- Email: [email protected]
Mastoparan blockade of currents through Ca2+-activated K+ channels in bovine chromaffin cells
0306.4522;92 S5.00 + 0.00 Pergamon Press Ltd ‘i‘ 1992 IBRO
Neuroscience Vol. 50, No. 3, pp. 675-684, 1992 Printed in Great Britain
MASTOPARAN BLOCKADE OF CURRENTS THROUGH Ca2+-ACTIVATED K+ CHANNELS IN BOVINE CHROMAFFIN CELLS M. I. Departments /iSecretory
GLAvINovI~,*t
b§ A. JosHIt
and J. M.
TRIFAR~/~
of *Anesthesia Research, tPhysiology, and $Biomedical Engineering, McGill University, 3655 Drummond Street, Montreal, P.Q., Canada H3G lY6; Process Research Program, Department of Pharmacology, Faculty of Medicine, University of Ottawa, 451 Smyth Road, Ottawa, Ontario, Canada KlH 8M5
Abstract-The action of mastoparan (a wasp venom peptide) on “maxi” Ca’+-activated K+ channels was studied in excised inside-out patch recordings from cultured bovine chromaffin cells, under normal conditions (160 mM K + inside, 154 mM Na+ outside). Mastoparan, when applied on the intracellular side of the membrane reduced the open channel probability in a concentration dependent manner. Changes in the channel kinetics were complex. The histograms of the open dwell times were all described by either one or two exponentials. Mastoparan shortened the mean duration of the major (long) component and to a lesser extent the minor (short) component. Closed dwell times, were described by three exponentials. While the short (major) component was prolonged by mastoparan, and the intermediate component was unaffected, the long component was shortened. Overall mean closed times were prolonged. The changes in channel kinetics could only partly be explained by a channel-blocking mechanism, even when assuming that mastoparan acts as both an intermediate and a slow channel blocker suggesting that it affects gating mechanism. The fact that mastoparan is a calmodulin inhibitor and a G-protein activator raises the possibility that in bovine chromaffin cells, either the membrane-bound calmodulin or a G-protein, plays a role in the modulation of Ca’+-activated K+ channels.
Mastoparan, a tetradecapeptide present in wasp venom (Ile-Asn-Leu-Lys-Ala-Leu-Ala-Ala-Leu-AlaLys-Lys-Ile-Leu NH,) has been shown to stimulate secretion from a variety of cell types including anterior pituitary cells,25 mast cells2 chromaffin cells,” platelets4’ and pancreatic beta cells.53 The effects of mastoparan on these systems could be the result of several actions, such as inhibition of calmodulin dependent processes,3 activation of GTP-binding proteins’“,” and subsequent stimulation of phosphoinoside-specific phospholipase C43.5”or direct stimulation of phospholipase A, with production of arachidonic acid.’ Moreover, it has been shown that some effects of mastoparan are inhibited by pertussis toxin, suggesting the involvement of G-proteins in the mechanism of action of the peptide.2’,40.43.53 Insect venom peptides (i.e. apamine) can also affect K + channels profoundly,’ but the molecular mechanism is not known. It has also been reported that several of these peptides inhibit, some of them very powerfully, calmodulin stimulated phosphodiesterase activity.3 This raises the possibility that calmodulin is involved in regulation of Ca2+-activated K+ channels. Indeed other evidence also suggests that a link between calmodulin and the K(Ca)-channel does
exist. Ca’+-dependent K+ hyperpolarization in mouse fibroblasts was shown to be blocked by intracellular perfusion of trifluoperazine (TFP). a potent inhibitor of calmodulin.3’ Similarly, on inside-out red blood cell vesicles, calmodulin antagonists block K(Ca) conductance by decreasing the rate of radioactive Rb+ uptake.42 On the other hand. no effect of calmodulin on Ca2+-dependent Rb + uptake observed in human blood cell inside-out vesicles argues against such a conclusion.’ Furthermore, in snail neurons, internal perfusion of the cell with calmodulin produced no effect on IK,ca,, nor did the inclusion of 50 PM of TFP in both the bath or in the internal of TFP comperfusate,18 although this concentration pletely blocks calmodulin-dependent enzymes in molluscan ganglia.14 This study represents an attempt to provide direct evidence on the effects of mastoparan on the gating of large conductance Ca2+ -activated K + channels in bovine chromaffin cells. Preliminary results have already been published.‘3
EXPERIMENTAL Isolation
of chromafin
PROCEDURES
cells
Bovine adrenal glands obtained from a slaughterhouse were freed from their cortices and perfused in rirro for 10 minG5 Perfusion was continued for 60 min with a fresh solution to which 04.5% collagenase (Sigma Chemical Co., St Louis. MO, U.S.A.) had been added and chromaffin cells
§To whom correspondence should be addressed. Abbreciations: ethyleneglycolbis(aminoethylEGTA, ether)tetra-acetate; HEPES, N-2_hydroxyethylpiperazine-N’-2-ethanesulphonic acid; TFP, trifluoperazine. 675
were isolated as described previously.‘h The cell preparation thus obtained was layered on a Percoll (Pharmacia) density gradient and chromaffin cells were isolated by centrifugation.“’ Cell cultures Chromaffin cells isolated as described above were plated on glass cover slips previously coated with collagen in Dulbecco’s serum, ascorbic acid (0.1 mM), glucose (1 mg/ml), HEPES buffer (3-mycostatin, 25 units/ml). The culture medium also contained lo-’ M 5-~urodeox~ridine and lo-’ M-cytosine arabinoside to prevent fibroblast proliferation by inhibiting cell division. The culture dishes were incubated at 37°C in a humidi~ed incubator (National) in an air + CO, (95 : 5%) atmosphere. The culture medium was changed every three days and cultures were also inspected at the same time by phase-contrast optics. The cultured cells were used for the preparation of the excised inside-out patches with 10 days of plating.rx Solutions and recording
Ca’+ -activated K + channels of large unit conductanc@
(130-220pS) from bovine adrenal meduliary chromaihn cell membranes were studied in excised inside-out membrane patches.” Patch pipets were prepared using Kimaxglass and were heat polished. The pipets had resistances between 2 and 7 MD when filled with Locke solution which consisted of (in mM): 154 NaCl, 2.2 CaCl,, 1.2 MgCl,, 2.5 KCl, 2.15 K,HPO,, 0.85 KH,PG,, and 10 dextrose. The internal solution contained (in mM): 160 KCljKOH, 10 HEPES, 1 MgCl,, 0.5 EGTA. The estimated free Ca concentrations and the corresponding total amounts of Ca present in the solutions were (in FM): 0.01 free Ca, 55.7 total Ca; 0.1 free Ca, 279 total Ca; 0.25 free Ca, 382 total Ca; 0.5 free Ca. 435 total Ca; 0.75 free Ca, 456 total Ca; 1 free Ca, 467 total Ca; 2 free Ca, 485 total Ca.29When mastoparan (lot No. 006346: Peninsula Laboratories) was used in the internal solution, a sufficient amount of 50 CM stock solution was added to the standard internal solution to make the desired concentration. The peptide was purified by high-performance liquid chromatography and the degree of purity of the preparation was confirmed by thin-layer chromatography (single spot run in three solvent systems), high-voltage electrophoresis (single spot observed) and amino acid analysis. Solutions were changed at the intracellular membrane surface of the patch by placing the recording pipette and attached membrane in a microchamber. All solutions were adjusted to a final pH of 7.2, Experiments were performed at room temperature (2@-23’C). Membrane currents were recorded with a patch-clamping system whose filter was set to 10 kHz, but whose actual band width was 8.9 kHz. The data were stored on a FM tape recorder (Hewlett-Packard. HP-3964A) which had a flat frequency‘response from 0 to 5 kHz. These analog data were then played back at l/16 the original recording speed and antialias filtered by a fourth-order Butterworth filter (Krohn-Hite 3323) set with a 3-dB cut-off frequency of 500Hz (or in some cases 250Hz) and then sampled at l-2 kHz by a micro-PDP-1 l/73 computer (Digital Equipment Corporation) equipped with a la-bit A/D. The equivalent samphng rate and antialiasing frequency was therefore 1632 kHz and 4-8 kHz, respectively. For analysis the effective sampling frequency was finally increased to 80-160 kHz by using a cubic spline interpolation. Analysis
To analyse changes in the gating kinetics produced by mastoparan, current records were displayed and a threshold was set at 50% of the current amplitude. Some of the rapid transitions were too fast to be measured accurately due to filtering introduced by the tape recorder and the Butterworth filter. Since the filter of the patch-clamp system was considerably above cut-off frequencies of the tape recorder
and the Butterworth filter, tt is not expected to t’urther ~Jow the transitions. For typical settings of the fourth-order Butterworth filter of 5 kHz and with frequency response of the FM tape recorder (O--5kHz), the dead time was 59 1~s. Therefore, it is possible to measure mean durations of rapidly declining exponentiai distributions accurately. as long as there is a sufficient number of events defining the curve at intervals longer than 59 ps. Nevertheless WCought to keep in mind that the missed events in such a distribution will tend to artificially prolong the complementary events (i.e. prolongation of open times in the case of missed channel closings). Therefore the duration of open and closed dwell time belonging to the long component was corrected for missed events.” A non-hnear, least-squares curve fitting method that uses a Lavenberg--Marquardt fitting algorithm was used to fit exponentials to data. A I? value that describes the cumulative error between the raw data and the calculated fit was calculated in each case. The number of significant exponential components was obtained from the relationship between the cumulative error and the dwell time duration When the increase in number of exponential components used to tit the observed data does not lead to greater cumulative error it was judged that the additional exponential component was not significant. Fitting an exponentiai to sampled data with least squares gives the correct time-constant but the magnitude of the fitted exponential will be increased over that of the true exponential by the sampling promotion ratio (P,).”
where M,,, is the magnitude of the exponential describing the true interval-duration in the absence of sampling promotion error and M,,, _._is the magnitude determined from fitting the sampled data by least squares. Since the sampling period was less than LO-20% of the fastest time-constant the sampling promotion error was in all cases negligible. Sampling detection error was also ne~i~ble in all cases since the sampling period was fess than 20% of the dead time and the time-constants of detected intervals were greater than dead times. However, the bin width relative to the time-constant of the true intervals was not small. The binning promotion error was, thus, not negligible and had to be taken into account by multiplying P, by the binning promotion ratio (P,,) which is: p,, = ~-~-^ A(O)exp( -O.ST(N,
+ NF)/z)
)
(2)
where Mb is the magnitude of the bin which combines the corrected sampling bins, divided by the magnitude of the exponential describing the distribution of true intervals at the midpoint of the combined bin. I -rw
(31 where r is a common ratio between the magnitudes of successive sampling bins, W is a number of sampling bins combined into one bin, Nr and NL are durations in integral sampling period of the first and last sampling bins in the combined bin, t is the duration of sampling period, z is the time-constant of the distribution of time-intervals and A (0) is the magnitude of the distribution of time-intervals at time zero. Note that A(O) in equation (2) will cancel A(O) in the denominator of equation (3). When data consisted oiseveral exponentials (as was most often the case) each component was corrected separately.
671
Mastoparan and Ca*+-activated K+ channels RESULTS
E#ect of varying intracellular free calcium
Recordings from excised inside-out patches when the internai surface of the membrane was exposed to varying [Ca*+]i at a constant potential of +40 mV are shown in Fig. 1. There were two active channels in this patch. At [Ca2+li 0.1 PM the probability of the channel being in the open state (P) was low, and openings were brief. When the [Ca’+], was increased to 0.25pM, P increased considerably so that the channels were open about 50% of the time. Openings were now much more frequent and separated by short closings. At 0.5pM [Ca*‘]i there was an additional but a small increase in P. The open-state probability (P) of the channel was calculated using the equation:
where k indicated the number of open channels for each peak in the amplitude histogram and A, the integral area of each individual peak (see Fig. 2). Integration was always between two minima on each side of every peak. N indicated the total number of channels functioning within the patch. The maximum number of peaks seen in a particular experiment was taken to indicate the total number of channels in the patch. Calcium dependence of the probability of the open state from four different patches is given in Fig. 2. Note that P increases markedly when [Ca2+ji increases from 0.1 to 0.25 FM. With further increase in [Ca’+], to 1 or 2pM, P remains essentially unchanged.
Effect of varying membrane potential
The open state probability was also modulated by the applied potential across the patch. The voltage dependence of the probabiIity of the open state for three concentrations of [Ca2+Ji are shown in Fig. 3. In low [Ca2+li (0.1 PM) channel openings were rare at -40 mV. They were more often seen at positive potentials but not markedly so. In medium [Ca**], (0.25 PM), at -40 mV, the probability of the channel being open was higher and increased with depolarization up to 0 or +20 mV, but decreased at more depolarized levels. Similar bell-shaped voltage dependence was observed at higher [Ca’+], (0.5 PM). This has been suggested to be due to Ca*+ blockade of K + channel!’ E#ect state
of mastop~ra~
on the probabilit.~
In order to determine whether mastoparan directly affects the channel gating, the amplitude histograms (Fig. 4C) were estimated from the current records before and after mastoparan application (Fig. 4A, B). In the absence of mastoparan, P was 0.45, and 0.26 in its presence (Fig. 4C; holding potential was 0 mV). The ratio of the probabilities of channel opening r (r = P,,,/P) was 0.58. This showed that the probability of the open state was altered by mastoparan suggesting a possible action on channel gating. Figure 5 plots concentration dependence of the effect of mastoparan on the probability of channel opening for mastoparan concentrations ranging from 0.1 to 1OpM (six different patches). It is evident from the figure that the probability of channel opening was reduced as the mastoparan concentration was raised. Kinetic analysis of ma~toparan block
The effect of mastoparan on the channel kinetics was studied further by determining how it alters the
C
0. IyM 01
0.25 ~JM
IlOpA 20ms
c
qf open
OSyM
Oi Fig. 1. Effect of [Ca2+], on the activity of Ca*+ -activated K + channels in an excised chromaffin membrane patch. Records of membrane currents at different [Ca’+], (0.1, 0.25 and 0.5 JLM) at a holding potential of +40 mV are depicted. Two channels are present. C indicates all channels closed (baseline); 0, and 0, indicate one or two open channels.
M. 1.(;LAVINOVI~. er a/.
678
CONTROL
‘t CURRENT
I
OIL 00 5
i 01
I
02 lN%ACELLULAR
05
I Cc”
I
,
2
5
(PM)
Fig. 2. Calcium dependence of the probability of the open state. Holding potential was +40 mV. Note the logarithmic scale. Different symbols indicate different patches.
frequency histograms of the open and the closed dwell times. The effect of mastoparan (10 PM) on the frequency histogram of the open dwell times is shown in Fig. 6A (holding potential was + 20 mV; [Ca*+],
was 0.25pM; this patch had only one channel). In both cases, the histogram of the open dwell times could be well fitted by a single exponential. Clearly the addition of mastoparan shortens the mean open time-constant.
100
80
1 MEMBRANE
POTENTIAL (mV)
Fig. 3. Voltage dependence of the probability of the open state. The solution bathing the intracellular side of the membrane had 0.1 PM (e), 0.25 ph4 (0, V, ‘I, 0) and 0.5 PM (m, a) free calcium.
( PA 1
Fig. 4. Effect of mastoparan on the current flow through Caz+-activated K+ channels. Channel recordings in the absence (A) and in the presence of mastoparan (10 PM) (B). The membrane holding potential was OmV; filtered at 2 kHz. (C) Amplitude distribution for recordings (100 s) of channel openings, from the same patch, plotted as a percentage on the ordinate vs amplitude @A) on the abscissa. Control, (0.25 pM free CaZ+). Mastoparan, (10 PM). Two channels were present in the patch. Marked shift of the amplitude histogram to the left shows that mastoparan powerfully reduces the probability of the open state.
The frequency histograms of the open dwell times are in many cases bi-exponential with a clearly visible minor (short) component. In such a case the short component is also reduced in duration by mastoparan (10 FM) although less so than the long component (Fig. 6B, C; in this experiment as previously only one channel was present in the patch; mean time-constants for both are given in Table 1). Similar changes were observed in three additional patches (all with only one channel) in which high concentrations (5 and 10 PM) of mastoparan were used (see Table 1). Three exponentials were necessary to fit the frequency histograms of the closed dwell times: fast, intermediate and long (Fig. 7). The time-constant of the fast (major) component increased with the addition of mastoparan, that of the intermediate component changed only marginally, while the timeconstant of the long (minor) component decreased. Similar changes were found in five other patches (same as those used to estimate the frequency histograms of the open dwell intervals) where mastoparan (5-10yM) was applied (Table 2). The number of exponentials necessary to fit the frequency histograms of the open or closed dwell times is dependent on the record’s length. With longer lengths the number of exponentials increases.35 In addition, some exponential components
679
Mastoparan and Ca’+-activated K+ channels
0
I
01
02
t
0.5 CONCENTRATION
f
I
2D
I.0
t
5.0
I
IOU
tpMf
Fig. 5. Concentration dependence of the blocking action of mastoparan cm Ca’+-activated K + channels. The probability of the open state was in each case normalized to I.0 in the control solution. The intracellular solution had 0.25pm free Cast m each case. Holding potential was OmV. The different symbols indicate different patches.
may not be very prominent at all Ca’* concentrations.“O Since in this study the intricacies of the mastoparan induced changes in the stochastic nature of gating were not explored, relatively short record lengths (less than 5000 openings or closures) were used t
The effect of filtering of the recording system on the frequency distributions of open and closed dwell times was corrected for as described in Experimental Procedures. In addition, when the data recording and analysis system fails to detect brief closed intervals, the observed mean open time, r”, will overestimate the true mean open time since the adjacent openings, separated by undetected closed intervals will be measured as single openings of longer duration. Activity of maxi Ca’+-activated K + channels appears in bursts in a variety of tissues,“,“’ with brief openings being preceded and followed by long closures and brief closures being preceded and foliowed by long openings.‘7~34 All our estimates of rc, and r, (Tables 1, 2) were corrected for missed events as suggested previously.30 The correction for missed events is only approximate. it can be argued that the closed dwell times become prolonged with mastoparan, at least pa&y, because of the unaccounted, progressivefy greater fraction of missed brief openings. This, however, is very unlike@ because the short openings are associated with fang
closures” and the long closures were shortened rather than further prolonged. Open dwell times could also have been shortened because fewer closures were missed (the closures belonging to the short component were prolonged with the addition of mastoparan; see Table 2). However, such a contribution is very small since even in control conditions only a small fraction of closures belonging to the short component were missed (~6%). With mastoparan this fraction decreases further but even the maximal decrease observed in patch No. 1 was modest (from 5.3% to 4.0%, Table 2).
This study shows that the current Row through the individual large-conductance Ca’+-activated K + channels is reduced by mastoparan. The effect is evident at micromolar concentrations and appears to be concentration dependent. The lower current flow is not due to lower single channel current amplitudes (masto~~ran is thus not a fast blocker) but rather due to the lower probability of the channel opening. Changes in the channel kinetics associated with lower probability of the open state are complex. Shorter duration of the long compo~~ent of open dwell intervals is compatible with slow btockade, but the duration of the long component of closed dwell intervals instead of being longer becomes
M. I.
680
GLAVINOVN?
et
al.
MASTOPARAN
0
IO
20
30
OPEN
40
INTERVAL
so
60
MASTOPABAAN
CONTROL
1
I
,
5
IO
45
70
(mr)
I 20
OPEN
INTERVAL
L
I
I
5
IO
I5
I 20
(ms)
Fig. 6. Frequency distribution of the channel open times in the absence and in the presence of mastoparan (10 @I) in the solution bathing the intraeehular side of the membrane. Holding potential was +2OmV. Filtered at 4 kHz. The intraecll~at solution had 0.25 PM free Ca2+. (A) One exponential could fit the data in both cases suggesting the pmdominancc of one open state. Note that mastoparan shortens the channel opening time. (a) When two exponcntials were ncecssary to fit the data nnu@oparan shortened both the long (major) and the short (minor) components, but the shortening of the long eomponent was more pronounced.
Table 1. EiTcct of mastoparan on open dwell times (T,,) Experiment No. 1 2 3 4 5 6 *S, short; L, long.
Mastoparan WW 10 5 5 10 10 10
No. of exponcntials
I 1 1
1
Components* L L L L
Open dwell times (ms) Before After 12.2 11.3 12.6 10.4
2
S
1.2
2
L S L
11.9 1.3 12.8
7.0 8.0 9.4 6.3 0.95 6.2 1.0 6.9
Inhibition (“/) 42.6 29.2 25.4 39.4 20.0 47.9 23.0 46.1
Mastoparan
and Cd2+-activated
K + channels
3
i;
E m
.
:c
.
(J.
MASTCPARAN
’
I 2
z
3
i. -
IS CLOSED
TIME
(ma)
‘0 -
!O
\.
I
I
L
0
20
30 CLCSEC
40
TIME
50
60
7c
[ ms)
Fig. 7. Mastoparan shortens the long ciosed intervals, and the short intervals are prolonged but the time-constant of the intermediate component appears unaffected. Frequency distributions of the closed dwell times in the absence and in the presence of mastoparan (IOpM) in the solution bathing the intracellular side of the membrane. Holding potential was OmV. Filtered at 4 kHz. The intracellular solution had 0.25 FM free Cal+. (A) Histograms of all closed intervals. The lines describe the slow component of the data. (B) The difference between the distribution of all closed intervals and the distribution of long closed interval. Two exponentials are clearly present. The lines through the slower component indicate the distribution of intermediate short intervals. The faster component obtained after subtraction of intermediate closed intervals indicates the distribution of short dosed intervals. All lines were best fitted using the least squares method. At least three closed states appear to be present since three exponentials were required to fit the data in each case.
Table 2. Effect of mastoparan Experiment NO.
Mastoparan (PM)
No. of exponentials
I
10
3
2
5
3
3
5
3
4
IO
3
5
10
3
6
10
3
*S, short:
I, intermediate:
L, long.
on closed dwell times (T,)
Components* S I L s I L S I L S I L S I L S I L
Closed dwell times (ms) After Before 1.1 5.5 34.7 1.4 4.7 41.3 1.3 4.3 40.2 1.5 4.8 45.3 1.4 5.1 86.3 1.3 4.1 58.4
1.5 5.7 13.2 1.6 4.5 25.2 I.5 4.0 21.3 1.9 4.4 24.0 1.7 4.8 57.8 1.6 4.5 40.0
Change PO) +36 f4.0 -72 fli -4.0 -39 +l6 -7.0 -32 +27 -8.0 -47 +21 - 6.0 -51 +23 -4.0 -44
682
M. 1. GLAWNWIC
shorter. A reduction in the duration of the short component of the open dwell intervals can be explained by a flickery block. However, a flickery block is expected to lead to a new (short) component in the frequency histogram of closed dwell intervals, and that was not observed. This suggests that mastoparan has a direct effect on the gating mechanism. Comments on a possible role of mastoparan in regulating gating of Ca*+ -activated channels
Evidence, although indirect, which suggests that mastoparan affects the gating of Ca2+-activated K+ channels raises the possibility that the gating mechanism is regulated by a Ca’+-calmodulin and that the mastoparan induced changes of the channel kinetics are due to its effect on calmodulin. This is a plausible proposition although at present we do not know the nature of such a calmodulin regulation of these channels. Furthermore, the mastoparan concentration necessary to inhibit phosphodiesterase was much lower than the concentration necessary to have 50% of the open channel time (50% inhibition was observed at 0.02 PM of mastoparan and we observed that P is 0.5 at 0.25 PM of masto-paran).’ Since the target of regulation must be a component which remains associated when the patch of the membrane is excised this would argue that it is the membrane bound calmodulin that is the agent involved in the regulation of channel gating. Calmodulin was indeed found to be present not only free in the cytoplasm of the chromaffin cell, but also in the membrane-bound form.22 Ca*+-calmodulin was found to induce the activation of a number of different enzymes including phosphatases and protein kinases.’ In addition, phosphorylation of ion channels or their protein subunits has been shown to be a major mechanism in their modulation.*’ Examples include K(Ca)-channels in snail neurons,” sodium channels from various cellular preparations,“,” and voltagesensitive calcium channels from rat skeletal muscle.13 The possibility of a calmodulin regulated K(Ca)-channel has been suggested by studies of putative calmodulin antagonists.3y,4’ Furthermore, a recent report showing that a Paramecium mutant with altered calmodulin has a defective calcium-dependent potassium conductance supports such a conclusion.44 Lack of such an effect in snail neurons could have occurred because of species and/or tissue differences, but also due to different contributions of two types of Ca?+-activated K+ channels (maxi and small). Recent single channel studies have shown that TFP affects the current flow through Ca’+-activated K + channels.3’ It is also known that mastoparan activates Gproteins. *‘,*’The peptide stimulates guanine nucleotide exchange by GTP-binding proteins in a manner similar to that of G-protein-coupled receptors.?” Moreover, ADP-ribosylation by pertussis toxin, which uncouples receptors from G-proteins, selectively inhibits mastoparan activation of G-proteins” as well as mastoparan effects on several cell
et al.
types,2.3X.40.53 Therefore, the effects of mastoparan on Ca* +-activated K + channels described here could also be due to an action of the peptide on a Gprotein, whose activity is coupled to the channel. Guanine nucleotides have been shown to affect the activity of different ion channels.“,’
Calcium and voltage sensitivity
of K(Ca)-channels
Calmodulin has four binding sites for calcium.” If calmodulin is involved in regulating maxi Ca2’-activated K+ channels it is desirable to have at least some evidence about the calcium sensitivity of these channels. This study shows that in bovine chromaffin cell membranes, [Ca2+], required to produce P = 0.5 at OmV (P, probability of the open state) was 0.24.3 FM thus in the same range as the affinity for calcium of calmodulin.24 This is comparable to [Ca’ ‘1, required to half-activate the channels in smooth muscle cells (0.5 PM),’ but is below the value for cultured rat muscle4,36 which was approximately 5 PM, particularly when these channels were incorporated into lipid bilayers. In such a case the concentration required was even higher (50 PM).~‘.~~ However, the calcium sensitivity of channels in other secretory cells is even higher, the comparable concentrations being 0.01 PM (mouse lacrimal glands)16 and 0.02 PM (pig pancreatic acinar cells).31 Thus, among calcium-activated K + channels from secretory cells, the channels from bovine cell membranes are not particularly sensitive to calcium. Gating of maxi K+(Ca*+) channels was found to be voltage dependent, as has been previously reported in rabbit transverse tubule,26 rat pituitary,” mammalian salivary gland,3’ rat muscle,4’ canine airway smooth muscle,32 smooth muscle of rabbit jejunum and guinea-pig mesenteric artery.5 However, unlike most other cases the probability of the open state instead of becoming progressively greater as the membrane becomes more depolarized, decreased at holding potentials above 0 to +20 mV as was reported for the smooth muscle.‘.” CONCLUSION
If calmodulin modulates calcium sensitivity of maxiK(Ca) channels from different cells their very variable Ca2+ -sensitivity would argue that the affinity of calmodulin for Ca2+ varies considerably from one type of cell to another, and that it is modest in bovine chromaffin cell membranes. Alternatively, and more likely, other factors in addition to calmodulin, such as G-proteins, probably influence the calcium sensitivity of maxi channels. Further studies should provide answers to these questions. Acknowledgements-This work
was supported by grants PG-20 (J.M.T.) and MA-9160 (M.I.G.) from the Medical Research Council of Canada. We thank Mrs R. Tang and
Mr D. Clayton
for their technical
assistance.
Mastoparan
and Cal+-activated
K + channels
683
REFERENCES 1. Alvarez J., Garcia-Sancho J. and Herreros B. (1986) The role of calmodulin on Ca’+-dependent K+ transport regulation in the human red cell. Biochem. biophys. Acta 860, 25-34. 2. Argiolas A. and Pisano J. J. (1983) Facilitation of phospholipase A, activity by mastoparans, a new class of mast cell degranulation peptides from wasp venom. J. biol. Chem. 258, 13,697713,702. 3. Barnette M. S., Daly R. and Weiss B. (1983) Inhibition of calmodulin activity by insect venom peptides. Biochem. Pharmac. 32, 2929-2933. 4. Barrett J. N., Magleby K. L. and Pallotta B. S. (1982) Properties of single calcium-activated potassium channels in cultured rat muscle. J. Physiol., Lond. 331, 21 l-230. 5. Benham C. D., Bolton T. B., Lang R. J. and Takewaki T. (1986) Calcium-activated potassium channels in single smooth muscle cells of rabbit jejunum and guinea-pig mesenteric artery. J. Physiol., Lond. 371, 4567. 6. Birnbaumer L., Abramowitz J. and Brown A. M. (1990) Receptor-effector coupling by G proteins. Biochem. bioph>,.v. Acta 1031, 163-224. 7. Blatz L. A. and Magleby K. L. (1986) Single apamin-blocked Ca-activated K+ channels of small conductance in cultured rat skeletal muscle. Nature 323, 718-720. 8. Brown D. A. (1990) G-proteins and potassium currents in neurons. A. Rev. Physiol. 52, 215-242. 9. Cheung W. Y. (1980) Calmodulin plays a pivotal role in cellular regulation. Science 207, 19-27. IO. Colquhoun D. and Sigworth F. J. (1983) Fitting and statistical analysis of single-channel records. In Single-Channel Recording (eds Sakmann B. and Neher E.), pp. 191-263. Plenum Press, New York. I 1. Costa M. R., Casnellie J. E. and Catterall W. A. (1982) Selective phosphorylation of the alpha subunit of the sodium channel by cALMP-dependent protein kinase. J. bioi. Chem. 257, 791Sp7921. 12. Costa M. R. and Catterall W. A. (1984) Phosphorylation in the subunit of the sodium channel by protein kinase C. Cell. molec. Neurobiol. 4, 291-297. 13. Curtis B. M. and Catterall W. A. (1985) Phosphorylation of the calcium antagonist of the voltage-sensitive calcium channel by CAMP-dependant protein kinase. Proc. num. Acad. Sci. C/.S.A. 82, 2528-2532. 14. DeReimer S. A., Kaczmarek L. K., Lai Y., McGuinness T. L. and Greengard P. (1984) Calcium/Calmodulin-dependent protein phosphorylation in the nervous system of Aplysia. J. Neurosci. 4, 1618-1625. 15. dePayer J. E.. Cachelin A. B.. Levitan I. B. and Reuter H. (1982) Ca’+-activated K + conductance in internally perfused snail neurons is enhanced by protein phosphorylation. Proc. natn. Acad. Sci. U.S.A. 79, 4207421 I, 16. Findlay I. (1984) A patch clamp study of potassium channels and whole-cell currents in acinar cells of the mouse lacrimal gland. J. Physiol., Land. 350, 179-195. 17. Glavinovic M. I. (1986) Oscillations of Ca2+ -activated K _ channels around various closed states in bovine chromaffin cells. Sot. Neurosci. Abstr. 12, 1198. IS. Glavinovic M. I. and Trifaro J.-M. (1988) Quinine blockade of currents through Ca’+-activated K+ channels in bovine chromaffin cells. J. Physiol., Lond. 399, 139-152. 19. Hamill 0. P., Marty A., Neher E., Sakmann B. and Sigworth F. J. (1981) Improved patch-clamp technique for high-resolution current recording from cells and cell-free membrane patches. Pfliigers Arch. ges Physiol. 391, 85-100. 20. Higashijima T., Bournier J. and Ross E. M. (1990) Regulation of G, and G, by mastoparan, related amphiphilic peptides, and hydrophobic amines. J. biol. Chem. 265, 14,17614,186. 21. Higashijima T., Uzu S., Nakajima T. and Ross E. M. (1988) Mastoparan, a peptide toxin from wasp venom, mimics receptors by activating GTP-binding regulatory proteins (G-proteins). J. biol. Chem. 263, 64914494. 22. Hikita T., Bader M.-F. and Trifaro J.-M. (1984) Adrenal chromaffin cell calmodulin: its subcellular distribution and binding to chromaffin granule membrane proteins. J. Neurochem. 43, 1087-1097. 23. Joshi A., Trifaro J.-M. and Glavinovic M. I. (1986) Effect of mastoparan on Ca’+ activated K + channels in bovine chromaffin cells. Sot. Neurosci. Abstr. 12, 1076. 24. Klee C. B.. Crouch T. H. and Richman P. G. (1980) Calmodulin. A. Rec. Biochem. 49, 489-515. 25. Kurihara H., Kitajima K., Senda T., Fujita H. and Nakajima T. (1986) Multigranular exocytosis induced by phospholipase AZ-activators, melittin and mastoparan, in rat anterior pituitary cells. Cell Tiss. Res. 243, 31 I-316. 26. Latorre R., Vergara C. and Hidalgo C. (1982) Reconstitution in planar lipid bilayers of a Ca’+-dependent K + channel from transverse tubule membranes isolated from rabbit skeletal muscle. Proc. natn. Acad. Sci. CT.S.A. 79, 805-809. 27. Levitan I. B. (1985) Phosphorylation of ion channels. J. Membrane Biol. 87, 177-190. 28. Levitan E. S. and Levitan I. B. (1986) Apparent loss of calcium-activated potassium current in internally perfused snail neurons is due to accumulation of free intracellular calcium. J. Membrane Biol. 90, 5945. 29. Magleby K. L. and Pallotta B. S. (1983) Calcium dependence of open and shut interval distributions from calcium-activated potassium channels in cultured muscle. J. Physiol., Lond. 344, 585-604. 30. Magleby K. L. and Pallotta B. S. (1983) Burst kinetics of single calcium-activated potassium channels in cultured rat muscle. J. Physiol., Lond. 344, 605423. 3 I. Maruyama Y., Gallacher D. W. and Peterson 0. H. (1983) Voltage and Ca’+-activated K + channel in basolateral acinar cell membranes of mammalian salivary glands. Nature 302, 827-829. 32. McCann J. D. and Welsh M. J. (1987) Neuroleptics antagonize a calcium-activated potassium channel in airway smooth muscle. J. gen. Physiol. 89, 339-352. 33. McManus 0. B., Blatz A. L. and Magleby K. L. (1987) Sampling. log binning, fitting, and plotting durations of open and shut intervals from single channels and the effects of noise. PJtigers Arch. ges Physioi. 410, 530-553. 34. McManus 0. B. and Magleby K. L. (1985) Inverse relationship of the durations of adjacent open and shut intervals for Cl- and K+ channels. Nature 317, 625427. 35. McManus 0. B. and Magleby K. L. (1988) Kinetic states and modes of single large-conductance calcium-activated potassium channels in cultured rat skeletal muscle. J. Physiol., Lond. 402, 79-120. 36. Methfessel G. and Boheim G. (1982) The gating of single calcium-dependent potassium channels is described by an activation/blockade mechanism. Biophys. sfruct. Mech. 9, 3540. 37. Moczydlowski E. and Latorre R. (1981) Gating kinetics of Ca’+-activated potassium channels from rat muscle incorporated into planar lipid bilayers: evidence for two voltage-dependent Ca2+ binding reactions. J. gen. Physiol. 82, 51 l-542.
684
M. I. GLAVINOVI~. ct ul.
38. Mousli M., Brommer C., Bueb J.-L., Tschirhart E., Gies J.-P. and Landry Y. (1989) Activation of rat peritoneal macr cells by substance P and mastoparan. J. Pharmac. exp. Ther. 250, 329-335. 39. Okada Y., Yada T., Ohno-Shosaku T. and Oiki S. (1987) Evidence for the involvement of calmodulin in the operation of Cal+-activated K+ channels in mouse fibroblasts. J. Membrane Biol. 96, 121-128. 40. Ozaki Y., Matsumoto Y.. Yatomi Y., Higashihara M., Kariya T. and Kume S. (1990) Mastoparan, a wasp venom, activates platelets via pertussis toxin-sensitive GTP-binding proteins. Biochem. biophys. Res. Commun. 170, 779 785. 41. Pallotta B. S., Magleby K. L. and Barrett J. N. (1981) Single channel recordings of Ca*+-activated K+ currents in rat muscle cell culture. Nature 293, 471474. 42. Pape L. and Kristensen B. 1. (1984) A calmodulin activated CaLf-dependent K+ channel in human erythrocyte membrane inside-out vesicles. Biochem. biophp. Acfu 770, l-6. 43. Perianin A. and Snyderman R. (1989) Mastoparan, a wasp venom peptide, identifies two discrete mechanisms for evaluating cytosolic calcium and inositol triphosphates in human polymorphonuclear leukocytes. J. Immunol. 143, 1669.-1673. 44. Schaefer W. H., Hinrichsen R. D., Burgess-Casler A., Kung C., Blair 1. A. and Watterson D. M. (1987) A mutant Paramecium with a defective calcium-dependent potassium conductance has an altered calmodulin: a nonlethal selective alteration in calmodulin regulation. Proc,. nafn. Acud. Ski. U.S.A. 84, 3931--3935. 45. Trifaro J. M., Poisner A. M. and Douglas W. W. (1967) The fate of chromaffin granules during catecholamine release from the adrenal medulla. I. Unchanged efflux of phospholipid and cholesterol. Biochem. Pharmacol. 16, 2095.-2100. 46. Trifaro J.-M., Ulpian C. and Preiksaitis H. (1979) Antimyosin stains chromaffin cells. Experientia 34, 1568-1571. 47. Trifaro J.-M., and Lee R. W. H. (1980) Morphological characteristics and stimulus-secretion coupling involving adrenal chromaffin cell cultures. Neuroscience 5, 1433 -1546. into planar 48. Vergara C. and Latorre R. (1983) Kinetics of Ca*+ -activated K+ channels from rabbit muscle incorporated bilayers. J. Physiol., Lond. 82, 543~.568. 49. Vergara C., Moczydlowski E. and Latorre R. (1984) Conduction, blockade and gating in a Ca’ +-activated K ’ channel incorporated into planar lipid bilayer. Biophys. J. 45, 73 -76. 50. Wallace M. A. and Carter H. R. (1989) Effects of wasp venom peptide, mastoparan, on a phosphoinositide-specific phospholipase C purified from rabbit brain membranes. Biochem. biophys. Acta 1006, 31 I-316. 51, Wilson S. P. (1989) Effects of mastoparan on catecholamine release from chromaffin cells. Fedn Eur. biochem. Sots Lerr. 247, 239-24 1. 52. Wong B. S., Lecar H. and Adler M. (1982) Single Ca2+-dependent K+ channels in clonal pituitary cells. Biophys. J. 39, 313-317. 53. Yokokawa N., Komatsu M., Takeda T., Aizawa T. and Yamada T. (1989) Mastoparan, a wasp venom, stimulates insulin release by pancreatic islets through pertussis toxin sensitive GTP-binding protein. Biochem. biophw Res. Commun. 158, 712.-716. (Accepted
14 May 1992)
Neuroscience Vol. 50, No. 3, pp. 675-684, 1992 Printed in Great Britain
MASTOPARAN BLOCKADE OF CURRENTS THROUGH Ca2+-ACTIVATED K+ CHANNELS IN BOVINE CHROMAFFIN CELLS M. I. Departments /iSecretory
GLAvINovI~,*t
b§ A. JosHIt
and J. M.
TRIFAR~/~
of *Anesthesia Research, tPhysiology, and $Biomedical Engineering, McGill University, 3655 Drummond Street, Montreal, P.Q., Canada H3G lY6; Process Research Program, Department of Pharmacology, Faculty of Medicine, University of Ottawa, 451 Smyth Road, Ottawa, Ontario, Canada KlH 8M5
Abstract-The action of mastoparan (a wasp venom peptide) on “maxi” Ca’+-activated K+ channels was studied in excised inside-out patch recordings from cultured bovine chromaffin cells, under normal conditions (160 mM K + inside, 154 mM Na+ outside). Mastoparan, when applied on the intracellular side of the membrane reduced the open channel probability in a concentration dependent manner. Changes in the channel kinetics were complex. The histograms of the open dwell times were all described by either one or two exponentials. Mastoparan shortened the mean duration of the major (long) component and to a lesser extent the minor (short) component. Closed dwell times, were described by three exponentials. While the short (major) component was prolonged by mastoparan, and the intermediate component was unaffected, the long component was shortened. Overall mean closed times were prolonged. The changes in channel kinetics could only partly be explained by a channel-blocking mechanism, even when assuming that mastoparan acts as both an intermediate and a slow channel blocker suggesting that it affects gating mechanism. The fact that mastoparan is a calmodulin inhibitor and a G-protein activator raises the possibility that in bovine chromaffin cells, either the membrane-bound calmodulin or a G-protein, plays a role in the modulation of Ca’+-activated K+ channels.
Mastoparan, a tetradecapeptide present in wasp venom (Ile-Asn-Leu-Lys-Ala-Leu-Ala-Ala-Leu-AlaLys-Lys-Ile-Leu NH,) has been shown to stimulate secretion from a variety of cell types including anterior pituitary cells,25 mast cells2 chromaffin cells,” platelets4’ and pancreatic beta cells.53 The effects of mastoparan on these systems could be the result of several actions, such as inhibition of calmodulin dependent processes,3 activation of GTP-binding proteins’“,” and subsequent stimulation of phosphoinoside-specific phospholipase C43.5”or direct stimulation of phospholipase A, with production of arachidonic acid.’ Moreover, it has been shown that some effects of mastoparan are inhibited by pertussis toxin, suggesting the involvement of G-proteins in the mechanism of action of the peptide.2’,40.43.53 Insect venom peptides (i.e. apamine) can also affect K + channels profoundly,’ but the molecular mechanism is not known. It has also been reported that several of these peptides inhibit, some of them very powerfully, calmodulin stimulated phosphodiesterase activity.3 This raises the possibility that calmodulin is involved in regulation of Ca2+-activated K+ channels. Indeed other evidence also suggests that a link between calmodulin and the K(Ca)-channel does
exist. Ca’+-dependent K+ hyperpolarization in mouse fibroblasts was shown to be blocked by intracellular perfusion of trifluoperazine (TFP). a potent inhibitor of calmodulin.3’ Similarly, on inside-out red blood cell vesicles, calmodulin antagonists block K(Ca) conductance by decreasing the rate of radioactive Rb+ uptake.42 On the other hand. no effect of calmodulin on Ca2+-dependent Rb + uptake observed in human blood cell inside-out vesicles argues against such a conclusion.’ Furthermore, in snail neurons, internal perfusion of the cell with calmodulin produced no effect on IK,ca,, nor did the inclusion of 50 PM of TFP in both the bath or in the internal of TFP comperfusate,18 although this concentration pletely blocks calmodulin-dependent enzymes in molluscan ganglia.14 This study represents an attempt to provide direct evidence on the effects of mastoparan on the gating of large conductance Ca2+ -activated K + channels in bovine chromaffin cells. Preliminary results have already been published.‘3
EXPERIMENTAL Isolation
of chromafin
PROCEDURES
cells
Bovine adrenal glands obtained from a slaughterhouse were freed from their cortices and perfused in rirro for 10 minG5 Perfusion was continued for 60 min with a fresh solution to which 04.5% collagenase (Sigma Chemical Co., St Louis. MO, U.S.A.) had been added and chromaffin cells
§To whom correspondence should be addressed. Abbreciations: ethyleneglycolbis(aminoethylEGTA, ether)tetra-acetate; HEPES, N-2_hydroxyethylpiperazine-N’-2-ethanesulphonic acid; TFP, trifluoperazine. 675
were isolated as described previously.‘h The cell preparation thus obtained was layered on a Percoll (Pharmacia) density gradient and chromaffin cells were isolated by centrifugation.“’ Cell cultures Chromaffin cells isolated as described above were plated on glass cover slips previously coated with collagen in Dulbecco’s serum, ascorbic acid (0.1 mM), glucose (1 mg/ml), HEPES buffer (3-mycostatin, 25 units/ml). The culture medium also contained lo-’ M 5-~urodeox~ridine and lo-’ M-cytosine arabinoside to prevent fibroblast proliferation by inhibiting cell division. The culture dishes were incubated at 37°C in a humidi~ed incubator (National) in an air + CO, (95 : 5%) atmosphere. The culture medium was changed every three days and cultures were also inspected at the same time by phase-contrast optics. The cultured cells were used for the preparation of the excised inside-out patches with 10 days of plating.rx Solutions and recording
Ca’+ -activated K + channels of large unit conductanc@
(130-220pS) from bovine adrenal meduliary chromaihn cell membranes were studied in excised inside-out membrane patches.” Patch pipets were prepared using Kimaxglass and were heat polished. The pipets had resistances between 2 and 7 MD when filled with Locke solution which consisted of (in mM): 154 NaCl, 2.2 CaCl,, 1.2 MgCl,, 2.5 KCl, 2.15 K,HPO,, 0.85 KH,PG,, and 10 dextrose. The internal solution contained (in mM): 160 KCljKOH, 10 HEPES, 1 MgCl,, 0.5 EGTA. The estimated free Ca concentrations and the corresponding total amounts of Ca present in the solutions were (in FM): 0.01 free Ca, 55.7 total Ca; 0.1 free Ca, 279 total Ca; 0.25 free Ca, 382 total Ca; 0.5 free Ca. 435 total Ca; 0.75 free Ca, 456 total Ca; 1 free Ca, 467 total Ca; 2 free Ca, 485 total Ca.29When mastoparan (lot No. 006346: Peninsula Laboratories) was used in the internal solution, a sufficient amount of 50 CM stock solution was added to the standard internal solution to make the desired concentration. The peptide was purified by high-performance liquid chromatography and the degree of purity of the preparation was confirmed by thin-layer chromatography (single spot run in three solvent systems), high-voltage electrophoresis (single spot observed) and amino acid analysis. Solutions were changed at the intracellular membrane surface of the patch by placing the recording pipette and attached membrane in a microchamber. All solutions were adjusted to a final pH of 7.2, Experiments were performed at room temperature (2@-23’C). Membrane currents were recorded with a patch-clamping system whose filter was set to 10 kHz, but whose actual band width was 8.9 kHz. The data were stored on a FM tape recorder (Hewlett-Packard. HP-3964A) which had a flat frequency‘response from 0 to 5 kHz. These analog data were then played back at l/16 the original recording speed and antialias filtered by a fourth-order Butterworth filter (Krohn-Hite 3323) set with a 3-dB cut-off frequency of 500Hz (or in some cases 250Hz) and then sampled at l-2 kHz by a micro-PDP-1 l/73 computer (Digital Equipment Corporation) equipped with a la-bit A/D. The equivalent samphng rate and antialiasing frequency was therefore 1632 kHz and 4-8 kHz, respectively. For analysis the effective sampling frequency was finally increased to 80-160 kHz by using a cubic spline interpolation. Analysis
To analyse changes in the gating kinetics produced by mastoparan, current records were displayed and a threshold was set at 50% of the current amplitude. Some of the rapid transitions were too fast to be measured accurately due to filtering introduced by the tape recorder and the Butterworth filter. Since the filter of the patch-clamp system was considerably above cut-off frequencies of the tape recorder
and the Butterworth filter, tt is not expected to t’urther ~Jow the transitions. For typical settings of the fourth-order Butterworth filter of 5 kHz and with frequency response of the FM tape recorder (O--5kHz), the dead time was 59 1~s. Therefore, it is possible to measure mean durations of rapidly declining exponentiai distributions accurately. as long as there is a sufficient number of events defining the curve at intervals longer than 59 ps. Nevertheless WCought to keep in mind that the missed events in such a distribution will tend to artificially prolong the complementary events (i.e. prolongation of open times in the case of missed channel closings). Therefore the duration of open and closed dwell time belonging to the long component was corrected for missed events.” A non-hnear, least-squares curve fitting method that uses a Lavenberg--Marquardt fitting algorithm was used to fit exponentials to data. A I? value that describes the cumulative error between the raw data and the calculated fit was calculated in each case. The number of significant exponential components was obtained from the relationship between the cumulative error and the dwell time duration When the increase in number of exponential components used to tit the observed data does not lead to greater cumulative error it was judged that the additional exponential component was not significant. Fitting an exponentiai to sampled data with least squares gives the correct time-constant but the magnitude of the fitted exponential will be increased over that of the true exponential by the sampling promotion ratio (P,).”
where M,,, is the magnitude of the exponential describing the true interval-duration in the absence of sampling promotion error and M,,, _._is the magnitude determined from fitting the sampled data by least squares. Since the sampling period was less than LO-20% of the fastest time-constant the sampling promotion error was in all cases negligible. Sampling detection error was also ne~i~ble in all cases since the sampling period was fess than 20% of the dead time and the time-constants of detected intervals were greater than dead times. However, the bin width relative to the time-constant of the true intervals was not small. The binning promotion error was, thus, not negligible and had to be taken into account by multiplying P, by the binning promotion ratio (P,,) which is: p,, = ~-~-^ A(O)exp( -O.ST(N,
+ NF)/z)
)
(2)
where Mb is the magnitude of the bin which combines the corrected sampling bins, divided by the magnitude of the exponential describing the distribution of true intervals at the midpoint of the combined bin. I -rw
(31 where r is a common ratio between the magnitudes of successive sampling bins, W is a number of sampling bins combined into one bin, Nr and NL are durations in integral sampling period of the first and last sampling bins in the combined bin, t is the duration of sampling period, z is the time-constant of the distribution of time-intervals and A (0) is the magnitude of the distribution of time-intervals at time zero. Note that A(O) in equation (2) will cancel A(O) in the denominator of equation (3). When data consisted oiseveral exponentials (as was most often the case) each component was corrected separately.
671
Mastoparan and Ca*+-activated K+ channels RESULTS
E#ect of varying intracellular free calcium
Recordings from excised inside-out patches when the internai surface of the membrane was exposed to varying [Ca*+]i at a constant potential of +40 mV are shown in Fig. 1. There were two active channels in this patch. At [Ca2+li 0.1 PM the probability of the channel being in the open state (P) was low, and openings were brief. When the [Ca’+], was increased to 0.25pM, P increased considerably so that the channels were open about 50% of the time. Openings were now much more frequent and separated by short closings. At 0.5pM [Ca*‘]i there was an additional but a small increase in P. The open-state probability (P) of the channel was calculated using the equation:
where k indicated the number of open channels for each peak in the amplitude histogram and A, the integral area of each individual peak (see Fig. 2). Integration was always between two minima on each side of every peak. N indicated the total number of channels functioning within the patch. The maximum number of peaks seen in a particular experiment was taken to indicate the total number of channels in the patch. Calcium dependence of the probability of the open state from four different patches is given in Fig. 2. Note that P increases markedly when [Ca2+ji increases from 0.1 to 0.25 FM. With further increase in [Ca’+], to 1 or 2pM, P remains essentially unchanged.
Effect of varying membrane potential
The open state probability was also modulated by the applied potential across the patch. The voltage dependence of the probabiIity of the open state for three concentrations of [Ca2+Ji are shown in Fig. 3. In low [Ca2+li (0.1 PM) channel openings were rare at -40 mV. They were more often seen at positive potentials but not markedly so. In medium [Ca**], (0.25 PM), at -40 mV, the probability of the channel being open was higher and increased with depolarization up to 0 or +20 mV, but decreased at more depolarized levels. Similar bell-shaped voltage dependence was observed at higher [Ca’+], (0.5 PM). This has been suggested to be due to Ca*+ blockade of K + channel!’ E#ect state
of mastop~ra~
on the probabilit.~
In order to determine whether mastoparan directly affects the channel gating, the amplitude histograms (Fig. 4C) were estimated from the current records before and after mastoparan application (Fig. 4A, B). In the absence of mastoparan, P was 0.45, and 0.26 in its presence (Fig. 4C; holding potential was 0 mV). The ratio of the probabilities of channel opening r (r = P,,,/P) was 0.58. This showed that the probability of the open state was altered by mastoparan suggesting a possible action on channel gating. Figure 5 plots concentration dependence of the effect of mastoparan on the probability of channel opening for mastoparan concentrations ranging from 0.1 to 1OpM (six different patches). It is evident from the figure that the probability of channel opening was reduced as the mastoparan concentration was raised. Kinetic analysis of ma~toparan block
The effect of mastoparan on the channel kinetics was studied further by determining how it alters the
C
0. IyM 01
0.25 ~JM
IlOpA 20ms
c
qf open
OSyM
Oi Fig. 1. Effect of [Ca2+], on the activity of Ca*+ -activated K + channels in an excised chromaffin membrane patch. Records of membrane currents at different [Ca’+], (0.1, 0.25 and 0.5 JLM) at a holding potential of +40 mV are depicted. Two channels are present. C indicates all channels closed (baseline); 0, and 0, indicate one or two open channels.
M. 1.(;LAVINOVI~. er a/.
678
CONTROL
‘t CURRENT
I
OIL 00 5
i 01
I
02 lN%ACELLULAR
05
I Cc”
I
,
2
5
(PM)
Fig. 2. Calcium dependence of the probability of the open state. Holding potential was +40 mV. Note the logarithmic scale. Different symbols indicate different patches.
frequency histograms of the open and the closed dwell times. The effect of mastoparan (10 PM) on the frequency histogram of the open dwell times is shown in Fig. 6A (holding potential was + 20 mV; [Ca*+],
was 0.25pM; this patch had only one channel). In both cases, the histogram of the open dwell times could be well fitted by a single exponential. Clearly the addition of mastoparan shortens the mean open time-constant.
100
80
1 MEMBRANE
POTENTIAL (mV)
Fig. 3. Voltage dependence of the probability of the open state. The solution bathing the intracellular side of the membrane had 0.1 PM (e), 0.25 ph4 (0, V, ‘I, 0) and 0.5 PM (m, a) free calcium.
( PA 1
Fig. 4. Effect of mastoparan on the current flow through Caz+-activated K+ channels. Channel recordings in the absence (A) and in the presence of mastoparan (10 PM) (B). The membrane holding potential was OmV; filtered at 2 kHz. (C) Amplitude distribution for recordings (100 s) of channel openings, from the same patch, plotted as a percentage on the ordinate vs amplitude @A) on the abscissa. Control, (0.25 pM free CaZ+). Mastoparan, (10 PM). Two channels were present in the patch. Marked shift of the amplitude histogram to the left shows that mastoparan powerfully reduces the probability of the open state.
The frequency histograms of the open dwell times are in many cases bi-exponential with a clearly visible minor (short) component. In such a case the short component is also reduced in duration by mastoparan (10 FM) although less so than the long component (Fig. 6B, C; in this experiment as previously only one channel was present in the patch; mean time-constants for both are given in Table 1). Similar changes were observed in three additional patches (all with only one channel) in which high concentrations (5 and 10 PM) of mastoparan were used (see Table 1). Three exponentials were necessary to fit the frequency histograms of the closed dwell times: fast, intermediate and long (Fig. 7). The time-constant of the fast (major) component increased with the addition of mastoparan, that of the intermediate component changed only marginally, while the timeconstant of the long (minor) component decreased. Similar changes were found in five other patches (same as those used to estimate the frequency histograms of the open dwell intervals) where mastoparan (5-10yM) was applied (Table 2). The number of exponentials necessary to fit the frequency histograms of the open or closed dwell times is dependent on the record’s length. With longer lengths the number of exponentials increases.35 In addition, some exponential components
679
Mastoparan and Ca’+-activated K+ channels
0
I
01
02
t
0.5 CONCENTRATION
f
I
2D
I.0
t
5.0
I
IOU
tpMf
Fig. 5. Concentration dependence of the blocking action of mastoparan cm Ca’+-activated K + channels. The probability of the open state was in each case normalized to I.0 in the control solution. The intracellular solution had 0.25pm free Cast m each case. Holding potential was OmV. The different symbols indicate different patches.
may not be very prominent at all Ca’* concentrations.“O Since in this study the intricacies of the mastoparan induced changes in the stochastic nature of gating were not explored, relatively short record lengths (less than 5000 openings or closures) were used t
The effect of filtering of the recording system on the frequency distributions of open and closed dwell times was corrected for as described in Experimental Procedures. In addition, when the data recording and analysis system fails to detect brief closed intervals, the observed mean open time, r”, will overestimate the true mean open time since the adjacent openings, separated by undetected closed intervals will be measured as single openings of longer duration. Activity of maxi Ca’+-activated K + channels appears in bursts in a variety of tissues,“,“’ with brief openings being preceded and followed by long closures and brief closures being preceded and foliowed by long openings.‘7~34 All our estimates of rc, and r, (Tables 1, 2) were corrected for missed events as suggested previously.30 The correction for missed events is only approximate. it can be argued that the closed dwell times become prolonged with mastoparan, at least pa&y, because of the unaccounted, progressivefy greater fraction of missed brief openings. This, however, is very unlike@ because the short openings are associated with fang
closures” and the long closures were shortened rather than further prolonged. Open dwell times could also have been shortened because fewer closures were missed (the closures belonging to the short component were prolonged with the addition of mastoparan; see Table 2). However, such a contribution is very small since even in control conditions only a small fraction of closures belonging to the short component were missed (~6%). With mastoparan this fraction decreases further but even the maximal decrease observed in patch No. 1 was modest (from 5.3% to 4.0%, Table 2).
This study shows that the current Row through the individual large-conductance Ca’+-activated K + channels is reduced by mastoparan. The effect is evident at micromolar concentrations and appears to be concentration dependent. The lower current flow is not due to lower single channel current amplitudes (masto~~ran is thus not a fast blocker) but rather due to the lower probability of the channel opening. Changes in the channel kinetics associated with lower probability of the open state are complex. Shorter duration of the long compo~~ent of open dwell intervals is compatible with slow btockade, but the duration of the long component of closed dwell intervals instead of being longer becomes
M. I.
680
GLAVINOVN?
et
al.
MASTOPARAN
0
IO
20
30
OPEN
40
INTERVAL
so
60
MASTOPABAAN
CONTROL
1
I
,
5
IO
45
70
(mr)
I 20
OPEN
INTERVAL
L
I
I
5
IO
I5
I 20
(ms)
Fig. 6. Frequency distribution of the channel open times in the absence and in the presence of mastoparan (10 @I) in the solution bathing the intraeehular side of the membrane. Holding potential was +2OmV. Filtered at 4 kHz. The intraecll~at solution had 0.25 PM free Ca2+. (A) One exponential could fit the data in both cases suggesting the pmdominancc of one open state. Note that mastoparan shortens the channel opening time. (a) When two exponcntials were ncecssary to fit the data nnu@oparan shortened both the long (major) and the short (minor) components, but the shortening of the long eomponent was more pronounced.
Table 1. EiTcct of mastoparan on open dwell times (T,,) Experiment No. 1 2 3 4 5 6 *S, short; L, long.
Mastoparan WW 10 5 5 10 10 10
No. of exponcntials
I 1 1
1
Components* L L L L
Open dwell times (ms) Before After 12.2 11.3 12.6 10.4
2
S
1.2
2
L S L
11.9 1.3 12.8
7.0 8.0 9.4 6.3 0.95 6.2 1.0 6.9
Inhibition (“/) 42.6 29.2 25.4 39.4 20.0 47.9 23.0 46.1
Mastoparan
and Cd2+-activated
K + channels
3
i;
E m
.
:c
.
(J.
MASTCPARAN
’
I 2
z
3
i. -
IS CLOSED
TIME
(ma)
‘0 -
!O
\.
I
I
L
0
20
30 CLCSEC
40
TIME
50
60
7c
[ ms)
Fig. 7. Mastoparan shortens the long ciosed intervals, and the short intervals are prolonged but the time-constant of the intermediate component appears unaffected. Frequency distributions of the closed dwell times in the absence and in the presence of mastoparan (IOpM) in the solution bathing the intracellular side of the membrane. Holding potential was OmV. Filtered at 4 kHz. The intracellular solution had 0.25 FM free Cal+. (A) Histograms of all closed intervals. The lines describe the slow component of the data. (B) The difference between the distribution of all closed intervals and the distribution of long closed interval. Two exponentials are clearly present. The lines through the slower component indicate the distribution of intermediate short intervals. The faster component obtained after subtraction of intermediate closed intervals indicates the distribution of short dosed intervals. All lines were best fitted using the least squares method. At least three closed states appear to be present since three exponentials were required to fit the data in each case.
Table 2. Effect of mastoparan Experiment NO.
Mastoparan (PM)
No. of exponentials
I
10
3
2
5
3
3
5
3
4
IO
3
5
10
3
6
10
3
*S, short:
I, intermediate:
L, long.
on closed dwell times (T,)
Components* S I L s I L S I L S I L S I L S I L
Closed dwell times (ms) After Before 1.1 5.5 34.7 1.4 4.7 41.3 1.3 4.3 40.2 1.5 4.8 45.3 1.4 5.1 86.3 1.3 4.1 58.4
1.5 5.7 13.2 1.6 4.5 25.2 I.5 4.0 21.3 1.9 4.4 24.0 1.7 4.8 57.8 1.6 4.5 40.0
Change PO) +36 f4.0 -72 fli -4.0 -39 +l6 -7.0 -32 +27 -8.0 -47 +21 - 6.0 -51 +23 -4.0 -44
682
M. 1. GLAWNWIC
shorter. A reduction in the duration of the short component of the open dwell intervals can be explained by a flickery block. However, a flickery block is expected to lead to a new (short) component in the frequency histogram of closed dwell intervals, and that was not observed. This suggests that mastoparan has a direct effect on the gating mechanism. Comments on a possible role of mastoparan in regulating gating of Ca*+ -activated channels
Evidence, although indirect, which suggests that mastoparan affects the gating of Ca2+-activated K+ channels raises the possibility that the gating mechanism is regulated by a Ca’+-calmodulin and that the mastoparan induced changes of the channel kinetics are due to its effect on calmodulin. This is a plausible proposition although at present we do not know the nature of such a calmodulin regulation of these channels. Furthermore, the mastoparan concentration necessary to inhibit phosphodiesterase was much lower than the concentration necessary to have 50% of the open channel time (50% inhibition was observed at 0.02 PM of mastoparan and we observed that P is 0.5 at 0.25 PM of masto-paran).’ Since the target of regulation must be a component which remains associated when the patch of the membrane is excised this would argue that it is the membrane bound calmodulin that is the agent involved in the regulation of channel gating. Calmodulin was indeed found to be present not only free in the cytoplasm of the chromaffin cell, but also in the membrane-bound form.22 Ca*+-calmodulin was found to induce the activation of a number of different enzymes including phosphatases and protein kinases.’ In addition, phosphorylation of ion channels or their protein subunits has been shown to be a major mechanism in their modulation.*’ Examples include K(Ca)-channels in snail neurons,” sodium channels from various cellular preparations,“,” and voltagesensitive calcium channels from rat skeletal muscle.13 The possibility of a calmodulin regulated K(Ca)-channel has been suggested by studies of putative calmodulin antagonists.3y,4’ Furthermore, a recent report showing that a Paramecium mutant with altered calmodulin has a defective calcium-dependent potassium conductance supports such a conclusion.44 Lack of such an effect in snail neurons could have occurred because of species and/or tissue differences, but also due to different contributions of two types of Ca?+-activated K+ channels (maxi and small). Recent single channel studies have shown that TFP affects the current flow through Ca’+-activated K + channels.3’ It is also known that mastoparan activates Gproteins. *‘,*’The peptide stimulates guanine nucleotide exchange by GTP-binding proteins in a manner similar to that of G-protein-coupled receptors.?” Moreover, ADP-ribosylation by pertussis toxin, which uncouples receptors from G-proteins, selectively inhibits mastoparan activation of G-proteins” as well as mastoparan effects on several cell
et al.
types,2.3X.40.53 Therefore, the effects of mastoparan on Ca* +-activated K + channels described here could also be due to an action of the peptide on a Gprotein, whose activity is coupled to the channel. Guanine nucleotides have been shown to affect the activity of different ion channels.“,’
Calcium and voltage sensitivity
of K(Ca)-channels
Calmodulin has four binding sites for calcium.” If calmodulin is involved in regulating maxi Ca2’-activated K+ channels it is desirable to have at least some evidence about the calcium sensitivity of these channels. This study shows that in bovine chromaffin cell membranes, [Ca2+], required to produce P = 0.5 at OmV (P, probability of the open state) was 0.24.3 FM thus in the same range as the affinity for calcium of calmodulin.24 This is comparable to [Ca’ ‘1, required to half-activate the channels in smooth muscle cells (0.5 PM),’ but is below the value for cultured rat muscle4,36 which was approximately 5 PM, particularly when these channels were incorporated into lipid bilayers. In such a case the concentration required was even higher (50 PM).~‘.~~ However, the calcium sensitivity of channels in other secretory cells is even higher, the comparable concentrations being 0.01 PM (mouse lacrimal glands)16 and 0.02 PM (pig pancreatic acinar cells).31 Thus, among calcium-activated K + channels from secretory cells, the channels from bovine cell membranes are not particularly sensitive to calcium. Gating of maxi K+(Ca*+) channels was found to be voltage dependent, as has been previously reported in rabbit transverse tubule,26 rat pituitary,” mammalian salivary gland,3’ rat muscle,4’ canine airway smooth muscle,32 smooth muscle of rabbit jejunum and guinea-pig mesenteric artery.5 However, unlike most other cases the probability of the open state instead of becoming progressively greater as the membrane becomes more depolarized, decreased at holding potentials above 0 to +20 mV as was reported for the smooth muscle.‘.” CONCLUSION
If calmodulin modulates calcium sensitivity of maxiK(Ca) channels from different cells their very variable Ca2+ -sensitivity would argue that the affinity of calmodulin for Ca2+ varies considerably from one type of cell to another, and that it is modest in bovine chromaffin cell membranes. Alternatively, and more likely, other factors in addition to calmodulin, such as G-proteins, probably influence the calcium sensitivity of maxi channels. Further studies should provide answers to these questions. Acknowledgements-This work
was supported by grants PG-20 (J.M.T.) and MA-9160 (M.I.G.) from the Medical Research Council of Canada. We thank Mrs R. Tang and
Mr D. Clayton
for their technical
assistance.
Mastoparan
and Cal+-activated
K + channels
683
REFERENCES 1. Alvarez J., Garcia-Sancho J. and Herreros B. (1986) The role of calmodulin on Ca’+-dependent K+ transport regulation in the human red cell. Biochem. biophys. Acta 860, 25-34. 2. Argiolas A. and Pisano J. J. (1983) Facilitation of phospholipase A, activity by mastoparans, a new class of mast cell degranulation peptides from wasp venom. J. biol. Chem. 258, 13,697713,702. 3. Barnette M. S., Daly R. and Weiss B. (1983) Inhibition of calmodulin activity by insect venom peptides. Biochem. Pharmac. 32, 2929-2933. 4. Barrett J. N., Magleby K. L. and Pallotta B. S. (1982) Properties of single calcium-activated potassium channels in cultured rat muscle. J. Physiol., Lond. 331, 21 l-230. 5. Benham C. D., Bolton T. B., Lang R. J. and Takewaki T. (1986) Calcium-activated potassium channels in single smooth muscle cells of rabbit jejunum and guinea-pig mesenteric artery. J. Physiol., Lond. 371, 4567. 6. Birnbaumer L., Abramowitz J. and Brown A. M. (1990) Receptor-effector coupling by G proteins. Biochem. bioph>,.v. Acta 1031, 163-224. 7. Blatz L. A. and Magleby K. L. (1986) Single apamin-blocked Ca-activated K+ channels of small conductance in cultured rat skeletal muscle. Nature 323, 718-720. 8. Brown D. A. (1990) G-proteins and potassium currents in neurons. A. Rev. Physiol. 52, 215-242. 9. Cheung W. Y. (1980) Calmodulin plays a pivotal role in cellular regulation. Science 207, 19-27. IO. Colquhoun D. and Sigworth F. J. (1983) Fitting and statistical analysis of single-channel records. In Single-Channel Recording (eds Sakmann B. and Neher E.), pp. 191-263. Plenum Press, New York. I 1. Costa M. R., Casnellie J. E. and Catterall W. A. (1982) Selective phosphorylation of the alpha subunit of the sodium channel by cALMP-dependent protein kinase. J. bioi. Chem. 257, 791Sp7921. 12. Costa M. R. and Catterall W. A. (1984) Phosphorylation in the subunit of the sodium channel by protein kinase C. Cell. molec. Neurobiol. 4, 291-297. 13. Curtis B. M. and Catterall W. A. (1985) Phosphorylation of the calcium antagonist of the voltage-sensitive calcium channel by CAMP-dependant protein kinase. Proc. num. Acad. Sci. C/.S.A. 82, 2528-2532. 14. DeReimer S. A., Kaczmarek L. K., Lai Y., McGuinness T. L. and Greengard P. (1984) Calcium/Calmodulin-dependent protein phosphorylation in the nervous system of Aplysia. J. Neurosci. 4, 1618-1625. 15. dePayer J. E.. Cachelin A. B.. Levitan I. B. and Reuter H. (1982) Ca’+-activated K + conductance in internally perfused snail neurons is enhanced by protein phosphorylation. Proc. natn. Acad. Sci. U.S.A. 79, 4207421 I, 16. Findlay I. (1984) A patch clamp study of potassium channels and whole-cell currents in acinar cells of the mouse lacrimal gland. J. Physiol., Land. 350, 179-195. 17. Glavinovic M. I. (1986) Oscillations of Ca2+ -activated K _ channels around various closed states in bovine chromaffin cells. Sot. Neurosci. Abstr. 12, 1198. IS. Glavinovic M. I. and Trifaro J.-M. (1988) Quinine blockade of currents through Ca’+-activated K+ channels in bovine chromaffin cells. J. Physiol., Lond. 399, 139-152. 19. Hamill 0. P., Marty A., Neher E., Sakmann B. and Sigworth F. J. (1981) Improved patch-clamp technique for high-resolution current recording from cells and cell-free membrane patches. Pfliigers Arch. ges Physiol. 391, 85-100. 20. Higashijima T., Bournier J. and Ross E. M. (1990) Regulation of G, and G, by mastoparan, related amphiphilic peptides, and hydrophobic amines. J. biol. Chem. 265, 14,17614,186. 21. Higashijima T., Uzu S., Nakajima T. and Ross E. M. (1988) Mastoparan, a peptide toxin from wasp venom, mimics receptors by activating GTP-binding regulatory proteins (G-proteins). J. biol. Chem. 263, 64914494. 22. Hikita T., Bader M.-F. and Trifaro J.-M. (1984) Adrenal chromaffin cell calmodulin: its subcellular distribution and binding to chromaffin granule membrane proteins. J. Neurochem. 43, 1087-1097. 23. Joshi A., Trifaro J.-M. and Glavinovic M. I. (1986) Effect of mastoparan on Ca’+ activated K + channels in bovine chromaffin cells. Sot. Neurosci. Abstr. 12, 1076. 24. Klee C. B.. Crouch T. H. and Richman P. G. (1980) Calmodulin. A. Rec. Biochem. 49, 489-515. 25. Kurihara H., Kitajima K., Senda T., Fujita H. and Nakajima T. (1986) Multigranular exocytosis induced by phospholipase AZ-activators, melittin and mastoparan, in rat anterior pituitary cells. Cell Tiss. Res. 243, 31 I-316. 26. Latorre R., Vergara C. and Hidalgo C. (1982) Reconstitution in planar lipid bilayers of a Ca’+-dependent K + channel from transverse tubule membranes isolated from rabbit skeletal muscle. Proc. natn. Acad. Sci. CT.S.A. 79, 805-809. 27. Levitan I. B. (1985) Phosphorylation of ion channels. J. Membrane Biol. 87, 177-190. 28. Levitan E. S. and Levitan I. B. (1986) Apparent loss of calcium-activated potassium current in internally perfused snail neurons is due to accumulation of free intracellular calcium. J. Membrane Biol. 90, 5945. 29. Magleby K. L. and Pallotta B. S. (1983) Calcium dependence of open and shut interval distributions from calcium-activated potassium channels in cultured muscle. J. Physiol., Lond. 344, 585-604. 30. Magleby K. L. and Pallotta B. S. (1983) Burst kinetics of single calcium-activated potassium channels in cultured rat muscle. J. Physiol., Lond. 344, 605423. 3 I. Maruyama Y., Gallacher D. W. and Peterson 0. H. (1983) Voltage and Ca’+-activated K + channel in basolateral acinar cell membranes of mammalian salivary glands. Nature 302, 827-829. 32. McCann J. D. and Welsh M. J. (1987) Neuroleptics antagonize a calcium-activated potassium channel in airway smooth muscle. J. gen. Physiol. 89, 339-352. 33. McManus 0. B., Blatz A. L. and Magleby K. L. (1987) Sampling. log binning, fitting, and plotting durations of open and shut intervals from single channels and the effects of noise. PJtigers Arch. ges Physioi. 410, 530-553. 34. McManus 0. B. and Magleby K. L. (1985) Inverse relationship of the durations of adjacent open and shut intervals for Cl- and K+ channels. Nature 317, 625427. 35. McManus 0. B. and Magleby K. L. (1988) Kinetic states and modes of single large-conductance calcium-activated potassium channels in cultured rat skeletal muscle. J. Physiol., Lond. 402, 79-120. 36. Methfessel G. and Boheim G. (1982) The gating of single calcium-dependent potassium channels is described by an activation/blockade mechanism. Biophys. sfruct. Mech. 9, 3540. 37. Moczydlowski E. and Latorre R. (1981) Gating kinetics of Ca’+-activated potassium channels from rat muscle incorporated into planar lipid bilayers: evidence for two voltage-dependent Ca2+ binding reactions. J. gen. Physiol. 82, 51 l-542.
684
M. I. GLAVINOVI~. ct ul.
38. Mousli M., Brommer C., Bueb J.-L., Tschirhart E., Gies J.-P. and Landry Y. (1989) Activation of rat peritoneal macr cells by substance P and mastoparan. J. Pharmac. exp. Ther. 250, 329-335. 39. Okada Y., Yada T., Ohno-Shosaku T. and Oiki S. (1987) Evidence for the involvement of calmodulin in the operation of Cal+-activated K+ channels in mouse fibroblasts. J. Membrane Biol. 96, 121-128. 40. Ozaki Y., Matsumoto Y.. Yatomi Y., Higashihara M., Kariya T. and Kume S. (1990) Mastoparan, a wasp venom, activates platelets via pertussis toxin-sensitive GTP-binding proteins. Biochem. biophys. Res. Commun. 170, 779 785. 41. Pallotta B. S., Magleby K. L. and Barrett J. N. (1981) Single channel recordings of Ca*+-activated K+ currents in rat muscle cell culture. Nature 293, 471474. 42. Pape L. and Kristensen B. 1. (1984) A calmodulin activated CaLf-dependent K+ channel in human erythrocyte membrane inside-out vesicles. Biochem. biophp. Acfu 770, l-6. 43. Perianin A. and Snyderman R. (1989) Mastoparan, a wasp venom peptide, identifies two discrete mechanisms for evaluating cytosolic calcium and inositol triphosphates in human polymorphonuclear leukocytes. J. Immunol. 143, 1669.-1673. 44. Schaefer W. H., Hinrichsen R. D., Burgess-Casler A., Kung C., Blair 1. A. and Watterson D. M. (1987) A mutant Paramecium with a defective calcium-dependent potassium conductance has an altered calmodulin: a nonlethal selective alteration in calmodulin regulation. Proc,. nafn. Acud. Ski. U.S.A. 84, 3931--3935. 45. Trifaro J. M., Poisner A. M. and Douglas W. W. (1967) The fate of chromaffin granules during catecholamine release from the adrenal medulla. I. Unchanged efflux of phospholipid and cholesterol. Biochem. Pharmacol. 16, 2095.-2100. 46. Trifaro J.-M., Ulpian C. and Preiksaitis H. (1979) Antimyosin stains chromaffin cells. Experientia 34, 1568-1571. 47. Trifaro J.-M., and Lee R. W. H. (1980) Morphological characteristics and stimulus-secretion coupling involving adrenal chromaffin cell cultures. Neuroscience 5, 1433 -1546. into planar 48. Vergara C. and Latorre R. (1983) Kinetics of Ca*+ -activated K+ channels from rabbit muscle incorporated bilayers. J. Physiol., Lond. 82, 543~.568. 49. Vergara C., Moczydlowski E. and Latorre R. (1984) Conduction, blockade and gating in a Ca’ +-activated K ’ channel incorporated into planar lipid bilayer. Biophys. J. 45, 73 -76. 50. Wallace M. A. and Carter H. R. (1989) Effects of wasp venom peptide, mastoparan, on a phosphoinositide-specific phospholipase C purified from rabbit brain membranes. Biochem. biophys. Acta 1006, 31 I-316. 51, Wilson S. P. (1989) Effects of mastoparan on catecholamine release from chromaffin cells. Fedn Eur. biochem. Sots Lerr. 247, 239-24 1. 52. Wong B. S., Lecar H. and Adler M. (1982) Single Ca2+-dependent K+ channels in clonal pituitary cells. Biophys. J. 39, 313-317. 53. Yokokawa N., Komatsu M., Takeda T., Aizawa T. and Yamada T. (1989) Mastoparan, a wasp venom, stimulates insulin release by pancreatic islets through pertussis toxin sensitive GTP-binding protein. Biochem. biophw Res. Commun. 158, 712.-716. (Accepted
14 May 1992)