Abstracts

Rationale: GABA-A receptors mediate CNS inhibition and are composed of 5 subunits arranged around a chloride channel. The mutation γ2R43Q at the γ2-β2 inter-subunit interface is associated with absence epilepsy (Wallace, Nat Genet 28:49), and has been shown to disrupt receptor trafficking (Kang, J Neurosci 24:8672; Sancar, JBC 279:47034; Hales, Mol Cell Neurosci 29:120), to alter synaptic and tonic inhibition (Tan, PNAS 104:17536; Eugène, J Neurosci 27:14108), and may participate in a salt bridge network with γ2E178 and β2R117 (Cromer, TIBS 27:280). However, there is disagreement over whether the mutation alters receptor kinetics or benzodiazepine pharmacology (Bianchi, J Neurosci 22:5321; Bowser, PNAS 99:15170). To clarify this issue, we mutated the putative salt bridge residues and measured the kinetics of GABA responses and of diazepam (DZ) dissociation. This latter measure arises solely from DZ-sensitive receptors, which must contain a γ2 subunit (mutant or otherwise), and thus cannot be explained by deficits in trafficking. Methods: HEK-293 cells were transfected with 1-4 μg total cDNA of wild type or mutant α1, β2 and γ2 subunits in the ratio 1:1:1. Receptor kinetics were examined in outside-out patches (voltage clamped at -60 mV, room temp.) using rapid (<200 μs) solution exchange techniques (Jones, J Neurosci 18:8590). DZ unbinding was assayed by pre-equilibrating in 10 μM DZ, then rapidly switching to wash solution and evaluating the decline of response potentiation at varying intervals, using a 20-30 ms pulse of subsaturating (30 μM) GABA. Results: The wild type α1β2γ2 receptor had a weighted deactivation time constant of 45 ± 4 ms (n = 24) and weighted desensitization time constant of 142 ± 10 ms (n = 27). Mutations that altered the size or charge of the putative salt bridge residues significantly prolonged GABA-A receptor deactivation by about two-fold (β2R117K), sped desensitization by almost two-fold (γ2R43K, γ2E178K) or had both effects (γ2R43Q, γ2E178Q) (n = 5-20; p < 0.05 by ANOVA; see Fig. 1 for deactivation). Fitting of traces to a detailed kinetic model suggests that these changes were due to slowed GABA unbinding and slowed recovery from desensitization. In DZ unbinding experiments, plots of percent potentiation vs. unbinding interval (Fig. 2) had a small, fast rising phase probably reflecting the relief of block by the high DZ concentration (τ-rise: WT = 200 ± 120 ms, n = 8; R43Q = 470 ± 150 ms, n = 3), and a large, slow decay phase reflecting the timecourse of DZ dissociation (τ-decay: WT = 2.9 ± 0.4 s; R43Q = 1.3 ± 0.7 s; R117K = 1.1 s, n = 1). Conclusions: Several mutations at the γ2/β2 interface slow GABA unbinding, which occurs at the β2/α1 interface. Furthermore, γ2R43Q and β2R117K speed DZ unbinding, which occurs at the α1/γ2 interface. Therefore the γ2/β2 interface, disrupted in heriditary absence epilepsy, appears to communicate allosterically with distant interfaces to modulate both GABA and benzodiazepine interactions with the receptor.