Abstract
Spider venoms are actively being investigated as sources of novel insecticidal agents for biopesticide engineering. After screening 37 theraphosid spider venoms, a family of three new “short-loop“ inhibitory cystine knot insecticidal toxins (κ-TRTX-Ec2a, κ-TRTX-Ec2b, and κ-TRTX-Ec2c) were isolated and characterized from the venom of the African tarantula Eucratoscelus constrictus. Whole-cell patch-clamp recordings from cockroach dorsal unpaired median neurons revealed that, despite significant sequence homology with other theraphosid toxins, these 29-residue peptides lacked activity on insect voltage-activated sodium and calcium channels. It is noteworthy that κ-TRTX-Ec2 toxins were all found to be high-affinity blockers of insect large-conductance calcium-activated K+ (BKCa) channel currents with IC50 values of 3 to 25 nM. In addition, κ-TRTX-Ec2a caused the inhibition of insect delayed-rectifier K+ currents, but only at significantly higher concentrations. κ-TRTX-Ec2a and κ-TRTX-Ec2b demonstrated insect-selective effects, whereas the homologous κ-TRTX-Ec2c also resulted in neurotoxic signs in mice when injected intracerebroventricularly. Unlike other theraphosid toxins, κ-TRTX-Ec2 toxins induce a voltage-independent channel block, and therefore, we propose that these toxins interact with the turret and/or loop region of the external entrance to the channel and do not project deeply into the pore of the channel. Furthermore, κ-TRTX-Ec2a and κ-TRTX-Ec2b differ from other theraphotoxins at the C terminus and positions 5 to 6, suggesting that these regions of the peptide contribute to the phyla selectivity and are involved in targeting BKCa channels. This study therefore establishes these toxins as tools for studying the role of BKCa channels in insects and lead compounds for the development of novel insecticides.
Introduction
Conventional agrochemicals are used currently to combat phytophagous pest insects and vectors of human and livestock diseases. In addition to the issue that many have been deregistered because of human health risks and environmental concerns, one of the major dilemmas associated with these agents is the increasing development of insecticide resistance (Nicholson, 2007a). Unfortunately, wide use and the limited target range of these compounds have resulted in an increasing incidence of insecticide resistance and a lack of effective insecticide control. Therefore, there is now a push to pursue novel insect-selective compounds with alternate targets as leads for new insecticidal agents (Nicholson, 2007b; King et al., 2008a). The validation of a novel insecticidal target and insect-selective lead compounds would therefore provide a fresh approach in the development of more effective insect control agents.
It is already well established that ion channel toxins, particularly those targeting voltage-activated sodium (NaV) channels, are useful insecticidal compounds, given their rapid lethal activity, high potency, and relative insect selectivity. Nevertheless, with an increasing incidence of insecticide resistance, there is a need to explore novel targets, such as insect voltage-activated calcium (CaV) and potassium (KV) channels (King et al., 2008a). It is noteworthy that potassium channels are a diverse group of ion channels represented in both electrically excitable and nonexcitable tissues and as such represent a novel target mediating the effects of potential insecticidal agents.
Peptide toxins isolated from spider venoms have long been recognized as valuable pharmacological tools for probing mammalian voltage-activated ion channel structure and function (Swartz, 2007). However, attention has been turning to their use as potential sources of insecticidal agents. The evolutionary process in spiders has led to the development of rich combinatorial peptide libraries, mostly comprising ion channel toxins specifically designed to rapidly immobilize and kill insect prey (Escoubas, 2006; King et al., 2008a). Indeed, recent evidence would suggest that some venoms from “primitive” mygalomorph spiders contain more than 1000 peptides (Escoubas et al., 2006). Although mygalomorph venoms potentially contain toxins for a wide range of insect targets, it is likely that toxins targeting voltage-activated ion channels are mainly responsible for the fast lethal effects of these venoms in insects. It is this rapid and pernicious insect-selective activity that makes these toxins ideal candidates as leads for the design of novel insecticides.
κ-Hexatoxin-Hv1c (κ-HXTX-Hv1c, formerly J-atracotoxin-Hv1c; King et al., 2008b) from the venom of the Australian funnel-web Hadronyche versuta (Araneae: Mygalomorphae: Hexathelidae) has been identified as the first toxin to selectively and potently target an insect K+ channel, with selective actions to block insect but not vertebrate large-conductance BKCa channels (Gunning et al., 2008). κ-HXTX-Hv1c exhibits lethal activities in a number of insect orders, including Coleoptera, Diptera, Lepidoptera, and Orthoptera, whereas lacking activity in mouse, rabbit, rat, and chick preparations (Maggio and King, 2002). Thus κ-HXTX-Hv1c validates K+ channels as potential insecticidal targets.
Although several insect-selective toxins have been isolated from the venom of spiders belonging to the family Hexathelidae (King, 2007; Nicholson, 2007b), there has been only a limited number of investigations of insecticidal toxins from spiders of the family Theraphosidae (Li et al., 2003; Corzo et al., 2008). This is despite a number of toxins from theraphosid (“tarantula”) spiders targeting mammalian acid-sensing ion channels, mechanosensitive, KV, NaV, and CaV channels having been characterized (see the ArachnoServer 2.0 Spider Toxin Database, http://www.arachnoserver.org/) (Herzig et al., 2011). In this study, we describe a family of toxins (theraphotoxins) from the venom of the East African tarantula Eucratoscelus constrictus (Gerstäcker 1873; Araneae: Mygalomorphae: Theraphosidae) that block insect BKCa channels and fail to significantly affect other insect KV, NaV, and both mid- to low-voltage-activated (M-LVA) and high-voltage-activated (HVA) CaV channels. It is noteworthy that although these toxins share their target with κ-HXTX-Hv1c, they possess no obvious sequence homology, implying that they may interact with distinct portions of the insect BKCa channel. These theraphotoxins will be useful for probing the biological role of BKCa channels in insects and are potential lead compounds for the development of insect-selective biopesticides.
Materials and Methods
Venom Supply.
E. constrictus (formerly E. longiceps; World Spider Catalog, http://research.amnh.org/entomology/spiders/catalog/index.html) venom was purchased from a commercial supplier (Invertebrate Biologics, Los Gatos, CA). Venom was obtained using electrical stimulation of the chelicera. Crude venom was first diluted up to two thirds of its final volume in ultrapure distilled water, centrifuged at 14,000 rpm for 30 min at 4°C, and filtered using a 0.45-μm microfilter (Millipore Corporation, Billerica, MA). The filter was rinsed with water (one third of final volume), and the venom was diluted to 10 times its initial volume and stored at −20°C until further use.
Toxin Purification and Biochemical Characterization.
Bioassay-guided fractionation of E. constrictus venom was performed using both reversed-phase and cation-exchange high-pressure liquid chromatography (HPLC). A volume of 20 to 25 μl of crude venom equivalent was diluted with 0.1% aqueous trifluoroacetic acid (TFA) and fractionated by reversed-phase HPLC (RP-HPLC) on a semipreparative C8 column (10 × 250 mm; Nacalai Tesque, Kyoto, Japan). The column was equilibrated in 0.1% TFA, and peptides were eluted with a 5 to 60% acetonitrile/0.1% TFA gradient at a flow rate of 2 ml/min (see Fig. 2 for details). Fractions were collected manually by monitoring the absorbance at 215 nm. The fractions were vacuum-dried, dissolved in 200 μl of ultrapure water, and tested for insecticidal activity (see below). Insecticidal fractions were further purified by cation-exchange HPLC using a TSKgel sulfopropyl column (4.6 × 75 mm; Tosoh Corporation, Tokyo, Japan). The fractions were diluted to 200 μl with 20 mM ammonium acetate and were analyzed using a linear gradient from 20 mM to 2 M ammonium acetate at 0.5 ml/min (see Fig. 2 for details). Absorbance was monitored at 280 nm, and fractions were collected manually. Fractions were dried under vacuum, and a final purification step was undertaken using an RP-HPLC C18 column (4.6 × 250 mm; Waters Symmetry, Waters, Milford, MA) and a water/acetonitrile/0.1% TFA gradient at 1 ml/min (see Fig. 2 for details). Purified toxins were freeze-dried and stored at −20°C.
Insect Lethality Assays.
Juvenile crickets (Gryllus bimaculatus third instar nymphs; average weight, 65–70 mg) were injected intrathoracically with 1 to 5 μl using a 10-μl precision syringe. Samples were dissolved in distilled water, and controls were injected with water only. For more rapid toxin identification, pools of fractions were tested. Each RP-HPLC fraction (see below) was dried, reconstituted in 200 μl of distilled water, and 1-μl aliquots of each fraction were mixed to reconstitute pools of 8 to 10 fractions. Final pool volumes varied from 8 to 14 μl depending on the number of aliquots mixed for a given pool. 2 μl of each pool was injected into crickets as described above (n = 3). For individual RP-HPLC fractions, aliquots of 1 μl were diluted 10 times in distilled water, and 2 μl was injected into crickets (n = 3). Cation-exchange HPLC fractions were freeze-dried, redissolved in distilled water, and 1:20 aliquots were injected into crickets (n = 3). Observation of paralysis and/or death was made at 5, 15, and 60 min after injection with the assay endpoint for calculation of median effective doses (ED50) being 15 min. ED50 values were evaluated using a range of five to six venom or toxin concentrations, with five crickets per dose (n = 3). Results were analyzed using the SOFTTOX program (WindowChem, Fairfield, CA) using the probit analysis method.
Mouse Bioassays.
Activity against vertebrates was evaluated by intracerebroventricular injection in female C57BL/6 mice. Mice (approximately 20 g average weight) were lightly anesthetized with diethyl ether, injected in the left cerebral ventricle with 5 μl of sample dissolved in a solution of bovine serum albumin [2% (w/v) in saline], and placed in glass jars for observation. Development of signs of toxicity was noted continuously during the first hour after injection and was monitored at regular intervals for 24 h or until death. A dose of 500 pmol/mouse was used to discriminate insect-selective peptides from peptides with both insect and vertebrate activity (n = 2).
Edman Peptide Sequencing.
Before sequencing, toxins were reduced with 20 mM β-mercaptoethanol (30 min, 37°C) and were alkylated with 4-vinylpyridine (Wako Pure Chemicals, Osaka, Japan), in 250 mM Tris-HCl/6 M guanidine buffer, pH 8.3, for 30 min at 37°C in the dark under argon. The pyridylethylated toxins were desalted by C18 RP-HPLC using a linear gradient of acetonitrile/water/0.1% TFA. Toxin sequencing was performed on an Applied Biosystems 477A automated gas-phase sequencer (Applied Biosystems, Foster City, CA). Pyridylethylated toxins were dissolved in HPLC-grade water, applied to TFA-treated glass fiber membranes precycled with Biobrene (Applied Biosystems), and subjected to N-terminal (Edman) sequencing. Sequence homologies were determined from a BLAST search of the ArachnoServer 2.0 Spider Toxin Database server.
Mass Spectrometry.
Mass spectra were recorded on an Applied Biosystems Voyager DE-Pro matrix-assisted laser desorption/ionization time-of-flight mass spectrometer (Applied Biosystems), in positive ion linear or reflector mode using α-cyano-4-hydroxycinnamic acid matrix. Calibrations were performed either in close external or internal mode, with a mixture of synthetic peptide calibrants (LaserBio Labs, Sophia-Antipolis, France). Data were analyzed with the manufacturer's DATA EXPLORER software. To confirm sequencing data, trypsin cleavage of κ-TRTX-Ec2b was performed in Tris-HCl buffer, pH 8.5, at 37°C for 4 h, with a 1:50 enzyme-to-peptide ratio. Trypsin was purchased from Promega (Charbonnières-les-Bains, France). Theoretical molecular masses were then calculated with version 7.10 of the GPMAW program (http://www.gpmaw.com/index.html) and compared with measured masses.
Homology Modeling.
Three-dimensional models of the κ-TRTX-Ec2 toxins were calculated using the NMR structure coordinates of κ-TRTX-Gr2a (Protein Data Bank ID 1LUP) as a template via submission to the Swiss Model server (http://www.expasy.ch/swissmod/SWISS-MODEL.html) and analyzed with PYMOL software version 1.3 (http://www.pymol.org).
Patch-Clamp Electrophysiology.
Whole-cell currents were recorded in voltage-clamp mode using the whole-cell patch-clamp technique using version 9 of the pCLAMP data acquisition system (Molecular Devices, Sunnyvale, CA). Data were filtered at 5 kHz with a low-pass Bessel filter with leakage and capacitative currents subtracted using P-P/4 procedures. Digital sampling rates were set between 15 and 25 kHz depending on the length of the voltage protocol. Using a Flaming-Brown micropipette puller (Sutter Instrument Company, Novato, CA), single-use electrodes were pulled from borosilicate glass to direct current resistances of approximately 1, 1.5, and 2.5 MΩ for sodium, calcium, and potassium channel current recordings, respectively. Liquid junction potentials for the various combinations of internal pipette and external bath solutions were calculated using the JPCalc software program (Molecular Devices), and all data were compensated for these values. Series resistance compensation was >80% for all cells. Cells were bathed in external solution through a continuous pressurized perfusion system at 1 ml/min, whereas toxin solutions were introduced via direct pressurized application via a perfusion needle at approximately 50 μl/min (Automate Scientific, San Francisco, CA). All experiments were performed at ambient room temperature (20–23°C).
Primary Cell Culture.
Dorsal unpaired median (DUM) neurons were isolated from unsexed adult American cockroaches (Periplaneta americana) as described previously (Gunning et al., 2008). In brief, DUM neurons were enzymatically and mechanically isolated from the terminal abdominal ganglia of the cockroach. Extracted ganglia were removed and placed in normal insect saline (NIS) containing 180 mM NaCl, 3.1 mM, 10 mM HEPES, and 20 mM d-glucose. Ganglia were then incubated in 1 mg/ml collagenase (type IA) for 40 min at 29°C. After enzymatic treatment, ganglia were washed three times in NIS and triturated through a fire-polished pipette to dissociate individual neurons. The cell suspension was then distributed onto 12-mm diameter glass coverslips precoated with 1 mg/ml concanavalin A (type IV). Cells were maintained for no longer than 24 h in NIS supplemented with 5 mM CaCl2, 4 mM MgCl2, 5% fetal bovine serum, and 1% penicillin and streptomycin and maintained at 30°C, 100% humidity.
Electrophysiological Recording Solutions.
To record INa, the external bath solution contained 80 mM NaCl, 5 mM CsCl, 1.8 mM CaCl2, 50 mM tetraethylammonium (TEA) chloride, 5 mM 4-aminopyridine (4-AP), 10 mM HEPES, 0.1 mM NiCl2, and 1 mM CdCl2, adjusted to pH 7.4 with 1 M NaOH. The pipette solution contained 34 mM NaCl, 135 mM CsF, 1 mM MgCl2, 10 mM HEPES, 5 mM EGTA, and 3 mM ATP-Na2, adjusted to pH 7.4 with 1 M CsOH.
Because of the report of ICa rundown with calcium as a charge carrier and greater success when barium was used as the charge carrier (Wicher and Penzlin, 1997), BaCl2 replaced CaCl2 in all experiments. The external bath solution for barium current (IBa) recordings contained 140 mM sodium acetate, 30 mM TEA-bromide, 3 mM BaCl2, and 10 mM HEPES, adjusted to pH 7.4 with 1 M TEA-OH. The external solution also contained 300 nM tetrodotoxin (TTX) to block NaV channels. Pipette solutions contained 10 mM sodium acetate, 110 mM CsCl, 50 mM TEA-bromide, 2 mM ATP-Na2, 0.5 mM CaCl2, 10 mM EGTA, and 10 mM HEPES, adjusted to pH 7.4 with 1 M CsOH.
The external bath solution for recording macroscopic KV channel currents contained 150 mM NaCl, 30 mM KCl, 5 mM CaCl2, 4 mM MgCl2, 0.3 mM TTX, 10 mM HEPES, and 10 mM d-glucose, adjusted to pH 7.4 with 1 M NaOH. The pipette solution consisted of 135 mM KCl, 25 mM KF, 9 mM NaCl, 0.1 mM CaCl2, 1 mM MgCl2, 1 mM EGTA, 10 mM HEPES, and 3 mM ATP-Na2, adjusted to pH 7.4 with 1 M KOH.
To eliminate any influence of differences in osmotic pressure, all internal and external solutions were adjusted to 400 ± 5 mOsM with sucrose. Experiments were rejected if there were large leak currents or currents showed signs of poor space clamping.
Curve-Fitting and Statistical Analysis.
Data analyses were completed offline at the conclusion of experiments using AxoGraph X version 1.1 (Molecular Devices). Mathematical curve-fitting was accomplished using PRISM version 5.00b for Macintosh (GraphPad Software Inc., San Diego, CA). All curve-fitting routines were performed using nonlinear regression analysis using a least-squares method. Comparisons of two sample means were made using a paired Student's t test. Multiple comparisons were assessed by repeated-measures analysis of variance with a Bonferroni's multiple comparison post hoc test; differences were considered to be significant if p < 0.05. All data are presented as mean ± S.E.M. of n independent experiments, unless stated otherwise.
The following equation was employed to fit current-voltage curves: where I is the amplitude of the peak current (either IBa, INa, or IK) at a given test potential V, gmax is the maximal conductance, V1/2 is the voltage at half-maximal activation, s is the slope factor, and Vrev is the reversal potential. Concentration-response curves were fitted using the following logistic equation: where x is the toxin dose, nH is the Hill coefficient (slope parameter), and IC50 is the median inhibitory concentration to block channel currents.
Results
Purification of Insecticidal Neurotoxins.
As part of a search for novel insecticidal peptide toxins, a large-scale screening of theraphosid venoms was pursued. Venom extracted from the African “tarantula” E. constrictus was selected for its particularly potent paralytic activity in crickets (ED50 = 0.017 μl/g; Fig. 1). Bioassay-guided fractionation of crude E. constrictus venom by C8 RP-HPLC yielded 26 fractions, with fraction 14 (f14) being selected for its high insecticidal toxicity (Fig. 2A). Because fractions f14 to f16 were inadequately resolved in the first HPLC separation, they were pooled and analyzed together by cation-exchange HPLC (Fig. 2B). This led to the separation of seven fractions (f14.1-f14.7) that were then tested for insect and mammalian activity. These seven fractions yielded three insecticidal peptides (f14.2, f14.3, and f14.5), which were fully purified after an additional chromatography step using C18 RP-HPLC (Fig. 2C). In the insect assay, all three fractions demonstrated significant neurotoxicity at a dose of 1.1 nmol/g (4 μg/g). At this dose, the three toxins caused rapid insecticidal activity, with complete paralysis within 5 min, and death within 15 min (n = 3). For f14.2 and f14.3, intracerebroventricular injection of 500 pmol (1.8 μg) into 20-g mice (n = 2) also did not result in any neurotoxicity or behavioral symptoms. All aspects of mouse posture and behavior appeared to be totally unaffected by toxin injection, indicating that f14.2 and f14.3 are phyla-selective neurotoxins. On the other hand, injection of 500 pmol f14.5 (1.8 μg) into 20-g mice (n = 2) resulted in strong neurotoxicity symptoms, including convulsions, tonic paralysis, general ataxia, extension of legs and toes, tail erection, and respiratory paralysis resulting in very marked cyanosis of the animal. These symptoms seemed to be reversible because animals recovered from cyanosis and tonic paralysis after approximately 4 h and started to resume normal posture and activity.
Edman degradation of reduced and alkylated peptides yielded the complete N-terminal sequence of the three toxins from single sequencing runs. All cysteine residues were definitively established after reduction and alkylation using 4-vinylpyridine. Based on their mass and amino acid sequence, these toxins belong to a class of inhibitor cystine knot (ICK) toxins (Pallaghy et al., 1994) with six cysteines (CI–CVI). The κ-TRTX-Ec2 toxins also clearly belong to the family of “short-loop” ICK spider peptides, distinguished from the more common “long-loop” ICK spider toxins, by the presence of only three residues in the CV–CVI loop (Escoubas and Rash, 2004). Because of high homology with a number of other previously characterized short-loop tarantula, we propose the presence of three disulfide bonds (Fig. 3A) linked in the CI–CIV, CII–CV, and CIII–CVI pattern resulting in a disulfide-linked pseudoknot described previously for this toxin family.
To verify sequencing data, f14.3 was also digested with trypsin and the resulting peptide mixture analyzed with matrix-assisted laser desorption/ionization time-of-flight mass spectrometry for confirmation of sequence ambiguities, particularly the presence of Trp24. In addition, the measured masses of the peptides were matched with calculated theoretical mass values of the amino acid sequence, using the software program GPMAW (http://www.gpmaw.com/). This allowed complete unambiguous sequence confirmation (data not shown).
Each peptide consists of 29 amino acids with observed monoisotopic molecular masses of 3626.73, 3676.71, and 3627.16 Da for f14.2, f14.3, and f14.5, respectively. This compares very well with the monoisotopic mass predicted from the sequence of 3626.53, 3676.51, and 3627.52 Da, respectively. This indicates that there are no apparent post-translational modifications, such as C-terminal amidation, observed with certain other spider toxins (e.g., β-TRTX-Gr1b, formerly GsAFI; see Fig. 3A). The three toxins share more than 80% sequence homology with a number of spider toxins, the majority of which are short-loop toxins from the venom of another theraphosid spider, Grammostola rosea (Fig. 3A). Homologous peptides exhibited a range of activities including high-affinity inhibition of NaV, CaV1.2, CaV3.1, KV4.1, KV4.2, KV4.3, and KVAP channels and weak affinity for mechanosensitive and KV2.1 channels (see the ArachnoServer 2.0 database for a full list of references). Furthermore, the positions of all cysteines were strictly conserved with all short-loop theraphosid spider toxins (Fig. 3A).
Using the rational nomenclature system for spider toxins proposed by King et al. (2008), f14.2, f14.3, and f14.5 were named κ-theraphotoxin (TRTX)-Ec2a, κ-TRTX-Ec2b, and κ-TRTX-Ec2c, respectively. The activity descriptor prefix “κ” indicates modulators of K+ channels (evidence presented below), “theraphotoxin” is the generic name for toxins from the family Theraphosidae (see Supplemental Table 4 in King et al., 2008b), “Ec” is the genus and species descriptor for E. constrictus, 2 is consistent with the majority of other short-loop ICK tarantula toxins (see below), and a, b, and c denote the first three paralogs (isoforms) discovered.
Target Determination of κ-TRTX-Ec2 Toxins.
Although κ-TRTX-Ec2 toxins share a similar sequence and cysteine spacing arrangement with short-loop theraphotoxins, this provided few clues as to the potential target of κ-TRTX-Ec2 toxins. This is because within the short-loop theraphotoxin group, the biological activities are quite diverse with activities on a wide variety of voltage-activated and mechanosensitive ion channel targets. Nevertheless, given the considerable sequence homology of κ-TRTX-Ec2 toxins with β-TRTX-Gr1a (formerly GrTx-I) from G. rosea (Fig. 3A), a toxin known to target mammalian NaV channels, initial experiments investigated whether κ-TRTX-Ec2 toxins modulated insect NaV channel currents (INa). Whole-cell TTX-sensitive INa were recorded from cockroach DUM neurons in the voltage-clamp configuration. Test pulses to −10 mV (Fig. 4D) from a holding potential (Vh) of −90 mV elicited a fast activating and inactivating inward INa that could be abolished by the addition of 300 nM TTX (Fig. 4A). The application of 330 nM κ-TRTX-Ec2b failed to significantly alter INa amplitude or time course over a 10-min perfusion period (92 ± 4% of control peak INa, p > 0.05, n = 3; Fig. 4A).
Experiments then investigated whether κ-TRTX-Ec2 toxins modulated the activity of insect CaV channels. This was based on their significant sequence homology with β/ω-TRTX-Tp2a (formerly ProTx-II) from Thrixopelma pruriens (Fig. 3A) shown previously to target mammalian CaV1.2 and CaV3.1 channels, as well as NaV channels (Middleton et al., 2002; Edgerton et al., 2007). We investigated the action of κ-TRTX-Ec2 toxins on both HVA and M-LVA CaV channels that have been characterized previously in cockroach DUM neurons (Chong et al., 2007). Unfortunately, despite differences in the kinetic and pharmacological properties of M-LVA and HVA CaV channels, there remains no mechanism for recording one current in isolation from the other because no peptide or organic blockers are available that exclusively block one channel subtype. As described previously, depolarizing pulses to two potentials (−20 and +30 mV; Fig. 4D) from a Vh of −90 mV were used to investigate the actions of the κ-TRTX-Ec2 toxins on M-LVA and HVA CaV channels, respectively (Chong et al., 2007). The elicited barium current (IBa) was abolished by the addition of 1 mM CdCl2, confirming that currents were carried via CaV channels (Fig. 4Ba). A 10-min application of 330 nM κ-TRTX-Ec2b failed to significantly reduce either M-LVA or HVA CaV channel currents, with only a 10 ± 6% (p > 0.05, n = 3) and 7 ± 6% (p > 0.05, n = 3) inhibition of control peak IBa, respectively (Fig. 4B, a and b).
To determine whether κ-TRTX-Ec2 toxins alter the voltage-dependence of channel activation, voltage protocols from a Vh of −90 to +70 mV in 10-mV increments were delivered at a frequency of 0.1 Hz to elicit families of INa and IBa. Resultant peak INa-V and peak IBa-V relationships were fitted with eq. 1 (see Materials and Methods). Consistent with a lack of inhibition of INa or IBa amplitude, there was no significant shift in the threshold, or voltage midpoint, of activation (V1/2) of NaV or CaV channels in the presence of 330 nM toxin (data not shown). These experiments therefore indicate that κ-TRTX-Ec2 toxins fail to affect invertebrate NaV and CaV channels in the model studied.
Global KV channel currents were recorded in the presence of 300 nM TTX and 1 mM CdCl2 to block NaV and CaV channel currents, respectively. Currents were elicited by depolarizing test pulses to +30 mV (Fig. 4D) and upon the addition of 50 mM TEA-chloride, a nonselective KV blocker, the channel current was completely abolished indicating that these currents were carried by K+ channels (data not shown). In contrast to the lack of overt activity on insect CaV and NaV channels, a 10-min perfusion with 330 nM κ-TRTX-Ec2a, -Ec2b, or -Ec2c resulted in a significant reduction in DUM neuron global KV channel current amplitude (Fig. 4C). This effect was partially reversible with approximately 50% recovery observed after a 10-min washout with toxin-free external solution.
Determining the K+ Channel Subtype Specificity of κ-TRTX-Ec2 Toxins.
Block of global outward IK indicates that κ-TRTX-Ec2 toxins target at least one of the four major K+ channel current subtypes in cockroach DUM neurons. These are identified according to their current kinetics and pharmacological properties and include slowly activating, noninactivating delayed-rectifier [IK(DR)], transient “A-type” [IK(A)], transient Na+-activated [IK(Na)], as well as large-conductance “late-sustained” and “fast-transient” Ca2+-activated [IBK(Ca)] K+ channel currents (Grolleau and Lapied, 1995). The fast-transient BKCa channel differs from the late-sustained BKCa channel in that it inactivates rapidly after activation and displays a voltage-dependent resting inactivation (Grolleau and Lapied, 1995). The KCa current in cockroach DUM neurons is of the large-conductance (BKCa) subtype because the current is voltage-activated, whereas SKCa and IKCa channel currents are voltage-insensitive, and no apamin-sensitive SKCa channels have been found in isolated cockroach DUM neurons (Grolleau and Lapied, 1995). As a consequence of the inhibition of macroscopic outward IK, all subtypes except for KNa channels, which contribute only a minor portion of the total current, were investigated as potential targets of κ-TRTX-Ec2 toxins.
For isolation of K+ channel subtypes, various channel blockers and current subtraction routines were used. To isolate outward IK(DR) in DUM neurons, IK(A) were blocked with 5 mM 4-AP (Grolleau and Lapied, 1995; Gunning et al., 2008) whereas BKCa channels were blocked with 1 mM CdCl2 and 30 nM charybdotoxin (ChTx) (Gunning et al., 2008). Both κ-TRTX-Ec2b and κ-TRTX-Ec2c (330 nM) failed to significantly inhibit IK(DR) (p > 0.05, n = 3; Fig. 5C b and c). However κ-TRTX-Ec2a produced a 31 ± 12% inhibition of control IK(DR) (p < 0.01, n = 3; Fig. 5Ca). To assess whether κ-TRTX-Ec2 toxins alter the voltage-dependence of IK(DR) activation, the current-voltage [IK(DR)-V] relationship in the presence of each toxin was determined using a series of 100-ms voltage pulses from a Vh of −90 to +40 mV in 10-mV increments, delivered at a frequency of 0.2 Hz. The threshold of IK(DR) activation was at membrane potentials more depolarized than −50 mV, in agreement with previous data from IK(DR) recordings in DUM neurons (Grolleau and Lapied, 1995; Gunning et al., 2008). Data were normalized against maximal late current (measured at 100 ms) recorded in the absence of toxin and subsequently fitted with eq. 1 (see Materials and Methods). It is noteworthy that there were no significant shifts in the voltage-dependence of IK(DR) activation for any κ-TRTX-Ec2 toxin (ΔV1/2 < 5.5 mV for all toxins, p > 0.05, n = 3–7) (Fig. 5D). In summary, κ-TRTX-Ec2b and κ-TRTX-Ec2c failed to significantly inhibit IK(DR), although in the case of κ-TRTX-Ec2a, this may account for a statistically significant proportion of the block of global KV current (Fig. 4C).
In insects, the shaker, shab, shaw, and shal genes encode for the equivalents of mammalian KV1–KV4 channels, respectively. Although it is believed the Shab and Shaw channels are responsible for IK(DR) in insects (Wei et al., 1990), little work has been done on native neurons, apart from Drosophila melanogaster to determine the various contributions of these two channels. Therefore, it is quite possible that κ-TRTX-Ec2a targets, with moderate affinity, one of the two possible KV channel subtypes contributing to IK(DR). However, at present, the relative expression of Shab versus Shaw channel subtypes in cockroach DUM neurons is unknown, and given the potent action on BKCa channels (see below), such an action is unlikely to contribute significantly to the overall activity of κ-TRTX-Ec2a.
Of particular interest was the effect of κ-TRTX-Ec2 toxins on IK(A) as a result of their homology with a number of theraphosid spider toxins targeting the mammalian KV4 channel (Fig. 3A). The KV4 channel is the ortholog of the insect Shal channel that is believed to underlie a significant portion of the IK(A) (Salkoff et al., 1992). IK(A) cannot be recorded in isolation from IK(DR) because there are no selective blockers of insect IK(DR). To record IK(A) in isolation from other KV channel currents, a subtraction routine was therefore employed using 4-AP. To isolate IK(A), neurons were continuously perfused with 30 nM ChTx and 1 mM CdCl2 to block BKCa channels (Gunning et al., 2008). At the conclusion of experiments, 5 mM 4-AP was then perfused to eliminate IK(A), allowing for offline subtraction of the remaining IK(DR). This involved separate experiments with κ-TRTX-Ec2a, -Ec2b, or -Ec2c, which were perfused for a period of 10 min or until equilibrium was achieved. 4-AP (5 mM) was then applied to block IK(A). The remaining IK(DR), recorded in the presence of 4-AP, was then digitally subtracted from both controls and currents recorded in the presence of each κ-TRTX-Ec2 toxin. Although there was a small reduction in the amplitude of IK(A) in the presence of κ-TRTX-Ec2a, IK(A) at +30 mV was not significantly reduced by any of the κ-TRTX-Ec2 toxins (p > 0.05, two-way analysis of variance, n = 3; Fig. 5A). To assess whether κ-TRTX-Ec2 toxins alter the voltage-dependence of IK(A) channel activation, the current-voltage (IK(A)-V) relationship in the presence of each toxin was determined using the same voltage protocol as for IK(DR)-V curves. Again, in agreement with previous data from IK(A) recordings in DUM neurons, the threshold of IK(A) activation was at membrane potentials more depolarized than −50 mV (Grolleau and Lapied, 1995; Gunning et al., 2008). It is noteworthy that there were no significant shifts in the voltage-dependence of KA channel activation for any κ-TRTX-Ec2 toxin (ΔV1/2 < 5.5 mV for all toxins, p > 0.05, n = 3–7; Fig. 5B). In summary, the κ-TRTX-Ec2 toxins failed to inhibit IK(A) to an extent that could account for the 20 to 40% decrease seen in global KV currents (Fig. 4C).
Like IK(A), BKCa channel currents [IBK(Ca)] cannot be recorded in isolation from IK(DR) because, as mentioned previously, there are no selective blockers of insect IK(DR). Therefore, to isolate IBK(Ca), 5 mM 4-AP was included in external recording solutions. Both 30 nM ChTx and 1 mM CdCl2 were also perfused at the conclusion of current recordings to eliminate IBK(Ca) allowing for offline subtraction of the remaining IK(DR). Although Cd2+ may modify many mammalian KV channel currents, there is no evidence to suggest that this is the case in DUM neurons. Unfortunately, the use of ChTx, in the absence of Cd2+, is insufficient to completely block BKCa channels, and previous studies of BKCa channel blockers in DUM neurons have all found it necessary to use Cd2+ to inhibit Ca2+ entry (Derst et al., 2003). In separate experiments, κ-TRTX-Ec2a, -Ec2b, or -Ec2c were then perfused for a period of 10 min, or until equilibrium was achieved. CdCl2 (1 mM) and 30 nM ChTx were then applied to block BKCa channels. The remaining IK(DR), recorded in the presence of BKCa channel blockers, was then digitally subtracted from both control and currents recorded in the presence of each κ-TRTX-Ec2 toxin (Fig. 6B). All three toxins caused a similar concentration-dependent block of both transient and sustained IBK(Ca) at all concentrations tested. The IC50 values for block of IBK(Ca) by κ-TRTX-Ec2a, -Ec2b, and -Ec2c were 3.7, 25.3, and 24.6 nM for peak IBK(Ca) and 3.7, 25.8, and 19.1 nM for late IBK(Ca), respectively (Fig. 7, E and F). The approximately 6-fold higher potency of κ-TRTX-Ec2a versus Ec2b/Ec2c on peak and late IBK(Ca) maybe partially explained by the significant block of IK(DR) by κ-TRTX-Ec2a (Fig. 5, Ca and Da), which would cause an apparent increase in potency after IK(DR) subtraction routines in IBK(Ca) isolation procedures. Thus, although κ-TRTX-Ec2 toxins share considerable sequence homology with other spider toxins targeting NaV, KV, or CaV channels, some of which are promiscuous in their actions on multiple families of ion channel (e.g., β/ω-TRTX-Tp2a), the evidence from the present study indicates that κ-TRTX-Ec2 toxins act selectively on BKCa channels.
To assess whether κ-TRTX-Ec2 toxins block DUM neuron BKCa channels in a voltage-dependent manner, the IBK(Ca)-V relationship in the presence of each toxin was determined. Families of IBK(Ca) were elicited by a series of 100-ms voltage pulses from a Vh of −80 to +40 mV in 10-mV increments at a frequency of 0.2 Hz. The threshold of IBK(Ca) activation was at membrane potentials greater than −40 mV, which is in agreement with previous data from BKCa channel recordings in DUM neurons (Grolleau and Lapied, 1995; Gunning et al., 2008). In the presence of 6 nM κ-TRTX-Ec2a, 25 nM κ-TRTX-Ec2b, or 25 nM κ-TRTX-Ec2c, there was no shift in the threshold or V1/2 of IBK(Ca) activation of either peak or late IBK(Ca) (Fig. 7). Furthermore, channel block was not relieved at increasing potentials (Fig. 7, Ac, Bc, and Cc), as noted for other voltage-dependent toxins, including many theraphosid spider toxins (Swartz, 2007).
Discussion
κ-TRTX-Ec2 Toxins Target Insect BKCa.
Although all three of the κ-TRTX-Ec2 toxins share significant homology with a number of short-loop theraphotoxins acting on mammalian voltage-activated ion channels, they are unique from two perspectives. First, κ-TRTX-Ec2a and κ-TRTX-Ec2b are strictly insect-selective and thus show invertebrate phyla-selectivity not described previously for this group of toxins. Second, κ-TRTX-Ec2 toxins display high affinity and selectivity for the insect BKCa channel, a pharmacology previously only reported with κ-hexatoxin-Hv1c (Gunning et al., 2008), with whom they share no significant homology (Fig. 3B). As potent ligands for the insect BKCa channel, these toxins therefore represent valuable tools in the exploration of BKCa channel structure and characterizing the biological functions of insect BKCa channels. Indeed, these two families of toxins may become the defining pharmacology for insect BKCa channels.
Mode of Action.
Several lines of evidence suggest that κ-TRTX-Ec2 toxins interact with their target via mechanisms unique from other theraphotoxins. Until the present study, KV channel modulators from tarantula venoms were all characterized as gating-modifier toxins (Swartz, 2007). These toxins inhibit conductance by interfering with the voltage-sensing region of ion channels to cause a depolarizing shift in the voltage-dependence of activation thus stabilizing the closed or inactivated states of the channel. However, unlike all other theraphotoxins acting on KV channels, BKCa current inhibition induced by the κ-TRTX-Ec2 toxins failed to exhibit any depolarizing shifts in the threshold, or voltage midpoint (V1/2), of channel activation, clearly indicating that the mechanism of channel inhibition is not via the modulation of channel gating.
Other toxins interacting with KV channels block ion conduction by physically occluding the pore of the channel near the selectivity filter. “Pore blockers” interact with amino acids situated in the S5–S6 linker and physically occlude the pore (Gross and MacKinnon, 1996). Among structurally dissimilar KV channel pore blockers from sea anemone and scorpion venoms, a “functional dyad” is common to the bioactive surface of the molecules. This comprises an invariant lysine residue situated within 6.6 ± 1.0Å of an aromatic residue such as tyrosine or phenylalanine (Dauplais et al., 1997). This mode of action is similar to that reported for the insect-selective neurotoxin κ-HXTX-Hv1c, from the venom of the hexathelid spider H. versuta, that also inhibits BKCa channels (Gunning et al., 2008). Homology modeling of κ-TRTX-Ec2 toxins on the known structure of κ-TRTX-Gr2a (formerly GsMTx-2, Oswald et al., 2002; Protein Data Bank ID 1LUP), revealed several pairs of basic and aromatic residues that could satisfy the requirements of a functional dyad (Fig. 8). Possible functional dyads include Lys26-Trp24, Lys4-Phe5 (Lys5-Phe4 in κ-TRTX-Ec2c) and Lys14-Tyr1. However, none of these potential dyads are unique from residues in other β- or β/ω-TRTX toxins that do not target KV channels (Fig. 3A). Moreover, there was no apparent voltage-dependence to the degree of block at various membrane potentials, with no reduction in block at higher potentials. Classic pore blockers such as charybdotoxin (α-KTx 1.1), κ-conotoxin PVIIA, and, to a lesser extent, κ-HXTX-Hv1c all demonstrate an alleviation of channel block at increasingly depolarized potentials as potassium driving force increases. As a consequence, these classic pore blockers all show moderate to strong voltage-dependent dissociation (Goldstein and Miller, 1993; Terlau et al., 1999; Gunning et al., 2008). Thus, the lack of voltage-dependent block by κ-TRTX-Ec2 toxins indicates that the toxins do not bind deeply into the extracellular mouth of the ion channel to plug the pore in a manner similar to other KV and BKCa channel blockers.
Scorpion toxins that block small-conductance SKCa (KCa2.x) channels lack the conserved functional dyad present in peptide pore blockers of KV, IKCa, and BKCa channels. These toxins include, among others, scyllatoxin (α-KTx 5.1) and tamapin (α-KTx 5.4). Like κ-TRTX-Ec2 toxins, α-KTx 5.1 and α-KTx 5.4 have been shown to display a voltage-independent block of SKCa channels (Shakkottai et al., 2001; Pedarzani et al., 2002) and are believed to interact with the turret and loop region of the SKCa channel (Rodríguez de la Vega et al., 2003; Andreotti et al., 2005). Hence, these types of toxins are known as “turret blockers” to distinguish them from classic pore blockers. Unlike Lys27 within the functional dyad of α-KTx 1.1, which plugs the pore of KV and BKCa channels and senses approximately 20% of the transmembrane electrical field (Goldstein and Miller, 1993), these SKCa toxins and κ-TRTX-Ec2 toxins all lack a Lys27 equivalent. Therefore, we predict that κ-TRTX-Ec2 toxins, like α-KTx 5 toxins, are likely to bind to the extracellular channel surface and act as turret blockers to completely inhibit ion flux. There exists the possibility that κ-TRTX-Ec2 toxins act to block the channel by acting as an electrostatic barrier rather than via steric occlusion of the channel. However, other toxins that cause such an electrostatic inhibition of channels do not completely inhibit current flow, resulting in a residual current (Hui et al., 2002). Given that κ-TRTX-Ec2 toxins can completely inhibit BKCa channel currents, we believe that they cause a direct steric occlusion of the channel.
Structure and Function.
Although the key residues comprising the bioactive surface (pharmacophore) of κ-TRTX-Gr2a are currently unknown, the pharmacophore of the highly homologous β/ω-TRTX-Tp2a (formerly ProTx-II) for NaV channels has been identified (Smith et al., 2007). Comparing the pharmacophore of the NaV and CaV channel-selective β/ω-TRTX-Tp2a (Fig. 8A) and the surface of the KV4-selective κ-TRTX-Gr2a and κ-TRTX-Cj2a (Fig. 8, B and C) with BKCa-selective κ-TRTX-Ec2 toxins, it can be seen that the major surface residues that differ for the κ-TRTX-Ec2 toxins are located at the C-terminal region (particularly residues 28–29), with additional mutations at positions 4 to 6, 11, 18, and 22 (Fig. 8, D–F). In particular, the presence of a glutamic acid residue at either position 28 or 29 in κ-TRTX-Ec2a and κ-TRTX-Ec2b is likely to confer insect-selectivity, because all of the remaining mammalian-active theraphotoxins have a basic or hydrophobic residue at one, or both, of these positions (Fig. 3A). Furthermore, affinity for the BKCa channel displayed by κ-TRTX-Ec2 toxins is likely to result from substitutions in the highly conserved Lys4-Trp5-Met6 of short-loop theraphotoxins (Fig. 3A), particularly because the apparent spatial location of Lys4-Phe5 of κ-TRTX-Ec2a and κ-TRTX-Ec2b (Phe4-Lys5 in κ-TRTX-Ec2c) is different from the other theraphotoxins (Fig. 8).
It is noteworthy that differences in the sequence of the KV/BKCa channel target can also drastically alter the affinity of toxins for the channel complex. Although the selectivity filter and pore helix are well conserved across all potassium channels, the solvent-exposed turret and loop regions display high sequence variability and contribute to differences in α-KTx specificity (Giangiacomo et al., 2008). This can also extend to a variation in the size of the turret region. For example, the BKCa (Slo1) channel turret has six more residues with very limited sequence homology to the turret region of KV1 channels (Giangiacomo et al., 2008). The size and sequence variability of the BKCa channel turret and loop regions could therefore explain the high selectivity of κ-TRTX-Ec2 toxins for BKCa over KV channels. These sequence and size variations in the turret region also extend to subtle differences between the turret regions of various BKCa channels across different phyla. Although the size of the turret region of BKCa channels from vertebrates, marine invertebrates, and nematodes is believed to be the same, insect BKCa turret regions are typically one to three amino acids shorter (Giangiacomo et al., 2008; Gunning et al., 2008). Furthermore, single-point mutations in this region can result in dramatic changes in the phyletic selectivity of toxins (Myers and Stampe, 2000; Derst et al., 2003). Thus, the amino acid variation in the turret and loop regions of insect versus vertebrate BKCa channels seems to be sufficient to explain the insect selectivity of κ-TRTX-Ec2 toxins and κ-HXTX-Hv1c (Gunning et al., 2008).
It is noteworthy that subtle changes in toxin sequence can also drastically alter the affinity of the toxin for specific channel subtypes. For example, point mutations of Lys32 in α-KTx 1.1 result in mutants with markedly reduced affinity for KV1.2 and KV1.3 channels but little change in affinity for BKCa or IKCa channels (Rauer et al., 2000). In addition, a chimeric toxin construct involving residues from the α/β turn of iberiotoxin (α-KTx 1.3) grafted into noxiustoxin (α-KTx 2.1) causes a 50-fold weaker block of KV1.3 channels but a >150-fold higher affinity for block of BKCa channels (Mullmann et al., 2001). These variations highlight that subtle differences in toxin sequence can have profound effects on channel subtype selectivity or specificity. This, together with variations in the amino acid sequence of channel subtypes, probably explains the different pharmacologies and modes of action of BKCa-selective κ-TRTX-Ec2 toxins versus other short-loop theraphosid spider toxins.
Novel Lead Compounds that Validate New Insecticide Targets.
The κ-TRTX-Ec2 toxins represent the second family of toxins, along with κ-HXTX-Hv1c, to validate insect BKCa channels as potential targets for insecticides. As potent and selective blockers of insect BKCa channels, κ-TRTX-Ec2a and κ-TRTX-Ec2b toxins also represent valuable lead compounds in the development of future insect-selective biopesticides with a novel mode of action. Nevertheless, to undertake any structure-based rational insecticide design program, there is also a need to identify the residues responsible for phyla and target selectivity. Therefore, the close sequence homology between diverse short-loop theraphotoxins, displaying both phyla-selective neurotoxicity and diverse pharmacology, has afforded the unique opportunity to hypothesize on the molecular determinants for insecticidal and target selectivity. As a result, future mutagenesis experiments to determine the pharmacophore of the κ-TRTX-Ec2 toxins, guided by these hypotheses, should enable the determination of the residues responsible for insect selectivity and assist in the design of phyla-selective peptidomimetic insecticidal compounds.
Authorship Contributions
Participated in research design: Windley, Escoubas, Valenzuela, and Nicholson.
Conducted experiments: Windley, Escoubas, and Nicholson.
Performed data analysis: Windley, Escoubas, and Nicholson.
Wrote or contributed to the writing of the manuscript: Windley, Escoubas, Valenzuela, and Nicholson.
Footnotes
This work was supported in part by the Australian Research Council [Grant DP0559396]; the Department of Education, Science and Training [Grant FR050106]; the Centre National de la Recherche Scientifique–Projet International de Coopération Scientifique [Grant 3603]; and an Australian Postgraduate Award.
Article, publication date, and citation information can be found at http://molpharm.aspetjournals.org.
doi:10.1124/mol.110.070540.
-
ABBREVIATIONS:
- NaV channel
- voltage-activated sodium channel
- 4-AP
- 4-aminopyridine
- BKCa channel
- large-conductance Ca2+ and voltage-activated K+ channel (KCa1.1, Maxi-K, BK, Slo1)
- CaV channel
- voltage-activated Ca2+ channel
- ChTx
- charybdotoxin (potassium channel scorpion toxin 1.1)
- DUM
- dorsal unpaired median
- HVA
- high-voltage-activated
- HXTX
- hexatoxin (from the venom of spiders belonging to the family Hexathelidae)
- IKCa
- intermediate-conductance Ca2+-activated K+ channel (KCa3.1, IKCa1)
- IK(A)
- transient A-type K+ current
- IBK(Ca)
- Ca2+-activated K+ channel current
- IK(DR)
- delayed-rectifier K+ current
- α-KTx
- potassium channel scorpion toxin
- KV channel
- voltage-activated K+ channel
- M-LVA
- mid- to low-voltage-activated
- NIS
- normal insect saline
- SKCa channel
- small-conductance Ca2+-activated K+ channel (KCa2.x)
- TEA
- tetraethylammonium
- TRTX
- theraphotoxin (from the venom of spiders belonging to the family Theraphosidae)
- TTX
- tetrodotoxin
- HPLC
- high-performance liquid chromatography
- TFA
- trifluoroacetic acid
- RP-HPLC
- reversed-phase high-performance liquid chromatography
- ICK
- inhibitor cystine knot
- f
- fraction.
- Received December 14, 2010.
- Accepted March 14, 2011.
- Copyright © 2011 The American Society for Pharmacology and Experimental Therapeutics