Journal of Physiology

Pfl/igers Arch (1990) 415:387 - 394

9 Springer-Verlag1990

Pinacidil activates the ATP-sensitive K + channel in inside-out and cell-attached patch membranes of guinea-pig ventricular myocytes Zheng Fan, Keiko Nakayama, and Masayasu Hiraoka Department of Cardiovascular Diseases, Medical Research Institute, Tokyo Medical and Dental University, 1-5-45 Yushima, Bunkyo-ku, Tokyo-ll 3, Japan Received March 30/Received after revision August 16/Accepted September 6, 1989

Abstract. Patch-clamp techniques were used to study the effects of pinacidil on the adenosine-5'-triphosphate (ATP)sensitive K + channel current in guinea-pig ventricular myocytes. In inside-out patches, the ATP-sensitive K + channel current could be recorded at an internal ATP concentration of 0.5 mM or tess and almost complete inhibition was achieved by raising the concentration to 2 raM. Application of pinacidil ( 1 0 - 3 0 ~M) in the presence of 2 mM ATP restored the current, whereas 5 mM ATP antagonized the effect of pinacidil. The conductance of the channel at symmetrical K + concentrations of 140 mM was 75 pS with a slight inward rectification at voltages positive to + 40 mV. There was no significant change in the conductance after application of pinacidil. In 0.5 mM ATP, at - 8 0 mV, both the distributions of the open time and the life-time of bursts could be fitted by a single exponential. An increase in ATP concentration decreased the mean life-time of bursts, whereas pinacidil increased it with little increase in the mean open time. Closed time distributions of the channel were fitted by at least two exponentials, with a fast and a slow time constant. An increase in ATP concentration markedly increased the slow time constant associated with a decrease in the number of bursts, whereas the effect of pinacidil was opposite to that of increased ATP. These results indicate that pinacidil increases the open-state probability of the ATP-sensitive K § channel. In cell-attached patches, application of pinacidil (100-200 gM) to the extracellular solution reversibly induced the channel activity, which showed similar properties to those of the ATP-sensitive K § channel recorded in cell-free patches. Key words: Patch-clamp technique brane - K + Current - ATP

Heart cell mem-

Introduction The adenosine-5~-triphosphate (ATP)-sensitive K + channel has been described in the membranes of cardiac muscle cells (Noma 1983; Trube and Hescheler 1984; Kakei et al. 1985), skeletal muscle cells (Spruce et al. 1987) and pancreatic/?cells (Cook and Hales 1984; Findlay et al. 1985; Rorsman and Trube 1985). This channel is characterized by a pronounced inhibition of channel opening when the intracelluOffprint requests to: M. Hiraoka

lar ATP concentration is increased to millimolar levels. The channel also has a high K § selectivity with a unitary conductance of 50 - 80 pS and an inward rectification at high positive voltages. A number of studies have focused on the inhibitory mechanisms of the ATP-sensitive K § channel, using nucleotides, ions and various pharmacological agents (for review, see Ashcroft 1988). In contrast, little work has been done on promoting agents and their mechanism of channel opening. Recently, it was found that diazoxide, a sulfonamide, can activate the ATP-sensitive K + channel in pancreatic /?-cells (Trube et al. 1986; Dunne et al. 1987). More recently, two potent K § channel openers, cromakalim (BRL 34915) and nicorandil, which produce relaxation in vascular smooth muscle and shorten the action potential duration in cardiac muscle, have been found to activate the ATP-sensitive K + channel in the heart (Escande et al. 1988; Sanguinetti et al. 1988; Hiraoka and Fan 1989). Among the family of so-called K + channel openers (Cook 1988), there is another type of drug named pinacidil, a potent vasodilator (Arrigoni-Martelli et al. 1980), which was suggested to activate the ATP-sensitive K § channel in a preliminary study (Escande et al. 1989). However, the precise mechanism for pinacidil to activate the ATP-sensitive K + channel has not been clarified. In the present study, the activation mechanism of the ATP-sensitive K § channel by pinacidil was explored using cell-free and cell-attached modes of the patch-clamp technique (Hamill et al. 1981).

Materials and methods Myocyte preparation. Single ventricular myocytes were isolated from guinea-pig hearts by enzymatic dissociation, as described previously (Hirano and Hiraoka 1988). Briefly, the animals were anaesthetized with pentobarbital sodium ( 4 0 - 5 0 n g 9 kg-1) after heparin administration (300 IU 9 kg- 1). The chest was opened under artificial respiration and the aorta was cannulated before removal of the heart. Using a Langendorff apparatus, the excised heart was perfused with 0.04% collagenase (Type 1, Sigma, St. Louis, MO) dissolved in low-Ca 2 § Tyrode's solution (30 gM). The heart was then stored in kraftbrfihe (KB) solution (Isenberg and K16ckner 1982) at room temperature for 60 rain. Single cells were obtained by gentle agitation of small pieces of ventricular tissue in a beaker containing KB solution. The preparations were then transferred to a recording chamber placed on the stage of an inverted phase-contrast microscope

388 (Diaphot TMD; Nikon, Tokyo). A single isolated cell having a smooth surface with clear striations was then selected for electrical measurement, as described below.

A

Recording methods. The patch-clamp technique (Hamill et al. 1981) was applied to record the current through single channels using a patch-clamp amplifier (Axopatch-lC, Axon Instruments, Burlingame). The current signals were stored on a video cassette recorder (HR-S 7000, Victor, Tokyo) through a PCM converter system (RP-882, NF Instruments, Yokohama) at a conversion rate of 40 kHz. The recorder signals were filtered off-line through a low-pass filter (48 dB/oct., FV-665, NF Instruments, Yokohama) and digitized at 30 kHz on to the disc of a computer (IBMPC/AT) using an analog-to-digital converter (CED 1401, Cambridge Elect. Design, Cambridge, UK).

Data analysis. A "50% threshold" criterion was used to detect events with the help of manual confirmation. The unitary current amplitude of the ATP-sensitive K + channel current was measured by two different histogram methods. At a membrane potential positive to 0 mV, a histogram was constructed from all the data points for recordings longer than 120 s, including the baseline and open level (Spruce et al. 1987). Since both sides of the membrane faced symmetrical K + concentrations, the inward rectifier channel current (Ik0 could be induced together with the ATP-sensitive K + channel current at a potential negative to 0 mV (Trube and Hescheler ~984; Kakei et al. 1985). In order to discriminate the distribution of unitary amplitude of the ATP-sensitive K + current from that of Ik~, a histogram was formed using averaged amplitudes of openings. Openings shorter than 0.3 ms were not included. Thereafter, the histogram was expressed by a sum of several gaussian distributions with mean and variance (Bhattacharya 1967). The difference of the means of two adjacent gaussian peaks was taken as a measure of the unitary current amplitude. Open and closed times were measured from records where only a single channel was activated. Each apparent distribution histogram of open or closed time was formed from a length of continuous recordings lasting more than 120 s. The distribution of open time was obtained by measuring the life-time of open events and of bursts separately. For measuring the life-time of open events, the cut-off frequency of the filter (fo) was set at 10 kHz to increase the resolution, whereas for measuring

ATP 2 mM

j

,4 C ~

Solutions. For inside-out patch-clamp (cell-free) experiments, the bathing solution (intracellular solution) contained (raM): 140 KC1, 5 HEPES, 1 EGTA, 5.5 glucose, 0.5 KzATP, and the pH was adjusted to 7.3 - 7 . 4 with KOH. When the KzATP concentration was changed, the final K + concentration was maintained constant at 141 m M by varying the KC1. The pipette solution (extracellular medium) contained (mM): 140 KC1, 0.53 CaCla, 5 HEPES, 5.5 glucose, and the pH was adjusted to 7 . 3 - 7 . 4 with KOH. In cell-attached patch-clamp experiments, the bathing Tyrode's solution had the following composition (mM): 144 NaC1, 0.33 NaHzPO4, 4.0 KC1, 1.8 CaCI2, 0.53 MgClz, 5.5 glucose, and the pH was adjusted to 7 . 3 - 7 . 4 by addition of NaOH. The pipette solution had the same composition as that used in cell-free experiments. The bathing solution could be replaced completely within 30 s when switching from one solution to another. Experiments were done at room temperature.

ATP 0.5 mM

..................

B

.....I...................................... ..l................ L . . . L . . . . . . . . . ~ r . I , ~ . . ~ a . ~ I . ~ , ~ I ~ I W ~ , # ,

ATP 2 mM

t

Pinacidil 3 0 .uM

.......

1"~1/

]'

9 fk

t

~

[ 5 pA

';: ~} !):i !ill! :

~!

,

Pinacidil 30 ATP 2 mM ~,~,q~,,i

~lJ~ i~*~N~ j ~ , ! ~ l

]

ATP 5 mM

I ~ .~ i~r ~

Fig. 1 A - C. Effects of ATP and pinacidil on ATP-sensitive singlechannel current recorded from an inside-out patch membrane. Membrane potential was held at + 60 inV. At the top of each record, the ATP concentration and application time of pinacidil in the internal solution are shown. A Prominent openings of the singlechannel currents up to three levels at 0.5 mM ATP were promptly inhibited by raising the ATP concentration to 2 mM. B Application of 30 gM pinacidil at 2 mM ATP produced activation of the channel current with the same unitary amplitude as that at 0.5 mM ATP. C Increasing the internal ATP concentration from 2 mM to 5 mM in the presence of pinacidil quickly abolished the channel activity. Upward direction indicates the outward current in this and the following figures. C represents the level of the closed state

the life-time of bursts, the fc was set at 0.1 kHz to decrease the errors in detection caused by flickering (see Sakmann and Trube 1984b). The distribution of closed time was obtained in a similar way, but measurements of shut time within bursts were made instead of the life-time of open events (any shut time longer than 20 ms being discarded), similarly the closed time between bursts instead of the lifetime of bursts (any closed time longer than 600 ms being discarded) was measured. A simplex method of least-squares analysis (Nelder and Mead 1965) was applied to fit a probability density function to open or closed times with a form of single or double exponentials. The number of openings was also analysed to estimate the kinetics of the channel activity. Generally, the openings of an ATP-sensitive K § channel are usually grouped in clusters of bursts, and each burst is often divided into several short openings by some brief closures. In order to analyse the number of bursts per cluster (see Discussion), the f~ was also set at 0.1 kHz to eliminate these very brief closures within bursts. A critical closure time of 600 ms was chosen in order to separate one cluster from another. Results

Identification of the pinacidil-activated single channel current The single channel current which was sensitive to the intracellular A T P concentration is shown in Fig. 1, being inhibited by 2 m M ATP. The single channel current activated by pinacidil had the same amplitude as that activated by 0.5 mM ATP. A further increase in ATP concentration to 5 mM abolished the channel activity induced by pinacidil. These changes were reversible (not shown). Results similar to those shown in Fig. I were confirmed in eight preparations. The conductance and permeability of the pinacidil-

389 strong inward rectification at positive voltages, so that the current amplitude was too small to be detected at voltages positive to 0 mV. Because of the magnitude of its conductance and its inward rectification, this channel was considered to be the inwardly rectifying K § channel (Kameyama et al. 1983; Sakmann and Trube 1984a). A histogram analysis of the larger amplitude current activated under both conditions is shown in Fig. 2 B and from this, the unitary single-channel current amplitude was obtained. The current-voltage relationships (I-V) for the channel are shown in Fig. 2C. The current had a reversal potential of 0 mV and the I-V at negative voltages was linear with a slope conductance of 74.5 + 5.7 pS (mean _+ SD; n = 6) and 73.3 _+ 5.9 pS, in the presence and absence of pinacidil respectively. The permeability for K + was calculated according to the Goldman-Hodgkin-Katz constant-field equation, giving mean values of 1.30 x 10-13 cm 3 "s 1 in 0.5 m M ATP and 1.28 x 10 ~3 cm 3 . s- 1 in 2 m M ATP plus pinacidil. The linearity of the I-V was lost at voltages positive to + 40 mV. Therefore the current showed an inward rectification under both conditions. These results strongly support the idea that pinacidil activates the ATP-sensitive K + channel.

A Control, ATP 0.5 mM

Pinacidil 30 uM, ATP 2 mM

L~'....I. L _ ~

~.

1 .....................

c~

40 mV

0 mv

-40 rnv

-80 mV

~ ~ ~-c/

" " ~ , ; r i l .................. i:

"

~s

B

-

C (a)

Control, ATP 0.5 rnM

150

Vm: -80 mv

2000

f

I (pA)

5.O

I000

-lO

(b)

-5

o

5

z

_'0

oJ

_20 / -2.5

PinacidW 30/aM, ATP 2 m M

v~ (mV)

-5.0

Vm= 80 mV 750

- - - O - - : Control. ATP 0.5 --Q--: 10

5

0

5

OJ

0

5

10

15

PiI~acid~ 30 .uM. ATP 2 mM

CURRENT(pA)

F i g . 2 A - C . Effects of pinacidil on activation of ATP-sensitive channel currents recorded from an inside-out patch membrane. A Voltage-dependent activation. Membrane potential [Vm (mV)] is indicated at the left of each trace. The leftpanel shows records taken during the control experiment with 0.5 mM ATP in the internal solution, and the right panel presents those with a solution containing 30 gM pinacidil and 2 mM ATP. The current direction is outward at positive voltages and inward at negative voltages. B n Amplitude histograms (top panel) of the single-channel current at Vm - 80 mV (Ic[t panel) and Vm = + 80 mV (right panel) obtained using the control solution containing 0.5 mM ATP. Bb Amplitude histograms (bottom panel) of the single-channel current at Vm = 80 mV (left panel) and V,, = + 80 mV (right panel) obtained using solution containing 30 gM pinacidil and 2 rnM ATP. Note that the current ampiitude of the single channel is not changed by pinacidil. C Current-voltage relationship (I-V) and the effect of pinacidil. Pinacidil does not change the conductance of this channel current

induced channels were compared with those of the ATPsensitive K § channel in the absence ofpinacidil. The voltagedependence of membrane current in 0.5 m M ATP solution and 2 m M ATP plus 30 gM pinacidil solution is shown in Fig. 2 A. When the membrane potential was held at a positive voltage, there was only one kind of channel opening with a long-lasting period under both conditions and the current reversed at around 0 mV. At negative voltages, two kinds of channel opened in both solutions. One kind of channel passed a current of larger amplitude and showed rapid opening and closing behaviour, whereas the other passed a current of smaller amplitude and showed longlasting opening. The former was sensitive to ATP concentration, whereas the latter was insensitive (not shown). The latter had a single channel conductance of 30 pS with a

Modulation of channel kinetics by pinacidil The open and closed time distributions were analysed to estimate the actions of pinacidil on the kinetic properties of the ATP-sensitive K + channel current. The experimental protocol was in principle the same as that shown in Fig. 1 except that the patch membrane potential was held at - 80 mV instead of + 60 inV. This voltage was chosen because it made fiickerings within a burst easily visible (see Fig. 2 A), while their durations were long enough to be determined and the signal/noise ratio was high. The histogram of the open time, which was analysed from the current records filtered at an fo of 10 kHz, showed a single-exponential distribution function in control with 0.5 m M ATP, 2 m M ATP and 2 m M ATP plus 30 ~tM pinacidil. The mean open time did not differ greatly under these three conditions. The histogram of the life-time of bursts was obtained from the records filtered at an fc of 0.1 kHz (see Sakmann and Trube 1984b). The value was best fitted by a single-exponential function under the three conditions. Its time constant, designated as Tb, was shortened by raising the ATP concentration and was prolonged markedly by pinacidil (Fig. 3 B). The analysis of the closed times is illustrated in Fig. 4. The histogram of the shut time within bursts was best fitted by a single-exponential function, when the analysis was done using closed times shorter than 20 ms, filtered at anf~ of 10 kHz. The time constant obtained by this method was designated as zo.f. When the closed times having durations up to 600 ms were analysed from the records filtered at an fo of 0.1 kHz, the histogram was best fitted by a double-exponential function. The time constant of the faster component was called to.f, and that of the slower component Tc.~. The same record of the closed times filtered at 10 kHz was also analysed and the histogram showed a best fit by a double-exponential function (not shown). Therefore, rc.c was most probably caused by the distortion of the fast component due to heavy filtering. Neither the mean closed time within bursts nor zc.f was changed markedly by variation of ATP concentration or application of pinacidil. With regard to the slower component, both zc.~ and the proportional area under the total distribution were largely in-

390 A

(a)

(b)

rc = 10 kHz

C--;]

L~__ I ;L--J

(a)

fo = 0.1 kHz

]L_~IIU--~JU~

Control, ATP 0.5 mM 40

700 Tc,f= 0.27 ms

20

350 50

5 ms

B

(a)

ms

0 ]i~'~

0

Control, ATP 0.5 mM 4O

80

(b) i

0

10

20

0

,,z, ,,>, 0

~

0

0

300

600

0

300

600

0

300

600

ATP 2 mM

= 52 ms

2O

40

215

1 600

300

10-

250 l"c,f= 0.33 ms

,,=,

125 4 II

5

m (b)

P-

z,, ,,-,'

ATP 2 mM 40

2O

20

10

0

(]

z

= 31 ms

01 ...... ~

o

(C)

i.s

;

Pinacidi130 .uM, ATP 2 mM

800 l [

(C)

0

10

20

0

300

] 600 400 ~ ~_

Pinacidi130 .uM, ATP 2 mM

80

40

40

20

o

0

0

10 OPEN TIME (ms)

20

C

0

40 % f = 0.28 ms '

2i5

20

g

CLOSED TIME (ms)

0

300 600 BURST DURATION (ms)

Fig. 3A, B. Effects of ATP and pinacidil on the mean open time of the ATP-sensitive single-channel current recorded from an insideout patch membrane. A a, b Filtered current records with high (a left panel; 10 kHz) and low (b right panel; 0.1 kHz) cut-off frequencies (flo).In both panels, the top trace is an idealization of the original current record (lower trace). B a - e Histograms of open time, analysed with highfc (left column) and histogram of life-time of bursts analysed with lowf~ (right column), with (a) 0.5 mM ATP and (b) 2 mM ATP in the internal solution and (e) 30 gM pinacidil and 2 mM ATP in the internal solution. The time constant (%) of the open time and that of bursts ('Cb) obtained under three different conditions is indicated

creased by an increase in ATP concentration, whereas they were decreased by pinacidil. A t this point, it is necessary to state that in 2 m M ATP without pinacidil, there was such a long-lasting closed state that the d a t a pertaining to this state were insufficient for statistical analysis. Therefore, the analysis of bursts had to be made from the long period of the records taken after switching the solution to that containing 2 m M ATP. Pinacidil apparently reduced the probability of the channel entering this long-resting state. In order to give a quantitative estimation o f the probability of the channel entering this closed state, the number o f bursts per cluster was counted. In 0.5 m M ATP, the mean number o f bursts per cluster was 30.3 at - 8 0 mV, and in 2 m M ATP the number was reduced to 3.2. A d d i t i o n o f 20 btM pinacidil increased the value to 43.0.

Pinacidil activation of ATP-sensitive K + channels in cell-attached patch membranes M e m b r a n e current was recorded in a cell-attached patch using a pipette filled with 140 m M K + solution. U n d e r these

Fig. 4 a - c . Effects of ATP and pinacidil on mean closed times of the ATP-sensitive channel current recorded from an inside-out patch membrane. The left panel shows analysis of closed times shorter than 20 ms at anfc of 10 kHz and the rightpanel shows the analysis of those up to 600 ms at anfo of 0.1 kHz. Raising the ATP concentration and the presence of pinacidil does not alter the fast component time constant (re.f) of distribution of short (left panel) or long (right panel) closed times. The slow component time constant (zo.s) of distribution of long closed times (right panel) is prolonged by 2 mM ATP, and shortened by pinacidil. Data obtained from the same patch as Fig. 3 and showing only a single level of the channel activity throughout the three conditions. It was not a usual finding that the ATP-sensitive K § channel was active in 2 mM ATP. However, we could see sporadic openings of the channel activity in the 2 mM ATP solution in about 20% of the total of 20 patches. The analysis of the open and the closed time distributions in the 2 mM ATP solution was made using long recordings from the one patch of such examples

conditions, only one type of current with a 30-pS channel conductance and a strong inward rectification at positive voltages could be recorded (il in Fig. 5A). Because of the magnitude of its conductance and its inward rectification, the channel was considered to be the inward rectifier K + channel. W h e n 0.2 m M pinacidil was a d d e d to the external perfusate, another kind of channel showing a conductance larger than that of the inward rectifier K § channel could sometimes be activated (i2 in Fig. 5 B). The incidence o f this channel activation was a b o u t 60% (four out of seven cells). The channel activity was lost again after washout o f pinacidil (Fig. 5 C). The conductance and kir/etic properties o f the pinacidil-activated channel are presented in Fig. 6. The current had a reversal potential at a r o u n d 0 mV (assuming that the resting m e m b r a n e potential of the cell was - 8 0 mV), which was close to the equilibrium potential of K § across the patch membrane. The conductance o f the channel was 69.8 _+ 3.6 pS (n = 4) calculated from the slope in regions

391 A

Control

A

(b)

(a)

l(pA)

100

; I ~ ' ; ' % , " ' " - F ; ~ , ; ; ............:...............: ; ~ ..................;I.......................:::::::::::::::::::::::::::::::::::............ ...........C

I-

50

500 ms

0

:........:_.:

. . . . . . . . . . . . . . . . .

. . . . . .

. . . . . .

B

I

40

/~

Ill Pinacidil 0.2 mM -10

-5 0 CURRENT (pA)

(b)

(a) f~=10mz

c ~J]~l_~

J~

5

JJU~]

O~ O

,,Oj ~ I -4~:J

-80

$ B

2.5"

(mY) RP-Vh~

.T,

-2.5 - -5.0

fo = 0.1 k H z

q~l~Lllj2~,q~U

r

l spA

Washout

C

5

Fig. 5 A - C . Activation of ATP-sensitive channel current by pinacidil recorded from a cell-attached patch membrane. Recordings were made with both the pipette and external solution containing 140 mM K +. A Control. One type of channel activity with a small unitary amplitude (i~) is predominantly activated. B Application of 0.2 mM pinacidil to the external solution activates another type of channel of large unitary amplitude with rapid flickering (i2) in addition to activation of the small-conductance channel. The channel activation was achieved within t min of the pinacidil application. C Wash-out of pinacidil quickly abolished the large-conductance channel activity within 1 min but the small-conductance channel activity persisted, i~ and i2 represent the amplitudes of the small- and large-conductance channel currents, respectively (Vm = - 80 mV)

C

Discussion

The present study demonstrated that the vasodilatory agent pinacidil activated a single channel current in cell-free and cell-attached patch membranes, which was sensitive to intracellular ATP concentration and was inhibited by raising its concentration above the millimolar level. The m o d e o f activation by pinacidil was p r o l o n g a t i o n o f the mean lifetime o f bursts, shortening o f the closed time between bursts and increasing burst numbers without any change in its

ms

50 ms

5O

40 I~

"to= 2.39 ms 20

0 ,,m,

10 OPEN TIME (ms)

20

0

10

rm~.,,=_,_ _ , 2.5 CLOSED TIME (ms)

300

600

BURST DURATION(ms)

20 ~H

04-u 0

o f negative voltage, and the current revealed an inward rectification at positive voltages (Fig. 6A). The calculated permeability was 1.22 x 10-~3 cm 3 . s-~ a value very close to that o f the ATP-sensitive K + channel current obtained from cell-free patches. Analysis o f the kinetic properties o f the channel revealed that the distributions o f b o t h the lifetime o f open events and the shut time within bursts were well fitted by single-exponential functions. The life-time o f the bursts also followed a single-exponential distribution. The distribution of closed time between bursts was best fitted by a double-exponential function (Fig. 6B, C), the time constants o f which were quite c o m p a r a b l e to those o f the ATP-sensitive K § channels shown in Figs. 3 and 4. Therefore, the pinacidil-activated channel current in intact cells was identified as the ATP-sensitive K + current. Similar types of experiments were done in solutions containing either 20 or 100 g M pinacidil. A t the former concentration, the opening o f the ATP-sensitive K + channel was seldom observed (two out of six patches), but at the latter the activity of the channel was easily recorded (three occasions in two out of four patches).

I

80

l'

0

r c f = 31.3 ms

r 0,s=160 ms

300

600

CLOSED TIME (ms)

6 A - C . Kinetic analysis of ATP-sensitive single-channel K + current activated by 200 gM pinacidil recorded from a cell-attached patch membrane. A a The amplitude histogram (Vm, - 8 0 mV) and A b the current-voltage (I-V) relationship of a single-channel current. Ba, b Current records filtered at different cut-off frequencies (fc), with the top trace an idealization of the original current record (lower trace). Note that the time scales differ in a and b. C Histograms to show the distribution of open time, burst duration and short (< 20 ms) and long (< 600 ms) closed times. Distributions of open time, burst duration and short closed time are all fitted by a singleexponential the time constants of which are To, % and zc.f respectively. The distribution of long closed-times is fitted by a doubleexponential, the time constants of which are designated r'c.f and zc.~ Fig.

conductance. It is conceivable, therefore, that pinacidil increases the open-state probability o f the ATP-sensitive K + channel. The properties o f the ATP-sensitive K + channel current have been described in cardiac and skeletal muscles, and also in pancreatic /?-cells (see Ashcroft 1988). The single channel current recorded in the present study during the control experiment using a low (0.5 m M ) ATP concentration had properties similar to those described in previous reports, i.e. a single channel conductance o f 5 0 - 80 pS with a symmetrical K + concentration o f about 1 4 0 - 1 5 0 m M on both sides of the membrane, an inward rectification at high positive voltages, and a reversal potential at the presumed K § equilibrium potential, in addition to its inhibition by raising the intracellular ATP concentration. There are still several unclarified problems regarding the activation mechanism o f the channel and its kinetics. Kakei et al. (1985) concluded that the gating kinetics of the ATP-sensitive K + channel

392 were almost independent of the membrane potential. We observed fast flickering at negative voltages, whereas no such activity was recognized at positive voltages (Fig. 2). This suggests that the channel kinetics may have a weak voltage dependence, or that at positive potentials they become too fast to be resolved by our recording system. Trube and Hescheler (1984) and Spruce et al. (1987) also described flickering of the channel current between the open state and short closed state at negative voltages, indicating voltagedependent kinetics. Zilberter et al. (1988) agreed with these observations and showed that the channel mean open and closed times during bursts depended primarily on the electromotive force for potassium ions. They concluded that these fast gating properties of the ATP-sensitive K + channels depend on the ion flux parameters. Our present result showed that the flickering was independent of ATP and pinacidil. Despite these observations, it seems probable that the voltage dependence of the channel kinetics contributes little to the amplitude of the averaged channel current, since the I-V curve of the latter was quite similar to that of the unitary current (Kakei et al. 1985). There is general agreement that at least one open state and two closed states are required to account for the kinetics of the ATP-sensitive K § channel (Ashcroft 1988). According to the fits applied to the distributions of open and closed times analysed in the present study, the following kinetic scheme could be proposed for the ATP-sensitive K § channel: K3... K2 K1 C3~ C2 ~ O z - ~ 7 C1 K-3 K 2 K-i where the rate constant K~ was considered to be much larger than that of K - 2 . For the present analysis, we combined the use of differentfc values and critical times for discriminating the different states. Assuming that the channel stays in the closed state C1 for a shorter time than in any other state, the open-time distribution analysed using an fo value of 10 kHz is therefore considered to reflect mainly the transition from the open state (O) to C1, because of the higher probability that it will enter this state rather than C2 when the channel leaves state O. When thefo was set at 0.1 kHz, the numbers of brief closures were filtered out so that the transition from state O to state C2 was picked up by revealing prolonged open events. However, this would not be accounted for by the presence of two open states, since the total open time distribution was best fitted by a singleexponential function. For the closed time analysis, we employed a different fc and divided the shut time events of the channel into different time scales. The histogram of the closed time shorter than 20 ms filtered at t0 kHz showed a single-exponential distribution, which was considered to reflect the transition from states C1 to O. When fc was reduced to 0.1 kHz and the closed time duration up to 600 ms was included in the histogram, the experimental result was best fitted by a double-exponential. Since two time constants were also obtained from the records filtered at an fc of 10 kHz, these results indicated the presence of another closed state, C2. Finally, when a rest time of greater than 600 ms was taken to separate clusters of bursts from the other events, another closed state, C3, was revealed. While, as mentioned above, the transition between states O and C1 are insensitive to changes in the ATP concentration or pinacidil, the ligand-regulated processes seem to affect

both the transition between states C2 and C3. According to the present analysis and the kinetic model proposed above, the values of K _ > K2 and K - 3 , and their changes can be estimated from %, z~.~ (Figs. 3, 4) and the mean number of bursts per cluster (Colquhoun and Hawkes 1983). An increase in ATP concentration decreased the forward rate constant, K2, and increased both backward rate constants, K_ 2 and K_ 3. Pinacidil antagonized the changes induced by raising the ATP concentration. The changes in K3 produced either by ATP or pinacidil were not estimated due to a lack of data to form a histogram of the burst duration. So far, several substances have been found to induce activity of the ATP-sensitive K + channel when there is sufficient ATP to block the channel activity. These are diazoxide, cromakalim and nicorandil (Trube et al. 1986; Dunne et al. 1987; Escande et al. 1988; Sanguinetti et al. 1988; Hiraoka and Fan 1989). In our experiments, exposure of the internal face of the membrane to pinacidil in the presence of 2 mM ATP restored the channel activity without changing the unitary amplitude, which is comparable to the action of a low ATP concentration on the ATP-sensitive K + current. Such activation induced by pinacidil could again be inhibited by raising the ATP concentration (see Fig. 1). Furthermore, the role of current activation of pinacidil seems to be similar to those of the other three agents, at least qualitatively, although the actions of the other agents on the channel kinetics have not been studied in detail. As described above, pinacidil increased the forward rate constants K2 and K3, and decreased the backward rate constant, K--2, the effects of which are opposite to those of raising the ATP concentration. It was also shown that pinacidil did not change the profile of the channel permeability or apparent reversal potential. These results rule out the possibility that the drug simply enters the channel and impedes the passage of carrier ions, and suggest that it exhibits a functional antagonism at the same site or process as that of the ATPchannel receptor interaction. Interestingly, the interaction between pinacidil and ATP on the gating process of the channel did not require the presence of Mg 2+ because we used an Mg2+-free intracellular solution throughout the experiments. In insulin-secreting cell lines, however, it was reported that diazoxide was unable to activate the ATPsensitive K + channel closed by 1 mM ATP in the absence of Mg 2+ (Dunne et al. 1987). The cell-attached patch-clamp configuration is well suited for investigating the physiological regulation of single-channel activity because the cell remains intact (Hamill et al. 1981). Pinacidil was able to activate the same type of single-channel current in the cell-attached patch configuration as that shown in the inside-out configuration. However, as the intracellular ATP concentration was not controlled, the threshold concentration of pinacidil for activating the channel varied from cell to cell. The threshold concentration was roughly estimated to be between 20 and 100 ~M. This value was somewhat higher than the concentrations required to shorten the action potential duration of Purkinje fibres (Smallwood and Steinberg 1988) or to increase the K + current in ventricular myocytes (Iijima and Taira 1988). A possible explanation for this discrepancy is that the effects of externally applied pinacidil are temperature-dependent, as is the case for cromakalim (Sanguinetti et al. 1988), since our experiments using the cell-attached patch configuration were performed at room temperature. A stronger inward rectification of the channel

393 current in intact cells than that in cell-free membranes m a y be explained by a higher concentration o f m o n o v a l e n t and divalent cations such as N a + and M g 2+ (Horie et al. 1987; F i n d l a y 1987) in the cytoplasm of the former than those in the internal solution that we used. The present finding that pinacidil activates the ATPsensitive K + channel current is undoubtedly of physiological and pharmacological significance. Pinacidil is a new vasodilatory agent currently under clinical investigation regarding its antibypertensive efficacy ( R a m s a y et al. 1983; W a r d 1984). A l t h o u g h pinacidil has been classified as a K + channel opener in s m o o t h muscle cells (Cook et al. 1988; Wilson et al. 1988), the type of channel affected by the drug in these cells remains unclarified. The present results obtained from cardiac myocytes m a y help to solve this issue. Furthermore, the effects o f pinacidil on the ATP-sensitive K § channel current could well explain its ability to shorten the action potential duration in cardiac tissues (Smallwood and Steinberg 1988) and to modify the shape o f the T wave on surface E C G (Goldberg 1988). In addition, pinacidil would be expected to be more effective in the ischaemic heart, where the intracellular ATP concentration is lower than that of the n o r m a l heart. Our d a t a m a y thus draw attention to the action of pinacidil on the ischaemic heart from the viewpoint o f b o t h basic and clinical pharmacology. A n o t h e r interesting aspect is why the class o f agents known as " K + channel openers" ( C o o k 1988), including pinacidil, cromakalim, and nicorandil, are consistently found to have an ability to activate the ATP-sensitive K § channel even though their chemical structures differ so markedly. It m a y therefore be helpful to clarify the details o f their action in order to understand b o t h the pharmacological mechanism of this type o f drug and also the regulatory mechanism o f the ATP-sensitive K + channel itself.

Acknowledgements. Pinacidil was a gift from Shionogi Pharmaceutical Co., Osaka. The authors thank Ms N. Fujita for her secretarial assistance.

References

Arrigoni-Martelli E, Kaergaard Nielsen C, Bang Olsen V, Petersen HJ (1980) N'-cyano-N-pyridyl-N'-l,2,2-trimethytpropyl-guanidine monohydrate (1134): a new potent vasodilator. Experientia 36:445- 447 Ashcroft FM (1988) Adenosine 5'-triphosphate-sensitive potassium channels. Annu Rev Neurosci 11:97 - 1t8 Bhattacharya CG (1967) A simple method of resolution of a distribution into Gaussian components. Biometrics 23 : 115-135 Colquhoun D, Hawkes AG (1983) The principles of the stochastic interpretation of ion-channel mechanisms. In: Sakman B, Neher E (eds) Single-channel recording. Plenum Press, New York, pp 135-175 Cook NS (1988) The pharmacology of potassium channels and their therapeutic potential. Trends Pharmacol Sci 9: 21 - 28 Cook DL, Hales CN (1984) Intracellular ATP directly blocks K + channels in pancreatic B-cells. Nature 3 l 1:27t -273 Cook NS, Quast u, Hof RP, Baumlin Y, PaIly C (1988) Similarities in the mechanism of action of two new vasodilator drugs: pinacidil and BRL 34915. J Cardiovasc Pharmacol 11:90-99 Dunne M J, Illot MC, Petersen OH (1987) Interaction of diazoxide, tolbutamide and ATP 4- on nucleotide-dependent K § channels in an insulin-secreting cell line. J Membr Biol 99:215-224

Escande D, Thuringer D, Leguern S, Cavero I (1988) The potassium channel opener cromakalin (BRL 349J 5) activates ATP-dependent K + channels in isolated cardiac mycoytes. Biochim Biophys Res Commun 154: 620- 625 Escande D, Thuringer D, Mestre M, Cavero I (1988) Potassium channel openers activate the ATP-modulated K + channels in guinea-pig cardiac myocytes. Circulation [Suppl II] 78 : 26 Findlay I (1987) The effects of magnesium upon adenosine triphosphate-sensitive potassium channels in a rat insulinsecreting cell line. J Physiol (Lond) 391:611 - 6 2 9 Findlay I, Dunne M J, Peterson OH (1985) ATP-sensitive inward rectifier and voltage- and calcium activated K + channels in cultured pancreatic islet cells. J Membr Biol 88:165-172 Goldberg MR (1988) Clinical pharmacology of pinacidil, a prototype for drugs that affect potassium channels. J Cardiovasc Pharmacol [Suppl II] 12: S 41 - S 47 Hamill OP, Marty A, Neher E, Sakmann B, Sigworth F (1981) Improved patch-clamp techniques for high-resolution current recordings from cells and cell-free membrane patches. Pfliigers Arch 351:85 - 100 Hirano Y, Hiraoka M (1988) Barium-induced automatic activity in isolated ventricular myocytes from guinea-pig hearts. J Physiol (Lond) 395:455-472 Hiraoka M, Fan Z (I989) Activation of ATP-sensitive outward K § current by nicorandil (2-nicotinamidoethyl nitrate) in isolated ventricular myocytes. J Pharmacol Exp Ther 250:278- 285 Horie M, Irisawa H, Noma A (1987) Voltage-dependent magnesium block of adenosine-triphosphate-sensitive potassium channel in guinea-pig ventricular cells. J Physiol (Lond) 387:251-272 Iijima T, Taira N (1988) Pinacidil increases the background potassium current in single ventricular cells. Eur J Pharmacol 141:I39--141 Isenberg G, Klockner U (1982) Calcium tolerant ventricular myocytes prepared by preincubation in a "KB medium". Pflfigers Arch 395:6-18 Kakei M, Noma A, Shibasaki T (1985) Properties of adenosinetriphosphate-regulated potassium channels in guinea-pig ventricular cells. J Physiol (Lond) 363:441-426 Kameyama M, Kiyosue T, Soejima M (1983) Inward rectifier K channel in the rabbit ventricular cells. Jpn J Physiol 33 : 10391056 Nelder JA, Mead R (1965) A simplex method for function minimization. Comput J 7 : 308 - 313 Noma A (1983) ATP-regulated K + channels in cardiac muscle. Nature 305 : 147-- 148 Ramsey LE, Freestone S (1983) Preliminary evaluation of pinacidil in hypertension. Br J Clin Pharmacol 16 : 3 3 6 - 338 Rorsman P, Trube G (1985) Glucose dependent K + channels in pancreatic B-cells are regulated by intracellular ATP. Pfl/.igers Arch 495 : 305- 309 Sakmann B, Trube G (1984a) Conductance properties of single inwardiy rectifying potassium channels in ventricular cells from guinea-pig hearts. J Physiol (Lond) 347:641-657 Sakmann B, Trube G (1984b) Voltage-dependent inactivation of inward-rectifying single-channel currents in the guinea-pig heart cell membrane. J Physiol (Lond) 347:659--683 Sanguinetti MC, Scott AL, Zingaro GJ, Siegel PK (1988) BRL 34915 (cromakalim) activates ATP-sensitive K + current in cardiac muscle. Proc Natl Acad Sci USA 85: 8360- 8364 Smallwood JK, Steinberg MI (1988) Cardiac electrophysiological effects of pinacidil and related pyridylcyanoguanidines: relationship to antihypertensive activity. J Cardiovasc Pharmacol 12:102-109 Spruce AE, Standen NB, Stanfield PR (1987) Studies on the unitary properties of adenosine-5'triphosphate-regulated potassium channels of frog skeletal muscle. J Physiol (Lond) 382:213-237 Trube G, Hescheler J (1984) Inward-rectifying channels in isolated patches of the heart cell membrane: ATP-dependence and comparison with cell-attached patches. Pfl/igers Arch 401 : 178 - 184

394 Trube G, Rorsmann P, Ohno-Shosaku T (1986) Opposite effects of tolbutamide and diazoxide on the ATP-dependent K + channel in mouse pancreatic/~-cell. Pflfigers Arch 407:493--499 Ward JW (1984) Pinacidil monotherapy for hypertension. Br J Pharmacol 18:223-225 Wilson C, Coldwell MC, Howlett DR, Cooper SM, Hamilton TC (1988) Comparative effects of K + channel blockade on the

vasorelaxant activity of cromakalim, pinacidil and nicorandil. Eur J Pharmacol 152 : 331 - 339 Zilberter Y, Burnashev N, Papin A, Portnov V, Khodorov B (1988) Gating kinetics of ATP-sensitive single potassium channels in myocardial cells depends on electromotive force. Pfliigers Arch 411 : 584-- 589

Note added in proof After the submission of our manuscript, an important paper appeared that described the activation of the ATP-sensitive K + channel by cromakalim and the same mechanism of action by pinacidil as a basis of its vasodilatory action in arterial smooth muscle cells [Standen NB, Quayle JM, Davies NW, Brayden JE, Huang Y, Nelson MT (1989) Hyperpolarizing vasodilators activate ATP-sensirive K + channels in arterial smooth muscle. Science 245 : 177-180].

Announcement

FASEB Summer Research Conference "Physiology and Pathophysiology of the Splanchnic Circulation" July 2 2 - 2 7 , 1990 Copper Mountain, Colorado; Chairman: A. P. Shepherd, Vice-Chairman: C. C. Chou. Inquiries: Federation of American Societies for Experimental Biology, Splanchnic Circulation Conference, 9650 Rockville Pike, Bethesda, MD 20814, USA

Conference Program Session I Session II Session III Session IV Session V Session VI Session VII Session VIII Session IX

Advances in instrumentation and techniques Developmental aspects of the splanchnic circulation Control of the splanchnic circulation: recent advances Lymphatic circulation Oxygen-derived free radicals Actions of inflammatory mediators on the microvasculature Endothelial cells Leukocytes and the microvasculature Splanchnic ischemia and multiple organ failure