A Change in the Ion Selectivity of Ligand-Gated Ion Channels Provides a Mechanism to Switch Behavior

Abstract

Behavioral output of neural networks depends on a delicate balance between excitatory and inhibitory synaptic connections. However, it is not known whether network formation and stability is constrained by the sign of synaptic connections between neurons within the network. Here we show that switching the sign of a synapse within a neural circuit can reverse the behavioral output. The inhibitory tyramine-gated chloride channel, LGC-55, induces head relaxation and inhibits forward locomotion during the Caenorhabditis elegans escape response. We switched the ion selectivity of an inhibitory LGC-55 anion channel to an excitatory LGC-55 cation channel. The engineered cation channel is properly trafficked in the native neural circuit and results in behavioral responses that are opposite to those produced by activation of the LGC-55 anion channel. Our findings indicate that switches in ion selectivity of ligand-gated ion channels (LGICs) do not affect network connectivity or stability and may provide an evolutionary and a synthetic mechanism to change behavior.

Fig 1. LGC-55 cation channel mutants gate sodium.

(A) Cys-loop LGICs are homopentameric channels, each subunit containing four transmembrane domains. Depicted is a schematic representation of an LGIC with transmembrane domains 1 and 2 (M1, M2) in light gray. In blue is the intracellular loop that links M1 and M2, which determines the ion selectivity of the channel. (B) Alignment of M1–M2 loop region of LGC-55 with structurally related Cys-loop LGICs. Identities are shaded in dark gray, while similarities are light gray. The blue boxes indicate residues that determine selectivity of anions, while red boxes indicate those for cation selectivity. The engineered LGC-55 cation-I and cation-II channels contain the M1 loop of the cationic 5HT3a receptor. The LGC-55 cation-II channel also contains an additional mutation at the 20ʹ residue, which is predicted to enhance cation selectivity (see text for details). (C) Ion selectivity of the LGC-55 anion (left) and LGC-55 cation-II (right) receptor in cultured C. elegans muscle cells. Tyramine (TA) evoked (0.5 mM, 250 ms) currents were recorded at the holding potentials shown. Black circles: ES1 (standard solution: 150 mM Na⁺, 165 mM Cl^-), LGC-55 anion: E_rev = -26.8 ± 3.1mV (n = 4), LGC-55 cation-II: E_rev = 2.4 ± 1.2 mV (n = 5); red squares: ES2 (low Na⁺: 15 mM Na⁺, 165 mM Cl^-), LGC-55 anion: = -24.3 ± 1.6 mV (n = 4), LGC-55 cation-II: -21.9 ± 2.6 mV (n = 5); blue triangles: ES3 (low Cl^-: 150 mM Na⁺, 30 mM Cl^-), LGC-55 anion: -1.9 ± 2.3 mV (n = 4) LGC-55 cation-II: 1.7 ± 0.9 mV (n = 5). The insets show representative macrocurrents of LGC-55 anion (left) and LGC-55 cation-II (right) elicited after perfusion of 0.5 mM tyramine at membrane holding potentials ranging from -60 to +60 mV in 20 mV steps in the standard solution.

Fig 2. Engineered LGC-55 cation channels are functional in vivo.

(A) Still images of wild-type and transgenic animals expressing LGC-55 anion or cation-II ectopically in all body wall muscle cells, on exogenous tyramine. LGC-55 anion animals paralyze in relaxed extended posture, while LGC-55 cation-II animals are hypercontracted. Scale bar = 0.25 mm. (B) Quantification of body length on exogenous tyramine (wild type, n = 57, Pmyo-3::LGC-55 anion(zfEx31), n = 53; Pmyo-3::LGC-55 cation-I(zfEx120), n = 59; Pmyo-3::LGC-55 cation-II(zfEx41), n = 55). Error bars represent the standard error of the mean (SEM). Statistical significance as indicated, *** p < 0.0001.

Fig 3. LGC-55 cation channels localize to postsynaptic specializations in the nerve ring.

(A) Right: schematic diagram depicting the location of the RIM cell bodies (red) and LGC-55 expressing neurons and neck muscles in the head (green). Left: side view of the main synaptic outputs of the RIM with the AVB in its ventral process and neck muscles and the RMD/SMD motor neurons in the nerve ring. (B) Representative images of animals coexpressing the synaptic vesicle marker mCherry::RAB-3 in the tyraminergic RIM neurons (left), a translational LGC-55 anion::GFP (first row), LGC-55 cation-II::GFP reporter (center row), and a translational LGC-55 anion::GFP in tyramine-deficient, tdc-1 mutant background (bottom row). Merge (right) shows synaptic contacts between the RIM and LGC-55 expressing neurons. The anterior is left, the nerve ring is indicated by a dashed line, and arrowheads indicate synaptic contacts between the RIM and its postsynaptic partner, the AVB neuron. Stars indicate neuromuscular junctions between the RIM and neck muscles and RMD/SMD motor neurons. The scale bar is 3 μm.

Fig 4. Model of the neural circuit for tyraminergic signaling in the neural escape response circuit that controls the coordination of head movements and locomotion in response to gentle anterior touch.

Tyramine release from the RIM (blue) activates LGC-55 anion channel, which is expressed in the neck muscles, RMD/SMD motor neurons, and the AVB forward premotor interneuron (purple). Hyperpolarization of the neck muscles and RMD/SMD motor neurons induces neck relaxation and the suppression of head movements; hyperpolarization of the AVB forward premotor interneuron promotes backward locomotion. Tyramine signaling is induced through activation of the anterior touch sensory neurons (ALM/AVM), which activate premotor interneurons (AVD/AVA) that drive backward locomotion and are electrically coupled to the RIM (AVA-RIM). Sensory neurons are shown as triangles, premotor interneurons required for locomotion as hexagons, motor neurons as circles, and muscles as an oval.

Fig 5. Exogenous tyramine induces long forward runs and neck contractions in animals that express the LGC-55 cation.

(A) Top: still images of the locomotion pattern of transgenic animals expressing LGC-55 anion and LGC-55 cation-II prior to immobilization on 30 mM tyramine. The x marks the starting location, and the dashed red line indicates the forward locomotion, while the dashed blue line indicates backward locomotion. Bottom: still images of animals that express the LGC-55 anion or LGC-55 cation-II after five min on exogenous tyramine. The arrow-headed line indicates head length. Animals that express the LGC-55 anion exhibit a relaxation of the head muscles causing an elongation of the neck, while the expression of the LGC-55 cation channel causes contraction of the head muscles and a shortening of the neck. Scale bar, 0.2 μm. (B) LGC-55 cation transgenic animals hypercontract their neck on exogenous tyramine. Shown is the quantification of head lengths on exogenous tyramine. The length of the neck was measured from the posterior of the pharynx to the tip of the nose (inset) after 5 min on 30 mM tyramine (dark grey bars) or 0 mM tyramine (light grey bars) of wild type, n = 68; Plgc-55::LGC-55 (lgc-55(tm2913); zfEx2), n = 75; lgc-55(tm2913), n = 65; Plgc-55::LGC-55 cation-I (lgc-55(tm2913); zfEx8), n = 49; Plgc-55::LGC-55 cation-II (lgc-55(tm2913); zfEx40), n = 49. The error bars represent SEM. Statistical difference as indicated; *** p < 0.0001, two-tailed Student’s t test. (C) LGC-55 cation animals immobilize more slowly on exogenous tyramine. Shown is the percentage of animals immobilized by tyramine each minute for 20 min. Each data point is the mean +/- SEM for at least four trials totaling 40 or more animals. (D) LGC-55 cation animals make long forward runs on exogenous tyramine. Shown is the number of backward (dark grey bars) and forward (light grey bars) body bends made before paralysis on 30 mM tyramine of wild type, n = 40; Plgc-55::LGC-55 (lgc-55(tm2913); zfEx2), n = 29; lgc-55(tm2913), n = 34; Plgc-55::LGC-55 cation-I (lgc-55(tm2913); zfEx8), n = 28; Plgc-55::LGC-55 cation-II (lgc-55(tm2913); zfEx40), n = 39. Error bars represent SEM. Statistical difference from anion, ** p < 0.001, *** p < 0.0001, two-tailed Student’s t test.

Fig 6. A switch in LGC-55 ion selectivity reverses behavioral output.

(A) Touch induces neck relaxation in LGC-55 anion and contraction in LGC-55 cation transgenic animals. Still images of the animal’s head before (top) and after (bottom) touch stimulus. Scale bar, 0.1 mm. Arrow indicates neck length. (B) As measured in A, neck length from posterior of the pharynx to the tip of the nose before (light gray bars) and after (dark gray bars) anterior touch of wild type (n = 39); lgc-55(tm2913) (n = 32); Plgc-55::LGC-55 cation-I (lgc-55(tm2913); zfEx8), n = 32; Plgc-55::LGC-55 cation-II (lgc-55(tm2913); zfEx40), n = 26. Analyses were performed in an unc-3 mutant background to prevent backward locomotion in response to touch and to maintain the animal in the field of view at high magnification. Error bars represent SEM. Statistical difference as indicated, ** p < 0.001, *** p < 0.0001, two-tailed Student’s t test. (C) LGC-55 cation animals fail to execute a long reversal in response to touch. Shown is the average number of backward body bends in response to anterior touch of wild type n = 100; Plgc-55::lgc-55 (lgc-55(tm2913); zfEx2), n = 100, lgc-55(tm2913), n = 100; Plgc-55:: LGC-55 cation-I (lgc-55(tm2913); zfEx8), n = 100; Plgc-55::LGC-55 cation-II (lgc-55(tm2913); zfEx40), n = 100. Error bars represent SEM. Statistical difference from anion, * p < 0.01, *** p < 0.0001, two-tailed Student’s t test. (D) Tyramine release from the RIM activates the LGC-55 cation channel. Shown is the length of the neck before (light gray bars) and after (dark grey bars) exposure to blue light in retinal fed animals expressing the light-activated cation channel, ChannelRhodopsin 2 (ChR2), in the RIM in a wild-type background (Ptdc-1::ChR2(zfIs9), n = 28); TA deficient (tdc-1(n3420); Ptdc-1::ChR2(zfIs9), n = 25); receptor deficient (lgc-55(tm2913); Ptdc-1::ChR2(zfIs9), n = 28; LGC-55 cation-II (lgc-55(tm2913); Plgc-55::LGC-55 cation-II; Ptdc-1::ChR2 (zfEx213), n = 20); TA deficient; LGC-55 cation-II (tdc-1(n3420); lgc-55(tm2913); Plgc-55::LGC-55 cation-II; Ptdc-1::ChR2(zfEx275), n = 16) animals. Analyses were performed in an unc-3 mutant background. Blue light causes activation of the RIM and release of tyramine. Tyraminergic activation of the LGC-55 anion causes a relaxation of the neck muscles, while activation of LGC-55 cation-II causes a hypercontraction of the neck muscles. There is no response in animals that are raised on plates without all-trans retinal. Error bars represent SEM. Statistical difference as indicated, ** p < 0.001, *** p < 0.0001, two-tailed Student’s t test.

Fig 7. Phylogenetic comparisons of ion channel domains of members of the Cys-loop family of LGICs.

LGIC phylogenetic comparison was performed on the ion channel domains of human and invertebrate LGICs. The neurotransmitter identities are indicated on the right. Blue shading indicates anionic channels, while red shading indicates cationic channels. Ce, C. elegans; Ls, Lymnaea stagnalis; Hs, Homo sapiens. Protein alignments were performed with ClustalW [40]. Phylogenetic analysis was performed using the neighbor joining method and midpoint rooted. Alignments and phylogenetic analyses were carried out using MacVector Software (Accelrys). GenBank accession number for the sequences used are as follows: LGC-55 Ce, NM_075469; ACC-1 Ce, NM_069314; ACC-3 Ce, NM_076409; LGC-53 Ce, NM_171813; MOD-1 Ce, N_741580; UNC-49 Ce, NM_001027610; EXP-1 Ce, NP_495229; LGC-35 Ce, NM_001027268; ACR-16 Ce, NM_001028676; UNC-29 Ce, NM_09998; UNC-38 Ce, NM_059071; GABAα Ls, X58638; GABAζ Ls, X71357; AchF Ls, DQ167349; AchI Ls, DQ167352; AchB Ls, DQ167345; AchK Ls, DQ167353; AchH Ls, DQ167351; AchG Ls, DQ167350; AchD Ls, DQ167347; AchA Ls, DQ167344; AchJ Ls, DQ167348; AchC Ls, DQ167344; AchE Ls, DQ167348; GABAθ Hs, NP_061028; GABAγ Hs, NP_775807; GABAα1 Hs, NP_000797; GABAρ Hs, NP_002033; GABAβ Hs, NP_000803; ACHα9 Hs, NP_060051; ACHα7 Hs, P36544; ACHδ Hs, NP_000742; ACHβ Hs, NP_000738; 5HT3A Hs, AAH04453; 5HT3B Hs, AAH46990.