Nitration and Functional Loss of Voltage-Gated K⁺ Channels in Rat Coronary Microvessels Exposed to High Glucose

Seven-week-old male Sprague-Dawley rats (Harlan, Madison, WI) were anesthetized with sodium pentobarbital (60 mg/kg i.p.). RSCAs (internal diameter 150–200 μm) were dissected from the left ventricle. Some RSCAs were incubated in culture media supplemented with either 5.5 mmol/l d-glucose (normal glucose [NG]) or 23 mmol/l d-glucose (high glucose [HG]), or 5.5 mmol/l d-glucose plus 17.5 mmol/l l-glucose (LG) for osmotic control as described previously (11). All rats were housed in the Association for Assessment and Accreditation of Laboratory Animal Care–approved Biomedical Resource Center at the Medical College of Wisconsin, and all protocols were approved by the Animal Care Committee.

Authentic ONOO⁻ was synthesized according to published methods (5,6). The amount of ONOO⁻ in the stock solution was determined spectrophotometrically using the reported extinction coefficient for ONOO⁻ (1,670 mol · l⁻¹ · cm⁻¹). Before each application, an aliquot of the stock solution was diluted in 1 mmol/l NaOH and rapidly added to the vessel chamber to achieve a final concentration of 5 μmol/l. Decomposed ONOO⁻ (DC-ONOO⁻) was made by leaving ONOO⁻ at room temperature for at least 2 h. The decay of ONOO⁻ was confirmed spectrophotometrically.

RSCAs were cannulated on glass micropipettes in an organ chamber filled with physiological salt solution as described previously (11). The physiological salt solution was warmed to 37°C, continuously circulated, and bubbled with 21% O₂, 5% CO₂, and 74% N₂. Most vessels that were equilibrated for 1 h at an intraluminal pressure of 60 mmHg developed spontaneous tone averaging 70% of the passive diameter, which was defined as the maximal diameter assessed in Ca²⁺-free solution at the end of the experiment. RSCAs that did not develop this degree of spontaneous tone were constricted by U46619 (10 nmol/l) to 70% of the passive diameter.

Enzymatic isolation of single VSMCs was performed according to published methods (12). Whole-cell patch-clamp recordings were obtained using standard pulse protocols and instrumentation (5). Briefly, families of K⁺ currents were generated by stepwise 10-mV depolarizing pulses (400-ms duration, 5-s intervals) from a holding potential of −60 mV to 50 mV in cells that were dialyzed with 10 nmol/l calcium. Trials were performed in triplicate, and peak current amplitudes were divided by membrane capacitance to obtain K⁺ current density (12).

Unitary K_v currents were obtained in inside-out membrane patches of VSMCs from RSCAs, which were bathed in symmetrical 145 mmol/l K⁺ solutions. For eliminating interfering currents from K_Ca and ATP-sensitive K⁺ channels, 100 nmol/l iberiotoxin was added to the pipette solution and 1 μmol/l apamin and glibenclamide was added to the bath solution. The effect of authentic ONOO⁻ on unitary K_v currents was evaluated by 2-min recording intervals before and after drug application. Averaged current amplitudes were obtained at stepwise patch potentials for calculation of single-channel conductance.

For determining the effect of HG on nitrotyrosine formation, RSCAs were exposed to NG or HG for 24 h. In some arteries, urate (100 μmol/l) was applied every 8 h to scavenge ONOO⁻. The joint effect of the O₂˙⁻ scavenger manganese [III] tetrakis 4-benzoic acid porphyrin (MnTBAP; 300 μmol/l) and the NO synthase inhibitor N^ω-nitro-l-arginine methyl ester (l-NAME; 100 μmol/l) on nitrotyrosine formation was examined in other arteries. Freshly isolated RSCAs were used as controls. All RSCAs were prepared as frozen sections (5 μm thick). The immunodetection of nitrotyrosine residues has been described (5,6).

For obtaining adequate vascular tissue for immunoanalysis, VSMCs were enzymatically dissociated from RSCAs and cultured in RPMI-1640 medium that contained 20% FBS, 100 units/ml penicillin G, and 100 μg/ml streptomycin. The first passage of VSMCs was divided into six plates and permitted to reach 80% confluence. In other plates, the purity of VSMCs was verified by immunostaining with α-smooth muscle actin. Subsequently, cells were treated for 24 h with DMEM that contained NG, HG, or LG. Urate (100 μmol/l) was added to some HG plates to scavenge ONOO⁻. After incubation, total or membrane proteins were collected for immunoprecipitation and Western blot analysis, respectively (13).

Proteins (100–200 μg/reaction) were immunoprecipitated by incubations at 4°C with antinitrotyrosine (4 μl/reaction) for 2 h, and then incubated in protein A for 1 h. The immunoprecipitates were centrifuged and washed three times with ice-cold lysis buffer, and Western blots were performed using antibodies directed against K_v1.2 and K_v1.5 channel α-subunits (Upstate Biotechnology, Lake Placid, NY) (13). Anti-K_v1.2 and anti-K_v1.5 were also used to compare K_v1.2 and K_v1.5 expression at different glucose levels. β-Actin was used as an internal standard to ensure equal lane loading.

All chemicals were purchased from Sigma. Correolide was provided by Drs. Maria Garcia and Gregory Kaczorowski (Merck Research Laboratories, Rahway, NJ). Correolide was prepared as a 20-mmol/l stock in DMSO and stored at 4°C.

All data are expressed as mean ± SE. Percentage of constriction was defined as the percentage of reduction from control internal diameter. Data from videomicroscopy, patch-clamp, and Western blot studies were compared using one-way ANOVA, with repeated measures for dose and condition. Differences between individual means were determined by Newman-Keuls test. All differences were judged to be significant at P < 0.05.

Initial experiments were designed to determine whether ONOO⁻ inhibited K_v channel function in RSCAs. Graded concentrations of 4-aminopyridine (4-AP), a broad-spectrum K_v channel blocker, constricted RSCAs in control conditions (Fig. 1; n = 5). These findings suggested that K⁺ efflux through K_v channels contributed significantly to resting diameter. Notably, decomposed ONOO⁻ (DC-ONOO⁻) had no effect on the constriction of RSCAs to 4-AP (Fig. 1A). However, 4-AP–induced contractions were attenuated by incubation of RSCAs in authentic ONOO⁻ (5 μmol/l), indicating a loss of K_v channel dilator function (Fig. 1B). Figure 2A shows similar U46619-induced constriction (10⁻⁹ to 10⁻⁷ mol/l) in RSCAs that were exposed to DC-ONOO⁻ or ONOO⁻. Arteries that were incubated in DC-ONOO⁻ or ONOO⁻ showed similar dilation to the l-type Ca²⁺ channel antagonist nifedipine (10⁻⁸ to 10⁻⁶ mol/l; Fig. 2B). These results indicate that ONOO⁻ selectively impairs vasoactive responses that are dependent on functional K_v channels.

K_v1 “Shaker-type” gene family members are redox targets in some VSMCs (8). To determine whether ONOO⁻ specifically impairs K_v1 channels in RSCAs, we used the specific K_v1 channel blocker correolide (COR) (13). DC-ONOO⁻ did not significantly affect K_v current in freshly isolated VSMCs, but COR (1 μmol/l) blocked a large component, indicating the presence of K_v1 family channels (Fig. 3A). In contrast, authentic ONOO⁻ (5 μmol/l) reduced total K_v current amplitude, and COR-sensitive K_v1 current was not detected after exposure to ONOO⁻ (Fig. 3B). Current-voltage relations averaged from five cells verified that DC-ONOO⁻ did not significantly alter K_v current density in coronary VSMCs (Fig. 3C), whereas authentic ONOO⁻ fully eliminated the COR-sensitive K_v1 current component (Fig. 3D).

Subsequent studies determined whether ONOO⁻ directly inhibited K_v channels in cell-free, inside-out patches of coronary VSMCs. A prominent K_v channel that progressively activated in response to stepwise patch depolarization was detected (Fig. 4A) and showed a unitary conductance of ∼75 pS (Fig. 4B). Exposing the cytosolic patch surface to ONOO⁻ (5 μmol/l) profoundly inhibited K_v channel activity, whereas the unitary amplitude was unaffected (Fig. 4C). In five different cells, ONOO⁻ reduced the open-state probability of single K_v channels by 40 ± 8% (Fig. 4D).

On the basis of initial findings indicating that authentic ONOO⁻ impaired K_v1 channel function, new studies investigated whether HG promotes ONOO⁻ formation in RSCAs. Nitrotyrosine, a footprint of ONOO⁻, was used to assess ONOO⁻ production in RSCAs that were exposed to NG or HG (5). Sample images from five representative experiments indicate that staining was more prominent in RSCAs that were incubated in HG media (Figs. 5A and B). Nitrotyrosine levels were reduced after treatment with 100 μmol/l urate, a scavenger of ONOO⁻ (Figs. 5C and D). The nitrotyrosine signal also was decreased after dual incubation with MnTBAP (300 μmol/l) and l-NAME (100 μmol/l), drugs that scavenge O₂˙⁻ and inhibit NO synthesis, respectively (Figs. 5E and F). A low level of nitrotyrosine reactivity was observed in freshly isolated RSCAs (Fig. 5G), and elimination of the primary antinitrotyrosine antibody resulted in signal loss (Fig. 5H). These studies suggest that HG increases tyrosine nitration via endogenous ONOO⁻ that is formed from the reaction of O₂˙⁻ with NO.

The K_v1 channels in VSMCs represent tetrameric structures composed of K_v1.2 and K_v1.5 α-subunits (13,14). Western blots indicated that the expression of K_v1.2 and K_v1.5 α-subunits was similar between RSCAs that were incubated for 24 h in NG, LG, or HG media (Figs. 6A and B). Averaged densitometric values from five studies confirmed these findings in which the expression of β-actin, a 42-kDa structural protein, provided an internal standard (Fig. 6C). However, nitrotyrosine immunoprecipitates revealed an elevated nitration of K_v1.2 α-subunits in RSCAs that were incubated in HG for 24 h, a response that was partially normalized by urate (Fig. 6D). In contrast, HG did not promote nitration of K_v1.5 α-subunits, and urate had little effect on basal nitration of this protein (Fig. 6E). Averaged data from five samples of immunoprecipitates indicated that HG enhanced nitration of K_v1.2 but not K_v1.5 α-subunits (Fig. 6F). Increased nitration of K_v1.2 α-subunits was also observed in K_v1.2 immunoprecipitates that were probed with nitrotyrosine antibody (data not shown).

A final set of experiments investigated whether incubation in HG for 24 h impairs K_v channel dilator function via generation of ONOO⁻. Compared with RSCAs that were exposed to NG levels (Fig. 7A), those that were exposed to HG constricted less to 4-AP (Fig. 7B), indicating a reduced dilator influence of K_v channels. Scavenging of ONOO⁻ by urate (100 μmol/l) restored 30% constriction to 4-AP in HG but not NG arteries. Ebselen (100 μmol/l), a chemically distinct ONOO⁻ scavenger, also partially (25%) and selectively restored the impaired constrictor response to 4-AP observed in RSCAs that were incubated in HG (Figs. 7C and D).

The present study provides several new findings. First, our data demonstrate that authentic ONOO⁻ attenuates the dilator function of K_v channels in RSCAs and reduces macroscopic and single-channel K_v1 current. Second, short-term exposure of RSCAs to HG promotes ONOO⁻ formation, an event that is associated with the nitration of K_v1.2 but not K_v1.5 pore-forming α-subunits. Third, RSCAs that are incubated in HG demonstrate impaired dilator function of K_v channels, and this defect is partially restored by ONOO⁻ scavengers. Overall, these findings suggest potentially important roles for ONOO⁻ production directly or through elevations in ambient glucose in the control of coronary vasomotor tone.

The formation of ONOO⁻, driven by the interaction of O₂˙⁻ and NO, has been proposed as a contributing factor in the pathogenesis of vasodilator dysfunction (5). Recently, our laboratory reported that vasodilator responses that are dependent on cAMP activation of K_v channels were attenuated in RSCAs that were exposed to HG for 24 h (7). The findings of the present study directly demonstrate that authentic ONOO⁻ reduces K_v channel dilator function in cannulated and pressurized RSCAs. In addition, patch-clamp studies revealed that ONOO⁻ blocks a COR-sensitive component of K_v1 current in single VSMCs from RSCAs, suggesting that K_v1 family channels are susceptible to ONOO⁻ inhibition. Finally, our data demonstrate that the open-state probability of K_v channels in cytosol-free membrane patches of RSCAs is reduced in response to authentic ONOO⁻, suggesting that ONOO⁻ may directly inhibit K_v1 channels in VSMC membranes independent of intracellular signaling mechanisms. Thus, it seems that ONOO⁻ may powerfully inhibit coronary K_v1 channel activity.

In this study, urate effectively prevented formation of ONOO⁻ in RSCAs that were exposed to HG (Fig. 5) but partially restored the impaired 4-AP constriction in HG vessels. This may be due to a higher threshold for detection of ONOO⁻ by immunohistochemistry than for its inhibition of K_v channels, which may occur at concentrations not detected by immunohistochemistry. Alternatively, it may be that factors other than ONOO⁻ contribute to impaired K_v channel function during hyperglycemia. We previously demonstrated that superoxide dismutase, a O₂˙⁻ scavenger, partially restored impaired K_v channel activation in RSCAs that were exposed to HG, indicating an inhibitory role of O₂˙⁻ on K_v channels (11). The relative contribution of O₂˙⁻ and ONOO⁻ in the reduction of K_v current in HG is not known. On the basis of our previous and present studies, we believe that both O₂˙⁻ and ONOO⁻ contribute to the impaired K_v channel function during hyperglycemia.

The present study also provides initial evidence that HG promotes ONOO⁻ formation in RSCAs, an event associated with nitration of tyrosine residues in VSMCs. In support of this hypothesis, scavenging of ONOO⁻ by urate reduced the immunoreactivity correlating to nitrotyrosine residues in RSCAs that were exposed to HG for 24 h. Similarly, the use of MnTBAP and l-NAME jointly to reduce the availability of O₂˙⁻ and NO to form ONOO⁻-attenuated nitrotyrosine formation. Subsequent findings revealed that elevation of ONOO⁻ by HG selectively increased the number of nitrotyrosine residues in K_v1.2 but not K_v1.5 α-subunits, an event that was partially reversed by the ONOO⁻ scavenger urate. These findings suggest that ONOO⁻ may specifically nitrate particular proteins that compose the K_v1 channel pore, providing initial insight into the molecular basis of ONOO⁻-induced inhibition of coronary K_v channels. Notably, short-term exposure to HG has been linked to an increased ONOO⁻ production in the endothelial cells of human aorta and rat retinal arteries (15–17), and ONOO⁻ may inhibit prostacyclin synthase in these endothelial cells (15). Thus, several lines of evidence suggest that the elevation of ONOO⁻ in response to HG may impair the function of at least several vasodilator mechanisms in the arterial wall, including K_v channels in VSMCs.

Several limitations of the present study should be acknowledged. First, our findings have not established a causal relationship between ONOO⁻-induced nitration of K_v1.2 α-subunits and impaired K_v1 channel function. Mutagenesis will be required to confirm the molecular basis of inhibition, which may involve single or multiple nitrations of the 17 tyrosine residues that reside in the K_v1.2 α-subunit. Second, although nitrotyrosine is widely used as a marker for ONOO⁻ formation, other mechanisms of tyrosine nitration exist, including H₂O₂-NO₂-hemeperoxidase and NO₂-mediated nitrosylation (18). However, because H₂O₂-NO₂-hemeperoxidase occurs primarily in inflammatory cells (19) and nitrotyrosine formation by NO₂ is a limited reaction (20), ONOO⁻ most likely was responsible for K_v channel nitration in our study. Finally, short-term exposure to HG may not produce the same complexity or degree of vascular dysfunction as in diabetes, where alternative mechanisms could mask or aggravate vasodilator dysfunction. However, the advantage of this study is the ability to directly examine K_v channel function in response to HG without confounding influences, such as circulating plasma constituents and neurohumoral factors. Future studies will examine the overall effect of ONOO⁻ in diabetes-induced K-channel–mediated vascular dysfunction.

K_v channels are a major contributor to the resting tone of small coronary arteries and represent a powerful dilator influence in the coronary circulation in vivo (for review, see 21). These channels participate in coronary vasodilator responses to acidosis, cAMP-dependent agonists such as β-adrenergic activation, and endothelial factors. Under these conditions, in which resting vascular tone relies on K_v channel function, the generation of ONOO⁻ during hyperglycemia impairs channel availability and may compromise coronary blood flow and myocardial perfusion.

This work was supported by P50 HL65203, P01 HL68769, and a VA Merit Grant (to D.D.G.); NIH R01 HL59238 (to N.J.R.); NIH RO1 HL067948; and a Scientist Development Grant from the American Heart Association (to Y.L.).