Insecticidal toxins from black widow spider venom

Abstract

The biological effects of Latrodectus spider venom are similar in animals from different phyla, but these symptoms are caused by distinct phylum-specific neurotoxins (collectively called latrotoxins) with molecular masses ranging from 110 to 140 kDa. To date, the venom has been found to contain five insecticidal toxins, termed alpha, beta, gamma, delta and epsilon-latroinsectotoxins (LITs). There is also a vertebrate-specific neurotoxin, alpha-latrotoxin (alpha-LTX), and one toxin affecting crustaceans, alpha-latrocrustatoxin (alpha-LCT). These toxins stimulate massive release of neurotransmitters from nerve terminals and act (1) by binding to specific receptors, some of which mediate an exocytotic signal, and (2) by inserting themselves into the membrane and forming ion-permeable pores. Specific receptors for LITs have yet to be identified, but all three classes of vertebrate receptors known to bind alpha-LTX are also present in insects. All LTXs whose structures have been elucidated (alpha-LIT, delta-LIT, alpha-LTX and alpha-LCT) are highly homologous and have a similar domain architecture, which consists of a unique N-terminal sequence and a large domain composed of 13-22 ankyrin repeats. Three-dimensional (3D) structure analysis, so far done for alpha-LTX only, has revealed its dimeric nature and an ability to form symmetrical tetramers, a feature probably common to all LTXs. Only tetramers have been observed to insert into membranes and form pores. A preliminary 3D reconstruction of a delta-LIT monomer demonstrates the spatial similarity of this toxin to the monomer of alpha-LTX.

Fig. 1

Comparison of latroinsectotoxins (LITs) with α-latrotoxin α-LTX. (A) All latrotoxins are post-translationally processed to yield mature, active toxins. Black arrowheads: potential sites for proteolysis of α-LTX, α-LIT and δ-LIT, which share the consensus sequence K/R−Φ_1–3−K/R₀₋₁−R↓, where K and R are lysines and arginines, respectively; Φ is a hydrophobic/aromatic amino acid and ↓ denotes the cleaved peptide bond. (B) Domain organisation of α-LIT and δ-LIT in comparison to α-LTX. Top, the amino acid position scale and 3D domain organisation of α-LTX (wing/body/head). The amino acid sequences, presented diagrammatically below, were aligned using a method best suited for sequences related by descent (Hein, 1990). The unique N-terminal domains are depicted as solid black lines. ARs are shown as boxes and are numbered above the α-LTX structure; sequences not recognised by the PFAM database algorithm are coloured grey. Gaps in the alignment are indicated with dots. The four-residue insertion in the α-LTX^N4C mutant (see text) is denoted by an open arrowhead. Numbers between the diagrams indicate percentage identities between the respective domains (numbers below the δ-LIT structure are relative to α-LTX).

Fig. 2

The 3D structures of α-LTX and δ-LIT monomers. Comparison of the side and top views of the α-LTX monomer obtained by single-particle analysis of cryo-electron micrographs (Orlova et al., 2000) with similar views of the δ-LIT monomer, reconstructed by single-particle analysis of negative-stain electron micrographs. The domains of α-LTX are marked as in (Orlova et al., 2000); tentative domain assignments in the δ-LIT reconstruction are based on the interactive overlap of the 3D maps, fitting of ARs and volumetric analysis. Both volumes were low-pass filtered to 20 Å. (J. Nield, R. Abbondati, B. Odier, A. Rohou, Y. Ushkaryov, unpublished results).

Fig. 3

Dimers and tetramers of α-LTX and the mechanism of membrane pore formation. (A, B) Top views of the two major multimeric species of α-LTX. The monomers are coloured alternately. Both dimers and tetramers are able to bind receptors but only the tetramer forms integral membrane pores. (C) Cut-open side view of the α-LTX tetramer inserted into a lipid bilayer. The pore in the centre of the tetramer is permeable to ions and cytosolic neurotransmitters. Approximate positions of trypsin cleavage sites A and B (see text) are indicated with arrows. The locations of ARs 14–22 are shown as numbers in the cross-section of one monomer. (D) A predicted steric hindrance mechanism which may prevent membrane pore formation by N-terminal fusion constructs of α-LTX. Sections through glutathione-S-transferase (GST) are shown to scale.

Fig. 4

Evolutionary conservation of the three types of α-LTX receptors. (A) Neurexins in cow and Drosophila. Vertebrates possess three neurexin genes, each of which encodes a long (α) and a short (β) form of neurexin. (B) Three latrophilins are found in vertebrates (LPH1 from the cow, Bos taurus, is shown), whereas only one form is present in insects (Drosophila and house fly, M. domestica). The nematode (C. elegans) LPH is illustrated for comparison because it binds ε-LIT from Latrodectus venom (Mee et al., 2004). (C) Multiple sequence alignment of the GPS domains from cow, C; Drosophila, D; housefly, HF; and C. elegans, CE. Identical residues are highlighted. (D) Receptor-like protein tyrosine phosphatase σ (PTPσ) in mouse and Drosophila. (A–D) Percentage identities are indicated by numbers between the respective domains (values below the C. elegans structure correspond to the cow sequence). The minimal protein regions required for α-LTX binding are indicated by square brackets above the diagrams. Domain names are abbreviated as follows: CH, O-linked carbohydrate domain; EGF, epidermal growth factor-like domains; FN3, fibronectin III-like domain; GBL, galactose-binding lectin-like domain; GPS, G-protein-coupled receptor proteolysis site; HRM, hormone receptor motif; Ig, immunoglobulin-like domain; LNS, laminin G-domain/neurexin/sex hormone binding protein repeat; SP, signal peptide; STP, serine/threonine/proline-rich region; Pro, proline-rich region; PTPase, protein tyrosine phosphatase; TMR, transmembrane region. Black arrows in A denote the site of alternative splicing #4 within the last LNS repeat; small black arrowheads in B and D indicate the sites of constitutive proteolysis.