Ciguatoxin from the flesh and viscera of the barracuda, Sphyraena jello

Taziaoa, Vd. 22, No. S, pp. 80i-810, 1984 . Printed ia Great aritain.

0041-0101/84 ß.00+ .00 O 1984 Pergamoa Pres Ltd.

CIGUATOXIN FROM THE FLESH AND VISCERA OF THE BARRACUDA, SPHYRAENA JELLO RICHARD J . LEWIS and R . ENDEAN Department of Zoology, University of Queensland, St . Lucia, Brisbane 4067, Queensland, Australia (Accepted jor publirntion 26 April 1984) R . J . LBwls and R . ErmF.~ri . Ciguatoxin from the flesh and viscera of the barracuda, Sphyraena jello. Toxicon 22, 805 - 810, 1984 . - Five of six barracuda, Sphyraenajello, captured in the same area of Queensland as toxic Spanish mackerel (Srnmberomorus rnmmersoní) were found to be toxic . One major lipid-soluble toxin, present in both the flesh and viscera of a pooled sample of barracuda, was chromatographically indistinguishable from Spanish mackerel flesh ciguatoxin. Signs and symptoms induced by toxic barracuda in humans were typical of ciguatera and cats and mice displayed signs indistinguishable from those induced by Spanish mackerel ciguatoxin . Purified barracuda ciguatoxin had an i .p . Ln to mice of SS Ng/kg . Barracuda ciguatoxin had a narrow range (0.7-1 .4 mouse units) of doses between 0% and 100 lethality and its dose vs . death-time relationship was similar to the relationship produced by moray eel and Spanish mackerel ciguatoxin. No water-soluble toxins were detected in the viscera of these barracuda, but a less polar lipid-soluble toxin was found in the viscera. This minor toxin represented 14ío of the total viscera lethality and induced signs in mice similar to the major toxin . Barracuda could be an important link in the transfer of ciguatoxin to large Spanish mackerel in southern Queensland waters .

INTRODUCTION CIGUATERA, a serious form of fresh fish poisoning, is caused by the ingestion of toxic individuals of many species of tropical and sub-tropical fishes . The lipid-soluble nature of a toxin causing food poisoning was detected in barracuda (Sphyraena picuda) by HASHIMOTO (1956) and barracuda (Sphyraenidae) have been implicated with ciguatera throughout the world (BARKIN, 1974; DE SYLVA, 1982; MooN, 1981 ; MORTON and BURKLEw, 1970; SOROKIN, 1975). Ciguatoxin, as isolated from moray cel flesh (SCHEUER et al., 1%7), is considered to be the principal toxin responsible for ciguatera . The toxic material present in ciguateric barracuda has not been chemically characterized and previous studies indicate the possible existence of more than one toxin in barracuda flesh (BANNER, 1%7; HASHIMOTO, 1979; ROYAL et al., 1982) . Although barracuda are not popular food fishes in Queensland, Australia, at least 9 cases of ciguatera have been attributed to barracuda captured from Hervey Bay, Queensland, Australia, since 1976 (unpublished observations) . After four recent cases of barracuda poisoning, in which the signs and symptoms displayed by the victims were typical of ciguatera, a further six whole barracuda also caught in Hervey Bay were obtained . These fish were identified as Sphyraenaje11o Cuvier . In this paper we report on the toxins present in the flesh and viscera of these toxic barracuda .

806

R. J. LEWIS and R. ENDFAN

MATERIALS AND METHODS Six whole frozen barracuda, Sphyraena jello, (mean weight 1 .7 kg) were obtained and stored at -18°C prior to testing. A portion of the flesh of each barracuda was fed separately to cats at 6-10~ of the body weight of each cat . Cats were observed intermittently for two days . The flesh of five of the six barracuda was toxic. All affected cats displayed signs typical of ciguatera (LEwts and ENne~v, 1983). Mouse bioassay Mice (16-25 g, Quackenbush) were injected i.p . with toxin emulsified in 1% Tween 60 saline and signs and death-time recorded . Ln values are expressed per kg of mouse and total lethality is expressed in mouse units (MU) . One MU equals one Ln,°+50 dose . Total lethality of each fraction was determined by a modification of the method of Tachibana, K. (Structural studies on marine toxins, Ph .D . Thesis, University of Hawaii, 1980) using the equation, log(MLT) = 21og(1 + death-time(hr)-'), which relates death-time to lethality for20 g mice (at least 2 mice per determination) . The accuracy of these results was confu-med by injecting three mice at four serial doses and then determining the Ln,° from the dose-response curve. A dose vs death-time relationship was obtained for barracuda ciguatoxin and Spanish mackerel ciguatoxin (using toxin extracted by Lswis and ENDEAN, 1983) using mice weighing 19-21 g. From the equation of Tachibana (lot. tit.) a curve of dose vs death-time was constructed and superimposed on the data for Spanish mackerel and barracuda ciguatoxins (Fig. 1). Theshortest tíme of survival was calculated for each toxin from the dose vs death-time values following the method of Mot.trrEtvao (1979) . Preparation of a crudt, methanol-soluble jraction This method followed the procedure used by LEwrs and Ermsnty (1984) . The flesh (4 .65 kg) of throe barracuda was heated at 150°C for 100 min and then extracted twice with 14 liters of acetone for 15 min using an Ika-UhraTurrax T 45 DX disperser at room temperature. The viscera from these barracuda (330 g, including livers and empty guts) were cooked in the same manner and then extracted twice with one liter of acetone using the disperses . Acetone-insoluble materials from flesh and viscera were removed after precipitation at -18°C overnight. The acetone-soluble residues were partitioned between diethyl ether and aqueous 20% NaCI (4:1 v/v, x 3) . The diethyl ether residues were further partitioned between 80go methanol-water and n-hexane (1 :1 v/v, x 3) . The methanol-soluble residues yielded 8.0 g (flesh) and 1 .1 g (viscera) andwere tested for lethality to mice . Preparation of water soluble, acetone soluble and acetone~insoluble fractions The method used was a modification of the method of YASUMOTO and Knivwo (1976). Viscera (172 g, empty gut, no liver) and liver (35 g) were removed from the remaining two toxic barracuda, minced and extracted with boiling methanol (2:1 v/w, x 3) . The aqueous residue remaining was extracted with diethyl ether (3 :1 v/v, x 2) and then with n-butanol (2 :1 v/v, x 3) . The aqueous phase remaining was designated the water-soluble fraction . The n-butanol residue was further separated into an acetone-soluble fraction and an acetone-insoluble fraction after the addition of acetone. The acetone-insoluble fraction was dialyzed against water and the retentate lyophilyzed. Portions of each fracdon, equivalent to S g of viscera or 5 g of liver, were injected i.p . into mice (18-22 g) . Mice were observed intermittently for three days . Pur~cation of the methanol-soluble fractions The methanol-soluble fractions wen purified by chromatography using a modification of the method of Lawts and Ermrrw (1983) . Columns of 100 mesh silicic acid (Mallinckrodt, St Louis), DEAF-cellulose (acetate form) (Brown, New Hampshire), Sephadex LH-20 (Pharmacia, Uppsala) and reverse-phase C-18 (semipreparative HPLC) (Waters Associates, Milford) were used and each complete toxic fraction was applied (less approximately 15 MU required for assay) to each column in succession . Details of elution conditions aregiven in Table 1 . The C-18 column was eluted at 22°C with methanol-water (4 :1) at 2 ml/min using a Waters 6000A pump. Elution from Sephadex LH-20 and C-18 columns was monitored at 206 nm using an LKH Uvicord S. Toxic zones were located by determining the lethaflty of fractions to mice. Fractions not lethal to mice at 1 g/kg were regarded as non-toxic. To test thestability of barracuda ciguatoxin to repeated chromatography on DEAEcellulose, portions of barracuda flesh toxin were reapplied several times to a DEAF-cellulose column (acetate form). Thin-layer chromatography

Silica ge160 precasted plates, 0.25 mm thick, (Mark, Darmstadt) were activated for 30 atin at 110°C and used for analytical separations. The solvent system chloroform-methanol-6 N ammonium hydroxide (90:9 .5 :0 .5) was used and bioassay was performed as previously described (LEWIS and ExnEntv, 1983). Chemicals used The solvents n-hexane, chloroform and methanol were HPLC grade (Waters, Milford, U.S.A .) and were used as provided . All other solvents used were analytical reagent grade. For silica gel, DEAF-cellulose and Sephadex LH-20 chromatography, 4-methyl-2,6-di-t-butylphenol (Koch-Light, U.K .) was added at a concentration of 0.005% w/v.

Ciguatoxin from Barracuda

807

~o,

e_ 8

32_ ~7/-

s

D~~th-Tim~ (Hr) FtG . 1 . DOSE VS . DEATH-TIME RELATIONSHIPS FOR CIOUATOXINS FROM THREE SOURCES.

The continuous curve (-) was calculated from the equation of Tachibana (Structural studies on marine toxins, Ph .D . Thesis, University of Hawaü, 1980) for classical ciguatoxin . Data for barracuda flesh toxin (" ) and Spanish mackerel ciguatoxin ( ~ ) clearly are comparable with those for classical ciguatoxin. ODe mouse tested at each point. MU = mouse unit (one MU = one LD, a + 30). RESULTS

Mouse bioassay Mice injected with the methanol-soluble fractions from the flesh and viscera of barracuda displayed signs typical of ciguatera (LEwls and ENDFAN, 1983). Both barracuda and Spanish mackerel ciguatoxin showed a sigmoid dose - response curve with a dose range of 0.7 -1 .4 MU between 0% and 100% lethality. Ciguatoxin from barracuda and from Spanish mackerel showed dose vs . death-time relationships (Fig . 1) overlapping the curve calculated for moray eel viscera ciguatoxin. However, the shortest time of survival for moray eel ciguatoxin (20 min) was different from that for barracuda (36 min) and Spanish mackerel (37 .5 min) toxins . Potential errors in calculating LDso values for barracuda and Spanish mackerel toxins were avoided by using death-times greater than 45 min. The minor toxin isolated from the barracuda viscera induced signs in mice similar to barracuda flesh toxin . No toxic signs were elicited in mice by water-soluble, acetonesoluble and acetone-insoluble fractions from either the viscera (excluding the liver) or the liver. Purification of methanol-soluble fractions The methanol-soluble fractions from the flesh and viscera had LDso values of 380 mg/kg and 97 mg/kg, respectively . The yield and lethal potency of fractions at each stage during purification are given in Table 1 . The single toxin found in the flesh was indistinguishable from Spanish mackerel ciguatoxin (LEWIS and ENDFAN, 1983). Two toxins were found in the viscera. The major toxin was indistinguishable from the flesh toxin, while the minor toxin (approximately 14% of total viscera lethality) was less polar than the flesh toxin (Table 1) . The minor toxin was not distinguished from flesh toxin by DEAF-cellulose column chromatography. The major toxin from the barracuda flesh and viscera was eluted from a C-18 column in 28 - 32 min, as is Spanish mackerel ciguatoxin (LFwls and ENDFAN, 1984). The flesh toxin finally purified by C-18 chromatography had an LDso of

808

R. J. LEWIS and R. ENDEAN

55 Fig/kg . The methanol-soluble barracuda toxin was stable to repeated elution from a DEAF-cellulose column (acetate form). Thin-layer chromatography The RJ values on silica gel plates for barracuda flesh toxin (0.1-0.3) and the major viscera toxin (0.08 - 0.25) were similar to the band for Spanish mackerel ciguatoxin (0.08 - 0.2) (LEWIS and ENDEAN, 1983). The minor toxin from barracuda viscera had a different Rf value of 0.25 - 0.4. DISCUSSION

When ingested by humans, barracuda flesh induced signs typical of ciguatera . Signs elicited in mice and cats by the major lipid-soluble toxin from the barracuda were similar to those reported for moray eel ciguatoxin and Spanish mackerel ciguatoxin. The doserange between 0% and 100% lethality to mice was narrow, as has been reported for the toxin from ciguaterie Lutjanis óuccanella (HOFFTvIAN et al., 1983). Also, the dose vs death-time relationship for barracuda and Spanish mackerel toxins were similar to the relationship calculated for moray eel ciguatoxin. The actions of barracuda ciguatoxin TABLE 1 . YIELD. LDP AND TOTAL LETHALITY OF TOXINS FROM BARRACUDA FLESH

Stage of purification Flesh Crude, methanol-soluble fraction Silicic acid column chromatography chloroform-methanol (97:3) eluate chloroform-methanol (9:1) eluate chloroform-methanol (1 :1) eluate DEAF-cellulose column chromatography chloroform eluate chloroform-methanol (1 :1) eluate Sephadex LH-20 gel filtration column 91 x 2.6 cm eluted with methanol toxic fraction 300-364 ml Viscera Crude, methanol-soluble fraction Silicic acid column chromatography chloroform-methanol (97:3) eluate chloroform-methanol (9:1) eluate chloroform-methanol (1 :1) eluate DEAF-cellulose column chromatography (a) minor toxin chloroform eluate chloroform-methanol (1 :1) eluate (b) major toxin chloroform eluate chloroform-methanol (1 :1) eluate Sephadex LH-20 gel filtration of major toxin column 91 x 2.6 cm eluted with methanol toxic fraction 290-350 ml . "Non-toxic at lg/kg. tMinor toxin. #Major toxin.

(4.óS kg) AND VISCERA (330 g)

Yield (g)

(mg/kg)

Total lethality (MU)

8.0

376

1055

0.32 0.78 1 .73

non-toxic" 61 non-toxic'

644

0.05 0.157

non-toxic' 14 .1

556

0.028

2.4

588

1 .1

97

565

0.2 0.09 0.4

125 22 .5 non-toxic"

80t 200#

0.067 0.101

non-toxic" 95

53

0.045 0.037

non-toxic" 8 .9

209

0.005

1 .3

190

LDeo

Ciguatoxin from Barracuda

80 9

(unpublished observations) and moray eel ciguatoxin (MIYAHARA et al., 1979) on the guinea-pig left atria are similar. The actions of barracuda ciguatoxin (unpublished observations) and Spanish mackerel ciguatoxin (LEWIS and ENDEAN, 1984) on guinea-pig ileum are indistinguishable. However, the shortest time of survival results for mice indicate that the toxins from barracuda and Spanish mackerel may be different from moray eel viscera ciguatoxin . For death-times greater than 45 min the equation of Tachibana (loc. cit.) can accurately and rapidly quantify ciguatoxin from different fishes . This method uses a minimum number of mice . Only one major methanol-soluble toxin was isolated from the flesh and viscera of a pooled sample of specimens of the barracuda, Sphyraena jello. The concentration of toxin in the barracuda viscera (1 .7 MU/g) was 7-fold that in the flesh (0.23 MU/g). This major toxin is identical with Spanish mackerel ciguatoxin on a variety of chromatographic systems, but had an Rf different from moray eel flesh ciguatoxin (0.25 - 0.38) (CHUNGUE et al., 1977) . The difference in Rj value between barracuda ciguatoxin and moray eel ciguatoxin may stem from differences in laboratory conditions and procedures employed for thin-layer chromatography or could be a real difference. Barracuda (S. jello) ciguatoxin is different from the two toxins in the flesh of ciguateric Scares gibbus (Chungue, E., Le complexe toxinique des poissons perroquets, Ph.D . Thesis, Université des Sciences et Techniques du Languedoc, Academie de Montpellier, 1977) which are unstable to repeated chromatography on DEAE-cellulose . Moray eel ciguatoxin (Chungue, loc. cit.) and barracuda toxin are stable to repeated elution on DEAEcellulose. A minor, lipid-soluble toxin was isolated from the viscera of barracuda. This toxin had a higher Rf value than the flesh toxin and accounted for 14% of the total lethality of the barracuda viscera. A minor, less-polar toxin was also found in the viscera of moray eels (Tachibana, loc. cit.). Failure to detect water-soluble toxins, such as maitotoxin, in the viscera of the piscivorous barracuda was not unexpected . These toxins are usually found in large amounts in the viscera and gut contents of herbivorous and coralivorous fishes (YASUMOTO et al., 1971, 1977). We have shown that barracuda and Spanish mackerel caught in the same area in Hervey Bay near Fraser Island contain identical toxins . Many of the toxic Spanish mackerel from this area are larger than 10 kg and might be expected to prey on barracuda of this size (1.7 kg). The incidence of ciguatoxic barracuda in Hervey Bay (five of six fish were toxic) is high. We propose that barracuda may be an important link in the transfer of ciguatoxin to large Spanish mackerel in this area . Recently, VERNOUX et al. (1982) found a toxin in ciguateric Caranx bartholomaei which was chemically different from ciguatoxins of moray eel, Spanish mackerel and barracuda. The presence of chemically different toxins in fishes may explain the wide variability of symptoms reported for ciguatera (BANNER et al., 1963 ; ENGLEBERG et al., 1983). Further investigations are required to determine the number of toxins involved in ciguatera. Acknowlaigements - This research wes supported by the Fishing Industry Research Council, Australia . We are indebted to W . J . Sm~.ns, Chief Inspector of Foods, and R . V . Hot,~s, Senior Inspector of Foods, of the Queensland Health Department for providing the barracuda and case histories for the victims of barracuda poisoning . We thank W. STABLUM for technical assistance . REFERENCES B~rMea, A. H . (196 Marine towns from the Pacific, 1 - Advances in the investigation of fish toxins . In : Anima! Toxins, p . IS7 (RussBt,L, F . E . and SwtnmBas, P . R ., Eds). Oxford : Pcrgamon Press . B~tvxsx, A. H., Stew, S ., At.swnsa, C . and Hst.muCH, P . (1963) Fish intoxication: notes on clguatera, its mode of action and a suggested therapy. South Pacjjîc Commission Technical Paper no. 141 .

81 0

R. J.

LEWIS and

R.

ENDFAN

BwwuN, R. M. (1974) Ciguatera poisoning: a common source outbreak . Sth. med. J., Nashville 67, 13 . CHUNGUE, E., BwGNts, R., FUSETANt, N. and HASH~MOTO, Y. (1977) Isolation of two toxins from a parrotfish

Scams gibbus. Taxirnn 15, 89 .

DE SYLVw, D. P. (1982) Barracuda - dangerous alive or dead. Oceans 15, S. ENGLEeEaG, N. C., MORRIS, J. G., LEWIS, J., McMaLwN, J. P., POLLARD, R. A. and BLAKE, P. A. (1983) Ciguatera fish poisoning: a major common-source outbreak in the U.S . Virgin Islands. Ann. intern. Med. 98,

336.

Y. (1956) A note on the poison of a barracuda, Sphyraenapicuda Bloch and Schneider. Bull. Jap. Soc. scient . Fïsh . 21, 1153 . HASHIMOTO, Y. (1979) Marine Toxins andOtherBioactive Marine Metabolites, p. 110. Tokyo: Japan Scientific HASHIMOTO,

Societies Press. HOFFMAN, P. A., GRANADE, H. R. and McMiLLAN, J. P. (1983) The mouse ciguatoxin bioassay : a dose-response curve and symptomatology analysis . Toxicon 21, 363 . LBwiS, R. J. and ENDEAN, R. (1983) Occurrence of a ciguatoxin-like substance in the Spanish mackerel

(Scomberomorus commerson~). Toxicon 21, 19 . R. J. and ENDEAN, R. (1984) Mode of action of ciguatoxin from the Spanish mackerel, Scomberomorus cnmmersoni, on the guinea-pig ileum and vas deferens . J. Pharmac. exp. Then, 228, 756. MIYAHAAA, J. T., AKwu, C. K. and YASUMOTO, T. (1979) Effect of ciguatoxin and maitotoxin on the isolated guinea-pig atria. Res. Commun . them. Path. Pharmac. 7á, 177. MOLINENGO, L. (1979) Thedose vs survival time in the evaluation of acute toxicity . J. Pharm. Pharmac. 31, 343 . MooN, A. J. (1981) Ciguatera poisoning. Practitioner 225, 1176 . MORTON, R. A. and BuxKLew, M. A. (1970) Incidence of ciguatera in barracuda from the west coast of Florida. Toxicon 8, 317. ROYAL, S. C., Pou, M. A., MENDS, T. J., BADEN, D. G., GALHRAITH, R. M., HIGERD, T. B., ENRIQUEZ, M. and JANOWSKI, H. J. (1982) Ciguatera fish poisoning -Bahamas, Miami. MorbidityMortality wkly Rep. 31, 391. ScHauEn, P. J., TAKA}IASHI, W., TSUT'SUMI, J. and YOSHIDA, T. (1967) Ciguatoxin : isolation and chemical nature . Science, N. Y. 155, 1267. SoxoKiN, M. (1975) Ciguatera poisoning in north-west Viti Levu, Fiji Islands. Hawaii med. J. 34, 207. VERNOUX, J. P., GAIGN, M., RIYECHE, N., TAOMOUTI, F., MwGRws, L. P. and NOLEN, J. (1982) Mise en évidence d'une toxine liposoluble de type ciguatérique chez Charanx bartholomaei péché aux Antilles françaises . Biochimie 64, 933. YASUMGTO, T., HASHIMOTO, Y., BAGNIS, R., RANDALL, J. E. änd BANNER, A. H . (1971) TOX1Clty Of the surgeonfishes. Bull. Jap. Soc. scient. Fish. 37, 724. YASUMOTO, T. and KANNO, K. (1976) Occurrence of toxins resembling ciguatoxin, scaritoxin, and maitotoxin in a turban shell. Bull. Jap. Soc. scient . Fish . 42, 1399 . YASUMOTO, T., NAKAJIMA, L, CHUNGUE, E. and BAGNIS, R. (1977) Toxins in the gut contents of Parrotfish . Bull. Jap. Soc. scient. Fish. 43, 69. LEwrs,