Search

K+ currents through ATP-dependent channels were recorded from inside-out patches of ß-cell membrane as previously described (Rorsman and Trube 1985). Channels were opened by removing ATP from the intracellular side of the membrane. The open probability and/or the number of active channels declined spontaneously (“run-down”) when ATP was absent for periods longer than about 30 s. Channels subject to the run-down could be activated again after applying a blocking concentration (>0.1 mM) of ATP in presence of 1 mM MgCl2 for at least 2 min. ATP in absence of Mg and the ATP-analogues AMP-PNP, AMP-PCP and ATP?S were ineffective in reactivating the channels. This suggests that phosphorylation of the channels or associated proteins of hydrolysis of ATP may be necessary for keeping the channels available. In contrast to the differential effects on the run-down, ATP in presence and absence of Mg and the ATP analogues were similarly effective in blocking the channels at concentrations above 0.1 mM. Using an experimental protocol avoiding the run-down ghe dose-inhibition curve for ATP was found to reach 50% at 18 µM.

The influence of the antidiabetic sulphonylurea tolbutamide on K+ channels of mouse pancreatic ß-cells was investigated using different configurations of the patch clamp technique. The dominant channel in resting cells is a K+ channel with a single-channel conductance of 60 pS that is inhibited by intracellular ATP or, in intact cells, by stimulation with glucose. In isolated patches of ß-cells membrane, this channel was blocked by tolbutamide (0.1 mM) when applied to either the intracellular or extracellular side of the membrane. The dose-dependence of the tolbutamide-induced block was obtained from whole-cell experiments and revealed that 50% inhibition was attained at approximately 7 µM. In cell-attached patches low concentrations of glucose augmented the action of tolbutamide. Thus, the simultaneous presence of 5 mM glucose and 0.1 mM tolbutamide abolished channel activity and induced action potentials. These were not produced when either of these substances was added alone at these concentrations. The inhibitory action of tolbutamide or glucose on the K+ channel was counteracted by the hyperglycaemic sulphonamide diazoxide (0.4 mM). Tolbutamide (1 mM) did not affect Ca2+-dependent K+ channels. It is concluded that the hypo- and hyperglycaemic properties of tolbutamide and diazoxide reflect their ability to induce the closure or opening, respectively, of ATP-regulated K+ channels.

Inward currents through single K+ channels in isolated ventricular heart cells of the guinea‐pig were recorded using the patch‐clamp technique (Hamill, Marty, Neher, Sakmann & Sigworth, 1981). The voltage‐dependent gating properties of the channels were examined in the potential range between 0 and ‐120 mV with 145 mM‐KCl on the extracellular side of the membrane patch, i.e. with approximately symmetrical transmembrane K+ concentrations. When voltage pulses from 0 mV to negative test potentials were applied to patches containing several channels, more channels were open at the beginning of the pulses than in the steady state. Averages of many current responses showed inactivation of the mean current in response to the hyperpolarizing voltage pulses. The inactivation was stronger and faster at larger hyperpolarization. The lifetimes of the open and closed states of the channel and the probability of the open state p were estimated from records of the elementary currents at various constant potentials. As indicated by the inactivation of the averaged currents, the value of p was smaller at more negative potentials, approximately 0.15 at ‐50 mV and 0.02 at ‐110 mV. This caused a negative slope in the current‐voltage relation of the time‐averaged current at potentials more negative than ‐50 mV. The channel openings were grouped in complex bursts. At least three exponentials were needed to fit the frequency histogram of the lifetimes of all closed states (time constants at ‐50 mV: 1.1 ms, 16 ms and 3.2 s). The lifetimes of the individual openings were exponentially distributed (time constant: 70 ms). The kinetics of the channel were interpreted by two different models involving three states of a channel (closed‐closed‐open or closed‐open‐closed). The rate constants and their voltage dependence were estimated for both models. Both models describe the data equally well; the reason for this ambiguity is discussed. The channels are blocked by Cs+ or Ba2+. Cs+ (0.1 mM) caused frequent and short interruptions of the individual channel openings. Ba2+ (0.5 mM) also shortened the openings and in addition decreased the number of openings per burst. The results suggest that the inward‐rectifying current IK1 in heart ventricular cells is partially inactivated by hyperpolarization. The inactivation could account for part of the time‐dependent decrease in the whole‐cell current previously ascribed to depletion of K+.

Single ventricular cells were enzymatically isolated from adult guinea‐pig hearts (Isenberg & Klöckner, 1982). The patch‐clamp technique (Hamill, Marty, Neher, Sakmann & Sigworth, 1981) was used to examine the conductance properties of an inward‐rectifying K+ channel present in their sarcolemmal membrane. When the K+ concentration on the extracellular side of the patch was between 10.8 and 300 mM, inward current steps were observed at potentials more negative than the K+ equilibrium potential (EK). At more positive potentials no current steps were detectable, demonstrating the strong rectification of the channel. The zero‐current potential extrapolated from the voltage dependence of the inward currents depends on the external K4 concentration [K+]o in a fashion expected for a predominantly K+‐selective ion channel. It is shifted by 49 mV for a tenfold change in [K+]o. The conductance of the channel depends on the square root of [K+]o. In approximately symmetrical transmembrane K+ concentrations (145 mM‐external K+), the single‐channel conductance is 27 pS (at 19‐23 degrees C). In normal Tyrode solution (5.4 mM‐external K+) we calculate a single‐channel conductance of 3.6 pS. The size of inward current steps at a fixed negative membrane potential V increases with [K+]o. The relation between step size and [K+]o shows saturation. Assuming a Michaelis‐Menten scheme for binding of permeating K+ to the channel, an apparent binding constant of 210 mM is calculated for a membrane potential of ‐100 mV. For this potential the current at saturating [K+]o is estimated as 6.5 pA. The rectification of the single‐channel conductance at membrane potentials positive to EK occurs within 1.5 ms of stepping the membrane potential from a potential of high conductance to one of low conductance. In addition to the main conductance state, the channel can adopt several substates of conductance. The main state could be the result of the simultaneous opening of four conducting subunits, each of which has a conductance of about 7 pS in 145 mM‐external K+. The density of the inward‐rectifying K+ channels in the ventricular sarcolemma is 0‐10 channel/10 micron2 of surface membrane; the average of twenty‐eight patches was 1 channel/1.8 micron2. It is concluded that the inward‐rectifying K+ channels mediate the resting K+ conductance of ventricular heart muscle and the current termed IK1 in conventional voltage‐clamp experiments.

Pancreatic islets of NMRI mice were dissociated into single cells which were kept in tissue culture for 1‐3 days. The whole‐cell configuration of the patch‐clamp technique was used to study inward and delayed outward currents of beta‐cells under voltage‐clamp conditions at 20‐22 degrees C. Outward currents were suppressed by substituting the impermeant cation N‐methyl‐D‐glucamine for intracellular K+. The remaining inward current had a V‐shaped current‐voltage relation reaching a peak value of 39 +/‐ 4 pA (mean +/‐ S.E. of mean) around ‐15 mV. It was identified as a Ca2+ current, because the peak amplitude was increased 1.6 times by increasing external [Ca2+] ([Ca2+]o) from 2.6 mM to 10 mM and it was blocked by Co2+ (5 mM) or nifedipine (5 microM) but not by TTX (20 microM). The activation time constant of the inward current at ‐10 mV was 1.28 +/‐ 0.08 ms. The relation between the degree of activation (estimated from the size of the tail currents) and membrane potential V followed the sigmoidal function f = 1/(1 + exp [(Vh‐V)/k]) with half‐maximal activation potential, Vh = 4 +/‐ 1 mV and slope factor, k = 14 +/‐ 1 mV (for [Ca2+]o 10 mM). The inward current inactivated only weakly during depolarizing pulses of 0.1‐1 s duration. The delayed outward current (in experiments with 155 mM‐internal [K+] ([K+]i)) had a linear voltage dependence at potentials above ‐20 mV; its amplitude at ‐10 mV was 210 +/‐ 30 pA. Tail currents related to the activation of the outward current had K+‐dependent reversal potentials. The current was blocked by extracellularly applied tetraethylammonium (20 mM) and 4‐aminopyridine (2 mM). It was not affected by glibenclamide (3 microM), tolbutamide (0.2 mM) and alterations of intracellular [Ca2+] (1 nM‐1 microM). The activation time constant of the outward current at ‐10 mV was 21 +/‐ 3 ms. The voltage dependence of activation could be described by the sigmoidal function (see above) with Vh = 19 +/‐ 1 mV and k = 5.6 +/‐ 0.4 mV. The outward current inactivated during long (15 s) depolarizing pre‐pulses (time constant at ‐10 mV: 2.6 +/‐ 0.6 s). 50% inactivation occurred at Vh = ‐36 +/‐ 2 mV, k was ‐4.1 +/‐ 0.2 mV. Inward and outward currents during depolarizing voltage pulses in beta‐cells are similar to Ca2+ and delayed K+ currents in other cell types. These currents seem sufficient to generate the action potentials of the beta‐cell.

The resting conductance of culturedß-cells from murine pancreases was investigated using the whole-cell, cell-attached and isolated patch modes of the patch-clamp technique. Whole-cell experiments revealed a high input resistance of the cells (>20 GO per cell or>100 kO·cm2), if the medium dialysing the cell interior contained 3 mM ATP. The absence of ATP evoked a large additional K+ conductance. In cell-attached patches single K+-channels were observed in the absence of glucose. Adition of glucose (20 mM) to the bath suppressed the channel activity and initiated action potentials. Similar single-channel currents were recorded from isolated patches. In this case the channels were reversibly blocked by adding ATP (3 mM) to the solution at the intracellular side of the membrane. The conductances (51 pS and 56 pS for [K+]0=145 mM, T=21° C) and kinetics (at -70 mV: topen=2.2 ms and tclosed=0.38 ms and 0.33 ms) of the glucose- and ATP-dependent channels were found to be very similar. It is concluded that both channels are identical. The result suggests that glucose could depolarize theß-cell by increasing the cytoplasmic concentration of ATP.

K+ channels in the membrane of murine pancreatic ß-cells were studied using the patch-clamp technique. The delayed outward current was activated in whole-cell experiments by depolarizing voltage pulses to potentials between -30 mV and 0 mV. Forskolin blocked the current rapidly (<5 s) and reversibly with 50% inhibition at 13 µM. The inhibition did not depend on a stimulation of the adenylate cyclase since it occurred even in presence of 1 mM cAMP in the pipette solution which replaced the cytoplasm. Membrane permeant cAMP analogues and phosphodiesterase inhibitors did not influence the delayed outward current. In experiments on outside-out patches forskolin (100 µM) shortened the openings of a channel of about 10 pS conductance at 0 mV and a time course of activation and inactivation similar to the whole-cell current. Another smaller, slowly activating channel and the Ca2+- and ATP-dependent K+ channels were influenced only weakly or not at all. It is therefore concluded that the 10-pS channel generates most of the delayed outward K+ current in murine pancreatic ß-cells. The Ca2+-independent part of the delayed outward current in bovine adrenal chromaffin cells was also blocked by forskolin (100 µM).

The influence of the hypoglycemic drugs tolbutamide, meglitinide, glipizide and glibenclamide on ATP-dependent K+ currents of mouse pancreatic B-cells was studied using the whole-cell configuration of the patch-clamp technique. In the absence of albumin, tolbutamide blocked the currents half maximally at 4.1 µmol/l. In the presence of 2 mg/ml albumin half maximal inhibition of the currents was observed at 2.1 µmol/l meglitinide, 6.4 nmol/l glipizide and 4.0 nmol/1 glibenclamide. The hyperglycemic sulfonamide diazoxide opened ATP-dependent K+channels. Half maximally effective concentrations of diazoxide were 20 µmol/l with 0.3 mmol/1 ATPand102 µmol/l with 1 mmol/1 ATP in the recording pipette. Thus, the action of diazoxide was dependent on the presence of ATP in the recording pipette. The free concentrations of the drugs which influenced ATP-dependent K+ currents were comparable with the free plasma concentrations in humans and the free concentrations which affected insulin secretion in vitro. The results support the view that the target for the actions of sulfonylureas and of diazoxide is the ATP-dependent K+ channel of the pancreatic B-cell or a structure closely related to this channel.

Two types of exogenous K channel were identified in Xenopus laevisoocytes injected with poly(A +) mRNA from the insulin-secreting cell line HIT T15. One of these was the ATP-regulated K channel (G channel) as evidenced by its conductance and inhibition by tolbutamide. The other resembled the Ca-activated K channel from β-cells.

The effects of intracellular purine nucleotides on tolbutamide-induced block of ATP-dependent K +channels from mouse pancreatic B-cells were studied using the patch-clamp technique. When applied to the inside of excised patches tolbutamide alone blocked channel activity half-maximally at 55μM and the concentration-response curve for the inhibition of K +channels by tolbutamide was flat. ADP (1 mM), but not other nucleotides (AMP, GTP or GDP) increased the steepness of the concentration-response curve and decreased the half-maximally effective tolbutamide concentration to 4.2 μM. It is suggested that the ATP-dependent K +channel or a closely related structure contains a receptor which is accessible for cytosolic ADP and controls the sensitivity to tolbutamide.

Two types of exogenous K channel were identified in Xenopus laevisoocytes injected with poly(A+) mRNA from the insulin‐secreting cell line HIT T15. One of these was the ATP‐regulated K channel (G channel) as evidenced by its conductance and inhibition by tolbutamide. The other resembled the Ca‐activated K channel from β‐cells.

The effects of intracellular purine nucleotides on tolbutamide‐induced block of ATP‐dependent K+channels from mouse pancreatic B‐cells were studied using the patch‐clamp technique. When applied to the inside of excised patches tolbutamide alone blocked channel activity half‐maximally at 55μM and the concentration‐response curve for the inhibition of K+channels by tolbutamide was flat. ADP (1 mM), but not other nucleotides (AMP, GTP or GDP) increased the steepness of the concentration‐response curve and decreased the half‐maximally effective tolbutamide concentration to 4.2 μM. It is suggested that the ATP‐dependent K+channel or a closely related structure contains a receptor which is accessible for cytosolic ADP and controls the sensitivity to tolbutamide.