CSTX-13 and the Multi-Component Venom of the Spider Cupiennius Salei

CCSSTTXX--1133 aanndd tthhee mmuullttii--ccoommppoonneenntt vveennoomm ooff tthhee ssppiiddeerr CCuuppiieennnniiuuss ssaalleeii

Inauguraldissertation der Philosophisch-naturwissenschaftlichen Fakultät der Universität Bern

vorgelegt von

Benno Wullschleger

von Zofingen

Leiter der Arbeit: Prof. Dr. Wolfgang Nentwig Zoologisches Institut der Universität Bern

CSTX-13 and the multi-component venom of the spider Cupiennius salei

Inauguraldissertation der Philosophisch-naturwissenschaftlichen Fakultät der Universität Bern

vorgelegt von

Benno Wullschleger

von Zofingen

Leiter der Arbeit: Prof. Dr. Wolfgang Nentwig Zoologisches Institut der Universität Bern

Von der Philosophisch-naturwissenschaftlichen Fakultät angenommen.

Bern, den 16.12.2004 Der Dekan: Prof. Dr. P. Messerli

für

Monika INHALTSVERZEICHNIS

INHALTSVERZEICHNIS

Einleitung 001

Publikationen 003

MANUSKRIPT I 004 Benno Wullschleger, Lucia Kuhn-Nentwig, Jan Tromp, Urs Kämpfer, Johann Schaller, Stefan Schürch & Wolfgang Nentwig (2004). CSTX-13, a highly synergistically acting two-chain neurotoxic enhancer in the venom of the spider Cupiennius salei (Ctenidae). Proc. Natl. Acad. Sci. USA 101, 11251–11256.

MANUSKRIPT II 015 Benno Wullschleger, Wolfgang Nentwig & Lucia Kuhn-Nentwig. Spider venom: Enhancement of venom efficacy mediated by different synergistic strategies in Cupiennius salei. Submitted.

MANUSKRIPT III 032 Benno Wullschleger, Wolfgang Nentwig & Lucia Kuhn-Nentwig. Different effects of neurotoxic and cytolytically acting peptides of Cupiennius salei venom to insects.

Zusammenfassung 046

Literatur 048

Danksagung 051

Curriculum vitae 052

EINLEITUNG

Spinnen gehören mit 38'663 rezenten Arten (PLATNICK, 2004) nach den Insekten zu der erfolgreichsten Tiergruppe. Die ebenfalls zu den Arachniden zählenden Skorpione gelten als die ältesten Landtiere, welche sich wahrscheinlich schon vor 400 Millionen Jahren entwickelt haben. Die ältesten fossilen Spinnen stammen aus dem Devon und dürften über 350 Millionen Jahre alt sein (WUNDERLICH, 1986). Trotz unterschiedlicher Habitattypen und Verhaltensmustern ist allen Spinnen gemeinsam, dass sie Prädatoren sind und sich in erster Linie von Arthropoden (v. a. Insekten) ernähren. Wohl nur ausnahmsweise werden Wirbeltiere wie Kaulquappen, kleine Fische, Reptilien, Nagetiere oder Vögel erbeutet (FOELIX, 1996). Ihrer Jagdtechnik entsprechend lassen sich Spinnen in Netzspinnen und Laufspinnen einteilen, wobei sich beide Gruppen durch eine grosse Variabilität auszeichnen. Allen gemeinsam ist aber, dass Beutetiere im Endeffekt mit Gift, welches mit Hilfe von Cheliceren inijziert wird, immobilisiert werden. Dabei nimmt die letale Wirkung des Giftes, im Gegensatz zur Lähmung, eher eine sekundäre Rolle ein (FRIEDEL & NENTWIG, 1989). Als giftlose Spinnen gelten lediglich die sehr ursprünglichen Liphistiidae, welche noch keine Giftdrüsen entwickelt haben (HAUPT, 2003) und die Uloboridae respektive Holarchaea (MARETIC, 1987), welche vermutlich ihre Giftdrüsen sekundär reduziert haben. Nur gerade von vier Gattungen sind Arten bekannt, die auch für den Menschen als potentiell gefährlich einzustufen sind. Hierzu gehören Latrodectus (Theridiidae), Atrax (Hexathelidae), Loxosceles (Sicariidae) und Phoneutria (Ctenidae) (HABERMEHL, 1976). Dagegen sind einheimische Spinnen nicht von medizinischer Bedeutung. Spinnengifte sind Multi-Komponenten-Systeme mit einer noch nicht gut verstandenen Komplexizität. Die Giftkomponenten können grob in drei chemische Klassen eingeteilt werden: Organische Moleküle und Ionen mit einem geringen Molekulargewicht (< 1 kDa), Polypeptide (3–10 kDa) und Proteine (> 10 kDa) (ESCOUBAS et al., 2000). Einige dieser Komponenten haben sich als ideale Werkzeuge erwiesen, um Strukturen und Funktionen von Ionenkanälen zu erforschen (SACCOMANO & AHLIJANIAN, 1994; SWARTZ & MACKINNON, 1997; SUCHYNA et al., 2004; LEE & MACKINNON, 2004). Andere haben das Potential, als Insektizide eingesetzt zu werden (KING et al., 2002; TEDFORD et al., 2004), sie werden verwendet, um Antiseren zu entwickeln (FISHER et al., 1981; SUTHERLAND, 1983) oder könnten Basis für zukünftige Medikamente sein (BODE et al., 2001). KING et al. (2002) bezeichnen Spinnengifte als pharmakologische Goldminen und die Autoren gehen bei der momentan bekannten Anzahl Spinnenarten von mehr als einer Million verschiedener,

1 EINLEITUNG pharmakologisch aktiver Peptide aus. Um so erstaunlicher ist es, dass Spinnengifte im Gegensatz zu anderen Tiergiften vergleichsweise wenig intensiv untersucht wurden. Cupiennius salei (Ctenidae; KEYSERLING, 1877), eine polyphage Spinne aus den tropischen Regionen Zentralamerikas, ist als Jäger in höheren Vegetationsschichten besonders auf ein effektives Gift angewiesen, um ihre Beute nicht zu verlieren (BARTH & SEYFARTH, 1979; NENTWIG, 1986; BARTH, 2001; KUHN-NENTWIG et al., 2004). Da Gift eine wertvolle Ressource ist und die Regeneration bis zu 16 Tage in Anspruch nimmt (BOEVE et al., 1995), geht die Spinne haushälterisch mit ihrem Gift um (MALLI et al., 1999; WIGGER et al., 2002; WULLSCHLEGER & NENTWIG, 2002). Das Gift von C. salei beinhaltet im Wesentlichen Cupiennius salei Toxine (CSTX-1 bis 13), zytolytisch aktive Cupiennine, niedermolekulare Substanzen wie Ionen und einzelne Aminosäuren und hochmolekulare Substanzen (Bsp. Hyaluronidase). Insgesamt setzt sich das Rohgift wohl aus über 100 verschiedenen Substanzen zusammen (KUHN-NENTWIG et al., 1994; KUHN-NENTWIG et al., 2004). Bei diesem ausgesprochen komponentenreichen Gift stellt sich natürlich die Frage nach dessen Zweck. Da C. salei ein breites Beutespektrum hat, wären beispielsweise beutetierspezifische Komponenten denkbar. Die hier vorliegende Dissertation zeigt indes, dass die Funktion der einzelnen Komponenten wohl eher darin liegt, mit anderen Komponenten synergistisch zu interagieren. Allerdings lassen die einzelnen Sensitivitätsexperimente mit Rohgift an verschiedenen Beutetierarten bereits den Schluss zu (KUHN-NENTWIG et al., 1998), dass Komponenten je nach Testorganismus unterschiedlich gut funktionieren.

Die drei vorliegenden Manuskripte geben einerseits Auskunft über Struktur und Wirkung von CSTX-13 und stellen andererseits einen Ansatz dar, die Komplexität des Giftes von Cupiennius salei besser verstehen zu können.

2 PUBLIKATIONEN

PUBLIKATIONEN

3 MANUSKRIPT I

MANUSKRIPT I

Wullschleger, B., Kuhn-Nentwig, L., Tromp, J., Kämpfer, U., Schaller, J., Schürch, S. & Nentwig, W. (2004). CSTX-13, a highly synergistically acting two-chain neurotoxic enhancer in the venom of the spider Cupiennius salei (Ctenidae). Proc. Natl. Acad. Sci. USA 101, 11251–11256.

4 CSTX-13, a highly synergistically acting two-chain neurotoxic enhancer in the venom of the spider Cupiennius salei (Ctenidae)

Benno Wullschleger*, Lucia Kuhn-Nentwig*†, Jan Tromp‡, Urs Ka¨ mpfer‡, Johann Schaller‡, Stefan Schu¨ rch‡, and Wolfgang Nentwig*

*Zoological Institute, University of Bern, Baltzerstrasse 6, CH-3012 Bern, Switzerland; and ‡Department of Chemistry and Biochemistry, University of Bern, Freiestrasse 3, CH-3012 Bern, Switzerland

Edited by Jerrold Meinwald, Cornell University, Ithaca, NY, and approved June 22, 2004 (received for review April 1, 2004) The survival of the spider Cupiennius salei depends on its hunting cysteine in their composition. Some peptides have already been success, which largely relies on its immediately paralyzing multi- characterized and named the cupiennin 1 family. These peptides component venom. Here, we report on the isolation and charac- exhibit strong bactericidal activities in the submicromolar terization of CSTX-13, a neurotoxic enhancer in the spider venom. range, but also cytolytic and insecticidal activities. In addition, De novo elucidation of the disulfide bridge pattern of CSTX-13 and cupiennin 1a shows a high synergistic effect with the main the neurotoxin CSTX-1 by tandem MS revealed an identical ar- neurotoxin CSTX-1, facilitating a rapid paralysis (12, 13). Pos- rangement. However, in contrast to CSTX-1, CSTX-13 is a two-chain itive insecticidal cooperativity between the cytolytically active peptide with two interchain and two intrachain disulfide bridges. oxyopinins and neurotoxins is also reported for the spider Furthermore, the insecticidal activity of CSTX-13 is synergistically Oxyopes kitabensis (14). ؉ increased in the presence of K ions as well as of the cytolytic The second group involves neurotoxically active peptides with peptide cupiennin 1a. We demonstrated that the weakly neuro- molecular masses of Ϸ8 kDa, which we named Cupiennius salei toxic CSTX-13 enhances the paralytic activity of the neurotoxin toxins (CSTX-1 to -13) (11). There is evidence that CSTX-1 ϩ CSTX-1 by 65% when it is administered with the latter at its entirely inhibits L-type Ca2 channels of GH3 cells (J. S. Cruz, personal nontoxic physiological concentration, which is 440 times below its communication). To date, sequence data for the neurotoxins LD50 concentration. CSTX-1 and CSTX-9 are available. The four disulfide bridges of CSTX-9 form linkages between C1–C4; C2–C5; C3–C8; and piders and scorpions use their venom to paralyze prey and͞or C6–C7 (15–17). This arrangement is also found in other spider Sto defend against predators. These venoms are complex neurotoxins belonging to the inhibitor cystine knot (ICK) struc- mixtures of different components, and the knowledge about tural motif (18). their interactions and role in the envenomation process is still So far, a roughly comparable two-chain peptide structure has limited. During their evolution, these arthropods have developed only been reported for the spider Agelenopsis aperta: the ␻-aga- a large number of neurotoxins that act simultaneously on various toxins IA and G block presynaptic calcium channels in insect invertebrate and͞or vertebrate membrane-bound sodium, po- neuromuscular junction (19, 20). A two-chain structure has also tassium, and calcium channels. Also, interactions with acid- been proposed for the Hololena toxin, a presynaptic antagonist sensing ion channels, glutamate receptors, and as yet unidenti- of insect neuromuscular transmission (21). fied targets lead to rapid paralysis or death of the envenomed Here, we present the amino acid sequence of CSTX-13, an animals (1, 2). enhancer peptide from the venom of C. salei, and the de novo Cupiennius salei (Keyserling, 1877) is a nocturnal hunting determination of the disulfide bridge pattern of CSTX-1 and spider living in the Central American rain forest (3). The spider CSTX-13. Despite its sequence similarity to the neurotoxins BIOCHEMISTRY relies on an immediately paralyzing venom activity because, in its CSTX-1 and CSTX-9, CSTX-13 acts as a neurotoxic enhancer. arboreal environment, a prey item that escapes is lost. The spider Moreover, its low neurotoxic activity is also augmented by other also loses its venom investment and reduces its chance of venom compounds. successfully subduing a subsequent prey item, because its venom storage is limited, regeneration takes Ϸ16 days (4), and its Materials and Methods production involves high metabolic costs. Behavioral, ecological, Chemicals. Chemicals were of analytical grade and purchased and biochemical investigations of the venom economy of C. salei from Merck unless otherwise specified. indicate that it alters the amount of venom injected according to the size, mobility, and defense behavior of its prey (5–8). Isolation of CSTX-13. Spider maintenance, venom collection, and This economical venom use is paralleled on the physiological separation of 425 ␮l of venom by FPLC and HPLC methods were and biochemical levels by the interactions of different venom performed as described (ref. 11; see also Supporting Text, which components (9). In the venom, low molecular mass compounds is published as supporting information on the PNAS web site). such as histamine (5.7 mM) and free amino acids, basically Final purification of CSTX-13 was achieved by RP-HPLC on a ␮ ␮ ϩ ϩ ϫ taurine (70 M) and glycine (43.3 M), are present. K ,Na , nucleosil 100–5 C8 column (4 250 mm, Macherey & Nagel) 2ϩ ϩ and Ca ions have also been identified. Remarkably, K ions using 22% solvent B (0.1% trifluoroacetic acid in acetonitrile) in are abundant in the venom and rare in the hemolymph (10, 11). Furthermore, C. salei possesses a complex multicomponent system consisting of a few proteins with molecular masses Ͼ10 This paper was submitted directly (Track II) to the PNAS office. kDa, among them a highly active hyaluronidase. About 100 Abbreviation: ESI, electrospray ionization different peptides with molecular masses between 2 and 8 kDa Data deposition: The sequences reported in this paper have been deposited in the Swiss- have been detected by electrospray ionization (ESI)-MS. The Prot and TrEMBL databases [accession nos. P83919 (CSTX-13 chain A) and P83920 (CSTX-13 peptides can be roughly divided into two groups. The first group chain B)]. contains the smaller peptides with molecular masses of Ϸ3–4 †To whom correspondence should be addressed. E-mail: [email protected]. kDa, which are mainly highly cationic ␣-helical peptides without © 2004 by The National Academy of Sciences of the USA www.pnas.org͞cgi͞doi͞10.1073͞pnas.0402226101 PNAS ͉ August 3, 2004 ͉ vol. 101 ͉ no. 31 ͉ 11251–11256 Fig. 1. Isolation of CSTX-13 from the venom of C. salei.(A) Crude venom was first separated by gel filtration on a Superdex 75 column. (B) Further separation of the pooled fraction was achieved by cationic exchange chromatography on a Mono S HR column. (C) When RP-HPLC was used, CSTX-13 was finally isolated on a nucleosil 100–5C8 column, and the purity was controlled by SDS͞PAGE. (D) Lanes: 1, native CSTX-13; 2, reduced CSTX-13; 3, native CSTX-1; 4, molecular mass markers (14.4–97.4 kDa; Bio-Rad); and 5, molecular mass markers (2.5–16.9 kDa; Amersham Pharmacia). (E) RP-HPLC of reduced and alkylated CSTX-13 on a nucleosil 100–5C8 column resulted in the separation of chain A (first arrow) and chain B (second arrow). solvent A (0.1% trifluoroacetic acid in water) with a flow rate of supernatant was recovered for separation by RP-HPLC. The 0.5 ml͞min. Directly after injection of the sample, the gradient digested and purified peptides were subjected to collision- (22–28% solvent B) was started for 20 min (Fig. 1C). This step induced dissociation using nitrogen as the collision gas. Collision was repeated several times to obtain CSTX-13 (variability be- energies were in the range of 20–80 eV. tween different preparations: 0.5–1.3 mg). Amino Acid Analysis and Amino Acid Sequence Analysis. Samples PAGE. SDS͞PAGE and silver staining of native and reduced were hydrolyzed in the gas phase with 6 M hydrochloric acid (2-mercaptoethanol) CSTX-13 were performed with the Phast- containing 0.1% (by volume) phenol for 24 h at 115°C under N2 System using high density PhastGel (Amersham Pharmacia). vacuum according to Chang and Knecht (22). N-terminal sequence analysis was carried out either in a Procise cLC 492 Reduction and Alkylation. Fifty micrograms of CSTX-13 was protein sequencer or in a pulsed liquid-phase sequencer 477A, reduced and alkylated according to the published procedure (ref. both from Applied Biosystems. 15, see Supporting Text). Chains A and B were further desalted ϫ and separated by RP-HPLC on a nucleosil 100-5 C8 column (4 Experiments with Spider Digestive Liquid. Digestive liquid from C. 250 mm; Macherey & Nagel) using 100% solvent A with a flow salei was obtained by electrical stimulation and collected in glass rate of 0.5 ml͞min for 0–5 min followed by a 55-min gradient of capillary tubes, and 8 ␮l of diluted digestive liquid (1:100 with 0.73% solvent B in solvent A per min (Fig. 1E). water) was mixed with 17 ␮g of CSTX-13 in 8 ␮l of water. The mixture was kept at 24°C and, after 0.5, 1, and 24 h, aliquots of MS. Mass spectrometric analyses were performed on a QSTAR 2 ␮l were analyzed by ESI-MS. Pulsar hybrid quadrupole time-of-flight mass spectrometer (Ap- For further experiments, 50 ␮g of CSTX-13 was dissolved in plied Biosystems) equipped with a nanoelectrospray ion source. 31.8 ␮l of water, mixed with 31.8 ␮l of diluted digestive liquid The instrument was tuned for a mass resolving power of 12,000 (1:100 with water), and incubated for 24 h at 24°C, and the ͞⌬ (m m, full width at half maximum) and calibrated with caesium fragment was isolated by RP-HPLC on a nucleosil 120–5C18 iodide and reserpine (Sigma). Samples were dissolved in meth- column (2 ϫ 125 mm; Macherey & Nagel) using a gradient of anol͞water (1:1 vol͞vol) containing 1% formic acid. The final 0.2%BinA͞min for 200 min and a flow rate of 0.5 ml͞min. For peptide concentration was 5 pmol͞␮l. All analyses were per- ESI-MS analysis, the CSTX-13 fragment was again dissolved in formed in the positive ion mode. Numbers represent monoiso- 100 ␮l of buffer [100 mM Tris⅐HCl, pH 8.0, containing 135 ␮M topic masses. TLCK (N␣-p-tosyl-L-lysine chloromethyl ketone, Sigma) and 220 For elucidation of the disulfide bridge pattern, 50 ␮g of native ␮M TPCK (N-tosyl-L-phenylalanine chloromethyl ketone, CSTX-13 was cleaved with immobilized trypsin (23 ␮l wet gel Sigma)], reduced, alkylated, and separated as described above. containing 0.5 units of trypsin (Sigma) in 50 ␮l of 0.1 M Tris⅐HCl buffer, pH 8.1, and 1.0 mM iodoacetamide, Fluka) under gentle Bioassays and Calculations. Bioassays were performed according shaking for 17 h at 24°C. The suspension was centrifuged, and the to Escoubas et al. (23) using 1- to 3-day-old Drosophila melano-

11252 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0402226101 Wullschleger et al. Fig. 2. Sequence comparison and disulfide bridge arrangement of CSTX-13, CSTX-1, CSTX-9 (C. salei), and ␻-agatoxin IA (A. aperta). Identical amino acids are shaded gray, the disulfide bridges are represented by lines, and the corresponding cysteine residues are shaded black. The disulfide bridge patterns of CSTX-13 and CSTX-1 were determined by nanoelectrospray tandem MS of the corresponding disulfide-linked tryptic fragment. gaster female flies. The injected volume was 0.05 ␮lof0.1M successive RP-HPLC (Fig. 1C). The retention profile of ammonium acetate, pH 6.1 (control), and all further injections CSTX-13 revealed no impurities. CSTX-13 was characterized by with different components were carried out in this solution, ESI-MS and amino acid composition. The yield of CSTX-13 except nifedipine (Calbiochem), which was injected in Ն0.06 M obtained by purification of crude venom was 1.2–3.0 ␮g͞␮l ammonium acetate, pH 6.1, and Յ5.6 M dimethyl sulfoxide. To depending on the separation protocol. estimate the LD50 (24 h after injection), 20 flies were used as N-terminal sequence analysis of native CSTX-13 provided control, and 20 flies were used for each concentration. evidence for a two-chain molecule: Ser͞Ala–Asp͞Lys–Xaa͞Lys– To investigate synergistic effects between CSTX-13 and fur- Thr͞Glu–Leu͞Leu–Arg͞Xaa–Asn͞Thr–His͞Xaa–Asp͞Gln– ther venom components such as cupiennin 1a (9.6 ␮M) (in Xaa͞Gln. Therefore, CSTX-13 was reduced and alkylated, and nontoxic concentration), histamine (5.7 mM, Sigma), taurine chains A and B were separated by RP-HPLC. Both chains were (0.07 mM, Sigma), and KCl (215 mM) (all in physiological venom sequenced by Edman degradation from the N to the C termini concentrations), bioassays were performed with 12.6 pmol of without any ambiguity. Chain A is composed of 34 residues CSTX-13 per mg of fly. We tested CSTX-13 alone and in (measured, 4,342.73 Da; calculated, 4,342.76 Da), and chain B is combination with each of the above mentioned venom compo- composed of 29 residues. ESI-MS of chain B gave a monoiso- nents [n ϭ 2 ϫ (15 ϫ 5) for each assay]. Venom components in topic mass of 3,475.80 Da, which is one mass unit less than the above-mentioned concentrations were injected alone as control expected theoretical mass of 3,476.83 Da, thus indicating C- (n ϭ 20). The paralytic activity of physiological KCl concentra- terminal amidation of chain B. The determined amino acid tion was measured by comparing the awake time of a control sequences of both chains agree well with the amino acid com- group (n ϭ 20) and a treated group (n ϭ 20) (Mann–Whitney U position of native CSTX-13 as well as with the individual chains. test, SPSS 10 software). Taking into account the four disulfide bridges of both peptide To highlight synergistic effects between the neurotoxins chains and the amidation, the calculated monoisotopic mass of CSTX-1 and CSTX-13, corresponding to their molar ratio in the CSTX-13 is in agreement with the measured mass of native crude venom (9:1), bioassays were performed with 0.315 pmol of CSTX-13 (measured, 7,354.51 Da; calculated, 7,354.37 Da) (Fig. CSTX-1 per mg of fly alone and in combination with 0.035 pmol 2 and Fig. 5, which is published as supporting information on the of CSTX-13 per mg of fly, and three further concentrations down PNAS web site). to 0.63 fmol of CSTX-13 per mg of fly [n ϭ 2 ϫ (12 ϫ 5) for each SDS͞PAGE analysis of purified native CSTX-13 revealed a assay]. CSTX-13 alone was used in the above mentioned con- single band at 12 kDa, whereas reduced CSTX-13 revealed a centrations as a control (n ϭ 2 ϫ 20). single band at Ϸ3 kDa, obviously containing the peptide chains BIOCHEMISTRY The influence of CSTX-13 on two different Ca2ϩ channel A and B. This supports the two-chain structure of the native blockers was further investigated. NiCl2 was administered in a CSTX-13 (Fig. 1 C and D). concentration of 5.26 nmol͞mg of fly alone and in combination To exclude the possibility that the two-chain structure of with 0.035 pmol of CSTX-13 per mg of fly [n ϭ 2 ϫ (12 ϫ 5) for CSTX-13 is a proteolytic artifact because of contamination of the each assay]. Nifedipine was tested in a concentration of 0.105 venom with digestive liquid (24), CSTX-13 was incubated with nmol͞mg of fly alone and in combination with 0.035 pmol of fresh digestive liquid. After 0.5, 1, and 24 h of incubation, the CSTX-13 per mg of fly [2 ϫ (6 ϫ 5) for each assay], and three obtained mass indicates a proteolytic degradation of the 14 further concentrations up to 5.5 pmol of CSTX-13 per mg of fly C-terminal amino acid residues of chain B (measured, 5,746.51 [n ϭ 6 ϫ 5 for each assay]. Da; calculated, 5,746.51 Da). The CSTX-13 fragment was puri- The relative mortality of D. melanogaster was arcsin square fied by RP-HPLC, and reduced and alkylated in the presence of root-transformed and treated as the dependent variable, two protease inhibitors. ESI-MS analysis of the purified com- whereas the venom components or CSTX-13 were treated as pounds revealed an intact chain A (measured, 4,342.80 Da; nominal independent variables. The experiment was analyzed by calculated, 4,342.76 Da) and a truncated chain B (measured, generalized linear models. The means of the nominal indepen- 1,868.03 Da; calculated, 1,867.98 Da). These findings are in dent variables venom components or CSTX-13, respectively, accordance with the result described above and support the were compared pairwise by the Bonferroni method. Fulfillment assumption of a native two-chain structure of CSTX-13 in the of the model assumptions was checked by visual inspection of the venom. CSTX-13 seems to be present in the venom as a residuals distribution for every statistical test conducted. Statis- two-chain molecule and to the best of our knowledge does not tics were performed with S-PLUS 6.0 PROFESSIONAL software. represent a purification artifact. Because of the unique amino acid sequences of CSTX-1 and Results CSTX-13 with cysteine residues arranged in close proximity, Purification and Sequence Analysis of CSTX-13. The crude venom classical approaches to determine the disulfide bridge pattern, (425 ␮l) was separated in a four-step protocol using gel filtration based on specific enzymatic or chemical cleavages, failed. Con- (Fig. 1A), cationic exchange chromatography (Fig. 1B), and sequently, the disulfide bridge patterns of CSTX-1 and CSTX-13

Wullschleger et al. PNAS ͉ August 3, 2004 ͉ vol. 101 ͉ no. 31 ͉ 11253 Fig. 3. Sequence of the cystine-containing fragment obtained from tryptic digest of native CSTX-13 (4,473.73 Da). The asterisk (*) indicates the [M ϩ 8H]8ϩ ion with m͞z 560.23, which was selected as the precursor for CID. Typical fragmentation pathways include loss of terminal amino acids in conjunction with loss of water (ions I, I, and III), disulfide cleavage (ion ␴), and cleavage of peptide bonds between cystines (disulfide bridge-defining ions A–Z). The enlargement shows the isotopic pattern of fragment ion C with m͞z 500.22, which defines, in combination with fragment ion A, the Cys 1–Cys 4 and Cys 2–Cys 5 bridges. were identified de novo by nanoelectrospray tandem MS. Diges- calculated masses. Fig. 3 also shows a section of the product ion tion of the native toxin with immobilized trypsin yielded main spectrum obtained by CID of the [M ϩ 8H]8ϩ precursor ion of fragments consisting of five short peptide chains cross-linked by the tryptic CSTX-13 fragment. The high mass accuracy and four disulfide bridges. resolving power of the tandem mass spectrometer allow unam- Multiply charged [M ϩ nH]nϩ ions (n ϭ 3–9) of the cystine biguous peak assignment, as demonstrated for the quadruply containing tryptic fragment (measured, 4,473.75 Da; calculated, charged fragment ion C. Additional information was obtained by 4,473.76 Da) of CSTX-13 were selected as precursor ions for assigning peaks generated by disulfide bridge cleavage. The same subsequent collision-induced dissociation (CID). The resulting strategy was applied for the elucidation of the disulfide bridge product ion spectra are characterized by abundant peaks of pattern of CSTX-1, which exhibits identical disulfide bridges fragment ions generated by cleavage of the disulfide bridges. (Fig. 2). These ions define the order of peptide chains. Further abundant peaks indicate repetitive loss of amino acids from the termini of Synergistic Insecticidal Effects. To evaluate the biological impor- the peptide chains, often occurring in conjunction with the loss tance of CSTX-13, comparative bioassays with D. melanogaster ͞ of water. Detailed information on the disulfide bridge pattern were performed. The LD50 of 16.3 pmol mg of fly (14.5–27.1; was obtained by detection of the less abundant fragment ions of 95% confidence limits) indicates a lower toxicity than other mass 968.43 Da (A), 1,895.80 Da (B), 1,995.85 Da (C), 1,289.45 neurotoxins of C. salei. CSTX-13 is Ϸ49 times less toxic than the Da (Y), and 1,390.49 Da (Z), generated by cleavage of the neurotoxin CSTX-1, and 2.8 times less toxic than the cytolytically peptide backbone between adjacent cystines. The corresponding active peptide cupiennin 1a, both key components identified in cleavage sites are indicated in Fig. 3 (see Fig. 6, which is the venom of C. salei (24, 12) (Table 1). published as supporting information on the PNAS web site). Synergistic interactions of different venom components (taurine, Measurements exhibit a maximum deviation of 0.02 Da from the histamine, KCl) with the paralytic activity of CSTX-13 were ana-

Table 1. Insecticidal activity of spider venom components

Venom Physiological venom LD50 Conﬁdence Synergistically tested components concentration, mM concentration limits, 95% concentration

CSTX-1 1.4–3.3 6.3 ␮M 5.2–7.6 CSTX-9 0.2–1.1 211 ␮M 190–273 CSTX-13 0.2–0.4 309 ␮M 275–514 0.01–0.7 ␮M Cupiennin 1a Ϸ1.2 106 ␮M75–149 9.6 ␮M Taurine 0.07 Ͼ160 mM 0.07 mM Histamine 5.7 917 mM 794–1,100 5.7 mM KCl 215 1,845 mM 1,650–2,131 215 mM

Estimation of the lethal doses (LD50)inaDrosophila bioassay, where 50% of the test ﬂies died of intoxication 24 h after injection. Different amounts of peptides, histamine, taurine, and KCl were dissolved in 0.1 M ammonium acetate, at a pH of 6.1, and 0.05 ␮l was injected into the ﬂies. The physiological concentrations of CSTX-1 (13), taurine (13), histamine (13), KCl (13), CSTX-9 (15), and cupiennin 1a (12) in the venom were reported.

11254 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0402226101 Wullschleger et al. ␮ only in high concentrations (LD50 309 M) when applied alone, but synergistically enhances the paralytic activity of the main neurotoxin CSTX-1 at low concentrations (0.7 ␮M). In the venom, CSTX-13 is constitutively present at a 7–8 times lower concentration than the main neurotoxin CSTX-1 and in an up to 2.8 times lower concentration than a further neurotoxin CSTX-9. Similarly, its insecticidal activity, expressed as a LD50 value, is 49 times lower than the activity of CSTX-1 and 1.5 times lower than that of CSTX-9 (Table 1) (24). Protein database search using BLASTP 2.2.8 (29) resulted in a high sequence identity of 56% (70% similarity) between CSTX-1 and CSTX-9, but in lower sequence identities of 35% between CSTX-13 and CSTX-1 (51% similarity), and of 31% between CSTX-9 and CSTX-13 Fig. 4. Synergistic effects between CSTX-13 and venom components. (A) (49% similarity). Nevertheless, all three peptides exhibit iden- Synergistic effects between CSTX-13 and low molecular venom components. tical disulfide bridge patterns (15–17) (Fig. 2). Sequence com- In a Drosophila bioassay, the lethal effect of CSTX-13 injected alone (239.4 ␮M) parison implies that, upon processing, a short peptide is excised was compared with the lethal effect of coinjected CSTX-13 (239.4 ␮M) with in the loop forming the disulfide bridge C6–C7, thus leading to taurine (0.07 mM; not significant), histamine (5.7 mM; not significant), KCl the two-chain structure of CSTX-13. (215 mM; *, P Ͻ 0.05), or cupiennin 1a (9.6 ␮M; ***, P Ͻ 0.001). As controls, taurine, histamine, KCl, and cupiennin 1a showed no toxic effect when No sequence similarities were detected with further neurotoxins and the two other sequenced two-chain calcium channel administered alone. (B) Synergistic effects of CSTX-13 on the toxicity of ␻ ␻ CSTX-1. The lethal effect of CSTX-1 (5.99 ␮M) was compared with the lethal blockers -agatoxin IA (66 and 3 residue chains) and -agatoxin effect of coinjected CSTX-1 (5.99 ␮M) with CSTX-13 (0.67 ␮M͞mg of fly; ***, G (62 and 3 residue chains) (20) from the spider Agelenopsis P Ͻ 0.001) (molar ratio of 9:1) corresponding to their concentrations in the aperta. In contrast to CSTX-13, which contains two interchain venom. CSTX-13 was also injected alone as control and showed no toxic effect and two intrachain disulfide linkages, these two neurotoxins on the flies. Statistical analysis was done by using the Bonferroni method. possess four intrachain and one interchain disulfide linkage. ␻- Standard error bars are shown for every treatment. agatoxin IA is formed from its precursor by excision of an internal heptapeptide leading to a major peptide chain that is connected to the minor peptide chain (three residues) by one lyzed. Taurine itself was not toxic to D. melanogaster up to 8.9 disulfide bridge (19) (Fig. 2). nmol͞mg of fly. The LD of histamine was 51.0 nmol͞mg of fly 50 Unlike the above mentioned ␻-agatoxins, CSTX-13 is neuro- (44.2–61.2; 95% confidence limits) and KCl showed a LD of 102.5 50 toxic by itself only at a high micromolar concentration. This nmol͞mg of fly (91.7–118.4; 95% confidence limits) (Table 1). circumstance raised the question of whether the two-chain When tested alone at physiological concentrations, taurine structure of CSTX-13 might be the result of a purification artifact (3.88 pmol͞mg of fly) and histamine (316.67 pmol͞mg of fly) showed no effects in a Drosophila bioassay. Injection of KCl caused by contamination with proteases. The purification (11.94 nmol͞mg of fly) showed a short significant paralytic effect protocols of CSTX-13 over the last 5 years have always resulted (564.7 s Ϯ SD 288.9 s; P Ͻ 0.001) when compared with the in a pure peptide with identical molecular masses. Experiments control group (234.5 s Ϯ SD 116.5 s). An injection of 12.6 pmol with spider digestive liquid, which could be the major source of of CSTX-13͞mg of fly resulted in a mortality of 39%. No protease contaminations, resulted only in a C-terminal trunca- statistically significant differences were observed by coinjection tion of 14 residues of chain B. As shown previously, C-terminal of CSTX-13 with taurine (mortality of 34%) or with histamine proteolytic degradation of CSTX-1 by spider digestive liquid (mortality of 42%). In contrast, coinjection of CSTX-13 with stopped at position 49 (Gly) (24). Therefore, we conclude that KCl significantly increased the mortality to 59% (P Ͻ 0.05) (Fig. the two-chain structure of CSTX-13 is not a purification artifact or protease degradation product, but a valid constitutive 4A). When injected alone, the cytolytic cupiennin 1a is not toxic BIOCHEMISTRY to D. melanogaster in a concentration of 0.53 pmol͞mg of fly, but component of the C. salei venom. Whether the two-chain it increases the mortality of CSTX-13 from 39% to 97% (P Ͻ structure is posttranslationally generated, however, remains 0.001) (Fig. 4A). to be investigated. In addition, we investigated the synergistic effect of CSTX-13 on the toxicity of CSTX-1. At physiological concentrations in the Biological Function of CSTX-13 in the Venom. To analyze the bio- venom, the molar ratio of CSTX-1 and CSTX-13 is 9:1. With logical function of CSTX-13, we investigated possible interac- administration of one peptide alone, 0.315 pmol of CSTX-1 per tions between CSTX-13 and different venom components in a mg of fly caused a mortality of 31%, and injection of 0.035 pmol Drosophila bioassay. Previously, we have shown that the neuro- of CSTX-13 per mg of fly had no effect. Surprisingly, coinjection toxicity of the main neurotoxin CSTX-1 to blow flies (Proto- of CSTX-1 and CSTX-13 in the above mentioned molar ratio of phormia sp.) could be increased when coinjected with taurine 9:1 significantly increased the mortality to 96% (P Ͻ 0.001) (Fig. and histamine (9). However, coinjection of CSTX-13 with 4B), and, even in a molar ratio of 500:1, the enhancing effect of taurine or histamine in its physiological venom concentration did CSTX-13 was observed (45%, not significant). not increase its insecticidal activity. Nevertheless, histamine as a ϩ In view of the fact that CSTX-1 inhibits L-type Ca2 channels neurotransmitter and taurine as a neuromodulator play an and that, in Drosophila muscle, a 1,4-dihydropyridine-sensitive important role in the insect nerve system (30–33). ϩ (25) homolog of the mammalian L-type͞␣1D (Dmca1D) subunit Remarkably, the venom of C. salei exhibits a very high K ion gene is expressed (26), the influence of CSTX-13 on the activity concentration that is 32-fold higher than in the hemolymph, and of the L-type calcium channel blocker nifedipine (27) as well as even 2.7-fold higher than in the prevenom of the scorpion NiCl2, a general inhibitor of calcium channels (28), was inves- Parabuthus transvaalicus (11, 34). Hammock and coworkers (34) tigated. No synergistic effects between NiCl2 or nifedipine and suggest an economically motivated strategy in venom utilization CSTX-13 were detected. for this scorpion. P. transvaalicus first secretes a prevenom containing a high Kϩ ion concentration at a low protein content, Discussion whereas the subsequently secreted venom is characterized by a The Structure of CSTX-13. In CSTX-13, we have characterized a high protein content and a 15-fold lower Kϩ ion concentration. two-chain peptide from the venom of C. salei. It paralyzes flies The synergistic activity in the prevenom between the ‘‘inexpen-

Wullschleger et al. PNAS ͉ August 3, 2004 ͉ vol. 101 ͉ no. 31 ͉ 11255 sive’’ Kϩ ion and the assumed inhibitors of rectifier Kϩ channels neurotoxic enhancer CSTX-13 show that it enhances the efficacy is proposed as a means of conserving metabolically expensive of the neurotoxin CSTX-1 at a concentration of 440 times below neuropeptides in the venom (34). In part, C. salei also uses this its LD50. Tests with different concentrations of CSTX-13 re- strategy to enhance its venom efficacy. Coinjection of CSTX-13 vealed a positive correlation between the amount of CSTX-13 with Kϩ ions increases the mortality of the flies by 20%. The and the efficacy of CSTX-1. The cooperation between CSTX-1 synergistic cooperation of Kϩ ions is also detectable when and CSTX-13 seems to be highly specific, because no synergistic ϩ 2ϩ applied together with CSTX-1, a suggested L-type Ca2 channel interactions between CSTX-13 and other Ca channel blockers, ϩ blocker (B.W. and L.K.-N., unpublished data). The high K such as nifedipine and NiCl2, were found. concentration in the venom alone caused an immediate short Conclusions paralysis, and there seems to be a general cooperation between Kϩ ions and various ion channel blockers described here for a In summary, the structural and biological characterization of labidognath spider. CSTX-13 provide further insight into the complexity of C. salei Enhancement of insecticidal efficacy through the cooperative venom as more multiple interactions between different venom interaction of different venom peptide neurotoxins in spiders components become apparent. After venom injection into a (35) and scorpions (36, 37) has been well investigated. Addi- prey animal, the hyaluronidase seems to act as a spreading tionally, synergistic interactions between acylpolyamines and factor, followed by the dual cytolytic activity of the cupiennins. cysteine-rich peptide neurotoxins (38) as well as between cyto- They facilitate the activity of the neurotoxins and at the same time protect the venom duct and glands against bacterial lytic peptides and neurotoxins have been described (12–14). invasion by membrane disturbance and pore building. Addi- These positive interactions were principally demonstrated by tionally, antimicrobial peptides may also modulate intracellu- applying both components in toxic concentrations. ϩ lar signaling by increasing intracellular Ca2 , as reported for In contrast, a nontoxic concentration of the cytolytically active parabutoporin and opistoporin from scorpion venoms (39). cupiennin 1a (20 times lower than its LD50) dramatically en- Simultaneously, the inhibition of ion channels by the neuro- hances the efficacy of CSTX-1 (12, 13). The same effect was toxins is further enhanced by the high Kϩ ion concentration in observed when testing CSTX-13 and cupiennin 1a. It is assumed the venom, shifting the Kϩ equilibrium potential (34). Finally, that, in both cases, mainly through the nonspecific cytolytic the neurotoxins act on different ion channels with a concom- activity of cupiennin 1a, CSTX-1 and CSTX-13 have better itant enhancement by CSTX-13. access to their targets. Surprisingly, when CSTX-1 and CSTX-13 were administered We thank Dr. Patrik Kehrli and Dr. Sven Bacher for statistical advice, together at their venom concentrations, a strong positive coop- Dr. Heather Murray for critical comments on the manuscript, and the eration was found. The data presented here on the two-chain Swiss National Science Foundation for funding.

1. Loret, E. & Hammock, B. (2001) in Scorpion Biology and Research, eds. 22. Chang, J.-Y. & Knecht, R. (1991) Anal. Biochem. 197, 52–58. Brownell, P. & Polis, G. (Oxford Univ. Press, New York), pp. 204–233. 23. Escoubas, P., Palma, M. F. & Nakajima, T. (1995) Toxicon 33, 1549– 2. Escoubas, P., Diochot, S. & Corzo, G. (2000) Biochimie 82, 893–907. 1555. 3. Barth, F. G. (2002) A Spider’s World: Senses and Behavior (Springer, New York). 24. Kuhn-Nentwig, L., Schaller, J., Ka¨mpfer, U., Imboden, H., Malli, H. & 4. Boeve´, J.-L., Kuhn-Nentwig, L., Keller, S. & Nentwig, W. (1995) Toxicon 33, Nentwig, W. (2000) Arch. Insect Biochem. Physiol. 44, 101–111. 1347–1357. 25. Gielow, M. L., Gu, G.-G. & Singh, S. (1995) J. Neurosci. 15, 6085–6093. 5. Malli, H., Imboden, H. & Kuhn-Nentwig, L. (1998) Toxicon 36, 1959–1969. 26. Ren, D., Xu, H., Eberl, D. F., Chopra, M. & Hall, L. M. (1998) J. Neurosci. 18, 6. Malli, H., Kuhn-Nentwig, L., Imboden, H. & Nentwig, W. (1999) J. Exp. Biol. 2335–2341. 202, 2083–2089. 27. Cohen, C. J., Ertel, E. A., Smith, M. M., Venema, V. J., Adams, M. E. & 7. Wigger, E., Kuhn-Nentwig, L. & Nentwig, W. (2002) Toxicon 40, 749–752. Leibowitz, M. D. (1992) Mol. Pharmacol. 42, 947–951. 8. Wullschleger, B. & Nentwig, W. (2002) Funct. Ecol. 16, 802–807. 28. Wakamori, M., Strobeck, M., Niidome, T., Teramoto, T., Imoto, K. & Mori, Y. 9. Kuhn-Nentwig, L., Bu¨cheler, A., Studer, A. & Nentwig, W. (1998) Naturwis- (1998) J. Neurophysiol. 79, 622–634. senschaften 85, 136–138. 29. Altschul, S. F., Madden, T. L., Scha¨ffer, A. A., Zhang, J., Zhang, Z., Miller, W. 10. Loewe, R., Linzen, B. & von Stackelberg, W. (1970) Z. Vergl. Physiol. 66, 27–34. & Lipman, D. J. (1997) Nucleic Acids Res. 25, 3389–3402. 11. Kuhn-Nentwig, L., Schaller, J. & Nentwig, W. (1994) Toxicon 32, 287–302. 30. Zheng, Y., Hirschberg, B., Yuan, J., Wang, A. P., Hunt, D. C., Ludmerer, S. W., 12. Kuhn-Nentwig, L., Mu¨ller, J., Schaller, J., Walz, A., Dathe, M. & Nentwig, W. Schmatz, D. M. & Cully, D. F. (2002) J. Biol. Chem. 277, 2000–2005. (2002) J. Biol. Chem. 277, 11208–11216. 31. Witte, I., Kreienkamp, H.-J., Gewecke, M. & Roeder, T. (2002) J. Neurochem. 13. Kuhn-Nentwig, L., Schaller, J. & Nentwig, W. (2004) Toxicon 43, 543–553. 83, 504–514. 14. Corzo, G., Villegas, E., Go´mez-Lagunas, F., Possani, L. D., Belokoneva, O. S. 32. Buchner, E., Buchner, S., Burg, M. G., Hofbauer, A., Pak, W. L. & Pollack, I. & Nakajima, T. (2002) J. Biol. Chem. 277, 23627–23637. (1993) Cell Tissue Res. 273, 119–125. 15. Schaller, J., Ka¨mpfer, U., Schu¨rch, S., Kuhn-Nentwig, L., Haeberli, S. & 33. Bicker, G. (1991) Brain Res. 560, 201–206. Nentwig, W. (2001) Cell. Mol. Life Sci. 58, 1538–1545. 34. Inceoglu, B., Lango, J., Jing, J., Chen, L., Doymaz, F., Pessah, I. N. & 16. Schaller, J., Kuhn-Nentwig, L., Schu¨rch, S., Ka¨mpfer, U., Mu¨ller, J. & Nentwig, Hammock, B. D. (2003) Proc. Natl. Acad. Sci. USA 100, 922–927. W. (2001) Chimia 55, 1058–1062. 35. Bindokas, V. P., Venema, V. J. & Adams, M. E. (1991) J. Neurophysiol. 66, 17. Schu¨rch, S., Schaller, J., Ka¨mpfer, U., Kuhn-Nentwig, L. & Nentwig, W. (2001) 590–601. Chimia 55, 1063–1066. 36. Herrmann, R., Moskowitz, H., Zlotkin, E. & Hammock, B. D. (1995) Toxicon 18. Norton, R. S. & Pallaghy, P. K. (1998) Toxicon 36, 1573–1583. 33, 1099–1102. 19. Santos, A. D., Imperial, J. S., Chaudhary, T., Beavis, R. C., Chait, B. T., 37. Regev, A., Rivkin, H., Inceoglu, B., Gershburg, E., Hammock, B. D., Gurevitz, Hunsperger, J. P., Olivera, B. M., Adams, M. E. & Hillyard, D. R. (1992) J. Biol. M. & Chejanovsky, N. (2003) FEBS Lett. 537, 106–110. Chem. 267, 20701–20705. 38. Adams, M. E., Herold, E. E. & Venema, V. J. (1989) J. Comp. Physiol. A 164, 20. Saccomano, N. A. & Ahlijanian, M. K. (1994) Drug Dev. Res. 33, 319–343. 333–342. 21. Bowers, C. W., Phillips, H. S., Lee, P., Jan, Y. N. & Jan, L. Y. (1987) Proc. Natl. 39. Moerman, L., Verdonck, F., Willems, J., Tytgat, J. & Bosteels, S. (2003) Acad. Sci. USA 84, 3506–3510. Biochem. Biophys. Res. Commun. 311, 90–97.

11256 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0402226101 Wullschleger et al. Wullschleger et al. 10.1073/pnas.0402226101.

Supporting Information

Files in this Data Supplement:

Supporting Text Supporting Figure 5 Supporting Figure 6

Supporting Text Isolation of CSTX-13. Briefly, 425 µl of crude venom was centrifuged (14 000 x g) for 10 min to remove insoluble proteins and cellular debris. The venom was divided into 50-µl aliquots, diluted with 200 µl of 200 mM ammonium acetate buffer, pH 5.5 (buffer A), and subsequently separated by gel filtration on a Superdex 75 HR 10/30 column (Amersham

Pharmacia). Further separation of the pooled fractions was achieved by cationic exchange chromatography on a Mono S HR 10/10 column (Amersham Pharmacia) in buffer A. The peptides were eluted with a salt gradient (2 M NaCl in buffer A, pH 5.5) as shown in Fig. 1B.

Afterward, the fractions were desalted and separated by RP-HPLC on a nucleosil 300-5 C4 column (4.6 x 250 mm; Macherey & Nagel) using 100% solvent A with a flow rate of 0.5 ml/min for 0-15 min, followed by a first 10-min gradient of 1% solvent B in A/min, a second

55-min gradient of 0.43% solvent B in A/min and a third 10-min gradient of 6.6% solvent B in A/min (solvent A, 0.1% trifluoroacetic acid in water; solvent B, 0.1% trifluoroacetic acid in acetonitrile). Further purification was achieved by RP-HPLC on a nucleosil 100-5 C8 column

(4 x 250 mm; Macherey & Nagel) using 22% solvent B in solvent A with a flow rate of 0.5 ml/min. Directly after injection of the sample, the gradient (22%-28% solvent B) was started for 20 min (Fig. 1C). This step was repeated several times to obtain a homogeneous peptide.

Reduction and Alkylation of CSTX-13. Fifty micrograms CSTX-13 was reduced with DTT

(10-fold molar excess over disulfides) in 100 µl of 0.1 M Tris-HCl buffer, pH 8.0, containing 6 M guanidine-HCl for 90 min at 37°C under nitrogen and alkylated with iodacetamide (5- fold molar excess over total thiols) for 1 h at 37°C in the dark. The reaction was stopped with

2-mercaptoethanol (10-fold molar excess over iodacetamide), and the solution acidified with triflouroacetic acid. The obtained peptides were further desalted and separated by RP-HPLC on a nucleosil 100-5 C8 column (4 x 250 mm; Macherey & Nagel) using 100% solvent A with a flow rate of 0.5 ml/min for 0-5 min followed by a 55-min gradient of 0.73% B in A/min

(Fig. 1E). Supporting Figure 5

Fig. 5. Experimental (Left) and calculated (Right) isotopic patterns of CSTX-13 chain A (4,342.80 Da) and native CSTX-13 (7,354.51 Da). Supporting Figure 6

Fig. 6. Product ion spectrum of the cystine-containing tryptic fragment of CSTX-13 (4,473.75 Da). The [M+7H]7+ ion was selected as the precursor. Triply charged fragment ions Y (m/z 464.52) and Z (m/z 430.84) define the Cys 3–Cys 8 bridge. The peak σ with m/z 443.52 corresponds to the triply charged ion formed by disulfide cleavage between the second and third peptide chain (Cys 2–Cys 5). MANUSKRIPT II

MANUSKRIPT II

Wullschleger, B., Nentwig, W. & Kuhn-Nentwig, L. Spider venom: Enhancement of venom efficacy mediated by different synergistic strategies in Cupiennius salei. Submitted.

15 Spider venom: Enhancement of venom efficacy mediated by different synergistic strategies in Cupiennius salei

Benno Wullschleger, Wolfgang Nentwig and Lucia Kuhn-Nentwig*

Zoological Institute, University of Bern, Baltzerstrasse 6, CH-3012 Bern, Switzerland

*Author for correspondence ([email protected]). Tel. 0041-31-631 4532 Fax 0041-31-631 4888 Short title: Spider venom components and their interactions MANUSKRIPT II Spider venom components and their interactions

ABSTRACT Besides the power of the chelicerae, synergistic interactions between different components in the venom of Cupiennius salei ensure the hunting success of this spider. The main components were tested alone and in combination according to their physiological venom concentrations in Drosophila bioassays. The high K+ ion content of the venom synergistically increases the insecticidal activity of the neurotoxins CSTX-1, CSTX-9 and CSTX-13 by 20% but does not influence the insecticidal effectiveness of the antimicrobially and cytolytically acting cupiennin 1a. Histamine only enhances the activity of the main neurotoxin CSTX-1. An important role in the envenomation process is exhibited by cupiennin 1a, which increases the insecticidal activity of the above-mentioned neurotoxins by up to 65%. Additionally, the highly synergistic effect of the enhancer CSTX-13 on CSTX-1, provoked in non-toxic physiological concentrations, could be verified for CSTX-9, but not for cupiennin 1a. CSTX-1 and CSTX-9 show positive interactions only when both are injected in toxic non- physiological concentrations.

Keywords: neurotoxin; Drosophila melanogaster; bioassay; synergism; multicomponent venom

17 MANUSKRIPT II Spider venom components and their interactions

1. INTRODUCTION Spider venoms are complex mixtures of components, used to paralyse prey items and defend against predators. Rapid paralysis results from the modifications of various ion channel targets by low molecular mass compounds, neurotoxic peptides and proteins. However, much less is known about the interactions of venom components within a prey organism (Adams 2004; Kuhn-Nentwig et al. 2004). The Central American spider Cupiennius salei (Keyserling 1877) uses its venom as economically as possible. The amount of venom injected varies depending on size, activity, defense behaviour and venom sensitivity of a prey item (Malli et al. 1999; Wigger et al. 2002; Wullschleger & Nentwig 2002). This optimal venom dosage is continued on the biochemical level through positive interactions among various venom components (Kuhn-Nentwig et al. 1998; Wullschleger et al. 2004). In C. salei venom, proteins with molecular masses above 10 kDa have been identified. Additionally, disulfide-rich neurotoxins, highly cationic peptides with molecular masses between 3 and 10 kDa, and low molecular mass compounds such as ions, biogenic amines, polyamines, and neurotransmitters are present (Kuhn-Nentwig et al. 2004). To date, toxicological information and sequence data for the neurotoxins CSTX-1, CSTX-9, and the neurotoxic, two-chain enhancer peptide CSTX-13 have been reported (Kuhn-Nentwig et al., 2004; Wullschleger et al. 2004). Furthermore, antimicrobially and cytolytically acting cupiennins have been identified. These highly cationic, α-helical, cysteine-free peptides may play a dual role in the venom: protection of the venomous apparatus against microbial invaders and, after venom injection into prey, an enhancement of the paralytic effect of the neurotoxins (Kuhn-Nentwig 2003). Insecticidal activities of similar cytolytically acting peptides have also been reported for the spider Lycosa carolinensis (Yan & Adams 1998) and the ant Pachycondyla goeldii (Orivel et al. 2001). Beyond these insecticidal activities, additional synergistic interactions with neurotoxins have been demonstrated for the spiders Oxyopes kitabensis (Corzo et al. 2002) and C. salei (Kuhn-Nentwig et al. 2004). Only limited information is available about possible synergistic effects of low molecular mass substances with neurotoxins immediately after venom injection into a prey item (Chan et al. 1975; Inceoglu et al. 2003; Wullschleger et al. 2004). It was previously shown that histamine and taurine facilitate the neurotoxic activity of CSTX-1 from C. salei (Kuhn-Nentwig et al. 1998). Accordingly, it was also hypothesised that µ-agatoxins, which are disulfide-bridged short peptides modifying Na+ channels, enhance the short term action of α-agatoxins (acylpolyamines). Both agatoxins have been identified in the venom of the spider Agelenopsis

18 MANUSKRIPT II Spider venom components and their interactions aperta. Furthermore, additive interactions among different ω-agatoxins, which are disulfide- bridge rich voltage activated Ca2+ channel inhibitors, have been reported for A. aperta (McDonough et al. 2002; Adams 2004). However, these findings were mainly obtained through neurophysiological investigations and bioassays of different venom components without considering their physiological ratios as they occur in the venom. In the study reported here, we analysed interactions among low molecular mass components and peptides as well as interactions between these peptides. Finally, we compared the paralytic activity of crude venom to the synergistic activity caused by defined venom components. The findings show multiple interactions of venom components in different molar ratios and help us to understand the complex nature of spider venom.

2. MATERIAL AND METHODS (a) Peptide isolation Spider maintenance, venom collection and isolation of CSTX-1, CSTX-9, CSTX-13, and cupiennin 1a were carried out as previously described (Kuhn-Nentwig et al. 1994, 2002). (b) Bioassays

The LD50 bioassays were performed according to Escoubas et al. (1995) using one to three day old Drosophila melanogaster female flies. The injected volume was 0.05 µl of 0.1 M ammonium acetate, pH 6.1 (control: 20 flies), and all further injections with different components were carried out in this solution (four concentrations, 20 flies each). Mortality rates were recorded 24 h after injection. LD50 bioassays were performed with (i) crude venom, (ii) CSTX-1, and (iii) CSTX-1 in combination with CSTX-9, CSTX-13, cupiennin 1a, histamine and KCl in their physiological venom concentrations. (c) Interactions between low molecular mass components and peptides Interactions between venom peptides and low molecular mass components were investigated in bioassays with 0.32 pmol CSTX-1/mg fly, 7.95 pmol CSTX-9/mg fly and 5.0 pmol cupiennin 1a/mg fly. We tested the peptides alone and in combination with histamine (5.7 mM), taurine (0.07 mM), and KCl (215 mM) in their physiological venom concentrations (control: 20 flies; bioassays: 15 groups with 5 flies and one repetition). (d) Interactions between peptides First, we analysed possible synergistic effects between cupiennin 1a and CSTX-1 or CSTX-9. Cupiennin 1a (0.53 pmol/mg fly) was applied in a non-toxic concentration alone as well as in combination with the CSTX-peptides in the above-mentioned (c) concentrations. In a second series of experiments, 0.32 pmol CSTX-1/mg fly alone and in combination with 0.06 pmol

19 MANUSKRIPT II Spider venom components and their interactions

CSTX-9/mg fly were injected, corresponding to their molar ratio in the crude venom (5.7:1). In the second bioassay, CSTX-1 (0.47 pmol/mg fly) and CSTX-9 (7.95 pmol/mg fly) were injected separately and in combination. In a third series of experiments bioassays were performed with 0.32 pmol CSTX-1/mg fly alone and in combination with 0.04 pmol CSTX-13/mg fly corresponding to their molar ratio in the crude venom (9:1) (repetition of Wullschleger et al. 2004). Next, the mortality of 7.95 pmol CSTX-9/mg fly was compared with the mortality of 7.95 pmol CSTX-9/mg fly combined with either 0.04 pmol CSTX-13/mg fly (molar ratio of 227:1) or 4.97 pmol CSTX- 13/mg fly (physiological molar ratio 1.6:1), respectively. Finally, the mortality of 5.0 pmol cupiennin 1a/mg fly was compared with the mortality of 5.0 pmol cupiennin 1a/mg fly combined with either 2.78 pmol CSTX-9/mg fly (physiological molar ratio 1.8:1) or 1.25 pmol CSTX-13/mg fly (physiological molar ratio 4:1), respectively. All bioassays were performed in 12 groups with 5 flies and were repeated once. Twenty flies were used as control for each bioassay. (e) Calculations and statistics

LD50 calculations were done using Proban software (Version 1.1, Jedrychowski, 1991, shareware). The relative mortality of D. melanogaster was arcsinus square root-transformed and treated as the dependent variable, whereas the venom components or the co-injected peptides were treated as nominal independent variables. The experiments were analysed using generalised linear models. The means of the nominal independent variables venom components or co-injected peptides, respectively, were compared pairwise by the Fisher LSD method. Fulfilment of model assumptions was checked by visual inspection of the residuals distribution for every statistical test conducted. Statistics were done with S-PLUS 6.0 Professional software.

3. RESULTS (a) Interactions between low molecular mass components and peptides Control injections of taurine (3.88 pmol/mg fly) or histamine (316.67 pmol/mg fly) alone, corresponding to their venom concentrations, showed no effects in the Drosophila bioassays. In contrast, KCl (11.94 nmol/mg fly) caused a short paralytic effect for 4-5 minutes, as reported previously (Wullschleger et al. 2004). The concentrations of the peptides co-injected with these low molecular mass components were chosen to produce mortalities below their

LD50 values. The injection of 0.32 pmol CSTX-1/mg fly generated a mortality rate of 27%. No significant difference in mortality increase was observed when taurine was co-injected

20 MANUSKRIPT II Spider venom components and their interactions

(36% mortality). However, co-injection of CSTX-1 with histamine (47% mortality; p < 0.01) or KCl (47% mortality; p < 0.01), significantly increased the mortality (figure 1a). Injection of 7.95 pmol CSTX-9/mg fly resulted in mortality rate of 35%. No significant increase was observed for co-injection with taurine (34% mortality) or with histamine (44% mortality). However, co-injection with KCl (54% mortality; p < 0.01) significantly increased the mortality (figure 1b). Injection of 5.0 pmol cupiennin 1a caused a mortality rate of 58%. No significant increase was observed for co-injection with taurine (57% mortality), histamine (56% mortality) or KCl (47% mortality) (figure 1d). (b) Interactions between peptides In a concentration of 0.53 pmol/mg fly, cupiennin 1a caused no insecticidal activities in the bioassay. Injection of CSTX-1 or CSTX-9 alone caused mortalities below their LD50 values. Co-injection of cupiennin 1a with CSTX-1 increased the mortality from 27% to 92% (p < 0.001) and co-injection with CSTX-9 increased the mortality from 35% to 78% (p < 0.001) (figure 2, left and middle). In a second series of experiments, a possible cooperativity between CSTX-1 and CSTX-9 was analysed first according to their molar ratio in the venom of 5.7:1 (0.32 pmol CSTX-1/mg fly : 0.06 pmol CSTX-9/mg fly). Injection of 0.06 pmol CSTX-9/mg fly had no effect on the flies. No enhanced mortality was observed between the injection of 0.32 pmol CSTX-1/mg fly alone (16% mortality) and in combination with CSTX-9 (17% mortality). Secondly, cupiennin 1a and CSTX-9 appear in the venom in a molar ratio of 1.8:1 (5.0 pmol cupiennin 1a/mg fly : 2.78 pmol CSTX-9/mg fly). Injection of CSTX-9 in this concentration is non-toxic. Furthermore, a mortality rate of 56% by injection of cupiennin 1a alone is not increased by co-injection with CSTX-9 (mortality of 61%) (figure 3a).

CSTX-1 and CSTX-9 were applied in toxic concentrations but below their LD50 values and when administered alone, 0.47 pmol CSTX-1/mg fly caused a mortality of 33% and 7.95 pmol CSTX-9/mg fly generated a mortality of 12%. Interestingly, co-injection of both toxins resulted in a mortality of 98% and differed significantly from the theoretical sum of both toxins (45%; p < 0.001) (figure 3b). In a third series of experiments possible enhancer effects of CSTX-13 on the insecticidal activity of CSTX-9 and cupiennin 1a, corresponding to their venom concentrations, were investigated. The molar ratio of CSTX-9 and CSTX-13 in the venom is 1.6:1 (7.95 pmol CSTX-9/mg fly : 4.97 pmol CSTX-13/mg fly). CSTX-13 alone was used in the above- mentioned concentration and was non-toxic. Co-injection of both enhanced the mortality caused by CSTX-9 alone from 37% to 98% (p < 0.001). Surprisingly, co-injection of CSTX-9

21 MANUSKRIPT II Spider venom components and their interactions and CSTX-13 even in a molar ratio of 227:1 (7.95 pmol CSTX-9/mg fly : 0.04 pmol CSTX- 13/mg fly) increased the mortality nearly as much (91%; p < 0.001, not shown). In contrast, co-injection of cupiennin 1a (5.0 pmol/mg fly) and CSTX-13 (1.25 pmol/mg fly) in their physiological venom ratio of 4:1 failed to significantly increase the mortality rate (50%), above that of cupiennin 1a alone (56%) (figure 4).

The main neurotoxin CSTX-1 shows an LD50 value of 0.45 pmol/mg fly (95% confidence limits: 0.40–0.55). A combined injection of CSTX-1 with CSTX-9, CSTX-13, cupiennin 1a, histamine, and KCl, corresponding to their physiological venom concentrations (Wullschleger et al. 2004), resulted in a decreased LD50 value of 0.10 pmol/mg fly (95% confidence limits:

0.09–1.06). The LD50 of crude venom amounts to 0.017 nl/mg fly (95% confidence limits: 0.016–0.021), corresponding to 0.02–0.06 pmol CSTX-1/mg fly.

4. DISCUSSION (a) Interactions between low molecular mass components and peptides The combined injections of KCl and the neurotoxins CSTX-1 or CSTX-9 resulted in an increase of approximately 20% mortality, a finding which confirms our previous results with CSTX-13 (Wullschleger et al. 2004) and indicates a general synergistic effect of potassium ions on the paralytic activity of diverse neurotoxins. The strategy of combining a high K+ concentration with specific neurotoxins in the prevenom, thus enhancing paralytic activity, was first reported for the scorpion Parabuthus transvaalicus (Inceoglu et al. 2003). It was hypothesised that the potassium equilibrium potential is locally shifted and the resulting paralytic effects are further amplified through the presence of peptide toxins, inhibiting ion channels which are responsible for the regeneration of the K+ equilibrium potential. C. salei exhibits a comparable strategy but uses a 2.7-fold higher concentration of potassium ions in its venom than P. transvaalicus to enhance its venom efficacy. This high K+ ion concentration can also provoke a nerve depolarisation thus affecting Ca2+ channels, which are in turn inhibited by CSTX-1, a known L-type calcium channel inhibitor (Mafra et al. 2001). As for cupiennin 1a, no synergistic effect of KCl on its insecticidal activity was detected. In contrast to the neurotoxins CSTX-1, CSTX-9 and CSTX-13, the cytolytically active cupiennin 1a adopts an α-helical conformation in the presence of negatively charged membranes, accumulates at the membrane surface and inserts itself into the lipid bilayer resulting in a destruction of cell membranes (Kuhn-Nentwig et al. 2002). Histamine, a neurotransmitter in

22 MANUSKRIPT II Spider venom components and their interactions insect nerve systems (Nässel 1999) and present in the spider venom, caused a significant mortality increase of 20% when co-injected with CSTX-1, but was less effective in combination with CSTX-9 or CSTX-13 in Drosophila flies (Wullschleger et al. 2004). Taurine, a neuromodulator in insects (Bicker 1991) and also present in the spider venom was totally effectless when co-injected with CSTX-1, CSTX-9 or CSTX-13. In contrast, we were earlier able to show in a blow fly bioassay that the neurotoxicity of CSTX-1 was enhanced by both taurine and histamine when injected in its physiological venom concentrations (Kuhn- Nentwig et al. 1998). This could indicate that synergistic interactions are highly species and neurotoxin specific, despite the close relationship between both fly families. (b) Interactions between peptides As shown previously, cupiennin 1a dramatically enhances the efficacy of the neurotoxins CSTX-1 and CSTX-13 although it is applied in a completely non-toxic concentration, or even

20-fold below its LD50 (Kuhn-Nentwig et al. 2004; Wullschleger et al. 2004). This synergistic effect was additionally proven for the neurotoxin CSTX-9 (figure 2). Positive insecticidal cooperativity between the cytolytically active oxyopinins and the neurotoxin oxytoxin 1 is also reported for the spider Oxyopes kitabensis (Corzo et al. 2002). There is evidence that these highly cationic peptides, found in the venom of O. kitabensis, afford diverse neurotoxins better access to their targets through their cytolytic activities. In contrast, the insecticidal activity of cupiennin 1a was definitely not enhanced when administered with the neurotoxins CSTX-9 or CSTX-13. The dramatic enhancer effect of the two-chain peptide CSTX-13 on the insecticidal activity of CSTX-1 in a concentration 440-fold below its LD50 value as reported recently (Wullschleger et al. 2004) could also be enlarged on the neurotoxin CSTX-9. Furthermore, the synergistic activity of CSTX-13 and the neurotoxins CSTX-1 and CSTX-9 is highly specific. The strong synergistic activities generated by cupiennin 1a and CSTX-13 are based on their physiological venom concentrations, which implies that both components were applied in non-toxic concentrations together with the neurotoxins CSTX-1 or CSTX-9. In contrast to these results, CSTX-9 did not enhance the neurotoxic activity of CSTX-1 when co-injected corresponding to its molar ratio in the venom. However, co-injection of CSTX-1 and CSTX- 9, both in toxic concentrations, increased the toxicity by more than 50% when compared to the theoretical toxicity sum of CSTX-1 and CSTX-9 injected alone, thus exhibiting a positive cooperativity. These findings are in agreement with other reports in which positive cooperativities between different neurotoxins were demonstrated by applying the components in a 1:1 molar ratio or in concentrations in which both components alone cause intoxications,

23 MANUSKRIPT II Spider venom components and their interactions or by using toxins from different sources (Bindokas et al. 1991; Herrmann et al. 1995; Shu & Liang 1999; Regev et al. 2003; Adams 2004).

9, CSTX-13, cupiennin 1a, histamine and KCl into Drosophila flies resulted in an LD50 value which is 4.5 fold lower than a single injection of CSTX-1. Taking this and the LD50 value obtained by injection of native venom into account, the venom components mentioned above are responsible for up to 57% of the crude venom efficacy. Obviously, still other unknown components are important in the envenomation process and cause at least 43% of the toxicity of C. salei venom. The interactions of different venom components presented here are extremely complex: histamine and taurine seem to enhance the activity of CSTX-1 highly specifically, and their effects are in part species specific. The high K+ ion concentration in the venom facilitates the neurotoxin activity, but not that of cupiennin 1a. However, this group of cytolytic peptides dramatically enhances the activity of the neurotoxins. In addition, the neurotoxins are further amplified by the two-chain enhancer CSTX-13. Differences in LD50 values obtained by injection of crude venom into various arthropods (Kuhn-Nentwig et al. 1998) lead us to assume that the interactions presented here cannot be generalised and are only validated for Drosophila melanogaster.

Acknowledgements We thank Dr. H. Murray for critical comments on the manuscript, and the Swiss National Science Foundation for funding.

24 MANUSKRIPT II Spider venom components and their interactions

REFERENCES Adams, M. E. 2004 Agatoxins: ion channel specific toxins from the American funnel web spider, Agelenopsis aperta. Toxicon 43, 509–525. (DOI 10.1016/j.toxicon.2004.02.004)

Bicker, G. 1991 Taurine-like immunoreactivity in photoreceptor cells and mushroom bodies: a comparison of the chemical architecture of insect nervous systems. Brain Res. 560, 201– 206.

Bindokas, V. P., Venema, V. J. & Adams, M. E. 1991 Differential antagonism of transmitter release by subtypes of ω-agatoxins. J. Neurophysiol. 66, 590–601.

Chan, T. K., Geren, C. R., Howell, D. E. & Odell, G. V. 1975 Adenosine triphosphate in tarantula spider venoms and its synergistic effect with the venom toxin. Toxicon 13, 61–66.

Corzo, G., Villegas, E., Gómez-Lagunas, F., Possani, L. D., Belokoneva, O. S. & Nakajima, T. 2002 Oxyopinins, large amphipathic peptides isolated from the venom of the wolf spider Oxyopes kitabensis with cytolytic properties and positive insecticidal cooperativity with spider neurotoxins. J. Biol. Chem. 277, 23627–23637. (DOI 10.1074/jbc.M200511200)

Escoubas, P., Palma, M. F. & Nakajima, T. 1995 A microinjection technique using Drosophila melanogaster for bioassay – guided isolation of neurotoxins in arthropod venoms. Toxicon 33, 1549–1555.

Herrmann, R., Moskowitz, H., Zlotkin, E. & Hammock, B. D. 1995 Positive cooperativity among insecticidal scorpion neurotoxins. Toxicon 33, 1099–1102.

Inceoglu, B., Lango, J., Jing, J., Chen, L., Doymaz, F., Pessah, I. N. & Hammock, B. D. 2003 One scorpion, two venoms: Prevenom of Parabuthus transvaalicus acts as an alternative type of venom with distinct mechanism of action. Proc. Natl. Acad. Sci. USA 100, 922–927. (DOI 10.1073/pnas.242735499)

Kuhn-Nentwig, L., Schaller, J. & Nentwig, W. 1994 Purification of toxic peptides and the amino acid sequence of CSTX-1 from the multicomponent venom of Cupiennius salei (Araneae: Ctenidae). Toxicon 32, 287–302.

25 MANUSKRIPT II Spider venom components and their interactions

Kuhn-Nentwig, L., Bücheler, A., Studer, A. & Nentwig, W. 1998 Taurine and histamine: Low molecular compounds in prey hemolymph increase the killing power of spider venom. Naturwissenschaften 85, 136–138.

Kuhn-Nentwig, L., Müller, J., Schaller, J., Walz, A., Dathe, M. & Nentwig, W. 2002 Cupiennin 1, a new family of highly basic antimicrobial peptides in the venom of the spider Cupiennius salei (Ctenidae). J. Biol. Chem. 277, 11208–11216. (DOI 10.1074/jbc.M111099200)

Kuhn-Nentwig, L. 2003 Antimicrobial and cytolytic peptides of venomous arthropods. Cell. Mol. Life Sci. 60, 2651–2668. (DOI 10.1007/s00018-003-3106-8)

Kuhn-Nentwig, L., Schaller, J. & Nentwig, W. 2004 Biochemistry, toxicology and ecology of the venom of the spider Cupiennius salei (Ctenidae). Toxicon 43, 543–553. (DOI 10.1016/j.toxicon.2004.02.009)

Mafra, R. A., Kuhn-Nentwig, L., Araújo, D. A., Beirão, P. S., & Cruz, J. S. 2001 Effect of CSTX-1, a toxin from Cupiennius salei spider, on L-type calcium currents. XVI FESBE Annual Meeting, Abstract #20.030, p. 237.

Malli, H., Kuhn-Nentwig, L., Imboden, H. & Nentwig, W. 1999 Effects of size, motility and paralysation time of prey on the quantity of venom injected by the hunting spider Cupiennius salei. J. Exp. Biol. 202, 2083–2089.

McDonough, S. I., Boland, L. M., Mintz, I. M. & Bean, B. P. 2002 Interactions among toxins that inhibit N-type and P-type calcium channels. J. Gen. Physiol. 119, 313–328.

Nässel, D. R. 1999 Histamine in the brain of insects: a review. Microsc. Res. Tech. 44, 121– 136.

Orivel, J., Redeker, V., Le Caer, J.-P., Krier, F., Revol-Junelles, A.-M., Longeon, A., Chaffotte, A., Dejean, A. & Rossier, J. 2001 Ponericins, new antibacterial and insecticidal

26 MANUSKRIPT II Spider venom components and their interactions peptides from the venom of the ant Pachycondyla goeldii. J. Biol. Chem. 276, 17823–17829. (DOI 10.1074/jbc.M100216200)

Regev, A., Rivkin, H., Inceoglu, B., Gershburg, E., Hammock, B. D., Gurevitz, M. & Chejanovsky, N. 2003 Further enhancement of baculovirus insecticidal efficacy with scorpion toxins that interact cooperatively. FEBS Lett. 537, 106–110. (DOI 10.1016/S0014- 5793(03)00104-2)

Shu, Q. & Liang, S. P. 1999 Purification and characterization of huwentoxin-II, a neurotoxic peptide from the venom of the Chinese bird spider Selenocosmia huwena. J. Peptide Res. 53, 486–491.

Wigger, E., Kuhn-Nentwig, L. & Nentwig, W. 2002 The venom optimisation hypothesis: a spider injects large venom quantities only into difficult prey types. Toxicon 40, 749–752.

Wullschleger, B. & Nentwig, W. 2002 Influence of venom availability on a spider’s prey- choice behaviour. Funct. Ecol. 16, 802–807.

Wullschleger, B., Kuhn-Nentwig, L., Tromp, J., Kämpfer, U., Schaller, J., Schürch, S. & Nentwig, W. 2004 CSTX-13, a highly synergistically acting two-chain neurotoxic enhancer in the venom of the spider Cupiennius salei (Ctenidae). Proc. Natl. Acad. Sci. USA 101, 11251– 11256. (DOI 10.1073/pnas.0402226101)

Yan, L. & Adams, M. E. 1998 Lycotoxins, antimicrobial peptides from venom of the wolf spider Lycosa carolinensis. J. Biol. Chem. 273, 2059–2066. (DOI 10.1074/jbc.273.4.2059)

27 MANUSKRIPT II Spider venom components and their interactions

FIGURES

Figure 1

100 ()a 100 ()b

75 CSTX-1 75 CSTX-9 ** ** ** 50 50

25 25

0 0 Balone +Taurine +Histamine +KCl Balone +Taurine +Histamine +KCl 100 ()c 100 ()d 75 CSTX-13 ** 75 Cupiennin 1a 50 50

25 25

0 0 Balone +Taurine +Histamine +KCl Balone +Taurine +Histamine +KCl

Fig. 1. Synergistic effects between low molecular mass components and CSTX-1, CSTX-9, CSTX-13, or cupiennin 1a as measured in a Drosophila bioassay: (a) The mortality caused by CSTX-1 alone (5.99 µM) was compared with the mortality due to co-injection with taurine (0.07 mM; n.s.), histamine (5.7 mM; ** p < 0.01), or KCl (215 mM; ** p < 0.01). (b) The mortality caused by CSTX-9 alone (143.14 µM) was compared with the mortality due to co- injection with taurine (0.07 mM; n.s.), histamine (5.7 mM; n.s.), or KCl (215 mM; ** p < 0.01). (c) The mortality caused by CSTX-13 alone (239.4 µM) was compared with the mortality due to co-injection with taurine (0.07 mM; n.s.), histamine (5.7 mM; n.s.), or KCl (215 mM; ** p < 0.01) (data from Wullschleger et al. 2004). (d) The mortality caused by cupiennin 1a alone (90 µM) was compared with the mortality due to co-injection with taurine (0.07 mM; n.s.), histamine (5.7 mM; n.s.) or KCl (215 mM; n.s.). Taurine, histamine or KCl injected in physiological venom concentrations showed no toxic effect when administered alone as controls (not shown). All data with mean ± s.e.; Fisher LSD test.

28 MANUSKRIPT II Spider venom components and their interactions

Figure 2

*** *** 100 *** 75

0 CSTX-1 CSTX-9 CSTX-13

Fig. 2. Synergistic effects between cupiennin 1a and CSTX-1, CSTX-9, or CSTX-13 as measured in a Drosophila bioassay: (Left) The mortality caused by CSTX-1 alone (5.99 µM) was compared with the mortality due to co-injection with cupiennin 1a (0.96 µM; *** p < 0.001). (Middle) The mortality caused by CSTX-9 alone (143.14 µM) was compared with the mortality due to co-injection with cupiennin 1a (0.96 µM; *** p < 0.001). (Right) the mortality of CSTX-13 alone (239.4 µM) was compared with the mortality due to co-injection with cupiennin 1a (0.96 µM; *** p < 0.001) (data from Wullschleger et al. 2004). Cupiennin 1a showed no toxic effect when administered alone as control (not shown). All data with mean ± s.e.; Fisher LSD test.

29 MANUSKRIPT II Spider venom components and their interactions

Figure 3

*** 100 ()a ()b 75 CSTX-1 Cupiennin 1a

Fig. 3. Interactions between CSTX-9 and CSTX-1, or cupiennin 1a as measured in a Drosophila bioassay: (a) The mortality caused by CSTX-1 alone (black bar; 5.99 µM) was compared with the mortality due to co-injection with CSTX-9 (1.05 µM; n.s.) corresponding to their physiological venom ratio. The mortality caused by cupiennin 1a alone (black bar; 90 µM) was compared with the mortality due to co-injection with CSTX-9 (50 µM; n.s.). CSTX- 9 showed no toxic effect when administered alone (control, not shown). (b) The mortality caused by CSTX-1 alone (white bar; 8.99 µM) or CSTX-9 alone (hatched bar; 143.14 µM), or their theoretical sum respectively, was compared with the mortality due to the co-injection of CSTX-1 (8.99 µM) and CSTX-9 (effective sum; 143.14 µM; *** p < 0.001). All data with mean ± s.e.; Fisher LSD test.

30 MANUSKRIPT II Spider venom components and their interactions

Figure 4

100 *** ***

0 CSTX-9 Cupiennin 1a CSTX-1

Fig. 4. Synergistic effects between CSTX-13 and CSTX-9, cupiennin 1a, or CSTX-1, corresponding to their molar concentrations in the venom, as measured in a Drosophila bioassay: (Left) The mortality caused by CSTX-9 alone (black bar; 143.14 µM) was compared with the mortality due to co-injection with CSTX-13 (white bar; 95.14 µM, *** p < 0.001) in a molar ratio of 1.6:1. (Middle) The mortality caused by cupiennin 1a (black bar; 90 µM) was compared with the mortality due to co-injection with CSTX-13 (white bar; 22.5 µM, n.s.) in a molar ratio of 4:1. (Right) The mortality caused by CSTX-1 (black bar; 5.99 µM) was compared with the mortality due to co-injection with CSTX-13 (white bar; 0.67 µM, *** p < 0.001) in a molar ratio of 9:1. CSTX-13 showed no toxic effect when administered alone as control (not shown). All data with mean ± s.e.; Fisher LSD test.

31 MANUSKRIPT III

MANUSKRIPT III

Wullschleger, B., Nentwig, W. & Kuhn-Nentwig, L. Different effects of neurotoxic and cytolytically acting peptides of Cupiennius salei venom to insects.

32 Different effects of neurotoxic and cytolytically acting peptides of Cupiennius salei venom to insects

Benno Wullschleger*, Wolfgang Nentwig and Lucia Kuhn-Nentwig

Zoological Institute, University of Bern, Baltzerstrasse 6, CH-3012 Bern, Switzerland

*Corresponding author. Tel. 0041-31-631 4532; fax 0041-31-631 4888

E-mail address: [email protected]

MANUSKRIPT III Effects of spider venom peptides to insects

Abstract The hunting success of the spider Cupiennius salei depends on the efficacy of its venom. Here we report on LD50 bioassays testing different neurotoxic and cytolytically acting venom peptides alone and in combinations to insects. The reactions of Drosophila melanogaster, D. melanogaster Oregon R, Protophormia terraenovae, Acheta domesticus and Blattella germanica to the main neurotoxin CSTX-1 differ with a factor of 35 in their LD50 values. In contrast to CSTX-1, LD50 values of the cytolytically acting cupiennin 1a varied only with a factor of 4. Cupiennin 1a shows additionally synergistic effects together with CSTX-1 in all tested insects (not significant in P. terraenovae). CSTX-9 has no enhancing effect on CSTX-1 whereas CSTX-13 increases the toxicity of CSTX-1 only in the flies. This could argue for a species-specific effect.

Keywords: Acheta domesticus; LD50-bioassay; Blattella germanica; venom peptides; Drosophila melanogaster; Protophormia terraenovae; spider venom; synergism.

34 MANUSKRIPT III Effects of spider venom peptides to insects

Introduction The ctenid Cupiennius salei (Keyserling, 1877) is a polyphagous predator living in the Central American rain forest (Barth & Seyfarth, 1979; Nentwig, 1986). As an ambush spider in the vegetation, C. salei does not build webs (Barth, 2002). In fact, the power of the chelicerae and over all the effectiveness of the venom is responsible for a successful overwhelming of a prey item (Nentwig, 1986). Several investigations have shown that the spider uses its venom as economically as possible. Therefore it had been assumed that venom represents an expensive resource (Malli et al., 1999; Wigger et al., 2002; Wullschleger & Nentwig, 2002). Additionally the venom storage is limited to approximately 10 µl and venom regeneration takes 16 days (Boevé et al., 1995). The crude venom of C. salei itself is a complex mixture consisting of proteins (> 10 kDa), peptides (3–9 kDa) and low molecular mass components (< 1 kDa) (Kuhn-Nentwig et al., 2004). There is evidence that synergistic interactions between the mentioned components are essential parts for the venom effectiveness. As recently published, CSTX-13, a highly synergistically acting two-chain neurotoxic enhancer peptide, increases the lethal effect of the main neurotoxic peptide CSTX- 1 and of CSTX-9 more than 50% in a Drosophila bioassay (Wullschleger et al., 2004; Wullschleger et al., submitted). The cytolytically acting peptide cupiennin 1a as well as the low molecular mass component KCl showed also synergistic effects together with CSTX-1 and CSTX-9 in our Drosophila bioassays (Wullschleger et al., submitted). But can these results be generalized to other insects? It is well known that several spider and scorpion toxins only effect selected prey animals. For instance, several latrotoxins, large protein toxins from the black widow spider (Latrodectus sp.) venom, effect selectively different classes of animals like insects, crustaceans and vertebrates (Fritz et al., 1980; Knipper et al., 1986; Ushkaryov et al., 2004). Also the toxins of different scorpion species are selectively targeted against different animals. Exponents of the scorpion α-toxin class as well as the β-toxin class include toxins that are highly active on mammals or insects (Gordon et al., 1998; Gilles et al., 2003; Leipold et al., 2004; Cohen et al., 2004). Only little is known about the cooperation of spider or scorpion venom components when applied to different species. Herrmann et al. (1995) tested scorpion toxins from two species, Androctonus australis and Leiurus quinquestriatus hebraeus, always in a 1:1 ratio, on flies, moths and mice. Whereas the positive change in potency by combinations of two different toxins in insects were comparable, no positive interactions could be found in mice. The results from co-injections of acylpolyamines, α-agatoxins and µ-agatoxins from the venom of the spider Agelenopsis aperta into house flies, showed a strong synergistic effect (Adams et al., 1989; Adams, 2004).

35 MANUSKRIPT III Effects of spider venom peptides to insects

Further additive interactions among different ω-agatoxins, disulfide-bridge rich voltage activated Ca2+ channel inhibitors, have been reported from A. aperta (Bindokas et al., 1991; Adams, 2004). These results were mainly obtained through neurophysiological investigations and bioassays of different venom components without considering their physiological ratios as they occur in the venom. Up to now, bioassays of different venom components considering their physiological ratios as they occur in the venom to different animals are not available. Our first investigations concerning interactions between different venom components of C. salei on D. melanogaster (Kuhn-Nentwig et al., 2004; Wullschleger et al., 2004; Wullschleger et al., submitted) initiated further bioassays with C. salei venom peptides on other insects. To discover differences in the sensitivity to several spider venom peptides and to check a generality for synergistic effects, LD50 bioassays were performed with D. melanogaster, an undefined strain from our breeding stock and the wild-type strain Oregon R, P. terraenovae, A. domesticus, and B. germanica. The species-unspecific effect of the cytolytically acting cupiennin 1a, the fly-specific enhancer activity of CSTX-13 as well as the neurotoxicity of CSTX-1, with LD50 value differences until 35 times between several species, explain us roughly the effect of the crude venom. Our results let assume that the reason for a multi-component venom system lies in cooperations of individual components.

Materials and Methods Chemicals. Ammonium acetate was of analytical grade and purchased from Merck. Isolation of the peptides. Spider maintenance, venom collection and isolation of CSTX-1, CSTX-9, CSTX-13, and cupiennin 1a by FPLC and HPLC methods were performed as described previously (Kuhn-Nentwig et al., 1994; Schaller et al., 2001; Kuhn-Nentwig et al., 2002; Wullschleger et al., 2004).

Insects. Four different insect species were chosen for the LD50 bioassays: Drosophila melanogaster Meigen (Diptera: Drosophilidae), an undefined strain from our breeding stock and the wild-type strain Oregon R from the Institute of Cell Biology, University of Bern, Switzerland; the blow fly Protophormia terraenovae (Robineau-Desvoidy) (Diptera: Calliphoridae) from a French supplier; Acheta domesticus (L.) (Saltatoria: Gryllidae) from Grigfarm, Switzerland, and Blattella germanica (L.) (Blattodea: Blattellidae) from Bayer, Germany.

LD50 bioassays. The LD50 bioassays were performed according to Escoubas et al. (1995) using one to three day old flies and L1 crickets or cockroaches. The injected volume was 0.05

36 MANUSKRIPT III Effects of spider venom peptides to insects

µl of 0.1 M ammonium acetate, pH 6.1 (control, 20 insects) and all further injections with different components were carried out in this solution (at least three concentrations, 20 insects each). The mortality rates were recorded 24 h after injection and LD50 calculations were done using the Proban software (Version 1.1, Jedrychowski, 1991, shareware). The LD50 bioassays were performed with the following venom components: (1) CSTX-1, (2) cupiennin 1a, (3) CSTX-1 in combination with cupiennin 1a, (4) CSTX-1 in combination with CSTX-9, and (5) CSTX-1 in combination with CSTX-13. All combinations were injected in their physiological venom concentrations (Kuhn-Nentwig et al., 1994; Schaller et al., 2001; Kuhn-Nentwig et al., 2002; Wullschleger et al., 2004).

Results

Bioassays with CSTX-1 or cupiennin 1a. The LD50 values of the neurotoxically CSTX-1 varied between 0.37 pmol/mg D. melanogaster Oregon R to 12.80 pmol /mg A. domesticus corresponding to a factor of 35. In contrast, the LD50 values of cupiennin 1a differ only slightly with a factor of 4. Thereby D. melanogaster was the most sensitive species with a

LD50 value of 4.78 pmol cupiennin 1a/mg whereas the LD50 value of P. terraenovae was ~20 pmol cupiennin 1a/mg (Table 1). Bioassays with combinations of CSTX-1, CSTX-9, CSTX-13 and cupiennin 1a. Synergistic interactions between CSTX-1 and cupiennin 1a corresponding to their physiological venom concentration (2:1) generate LD50 values between 0.19 pmol CSTX- 1/mg D. melanogaster Oregon R and 2.38 pmol/mg B. germanica. Thus, through the activity of cupiennin 1a the LD50 value of CSTX-1 was reduced up to 6 times (A. domesticus) and the ratio between the most sensitive and the most insensitive insect was 1:13. The combination of CSTX-1 and CSTX-9, applied in their physiological venom concentration

(5.7:1), caused LD50 values between 0.34 pmol CSTX-1/mg D. melanogaster Oregon R and

12.48 pmol CSTX-1/mg A. domesticus. The ratio between these two LD50 values was 1:37 and therefore corresponds to the values of CSTX-1 alone. The co-injection of CSTX-1 and CSTX-13, corresponding to their physiological venom concentration (9:1), resulted in LD50 values between 0.10 pmol CSTX-1/mg D. melanogaster Oregon R and 12.75 pmol CSTX-1/mg A. domesticus. Thus, D. melanogaster Oregon R is 128 times more sensitive to the combination of CSTX-1 and CSTX-13 than A. domesticus (Table 1).

37 MANUSKRIPT III Effects of spider venom peptides to insects

Discussion The crude venom of the spider Cupiennius salei is a very complex mixture containing more than hundred components (Kuhn-Nentwig et al., 1994). This fact prompted us to test several of these components alone or in combinations. Our here presented toxicity data of CSTX-1 alone varied from 0.37 to 12.80 pmol CSTX-1/mg insect. In consideration of the amount of CSTX-1 in the crude venom, the CSTX-1 data explain between 10% and 40% of the crude venoms toxicity (Kuhn-Nentwig et al., 1994; Kuhn-Nentwig et al., 1998; Wullschleger & Nentwig, 2002). Because reactions of tested arthropods to the spider’s crude venom differ with LD50 values from 0.01 to >20 nl/mg arthropod over three logarithmic units (Kuhn- Nentwig et al., 1998; Wullschleger & Nentwig, 2002), we conclude that several components or combinations of them act in a species-specific manner. The LD50 data of the crude venom are interesting for two reasons: (1) In the tested insects orders, we found species with similar

LD50 values (e.g. Diptera) and others which differ in their LD50 values up to a factor of 50 (e.g. Blattodea). (2) Insects sensitive to crude venom are also sensitive to CSTX-1 (e.g. dipterans) and vice versa (e.g. crickets) (Kuhn-Nentwig et al., 1998). Nevertheless, at least more than a half of the crude venom toxicity remains unexplained with our CSTX-1 data and other venom components have to be responsible for the full toxicity. In contrast to CSTX-1, the cytolytically acting peptide cupiennin 1a seems to act more in a species independent way as the similar range of the LD50 values in a similar range showed. This could argue that cupiennin 1a disturb more membranes than interacting with specific targets like ion channels. Cupiennin 1a is acting insecticidal in a concentration of about 10 pmol/mg insect. Consequently, A. domesticus or B. germanica are similar or less sensitive to the main toxin CSTX-1 than to cupiennin 1a. Because co-injection of CSTX-1 and cupiennin

1a reduced the LD50 values up to six times, especially for insects less sensitive to CSTX-1 this combination seems to be a successful strategy. The tested dipterans, have a very low CSTX-1

LD50 value and cupiennin 1a has only a minor influence of a factor between 1 to 2. Particularly for P. terraenovae the synergism between CSTX-1 and cupiennin does not work. Because P. terraenovae is around 40 times heavier than all the other tested insects we cannot exclude weight specific factors and therefore further investigations are necessary. Synergistic interactions between cytolytic peptides and neurotoxins are also well described for D. melanogaster (Kuhn-Nentwig et al., 2002; Kuhn-Nentwig et al., 2004) as well as for Spodoptera litura (Corzo et al., 2002). CSTX-9 had no synergistic effect to CSTX-1 and therefore the role of this common peptide in the envenomation process is still unclear. Nevertheless, CSTX-9 seems to be important as the

38 MANUSKRIPT III Effects of spider venom peptides to insects amount and the toxicity of this peptide indicate (Schaller et al., 2001). The synergistic effects of CSTX-13 to CSTX-1 seem to be independent of the tested Drosophila strains. Although P. terraenovae is part of the same order as D. melanogaster, we could not find the same effect. However, CSTX-13 had a significant enhancing effect of 30% to CSTX-1 in a Protophormia bioassay when we increase the amount of CSTX-13 to an unnatural molar ratio of 1:1, where CSTX-13 alone was also not toxic (data not shown). Within the dipteran group qualitative differences in the neurochemical organization of the sensory system are known (Buchner et al., 1993). These differences could perhaps explain these LD50 results. Other positive interactions are known between different spider (Chan et al., 1975; Bindokas et al., 1991; Shu & Liang, 1999) as well as scorpion (Regev et al., 2003) venom components. Herrmann et al. (1995) co-injected even different peptides from scorpion venoms in two different insect species and mice but like all the others cited investigations venom components were used in a non-physiological ratio or originated from different animal species. Our first Drosophila bioassays with co-injection of CSTX-1 with CSTX-9, CSTX-13, cupiennin 1a, KCl and histamine showed a 4.5-fold potency of the effect of CSTX-1 alone (Wullschleger et al., submitted). We present here interactions between peptides, injected in physiological venom ratios to different insect species. The results let us anticipate that the effect of the crude venom, depending on insect species, is accomplished by specificly and / or unspecificly acting peptides. Because not all venom components have the same effect to all species the large range of the LD50 values could be explained. In conclusion: a multifunctional crude venom is essential for a polyphagous predator because single venom component alone has only a low toxicity and therefore we assume a multicomponent venom could be an evolutionary answer to a wide spectrum of prey animals. But to understand better the complexity of this context more venom combinations and arthropod species have to be tested. These continuative investigations could exhibit further insights into spider-prey-interactions on the venom level.

Acknowledgements Many thanks go to Manfred Arnold, Bayer AG (Monheim, Germany) for providing cockroaches and Prof. Dr. Beat Suter, Institute of Cell Biology, University Bern (Bern, Switzerland) for D. melanogaster Oregon R. Further we thank Bernhard Merz, Muséum d'Histoire Naturelle (Genève, Switzerland) for the determination of the blow fly species. This work has been supported by the Swiss National Science Foundation.

39 MANUSKRIPT III Effects of spider venom peptides to insects

References Adams, M. E., Herold, E. E., Venema, V. J. 1989. Two classes of channel-specific toxins from funnel web spider venom. J. Comp. Physiol. [A] 164, 333–342.

Adams, M. E. 2004. Agatoxins: ion channel specific toxins from the American funnel web spider, Agelenopsis aperta. Toxicon 43, 509–525.

Barth, F. G., Seyfarth, E.-A. 1979. Cupiennius salei Keyserling (Araneae) in the highlands of central Guatemala. J. Arachnol. 7, 255–263.

Barth, F. G. 2002. A spider’s world: senses and behaviour. Springer, New York.

Bindokas, V. P., Venema, V. J., Adams, M. E. 1991. Differential antagonism of transmitter release by subtypes of omega-agatoxins. J. Neurophysiol. 66, 590–601.

Boevé, J.-L., Kuhn-Nentwig L., Keller, S., Nentwig, W. 1995. Quantity and quality of venom released by a spider (Cupiennius salei, Ctenidae). Toxicon 33, 1347–1357.

Buchner, E., Buchner, S., Burg, M. G., Hofbauer, A., Pak, W. L., Pollack, I. 1993. Histamine is a major mechanosensory neurotransmitter candidate in Drosophila melanogaster. Cell Tissue Res. 273, 119–125.

Chan, T. K., Geren, C. R., Howell, D. E., Odell, G. V., 1975. Adenosine triphosphate in tarantula spider venoms and its synergistic effect with the venom toxin. Toxicon 13, 61–66.

Cohen, L., Karbat, I., Gilles, N., Froy, O., Corzo, G., Angelovici, R., Gordon, D., Gurevitz, M. 2004. Dissection of the functional surface of an anti-insect excitatory toxin illuminates a putative "hot spot" common to all scorpion beta-toxins affecting Na+ channels. J. Biol. Chem. 279, 8206–8211.

Corzo, G., Villegas, E., Gomez-Lagunas, F., Possani, L. D., Belokoneva, O. S., Nakajima, T. 2002. Oxyopinins, large amphipathic peptides isolated from the venom of the wolf spider Oxyopes kitabensis with cytolytic properties and positive insecticidal cooperativity with spider neurotoxins. J. Biol. Chem. 277, 23627–23637.

40 MANUSKRIPT III Effects of spider venom peptides to insects

Escoubas, P., Palma, M. F., Nakajima, P. 1995. A microinjection technique using Drosophila melanogaster for bioassay – guided isolation of neurotoxins in arthropod venom. Toxicon 33, 1549–1555.

Fritz, L. C., Tzeng, M.-C., Mauro, A. 1980. Different components of black widow spider venom mediate transmitter release at vertebrate and lobster neuromuscular junctions. Nature 283, 486–487.

Gilles, N., Gurevitz, M., Gordon, D. 2003. Allosteric interactions among pyrethroid, brevetoxin, and scorpion toxin receptors on insect sodium channels raise an alternative approach for insect control. FEBS Lett. 540, 81–85.

Gordon, D., Savarin, P., Gurevitz, M., Zinn-Justin, S. 1998. Functional anatomy of scorpion toxins affecting sodium channels. J. Toxicol. 17, 131–159.

Herrmann, R., Moskowitz, H., Zlotkin, E. Hammock, B. D. 1995. Positive cooperativity among insecticidal scorpion neurotoxins. Toxicon 33, 1099–1102.

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 (α-latrotoxin), insect synaptosomes by multiple components. Neurosci. 19, 55–62.

Kuhn-Nentwig, L., Schaller, J., Nentwig, W. 1994. Purification of toxic peptides and the amino acid sequence of CSTX-1 from the multicomponent venom of Cupiennius salei (Araneae: Ctenidae). Toxicon 32, 287–302.

Kuhn-Nentwig, L., Bücheler, A., Studer, A., Nentwig, W. 1998. Taurine and Histamine: Low molecular compounds in prey hemolymph increase the killing power of spider venom. Naturwissenschaften 85, 136–138.

Kuhn-Nentwig, L., Müller, J., Schaller, J., Walz, A., Dathe, M., Nentwig, W. 2002. Cupiennin 1, a new family of highly basic antimicrobial peptides in the venom of the spider Cupiennius salei (Ctenidae). J. Biol. Chem. 277, 11208–11216.

41 MANUSKRIPT III Effects of spider venom peptides to insects

Kuhn-Nentwig, L., Schaller, J., Nentwig, W. 2004. Biochemistry, toxicology and ecology of the venom of the spider Cupiennius salei (Ctenidae). Toxicon 43, 543–553.

Leipold, E., Lu, S., Gordon, D., Hansel, A., Heinemann, S. H. 2004. Combinatorial interaction of scorpion toxins Lqh-2, Lqh-3, and LqhαIT with sodium channel receptor sites- 3. Mol. Pharmacol. 65, 685–691.

Malli, H., Kuhn-Nentwig, L., Imboden, H., Nentwig, W. 1999. Effects of size, motility and paralysation time of prey on the quantity of venom injected by the hunting spider Cupiennius salei. J. Exp. Biol. 202, 2083–2089.

Nentwig, W. 1986. Non-webbuilding spiders: prey specialists or generalists? Oecologia 69, 571–576.

Regev, A., Rivkin, H., Inceoglu, B., Gershburg, E., Hammock, B. D., Gurevitz, M., Chejanovsky, N. 2003. Further enhancement of baculovirus insecticidal efficacy with scorpion toxins that interact cooperatively. FEBS Lett. 537, 106–110.

Schaller, J., Kämpfer, U., Schürch, S., Kuhn-Nentwig, L., Haeberli, S., Nentwig, W. 2001. CSTX-9, a toxic peptide from the spider Cupiennius salei: amino acid sequence, disulphide bridge pattern and comparison with other spider toxins containing the cystine knot structure. Cell. Mol. Life Sci. 58, 1538–1545.

Shu, Q., Liang, S. P. 1999. Purification and characterization of huwentoxin-II, a neurotoxic peptide from the venom of the Chinese bird spider Selenocosmia huwena. J. Peptide Res. 53, 486–491.

Ushkaryov, Y. A., Volynski, K. E., Ashton, A. C. 2004. The multiple actions of black widow spider toxins and their selective use in neurosecretion studies. Toxicon 43, 527–542.

Wigger, E., Kuhn-Nentwig, L., Nentwig, W. 2002. The venom optimisation hypothesis: A spider injects large venom quantities only into difficult prey types. Toxicon 40, 749–752.

42 MANUSKRIPT III Effects of spider venom peptides to insects

Wullschleger, B., Nentwig, W. 2002. Influence of venom availability on a spider’s prey- choice behaviour. Funct. Ecol. 16, 802–807.

Wullschleger, B., Kuhn-Nentwig, L., Tromp, J., Kämpfer, U., Schaller, J., Schürch, S., Nentwig, W. 2004. CSTX-13, a highly synergistically acting two-chain neurotoxic enhancer in the venom of the spider Cupiennius salei (Ctenidae). Proc. Natl. Acad. Sci. USA 101, 11251–11256.

Wullschleger, B., Nentwig, W., Kuhn-Nentwig, L. Spider venom: Enhancement of venom efficacy mediated by different synergistic strategies in Cupiennius salei. Submitted.

43 MANUSKRIPT III Effects of spider venom peptides to insects

Table 1. Different amounts of peptides alone or in combinations were dissolved in 0.1 M ammonium acetate, at pH 6.1 and 0.05 µl was injected into the insects. Lethal doses (LD50) and 95% confidence limits (Cl) were estimated by bioassays, where 50% of the test insects died of intoxication 24h after injection.

Table 1. Insecticidal activity (pmol/mg insect) of neurotoxins (CSTX-1, CSTX-9, CSTX-13) and the cytolytic peptide (Cupiennin 1a) from the venom of C. salei on different insects.

CSTX-1 Cupiennin 1a CSTX-1 CSTX-1 CSTX-1

+ Cupiennin 1a + CSTX-9 + CSTX-13

(molar ratio 2:1) (molar ratio 5.7:1) (molar ratio 9:1)

Insect LD50 95% Cl LD50 95% CI LD50 95% CI LD50 95% CI LD50 95% CI

Drosophila melanogaster 0.58 (0.54 – 0.63) 4.78 4.24 – 5.30 0.40 0.34 – 0.46 0.61 0.51 – 0.71 0.15 0.10 – 0.19

(Diptera)

Drosophila melanogaster 0.37 0.30 – 0.51 6.19 4.62 – 7.01 0.19 0.16 – 0.21 0.34 0.30 – 0.38 0.10 0.10 – 0.11

Oregon R (Diptera)

Protophormia terraenovae 0.75 0.66 – 0.86 ~20 0.66 0.63 – 0.70 0.77 0.71 – 0.85 0.60 0.50 – 0.70

(Diptera)

Acheta domesticus 12.80 9.63 – 18.02 5.98 4.50 – 7.33 2.15 1.59 – 3.27 12.48 10.19 – 15.06 12.75 10.96 – 15.06

(Saltatoria)

Blattella germanica 8.52 6.71 – 10.21 10.20 7.74 – 16.23 2.38 1.98 – 2.74 10.34 8.58 – 12.73 10.04 8.19 – 12.50

(Blattodea)

ZUSAMMENFASSUNG

ZUSAMMENFASSUNG Seit 1960, als Cupiennius salei mit einer Fracht Bananen den Weg in die Münchener Grossmarkthalle fand und von Mechthild Melchers erforscht wurde (MELCHERS, 1963 a; MELCHERS, 1963 b; MELCHERS, 1967), gilt C. salei als die am besten untersuchte Spinne. Die beträchtliche Grösse, die Asaisonalität, die hohe Reproduktion und die einfache Haltung waren entscheidende Faktoren, um diese Spinne zu einem erfolgreichen Forschungsobjekt zu machen. Da C. salei mit ca. 10 µl zudem über ein grosses Giftvolumen verfügt (MALLI et al., 1993), hat sich Forschung zum Gift dieser Spinne in den letzten 15 Jahren ebenfalls etabliert (KUHN-NENTWIG et al., 2004).

Nachdem bereits die Toxine CSTX-1 und CSTX-9 sequenziert wurden (KUHN-NENTWIG et al., 1994; SCHALLER et al., 2001), wurde mit CSTX-13 ein weiteres Toxin sequenziert. Dieses zweikettige Peptid ist mit keinem anderen bis heute bekannten Spinnentoxin vergleichbar. Leider stellte sich CSTX-13 in Drosophila-Bioassays als äusserst ungiftig heraus, was zwangsläufig eine Erklärung für das Vorkommen im Rohgift verlangte. Mit unseren Bioassays liess sich zeigen, dass KCl und Cupiennin 1a die Toxizität von CSTX-13 verstärkten und dieses wiederum die Wirkung des Haupttoxins CSTX-1 um 60% zu steigern vermochte. Der Synergismus zwischen CSTX-1 und CSTX-13 ist insofern erstaunlich, als

CSTX-13 diese Wirkungssteigerung mit einer im Vergleich zum LD50 Wert 440fach geringeren Konzentration hervorruft (Manuskript I).

Drosophila-Bioassays mit CSTX-9 ergaben in Kombination mit CSTX-13, in denen ebenfalls die im Rohgift molaren Verhältnisse berücksichtigt wurden, eine Zunahme der CSTX-9- Mortalität um 50%. Das zytolytisch wirkende Peptid Cupiennin 1a steigerte die Wirkung von CSTX-1, CSTX-9 und von CSTX-13 um bis zu 60%. Die Wirkung von Cupiennin 1a liess sich seinerseits weder mit einer Co-Injektion von CSTX-Peptiden noch mit den niedermolekularen Substanzen Taurin, Histamin und KCl steigern. Hingegen konnte KCl die Wirkung aller CSTX-Peptide um ca. 20% verstärken, währenddessen Histamin nur mit CSTX-1 einen synergistischen Effekt zeigte. Eine Kombination all dieser Komponenten erklärt 57% der Rohgift-Toxizität (Manuskript II).

Da alle hier erwähnten Erkenntnisse nur auf Drosophila melanogaster basierten, wurden in einer nächsten Phase die Bioassays auf weitere Insektenarten ausgedehnt. Neben unserem undefinierten D. melanogaster Stamm wurden D. melanogaster (Oregon R), Protophormia

46 ZUSAMMENFASSUNG terraenovae, Acheta domesticus und Blattella germanica eingesetzt. Es zeigte sich, dass die Fliegen bis 35 mal empfindlicher auf CSTX-1 reagierten als die beiden hemimetabolen Insekten. Zusätzlich scheint der Synergismus zwischen CSTX-1 und CSTX-13 nur bei den getesteten Fliegen zu funktionieren. Im Gegensatz zu CSTX-1 liegen die LD50 Werte von Cupiennin 1a nahe beieinander und unterscheiden sich maximal um das Vierfache. Cupiennin 1a kann die Toxizität von CSTX-1 um bis das Sechsfache verstärken. Insbesondere bei den für CSTX-1 weniger empfindlichen Arten A. domesticus und B. germanica führt das zytolytische Peptid zu einer bedeutenden Wirkungssteigerung. Hingegen konnte CSTX-9 in Kombination mit CSTX-1 bei keiner getesteten Insektenart eine höhere Mortalität erzielen (Manuskript III).

Die Resultate meiner Dissertation bieten einen Ansatz, das Multi-Komponenten-Gift der Spinne C. salei zu erklären. Dabei scheint die Hauptwirkung des Rohgifts durch das am häufigsten vorkommende Toxin CSTX-1 bestimmt zu werden. Da CSTX-1 allerdings deutlich weniger als die Hälfte der Toxizität ausmacht, müssen weitere Giftbestandteile involviert sein. Dabei konnte gezeigt werden, dass sowohl niedermolekulare Substanzen als auch Peptide je nach Insektenart unterschiedlich gut synergistisch interagieren. Um allerdings ein weitreichenderes Bild zu erhalten, müssten zusätzliche Giftbestandteile an weiteren Arthropoden getestet werden. Diese neuen Erkenntnisse könnten helfen, die Komplexizität des Giftes mit seinen je nach Testorganismus sehr unterschiedlichen LD50 Werten auf verschiedenen Interaktionsebenen besser zu verstehen.

47 LITERATUR

LITERATUR the common mealbeetle. Toxicon 27, 305–316.

Habermehl, G. (1976) Gift-Tiere und ihre Barth, F. G. & Seyfarth, E.-A. (1979) Waffen. Springer. Berlin. Cupiennius salei Keyserling Haupt, J. (2003) The Mesothelae–a (Araneae) in the highlands of central monograph of an exceptional group Guatemala. J. Arachnol. 7, 255–263. of spiders (Araneae: Mesothelae). Barth, F. G. (2001) Sinne und Verhalten: Zoologica 154. aus dem Leben einer Spinne. Keyserling, E. (1877) Über amerikanische Springer. Berlin. Spinnenarten der Unterordnung Bode, F., Sachs, F. & Franz, M. R. (2001) Citigradae. Verh. zool.-bot. Ges. Tarantula peptide inhibits atrial Wien 26, 609–708. fibrillation. Nature 409, 35–36. King, G. F., Tedford, H. W. & Maggio, F. Boevé, J.-L., Kuhn-Nentwig L., Keller, S. (2002) Structure and function of & Nentwig, W. (1995) Quantity and insecticidal neurotoxins from quality of venom released by a spider Australian funnel-web spiders. J. (Cupiennius salei, Ctenidae). Toxicol. Toxin Rev. 21, 359–389. Toxicon 33, 1347–1357. Kuhn-Nentwig, L., Schaller, J. & Nentwig, Escoubas, P., Diochot, S. & Corzo, G. W. (1994) Purification of toxic (2000) Structure and pharmacology peptides and the amino acid sequence of spider venom neurotoxins. Bio- of CSTX-1 from the multicomponent chimie 82, 893–907. venom of Cupiennius salei (Araneae: Fisher, M. M., Raftos, J., McGuiness, R. Ctenidae). Toxicon 32, 287–302. T., Dicks, I. T., Wong, J. S., Burgess, Kuhn-Nentwig, L., Bücheler, A., Studer, K. R. & Sutherland, S. K. (1981) A. & Nentwig, W. (1998) Taurine Funnel-web spider (Atrax robustus) and histamine: low molecular antivenom. 2. Early clinical compounds in prey hemolymph experience. Med. J. Aust. 2, 525– increase the killing power of spider 526. venom. Naturwissenschaften 85, Foelix, R. F. (1996) Biology of Spiders. 136–138. Oxford University Press. New York. Kuhn-Nentwig, L., Schaller, J. & Nentwig, Friedel, T. & Nentwig, W. (1989) W. (2004) Biochemistry, toxicology Immobilizing and lethal effects of and ecology of the venom of the spider venoms on the cockroach and

LITERATUR

spider Cupiennius salei (Ctenidae). (Ctenidae). Z. Morph. Ökol. Tiere 58, Toxicon 43, 543–553. 321–346. Lee, S.-Y. & MacKinnon, R. (2004) A Nentwig, W. (1986) Non-webbuilding membrane-access mechanism of ion spiders: prey specialists or channel inhibition by voltage sensor generalists? Oecologia 69, 571–576. toxins from spider venom. Nature Platnick, N. I. (2004) The World Spider 430, 232–235. Catalog, Version 5.0. Malli, H., Vapenik, Z. & Nentwig, W. http://research.amnh.org/entomology/ (1993) Ontogenetic changes in the spiders/catalog/COUNTS.html. toxicity of the venom of the spider Saccomano, N. A. & Ahlijanian, M. K. Cupiennius salei (Araneae: (1994) Ca2+ channel toxins: tools to Ctenidae). Zool. J. Physiol. 97, 113– study channel structure and function. 122. Drug Develop. Res. 33, 319–343. Malli, H., Kuhn-Nentwig, L., Imboden, H. Schaller, J., Kämpfer, U., Schürch, S., & Nentwig, W. (1999) Effects of Kuhn-Nentwig, L., Haeberli, S. & size, motility and paralysation time of Nentwig, W. (2001) CSTX-9, a toxic prey on the quantity of venom peptide from the spider Cupiennius injected by the hunting spider salei: amino acid sequence, Cupiennius salei. J. Exp. Biol. 202, disulphide bridge pattern and 2083–2089. comparison with other spider toxins Maretić, Z. (1987) Spider venoms and their containing the cystine knot structure. effect. In: Ecophysiology of spiders. Cell. Mol. Life Sci. 58, 1538–1545. Ed. Nentwig, W. Springer Verlag. Suchyna, T. M., Tape, S. E., Koeppe II, R. Berlin. E., Andersen, O. S., Sachs, F. & Melchers, M. (1963 a) Cupiennius salei Gottlieb, P. A. (2004) Bilayer- (Ctenidae) – Kokonbau und Eiablage. dependent inhibition of mechano- Encyclop. Cinematogr., Film E 363. sensitive channels by neuroactive Melchers, M. (1963 b) Zur Biologie und peptide enantiomers. Nature 430, zum Verhalten von Cupiennius salei 235–240. Keyserling (Ctenidae), einer amerika- Sutherland, S. K. (1983) Australian animal nischen Ctenide. Zool. Jb. Syst. 91, toxins. Oxford Univ. Press. 1–90. Melbourne. Melchers, M. (1967) Der Beutefang von Swartz, K. J. & MacKinnon, R. (1997) Cupiennius salei Keyserling Mapping the receptor site for

LITERATUR

hanatoxin, a gating modifier of venom quantities only into difficult voltage-dependent K+ channels. prey types. Toxicon 40, 749–752. Neuron 18, 675–682. Wunderlich, J. (1986) Spinnenfauna Tedford, H. W., Sollod, B. L., Maggio, F. gestern und heute: Fossile Spinnen in & King, G. F. (2004) Australian Bernstein und ihre heute lebenden funnel-web spiders: master Verwandten. Verlag Jörg Wunder- insecticide chemists. Toxicon 43, lich. Straubenhardt. 601–618. Wullschleger, B. & Nentwig, W. (2002) Wigger, E., Kuhn-Nentwig, L. & Nentwig, Influence of venom availability on a W. (2002) The venom optimisation spider’s prey-choice behaviour. hypothesis: a spider injects large Funct. Ecol. 16, 802–807.

DANKSAGUNG

Ich bedanke mich recht herzlich bei Prof. Dr. Wolfgang Nentwig für die Betreuung und die Möglichkeit, mich während einer Dissertation mit „meinen“ Spinnen weiter wissenschaftlich auseinandersetzen zu können.

Dr. Lucia Kuhn-Nentwig war mir mit Ihrem Engagement stets Vorbild und hat mit vielen wertvollen Ideen und kritischer Auseinandersetzung massgeblichen Anteil an der hier vorliegenden Dissertation. Ich konnte sehr viel von Dir lernen. Danke!

Prof. Dr. Johann Schaller, Urs Kämpfer, Adrian Schindler, Dr. Stefan Schürch und Jan Tromp für unzählige Massenspektrometrie- und Aminosäurenanalysen. Herzlichen Dank für den stets freundschaftlichen Kontakt und die wertvolle Zusammenarbeit.

Dr. Patrik Kehrli, der für jedes statistische Problem einen Ansatz wusste und mir als hilfsbereiter Bürokollege in Erinnerung bleibt.

Ein herzliches Dankeschön meinem Laborkollegen Sathyan Chandru.

Der Abteilung Synökologie für ein angenehmes und hilfreiches Umfeld.

Monika, meinem Schatz, die sich stets für meine Arbeit interessiert und bei der Spinnenzucht mitgeholfen hat. Merci, dass Du meiner Spinnenleidenschaft so viel Verständnis entgegenbringst.

Meinen Eltern, die mich während der ganzen Zeit unterstützt haben und mir mit all den 8- beinern ein Zuhause geboten haben, um schliesslich auch vom „Spinnenvirus“ infiziert zu werden.

Meinem Bruder Jürg ebenfalls ein herzliches Dankeschön.

PD Dr. Rainer F. Foelix und Bruno Erb, mit denen ich einen Einblick in Anatomie und Mechanik von Spinnencheliceren gewonnen habe und somit eine willkommene Abwechslung zu meiner Forschung hatte. Vielen Dank!

Allen Freunden und Bekannten/Verwandten, die mich in irgend einer Form unterstützt haben.

Dem Schweizerischen Nationalfond, der dieses Projekt finanziell unterstützt hat.

51 CURRICULUM VITAE

CURRICULUM VITAE

Name: Benno Wullschleger Geburtsdatum: 30.08.1977 Heimatort: Zofingen (AG)

Ausbildung

1984–1989 Primarschule in Rothrist 1989–1993 Bezirksschule in Rothrist 1993–1997 Psychologisch-Soziales Gymnasium an der neuen Kantonsschule in Aarau 1997–2000 Biologiestudium an der Universität Bern 2000–2001 Diplom: Influence of venom availability on the prey choice behaviour of a spider. Zoologisches Institut der Universität Bern, Abteilung Synökologie. Leiter: Prof. Dr. Wolfgang Nentwig 2002–2004 Doktorand am Zoologischen Institut der Universität Bern, Abteilung Synökologie. Leiter: Prof. Dr. Wolfgang Nentwig

Sonstiges

2000– Fachlehrer Biologie an der Bezirksschule in Aarau 2004–2005 AHL, Universität Bern