Advertisement

Functional genomics of epilepsy-associated mutations in the GABAA receptor subunits reveal that one mutation impairs function and two are catastrophic

  • Nathan L. Absalom
    Affiliations
    From the Brain and Mind Centre, University of Sydney, 94 Mallett Street, Camperdown, New South Wales 2050, Australia

    School of Pharmacy, University of Sydney, Camperdown, New South Wales 2006, Australia
    Search for articles by this author
  • Philip K. Ahring
    Affiliations
    From the Brain and Mind Centre, University of Sydney, 94 Mallett Street, Camperdown, New South Wales 2050, Australia

    School of Pharmacy, University of Sydney, Camperdown, New South Wales 2006, Australia
    Search for articles by this author
  • Vivian W. Liao
    Affiliations
    From the Brain and Mind Centre, University of Sydney, 94 Mallett Street, Camperdown, New South Wales 2050, Australia

    School of Pharmacy, University of Sydney, Camperdown, New South Wales 2006, Australia
    Search for articles by this author
  • Thomas Balle
    Affiliations
    From the Brain and Mind Centre, University of Sydney, 94 Mallett Street, Camperdown, New South Wales 2050, Australia

    School of Pharmacy, University of Sydney, Camperdown, New South Wales 2006, Australia
    Search for articles by this author
  • Tian Jiang
    Affiliations
    From the Brain and Mind Centre, University of Sydney, 94 Mallett Street, Camperdown, New South Wales 2050, Australia

    School of Pharmacy, University of Sydney, Camperdown, New South Wales 2006, Australia
    Search for articles by this author
  • Lyndsey L. Anderson
    Affiliations
    From the Brain and Mind Centre, University of Sydney, 94 Mallett Street, Camperdown, New South Wales 2050, Australia

    Discipline of Pharmacology, Faculty of Medicine and Health, University of Sydney, Camperdown, New South Wales 2006, Australia

    Lambert Initiative for Cannabinoid Therapeutics, University of Sydney, 94 Mallett Street, Camperdown, New South Wales 2050, Australia
    Search for articles by this author
  • Jonathon C. Arnold
    Affiliations
    From the Brain and Mind Centre, University of Sydney, 94 Mallett Street, Camperdown, New South Wales 2050, Australia

    Discipline of Pharmacology, Faculty of Medicine and Health, University of Sydney, Camperdown, New South Wales 2006, Australia

    Lambert Initiative for Cannabinoid Therapeutics, University of Sydney, 94 Mallett Street, Camperdown, New South Wales 2050, Australia
    Search for articles by this author
  • Iain S. McGregor
    Affiliations
    From the Brain and Mind Centre, University of Sydney, 94 Mallett Street, Camperdown, New South Wales 2050, Australia

    Lambert Initiative for Cannabinoid Therapeutics, University of Sydney, 94 Mallett Street, Camperdown, New South Wales 2050, Australia

    the School of Psychology, Faculty of Science, University of Sydney, Camperdown, New South Wales 2006, Australia
    Search for articles by this author
  • Michael T. Bowen
    Affiliations
    From the Brain and Mind Centre, University of Sydney, 94 Mallett Street, Camperdown, New South Wales 2050, Australia

    Lambert Initiative for Cannabinoid Therapeutics, University of Sydney, 94 Mallett Street, Camperdown, New South Wales 2050, Australia

    the School of Psychology, Faculty of Science, University of Sydney, Camperdown, New South Wales 2006, Australia
    Search for articles by this author
  • Mary Chebib
    Correspondence
    To whom correspondence should be addressed: Brain and Mind Centre, University of Sydney, 94 Mallett St., Camperdown, New South Wales 2050, Australia. Tel.: 612-9351-8584; Fax: 612-9351-4391
    Affiliations
    From the Brain and Mind Centre, University of Sydney, 94 Mallett Street, Camperdown, New South Wales 2050, Australia

    School of Pharmacy, University of Sydney, Camperdown, New South Wales 2006, Australia
    Search for articles by this author
  • Author Footnotes
    2 The abbreviations used are: GABAAγ-aminobutyric acid type AcRNAcomplementary RNAEst. Po(max)estimated maximum open probabilitycryo-EMcryogenic EMANOVAanalysis of variancePDBProtein Data Bank.
Open AccessPublished:February 06, 2019DOI:https://doi.org/10.1074/jbc.RA118.005697
      A number of epilepsy-causing mutations have recently been identified in the genes of the α1, β3, and γ2 subunits comprising the γ-aminobutyric acid type A (GABAA) receptor. These mutations are typically dominant, and in certain cases, such as the α1 and β3 subunits, they may lead to a mix of receptors at the cell surface that contain no mutant subunits, a single mutated subunit, or two mutated subunits. To determine the effects of mutations in a single subunit or in two subunits on receptor activation, we created a concatenated protein assembly that links all five subunits of the α1β3γ2 receptor and expresses them in the correct orientation. We created nine separate receptor variants with a single-mutant subunit and four receptors containing two subunits of the γ2R323Q, β3D120N, β3T157M, β3Y302C, and β3S254F epilepsy-causing mutations. We found that the singly mutated γ2R323Q subunit impairs GABA activation of the receptor by reducing GABA potency. A single β3D120N, β3T157M, or β3Y302C mutation also substantially impaired receptor activation, and two copies of these mutants within a receptor were catastrophic. Of note, an effect of the β3S254F mutation on GABA potency depended on the location of this mutant subunit within the receptor, possibly because of the membrane environment surrounding the transmembrane region of the receptor. Our results highlight that precise functional genomic analyses of GABAA receptor mutations using concatenated constructs can identify receptors with an intermediate phenotype that contribute to epileptic phenotypes and that are potential drug targets for precision medicine approaches.

      Introduction

      Epileptic encephalopathies are a devastating group of severe childhood epilepsies with poor developmental outcomes that are often resistant to pharmacological treatment (
      • Howard M.A.
      • Baraban S.C.
      Catastrophic epilepsies of childhood.
      ). In many cases, the causes are genetic, and recent advances in whole-genome sequencing have identified a series of de novo and inherited mutations in various genes. Several mutations in genes that encode for the α1 (
      • Cossette P.
      • Liu L.
      • Brisebois K.
      • Dong H.
      • Lortie A.
      • Vanasse M.
      • Saint-Hilaire J.M.
      • Carmant L.
      • Verner A.
      • Lu W.Y.
      • Wang Y.T.
      • Rouleau G.A.
      Mutation of GABRA1 in an autosomal dominant form of juvenile myoclonic epilepsy.
      • Gontika M.P.
      • Konialis C.
      • Pangalos C.
      • Papavasiliou A.
      Novel SCN1A and GABRA1 gene mutations with diverse phenotypic features and the question on the existence of a broader spectrum of Dravet syndrome.
      ,
      • Hernandez C.C.
      • Klassen T.L.
      • Jackson L.G.
      • Gurba K.
      • Hu N.
      • Noebels J.L.
      • Macdonald R.L.
      Deleterious rare variants reveal risk for loss of GABAA receptor function in patients with genetic epilepsy and in the general population.
      ,
      • Johannesen K.
      • Marini C.
      • Pfeffer S.
      • Møller R.S.
      • Dorn T.
      • Niturad C.E.
      • Gardella E.
      • Weber Y.
      • Søndergård M.
      • Hjalgrim H.
      • Nikanorova M.
      • Becker F.
      • Larsen L.H.
      • Dahl H.A.
      • Maier O.
      • et al.
      Phenotypic spectrum of GABRA1: from generalized epilepsies to severe epileptic encephalopathies.
      ,
      • Kodera H.
      • Ohba C.
      • Kato M.
      • Maeda T.
      • Araki K.
      • Tajima D.
      • Matsuo M.
      • Hino-Fukuyo N.
      • Kohashi K.
      • Ishiyama A.
      • Takeshita S.
      • Motoi H.
      • Kitamura T.
      • Kikuchi A.
      • Tsurusaki Y.
      • et al.
      De novo GABRA1 mutations in Ohtahara and West syndromes.
      ,
      • Lachance-Touchette P.
      • Brown P.
      • Meloche C.
      • Kinirons P.
      • Lapointe L.
      • Lacasse H.
      • Lortie A.
      • Carmant L.
      • Bedford F.
      • Bowie D.
      • Cossette P.
      Novel α1 and γ2 GABAA receptor subunit mutations in families with idiopathic generalized epilepsy.
      ,
      • Møller R.S.
      • Larsen L.H.
      • Johannesen K.M.
      • Talvik I.
      • Talvik T.
      • Vaher U.
      • Miranda M.J.
      • Farooq M.
      • Nielsen J.E.
      • Svendsen L.L.
      • Kjelgaard D.B.
      • Linnet K.M.
      • Hao Q.
      • Uldall P.
      • Frangu M.
      • et al.
      Gene panel testing in epileptic encephalopathies and familial epilepsies.
      • Stosser M.B.
      • Lindy A.S.
      • Butler E.
      • Retterer K.
      • Piccirillo-Stosser C.M.
      • Richard G.
      • McKnight D.A.
      High frequency of mosaic pathogenic variants in genes causing epilepsy-related neurodevelopmental disorders.
      ), β3 (
      • Hernandez C.C.
      • Klassen T.L.
      • Jackson L.G.
      • Gurba K.
      • Hu N.
      • Noebels J.L.
      • Macdonald R.L.
      Deleterious rare variants reveal risk for loss of GABAA receptor function in patients with genetic epilepsy and in the general population.
      ,
      • Lachance-Touchette P.
      • Brown P.
      • Meloche C.
      • Kinirons P.
      • Lapointe L.
      • Lacasse H.
      • Lortie A.
      • Carmant L.
      • Bedford F.
      • Bowie D.
      • Cossette P.
      Novel α1 and γ2 GABAA receptor subunit mutations in families with idiopathic generalized epilepsy.
      ,
      • Møller R.S.
      • Larsen L.H.
      • Johannesen K.M.
      • Talvik I.
      • Talvik T.
      • Vaher U.
      • Miranda M.J.
      • Farooq M.
      • Nielsen J.E.
      • Svendsen L.L.
      • Kjelgaard D.B.
      • Linnet K.M.
      • Hao Q.
      • Uldall P.
      • Frangu M.
      • et al.
      Gene panel testing in epileptic encephalopathies and familial epilepsies.
      ,
      • Allen A.S.
      • Berkovic S.F.
      • Cossette P.
      • Delanty N.
      • Dlugos D.
      • Eichler E.E.
      • Epstein M.P.
      • Glauser T.
      • Goldstein D.B.
      • Han Y.
      • Heinzen E.L.
      • Hitomi Y.
      • Howell K.B.
      • et al.
      Epi4K Consortium, Epilepsy Phenome/Genome Project
      De novo mutations in epileptic encephalopathies.
      ,
      • Janve V.S.
      • Hernandez C.C.
      • Verdier K.M.
      • Hu N.
      • Macdonald R.L.
      Epileptic encephalopathy de novo GABRB mutations impair GABAA receptor function.
      • Papandreou A.
      • McTague A.
      • Trump N.
      • Ambegaonkar G.
      • Ngoh A.
      • Meyer E.
      • Scott R.H.
      • Kurian M.A.
      GABRB3 mutations: a new and emerging cause of early infantile epileptic encephalopathy.
      ), and γ2 subunits (
      • Hernandez C.C.
      • Klassen T.L.
      • Jackson L.G.
      • Gurba K.
      • Hu N.
      • Noebels J.L.
      • Macdonald R.L.
      Deleterious rare variants reveal risk for loss of GABAA receptor function in patients with genetic epilepsy and in the general population.
      ,
      • Lachance-Touchette P.
      • Brown P.
      • Meloche C.
      • Kinirons P.
      • Lapointe L.
      • Lacasse H.
      • Lortie A.
      • Carmant L.
      • Bedford F.
      • Bowie D.
      • Cossette P.
      Novel α1 and γ2 GABAA receptor subunit mutations in families with idiopathic generalized epilepsy.
      ,
      • Stosser M.B.
      • Lindy A.S.
      • Butler E.
      • Retterer K.
      • Piccirillo-Stosser C.M.
      • Richard G.
      • McKnight D.A.
      High frequency of mosaic pathogenic variants in genes causing epilepsy-related neurodevelopmental disorders.
      ,
      • Baulac S.
      • Huberfeld G.
      • Gourfinkel-An I.
      • Mitropoulou G.
      • Beranger A.
      • Prud'homme J.F.
      • Baulac M.
      • Brice A.
      • Bruzzone R.
      • LeGuern E.
      First genetic evidence of GABAA receptor dysfunction in epilepsy: a mutation in the γ2-subunit gene.
      • Boillot M.
      • Morin-Brureau M.
      • Picard F.
      • Weckhuysen S.
      • Lambrecq V.
      • Minetti C.
      • Striano P.
      • Zara F.
      • Iacomino M.
      • Ishida S.
      • An-Gourfinkel I.
      • Daniau M.
      • Hardies K.
      • Baulac M.
      • Dulac O.
      • et al.
      Novel GABRG2 mutations cause familial febrile seizures.
      ,
      • Shen D.
      • Hernandez C.C.
      • Shen W.
      • Hu N.
      • Poduri A.
      • Shiedley B.
      • Rotenberg A.
      • Datta A.N.
      • Leiz S.
      • Patzer S.
      • Boor R.
      • Ramsey K.
      • Goldberg E.
      • Helbig I.
      • Ortiz-Gonzalez X.R.
      • et al.
      De novo GABRG2 mutations associated with epileptic encephalopathies.
      • Wallace R.H.
      • Marini C.
      • Petrou S.
      • Harkin L.A.
      • Bowser D.N.
      • Panchal R.G.
      • Williams D.A.
      • Sutherland G.R.
      • Mulley J.C.
      • Scheffer I.E.
      • Berkovic S.F.
      Mutant GABAA receptor γ2-subunit in childhood absence epilepsy and febrile seizures.
      ) of the GABA type A (GABAA)
      The abbreviations used are: GABAA
      γ-aminobutyric acid type A
      cRNA
      complementary RNA
      Est. Po(max)
      estimated maximum open probability
      cryo-EM
      cryogenic EM
      ANOVA
      analysis of variance
      PDB
      Protein Data Bank.
      receptor (GABRA1, GABRB3, and GABRG2, respectively) have been identified that result in epileptic encephalopathies.
      GABAA receptors are essential mediators of neurotransmission in both the developing and adult brain (
      • Khazipov R.
      GABAergic synchronization in epilepsy.
      ). These receptors are ion channels composed of five subunits that arrange around a central ion pore (
      • Miller P.S.
      • Aricescu A.R.
      Crystal structure of a human GABAA receptor.
      ). When GABA is released from the synapse, it binds to these receptors anchored at the postsynaptic membrane to open an ion channel, allowing chloride ions to pass, hyperpolarizing and inhibiting the cell (
      • Olsen R.W.
      GABAA receptor: positive and negative allosteric modulators.
      ).
      Many genes encoding for different subunits of the GABAA receptor are present in the mammalian brain, including six α (α1–6), three β (β1–3), three γ (γ1–3), and a δ, Ε, and π, and the majority of receptors are thought to contain two α, two β, and one γ subunit where they are anchored at the synapse, responding to high concentrations of GABA (
      • Hevers W.
      • Lüddens H.
      The diversity of GABAA receptors: pharmacological and electrophysiological properties of GABAA channel subtypes.
      ). Other combinations of receptors, often containing δ subunits, are found extrasynaptically, where they respond to low concentrations of GABA or spillover from the synapse (
      • Farrant M.
      • Nusser Z.
      Variations on an inhibitory theme: phasic and tonic activation of GABAA receptors.
      ). Each individual subunit consists of a large extracellular domain, four transmembrane domains where the second (M2) lines the channel pore, two short and one large loop linking transmembrane domains, and a short C terminus. At synaptic receptors, GABA binds within the β–α interface located between adjacent extracellular domains to trigger an activation pathway through a series of conformational changes that ultimately open the channel pore. These conformational changes are transmitted through interactions at the coupling region, where loops in the extracellular domain in close proximity to the membrane interact with the pre-M1 and M2-M3 loops that connect transmembrane domains (
      • Bouzat C.
      New insights into the structural bases of activation of Cys-loop receptors.
      ). This results in tilt of the M2 domain to open the pore. Epilepsy-causing mutations identified in the α1, β3, and γ2 subunits are located at different regions throughout the protein, including amino acids throughout the activation pathway from the ligand-binding pocket and extracellular structural β-sheets, through to the coupling and transmembrane M1 and M2 regions.
      Mutations in the GABAA receptor that cause epilepsy typically impair this process, either through misfolding of protein to reduce the number of receptors at the cell surface or disturbing the ability of the receptor to open in response to GABA (
      • Møller R.S.
      • Larsen L.H.
      • Johannesen K.M.
      • Talvik I.
      • Talvik T.
      • Vaher U.
      • Miranda M.J.
      • Farooq M.
      • Nielsen J.E.
      • Svendsen L.L.
      • Kjelgaard D.B.
      • Linnet K.M.
      • Hao Q.
      • Uldall P.
      • Frangu M.
      • et al.
      Gene panel testing in epileptic encephalopathies and familial epilepsies.
      ). In all cases, the mutations that cause epileptic encephalopathies are dominant, with patients carrying one copy of the WT allele and one copy of the mutant allele (
      • Cossette P.
      • Liu L.
      • Brisebois K.
      • Dong H.
      • Lortie A.
      • Vanasse M.
      • Saint-Hilaire J.M.
      • Carmant L.
      • Verner A.
      • Lu W.Y.
      • Wang Y.T.
      • Rouleau G.A.
      Mutation of GABRA1 in an autosomal dominant form of juvenile myoclonic epilepsy.
      • Gontika M.P.
      • Konialis C.
      • Pangalos C.
      • Papavasiliou A.
      Novel SCN1A and GABRA1 gene mutations with diverse phenotypic features and the question on the existence of a broader spectrum of Dravet syndrome.
      ,
      • Hernandez C.C.
      • Klassen T.L.
      • Jackson L.G.
      • Gurba K.
      • Hu N.
      • Noebels J.L.
      • Macdonald R.L.
      Deleterious rare variants reveal risk for loss of GABAA receptor function in patients with genetic epilepsy and in the general population.
      ,
      • Johannesen K.
      • Marini C.
      • Pfeffer S.
      • Møller R.S.
      • Dorn T.
      • Niturad C.E.
      • Gardella E.
      • Weber Y.
      • Søndergård M.
      • Hjalgrim H.
      • Nikanorova M.
      • Becker F.
      • Larsen L.H.
      • Dahl H.A.
      • Maier O.
      • et al.
      Phenotypic spectrum of GABRA1: from generalized epilepsies to severe epileptic encephalopathies.
      ,
      • Kodera H.
      • Ohba C.
      • Kato M.
      • Maeda T.
      • Araki K.
      • Tajima D.
      • Matsuo M.
      • Hino-Fukuyo N.
      • Kohashi K.
      • Ishiyama A.
      • Takeshita S.
      • Motoi H.
      • Kitamura T.
      • Kikuchi A.
      • Tsurusaki Y.
      • et al.
      De novo GABRA1 mutations in Ohtahara and West syndromes.
      ,
      • Lachance-Touchette P.
      • Brown P.
      • Meloche C.
      • Kinirons P.
      • Lapointe L.
      • Lacasse H.
      • Lortie A.
      • Carmant L.
      • Bedford F.
      • Bowie D.
      • Cossette P.
      Novel α1 and γ2 GABAA receptor subunit mutations in families with idiopathic generalized epilepsy.
      ,
      • Møller R.S.
      • Larsen L.H.
      • Johannesen K.M.
      • Talvik I.
      • Talvik T.
      • Vaher U.
      • Miranda M.J.
      • Farooq M.
      • Nielsen J.E.
      • Svendsen L.L.
      • Kjelgaard D.B.
      • Linnet K.M.
      • Hao Q.
      • Uldall P.
      • Frangu M.
      • et al.
      Gene panel testing in epileptic encephalopathies and familial epilepsies.
      ,
      • Stosser M.B.
      • Lindy A.S.
      • Butler E.
      • Retterer K.
      • Piccirillo-Stosser C.M.
      • Richard G.
      • McKnight D.A.
      High frequency of mosaic pathogenic variants in genes causing epilepsy-related neurodevelopmental disorders.
      ,
      • Allen A.S.
      • Berkovic S.F.
      • Cossette P.
      • Delanty N.
      • Dlugos D.
      • Eichler E.E.
      • Epstein M.P.
      • Glauser T.
      • Goldstein D.B.
      • Han Y.
      • Heinzen E.L.
      • Hitomi Y.
      • Howell K.B.
      • et al.
      Epi4K Consortium, Epilepsy Phenome/Genome Project
      De novo mutations in epileptic encephalopathies.
      ,
      • Janve V.S.
      • Hernandez C.C.
      • Verdier K.M.
      • Hu N.
      • Macdonald R.L.
      Epileptic encephalopathy de novo GABRB mutations impair GABAA receptor function.
      ,
      • Papandreou A.
      • McTague A.
      • Trump N.
      • Ambegaonkar G.
      • Ngoh A.
      • Meyer E.
      • Scott R.H.
      • Kurian M.A.
      GABRB3 mutations: a new and emerging cause of early infantile epileptic encephalopathy.
      ,
      • Baulac S.
      • Huberfeld G.
      • Gourfinkel-An I.
      • Mitropoulou G.
      • Beranger A.
      • Prud'homme J.F.
      • Baulac M.
      • Brice A.
      • Bruzzone R.
      • LeGuern E.
      First genetic evidence of GABAA receptor dysfunction in epilepsy: a mutation in the γ2-subunit gene.
      ,
      • Boillot M.
      • Morin-Brureau M.
      • Picard F.
      • Weckhuysen S.
      • Lambrecq V.
      • Minetti C.
      • Striano P.
      • Zara F.
      • Iacomino M.
      • Ishida S.
      • An-Gourfinkel I.
      • Daniau M.
      • Hardies K.
      • Baulac M.
      • Dulac O.
      • et al.
      Novel GABRG2 mutations cause familial febrile seizures.
      ,
      • Shen D.
      • Hernandez C.C.
      • Shen W.
      • Hu N.
      • Poduri A.
      • Shiedley B.
      • Rotenberg A.
      • Datta A.N.
      • Leiz S.
      • Patzer S.
      • Boor R.
      • Ramsey K.
      • Goldberg E.
      • Helbig I.
      • Ortiz-Gonzalez X.R.
      • et al.
      De novo GABRG2 mutations associated with epileptic encephalopathies.
      • Wallace R.H.
      • Marini C.
      • Petrou S.
      • Harkin L.A.
      • Bowser D.N.
      • Panchal R.G.
      • Williams D.A.
      • Sutherland G.R.
      • Mulley J.C.
      • Scheffer I.E.
      • Berkovic S.F.
      Mutant GABAA receptor γ2-subunit in childhood absence epilepsy and febrile seizures.
      ). For mutations in the γ2 subunit, the resultant receptors will either be a WT or contain a single mutation. However, for β3 mutations, a more complicated mixture of receptors will be expressed. A WT containing two normal β3 subunits, two heteromutant receptors containing a single mutation at either of the two β3 subunit locations within the pentamer, and a homomutant receptor containing the mutation at both β3 subunit locations within the pentamer can be formed. If the surface expression is unaffected and the distribution of mutant receptors into the complex is random, some 50% of the resultant receptors will contain mutations with just a single copy of the mutation. To date, there is a lack of research that assesses the effect of a single copy of the mutation on GABAA receptor function, an important component that contributes to understanding the epilepsy phenotype of individuals. Therefore, it is vital to determine how single copies of the mutation, as well as two copies, alter the function of the receptor to properly characterize the molecular phenotype of the mutation.
      We chose five mutations to investigate using the concatenated construct including one in the γ2 subunit (γ2R323Q) and four in the β3 subunit (β3D120N, β3T157M, β3S254F, and β3Y302C) (Fig. 1A). These mutations were chosen as they were located in different regions along the activation pathway of the receptor, including the M2-M3 coupling loop of the γ2 subunit (γ2R323Q). The β3 mutations were located in the area surrounding the ligand-binding site (β3D120N), a β-sheet within the extracellular domain (β3T157M), the M2-M3 coupling loop (β3Y302C), and the M1 transmembrane region (β3S254F) (Fig. 1, B and C). Previous functional genomic analysis of these mutations in Xenopus oocytes and HEK293 cells have demonstrated that the γ2R323Q, β3D120N, and β3Y302C mutations substantially reduce either the potency of GABA or the magnitude of GABA-activated currents when expressed in α1β3γ2 or α5β3γ3 receptors (
      • Møller R.S.
      • Larsen L.H.
      • Johannesen K.M.
      • Talvik I.
      • Talvik T.
      • Vaher U.
      • Miranda M.J.
      • Farooq M.
      • Nielsen J.E.
      • Svendsen L.L.
      • Kjelgaard D.B.
      • Linnet K.M.
      • Hao Q.
      • Uldall P.
      • Frangu M.
      • et al.
      Gene panel testing in epileptic encephalopathies and familial epilepsies.
      ,
      • Janve V.S.
      • Hernandez C.C.
      • Verdier K.M.
      • Hu N.
      • Macdonald R.L.
      Epileptic encephalopathy de novo GABRB mutations impair GABAA receptor function.
      ,
      • Shen D.
      • Hernandez C.C.
      • Shen W.
      • Hu N.
      • Poduri A.
      • Shiedley B.
      • Rotenberg A.
      • Datta A.N.
      • Leiz S.
      • Patzer S.
      • Boor R.
      • Ramsey K.
      • Goldberg E.
      • Helbig I.
      • Ortiz-Gonzalez X.R.
      • et al.
      De novo GABRG2 mutations associated with epileptic encephalopathies.
      ), whereas the β3T157M mutation caused only subtle changes at α5β3γ3 receptors (
      • Møller R.S.
      • Larsen L.H.
      • Johannesen K.M.
      • Talvik I.
      • Talvik T.
      • Vaher U.
      • Miranda M.J.
      • Farooq M.
      • Nielsen J.E.
      • Svendsen L.L.
      • Kjelgaard D.B.
      • Linnet K.M.
      • Hao Q.
      • Uldall P.
      • Frangu M.
      • et al.
      Gene panel testing in epileptic encephalopathies and familial epilepsies.
      ). However, these experiments did not fully describe the molecular phenotype of the mutations, as they were unable to distinguish how receptors that contain one or two copies of the mutation differ from the WT receptor or from each other.
      Figure thumbnail gr1
      Figure 1A, pentameric structure of the GABA and diazepam-bound α1β3γ2 receptor (PDB code 6HUP) from above showing the orientation of the subunits. The subunits are colored with respect to their order in the pentamer: first γ2 (green), second β3 (maroon), third α1 (blue), fourth β3 (red), and fifth α1 (dark blue). B, side view showing the first γ2 subunit adjacent to the second β3 subunit. The γ2R323 residue on the M2-M3 loop is depicted with the side chain in black in a transparent sphere. C, side view showing the fourth β3 subunit adjacent to the fifth α1 subunit, with the GABA-binding site highlighted. The side chains of the β3D120N residue at the GABA-binding site, the β3T157M residue within an internal β-sheet, the β3Y302C residue in the M2-M3 loop, and the β3S254F residue in the M1 region are shown in black.
      To resolve this question, we created a concatenated α1β3γ2 GABAA receptor construct with five linked subunits in the sequence γ2-β3-α1-β3-α1. For each mutation, we then created a set of receptor constructs that resembled the expressed receptors from an individual with a dominant mutation. Typically, but not always, a single copy of the mutation impaired the activation properties of the receptor, whereas a second copy intensified the effect of the mutation to be catastrophic.
      We propose that precise functional genomic analysis using concatenated receptors can identify the phenotype of the individual receptors that are expressed by patients with dominant mutations. This information may ultimately assist in precision medicine approaches, where mutant GABAA receptors containing a single mutation can be targeted to treat individual patients.

      Results

      Activation properties of WT concatenated receptor

      To determine how receptors with either a single or two copies of an epilepsy-causing mutation differ from WT receptors, we created a concatenated receptor construct with five subunits linked by AGS repeats (Fig. 2A). Although concatenated ligand-gated ion channels have previously been created, many of these do not reliably form in the standard orientation (
      • Ahring P.K.
      • Liao V.W.Y.
      • Balle T.
      Concatenated nicotinic acetylcholine receptors: a gift or a curse?.
      ) or are mixtures of dimeric and trimeric concatenated constructs (
      • Baumann S.W.
      • Baur R.
      • Sigel E.
      Forced subunit assembly in alpha1beta2gamma2 GABAA receptors: insight into the absolute arrangement.
      ). When injected alone, these dimeric and trimeric constructs can result in low currents that could potentially confound the analysis of mutations that impair receptor function (
      • Sigel E.
      • Kaur K.H.
      • Lüscher B.P.
      • Baur R.
      Use of concatamers to study GABAA receptor architecture and function: application to δ-subunit-containing receptors and possible pitfalls.
      ). However, constructs of concatenated pentameric GABAA receptors have also been described that contain β2 subunits (
      • Baur R.
      • Minier F.
      • Sigel E.
      A GABAA receptor of defined subunit composition and positioning: concatenation of five subunits.
      ). Therefore, the new construct was designed with five subunits in the same order so that the subunits would arrange themselves primarily counter-clockwise when viewed from the extracellular side of the membrane, forming two GABA-binding sites at the β3-α1 interfaces and a benzodiazepine-binding site at the α1–γ2 interface (Fig. 2B). Sequence encoding for the signal peptide of the α1 and β3 subunits were removed, and four different linkers were incorporated with lengths calculated with the same methodology as previously (
      • Ahring P.K.
      • Liao V.W.Y.
      • Balle T.
      Concatenated nicotinic acetylcholine receptors: a gift or a curse?.
      ). To ensure predominately counter-clockwise expression, the first linker was designed to be relatively short, and subsequent linkers were longer with similar lengths when N and C termini were taken into account. These linkers contained peptide sequences of (AGS)5 between the γ2 and β3 subunits ((AGS)5LGS(AGS)3 between the first β3 and α1 subunits, AGT(AGS)5 between the α1 and β3 subunits, and (AGS)4ATG(AGS)4 between the final β3 and α1 subunits) to form the DNA construct encoding the five subunits in the order of γ2-β3-α1-β3-α1 (Fig. 2, A and B).
      Figure thumbnail gr2
      Figure 2A, schematic of the coding region of concatenated receptor containing the DNA construct. Linker lengths are 15 amino acids ((AGS)5), 27 amino acids ((AGS)5LGS(AGS)3), 18 amino acids (AGT(AGS)5), and 27 amino acids ((AGS)4AGT(AGS)4). B, schematic of the expected arrangement of the concatenated receptor where the subunits arrange in a counter-clockwise orientation. GABA- and clobazam-binding sites are shown. C, representative data (above) from a single two-electrode voltage clamp experiment where different concentrations of GABA (open bars) were applied to construct a concentration–response curve to GABA (below). Filled bars, reference 3 mm GABA applications; open bars, GABA applications at concentrations shown. Peak currents were measured, and the mean ± S.E. (error bars) was plotted (open circles) and fitted to the Hill equation (below). D, representative data (above) from a single two-electrode voltage clamp experiment constructing a modulation curve to clobazam (below). Three pulses of reference 10 μm GABA (closed bars) were applied prior to co-application of 10 μm GABA and clobazam at concentrations shown (closed bars). Percent modulation of the control GABA response was calculated, and the mean ± S.E. was plotted and fitted to the Hill equation (below). The fitted EC50 of clobazam was 86 nm (log EC50 = −4.03 ± 0.06, mean ± S.E., n = 10), and the fitted Emax was 306% (320 ± 32, mean ± S.E., n = 10).
      cRNA (2 ng) of the WT concatenated construct was injected into Xenopus oocytes, and the oocytes were incubated for 2–4 days. Denaturing agarose gel electrophoresis was performed on the RNA to ensure that a single band at the correct size only was transcribed, and Western blotting was performed to ensure that the protein was properly translated and degradation products were not observed (Fig. S1). Peak currents were measured using two-electrode voltage clamp electrophysiology upon application of a range of GABA solutions, and the measured responses were used to construct the concentration–response curve (Fig. 2C). Injection of the WT cRNA resulted in robust GABA-activated currents, with 3 mm GABA eliciting an average current of 2 μA (Fig. 2C and Table 1). GABA activated the WT concatemer with an EC50 of 69 μm, similar to our previously published value of 53 μm and other previously published reports (e.g. 74 μm (
      • Ramerstorfer J.
      • Furtmüller R.
      • Sarto-Jackson I.
      • Varagic Z.
      • Sieghart W.
      • Ernst M.
      The GABAA receptor α+β-interface: a novel target for subtype selective drugs.
      )) where unlinked α1, β3, and γ2 subunits were injected into Xenopus oocytes (
      • Che Has A.T.
      • Absalom N.
      • van Nieuwenhuijzen P.S.
      • Clarkson A.N.
      • Ahring P.K.
      • Chebib M.
      Zolpidem is a potent stoichiometry-selective modulator of α1β3 GABAA receptors: evidence of a novel benzodiazepine site in the α1-α1 interface.
      ).
      Table 1Concentration–response curves of α1β3γ2 receptors
      ConstructEC50 μm (log EC50)I3 mm_GABAnHnEst. Po(max)n
      nA
      γ2-β3-α1-β3-α169.0 (−4.12 ± 0.06)2095 ± 1261.3 ± 0.1130.95 ± 0.0410
      γ2R323Q-β3-α1-β3-α1315 (−3.42 ± 0.04)***1395 ± 2140.9 ± 0.1***100.85 ± 0.0310
      γ2-β3D120N-α1-β3-α11144 (−2.77 ± 0.05)***701 ± 92***1.1 ± 0.1101.12 ± 0.0710
      γ2-β3T157M-α1-β3-α1422 (−3.37 ± 0.05)***1146 ± 210*1.2 ± 0.1100.90 ± 0.0510
      γ2-β3S254F-α1-β3-α1181 (−3.68 ± 0.04)***1230 ± 134*1.3 ± 0.1101.00 ± 0.0410
      γ2-β3Y302C-α1-β3-α1164 (−3.71 ± 0.05)***1826 ± 1651.0 ± 0.1*100.40 ± 0.06***10
      γ2-β3-α1-β3D120N-α11473 (−2.87 ± 0.11)***969 ± 158***0.8 ± 0.1***101.00 ± 0.0510
      γ2-β3-α1-β3T157M-α1279 (−3.53 ± 0.06)***1995 ± 2561.1 ± 0.1100.96 ± 0.0410
      γ2-β3-α1-β3S254F-α134.2 (−4.47 ± 0.07)**2480 ± 2481.2 ± 0.1100.95 ± 0.0310
      γ2-β3-α1-β3Y302C-α1471 (−3.35 ± 0.09)***1336 ± 156*1.1 ± 0.1*100.74 ± 0.05*10
      γ2-β3D120N-α1-β3D120N-α1ND
      ND, currents too low to determine concentration–response curve.
      NDND10ND
      γ2-β3T157M-α1-β3T157M-α1NDNDND10ND
      γ2-β3S254F-α1-β3S254F-α129.6 (−4.52 ± 0.05)***1528 ± 2101.3 ± 0.1100.89 ± 0.0310
      γ2-β3Y302C-α1-β3Y302C-α16806 (−2.16 ± 0.08)***77 ± 15***0.9 ± 0.1***100.24 ± 0.04***10
      a ND, currents too low to determine concentration–response curve.
      To ensure that the receptors were arranging in the correct orientation, the modulation of GABA-elicited currents of our concatenated receptors was measured using a benzodiazepine, clobazam, that binds selectively to the α1–γ2 interfaces. When clobazam was co-applied with 10 μm GABA, the response of the activated receptors was increased with increasing clobazam concentrations (Fig. 2D). We constructed a concentration–response curve of clobazam modulation of 10 μm GABA-activated currents. The maximum modulation by clobazam was 306%, and the EC50 value of clobazam was 86 nm, similar to previously published values of 256% and 132 nm at nonconcatenated α1β2γ2 receptors (
      • Hammer H.
      • Ebert B.
      • Jensen H.S.
      • Jensen A.A.
      Functional characterization of the 1,5-benzodiazepine clobazam and its major active metabolite N-desmethylclobazam at human GABAA receptors expressed in Xenopus laevis oocytes.
      ) (Fig. 2D). Taken together, the GABA and clobazam concentration–response curves demonstrate that the concatenated receptor reliably replicates the activation properties of its respective unlinked receptor. We then used this construct as a backbone so that mutant β3 and/or γ2 subunit(s) can be inserted at specific regions of the pentameric construct to analyze the effects of epilepsy-causing mutations. For γ2 mutations, a single copy of the mutation was inserted into the receptor, and for β3 mutations, either a single copy of the mutation was inserted within different subunits or two copies of the mutation were inserted into the receptor.

      Absolute expression levels of mutant receptors

      We chose five mutations to investigate using concatemers, including one in the γ2 subunit (γ2R323Q) and four in the β3 subunit (β3D120N, β3T157M, β3S254F, and β3Y302C). These mutations were chosen as they are located in different regions along the activation pathway of the receptor. In order from the extracellular to transmembrane domains, the amino acids included the area surrounding the ligand-binding site (β3D120N), a β-sheet located in the extracellular domain of the β3 subunit (β3T157M), the M2-M3 coupling loop of both the γ2 (γ2R323Q) and β3 subunit (β3Y302C), and the M1 region (β3S254F). The β3D120 and β3Y302 residues are located at the interface of the α1 and β3 subunits, the β3T157M residue is located within the β-sheet of the β3 subunits, and the β3S254 residue is located at the interface of a β3 and α1 subunit or at the interface of a β3 and γ2 subunit.
      For the γ2 mutation, we introduced a single copy of the γ2R323Q mutation into the first subunit of the concatenated construct. For each of the β3 mutations, we created a set of three constructs with a mutation in either the second or fourth β3 subunit and a construct with a mutation in both the second and fourth β3 subunits (Fig. 3A). We injected 2 ng of cRNA encoding for each of the constructs and then compared the absolute currents elicited by 3 mm GABA at the mutant receptors with the WT (Fig. 3B).
      Figure thumbnail gr3
      Figure 3A, schematic of concatenated receptor indicating the location of mutations when they are introduced into the γ2 or distinct β3 subunits. Red circles indicate the location of mutations on the first γ2 subunit, and red circles, purple squares, and blue circles indicate location of mutations on the second, fourth, or both second and fourth β3 subunits, respectively. B, representative traces of WT and γ2R323Q, β3D120N, β3T157M, β3S254F, and β3Y302C mutant receptors with mutation(s) in the labeled locations after application of reference 3 mm GABA (filled bars). Scale bars, 500 nA and 100 s. C, absolute current elicited by 3 mm GABA after injection of 2 ng of RNA. Individual data points are depicted as either open circles or squares with WT as black bars and gray circles and a color and pattern scheme identical to that in A. Bars and error bars represent mean ± S.D. of 10–13 individual cells. *, p < 0.05; **, p < 0.01; ***, p < 0.001 compared with WT, one-way ANOVA with Tukey's post hoc test.
      Strikingly, the incorporation of two mutations into the receptor was catastrophic for three of the β3 mutations. When two copies of the β3D120N or β3T157M mutations were incorporated into the concatemer, the GABA-elicited currents were too small to be measured, whereas the incorporation of two β3Y302C mutations significantly reduced the GABA-elicited currents (I3 mm_GABA = 77 nA, γ2-β3Y302C-α1-β3Y302C-α1; I3 mm_GABA = 2.1 μA, WT). In contrast, there was no significant difference in the current amplitudes compared with WT when two copies of the β3S254F mutation were incorporated into the receptor (I3 mm_GABA = 1.5 μA, γ2-β3S254F-α1-β3S254F-α1) (Fig. 3, B and C).
      In contrast, the incorporation of a single mutation into the receptor did not cause the same marked effects on the magnitude of absolute currents, with no more than a 3-fold reduction at any mutated concatemer. A single γ2R323Q mutation in the first subunit of the concatemer had similar GABA-activated currents (I3 mm_GABA = 1.4 μA, γ2R323Q-β3-α1-β3-α1) to WT, whereas the introduction of a single β3D120N mutation at either subunit location significantly reduced current amplitudes (I3 mm_GABA = 700 and 970 nA, γ2-β3D120N-α1-β3-α1 and γ2-β3-α1-β3D120N-α1, respectively). When introduced at the first β3 subunit, a single β3T157M or β3S254F mutation significantly reduced the current amplitudes (I3 mm_GABA = 1.1 and 1.2 μA, γ2-β3T157M-α1-β3-α1 and γ2-β3S254F-α1-β3-α1, respectively), but not when introduced in the second (I3 mm_GABA = 2.0 and 2.5 μA, γ2-β3-α1-β3T157M-α1 and γ2-β3-α1-β3S254F-α1 concatemers, respectively). A single β3Y302C mutation significantly reduced GABA-activated currents when introduced in the second β3 subunit, but not the first (I3 mm_GABA = 1.3 and 1.8 μA, γ2-β3-α1-β3Y302C-α1 and γ2-β3Y302C-α1-β3-α1, respectively) (Fig. 3, B and C).
      Although several concatemers containing single mutations had significant reductions in the maximum absolute currents elicited by 3 mm GABA, this crude approach is a poor measure of how mutations alter receptor properties. Variation in the maximum absolute current can be introduced in several ways that are a consequence of experimental conditions. These include large rightward shifts in the EC50 and small changes in the RNA concentration, the incubation time, and the rate at which the individual oocytes form protein and express receptors at the cell surface. However, mutations may also cause changes in the intrinsic activation properties of the receptor that reduce the current passing across the synapse. This can occur through changes in the potency of GABA or changes in the efficacy, where GABA reverts to a more partial agonist, or both. To determine whether the mutations changed either the potency or efficacy of GABA activation, we next constructed concentration–response curves to GABA and estimated the maximum open probability of GABA to determine whether the mutations changed these intrinsic activation properties of the receptor, when either one or two copies of the mutation were present.

      γ2R323Q impairs GABA activation properties of the receptor

      We therefore constructed concentration–response curves to GABA and standardized the response against an estimated maximum open probability of the receptor for the WT and each mutation. We initially compared the γ2R323Q-β3-α1-β3-α1 mutant receptor with the WT, as this mutation is incorporated into the receptor within the only γ2 subunit (Fig. 4A). The potency of GABA has previously been shown to be reduced at α1β2γ2 receptors when expressed in HEK293 cells, and we would expect comparable results in our concatenated construct (
      • Shen D.
      • Hernandez C.C.
      • Shen W.
      • Hu N.
      • Poduri A.
      • Shiedley B.
      • Rotenberg A.
      • Datta A.N.
      • Leiz S.
      • Patzer S.
      • Boor R.
      • Ramsey K.
      • Goldberg E.
      • Helbig I.
      • Ortiz-Gonzalez X.R.
      • et al.
      De novo GABRG2 mutations associated with epileptic encephalopathies.
      ).
      Figure thumbnail gr4
      Figure 4A, schematic of concatenated receptor indicating the location of the γ2R323Q mutation (red circle) within the concatenated construct (left). Shown are representative data (right) from a single two-electrode voltage clamp experiment where different concentrations of GABA (open bars) were applied to construct a concentration–response curve to GABA at γ2R323Q-β3-α1-β3-α1 receptors. Filled dark red bars and traces represent reference 3 mm GABA applications, and open red bars and traces represent GABA applications at the concentrations shown. Shown is a concentration–response curve to GABA (below) of WT γ2-β3-α1-β3-α1 (○) and γ2R323Q-β3-α1-β3-α1 (○) receptors normalized to the Est. Po(max) and fitted to the Hill equation. Dots, mean ± S.E. (error bars) of 10–13 individual experiments. The EC50 of the mutant receptor derived from the curve fit is shown. *, p < 0.05; **, p < 0.01; ***, p < 0.001 compared with WT, one-way ANOVA with Tukey's post hoc test. B, representative traces of WT (black) and γ2R323Q (red) mutant receptors after application of reference 3 mm GABA (filled bars) and 10 mm GABA, 1 μm diazepam, and 3 μm etomidate (open bars), respectively. Scale bars, 500 nA and 100 s. The Est. GABA Po(max) of WT (○) and γ2R323Q (○) mutant receptors (below) was determined by dividing the current elicited by 3 mm GABA by the current elicited by 10 mm GABA, 1 μm diazepam, and 3 μm etomidate and corrected for the reference 3 mm GABA current. Lines and error bars represent the mean ± S.D. of 10 individual cells.
      We first tested whether the γ2R323Q mutation altered the potency of GABA. We constructed concentration–response curves to GABA at γ2R323Q-β3-α1-β3-α1 receptors to determine the EC50 of the mutant receptors. These experiments were run on an automated protocol, where 3 mm GABA was applied as a reference and internal standard three times during the experiment for all receptors (Fig. 4A). Similar to the results reported using this mutation with free α1 and β3 subunits in HEK293 cells (
      • Shen D.
      • Hernandez C.C.
      • Shen W.
      • Hu N.
      • Poduri A.
      • Shiedley B.
      • Rotenberg A.
      • Datta A.N.
      • Leiz S.
      • Patzer S.
      • Boor R.
      • Ramsey K.
      • Goldberg E.
      • Helbig I.
      • Ortiz-Gonzalez X.R.
      • et al.
      De novo GABRG2 mutations associated with epileptic encephalopathies.
      ), there was a decrease in the potency of GABA with a significant 4.5-fold decrease in the potency of GABA (EC50 = 315 μm) (Fig. 4A and Table 1), demonstrating that the activation of receptors by GABA is impaired by the γ2R323Q mutation.
      We then determined whether the maximal efficacy of GABA was impaired by the γ2R323Q mutation by estimating the maximum Po (Est. Po(max)) at WT and γ2R323Q-β3-α1-β3-α1 receptors using a pharmacological technique similar to Shin et al. (
      • Shin D.J.
      • Germann A.L.
      • Johnson A.D.
      • Forman S.A.
      • Steinbach J.H.
      • Akk G.
      Propofol is an allosteric agonist with multiple binding sites on concatemeric ternary GABAA receptors.
      ). At oocytes expressing either WT or mutant receptors, we applied the 3 mm GABA reference and then co-applied 10 mm GABA with 1 μm etomidate and 3 μm diazepam to shift as many receptors as possible to the open state (Fig. 4B). We assumed that the combination of GABA with etomidate and diazepam opened the receptors with a probability approaching 1. The Est. Po(max) of GABA for each receptor was then calculated by dividing the current elicited by 3 mm GABA by the current elicited by GABA, etomidate, and diazepam and corrected to account for shifts in the concentration–response curves (Fig. 4B and Table 1). We refer to Est. Po(max) as an estimated maximum open probability as the true current amplitude may be underestimated by mutations that greatly impair receptor activation or modulation or change desensitization kinetics. As expected, GABA elicited a very high Est. Po(max) of 0.95 at WT receptors, consistent with single-channel recordings where the channel enters a long-lived open state (
      • Fisher J.L.
      • Macdonald R.L.
      Single channel properties of recombinant GABAA receptors containing γ2 or δ subtypes expressed with α1 and β3 subtypes in mouse L929 cells.
      ). The γ2R323Q mutation did not significantly alter the Est. Po(max) with a value of 0.85 (Fig. 4B and Table 1).
      When incorporated into the first subunit of the concatemer, the γ2R323Q mutation reduced the potency of GABA without causing a significant reduction in the efficacy. These changes in the activation properties of the receptor caused by the γ2R323Q mutation in the concatenated construct were similar to the reported effects of receptors composed of unlinked subunits. Therefore, the concatenated construct is a suitable method of analyzing the effect of mutations on the activation properties of the receptor.

      A single β3D120N or β3T157M mutation impairs GABA potency, and two mutations are catastrophic

      We next assessed the effects of two β3 mutations, β3D120N and β3T157M, that are both located in the extracellular domain at the earlier stages of the activation pathway. The β3D120N mutation has previously been expressed in combination with either α1, β3, and γ2 or α1 and γ2 free subunits to determine the effects of heterozygous or homozygous expression, respectively, on receptor function (
      • Janve V.S.
      • Hernandez C.C.
      • Verdier K.M.
      • Hu N.
      • Macdonald R.L.
      Epileptic encephalopathy de novo GABRB mutations impair GABAA receptor function.
      ). Both the gating properties and the absolute expression levels of the receptor were reduced in both cases, whereas in a separate study, the incorporation of the β3T157M mutation with α5 and γ2 subunits only made subtle changes to the activation properties of the receptor (
      • Møller R.S.
      • Larsen L.H.
      • Johannesen K.M.
      • Talvik I.
      • Talvik T.
      • Vaher U.
      • Miranda M.J.
      • Farooq M.
      • Nielsen J.E.
      • Svendsen L.L.
      • Kjelgaard D.B.
      • Linnet K.M.
      • Hao Q.
      • Uldall P.
      • Frangu M.
      • et al.
      Gene panel testing in epileptic encephalopathies and familial epilepsies.
      ). To determine the effect of the mutations when they were expressed in a single subunit within the receptor, we constructed concentration–response curves to GABA and measured the Est. Po(max) at concatenated receptors containing a single copy of a mutation.
      We created concatenated constructs by introducing β3D120N or β3T157M mutations in the second, the fourth or both the second and fourth subunits in the concatemer and constructed concentration–response curves to GABA (Fig. 5A). A single copy of the β3D120N mutation significantly reduced the potency of GABA by 16–20-fold, regardless of whether it was introduced at the second or fourth subunit (EC50 = 1.14 and 1.47 mm, γ2-β3D120N-α1-β3-α1 and γ2-β3-α1-β3D120N-α1, respectively) (Fig. 5B and Table 1). These EC50 values were not significantly different from each other, strongly suggesting that the subunit location of the β3D120N mutation within the pentameric structure did not affect how the mutation altered receptor activation properties.
      Figure thumbnail gr5
      Figure 5A, schematic of concatenated receptor indicating the location of β3D120N and β3T157M mutations introduced within the second (closed red circle), fourth (closed purple square) or the second and fourth (closed blue circles) subunits within the resulting pentameric receptor. B, concentration–response curve to GABA of WT γ2-β3-α1-β3-α1 (open black circle), γ2-β3D120N-α1-β3-α1 (open red circle), γ2-β3-α1-β3D120N-α1 (open purple square), and γ2-β3D120N-α1-β3D120N-α1 (closed blue circle) receptors; C, Concentration–response curve to GABA of WT γ2-β3-α1-β3-α1 (open black circle), γ2-β3T157M-α1-β3-α1 (open red circle), γ2-β3-α1-β3T157M-α1 (open purple square), and γ2-β3T157M-α1-β3T157M-α1 (closed blue circle) receptors normalized to the Est. Po(max) and fitted to the Hill equation. The EC50 of the mutant receptor derived from the curve fit is shown. *, p < 0.05; **, p < 0.01; ***, p < 0.001 compared with WT, one-way ANOVA with Tukey's post hoc test. D, representative traces of β3D120N and β3T157M receptors after application of 3 mm GABA (filled bars) and 10 mm GABA, 1 μm diazepam, and 3 μm etomidate (open bars), respectively. Red traces indicate mutation at the second subunit location, and purple traces indicate mutation at the fourth. Scale bars, 500 nA and 100 s. E, estimated GABA Po(max) of WT, β3D120N, and β3T157M receptors was determined by dividing the current elicited by 3 mm GABA by the current elicited by 10 mm GABA, 1 μm diazepam, and 3 μm etomidate and corrected for the reference 3 mm GABA. Open gray circles, WT; open red circles, receptors with a mutation in the second position; open purple squares, receptors with a mutation in the fourth position. Lines and error bars represent mean ± S.D. of 10 individual cells.
      Similarly, the β3T157M mutations significantly reduced the potency of GABA by 4–6-fold when introduced at either the second or fourth subunit (EC50 = 422 and 279 μm, γ2-β3T157M-α1-β3-α1 and γ2-β3-α1-β3T157M-α1, respectively) (Fig. 5C and Table 1). Again, the EC50 values were not significantly different from each other, demonstrating that the subunit location of the β3T157M did not affect how the mutation altered the activation properties of GABA. Neither receptor expressed measurable currents when a mutation was introduced in both β3 subunits, and as such, concentration–response curves could not be constructed.
      We next measured the Est. Po(max) of two receptors with a single copy of the mutation in the second or fourth subunit, respectively, to determine whether the efficacy of GABA had been altered (Fig. 5D). Despite the absolute current levels being reduced by the β3D120N mutation, the combination of etomidate and diazepam failed to appreciably increase the maximal response to GABA at either receptor containing a single mutation (Fig. 5, D and E). This is likely due in part to the large rightward shift of the concentration–response curve where 3 mm GABA no longer elicits the maximum response. Similarly, the combination of etomidate and diazepam had little effect on the maximal GABA current elicited at the two receptors containing a single β3T157M mutation (Fig. 5, D and E). Consequently, there was no significant difference in the Est. Po(max) at the four receptors (Est. Po(max) = 1.12, 1.00, 0.9, and 0.96, γ2-β3D120N-α1-β3-α1, γ2-β3-α1-β3D120N-α1, γ2-β3T157M-α1-β3-α1, and γ2-β3-α1-β3T157M-α1, respectively) (Fig. 5E and Table 1).
      Taken together, single copies of the β3D120N and β3T157M mutations at the earlier stages of the activation pathway both impair GABA activation of the receptor by reducing the potency of GABA by ∼20- and 5-fold, respectively, without altering the maximal efficacy of GABA. There was little difference in the effect of either single mutation when located at different subunits β3 within a pentamer. A second copy of the β3D120N or β3T157M mutation intensifies the effect of the mutation and appears to be catastrophic, leading to little to no functional receptor expression.

      β3S254F and β3Y302C mutation effects are dependent on location and number of mutations

      We next assessed the effects of two other β3 mutations, β3Y302C and β3S254F. The β3Y302 residue is located in the M2-M3 coupling loop, a key motif in the activation pathway that links extracellular and transmembrane domains. The β3S254 residue is located in the transmembrane regions within the M1 transmembrane helix that moves late in the activation process of ligand-gated ion channels (
      • Purohit P.
      • Gupta S.
      • Jadey S.
      • Auerbach A.
      Functional anatomy of an allosteric protein.
      ). The β3Y302C mutation has been shown to impair receptor activation when expressed with either α1 and γ2 subunits or α5 and γ2 receptors (
      • Møller R.S.
      • Larsen L.H.
      • Johannesen K.M.
      • Talvik I.
      • Talvik T.
      • Vaher U.
      • Miranda M.J.
      • Farooq M.
      • Nielsen J.E.
      • Svendsen L.L.
      • Kjelgaard D.B.
      • Linnet K.M.
      • Hao Q.
      • Uldall P.
      • Frangu M.
      • et al.
      Gene panel testing in epileptic encephalopathies and familial epilepsies.
      ,
      • Janve V.S.
      • Hernandez C.C.
      • Verdier K.M.
      • Hu N.
      • Macdonald R.L.
      Epileptic encephalopathy de novo GABRB mutations impair GABAA receptor function.
      ), whereas there are no functional data on how the β3S254F mutation affects receptor activation.
      A single copy of the β3Y302C mutation, introduced at either the second or the fourth subunit, significantly reduced the potency of GABA between 2- and 7-fold (EC50 = 167 and 471 μm, γ2-β3Y302C-α1-β3-α1 and γ2-β3-α1-β3Y302C-α1, respectively), and these EC50 values differed significantly from each other (Fig. 6A and Table 1). Two copies of the β3Y302C mutation were catastrophic, further reducing the potency of GABA by nearly 100-fold, an order of magnitude greater than either of the single mutations (EC50 = 6.81 mm, γ2-β3Y302C-α1-β3Y302C-α1). This EC50 value was significantly greater than the EC50 value of the WT or the two concatemers containing a single β3Y302C mutation (Fig. 6A and Table 1).
      Figure thumbnail gr6
      Figure 6A, concentration–response curve to GABA of WT γ2-β3-α1-β3-α1 (open black circles), γ2-β3Y302C-α1-β3-α1 (open red circles), γ2-β3-α1-β3Y302C-α1 (open purple squares) and γ2-β3Y302C-α1-β3Y302C-α1 (closed blue circles) receptors normalized to the Est. Po(max) and fitted to the Hill equation. Dots, mean ± S.E. (error bars) of 10–13 individual experiments. The EC50 of the mutant receptor derived from the curve fit is shown. *, p < 0.05; **, p < 0.01; ***, p < 0.001 compared with WT, one-way ANOVA with Tukey's post hoc test. B, representative traces of β3Y302C mutant receptors after application of reference 3 mm GABA (filled bars) and 10 mm GABA, 1 μm diazepam, and 3 μm etomidate (open bars), respectively. Red traces indicate mutation at the second subunit location, purple indicates mutation at the fourth subunit location, and blue indicates mutation at both the second and fourth locations. Scale bars, 500 nA and 100 s. C, estimated GABA Po(max) of WT γ2-β3-α1-β3-α1 (open black circles), γ2-β3Y302C-α1-β3-α1 (open red circles), γ2-β3-α1-β3Y302C-α1 (open purple squares) and γ2-β3Y302C-α1-β3Y302C-α1 (closed blue circles) mutant receptors. Est. Po(max) was determined by dividing the current elicited by 3 mm GABA by the current elicited by 10 mm GABA, 1 μm diazepam, and 3 μm etomidate and corrected where 3 mm GABA was not at the maximum of the concentration–response curves. Lines and bars, mean ± S.D. of 10 individual cells. *, p < 0.05; **, p < 0.01; ***, p < 0.001 compared with WT, one-way ANOVA with Tukey's post hoc test. D, concentration–response curve to GABA of WT γ2-β3-α1-β3-α1 (open black circles), γ2-β3S254F-α1-β3-α1 (open red circles), γ2-β3-α1-β3S254F-α1 (open purple squares), and γ2-β3S254F-α1-β3S254F-α1 (closed blue circles) receptors normalized to the Est. Po(max) and fitted to the Hill equation. Dots, mean ± S.E. of 10–13 individual experiments. The EC50 of the mutant receptor derived from the curve fit is shown. *, p < 0.05; **, p < 0.01; ***, p < 0.001 compared with WT, one-way ANOVA with Tukey's post hoc test. E, representative traces of β3Y302C mutant receptors after application of reference 3 mm GABA (filled bars) and 10 mm GABA, 1 μm diazepam, and 3 μm etomidate (open bars), respectively. Red traces, mutation at the second subunit location; purple traces, mutation at the fourth subunit location; blue traces, mutation at both the second and fourth locations. Scale bars, 500 nA and 100 s. F, estimated GABA Po(max) of WT γ2-β3-α1-β3-α1 (open gray circles), γ2-β3S254F-α1-β3-α1 (open red circles), γ2-β3-α1-β3S254F-α1 (open purple squares), and γ2-β3S254F-α1-β3S254F-α1 (closed blue circles) mutant receptors. Est. Po(max) was determined by dividing the current elicited by 3 mm GABA by the current elicited by 10 mm GABA, 1 μm diazepam, and 3 μm etomidate and corrected where 3 mm GABA was not at the maximum of the concentration–response curves. Lines and error bars, mean ± S.D. of 10 individual cells. *, p < 0.05; **, p < 0.01; ***, p < 0.001 compared with WT, one-way ANOVA with Tukey's post hoc test.
      Additionally, the introduction of the β3Y302C mutation significantly reduced the Est. Po(max) compared with WT, regardless of the subunit location of the mutation or whether one or two copies of the mutation were introduced. The efficacy of GABA was least affected by one copy of the β3Y302C mutation at the fourth subunit (Est. Po(max) = 0.74, γ2-β3-α1-β3Y302C-α1), whereas a single copy of the β3Y302C mutation at the second subunit significantly reduced the efficacy of GABA compared with the WT or the receptor with a single mutation at the fourth subunit (Est. Po(max) = 0.4, γ2-β3Y302C-α1-β3-α1). Two copies of the β3Y302C mutation resulted in the lowest efficacy of GABA (Po(max) = 0.24, γ2-β3Y302C-α1-β3Y302C-α1) (Fig. 6 (B and C) and Table 1). This demonstrates that, unique among the mutations that we have investigated, the single β3Y302C mutation impairs activation of the receptor to decrease both the potency and maximum efficacy of GABA activation. Differences in the magnitude of the reduction in the maximal efficacy suggest that these residues may not be equivalent when located in different subunits.
      The introduction of the β3S254F mutation did not follow the same pattern as the other mutations, which reduced the potency of GABA regardless of the subunit location of the mutation. Instead, when the mutation was introduced at the second subunit, the EC50 value of GABA was significantly increased nearly 3-fold compared with the WT (EC50 = 181 μm, γ2-β3S254F-α1-β3-α1), demonstrating that the activation properties of this concatemer were impaired (Fig. 6D and Table 1). However, when the mutation was introduced at the fourth subunit, the EC50 value of GABA significantly decreased 2-fold compared with the WT, as did the EC50 when the β3S254F mutation was introduced at both the second and fourth subunits (EC50 = 34.2 and 29.6 μm, γ2-β3-α1-β3S254F-α1 and γ2-β3S254F-α1-β3S254F-α1, respectively) (Fig. 6D and Table 1). This demonstrates that the subunit location of the β3S254F mutation defines the functional effect of the receptor, determining whether the potency of GABA has increased or decreased.
      The maximal efficacy of GABA was not changed by the β3S254F, regardless of whether one or two copies of the mutation were incorporated into the concatemer (Est. Po(max) = 1.00, 0.95, and 0.89, γ2-β3S254F-α1-β3-α1, γ2-β3-α1-β3S254F-α1, and γ2-β3S254F-α1-β3S254F-α1, respectively) (Fig. 6 (E and F) and Table 1). The differences in the EC50 value at receptors with a single β3S254F mutation at different subunit locations suggest that these locations are not equivalent in the activation pathway.

      Discussion

      Recent advances in whole-genome sequencing have enabled the identification of a large number of de novo mutations that cause a range of severe childhood epilepsies. In all of these cases, the mutations are dominant (
      • Cossette P.
      • Liu L.
      • Brisebois K.
      • Dong H.
      • Lortie A.
      • Vanasse M.
      • Saint-Hilaire J.M.
      • Carmant L.
      • Verner A.
      • Lu W.Y.
      • Wang Y.T.
      • Rouleau G.A.
      Mutation of GABRA1 in an autosomal dominant form of juvenile myoclonic epilepsy.
      • Gontika M.P.
      • Konialis C.
      • Pangalos C.
      • Papavasiliou A.
      Novel SCN1A and GABRA1 gene mutations with diverse phenotypic features and the question on the existence of a broader spectrum of Dravet syndrome.
      ,
      • Hernandez C.C.
      • Klassen T.L.
      • Jackson L.G.
      • Gurba K.
      • Hu N.
      • Noebels J.L.
      • Macdonald R.L.
      Deleterious rare variants reveal risk for loss of GABAA receptor function in patients with genetic epilepsy and in the general population.
      ,
      • Johannesen K.
      • Marini C.
      • Pfeffer S.
      • Møller R.S.
      • Dorn T.
      • Niturad C.E.
      • Gardella E.
      • Weber Y.
      • Søndergård M.
      • Hjalgrim H.
      • Nikanorova M.
      • Becker F.
      • Larsen L.H.
      • Dahl H.A.
      • Maier O.
      • et al.
      Phenotypic spectrum of GABRA1: from generalized epilepsies to severe epileptic encephalopathies.
      ,
      • Kodera H.
      • Ohba C.
      • Kato M.
      • Maeda T.
      • Araki K.
      • Tajima D.
      • Matsuo M.
      • Hino-Fukuyo N.
      • Kohashi K.
      • Ishiyama A.
      • Takeshita S.
      • Motoi H.
      • Kitamura T.
      • Kikuchi A.
      • Tsurusaki Y.
      • et al.
      De novo GABRA1 mutations in Ohtahara and West syndromes.
      ,
      • Lachance-Touchette P.
      • Brown P.
      • Meloche C.
      • Kinirons P.
      • Lapointe L.
      • Lacasse H.
      • Lortie A.
      • Carmant L.
      • Bedford F.
      • Bowie D.
      • Cossette P.
      Novel α1 and γ2 GABAA receptor subunit mutations in families with idiopathic generalized epilepsy.
      ,
      • Møller R.S.
      • Larsen L.H.
      • Johannesen K.M.
      • Talvik I.
      • Talvik T.
      • Vaher U.
      • Miranda M.J.
      • Farooq M.
      • Nielsen J.E.
      • Svendsen L.L.
      • Kjelgaard D.B.
      • Linnet K.M.
      • Hao Q.
      • Uldall P.
      • Frangu M.
      • et al.
      Gene panel testing in epileptic encephalopathies and familial epilepsies.
      ,
      • Stosser M.B.
      • Lindy A.S.
      • Butler E.
      • Retterer K.
      • Piccirillo-Stosser C.M.
      • Richard G.
      • McKnight D.A.
      High frequency of mosaic pathogenic variants in genes causing epilepsy-related neurodevelopmental disorders.
      ,
      • Allen A.S.
      • Berkovic S.F.
      • Cossette P.
      • Delanty N.
      • Dlugos D.
      • Eichler E.E.
      • Epstein M.P.
      • Glauser T.
      • Goldstein D.B.
      • Han Y.
      • Heinzen E.L.
      • Hitomi Y.
      • Howell K.B.
      • et al.
      Epi4K Consortium, Epilepsy Phenome/Genome Project
      De novo mutations in epileptic encephalopathies.
      ,
      • Janve V.S.
      • Hernandez C.C.
      • Verdier K.M.
      • Hu N.
      • Macdonald R.L.
      Epileptic encephalopathy de novo GABRB mutations impair GABAA receptor function.
      ,
      • Papandreou A.
      • McTague A.
      • Trump N.
      • Ambegaonkar G.
      • Ngoh A.
      • Meyer E.
      • Scott R.H.
      • Kurian M.A.
      GABRB3 mutations: a new and emerging cause of early infantile epileptic encephalopathy.
      ,
      • Baulac S.
      • Huberfeld G.
      • Gourfinkel-An I.
      • Mitropoulou G.
      • Beranger A.
      • Prud'homme J.F.
      • Baulac M.
      • Brice A.
      • Bruzzone R.
      • LeGuern E.
      First genetic evidence of GABAA receptor dysfunction in epilepsy: a mutation in the γ2-subunit gene.
      ,
      • Boillot M.
      • Morin-Brureau M.
      • Picard F.
      • Weckhuysen S.
      • Lambrecq V.
      • Minetti C.
      • Striano P.
      • Zara F.
      • Iacomino M.
      • Ishida S.
      • An-Gourfinkel I.
      • Daniau M.
      • Hardies K.
      • Baulac M.
      • Dulac O.
      • et al.
      Novel GABRG2 mutations cause familial febrile seizures.
      ,
      • Shen D.
      • Hernandez C.C.
      • Shen W.
      • Hu N.
      • Poduri A.
      • Shiedley B.
      • Rotenberg A.
      • Datta A.N.
      • Leiz S.
      • Patzer S.
      • Boor R.
      • Ramsey K.
      • Goldberg E.
      • Helbig I.
      • Ortiz-Gonzalez X.R.
      • et al.
      De novo GABRG2 mutations associated with epileptic encephalopathies.
      • Wallace R.H.
      • Marini C.
      • Petrou S.
      • Harkin L.A.
      • Bowser D.N.
      • Panchal R.G.
      • Williams D.A.
      • Sutherland G.R.
      • Mulley J.C.
      • Scheffer I.E.
      • Berkovic S.F.
      Mutant GABAA receptor γ2-subunit in childhood absence epilepsy and febrile seizures.
      ), whereby patients will contain one WT and one mutant copy of the gene. In cases where the mutations are in the β3 subunit of the GABAA receptor, this will lead to several potential receptors being expressed in each subtype, with heteromutant receptors containing single mutations at either of the two β3 subunits and a homomutant receptor containing mutations at both β3 subunits. Precise functional genomic analysis requires the understanding of how each of these individual receptors are affected by the mutation, as these receptors could be expressed and contribute to the pathology of the disorder or even be targeted by GABAergic drugs to treat the seizures.
      The in vitro analysis of these mutations has, to date, relied solely on injection or transfection of WT and/or mutant subunits in heterologous systems and quantification of receptor expression levels complemented with whole-cell recording and, at times, single-channel analysis (
      • Møller R.S.
      • Larsen L.H.
      • Johannesen K.M.
      • Talvik I.
      • Talvik T.
      • Vaher U.
      • Miranda M.J.
      • Farooq M.
      • Nielsen J.E.
      • Svendsen L.L.
      • Kjelgaard D.B.
      • Linnet K.M.
      • Hao Q.
      • Uldall P.
      • Frangu M.
      • et al.
      Gene panel testing in epileptic encephalopathies and familial epilepsies.
      ,
      • Janve V.S.
      • Hernandez C.C.
      • Verdier K.M.
      • Hu N.
      • Macdonald R.L.
      Epileptic encephalopathy de novo GABRB mutations impair GABAA receptor function.
      ,
      • Shen D.
      • Hernandez C.C.
      • Shen W.
      • Hu N.
      • Poduri A.
      • Shiedley B.
      • Rotenberg A.
      • Datta A.N.
      • Leiz S.
      • Patzer S.
      • Boor R.
      • Ramsey K.
      • Goldberg E.
      • Helbig I.
      • Ortiz-Gonzalez X.R.
      • et al.
      De novo GABRG2 mutations associated with epileptic encephalopathies.
      ). However, this approach is inadequate to describe the entire molecular phenotype, as mixed populations of receptors with one or two mutations will form at a ratio of 2:1. This will be particularly problematic when the maximal efficacy is reduced by the mutation, as the higher efficacy of the WT receptor will dominate the signal in whole-cell recordings. Therefore, by using a concatenated receptor construct, we have derived results from heteromutant β3 receptors that provide significant insights into the molecular phenotypes of epilepsies caused by GABAA receptor mutations as well as knowledge about the activation mechanisms of these receptors.

      Mutations impair synaptic transmission through efficacy or potency of GABA activation

      There is an enormous amount of understanding of how GABAAR opening is triggered through an activation pathway initiated by GABA binding that then opens the intrinsic ion channel to mediate neuronal inhibition (
      • Bouzat C.
      New insights into the structural bases of activation of Cys-loop receptors.
      ). Briefly, the agonist binds at the interface between a β3 and an α1 subunit at the extracellular domain, causing a series of conformational changes within the receptor that lead to the transmembrane domains. At the interface of the extracellular and transmembrane domains, a set of interacting loops, including the β1-β2, β6-β7, and β8-β9 in the extracellular domain and pre-M1 and M2-M3 loops connecting the transmembrane domain, alter their conformation during receptor activation (
      • Kash T.L.
      • Dizon M.J.
      • Trudell J.R.
      • Harrison N.L.
      Charged residues in the β2 subunit involved in GABAA receptor activation.
      ,
      • Kash T.L.
      • Jenkins A.
      • Kelley J.C.
      • Trudell J.R.
      • Harrison N.L.
      Coupling of agonist binding to channel gating in the GABAA receptor.
      ). This leads to a tilting of the M2 helices and ultimately channel opening. Additionally, the number of molecules bound is important, where two molecules of agonist are required to be bound to fully activate the receptor (
      • Sigel E.
      • Steinmann M.E.
      Structure, function, and modulation of GABAA receptors.
      ).
      Synaptic receptors that contain a γ2 subunit typically have intrinsic activation properties distinct from those of extrasynaptic receptors that contain a δ subunit, including a high efficacy where the maximal open probability elicited by GABA approaches 1 and lower potency (
      • Fisher J.L.
      • Macdonald R.L.
      Single channel properties of recombinant GABAA receptors containing γ2 or δ subtypes expressed with α1 and β3 subtypes in mouse L929 cells.
      ,
      • Ahring P.K.
      • Bang L.H.
      • Jensen M.L.
      • Strøbaek D.
      • Hartiadi L.Y.
      • Chebib M.
      • Absalom N.
      A pharmacological assessment of agonists and modulators at α4β2γ2 and α4β2δ GABAA receptors: the challenge in comparing apples with oranges.
      ). The maximal efficacy of an agonist, or whether the agonist is partial or full, is largely defined by transitional conformational states, known as pre-activated or “flip” states, that precede the final conformational change that opens the channel gate (
      • Lape R.
      • Colquhoun D.
      • Sivilotti L.G.
      On the nature of partial agonism in the nicotinic receptor superfamily.
      ). Residues at the earlier stage of the activation pathway (
      • Janve V.S.
      • Hernandez C.C.
      • Verdier K.M.
      • Hu N.
      • Macdonald R.L.
      Epileptic encephalopathy de novo GABRB mutations impair GABAA receptor function.
      ) appear to have no influence on the maximal efficacy of GABA, with β3D120N and β3T157M single mutations having an efficacy similar to that of WT. At these mutations, located near the ligand-binding domain and within an extracellular structural β-sheet, respectively, GABA remains acting as essentially a “full” agonist, activating the receptor with a very high maximal open probability.
      However, the β3Y302C mutations in the M2-M3 loop change the intrinsic property of receptor activation such that the maximal efficacy of GABA is significantly reduced. Essentially, GABA has become a “partial” agonist at these receptors, where the mutation in the coupling region may be destabilizing transitional conformational states. Further, the efficacy of GABA was different, depending on the subunit location of the β3Y302C mutation, suggesting that the two mutations are not entirely equivalent.

      Asymmetrical effects of the same mutation at different subunit locations

      A notable feature of these mutations is the differential effect of the β3S254F mutation when located at different subunit locations and slightly different levels of reductions in the efficacy of GABA at receptors with a β3Y302C mutation in different locations. This is in contrast with the β3D120N and β3T157M mutations, where the location of the individual mutations had no significant effect. It is known that the pseudosymmetry of the pentameric receptor can cause positional effects of mutations (
      • Sigel E.
      • Baur R.
      • Boulineau N.
      • Minier F.
      Impact of subunit positioning on GABAA receptor function.
      ), and there are two possible reasons for these differences. First, the local environment surrounding mutated residues may be identical at the two subunit locations for some mutations but not others. Second, the conformational changes during the activation process may not be symmetrical from ligand binding to channel opening. Furthermore, in epilepsy-causing mutations that impair surface expression, the location of the mutation can also determine the severity of the effect (
      • Gallagher M.J.
      • Song L.
      • Arain F.
      • Macdonald R.L.
      The juvenile myoclonic epilepsy GABAA receptor α1 subunit mutation A322D produces asymmetrical, subunit position-dependent reduction of heterozygous receptor currents and α1 subunit protein expression.
      ).
      Recent advances in cryogenic EM (cryo-EM) have enabled the solving of many membrane-bound proteins to a very fine resolution. During the preparation of this manuscript, a cryo-EM structure of the α1β3γ2 receptor with and without GABA and diazepam bound and a structure of the α1β1γ2 receptor were published (
      • Phulera S.
      • Zhu H.
      • Yu J.
      • Claxton D.P.
      • Yoder N.
      • Yoshioka C.
      • Gouaux E.
      Cryo-EM structure of the benzodiazepine-sensitive α1β1γ2S tri-heteromeric GABAA receptor in complex with GABA.
      ,
      • Laverty D.
      • Desai R.
      • Uchański T.
      • Masiulis S.
      • Stec W.J.
      • Malinauskas T.
      • Zivanov J.
      • Pardon E.
      • Steyaert J.
      • Miller K.W.
      • Aricescu A.R.
      Cryo-EM structure of the human α1β3γ2 GABAA receptor in a lipid bilayer.
      ). The sequences between the β2 and β3 subunits are highly homologous and identical within the regions containing the mutations studied. Hence, we can utilize the cryo-EM structure to explain the positional effects, or lack thereof, of our mutations.
      Within the GABA and diazepam-bound α1β3γ2 structure, the β3120 residue was located within the general vicinity of the ligand-binding site at the interface of the β3 and α1 subunits (Fig. 7A) (
      • Janve V.S.
      • Hernandez C.C.
      • Verdier K.M.
      • Hu N.
      • Macdonald R.L.
      Epileptic encephalopathy de novo GABRB mutations impair GABAA receptor function.
      ,
      • Laverty D.
      • Desai R.
      • Uchański T.
      • Masiulis S.
      • Stec W.J.
      • Malinauskas T.
      • Zivanov J.
      • Pardon E.
      • Steyaert J.
      • Miller K.W.
      • Aricescu A.R.
      Cryo-EM structure of the human α1β3γ2 GABAA receptor in a lipid bilayer.
      ). The cryo-EM structure indicated that the local environment surrounding the β3120 residues is identical, making contacts with the adjacent α1M141 and α1P142 residues regardless of the subunit the mutation is in (Fig. 7A). Similarly, the β3T157 residue is located within a β-sheet within the interior of the β3 subunit, where the amino acid side chains interact entirely with residues within the β3 subunit, and the local environment that surrounds the amino acid residues is identical (Fig. 7B). Therefore, the similarity of the functional changes caused by mutations at these residues suggests that both of the local environments are identical and that the conformational changes at these initial stages of the activation pathway are symmetrical.
      Figure thumbnail gr7
      Figure 7A–C, enlarged view of the α1β3γ2 cryo-EM structure (PDB code 6HUP) (
      • Laverty D.
      • Desai R.
      • Uchański T.
      • Masiulis S.
      • Stec W.J.
      • Malinauskas T.
      • Zivanov J.
      • Pardon E.
      • Steyaert J.
      • Miller K.W.
      • Aricescu A.R.
      Cryo-EM structure of the human α1β3γ2 GABAA receptor in a lipid bilayer.
      ) showing the β3 mutant residues β3D120 (A), β3T157 (B), and β3Y302 (C) in the two different subunit locations of the second β3 subunit (left) and fourth β3 subunit (right). The subunits are colored with the first γ2 subunit in green, the second β3 subunit in maroon, the third α1 subunit in light blue, the fourth β3 subunit in red, and the fifth α1 subunit in dark blue and the GABA molecule in blue, red, and green. Residues from the adjacent α1 or γ2 subunits are indicated. At the β1D120 (A), β1T1577 (B), and β1Y302 (C) residues, the interacting partners are identical residues either on the adjacent subunit or within the β3 subunit itself.
      The β3Y302 residue is located in the M2-M3 loop, a key motif in the coupling of ligand binding to channel gating that moves considerably during the channel activation process. Although there are differences in the activation processes of single β3Y302C mutant receptors, depending on the subunit that the mutation is located in, the local environment surrounding the β3Y302 residue is similar at the two subunits (Fig. 7C). There are slightly different poses for the β3Y302 residue in each of the different locations, but it is known that the M2-M3 region alters conformation during the gating process (
      • Absalom N.L.
      • Lewis T.M.
      • Schofield P.R.
      Mechanisms of channel gating of the ligand-gated ion channel superfamily inferred from protein structure.
      ), and it may be differences in the conformational changes that cause subtle differences in the effect of the mutation when it is present at the different locations.
      The cryo-EM structure of the α1β1γ2 GABAA receptor identified asymmetry within the transmembrane regions of the receptor (
      • Phulera S.
      • Zhu H.
      • Yu J.
      • Claxton D.P.
      • Yoder N.
      • Yoshioka C.
      • Gouaux E.
      Cryo-EM structure of the benzodiazepine-sensitive α1β1γ2S tri-heteromeric GABAA receptor in complex with GABA.
      ). Whereas the distance between subunits was the same, the angles in the pentamer substantially differed along with the tilt of the transmembrane helices when compared with the β3 subunit. This asymmetry of key secondary structures may contribute to the β3Y302C mutation having different effects when introduced in different regions within the transmembrane regions of GABAA receptors, where the transitional or pre-activated states are subtly different at the coupling regions, depending on where the first GABA molecule is bound.
      The β3S254F mutation caused markedly different effects, depending on the subunit where the mutation was introduced. The β3S254 residue is located deep within the transmembrane region of the M1 helix, where it can interact with other transmembrane helices, including the M2 helix of the adjacent subunit.
      The M1 helix of one β3 subunit is adjacent to the M3 of an α1 subunit, whereas the M1 helix of the second β3 subunit is adjacent to the M3 of the γ2 subunit. This difference may underlie the different functional changes when the mutation is introduced at different locations. The side chains of the β3S254 residues themselves make intrasubunit interactions within the β3 subunit, and the introduction of the phenylalanine residue is unlikely to make different side chain interactions when introduced at the two locations (Fig. 8). However, the bulky side chain is likely to cause structural rearrangements when occupying a larger volume, where the backbone of the M1 helix backs onto the M3 helix of either the γ2 or α1 subunit. At a critical part of the M3 helix, the two subunits have different amino acid sequences, and the M3 helices are in markedly different conformations in different cryo-EM structures.
      Figure thumbnail gr8
      Figure 8A–C, enlarged view of the GABA and diazepam-bound α1β3γ2 cryo-EM structure (PDB code 6HUP) (A), apo-α1β3γ2 (PDB 6i53) (B), and α1β1γ2 (PDB code 6DW0) (C) (
      • Phulera S.
      • Zhu H.
      • Yu J.
      • Claxton D.P.
      • Yoder N.
      • Yoshioka C.
      • Gouaux E.
      Cryo-EM structure of the benzodiazepine-sensitive α1β1γ2S tri-heteromeric GABAA receptor in complex with GABA.
      ,
      • Laverty D.
      • Desai R.
      • Uchański T.
      • Masiulis S.
      • Stec W.J.
      • Malinauskas T.
      • Zivanov J.
      • Pardon E.
      • Steyaert J.
      • Miller K.W.
      • Aricescu A.R.
      Cryo-EM structure of the human α1β3γ2 GABAA receptor in a lipid bilayer.
      ) showing the β3S254 or β1S254 mutant residues in the two different subunit locations of the second β3 subunit (left) and fourth β3 subunit (right). The subunits are colored with the first γ2 subunit in green, the second β3 subunit in maroon, the third α1 subunit in light blue, the fourth β3 subunit in red, and the fifth α1 subunit in dark blue. Residues from adjacent α1 or γ2 subunits are indicated. Although the interacting partners of the Ser-254 residue are within the β subunit, the increased volume of the phenylalanine residue that substitutes for the serine at position 254 will cause the helix to occupy the space closer to the M3 helix of the adjacent subunit. When in the second position in the apo or α1β1γ2 structures, the M3 helix of the γ2 helix is kinked rather than parallel to the M1 helix of the β subunit. The residues of the M1 and M3 that face each other are the β1M253, β1L256, and β1I259 residues for both subunits; the γ2V341, γ2I344, γ2F345, and γ2S348 residues of the first subunit (left); and the α1Y321, α1F323, and α1Y325 residues of the third subunit (right). At the locations of all of these mutations, the sequence between β1 and β3 is identical.
      The M3 helix of the α1 subunit is parallel to the M1 helix of the β3 subunit in the apo and GABA and diazepam-bound α1β3γ2 structures (Fig. 8, A and B). The M3 helix of the γ2 subunit is similarly parallel in the GABA and diazepam-bound α1β3γ2 structure, but in the apo-structure, the M3 helix of the γ2 subunit is tilted, angling away from the M1 helix of the β3 subunit at the extracellular end. Further, the γ2F343 residue has an altered side-chain conformation. A similar change in the M3 conformation of the M3 helix is also seen in the α1β3γ2 cryo-EM structure (
      • Phulera S.
      • Zhu H.
      • Yu J.
      • Claxton D.P.
      • Yoder N.
      • Yoshioka C.
      • Gouaux E.
      Cryo-EM structure of the benzodiazepine-sensitive α1β1γ2S tri-heteromeric GABAA receptor in complex with GABA.
      ) (Fig. 8C). A plausible reason for the different activation properties of the two single β3S254F mutant receptors is that when the residue is mutated adjacent to the γ2 subunit, the subsequent rearrangements introduce twisting or tilting of the α-helices to favor a closed conformation, but when adjacent to an α1 subunit, the interactions stabilize an intermediate or open state.

      Molecular phenotype of the mutations

      There is a wealth of information that suggests that impairment of GABAA receptor–mediated inhibition can lead to seizures. This includes pharmacological evidence, where antagonists of the GABAA receptor, such as bicuculline, induce seizures (
      • Wood J.D.
      The role of γ-aminobutyric acid in the mechanism of seizures.
      ), and genetic evidence, where mutations in GABAA receptor subunits are known to reduce receptor translocation to the cell surface (
      • Kang J.Q.
      • Shen W.
      • Zhou C.
      • Xu D.
      • Macdonald R.L.
      The human epilepsy mutation GABRG2(Q390X) causes chronic subunit accumulation and neurodegeneration.
      ) or impair the activation properties of the receptor (
      • Møller R.S.
      • Larsen L.H.
      • Johannesen K.M.
      • Talvik I.
      • Talvik T.
      • Vaher U.
      • Miranda M.J.
      • Farooq M.
      • Nielsen J.E.
      • Svendsen L.L.
      • Kjelgaard D.B.
      • Linnet K.M.
      • Hao Q.
      • Uldall P.
      • Frangu M.
      • et al.
      Gene panel testing in epileptic encephalopathies and familial epilepsies.
      ,
      • Janve V.S.
      • Hernandez C.C.
      • Verdier K.M.
      • Hu N.
      • Macdonald R.L.
      Epileptic encephalopathy de novo GABRB mutations impair GABAA receptor function.
      ,
      • Baulac S.
      • Huberfeld G.
      • Gourfinkel-An I.
      • Mitropoulou G.
      • Beranger A.
      • Prud'homme J.F.
      • Baulac M.
      • Brice A.
      • Bruzzone R.
      • LeGuern E.
      First genetic evidence of GABAA receptor dysfunction in epilepsy: a mutation in the γ2-subunit gene.
      ,
      • Shen D.
      • Hernandez C.C.
      • Shen W.
      • Hu N.
      • Poduri A.
      • Shiedley B.
      • Rotenberg A.
      • Datta A.N.
      • Leiz S.
      • Patzer S.
      • Boor R.
      • Ramsey K.
      • Goldberg E.
      • Helbig I.
      • Ortiz-Gonzalez X.R.
      • et al.
      De novo GABRG2 mutations associated with epileptic encephalopathies.
      ). In this study, we focused on five distinct mutations, four of which clearly impaired the activation by GABA when only a single copy of the mutation was present. Our approach using the concatenated receptor enabled us to assess the complexity of receptors that contain mutations in one or both β3 subunits.
      For three of the mutations, the resulting receptors followed a predictable pattern. A single copy of the mutation, introduced at either β3 subunit, caused a substantial impairment of the activation of the receptor by GABA. When a second mutation was introduced to the receptor, there was a catastrophic change to the receptor such that the magnitude of currents was either too low to be measured or markedly reduced. Regardless of whether the expression of these receptors is affected by the mutation, we would expect synaptic transmission of GABA-elicited currents to be impaired in the inhibitory pathways within these patients, with receptors containing single mutations mediating reduced neurotransmission and receptors with two copies of the mutation mediating little, if any, chloride current.
      The effects of the β3S254F mutation were very different, having markedly different shifts in the potency of GABA, depending on the location of the mutation. Notably, mutations in the proline residue that precedes the same serine in the β1 subunit also cause the potency of GABA to increase, but the same mutation in an α1 subunit failed to assemble in HEK293 cells (
      • Greenfield Jr, L.J.
      • Zaman S.H.
      • Sutherland M.L.
      • Lummis S.C.
      • Niemeyer M.I.
      • Barnard E.A.
      • Macdonald R.L.
      Mutation of the GABAA receptor M1 transmembrane proline increases GABA affinity and reduces barbiturate enhancement.
      ). Although we found robust activation of the receptors in our oocyte expression system, it is possible that the mutations also impair assembly or trafficking to the cell surface in the mammalian cell. We also cannot rule out that the mutant subunit preferentially arranges itself in the location where the potency of GABA is reduced, leading to impaired GABA activation. Importantly, we have only considered the effect of the mutation on the synaptic α1β3γ2, and the effect of the mutation in other GABA receptor subtypes, including α5β3γ2 or α4β3δ receptors, may contribute to the overall phenotype, including seizures. Indeed, all β3 mutations would be expressed in these subtypes, and as such, the effects of the β3D120N, β3T157M, and β3Y302C on extrasynaptic receptors also need to be considered, particularly when the β3T157M mutation has been reported to have little effect when expressed in the α5β3γ2 subtype (
      • Møller R.S.
      • Larsen L.H.
      • Johannesen K.M.
      • Talvik I.
      • Talvik T.
      • Vaher U.
      • Miranda M.J.
      • Farooq M.
      • Nielsen J.E.
      • Svendsen L.L.
      • Kjelgaard D.B.
      • Linnet K.M.
      • Hao Q.
      • Uldall P.
      • Frangu M.
      • et al.
      Gene panel testing in epileptic encephalopathies and familial epilepsies.
      ). It is also possible that the β3S254F mutation is not truly pathogenic in itself; however, this mutation has been subsequently identified de novo in another patient and is therefore very unlikely not to be pathogenic (
      • Liu J.
      • Tong L.
      • Song S.
      • Niu Y.
      • Li J.
      • Wu X.
      • Zhang J.
      • Zai C.C.
      • Luo F.
      • Wu J.
      • Li H.
      • Wong A.H.C.
      • Sun R.
      • Liu F.
      • Li B.
      Novel and de novo mutations in pediatric refractory epilepsy.
      ).

      Conclusions and future directions

      Genetic epileptic encephalopathies are a devastating group of severe childhood epilepsies that are often resistant to pharmacological treatment and include patients with mutations in genes that encode for the GABAA receptor. Introducing mutations to concatenated receptors demonstrates that the number of mutations within the receptor matter. Typically, the incorporation of one mutation impairs the activation properties of the receptor, reducing the GABA potency, efficacy, or both, and the incorporation of two mutations is often catastrophic to receptor function. However, the mutations are complex, where individual mutations can also increase the potency of GABA, and the subunit location of the mutation can also determine the functional change in the mutation. The resultant molecular phenotype is likely a complex mixture of receptors with a mix of WT receptors, receptors containing a single mutation, and receptors containing two mutations. Furthermore, receptors containing one mutation that have an intermediate effect on the activation process may be a useful target for GABAergic drugs to confer the most benefit for their specific mutation.

      Experimental procedures

      Molecular biology

      Human cDNAs for monomeric α1, β3, and γ2 GABAAR subunits were kind gifts from Saniona A/S (Copenhagen, Denmark). The γ2, β3, and α1 subunits were initially subcloned into five separate in-house vectors. Linker sequences were then added through standard PCRs, where the antisense oligonucleotides caused deletion of the stop codon and in-frame fusion to the AGS linker sequence, and the sense β3 or α1 oligonucleotides caused omission of the respective β3 or α1 signal peptide and in-frame fusion to the AGS linker sequence. The remaining sequences of the sense and antisense oligonucleotides were designed to match the respective WT sequences and included unique restriction sites within each linker region and at the beginning and end of each gene sequence. Standard PCRs with the γ2, β3, or α1 sequences as template were performed using Q5 polymerase (Genesearch, Gold Coast, Australia), and PCR products were cloned into in-house vectors using restriction enzyme digestion and ligation. Correct introduction of linker sequences and fidelity of all coding sequences were verified by double-stranded sequencing. The γ2-β3-α1-β3-α1 concatenated construct was then created by a restriction enzyme digest of the five vectors and ligation of the five subunits with linker sequences and subcloned into an in-house vector. The resulting construct contained the subunits with linker sequences in the order of γ2-(AGS)5-β3-(AGS)5LGS(AGS)3-α1- AGT(AGS)5-β3-(AGS)4ATG(AGS)4-α1. The vector was transformed into Escherichia coli 10-β bacteria for plasmid amplification, and purifications were performed with standard kits (Qiagen, Chadstone, Australia). Mutations were made using the QuikChange II site-directed mutagenesis kit (Agilent Technologies, Mulgrave, Australia) in vectors containing single subunits and flanking linker sequences and then confirmed through dsDNA sequencing. A restriction digest was then performed on subunit DNA containing the mutation and the concatenated construct to remove the appropriate WT subunit and then ligated to introduce the mutant subunit. DNA gel electrophoresis was performed to ensure incorporation of the five subunits. cRNA was produced from linearized cDNA using the mMessage mMachine T7 Transcription kit (Thermo Fisher, Scoresby, Australia) according to the manufacturer's description and stored at −80 °C until use.

      Xenopus surgery and oocyte preparation

      All procedures using Xenopus laevis frogs were approved by the animal ethics committee of the University of Sydney (AEC number 2013/5269) and are in accordance with the National Health and Medical Research Council (NHMRC) of Australia code for the care and use of animals. In brief, a section of ovarian lobe from X. laevis was surgically removed while the frog was under anesthesia induced by tricaine, cut into smaller portions, and digested with 35 mg of collagenase-A diluted in 15 ml of OR2 (82.5 mm NaCl, 5 mm HEPES, 2 mm MgCl2, and 2 mm KCl, pH 7.4) at 18 °C for ∼1 h until the oocytes were fully detached from the follicles and the ovary tissue. Oocytes were then injected with a total of 2 ng of cRNA per cell that encoded concatenated WT or mutant receptors and were incubated for 2–4 days on an oscillator at 18 °C in ND96 solution (96 mm NaCl, 2 mm KCl, 1 mm MgCl2, 1.8 mm CaCl2, 5 mm HEPES, pH 7.4) supplemented with 2.5 mm pyruvate, 0.5 mm theophylline, and 50 μg/ml gentamycin.

      Two-electrode voltage-clamp recording

      Cell currents were recorded using the two-electrode voltage-clamp method as described previously (
      • Chua H.C.
      • Absalom N.L.
      • Hanrahan J.R.
      • Viswas R.
      • Chebib M.
      The direct actions of GABA, 2′-methoxy-6-methylflavone and general anaesthetics at β3γ2L GABAA receptors: evidence for receptors with different subunit stoichiometries.
      ). Briefly, oocytes were continuously superfused at room temperature with ND96 at ∼5 ml/min. Cells were impaled with microelectrodes fashioned from capillary glass (Harvard Apparatus, Holliston, MA) that were prepared with a micropipette puller (Narishige, Tokyo, Japan) and filled with 3 m KCl (0.3–2.0 megaohms) and then voltage-clamped at −60 mV. A semi-automated three-channel oocyte recording system was used, where the application of solution was controlled through programming of a Powerlab 8/36 data acquisition system (ADI Instruments, Sydney, Australia) that switched solutions through a VC-8 eight-channel perfusion system (Warner Instruments LLC, Hamden, CT) and then applied solution to three recording chambers. Currents were recording using a GeneClamp 500B (Axon Instruments, Foster City, CA) or OC-725C amplifier clamp (Warner Instruments) and digitized with a Powerlab 8/36 and LabChart version 8.03 (ADInstruments, Sydney, Australia).
      For clobazam concentration–response curves, a 10 μm concentration was applied as a reference, and the responses were normalized to the mean current of the second two GABA concentrations. For all other experiments, a 3 mm concentration of GABA was applied as a reference three times during the experiment, and for concentration–response curves, peak currents were normalized to the mean current of the second two GABA applications. When estimating the maximal PO, after three consecutive applications of the reference 3 mm GABA solution, the solution containing 10 mm GABA, 1 μm diazepam, and 3 μm etomidate was applied, and peak currents were normalized to the mean current of the second two GABA applications. A washout period of 10–12 min was performed between GABA applications to prevent effects from desensitization. All experiments were performed over a minimum of two different batches of oocytes, and a minimum of 10 individual experiments were performed. The GABA applications were applied in a sequence identical to those shown in the representative data figures. Data were acquired at 1 kHz, and, for the purposes of displaying representative traces, the data were converted to 10 Hz offline through Microsoft excel.

      Data analysis and statistics

      Concentration–response curves were fitted using GraphPad Prism version 7 to a monophasic Hill equation of the form,
      I=Imax([A]nH[A]nH+EC50nH
      (Eq. 1)


      where Imax is the maximum current, EC50 is the concentration that produces the half-maximum response, [A] is the concentration of ligand, and nH is the Hill slope. Individual oocytes where a complete concentration–response curve was taken are recorded as a single n. Responses were normalized to the fitted maximum response of individual concentration–response curves. The EC50 shown is from the fitting of Hill equations to all data, whereas the log EC50, Imax, and nH values are the mean and S.E. derived from fitting curves to individual experiments.
      For clobazam concentration–response curves, the percent modulation of clobazam was derived by the equation.
      Percent modulation=100×IclobazamI10μM GABAI10μM GABA
      (Eq. 2)


      These data were then fitted to the Hill equation, as above, to determine the parameters of the concentration–response curves.
      The estimated Po(max) for individual experiments was derived by the equation.
      Est.P0(max()=correction factor×I3mM GABAI10mM GABA,1μM diazepam,3μM etomidate
      (Eq. 3)


      The correction factor was determined for each mutation to correct for the fact that 3 mm GABA did not always elicit the maximum response to GABA. This was derived from rearranging the Hill equation,
      Correction factor=1+0.003nHEC50nH
      (Eq. 4)


      where the EC50 was in m.
      When comparing the WT and mutation concentration–response curves, all data were transformed to the Est. Po(max).
      For statistical analysis, Est. Po(max) values and parameters derived from concentration–response curves were compared with a one-way ANOVA with Tukey's post hoc test. Significance values of p < 0.05, p < 0.01, and p < 0.001 are shown under “Results.”

      Author contributions

      N. L. A., P. K. A., J. C. A., I. S. M., M. T. B., and M. C. conceptualization, N. L. A. data curation; N. L. A., T. B., and T. J. formal analysis; N. L. A., P. K. A., T. B., T. J., L. L. A., and M. C. investigation; N. L. A., P. K. A., V. W. L., T. B., T. J., and L. L. A. methodology; N. L. A., M. T. B., and M. C. writing-original draft; N. L. A., J. C. A., I. S. M., M. T. B., and M. C. project administration; N. L. A., P. K. A., V. W. L., T. B., J. C. A., I. S. M., M. T. B., and M. C. writing-review and editing; P. K. A., V. W. L., and T. B. resources; J. C. A., I. S. M., M. T. B., and M. C. funding acquisition; M. C. supervision.

      References

        • Howard M.A.
        • Baraban S.C.
        Catastrophic epilepsies of childhood.
        Annu. Rev. Neurosci. 2017; 40 (28772100): 149-166
        • Cossette P.
        • Liu L.
        • Brisebois K.
        • Dong H.
        • Lortie A.
        • Vanasse M.
        • Saint-Hilaire J.M.
        • Carmant L.
        • Verner A.
        • Lu W.Y.
        • Wang Y.T.
        • Rouleau G.A.
        Mutation of GABRA1 in an autosomal dominant form of juvenile myoclonic epilepsy.
        nat. Genet. 2002; 31 (11992121): 184-189
        • Gontika M.P.
        • Konialis C.
        • Pangalos C.
        • Papavasiliou A.
        Novel SCN1A and GABRA1 gene mutations with diverse phenotypic features and the question on the existence of a broader spectrum of Dravet syndrome.
        Child Neurol. Open. 2017; 4 (2329048X17706794, 28540321)
        • Hernandez C.C.
        • Klassen T.L.
        • Jackson L.G.
        • Gurba K.
        • Hu N.
        • Noebels J.L.
        • Macdonald R.L.
        Deleterious rare variants reveal risk for loss of GABAA receptor function in patients with genetic epilepsy and in the general population.
        PLoS One. 2016; 11 (27622563): e0162883
        • Johannesen K.
        • Marini C.
        • Pfeffer S.
        • Møller R.S.
        • Dorn T.
        • Niturad C.E.
        • Gardella E.
        • Weber Y.
        • Søndergård M.
        • Hjalgrim H.
        • Nikanorova M.
        • Becker F.
        • Larsen L.H.
        • Dahl H.A.
        • Maier O.
        • et al.
        Phenotypic spectrum of GABRA1: from generalized epilepsies to severe epileptic encephalopathies.
        Neurology. 2016; 87 (27521439): 1140-1151
        • Kodera H.
        • Ohba C.
        • Kato M.
        • Maeda T.
        • Araki K.
        • Tajima D.
        • Matsuo M.
        • Hino-Fukuyo N.
        • Kohashi K.
        • Ishiyama A.
        • Takeshita S.
        • Motoi H.
        • Kitamura T.
        • Kikuchi A.
        • Tsurusaki Y.
        • et al.
        De novo GABRA1 mutations in Ohtahara and West syndromes.
        Epilepsia. 2016; 57 (26918889): 566-573
        • Lachance-Touchette P.
        • Brown P.
        • Meloche C.
        • Kinirons P.
        • Lapointe L.
        • Lacasse H.
        • Lortie A.
        • Carmant L.
        • Bedford F.
        • Bowie D.
        • Cossette P.
        Novel α1 and γ2 GABAA receptor subunit mutations in families with idiopathic generalized epilepsy.
        Eur. J. Neurosci. 2011; 34 (21714819): 237-249
        • Møller R.S.
        • Larsen L.H.
        • Johannesen K.M.
        • Talvik I.
        • Talvik T.
        • Vaher U.
        • Miranda M.J.
        • Farooq M.
        • Nielsen J.E.
        • Svendsen L.L.
        • Kjelgaard D.B.
        • Linnet K.M.
        • Hao Q.
        • Uldall P.
        • Frangu M.
        • et al.
        Gene panel testing in epileptic encephalopathies and familial epilepsies.
        Mol. Syndromol. 2016; 7 (27781031): 210-219
        • Stosser M.B.
        • Lindy A.S.
        • Butler E.
        • Retterer K.
        • Piccirillo-Stosser C.M.
        • Richard G.
        • McKnight D.A.
        High frequency of mosaic pathogenic variants in genes causing epilepsy-related neurodevelopmental disorders.
        Genet. Med. 2018; 20 (28837158): 403-410
        • Allen A.S.
        • Berkovic S.F.
        • Cossette P.
        • Delanty N.
        • Dlugos D.
        • Eichler E.E.
        • Epstein M.P.
        • Glauser T.
        • Goldstein D.B.
        • Han Y.
        • Heinzen E.L.
        • Hitomi Y.
        • Howell K.B.
        • et al.
        • Epi4K Consortium, Epilepsy Phenome/Genome Project
        De novo mutations in epileptic encephalopathies.
        nature. 2013; 501 (23934111): 217-221
        • Janve V.S.
        • Hernandez C.C.
        • Verdier K.M.
        • Hu N.
        • Macdonald R.L.
        Epileptic encephalopathy de novo GABRB mutations impair GABAA receptor function.
        Ann. Neurol. 2016; 79 (26950270): 806-825
        • Papandreou A.
        • McTague A.
        • Trump N.
        • Ambegaonkar G.
        • Ngoh A.
        • Meyer E.
        • Scott R.H.
        • Kurian M.A.
        GABRB3 mutations: a new and emerging cause of early infantile epileptic encephalopathy.
        Dev. Med. Child. Neurol. 2016; 58 (26645412): 416-420
        • Baulac S.
        • Huberfeld G.
        • Gourfinkel-An I.
        • Mitropoulou G.
        • Beranger A.
        • Prud'homme J.F.
        • Baulac M.
        • Brice A.
        • Bruzzone R.
        • LeGuern E.
        First genetic evidence of GABAA receptor dysfunction in epilepsy: a mutation in the γ2-subunit gene.
        nat. Genet. 2001; 28 (11326274): 46-48
        • Boillot M.
        • Morin-Brureau M.
        • Picard F.
        • Weckhuysen S.
        • Lambrecq V.
        • Minetti C.
        • Striano P.
        • Zara F.
        • Iacomino M.
        • Ishida S.
        • An-Gourfinkel I.
        • Daniau M.
        • Hardies K.
        • Baulac M.
        • Dulac O.
        • et al.
        Novel GABRG2 mutations cause familial febrile seizures.
        Neurol. Genet. 2015; 1 (27066572): e35
        • Shen D.
        • Hernandez C.C.
        • Shen W.
        • Hu N.
        • Poduri A.
        • Shiedley B.
        • Rotenberg A.
        • Datta A.N.
        • Leiz S.
        • Patzer S.
        • Boor R.
        • Ramsey K.
        • Goldberg E.
        • Helbig I.
        • Ortiz-Gonzalez X.R.
        • et al.
        De novo GABRG2 mutations associated with epileptic encephalopathies.
        Brain. 2017; 140 (27864268): 49-67
        • Wallace R.H.
        • Marini C.
        • Petrou S.
        • Harkin L.A.
        • Bowser D.N.
        • Panchal R.G.
        • Williams D.A.
        • Sutherland G.R.
        • Mulley J.C.
        • Scheffer I.E.
        • Berkovic S.F.
        Mutant GABAA receptor γ2-subunit in childhood absence epilepsy and febrile seizures.
        nat. Genet. 2001; 28 (11326275): 49-52
        • Khazipov R.
        GABAergic synchronization in epilepsy.
        Cold Spring Harb. Perspect. Med. 2016; 6 (26747834): a022764
        • Miller P.S.
        • Aricescu A.R.
        Crystal structure of a human GABAA receptor.
        nature. 2014; 512 (24909990): 270-275
        • Olsen R.W.
        GABAA receptor: positive and negative allosteric modulators.
        Neuropharmacology. 2018; 136 (29407219): 10-22
        • Hevers W.
        • Lüddens H.
        The diversity of GABAA receptors: pharmacological and electrophysiological properties of GABAA channel subtypes.
        Mol. Neurobiol. 1998; 18 (9824848): 35-86
        • Farrant M.
        • Nusser Z.
        Variations on an inhibitory theme: phasic and tonic activation of GABAA receptors.
        nat. Rev. Neurosci. 2005; 6 (15738957): 215-229
        • Bouzat C.
        New insights into the structural bases of activation of Cys-loop receptors.
        J. Physiol. Paris. 2012; 106 (21995938): 23-33
        • Ahring P.K.
        • Liao V.W.Y.
        • Balle T.
        Concatenated nicotinic acetylcholine receptors: a gift or a curse?.
        J. Gen. Physiol. 2018; 150 (29382698): 453
        • Baumann S.W.
        • Baur R.
        • Sigel E.
        Forced subunit assembly in alpha1beta2gamma2 GABAA receptors: insight into the absolute arrangement.
        J. Biol. Chem. 2002; 277 (12324466): 46020-46025
        • Sigel E.
        • Kaur K.H.
        • Lüscher B.P.
        • Baur R.
        Use of concatamers to study GABAA receptor architecture and function: application to δ-subunit-containing receptors and possible pitfalls.
        Biochem. Soc. Trans. 2009; 37 (19909272): 1338-1342
        • Baur R.
        • Minier F.
        • Sigel E.
        A GABAA receptor of defined subunit composition and positioning: concatenation of five subunits.
        FEBS Lett. 2006; 580 (16494876): 1616-1620
        • Ramerstorfer J.
        • Furtmüller R.
        • Sarto-Jackson I.
        • Varagic Z.
        • Sieghart W.
        • Ernst M.
        The GABAA receptor α+β-interface: a novel target for subtype selective drugs.
        J. Neurosci. 2011; 31 (21248110): 870-877
        • Che Has A.T.
        • Absalom N.
        • van Nieuwenhuijzen P.S.
        • Clarkson A.N.
        • Ahring P.K.
        • Chebib M.
        Zolpidem is a potent stoichiometry-selective modulator of α1β3 GABAA receptors: evidence of a novel benzodiazepine site in the α1-α1 interface.
        Sci. Rep. 2016; 6 (27346730): 28674
        • Hammer H.
        • Ebert B.
        • Jensen H.S.
        • Jensen A.A.
        Functional characterization of the 1,5-benzodiazepine clobazam and its major active metabolite N-desmethylclobazam at human GABAA receptors expressed in Xenopus laevis oocytes.
        PLoS One. 2015; 10 (25798598): e0120239
        • Shin D.J.
        • Germann A.L.
        • Johnson A.D.
        • Forman S.A.
        • Steinbach J.H.
        • Akk G.
        Propofol is an allosteric agonist with multiple binding sites on concatemeric ternary GABAA receptors.
        Mol. Pharmacol. 2018; 93 (29192122): 178-189
        • Fisher J.L.
        • Macdonald R.L.
        Single channel properties of recombinant GABAA receptors containing γ2 or δ subtypes expressed with α1 and β3 subtypes in mouse L929 cells.
        J. Physiol. 1997; 505 (9423172): 283-297
        • Purohit P.
        • Gupta S.
        • Jadey S.
        • Auerbach A.
        Functional anatomy of an allosteric protein.
        nat. Commun. 2013; 4 (24352193): 2984
        • Kash T.L.
        • Dizon M.J.
        • Trudell J.R.
        • Harrison N.L.
        Charged residues in the β2 subunit involved in GABAA receptor activation.
        J. Biol. Chem. 2004; 279 (14610076): 4887-4893
        • Kash T.L.
        • Jenkins A.
        • Kelley J.C.
        • Trudell J.R.
        • Harrison N.L.
        Coupling of agonist binding to channel gating in the GABAA receptor.
        nature. 2003; 421 (12529644): 272-275
        • Sigel E.
        • Steinmann M.E.
        Structure, function, and modulation of GABAA receptors.
        J. Biol. Chem. 2012; 287 (23038269): 40224-40231
        • Ahring P.K.
        • Bang L.H.
        • Jensen M.L.
        • Strøbaek D.
        • Hartiadi L.Y.
        • Chebib M.
        • Absalom N.
        A pharmacological assessment of agonists and modulators at α4β2γ2 and α4β2δ GABAA receptors: the challenge in comparing apples with oranges.
        Pharmacol. Res. 2016; 111 (27178730): 563-576
        • Lape R.
        • Colquhoun D.
        • Sivilotti L.G.
        On the nature of partial agonism in the nicotinic receptor superfamily.
        nature. 2008; 454 (18633353): 722-727
        • Sigel E.
        • Baur R.
        • Boulineau N.
        • Minier F.
        Impact of subunit positioning on GABAA receptor function.
        Biochem. Soc. Trans. 2006; 34 (17052217): 868-871
        • Gallagher M.J.
        • Song L.
        • Arain F.
        • Macdonald R.L.
        The juvenile myoclonic epilepsy GABAA receptor α1 subunit mutation A322D produces asymmetrical, subunit position-dependent reduction of heterozygous receptor currents and α1 subunit protein expression.
        J. Neurosci. 2004; 24 (15201329): 5570-5578
        • Phulera S.
        • Zhu H.
        • Yu J.
        • Claxton D.P.
        • Yoder N.
        • Yoshioka C.
        • Gouaux E.
        Cryo-EM structure of the benzodiazepine-sensitive α1β1γ2S tri-heteromeric GABAA receptor in complex with GABA.
        Elife. 2018; 7 (30044221): e39383
        • Laverty D.
        • Desai R.
        • Uchański T.
        • Masiulis S.
        • Stec W.J.
        • Malinauskas T.
        • Zivanov J.
        • Pardon E.
        • Steyaert J.
        • Miller K.W.
        • Aricescu A.R.
        Cryo-EM structure of the human α1β3γ2 GABAA receptor in a lipid bilayer.
        nature. 2019; 565 (30602789): 516-520
        • Absalom N.L.
        • Lewis T.M.
        • Schofield P.R.
        Mechanisms of channel gating of the ligand-gated ion channel superfamily inferred from protein structure.
        Exp. Physiol. 2004; 89 (15123543): 145-153
        • Wood J.D.
        The role of γ-aminobutyric acid in the mechanism of seizures.
        Prog. Neurobiol. 1975; 5 (830081): 77-95
        • Kang J.Q.
        • Shen W.
        • Zhou C.
        • Xu D.
        • Macdonald R.L.
        The human epilepsy mutation GABRG2(Q390X) causes chronic subunit accumulation and neurodegeneration.
        nat. Neurosci. 2015; 18 (26005849): 988-996
        • Greenfield Jr, L.J.
        • Zaman S.H.
        • Sutherland M.L.
        • Lummis S.C.
        • Niemeyer M.I.
        • Barnard E.A.
        • Macdonald R.L.
        Mutation of the GABAA receptor M1 transmembrane proline increases GABA affinity and reduces barbiturate enhancement.
        Neuropharmacology. 2002; 42 (11955521): 502-521
        • Liu J.
        • Tong L.
        • Song S.
        • Niu Y.
        • Li J.
        • Wu X.
        • Zhang J.
        • Zai C.C.
        • Luo F.
        • Wu J.
        • Li H.
        • Wong A.H.C.
        • Sun R.
        • Liu F.
        • Li B.
        Novel and de novo mutations in pediatric refractory epilepsy.
        Mol. Brain. 2018; 11 (30185235): 48
        • Chua H.C.
        • Absalom N.L.
        • Hanrahan J.R.
        • Viswas R.
        • Chebib M.
        The direct actions of GABA, 2′-methoxy-6-methylflavone and general anaesthetics at β3γ2L GABAA receptors: evidence for receptors with different subunit stoichiometries.
        PLoS One. 2015; 10 (26496640): e0141359