Conotoxins as Tools to Understand the Physiological Function of Voltage-Gated Calcium (CaV) Channels

Voltage-gated calcium (CaV) channels are widely expressed and are essential for the completion of multiple physiological processes. Close regulation of their activity by specific inhibitors and agonists become fundamental to understand their role in cellular homeostasis as well as in human tissues and organs. CaV channels are divided into two groups depending on the membrane potential required to activate them: High-voltage activated (HVA, CaV1.1–1.4; CaV2.1–2.3) and Low-voltage activated (LVA, CaV3.1–3.3). HVA channels are highly expressed in brain (neurons), heart, and adrenal medulla (chromaffin cells), among others, and are also classified into subtypes which can be distinguished using pharmacological approaches. Cone snails are marine gastropods that capture their prey by injecting venom, “conopeptides”, which cause paralysis in a few seconds. A subset of conopeptides called conotoxins are relatively small polypeptides, rich in disulfide bonds, that target ion channels, transporters and receptors localized at the neuromuscular system of the animal target. In this review, we describe the structure and properties of conotoxins that selectively block HVA calcium channels. We compare their potency on several HVA channel subtypes, emphasizing neuronal calcium channels. Lastly, we analyze recent advances in the therapeutic use of conotoxins for medical treatments.


Introduction
Venomous cone snails (Conus) produce several toxic peptides, conopeptides, which target the neuromuscular system of their prey, worms, mollusks, snails and fishes [1,2]. Conotoxins are peptides of 20-30 residues whose main structural characteristic is a rigid backbone formed by disulfide bonds between six cysteines. Conotoxins can be classified according to several criteria, including: a. the gene superfamily they belong to; b. the pattern of cysteine distribution, cysteine framework; and c. their molecular targets. Table 1 summarizes the known groups of conotoxins and their protein targets.
The toxins produced by the genus Conus are numerous and diverse, and approximately 6200 different toxins have been isolated and identified from more than 100 different species thus far [14,15]. The target of most of these toxins are ion channels, including voltage-and ligand-gated channels, as well as G-protein coupled receptors [16,17]. In this review, we will focus on ω-conotoxins, which modulate Ca V 2.X channels. ω-conotoxins prevent entry of calcium (Ca 2+ ) through these voltage-activated Ca V , channels at the presynaptic nerve terminal, thereby, interfering with the release of vesicles containing acetylcholine and neurotransmission [13]. In general, ω-conotoxins impede Ca 2+ flux by physically occluding the channel pore [18]. The kinetics of the binding is variable and can show slow dissociation rates, generating poorly reversible blockage and long term inhibition [18].

Voltage-Gated Calcium Channels
Voltage-gated Ca 2+ (Ca V ) channels are transmembrane proteins that belong to the same transmembrane gene superfamily as the Na V and the K V channels. Ca V channels can be organized into two groups according to the voltage changes required for activation: Ca 2+ channels that require "larger" depolarizations to be opened (when compared with the current-voltage relation for I Na ) are known as high-voltage activated (HVA) channels, whereas Ca 2+ channels that open at more negative potentials are known as low-voltage activated (LVA) [19]. Ca V s are composed of a pore forming subunit, α 1 , encoded by the CACNA1x genes (see Table 2). L-Type Ca V s, Ca V 1.1-1.4, are known as α 1 S, α 1 C, α 1 D, and α 1 F. The P/Q-, N-and R-type, Ca V 2.1-Ca V 2.3, are termed as α 1 A, α 1 B, and α 1 E. Finally, the T-Type, Ca V 3, are composed of α 1 G, α 1 H, and α 1 I (Table 2 [20,21]). Depolarizations provoked by the opening of Ca V channels shape the action potential in the heart, regulate muscle contraction, and modulate neurotransmitter secretion at nerve terminals. In general, "excitable cells translate their electricity into action by Ca 2+ fluxes modulated by voltage-sensitive, Ca 2+ -permeable channels" [19]. Once Ca 2+ ions gain access to the cytosol, they act as second messengers, capable of binding thousands of proteins affecting their localization and function. Variations of intracellular Ca 2+ concentrations influence many cell functions such as transcription, motility, apoptosis and initiation of development [22]. The expression and properties of the pore forming α subunit are modified by two main auxiliary or accessory subunits: α 2 δ and β, which regulate the channel's biophysical properties, its trafficking, and membrane expression. Ca V 1 and Ca V 2 channels can form heteromeric complexes co-assembling with different α 2 δ subunits, which are encoded by CACNA2D1-4 genes, and β subunits, which are encoded by CACNB1-4 genes. The stoichiometry of this assembly is of one β subunit and one α 2 δ accessory subunit. An additional accessory subunit, γ, has been reported only in skeletal muscle [20].

Ca V 2.X Channels
The channels of the Ca V 2 family is formed by a pore-forming Ca V α1 subunit plus the auxiliary subunits Ca V β and Ca V α 2 δ, with the Ca V α1 subunit defining the channel subtype, as shown previously (see Table 2). The Ca V 2.1 channels conduct currents classified as P-type and Q-type that are well described in neurons, whereas the Ca V 2.2 and Ca V 2.3 channels underpin the N-type and R-type currents, respectively, also characterized in neurons [23].
Ca v 2 channels are responsible for the Ca 2+ influx required for the fast release of neurotransmitters as well as for the release of hormones from secretory-type cells such as chromaffin cells [24]. Ca V 2 channels also regulate neuronal excitability via activation of the Ca 2+ activated K + channels that in turn control repolarization and hyperpolarization [25]. Consistent with these functions, Ca V 2.1 null mice exhibit ataxia and die around 4 weeks after birth [26]. Mice deficient in Ca V 2.2 channels, N-type, showed suppressed response to pain, which is consistent with the use of conotoxins as analgesics [27], and with the expression of Ca V 2.2 channels in nerve terminals in association with pain receptors. Ca V 2.2 channels are involved in neurotransmitter release of nociceptive pathways from afferent terminals in the ventral and dorsal horn of the spinal cord and dorsal root ganglion neurons [20,28]. Ca V 2.3 null mice also show reduced pain sensitivity [23].

General Properties of ω-Conotoxins
ω-conotoxins are small peptides ranging in size from 13 to 30 amino acids. They have net charges between +5 and +7 [17], are mostly polar and are highly water soluble. They show three disulfide bridges that are formed between conserved cysteine residues that are arranged in the following organization, C-C-CC-C-C [29]; they form a common structural motif consisting of a cysteine knot, which is also present in toxic and inhibitory polypeptides [30]. The ω-conotoxins family exhibit a characteristic pattern signature described in the PROSITE database [31,32] (see Figure 1). The expression and properties of the pore forming α subunit are modified by two main auxiliary or accessory subunits: α2δ and β, which regulate the channel's biophysical properties, its trafficking, and membrane expression. CaV1 and CaV2 channels can form heteromeric complexes co-assembling with different α2δ subunits, which are encoded by CACNA2D1-4 genes, and β subunits, which are encoded by CACNB1-4 genes. The stoichiometry of this assembly is of one β subunit and one α2δ accessory subunit. An additional accessory subunit, γ, has been reported only in skeletal muscle [20].

CaV2.X Channels
The channels of the CaV2 family is formed by a pore-forming CaVα1 subunit plus the auxiliary subunits CaVβ and CaVα2δ, with the CaVα1 subunit defining the channel subtype, as shown previously (see Table 2). The CaV2.1 channels conduct currents classified as P-type and Q-type that are well described in neurons, whereas the CaV2.2 and CaV2.3 channels underpin the N-type and R-type currents, respectively, also characterized in neurons [23].
Cav2 channels are responsible for the Ca 2+ influx required for the fast release of neurotransmitters as well as for the release of hormones from secretory-type cells such as chromaffin cells [24]. CaV2 channels also regulate neuronal excitability via activation of the Ca 2+ activated K + channels that in turn control repolarization and hyperpolarization [25]. Consistent with these functions, CaV2.1 null mice exhibit ataxia and die around 4 weeks after birth [26]. Mice deficient in CaV2.2 channels, N-type, showed suppressed response to pain, which is consistent with the use of conotoxins as analgesics [27], and with the expression of CaV2.2 channels in nerve terminals in association with pain receptors. CaV2.2 channels are involved in neurotransmitter release of nociceptive pathways from afferent terminals in the ventral and dorsal horn of the spinal cord and dorsal root ganglion neurons [20,28]. CaV2.3 null mice also show reduced pain sensitivity [23].

General Properties of ω-Conotoxins
ω-conotoxins are small peptides ranging in size from 13 to 30 amino acids. They have net charges between +5 and +7 [17], are mostly polar and are highly water soluble. They show three disulfide bridges that are formed between conserved cysteine residues that are arranged in the following organization, C-C-CC-C-C [29]; they form a common structural motif consisting of a cysteine knot, which is also present in toxic and inhibitory polypeptides [30]. The ω-conotoxins family exhibit a characteristic pattern signature described in the PROSITE database [31,32] (see Figure 1). Figure 1. ω-Conotoxins family pattern (PROSITE ID: PS60004). The pattern is described using the following conventions: 'x' is used for a position where any amino acid is accepted; ambiguities are indicated by listing the acceptable amino acids for a given position, between square parentheses '[ ]', i.e., [ALT] stands for Ala or Leu or Thr. Each element in the pattern is separated from its neighbor by a '-'. Repetition of a pattern element can be indicated by following that element with a numerical value or a numerical range in brackets. Examples: Most ω-conotoxins characterized to date are selective for N-type CaV channels. As indicated, the main mechanism of action of ω-conotoxins' is by blocking the channel pore [33], which is accomplished by tight binding of the toxin to the channel pore [18]. The most studied and defined ω-conotoxin is GVIA isolated from Conus geographus [34]. Its specific activity against N-type Ca 2+ channels-CaV2.2 Figure 1. ω-Conotoxins family pattern (PROSITE ID: PS60004). The pattern is described using the following conventions: 'x' is used for a position where any amino acid is accepted; ambiguities are indicated by listing the acceptable amino acids for a given position, between square parentheses '[ ]', i.e., [ALT] stands for Ala or Leu or Thr. Each element in the pattern is separated from its neighbor by a '-'. Repetition of a pattern element can be indicated by following that element with a numerical value or a numerical range in brackets. Examples: Most ω-conotoxins characterized to date are selective for N-type Ca V channels. As indicated, the main mechanism of action of ω-conotoxins' is by blocking the channel pore [33], which is accomplished by tight binding of the toxin to the channel pore [18]. The most studied and defined ω-conotoxin is GVIA isolated from Conus geographus [34]. Its specific activity against N-type Ca 2+ channels-Ca V 2.2 channels- [35] was established in neuronal cell types [36]. Other ω-conotoxins from the venom of different Conus species include CVID from the venom of Conus catus, CNVIIA from Conus consors and MVIIA, MVIIIV, and MVIID from Conus magus have been identified [37]. Additional ω-conotoxins have been isolated from other Conus such as striatus [38], and magus [39].

C. geographus-GVIA
GVIA. It consists of 27 amino acids with a backbone constrained by the formation of three disulfide bonds (Cys 1 -Cys 16 , Cys 9 -Cys 20 , and Cys 15 -Cys 26 ). The possible toxic effect of GVIA and of the other members of the family such as GVIB, GVIC, GVIIA, and GVIIB was determined by performing intracerebral injections in mice, which provoked involuntary movements ("shaking") in the animals [34] (Table 3). In vitro studies were first performed on nerve-muscle preparations of frogs where GVIA irreversible blocked the voltage-activated Ca 2+ channels of the presynaptic terminal preventing acetylcholine exocytosis [13]. Together, these studies showed that GVIA selectively inhibits Ca V 2.2 channels in an irreversible manner. The site of action of GVIA on Ca V 2.2 was found to be on the large extracellular domain III between the S5-S6 trans-membrane regions [18]; mutagenesis studies further showed that the reversibility of the block induced by GVIA and MVIIA was dramatically enhanced by swapping a glycine residue at position 1326 for a proline. GVIA also binds the α1 subunit of the Torpedo nAChR [13]. The 3D structures of GVIA resolved by NMR spectroscopy deposited in the Protein Data Bank (PDB) are: 2CCO [40], 1TTL [41], and 1OMC [42].

C. magus-MVIIA and MVIIC
MVIIA. Also known as ziconotide is a 25 amino acid peptide that also blocks the pore of Ca V 2.2 channels (Table 3) and induced potent analgesia in rodents [43] and human patients with persistent cancer pain [44]. In December 2004, the Food and Drug Administration (FDA) approved Prialt ® (commercial name for MVIIA) for the treatment of severe chronic pain using an intrathecal pump system to deliver the drug into the cerebrospinal fluid. Consistent with this action, injection of MVIIA into mammals caused important neuromuscular effects such as decrease of spontaneous and coordinated locomotor activity and tremors [45]. It was shown that these effects and the pain relief caused by delivery of MVIIA into the cerebrospinal fluid are mediated by inhibition of the release of pro-nociceptive neurochemicals such as glutamate, calcitonin gene-related peptide (CGRP), and substance P into the brain and spinal cord [46,47]. Site-mutagenesis studies revealed that the Met 12 residue in loop 2 ( Figure 2) is the responsible for the toxicity of MVIIA. Met 12 interacts with the hydrophobic pocket residues Ile 300 , Phe 302 , and Leu 305 , located between repeats II and III of Ca V 2.2 channels; this interaction disrupts the normal function of the channel [45]. Systematic mutations of the residues in the loop 2 of MVIIA as well as of other ω-conotoxins may be used for future drug design to develop modulators of Ca V 2.2 with lower side effects and higher effectiveness [45]. The 3D structures of MVIIA resolved by NMR spectroscopy deposited in the PDB are: 1OMG [48], 1MVI [49], 1TTK [50], 1DW4 [51], and 1DW5 [51].
MVIIC. This toxin blocks Ca V 2.1 and Ca V 2.2 channels (Table 3). It possesses similar characteristics to those described for MVIIA and its intracerebral injection in mice caused progressive decrease in respiration rates with marked signs of gasping for breath. The peptide was lethal at low doses (0.1-0.4 µg [52]). The 3D structures of MVIIC resolved by NMR spectroscopy deposited in the PDB are: 1OMN [53] and 1CNN [54].

C. striatus-SVIA and SVIB SO-3
SVIA and SVIB. The SVIA toxin contains 24 amino acids. Its administration into lower vertebrates such as fish and frogs provokes paralysis [38], although it has relatively poor activity against mammalian Ca 2+ channels. While SVIA blocks only Ca V 2.2, SVIB blocks P/Q type and N-type channels (Table 3). SVIB induces respiratory distress in mice when injected intracranially at concentrations of 70 pmol/g mouse and it is lethal around 300 pmol/g mouse; SVIA administration does not kill mice even at extremely high doses [38].

SO-3.
This ω-conotoxin shows analgesic activity similar to that of MVIIA when tested in models of acute and chronic pain in rodents, however, it has fewer adverse effects than MVIIA [45,55]. The 3D structure of SO-3 resolved by NMR spectroscopy deposited in the PDBe is: 1FYG [56].

C. catus-CVID
CVID. The sequence of its loop 4 is less conserved than other of ω-conotoxins. It displays the highest selectivity for N-type over P/Q-type Ca 2+ channels (radioligand binding assays) [57] and because of this it has been tested in clinical trials as analgesic [58]. The 3D of CVID structure resolved by NMR spectroscopy deposited in the PDB is: 1TT3 [50].

FVIA.
It is reported to only be effective against Ca V 2.2 channels [59]. The 3D structure of FVIA resolved by NMR spectroscopy deposited in the PDB is 2KM9 (to be published).

C. textile-TxVII and CNVIIA
TxVII. This conopeptide is very hydrophobic and has net negative charge of −3. The sequence of TxVII is 58% identical to that of δ-conotoxin-TxVIA, which targets Na + channels. This toxin blocks the slowly inactivating, dihydropyridine-(DHP-) sensitive current [60]. The 3D of TxVII structure resolved by NMR spectroscopy deposited in the PDB is: 1F3K [61].
CNVIIA. This toxin is closely related to the CnVIIH toxin (Table 3), which possesses an unprocessed final glycine and therefore lacks amidation of its C-terminal end [62,63]. CNVIIA blocks Ca V 2.2 channels but surprisingly it does not block the neuromuscular junction of amphibians. Intracerebroventricular injection of CNVIIA in mice causes shaking movements and mild tremors, depending on dosage, whereas when injected intramuscularly into fish it causes paralysis and death at higher doses [62].

Structural Characteristics of ω-Conotoxins and Blockade Site on the Ca V Channels
ω-conotoxins share several structural characteristics that allow them to block multiple Ca V s on diverse cell types. Here we explore in detail four well known ω-conotoxins, CVID, SVIB, GVIA, and MVIIA, whose 3D structures have been resolved by NMR except for CVID (PDB IDs 1MVJ [49], 2CCO [40], and 1MVI [49], respectively); their multiple sequence alignments as well as their 3D structures are shown in Figure 2. As previously noted, they share four loops and three disulfide bonds (Figure 2A,C), giving them the same structural pattern ( Figure 2B). These similarities are evident between CVID and SVIB (RMSD backbone = 0.109 Å), although they are more subtle between CVID and GVIA (RMSD backbone = 1.635 Å) ( Figure 2D). The main structural differences between loops 2 and 3 (structural difference 1) and 4 (structural difference 2), where ω-conotoxins residues are not highly conserved, are highlighted with gray boxes ( Figure 2B). Despite the structural similarities, there are differences in the selectivity of targets between these toxins. To understand the selectivity of these toxins at the structural basis, using NMR spectroscopy, researchers have determined the secondary and tertiary structures [50]. Adams et al. found a correlation between the solvent accessible surface area and the selectivity of ω-conotoxins, where the most exposed residue, R10 in MVIIA, play a crucial role in binding to Ca V s [50]. The residue(s) on Ca V channels that interacts with ω-conotoxins is not yet elucidated, although the extracellular linker region between the P-region and S5 in domain III, the pore of Ca V 2.2, is reported to be the area where the toxins bind channels [17,33]. In this region, G1326 appears to be the essential residue, as its mutation modifies the access of GVIA and MVIIA to the active site [69].
The structure-activity relationship (SAR) studies conducted in conotoxins identified key residues involved in the interaction with protein targets as well as identification of specific amino acids involved in their structural arrangement. These studies have been used to design small bioactive mimetics to selectively block Ca V 2.2 over Ca V 2.1 channels [70,71]. Bioactive mimetics have become promising candidates in the search for novel drugs for the treatment of chronic pain [21]. For example, based on the 3D structure of MVIIA [48] and identification of key residues such as K2, R10, L11, Y13, and R21 involved in the binding of MVIIA [72]. The data collected gave fundamental information for the design of the first bioactive mimetic of MVIIA in 1998, including the draft of small structures to mimic the residues R10, L11, and Y13 [73]. Although this bioactive mimetic showed poor inhibition against Ca V 2.2 (19% at 10 µM), a second generation of mimetics was produced and two of these compounds showed promising activities against Ca V 2.2 (IC 50 = 3.3 and 2.7 µM) [74]. Since then, others ω-conotoxins mimetics have been reported using SAR information [70,75,76].

Therapeutic Uses of Conotoxins
The therapeutic and pharmacological potential of the conotoxins is well-known [1,47,77]. Nevertheless, their intrinsic physic-chemical and therapeutic characteristics such as molecular weight and low bioavailability due to their susceptibility to peptidase degradation has prevented the widespread use of conotoxins in the clinic. Importantly, and despite these limitations, their ability to selectively bind closely related molecular targets is an important strength of these marine conopeptides. Another advantage of Conotoxins is the diversity of targets, as they can act upon ion channels such as KV, NaV, and CaV channels, as well as on several G-protein coupled receptors including neurotensin, α-adrenergic, and vasopressin receptors and also on ligand-gated receptors such as AChRs, 5HT3Rs, and NMDARs [77]. These properties make them excellent candidates to develop new bio-compounds and derivatives against pathologies such as pain, stroke, and convulsive disorders. Especially interesting is their specific affinity for N-type, CaV2.2, CaV channels, which is a useful pharmacological characteristic for the validation of molecular targets, for example, in neuropathic pain. CaVs channel-mediated cellular events can be modulated for therapeutic purposes by direct block of CaV2.2,

Therapeutic Uses of Conotoxins
The therapeutic and pharmacological potential of the conotoxins is well-known [1,47,77]. Nevertheless, their intrinsic physic-chemical and therapeutic characteristics such as molecular weight and low bioavailability due to their susceptibility to peptidase degradation has prevented the widespread use of conotoxins in the clinic. Importantly, and despite these limitations, their ability to selectively bind closely related molecular targets is an important strength of these marine conopeptides. Another advantage of Conotoxins is the diversity of targets, as they can act upon ion channels such as K V , Na V, and Ca V channels, as well as on several G-protein coupled receptors including neurotensin, α-adrenergic, and vasopressin receptors and also on ligand-gated receptors such as AChRs, 5HT3Rs, and NMDARs [77]. These properties make them excellent candidates to develop new bio-compounds and derivatives against pathologies such as pain, stroke, and convulsive disorders. Especially interesting is their specific affinity for N-type, Ca V 2.2, Ca V channels, which is a useful pharmacological characteristic for the validation of molecular targets, for example, in neuropathic pain. Ca V s channel-mediated cellular events can be modulated for therapeutic purposes by direct block of Ca V 2.2, i.e., small peptides as conotoxins; by activation of GPCRs, or by direct interference with the channel trafficking [23].
In nature the proteins targeted by the cone snails on the preys are closely related to the proteins targeted in humans; however small structural and physiological differences can modify the efficacy, selectivity, and potency of conotoxins. Moreover, the target protein in cone snail's preys may serve functions that are distinct to those in humans. Further, in humans and mammals the target proteins may be found in protected physiological spaces such as the Central Nervous System [77].
The pharmacological and therapeutically pre-clinical efficacy of MVIIA and CVID, along with the FDA approval of Prialt ® , have established ω-conotoxins (and conotoxins in general) as viable platforms for the design of new and specific drugs to alleviate pain by aiming N-type Ca V channels.

Conclusions
Neuronal Ca V channels have potential as targets for treatments of pain and the selectivity of conotoxins for these channels render conopeptides valuable therapeutic tools. ω-conotoxins display an inhibitory cysteine knot which is also present in other toxic peptides. This motif, along with other common structural characteristics, is the basis of their potent and selective blocking activity on the pore of Ca V channels. A ω-conotoxin, MVIIA, has been approved by the FDA for therapeutic use under the commercial name of Prialt ® . Going forward, however, more widespread applications of conotoxins will require improvements to enhance their transport across the blood-brain barrier as well as modification to increase their chemical stability.
The association between the structure of ω-conotoxins and their activity against Ca V channels remains undetermined and such knowledge will be fundamental to improve their use as therapeutic agents. Techniques such as circular dichroism and NMR spectroscopy have been helpful in the development of SAR studies, which have aided in the design of MVIIA [84], and GVIA [85] analogues. Additionally, the combination of electrophysiology, computational biophysics approaches, and SAR studies has provided new insights into the molecular binding mechanism of ω-conotoxins to their targets. This knowledge now places the drug design processes targeting chronic pain in a robust position to develop novel therapeutic agents. The design of small mimetics requires the identification of the correct scaffolds as well as of key residues to mimic. Towards this end, non-peptide mimetics containing the scaffolds of dendritic, 8-hydroxy-2-(1H)-quinolinone and the 5-hydroxymethyl resorcinol and the residues Leu, Arg, and Tyr, which matched the pharmacophore found in the conotoxin, were developed as MVIIA mimetics and show promissory biological activities against Ca V 2.2. Conotoxins remain an attractive option for the development of new therapeutic strategies using bioactive mimetics against chronic pain. Nevertheless, additional work involving both experimental and theoretical approaches are needed to unravel at the structural level the mechanisms modulating the protein targets of these peptides.