Local Regulation of Arterial L-Type Calcium Channels by Reactive Oxygen Species

Abstract

Rationale

Reactive oxygen species (ROS) are implicated in the development of cardiovascular disease and oxidants are important signaling molecules in many cell types. Recent evidence suggests that localized subcellular compartmentalization of ROS generation is an important feature of ROS signaling. However, mechanisms that transduce localized subcellular changes in redox status to functionally-relevant changes in cellular processes such as Ca²⁺ influx are poorly understood.

Objective

To test the hypothesis that ROS regulate L-type Ca²⁺ channel activity in cerebral arterial smooth muscle.

Methods and Results

Using a TIRF (total internal reflection fluorescence) imaging-based approach, we found that highly-localized subplasmalemmal generation of endogenous ROS preceded and colocalized with sites of enhanced L-type Ca²⁺ channel sparklet activity in isolated cerebral arterial smooth muscle cells. Consistent with this observation and our hypothesis, exogenous ROS increased localized L-type Ca²⁺ channel sparklet activity in isolated arterial myocytes via activation of protein kinase C alpha (PKCα) and when applied to intact cerebral arterial segments, exogenous ROS increased arterial tone in an L-type Ca²⁺ channel-dependent fashion. Furthermore, angiotensin II-dependent stimulation of local L-type Ca²⁺ channel sparklet activity in isolated cells and contraction of intact arteries was abolished following inhibition of NADPH oxidase.

Conclusions

Our data support a novel model of local oxidative regulation of Ca²⁺ influx where vasoconstrictors coupled to NAPDH oxidase (e.g., angiotensin II) induce discrete sites of ROS generation resulting in oxidative activation of adjacent PKCα molecules that in turn promote local sites of enhanced L-type Ca²⁺ channel activity resulting in increased Ca²⁺ influx and contraction.

Introduction

Oxidative and reductive biochemical processes are integral components of cellular biology. However, disruption of cellular redox status beyond physiological parameters is correlated with disorders ranging from cardiovascular disease to neurodegeneration to cancer. The cardiovascular system appears to be particularly sensitive to perturbations in redox balance with increased oxidative stress implicated in the development of cardiovascular diseases including heart failure, atherosclerosis, stroke and hypertension ^1–5. The apparent sensitivity of the cardiovascular system to redox imbalance may reside in the relative functional importance of redox signaling under physiological conditions. Redox biochemistry has been studied extensively in the vasculature ^{4, 6–9} with specialized subcellular compartmentalization, regulation, and generation of reactive oxygen species (ROS) emerging a common theme ^10–12. However, mechanisms that transduce localized subcellular changes in redox signaling to functionally-relevant changes in cellular processes such as Ca²⁺ influx are less clear.

Peripheral arterial resistance is determined by the contractile state of arterial smooth muscle which is tightly coupled to Ca²⁺ influx through L–type voltage-dependent Cav1.2 Ca²⁺ channels ^13–15. Total internal reflection fluorescence (TIRF) microscopy has been used to image L-type Ca²⁺ channel (Ca²⁺ sparklet) activity in arterial smooth muscle cells with high temporal and spatial resolution ^16–19. Using this approach, a high-activity mode of Cav1.2 Ca²⁺ channel function was identified. These studies revealed that high-activity Ca²⁺ sparklets resulted from spatially restricted enhancement of Cav1.2 Ca²⁺ channel function by protein kinase C alpha (PKCα). PKCα-dependent Ca²⁺ sparklet activity was found to be necessary for normal arterial function ^{16, 19, 20}. In addition, increased PKCα-dependent Ca²⁺ sparklet activity was associated with altered gene expression and arterial dysfunction in experimental models of hypertension ^{19, 20}. Despite these intriguing observations, the molecular mechanisms underlying high-activity localized PKCα-dependent L-type Ca²⁺ channel function are poorly understood.

Using a novel TIRF imaging-based approach, we identified and characterized a mechanism that functionally links local oxidant signaling to sustained colocalized L-type Ca²⁺ channel activity. We demonstrate for the first time that highly-localized subplasmalemmal generation of endogenous ROS precede and colocalize with enhanced L-type Ca²⁺ channel activity in cerebral arterial smooth muscle. Consistent with this observation, ROS increased L-type Ca²⁺ channel sparklet activity via activation of PKCα and increased arterial tone in an L-type Ca²⁺ channel-dependent fashion. Taken together, our data suggest that local oxidative activation of PKCα-dependent L-type Ca²⁺ influx represents a functionally relevant convergence of oxidant and Ca²⁺ signaling pathways in cerebral arterial smooth muscle with physiological and pathological implications.

Methods

Male Sprague-Dawley rats were euthanized with sodium pentobarbital (200 mg/kg intraperitoneally) as approved by the Institutional Animal Care and Use Committee of Colorado State University. Smooth muscle cells were isolated from basilar and cerebral arteries.

Arteries for intact tissue experiments were cannulated, pressurized with bicarbonate-based physiological saline solution, and superfused with aerated physiological saline solution at 37°C. To block the effects of endothelial-derived nitric oxide, we used the nitric oxide synthase inhibitor N^G-nitro-L-arginine (L-NNA; 300 μmol/L). Intravascular pressure was maintained at 80 mmHg and inner diameter was continuously monitored.

We used the conventional whole-cell patch-clamp technique to voltage clamp freshly isolated arterial myocytes. L-type Ca²⁺ channel sparklets were recorded with a through-the-lens TIRF system with 60X (numerical aperture = 1.49) and 100X (numerical aperture = 1.45) TIRF oil-immersion objectives. All TIRF experiments were performed in the presence of thapsigargin (1 μmol/L). To monitor Ca²⁺ influx, myocytes were loaded with the Ca²⁺ indicator fluo-5F (200 μmol/L) via dialysis through the patch pipette. L-type Ca²⁺ channel sparklets were visualized and recorded at a holding potential of −70 mV with elevated external [Ca²⁺] (20 mmol/L) to facilitate the detection of Ca²⁺ sparklet events and provide fluorescent signals of sufficient amplitude ¹⁷ to permit quantal analysis of Ca²⁺ sparklet activity. Fluo-5F fluorescence was converted to [Ca²⁺] and analyzed as previously reported using a custom automated algorithm ^{17, 21}. Ca²⁺ sparklet activity was quantified ^{17, 18} by calculating the nP_s of each sparklet site, where n is the number of quantal levels, and P_s is the probability that a given Ca²⁺ sparklet site is active. As with previous reports ^{17, 18}, Ca²⁺ sparklet activity was bimodally distributed with sites of low activity (nP_s 0 to 0.2) and high activity (nP_s>0.2). TIRF microscopy was also used to visualize subplasmalemmal ROS generation in isolated arterial myocytes using the cell-permeant ROS indicator 5-(and-6)-chloromethyl-2′7′-dichlorodihydrofluorescein diacetate acetyl ester (DCF; 10 μmol/L).

For our immunofluorescence studies we used a mouse monoclonal PKCα antibody (Abcam) with an Alexa Fluor 488-conjugated rabbit anti-mouse secondary antibody on freshly prepared myocytes. Fluorescence was undetectable in control experiments where either the primary or the secondary antibody was omitted (data not shown). Membrane and cytosolic PKCα-associated fluorescence was quantified by measuring the intensity of pixels above a set threshold. We determined the ratio of plasma membrane to cytosolic PKCα-associated fluorescence and used this as an indicator of PKCα translocation and activity ^{19, 22}.

Normally distributed data are presented as the mean ± standard error of the mean (SEM) with comparisons performed using parametric tests. For bimodally distributed Ca²⁺ sparklet activity (i.e. nP_s) datasets comparisons were performed using non-parametric tests. Arithmetic means of nP_s datasets are indicated in the figures (solid red horizontal lines) for non-statistical purposes. P values less than 0.05 were considered significant and asterisks (*) in the figures indicate a significant difference between groups.

An expanded Materials and Methods section can be found in the online data supplement at http://circres.ahajournals.org.

Results

To test the overall hypothesis that ROS regulate L-type Ca²⁺ channel activity in cerebral arterial smooth muscle we formulated four specific requisite criteria: 1) Exogenous ROS should increase arterial tone in a L-type Ca²⁺ channel-dependent manner; 2) exogenous ROS should stimulate Ca²⁺ sparklet activity in isolated myocytes; 3) inhibition of endogenous ROS generation should limit arterial contraction and prevent stimulation of Ca²⁺ sparklet activity; and 4) endogenous ROS production should colocalize with Ca²⁺ sparklet activity.

Exogenous ROS increase arterial tone in pressurized cerebral arteries

We examined the effects of ROS on arterial smooth muscle function by exposing pressurized cerebral arteries (80 mmHg at 37°C) to the ROS generating system xanthine oxidase (XO; 0.2 mU/mL) plus hypoxanthine (HX; 250 μmol/L; Figure 1). The nitric oxide synthase inhibitor L-NNA (300 μmol/L) was present to preclude contractile responses from ROS-dependent reductions in nitric oxide ²³. Following the development of a stable level of myogenic tone (24.8±2.5 %, n=8 arteries), we exposed arteries to HX followed by XO/HX and monitored changes in luminal diameter. While HX alone had no effect (P>0.05, n=8 arteries), addition of XO reversibly reduced luminal diameter by 18.1±6.9 μm corresponding to a 10.7±4.1 % increase in arterial tone (Figure 1A and B; P<0.05, n=8 arteries). Transient dilations were occasionally observed (in 3 out of 8 arteries) following addition of XO. In the presence of the L-type Ca²⁺ channel antagonist diltiazem (10 μmol/L), which nearly abolished arterial tone (from 23.6±7.1 % to 4.4±1.7 %), XO/HX was without effect (Figure 1C and D; P>0.05, n=3 arteries). Although multiple mechanisms are certainly involved in arterial contractile responses to XO/HX, these data are consistent with the hypothesis that ROS increase arterial tone by stimulating L-type Ca²⁺ channel function.

Figure 1. Reactive oxygen species increase tone in pressurized cerebral arteries.

A, Representative time course showing the luminal diameter of a pressurized (80 mm Hg) cerebral artery exposed to hypoxanthine (HX; 250 μmol/L) followed by xanthine oxidase (XO; 0.2mU/mL) plus HX. B, Plot of the mean ± SEM change in arterial tone (% Δ arterial tone) during HX and XO/HX exposure (n=7 arteries). C, Representative time course showing the luminal diameter of a pressurized (80 mm Hg) cerebral artery exposed to HX followed by XO/HX in the presence of the L-type Ca²⁺ channel blocker diltiazem (10 μmol/L). D, Plot of the mean ± SEM change in arterial tone (% Δ arterial tone) during diltiazem + HX and diltiazem + XO/HX exposure (n=3 arteries). *P<0.05

Exogenous ROS stimulate L-type Ca²⁺ channel sparklet activity in isolated cerebral arterial smooth muscle cells

To test the hypothesis that ROS stimulate L-type Ca²⁺ channel activity, we recorded Ca²⁺ sparklets in isolated voltage-clamped (−70 mV) arterial myocytes before and after XO/HX (2 mU/mL/250 μmol/L). As shown in Figure 2A, XO/HX exposure increased Ca²⁺ sparklet activity. To further characterize the effects of ROS on Ca²⁺ sparklets we constructed Ca²⁺ sparklet amplitude histograms under control conditions and after XO/HX (Figure 2B). Fitting these distributions with a multi-component Gaussian function revealed that Ca²⁺ sparklet amplitudes were quantal and that stimulation by ROS increased the number of quanta activated but not the amplitude of the quantal event (34 nmol/L [Ca²⁺] for control and 36 nmol/L [Ca²⁺] for XO/HX). Note that the Ca²⁺ sparklet quantal amplitudes before and after XO/HX approximate those previously reported for arterial smooth muscle L-type Ca²⁺ channels and for heterologously expressed Cav1.2 L-type Ca²⁺ channels ^{17, 18, 20, 24}. From these data we conclude that ROS increase Ca²⁺ influx by stimulating localized L-type Ca²⁺ channel activity.

Figure 2. Reactive oxygen species increase L-type Ca²⁺ channel sparklet activity in isolated cerebral arterial smooth muscle cells.

A, Representative TIRF images showing Ca²⁺ influx in an arterial myocyte before and after application of XO and HX (2mU/mL and 250 μmol/L, respectively). Traces show the time course of Ca²⁺ influx at the three circled sites before and after XO/HX. B, Ca²⁺ sparklet amplitude histograms before (white) and after XO/HX exposure (red). The solid black lines are best-fits to the control (q=34 nmol/L) and to the XO/HX (q=36 nmol/L) histograms with a multi-component Gaussian function where q is the quantal unit of Ca²⁺ influx (see Detailed Methods in the Online Supplement). C, Plot of Ca²⁺ sparklet site activities (nP_s) before and after XO/HX (n=8 cells). The red solid lines are the arithmetic means of each group and the dashed line marks the threshold for high-activity Ca²⁺ sparklet sites (nP_s≥0.2). D, Plot of the mean ± SEM Ca²⁺ sparklet density (Ca²⁺ sparklet sites/μm²) before and after XO/HX (n=8 cells). *P<0.05

Next we quantified L-type Ca²⁺ channel activity by determining the nP_s of each Ca²⁺ sparklet site, where n is the number of quantal levels detected and P_s is the probability that a given Ca²⁺ sparklet site is active. As evident in the histogram in Figure 2B, Ca²⁺ sparklet activity (nP_s) increased after XO/HX exposure with the number of high-activity Ca²⁺ sparklet sites (nP_s ≥2) increasing from 2 to 16 (Figure 2C; P<0.05, n=8 cells). In addition to increasing Ca²⁺ sparklet activity, XO/HX also increased Ca²⁺ sparklet site density ≈3.5-fold by promoting Ca²⁺ influx at regions previously devoid of activity (Figure 2D; P<0.05, n=8 cells). These data suggest that ROS stimulate localized L-type Ca²⁺ channel function by increasing the occurrence of high-activity Ca²⁺ sparklet sites and by stimulating nascent Ca²⁺ sparklet activity.

Stimulation of L-type Ca²⁺ channel sparklet activity by exogenous ROS is PKCα-dependent

PKCα, which is subject to oxidative activation ²⁵, stimulates L-type Ca²⁺ channel sparklet activity ^{17, 18}. Accordingly, we tested the effects of XO/HX on PKCα activation in arterial myocytes with immunofluorescence by using the level of plasma membrane-associated PKCα as an indicator of activation ^{19, 22} (Figure 3A). Under control conditions (HX; 250 μmol/L), PKCα-associated fluorescence was diffusely cytosolic with few discrete elevations along the plasma membrane. In contrast, following XO/HX exposure (2 mU/mL/250 μmol/L), and indicative of activation, PKCα-associated fluorescence was characterized by prominent focal elevations near (≤1 μm) the plasma membrane (Figure 3B; P<0.05, n=15 cells).

Figure 3. Reactive oxygen species increase L-type Ca²⁺ channel sparklet activity via stimulation of PKC.

A, Representative surface plots of PKCα-associated immunofluorescence in cerebral arterial myocytes exposed to either HX alone (250 μmol/L) or XO and HX (2 mU/mL and 250 μmol/L, respectively). B, Bar plot of the mean ± SEM membrane-to-cytosol PKCα-associated fluorescence ratios in HX- and XO/HX-exposed myocytes (n=15 cells from three independent experiments). C, Representative TIRF images showing Ca²⁺ influx in an arterial myocyte under control conditions and after exposure to XO and HX (2 mU/mL and 250 μmol/L, respectively) in the presence of the PKC inhibitor Gö6976 (100 nmol/L). D, Plot of Ca²⁺ sparklet site activities (nP_s) and mean ± SEM Ca²⁺ sparklet density (Ca²⁺ sparklet sites/μm²) under control conditions and after exposure to XO and HX in the presence of Gö6976 (n=5 cells). *P<0.05

Our PKCα immunofluorescence data suggest that ROS could increase L-type Ca²⁺ channel sparklet activity by activating PKCα. To examine this possibility, we tested the effects of XO/HX on Ca²⁺ sparklets in the presence of the PKC inhibitor Gö6976 (100 nmol/L). In contrast to experiments in the absence of Gö6976 (e.g., Figure 2), XO/HX had no effect on Ca²⁺ sparklet activity or density following inhibition of PKCα (Figure 3C and D; P>0.05, n=5 cells). These data suggest that ROS stimulate L-type Ca²⁺ channel sparklet activity via activation of PKCα.

NADPH oxidase inhibition prevents Ang II-dependent stimulation of L-type Ca²⁺ channels and arterial constriction

Ang II stimulates ROS-generating NADPH oxidase signaling complexes. Therefore, we used Ang II to examine the role of endogenous ROS generation on Ca²⁺ sparklet activity and arterial constriction. As shown previously ^{19, 20}, Ang II (100 nmol/L) increased L-type Ca²⁺ channel sparklet activity (nP_s) and the number of active Ca²⁺ sparklet sites (Figure 4A and B; P<0.05, n=8 cells). This effect was blocked by the AT1 receptor antagonist ZD7155 (500 nmol/L; control nP_s=0.35±0.32, Ang II + ZD7155 nP_s=0.30±0.11, P>0.05, n=4 cells). Inhibition of NADPH oxidase-mediated ROS generation with apocynin (25 μmol/L) ⁹ abolished the effect of Ang II on Ca²⁺ sparklet activity and density (Figure 4C and D; P>0.05, n=6 cells). Unlike PKC inhibition with Gö6976, apocynin prevented stimulation of Ca²⁺ sparklets by Ang II but had no effect on Ca²⁺ sparklet sites previously active under control conditions (P>0.05, n=5 cells). This indicates that the effects of apocynin are not the result of PKC inhibition or by direct blockade of L-type Ca²⁺ channels as suggested for other NADPH oxidase inhibitors (e.g. diphenyleneiodonium ²⁶). Note that Ang II and apocynin (and XO/HX) produced similar results in conventional recordings of macroscopic arterial L-type Ca²⁺ currents (see Online Supplement Figure I).

Figure 4. Inhibition of NADPH oxidase with apocynin prevents angiotensin II-dependent stimulation of L-type Ca²⁺ channel sparklet activity.

A, Representative TIRF images showing Ca²⁺ influx in an arterial myocyte before and after application of angiotensin II (Ang II; 100 nmol/L). Traces show the time course of Ca²⁺ influx at the three circled sites. B, Plot of Ca²⁺ sparklet site activities (nP_s) and plot of mean ± SEM Ca²⁺ sparklet density (Ca²⁺ sparklet sites/μm²) before and after Ang II (n=8 cells). C, Representative TIRF images show Ca²⁺ influx in an arterial myocyte before and after application of Ang II (100 nmol/L) in the presence of the NADPH oxidase inhibitor apocynin (25 μmol/L). Traces showing the time course of Ca²⁺ influx at the three circled sites. D, Plot of Ca²⁺ sparklet site activities (nP_s) and plot of mean ± SEM Ca²⁺ sparklet density (Ca²⁺ sparklet sites/μm²) before after Ang II in the presence of apocynin (n=6 cells). *P<0.05

Ang II produces endothelium-independent arterial constriction via stimulation of AT1 receptors ^{27, 28} (also see Figure 5A). Similar to our Ca²⁺ sparklet experiments, inhibition of NADPH oxidase with apocynin (and PKC with Gö6976) abolished contractile responses to Ang II (1 nmol/L; Figure 5B, C and D; P<0.05, n=3 arteries). Taken together, these data support our hypothesis that generation of endogenous ROS is necessary for AT1-dependent Ang II stimulation of L-type Ca²⁺ channels and arterial constriction.

Figure 5. Inhibition of NADPH oxidase and PKC prevents angiotensin II-dependent constriction of cerebral arteries.

A, B, & C, Representative time courses showing luminal diameters of pressurized (80 mm Hg) cerebral arteries exposed to angiotensin II (Ang II; 1 nmol/L) in the absence (panel A) or presence of the NADPH oxidase inhibitor apocynin (25 μmol/L; panel B) and the PKC inhibitor Gö6976 (100 nmol/L; panel C). D, Plot of the mean ± SEM induced constriction (%) by Ang II in the absence or presence of apocynin and Gö6976 (n=3 arteries). *P<0.05

Ang II stimulates L-type Ca²⁺ channel sparklet activity via local production of ROS

If generation of endogenous ROS is necessary for Ang II-dependent stimulation of L-type Ca²⁺ channels, then sites of Ca²⁺ sparklet activity should colocalize with endogenous sites of ROS production. We used TIRF microscopy to visualize subplasmalemmal generation of ROS in response to Ang II. Isolated myocytes were loaded with the ROS indicator DCF (10 μmol/L). Figure 6A shows TIRF images of DCF fluorescence under control conditions and after application Ang II (100 nmol/L; both in Ca²⁺-free buffer). Interestingly, discrete sites (puncta) of elevated DCF fluorescence were apparent under control conditions (0.006±0.002 puncta per μm²) and increased in number ≈3.5-fold with Ang II (0.021±0.002 puncta per μm²; P<0.05; n=6 cells). The amplitude of the DCF elevations before and after Ang II were not different (P<0.05; n=6 cells). Ca²⁺ sparklet site and ROS puncta densities (0.025±0.005 and 0.021±0.002, respectively) were similar (P>0.05).

Figure 6. Angiotensin II promotes spatially restricted generation of reactive oxygen species in isolated cerebral arterial smooth muscle cells.

A, Representative TIRF images showing punctate elevations of dichlorofluorescein (DCF) fluorescence (indicating reactive oxygen species formation) in an arterial myocyte before and after application of Ang II (100 nmol/L). B, Plot of the mean ± SEM reactive oxygen species (ROS) puncta density (ROS puncta/μm²) before and after Ang II (n=6 cells). C, Plot of the mean ± SEM ROS puncta amplitude (arbitrary units; AU) before and after Ang II (n=6 cells). *P<0.05

Having demonstrated that increased ROS production (as with Ca²⁺ sparklet activity) in response to Ang II is localized, we developed an experimental approach to visualize the spatial distributions of ROS generation (with DCF) and L-type Ca²⁺ channel sparklet activity (with fluo-5F) in the same cell. First, we loaded isolated myocytes with DCF in Ca²⁺ free buffer (as above), placed a patch pipette on the cell, formed a GΩ seal (i.e., established the cell-attached patch configuration), and applied Ang II (100 nmol/L). Upon visualization of sites of punctate elevations in DCF fluorescence (Figure 7A, panel 1, time course i), negative pressure was applied to the interior of the pipette to rupture the plasma membrane (i.e., establish the whole-cell patch configuration) thus dialyzing the cell with fluo-5F. Next, we replaced the Ca²⁺-free external solution with one containing 20 mmol/L Ca²⁺ and monitored for Ca²⁺ sparklet activity. Strikingly, we found that the sites of ROS generation colocalized with high-activity (nP_s=0.56±0.26) Ca²⁺ sparklet sites (Figure 7A, panel 2, time course ii; n=6 sites from 5 cells). While Ca²⁺ sparklet sites were always associated with a preceding site of ROS generation, we did observe three sites of ROS generation that were not associated with subsequent Ca²⁺ sparklet activity (data not shown; n=3 ROS sites in 7 cells).

Figure 7. Local sites of elevated reactive oxygen species generation precede and colocalize with Ca²⁺ sparklet activity in cerebral arterial smooth muscle cells.

A, Representative TIRF images of an arterial myocyte exposed to Ang II showing sequential DCF fluorescence (indicating ROS formation; panel 1) and fluo-5F fluorescence (indicating Ca²⁺ influx; panel 2). DCF fluorescence was obtained in Ca²⁺-free solution prior to cell rupture and dialysis with fluo-5F and subsequent introduction of Ca²⁺ (20 mmol/L). Panel 4 shows an overlay of fluo-5F fluorescence thresholded to isolate Ca²⁺ sparklet activity (red; panel 3) with the DCF fluorescence (green; panel 1) to demonstrate colocalization (yellow) of punctate ROS generation and Ca²⁺ sparklet activity. Traces showing the time course of DCF fluorescence (i.e. ROS generation) at the site circled in panel 1 (i) and the time course of fluo-5F fluorescence (i.e. Ca²⁺ influx) at the site circled in panel 2 (ii). B, Plot of the mean ± SEM distance (in μm) between the peaks of ROS puncta and adjacent Ca²⁺ sparklet sites (n=6 ROS/Ca²⁺ sparklet sites from 5 cells). C, plot of the mean ± SEM cell area imaged (in μm²) during the DCF/fluo-5F experiments (n=5 cells). D, Proposed mechanism by which local generation of ROS stimulate PKC-dependent L-type Ca²⁺ channel sparklet activity in cerebral arterial smooth muscle cells (see Discussion).

To further establish the proximity of punctate ROS generation and Ca²⁺ sparklet activity we thresholded (mean basal fluo-5F fluorescence plus three times its standard deviation) our fluo-5F fluorescence images to isolate Ca²⁺ sparklet fluorescence and merged them (Figure 7A, panel 3) with the DCF images (Figure 7A, panel 1). As evident in the composite image (Figure 7A, panel 4), sites of punctate ROS generation and Ca²⁺ sparklet activity were closely apposed. Indeed, the average distance between the peaks of ROS puncta and Ca²⁺ sparklet sites was less than 1 μm (0.91±0.24 μm; n=6 sites from 5 cells). To place this distance into perspective, the average cell area imaged during these experiments was 99.2±13.7 μm² and the probability, by chance alone, that we would observe 6 ROS puncta and 6 Ca²⁺ sparklet sites ≈1 μm apart in this area is less than 1 in 1,000,000 (see Detailed Methods in the Online Supplement for details). From these data we conclude that sites of ROS generation precede and colocalize with sites of L-type Ca²⁺ channel sparklet activity.

Discussion

In this study we combined TIRF imaging with conventional techniques to test the hypothesis that ROS regulate L-type Ca²⁺ channel activity in cerebral arterial smooth muscle. The major findings in support of this hypothesis are: 1) Exogenous ROS generated by xanthine oxidase constrict pressurized cerebral arteries in an L-type Ca²⁺ channel-dependent manner; 2) exogenous ROS increase PKCα-dependent L-type Ca²⁺ channel sparklet activity in isolated arterial myocytes; 3) generation of endogenous ROS by NADPH oxidase is necessary for stimulation of Ca²⁺ sparklets and contraction by Ang II; and 4) Ang II induces localized sites of ROS production that precede and colocalize with subsequent Ca²⁺ sparklet activity. These observations support a model (see Figure 7D) of local oxidative regulation of Ca²⁺ influx where vasoconstrictors coupled to NAPDH oxidase (e.g., Ang II) induce discrete sites of ROS generation resulting in activation of adjacent PKCα molecules. PKCα activation in turn promotes localized L-type Ca²⁺ channel (i.e., sparklet) activity resulting in increased Ca²⁺ influx and arterial contraction. To the best of our knowledge, our model provides the first experimentally-based mechanistic framework whereby local changes in redox signaling result in changes in Ca²⁺ influx and arterial function.

ROS impairment of endothelial function leads to increased arterial tone ^{23, 29}. Our observation that XO/HX constricts cerebral arteries (presumably via PKC activation) after endothelial nitric oxide synthase inhibition with L-NNA suggests that ROS also increase tone through smooth muscle-specific mechanisms. Activation of arterial smooth muscle PKC by ROS could induce contraction by a minimum of four non-mutually exclusive mechanisms: 1) Increased localized L-type Ca²⁺ channel sparklet activity as characterized here; 2) modulation of other plasmalemmal ion channels (e.g., inhibition of hyperpolarizing potassium channels ³⁰ or activation of depolarizing transient receptor potential channels ³¹); 3) decreased hyperpolarizing ryanodine receptor-dependent Ca²⁺ spark activity ³²; and 4) sensitization of the arterial smooth muscle contractile apparatus to Ca^{2+ 33}. The discrete subcellular nature of Ang II-dependent ROS generation demonstrated here suggests that the relationship between ROS and these diverse regulatory mechanisms could be determined in part by coincident localization of ROS generating and smooth muscle regulatory signaling complexes. In addition, differential subcellular distributions of signaling complexes and ROS generating enzymes could account for contradictory data regarding the effect of ROS on arterial tone (i.e., contraction ³⁴ versus relaxation ³⁵).

PKC stimulation of NADPH oxidase is well-documented ^{36, 37}. However, reciprocal activation of PKC by ROS in arterial smooth muscle, as shown here, has not been reported. Oxidative activation of PKCα leads to sustained cofactor-independent kinase activity ^{25, 38}. Sustained PKC activity is consistent with the observation that established high-activity L-type Ca²⁺ channel sparklet sites are abolished by PKC inhibition ¹⁷. Application of exogenous ROS did not produce a uniform translocation of PKCα to the plasma membrane. Rather, accumulation of PKCα at the plasma membrane was irregular with punctate elevations in PKCα-associated fluorescence (Figure 3). Similarly, the increase in Ca²⁺ sparklet site density by exogenous ROS was far less than predicted if ROS exposure resulted in a generalized non-specific translocation of PKC to the plasma membrane. Importantly, previous work in arterial smooth muscle cells has shown that the scaffold protein AKAP150 is necessary for punctate membrane localization of PKCα and Ca²⁺ sparklet activity ¹⁹. In this context, our observation that ROS exposure produced spatially restricted PKCα-dependent Ca²⁺ sparklet activity suggests that while ROS may induce PKC activation, additional components such as AKAP150 are necessary for efficient targeting of the kinase to the plasma membrane to stimulate localized L-type Ca²⁺ channel activity.

Our imaging of punctate sites of endogenous ROS generation in isolated cells with DCF indicates that stimulation of L-type Ca²⁺ channels by ROS would also be restricted by the localized nature of ROS production and not just by limited membrane targeting of PKCα per se. However, PKC phosphorylation of p47phox is an important step in the initial activation of NADPH oxidase by Ang II ^{9, 39}. Thus, PKC regulation of L-type Ca²⁺ channels likely involves (at least) two events where PKC first stimulates the production of NADPH oxidase-derived ROS which in turn oxidatively activates PKCα resulting in enhanced Ca²⁺ sparklet activity. Our data do not address this hypothesis directly. However, during our ROS imaging experiments the external solution was free of Ca²⁺ and we incubated the cells with the sarcoplasmic reticulum Ca²⁺-ATPase inhibitor thapsigargin. Thus, neither Ca²⁺ entry into the cell or release from of Ca²⁺ from internal stores appear to be necessary for ROS generation in response to Ang II in isolated arterial myocytes.

In combination with supporting data, our sequential ROS/Ca²⁺ imaging experiments provide compelling evidence suggesting that endogenous ROS locally regulate L-type Ca²⁺ channel activity. Using this unique approach we found that ROS production preceded and colocalized with sites of L-type Ca²⁺ channel sparklet activity (respective peaks less than 1 μm apart). It is possible that the molecular components necessary for ROS production and L-type Ca²⁺ channel activity share the same subcellular location but are independent and not functionally linked. However, our data showing that: 1) Exogenous ROS promote PKCα activation and translocation to the plasma membrane; 2) exogenous ROS stimulate PKC-dependent Ca²⁺ sparklet activity and; 3) inhibition of endogenous ROS generation with apocynin abolished stimulation of Ca²⁺ sparklets by Ang II renders this possibility unlikely. Note that whether apocynin is acting as a bone fide inhibitor of NADPH oxidase or as an antioxidant ⁴⁰ does not contradict our overall hypothesis that localized generation of ROS stimulate L-type Ca²⁺ channel activity.

To conclude, in this study we provide compelling evidence in support of the hypothesis that local changes in redox status are transduced to sustained Ca²⁺ influx through L-type Ca²⁺ channels (i.e. high-activity Ca²⁺ sparklets). The implications and questions raised by our observations are intriguing. What molecular components are necessary for the formation of this novel signalosome? Are other signaling pathways subject to or influenced by similar local regulation by ROS? Does localized oxidative activation of L-type Ca²⁺ channels contribute to enhanced Ca²⁺ influx during hypertension? The data presented here may provide a mechanistic starting point for future efforts aimed at answering these important questions.

Novelty and Significance.

What Is Known?

• Reactive oxygen species (ROS) are important signaling molecules in cardiovascular cells.

• Because of their reactive nature and as a mechanism conferring specificity, ROS production is thought to be localized in order for them to function effectively.

• Highly localized L-type calcium channel activity has been observed in arterial smooth muscle cells and shown to contribute to arterial contraction.

What New Information Does This Article Contribute?

• Exogenous ROS increase localized protein kinase C-dependent L-type calcium channel activity in isolated arterial smooth muscle cells and constrict intact arteries in an L-type calcium channel-dependent manner.

• Generation of endogenous ROS by NADPH oxidase is necessary for stimulation of L-type calcium channels and arterial contraction in response to the vasoconstrictor angiotensin II.

• Angiotensin II induces punctate sites of ROS generation that precede and colocalize with L-type calcium channel activity in isolated arterial smooth muscle cells.

Summary.

ROS and calcium are essential components of arterial function under physiological and pathophysiological conditions. However, the relationship between these two signaling modalities is unclear. In this study we investigated the functional and spatial relationship between ROS and L-type calcium channels, a major source of cytoplasmic calcium in arterial smooth muscle. We found that ROS stimulated local sites of L-type calcium channel activity in isolated arterial smooth muscle cells and induced contraction of intact arterial segments. Using a novel imaging-based approach, we visualized punctate sites of endogenous ROS formation in isolated arterial smooth muscle cells. We also show, for the first time, that the spatial distribution of these ROS puncta overlaps with that of local L-type calcium channel activity. These observations indicate that discrete sites of ROS generation are functionally and spatially coupled to local calcium influx through single L-type calcium channels in arterial smooth muscle cells. ROS are widely implicated in the pathogenesis of hypertension. As the relationship between ROS and L-type calcium channels results in increased channel function, our findings should provide mechanistic insight into events underlying increased calcium influx during hypertension and lead to the development of new rational therapies for managing and preventing disease.

Non-standard Abbreviations and Acronyms

Ang II

angiotensin II

DCF

5-(and-6)-chloromethyl-2′7′-dichlorodihydrofluorescein diacetate acetyl ester, HX, hypoxanthine

L-NNA

N^G-nitro-L-arginine

PKCα

protein kinase C alpha

ROS

reactive oxygen species

TIRF

total internal reflection fluorescence

XO

xanthine oxidase