Mitochondria are proposed to be a major oxygen sensor in hypoxic pulmonary vasoconstriction (HPV), a unique response of the pulmonary circulation to low oxygen tension. Mitochondrial factors including reactive oxygen species, cytochrome c, ATP, and magnesium are potent modulators of voltage-gated K+ (Kv) channels in the plasmalemmal membrane of pulmonary arterial (PA) smooth muscle cells (PASMCs). Mitochondria have also been found close to the plasmalemmal membrane in rabbit main PA smooth muscle sections. Therefore, we hypothesized that differences in mitochondria localization in rat PASMCs and systemic mesenteric arterial smooth muscle cells (MASMCs) may contribute to the divergent oxygen sensitivity in the two different circulations. Cellular localization of mitochondria was compared with immunofluorescent labeling, and differences in functional coupling between mitochondria and Kv channels was evaluated with the patch-clamp technique and specific mitochondrial inhibitors antimycin A (acting at complex III of the mitochondrial electron transport chain) and oligomycin A (which inhibits the ATP synthase). It was found that mitochondria were located significantly closer to the plasmalemmal membrane in PASMCs compared with MASMCs. Consistent with these findings, the effects of the mitochondrial inhibitors on Kv current (IKv) were significantly more potent in PASMCs than in MASMCs. The cytoskeletal disruptor cytochalasin B (10 μM) also altered mitochondrial distribution in PASMCs and significantly attenuated the effect of antimycin A on the voltage-dependent parameters of IKv. These findings suggest a greater structural and functional coupling between mitochondria and Kv channels specifically in PASMCs, which could contribute to the regulation of PA excitability in HPV.
- pulmonary artery
- vascular smooth muscle cells
- mesenteric artery
- K+ channel activation
- K+ channel inactivation
- confocal imaging
- patch-clamp technique
the pulmonary circulation represents a vascular bed with unique properties. Under physiological conditions this vasculature is a highly oxygenated, low-pressure system. When oxygen tension drops during, for example, chronic obstructive pulmonary disease, a response known as hypoxic pulmonary vasoconstriction (HPV) occurs, diverting the blood flow to the higher oxygenated areas to match ventilation to perfusion. The mechanisms involved in this response are complex and currently incompletely understood (14, 48, 52). Such responses do not occur in the systemic circulation, which dilates when oxygen tension is reduced to maximize blood supply to all parts of the body.
Smooth muscle cell contraction is critical to vasoconstriction; however, to date no definitive mechanism can account for how these cells sense changes in oxygen tension or how increased contractility occurs. The response is considered to be a complex, multifactorial phenomenon ultimately contributing to elevated intracellular Ca2+ concentration ([Ca2+]i), essential for formation of the calcium-calmodulin complex and activation of the myosin light chain kinase. Hypoxic inhibition of voltage-gated K+ (Kv) channels, which results in membrane depolarization and voltage-dependent Ca2+ influx in pulmonary arterial (PA) smooth muscle cells (PASMCs), was originally suggested as one of the important factors leading to increased [Ca2+]i during HPV (3, 36, 47, 51). Currently, mitochondria, the main consumers of cellular oxygen during respiration, are considered to be one of the putative oxygen sensors in pulmonary smooth muscle cells (18, 22, 50, 53), and it has been postulated that Kv channels may be regulated by an effector of mitochondrial origin (3, 22). This idea is supported by the observations that 1) inhibitors of the mitochondrial electron transport chain (mETC) such as rotenone and antimycin A, acting at complexes I and III, respectively, and the mitochondrial uncoupler FCCP mimic the effect of hypoxia on Kv current (IKv) in PASMCs (3, 57) and ductus arteriosus (23) and 2) mice lacking the active gp91phox subunit of the membrane-bound NADPH oxidase preserve the effect of hypoxia on IKv in PASMCs (5). Numerous products of mitochondrial metabolism have also been shown to activate Kv channels, including cytochrome c (34) and hydrogen peroxide in pulmonary and coronary arterial smooth muscle cells (39, 40), while reducing agents inhibit Kv channels (27, 28, 38). In addition, we recently reported (9) that that the voltage-dependent characteristics of IKv in PASMCs are potently modulated via a mitochondrion-mediated Mg2+-dependent mechanism.
Although products of mitochondrial metabolism are currently established to have effects on Kv channels, the exact reasons why these mediators and channels existing in all circulatory beds should behave differently in the pulmonary circulation in response to hypoxia remain unknown. One possibility could be functional differences of the mitochondria present in different vascular beds (22, 37). Another possibility could be that the intracellular locality of mitochondria in relation to other organelles and the cellular membrane may vary in the different vascular beds. Using electron microscopy imaging, Vallières et al. (46) demonstrated the presence of mitochondria close to the plasmalemmal membrane in rabbit main PA smooth muscle sections. Such close proximity between mitochondria and the plasmalemmal membrane, if it specifically exists in PASMCs, could be responsible for the opportunistic detection of a mitochondrion-dependent mediator by Kv channels.
No systematic comparison of mitochondrial distribution in smooth muscle cells (SMCs) isolated from resistance PAs and systemic arteries has yet been performed. Therefore, the main aim of this study was to investigate the differences in cellular localization of mitochondria in PASMCs and mesenteric arterial SMCs (MASMCs, as an example of systemic circulation) with fluorescence confocal imaging. The functional differences between the inhibition of mitochondria and IKv characteristics in both cell types were investigated with the patch-clamp technique using the inhibitor of complex III of the mETC antimycin A and the inhibitor of the mitochondrial F0/FI ATP synthase oligomycin. These inhibitors were chosen because they have a most pronounced effect on IKv in rat PASMCs (9).
MATERIALS AND METHODS
All standard chemicals were purchased from BDH Merck or Fisher. Collagenase type XI and type P were obtained from Sigma and Roche, respectively. Papain, dithiothreitol, mETC inhibitor antimycin A (a mixture of antimycins from Streptomyces species), and ATP synthase inhibitor oligomycin were all obtained from Sigma; cytochalasin B was purchased from Calbiochem. Brefeldin A BODIPY 558/568, MitoTracker Green FM, di-8-ANEPPS, and BAPTA-AM were purchased from Invitrogen.
Male Wistar rats (225–300 g) were killed by cervical dislocation as approved by the local United Kingdom Home Office inspector, in accordance with Schedule 1 as prescribed and approved by the Home Office and the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health (NIH Pub. No. 85-23, revised 1996). Small intrapulmonary arteries (3rd–5th order) were microdissected. Isolation of PASMCs (using 1 mg/ml type XI collagenase, 0.5 mg/ml papain, and 1 mM dithiothreitol and 20 min incubation at 37°C) were performed as previously described (43). SMCs from small mesenteric arteries (MA) (3rd–4th order) (MASMCs) were isolated in a similar manner except that 2 mg/ml collagenase and 1 mg/ml papain were used and the tissue incubation in the enzyme solution was increased to 30 min.
Small aliquots of the suspension of freshly isolated PASMCs or MASMCs were transferred to the experimental chambers and diluted with a saline buffer composed of (mM) 120 NaCl, 6 KCl, 0.5 CaCl2, 1.2 MgCl2, 12 glucose, and 10 HEPES, pH adjusted to 7.4 with NaOH. The sarcoplasmic reticulum (SR) in the SMCs was stained by 30-min incubation of the myocytes with 2 μM brefeldin A BODIPY 558/568 [absorbance (Abs)/emission (Em) = 559 nm/568 nm], which was previously demonstrated to stain the SR in rat gastric myocytes (54). The validity of this staining was further confirmed in myocytes in which intracellular calcium stores were visualized with the low-affinity [kd(Ca) = 42 μM] fluorescent Ca2+ indicator fluo-3FF (12). This approach was demonstrated previously to reveal a well-developed subplasmalemmal SR network in both vascular (11) and visceral (12) SMCs.
Mitochondria were stained by 30-min incubation of SMCs with 1 μM MitoTracker Green FM (Abs/Em = 490 nm/516 nm) (24). The plasmalemmal membrane was labeled by incubating the cells for 3 min with 10 μM di-8-ANEPPS (Abs/Em = 488 nm/605 nm) (41). An experimental chamber containing cells was placed on the stage of an Axiovert 100 M inverted microscope attached to a LSM 510 laser-scanning unit (Zeiss, Oberkochen, Germany). Confocal imaging was performed with a Zeiss plan-Apochromat 40× 1.3 numerical aperture oil-immersion objective. The SCSi interface of the confocal microscope was hosted by a Pentium PC (32-bit Windows NT 4.0 operating system) running LSM510 software (Zeiss). To visualize the three-dimensional distribution of the SR and mitochondria in the cell, z-sectioning (series of 40 x-y images taken from a confocal optical section <0.6 μm and with a z-step of 0.3 μm) was performed. Brefeldin A BODIPY fluorescence was excited by the 488-nm line of a 200-mw argon ion laser (Laser-Fertigung, Hamburg, Germany), and the emitted fluorescence was detected with 505- to 550-nm band-pass emission filter. MitoTracker Green fluorescence was excited by the 543-nm line of a 5-mw HeNe ion laser (Laser-Fertigung), and the emitted fluorescence signal was captured at wavelengths >560 nm. Di-8-ANEPPS fluorescence was excited by the 458-nm line of a 200-mw argon ion laser, and the emitted fluorescence was captured at wavelengths >585 nm. For all desired laser lines the illumination intensity was set with an acoustooptical tunable filter. Averaging of two frames was used to reduce noise. The gains were such that the red and green fluorescent signals were approximately of the same intensity but below the saturation of the photomultiplier. To avoid any bleed-through, imaging was performed with the multitrack mode of an LSM 510 (sequential capturing of images with 2 separate photomultipliers). A threshold was set to remove subsignal noise from the images. The adequacy of the imaging protocol applied to the double-labeled myocytes was confirmed by control experiments on the single-labeled cells. It is worth noting that, although the z-sectioning was performed through the entire cell volume, the distribution of the SR and mitochondria was analyzed only in x-y images taken from the half of each SMC located closer to the objective lens. This shortens the length of the optical path and thus minimizes erroneous measurements caused by waveguide dispersion of light, which occurs when the wave propagates through any inhomogeneous structure such as a cell.
Cells were placed in a chamber with a volume of 100–200 μl and continually superfused (∼1 ml/min) with a physiological saline solution (PSS) or a test solution via a five-barrel pipette at room temperature. PSS contained (mM) 140 NaCl, 5 KCl, 1.5 CaCl2, 1.2 MgCl2, 10 HEPES, and 10 glucose, pH = 7.2. Control pipette solution contained (mM) 140 KCl, 0.5 MgCl2, 10 HEPES, 10 EGTA, and 0.5 CaCl2, pH = 7.2, and was used for recording unless otherwise stated. Cells were dialyzed with pipette solution for 5 min before currents were recorded. The effects of inhibitors were studied a minimum of 5 min after addition. Paxilline (1 μM) and glibenclamide (10 μM) to block Ca2+-activated and ATP-sensitive K+ currents, respectively, were present in all extracellular recording solutions, isolating the voltage-dependent K+ channels from other K+ conductances as previously described (43). To evaluate cell size, whole cell capacitance transient currents were measured with a 10-mV hyperpolarizing step after stable whole cell access was achieved in each cell. Cell membrane capacitance (Cm) was then calculated as an area under a capacitance transient and expressed in picofarads.
IKv amplitude was measured with a 200-ms voltage step applied from the holding potential of −80 mV to membrane potentials between −100 and +50 mV in 10-mV increments with a frequency of 0.1 Hz. After membrane depolarization cells were repolarized to −20 (PASMCs) or 0 (MASMCs) mV for 160 ms to measure tail currents, which were used to assess changes in the steady-state activation of IKv. A more positive step potential in MASMCs was required to increase the size of tail currents because of the significantly smaller current amplitude in these cells. Current-voltage (I-V) curves were constructed from the tail current measures 2–3 ms after repolarization and fitted with the following equation: where Va and ka are the half-activation potential and the e-fold steepness of the activation dependence (the slope factor of activation), respectively, and Vm is membrane potential. IKv block was calculated as a percentage of the current at +50 mV in control conditions from the same cell.
IKv inactivation was measured with a two-pulse protocol consisting of a 10-s conditioning pulse stepping in 10-mV increments from −100 to +40 mV, followed by a test potential to +60 mV for 200 ms. Current at the end of the test pulse was measured, normalized to the maximal current recorded, and plotted versus conditioning potential. The resulting I-V curve was fitted to the following equation: where Vh and kh are the half inactivation potential and the e-fold steepness of the inactivation dependence (the slope factor of inactivation), respectively, and A denotes the noninactivating component of the current.
Data analysis and statistics.
Data were analyzed and presented with pCLAMP 8 (Axon Instruments), Microsoft Excel, and Microcal Origin 6.0 software. Data are expressed as means ± SE. Statistical comparisons were performed with a paired or unpaired two-tailed t-test, with P < 0.05 deemed significant unless stated otherwise. When normality test failed (Shapiro-Wilk test) the Wilcoxon signed rank test and the Mann-Whitney test were used for comparison of paired and unpaired data, respectively. Paired statistical tests were used to compare the difference in parameters obtained in the control condition and in the presence of mitochondrial inhibitor in the same cell, whereas unpaired tests were used to compare differences between two groups of cells (e.g., PASMCs vs. MASMCs).
Differences in spatial distribution of mitochondria in PASMCs and MASMCs.
The distribution of mitochondria [stained with MitoTracker Green FM (24)] in relation to the SR [stained with brefeldin A BODIPY 558/568 (54)] was compared in SMCs isolated from rat small PAs and MAs. The association between mitochondria and SR elements was assessed with the confocal z-sectioning protocol described in materials and methods. This approach revealed that in PASMCs, on movement from the cell bottom toward the middle of the myocyte depth (Fig. 1A), the fluorescent signal from MitoTracker Green appeared in the x-y confocal optical slices (Fig. 1C) before that of brefeldin A BODIPY (Fig. 1B). In PASMCs, in individual x-z and y-z optical cross sections of the myocyte MitoTracker Green fluorescence (mitochondria) was detected distinctly closer to the cell periphery than that of brefeldin A BODIPY (SR) (Fig. 1, D and E). In MASMCs, however, brefeldin A BODIPY fluorescence staining the SR appeared before or simultaneously with MitoTracker Green fluorescence (Fig. 2B). Such opposing juxtaposition between the SR and mitochondria in MASMCs compared with PASMCs becomes more evident from the comparison of individual x-z and y-z optical cross sections of the myocyte MitoTracker Green (mitochondria) and brefeldin A BODIPY (SR) fluorescent signals (Fig. 2, C and D).
To quantify these differences, the total area occupied by pixels showing MitoTracker Green and brefeldin A BODIPY fluorescence was analyzed in four PASMC and MASMCs and plotted against the optical slice number in Fig. 3. This analysis clearly shows the existence of an area close to the surface of PASMCs that is enriched with mitochondria (proportional to the area with MitoTracker Green fluorescence), while in MASMCs a similar area is predominantly occupied with the SR (proportional to the area with brefeldin A BODIPY fluorescence). Figure 3 also shows diagrammatic representations of these areas, comparing PASMCs with MASMCs.
The existence of the subplasmalemmal pool of mitochondria in PASMCs, but not in MASMCs, was further studied by double staining of both types of cells with MitoTracker Green and di-8-ANEPPS, selective fluorescent probes for mitochondria and the plasmalemmal membrane, respectively (Fig. 4). Figure 4A shows x-y confocal images taken though the middle of a representative PASMC. As can be seen clearly in Fig. 4A, MitoTracker Green fluorescence often appeared to be “fused” (within ∼200 nm, the limit of lateral resolution of our confocal microscope) with di-8-ANEPPS fluorescence shown in red (plasmalemma). Conversely, in MASMCs, a clear gap between the two fluorescent signals was observed (Fig. 4B). To quantitatively compare relative distances between mitochondria and the plasma membrane in the two cell types, the spatial profiles of fluorescent signals recorded for di-8-ANEPPS (red) and MitoTracker Green (green) were constructed across the shortest distance between the center of each mitochondrion and the plasma membrane (shown by yellow lines in Fig. 4, C and D, for PASMC and MASMC, respectively). Histograms of distribution of mitochondria according to their distance from the plasma membrane measured in six PASMCs and six MASMCs are compared in Fig. 4E. The histogram bin size was set to 0.5 μm, which is 2.5-fold greater than the lateral optical resolution of our imaging system determined with 0.2-μm fluorescent beads. This analysis clearly shows that in PASMCs the majority of peripheral mitochondria are localized between 0.5 and 1.5 μm from the plasma membrane, compared with ∼1.5–3.5 μm in MASMCs. On average, the mean distance between peripheral mitochondria and the plasma membrane calculated as described above was twofold greater in MASMCs (2.4 ± 0.05 μm, n = 405) than in PASMCs (1.2 ± 0.05 μm, n = 235) (P < 0.0001; Fig. 4F).
Comparison of IKv characteristics in PASMCs and MASMCs.
In view of the locality of the mitochondria to the plasma membrane demonstrated above, it could be postulated that the mitochondrion-mediated effects on Kv channels that occur in PASMCs may exist in MASMCs but would be less effective. Consequently, IKv was isolated pharmacologically (as described in materials and methods), and its electrophysiological properties were compared in PASMCs and MASMCs under identical experimental conditions.
With the IKv activation protocol featured in Fig. 5A, whole cell IKv was recorded in both PASMCs and MASMCs. The amplitude of the current recorded from a representative MASMC (Fig. 5C) was considerably smaller than in a PASMC (Fig. 5B). On average, the differences in IKv amplitude were significantly different between −40 and +50 mV. For example, at 0 and +50 mV, the mean amplitude of IKv was equal to 69 ± 5 and 270 ± 17 pA (n = 60), respectively, in MASMCs and 446 ± 20 and 1,369 ± 59 pA (n = 184), respectively, in PASMCs (P < 0.0001). On the other hand, the size of MASMCs, calculated from Cm, was significantly larger than that of PASMCs [16.3 ± 0.6 pF (n = 60) vs. 11.1 ± 0.3 pF (n = 184); P < 0.0001]. The smaller IKv and the larger cell size in MASMCs result in significant differences in the whole cell current density in the two cell types (Fig. 5D), which was 10- to 18-fold greater between −30 and 0 mV and 8- to 9-fold greater between 0 and +50 mV in PASMCs compared with MASMCs.
Comparison of the steady-state activation of IKv measured from the normalized I-V curves (which were derived from the IKv tail current measurements) from the representative PASMC and MASMC shown in Fig. 5, B and C, indicates that IKv activation in MASMC was shifted by ∼6 mV to more positive membrane voltages and was steeper than in PASMC (Fig. 5E). These differences in the activation parameters were significant (Table 1).
Comparison of IKv inactivation measured with the voltage protocol shown in Fig. 6A from the representative PASMC (Fig. 6B) and MASMC (Fig. 6C) demonstrates that IKv inactivation in PASMCs is relatively slow and incomplete compared with that in MASMCs. The normalized IKv recorded during the test pulse was plotted against the prepulse potential, and resulting dependencies were fitted to the equation described in materials and methods, yielding the parameters of IKv inactivation, the half-inactivation potential Vh, the slope factor kh, and the noninactivating component A (Fig. 6D). Statistical comparison of these parameters showed that IKv in MASMCs is inactivated at more negative membrane potentials, its dependence on the conditional prepulse potential is steeper, and it has less noninactivated current than IKv in PASMCs; these differences in inactivation characteristics were significant (Table 1).
Therefore, direct comparison of the steady-state activation and inactivation of IKv measured under identical experimental conditions in both cell types clearly demonstrates that PASMCs have significantly greater current amplitude and more available and less inactivated Kv channels in the negative range of membrane potentials compared with MASMCs.
Mitochondrial inhibition regulates IKv more potently in PASMCs.
Inhibition of the mETC was previously shown to inhibit IKv in PASMCs (3, 23, 57). We recently showed (9) that steady-state activation and IKv block were significantly affected by mitochondrial inhibition in PASMCs, with the most pronounced effects observed for antimycin A (inhibiting the mETC at complex III) and oligomycin (blocking the ATP synthase). These two mETC inhibitors therefore were chosen to verify whether functional coupling between mitochondria and Kv channels in PASMCs is greater than in MASMCs, as would be predicted based on the difference in the distribution of mitochondria in the two cell types as demonstrated in Figs. 1–4.
The different sites of action of antimycin A and oligomycin in mitochondria are schematically depicted in Fig. 7A. The differences in the mean parameters for IKv activation and inactivation due to mitochondrial inhibition with antimycin A and oligomycin in PASMCs and MASMCs are comprehensively analyzed in Table 2, while Fig. 7,B–D, compares the relative changes in IKv activation, inactivation, and block in PASMCs to those in MASMCs. Although both inhibitors caused significant negative shifts in IKv inactivation (Fig. 7C) and significant decreases in IKv amplitude at +50 mV (Fig. 7D) in both cell types, all these effects were significantly smaller in MASMCs than in PASMCs (Fig. 7, B–D). Similarly, antimycin A had a significantly greater effect in PASMCs than in MASMCs (Fig. 7B). Oligomycin, on the other hand, caused significant differences in IKv activation only in PASMCs and not in MASMCs (Fig. 7B and Table 2). Furthermore, inhibitor-induced changes in other parameters, such as ka, kh, and A, were less pronounced and overall not significantly different in MASMCs from those in PASMCs (Table 2). Thus differences in potency of the effects of the two different mitochondrial inhibitors in two cell types correlate with the difference in mitochondrial distribution described above.
Disruption of cytoskeleton with cytochalasin B alters both peripheral mitochondria distribution and antimycin A-dependent modulation of IKv.
A link between mitochondria and the cytoskeleton has been demonstrated previously in cardiomyocytes (2), neurons (25), human coronary arterioles (19), and pulmonary endothelial cells (1) but not in PASMCs. If the cytoskeleton is involved in juxtaposition of mitochondria and the plasma membrane in PASMCs, then disruption of the cytoskeleton should alter the association between mitochondria and the plasma membrane and at the same time attenuate the effect of the mitochondrial inhibitors on IKv. The effect of treatment of PASMCs with 10 μM cytochalasin B, a disrupter of F-actin microfilaments (1, 17, 19), on distribution of mitochondria was visualized with MitoTracker Green and confocal imaging. Figure 8, A and B, compare distribution of mitochondria (shown in green) in relation to the plasma membrane (stained with di-8-ANEPPS and shown in red) in a single PASMC before and 10 min after incubation with cytochalasin B, respectively. To reduce possible errors resulting from changes in the cell shape due to a disruption of the cytoskeleton, partially contracted PASMCs were used in these experiments. As can be seen in Fig. 8, cytochalasin B did not significantly change the cell shape; however, the pattern of green fluorescence across the whole cell was altered, suggesting that the mitochondria are associated with F-actin filaments in PASMCs. Importantly, the distribution of mitochondria at the submembrane regions was altered, as can be seen at an increased magnification in Fig. 8, A and B, bottom. Similar effects were observed in the other eight PASMCs studied.
The above results suggest that if Kv channels in the plasma membrane and peripheral mitochondria are functionally coupled in PASMCs as the electrophysiological evidence indicates (Fig. 7; Ref. 9), then the cell treatment with cytochalasin B should also modulate the effect of the mitochondrial inhibitors on IKv characteristics. To test this prediction, the effect of 10 μM cytochalasin B on antimycin A-induced changes in IKv activation, inactivation, and block was assessed. Cytochalasin B caused a decrease in IKv amplitude by 35.7 ± 6.1% (n = 5, P < 0.01) but itself did not have any significant effect on the steady-state activation or inactivation parameters of IKv (Table 3). Importantly, the effect of antimycin A on IKv activation was completely abolished (Fig. 8C) and antimycin A-induced changes in IKv inactivation were significantly attenuated (Fig. 8D) in the presence of cytochalasin B. Thus the ability of cytochalasin B to inhibit the antimycin A-induced changes in the voltage-dependent IKv characteristics correlates well with cytochalasin B-mediated changes in mitochondrial distribution in the vicinity of the plasma membrane, providing direct evidence for close functional coupling between Kv channels and peripheral mitochondria in PASMCs. Pretreatment of cells with the cytoskeletal disruptor did not, however, significantly change the antimycin A-mediated inhibition of IKv (Fig. 8E), suggesting that other factors are likely to contribute to the IKv block caused by the mitochondrial inhibitors.
Intracellular Mg2+ is involved in antimycin A-mediated changes in IKv inactivation.
We recently demonstrated (9) that dialysis of PASMCs with Na2ATP and EDTA, which are more potent chelators of Mg2+ than Ca2+, compared with EGTA present in the pipette solution, significantly attenuated the effects of the mitochondrial inhibitors on IKv activation and IKv block. Considering that the effect of the mitochondrial inhibitors on the steady-state inactivation of IKv has not been studied previously, it is important to demonstrate whether changes in IKv inactivation induced by the mitochondrial inhibitors are also mediated by intracellular Mg2+. To test this possibility, the effect of antimycin A and oligomycin was studied in cells dialyzed with pipette solutions containing 5 mM EDTA or 5 mM Na2ATP, respectively. Figure 9, A and B, clearly demonstrate that the relative changes in Vh caused by both inhibitors were significantly reduced under these conditions.
We previously found (9) that cell dialysis with increased Mg2+ concentration also mimicked the effects of the mitochondrial inhibitors on IKv activation and block. When the steady-state inactivation of IKv was measured in cells dialyzed with 5 mM instead of 0.5 mM MgCl2 in the pipette solution but in the absence of mitochondrial inhibitors, a significant leftward shift by ∼7 mV in the inactivation dependence was observed. The mean Vh was changed from −23.9 ± 1.1 mV for 0.5 mM MgCl2 (n = 73) to −30.7 ± 2.8 mV for 5 mM MgCl2 (n = 12, P < 0.026). The noninactivating component A of IKv was also decreased under this condition, from 0.28 ± 0.01 (0.5 mM) to 0.21 ± 0.02 (5 mM) (P < 0.03). No significant changes in kh were observed [11.2 ± 0.4 (n = 72) in 0.5 mM MgCl2 vs. 10.4 ± 0.5 mV (n = 12) in 5 mM MgCl2]. Furthermore, antimycin A-induced changes in Vh were significantly attenuated (Fig. 9C), whereas changes in kh and A remained similar (compare Tables 2 and 4) under these conditions.
Subplasmalemmal localization of mitochondria provides structural basis for functional interaction with Kv channels in PASMCs.
Although the amount of intracellular diffusible mediators in close vicinity to the Kv channel cannot be precisely quantified, the overall local cellular changes could be substantial, providing that diffusion processes in the vicinity of the Kv channels are limited. The existence of a submembrane compartment formed by the SR and called the “superficial barrier” has been previously proposed in systemic vascular SMCs [recently reviewed by Poburko et al. (35)]. Indeed, a tight juxtaposition between the plasmalemmal membrane and the SR has been directly visualized in vascular (11) and visceral (12) SMCs. In myocytes isolated from intralobar rat PAs, we previously demonstrated (42) the presence of a restricted diffusion space by measuring the changes in the K+ gradient during the development of IKv. This phenomenon has not been reported in other vascular SMCs. Comparison of the relative distribution of mitochondria and the SR in PASMCs and MASMCs clearly demonstrates the presence of a distinctive population of mitochondria close to the plasmalemmal membrane and relatively separate from the SR in PASMCs but not in MASMCs, where mitochondria are more closely associated with the SR (Figs. 1–4). This novel and important observation demonstrates the presence of a subplasmalemmal pool of mitochondria in PASMCs that are potentially functionally associated with the Kv channels, as our electrophysiological evidence shows. Such compartmentalization, similar to that formed by the SR in other types of vascular SMCs (35), could be primarily responsible for the lack of sensitivity of IKv to hypoxia in mesenteric arteries previously reported by others (31, 56, 58). Notably, close association between the plasma membrane and mitochondrion has previously been shown by electron microscopy (46). Although the full physiological significance of the existence of the distinctive subplasmalemmal population of mitochondria in PASMCs remains to be established, it is possible that a close juxtaposition of the plasma membrane and mitochondria in PASMCs might be important in polarization of mitochondrion-mediated signals [e.g., reactive oxygen species (ROS) release] in a similar manner to the “mitochondrial belt” involvement in cytosolic Ca2+ signal polarization in pancreatic acinar cells (45).
The ability of the F-actin filament disrupter cytochalasin B to alter mitochondrial distribution in PASMCs suggests that the cytoskeleton contributes to cellular organization of mitochondria in PASMCs. Whether mitochondria are similarly associated with the cytoskeleton in MASMCs remains to be established. Interestingly, inhibition of F-actin filament polymerization with cytochalasin suppressed mitochondrial ROS production induced by stretch in bovine pulmonary endothelial cells (1) and by flow in human coronary arterioles (19). Whether the cytoskeleton is involved in the generation of ROS in PASMCs has not been yet investigated. Our data clearly suggest that in PASMCs mitochondria are spatially arranged in the distinctive layer separate from the SR, forming a compartment with restricted diffusion in the subplasmalemmal space, thus providing a structural basis for a better functional coupling between mitochondria and Kv channels in PASMCs than in systemic arterial myocytes. The correlation between the disruption of the distribution of peripheral mitochondria by cytochalasin B and significant attenuation of the effects of antimycin A on IKv activation and inactivation directly supports this conclusion.
Mechanisms involved in Kv channel regulation by mitochondria in PASMCs.
The electrophysiological evidence presented in Fig. 9, demonstrating that addition of Mg2+-chelating agents into the pipette solution inhibits changes in Vh caused by antimycin A and oligomycin, clearly suggests that an increase in intracellular concentration of Mg2+ plays a key role in the modulation of IKv inactivation by mitochondrial inhibitors. The ability of the increased pipette Mg2+ concentration first to mimic and then to attenuate the effect of antimycin A (Fig. 9C) also supports this notion. We previously demonstrated (9) that similar pipette interventions significantly attenuated changes in the IKv activation and IKv block caused by antimycin A and oligomycin in PASMCs, thus suggesting that changes in mitochondrion-mediated changes in intracellular Mg2+ concentration represent one of the major mechanisms controlling the functional activity of Kv channels. On the other hand, the failure of cytochalasin B to attenuate IKv block, while inhibiting changes in IKv activation and inactivation (Fig. 8, C–E), suggests the presence of other factors (e.g., ROS) that may influence the ability of Mg2+ to interact with Kv channels. The identification of these factors and the establishment of their importance in mitochondrion-mediated and Mg2+-dependent modulation of IKv in PASMCs in normoxic and hypoxic conditions will entail further experimental work.
Stronger functional coupling exists between mitochondria and Kv currents in PASMCs than in MASMCs.
To analyze the differences in functional coupling between mitochondria and Kv channels in PASMCs and MASMCs, the inhibitors of complex III (antimycin A) and of the ATP synthase (oligomycin) were used because, as we previously reported (9), their effects on IKv activation and IKv block in PASMCs are substantially stronger compared with other mETC inhibitors (rotenone, myxothiazol, and cyanide) and the mitochondrial uncoupler CCCP. Comparison of the effects of antimycin A and oligomycin on IKv activation and IKv block in the two cell types clearly demonstrated significantly lesser effects in MASMCs than in PASMCs (Fig. 7, B and D). Notably, the effect of oligomycin on IKv activation was not significant in MASMCs. In addition, we also compared the effect of the two inhibitors on IKv inactivation, which represents another important voltage-dependent characteristic of Kv channels that, together with the steady-state activation, determines the availability of Kv channels at membrane potentials negative to 0 mV (26). Similar to IKv activation and IKv block, changes in IKv activation induced by both inhibitors were also significantly smaller in MASMCs than in PASMCs (Fig. 7C), as would be predicted from the different distribution of mitochondria in PASMCs and MASMCs.
The tighter spatial and functional coupling between Kv channels and mitochondria, one of the proposed oxygen sensors (18, 49), in PASMCs can contribute to the difference in sensitivity of IKv to hypoxia in the pulmonary and systemic arterial SMC reported previously (22, 31, 56, 58). This could be further enhanced by the difference in the functional activity of mitochondria. Thus Michelakis et al. (22) demonstrated a significantly greater rate in basal production of mitochondrial ROS in PASMCs than in renal arterial SMCs, while Rathore et al. (37) found that mitochondrial ROS were selectively increased by hypoxia in PASMCs but not in MASMCs. It is currently difficult to directly measure changes in the concentration of mitochondrial ROS or other mitochondrial metabolites such as ATP, cytochrome c, or magnesium ions [all of which modulate the activity of Kv channels in PASMCs (3, 8, 9, 34, 38, 59)] in submembrane compartments in intact cells. In this instance, ion channel characteristics, which can be measured, quantified, and even calibrated with different intracellular solutions, can serve as useful markers. Indeed, ATP-sensitive K+ (KATP) channels and Ca2+-sensitive K+ (BKCa) channels have been studied to assess changes in ATP concentration and [Ca2+]i, respectively, directly beneath the plasma membrane (13, 44). Similarly, as our results demonstrate, voltage-dependent parameters of IKv, a dominant K+ current in PASMCs, can serve as useful markers of changes in the confined subplasmalemmal environment in close vicinity to the Kv channel in PASMCs.
Distinct differences in Kv channel activity between PASMCs and MASMCs.
To investigate mechanisms of HPV, which are specific to the pulmonary circulation, MAs are often used as a negative control. Although the voltage-dependent characteristics of IKv, such as steady-state activation and inactivation, have been reported previously in rat PASMCs (27, 43) and MASMCs (15, 21, 30), no direct comparison between the two cell types has been performed under identical experimental conditions to determine whether there is a significant difference in IKv properties in the two cell types. Also, in this study, IKv was isolated in both PASMCs and MASMCs by elimination of BKCa and KATP currents with paxilline and glibenclamide, respectively. In other studies a low extracellular Ca2+ concentration (15, 21) and a high concentration of ATP in the pipette (15, 21, 30), respectively, were used in MASMCs for this purpose. It is noteworthy that the amplitude and the parameters for activation and inactivation of IKv measured under our experimental conditions in MASMCs were not dissimilar to those reported by others. For example, the mean midpoints for IKv activation and inactivation of −10 and −39.4 mV (30), −11 and −40 mV (21), and −3.5 and −33.9 mV (15) are comparable with the −9.4 and −37 mV, respectively, we report in the present study (Table 1), suggesting that that our experimental conditions and pharmacological agents used for current isolation do not have a substantial effect on the voltage-dependent properties of IKv in MASMCs.
Comparison of the IKv amplitude, cell size, and steady-state activation and inactivation characteristics of the current measured in identical experimental conditions revealed significant differences in all three groups of characteristics in PASMCs and MASMCs. The most significant differences were found in the whole cell current amplitude (larger in PASMCs) and the cell size (larger for MASMCs), giving between 8- and 18-fold differences in the range of membrane potential between −30 and +50 mV of cells with PASMCs. Also, the midpoint for IKv inactivation was significantly shifted to more negative potentials by ∼13 mV in MASMCs compared with PASMCs. Although a relatively small shift to more positive membrane potentials (∼4 mV) was seen in the midpoint for IKv activation, the effect was nevertheless significant. Finally, the differences in the slope factors for both activation and inactivation dependencies were also significantly different between PASMCs and MASMCs.
Since native Kv channels are believed to be heteromultimers that are formed by more than one Kv1.x α-subunit in both PASMCs (6, 7, 33, 60) and MASMCs (10, 20, 30, 55), the differences in the voltage-dependent characteristics may be due to a different heteromultimeric assembly of the Kv channels in the two cell types. In addition, differences in levels of expression or isoform of accessory Kv β-subunits, found in both cell types (7, 10, 30, 33, 60), could also contribute to the differences in IKv parameters we observed between two cell types. Also, the presence of functional Kv channels formed by the Kv2.1 α-subunit, a member of another Kv channel family, cannot be ruled out (6, 10, 20, 29, 60). Indeed, between two and four small-amplitude single-channel Kv currents were found in rat (60), human (32), and mouse (16) PASMCs. Although heterogeneity in distribution of different Kv currents was also observed in different PASMCs (4, 43), the presence of multiple voltage-dependent Kv components in a single cell was not obvious because of the homogeneity of the activation and inactivation dependencies (Figs. 5 and 6). It is possible that Kv channel heterogeneity observed at the single-channel level may contribute to some variability in the voltage-dependent parameters for IKv between cells.
It is worth mentioning that the 2- to 3-fold difference in single-channel conductance for the delayed currents encoded by Kv1.x and Kv2.x α-subunits is unlikely to explain the at least 10-fold increase in the whole cell IKv density between −40 and 0 mV in PASMCs compared with MASMCs. Similarly, the differences in the IKv activation and inactivation parameters alone between the two cell types could only predict a 2.6- to 3-fold increase in the steady-state open probability of Kv channels available in the same voltage range (calculated with the mean values for IKv activation and inactivation parameters listed in Table 1 and the theoretical model described in Ref. 26). It is therefore likely that the number of functional Kv channels expressed per unit of the cell membrane area should also be greater in PASMCs than in MASMCs.
In conclusion, our findings demonstrate the presence of a distinctive layer of mitochondria at the cell periphery in PASMCs, forming a restricted diffusion cytosolic space between the mitochondria and plasmalemmal membrane and thus providing a structural basis for a close functional interaction between the Kv channels and mitochondria as supported by the electrophysiological evidence. The smaller density of Kv channels in MASMCs alongside the greater diffusion space between the mitochondria and the cell membrane may provide one plausible explanation for the differences in cellular responses during hypoxia.
This work was supported by the British Heart Foundation (Grants PG03/059 and PG04/069 supporting K. H. Yuill and PG/08/062/25382 awarded to D. V. Gordienko) and the Wellcome Trust (075112 awarded to D. V. Gordienko).
Present address of A. L. Firth: Div. of Pulmonary and Critical Care Medicine, Dept. of Medicine, University of California, San Diego, La Jolla, CA 92093.
↵* A. L. Firth, D. V. Gordienko, and K. H. Yuill contributed equally to this work.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2009 the American Physiological Society