- Email: [email protected]
Inactivating and non-inactivating outward current channels in cell-attached patches ofHelix neurons
]6
Brain tge~earch. 41)5 i !~?;,~vi I(~ 2" !iGcxic~
BRE 12363
Inactivating and non-inactivating outward current channels in cell-attached patches of Hel& neurons Jeffrey L. Ram 1'2 and Daniel Dagan 1 t Rappaport Family Institute, Faculty of Medicine, Technion, Haifa (Israel) and eDepartment of Physiology, Wayne State University, Detroit, M148201 (U.S.A.) (Accepted 1 July 1986) Key words: Potassium channel; Patch clamping; Inactivation; Voltage-dependent channel; Non-inactivating current; Helix neuron
Two species of inactivating outward current channels and a non-inactivating voltage-dependent current were seen in cell-attached patches of Helix neurons. Large, slowly inactivating channels had a slope conductance of 44 pS as measured with patch pipets containing the normal extracellular ion concentrations, including 4 mM potassium. Latency to maximal opening was 50-220 ms, and the inactivation time constant averaged 350 ms. Channel opening was decreased by preceding depolarization. The channels were selective for potassium and inhibited by 50 mM TEA. Small, quickly inactivating channels were 14 pS and had kinetics and voltage dependence similar to IA. Patch depolarization also activated a non-inactivating voltage-dependent outward current having channel conductance and/or kinetics such that individual channel openings and closings could not be distinguished. Such current was also seen in the presence of 50 mM TEA, but not in the presence of Co2+, characteristics which are similar to outward hydrogen ion currents, described by others in Helix neurons.
INTRODUCTION V o l t a g e - d e p e n d e n t o u t w a r d currents in molluscan neurons flow through a complex combination of ion specific channels, only a few of which have been analyzed with patch clamp methods. In Helix, known outward currents include delayed potassium (IK), Ca2+-activated potassium (1K(Ca)), fast transient potassium (IA), and hydrogen ion 26 currents. Fluctuation analysis methods have indicated that the channels underlying 1K(C~) are 15-20 pS u and that steady state I K channels are only about 2.5 pS 22. Only IK(Ca) has been identified in giga-seal patch recordings; Lux et al. 17 r e p o r t e d a non-inactivating 18 pS channel which had kinetic properties expected of IK(Cal channels. Cottrell et al. 7 have observed a non-inactivating 52 pS channel which they suggest is also calcium-activated. Identification of channels underlying the other outward currents has not been reported. None of the outward channels previously described in Helix exhibited inactivation, a p r o p e r t y of
both I K and IA, and, in some preparations 19 of IK(Ca)" Recently, it has been suggested in several different preparations that I K m a y be d e p e n d e n t on two different channels, one a non-inactivating channel and the other a channel exhibiting t i m e - d e p e n d e n t inactivation 3'~25. As a means of identifying the channels underlying the v o l t a g e - d e p e n d e n t outward currents in Helix, particularly including an identification of channels underlying inactivating currents, giga-seal patch clamp recording was done on ceU-attached patches of Helix neurons. A t least two species of inactivating outward current channels were observed. In addition, a non-inactivating voltage d e p e n d e n t outward current having unresolvable channels was observed. A preliminary r e p o r t of some of these data has been published 21. MATERIALS AND METHODS Experiments were p e r f o r m e d on unidentified 50-150 p m neurons in parietal and pedal ganglia of
Correspondence: J.L. Ram, Department of Physiology, Wayne State University, Detroit, MI 48201, U.S.A. 0006-8993/87/$03.5(I © 1987 Elsevier Science Publishers B.V. (Biomedical Division)
17
• 4
-
-
Normal
4. LOW C I -
"
.ig.K*
.
.../
=
S"@
•
• 0 No*
~///
"
3w r~ v L
2-
(J 90---
1 --
!
i
,
i
r
I
!
i
5O DepolorizQtion
70 ~
I
i
I
I
100
fFom
rest. (mV)
3-2--
55
~
I
0 ,
.~t
, -=:
25
"- J
,
,
'
I
'
0.5
'
1.0
Time(s) 5pA
I ls
Fig. 1. Large, slowly inactivating potassium channels. A: typical recordings of inactivating outward current channels in cell-attached patches of Helix neurons. Numbers at left show amount of patch depolarization from resting potential (about -50 mV when measured in several neurons), produced by hyperpolarizing the patch pipet. B: current-voltage relationship for single channels. The change in current when single channels opened or closed (the obvious current steps in multi-channel records such as those in A) was measured at various levels of depolarization. Solid lines are least squares fits to current-voltage data (left) for the patch in A, having a slope of 39.5 pS, and (right) for a patch in which potassium in the pipet solution was raised to 84 mM by substituting potassium for all of the sodium in the normal pipet solution (high K+), having a slope of 55 pS. Dashed lines show current calculated with the Goldman-HodgkinKatz current equation (for details, see text). Also shown are current-voltage relationships when Tris was substituted for sodium (0 Na ÷) and when chloride was reduced by replacing NaC1 and KCI in the pipet solution with sodium acetate and potassium acetate (low CI-). C: mean number of channels open at various times during 8 1.1 s pulses to 80 mV above rest for the same patch as in A.
Helix aspersa. After microdissection of outer and inner sheaths, ganglia were immersed in 5 mg/ml tryosin (Sigma, Type IX) at 37 °C for 10-40 rain. Experiments were done at ambient temperature (22-24 °C). Except as noted, patch electrodes were polished, coated with Sylgard, and filled with the extracellular solution, containing (in mM) NaCI 80, KC1 4, MgCI 2 5, CaCI 2 7, and Tris-HC1 10 (pH 7.8). Membrane patches were clamped in the cell-attached configuration using a Yale MKV patch clamp amplifier (New Haven, CT) the output of which was filtered at 750 Hz (low pass). Except as noted, patch membrane potential was changed by changing the patch pipet potential, and membrane potentials are given as amounts of depolarization from rest. Data were usually recorded directly on a Gould 220 recorder and hence may be considered heavily filtered (minimal duration opening or closing detectable was 1.5 ms). In some cases (noted in figure captions) better time resolution was obtained either by recording on a T E A C FM recorder (good frequency response up to about 600 Hz) or by catching the record on a Gould OS1420 digital oscilloscope. In either case the recorded data was then played back on to the Gould 220 at slower than real time, to obtain adequate resolution.
RESULTS
Large, slowly inactivating channels As illustrated in Fig. 1A, membrane depolarization opened inactivating outward current channels. Such channels were seen in about one-quarter of more than 100 patches studied to date. Except where stated otherwise statistical statements are based on analysis of 7 representative patches having these channels. Slope conductance for depolarizations of 25-100 mV from resting potential (Fig. 1B) averaged 44 + 11 pS (mean + S.D.). Ionic selectivity of the channels was investigated by varying the ionic composition of solutions in the patch electrode (Fig. 1B). Lowering chloride to 25% of its normal level had no effect on single channel current, indicating that chloride is not the charge carrier. When Tris was substituted for sodium, no shift in the curve occurred; however, when potassium was substituted for sodium, elevating potassium to 84
mM, the single channel I - V curve moved in the depolarizing direction. The position of the single channel I - V curve with normal potassium in the pipet and its shift with high potassium is approximately as predicted by the Goldm a n - H o d g k i n - K a t z current equation for ion selective channels 12. Currents predicted by the Goldm a n - H o d g k i n - K a t z current equation for a potassium selective channel were calculated with ionic permeability, P = 19 × 10 ,4 cm3/s and [K+]i = 88.1 mM 15 and were plotted in Fig. 1B assuming a resting potential o f - 5 0 inV. This particular value for P was chosen empirically by trying several values in the range of 15-25 x 10 -14 cm3/s and selecting the value that appeared to fit the data best by eye. The qualitative features of the curvature at normal potassium and the shift with high potassium is present at all values of P. In normal medium ([K+],, = 4 raM) calculated reversal potential is 28 mV below rest. Since the slowly inactivating channels rarely opened at less than 40 mV above rest it was not possible to get data at potentials close to the reversal potential in normal medium. In high K + medium ([K+]o = 84 mM) the calculated reversal potential is 49 mV above rest, a depolarizing shift close to that observed. Thus, based on the lack of effect of changing ions other than potassium on the I - V curve and the agreement of the effect of changing potassium with theoretical predictions, it is concluded that the major premeant ion of these channels is potassium. Mean delay from onset of membrane depolarization to maximal number of open channels in response to each depolarizing pulse ranged, for different patches, from 50 to 220 ms (mean + S.D. = 105 + 60 ms). Inactivation was often nearly complete within a second (e.g. Fig. IC). Inactivation time constants were determined by counting the number of open channels at 50-100 ms intervals during several repetitions of identical depolarizations and then measuring the interval required for the mean number of open channels to fall to 1/e of its peak value. The mean time constant of inactivation was 350 ms and ranged from 200 to 650 ms. In the example illustrated in Fig. 1C, the inactivation time constant was 300 ms. Similar kinetics of activation and inactivation were also seen in two patches which apparently had only a single channel of this type. Recovery from inactivation produced by a 1.2 s de-
19
A
B
Interval (s) 0.4
1.2"
i
d
1.0"
_ 0.8" IE -,,
#
0.8
0.6" O.4"
1,6
-
-
0
1.'o
0
5 pal
Interval
21o
31o
"
~ 4 9'
(s)
Is Fig. 2. Recovery of slowly inactivating channel from inactivation. Two identical depolarizing pulses (holding potential = resting potential. Clamp potential = 80 mV depolarized from rest. Pulse duration = 1.2 s), separated by various intervals, were given. In all cases the first pulse of a pair was given at least 4.9 s after any preceding pulse. A: sample recordings, showing a range of recovery from very little, at a pulse interval of 0.4 s (top), to almost complete, at a pulse interval of 1.6 s (bottom). B: graph of data from 41 presentations of pulse pairs. MaxJMax 1, for each pulse pair, the maximum number of channels open during the second pulse divided by the maximum number of channels open during the first pulse. For the 41 pulse pairs, the maximum number of channels open during the first pulse averaged 6.1 + 0.8 channels (mean + S.D.).
polarization was analyzed in two patches. Both showed complete recovery within 5 s (illustrated in Fig. 2). Activation of these channels was voltage dependent. Probability of opening rose from very low levels at less than 40 m V depolarization above rest to near maximal levels with depolarizations of 80 mV or more (Fig. 1A). The minimal depolarization at which openings were seen averaged 42 + 8 mV above rest. Depolarization of the holding potential decreased opening probability of these channels. In 4 patches on different cells channel opening in response to t Jot-reel
R
TEA
pulses 80 mV above rest was studied. Depolarization of holding potential to +30 to + 4 0 mV above rest reduced the maximal n u m b e r of channels opening to 25 + 17% of that observed in the same patches with holding potential equal to rest. As expected of a potassium channel, these channels were inhibited by 50 m M t e t r a e t h y l a m m o n i u m ( T E A ) in the pipet. Cells which had inactivating delayed channels were located and then subsequent patches were obtained on the same cells using electrodes that contained either normal Ringer or Ringer containing 50 m M T E A . On 3 cells 6 out of 8 n o r m a l Ringer patches contained inactivating delayed channels. Channels of 4 0 - 5 0 pS were never seen with T E A in the electrode (6 patches), although 3 of these patches contained 25 pS channels (Fig. 3), a channel 0 Ca**
O Ca "* • Ba *+
~pA I ls
Fig. 3. Effect of tetraethylammonium (TEA). Left: channels obtained with normal extracellular medium in patch pipet. Slope conductance = 47 pS. Right: another patch obtained on the same neuron with a patch pipet containing 50 mM TEA (pipet solution made by 1:20 dilution of 1 M TEA with normal pipet solution). Slope conductance = 25 pS. For both traces, holding potential = resting potential; clamp potential = 100 mV depolarized from rest.
~pAI
15
Fig. 4. Responses to depolarizing pulses to 75 mV above resting potential of patches obtained with CaCI2 not included in the pipet solution (left) and with 10 mM BaCI2 used in place of CaCI2 in the pipet solution (right).
2O
size not otherwise obtained with these techniques. The slowly inactivating outward current channels were present when calcium had been left out of the pipet solution (Fig. 4). This was seen in 5 patches on 3 cells, of which two of the cells were shown to have these channels with pipets containing normal calcium. In addition, two patches obtained when calcium had been replaced by barium in the pipet solution "-
also had these channels (e.g. Fig. 4). W h e n pipets
r s for b;d ,q
contained l0 mM cobalt, in place of calcium, the
.
,"s Vr: 3 0
40
48 5 p A I ~ ls
Fig. 5. Patches with large, slowly inactivating inward current. A: slowly inactivating outward channels can be elicited in such patches. Holding potential = resting potential. Clamp potential = 60 mV depolarized from rest. Slope conductance = 55 pS. B: voltage-dependent activation of slowly inactivating inward current in absence of sodium (Tris was substituted for all sodium in the pipet solution).
channels were not seen in 5 patches on 3 cells; however, it was not demonstrated that these cells had the channels in patches obtained with normal calcium. There does not seem to be any particular correlation with the presence of putative calcium currents in the patch. Thus, the slowly inactivating outward current channels were seen in patches that had little apparent voltage d e p e n d e n t inward current (e.g. Fig. 1A) and also where a large, slowly inactivating inward current was present (Fig. 5A). In Fig. 5A inacti-
5 pA
~1
ls
I
t 15pA I 50 mV ls
t 15 pA
12SmV SOres
Fig. 6. Slowly inactivating outward current channels activated by changing either pipet potential or intracellular potential. Measured resting potential of the cell was -50 inV. A: effect of depolarizing patch membrane by hyperpolarizing patch pipet by 70 mV (starting at arrow). B: upper trace: patch pipet current recording. Lower trace: intracellular potential. Patch pipet potential was hyperpolarized, relative to bath potential, by 70 mV throughout. Between arrows the cell was depolarized by passing current through the intracellular electrode, thereby depolarizing the cell body, activating action potentials, and activating inactivating outward current channels. C: same recording as sections of B, with a longer time-base, to illustrate action potential broadening and its relationship to channel activity. The data in B, which had been recorded on an FM recorder, was digitized on a Gould oscilloscope and played back at slower than real time. Numbers are the order of the action potentials in the sequence.
21 vation of outward current channels by a preceding depolarizing pulse occurred even when there was little diminution of the inward current. The slowly inactivating inward current is likely to be calcium since it was seen in some patches where the patch pipet contained no Na ÷ (Fig. 5B) and was not seen when patch pipet solutions did not contain Ca 2÷ (data not shown). The unitary channel current of Helix calcium channels under conditions shown in Fig. 5 is only about 0.4 p A ]6, and multi-channel patches would have been expected to produce the 'noisy current' records such as those we observed. Slow inactivation of outward current channels also occurred when these channels were activated by depolarizing the cell body, rather than by hyperpolarizing the patch pipet (Fig. 6). Fig. 6A shows activation of the channels by patch pipet hyperpolarization. Fig. 6B shows similar channels (slope conductance = 48 pS) activated by passing depolarizing current through an intraceUular electrode in the same cell. Fig. 6C (playback from tape at an expanded time scale of portions of Fig. 6B) illustrates that this cell had action potentials that broadened with repetitive firing. The occurrence of the inactivating channels is correlated with the initial narrow action potentials; the channels are inactivated when the action potentials have become much broader (e.g. action potential number 13).
trated in Fig. 7. Data at the left (time points 26.2, 26.8, and 27.6) show that the transient current could be increased by hyperpolarizing the holding potential. Data at the right (time points 41.3, 41.8, and 42.3) illustrate the voltage dependence of the current with the holding potential 50 m V below rest. Equal size pulses at different holding potentials (time points 42.3 and 43.0) demonstrate that these transient currents were not uncompensated capacitative artifacts. The current steps of the underlying channels could not be distinctly identified but was clearly less than 2 p A at 75 mV above rest (data not shown). In two patches a small enough number of channels with quickly inactivating kinetics enabled analysis of single channel properties. Analysis of one of them is illustrated in Fig. 8. As with the macroscopic currents described above, the channels inactivated completely in less than 100 ms (Fig. 8 (1)), were much 1 Kmetlcs
2 VoPtcge dependence
A V m VH = - 2 5 m V a
4°I~
~¢,-,r~,~l _j2oA 50ms
26.2
,oo, 100.
(mY)-50- 1set
.
26.8
.
41 3
276
.
~
JL
41.Bh
.
4:10
42.3
.
.
.
-
Fig. 7. Fast, transient outward current in a membrane patch. A Vmis the change in membrane potential from rest. Numbers above each trace indicate the number of minutes after the patch giga-seal was obtained. Left: effect of changing holding potential. Right: effect of changing clamp potential,
r
AV ~ VH 0
m [ Vc 75
3
2 5 150 150 - ~
75
75
25
150 150
5 0 75 9 0
Single channel c o n d u c t a n c e
~30
~ 20
I glo
O 10
Small, quickly inactivating channels In approximately 20% of the patches an outward transient current which inactivated in less than 100 ms was present. Except in a few cases to be described below there were either too many channels present or the presence of other channels prevented identification of the underlying single channel openings. Typically, these transient currents appeared as illus-
clcimpV
"~5o ~4o E3o ~ 20
b
~ ~ ¢
E f f e c t of holdmgV
VC - 7 5 m Y
0
0
20
40 60 Time (ms)
•
80
i
~[ os
ps
100 20 40 60 80 AVm(mV)
O0
Fig. 8. Fast transient, outward current channels. 1: sample recordings (a and b) and time course analysis (c) for depolarizations from hold = 25 mV below rest to clamp potential = 75 mV depolarized from rest. In a, there are openings of at least two of the fast channels and possibly an unresolved opening of a third. In b, in addition to rapid openings of at least two of the fast channels, there are long latency openings of the much larger, slowly inactivating channels, also present in this patch, c: average current, as a function of time after depolarization, due to openings of the fast, small channels. Note that the fast channels inactivate completely before the slowly inactivating channels reach peak opening probability. In this patch, opening probability of the slow channels reached its maximum at 300 ms after onset of depolarization and inactivated with a time constant of 350 ms. 2: voltage dependence of fast transient channels. 3: current-voltage relationship for single channels. The change in current when single channels opened or closed (the obvious current steps in multi-channel records such as those in la) was measured at various levels of depolarization. AV M, amount of depolarization from rest; VH, holding potential; Vo clamp potential.
more effectively activated from a hyperpolarized holding potential (Fig. 8 (2)), and could occur in the same patch with the larger slowly, inactivating channels (Fig. 8 (1, b)). The single channel slope conductance was 14 pS (Fig. 8 (3)). The channels in the other patch had similar properties.
:, : ,;',
i
/-
! ,q :~I"
[[ (,wfv'
,I
1:)%
,r~'
Non-inactivating voltage-dependent current Previous fluctuation analysis experiments on Hel& neurons by Reuter and Stevens 22 had indicated the presence of non-inactivating delayed potassium channels having a mean single channel conductance of 2.5 pS. This size of channel would have been too small to be seen as distinct channel openings and closings with the methods used in the present study; however, such channels would be expected to cause a voltage-dependent increase in patch mean conductance as the patch is depolarized. As described below, a voltage-dependent increase in patch mean conductance was observed, but it is unclear whether it should be attributed to the same channels analyzed by Reuter and Stevens 22. To observe non-inactivating current, patch membrane potential was slowly changed at approximately 5 mV/s. During depolarization and repolarization the slope of the current curve (i.e. membrane conductance) increased when the membrane was depolarized more than 50 mV above rest (Fig. 9A). In some patches, inward current was activated during depolarization but was inactive 30 s later, during repolarization (Fig. 9A, left), leaving just the inflection due to activation of outward current. Inward current was probably carried mainly by calcium, since it was eliminated when the pipet solution did not contain calcium (Fig. 9A, right). Fluctuations of the current (seen as an increased width in the baseline) also increased when the patch was depolarized. The amount of voltage dependent non-inactivating outward current ranged, in different patches, from 0 to 10 pA/10 mV, with a median of 0.8 pA/10 mV in a typical set of 10 patches. In response to depolarizing pulses in normal medium a voltage-dependent current with increased fluctuations was elicited by a 70 mV depolarization from a holding potential of +50 mV (Fig. 9B, middle), but not from a holding potential o f - 2 5 mV (Fig, 9B, left). No inactivation of the voltage-dependent current occurred within 800 ms. Inactivation
--
?. . . .
F- ~ [ /
L _ -0- ~ --]
_.
~
I - O - t-
+
I.%0
-'+1oo - - * .5(; -
0 .5()
?:¢v 'trcrT,' ( r e s t ,)
5 p A [ ~ 1$
Fig. 9. Voltage-dependent non-inactivating outward current. A (left): slow membrane depolarization (approximately 5 mV/s) with normal medium in the patch pipet elicited inward current (negative slope region) followed by an outward current when the membrane was depolarized from rest more than 50 mV. The inward current was absent 30 s later during repolarization, while the outward current remained unaltered. A (right): the inward current was absent when the pipet solution did not contain CaCI2. Upper curves show current; lower curves show amount of depolarization of patch membrane potential from rest. Vertical lines provide a guide between the voltage records and the inflection points on the current curves. B: responses to depolarizing pulses. This patch also contained a single slowly inactivating channel (maximum number ever opened while patch was held), one opening of which is seen at right. C: continuous record of responses to repeated 80 mV depolarizing pulses of same patch as in Fig. 1A. In B and C shading shows current attributable to leakage (proportional to current elicited by a 25 mV hyperpolarization from rest).
also did not occur when the potential was stepped from -25 mV below rest to + 130 mV above rest (Fig. 9B, right). With repeated depolarizations the amount of 'non-inactivating' current decreased somewhat (Fig. 9C). This decrease in current was always small compared to inactivation of the large slowly inactivating channels, such as occurred in the same patch in Fig. 9C. Experiments with T E A and cobalt in the patch pipet were done in order to test whether the non-inactivating voltage-dependent current had properties expected of a delayed potassium current. Fifty mM T E A in the patch pipet did not block non-inactivating
•
23
A
B
O Co+++ 50 TEA
0 Ca** ÷10 Co÷~
2pA
12pA
50 mV
150mY / m
lOs
Fig. 10. Effects of ions and tetraethylammonium (TEA) on non-inactivating voltage-dependent current. Upper curves, current; lower curves, amount of depolarization of patch membrane from rest. A: 0 Ca2÷ + 50 TEA: pipet solution did not contain CaCI2and was diluted 1:20 with 1 M TEA. B: 0 Ca2÷ + 10 Co2÷: 10 mM CoCI2 was used in place of CaCI2 in the pipet solution.
voltage-dependent current (Fig. 10A). The range of currents obtained with and without T E A (e.g. compare Fig. 9A right and Fig. 10A) overlap (for 7 T E A patches, range is 0 to 5 pA/10 mV; median is 0.7 pA/10 mV). When 10 mM CoC12 was substituted for CaCI 2 in the pipet solution, the non-inactivating current was not present; in fact, few channels or currents other than a linear 'leakage' current were present (Fig. 10B). DISCUSSION The properties of unitary channels underlying inactivating voltage-dependent outward currents in Helix neurons have not previously been described. This paper provides data on two types of inactivating outward current channels, which we will attempt here to relate to whole cell currents. In addition noninactivating voltage-dependent current, carried by channels clearly distinct from the two species of inactivating channels, has been observed, which we will also attempt to relate to whole cell currents. First, we consider the small, quickly inactivating channels, because their relationship to whole cell currents seems most obvious. These channels probably underlie the fast transient current 19, also known as IA4. Properties of I A that appear to be similar to the channels described here include (a) I A usually requires a preceding hyperpolarization below resting
potential in order to be elicited; (b) I A reaches peak activation within 15 ms of depolarization; and (c) the inactivation time constant for IA is less than 150 ms. These properties not only have clear correlations with characteristics of the quickly inactivating channels described here, but also serve to distinguish IA clearly from the slowly inactivating channels reported in this paper. Other laboratories have also recently described putative I A channels. We report here a conductance of 14 pS. Other reported conductances include a smaller size in Drosophila neurons 24, 20 pS in cultured dorsal root ganglion neurons 13, and 22 pS in cultured nodose ganglia neurons 23. The slowly inactivating channels clearly underlie a potassium current, but it is presently uncertain whether these channels are delayed potassium or calcium-dependent potassium channels. It has been proposed, based on whole cell current kinetics and pharmacology 3'5"25, that delayed potassium current may utilize separate populations of channels for the inactivating and non-inactivating fractions of the total outward current. Several different sizes of delayed potassium channels have been seen in single channel studies 14. A fluctuation analysis model for flog node predicts a unitary conductance for the inactivating channel of 40-70 pS 5, in agreement with the channels reported here. It should be pointed out, however, that the frog node study was done in the presence of calcium and the inactivating channels could conceivably be calcium-activated. Arguments against the slowly inactivating channels reported here being calcium activated include the following: (a) no correlation with slow inward current (compare Figs. 1 and 5); (b) present when no CaCl 2 was used in pipet solution and when barium was substituted for calcium (Fig. 4); (c) when elicited by cell depolarization the inactivating channels are active when action potentials are relatively narrow, and have inactivated when the action potentials are broad, indicative of larger calcium current (Fig. 6); and (d) calcium-activated channels, studied in Helix in three other laboratories by both fluctuation analysis methods 11 and single channel methods 9A7, have an estimated single channel conductance of 15-20 pS, less than half the conductance of channels described here, and have been reported not to inactivate in Helix ~7. In unpublished studies from our laboratory 18 pS channels have been seen in Helix neurons and may
24 correspond to these calcium-activated channels. On the other hand, these data do not conclusively exclude calcium dependence of the slowly inactivating channels. Since the ~0 calcium' solution was made by leaving CaCI2 out of the pipet solution, it probably contained a small amount of free calcium (according to B. Frankenhaeuser m. between 2 × 1()-s M and 10-~ M). The time course of inactivation of putative calcium current (Fig. 5 and ref. 16) resembles that of the inactivation of the slowly inactivating channels. The observation by 3 laboratories of calcium-dependent channels of 15-20 pS does not exclude a species of 45 pS from also being calcium-dependent. Non-inactivating 52 pS channels in Helix neuron F2 have been suggested on indirect evidence to be calcium-dependent 7. Whether these channels are indeed calciumdependent and whether they may be another mode of activity of the slowly inactivating channels described here is a question for future research. Non-inactivating voltage-dependent current was looked for initially to corroborate predictions based on fluctuation analysis of delayed potassium current by Reuter and Stevens 22. Although the expected current was seen, it had pharmacological properties that suggest that all or part of the current in the present study may be distinguishable from delayed potassium current. In particular, the current was not blocked by 50 mM TEA, a concentration which Meech and Standen TM show will block 80% of the delayed potassium current, and it was blocked by 10 mM cobalt in the electrodes. These properties are similar to those described by Thomas and Meech 26 and Byerly et al. 2 for voltage-dependent hydrogen ion currents in snail neurons. In some neurons hydrogen ion currents may therefore play a larger role than previously suggested. Further studies on the pH dependence, tail current reversal potential, and pharmacology of
REFERENCES 1 Aldrich, R.W., Jr., Getting, P.A. and Thompson, S.H., Mechanisms of frequency-dependent broadening of molluscan neurone soma spikes, J. Physiol. (London), 291 (1979) 531-544. 2 Byerly, L., Meech, R. and Moody, W., Jr., Rapidly activating hydrogen ion currents in perfused neurones of the snail, L ymnaea stagnalis, J. Physiol. (London), 351 (1984) 199-216. 3 Chabala, L.D., The kinetics of recovery and development of potassium channel inactivation in perfused squid (Loligo
these voltage-dependent currents must be done to test this suggestion. Whatever the exact ion dependence and selectivity of the channels and currents described here, an evident conclusion from this paper is that in Helix inactivating outward currents are carried by distinguishable and much larger channels than the non-inactivating currents. Fig. 9C illustrates the integration of the two types of channels, as seen in a number of patches. The function of the inactivating portion of the outward current has an important role in allowing action potential broadening in cells with calcium currents 1. As outward current channels inactivate, there is a smaller repolarization current, thereby leading to longer duration action potentials, consequent greater activation of calcium channels, and broadening of the calcium-dependent shoulder. This conclusion would be true whether the inactivating channels were calcium dependent or not. This model is fully consistent with inactivation of the large outward channels in Fig. 6 as action potentials became broader. Utilization of separate channels for the inactivating portion of the outward current may make modulation of its function less complex.
ACKNOWLEDGEMENTS We thank the Lady Davis Trust for financial assistance to J.L.R. while he was on sabbatical leave as a Visiting Associate Professor at the Technion. We thank Anna Levi for technical assistance, Boaz Gillo, Douglas Ewald, and Steven Siegelbaum for helpful discussions and advice on patch clamping technique, Yoram Palti for comments on an earlier draft of this paper, and Kibbutz Ma'ayan Zwi for use of an Apple lie computer.
pealie) giant axons, J. Physiol. (London), 356 (1984) 193-220. 4 Connor, J.A. and Stevens, C.F., Voltage clamp studies of a transient outward membrane current in gastropod neural somata, J. Physiol. (London), 2!3 (1971) 21-30. 5 Conti, F., Hille, B. and Nonner, W., Non-stationary fluctuations of the potassium conductance at the node of Ranvier of the frog, J. Physiol. (London), 353 (1984) 199-230. 6 Conti, F. and Neher, E., Single channel recordings of K÷ currents in squid axons, Nature (London), 285 (1980) 140-143. 7 Cottrell, G.A., Davies, N.W. and Green, K.A., Multiple
25 actions of a molluscan cardioexcitatory neuropeptide and related peptides on identified Helix neurones, J. Physiol. (London), 356 (1984) 315-333. 8 Dubois, J.M., Evidence for the existence of three types of potassium channels in the frog Ranvier node membrane, J. Physiol. (London), 318 (1981) 297-316. 9 Ewald, D.A., Williams, A. and Levitan, I.B., Modulation of single Ca2+-dependent K + channel activity by protein phosphorylation, Nature (London), 315 (1985) 503-506. 10 Frankenhaeuser, B., The effect of calcium on the myelinated nerve, J. Physiol. (London), 137 (1957) 245-260. 11 Hermann, A. and Hartung, K., Noise and relaxation measurements of the Ca 2+ activated K + current in Helix neurones, Pfliigers Arch., 393 (1982) 254-261. 12 Hille, B., Single Channels in Excitable Cells, Sinauer Assoc., Sunderland, MA, 1984, pp. 230-236. 13 Kasai, H., Single transient potassium currents in mammalian sensory neurons studied using patch clamp techniques, Soc. Neurosci. Abstr., 11 (1985) 955. 14 Llano, I. and Bezanilla, F., Two types of potassium channels in the cut-open squid giant axon, Biophys. J., 47 (1985) 221a. 15 Leefmans, F.J. and t3amino, S.M., Intracellular free potassium concentration in snail neurones measured with ion-selective microelectrodes, Soc. Neurosci. Abstr., 8 (1982) 684. 16 Lux, H.D. and Brown, A.M., Patch and whole cell calcium current recorded simultaneously in snail neurones, J. Gen. Physiol., 83 (1984) 727-750. 17 Lux, H.D., Neher, E. and Marty, A., Single channel activity associated with the calcium dependent outward current
in Helix pomatia, Pfliigers Arch., 389 (1981) 293-295. 18 Meech, R.W. and Standen, N.B., Potassium activation in Helix aspersa neurones under voltage clamp: a component mediated by calcium influx, J. Physiol. (London), 249 (1975) 211-239. 19 Neher, F., Two fast transient current components during voltage clamp in snail neurons, J. Gen. Physiol., 58 (1971) 36-53. 20 Pallotta, B.S., Calcium-activated potassium channels in rat muscle inactivate from a short duration open state, J. Physiol. (London), 363 (1985) 501-516. 21 Ram, J.L. and Dagan, D., Two types of delayed potassium channels: inactivating and non-inactivating channels, Soc. Neurosci. Abstr., 11 (1985) 956. 22 Reuter, H. and Stevens, C.F., Ion conductance and ion selectivity of potassium channels in snail neurons, J. Membr. Biol., 57 (1980) 103-118. 23 Shrier, A. and Cooper, E., Single channel analysis of I a currents from mammalian sensory neurons, Biophys. J., 47 (1985) 385a. 24 Solc, C.K. and Aldrich, R.W., Patch clamp analysis of inactivating outward currents in dissociated CNS neurons of wild-type and shaker Drosophila, Soc. Neurosci. Abstr., 11 (1985) 954. 25 Strong, J.A. and Kaczmarek, L.K., Modulation of multiple potassium currents in dialyzed bag cell neurons of Aplysia, Soc. Neurosci. Abstr., 10 (1984) 1101. 26 Thomas, R.C. and Meech, R.W., Hydrogen ion currents and intracellular pH in depolarized voltage-clamped snail neurones, Nature (London), 299 (1982) 826-828.
Brain tge~earch. 41)5 i !~?;,~vi I(~ 2" !iGcxic~
BRE 12363
Inactivating and non-inactivating outward current channels in cell-attached patches of Hel& neurons Jeffrey L. Ram 1'2 and Daniel Dagan 1 t Rappaport Family Institute, Faculty of Medicine, Technion, Haifa (Israel) and eDepartment of Physiology, Wayne State University, Detroit, M148201 (U.S.A.) (Accepted 1 July 1986) Key words: Potassium channel; Patch clamping; Inactivation; Voltage-dependent channel; Non-inactivating current; Helix neuron
Two species of inactivating outward current channels and a non-inactivating voltage-dependent current were seen in cell-attached patches of Helix neurons. Large, slowly inactivating channels had a slope conductance of 44 pS as measured with patch pipets containing the normal extracellular ion concentrations, including 4 mM potassium. Latency to maximal opening was 50-220 ms, and the inactivation time constant averaged 350 ms. Channel opening was decreased by preceding depolarization. The channels were selective for potassium and inhibited by 50 mM TEA. Small, quickly inactivating channels were 14 pS and had kinetics and voltage dependence similar to IA. Patch depolarization also activated a non-inactivating voltage-dependent outward current having channel conductance and/or kinetics such that individual channel openings and closings could not be distinguished. Such current was also seen in the presence of 50 mM TEA, but not in the presence of Co2+, characteristics which are similar to outward hydrogen ion currents, described by others in Helix neurons.
INTRODUCTION V o l t a g e - d e p e n d e n t o u t w a r d currents in molluscan neurons flow through a complex combination of ion specific channels, only a few of which have been analyzed with patch clamp methods. In Helix, known outward currents include delayed potassium (IK), Ca2+-activated potassium (1K(Ca)), fast transient potassium (IA), and hydrogen ion 26 currents. Fluctuation analysis methods have indicated that the channels underlying 1K(C~) are 15-20 pS u and that steady state I K channels are only about 2.5 pS 22. Only IK(Ca) has been identified in giga-seal patch recordings; Lux et al. 17 r e p o r t e d a non-inactivating 18 pS channel which had kinetic properties expected of IK(Cal channels. Cottrell et al. 7 have observed a non-inactivating 52 pS channel which they suggest is also calcium-activated. Identification of channels underlying the other outward currents has not been reported. None of the outward channels previously described in Helix exhibited inactivation, a p r o p e r t y of
both I K and IA, and, in some preparations 19 of IK(Ca)" Recently, it has been suggested in several different preparations that I K m a y be d e p e n d e n t on two different channels, one a non-inactivating channel and the other a channel exhibiting t i m e - d e p e n d e n t inactivation 3'~25. As a means of identifying the channels underlying the v o l t a g e - d e p e n d e n t outward currents in Helix, particularly including an identification of channels underlying inactivating currents, giga-seal patch clamp recording was done on ceU-attached patches of Helix neurons. A t least two species of inactivating outward current channels were observed. In addition, a non-inactivating voltage d e p e n d e n t outward current having unresolvable channels was observed. A preliminary r e p o r t of some of these data has been published 21. MATERIALS AND METHODS Experiments were p e r f o r m e d on unidentified 50-150 p m neurons in parietal and pedal ganglia of
Correspondence: J.L. Ram, Department of Physiology, Wayne State University, Detroit, MI 48201, U.S.A. 0006-8993/87/$03.5(I © 1987 Elsevier Science Publishers B.V. (Biomedical Division)
17
• 4
-
-
Normal
4. LOW C I -
"
.ig.K*
.
.../
=
S"@
•
• 0 No*
~///
"
3w r~ v L
2-
(J 90---
1 --
!
i
,
i
r
I
!
i
5O DepolorizQtion
70 ~
I
i
I
I
100
fFom
rest. (mV)
3-2--
55
~
I
0 ,
.~t
, -=:
25
"- J
,
,
'
I
'
0.5
'
1.0
Time(s) 5pA
I ls
Fig. 1. Large, slowly inactivating potassium channels. A: typical recordings of inactivating outward current channels in cell-attached patches of Helix neurons. Numbers at left show amount of patch depolarization from resting potential (about -50 mV when measured in several neurons), produced by hyperpolarizing the patch pipet. B: current-voltage relationship for single channels. The change in current when single channels opened or closed (the obvious current steps in multi-channel records such as those in A) was measured at various levels of depolarization. Solid lines are least squares fits to current-voltage data (left) for the patch in A, having a slope of 39.5 pS, and (right) for a patch in which potassium in the pipet solution was raised to 84 mM by substituting potassium for all of the sodium in the normal pipet solution (high K+), having a slope of 55 pS. Dashed lines show current calculated with the Goldman-HodgkinKatz current equation (for details, see text). Also shown are current-voltage relationships when Tris was substituted for sodium (0 Na ÷) and when chloride was reduced by replacing NaC1 and KCI in the pipet solution with sodium acetate and potassium acetate (low CI-). C: mean number of channels open at various times during 8 1.1 s pulses to 80 mV above rest for the same patch as in A.
Helix aspersa. After microdissection of outer and inner sheaths, ganglia were immersed in 5 mg/ml tryosin (Sigma, Type IX) at 37 °C for 10-40 rain. Experiments were done at ambient temperature (22-24 °C). Except as noted, patch electrodes were polished, coated with Sylgard, and filled with the extracellular solution, containing (in mM) NaCI 80, KC1 4, MgCI 2 5, CaCI 2 7, and Tris-HC1 10 (pH 7.8). Membrane patches were clamped in the cell-attached configuration using a Yale MKV patch clamp amplifier (New Haven, CT) the output of which was filtered at 750 Hz (low pass). Except as noted, patch membrane potential was changed by changing the patch pipet potential, and membrane potentials are given as amounts of depolarization from rest. Data were usually recorded directly on a Gould 220 recorder and hence may be considered heavily filtered (minimal duration opening or closing detectable was 1.5 ms). In some cases (noted in figure captions) better time resolution was obtained either by recording on a T E A C FM recorder (good frequency response up to about 600 Hz) or by catching the record on a Gould OS1420 digital oscilloscope. In either case the recorded data was then played back on to the Gould 220 at slower than real time, to obtain adequate resolution.
RESULTS
Large, slowly inactivating channels As illustrated in Fig. 1A, membrane depolarization opened inactivating outward current channels. Such channels were seen in about one-quarter of more than 100 patches studied to date. Except where stated otherwise statistical statements are based on analysis of 7 representative patches having these channels. Slope conductance for depolarizations of 25-100 mV from resting potential (Fig. 1B) averaged 44 + 11 pS (mean + S.D.). Ionic selectivity of the channels was investigated by varying the ionic composition of solutions in the patch electrode (Fig. 1B). Lowering chloride to 25% of its normal level had no effect on single channel current, indicating that chloride is not the charge carrier. When Tris was substituted for sodium, no shift in the curve occurred; however, when potassium was substituted for sodium, elevating potassium to 84
mM, the single channel I - V curve moved in the depolarizing direction. The position of the single channel I - V curve with normal potassium in the pipet and its shift with high potassium is approximately as predicted by the Goldm a n - H o d g k i n - K a t z current equation for ion selective channels 12. Currents predicted by the Goldm a n - H o d g k i n - K a t z current equation for a potassium selective channel were calculated with ionic permeability, P = 19 × 10 ,4 cm3/s and [K+]i = 88.1 mM 15 and were plotted in Fig. 1B assuming a resting potential o f - 5 0 inV. This particular value for P was chosen empirically by trying several values in the range of 15-25 x 10 -14 cm3/s and selecting the value that appeared to fit the data best by eye. The qualitative features of the curvature at normal potassium and the shift with high potassium is present at all values of P. In normal medium ([K+],, = 4 raM) calculated reversal potential is 28 mV below rest. Since the slowly inactivating channels rarely opened at less than 40 mV above rest it was not possible to get data at potentials close to the reversal potential in normal medium. In high K + medium ([K+]o = 84 mM) the calculated reversal potential is 49 mV above rest, a depolarizing shift close to that observed. Thus, based on the lack of effect of changing ions other than potassium on the I - V curve and the agreement of the effect of changing potassium with theoretical predictions, it is concluded that the major premeant ion of these channels is potassium. Mean delay from onset of membrane depolarization to maximal number of open channels in response to each depolarizing pulse ranged, for different patches, from 50 to 220 ms (mean + S.D. = 105 + 60 ms). Inactivation was often nearly complete within a second (e.g. Fig. IC). Inactivation time constants were determined by counting the number of open channels at 50-100 ms intervals during several repetitions of identical depolarizations and then measuring the interval required for the mean number of open channels to fall to 1/e of its peak value. The mean time constant of inactivation was 350 ms and ranged from 200 to 650 ms. In the example illustrated in Fig. 1C, the inactivation time constant was 300 ms. Similar kinetics of activation and inactivation were also seen in two patches which apparently had only a single channel of this type. Recovery from inactivation produced by a 1.2 s de-
19
A
B
Interval (s) 0.4
1.2"
i
d
1.0"
_ 0.8" IE -,,
#
0.8
0.6" O.4"
1,6
-
-
0
1.'o
0
5 pal
Interval
21o
31o
"
~ 4 9'
(s)
Is Fig. 2. Recovery of slowly inactivating channel from inactivation. Two identical depolarizing pulses (holding potential = resting potential. Clamp potential = 80 mV depolarized from rest. Pulse duration = 1.2 s), separated by various intervals, were given. In all cases the first pulse of a pair was given at least 4.9 s after any preceding pulse. A: sample recordings, showing a range of recovery from very little, at a pulse interval of 0.4 s (top), to almost complete, at a pulse interval of 1.6 s (bottom). B: graph of data from 41 presentations of pulse pairs. MaxJMax 1, for each pulse pair, the maximum number of channels open during the second pulse divided by the maximum number of channels open during the first pulse. For the 41 pulse pairs, the maximum number of channels open during the first pulse averaged 6.1 + 0.8 channels (mean + S.D.).
polarization was analyzed in two patches. Both showed complete recovery within 5 s (illustrated in Fig. 2). Activation of these channels was voltage dependent. Probability of opening rose from very low levels at less than 40 m V depolarization above rest to near maximal levels with depolarizations of 80 mV or more (Fig. 1A). The minimal depolarization at which openings were seen averaged 42 + 8 mV above rest. Depolarization of the holding potential decreased opening probability of these channels. In 4 patches on different cells channel opening in response to t Jot-reel
R
TEA
pulses 80 mV above rest was studied. Depolarization of holding potential to +30 to + 4 0 mV above rest reduced the maximal n u m b e r of channels opening to 25 + 17% of that observed in the same patches with holding potential equal to rest. As expected of a potassium channel, these channels were inhibited by 50 m M t e t r a e t h y l a m m o n i u m ( T E A ) in the pipet. Cells which had inactivating delayed channels were located and then subsequent patches were obtained on the same cells using electrodes that contained either normal Ringer or Ringer containing 50 m M T E A . On 3 cells 6 out of 8 n o r m a l Ringer patches contained inactivating delayed channels. Channels of 4 0 - 5 0 pS were never seen with T E A in the electrode (6 patches), although 3 of these patches contained 25 pS channels (Fig. 3), a channel 0 Ca**
O Ca "* • Ba *+
~pA I ls
Fig. 3. Effect of tetraethylammonium (TEA). Left: channels obtained with normal extracellular medium in patch pipet. Slope conductance = 47 pS. Right: another patch obtained on the same neuron with a patch pipet containing 50 mM TEA (pipet solution made by 1:20 dilution of 1 M TEA with normal pipet solution). Slope conductance = 25 pS. For both traces, holding potential = resting potential; clamp potential = 100 mV depolarized from rest.
~pAI
15
Fig. 4. Responses to depolarizing pulses to 75 mV above resting potential of patches obtained with CaCI2 not included in the pipet solution (left) and with 10 mM BaCI2 used in place of CaCI2 in the pipet solution (right).
2O
size not otherwise obtained with these techniques. The slowly inactivating outward current channels were present when calcium had been left out of the pipet solution (Fig. 4). This was seen in 5 patches on 3 cells, of which two of the cells were shown to have these channels with pipets containing normal calcium. In addition, two patches obtained when calcium had been replaced by barium in the pipet solution "-
also had these channels (e.g. Fig. 4). W h e n pipets
r s for b;d ,q
contained l0 mM cobalt, in place of calcium, the
.
,"s Vr: 3 0
40
48 5 p A I ~ ls
Fig. 5. Patches with large, slowly inactivating inward current. A: slowly inactivating outward channels can be elicited in such patches. Holding potential = resting potential. Clamp potential = 60 mV depolarized from rest. Slope conductance = 55 pS. B: voltage-dependent activation of slowly inactivating inward current in absence of sodium (Tris was substituted for all sodium in the pipet solution).
channels were not seen in 5 patches on 3 cells; however, it was not demonstrated that these cells had the channels in patches obtained with normal calcium. There does not seem to be any particular correlation with the presence of putative calcium currents in the patch. Thus, the slowly inactivating outward current channels were seen in patches that had little apparent voltage d e p e n d e n t inward current (e.g. Fig. 1A) and also where a large, slowly inactivating inward current was present (Fig. 5A). In Fig. 5A inacti-
5 pA
~1
ls
I
t 15pA I 50 mV ls
t 15 pA
12SmV SOres
Fig. 6. Slowly inactivating outward current channels activated by changing either pipet potential or intracellular potential. Measured resting potential of the cell was -50 inV. A: effect of depolarizing patch membrane by hyperpolarizing patch pipet by 70 mV (starting at arrow). B: upper trace: patch pipet current recording. Lower trace: intracellular potential. Patch pipet potential was hyperpolarized, relative to bath potential, by 70 mV throughout. Between arrows the cell was depolarized by passing current through the intracellular electrode, thereby depolarizing the cell body, activating action potentials, and activating inactivating outward current channels. C: same recording as sections of B, with a longer time-base, to illustrate action potential broadening and its relationship to channel activity. The data in B, which had been recorded on an FM recorder, was digitized on a Gould oscilloscope and played back at slower than real time. Numbers are the order of the action potentials in the sequence.
21 vation of outward current channels by a preceding depolarizing pulse occurred even when there was little diminution of the inward current. The slowly inactivating inward current is likely to be calcium since it was seen in some patches where the patch pipet contained no Na ÷ (Fig. 5B) and was not seen when patch pipet solutions did not contain Ca 2÷ (data not shown). The unitary channel current of Helix calcium channels under conditions shown in Fig. 5 is only about 0.4 p A ]6, and multi-channel patches would have been expected to produce the 'noisy current' records such as those we observed. Slow inactivation of outward current channels also occurred when these channels were activated by depolarizing the cell body, rather than by hyperpolarizing the patch pipet (Fig. 6). Fig. 6A shows activation of the channels by patch pipet hyperpolarization. Fig. 6B shows similar channels (slope conductance = 48 pS) activated by passing depolarizing current through an intraceUular electrode in the same cell. Fig. 6C (playback from tape at an expanded time scale of portions of Fig. 6B) illustrates that this cell had action potentials that broadened with repetitive firing. The occurrence of the inactivating channels is correlated with the initial narrow action potentials; the channels are inactivated when the action potentials have become much broader (e.g. action potential number 13).
trated in Fig. 7. Data at the left (time points 26.2, 26.8, and 27.6) show that the transient current could be increased by hyperpolarizing the holding potential. Data at the right (time points 41.3, 41.8, and 42.3) illustrate the voltage dependence of the current with the holding potential 50 m V below rest. Equal size pulses at different holding potentials (time points 42.3 and 43.0) demonstrate that these transient currents were not uncompensated capacitative artifacts. The current steps of the underlying channels could not be distinctly identified but was clearly less than 2 p A at 75 mV above rest (data not shown). In two patches a small enough number of channels with quickly inactivating kinetics enabled analysis of single channel properties. Analysis of one of them is illustrated in Fig. 8. As with the macroscopic currents described above, the channels inactivated completely in less than 100 ms (Fig. 8 (1)), were much 1 Kmetlcs
2 VoPtcge dependence
A V m VH = - 2 5 m V a
4°I~
~¢,-,r~,~l _j2oA 50ms
26.2
,oo, 100.
(mY)-50- 1set
.
26.8
.
41 3
276
.
~
JL
41.Bh
.
4:10
42.3
.
.
.
-
Fig. 7. Fast, transient outward current in a membrane patch. A Vmis the change in membrane potential from rest. Numbers above each trace indicate the number of minutes after the patch giga-seal was obtained. Left: effect of changing holding potential. Right: effect of changing clamp potential,
r
AV ~ VH 0
m [ Vc 75
3
2 5 150 150 - ~
75
75
25
150 150
5 0 75 9 0
Single channel c o n d u c t a n c e
~30
~ 20
I glo
O 10
Small, quickly inactivating channels In approximately 20% of the patches an outward transient current which inactivated in less than 100 ms was present. Except in a few cases to be described below there were either too many channels present or the presence of other channels prevented identification of the underlying single channel openings. Typically, these transient currents appeared as illus-
clcimpV
"~5o ~4o E3o ~ 20
b
~ ~ ¢
E f f e c t of holdmgV
VC - 7 5 m Y
0
0
20
40 60 Time (ms)
•
80
i
~[ os
ps
100 20 40 60 80 AVm(mV)
O0
Fig. 8. Fast transient, outward current channels. 1: sample recordings (a and b) and time course analysis (c) for depolarizations from hold = 25 mV below rest to clamp potential = 75 mV depolarized from rest. In a, there are openings of at least two of the fast channels and possibly an unresolved opening of a third. In b, in addition to rapid openings of at least two of the fast channels, there are long latency openings of the much larger, slowly inactivating channels, also present in this patch, c: average current, as a function of time after depolarization, due to openings of the fast, small channels. Note that the fast channels inactivate completely before the slowly inactivating channels reach peak opening probability. In this patch, opening probability of the slow channels reached its maximum at 300 ms after onset of depolarization and inactivated with a time constant of 350 ms. 2: voltage dependence of fast transient channels. 3: current-voltage relationship for single channels. The change in current when single channels opened or closed (the obvious current steps in multi-channel records such as those in la) was measured at various levels of depolarization. AV M, amount of depolarization from rest; VH, holding potential; Vo clamp potential.
more effectively activated from a hyperpolarized holding potential (Fig. 8 (2)), and could occur in the same patch with the larger slowly, inactivating channels (Fig. 8 (1, b)). The single channel slope conductance was 14 pS (Fig. 8 (3)). The channels in the other patch had similar properties.
:, : ,;',
i
/-
! ,q :~I"
[[ (,wfv'
,I
1:)%
,r~'
Non-inactivating voltage-dependent current Previous fluctuation analysis experiments on Hel& neurons by Reuter and Stevens 22 had indicated the presence of non-inactivating delayed potassium channels having a mean single channel conductance of 2.5 pS. This size of channel would have been too small to be seen as distinct channel openings and closings with the methods used in the present study; however, such channels would be expected to cause a voltage-dependent increase in patch mean conductance as the patch is depolarized. As described below, a voltage-dependent increase in patch mean conductance was observed, but it is unclear whether it should be attributed to the same channels analyzed by Reuter and Stevens 22. To observe non-inactivating current, patch membrane potential was slowly changed at approximately 5 mV/s. During depolarization and repolarization the slope of the current curve (i.e. membrane conductance) increased when the membrane was depolarized more than 50 mV above rest (Fig. 9A). In some patches, inward current was activated during depolarization but was inactive 30 s later, during repolarization (Fig. 9A, left), leaving just the inflection due to activation of outward current. Inward current was probably carried mainly by calcium, since it was eliminated when the pipet solution did not contain calcium (Fig. 9A, right). Fluctuations of the current (seen as an increased width in the baseline) also increased when the patch was depolarized. The amount of voltage dependent non-inactivating outward current ranged, in different patches, from 0 to 10 pA/10 mV, with a median of 0.8 pA/10 mV in a typical set of 10 patches. In response to depolarizing pulses in normal medium a voltage-dependent current with increased fluctuations was elicited by a 70 mV depolarization from a holding potential of +50 mV (Fig. 9B, middle), but not from a holding potential o f - 2 5 mV (Fig, 9B, left). No inactivation of the voltage-dependent current occurred within 800 ms. Inactivation
--
?. . . .
F- ~ [ /
L _ -0- ~ --]
_.
~
I - O - t-
+
I.%0
-'+1oo - - * .5(; -
0 .5()
?:¢v 'trcrT,' ( r e s t ,)
5 p A [ ~ 1$
Fig. 9. Voltage-dependent non-inactivating outward current. A (left): slow membrane depolarization (approximately 5 mV/s) with normal medium in the patch pipet elicited inward current (negative slope region) followed by an outward current when the membrane was depolarized from rest more than 50 mV. The inward current was absent 30 s later during repolarization, while the outward current remained unaltered. A (right): the inward current was absent when the pipet solution did not contain CaCI2. Upper curves show current; lower curves show amount of depolarization of patch membrane potential from rest. Vertical lines provide a guide between the voltage records and the inflection points on the current curves. B: responses to depolarizing pulses. This patch also contained a single slowly inactivating channel (maximum number ever opened while patch was held), one opening of which is seen at right. C: continuous record of responses to repeated 80 mV depolarizing pulses of same patch as in Fig. 1A. In B and C shading shows current attributable to leakage (proportional to current elicited by a 25 mV hyperpolarization from rest).
also did not occur when the potential was stepped from -25 mV below rest to + 130 mV above rest (Fig. 9B, right). With repeated depolarizations the amount of 'non-inactivating' current decreased somewhat (Fig. 9C). This decrease in current was always small compared to inactivation of the large slowly inactivating channels, such as occurred in the same patch in Fig. 9C. Experiments with T E A and cobalt in the patch pipet were done in order to test whether the non-inactivating voltage-dependent current had properties expected of a delayed potassium current. Fifty mM T E A in the patch pipet did not block non-inactivating
•
23
A
B
O Co+++ 50 TEA
0 Ca** ÷10 Co÷~
2pA
12pA
50 mV
150mY / m
lOs
Fig. 10. Effects of ions and tetraethylammonium (TEA) on non-inactivating voltage-dependent current. Upper curves, current; lower curves, amount of depolarization of patch membrane from rest. A: 0 Ca2÷ + 50 TEA: pipet solution did not contain CaCI2and was diluted 1:20 with 1 M TEA. B: 0 Ca2÷ + 10 Co2÷: 10 mM CoCI2 was used in place of CaCI2 in the pipet solution.
voltage-dependent current (Fig. 10A). The range of currents obtained with and without T E A (e.g. compare Fig. 9A right and Fig. 10A) overlap (for 7 T E A patches, range is 0 to 5 pA/10 mV; median is 0.7 pA/10 mV). When 10 mM CoC12 was substituted for CaCI 2 in the pipet solution, the non-inactivating current was not present; in fact, few channels or currents other than a linear 'leakage' current were present (Fig. 10B). DISCUSSION The properties of unitary channels underlying inactivating voltage-dependent outward currents in Helix neurons have not previously been described. This paper provides data on two types of inactivating outward current channels, which we will attempt here to relate to whole cell currents. In addition noninactivating voltage-dependent current, carried by channels clearly distinct from the two species of inactivating channels, has been observed, which we will also attempt to relate to whole cell currents. First, we consider the small, quickly inactivating channels, because their relationship to whole cell currents seems most obvious. These channels probably underlie the fast transient current 19, also known as IA4. Properties of I A that appear to be similar to the channels described here include (a) I A usually requires a preceding hyperpolarization below resting
potential in order to be elicited; (b) I A reaches peak activation within 15 ms of depolarization; and (c) the inactivation time constant for IA is less than 150 ms. These properties not only have clear correlations with characteristics of the quickly inactivating channels described here, but also serve to distinguish IA clearly from the slowly inactivating channels reported in this paper. Other laboratories have also recently described putative I A channels. We report here a conductance of 14 pS. Other reported conductances include a smaller size in Drosophila neurons 24, 20 pS in cultured dorsal root ganglion neurons 13, and 22 pS in cultured nodose ganglia neurons 23. The slowly inactivating channels clearly underlie a potassium current, but it is presently uncertain whether these channels are delayed potassium or calcium-dependent potassium channels. It has been proposed, based on whole cell current kinetics and pharmacology 3'5"25, that delayed potassium current may utilize separate populations of channels for the inactivating and non-inactivating fractions of the total outward current. Several different sizes of delayed potassium channels have been seen in single channel studies 14. A fluctuation analysis model for flog node predicts a unitary conductance for the inactivating channel of 40-70 pS 5, in agreement with the channels reported here. It should be pointed out, however, that the frog node study was done in the presence of calcium and the inactivating channels could conceivably be calcium-activated. Arguments against the slowly inactivating channels reported here being calcium activated include the following: (a) no correlation with slow inward current (compare Figs. 1 and 5); (b) present when no CaCl 2 was used in pipet solution and when barium was substituted for calcium (Fig. 4); (c) when elicited by cell depolarization the inactivating channels are active when action potentials are relatively narrow, and have inactivated when the action potentials are broad, indicative of larger calcium current (Fig. 6); and (d) calcium-activated channels, studied in Helix in three other laboratories by both fluctuation analysis methods 11 and single channel methods 9A7, have an estimated single channel conductance of 15-20 pS, less than half the conductance of channels described here, and have been reported not to inactivate in Helix ~7. In unpublished studies from our laboratory 18 pS channels have been seen in Helix neurons and may
24 correspond to these calcium-activated channels. On the other hand, these data do not conclusively exclude calcium dependence of the slowly inactivating channels. Since the ~0 calcium' solution was made by leaving CaCI2 out of the pipet solution, it probably contained a small amount of free calcium (according to B. Frankenhaeuser m. between 2 × 1()-s M and 10-~ M). The time course of inactivation of putative calcium current (Fig. 5 and ref. 16) resembles that of the inactivation of the slowly inactivating channels. The observation by 3 laboratories of calcium-dependent channels of 15-20 pS does not exclude a species of 45 pS from also being calcium-dependent. Non-inactivating 52 pS channels in Helix neuron F2 have been suggested on indirect evidence to be calcium-dependent 7. Whether these channels are indeed calciumdependent and whether they may be another mode of activity of the slowly inactivating channels described here is a question for future research. Non-inactivating voltage-dependent current was looked for initially to corroborate predictions based on fluctuation analysis of delayed potassium current by Reuter and Stevens 22. Although the expected current was seen, it had pharmacological properties that suggest that all or part of the current in the present study may be distinguishable from delayed potassium current. In particular, the current was not blocked by 50 mM TEA, a concentration which Meech and Standen TM show will block 80% of the delayed potassium current, and it was blocked by 10 mM cobalt in the electrodes. These properties are similar to those described by Thomas and Meech 26 and Byerly et al. 2 for voltage-dependent hydrogen ion currents in snail neurons. In some neurons hydrogen ion currents may therefore play a larger role than previously suggested. Further studies on the pH dependence, tail current reversal potential, and pharmacology of
REFERENCES 1 Aldrich, R.W., Jr., Getting, P.A. and Thompson, S.H., Mechanisms of frequency-dependent broadening of molluscan neurone soma spikes, J. Physiol. (London), 291 (1979) 531-544. 2 Byerly, L., Meech, R. and Moody, W., Jr., Rapidly activating hydrogen ion currents in perfused neurones of the snail, L ymnaea stagnalis, J. Physiol. (London), 351 (1984) 199-216. 3 Chabala, L.D., The kinetics of recovery and development of potassium channel inactivation in perfused squid (Loligo
these voltage-dependent currents must be done to test this suggestion. Whatever the exact ion dependence and selectivity of the channels and currents described here, an evident conclusion from this paper is that in Helix inactivating outward currents are carried by distinguishable and much larger channels than the non-inactivating currents. Fig. 9C illustrates the integration of the two types of channels, as seen in a number of patches. The function of the inactivating portion of the outward current has an important role in allowing action potential broadening in cells with calcium currents 1. As outward current channels inactivate, there is a smaller repolarization current, thereby leading to longer duration action potentials, consequent greater activation of calcium channels, and broadening of the calcium-dependent shoulder. This conclusion would be true whether the inactivating channels were calcium dependent or not. This model is fully consistent with inactivation of the large outward channels in Fig. 6 as action potentials became broader. Utilization of separate channels for the inactivating portion of the outward current may make modulation of its function less complex.
ACKNOWLEDGEMENTS We thank the Lady Davis Trust for financial assistance to J.L.R. while he was on sabbatical leave as a Visiting Associate Professor at the Technion. We thank Anna Levi for technical assistance, Boaz Gillo, Douglas Ewald, and Steven Siegelbaum for helpful discussions and advice on patch clamping technique, Yoram Palti for comments on an earlier draft of this paper, and Kibbutz Ma'ayan Zwi for use of an Apple lie computer.
pealie) giant axons, J. Physiol. (London), 356 (1984) 193-220. 4 Connor, J.A. and Stevens, C.F., Voltage clamp studies of a transient outward membrane current in gastropod neural somata, J. Physiol. (London), 2!3 (1971) 21-30. 5 Conti, F., Hille, B. and Nonner, W., Non-stationary fluctuations of the potassium conductance at the node of Ranvier of the frog, J. Physiol. (London), 353 (1984) 199-230. 6 Conti, F. and Neher, E., Single channel recordings of K÷ currents in squid axons, Nature (London), 285 (1980) 140-143. 7 Cottrell, G.A., Davies, N.W. and Green, K.A., Multiple
25 actions of a molluscan cardioexcitatory neuropeptide and related peptides on identified Helix neurones, J. Physiol. (London), 356 (1984) 315-333. 8 Dubois, J.M., Evidence for the existence of three types of potassium channels in the frog Ranvier node membrane, J. Physiol. (London), 318 (1981) 297-316. 9 Ewald, D.A., Williams, A. and Levitan, I.B., Modulation of single Ca2+-dependent K + channel activity by protein phosphorylation, Nature (London), 315 (1985) 503-506. 10 Frankenhaeuser, B., The effect of calcium on the myelinated nerve, J. Physiol. (London), 137 (1957) 245-260. 11 Hermann, A. and Hartung, K., Noise and relaxation measurements of the Ca 2+ activated K + current in Helix neurones, Pfliigers Arch., 393 (1982) 254-261. 12 Hille, B., Single Channels in Excitable Cells, Sinauer Assoc., Sunderland, MA, 1984, pp. 230-236. 13 Kasai, H., Single transient potassium currents in mammalian sensory neurons studied using patch clamp techniques, Soc. Neurosci. Abstr., 11 (1985) 955. 14 Llano, I. and Bezanilla, F., Two types of potassium channels in the cut-open squid giant axon, Biophys. J., 47 (1985) 221a. 15 Leefmans, F.J. and t3amino, S.M., Intracellular free potassium concentration in snail neurones measured with ion-selective microelectrodes, Soc. Neurosci. Abstr., 8 (1982) 684. 16 Lux, H.D. and Brown, A.M., Patch and whole cell calcium current recorded simultaneously in snail neurones, J. Gen. Physiol., 83 (1984) 727-750. 17 Lux, H.D., Neher, E. and Marty, A., Single channel activity associated with the calcium dependent outward current
in Helix pomatia, Pfliigers Arch., 389 (1981) 293-295. 18 Meech, R.W. and Standen, N.B., Potassium activation in Helix aspersa neurones under voltage clamp: a component mediated by calcium influx, J. Physiol. (London), 249 (1975) 211-239. 19 Neher, F., Two fast transient current components during voltage clamp in snail neurons, J. Gen. Physiol., 58 (1971) 36-53. 20 Pallotta, B.S., Calcium-activated potassium channels in rat muscle inactivate from a short duration open state, J. Physiol. (London), 363 (1985) 501-516. 21 Ram, J.L. and Dagan, D., Two types of delayed potassium channels: inactivating and non-inactivating channels, Soc. Neurosci. Abstr., 11 (1985) 956. 22 Reuter, H. and Stevens, C.F., Ion conductance and ion selectivity of potassium channels in snail neurons, J. Membr. Biol., 57 (1980) 103-118. 23 Shrier, A. and Cooper, E., Single channel analysis of I a currents from mammalian sensory neurons, Biophys. J., 47 (1985) 385a. 24 Solc, C.K. and Aldrich, R.W., Patch clamp analysis of inactivating outward currents in dissociated CNS neurons of wild-type and shaker Drosophila, Soc. Neurosci. Abstr., 11 (1985) 954. 25 Strong, J.A. and Kaczmarek, L.K., Modulation of multiple potassium currents in dialyzed bag cell neurons of Aplysia, Soc. Neurosci. Abstr., 10 (1984) 1101. 26 Thomas, R.C. and Meech, R.W., Hydrogen ion currents and intracellular pH in depolarized voltage-clamped snail neurones, Nature (London), 299 (1982) 826-828.