Venom-Derived Peptide Modulators of Cation-Selective Channels: Friend, Foe or Frenemy

Ion channels play a key role in our body to regulate homeostasis and conduct electrical signals. With the help of advances in structural biology, as well as the discovery of numerous channel modulators derived from animal toxins, we are moving toward a better understanding of the function and mode of action of ion channels. Their ubiquitous tissue distribution and the physiological relevancies of their opening and closing suggest that cation channels are particularly attractive drug targets, and years of research has revealed a variety of natural toxins that bind to these channels and alter their function. In this review, we provide an introductory overview of the major cation ion channels: potassium channels, sodium channels and calcium channels, describe their venom-derived peptide modulators, and how these peptides provide great research and therapeutic value to both basic and translational medical research.


INTRODUCTION
Cone snails whose shells are coveted for their elaborate patterns, yellow dart frogs measuring just a few centimeters long, and transparent bell-shaped jellyfish with delicate tentacles might all seem unlikely candidates, but they are among the deadliest animals in the world. Like the more obvious perilous creatures -venomous snakes, spiders and scorpions -these animals release toxins that dramatically modulate the activity of various targets including ion channels, thereby affecting cellular communication and disrupting normal biochemical and physiological processes in prey or predator.
Animal venom is a complex mixture of various components -inorganic salts, organic molecules like alkaloids, proteins and peptides (King, 2011). While this concoction enables a multi-pronged attack upon the target organism, it has also led to an entire collection of bio-active compounds being available to researchers for probing the structural and functional properties of their molecular targets. Since ion channels play an essential role in neuronal signaling and muscle contractions, it is unsurprising that many venom toxins have evolved to block or modulate the function of ion channels (Dutertre and Lewis, 2010). Not only have venom-derived peptides been used extensively in probing ion channels, the understanding of the mechanism of this interaction has also led to the development of venom-based therapeutics targeting various ion channels. In fact, the recognition of animal venom having medicinal benefits is not a recent phenomenon. Venom from various animals had been used as medicines for centuries, in civilizations all over the world (Bhattacharjee and Bhattacharyya, 2014;Utkin, 2015).
Modern medicine has shown conclusively that venoms contain compounds with therapeutic potential. Many of these have been isolated, analyzed for structure and function, and have served as scaffolds for the development of various drugs. Venom peptides have evolved to be highly stable, being able to withstand degradation by proteolytic enzymes in the foreign environment they are  Kubo et al., 1993). As the name suggests, Kir channels inwardly rectify outward K + current, allowing extracellular K + to readily flow into the cells. The unique molecular mechanism is due to the intracellular binding of Mg 2+ and polyamines (Lu, 2004). Kir channels are homo-or hetero-tetrameric structures assembled from four Kir subunits, containing two TM segments separated by a selectivity filter region (Whorton and MacKinnon, 2011;Li et al., 2017). Structural, functional and pathophysiological details of four specific types of Kir channels have been detailed elsewhere (Hibino et al., 2010).
The peptides that show high affinity toward Kir channels (IC 50 < 0.5 µM) are scorpion toxin ChTx2 (α-KTx1.2), snake toxin δ-dendrotoxin (δ-DTX), and honey bee toxin Tertiapin (TPN) (Lu and MacKinnon, 1997;Imredy et al., 1998;Jin and Lu, 1998;Doupnik, 2017). Like many other venom toxins, these three molecules are rich in cysteine and positively charged residues. Computational simulation and docking studies have hypothesized binding mechanisms of these toxins (Li et al., 2016). Positively charged residues from toxin come into close contact with negatively charged residues on channel pore region, strengthening electrostatic interactions between the two. Hydrophobic forces between aliphatic residues also count into binding affinity.
TPN and TPN M13Q are considered the most potent inhibitors. TPN binds to Kir1.1 and Kir3.1/3.4 at 2-8 nM (EC 50 ), thus being an ideal tool for investigations of Kir channels' functional and pharmacological properties (Dobrev et al., 2005;Walsh, 2011). TPN has shown potential therapeutic use in a canine model, treating atrial fibrillation, without causing ventricle arrhythmia (Hashimoto et al., 2006). More recently, TPN, together with sodium channel blockers, has been shown to have synergistic effects in preventing atrial fibrillation and prolonging atrial effective refractory period. The combination formula has been patented for medication manufacturing by Gilead Sciences.
Voltage-gated potassium (K v ) channels tightly control membrane permeability of K + by sensing voltage change, thereby playing a key role in regulating action potential and propagating electrical signals in excitable cells (Yellen, 2002). In non-excitable cells, K v channels modulate cellular metabolism and facilitate downstream signaling cascade; for example, K v 1.3 in T lymphocytes (Cahalan and Chandy, 2009). 40 K v channels in 12 subfamilies have been found and many extensively studied (Alexander et al., 2017). K v channels are homo-or heterotetramers, made up of four subunits each consisting of six TM helices. Voltage sensing domain (VSD) (S1-S4) is connected to pore domain (S5-S6) through S4-S5 intracellular loop, driving the pore to open or close (Long et al., 2005).
Research on venom peptide modulators of K v channels started in 1980s, and to date more than 200 peptides with inhibitory effect on K v channels have been identified (Carbone et al., 1982;Kuzmenkov et al., 2016). These polypeptides usually bind to K v channels in two unique mechanisms. The pore blockers sit in the shallow vestibule at extracellular pore region, while the gating modifiers bind to the so-called "paddle motif " of the VSD accessible from the extracellular side.
Scorpion toxin charybdotoxin (ChTx) was one of the earliest venom toxins used as an important research tool to understand Frontiers in Pharmacology | www.frontiersin.org   Chuang et al., 1998;Sidach and Mintz, 2002 ProTx1 Thrixopelma pruriens 200 nM Ohkubo et al., 2010 Ca v 3.2 Kurtoxin Parabuthus transvaalicus 25-50 nM Chuang et al., 1998;Sidach and Mintz, 2002 K v channel subunit stoichiometry (MacKinnon, 1991), auxiliary beta subunits (Garcia et al., 1995), as well as its overall architecture (Hidalgo and MacKinnon, 1995). Sea anemone toxin ShK blocks K v channels at nanomolar to sub-nanomolar potency (Castañeda et al., 1995;Kalman et al., 1998). ShK and its analogs are blockers of the K v channel pore. They bind to all four subunits in the channel tetramer by two key interactions within the external vestibule -Lys22 occludes the channel pore like a "cork in a bottle, " and Tyr23, together with Lys22, forms a "functional dyad" required for channel block. Many K + channel-blocking peptides exhibit similar blocking mechanism, consisting of a dyad of lysine and neighboring aromatic/aliphatic residue (Chang et al., 2018). With the goal of developing a highly selective K v 1.3 inhibitor, nearly 10 years of effort was made to re-engineer the native ShK. In 2006, a stable analog, ShK-186 demonstrated specific binding to K v 1.3 at 69 pM, which is 100-fold selective to other K v channels . ShK-186 (Dalazatide), now being developed by Kineta, has passed phase I clinical trials It is the only venom-derived peptide blocking K + channels that is being developed as a therapeutic , Tarcha et al., 2017.
The hERG channel (or K v 11.1) plays a crucial role in the cardiac action potential by repolarizing IKr current, the rapid component of the delayed rectifier potassium current. While selective K v 11.1-blockers are available (e.g., BeKm-1 from scorpion Mesobuthus eupeus) (Korolkova et al., 2001), it warrants special attention as many drugs/peptides intended for other targets, can exhibit non-selective binding to it, with potentially fatal consequences. Inhibition of hERG by drugs can lead to lengthening of the electrocardiographic QT interval, while hERG channel activators can cause drug-induced short QT syndrome. Both cases can lead to potentially fatal arrhythmias. Hence, FDA guidelines recommend that all drugs that are intended for human use be evaluated for anti-hERG activity (Vandenberg et al., 2012).
Calcium (Ca 2+ )-activated potassium channels (K Ca ) are broadly divided into three subtypes based on their single channel conductance -big conductance (BK Ca ), intermediate conductance (IK Ca ) and small conductance (SK Ca ). While the BK Ca channels are activated by both voltage and increase in cytosolic Ca 2+ , the IK Ca and SK Ca channels are activated exclusively by the latter. Like Kir and K v channels, the K Ca channels are tetramers made up of four α subunits. BK Ca requires additional regulatory subunits, and is made up of 6/7 TM segments, while SK Ca and IK Ca contain 6 TM segments, with a calmodulin molecule bound to each subunit, serving as the Ca 2+ sensor. One of the first peptide toxins that were found to inhibit K + channels included apamin (derived from bee venom) and charybdotoxin (ChTX, derived from the scorpion venom) (Hugues et al., 1982;Rauer et al., 2000). Apamin blocks SK channels (K Ca 2), and served as a primary pharmacological tool to distinguish between K Ca 2 channels and K Ca 1.1/K Ca 3.1. ChTX inhibits both K Ca channels (K Ca 1.1 and K Ca 3.1) and K v channels (K v 1.2, K v 1.3, and K v 1.6). Another scorpion toxin iberiotoxin is selective for BK channel (K Ca 1.1) (Candia et al., 1992).

VENOM PEPTIDES TARGETING VOLTAGE-GATED SODIUM CHANNELS
Voltage-gated sodium (Na v ) channels are present in the membranes of most excitable cells and are responsible for initiation and propagation of action potentials. Studies elucidating details of ion selectivity, hypothesizing the Na v pore diameter and binding mechanism of sodium-channel acting local anesthetics and related drugs, were bolstered by the availability of ion channel toxins, like the alkaloids tetrodotoxin (TTX) and saxitoxin (STX) (Hille, 1971(Hille, , 1975(Hille, , 1977Armstrong et al., 1973). Studies to isolate and purify the Na v channel protein were pioneered by William Catterall and co-workers using, besides TTX and STX, scorpion toxin (ScTx) neuropeptides (Agnew et al., 1978;Beneski and Catterall, 1980;Hartshorne and Catterall, 1981).
Na v channels are divided into nine subtypes (Na v 1.1-Na v 1.9) based on their sequence, TTX binding and tissue expression. The 250 kDa channel-forming α-subunits are pseudo-tetrameric, wherein a single polypeptide chain folds into four homologous, non-identical domains (DI-IV), each containing six TM segments (S1-S6). The S5-S6 segments from all four domains form the central ion pore, while the S1-S4 segments in each domain form the VSD. A single channel is composed of one poreforming α subunit, which may be associated with either one or two β subunits. The α subunit is functional on its own, and forms the core of the channel.
The venom of various animals contain toxins that target Na v channels to attack the neuromuscular systems of their adversaries and prey. Toxins that modulate Na v channel function generally do so in two ways -either by blocking the flow of Na + ions through the pore, or by modifying the gating mechanisms.
One of the best studied pore blockers for Na v channels are the µ-conotoxin peptides from cone snails. Conotoxins are disulfiderich peptides that are isolated from the venom of cone snails (genus Conus). Venom derived from cone snails is a treasure trove of peptide toxins for different ion channels and other receptor proteins (Olivera et al., 1985(Olivera et al., , 1990. M-conotoxins demonstrate the best binding with the skeletal muscle isoform of Na v channel, Na v 1.4, with variable binding to other isoforms. These variations in targeting selectivity and affinity of each peptide for the different Na v isoforms constitute an important tool for distinguishing between different isoforms (Zhang et al., 2013). On the other hand are toxin peptides that modify Na v channel gating by interacting with the voltage sensors. Various classes of conotoxins interact with the voltage sensors of Na v channels and influence their gating properties. -conotoxins are ubiquitously expressed in a range of cone snail venoms and inhibit fast inactivation of channels. While the µ-conotoxins are pore-blocking peptides, the µO-conotoxins are gating modifiers that target the voltage sensors and inhibit channel opening (Daly et al., 2004;Zorn et al., 2006;Leipold et al., 2007). MO-conotoxins were evaluated for their pain-relieving activity and found to be anti-nociceptive in animal models of pain (Teichert et al., 2012).
Several spider toxins are in pre-clinical development stage as antagonists of Na v 1.7, an attractive target for development of non-opiod pain medication. Protoxin-II (ProTX-II), derived from the tarantula Thrixopelma pruriens, inhibits channel activation by shifting to positive potentials the voltage dependence of channel activation. Using ProTX-II as a scaffold, a highly potent and selective Na v 1.7 blocking peptide (JNJ63955918) has been developed, the effect of which mirrors features of the Na v 1.7-null phenotype (Flinspach et al., 2017). Another venom peptide, huwentoxin IV, is derived from the Chinese bird-eating spider Selenocosmia huwena (Peng et al., 2002). This peptide preferentially inhibits Na v 1.7 by binding one of the four VSDs of the channel, making it more selective as compared to the local anesthetics that bind the conserved channel pore (Ragsdale et al., 1996;Xiao et al., 2008Xiao et al., , 2011. Various mutational studies led to a triple mutant of huwentoxin IV (E1G, E4G, and Y33W) being developed with a very high potency toward Na v 1.7 blocking (Revell et al., 2013).
Na v 1.5 is expressed mainly in cardiac muscle, where it mediates fast depolarization phase of the cardiac action potential and is a target for class I anti-arrhythmic agents. Jingzhaotoxin-III (from the Chinese tarantula Chilobrachys jingzhao) selectively inhibits the activation of Na v 1.5 in heart cells (IC50 ∼ 350 nM), but not Na v neuronal subtypes (Rong et al., 2011).
Sea anemones are another source of Na v -targeting peptides. Some key toxins are ATX-II (from Anemonia sulcata), AFT-II (from Anthopleura fuscoviridis) and Bc-III (from Bunodosoma caissarum). ATX-II strongly affects Na v 1.1 and Na v 1.2, while AFT-II affects Na v 1.4 and Na v 1.5. Given that these two differ in a single amino acid (ATX-II→K36A→AFTII), indicates that the lysine at position 36 is important for the very strong effects of ATX-II on Na v 1.1/2 channels (Oliveira et al., 2004;Moran et al., 2009).

VENOM PEPTIDES TARGETING VOLTAGE-GATED CALCIUM CHANNELS
Voltage-gated calcium channels (Ca v ) facilitate cellular calcium influx in response to membrane depolarization. They regulate hormone secretion, neurotransmitter release, propagation of cardiac action potential, muscle contraction and gene expression in different cell types (Catterall, 2011).
Similar to the Na v channels, the α1 subunit of Ca v channels is organized in four homologous domains (I-IV), each containing six TM segments (S1-S6). The S1-S4 segments constitute the voltage sensor, while S5-S6 constitute the pore. Auxiliary subunits usually associate with α1, regulating channel expression and function. Ca v channels are grouped into various types based on their electrophysiological and pharmacological properties and tissue distribution -L-type (Ca v 1 subfamily: Ca v 1.1-Ca v 1.4); P/Q-, N-, and R-types (Ca v 2.1, Ca v 2.2 and Ca v 2.3, respectively) and T-type (Ca v 3 subfamily: Ca v 3.1-Ca v 3.3). Venom toxins have played a vital role in the discovery of, and in deciphering the structure and function of, many Ca v channels. Chief among them are the ω-conotoxins and ω-agatoxins.
-conotoxins are ∼24-30 residues in length and contain three intramolecular disulfide bonds. They target Ca v channels via blocking the ion pore. -conotoxin GVIA, from the venom of Conus geographus, was the first of the ω-conotoxins to be isolated and characterized (Kerr and Yoshikami, 1984;Olivera et al., 1985). Studies with GVIA showed inhibition of Ca 2+ entry (voltage-activated), and GVIA was a powerful probe to explore the presynaptic terminal, linking Ca v (N-type) channels to neurotransmitter release and synaptic transmission (Kerr and Yoshikami, 1984;Olivera et al., 1984). Molecular identity of the N-type and L-type channel subunit composition was determined using GVIA binding (Williams et al., 1992).
Subsequent to GVIA, many other ω-conotoxins were identified. One of the most prominent ones is MVIIA, from the Magician's cone snail, Conus magus (Olivera et al., 1987), which was tested and developed as a therapeutic agent against pain. Ziconotide (Prialt R ) has been clinically approved for the treatment of severe chronic pain associated with cancer and neuropathies, and is currently the only venom peptide drug targeting a voltage-gated ion channel (Ca v 2.2) that is in clinical use (Miljanich, 2004). A more selective ω-conotoxin, CVID, was isolated from Conus catus (Lewis et al., 2000), and was being developed as leconotide for pain treatment. However, it failed clinical trials due to adverse side-effects (Kolosov et al., 2010).

CONCLUDING REMARKS
Given that there are many species whose toxic venom are yet to be fully explored, the collection of venom-derived peptides to be discovered is immense. Also, with the advent of technology in drug design, based on currently available toxin peptides, new drugs will be developed into more stable and selective biologics. While venomous species developed toxins to incapacitate prey and predators, and envenomation is a public health hazard for us humans, the toxins have proven to be an excellent source of research and therapeutic tools.

AUTHOR CONTRIBUTIONS
All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication.

FUNDING
SB has received funding from the People Programme (Marie Curie Actions) of the European Union's Seventh Framework Programme (FP7/2007(FP7/ -2013 under REA grant agreement no 608765. The authors also acknowledge support from Singapore Ministry of Education Academic Research Fund Tier 2 (MOE2016-T2-2-032) and Lee Kong Chian School of Medicine, Nanyang Technological University Singapore Start-Up Grant to Prof. George K. Chandy.