Nicotinic acetylcholine receptors (nAChRs) are pentameric ligand-gated ion channels that play crucial roles in neurotransmission and regulate complex processes in brain functions, including anxiety, learning and memory, food intake, drug addiction, cognition and nociception. To perform these and other functions, a diverse array of nAChR subtypes are generated by homomeric or heteromeric assembly of 17 homologous nAChR subunits. Agonists, acetylcholine and nicotine, bind to the interface formed between two α subunits and between α and non-α subunits to activate the nAChR and allow cation influx. The diversity of subunit interfaces determines the channel properties, the responses to different agonists/antagonists, desensitization and downstream signaling and thus, define specialized properties and functions. Over the last several decades, snake venom neurotoxins have contributed to the purification, localization and characterization of molecular details of various nAChRs. Utkin et al. have described the purification and characterization of αδ-bungarotoxins, a novel class of neurotoxins in a recent paper published in the Biochemical Journal [Biochem. J. (2019) 476, 1285–1302]. These toxins from Bungarus candidus venom preferably bind to α–δ site with two orders of magnitude higher affinity compared with α–γ or α–ε sites. The subtle changes in the structure of αδ-bungarotoxins led to variation in interface selectivity. Such new classes of antagonists will offer us great opportunity to delineate the pharmacophores and design new highly selective antagonists. Thus, their findings provide new impetus to re-evaluate molecular details of pharmacological properties of α-neurotoxins with careful consideration towards subtype-, interface- and species-selectivity.
Toxins cause death and debilitation and hence, they are thought as villains. In reality, they have made more contributions to improving our lives than causing death. Because of their high affinity, exquisite specificity and selectivity, toxins have played crucial roles in the discovery and development of therapeutic and diagnostic agents for human diseases. They have also served as research tools to understand molecular mechanisms of normal physiological processes and to probe the structure of the target receptor, ion channel or enzyme.
Snake venom toxins have facilitated our understanding of neurotransmission. In 1963, the first neurotoxin, α-bungarotoxin, was isolated from the venom of Taiwanese many-banded krait (Bungarus multicinctus) and characterized . This postsynaptic toxin belongs to the three-finger toxin family and irreversibly binds to peripheral muscle nicotinic acetylcholine receptor (nAChR) (Kd ∼ 10−9 to 10−11 M) and neuronal α7, α9 and α9α10 nAChRs (Kd ∼ 10−8 to 10−9 M). α-Bungarotoxin enabled the isolation of nAChR, as the bait in the affinity chromatography from the electric organ of the electric eel and other sources . Subsequently, α-bungarotoxin and related curare-mimetic neurotoxins were used as ideal ligands to probe the structure and localization of various subtypes of nAChRs and evaluate molecular interaction with agonists and antagonists.
nAChRs are a superfamily of pentameric ligand-gated ion channels that act as Nature's ultimate molecular sensors and convert chemical signals of acetylcholine (ACh) into electrical impulses of cation influx. In vertebrates, these Cys-loop receptors are formed by homomeric (only α7 and α9 subunits) or heteromeric assembly of 17 homologous nAChR subunits (α1–α10, β1–β4, δ, γ and ε) around a central ion-conducting pore. Loop C of only α subunits have cysteines 192–193 located near the entrance to transmembrane helix 1 (TM1) that play an important role in ACh-binding; the other subunits lack these cysteines [3,4]. The peripheral muscle-type nAChR, expressed in the neuromuscular junction, comprises two α1 subunits and one each of the β1, δ and γ/ε subunits (γ in fetal and ε in adult) in a clockwise α1–β1–δ–α1–γ/ε arrangement (when viewed from the extracellular space). ACh-binding sites are located at α1–δ and α1–γ/ε subunit interfaces of the extracellular domains. The neuronal nAChRs, expressed in the brain, are either heteropentamers comprising of great diverse combination of two or three types of subunits (Figure 1) or homopentamers (α7) [3,5]. In these receptors, ACh-binding sites are located at α–β and α–α interfaces, respectively. The subunits (α5, β3, α3, α4, β2 and β4), which do not directly participate in ACh-binding, are identified as accessory subunits. These accessory subunits alter the pharmacological and biophysical properties, the sensitivity to allosteric modulators and the sensitivity to up-regulation of the nAChRs .
Schematic representation of distinct subunit interfaces in nAChRs.
The hydrophobic agonist-binding sites, at the interface between adjacent subunits, are formed by six noncontiguous regions (loops A–F) three loops each from both subunits. The α subunit, which acts as the principal (+) component, contributes three loops of highly conserved residues (loops A–C), while a complementary (–) component of the adjacent, mostly non-α subunit contributes three loops (D–F) that have lower levels of sequence conservation between subunits . Consequently, the two components of the agonist-binding sites are nonequivalent, and the loops contribute differently to receptor function . Several classes of antagonists, evolved in animal venoms, bind to these agonist-binding sites and compete with the agonist to block the neurotransmission by various nAChR subtypes. Thus, the distinct microenvironment at the agonist-binding interfaces defines affinities and selectivity towards various agonists and antagonists . In this issue of Biochemical Journal, Utkin et al.  have described a new class of α-neurotoxins that can distinguish two ACh-binding sites in the Torpedo californica and mouse muscle nAChRs and preferably binds to α–δ site with two orders of magnitude higher affinity compared with α–γ or α–ε sites. They identified three isoforms (αδ-BgTx-1-3) in the Malayan Krait (Bungarus candidus) by genomic DNA analysis and isolated two of them (αδ-BgTx-1 and αδ-BgTx-2) from the venom. This group of long-chain α-neurotoxins, named as αδ-bungarotoxins, is structurally and functionally similar to α-bungarotoxin. In contrast, typical α-neurotoxins, such as α-bungarotoxin (long-chain or Type I), erabutoxin (short-chain or Type II) and other related toxins, bind to α–δ, α–γ or α–ε sites with equal affinity and cannot distinguish them . Therefore, structure–function relationships of αδ-bungarotoxins will help us identify specific residues that contribute to interface selectivity towards α–δ site.
nAChRs are fast cationic receptors that play crucial physiological roles in neurotransmission and in regulatory processes (cell excitability and neuronal integration) and in complex brain functions (anxiety, learning and memory, food intake, nicotine addiction, cognition, nociception and attention) [10,11]. Their dysfunction leads to many neurological disorders including Alzheimer's disease, Parkinson's disease, autism, schizophrenia and neuropathic pain [11,12]. nAChRs play an important role in the addiction circuitry for all drugs of abuse . nAChRs facilitate cell adhesion and proliferation, angiogenesis in endothelial cells and cholinergic modulation of immune cells [14,15]. Therefore, nAChRs are targets for cancer and inflammatory diseases such as Crohn's disease, epilepsy, rheumatoid arthritis and ulcerative colitis [16,17].
Over 100 neuroanatomical regions are described in the human brain. Each region has its own physiological role, and precise assembly and distribution of various nAChR subtypes contribute to pre-, post- and extra-synaptic signaling that regulate various physiological processes. To develop highly selective research tools and therapeutic agents, the following three factors should be considered carefully.
Subunit composition: The diversity of nAChR subtypes determines their specialized properties and functions. In mammals, in addition to muscle-type nAChRs, wide varieties of nAChR subtypes are formed by diverse combinations of 11 neuronal subunits (all but α8) (Figure 1). Various combinations of subunits alter the opening and closing kinetics, single-channel properties, desensitization, the responses to different agonists and antagonists and downstream signaling. The differences in subunit interfaces determine the characteristics of ligand-binding sites . They also alter the cation permeability (from monovalent Na+ and K+ ions to divalent Ca2+ ions). Ca2+ ions, in turn, affect signal transduction. The differences in the extracellular domains may affect their modulation by external Ca2+ and Zn2+ cations. Similarly, the diversity of their cytoplasmic domains contributes to further diversity in their interactions with the cytoplasmic machinery. The expression of various subunits and nAChR assembly mechanisms and checkpoints show high fidelity in the brain; the α4, α7 and β2 subunits interact with each other to form functional nAChRs in heterologous systems but in hippocampal neurons they mostly combine to form α4β2 and α7 nAChRs, while the α3, α4, β2 and β4 subunits show preference to form mostly α3β4 and α4β2 receptors in the brain and ganglia . The rich diversity of nAChRs in human brain makes it difficult to determine finer aspects of their composition, stoichiometry and function. Although each subtype exhibits unique functional properties, there is sufficient overlap when the subtypes share common subunits or subunits with a high similarity. The composition and stoichiometry of significant number of receptors are unclear and they are represented with ‘*’ as in α3β4* to indicate this fact. At times, nAChRs are associated with proteins, such as rapsyn, which affect their function .
Localization: Distinct functional properties of nAChR subtypes support various physiological processes in respective tissues and hence, fit their roles. The muscle-type nAChRs are localized opposite motor nerve terminals (endplate) at a density of ∼10 000 receptors/μm2 . To understand the involvement of specific nAChR subtypes in normal physiology and in disease states and to develop interventional strategies to treat several neurological diseases, it is vital to define the precise localization of various neuronal nAChR subtypes. Although the regional distribution of neuronal nAChRs in the rodent brain has been established to an extent, the data on human brain are scanty. The differential distribution of neuronal nAChR subtypes over the cell surface (cell soma, dendrites, preterminal axon regions and axon terminals), precise cytoplasmic regulatory mechanisms, ontogenetic changes in the development and aging are not yet properly established. The impact of subunit composition and stoichiometry is only beginning to be understood. Until recently, α7 subunits were thought to form only homomeric receptors. Recent studies have shown that the heteromeric nAChRs formed by α7 subunits with α5, β2, β3 or β4 subunits exhibit altered pharmacological properties . Interestingly, α7β2 subtype is highly sensitive to inhibition by oligomeric forms of amyloid Aβ1–42 and may be relevant in Alzheimer's disease . Similarly, incorporation of accessory subunit α5 alters the properties of α3β2 and α4β2 nAChRs .
Compartmentalization and accessibility: It is important to understand the distribution of the newly designed agonists and antagonists through various routes of administration. This will help in understanding the compartmentalization and access to relevant receptors in healthy and afflicted animal (preferably primate) models.
Snake venoms are a rich source of antagonists of nAChRs. Both long-chain and short-chain neurotoxins bind to the muscle-type nAChRs. They exhibit species-specific reversibility: their binding to rat, mouse and Torpedo AChRs is nearly irreversible; their binding to human and hedgehog AChRs is slowly reversible, while their binding to snake and mongoose AChRs is negligible. Only long-chain neurotoxins bind to α7, α9 and α9α10 nAChRs. Thus, these groups of neurotoxins exhibit clear subtype selectivity. κ-Bungarotoxins, although structurally similar to long-chain neurotoxins, do not bind to the muscle-type, α7, α9 and α9α10 nAChRs, but selectively block nAChRs containing α3 and β4/β2 subunits. The β subunit appears to control their specificity; β2-containing AChRs are less sensitive than β4-containing nAChRs. Denmotoxin, irditoxin and fulgimotoxin from colubrid venoms exhibit excellent species-specific neurotoxicity [24–26]. Waglerin, a 22-residue and single disulfide peptide, binds selectively (3700-fold higher affinity) to α–ε site compared with α–γ site and fails to exert neurotoxicity in ε subunit knockout mice [27,28]. This high selectivity is due to its interactions with charged and hydrophobic residues from the complementary (–) face . Azemiopsin, a 21-residue peptide with no cysteines, shows slightly increased affinity towards adult muscle-type nAChR with α–ε site compared with fetal nAChR . α-Conotoxins from Conus venoms also exhibit subtype and interface selectivity. α-Conotoxin MI selectively (∼10 000-fold) binds the α–δ site compared with the α–γ site . Interestingly, loop G at the complementary (–) face is involved in its binding to α-conotoxin MI . An antagonist bound to one subunit interface is enough to block the rotational torque and channel opening.
In this issue, Utkin et al.  show that subtle changes in the structure of α-neurotoxins led to variation in interface selectivity. Their data provide new impetus to re-evaluate molecular details of pharmacological properties of α-neurotoxins with careful consideration towards subtype-, interface- and species-selectivity. Such new classes of antagonists will offer us great opportunity to delineate the pharmacophores and design new highly selective antagonists. Equipped with the mechanistic details of the interaction of these antagonists with various nAChRs and the intent to convert them as agonists will lead to the design of highly selective agonists. Overall, these new classes of α-neurotoxins will contribute to the next generation of nAChR ligands that are useful in the development of novel research tools and therapeutic agents.
This work is supported by Ministry of Education grant MOE2017-T2-1-045 to R.M.K.
I thank Cho Yeow Koh for preparing the figure.
The Author declares that there are no competing interests associated with this manuscript.