J. exp. Biol. (1976), 64, 477-487
With 7 figures
Printed in Great Britain
477
THE DIFFERENT ACTIONS OF CHLORIDE AND POTASSIUM
ON POSTSYNAPTIC INHIBITION OF AN
ISOLATED NEURONE
BY HARTMUT MEYER
Max-Planck-Institut fiir Psychiatric, Z)-8ooo Munchen 40, Kraepelmstrasse 2, G.F.R.
(Received 16 September 1975)
SUMMARY
Isolated stretch receptor neurones from freshwater crayfish were examined
in solutions containing different concentrations of chloride and potassium.
In normal solution the inhibitory reversal potential (£n>gp) of this preparation
was strictly negative with respect to the resting potential and even to the
reversal potential of spike after-hyperpolarization. The time courses of
resting potential and 2?rpsp following rapid solution change suggest that the
current generating the IPSP is mainly carried by chloride ions and that the
participation of potassium is very small. This has also been confirmed by
the calculated conductances of the activated inhibitory membrane in the
different solutions. The results add further evidence for an outwardly
directed pumping of chloride ions which keeps the intracellular concentration of this anion at the low level necessary for hyperpolarizing inhibition.
INTRODUCTION
Recent work on cat motoneurones (Llinas & Baker, 1972; Lux, Loracher & Neher,
1970; Lux, 1971) as well as crayfish stretch receptors (Ozawa & Tsuda, 1973;
Meyer & Lux, 1974) indicates a predominating contribution of chloride to inhibitory
electrogenesis in these preparations. In the neuromuscular system of the lobster
(Grundfest, Reuben & Rickles, 1959; Motokizawa, Reuben & Grundfest, 1969) as
well as for the crayfish (Takeuchi & Takeuchi, 1967) it has been demonstrated that
the inhibitory postsynaptic membrane appears to become highly anion-selective
when synaptically activated. Inhibition of the crayfish motor giant fibre also appears
to be largely determined by chloride ions (Ochi, 1969), although in the stretch
receptor neurone potassium (Edwards & Hagiwara, 1959) as well as chloride (Ozawa &
Tsuda, 1973; Meyer & Lux, 1974) has been proposed to be mainly responsible for
the IPSP.
The experiments reported here were performed to discriminate between the
individual contributions of potassium and chloride to postsynaptic inhibitory currents
of the abdominal tonic stretch receptor of the crayfish.
31-2
478
HARTMUT MEYER
METHODS
Complete abdominal stretch receptors were dissected from male specimens of the
freshwater crayfish, Astacus fluviatilis L., and mounted unstretched in their original
orientation in a plexiglass chamber containing physiological saline. Only the receptors
of the first, second and third segments were used. A special arrangement allowed
rapid, temperature-controlled exchange of solutions in the bath. The ends of both
receptor muscle strands were held by a pair of fine forceps mounted on a mechanical
stretching device. The sensory nerve was sucked into a micromanipulated suction
electrode for antidromic stimulation. The nerve leading to the phasic receptor cell
(which contains a branch of the inhibitory axon supplying both the phasic and the
tonic cells) was stimulated with suprathreshold pulses, by means of a second suction
electrode, to generate IPSPs in the tonic, so-called 'slow', receptor cell (Kuffler,
1954; Kuffler & Eyzaguirre, 1955).
For intracellular potential recording and simultaneous current injections doublebarrelled glass microelectrodes were used. The recording barrel contained a mixture
of 85 % of saturated (~ o-6 M) KS,SO 4 solution and 15 % of 1-5 M-KC1 solution. The
addition of KCl was found to stabilize the potentials recorded via a sintered Ag-AgCl
pellet (Lux et al. 1970), the KCl concentration being low enough to avoid appreciable
Cl~ leakage into the cell. For Cl~ injections the other barrel was filled with 0-5 mMKC1 solution. The reference electrode was another Ag-AgCl pellet which was connected with the solution in the bath via a second glass microelectrode, filled with a
3 M-KCI solution. To minimize uncontrolled potential changes caused by KCl
diffusion from the tip, the diameter of the latter was kept small (3-5 /im). Electrophoresis current was delivered by a floating amplifier configuration which allowed
controlled current injection with an accuracy of 5 % (Lux et al. 1970). Temperature
in the bath was adjusted to 14 ± 1 °C and kept constant by means of a cooling device.
The standard solution employed was that described by van Harreveld (1936) and
contained: NaCl 205; KCl 5-4; CaCl2 13-5; MgCl2 2-6; NaHCO3 2-3 mM. In the
experiments where external ionic concentrations were varied, isethionate was exchanged for chloride and sodium for potassium in equal amounts to maintain constant
ionic strength and osmolarity. No differences could be observed when acetate was
exchanged for isethionate (i.e. the membrane permeability of this neurone for acetate
seems to be negligibly small). The pH of all solutions was carefully adjusted to
7-5 ± 0-2 with 1 M-Tris-maleate buffer, since preliminary experiments had revealed a
considerable effect of the pH upon both resting potential and spike generation
(Meyer & Prince, 1973). For long-term observations the membrane potentials and
the bath temperature (measured directly by means of a thermistor in conjunction
with a bridge circuit) were continuously monitored by a strip chart recorder. The
stimulation and recording equipment was similar to that described elsewhere (Meyer &
Prince, 1973; Meyer & Lux, 1974).
RESULTS
RP and EIPSF under normal conditions
Resting membrane potential (RP) and IPSP reversal potential (EjpSV) were recorded
from 83 neurones with resting potentials of at least — 50 mV. These cells were
Cl~ and K+ actions on postsynaptic inhibition
479
-70
I
-80
50 msec
-90
-90
-80
-70
Membrane potential (mV)
-60
-i
20 mV
Fig. i. A. Scatter diagram for resting potentials and En-si* of 83 different cell preparations in
normal saline. The E1PBPs are consistently negative with respect to the resting potentials. The
45 0 line indicates values expected for coincident RPs and Emrs.
B. Reversal potentials of spike after-hyperpolarization (£AHP) and of IPSP CEIFSP) obtained when
the resting potential was varied by trans-membrane current pulses, each about 1 s after the
preceding. Upper beam indicates magnitude of currents flowing across the membrane during
the single current steps.
selected from a larger sample on the basis of the stability of potentials and the complete recovery of RP and IPSP after exchange of solutions. After insertion of the
electrode into the cell body the resting potential was allowed to stabilize for at least
5 min. The resting potentials averaged — 64-3 ±1-3 mV (mean ±s.D.) and the
inhibitory reversal potentials — 8i-i±i-8mV. A mean input resistance of 6-5 ±
0-9 MCI was calculated from intracellular transmembrane current pulses and the
resulting changes in E^ (Fig. 1).
Effects of variations in external Cl~ concentration
In these experiments variations in RP and £rp 9 p were recorded from 21 cells during
and following rapid exchange of normal saline by Ringer solution containing only
I
5% (36 miw) of the normal chloride contents, the remainder being replaced by
equivalent amounts of isethionate. Lower Cl~ concentrations were not used since
the inhibitory synaptic transmission was frequently blocked in solutions containing
less than ~ 35 mM. Figs. 2 A and B show that RP changes following solution exchange
were comparatively slow and small (4-8 ±3-2 mV), whereas the £n>sp altered much
faster with a biphasic time course and a peak amplitude of 22-0 + 5-1 mV. To determine the total input resistance in different extracellular Cl~ concentrations, the cells
were soaked for at least 10 min in solutions containing 100, 75, 50 or 15% of the
standard Ringer's Cl~ contents (243 mM 1 = 100%), the removed chloride being
replaced by isethionate. The input resistance was slightly altered, the maximal change
during passage from normal to 15% Cl~ Ringer and vice versa was found to be
0-9 MQ (Fig. 2C).
Effects of intracellxdarly injected Cl~ ions
In these experiments seven cells were penetrated with double-barrelled electrodes, one barrel being filled with 0-5 M-KC1 solution and the other with the normal
480
HARTMUT MEYER
II
-50
-60
I -70
Is
a
-70
-80
-90
3
4
Time (min)
-2
nA
10
^ * v
5
0 Normal, RP - 6 4 mV
• 15 % C1-, RP - 7 3 mV
10
•
Depolarization (mV)
20
I
I
I
Hyperpolarization (mV)
-2
•4
Fig. 2. Simultaneous recordings of potential changes during and following perfusion of the
neurone with 15% Cl" solution (A) and with normal saline again (B). Arrows indicate
beginning and end of solution change. C. Current-voltage relations for a neurone in solutions
containing different concentrations of Cl~ for 10 min. The respective resting potentials are
indicated in the inset.
mixture. For iontophoretic injections a hyperpolarizing current of 40 nA
was applied for 90 s. The RP and E^pgp were continuously monitored from termination of the injection until equilibrium conditions had recovered. Normally three to six
subsequent injections were possible, each about 30 min after the other. The most
common effect found was an immediate shift of the .Eipgp t 0 values about o mV or
further, whereas RP remained essentially unaffected aside from minor initial depolarizations. As a result the IPSPs became depolarizing with amplitudes of about
25 mV. Normally this was enough to reach the firing level and to evoke an action
potential (Fig. 3 A, B). It is noteworthy that insertion of the iontophoresis electrode
itself caused in many cases a slight depolarizing shift of the EIPSF which apparently
was due to Cl~ leakage from the tip of the electrode into the cell. The half time of
the nearly exponential recovery of the iSjpgp from the Cl~ injections was 2-9 ± 0-5 min.
Cl~ and K+ actions on postsynaptic inhibition
3 min
5 min
10 min
* •
20 mV
50 msec
After chloride injection
+ 10r
.2
<£
-30
-50
-70
1PSP
8
Time (min)
12
16
Fig. 3. Effects of a chloride injection (40 nA, 90 s) on a neurone bathed in normal saline.
A. Immediately after injection the IPSP became depolarizing, the amplitude being large
enough to reach the firing level. B. Time courses of resting potential and ElPSf following
chloride injection (indicated by arrow).
Effects of externally varied K+ concentration
Variations of RP and EIV8F were recorded during and following rapid exchange
(30 ml/min) of normal saline by solutions containing different amounts of potassium.
A decrease or increase of extracellular K + concentration [K+]o caused an immediate
shift of the resting potential (Fig. 4E) and a simultaneous but slower change in the
J? Ipgp (Fig. 4A-D). To measure the total input resistance, the cells were exposed for
at least 10 min to potassium-free solutions or those containing 5-4 (normal) and
20 mM — K + . The total quantity of cations was kept constant by addition or removal
of sodium. An increase of [K + ] o from 0 to 20 mM resulted in a decrease of the input
resistance from about 11 to 4 MQ (Fig. 4E). The respective £n>sp values under
steady-state conditions were —104 and —57 mV (cf. Fig. 6B).
The fast time courses of the EiFSP shifts following variations of the extra- and
intracellular Cl~ concentration indicate a very important contribution of chloride
HARTMUT MEYER
H
-40
-60
-80
Resting potential
& -100
A
B
C
D
-45
Norma!-+20 mM-K+
K+-free-»normal
20 mM-K.+-mormal
Normal-»K+-free
-65
-85
D
-105 •
2
4
Tune (min)
nA
2
10
Depolarization (mV)
• K.+-free, RP - 9 4 mV
o Normal, RP - 6 5 mV
a 20 mM-K+, RP - 4 9 mV
10
20
30
Hyperpolarization (raV)
E
Fig. 4. A-D. Simultaneous recordings of potential shifts during and following perfusion
of the neurone with salines containing different concentrations of potassium. Arrows indicate
beginning and end of solution change. E. Current-voltage relations for a neurone exposed to
different external potassium concentrations for 10 min, indicating a decrease of total input
resistance with increasing external K+ concentration. The respective resting potentials are
indicated in the inset.
to the generation of the inhibitory e.m.f. Effects on the resting potential were comparably small and slow (Figs. 2, 3). On the other hand a change of the extracellular
K+ concentration produced a sudden shift of the resting potential whereas the Erpgp
markedly lagged behind (Fig. 4). Since the resting potential obviously depends
directly on the transmembrane K+ gradient, these different time courses of RP and
JFjpsp seem to rule out an appreciable contribution of potassium to inhibitory electrogenesis and of chloride to the generation of the resting potential. However, there is
a slow .ETPSP shift during [K+] variation in the bathing solution, which indicates an
Cl~ and K+ actions on postsynaptic inhibition
483
58
40
8
20 <
15%
36
50%
100%
243
0
a-
360 HIM
Fig. 5. Dependence of isipsp on transmembrane chloride distribution. Absolute £irsp changes
are plotted against the respective external chloride concentrations. The different symbols
represent data from different neurones. The dashed line indicates the values expected for the
Nernst equation (58 mV per tenfold concentration change).
+ 10r
. K+-free
o Normal
0 20 mu K+
t
-120
-100
-80
-60
-40
Membrane potential (mV)
Fig. 6. IPSP peak amplitudes plotted against the respective membrane potentials at which
they were recorded. Solid lines: resulting curves for presynaptically evoked IPSPs. Dashed
lines: resulting curves for IPSPs elicited by GABA pulses. The intersection points with
zero line indicate IPSP reversal potentials (open arrows). Respective resting potentials are
marked by filled arrows.
indirect effect of potassium on the generation of the IPSP. It is conceivable that the
variation of the external K + concentration is followed by a redistribution of potassium
across the membrane. This would produce a compensating redistribution of anions,
presumably of chloride to achieve electrochemical equilibrium. There is in addition
good evidence for an outwardly directed chloride pump (Meyer & Lux, 1974) which
maintains the chloride gradient negative to RP for hyperpolarizing inhibition. The
time course of the £rpgP could, therefore, reflect the time course of botR passive and
active chloride movement across the membrane.
IPSP reversal potential and Cl~ gradient
If the E1PSP is due solely to the transmembrane Cl~ gradient (E1¥BF ~ EC1), the
intracellular fraction of electrochemically active chloride [Cl~]i can be estimated by
484
HARTMUT MEYER
Table 1. Effect of ionic concentrations on inhibitory conductance change
Solution
K+-free
Normal
20 tnM-K
is% ci-
£ M (mV)
+
— I2O
-90
-70
-90
AEi(mV)
£buP(mV)
fti(nS)
+ 9-4
-108
-82
-54-5
153
270
90
-67
467
153
416-1
44-1
means of the Nernst equation. With a mean £ IP gp of — 80 mV and a [Cl~]0 of 243 mM,
the value of [Cl~]j in normal Ringer solution was calculated to be about i0'5 mM.
In K+-free solution [Cl~]i decreased to 3-7 mM and even to 1-5 mM when 85% of
the normal Cl~ contents were replaced by isethionate. The EIF3P changes resulting
from variations in [Cl~]0 were smaller, by about 42 % (Fig. 5), than those predicted
from the Nernst equation if only a single ionic battery is assumed for inhibitory
electrogenesis. This deviation can be explained either by an increase in other ionic
conductances (K+) during an IPSP or by a transmembrane chloride redistribution
following a change of [Cl~]0 due to the membrane permeability for KC1. This
chloride redistribution reduces the difference between [Cl~]t and [Cl~]0, and thus
prevents the ElPSP from reaching its theoretical value.
Effect of ionic concentrations on the inhibitory membrane conductance
To decide whether part of the inhibitory current might also be carried by potassium
ions, the conductance changes of the inhibitory subsynaptic membrane were determined in solutions of different ionic concentrations by the following two methods.
1. The synaptic conductance change was calculated, assuming that the synaptic
and nonsynaptic current are in equilibrium at the peak of an IPSP, as follows:
(a) Aft (Eu + AEj-
(*) Agl = E f g
£IPSP) = - f t j - ^ i ,
,
i l M + Ail T — i i I P S p
where Agt = synaptic conductance change, EM = membrane potential, AEt = IPSP
peak amplitude, £ M = resting membrane conductance.
To determine these values in the different Cl~ and K+ solutions, the IPSP peak
amplitudes were plotted against the respective membrane potentials (Fig. 6). The
resulting curves approximate straight lines within the potential ranges given in the
figure. Since the curves relate the IPSP peak amplitudes and the respective driving
potentials, a change in the slope can be interpreted as a change in the effectiveness
of the IPSP ( A B ^ M - Ejpsr)).
To ensure that the calculated conductance changes did not result from changes
in release of transmitter from the presynaptic terminals (Gage & Quastel, 1965;
Liley, io,56;Takeuchi & Takeuchi, 1961, 1962; Edwards & Ikeda, 1962; Hubbard &
Willis, 1962), the inhibitory receptors were also activated by GAB A (which was
released electrophoretically from an extracellular electrode). The resulting potentials
are given by the dashed lines in Fig. 6.
2. If it is assumed that GABA acts only on the inhibitory postsynaptic membrane,
and produces additional ionic fluxes in parallel with the nonsynaptic membrane
Cl~ and K+ actions on postsynaptic inhibition
485
Table 2. Effect of ionic concentrations on inhibitory conductance: IPSPs are
elicited by GABA pulses
Solution
K+-frec
Normal
20 mM-K
i(mV)
i(mV)
— 120
-OO
+
-70
-90
15% ci-
gu (nS)
-II2-5
+ 29
+ 2-8
+ 2-5
+ 97
-82
-56
-685
90
56-7
153
270
151
824
58-7
124-1
-nA
~ K+-frcc
- 5 10 mV
1
1
10 5
Normal
1
\\
P
-v
Fig. 7. Current-voltage relations for a neurone with different external ionic concentrations
(solid lines) and in the same solutions with 4 x io~* mg/ml GABA.
Table 3. Effect of ionic concentrations on resting membrane conductance in
solutions with and without GABA added
Solution
+
K -free
Normal
20 mM-K +
15% Cl-
g* (nS) ft, + G A B A (nS)
83
147
133
228
128
270
280
269
(nS)
64
gl
137
52
141
(Takeuchi & Takeuchi, 1965, 1967), then the GABA-induced conductances change
of the inhibitory membrane can be determined by subtracting the resting conductance
Figs. 2E, 3C) from the conductance in the presence of GABA in the bath. The
values were obtained from three cells, like those of Figs. 2E and 3 C, with and without
4 x io~3 mg/ml (~4 x io" 5 M) GABA added to the different Cl~ and K + solutions
(Fig. 7).
It is shown that the conductance changes of the inhibitory synaptic membrane
obtained by GABA pulses (Table 2) agree well with those calculated by the subtraction method (Table 3). Thus the differences in the values obtained from presynaptically evoked IPSPs could arise from a reduction in presynaptic transmitter release in
solutions with lowered chloride concentrations (Fig. 6A) and an increase of transmitter release when the extracellular K+ concentration is raised (Fig. 6B). This
shows that the great differences in inhibitory conductance changes given in Table 1
are not real, but are brought about by changes in the release of inhibitory transmitter
under different ionic conditions.
DISCUSSION
It is demonstrated that the IPSP reversal potential of the stretch receptor neurone
of Astacus fluviatilis is normally about 7 mV negative with respect to RP and even
486
HARTMUT MEYER
slightly negative to reversal potential of the spike after-hyperpolarization
Since the £ AH p is known to depend mainly, or even exclusively, on the transmembrane
K + gradient (EASF ~ E^), the difference between EAap and ElP8F already indicates that
the inhibitory potential must be dominated by a different ionic gradient, which can
drive EM to values more negative than EK during synaptic activation. Changes in
external Cl~ concentrations caused an immediate shift of the Ejj^p which was followed
by a slow change of RP in the range of some mV. Such a change of the resting potential
would be expected, with a membrane permeable to KC1, because a redistribution of
Cl~ across the membrane would be associated with a corresponding flux of potassium
to maintain electrochemical equilibrium. Comparable effects on jEjpgp and RP were
brought about by intracellular chloride injections. With a slope of 42 mV/decade
change of [Cl~]0 (Fig. 5), an intracellular concentration of free Cl~ of about 4-7 mM/1
can be estimated. Comparing this value with those given in the literature for related
material (~5omM/l in the lobster muscle fibre, Dunham & Gainer, 1968) the
question arises whether this marked difference between [Cl~]0 and free [Cl~]j is
restricted only to one or more subsynaptic compartments as suggested for other
preparations (Dunham & Gainer, 1968). Since this would include boundary structures
with a low permeability for chloride, one would expect a delayed shift of the 2?rpsP
following intracellular chloride injections due to aggravated intercompartmental Cl~
distribution which could not be observed. Whether a compartmentation due to
partial Cl~ immobilization in chemical compounds is realized in this preparation
cannot be decided from these experiments. Alteration of the external potassium
concentration caused an immediate shift in RP whereas the changes of EIPSP occurred
more slowly (Fig. 4). This indicates that the transmembrane [K+] gradient is unlikely
to contribute to the inhibitory e.m.f.
A predominant role of chloride ions in the generation of the IPSP was, however,
indicated by the calculated inhibitory membrane conductance changes by two independent methods (Tables 2 and 3). It was also shown that any change of [K + ] 0)
especially an increase, resulted in a considerable reduction of AgIt whereas a reduction of [Cl~]0 had an opposite effect. Nevertheless, a minor contribution of EK to
the inhibitory e.m.f. cannot be excluded. Since a sudden change of [K + ] o implies a
redistribution of potassium across the membrane, the observed shift of the EIF3P
must obviously result from a simultaneous redistribution of chloride which maintains
the EjpSP consistently negative to the RP, even in K+-free conditions with an RP of
— 90 mV or more (cf. Fig. 4D). These experiments do, therefore, provide evidence of
an active mechanism which keeps the EIVSP negative with respect to RP, presumably
by means of an outwardly directed chloride pump as already proposed for other
preparations (Llinas & Baker, 1972; Lux, 1971; Motokizawa et al. 1969; Ozawa &
Tsuda, 1973; Meyer & Lux, 1974). According to this interpretation the time courses
of the Ejpsp shifts during external [K+] change depend on both passive membrane
conductance for chloride and the rate of active Cl~ extrusion. The chloride redistribution obtained with increasing [K + ] o results in an increase in [Cl-]j, which reduces the
difference between [Cl~]£ and [Cl~]0. Additionally there is a decrease in total input
resistance (Fig 4E) which could facilitate an influx of chloride. On the other hand
[Cl"]! decreases upon reduction of [K + ] o (Fig. 4C, D) and thus increases the difference
between intra- and extracellular Cl~ concentrations. A simultaneous rise in the total
Cl~ and K+ actions on postsynaptic inhibition
487
input resistance also reduces chloride extrusion. This explains the observed changes
•n the rates of chloride redistribution obtained upon raising (~ 30 ± 10 s) and diminishing [K+]o (~ 80 ± 20 s).
The exactly similar time courses of the ^jpgp (upon change from K+-free to normal
and from normal to 20 mM K + solution) indicate that the passive chloride conductance and the rate of the active chloride extrusion are not essentially influenced by EM.
This investigation was partly supported by the Deutsche Forschungsgemeinschaft.
REFERENCES
DUNHAM, P. B. & GAINER, H. (1968). The distribution of inorganic ions in lobster muscle. Biochim.
Biophyt. Acta 150, 488-99.
EDWARDS, C. & HAOIWARA, S. (1959). Potassium ions and the inhibitory process in the crayfish stretch
receptor. J. gen. Physiol. 43, 313-21.
EDWARDS, C. & IKEDA, K. L. (1962). Effect* of 2-PAN and succinyl choline on neuromuscular transmission in the frog. J. Pharmacol. 138, 332-9.
GAGE, P. W. & QUASTBL, D. M. J. (1965). Dual effect of potassium on transmitter release. Nature,
Lond. 206, 625-6.
GRUNDFEST, H., REUBEN, J. P. & RICKLES, W. H. JR. (1959). The electrophysiology and pharmacology
of lobster neuromuscular synapses. J. gen. Phytiol. 43, 1301-23.
HUBBARD, J. I. & WILLIS, W. D. (1962). Reduction of transmitter output by depolarization. Nature,
Lond. 193, 1294-5.
KUFFLER, S. W. (1954). Mechanisms of activation and motor control of stretch receptors in lobster
and crayfish. J. Neurophysiol. 17, 558-74.
KUFFLER, S. W. & EVZACUIRRE, C. (1955). Synaptic inhibition in an isolated nerve cell. J. gen. Phytiol.
39. I55-84LILEY, A. W. (1956). An investigation of spontaneous activity at the neuro-muscular junction of the
rat. J. Phytiol. 13a, 650-66.
LLINAS, R. & BAKER, R. (1972). A chloride-dependent inhibitory postsynaptic potential in cat trochlear
motoneurons. J. Neuropltytiol. 35, 484-92.
Lux, H. D., LORACHER, C. & NEHER, E. (1970). The action of ammonium on postsynaptic inhibition
of cat spinal motoneurons. Exp. Brain Ret. 11, 431-47.
Lux, H. D. (1971). Ammonium and chloride extrusion: hyperpolarizing synaptic inhibition in spinal
motoneurons. Science, N.Y. 173, 555-7.
MBYER, H. & PRINCE, D. A. (1973). Convulsant actions of penicillin: effects on inhibitory mechanisms.
Brain Ret. 53, 477-82.
MEYER, H. & Lux, H. D. (1974). Action of ammonium on a chloride pump: removal of hyperpolarizing
inhibition in an isolated neuron. PflOgert Arch. 350, 185-95.
MOTOKIZAWA, F., REUBEN, J. P. & GRUNDFEST, H. (1969). Ionic permeability of the inhibitory postsynaptic membrane of lobster muscle fibers. J. gen. Physiol. 54, 437-61.
OCHI, R. (1969). Ionic mechanism of the inhibitory postsynaptic potential of crayfish giant motor
fiber. Pflkgers Arch. 311, 131-43.
OZAWA, S. & TSUDA, K. (1973). Membrane permeability change during inhibitory transmitter action
in crayfish stretch receptor cell. J. Neurophytiol. 36, 805-16.
TAKBUCHI, A. & TAKEUCHI, N. (1961). Changes in potassium concentration around motor nerve
terminals produced by current flow, and their effects on neuromuscular transmission. J. Phytiol.
155. 46-58.
TAKEUCHI, A. & TAKEUCHI, N. (1962). Electrical changes in pre- and post-synaptic axons of the giant
synapse of Loligo. J. gen. Phytiol. 45, 1181-93.
TAKEUCHI, A. & TAKEUCHI, N. (1965). Localized action of gamma-aminobutyric acid on the crayfish
muscle. J. Pliytiol. 177, 225-38.
TAKEUCHI, A. & TAKEUCHI, N. (1967). Anion permeability of the inhibitory post-synaptic membrane
of the crayfish neuromuscular junction. J. Phytiol. 191, 575-90.
VAN HARRBVELD, A. (1936). A physiological solution for freshwater crustaceans. Proc. Soc. exp. Biol.
(N.Y.) 34, 428-32.