Structure-function studies of GABA-C receptor ligands

Abstract

Throughout the central nervous system (CNS), the Cys-loop superfamily of ligand-gated ion channels {LGICs), including nicotinic acetylcholine, serotonin type-3A, strychnine-sensitive glycine and y-aminobutyric acid A/C receptors, play important roles in synaptic transmission by converting chemical signals into electric signals. Designing potent and subtype-selective ligands with therapeutic value requires knowledge about how ligands interact with their binding sites. y-Aminobutyric acid (GABA) is the predominant inhibitory neurotransmitter in the mammalian CNS and its binding modes at GABA receptors have not been fully elucidated. GABAc receptors consist of p subunits (p1-p3) and they are known to form homomeric receptors. The five subunits are arranged around a central chloride selective ion channel pore. Each subunit contains a large extracellular N-terminal domain, four transmembrane domains {Ml-M4) of which the second (M2) lines the channel pore and a large M3-M4 intracellular loop. The orthosteric binding site is located at the interface between two subunits in the N-terminal domain and the key residues for ligand binding are found at the five discontinuous loops (A-E). This thesis describes how ligand binding and receptor gating are closely related and explores the effect of receptor conformational changes upon ligand binding. A series of point mutations in the N-terminal domain of the GABAc p1 receptor were created and expressed in Xenopus oocytes. The mutant receptors were then examined using a range of pharmacological tools to probe function which was measured using the two-electrode voltage clamp method. The GABA binding mode was explored at GABA receptors using the enantiomers of 3-fluoro-y-aminobutyric acid (3F-GABA) and stereoisomers of 2,3-difluoro-4-aminobutyric acids as conformational probes. Both enantiomers of 3F­GABA were full agonists, with the R-3F-GABA being approximately 3-fold more potent than 5-3F-GABA at GABAc receptors. In contrast, both enantiomers were partial agonists with similar efficacy and potency at GABAA receptors. These results suggest a different GABA binding mode at GABAc receptors to that found in the related but pharmacologically distinct GABAA receptors. The effect of the different stereoisomers of 2,3-difluoro-4-aminobutyric acids were also examined at GABAA, GABA8 and GABAc receptors. In the study, two enantiomeric GABAc receptor ligands were identified, one of which is an agonist (25,35-2,3-difluoro-4-aminobutyric acid) while the other is an antagonist (2R,3R-2,3-difluoro-4-aminobutyric acid). 4-Amino-3-hydroxybutanoic acid (GABOB) is an endogenous ligand found in the CNS in mammals and a metabolite of GABA. Homology modeling of the GABAc Pi receptor revealed a potential hydrogen (H-bond) interaction between the hydroxyl group of GABOB and threonine 244 (T244) located on loop C of the ligand binding site. Using site-directed mutagenesis, the effect of mutating T244 on the efficacy and pharmacology of GABOB and various ligands were examined. It was found that mutating T244 to amino acids that lacked a hydroxyl group in the side chain produced GABA insensitive receptors. Only by mutating PiT244 to serine (PiT2445) produced a GABA responsive receptor, albeit 39-fold less sensitive to GABA than Pi wild-type. It was also found that this mutation also changed the activity of GABAc receptor partial agonists, muscimol and imidazole-4-acetic acid (I4AA). At the concentrations tested, both muscimol and I4AA antagonized the currents produced by GABA at PiT2445 mutant receptors (Muscimol: PiWild-type, EC50 = 1.4 µM; PiT2445, IC50 = 32.8 µM. I4AA: Pi wild-type, EC50 = 8.6 µM; PiT2445, IC50 = 21.4 µM). This indicates that T244 is predominantly involved in channel gating. R-(-)-GABOB and 5-(+)-GABOB are full agonists at Pi wild-type receptors. In contrast, R-(-)-GABOB was a weak partial agonist at PiT2445 (lmM activates 26 % of the current produced by GABA ECso versus Pi wild-type, EC50 = 19 µM; lmax 100%), and 5-(+)-GABOB was a competitive antagonist at PiT2445 receptors (Pi wild-type, EC50 = 45 µM versus PiT2445, IC50 = 417.4 µM, Ks = 204 µM). This highlights that the interaction of GABOB with T244 is enantioselective. In contrast, the potencies of a range of antagonists tested, 3-aminopropyl(methyl)phosphinic acid (3-APMPA), 3-aminopropylphosphonic acid (3-APA), 5- and R-(3-amino-2-hydroxypropyl)methylphosphinic acid (5-(-)-CGP44532 and R-(+)-CGP44533), were not altered. This suggests that T244 is not critical for antagonist binding. Receptor gating is dynamic and this study highlights the role of loop C in agonist-evoked receptor activation, coupling agonist binding to channel gating. Ligands acting on receptors are considered to induce a conformational change within the ligand-binding site by interacting with specific amino acids. In this study, tyrosine 102 (Y102) located in the GABA binding site of the Pi subunit of the GABAc receptor was mutated to alanine (piY102A), serine (piY102S) and cysteine (piY102C) to assess the role of this amino acid plays on the action of 12 known and 2 novel antagonists. Of the mutated receptors, piY102S was constitutively active providing an opportunity to assess the activity of the antagonists on Pi receptors with a proportion of receptors existing in the open conformational state compared to those existing predominantly in the closed conformational state (pi wild-type, PiY102C and PiY102A). It was found that the majority of antagonists studied were able to inhibit the constitutive activity displayed by PiY1025, thus displaying inverse agonist activity. The exception was (±)-4-aminocyclopent-1-enecarboxamide ((±)-4-ACPAM) not exhibiting any inverse agonist activity, but acting explicitly on the closed conformational state of Pi receptors. It was also found that GABA antagonists were more potent at the closed compared to the open conformational states of Pi receptors suggesting that they may act by stabilizing the closed conformational state and thus reducing activation by agonists. Furthermore, of the antagonists tested, Y102 was found to have the greatest influence on the antagonist activity of gabazine (SR-95531) and its analogue (SR-95813). Our GABAc Pi receptor homology model identified a novel cavity, which extended beyond the GABA binding site. The model predicted phenylalanine 124(F124), one of the residues lining the cavity, was pointing towards the orthosteric binding site. In this study, F124 was mutated to various amino acids and only a modest effect on receptor pharmacology was observed. However, the mutations had a significant effect on the channel deactivation rate ('toeactivation)- This finding suggests that F124 may play a role in channel gating or stabilizing the open conformation of the receptor. Designing potent selective agents are the key for the further understanding of the physiological roles of GABAc receptors. Gabazine (SR-95531) is a potent GABAA receptor competitive antagonist. In this study, a series of novel gabazine analogues were tested at GABAA and GABAc receptors. Of the compounds studied, (p)-methoxy analogue without the butyric acid side-chain was 20-fold more potent at GABAc over GABAA receptors. As there was no butyric acid side chain, it is suggested that the carboxylic acid is not important for gabazine activity at this receptor. Establishing the structure-activity relationship based on this analogue will facilitate the development of selective GABAc receptor antagonists with possible physiological effects including memory-enhancement. Overall, our studies describe agonist and GABAc receptor antagonist induced conformational changes within the ligand binding site. Our findings also highlight the dynamic nature of receptor gating, initiated by ligand binding at a site physically distinct from the ion channel.