European Journal of Pharmacology - Molecular Pharmacology Section, 247 (1993) 23-27 23 © 1993 Elsevier Science Publishers B.V. All rights reserved 0922-4106/93/$06.00

EJPMOL 90488

Natural mutation of G A B A A receptor a6 subunit alters benzodiazepine affinity but not allosteric GABA effects

Esa R. Korpi a.b and P e t e r H. Seeburg b

" Biomedical Research Center, Alko Ltd., HelsinkL Finland, and b Laboratory of Molecular Neuroendocrinology, ZMBH, Unil,ersity of Heidelberg, Heidelberg, Germany

Received 31 March 1993; revised MS received 10 May 1993; accepted 18 May 1993

The binding of the imidazobenzodiazepine, [3H]Ro 15-4513, to cerebellar granule cell-specific GABAA/benzodiazepine receptors is typically insensitive to benzodiazepine receptor agonists such as diazepam. A mutation in the a6 subunit, causing replacement of the arginine at the 100 position by glutamine (Q100), has recently been found in an alcohol- and benzodiazepine-sensitive rat line. The mutant a6(Q100)132y2 recombinant receptors are sensitive to diazepam. The binding of [3H]Ro 15-4513 to cerebellar diazepam-insensitive receptors is enhanced by GABA, whereas binding to diazepam-sensitive receptors is inhibited. Recombinant receptors consisting of 132 and 3,2 subunits together with the wildtype a6 or mutant a6(Q100) subunit showed positive modulation of [3H]Ro 15-4513 binding by GABA, whereas a113272 receptors showed negative modulation. The picrotoxin-sensitive binding of a convulsant, t-butylbicyclophosphoro[35S]thionate ([35S]TBPS), was inhibited in the a613272 and a6(Q100)13272 receptors by GABA at concentrations less than one-tenth of those required in the a113272 receptors. GABA effects on [35S]TBPS binding were only slightly affected by diazepam in the a6(Q100)132y2 receptors, while profound effects were seen in the a113272 receptors in the presence of diazepam. The results with the mutant receptor suggest that the a l and a6 subunits are responsible for differential allosteric actions by GABA on other binding sites, independently of the structures defining the benzodiazepine binding pharmacology.

Ro 15-4513; Cerebellum; GABA A receptor subtypes; Recombinant mutant receptors; (Selected rat lines)

I. Introduction

Central GABA A receptors are pharmacologically heterogeneous, especially at their benzodiazepine recognition sites (Liiddens and Wisden, 1991). These receptors are thought to be composed of several sub- units belonging to different subunit families, a , /3 , y, and p (Wisden and Seeburg, 1992). The a and /3 subunits are sufficient to form functional GABA ago- nist-sensitive ion channels, which are affected by con- vulsants (such as picrotoxin), barbiturates and neuro- steroids but not by benzodiazepines (Schofield et al., 1987; Puia et al., 1990). Benzodiazepine modulation appears to require the assembly of still another sub- unit, y. The 3,2 subunit, in particular, is important for the benzodiazepine binding domain (Pritchett et al., 1989), although benzodiazepine receptor pharmacol-

Correspondence to: E.R. Korpi, Biomedical Research Center, Alko Ltd., P.O. Box 350, SF-00101 Helsinki, Finland. Tel. 358.0.1332849; Fax 358.0.1332781.

ogy, as such, is mainly determined by the a subunit variant (see Lfiddens and Wisden, 1991).

The binding of a cage convutsant, t-butylbi- cyclophosphoro[35S]thionate ([35S]TBPS), to picro- toxin-sensitive sites on membranes from whole adult rat forebrain is potently inhibited by GABA and all known GABA A receptor agonists in a way reversible by all known GABA A receptor antagonists (Squires et al., 1983; Squires and Saederup, 1987). In situ, GABA more potently inhibits the binding of [3sS]TBPS to the cerebellar granule layer than to the molecular layer of rat brain (Korpi et al., 1992a; Sapp et al., 1992). We have recently found that this regional heterogeneity of GABA responses is largely determined by the a sub- unit variant, a6/32y2 receptors being much more sen- sitive than a1/32y2 receptors (Kleingoor et al., 1993; Korpi and Lfiddens, 1993). This differential sensitivity can be demonstrated both by [35S]TBPS binding (Korpi and Ltiddens, 1993) and by electrophysiological mea- surements using recombinant receptors (Kleingoor et al., 1993). This is consistent with the cell-specific ex- pression of the a6 subunit in cerebellar granule cells,

24

as opposed to the more universal expression of the a l subunit (Liiddens et al., 1990; Laurie et al., 1992; Wisden et al., 1992). It is not known what parts of the a subunit proteins are responsible for this dramatic change in GABA sensitivity.

Wieland et al. (1992) identified in the a subunits a histidine residue (His-101 in a l ) which is crucial for the construction of binding domains for benzodi- azepine agonists. There are also benzodiazepine ago- nist-insensitive binding sites which can be labelled with the imidazobenzodiazepine [3H]Ro 15-4513 (Mal- miniemi and Korpi, 1989). If this histidine is replaced by an arginine residue, as is the case in the a6 and a4 subunits, agonist-insensitive binding sites are produced in recombinant receptors (Lfiddens et al., 1990; Wis- den et al., 1991). Recently, we found a natural point mutation in this position in the GABA A receptor a6 subunit, replacing Arg-100 with glutamine (Korpi et al., 1993). The latter residue, too, led to the production of benzodiazepine agonist-sensitive binding sites and po- tentiation of GABA responses by these compounds. In the present report, we demonstrate that GABA modu- lation of convulsant and benzodiazepine binding in recombinant, mutant a6(Q100)/3272 receptors is simi- lar to that found in receptors containing the wild-type a6 subunit. This suggests that GABA modulation is not determined by structures responsible for the phar- macological specificity of benzodiazepine binding sites.

2. Materials and methods

2.1. Chemicals

Radiolabelled t-butylbicyclophosphorothionate ([35S]TBPS), ethyl 8-azido-5,6-dihydro-5-methyl-6-oxo- 4H-imidazo[1,5-a][1,4]benzodiazepine-3-carboxylate ([3H]Ro 15-4513) and [3H]muscimol were purchased from Du Pont de Nemours, NEN Division, Germany. Unlabelled Ro 15-4513 was obtained from F. Hoff- mann-LaRoche (Basle, Switzerland), GABA from Serva Feinbiochemica (Heidelberg, Germany) and picrotox- inin from Sigma Chemicals (St. Louis, MO, USA).

2.2. Recombinant receptors

Human embryonic kidney 293 ceils were transfected with rat al-, a6-, a6(Q100)-, /32- and y2-encoding cDNAs, subcloned individually into eukaryotic expres- sion vectors (Lfiddens et al., 1990; Pritchett et al., 1989; Ymer et al., 1989; Korpi et al., 1993), using the calcium phosphate precipitation method (Chen and Okayama, 1987). Based on optimising [35S]TBPS binding, the quantitative ratios of the cDNAs for the a, /3 and y subunits were 1 (= 5 /zg per 4 × 10 6 cells):5:0.1 (Hartmut Liiddens, in preparation). Briefly, the cells

were plated on 15-cm plastic dishes (Becton Dickinson Labware, Lincoln Park, N J, USA) to be transfected 2 to 3 days later. About 20 h after transfection, the medium was changed, and cells were washed once and then harvested into phosphate-buffered saline. The cell pellets were homogenised in 50 mM Tris-citrate (pH 7.4) buffer with a Polytron, followed by two centrifuga- tion-resuspension cycles for binding experiments with flesh homogenate, or by one centrifugation for freezing at - 80°C.

2.3. Binding assays

Both flesh and frozen membranes were used with similar results. Frozen membranes were thawed, resus- pended and centrifuged once, before final resuspen- sion in 50 mM Tris-citrate, giving a protein concentra- tion of 50-120 /~g in 0.5 ml (Bio-Rad method using bovine serum albumin as standard). Each sample was homogenised briefly with a Polytron before adding it to the binding mixture. After incubation with duplicate samples, the bound ligand was separated from the free one by rapidly filtrating the membranes onto Schle- icher and Schuell (No. 52) or Whatman (GF/B) glass fibre filters, using a manifold for 24 simultaneous sam- ples operating with reduced pressure or a Brandel M-48R filtration unit. The samples were rinsed twice with 5 ml of ice-cold 10 mM Tris-HC1 (pH 7.4) or incubation solution. The filters were air-dried and were immersed in 4 ml of Packard Ultima Gold or Wallac Optiphase Hisafe II scintillation fluid. The radioactiv- ity of the filters was counted with in a Beckmann or Wallac liquid scintillation counter using external stan- dardization.

[3H]Ro 15-4513 binding (6 nM) was determined in Krebs-Tris buffer (composition in mM: NaC1 116, KCI 4.8, CaC12 1.2, MgC12 1.2, Tris base 15; pH 7.4 ad- justed with HC1 at room temperature) with 60-min incubations at 0°C in the dark as described by Korpi et al. (1992b). The amount of nonspeciflc binding was determined in the presence of 10/.LM Ro 15-4513.

[35S]TBPS binding (6 nM)was determined by 90-min incubation at 22°C in 50 mM Tris-citrate supplemented with 200 mM NaCI. Nonspeciflc binding was deter- mined in the presence of 10 /~M picrotoxinin. GABA was added to the incubation mixture at final concentra- tions ranging from 10 nM to 10/~M with or without 10 /zM diazepam.

[3H]Muscimol binding (1 to 48 nM) was determined in Tris-citrate buffer with 60-min incubation at 0°C, with 100/xM GABA defining the nonspecific binding.

The Graph Pad Inplot program was used to calcu- late the best-fitting values for the parameters of the saturation isotherms and displacement curves. Statisti- cal significances of differences were assessed by analy- sis of variance and the paired t-test, using the Graph

25 ~ f ~

150

125

i00

75

50

i~.%< ~ / a6(Q100)

I I - ~ ' ' b 5 I --8 --7 --6 -- --4

log [GABA]

Fig. l. GABA modulation of [3H]Ro 15-4513 binding in recombinant o~6/3272, a6(Q100)/3272 and a l /3272 receptors. The values are means + S.E.M. The basal binding was 1.2 + 0.3 (mean + S.E.M., n = 5), 2.0+0.6 and 0.67_+0.08 pmol /mg protein for the a6, o~6(Q100) and a l subunit-containing receptors, respectively. * P < 0.05, ** P

< 0.01, for difference from basal level.

Pad Instat program. The number of experiments (n) quoted in the figure legends refers both to the number of independent transfections and to that of binding assays.

3. Results

[3H]Ro 15-4513 binding was differently modulated by GABA in three recombinant receptors (fig. 1). In the cd/32T2 receptors, GABA acted as a negative modulator of the binding, which is in agreement with the classification of Ro 15-4513 as an inverse agonist (Bonetti et al., 1989). In the a6/3272 receptors, GABA enhanced the binding significantly (fig. 1), which con- curs with our previous findings on the wildtype recep- tors of cultured cerebellar granule cells and cerebellar homogenates (Malminiemi and Korpi, 1989; Uusi- Oukari and Korpi, 1992). The mutant a6(Q100)#272 receptor also showed enhancement of binding by GABA. [3I--I]Ro 15-4513 binding was sensitive to di- azepam in a l#272 (K~ 0.016 IzM; Ltiddens et al., 1990) and a6(Q100)/3272 (K i 1.3 /xM; Korpi et al., 1993) receptors but insensitive in a6/3272 receptors.

[35S]TBPS binding to the convulsant sites at the GABA A receptors were inhibited by GABA at lower concentrations in the c~6/32,/2 and a6(Q100)/3272 re- ceptors than in the al/3272 receptors (fig. 2). The difference in the present conditions is more than one magnitude in the concentration of GABA. Diazepam at 10 izM enhanced both the stimulation of the binding by low GABA concentrations and the inhibition of the binding by higher G A B A concentrations in the al/3272 receptors (fig. 3). On the other hand, diazepam had no significant (P > 0.05) effects in the wildtype a6#2T2

25

150

.~ .~ 100

rr~ m

~ 50

i i i i

-8 -7 -6 -5

log [GABA]

Fig. 2. GABA modulation of [35S]TBPS binding in recombinant a6#272 , a6(Q100)/3272 and a l /3272 receptors. The results are means-+ S.EM. The basal binding was 47 + 14 (mean + S.E.M., n = 4), 46_+7 (n = 4) and 77_+ 11 (n = 3) fmol /mg protein for the or6, ot6- (Q100) and a l subunit-containing receptors, respectively. * * * P < 0.001, for difference from the a6 and a6(Q100) subunit-containing

receptors.

receptors, and only slightly enhanced (P = 0.04) the inhibition by 100 nM GABA in the a6(Q100)/3272 receptors.

150

150

100

v

t~ 50 t -

O u ~

m ~ 200

a 150 fn

(16

100 ~ ~ , ~

5O

0

100

50

a6(QlO0)

al

- 8 - 7 6 - 5

log [GABA]

Fig. 3. Diazepam modulation of GABA effects on [35S]TBPS binding in recombinant a6/32T2, a6(Q100)/32T2 and a l /3272 receptors. The results are means+S.E.M. (n = 3-4). Open symbols denote values in the presence of GABA only, and closed symbols in the presence of GABA and 10 ,aM diazepam. For the basal binding values, see the legend to fig. 2. * P < 0.05, ** P <0.01, for differ-

ence from corresponding value without diazepam.

26

High-affinity [3H]muscimol binding was similar in the a6/32y2 and a6(Q100)/32y2 receptors, with K D values of 3.2 _+ 0.4 (mean + S.E.M., n = 3) and 3.0 + 0.2 riM, respectively. The a l / 3 2 y 2 receptors bind [3H]- muscimol at about the same affinity (K D 8 nM; Korpi and Lfiddens, 1993).

4. Discussion

Alcohol-sensitive ANT rats produced by selective breeding for enhanced motor impairment by moderate doses of ethanol (Eriksson and Rusi, 1981) are also abnormally sensitive to the motor-impairing effects of benzodiazepine agonists (Hellevuo et al., 1989). The ANT rats carry a guanine to adenine mutation in a pharmacologically critical position in their GABA A receptor a6 subunits (Korpi et al., 1993). The resulting amino acid change from Arg-100 to Gln-100 converts the cerebellar granule cell-specific receptors from di- azepam insensitive to diazepam sensitive (Uusi-Oukari and Korpi, 1990; Korpi et al., 1992b). Since the cere- bellum is important for motor coordination, the GABA A receptor mutation apparently explains the abnormal behavioral benzodiazepine sensitivity. In the present study, we demonstrate that while benzodi- azepine sensitivity is altered by the a6(Q100) mutation in recombinant receptors, certain other characteristics of the a6 subunit-containing receptors are not af- fected.

We have previously reported two unique properties of the cerebellar granule cell layer, both of which can be accounted for by the c~6 subunit-containing GABA A receptors. First, the diazepam-insensitive binding of [3H]Ro 15-4513, specific for a6 (Liiddens et al., 1990), is enhanced by GABA whereas diazepam-sensitive binding is not (Malminiemi and Korpi, 1989; Uusi- Oukari and Korpi, 1992). Second, the sensitivity of convulsant binding to inhibition by GABA is much higher in GABA A receptors of granule cell than in the GABA A receptors of other brain regions (Korpi et al., 1992a), a phenomenon to which the a6 subunit also appears to contribute (Korpi and Lfiddens, 1993). In the present study, both of these features were unaf- fected by the a6(Q100) mutation in recombinant re- ceptors. This indicates that GABA stimulation of [3H]Ro 15-4513 binding takes place within the oligomeric receptor complex, and that the allosteric coupling(s) of GABA sites with benzodiazepine and convulsant sites do not depend on the pharmacological specificity of the benzodiazepine binding sites. The latter conclusion is supported by our recent study (Kleingoor et al., 1993) showing that benzodiazepine pharmacology, but not GABA sensitivity, can be changed in the ax /32y2 receptors by reciprocal histi-

dine and arginine substitutions in a6 and a l subunits, at the 100th or 101st amino acid residue, respectively.

GABA inhibition of [35S]TBPS binding in the mu- tant a6(Q100)/32y2 receptors was only marginally en- hanced by diazepam, even though most of the benzodi- azepine receptors were presumably occupied by 10 g M diazepam. Marked diazepam-induced potentiation of GABA currents has been demonstrated in H E K 293 cells expressing the mutant receptors (Korpi et al., 1993). This suggests that the conditions at which the convulsant binding was carried out in the present study may not have been optimal for the detection of benzo- diazepine modulation of GABA effects. It may also suggest that the allosteric modulation of [35S]TBPS binding by GABA and benzodiazepines differs from that of GABA-activated whole cell currents. Benzodi- azepine modulation might have been more evident if the binding had been assayed earlier in the pre-equi- librium phase (Maksay and Simonyi, 1986).

On the basis of the present and previous findings (Korpi et al., 1993), the ANT rats probably have the most sensitive GABA A receptors in a defined cell population of all living rodents. Electrophysiological experiments on cerebellar sections from these animals are needed to verify the functional characteristics of the intact mutant receptors.

In conclusion, the GABA A receptors with the cere- bellum-specific wildtype or naturally occurring mutant c~6(Q100) subunits have the greatest sensitivity to their natural transmitter, GABA, irrespective of their benzo- diazepine pharmacology.

Acknowledgements

The authors wish to thank Pirkko Johansson for the binding assays and Sabine Griinewald for the cell culture. This study was supported by the Alexander von Humboldt Foundation, the Deutsche Forschungsgemeinschaft (SFB 317/B9), and the Academy of Fin- land.

References

Bonetti, E.P., W.P. Burkard, M. Gabl, W. Hunkeler, H.P. Lorez, J.R. Martin, H. Moehler, W. Osterrieder, L. Pieri, P. Polc, J.G. Richards, R. Schaffner, R. Scherschlicht, P. Schoch and W.E. Haefely, 1989, Ro 15-4513: Partial inverse agonism at the BZR and interaction with ethanol, Pbarmacol. Biochem. Behav. 31, 733.

Chen, C. and H. Okayama, 1987, High-efficiency transformation of mammalian cells by plasmid DNA, Mol. Cell. Biol. 7, 2745.

Eriksson, K. and M. Rusi, 1981, Finnish selection studies on alcohol-related behaviors: General outline, in: Development of Animal Models as Pharmacogenetic Tools, NIAAA Research Monograph No. 6, eds. G.E. McClearn, R.A. Deitrich and G. Erwin (U.S. Government Printing Office, Washington, DC), p. 87.

Hellevuo, K., K. Kiianmaa and E.R. Korpi, 1989, Effect of GABAer-

gic drugs on motor impairment from ethanol, barbital and lo- razepam in rat lines selected for differential sensitivity to ethanol, Pharmacol. Biochem. Behav. 34, 399.

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

Korpi, E.R., C. Kleingoor, H. Kettenmann and P.H. Seeburg, 1993, Benzodiazepine-induced motor impairment linked to point muta- tion in cerebellar GABA A receptor, Nature 36l, 356.

Korpi, E.R. and H. Liiddens, 1993, Regional GABA sensitivity of [35S]TBPS binding depends on GABA A receptor a subunit. Mol. Pharmacol. (in press).

Korpi, E.R., H. Liiddens and P.H. Seeburg, 1992a, GABA A antago- nists reveal binding sites for [35S]TBPS in cerebellar granular cell layer, Eur. J. Pharmacol. 211,427.

Korpi, E.R., M. Uusi-Oukari and K. Wegelius, 1992b, Substrate specificity of diazepam-insensitive cerebellar [3H]Ro 15-4513 binding sites, Eur. J. Pharmacol. 213, 323.

Laurie, D.J., P.H. Seeburg and W. Wisden, 1992, The distribution of 13 GABA A receptor subunit mRNAs in the rat brain. II. Olfac- tory bulb and cerebellum. J. Neurosci. 12, 1063.

Liiddens, H., D.B. Pritchett, M. Kohler, 1. Killisch, K. Keiniinen, H. Monyer, R. Sprengel and P.H. Seeburg, 1990, Cerebellar GABA A receptor selective for a behavioural alcohol antagonist, Nature 346, 648.

Liiddens, H. and W. Wisden, 1991, Function and pharmacology of multiple GABA A receptor subunits, Trends Pharmacol. Sci. 12, 49.

Maksay, G. and M. Simonyi, 1986, Kinetic regulation of convulsant (TBPS) binding by GABAergic agents, Mol. Pharmacol. 30, 321.

Malminiemi, O. and E.R. Korpi, 1989, Diazepam-insensitive [3H]Ro 15-4513 binding in intact cultured cerebellar granule cells, Eur. J. Pharmacol. 169, 53.

Pritchett, D.B., H. Sontheimer, B.D. Shivers, S. Ymer, H. Ketten- mann, P.R. Schofield and P.H. Seeburg, 1989, Importance of a novel GABA A receptor subunit for benzodiazepine pharmacol- ogy, Nature 338, 582.

Puia, G., M.R. Santi, S. Vicini, D.B. Pritchett, R.H. Purdy, S.M. Paul, P.H. Sceburg and E. Costa, 1990, Neurosteroids act on recombinant human GABA A receptors, Neuron 4, 759.

27

Sapp, D.W., U. Witte, D.M. Turner, B. Longoni, N. Kokka and R.W. Olsen, 1992, Regional variation in steroid anesthetic modulation of [3sS]TBPS binding to y-aminobutyric acid A receptors in rat brain, J. Pharmacol. Exp. Ther. 262, 801.

Schofield, P.R., M.G. Darlison, N. Fujita, D.R. Burt, F.A. Stephen- son, H. Rodriguez, L.M. Rhee, J. Ramanchandran, V. Reale, T.A. Glencorse, P.H. Seeburg and E.A. Barnard, 1987, Sequence and functional expression of the GABA A receptor shows a lig- and-gated receptor super-family, Nature 328, 221.

Squires, R.F., J.E. Casida, M. Richardson, E. Saederup, 1983, [35S]t-Butylbicyclophosphorothionate binds with high affinity to brain-specific sites coupled to y-aminobutyric acid-A and ion recognition sites, Mol. Pharmacol. 23, 326.

Squires, R.F., E. Saederup, 1987, GABA A receptor blockers reverse the inhibitory effect of GABA on brain-specific [35S]TBPS bind- ing, Brain Res. 414, 357.

Uusi-Oukari, M. and E.R. Korpi, 1990, Diazepam sensitivity of the binding of an imidazobenzodiazepine, [3H]Ro 15-4513, in cere- bellar membranes from two rat lines developed for high and low alcohol sensitivity, J. Neurochem. 54, 1980.

Uusi-Oukari, M. and E.R. Korpi, 1992, Functional properties of GABA A receptors in two rat lines selected for high and low alcohol sensitivity, Alcohol 9, 261.

Wieland, H.A., H. Liiddens and P.H. Seeburg, 1992, A single histi- dine in GABA A receptors is essential for benzodiazepine agonist binding, J. Biol. Chem. 267, 1426.

Wisden, W., A. Herb, H. Wieland, K. Kein/inen, H. Lfiddens and P.H. Seeburg, 1991, Cloning, pharmacological characteristics and expression pattern of the rat GABA A receptor a4 subunit. Febs Lett. 289, 227.

Wisden, W, DJ. Laurie, H. Monyer and P.H. Seeburg, 1992, The distribution of 13 GABA A receptor subunit mRNAs in the rat brain. I. Telencephalon, diencephalon, mesencephalon, J. Neu- rosci. 12, 1040.

Wisden, W. and P.H. Seeburg, 1992, GABA A receptor channels: from subunits to functional entities, Curr. Opin. Neurobiol. 2, 263.

Ymer, S., P.R. Schofield, A. Draguhn, P. Werner, M. K6hler and P.H. Seeburg, 1989, GABA A receptor /3 subunit heterogeneity: functional expression of cloned cDNAs, EMBO J. 8, 1665.