Characterization of a Choline-Gated Chloride Channel (LGC-40) from Haemonchus contortus Highlights a Novel Cholinergic Binding Site

Abstract

Nematode cys-loop ligand-gated chloride channels are important targets for anthelmintic drugs as many subtypes are not present in mammals. We report the isolation and functional characterization of a novel member of the cys-loop ligand-gated chloride channel family (Hco-LGC-40) from the parasitic nematode Haemonchus contortus. Electrophysiological analysis of the channel expressed in Xenopus oocytes revealed that it responds to both acetylcholine and choline with EC₅₀ values in the low micromolar range. Hco-LGC-40 can also associate with a member of a separate family of acetylcholine-gated chloride channels, Hco-ACC-1, to produce a heteromeric channel with a lower sensitivity to acetylcholine. In silico analysis reveals several residues that may play a role in forming a unique cholinergic binding pocket. Overall, these results suggest that choline may act as a neurotransmitter by binding to cys-loop receptors in parasitic nematodes.

1. Introduction

Cys-loop (cystine-loop) ligand-gated chloride channels (LGCCs) are a group of ion channel receptors that are activated when bound to ligands (e.g., neurotransmitters). They are made up of five protein subunits that interact to form functional channels. In the Caenorhabditis elegans and Haemonchus contortus genomes, there are several groups of cys-loop ligand-gated chloride channel receptors that respond to a diverse set of neurotransmitters and are targets for well known antiparasitics. For example, one well-characterized group, the AVR-14 group, represents several different subunits that form glutamate-gated chloride channels and are targets for macrocyclic lactone anthelmintics such as ivermectin [1]. Other groups include several subunits that form chloride channels that respond to a diverse set of neurotransmitters such as serotonin and dopamine (MOD-1 Group), GABA (UNC-49 Group) and acetylcholine (ACC-1 Group) [2,3,4]. As a whole, these groups have been relatively well characterized. However, there are other groups of cys-loop receptors where functional information is relatively limited. One such group is the LGC-57 group which appears to contain subunits that form unique types of acetylcholine-gated chloride channels [5,6]. We have recently shown that one member, LGC-39, forms a homomeric channel that is activated by several cholinergic agonists including acetylcholine and the muscarinic receptor antagonist atropine [5]. These findings highlight the diversity of nematode cys-loop receptors that exhibit unique ligand-binding pockets.

In an effort to characterize additional cys-loop ligand-gated chloride channels in parasitic nematodes that respond to cholinergic agonists, we report the isolation of LGC-40 from H. contortus (Hco-LGC-40). The channel has a high sensitivity to choline, suggesting that choline is a neurotransmitter or neuromodulator in H. contortus that may act at LGC-40. Analysis of the parasitic nematode genomes indicates that LGC-40 is present in clade C, III, IV and V genomes, but we cannot confirm its presence in clade I and II genomes. Compared to LGC-40 from C. elegans, Hco-LGC-40 is 3× more sensitive to choline and 24× more sensitive to acetylcholine. In silico analysis has identified several binding pocket residues that may directly play a role in forming a novel cholinergic binding pocket.

2. Methods

2.1. Cloning of Hco-Lgc-40 and Sequence Analysis

Total RNA was isolated from H. contortus adult females using Trizol (Invitrogen, Carlsbad, CA, USA). Complementary DNA (cDNA) was synthesized using the Quantitect Reverse Transcriptase kit from Qiagen (Dusseldorf, Germany) with a unique 3′ oligo-dT anchor primer sequence:

(5′CCTCTGAAGGTTCACGGATCCACATCTAGATTTTTTTTTTTTTTTTTTVN3′)

[where V is either A, C, or G and N is either A, C, G, or T] [7]. Amplification of the hco-lgc-40 complete sequence was conducted using two sets of primers targeting the full- length coding sequence. Amplification of the full-length coding sequence was performed using primers that targeted the 5′ and 3′ ends of the gene with added sequences for BamHI and XbaI restriction sites to facilitate cloning into the pGEMHE Xenopus laevis expression vector. This vector was then linearized and used as a template for in vitro cRNA synthesis, conducted using the mMESSAGE mMACHINE T7 Transcription Kit (Ambion, Austin, TX, USA). A Phylogenetic tree was constructed using PhyML 3.0 and a tree was constructed using FigTree 1.4.4.

2.2. Expression of Hco-LGC-40 and Electrophysiological Analysis

All animal procedures followed the Ontario Tech University Animal Care Committee (Animal Use Protocol # 14290, approval date 18 March 2018) and the Canadian Council on Animal Care guidelines and oocyte procedures according to [8]. Xenopus laevis female frogs were supplied by Nasco (Fort Atkinson, WI, USA). The frogs were kept in a climate-controlled, light-cycled room and stored in regularly cleaned tanks.

X. laevis were anesthetized with 0.15% 3-aminobenzoic acid ethyl ester methanesulphonate salt (MS-222) buffered with NaHCO₃ to pH 7 (Sigma-Aldrich, Oakville, ON, Canada). During surgery, a section of the frog ovary was removed and the lobe was defolliculated with OR-2, calcium-free oocyte Ringers solution [82 mM NaCl, 2 mM KCl, 1 mM MgCl₂, 5 mM HEPES pH 7.5 (Sigma-Aldrich)] and 2 mg/mL collagenase-II (Sigma-Aldrich). The oocytes were incubated in the defolliculation solution for 2 h at room temperature under agitation. Collagenase was washed from the oocytes with ND-96 solution (96 mM NaCl, 2 mM KCl, 1 mM MgCl₂, 1.8 mM CaCl₂) and oocytes were incubated in ND96 supplemented with 0.275 µg/mL pyruvate and 50 µg/mL gentamycin for one hour at 18 °C prior to injection.

X. laevis oocytes were injected with 50 nL cRNA encoding either a single subunit, to test the presence of a homomeric channel, or two subunits, to test the presence of a heteromeric channel. The injected oocytes were incubated at 18 °C in a supplemented ND96 solution. The oocytes were incubated for a minimum of 48 h and a maximum of 72 h following injection and supplemented ND96 solution was changed every 24 h until electrophysiology was performed.

Two-electrode voltage clamp (TEVC) electrophysiology was performed using the Axoclamp 900A voltage clamp (Molecular Devices, Sunnyvale, CA, USA). Glass electrodes were made using a P-97 Micropipette Puller (Sutter Instrument Co. Novato, CA, USA). The electrodes were backfilled with 3M KCl and contained Ag|AgCl wires, which connected the electrodes to their respective Axon Instrument Headstage (Molecular Devices). Oocytes were clamped at a holding potential of −60 mV and a second electrode was used to measure current changes during channel activation.

Solutions containing acetylcholine and choline were dissolved in non-supplemented ND96 and were washed over injected oocytes. Channel activation was observed as the presence of a current. Before testing of each compound, oocytes were washed with non-supplemented ND96 until the stable resting membrane potential was reached. Washing of solution over the oocytes was accomplished using the RC-1Z perfusion chamber (Warner Instruments, Holliston, MA, USA) and a Fisherbrand Variable-Flow Peristaltic Pump (Fisher Scientific, Hampton, NH, USA) to remove liquid waste. Responses to each ligand (e.g., choline) at increasing concentrations were recorded in individual oocytes, and dose–response curves were generated using GraphPad Prism (GraphPad Software v5.0, San Diego, CA, USA) with data fit to the following equation:

$$I_{max}={\displaystyle \frac{1}{1+{\left(\frac{{EC}_{50}}{\left[D\right]}\right)}^h}}$$

In this equation, $I_{max}$ is the maximal response, EC₅₀ is the concentration of the agonist that produces 50% of the maximal response, [D] is the agonist concentration, and h is the hill coefficient.

To confirm that Hco-LGC-40 is a chloride channel, current–voltage relationships were recorded. This was performed by changing the holding potential from −40 mV to +40 mV in 20 mV increments. At each 20 mV increase, the oocyte was exposed to the EC₅₀ concentration of the responding ligand. These tests were performed using ND96 as well as reduced Cl⁻ ND96, where NaCl was partially replaced with Na-gluconate (Sigma), creating a Cl⁻ concentration of 62.5 mM. Current–voltage graphs were generated using Graphpad Prism Software v5.0 (San Diego, CA, USA).

2.3. Modelling of the Hco-LGC-40 Binding Pocket

To further analyze the structure of the Hco-LGC-40 receptor and how choline binds within the binding pocket, a homodimer model was generated. The template used to model the homodimers was the Danio rerio alpha-1 glycine receptor (3JAD), as it was the template with the highest sequence identity. The chosen template was used to generate a homodimer model in MODELLER v9.20 using automated scripts. The most energetically favoured model was chosen based on the DOPE score and selected for computational agonist docking.

Choline, in its energy reduced form, was downloaded from the Zinc database http://zinc.docking.org. Homodimer models, previously created using MODELLER [9], were prepared for ligand docking using AutoDock tools [10]. AutoDock Vina was used to simulate docking of a ligand in the binding site of the homodimers [11]. Pymol was used to visualize the docked ligand on its corresponding homodimer, and Chimera v1.6.1 [12] was used to determine the distance between the amino acid residues and the choline ligand.

3. Results and Discussion

3.1. Cloning of Hco-Lgc-40 and Sequence Analysis

Amplification of the complete hco-lgc-40 sequence yielded a 1422 base pair sequence, submitted to GenBank (accession number QCU71395.1), which encodes for a protein 473 amino acids in length. This protein has a 73% similarity with the Cel-LGC-40 protein and contains the characteristic cys-loop and seven binding loops in the extracellular domain as well as four transmembrane domains. LGC-40 sequences were also found in nematodes from clade III (Toxocara canis), clade IV (Steinernema carpocapsae) and clade C (Plectus sambesii) but not found in nematodes from clades I and II when searched in the WormBase Parasite Database (https://parasite.wormbase.org) (Figure 1). We did find two related subunits from the clade I nematode Mesodorylaimus sp. and clade II nematode Trissonchulus latispiculum that groups with the LGC-57 and 58 subunits. However, at this time, we have not designated them a formal name. We were not able to find LGC-40 sequences in another clade III nematode, Brugia malayi, whose genome is also sequenced. It would be interesting to conduct a more comprehensive examination of the LGCC family in clade III nematodes to determine whether there are other differences in the repertoire of cys-loop receptor gene sequences.

3.2. Expression of Hco-LGC-40 and Electrophysiological Analysis

We expressed Hco-LGC-40 in Xenopus laevis oocytes and analyzed for channel activity using similar methods as [5] and found that choline and acetylcholine produced currents. The channel responded in a dose-dependent manner to both molecules (Figure 2A). The choline EC₅₀ value for Hco-LGC-40 was 2.3 ± 0.37 µM (n = 6) with a Hill coefficient of 2.8 ± 0.12, which suggests that more than one choline molecule is required to open the channel. The acetylcholine EC₅₀ value for Hco-LGC-40 is 4.9 ± 0.24 µM (n = 4) with a Hill coefficient of 3.99 ± 0.88, suggesting the binding of at least two acetylcholine molecules for channel activation (Figure 2B). Current–voltage analysis suggests that Hco-LGC-40 is a chloride channel (Figure 2C).

The LGC-40 homomeric channel (Cel-LGC-40) was recently characterized from the free-living nematode C. elegans using a similar methodology as presented here [6]. Overall, we have found that compared to Cel-LGC-40, Hco-LGC-40 is 6.5× more sensitive to choline (EC₅₀ 2.3 vs. 15 µM) and 23× more sensitive to acetylcholine (EC₅₀ 4.9 vs. 115 µM). It is possible that the differences in ligand sensitivities (particularly acetylcholine) of LGC-40 between H. contortus and C. elegans are linked to functional differences between species. In C. elegans, LGC-40 is expressed in some pharyngeal neurons at possible extra synaptic sites. The high sensitivity of this channel to choline suggests that choline (a metabolite of acetylcholine) is a bonafide in vivo neurotransmitter that acts as a neuromodulator [6]. However, at this time, the exact function of LGC-40 in parasitic nematodes is unknown. Indeed, we have found other examples of functional differences between cys-loop receptor orthologues from C. elegans and H. contortus. For example, ACC-1 from C. elegans can form functional homomeric channels in Xenopus oocytes and is expressed in pharyngeal neurons. In contrast, ACC-1 from H. contortus cannot form functional homomeric channels in oocytes and is expressed in pharyngeal muscle [14].

An unexpected result from this study is that Hco-LGC-40 also co-assembles with a the ACh-gated LGCC, Hco-ACC-1, to produce a channel with a lower sensitivity to ACh [EC₅₀ of 25 ± 2 µM]. Hco-ACC-1 is an ACh-gated LGCC from a different subfamily as compared to Hco-LGC-40. We have previously shown that Hco-ACC-1 can also co-assemble with two other members of the ACC-1 subfamily, ACC-2 and LGC-46, to produce channels with increased sensitivity to ACh [14,15]. The assembly of LGC-40 with ACC-1 suggests that ACC-1 may be able to co-assemble, at least in the oocyte expression system, with a variety of cys-loop receptor subunits. Whether this may occur in vivo will be an interesting avenue of future research.

3.3. Modelling of the Hco-LGC-40 Binding Pocket

Analysis of the choline binding pocket can provide some clues on the key determinants for agonist recognition. We generated a model of the Hco-LGC-40 receptor and docked choline into the predicted binding pocket. Figure 3 shows the residues within 5 Å of the choline molecule. The binding pocket has the signature tryptophan in loop C, which is also found several unique types of cys-loop receptors such as serotonin, tyramine, dopamine and acetylcholine-gated chloride channels from the ACC-1 family. We have previously shown that this residue in the ACh-gated LGCC, Hco-ACC-2, is absolutely required for channel function [16]. This tryptophan is also present in Hco-LGC-40. However, while Hco-ACC-2 and LGC-40 are both cholinergic receptors, they have very different sensitivities to choline, with Hco-ACC-2 being >500× less sensitive. Further comparison of the binding pocket residues does reveal some differences between the two receptors. One difference is the presence of a loop B tyrosine (Y) in LGC-40 with a phenylalanine (F) in the analogous position in ACC-2. We previously showed that changing this position from F to Y in ACC-2 increases the sensitivity of choline by 50%. Other differences include various loop E residues with LGC-40 having two tryptophan residues facing the binding pocket with the analogous positions in ACC-2 having methionine. Given the importance of aromatic residues in cys-loop receptor binding pockets, it is likely that the two tryptophan residues in loop E are playing an important role in agonist recognition in LGC-40. Interestingly, there are two other cys-loop receptors (LGC-57 and 58) that also have a high sensitivity to choline. Both of these receptors share many of the same binding site residues as LGC-40 (Figure 3).

In conclusion, we have isolated and characterized an additional cys-loop receptor from H. contortus. This receptor appears to be a novel cholinergic receptor subtype that has a high affinity for choline and acetylcholine. Whether choline functions as a bonafide neurotransmitter in parasitic nematodes will be an interesting avenue for future research.