(Received for publication, January 21, 1997)
From the Max Planck Institute of Psychiatry, Clinical Institute, Kraepelinstrasse 10, 80804 Munich, Federal Republic of Germany
-Aminobutyric acid type A
(GABAA) receptor subtypes containing the
6-subunit are generally thought to be insensitive to the
action of benzodiazepine agonists. We describe the specific binding of
the benzodiazepine agonist flunitrazepam to
6
2
2-containing GABAA receptors, which has not been observed before and
differs from previous reports. With the whole-cell voltage-clamp
technique, we observed a functional discrimination between
1
2
2- and
6
2
2-receptors. Different
benzodiazepines had different effects on GABA-evoked chloride currents.
The agonist flunitrazepam had an inverse agonistic effect, whereas the
antagonist flumazenil increased GABA-induced chloride currents. The
action of flunitrazepam on the channel activity of
6
2
2-receptors was opposite
to its action on
1
2
2-receptors. We conclude
that flunitrazepam can act as either an agonist or an inverse agonist,
depending on the GABAA receptor configuration.
Benzodiazepines (BZs)1 act via the
-aminobutyric acid type A (GABAA) receptor, thereby
influencing the chloride influx into the cell (1, 2). Studies of a
variety of GABAA receptor compositions have led to the
conclusion that ternary receptors reconstituted from
-,
-, and
-subunits have binding sites for BZs, with the high-affinity site
being located on the
-subunit (3, 4). These subunits are
differentially distributed throughout the central nervous system (5,
6). It is generally thought that the subunit composition of the
GABAA receptor complex organizes the receptor's
pharmacology. The best known concept is that the
1
x
2-receptors, called type I
BZ receptors, preferentially bind full BZ agonists with high affinity
(3) and that the
2/3/5
x
2-receptors, classified
as type II BZ receptors, have a 10-fold lower affinity for BZ agonists
(7-10). The view is widely held that
6- and
4-containing receptors have virtually no affinity for
full BZ agonists (11-13), although they retain a high affinity for the
partial inverse BZ agonist Ro 15-4513 (14, 15). In the mature mammalian
brain, the
6-subunit is expressed exclusively in the
cerebellum (6), where it is preferentially found in granular cells
(16).
The studies we report here focus on the BZ pharmacology of the
6
2
2-receptor. We show that
6
2
2-GABAA
receptors have high affinity for the BZ agonist flunitrazepam, although
the histidine at position 100 in the ligand-binding domain, which is
thought to be crucial for BZ binding, is missing (13). Furthermore, using the whole-cell voltage-clamp technique on
6
2
2-transfected cells, we
demonstrate that flunitrazepam has an inverse agonistic effect on
GABA-induced chloride currents, whereas the BZ antagonist Ro 15-1788 (flumazenil) increases GABA-mediated chloride currents.
Human embryonic kidney cells
(HEK 293) were grown and transfected as described (17). In a
modification of that method, cells were cotransfected with different
rat -variants (
1,
1(Arg-101),
6, and
6(His-100)) in combination with
2- and
2-cDNAs (11, 13, 18), using an
electroporation system (Biotechnologies & Experimental Research Inc.,
San Diego, CA) after determination of the optimal electric field
strength (19). For optimal receptor expression, final concentrations
were as follows:
1,
1(Arg-101),
6, and
6(His-100), 3.3 µg of vector
DNA/10-cm tissue culture dish;
2, 16.6 µg; and
2, 0.3 µg.
HEK 293 cells were prepared as described previously (20). For saturation binding analysis with [3H]flunitrazepam or [3H]Ro 15-4513, concentrations between 1.2 and 40 nM were used. Binding experiments were performed in plastic microtiter plates using ~106 cells/well in a total volume of 0.25 ml. The membranes were incubated with the radioactive ligand at 4 °C for 1 h with continuous shaking. In competition experiments, the competing reagent was coapplied with the radioactive ligand. The binding reaction was terminated by rapid filtration through Whatman GF/C filters with a Titertek cell harvester (Skatron Instruments, Newmarket, United Kingdom). The filters were washed twice with 5 ml of ice-cold incubation buffer. Filter-retained radioactivity was determined by liquid scintillation counting. Nonspecific binding was determined in the presence of either 50 µM diazepam or 50 µM Ro 15-4513. Binding assays were analyzed as described (17).
Electrophysiological RecordingsHEK 293 cells were recorded in the whole-cell voltage-clamp configuration of the giga-seal technique (21) under visual control using an inverted microscope (Zeiss, Jena, Germany). The cells were kept in a bath solution containing 150 mM NaCl, 2.5 mM KCl, 1.25 mM NaH2PO4, 2 mM CaCl2, 2 mM MgSO4, 10 mM glucose, 20 mM sucrose, and 10 mM HEPES. The pH was adjusted to 7.4 with NaOH. Patch electrodes were pulled from borosilicate glass (Hilgenberg, Malsfeld, Germany) using a horizontal pipette puller (Zeitz Instruments, Augsburg, Germany) to yield pipettes with a resistance of 3-6 megaohms. Pipettes were filled with a solution containing 130 mM KCl, 15 mM NaCl, 2 mM MgCl2, 1 mM CaCl2, 2 mM ATP, 0.25 mM cAMP, 5 mM glucose, 10 mM HEPES, and 11 mM EGTA. The pH was adjusted to 7.25 with KOH.
After the whole-cell configuration had been established, the cells were lifted from the glass substrate, and GABA or BZs were applied at the indicated concentrations using a fast superfusion device. A piezo-translator-driven double-barreled application pipette was used to expose the lifted cell to either GABA-free or GABA-containing solution for control experiments (flow rate of 200 µl/min). A 2.5-s GABA pulse was delivered every 60 s. BZs were present at the indicated concentrations in both GABA-free and GABA-containing solutions so that their modulatory properties could be studied. Flunitrazepam and Ro 15-4513 were dissolved in ethanol and diluted with bath solution to the desired concentration. The ethanol concentration was 0.1% in all solutions to avoid any confounding solvent effects. Recordings were obtained from 6 to 12 independent experiments.
Current signals were recorded at a holding potential of 50 mV with an
EPC-9 amplifier (Heka, Lambrecht, Germany) using Pulse software on a
Macintosh II computer. The data were analyzed using PulseFit (Heka) and
IgorPro (Wacemetrics, Lake Oswego, OR) software.
Classical BZ agonists are generally thought to bind to all ternary
GABAA receptor complexes reconstituted from -,
-, and
-subunits, except when the
6- or
4-subunit is involved (13). This conclusion was drawn
indirectly from competition studies with classical BZ agonists as
competitors for the inverse agonist-binding site of
[3H]Ro 15-4513. A single histidine was found to be
essential for prototypic BZ agonist binding (13). We analyzed the
direct binding of [3H]flunitrazepam to recombinant
1
2
2- and
6
2
2-GABAA
receptors. The results for the recombinant
6
2
2-receptor were
unexpected. Scatchard analysis with [3H]flunitrazepam in
concentrations between 1.2 and 40 nM revealed a single
category of saturable high-affinity binding sites, with a
Kd value of 8.7 nM for
6
2
2-receptors compared
with a Kd value of 3.3 nM for
1
2
2-receptors (Fig.
1, A and B). With this finding for
6
2
2-receptors, we
postulate a binding site for flunitrazepam on this receptor subtype,
which has not been detected previously. Whereas
1
2
2-receptors are found
throughout the central nervous system,
6
2
2-receptors are
localized predominantly on cerebellar granule cells in the neighborhood of a variety of excitatory ion channels (12, 16).
In competition binding studies, the
1
2
2-receptors, which are
the "classical" GABAA receptors, exhibited the expected
displacement of flunitrazepam binding by other BZ agonists (11),
whereas for the
6
2
2-receptors, we obtained
Ki values 150-1000-fold higher than the observed
Kd (Table I). The weak
displacement of flunitrazepam binding by classical BZ agonists
suggested that we had found a binding site on the
6
2
2-receptor with a
preference for flunitrazepam. We then looked at whether the peripheral
BZ receptor ligand 4
-chlorodiazepam (Ro 5-4864), which interacts with
a unique low-affinity site on the GABAA receptor, and a
peripheral BZ receptor antagonist, the isoquinoline carboxamide
derivative PK 11195 (22), bound to the flunitrazepam-binding site. We
did not detect any flunitrazepam displacement (Table I). Finally, we
examined whether other BZ agonists, such as alprazolam and chlordiazepoxide, or BZ inverse agonists, such as
methyl-6,7-dimethoxy-4-ethyl-
-carboline 3-carboxylate,
n-butyl-
-carboline 3-carboxylate, and desmethyldiazepam, were able to compete for the flunitrazepam-binding site on the
6-subunit, but found only weak displacement or none at
all (Table I).
|
|
We also wanted to test whether the Ro 15-4513 binding site on
6
2
2 receptors had any influence on the
newly described flunitrazepam binding site. Analysis of
1(Arg-101)
2
2 and
6(His-100)
2
2 receptors along with the
wild-type receptors showed that the sensitivity of [3H]Ro
15-4513 binding to flunitrazepam as well as to diazepam coincided with
the presence of the crucial amino acid histidine (Table II). The newly
proposed high affinity binding site of flunitrazepam on
6
2
2 receptors did not compete with the
Ro 15-4513 binding site. The BZ antagonist flumazenil exhibited a
100-300-fold lower affinity for mutant receptors and
6
2
2 receptors than for
1
2
2 receptors which is in agreement
with data previously published (13).
To investigate whether an exchange of the crucial amino acid histidine,
responsible for high-affinity BZ agonist binding, would change the
flunitrazepam affinity for
6
2
2-receptors, we used two
mutants, already described elsewhere (13), for
[3H]flunitrazepam binding studies. These mutant
receptors,
1(Arg-101)
2
2 and
6(His-100)
2
2, showed a
single class of binding sites, with mean Kd values
of 19.7 and 5.4 nM, respectively (Fig. 1, A and
B). This gives evidence that histidine 100 is not the exclusive determinant of flunitrazepam binding as postulated earlier (13).
The response to GABA characterized by the electrophysiological
investigation of recombinant GABAA receptors is known to be enhanced consistently by flunitrazepam (23, 24), and this effect can be
blocked completely by flumazenil (3, 24). In the case of the rat
1
2
2-receptor, 1 µM flunitrazepam has been shown to enhance 1 µM GABA-evoked currents by 64% (25). We observed that
1
2
2-receptors reacted as
expected to the coapplication of GABA and flunitrazepam. Flunitrazepam
(10 µM) markedly affected the amplitude of the peak
current at both GABA concentrations used and in a
dose-dependent manner. At 1 µM GABA, we
detected a 117 ± 46% increase in the amplitude, whereas
coapplication with a higher concentration of GABA (10 µM) elicited a smaller potentiation of 46 ± 6%
(Fig. 2A).
For 6
2
2-receptors,
however, 10 µM flunitrazepam and 10 µM GABA
resulted in a significant reduction of control currents to 76 ± 12%, whereas coapplication of 1 µM GABA and 10 µM flunitrazepam did not produce a significant change in
amplitude (Fig. 2, B and C). This effect of
flunitrazepam was shown to be dependent on the concentration of GABA
and became significant at a concentration of 6 µM GABA
(Fig. 2C). Our data thus indicate that flunitrazepam acts on
6
2
2-receptors in an
inverse agonistic manner.
Our investigation of the effects of flumazenil on GABA-evoked chloride
currents showed opposite effects for the two receptors studied. For
cells expressing
1
2
2-receptors, 10 µM flumazenil reduced the amplitude of chloride currents
induced by application of 10 µM GABA to 89 ± 3% of
control, whereas for cells expressing
6
2
2-receptors, there was
an increase to 118 ± 3% of control induced by application of 1 µM GABA and 10 µM flumazenil (Figs. 3 (A and B) and
4A). We have used different GABA
concentrations depending on the receptor subtype because we found in an
earlier investigation (17) that
6
2
2-receptors effectively
responded to GABA at lower concentrations than
1
2
2-receptors. The effect of flumazenil on
6
2
2-receptor-mediated
responses is low compared with previous reports on Sf-9 cells and HEK
293 cells (26). In an earlier investigation, an apparent agonistic
activity of flumazenil was reported for a binary GABAA
receptor using the rat
1- and
2-subunits
expressed in Xenopus oocytes (23). Here, 10 µM
flumazenil potentiated the response induced by 2-10 µM
GABA to 175% of control.
Coapplication of flunitrazepam and flumazenil to recombinant
1
2
2- and
6
2
2-receptors (Fig.
4B) showed that in the case of the
6
2
2-receptors, the inverse
agonistic properties of flunitrazepam were reduced by the agonistic
action of flumazenil. As expected, for the
1
2
2-receptors, the
agonistic properties of flunitrazepam were almost completely suppressed
by flumazenil.
The partial inverse agonist Ro 15-4513 had the opposite effect on
GABA-induced currents in
1
2
2- and
6
2
2-receptors. In cells
expressing
1
2
2-receptors,
1 µM Ro 15-4513 decreased the amplitude of 10 µM GABA-induced chloride currents. For
6
2
2-receptors with 1 µM GABA, there was a comparable change in current
amplitude, but in the opposite direction (Figs. 3 (A and
B) and 4A).
In this study, we demonstrate that the two receptor subtypes
1
2
2 and
6
2
2 react with the BZ
receptor ligands flunitrazepam, flumazenil, and Ro 15-4513 in a
functionally opposite manner. The exact mechanism underlying the
distinct BZ pharmacology of the
6
2
2-receptor is not clear
at present. Our findings, together with those of earlier investigations
(13, 26), lead us to conclude that distinct amino acids of the
BZ-binding pocket determine the BZ agonistic or inverse agonistic
activity of the GABAA receptor subtype. For the wild-type
6
2
2-receptor, we postulate
that after binding of flunitrazepam, the bulky amino acid arginine induces a structural change that modifies the channel in a way that
decreases chloride ion flux. If the smaller amino acid histidine (wild-type
1
2
2-receptor)
is present, such a hindrance of influx may be absent. Studies involving
site-directed mutagenesis should be conducted to elucidate the
molecular characteristics of the flunitrazepam-binding site on the
6
2
2-receptor.
We are grateful to Peter Seeburg and Heike Wieland for providing the GABAA receptor cDNAs, Hartmut Lüddens for supplying several BZs, and Thorsten Trapp for critical comments on the manuscript. We are also grateful to Thomas Kuckuk and Bettina Burkart-Lauer for excellent assistance with the artwork.