Articles, Cellular/Molecular

State-Dependent Cross-Linking of the M2 and M3 Segments: Functional Basis for the Alignment of GABAA and Acetylcholine Receptor M3 Segments

Michaela Jansen and Myles H. Akabas
Journal of Neuroscience 26 April 2006, 26 (17) 4492-4499; https://doi.org/10.1523/JNEUROSCI.0224-06.2006

Abstract

Construction of a GABAA receptor homology model based on the acetylcholine (ACh) receptor structure is complicated by the low sequence similarity between GABAA and ACh M3 transmembrane segments that creates significant uncertainty in their alignment. We determined the orientation of the GABAA M2 and M3 transmembrane segments using disulfide cross-linking. The M2 residues α1M266 (11′) and α1T267 (12′) were mutated to cysteine in either wild type or single M3 cysteine mutant (α1V297C, α1A300C to α1A305C) backgrounds. We assayed spontaneous and induced disulfide bond formation. Reduction with DTT significantly potentiated GABA-induced currents in α1T267C-L301C and α1T267C-F304C. Copper phenanthroline-induced oxidation inhibited GABA-induced currents in these mutants and in α1T267C-A305C. Intrasubunit disulfide bonds formed between these Cys pairs, implying that the α-carbon separation was at most 5.6 Å. The reactive α1M3 residues (L301, F304, A305) lie on the same face of an α-helix. The unresponsive ones (A300, I302, E303) lie on the opposite face. In the resting state, the reactive side of α1M3 faces M2-α1T267. In conjunction with the ACh structure, our data indicate that alignment of GABAA and ACh M3 requires a single gap in the GABAA M2–M3 loop. In the presence of GABA, oxidation of α1T267C-L301C and α1T267C-F304C had no effect, but oxidation of α1T267C-A305C caused a significant increase in spontaneous channel opening. We infer that, as the channel opens, the distance and/or orientation between M2-α1T267 and M3-α1A305 changes such that the disulfide bond stabilizes the open state. This begins to define the conformational motion that M2 undergoes during channel opening.

Introduction

GABAA receptors are members of the Cys-loop receptor superfamily of neurotransmitter-gated ion channels that includes acetylcholine, glycine, and 5-HT3 receptors (Hevers and Lüddens, 1998; Karlin, 2002; Lester et al., 2004). Many general anesthetics and drugs used to treat anxiety, epilepsy, and sleep disorders target GABAA receptors, which mediate fast inhibitory synaptic transmission. Five homologous subunits, each with an extracellular N-terminal ligand-binding domain and four α-helical transmembrane segments (M1–M4), assemble around the central channel axis to form the receptors. The channel is lined by the five M2 segments, which are surrounded by an outer ring of helices, formed by the M1, M3, and M4 segments (Xu and Akabas, 1996; Miyazawa et al., 2003).

Recently, GABAA receptor homology models (Cromer et al., 2002; Trudell and Bertaccini, 2004; Ernst et al., 2005) were constructed based on the acetylcholine binding protein and Torpedo ACh receptor structures (Brejc et al., 2001; Miyazawa et al., 2003; Celie et al., 2004; Unwin, 2005). Homology models depend on the correct alignment of subunit sequences from different superfamily members. Uncertainty in sequence alignment limits the veracity of models. Alignment of ACh and GABAA M1 and M2 transmembrane segments is facilitated by highly conserved residues corresponding to GABAA α1R221, α1P233, α1F245, α1R255, α1L264, and α1P278 (Fig. 1) (Le Novere and Changeux, 1999; Ernst et al., 2005). M3 segment alignment is more complicated, because they are mostly hydrophobic with very limited sequence identity (Fig. 1). Thus, multiple sequence alignment programs introduce a variable number of gaps in the GABAA M2–M3 loop relative to the ACh receptor depending on the number of sequences and the specific subunits that are aligned. Progress in generating correct homology models requires an experimental determination of both the face of the GABAA M3 segment that interacts with M2 and the vertical position of M3 relative to defined levels in M2. Having the correct structural model is essential to understanding the molecular basis for general anesthetic and alcohol action on GABAA receptors.

Figure 1.
Figure 1.

Alignment of GABAA, Gly, 5-HT3, and ACh receptor subunit sequences. The lines on top indicate TM regions based on the extents of the TM segments in the 4 Å ACh receptor structure (Unwin, 2005). Residues in the GABAA α1 subunit that were mutated in this study are underlined and in bold. The extent of amino acid conservation at each position is indicated below the sequences: absolutely conserved (*), highly conserved (:), and moderately conserved (.). All sequences are the rat sequences except for the bottom four AChR subunits labeled “torma,” which are from T. marmorata.

In the current work, we determined the proximity of M2 and M3 residues using disulfide cross-linking between pairs of engineered cysteines. We used two M2 positions, α1M266 (11′) and α1T267 (12′), that we showed lie opposite to the M2 channel-lining face (Xu and Akabas, 1996). In the ACh receptor structure, the aligned residues are in close proximity to M1 and M3, respectively (Miyazawa et al., 2003). Based on the ACh receptor structure, GABAA M3 residues between α1A300 and α1A305 should lie in close proximity to α1T267. We constructed all of the double cysteine mutants between these M2 and M3 positions and assayed for spontaneous and induced disulfide bond formation in the absence and presence of GABA. Two M3 positions formed disulfide bonds in the closed state but not in the presence of GABA. Disulfide bond formation with a third position in the presence of GABA significantly increased the channels' spontaneous open probability. These results provide insight into the resting state structure and the conformational motion between M2 and M3 during channel gating in the Cys-loop receptor superfamily.

Materials and Methods

Alignment.

Sequences were obtained from UniProt (Universal Protein Resource; http://www.ebi.uniprot.org/index.shtml) or RCSB PDB (Research Collaboratory for Structural Bioinformatics; http://www.rcsb.org/pdb/), and aligned using ClustalW (http://www.ebi.ac.uk/clustalw/index.html). To improve the reliability of the alignment, we performed a multiple alignment with several other members of the Cys-loop superfamily of ligand-gated ion channels.

Mutagenesis and oocyte expression.

M2 segment residues can be named using an index numbering system that facilitates comparison between M2 segments of different members of the superfamily. At the cytoplasmic end, the 0′ position is defined as the conserved positively charged residue aligned with GABAA α1R255, and the 20′ position (Miller, 1989) is the residue aligned with the ACh receptor extracellular ring of charge (Imoto et al., 1988). The rat GABAA receptor α1, β1, and γ2S subunit cDNA constructs in the pGEMHE vector were used. The M2 mutations were introduced using α1 wild-type (wt) or α1M3 single mutants (V297C, A300C, L301C, I302C, E303C, F304C, A305C) as templates (Williams and Akabas, 1999) and the QuikChange mutagenesis kit (Stratagene, La Jolla, CA). Mutational primers (Sigma, St. Louis, MO) introduced M266C (5′-GTG ACG ACC GTT CTG ACC TGT ACA ACC TTG AGT ATC AGT GC-3′ and 3′-GCA CTG ATA CTC AAG GTT GTA CAG GTC AGA ACG GTC GTC AC-5′) or T267C (5′-ACG ACC GTT CTG ACC ATG TGT ACA TTG AGT ATC AGT GCC AG-3′ and 3′-CTG GCA CTG ATA CTC AAT GTA CAC ATG GTC AGA ACG GTC GT-5′) and a silent BsrGI site (mutated base pairs in bold; BsrGI recognition site underlined). The identity was verified by restriction digestion and DNA sequencing of the complete coding region. Plasmids were linearized with NheI before in vitro mRNA transcription with T7 RNA polymerase (Amplicap T7 High Yield Message Maker; Epicenter Technologies, Madison, WI; or mMessage mMachine kit; Ambion, Austin, TX). mRNA was dissolved in diethylpyrocarbamate-treated water and stored at −80°C. Female Xenopus laevis were purchased from Nasco Science (Fort Atkinson, WI). Stage V–VI oocytes were defolliculated with a 75 min incubation in 2 mg/ml Type 1A collagenase (Sigma) in OR2 (82.5 mm NaCl, 2 mm KCl, 1 mm MgCl2, and 5 mm HEPES; pH adjusted to 7.5 with NaOH). Oocytes were washed thoroughly in OR2 and kept in SOS medium (82.5 mm NaCl, 2.5 mm KCl, 1 mm MgCl2, 5 mm HEPES, pH 7.5) supplemented with 1% antibiotic–antimycotic (100×) liquid (10,000 IU/ml penicillin, 10,000 μg/ml streptomycin, and 25 μg/ml amphotericin B; Invitrogen, Carlsbad, CA) and 5% horse serum (Sigma). Oocytes were injected 24 h after isolation with 50 nl (10 ng) of a 1:1:1 mixture of rat α1:β1:γ2S subunit mRNA and were kept in horse serum medium for 2–10 d at 17°C. Mutant subunit mRNA was substituted for wt α1 subunit where necessary (Horenstein et al., 2001).

Electrophysiology.

Electrophysiological recordings were conducted at room temperature in a ∼250 μl chamber continuously perfused at a rate of 5–6 ml/min with Ca2+-free frog Ringer's buffer (CFFR) (115 mm NaCl, 2.5 mm KCl, 1.8 mm MgCl2, 10 mm HEPES, pH 7.5 with NaOH) using equipment and procedures described previously (Horenstein and Akabas, 1998). Currents were recorded from individual oocytes using two-electrode voltage-clamp at a holding potential of −60 mV. The ground electrode was connected to the bath via a 3 m KCl/Agar bridge. Glass microelectrodes had a resistance of <2 MΩ when filled with 3 m KCl. Data were acquired and analyzed using a TEV-200 amplifier (Dagan Instruments, Minneapolis, MN), a Digidata 1322A data interface (Molecular Devices, Union City, CA), and pClamp 8 software (Molecular Devices). Currents (IGABA) elicited by GABA applications were separated by at least 6 min of CFFR wash to allow complete recovery from desensitization. Currents were judged to be stable if the variation between consecutive GABA pulses was ≤10%.

Reagents.

GABA (Sigma) was prepared as a 100 mm stock solution in water. Dithiothreitol (DTT) (Sigma) was dissolved in water to obtain a 1 m stock solution and diluted into CFFR before each experiment. o-Phenanthroline (Sigma) was made as a 1 m stock solution in DMSO and CuSO4 as a 100 mm stock solution in water. CuSO4 and o-phenanthroline were mixed in CFFR before use to a final concentration of 100:200 μm or 10:20 μm Cu:Phen. Tris(2-carboxyethyl)phosphine HCl (TCEP) (Sigma) was dissolved in water to obtain a 100 mm stock and diluted into CFFR before each experiment. Picrotoxin was dissolved in DMSO at a 100 mm concentration and diluted in CFFR.

DTT effect.

Once stable GABA EC30–50 test currents (I) were achieved, 10 mm DTT was applied for 2 min. After a washout period of 6 min the GABA EC30–50 test current was recorded again (IDTT). Potentiation by application of DTT was calculated as follows: effect % = {(IDTT − I)/IDTT} × 100.

Cu:Phen effect.

Oocytes were pretreated with DTT and GABA test pulses (IDTT) were applied. Oocytes were then treated with 100:200 μm Cu:Phen for 2 min, and two more GABA test pulses were applied (ICu:Phen). A second application of 10 mm DTT for 2 min followed by GABA test pulses was used to test for the reversibility of the Cu:Phen effect. Inhibition by Cu:Phen induced disulfide bonds was calculated as follows: effect % = {(ICu:Phen − IDTT)/IDTT} × 100.

Cu:Phen effect in the presence of GABA, open/desensitized states.

For this set of experiments, we first determined the minimum application duration of 10:20 μm Cu:Phen that produces inhibition comparable with a 2 min application of 100:200 μm Cu:Phen. Preliminary experiments with the α1T267C L301C double mutant on initially fully reduced (10 mm DTT; 2 min) oocytes using alternating applications of 10:20 μm Cu:Phen for 1 min and GABA EC30 test currents showed that a comparable effect to the application of 100:200 μm Cu:Phen for 2 min was reached after a cumulative exposure to 10:20 μm Cu:Phen of 2 min.

To determine whether disulfide bonds were formed in the open state, we performed the following experiment: Once stable GABA EC30–50 test currents (I) were achieved, we applied 10 mm DTT for 2 min. The GABA EC30–50 test current was recorded again (IDTT) to determine the extent of spontaneously formed disulfide bonds. After a 20 s GABA ECmax pulse, cells were immediately treated with a mixture of the same GABA concentration and 10:20 μm Cu:Phen for 2 min. After washout of the reagents, two more GABA test pulses were applied (ICu:Phen + GABA). A second application of 10 mm DTT for 2 min followed by GABA test pulses was used to test for the reversibility of the Cu:Phen effect. Inhibition by disulfide bonds formed in the presence of GABA and Cu:Phen was calculated as follows: effect % = {(ICu:Phen + GABA − IDTT)/IDTT} × 100.

Cu:Phen induced oxidation was performed for each oocyte in the absence and in the presence of GABA. One-half of the oocytes were first oxidized in the presence and then in the absence of GABA, the other one-half in the reverse order.

Homology modeling.

The membrane-spanning domain of the rat GABAA receptor α1 subunit was modeled on the basis of the 4 Å structure obtained by electron microscopy of crystalline postsynaptic membranes with pentameric ACh receptors of Torpedo marmorata [Protein Data Bank (PDB) number 2BG9]. Using Deep View/Swiss Pdb-Viewer, version 3.7 (Guex and Peitsch, 1997; Schwede et al., 2003), we aligned the GABA α1 subunit sequence with the ACh E chain (γ chain) manually, taking the alignment that we obtained with ClustalW, as outlined above, into account. We generated two different models, one without a gap and one with a 1 aa gap in the M2–M3 loop in the GABA α1 sequence compared with the ACh E chain (γ chain) sequence. The ACh structure aligned with the raw sequence was submitted as a modeling request in Swiss Model. The resulting homology modeled α1 structure was subsequently subjected to energy minimization using the GROMOS 43B1 force field in vacuo that is part of Swiss-PdbViewer. Figures were rendered with POV-Ray v3.6 (http://www.povray.org/) (Bali and Akabas, 2004).

Results

Effect of DTT on wt and mutant 12′ Cys receptors

In electrophysiological experiments, we recorded GABA-induced currents (IGABA) from oocytes expressing GABA α1β1γ2 receptors. Where applicable, the α1 wt subunit was substituted by mutant α1 subunits, containing a single engineered Cys in M2 (11′ or 12′ Cys) or two engineered Cys [one in M2 (11′ or 12′ Cys) and one in M3]. DTT is a disulfide bond-reducing agent. It contains two sulfhydryl moieties and is thus able to convert protein disulfides to sulfhydryls via a thiol–disulfide exchange reaction. In wt α1β1γ2 receptors, DTT potentiated GABA-induced currents by 14 ± 4% (n = 6) (Fig. 2). Under the same conditions, the single mutant α1T267C, and the double mutants α1T267C-A300C, α1T267C-I302C, α1T267C-E303C, and α1T267C-A305C were not significantly different from wt (one-way ANOVA; Dunnett's multiple comparison test; p > 0.05). However, the currents in the double mutants α1T267C-L301C and α1T267C-F304C were potentiated by 71 ± 25% (n = 6) and 81 ± 32% (n = 8), respectively. The potentiation in the double mutants α1T267C-L301C and α1T267C-F304C was significantly different from wt receptors (one-way ANOVA; Dunnett's multiple comparison test; p < 0.05 and p < 0.01) (Fig. 2) and may be explained by the reduction of a spontaneously formed disulfide bond between the engineered Cys in M2 and M3. If a disulfide bond were to be formed between these Cys, the relative mobility of at least the M2 and M3 α-helices would be diminished, thus impairing receptor function.

Figure 2.
Figure 2.

The average effect of a 2 min application of 10 mm DTT on wt and 12′ Cys mutants. A positive effect indicates potentiation, and a negative effect indicates inhibition of subsequent GABA-induced currents. The asterisks indicate effects that are significantly different from the effect on wt by a one-way ANOVA, using Dunnett's multiple comparison test, p < 0.001. For each condition, the results of three to eight oocytes are averaged. Means ± SEM are shown.

Effect of Cu:Phen on wt and mutant 12′ Cys mutants

Under aerobic conditions (oxygen dissolved in CFFR) (with Cu:Phen in a molar ratio of 1:2 serving as a catalyst), oxygen oxidizes the sulfhydryl groups selectively to disulfides but rarely to higher oxidized states (Kobashi, 1968). We tested the effect of Cu:Phen (100:200 μm; 2 min) on fully reduced wt and 12′ Cys single and double mutant receptors. To reduce spontaneously formed disulfide bonds between engineered Cys pairs or between engineered Cys and small endogenous sulfhydryls like cysteine and glutathione, oocytes were first treated with DTT (10 mm; 2 min). Cystine and GSSG (oxidized glutathione) participate in disulfide interchange reactions and are capable of forming mixed disulfides with proteins. After DTT application, oocytes were allowed to equilibrate with CFFR, and GABA pulses were recorded with a frequency of one per 6 min. Subsequently, cells were treated with Cu:Phen (100:200 μm; 2 min), and after a washout period, the resulting GABA-induced currents were recorded (Fig. 3). There was no effect of oxidation neither in wt receptors, nor the 12′ single mutant α1T267C, nor in the double mutants α1T267C-V297C, α1T267C-A300C, α1T267C-I302C, and α1T267C-E303C. Cu:Phen inhibited GABA-induced currents in wt, α1T267C, α1T267C-A300C, α1T267C-I302C, and α1T267C-E303C by 11 ± 4% (n = 6), 8 ± 7% (n = 5), 3 ± 4% (n = 3), 15 ± 5% (n = 5), 1 ± 16% (n = 5) (one-way ANOVA; p > 0.05). However, the double mutants that were potentiated by DTT, namely α1T267C-L301C and α1T267C-F304C, and in addition also α1T267C-A305C (Fig. 3E) showed an inhibition of 45 ± 2% (n = 18), 45 ± 4% (n = 11), and 55 ± 7% (n = 5), respectively, that was significantly different from wt (one-way ANOVA; Dunnett's multiple comparison test; p < 0.001) (Fig. 4). The Cu:Phen-induced inhibition of GABA currents could be fully reversed by a second application of DTT. The ability to reverse the Cu:Phen inhibition with DTT implies that it is attributable to disulfide bond formation and not attributable to oxidation of one or both of the Cys to a higher-order oxidation state that would not be reducible by DTT. We infer that, by alternatively applying DTT and Cu:Phen, we cycle between fully reduced (free sulfhydryls) and fully oxidized (disulfide-bonded) receptors and that the inhibition in the oxidized state is caused by a disulfide bond that cross-links the engineered Cys in M2 and M3.

Figure 3.
Figure 3.

DTT/TCEP and Cu:Phen effect on wt and α1–12′ Cys mutants. GABA-induced current traces before and after 2 min applications of 10 mm DTT or 100:200 μm Cu:Phen from wt and mutant receptors are shown. A, Wild-type α1β1γ2S are not significantly affected by either DTT or Cu:Phen. B, α1T267C-E303Cβ1γ2S are not significantly affected by either DTT or Cu:Phen. C, D, α1T267C-F304C (C) and α1T267C-L301C (D) show spontaneous disulfide bond formation that is reversed by DTT and induced by Cu:Phen. In α1T267C-L301C-containing receptors, the current size decreases with each GABA pulse. After reaching a stable plateau, current can be recovered to the initial size after reduction with DTT. A Cu:Phen application inhibits currents to the same extent as reached after repetitive GABA applications. This inhibition is DTT reversible. E, In α1T267C-A305C-containing receptors, DTT application does not significantly alter the subsequent current. Cu:Phen, however, significantly inhibits these receptors. This inhibition is reversed by DTT application. F, α1T267C-L301C can be cycled between reduced and oxidized states by reducing with TCEP (1 mm; 2 min) and oxidizing with Cu:Phen (100:200 μm; 2 min). DTT (10 mm; 2 min) does not significantly potentiate TCEP-treated oocytes. All traces are separated by 6 min washes with CFFR from reagent or GABA applications to allow for washout of excess reagents or complete recovery from desensitization. All GABA current traces represent stable currents: two subsequent GABA currents were not different by >10%. Currents during reagent application are not shown. The bars above the traces indicate the duration of GABA applications; the arrows indicate the application of the indicated reagent.

Figure 4.
Figure 4.

The average effect of a 2 min application of 100:200 μm Cu:Phen on wt and 12′ Cys mutants. A positive effect indicates potentiation, and a negative effect indicates inhibition of subsequent GABA-induced currents. The asterisks indicate effects that are significantly different from the effect on wt by a one-way ANOVA, using Dunnett's multiple comparison test, p < 0.001. For each condition, the results of 3–18 oocytes are averaged. Means ± SEM are shown.

Spontaneous formation of disulfide bonds, in the absence of Cu:Phen, was often recorded after GABA applications in the receptors α1T267C L301C and α1T267C F304C (Fig. 3D). After stabilization of the current, further oxidation using Cu:Phen (100:200 μm; 2 min) did not lead to any additional inhibition of GABA-induced currents. Reduction by DTT increased the GABA-induced currents to at least their initial size. In turn, this increase could be reversed by oxidation with Cu:Phen. This spontaneous decrease in the GABA-induced current was not observed in all the oocytes. In some oocytes, the initial current was as big as after DTT application, whereas in others the initial current was as inhibited as after Cu:Phen application. We infer that these differences are attributable to changes in the redox environment of different batches of oocytes, thus leading to varying degrees of spontaneously formed disulfide bonds.

For α1T267C-L301C and α1T267C-F304C, we also determined whether the inhibition by Cu:Phen (100:200 μm; 2 min) could be reversed by application of TCEP (Fig. 3E). TCEP is a reducing agent that, contrary to DTT, forms only weak complexes with heavy metals (Krezel et al., 2003). Thus, if after TCEP application GABA-induced currents are potentiated, this can only be caused by reduction of disulfide bonds and not by chelating heavy metals. After a 2 min Cu:Phen application (100:200 μm), α1T267C-A301C receptors were inhibited by 52 ± 2% (n = 3) and α1T267C-F304C receptors by 60 ± 2% (n = 5). This inhibition could be reversed by a 1 min treatment with 2 mm TCEP. A subsequent application of DTT did not potentiate GABA currents any further (α1T267C-A301C: 0 ± 5%, n = 3; α1T267C-F304C: −7 ± 2%, n = 5). We infer that both the α1T267C-A301C and the α1T267C-F304C disulfide bond are accessible for both the smaller and membrane-permeant DTT and for the larger and triple negatively charged, and thus membrane-impermeant, TCEP.

Cu:Phen effect on the 12′ mutants α1T267C-L301C, α1T267C-F304C, and α1T267C-A305C in the absence and presence of GABA

In the absence of GABA, the channels are primarily in the closed, unliganded state. In the presence of GABA, the channels undergo rapid transitions between the open, desensitized and closed states. In addition, there are multiple open and desensitized states. Given the timescale of the transitions (milliseconds) and the timescale of the experiments (tens of seconds), we cannot distinguish in which state disulfide bond formation is occurring in the presence of GABA.

We examined the effect of agonist coapplication on Cu:Phen-induced disulfide bond formation. The effect of 10:20 μm Cu:Phen on α1T267C-L301C and α1T267C-F304C containing receptors was abolished when it was coapplied with GABA. The double mutant α1T267C-L301C was inhibited by 36 ± 2% (n = 4) by applying Cu:Phen (10:20 μm; 2 min) alone compared with 7 ± 4% (n = 4) by applying Cu:Phen (10:20 μm; 2 min) in the presence of a saturating GABA concentration. When compared with a paired t test, the effect in the closed state is significantly different from the one in the open state (p = 0.015). The α1T267C-F304C receptors were inhibited by 31 ± 1% (n = 4) by applying Cu:Phen alone and by −12 ± 6% (n = 4) by applying Cu:Phen together with GABA (paired t test; p = 0.0057). The propensity to form a disulfide bond in the closed state is increased for α1T267C-L301C and α1T267C-F304C.

In contrast to these two double mutants, the amount of inhibition of the GABA peak currents for the α1T267C-A305C mutant increased from 26 ± 3% (n = 4) with Cu:Phen alone to 67 ± 6% (n = 4) for Cu:Phen in the presence of GABA (paired t test; p = 0.0012). We conclude that the distance/relative orientation of the 12′ Cys in M2 and the 305Cys in M3 is different in the closed (absence of GABA) and open/desensitized (presence of GABA) states. In the double mutant, α1T267C-A305C, a significant change in leak current was induced by coapplication of Cu:Phen with GABA, but not by Cu:Phen alone (Fig. 5C). Subsequent treatment with DTT reversed both the increase in leak current and also the decrease in GABA-induced current amplitude. The increase in leak current could be transiently blocked by applying the open channel blocker picrotoxin (50 μm) (Fig. 5D). We infer that, when α1T267C-A305C is oxidized in the open state, the receptor is trapped in a conformation for which the spontaneous open probability is increased. Therefore, the current amplitude that can result from GABA application is decreased. If the increase in leak current after a combined GABA–Cu:Phen application is added to the GABA current amplitude after the Cu:Phen plus GABA application, the inhibition would be 24 ± 7% (n = 4), which is comparable with the inhibition caused by application of Cu:Phen alone.

Figure 5.
Figure 5.

Effect of a 2 min application of 10:20 μm Cu:Phen applied in the presence or in the absence of GABA on GABA-induced currents from oocytes expressing 12′ Cys mutants. A, B, In α1T267C-L301C (A)- and in α1T267C-F304C (B)-containing receptors, a 2 min application of Cu:Phen in the absence of GABA inhibited the subsequent GABA currents. This inhibition was reversed by DTT. A subsequent 2 min application of Cu:Phen in the presence of EC80 GABA did not inhibit the subsequent GABA currents. An additional DTT application did not change the current size. Reagents were applied at the downward arrows. Currents are not shown during reagent applications. C, In α1T267C-A305C receptors, a 2 min application of Cu:Phen in the absence of GABA inhibited the subsequent GABA current. Reduction with DTT reversed the inhibition of the GABA current. A subsequent 2 min application of Cu:Phen in the presence of GABA had two effects on the following GABA currents. A large increase in the holding current was observed (note the level at which the subsequent GABA current begins), and the magnitude of the subsequent GABA current was significantly reduced. It is notable that the magnitude of the total current, the sum of the holding current and the GABA-induced current, was similar to the GABA current before the Cu:Phen/GABA application. The increased holding current and the reduced GABA current were reversed by a subsequent reduction with DTT. D, Currents from an oocyte expressing α1T267C-A305C. Application and washout of Cu:Phen in the presence of GABA caused a significant increase in the holding current and a reduction in the GABA test current. The increased holding current was inhibited by application of 50 μm picrotoxin. All GABA current traces represent stable currents: two subsequent GABA currents were not different by >10%. Only one is shown for the sake of clarity. Currents during reagent application are not shown. The bars above the traces indicate the duration of application of GABA; the arrows indicate the application of the given reagent.

Effect of DTT and Cu:Phen on α1T267C-V297C

We sought to test the extent of vertical motion of M3 relative to the 12′ M2 position. V297 lies on the same face of the M3 helix as the reactive residues but one helical turn closer to the extracellular side. Application of 10 mm DTT for 2 min or 100:200 μm Cu:Phen for 2 min had no effect on the GABA-induced currents of the double mutant α1T267C-V297C. We infer that a disulfide bond does not form between these two residues, implying that there is limited vertical motion of the M3 segment relative to the M2 segment.

Cu:Phen effect on the 11′ mutants α1M266C-L301C, α1M266C-F304C, and α1M266C-A305C in the absence and presence of GABA

As a control for possible reaction of the M3 Cys with endogenous cysteines, we tested the effects of DTT and Cu:Phen on the redox-sensitive M3 Cys mutants in the presence of the 11′ Cys mutant α1M266C. A 2 min application of 10 mm DTT had no significant effects on the subsequent GABA-induced currents in oocytes expressing the mutants α1M266C, α1M266C-L301C, α1M266C-F304C, or α1M266C-A305C compared with wild type (one-way ANOVA with Dunnett's multiple comparison test; p > 0.05) (data not shown). We infer that no disulfide bonds formed spontaneously between these pairs of Cys mutants or between these Cys mutants and the endogenous cysteines.

We also tested the effect of oxidation on these mutants with a 2 min application of 100:200 μm Cu:Phen to oocytes expressing these mutants in the absence and presence of GABA. There was no significant effect on the subsequent GABA-induced currents compared with wild type (one-way ANOVA with Dunnett's multiple comparison test; p > 0.05) (data not shown). We infer that disulfide bonds were not formed between the M2–11′ Cys and these M3 residues. Furthermore, we infer that these M3 Cys mutants did not form disulfide bonds with endogenous cysteine residues. Thus, we conclude that the effects described in the previous sections with the double Cys mutants containing the M2–12′ Cys arose because of disulfide bond formation between the M2–12′ Cys and the corresponding M3 Cys.

Homology modeling

The first step in constructing a homology model is to align the respective sequences. As noted in Introduction, this is easy for ACh and GABAAM1 and M2 segments because of the presence of a series of absolutely conserved residues (Fig. 1). Alignment of the M3 segments is more difficult. Depending on the number of sequences analyzed by multiple sequence alignment programs, either no gap or a one residue gap is created in the M2–M3 loop of the GABAA receptor sequences compared with the ACh receptor (AChR) sequences. Having identified the GABAA α1M3 residues that are in close proximity to α1M2–12′ as L301, F304, and A305, we can conclude that these residues should align with AChR residues in close proximity to αM2–12′ in the AChR structure (Unwin, 2005). These residues are ACh M3 αI289, αT292, and αV293. These residues align with the GABAA receptor M3 residues that we identified if one assumes a one residue gap in the M2–M3 loop as shown in Figure 1.

We constructed homology models (see Materials and Methods) with no or one residue gaps in the M2–M3 loop (Fig. 6, compare A, B). The channel-lining M2 residues are in good agreement with our previous substituted cysteine accessibility results (Xu and Akabas, 1993, 1996). The lack of the 1 aa gap in the M2–M3 loop introduces a 100° rotation of the M3 segment residues. Consequently, the face of M3 that apposes M2 is made up of different amino acids in both models. Based on our cross-linking results, we infer that the alignment with the 1 aa gap in the M2–M3 loop must be used to thread the GABAA receptor α1 subunit onto the AChR structure. This suggests that the M2–M3 loop is one residue shorter in the GABAA receptor compared with the ACh and 5-HT3 receptor subunits. The functional consequences of this remain to be investigated.

Figure 6.
Figure 6.

Homology model. A homology model of the GABAA receptor α1 subunit was generated using the nicotinic ACh receptor structure (PDB number 2BG9) as a template. The M2 and M3 helices and the connecting M2–M3 loop are shown in wireframe representation (green). Positions at which disulfide bonds formed are shown in CPK colors. In A–F, two views of the M2 and M3 segments of wt and mutant receptors are shown from the side (top) and top (bottom). A shows a model with a 1 aa gap in the M2–M3 loop; note the proximity of the M3 positions that formed disulfide bonds with the M2–12′ position. M2 is on the right. B shows the same residues in a model with no gap in the M2–M3 loop. Note that the residues that formed disulfide bonds are not in close proximity in this model. C–G were generated using the model containing a single gap in the M2–M3 loop as shown in A. C–E illustrate the proximity of the pairs of residues that formed disulfide bonds. van der Waals surfaces of Cys substituted at these positions are shown. F, Mutations at M2–15′ S270 and M3 A291 alter the action of general anesthetics. van der Waals surfaces of Cys substituted at these positions are shown. The α carbons are separated by ∼15 Å, making it unlikely that they are part of a single binding site. G shows the α-carbon separation of the M2–12′ residue and several M3 residues.

Discussion

These results have important implications for Cys-loop receptor structure–function. First, the closed state disulfide bonds provide an experimental basis for alignment of GABAA and ACh M3 segments, an essential step to constructing realistic GABAA homology models based on the 4 Å AChR structure. This will provide a molecular framework for interpreting the effects of GABAA mutations on general anesthetic efficacy. Second, disulfide bond formation in the presence of GABA that increases spontaneous channel opening may provide insight into the conformational changes that M2 channel-lining segments undergo during channel opening in GABAA and other Cys-loop receptors.

Homology modeling offers the possibility of applying to GABAA receptors the recent advances in our understanding of ACh receptor structure. It is, however, essential to align correctly the ACh and GABAA sequences. In the transmembrane domain, absolutely conserved residues in and flanking M1 and M2 facilitate correct alignment but depending on the number of sequences aligned, multiple sequence alignment programs introduce a gap in the M2–M3 loop (Fig. 1). This, coupled with the low sequence identity in M3, makes it difficult to align the ACh and GABAA M3 segments. We used an unbiased search to determine the proximity between M2 and M3 residues. In the closed state, disulfide bonds formed spontaneously between M2–12′, α1T267C, and both M3 α1L301C and α1F304C and could be induced with α1A305C. These three residues lie on one face of the M3 helix. In the closed state, this face of M3 lies in close proximity to M2 (Fig. 6). When the M2 Cys was moved to the adjacent position α1M266C (11′), which should face M1, oxidation had no effect on any of these M3-Cys mutants, confirming the specificity of the reaction between M2–12′Cys and these M3 cysteines. In our homology model, based on the AChR structure (Unwin, 2005), the M2–12′ α carbon is within 10 Å of the α carbons of these three M3 residues (Fig. 6). In a disulfide bond, the α carbons can be separated by at most 5.6 Å (Careaga and Falke, 1992). This implies that 4–5 Å relative movements of M2 and M3 would bring these positions into sufficiently close proximity to permit disulfide bond formation. Previously, we showed that M2 undergoes even larger translational movements at the 20′ level (Horenstein et al., 2005) that are facilitated by the loose protein packing around the extracellular one-half of M2 (Goren et al., 2004; Unwin, 2005). Because disulfide bonds formed spontaneously with L301C and F304C, we infer that the collision frequency of M2–12′ with these positions was higher and/or the orientation was more favorable than with A305C (Fig. 6C–E,G). In the Torpedo AChR structure, the α subunit M3 residues in close proximity to M2–12′ are I289, T292, and V293 (Unwin, 2005). To align these ACh and GABAA residues, a single gap must be introduced into the M2–M3 loop (Figs. 1, 6). Thus, despite the cogent rationale offered recently for a two residue gap in the M2–M3 loop (Ernst et al., 2005), it is not supported by our current experiments. Furthermore, our results imply that the relative depths of the M2 and M3 segments in the 2BG9 AChR structure are correct (Unwin, 2005).

Our results place limits on the extent of thermal motion occurring in this region of M3. The absence of disulfide bond formation with residues on the backside of M3 implies that M3 does not undergo significant rotational motion in the closed state. The absence of disulfide bond formation between M2–12′C and M3-V297C, the residue on the reactive face one helical turn more extracellular than L301, implies that in the closed state the upward displacement (i.e., perpendicular to the membrane plane) of M2 relative to M3 is less than ∼5 Å. The α-carbon separation between M2–12′ and V297 is 10 Å (Fig. 6G). Furthermore, α1Cys293, an endogenous cysteine on the same helical face one turn above V297 (Fig. 6G), did not form a disulfide bond with the M2–12′ Cys because neither DTT nor Cu:Phen had any effect on the single M2–12′Cys mutant. These results, combined with the lack of reaction of sulfhydryl-reactive reagents with M3 engineered Cys residues below α1Y294C in the closed state (Williams and Akabas, 1999) suggest that M3 has limited thermal motion and water surface accessibility in the closed state. In contrast, M2 undergoes significant thermal motion in the closed state, particularly above 12′ (Horenstein et al., 2001; Bera and Akabas, 2005; Horenstein et al., 2005). These results also place limits on the potential rotational movement of M2, because no disulfide bonds formed between the M2–11′ position and any of the M3 Cys mutants.

GABA activation altered the orientation and/or relative position of M2 and M3. Previously, we showed that GABA activation increased the water accessibility of several M3 positions including L301, suggesting that M3 and/or its surroundings underwent a conformational change during gating (Williams and Akabas, 1999). With GABA, disulfide bonds did not form or formed at a significantly slower rate between M2–12′C and either L301C or F304C (Fig. 5A,B). In contrast, an α1T267C-A305C disulfide bond formed in the presence and absence of GABA, but the functional effect was quite different: When induced in the absence of GABA, the subsequent GABA currents were inhibited, whereas in the presence of GABA, it caused a significant increase in the spontaneous open probability that was inhibited by picrotoxin and reversed by DTT (Fig. 5C,D). We infer that, in the closed and activated states, disulfide bond formation trapped the receptor in two distinct conformations that were not interchangeable after bond formation. This conformational change could involve an ∼5 Å translational movement of M2 toward the M1 segment in the counterclockwise adjacent subunit (looking from the extracellular side).

How could an α1T267C-A305C disulfide bond increase spontaneous opening? In the closed state, the M2 segments probably fluctuate continuously between their closed and open conformations, although predominantly in the closed conformation. Presumably, spontaneous opening and ion conduction only occurs when all five M2 segments are in the open conformation, which rarely occurs in the absence of GABA. We infer that the α1T267C-A305C disulfide bond formed in the presence of GABA holds the two α1M2 segments in a conformation similar to their open conformation. Thus, after disulfide bond formation, spontaneous opening would only require the remaining three M2 segments to simultaneously enter the open conformation. This would significantly increase the spontaneous open probability. Alternatively, the conformation with the α subunit disulfide bonds might induce a conformational change in the M2 segments of the neighboring subunits. We cannot distinguish between these possibilities, although we favor the former. Previously, we found that spontaneous channel opening increased when a disulfide bond formed in the presence of GABA between M2–6′ Cys in adjacent subunits (Horenstein et al., 2001). Thus, there are multiple positions where disulfide bond formation in the GABA-activated state leads to increased spontaneous channel opening. These may provide insight into the conformational changes that occur during channel activation.

GABAA M2–15′ mutations alter general anesthetic potentiation in both heterologous expression systems and knock-in mice (Belelli et al., 1997; Mihic et al., 1997; Jurd et al., 2003). Mutations of residues aligned with M3-α1A291 also alter general anesthetic potentiation (Mihic et al., 1997; Krasowski et al., 2001; Bali and Akabas, 2004). These two positions were widely assumed to be part of a single anesthetic binding site; however, the intravenous anesthetic propofol only protected a Cys at the β2M3 site from modification by sulfhydryl reagent, suggesting that they were separate sites (Bali and Akabas, 2004). Using unrelated four-helix-bundle proteins as templates, models were constructed that placed these two positions in close proximity (Trudell and Bertaccini, 2002). Other investigators, using the effects of Hg2+ on Cys mutants substituted at the aligned positions in glycine receptors, concluded that a single Hg2+ ion could bind simultaneously to engineered Cys at both positions (Lobo et al., 2004). In our homology model, based on the 2BG9 coordinates and our current results, the α carbons of α1M2–15′ and M3-α1A291 are separated by 15.6 Å (Fig. 6F). This suggests that they are not part of a single binding site. The lack of rotation of M3 and the limited vertical movement make it unlikely that thermal motion could bring these positions into close proximity. Given their separation, we doubt that a single Hg2+ could simultaneously bind to Cys at both sites (Lobo et al., 2004). More likely, each Cys bound separate Hg2+ ions leading to the reported effects.

Our results provide experimental evidence for the correct threading of the GABAA M3 segment onto the AChR structure. They support the ClustalW sequence alignment with a one residue gap in the GABAA and glycine M2–M3 loops compared with the ACh and 5-HT3 receptors. How this alters interactions between this loop and the extracellular domain that are critical for channel gating (Kash et al., 2003; Lummis et al., 2005; Reeves et al., 2005) remain to be elucidated. The development of more realistic GABAA receptor homology models will lead to a better understanding of general anesthetic binding and of the conformational changes induced by agonist activation and by anesthetics.

Footnotes

  • This work was supported in part by National Institutes of Health Grants NS30808 and GM77660 (M.H.A.). M.J. was supported in part by a fellowship from the Deutsche Forschungsgemeinschaft. We thank Moez Bali, David Liebelt, Nicole McKinnon, David Reeves, and Paul Riegelhaupt for helpful discussions and also for comments on this manuscript.

  • Correspondence should be addressed to Dr. Myles H. Akabas, Department of Physiology and Biophysics, Albert Einstein College of Medicine, Bronx, NY 10461. Email: makabas{at}aecom.yu.edu

References

  1. ↵
    Bali M, Akabas MH (2004) Defining the propofol binding site location on the GABAA receptor. Mol Pharmacol 65:68–76.
    OpenUrlAbstract/FREE Full Text
  2. ↵
    Belelli D, Lambert JJ, Peters JA, Wafford K, Whiting PJ (1997) The interaction of the general anesthetic etomidate with the gamma-aminobutyric acid type A receptor is influenced by a single amino acid. Proc Natl Acad Sci USA 94:11031–11036.
    OpenUrlAbstract/FREE Full Text
  3. ↵
    Bera AK, Akabas MH (2005) Spontaneous thermal motion of the GABA(A) receptor M2 channel-lining segments. J Biol Chem 280:35506–35512.
    OpenUrlAbstract/FREE Full Text
  4. ↵
    Brejc K, van Dijk WJ, Klaassen RV, Schuurmans M, van Der Oost J, Smit AB, Sixma TK (2001) Crystal structure of an ACh-binding protein reveals the ligand-binding domain of nicotinic receptors. Nature 411:269–276.
    OpenUrlCrossRefPubMed
  5. ↵
    Careaga CL, Falke JJ (1992) Thermal motions of surface alpha-helices in the d-galactose chemosensory receptor. Detection by disulfide trapping. J Mol Biol 226:1219–1235.
    OpenUrl
  6. ↵
    Celie PH, van Rossum-Fikkert SE, van Dijk WJ, Brejc K, Smit AB, Sixma TK (2004) Nicotine and carbamylcholine binding to nicotinic acetylcholine receptors as studied in AChBP crystal structures. Neuron 41:907–914.
    OpenUrlCrossRefPubMed
  7. ↵
    Cromer BA, Morton CJ, Parker MW (2002) Anxiety over GABA(A) receptor structure relieved by AChBP. Trends Biochem Sci 27:280–287.
    OpenUrlCrossRefPubMed
  8. ↵
    Ernst M, Bruckner S, Boresch S, Sieghart W (2005) Comparative models of GABAA receptor extracellular and transmembrane domains: important insights in pharmacology and function. Mol Pharmacol 68:1291–1300.
    OpenUrlAbstract/FREE Full Text
  9. ↵
    Goren EN, Reeves DC, Akabas MH (2004) Loose protein packing around the extracellular half of the GABAA receptor β1 subunit M2 channel-lining segment. J Biol Chem 279:11198–11205.
    OpenUrlAbstract/FREE Full Text
  10. ↵
    Guex N, Peitsch MC (1997) SWISS-MODEL and the Swiss-PdbViewer: an environment for comparative protein modeling. Electrophoresis 18:2714–2723.
    OpenUrlCrossRefPubMed
  11. ↵
    Hevers W, Lüddens H (1998) The diversity of GABAA receptors. Pharmacological and electrophysiological properties of GABAA channel subtypes. Mol Neurobiol 18:35–86.
    OpenUrlCrossRefPubMed
  12. ↵
    Horenstein J, Akabas MH (1998) Location of a high affinity Zn2+ binding site in the channel of alpha1beta1 gamma-aminobutyric acid A receptors. Mol Pharmacol 53:870–877.
    OpenUrlAbstract/FREE Full Text
  13. ↵
    Horenstein J, Wagner DA, Czajkowski C, Akabas MH (2001) Protein mobility and GABA-induced conformational changes in GABAA receptor pore-lining M2 segment. Nat Neurosci 4:477–485.
    OpenUrlPubMed
  14. ↵
    Horenstein J, Riegelhaupt P, Akabas MH (2005) Differential protein mobility of the gamma-aminobutyric acid, type A, receptor alpha and beta subunit channel-lining segments. J Biol Chem 280:1573–1581.
    OpenUrlAbstract/FREE Full Text
  15. ↵
    Imoto K, Busch C, Sakmann B, Mishina M, Konno T, Nakai J, Bujo H, Mori Y, Fukuda K, Numa S (1988) Rings of negatively charged amino acids determine the acetylcholine receptor channel conductance. Nature 335:645–648.
    OpenUrlCrossRefPubMed
  16. ↵
    Jurd R, Arras M, Lambert S, Drexler B, Siegwart R, Crestani F, Zaugg M, Vogt KE, Ledermann B, Antkowiak B, Rudolph U (2003) General anesthetic actions in vivo strongly attenuated by a point mutation in the GABA(A) receptor beta3 subunit. FASEB J 17:250–252.
    OpenUrlAbstract/FREE Full Text
  17. ↵
    Karlin A (2002) Emerging structure of the nicotinic acetylcholine receptors. Nat Rev Neurosci 3:102–114.
    OpenUrlCrossRefPubMed
  18. ↵
    Kash TL, Jenkins A, Kelley JC, Trudell JR, Harrison NL (2003) Coupling of agonist binding to channel gating in the GABA(A) receptor. Nature 421:272–275.
    OpenUrlCrossRefPubMed
  19. ↵
    Kobashi K (1968) Catalytic oxidation of sulfhydryl groups by o-phenanthroline copper complex. Biochim Biophys Acta 158:239–245.
    OpenUrlPubMed
  20. ↵
    Krasowski MD, Nishikawa K, Nikolaeva N, Lin A, Harrison NL (2001) Methionine 286 in transmembrane domain 3 of the GABAA receptor beta subunit controls a binding cavity for propofol and other alkylphenol general anesthetics. Neuropharmacology 41:952–964.
    OpenUrlCrossRefPubMed
  21. ↵
    Krezel A, Latajka R, Bujacz GD, Bal W (2003) Coordination properties of tris(2-carboxyethyl)phosphine, a newly introduced thiol reductant, and its oxide. Inorg Chem 42:1994–2003.
    OpenUrlCrossRefPubMed
  22. ↵
    Le Novere N, Changeux JP (1999) The ligand gated ion channel database. Nucleic Acids Res 27:340–342.
    OpenUrlAbstract/FREE Full Text
  23. ↵
    Lester HA, Dibas MI, Dahan DS, Leite JF, Dougherty DA (2004) Cys-loop receptors: new twists and turns. Trends Neurosci 27:329–336.
    OpenUrlCrossRefPubMed
  24. ↵
    Lobo IA, Trudell JR, Harris RA (2004) Cross-linking of glycine receptor transmembrane segments two and three alters coupling of ligand binding with channel opening. J Neurochem 90:962–969.
    OpenUrlCrossRefPubMed
  25. ↵
    Lummis SC, Beene DL, Lee LW, Lester HA, Broadhurst RW, Dougherty DA (2005) Cis-trans isomerization at a proline opens the pore of a neurotransmitter-gated ion channel. Nature 438:248–252.
    OpenUrlCrossRefPubMed
  26. ↵
    Mihic SJ, Ye Q, Wick MJ, Koltchine VV, Krasowski MD, Finn SE, Mascia MP, Valenzuela CF, Hanson KK, Greenblatt EP, Harris RA, Harrison NL (1997) Sites of alcohol and volatile anaesthetic action on GABA(A) and glycine receptors. Nature 389:385–389.
    OpenUrlCrossRefPubMed
  27. ↵
    Miller C (1989) Genetic manipulation of ion channels: a new approach to structure and mechanism. Neuron 2:1195–1205.
    OpenUrlCrossRefPubMed
  28. ↵
    Miyazawa A, Fujiyoshi Y, Unwin N (2003) Structure and gating mechanism of the acetylcholine receptor pore. Nature 423:949–955.
    OpenUrlCrossRefPubMed
  29. ↵
    Reeves DC, Jansen M, Bali M, Lemster T, Akabas MH (2005) A role for the β1–β2 loop in the gating of 5-HT3 receptors. J Neurosci 25:9358–9366.
    OpenUrlAbstract/FREE Full Text
  30. ↵
    Schwede T, Kopp J, Guex N, Peitsch MC (2003) SWISS-MODEL: an automated protein homology-modeling server. Nucleic Acids Res 31:3381–3385.
    OpenUrlAbstract/FREE Full Text
  31. ↵
    Trudell JR, Bertaccini E (2002) Molecular modelling of specific and non-specific anaesthetic interactions. Br J Anaesth 89:32–40.
    OpenUrlAbstract/FREE Full Text
  32. ↵
    Trudell JR, Bertaccini E (2004) Comparative modeling of a GABAA alpha1 receptor using three crystal structures as templates. J Mol Graph Model 23:39–49.
    OpenUrlCrossRefPubMed
  33. ↵
    Unwin N (2005) Refined structure of the nicotinic acetylcholine receptor at 4Å resolution. J Mol Biol 346:967–989.
    OpenUrlCrossRefPubMed
  34. ↵
    Williams DB, Akabas MH (1999) Gamma-aminobutyric acid increases the water accessibility of M3 membrane-spanning segment residues in gamma-aminobutyric acid type A receptors. Biophys J 77:2563–2574.
    OpenUrlPubMed
  35. ↵
    Xu M, Akabas MH (1993) Amino acids lining the channel of the gamma-aminobutyric acid type A receptor identified by cysteine substitution. J Biol Chem 268:21505–21508.
    OpenUrlAbstract/FREE Full Text
  36. ↵
    Xu M, Akabas MH (1996) Identification of channel-lining residues in the M2 membrane-spanning segment of the GABAA receptor alpha1 subunit. J Gen Physiol 107:195–205.
    OpenUrlAbstract/FREE Full Text
Back to top

In this issue

Email
Print
View Full Page PDF
Citation Tools
Respond to this article
State-Dependent Cross-Linking of the M2 and M3 Segments: Functional Basis for the Alignment of GABAA and Acetylcholine Receptor M3 Segments
Michaela Jansen, Myles H. Akabas
Journal of Neuroscience 26 April 2006, 26 (17) 4492-4499; DOI: 10.1523/JNEUROSCI.0224-06.2006
Twitter logo Facebook logo Mendeley logo

Jump to section

Responses to this article

Respond to this article

Jump to comment:

No eLetters have been published for this article.