The k-atracotoxins (k -ACTXs, previously the Janus-faced atracotoxins ) are a family of five insect-selective excitatory peptide neurotoxins containing 36-37 residues with four disulfide bonds. Toxins from this family were isolated from the venom of the Blue Mountains funnel-web spider (Hadronyche versuta) and Toowooba funnel-web spider (Hadronyche infensa). The NMR solution structure and primary sequence of the prototypic member k -ATCT-Hv1c provided few clues as to the likely molecular target. In order to characterise the site of action and phylogenetic specificity of these toxins, whole-cell patch-clamp electrophysiology was employed using isolated DUM neurons from the American cockroach (Periplaneta americana). k -ACTX-Hv1c had no effect on the gating or kinetics of INa or ICa at concentrations up to 1 μM. However, at the same concentration, k -ATCT-Hv1c reduced Kv channel currents by 56 ± 7% (n = 5). Subsequent experiments in insect DUM neurons indicated that inhibition of the macroscopic IK was due to a block of calcium-activated Kv (KCa) channels, with an IC50 of 2.3 nM and 2.9 nM for peak and late IK(Ca) respectively (n = 5), and not ‘A-type’ or delayed-rectifier Kv channels. Insect selectivity was confirmed by a lack of activity on rat dorsal root ganglion (DRG) neuron global IK as well as IK(Ca) at doses up to 1 μM. k -ACTX-Hv1c is a selective insect KCa (BKCa) channel pore-blocker, not a gating modifier, as inhibition of insect IK(Ca) occurred in the absence of any voltage-dependent actions on channel activation. Specificity for the insect BKCa channel was validated by k -ACTX-Hv1c induced inhibition of IK(Ca) from the cloned insect KCa channel a-subunit (pSlo) expressed in HEK293 cells (IC50 of 240 nM). The 80-fold reduction in IC50, most likely indicates that k-ACTX-Hv1c interacts with the auxillary subunits that form part of the wild-type channel, in a manner similar to the BKCa blocker, charybdotoxin (ChTX), as previously reported. Phyletic selectivity of k-ACTX-Hv1c was confirmed by the 9776-fold
increase in IC50 against mSlo channels. Interestingly k-ACTX-Hv1c, like ChTX, failed to potently block the dSlo channel with the IC50 >10 μM.
Additional experiments on DUM neuron IK(Ca) using alanine mutants confirmed the pharmacophore of bioactive residues k-ACTX-Hv1c comprises Arg8, Pro9, Val29 and Tyr31, previously identified by acute toxicity tests in house flies (Musca domestica). Interestingly, the functionally critical Arg8 and Tyr31 residues align extremelyç well with the Lys-Phe/Tyr diad conserved amongst structurally dissimilar Kv channel toxins, providing a possible basis for targeting of the toxin to K+ channels. Using a panel of 8 mutants (R8E, R8Q, R8K, R8H, Y31W, Y31F, Y31L and Y31V) the mechanism of interaction was investigated further. The Arg8 residue appears to interact with the channel via hydrogen bonding from the s-guanido group to carbonyl groups on the extracellular surface of the channel, as evidenced by the high potency of the R8H mutant. The imidazole group of His is an adequate substitute for the s-guanido group of arginine. In contrast the R8E, R8Q and R8K had reduced potency indicating that the positive charge of the amino group of Arg does not directly interact with the target nor is the alkyl group of Arg critical for binding to the target. The critically important Tyr31 interacts with the channel via non-specific hydrophobic interactions as substitution for an aromatic ring (Y31F & Y31W) maintains the potency of the toxin. In contrast substitution to small less hydrophobic side chains (Y31V, Y31L and Y31A) reduced potency. It appears therefore that Tyr31 in conjunction with IIe2 and Val29, that lie at either side of the primary pharmacophore, appear to act as ‘gasket’ residues to exclude bulk solvent from disrupting the Arg8-channel interaction.
This study has identified k-atracotoxins as potential lead compounds in the development of new biopesticides and validates insect BKCa channels as potential insecticide targets.
The second part of this thesis was to determine the target site for the ‘hybrid’ toxin FW178 from the venom of the Blue Mountains funnel-web spider (H. versuta). FW178 is a unique toxin that shares little homology to other known atracotoxins. In order to identify the site of action of this toxin, whole-cell patch-clamp electrophysiology was employed using isolated DUM neurons from the American cockroach (Periplaneta americana). FW178 failed to inhibit insect INa , IK(DR) or IK(A) at doses up to 1 μM. However, further studies demonstrated that Ƞ-ACTX-Hv1a blocks voltage-gated calcium (CaV) channel currents in DUM neurons as well as K(Ca) channel currents carried by pSlo channels with IC50 values of 409 nM and 671 nM, respectively. FW178 therefore blocks cockroach CaV currents with approximately the same potency as Ѡ -ACTX-Hv1a, a known insect M-LVA and HVA CaV channel blocker, while it blocks cockroach pSlo channels with about a 4-fold lower potency than k-ACTX-Hv1c.
Interestingly FW178 has an LD50 of 38 ± 3 pmol/g when injected into M. domestica as compared to the LD50 values for Ѡ -ACTX-Hv1a (86.5 ± 1.3 pmol/g) and Ѡ -ACTX-Hv1c (91 ± 5 pmol/g). This makes FW178 at least two-fold more potent than any other atracotoxin isolated from Australian funnel-web spiders. Despite this, FW178 only blocks cockroach CaV channels with a similar potency to k-ACTX-Hv1a, and blocks cockroach BK(Ca) with 4-fold less potency than k-ACTX-Hv1c. Therefore the striking potency of FW178 may result from a synergistic action to block insect CaV and BK(Ca) channels. Not surprising the pharmacophore of FW178 (refer section 4.3) contains elements of the pharmacophore of both Ѡ-ACTX-Hv1a and k-ACTX-Hv1c. Thus FW178 directly block insect K(Ca) channels, but the toxin also enhances this action by indirectly reducing current through these channels by block of the transient inward flow of calcium through CaV channels. Therefore FW178 represents the first known dual-target, self-synergizing toxin and is an excellent lead compound for the development of a novel insecticide.