GABA Antagonists Differentiate Between Recombinant GABAJ Benzodiazepine Receptor Subtypes

The Journal of Neuroscience, October 1995, 75( 10): 6957-6962

GABA Antagonists Differentiate between Recombinant GABAJ Benzodiazepine Receptor Subtypes

Hattmut Liddens’ and Esa R. Korpiizz ‘Laboratory of Molecular Neuroendocrinology, Center for Molecular Biology, D-69120 Heidelberg, Germany and 2Biomedical Research Center, Alko Group Ltd., SF-00101 Helsinki, Finland

Seventeen rat GABA, receptor subtypes were transiently 1991; Harvey et al., 1994; Korpi et al., 1994), the assemblywith expressed in the human embryonic kidney 293 cell line the 6 (Shivers et al., 1989) or p subunits(Cutting et al., 1991, from al, CX~, ar3, (~5, or ar6 variants with any of the three p 1992). All natural GABA, receptors appear to recognize the subunits and ~2s or ~3. We obtained fingerprints in the GABA analog muscimol, the GABA antagonistsSR 95531 and form of subtype characteristic concentration-response bicuculline and the cage convulsant %S-TBPS (Olsen et al., curves of %-TBPS binding to GABA and the GABA, an- 1990; Korpi and Liiddens, 1993). tagonists SR 95531 and bicuculline. %-TBPS binding is modulatedby most, if not all, compounds cw363y2Si3 and cw563y2Sl3 containing receptors effec- interacting with GABA, receptors(for review, see Liiddens et tively recognized 35S-TBPS but not when 63 was replaced al., 1994). Therefore, 35S-TBPSbinding lends itself as a univer- by the 61 or 62 subunit. This indicates a specific interac- sal tool to analyze GABA, receptor heterogeneity in combination of (Y and p variants to form high-affinity 35S-TBPS bind- tion with compoundsrecognizing recombinantly expressedre- ing sites. ceptors (Korpi and Ltiddens, 1993; Im et al., 1994; Liiddenset At low levels GABA allosterically increased %-TBPS al., 1994; Wafford et al., 1994; Korpi et al., 1995) or native binding to all receptors with the concentration and mag- receptors(Gee et al., 1988; Peris et al., 1991; Korpi et al., 1992; nitude depending on the subunit combination. Exchange of Sapp et al., 1992; Liljequist and Tabakoff, 1993). Physiologi- the p variant did not alter the concentration-response cally relevant concentrationsof GABA agonistslike GABA it- curves for al and a6 containing receptors but did so for self and muscimolinhibit 35S-TBPSbinding and GABA, antag- a2 containing receptors. ~11261~3 receptors displayed onists block the GABA-induced decreaseof this binding in a strong GABA-induced stimulation of %&TBPS binding, concentration-dependentmanner (Ticku and Ramanjaneyulu, whereas binding to ~1263~3 receptors was marginally in- 1984; Squiresand Saederup,1987; Korpi et al., 1992; Korpi and creased. Liiddens, 1993). Whereasthe binding constantsfor 3H-muscimol SR 95531 and bicuculline decreased 35S-TBPS binding to and 3H-GABA are similar for a number of native (Olsen et al., all y3 containing receptors. In addition, bicuculline was ef- 1990; Bureau and Olsen, 1993) and recombinantreceptors (Kor- fective on cwl pxy2 receptors. SR 95531 was threefold more pi and Ltiddens, 1993; Ltiddens et al., 1994), 35S-TBPSbinding potent than bicuculline in reversing GABA-induced modu- can be usedto differentiate GABA, receptor subtypesdue to its lation of “S-TBPS binding in most receptor types, but was sensitivity to allosteric modulation by GABA (Korpi and Ltid- 30-fold more potent on ~~261~3 and 01661~2s receptors. dens, 1993; Ltiddens et al., 1994). 35S-TBPSbinding to cere- We conclude that the recognition site for GABA analogs bellar granule cell-specific GABA, receptorsis highly sensitive on GABA, receptors is jointly affected by all three subunit to GABA and can be blocked by residual endogenousGABA classes present in a receptor. concentrations,but at least IO-fold higher GABA concentrations [Key words: GABA, SR 95531, bicuculline, 35S-TBPS, al- are neededto inhibit 35S-TBPSbinding in other brain areas.The losteric coupling, subunit interaction] high GABA sensitivity can be mimicked by the coassemblyof a6, B2, and y2 subunitsin the heterologoushuman embryonic GABA, receptorsare pentameric proteins, their subunitsbeing kidney 293 (HEK 293) cell system (Korpi and Ltiddens, 1993), encodedby different genes(Seeburg et al., 1990). More than 54 but lower GABA sensitivity of 35S-TBPSbinding can be asso- GABA,/benzodiazepine (BZ) receptorsmay assemblein mam- ciated with a number of different GABA, receptors. mals in subunit combinationsof aiBjyk, with i = 1-6, j = l- The affinities for the antagonistsSR 95531 and bicuculline 3, and k = 1-3, assuminga single stoichiometry for each as- differ between native GABA, receptors (Olsen et al., 1984; sembly (Ltiddens and Wisden, 1991) and neglecting splice var- McCabe et al., 1988). Therefore, it appearedpossible to further iants (Whiting et al., 1990; Bateson et al., 1991; Kofuji et al., subtype low GABA-sensitive 35S-TBPSreceptors. A large number of recombinant ternary GABA, receptors differing in the three subunit components(a, B, or r), was investigated for the Received Feb. 13, 1995; revised June 9, 1995; accepted June 14, 1995. This study was funded, in part, by the DFG (SFB 3 17) and the Academy of propertiesof 35S-TBPSbinding as affected by GABA, SR 95531 Finland. We thank Sabine Grttnewald for help in cell culture work and Peter and bicuculline. The identified characteristicsof the recombinant H. Seeburg for his support. receptorsallowed us to ascribepharmacological fingerprints to Correspondence should be addressed to Hartmut Ltiddends, Clinical Re- search Group, Department of Psychiatry, Uni Maim, Verfttgungsgebaude, Ob- most subunit combinations which has been impossiblewhen ere Zahlbacher Strasse 63, D-55101 Mainz, Germany. basedsolely on the benzodiazepinepharmacology of thesere- Copyright 0 1995 Society for Neuroscience 0270-6474/95/156957-06$05.00/O ceptors (Ltiddens et al., 1994). 6956 Ltiddens and Korpi - Effect of GABA Antagonists on Recombinant GABA, Receptors

Table 1. Effect of exchanging p variants in (u2pxy213 and a3pxy213 receptors on 3H-Ro15-4513, 35S-TBPS, and 3H-muscimol binding

Y2 Y3 Pl I32 P3 Pl P2 P3 a2 3H-Ro15-4513 113 ? 20 91 * 4 212 2 3 51 5 3 10 f 5 17 ? 8 ‘S-TBPS 26 + 6 41 2 7 56 k 7 15 t 2 35 k 13 74 t 28 3H-Muscimol 60 t- 18 49 2 11 62 + 6 85 t 25 70 t 22 119 ? 46 a3 3H-Ro15-4513 151 5 35 83 t- 15 222 k 136 94 2 48 17 t 5 56 t 12 3%TBPS 724 353 27 2 5 4+2 425 34 t 2 3H-Muscimol 371 _t 80 51 t 11 191 ? 75 145 -e 52 48 2 7 320 t 80 012@y2, 012pxy3, cy3@q2, and u3pxy3 subunit combinations were expressed in HEK 293 cells and membranes were prepared as described. Binding was performed with 6 nM WRo15-45 13, ‘STBPS, or ?H-muscimol. Results, expressed in fmol/mg protein, are the means 2 SEM of three experiments.

Materials and Methods recognition sites. All a2 receptors recognized 35S-TBPS,but Materials. All radioligands were purchased from Du Pont-New England only thosea3 receptorsthat containedthe B3 subunitbound this Nuclear (Dreieich, Germany). GABA, picrotoxinin, and bicuculline ligand to a substantialamount. We therefore characterized the were obtained from Sigma Chemicals (St. Louis, MO) and SR 95531 regulation of 35S-TBPS binding to (-wlBxy2/3, or2Bxy2/3, from Research Biochemicals Inc. (Natick, MA). Ro 15-4513 and Ro 15-1788 were kindly donated by Dr. Hunkeler, Hofmann-LaRoche (Ba- a3/33y2/3, a5B3y2/3, and a6Bxy2 receptors by GABA, SR sle, Switzerland). 95531, and bicuculline to establishunique properties of these Transfection and membrane preparation. Expression vectors (Pritch- receptor subtypes. ett and Seeburg, 1990) for the (Y, B, and y subunits were transfected in triple combination into HEK 293 cells (ATCC CRL 1573) as described Regulation of j5S-TBPSbinding by GABA previously (Korpi and Liiddens, 1993). For optimal receptor expression, final concentrations (p,g of vector DNA per 15 cm tissue culture plate) GABA modulated the 35S-TBPSbinding to all recombinantly were: ol, 5; u2 12; a3, 3; 015, 2; Bl, 3; B2, 25; B3, 1; y2S, 0.5; and expressedGABA, receptor combinations tested (Fig. 1). For ~3, 0.2. The y2S variant is abbreviated y2 in the remainder of the text. most receptors the neurotransmitterled to a sharp rise in 35S- The cells were washed 40 hr after transfection with phosphate-buffered TBPS binding at concentrationsvarying from 100 nM (al B2y3) saline, pH 7.4, at 37”C, harvested in ice-cold phosphate-buffered saline and 1 FM (alBxy3 and a2Bly3) to 10 PM (o3B3y2 and and centrifuged at 150 X g. The cell pellets were homogenized in an Ultraturrax homogenizer for 15 set, pelleted at 23,000 X g and used o3B3y3). The stimulation of the binding by GABA was most immediately or frozen at -80°C and recentrifuged with identical results. prominent on a2Bly3 receptors (Fig. lB), but very modest in The membrane pellets were resuspended in 50 mu T&citrate buffer, a2B3y3 receptors (Fig. 1B). Exchanging the B variants in pH 7.3. alBxy2 or o6Bxy2 receptorsdid not alter the GABA concen- Binding assays. Resuspended cell membranes (50-200 pg protein per tube) were incubated in a final volume of 0.5 ml of 50 mrvr TrYcitrate tration dependencyof the binding (Fig. lA,D), but GABA stim- buffer, pH 7.3, for 3H-Ro 15-4513 (6 nM) and “H-muscimol (6 nM), or ulation was pronounced in alB2y3 receptors and reduced in in 50 mu Tris/citrate buffer, pH 7.3, supplemented with 0.2 M NaCl for al B3y3 receptors. 35S-TBPSbinding to the three o6Bxy2 re- %-TBPS (2-6 nM). Nonspecific binding was determined by 10 KM Ro ceptors was completely blocked by 1 PM GABA. For all recep- 15-1788, 100 PM GABA,-and 20 FM picrotoxinin for the three radioli- tors besides(~3B3y3 receptors 10 FM GABA significantly re- gands, respectively. After 1 hr at 4°C (3H-Ro 15-4513 and 3H-muscimol) or 90 min at room temperature (24°C %S-TBPS), the assay mixtures duced 35S-TBPSbinding as compared to the respective binding were rapidly diluted to 5 ml with ice-cold 10 mu TrisMCl, pH 7.5, and maximum, but binding to o3 or CX~subunit containing receptors filtered through glass fiber filters (Schleicher and Schuell. no. 52). Fil- was not fully abolishedby up to 1 mu GABA. Thus, they rep- ters were imkerged in 4 ml of Packard Ultima Gold scintillation’fluid, resent a recognizable subsetof GABA, receptors. and the radioactivity determined in a Beckman liquid scintillation coun- ter using external standardization. Nonlinear regression was performed Since “YS-TBPS binding was greatly stimulated by GABA in on the INPLOT and PRISM programs (GraphPad Software, San Diego, somereceptor subtypes,we testedwhether (~3and a5 containing CA) to calculate the parameters of saturation isotherms and inhibition receptorsdevoid of evident basalbinding, that is, 013Bl/2y2/3 curves. and a5p1/2y2/3 receptors,would bind the cage convulsantwhen 10 PM GABA was added. We could not detect any 35S-TBPS Results binding to a5pll2y2l3 receptorsin the presenceof GABA (data p variant selectivity of 1x2and (~3 containing GABA, receptors not shown).However, the small amountof picrotoxinin-sensitive We expressed ternary GABA, receptors in HEK 293 cells con- ‘jS-TBPS binding to o3Bly2 receptors(Table 1) increasedthree- figured from a2 or a3 subunits with any of the three p subunits fold with 10 PM GABA, comparableto the relative magnitude and either the y2 or y3 variant. These receptors were tested for of the GABA effect on a3B3y2 receptors. their ability to bind 3H-Ro 15-45 13, an imidazobenzodiazepine recognizing all known BZ receptors (Liiddens et al., 1990; Olsen Intrinsic modulation of 35S-TBPSbinding by SR 95.531and et al., 1990), 35S-TBPS, the prototypic ligand for the GABA, bicuculline receptor channel and 3H-muscimolto identify the GABA-rec- 35S-TBPSbinding to most of the GABA, receptor subtypestest- ognition site directly. We observed that all receptorsbound high ed in the nominal absenceof GABA was not significantly af- levels of 3H-muscimoland moderateto high amountsof “H-R0 fected by 30 PM SR 95531 or 100 FM bicuculline (Figs. 2, 3). 15-4513 (Table I), indicating the presenceof GABA and BZ However, 35S-TBPSbinding to alB2y3, (~2B3y3, cr3B3y3, and The Journal of Neuroscience, October 1995, 15(10) 6959

Figure 1. Modulation of 35S-TBPS binding to recombinant receptors by - a2p3y3 GABA. HEK 293 cells were transfect- o:ool0101 Ii.1 i lb lob 10&l i Id0 lo&l ed with the subunit combi. nations indicated, harvested and the membranes n L D prepared as described in Materials and 150 Methods. The cell membranes were incubated with 2 IBM YS-TBPS for 90 min - a3p3y2 loo - a6filyZ at room temperature with GABA con- - a3p3y3 centrations ranging from 1 nh4 to 1 mu. + a6p2y2 SO The values are expressed as percentage - a5p3y2 -+-- a6p3y2 of control binding in the absence of - a5!3373 0 GABA. Experiments were performed three or four times with duplicate sam- 0.0010.01 0.1 1 10 loo loo0 ples and the results expressed as the av- [CABAL PM WLVd PM erage 2 SEM.

(-wSp3y3 receptors was reduced by SR 95531 in a concentration- 1993). We therefore studied %S-TBPS binding of recombinant dependent manner (Figs. 2A,B, 3), that is, from the group of y3 GABA, receptors in the presence of SR 95531 and bicuculline containing receptors only a2ply3 receptors were unaffected. Bi- plus GABA concentrations which block Yi-TBPS binding to cuculline differed from SR 95531 in its effects on certain recep- most receptor types (10 PM GABA in al, c~2, a3, and a5 con- tor subtypes. It reduced the binding in all al containing recep- taining receptors and 1 pM GABA in (~6 containing receptors) tors (Fig. 2C), but was less efficacious than SR 95531 in a3P3y3 and obtained concentration-response curves which differed receptors (Fig. 3). Thus, the spectrum of receptors that were among the receptors. intrinsically affected by SR 95531 was a subset of those affected The shape of the SR 9553 1 concentration-response curves for by bicuculline. ollp1/2y2 receptors were congruent (Fig. 4A), but the curves for ollp3y2 and alp2y3 receptors were right-shifted (Fig. 4A) as Antagonism of GABA modulated35S-TBPS binding by SR compared to (Y1 p 1/2y2 receptors. 95531 and bicuculline As little as 300 nM SR 95531 was sufficient (Fig. 4B) to It is nearly impossible to examine native GABA, receptors free convert the inhibitory effect of 10 FM GABA on a2ply3 re- of contaminating endogenous GABA (Korpi and Lilddens, ceptors to a stimulatory effect equivalent to 1 FM GABA (Fig. lB), whereas a2P3y3 receptors required higher antagonist concentrations to recover from GABA inhibition. We found a sim- p2y2 - aq33yZ - a1pzy3 ilar tendency for the three c~2pxy2 receptors (Fig. which 160 C 160 4B), A T points to a higher potency of SR 95531 on pl containing 120 012pxy213 receptors than on p2 or @3 containing receptors.

- a3p3y2 + a3p3y3 - a5p3y2

- a5@3y3 - a6ply2 --+- c&6&2 -@-- a6P3y2 Oil i lb A B

1604 I 160-J I

0:1 i lb 0.1 1 10 1 10 100 LSR955311, pM [bicucuUinel.pM [SR 955311, KM [Bicucullinel.pM Figure 2. Intrinsic effects of SR 95531 and bicuculline on 15S-TBPS Figure 3. Intrinsic effect of SR 95531 and bicuculline on %-TBPS binding of cwl and (~2 subunit containing recombinant receptors. HEK binding of a3, a5 and a6 subunit containing recombinant receptors. 293 cells were transfected with oil and a2 subunit combinations as HEK 293 cells were transfected with (~3, a5, and (-u6 subunit combi- indicated, harvested and the membranes prepared as described in Ma- nations as indicated and harvested, and the membranes were prepared terials and Methods. The cell membranes were incubated with 2 nM or as described in Materials and Methods. The cell membranes were in- 6 nM ((r2 containing receptors) YS-TBPS for 90 min at room temper- cubated with 2 nM ?S-TBPS for 90 min at room temperature with 0.1 ature with 0.1 FM to 30 FM SR 95531 (A, B) or 0.3 FM to 100 PM pM to 30 pM SR 95.531 (leftpanel) or 0.3 p,M to 100 p,M bicuculline bicuculline (C, D). The values are expressed as percentage of control (rightpanel). The values are expressed as percentage of control binding binding in the absence of GABA. Experiments were performed three in the absence of GABA. Experiments were performed three or four or four times with duplicate samples and the results expressed as av- times with duplicate samples and the results expressed as average 2 erage 2 SEM. SEM. 6660 Liiddens and Korpi l Effect of GABA Antagonists on Recombinant GABA, Receptors

Figure 4. Modulation of YS-TBPS 200 a1P1yZ 300- binding to recombinant receptors by - aw. SR 95531 in the presence of GABA. E 150 al w HEK 293 cells were transfected with s zoo- - @433@ the subunit combinations indicated and g loo a1l33Y2 loo- - 4w harvested, and the membranes were 50 alf32y3 prepared as described in Materials and - a33r3 0 0’ , I Methods. Cell membranes were incu- 0:1 i lb bated with 2 nM or 6 nM ((~2 containing receptors) YS-TBPS for 90 min at room D 150 temperature with 1 pM GABA (~6 containing receptors) or 10 pM GABA (all - u3!33y2 - ww other receptors) plus SR 9553 1 concen- - ci3p3y3 lo+J trations ranging from 0.1 pM to 30 pM. h 432y2 The values are expressed as percentage - w33yz of control binding in the absence of 50 - m33G GABA. Experiments were performed - a336 three or four times with duplicate samples and the results expressed as the av- 0.1 1 10 erage 2 SEM. [SR 955311. PM [SR 955311, WM

In a3 and 01.5containing receptors,SR 95531 was more potent functional allosteric interactions. Furthermore, they can be ap- in y2 than in y3 containing subtypes in reversing the GABA plied to simulate pharmacologicalbrain regional heterogeneity effect (Fig. 4C). However, GABA sensitivity was greater in y3 and predict subunit combinationsof native GABA, receptors. than in y2 containing receptors(Fig. IC), which could indicate The formation of the ionophore binding site recognized by that the affinity of SR 95531 is similar for the four receptor 35S-TBPSin a3y2l3 and or57213containing receptorsdepended subtypes. on the presenceof the l33 variant. Receptorscontaining the y3 Already 100 nM SR 95531 reversedthe inhibitory effect of 1 variant were generally more sensitive to GABA than y2 con- pM GABA on cx6ply2 and a6P2y2 receptors(Fig. 40). Tenfold taining receptors but the former receptors were more strongly higher antagonistconcentrations were neededto achieve a sim- modulatedby the GABA, antagonistsSR 9553 1 and bicuculline ilar result for (w6p3y2 receptors. in the absenceof GABA. The presentresults, thus,indicate roles In the presenceof GABA threefold higher concentrationsof for the p as well as the y subunitsin the transduction of the bicuculline were neededfor most recombinantGABA, receptors ligand binding processfrom the GABA recognition site to the to obtain concentration-responsecurves congruent to those of Cl--ionophore, and agree with the idea of a ternary receptor SR 95531 (Fig. 5A-D), which indicates a constant ratio of po- complex with intimate interactions between the subunits. The tency for SR 95531 and bicuculline. The two most pronounced pharmacologicalfingerprint of a2P 1~3 receptorsdeviated from exceptionswere a2ply3 receptors(Figs. 40, 5D) and a6ply2 receptors(Fig. 40,5D), where SR 95531 was much more potent that of other recombinantreceptors as its 35S-TBPSbinding was than bicuculline. strongly enhancedby GABA, and SR 95531 antagonizedthe GABA action 30-fold more potently than bicuculline. This Discussion makesthe a2ply3 subunit combination an excellent candidate This study presentsa novel pharmacologicalclassification of a to explain someof the rat brain regional heterogeneitythat can number of GABA, receptor subtypes.The data outline the im- be seenin the actions of the antagonistson ?S-TBPS binding. portance of particular subunitson receptor domainsand their The clear features of a2ply3 receptorsshould make it possible

A B Figure 5. Modulation of 35S-TBPS binding to recombinant receptors by bicuculline in the presence of GABA. - alply HEK 293 cells were transfected with - a1pzy2 the subunit combinations indicated and harvested, and the membranes were - n1psy2 prepared as described under Materials and Methods. Cell membranes were in- - a1pzy3 cubated with 2 nM or 6 nM (012 containing receptors) YS-TBPS for 90 min at room temperature with 1 pM GABA C 400 (o6 containing receptors) or 10 pM GABA (all other receptors) plus bicu- - 300 culline concentrations ranging from 0.3 & JLM to 100 PM. The values are ex- B pressed as percentage of control bind- *g 200 - ing in the absence of GABA. Experi- ments were performed three or four - times with duplicate samples and the loo m ,. results expressed as the average + i lb l&l 1 10 100 SEM. [Bicudlinel,pM [Bicuculline],~ The Journal of Neuroscience, October 1995, 15(10) 6961 to identify brain regions enriched in this subunit combination receptorsall y3 containing GABA, receptorsare affected by one using autoradiography on rat brain sections. or both of the GABA antagonistsin the absenceof GABA. Bicuculline seemsto have the samepotency as SR 95531 but a Role of /3 variants in forming 35S-TBPS binding site broader intrinsic specificity than SR 95531, leaving only We previously reported on the B variant selectivity of 015 con- a3B3y3 receptorsas more sensitive to SR 95531 than bicucul- taining 35S-TBPS receptors (Ltiddens et al., 1994). Neither al line. In the absenceof GABA the antagonistscould act on the nor cr6 containing receptors are B subunit selective (Ltiddens et GABA recognition site in a manner analogousto the inverse al., 1994; Korpi et al., 1995), and now we show that the (-wZ agonistaction at BZ recognition site, but the possibility remains variant could also form ?S-TBPS recognizing receptors with all that the action(s) proceedsthrough a novel, independentrecog- three B subunits (Table 1). The u3 variant resembles the a5 nition site(s). variant since a3B1/2y2/3 and aSB1/2y2/3 receptors bound little We evaluated the effects of SR 95531 and bicuculline on 35S- YS-TBPS when compared to the corresponding B3 containing TBPS binding in the presenceof 1 pM GABA (a6Bxy2) or 10 receptors. However, in contrast to a5B1/2y2/3 receptors %- FM GABA (all other receptors) to comparethe potency of the TBPS binding to the corresponding (~3 receptors was signifi- two antagonistsin reversing the GABA induced effect on 35S- cantly increased by 10 mu GABA which could indicate a lower TBPS binding. For most receptors threefold higher concentra- affinity of %-TBPS to a3Bll2y2/3 receptors than to a3B3y2/3 tions of bicuculline than SR 95531 were neededto obtain con- receptors rather than differences in the number of binding sites. gruent SR 95531 and bicuculline concentration-responsecurves Though it is still open whether our findings are specific for HEK in the presenceof GABA (compareFigs. 4, 5). A ratio of three 293 cells (compare Im et al., 1994) or for GABA, receptor sub- for the potency of SR 95531 and bicuculline has been reported units from rodents (compare Wafford et al., 1994), our data in- earlier for 35S-TBPSbinding to whole rat brain membranes dicate that the ability of GABA, receptors to bind 35S-TBPS (Squiresand Saederup,1987). However, all B3 variant receptors with high affinity requires the proper interaction of two of the exhibited a lower sensitivity to SR 95531 than the corresponding three different subunits participating in the formation of the pen- Bl- or BZcontaining receptors.In contrast,a2Bly3 and a6Bly2 tamer and hence is prone to subtle alterations in any of the receptors were striking examplesof highly SR 95531 sensitive variants. receptor subtypes.For thesereceptor types a 30-fold lower concentration of SR 95531 than of bicuculline was neededto reach Role of subunits in the allosteric interaction with the 35S-TBPS similar levels of GABA-modulated 35S-TBPSbinding. binding site The concentration-responsecurves for bicuculline (plus con- Modulation of 35S-TBPS binding by GABA or its analogs ap- stant GABA) were generally similar when the B variants were pears to be a function of all participating subunit classes. Most exchanged in a axBxyx combination, arguing for similar bicu- y3 containing receptors were more sensitive to GABA modu- culline affinities of receptor combinations differing in the B lation of 35S-TBPS binding than y2 containing receptors. The component. Indeed,the affinity of GABA and bicuculline seems transition from y2 to y3 caused a left shift of otherwise fairly to vary within a narrow range (Ltiddens et al., 1990, 1994; Bu- congruent concentration-response curves in the al Bxy2l3, reau and Olsen, 1993). Together with our data on SR 95531 a3B3y2/3 and &p3y2/3 receptor pairs (Fig. lA,C), but affected concentration-response,this leadsto the conclusionthat the af- only the magnitude of the GABA effect in a2 receptors (Fig. finity of Bl containing receptorsfor SR 95531 is lo-fold higher IB). than of the correspondingB2 or B3 containing receptors. alB213y3 receptors were more sensitive to GABA than the The y3 subunit was associatedwith high GABA sensitivity corresponding y2 containing receptors, but the affinity of these (see above), which was corroborated by the concentration-re- receptor subtypes for GABA is identical (Korpi and Ltiddens, sponsecurves of SR 95531 and bicuculline, that is, generally 1993; Liiddens et al., 1994). Hence, the observed difference be- higher concentrations of GABA antagonistswere needed to tween the receptors reflects variations in allosteric coupling be- counteract the GABA modulation of 35S-TBPS binding in tween GABA and 35S-TBPS binding site. The same mechanism oxBxy3 receptorsthan in c-uxBxy2receptors. Since most y3 con- of action may hold true for the higher sensitivity of y3 receptors taining receptorsdisplayed a more pronouncedinhibition of 35S- in the a3P3y2/3 and a5/33y2/3 receptor pairs. TBPS binding by the antagonistsin the absenceof GABA, the Only limited data are available on GABA concentration-re- resultsclearly dissociatethe intrinsic actions of the antagonists sponseof recombinantGABA, receptors, but the known EC,, from their GABA antagonism. values determined in Xenopus laevis oocytes, alB2y2, 20 pM GABA antagonist projiles as an aid to identify native GABA, (Korpi et al., 1995); a3fkill2y2, 300 FM (Sigel et al., 1990); receptor subtypes a5BU2y2, 15 p,M (Sigel et al., 1990); and a6B2y2, 2 FM (Korpi et al., 1995) correlate well with the deflection point of the con- a6Bly2 receptor and a2Bly3 receptors are the only recombi- centration-responsecurves of YS-TBPS binding for GABA, that nant receptorswith a potency ratio of SR 95531 to bicuculline is, they are lo-30-fold higher than the GABA concentration around 30, but in a study employing furosemide as a tool to needed to achieve maximal %S-TBPS binding. If this pattern differentiate GABA, receptors we could exclude the (w6Bly2 holds true, an EC,, value of approximately 1 PM can be esti- combination as a naturally occurring GABA, receptor subtype mated for cylBxy3 receptors, that is, IO-fold lower than for (Korpi et al., 1995). Therefore, native GABA, receptors, which alBxy2 receptors. The EC,, values for a3B3y3 and cx5B3y3 are drastically more sensitive to SR 95531 than to bicuculline receptorswould then be about 30 and 1 PM, respectively. These should be composedof (~2, Bl, and y3 subunits. We are cur- predictions, except for olBxy3 receptors, which was not eval- rently investigating whether the properties of recombinant uated, have been recently confirmed (Ebert et al., 1994). a2Bly3 receptorsare present in rat brain regions with high (-w2 Most receptor subtypesare insensitive to the intrinsic effects mRNA expression.These receptors should constitute a minor of SR 95531 and bicuculline, but with the exception of a2Bly3 fraction of all brain GABA, receptors, but could be identified 6962 Likkiens and Korpi . Effect of GABA Antagonists on Recombinant GABA, Receptors by comparison of pharmacological fingerprints obtained by 35S- subunit abolishes GABA, receptor function. J Neurochem 63: 1167- 1170. TBPS autoradiography of rat brain sections and by ligand bind- Korpi ER, Kuner T, Seeburg PH, Liiddens H (1995) A selective an- ing assays of recombinant GABA, receptors. Our data indicate tagonist for the cerebellar granule cell-specific y-aminobutyric acid that SR 95531 differentiates this receptor from other GABA, type A receptor. Mol Pharmacol 47:283-289. receptor subtypes based on its potent GABA antagonism and Liljequist S, Tabakoff B (1993) Bicuculline-produced regional differ- not on its strong intrinsic activity on 35S-TBPS binding. ences in the modulation of %S-TBPS binding by GABA, pentobar- bital and diazepam in mouse cerebellum and cortex. J Pharmacol Exp GABA analogs regulate 35S-TBPS binding of a wide range of Ther 264:638-647. recombinant GABA, receptors. Besides being used to determine Ltiddens H, Wisden W (1991) Function and pharmacology of multiple the subunit composition of native GABA, receptors, our data GABA, receptor subunits. Trends Pharmacol Sci 12:49-5 1. yielded new insights into the pharmacology of these receptors Ltiddens H, Pritchett DB, Kohler M, Killisch I, Keinanen K, Monyer H, Sprengel R, Seeburg PH (1990) Cerebellar GABA, receptor se- which provide starting points to analyze the contribution of sin- lective for a behavioural alcohol antagonist. Nature 346:648-651. gle subunits to the GABA recognition site and to develop novel Liiddens H, Seeburg PHS, Korpi ER (1994) Impact of B and y variants GABA, receptor subtype selective compounds not recognizing on ligand binding properties of y-aminobutyric acid type A receptors. the BZ binding site. Mol Pharmacol 45:810-814. McCabe RT, Wamsley JK, Yezuita JP, Olsen RW (1988) A novel GA- References BA, antagonist ‘H-SR 9553 1: microscopic analysis of binding in the rat brain and allosteric modulation by several benzodiazepine and Bateson AN, Lasham A, Dar&on MG (1991) y-Aminobutyric acid, barbiturate receptor ligands. Synapse 2:163-173. receptor heterogeneity is increased by alternative splicing of a novel Olsen RW, Snowhill EW Wamsley JK (1984) Autoradiographic local- B-subunit gene transcript. J Neurochem 56:1437-1440. ization of low affinitv GABA receutors with 3H-bicuculline methoch- Bureau MH, Olsen RW (1993) GABA, receptor subtypes: ligand bind- loride. Eur J Pharmacol 99:247-8: ing heterogeneity demonstrated by photoaffinity labeling and auto- Olsen RW, McCabe RT, Wamsley JK (1990) GABA, receptor sub- radiography. J Neurochem 61: 1479-149 1. types: autoradiographic comparison of GABA, benzodiazepine, and Cutting GR, Lu L, Ohara BE Kasch LM, Montrose-Rafizadeh C, Don- convulsant binding sites in the rat central nervous system. J Chem ovan DM, Shimada S, Antonarakis SE, Guggino WB, Uhl GR, Ka- Neuroanat 3:59-76. zazian HH (1991) Cloning of the y-aminobutyric acid (GABA) pl Peris J, Shawley A, Dawson R, Abendschein KH (1991) Regulation cDNA: a GABA receptor subunit highly expressed in the retina. Proc of ?S-TBPS binding by bicuculline is region specific in rat brain. Nat1 Acad Sci USA g&2673-2677.- _ - Life Sci 49:54. Cutting GR. Curristin S. Zoghbi H. O’Hara B. Seldin ME Uhl GR Pritchett DB, Seeburg PH (1990) y-Aminobutyric acid, receptor o5- (1962) Identification of a iutative’ y-aminobutyric acid (GABA) re- subunit creates novel type II benzodiazepine receptor pharmacology. ceptor subunit p2 cDNA and colocalization of the genes encoding p2 J Neurochem 54: 1802-l 804. (GABRR2) and pl (GABRRl) to human chromosome 6q14-q21 and Sapp DW, Witte U, Turner DM, Longoni B, Kokka N, Olsen RW mouse chromosome 4. Genomics 12:801-806. (1992) Regional variation in steroid anesthetic modulation of %- Ebert B, Wafford KA, Whiting PJ, Krogsgaard-Larsen P, Kemp JA TBPS’ bind&g to y-aminobutyric acid, receptors in rat brain. J Phar- (1994) Molecular pharmacology of y-aminobutyric acid type A re- macol Exp Ther 262:801-808. ceptor agonists and partial agonists in oocytes injected with different Seeburg PH, Wisden W, Verdoorn TA, Pritchett DB, Werner P Herb A, ~1, B, y receptor subunit combinations. Mol Pharmacol 46:957-963. Ltiddens H, Sprengel R, Sakmann B (1990) The GABA, receptor Gee KW, Brinton RE, McEwen BS (1988) Regional distribution of a family: molecular and functional diversity. Cold Spring Harbor Symp Ro5-4864 binding site that is functionally coupled to the y-amino- Quant Biol 55:29+4. butyric acid/benzodiazepine receptor complex in rat brain. J Phar- Shivers BD, Killisch I, Sprengel R, Sontheimer H, Kohler M, Schofield macol Exp Ther 244:379-383. PR, Seeburg PH (1989) Two novel GABA, receptor subunits exist Harvey RJ, Chinchetru MA, Darlison MC (1994) Alternative splicing in distinct neuronal subpopulations. Neuron 3:327-337. of a Sl-nucleotide exon that encodes a putative protein kinase C Sigel E, Baur R, Trube G, MGhler H, Malherbe P (1990) The effect of phosphorylation site generates two forms of the chicken y-amino- subunit composition of rat brain GABA, receptors on channel func- butyric acid, receptor B2 subunit. J Neurochem 62: l&16. tion. Neuron 5:703-711. Im WB, Pregenzer JE Thomsen DR (1994) Effects of GABA and var- Squires RE Saederup E (1987) GABA, receptor blockers reverse the ious allosteric ligands on TBPS binding to cloned rat GABA, recep- inhibitory effect of GABA on brain-specific ?-TBPS binding. Brain tor subtypes. Br J Pharmacol 112:1025-1030. Res 414:357-364. Kofuji P, Wang JB, Moss SJ, Huganir RL, Burt DR (1991) Generation Ticku MK, Ramanjaneyulu R (1984) Differential interactions of of two forms of the y-aminobutyric acid, receptor y2-subunit in mice GABA agonists, depressant and convulsant drugs with [?S]-t-butyl- by alternative splicing. J Neurochem 56:713-715. bicyclophosphorothionate binding sites in cortex and cerebellum. Korpi ER, Ltlddens H (1993) Regional y-aminobutyric acid sensitivity Pharmacol Biochem Behav 21:151-158. of t-butylbicyclophosphoro[35S]-thionate binding depends on y-ami- Wafford KA, Bain CJ, Quirk K, Mckernan RM, Wingrove PB, Whiting nobutyric acid, receptor o( subunit. Mol Pharmacol 44:87-92. PJ, Kemp JA (1994) A novel allosteric modulatory site on the GA- Korpi ER, Liiddens H, Seeburg PH (1992) GABA, antagonists reveal BA, receptor B subunit. Neuron 12:775-782. . binding sites for “?-TBPS in cerebellar granular cell layer. Eur J Whiting P McKernan RM, Iversen LL (1990) Another mechanism for Pharmacol 211:427-428. creating diversity in y-aminobutyrate type A receptors: RNA splicing Korpi ER, Kuner T, Kristo P, Kohler M, Herb A, Liiddens H, Seeburg directs expression of two forms of y2 phosphorylation site. Proc Nat1 PH (1994) Small N-terminal deletion by splicing in cerebellar a6 Acad Sci USA 87:9966-9970.