Rotenone-stimulated superoxide release from mitochondrial complex I acutely augments L-type Ca²⁺ current in A7r5 aortic smooth muscle cells

Rotenone induced superoxide release in mitochondria and acutely increased L-type Ca²⁺ channel currents (I_Ca,L) in A7r5 arterial smooth muscle cells. Superoxide is dismutated to hydrogen peroxide (H₂O₂), which increases I_Ca,L. However, the increase occurred in the presence of dithiothreitol or H₂O₂, indicating that superoxide per se caused an increase in I_Ca,L. The increase was suppressed by protein kinase inhibitor staurosporine.

Abstract

Voltage-gated L-type Ca²⁺ current (I_Ca,L) induces contraction of arterial smooth muscle cells (ASMCs), and I_Ca,L is increased by H₂O₂ in ASMCs. Superoxide released from the mitochondrial respiratory chain (MRC) is dismutated to H₂O₂. We studied whether superoxide per se acutely modulates I_Ca,L in ASMCs using cultured A7r5 cells derived from rat aorta. Rotenone is a toxin that inhibits complex I of the MRC and increases mitochondrial superoxide release. The superoxide content of mitochondria was estimated using mitochondrial-specific MitoSOX and HPLC methods, and was shown to be increased by a brief exposure to 10 μM rotenone. I_Ca,L was recorded with 5 mM BAPTA in the pipette solution. Rotenone administration (10 nM to 10 μM) resulted in a greater I_Ca,L increase in a dose-dependent manner to a maximum of 22.1% at 10 μM for 1 min, which gradually decreased to 9% after 5 min. The rotenone-induced I_Ca,L increase was associated with a shift in the current-voltage relationship (I-V) to a hyperpolarizing direction. DTT administration resulted in a 17.9% increase in I_Ca,L without a negative shift in I–V, and rotenone produced an additional increase with a shift. H₂O₂ (0.3 mM) inhibited I_Ca,L by 13%, and additional rotenone induced an increase with a negative shift. Sustained treatment with Tempol (4-hydroxy tempo) led to a significant I_Ca,L increase but it inhibited the rotenone-induced increase. Staurosporine, a broad-spectrum protein kinase inhibitor, partially inhibited I_Ca,L and completely suppressed the rotenone-induced increase. Superoxide released from mitochondria affected protein kinases and resulted in stronger I_Ca,L preceding its dismutation to H₂O₂. The removal of nitric oxide is a likely mechanism for the increase in I_Ca,L.

NEW & NOTEWORTHY

Rotenone induced superoxide release in mitochondria and acutely increased L-type Ca²⁺ channel currents (I_Ca,L) in A7r5 arterial smooth muscle cells. Superoxide is dismutated to hydrogen peroxide (H₂O₂), which increases I_Ca,L. However, the increase occurred in the presence of dithiothreitol or H₂O₂, indicating that superoxide per se caused an increase in I_Ca,L. The increase was suppressed by protein kinase inhibitor staurosporine.

influx of calcium ions through L-type Ca²⁺ channel currents (I_Ca,L) is the major source of higher [Ca²⁺ ]_i, which initiates contraction of arterial smooth muscle cells (ASMCs) and regulates blood pressure (30). I_Ca,L is modulated by neurotransmitters, hormones, and autacoids in the control of cardiovascular function (21, 24, 32). Regulation of hypoxic vasodilatation of systemic arteries and hypoxic vasoconstriction of the pulmonary artery are fundamental regulatory mechanisms to ensure the efficient supply of O₂ throughout the human body. Hypoxia in tissues is detected by a change in reactive oxygen species (ROS) (18). Superoxide and H₂O₂ are produced in tissues and regulate cellular function in various physiological and pathophysiological conditions (8, 14, 18, 22, 40, 42, 50). The main sources of superoxide are NADPH oxidase (NOX) localized in the plasma membrane and the mitochondrial respiratory chain (MRC) in the mitochondrial inner membrane (11, 14, 33, 51, 56).

ANG II binds to the angiotensin type-1 receptor and augments the release of superoxide by activating NOX in the vasculature (33). H₂O₂ dismutated from superoxide generated by ANG II-stimulated NOX and from superoxide produced by exogenous xanthine oxidase plus hypoxanthine augments I_Ca,L in rat cerebral ASMCs (5); this increase is mediated by PKCα activation (5, 12, 32). H₂O₂ also increases I_Ca,L via CaMKII-induced phosphorylation of Ca_V1.2 channels in rabbit (52) and rat ventricular myocytes (41). Activation of Ca_V1.2 channels regulates mitochondrial metabolism in guinea pig ventricular myocytes (46) via the actin cytoskeleton, which is connected to voltage-dependent anion channels (VDACs) in the mitochondrial outer membrane (45). In rat cerebral ASMCs, ROS generated from a subpopulation of mitochondria increases H₂O₂ in microdomains adjacent to Ca_V1.2 channels, thus increasing Ca_V1.2b channel opening (13).

Complex I, a proton-pumping NADH:ubiquinone oxidoreductase with a molecular mass of approximately 1 MDa (54), is the first complex of the MRC in the inner membrane. MRC releases a considerable amount of superoxide during normal oxidative metabolism (11, 56). Rotenone, a phytogenic toxin, binds to complex I and increases the release of superoxide into the mitochondrial matrix (44, 56). In bovine coronary ASMC mitochondria, rotenone stimulates the release of superoxide, detectable by MitoSOX fluorescence (16). Chronic hypoxia and rotenone treatment increases I_Ca,L in HEK 293 cells that express a human cardiac L-type Ca²⁺ channel α₁ subunit (9). Paraquat-generated superoxide acutely increases high-K⁺-induced I_Ca,l-dependent intracellular Ca²⁺ increase in rat renal afferent arterioles (47). However, it remains unclear how superoxide acutely regulates I_Ca,L, probably because modulation by superoxide partly overlaps with H₂O₂-induced modulation.

We studied acute superoxide-induced modulation of I_Ca,L, using rotenone as a stimulator of superoxide release from mitochondria, inhibitors of H₂O₂-mediated modulation, and H₂O₂ in cultured A7r5 ASMCs derived from embryonic rat thoracic aorta (25). Rotenone resulted in greater mitochondrial superoxide release, as estimated by the oxidation product of MitoSOX using HPLC, and an acute increase in I_Ca,L. The increase was associated with a characteristic shift of the current-voltage (I-V) relationship to the hyperpolarizing direction. Rotenone-induced increase in I_Ca,L was strongly inhibited by Tempol, a SOD mimetic (48), and by staurosporine, a broad-spectrum protein-kinase inhibitor (31, 36). The increase was not affected by DTT or H₂O₂, indicating that per se was responsible for the increase.

MATERIALS AND METHODS

Cell culture.

A7r5 smooth muscle cells were purchased from the American Type Culture Collection (Manassas, VA) and maintained under 5% CO₂ at 37°C in DMEM supplemented with L-glutamine, 4.5 g/l glucose, and 10% fetal bovine serum (Invitrogen, Carlsbad, CA). Cells at ∼70% confluence were subcultured weekly using 0.05% trypsin-EDTA (Gibco 25300-054; Thermo Fischer Scientific, Grand Island, NY) for up to 20 passages. A portion of detached cells from a culture of more than five passages was transferred to a tube containing normal Tyrode (NT) solution at room temperature for patch-clamp experiments and aliquoted into microtubes for pretreatment with Tempol or staurosporine.

Superoxide measurement using HPLC.

2-Hydroxy-Mito-ethidium (2-OH-Mito-ethidium), the superoxide-specific product of MitoSOX (Mito-2-hydroxyethidine, MitoSOX Red; Thermo Fischer Scientific) was measured using a previously-described HPLC method (3, 55). Increasing concentrations of 2-OH-Mito-ethidium were loaded onto the column to generate a standard curve. Confluent A7r5 cells in 6-well plates were treated with 5 μM MitoSOX for 1 h and rotenone (10 μM) for 3 min. Wells that were treated with MitoSOX but not with rotenone served as controls. Cells were washed with PBS, and 100% acetonitrile (HPLC grade; Sigma-Aldrich, St. Louis, MO) was added. Following this, cells were scraped from the 6-well plates and collected in 1-ml Eppendorf tubes, incubated at −20°C for 1 h, and centrifuged at 11,000 rpm for 10 min at 4°C. The resulting pellets were assayed for protein concentration using the Bradford method. 2-OH-Mito-ethidium in supernatant was quantified using an HPLC system with an FP-1520 fluorescence detector (JASCO Analytical Instruments, Easton, MD) and a Beckman Ultrasphere column (C18) (5 μm, 250 × 4.6 mm; Thermo Fischer Scientific). HPLC data were normalized by protein content.

Solutions and drugs.

NT solution contained the following (in mM): NaCl, 135; KCl, 5.4; CaCl₂, 1.8; MgCl₂, 1; HEPES, 5; and glucose, 5.5; pH was adjusted to 7.4 using NaOH. The Ba²⁺ solution contained the following (in mM): NaCl, 108; TEACl, 20; CsCl, 5.4; BaCl₂, 10; MgCl₂, 1; HEPES, 5; and glucose 5.5; pH was adjusted to 7.4 using NaOH. To obtain the background current for subtraction, 30 μM CdCl₂ in a Ba²⁺ solution was prepared. The pipette solution contained the following (in mM): Cs-aspartate, 115; TEACl, 20; MgCl₂, 1; BAPTA, 5; Mg·ATP, 3; GTP, 0.2; and HEPES, 10; pH was adjusted to pH 7.2 using CsOH. All chemicals used to make solutions, and rotenone, H₂O₂, DTT, Tempol, and staurosporine were from Sigma-Aldrich.

Patch-clamp.

A drop of cell suspension was placed in a chamber on a wide mechanical stage with XY control. Cells attached to the glass bottom were observed with an upright microscope (BX51; Olympus, Tokyo, Japan) using a ×10 objective and ×40 water-immersion objective with a long working distance. Solutions flowed continuously to the chamber by gravity with a flow rate of 2–4 ml/min during patch-clamp and were removed using an adapted U-tube. The first solution used was NT. After establishing patch-clamp, the NT solution was exchanged to the Ba²⁺ solution. Solution exchange was conducted by manually switching the inlet tube to another solution container (time 0 for the exchange, indicated with an arrow in figures). It took ∼30 s for the new solution to enter the chamber. Recording was performed near the entry point of the solution in the chamber. To minimize mechanical issues related to flow, the ×40 objective was used only to select the target cell, position the patch pipette, and examine cells at the end of the experiment. Patch-clamp was conducted under the ×10 objective. Patch pipettes were prepared from hard glass capillary tubing containing a glass filament using a micropipette puller (P-97; Sutter Instrument, Novato, CA). Pipette resistance was relatively high (15–25 MΩ) when filled with pipette solution. Following establishment of a giga-ohm seal by gentle suction, the membrane was ruptured using the Zap function of the amplifier. Junction potential was adjusted before establishment of the seal, and experimental drift was estimated after each experiment by removing the cell from the pipette in NT solution. Changes in junction potential were within ±2 mV. The voltage-clamp amplifier (Axopatch 200B; Molecular Devices, Sunnyvale, CA) was driven by Clampex 9 software via a digital interface (Digidata 1332A; Molecular Devices). Currents were filtered at 2 kHz using the amplifier's low-pass eight-pole Bessel filter. Membrane capacitance was estimated using the built-in program in Clampex. The average membrane capacitance was 66 pF. I_Ca,L for I-V was obtained by applying 500-ms depolarization steps in 10-mV increments at 0.2 Hz from −40 to 50 mV, starting from a holding potential (HP) of −80 mV. Test depolarization steps were preceded by a 50- or 70-ms prepulse to detect and inactivate the T-type Ca²⁺ channel current, which appeared infrequently and with relatively small amplitude. About 50 s were necessary to obtain one I-V and the series was repeated at approximately 2-min intervals. Following the application of rotenone and other agents, I-V was obtained starting after 1 min of perfusion. Quasi-steady-state inactivation curves (f_∞-V) were generated using a gapped double-pulse protocol at 0.1 Hz. Two seconds of prepulse to potentials between −100 and 30 mV in 10-mV increments from an HP of −80 mV were followed by a 50- or 70-ms step to −40 mV, then a 500-ms test pulse to 0 mV. In the presence of rotenone, f_∞-V was obtained following a 5-min perfusion. Time lapse change in I_Ca,L was obtained with repetitive application of 500-ms pulses to −15 mV (unless otherwise noted) preceded by a short prepulse at a rate of 1/20 s from a holding potential of −80 mV. I_Ca,L was quantified after subtracting the current obtained in the presence of 30 μM Cd²⁺. Current traces shown in all figures display subtracted currents in which capacitive transients have sometimes been truncated for illustrative purposes. The effect of Tempol was examined after pretreatment with 1 mM Tempol at room temperature for periods longer than 40 min. In several experiments, additional overnight exposure at 4°C was performed to equilibrate mitochondria with Tempol. Patch-clamp experiments were conducted at room temperature.

Data analysis.

I_Ca,L was analyzed using low-pass filtering with a cutoff frequency of 1 kHz with Clampfit v9 or v10 software (Molecular Devices). Statistical analysis was performed using Prism software (v6; GraphPad Software, San Diego, CA). Igor Pro (v6; Wavemetrics, Portland, OR) was used for curve fitting and illustration. Subtraction of Cd²⁺ resistant current was performed using the Table function of Igor Pro and Microsoft Excel. The I-V relationship of I_Ca,L was obtained by calculating means ± SE. I-V curves were obtained by fitting the data with the following equation, adapted from the Boltzmann equation: I_Ca,L = G_max(V − E_rev)/{1 + exp[(V_0.5 − V)/k]}, where V is the membrane potential, G_max is the maximal conductance, E_rev is the reversal potential, V_0.5 is the half-activation potential, and k is the slope factor (34). The parameter of the curve obtained by fitting 10 points is given by means ± SD. The f_∞-V relationship was fitted to 14 points given by means ± SE with the following Boltzmann equation: f_∞ = c₀ + (c₁ − c₀)/{1 + exp [−(V − V_0.5)/k]}, where c₀ is a voltage-independent constant, c₁ − c₀ is the maximal availability of the voltage-dependent component, V is the membrane potential of prepulse, and V_0.5 is the voltage required for 50% inactivation of the voltage-dependent component.

Statistical analyses.

Statistical results of the experimental data are presented as means ± SE, and their statistical comparisons were performed using a paired t-test in Fig. 1, two-way ANOVA followed by Dunnett's multiple comparison test (Figs. 3 and 9B), followed by Tukey's multiple comparison (Figs. 2, 5, 7, and 9A), Sidak's test (Fig. 8A), and both Dunnett's and Sidak's test in Fig. 8B. Parameters obtained by fitting with the Boltzmann equation were compared with one-way ANOVA followed by Tukey's multiple comparison.

Fig. 1.

Effect of rotenone on mitochondrial superoxide in A7r5 cells. Mitochondrial superoxide content was estimated by measuring 2-OH-Mito-ethidium, the superoxide-specific product of MitoSOX, using HPLC analysis. Control represents a time-matched background product. The value for rotenone represents the product from cells treated with 10 μM rotenone for 3 min. Values are shown as means ± SE, n = 5. **P < 0.01 compared with control.

Fig. 3.

Voltage-dependence of rotenone (Rot)-induced increase in I_Ca,L. a: command pulse, b–d, family of typical current traces [b, Con 1, recorded 2 min before c; c, Con 2 (control for Rot); d, Rot, recorded 1 min after application of 10 μM Rot]; traces are shown in black from −40 to 0 mV, in gray from 10 to 50 mV, and by a thick line at −20 mV. A and B: I-V relationship of I_Ca,L (A, current density; B, amplitude normalized by the maximal amplitude of control). Values shown are means ± SE, n = 18. Rot, 10 μM rotenone. The amplitude before and after exchange of the solution at each potential was compared by two-way ANOVA. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 between Con 2 and Rot. I-V was fitted using the Boltzmann equation. In A, G_max (pS/pF) Con 1, 80.5 ± 6.6; Con 2, 82.0 ± 6.8; Rot, 91.8 ± 8.1; n = 10; P < 0.01 compared Con 2 with Rot: V_0.5 (mV); Con 1, −19.7 ± 1.2; Con 2, −19.8 ± 1.2; Rot, −22.2 ± 1.3; n = 10, P < 0.001 compared Con 2 with Rot. In B, V_0.5 (mV) Con 1, −19.4 ± 1.3; Con 2, −19.6 ± 1.3; Rot, −22.2 ± 1.3, n = 10, P < 0.001 compared Con 2 with Rot.

Fig. 9.

Effect of staurosporine on I_Ca,L and its modulation by rotenone (Rot). After pretreatment with 100 nM staurosporine (S), the I-V relationship was determined in 2-min intervals. Rot (10 μM) was introduced after S-induced inhibition of I_Ca,L. a–c: typical records (a: S1, 1–2 min before S2; b: S2, control; c: S + Rot, 1–2 min after addition of Rot). A: traces are shown in black from −40 to 0 mV, in gray from 10 to 50 mV, and by a thick line at −20 mV. I-V of I_Ca,L (I_peak and I₅₀₀) in pA/pF. Values shown are means ± SE; control from cells in Figs. 3, 5, and 7 (n = 40), S and S + Rot (n = 12). *P < 0.05, ****P < 0.0001 S1 compared with control. B: I-V normalized by the maximal I_peak of S2. ##P < 0.01, ###P < 0.001 S1 compared with S2; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 S2 compared with S + Rot. I-V fitting with the Boltzmann equation. G_max: control, 81.5 ± 6.7; S1, 51.8 ± 4.4 pS/pF (P < 0.001 S1 compared with control); S2, 49.3 ± 4.3; S + Rot, 46.4 ± 4.6 pS/pF. V_0.5: control, −19.6 ± 1.2; S1, −22.3 ± 1.4 mV (P < 0.0001 S1 compared with control); S2, 22.6 ± 1.4; S + Rot, −23.8 ± 1.6 mV.

Fig. 2.

Rotenone-induced dose- and time-dependent increase in I_Ca,L. a: test pulse, b–e: typical current traces with superposition of control (black) and trace with maximal increase of I_Ca,L (gray). b: 10 nM, c: 100 nM, d: 1 μM, e: 10 μM. A: Change in peak amplitude of I_Ca,L. B: change in I_Ca,L amplitude at 500 ms (I₅₀₀) before and during application of various concentrations of rotenone (Rot). Test pulse to −15 mV was applied at 20-s intervals. Values shown are means ± SE of the ratio to the control amplitude obtained before switching to rotenone. n = 11, 10 nM; 12, 100 nM; 13, 1 μM; and 18, 10 μM. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 compared with the values obtained in the presence of 10 nM rotenone; #P < 0.05 compared with 1 μM rotenone.

Fig. 5.

Effect of DTT and rotenone (Rot) on I-V relationship of I_Ca,L. a–d: typical current traces (a: Con 1; b: Con 2; c: 2 mM DTT; d: 2 mM DTT + 10 μM Rot recorded with a 2-min interval). Traces are shown in black from −40 to 0 mV, in gray from 10 to 50 mV, and by a thick line at −20 mV. In I–V, normalized amplitude by the maximal amplitude of control (values shown are means ± SE, n = 10) is plotted vs. voltage. The curves of Con 2 and Con 1, obtained 1–2 min before Con 2, almost overlap. Statistical comparison was performed between values obtained before and after the exchange. *P < 0.05, ****P < 0.0001 for the exchange from control to DTT, ##P < 0.01, ###P < 0.001, ####P < 0.0001 for that from DTT to DTT + Rot. I-V was fitted using the Boltzmann equation. V_0.5 (mV): Con 1, −19.0 ± 1.4; Con 2, −18.9 ± 1.4; 2 mM DTT, −17.8 ± 1.3; 2 mM DTT + 10 μM Rot, −19.3 ± 1.3, n = 10.

Fig. 7.

Changes in I-V relationship produced by H₂O₂, rotenone, and DTT. a–d: typical traces (a: control; b: after 5 min in the presence of 0.3 mM H₂O₂; c: after 1–2 min of addition of 10 μM Rot to H₂O₂; d: after 1–2 min of addition of 2 mM DTT to H₂O₂ + Rot). Traces are shown in black from −40 to 0 mV, in gray from 10 to 50 mV, and by a thick line at −20 mV. I–V: peak amplitude and I₅₀₀ normalized by maximal amplitude in control; values shown are means ± SE, n = 11. Statistical comparison before and after solution change: *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, control vs. H₂O₂, #P < 0.05, ##P < 0.01, ####P < 0.0001, H₂O₂ vs. H₂O₂ + Rot, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, H₂O₂ + Rot vs. H₂O₂ + Rot + DTT. I-V was fitted using the Boltzmann equation. V_0.5 (mV): Con, −19.8 ± 1.2; H₂O₂, −20.2 ± 1.3; H₂O₂ + Rot, −23.1 ± 1.5; P < 0.001 compared with H₂O₂, H₂O₂ + Rot + DTT, −21.3 ± 1.3; P < 0.05 compared with H₂O₂ + Rot.

Fig. 8.

Effect of Tempol on I_Ca,L and its rotenone-induced modulation. A: effect of 10 μM rotenone on I_Ca,L elicited by repetitive depolarization to −15 mV in the presence of 1 mM Tempol (T, means ± SE, n = 16) and in the absence of T (n = 18) from Fig. 1. *P < 0.05, **P < 0.01, ***P < 0.001, time-matched comparison. a: typical record in the presence of T; control (black), maximal increase by rotenone (Rot) (dark gray), and 5 min after Rot (light gray). B: effect of T and that of Rot in the presence of T on I-V of I_Ca,L (pA/pF, means ± SE, n = 14). b–d: typical records (b: T1 3–4 min before; c: T2 1–2 min before switching to T + 10 μM Rot; d: 1–2 min after switching to T + Rot). Traces are shown in black from −40 to 0 mV, in gray from 10 to 50 mV, and by a thick line at −20 mV. Control, from cells in Figs. 3, 5, and 7 (n = 40). #P < 0.05 compared T1 with control (Sidak's test), **P < 0.01, ****P < 0.00001, compared T + Rot with T2 in pared data (Dunnett's test). I-V was fitted using the Boltzmann equation. G_max: control, 81.5 ± 6.7; T1, 99.4 ± 8.6 pS/pF (P < 0.001 compared with control); T2, 97.4 ± 8.4; Tempol + Rot, 100.5 ± 9.1 pS/pF. V_0.5: control, −19.6 ± 1.2; T1, −21.4 ± 1.2 (P < 0.01 compared with control); T2, −21.8 ± 1.2 mV; Tempol + Rot, −23.4 ± 1.3 mV (P < 0.05 compared with T2).

RESULTS

Effect of rotenone on mitochondrial superoxide release.

Superoxide in the mitochondrial matrix oxidizes MitoSOX and generates 2-OH-Mito-ethidium (55). It was quantified using HPLC analysis (3, 55). Abundance of 2-OH-Mito-ethidium in control cells and in cells treated with 10 μM rotenone for 3 min is shown in Figure 1. Oxidation of MitoSOX by superoxide should proceed during its loading (1 h) and application of rotenone for 3 min. Rotenone administration led to a significant increase in abundance of 2-OH-Mito-ethidium compared with control (P < 0.01), indicating that it resulted in an acute increase in mitochondrial superoxide release in A7r5 cells.

Rotenone induces a dose-dependent increase in I_Ca,L.

I_Ca,L was identified from its voltage dependence and marked increase by 100 nM (−) Bay-K 8644 in preliminary experiments. The run-down of I_Ca,L was small in the present recordings with a thin pipette containing 5 mM BAPTA [cf. (38)]. To study time- and dose-dependent modulation of I_Ca,L by rotenone, various concentrations of rotenone (10 nM to 10 μM) were applied during repetitive depolarization to −15 mV at 20-s intervals (Fig. 2). Representative records at various concentrations show that I_Ca,L amplitude was clearly increased by 100 nM (Fig. 2c), 1 μM (Fig. 2d), and 10 μM (Fig. 2e) rotenone and that the increase rapidly decays during pulses, except at 100 nM. A plot of the time-lapse change in amplitude of I_Ca,L shows that the change mediated by 10 nM rotenone was negligible, 100 nM induced a rapid and transient increase, 1 μM produced a larger and more sustained increase, and that the largest increase was produced by 10 μM (Fig. 2A). The changes induced by 100 nM and 10 nM rotenone were different, with statistical significance only at 1 min, whereas the differences between 1 μM or 10 μM and 10 nM were highly significant, starting from 40 s until 5 min. The greatest increase was caused by 10 μM rotenone (22.1 ± 3.4%, n = 18) at 80 s, and the increase gradually decreased to 9% after 5 min. In the presence of rotenone, the amplitude at the end of the pulse (I₅₀₀) initially increased only slightly at 100 nM, and gradually decreased in the presence of rotenone in a dose-dependent manner by 9% at 10 nM, 11% at 100 nM and 1 μM, and 19% at 10 μM. I₅₀₀ in the presence of 10 μM and 10 nM rotenone differed with statistical significance (P < 0.05) after 3 min (Fig. 2B). Peak amplitude of I_Ca,L elicited by repetitive depolarization pulses to 0 mV was significantly increased by 16.4 ± 2.2% (n = 22) with 10 μM rotenone (data not shown).

Effect of rotenone on current voltage relationship of I_Ca,L.

After the maximal amplitude of I_Ca,L was stabilized, two control I-V measurements were recorded and 10 μM rotenone was introduced. The recording of test series in the presence of rotenone started 1 min after changing the solution. The difference between the two control I-V measurements (Con 1 and Con 2) was negligible (Fig. 3, b and c). Rotenone produced an acute increase in I_Ca,L at each potential between −30 and 10 mV (Fig. 3, A and B). The relative increase was greater at a small depolarization of −30 and −20 mV compared with that at a large depolarization. The larger increase at negative voltages is also apparent when comparing the current traces of −20 mV (Fig. 3c vs. Fig. 3d). Fitting of I-V curves using the Boltzmann equation showed that rotenone increased macroscopic conductance (G_max) by 13.0% (P < 0.01) and induced a 2.6 mV leftward shift of V_0.5 from −19.6 ± 1.3 mV (SD) to −22.2 ± 1.4 mV (SD, P < 0.001) in Figure 3A. Similar percentage increases in G_max and a leftward shift of V_0.5 were also obtained in I-V measurements of normalized amplitude (Fig. 3B).

Effect of DTT and H₂O₂ on rotenone-induced increase in I_Ca,L.

To examine the role of thiol-oxidation, thiol-glutathionylation, and thiol-nitrosylation in the rotenone-induced increase in I_Ca,L, the effect of DTT was examined. DTT (2 mM) applied during repetitive depolarization to −15 mV increased peak amplitude of I_Ca,L by 17.9 ± 2.5% (n = 13, Fig. 4A). Thereafter, 10 μM rotenone applied in the presence of DTT increased I_Ca,L by 18% in 1–2 min, resulting in an overall increase of 29.4 ± 4.6%. When the application of rotenone and DTT was conducted in a reverse order (Fig. 4B), DTT resulted in an I_Ca,L increase to a similar extent as in Figure 4A. In the I-V relationship, DTT produced a significant increase in I_Ca,L at voltages between −10 mV and 20 mV (Fig. 5), increased the G_max by 18.1%, and shifted V_0.5 slightly in a positive direction by 0.9 mV. DTT did not induce significant changes in I₅₀₀. Rotenone added in the presence of DTT resulted in a significant increase in I_Ca,L in both peak amplitude and I₅₀₀ at potentials between −20 mV and 0 mV, increased G_max by 12.6%, and slightly shifted V_0.5 to a hyperpolarizing direction by 1.5 mV.

Fig. 4.

Time-dependent changes in I_Ca,L induced by DTT and rotenone. A: effect of 2 mM DTT on basal I_Ca,L and on 10 μM rotenone (Rot)-induced modulation of I_Ca,L. Test pulse to −15 mV was applied at 20-s intervals. Values shown are means ± SE from 13 cells plotted vs. time. a: superposed traces of control (black) in the presence of DTT (dark gray) and DTT + 10 μM Rot (gray). B: effect of 10 μM rotenone (shown in Fig. 1) and 2 mM DTT in the presence of rotenone. Values shown are means ± SE, n = 18. a and b: typical traces [a, control (black) after 1 min in the presence of rotenone (dark gray) and after 5 min (gray); b, after 5 min in the presence of rotenone (black) and in the presence of rotenone and DTT (dark gray)].

The ineffectiveness of DTT in influencing the rotenone-induced increase in I_Ca,L suggests that H₂O₂ is not involved in this process because it targets oxidation-sensitive amino acid residues, and the H₂O₂-induced increase in I_Ca,L due to CaMKII-dependent protein phosphorylation is suppressed by DTT (41). Rotenone was applied in the presence of H₂O₂ to further clarify the involvement of H₂O₂ in this process. Externally applied H₂O₂ (0.3 mM) inhibited I_Ca,L with a gradual onset, decreasing peak amplitude by 13.1 ± 3.6% (means ± SE, n = 12) and I₅₀₀ by 43.7 ± 4.5% in 5 min (Fig. 6a). Following inhibition, 10 μM rotenone in the presence of H₂O₂ produced a rapid increase in I_Ca,L by 19% (Fig. 6b), which gradually declined. The subsequent application of DTT produced a large (Fig. 6c) and variable increase of 26% in peak amplitude. In the I-V relationship, 0.3 mM H₂O₂ produced a comparable inhibition of I_Ca,L and I₅₀₀ at voltages between −20 mV and 30 mV with statistical significance (P < 0.001) (Fig. 7). Fitting I-V curves with the Boltzmann equation revealed an H₂O₂-induced decrease of 19% in G_max (P < 0.0001) without a significant shift in V_0.5. In the presence of H₂O₂, 10 μM rotenone resulted in an I_Ca,L increase at negative potentials between −30 and −10 mV with statistical significance, increased G_max by 12% (P < 0.01) and shifted V_0.5 by 2.9 mV (P < 0.001) in a negative direction. In the presence of H₂O₂ and rotenone, DTT led to a further increase in I_Ca,L at voltages between −10 mV and 40 mV with statistical significance, with a 12% increase in G_max (P < 0.05) and a V_0.5 shift in a positive direction by 1.8 mV (P < 0.05). The decrease in I₅₀₀ induced by H₂O₂ was not significantly affected by rotenone or by DTT.

Fig. 6.

I_Ca,L inhibition by H₂O₂ and increase by rotenone and DTT in the presence of H₂O₂. H₂O₂ (0.3 mM), H₂O₂ + 10 μM Rot, and H₂O₂ + Rot + 2 mM DTT were sequentially applied during repetitive depolarization to −15 mV at 20-s intervals. a: superposed traces of control (black) and H₂O₂ (gray); b: H₂O₂ (black) and 1 min after rotenone (gray); c: the current at the end of H₂O₂ + Rot (black) and H₂O₂ + Rot + 2 mM DTT (gray). Values shown are means ± SE (n = 12) of normalized peak amplitude of I_Ca,L and I₅₀₀ was plotted vs. time.

Reduction of rotenone-induced increase in I_Ca,L by Tempol.

To determine the involvement of superoxide in a rotenone-induced increase in I_Ca,L, cells were treated with Tempol (T), a membrane-permeable SOD mimetic of weak catalase-like activity (48). Cells were treated with 1 mM Tempol for periods of time greater than 40 min (usually 4 to 6 h) at room temperature, with or without an additional 15- to 20-h exposure period at 4°C; I_Ca,L was then measured in the presence of Tempol. Rotenone (10 μM) applied during repetitive depolarization increased I_Ca,L amplitude transiently by 5.2% at its maximum (Fig. 8A). This rotenone-induced increase was significantly smaller than that obtained in the absence of Tempol. The change in I₅₀₀ induced by rotenone was not different from that observed in control cells. When compared with the I-V relationship of I_Ca,L current density, Tempol-treated cells exhibited significantly greater current density at −10 mV (Fig. 8B). In the presence of Tempol, I_Ca,L was increased by 10 μM rotenone slightly but significantly at voltages between −30 and −10 mV. G_max was increased from 81.5 ± 6.7 pS/pF (control 40 cells) to 99.4 ± 8.6 pS/pF by Tempol (T1, P < 0.001 compared with control). However, the change in G_max from T2, 97.4 ± 8.1 to T +10 μM rotenone, 100.5 ± 9.1 pS/pF was not statistically significant. Thus rotenone-induced increase in I_Ca,L was considerably suppressed by Tempol. In the presence of Tempol, rotenone shifted V_0.5 from T2, −21.8 ± 1.2 mV to T + Rot, −23.4 ± 1.3 mV (P < 0.05).

Effect of staurosporine on I_Ca,L and rotenone-induced increase in I_Ca,L.

I_Ca,L increase by a β-adrenergic agonist through PKA-dependent channel phosphorylation is associated with a negative shift of I-V in rabbit ventricular myocytes (43). To determine the involvement of phosphorylation in the superoxide-induced increase in I_Ca,L, we examined the effect of staurosporine (S), a broad-spectrum protein kinase inhibitor (31). Because I_Ca,L is modulated by the coordinated activity of protein kinases and protein phosphatases, S inhibits I_Ca,L in the presence of active phosphatases in mice ventricular myocytes (15). To obtain a steady-state effect of S, cells were pretreated with 100 nM S for 40–450 min (average 165 min). I_Ca,L was inhibited by the sustained application of 100 nM S, with a decrease of the maximal density in I-V from 4.21 ± 0.30 pA/pF (control, means ± SE, n = 40) to 2.67 ± 0.40 pA/pF (S, n = 12, P < 0.05) (Fig. 9A). Compared with control, S led to a decrease in G_max from 81.5 ± 6.7 to 51.8 ± 4.4 pS/pF (S1, P < 0.001) and V_0.5 of activation was shifted by S in a negative direction from control by −19.6 ± 1.2 mV to −22.3 ± 1.4 mV (S1, P < 0.001). Figure 9, a–c, is a representative record showing that 10 μM rotenone had little effect on I_Ca,L in the presence of S. An I-V relationship of normalized amplitude (Fig. 9B) shows that S2 decreased slightly compared with S1, and that 10 μM rotenone applied after S2 did not affect the peak amplitude of I_Ca,L at the negative membrane potentials but slightly decreased it at large positive potentials. Rotenone decreased I₅₀₀ at voltages between −10 and 20 mV.

Effect of rotenone on steady-state inactivation and window current of I_Ca,L.

The rotenone-induced increase in I_Ca,L gradually decreased along with I₅₀₀ over 5 min (Fig. 2). These findings suggest that rotenone augments the voltage-dependent inactivation of I_Ca,L. We examined the effect of rotenone on a quasi-steady state inactivation curve (f_∞-V). In Figure 10, control f_∞-V preceded control I-V before the application of rotenone, and f_∞-V in the presence of rotenone was obtained after 5–10 min treatment with rotenone. f_∞-V was shifted by 10 μM rotenone in a hyperpolarizing direction. Fitting of f_∞-V using the Boltzmann equation showed a change in V_0.5 from −33.0 mV in the control to −38.4 mV in the presence of rotenone, and a rotenone-mediated decrease in c₀, a voltage-independent constant, from 0.11 to 0.06 (Fig. 10, a, b, and A). From the I-V curve shown in Figure 3A and the f_∞-V relationships, the window current (I_WD) was calculated as the product of the simulated I-V and f_∞-V curves at the voltage range between −40 and 0 mV [Fig. 10B (34)]. In the control, the I_WD density was maximal at a voltage between −20 and −10 mV. Rotenone shifted the peak of I_WD to −20 mV and increased the density of I_WD by 20–30% at a voltage range between −30 mV and −20 mV (Fig. 10B, Rot 1). The shift of f_∞-V to a hyperpolarizing direction in the presence of rotenone decreased window I_WD at all voltage ranges, a decrease more prominent at voltages more positive than −20 mV. Because I_Ca,L decreases over time in the presence of rotenone (Fig. 2), the actual decrease in I_WD at 5 min was expected to be more extensive than the curve shown as Rot 2 in Figure 10B.

Fig. 10.

Effect of rotenone on steady-state inactivation and window current of I_Ca,L. a and b show typical current traces during test pulse to 0 mV after conditioning steps shown to the left (a: control, b: in the presence of 10 μM rotenone). A: quasi-steady-state inactivation curves (f_∞-V) in control and after 5- to 10-min application of 10 μM rotenone; values shown are means ± SE, n = 17. Curves were obtained by fitting with the Boltzmann equation. V_0.5 for 50% reduction of the voltage-dependent component was −33.0 mV for control and −38.4 mV for rotenone. B: effect of 10 μM rotenone on window I_Ca,L (I_WD) obtained as the product of the fitting curve of I-V of the peak amplitude in Fig. 3A and of f_∞-V in Fig. 10A. Control I_WD represents the product of control I-V and control f_∞-V. I_WD of Rot 1 and Rot 2 were obtained as the product of I-V obtained from 1–2 min in the presence of rotenone and control f_∞-V and the product of the same I-V and f_∞-V obtained after a 5-min application of rotenone, respectively.

DISCUSSION

Rotenone augmented mitochondrial superoxide release estimated by mitochondrial-specific MitoSOX products using HPLC in A7r5 cells (Fig. 1), as shown previously by the increase in MitoSOX fluorescence in bovine coronary ASMCs (16). In this study we found that rotenone resulted in an acute increase in I_Ca,L. BAPTA (5 mM) in the pipette solution separated the superoxide-dependent increase in I_Ca,L from the increase produced by H₂O₂ mediated by CaMKII in rat ventricular myocytes (41) and PKCα in rat cerebral artery ASMCs (5, 12). Furthermore, I_Ca,L was increased by rotenone in the presence of H₂O₂, which induced weak inhibition of I_Ca,L. In addition, a rotenone-induced increase occurred in the presence of DTT, which suppresses the H₂O₂-induced increase in I_Ca,L in rat ventricular myocytes (41). The inhibition of rotenone-induced I_Ca,L increase by Tempol, a membrane-permeable SOD mimetic (48), further supports the notion that superoxide can increase I_Ca,L prior to its dismutation to H₂O₂.

The percentage of O₂ converted to superoxide in normal oxidative metabolism of the MRC is known to be ∼2% (11, 44). Complexes I and III of the MRC are major mitochondrial superoxide release sites, into the matrix, and into the matrix and intermembrane space, respectively (7), although complex II and other proteins also contribute to the release (7, 17). Rotenone is a complex I inhibitor and augments superoxide release in neurons (44), ASMCs from bovine coronary artery (16), rat renal artery (29), and A7r5 cells (Fig. 1). Superoxide released from complex I is dismutated to H₂O₂ by MnSOD in the matrix and by Cu/ZnSOD in the cytoplasm and intermembrane space (44). Superoxide in the matrix must traverse the inner and outer mitochondrial membranes to reach the cytoplasm (Fig. 11). We hypothesize that the superoxide efflux is facilitated by physiological openings in the mitochondrial permeability transition pores composed of the c-subunit ring of F1F0 ATP synthase in the inner mitochondrial membrane (1) and VDACs in the outer mitochondrial membrane, which is inhibited by 4′-diisothiocyano-2,2′-disulfonic acid stilbene (DIDS) (19). It has been shown that rotenone augments the release of superoxide from the matrix to the outside of mitochondria isolated from mice skeletal muscle, and that the release is largely enhanced by knockdown of the MnSOD gene and suppressed by DIDS (28). Because a population of mitochondria is located close to the Ca_V1.2 channel complex in rat cerebral ASMCs (13) and Ca_V1.2 is mechanically connected via cytoskeletal proteins to VDACs in the outer membrane of mitochondria in guinea pig ventricular myocytes (45), it seems possible that superoxide could reach targets outside of the mitochondria and modulate Ca_V1.2 function.

Fig. 11.

Model for the superoxide-induced increase in I_Ca,L in A7r5 aortic smooth muscle cells. Rotenone stimulates superoxide (O₂^·−) release from complex I in the inner mitochondrial membrane to the matrix. Superoxide diffuses out from the matrix to the cytoplasm through the inner mitochondrial membrane, intermembrane space (IMS), and outer mitochondrial membrane. Superoxide is partly dismutated to H₂O₂ by SOD in the matrix, IMS, and in the cytoplasm. H₂O₂ activates PKCα and increases I_Ca,L. However, DTT prevents thiol-oxidation by H₂O₂ and prevents PKCα activation, and BAPTA from the pipette solution decreases cytoplasmic [Ca²⁺] and inhibits the PKCα activation. Superoxide binds nitric oxide (NO) and generates peroxynitrite (ONOO⁻), which can increase I_Ca,L. However, DTT inhibits ONOO⁻-induced thiol-nitrosylation and inhibits the increase. NOS-derived NO activates guanylate cyclase (GC) and generates cGMP, which activates PKG, resulting in inhibition of I_Ca,L. Reduction of NO by superoxide eventually downregulates PKG and increases I_Ca,L. Tempol decreases the concentration of superoxide. Staurosporine inhibits PKCα and PKG and eventually inhibits the superoxide-induced increase in I_Ca,L.

The suppression of the rotenone-induced I_Ca,L increase by Tempol indicates that superoxide is involved in this process. Although further study is necessary, the Tempol-induced increase in I_Ca,L and shift of I-V to a negative direction could be explained by the direct action of Tempol on the Ca_V1.2 channel, since Tempol increases the BK_Ca current by direct action on the BK_Ca α subunit in ASMSCs from rat mesenteric artery (53). I_Ca,L is closely regulated by protein phosphorylation, even under basal conditions, through the balanced activity of protein kinases and protein phosphatases in cardiac myocytes and smooth muscle cells (20, 21, 24, 32). Staurosporine, originally isolated from the culture medium of Actinomyces, exhibits broad-spectrum protein kinase inhibitor activities (31, 36). The pIC₅₀ of staurosporine against protein kinases ranges from 8 to 9 for most PKC, PKA, CaMKII, and tyrosine kinases (31). Staurosporine decreases I_Ca,L in the presence of phosphatase activity in mouse ventricular myocytes (15). Here, staurosporine (100 nM) progressively inhibited I_Ca,L and suppressed the rotenone-induced I_Ca,L increase (Fig. 9). This finding indicates that protein phosphorylation supports basal I_Ca,L in A7r5 cells and that superoxide modifies protein kinase-dependent modulation of I_Ca,L. The staurosporine-induced suppression of the increase in I_Ca,L rules out any direct interaction of superoxide or rotenone with Ca_V1.2 channels from the mechanism of the rotenone-induced increase in I_Ca,L.

Thiol residues of channel proteins or their modulatory proteins are oxidized or nitrosylated in the initial step of ROS-induced regulation of protein function (22, 50, 51). H₂O₂-induced inhibition of I_Ca,L is likely due to protein thiol oxidation because it is reversible by DTT (cf. Figs. 6 and 7). NO synthase (NOS) is expressed in ASMCs and in endothelial cells (6). Because superoxide rapidly binds NO to produce peroxynitrite (ONOO⁻), it has dual roles in the presence of NO as a scavenger of NO and as a precursor of ONOO⁻ (51) (Fig. 11). DTT inhibits ONOO⁻-induced increases of high K⁺-induced contraction in guinea pig gall bladder smooth muscle strips (2). ONOO⁻ binds with thiol residues of proteins to result in S-nitrosylation and S-nitrosoglutathionilation of proteins (10). Extracellular application of S-nitrocysteine and S-nitrosoglutathione (GSNO) augments I_Ca,L in ferret ventricular myocytes (10), while GSNO induces inhibition of I_Ca,L in Ca_V1.2b-expressing HEK293 cells (37). These GSNO-induced I_Ca,L changes are inhibited by DTT (10, 37) (Fig. 11). Moreover, protein S-nitrosylation in the diabetic rat heart is reduced by DTT (35). Because DTT had little effect on the superoxide-induced increase in I_Ca,L, the contribution of thiol oxidation, S-nitrosylation, and S-nitroglutathionylation to the increase in I_Ca,L could be ruled out. H₂O₂ transfers two electrons during oxidation while superoxide transfers one (49), suggesting that difference in the chemical reactivity may explain the lower sensitivity of superoxide-induced modification to DTT.

NO activates guanylate cyclase (GC) to generate cGMP, which in turn activates PKG. Application of 8-BrcGMP, a membrane-permeable cGMP analog, inhibits I_Ca,L in A7r5 cells (27) and the PKG inhibitor Rp-8-Br-PET-cGMPS rapidly induces a moderate increase in I_Ca,L in rabbit portal vein smooth muscle cells (39). Staurosporine-induced inhibition of PKG could abolish NO-induced inhibition prior to the application of rotenone, thus eliminating the NO-scavenging effect of superoxide to increase I_Ca,L. The hypothetical cGMP-PKG-mediated mechanism of rotenone-induced I_Ca,L increase (shown in Fig. 11) proceeds as follows:

Rotenone → complex I superoxide release ↑ → matrix superoxide ↑→ cytoplasmic superoxide ↑→ cytoplasmic NO↓→ GC activity↓→ cGMP↓→ PKG↓→ I_Ca,L↑.

The notion that a superoxide-induced leftward shift n the I-V relationship by the decrease of NO is seemingly not compatible with the absence of a detectable I-V shift in 8-BrcGMP-induced inhibition of I_Ca,L in rabbit portal vein smooth muscle cells (23), and an NO donor-induced inhibition of I_Ca,L in rat vestibular hair cells (4). The sustained treatment by staurosporine shifted the V_0.5 to a negative direction, however. If V_0.5 contributed by basal PKG activation is more positive than in its absence, suppression of PKG by superoxide or staurosporine could shift V_0.5 to a negative direction. Superoxide generated by the xanthine/xanthine oxidase system augments the activity of purified PKC isolated from rat brain, and the activation is associated with a marked depletion of Zn²⁺, thereby, superoxide-induced activation of PKC was partially (about 30%) suppressed by DTT (26). Direct modification of such regulatory proteins by superoxide may contribute to the mechanism of superoxide-induced increase in I_Ca,L.

The superoxide-induced transient I_Ca,L increase associated with a shift in the I-V relationship in a hyperpolarizing direction increased window I_Ca,L close to the resting potential. Increased I_WD is expected to transiently augment the contraction of ASMCs. However, the effect of superoxide on [Ca²⁺]_i increase of ASMCs is overlapped with that of H₂O₂ and ROS-induced modulation of potassium current. Sustained pretreatment by rotenone did not affect high-K⁺-induced contraction of bovine coronary artery rings (16), and rotenone-induced ROS production results in relaxation of rat renal arterial rings by increasing voltage-dependent K⁺ current, while it induces constriction of rat pulmonary artery by decreasing the K⁺ current (29). The effect of modulation of I_Ca,L by mitochondrial ROS on arterial contraction is considered to be variable among ASMSCs from different vascular beds and this may originate from the difference in mechanisms influenced by superoxide and/or hydrogen peroxide.

Limitation of the present study.

Staurosporine is a broad-spectrum protein kinase inhibitor. Although suppression by staurosporine of the superoxide-induced increase in I_Ca,L suggests the involvement of one or more protein kinases, no responsible protein kinase was identified.

In conclusion, the complex I inhibitor rotenone resulted in stimulating superoxide release in mitochondria and increasing I_Ca,L recorded with a BAPTA-containing pipette associated with a shift in I-V to a hyperpolarizing direction. Because the increase was suppressed by the SOD mimetic Tempol, and was not affected by H₂O₂ or DTT, it was concluded that superoxide (but not H₂O₂) was responsible for this increase. From its occurrence in the presence of DTT, the involvement of peroxynitrite or S-nitrosylation was excluded from the mechanism. Staurosporine suppression of the increase in I_Ca,L indicates that protein kinase is involved in this mechanism. Although further experiments are necessary to confirm the NO-scavenging effect of superoxide as the mechanism of I_Ca,L increase, with other mechanisms possible, this study has provided novel evidence of mitochondrial superoxide-mediated increase in I_Ca,L.

GRANTS

Support for this study was provided by National Heart, Lung, and Blood Institute Grants R01 HL-115124 to M.S. Wolin and R01 HL-085352 to S.A. Gupte.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

R.O. and S.A.G. conception and design of research; R.O., V.D., A.L., and D.P. performed experiments; R.O. and A.L. analyzed data; R.O., M.S.W., and S.A.G. interpreted results of experiments; R.O. prepared figures; R.O. drafted manuscript; R.O., M.S.W., and S.A.G. edited and revised manuscript; R.O., V.D., A.L., D.P., M.S.W., and S.A.G. approved final version of manuscript.