Protection against cardiac injury by small Ca²⁺-sensitive K⁺ channels identified in guinea pig cardiac inner mitochondrial membrane

Abstract

We tested if small conductance, Ca²⁺-sensitive K⁺ channels (SK_Ca) precondition hearts against ischemia reperfusion (IR) injury by improving mitochondrial (m) bioenergetics, if O₂–derived free radicals are required to initiate protection via SK_Ca channels, and, importantly, if SK_Ca channels are present in cardiac cell inner mitochondrial membrane (IMM). NADH and FAD, superoxide (O₂^•−), and m[Ca²⁺] were measured in guinea pig isolated hearts by fluorescence spectrophotometry. SK_Ca and IK_Ca channel opener DCEBIO (DCEB) was given for 10 min ending 20 min before IR. Either TBAP, a dismutator of O₂^•−, NS8593, an antagonist of SK_Ca isoforms, or other K_Ca and K_ATP channel antagonists, was given before DCEB and before ischemia. DCEB treatment resulted in a 2-fold increase in LV pressure on reperfusion and a 2.5 fold decrease in infarct size vs. non-treated hearts associated with reduced O₂^•− and m[Ca²⁺], and more normalized NADH and FAD during IR. Only NS8593 and TBAP antagonized protection by DCEB. Localization of SK_Ca channels to mitochondria and IMM was evidenced by a) identification of purified mSK_Ca protein by Western blotting, immuno-histochemical staining, confocal microscopy, and immuno-gold electron microscopy, b) 2-D gel electrophoresis and mass spectroscopy of IMM protein, c) [Ca²⁺]–dependence of mSK_Ca channels in planar lipid bilayers, and d) matrix K⁺ influx induced by DCEB and blocked by SK_Ca antagonist UCL1684. This study shows that 1) SK_Ca channels are located and functional in IMM, 2) mSK_Ca channel opening by DCEB leads to protection that is O₂^•− dependent, and 3) protection by DCEB is evident beginning during ischemia.

1. Introduction

Depressed mitochondrial (m) bioenergetics, excess reactive oxygen species (ROS) generation, and mCa²⁺ loading are major factors underlying ischemia and reperfusion (IR) injury [1]. Prophylactic measures targeted in part to mitochondria that reduce cardiac IR injury [2, 3] include ischemic preconditioning (IPC, i.e., brief pulses of ischemia and reperfusion before longer ischemia) and pharmacologic pre-conditioning (PPC), i.e., cardiac protection elicited some time after the drug is washed out. PPC is theoretically a better approach because it does not require the heart to first undergo brief ischemia. We reported previously that activation of a large (big) conductance Ca²⁺ –sensitive K⁺ channel (mBK_Ca), which may be located in the cardiac myocyte inner mitochondrial membrane (IMM), can induce PPC [4]. The BK_Ca channel has not been found in the cardiac myocyte plasma membrane, but we have shown that a BK_Ca channel opener, NS1619, has biphasic effects on mitochondrial respiration, membrane potential (ΔΨ_m), and superoxide radical (O₂^•−) production in isolated mitochondria [5, 6]. This suggested that opening of other mitochondrial K⁺ channels could also elicit PPC.

There are other K_Ca channels of intermediate or small conductances identified in non-cardiac cells [7–10] that are membrane bound, calmodulin (CaM) –dependent and gated by Ca²⁺ and other factors. These channels have smaller unit conductances of 3–30 (small, SK_Ca) and 20–90 (intermediate IK_Ca) pS [11]. The opening of SK_Ca channels is initiated by Ca²⁺ binding to calmodulin at the C terminus of the channel [12, 13]. Of the known isoforms of SK_Ca channels that have been identified in endothelial cells, one is K_Ca2.3, which was found to exert a potent, tonic hyperpolarization that reduced vascular smooth muscle tone [14]. Moreover, there is evidence for the K_Ca2.2 isoform in rat and human hearts using Western blot analysis and reverse transcription -polymerase chain reaction [15]. Clones of the channel from atria and ventricles showed much greater expression in atria compared to ventricles, and electrophysiological recordings exhibited much greater atrial than ventricular sensitivity to AP repolarization by apamin, a selective SK_Ca antagonist [15, 16].

We postulated that activation of SK_Ca channels induces a preconditioning effect similar to that elicited by a BK_Ca (K_Ca1.1, maxi-K) opener, and that this effect is mediated via channels located in the IMM, i.e., they promote K⁺ entry into the mitochondrial matrix. We tested if the K_Ca3.1 (IK_Ca1) [7, 9, 17] and K_Ca2.2 and K_Ca2.3 (SK_Ca) [18–21] opener DCEBIO (DCEB), given transiently before ischemia, elicits PPC in a manner similar to that of the mBK_Ca channel opener NS1619 [4]. We specifically examined the role of DCEB in attenuating the deleterious effects of IR injury on mitochondrial bioenergetics by near continuous measurement of m[Ca²⁺], NADH and FAD, and O₂^•− in isolated perfused hearts. We infused NS8593 to antagonize SK_Ca channel opening [22, 23], and several other K⁺ channel blockers to rule out effects of DCEB on other putative mK⁺ channels, i.e., IK_Ca (K_Ca3.1) BK_Ca, and K_ATP channels. Because protective effects of putative K_ATP [24] and BK_Ca [4] channel openers can be abolished by ROS scavengers, we similarly bracketed DCEB with a matrix targeted dismutator of O₂^•− to assess the role of SK_Ca channel opening on O₂^•− production, presumably by mitochondrial respiratory complexes. We used several approaches to furnish solid evidence for the presence and functionality of SK_Ca channel proteins in the IMM of guinea pig isolated cardiac mitochondria, and in isolated IMM.

2. Materials and methods

2.1. Isolated heart model

The investigation conformed to the Guide for the Care and Use of Laboratory Animals (NIH Publication 85-23, revised 1996). Guinea pig hearts were isolated and prepared as described in detail [4, 25–31] with care to minimize IPC. These were pre-oxygenation, maintained respiration after anesthesia with ketamine (50 mg/kg), and immediate aortic perfusion with cold perfusate. Hearts were instrumented with a saline filled balloon and transducer to measure left ventricular pressure (LVP) and an aortic flow probe to measure coronary flow (CF). Heart rate and rhythm were measured via atrial and ventricular electrodes. Hearts were perfused at constant pressure with modified Krebs-Ringer’s solution at 37°C. Heart rate (HR) and rhythm, myocardial function (isovolumetric LVP), coronary flow and venous pO₂ were measured continuously. %O₂ extraction, myocardial O₂ consumption (MVO₂) and cardiac efficiency (HR•LVP/MVO₂) were calculated. At 120 min reperfusion, hearts not isolated for mitochondria were stained with 2,3,5-triphenyltetrazolium chloride (TTC) and infarct size was determined as a percentage of ventricular heart weight [4, 26, 30].

2.2. Cardiac fluorescence measurements

Either m[Ca²⁺], NADH and FAD, or ROS (principally O₂^•−) was measured near continuously or intermittently in the heart using one of four excitation (λ_ex) and emission (λ_em) fluorescence spectra described below. NADH and FAD were measured in the same heart; m[Ca²⁺] and ROS were measured in different subsets of hearts. A trifurcated fiberoptic probe (3.8 mm² per bundle) was placed against the LV to excite and to record light signals at specific λ’s using spectrophotofluorometers (SLM Amico-Bowman and Photon Technology International). The incident polychromic light was filtered at 350 or 490 nm and recorded at 390/450 or 540 nm, respectively, to measure NADH [25, 28, 30, 32, 33] and FAD [30, 32] tissue autofluorescence. Alternatively, hearts assigned to measure Ca²⁺, were loaded with 6 µM indo 1 AM for 30 min followed by washout of residual dye for 20 min. Ca²⁺ transients were recorded at the same wavelengths as for NADH. Then hearts were perfused with MnCl₂ to quench cytosolic Ca²⁺ to reveal non-cytosolic [Ca²⁺], mostly [mCa²⁺] [25, 29, 33]. In other hearts, as reported earlier [4, 26, 27, 31, 33, 34], dihydroethidium (10 µM, DHE), which is used to measure intracellular superoxide (O₂^•−) level, was loaded for 30 min and washed out of residual dye for 20 min. The LV wall was excited with light λ_ex 540 nm; λ_em 590 nm) to measure a fluorescence signal that is primarily a marker of the free radical O₂^•−.[31, 35] DHE enters cells and is oxidized by O₂^•− where it is converted to the labile cation, 2-hydroxyethidium (2-HE⁺), which causes a red-shift in the EM light spectrum [36, 37].

Myocardial fluorescence intensity was recorded in arbitrary fluorescence units (afu) during 35 discrete sampling periods throughout each experiment at a sampling rate of 100 points/s (100 Hz, pulse width 1 µs) during a 12 s triggered period for O₂^•− and for a 2.5 s triggered period for NADH and FAD, and m[Ca²⁺]. For each fluorescence study, no drug alone had any effect on background autofluorescence. Signals were digitized and recorded at 200 Hz (Power LAB/16sp, Chart and Scope version 3.6.3. AD Instruments) on G5 Macintosh computers for later analysis using specifically designed programs with MATLAB (MathWorks) and Microsoft Excel software. All variables were averaged over the 2.5 or 12 s sampling period.

2.3. Protocol

Hearts were infused with 3 µM DCEBIO (DCEB) for 10 min ending 20 min before the onset of 30 min global ischemia. DCEB is derived from the benzimidazolone class of compounds, which are known to stimulate chloride secretion in epithelial cells [7, 8, 38]. DCEB non-selectively opens K_Ca2.2 and 2.3 channels [7, 18–21]. In most hearts DCEB was bracketed either with 40 µM PAX (paxilline), a blocker of BK_Ca channels [39], 20 µM TBAP, a chemical dismutator of O₂^•− that can enter the matrix, 200 µM GLIB (glibenclamide), a K_ATP channel blocker, or 100 nM TRAM (TRAM-34), an established blocker of IK_Ca conductance channels [9]. TRAM was selected because DCEB also opens IK_Ca channels [7, 9, 17, 21]. PAX, TBAP, GLIB, or TRAM was given 5 min before, during DCEB perfusion, and for 5 min after stopping DCEB. In a separate study DCEB was bracketed with 10 µM NS8593, a specific antagonist of SK_Ca channels [22, 23] to compare with a no drug IR control. Drug exposure was discontinued 15 min before the onset of global ischemia. lasted 120 min. NS8593 caused a transient fall in systolic (and developed) LVP and an increase in coronary flow. Additional studies (not displayed) showed that each of these drugs, except for NS8593, given alone (without DCEB) for 20 min before ischemia elicited no appreciable effects and had no different effect on IR injury than the drug-free controls.

2.4. Statistical Analyses

A total of 155 isolated heart experiments were divided into 7 groups, a drug-free control, and DCEB alone or plus NS8593, PAX, TBAP, GLIB or TRAM. Functional data were recorded from 12–15 hearts per group. Infarct size was measured in a blinded manner in 8 hearts per group. NADH and FAD were measured in approximately 6–8 hearts per group, O₂^•− in 5–7 hearts per group, and m[Ca²⁺] in 6–8 hearts per group. Because functional studies showed trends that PAX, GLIB, or TRAM did not block protective effects of DCEB, only four groups were compared in NADH and FAD experiments and three groups were compared in O₂^•− and m[Ca²⁺] experiments. All data were expressed as means ± standard error of means. Appropriate comparisons were made among groups that differed by a variable at a given condition or time, and within a group over time compared to the initial control data. Statistical differences were measured across groups at specific time points (20, 50, 85, 145, and 200 min). Differences among variables were determined by two-way multiple ANOVA for repeated measures (Statview® and CLR anova® software programs for Macintosh®); if F tests were significant, appropriate post-hoc tests (e.g., Student-Newman-Keul’s, SNK) were used to compare means. The incidence of ventricular fibrillation (VF) vs. sinus rhythm per group, and the number of VF’s per heart per group, were determined by Fisher’s Exact Test. In mitochondria K⁺ flux experiments drug treatments were compared to control using the same statistical tests. Mean values were considered significant at P values (two-tailed) <0.05.

2.5. Isolation of cardiac mitochondria and inner mitochondrial membranes (IMMs)

Mitochondria were freshly isolated from 25 guinea pig hearts by differential centrifugation as described previously [5, 6, 34, 40, 41]. To test mitochondrial viability and function in each preparation, the respiratory control index (RCI, state 3/state 4) was determined under both pyruvate (P, 10 mM), and succinate (S, 10 mM) + rotenone (R, 4 µM) conditions. State 3 respiration was determined after adding 250 µM ADP. Intact mitochondrial preparations were discarded if the RCI was less than 3 with succinate + R or less than 9 with pyruvate.

To isolate fraction-enriched IMMs, isolated mitochondria were shocked osmotically by incubating in 10 mM phosphate buffer saline (PBS) (pH 7.4) for 20 min, and then in 20% sucrose for another 15 min. The IMMs were sonicated for 30 s, 3 times, and then centrifuged at 8,000 g for 10 min. The supernatant containing sub-mitochondrial particles was fractionated using a continuous sucrose gradient (30% to 60%) and centrifuged at 80,000 g overnight in a SW28 rotor. The IMMs (enriched in the heavy fractions) were suspended with the isolation medium without EGTA and centrifuged at 184,000 g for 30 min. The final pellet enriched IMMs were suspended in isolation medium without EGTA and BSA and stored at −80°C in small aliquots until use.

2.6. Enhancement of calmodulin-binding proteins from IMM

Calmodulin binds to SK_Ca channels so the calmodulin binding proteins obtained from the IMMs were concentrated to enhance the sensitivity of detection of mSK_Ca channels by Western blotting and by mass spectrometry. For calmodulin column chromatography (calmodulin-sepharose beads) the IMMs (5 mg protein) were solubilized for 2 h at 4°C in washing buffer, 200 mM KCl, 1 mM MgCl₂, 200 µM CaCl₂, 20 mM HEPES, pH 7.4, and 0.5% CHAPS with protease inhibitors. After centrifugation at 50,000 g for 30 min, the supernatant was applied to a calmodulin-sepharose column (10 by 1.5 cm) pre-equilibrated with the solubilization buffer containing 0.1% CHAPS. The column was washed rapidly with 500 mL of washing buffer as above. The proteins were eluted from the column by 2 mM EGTA in the elution buffer (200 mM KCl, 20 mM HEPES, pH 7.4, 0.1% (w/v) CHAPS) after washing. The fractions collected were concentrated and the proteins were separated by 2-D gel electrophoresis as follows.

2.7. Purification of SK_Ca channel proteins from IMM by isoelectric focusing

After isolating the IMM protein fraction (2.5, 2.6) the first dimension of isoelectric focusing (IEF) during 2-D gel electrophoresis was done in native gel buffer on an Immobilon Drystrips (Amersham) with pH 4~7 gradient. The antibody was targeted to K_Ca2.3 (aka, hSK3, KCN3, Osenses Pty, Ltd.). The second dimension was done in a 10% Criterion® tris-SDS gel (Bio-Rad). Two identical gels were run at the same time, with one used for transfer to nitrocellulose membrane for Western blot analysis, and the other for silver staining for visualization.

2.8. IMM protein identification using electrospray LC/MS

IMM proteins (from 2.5, 2.6) were digested with trypsin and subjected to pH focusing into 10 fractions over pH 3–10 and each fraction was directly analyzed using a NP LC/ESI mass spectrometer (Finnigan™ LTQ™ Ion Trap MS, Thermo Electron Corporation) to generate specific mass spectra typical for a given protein. The instrument utilizes stepped normalized collision energy (SNCE) to improve fragmentation efficiency over a wide mass range. This increases the capacity of a linear trap and the accuracy and sensitivity of peptide detection in the fmol range. A mass database (NCBI Entrez Pubmed protein) was searched for matching proteins and consequently the SK_Ca channel protein of interest was tentatively identified in IMM.

2.9. Purification of intact mitochondria by Percoll gradient fractionation

To further verify localization of SK_Ca channel protein in an intact mitochondria preparation, the Percoll gradient technique [42, 43], with slight modifications, was used to purify intact mitochondria and immuno-histochemical staining was utilized to identify SK_Ca channel protein. In brief, mitochondria isolated as previously described [5, 6, 34, 40, 41] were layered over 30% Percoll (in buffer A containing 450 mM mannitol, 50 mM HEPES, 2 mM EDTA, pH adjusted to 7.4 followed by addition of 50 mg BSA), and centrifuged at 95,000 g for 30 min. The lower dense band observed at the bottom of the tube, enriched in mitochondria, was collected using a long tip glass pipette. Collected mitochondria (~4 ml) were resuspended in the same buffer used to dilute Percoll, and centrifuged again at 6300 g. The resulting pellet was suspended in the same buffer without BSA (buffer B) and re-centrifuged at 6300 g. The mitochondrial pellet was resuspended in a small volume (~0.3 ml) of buffer B and stored until further use.

2.10. Identification and localization of SK_Ca channel protein in purified mitochondria

Immuno-histochemical staining with an anti-SK_Ca antibody and confocal microscopy were used, in part, to verify that SK_Ca channel protein resides in mitochondria. Briefly, mitochondria, isolated and purified as described above (2.9), were fixed onto poly-lysine coated coverslips. Mitochondrial structures were then fixed using paraformaldehyde and membranes were permeabilized using Triton X-100 and non-specific binding sites blocked by goat serum albumin. Coverslips were then incubated in solution containing anti-K_Ca2.2 (anti-SK2, ETQMENYDKHVTYNAERS, Alomone Labs (1:1000 in 5% milk)) and anti-ANT (adenine nucleotide translocase, Invitrogen) antibodies for 30 min followed by three washes in 0.1 M PBS. Coverslips were then incubated in appropriate secondary antibodies (Alexa Flour 455 and 546 respectively, Invitrogen (1:3000 in 2% milk)) for another 30 min and were then transferred onto microscope slides and visualized using a Leica confocal microscope (TCS SP5). Alternatively, mitochondria were utilized for immuno-gold labeling to localize SK_Ca channel protein in individual mitochondria.

2.11. Localization of SK_Ca channel protein by immuno-gold labeling and electron microscopy

Immuno-electron microscopy (IEM) was used to localize SK_Ca protein in purified cardiac mitochondria similar to the technique used by Douglas et al. [44] to localize BK_Ca channel protein in mitochondria. The final mitochondrial pellet, prepared as described above (2.9), was resuspended in 500 µL isolation buffer before centrifugation at 16,000 g for 20 min. The supernatant was discarded and an EM fixative containing 0.1% glutaraldehyde + 2% paraformaldehyde in 0.1 M NaH₂PO₄ buffer (pH 7.4) was added. After 1h fixation at room temperature the pellet was gently detached from the tube with a 25G needle and processed following the protocols of Berryman and Rodewald [45] and Burleigh et al. [45]. Pellets were washed 3 × 20 min in 0.1M NaH₂PO₄ buffer containing 3.5% sucrose and 0.5 mM CaCl₂, then rinsed in 0.1 M glycine in NaH₂PO₄ buffer for 1 h on ice before returning to NaH₂PO₄ buffer. The pellets were cut into 1 mm cubes and then washed 4 × 15 min in 0.1M tris maleate buffer + 3.5% sucrose, pH 6.5, at 4°C followed by post fixation in 2% Uanyl acetate (w/v) in tris buffer, pH 6, for 2 h at 4°C; specimens were then given a final rinse 2 × 5 min in Tris maleate buffer, pH 6.5. The specimens were then processed by the progressive lowering-of-temperature method into Lowicryl K4M resin and the resin was polymerized by UV irradiation. Ultrathin sections (70 nm) were cut onto Formvar/carbon coated grids. Immuno-labeling was performed by floating grids on droplets of 0.1 M NaH₂PO₄ buffer containing 5% BSA (PB-BSA), then incubating with rabbit polyclonal anti-K_Ca2.2 (anti-SK2, Alomone Labs) diluted 1:50 for 90 min, or with the positive control mitochondrial marker, cytochrome c oxidase (anti-COX1: Complex IV, subunit 1) mouse monoclonal antibody diluted 1:500. Non-immune rabbit polyclonal serum was used as the negative control. This step was followed by 3 × 5 min washes in PB-BSA. The sections were then incubated with goat anti-rabbit IgG, or goat anti-mouse IgG, conjugated to 10 nM colloidal gold [46] for 90 min at room temp, rinsed in distilled water, and then stained with 2% aqueous uranyl acetate. Sections were examined in a JEOL JEM2100 TEM at 80 kV.

2.12. Purification/identification of SK_Ca channel protein by isoelectric focusing (IEF) and Western blotting

Total mitochondrial protein, once isolated and purified as above (2.9), was partitioned by IEF and the resulting fractions analyzed for mSK_Ca protein. Mitochondria (1 mg) were suspended in 3 mL electrophoresis buffer (0.1% w/v CHAPS, 0.1% w/v dodecyl maltoside, 5% (v/v) glycerol, 10 mg dithiothreitol) and IEF was performed using the Micro-Rotofor system (BioRad, CA) for 4 h at 400 mA constant current. The fractions thus obtained were collected and analyzed for SK_Ca protein by Western blot using the anti-K_Ca2.2 (anti-SK2) antibody. Briefly, equal volumes of the 10 fractions obtained by IEF were suspended in Laemmli buffer and resolved using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) [47], as originally described by Laemmli [48], and transferred onto poly vinilidine difluoride membranes using Transblot System (Bio-Rad) in 50 mM tricine and 7.5 mM imidazole transfer buffer. Membranes were blocked with 10% non fat dry milk in tris buffered saline-TBS_t (25 mM Tris-HCl at pH 7.5, 50 mM NaCl and 0.1% Tween 20) by incubating for 1 h followed by incubation in the anti-K_Ca2.2 antibody (anti-SK2) solution overnight at 4°C. After three washes with TBS_t the membrane was incubated with an appropriate secondary antibody conjugated to horseradish peroxidase for 3 h. After five washes with TBS_t the membrane was incubated in enhanced chemiluminescence detection solution (ECL-Plus, GE-Amersham) and exposed to X-ray film for autoradiography. The protein fraction containing the largest amount of SK_Ca was used for single channel recordings.

2.13. Enriching and incorporating mSK_Ca channel protein into lipid bilayers

Channel activity of the purified and enriched mSK_Ca protein was monitored by incorporating it into a planar lipid bilayer, as previously described [49]. Briefly, phospholipids were prepared by mixing phosphatidyl-ethanolamine, phosphatidyl-serine, phosphatidyl-choline, and cardiolipin (Avanti Polar Lipids) in a ratio of 5:4:1:0.3 (v/v). The phospholipids were dried under N₂ and re-suspended in n-decane to a final concentration of 25 mg/mL. The cis/trans chambers contained symmetrical solutions of 10 mM HEPES, 200 mM KCl and 100 µM CaCl₂ at pH 7.4. The cis chamber was held at virtual ground and the trans chamber was held at the command voltages. SK_Ca protein was added into the cis chamber. The effect of the SK_Ca blocker apamin, 100 nM, on channel activity was tested by adding it to the cis chamber in the presence of 100 µM CaCl₂. To test for Ca²⁺ dependence of the SK_Ca channel, [Ca²⁺] was serially increased (1, 50 and 100 µM) in the cis chamber. Currents were sampled at 5 kHz and low pass filtered at 1 kHz using a voltage clamp amplifier (Axopatch 200B, Molecular Devices) connected to a digitizer (DigiData 1440, Molecular Devices), and recorded in 1 min segments. The pClamp software (version 10, Molecular Devices) was used for data acquisition and analysis. Additional analyses were conducted using Origin 7.0 (OriginLab).

2.15. Matrix K⁺ measured in isolated mitochondria

Cardiac isolated mitochondria (0.5 mg protein/mL) were suspended in respiration buffer containing 130 mM KCl, 5 mM K₂HPO₄, 20 mM MOPS, 2.5 mM EGTA, 1 µM Na₄P₂O₇, 0.1% BSA, pH 7.15 adjusted with KOH. Buffer [Ca²⁺] was less than 100 nM as assessed by the fluorescence dye indo 1. Matrix K⁺ was monitored during state 4 respiration (200 µM ATP) with substrate Na-pyruvate (10 mM) in a cuvette-based spectrophotometer (QM-8, Photon Technology International, PTI) with light (λ_ex 340 and 380 nm; λ_ex 500 nm) in the presence of the fluorescence dye PBFI (1 µM per mg/mL protein, Invitrogen) [50]. PBFI, in the acetylated methyl-ester (AM) form, was added to the mitochondrial preparation and incubated at 25°C for 20 min. After entering the matrix PBFI is retained in the matrix after it is cleaved from the methyl-ester. During the last pellet wash the extra-matrix residual dye was washed out. Most experiments were conducted in the presence of 500 µM quinine to block the mitochondrial K⁺/H⁺ exchanger (mKHE) and extrusion of the K⁺ [50]. In some experiments 0.25 nM valinomycin, a K⁺ ionophore, was given to verify an increase in matrix K⁺ influx, and to be used as a reference for the change of K⁺ influx by DECB ± its antagonist UCL1684.

3. Results

3.1. DCEB protects isolated heart against IR injury

Spontaneous heart rate averaged 242±13 beats/min before ischemia for all groups; this was not statistically different at 120 min reperfusion for all groups (data not displayed). If ventricular fibrillation (VF) occurred, it was only once within the first 5 min of reperfusion in any heart; all were converted to sinus rhythm with intracoronary lidocaine. After 5 min reperfusion all hearts remained in sinus rhythm, some with occasional pre-ventricular excitations. In data not displayed the incidence of VF on reperfusion was CONTROL 100%, DCEB+TBAP 100%, DCEB 76%, DCEB+TRAM 72%, DCEB+PAX 77%, DCEB+GLIB 77% (all nonsignificant vs. control group).

Figs. 1–4 show the marked degree of dysfunction or damage in the untreated control group during and after global ischemia and the beneficial effects of PPC elicited by DCEB treatment before ischemia. Developed LVP (Fig. 1A) and coronary flow (Fig. 1B) were reduced in each group after ischemia compared to before ischemia, but these changes were much larger in the CONTROL and DCEB+TBAP groups than in the other groups. Similarly, cardiac efficiency (Fig. 2A) was lower and infarct size (Fig. 2B) was largest in the CONTROL and DCEB+TBAP groups than in all other groups. The drug treatments before ischemia had no effects by themselves on any of the functional variables. These figures indicate that these variables were markedly improved on reperfusion after treatment with DCEB and that these improvements were reversed by TBAP, but not by PAX, TRAM, or GLIB.

Fig. 1.

Improved (A) developed (systolic-diastolic) LV pressure and (B) coronary flow after preconditioning with 3 µM DCEB. Note that TBAP (synthetic superoxide dismutase mimetic) reversed the protective effects of DCEB whereas antagonists of big (PAX, paxilline) and intermediate (TRAM) conductance K_Ca channels did not.

Fig. 4.

Reduced (A) O₂^•− (DHE fluorescence) and (B) mitochondrial [Ca²⁺] (indo 1 fluorescence) after preconditioning with DCEB. Note the increases in these signals during ischemia and the slow decline during reperfusion. DCEB attenuated the increase in these signals during ischemia and reperfusion and this was reversed by TBAP.

Fig. 2.

A: Improved cardiac efficiency (developed LV pressure (mmHg)•heart rate (beats/min))/MVO₂ (µL O₂•g⁻¹•min⁻¹) after preconditioning with DCEB. Note that TBAP reversed the protective effects of DCEB whereas antagonists of big (PAX) and intermediate (TRAM) conductance K_Ca channels did not. B: Marked decrease in infarct size after preconditioning with DCEB. Note that TBAP reversed the anti-infarction effect of DCEB whereas antagonists of big (PAX) and intermediate (TRAM) conductance K_Ca channels and K_ATP channels (glibenclamide, GLIB) did not.

There was no detectable change in NADH and FAD autofluorescence in any group by drugs given and terminated before ischemia (Fig 3A,B). NADH (Fig. 3A) remained higher at the end of ischemia and fell less during reperfusion after treatment with DCEB; this was reversed by TBAP but not by TRAM. FAD remained lower at the end of ischemia and rose less during reperfusion after treatment with DCEB (Fig 3B); this was reversed by TBAP, but not by TRAM. In other experiments there were no detectable change in NADH or FAD on reperfusion after DCEB+PAX or +GLIB treatment vs. DCEB alone.

Fig. 3.

Improved redox state (A: NADH and B: FAD autofluorescence) after preconditioning with DCEB. Note the inverse changes in NADH and FAD during ischemia and reperfusion and the more normalized responses in the DCEB group. TBAP reversed the protective effects DCEB whereas paxilline (PAX), an antagonist of big conductance BK_Ca channels did not.

DHE fluorescence (O₂^•− formation) (Fig. 4A) and indo 1 fluorescence (m[Ca²⁺]) (Fig. 4B) rose markedly in each group during the course of ischemia. TBAP caused a small, but insignificant, decrease in DHE fluorescence before ischemia. TBAP reversed the effect of DCEB to reduce O₂^•− and m[Ca²⁺] on reperfusion. Other experiments (not shown) did not demonstrate detectable changes in ROS formation or m[Ca²⁺] on reperfusion after DCEB+PAX, +GLIB or +TRAM treatments vs. DCEB alone.

In companion experiments (Fig. 5A–D) the protective effects of DCEB were abolished or antagonized by the SK_Ca channel antagonist NS8593, thus demonstrating that DCEB protected via activation of SK_Ca channels. DCEB–induced maintenance of developed LVP was completely blocked, while the maintenance of coronary flow and the reduction of diastolic LVP and FAD oxidation by DCEB were all markedly reversed by NS8593. NS8593 alone significantly depressed developed LVP when given before ischemia and tended (non significantly) to slightly increase coronary flow, possibly indirectly due to reduced ventricular compression; thus the small increase in flow (Fig. 5B) noted in the presence of DCEB is due to NS8593 rather than DCEB. Generally, cardiac depression before ischemia is cardioprotective, but giving NS8593 with DCEB before ischemia, resulted in a worsening of contractile function on reperfusion.

Fig. 5.

Improved (A) developed LV pressure and coronary flow (B), and decreased diastolic LV pressure (C) and FAD oxidation (D), after preconditioning with DCEB. Note that 10 µM NS8593 (a specific SK_Ca antagonist) abrogated these protective effects of DCEB.

These studies demonstrated that DCEB had protective effects against cardiac IR injury mediated by the SK_Ca channel, and that cardiac mitochondria appeared to be involved in mediating this protection. Studies were then undertaken to isolate and identify the target of DCEB, the SK_Ca protein, in cardiac isolated mitochondria and in IMM, and to verify the functionality of the protein in an artificial lipid bilayer.

3.2. Isoelectric focusing and peptide sequences identify SK_Ca in isolated IMM

IMM protein, enhanced for calmodulin-binding residues, was separated by 2-D electrophoresis after silver staining. Three peptide spots of approximately 70 kD at pH 5.2–5.5 were detected as SK_Ca using the anti K_Ca2.3 (anti-hSK3) (Fig. 6, panels a–c). Complementing this finding, a K_Ca2.3 protein was identified by ESI-mass spectrometry from five matching peptides with an amino acid coverage of 10.73% (Table). There was no evidence for peptides matching Na⁺/K⁺ ATPase or Ca²⁺ ATPase suggesting the absence of sarcolemmal and t-tubular membranes in the mitochondrial fraction. The mass spectrum of one of these peptide sequences is shown (Fig. 7). These results demonstrated that SK_Ca channels were present in the IMM.

Fig. 6.

Identity of small-conductance K_Ca channels in IMM from guinea pig heart. Top panel: Silver staining of calmodulin affinity column-purified protein fractions after 2-D gel fractionation. The square indicates the area of interest, which was magnified and is shown in the middle panel. The arrows indicate position of K_Ca2.3 proteins. Bottom panel: Western blot with an antibody targeting SK_Ca (anti-hSK3) channel detection at 3 spots at 70 kDa (arrow) between pH 5.2 and 5.5. Negative control was done by pre-incubating K_Ca2.3 antibodies with blocking peptide (not shown).

Table.

Protein coverage matched to an SK_Ca subunit 6 isoform by NP LC/SI mass spectrometry.

Reference: gi|21361129|ref|NP_002240small conductance calcium-activated potassium protein 6. Database: C:\Xcalibur\database\human_ref.fasta Number of Amino Acids: 736 Average MW: 82026.4 pI:10.07. (matched peptide sequences). Aka: SKCa3, KCNN3
Note the matching of peptide sequences covering subunit 6 and calmodulin binding domain.
MDTSGHFHDS	GVGDLDEDPK	CPCPSSGDEQ	QQQQQQQQQQ	QPPPPAPPAA
PQQPLGPSLQ	PQPPQLQQQQ	QQQQQQQQQQ	QQQQQPPHPL	SQLAQLQSQP
VHPGLLHSSP	TAFRAPPSSN	STAILHPSSR	QGSQLNLNDH	LLGHSPSSTA
TSGPGGGSRH	RQASPLVHRR	DSNPFTEIAM	SSCKYSGGVM	KPLSRLSASR
RNLIEAETEG	QPLQLFSPSN	PPEIVISSRE	DNHAHQTLLH	HPNATHNHQH
AGTTASSTTF	PKANKRKNQN	IGYKLGHRRA	LFEKRKRLSD	YALIFGMFGI
VVMVIETELS	WGLYSKDSMF	SLALKCLISL	STIILLGLII	AYHTREVQLF	code:
VIDNGADDWR	IAMTYERILY	ISLEMLVCAI	HPIPGEYKFF	WTARLAFSYT	BOLD = peptides identified
PSRAEADVDI	ILSIPMFLRL	YLIARVMLLH	SKLFTDASSR	SIGALNKINF	blue = transmembrane region 1
NTRFVMKTLM	TICPGTVLLV	FSISLWIIAA	WTVRVCERYH	DQQDVTSNFL	green = selectivity pore
GAMWLISITF	LSIGYGDMVP	HTYCGKGVCL	LTGIMGAGCT	ALVVAVVARK	grey = transmembrane region 6
LELTKAEKHV	HNFMMDTQLT	KRIKNAAANV	LRETWLIYKH	TKLLKKIDHA	yellow = calmodulin binding domain
KVRKHQRKFL	QAIHQLRSVK	MEQRKLSDQA	NTLVDLSKMQ	NVMYDLITEL	(CaMD)
NDRSEDLEKQ	IGSLESKLEH	LTASFNSLPL	LIADTLRQQQ	QQLLSAIIEA
RGVSVAVGTT	HTPISDSPIG	VSSTSFPTPY	TSSSSC
Protein Coverage:			MH+	% Mass	AA		% AA
RALFEKR			920.09	1.12	279 – 285		0.95
GVCLLTGIMGAGCTALVVAVVARKLELTK			2888.59	3.52	527 – 555		3.94
HVHNFMMDTQLTKR			1759.05	2.14	559 – 572		1.90
IKNAAANVLRETWLIYK			2004.36	2.44	573 – 589		2.31
FLQAIHQLRSVK			1440.72	1.76	609 – 620		1.63
		Totals:	8936.73	10.89	79		10.73

Letters in bold represent those peptides identified based on their MS/MS profiles. The list below the sequence represents all peptides identified by mass spectrometry. The gray-filled sections of the bar above the sequence represent the position of the identified peptides in the sequence. Also shown is the molecular weight and pI of the protein. There are 4 peptides covering 10.89% of the total amino acids of this protein segment. Note that the mass spectral analysis showed a 70% amino acid coverage in the calmodulin binding domain (CaMBD), a peptide sequence just before (N terminus) the a pore forming subunit 1, and the entire sequence of subunit 6.

Fig. 7.

Identification of one peptide, FLQAIHQLRSVK (in CaMBD), from data obtained using nano-LC/MS. The b-ions and y-ions are fragment masses of the above peptide upon its collision fragmentation. Peptides were identified by searching the rodent subset of Uniprot databases. This protein was identified based on the 5 matching peptide sequences shown in the Table.

3.3. Western blots of serially purified mitochondria demonstrate SK_Ca channel protein

Mitochondria exhibited increasing band densities for both SK_Ca and ANT protein (Fig. 8) when enriched by Percoll gradient serial purification. This furnished compelling Western blot evidence that SK_Ca channel protein increases in abundance with ANT, which is present only in the IMM.

Fig. 8.

Western blots of serially purified mitochondria showed increasing amounts of SK_Ca protein. Equal amounts of protein were loaded in the gel. Total homogenate (lane 1, TH) showed least band intensity, followed by mitochondria isolated by differential centrifugation (lane 2 RC); mitochondrial purified further by Percoll gradient purification (lane 3, PP) had the highest band intensity. Protein bands of SK_Ca are approximately 68 KDa. Purity of mitochondria was followed by assaying the increased amount of ANT, along with SK_Ca, protein in their respective purification fractions.

3.4. Immunochemistry and confocal microscopy identify SK_Ca channel protein in mitochondria

Confocal microscopy was used to localize SK_Ca protein to intact mitochondria. Cardiac mitochondria were visualized as stained by an antibody against ANT (green), and SK_Ca channel protein was visualized using the anti-K_Ca2.2 (anti-SK2) antibody (red) (Fig. 9). Overlay of the two images (yellow) shows co-localization of SK_Ca and ANT proteins in cardiac mitochondria. Since ANT localizes only to the IMM, this suggested that SK_Ca channel protein also localizes to the IMM.

Fig. 9.

SK_Ca protein identified in isolated mitochondria and visualized by confocal microscopy. Overlay of the two images (anti-ANT, green and anti-SK_Ca, red) demonstrates co-localization (yellow) of the SK_Ca protein in mitochondria.

3.5. Immuno-gold labeling and EM show localization of SK_Ca channels in mitochondrial matrix

To further confirm the presence and localization of the SK_Ca channels on the IMM, mitochondria were visualized at high resolution using IEM. A large field EM view shows largely normal appearing cardiac mitochondria with intact outer membranes and cristae (Fig. 10). Enhanced resolution of immuno-gold labeled mitochondria show gold particles attributed to SK_Ca channels (Fig. 11A, B) or cytochrome c oxidase (COX) (Fig. 11C) within the matrix; in detailed examination of electron micrographs approximately 50% contained at least 2 gold particles. Negative controls (Fig. 11D) (non-immune rabbit polyclonal serum showed no gold particles in any field. Figs. 8, 9, and 11 confirm that SK_Ca channels are located in mitochondria and most likely in the IMM.

Fig. 10.

Electron micrograph of isolated mitochondria. Larger field view of untreated mitochondria shows largely intact structural characteristics after isolation from guinea pig hearts.

Fig. 11.

Immuno-electron microscopy of isolated cardiac mitochondria. A, B: SK_Ca protein as visualized in two mitochondria; 50% of all viewed mitochondria exhibited gold labeling. Gold labeling was obtained by immuno-gold secondary antibody conjugated to primary rabbit polyclonal anti-K_Ca2.2 (anti-SK2). C: Positive control was anti-cytochrome c oxidase (COX1) conjugated to goat anti-rabbit or mouse conjugated to colloidal gold; each mitochondrion in a large field view exhibited at least two gold particles. D: Negative control was only secondary polyclonal rabbit antibody conjugated to gold; there was no gold labeling of any mitochondria in any views.

3.6. Mitochondrial SK_Ca protein forms a functional channel

To test if purified mitochondrial SK_Ca protein forms a functional channel, SK_Ca protein, isolated as noted above (2.9), was incorporated into a planar lipid bilayer for electrophysiological measurements. In the lipid bilayer, the SK_Ca protein exhibited robust activity in the presence of 100 µM [Ca²⁺] (Fig. 12A). Two conducting states with chord conductances of 230 and 730 pS were observed when recorded in an ionic condition of equimolar 200 mM KCl. Adding apamin blocked channel activity (Fig. 12B) indicating that the functional channel formed by the mSK_Ca protein was inhibited by this SK_Ca channel blocker. The mSK_Ca channel protein incorporated into the planar lipid bilayer also displayed Ca²⁺–dependent activity (Fig. 13). The mSK_Ca channel exhibited increasing activity as [Ca²⁺] was serially increased from 1 to 100 µM. As shown, channel open probability (P_o) increased from P_o=0.5 at 1 µM [Ca²⁺] to P_o=1.0 at 50 and 100 µM Ca²⁺. A notable observation was also the [Ca²⁺] dependent increase in the number of conducting states. At 1 µM Ca²⁺ the predominant conductance was 180 pS; however, at 50 and 100 µM Ca²⁺ multiple, larger conductances were revealed. Thus, as [Ca²⁺] was increased the mSK_Ca channel exhibited greater conducting states while at lower [Ca²⁺], low conductance states dominated. This observation is further supported by the existence of a smaller conductance channel of 70 pS which was detected, albeit infrequently, at 1 µM Ca²⁺ (Fig. 13, inset).

Fig. 12.

mSK_Ca channel protein activity. Purified mitochondrial SK_Ca protein was incorporated into a planar lipid bilayer and channel activity was recorded at a membrane potential of −10 mV in the presence of 100 µM CaCl₂. Dotted lines denote zero current levels and downward deflections denote channel openings. A: Two primary conductance states with chord conductances of 230 and 720 pS were observed under control conditions. The current recording is also depicted in an expanded time scale. Corresponding all-points amplitude histogram is also shown. B: Channel activity was blocked by adding100 nM apamin.

Fig. 13.

mSK_Ca channel sensitivity to [Ca²⁺]. Channel activity of purified mitochondrial SK_Ca protein, incorporated into the planar lipid bilayer, was recorded at a membrane potential of −10 mV. [Ca²⁺] was incrementally increased; dotted lines denote zero current levels and downward deflections denote channel openings. The corresponding all-points amplitude histogram is also shown. The predominant conductance was 180 pS when channel activity was recorded in 1 µM [Ca²⁺]. However, we have also observed, infrequently, a smaller conducting state of 70 pS at 1 µM [Ca²⁺]. A sample tracing is depicted in the inset in which the calibration for the x- and y-axis is 200 ms and 1 pA, respectively; C and O denote the closed and open states, respectively.

3.7. DCEB–induced increased matrix [K⁺] is blocked by UCL1684

The consequence of opening of SK_Ca channels to changes in mitochondrial matrix [K⁺] was also determined. In isolated mitochondria the SK_Ca channel opener DCEB increased matrix [K⁺] in the presence of quinine to inhibit KHE and thus counter K⁺ extrusion (Fig. 14A, B). The observed influx of K⁺ into the matrix was confirmed by similar K⁺ influx induced by the K⁺ ionophore valinomycin. The effect of DCEB was blocked by UCL1684 (a SK_Ca blocker) but not by iberotoxin (IBX) (a blocker of BK_Ca but not SK_Ca channels) (Fig. 14B). The increase in matrix K⁺ uptake induced by DCEB and blocked by UCL1684 (Fig. 14), and the Ca²⁺ induced increases in K⁺ current and inhibition by apamin in lipid bilayers (Figs. 12, 13) functionally linked DCEB’s cardiac effects to SK_Ca channels presence and activity in cardiac mitochondria.

Fig. 14.

A: Sample time tracing showing effect of 30 µM DCEB in the presence of 500 µM quinine (KHE inhibitor) to increase matrix K⁺ (PBFI fluorescence) in mitochondria isolated from a guinea pig heart. No change was observed in the absence of quinine. The DCEB-induced increase in K⁺ flux was completely blocked by 100 nM UCL1684. Note larger but similar effect of 1 nM valinomycin, a K⁺ ionophore, to DCEB B: Summary effects (n = 10 mitochondrial preparations) of DCEB, expressed as a % of valinomycin effect, on increasing matrix K⁺ in the presence of quinine. This increase in K⁺ was blocked by SK_Ca channel blocker UCL1684 but not by 200 nM iberiotoxin (IBX), a blocker of BK_Ca channels.

4. Discussion

Our results suggest a novel role for the SK_Ca channel in cardiac myocyte preconditioning, likely mediated via altered mitochondrial function due to opening of SK_Ca channels located in myocyte mitochondrial IMM (mSK_Ca). Our comprehensive experimental approach shows that the well-known SK_Ca (and IK_Ca) channel activator DCEB preconditioned hearts and that this was fully reversible by bracketing DCEB with the intra-matrix O₂^•− dismutator TBAP. Cardioprotective effects of DCEB were attributed specifically to activation of SK_Ca channels and not to activation of K_ATP, IK_Ca, or BK_Ca channels because NS8593, but not GLIB, TRAM, or PAX blocked its effects. DCEB increased K⁺ flux in isolated mitochondria and the purified SK_Ca protein formed a functional channel when incorporated into lipid bilayers. Thus mSK_Ca channel opening, similar to that of mK_ATP and mBK_Ca channel opening, appears to induce PPC by an as yet unclear mechanism related to enhanced matrix K⁺ entry. Moreover, mitochondrial-derived O₂^•− is required to initiate PPC by DCEB, because if O₂^•− is rapidly converted to downstream products the protection by DCEB is lost.

Just as we supported our evidence that the SK_Ca channel is specifically involved in PPC of isolated hearts, we sought to support specifically that the mSK_Ca channel was associated with cardioprotection. To provide evidence that DCEB’s has protective effects mediated by mSK_Ca channels, it was necessary to rigorously identify mSK_Ca protein in purified mitochondria, and specifically in the IMM. To do so we utilized Western analysis and immuno-histochemistry, confocal microscopy, electron microscopy, and mass spectrometry of purified mitochondrial proteins derived from IMM. Organelle location was accompanied by channel functionality in isolated mitochondria and in lipid bilayers, thus supporting that this channel may play a role in cardiac protection via a mitochondrial mechanism. Overall our study indicates that the SK_Ca channel, localized in the cardiac cell IMM, mediates the effect of DCEB in preconditioning of the heart via a mitochondrial mechanism related to mK⁺ flux and O₂^•− generation.

The dequalinium analogue, UCL1684, is known to block opening of apamin-sensitive SK_Ca channels in mammalian cell lines [51]. We support involvement of the mitochondrion in DCEB’s PPC effect because TBAP could block the protective functional effects of DCEB as well as block DCEB’s effect to decrease ischemia-induced levels of mCa²⁺ and of O₂^•−, presumably generated by complexes along the electron transport system [52]. Also, DCEB directly induced an increase in matrix K⁺ uptake. Moreover, since DCEB had no significant effects on coronary flow (Fig. 1), automaticity, or contractility, this suggests DCEB had no effect on endothelial/vascular or ventricular cell function. Overall, our results demonstrate the marked cardioprotective effects after preconditioning with DCEB and implicate mSK_Ca channel opening and generation of O₂^•−, or its products, as initiators and inhibitors of mitochondrial as well as cardiac myocyte PPC.

The improvement in cardiac function by DCEB was accompanied by reduced formation of O₂^•−, reduced m[Ca²⁺], and improved redox state (more normal NADH and FAD levels) during both ischemia and reperfusion. These cardioprotective effects of DCEB were blocked only by either TBAP or NS8593. We suggest that initial formation of O₂^•− is essential for the triggering mechanism of PPC by mSK_Ca channel activation. However, a downstream product of O₂^•−, e.g. H₂O₂, might actually mediate the protective effects of DCEB. A similar dependence for O₂^•− has been observed for the mK_ATP channel opener diazoxide [24], the BK_Ca channel opener NS1619 [4], and volatile anesthetics [27, 34]. Drug lipophilicity with mitochondrial membrane penetration may be an important common denominator for the activity of drugs such as diazoxide, a putative mK_ATP channel agonist, and DCEB.

4.1 Distribution and function of Ca²⁺ -sensitive K⁺ channels

The cell membranes of vascular smooth muscle, neural, and secretory cells contain large conductance (200–300 pS). i.e. Big Ca²⁺-sensitive K⁺ (BK_Ca, aka maxi-K_Ca) channels that when opened produce vasodilation, hyperpolarization, and secretion. BK_Ca channel opening is activated by increased [Ca²⁺]_i and by cell membrane depolarization [53]. Activation of BK_Ca over a range of [Ca²⁺] is mediated at several binding sites within the channel [54] so that there is a wide range of [Ca²⁺] responsiveness (K_a 10–1000 µM) [55]. As K⁺ exits the cell with BK_Ca channel opening, this elicits cell membrane repolarization or hyperpolarization, which in turn reduces Ca²⁺ entry by closing voltage-dependent Ca²⁺ channels. Altered redox potential in smooth muscle [56] suggested mitochondrial involvement. Xu et al. [57] first furnished evidence that BK_Ca channels are located in cardiac mitochondria.

The membrane bound, but non-voltage-gated, K_Ca channels, i.e., SK_Ca and IK_Ca [11], are gated by Ca²⁺ and other factors. SK_Ca (a.k.a. K_Ca2.1–2.3, KCNN1–3 -gene symbol) channels have characteristics that largely differ from the BK_Ca channels i.e. small unitary conductance (10–30 pS), voltage independence with activation only by Ca²⁺ at very low K_a (0.3 µM with steep I/V slopes), a sensitivity to apamin, heteromeric assembly of the SK_Ca pore forming subunits with calmodulin (CaM), N rather than C terminal EF hand domain for Ca²⁺ binding, and Ca²⁺ gating near the K⁺ selectivity filter [58]. SK_Ca’s are unique in that calmodulin forms an integral part of the channel, forming its Ca²⁺-sensitive subunits [59]. In the presence of Ca²⁺, two calmodulin binding domains form a dimer, which allows the channel to open [12].

Antibodies against K_Ca2.2 and K_Ca2.3 were both used for immuno-staining and Western blot characterizations in purified mitochondria and in the enriched IMM fraction, respectively. We found that the identified protein was positive to both sets of antibodies. Because the sequence homology of the SK_Ca family of channels (K_Ca2.1, K_Ca2.2, K_Ca2.3) is highly conserved [60], the commercial antibodies we used might not have been selective enough to definitely identify the molecular identity of the specific mSK_Ca isoform. But we did confirm that the purified mSK_Ca protein formed a functional channel by recording channel activity after incorporating the protein into a planar lipid bilayer. The channel was inhibited by apamin, a blocker of plasmalemmal membrane SK_Ca channels. However, the recorded conductance states were higher than those reported for cell membrane SK_Ca channels, which are in the range of 10–14 pS [11, 58]. The underlying cause of this discrepancy is unclear. It is possible that mSK_Ca channels have biophysical attributes that are different from their cell membrane counterparts. In particular, the mSK_Ca channels exhibited multiple conducting states that appear to be Ca²⁺–dependent. As the [Ca²⁺] increased, the channel’s conductance also increased. Thus, at very low [Ca²⁺], lower conductances may be revealed that are closer to those reported for the plasmalemmal SK_Ca channel. In support of this, in some recordings we infrequently observed a conductance of 70 pS at 1 µM [Ca²⁺]. However, the mechanism that underlies this Ca²⁺–dependent gating of the mSK_Ca channel has yet to be delineated. Though it is premature to speculate on the structural homology between the mitochondrial and plasmalemmal SK_Ca channels, based on our MS data and the Ca²⁺ sensitivity of the mSK_Ca channel, the Ca²⁺calmodulin–binding domain and the S6 transmembrane region appear to be conserved. Indeed, the block of channel activity by apamin showed that this mSK_Ca channel does share a pharmacological property similar to the plasmalemmal SK_Ca.

However, the planar lipid bilayer experiments appear to indicate that the apamin binding site and the Ca²⁺ binding site are both localized to the same side of the mSK_Ca channel. This was an unexpected finding because apamin has been reported to be an external pore blocker that binds to the outer pore region of the plasmalemmal SK_Ca channel [61], whereas the Ca²⁺ sensing region was believed to be on the intracellular side, conferred by calmodulin that is constitutively bound to the C-terminus of the channel [12]. Consequently, our findings would imply that the position of the C-terminus in the mSK_Ca channel differs from that in the plasmalemmal SK_Ca channel. Therefore, based on the sidedness of the apamin effect and Ca²⁺ sensitivity, together with our observed biophysical properties, the mSK_Ca channel may exhibit some functional and structural differences from the plamalemmal SK_Ca channel. Additional experiments will be needed to confirm this possibility.

Our study represents the first conclusive report that identifies SK_Ca channels in cardiac myocyte mitochondria. Their presence in the IMM would indicate that they have an important function in fine-tuning regulation of mitochondrial bioenergetics, perhaps via volume control, which is largely controlled by K⁺ flux. In contrast, the voltage and Ca²⁺-dependent BK_Ca channels may open only when ΔΨ_m is high (state 4 respiration) or in response to a large imbalance in m[Ca²⁺] or cytosolic [Ca²⁺], much as BK_Ca channels regulate cell membrane potential in excitable cells.

Three genes encode the SK_Ca family; all have been cloned, and the amino acid sequences predict subunits similar to those in other K⁺ channels. Channel specificity resides in the C terminal domains where each SK_Ca subtype interacts with the ubiquitous Ca²⁺ sensor CaM. This constitutive binding domain is called CaMBD [62]. Crystal structures show that SK_Ca+CaMBD contains two EF hand motifs within each of the globular N and C terminal regions separated by a flexible linker [62]. The C terminus is required to establish the link of SK_Ca and CaM. Substitution of neutral amino acids for aspartate and glutamate only in N terminus EF hand region blocks Ca²⁺ gating. This binding site is positioned just below the K⁺ selectivity filter, which suggests that conformational changes near or even in the selectivity filter itself function to gate SK_Ca channels. The 18 amino acid bee venom toxin apamin is highly selective for SK2 by docking at the pore entrance and between the S3–S4 loops [63].

SK_Ca channels in neurons lie adjacent to Ca²⁺ stores and Ca²⁺ channels. In nerve cells SK_Ca channels play a role in setting the intrinsic firing frequency, while BK_Ca channels regulate action potential shape and may contribute to the unique climbing fiber response [15]. The K⁺ flux mediated by BK_Ca and SK_Ca channels in mitochondria may be differentiated by both their sensitivities to Ca²⁺ and dependence on ΔΨ_m during states 3 and 4 respiration. Because there are many differences between these channels, the functional effects of opening these channel are expected to differ; e.g., unlike BK_Ca, SK_Ca channel opening may “fine tune” matrix K⁺ influx due to changing Ca²⁺ levels independent of changes in ΔΨ_m during the variable rate of oxidative phosphorylation.

4.2. mSK_Ca channel opening triggers preconditioning via ROS

The presence of both SK_Ca and BK_Ca channels in cardiac myocyte IMM indicates a functional importance for these channels during excess mCa²⁺ loading; moreover their endogenous opening during IPC, or as a pharmacological therapy, may be an important trigger for cardioprotection. It is unclear if these drugs actually open these K⁺ channels directly to elicit preconditioning, or if they themselves alter mitochondrial bioenergetics (as mild uncouplers of oxidative phosphorylation), which mediates the memory of preconditioning by other downstream effectors. Although the mitochondrial preconditioning effect of DCEB appears to require both mSK_Ca channel opening and generation of O₂^•−, these factors are not effectors of PPC as the DCEB and TBAP are washed out before ischemia.

There is ample evidence that O₂^•− is necessary to trigger PPC by mK⁺ channel openers but the mechanism of O₂^•−, and its products or reactants, in mediating PPC is unknown. An increase in redox state (increased NADH, decreased FAD) at a given [O₂] can result in increased O₂^•− generation [64]. O₂ derived free radical “bursts” are known to occur during reperfusion when excess O₂ is available. Our group [4, 26, 27, 33] and others [35, 65] have shown, moreover, that ROS are also formed in excess during ischemia before reperfusion when tissue O₂ tension decreases, the redox state increases and then decreases, and cytochrome c oxidase (complex IV) activity is low [66]. The putative mK_ATP channel opener diazoxide [67] mimicked IPC on reducing infarct size and the ROS scavengers, N-acetylcysteine [68] or N-mercapto-propionyl-glycine [24], blocked the preconditioning effect of diazoxide. It has been suggested that mK_ATP channel opening can cause a small increase in ROS formation [69], which may trigger cardioprotection through activation of protein kinases. Conversely, ROS have also been proposed to activate the sarcolemmal K_ATP channel by modulating its ATP binding sites as this effect is blocked by GLIB or by ROS scavengers [70]. Others have proffered that ROS produced during IPC may afford cardioprotection on reperfusion directly, or via a feed forward mechanism for K_ATP channel-induced ROS production [71, 72].

In the present study evidence that the protective effect of DCEB is mediated by ROS is indicated by reversal of the protection in the presence of TBAP. O₂^•− or OH^•, or even non-radical reactants like H₂O₂ or ONOO⁻ (formed in the absence or presence of NO^•, respectively) may actually produce the preconditioning responses, but O₂^•− appears necessary to initiate the response. It is also possible that mSK_Ca, mBK_Ca, and or mK_ATP channel activation is altered by ROS as a feed forward controller of mitochondrial function. Enhanced electron transfer before ischemia may minimize respiratory inefficiency, i.e., reduced matrix contraction and improved respiration on reperfusion. mSK_Ca channel opening, like mK_ATP channel opening, and indeed opening of any mK⁺ channel, could induce PPC by mildly enhancing O₂^•− generation, which stimulates enzymatic pathways that help to protect the cell from IR injury. Interestingly, we observed that O₂^•− dismutation by TBAP blocked protection by DCEB.

When DCEB was given alone or with any of the inhibitors, it had no direct detectable effect on mechanical function or mitochondrial bioenergetics (redox state, O₂^•− levels) in isolated hearts. In our related study [4] neither NS1619 nor its antagonist PAX showed any direct effect on measured variables. In the ex vivo, intact heart perfused with adequate substrates and O₂, mitochondria are mostly respiring in the non-resting state 3, so small changes in O₂^•− between states 3 (ample ADP) and 4 (consumed ADP) respiration can not be observed. However, in isolated cardiac mitochondria we reported that low, but not high, concentrations of the BK_Ca channel opener NS1619 can increase resting state 4 respiration and ROS generation while maintaining IMM potential (ΔΨ_m) [6].

We propose similarly that DCEB, like NS1619, increases intramatrix K⁺, which is replaced immediately with H⁺ via KHE. We suggest that at low concentrations of these openers a transient increase in matrix acidity, i.e., via a proton leak, stimulates respiration but maintains ΔΨ_m so that a greater amount of O₂^•− is generated at mitochondrial respiratory complexes due to impaired electron transport. O₂^•− itself, or a reactant, may in turn stimulate downstream-induced phosphorylation pathways fostering K⁺ channel opening as necessary when ischemia occurs. The net effect could be preservation of mitochondrial bioenergetics during ischemia as evidenced by better maintenance of the reduced state (high NADH and low FAD) and smaller increases in O₂^•− generation and less m[Ca²⁺] overload. This could lead to better preservation of oxidative phosphorylation and ATP turnover leading to better utilization of ATP on initial reperfusion after ischemia.

4.3. Putative mechanism of mitochondrial K⁺ flux on mitochondrial protection during IR injury

BK_Ca and SK_Ca channel openers appear to have a profound ability to induce PPC but the mechanism is unclear. It is possible that brief ischemia as in IPC causes a slight elevation of mCa²⁺ that induces mSK_Ca and mBK_Ca channel opening and, like mK_ATP channel opening, leads to partial dissipation of ΔΨ_m and or matrix swelling as a protective mechanism against subsequent IR injury. It is now clear that K⁺ is required for optimal functioning of oxidative phosphorylation because matrix K⁺ flux largely regulates matrix volume and can modulate ΔΨ_m [73–75]. Trans-matrix K⁺ flux can also modulate ROS production [6]. mSK_Ca channels, like mBK_Ca channels [57, 76], may act to modulate matrix volume during times of increased matrix Ca²⁺ load, such as occurs during IR injury [25, 29]. Xu et al. [57] first suggested that opening mBK_Ca channels to enhance matrix K⁺ influx is an important factor in mitigating IR injury in a manner similar to mK_ATP channels. They proposed [57] that the function of mBK_Ca channels was to improve the efficiency of mitochondrial energy production.

As with the other two K⁺ channels reported in mitochondria, K_ATP and BK_Ca, once the K⁺ channel is opened the increase in K⁺ uptake leads to changes in the matrix as described by Garlid et al. and Beavis et al. [77, 78]. Electrogenic H⁺ efflux driven by the respiratory chain is balanced by electrophoretic K⁺ influx. If this were uncompensated, it would cause a very large increase in matrix pH of about 2 pH units. Partial compensation is provided by electroneutral uptake of substrate anions, such as phosphate. The compensation is partial because the concentration of phosphate in the cytosol is much lower than that of K⁺, and this imbalance leads to matrix alkalinization [79, 80]. Matrix alkalinization now releases the K⁺/H⁺ antiporter from inhibition by matrix protons [75], causing K⁺ efflux to increase in response to increased K⁺ uptake until a new K⁺ steady state is achieved. An increase in futile K⁺ cycling is believed to produce mild uncoupling [78] and regulates mitochondrial bioenergetics and ROS emission.

Although mK_Ca channels likely play a role in regulating mitochondrial bioenergetics, it is unknown how opening of these channels leads to more normalized NADH/FAD levels, reduces excess ROS, and decreases Ca²⁺ loading during IR. Just as the existence and function of the mK_Ca channel in IMM on mitochondrial respiration is unclear, so is the mechanism of K⁺ influx via mK_ATP channels in IMM [73, 76, 81–84]. For the K_ATP channel it was proposed that its opening depolarizes the IMM to cause uncoupling and hasten respiration [76, 81, 85]. Subsequent ischemia would then reduce the driving force for Ca²⁺ influx through the mCa²⁺ uniporter; this could attenuate mCa²⁺ overload [86, 87] so that energized mitochondria on reperfusion would perform more efficiently. Indeed the putative mK_ATP channel opener diazoxide is reported to reduce the rate of mCa²⁺ uptake by depolarizing the IMM and decreasing the driving force for mCa²⁺ entry [76, 81, 85], although this could be due to respiratory inhibition distinct from K_ATP channel opening [73, 88].

Garlid’s group [73, 88], however, proposed that the physiological role of potential mK⁺ channels is control of matrix volume rather than dissipation of ΔΨ_m and uncoupling. They postulated that matrix swelling by K⁺ uptake is caused by concomitant uptake of Cl⁻ and water by osmosis. But subsequent activation of mKHE may only slightly dissipate the proton gradient (ΔµH) by increasing matrix acidity (proton leak) without significantly altering ΔΨ_m [74, 88]. In turn, mitochondrial swelling might optimize mitochondrial function because partial uncoupling was seen to improve efficiency of oxidative phosphorylation [89]. More specifically, during hypoxia matrix K⁺ influx appears to maintain a normal matrix volume, which preserves a narrow intermembrane space and helps to facilitate energy transfer to ATP-utilizing sites, to reduce outer membrane permeability to nucleotides, and to slow ATP hydrolysis [73–75]. The end result of mSK_Ca channel opening, like mK_Ca channel opening, may be to improve mitochondrial efficiency, reduce m[Ca²⁺] and ROS production, and thereby to protect overall mitochondrial function during IR.

4.4. Interrelationship and timing of mCa²⁺ loading, ΔΨ_m, redox state, and ROS during cardiac IR injury

Prolonged mitochondrial ischemia is marked by the following: decreasing ΔΨ_m, an oxidized redox state, excess ROS, matrix contraction, and increasing mCa²⁺ loading. Ca²⁺ overload due to leaky IMM could impede normal electron transfer so that greater amounts of ROS are produced during IR. Alternatively, ROS can damage membranes by lipid peroxidation; this can hamper selective permeability to ions and allow cytosolic and mCa²⁺ uptake as a result of increased reverse mode sarcolemmal Na⁺/Ca²⁺ exchange (NCE) [90, 91].

Our studies in the intact heart model show an interrelationship between O₂^•− produced, redox state, and mCa²⁺ influx during ischemia. Continuously measured NADH and O₂^•− changed together during ischemia as well as during reperfusion. Ischemia-induced rises in NADH [4, 25, 28, 30, 33], ROS [4, 26, 27, 33], and m[Ca²⁺] [4, 25, 29, 33] returned closer to normal values on reperfusion after PPC. These effects were reversed by ROS scavengers or by blocking sarcolemmal K_ATP and/or mK_ATP channel opening with GLIB or 5-hydroxydecanoate [27, 29]. Preconditioning also led to reduced ROS generation and improved ATP synthesis in isolated mitochondria [92]. These studies suggest that temporary exposure to distinct cardioprotective drugs before ischemia causes ROS-dependent changes in mitochondrial bioenergetics that initiates a preconditioning effect. mK_Ca is likely to be activated endogenously as matrix Ca²⁺ rises in response to an increase in Ca²⁺ load, such as occurs during ischemia; opening these channels pharmacologically before ischemia may lead to added protection.

4.5 Summary and limitations

We have furnished ample evidence for the presence of SK_Ca channels in purified mitochondria and in IMM from cardiac cells, for the functional effects of the IK_Ca and SK_Ca channel opener DCEB on K⁺ flux in isolated mitochondria, and for the channel conductance of SK_Ca proteins incorporated into planar lipid bilayers. Moreover, we have demonstrated that SK_Ca channel opener DCEB initiates cardiac PPC as shown by marked metabolic and functional improvements during reperfusion. These are supported by better preserved reduced redox state (high NADH and low FAD), decreased O₂^•− production, reduced mCa²⁺ loading during IR, and reduced infarct size. The protection by DCEB was blocked by dismutation of O₂^•− with TBAP and by the SK_Ca antagonist NS8593. It is possible that mSK_Ca channel opening induces a mild proton leak due to mKHE, which accelerates respiration, but maintains ΔΨ_m, so that small amounts of generated O₂^•− triggers a downstream protective pathway.

All of the K⁺ channel agonists may converge on a pathway that stimulates a small amount of ROS. That TBAP blocks protection by this drug and that mitochondria are a major source of ROS, suggest that DCEB exerts its effects primarily in mitochondria. Only relative changes in NADH and FAD levels and ROS formation can be assessed in our model. We did not test if the mSK_Ca channel is open during IR injury, although we have preliminary evidence that the BK_Ca channel is open during reperfusion [93]. It is plausible that some factors that induce preconditioning, like small increases in ROS or m[Ca²⁺], are the same factors, albeit at much greater levels, that cause IR damage. Thus the individual stages of triggering, activation and end-effect must be well delineated to unravel the complicated mechanism underlying the cardiac protection afforded by preconditioning.

Highlights.

• We identify small Ca2+-sensitive K+ channels in guinea pig cardiac mitochondria.

• Identity based on Westerns, confocal microscopy, immune-gold EM, and MS.

• Function based on K+ flux, Ca2+, apamin, and UCL1684–sensitive conductance.

• Channel agonist, DCEBIO protects hearts against IR injury, and antagonist NS8593 blocks protection.

• Mechanism of protection in isolated hearts is mediated by the superoxide radical.

Abbreviations

IR

ischemia reperfusion

SK_Ca

small conductance Ca²⁺ -sensitive K⁺ channel

BK_Ca

big conductance Ca²⁺ -sensitive K⁺ channel

K_ATP

ATP -sensitive K⁺ channel

DCEB

5,6-dichloro-1-ethyl-1,3-dihydro-2H-benzimidazol-2-one

IMM

inner mitochondrial membrane

TBAP

Mn(III) tetrakis (4-benzoic acid) porphyrin

PPC

pharmacological preconditioning

TRAM

TRAM-34: 1-[(2-chlorophenyl) (diphenyl)methyl]-1H-pyrazole

GLIB

glibenclamide

PAX

paxilline

BSA

bovine serum albumin

IEM

immune-electron microscopy

MS

mass spectroscopy

NS8593

N-[(1R)-1,2,3,4-tetrahydro-1-naphthalenyl]-1H-benzimidazol-2-amine hydrochloride

UCL 1684

6,10-diaza-3(1,3)8,(1,4)-dibenzena-1,5(1,4)-diquinolinacy clodecaphane