PIP2 and PIP as Determinants for ATP Inhibition of KATP Channels

See allHide authors and affiliations

Science  06 Nov 1998:
Vol. 282, Issue 5391, pp. 1141-1144
DOI: 10.1126/science.282.5391.1141


Adenosine triphosphate (ATP)–sensitive potassium (KATP) channels couple electrical activity to cellular metabolism through their inhibition by intracellular ATP. ATP inhibition of KATP channels varies among tissues and is affected by the metabolic and regulatory state of individual cells, suggesting involvement of endogenous factors. It is reported here that phosphatidylinositol-4,5-bisphosphate (PIP2) and phosphatidylinositol-4-phosphate (PIP) controlled ATP inhibition of cloned KATP channels (Kir6.2 and SUR1). These phospholipids acted on the Kir6.2 subunit and shifted ATP sensitivity by several orders of magnitude. Receptor-mediated activation of phospholipase C resulted in inhibition of KATP-mediated currents. These results represent a mechanism for control of excitability through phospholipids.

Modulation of KATPchannels by activation of metabotropic receptors and cell metabolism is an important pathway for regulation of cell excitability (1). A common feature of these regulatory effects is that inhibition of KATP channels by ATP can be antagonized and activation can be mimicked by so-called K channel openers (2). These drugs are known to activate KATPchannels even in the presence of millimolar concentrations of ATP (3) and involve the sulfonylurea receptor (SUR) to exert their effect (4, 5).

The effect of the K channel opener diazoxide on the current mediated by KATP channels in response to voltage steps in giant inside-out patches from Xenopus oocytes expressing Kir6.2 and SUR1 subunits is demonstrated in Fig. 1A (6). Inhibition of the KATP-mediated current by an initial application of 100 μM ATP was partly reversed by the addition of 100 μM diazoxide. After wash-out of both ATP and diazoxide, the current amplitude rapidly recovered to the level present before ATP application. This protocol was repeated four times after the patch had been intermittently exposed to the phospholipid PIP2. Besides its known effect of inhibiting run-down (7, 8), exposure to 5 μM PIP2 reduced the inhibitory effect of ATP and removed activation of channel activity by diazoxide (n = 3). Patch excision into Mg-ATP–free solution resulted in substantial run-down of KATP channel activity (Fig. 1B). This phenomenon is known for a variety of native and cloned Kirand KATP channels (7–10) and has recently been shown to be induced by wash-out of phospholipids such as PIP2 and PIP (8). Run-down of KATPchannels was accompanied by a marked increase in ATP sensitivity (n = 6). Immediately after patch excision, 10 μM ATP blocked about 30% of the current. This block increased to about 70% after several minutes, suggesting that the change in ATP sensitivity results from wash-out of phospholipids (Fig. 1B). The effect of PIP2 on ATP inhibition was further characterized in experiments with single channels (Fig. 1C) and metabolites of PIP2 (Fig. 1, D to F). Prolonged application of 10 μM PIP2 completely removed inhibition of KATPchannels by 1 mM ATP (Fig. 1D; n = 3). The current amplitude in the absence of ATP remained unchanged in this type of experiment, indicating that the PIP2-mediated decrease in ATP sensitivity can be separated from the effect on run-down. Moreover, PIP2 had no effect on the single-channel amplitude but increased the channels' open probability with respect to that observed after run-down of channel activity (Fig. 1C). Similar to PIP2, PIP was also able to reduce ATP inhibition, although on a slower time scale (Fig. 1E; n = 6), whereas phosphatidylinositol (PI) had no obvious effect on KATPchannels even at a 10 times higher concentration (Fig. 1F;n = 3). Furthermore, phospholipids were only active when applied to the cytoplasmic side of the membrane; addition of 50 μM PIP2 into the pipette had no effect on ATP sensitivity in inside-out patches (11).

Figure 1

Polyphospholipids act as K channel openers. (A) Responses to voltage steps from −80 to 20 mV in a giant inside-out patch with Kir6.2/SUR1 channels. Solutions were exchanged with a multibarrel application system. One hundred micromolar ATP (in K-Int1.0Mg) blocked almost all channel activity (left). Addition of 100 μM diazoxide activated channels in the presence of Mg-ATP. Wash-out of diazoxide and Mg-ATP recovered current to the level present before Mg-ATP application before run-down continued. Application of 5 μM PIP2 stopped current run-down and slightly increased current. Reapplication of Mg-ATP produced less current inhibition, and diazoxide was less effective in antagonizing ATP inhibition (middle panel). Three subsequent applications of PIP2 further decreased ATP sensitivity and finally abolished the diazoxide effect (right). Time scales are as indicated. (B) Repetitive dose-response measurements after patch excision with ATP concentrations as indicated. The first and last dose responses are shown on an expanded time scale. The first application of 10 μM ATP produced about 30% block, whereas the last application produced about 70% block. (C) Inside-out patch with four active channels at −80 mV. Initial application of 1 mM ATP (in K-Int0Mg) blocked channel activity, application of 100 μM PIP2 for 10 s increased channel activity, and reapplication of 1 mM ATP produced only partial channel inhibition. PIP2 had no obvious effect on the single-channel amplitude. (D toF) Application of 10 μM PIP2 (D), 10 μM PIP (E), or 100 μM PI (F) on inside-out patches with KATPchannels in the presence of 1 mM ATP (in K-Int0Mg). PIP2 and PIP recovered KATP-mediated currents, whereas PI failed.

For a more quantitative investigation of the phospholipid effect, PIP2 was applied to giant inside-out patches for increasing intervals, and the ATP sensitivity of channels was determined. The dose-response curves for ATP inhibition were gradually shifted toward higher concentrations, without changes in their steepness (Fig. 2, A and B). This result suggests that increasing concentrations of PIP2 gradually change ATP sensitivity of individual channels rather than render channels ATP insensitive in an all-or-none manner. To analyze the underlying mechanism, we determined the kinetics of ATP inhibition before and after exposure to 10 μM PIP2, using a piezo-controlled application system (12). The release of ATP inhibition exhibited a monoexponential time course (time constant: 69.1 ± 10.4 ms; n = 5) that did not change with PIP2 application (Fig. 2C). The steady-state ATP inhibition, however, was markedly reduced by the phospholipid (Fig. 2, B and C). As a consequence, the apparent blocking rate constants (k on) calculated from steady-state inhibition and release rates (k off) changed by as much as three orders of magnitude in response to the PIP2exposure, whereas K off remained unchanged (Fig. 2D). This change suggests that PIP2 decreases the probability of ATP binding to its receptor site without affecting the stability of the ATP-receptor interaction.

Figure 2

Effect of PIP2 on steady state and time course of ATP inhibition. (A) Dose-response measurements of ATP inhibition (ATP concentrations as indicated, voltage steps from 0 to −80 mV) in an inside-out patch before (top) and 1 s (middle) or 405 s (bottom) after PIP2 (100 μM) application. (B) Dose-response curves from experiments as in (A); data points are mean ± SD from three experiments. Continuous lines represent fit of a Hill equation:I/I max = 1/[1+([ATP]/IC50(ATP))n], whereI is the current in the presence of ATP,I max is the current amplitude in the absence of ATP, IC50(ATP) is the concentration for half-maximal inhibition, and n is the Hill coefficient. PIP2(cumulative application time as indicated) shifted ATP sensitivity of KATP channels without affecting the Hill coefficient. (C) On and off kinetics of ATP inhibition measured with a piezo-driven application system allowing solution exchange at an inside-out patch in less than 2 ms. Time course of block and block release by 100 μM ATP (in K-Int0Mg) as measured before and after application of 10 μM PIP2 (duration as indicated). Each trace is the average current from five subsequent ATP applications. Currents were normalized to the current preceding ATP application. Time course was slowed and steady state of ATP inhibition was decreased by PIP2, whereas block release remained unchanged. (Inset) Current traces normalized to their amplitude for better comparison of time courses for block and unblock. (D) Time dependence of PIP2 effect on ATP inhibition kinetics. Off-rates were calculated from the time constants obtained by monoexponential fits to the release of ATP inhibition (τoff) as k off = 1/τoff. Apparent on-rates were calculated fromk off, steady-state block (b), and ATP concentration ([ATP]) as k on = (bk off)/{(1 − b)[ATP]}; this procedure was used because fit of the on-reaction required more than a single exponential. The last four data points were obtained with 1 mM ATP to obtain sufficient current inhibition. ○,k on; •, k off. Data points are mean ± SD from two to four experiments.

This raised the question of which subunit of the KATPchannel complex is involved in action of PIP2 and PIP. Experiments as in Fig. 2B were performed with a mutant of Kir6.2 that has been deleted by the COOH-terminal 26 amino acids [Kir6.2(ΔC26)] and that forms ATP-sensitive channels in the absence of the SUR subunit (13). Kir6.2(ΔC26) channels displayed PIP2-dependent decrease of ATP inhibition in the absence and presence of SUR1 (Fig. 3, A and B). As described before (13), ATP sensitivity determined before PIP2 application was substantially lower in homomeric Kir6.2(ΔC26) channels than in Kir6.2(ΔC26)/SUR1 channels. However, longer applications of PIP2 were required to shift ATP inhibition (Fig. 3, A, B, and D). In addition, the PIP2 effect was less stable in Kir6.2(ΔC26) channels in the absence of SUR1 as judged from the faster wash-out of the phospholipid effect (11). These results indicate that PIP2 basically exerts its effect on ATP inhibition through interaction with the Kir6.2 subunit. Morever, SUR1 increases sensitivity of KATP channels for ATP and stabilizes binding of PIP2. Correlation between binding affinity for PIP2 and channel inhibition by ATP was further investigated in a Kir6.2 mutant channel, in which arginine (R) 176 was changed to alanine (A). This and the adjacent residue have recently been shown to reduce interaction between Kir6.2 and Kir1.1 with PIP2 (8, 10). Longer PIP2 applications were indeed necessary to induce a given shift in ATP sensitivity compared with Kir6.2 wild-type channels, although the ATP inhibition measured before PIP2 exposure was similar in both channels (Fig. 3, C and D).

Figure 3

Contribution of KATP subunits to ATP inhibition and PIP2 effect. (A toC) Dose-response curves from experiments as in Fig. 2A with Kir6.2(ΔC26) channels in the presence (A) and absence (B) of SUR1 or with Kir6.2(R176A)/SUR1 channels (C). Data points are mean ± SD from four experiments; cumulative application time of PIP2 is as indicated. (D) Dependence of IC50 for ATP inhibition as a function of PIP2 application time for the channels indicated. Data points are taken from the results shown in Fig. 2B and (A) to (C).

To ensure that PIP2 plays a role in the activity of KATP channels in the cellular environment, we injected PIP2 and PIP into oocytes expressing KATPchannels (14). Injection of about 50 nl of phospholipids at a concentration of 10 or 1 mM resulted in an increase of KATP-mediated currents (Fig. 4, A and B). PIP2 was more effective than PIP, and intracellular calcium had no effect on the KATP-mediated current (Fig. 4B). We further tested the effect of transient stimulation of a coexpressed metabotropic purino-receptor of the P2Y2 subtype, which is known to activate phospholipase C (PLC) (15, 16) and thus decrease the concentration of PIP2. Such receptor-activated stimulation of PLC reduces PIP2concentrations in cultured cells within minutes by about a factor of 2 (17). Application of 300 μM ATP to an oocyte coexpressing KATP channels and P2Y2 receptors resulted in a reversible reduction of the potassium current by 59 ± 3% (Fig. 4C; n = 2). To exclude effects of protein kinase C, we performed these experiments in the presence of staurosporine (100 nM). The initial increase in outward current observed in the first few seconds after ATP application is due to transient activation of calcium-dependent chloride channels endogenous to Xenopus oocytes through the PLC-IP3 pathway (15, 18). The P2Y2-mediated increase in ATP inhibition was confirmed by dose-response experiments performed with excised patches from oocytes that had been incubated with ATP (1 mM) before the measurements. As shown in Fig. 4D, P2Y2stimulation resulted in an increase in ATP inhibition by about 70%, which was independent of intracellular ATP at concentrations > 100 μM.

Figure 4

Modulation of KATPcurrents by PIP2 injection or stimulation of a coexpressed P2Y receptor. (A) Two electrode voltage-clamp measurements from oocytes expressing cloned KATP channels in response to voltage ramps from −120 to 50 mV. The external solution was intermittently changed from Ext90K to Ext2.5K to monitor both leakage and KATPcurrent. Injection of 50 nl of 10 mM PIP2 plus 10 mM BAPTA produced an initial transient increase of the leakage current (visible at 2.5 mM K+) and a delayed but continuous increase in KATP-mediated (I KATP) current from about 1 to about 8 μA over 4.5 hours. (B) Relative increase of KATP-mediated current measured 10 hours after injection of either 10 mM BAPTA, 1 mM PIP and 10 mM BAPTA, or 1 mM PIP2 and 10 mM BAPTA. Bars represent mean ± SD of three to five experiments. (Cand D) Stimulation of a coexpressed P2Y receptor decreased KATP-mediated currents by increasing ATP sensitivity. (C) Responses to voltage ramps from −120 mV to 50 mV in 2.5 s in Ext90K (traces in black) or Ext2.5K (traces in grey) to which 100 nM staurosporine was added. Application of ATP evoked a transient increase in outward current followed by a reversible decrease in KATP-mediated current. Measurements of KATP-mediated currents in whole oocytes were enabled by high expression levels of the channel protein. (D) Dose response for ATP inhibition of KATP channels measured in patches from oocytes before (control) and after stimulation of P2Y receptors (P2Y stimulation). Data points are mean ± SD from five experiments. (Inset) Relative inhibition (calculated from the ratio of the two dose responses) of KATP currents by P2Y receptor stimulation was independent of intracellular ATP for concentrations > 100 μM.

In conclusion, inhibition of KATP channels by intracellular ATP depends on the concentration of phospholipids such as PIP and PIP2 in the cell membrane. At low concentrations of phospholipids, KATP channels are blocked by micromolar concentrations of ATP, whereas prolonged application of PIP2 renders channels ATP insensitive. This observation might underlie the variability observed for ATP sensitivity of KATP channels in various tissues and studies (19, 20) and may contribute to the mechanism or mechanisms by which KATP channels can overcome the high physiological concentrations of intracellular ATP.

Mechanistically, PIP2 most likely binds to the Kir6.2 subunit, although SUR increases binding affinity of the phospholipids. It exerts its effect either by stabilizing a state in which the channel cannot interact with ATP (21) or by rendering the ATP-binding site inaccessible for ATP, for example, by competitive binding. However, the finding that PIP2 changes the apparent on-rate for ATP inhibition by several orders of magnitude without having any effect on the off-rate argues against a change in conformation of the ATP-binding site, which would also affect the off-rate.

The observation that a metabotropic receptor coupling to PLC can control activity of KATP channels might point toward the physiological role of the phospholipid effect presented here. This role is supported by the finding that ATP sensitivity of cardiac KATP channels is regulated through a G-protein–related pathway (22). This was observed in a membrane patch and did not involve a diffusible second messenger. Thus, phospholipid-mediated opening of otherwise ATP-blocked KATP channels might represent a new mechanism to control excitability in a wide variety of cells.

  • * These authors contributed equally to this work.

  • To whom correspondence should be addressed. E-mail: bernd.fakler{at}


View Abstract

Navigate This Article