Advertisement

Unique Contributions of an Arginine Side Chain to Ligand Recognition in a Glutamate-gated Chloride Channel*

  • Timothy Lynagh
    Affiliations
    Center for Biopharmaceuticals, Department of Drug Design and Pharmacology, University of Copenhagen, 2100 H Copenhagen, Denmark
    Search for articles by this author
  • Vitaly V. Komnatnyy
    Affiliations
    Center for Biopharmaceuticals, Department of Drug Design and Pharmacology, University of Copenhagen, 2100 H Copenhagen, Denmark
    Search for articles by this author
  • Stephan A. Pless
    Correspondence
    To whom correspondence should be addressed. Tel.: 45-23-649066.
    Affiliations
    Center for Biopharmaceuticals, Department of Drug Design and Pharmacology, University of Copenhagen, 2100 H Copenhagen, Denmark
    Search for articles by this author
  • Author Footnotes
    * This work was supported by the Center for Biopharmaceuticals (University of Copenhagen), the Carlsberg Foundation, the Lundbeck Foundation, and the Danish Council for Independent Research (Det Frie Forskningsråd). The authors declare that they have no conflicts of interest with the contents of this article.
    This article contains supplemental Fig. S1.
Open AccessPublished:January 17, 2017DOI:https://doi.org/10.1074/jbc.M116.772939
      Glutamate recognition by neurotransmitter receptors often relies on Arg residues in the binding site, leading to the assumption that charge-charge interactions underlie ligand recognition. However, assessing the precise chemical contribution of Arg side chains to protein function and pharmacology has proven to be exceedingly difficult in such large and complex proteins. Using the in vivo nonsense suppression approach, we report the first successful incorporation of the isosteric, titratable Arg analog, canavanine, into a neurotransmitter receptor in a living cell, utilizing a glutamate-gated chloride channel from the nematode Haemonchus contortus. Our data unveil a surprisingly small contribution of charge at a conserved arginine side chain previously suggested to form a salt bridge with the ligand, glutamate. Instead, our data show that Arg contributes crucially to ligand sensitivity via a hydrogen bond network, where Arg interacts both with agonist and with a conserved Thr side chain within the receptor. Together, the data provide a new explanation for the reliance of neurotransmitter receptors on Arg side chains and highlight the exceptional capacity of unnatural amino acid incorporation for increasing our understanding of ligand recognition.

      Introduction

      Neurotransmitter receptors are vital signaling proteins that are embedded in the cell membrane and trigger intracellular changes in response to extracellular chemical signals. The two classical receptor types are metabotropic, G-protein-coupled receptors (GPCRs) that act over seconds or minutes via intracellular second messengers (
      • Rajagopal S.
      • Rajagopal K.
      • Lefkowitz R.J.
      Teaching old receptors new tricks: biasing seven-transmembrane receptors.
      ), and ionotropic, ligand-gated ion channels (LGICs)
      The abbreviations used are: LGIC
      ligand-gated ion channel
      pLGIC
      pentameric ligand-gated ion channel
      Can
      canavanine
      GluCl
      glutamate-gated chloride channel
      iGluR
      ionotropic glutamate receptor
      Nvoc
      4,5-dimethoxy-2-nitrobenzyl
      pdCpA
      5′-O-phosphoryl-2′-deoxycytidylyl-(3′→5′)adenosine
      ESI
      electrospray ionization
      IVM
      ivermectin
      Glu
      glutamate
      ANOVA
      analysis of variance.
      that mediate ion flux across the membrane on the millisecond timescale (
      • Smart T.G.
      • Paoletti P.
      Synaptic neurotransmitter-gated receptors.
      ). The rapid chemo-electric signaling of LGICs is perfectly suited to the nervous system, where activation of sodium channels and chloride channels mediates excitatory and inhibitory signals, respectively (
      • Smart T.G.
      • Paoletti P.
      Synaptic neurotransmitter-gated receptors.
      ). The first step in the process of activation is the recognition of a specific ligand, which in the case of the animal nervous system is very often the neurotransmitter glutamate (
      • Moroz L.L.
      • Kocot K.M.
      • Citarella M.R.
      • Dosung S.
      • Norekian T.P.
      • Povolotskaya I.S.
      • Grigorenko A.P.
      • Dailey C.
      • Berezikov E.
      • Buckley K.M.
      • Ptitsyn A.
      • Reshetov D.
      • Mukherjee K.
      • Moroz T.P.
      • Bobkova Y.
      • et al.
      The ctenophore genome and the evolutionary origins of neural systems.
      ).
      Glutamate binding to neurotransmitter receptors has been studied in great detail, reflected in X-ray structures of ligand-receptor complexes of both LGICs (
      • Armstrong N.
      • Gouaux E.
      Mechanisms for activation and antagonism of an AMPA-sensitive glutamate receptor: crystal structures of the GluR2 ligand binding core.
      ,
      • Hibbs R.E.
      • Gouaux E.
      Principles of activation and permeation in an anion-selective Cys-loop receptor.
      ) and GPCRs (
      • Kunishima N.
      • Shimada Y.
      • Tsuji Y.
      • Sato T.
      • Yamamoto M.
      • Kumasaka T.
      • Nakanishi S.
      • Jingami H.
      • Morikawa K.
      Structural basis of glutamate recognition by a dimeric metabotropic glutamate receptor.
      ). Perhaps not surprisingly, each complex contains an Arg side chain in close proximity to at least one of the glutamate carboxylates, suggestive of ionic interactions between negatively charged carboxylate and positively charged guanidino groups. Reduced function upon Ala substitution confirms the importance of these Arg residues in glutamate recognition in each receptor subfamily (
      • Daeffler K.N.
      • Lester H.A.
      • Dougherty D.A.
      Functional evaluation of key interactions evident in the structure of the eukaryotic Cys-loop receptor GluCl.
      ,
      • Hampson D.R.
      • Huang X.P.
      • Pekhletski R.
      • Peltekova V.
      • Hornby G.
      • Thomsen C.
      • Thøgersen H.
      Probing the ligand-binding domain of the mGluR4 subtype of metabotropic glutamate receptor.
      • Kawamoto S.
      • Uchino S.
      • Xin K.Q.
      • Hattori S.
      • Hamajima K.
      • Fukushima J.
      • Mishina M.
      • Okuda K.
      Arginine-481 mutation abolishes ligand-binding of the AMPA-selective glutamate receptor channel α1-subunit.
      ), but despite this apparent functional evidence for a charge-charge interaction, replacing a large Arg side chain with a much smaller Ala side chain involves more physico-chemical changes than merely removing a positive charge. As such, the precise contribution of highly conserved Arg side chains in ligand recognition remains unknown.
      Here, we have sought experimental evidence for charge-charge interactions in ligand recognition, focusing on the AVR-14B glutamate-gated chloride channel (GluCl (
      • McCavera S.
      • Rogers A.T.
      • Yates D.M.
      • Woods D.J.
      • Wolstenholme A.J.
      An ivermectin-sensitive glutamate-gated chloride channel from the parasitic nematode Haemonchus contortus.
      )). GluCls are invertebrate-specific members of the pentameric ligand-gated ion channel (pLGIC or “Cys-loop receptor”) family, sharing significant homology with vertebrate GABA and acetylcholine receptors and constituting an important antiparasitic drug target (
      • Wolstenholme A.J.
      Glutamate-gated chloride channels.
      ). Despite the fact that a Caenorhabditis elegans GluCl was the first eukaryotic pLGIC to be visualized by X-ray crystallography (
      • Hibbs R.E.
      • Gouaux E.
      Principles of activation and permeation in an anion-selective Cys-loop receptor.
      ), the molecular basis for neurotransmitter recognition in GluCls has received little experimental interrogation, as compared with vertebrate homologs. It has recently been shown, however, that in GluCls, glutamate recognition involves interactions between aromatic residues on the principal face of the binding site with the glutamate amine (
      • Daeffler K.N.
      • Lester H.A.
      • Dougherty D.A.
      Functional evaluation of key interactions evident in the structure of the eukaryotic Cys-loop receptor GluCl.
      ), as well as interactions between Arg residues in the binding site with the glutamate carboxylate groups (
      • Lynagh T.
      • Beech R.N.
      • Lalande M.J.
      • Keller K.
      • Cromer B.A.
      • Wolstenholme A.J.
      • Laube B.
      Molecular basis for convergent evolution of glutamate recognition by pentameric ligand-gated ion channels.
      ), as illustrated in supplemental Fig. S1. We replaced Arg76 in the glutamate binding site with a titratable amino acid, providing us with the unprecedented opportunity to test glutamate sensitivity when an isosteric side chain is present but charged or uncharged. The results indicate only a small role of positive charge and unveil another unique property of Arg side chains that contributes to glutamate sensitivity, namely the ability to form hydrogen bonds both with the agonist and with vicinal receptor side chains.

      Results

      To test whether positive charge of Arg residues 76 and 95 is sufficient for glutamate recognition in the glutamate binding site of the AVR-14B GluCl, we replaced these individually with Lys via site-directed mutagenesis and measured glutamate-gated chloride currents with two electrode voltage clamp experiments (Fig. 1, A and B). Given the water-accessible location of these positions in GluCls (
      • Althoff T.
      • Hibbs R.E.
      • Banerjee S.
      • Gouaux E.
      X-ray structures of GluCl in apo states reveal a gating mechanism of Cys-loop receptors.
      ) and the positive charge on Lys side chains in such environments (
      • Cymes G.D.
      • Ni Y.
      • Grosman C.
      Probing ion-channel pores one proton at a time.
      ), one would expect glutamate sensitivity of mutant receptors to reflect that of WT receptors if positive charge were the main contribution of these side chains. However, we observed that glutamate sensitivity was practically abolished at R76K and R95K mutants (Fig. 1B), suggesting that positive charge alone at these positions is not sufficient for high glutamate sensitivity.
      Figure thumbnail gr1
      FIGURE 1Arg-Lys substitutions drastically reduce glutamate recognition. A, X-ray structure of GLC-1 GluCl (PDB 3RIF; gray shading, notional cell membrane). Magnified view shows glutamate binding site and selected amino acid side chains. These include GLC-1 arginine residues 37 and 56, which are labeled Arg76 and Arg95 to describe the equivalent residues from the AVR-14B GluCl used in the present study. B, left, example recordings of glutamate (Glu) and ivermectin (IVM, 1 μm) responses at oocytes expressing mutant AVR-14B GluCls (scale bars: x, 5 s; y, 2 μA). Activation by IVM, which binds elsewhere on the receptor, confirms cell surface expression in the absence of Glu-gated currents. Right, mean ± S.E. (n = 4–8) peak current responses to increasing concentrations of Glu, normalized to maximum Glu-gated current (WT) or maximum IVM-gated current (mutants).
      Although Arg-to-Lys substitution largely retains side chain size and charge, it also involves the loss of two potentially hydrogen-bonding (H-bonding) nitrogen atoms (Fig. 2A), which could also underlie the loss of glutamate sensitivity we observed in R76K and R95K mutants. Assessing the precise contributions of charge or H-bonding of the Arg side chain to glutamate recognition would thus require substitution with an uncharged but otherwise isosteric analog. To this end, we sought to replace Arg76 or Arg95 with canavanine (Can), an isosteric Arg analog with a pKa of 7 (
      • Boyar A.
      • Marsh R.E.
      l-Canavanine, a paradigm for the structures of substituted guanidines.
      ) (Fig. 2A), reasoning that experiments at low and high pH would assay glutamate sensitivity when a single guanidino side chain is protonated or deprotonated, respectively. Site-specific incorporation of Can was achieved by the in vivo nonsense suppression method (
      • Dougherty D.A.
      • Van Arnam E.B.
      In vivo incorporation of non-canonical amino acids by using the chemical aminoacylation strategy: a broadly applicable mechanistic tool.
      ) (Fig. 2B). Successful incorporation of Can at position Arg76 was evident in robust glutamate-gated currents through Can76, receptors, but although currents were also observed for Can95, these were not significantly greater than controls lacking Can (Fig. 2C; 18 ± 5 nA, n = 10; 4 ± 2 nA, n = 8). This made further characterization of Can95 receptors difficult, and we chose not to investigate these further.
      Figure thumbnail gr2
      FIGURE 2Incorporation of titratable arginine analog, canavanine. A, l-lysine (Lys), l-arginine (Arg), and l-canavanine (Can). B, graphic illustrating nonsense suppression of Arg76UAG mRNA by co-injection of Can-ligated tRNA into Xenopus laevis oocytes (yellow/brown spheres). C, mean peak current amplitude (± S.E.) in response to 10 mm glutamate at oocytes injected with mRNA and tRNA ± Can (n = 8–15; *, p < 0.05; n.s. not significant, ANOVA). Inset shows example responses to 10 mm glutamate at oocytes (pH = 7) injected with indicated RNA combinations (scale bars: x, 5 s; y, 20 nA). D, example recordings of current responses to increasing concentrations of glutamate and mean ± S.E. peak current responses (n = 5–8; normalized to maximum glutamate-gated current) at Can76 GluCls (Can76). Scale bars: x, 10 s; y, 100 nA. E, left, glutamate-gated currents at oocytes expressing wild-type or Can76 GluCls continuously perfused at pH 7.0 or 5.8, as indicated (scale bars: x, 30 s; y, 100 nA). Right, mean ± S.E. peak current responses to 30 μm glutamate, normalized to maximum glutamate-gated current (I30 μm/Imax; n = 5–8; *, p < 0.05, ***, p < 0.001, Student's t test). Gray arrows illustrate inhibition of 30 μm glutamate-gated current. Can76 pH 7.0 recording in E is repeated from D.
      When we measured glutamate sensitivity of Can76 receptors at pH 7.0, at which one would expect only ∼50% of Can76 side chains to be protonated, we were surprised to find that the EC50 for activation by glutamate was 48 ± 7 μm (n = 6), and thus not significantly different from WT receptors (41 ± 11 μm, n = 8). At pH 8.5, when even fewer, if any, Can76 side chains are expected to be protonated, we observed only a modest decrease in glutamate sensitivity (Fig. 2D), with the EC50 significantly increased to 119 ± 11 μm (Table 1). A further increase in pH to 9.2 saw no additional rise in EC50 value (Fig. 2D; Table 1), suggesting that the effect was saturated around pH 8.5. No such pH-dependent shift was seen for WT receptors (Table 1), in which Arg76 residues are always protonated. This indicates that a 2-fold decrease in glutamate sensitivity in conditions that deprotonate the Can (not Arg) side chains is specific for receptors incorporating a Can residue in the glutamate binding site. Unfortunately, we could not measure the effects of fully protonated Can76 side chains, as acidic pH causes significant inhibition of function even in WT receptors (Fig. 2E), as is the case in structurally related GABA and glycine receptors (
      • Chen Z.
      • Dillon G.H.
      • Huang R.
      Molecular determinants of proton modulation of glycine receptors.
      ,
      • Huang R.Q.
      • Dillon G.H.
      Effect of extracellular pH on GABA-activated current in rat recombinant receptors and thin hypothalamic slices.
      ). To verify the 2-fold decrease in glutamate sensitivity observed with an uncharged analog at position 76, we attempted to replace Arg76 with citrulline, an analog in which one η nitrogen is replaced by an oxygen and which is uncharged (
      • Judice J.K.
      • Gamble T.R.
      • Murphy E.C.
      • de Vos A.M.
      • Schultz P.G.
      Probing the mechanism of staphylococcal nuclease with unnatural amino acids: kinetic and structural studies.
      ), via nonsense suppression, but this was not successful (data not shown). Thus, and although a complete titration could not be completed, our data show that in conditions in which theoretically only 50% of Can76 side chains carry a positive charge, Can76 receptors show very similar glutamate sensitivity to WT Arg76 receptors (Table 1). Perhaps more strikingly, upon deprotonation and loss of positive charge in Can76 receptors, a mere 2-fold reduction in glutamate sensitivity is observed. This is a modest reduction in agonist sensitivity as compared with the 10,000-fold reduction caused by R76N or even R76K mutations in this very receptor (
      • Lynagh T.
      • Beech R.N.
      • Lalande M.J.
      • Keller K.
      • Cromer B.A.
      • Wolstenholme A.J.
      • Laube B.
      Molecular basis for convergent evolution of glutamate recognition by pentameric ligand-gated ion channels.
      ) (Fig. 1), raising the possibility that a more substantial contribution to glutamate binding derives from some property of the Arg (or, indeed, Can) side chain other than positive charge.
      TABLE 1Glutamate sensitivity of Can76 and Arg76 GluCls
      EC50
      EC50 value was calculated by plotting current amplitude against glutamate concentration and fitting with Hill equation for experiments at individual oocytes and then averaged. Data are mean ± S.E.
      Imaxn
      μmnA
      Can76
      pH 7.048 ± 760 ± 406
      pH 8.5119 ± 11***214 ± 915
      pH 9.2118 ± 17***110 ± 276
      Arg76 (WT)
      pH 7.041 ± 11241 ± 518
      pH 8.543 ± 5170 ± 478
      pH 9.257 ± 7428 ± 917
      a EC50 value was calculated by plotting current amplitude against glutamate concentration and fitting with Hill equation for experiments at individual oocytes and then averaged. Data are mean ± S.E.
      Although some of that contribution is presumably via H-bonds between Arg (or Can) ηNH2 group(s) and glutamate (Fig. 3A), we considered that the Arg ϵNH group could also be important. Indeed, we noticed in the GLC-1 GluCl-glutamate X-ray structure (Protein Data Bank (PDB) 3RIF (
      • Hibbs R.E.
      • Gouaux E.
      Principles of activation and permeation in an anion-selective Cys-loop receptor.
      )) that although not in direct contact with the bound agonist, the hydroxyl oxygen atom of a Thr residue (equivalent to Thr93 in the AVR-14B GluCl) is located in close (2.9 Å) proximity to the ϵNH of the Arg equivalent to Arg76 (Fig. 3A). If this potential H-bond were important for glutamate sensitivity, we reasoned that the T93S substitution should retain glutamate sensitivity, as Ser also possesses a β-hydroxyl group. In contrast, Val is sterically similar to Thr but devoid of the hydroxyl, and mutant T93V receptors would be expected to show decreased glutamate sensitivity. Indeed, when we measured glutamate-gated currents at these mutants, T93V receptors showed drastically reduced glutamate sensitivity, barely responding to millimolar concentrations (Fig. 3, B and C). By contrast, T93S receptors showed a 370-fold increase in glutamate sensitivity as compared with WT (Fig. 3, B and C; Table 2), confirming that the hydroxyl at this position is required for regular (or increased) glutamate sensitivity, likely through an interaction with Arg76.
      Figure thumbnail gr3
      FIGURE 3Conventional mutagenesis shows the importance of Thr93 in glutamate recognition. A, magnified view of GLC-1 crystal structure illustrating proximity of Thr93 hydroxyl oxygen to Arg76 ϵ nitrogen (2.9 Å; orange dashed line; numbers refer to equivalent residues in AVR-14B). Dashed lines indicate inter-atomic distances ≤3.5 Å. Amino acid sequence alignment shows selected Loop G and Loop D residues from ecdysozoan GluCls (dark font), lophotrochozoan GluCls and vertebrate and invertebrate GABA and glycine receptors (light font). B, example recordings of glutamate (Glu) and ivermectin (IVM) responses at oocytes expressing mutant AVR-14B GluCls (scale bars: x, 5 s; y, 2 μA). C, mean ± S.E. (n = 6–8) peak current responses to increasing concentrations of glutamate, normalized to maximum glutamate-gated current (WT and T93S) or maximum IVM-gated current (T93V).
      TABLE 2Glutamate sensitivity of WT and conventional mutants
      EC50
      EC50 value was calculated by plotting current amplitude against glutamate concentration and fitting with Hill equation for experiments at individual oocytes and then averaged. Data are mean ± S.E.
      Imax
      Maximum glutamate-gated peak currents. Mean ± S.E.
      n
      μmμA
      WT17 ± 35.0 ± 0.54
      R76K≫100 mm
      For certain mutants, saturation in the concentration-response relationship was not reached, up to 30 mm glutamate. EC50 values are therefore estimated to be well over 100 mm.
      0.10 ± 0.028
      T93S0.046 ± 0.011***4.8 ± 0.76
      R95K≫100 mm
      For certain mutants, saturation in the concentration-response relationship was not reached, up to 30 mm glutamate. EC50 values are therefore estimated to be well over 100 mm.
      0.05 ± 0.017
      R76K/T93S≫100 mm
      For certain mutants, saturation in the concentration-response relationship was not reached, up to 30 mm glutamate. EC50 values are therefore estimated to be well over 100 mm.
      0.44 ± 0.098
      R95K/T93S530 ± 1401.3 ± 0.49
      a EC50 value was calculated by plotting current amplitude against glutamate concentration and fitting with Hill equation for experiments at individual oocytes and then averaged. Data are mean ± S.E.
      b Maximum glutamate-gated peak currents. Mean ± S.E.
      c For certain mutants, saturation in the concentration-response relationship was not reached, up to 30 mm glutamate. EC50 values are therefore estimated to be well over 100 mm.
      Seeking specific evidence for this interaction, we performed double mutant cycle analysis, according to which the T93S mutation should not restore high glutamate sensitivity on R76K mutant receptors because these WT residues are coupled and the effects of their mutation are not simply additive (
      • Hidalgo P.
      • MacKinnon R.
      Revealing the architecture of a K+ channel pore through mutant cycles with a peptide inhibitor.
      ,
      • 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.
      • Wilkinson A.J.
      • Fersht A.R.
      • Blow D.M.
      • Winter G.
      Site-directed mutagenesis as a probe of enzyme structure and catalysis: tyrosyl-tRNA synthetase cysteine-35 to glycine-35 mutation.
      ). In contrast, the combined effects of mutating independent residues are additive, and the T93S mutation might be expected to confer higher glutamate sensitivity on R95K receptors, as we do not expect coupling between these two residues. Indeed, double mutant T93S/R95K receptors responded to glutamate in a concentration range between T93S and R95K single mutants (Fig. 4A; Table 2), indicative of independence. R76K/T93S receptors, however, showed no greater glutamate sensitivity than single mutant R76K receptors (Fig. 4A), and analysis indicated an energetic coupling of ∼14 kJ/mol between Arg76 and Thr93 (Fig. 4B), which we interpret as evidence for a strong H-bond between the Arg ϵNH and Thr OH groups.
      Figure thumbnail gr4
      FIGURE 4Double mutant cycle analysis. A, glutamate-gated currents at oocytes expressing double mutant T93S/R95K or R76K/T93S receptors and mean (± S.E.) data. Robust activation by IVM indicates successful expression of R76K/T93S despite small responses to glutamate. WT and single mutant data are repeated from earlier figures for comparison. B, principles of double mutant cycle analysis and analysis of Arg76–Thr93 and Arg95–Thr93 coupling. For two residues “A” and “B,” EC50 values of WT (AB), single mutant A′B, single mutant AB′, and double mutant A′B′ are used to calculate the coupling coefficient, Ω, from which the coupling energy, ΔΔG, can be calculated (
      • Wilkinson A.J.
      • Fersht A.R.
      • Blow D.M.
      • Winter G.
      Site-directed mutagenesis as a probe of enzyme structure and catalysis: tyrosyl-tRNA synthetase cysteine-35 to glycine-35 mutation.
      ). Our final values for Ω and ΔΔG are only estimates because the EC50 values of certain single and double mutants could not be calculated: here the EC50 values have been estimated from A.

      Discussion

      Taken together, these results suggest that the positive charge of Arg76 contributes little to glutamate binding in GluCls. Instead, our results show that two other aspects of the Arg side chain contribute to effective glutamate recognition. First, the data suggest that H-bonds between the Arg ηNH2 group(s) and the agonist α-carboxylate are important, as conventional Asn and Lys mutations that remove this moiety of Arg are severely detrimental to glutamate recognition, whereas the non-canonical substitution of Can, which retains ηNH2 groups, retains WT-like glutamate sensitivity. Second, the ϵNH of Arg appears to interact closely with a vicinal receptor hydroxyl side chain, the removal of which via Val substitution drastically reduces glutamate sensitivity. Based on available GluCl structures, this interaction is likely to stabilize Arg for its interaction(s) with the glutamate α-carboxylate (Fig. 3A). Notably, this Loop D hydroxyl side chain is highly conserved in GluCls that possess the Loop G Arg (Arg76 in AVR-14B), which interacts with the glutamate α-carboxylate. By contrast, the Loop D hydroxyl is absent in GluCls where instead a Loop A Arg (on the opposing face of the binding site, equivalent to Q141 in AVR-14B) interacts with the glutamate α-carboxylate (Fig. 3A) (
      • Lynagh T.
      • Beech R.N.
      • Lalande M.J.
      • Keller K.
      • Cromer B.A.
      • Wolstenholme A.J.
      • Laube B.
      Molecular basis for convergent evolution of glutamate recognition by pentameric ligand-gated ion channels.
      ).
      Conventional mutagenesis is an indispensable tool for dissection of protein structure and function, but in fine-tuning the details of ligand recognition, it is limited by the numerous physico-chemical changes involved in most substitutions (
      • Pless S.A.
      • Ahern C.A.
      Unnatural amino acids as probes of ligand-receptor interactions and their conformational consequences.
      ). This is perhaps especially the case for Arg, where conventional analogs Lys and His are noticeably smaller and more frequently than Arg uncharged in physiological settings (
      • Cymes G.D.
      • Ni Y.
      • Grosman C.
      Probing ion-channel pores one proton at a time.
      ,
      • Mehler E.L.
      • Fuxreiter M.
      • Simon I.
      • Garcia-Moreno E.B.
      The role of hydrophobic microenvironments in modulating pKa shifts in proteins.
      ). Previous attempts to circumvent the limitations of conventional mutagenesis by incorporating unnatural Arg analogs have been few (
      • Judice J.K.
      • Gamble T.R.
      • Murphy E.C.
      • de Vos A.M.
      • Schultz P.G.
      Probing the mechanism of staphylococcal nuclease with unnatural amino acids: kinetic and structural studies.
      ,
      • Akahoshi A.
      • Suzue Y.
      • Kitamatsu M.
      • Sisido M.
      • Ohtsuki T.
      Site-specific incorporation of arginine analogs into proteins using arginyl-tRNA synthetase.
      ,
      • Le D.D.
      • Cortesi A.T.
      • Myers S.A.
      • Burlingame A.L.
      • Fujimori D.G.
      Site-specific and regiospecific installation of methylarginine analogues into recombinant histones and insights into effector protein binding.
      ) and arguably difficult (
      • Krishnakumar R.
      • Ling J.
      Experimental challenges of sense codon reassignment: an innovative approach to genetic code expansion.
      ). Our use of the in vivo nonsense suppression method was successful for one of the two positions tested, and despite limited efficiency (currents through Can76 were substantially smaller than conventional mutant receptors; compare Imax in TABLE 1, TABLE 2), the data strongly support the notion of robust and specific Can incorporation in three ways. First, the level of nonspecific incorporation of endogenous amino acids, as inferred from the very low current levels upon co-injection of uncharged tRNA, was very low (Fig. 2C). Second, although all conventional replacements at position 76 tested here and elsewhere (
      • Daeffler K.N.
      • Lester H.A.
      • Dougherty D.A.
      Functional evaluation of key interactions evident in the structure of the eukaryotic Cys-loop receptor GluCl.
      ,
      • Lynagh T.
      • Beech R.N.
      • Lalande M.J.
      • Keller K.
      • Cromer B.A.
      • Wolstenholme A.J.
      • Laube B.
      Molecular basis for convergent evolution of glutamate recognition by pentameric ligand-gated ion channels.
      ) resulted in drastic losses in glutamate sensitivity, we found Can incorporation to yield WT-like glutamate sensitivity (at pH 7.0; Table 1). Lastly, and perhaps most significantly, we found the agonist sensitivity of Can76 receptors to be titratable between pH 7.0 and 8.5, a property that would be highly unlikely in the event of endogenous amino acids incorporating into this position. We note here that in our hands the incorporation of Can was more successful than citrulline at positions 76 and 95. However, we cannot assess with certainty the reason for this difference, which could be due to ribosomal recognition of the charged tRNA, protein folding, or protein function.
      In conclusion, Can incorporation, together with subsequent conventional mutagenesis, has unveiled crucial determinants of ligand recognition that had previously escaped identification. Our results suggest that the common occurrence of Arg residues in glutamate binding sites is related to the ability of Arg side chains to participate in H-bonds both with ligand and with vicinal receptor side chains simultaneously. Thus, the unique propensity of Arg for forming multiple H-bonds, as described previously in the context of intramolecular interactions in other protein families (
      • Borders Jr, C.L.
      • Broadwater J.A.
      • Bekeny P.A.
      • Salmon J.E.
      • Lee A.S.
      • Eldridge A.M.
      • Pett V.B.
      A structural role for arginine in proteins: multiple hydrogen bonds to backbone carbonyl oxygens.
      ), seems to have been utilized by GluCls for the specific functional requirements of ligand recognition. The results also complement work on tetrameric ionotropic glutamate receptors (iGluRs), regarding both the unique role of Arg and the occurrence of vicinal hydroxyl side chains that could stabilize binding site architecture. In NMDA-type iGluRs, for example, Lys substitution of the Arg residue that binds the ligand α-carboxylate abolishes glutamate sensitivity (
      • Chen P.E.
      • Geballe M.T.
      • Stansfeld P.J.
      • Johnston A.R.
      • Yuan H.
      • Jacob A.L.
      • Snyder J.P.
      • Traynelis S.F.
      • Wyllie D.J.
      Structural features of the glutamate binding site in recombinant NR1/NR2A N-methyl-d-aspartate receptors determined by site-directed mutagenesis and molecular modeling.
      ,
      • Laube B.
      • Hirai H.
      • Sturgess M.
      • Betz H.
      • Kuhse J.
      Molecular determinants of agonist discrimination by NMDA receptor subunits: analysis of the glutamate binding site on the NR2B subunit.
      ), much like R76K and R95K substitutions in the GluCl. Remarkably, NMDA-type iGluRs also contain Ser and Thr residues, whose side chain hydroxyl oxygens are, similar to Thr93 in the GluCl, as close as 2.7 Å to other side chains that form the glutamate binding site (
      • Chen P.E.
      • Geballe M.T.
      • Stansfeld P.J.
      • Johnston A.R.
      • Yuan H.
      • Jacob A.L.
      • Snyder J.P.
      • Traynelis S.F.
      • Wyllie D.J.
      Structural features of the glutamate binding site in recombinant NR1/NR2A N-methyl-d-aspartate receptors determined by site-directed mutagenesis and molecular modeling.
      ,
      • Furukawa H.
      • Singh S.K.
      • Mancusso R.
      • Gouaux E.
      Subunit arrangement and function in NMDA receptors.
      ), and whose substitution for Ala drastically reduces glutamate sensitivity (
      • Chen P.E.
      • Geballe M.T.
      • Stansfeld P.J.
      • Johnston A.R.
      • Yuan H.
      • Jacob A.L.
      • Snyder J.P.
      • Traynelis S.F.
      • Wyllie D.J.
      Structural features of the glutamate binding site in recombinant NR1/NR2A N-methyl-d-aspartate receptors determined by site-directed mutagenesis and molecular modeling.
      ).
      We present here, to our knowledge, the first example of Arg analog incorporation into membrane-bound receptors, and as such, these results describe an incisive approach to dissecting chemical interactions in a broad and therapeutically relevant family of membrane proteins. Interestingly, an H-bond network adjacent to the ligand binding site has been proposed for the structurally related glycine receptor (
      • Yu R.
      • Hurdiss E.
      • Greiner T.
      • Lape R.
      • Sivilotti L.
      • Biggin P.C.
      Agonist and antagonist binding in human glycine receptors.
      ), suggesting that stabilization of ligand binding by such H-bond networks could be a conserved feature of ligand recognition by pLGICs.

      Author Contributions

      All authors conceptualized the design and developed the methodology. T. L. and V. V. K. performed investigations. T. L. wrote the original draft, and all authors reviewed and edited the manuscript; T. L. performed visualization. T. L. and S. A. P. were responsible for funding acquisition; S. A. P. supervised the study.

      References

        • Rajagopal S.
        • Rajagopal K.
        • Lefkowitz R.J.
        Teaching old receptors new tricks: biasing seven-transmembrane receptors.
        Nat. Rev. Drug Discov. 2010; 9: 373-386
        • Smart T.G.
        • Paoletti P.
        Synaptic neurotransmitter-gated receptors.
        Cold Spring Harb. Perspect. Biol. 2012; 4: a009662
        • Moroz L.L.
        • Kocot K.M.
        • Citarella M.R.
        • Dosung S.
        • Norekian T.P.
        • Povolotskaya I.S.
        • Grigorenko A.P.
        • Dailey C.
        • Berezikov E.
        • Buckley K.M.
        • Ptitsyn A.
        • Reshetov D.
        • Mukherjee K.
        • Moroz T.P.
        • Bobkova Y.
        • et al.
        The ctenophore genome and the evolutionary origins of neural systems.
        Nature. 2014; 510: 109-114
        • Armstrong N.
        • Gouaux E.
        Mechanisms for activation and antagonism of an AMPA-sensitive glutamate receptor: crystal structures of the GluR2 ligand binding core.
        Neuron. 2000; 28: 165-181
        • Hibbs R.E.
        • Gouaux E.
        Principles of activation and permeation in an anion-selective Cys-loop receptor.
        Nature. 2011; 474: 54-60
        • Kunishima N.
        • Shimada Y.
        • Tsuji Y.
        • Sato T.
        • Yamamoto M.
        • Kumasaka T.
        • Nakanishi S.
        • Jingami H.
        • Morikawa K.
        Structural basis of glutamate recognition by a dimeric metabotropic glutamate receptor.
        Nature. 2000; 407: 971-977
        • Daeffler K.N.
        • Lester H.A.
        • Dougherty D.A.
        Functional evaluation of key interactions evident in the structure of the eukaryotic Cys-loop receptor GluCl.
        ACS Chem. Biol. 2014; 9: 2283-2290
        • Hampson D.R.
        • Huang X.P.
        • Pekhletski R.
        • Peltekova V.
        • Hornby G.
        • Thomsen C.
        • Thøgersen H.
        Probing the ligand-binding domain of the mGluR4 subtype of metabotropic glutamate receptor.
        J. Biol. Chem. 1999; 274: 33488-33495
        • Kawamoto S.
        • Uchino S.
        • Xin K.Q.
        • Hattori S.
        • Hamajima K.
        • Fukushima J.
        • Mishina M.
        • Okuda K.
        Arginine-481 mutation abolishes ligand-binding of the AMPA-selective glutamate receptor channel α1-subunit.
        Brain Res. Mol. Brain Res. 1997; 47: 339-344
        • McCavera S.
        • Rogers A.T.
        • Yates D.M.
        • Woods D.J.
        • Wolstenholme A.J.
        An ivermectin-sensitive glutamate-gated chloride channel from the parasitic nematode Haemonchus contortus.
        Mol. Pharmacol. 2009; 75: 1347-1355
        • Wolstenholme A.J.
        Glutamate-gated chloride channels.
        J. Biol. Chem. 2012; 287: 40232-40238
        • Lynagh T.
        • Beech R.N.
        • Lalande M.J.
        • Keller K.
        • Cromer B.A.
        • Wolstenholme A.J.
        • Laube B.
        Molecular basis for convergent evolution of glutamate recognition by pentameric ligand-gated ion channels.
        Sci. Rep. 2015; 5: 8558
        • Althoff T.
        • Hibbs R.E.
        • Banerjee S.
        • Gouaux E.
        X-ray structures of GluCl in apo states reveal a gating mechanism of Cys-loop receptors.
        Nature. 2014; 512: 333-337
        • Cymes G.D.
        • Ni Y.
        • Grosman C.
        Probing ion-channel pores one proton at a time.
        Nature. 2005; 438: 975-980
        • Boyar A.
        • Marsh R.E.
        l-Canavanine, a paradigm for the structures of substituted guanidines.
        J. Am. Chem. Soc. 1982; 104: 1995-1998
        • Dougherty D.A.
        • Van Arnam E.B.
        In vivo incorporation of non-canonical amino acids by using the chemical aminoacylation strategy: a broadly applicable mechanistic tool.
        Chembiochem. 2014; 15: 1710-1720
        • Chen Z.
        • Dillon G.H.
        • Huang R.
        Molecular determinants of proton modulation of glycine receptors.
        J. Biol. Chem. 2004; 279: 876-883
        • Huang R.Q.
        • Dillon G.H.
        Effect of extracellular pH on GABA-activated current in rat recombinant receptors and thin hypothalamic slices.
        J. Neurophysiol. 1999; 82: 1233-1243
        • Judice J.K.
        • Gamble T.R.
        • Murphy E.C.
        • de Vos A.M.
        • Schultz P.G.
        Probing the mechanism of staphylococcal nuclease with unnatural amino acids: kinetic and structural studies.
        Science. 1993; 261: 1578-1581
        • Hidalgo P.
        • MacKinnon R.
        Revealing the architecture of a K+ channel pore through mutant cycles with a peptide inhibitor.
        Science. 1995; 268: 307-310
        • 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: 272-275
        • Wilkinson A.J.
        • Fersht A.R.
        • Blow D.M.
        • Winter G.
        Site-directed mutagenesis as a probe of enzyme structure and catalysis: tyrosyl-tRNA synthetase cysteine-35 to glycine-35 mutation.
        Biochemistry. 1983; 22: 3581-3586
        • Pless S.A.
        • Ahern C.A.
        Unnatural amino acids as probes of ligand-receptor interactions and their conformational consequences.
        Annu. Rev. Pharmacol. Toxicol. 2013; 53: 211-229
        • Mehler E.L.
        • Fuxreiter M.
        • Simon I.
        • Garcia-Moreno E.B.
        The role of hydrophobic microenvironments in modulating pKa shifts in proteins.
        Proteins. 2002; 48: 283-292
        • Akahoshi A.
        • Suzue Y.
        • Kitamatsu M.
        • Sisido M.
        • Ohtsuki T.
        Site-specific incorporation of arginine analogs into proteins using arginyl-tRNA synthetase.
        Biochem. Biophys. Res. Commun. 2011; 414: 625-630
        • Le D.D.
        • Cortesi A.T.
        • Myers S.A.
        • Burlingame A.L.
        • Fujimori D.G.
        Site-specific and regiospecific installation of methylarginine analogues into recombinant histones and insights into effector protein binding.
        J. Am. Chem. Soc. 2013; 135: 2879-2882
        • Krishnakumar R.
        • Ling J.
        Experimental challenges of sense codon reassignment: an innovative approach to genetic code expansion.
        FEBS Lett. 2014; 588: 383-388
        • Borders Jr, C.L.
        • Broadwater J.A.
        • Bekeny P.A.
        • Salmon J.E.
        • Lee A.S.
        • Eldridge A.M.
        • Pett V.B.
        A structural role for arginine in proteins: multiple hydrogen bonds to backbone carbonyl oxygens.
        Protein Sci. 1994; 3: 541-548
        • Chen P.E.
        • Geballe M.T.
        • Stansfeld P.J.
        • Johnston A.R.
        • Yuan H.
        • Jacob A.L.
        • Snyder J.P.
        • Traynelis S.F.
        • Wyllie D.J.
        Structural features of the glutamate binding site in recombinant NR1/NR2A N-methyl-d-aspartate receptors determined by site-directed mutagenesis and molecular modeling.
        Mol. Pharmacol. 2005; 67: 1470-1484
        • Laube B.
        • Hirai H.
        • Sturgess M.
        • Betz H.
        • Kuhse J.
        Molecular determinants of agonist discrimination by NMDA receptor subunits: analysis of the glutamate binding site on the NR2B subunit.
        Neuron. 1997; 18: 493-503
        • Furukawa H.
        • Singh S.K.
        • Mancusso R.
        • Gouaux E.
        Subunit arrangement and function in NMDA receptors.
        Nature. 2005; 438: 185-192
        • Yu R.
        • Hurdiss E.
        • Greiner T.
        • Lape R.
        • Sivilotti L.
        • Biggin P.C.
        Agonist and antagonist binding in human glycine receptors.
        Biochemistry. 2014; 53: 6041-6051
        • Pajpanova T.
        • Stoev S.
        • Golovinsky E.
        • Krauß G.-J.
        • Miersch J.
        Canavanine derivatives useful in peptide synthesis.
        Amino Acids. 1997; 12: 191-204
        • Nowak M.W.
        • Gallivan J.P.
        • Silverman S.K.
        • Labarca C.G.
        • Dougherty D.A.
        • Lester H.A.
        In vivo incorporation of unnatural amino acids into ion channels in Xenopus oocyte expression system.
        Methods Enzymol. 1998; 293: 504-529
        • Dahan D.S.
        • Dibas M.I.
        • Petersson E.J.
        • Auyeung V.C.
        • Chanda B.
        • Bezanilla F.
        • Dougherty D.A.
        • Lester H.A.
        A fluorophore attached to nicotinic acetylcholine receptor βM2 detects productive binding of agonist to the αδ site.
        Proc. Natl. Acad. Sci. U.S.A. 2004; 101: 10195-10200
        • Edgar R.C.
        MUSCLE: multiple sequence alignment with high accuracy and high throughput.
        Nucleic Acids Res. 2004; 32: 1792-1797