Kaliotoxin, a Novel Peptidyl Inhibitor of Neuronal BK-Type

A peptidyl inhibitor of the high conductance Ca2+activated K+ channels (KCa) has been purified to homogeneity from the venom of the scorpion Androctonus mauretanicus mauretanicus. The peptide has been named kaliotoxin (KTX). It is a single 4-kDa polypeptide chain. Its complete amino acid sequence has been determined. KTX displays sequence homology with other scorpion-derived inhibitors of Ca2+-activated or voltage-gated K+ channels: 44% homology with charybdotoxin (CTX), 62% with noxiustoxin (NTX), and 44% with iberiotoxin (IbTX). Electrophysiological experiments performed in identified nerve cells from the mollusc Helixpomatia showed that KTX specifically suppressed the whole cell Ca2+-activated K+ current. KTX had no detectable effects on voltagegated K+ currents (delayed rectifier and fast transient A current) or on L-type Ca2+ currents. KTX interacts in a one-to-one way with KCa channels with a K d of 20 nM. Single channel experiments were performed on high conductance KCa channels excised from the above Helix neurons and from rabbit coeliac ganglia sympathetic neurons. KTX acted exclusively at the outer face of the channel. KTX applied on excised outside-out KCa channels induced a transient period of fast-flicker block followed by a persistent channel blockade. The KTX-induced block was not voltage-dependent which suggests differences in the blockade of KCa channels by KTX and by CTX. Comparison of KTX and CTX sequences leads to the identification of a short amino acid sequence (26-33) which may be implicated in the toxin-channel interaction. KTX therefore appears to be a useful tool for elucidating the molecular pharmacology of the high conductance Ca2+-activated K+ channel.

A current) or on L-type Ca2+ currents. KTX interacts in a one-to-one way with KCa channels with a K d of 20 nM.
Single channel experiments were performed on high conductance KCa channels excised from the above Helix neurons and from rabbit coeliac ganglia sympathetic neurons. KTX acted exclusively at the outer face of the channel. KTX applied on excised outside-out KCa channels induced a transient period of fast-flicker block followed by a persistent channel blockade. The KTX-induced block was not voltage-dependent which suggests differences in the blockade of KCa channels by KTX and by CTX. Comparison of KTX and CTX sequences leads to the identification of a short amino acid sequence (26-33) which may be implicated in the toxin-channel interaction. KTX therefore appears to be a useful tool for elucidating the molecular pharmacology of the high conductance Ca2+-activated K+ channel.
* This work was supported in part by DRET Contract 88/139 and by the Centre National de la Recherche Scientifique. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Potassium-selective channels are extraordinarily diverse as regards to their gating mechanism, pharmacology, ionic conduction properties, and regulation. They are involved in a number of physiological processes such as neuronal electrical activity, muscle contraction, secretory processes, cell proliferation, and cell volume regulation (for reviews see Refs. 1-

4).
Toxins isolated from the venom of insects, scorpions, snakes, and other species are useful tools for probing the structural differences between these channels and evaluating their physiological contribution to the cell behavior (5,6).
Within the wide class of Ca2+-activated K+ channels, two main subtypes have been recognized. These channels differ not only in their unitary conductance and gating mechanisms, but also in their affinity to toxins: charybdotoxin (CTX),' a protein isolated from the venom of the scorpion Leiurus quinquestriatus (7), is a potent inhibitor of the high conductance Ca2+-activated BK channel and apamin, which has been isolated from the bee venom, specifically blocks small-sized Ca2+-activated SK channels (8-10). Owing to its high selectivity for BK-type channels, CTX has been intensively used to investigate the structure and function of BK channels in a variety of membranes. CTX blocks BK channels in muscles ( l l ) , epithelia (12), and neurons (13). More recently, it has been reported that CTX also binds to voltage-dependent K+ channels in synaptosomes (14), dorsal root neurons ( E ) , lymphocytes (16), and myotubes from Drosophila mutants (17). The high affinity binding site for CTX therefore seems to be shared by various voltage-or Ca2+-activated K+ channels, which may have structural homologies (18,19).
In the present paper, we describe the purification, the sequence and the molecular effects on high conductance BKtype neuronal channels of a new K' channel inhibitor purified from a pool of Androctonus mauretanicus mauretanicus venom.

Materials
The venom of Androctonus mauretanicus mauretanicus was obtained by manually stimulating the animals.

1640
This is an Open Access article under the CC BY license.
gave 250 pl of venom (every 100 p1 of venom yielded 8.24 mg of protein). UV-grade acetonitrile was from Fisons Scientific (United Kingdom), trifluoroacetic acid was from Baker (United Kingdom), and all the other analytical reagents were from Merck (Germany).
Bovine serum albumin (BSA) fraction V and cyanogen bromide (CNBr) were from Sigma. Spectrapor 3 dialysis membrane was from Spectrum Medical Industries. The water used to prepare solvents, buffers and dialysis liquids was obtained with a Milli/RO/Milli/Q system from Millipore. Carboxypeptidase Y was from Pierce (The Netherlands) and charybdotoxin was from Latoxan (France). TTX, TEA, EGTA and salts for electrophysiological recordings were from Sigma.
High-performance Liquid Chromatography A Millipore/Waters Associates system was used, including two model 510 pumps, a U6K injector, an automated gradient controller, a 490 spectrophotometric detector, and a data integrator/recorder module. Reverse-phase HPLC was carried out at 25 "C on a Beckman 4 X 250-mm analytic column prepacked with 5 Fm of Lichrosphere 100 RP-18. Solvent A was 0.1% trifluoroacetic acid (v/v) in water, and solvent B was acetonitrile. Additional details concerning all the chromatographic steps are given in the text and figure legends. A Gilson (France) model 202 fraction collector was used with Corning glass tubes at the detector output. The fractions pooled were lyophilized twice in order to eliminate the solvents. Samples for toxicity assays were lyophilized in the presence of 0.1% BSA.

Polyacrylamide Gel Electrophoresis
Polyacrylamide gel electrophoresis of basic proteins was performed at pH 4.1 on 20% homogeneous Phast-Gel using a Phast-System (Pharmacia, Sweden) as defined in the Pharmacia application file No. 300. Proteins were stained with Coomassie Blue under nondenaturating conditions using the Pharmacia development technique No. 200 for native polyacrylamide gel electrophoresis.

Reduction and S-Carboxymethylation
The toxin was reduced with dithioerythritol and S-alkylated with iodoacetic acid as described previously (20). The reduced carboxymethylated toxin (RCM-toxin) was desalted by performing dialysis against 50 mM ammonium bicarbonate, pH 8.0.

Amino Acid Analysis
Acid hydrolyses with 6 N HC1 were carried out on 1-nmol samples of RCM-toxin for 20 and 70 h at 110 "C and under vacuum using a Pico-Tag work station from Millipore/Waters Associates. The amino acid composition was calculated from the analyses made on a Beckman 6300 amino acid analyzer.
CNBr Cleavage at Methionine The RCM-toxin was cleaved by CNBr (21). CNBr-cleaved peptides were separated by HPLC on column prepacked with 5 pm of Lichrosphere 100 RP-18 with isocratic desalt in 100% solvent A (0.1% trifluoroacetic acid) in 70 min followed by a linear gradient from 0 to 40% solvent B (acetonitrile) in solvent A in 100 min; flow rate 1 ml/ min; absorbance reading at 215 nm.

Amino Acid Sequence Analysis
Automated Edman degradations were performed in a Beckman 890" sequencer, and the characterization of phenylthiohydantoinderivatives was carried out by C18-HPLC using the method by Hawke et al. (22).

Lethality Assay on Mice
Venom and toxin toxicities in uiuo were tested on male mice C57 BI/6 (produced at the laboratory) weighing 20 & 3 g, by performing intracerebroventricular injections, and the lethal dose killing 50% of the animals ( L D d was determined as described previously (23).

Electrophysiological Tests
Electrophysiological experiments were performed on invertebrate neurons in the perioesophageal ganglion of the snail Helix pomatia and on vertebrate sympathetic neurons in the coeliac ganglion of rabbit.
Invertebrate Neurons-Two groups of identified neurons were selected on the basis of their various potassium currents (24): (a) U cells (25), these cells generate purely calcium-dependent spikes, and most of the outward current flows through Ca2+-dependent potassium channels (KCa); (b) P cells possess a calcium-and sodium-dependent spike (26). The outward repolarizing current flows in almost equal amounts through Ca2+-dependent (KCa) and voltage-activated (KV) potassium channels. The KCa channel in P cells has similar gating and pharmacological properties to those of the U cell KCa channel.
It displays, however, a long-lasting Ca2+-dependent inactivation which is absent in U cells (27). P cells also have A channels.
The three major K' current components (KCa, KV, and A) were specifically identified by performing either intracellular injections of EGTA (KCa current) or intracellular injections of TEA+ (KV current) or bath applications of 10 mM 4-aminopyridine (A current) (for details, see Ref. 27).
In most experiments, the inward Na+ current was suppressed by either adding M tetrodotoxin (TTX) or bathing the cell in a Na+free saline. Ca2+ currents and Caz+-dependent currents were suppressed by adding 1 mM Cd" to the bath.
Sympathetic Neurons-These neurons were selected because they have a high density of KCa channels with a high unitary conductance (28). Details of the experimental procedure will be published elsewhere. In brief, coeliac ganglia were excised and placed in a Ringer saline buffered with bicarbonate-C02, and the connective sheet was softened by a short (15-20 min) protease treatment as used with invertebrate neurons. This treatment was followed by a 2-3-h cleaning of the neuron surface with a gentle stream of Ringer saline. With this procedure, most of the ganglionic connectivity was preserved.
Electrophysiological Recordings-Helix nerve cells were studied under conventional two-electrode voltage clamp and patch clamp conditions, and sympathetic neurons under patch clamp conditions, as described previously (26,29). Voltage clamp data were sampled at 0.5-1 kHz through a 16-bit A/D converter and stored on floppy disks.
The holding potential was set at -50 mV (-90 mV for evoking A currents). Patch clamp data were low-pass filtered at 2-5 kHz using a 6-pole Bessel filter and continuously recorded in a video cassette recorder after 16-bit digitization at 44 kHz with a pulse code modulator (Biologic, France). Stored data were further digitized at 1-5 kHz and transferred to an Olivetti M28 PC computer for further analysis. All voltages given are from the bath potential. For outside-out patches, electrodes were filled with a KC1-rich saline (80 mM for Helix and 150 mM for rabbit), to IO-' M free calcium adjusted by EGTA buffer (30), 1 mM MgClZ, and 5 mM Tris, pH 7.5.
Lyophilized toxins and venoms were dissolved in 1 ml of normal saline supplemented with 1 mg of BSA. These stock solutions were diluted at the desired concentration with salines containing 1 mg/ml BSA.

Purification and Characterization of Kaliotoxin
The venom was dialyzed against distilled water in order to eliminate salts and small peptides (PM c 3500). Part of the dialyzed venom was diluted in the standard Helix Ringer saline in the presence of 1 mg/ml BSA. Electrophysiological tests were performed under voltage clamp conditions in the U cell group. In these cells, pulse depolarizations at positive levels successively activated a Ca2+ current and a Ca2+-dependent K' current (27).
The dialyzed venom almost completely blocked the U cell outward current without having any noticeable effects on the inward Ca2+ current ( Fig. 1B) Fig. 2 for the criteria used to identify the KCa current). The blockade of the outward current unmasked the slowly inactivating inward Ca2+ current which was not affected by the venom. C, peaks labeled 1-6 in A were tested individually to determine their ability to block the KCa current. Only peak 3 was found to effectively block the outward current. might contain a charybdotoxin-like toxin. By pooling several adjacent peaks from the HPLC chromatogram, we first determined that the fraction active on the Helix KCa channel was located only in the early peaks (retention time: 40-60 min; To prevent protein denaturation and adhesion to the recipient wall, lyophilized fractions were diluted (approximately 0.5 pg/ml) in BSA-containing Helix salines, stored, and used in plastic flasks.
The active fraction was found to be located in the peak labeled 3 in the HPLC chromatogram; within 1 min of perfusion (1.5 ml/min) this peak selectively blocked ( n = 3) the U cell KCa current (Fig. 1G). Adjacent peaks (Fig. lCz) had no blocking effects, which ruled out the possibility that the blockade might be attributable to cross-contamination and unspecific effects of the solvents used to separate the venom proteins.
A reverse-phase chromatography of peak 3 resulted in an active major fraction (eluting at 50 min) (Fig. 2 A ) which gave a single band under polyacrylamide gel electrophoresis in homogeneous Phast-gel 20% (not illustrated). The electrophoretic mobility suggested the existence of a small basic peptide. The purified peptide was analyzed to determine its amino acid content after reduction and carboxymethylation ( Table I). Tyrosine and tryptophane were lacking. The molecular mass calculated from the amino acid composition (37 residues) was 4024 Da, which was almost the same as that obtained with the other K' channel peptidic inhibitors characterized so far. The amino acid composition was very different, however, from that of the previous peptides (see below).
The LD50 of the pure peptide (after intracerebroventricular injection into the mouse) was 1.2 pg/kg. The active peptide amounted to about 1% (in optical density units at 215 nm) of  the dialyzed venom loaded onto the column at the first reverse-phase chromatographic step. This peptide was called kaliotoxine (KTX) due to its ability to block potassium (or kalium)-selective channels.
Automated Edman degradation of 2 nmol of RCM-KTX led to the identification of the 29 amino-terminal residues (line a in Fig. 2C). The initial Edman degradation yield was 25%, and the repetitive yield was 90% for the first 21 steps. To further determine the amino acid sequence, 6 nmol of RCM-KTX were cleaved at the methionine level by a BrCN treatment, and the resulting peptide mixture was separated by reverse-phase HPLC (data not shown). Edman degradation of these peptides made it possible to identify residues 24-28 and 30-37 (lines b and c in Fig. ZC). The complete amino acid sequence agreed with the amino acid composition of KTX shown in Table I. An attempt at sequencing KTX from its carboxyl-terminal end with carboxypeptidase Y was unsuccessfull.

Blockade of Ca2+-activated K' Currents in Helix Neurons by Kaliotoxin
The toxin applied to a U cell had TEA-like effects; the Ca2+-dependent U cell spike lengthened, its repolarizing phase was delayed, and the fast phase of the post-spike hyperpolarization was suppressed (Fig. 2B). These effects were partly reversible upon washing. They were similar to those induced by charybdotoxin on Aplysia neurons (32) and pyramidal cells (33).
Evidence that the spike lengthening resulted from a specific blockade of the Ca2+-dependent component of the repolarizing current is provided by the data in Fig. 3, A and B, which were obtained under voltage clamp conditions in two different U cells. Fig. 3A1 shows a set of currents induced by pulse depolarizations at potentials ranging from -10 to +90 mV. The outward current at large positive potentials had two distinct phases: a fast phase, the amplitude of which increased almost linearly with the pulse potential, and a slower phase, which was particularly prominent at +70 and +90 mV. Just after recording the current set in Fig. 3.41, the cell was impaled with a third microelectrode filled with 0.7 M EGTA. Intracellularly injected EGTA (20-50 nA for 2-5 min) specifically suppressed the slow current component, thus unmasking the fast phase of voltage-gated K' current (Fig. 3A2). The EGTAsensitive current (Fig. 3A3) had similar properties to those of the Ca2+-activated K' current originally described (34) in molluscan neurons (35). Its activation rate decreased characteristically at positive potentials approaching the equilibrium potential for calcium ions.
In spite of its reduced unitary conductance (40-60 picosiemens) (29), the main properties of the KCa channel in Helix nerve cells were similar to those of the large BK channels (36). This channel was blocked by relatively low concentrations of extracellularly applied TEA' (Kd z 2 mM) and by charybdotoxin ( K d = 50 nM). It was insensitive to apamine (100 nM), to the Mast cell degranulating peptide (200 nM), and to the venom from Dendroaspis angusticeps and Dendroaspis polylepis.
The recordings in Fig. 3B were obtained from another U cell subjected to the same voltage program as in the series in Fig. 3A. KTX mimicked the effects of intracellular EGTA.
The KTX-sensitive current (Fig. 3B3) had similar properties to those of the EGTA-sensitive current, i.e. KTX blocked the Ca2+-activated K' component of the outward current with apparently no effect on the voltage-gated K' component.

Blockade of Single KCa Channels by KTX KCa Channels Excised from Helix
Neurons-Channels were excised from Helix U cells in either the inside-out or outsideout configuration. Pipettes were filled with a KC1-rich saline (80 mM KCl) supplemented with 0.1 mM CaC12, 1 mM MgC12, 5 mM Tris. Several criteria were used to identify the excised channels as Ca2+-activated K+ channels. 1 ) KCa channels had a voltage-dependent opening probability which increased efold ( e = 2.72) per =15 mV depolarization (29); 2 ) KCa channels in outside-out patches bathed in the physiological Helix saline were reversibly blocked by adding 5 mM TEA' to the bath; 3) with inside-out patches, the bath saline was replaced by the above KC1-rich saline in which the Ca2+ concentration was varied from nominally 0 (by adding 1 mM EGTA) to M. KCa channel openings were prevented by the saline containing EGTA. Using these criteria, we found that kaliotoxin (50-100 nM) applied to the cytoplasmic face of the patch had no effect on the opening probability of KCa channels.
The recordings in Fig. 4 were obtained on an outside-out patch fitted with two KCa channels. In controls, the channel had long lasting openings at 0 mV, the duration of which increased at positive patch potentials (+20 mV). KTX (40 nM) decreased the channel opening probability, mainly by reducing the opening time duration ( n = 3). Only brief openings persisted in the presence of KTX. The blockade was partly relieved after 15-20-min washing with the physiological saline. brief closures (Fig. 5A). TEA+ (1 mM) added to the bath saline reduced the unitary current amplitude (from 4.8 to 3.6 PA) and induced a characteristic fast-flicker block (Fig. 5B) (12,  28, 37). These effects could be readily reversed by washing (Fig. 5C). Immediately upon treatment with 50 nM kaliotoxin, the channel displayed an erratic behavior with frequent closures and incomplete openings (Fig. 5D1). After a few seconds, the channel remained closed (Fig. 5D2). The KTX block was only partly reversible by prolonged washing: after 12 min (Fig.  5 E ) the unitary current remained depressed and the channel still showed numerous short-lived and long-lasting periods of closure.

KCa Channels Excised from Rabbit Sympathetic Neurons-
In sympathetic neurons ( n = 4) as well as in Helix neurons ( n = 3), a brief application of KTX resulted in full recovery, whereas prolonged KTX applications were poorly reversible.

Dose Dependence of KCa Current Block by K T X
The sensitivity of the Helix KCa channel to KTX was determined on the macroscopic KCa current in voltageclamped U cells. The Fig. 6A1 shows a  from -10 to +70 mV). Insets, currents in a U cell before and after application of 50 nM KTX. B2, the channel blockade was also independent of the holding potential. Data from a U cell in the presence of 50 nM KTX. experiment, the cell was injected with EGTA in order to precisely evaluate the KCa component remaining in the presence of 250 nM KTX. The EGTA-resistant current (mainly calcium and KV currents) was then subtracted from the currents recorded in the presence of various KTX concentrations, which yielded the current set in Fig. 6A1. A plot of the relative KCa current amplitude from this series uersus the KTX concentration is displayed in Fig. 6A2. The experimental points have been fitted with theoretical curves for n KTX molecules interacting with one channel; the continuous curve in Fig. 6Az was the best fit obtained with n = 1, i.e. for one molecule of toxin interacting with one channel. The corresponding dissociation constant of the KTX-channel complex was 20 nM. The dotted line in Fig. 6A2 shows the sensitivity of the Helix KCa channel to CTX. The channel affinity for CTX was slightly smaller (Kd = 50 nM) than for KTX.
The fraction of KCa current blocked by KTX appeared to be independent of both the potential pulse level (range: -10 to +70 mV in Fig. 6B1) and on the holding potential (Fig.  6B2). These results contrast with the voltage-dependent blockade of the channel induced by CTX in Aplysia neurons (32).

Kaliotoxin Does Not Affect Ca' Currents or
Voltage-dependent K+ Currents The data in Fig. 3 indicate that kaliotoxin apparently had no effects on Ca2+ channels or on voltage-gated K+ channels. This point was then definitely confirmed by the experiments illustrated in Fig. 7.
The KV current set in Fig. 7A  KTX had no effect on either Ca2+ currents or voltage-dependent K+ currents. A, voltage-gated KV currents from a Helix P cell. Ca2+ currents and Ca2+-activated K+ currents were suppressed by adding 1 mM Cd" to the normal saline. B, Ca2+ currents from a U cell bathed in normal saline containing M TTX and 20 mM TEA+. C, fast transient A currents from a P cell; A currents were elicited by pulse depolarizations at the levels indicated from a -90 mV conditioning potential applied in order to remove A current inactivation. KTX (right-hand series) had no detectable effects on these currents. potential. The cell was bathed in the physiological saline; Ca2+ current and Ca2+-dependent current were suppressed by adding 1 mM Cd" to the bath saline. KTX had no effect on the KV current induced by either moderate or large depolarizations.
The Ca2+ current set in Fig. 7B was from a U cell bathed in a saline containing TEA+ (20 mM) and TTX M). At this concentration, TEA' suppressed the KCa current (Kd =2 mM) and most of the KV current (Kd =lo mM). The fact that KTX had no effect on the Ca2+ current ruled out the possibility that the blockade of the KCa current may have resulted merely from that of the Ca2+ channels.
The sensitivity of the A current to KTX was assessed in P cells. To relieve the inactivation of A channels, the holding potential was set at -90 mV. The other K+ currents were blocked by 20 mM TEA+ which had no effect on the A channels. The A current was then identified by assessing its blockade by 5-10 mM 4-aminopyridine. The A current remained unaltered in the presence of 50-100 nM of KTX (Fig.  7 C ) .

Two KCa Channel Subtypes in Helix Neurons
Are Blocked by KTX It has been shown that two KCa channels are present in Helix nerve cells (27). In most neurons the macroscopic KCa .activated K+ Channels 1645 current in these cells persisted during prolonged depolarizations, i.e. there was no evidence for the existence of inactivating processes (29). In a few nerve cells, exemplified by the P cell group, the macroscopic KCa current displayed a pronounced fast relaxation upon step depolarization. We have demonstrated 1) that both currents have identical gating and pharmacological properties and 2) that the KCa inactivation in P cells is Ca2+-dependent (26). The KCa component in P cells could therefore be easily and reversibly separated from the KV component by just producing a brief Ca2+ entry. This is depicted in the inset in Fig. 8B which shows two superimposed current traces obtained in response to the attached voltage program; a brief voltage pulse at +10 mV, aimed at inducing an entry of Ca2+ ions, specifically suppressed the slow KCa component of the total outward current induced by a test pulse applied 0.8s later. The current set in Fig. 8B represents the current part suppressed by the brief Ca2+ entry. This current has similar properties to those of the U cell KCa current including a characteristic slowing down at large positive potentials. The series in Fig. 8A was obtained from a P cell. In control saline (Fig. 8A1) the outward current showed the two phases typical of KV and KCa current. KTX specifically suppressed the slow KCa component (Fig. 8A3), thus unmasking the fast-activating KV component ( Fig. 8 A 2 ) .

DISCUSSION
We have demonstrated here that the venom from A. mauretanicus mauretanicus contains a peptide kaliotoxin that specifically blocks Ca2+-activated potassium channels present in nerve cells from two different animal species, the rabbit and the mollusc H. pomatia. In spite of differences in their unitary conductance, KCa channels in Helix neurons (unitary conductance under symmetrical ionic conditions: 60 picosiemens) and in rabbit sympathetic neurons (160-200 picosiemens) belong to the wide group of Ca2+-dependent voltage-gated potassium channels generally referred to as BK (big) channels. The intermediate-conductance channel in Helix neurons is sometimes referred to as IK channel, although IK channels in GH3 cells (38) have properties that lie between those of BK and SK (small-sized) channels, whereas Helix channels have the same pharmacological and gating properties as the BK-type (29). KTX at concentrations ranging from 10 to 250 nM did not affect either voltage-gated K+ channels (delayed rectifier and fast transient A channels) or L-type Ca2+ channels in Helix neurons. These channels are also insensitive to charybdotoxin (see also Ref. 32).
Highly specific channel blockers are useful tools for physiologists and biochemists attempting to elucidate the role and structure of ionic channels. This is particularly obvious in the case of K+-selective channels, in view of their wide diversity. In this context, CTX has been intensively used to characterize BK-type channels. It has been reported, however, that CTX blocks Ca2+-insensitive, V-dependent K+ channels in synaptosomes (14, 39), in dorsal root ganglion cells (15), and in lymphocytes (16, 40). Whether KTX has similar effects still remains to be established.
It appears likely that the mechanism underlying the channel blockade induced by KTX may differ to some extent from that involving CTX. The binding of both toxins to the channel is a bimolecular process. KTX, however, induces a transient period of fast flickering in the channel openings, which is not observed with CTX (32). This transient period induced by KTX is followed by an almost complete blockade of the channel. The fast flickering period is reminiscent of that induced by tetraalkylammonium ions. According to Guggino et al. (12), KTX and CTX might be classified as fast and slow channel blockers, respectively. Another difference between the blockade induced by KTX or CTX lies in the sensitivity to voltage. The blockade of Ca2+-activated K+ channels from skeletal muscles (11, 41) and Aplysia neurons (32) by CTX is voltage-dependent. No evidence for this dependence was observed in the whole-cell KCa current in Helix neurons at voltages ranging from -100 to +90 mV. Additional data at the single channel level are required to clear up these points. For the first time, a potent inhibitor of the high conductance Ca2+-activated K+ channel has been purified to homogeneity from a scorpion venom obtained by manually stimulating the animals. This mode of extraction prevents the soluble material constituting the physiological secretion released by the scorpion during a sting from being contaminated by insoluble mucoproteins and small peptides released by the secretory cells during stimulation with electric shocks. Thus, only one HPLC step was necessary to isolate with a very good yield the toxin from the venom of A. mauretanicus mauretanicus.
In Fig. 9A, the KTX amino acid sequence has been aligned with the sequence of other known peptidyl inhibitors of K+ channels, CTX (42), NTX (43), and IbTX (44), in accordance with the CLUSTAL alignment program (45). The highest level of similarity was detected between KTX and NTX (52% of homology) and lesser sequence homology was observed between KTX and CTX or IbTX (44% of homology).
The N-terminal end of KTX was not blocked, contrary to those of CTX and IbTX. A nonamidated C-terminal end was detected in CTX (42) and apparently also in NTX and IbTX. When S-carboxymethylated KTX was digested by carboxypeptidase Y, no amino acid was released the C-terminal Pro present in KTX sequence, however, is a good explanation for this result. Tyr and Trp are lacking in KTX and only one Phe is present (Table I, and Fig. 2C) which makes necessary to use a wavelength of 215 nm to detect it during HPLC runs. This might also explain that CTX is eluted long after KTX when the two toxins are loaded on a 100 RP-18 column using an acetonitrile gradient.
It should be noted that few amino acid positions (boxed residues in Fig. 9A) are conserved among these four polypeptides: 5 residues in addition to half-cystines. In other respects, KTX and CTX appear pharmacologically very close to each other (this work) and different from NTX (19). Fig. 9B shows the hypothetical structure of KTX based on the disulfide linkages in CTX (46). The shaded units in this figure are the invariant residues (in addition to half-cystines, black units) between KTX and CTX. This figure reveals the presence of a cluster of identical or highly conserved positions from residues 26 to 33. We propose that this sequence plays a prominent role in the specific interaction of KTX and CTX with the BK-type Ca2+-activated K+ channel. This hypothesis seems to be supported by the recent findings obtained with 'H NMR spectroscopy of CTX in solution (47). In this work, the authors propose that CTX has an a-helix (residues 10-20) linked, by two disulfide bridges, to an antiparallel @-sheet made of two short strands (residues 26-28 and 33-35) connected by a turn (residues 29-32). A similar spatial arrangement has been previously found in a and @ scorpion toxins acting on voltage-dependent sodium channels (48-50). Both a-and @-toxins have a loop protruding from the structured region, and in a-toxin this loop is five amino acid longer than in @-toxin (49). This structural difference may be related to the ability for these toxins to bind to two distinct sites. It is tempting to suggest that the loop in a-and @-toxins correspond to the one made by the highly conserved residues 29-32 in CTX and KTX. Consequently, we propose that these 4 residues might play a major role in the pharmacological activity of CTX and KTX. This attractive hypothesis implies that KTX and CTX have a similar three-dimensional structure. Two proline residues, however, are present in positions 12 and 17 in KTX, i.e. in the region corresponding to the a-helix in CTX (residues 10-20). Although proline residues are generally known to destabilize the a-helix structure, a-helices containing proline residues in their sequence have been described (51). More data concerning the three dimensional structure of KTX and the KTX-channel intermtion are required in order to test our hypothesis.