The Chemical Nature of Scaritoxin
YONG-GOE JOH and PAUL J. SCHEUER
Introduction
One of the many questions associatedwith ciguatera research has been that ofthe multiple nature of the toxins thatgive rise to ciguatera symptoms in man.The lipophilic ciguatoxin (Scheuer etal., 1967) and the hydrophilic maitotoxin (Yasumoto et al., 1976) are distinctand reasonably well characterized molecules, though their specific contributions to the ciguatera syndrome have notbeen assessed.
Bagnis et al. (1974) conducted an epidemiological survey of ciguatera intoxication in the Gambier islands andobserved that the most frequently implicated fish were parrotfish (Scaridae).Afflicted persons would, in addition tothe conventional immediate ciguaterasymptoms, suffer significantly fromdelayed (by 5-10 days) and prolonged (1or more months) episodes of disturbedequilibrium, locomotor difficulties, andkinetic tremors.
An obvious explanation is the presence in parrotfish of a new toxin calledscaritoxin, or of two distinct toxic entities (Chungue et al., 1977a, b). Fromthe flesh of Scarus gibbus, Chungue etal. (1977a, b) successfully separated twotoxins by chromatography on DEAE
ABSTRACT-Two toxins were isokitedfromScams sordidus (Family Scaridae) collectedat Tarawa atoll, Republic of Kiribati. Bothtoxins were present in the flesh and viscera,but in different ratios. They correspond tothe previously described scaritoxin andciguatoxin. The toxins can be interconvertedand evoke parallel symptoms in mice. Scaritoxin is probably identical with less polarciguatoxin, but lack of material preventedunequivocal proofby spectral comparison.
48(4), 1986
cellulose. A toxin eluted with chloroform and designated SG-l evoked sluggishness and severe hind limb paralysisin mice, different from typical ciguatoxin symptoms. A second toxin elutedwith chloroform-methanol (1:1) anddesignated SG-2 produced conventionalciguatoxin symptoms in mice, such asdiarrhea, lachrymation, salivation, andrespiratory difficulties.
Both toxin entities were further purified on Sephadex LH-20. SG-l was ayellowish oil with an LDso of 0.03 mg/kg. The polar toxin SG-2 was comparedchromatographically with moray eelciguatoxin and was found to be indistinguishable. Both toxins, SG-l and SG-2,responded negatively to reaction withiodine and to Dragendorff and Jaffereagents. A chromatographic estimate ofthe molecular weight of the new, lesspolar SG-l toxin was approximately 800daltons.
Chungue (1977) in a detailed investigation substantiated these results for sixadditional species of parrotfish. Sheconcluded that scaritoxin is a ciguateracausing toxin that is characteristic of theScaridae.
In a more recent study, Nukina et al.(1984) showed that ciguatoxin isolatedfrom moray eel, Lycodontis javanicus(= Gymnothorax), viscera can be separated on basic alumina into two entities, differing in polarity but indistinguishable in symptoms or in LDso (i.p.mice). The two toxins show only minordifferences in their high field 'H NMFspectra. It was further shown that thetwo toxins are interconvertible. It oc-
The authors are with the Department of Chemistry,University of Hawaii at Manoa, 2545 The Mall,Honolulu, HI 96822. The pennanent address ofYong-Goe Joh is: Department of Food Science andNutrition, Dong-A University, Pusan, Korea.
curred to us that these two chromatographic forms of ciguatoxin might beidentical with the scaritoxin-ciguatoxinpair recognized in the flesh of parrotfish by Chungue (1977). We had an opportunity to collect parrotfish, S. sordidus, on Tarawa atoll, Republic ofKiribati, and thus subject this hypothesis to an experimental test.
Materials and Methods
Fish were collected 19-17 May 1983at Taborio, Betio, Bikenibeu, and Teaoraereke (Fig. 1). Small-scale extractionand testing showed that only S. sordiduscaught at Bikenibeu were toxic. Flesh(5.34 kg) and viscera (0.876 kg) wereworked up separately.
Fish flesh was mixed with acetoneand reduced to a brei in a I-gallon Waring Blendor'. The mixture was transferred to a 4 liter Erlenmeyer flask andallowed to stand in excess acetone for2 days, then with fresh acetone foranother 4 days. The combined acetoneextracts were concentrated to an aqueous suspension, which was treated twicewith equal volumes of hexane. The hexane layer was back-washed with 3 X 100rn1 of methanol/water, 8:2. The washings were added to the aqueous suspension, and the combined aqueous material was extracted three times with equalvolumes of ethyl acetate. Evaporation ofethyl acetate resulted in crude toxin.
Crude toxin was dissolved in a minimum amount of chloroform and appliedto the top of a column packed withSilicar (200-425 mesh, Mallinckrodt,28 gig sample). Successive elution withchloroform (10 ml/g silica), chloroform!
IReference to trade names or commercial finnsdoes not imply endorsement by the NationalMarine Fisheries Service, NOAA.
19
( ) Sand covered reef, mainly dry at low tide~----~
- - - - - - - - - - - - - - Approximate edge of lagoon
The toxic fractions (polar toxin, 450ng; less polar toxin (810 ng) were further purified on a Sephadex LH 20(Pharmacia) column (103 x 1.4 cm) byelution with cWoroformlmethanol (2:1).Fractions (4.9 ml) were collected every10 minutes. Elution was monitored at254 nm (ISCa Model UA-5).
Aluminum plates coated with silicagel 60F-254 (0.2 mm) and glass platesspread with silica gel H were used forthin-layer chromatography (TLC).Plates were developed with cWoroformlmethanol/water/acetic acid (90:9.5:0.3:0.2) and visualized with iodine vapor.Spots or bands were scraped off and extracted with chloroform/methanol, 4:1for the less polar and 1:1 for the polartoxin. For bioassay the recovered TLCfractions were homogenized in 1 percentTween 80.
For bioassay, each sample was dissolved in a known volume of methanol.Aliquots were removed from this stocksolution and solvent was removed by astream of nitrogen. Tween 80 (0.5 mlof 1 percent or 0.1 ml of 5 percent) wasadded and homogenized in a Vortexmixer. Swiss Webster mice (male orfemale, 16-22 g) were injected intraperitoneally. Death times of less than 3hours were used to estimate toxicity asdescribed by Tachibana (1980).
173"10'
~-N-
~
173"05'
SCAlE
L-'9_-'----'1_--'----"----'STAIUIE MILES
o ,
"
(~ ~) Dry land above sea level
17'5trJ'
" ...~ '.
~ :.'~~TABORIO
I;::> , ~ ••~TEAORAEREKE
172"55'
,I,,
(I' ~t::~~··; t:\,,\II\,I
II
Figure I.-Tarawa atoll, Republic of Kiribati.
methanol (9:1, 14 ml/g silica), chloroform/methanol (1:1, 10 ml/g silica), andmethanol (10 ml/g silica) separated thetoxin. All fractions were evaporated todryness immediately after elution.
The chloroform/methanol fraction(9:1) was further purified on a DEAEcelulose column (Cellex D, acetateform, BioRad Laboratories, 9.3 gigsample), which was prepared as previously described (Rouser et aI., 1963).Eluting solvents were chloroform (60ml/g adsorbent), chloroform/methanol(1:1, 90 ml/g adsorbent), and methanol(60 ml/g adsorbent). Guaiazulene (ablue pigment, Aldrich Chemical Co.)was used to determine the bed volumeand to detect imperfections in the packing.
An alumina (activity grade I, Woelm)column was prepared by suspendingalumina (100 gig sample) in cWoroform.Toxic fractions were added as chloroform solutions and were eluted successively with chloroform/methanol (100:1, 9:1, 1:1), methanol, methanol/water(1:1), TI5 ml/g alumina.
Alumina activity grade V was prepared according to the Brockmannscale. Grade I and V alumina columns(18.5 X 0.7 cm) were packed with 10 galumina each and the toxins were elutedwith chloroform (40 ml), chloroformlmethanol (36 + 4 ml), chloroformlmethanol (1:1), methanol (40 ml), andmethanol/water (1:1). The eluates wereconcentrated and the residues weredissolved.
Results
Small-scale extractions and bioassayshowed that S. sordidus was the onlytoxic species; S. jrenatus, S. scaber, andS. pectoralis were nontoxic.
Yields and lethalities of extracts andchromatographic fractions are shown inTables 1 (flesh) and 2 (viscera). Theviscera proved to have a higher concentration of toxin, a phenomenon that hadpreviously been observed in the morayeel (Yasumoto and Scheuer, 1969). Toxin could be eluted from silicic acidcleanly in cWoroform/methanol (9:1).The toxin was separable into polar andless polar entities on DEAE cellulose.In S. sordidus flesh the polar toxin predominated, while in the viscera thereverse was so. When the chloroformlmethanol (1:1) fraction from the DEAEcellulose column was rechromatographed on alumina activity grade I and
20 Marine Fisheries Review
Figure 2.-Interconversion of Scarus sordidus flesh toxins on alumina.
I After treatment with IN NaOH in aqueous methanol.
28: chloroform-methanol (9:1). C: chloroform-methanol (1:1).E: methanol-water ( I: I ).
Column charge ST-I,I.I~g ST_II, 1.13 ~g ST-2, I. 2 ~g
Activity grade I I II.
Eluates 2I I I I I I I I I
~I I
8 C E 8 8 EAmounts
0.12 0 . 15of toxin 0.55 0.12 0.34 0.21 0.80recovered ~g ~g ~g ~g ~g ~g ~g
Table 3.-Comparlson of Rr-values of S. 50rd/dus flesh toxins with PCTX (Tachlbana, 1980)and scaritoxln (Chungue, 1977). Aluminumplates were coated with silica gel 60F·254 (0.2mm): solvent system was chloroform/methanol!water/acetic acid (90:95:0.2:0.3).
to ST-1 by chromatography over deactivated (Grade V, 15 percent water)alumina. When ST-1 is first treated withaqueous methanolic sodium hydroxide,then chromatographed on alumina of activity I, partial conversion to ST-2 alsotakes place, but the loss of material isnecessarily severe. Paucity of availabletoxin prevented us from carrying out afull-scale experiment. Results are shownin Figure 2. The less polar toxin isdesignated ST-l, the polar toxin, ST-2.
On subsequent chromatography onSephadex LH 20, ST-l and ST-2 areeluted in parallel fractions (88-110 ml forST-1 and 86-108 ml for ST-2) , thusdemonstrating the identical size of thetwo toxins.
To show the relationship of the twoS. sordidus toxins to ciguatoxin (Nukinaet al., 1984) and to scaritoxin (Chungue,1'177), we carried out thin layer chromatography under Chungue's conditions.Within the normal variability of TLCwith time and place, and with a foursolvent developer, ST-1 and scaritoxinare very likely to be identical, as areST-2 and ciguatoxin (Table 3). The TLCspots are egg-shaped rather than circular. Significantly, though, the Rrvalues do not overlap. Amounts of toxinsufficiently large for 'H NMF comparison are needed for unequivocalproof.
21
Discussion
The demonstration of two chromatographically distinct toxins in a ciguatoxic fish, while interesting by itself,raised another question concerning theprecursor(s) of the toxins. Parrotfish, S.gibbus, feed primarily on coral. To shedlight on the origin of the toxin, Yasumoto in collaboration with workers inTahiti (Yasumoto et aI., 1'177) examinedthe gut contents and liver of S. gibbus.The gut, surprisingly, contained noscaritoxin, but contained ciguatoxin andmaitotoxin in addition to a fast-actingacetone-soluble paralysis-causing toxin.Only ciguatoxin was isolated from theliver. Since both ciguatoxin and scaritoxin are routinely isolated from flesh,these results suggest that the parrotfishmay have the ability to transform ciguatoxin into scaritoxin, yet the absence ofscaritoxin in the liver was puzzling.
0.92
0.78-
ScaritoxinPCTX
0.28-
0.54
ST-2
0.54
0.30-
ST-1
0.60-
0.75
eluted with a gradient beginning withchloroform/MeOH (100:1) and endingwith methanol/water (1:1),95.5 percenttoxicity was eluted with methanol/water(1:1). The two toxins gave rise to parallel symptoms in mice.
In analogy with the recent demonstration (Nukina et al., 1984) that ciguatoxinfrom moray eel viscera can be separatedinto two distict entities of differentpolarity by alimina chromatography andthat the two toxins are interconvertible,we were able to show that the two S.sordidus toxins obtained by DEAEcellulose chromotography can also bepartially interconverted. Figure 2 showsthat ST-1 is partially converted to ST-2when passed through a column of highlyactive (Grade I) alumina. ST-2, on theother hand, can be partially converted
Table 1.-Yields and toxicity of extracts and of chrom-atographic fractions of S. sordidu5 flesh (5.34 kg).
TotalYield LDso toxicity
Purification state (g) (mg/kg) (M.U.)
Ethyl acetate 8.99
Silicic acidChloroform 3.52 NontoxicChloroform/methanol 1.53 23.0 3,319(9:1)(1:1) (9:1) 0.63 12.5 2,520
(1:1) 1.13 NontoxicMethanol 2.20 Nontoxic
DEAE-celluloseChloroform 2.74 29.0 556Chloroform/methanol 0.52 22.5 1,160(1:1)Methanol 0.01 Nontoxic
Table 2.-Ylelds of extracts and of chromatographicfractions of S. 50rdidu5 viscera (0.876 kg).
TotalYield LDso toxicity
Purification state (g) (mg/kg) (M.U.)
Ethyl acetate 13.81
Silicic acidCHCI3 4.46 NontoxicCHCI,.CH3OH 5.05 29 8,763(9:1)CHCI,.CH,oH 1.48 Nontoxic(1:1)CH,oH 1.58 Nontoxic
DEAE-celluloseCHCI3 2.94 29.5 5,034CHCI,.CH,oH 1.00 27.8 1,799CH,oH 0.08 Nontoxic
48(4), 1986
Our findings that toxic parrotfishfrom Tarawa atoll come from a narrowgeographic area (Fig. 1) is consistentwith the traditional ciguatera phenomenon (e.g., Withers, 1982). We wereboth surprised, though, that of fourspecies examined, only S. sordidusproved toxic. We were able to isolate twotoxins, parallel with Chungue's (1fJ77)work on S. gibbus. In contrast to Yasumoto's (Yasumoto et al., IfJ77) results,we were able to show that both toxinsare present in the flesh and in the viscera, albeit in different proportions. Weobserved no evidence of a third acetonesoluble and fact-acting toxin (Yasumotoet al., IfJ77). This is not a critical point,as coral-feeding parrotfish have manyopportunities to ingest other toxins.Bagnis' (1fJ74) original observation ofthe different symptomology in man
22
following intoxication by parrotfish remains unexplained.
Acknowledgments
We thank Mark R. Hagadone andDominic McCarthy for collections, JohnE. Randall for identification of fishes,and the National Marine Fisheries Service (NA80AA-OOlOl) through a subcontract with the Medical University ofSouth Carolina for financial support.
Literature Cited
Bagnis, R .• E. Loussan, and S. Thevenin. 1974.Les intoxications par poissons perroquets auxIles Gambier. Med. Trop. 34:523-527.
Chungue, E. 1977. Le complexe toxinique despoissons perroquets. Ph.D. thesis. Academiede Montpelier, Universite de Sciences et Techniques de Languedoc, Montpelier, France.
____ , R. Bagnis, N. Fusetani, and Y.Hashimoto. 1977a. Isolation of two toxins froma parrotfish Scarus gibbus. Toxicon 15: 89-93.
, and T. Yasumoto,
1977b. Le complexe toxinique des poissonsperroquets. Biochimie 59:739-741.
Nukina, M., L. Koyanagi, and P. 1. Scheuer. 1984.Two interchangeable forms of ciguatoxin. Toxicon 22:169-176.
Rouser, G., G. Kritchevsky, D. Heller, and E.Lieber. 1963. Lipid composition of beef brain,beef liver, and the sea anemone: Two approaches to quantitative fractionation of complex lipid mixtures. 1. Am. Oil Chern. Soc.40:425-454.
Scheuer, P. 1., W. Takahashi, 1. Tsutsumi, and T.Yoshida. 1967. Ciguatoxin: Isolation and chemical nature. Science 155:1267-1268.
Tachibana, K. 1980. Structural studies on marinetoxins. Ph.D. Thesis, Univ. Hawaii, Honolulu,157 p.
Withers, N. W. 1982 Ciguatera fish poisoning.Ann. Rev. Med. 33:97-111.
Yasumoto, T., R. Bagnis, and J. P. Vernoux. 1976.Toxicity of surgeonfishes. - II. Properties ofthe principal water soluble toxin. Bull. Jpn. Soc.Sci. Fish. 43:359-365.
----,=-----,-_ , I. Nakajima, E. Chungue, and R.Bagnis. 1977. Toxins in the gut contents ofparrotfish. Bull. Jpn. Soc. Sci. Fish. 43:69-74.
_,------,--,- , and P. 1. Scheuer. 1969. Marine toxins of the Pacific - VIII. Ciguatoxin from morayeel livers. Toxicon 7:273-276.
Marine Fisheries Review