Neuron,

Vol. 12, 117-126,

January,

1994, CopyrIght

0 1994 by Cell Press

Functional Diversity of GABA-Activated

Cl- Currents in Purkinje versus Granule
Neurons in Rat Cerebellar Slices
Ciulia Puia, Erminio Costa, and Stefano Vicini
FIDIA
Georgetown
Institute for the Neurosciences
Georgetown
University School of Medicine
Washington,
D.C. 20007

Summary
In rat cerebellar
slices, we compared
wholecell
recordings of spontaneous
inhibitory
postsynaptic
currents (sIPSCS) with Cl- currents resulting from pulses
of GABA (1 mM, <2 ms) to outside-out patches from
Purkinje and granule neurons. slPSCs in Purkinje cells
decayed with a single fast exponential,
as previously
reported, whereas in granule cells slPSC decay was best
described by the sum of a fast and a slow exponential
curve, with a variable contribution
of the slow component to the peak current. CABA pulses to nucleated
patches from granule cells elicited Cl- currents with de
cays similar to slPSC decays, whereas in patches from
Purkinje neurons GABA pulses produced Cl- currents
decaying largely with a fast component, but often followed by a slower exponential.
CABA concentration
steps produced rapidly desensitizing currents in patches
from both cerebellar neurons. In distinct cerebellar neurons, specific functional properties of CABAA receptors
may relate to the presence of distinct receptor subtypes.
Introduction
yhminobutyric
acid (GABA), the major inhibitory
neurotransmitter
in the mammalian
CNS, hyperpolarizes
neuronal membranes byopeningacl-channel
intrinsic to the GABAA receptor (Sivilotti and Nistri, 1991).
This receptor
is a hetero-pentameric
complex that
contains two allosteric modulatory
sites for clinically
relevant drugs (Costa, 1989; Sieghart, 1992).
Molecular
cloning studies have established
that 16
genes encode the subunits assembled
in GABAA receptors (reviewed in Olsen and Tobin, 1990, and Burt
and Kamatchi, 1991). The combinatorial
manner by
which this large subunit diversity is assembled to form
GABAA receptors
suggests that a high number
of
structurally
different
receptor subtypes may exist in
vivo. This great heterogeneity
might correspond
to
functionally
distinct postsynaptic
receptors in inhibitory synapses.
A number
of studies have investigated
the functional properties of GABAA receptors at inhibitory
synapses by studying spontaneous
and evoked inhibitory
postsynaptic currents (IPSCs) in voltage-clamped
neurons. Among these studies Cl--mediated
IPSCs were
described
in hippocampal,
cortical, and cerebellar
neurons in primary culture (Segal and Barker, 1984;
Vicini et al., 1986; Cull-Candy
and Usowicz, 1987).
IPSCswere also studied in hippocampal
neurons (Col-

lingridge
et al., 1984; Edwards et al., 1990; Ropert et
al., 1990; Otis and Mody, 1992; Pearce, 19931, septal
neurons (Schneggenburger
et al., 1992), and Purkinje
neurons (Farrant and Cull-Candy,
1991; Vincent et al.,
1992) from rat brain slices. When the results of these
studies are compared, the major difference found between GABA-mediated
IPSCs in various preparations
is the time constant of the decay of the synaptic current. This constant is largely variable, with values ranging from 4 to 30 ms, and in some inhibitory
synapses
becomes biphasic owing to the presence of a slower
time constant. After taking into account the variability
of these results related to various experimental
conditions, such as temperature,
holding potential value,
and direction
of Cl- flow, one cannot rule out the
possibility that at least part of the large range of IPSC
kinetic values is related to distinct
properties
of
GABA* receptors, perhaps corresponding
to specific
functional
characteristics
associated with selected
molecular
forms of GABAA receptor subunits.
To investigate further the functional
heterogeneity
of GABAergic synapses, we chose to study GABAergic
inhibitory
synapses in cerebellar Purkinjeand
granule
neurons.
In the mammalian
cerebellum
two major
types of GABA-mediated
synaptic inhibition
occur:
the first is on Purkinje
cells by stellate and basket
cells, and the second is operated by the type II Golgi
cells on granule cells (Palay and Chan-Palay,
1974).
The circuits in which the Golgi, stellate, and basket
cells participate
have under tight control the basic
mossy fiber-granule
cell-Purkinje
cell pathway and
are deeply involved in motor coordination
(Eccles et
al., 1967; Ito, 1984). In situ hybridization
studies (Laurie
et al., 1992) and immunocytochemical
techniques
(Thompson
et al., 1992; Fritschy et al., 1992) have
shown a different distribution
for the mRNAs encoding GABAA subunits and the subunit proteins themselves in Purkinje versus granule cells. In particular,
the presence of a6 as well as 8 subunits is detected
in cerebellar
granule cells but not in Purkinje cells.
Furthermore,
a recent study with specific antibodies
(Baude et al., 1992) indicates that immunoreactivity
for both al and a6 is present at synapses innervated
by type II Golgi cell terminals in granule neuron dendrites (Baudeet al., 1992), implying that GABA released
from Golgi neurons may act on several GABA,+ receptor subtypes. Thus, the structure of the postsynaptic
GABAA receptor may become the agent that characterizes the inhibitory
message carried by GABA released
from type II Golgi neurons.
The goal of our work was to characterize the properties of the GABAA receptor channels at the inhibitory
synapses located in cerebellar
Purkinje and granule
neurons by a comparison
of the biophysical
properties of slPSCs and GABA-activated
currents in outsideout patches excised from these neurons. We will describe functional
differences
of inhibitory
synapses

ii8

PURKINJE
r*-

GRANULE

r-

Figure 1. Whole-Cell
Recordingsof
Recorded
from Cerebellar
Neurons

slPSCs

(A) slPSCs from a Purkinje neuron (left) and

from agranule
neuron (right). (B)Theeffect
of bath perfusion
with bicuculline
methiodide (20 PM). (C)The average of 20 synaptic
currents
superimposed;
the curve
fitting
used single and double
exponentials.

-I

250

pA

-1

100 ms

IOOP

100 ms

T 59ms
90 pA

1Oms

in granule
slices.

versus

Purkinje

neurons

% Slow

in rat cerebellar

Results
We have investigated
the characteristics
of GABAactivated Cl- currents in specific neurons in rat cerebellar slices. Purkinje
and granule
neurons were
visually identified
by their location
and by easily
detectable morphological
characteristics,
such as the
size of the cell body. In many Purkinje cells most parts
of the large dendritic tree were clearly seen in slightly
different focal planes. Care was taken to record from
superficial
cells to minimize access resistance, which
was typically less than 10 MD. Our results derive from
a data set of 49 Purkinje cells and 42 granule cells.
The average resting potential measured immediately
after establishing
the whole-cell
recording configuration of Purkinje cells was -61 f 7 mV (mean + SD)
and the input resistance was 431 + 54 MQ; granule
cells had a resting potential
of -67 k 8 mV and an
input resistance of 3.9 + 0.9 CD.

Spontaneous IPSCs in Purkinje versus Granule

Neurons in Cerebellar Slices
We investigated
the characteristics
of spontaneous
IPSCs (slPSCs) with whole-cell
recordings
from Purkinje and granule neurons voltage clamped at -60
mV. A sample of slPSCs recorded from a Purkinje cell
is shown in Figure IA. As previously
reported (Farrant
and Cull-Candy,
1991; Vincent et al., 1992), in most
Purkinje neurons sIPSCS were highly variable in the
peak amplitude,
being 172 f 123 pA in our data base.
Bath perfusion
with the GABAA receptor antagonist
bicuculline
methiodide
(10 PM) completely
and reversibly abolished sIPSCS (Figure IB). In Purkinje neu-

54

lOIll*

rons under adequate voltage-clamp

control, as estimated by the analysis of the capacitative
currents to
a 10 mV hyperpolarizing
voltage step, the decay time
oftheaveraged
sIPSCswasdescribed
byasingleexponential curve (Figure IC) with decay time constant of
7.9 f 1.9 ms (mean f SD, n = 7). Similar results have
been previously
published
(Vincent et al., 1992).
In Figure IB, slPSCs from a granule neuron are
shown. The peak amplitude
of the average of 50
sIPSCS recorded in each granule neuron ranged from
11 to 151 pA, with a mean of 55 f 37 pA in our data
base42 granule cells. Bath perfusion with 10 uM bicuculline methiodide
completely
and reversibly
abolished slPSCs in granule neurons (Figure IB). In these
neurons, averaged sIPSCS had a rise time of 0.6 f 0.3
ms (range 0.3 to 1.3 ms). In granule neurons, the decay
phase of sIPSCS recorded
at a holding potential
of
-60 mV was best described
by the sum of two exponentials (Figure IC), with a fast time constant of 7.0
+ 1.6 ms and a slow time constant of 59 f 16 ms. In
the presence of tetrodotoxin
CrrX; 1 PM) sIPSCS in
granule neurons were reduced in both amplitude and
frequency
of occurrence
but mantained
two kinetic
component
decay, with a fast time constant of 6.4 f
0.9 ms and a slow time constant of 65 f 19 ms (n =
3), and the relative proportion
of the slow component
to the average sIPSC was on average of 57% * 19%.
Cerebellar
granule cells were electronically
compact, as suggested by the poor correlation
between
the distribution
of the amplitude
versus the various
kinetic parameters (Figure 2) investigated
in a sample
of 100 sIPSCS recorded from a single granule neuron
and by total electrotonic
length measurements
(Silver
et al., 1992). Therefore,
the slower component
of
sIPSC decays could not arise solely from unclamped
synaptic sites. A slow component
was always present,
and its average contribution
to the peak current was

GABA-Mediated
119

Currents

in Cerebellar

Slices

Figure
plitude

2. Correlation
and Kinetic

between
Parameters

slPSC

Am-

Relationship
of the amplitude
versus the fast
decay time constant
(A), the slow decay time
constant
(B), the rise time (C), and the percent slow component
(D) derived from the
doubleexponential
fitting of 100 slPSCs recorded
in a cerebellar
granule neuron.
The
linear regression
fit is superimposed
with
the data, and the correlation
coefficient
(r)
is also indicated.

=0.02

r023

3 &-FYTTr

Amplitude (pA)

a hallmark of each granule cell investigated,

ranging
from 15% to 100% of the total peak amplitude,
with
an average of 60% L- 23%. In Figure 3A we report the
distribution
of the relative contribution
of the slow
component
to the peak current investigated
in a sample of 100 slPSCs recorded from a single granule neuron. Figure 3B shows the distribution
of the averaged
contribution
of the slow component
to slPSCs in our
global sample of 42 granule neurons.
CABA-Activated
Currents in Outside-Out
Patches
from Purkinje and Granule Neurons
We studied GABA-activated
Cl- currents in outsideout patches excised from Purkinje and granule neurons of rat cerebellar
slices. The fast application
of
brief (<2 ms) pulses of GABA (1 mM concentration
in
the extracellular
solution) combined
with TTX (1 PM)
elicited Cl-currents
in both Purkinjeand
granule neurons. However, whereas in Purkinje neurons we obtained peak Cl- currents ranging from 30 to 160 pA
(see an example in Figure 4A), in granule neurons the
channel
density was much lower, yielding currents
of only a few picoamperes
in the best cases. To overcome this limitation,
we used outside-out
nucleated
patches of granule
cells (see Experimental
Procedures), in which we could record currents from 50 to
700 pA. In 10 granule cells (Figure 4A), the fast application of GABA (1 mM) elicited Cl- currents with a fast
rise time (0.4 f 0.2 ms) and a rapid decay that resembled the slPSCs measured
in whole-cell
recordings
(Figure IA). In fact, from the analysis of the average of
5 repetitive fast GABA applications
on each nucleated
outside-out
patch, the decay of the Cl- currents was
best described by the sum of two exponentials
(Figure
4B), with a fast time constant of 3.7 + 1.82 ms and a

slow time constant of 102 f 48 ms. The contribution

of the slow component
to the peak amplitude
of the
average Cl-currents
in nucleated outside-out
patches
excised from granule neurons was 42% f 8.4%, with
a variability
of 27%-57%.

A I

% Slow component
Figure 3. VariabilityoftheContributionoftheSlowComponent
to SIPSC
(A) The distribution
of the percent slow component
derived from
the double-exponential
fitting of 100 slPSCs in a cerebellar
granule neuron.
(B) The distribution
of the percent
slow component
derived
from
the double-exponential
fitting
of the average
slPSCs in 42 cerebellar
granule
neurons.

PURKINJE

GRANULE

-r-1
zs

of Inward
Currents
of CABA (1 mM) on
Excised from a PurNucleated
Outsidea Granule
Neuron

In (A) 4-5 traces are shown superimposed;

(B) shows their average, with the decay fitted by a double exponential.
Above the superimposed
traces in (A)are shown thecurrents
generated
by the liquid
junction
potential
due to a 50~1 dilution
of the
CABA-containing
solution
measured
after
blowing
out the patch. This gives an indication of the duration
of the pulse application. Calibration
bars apply to all traces,
but not to the 10 pA current
pulse.

100 pA

3ms
T s 70 ms

Zf

31ms

% Slow

Figure 4. Comparison
Elicited by Brief Pulses
an Outside-Out
Patch
kinje Neuron and on a
Out Patch Excised from

9
i

% Slow

In outside-out
patches from Purkinje neurons (n =
6), however, the decay of currents produced
by GABA
pulses did not completely
match the slPSC decay in
these neurons because of the presence of a slowly
decayingcomponent.Fromtheanalysisoftheaverage
of 5 repetitive GABA pulses on each outside-out
patch
(Figure 4A), the decay of the Cl- currents was best
described by the sum of two exponentials
(Figure 4B),
with a fast time constant of 5.1 f 2.3 ms and a slow
time constant of 95 f 36 ms. However, in Purkinje
cells, the contribution
of the slow component
to the
peak amplitude
of the average Cl- currents was only
10% + 7.4%, ranging from 0 to 25%. The rise time of
these averaged Cl- currents in Purkinje neurons was
0.7 + 0.4 ms.
Fast Desensitization
of the GABAA Receptor
To characterize
further the differences
in the intrinsic
properties of GABA* receptors in outside-out
patches
excised from cerebellar
neurons and to study the response to GABA pulses (<2 ms), we also investigated
the Cl- currents elicited by a 1 mM GABA step lasting
for a few seconds (Figure 5A). In 14 Purkinje neurons,
the step applications
of CABA elicited Cl- currents
with a fast rise time (0.5 f 0.3 ms) and a biphasic decay
(Figure 5). In these cells the fast component
of Clcurrent decay had a time constant of 5.6 of:2.7 ms and
an initial slow time constant of 158 f 97 ms, followed
by a sustained plateau (which we did not investigate).
The contribution
of the slow component
to the peak
amplitude
of the average Cl- currents was 58% +
14.9%.
In 18 nucleated
outside-out
patches excised from
granule cells, step applications
of GABA also produced Cl- currents with a fast rise time (0.4 f 0.2 ms)
and a biphasic decay (Figure 5B). In these cells the
fast component
of the Cl- current decay had a time

34

constant of 5 + 2.1 ms and an initial slow time constant

of 184 f 173 ms, also followed
by a sustained plateau
(which we did not investigate).
The contribution
of
the slow component
to the peak amplitude
of the
average Cl- currents was 73% f 14.9%.
In 4additional
Purkinje and 3 granule cells, the step
application
of GABA had a slower onset, as measured
at the end of the recording
(see Experimental
Procedures),therebyproducingCI-currentswitharisetime
>5 ms. In these patches, the current showed only a
slow decaying component,
followed by the sustained
plateau (data not shown).
slPSCs in Nucleated Outside-Out
Patches
from Granule Neurons
In 5 outside-out
nucleated
patches, we noticed that,
after the excision, we could detect an increase of the
background
noise and record clearly distinct slPSCs
by bringing the patch once.again close to the surface
oftheslicewherethecell
bodyofthegranuleneurons
was located (Figure 6). These currents ranged from 10
to 150 pA and were reversibly blocked by bicuculline
methiodide
(20 PM; data not shown). Perhaps they
reflected the spontaneous
vesicular release of GABA
from the synaptic terminals
in the slices being exposed by the cell enucleation
procedure.
The kinetics
of these sIPSCS in the nucleated patches remarkably
resembled those of the sIPSCS recorded in the wholecellvoltage-clampconfiguration.Theirdecaywas
best
described by the sum of two exponentials,
with a fast
time constant of 7.5 + 0.9 ms and a slow time constant
of 69 f 23.1 ms. In each patch, the slow component
of sIPSC decay was always present, and its contribution to the peak current was the same as in the sIPSCS
recorded in the whole-cell
mode, ranging from 15%
to 100% of the total peak amplitude,
with an average
of 60 + 23.2 pA.

CABA-Mediated

Currents

in Cerebellar

Slices

121

PURKINJE

Figure 5. Comparison
of Inward Currents
Elicited by Concentration
Steps of CABA
(1 mM) on an Outside-Out
Patch Excised
from a Purkinje
Neuron
and on a Nucleated Outside-Out
Patch Excised
from a
Granule
Neuron

GRANULE

-J

50 ms

B
Zf

Ts

6.6 ms
59 ms

% Slow

46

Diazepam Differentially
Affects the Time Course
of slPSCs in Purkinje versus Granule Neurons
In 7 Purkinje and 6 granule neurons, we investigated
the action of diazepam (IO PM) on the time course
of slPSCs recorded
in voltage-clamped
neurons. As
previously described, in both cell types these currents
had very variable peak amplitudes.
Therefore, we did
not investigate
the diazepam
effect on slPSC peak
amplitude.
Upon bath perfusion
with a saturating
concentration
of the benzodiazepine,
we noticed a
prolongation
of the time course of the synaptic currents in both Purkinje and granule neurons that fully
recovered after prolonged
perfusion with the control
solution
(data not shown). As shown in Figure 7, a
detailed analysis of the decay time of the average of
50 slPSCs in the cerebellar
neurons investigated
revealed that, whereas in Purkinje cells (n = 7) the prolongation of the average current was achieved by increasing the single decay time constant from 10 f 1
to 18 f 0.6, in granule cells (n = 6) the fast decay time
constant increased from 4 f 1.7 to 8 * 2.5, but the slow
component
remained unchanged
after the diazepam
application
(48 f IO before and 47 f 7 after diazepam).
In granule neurons, the relative contribution
of the
slow component
to slPSC peak amplitude
was also
not affected by diazepam (71 f 15 before and 75 k
15 after, n = 6). Recently, furosemide
(500 PM) was
shown to cancel selectively the fast component
of
IPSCs in hippocampal
neurons (Pearce, 1993). However, the slPSCs from both cerebellar
neurons failed
to be affected by furosemide
(500 PM, n = 4 Purkinje
neurons and n = 5 granule neurons; data not shown).
Discussion
In the mammalian cerebellum,
the synaptic inhibition
impinging
on cell bodies and dendrites
of Purkinje
neurons
originates
from basket and stellate inter-

50

ms

In (A) 3 traces are shown superimposed;

(B)
shows their average, with the decay fitted
by a double exponential.
Above the superimposed
traces in (A) are shown
the currents generated
by the liquid junction
potential
due to a 5O:l dilution
of the
GABA-containing
solution
measured
after
blowing out the patch. This gives an indication of the duration
of the pulse application. Calibration
bars apply to all traces,
but not to the 10 pA current
pulse.

Cf
4.8ms
Ts 107ms
% Slow 55

neuron terminals (Palay and Chan-Palay, 1974; Somogyi et al., 1989), whereas the inhibitory
Golgi type II
interneuron
terminals
are afferents to the granule
neuron dendrites
in cerebellar
glomeruli
(Palay and
Char+Palay, 1974; Somogyi et al., 1989). Additional
inhibitory synapses to granule cell dendrites
have also
been shown to arise from axon collaterals of Purkinje
neurons (Palay and Chan-Palay, 1974) as well as from
GABAergic
neurons
located in the deep cerebellar
nuclei (Hamori and Takacs, 1989). Our results show
a functional
diversity between inhibitory
synapses in

B
C

Figure
6. A Comparison
of slPSCs Recorded
in a VoltageClamped
Granule
Neuron
in Rat Cerebellar
Slices Before and
After Excison of a Nucleated
Outside-Out
Patch
(A)slPSCsinagranuleneuron.(B)Thecurrenttracea~erexcision
of the outside-out
nucleated
patch from the granule
neuron
shown in (A). (C) slPSCs produced
in the outside-out
nucleated
patch shown
in (B) by repositioning
the patch where the cell
body was located before excision.

Neuron

122

Figure 7. Whole-Cell
Recordings
of slPSCs
Recorded
from Cerebellar
Neurons
in the
Presence
and Absence
of Diazepam

GRANULE

+ Diazepam

(A) slPSCs in a Purkinje

neuron
(left) and
in a granule
neuron
(right).
(B) The effect
of bath perfusion
with diazepam
(IO FM).
(C) Superimposed
average
of 20 synaptic
currents
in the two experimental
conditions. Averaged
currents
are scaled to the
same peak amplitude.
The vertical
calibration bars in (C) refer to the currents
in the
presence
of diazepam.

10 FM

B
50 pA
500 Ins

C
150 pA

20pA
I
70 ms

voltage-clamped
was documented

Purkinje and granule neurons, which

by the slPSC kinetic characteristics.

Kinetically Distinct slPSCs in Cerebellar Neurons

slPSCs have been previously
described
in voltageclamped Purkinje neurons (Farrant and Cull-Candy,
1991; Vincent et al., 1992). We confirm these results
by showing that in these neurons slPSCs had a decay
time constant of 7-12 ms. To our knowledge,
there
are not reports of voltage-clamp
studies of slPSCs in
cerebellar
granule neurons. Our data show that the
decay of slPSCs in these neurons is best described
by two exponentials,
with time constants of 7 and 60
ms, with a variable contribution
of the slower decay
component
to the total peak amplitude.
These results
indicate an intriguing
similarity between slPSC decay
in granule neurons of the cerebellum
and granule
neurons of the hippocampus
(Edwards et al., 1990).
Vincent et al. (1992), discussing
the differences
between their resultson
slPSCs in Purkinje neuronsand
those of Edwards et al. (1990) on slPSCs in hippocampal granule
neurons,
proposed
that these kinetic
differences
might be related to distinct intrinsic properties of the GABAA receptor channels in thetwo preparations. We believe that an analogous
interpretation
could be extended to the difference we observed between GABAergic currents generated
in Purkinje and
granule neuron synapses. In fact, the poor correlation
of rise and decay times versus amplitude
that we observed for slPSCs in cerebellar
granule neurons (Figure 2) was also shown for slPSCs recorded
in hippocampal granule neurons by Edwards et al. (1990), who
considered
it to indicate that slPSCs were generated
in close proximity to the cell body in a region under
adequate voltage clamp. Presumably a similar consideration also applies to cerebellar
granule neurons, in
which synapses are located on the short dendrites

70 rns

(Palay and Chan-Palay, 1974). However, recent modeling of space-clamp

errors associated with whole-cell
recording
of synaptic currents
in neurons
revealed
the possibility
that in some cases the correlation
of
kinetic parameters
may provide misleading
conclusions (Spruston et al., 1993). In any case, total electrotonic length measurements
and recording of excitatory
synaptic currents support the notion that a high quality space clamp can be achieved
in whole-cell
recordings from cerebellar
granule neurons (Silver et
al., 1992).
The proposal that the presence of fast and slow decay components
in slPSCs in granule
neurons
is
related to distinct intrinsic properties
of the GABAA
receptor channels is further supported
by the observation that in a nucleated patch excised from a granule
neuron, in which the voltage control is optimal, the
slPSCs were similar to those recorded
in the wholecell mode.
slPSC Decay Times Relate to the Properties
of CABA Receptor Channels
In GABAergic
synaptic transmission,
the decay of
slPSCs is dependent
upon a complex series of events.
Neurotransmitter
diffusion
and reuptake,
together
with the kinetic properties
of the postsynaptic
channels, might concur to determine
the time course of
the synaptic current. To dissect out the various components
that determine
the slPSC decay, we attempted
to mimic synaptic transmission
by using
GABA pulses to outside-out
membrane
patches excised from cerebellar
neurons. In doing so, we assumed that the transmitter
concentration
which produces an slPSC is in the millimolar
range and that the
duration of the vesicular release at the synapse is very
short (- 1 ms) and similar to that recently calculated
for the vesicular
release of glutamate
from gluta-

CABA-Mediated
123

Current5

in Cerebellar

Slices

matergic synapses (Clements

et al., 1992). Furthermore, we also surmised that the properties
of the
CABA* receptors in the outside-out
patches excised
from the cell body were similar to those of the channels operating
at synaptic sites. While these assumptions are not yet validated,
our results show that in
granule neurons there is a good correlation
between
the kinetic properties
of slPSCs and those of the currents evoked by GABA pulses to outside-out
nucleated patches. In fact, in these neurons the fast Clcurrents activated by GABA decayed with two exponential components,
with kinetics and relative proportion of the slow component
similar to those of
the slPSCs. In Purkinje neurons, however, while we
confirmed
the results of Vincent et al. (1992) showing
the absence of slPSCs with biphasic decay, we described a biphasic decay of the Cl-currents
generated
by GABA pulses on outside-out
patches. The relative
proportion
of the slow decay component
never exceeded 25% of the total peakamplitude.
Distinct properties of synaptic and extrasynaptic
GABAA receptors
could be responsible
for the differences
observed between slPSCs and channel
currents
recorded
in
patches from Purkinje
neurons,
although
this proposal requires further investigation.
In any case, our
results indicate that, at least for granule neurons, distinct relaxation properties
of GABAA receptor channel
currents in outside-out
patches approximate
the distinctkineticsobservedintheaveragesIPSCandtherefore lend support to the hypothesis that functionally
diverse CABAA receptor subtypes underlie the properties of GABA-mediated
inhibitory
synaptic currents.
In addition,
we cannot rule out the possibility
that
synaptic and extrasynaptic
GABA,+ receptors are distinct also in granule neurons and that in these neurons
the slow slPSC decay component
results from synaptically released GABA on extrasynaptic
receptors.
Fast Desensitization
of CABAA Receptors
To investigate further the properties
of GABA* receptors expressed
in cerebellar
neurons, we compared
theCI-current
induced by sustained stepapplications
of high GABA concentrations
to patches excised from
Purkinje and granule neurons. An important
result of
this experimental
protocol was that the Cl- current
desensitized
during the GABA application
(see Figure
5). The GABAA receptor desensitization
was similar
in both cell types. A slow time course (hundreds
of
milliseconds)
of desensitization
of the GABA,+ receptor response has been described
previously
(Mierlak
and Farb, 1988; Oh and Dichter, 1992; Frosch et al.,
1992). We report a very rapid (a few milliseconds)
component of desensitization
for the GABAA receptor response induced
by a high concentration
of GABA,
which is followed
by the previously described slower
component
and a sustained plateau. We, as well as
others, failed to observe a fast desensitizing
component when the onset of the GABA concentration
step
was longer lasting (a few milliseconds).
A slower rising
GABA step might be the reason why a fast GABAA

receptor
desensitization
was not described
previously. In fact, one may suggest that a subgroup
of
the receptor population
enters into a desensitized
state during the slow onset of GABA application,
preventing the observation
of the fast desensitizing
response. This finding is reminiscent
of the fast desensitization of the response to excitatory amino acids
(Tang et al., 1989; Trussell and Fischbach, 1989; Mayer
and Vyklicky, 1989). We do not know whether
the
fast desensitization
observed during sustained GABA
steps also occurs during the physiological
release of
GABA at inhibitory
synapses. Before this question can
be addressed, the concentration
dependence
limits
of fast desensitization
as well as the GABA concentration at the synapses has to be thoroughly
investigated.
Our results, however, indicate that GABA response
desensitization
is similar in Purkinjeand
granulecells.
In conclusion,
from the experiments
in excised membrane patches, it is evident that the onset and the
duration of the GABA application
are crucial determinants of the kinetics of the Cl- current. As a consequence, at inhibitory
synapses of granule but not of
Purkinje neurons, the slow component
in the decay
phase of slPSCs might alternatively
relate to distinct
relaxation properties
of the postsynaptic
channels or
to a kinetically distinct decrease in neurotransmitter
concentration
regulated by GABA reuptake (Isaacson
et al., 1993).
Different Subtypes of GABA* Receptors
in Cerebellar Synapses
Since in granule neurons we observed both fast and
slow GABA-activated
Cl- currents, one might surmise
that they are the product of a mixed activation of two
GABAA receptor populations.
This is also supported
by the wide distribution
of the slow decay contribution to the total synaptic currents recorded from different granule
cells. Kinetically
distinct
GABA responses in Purkinjeand
granule neurons might derive
from GABAA receptor structural diversity (reviewed in
Olsen and Tobin, 1990, and Burt and Kamatchi, 1991),
which determines
specific biophysical
properties
in
the associated ion channel. Various studies (Bovolin
et al., 1992; Wisden et al., 1992; Thompson et al., 1992)
have shown that many subunits are common to granule and Purkinje neurons. However, the a6 subunit
is uniquely expressed by the granule cells (Wisden et
al., 1992). Interestingly,
in hippocampal
granule neurons, in which slPSCs also have a double exponential
decay (Edwards et al., 1990; but see Otis and Mody,
1992), the localization
of a4 subunits, which possess
a high degree of homology
with a6 subunits (Ymer
et al., 1989; Liiddens et al., 1990), has been reported
(Wisden et al., 1992; Laurie et al., 1992). The 8 subunit
mRNA is also selectively present in both cerebellar
and hippocampal
granule neurons (Laurieet al., 1992).
Slow GABA-activated
Cl-currents,
however, could be
independent
from the subunit assembly of the receptor and could be instead consequent
to reversible
posttranslational
modifications,
such a phosphoryla-

124

tion (Swope et al., 1992). In support of these alternatives, we observed a small proportion
of slower relaxing GABA-activated
Cl- current
in patches from
Purkinje
neurons, in which neither the a6 nor the 6
subunit has been detected. Further work with recombinant GABAA receptors will be required to determine
whether the different
functional
profiles detected at
cerebellar
inhibitory
synapses are related to the molecular structure of receptors specifically
located in
certain neuronal
populations.
Allosteric Regulation of GABAA Receptors
in Cerebellar
Neurons
An important
property of the GABAA receptor is that
it is allosterically
regulated
by clinically
important
compounds,
such as benzodiazepines
(Costa et al.,
1975). In this regard, it has been demonstrated
that
the molecular
composition
of the receptor is strongly
related to the potency and efficacy of the allosteric
modulation
by benzodiazepines
and betacarboline
derivatives
(Pritchett et al., 1989a, 1989b; Puia et al.,
1991). Therefore,
diversity in GABAA receptor responsiveness to benzodiazepines
arises from the structural composition
of these receptors. Among these
distinct benzodiazepine
receptor types (reviewed by
Doble and Martin, 1992), those lacking they2 subunit
as well as those comprising
the a6 or 6 subunit have
very poor sensitivity
to diazepam
(Pritchett
et al.,
1989a; Shivers et al., 1989; Puia et al., 1991; Kleingoor
et al., 1991; Ducic et al., 1993; Kleingoor
et al., 1993).
Our results, provide evidence of a selective prolongation of the fast slPSCs decay by diazepam in Purkinje
neurons.
The lack of effect of diazepam on slowly
decaying slPSCs in granule neurons indicates a specificity in the allosteric modulation
of GABAA receptors
located in postsynaptic
sites in granule neurons, probably related to a specific subunit combination
that
confers insensitivity
to benzodiazepines.
Our resultsdemonstrate
IPSCs in cerebellargranule
neuron with different
proportions
of fast and slow
components.
This functional
difference
might allow
inhibitory
interneurons
in the granule layer to produce specific patterns of inhibition
and to select granule cell groups. In s~.~pport of this view, Eccles et al.
(1967) demonstrated
the existence of nonoverlapping
compartments
of inhibited
granule cells, similar to
the glomeruli
of the olfactory bulb and the barrels in
the neocortex,
with structural and functional
unity.
Experimental

Procedures

Brain Slices
Sagittal slices of cerebellum
(200-300 pm) were prepared
from
14-to la-day-old
Sprague-Dawley
rats as described
by Edwards
et al. (1989). Cerebellar
neurons
were viewed
with an upright
microscope
equipped
with differential
interferencecontrast
Nomarski
optics
(UEM, Zeiss, Federal Republic
of Germany)
and
an electrically
insulated
water immersion
40x objective
with a
long working
distance
(2 mm).
Solutions and Drugs
Experiments
were performed

at room

temperature

(22C-24OC)

using an extracellular
medium
composed
of 120 mM NaCI, 3.1
mM KCI, 1.25 mM K2HPO+ 26 mM NaHCOI,
5.0 mM dextrose,
1.0 mM MgCI?, and 2.0 mM CaClz and containing
10 PM 6 cyano7-nitroquinoxaline-2,3dione
(CNQX,
Tocris,
UK) and 20 PM
3-[( f )-2carboxypiperazin4yl]-propyl-I-phosphonic
acid (CPP)
to block excitatory
amino acid-mediated
synaptic
transmission.
The solution
was maintained
at pH 7.4 by bubbling
with 5% COZ,
95% 0,. The slice was completely
submerged
in a total volume
of 500 ~1 and continuously
perfused
at a rateof 5 ml/min. Bicuculline methiodide,
TTX (Sigma, St. Louis, MO), diazepam
(in 0.1%
dimethylsulfoxide;
a gift of the late Dr. Haefely, Hoffman
La Roche
Laboratories,
Basel, Switzerland),
and furosemide
solution
(10
mg/ml; Abbott Laboratories,
IL) were diluted in the extracellular
medium
and were superfused
through
parallel
inputs to the
perfusion
chamber
until effective
replacement
of the solution
was obtained.
For fast application
of CABA, we used a piezoelectric
translator
(P-24530Stacked
Translator,
Physik Instrumente,
Federal Republic of Germany)
to position
double
barrel theta tubing
in front
of the excised
patch quickly.
One barrel contained
extracellular
medium
with added TTX, and the other contained
this solution
and 1 mM CABA, similar to the solution
described
for fast glutamate application
by Lester and Jahr (1992) and Colquhoun
et al.
(1993). After each patch recording,
on and off rates as well as
pulse duration
were measured
by blowing
out the patch and
recording
currents
generated
by the liquid junction
potential
due to a 5O:l dilution
of the CABA-containing
solution
(Lester
and Jahr, 1992). On and off rates of the system
were typically
less than 0.2 ms.
Outside-out
patches
from Purkinje
neurons
were excised followingtheproceduredescribed
byHamilletal.(1981).
Nucleated
outside-out
patches were instead isolated following
the detailed
protocol
described
for mouse forebrain
neurons
in primary
culture by Sather et al. (1992).
Electrophysiology
Voltage-clamp
recordings
of sIPSCS were performed
using the
whole-cell
recording
configuration
of the patch-clamp
technique (Hamill et al., 1981) with a patch-clamp
amplifier
(EPC 7,
List Electronics,
Darmstadt,
Federal Republic
of Germany)
after
capacitance
and series resistance
compensation.
Series resistance was checked
for constancy
throughout
the experiments.
Electrodes
were pulled from borosilicate
glass capillaries
(Wiretrol II, Drummond,
Broomall,
PA) and were filled with a solution
containing
145 mM CsCI, 1 mM MgCI,, 5.0 mM ECTA, 2.0 mM
Na-ATP,
and 10 mM HEPES (to pH 7.2 with CsOH).
We chose
Cs+ as the major cation to improve
the quality
of the voltage
clamp
and to prevent
GABAs
receptor-mediated
currents.
GABA-activated
currents
were recorded
in outside-out
patches
excised from neurons
in the slicewith
a patch pipettecontaining
the same solution
as for the whole-cell
recording
experiments.
Current
traces from whole-cell
and outside-out
patches
were
recorded
on a VR-10 data storage system (Instrutech
Co., Haverhill, MA).
Data Analysis
Current
traces were filtered
at 3 kHz (-3 dB, 8 pole, low pass
Bessel filter; Frequency
Devices)
and stored in an LSI II/73 computer (INDEC System, Sunny Vale, CA) after digitization
(IO kHz)
with a Data Translation
analog to digital converter.
Decay time
constants
of sIPSCS and CABA-activated
currents
were determined from exponential
fitting with the 11173 system by using
an entirely
automated
least squares
procedure
(see Vicini and
Schuetze,
1985, for further
details). This method
uses a Simplex
algorithm
(Caceci and Cacheris,
1984) to fit the data to either a
singleor double-exponential
equation
of the form
I(t) = I,
exp(-t/G
+ IJ exp(-t/r,),
where
If and I, are the amplitudes
of
the slPSC fast and slow components,
and G and T. are their
respective
decay time constants.
Peak amplitudes
were measured at the absolute
maximum
of the currents,
taking into account the noise of the baseline and noise around
the peak. Rise
times were measured
as the time elapsed
from 20% to 80% of
the peak amplitude
of the response

GAEA-Mediated
125

Currents

in Cerebellar

Slices

Acknowledgments
We wish to thank Charles T. Livsey for comments
on the manuscript.
This work
was supported
in part by NIMH
grant #ROl
MH49486-OIAI.
The costs of publication
of this article were defrayed
in part
by the payment
of page charges.
This article must therefore
be
hereby marked advertisement
in accordance
with 18 USC Section 1734 solely to indicate
this fact.
Received

June

14, 1993; revised

October

28, 1993.

aminobutyric
and triple
antibodies.

acid receptors
immunofluorescence
Proc. Natl. Acad.

Frosch, M. P., Lipton,

tion of CABA-activated
neurons.
J. Neurosci.

identified
in neurons
by double
staining
with subunit
specific
Sci. USA 89, 6726-6730.

S. A.,and Dichter,
M.A.
currents
and channels
72, 3042-3053.

(1992). Desensitizain cultured

cortical

Hamill, 0. P., Marty, A., Neher, E., Sakmann,

B., and Sigworth,
F. J. (1981). Improved
patch-clamp
techniques
for high resolution
current
recording
from cells and cell-free
membrane
patches.
Pflijgers
Arch. 397, 85-100.
Hamori,
J., and Takacs, J. (1989). Two types of GABA-containing
axon terminals
in cerebellar
glomeruli
of cat: an immunogold
study. Exp. Brain Res. 74, 471-479.

References

R. A. (1993).
hippocampus.

Baude, A., Sequier,

J. M., McKernan,
R. M., Olivier,
K. R., and
Somogyii,
P. (1992). Differential
subcellular
distribution
of the
a6 subunit
versus the al and p2/3 subunits
of the GABAA/benzodiazepine
receptor
complex
in granule
cells of the cerebellar
cortex.
Neuroscience
57, 739-748.

Isaacson,
J. S., Solis, J. M., and Nicoll,
diffuse
synaptic
actions
of CABA in the
70, 165-175.

Bovolin,

Kleingoor,
C., Ewert, M., von Blankenfeld,
C., Seeburg,
P. H.,
and Kettenmann,
H. (1991). Inverse but not full benzodiazepine
agonists
modulate
recombinant
a6P2y2 CABAA receptors
in
transfected
human embryonic
kidney cells. Neurosci.
Lett. 730,
169-172.

P., Santi, M. R., Puia, G., Costa, E., and Grayson,

D. R.
Expression
patterns
of y-aminobutyric
acid A receptor
subunit
mRNAs
in primary
cultures
of granule
neurons
and
astrocytes
from neonatal
rat cerebella.
Proc. Natl. Acad. Sci. USA

(1992).

89, 9344-9348.
Burt,
from

D. R., and Kamatchi,

G. L. (1991). GABAA receptor
subtypes:
pharmacology
to molecular
biology.
FASEB J. 5,2916-2923.

Caceci, M. S., and Cacheris,

Byte 9, 340-362.

W. P. (1984).

Fitting

curves

to data.

Clements,
J. D., Lester, R. A. J., Tong, C., Jahr, C. E., and Westbrook, G. L. (1992). The time course
of glutamate
in the synaptic
cleft. Science 258, 1498-1501.
Collingridge,
G. L., Gage, P. W., and Robertson,
tory post-synaptic
currents
in rat hippocampal
J. Physiol. 356, 551-564.

8. (1984). InhibiCA1 neurones.

Colquhoun,
D., Jonas, P., and Sakmann,
B. (1993). Action of brief
pulses of glutamate
on AMPA/kainate
receptors
in patches
excised from different
neurones
of rat hippocampal
slices. J. Physiol. 458, 261-287.
Costa,
amino

E. (1989). Allosteric
modulatory
centers
acid receptors.
Neuropsychopharmacology

of transmitter
2, 167-174.

Costa, E., Guidotti,

A., and Mao, C. C. (1975). Evidence
for the
involvement
of GABA in the action of benzodiazepines:
studies
in rat cerebellum.
In Mechanism
of Action of Benzodiazepines,
E. Costa and P. Creengard,
eds. (New York.: Raven Press), pp.
113-130.
Cull-Candy,
S. G., and Usowicz,
M. M. (1987). Glutamate
and
aspartate
activated
channels
and inhibitory
synaptic
currents
in
large cerebellar
neurons
grown
in culture.
Brain Res. 402, 182187.
Doble, A., and Martin,
I. L. (1992). Multiple
benzodiazepine
receptors:
no reason for anxiety. Trends Pharmacol.
Sci. 73, 76-81.
Ducic,
I., Puia, C., Mereu,
G., Vicini, S., and Costa, E. (1993).
Triazolam
is more efficacious
than diazepam
in a broad spectrum
of recombinant
CABA* receptors.
Eur. J. Pharmacol.
244,29-35.
Eccles, J. C., Ito, M., and Szentagothai,
J. (1967). The Cerebellum
as a Neuronal
Machine
(Berlin: Springer-Verlag).

Ito, M. (1984).The
Raven Press).

Cerebellum

and Neuronal

Control

Korpi,

(1993).
point
359.

Laurie, D. J., Seeburg, P. H., and Wisden,

W. (1992). The distribution of thirteen
CABAA receptor
subunit mRNAs in the rat brain.
II. Olfactory
bulb and cerebellum.
J. Neurosci.
12, 1063-1076.
Lester, R. A. J., and Jahr, C. E. (1992). NMDA channel
depends
on agonist affinity.
J. Neurosci.
72, 635-643.

Mayer, M. L., and Vyklicky,

L., Jr. (1989). Concanavalin
A selectively reduces
desensitization
of mammalian
neuronal
quisqualate receptors.
Proc. Natl. Acad. Sci. USA 86, 1411-1415.
Mierlack,
D., and Farb, D. H. (1988). Modulation
of neurotransmitter
receptor
desensitization:
chlordiazepoxide
stimulates
fading of the CABA response.
J. Neurosci.
8, 814-820.
Oh, D. J., and Dichter,
M. A. (1992). Desensitization
of GABAinduced
currents
in cultured
rat hippocampal
neurons.
Neuroscience 49, 571-576.
Olsen, R. W., and Tobin, A. J. (1990). Molecular
receptors.
FASEB J. 4, 1469-1480.
Otis, T. S., and Mody,
frequency
of GABAA
postsynaptic
currents
49, 13-32.

213-249.
Farrant, M., and Cull-Candy,
S. G. (1991). Excitatory
receptor-channels
in Purkinjecells
in thin cerebellar
Roy. Sot. Lond. (B) 244, 179-184.
Fritschy,
J. M., Benke, D., Mertens,
S., Oertel,
and Mohler,
H. (1992). Five subtypes
of

amino acid
slices. Proc.

W. H., Bachi, T.,

type
A gamma-

biology

of GABAA

I. (1992). Modulation
of decay kinetics and
receptor-mediated
spontaneous
inhibitory
in hippocampal
neurons.
Neuroscience

Pearce,
GABAn

Edwards,
F. A., Konnerth,
A., and Sakmann,
B. (1990). Quanta1
analysis of inhibitory
synaptic
transmission
in the dentate gyrus
of rat hippocampal
slices: a patch-clamp
study. 1. Physiol. 430,

behavior

Ltiddens,
H., Pritchett,
D. B., Kohler, M., Killisch,
I., Keinanen,
K., Monyer,
H., Sprengel,
R., and Seeburg, P. H. (1990). Cerebellar
GABAA receptor
selective
for a behavioral
alcohol
antagonist.
Nature 346, 648-651.

synaptically
connected
system.
Pfliigers
Arch.

A., Sakmann,
B., and Takahashi,
T.
for patch clamp recordings
from
neuronesof
mammalian
central nervous
474, 600-612.

York:

E. R., Kleingoor,
C., Kettenmann,
H., and Seeburg,
P. H.
Benzodiazepineinduced
motor
impairment
linked
to
mutation
in cerebellar
CABA* receptor.
Nature 367, 356-

Palay, S. L., and Chan-Palay,

V. (1974). Cerebellar
ogy and Organization
(Berlin: Springer-Verlag).

F. A., Konnerth,

(New

Kleingoor,
C., Wieland,
H. A., Korpi, E. R., Seeburg,
P. H., and
Kettenmann,
H. (1993). Current
potentiation
by diazepam
but
not GABA sensitivity
is determined
by a single histidine
residue.
NeuroReports
4, 187-190.

(1989). A thin slice preparation

Edwards,

Local and
Neuron

R. A. (1993). Physiological
responses
in rat hippocampus.

evidence
Neuron

Cortex,

Cytol-

for two distinct

70, 189-200.

Pritchett,
D. B., Sontheimer,
H., Shivers,
B., Ymer, S., Kettenmann, H., Schofield,
P. R., and Seeburg, P. H. (1989a). Importance
of a novel GABAA receptor
subunit
for benzodiazepine
pharmacology.
Nature 338, 582-585.
Pritchett,
D. B., Liiddens,
H., and Seeburg,
P. H. (198913). Type
I and type II GABArbenzodiazepine
receptors
produced
in
transfected
cells. Science 245, 1389-1392.
Puia, G., Vicini,
of recombinant

S., Seeburg,
y-aminobutyric

P. H., and Costa, E. (1991). Influence

acid A receptor
subunit
compo-

NWJKT

126

sition on the action of allosteric

modulators
of y-aminobutyric
acid-gated
Cl- currents.
Mol. Pharmacol.
39, 691-696.
Ropert, N., Miles, R., and Korn,
tureinhibitorypostsynapticcurrentsinCA1
of rat hippocampus.
J. Physiol.

H. (1990). Characteristic
pyramidal
428, 707-737.

of minianeurones

Sather, W., Dieudonne,

S., MacDonald,
j. F., and Ascher,
P.
(1992). Activation
and desensitization
of N-methyl-o-aspartate
receptors
in nucleated
outside-out
patches from mouse neurones.
J. Physiol. 450, 643-672.
Schneggenburger,
R., Lopez-Barneo,
J., and Konnerth,
A. (1992).
Excitatory
and inhibitory
synaptic
currents
and receptors
in medial septal neurones.
J. Physiol. 445, 261-276.
Segal, M., and Barker,
culture:
voltage-clamp
tions. J. Neurophysiol.

J. L. (1984). Rat hippocampal

neurons
in
analysis
of inhibitory
synaptic
connec52, 469-487.

Shivers,
B. D., Killisch, I., Sprengel,
R., Sontheimer,
H., Kohler,
M., Schofield,
P. R., and Seeburg,
P. H. (1989). Two novel GABAA
receptor
subunits
exist in distinct
neuronal
subpopulations.

Neuron

3,327-337.

Siegel, R. E. (1988). The mRNA encoding

GABAAIbenzodiazepine
receptor
subunits
are localized
in different
cell populations
the bovine cerebellum.
Neuron
7, 579-584.

of

Sieghart, W. (1992). GABAA receptors:

ligand-gated
Cl- ion channels modulated
by multiple
drug-binding
sites. Trends
Pharmacol. Sci. 73, 446-450.
Silver, R. A., Traynelis,
S. F., and Cull-Candy,
S. G. (1992). Rapidtime-course
of miniature
and evoked
excitatory
currents
at cerebellar synapses
in situ. Nature 355, 163-166.
Sivilotti, L., and Nistri, A. (1991). CABA receptor
mechanisms
the central
nervous
system.
Prog. Neurobiol.
36, 35-92.

in

Somogyi,
P., Takagi, H., Grayson,
R. J., and Mohler,
H. (1989).
Subcellular
localization
of benzodiazepine/GABAA
receptors
in
the cerebellum
of rat, cat and monkey
using monoclonal
antibodies. J. Neurosci.
9, 2197-2209.
Spruston,
N., Jaffe, D. B., Willims,
Voltageand space-clamp
errors
ments of electrotonically
remote
iol. 70, 781-802.

S. H., and Johnston,

D. (1993).
associated
with the measuresynaptic
events. J. Neurophys-

Swope,
S. L., Moss,
S. J., Blackstone,
C. D., and
R. L. (1992). Phosphorylation
of ligand-gated
channels:
mode of synaptic
plasticity.
FASEB J. 6, 2514-2523.

Huganir,
a possible

Tang, C. M., Dichter,

M., and Morad,
M. (1989). Quisqualate
activates
a rapidly
inactivating
high conductance
ionic channel
in hippocampal
neurons.
Science 243, 1474-1477.
Thompson,
C. L., Bodewitz,
C., Stephenson,
F. A., and Turner,
J. D. (1992). Mapping
of GABA* receptor
a5, a6 subunit-like
immunoreactivity
in rat brain. Neurosci.
Lett. 744, 53-56.
Trussell,
L. O., and Fischbach,
G. D. (1989). Glutamate
desensitization
and its role in synaptic
transmission.
209-218.

receptor
Neuron
3,

Vicini, S., and Schuetze,

S. M. (1985). Cating properties
of acetylcholine
receptors
at developing
rat endplates.
J. Neurosci.
5,
2212-2224.
Vicini, S., Alho, H., Costa, E., Mienville,
J. M., Santi, M. R., and
Vaccarino,
F. M. (1986). Modulation
of y-aminobutyric
acidmediated
inhibitory
synaptic
currents
in dissociated
cortical
cell
cultures.
Proc. Natl. Acad. Sci. USA 83, 9269-9273.
Vincent,
P., Armstrong,
C. M., and Marty, A. (1992). Inhibitory
synaptic
currents
in rat cerebellar
Purkinje
cells: modulation
by
postsynaptic
depolarization.
J. Physiol. 456, 453-471.
Wisden,
W., Laurie, D. J., Monyer,
H., and Seeburg,
P. H. (1992).
The distribution
of 13 CABAA subunit
mRNAs in the rat brain.
I. Telencephanon,
diencephalon,
mesencephalon.
J. Neurosci.
72, 1040-1062.
Ymer, S., Draguhn,A.,
Kohler, M., Schoefield,
P. R., and Seeburg,
P. H. (1989). Sequence
and expression
of CABAA receptor
a4
subunit.
FEBS Lett. 258, 119-122