Toxicon 43 (2004) 527–542 www.elsevier.com/locate/toxicon
Review
The multiple actions of black widow spider toxins and their selective use in neurosecretion studies Y.A. Ushkaryov*, K.E. Volynski, A.C. Ashton Department of Biological Sciences, Imperial College, London, SW7 2AY, UK
Abstract The black widow spider venom contains several large protein toxins—latrotoxins—that are selectively targeted against different classes of animals: vertebrates, insects, and crustaceans. These toxins are synthesised as large precursors that undergo proteolytic processing and activation in the lumen of the venom gland. The mature latrotoxins demonstrate strong functional structure conservation and contain multiple ankyrin repeats, which mediate toxin oligomerisation. The three-dimensional structure has been determined for a-latrotoxin (aLTX), a representative venom component toxic to vertebrates. This reconstruction explains the mechanism of aLTX pore formation by showing that it forms tetrameric complexes, harbouring a central channel, and that it is able to insert into lipid membranes. All latrotoxins cause massive release of neurotransmitters from nerve terminals of respective animals after binding to specific neuronal receptors. A G protein-coupled receptor latrophilin and a single-transmembrane receptor neurexin have been identified as major high-affinity receptors for aLTX. Latrotoxins act by several Ca2þ-dependent and -independent mechanisms based on pore formation and activation of receptors. Mutant recombinant aLTX that does not form pores has been used to dissect the multiple actions of this toxin. As a result, important insights have been gained into the receptor signalling and the role of intracellular Ca2þ stores in the effect of aLTX. q 2004 Elsevier Ltd. All rights reserved. Keywords: Latrotoxin; Spider venom; Latrophilin; Neurexin; Pore formation
Contents 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. The specialisations of latrotoxins. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. The LTX family portrait . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Primary structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Oligomerisation and three-dimensional structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. Pore formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. aLTX receptors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6. Modes of action of latrotoxins. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.1. Ca2þ-dependent and Ca2þ-independent actions of aLTX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2. Effects mediated by aLTX pores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3. Receptor-mediated effects. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.4. Probable Ca2þ-independent mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7. Latrotoxins-versatile tools to study neuroexocytosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
527 528 529 529 531 532 534 535 535 536 537 538 539 539
1. Introduction * Corresponding author. Tel.: þ44-20-7594-5237; fax: þ 44-207594-5207. E-mail address:
[email protected] (Y.A. Ushkaryov). 0041-0101/$ - see front matter q 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.toxicon.2004.02.008
The notorious black widow spider (genus Latrodectus) belongs to the family of comb-footed cobweb spiders
528
Y.A. Ushkaryov et al. / Toxicon 43 (2004) 527–542
(Theridiidae: Arthropoda, Arachnida, Araneae) (Platnick, 1993; Platnick, 1997). Black widows are cosmopolitan and found throughout much of the world: China, Central Asia (L. mactans tredecimguttatus and L. lugubris, or karakurt), southern Europe (L. mactans tredecimguttatus), North and South America (L. mactans, L. geometricus, L. hesperus), India and Australia (L. hasselti, or red-back spider). The venom of Latrodectus spp. is a powerful stimulant of neurosecretion in different classes of animals (Longenecker et al., 1970; Frontali et al., 1972; Griffiths and Smyth, 1973; Kawai et al., 1972). In addition to Latrodectus, the Theridiidae family includes more than 50 other genera; two of these, Steatoda and Achaearanea, are closely related to Latrodectus and, being particularly synanthropic, are routinely misidentified as black widows. The symptoms of Steatoda spp. envenomation resemble those of latrodectism (Warrell et al., 1991) and can be treated successfully with red-back spider antivenom (South et al., 1998; Graudins et al., 2002). However, antibodies against the principal component of the black widow venom, a-latrotoxin (aLTX), do not crossreact with Steatoda venom (Cavalieri et al., 1987), suggesting that some other functional constituents of these theridiid venoms are similar (see also Gillingwater et al., 1999). Over the last thirty years, the toxins from the black widow spider venom—latrotoxins—have been extensively used to study the molecular mechanisms of neurosecretion in vertebrates, insects and crustaceans. Recent advances in the structural and functional analysis of some latrotoxins, especially aLTX, have radically improved our understanding of the complex actions of these toxins and brought about important insights into the mechanisms regulating neurotransmitter release. This review will summarise the contemporary views regarding the structures and modes of action of latrotoxins and their application as neurobiology tools.
2. The specialisations of latrotoxins Most of the work on the biochemistry and physiology of latrotoxins was carried out using venoms from the Mediterranean and Central Asian varieties of L. mactans. The first attempt to separate the venom from L. mactans tredecimguttatus, using gel-filtration followed by ionexchange chromatography, without achieving complete purity, clearly indicated the presence of several proteins selectively toxic to vertebrates, insects or crustaceans (Granata et al., 1972; Frontali et al., 1976; Knipper et al., 1986). To date, at least seven different latrotoxins—principal toxic components of the venom—have been isolated from the Central Asian L. mactans tredecimguttatus, and this required a combination of two high-pressure ion-exchange chromatographies with hydrophobic chromatography
(Krasnoperov et al., 1990a; Krasnoperov et al., 1990b). All latrotoxins studied are large acidic proteins (pI , 5.0 – 6.0) with molecular masses ranging between 110 and 130 kDa. Most of them are targeted against insects, the spider’s natural prey, and are called latroinsectotoxins (LITs) (a, b, g, d, and e). LIT have LD50 for insects in the range of 15 – 1000 mg/kg of body weight (determined using wax moth larvae, Galleria mellonella) and do not affect vertebrates or crustaceans even at doses up to 5 mg/kg of body weight (Krasnoperov et al., 1990b). One toxic protein, a-latrocrustatoxin (aLCT), is active only in crustaceans, with LD50 for crayfish (Procambarus cubensis) being 100 mg/kg of body weight and doses up to 5 mg/ kg being harmless for insects and vertebrates (Krasnoperov et al., 1990a). Finally, by far the most studied toxin is aLTX, the only known venom component aimed specifically at vertebrates (Grasso, 1976). Its LD50 in mice is 20 – 40 mg/kg of body weight (Ushkarev and Grishin, 1986), and it is ineffective in insects and crustaceans (Magazanik et al., 1992; Ushkaryov, unpublished observations). It is interesting to note that the venoms from the Central Asian L. mactans and L. lugubris produce very similar chromatographic profiles, which resemble those of the European L. mactans tredecimguttatus (Ushkaryov, unpublished observations). In contrast, the composition of the L. hesperus venom is different, although aLTX purified from it is indistinguishable biochemically from the L. mactans toxin. Electrophysiological experiments confirm the high specificity of different latrotoxins to their target animals: at low nanomolar concentrations, aLIT causes a hundredfold increase in the frequency of miniature end-plate potentials (mepps) in blowfly (Calliphora vicina) larvae neuromuscular preparations but at concentrations up to 50 nM does not affect mepps frequency of frog neuromuscular junctions (NMJs) (Kovalevskaia et al., 1990; Magazanik et al., 1992). Reciprocally, aLTX is very efficacious in frog NMJs, but even at 50 nM does not display any activity in blowflies (Magazanik et al., 1992). There have been some reports of aLTX-evoked transmitter release in Drosophila melanogaster larvae (Umbach et al., 1998). These results could be explained by some, albeit low, affinity of aLTX to Drosophila neurones, by an extremely high sensitivity of the larval NMJ to even low amounts of bound aLTX or by an incomplete removal of some insectotoxins. In fact, a low-molecular mass protein— latrodectin—persistently co-purifies with aLTX (Kiyatkin et al., 1992; Pescatori et al., 1995), although it seems to have no specific effect (Volkova et al., 1995) and is not required for aLTX action (Volynski et al., 1999). In addition, even the comprehensive purification methods fail to completely separate aLTX from dLIT. Only recently was aLTX highly purified from the venom of L. lugubris, using successive high-pressure gel-filtration, ion-exchange chromatography and preparative native electrophoresis (Ashton et al., 2000).
Y.A. Ushkaryov et al. / Toxicon 43 (2004) 527–542
As a result, the toxin became less active in Drosophila (J.A. Umbach, C. Gundersen, personal communication) without any loss of toxicity for vertebrates. Despite being selective for distinct animals, all latrotoxins target only neuronal cells (with the exception of dLIT, which was shown by Dulubova et al. (1996) to also affect insect muscles). Both the animal-selectivity and tissuespecificity stem from the fact that, in order to act, these toxins must bind to cell-surface receptors, which are neurone-specific but distinct in different classes of animals. Indeed, radioactively labelled aLTX, aLIT and aLCT have strong affinities for neuronal membranes from vertebrates, insects, and crustaceans, respectively, (Meldolesi, 1982; Knipper et al., 1986; Magazanik et al., 1992; Krasnoperov et al., 1991). At its biologically active concentrations, aLTX does not bind to non-neuronal mammalian tissues (Meldolesi, 1982; Ushkarev and Grishin, 1986) or cultured cells (Davletov et al., 1998; Volynski et al., 2000). Moreover, aLTX and aLIT do not compete with each other for their respective binding sites in the membranes from rat brain and insect heads (Magazanik et al., 1992). The venom of another important theridiid, Steatoda spp., also causes spontaneous neurotransmitter secretion in mouse NMJs (Gillingwater et al., 1999) and ‘excessive neurotransmitter release’ in vitro (South et al., 1998), although its effect is subtly distinct from that of aLTX (Gillingwater et al., 1999).
3. The LTX family portrait 3.1. Primary structures To date, four latrotoxins have been cloned, using cDNA or intron-less genomic DNA from L. mactans tredecimguttatus, and sequenced: aLTX, aLIT, dLIT, and aLCT (Kiyatkin et al., 1990, 1993; Dulubova et al., 1996;
529
Volynskii et al., 1999; Danilevich and Grishin, 2000). Surprisingly, the molecular masses of all these toxins deduced from their DNA sequences are substantially higher than those determined by SDS electrophoresis (Grasso, 1976; Krasnoperov et al., 1990a,b) and mass spectrometry (Dulubova et al., 1996) (Table 1). This indicates that latrotoxins are synthesised as larger precursor proteins that undergo post-translational processing. The full-size toxins are, in fact, inactive (Dulubova et al., 1996). The N-termini of the four toxins have been experimentally determined (Volkova et al., 1991; Bulgakov et al., 1992; Dulubova et al., 1996; Volynskii et al., 1999), and unexpectedly, they appear to be preceded not by defined signal peptides but by relatively short hydrophilic sequences ending with a cluster of basic amino acids (Table 1). These clusters resemble the recognition site of furin, a subtilisin-like proteolytic enzyme involved in the processing of many protein precursors. Furin has been shown to cleave and activate, among other proteins, such toxins as Shiga toxin (Garred et al., 1995), Clostridium septicum a-toxin (Gordon et al., 1997), and diphtheria toxin (Tsuneoka et al., 1993). Therefore, it has been proposed (Volynski et al., 1999) that, during maturation, a furin-like protease hydrolyses LTX precursors, thus, activating the toxins. The N-terminal cleavage of LTX precursors cannot, however, explain the size of mature proteins, suggesting that another proteolytic site must be present in the C-terminal part of the protoxins. Other potential sites for cleavage by furin-like proteases have indeed been found in the C-termini of the precursors (Table 1). Proteolysis at both the N- and C-terminal sites would produce proteins with molecular masses identical to those of mature latrotoxins (Fig. 1A), and expression of the so truncated proteins has produced toxins identical to their natural prototypes both by SDS electrophoresis and specific activity (Dulubova et al., 1996; Ichtchenko et al., 1998; Volynski et al., 1999). Analysis of the amino acid sequences of latrotoxins using protein sorting algorithms (Nakai and Horton, 1999),
Table 1 Molecular masses and proteolytic processing sites of latrotoxins a-LTX Molecular massesa Precursor 156.9 Mature toxin 131.5 Cleavage sitesb N-terminal RMRRlE1 GED C-terminal KFRR1179 lEYKS a
a-LCT
a-LIT
d-LIT
156.1 ,130
157.8 ,130
132.6 110.9
RVISKRlE1 MSK RRFFR1173 lNESP
SAISKRlE1 MSR RFLR1167 lSGHS
RLKRlD1 EED MFRR987 lTLPE
Furin consensusc 6 5 4 3 X x X x
2 X
1 R l
Molecular masses are given in kDa. Amino acid residues adjacent to the cleavage site (vertical bar) are shown in the single-letter code; numbers below mark residue positions in the mature proteins. All N-termini have been determined experimentally. C-terminal cleavage positions are proposed based on the molecular masses of mature proteins and the presence of sequences homologous to the site of cleavage by furin. c The furin consensus site contains an essential Arg at position 1 and positively charged residues at positions 2, 4 or 6 (Nakayama, 1997). b
530
Y.A. Ushkaryov et al. / Toxicon 43 (2004) 527–542
Fig. 1. Analysis of the primary structures of latrotoxins. (A) All known latrotoxins are synthesised as precursor proteins that are cleaved posttranslationary at two furin-like sites located at the N- and C-termini. The ruler above shows amino acid positions, the cleavage sites are indicated by small arrows, and the cleaved fragments are represented in black. (B) Schematic diagrams of aligned aLTX, aLIT, dLIT, and aLCT. The ruler above shows amino acid positions; the borderlines between the wing, body and head domains are indicated by vertical bars (see also Fig. 2). Ankyrin repeats are shown as boxes (numbered at the top) of different shade of grey, which indicates the degree of homology to the ankyrin repeat consensus (E-values: white, 1–10210; light grey, 1–3; dark grey, 3–30); narrow dark grey boxes denote sequences that are not recognised by the software but are demarcated by key residues. (C) Similarity (top) and complexity (bottom) plots of the four aligned latrotoxins. Similarity was calculated as a moving-average (using a 5-residue window) homology of each residue to the consensus sequence (identity attracted the score of 1; similarity, 0.5; dissimilarity, 0). Complexity was calculated as the mean pairwise residue substitution score for each amino acid position of the alignment; the scores were taken from the PAM250 substitution matrix used to optimise the alignment. (D) Phylogenetic tree of four compared latrotoxins (Myr, million years).
predicts that they have no signal peptides or transmembrane domains and must, therefore, be cytosolic (Table 2). Indeed, aLTX is synthesised by free ribosomes (Cavalieri et al., 1990). It may also associate with secretory granules
(possibly via the short hydrophobic stretch near the C-terminus of the precursor) but does not appear in their lumens. The cytoplasmic latrotoxins are probably released into the lumen of the venom gland when secretory
Y.A. Ushkaryov et al. / Toxicon 43 (2004) 527–542 Table 2 Probabilities of specific subcellular localisations of latrotoxins Subcellular locus
a-LTX
a-LCT a-LIT
d-LIT
Cytoplasmic Cytoskeletal Endoplasmic reticulum Golgi complex Mitochondrial Nuclear Peroxisomal Plasma membrane Secretory vesicles Vacuolar
30.4a –b 4.3
13.0 4.3 –
43.5 4.3 –
43.5 – 8.7
8.7 13.0 13.0 – 8.7
– 17.4 60.9 – –
– 4.3 39.1 – 4.3
– 21.7 17.4 4.3 –
17.4
4.3
4.3
4.3
Prediction
cytoplasmic nuclear cytoplasmic cytoplasmic
4.3
–
–
–
a
The probabilities (in %) are calculated for precursor toxins using the k-nearest neighbour algorithm for assessing the probability of protein localisation at each subcellular locus (PSORT II) (Horton and Nakai, 1997; Nakai and Horton, 1999). b Dashes denote the absence of respective predicted localisation signals in protein structures.
epithelial cells periodically burst and expel their contents (Smith and Russell, 1966). The proteolytic cleavage must then occur extracellularly, in the venom gland, which contains different proteases (Huidobro-Toro et al., 1982). This type of secretion is corroborated by the fact that all latrotoxins contain ankyrin repeats (see below) that are found only in intracellular proteins (Sedgwick and Smerdon, 1999), suggesting the evolutionary descent of latrotoxins from an ankyrin-like cytosolic protein. The most characteristic feature of all mature latrotoxins is the presence of multiple repeats (from 13 in dLIT to 22 in aLTX) that take up the C-terminal two-thirds of the molecule (Fig. 1B). These ,33 amino acid-long repeats are similar to those found in a cytoskeletal protein ankyrin and in hundreds of other unrelated proteins with diverse functions, where such repeats are thought to mediate intraand inter-molecular interactions and oligomerisation (Sedgwick and Smerdon, 1999; Andrade et al., 2001). Interestingly, not all ankyrin repeats in latrotoxins are perfectly conserved: both the first one in each protein and a few repeats near the C-termini of the longer toxins (aLTX, aLIT, and aLCT) have diverged so much that they cannot be detected by motif-recognising software (Fig. 1B). It is possible that these imperfect repeats have evolved to break up the rigid spring-like structure formed by the more conventional ankyrin repeats, thus separating and linking the three defined domains of the molecule: the wing, the body and the head (Figs. 1B and 2). The latter two domains are constructed mostly of ankyrin repeats, whilst
531
the N-terminal domain, containing a large proportion of a-helical structure, shows little homology to other proteins. The primary structures of latrotoxins show a moderate degree of amino acid homology, with only , 17% sequence identity in four aligned proteins. However, the movingaverage similarity plot displays a very robust overall conservation of the four sequences, with alternating short regions of higher and lower similarity (Fig. 1C, top). This propensity for mostly homologous substitutions is rather evenly distributed along the sequence, indicating a high probability of similar protein folding and functional conservation, although peaks of even higher similarity (up to 80%) coincide with the beginning of each ankyrin repeat. In contrast, pairwise comparison of these toxins demonstrates the average homology of only 34%, which is distributed unevenly, being higher in the beginning of each ankyrin repeat but rather low in the loops of repeats (Fig. 1C, bottom). Such loops usually bind to other proteins (Rubtsov and Lopina, 2000); in aLTX, they provide contacts between the body and head domains (Ushkaryov, unpublished observation) and, therefore, may be internally sequence-specific and dissimilar between latrotoxins. The lowest degree of conservation between any two sequences is found in the middle of the N-terminal domain, in the area of imperfect repeats and in the head domains. Despite the big difference in length, aLTX and dLIT are closer to each other than they are to the pair of aLIT and aLCT, which are 68% homologous (Fig. 1D). 3.2. Oligomerisation and three-dimensional structure aLTX and dLIT consist of single polypeptide chains with molecular masses of 130 and 111 kDa, respectively, and they have been thought to exist as monomers. However, under normal conditions, both toxins behave as larger proteins with apparent masses of 260 and 230 kDa, respectively (Ashton et al., 2000). Thus, normally both aLTX and dLIT exist as dimers, which are more stable in aLTX probably due to the presence of additional C-terminal ankyrin repeats known to mediate protein– protein interactions. Furthermore, both toxins can also tetramerise. The tetramerisation of aLTX requires the presence of divalent cations or amphipathic molecules (low [SDS] or lipids), whilst dLIT strongly tetramerises in the presence of SDS but does not depend on divalent cations. Divalent cations are also required for aLTX activity; and tetramers, which appear to be the predominant aLTX oligomer in the presence of Ca2þ or Mg2þ, most likely represent the active form of this toxin (Krasilnikov and Sabirov, 1992; Ashton et al., 2000). Other toxins have not been specifically studied in this respect, but they are also very likely to act in their oligomeric form (Shatursky et al., 1995). The tertiary organisation of aLTX has been studied by electron microscopy in vitrified ice (cryo-EM) and image analysis of individual molecular particles (Orlova et al., 2000). So far, the three-dimensional (3D) structure has been
532
Y.A. Ushkaryov et al. / Toxicon 43 (2004) 527–542
Fig. 2. Stereo views of aLTX monomer: tilted view from above (top) and side view (bottom).
determined only for aLTX; however, considering the strong functional sequence conservation (Fig. 1C), the 3D structure of the other latrotoxins is likely to be similar. The LTX monomer (Fig. 2) comprises three main domains: wing, body and head. The N-terminal wing connects to the vertical part of the massive L-shaped body. The latter is joined to the C-terminal head via a short thin neck. The three domains of aLTX are divided by narrow constrictions and form a highly bent molecule. The first 3D-reconstruction of aLTX tetramers (Orlova ˚ et al., 2000) has also been obtained by cryo-EM at 14 A resolution (Fig. 3). The tetrameric complex consists of four monomers symmetrically arranged around a central axis, ˚ resembling a four-blade propeller with a diameter of 250 A ˚ . The compact central mass of the and a thickness of 100 A tetramer is formed by the head domains, which are brought together and surrounded by the body domains. The wings protrude from the bodies perpendicular to the axis of the tetramer; they are connected to the central mass via narrow junctions with well-defined spatial orientations, indicating a rigid structure at the junction point. The base of the ˚ deep. The hydrophobic ‘propeller’ below the wings is 45 A base of the tetramer is thought to mediate aLTX attachment to the membranes.
4. Pore formation aLTX has long been known to form Ca2þ-permeable pores in the plasma membrane of cells sensitive to it (Grasso et al., 1980). However, only with the determination of its 3D structure has it become possible to explain how the hydrophilic aLTX is able to insert itself into the membrane (Orlova et al., 2000). Furthermore, the understanding of the mechanism of pore formation has helped to identify its specific effects on secretion and to study the receptormediated aLTX actions in isolation. The tetramer formed by aLTX is amphipathic, being able to adhere to hydrophobic surfaces with its bottom (Orlova et al., 2000). Toxin tetramers inserted into the membrane of liposomes have been directly visualised by cryo-EM (Orlova et al., 2000). The most relevant feature of the tetramer is the central channel surrounded by toxin monomers; at the lower end, this pear-shaped ˚ in diameter, then it widens to 36 A ˚ to be channel is 25 A ˚ at the top. Thus, the aLTX tetramer constricted to 10 A inserted into the membrane (Fig. 4) would permeabilise it to all substances that can pass through the central channel (small molecules and hydrated cations). Indeed, the pores formed by all latrotoxins studied are permeable for
Y.A. Ushkaryov et al. / Toxicon 43 (2004) 527–542
533
Fig. 3. Stereo views of aLTX tetramer: tilted view from above (top) and view from the side that attaches to cell surface (bottom).
divalent, and certain monovalent, cations (Robello et al., 1987; Wanke et al., 1986; Hurlbut et al., 1994) but not trivalent cations, such as lanthanides (Hurlbut et al., 1994; Chanturiya and Nikoloshina, 1994; Ashton et al., 2001). In addition, aLTX pores allow the exchange of neurotransmitters (McMahon et al., 1990; Davletov et al., 1998; Ashton et al., 2000), fluorescent dyes (Davletov et al., 1998; Volynski et al., 2000) and ATP (McMahon et al., 1990).
Association of aLTX with its receptors greatly helps the toxin to partition into the membrane phase (Volynski et al., 2000). However, in the absence of cell-surface receptors, the high-purity aLTX forms pores very inefficiently: it only inserts into the membrane at concentrations exceeding its normal active concentration by two orders of magnitude (Volynski et al., 2000). Although aLTX was reported to induce pores in protein-free lipid bilayers (Finkelstein et al., 1976), the toxin used in that study was not highly purified
Fig. 4. A model of the aLTX pore in the membrane bilayer. The tetramer is cut open to show the channel. The base of the tetramer fully penetrates the membrane, whilst the wings are attached to the outer membrane surface. Cations can enter the cytosol through the channel, as shown by the arrow.
534
Y.A. Ushkaryov et al. / Toxicon 43 (2004) 527–542
and probably contained proteins from the venom (e.g. latrodectin) that could assist its membrane insertion, as was suggested recently (Volynski et al., 2000). In contrast, when exogenous toxin receptors were expressed in cells (COS-7, HEK293 or BHK) that are normally insensitive to aLTX, it easily induced pores (Hlubek et al., 2000; Van Renterghem et al., 2000; Volynski et al., 2000). Furthermore, signallingdeficient (truncated) mutants of the receptors also facilitated pore formation that correlated with toxin binding irrespective of the receptor structure. Thus, pore formation does not require any signal transduction. The toxin inserted into the membrane always remains on the cell surface without dissociating from cells or penetrating completely into the cytosol (Volynski et al., 2000). Interestingly, when aLTX pore formation was induced by the expression of neurexin, a receptor that binds the toxin only in the presence of Ca2þ, the removal of this cation allowed aLTX to dissociate from the receptor but remain in the cell membrane and continue to form pores (Volynski et al., 2000). The conclusion from this study was that the toxin itself can form pores in the membrane, and neither the receptors nor other neuronal proteins physically contribute to the pore structure. Thus, any toxin binding proteins expressed on cell surface dramatically facilitate the toxin insertion into lipid membranes probably by increasing its near-membrane concentration and/or by properly orientating the toxin tetramer (Volynski et al., 2000; Hlubek et al., 2000; Van Renterghem et al., 2000). Studies with other latrotoxins suggest that they too must form pores in membranes. It has been suggested that an oligomeric form of aLIT creates channels in artificial lipid membranes (Shatursky et al., 1995). Likewise, dLIT, which also assembles into tetramers (Ashton et al., 2000), can form channels in artificial lipid bilayers and in locust muscle membranes (Dulubova et al., 1996). Steatoda paykulliana venom also forms Ca2þ-permeable channels in lipid bilayers (Mironov et al., 1986), although it is currently unclear which component is responsible for pore formation: a 5 kDa peptide (Gillingwater et al., 1999) or another component that is immunologically cross-reactive with Latrodectus venom proteins (South et al., 1998; Graudins et al., 2002). Interestingly, a recombinant mutant aLTX that cannot form membrane pores (aLTXN4C) was described recently (Ichtchenko et al., 1998; Ashton et al., 2001; Volynski et al., 2003). A four-amino acid insert (Val – Pro – Arg – Gly) representing a cleavage site for thrombin was introduced into aLTX between the presumed N-terminal domain and the ankyrin repeats (Fig. 1B, arrowhead) (Ichtchenko et al., 1998). The borderline between the domains was defined from the toxin sequence; however, cryo-EM has demonstrated (Fig. 2) (Orlova et al., 2000) that the domain structure of aLTX is different, and that the mutation landed inside a tightly packed ‘body’ domain. As a result, the 3D structure of the toxin has apparently changed, and although aLTXN4C is dimeric, it can no longer form cyclical tetramers (Volynski et al., 2003). Therefore, despite the fact that this mutant has the same
affinity for the receptors as the wild type aLTX (Ichtchenko et al., 1998; Volynski et al., 2003; Capogna et al., 2003), it fails to incorporate into the membrane and form pores as the wild type toxin does (Volynski et al., 2003). aLTXN4C was originally thought to be altogether inactive (Ichtchenko et al., 1998; Khvotchev and Sudhof, 2000) because it lacked the major-ionophore-activity of aLTX. However, it was later discovered to strongly stimulate transmitter release (Fig. 5) due to its ability to activate receptors (Ashton et al., 2001; Capogna et al., 2003; Volynski et al., 2003). Thus, aLTXN4C provides an ideal tool for investigating exocytotic signals transduced by LTX receptors (see below). Which domain of the toxin molecule participates in the interaction with receptors? It has been proposed that the receptor-binding sites are localised in the N-terminal domain of aLTX (Ichtchenko et al., 1998), and the toxin’s 3D structure (Orlova et al., 2000) is fully consistent with the N-terminal wing being capable of binding to cell-surface molecules (see Fig. 4). To test this hypothesis we have used ‘domain swapping’ to make chimerical LTX mutants. One of these constructs (LI) contained the wing domain from aLTX fused with the combined body/head domains from aLIT, whilst the other (IL) had the aLIT wing and the aLTX body/ head. Pre-incubation of bovine brain membranes with a 100fold excess of LI completely abolished specific binding of radioactively labelled aLTX; in contrast, the IL did not affect such binding (K.E. Volynski, T.M. Volkova, V.G. Lelianova, E.D. Nosyreva, K.A. Pluzhnikov, E.V. Grishin, and Y.A. Ushkaryov, unpublished observation). This finding implicates the wing domains of latrotoxins in the binding to their receptors.
5. aLTX receptors Latrotoxins trigger exocytosis only after binding to neuronal receptors, and the identification of such receptors is likely to reveal the important mechanisms of presynaptic regulation of neurotransmitter release. At present, only aLTX receptors have been identified and characterised. Surprisingly, several structurally and functionally unrelated cell-surface receptors for aLTX have been found. The first receptor to be discovered was neurexin Ia (NRX) (Ushkaryov et al., 1992), a neuronal protein with a single transmembrane domain. Multiple forms of NRXs have been identified; however, all of them bind aLTX only in the presence of Ca2þ (Davletov et al., 1995; Sugita et al., 1999). Because aLTX was also known to act in the absence of this cation (see below), this clearly indicated that another, Ca2þ-independent toxin receptor must also exist. Such a protein was later discovered and termed latrophilin (LPH) (Davletov et al., 1996), or CIRL (Ca2þ-independent receptor for LTX) (Krasnoperov et al., 1996). LPH is a heptahelical transmembrane protein that belongs to
Y.A. Ushkaryov et al. / Toxicon 43 (2004) 527–542
535
Fig. 5. aLTXN4C increases mEPSC frequency and enhances the amplitude of evoked EPSCs in a reversible manner. (A) Continuous whole-cell recordings of mEPSCs from a CA3 pyramidal cell in a slice culture in the presence of 1 mM TTX and 30 mM bicuculline. Focal application of 1 nM aLTXN4C reversibly increased the frequency of mEPSCs. (B) Single traces of EPSCs evoked in a CA3 neuron by pairs of stimuli in the dentate gyrus (50 ms intervals between stimuli) before and after aLTXN4C application and after washout. (Adapted from Capogna et al., 2003).
the secretin/calcitonin family of G protein-coupled receptors (Krasnoperov et al., 1997; Lelianova et al., 1997). Most members of this receptor family bind biologically active peptides and regulate various secretory processes (Jelinek et al., 1993; Lin et al., 1992; Chen et al., 1993). Three isoforms of LPH have been found in mouse, rat, cow and man, the mammalian species so far studied in that respect (Sugita et al., 1998; Ichtchenko et al., 1999; Matsushita et al., 1999; Ushkaryov, unpublished observation). Of these isoforms, LPH 1 (the only homologue that binds aLTX with high affinity) and LPH 3 are highly enriched in mammalian brain, whilst LPH 2 is expressed ubiquitously (Sugita et al., 1998; Ichtchenko et al., 1999; Matsushita et al., 1999). Recently, a third receptor for aLTX, protein tyrosine phosphatase s (PTPs) (Krasnoperov et al., 2002), has been described that binds toxin in a Ca2þ-independent manner. However, PTPs appears to represent a minor receptor component: much less of it is purified from brain extracts on aLTX columns compared to NRX or LPH, and its affinity for the toxin has not been determined (Krasnoperov et al., 2002). Receptor gene knockout was used to identify the functions and the relative roles of the aLTX receptors. Recently, a mouse lacking both LPH 1 and NRX Ia has been produced (Tobaben et al., 2002). The results of this study have confirmed that LPH 1 and NRX Ia represent the major LTX receptors in brain. aLTX exhibited only minor residual activity in the double knockout mice, and this could be due
not only to PTPs but also to LPH 2 and 3, as well as NRX Ib, IIa, IIb, IIIa and IIIb (Tobaben et al., 2002).
6. Modes of action of latrotoxins The vertebrate-specific aLTX has been studied in detail, and its mode of action will be discussed below with reference to the other latrotoxins. Much of the early work on the action of black widow spider venom and its constituent toxins has been extensively discussed in the previous reviews (e.g. Rosenthal and Meldolesi, 1989), and we will concentrate here on the major advances made over the last decade. Although the actions of aLTX are complex, they can be divided into two major pathways: after binding to specific neuronal receptors, aLTX may (i) insert into the plasma membrane, leading to pore formation and possibly other effects and (ii) activate the receptor, leading to intracellular signalling. Both these pathways can stimulate neurotransmitter release, although their mechanisms are different. The mechanisms of aLTX actions are summarised in Fig. 6. 6.1. Ca2þ-dependent and Ca2þ-independent actions of aLTX aLTX has long been known to act both in the presence and in the absence of Ca2þ (Longenecker et al., 1970), with
536
Y.A. Ushkaryov et al. / Toxicon 43 (2004) 527–542
Fig. 6. Multiple mechanisms of aLTX action. 1, Influx of Ca2þ through the pores formed by the toxin after its binding to LPH or NRX. 2, Efflux of neurotransmitters through toxin pores. 3, Ca2þ-dependent receptor-mediated signalling, leading to release of Ca2þ from internal stores. 4, A hypothetical direct interaction of aLTX with the exocytotic machinery (Ca2þ-independent).
the two activities exhibiting different features (Ceccarelli and Hurlbut, 1980; del Castillo and Pumplin, 1975; Fesce et al., 1986). The simplest belief propagated in literature over many years has been that the Ca2þ-dependent effect is due to this cation entering terminals through the toxin pore and triggering secretion, whilst the Ca2þ-independent effect results from receptor-mediated signalling. As we will see later, contrary to the popular view, the Ca2þ-independent pathway seems to be mainly associated with the toxin insertion into the membrane with subsequent pore formation and/or other effects (Khvotchev and Sudhof, 2000; Ashton et al., 2001). On the other hand, the receptorsignalling mechanisms identified so far seem to critically require extracellular Ca2þ (Ca2þ e ) (Ashton et al., 2001; Capogna et al., 2003; Volynski et al., 2003). Similar to aLTX, aLIT increases the frequency of mepps in insect larvae NMJs in both the presence and absence of Ca2þ (Magazanik et al., 1992). dLTX also seemed not to require Ca2þ or even any divalent cations to increase the frequency of miniature excitatory postsynaptic potentials in locust neuromuscular preparations, although in Ca 2þ-free media the toxin was used at very high concentrations (up to 1 mM) (Dulubova et al., 1996). 6.2. Effects mediated by aLTX pores As described above, the pores formed by latrotoxins in the plasma membranes of receptor-expressing cells are permeable to Ca2þ. Influx of extracellular Ca2þ into the cytosol of excitable cells can stimulate exocytosis directly
and efficiently. This cation influx is proportional to the number of bound toxin molecules that, in turn, is dependent on the number of cell surface receptors. Indeed, when PC12 or chromaffin cells, possessing endogenous receptors and normally sensitive to nanomolar aLTX concentrations, were transfected with recombinant exogenous receptors, they acquired a 10-fold higher sensitivity to the toxin (Krasnoperov et al., 1997; Sugita et al., 1999). Importantly, this effect did not depend on the receptor type. Moreover, even signalling-deficient receptor mutants supported this hypersensitisation of cells to the toxin, according to their ability to bind aLTX and facilitate its pore formation (Krasnoperov et al., 1999; Sugita et al., 1998; Hlubek et al., 2000; Volynski et al., 2000). These observations indicate that under normal conditions, i.e. when Ca2þ is present in the medium, the major effect of aLTX is based on the pore-mediated Ca2þ influx. Furthermore, Ca2þ strongly facilitates tetramerisation of aLTX and, consequently, its pore formation (Ashton et al., 2000). Some physiological effects of aLTX are inhibited by trivalent cations (La3þ, Gd3þ, Al3þ) (Scheer, 1989; Rosenthal et al., 1990). Although La3þ was known to block the toxin pores (Hurlbut et al., 1994), its inhibitory effect on the actions of the toxin was previously not understood (Ichtchenko et al., 1998). Using cryo-EM and biochemical assays, the molecular mechanism of La3þ interaction with aLTX has now been explained (Ashton et al., 2001). First, La3þ acts by causing the dissociation of aLTX tetramers and thus preventing the toxin from making
Y.A. Ushkaryov et al. / Toxicon 43 (2004) 527–542
pores. Moreover, even after some toxin pores have formed, La3þ, added to the bath, perturbs their structure and instantaneously blocks ion conduction through the central channel; this blockade is fully reversible (Ashton et al., 2001). An important conclusion from these experiments is that trivalent cations must block all components of the toxin action that depend on pore formation. It is necessary to remember that the large toxin pores (Fig. 5) can also pass small compounds that are normally present in the cytoplasm, including neurotransmitters. Careful studies with brain synaptosomes using radiolabelled or endogenous neurotransmitters, such as norepinephrine (NE), acetylcholine, GABA or glutamate (Glu), have revealed that the toxin pores cause massive leakage of the transmitter pool that is present in the cytosol (McMahon et al., 1990; Adam-Vizi et al., 1993; Davletov et al., 1998; Ashton et al., 2000, 2001). In order to identify and dissect out this type of apparent release, the following criteria may be used. As opposed to vesicular exocytosis, any leakage should be insensitive to the cleavage of the SNARE proteins by clostridial neurotoxins (Davletov et al., 1998; Ashton et al., 2000; Ashton et al., 2001). Depletion of synaptic vesicles by bafilomycin A1 should not affect this nonvesicular process (Ashton et al., 2001). The leakage also strongly correlates with aLTX tetramerisation and pore formation (Davletov et al., 1998; Rahman et al., 1999; Ashton et al., 2000; Orlova et al., 2000); it is abolished by La3þ, which blocks the toxin pores. Finally, this release cannot be induced by the mutant aLTXN4C, which does not form pores. Significantly, this transmitter leakage does not depend on the presence of true Ca2þ and is difficult to e distinguish biochemically from the truly Ca2þ-independent exocytosis, as it may obscure the latter (Davletov et al., 1998; Ashton et al., 2001). 6.3. Receptor-mediated effects It is very difficult to study any receptor signalling in the presence of the toxin pores, whose effect is usually much stronger. To avoid this complication and detect only the receptor-mediated aLTX actions, two approaches have been used. The first one was based on measuring the release in the presence of La3þ, which efficiently blocks toxin pores, as described above. However, this method is not applicable to all systems, as trivalent cations can interfere with normal physiology of some cells. A much better approach has been provided by the use of aLTXN4C, the mutant toxin, which does not form pores (see earlier) but binds to the receptors (Ichtchenko et al., 1998; Ashton et al., 2001; Capogna et al., 2003; Volynski et al., 2003). In contrast to the wild type aLTX, this mutant toxin can be washed from the cell membranes, and its action is fully reversible upon washout (Capogna et al., 2003; Volynski et al., 2003). Consistent with the mutant acting independently of pore formation, La3þ does not block its effect (Capogna et al., 2003; Volynski et al., 2003). Therefore, it has been concluded that
537
aLTXN4C stimulates the frequency of both spontaneous and miniature evoked excitatory postsynaptic currents (mEPSCs) (Fig. 5) only by activating receptors on the cell surface and without incorporating itself into the neuronal membrane. In contrast, the pore-forming aLTX actually inhibits evoked synaptic transmission (Ceccarelli and Hurlbut, 1980; Capogna et al., 1996b). The effect of mutant toxin has been studied in several secretory systems. aLTXN4C has been shown to efficiently release radiolabelled neurotransmitters from rat brain synaptosomes and endogenous catecholamines from bovine chromaffin cells; it causes a massive increase in the frequency of mEPSCs in rat hippocampal slice cultures and the frequency of mepps in mouse neuromuscular preparations. Curiously, in all these systems, the mutant toxin is active only in the presence of Ca2þ (Capogna et al., 2003; Volynski et al., 2003), logically leading to the suggestion that this toxin acts via NRX, a receptor that requires Ca2þ for binding. However, when the toxin interaction with NRX was abolished by replacing Ca2þ in the buffers with Sr2þ, the effect of aLTXN4C was not affected (Volynski et al., 2003), indicating that none of the multiple NRX forms is crucial for the receptor-mediated toxin action and implicating LPH as the main signalling aLTX receptor. A detailed investigation into the nature of the aLTX signalling pathway was conducted biochemically in synaptosomes and electrophysiologically in CA3 pyramidal cells (Davletov et al., 1998; Ashton et al., 2001; Capogna et al., 2003). The findings from these experiments suggest the following mechanism of the receptor-mediated action of aLTX. The toxin stimulates a presynaptic receptor, most likely LPH, which is a G protein-coupled receptor linked to Gaq/11 (Rahman et al., 1999). The downstream effector of Gaq/11 is phospholipase C (PLC). Indeed, aLTXN4C was demonstrated to activate PLC and stimulate Ca2þ e -dependent hydrolysis of phosphoinositides (Ichtchenko et al., 1998). Moreover, the inhibition of PLC by U73122 greatly attenuated the receptor-dependent aLTX action. Activated PLC increases the cytosolic concentration of IP3, which in turn induces release of Ca2þ from intracellular stores. This rise of cytosolic Ca2þ (similar to presynaptic residual Ca2þ) may increase the probability of release and, consequently, the rate of spontaneous exocytosis and the amplitude of evoked release (Capogna et al., 2003) (Fig. 5). In fact, mobilisation of Ca2þ appears to play a pivotal role in aLTX receptor signalling because the effect of aLTXN4C is abolished by both an inhibitor of IP3-induced calcium release, 2-aminoethoxydiphenyl borate, and by depletion of stores with thapsigargin (Ashton et al., 2001; Capogna et al., 2003). Chelation of cytosolic Ca2þ by membrane permeable BAPTA-AM also blocks the aLTXN4C action. Moreover, LTXN4C was directly shown to induce a rise in the presynaptic [Ca2þ] (Capogna et al., 2003). Thapsigargin blocks the aLTXN4C effect on both secretion and the rise of
538
Y.A. Ushkaryov et al. / Toxicon 43 (2004) 527–542
calcium, indicating that the latter must be the underlying mechanism of the receptor-mediated action. The question why the mutant toxin requires extracellular Ca2þ or Sr2þ to induce release of neurotransmitters currently remains unresolved. There could be several possible explanations for this phenomenon. Firstly, Ca2þ or Sr2þ can serve as extracellular co-factors for LPH and its signalling partners (e.g. PLC). It is also possible that these divalent cations participate in some other, receptor-independent signalling mechanism (e.g. cyclical Ca2þ amplification pathway (Masgrau et al., 2000)) that allows the activation of intracellular Ca2þ stores critical for aLTXN4C-induced release. Finally, they may enter the cell via some LPHlinked low-conductance cation channels and induce secretion directly. The latter hypothesis is interesting but seems unlikely because these hypothetical channels are undetectable electrophysiologically or biochemically and must be very unusual, being insensitive to La3þ, Cd2þ, SKF 96365 and other channel-blocking drugs (Capogna et al., 2003). Some authors, using cells expressing exogenous LPH or PTPs mutants incapable of signal transduction, argued that aLTX did not act through any receptor (Sugita et al., 1998; Krasnoperov et al., 2002). However, because all such receptor constructs allow toxin binding and pore formation (Volynski et al., 2000), transmitter release in those experiments was simply due to the pore-mediated Ca2þ influx, which does not require any signalling. The properties of receptor-mediated exocytosis described above for aLTXN4C are almost identical to those found for the action of aLCT on amino acid release from crayfish NMJ (Elrick and Charlton, 1999). These features include: (i) absolute Ca2þ e requirement; (ii) absolute Ca2þ requirement; (iii) occurrence of bursts; (iv) lack of i involvement of known voltage-gated Ca2þ channels; (v) insensitivity to La3þ; (vi) no cessation of activity after prolonged stimulation. These authors explained their data by unusual properties of the aLCT pores. However, in other studies, more consistent with pore-dependent toxin action (Burmistrov et al., 1997), intense transmitter release elicited by aLCT was very similar to that caused by wild type aLTX in vertebrate NMJs and eventually resulted in complete blockade of neuromuscular transmission and nerve terminal damage. Therefore, the effect of aLCT observed by Elrick and Charlton (1999) could possibly be due to their toxin preparation failing to form pores but still acting through a receptor. 6.4. Probable Ca2þ-independent mechanisms In Ca2þ-free buffers, both aLTX and aLIT can cause vesicular release, detected electrophysiologically, in neurones (Longenecker et al., 1970; Ceccarelli and Hurlbut, 1980; Capogna et al., 1996a) but not in endocrine cells (Bittner et al., 1998; Sugita et al., 1998); and even in neuronal cells only small synaptic vesicles, but not large dense-core vesicles, seem to be affected (Matteoli et al.,
1988). The Ca2þ-independent effect of aLTX is especially distinct in motor nerve terminals (Longenecker et al., 1970; Fesce et al., 1986), although central synapses can also be stimulated by this toxin in the nominal absence of Ca2þ (Capogna et al., 1996a). However, the features of Ca2þdependent and -independent secretion are quite distinct. The increase in the frequency of spontaneous release caused by aLTX at the NMJ in the absence of Ca2þ is slow and gradual; after reaching a maximum, the high frequency of minis subsides, and eventually they cease (Ceccarelli and Hurlbut, 1980; Tsang et al., 2000). In contrast, in the presence of Ca2þ, the toxin causes burst-like effects that start abruptly and are interspersed by periods of almost complete inactivity; this release continues much longer and the overall amount of released neurotransmitter is much higher than in Ca2þ-free saline (Fesce et al., 1986). In central neurones, due to experimental difficulties, the Ca2þ-independent actions of aLTX have not been characterised as extensively as at the NMJ, but most authors also find that the addition of toxin in the presence of Ca2þ causes a much bigger effect (Capogna et al., 1996a; Auger and Marty, 1997). In synaptosomes, Ca2þ-independent release can be easily measured (Ichtchenko et al., 1998; Davletov et al., 1998; Ashton et al., 2001), but it mostly represents nonvesicular transmitter efflux (see above) that may mask any vesicular component so far never identified as such biochemically (Davletov et al., 1998; Ashton et al., 2000). What could be the mechanism of the Ca2þ-independent aLTX-stimulated vesicular secretion? Until recently, this release was thought to be mediated by a receptor transduction pathway that theoretically would not require Ca2þ influx. The only receptor participating in such signalling could be LPH (and possibly PTPs), because NRX does not bind the toxin under these conditions. However, the receptor transduction pathway described above actually requires extracellular Ca2þ (Davletov et al., 1998; Ashton et al., 2001; Capogna et al., 2003) and cannot explain the Ca2þindependent aLTX-induced exocytosis. Some insight into the mechanism of this secretion has been obtained by comparing the mutant aLTXN4C (see above) with wild type toxin. The former does not insert into the membrane and does not form pores but only acts by activating receptors; the latter inserts itself into the membrane, forms pores and activates receptors (Capogna et al., 2003; Volynski et al., 2003). Similarly, the effects of these toxins partially overlap: the mutant induces only Ca2þ-dependent vesicular release, whilst the wild type aLTX causes both Ca2þ-dependent and -independent exocytosis. This suggests that vesicular secretion triggered by aLTX in the absence of Ca2þ must be due to its membrane penetration and/or pore formation. This is further substantiated by the requirement of Mg2þ for aLTX-stimulated release in Ca2þ-free conditions (Misler and Falke, 1987), which agrees with the fact that Mg2þ facilitates toxin tetramerisation and pore formation (see earlier). It is unclear whether secretion in the absence of extracellular Ca2þ can be induced by the pore-mediated
Y.A. Ushkaryov et al. / Toxicon 43 (2004) 527–542
influx/efflux of cations or membrane depolarisation. Mg2þ or Naþ could enter through aLTX pores and either trigger secretion by themselves or by stimulating Ca2þ release by mitochondria; however, the latter proved non-essential for synaptic vesicle exocytosis (Tsang et al., 2000). On the other hand, the tetrameric wild type toxin inserted into the neuronal plasma membrane could potentially interact directly with some components of the intracellular transmitter release apparatus, defining a novel hypothetical mechanism proposed recently (Khvotchev and Sudhof, 2000). aLIT can also induce release in the absence of Ca2þ, but Mg2þ is required (Magazanik et al., 1992), indicating that this action of the insectotoxins also depends upon pore formation (Shatursky et al., 1995) due to tetramerisation and membrane insertion.
539
instrumental for the elucidation of the endogenous functions of aLTX receptors.
Acknowledgements We thank all other (present and past) members of this laboratory and our collaborators for making important contributions to the structural and functional studies of latrotoxins. We thank A. Rohou for preparing the stereoimages of aLTX. The work is supported by a Wellcome Senior European Research Fellowship and a BBSRC project grant (to Y. A. U.)
References 7. Latrotoxins-versatile tools to study neuroexocytosis All known latrotoxins have been used to some extent to analyse the mechanisms involved in the regulation of neurotransmitter release, although most of our knowledge comes from the application of aLTX. Thus, among other important findings, this toxin has helped to: confirm the vesicular hypothesis of transmitter release (Ceccarelli and Hurlbut, 1980; Hurlbut et al., 1990), establish the requirement of Ca2þ for endocytosis (Ceccarelli and Hurlbut, 1980), characterise individual transmitter release sites in the central nervous system (Auger and Marty, 1997), and identify two families of important neuronal cell-surface receptors (Ushkaryov et al., 1992; Lelianova et al., 1997; Krasnoperov et al., 1997). However, detailed insights into the intracellular signalling have been hampered by the robust effects of the toxin pore, which have deceived many researchers. The advent of the mutant toxin aLTXN4C that is unable to form pores but acts via receptor stimulation has dramatically raised the research potential of the toxin and allowed one to explain some of the previously not understood actions of wild type aLTX. Thus, it has helped to approach the deciphering of the intracellular signal transduction machinery stimulated by aLTX (Ashton et al., 2001; Capogna et al., 2003; Volynski et al., 2003). By using this artificial toxin, one can study the nature and properties of intracellular Ca2þ stores implicated in the toxin receptor transduction pathway and their effect on evoked postsynaptic potentials (Capogna et al., 2003). Whilst aLTX is able to trigger the exocytosis of both the readily releasable pool and the depot pool of synaptic vesicles, aLTXN4C specifically induces only the readily releasable pool (Ashton et al., 2001). This study has already highlighted the biochemical differences between the two pools, and future experiments may reveal the precise localisation of such vesicles within the nerve terminal and their interaction with the cytoskeletal elements. Finally, the mutant toxin will be
Adam-Vizi, V., Deri, Z., Bors, P., Tretter, L., 1993. Lack of involvement of [Ca2þ]i in the external Ca2þ- independent release of acetylcholine evoked by veratridine, ouabain and a-latrotoxin: possible role of [Naþ]i. J. Physiol. Paris 87, 43–50. Andrade, M.A., Perez-Iratxeta, C., Ponting, C.P., 2001. Protein repeats: structures, functions, and evolution. J. Struct. Biol. 134, 117 –131. Ashton, A.C., Rahman, M.A., Volynski, K.E., Manser, C., Orlova, E.V., Matsushita, H., Davletov, B.A., van Heel, M., Grishin, E.V., Ushkaryov, Y.A., 2000. Tetramerisation of a-latrotoxin by divalent cations is responsible for toxin-induced non-vesicular release and contributes to the Ca2þ-dependent vesicular exocytosis from synaptosomes. Biochimie 82, 453 –468. Ashton, A.C., Volynski, K.E., Lelianova, V.G., Orlova, E.V., Van Renterghem, C., Canepari, M., Seagar, M., Ushkaryov, Y.A., 2001. a-latrotoxin, acting via two Ca2þ-dependent pathways, triggers exocytosis of two pools of synaptic vesicles. J. Biol. Chem. 276, 44695– 44703. Auger, C., Marty, A., 1997. Heterogeneity of functional synaptic parameters among single release sites. Neuron 19, 139–150. Bittner, M.A., Krasnoperov, V.G., Stuenkel, E.L., Petrenko, A.G., Holz, R.W., 1998. A Ca2þ-independent receptor for a-latrotoxin, CIRL, mediates effects on secretion via multiple mechanisms. J. Neurosci. 18, 2914–2922. Bulgakov, O.V., Volkova, T.M., Galkina, T.G., Pashkov, V.N., Pluzhnikov, K.A., Grishin, E.V., 1992. Study of the amino acid sequence of latroinsectotoxin from black widow spider venom. Bioorg. Khim. 18, 871–874. Burmistrov, Y.M., Shuranova, Z.P., Artiukhina, N.I., 1997. Effects of black widow spider venom and latrocrustatoxin on crustacean nerve cells: electrophysiological and ultrastructural study. Gen. Pharmacol. 28, 159–166. Capogna, M., Gahwiler, B.H., Thompson, S.M., 1996a. Calciumindependent actions of a-latrotoxin on spontaneous and evoked synaptic transmission in the hippocampus. J. Neurophysiol. 76, 3149–3158. Capogna, M., Gahwiler, B.H., Thompson, S.M., 1996b. Presynaptic inhibition of calcium-dependent and -independent release elicited with ionomycin, gadolinium, and a-latrotoxin in the hippocampus. J. Neurophysiol. 75, 2017–2028. Capogna, M., Volynski, K.E., Emptage, N.J., Ushkaryov, Y.A., 2003. The a-latrotoxin mutant LTXN4C enhances spontaneous
540
Y.A. Ushkaryov et al. / Toxicon 43 (2004) 527–542
and evoked transmitter release in CA3 pyramidal neurons. J. Neurosci. 23, 4044–4053. Cavalieri, M., D’Urso, D., Lassa, A., Pierdominici, E., Robello, M., Grasso, A., 1987. Characterization and some properties of the venom gland extract of a theridiid spider (Steatoda paykulliana) frequently mistaken for black widow spider (Latrodectus tredecimguttatus). Toxicon 25, 965–974. Cavalieri, M., Corvaja, N., Grasso, A., 1990. Immunocytological localization by monoclonal antibodies of a-latrotoxin in the venom gland of the spider Latrodectus tredecimguttatus. Toxicon 28, 341 –346. Ceccarelli, B., Hurlbut, W.P., 1980. Ca2þ-dependent recycling of synaptic vesicles at the frog neuromuscular junction. J. Cell Biol. 87, 297–303. Chanturiya, A.N., Nikoloshina, H.V., 1994. Correlations between changes in membrane capacitance induced by changes in ionic environment and the conductance of channels incorporated into bilayer lipid membranes. J. Membr. Biol. 137, 71–77. Chen, R., Lewis, K.A., Perrin, M.H., Vale, W.W., 1993. Expression cloning of a human corticotropin-releasing-factor receptor. Proc. Natl Acad. Sci. USA 90, 8967–8971. Danilevich, V.N., Grishin, E.V., 2000. The chromosomal genes for black widow spider neurotoxins do not contain introns. Bioorg. Khim. 26, 933 –939. Davletov, B.A., Krasnoperov, V., Hata, Y., Petrenko, A.G., Sudhof, T.C., 1995. High affinity binding of a-latrotoxin to recombinant neurexin Ia. J. Biol. Chem. 270, 23903–23905. Davletov, B.A., Shamotienko, O.G., Lelianova, V.G., Grishin, E.V., Ushkaryov, Y.A., 1996. Isolation and biochemical characterization of a Ca2þ-independent a-latrotoxin-binding protein. J. Biol. Chem. 271, 23239–23245. Davletov, B.A., Meunier, F.A., Ashton, A.C., Matsushita, H., Hirst, W.D., Lelianova, V.G., Wilkin, G.P., Dolly, J.O., Ushkaryov, Y.A., 1998. Vesicle exocytosis stimulated by a-latrotoxin is mediated by latrophilin and requires both external and stored Ca2þ. EMBO J. 17, 3909–3920. del Castillo, J., Pumplin, D.W., 1975. Discrete and discontinuous action of brown widow spider venom on the presynaptic nerve terminals of frog muscle. J. Physiol. (London) 252, 491–508. Dulubova, I.E., Krasnoperov, V.G., Khvotchev, M.V., Pluzhnikov, K.A., Volkova, T.M., Grishin, E.V., Vais, H., Bell, D.R., Usherwood, P.N., 1996. Cloning and structure of a-latroinsectotoxin, a novel insect-specific member of the latrotoxin family: functional expression requires C-terminal truncation. J. Biol. Chem. 271, 7535–7543. Elrick, D.B., Charlton, M.P., 1999. a-Latrocrustatoxin increases neurotransmitter release by activating a calcium influx pathway at crayfish neuromuscular junction. J. Neurophysiol. 82, 3550–3562. Fesce, R., Segal, J.R., Ceccarelli, B., Hurlbut, W.P., 1986. Effects of black widow spider venom and Ca2þ on quantal secretion at the frog neuromuscular junction. J. Gen. Physiol. 88, 59 –81. Finkelstein, A., Rubin, L.L., Tzeng, M.C., 1976. Black widow spider venom: effect of purified toxin on lipid bilayer membranes. Science 193, 1009–1011. Frontali, N., Granata, F., Parisi, P., 1972. Effects of black widow spider venom on acetylcholine release from rat cerebral cortex slices in vitro. Biochem. Pharmacol. 21, 969–974.
Frontali, N., Ceccarelli, B., Gorio, A., Mauro, A., Siekevitz, P., Tzeng, M.C., Hurlbut, W.P., 1976. Purification from black widow spider venom of a protein factor causing the depletion of synaptic vesicles at neuromuscular junctions. J. Cell Biol. 68, 462 –479. Garred, O., van Deurs, B., Sandvig, K., 1995. Furin-induced cleavage and activation of shiga toxin. J. Biol. Chem. 270, 10817–10821. Gillingwater, T.H., Kalikulov, D., Ushkaryov, Y., Ribchester, R.R., 1999. Comparison of effects of a-latrotoxin with a partially purified toxin from another theridiid spider,Steatoda paykulliana, on exocytosis at mouse neuromuscular junctions. J. Physiol. (London), 520P(see also p.40). Gordon, V.M., Benz, R., Fujii, K., Leppla, S.H., Tweten, R.K., 1997. Clostridium septicum a-toxin is proteolytically activated by furin. Infect. Immun. 65, 4130–4134. Granata, F., Paggi, P., Frontali, N., 1972. Effects of chromatographic fractions of black widow spider venom on in vitro biological systems. Toxicon 10, 551–555. Grasso, A., 1976. Preparation and properties of a neurotoxin purified from the venom of black widow spider (Latrodectus mactans tredecimguttatus). Biochim. Biophys. Acta 439, 406 –412. Grasso, A., Alema, S., Rufini, S., Senni, M.I., 1980. Black widow spider toxin-induced calcium fluxes and transmitter release in a neurosecretory cell line. Nature 283, 774–776. Graudins, A., Gunja, N., Broady, K.W., Nicholson, G.M., 2002. Clinical and in vitro evidence for the efficacy of Australian redback spider (Latrodectus hasselti) antivenom in the treatment of envenomation by a Cupboard spider (Steatoda grossa). Toxicon 40, 767–775. Griffiths, D.J., Smyth, T. Jr, 1973. Action of black widow spider venom at insect neuromuscular junctions. Toxicon 11, 369 –374. Hlubek, M.D., Stuenkel, E.L., Krasnoperov, V.G., Petrenko, A.G., Holz, R.W., 2000. Calcium-independent receptor for alatrotoxin and neurexin Ia facilitate toxin-induced channel formation: evidence that channel formation results from tethering of toxin to membrane. Mol. Pharmacol. 57, 519 –528. Horton, P., Nakai, K., 1997. Better prediction of protein cellular localization sites with the k nearest neighbors classifier. Proc. Int. Conf. Intell. Syst. Mol. Biol. 5, 147 –152. Huidobro-Toro, J.P., Chelala, C.A., Musacchio, J.M., 1982. Hydrolysis of substance P and bradykinin by black widow spider venom gland extract. Biochem. Pharmacol. 31, 3323–3328. Hurlbut, W.P., Iezzi, N., Fesce, R., Ceccarelli, B., 1990. Correlation between quantal secretion and vesicle loss at the frog neuromuscular junction. J. Physiol. (London) 425, 501 –526. Hurlbut, W.P., Chieregatti, E., Valtorta, F., Haimann, C., 1994. a-Latrotoxin channels in neuroblastoma cells. J. Membr. Biol. 138, 91–102. Ichtchenko, K., Khvotchev, M., Kiyatkin, N., Simpson, L., Sugita, S., Sudhof, T.C., 1998. a-Latrotoxin action probed with recombinant toxin: receptors recruit a-latrotoxin but do not transduce an exocytotic signal. EMBO J. 17, 6188– 6199. Ichtchenko, K., Bittner, M.A., Krasnoperov, V., Little, A.R., Chepurny, O., Holz, R.W., Petrenko, A.G., 1999. A novel ubiquitously expressed (-latrotoxin receptor is a member of
Y.A. Ushkaryov et al. / Toxicon 43 (2004) 527–542 the CIRL family of G-protein-coupled receptors. J. Biol. Chem. 274, 5491–5498. Jelinek, L.J., Lok, S., Rosenberg, G.B., Smith, R.A., Grant, F.J., Biggs, S., Bensch, P.A., Kuijper, J.L., Sheppard, P.O., Sprecher, C.A., 1993. Expression cloning and signaling properties of the rat glucagon receptor. Science 259, 1614– 1616. Kawai, N., Mauro, A., Grundfest, H., 1972. Effect of black widow spider venom on the lobster neuromuscular junctions. J. Gen. Physiol. 60, 650– 664. Khvotchev, M., Sudhof, T.C., 2000. a-Latrotoxin triggers transmitter release via direct insertion into the presynaptic plasma membrane. EMBO J. 19, 3250–3262. Kiyatkin, N.I., Dulubova, I.E., Chekhovskaya, I.A., Grishin, E.V., 1990. Cloning and structure of cDNA encoding a-latrotoxin from black widow spider venom. FEBS Lett. 270, 127 –131. Kiyatkin, N., Dulubova, I., Chekhovskaya, I., Lipkin, A., Grishin, E., 1992. Structure of the low molecular weight protein copurified with a-latrotoxin. Toxicon 30, 771 –774. Kiyatkin, N., Dulubova, I., Grishin, E., 1993. Cloning and structural analysis of a-latroinsectotoxin cDNA Abundance of ankyrinlike repeats. Eur. J. Biochem. 213, 121–127. Knipper, M., Madeddu, L., Breer, H., Meldolesi, J., 1986. Black widow spider venom-induced release of neurotransmitters: mammalian synaptosomes are stimulated by a unique venom component (a-latrotoxin), insect synaptosomes by multiple components. Neuroscience 19, 55–62. Kovalevskaia, G.I., Pashkov, V.N., Bulgakov, O.V., Fedorova, I.M., Magazanik, L.G., Grishin, E.V., 1990. Identification and isolation of the protein insect toxin (a-latroinsectotoxin) from venom of the spider Latrodectus mactans tredecimguttatus. Bioorg. Khim. 16, 1013–1018. Krasilnikov, O.V., Sabirov, R.Z., 1992. Comparative analysis of latrotoxin channels of different conductance in planar lipid bilayers. Evidence for cluster organization. Biochim. Biophys. Acta 1112, 124 –128. Krasnoperov, V.G., Shamotienko, O.G., Grishin, E.V., 1990a. A crustacean-specific neurotoxin from the venom of the black widow spider Latrodectus mactans tredecimguttatus. Bioorg. Khim. 16, 1567–1569. Krasnoperov, V.G., Shamotienko, O.G., Grishin, E.V., 1990b. Isolation and properties of insect-specific neurotoxins from venoms of the spider Lactodectus mactans tredecimguttatus. Bioorg. Khim. 16, 1138–1140. Krasnoperov, V.G., Shamotienko, O.G., Grishin, E.V., 1991. Interaction of a-125I-latrocrustotoxin with nerve cell membranes from the river crab Astacus astacus. Bioorg. Khim. 17, 716–718. Krasnoperov, V.G., Beavis, R., Chepurny, O.G., Little, A.R., Plotnikov, A.N., Petrenko, A.G., 1996. The calcium-independent receptor of a-latrotoxin is not a neurexin. Biochem. Biophys. Res. Commun. 227, 868 –875. Krasnoperov, V.G., Bittner, M.A., Beavis, R., Kuang, Y., Salnikow, K.V., Chepurny, O.G., Little, A.R., Plotnikov, A.N., Wu, D., Holz, R.W., Petrenko, A.G., 1997. a-Latrotoxin stimulates exocytosis by the interaction with a neuronal G-protein-coupled receptor. Neuron 18, 925–937. Krasnoperov, V., Bittner, M.A., Holz, R.W., Chepurny, O., Petrenko, A.G., 1999. Structural requirements for a-latrotoxin binding and a-latrotoxin-stimulated secretion. A study with
541
calcium-independent receptor of a-latrotoxin (CIRL) deletion mutants. J. Biol. Chem. 274, 3590–3596. Krasnoperov, V.G., Bittner, M.A., Mo, W., Buryanovsky, L., Neubert, T.A., Holz, R.W., Ichtchenko, K., Petrenko, A.G., 2002. Protein tyrosine phosphatase-a is a novel member of the functional family of a-latrotoxin receptors. J. Biol. Chem. 277, 35887–35895. Lelianova, V.G., Davletov, B.A., Sterling, A., Rahman, M.A., Grishin, E.V., Totty, N.F., Ushkaryov, Y.A., 1997. aLatrotoxin receptor, latrophilin, is a novel member of the secretin family of G protein-coupled receptors. J. Biol. Chem. 272, 21504– 21508. Lin, C., Lin, S.C., Chang, C.P., Rosenfeld, M.G., 1992. Pit-1dependent expression of the receptor for growth hormone releasing factor mediates pituitary cell growth. Nature 360, 765–768. Longenecker, H.E., Hurlbut, W.P., Mauro, A., Clark, A.W., 1970. Effects of black widow spider venom on the frog neuromuscular junction. Effects on end-plate potential, miniature end-plate potential and nerve terminal spike. Nature 225, 701– 703. Magazanik, L.G., Fedorova, I.M., Kovalevskaya, G.I., Pashkov, V.N., Bulgakov, O.V., Grishin, E.V., 1992. Selective presynaptic insectotoxin (a-latroinsectotoxin) isolated from black widow spider venom. Neuroscience 46, 181–188. Masgrau, R., Servitja, J.M., Sarri, E., Young, K.W., Nahorski, S.R., Picatoste, F., 2000. Intracellular Ca2þ stores regulate muscarinic receptor stimulation of phospholipase C in cerebellar granule cells. J. Neurochem. 74, 818–826. Matsushita, H., Lelianova, V.G., Ushkaryov, Y.A., 1999. The latrophilin family: multiply spliced G protein-coupled receptors with differential tissue distribution. FEBS Lett. 443, 348– 352. Matteoli, M., Haimann, C., Torri-Tarelli, F., Polak, J.M., Ceccarelli, B., De Camilli, P., 1988. Differential effect of a-latrotoxin on exocytosis from small synaptic vesicles and from large densecore vesicles containing calcitonin gene-related peptide at the frog neuromuscular junction. Proc. Natl Acad. Sci. USA 85, 7366–7370. McMahon, H.T., Rosenthal, L., Meldolesi, J., Nicholls, D.G., 1990. a-Latrotoxin releases both vesicular and cytoplasmic glutamate from isolated nerve terminals. J. Neurochem. 55, 2039–2047. Meldolesi, J., 1982. Studies on a-latrotoxin receptors in rat brain synaptosomes: correlation between toxin binding and stimulation of transmitter release. J. Neurochem. 38, 1559–1569. Mironov, S.L., Sokolov, Yu.V., Chanturiya, A.N., Lishko, V.K., 1986. Channels produced by spider venoms in bilayer lipid membrane: mechanisms of ion transport and toxic action. Biochim. Biophys. Acta 862, 185 –198. Misler, S., Falke, L.C., 1987. Dependence on multivalent cations of quantal release of transmitter induced by black widow spider venom. Am. J. Physiol. 253, C469–C476. Nakai, K., Horton, P., 1999. PSORT: a program for detecting sorting signals in proteins and predicting their subcellular localization. Trends Biochem. Sci. 24, 34–36. Nakayama, K., 1997. Furin: a mammalian subtilisin/Kex2p-like endoprotease involved in processing of a wide variety of precursor proteins. Biochem. J. 327, 625–635. Orlova, E.V., Rahman, M.A., Gowen, B., Volynski, K.E., Ashton, A.C., Manser, C., van Heel, M., Ushkaryov, Y.A., 2000. Structure of a-latrotoxin oligomers reveals that divalent cationdependent tetramers form membrane pores. Nat. Struct. Biol. 7, 48 –53.
542
Y.A. Ushkaryov et al. / Toxicon 43 (2004) 527–542
Pescatori, M., Bradbury, A., Bouet, F., Gargano, N., Mastrogiacomo, A., Grasso, A., 1995. The cloning of a cDNA encoding a protein (latrodectin) which co-purifies with the (-latrotoxin from the black widow spider Latrodectus tredecimguttatus (Theridiidae). Eur. J. Biochem. 230, 322– 328. Platnick, N.I., 1993. Advances in spider taxonomy 1988–1991: with synonymies and transfers 1940–1980, New York Entomological Society, New York. Platnick, N.I., 1997. Advances in spider taxonomy 1992–1995: with redescriptions 1940 –1980, New York Entomological Society, New York, pp. 1–976. Rahman, M.A., Ashton, A.C., Meunier, F.A., Davletov, B.A., Dolly, J.O., Ushkaryov, Y.A., 1999. Norepinephrine exocytosis stimulated by a-latrotoxin requires both external and stored Ca2þ and is mediated by latrophilin, G proteins and phospholipase C. Philos. Trans. R. Soc. Lond. B 354, 379–386. Robello, M., Fresia, M., Maga, L., Grasso, A., Ciani, S., 1987. Permeation of divalent cations through a-latrotoxin channels in lipid bilayers: steady-state current-voltage relationships. J. Membr. Biol. 95, 55–62. Rosenthal, L., Meldolesi, J., 1989. a-Latrotoxin and related toxins. Pharmacol. Ther. 42, 115– 134. Rosenthal, L., Zacchetti, D., Madeddu, L., Meldolesi, J., 1990. Mode of action of a-latrotoxin: role of divalent cations in Ca2þdependent and Ca2þ-independent effects mediated by the toxin. Mol. Pharmacol. 38, 917–923. Rubtsov, A.M., Lopina, O.D., 2000. Ankyrins. FEBS Lett. 482, 1–5. Scheer, H.W., 1989. Interactions between a-latrotoxin and trivalent cations in rat striatal synaptosomal preparations. J. Neurochem. 52, 1590–1597. Sedgwick, S.G., Smerdon, S.J., 1999. The ankyrin repeat: a diversity of interactions on a common structural framework. Trends Biochem. Sci. 24, 311–316. Shatursky, O.Y., Pashkov, V.N., Bulgacov, O.V., Grishin, E.V., 1995. Interaction of a-latroinsectotoxin from Latrodectus mactans venom with bilayer lipid membranes. Biochim. Biophys. Acta 1233, 14–20. Smith, D.S., Russell, F.E., 1966. Structure of the venom gland of the black widow spider Latrodectus mactans. A preliminary light and electron microscopic study. In: Russel, F.E., Saunders, P.R (Eds.), Animal Toxins, Pergamon, Oxford, pp. 1–15. South, M., Wirth, P., Winkel, K.D., 1998. Redback spider antivenom used to treat envenomation by a juvenile Steatoda spider. Med. J. Aust. 169, 642. Sugita, S., Ichtchenko, K., Khvotchev, M., Su¨dhof, T.C., 1998. a-Latrotoxin receptor CIRL/latrophilin 1 (CL1) defines an unusual family of ubiquitous G-protein-linked receptors. Gprotein coupling not required for triggering exocytosis. J. Biol. Chem. 273, 32715–32724. Sugita, S., Khvochtev, M., Su¨dhof, T.C., 1999. Neurexins are functional a-latrotoxin receptors. Neuron 22, 489 –496. Tobaben, S., Sudhof, T.C., Stahl, B., 2002. Genetic analysis of a-latrotoxin receptors reveals functional interdependence of
CIRL/Latrophilin 1 and neurexin Ia. J. Biol. Chem. 277, 6359–6365. Tsang, C.W., Elrick, D.B., Charlton, M.P., 2000. a-Latrotoxin releases calcium in frog motor nerve terminals. J. Neurosci. 20, 8685–8692. Tsuneoka, M., Nakayama, K., Hatsuzawa, K., Komada, M., Kitamura, N., Mekada, E., 1993. Evidence for involvement of furin in cleavage and activation of diphtheria toxin. J. Biol. Chem. 268, 26461– 26465. Umbach, J.A., Grasso, A., Zurcher, S.D., Kornblum, H.I., Mastrogiacomo, A., Gundersen, C.B., 1998. Electrical and optical monitoring of a-latrotoxin action at Drosophila neuromuscular junctions. Neuroscience 87, 913 –924. Ushkarev, Iu.A., Grishin, E.V, 1986. Neurotoxin of the black widow spider and its interaction with receptors from the rat brain. Bioorg. Khim. 12, 71–80. Ushkaryov, Y.A., Petrenko, A.G., Geppert, M., Sudhof, T.C., 1992. Neurexins: synaptic cell surface proteins related to the a-latrotoxin receptor and laminin. Science 257, 50 –56. Van Renterghem, C., Iborra, C., Martin-Moutot, N., Lelianova, V., Ushkaryov, Y., Seagar, M., 2000. a-Latrotoxin forms calciumpermeable membrane pores via interactions with latrophilin or neurexin. Eur. J. Neurosci. 12, 3953–3962. Volkova, T.M., Galkina, T.G., Kudelin, A.B., Grishin, E.V., 1991. Structure of tryptic fragments of a neurotoxin from black widow spider venom. Bioorg. Khim. 17, 437–441. Volkova, T.M., Pluzhnikov, K.A., Woll, P.G., Grishin, E.V., 1995. Low molecular weight components from black widow spider venom. Toxicon 33, 483 –489. Volynski, K.E., Nosyreva, E.D., Ushkaryov, Y.A., Grishin, E.V., 1999. Functional expression of a-latrotoxin in baculovirus system. FEBS Lett. 442, 25–28. Volynskii, K.E., Volkova, T.M., Galkina, T.G., Krasnoperov, V.G., Pluzhnikov, K.A., Khvoshchev, M.V., Grishin, E.V., 1999. Molecular cloning and primary structure of cDNA fragment for a-latrocrustatoxin from black widow spider venom. Bioorg. Khim. 25, 25–30. Volynski, K.E., Meunier, F.A., Lelianova, V.G., Dudina, E.E., Volkova, T.M., Rahman, M.A., Manser, C., Grishin, E.V., Dolly, J.O., Ashley, R.H., Ushkaryov, Y.A., 2000. Latrophilin, neurexin and their signaling-deficient mutants facilitate a-latrotoxin insertion into membranes but are not involved in pore formation. J. Biol. Chem. 275, 41175–41183. Volynski, K.E., Capogna, M., Ashton, A.C., Thomson, D., Orlova, E.V., Manser, C.F., Ribchester, R.R., Ushkaryov, Y.A., 2003. Mutant a-latrotoxin (LTXN4C) does not form pores and causes secretion by receptor stimulation.This action does not require neurexins. J. Biol. Chem. 278, 31058–31066. Wanke, E., Ferroni, A., Gattanini, P., Meldolesi, J., 1986. a Latrotoxin of the black widow spider venom opens a small, non-closing cation channel. Biochem. Biophys. Res. Commun. 134, 320–325. Warrell, D.A., Shaheen, J., Hillyard, P.D., Jones, D., 1991. Neurotoxic envenoming by an immigrant spider (Steatoda nobilis) in southern England. Toxicon 29, 1263–1265.