Introduction

Summary Introduction Materials & Methods Results Discussion References

Fast inhibitory neurotransmission in the mammalian central nervous system is mediated principally through the GABAR. The GABAR is composed of a pentameric combination of 1-6, 1-4, 1-3, 1, and/or 1 subunit subtypes that form an intrinsic chloride ion channel. GABAR subunits are thought to have a membrane topology similar to that of nicotinic acetylcholine receptor subunits, with a large extracellular amino-terminal domain, four TMs (TM1-4), a small extracellular domain between TM2 and TM3, and a large cytoplasmic domain between TM3 and TM4 (Fig. 1A). The activity of GABARs is modulated through a large number of allosteric regulatory sites. The properties of these allosteric sites are influenced by the subunit subtype composition of the GABAR (see review in Ref. 1). The identity of the  subtype affects the receptor responsiveness to GABA and pentobarbital and to many allosteric agents, including benzodiazepines, divalent and trivalent cations, and furosemide.

View larger version (11K): [in this window] [in a new window]   Fig. 1.   Location of the chimera splice site. A, GABAR subunit. The structure of GABAR subunits consists of four TMs (TM1-4), a large amino-terminal extracellular domain, and a large intracellular domain between TM3 and TM4. Arrow, splice site for the chimeras was located within TM1 (not drawn to scale). B, Comparison of the amino acid sequence of TM1 of wild-type, mutant, and chimera subtypes. The full sequence of the rat 1 subtype is shown. Only differing residues and residues surrounding the splice site are shown for the other subtypes. Dashes, residues identical to the wild-type 1 subtype. Bold, leucine residue changed to a threonine residue in the mutant 1 subtype. Solid line, location of the splice site between the proline and cysteine residues.

The 1 and 6 subtypes are among the most divergent of the  family subtypes in their functional properties, and they share the least structural homology of the family members (59% amino acid identity) (2, 3). The differences in the amino acid sequences between 1 and 6 subtypes occur primarily in the large extracellular amino-terminal domain and the intracellular loop between TM3 and TM4, whereas the TMs and the TM2-3 extracellular domain are well conserved (83% amino acid identity). Receptors containing the 6 subtype along with  and 2 subtypes are more sensitive to GABA and to direct activation by pentobarbital than are receptors containing the 1 subtype. Receptors containing the 1 subtype along with  and 2 subtypes are potentiated by benzodiazepine agonists and lanthanum but are relatively insensitive to inhibition by zinc and furosemide. In contrast, receptors containing the 6 subtype along with  and 2 subtypes are insensitive to benzodiazepines, more sensitive to inhibition by zinc, and strongly inhibited by lanthanum and furosemide (2, 4-7). The structural bases for the differences in allosteric regulation of the 1 and 6 subtypes are not known except for the benzodiazepines. Potentiation by benzodiazepine agonists requires a histidine residue in the large amino-terminal extracellular domain (H101 in rat 1); the benzodiazepine-insensitive 4 and 6 subtypes contain an arginine residue at the equivalent location (8). Other residues in this domain have also been shown to contribute to the functional properties of several benzodiazepine ligands (9). No studies on the zinc site have been reported for the GABAR, although a histidine residue located in the extracellular amino-terminal domain is required for zinc inhibition of the structurally related 1 subtype of GABA_C receptors (10). Although structural domains important for direct activation by pentobarbital have not been localized, it has been shown that pentobarbital acts through different sites than does GABA (11). No information is available on the sites of action of lanthanum or furosemide on GABARs. All these agents act after application to the extracellular surface of the receptor and therefore likely bind to extracellular sites or the pore-forming TM. The amino-terminal domain is the largest extracellular domain, and the 1 and 6 subtypes share relatively low sequence homology in this region. Therefore, structural differences in the amino-terminal extracellular domain may be predominantly responsible for the differences in responsiveness of GABARs containing 1 and 6 subtypes to these regulatory agents.

To determine the role of the large extracellular amino-terminal domain in the functional properties of these GABAR regulatory sites, we created chimeric constructs of the rat 1 and 6 subtypes by interchanging their amino-terminal extracellular domains (Fig. 1). The chimeric subtypes contained the large extracellular domain and approximately one half of the TM1 of one of the  subtypes, with the remainder of the subunit structure from the other subtype, including TM2-4, the extracellular bridge between TM2 and TM3, and the large intracellular loop between TM3 and TM4. To generate the splice site, a restriction site for DraIII was created in the 1 subtype cDNA, converting a leucine residue to the threonine residue (L258T) present at this location in the 6 subtype (Fig. 1B). This mutation did not affect any of the functional properties of the 1 subtype examined in this study. Plasmids containing these constructs were transfected along with 3 and 2L subtypes into L929 fibroblasts, and the responsiveness of the GABARs to GABA, pentobarbital, diazepam, furosemide, zinc, and lanthanum was measured with whole-cell recording. The structural differences between the 1 and 6 subtypes could cause functional differences in the response to these modulators through a large number of mechanisms, including changes in binding or transduction properties. Because we examined the functional responses to these agents, not their binding properties, we could not distinguish among these mechanisms.

	Materials and Methods

Summary Introduction Materials & Methods Results Discussion References

Construction of mutant and chimeric  subtype cDNAs. Full-length cDNAs for the rat GABAR 1 (Dr. A. Tobin, University of California, Los Angeles) and 6 (F. Tan, University of Michigan) subtypes were subcloned into the pBluescript II KS⁽⁾ vector. To create a restriction site for DraIII in the 1 subtype, primers were created to make six nucleotide substitutions, converting the sequence TCTGCCATGCAT to CACTCCGTGCAC. This resulted in the change of a single amino acid, Leu258 to Thr258. A region between restriction sites for BsmI and NsiI containing the sequences to be modified was amplified using polymerase chain reaction techniques with a 5 primer of 5-GGATGAAAGATTAAAATTCAAAGGGACCCATGAC-3 and a 3 primer of 3-GCCGATGAAACAATAAAGTTTGTATGTGAGGCACGTG-5, which amplified the sequence from nucleotide 290 to nucleotide 783 of the open reading frame sequence of the 1 subtype. Oligonucleotide primers were synthesized at the University of Michigan DNA synthesis core facility. This region was spliced out of the wild-type sequence through digestion with BsmI and NsiI. The amplified fragment was digested with BsmI and AspHI, gel-purified, and ligated into the wild-type cDNA. Introduction of the mutation into the 1 sequence was verified with restriction mapping using NsiI and DraIII enzymes. To form the chimeras, both the mutant 1 and wild-type 6 subtype cDNAs were digested with DraIII, releasing a portion of the vector and the amino-terminal region of the subtype to the restriction site within TM1. The resulting fragments were gel-purified, swapped, and re-ligated to the other subtype to form the chimeric sequences. The 1/6 chimera and the 1_(L258T) subtype cDNAs were released from their vectors using HindIII and PstI. The 6/1 chimera was released using SpeI and XhoI. T4 DNA polymerase was used to fill in the ends. A BglII linker was then ligated to the ends, and the cDNAs were digested with BglII to allow ligation into the pCMVNeo vector using the BglII site. The sequence of the inserted PCR fragment was verified for all constructs with DNA sequencing.

Transfection of L929 cells. Full-length cDNAs for the rat GABAR 1 (Dr. A. Tobin), 3 (Dr. D. Pritchett, University of Pennsylvania), and 6 and 2L (F. Tan) subtypes were subcloned into the pCMVNeo expression vector (12) and transfected into the mouse fibroblast cell line L929 (American Type Culture Collection, Rockville, MD). Chimeric constructs and the 1_(L258T) mutant subtype were prepared as described above. For selection of transfected cells, the plasmid pHook-1 (InVitrogen, San Diego, CA) containing cDNA encoding the surface antibody sFv was also transfected into the cells. L929 cells were maintained in Dulbecco's modified Eagle's medium plus 10% heat-inactivated horse serum, 100 IU/ml penicillin, and 100 µg/ml streptomycin. Cells were passaged by a 5-min incubation with 0.5% trypsin/0.2% EDTA solution in phosphate-buffered saline (10 mM Na₂HPO₄, 150 mM NaCl, pH 7.3).

Electrophysiological recording solutions and techniques. For whole-cell recording, the external solution consisted of 142 mM NaCl, 8.1 mM KCl, 6 mM MgCl₂, 1 mM CaCl₂, 10 mM glucose and 10 mM HEPES, pH 7.4, and osmolarity adjusted to 295-305 mOsM. Recording electrodes were filled with an internal solution of 153 mM KCl, 1 mM MgCl₂, 5 mM K-EGTA, 10 mM HEPES, and 2 mM MgATP, pH 7.4, and osmolarity adjusted to 295-305 mOsM. These solutions provided a chloride equilibrium potential near 0 mV. Patch pipettes were pulled from thick-walled borosilicate glass with an internal filament (World Precision Instruments, Pittsburgh, PA) on a P-87 Flaming Brown puller (Sutter Instrument Co., San Rafael, CA) and fire-polished to a resistance of 5-10 M. Drugs were applied to cells using a "multipuffer" system with a 10-90% rise time of 70-150 msec (16). Currents were recorded with a List EPC-7 (List Electronics, Darmstadt, Germany) patch-clamp amplifier and stored on  videotape (Sony). All experiments were performed at room temperature.

Analysis of whole-cell currents. Whole-cell currents were analyzed off-line using the programs Axotape (Axon Instruments, Foster City CA) and Prism (GraphPAD Software, San Diego, CA). Normalized concentration-response data for the different isoforms were fit with a four-parameter logistic equation: current = maximum current/{1 + (10 (logEC₅₀  log[drug])*n)}, where n represents the Hill coefficient. All fits were made to normalized data with the current expressed as a percentage of the maximum current elicited by saturating GABA concentrations for each cell or, in the case of modulators, a percent of the response to GABA alone. Statistical tests were performed using the Instat program (GraphPAD). Comparisons of the receptor properties were performed with one-way analysis of variance and Tukey-Kramer multiple comparisons test with a statistical significance level of 0.05.

	Results

Summary Introduction Materials & Methods Results Discussion References

Responsiveness to GABA. All  subtype constructs in combination with 3 and 2L subtypes produced GABA-sensitive currents in L929 fibroblasts (Fig. 2A). Cells were voltage-clamped at 50 mV, and whole-cell currents were recorded in response to varying concentrations of GABA. The amplitudes of the maximum currents evoked by GABA were similar for all GABAR isoforms, with average ± standard error values of 974.6 ± 154.7 pA (132L, 13 cells), 1293.9 ± 235.9 pA (1_(L258T)32L, seven cells), 812.0 ± 179.0 pA (632L, nine cells), 1123.6 ± 354.3 pA (1/632L, nine cells), and 782.0 ± 226.0 pA (6/132L, seven cells). The maximum currents were not significantly different among isoforms. The GABAR isoforms exhibited different GABA sensitivities (Fig. 2B). The 132L isoform was ~7-fold less sensitive to GABA (average EC₅₀ = 12.1 ± 2.5 µM, seven cells) than the 632L isoform (average EC₅₀ = 1.8 ± 0.2 µM, six cells). The 1_(L258T)32L receptor had the same sensitivity to GABA as the wild-type 132L receptor, with an average GABA EC₅₀ value of 12.1 ± 1.7 µM (four cells). Receptors containing the 1/6 chimera were ~6-fold less sensitive to GABA than the 132L receptor, with an average GABA EC₅₀ value of 69.2 ± 6.4 µM (five cells). In contrast, receptors containing the 6/1 chimera were ~5-fold more sensitive to GABA than the 632L receptor, with an average EC₅₀ value of 0.34 ± 0.055 µM (six cells). The logEC₅₀ values for GABA were significantly different among all isoforms except for receptors containing the 1(L58T) mutation or the wild-type 1, which were not different from each other.

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Sensitivity to GABA, A, Representative whole-cell traces from transfected L929 fibroblasts. Fibroblasts were transfected with the subtypes indicated, and the peak current to varying concentrations of GABA was measured. The peak currents increased in a concentration-dependent manner for all isoforms. GABA was applied for 7-12 sec as indicated (bar) to cells voltage-clamped at 50 mV. The same time scale applies to all traces. B, Concentration-response relationships were constructed by normalizing the peak response to each concentration of GABA as a percentage of the maximum current response for each cell. Values are mean ± standard error. Data were fit with a four-parameter logistic equation. EC50 values and Hill slopes (n) of these fits were 0.36 µM, n = 1.1 for 6/132L (six cells), 1.7 µM, n = 1.2 for 632L (six cells), 11.5 µM, n = 1.3 for 132L (seven cells), 11.3 µM, n = 1.3 for 1(L258T)32L (four cells), and 67.9 µM, n = 1.4 for 1/632L (five cells).

Responsiveness to pentobarbital. Pentobarbital can directly activate GABARs with an efficacy dependent on the subunit subtype composition (7). The 1-containing receptors are less responsive to pentobarbital than are 6-containing receptors. Pentobarbital is more effective as an agonist than GABA at 632L receptors, producing larger maximum currents than GABA. The structural dependence of pentobarbital activation is different from that of GABA because mutations in the  subunit that prevent activation by GABA do not affect pentobarbital activity (11). In addition, activation by pentobarbital is not prevented by bicuculline, a competitive antagonist at the GABA site, although currents are blocked by the noncompetitive antagonist picrotoxin (7). The 632L isoform was highly sensitive to pentobarbital, with an EC₅₀ value of 44 µM and an average maximum current evoked by 300 µM pentobarbital that is 218.5 ± 58.7% (six cells) of that evoked by 600 µM GABA (Fig. 3). In contrast, receptors containing the 1 and 1_(L258T) subunits were nearly equally activated by 300 µM pentobarbital or 600 µM GABA, with average currents in response to 300 µM pentobarbital that were 85.2 ± 7.9% (six cells, 132L) and 88.2 ± 8.0% (five cells, 1_(L258T)32L) of the response to GABA. EC₅₀ values for pentobarbital were 101 µM (132L) and 126 µM (1_(L258T)32L). The amino-terminal extracellular domain seemed to be principally responsible for the differential effect of pentobarbital. The 1/632L isoform was 1-like in its response to pentobarbital, with 300 µM pentobarbital evoking currents 95.0 ± 3.3% (five cells) of the current response to 600 µM GABA with an EC₅₀ value of 131 µM. The efficacy of pentobarbital was not significantly different among the 1-, 1_(L258T)- and 1/6-containing receptors. The 6/132L isoform was 6-like, with 300 µM pentobarbital approximately twice as effective as 600 µM GABA (196.3 ± 15.0%, four cells) and an EC₅₀ value of 35 µM. The efficacy of pentobarbital at the 6/132L isoform was not significantly different from the response of the 632L isoform.

View larger version (25K): [in this window] [in a new window]   Fig. 3.   Direct activation by pentobarbital. A, Representative whole-cell traces from transfected fibroblasts. The peak response to 300 µM pentobarbital alone was compared with the peak response to 600 µM GABA for fibroblasts transfected with the subtypes indicated. GABA or pentobarbital was applied for 10-15 sec as indicated (bar) to cells voltage-clamped at 50 mV. Traces shown are 50 sec in duration. The same time scale applies to all traces. B, Concentration-response relationships were constructed by expressing the peak current in response to pentobarbital as a percentage of the response to 600 µM GABA for each cell. Symbols and bars, mean ± standard error. Data were fit with a four-parameter logistic equation.

Responsiveness to diazepam. A histidine residue in the amino-terminal extracellular domain of the 1 subtype (H101 in rat 1) is required for sensitivity to the benzodiazepine agonist diazepam (8). The 6 subtype contains an arginine residue in this location, which accounts for its insensitivity to benzodiazepine agonists. Consistent with these reports, the 132L, 1_(L258T)32L, and 1/632L receptor currents were all equally sensitive to potentiation by diazepam, with EC₅₀ values of 38 nM (132L, four cells), 38 nM (1_(L258T)32L, three cells), and 42 nM (1/632L, three cells), whereas the 632L (three cells) and 6/132L (four cells) receptor isoform currents were insensitive to diazepam (Fig. 4). The logEC₅₀ values for diazepam and the degree of current potentiation by 800 nM diazepam were not significantly different for the sensitive isoforms. The effects of 2 µM diazepam were not significantly different for the insensitive 632L and 6/132L isoforms.

View larger version (18K): [in this window] [in a new window]   Fig. 4.   Responsiveness to diazepam. A, Representative whole-cell traces from transfected L929 fibroblasts. Fibroblasts were transfected with the subtypes indicated, and the response to GABA or GABA plus diazepam was measured. The GABA concentration used was near the EC10 value for each isoform. GABA or GABA plus diazepam was applied for 7-12 sec as indicated (bar) to cells voltage-clamped at 50 mV. All traces shown are 50 sec in duration. The same time scale applies to all traces. B, Concentration-response relationships were constructed by expressing the peak current in response to diazepam as a percentage of peak current response to GABA alone for each cell. Symbols and bars, mean ± standard error. Data for the 132L, 1(L258T)32L, and 1/632L isoforms were fit with a four-parameter logistic equation.

Responsiveness to furosemide. Furosemide, a loop diuretic, is also a specific inhibitor of GABARs containing 4 or 6 subtypes (4, 17). Furosemide is much less effective at receptors containing an 1 subtype. 32L receptors containing wild-type 1 or the 1_(L258T) mutant showed a low sensitivity to furosemide, with average inhibition by 3 mM furosemide to 51 ± 3.6% (132L, five cells), and 58 ± 4.3% (1_(L258T)32L, four cells) of the current evoked by GABA alone (Fig. 5). The degree of current inhibition by 3 mM furosemide of the mutant and wild-type 1 subtypes was not significantly different. Due to the poor solubility of furosemide in water, complete concentration-response relationships could not be determined for these isoforms. The 632L current was highly sensitive to inhibition by furosemide, with an IC₅₀ value of 27 µM and nearly complete inhibition of the current (six cells). The appearance of the currents during inhibition by furosemide also differed (Fig. 5A). For 632L currents, higher concentrations of furosemide (30 µM) caused an increase in the apparent rate of desensitization, with the peak current declining rapidly. In contrast, for 132L currents, furosemide caused a uniform decrease in the whole-cell current, with no apparent difference between peak and steady state current.

View larger version (21K): [in this window] [in a new window]   Fig. 5.   Inhibition by furosemide. A, Representative whole-cell traces from transfected L929 fibroblasts. Fibroblasts were transfected with the subtypes indicated and the peak current response to GABA or GABA plus furosemide was measured. The GABA concentration used was near the EC50 value for each isoform. GABA or GABA plus furosemide was applied for 7-12 sec as indicated (bar) to cells voltage-clamped at 50 mV. All traces shown are 50 sec in duration. The same time scale applies to all traces. B, Concentration-response relationships were constructed by expressing the inhibition of the peak current by furosemide as a percentage of response to GABA alone for each cell. Symbols and bars, mean ± standard error. Data for the 632L, 1/632L, and 6/132L isoforms were fit with a four-parameter logistic equation.

Inhibition by zinc. The divalent cation zinc inhibits GABAR currents in a subunit subtype-dependent manner. Both and currents are highly sensitive to zinc inhibition, with IC₅₀ values of <5 µM (5, 18). The addition of a  subunit decreases the sensitivity to zinc by ~10-100-fold. However, the  subtype also influences the zinc sensitivity; receptors with 5 and 6 subtypes, even in combination with the 2L subtype, are more sensitive to current inhibition by zinc than are 1 subtype-containing receptors (5, 19).

View larger version (21K): [in this window] [in a new window]   Fig. 6.   Inhibition by zinc. A, Representative whole-cell traces from transfected L929 fibroblasts. Fibroblasts were transfected with the subtypes indicated and the peak current response to GABA or GABA plus zinc was measured. The GABA concentration used was near the EC50 value for each isoform. GABA or GABA plus zinc was applied for 7-12 sec as indicated (bar) to cells voltage-clamped at 50 mV. All traces shown are 40 sec in duration. The same time scale applies to all traces. B, Concentration-response relationships were constructed by expressing the inhibition of the peak current by zinc as a percentage of response to GABA alone for each cell. Symbols and bars, mean ± standard error. Data were fit with a four-parameter logistic equation.

Effect of lanthanum. The trivalent cation lanthanum affects GABAR currents in an  subunit subtype-dependent manner (6). 132L currents are potentiated by lanthanum, whereas 632L and 63 currents are inhibited. The currents from the 132L and 1_(L258T)32L isoforms were potentiated by lanthanum ~2-fold, with EC₅₀ values of 233 µM (132L, four cells) and 263 µM (1_(L258T)32L, five cells) (Fig. 7). The logEC₅₀ value and degree of potentiation by lanthanum for these isoforms were not significantly different. The 632L isoform was inhibited by lanthanum with an IC₅₀ value of 193 µM (five cells) and a maximal inhibition to ~40% of the control current.

View larger version (22K): [in this window] [in a new window]   Fig. 7.   Effect of lanthanum. A, Representative whole-cell traces from transfected L929 fibroblasts. Fibroblasts were transfected with the subtypes indicated, and the peak current response to GABA or GABA plus lanthanum was measured. The GABA concentration used was near the EC50 value for each isoform. GABA or GABA plus lanthanum was applied for 7-12 sec as indicated (bar) to cells voltage-clamped at 50 mV. All traces shown are 60 sec in duration. The same time scale applies to all traces. B, Concentration-response relationships were constructed by expressing the peak current in response to GABA plus lanthanum as a percentage of response to GABA alone for each cell. Symbols and bars, mean ± standard error. Data for all isoforms except the 6/132L isoform were fit with a four-parameter logistic equation.

	Discussion

Summary Introduction Materials & Methods Results Discussion References

To examine the contribution of the amino-terminal extracellular domain to the differential regulation of the 1 and 6 subtypes by several allosteric agents, we created chimeric subunits containing the entire amino-terminal extracellular domain and part of the first TM of one subtype spliced with the remaining subunit structure of the other subtype. This separated the largest extracellular portion of the subunit from the TMs. The carboxyl-terminal domains of the chimeras also included all the intracellular regions of the subtype in addition to a short extracellular sequence between TM2 and TM3. To form the chimeras, mutations were made in the 1 subtype sequence to create a restriction site; this resulted in the conversion of Leu258 to a threonine residue. This mutation had no effect on any of the pharmacological properties of the receptor examined in this study. The chimeric constructs efficiently expressed functional GABARs when cotransfected with 3 and 2L subtypes in L929 fibroblasts; this indicates that the chimeric  subtypes assembled with 3 and 2L subtypes to form functional GABARs. We examined the responses to GABA, pentobarbital, diazepam, furosemide, zinc, and lanthanum to determine whether the amino-terminal domain and/or the remaining carboxyl-terminal domains were involved in the actions of these agents (Table 1). Structural differences in the subtypes could cause changes in the effect or effectiveness of the allosteric modulators through many different mechanisms. The structures could form part of the binding site or the signal transduction pathways or influence either of these through remote effects on secondary, tertiary, or quaternary structures of the receptor. We can conclude from these results only whether these structural domains contribute to the differences between the properties of the 1 and 6 subtypes. It is likely that other regions of the subtypes also contribute to the binding or transduction sites but are not responsible for the functional differences among the subtypes.

GABA sites. Several amino acid residues in the large extracellular domain of many of the GABAR subunit families have been implicated in GABA binding (9, 20). Therefore, it is not surprising that the primary determinant of the GABA EC₅₀ value for channel activity resided in the amino-terminal domain of the chimeric subtypes. The 1/632L isoform was closer in EC₅₀ value to the 132L isoform than to the 632L isoform, and the 6/132L isoform was closer in EC₅₀ value to the 632L isoform than to the 132L isoform. Interestingly, the EC₅₀ values for the chimeras were not equal or intermediate to the wild-type subtypes. Instead, the 6/1 chimera produced higher affinity for GABA than did the 6 subtype, whereas the 1/6 chimera conferred lower affinity than did the 1 subtype. This suggests that the amino-terminal domain alone does not determine the GABA EC₅₀ value. The remaining carboxyl-terminal domains of the 1 and/or 6 subtypes must also contribute to the ability of GABA to bind and/or activate the GABAR channel. The degree of shift in EC₅₀ value was nearly identical for both chimeric constructs, although they occurred in opposite directions. Therefore, it may be that the TM2-3 or carboxyl-terminal extracellular regions of the 1 subtype contribute to an increased sensitivity to GABA and/or that those of the 6 subtype reduce the sensitivity to GABA.

Direct activation by pentobarbital. Pentobarbital acts as an agonist at GABARs in addition to its action as a positive allosteric modulator of GABAR activity (21, 22). Although the allosteric action is apparently independent of the subunit subtype composition of the receptor, direct activation is  subtype dependent (7, 17). At 6-containing receptors, pentobarbital is more efficacious than GABA, producing larger maximal currents. With 1-containing receptors, pentobarbital is as or slightly less efficacious than GABA, with a higher EC₅₀ value for activation than 6-containing receptors. Our results suggest that structural differences in the extracellular amino-terminal domain are responsible for the differences in pentobarbital activity. The 1/6 chimera had a response similar to the 1 subtype, whereas the 6/1 chimera was 6-like in both EC₅₀ value and efficacy compared with GABA. It has been shown that structures that alter GABA binding do not affect activation by pentobarbital (11), but the functional properties of both of these agonists seem to be primarily determined by the same general domain, the extracellular amino terminus.

Diazepam site. Our results are consistent with the finding that H101 (in the rat 1 subtype), which is located in the amino-terminal extracellular domain, is required for diazepam sensitivity (8). The 1/6 chimera conferred complete sensitivity to diazepam with an EC₅₀ value and maximum potentiation indistinguishable from those of the wild-type 1 subtype. The 6/1 chimera was completely insensitive to diazepam. This suggests that the amino-terminal extracellular domain is principally responsible for the contribution of the 1 subtype to the functional domains for diazepam and that the remaining carboxyl-terminal domains do not contribute to the difference in diazepam sensitivity between the 1 and 6 subtypes.

Furosemide sites. Furosemide inhibits 6 subtype-containing receptors with high affinity, whereas 1 subtype-containing receptors are almost 100-fold less sensitive. A high affinity site seemed to require the amino-terminal extracellular domain of the 6 subtype because the 6/1 chimera conferred the same sensitivity to furosemide as the wild-type 6 subtype. However, receptors containing the 1/6 chimera were also somewhat sensitive to furosemide, with an intermediate IC₅₀ value. This suggests that a second, lower affinity inhibitory site for furosemide was associated with the carboxyl-terminal domains of the 6 subtype. The appearance of the inhibited currents was also different for the two chimeras. The high affinity inhibition seen with the 6/1 chimera was uniform, with equal block of early and late currents. The lower affinity inhibition of the 1/6 chimera increased through the time of drug application, causing an increase in the degree of apparent desensitization. This is a common characteristic of open-channel blockers, which would be consistent with the involvement of the TMs in the formation of this inhibitory site. Although receptors containing the 6 subtype also showed this enhanced apparent desensitization, it appeared at lower concentrations of furosemide than required for receptors containing the 1/6 chimera. This could result from a positive interaction between the two proposed furosemide sites, resulting in an increased sensitivity at the blocking site when the high affinity site is bound. Interestingly, the 4 subtype, which shares many functional characteristics with the 6 subtype, is less sensitive to furosemide than the 6 subtype, with an IC₅₀ value of 162 µM when coexpressed with 3 and 2S subtypes in Xenopus laevis oocytes (17). This is similar to results seen with the 1/6 chimera (IC₅₀ = 180 µM) and suggests that the 4 subtype may contain the lower affinity site for furosemide but not the higher affinity site associated with the extracellular amino-terminal domain.

Zinc site. The 6 subtype confers a greater sensitivity to zinc inhibition than the 1 subtype. This characteristic was associated with the carboxyl-terminal portion of the chimeras. Receptors containing the 1/6 chimera had 6 subtype-like affinity for zinc, whereas receptors containing the 6/1 chimera had 1 subtype-like affinity. This is in contrast to the findings with the 1 subtype of the GABA_C receptor in which a histidine in the amino-terminal extracellular domain was responsible for zinc sensitivity (10). The inhibition of GABAR currents by zinc is noncompetitive and voltage independent. The primary effect of zinc on single-channel kinetics is to reduce the frequency of channel opening rather than to reduce channel open time or conductance (23, 24). This is consistent with zinc acting through an extracellular allosteric regulatory site rather than as a channel blocker. It has been suggested that the TM2-3 extracellular domain may be important in the regulation of zinc binding by the  subunit (24). The 6 subtype contains a histidine residue (H292) in this region, whereas the 1 subtype contains an asparagine residue (N301) in the equivalent location. Because histidine residues frequently contribute to zinc binding sites, this residue may play an important role in the zinc sensitivity of receptors containing the 6 subtype.

Lanthanum sites. Lanthanum affects the activity of 1 and 6 subtype-containing receptors in opposite directions, enhancing the activity of the 132L isoform while inhibiting the activity of the 632L isoform. The positive modulatory site seemed to be associated with the amino-terminal extracellular domain of the 1 subtype, with the 1/6 chimera conferring the same sensitivity and degree of enhancement as the 1 subtype. However, the 6/132L isoform showed an intermediate sensitivity, with only a slight enhancement by lanthanum. Neither receptor containing a chimera exhibited the inhibition seen with receptors containing the 6 subtype. The simplest explanation for this result is that structural domains from both regions of the 6 subtype are required for the formation of the inhibitory site.

Structure-function relationships of GABAR subunits. The structure of GABARs is complex, with five different subunits contributing to the functional properties of the receptor. We examined the structural domains responsible for contribution of the 1 and 6 subtypes to several allosteric regulatory sites of the GABAR. Whether the same or homologous structures form these sites for the other  subtypes is not known. In addition, the  subunit is clearly not solely responsible for the functional properties of these sites. The identity of the  subtype also influences the response of the receptor to GABA, pentobarbital, and furosemide (4, 7, 19, 25), and the  subunit contributes to the functional properties of the GABA, diazepam, zinc, and lanthanum sites (5, 6, 18, 26, 27). For the benzodiazepine site, a threonine residue (T142 in human 2) in the amino-terminal extracellular region of the 2 subtype has been found to be important for sensitivity (28). This is in the same general domain as the histidine residue that confers benzodiazepine sensitivity in the 1 subtype. The regions of the , , and  subunits that contribute to the functional properties of the other allosteric modulators are not necessarily structurally homologous to those for the  subunit. To completely understand the structural properties of the regulatory sites of the GABAR, the contributions that the other subunits make to these sites must also be examined.