Toxiear Vol . 29, No. 10, pP . IY1~1233, 1991 . Printed in Grat Britain .

0041-0101/91 SS .00+00 ® 1991 Papmon Prw pk

ISOLATION OF NEUROTOXIC PEPTIDES FROM THE VENOM OF THE `ARMED' SPIDER PHONEUTRIA NIGRIVENTER LEONIDFS REZENDE JR,I MARTA N. CORDEQtO,Z EDUARDO B. OLIVEIRAS and CARLOS R. DINIZ'~Z" 'Departamento de Bioquimica e ImunoloBa, Instituto de Cü'tncias Biol6~cas, Universidade Federal de Minas Germs, Helo Horizonte, M .G., Brazil ; ~Centro de Pesquisa e Desenvolvimento, Fundaçi[o Ezequiel Dias, CP. 026, 30550 Belo Horizonte, M .G., Brazil ; and'Departamento de Bioquimica, Faculdade de Medicine de Ribeitäo Preto, Ribeirâo Preto, Silo Paulo, Brazil

(Received 19 December 1990 ; accepted 8 May 1991) L. REZENDE JR, M. N. CORDEIRO, E. B. OLIVEIRA and C. R . DII~TZ . Isolation of neurotoxic peptides from the venom of the `armed' spider Phoneutria nigriventer . Toxicon 29, 1225-1233, 1991 .-Three neurotozic fractions, lethal to mice, were isolated from the venom of the spider Phoneutria nigriventer, by gel filtration and reverse phase chromatography (Phoneutria toxins 1, 2 and 3) . These toxins have mol. wts in the range 6000-9000, and have different amino acid compositions and N-terminal amino acid sequences. The toxins also differ in the lethality and signs they cause in mice after intracerebro-ventricular injection. The median ~so being respectively for the whole venom, toxins 1, 2 and 3, 47 f 5 fig, 45 f4 iug, 1 .7 f 0.7 hg and 137 f 10 pgfkg mouse. Toxins 1 and 2 induce excitatory symptoms in mice and toxin 3 a flaccid paralysis with an EDso of 40 f 5 icg,/kg mouse as measured also by intracerebro-ventricular injection. The presence in the venom of a non-neurotoxic, smooth muscle active peptide is also described. INTRODUCTION

SPIDERS from the genus Phoneutria (PERTY, 1833), family Ctenidae, are widely distributed in the warm regions of South America. Several species have been described but P. nigriventer (KEYSERLING, 1891) is the most common one found in central and south Brazil (voN EICKSTEDT, 1978/79) . It is also the species most frequently involved in human accidents (LUCAS, 1988). Envenomation by P. nigriventer was first described by BRAZIL and VELLARD (1925) . On the basis of symptoms observed in human patients and the results of subcutaneous injection into experimental animals, they concluded that the venom of Phoneutria is mainly neluotoxic . BARRIO (1955) showed that the contraction of rat diaphragm induced by Phoneutria venom was due to the release of acetylcholine by a basic component that could be separated by paper electrophoresis. Two polypeptide fractions stimulating contraction of guinea pig ileum as well as L-histamine and 5-hydroxytryptamine were isolated by DINIZ (1963) by paper chromatography . ScI~N'Author to whom correspondence should be addrea4ed at : Centro de Pesquisas e Desenvolvimento, Fundaç~o Ezequiel Dias, C .P. 26 30550-Helo Horizonte, MG Brazil . 1225

1226

L. REZENDE et al.

and PERE)RA LnKw (1971) showed that several pharmacologically active polypeptides covering the apparent mol. wt range of 5000-7000 could be separated from P. nigriventer venom by gel filtration and ion-exchange chromatography. ENTWISTLE et al. (1982) isolated a number of peptides neurotoxic to insects including a neutral polypeptide of mol. wt 5200 that produced marked stimulatory effects on the neuromuscular preparation from the locust tibial muscle. More recently, FONTANA and VITAL BRAZIL (1985) observed that the whole P. nigriventer venom caused a depolarization of the rat diaphragm muscle fiber membrane and increased the frequence of miniature end plate potentials . These effects were abolished by tetrodotoxin. In the present work we describe a rapid procedure involving a combination of gel filtration and reverse phase fast protein liquid chromatography that was used to separate peptides which are neurotoxic in mice. The presence in the venom of substances which are active upon smooth muscle was also confirmed . A preliminary account of some of these results has been given (CoRD>=nto et al., 1988). B>?ItG

MATERIALS AND METHODS Venom and other materials

Venom from 200 adult live female P. nigriventer spiders was obtained by electrical stimulation of the fangs as described by BAwuo and VrrAt. Btuza (1949). The spiders were collected in the region of Santa Barbara, State of Minas Gerais . They were kept at the arachnidarium of Fundaçâo Ezequiel Dias (Bolo Horizonte, Brazil), fed with water and given insects ad~libitum except in the two weeks proceeding the venom extraction (once a month) . The venom (5-12 pl/spider, l60 mg protein/ml) was aspirated in a Pasteur pipet, transferred immediately to glass tubes in ice, diluted with an equal volume of distilled water and centrifuged at 4000a to remove insoluble proteins and cellular debris . If not used immediately the venom was lyophilized and stored at -20°C in a desiocator. Siliconizod glassware or nalgene material was used throughout the work to avoid loss of activity due to adsorption of toxins on the walls. Sephadex G-SO superfine was purchased from Pharmacia Fine Chemicals (Uppsala, Sweden). All of the chemicals and solvents were pro-analysis of HPLC grade purchased from Sigma Chemicals Co . (St. Louis, U.S.A .), Merck, (Darmstadt F.R.G .) and Aldrich Chemical Co . (Milwaukee, U.S .A.) . Fast liquid chromatography was performed in the Pharmacia (FPLC) and Varian (HPLC) systems.

Pur~cation and fractionation procedwes

Pooled samples of centrifuged venom (0 .8-3 .0 ml) or 15-35.0 mg of lyophilized venom dissolved in 0.5 ml of 0.15 M ammonium formate (pH 6.3) were subjected to conventional gel filtration on a column (1 .5 x 200 cm) of Sephadex G-50 Superfine equilibrated and eluted (5 .0 ml/tube) with 0.15 M formate buffer (pH 6.3) at 4°C. The neurotoxic and smooth muscle active components were recovered in a broad peak, P4, M, 6000-20,000 (Fig. la), lyophilized and kept at -20°C in a desiccator. For further purification samples of 10-20 mg of the lyophilized toxic pool (P Fig. la) were dissolved in 0.20 ml of 0.15 M ammonium formate buffer at pH 6.3 and then gel filtered on a column of Superase 12 hr (10/30, FPLC, Pharmacia) equilibrated and eluted with the same buffer at room temperature (23°C). The fractions eluted (0.8 ml/tube) were pooled as indicated by horizontal bars in Fig. Ib and assayed for neurotoxic and smooth muscle activities .

Reverse phase liquid chromatography

The lyophilized toxic and smooth muscle active fractions S, and SZ (Fig. lb) were subjected to liquid chromatography on an analytical peptide reverse phase chromatography column (C~/Ci~ (PEP RPC HR 5/5, Phannacia) in a FPLC system. The mobile phase was a mixture of solvent A (0 . l% of trifluoroacetic acid-TFA) and solvent B (acetonitrile) at 20°C. Elution of fraction Sr was performed with a linear gradient from 0~5% H in A (Fig. Ic) and fraction S, was eluted with a stepwise gradient from 0 to 30% of B in A (Fig. ld). The fractions eluted were pooled, dried under nitrogen and lyophilized for bioassay and further purification. The toxic and muscle active fractions were then rechromatographod on a reverse phase liquid chromatography rnlumn C,/C, (Pro RPC 5/10, FPLC Pharmacia) using a linear gradient from 0 to 50% B in A (Fig. 2 ate). For HPLC chromatography a Vydac C column 4.6 mm x 25 cm (218 TP 54, Technicol, Stockport, U.K.) was used (Fig . 2b, inset).

Neurotoxins from Spider Venom

122 7

E c O

m N H Q W U Z Q m

O

N m

EFFLUENT

VOLUME

(ml)

FIG. I . ELUTION PROFILES OF BIOLOGICAL ACTIVITIE4 ON GEL FILTRATION AND REVERSE PHASE LIQUID clntostATOGRAPHV of THE P. nigriventer vENO~t . (a) Sephadex G 50 gel filtration of P. nigriventer venom . A sample (150 mg) of crude lyophilized venom was dissolved in 1 ml of 0 .15 M ammonium formate buffer (pH 6.3) and applied to the column (1 .5 x 200 cm) at 4°C and eluted with the same buffer . The volume of the fractions collected was 5 ml . Fraction P, was lethal to mice, contracted the guinea pig ileum, and was inactivated by chymotrypsin . (b) Superase 12 HR gel filtration of 6 mg of lyophilized P, fraction obtained by Sephadex G-50 chromatography (Fig. la) . The sample was dissolved in 0.2 ml of 0.15 M ammonium formats buffer (pH 6 .3) and applied to a l .0 x 30 .0 cm column equilibrated and eluted at room temperature (23°C) with the same buffer (volume fraction 0 .8 ml). Fractions S, and S2 were toxic to mice and contracted the guinea pig ileum . ~ was inactive . (c) Reverse phase chromatography (FPLC) of fraction S Z (Fig . Ib) on a C,/C  column (PEP RPC 0.5 x 5 cm, Pharmacia). S Z (1 mg) was dissolved in I .0 ml of 0 .1 % triftuoracetic solution in water and the column was eluted using a linear gradient 0-50% of acetonitrile in 0 .1 % TFA . Fractions T, and TZ were lethal to mice . (d) Reverse phase FPLC of fraction S, (I .0 mg) from the Superase l2 HR column (Fig. ic) on a C7/C, a column (PEP RPC 0 .5 x 0 .5 cm) using a stepwise gradient 0-34% of acetonitrile in 0.1 TFA . Fraction T3 induced flaccid paralysis in mice. Fraction M contracted the guinea pig ileum . This activity was inactivated by chymotrypsin .

Bioassays Toxicity. The toxicities of whole venom and column eluates were assayed by intracerebro-ventricular injection (i .c .v .) of 30 pl samples of the lyophilized material dissolved in O.15 M saline containing l mg/ml of bovine serum albumin (BSA) into I8-20 g albino mice (HALEY and McCatstlclc, 1957) . Appearance of neurotoxic signs such as excitation, salivation, lachrymation, priapism, spastic or flaccid paralysis, scratching and tail elevation were observed. The median lethal dose 50% (~xi) was determined in 12 mice at each of five dosage levels and observed during a 4 hr period following i .c .v . injection. The calculation was done by the probit method (FtNNEY, 1952) using an Itautec computer. Fractions that produced flaccid paralysis were assayed by the observation of mice for 4 hr after i .c .v . injection of the eluates and determination of the effective dose that produced a flaccid paralysis in 50°/. of the animals (®~ . Smooth ttnrtck assay. Guinea pigs were killed by cervical dislocation followed by exsanguination . Pieces of the guinea pig ileum were suspended in 10 ml Tyrode solution at 35°C and contraction registered with a frontal isotonic lever in a smoked drum . Samples of the column fractions were assayed in the guinea pig ileum as described by REZENDE Ja et al. (1976) . S-Hydroxytlyptamine and/or histamine were detected in column eluates described previously by I)INIZ (1%3) . Hradykinin, histamine and 5-hydroxytryptamine were usod as comparison standards.

1228

L. REZENDE et al.

EFFLUENT

VOLUME

(ml) F~c. 2 . RECHROMAT'OCRAPtiY OF BIOLOGICAL ACI7VE PRACTIONS SEPARA7ED FROLt P. nigriventer veivoM (FAG . 1). (a) Reverse phase FPLC of the T, fraction (Fig . lc) 0 .5 mg on a 0 .5 x 10 cm C,/Ce PRO RPC column . A linear gradient 0-50% of acetonitrile in 0 .1 % TFA was used. The fraction i?hTx, was lethal to mice and inactive on the guinea pig ileum. (b) Reverse phase FPLC of fraction TZ (Fig. lc, 1 mg) on a PRO RPC C,/C, column (0 .5 x l0 cm). A linear gradient 0-50% of acetonitrile in 0 .1 % TFA was employed. The fraction PhTx2 was lethal to mice (intracerebro-ventricular injection) and contracted the guinea pig ileum. (Inset) Rechromatography of fraction PhTx2 (Fig . 2b) on an HPLC Vydac C,e column (0.46 x 25 cm) . A linear gradient of 255% acetonitrile concentration in 0.1% TFA was used for the elution. The four fractions collected were lethal to mice and contracted the guinea pig ileum . (c) Rechromatography of fraction T3 (Fig . ld) on a PRO RPC column (0 .5 x 10 cm). A linear gradient of 0-50% of acetonitrile in 0.1 % TFA was used. PhTx~ was the fraction that induced flaccid paralysis in mice and was inactive on the guinea pig ileum . (d) Rechromatography of the guinea pig ileum contracting peptide M obtained from the PEP RPC column (Fig . lc) on the PRO RPC C,/CB column . A linear gradient 0-20% of acetonitrile in 0 .1 TFA was used . The active fraction is indicated as M . Chymotrypsin digestion . Lyophilized toxic and/or smooth muscle active fractions were dissolved in 200 pl of 0.2 M Tris buffer, pH 8 .5, containing 20 pg of chymotrypsin and incubated for 30 min at 37°C . The samples were subsequently tested for activity by i .c.v . injection into mice and in the guinea pig ileum as described above.

Protease assay in venom and fractions Proteolytic activity of venom and fractions was detected by electrophoresis on polyacrylamide gels containing co-polymerized gelatin or casein in the presence of SDS at different pH values from 3 .5 to 8 .0 (HANSEN and DownLe, 1980) . Electrophoresis . Electrophoresis of the whole spider venom or fractions was performed on a gel mixture containing 22.5 g of polyacrylamide monomer, 0.8 g of bis acrylamide, 36 g of urea and 2 ml of propionic acid made to l00 ml with distilled water as described by Ct~rtiat and L~wRexce (1989). The mobility was determined by the ratio of net positive charges to mass. Ribonuclease and aotamine were used as roferonce standards . Determination of apparent molecular weights . Apparent mol. wts (M,) were determined by gel filtration on a Superose HR IO/30 FPLC column, previously equilibrated and elutod with guanidine chloride (6 M) or by polyacrylamide gel electrophoresis in the presence of SDS (SDS-PAGE) according to the method of L~ea~aa (1970) using Pharmaeia protein standards of known mol . wt .

1229

Neurotoxins from Spider Venom TABLE

1.

AMINO

ACm

FMCTION3

OF

PhTx,

PhTx 2

PhTx3

10.0 (10) 1.0 (I) 2.9 (3) 3.4 (3) 2.9 (3) 8.6 (9) 5.8 (6) 13.7 (14) 4.5 (5) 0 (0) 0.8 (1) 2.2 (2) 2.9 (3) 1 .! (1) 2.l (2) 6.1 (6) 3.5 (4) n.d .

6.3 (6) 2.7 (3) 3.8 (4) 6.7 (7) 2.8 (3) 7.8 (8) 3.4 (3) 17 .7 (18) 2.7 (3) l.2 (1) 2.5 (3) 2.7 (3) 2.l (2) 2.3 (2) - (0) 5.7 (6) 3.6 (4) n.d .

5.2 (5) 2.2 (2) 3.3 (3) 5.2 (5) 1 .6 (2) 4.6 (5) 2.5 (3) 15 .8 (16) 0.9 (1) I .1 (1) 1 .9 (2) (.6 (2) 2.1 (2) 1.2 (l) 0.6 (1) 4.4 (4) 2.0 (2) n.d .

Amino acid ASX THR SER GLY PRO GLY ALA CYS' VAL MET" ILE LEU TYR PHE HIS LYS ARG TRP

COMPOBI770N OF 1~OXIC nigriventer vENOM

P.

Determined as cysteic acid or methionine sulfone. n.d .-Not determined . Numbers in parentheses represent the closest integer. '

Amino acid analysis. Proteins were hydrolysed in 50 ~l of 4 N methanesulfonic acid containing 0.2% tryptamine in evacuated, sealed tubes at 115°C for 24 hr (SIMP90N et al., 1976) and the amino acid contents of hydrolysates determined on a standard amino acid analyzer equipped with a single column (Beckman W-3 resin) . The values for the labile residues were not corrected for destruction during hydrolysis, and those for half-cystine and methionine were determined as cysteic acid and methionine sulfone, respectively, after performic acid oxidation of the methanesulfonic acid hydrolyzates (OLIVEIRA et al., 1979). Determination of N-terminal amino acid sequences. The DABITC/PI'fC double coupling microsequencing method as described by C~IANO et al. (1978) was used to determine the N-terminal amino acid sequences of the native and reduced S~arboxymethylated toxins . Reduction and S-carboxymethylation . The toxins were reduced and S~arboxymethylated as described by

TABLE

2.

PuwFICA~rION

Purification step Lyophilized venom Sephadex G-50 Superase 12 HR Pcp RPC Pro RPC

aar rrm P. nigriventer roxlNS: Y>ELns, a~F FaACnoNs

Toxic fraction P, S, S~ T, Tz T3 PhTx, PhTx= PhTx3

References Protein to Fig. 1 yield % 1a lb 1b lc lc lg Id le lh

100.00 56 .00 28 .40 5.80 3.10 0.72 5 .23 0.22 0.26 0.33

rErxwLIrr AND srECIFIC ACI7VITY

Specific Yield of lethal potency lethality % (~so/A~ 100.00 108.00 19 .00 20 .00 (.00 8.29 I .23 0.20 4.80 0.08

21 .3 43 .4 15 .1 77 .5 7.5 258.E 5.3 22.0 588.2 7.2

Protein yield expressed in optical density units at 280 nm from 16.0 mg of lyophilized venom dissolved in 0.9'/o NaCI solution. Lethality determined by r .n, ° assay in mice as described in Methods.

1230

L. REZENDE et al.

Toxin

FIG.

3.

ßsqusncs

PhTxi

Ala .Olu .L~u .Thr .ßsr .Cys .Phs .Pro .Val .ßlY-

PhTxQ

Ala .Thr .CYS .Ala .ßlY .ßln .Asp .ßln .Thr .Cys .Lys-

PhTx3 a

Oly .Cys .Ils .ßly .Arp .Asn .Olu .Bsr .ßln .LysLys .Asp .Asn .Val .Tyr .Lys .Phs .Lys .Olu-

THE AM1N0 TERMINAL SEQUENCE OF AMINO ACIDS IN THE PURIFIED TOXIC FRACTIONS OF P. nigriventer vraroM .

Sequences determined by manual DAHTfC/PITC double coupling microsequence method . 'Native neurotoxin reduced and S~arboxymethylated . et al. (1963) . Protein determination . Protein content was determined according to its absorbance at 280 nm and by the Lowry procedure (Lowxv et al ., 1951) using HSA as the standard . CRFSrF1ELD

RESULTS AND DISCUSSION

The toxic components in the crude venom of the spider P. nigriventer were initially obtained as a broad peak by conventional gel filtration on a Sephadex G-50 column. A pool of the toxic material (P4 , Fig. 1a) was thus obtained without loss of toxic activity .

A B

C

D

E

F

P. rligriVeAter VENOM AND PURIFIED TOXINS . Channels: (A) whole venom (20 l+g), (H) PhTx, (20 Ng), (C) PhTxZ (20 t?tsx and Bnz-ty, 1963). Additional gel filtration of the toxic pool by FPLC on a Superose 12 column resulted in the separation of two toxic and smooth muscle active fractions, S, and SZ (Fig. lb). Both fractions were inactivated by digestion with chymotrypsin . Fraction SZ (Fig. 1 b) was further purified by reverse phase chromatography on a Pep RPC column . Two major toxic fractions T, and T2 were separated (Fig. lc). After a second rechromatography on a Pep RPC column the T, fraction was further fractionated by reverse phase chromatography on a PRO RPC column into several non-toxic and one toxic component PhTx, (Fig. 2a). Electrophoresis of PhTx, in a high resolution 22 .5% polyacrylamide propionic acid urea system (Ct-i~rrtst and LnwttEtvct:, 1989) is shown in Fig. 4. The partial amino terminal sequence of this fraction is shown in Fig. 3 . A complete amino acid sequence of this neurotoxin has been recently determined at the University of Durham (Dtxtz et al., 1990). The results indicated that PhTx, is a single chain peptide. No homologies were found when the sequence was compared with those of other proteins stored in the U.S. National Biomedical Research Foundation 1988 data bank. The mol. wt is compatible with the results of amino acid analysis shown in Table 1 . The neurotoxic symptoms induced by i .c.v . injection of PhTx, in mice were tail elevation, excitation and spastic paralysis of the posterior extremities such that the mice crawl supported only by their front legs . Salivation, lachrymation, priapism or scratching were not observed and no contraction was produced in the guinea pig ileum preparation at a concentration of Cr24 pg/ml (n = 3). The i.c.v. trn~ in mice was 45 f 4 faglkg mouse. Fraction TZ after rechromatography on a Pep RPC column was further purified on a Pro RPC column (Fig. 2b). A broad bifurcated toxic peak (PhTxZ) was separated from a small number of non-toxic peptides . Polyacrylamide gel electrophoresis of PhTxZ on 22.5% gels gave four bands which migrated to the cathode at pH 3.5 (Fig. 4). An additional chromatography by HPLC on a Vydac C,8 column resolved the PhTxZ fraction into four lethal components (Fig. 2b, inset) that caused the above signs in mice after i .c.v . injection. The mol. wt of PhTx Z by SDS-PAGE and gel filtration in a Superose calibrated column was in the range 8000-8500. This value is compatible with the amino acid analysis composition of the PhTxZ reaction shown in Table 1 . A partial amino terminal sequence of one of the PhTxZ components isolated by HPLC is shown in Fig. 3. These results indicated that PhTxZ isolated by FPLC (Fig. 2b) is probably a mixture of isoforms of the same toxin. PhTxZ or the components separated from it by HPLC produced excitatory signs in mice (i.c.v.) including salivation, lachrymation, priapism, convulsions and spastic paralysis of the anterior and posterior extremities . Added to the guinea pig ileum preparation (0.2-1 .0 ug/ml, 4 experiments), PhTxZ produced a slow contraction of the smooth muscle which was resistant to washing, antagonized completely by atropine (1 ~eg/ml) and tetrodotoxin (2 ~g/ml) . The t.n~ by i.c.v. injection found for the PhTxZ fraction was 1 .7 t 0.7 ~g/kg mouse. This fraction at a concentration of 0.25 ~g/ml (n = 6) reproduced the effects of the total venom Observed by FONTANA and VrrAL Bttaztt . (1985) on the nerve diaphragm muscle preparation of the rat: non-uniform depolarization of the rat diaphragm muscle fiber membrane and increase in the frequency of miniature end plate potentials (M . D. Fontana and D. Vital Brazil, personal communication) . These

1232

L. REZENDE et al.

effects of PhTxZ fraction were suppressed by tetrodotoxin (3.0 pM) which indicated that they were caused by the activation by PhTxZ of the voltage dependent sodium channel in nerve and muscle cell membrane . Fraction S, obtained after Superose gel filtration (Fig. lb) was resolved into two active components (T3 and M, Fig. ld) on the Pep RPC column using a stepwise gradient from 0 to 34% of acetonitrile in aqueous 0.1 % TFA solution . The fraction T3 induced neurotoxic effects by i.c.v. in mice characterized by a progressive flaccid paralysis of all legs of the mouse for 24 hr or more . This effect was also observed by SCI~xBERG and PÉREIRA LIMw (1976) after s.c. injection of one fraction obtained by ion exchange chromatography of the venom. T3 was further purified on a ProRPC column using a linear gradient of acetonitrile (0-50%) in an aqueous 0.1 % solution of trifluoracetic acid. The paralysing fraction obtained and named PhTx3 (Fig. 2c) was lethal to mice (Ln~ 137 f 9 pg~g) and the flaccid paralyzing effect showed an moo of 0.04 I