L NaCl, 20 mmol/L KCl, 3 mmol/L ATP, 3 mmol/L MgCl₂, and 30 mmol/L imidazole with and without the addition of 5 mmol/L ouabain. After 30 minutes incubation period at 37 °C, the reaction was terminated with 1% trichloroacetic acid, and the precipitated proteins were pelleted. Inorganic phosphate (Pi) was determined by utilizing molybdate-H₂SO₄ solution with Fiske reducing agent [38]. The Na⁺-K⁺ ATPase activity was calculated as the difference in ATPase activity between cells exposed to ouabain and those not exposed.

Organ bath vasomotor studies

Rabbit aortic rings measuring ∼5 mm in length were mounted on two stainless steel hooks immersed in 25 ml organ bath chambers. The isometric contractions were measured in organ chambers at 37°C by using a force transducer (Radnoti) and recording with Powerlab software (AD Instruments, United States). Vascular rings were allowed to equilibrate for 0.5 hr and tension was gradually increased to 2 g. Vessels were then contracted with the addition of 80 mmol/L KCl to determine integrity of the rings. After maximum contraction was achieved, the aortic rings were washed with KHB and re-equilibrated, followed by replacement of KHB with KHB that was nominally K⁺-free. After 20 min in K⁺-free KHB, rings were pre-contracted with phenylephrine (100 nmol/L). Once a plateau was achieved, increasing concentrations of K⁺ were administered as previously described [39]. As the Na⁺-K⁺ pump is inhibited in the absence of extracellular K⁺, and reactivated by its reintroduction at low concentrations, Webb and Bohr characterized K⁺-induced relaxation as an index of Na⁺-K⁺ ATPase activity in arteries; demonstrating it to be dependent on external K⁺, intracellular Na⁺ and Mg²⁺, and inhibited by ouabain [39]. Because K⁺ concentrations higher than 6 mmol/L may induce depolarization and contraction, the response to addition of 6 mmol/L KCl was the maximum recorded. To examine the effects of in vitro exposure to Ang II, aortic rings from the same animal were exposed to Ang II (100 nmol/L, 37 °C, 1h) or control solutions in the organ bath. Ang II was then washed out prior to tensioning and measurement of K⁺-induced vasorelaxation. In some experiments aortic rings were pre-incubated with the rFXYD3 (500 nM) or vehicle before the Ang II incubation step. Pre-incubation in Ang II, or rFXYD3 had no significant effect on absolute phenylephrine-induced contraction (2.1 ± 0.3 g and 2.4 ± 0.2 g respectively) compared with control (2.3 ± 0.2 g). K⁺-induced relaxation is expressed as % of phenylephrine (100 nmol/L)- induced contraction. For mouse aortic studies, aortic rings were ∼4 mm in length with intact endothelium, and studied in physiological salt solution (PSS) of the following composition (mmol/L): KCl 4.7, CaCl₂ 2.5, NaCl 118.3, KH₂PO₄ 1.2, MgSO₄ 0.6, NaHCO₃ 25, and dextrose 5.5. Rings were stretched incrementally to an optimal tension of 2 g during one hour with repeated washing. After a 30-min equilibration, rings were contracted using 50 mmol/L KCl and rinsed. After another 30-min equilibration, rings were washed and maintained in nominally potassium free PSS for 20 min and contracted with phenylephrine to approximately 1 g. Relaxations to cumulative increasing concentrations of KCl (2 to 6 mM) were determined and expressed as percent decrease in the phenylephrine-induced force.

Confocal immunofluorescence microscopy

To detect glutathionylation of the β₁ pump subunit in situ, VSMCs grown on coverslips were first treated with s-nitrosoglutathione (GSNO, 0.2 mmol/L , 37 °C, 1 hr) or Ang II (100 nmol/L, 37 °C, 1hr). Cells were then washed in PBS to remove treatment and residual media, fixed with 3.7% (w/v) formaldehyde in PBS for 10 min at room temperature, permeabilized in 0.1% (w/v) Triton-X in PBS for a further 10 min and blocked in 5 % (w/v) bovine serum albumin in PBS at 4 °C overnight. Cells were then incubated with anti-GSH mouse monoclonal antibody (1:500) and in anti-Na⁺-K⁺ pump β₁ subunit rabbit polyclonal antibody (1:100) overnight at 4°C. Antigen-antibody complexes were visualized with fluorescein isothiocyanate (FITC) conjugated goat anti-mouse (1:500) and Alexa Fluor 594 chicken anti-rabbit conjugated secondary antibodies (1:500). The cover slips were mounted with ProLong Gold antifade reagent with 4’,6-diamidino-2-phenylindole (DAPI) to stain the nuclei. Confocal images were detected by a Leica TCS SP5 spectral confocal scanner and a Leica DMI 6000 CS microscope (Leica, Mannheim, Germany) equipped with a x60 oil-immersion objective. To avoid cross talk between the two fluorescence dyes, we used the sequential method in which the two laser lines 488 and 594 nm were applied to the cells alternately. Visualization and analysis were performed using Leica Confocal Microscope Systems software.

To detect O₂^•− level in the media of the aorta, cryosections (10 µm) of aorta in Tissue-Tek optimized cutting temperature (OCT) medium were allowed to thaw and incubated with O₂^•−-sensitive fluorescent dye dihydroethidium (DHE) (2 µM) at 37 °C for 20 minutes in the dark, . The sections were fixed by paraformaldehyde and images obtained by a laser-scanning confocal fluorescent microscope (Leica TCS SP5) using an excitation wavelength of 488 nm. Fluorescence was detected at 585 nm. The signal intensity within the media layer was analysed by measurement of the signal from 10 randomly-chosen regions of interest per section by a blinded researcher. The DHE fluorescence detected by this technique is thought to reflect reactivation of enzymatic sources of ROS, in this case NADPH oxidase [40] during thawing. Although distant from physiological conditions, microtopography of O₂^•− has an advantage over HPLC-based measurement of DHE oxidation in allowing focused examination of DHE-fluorescence from the media [41] and has been used extensively by researchers in the Field. As it is well recognized that DHE-fluorescence also detects H₂0₂-derived products [42], we examined the specificity of baseline DHE fluorescence for O₂^•− by incubating some sections in polyethelene glycol adducted superoxide dismutase (PEG-SOD; 1000 U/mL; 15 min prior to DHE). As shown in Online Figure 1, PEG-SOD substantially decreased DHE fluorescence from the media under baseline conditions. Auto-fluorescence of the elastic lamina was not affected and was excluded from analysis.

Materials

Angiotensin II, GSH ethyl ester, ouabain, DHE and PEG-SOD were purchased from Sigma. Sulfo-NHS biotin was obtained from Merck, streptavidin-sepharose beads from GE Healthcare Bio-Sciences, the chemiluminescence kit from Pierce; and protein A/G plus agarose beads from Santa Cruz Biotechnology. Antibodies were obtained from the following vendors: α₁- and β₁-subunits of Na⁺-K⁺-ATPase from Upstate Biotechnology, FXYD1 and FXYD3 from Abcam, goat anti-mouse IgG-HRP secondary antibody from Pierce, anti-GSH and Alexa Fluor 594 chicken anti-rabbit-conjugated secondary antibody from Invitrogen; and and α-tubulin antibody as well as FITC-conjugated goat anti-mouse antibody from Santa Cruz Biotechnology.

Data analysis

Each presented immunoblot is representative of at least three separate experiments. The band densities were quantified by densitometry using an LAS-3000 (Fujifilm, Japan). Data are expressed as the mean ± standard error of the mean (SEM). Statistical comparisons were made with non-paired Student’s t test between two groups or one-way ANOVA for multiple comparisons. K⁺-induced relaxation data were analysed by 2-way ANOVA followed with Sidak’s (for Ang II and captopril experiments) or Dunnett’s (for FXYD effects) post-hoc analyses for multiple comparisons. P<0.05 was considered to be statistically significant.

RESULTS

Na⁺-K⁺ pump β₁ subunit is glutathionylated in VSMCs

Susceptibility of the Na⁺-K⁺ pump’s α₁ and β₁ subunits to glutathionylation was examined in primary cultured aortic VSMCs from rabbits using the BioGEE technique. The β₁ subunit was detected in the biotin-tagged glutathionylated subfraction at baseline. Exposure of the VSMCs to GSNO (0.2 mM for 1 hour), which is known to induce glutathionylation of susceptible proteins by permitting nucleophilic attack of its sulfur residue by the protein thiolate anion [43], increased the signal of GSS-β₁ by ∼70% (Figure 1A). However, there was no evidence of α₁ subunit glutathionylation at baseline, or after GSNO exposure (data not shown). The signal for glutathionylated β₁ subunit was not detected when the lysate was incubated with 1 mM DTT prior to precipitation by streptavidin. This is supportive of a mixed disulfide bond between the β₁ subunit and GSH.

Figure 1. Glutathionylation of Na⁺-K⁺ pump subunits in VSMCs under oxidative stress.

β₁ subunit glutathionylation with and without GSNO detected using the BioGEE (A) and the GSH antibody technique (B). Immunoprecipitation (IP) with non-immune IgG was used as a negative control to ensure the specificity of the β₁ subunit IP. C. Confocal micrographs of VSMCs demonstrating β₁ subunit (left column, red fluorescence), and glutathionylation (second column, green fluorescence) in control rabbit primary VSMCs and cells incubated with GSNO for 1 hour. The third column shows the merged glutathionylation/β₁ subunit image along with 4′,6-diamidino-2-phenylindole (DAPI) staining of the nucleus (blue). Yellow fluorescence reflects co-localization of GSH and β₁ subunit, thus the glutathionylated population of the β₁ subunit. Images are representative of 3 independent experiments. D. Ouabain-sensitive ATPase activity in VSMCs exposed to GSNO or control solutions for 1 hour. n=5. * indicates p<0.05. TL: total cell lysate; GSSPs: total glutathionylated proteins, sampled as the biotinylated fraction; IP: immunoprecipitate; and IB: immunoblot.

To examine glutathionylation of β₁ subunit in VSMCs with a separate technique, β₁ subunit immunoprecipitate was immunoblotted with anti-GSH antibody. Similar to the BioGEE technique, a moderate degree of β₁ glutathionylation was present at baseline, and exposure of the cells to GSNO increased the glutathionylation signal (Figure 1B).

To determine the subcellular localization of the glutathionylated population of the pump’s β₁ subunit, immunofluorescence studies were performed. β₁ subunit was detected at the plasma membrane, and was unaffected by GSNO (Figure 1C, red fluorescence in left column). The β₁ subunit was also detected in the peri-nuclear region consistent with previous reports of β subunit assembly and maturation occurring in the Golgi and sarcoplasmic reticulum [44, 45]. A small signal for total cellular glutathionylation was observed under baseline conditions by GSH antibody (Figure 1C, green fluorescence in middle column). This was substantially increased after exposure to GSNO, and co-localized with the β₁ subunit at both the plasma and intracellular pools, reflected by the yellow fluorescence signal (Figure 1C, right column).

Consistent with the causal relationship between glutathionylation of β₁ subunit and inhibition of the Na⁺-K⁺ ATPase shown previously in mutational studies [20], GSNO caused a reduction in the activity of the enzyme measured by ouabain-sensitive ATPase activity in VSMCs (Figure 1D).

Ang II increases glutathionylation of the β₁ subunit and inhibits the Na⁺-K⁺ pump via NADPH oxidase in VSMCs

We have previously reported that the activated NADPH oxidase complex associates with the Na⁺-K⁺ pump in cardiac myocytes [27]. Since Ang II activates NADPH oxidase in VSMCs [46], we examined co-immunoprecipitation of the Na⁺-K⁺ pump with p47^phox NADPH oxidase subunit in VSMCs. p47^phox/α₁ subunit co-immunoprecipitation was minimal under baseline conditions, and this was increased significantly with Ang II exposure (Figure 2A), consistent with translocation of the p47^phox subunit to the membrane known to occur upon phosphorylation and required for activation of the oxidase [47, 48].

Figure 2. Ang II and NADPH oxidase-dependent β₁ subunit glutathionylation in VSMCs.

A. p47^phox subunit immunoblot (IB) of VSMC lysate (TL); α₁ subunit immunoprecipitate (IP); and non-specific IgG IP showing increased co-immunoprecipitation in response to Ang II, consistent with translocation of this cytosolic NADPH oxidase subunit to the membrane and its association with the Na⁺-K⁺ pump molecular complex (n=5). α₁-IB of α₁ IP was used as a control. B. Glutathionylation of Na⁺-K⁺ pump β₁ subunit in VSMCs after exposure to Ang II (100 nM for 1 hour), with or without preincubation in the membrane permeable NADPH oxidase inhibitor peptide tat-gp91ds (5 µM for 1 hour; n=5 per group). C. Confocal micrographs of VSMCs demonstrating β₁ subunit (left column, red fluorescence), and glutathionylation (second column, green fluorescence) in control rabbit primary VSMCs and cells incubated with Ang II (100 nM) for 1 hour. The third column shows the merged glutathionylation/β₁ subunit image along with DAPI staining of the nucleus (blue). Yellow fluorescence reflects colocalization of GSH and β₁ subunit. Images are representative of 3 independent experiments. D. Na⁺-K⁺ ATPase activity in VSMCs exposed to Ang II (100 nM for 1 hour) or control solutions, with and without co-incubation in tat-gp91ds (n=5). * indicates p<0.05.

We next examined the effect of Ang II on β₁ pump subunit glutathionylation. Exposure of VSMCs to Ang II increased β₁ subunit glutathionylation (Figure 2B) detected by the BioGEE technique. The increase was abolished by pre-incubating the cells in cell-permeable inhibitory peptide of NADPH oxidase, tat-gp91ds, thus demonstrating NADPH-oxidase-dependence of the effect. Ang II-induced β₁ subunit glutathionylation and its subcellular localization were also demonstrated with immunofluorescence. Ang II increased the signal representing the co-localized GSH and β₁ subunit (Figure 2C). In functional studies, Ang II-induced β₁ subunit glutathionylation was associated with a reduction in Na⁺-K⁺ ATPase activity (Figure 2D); an effect that was also NADPH oxidase-dependent as demonstrated by abolishment of the inhibition by tat-gp 91ds.

Ang II-induced oxidative inhibition of the Na⁺-K⁺ pump occurs in intact aorta and human arteries and decreases K⁺-induced vasorelaxation

Ang II is well known to increase vasomotor tone. In order to explore the possible role of NADPH oxidase-dependent inhibition of the Na⁺-K⁺ pump in this phenomenon we studied the effect of Ang II on β₁ subunit glutathionylation in rabbit intact aortic rings (Figure 3A). β₁ subunit glutathionylation was detected at baseline and was increased by ∼60% by exposure of the rings to Ang II.

Figure 3. Ang II and oxidative inhibition of the Na⁺-K⁺ pump in aorta.

A. Effect of Ang II on β₁ subunit glutathionylation in intact rabbit aortic rings as determined by the GSH antibody technique (n=9). Aortic rings were exposed to Ang II (100 nM) or control solution for 1 hour prior to lysis. B. Effect of Ang II on K⁺-induced vasorelaxation in rabbit aorta. Relaxation is expressed as % of phenylephrine induced contraction (n=5). C. Effect of Ang II on β₁ subunit glutathionylation in human arterial segments (n=6). * indicates p<0.05.

To assess the effect of Ang II-induced oxidative modification on Na⁺-K⁺ pump activity and Na⁺-K⁺ pump-dependent alterations in vascular tone, we used the method described by Webb and Bohr [39]. In rings pre-incubated in K⁺-free solutions for 20 min and pre-contracted with phenylephrine, addition of KCl in a cumulative fashion induced a concentration-dependent relaxantion response as previously described [39] that was independent of the presence of endothelium (n=6; p=0.88; data not shown). Consistent with its effect on glutathionylation of the β₁ subunit and inhibition of Na⁺-K⁺ ATPase activity in VSMCs, pre-incubation in Ang II decreased K⁺-induced vasorelaxation (Figure 3B).

To examine the relevance of findings in cultured cells and rabbit aorta to humans, we examined oxidative modification of the pump in human arterial segments. We detected β₁ subunit glutathionylation at baseline. The signal was increased by exposing the vessels to Ang II ex vivo (Figure 3C), demonstrating that the effect of Ang II-mediated signaling on downstream oxidative modification of the Na⁺-K⁺ ATPase also occurs in the human vasculature.

In vivo captopril treatment decreases vascular β₁ subunit glutathionylation, increases Na⁺-K⁺ pump activity, and improves K⁺-induced vasorelaxation

The renin-angiotensin system is active under basal physiological conditions and since we detected a degree of baseline β₁ subunit glutathionylation in VSMCs, we examined the effect of renin-angiotensin system disruption at baseline on β₁ subunit glutathionylation and Na⁺-K⁺ pump activity by administering the ACE inhibitor captopril to healthy rabbits. Consistent with the effect of Ang II in activation of NADPH oxidase, attenuation of Ang II signaling by captopril in vivo decreased O₂^•−-sensitive DHE fluorescence in the media of aortic rings (Figure 4A). Treatment with captopril decreased β₁ subunit glutathionylation (Figure 4B), and markedly enhanced K⁺-induced vasorelaxation suggesting a significant increase in the activity of the pump by ACE inhibition (Figure 4C). Taken together, these results suggest that the signaling pathways for oxidative inhibition of Na⁺-K⁺ pump shown in in vitro experiments are also present in the more complex in vivo milieu and are therefore of physiological significance.

Figure 4. Effect of ACE inhibition by captopril in vivo on oxidative regulation of the Na⁺-K⁺ pump in rabbit aorta.

A. DHE fluorescence in media of aorta from rabbits fed captopril (8 mg/kg) or control drinking solutions for 7 days. The histogram shows average DHE fluorescence signal from the media, excluding elastic lamina (n=4 rabbits per group). B. β₁ subunit glutathionylation in captopril-fed and control rabbits (n=4). C. K⁺-induced vasorelaxation of aortic rings from captopril versus control rabbits (n=4–6). * indicates p<0.05.

FXYD1 modulates oxidative inhibition of the Na⁺-K⁺ pump in vessels

FXYD1, but not FXYD3 was detected in rabbit aortic primary VSMCs by standard immunoblotting techniques (Figure 5A). Exposure of rabbit aortic VSMCs to rFXYD3, chosen to allow immunodetection against the background expression of FXYD1, resulted in the co-immunoprecipitation of rFXYD3 with the native α₁ subunit (Figure 5B). This is consistent with results in isolated cardiac myocytes in which incubation of rFXYD3 resulted in competitive displacement of native FXYD1 [30] from its association with native α subunit. Incubation of rabbit aortic VSMCs with rFXYD3 had no effect on baseline β₁ subunit glutathionylation, but protected against Ang II-induced increase in β₁ subunit glutathionylation (Figure 5C), and associated oxidative inhibition of the Na⁺-K⁺ pump (Figure 5D).

Figure 5. FXYD proteins and oxidative regulation of the Na⁺-K⁺ pump in aorta and VSMCs.

A. Immunoblot (IB) of FXYD1 in primary rabbit aortic VSMC lysate, with recombinant FXYD1 (rFXYD) loaded as a positive control. FXYD1 is detected at ∼10 kDa (shown with arrow). B. α₁ and FXYD3 immunoblot of VSMC total lysate (TL) and α₁ subunit immunoprecipitate (IP) in VSMCs preincubated with or without rFXYD3. IB of α-tubulin has been used as a loading control. C. The effect of rFXYD3 on Ang II-induced β₁ subunit glutathionylation in primary aortic VSMCs (pre-incubated with rFXYD3 before exposure to Ang II (100 nM) or control solutions for 1 hour). D. The effect of rFXYD3 on Ang II-induced inhibition of Na⁺-K⁺ ATPase activity in VSMCs. The cells were incubated with rFXYD with or without subsequent exposure to 100 nM Ang II for 1 hour.

Given the observed effect of rFXYD3 in decreasing Ang II-induced β₁ subunit glutathionylation in VSMCs, we examined the effect of FXYD proteins on K⁺-induced vasorelaxation. Pre-incubation of rabbit aortic rings in rFXYD3 protected against Ang II-induced decrease in K⁺-induced vasorelaxation (Figure 6A). This effect was not seen when the aortic rings were incubated in Cys-less rFXYD3 (Figure 6A). We also examined K⁺-induced relaxation in the aorta of male FXYD1 KO mice. There was no difference in KCl or PE-induced contractions of aortic rings in the WT and KO mice (data not shown). However, consistent with the role of FXYD protein in decreasing oxidative inhibition of the vascular Na⁺-K⁺ pump, FXYD1 KO markedly decreased K⁺-induced relaxation under basal conditions (Figure 6B).

Figure 6. Effect of FXYD proteins on K⁺-induced relaxation.

A. Effect of Ang II on K⁺-induced vasorelaxation in rabbit aortic rings after preincubation in rFXYD3 or cys-less rFXYD3 (500 nM) (n=5–6 per group). B. Effect of FXYD1 KO on K⁺-induced vasorelaxation in mouse aorta under baseline conditions (n=3 per group). Relaxation is expressed as % of phenylephrine induced contraction. * indicates p<0.05.

DISCUSSION

In this study we have shown that Ang II increases glutathionylation of the Na⁺-K⁺ pump’s β₁ subunit and decreases the activity of Na⁺-K⁺ ATPase in VSMCs in an NADPH oxidase-dependent manner. This novel oxidative pathway is also observed in both rabbit and human vessels. The significance in regulation of the pump activity in vessels is suggested by the associated reduction in K⁺-induced vasorelaxation in aortic rings exposed to Ang II in vitro, and increase in K⁺-induced vasorelaxation by disruption of the renin-angiotensin system by ACE inhibition in vivo. Localization of Na⁺-K⁺ pump in caveolae [27, 49] is critical for its physiological coupling to NADPH oxidase and nitric oxide synthase (NOS) [20, 27, 50, 51]. In cardiac myocytes we have shown this to be important for hormonal regulation of the Na⁺-K⁺ pump [27, 28, 50, 52]. The current study is the first to demonstrate oxidative regulation of the pump in the vasculature. The results cannot be assumed from our previous work in cardiac myocytes, given tissue-specificity of Na⁺-K⁺ pump β subunit expression, the lack of susceptibility of the β₂ and β₃ subunits to glutathionylation, and the necessity for receptor-coupled activation of NADPH oxidase within the specific signaling microdomain. The effects of the oxidative pathway on the pump function in VSMCs we show here are expected to mediate alterations in vascular contractility through the effects of [Na⁺]_i on [Ca²⁺]_i [5–8]. This may have significance for altered regulation of vascular tone in pathophysiological conditions such as diabetes mellitus, obesity and hypertension that are characterized by neurohormonal abnormalities and increased oxidative stress.

Treatment with angiotensin receptor blockers or ACE inhibitors is recognized as one of the most effective therapies in reducing blood pressure; and in preventing progression and complications of vascular disease, particularly the atherosclerotic process [53–56]. The data presented in this study, that Ang II results in NADPH oxidase-dependent inhibition of the Na⁺-K⁺ pump via glutathionylation of its β₁ subunit, and that treatment with an ACE inhibitor reverses oxidative inhibition of the Na⁺-K⁺ pump, suggests an additional mechanism for the observed benefit of antagonizing renin angiotensin system in disease states associated with oxidative stress and abnormal vascular function.

A limitation of the study, that is shared by all studies of the Na⁺-K⁺ pump in the vasculature, is that although K⁺-induced vasorelaxation is a well-validated index of Na^+-K⁺ pump activity, we cannot determine the precise contribution of Ang II-induced pump inhibition to the total increase in vascular tone in vivo. One targeted approach that we are pursuing that may clarify this important question is to develop a mouse model with a point knock-in mutation of the reactive cysteine of the β₁ pump subunit, and then to study the effects of Ang II signaling in vivo. In addition, our study has not specifically focused on the NADPH oxidase (NOX) isoform involved in Ang II-induced glutathionylation of the vascular Na⁺-K⁺ pump. Although VSMCs express NOX1, NOX4 and NOX5, it is NOX1 that has a firmly established role in a multitude of disease states characterized by neurohormonal abnormalities: Ang II-induced hypertension and vascular O₂^•− production are exacerbated in transgenic mice overexpressing the NOX1 subunit [57, 58]; and are blunted in mice that are globally deficient in the NOX1 subunit [59]. Our data showing an Ang II-induced translocation of p47^phox subunit to the microdomain of the Na⁺-K⁺ pump, and blockade of the Ang II-induced inhibition of the pump by tat-gp91ds (Figure 2) also suggest NOX1 as the major isoform mediating Ang II effects, since the regulatory subunits p22^phox and p47^phox are a critical component of NOX1 complex, but not NOX4 and 5 [60].

NADPH oxidase has previously been reported to be electrogenic in leukocytes [61]. Hyperpolarization of the cell membrane due to the outward electron leakage upon activation of NADPH oxidase, shown in leukocytes, may contribute to inhibition of the voltage-dependent Na⁺-K⁺ pump in VSMCs, independent of redox modifications. The amplitude of this NADPH oxidase-related current has been measured at ∼5 pA in phagocytes [61] in which NADPH oxidase is highly abundant. Since the membrane potential of VSMCs is approximately −72 mV, with a transient current of –20 pA [62], the current that may result from NADPH oxidase activation in VSMCs is predicted to be small (< 5 pA as opposed to leukocytes given lower abundance of NADPH oxidase), and is unlikely to hyperpolarize the VSMCs enough to alter Na⁺-K⁺ pump activity. Our previous observation that Ang II inhibition of the Na⁺-K⁺ pump in cardiac myocytes is completely abolished by the inclusion of SOD in patch pipette solutions [27] also lends further support for the ROS-dependent rather than electrogenic, hyperpolarizing effect of NADPH oxidase on pump activity in VSMCs.

Since there is a moderate degree of baseline β₁ subunit glutathionylation, signaling pathways that promote deglutathionylation may result in Na⁺-K⁺ pump stimulation, and thus indirectly result in enhanced vasorelaxation. NO, well known for its vasorelaxant properties [63–65], stimulates the Na⁺-K⁺ pump in a number of tissues [66–70]. In the cardiac myocytes, activation of NOS by β₃ adrenergic receptor agonists results in a decrease in β₁ subunit glutathionylation and soluble guanylyl cyclase (sGC)-dependent Na⁺-K⁺ pump stimulation [51]. The mechanism by which NO/sGC/cyclic guanidine monophosphate (cGMP) signaling results in a reduction in a functionally significant oxidative modification of a key caveolar protein remains to be determined and is the focus of ongoing research in our laboratory. However, if such a pathway also exists in VSMCs, it has important implications for further understanding of the sGC/cGMP-dependent vasorelaxant effects of NO.

The effect of the rFXYD3 protein to protect against Ang II-mediated reduction in K⁺-induced vasorelaxation paralleled the reduced K⁺-induced vasorelaxation in the FXYD1 knockout mice (Figure 6). It is consistent with our prior work in cardiac tissue, which showed a more than 3-fold increase in β₁ subunit glutathionylation over baseline in the FXYD1-kockout hearts [30], as well as the 50% lower level of Na⁺-K⁺ ATPase activity measured in cardiac membranes [32]. The data in Figure 6 further supports the interpretation of K⁺-induced relaxation as an indirect, but physiologically relevant measure of vascular sodium pump activity [39], as well as point to a potential role for FXYD proteins in novel therapeutic approaches to ROS-induced alterations in vascular function.

Despite the well-accepted pathophysiological role of ROS in many disease states, non-specific antioxidants such as Vitamin E have had disappointing results in clinical trials [71, 72]. It has been proposed that the effects of ROS are dependent on the site of their synthesis, and that a targeted inhibition is likely to be a more successful approach than the use of broad-acting antioxidants [73, 74]. The ability of rFXYD3 to incorporate in the membrane with the native Na⁺-K⁺ pump in VSMCs when applied outside the cell is similar to what we have demonstrated in cardiac myocytes [30], and consistent with the observed co-immunoprecipitation of α subunits in kidney membrane fragments exposed to exogenous FXYD10 from shark rectal gland [35]. The lipophilic nature of this peptide likely facilitates this ability and suggests the potential for FXYD proteins to be adapted for therapeutic use in disease states of high oxidative stress and vascular dysfunction, reaching the microdomain of the Na⁺-K⁺ pump that non-specific antioxidants fail to influence. Identifying techniques to enhance expression of endogenous FXYD1 in vascular tissue may be an alternative approach for protecting caveolar proteins from ROS effects, and further understanding of the mechanism by which the FXYD proteins facilitate deglutathionylation may identify novel treatment targets.

The mechanisms by which FXYD proteins with a reactive cysteine residue, either endogenous or recombinant, interact with Cys46 on the β₁ subunit of the pump and facilitate deglutathionylation remain unclear. Although we know that the FXYD “C2” (the second cysteine from the membrane, which is highly conserved and flanked by basic amino acids) is critical for the functional partnership [30], and is glutathionylated, it is uncertain whether this glutathionylation is a primary event in the regulatory pathway, or whether it just reflects the reactivity of this cysteine, and it acts via an alternative mechanism. As we have previously discussed [30], there is a large distance between the reactive cysteines of the two proteins in the three-dimensional structure [75]. It is possible that glutathionylation of the FXYD protein associated with the pump complex results in weakening of the FXYD-α subunit bonds, and that this allows the β subunit to move to a position that facilitates deglutathionylation, such as via enzymatic catalysis of de-gluatathionylation by glutaredoxin 1. The functional role of palmitoylation of this FXYD cysteine, which has also been observed [76], is uncertain in the context of the regulation of β₁ subunit glutathionylation and oxidative regulation of the Na⁺-K⁺ pump in either the heart or the vasculature.

Endogenous cardiotonic steroids are thought to contribute to elevations in blood pressure partially through a direct inhibitory effect on Na⁺-K⁺ pump activity [5, 6, 77, 78]. However, additional longer term mechanisms may also play a role. These include actions of the Na⁺-K⁺ pump as a signaling molecule coupled to intracellular kinases and gene transcription driving vascular smooth muscle hypertrophy and proliferation [79, 80]. Glutathionylation of the β₁ subunit, which slows E1→E2 conformational change of the pump [20], is predicted to increase the proportion of pump complexes in the E1 conformational state. Since E1 Na⁺-K⁺ ATPase can bind both the Src homology 2 (SH2) domain and kinase domains simultaneously, and has been shown to keep Src in an inactive state, β₁ subunit glutathionylation might therefore decrease Na⁺-K⁺ ATPase-dependent Src kinase activation [81, 82]. Thus the oxidative inhibition of the Na⁺-K⁺ pump in VSMCs that we describe here may have chronic effects on proliferation and hypertrophy of VSMCs that require further investigation.

In summary, we demonstrate redox-regulation of Na⁺-K⁺ pump function in the vasculature in response to a major neurohormonal regulator of vascular function. Our data may have implications for vascular dysfunction in a variety of disease states characterized by elevated oxidative stress. Furthermore, the ability of recombinant FXYD protein to protect against oxidative Na⁺-K⁺ pump inhibition may present a potential therapeutic strategy.

Highlights.

• -Glutathionylation of the Na⁺-K⁺ pump β₁ subunit (β₁-GSS) inhibits the pump in heart.
• -Regulation of Na⁺-K⁺ pump function in the vasculature is not fully understood.
• -We show that Ang II inhibits the vascular pump via NADPH oxidase and ↑ β₁-GSS.
• -FXYD protein protected against Ang II effects on β₁-GSS and vascular pump activity.
• -Redox-dependent pump inhibition may be significant in regulation of vascular tone.

Abbreviations

[Ca²⁺]_i

intracellular calcium concentration

VSMCs

vascular smooth muscle cells

Ang II

angiotensin II

NADPH oxidase

nicotinamide adenine dinucleotide phosphate-oxidase

DHE

dihydroethidium

β1-GSS

glutathionylated β1 subunit of Na⁺-K⁺ pump

[Na⁺]_i

intracellular calcium concentration

NO

nitric oxide

NOX

NADPH oxidase

ACE

angiotensin converting enzyme

KHB

Krebs-Henseleit buffer

BioGEE

biotinylated glutathione ethyl ester

GSH

glutathione

DMSO

dimethyl sulfoxide

PBS

phosphate buffer saline

BCA

bicinchoninic acid

GSNO

s-nitrosoglutathione

FITC

fluorescein isothiocyanate

DAPI

4’,6-diamidino-2-phenylindole

PEG-SOD

pegylated superoxide dismutase

RAS

renin-angiotensin system

Pi

inorganic phosphate

sGC

soluble guanylyl cyclase

cGMP

cyclic guanidine monophosphate