Characterization of the Effects of Charged Residues in the Intracellular Loop on Ion Permeation in α1 Glycine Receptor Channels

Abstract

The Cys loop receptor channels mediate fast synaptic transmission in the nervous system. The M2-demarcated transmembrane pore is an important determinant of their ion permeation properties. Portals within the intracellular domain are also part of the permeation pathway in cationic Cys loop receptors, with charged residues in a helical MA stretch partially lining these openings profoundly affecting channel conductance. It is unknown whether analogous portals contribute to the permeation pathway in anionic Cys loop receptors. We therefore investigated the influence of charged residues within the proposed MA stretch on functional properties of the homomeric glycine α1 receptor. Up to eight basic residues in the MA stretch were concurrently mutated to a negatively charged glutamate, and wild-type and mutant subunits were expressed in HEK-293 cells. Mutation of all eight residues produced a non-functional receptor. The greatest reduction in conductance at negative membrane potentials (from 92.2 ± 2.8 to 60.0 ± 2.2 picosiemens) was observed with glutamate present at the 377, 378, 385, and 386 positions (the 4E subunit). Inclusion of additional glutamate residues within this subunit did not decrease conductance further. Neutralizing these residues (the 4A subunit) caused a modest decrease in conductance (80.5 ± 2.3 picosiemens). Outward conductance at positive potentials was not markedly affected. Anion to cation selectivity and concentration-response relationships were unaffected by the 4A or 4E mutations. Our results identify basic residues affecting conductance in the glycine receptor, suggesting that portals are part of the extended permeation pathway but that the M2-demarcated channel pore is the dominant determinant of permeation properties in glycine receptors. The Cys loop receptor channels mediate fast synaptic transmission in the nervous system. The M2-demarcated transmembrane pore is an important determinant of their ion permeation properties. Portals within the intracellular domain are also part of the permeation pathway in cationic Cys loop receptors, with charged residues in a helical MA stretch partially lining these openings profoundly affecting channel conductance. It is unknown whether analogous portals contribute to the permeation pathway in anionic Cys loop receptors. We therefore investigated the influence of charged residues within the proposed MA stretch on functional properties of the homomeric glycine α1 receptor. Up to eight basic residues in the MA stretch were concurrently mutated to a negatively charged glutamate, and wild-type and mutant subunits were expressed in HEK-293 cells. Mutation of all eight residues produced a non-functional receptor. The greatest reduction in conductance at negative membrane potentials (from 92.2 ± 2.8 to 60.0 ± 2.2 picosiemens) was observed with glutamate present at the 377, 378, 385, and 386 positions (the 4E subunit). Inclusion of additional glutamate residues within this subunit did not decrease conductance further. Neutralizing these residues (the 4A subunit) caused a modest decrease in conductance (80.5 ± 2.3 picosiemens). Outward conductance at positive potentials was not markedly affected. Anion to cation selectivity and concentration-response relationships were unaffected by the 4A or 4E mutations. Our results identify basic residues affecting conductance in the glycine receptor, suggesting that portals are part of the extended permeation pathway but that the M2-demarcated channel pore is the dominant determinant of permeation properties in glycine receptors. Rapid and coordinated communication between neurons at synapses in the central nervous system is critical for the maintenance of normal brain function. Key to this process is the Cys loop superfamily of ligand-gated ion channels (LGICs), 4The abbreviations used are: LGICs, ligand-gated ion channels; ACh, acetylcholine; 5-HT3, 5-hydroxytryptamine type 3; GABAA, γ-aminobutyric acid type A; HEK, human embryonic kidney; MA, membrane associated; WT, wild-type; GHK, Goldman-Hodgkin-Katz; pS, picosiemens. which includes the cation-selective nicotinic acetylcholine (ACh) and 5-hydroxytryptamine type 3 (5-HT3) and the anion-selective glycine and γ-aminobutyric acid type A (GABAA) receptors. LGICs convert the chemical messages conveyed by neurotransmitters into excitatory or inhibitory electrical signals via the selective conduction of ions. The polarity of the most permeant ion (ion charge selectivity) and the rate at which the ions permeate (the single channel conductance, γ) are major determinants of how this synaptic signal affects neuronal excitability. Cys loop LGICs are composed of five protein subunits arranged around a central ion-conducting channel pore. Each of the five subunits consists of a large extracellular domain harboring the neurotransmitter binding site, four membrane-spanning domains (M1–M4) and, in mammalian LGICs at least, a large intracellular loop connecting M3–M4. The electron microscopic images of the Torpedo marmorata nicotinic ACh receptor (1Miyazawa A. Fujiyoshi Y. Stowell M. Unwin N. J. Mol. Biol.. 1999; 288: 765-786Google Scholar, 2Miyazawa A. Fujiyoshi Y. Unwin N. Nature.. 2003; 423: 949-955Google Scholar, 3Unwin N. J. Mol. Biol.. 2005; 346: 967-989Google Scholar) provide our most complete picture of the structure of the ion permeation pathway of mammalian LGICs. Ions entering the wide external vestibule are funneled through a narrow transmembrane pore lined by residues within the α-helical second transmembrane (M2) domain. This M2-demarcated channel pore extends into an intracellular vestibule perforated by five openings, or portals, with a narrow diameter (8Å in the closed Torpedo nicotinic ACh receptor (3Unwin N. J. Mol. Biol.. 2005; 346: 967-989Google Scholar)), comparable with that of the open M2 pore. Such portals provide the only pathway for the passage of ions into the cytoplasm and are lined partially by residues from an amphipathic helix within the intracellular M3–M4 loop known as the membrane-associated helix (MA stretch). Traditionally, the determinants of LGIC ion conductance and selectivity were thought to be limited to the M2 domain that lines the integral channel pore and immediate flanking sequences (reviewed in Refs. 4Keramidas A. Moorhouse A.J. Schofield P.R. Barry P.H. Prog. Biophys. Mol. Biol.. 2004; 86: 161-204Google Scholar and 5Jensen M.L. Schousboe A. Ahring P.K. J. Neurochem.. 2005; 92: 217-225Google Scholar). Cumulating recent evidence, however, for cation-selective LGICs also indicates a key contribution from the M3–M4 loop and/or MA stretch residues in cation conductance (6Kelley S.P. Dunlop J.I. Kirkness E.F. Lambert J.J. Peters J.A. Nature.. 2003; 424: 321-324Google Scholar, 7Deeb T.Z. Carland J.E. Cooper M.A. Livesey M.R. Lambert J.J. Peters J.A. Hales T.G. J. Biol. Chem.. 2007; 282: 6172-6182Google Scholar, 8Hales T.G. Dunlop J.I. Deeb T.Z. Carland J.E. Kelley S.P. Lambert J.J. Peters J.A. J. Biol. Chem.. 2006; 281: 8062-8071Google Scholar, 9Jansen M. Bali M. Akabas M.H. J. Gen. Physiol.. 2008; 131: 137-146Google Scholar, 10Gee V.J. Cooper S.T. Gibb A.J. Millar N.S. Br. J. Pharmacol.. 2007; 152: 501-512Google Scholar, 11Reeves D.C. Jansen M. Bali M. Lemster T. Akabas M.H. J. Neurosci.. 2005; 25: 9358-9366Google Scholar). The homomeric 5-HT3A receptor, for example, has a very low γ (0.4–1.0 pS; (12Brown A.M. Hope A.G. Lambert J.J. Peters J.A. J. Physiol. (London).. 1998; 507: 653-665Google Scholar)) relative to most other LGICs. Co-expression of 5-HT3A and 5-HT3B subunits produces an increased γ, and a series of 5-HT3A/3B receptor chimeras and mutant 5-HT3A receptors revealed that the major determinants of homomeric 5-HT3A receptor sub-pS γ were located within the MA stretch. In particular, replacement of three positively charged arginine residues (Arg-432, Arg-436, and Arg-440 in the human 5-HT3A subunit sequence) within the MA stretch by their human 5-HT3B counterparts (QDA, respectively) increased γ to about 35–40 pS, with R436D exerting the greatest influence (6Kelley S.P. Dunlop J.I. Kirkness E.F. Lambert J.J. Peters J.A. Nature.. 2003; 424: 321-324Google Scholar, 7Deeb T.Z. Carland J.E. Cooper M.A. Livesey M.R. Lambert J.J. Peters J.A. Hales T.G. J. Biol. Chem.. 2007; 282: 6172-6182Google Scholar, 8Hales T.G. Dunlop J.I. Deeb T.Z. Carland J.E. Kelley S.P. Lambert J.J. Peters J.A. J. Biol. Chem.. 2006; 281: 8062-8071Google Scholar, 11Reeves D.C. Jansen M. Bali M. Lemster T. Akabas M.H. J. Neurosci.. 2005; 25: 9358-9366Google Scholar). The Arg-436 residue is also an important determinant of the relative Ca2+ permeability of the 5-HT3A receptor (13Livesey M.R. Cooper M.A. Deeb T.Z. Carland J.E. Kozuska J. Hales T.G. Lambert J.J. Peters J.A. J. Biol. Chem.. 2008; 283: 19301-19313Google Scholar). Complete deletion of the 115-residue 5-HT3A M3–M4 loop (and replacement with a 7-residue loop from a bacterial LGIC homologue) produced no further increase in γ, suggesting that residues within the MA stretch were solely responsible for the low γ of the 5-HT3A receptor (9Jansen M. Bali M. Akabas M.H. J. Gen. Physiol.. 2008; 131: 137-146Google Scholar). Subsequent mutagenesis work indicates the impact of the M3–M4 loop region on γ is a conserved feature of the ion conduction properties of nicotinic ACh receptors (8Hales T.G. Dunlop J.I. Deeb T.Z. Carland J.E. Kelley S.P. Lambert J.J. Peters J.A. J. Biol. Chem.. 2006; 281: 8062-8071Google Scholar, 10Gee V.J. Cooper S.T. Gibb A.J. Millar N.S. Br. J. Pharmacol.. 2007; 152: 501-512Google Scholar). Glutamate to arginine mutations to invert residue charge at analogous positions in the α4 and β2 nicotinic ACh subunit MA stretch can reduce the γ of the α4β2 nicotinic ACh receptor from about 30 to 12 pS (8Hales T.G. Dunlop J.I. Deeb T.Z. Carland J.E. Kelley S.P. Lambert J.J. Peters J.A. J. Biol. Chem.. 2006; 281: 8062-8071Google Scholar). The α7 nicotinic ACh receptor subunit M3–M4 loop also presents a high conductance phenotype and only neutral or negatively charged residues within homologous regions of its MA stretch (10Gee V.J. Cooper S.T. Gibb A.J. Millar N.S. Br. J. Pharmacol.. 2007; 152: 501-512Google Scholar, 14Peters J.A. Hales T.G. Lambert J.J. Trends Pharmacol. Sci.. 2005; 26: 587-594Google Scholar). Thus, critical residues in the MA stretch lining the portals and affecting conductance appear to be conserved among mammalian cationic Cys loop LGICs. Whether such a structural determinant of conductance extends to the anionic LGICs is unknown. Although the α1 glycine receptor subunit contains a large number of basic residues in the region of the intracellular M3–M4 loop putatively corresponding to the cationic MA stretch (3Unwin N. J. Mol. Biol.. 2005; 346: 967-989Google Scholar), earlier studies of anionic Cys loop LGICs suggested that these receptors lacked the MA stretch (15Finer-Moore J. Stroud R.M. Proc. Natl. Acad. Sci. U. S. A.. 1984; 81: 155-159Google Scholar, 16Grenningloh G. Rienitz A. Schmitt B. Methfessel C. Zensen M. Beyreuther K. Gundelfinger E.D. Betz H. Nature.. 1987; 328: 215-220Google Scholar, 17Grenningloh G. Gundelfinger E. Schmitt B. Betz H. Darlison M. Barnard E.A. Schofield P.R. Seeburg P.H. Nature.. 1987; 330: 25-26Google Scholar). Nevertheless, the high conductance phenotype of the α1 glycine receptor homomers (∼90 pS) is consistent with some of these positively charged basic residues framing putative intracellular portals and supporting high anion flux rates. In this study, we directly investigated the influence of charged residues within the proposed MA stretch on functional properties of the α1 glycine receptor homomeric channels, a high conductance anionic Cys loop LGIC with well characterized ion permeation properties (4Keramidas A. Moorhouse A.J. Schofield P.R. Barry P.H. Prog. Biophys. Mol. Biol.. 2004; 86: 161-204Google Scholar, 18Barry P.H. Lynch J.W. IEEE Trans. Nanobioscience.. 2005; 4: 70-80Google Scholar). Our results identify particular sets of basic residues affecting conductance, suggesting that the portals are indeed features of the extended glycine receptor permeation pathway. However, the modest magnitude of their effect on conductance and the lack of effect on ion charge selectivity suggest that the M2-demarcated channel pore remains the dominant determinant of α1 glycine receptor permeation properties, where the strongest interactions with permeant ions occur. Molecular Biology and Transient Transfection of HEK-293 Cells—cDNAs encoding human wild type (WT) and mutant α1 glycine receptor subunits were cloned into pcDNA1.1. Point mutations were introduced using standard molecular biological techniques, and all constructs were sequenced to confirm fidelity. Transient transfection of HEK-293 cells with cDNA was performed using polyethylenimine (jetPEI; Poly-plus transfection). All WT and mutant subunits were expressed as homomeric receptors. HEK-293 cells were co-transfected with the expression marker, CD4 surface antigen, and CD4 antibody-coated were used to identify cells were used for were and in a of and were at in an of relative were used to functional glycine receptor expression and to the concentration-response properties and anion to cation selectivity of the WT and 4A and 4E mutant glycine receptors for a of the The was used to single glycine receptor channel and to were performed at a of ± All were using and using a were to of for and for The of used for were also with were with or and an or were and with a or at and were and at and All were the of a for the series from between and was by of positively charged residues in the of α1 glycine receptors conductance, ± ± ± ± ± ± ± ± in a of standard external and with The standard used to the and with standard to a of using the of P.H. J. Scholar), which was for γ standard were used for all of γ and for the concentration-response relative selectivity a number of were The external and The and the external was with an of the external was with an of external were The of all these was to with a known of The potentials from these three were and for the and and were for permeability was in the and via a system. glycine and the external were using a to the from to an using open of of from in which were for of conductance using or were at or for of such additional did not the of but did to were from of using a of and a of were to the using a with to The closed was from the open The number of to the open was increased was no further in the conductance were from the γ in which is the of single channel is the membrane and the which was as for the standard with in to glycine were as the between the and the to glycine from in which the of was and that and a of the were for further concentration-response The between and glycine in was to the to the the a of the and the were for glycine receptor as of the concentration-response in the from were to the and these were also to of relationships in the and external were by in from to between and for a of The was for the of the were and glycine and from in the of glycine all the were to or and the was directly from the The potentials in were the external and the were with the to the ion charge selectivity where is the at which the is is the is in is is the permeability of the and is the of the ion in the extracellular or intracellular The was for and were used for all and by and of at Scholar). are as the ± sets were using of with a was to be within the MA for stretch residues of the cation-selective 5-HT3A and α4β2 nicotinic ACh receptors to γ by and T.Z. Carland J.E. Cooper M.A. Livesey M.R. Lambert J.J. Peters J.A. Hales T.G. J. Biol. Chem.. 2007; 282: 6172-6182Google Scholar, 8Hales T.G. Dunlop J.I. Deeb T.Z. Carland J.E. Kelley S.P. Lambert J.J. Peters J.A. J. Biol. Chem.. 2006; 281: 8062-8071Google Scholar). In particular, a series of positively charged arginine residues as key determinants of the γ of 5-HT3A receptors (6Kelley S.P. Dunlop J.I. Kirkness E.F. Lambert J.J. Peters J.A. Nature.. 2003; 424: 321-324Google Scholar). 5-HT3A receptors, α1 glycine receptors a γ (∼90 of the proposed MA stretch of the α1 glycine receptor subunit with that of the 5-HT3A subunit and other subunits revealed that this region contains a number of positively charged residues We therefore investigated the influence of positively charged residues on the anion flux of α1 glycine receptors. We mutated the eight positively charged or arginine residues within the putative MA stretch to negatively charged glutamate We concurrently from between and to all of the or arginine residues to glutamate, on the region with the series of positively charged arginine residues to cation flux of 5-HT3A receptors (6Kelley S.P. Dunlop J.I. Kirkness E.F. Lambert J.J. Peters J.A. Nature.. 2003; 424: 321-324Google Scholar, 7Deeb T.Z. Carland J.E. Cooper M.A. Livesey M.R. Lambert J.J. Peters J.A. Hales T.G. J. Biol. Chem.. 2007; 282: 6172-6182Google Scholar) The mutations and their are in and in subunits were expressed in HEK-293 and were from membrane The dominant conductance from to are were observed in WT and mutant receptors, but the dominant were also of the of glycine to in at for glycine receptor with to to mutations the 4E to Mutation of all positively charged residues to glutamate the subunit) in a non-functional receptor This that the replacement of by glutamate is not or the effect of all eight mutations is not It is that an arginine residue is conserved at the homologous all glycine and most receptor The of glutamate residues into the proposed MA stretch of the α1 glycine receptor in a γ with from pS to between and pS, as in The decrease in γ pS) for sets of mutations or introduced Glutamate mutations to of positively charged residues within the putative MA stretch the 4E were the mutations that in this γ, and we further on these residues to further studies on ion permeation properties of the mutant 4E receptors, we their concentration-response relationships using Point mutations to regions within the MA stretch of the nicotinic ACh receptor to channel K. M. A. K. A.G. J. Gen. Physiol.. Scholar), and our to for the in glycine receptors whether charged residues in the putative MA stretch glycine was important to conductance properties at of are in for the WT, and 4E receptors, no major in in and the concentration-response for the WT, and 4A glycine receptor constructs with no in the or of glycine at or glycine receptors not from that of WT glycine ± ± ± ± ± ± ± ± ± in a in the MA a of further the γ properties of the 4E mutant glycine receptor relative to the WT receptor by relationships between and and and The WT glycine receptor an as A.J. A. A. Schofield P.R. Barry P.H. J. Gen. Physiol.. Scholar), with γ of pS at and pS at A and We the of by the γ by the γ to a of The 4E mutant glycine receptor a γ at negative membrane potentials pS at but only a very modest reduction in γ, in a of and between of in charge or at these residue we also mutated the residues to a acid consists of a as well as The for this 4A mutant glycine receptor a reduction in γ pS at and γ, in a of and the effect of the 4E mutations was to decrease γ by the charge at these residue the charge in a of the WT to a for the 4E glycine receptor. in the α1 MA on results that positive on critical residues within the putative MA helix influence whether these charged residues with permeant ions to the that can also ion were for the WT, and 4E glycine receptors. of the external results in a of the to positive membrane potentials and a in the to negative membrane in the potentials for all three glycine receptors to positive membrane potentials as the external was all three were anion Anion to cation permeability were in by the as a of the external to the was no in the for WT, and 4A receptors. The were in external and of can influence permeation for example, external can increase the glycine receptor J. E. T. M. A. J. and H. in We the results in external and and were also no in between WT ± A. Moorhouse A.J. Schofield P.R. Barry P.H. Biophys. Scholar), 4A ± and 4E ± glycine receptors. The of the intracellular M3–M4 loop in Cys loop receptor is to be as this region the greatest of among receptor subunits and a complete structural of this loop is in this on located within the loop and the of residues in receptor and of the receptor to the N. Pharmacol. Scholar, J. J. Pharmacol. 2005; Scholar, Scholar). studies of the cation-selective 5-HT3A and α4β2 nicotinic ACh receptors a series of residues at the of the M3–M4 loop that a key in receptor γ (6Kelley S.P. Dunlop J.I. Kirkness E.F. Lambert J.J. Peters J.A. Nature.. 2003; 424: 321-324Google Scholar, 8Hales T.G. Dunlop J.I. Deeb T.Z. Carland J.E. Kelley S.P. Lambert J.J. Peters J.A. J. Biol. Chem.. 2006; 281: 8062-8071Google Scholar). residues part of an thought to partially a series of portals that the transmembrane channel pore (3Unwin N. J. Mol. Biol.. 2005; 346: 967-989Google Scholar). portals an pathway for ion with and of these MA stretch residues an impact on γ in cationic LGICs T.Z. Carland J.E. Cooper M.A. Livesey M.R. Lambert J.J. Peters J.A. Hales T.G. J. Biol. Chem.. 2007; 282: 6172-6182Google Scholar). Whether a intracellular portals by residues in a MA in anionic LGICs is not is important to whether this is indeed a conserved structural and functional feature among the Cys loop LGIC this is for a complete of the molecular determinants of ion the M2 domain has as the determinant of ion permeation properties in the Cys loop LGICs (4Keramidas A. Moorhouse A.J. Schofield P.R. Barry P.H. Prog. Biophys. Mol. Biol.. 2004; 86: 161-204Google Scholar). It was that deletion of the M3–M4 loop of the homomeric receptor (and replacement by the 7-residue bacterial Cys loop LGIC sequence) in no of ion charge the permeability were not (9Jansen M. Bali M. Akabas M.H. J. Gen. Physiol.. 2008; 131: 137-146Google Scholar). The of mutations on ion charge selectivity or the of mutations to the putative MA stretch on conductance not for anionic The investigated the of charged residues in the putative MA stretch at the of the α1 glycine subunit M3–M4 loop on receptor γ, ion charge and that critical charged residues in and to the M2 domain of all the Cys loop LGICs on these permeation and that charged residues in the MA stretch of 5-HT3A and α4β2 nicotinic ACh receptors influence on the positively charged residues within this region of the α1 glycine receptor. We proposed that positive charge in the putative MA stretch the passage of negative ions through the as of the at negative membrane of of these positive or arginine residues to the negative glutamate residue did indeed reduce γ at negative potentials by about A reduction was also with mutations of or residues that all mutations to sets of positively charged and residues to a series of 3 well conserved arginine residues responsible for the γ of 5-HT3A receptors J.A. Carland J.E. Cooper M.A. Livesey M.R. Deeb T.Z. Hales T.G. Lambert J.J. 2006; Scholar). we that intracellular portals partially lined by charged M3–M4 loop residues that can influence conductance are a conserved feature among anionic and cationic LGICs. of the of these residues on ion permeation that the charge was the critical responsible for the with a on γ of charge by to The of arginine and glutamate are and that of Prog. Biophys. Mol. Biol.. Scholar). The modest in γ observed with the of residues into the MA stretch that the properties of residues at the positions investigated are not major determinants of ion the properties of these residues the greatest impact on to was with or on It is to how a influence γ that the of anion We such conductance by that γ is by the rate of ion flux a dominant selectivity at the of the M2 pore. the intracellular portals to an intracellular a of in this in the of a negative membrane increase the flux of The was by the with our that the M2 domain is the dominant determinant of ion permeation was the that the 4E no effect on ion charge selectivity of the glycine receptor. We that ions not as with charged residues lining the portals as with charged residues at the dominant selectivity at the intracellular of This be a of relative of the M2 pore and the intracellular portals, with the for the glycine receptor. The proposed diameter of the intracellular portals in the closed of the Torpedo nicotinic ACh receptor portals is comparable with that proposed for the M2 pore of that channel (3Unwin N. J. Mol. Biol.. 2005; 346: 967-989Google Scholar). It be to whether the of the portals of anionic LGICs are to of Torpedo or whether the portals of the open of LGICs are in the closed was that of the three MA stretch that are critical for the 5-HT3A receptor conductance also no effect on ion charge selectivity in channels (13Livesey M.R. Cooper M.A. Deeb T.Z. Carland J.E. Kozuska J. Hales T.G. Lambert J.J. Peters J.A. J. Biol. Chem.. 2008; 283: 19301-19313Google Scholar). In our results that intracellular portals lined by that influence γ are conserved the Cys loop LGIC and molecular of permeation through these channels to be to this extended permeation pathway. However, the lack of on ion charge selectivity and the in that we observed are consistent with an selectivity the dominant determinant of ion permeation properties in glycine receptors. Whether a for other anion selective Cys loop LGICs remains to be