Functional Distinctions between the Mitochondrial ATP-dependent K (cid:1) Channel (mitoK ATP ) and Its Inward Rectifier Subunit (mitoKIR)*

The ATP-sensitive potassium channel from the inner mitochondrial membrane (mitoK ATP ) is a highly selec- tive conductor of K (cid:1) ions. When isolated in the presence of nonionic detergent and reconstituted in liposomes, mitoK ATP is inhibited with high affinity by ATP ( K 1/2 (cid:2) 20–30 (cid:3) M ). We have suggested that holo-mitoK ATP is a heteromultimer consisting of an inwardly rectifying K (cid:1) channel (mitoKIR) and a sulfonylurea receptor (Grover, G. J., and Garlid, K. D. (2000) J. Mol. Cell. Cardiol. 32, 677–695). Here, we show that a 55-kDa protein isolated by ethanol extraction and reconstituted in bilayer lipid membranes and liposomes is the mitoKIR. This protein, which lacks the sulfonylurea receptor subunit, is inhibited with low affinity by ATP, with K 1/2 (cid:1) 550 (cid:3) M . ATP inhibition of both mitoKIR and holo-mitoK ATP is re- versed by UDP ( K 1/2 (cid:2) 10–15 (cid:3) M ). Holo-mitoK ATP is opened Electrophysiological mitochondria protein/ml Tris-HCl (cid:5) The pellet resuspended m M Tris-HCl (pH protein/ml, 66% and super- natant dialyzed Tris-HCl (pH The dialyzed particulate onto 1-ml DEAE-cellulose pre-equilibrated M Tris-HCl, g/liter mercaptoethanol, and m M EDTA (pH The bound were eluted with a KCl step and the active was eluted with m M KCl column for 3 h, a and again eluted with a KCI step gradient, be recovered in the m M KCl ATP in Liposomes— Rat brain mitochondria were solubilized in Triton X-100 and purified on DEAE-cellulose, as described previously (12). The active fraction, elut- ing with 250 m M KCl, was dialyzed overnight at 4 °C against column buffer and reconstituted into liposomes containing potassium-binding benzofuran isophthalate (PBFI) (12). Proteins in the fraction were visualized by Coomassie Blue R-250 staining of SDS-PAGE gels (18). The internal medium contained tetraethylammonium cation (TEA (cid:2) ) salts of sulfate (100 m M ), EDTA (1 m M ), HEPES (25 m M ), pH 6.8, and 2 m M PBFI. 15 (cid:1) l of stock vesicles (50 mg phospholipid/ml of stock) were diluted in 2 ml of external medium containing 150 m M KCl, 1 m M TEA-EDTA, 1 m M MgCl 2 , and 25 m M TEA-HEPES (pH 7.2). Electrophoretic K (cid:2) flux was initiated by the addition of 0.3 (cid:1) M carbonyl cyanide m -chlorophe-nylhydrazone (to provide charge compensation via H (cid:2) efflux), and the K (cid:2) influx was recorded as increased fluorescence of intraliposomal PBFI using an SLM/Aminco 8000C fluorescence spectrophotometer ( (cid:2) ex / (cid:2) em (cid:7) 345/485 nm). The K (cid:2) response of intravesicular PBFI was cali- brated by stepwise additions of KCl to proteoliposomes in the internal medium after the addition of 0.5 (cid:1) M nigericin and 1 (cid:1) M tributyltin (19). Results are plotted as a normalized flux value (cid:3) J / (cid:3) J control , where (cid:3) J control is the difference between fluxes in the absence and presence of 200 (cid:1)

by diffusion and by means of ATP-sensitive potassium channel from the inner mitochondrial membrane (mitoK ATP ). At the high values of ⌬⌿ maintained by mitochondria, both of these processes increase exponentially with ⌬⌿ (3,4) and are consequently very sensitive to fluctuations in ⌬⌿. These fluctuations, in turn, are high in tissues such as heart, which undergo large variations in energy demand and ATP synthesis rates (5). Thus, regulation of K ϩ influx and efflux pathways can be seen as a means of regulating volume in the face of the changing energy requirements of the cell. MitoK ATP plays more than a housekeeping role in cell physiology. There is now general agreement that mitoK ATP plays a key role in cardioprotection against ischemia-reperfusion injury (6,7). The proposed mechanisms of this protective effect of mitoK ATP opening (5,8,9) are plausible; however, it is evident that more needs to be known about the functional properties of mitoK ATP before its role in vivo can be ascertained.
By using a novel ethanol extraction technique, Mironova et al. (10) were the first to report reconstitution in lipid bilayer membranes of a 55-kDa K ϩ channel from mitochondria.
Paucek et al. (3) used a detergent extraction technique and were the first to report reconstitution of mitoK ATP in liposomes. The latter channel was associated with two proteins of molecular mass 55 and 63 kDa, and we hypothesized that mitoK ATP is a heteromultimeric complex consisting of a 55-kDa inwardly rectifying K ϩ channel (mitoKIR) and a 63-kDa sulfonylurea receptor (mitoSUR), analogous to the plasma membrane ATPdependent K ϩ channel (cellK ATP ) (11,12).
In this report, we focus on three interactions that address the key question of whether the 55-kDa K ϩ channel protein observed in the ethanol purification is the same as the 55-kDa protein purified with detergents. First, we show that UDP reverses ATP-inhibition of K ϩ flux mediated by both mitoKIR and mitoK ATP reconstituted in liposomes. Moreover, UDP exerts the same action in isolated mitochondria, and the affinities for the opening effect of UDP are about the same in each preparation. Second, we show that the mitoKIR opener p-diethylaminoethylbenzoate (DEB) also activates K ϩ flux via mi-toK ATP in isolated mitochondria and this effect is inhibited by 5-HD and glibenclamide.
Third, we show that tetraphenylphosphonium ion (TPP ϩ ) inhibits reconstituted mitoKIR and mitoK ATP with similar affinities. We conclude that the 55-kDa protein obtained by ethanol extraction is the channel component of mitoK ATP .

EXPERIMENTAL PROCEDURES
Mitochondrial Preparations-The experiments in this paper utilized mitochondria isolated from rat liver (13), rat heart (14), and rat brain cortex (15). Different mitochondrial preparations were used in different protocols, largely for practical reasons. We have shown previously that the properties of liver, heart, and brain mitochondria are qualitatively and quantitatively similar (12). * This work was supported by National Institutes of Health Grants HL67842, HL36573, and TW01116 (to K. D. G.) and Russian Foundation for Basic Research Grants 01-04-48555, 04 -04-97281 (to G. D. M.), and AHA 963 0004N (to P. P.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Electrophysiological Measurements of mitoKir in Lipid Bilayer Membranes-Rat liver mitochondria were diluted to 7 mg of protein/ml with 20 mM Tris-HCl (pH 7.4), stirred for 20 min at 4°C, and centrifuged at 7000 ϫ g for 15 min at 4°C. The pellet was resuspended in 10 mM Tris-HCl (pH 7.4) to 44 mg of protein/ml, and this was diluted with 10 parts of 66% ethanol in water (Ϫ20°C). After stirring for 30 min at 4°C, the suspension was centrifuged at 7000 ϫ g for 15 min, and the supernatant was dialyzed overnight against 5 mM Tris-HCl (pH 7.4), 0.5 g/liter mercaptoethanol at 4°C. The dialyzed sample was centrifuged at 100,000 ϫ g for 30 min to remove remaining particulate matter and then loaded onto a 1-ml DEAE-cellulose column pre-equilibrated with column buffer containing 50 mM Tris-HCl, 0.5 g/liter mercaptoethanol, and 1 mM EDTA (pH 7.4). The bound proteins were eluted with a KCl step gradient, and the active fraction was eluted with 250 mM KCl (16). This fraction was dialyzed against column buffer for 3 h, applied to a second DEAE column (1 ml), and again eluted with a KCI step gradient, to be recovered in the 250 mM KCl fraction.
Ion conductance of the partially purified 55-kDa protein, mitoKIR, was measured in lipid bilayer membranes (17) using bovine brain lipid (20 mg/ml) plus 10% cardiolipin dissolved in decane. The solution on both sides of the bilayer contained 100 mM KCl, 20 mM Tris-HCl (pH 7.4). Potentials (10 -100 mV) were applied across the membrane, and currents were registered using an operational amplifier (AD 711C, Analog Devices).
15 l of stock vesicles (50 mg phospholipid/ml of stock) were diluted in 2 ml of external medium containing 150 mM KCl, 1 mM TEA-EDTA, 1 mM MgCl 2 , and 25 mM TEA-HEPES (pH 7.2). Electrophoretic K ϩ flux was initiated by the addition of 0.3 M carbonyl cyanide m-chlorophenylhydrazone (to provide charge compensation via H ϩ efflux), and the K ϩ influx was recorded as increased fluorescence of intraliposomal PBFI using an SLM/Aminco 8000C fluorescence spectrophotometer ( ex / em ϭ 345/485 nm). The K ϩ response of intravesicular PBFI was calibrated by stepwise additions of KCl to proteoliposomes in the internal medium after the addition of 0.5 M nigericin and 1 M tributyltin (19). Results are plotted as a normalized flux value ⌬J/⌬J control , where ⌬J control is the difference between fluxes in the absence and presence of 200 M ATP, and ⌬J is the difference between fluxes in the presence or absence of the K ATP channel opener or blocker tested. K 1/2 values and Hill coefficients were determined from non-linear least-square-fitting to the Hill equation.
Potassium Flux Measurements of MitoKIR in Liposomes-Rat brain mitochondria were diluted to 10 mg/ml in distilled water, then diluted with 10 parts of 66% ethanol solution in water (Ϫ20°C). The remainder of the extraction and purification on DEAE-cellulose followed the procedure described for mitoKIR studies in lipid bilayer membranes (see above). This active fraction was visualized by silver staining (Bio-Rad) of SDS-PAGE gels (18). The final mitoKir fraction was incorporated into PBFI-containing liposomes using the same technique and buffers described above for mitoK ATP obtained by Triton X-100 extraction.
Chemicals-PBFI was purchased from Molecular Probes; electrophoresis chemicals and SM-2 Bio-Beads from Bio-Rad; L-soybean lecithin from Avanti; and column resins, nucleotides, ionophores, buffers, and other chemicals were purchased from Sigma.

Isolation and Partial Purification of Holo-mitoK ATP and
MitoKIR from Rat Liver Mitochondria- Fig. 1 contains the SDS-PAGE gel patterns of the two preparations described under "Experimental Procedures." Fig. 1, lane 1 contains standards; lane 2 contains the reconstitutively active mitoK ATP fraction isolated and purified in the presence of Triton X-100. This fraction contains two bands, at 55 and 63 kDa, and presents the same appearance as mitoK ATP isolated from rat brain mitochondria (12). Fig. 1, lane 3 contains the reconstitutively active mitoKIR fraction extracted with ethanol and purified in the absence of detergents. This fraction contains a single 55-kDa protein band. It should be noted that 2-dimensional electrophoresis reveals that each of these bands contains at least five different proteins.
The Effects of ATP and Mg 2ϩ on MitoKIR and Holo-mi-toK ATP - Fig. 2 contains the concentration dependence of ATP inhibition of the channels in the presence (q) or absence (E) of 500 M Mg 2ϩ . These experiments were performed on four different preparations: lipid bilayer membranes containing mi-toKIR ( Fig. 2A), liposomes containing mitoKIR (Fig. 2B), liposomes containing holo-mitoK ATP (Fig. 2C), and isolated mitochondria containing holo-mitoK ATP (Fig. 2D).
We note the following general characteristics of these preparations with respect to ATP inhibition: (i) the apparent affinities for ATP differ widely, with K 1/2 values of about 600 M in liposomes or lipid bilayer membranes containing mitoKIR, 20 -30 M in liposomes containing holo-mitoK ATP (3,23), and 1 M in mitochondria; (ii) ATP inhibition of holo-mitoK ATP in liposomes or mitochondria exhibits an absolute requirement for Mg 2ϩ (3,23), whereas ATP inhibition of mitoKIR is indifferent to Mg 2ϩ ; (iii) we always observed a 10-fold higher ATP affinity in isolated mitochondria over liposomes (24), and we have no explanation for this finding at present.
The Effects of UDP upon MitoKIR Reconstituted in Lipid Bilayer Membranes-When reconstituted in lipid bilayer membranes, the 55-kDa protein forms potassium channels with a unitary conductance of about 10 pS (16,26,27). The current-voltage plot of the channel exhibits characteristic rectifying properties, as shown in Fig. 3A. We frequently observed conductances of up to 100 pS (Fig. 3B), which we have interpreted as clusters of 10 elementary channels (28). Note that in Fig. 3B, current is inhibited by 1 mM ATP, and the ATP-inhibited current is restored by 20 M UDP. Therefore, UDP is a mitoKIR opener. It can be seen in Fig. 3B that UDP restored K ϩ conductance to levels higher than those in the control. We attribute this overshoot to UDP activation of channels that had lost their activity with time. Spontaneous decline in channel activity (rundown) has been described previously for both mitoKIR (29) and cellK ATP (30). We discovered that the loss of channel activity because of rundown could be reversed by the addition of 100 -300 M UDP (Fig. 3C). UTP was less effective than UDP, and UMP and CDP had no effect (results not shown).
UDP Opens MitoKIR and Holo-mitoK ATP - Fig. 4A contains representative K ϩ flux traces from liposomes reconstituted with holo-mitoK ATP . Electrophoretic K ϩ flux (Fig. 4A, trace a) was inhibited by 200 M ATP (Fig. 4A, trace d), and flux was partially restored by 100 M UDP (Fig. 4A, trace b). The reactivation by UDP was largely abolished in the presence of 200 M 5-HD, a specific inhibitor of mitoK ATP (22) (Fig. 4A, trace c). Fig. 4B contains representative swelling traces from isolated mitochondria containing holo-mitoK ATP . As reported previously (22), matrix swelling secondary to K ϩ uptake (Fig. 4B,  trace a) was inhibited in the presence of 100 M ATP (Fig. 4B,  trace d). Swelling was restored by the addition of 30 M UDP (Fig. 4B, trace b) and re-inhibited by 200 M 5-HD (Fig. 4B,  trace c). In the absence of added ATP, UDP did not enhance mitochondrial K ϩ uptake (data not shown). Each of these traces was duplicated in medium containing TEA ϩ instead of K ϩ . As we showed previously for ATP (22,24), none of these agents had any effect upon the light-scattering signal in TEA ϩ medium, showing that the matrix swelling was due specifically to uptake of K ϩ salts and water.
The data in Fig. 5 compare the UDP concentration dependences for reactivation of mitoKIR in lipid bilayer membranes (Fig. 5A) and of holo-mitoK ATP in liposomes (Fig. 5B). The K 1/2 (UDP) is 10 -13 M in both preparations. In additional experiments (not shown), we found that the K 1/2 (UDP) was not affected by increasing the ATP concentration to 500 M in liposomes and 3 mM in lipid bilayer membranes. Thus, UDP interaction with mitoK ATP seems to be non-competitive with ATP.
DEB Opens MitoKIR and Holo-mitoK ATP -DEB, functioning as an electron donor, was previously shown to reverse ATP inhibition of mitoKIR at micromolar concentrations (29). In the present study, we found that DEB also reversed ATP inhibition of K ϩ flux through holo-mitoK ATP in isolated rat heart mitochondria (Fig. 6A). Opening by DEB was inhibited by 5-HD (50 M) and glibenclamide (2 M) (Fig. 6A). The concentrationdependence of DEB opening of mitoK ATP in mitochondria is shown in Fig. 6B, with K 1/2 about 10 M.
TPP ϩ Inhibits MitoKIR and Holo-mitoK ATP -100 nM TPP ϩ was found to inhibit the conductance of lipid bilayer membranes containing mitoKIR (Fig. 7A). TPP ϩ also inhibited mi-toKIR and holo-mitoK ATP reconstituted in liposomes, with K 1/2 of 44 -50 nM (Fig. 7B). DISCUSSION For insights into mitoK ATP function, it is useful to consider the salient aspects of plasma membrane K ATP channels (cellK ATP ). CellK ATP consists of two subunits, an inward-rectifying potassium channel, Kir6.1 or Kir6.2, and a regulatory sulfonylurea receptor, SUR1, SUR2A, or SUR2B (see review by Ashcroft  This experiment is representative of more than five independent assays. B, UDP reverses ATP inhibition of K ϩ flux in isolated mitochondria. Shown are typical traces of the light-scattering parameter versus time in isolated rat heart mitochondria respiring in medium described in "Experimental Procedures." Changes in ␤ are proportional to changes in mitochondrial volume, and mitochondrial volume increases secondarily to uptake of K ϩ salts and osmotically obligated water (20). Control flux (trace a) reflects K ϩ uptake via diffusion and mitoK ATP . The mitoK ATP component was inhibited by 100 M ATP (trace d). ABC-binding cassette family (32), and tetramers of KIRx and SURx peptides form functional hetero-octamers in the plasma membrane (33). An RKR domain in the C terminus of KIR6.x and in the 6th intracellular loop of SUR prevents membrane expression of either of these proteins in the absence of its partner (34). However, C-terminal truncations of KIR6.2 permit expression and measurements of channel activities (35). These studies have shown that ATP inhibits cellK ATP by interacting with KIR and that SUR1 enhances the direct blockage of KIR6.2 by ATP by an unknown mechanism that involves the N terminus of KIR6.2. Sulfonylureas and K ATP channel openers interact with sites on SUR (35,36).
MitoK ATP and cellK ATP react with the similar ligands, including K ϩ channel openers, sulfonylureas, nucleotides, and long-chain CoA esters. On this basis, we have proposed that they belong to the same gene family (37). Although verification of this hypothesis awaits the molecular structure of mitoK ATP , the functional parallels with cellK ATP lead to the working model contained in Fig. 8. Thus, the 55-kDa band seen in both preparations of Fig. 1 contains mitoKIR, and the 63-kDa band seen only in the detergent-extracted preparation contains mi-toSUR (12). A key question addressed by this report is whether the channel activity of the detergent-extracted mitoK ATP (3) is the same as that of the ethanol-extracted mitoKIR (10).
There is considerable evidence favoring this proposal: both fractions contain the 55-kDa band (Fig. 1), both channels are K ϩ -selective (3,16), both manifest an ϳ10-pS single-channel conductance in 100 mM KCl (16,38), and both channels are inhibited by ATP (Fig. 2). The 55-kDa protein isolated with ethanol is an inwardly rectifying K ϩ channel, mitoKIR (Fig.   FIG. 5. Concentration dependence of UDP reactivation of mi-toKIR and holo-mitoK ATP . Relative activation of the ATP-inhibited channels is plotted versus UDP concentration for two different preparations. A, mitoKIR reconstituted in lipid bilayer membranes. Channel current with applied voltage of 50 mV was inhibited by 1 mM ATP and reactivated by UDP. ⌬J/⌬J control is the normalized K ϩ current. ⌬J ϭ J ATPϩUDP Ϫ J ATP , and ⌬J control ϭ J o Ϫ J ATP , where J 0 is the current in nucleotide-free medium, and J ATP is the current in medium containing 1 mM ATP. K 1/2 (UDP) from three independent experiments was 9.6 Ϯ 0.5 M, with N H ϭ 1.0. Note that current is restored to greater than 100% of control, an effect that may be due to the effect of UDP upon rundown (see Fig. 3). B, holo-mitoK ATP reconstituted in liposomes. K ϩ flux was inhibited by 200 M ATP and reactivated by UDP. ⌬J/⌬J control is the normalized K ϩ flux, as described for A, except that J ATP is the flux in medium containing 0.2 mM ATP. K 1/2 (UDP) from three independent experiments was 13 Ϯ 2 M, with N H ϭ 1.

FIG. 6. DEB reverses ATP inhibition of holo-mitoK ATP.
A, in respiring rat heart mitochondria. Shown are relative rates of K ϩ uptake into mitochondria, from light-scattering experiments performed in K ϩ medium (as in Fig. 4B). 100 M ATP inhibited potassium influx in mitochondria, and this inhibition was reversed by 100 M DEB. Mi-toK ATP opening by DEB was prevented by 50 M 5-HD and 2 M glibenclamide. Three experiments were performed. B, in liposomes containing rat brain mitoK ATP. K ϩ flux was inhibited by 200 M ATP and reactivated by DEB. ⌬J/⌬J control is the normalized K ϩ flux, as described for Fig. 5B. K 1/2 (DEB) was 10 Ϯ 2 M, with N H ϭ 1 (two independent experiments). 3A), and the 63-kDa band is labeled with high affinity by sulfonylurea (12), consistent with identification of the 63-kDa band with mitoSUR (11).
It is also clear that the detergent-isolated fraction (holo-mitoK ATP ) differs from the ethanol-extracted fraction mitoKIR) in its regulation. Thus, ATP inhibition of holo-mitoK ATP exhibits an absolute requirement for Mg 2ϩ ions, and the apparent affinity for ATP is an order of magnitude higher (Fig. 2). Pharmacological blockers, such as glibenclamide and 5-HD, and pharmacological openers, such as cromakalim and diazoxide (6,7,22), are potent regulators of holo-mitoK ATP ; however, these agents have no effect upon channel activity of the mitoKIR subunit (28,29). CellK ATP exhibits similar properties. Thus, co-expression of SUR1 and KIR6.2⌬C36 resulted in ATP inhibition with higher affinity than expression of KIR6.2⌬C36 alone (35). Moreover, holo-cellK ATP was sensitive to sulfonylureas and diazoxide, whereas KIR6.2⌬C36 was not (35). Thus, the differences in regulation of mitoKIR and holo-mitoK ATP can be rationalized by assuming that mitoK ATP is a close relative of cellK ATP .
For a more direct test of the hypothesis that the 55-kDa protein is the channel component of mitoK ATP , we focused on three new regulatory interactions with mitoKIR and holo-mi-toK ATP . We found that 20 M UDP reversed ATP inhibition of mitoKIR in lipid bilayer membranes (Fig. 3B), and 100 -300 M UDP rescued mitoKIR from channel rundown (Fig. 3C). UDP also reversed ATP inhibition of holo-mitoK ATP in liposomes (Fig. 4A) and in intact mitochondria (Fig. 4B). MitoKIR and holo-mitoK ATP were opened by UDP with similar affinities of 10 -13 M (Fig. 5). UDP seems to be non-competitive with ATP. UDP is the first nucleotide shown to open mitoK ATP via the mitoKIR subunit, but there is precedent for this action from studies on cellK ATP : oleoyl CoA opens ␤-cell K ATP (39), and this effect has been shown to be mediated by means of the KIR 6.2 subunit (40,41). FIG. 8. Model of the interactions of regulatory ligands with mitoK ATP structure. The known structure-function of cellK ATP is used to interpret the known functional and binding properties of mitoK ATP and mitoKIR. The scheme summarizes the results of several studies, including those reported in this paper. MitoK ATP is thought to be composed of four subunits of a 55-kDa inwardly rectifying K ϩ channel (mitoKIR) and at least four subunits of a 63-kDa sulfonylurea-binding regulatory subunit (mitoSUR). MitoSUR is thought to contain one or two nucleotide binding folds (NBF). ATP binds to both subunits. ATP inhibition of holo-mitoK ATP exhibits an absolute requirement for Mg 2ϩ , whereas ATP inhibition of mitoKIR does not, suggesting that Mg 2ϩ binds to mitoSUR. GTP, which reverses ATP inhibition (42), interacts with mitoSUR. The classical K ATP channel openers (diazoxide, cromakalim, and P1075) and blockers (glibenclamide and 5-HD) interact with mitoSUR and have no effect upon mitoKIR. Some evidence (G. Mironova, P. Paucek, unpublished data) suggests that diazoxide and cromakalim interact with different regions of mitoSUR, as they have been shown to do with plasma membrane SURs (43). The openers UDP and DEB, and the blocker TPP ϩ , were shown to act upon mitoKIR. Because they interact directly with the channel, they exert similar effects upon holo-mitoK ATP .
FIG. 7. TPP ؉ inhibits mitoKIR and holo-mitoK ATP . A, mitoKIR reconstituted in lipid bilayer membranes. Inhibition of channel current was measured with an applied voltage of 50 mV. 100 -200 nM TPP ϩ caused complete inhibition of the channel. B, mitoKIR (q) and holo-mitoK ATP (E) reconstituted in liposomes. K ϩ flux was inhibited by the indicated concentrations of TPP ϩ . ⌬J/⌬J control is the normalized K ϩ flux, as described for Fig. 5B. The value for K 1/2 (TPP ϩ ) from the combined data was 44 nM, and the curve was drawn with N H ϭ 1. The results are typical of two experiments for each preparation. TPP ϩ is commonly used to estimate mitochondrial membrane potential. It is also a potent inhibitor of holo-mitoK ATP and mitoKIR (Fig. 7). Indeed, the apparent affinities of mi-toKIR and holo-mitoK ATP for TPP ϩ are identical: ϳ60 nM.
We suggested previously (29) that some agents may regulate the activity of mitoK ATP via intraprotein electron transport, and we showed that micromolar concentrations of the electroneutral reducing agent, DEB, opens the ATP-inhibited channel in lipid bilayer membranes. The DEB-activated mi-toKIR can be re-inhibited by pelargonidine, an electron acceptor. Here we show that DEB also opens holo-mitoK ATP in isolated mitochondria (Fig. 6A) and in liposomes reconstituted with mitoK ATP (Fig. 6B). Whereas DEB-activated mi-toKIR is not inhibited by 5-HD or glibenclamide, DEB-activated mitoK ATP is inhibited by both agents, consistent with their presumed effect upon mitoSUR.
Our strategy in these studies was to identify agents that act directly on mitoKIR and then to determine whether they have similar actions upon holo-mitoK ATP . This was found to be the case for UDP, DEB, and TPP ϩ (summarized in Fig. 8).
On the basis of these results, we conclude that the ethanolextracted mitoKIR is the channel responsible for the channel activity of mitoK ATP .