![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
Vol. 56, Issue 3, 464-472, September 1999
Department of Neurochemistry, Max-Planck Institute for Brain Research, Frankfurt/Main, Federal Republic of Germany
 |
Summary |
|---|
 Top
 Summary
 Introduction
Materials and Methods
Results
Discussion
References
|
|---|
Different amino side chains in the N-terminal extracellular
region of the inhibitory glycine receptor (GlyR) have been shown to be
crucial for ligand recognition. Here we describe a novel domain of the
GlyR
1 subunit that constitutes an important determinant of
antagonist activity. The antagonists strychnine, nipecotic acid, and
isobutyric acid displayed reduced potencies at recombinant GlyRs formed
from 1 subunits, in which lysine 104, phenylalanine 108, or
threonine 112 were replaced by alanine. Agonist affinities, in
contrast, were slightly increased at these mutant receptors. Taurine
and 
-aminoisobutyric acid, which are partial agonists at the
wild-type GlyR, behaved as full agonists at the mutant GlyRs and failed
to inhibit glycine-induced currents. This is consistent with apolar
residues at positions 104, 108, and 112 of the 1 subunit reducing
the antagonistic, but not the agonistic, binding of -amino acids.
Our data support a model in which the partial agonism of -amino
acids results from their self-inhibitory activity.
| |
Introduction |
|---|
|
Top
Summary
Introduction
Materials and Methods
Results
Discussion
References
|
|---|
The
inhibitory glycine receptor (GlyR) is a member of the ligand-gated ion
channel family that mediates synaptic inhibition by increasing the
chloride permeability of the postsynaptic membrane. Biochemical and
molecular studies indicate that the GlyR is a pentameric membrane
protein composed of ligand-binding and structural subunits. The
subunit exists in various isoforms (1-4), which all form
functional homo-oligomeric receptor channels upon heterologous
expression in Xenopus laevis oocytes or mammalian cell lines
(reviewed in Kuhse et al., 1995
).
Site-directed mutagenesis indicates that different domains within the
extracellular amino-terminal region of the GlyR 1 subunit contribute
to ligand binding. These include amino acid residues at positions 52 (Ryan et al., 1994; Saul et al., 1994), 159 to 161 (Vandenberg
et al., 1992a; Schmieden et al., 1993), and 200 to 206 (Vandenberg et
al., 1992b), respectively. Substitution of these positions has been
shown to alter the apparent affinities of agonists and/or competitive
antagonists. Interestingly, the homologous positions of type A
-aminobutyric acid receptors (GABAARs) and
nicotinic acetylcholine receptors also have been found to be crucial
for ligand binding (Galzi and Changeux, 1995). In addition, mutations
causing hereditary hyperekplexia have been located in the short loop
connecting transmembrane segments 2 and 3 and shown to drastically
alter both agonist affinity and channel gating (Langosch et al., 1994;
Rajendra et al., 1995; Lewis et al., 1998).
Previous comparisons of the agonist response properties of 1
and 2 GlyRs have identified residue 111 of the 1 subunit as a
crucial determinant of activation by the partial agonist taurine (Schmieden et al., 1992). Taurine displays a highly variable efficacy of GlyR gating in different preparations and has been proposed to act
as a GlyR subtype- or cell type-specific ligand (Lewis et al., 1991).
However, some of the reported differences in agonist efficacy may
result from the antagonistic properties of -amino acids (Horikoshi
et al., 1988; Schmieden and Betz, 1995). Here, we mutated
residues around position 111 of the GlyR 1 subunit and found that
substitution of aromatic, polar, and charged side chains at positions
104, 108, and 112 decreases antagonist but increases agonist potencies.
Our data are consistent with the low efficacy of -amino acid partial
agonists resulting from self-inhibition.
| |
Materials and Methods |
|---|
|
Top
Summary
Introduction
Materials and Methods
Results
Discussion
References
|
|---|
In Vitro Mutagenesis and RNA Synthesis.
Oligonucleotide-directed mutagenesis was performed on single-stranded
cDNA of the human GlyR 1 subunit cloned into pBluescript (Grenningloh et al., 1990) using an in vitro mutagenesis kit (In Vitro
Mutagenesis System II; Amersham, Amersham, UK). All mutants were
identified and verified by dideoxy sequencing of the mutated regions.
EcoRV linearized plasmid DNAs were used for synthesis of RNA
(mRNA Capping Kit; Stratagene, Inc., La Jolla, CA) with T3 RNA
polymerase as described (Schmieden et al., 1992).
Oocyte Expression and Electrophysiology.
The methods used
for analyzing the pharmacology of recombinant GlyRs have been described
previously (Schmieden et al., 1989). Briefly, Xenopus laevis
oocytes were removed from frogs anesthetized with urethane (Sigma,
Munich, Germany), dissected after collagenase (Sigma) treatment, and
injected with cRNAs (10-20 ng per oocyte) of the human GlyR 1
subunit and mutants thereof. Voltage-clamp recording of whole-cell
currents was performed 24 to 48 h after injection at a holding
potential of 
70 mV. Experimental values are presented as the
mean ± S.E.M. of peak current responses. For the evaluation of
half-maximal effective agonist concentrations (EC50) and Hill coefficients (h) from
dose-response curves, data from several oocytes were fitted by the
logistic equation:
|
(1) |
Data Analysis.
Analysis of the pharmacological data was
performed on a MacIntosh computer using a fitting program written by V. Schmieden. The algorithm used calculates the differences between the
measured data (Yx) and the values (Y) obtained by
equation fitting. By reiterating the parameters
EC50 from eq. 1, and
Ka, Kb, and h, respectively, from eq. 3, an optimal fit was established by minimizing the sum of the square of residuals (SS; Bowen and Jerman, 1995) according to:
|
(2) |
![<UP>I</UP>=<FR><NU><UP>I</UP><SUB><UP>max</UP></SUB>[<UP>L</UP>]<SUP><UP>h</UP></SUP></NU><DE><FENCE><IT>K</IT><SUB><UP>a</UP></SUB>+F<FR><NU><IT>K</IT><SUB><UP>a</UP></SUB>[<UP>L</UP>]</NU><DE><IT>K</IT><SUB><UP>b</UP></SUB></DE></FR></FENCE><SUP><UP>h</UP></SUP>+[<UP>L</UP>]<SUP><UP>h</UP></SUP></DE></FR>](/content/vol56/issue3/fulltext/464/img003.gif)
|
(3) |
-amino acids. [L] denotes the total concentration of -amino acid, h is the Hill
coefficient, and Imax represents the maximal
current evoked by the full agonist glycine.
| |
Results |
|---|
|
Top
Summary
Introduction
Materials and Methods
Results
Discussion
References
|
|---|
Characterization of Agonist Responses.
We first analyzed
oocytes injected with the GlyR 1 subunit mutants K104A, F108A, and
T112A for their responses to several -amino acids. This revealed
significant changes in agonist pharmacology resulting from these substitutions.
1 subunit GlyR. This indicates that the amino acid
substitutions introduced had no major effects on expression efficiency
or gating properties. The glycine concentrations generating
half-maximal responses (EC50) calculated from
dose-effect curves were 0.15 ± 0.01 and 0.11 ± 0.02 mM for
the K104A and T112A mutants, respectively (Fig. 1B). These values
resemble that found for the wt 1 subunit (EC50 = 0.20 ± 0.03 mM). Substitution of phenylalanine at position 108 by alanine in mutant F108A generated receptor channels displaying a
3-fold higher affinity for glycine (EC50 = 0.06 ± 0.01 mM). Hill coefficients ranged between 2.1 and 2.3 for
all these homo-oligomeric GlyRs.
|
; Schmieden and Betz,
1995), -amino acid derivatives with C substitutions displayed stereospecific agonist activity. L-alanine and
L-serine behaved as full agonists at all mutants tested
(data not shown), with EC50 values between 0.73 to 1.64 and 0.97 to 5.9 mM, respectively (see Table
1). This represents a 2- to 4-fold
increase in the agonistic potency of these amino acids as
compared with the wt 1 GlyR. In contrast, a strong increase in the
relative current response was obtained for the stereoisomers
D-alanine and D-serine. Whereas
D-serine does not gate the wt 1 GlyR channel (Schmieden and Betz, 1995), it elicited strong currents in oocytes expressing the
mutants K104A, F108A, and T112A. Compared to saturating concentrations of glycine, D-serine exhibited maximal responses of 64 ± 13% at K104A and of 90% at F108A and T112A, respectively (data not
shown), with EC50 values ranging between 6 and 22 mM (Table 1). Remarkably, the EC50 values of
D-alanine were found to be 6- to 13-fold lower with the
different mutants than that determined for the wt 1 subunit (Table
1). These data show that substitutions at positions 104, 108, or 112 alter the stereoselective binding of D-amino acids to the
inhibitory GlyR.
|
-Amino acids are known to behave as partial agonists at the 1
GlyR, with a rank order of -alanine > taurine > -aminobutyric acid (-ABA) when maximal current responses were
compared to that of glycine (Schmieden and Betz, 1995). Whereas
application of 10 mM taurine produced relative responses of about 0.32 at the wt 1 GlyR, the mutants K104A, F108A, and T112A were very
efficiently gated by taurine, with relative responses of 0.74, 0.90, and 0.86, respectively (Fig. 1A, Table 1). Figure 1C indicates that
this increased efficacy was not the result of a change in affinity, because the EC50 values of taurine were very
similar for the wt 1 subunit and the various mutants. Another
important difference from oocytes injected with wt 1 cRNA became
apparent when the mutant receptors were exposed to -ABA. This
-amino acid analog generated maximal currents of 63 to 77% of the
glycine Imax at all three mutants (Table 1); this
represents an increase in gating efficacy of about 10-fold. The
respective dose-effect curves differed much more from the wt 1 GlyR
than those obtained for taurine. Although concentrations of 4.3 ± 0.7 mM -ABA were sufficient to evoke a half-maximal response at
K104A, the mutants F108A and T112A exhibited EC50
values of 1.9 ± 0.5 and 9.7 ± 1.7 mM, respectively (Fig. 1D
and Table 1). The Hill coefficient for -ABA of about 1.3 at all
mutants was significantly lower than that obtained for glycine.
The third known partial agonist of the GlyR is -aminoisobutyric acid
(-AIBA), whose maximal responses at the wt 1 GlyR correspond to
only about 5% of the glycine Imax (Schmieden and Betz, 1995). Interestingly, even 50 mM -AIBA failed to evoke saturating responses at K104A-expressing oocytes. At 6 out of 10 oocytes, no current response to -AIBA was detectable, whereas the
respective glycine current was >1 µA. A least-squares fit of the
resulting data indicated a relative efficacy of 0.11 and an
EC50 value of 21 mM (Table 1). Thus, -AIBA was
significantly less potent at the mutant receptors than the related
molecule -ABA.
Antagonist Pharmacology. To investigate whether the substitutions described above also affect antagonist efficacy, we examined several antagonists for their potency to inhibit glycine responses.
Nanomolar concentrations of strychnine are known to potently suppress glycine-induced currents at the wt1 GlyR (Sontheimer et al., 1989;
Grenningloh et al., 1990). Here, strychnine was tested using glycine at
a concentration corresponding to its EC50 value.
Figure 2A shows that the mutants K104A
and F108A were about 6- to 10-fold less sensitive to the
alkaloid (IC50 values of 172 ± 9 and
293 ± 6 nM, respectively) than the 1 subunit receptor (IC50 of about 15 nM). A much stronger shift of
the inhibition curve to low affinity was obtained for T112A
(IC50 of 1.6 ± 0.1 µM; Table
2).
|
|
). At all mutants,
the potency of this heterocycle was significantly reduced (Fig. 2B), resulting in IC50 values of about 3 mM for K104A
and F108A, and of about 7 mM for T112A (Table 2). The piperidine
derivative isonipecotic acid was also analyzed on K104A and showed a
significant reduction in affinity (Table 2). Finally, inhibition of
glycine currents by the -amino acid isobutyric acid
(IC50 = 20.4 ± 1.3 mM; n = 3) was strongly reduced (Table 2).
A new class of GlyR antagonists is represented by the kynurenic acid
analog 5,7-dichloro-4-hydroxy-quinoline-3-carboxylic acid (5,7ClQA).
This compound inhibits in a mixed competitive/noncompetitive fashion,
which originates from the chloride substitution at the aromatic ring
system (Schmieden et al., 1996). On K104A and wt 1 GlyR expressing
oocytes, IC50 values for 5,7ClQA were similar (Table 2).
To examine whether the effects described above might be potentiated
upon multiple substitution, we designed a triple (K104A, F108A, T112A)
1 mutant. The resulting receptor was potently gated by agonists, but
showed a pharmacological profile related to that of the mutant F108A.
Briefly, the EC50 value for glycine was
0.063 ± 0.01 mM (n = 4); taurine and -ABA
exhibited EC50 values of 0.61 ± 0.09 mM
(n = 3) and 3.17 ± 0.78 mM (n = 4), respectively. Maximal responses for taurine were 100%, and for
-ABA 60% of the glycine Imax. Furthermore,
nipecotic acid and strychnine antagonized current responses with
IC50 values of 3.18 ± 0.9 mM
(n = 3) and 0.32 ± 0.1 µM (n = 3), respectively.
In conclusion, substitution of lysine 104, phenylalanine 108, or
threonine 112 by alanine reduced the potency of antagonists in a rank
order of strychnine > isobutyric acid > nipecotic acid, whereas the apparent affinities of agonists were inversely increased (Fig. 3). This indicates a critical role
for these amino acid side chains and/or the region defined by these
mutations in determining the consequences of ligand binding.
|
Modeling of Partial Agonist Function.
We have previously shown
that the -amino acids taurine and in particular -ABA and -AIBA
exhibit antagonistic activity when coapplied with glycine (Schmieden
and Betz, 1995). To test whether a similar behavior is found for the
mutants described above, we performed inhibition experiments with
taurine on oocytes expressing mutant K104A. As shown in Fig.
4A, increasing concentrations of taurine
generated strong currents when applied alone. In the presence of 100 µM glycine, both ligands acted synergistically as agonists. Similar
results were found for -ABA; again the dose-effect curve of glycine
was not shifted when 2 mM -ABA was added (Fig. 4B). These data
indicate a loss of the antagonistic properties of these -amino acids
upon mutation of lysine 104. In contrast, -AIBA still potently
inhibited glycine responses (Fig. 4C). The current elicited by 0.1 mM
glycine was drastically reduced when increasing concentrations of
-AIBA were coapplied. Analysis of the data indicated an
IC50 value for -AIBA of 2.4 mM (Table 2).
|
-ABA were strongly enhanced upon mutation,
whereas those of -AIBA were not. To unravel the basis of this
observation, we plotted the determined Imax
values against the respective EC50 values for
each mutant (Fig. 5A). Neither taurine
nor -ABA or -AIBA showed a good correlation between agonist
efficacy and apparent affinity. For example, the relative Imax value of -ABA found at the T112A mutant
was much larger (75%) than that (7%) obtained with the wt 1
subunit. The EC50 value for this ligand, however,
was lower at wt 1 than at T112A GlyRs. This analysis indicates that
the different binding affinities of -amino acids as defined by their
EC50 values are not sufficient to explain their
vastly different gating efficacies.
|
-amino acids might be
due to a dual action as both activators and competitive inhibitors (Schmieden and Betz, 1995). This was proposed to reflect different molecular conformations of the ligand. The dose-response profiles of
partial agonists may thus be considered as the sum of both an agonist
and an antagonist concentration-effect curve. Under conditions of full
binding site saturation, the relative maximal current
(Irel) obtained with a partial agonist should be
proportional to the ratio of
Ka/Kb, i.e.,
the binding constants for the activating and inhibiting conformers, as
follows:
|
(4) |
-ABA, and -AIBA could be described as a fraction of two
components (eq. 3; see Materials and Methods). Figure 5B shows a least-squares fit for the partial agonists tested, in which
Imax was normalized to the maximal glycine
current. For taurine, a ratio of
Ka/Kb of 0.63 (Ka = 0.51; Kb = 0.8) and a Hill coefficient of 2.6 were calculated. Whereas this
Ka value is comparable to the determined
EC50 value (Table.1), the resulting Hill
coefficient was significantly higher than the slope obtained by a
simple fit (eq. 1) and very similar to that found for glycine. Correspondingly, for -ABA and -AIBA, ratios of 0.707 (Ka = 2.9; Kb = 4.1)
and 1.88 (Ka = 17.2;
Kb = 9,13), respectively, were found.
Again, for both ligands the calculated Hill coefficients were 2.1 and
2.6. This result strengthens the view that the reduced agonist
efficacies of these partial agonists are due to competitive self-inhibition.
This approach was also used to model the low current responses of
partial agonists at the 1 wt GlyR and the high current responses
obtained for mutants F108A and T112A. To this end, the individual
dose-response relations of taurine, -ABA, and -AIBA were fitted
by eq. 3. The parameters calculated from this analysis are summarized
in Table 3. It is obvious that the
calculated Ka values and the experimentally
determined EC50 values were rather similar. A
high current response is thus the consequence of a low
Kb value. In other words, the mutations
described minimize the antagonistic binding of partial agonists and
increase the efficacy of agonistic interaction. In Fig. 5C, the maximal
currents evoked by saturating concentrations of the -amino acids
were plotted versus the logarithm of the calculated
Ka/Kb ratios.
All determined maximal currents showed a good correlation with the affinity ratios
Ka/Kb. The
sigmoidal shape of the data fit is consistent with eq. 4 and exhibits a
slope of 2.
|
1
subunit generate GlyRs displaying increased agonist and reduced
antagonist affinities. Notably, the efficacies of partial agonists were
significantly higher than at the wt 1 GlyR. This is consistent with
an impairment of -amino acid antagonism.
| |
Discussion |
|---|
|
Top
Summary
Introduction
Materials and Methods
Results
Discussion
References
|
|---|
In this study, we describe the pharmacology of three GlyR 1
subunit mutants, in which lysine 104, phenylalanine 108, and threonine
112, respectively, were replaced by alanine. Like the wt 1 subunit,
these mutants generated fully functional receptor proteins that
displayed moderate increases in agonist affinity. The largest
reductions in EC50 values were seen for
D-alanine and D-serine, whereas glycine,
L-alanine, L-serine, and taurine affinities
were detectably affected only upon substitution of phenylalanine 108. Notably, -ABA and -AIBA displayed rather low apparent affinities
at mutants T112A and K104A. This might be due to the exposed methyl
groups at their C and C atoms, which may restrict interactions
with the mutated binding pocket.
The most interesting result obtained with our GlyR mutants is that the
apparent changes in antagonist affinities were opposite to those found
for the -amino acid agonists. Similar observations have recently
been reported for the GABAAR (Ebert et al.,
1997). Strychnine, isobutyric acid, nipecotic acid, and isonipecotic acid all inhibited the glycine response only with high
IC50 values. This decrease in antagonist affinity
was most pronounced with mutant T112A. Interestingly, the partial
agonist -AIBA inhibited glycine responses at mutant K104A to >90%;
however, its IC50 value was 3-fold higher than
that obtained at the wt 1 subunit. A similarly reduced antagonistic
affinity was also found for nipecotic and isonipecotic acid. From this
data one may speculate that -AIBA and the other GlyR antagonists use
similar subsites within the binding pocket.
Our results imply that aromatic, polar, and charged amino acid residues
in the mutated region contribute to antagonist efficacy in a rank order
of threonine 112 > phenylalanine 108 > lysine 104. Agonist
affinities, in contrast, increased when introducing apolar side chains
at these positions. Interestingly, the pharmacological profile of the
triple mutant (K104A, F108A, T112A) resembled that of the single mutant
F108A. This suggests that the domain harboring positions 104 to 112 contributes only to a limited extent to antagonist/receptor interaction. Consequently, multiple substitutions in this region may
not cause further reductions in antagonist affinities. Similarly, substitution of histidine 107 and glutamate 110 had no significant effect on antagonist recognition (Schmieden et al., 1992).
An important role of aromatic residues in ligand recognition by
neurotransmitter receptors has been demonstrated in various studies.
For the GlyR, photoaffinity labeling by strychnine has been proposed to
involve energy transfer from aromatic side chains, such as tyrosine
(Graham et al., 1983). Mutational analysis identified tyrosines 161 and
202 as crucial determinants of strychnine binding (Vandenberg et al.,
1992a; Rajendra et al., 1995). Furthermore, the exchange of
phenylalanine 159 and tyrosine 161 has been found to increase agonist
but not antagonist affinity (Schmieden et al., 1993). In addition,
charged amino acid residues like lysine 200 (Rajendra et al., 1995)
promote both strychnine and glycine recognition, whereas mutations of
the polar residue threonine 204 selectively reduces agonist binding
(Rajendra et al., 1995). These findings lead us to propose that the
agonist and antagonist recognition site of the GlyR contains two
domains (positions 159-161 and 200-204), where agonists bind to
hydroxyl moieties, whereas antagonists interact directly with aromatic
ring systems (see also Rajendra et al., 1995). The domain harboring
positions 104 to 112 includes both hydroxylated and aromatic side
chains; however, it appears crucial primarily for antagonist efficacy.
The question whether the amino acid residues mutated here are directly
involved in ligand binding cannot be answered presently. Homologous
positions of the GABAAR have been implicated in
benzodiazepine recognition (Siegel and Buhr, 1997). Threonine 142 of
the subunit and histidine 101 of the 1 subunit of the
GABAAR are thought to contribute to the interface
between both subunits (Galzi and Changeux, 1995). Mutation of position
100 in the GABAAR 6 subunit (corresponding to
H101 of the 1 GABAAR subunit) caused a
complete loss of diazepam binding, whereas the affinity for GABA was
not altered (Korpi et al., 1993) Furthermore, tyrosine 93 in the
-subunit of the nicotinic acetylcholine receptor, involved in
acetylcholine binding, is equivalent to histidine 101 of the
GABAAR 1 subunit. Although benzodiazepines are
allosteric modulators rather than agonists or antagonists of the
GABAARs, it is interesting to speculate that this
binding domain of GABAARs may be equivalent to
one for GlyR competitive antagonists. Notably, quinolinic acid
compounds behave differently from the other GlyR antagonists. The
IC50 value of 5,7ClQA was similar for both mutant
and wt GlyRs. 5,7ClQA has been described previously as a mixed
competitive/noncompetitive antagonist (Schmieden et al., 1996). Neither
the competitive binding nor the allosteric interaction of 5,7ClQA were
found to be affected by the mutations analyzed in this study (V.S.,
unpublished data). Consequently, a separate subsite for quinolines may exist.
The maximal current responses elicited by taurine and -ABA were
significantly higher with the mutant receptors (70-90% of the maximal
glycine response) than with the wt 1 GlyR (30 and 7%,
respectively). In contrast, -AIBA appeared similarly effective at
both the mutants and the wt 1 subunit. To examine whether this
change in ligand efficacy reflected differences in agonist affinity, we
compared the EC50 values obtained for taurine,
-ABA, and -AIBA to the respective maximal current responses and
found no correlation. This may reflect a nonequivalence of measured EC50 value and true binding affinities. For
recombinant GlyRs, glycine affinities determined by displacement of
bound [3H]strychnine and
EC50 values calculated from dose-effect curves have been found to be similar (Rajendra et al., 1995). We therefore conclude that the low efficacy of -amino acids must be due to other
properties of these ligands.
In an previous report we speculated that the low gating efficacy of
partial agonists originates from their simultaneous agonistic and
antagonistic binding within the pharmacophore of the GlyR (Schmieden
and Betz, 1995). This assumption is based on the fact that -amino
acids can exist in different conformations. The trans configuration is structurally related to nipecotic acid, which is a
full antagonist, whereas the cis configuration has been
proposed to represent the agonistic conformation (Schmieden and Betz,
1995). Accepting these predictions and assuming that the binding site of the GlyR can be occupied by both the trans- and
cis-conformers, a simple competitive interaction may occur
between the latter. Here, we examined this proposal by fitting the
current responses of partial agonists with the logistic function 3, which considers the respective affinities of both conformers. Based on
visual inspection of the resulting fits it can be concluded that this procedure results in a good description of the observed current responses. Accordingly, the efficacy (Irel) of
the respective ligand is proportional to its
Ka/Kb ratio. It
is worth noting that the Kb, but not the
Ka, value depends critically on the mole fractions of agonistic and antagonistic conformers (F)
within the agonist solution. Here, we assumed the mole fraction ratio to be 1, because this value is close to the mole ratios obtained for
-alanine and taurine in NMR studies (Ham, 1974). As shown in Fig. 5,
this resulted in excellent fits of the predicted and our experimentally
determined relative current values.
According to the proposal made above, the high current responses of
taurine and -ABA found at our mutants reflect an increase in
agonistic binding (low Ka value) with a
simultaneous loss of GlyR antagonism (high
Kb value). Indeed, our competition
experiments clearly show that both ligands failed to inhibit
glycine-evoked currents. In contrast, -AIBA was still capable of
antagonizing glycine responses at mutant K104A. Although the determined
EC50 and IC50 values for
this ligand changed (both about 2.5-fold), the calculated
Ka/Kb ratios
and the evoked current responses appeared similar for both the mutants
and the wt 1 GlyR.
Our model also gives a reasonable explanation for the differences in
Hill coefficients reported between full and partial agonists. When
using a uniform cooperativity for all agonists of about h = 2.5, our model calculations generated dose-response curves for taurine and
-ABA with a slope <2. Similar results were observed in experiments
with adrenergic 2 autoreceptors (Feuerstein et al., 1994) and the
1 GlyR (Schmieden and Betz, 1995), in which partial agonist function
was mimicked by applying mixtures of a full agonist and an antagonist
at various concentration ratios.
Allosterical two-state receptor models (Monod et al., 1965; Leff, 1995)
have also been used to explain partial agonist activity. In these
models, ligands are thought to bind to active (R) or inactive (R')
receptor conformations with corresponding
Ka and Kb
values. The affinities of a particular ligand for R' and/or R then
define its full agonistic, partial agonistic, or antagonistic activities, respectively. Consequently, receptors in which the R:R'
ratios were changed ("L-phenotype" according to Galzi
et al., 1996) may yield different gating efficiencies. Using this approach, the low glycine responses of GlyR 1 mutants causing hyperekplexia (Rajendra et al., 1994; Langosch et al., 1994) have been explained theoretically (Galzi et al., 1996). However, these models consistently correlated low current responses with low affinities for agonists. Assuming that our mutations solely changed the
allosteric R/R' ratio of the resulting GlyRs by favoring a high open
probability as required for increased current responses, the resulting
EC50 values should be significantly lower than
for the wt GlyR. This prediction contradicts our findings. We
therefore conclude that the amino acid exchanges at positions 104, 108, and 112 had no major effect on allosteric transitions. Rather, these
amino acids appear crucial for the recognition of antagonists and the
trans-conformation of -amino acids. All presently
available data are compatible with the interpretation that the altered
gating efficiencies of partial agonists seen upon substitution of
residues 104, 108, and 112 reflect a loss of self-inhibition resulting from an increased
Ka/Kb ratio.
| |
Footnotes |
|---|
Received March 8, 1999; Accepted June 16, 1999
1 Present address: Department of Physiology, Campus Charité, Humboldt-University of Berlin, Tucholskystrasse 2, 10117 Berlin, Federal Republic of Germany. Phone: +49(30)28026183, Fax +49(30)28026669, e-mail: volker.schmieden{at}charite.de
2 Present address: Department of Anatomy and Cellular Neurobiology, University of Ulm, Albert-Einstein-Allee 11, D-89081 Ulm, Federal Republic of Germany.
Send reprint requests to: Dr. Volker Schmieden, Department of Physiology, Campus Charité, Humboldt-University of Berlin, Tucholskystrasse 2, 10117 Berlin, Germany. E-mail: volker.schmieden{at}charite.de
| |
Abbreviations |
|---|
-ABA, -aminobutyric acid;
GABA, -aminobutyric acid;
-AIBA, -aminoisobutyric acid;
GABAAR, type A -aminobutyric acid receptor;
GlyR, inhibitory glycine receptor;
wt, wild-type;
5,7ClQA, 5,7-dichloro-4-hydroxy-quinoline-3-carboxylic acid.
| |
References |
|---|
|
Top
Summary
Introduction
Materials and Methods
Results
Discussion
References
|
|---|
-aminobutyric acid type A receptors.
Mol Pharmacol
52:
1150-11561(K276E), expressed in Xenopus oocytes.
J Physiol (Lond)
507:
25-401 subunit of the inhibitory glycine receptor in the spasmodic mouse.
Nat Genet
7:
131-135[Medline].1 subunit in the spasmodic mouse affects agonist responses.
FEBS Lett
350:
71-76[Medline].-alanine receptor responsive to GABA.
Science (Wash DC)
262:
256-258 subunit form the strychnine-binding site.
Proc Natl Acad Sci USA
89:
1764-1769.This article has been cited by other articles:
|
|
 |
|
  A. Gameiro, F. Reimann, A. M. Habib, D. O'Malley, L. Williams, A. K. Simpson, and F. M. Gribble The neurotransmitters glycine and GABA stimulate glucagon-like peptide-1 release from the GLUTag cell line J. Physiol., December 15, 2005; 569(3): 761 - 772. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. S Miller, M. Beato, R. J Harvey, and T. G Smart Molecular determinants of glycine receptor {alpha}{beta} subunit sensitivities to Zn2+-mediated inhibition J. Physiol., August 1, 2005; 566(3): 657 - 670. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. B.P. van Kuilenburg, R. Meinsma, E. Beke, B. Assmann, A. Ribes, I. Lorente, R. Busch, E. Mayatepek, N. G.G.M. Abeling, A. van Cruchten, et al. {beta}-Ureidopropionase deficiency: an inborn error of pyrimidine degradation associated with neurological abnormalities Hum. Mol. Genet., November 15, 2004; 13(22): 2793 - 2801. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. W. Lynch Molecular Structure and Function of the Glycine Receptor Chloride Channel Physiol Rev, October 1, 2004; 84(4): 1051 - 1095. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Castaldo, P. Stefanoni, F. Miceli, G. Coppola, E. M. del Giudice, G. Bellini, A. Pascotto, J. R. Trudell, N. L. Harrison, L. Annunziato, et al. A Novel Hyperekplexia-causing Mutation in the Pre-transmembrane Segment 1 of the Human Glycine Receptor {alpha}1 Subunit Reduces Membrane Expression and Impairs Gating by Agonists J. Biol. Chem., June 11, 2004; 279(24): 25598 - 25604. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Z. Chen, G. H. Dillon, and R. Huang Molecular Determinants of Proton Modulation of Glycine Receptors J. Biol. Chem., January 9, 2004; 279(2): 876 - 883. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. L. Thio, A. Shanmugam, K. Isenberg, and K. Yamada Benzodiazepines Block {alpha}2-Containing Inhibitory Glycine Receptors in Embryonic Mouse Hippocampal Neurons J Neurophysiol, July 1, 2003; 90(1): 89 - 99. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||
), 
), and 
) by strychnine (A)
and nipecotic acid (nip; B). Glycine at half-saturating concentration
was coapplied with increasing concentrations of the indicated
antagonists. Current values were normalized to the glycine response
obtained in the absence of antagonist and were fitted by a single
sigmoidal curve. Error bars often were smaller than the symbols used,
and are not indicated. For comparison, the inhibition curves for
strychnine and nipecotic acid obtained at the wt 
) of 0.1 mM glycine
(n = 4). Note the high potency of 
), and 
),
respectively. Imax values were plotted as current relative
to that induced by a saturating glycine concentration. Bars represent
the mean ± S.E.M. of three to six experiments. B, dose response
curves of taurine (