Reversal of the ATP-liganded state of ATP-sensitive K+ channels by adenylate kinase activity.

The mechanism that promotes transition from the ATP- to the ADP-liganded state of ATP-sensitive K+ (KATP) channels and consequent channel opening in a cytosolic environment of high ATP concentration has yet to be understood. A mechanism examined here that could reverse the ATP-inhibited state is based on the action of adenylate kinase to catalyze phosphoryl transfer between ATP and AMP, resulting in transformation of ATP into ADP. In membrane patches excised from guinea pig cardiomyocytes, AMP alone did not affect channel behavior but increased the open probability of ATP-inhibited KATP channels. This required MgCl2 and a hydrolyzable form of ATP and was prevented by P1,P5-di-adenosine-5′-pentaphosphate, an inhibitor of adenylate kinase. The single channel amplitude and kinetics of channel openings induced by the ADP-generating substrates of adenylate kinase, AMP and MgATP, were indistinguishable from the biophysical properties of the KATP channel exhibited after addition of MgADP. In whole cell voltage-clamped cardiomyocytes, introduction of exogenous adenylate kinase along with millimolar MgATP and AMP induced a K+ current that was suppressed by a sulfonylurea blocker of KATP channels. Enriched sarcolemmal membrane preparations were found to possess ATP·AMP phosphotransferase activity with properties attributable to an extramitochondrial isoform of adenylate kinase. These results indicate that adenylate kinase is a naturally occurring component of sarcolemmal membranes that could provide dynamic governance of KATP channel opening through its phosphoryl transfer catalytic action in the microenvironment of the channel.

The mechanism that promotes transition from the ATP-to the ADP-liganded state of ATP-sensitive K ؉ (K ATP ) channels and consequent channel opening in a cytosolic environment of high ATP concentration has yet to be understood. A mechanism examined here that could reverse the ATP-inhibited state is based on the action of adenylate kinase to catalyze phosphoryl transfer between ATP and AMP, resulting in transformation of ATP into ADP. In membrane patches excised from guinea pig cardiomyocytes, AMP alone did not affect channel behavior but increased the open probability of ATP-inhibited K ATP channels. This required MgCl 2 and a hydrolyzable form of ATP and was prevented by P 1 ,P 5di-adenosine-5-pentaphosphate, an inhibitor of adenylate kinase. The single channel amplitude and kinetics of channel openings induced by the ADP-generating substrates of adenylate kinase, AMP and MgATP, were indistinguishable from the biophysical properties of the K ATP channel exhibited after addition of MgADP. In whole cell voltage-clamped cardiomyocytes, introduction of exogenous adenylate kinase along with millimolar MgATP and AMP induced a K ؉ current that was suppressed by a sulfonylurea blocker of K ATP channels. Enriched sarcolemmal membrane preparations were found to possess ATP⅐AMP phosphotransferase activity with properties attributable to an extramitochondrial isoform of adenylate kinase. These results indicate that adenylate kinase is a naturally occurring component of sarcolemmal membranes that could provide dynamic governance of K ATP channel opening through its phosphoryl transfer catalytic action in the microenvironment of the channel.
ATP-sensitive K ϩ (K ATP ) 1 channels are involved in signaling networks that transduce cellular metabolic events into membrane potential changes and have been implicated in glucoseinduced insulin secretion in pancreatic ␤ cells or ischemiaassociated action potential shortening in heart muscle (1)(2)(3)(4)(5). Although the defining property of K ATP channels is their inhi-bition by intracellular ATP (5), which is readily demonstrable in excised membrane patches, as is the effect of ADP to reverse this ATP-inhibited state (1, 2, 6 -9), the mechanism by which opening of this channel is governed in situ has not been elucidated. The major question still to be resolved is how transition from the ATP-to the ADP-liganded state is accomplished.
The common assumption that changes in the cytosolic concentrations of adenine nucleotides are the sole determinant of K ATP channel opening has been contested (2, 4, 6, 10 -13). In cardiac cells, the ATP concentration (ϳ5-10 mM) exceeds by over 100-fold the IC 50 value for K ATP channel closure (12,13). This translates into a requirement for a change of two orders of magnitude in the intracellular ATP concentration, which is incompatible with cell viability, to achieve a mass action-induced change in the state of ATP-liganding and channel opening.
Considering that altered concentrations of cytosolic ATP and/or ADP are not readily detectable nor correlated with predictable changes in K ATP channel function, transition from the closed to the open state may be governed by a dynamic, rather than a static, property of adenine nucleotide metabolism. Although a component of the K ATP channel complex belongs to the ATP-binding cassette (ABC) superfamily of transporters (14 -16), essentially all members of which possess ATPase activity, ATP-phosphohydrolase has not been reported as an inherent activity of the K ATP channel. Therefore, identifying another catalytic mechanism to accomplish transformation of ATP to ADP is in order. Adenylate kinase, which catalyzes the phosphoryl transfer reaction between ATP and AMP resulting in the generation of ADP (17), conforms to these requirements. This process appears suitable since adenylate kinase in muscle has been proposed to operate as a transducing system coupling ATP signaling with its glycolytic generation. A key feature of this process is the conversion, at a regulatory site in glycolysis, of inhibitory ATP by the catalytic action of adenylate kinase using AMP as a phosphoryl acceptor (18). A second characteristic of this regulatory process is that it depends on the rate at which transformation occurs with no change in the cellular steady state levels of adenine nucleotide reactants (19). Also relevant is the observation that the insulin secretory response to glucose is correlated to the rate of adenylate kinase-catalyzed phosphoryl transfer (20).
Possible involvement of adenylate kinase in regulating the opening of cardiac sarcolemmal K ATP channels was examined by determining if the ATP-inhibited state of these channels can be reversed when the second substrate required for adenylate kinase catalysis, AMP, is provided and whether adenylate kinase can be identified as a sarcolemmal membrane-associated activity. Evidence is presented that adenylate kinase activity is detectable in cardiac sarcolemma, a locale which permits it to assure an AMP-dependent increase in K ϩ current of ATP-inhibited channels. An adenylate kinase-promoted transition from the ATP-to the ADP-liganded state could provide a mechanism for K ATP channel opening even in a cytosolic environment of ATP concentrations sustained at levels that exceed the channel's sensitivity toward ATP-induced closure.
Statistics-Data were expressed as mean Ϯ S.E. Significance of differences between two means was determined with the Student's t test, and a value of p Ͻ 0.05 was considered significant.
Materials-Chemicals were from Sigma.

AMP-dependent Increase in ATP-inhibited K ATP Channel Activity in Sarcolemmal
Patches-Upon excision of a sarcolemmal membrane patch from a cardiac cell, multiple openings of K ATP channels appeared and were suppressed by 100 M ATP (Fig.  1A). This ATP-inhibited K ATP channel activity was gradually reversed by 100 M AMP (Fig. 1A). Consequently, K ATP chan- 100 M AMP was added to the 100 M ATP-inhibited channels (n ϭ 8; Fig. 1A 2 ). Even when 1 mM ATP was used to inhibit K ATP channel activity, addition of AMP (200 M) enhanced the probability of K ATP channel opening (Fig. 1B). AMP alone did not increase the probability of K ATP channel opening without prior ATP-induced channel inhibition (Fig. 2). On average, the NP o was 3.9 Ϯ 0.8 prior to and 3.7 Ϯ 0.7 following addition of 100 M AMP (n ϭ 4; p Ͼ 0.05).
Requirements for AMP-induced Opening of ATP-inhibited K ATP Channels-The AMP-dependent increase in ATP-inhibited K ATP channel opening required both the presence of MgCl 2 and a hydrolyzable form of ATP. Under Mg 2ϩ -free conditions, K ATP channel opening (NP o at 3.6 Ϯ 0. Sensitivity to an Inhibitor of Adenylate Kinase-A Mg 2ϩ -dependent process in which AMP and ATP, but not ATP␥S, appear to serve as reactants defines the requirements of ATP⅐AMP phosphotransfer catalyzed by adenylate kinase. Treatment of membrane patches with P 1 ,P 5 -di-adenosine-5Јpentaphosphate (Ap 5 A; 20 M), a potent transition-state inhibitor of adenylate kinase-catalyzed phosphotransfer, prevented the increase in ATP-inhibited K ATP channel opening dependent on AMP plus Mg 2ϩ (Fig. 3). On average, the mean NP o , which was 3.9 Ϯ 0.6 in the absence of nucleotides and 1.9 Ϯ 0.  (Fig. 4B). Thus, the amplitude and kinetic properties of K ATP channels induced by the substrates of adenylate kinase, AMP plus Mg⅐ATP, were similar to openings induced by Mg⅐ADP, the product of ATP⅐AMP phosphotransfer.
Adenylate Kinase-induced Glyburide-sensitive Outward Current-Introduction of purified adenylate kinase (200 units/ml), along with its two substrates ATP (5 mM) and AMP (1 mM), to the pipette solution perfusing the interior of voltage-clamped cardiomyocytes produced a pronounced outward current (Fig.  5). During the first minute after whole cell patch formation, no significant outward current could be measured ( Fig. 5A and 5B, left panel), but after a lag of 5-10 min a prominent outward current developed ( Fig. 5A and 5B, right panel). On average, the value of the outward current reached 0.92 Ϯ 0.15 nA at a membrane potential of 0 mV (n ϭ 3). This current was suppressed by 10 M glyburide, a sulfonylurea blocker of K ATP channels (Fig. 5A). When ATP (5 mM) alone was added to the pipette solution, the outward current at a membrane potential of 0 mV was essentially equal to zero (not illustrated; n ϭ 19). In the presence of 5 mM ATP, addition of AMP (1 mM) to the pipette solution produced, within 15 min following patch formation, a small glyburide-sensitive outward current at positive potentials (0.23 Ϯ 0.07 nA at ϩ40 mV; n ϭ 3; not illustrated). Thus, introduction of exogenous adenylate kinase to the ATP plus AMP-containing pipette solution promoted the induction of a glyburide-sensitive K ATP current.
Occurrence of Adenylate Kinase Activity in Enriched Sarcolemmal Fraction-Cardiac muscle membranes were isolated by a procedure that enriches the sarcolemmal fraction of disrupted heart muscle. The sarcolemma specific Na,K-ATPase activity was used as a reference standard to assess the degree of sarcolemmal membrane enrichment, and the presence of the mitochondrial marker, succinic dehydrogenase, was used to determine the magnitude of mitochondrial contamination. The procedure yielded a preparation of membranes 6-fold enriched in Na,K-ATPase activity (with a specific activity of 146 Ϯ 2 nmol⅐min Ϫ1 ⅐mg protein Ϫ1 in the membrane fraction versus 25 Ϯ 2 nmol⅐min Ϫ1 ⅐mg protein Ϫ1 in unfractionated heart homogenate), and with reduced succinic dehydrogenase activity (from 125 Ϯ 2 nmol⅐min Ϫ1 ⅐mg protein Ϫ1 in the unfractionated homogenate to 75 Ϯ 3 nmol⅐min Ϫ1 ⅐mg protein Ϫ1 in the enriched sarcolemmal fraction). The ATP⅐AMP phosphotransferase activity in the sarcolemmal fraction was proportional to the concentration of protein (from 3.75 to 15 g/ml; n ϭ 16) present and exhibited the substrate requirements (Fig. 6A) and reversibility (Fig. 6B) characteristic of adenylate kinase. In the sarcolemma-enriched fraction, the specific activity of adenylate kinase, assayed in the direction of ADP formation with 1 mM AMP and 1 mM ATP, was 30 Ϯ 4 nmol⅐min Ϫ1 ⅐mg protein Ϫ1 (n ϭ 4; Fig. 6A). This represented a specific activity approximately one-third of that found in the unfractionated heart homogenate (i.e. 98 Ϯ 8 nmol⅐min Ϫ1 ⅐mg protein Ϫ1 , n ϭ 4), which contained the vast majority of the total adenylate kinase activity as non-membrane associated. Raising the ionic strength of the buffer in which the sarcolemmal membrane fraction was suspended with potassium acetate (0.5-1.5 M, 90 min at 4°C), followed by pellet separation using centrifugation (100,000 ϫ g for 1 h) and sucrose (0.5 M) gradient did not remove adenylate kinase activity, which was 121 Ϯ 6 before (n ϭ 4) and 143 Ϯ 1 nmol⅐min Ϫ1 ⅐mg protein Ϫ1 after such treatment (assayed in the formation of ATP production, n ϭ 4; not illustrated). The persistence of ATP⅐AMP phosphotransferase as a membrane-associated activity after several serial dilutions, resuspensions, and rehomogenizations of the fractionation procedure indicated that it most probably represents a species of adenylate kinase tightly bound to cardiac membranes. The phosphotransferase activity was highly sensitive to the relatively specific inhibitor Ap 5 A (35 Ϯ 1 and 85 Ϯ 2% inhibition at 1 and 10 M Ap 5 A, respectively; n ϭ 7) and relatively insensitive to inhibition by the less potent analog P 1 ,P 4 -di-adenosine-5Ј-tetraphosphate (14 Ϯ 3% at 10 M, n ϭ 4), a characteristic of an extra-mito-chondrial isoform of adenylate kinase (i.e. AK1). In addition, the adenylate kinase activity in the enriched sarcolemmal fraction exhibited acid stability (i.e. 121 Ϯ 6 before and 127 Ϯ 8 nmol⅐min Ϫ1 ⅐mg protein Ϫ1 after a 120-min-long treatment at pH 6 for 2 h, n ϭ 4), which is another property of the extramitochondrial isoform. Also, in the presence of 0.2% Triton X-100 (for 120 min at 0°C) the adenylate kinase activity present in the isolated sarcolemmal fraction was solubilized resulting in a 2-fold increase in activity (n ϭ 12).

DISCUSSION
The present study reveals that opening of ATP-inhibited sarcolemmal K ATP channels can be promoted by adenylate kinase activity within the microenvironment of the channel. This conclusion is supported by two lines of evidence: 1) the open probability of K ATP channels was up-regulated, in excised patches, by ADP-generating substrates of adenylate kinase and 2) adenylate kinase activity itself was found associated with sarcolemmal membranes where it catalyzed transformation of FIG. 5. Adenylate kinase induced outward current in whole cell voltage-clamped cardiomyocytes. Adenylate kinase (200 units/ ml; purified from rabbit muscle), ATP (5 mM), and AMP (1 mM) were added to the pipette solution. A, time course of development of adenylate kinase-induced current measured at the end of a 1000-ms-long depolarizing pulse applied from a holding potential of Ϫ50 to 0 mV. Glyburide was added to the bathing solution. B, current records obtained at 1 min (left panel) and 10 min (right panel) following whole cell patch formation. Holding potential was Ϫ50 mV. Rectangular 1000-mslong pulses were applied (in 10-mV steps) from Ϫ100 to ϩ40 mV (every 3 s). In this patch, with development of outward current, suppression of the inward component was observed.
FIG. 6. Adenylate kinase activity in isolated sarcolemmal preparations. A, adenylate kinase activity was present only when both AMP and ATP (1 mM, left; n ϭ 4) were provided to the enriched sarcolemmal membrane fraction but not in ATP (center; n ϭ 4) or AMP (right; n ϭ 4) alone. Adenylate kinase was measured by oxidation of NADH as ADP was produced from ATP and AMP. No ADP production was detected when sarcolemma was omitted (n ϭ 4; not shown), which indicated absence of contamination with exogenous adenylate kinase activity. B, consistent with its bidirectional property, adenylate kinase activity was also measured when ADP (2 mM; closed circles) was provided to the sarcolemmal membrane fraction but not in the absence of ADP (open circles). Adenylate kinase was assayed by measuring formation of NADPH as ATP was produced from ADP. Adenylate kinase activity was essentially linear with time (tested up to 50 min; n ϭ 6). ATP into ADP. These results suggest that reversal from the close to the open state of K ATP channels could be regulated in situ through localized phosphoryl transfer between ATP and AMP, resulting in transformation of ATP, the inhibitory ligand, into ADP, the channel activator.
Substrates of Adenylate Kinase Enhance the Open Probability of K ATP Channels-In the absence of ATP, AMP alone did not alter cardiac K ATP channel behavior, which is in accord with the sensitivity of these channels toward adenine nucleotides (1,2,4,6,16). But in the presence of ATP, AMP could increase the open probability of K ATP channels as previously reported in excised membrane patches from pancreatic ␤ cells (7). It is unlikely that the AMP-dependent activation of K ATP channels resulted from AMP acting as a partial antagonist, in the presence of the full channel antagonist ATP, since adenine nucleotides inhibit cardiac K ATP channels irrespectively of whether Mg 2ϩ is present or not (27), yet AMP required MgCl 2 to activate the opening of ATP-inhibited K ATP channels. Also, while poorly hydrolyzable ATP analogs do substitute for ATP as inhibitory ligands (1,2,4,6,8), replacement of ATP with ATP␥S prevented the activating effect of AMP on K ATP channels. The requirements for both Mg 2ϩ as a cofactor and AMP and ATP as apparent substrates are indicative of a possible involvement of adenylate kinase, which catalyzes ATP⅐AMP phosphotransfer (17,18). It is well established that when AMP and Mg 2ϩ are provided, the adenylate kinase-catalyzed phosphotransfer converts ATP into ADP (17,26). Such reaction could underlie the observed AMP-dependent opening of ATPinhibited K ATP channels since the ADP-liganded state of K ATP channels is associated with channel opening either by antagonism of ATP inhibition or by direct channel activation (2, 6, 9, 28 -31). In line with such assumption was also the finding that the biophysical properties of K ATP channel opening induced by the substrates of adenylate kinase, AMP plus ATP, were virtually indistinguishable from those induced by the product of catalysis, ADP. A role for adenylate kinase is further supported by the loss of AMP-induced opening of ATP-closed channels in patches treated with a selective inhibitor of adenylate kinase, Ap 5 A, which has no effect on other AMP-dependent catalytic processes including the AMP-kinase activity (32). Furthermore, in clamped cardiomyocytes, introduction of exogenous adenylate kinase through the patch pipette, along with ATP and AMP, induced an outward current that was inhibited by glyburide, a sulfonylurea blocker of K ATP channels. Taken together, these observations may fulfill the established criteria for a catalytic process to regulate a specific ion channel behavior (33) and support the notion that adenylate kinase activity can modulate the opening of cardiac K ATP channels.
Adenylate Kinase Activity Associated with Cardiac Sarcolemma-To account for the activation of K ATP channels by ADPgenerating substrates in excised membrane patches, endogenous adenylate kinase activity must be in close proximity to the channels. Although adenylate kinase is among the most ubiquitous and diversely distributed of cellular enzyme activities (17,26,34), its occurrence as an associated protein of cardiac sarcolemmal membranes has never been reported. Thus far, the presence of adenylate kinase within a plasma membrane has only been described in erythrocytes and synaptosomes where it is tightly associated with the lipid bilayer (35)(36)(37). Evidence that this enzyme activity may be a constituent of cardiac sarcolemmal membranes was obtained, herein, in isolated sarcolemmal preparations that possessed ATP⅐AMP phosphotransferase activity with properties attributable to an extra-mitochondrial adenylate kinase isoform most probably tightly bound to cardiac membranes (26, 34, 38 -42). Close association of ion channel proteins with modulatory enzymes has been reported in the case of protein kinases (33) and may represent a mechanism by which cells achieve highly localized regulation of ion-channel function by otherwise ubiquitous biochemical processes.
Adenylate Kinase-dependent Regulation of K ATP Channels-It has been recently suggested that the microenvironment surrounding sarcolemmal K ATP channels may play a role in regulating the ATP-dependent channel gating (43). Herein, in an environment of millimolar concentrations of ATP, addition of AMP increased the probability of channel opening. This is in line with the reported property of adenylate kinase to catalyze rapid phosphorylation of AMP within restricted subcellular compartments of intact muscle cells with no measurable change in the overall cytosolic levels of AMP, ADP, or ATP (18, 19). An adenylate kinase-catalyzed transformation of ATP to ADP within a locale closely associated with sarcolemmal K ATP channels would provide a unique endogenous mechanism for channel opening even when cytosolic concentrations of ATP are at millimolar levels.
Adenylate kinase activity is an integral part of cellular energy phosphoryl transfer networks, which couple flux-generating (ATPases) and flux-responding (glycolysis and oxidative phosphorylation) processes and permit sequential transfer of phosphoryls between nucleotides (18, 19). This is of importance since evidence for functional interactions between K ATP channels with both the flux-generating Na,K-ATPase and the fluxresponding glycolytic enzymes have been obtained within membrane patches (13, 44 -46). Based on the current understanding of adenylate kinase activity (18, 20), it can be speculated that the switch between the ATP-and ADP-liganded states of K ATP channels could be governed by the rate at which AMP is generated at a remote signaling site and transferred by adenylate kinase phosphotransfer to the channel. Thus, adenylate kinase could be a determinant of the composition of adenine nucleotide species at the channel site and/or govern the duration of the ATP-liganded state of the channel by the rate it catalytically transforms ATP to ADP. In this regard, ADP could be viewed as a second messenger transducing adenylate kinase activity into K ATP channel behavior. Indeed, in addition to biophysical and biochemical data implicating this enzyme in the regulation of channel opening (present study; see also Refs. 7 and 20), evidence, obtained by analysis of mutated components of the K ATP channel protein complex, identifies domains within the channel complex responsible for the ADP-dependent channel activation (9).
At present, the precise relationship between K ATP channels and adenylate kinase is not known. One of the subunits of the K ATP channel complex is encoded by the sulfonylurea receptor, a member of the ABC superfamily (14 -16). Similar to other ABC proteins, the sulfonylurea receptor contains two nucleotide-binding folds (14,16). There is, however, limited homology between the respective nucleotide-binding domains in the sulfonylurea proteins and the adenylate kinase isoforms (47), which suggests that adenylate kinase activity may not be intrinsic to the K ATP channel complex per se. Yet, possible association between adenylate kinase activity and ABC proteins has been suggested. Recently, mutations in the adenylate kinase gene have been linked to loss of osmoprotection conferred by the ProU transporter, a member of the ABC family (48). Also, selective ligands of adenylate kinase (32) apparently can target certain members of the ABC superfamily, including the K ATP channel and the cystic fibrosis conductance regulator (49,50).
Finally, up-and down-regulation of adenylate kinase activity correlate with the opening and closing of K ATP channels, respectively. In intact muscle cells, adenylate kinase-catalyzed phosphoryl transfer is increased when oxidative phosphorylation or creatine kinase activity is impaired (18, 51), conditions known to enhance K ATP channel opening (2,13). Conversely, suppression of adenylate kinase activity accompanies glucoseinduced insulin secretion in pancreatic ␤ cells (20), which is associated with closure of K ATP channels (1). Further understanding of the mechanisms regulating the activity of adenylate kinase may provide a previously unrecognized approach to regulate the K ATP channel behavior.