The following is a brief overview of our research on the isolation and characterization of novel insecticidal neurotoxins. In addition to their utility as insecticide leads, these toxins have enabled the identification and validation of novel insecticide targets, and ultimately they should prove to be valuable pharmacological tools. The long-term goal of our research, which is funded by the U.S. National Science Foundation, is to develop safer biopesticides and chemical insecticides. Much of this summary is extracted from a recent review [King, G.F., Tedford, H.W. & Maggio, F. (2002) J. Toxicol. Toxin Reviews, in press] which should be consulted for further details.
1. The ongoing war against arthropod pests
2. Insecticide design: what can we learn from venomous spiders?
3. Australian funnel-web spiders: bad spider makes good
4. The omega-atracotoxin-1 family
5. The omega-atracotoxin-2 family
6. Invertebrate calcium channels: are they good insecticide targets?
7. The enigmatic Janus-faced atracotoxins
8. Atracotoxin evolution
9. Future prospects
10. Contact details
11. References
In addition to destroying an estimated 20-30% of the world’s food supply [1], arthropod pests are responsible for the transmission of many new and reemerging human diseases [2]. Mosquitoes are the most pernicious athropods from a human health perspective, being responsible for the transmission of dengue fever, yellow fever, West Nile virus, filariasis, various forms of encephalitis, and malaria, with the latter causing at least one million deaths annually [2,3].
Arthropod pests have largely been controlled by spraying broad-spectrum chemical insecticides (we currently spray more than one billion pounds of pesticide per year in the United States). However, the long-term application of a small armament of insecticides that act on a restricted number of invertebrate nervous system targets has inevitably led to the development of resistance in most agrochemically and medically important arthropods [2,4]. Furthermore, most chemical insecticides are relatively non-specific. It has been estimated that they kill about 70 million birds and 10 million fish per year. Moreover, there is now strong epidemiological [5,6] and experimental [7,8] evidence linking certain neurological disorders, such as Parkinson’s disease, to pesticide exposure. Thus, we believe that it is critical at this juncture to identify and characterize novel insecticidal compounds as well as new insecticide targets. These objectives are the current focus of the laboratory, while our long-term goal is to develop safer biopesticides and chemical insecticides.
Most spider venoms are likely to be rich sources of insecticidal compounds since their primary role is to kill or paralyze arthropod prey. Thus, it seems surprising that spider venoms, which as a whole are likely to contain more than a million different pharmacologically active peptides [9], have not been explored as thoroughly as those of other venomous creatures such as scorpions and cone snails. While numerous peptide neurotoxins have been isolated from spider venoms [10,11], most were purified based on activity against vertebrate targets. These toxins generally show little or no preference for invertebrates [12] and hence they are not particularly useful as insecticide leads. However, in contrast to most other studies on arachnid toxins, our laboratory has focused on the isolation and characterization of insect-specific spider toxins that might be useful as insecticide leads and for defining novel insecticide targets.
Back to top
Spiders are taxonomically divided into two major suborders, Araneomorphae and Mygalomorphae. The Araneomorphs, or modern spiders, evolved after the appearance of flying insects, and they now represent ~90% of the world’s spiders. Many of these spiders have devised elaborate means of using their silk to capture flying prey, the most spectacular being the orb-web weavers. Araneomorphs generally live for only 1-2 years, and, because of their small size, venom acquisition often requires electrical stimulation or complete removal of the venom glands.
The ancestral lineage of the Mygalomorphs, or primitive spiders, dates back 360 million years, before the evolution of flying insects [13]. These spiders have several qualities that recommend them as a venom source for insect-toxin screens. First, Mygalomorphs only use silk in a rudimentary manner, relying instead on their physical size and often massive venom apparatus to immobilize prey. Second, Mygalomorphs usually have long-life spans, with the females of many species living longer than 20 years [13]. Third, when provoked, many of these spiders adopt an aggressive/defensive "ready-to-strike" stance with forelegs and palps raised and fangs exposed. This enables venom to be aspirated directly from the fang tips (see Figure 1) without the need for electrical stimulation which can contaminate the venom with enzymes from saliva and digestive fluids [10]. The combination of long lifespan and facile milkings enables sufficient venom to be obtained from a relatively small cohort of captive spiders for detailed biochemical and biophysical characterization of individual venom components.
Australian funnel-web spiders (Araneae:Mygalomorphae:Hexathelidae:Atracinae) are currently divided into two genera, Atrax and Hadronyche [14]. There are at least 40 different species, although many remain undescribed. The spiders are nocturnal, and most are terrestial burrowers. They spend the day huddled in a small chamber at the bottom of a long burrow in the ground [13]. The burrow is lined with a silk tube, hence the name "funnel-web spider". At night the spiders ascend the silken tube and position themselves at the burrow entrance waiting for insect prey to stumble into striking range. As a group these are probably the world’s deadliest spiders [15], with envenomation by the Sydney funnel-web spider, Atrax robustus, having caused at least 14 deaths prior to the introduction of an antivenom in 1980 [16]. However, it is important to recognize that the vertebrate toxity of the venom is attributable to single component out of the more than 100 compounds that are present in the venom. The majority of venom components are not toxic to vertebrates, as one might expect given that these spiders prey mainly on arthropods.
Atkinson and coworkers were the first to recognize the potent insecticidal properties of Australian funnel-web spider venom [17,18]. By testing the venoms of a taxonomically diverse group of Australian spiders [18] they showed that funnel-web venom was the most lethal against larvae of the moth Helicoverpa armigera (cotton bollworm), a refractory agricultural pest. This led to the identification of a potent insect-specific neurotoxin [18] that is now referred to as w-atracotoxin-1 [19]. All toxins from Australian funnel-web spiders are referred to by the generic name atracotoxin as all of these spiders, regardless of genus, belong to the Atracinae subfamily [19].
We subsequently implemented a detailed screen of venom from the Blue Mountains funnel-web spider, Hadronyche versuta. This screen led to the discovery of several families of novel insect-specific neurotoxins [12,20] as well as additional members of the w-atracotoxin-1 family [21]. It should be stressed that these toxins are highly insect-specific; they have no activity in any vertebrate system that has been tested thus far.
The first family of insect-specific neurotoxins isolated from Australian funnel-web spiders was the w-ACTX-1 family [17]. These peptides each contain 36-37 residues with six strictly conserved cysteine residues that form three disulfide bonds. A single species of spider may contain six or more variants of the toxin, with some variants differing by only a single conservative residue substitution [21]. The w-ACTX-1 family of toxins are lethal to a wide range of insects, including members from the orders Coleoptera, Orthoptera, Lepidoptera, and Diptera, but they are harmless when injected at high doses into newborn mice [17,19,22]. Injection of toxin into American cockroaches (Periplaneta americana) causes a loss of locomotion, high-frequency twitching of limbs with loss of righting reflexes, followed by paralysis and death [19]. Direct application of toxin to the cockroach metathoracic ganglion abolishes hind-limb reflexes, whereas the forelimbs, which are not directly innervated by motor neurones of the metathoracic ganglion, are unaffected [19]. These peptides can therefore be classified as depressant neurotoxins. Electrophysiological studies [19] revealed that the phylogenetic specificity of the toxins derives from their ability to block insect, but not vertebrate, voltage-gated calciumchannels (VGCCs). The channel subtype remains to be definitely determined but our preliminary studies suggest that it is an N- or L-like VGCC.
We determined the three-dimensional solution structure of the prototypic family member, w-ACTX-Hv1a, using NMR spectroscopy [19]. The structure is not closely related to any of the currently available structures of vertebrate VGCC blockers such as w-agatoxin-IVA [23], w-conotoxin GIVA [24], and w-conotoxin TxVII [25].The structure of w-ACTX-Hv1a comprises a structurally disordered N-terminus (residues 1-3), a disulfide-rich globular core (residues 4-21), and a finger-like b-hairpin (residues 22-37) that protrudes from the globular domain. The three disulfide bonds form an inhibitory cystine knot (ICK) motif [36,37] in which the Cys17-Cys36 disulfide passes through a 13-residue ring formed by the Cys4-Cys18 and Cys11-Cys22 disulfide bridges and the intervening sections of polypeptide backbone. Except for the special case of cyclic ICK peptides, cystine knots are not true knots in the mathematical sense as they can be untied by a non-bond-breaking geometrical transformation [38]. Nevertheless, the cystine knot provides ICK toxins with tremendous stability and resistance to proteases, thereby ensuring that the injected toxin is not degraded before reaching its target.
We subsequently developed an efficient bacterial expression system for production of recombinant w-ACTX-Hv1a so that the functional importance of individual residues could be examined using alanine scanning mutagenesis [22]. Remarkably, this appears to have been the first recombinant expression system developed for a spider-derived neurotoxin. In alanine scanning mutagenesis, each amino acid residue is separately replaced by an alanine (except alanine, which we replace with serine) and the the activity of each resulting point mutant is examined. Alanine is generally chosen for scanning mutagenesis because it can be accommodated in most types of secondary structure (b sheet, a helix, and b turn), thus minimizing the chances that the point mutation will induce major structural perturbations. The mutagenesis study of w-ACTX-Hv1a revealed that the complete bioactive surface or "pharmacophore" of the toxin is restricted to just four residues that form a contiguous patch on a single face of the toxin, which is encouraging for the design of functional small-molecule mimics.
We recently isolated a second family of w-atracotoxins that specifically block insect VGCCs [12]. These toxins contain 41-45 residues with six conserved cysteines that form three disulfide bonds. The sequences are unrelated to the w-ACTX-1 toxins and there are no homologs in the protein/DNA sequence databases. As for the w-ACTX-1 family, a single spider may contain more than one variant of the toxin. These toxins cause a prolonged but reversible paralysis in a range of insects including dipterans. At moderate to high doses the paralysis is sustained for ~6 h in house crickets (Acheta domestica) and ~10 h in flies (Musca domestica).
The prototypic family member, w-ACTX-Hv2a, inhibits bee brain calcium currents with an EC50 of ~130 pM, whereas it has virtually no effect on calcium currents in rat trigeminal neurons at a concentration of 1 mM [12]. This indicates not only that these toxins are extremely potent blockers of insect VGCCs but that they have exceptional phylogenetic specificity, with at least a 10,000-fold preference for bee-brain versus rat-brain VGCCs. w-Agatoxin-IVA, the classical blocker of vertebrate P/Q-type VGCCs, has a marginal 2-fold preference for invertebrate VGCCs in the same assay [12]. Thus, to the best of our knowledge, w-ACTX-Hv2a is the most potent and specific inhibitor of insect calcium channels discovered to date.
There are several lines of evidence suggesting that the w-ACTX-1 and w-ACTX-2 toxins target different subtypes of insect VGCCs: (i) the w-ACTX-1 toxins are lethal to orthopterans and dipterans, whereas w-ACTX-Hv2a causes sustained but reversible paralysis in these insects; (ii) at saturating concentrations of toxin, w-ACTX-Hv1a and w-ACTX-Hv2a differ in the extent of calcium current blockage in various insect neurons [12,19]; (iii) the 3D structures of the toxins are vastly different (see below). These two families of VGCC blockers presumably play different but synergistic roles in funnel-web spider venom; it is likely that the w-ACTX-2 toxins are important for the initial rapid incapacitation of prey whereas the w-ACTX-1 toxins contribute to slower-onset, but irreversible, paralysis.
The solution structure of w-ACTX-Hv2a was determined using NMR spectroscopy. The disulfide-rich region of the toxin is organized into a compact globular domain containing a short 310 helix (residues 13-17), a b-hairpin comprising residues 23-30, and several well defined b-turns [12]. The three disulfide bridges form a cystine-knot in which the Cys17-Cys29 disulfide pierces a 15-residue ring formed by the two other disulfide bridges and the intervening sections of polypeptide backbone. The two N-terminal residues and the entire C-terminal region (residues 33-45) are completely disordered in solution. This is in striking contrast to most other ICK toxins, which are exemplars of economical protein engineering in that almost every single residue contributes in some way to the 3D fold.
The 3D structure of w-ACTX-Hv2a is very different to that of w-ACTX-Hv1a even though they both target VGCCs, which highlights the difficulty of predicting toxin function from structure. There are no close structural homologs of w-ACTX-Hv2a in the protein database but the toxin has certain structural, chemical, and functional similarities with w-agatoxin-IVA from the unrelated American funnel-web spider Agelenopsis aperta, viz: (i) the core ICK regions of the two toxins can be superimposed so that the central disulfide bridges and the b-strands overlay reasonably well; (ii) both toxins contain a highly disordered, lipophilic C-terminal extension that protrudes from the globular domain; (iii) both toxins target VGCCs although, as outlined above, w-ACTX-Hv2a has a much higher specificity for invertebrate VGCCs; (iv) deletion of the disordered C-terminal tail from either toxin completely abolishes their ability to block VGCCs [12,23].
The structural and functional similarities between w-ACTX-Hv2a and w-agatoxin-IVA suggest a similar mode of action [12]. A C-terminal truncation mutant of w-ACTX-Hv2a comprising only residues 1-32 of the toxin neither blocks invertebrate VGCCs nor competitively inhibits the wild-type toxin; this implies that the lipophilic C-terminal tail is essential for interaction with VGCCs and that it initiates toxin binding. Hence, it has been suggested [12] that the lipophilic tails of w-ACTX-Hv2a and w-agatoxin-IVA initiate toxin binding by penetrating the membrane either adjacent to the channel or, perhaps more likely, by intercalation between transmembrane segments of the channel protein. Anchoring of the C-terminal tail in the membrane might alter the channel conformation sufficiently to reveal an otherwise cryptic high-affinity binding site for the disulfide-rich core region of the toxin. While this hypothesis is highly speculative, it is salient to note that the insect-VGCC blocker PLTX-II from the spider Plectreurys tristis contains a C-terminal palmitoyl group and therefore it may function similarly.
In terms of insecticide development, the w-atracotoxins are of considerable interest because they act on an non-conventional target, namely insect VGCCs. In contrast, most commonly used insecticides target insect voltage-gated sodium channels (e.g., DDT, pyrethroids), acetylcholinesterase (e.g., organophosphates, carbamates), or GABA receptors (e.g., the arylheterocycles endosulfan and fipronil) [26].
Vertebrate VGCCs are well characterized at both the pharmacological and molecular level [27,28]. They are classified as high-voltage-activated (HVA) or low-voltage-activated (LVA) based on the voltage required for channel activation (about -30 and -60 mv, respectively). Cloning of channel genes has revealed that the subtype classifications originally based on the pharmacological properties of the channels correlate with the type of pore-forming a1 subunit present in the channel. Thus, HVA channels are subdivided into L-type (a1S, a1C, a1D, and a1F subunits), N-type (a1B), P/Q-type (a1A), and R-type (a1E), while LVA currents are carried by T-type channels (a1G, a1H, and a1I) [27].<
Invertebrate VGCCs are less well characterized, both pharmacologically and at the molecular level [27,28]. Sequencing of the Drosophila genome revealed four genes encoding VGCC a1 subunits, including homologs of vertebrate T-type, L-type, and N-type channels as well as a homolog of two novel C. elegans VGCC-coding genes. However, the homology between Drosophila and vertebrate a1 subunits can be as low as 50%, perhaps explaining why the pharmacological classifications developed for vertebrate VGCCs have proved unsuitable for invertebrates. This turns out to be a two-edged sword. On one hand, it makes it difficult to specifically identify the target of an invertebrate-specific VGCC toxin since the pharmacological distinctions used for vertebrate VGCCs are not applicable and cloned invertebrate VGCCs are not readily available. On the other hand, as demonstrated by the two families of w-atracotoxins, the invertebrate and vertebrate channels differ sufficiently at the molecular level such that toxins can be isolated that specifically target invertebrate, but not vertebrate, VGCCs.
The selective lethality of the w-atracotoxins demonstrates the validity of VGCCs as insecticide targets, as might be expected given the critical role of insect VGCCs in modulating cellular excitability and neurotransmitter release. The w-atracotoxins might also prove to be valuable tools for the pharmacological classification of invertebrate VGCCs since the w-ACTX-1 and w-ACTX-2 toxin families almost certainly target different calcium channel subtypes. Continued efforts at mapping the pharmacophores of these toxins using site-directed mutagenesis should help to elucidate the architecture of the extracellular surfaces of invertebrate VGCCs and, in particular, how they differ from their vertebrate counterparts.
The Janus-faced atracotoxins (J-ACTXs) are a recently discovered family of unusual insect-specific neurotoxins [20]. The toxins each contain 36-37 residues with eight conserved cysteine residues that are paired to form four disulfide bonds. Direct application of the prototypic family member J-ACTX-Hv1c to the cockroach metathoracic ganglion causes immediate fasciculations of the limbs and posterior sensory organs (cerci), indicating that it is an excitatory neurotoxin [20]. This contrasts with w-ACTX-Hv1a which blocks limb reflexes in this assay [19].
The 3D structure of J-ACTX-Hv1c consists of a disulfide-rich core (residues 3-19) with residues 20-34 forming a b-hairpin that projects from this globular region [20]. Residues 1-2 and 35-37 at the N- and C-termini, respectively, are disordered in solution. The three central disulfides form a cystine knot in which the Cys16-Cys32 disulfide passes through a 14-residue ring formed by the Cys10-Cys22 and Cys3-Cys17 disulfide bridges and the intervening sections of polypeptide backbone. An electrostatic surface map of J-ACTX-Hv1c revealed that almost all of the hydrophobic residues are conspicuously clustered on one face of the molecule, while the opposing face presents a ladder of positively and negatively charged residues. Thus, the toxins were named the Janus-faced atracotoxins after Janus, the two-faced Roman god [20]. Given that the likely target of the J-ACTXs is an invertebrate ion channel (see below), it is fitting that Janus was the god of gates and doors in Roman mythology.
The most remarkable feature of the structure of J-ACTX-Hv1c is a vicinal disulfide bridge between the sidechains of Cys13 and Cys14 [20]. There are currently no other examples of vicinal disulfides in peptide toxins, and J-ACTX-Hv1c is one of only three proteins that are known to contain a vicinal disulfide, the others being found at the active site of methanol dehydrogenase (MDH) [30] and at the acetylcholine binding site on the a subunit of the acetylcholine receptor (aAChR) [31]. A vicinal disulfide bridge (i.e., a disulfide bond between adjacent cystine residues) can only be formed by significantly distorting the planarity of the peptide bond between the connected residues. The Cys13-Cys14 peptide bond in J-ACTX-Hv1c adopts a highly distorted trans configuration with an w torsion angle of —135 degrees (c.f. 180 +/- 5 degrees for a tyical trans peptide bond).
The three disulfide bridges in the ICK motif of J-ACTX-Hv1c direct the tertiary fold of the peptide as they connect residues that are distal in the toxin sequence. However, the vicinal disulfide does not play an important architectural role since the 3D structure of a C13S,C14S double-mutant is essentially identical to that of native J-ACTX-Hv1c [20]. However, insecticidal activity is almost completely abrogated in the C13S,C14S double-mutant, demonstrating that the vicinal disulfide plays a key functional role [20]. The vicinal disulfides also play functional roles in MDH and aAChR where they are involved in enzyme catalysis [30] and acetylcholine binding [32], respectively.
We recently developed a recombinant bacterial expression system for J-ACTX-Hv1c, which enabled the complete toxin pharmacophore to be mapped using alanine-scanning mutagenesis [33]. This study represents the first complete mapping of the bioactive surface of a spider toxin. The J-ACTX-Hv1c pharmacophore is restricted to seven residues that form a bipartite surface patch on one face of the toxin. However, the primary pharmacophore (the "hot spot") is formed by just five residues (Arg8, Pro9, Tyr31, and the Cys13-Cys14 vicinal disulfide), with two flanking residues (Ile2 and Val29) being less important for insecticidal potency [33]. This is an important observation from the viewpoint of insecticide design, since restriction of the pharmacophore to a single surface of the toxin should increase the probability of successfully designing small-molecule mimics of the J-ACTXs.
These toxins belong to a functionally diverse fold-family that includes toxins that interact with VGCCs, VGSCs, and voltage-gated potassium channels. This highlights the point made above that the 3D fold of ICK toxins generally provides little insight into their mode of action. However, for cystine-knot and other disulfide-rich ion channel toxins, the topological disposition of key functional residues, regardless of three-dimensional scaffold, is often informative of function. Thus, it is intriguing that the Arg8-Tyr31 diad in J-ACTX-Hv1c pharmacophore superimposes very well on the functionally critical Lys-Tyr/Phe diad that is spatially conserved across a range of structurally dissimilar potassium channel blockers [34]. Thus, it is possible that the J-ACTXs target invertebrate potassium channels [33]. We are currently testing this hypothesis by employing a variety of biochemical, electrophysiological, and genetic screens in order to identify the exact molecular target of the J-ACTXs and other atracotoxins.
Pioneering studies by Olivera and coworkers (reviewed in Refs 35 & 36) have revealed that the remarkable diversity of cone-snail venoms is the product of a combinatorial peptide library strategy employed by the snails. Mature conotoxins are derived from a mRNA translation product that comprises an N-terminal signal sequence, a central propeptide, and a C-terminal mature-toxin sequence. The signal sequence is presumably critical for targeting the toxin to specific secretory pathways and consequently it is highly conserved within a particular conotoxin superfamily [36]. In striking contrast, the mature toxin sequence is hypermutated, with only the cystine framework being conserved amongst the different superfamily members [35,36]. This has enabled the cone snails to evolve a combinatorial peptide library in which hypermutable loops are displayed on the same cystine-stabilized scaffold.
While the available data is still scarce, recent analysis of cDNA libraries from Australian funnel-web spiders suggests that these spiders have employed a similar combinatorial peptide library strategy for diversifying their toxin pool. For example, the mature w-ACTX-2 toxins are derived from a prepropeptide that contains an N-terminal 23-residue signal sequence followed by a 33-residue propeptide sequence that precedes the mature toxin. The number and position of the cysteine residues in the mature toxins is strictly preserved, as seen in the cone snails, but the mature toxin sequences (53% identity across the four toxins) are significantly more variable than the signal sequences (78% identity, and 100% similarity if conservative substitutions are included). A key unanswered question is whether a single atracotoxin superfamily (as defined by the signal sequence plus the cystine framework of the mature toxin) can include toxins with different functions as would be expected if the spiders have employed a hypermutation strategy (as opposed, for example, to a biased hypermutation that conserves both the cystine framework and several key residues that confer binding to a specific target molecule).
From the viewpoint of insecticide design, the atracotoxins described above are useful lead compounds for several reasons: (i) they are active against a broad range of arthropod pests of agricultural and medical importance; (ii) they act on non-conventional neuronal targets, making them invaluable tools for the characterization of new insecticide targets; (iii) they are efficiently expressed in heterologous systems, and they do not require any posttranslational modifications, making them suitable for biopesticide applications; (iv) scanning mutagenesis has revealed that the pharmacophore of each toxin is restricted to a small number of spatially contiguous residues, making them useful leads for the rational design of novel chemical insecticides. Current work in the lab is focused on obtaining a more detailed definition of the toxin pharmacophores and elucidating the precise molecular targets of these toxins. We also attempting to address several fundamental questions regarding the expression and evolution of the atracotoxins.
Back to topIf you would like to learn more about these fascinating toxins, consult some of the primary references listed below and/or contact the graduate students working on these toxins: w-ACTX-1 (Billy Tedford: billy@psel.uchc.edu), J-ACTX (Frank Maggio: frank@psel.uchc.edu), and w-ACTX-2 (Brianna Sollod: brianna@psel.uchc.edu).
Back to top1. Oerke, E.-C. (1994) in Crop production and crop protection-estimated losses in major food and cash crops (Oerke, E.-C., Dehne, H.-W., Schönbeck, F., and Weber, A., eds), Elsevier, Amsterdam
2. Brogdon, W.G. & McAllister, J.C. (1998) Emerging Infectious Diseases 4, 605-613.
3. Greenwood, B. & Mutabingwa, T. (2002) Nature 415, 670-672.
4. Feyereisen, R. (1995) Toxicology Letters 82-83, 83-90.
5. Gorell, J.M., Johnson, C.C., Rybicki, B.A., Peterson, E L. & Richardson, R.J. (1998) Neurology 50, 1346-1350.
6. Le Couteur, D.G., McLean, A.J., Taylor, M.C., Woodham, B.L. & Board, P.G. (1999) Biomedicine and Pharmacotherapy 53, 122-130.
7. Betarbet, R., Sherer, T.B., MacKenzie, G., Garcia-Osuna, M., Panov, A.V. & Greenamyre, J.T. (2000) Nature Neuroscience 3, 1301-1306.
8. Uversky, V.N., Li, J. & Fink, A.L. (2001) FEBS Letters 500, 105-108.
9. King, G.F., Teford, H.W. & Maggio, F. (2002) Journal of Toxicology: Toxin Reviews, in press.
10. Escoubas, P., Diochot, S. & Corzo, G. (2000) Biochimie 82, 893-907.
11. Rash, L.D. & Hodgson, W.C. (2002) Toxicon 40, 225-254.
12. Wang, X.-H., Connor, M., Wilson, D., Wilson, H.I., Nicholson, G.M., Smith, R., Shaw, D., Mackay, J.P., Alewood, P.F., Christie, M.J. & King, G.F. (2001) Journal of Biological Chemistry 276, 40806-40812.
13. Brunet, B. (1998) Spiderwatch: a guide to Australian spiders, New Holland Publishers, Sydney
14. Gray, M. R. (1988) in Australian arachnology (Austin, A. D., and Heather, N. W., eds), The Australian Entomological Society, Brisbane
15. Miller, M.K., Whyte, I.M., White, J. & Keir, P.M. (2000) Toxicon 38, 409-427.
16. Sutherland, S. K. (1980) Medical Journal of Australia 2, 437-441.
17. Atkinson, R.K., Tyler, M.I., Vonarx, E J. & Howden, M.E.H. (1993) International Patent WO 93/15108.
18. Atkinson, R.K., Vonarx, E.J. & Howden, M.E.H. (1996) Comparative Biochemistry and Physiology 114C, 113-117.
19. Fletcher, J.I., Smith, R., O'Donoghue, S.I., Nilges, M., Connor, M., Howden, M.E.H., Christie, M.J. & King, G.F. (1997) Nature Structural Biology 4, 559-566.
20. Wang, X.-H., Connor, M., Smith, R., Maciejewski, M.W., Howden, M.E.H., Nicholson, G.M., Christie, M.J. & King, G.F. (2000) Nature Structural Biology 7, 505-513.
21. Wang, X.-H., Smith, R., Fletcher, J.I., Wilson, H., Wood, C.J., Howden, M.E H. & King, G.F. (1999) European Journal of Biochemistry 264, 488-494.
22. Tedford, H.W., Fletcher, J I. & King, G.F. (2001) Journal of Biological Chemistry 276, 26568-26576.
23. Kim, J.I., Konishi, S., Iwai, H., Kohno, T., Gouda, H., Shimada, I., Sato, K. & Arata, Y. (1995) Journal of Molecular Biology 250, 659-671.
24. Pallaghy, P. K. & Norton, R. S. (1999) Journal of Peptide Research 53, 343-351.
25. Kobayashi, K., Sasaki, T., Sato, K. & Kohno, T. (2000) Biochemistry 39, 14761-14767.
26. ffrench-Constant, R.H., Pittendringh, B., Vaughan, A. & Anthony, N. (1998) Philosophical Transactions of the Royal Society of London Series B 353, 1685-1693.
27. Jeziorski, M.C., Greenberg, R.M. & Anderson, P.A.V. (2000) Journal of Experimental Biology 203, 841-856.
28. Wicher, D., Walther, C. & Wicher, C. (2001) Progress in Neurobiology 64, 431-525.
29. Maggio, F. & King, G. F. (2002) Toxicon 40, 1355-1361.
30. Blake, C.C.F., Ghosh, M., Harlos, K., Avezoux, A. & Anthony, C. (1994) Nature Structural Biology 1, 103-105.
31. Kao, P.N. & Karlin, A. (1986) Journal of Biological Chemistry 261, 8085-8088.
32. Czajkowski, C. & Karlin, A. (1995) Journal of Biological Chemistry 270, 3160-3164.
33. Maggio, F. & King, G.F. (2002) Journal of Biological Chemistry 277, 22806-22813.
34. Dauplais, M., Lecoq, A., Song, J., Cotton, J., Jamin, N., Gilquin, B., Roumestand, C., Vita, C., Medeiros, C. L. C. d., Rowan, E. G., Harvey, A. L. & Ménez, A. (1997)Journal of Biological Chemistry 272, 4302-4309.
35. Olivera, B.M., Hillyard, D.R., Marsh, M. & Yoshikami, D. (1995) Trends in Biotechnology 13, 422-426.
36. Pallaghy, P.K., Nielsen, K.J., Craik, D.J. & Norton, R.S. (1994) Protein Science 3: 1833-1839.
37. Norton, R.S. and Pallaghy, P.K. (1998) The cystine knot structure of ion channel toxins and related polypeptides. Toxicon 36, 1573-1583.
38. Craik, D.J., Daly, N.L. & Waine, C. (2001) Toxicon 39, 43-60.