Oxidative Stress Inhibits Vascular KATP Channels by S-Glutathionylation*

The KATP channel is an important player in vascular tone regulation. Its opening and closure lead to vasodilation and vasoconstriction, respectively. Such functions may be disrupted in oxidative stress seen in a variety of cardiovascular diseases, while the underlying mechanism remains unclear. Here, we demonstrated that S-glutathionylation was a modulation mechanism underlying oxidant-mediated vascular KATP channel regulation. An exposure of isolated mesenteric rings to hydrogen peroxide (H2O2) impaired the KATP channel-mediated vascular dilation. In whole-cell recordings and inside-out patches, H2O2 or diamide caused a strong inhibition of the vascular KATP channel (Kir6.1/SUR2B) in the presence, but not in the absence, of glutathione (GSH). Similar channel inhibition was seen with oxidized glutathione (GSSG) and thiol-modulating reagents. The oxidant-mediated channel inhibition was reversed by the reducing agent dithiothreitol (DTT) and the specific deglutathionylation reagent glutaredoxin-1 (Grx1). Consistent with S-glutathionylation, streptavidin pull-down assays with biotinylated glutathione ethyl ester (BioGEE) showed incorporation of GSH to the Kir6.1 subunit in the presence of H2O2. These results suggest that S-glutathionylation is an important mechanism for the vascular KATP channel modulation in oxidative stress.

ATP-sensitive K ϩ (K ATP ) 3 channels are regulated by intracellular ATP/ADP and couple intermediary metabolic states to membrane excitability. The activity of these channels is low under physiological conditions and drastically rises during metabolic stress (1,2). The K ATP channels are expressed in almost all tissues. In vascular smooth muscle cells (SMCs), activation of K ATP channels by several vasodilators reduces SMC membrane excitability, leading to vasorelaxation (3,4). Activity of the channels is inhibited by vasoconstrictors (5), resulting in depolarization of the SMCs and vasoconstriction. Such a com-mon vascular regulator is thus targeted by a variety of cellular events in physiological and pathological conditions (1,6).
One important cellular event is oxidative stress that is known to play an important role in the development and maintenance of several cardiovascular diseases, such as hypertension, atherosclerosis, and diabetic vascular complications (7,8). During oxidative stress, excessive reactive oxygen species (ROS), such as superoxide (O . ), hydroxyl radical (OH⅐), and hydrogen peroxide (H 2 O 2 ), are overly produced causing vascular dysfunction and structural damages (9,10). Previous studies indeed have shown that O . suppresses K ATP channels and blunts the pial arterial dilation responses (11). In diabetic patients, in whom oxidative stress is evident, K ATP channel function is disrupted, leading to impaired vasodilation responses (12). In insulin-resistant rats, the K ATP channel-dependent vasodilation is also impaired, which is likely to be mediated by ROS (13). Although the dysfunction of K ATP channels in oxidative stress has been documented, the mechanism underlying channel modulation remains unknown (9,14). ROS can modulate proteins by intra-and intermolecular thiol oxidation (15,16), which may be the underlying cause for the modulation of vascular K ATP channels in oxidative stress. To test this hypothesis, we performed studies using a combined molecular biology, electrophysiology, and biochemistry approach. Our results showed that the Kir6.1/SUR2B channel, the major isoform of vascular K ATP channels, was inhibited by micromolar concentrations of H 2 O 2 as well as several other oxidants via S-glutathionylation.

MATERIALS AND METHODS
Chemicals and Reagents-Unless stated, all reagents and chemicals used in this study were purchased from Sigma. H 2 O 2 and glutathione (GSH) were freshly made and used within 4 h. Other reagents were prepared as concentrated stocks in double-distilled water or dimethyl sulfoxide (DMSO). In cases where DMSO was used, the final concentration of DMSO in solution was Ͻ0.1% (v/v). At this concentration, DMSO did not have any detectable effect on the channel activity.
Mesenteric Artery Preparation and Tension Measurements-All animal experiments were performed in compliance with an approved protocol by the Institutional Animal Care and Use Committees (IACUC) at Georgia State University. Male Sprague-Dawley rats (200 -250 g body weight) were deeply anesthetized and sacrificed. Mesenteric arteries were dissected, and connective tissues were removed in physiological saline solution (PSS) containing the following ( (Invitrogen Inc., Carlsbad, CA). To facilitate the identification of positively transfected cells, 0.4 g of green fluorescent protein (GFP) cDNA (pEGFP-N2; Clontech, Palo Alto, CA) was included in the cDNA mixture. One day after transfection, cells were disassociated with 0.25% trypsin, split, and transferred to coverslips for further growth. Experiments were performed on the cells in coverslips during the following 12-48 h.
Electrophysiology-Patch-clamp experiments were carried out at room temperature as described previously (3-5, 17, 18). In brief, fire-polished patch pipettes with 2-5 M⍀ resistance were made of 1.2-mm borosilicate glass capillaries. Whole-cell currents were recorded in single-cell voltage clamp with a holding potential of 0 mV and step to Ϫ80 mV. The bath solution contained the following (concentration in mM): KCl, 10; potassium gluconate, 135; EGTA, 5; glucose, 5; and HEPES, 10 (pH 7.4). The pipette was filled with a solution containing the following (concentration in mM): KCl, 10; potassium gluconate, 133; EGTA, 5; glucose, 5; K 2 ATP, 1; NaADP, 0.5; MgCl 2 ,1; and HEPES, 10 (pH ϭ 7.4). To avoid nucleotide degradation, all intracellular solutions were freshly made and used within 4 h. All the recordings were made with the Axopatch 200B amplifier (Axon Instruments Inc., Foster City, CA). The data were lowpass filtered (2 kHz, Bessel 4-pole filter, Ϫ3 dB) and digitized (10 kHz, 16-bit resolution) with Clampex 9 (Axon Instruments Inc.). Macroscopic currents were recorded from giant insideout patches, and single-channel currents were recorded from small inside-out patches with a constant single voltage of Ϫ80 or Ϫ60 mV. Symmetric high K ϩ (145 mM in total) was used in both bath and pipette solutions and K 2 ATP (1 mM), and NaADP (0.5 mM), were included in the bath solution to maintain the channel activity. Higher sampling rate (20 kHz) was used to digitize the currents recorded from inside-out patch. Data were analyzed using Clampfit 9 (Axon Instruments Inc.).
Immunochemistry-Immunochemistry was performed on the HEK293 cells with and without Kir6.1/SUR2B transfection. Two hours before the experiments, the culture medium was replaced with fresh medium. Biotinylated glutathione ethyl ester (BioGEE; 250 M; Invitrogen) was added to the medium and incubated for 1 h followed by an H 2 O 2 (750 M) challenge for 15 min. The medium was then discarded, and the cells were washed three times with phosphate-buffered saline (PBS) (containing 0.3% Triton X-100) to remove the excessive free Bio-GEE that had not conjugated with the proteins. Cells were then fixed with 4% paraformaldehdyde for 30 min followed by three washes with PBS. Dylight-488-conjugated streptavidin was diluted 1:1000 in PBS and added to the cells for 1 h of incubation at room temperature. After three PBS washes, the cells were examined under the LSM 510 confocal microscope (Zeiss). For double staining, the cells were further incubated with rabbit primary antibody against Kir6.1 for 2 h followed by three washes. Dylight-594-conjugated goat anti-rabbit secondary antibody (1:1000; Jackson ImmunoResearch) was used to visualize the Kir6.1 staining. Experiments were repeated three times.
Streptavidin Pull-down Assay and Western Blot-HEK293 cells expressing Kir6.1/SUR2B channels and the A10 smooth muscle cell line were used for this experiment. A10 cells were cultured in DMEM with 10% fetal bovine serum (FBS) at 37°C in humidified atmosphere with 5% CO 2 . BioGEE and H 2 O 2 treatments were performed as described above in Immunochemistry. The cells were then washed once and lysed using RIPA buffer (Sigma). Samples were run on 10% SDS-polyacrylamide non-reducing gel and then transferred to a nitrocellulose membrane (Bio-Rad). Rabbit primary antibodies against Kir6.1 (1:500; Sigma) and secondary antibodies conjugated with alkaline phosphatase were used in the Western blot (1:10,000; Jackson ImmunoResearch). Signals were visualized by SuperSignal West Pico Chemiluminescent Substrate (Thermo Scientific).
Biotin-GSH-conjugated proteins were pulled down using streptavidin-dynabeads according to the instructions provided by Invitrogen. Briefly, the beads were washed three times before the immobilization. Samples were then mixed with beads and incubated at room temperature with gentle rotation for 30 min. A magnet was used to separate the biotinylated molecules-bead complex from other unlabeled proteins. Supernatant containing unlabeled proteins was discarded, and the pellet was resuspended followed by three washes. The biotinylated moleculesbeads complex was then resuspended in the loading buffer with 0.1% SDS and boiled so that glutathionylated proteins were released into the solvent for further analysis. Experiments on HEK293 cells transfected with the Kir6.1/SUR2B channel were repeated four times, and experiments on A10 cells were repeated five times.
Data Analysis-Data were presented as means Ϯ S.E. Differences were evaluated using Student's t-tests or ANOVA, and statistical significance was accepted when p Ͻ 0.05.

H 2 O 2 Impaired Pinacidil-induced Vasodilation in Isolated
Mesenteric Rings-To define conditions whereby oxidative stress disrupts vascular K ATP channel function, we examined the vasodilating effects of the K ATP channel opener in the presence and absence of H 2 O 2 . Experiments were conducted on endothelium-intact (EI) and endothelium-denuded (ED) mesenteric rings. Vasoconstriction was first produced with 30 mM K ϩ . This was followed by treatments with increasing concentrations of pinacidil, a specific K ATP channel opener and a strong vasodilator (Fig. 1A). The vascular tones were measured with a force-electricity transducer. A pretreatment of the rings with H 2 O 2 (300, 600 M) impaired the pinacidil-induced vasodilation in both ED and EI rings. In ED rings, the IC 50 concentration of pinacidil for vasorelaxation was raised by 5-18-fold with the 300 M and 600 M H 2 O 2 treatments, respectively ( Fig.  1, A and B). Similar results were obtained in EI rings (Fig. 1C). Taken together, our data indicate that the function of vascular K ATP channels is disrupted in oxidative stress.
H 2 O 2 Inhibited Kir6.1/SUR2B Channel Activity in the Presence of GSH-The Kir6.1/SUR2B channel is the major isoform of vascular K ATP channels (6,19,20). Thus, we studied its modulation by expressing the Kir6.1/SUR2B channel in HEK293 cells. In wholecell voltage clamp, the baseline Kir6.1/SUR2B currents were small, and no obvious effect on the currents was observed when H 2 O 2 was applied. After the currents were activated by pinacidil, however, the Kir6.1/SUR2B channel was dose-dependently inhibited by H 2 O 2 with an IC 50 of 1.53 mM ( Fig. 2A, E, and F).
The results of whole-cell recordings may be affected by washout or inadequate controls of cytosolic soluble factors that can be potentially involved in the channel modulation, such as endogenous GSH and GSSG. Therefore, further studies were performed in excised patches. In giant inside-out patches, millimolar concentrations of H 2 O 2 were required to inhibit the Kir6.1/ SUR2B channel in the absence of cytosolic soluble components (Fig.  2, B, E, and F). Strikingly, the administration of small amount of glutathione (GSH) drastically enhanced the channel sensitivity to H 2 O 2 . In the presence of 100 M GSH, H 2 O 2 as low as 10 M began to inhibit channel activity, and clear concentration dependence was seen with an IC 50 of 25 M (Fig. 2, C, E, and F). The effect of H 2 O 2 was subsequently studied in the presence of 2 mM GSH and 40 M GSSG, a ratio that is close to the physiological concentrations of GSH/GSSG in the cytosol (21). Under this condition, H 2 O 2 also potently inhibited the channel with an IC 50 of 20 M (Fig. 2, D-F). Consistent with these observations in inside-out patches, a supplement of GSH/GSSG (2 mM and 40 M, respectively) to the pipette solution significantly enhanced the channel sensitivity to H 2 O 2 in whole-cell recordings. With GSH/GSSG in the pipette solution, 100 M H 2 O 2 inhibited the whole-cell currents by 36.2 Ϯ 8.8% (n ϭ 5) compared with 8.0 Ϯ 5.1% when GSH/ GSSG were absent (n ϭ 4, p Ͻ 0.05; Fig. 2G). In contrast, the same concentration of H 2 O 2 (100 M) inhibited the channel by Ͼ80% in inside-out patches in the presence of GSH or GSH/ GSSG (Fig. 2G). The difference in the H 2 O 2 sensitivity between inside-out patches and whole-cell recordings may be due to the diffusion kinetics across plasma membranes. Using the pseudofirst-order reaction analysis described by Tang et al. (22) in their study on H 2 O 2 -mediated BK channel inhibition, we calculated the pseudo first order constant to be 724 M Ϫ1 min Ϫ1 based on the average Kir6.1/SUR2B channel inhibition (36.2%) by 100 M H 2 O 2 (with GSH/GSSG) during a period of Ͻ5 min. With this constant, our further calculation showed that a 50% inhibition of the channel was achieved by ϳ23 M H 2 O 2 in ϳ30 min, which implies a protective role of the membrane barrier against a burst of H 2 O 2 (see discussion for details). The requirement of GSH for H 2 O 2 to produce its channel inhibition effect indicates that GSH-mediated protein modifications of the Kir6.1/SUR2B channel, such as S-glutathionylation, are likely to occur when H 2 O 2 is produced as an intermediary metabolite or a product of oxidative stress.
To understand the biophysical mechanisms underlying Kir6.1/SUR2B modulation, we analyzed single channel proper-  DECEMBER 3, 2010 • VOLUME 285 • NUMBER 49 ties recorded in regular inside-out patches in the presence of 2 mM GSH and 40 M GSSG. We found that the channel openstate probability was progressively suppressed with increased H 2 O 2 concentrations while the unitary conductance remained unchanged (Fig. 2D).

Regulation of K ATP Channel by S-Glutathionylation
GSH-dependent Modulation by Other Oxidants-Diamide, an oxidant that produces intra-and intermolecular disulfide bonds, is known to cause S-glutathionylation in the presence of GSH (23,24). In giant inside-out patches, we found that the joint application of diamide and GSH drastically inhibited channel activity within 2-8 min, regardless of the order of application, i.e. 78.4 Ϯ 5.1% inhibition with GSH first and 87.9 Ϯ 8.7% inhibition with diamide first (Fig. 3, A, B, and E). No statistical difference was found between the experiments (p Ͼ 0.05, n ϭ 12). In contrast, neither GSH (Fig. 3, C and E) nor diamide alone (Fig. 3, D and E) resulted in such channel inhibition over an extended time period (8 -14 min), suggesting that the channel inhibition is unlikely to be a result of the formation of disulfide bonds within a protein or between proteins.
Biochemical Evidence for the Kir6.1 S-Glutathionylation-Does S-glutathionylation occur on the channel protein or on another protein that regulates the channel activity? To address this question, we examined S-glutathionylation using membrane-permeable BioGEE. HEK293 cells transfected with Kir6.1/SUR2B or the expression vector alone were loaded with BioGEE (250 M) for 1 h followed by an H 2 O 2 (750 M) challenge for 15 min. Clear labeling was observed in the Kir6.1/SUR2Btransfected cells (Fig. 6, A-C), whereas a rather weak stain was seen with the mock transfection (data not shown). When the cells were double-stained with BioGEE (green) and Kir6.1 antibodies (red), co-localization was observed (Fig. 6, D-F).
The Kir6.1 S-glutathionylation was further tested with the streptavidin pull-down assay (26). The A10 vascular smooth muscle cell line, in which the Kir6.1/SUR2B channel was endogenously expressed (17,27), was loaded with BioGEE (250 M) for 1 h followed by an H 2 O 2 (750 M) challenge for 15 min. A strong Kir6.1-reactive band (ϳ32 kDa) was detected in the whole-cell lysate (Fig. 6G, lower panel). After pull-down with streptavidin, the cell lysate pretreated with a combination of BioGEE and H 2 O 2 showed a clear band of Kir6.1 immunoreactivity (Fig. 6G, upper panel), whereas no band was observed in cell lysates treated with either BioGEE or H 2 O 2 alone (Fig. 6G). Similar results were obtained using HEK293 cells transfected with Kir6.1/SUR2B (supplemental Fig. S1).
Oxidant-mediated Blockade of the K ATP Channel Activation by Natural Activators-Kir6.1/SUR2B channel activity is low under basal condition but increases significantly in the presence of several vasodilating hormones and neurotransmitters that are coupled to the adenylyl cyclase-cAMP-PKA pathway (3,18). Therefore, we examined the effect of S-glutathionylation on the Kir6.1/SUR2B currents activated by the vasoactive intestinal polypeptide (VIP; 100 nM) and ␤-adrenoreceptor agonist isoproterenol (Isop; 100 nM), both of which activate Kir6.1/SUR2B currents through the PKA pathway. In the whole-cell configuration, DTNP (50 M) or 2-DTP (50 M) FIGURE 3. The effect of diamide/GSH on the channel activity. Currents were studied in giant inside-out patches obtained from HEK293 cells expressing the Kir6.1/SUR2B channel with a holding potential of Ϫ60 mV. A, after channel activation by pinacidil, application of GSH (100 M) followed by DIA (100 M) for additional ϳ4 min strongly inhibited the channel activity. The inhibition was not reversed by high concentration of pinacidil (100 M). B, in another cell, the application of DIA followed by GSH also markedly inhibited the Kir6.1/SUR2B channel activity. C, a prolonged GSH (100 M) treatment did not have detectable effects on the Pin-activated currents. D, a prolonged treatment (8 min) with DIA (100 M) did not cause marked channel inhibition. E, summary of the effects of Pin, GSH alone, DIA alone, GSH followed by DIA, and DIA followed by GSH.

DISCUSSION
Inflammatory oxidative stress is a common pathogenesis of cardiovascular diseases, including hypertension, atherosclerosis, and diabetic vascular complications, which mostly result from the overproduction of ROS overwhelming the capacity of cellular antioxidant defense systems (7,8). When excessively produced, ROS can cause damages to lipids, proteins, and nucleotides, leading to cell dysfunction, structural injuries, and death (7). As the universal antioxidant treatment did not yield promising results in clinic trials (28), the identification of specific molecules and the understanding of the molecular mechanisms underlying the oxidant-mediated protein modulation becomes crucial for the development of novel therapeutic strategies. Our studies indicate that the Kir6.1/SUR2B channel is inhibited by micromolar concentrations of H 2 O 2 through S-glutathionylation. Such channel inhibition is of pathophysiological relevance and is likely to occur in vasculatures as the production of micromolar concentrations of H 2 O 2 has been shown during oxidative stress (10).
H 2 O 2 , O . , OH⅐, and peroxyl (RO . ) are the major oxidants produced endogenously in biological systems. Their concentrations in the cytoplasm are tightly controlled by several antioxidant systems (29). Early studies have shown that ROS including H 2 O 2 have modulatory effects on membrane proteins (22,30,31). Most of the studies, however, use millimolar concentrations of H 2 O 2 (32,33). In the present study, we have observed a clear inhibition of the Kir6.1/SUR2B channel by H 2 O 2 with IC 50 of ϳ1.5 mM in whole-cell recording. In contrast, the inhibition of K ATP currents can be clearly seen with micromolar concentrations of H 2 O 2 in vascular rings. The major reason for the discrepancy is likely to be the effect of washout or inadequate controls of the cytosolic soluble factors that play a role in the channel modulation in oxidative stress. Indeed, some of the cytosolic substances are identified to be GSH and GSSG in the present study.
Searching for the missing cytosolic factors to better characterize the effect of H 2 O 2 , we conducted experiments in giant inside-out patches. We have found that supplying a small amount of exogenous GSH dramatically reduces the concentration of H 2 O 2 needed for channel inhibition. Using 2 mM GSH and 40 M GSSG, which mimic the intracellular GSH/GSSG ratio (21,34), we have also observed clear channel inhibition by micromolar concentration of H 2 O 2 . The drastic effect of GSH/ GSSG on the channel sensitivity to H 2 O 2 is not only seen in inside-out patches, but also takes place in the whole-cell recordings. A supplement of GSH/GSSG to the pipette solution  Comparing the inside-out patch data with the whole-cell recordings, we have found that the lower H 2 O 2 sensitivity in the whole-cell recordings appears to be also related to the transmembrane diffusion kinetics. The limited capacity of transmembrane diffusion may act as a cellular protection mechanism, diminishing or even eliminating the effect of a burst of H 2 O 2 production in the interstitial fluid. In the vascular inflammation state, however, such a membrane barrier does not seem adequate to protect the cell against a long-lasting production of H 2 O 2 . Consequently, H 2 O 2 manages to pass through the plasma membranes and inhibits the channel from inside with a potency similar to that seen in inside-out patches.
Such channel inhibition is not limited to H 2 O 2 as diamide, another oxidant, also produces strong channel inhibition when applied together with GSH. Furthermore, channel inhibition can be produced by the general S-glutathionylation-inducer GSSG. These results thus indicate that S-glutathionylation is likely to be the underlying cause for Kir6.1/SUR2B channel inhibition by these oxidants under pathophysiological conditions.
Hence, several lines of evidence shown in the present study support S-glutathionylation of the Kir6.1/SUR2B channel. (i) The channel is inhibited by H 2 O 2 or diamide potently only when there is GSH. (ii) The S-glutathionylation inducer GSSG causes the channel inhibition. (iii) Several 2-PDSs reactive with thiol groups to form adaptors at cysteine residues inhibit the Kir6.1/SUR2B channel in micromolar concentrations. (iv) The oxidant-mediated channel inhibition can be reversed by the specific deglutathionylation reagent Grx1 and the general reducing reagent DTT. (v) Streptavidin pull-down assays reveal the incorporation of the GSH moiety to the Kir6.1 subunit in the presence but not in the absence of H 2 O 2 . This K ATP channel modulation mechanism is not limited to the pinacidil-induced currents. The K ATP currents activated by the natural vasodilators VIP and Isop are similarly inhibited by these oxidants.
The modulation of protein activity by S-glutathionylation is a newly recognized post-translational regulatory mechanism (35,36). This process, facilitated by oxidative stress and also seen in unstressed cells, can result in major changes to protein conformations and functions (16). Such modulation has been demonstrated in a large number of proteins using microarray (37) or proteomic analysis (38), while the physiological relevance of this modulation remains to be understood (36). Several recent studies indicate that the following membrane proteins are modulated by S-glutathionylation: the cystic fibrosis transmembrane conductance regulator (24), the ryanodine receptor (39), the sarco/endoplasmic reticulum Ca 2ϩ ATPase (40), and the Na ϩ -K ϩ pump (41). The present study has shown for the first time that an important vascular tone regulator, Kir6.1/SUR2B channel, is subject to S-glutathionylation produced by H 2 O 2 at physiological or pathophysiological concentrations. This channel modulation by H 2 O 2 can lead to impairment of the vasodilation responses in a variety of vascular complications. Therefore, the demonstration of S-glutathionylation as a regulatory mechanism has important clinical implications. With the information, new therapeutic strategies may be formulated by preventing the oxidative modulation of the Kir6.1/SUR2B channel.
Another similar post-translational regulation mechanism is S-nitrosylation through which a nitric oxide (NO) moiety is incorporated into the thiol group of a cysteine residue (25,42). However, our initial attempts using NO donors yielded inconsistent results (data not shown).
In conclusion, our studies indicate that the Kir6.1/SUR2B channel is inhibited by H 2 O 2 at micromolar concentrations owing to S-glutathionylation. The demonstration of the mechanism underlying the impairment of the vascular K ATP channel should have impacts on the treatment and prevention of several vascular diseases caused by oxidative stress.