Introduction

BK channels are a group of calcium-activated potassium channels with large conductance that are found in both excitable and non-excitable cells. BK channels play important roles in many cellular processes, such as the regulation of neuronal firing and neurotransmitter release1, smooth muscle contractions2, endocrine cell secretion3,4, and cell proliferation5,6,7. BK channels are also called Maxi-K, slo1, KCNMA1, or Kca1.18. In general, a functional BK channel is formed by a tetramer of four pore-forming α subunits alone or in association with auxiliary β subunits (β1–β4) that play a modulatory role in specific tissues9. In contrast to other voltage-gated potassium channels, which have six transmembrane helices (S1–S6), the α subunit of BK channels possesses an extra transmembrane domain (S0 at the N-terminus) and a unique long, intracellular tail at the C-terminus (Figure 1). The S0 segment assembles with the β subunit and may contribute to the regulation of the functions of the voltage-sensing domain10,11,12. The long intracellular tail at the C-terminus forms a gating ring that can be triggered by intracellular chemical ligands, including Ca2+, Mg2+, and lipid molecules13,14,15. BK channels act as molecular transducers that link intracellular calcium signals with extracellular stimuli, and an allosteric link between the opening of the voltage-activated channel and calcium sensitivity has been proposed16,17.

Figure 1
figure 1

Topology of the BK channel α subunit (KCNMA1). S0-S6 corresponds to the seven transmembrane segments of the α subunit, where S0 is thought to be important in facilitating the association of the α and β subunits. RCK1 and RCK2 in the cytosolic domains represent binding sites for Mg2+ and Ca2+, respectively.

PowerPoint slide

Full size image

Laboratory and clinical studies have suggested that the malfunction of BK channels may cause numerous disorders in humans, such as hypertension18, asthma19, diabetes20, epilepsy21 and cancer22. Either “gain-of-function” or “loss-of-function” mutations are identified in both the α and β1 subunits of BK channels. The β1 subunit appears to play a major role in regulating the sensitivity of BK channels intracellular calcium spikes. Impairment of the β1 subunit of BK channels results in an increase in arterial tone and blood pressure or hydrosaline retention in the kidney. These unique characteristics of BK channels represent specific targets for drug development.

A number of peptide toxins targeting BK channels have recently been isolated from various venoms, particularly scorpion venom. Because of their modest size, compact structure, and high potency and selectivity, these venom toxins are widely used as pharmacological probes to determine the molecular structure of BK channels and design therapeutic agents.

Peptide toxin blockers for BK channels

Among the different types of peptide toxins, the α-KTx peptides block potassium channels with high affinity and specificity. These peptides can be divided into at least 19 subfamilies based on their amino acid sequences, and several members of these subfamilies have a high affinity for BK, SK, and IK channels. Theses peptide toxins are pore blockers that are used to identify the pore domain of BK channels. BK blockers are used to probe BK channel function and therapeutic utility. The following section presents an overview of the peptide toxins and describes their different channel affinities and three-dimensional structures.

CSα/β toxins in scorpions

Scorpion venoms are dominated by so-called “cysteine-stabilized α/β (CSα/β)” toxins that are evolutionarily derived from CSα/β defensins. Most peptide toxins that target BK channels are from scorpions. These toxins are pore-blocking toxins that inhibit the flow of K+ ions through BK channels by plugging the ion conductance pathway with a highly conserved lysine residue. Based on the length of the amino acid sequence and the number of cysteine residues, these toxins can be classified into several types: short-chain peptide toxins of 29–39 amino acid residues with three disulfide bridges, including α-KTx1, α-KTx3, α-KTx16, and α-KTx19; short-chain peptides with eight disulfide bridges, including α-KTx12; and long-chain peptide toxins containing 60–70 amino acid residues with four disulfide bridges, including the recently discovered BmP09 peptide toxin.

The α-KTx 1.x subfamily

The α-KTx 1.x subfamily is a group of short peptides consisting of 36–37 amino acid residues and three disulfide bridges. This subfamily displays high affinity for BK channels23. The “x” refers to the chronological order in which the peptide was discovered. Within subfamily 1, only α-KTx 1.3 (iberiotoxin, IbTX)24,25 and α-KTx 1.4 (limbatustoxin, LbTX)26,27 have high selectivity for BK channels, whereas α-KTx 1.1 (charybdotoxin, ChTX)28 and α-KTx 1.2 (Lq2)29,30 block BK channels and other types of voltage-gated potassium channels. Although the selectivity and affinity of these toxins differ, they exhibit a high degree of sequence similarity (Figure 2). Each of these toxins have a W residue at position 14 and a K residue at position 27 and contain six conserved cysteine residues in three disulfide bridges with CI-CIV, CII-CV, and CIII-CVI connectivity to maintain the rigid structure28,31. Their structures share a classic motif containing a β-sheet formed by three anti-parallel β-strands on one side of the molecule and a helix on the other side32 (Table 1).

Figure 2
figure 2

Multiple sequence alignment of short-chain peptide toxins from scorpion venom. Conserved cysteine and lysine residues are designated by a dark background. The disulfide framework and the extra disulfide bridge in BuTx are labeled, along with their sequences.

PowerPoint slide

Full size image
Table 1 Three-dimensional structures and pharmacological properties of venom peptide blockers of BK channels.
Full size table

The binding of toxins to BK channels is very sensitive to electrostatic interactions between several basic residues on the toxin and negative charges in the outer vestibule of the channel33,34. Multiple residues, including Trp14 on the β-sheet face of the helix35, Thr23 and Arg25 in the turn between the helix and the second anti-parallel strand36, and Lys27, Arg34, and Tyr36 in the second and third strands of the β-sheet35,36,37, are critical for binding of toxins to the BK channel. This region is highly conserved within the α-KTx 1.x subfamily. Therefore, the primary sequence and the three-dimensional structure of the toxins are important determinants for recognizing and interacting with BK channels.

Charybdotoxin, also called ChTX, was the first potent peptide toxin used to identify and characterize a BK channel23. ChTX is a 37-amino-acid peptide from the venom of the scorpion Leiurus quinquestriatus (var. hebraeus). ChTX has eight positively charged residues, two negatively charged residues, and a blocked N-terminus (cyclization of glutamine at the N-terminus). ChTX can bind to the outer vestibule of a BK channel, but this binding is greatly decreased by an increased internal potassium concentration. Unfortunately, ChTX also blocks other subtypes of potassium channels, such as the small-conductance, calcium-activated potassium channel (SK)38, HLK3 (a voltage-gated potassium channel)39, and a modified Drosophila “Shaker” K+ channel40. The crystal structure of this peptide toxin in complex with a Kv channel indicates that ChTX binds to the channel in a lock and key manner and interacts directly with conducting ions inside the selectivity filter41.

Lq2 (α-KTx 1.2) is a ChTX homolog derived from the venom of the Israeli scorpion, Leiurus quinquestriatus hebraeus. In contrast to ChTX, Lq2 also contains 37 amino acid residues and a classic α/β scaffold30. Lq2 blocks BK channels with a KD of 43 nmol/L29 and can also block inwardly rectifying potassium channels42.

IbTX (α-KTx 1.3, iberiotoxin), a minor fraction of the crude venom of Buthus tamulus, was discovered by Galvez et al in 199024. IbTX also consists of 37 amino acids and possesses 68% sequence identity with ChTX24,25. IbTX has one less positively charged and four more negatively charged residues than ChTX. Functional studies have demonstrated that IbTX binds to the external opening of BK channels with higher affinity than ChTX as indicated by the lower dissociation rate of IbTX compared with ChTX24,25. The nuclear magnetic resonance (NMR) structure of synthetic IbTX suggests that the configuration of the peptide backbone is nearly identical to that of ChTX. IbTX is a specific blocker of BK channels and is widely used in structural and functional studies of BK channels24,43,44,45,46.

LbTX (α-KTx 1.4, limbatustoxin), was isolated from the scorpion Centruroides limbatus. LbTX shares 57% and 70% sequence identity with ChTX and IbTX, respectively. LbTX also exhibits high selectivity and affinity for BK channels26.

BmTx1 (α-KTx 1.5) and BmTx2 (α-KTx 1.6) are both 37-amino-acid peptide toxins purified from the venom of the Chinese Buthid (Buthus martensii). The solution structures of these two peptides are very similar47. They both contain a triple-stranded β-sheet connected to an α-helix, a classic motif in α-KTx 1 subfamily structures. The inhibitory effects of BmTx1 and BmTx2 on BK channels are similar, with IC50 values of 0.65 nmol/L and 0.3 nmol/L, respectively48. These toxins also block voltage-gated potassium channels.

Lqh 15-1 (α-KTx 1.7), also called ChTx2, was isolated from the scorpion venom of Leiurus quinquestriatus hebraeus49,50. Lqh 15-1 consists of 36 amino acid residues. ChTX (α-KTx 1.1) and Lqh 15-1 are homologues with similar three-dimensional structures. Depending on the conditions, Lqh 15-1 blocks BK channels with KD values of 50 nmol/L50 or 105 nmol/L49.

Slotoxin (α-KTx 1.11), which contains 37 amino acids, was isolated from Centruroides noxius Hoffmann scorpion venom. Slotoxin specifically blocks mammalian BK channels (hslo)51. Slotoxin also can differentiate among the three types of BK complexes, including α, α+β1, and α+β4. For example, slotoxin reversibly blocks the pore-forming α subunit with a KD of 1.5 nmol/L, but the blocking effects on the α+β4 complexes are much weaker, with a KD of 100 nmol/L.

Kaliotoxin (α-KTx 3.x)

Kaliotoxin(α-KTx 3.1) is a neurotoxin purified from the scorpion Androctonus mauretanicus mauretanicus52. Kaliotoxin (KTX) has 37 amino acid residues and shares 38% identity with ChTX and 44% identity with IbTX. Because of the presence of two prolines in the α-helix, the three-dimensional structure of KTX determined by NMR includes a shorter and distorted α-helix that differs slightly from those of its homologues53. Furthermore, the increased length of the extended structure preceding the helix alters the accessibility of the conserved Lys2754 and, consequently, its channel blocking effect55. KTX can block both Kv1.356 and BK channels52 with a KD of 20 nmol/L52.

Kaliotoxin 2 (KTX2, α-KTx 3.5) also contains 37 amino acid residues and shares 76% identity with KTX. These two peptides differ mainly in their N-terminal regions and residues, such as 31 and 34, that are located in the channel recognition region. These differences result in a 5-fold decrease in channel blocking by KTX2 compared with KTX57.

Martentoxin (α-KTx 16.2)

Martentoxin (α-KTx 16.2), also named MarTX or BmTx3B, is a potassium channel blocker derived from the Chinese scorpion Buthus martensii Karsch. Martentoxin has 37 amino acid residues and forms a typical CSαβ motif with three disulfide bridges involving Cys8-Cys29, Cys14-Cys34, and Cys18-Cys36. Although the structure of martentoxin is similar to that of the α-KTx1 subfamily, the electrostatic and hydrophobic surface distributions differ considerably58. Martentoxin blocks the currents of neuronal BK channels with an IC50 of 78 nmol/L in Ca2+-dependent manner and partially inhibits mKv1.3 channels59,60,61.

BmBKTX1 (α-KTx 19.1)

BmBKTX1 (α-KTx 19.1) is a 31-amino-acid peptide derived from the Chinese scorpion Buthus martensii Karsch. Two-dimensional NMR spectroscopy revealed a typical Csαβ motif for BmBKTX162. However, BmBKTX1 has only two β-sheet strands connected to an α-helix with three disulfide bridges involving Cys3-Cys22, Cys8-Cys27, and Cys12-Cys29. BmBKTX1 blocks the α subunits of the BK channels of the cockroach (pSlo) and fruit fly (dSlo) with IC50 values of 82 nmol/L and 194 nmol/L, respectively. However, BmBKTX1 does not affect the activity of human BK channels (hSlo) or other voltage-gated potassium channels63.

Butantoxin (α-KTx 12.1)

Butantoxin (α-KTx 12.1), also named BuTX or TsTX-IV, is present in the venom of three Brazilian scorpions, Tityus serrulatus, Tityus trivittatus, and Tityus stigmurus. Butantoxin consists of 40 amino acid residues with four disulfide bridges: Cys2-Cys5, Cys10-Cys31, Cys16-Cys36, and Cys20-Cys38. The secondary structure of butantoxin contains a short, two-turn α-helix (Glu15-Phe23) and a two-stranded β-sheet (Gly29-Met32 and Lys35-Cys38)64. Butantoxin contains more disulfide bonds than other CSα/β toxins. The disulfide bonds CIII–CVI, CIV–CVII, and CV–CVIII constitute the core disulfide framework of CSα/β defensins. The remaining N-terminal CI-CII disulfide bridge is unique to this toxin and does not appear to confer additional stability64. Butantoxin blocks Kv1.265 and BK channels66 with IC50 values of 165 nmol/L and 50 nmol/L, respectively.

The BmP09 long-chain toxin

BmP09, purified from the venom of the Chinese scorpion Buthus martensii Karsch, is a novel long-chain toxin. BmP09 contains 66 amino acid residues, including eight cysteines that specifically block the subunits of mSlo1 with an IC50 of 27 nmol/L. BmP09 exhibits greater specificity and reversibility than ChTX. The Met66 residue at the C-terminus is very important for maintaining selectivity for BK channels. A three-dimensional simulation suggests that the interaction between BmP09 and the BK channel is stabilized by aromatic π-π interactions. The Lys41 residue of BmP09 also blocks the pore of the entrance of the BK channel67.

α/β/α toxins from snakes

Although scorpion venoms are a rich source of peptide toxins that interact with BK channels, many potent and selective BK blockers have been extracted from other species, such as natrin from Naja atra venom68. Natrin is a member of the cysteine-rich secretory protein (CRISP) family that is found in many snake venoms69. It consists of 221 amino acid residues that form two separate domains, the N-terminal α/β/α-sandwich motif (PR-1 domain, residues 1-160) and the C-terminal cysteine-rich domain (CRD, residues 183–221). A compact hinge region (residues 161–182) connects these two domains. All members of the CRISP family contain 16 strictly conserved cysteines that form eight disulfide bonds. Some members of the CRISP family block L-type Ca2+ channels or cyclic nucleotide-gated ion channels. However, natrin can block BK channels in a concentration-dependent manner with an IC50 of 34.4 nmol/L68. The flexible CRD domain is thought to play an important role in channel blocking.

Use of peptide toxins for structural and functional analyses

Native toxins from animal venoms are usually present in small amounts, but chemical synthesis or recombinant expression in bacteria can provide large amounts of the toxin70,71. The expression of fusion-tagged toxins using a bacterial system is the most economical method of producing these toxins, but cleavage of the fusion tags normally results in extra residue(s) at the N-terminus72. The extra residues are thought to have little or no influence on toxin activity. We describe the preparation of IbTX as an example of a recombinant expressed toxin to illustrate the influence of extra residues at the N-terminus.

IbTX is a typical peptide toxin with high selectivity and affinity for BK channels. To obtain large amounts of IbTX, two SUMO-tagged IbTXs with either a SUMO protease or a Factor Xa cleavage site at the N-terminus of the IbTX peptide were cloned, expressed, and purified. Cleavage by SUMO protease introduced an extra serine encoded by the internal part of the BamHI restriction site. Thus, the N-terminal sequence of this SIbTX begins with SQ. Factor Xa cleavage produced an IbTX with an N-terminal Q that was cyclized to pyroglutamate (cIbTX). The toxins and their three disulfide bonds were verified by HPLC and ESI-MS. Patch-clamp electrophysiological assays of SIbTX or cIbTX against a BK channel expressed in Chinese hamster ovary cells demonstrated that 1 μmol/L cyclized IbTX (cIbTX) reduced the BK current to 17% of the control, with an IC50 of 14.9±1.4 nmol/L (n=5), whereas 1 μmol/L uncyclized IbTX (sIbTX) only reduced the BK current to 40% of the control with an IC50 of 105.8±1.9 nmol/L (n=5) at a holding potential of 120 mV (Figure 3). These results demonstrate that the affinity of cIbTX for BK channels is 10-fold higher than that of SIbTX, indicating that the presence of an extra residue at the N-terminus of IbTX strongly influences the self-cyclization, channel binding, and inhibitory activities of the toxin for the BK channel.

Figure 3
figure 3

Blocking activity of SIbTx and cIbTx on the BK channel. SIbTx denotes an extra S at the N-terminal region of IbTx, and cIbTx denotes cyclization of the N-terminal glutamine to pyroglutamate.

PowerPoint slide

Full size image

These results are consistent with the observation that the generation of pyroglutamic acid at the N-terminus can greatly enhance the blocking effect of toxins for their target channels. This modification is present in many types of peptide toxins from diverse origins. Most of these modifications occur in conotoxin, whereas only a few occur in spider and scorpion toxins. Therefore, the N-terminal generation of pyroglutamic acid is a major posttranslational modification of secretory proteins and peptides. This modification is important for biological function, such as mediating interactions with receptors or stabilizing against N-terminal degradation73. Our results provide evidence that the N-terminal generation of pyroglutamic acid by IbTX increases its ability to block BK channels.

Non-peptide blockers

Small-molecule blockers are another type of BK channel blockers. They are usually produced by various organisms, such as bacteria, fungi, plants, and animals, using various biosynthetic methods. In general, the vast majority of non-peptide blockers of BK channels are alkaloids, including indole diterpenes (paxilline, verruculogen, and penitrem), organic amines (tetraethylammonium, verapamil, and ketamine), quinoline derivatives (tetrandrine, quinine, and quinidine), and imidazole derivatives (clotrimazole). Although these small molecules are alkaloids, their structures are very diverse (Figure 4) and their interactions with ion channels vary dramatically74,75,76. The most potent and selective non-peptide blockers are tremorgenic indole-diterpenes from fungi of the genera Penicillium, Aspergillus, and Claviceps77.

Figure 4
figure 4

Molecular formulas of multiple non-peptide blockers of BK channels. Although these small-molecule blockers are alkaloids, with the exception of TEA, their structures vary dramatically.

PowerPoint slide

Full size image

Indole diterpenes

Tremorgenic indole diterpenes belong to a family of tremorgenic mycotoxins from fungi. They can cause neural disorders, such as staggers syndromes in ruminants78,79. These molecules are allosteric modulators, as determined by a [125I]-ChTX binding assay, and may act in either a positive or negative manner74. One group inhibits the [125I]-ChTX binding interaction, acting as a negative allosteric modulator, and includes paspalitrem A and C, aflatrem, penitrem A, and paspalinine. The other indole diterpenes are positive allosteric modulators that exhibit increased [125I]-ChTX binding in a concentration-dependent manner, such as paxilline, verruculogen, and paspalicine. Furthermore, the [125I]-ChTX binding assay revealed a unique interaction of these compounds through the β subunit of the BK channel.

Tetraethylammonium

Organic amines, such as tetraethylammonium (TEA) and its analogues, were originally used to investigate the structure of the ion conduction pathway of BK channels expressed in the rat muscle membrane80. TEA and its analogues block BK channels in a voltage-dependent manner. However, as a widely used open-channel blocker of different types of voltage-gated potassium channels, the applicability of TEA for BK channel studies is limited. Furthermore, TEA and its analogues block BK channels through either the internal or external side of the planar lipid membrane, suggesting a different interaction mode compared with peptide blockers of the BK channel.

Other non-peptide blockers

Alkaloids such as organic amines (verapamil and ketamine), quinoline derivatives (tetrandrine, quinine, and quinidine), and imidazole derivatives (clotrimazole) also possess BK channel blocking activity. Verapamil and its analogues gallopamil and D890, a type of calcium channel blocker, can induce a reversible flickering block of the BK channel activity in rat aortic myocytes81. Ketamine, a dissociative anesthetic, inhibits BK channels by an indirect mechanism. Quinine and quinidine, extracted from Cinchona species, can block both BK and IK channels82,83. Tetrandrine from the tuberous roots of Stephania tetrandra is a bis-benzylisoquinoline alkaloid that inhibits the activity of BK channels in isolated nerve terminals of rat neurohypophysis and cultured endothelial cells84,85. The antifungal clotrimazole, a member of the cytochrome P450 inhibitor family, can block BK activity without altering single-channel conductance86.

Additional endogenous and exogenous compounds have been reported to inhibit BK channel activity. For example, 2-methoxyestradiol (an endogenous metabolite of 17- β-estradiol) and ICI-182,780 (an intracellular estrogen receptor antagonist) can reduce channel activity by shifting the activation curve of the BK channel to a positive potential87,88.

Until recently, the pharmacology of the modulation of maxi-K channels was restricted to potent blockade by the K+ channel blocker TEA, effective and specific blockade by several scorpion toxin peptides, such as IbTX, and blockade by a series of tremorgenic alkaloids exemplified by paxilline89. Although none of these compounds have true potential as therapeutic agents, they continue to be very useful tools in the study of BK channels.

Summary and perspective

We have provided an overview of BK channel blockers, including the structures and pharmacological properties of venom peptides and the molecular formulas and pharmacological effects of non-peptide compounds.

The structural properties and blocking effects of peptide blockers for BK channels

In various scorpion venom peptide blockers of the BK channel, precise disulfide frameworks maintain the CSα/β motif, allowing the conserved lysine to block the entrance to the channel pore. The existence of an interaction surface, such as that of ChTX (α-KTx 1.1)28,31,90 and IbTX (α-KTx 1.3)91, assists binding to the extracellular opening of the BK channel conduction pathway and occludes the outer part of the channel pore, thereby optimizing the interaction between these venom peptides and their molecular targets. In addition to the typical α-KTx peptides that block potassium channels, snake venom peptides with different three-dimensional structures and disulfide bridge patterns also exhibit BK channel blocking properties. The CRD region of these peptides is thought to act as the channel blocker.

Structural properties and BK channel binding affinity and specificity of peptide blockers

The binding of α-KTx peptides to the BK channel is very sensitive to the electrostatic and hydrophobic surface distributions of the venom peptides. Sequence alignments of α-KTx peptides suggest a less positive net charge for IbTX, with four more negatively charged (Asp) and one less positively charged residue. In addition to the difference in electrostatic structure between ChTX and IbTX, Gly30 in IbTX is another determinant of its specificity for the BK channel92,93; this residue is also found in other BK channel-specific blockers, such as LbTX and Slotoxin. Furthermore, our studies of recombinant peptide toxins produced in bacteria have indicated that the pyroglutamic acid of the N-terminus of venom toxins may greatly enhance the blocking effects of toxins on target channels.

Thus, the surface charge distributions and three-dimensional structures of toxins are important determinants of their recognition and interactions with BK channels. This knowledge will facilitate the development of new toxin functions using similar three-dimensional scaffold and electrostatic surface distributions.

Potential clinical applications for BK channel peptide blockers

BK channels are important integrators in many biological systems, and their (patho) physiological functions determine the therapeutic utility of BK channel peptide blockers. Blockade of the BK channel induces an increase in neurotransmitter release94, which suggests that a brain-penetrant blocker of BK channels could increase the signal-to-noise ratio for the remaining synapses, resulting in cognition enhancement in a compromised neuronal system. Other conditions, such as ischemic reperfusion injury, which features abnormal enhancement of BK channel activity indicated by NOS activation95, epilepsy21 due to gain-of-function BK channel mutants, or cancer96 featuring amplifications of the BK channel gene, also represent potential therapeutic uses for BK channel blockers.

Potential clinical development of BK channel non-peptide blockers

For non-peptide compounds, it is difficult to determine a functional structure because of the large diversity of molecular formulas of these blockers. However, small-molecule blockers such as TEA are also used as pharmacological tools or as potential drugs to increase synaptic transmission and restore cognitive function in some types of neurodegenerative diseases. For example, Phenylindole 57, a BK channel blocker, has been used in a mouse model to reduce depression in a dose-dependent manner97 and yielded effects similar to those of antidepressant and psychostimulating drugs. Furthermore, Phenylindole 58 enhanced memory when administered by intraperitoneal injection animal models.

Blockers of BK channels may have potential therapeutic uses. However, none of these blockers has been used in clinical treatments due to limitations in synthesis, delivery, metabolism, tissue penetration and reversibility. Therefore, it is important to develop new drugs targeting BK channels that may facilitate the elucidation of the detailed relationships between the structure and function of BK channels. The structures of known blockers can also be used as templates to design new blockers with potential clinical applications.