gjtglATINGCIA'I ya Grail'9Fish Poisoning

n x a' 1 RRR e 0

Edited by T. R. Tosteson OcThomas R. Tosteson 1992. All rightsreserved.

Production Credits Proofreading;Carmen Tosteson, Val Hensley Designand Layout: Alexis Tosteson Garcia, Angel An.ibal Escobar Cover Art: Bonnie Dennis Bower CoverDesign: Angel Anibal Fscobar

ISBK 0-92 l 317-35-2

PolysciencePublications, inc. MorinHeights, Quebec, Canada Printed in Canada. Contents

Preface

Acknowledgements vu

Chemistry of Fish artd Dirtoflagellate Toxins

Structure Determination of Ciguatoxin of Moray Eels and Garrrbierdiscus toxicus. T. Yasumoto,M. Murata, Y. Ishibashi, M. Fukui and A.M. Legrand

Extraction and Purification of Toxic Fractions From Barracuda 13 Sphyraerra barracuda! Implicated in Ciguatera Poisoning. Pedro M. Gamboa,Douglas L. Park, and Jean-MareFrerny

Characterizationof Ciguatoxins from Different Fish Speciesand Wild Gambierdiscustoxicus. A-M. Legrand, M. Fukui, P. Cruchet, Y. Ishibashi and T. Yasumoto

Brevetoxinsand RelatedPolymeric Ether Toxins: 33 Structuresand 8iosynthesis. Y. Shimizu, A, V. Krishna Prasad, H-N. Chou, G. Wresnsford, and M. Alarn

Pharmacologyand Toxicology

Effectsof Toxins on the Voltage-GatedVa Channel. 47 M. T, Tosteson

PathomorphologicalStudies on ExperimentalMaitotoxicosis 55 and Ciguatoxicosisin Mice. K. Terao, E. Ito and T. Yasurnoto

Mechanismsof Pharmacologicaland Toxicological Action 7I of Ciguatoxin and Maitotoxin. Y. Ohizumi

Inhibition of SkeletalMuscle Responseto Acetylcholine 79 by Dinoflagellate and Ciguatoxic Fish Extracts. G. Escalona de Motta, J,A. Mercado, T.R. Tosteson and D.L. Ballantine CaribbeanCiguatera and Polyether DinoflagellateToxins: 89 Correlationof BarracudaCiguatoxins with StandardToxins. T.R. Tosteson,R.A. Edward and D.G. Baden

ComputationalModeling of the PolyetherLadder 103 ToxinsBrevetoxin and Ciguatoxin. D. G. Baden, K. S. Rein, M. Kinoshita and R. F. Gawley

Antibody Production and Developmentof a 115 Radioimmunoassayfor the PbTx-2-TypeBrevetoxins. M. A. Poli and J. F. Hewetson

Ecology

Excretionof Ciguatoxin from Moray Eels Muraenidae! 131 of the Central Pacific. R,J,Lewis, M. Sellin, R. Street,M.J. Holmes and N.C. Gillespie

SeasonalAbundance and Toxicity of Gambierdiscustoxicus 145 Adachi et Fukuyo from 0'ahu, 1 lawai'i. E.J.McCaffrey, M.M,K. Shimizu, P.J.Scheuer and J,T. Miyahara

SocietalImpact

Public Health, Epiderniologicaland SocioeconomicPatterns 157 of Ciguaterain Tahiti. R. Bagnis,A. Spiegel,L Nguyen and R. Plichart

Ciguatera:Clinical, Epiderniologicaland 169 Anthropo!ogica1 Aspects. P. Alvarez, C. Duval, T, Colon, G. Gonzalves, E. Martinez, J.Gomez, L. Cornpres,N. Pina and P.L.Castellanos

How CiguateraAffects Canadians 181 E.C.D. Todd

Reporton Ciguatera Fish Poisoning, Owia, November 23, 1985 197 K. Morris

LegalAspects of CiguateraFish Poisoning in PuertoRico 201 J V. Biaggi Preface

Ciguaterafish poisoningis a humanhealth problem that affectsall per- sonsliving in tropicalseas for whommarine fish representa significantsource of food.Ciguatera traditionally was limited to tropicalregions, however, modernimprovements in refrigeration and transport have augmented com- rnercialization of tropicalreef fishes and increased the frequency of thistype of fish poisoningin temperateportions of theearth Sincethe early reports of this poisoningin thetropical Pacific in the17th century, ciguatera has come to have an impact of global proportions, A broad and detailed treatment of the present status of researchon ciguaterafish poisoning is givenin thisvolume in 18papers presented at the Third InternationalConference on CiguateraFish Poisoningheld in I.a Parguera,Puerto Rico, April 30 throughMay 5, 1990.Laboratories, agencies and universitiesrepresented at this Conferencestretched from Australia,Ja- panand the islands of thesouth Pacific to Hawaii,the United States, Canada and throughthe islandsof the Caribbeanto Franceand the Mediterranean Sea.The Proceedings are divided into four sectionscovering; I! thechemis- try of fishand dinoflagellate toxins, ! thepharmacology and toxicology of ciguatoxicfish and dinoflagellatetoxins, including the actionof thesetoxins on membranechannels, their pathomorphologicaleffects on the organsand tissuesof the mouse,and a correlationof toxin potency with computer gener- atedconformational models of thesetoxins, ! ecologicaland fisheries aspects of ciguateraand the toxicdinoflagellate vectors of this poisoning,and ! the societalimpact of ciguaterain FrenchPolynesia, Canada, the DominicanRe- public, PuertoRico and St, Vincentsand Grenadines.These presentations addressglobal issuesdealing with epidemiology,public health,socioeco- nomicand legalaspects of ciguaterafish poisoning,a review of past and current trends, with a description of local and international programs pres- entlybeing implemented to dealwith ciguaterain theseareas. We hope that this volume will be of use to researchscientists working in this field as well as professionalsand studentsin fisheriesand public health interestedin the origin and impact of ciguatera fish poisoning. It would be difficult to name all of the personswho contributed to the successof the Third International Conferenceon CiguateraFish Poisoningancl theproduction of theseProceedings. Thanks are due to all of thecontributors to the Proceedingsfor their help in manuscript review,encouraging comments and patiencewith this process.Special gratitude goes to Mr. Alexis E. Tosteson,Administrative Director of the Conference,for his efforts and dili- gencein the preparationof theseProceedings

La Parguera Puerto Rico T.R. Tosteson Acknowledgements

Variousorganizattons contributed to the Third InternationalConference on CiguateraFish Poisoningand to the preparationof theseProceedings. We would like to acknowledgethe following:

SPOnSOrS U.S. Army ResearchOffice U. S.Army ChemicalResearch, Development and EngineeringCenter SeaGrant College Program of the Universityof PuertoRico National SeaCrant CollegeProgram

COntributorS

Fundacion Educativa Ana G. Mt'.ndez CaribbeanFishery ManagementCouncil IntergovernmentalOceanographic Commission of Unesco FxperimentalProgram to StimulateCompetitive Research in PuertoRico Facultyof Arts k Sciences,Mayaguez Campus, University of PuertoRico Agricultural DevelopmentAdministration, Department of Agriculture, Commonwealth of Puerto Rico Centerfor the Enhancementof Teaching,University of PuertoRico BancoPopular de Puerto Rico Serviciosde Segurosde Salud de Puerto Rico Corporation for the TechnologicalDevelopment of Tropical Resources, Commonwealth of Puerto Rico Chiral Corporation Upjohn Pharmaceutical RegionalCo-ordinating Unit of theCaribbean Environment Program of the United Nations Environment Program,Kingston, jamaica This publicationis a result,in part, of funds sponsoredby the U, S, Army Research Otfice, Grant Num, DAAL03-89-G-011?and NOAA, National Sea Grant College ProgramOffice NSGCPO!,Department of Commerce,under Grant No. UPR-SG-04- F-158-44030 Project No. ti/C-74-10!. Opiruons,interpretations, conclusions, and recommendationsare those of the authors and are not necessarilyendorsed by sponsoringand contributing agenciesand organizations. Recommendationsare those of the authors and are not necessarily endorsedby sponsonng and contributing agenciesand organizations. Chemistry of Fish and Dinof lag ellate Toxins Structure Determination of Ciguatoxin of Moray Eels and Garrfbierdi sees toxicus

T. Yasurnoto', M. Murata', Y, lshibashi', M. Fukui' and A.M. Legrand2 1Facultyof Agriculture,Tohoku University, 1-1 Tsutsumidori- Amarniyamachi,Aobaku, Sendai 981, Japan. InstitutTerritorial de RecherchesMc'dicales Louis Malarde, Papeete,Tahiti, French Polynesia

ABSTRACT Ciguatoxinisolated from visceraof the morayeel Gyrnnofhorax javanicus and its congener tentativecodename CTX-4B! isolated from the causative dinoflagellate Gambierdiscustoxicus were subjectedto structure eluci- dation. The molecular formulae of CslIHssOt9and C6sHs40~swere determined for ciguatoxinand CTX-48respectively, by HR-FABMSand ~H NMR mea- surements.~H NMR signalsof protonsaround a nine-memberedring were broadenedor disappeareddue to a slowconformational change of the ring, but spectraquality was improved by measuringat low temperature.By exten- sive use of rH NMR 2D-correlation and NOE experimentscarried out on 0.35 rng of ciguatoxinand 0,74 mg of CTX-48,the two toxinswere toundto have brevetoxin-likepolyether structures comprised of thirteenrings /6/6/7/7/ 9/7/6/8/6/7/6-spiro-5!. CTX-4Bwas in a formless oxidized than ciguatoxin at two terminals of the molecule. Their relative stereochemistries,except for C2of ciguatoxin,were clarified by detailedanalyses of 'H NMR NOEexperi- ments.The primary alcoholat the terminal of the ciguatoxinmolecule suggestedits usefulnessfor preparinga flourescentderivative or a toxin- proteinconjugate to be used for immunization,

INTRODUCTION Ciguatoxin CTX! is a termfirst proposedby Scheuerfor the principal toxin of the red snapperLutjanus bohar and then for the toxin in the liver of the moray eel Gymnofhorax =Lychodontis! javanicus'~. The purifled toxin was obtainedas fine crystals and suggestedby mass spectralmeasurements to have a molecular formula of either C53H~NOq4112.2! or Cs4H7s024 111.2!. The polyethernature of the molecule was indicated by 'H NMR measurements. However, the chemical structure of the toxin has remained undetermineddue to the extreme difhculty in obtainingthe toxin as well as the complexstructure of themolecule, The toxinwas presumed to be acquiredby fish through the food chain. As to the origin of the toxin, Yasumoto discov- T y'asumoto, et aL eredin collaborationwith the Tahitigroup, the epiphytic dinoflagellate Gtttrtbierdiscustoxicus as the probableprimary source. The toxin detectedin the dinoflagellateresembled CTX in somechromatographic properties but re- mainedambiguous with regard to structural relationship with CTX. ln the presentstudy structureelucidation was carriedout on CTX ob- cd from moray eel visceraand one of the toxins,code named CTX-4B previouslyGT4b!, isolated from G,toxicus, By extensiveuse of 'H NMR2D- correlation and NOE measurements, we determined the structures of the two toxins. They shareda brevetoxin-likeskeleton comprised of thirteenether rings,Their relative stereochemistries were determined, except for C2ot CTX, by detailed'H NMR NOEexperiments. The primary alcohol present at one term}nalof CTXmolecule seemed to be usefulfor fluorescencelabelling and for preparationof CTX-protein conjugatefor immunization,

METHODS Chromatographicallypure CTX 50 Itg! was obtained from viscera 45 kg! of the morayeel Gyrrtnothoraxjavatticus collected in FrenchPolynesia as describedpreviously3. Wild specimensof G. toxicuswere collected in the GambierIslands, French Polynesia, during 1977-1979by scrubbingthe surface of dead corals followed by sedimentationand sieving4~,Isolation of toxins from G. toxicuswas carried out following the method used for CTX with a slight modification. A CTX congener 40 !tg! purified from G. toxicus was usedfor structureelucidations7, This congenercorresponds to CTX-48ac- cording to our new code names proposed for CTX congenerss. NMR measurementswere carried out with a JEOLGSX 400 00 MHz!. Spectra weretaken in pyridine-d>,CDaCN solutions to achievebetter separation of heavilyoverlapping proton signals. Measurements of proton signals were doneat both25'C and at -25'C to observesignals of protonson and arounda nine-memberedring, which due to a conformationalchange causes broaden- ingor evendisappearance of thosesignals, FAB-massspectra were taken on a JEOLJMS DX-303HF mass spectrometer.

RESULTSAND DlSCUSSION

Toxicityartd physicochemical properties Thelethal potency of CTX in mice i.p.! wasestimated to be 0.35Itg/kg, andthat of CTX-4B4 ltg/kg a. CTX wasobtained as an amorphouswhite solid;negative reaction to ninhydrin or Dragendorff'stests; no UV maximum above210 nm; IR

CHsCN,e 22000!j;high resolutionFAR-!VIS m/z 1061.587MH+ calcu- lated for C6oHssOts,1061.504!. Molecularformula No evidencewas obtained by the color reagentsand IR spectrameasure- mentsto support the presenceof nitrogenatoms in CTX. The molecular formulaof CTXsuggested by high resolutionFAB-MS HRFABMS! was sup- portedin a subsequentNMR study. Partialstructures deduced from 'H and ' C KEIR spectracontained at least59 carbonsand thus left only 28daltons unaccountedfor. Hencethe presence of nitrogenor sulfur atomsin CTXwas ruled out. Functionalgroups Comparisonof 'H NMR spectra in pyridine-dsand CDsOD!with those reportedby Tachibanaconfirmed that our CTXwas identical with that of the Hawaiigroups, The presenceof five methyl I singletand 4 doublets!,five hydroxyls,and four doublebonds I transand 4 cis!was shown in the tH NMR spectra.To confirmthe numberof doublebonds, 1 Itg of CTX was subjectedto catalytic hydrogenation. The reduction product gave a cationated ion M+Na!+ at rn/z 1143in FAB-massspectra, suggesting the presenceof five double bonds in the CTX molecule.

4 D 35 3.0 20 j.5

Fig. 1. NMR Spectraof CTX-4B, 'H NMR spectra of CTX-4B 00 MHz! in CDsCN at 20'C. Two inset partial spectrawere measuredat -25'C. T. Yasumoto,efaf.

The olefinicunobserved by measurementsat 25'C, appearedin spectra measuredat -25'C Figure 1!. As there was no sign of carbonylin the lR spectrum,we inferred that unsaturation in the CTX molecule was ascribable mainlyto C=Cbonds and ether rings. HenceCTX probably has thirteen ether ringsfused in a laddershape,mimicking brevetoxin-B and yessotoxin'o 6.

pig.2. NMR-COSYMap of CTX-4B. H- H COSYmap of CTX-4B00MHz! in pyridine-dsat 25'Cand the assign- ment of cross peaks, Figures denote the carbon numbers of coupled protons givinga crosspeak. Asterisks express signals and cross peaks due to impurities. »guresin parentheseswere assigned with additional use of 2D-HOHAHAand relayedCOSY Structure ot Moray Eel Ciguatoxin

Prntottcortrtecfivify tH NMR chemicalshifts and couplingconstants were almostidentical between CTX and CTX-4Bexcept for the two terminal parts of the molecules. Therefore,the structure elucidation of the common part will be discussedusing thedata obtained with the latter.The skeletal structure was establishedmainly on thebasis of 'H-'H 2D NMR dataobtained from COSY Figure2!, relayedCOSY, and 2D HOHAHAexperiments. Interpretation of 'IH-'IH COSYmeasured in pyridine-dsis shownin Figure 3A, First,four fragments,CI-C32, C34-C38, C40-51 and C53-55 were assigned. The positions of OH groupswere clarihedby the COSYmeasured in pyridine-ds,2D- HOHAHA experimentsrevealed multtple relayed couplings of H-37/Me-57

re

Fig. 3. CTX-45: Resultsof 2D Correlatton and NOE Experiments. 3-A,Heavy lines indicate the connectivities assigned on the basis of 'H-'H COSY, relayedCOSY and 2D-HOHAHAat roomtemperature, while brokenlines de- note those as-25 C. 3-6, Arrows and figures denote irradiated protons tail!/ observedprotons head! in NOE differenceexperiments and NOE's in percent- age measured in CD3CN at -25'C. The figures with an asterisk are those measuredin pyridine-ds at -25'C. and Hbrevetoxins'>.MMp calculations also supportedthis T +asurnoto,et trl.

onclusion.'H NMRspectral simulations done for H2-38/H-39 Me-57!/H- 40 andH-39/HZ-40/H-41 using parameters obtained from the MM2, calcu- lationsand COSY exPeriments enabled us to establish the structure of' the ring 1;ncludingthe equatorial orientation ofMe-57. A longrange coupling of Me- 56/H-32, anda largeNOE on Me-56/H37 allowed us to connectH-32 through H 37 Couplingconstants of H-34 throughH-37 werealso typicalfor a 2 23,5,6-pentasubstituted tetrahydropyran. The presence of a signalat 8109.2, typtcaifor a spiro carbon of5/6 membered rings, in a ' C NMRspectrum and an >OE on MMO/H-53revealed the spiro-fused ring M. Etfter linkageattd stereochemistry positions of ether rings were clarified by analysesof coupling constants betweenangular oxymethine protons ca,9 Hz! and by NOE experimentson angular protonsor a methyl Me-56!.Geometries of the doublebonds were assignedby their vicinal couplingconstants. The couplingpattern of H-11 triplet,9Hz! indicated the 11-OHto be equatorial,and a couplingconstant of H-51 doublet,11Hz! revealeddiequatorial orientation of Me-59andMe-60. A largerNOE was observed on Me-60 than on H-54 when H8-53 was irradiated and this order in NOE intensities was reversed when Ha-54 was irradiated. Thecoupling pattern of H8-53 dd, 10, 2 Hz! and H-53 dd, 10, 4 Hz! correspondedto trans-vicinal and cis-vicinal couplings, respectively, on tetrahydrofuran.Thus, the hydroxyl group at C-54 was shown to be a 8- substituent.A probableW-type coupling observed between H-32 and Me-56 suggestedan equatorialorientation of OH-32. This was further supportedby a large NOE on H-32/H-34, A prominent NOE on H-48 observedupon irra- diation of Me-58 suggesteda 8-orientation of Me-58, The a orientation of

R> 7= 8> CHp--CH- Rr=H

~'8 4. Structuresof CTX and CTX-4B. Structureof Moray Eel Ciguatoxin

Table1. 'H NMR Che nicalShifts' and CouplingConstants' of C guatoxin' and GT4b2

poritn 2 bott ~ re ! pattern poelto l pettoro I pe 35 1.9! 38 !.54 3, 11. 8! 1.53 ! 4.80 < e ! 4.ee < ~ ! 1.843. 10. 4! 1,83< - ! 5.86 - ! 5.901, 3, 2! I. 91 q,t, 7. 8! 1.90 -! 5,8! - ! 5.77 ]7 5.76 3, 2, 2! 5.73 <3r.d, 13! 1.94 q.t, 6, 11! 2.01 - ! 16 s,ee < - > 5.89 r.d, ! 51 ! .60 21 5.673. 2. 2! 5,67 br.d, 13! 54d8,1,68 - ! 4.86 e ! 22 6.10 br,d. 13! e,led br.d. 13! da.>.90 - ! 23 4,03 br.d, ee. 8!' 4.03' - > l,el - ! 4.180, 1> 3.6ed 3.68 - ! 3.$8 -! 4.19 0, 5! 25 2.20 - ! 2,20' - 56 1.37 ~ ! 1.37 ~ ! 3.014 - ! 3,00 ~ ! 57 0.92 0,92 < 8 ! le 6.05' 3.30< 8 ! 27 6,07 ii, Ii. 5!" e,07d - ! 59 1.28! 1.32 6 ! le 2.36 <-! 2,36 - ! 60 0.97! 1.24< 7 ! 2.9e' < 8 > 2.98 ~ ! I-CW 6.40 e. 4! 29 3 Wed br.d, 9! 3.86 < - ! 2-011 6<1 d ! 30 3 68 <9. 8. 7! 3.67 ! 11-CW7 34. <2! 7 32 2 ! 31 2.65 <13,8. 6! 2.66 - ! 32-01<5. 29 I! %.29 3 ! 2.70 ' 2.70 - ! 47m e.le ! 6,77 3 ! 32 4,16 8, e, I! 4.16 ! 6.53< 4 ! 3.34 2, 4!d 3.33 - !

4ProtonNMR spectrawere ineasured with a JEOLSGx 400 00MHz! spec- trometerin pyridine-dsat 25'Cexcept for thosewith superscripts,d, e, g and h. bMultiplicity s, singlet;d, doublet:t, triplet; q, quartet;m, multiplet!and coupling constantsin Hz. 'Couplings were unassignabledue to heavy overlappingsof signals.OMeasured in pyndine-dsat -20'C.'Obtained from NOE differencespectra in pyridine-dsat -20'C.'Obtained from decoupling differencespectra. sCoupling constants in CDSCNat 25'C. !!Couplingcon- stants in CDSCN at -25'C. T. Yasumoto,ef ak

OH~7 was supportedby NOE's detectedboth on H-44/OH-47 and H-49/ OH-47. The orientation of C53 to rmg L was assigned to be equatorial by NQE's on Me-60/H53 and on H48/H-55. The structural difference between CTX and CTX-48 was readily eluci- dated by COSYanalyses. CTX-4B of G. toxic@ahas a trans-butadieneat one terminusof the moleculeand lacksa hydroxyl group at C54at the other end of CTX. Thus, we successfullydetermined the structures of CTX and CTX-4B Figure4!, except for the stereochemistry atC2 of CTX,with nomore than 1 mg of material. i H NMR chemicalshifts and coupling constantsof CTX and the congenerare given in Table I, Further details of spectraldata to support the structureassignments will be presentedin a coiningpaper . Thedifference in the extentof oxygenationbetween the two toxinsled us to hypothesizean oxidative modification of toxins during the course of the food chain. It is interesting to note that lethal potency of CTX-4Bincreased aboutten fold when it was oxidized to CTX, That may explain, at least partly, why the livers of inoray eelsare so toxic comparedwith thoseof herbivorous fish, Confirmation of the presenceof a priinary hydroxyl group at one termi- nus of CTX allowed us to designa fluorornetricHPLC methodfor detectionof nanogramsof CTX", Effortsare beingcontinued to elucidatestructures of other toxins. Acknowfedgements We are mostgrateful to ProfessorJ. Roux,Director of the InstituteLouis Malarde,and to Dr. R. Bagnisfor their continuedsupport of the ciguatera project, Thanksare also due to Mr, J. Bennettfor collecting the dinofiagellate and fish, and to Mr. P. Cruchet for his assistancein toxin purification. Discus- sionsand suggestionsprovided by professorM. Hirama of Tohoku University aregreatly acknowledged.This work wassupported by the FrenchMinistry of Researchand Industry,Pasteur Institute of Paris,French Polynesia Govern- ment, and by a Grant-in-Aid for Overseas Researchof the Ministry of Education,Science and Culture, Japan.

LITERATURECITED Scheuer,P.J., W. Takahasi,J. Tsutsumiand T.Yoshida. 1967. Cigua- toxin: Isolation and chemical nature. Science 10:1267-1268. Tachibana, K., M. Nukina, Y-G. Joh and P.J Scheuer. 1987. Recent developments in the molecular structure of ciguatoxin. BioL Bull. 172:] 22-127. 3. Legrand, A,M., M. Letandon, J. N, Genthon, R, Bagnis and T. Yasumoto, 1989. Isolation and some properties of ciguatoxin, J, Appl. Phycol.1:183-'I 88,

10 Structure of Moray Eel Ciguatoxin

4. Yasumoto,T., I. Nakajirna,R. Bagnisand R. Adachi.1977. Finding of a dinoflagellateasa likelyculprit of ciguatera.Bull. Jpn, Soc. Sci. Fish. 43:1021-1026, 5. Yasumoto,T., I. Nakajima,Y. OshiIna and R. Bagnis, 1979. A newtoxic dinoflagellate found in associaflonwith ciguatera. In: Toxic Dinoflagel- lateBlooms, Taylor and Seliger eds.!. Elsevier, New York. pp, 65-70. 6. Murata,M., M. Kumagai,J.S. Lee and T. Yasumoto.1987. Isolation andstructure of yessotoxin,a novel polyether compound implicated in diarrheticshellfish poisoning. Tetra. Let. 28:5869-5872. 7. Murata, MA-M, Legrand, Y. Ishibashi, M. Fukui and T. Yasumoto.1990. Structure and configurationsof ciguatoxinfrom the morayeel Gymuvthorax javanicus, and its likelyprecursor from the di- noflagellateGambierdiscus foricus. J. Am. Chem, Soc. In press. 8, Legrand,A-M., M. Fukui,P, Cruchet,Y. Ishibashiand Y. Yasumoto.1991. Characterization of ciguatoxinsfrom different fish spe- cies and wild Gambierdiscustoxicus, In: Proceedingsof the Third InternationalConference on CiguateraFish Poisoning, T.R. Tosteson ed.!.Polyscience Publishers, Quebec, Canada. See pp. 27-36, these Pro- ceedings 9. Tachibana,K, 1980.Structural Studies on MarineToxins. Ph.D. Disser- tation, University of Hawaii, 156pp. 10. Lin,Y.Y., M. Risk,S.M. Ray, D.V, Bngen and J, Clardy. 1981. Isolation and structure of brevetoxin B from the "red tide" dinoflagellate Pfychod~scusbrevis Gymnodinium breve!, J.Am. Chem. Soc. 103:6773-6775. I '1. Shimizu, Y., H,N, Chou, I I.Bando, G.V. Duyne and J. Clardy.1986. Structure of brevetoxinA GB-1toxin !, themost potent toxin irt the Floridared tide organismGymnodinium breve Ptychvdiscus brevis!,J. A. Chem.Soc. 108:514-515. Extraction and Purification of Toxic Fractions from Barracuda Sphyraena barracuda! Implicated irt Ciguatera Poisoning

PedroM. Camboa',Douglas L. Parkz, and Jean-Mare Fremy3 1Departmentof Chemistry, Universidad del Este, Cuidad Bolivar, Venezu- ela, 2Divisionof ContaminantsCheniistry, Center for Food Safety and AppliedNutrition, Food and Drug Administration, Washington, DC20204, Currentaddress: Department of Nutrition and Food Science, University of Arizona,Tucson, AZ 85721, CentreNational d'Etudes Veterinaires et Alimentaires,Laboratoire Central d'Hygiene Aliinentaire, 75015 Paris, France.

ABSTRACT Samplesof severalbarracuda fish Sphyraenabarracuda! implicated in ciguaterapoisoning were analyzed for toxicity. Toxic fractions were isolated and purified using column and thin layer TLC! chromatographic techniques.Crude extracts of fishsamples were prepared by extractingwith acetoneand partition purification with hexane,inethanol, and chloroforin,Partially purified extracts were applied to a silicagel coluinn and toxic fractionseluted with cMoroform:methanolmixtures of 100:0,95:5, and 50:50. The 5% methanol/chloroform was most efficient in separating toxic fractions, The fractions were purified further using silica gel TLC plates.Developing solvents used were chloroform:methanol at 95:5, 90:10, and50:50. Visualization of separatedcompounds was accomplished by spray- ingthe plate with sulfuric acid:acetic acid;methanol 1:I;9! mixture and heating theplate at 110 C for3 minutes.Three separate toxic fractions were observed. Toxicpotentials were confirmed in three fractions using intra peritoneal i,p.! injectionin miceand rectal temperature monitoring. All fractionswere tested forthe presence of decomposition products histamine, putrescine, cadaver- ine! and okadaic acid using high performance liquid chromatogra- phy, Okadaicacid was detected in one of the toxic TLC fractions.

INTRODVCTlON Knowledgethatcertain fish and shellfish can become poisonous and cause death or illness when eaten has been known for centuries, but only more recentlyhas the clinical nature and effology of thesefoodborne intoxications becomeapparent'. Garnbierdiscus toxicus has been identified as a sourceof P. M. Gamboa, et al.

ciguatoxinand more recently several strains of Prorocentrumsp.z 3. Theability to cultureselected organisms under large scale conditions has been accom- plished D.L Park,unpublished data; R, Dickey,unpublished datas. Over 400species of bony fish havebeen reported to causeciguatera 7. lt is believed that fish becomeciguatoxic through the food chain; herbivorous fish feed on algaeand detritus of coralreefs, then the largerreef carnivores prey on these herbivores"9. There appears to be a widedistribution of thetoxin in various hshtissues'"", Although ciguatoxin has been crystallized following extrac- tion from the livers of ciguatoxiceels, it hasonly recentlybeen completely structurallyidentified'z's Thereappear to be severaltoxins involved, i.e., ciguatoxin,rnaitotoxin, okadaic acid, and scaritoxin, associated with toxigenic dinoflageilatesor fish3'4's.The mouse assay has been used for ciguateraas well as the feedingot cats,chickens, mosquitoes and mongooses'7-z'.More recentlyseveral immunological methods have been developedzz z-', This study focusedon theanalysis of barracudafish Sphyraenabarracuda! implicated in a ciguatera poisoningoutbreak which occurredin Puerto Rico,

MATERIALSAND METHODS FishSamples Barracudafish Sphyraenabarracuda! implicated in a ciguaterapoisoning outbreak were harvested 4 miles off the southwest coast of Puerto Rico.Samples of thesuspect fish wereobtained by theFood and Drug Admin- istration, San Juan, Puerto Rico, from the RosasFish Market in Puerto Real, CaboRojo, PR 00623, on May 29, 1985,Following the poisoningoutbreak, suspectfish were maintained at -18'Cuntil ready for analysis. Extractionand Purification of 8arracudaToxins A modification of the methodreported by McMillan and co-workers was usedto extractthe toxins Figure 14. Thepartially purified extracts were dissolvedin methanol MeOH!, then purified further on a silicicacid column Mallinkrodt,100 Mesh! previously activated at 100C for 1 hour,as reported byNukina and co-workerszs. As solution solvents, binary systems of chloro- form-methanol00:0, 95:5 and 50:50! were tried in sequence. Thin Iayer Chromatography Silica gel 60 plates x 20 cm x 0.25mm and 20 x 20 crn x 0.25mm! were activatedat 100'Cfor 10 minutes. The plates were developed in a closednon- equilibrated system with developing solventsof 5, 10, and 50% MeOH in chloroform.Visualization of separated compounds was accomplished by sprayingwith a H>SO4-aceticacid-MeOH :I:9! solution and heating at 110C for 3 min.To heat the 20 x 20cm plates, the extraction zone was protected afterdevelopment> with glassplates 8 x 20an!, andthe plates inverted to exposethe developedside of the TLC plate. Applicationof theextract on the

14 EXtraCtianand PurifiCatiOnOf CiguatOXinS

FISH SAMPLE 00 G! fCDDNKD 15 MIN. AT 15 Psz!

HOMOGENIZEDNZTH ACETONE 00 ML!

RESIDUE EXTRACT DISCARD! I FILTER

DUE FILTRATE DISCARD! I CONCEHTRATK fUNDER !fi! I ADD METHANOL 00 Mt.!

SOLUTZOH !tEXANE EXTRACT DISCARD! CONC'ENTRATE fUNDER !f,!

CHLOROFORMPARTITION

CHLOROFORMEXTRACT METHANOLSOLUTION DISCARD! CONCKNTRATE UNDER Nz! I PARTIALLY PURIFIED EXTRACT STORK UNDERREFRIGERATION! I TLC 1 HPLC

Figure l. Extraction of BarracudaCiguatoxin. Extractionmethod used for isolating ciguatoxic fractionsfrom barracuda Sphyraena barracuda!. P. M. Gamboa,et al. TLCplate 95 lil! wasdone with a CamagLinomat Ill continuoustype dosemeter,equipped with a 100p.l syringe Camag! using a pressure of28 psi anda workingspeed of 9 sec/liLThe purified fractions were removed from theTLC plate by scraping off the silica gel and aspirating into a 10ml syringe witha Millex-SR Millipore Co, No. SLS2025 NS! filter attached to it. Liqut'dChromatography Crudeextracts of toxicand non-toxicfish were analyzedfor histamine usingthe AOACfluorometric method and for putrescineand cadaverine followingthe gas liquid chromatographic method of Staruszkiewiezand Bond26,Six purifled fractions isolated by TLC were analyzed for okadaic acid OA!using the method of Leeet al.>7. isolates were evaporated todryness and redissolvedin 500itl chloroform,100 lil of which were usedfor derivatization

34

t- 54 32 I 30 28 1015 20

TIME HQUR!

Figure2. Effectof CiguatoxicBarracuda Extracts on Rectal Temperature and Toxicityin lattice.Rectal temperature and toxic responses in mice dosed i.p,! with crudeextracts of ciguatoxicbarracuda

16 Extractionand Purificationof Ciguatoxins with the ADAM reagent. BeforeLC analysis,the derivatizedisolate was purifiedfurther with a silica gel SEP-PAKcolumn conditionedwith hexane- chloroform. The column was washed with hexane-chloroform, then chloroform. Derivatized OA was eluted off the column with chloroform- methanol 95:5!. Parametersfor the LC systemwere; oven temperature, 35 C; columns,Lichrocart 250-4, Superspher 100 RP-18 ltm pore size,250 mm length!and 4.4 Lichrocartpre-column RP-8 pm pore size!;eluting solvent, acetonitrile-water 80:20!;flow rate, 1 ml/min. A fluorescencedetecfion sys- tem 65 nm, excitation; 412 nm, emission! was used to monitor column fractions.

Mouse Assay The mousetemperature depression toxicity test asreported by Sawyeret al. wasemployed using CD-1, femalemice CharlesRiver! weighing between 19 and21 grams>s.One ml of extract,suspended in 5%Tween 80 solution preparedwith 0.9% NaCl solution!,was injected intraperitoneally i.p.! into the mouse and toxicityvalues monitored by takingrectal temperaturemea- surementsevery hour for 16hours with a digital thermometer YSI type!. Zero time temperaturewas taken just beforedosing.

RESULTSAND DlSCUSSION 0 Therewas a rapiddecrease in the rectaltemperature values of almost 10 C whenthe mice wereinjected with crudetoxic extractsof the contaminatedfish sample Figure 2!. Experimentalanimals exhibiting this response usually did not survive more than 4 hours with the typical symptoms of convulsions followedby respiratory failure, Figure 3 showsthe response of animalsdosed 0 with non-contaminatedextracts. A decreasein temperatureof around4 C wasobserved during the first hour; however,the animalrecovered and body temperaturereturned to normal. McMillan and co-workers reported that in mice treated either with control or contaminated extracts where the final solventwas chloroformthis "knockdown" phenomenonin rectal temperature wasobserved24, Traces of chloroformincorporated incorn oil andinjected to the miceelicited a similarphenomenon, Withrespect to the polarityof the toxicfraction s!isolated using a silica gelcolumn, the toxins were cluted with a 5%MeOH-chloroform system. For the other elution solvents,0 and50% MeOH-chloroform, the animalsexhibited notoxic response when exposed to anaverage dose representing 1.027 and 0.472rng extract/g fish flesh, respectively. Theaverage dose given the animals fromthe 5% MeOH-chloroform eluate was 1,37 rng extract/g fish flesh. This is inagreement with publishedresults of Lewisand Endean and Nukina et al., who havedemonstrated isolation of the toxicfraction from crudeextracts by slightlyincreasing the polarity of chloroform>'~~.These results demon- stratethat the toxic fractionis eluted with the methanol:chloroform ;95!

17 P. M. Gamboa, et aI.

38

37

36

L x 35 34 1015 20

TIRE HotjR!

Figure3. Effectof ControlExtracts on RectalTemperature. Effectof non-ciguatoxicbarracuda extracts on rectal teinperature in mice, Rectal temperatureresponse in roicedosed i.p.! with crudeextracts of non-ciguatoxic barracuda. binarysolvent system. An equallyimportant factor to noteis thatthe animal's deathoccurs following a rapid decreasein therectal temperature in the order of 10'C The toxicfraction eluted with 5%MeOH/chloroform was further sepa- ratedusing silica gel TLC plates.Excellent resolution of 5 componentsof differentpolarity was obtained,The toxic responsesin the mousefor each fractionare presented in Figures4 and5, lt is iinportantto note that to be considered toxic, the fractions must both cause a decrease in the rectal tem- perature as weil as produce death at the end of the evaluation time 6 hours>.Nukina and co-workersworked with a purifiedextract eluted with 10%methanol:chloroform froin the silicic acid column~s. They observed the existenceof two toxic fractions. One acidic fraction more polar! eluted with

18 Extractionand Purification of Ciguatoxins methanol:H>0 I:1! from an aluminumoxide columnand anotherfraction consideredneutral less polar! which eluted with methanol:chloroform l;Il!. Figures4 and5 showthat, utilizing the TLC system,it is possibleto have excellent resolution of at least five ! fractions, three ! of which caused death in mice at the concentrationstested. These results suggest that it is possibleto distinguishbetween the hypothermicfactor and the toxicfactor, It also shows that more than one compound is involved with the ciguatera phenomenon.This would explainthe array of multiplesymptoms associated with humanintoxication caused by theconsumption of ciguatoxict'ish. Non-detectablelevels of histamine, putrescine,and cadaverine were ob- served in the crude extracts of toxic and non-toxic fish analyzed. Five TLC

CONTROL TWEEN 80! FRACTION 1, 0.002 MG/G FRACTION 4y 0 019 MG/G

38

34

32

30 CLW 28

10

TIE HOUR! Figure 4, Purified Ciguatoxin Fractions: RectalTemperature Response and Toxicity. Rectal temperature and toxic responsesin mice dosed ip! with isolated fractions followrng column and thin-layer purification of extracts from ciguatoxic fish barra- cuda!, P.M,Gamboa, efat, COHTROt. THEEk 80!

FRACTION 3, 0. 014 MG/G FRACTIOk 2, 0, 009 HG/G ~ ~ ~ ~ ~ ma 0 Sot.VERT FROkTR 0. 125 HG/G

38

36

Lx 34 C 4 X IIEI3032 O 15

20

TIME HOUR!

Figure5. Purified Ciguatoxic Fractions:Rectal Temperature Response and Toxicity. Rectaltemperature and non-toxic responsesin micedosed ip! with isolated fractions following column and thin-layer purification of extractsfrom ciguatoxic fish barra- cuda! isolatesfrom the toxic fish Figures4 and5! were subjectedto HPLC analyses for OA. One of the isolates toxic to the mouse Fraction 1, Figure 4! tested positive for OA. Chromatogramsof this isolate,standard OA, anda samp]eof mussel containing OA as a positive control, are shown in Figure 6, Prorocentramlima and P, concavum,toxigenic dinoflagellates implicated as sources of ciguatera poisoning toxins have been reported tn produce OA " . P. !ima is widely distributed in the Caribbean Sea which allowsfor thepossibility that OA is takenup by herbivoresand passedup the food chain to carnivorous fish such as the barracuda. OA could contribute to

20 Extractionand Purification of Ciguatoxins thediarrhea frequently seen inciguatera fishpoisoning patients. This isthe firstreport of OAin fishtissue associated with a ciguaterapoisoning outbreak.OA has also been implicated as one of thetoxins responsible for diarrhoeicshellfish poisoning DSP! outbreaks', In 1983,Hoffman and co-workers published a study where they related thedegree of responseof the animal mice! when the dose injected variedl~.Among the most important symptotnatic aspects, they pointed out a

JL i I I t I I '.' c'I t ,0I Iat t Jo rr I'O 3 I. tarr a rtr I Toiicov I tlat 2 atilt octa O' S rrl/acr 3 'tacc aacid I ctar1 IXIXI tatpari 2I onmIt a nr OIIo contmlt Figure6.Chromatographic Comparison: Ciguatoxic Extracts, Shellfish Extracts and OkadaicAcid. Liquid Chromatograrns of par ta!lypurified ciguatoxic fish barracuda!extract,tnussel extract containing okadaic, and okadaic standard.

typicalbehavioral trend, characterized primarily bydecreases inrectal tem- perature,reflexes andmotor activity, Inthis study, the authors established an outlineofsymptoms andcategorized them to different degrees oftoxicity. The firstfour categories were related directly to rectal temperature. Sawyer etal, studiedthe effect of thetoxins associated with the ciguatera phenomenon and itsrelation with the changes in temperature ofanimals treated with partially purifiedextracts". Their results demonstrated theinability of the mouse to controlits corporal temperature when treated with ciguatoxic extracts. The authorsproposed that ciguatoxins area chemicallyrelated group of com- poundswith unique pharmacological properties and that the rectal temperaturemeasurement represents a viable alternative to difterentiate ciguatoxinsfrom the other biotoxins of marine origin.

21 P. M, Garnboa, et al,

Figure2 providesevidence on the toxic nature of the fish implicated in the ciguaterapoisoning incident. Another fish specimenfrom the samelot was extractedand testednegative to the mousebioassay Figure 3!, These results support the theory that for a given ciguaterapoisoning outbreak, not all the fish would be contaminated. This highlights the coinplexity and difficulty in setting up an effectivefish safetymonitoring program for ciguatera. Actcrtowledgem crt ts Agricultural ExperimentStation Journal Article No. 7296

LITERATUREC ITED 1. World Health Organization. 1984.Aquatic Marine and Freshwater! Biotoxins, Environmental Health Criteria 37, Geneva, Switzerland. 2. Bagnis, R., S. Chanteau and T. Yasumoto. 1977. Decouverte d'un agent etiologiquevraisemblable de la ciguatera.C.R. Acad, Sci. Paris!. 28: 105- 108. 4 3. Tindall, D.R., R.W. Dickey, R.D. Carlson and G. Morey- Gaines. 1984, Ciguatoxic dinoflagellates from the Caribbean Sea. In: SeafoodToxins. E.P.Ragelis, ed.!. ACS SymposiumSeries, 4262, Arneri- canChemical Society, Washington, DC, pp, 225-240. 4. Babinchak,J., D. J. Jollow, M.S. Voegtline and T.B. Higerd. 1985. Toxin production by Garrtbierdiscustoxicus isolated from the Florida Keys, Mar. Fish. Rev. 48: 53-64. 5, Durand-Clement, M. 1986, A study of toxin production by Gambierdiscus toricus in culture. Toxicon 241!:1153-1157. 6. Halstead, B,W. 1978. Poisonous and Venomous Marine Aniinals of the World, Vol. 2. Vertebrates.U.S. Governinent Printing Offices,Washing- ton, DC, 330 pp, 7. Banner,A,H. 1975. Ciguatera:A diseasefrom coral reef fish, In: Biology and Geologyof Coral Reefs, O.W.Jones and R. Endean,eds.!. Academic Press,New York, pp, 177-213, 8. Legrand, A-M., M. Fukai, P. Crachet, Y. Ishibashi and T. Yasumoto.1992. Characterizationof ciguatoxinsfrom different fish spe- cies and wild Garnbierdiscustoxicus. In: Proceedingsof the Third International Conferenceon CiguateraFish Poisoning T.R. Tosteson, ed.!, PolysciencePublishers, Quebec, Canada. See pp. 27-36,these Pro- ceedings. 9, Vernoux, J.P., L.P. Magras, S,A.E, Anadaloussiand N. Riyeche. 1986. A surveyof comparative toxicityin the ciguaterafood chain on Saint BarthelernyIsland in the Caribbean.Bull. Soc.Path. Ex. 79;275-283. 10. Vernoux,J.P., N, Lahlou, S,A.E, Anadaloussi,N. Riyeche and L.P. Magras.1985. A studyof the distributionof ciguatoxinin individual Caribbeanfish. Acta Tropica42:225-233.

22 Extraction and Purification of Ciguatoxins

11. Lewis, R.J.and R. Endean. 1984. Ciguatoxin from the flesh and viscera of the barracuda,Sphyraetia jello. Toxicon 22!:805-810. 12, Yong-Goe,J. andP,J, Scheuer. 1986. The chemical nature of scaritoxin. Mar. Fish, Rev, 4S:19-22. . Murata, M., A-M. Legrand, Y. Ishibashi and T. Yasumoto. 'I989. Structureof ciguatoxin and its congener.J. Ain. Chem. Soc. I 'I I:S929-8931, 14, Gamboa, P.M, and D.L. Park. 1985. Fractionation of extracts of Proroceritrumlima, Gambierdiscustoxicus and ciguatoxicfish using counter current chromatography. Paper presented at 2nd Conference on Ciguatera,San Juan, I'uerto Rico, April 23-25, 15. Vernoux,J.P. and S.A.E. Andaloussi, 1986, Heterogeneityof ciguatoxins extractedfrom fish caught at the coastof the FrenchAntilles. Biochemic 68:287-291. 16. Yasumoto, T., M. Murata, Y. Ishibashi, M. Fukai and A-M. Legrand. 1992, Structure determination of ciguatoxins of Moray Eels and Gambierdiscustoxicus . In: Proceedings of the Third Interational Conferenceon CiguateraFish Poisoning T.R.Tosteson ed.!. Polyscience Publishers,Quebec, Canada. See pp. 3-12,these Proceedings. 17. Hoffrnan, P.A., H. R. Granade and J,P, McMillan. 1983. The mouse ciguatoxin assay:A dose-responsecurve and syinptomology analysis. Toxicon 21: 363-369. 18. Pompon, A., E, Chungue, I, Chazelet and R, Bagnis. 1984. Ciguatera: Une methode rapide,simple et fiable de detectionde Ia ciguatoxine,Bull. World Health Organ.,62!:639-645, 19. Vernoux, J,P. and N. Lahlou. 1986.The chick ciguatoxin bioassay:A symptomology and sensitiveanalysis. Bull. Soc.Path. Ex. 79 l40-146. >20. Chungue,E., R. Bagnis,and F, Pare. 1984.The useof inosquitoes Aedes aegypti! to detect ciguatoxin in surgeon fishes Ctenoc/raetusstriatus!. Toxicon 22: 161-164. 21. Bagnis,R, and G, Fevai. 1971, La ciguatera feline experirnentalea Ta- hiti, Rev. Med. Vet., 122: 229-238. 22. Hokarna, Y., L.II. Kimura, M.A. Abad, L. Yokochi, P.J. Scheuer, M, Nukina, T. Yashumoto, D.G. Baden, and Y. Shimizu, Y. 'I984. An en- zyme immunoassay for the detection of ciguatoxin and competitive inhibition by related natural polyether toxins. In: SeafoodToxins E.P. Ragelis,ed.!. ACS SymposiumSeries ff 262,American ChemicalSociety, Washington,DC. pp. 307-320. 23. Emerson, D.L., R.M. Galbraith, J.P. McMillan and T.B. Higerd, 1983. Preliminary immunologic studies of ciguaterapoisoning, Arch, Intern, Med. 143: 1931-1933, 24. McMillan, J.,H. Ray and P. Hoffman. 1980.Cigua.tera fish poisoningin the UnitedStates Virgin Islands:Preliminary studies. J. Coll. Virgin Is,

23 P, M. Gamboa,et aL

6;84-107, 25, Nukina,M., L. Koyanagiand P. Scheuer,1984. Two interchangeable formsof ciguatoxin. Toxicon 22: 169-176, 26. Staruszkiewiez,W.and J, Bond. 1981, J. Assoc. Off. Anal.Chem. 64: 584. 591, 27. Lee,J.S., T. Yanagi, R. Kenma and T. Yasumoto. 1987, Fluorirnetric de- terminationof diarrheticshellfish toxins by high-performanceliquid chromatography,Agric. Biol. Chem., 51: 877-881, 28. Sawyer,PD. Jallow, P. Scheuer, R.York, J, McMillan, H. Withers,H, Fudenbergand T. Higerd. 1984, Effect of ciguatera-associated toxins on bodytemperature in mice.ln: SeafoodToxins E. Ragelis,ed.!. ACS SymposiumSeries ¹262, American Chemical Society, Washington, DC. pp, 321-329. 29. Lewis,R.J. and R, Endean.1983, Occurrence of a ciguatoxin-likesub- stancein Spanishmackerel Scomberomorus cammersani!. Toxicon 21: 19-24. 30. Dickey,R.W., S,C. Bobzin, D.J, Faulkner, F,A, Benesathand D. Andrzejewskb1990, Okadaic acid from Prorocentrumconcavum, Toxicon 28: 371-377. Characterization of Ciguatoxins from Different Fish Speciesand Wild Gambierdiscustoxicus

A-M. Legrand',M. Fukui~,P. Cruchet',Y. Ishibashirand T. Yasumoto2 I lnstitutTerritorial de Recherches IVIedicales Louis Malardd, B,P, 30, Papeete,Tahiti, FrenchPolynesia. Faculty of Agriculture,Tohoku Univer- sity, 1-1Tsutsumidori-Amarniyamachi, Aoba-ku, Sendai981, Japan.

ABSTRACT Subsequentto the isolationand structure elucidation of ciguatoxin CTX! and its congenerfrom the moray eel Gymnothoraxjavanicus, Bleeker, and the causative dinoflagellate Gambierdt'seesfoxirus, Adachi and Fukuyo, we care- fully examined all the toxic components in moray eel, parrotfish and G. toxicus. As many as 21 toxins, including the above two toxins, were detected and characterizedby high performanceliquid chromatography HPLC! and mass spectra. Thesetoxins vere separableinto four groups according to the order of elution from a reversed phase column. Polarities of toxins in G. toxr'cpsand flsh suggestedthat toxins underwent oxidative modifications dur- ing thecourse of thefood chain, CTX wasthe major toxin in the moray eelbut was insignificant or absentin the parrotfish and G, toxicus,Detection of CTX in other carnivorous fish suchas red snappersand grouperswas facilitatedby an HPLC method which detected fluorescenceafter derivatizing CTX with anthroylnitrile.

INTRODUCTION Ciguaterais an importanthealth and economic problemin the South Pacific area. In FrenchPolynesia, the Tuamotu, Gambierand MarquesasAr- chipelagosare the most concernedareas. A number of fish of varying feeding habitsare involvecl in the intoxication.Patients may show a wide rangeof symptomsincluding neurological, cardiovascular and gastrointestinaldisor- ders '. Becauseof the diversityof the complexhuman symptoms,there has beena questionas to whetherthe ciguatoxin CTX!isolated from the moray eel,purilied and designated by Scheuer'sgroup was solely responsible for the intoxication~.Another question which remained unsolved was whetherthe ether-soluble toxin detected in the epiphytic dinoflagellate Gambierdiscus toxicus,Adachi and Fukuyo, the putative origin of ciguatoxin,was chemically relatedto ciguatoxins4. AsG. toxicus produced little or noCTX under culture conditions,we isolatedpure CTX from themoray eel viscera s. Ether-soluble A-M Legrand,et aL toxinsof G.toxicus were also purified from wild specimenscollected in the GarnbierIslands. As describedseparately in theseProceedings, our previous studiesrevealed that CTX and a congenerisolated from C. toxicusshare the sameladder-shaped skeleton comprised of thirteenether ringss ~, Duringthe courseof toxinpurification, we noticed the presence of other toxins in addition to CTXand the congenerss. The difference in theextent of oxygenationbe- tweenthe two toxins also suggested that a seriesof oxidativemodifications of thetoxins during food chain transmission might yield a numberof congeners. In thepresent study, detailed analyses of toxincompositions were carried out onG. toxicus and fishes at different stages of thefood chain. As manyas 21 toxinswere detected and characterizedby high performanceliquid chroma- tography HPI.C! and massspectral measurements. The results presented heresupport the hypothesis of oxidativemodification of toxinsin fishes.

MATERIALS AND METHODS Thewild specimensof G,toxicus were collected in theGambier Islands, FrenchPolynesia, in 1979,Moray eels Cymnothorax =Lycodontis! java»icus S40specimens! were collected around the island of Tahiti and in theTuarnotu andMarquesas Archipelagos. The viscera 25 kg including43 kg of livers! were usedfor isolation of CTX and minor toxins. Red snappersLuj ta»us bohar, groupersPlectropomus leopardus and Epi»eyhelus fuscoguttatus, and parrotfish Scarusgibbus were collected in FrenchPolynesia, Micronesia Saipan!, Fiji around Suva!and Japan Okinawa!. Purificationof CTXfrom the morayeels is illustrated in FigureI, Details of theprocedure were described previously~. Purification of toxinsfrom other fishesor G,toxicus was carried out virtually in the sameway. Comparisonof retention times of toxins was achievedon a DevelosilODS-7 column 0x250 mm, Nomura Chemicals,Japan! by linear gradient elution starting from acetonitrile-water0:40! to acetonitrile .5%/min, 1 mL/min! for moray eel toxins,and from acetonitrile-water 85:15! to acetonitrile,5%/min, 1 mLj min!for toxinsin theparrot fish and in G.toxicus. The elution was continued at least for 60 min after the acetonitrile concentration had reached 100%. Elutionof toxinswas monitoredby both mousebioassays and by measuring theend absorption 10 nm!with a UVfIow monitor. Toxins thus separated werefurther purified on the same column in thecase of a relativelypolar toxin group,or on either a Capcellpak C-S column 0x250 rnm, Shiseido Co., Japan! or on AsahipakODV-50 column 0x250 rnm, Asahi Kasei Co., Japan! with similarsolvent systems. Mass spectra of thepurified toxins were measured on a JEOLJURIS DX-303HF spectrometer by using glycerol,thioglycerol or 4- mtrobenzyla!coholasa matrix,and 'H NMRspectra on a JEOLGSX-400 00 MHz!spectrometer. Fluorescence labelling of CTXwith anthroylnitrileand subsequentHPI.C determination ofthe fluorescent derivative were carried out

26 Ciguatoxins from Different Fish Species

Mora eel viscera 125 kg

Extracted with Acetone

PartitionedEt20/H20 Defatted with rt-C6H14

Toxic extract

Si02/CHC13-MeOH97:3; 9;1

CHCl3-Me3OH 9:1 ! Fiorisil/EtOAc; EtOAc-MeOH 9:1; 3:1

EtOAc-MeOH 9:1 ! DEAE-Cellulose/CHC13,CHC13-MeOH I:1

CHC1 -MeOH I:1

SephadexLH-20/CHC13-MeOH 6:4

Toxic fr.

ODS cartridge/MeOH-H20 5:5;6:4; 7;3;8:2

MeOH-H20 8:2 LiChroprepRP-18 /MeOH-H208:2 twice

/MeCN-H20 65:35 twice CTX 350 pg LD5p 35p ng/kg!

Figure l. Extractionof EelCiguatoxin. Purificationprocedure of ciguatoxinfrom morayeel viscera. according to the method previously developed for measurements of pectenotoxin-1"'~.

RESULTS As many as 21 toxins were detectedfrom G. foxicus,parrot fish and moray eels.They were separableinto four groups on the basis of their polar- ities. Eachgroup was further divided intoseveral components. Tentative

27 A-M Legrand,et al. codenames and mass unit of theprotonated molecular tons of all thecompo- nentsdetected arepresented inTable I, G.Ioxicus contained only group 3 and group4 toxinsbutno ciguatoxin, Theleast polar toxin C!, of which the molecularweight is still unknown, was the major toxin. In parrotfish, toxins weredistributed among the groups 1,3 and4. Although a toxin very close to

Table 1. TentativeCoding of Ciguatoxins

ORDER MORAY PARROT OF CODE NO MH+ EEL FISH ~to ~ ELUTION M/Z! LIVER! FlESH!

IA 0 I B C1X! 1111 1029 0 1117 1079 1061 1117 0 0 1061 1095 3A 0 3B 1077 3C 1085 1023 9D 941 1095 0 0 1029 1061 1023 1061 0 0 '4B GT-48! 1061 0 0 4C

': Stereo isomer ~: Maiortoxic fraction

Proposedcoding ofciguatoxins fromthe moray eel, Parrotfish andG. Ioxicus. 'Stereoisomer, solid circles are major toxic fractions. Ciguatoxins from Different Fish Species

CTXin the elutingposition from the columnwas present in the parrotfish, it was distinct from CTX in the massspectruin. The toxin 3Bwas the major toxin in the parrot fish. The toxins 4A and 48 isolated from the parrot fish were suggestedto be identical with thoseof G. toxictison the basisof their chrornato- grams Figure 2! and inass spectral data. Toxins of the moray eels showed higher polarities than those of G. foxicusand the parrot fish. CTX was the principal toxin, but as many as nine toxins were isolated from the moray eels. Trace amounts of CTX were identified by HPLC after derivatizing the toxin into an anthroylester. The minimumdetection level for CTX was I ng per injection. Fluorometric analyses Figure 3! show CTX as the major toxin in carnivorous fishessuch as the moray eels,red snappersand groupers.

'i0 Xi C gain!

tories Emmic toxic iaI ic 31 cia!

Figure 2. HPLC Analysesof G. toxicus and Parrotfish Toxins HPLC comparison between toxins of G. toxicus A! and toxins of the parrotfish Scarusgibbus 8!.

29 A-M Legrand,et af. ooloaa t 06v65065tOOS-S ~ olaoatt 45 ~ OaCO tloa : l.ool/ola ooaltor: Plaoroootot faacagattataa 8.565~ Ko.565m 5 " C. Jaaoaiaar 6. Oaooror ooO55 P. laoparoaa

Figure3. FluoroinetricHPLC Detection of CTX.

DISCUSSION Inprevious studies we isolated CTX from moray eels and then determined its structuretogether with that of a congenerisolated from the causative dinofiagellateG.foxicus s='. Our hypothesis, drawn from the difference in the degreeofoxygenation between the two toxins, that an oxidative modification oftoxins must be taking place during the food chain was supported by results of thepresent study. The data shown in Table1 clearlyindicate that the polarityof toxins increases asthey move up the food chain from G. foxicus to theparrot fish and then to themoray eel. The widely accepted previous assumptionthat CTX is the principal toxin in manyciguateric fish was proven tobe applicable tocarnivorous fishes but not to other fish species. The multi- plicityof thetoxins and different toxin compositions between the moray eel andthe parrot fish may provide a basistoexplain the diverse symptoms of ciguatera,especially toexplain the difference between patients intoxicated by carnivoresorby herbivores '. The fluorometric HPLC determination method for CTXwas developed for thefirst timein thisstudy. The method will be usefulto identify the toxin in smallsamples, although difficulties may arise fromthe extremely low concentrationof the toxin in fish.We have observed CTXconcentrations as low as 0.5ppb in fish Fleshsamples iinplicatcd in humanintoxication in Okinawa.Developing a properprocedure for detecting sucha low concentrationof CTX will be an extremelydifficult task, The structureof CTXsuggests the possibility of obtaining a specific anti-CTX antibodyby immunizing animals with a CTX-proteinconjugate prepared via CTX-hemisuccinate.Developing a highly specific enzyme linked imrnunoas- Ciguatoxins from DifFerentFish Species

say method seems possible, if enough CTX is made available. The cross reactivity of such an antibody to the congenersisolated in this study remains to be testedin the future. In the present study we demonstratedno lessthan 21 toxins in ciguatera samples. It may be possible, however, that there are additional toxins in different speciesor fish from other regions. Structural elucidation oFtoxins other than CTX and 48 is under way. Acknowfedgemenls This paperis dedicatedto ProfessorJ. Roux, Director of the Louis Malarde Institute, whosecontinued support to the collaborationbetween Tahitian and Japaneseteams led us to many successfulresults including the structural determination of CTX and its congener. The authors thank Dr.R. Bagnisfor his continued support to the ciguatera project, Messrs J, Bennett and M. Barsinasfor collecting samples,and MmesC. Lotte,J. Teore and T. Maubertfor technical assistances.Some of the samples were provided by ProfessorY. Hokama of the University of Hawaii and by Dr. U. Raj,a formerstaff member of the University of the South Pacific in Fiji. Special thanks go to Dr. M. Murata of Tohoku University, Japan for his valuable assistancein spectral measurements,This work was supported by the FrenchMinistry of Research and Industry, PasteurInstitute of Paris,French Polynesia Government and by a Grant-in-Aid for OverseasResearch of the Ministry of Education, Science and Culture, Japan.

LITERATURECITED Bagnis,R. 1968. Clinical aspectsof ciguatera fish poisoning! in French Polynesia. Hawaii Med. J. 28:25-28. Tachibana, K., M. Nukina, Y. Joh and P.J. Scheuer. 1987. Recent devel- oprnentsin the molecularstructure of ciguatoxin. Biol. Hull. 172:122-127, Yasumoto,T I, Nakajima, R, Bagnisand R. Adachi. 1977. Finding of a dinoflagellate as a likely culprit of ciguatera. Bull. Jpn. Soc. Sci. Fish. 43:1021-1026, 4, Bagnis,RS, Chanteau and T. Yasumoto. 1977.Un agent etiologique vraisemblablede la ciguatera.C.R. Acad Sc.I'aris, 285Ser ie D: 67l -674 Legrand, A-MM, Litaudon, J-N. Genthon, R. Hagnis and T, Yasumoto. 1989a.Isolation and someproperties of ciguatoxin. J. Appl. Phycol. 1 183-188. Murata, J., A-M. Legrand,Y lshibashi and T. Yasumoto. 1989. Structure of ciguatoxin and its congener. J. Am. Chem. Soc.Ill:8929-8931, Murata, MA-M. Legrand, Y. Ishibashi, M. Fukui and T. Yasumoto, 1990. Structuresand configurations of ciguatoxin from the moray eel Cymnothoraxjavanicus and its likely precursor from the di- noflagellateGambierdiscus toxicus. J. Am. Chem. Soc.in press. A-M Legrand,et al.

S. Legrand,A-M., P. Cruchet,R. Bagnis,M. Murata,Y, Ishibashiand T. Yasurnoto.1990. Characterization and spectralevidence for the pres- enceof multiple ciguatoxins. In: ToxicMarine Phytoplankton E.Graneli, B.G.Sundstrorn,L. Elderand D.M Anderson,eds.!, Elsevier,NewYork. pp. 374-37S. 9. Goto, J., V. Goto, F. Sharnsa,M. Saito, S. Komatsu,K, Suzaki and T. Narnbara.1983. New sensitivederivatization of hydroxysteroidsfor high performanceliquid chromatographywith fluorescencedetection, Anal, Chimica Acta. 147:397-400. 10. Lee,J.S., M, Murataand T, Yasumoto.1989, Analytical methods for determination of diarrhetic shellfish toxins. In: Mycotoxins and Phycotoxins'88, Natori, Hashimoto and Ueno, eds.!. Elsevier, Amsterdam. pp, 327-334. BrevetOxirtsartd RelatedPOlymeriC Ether TOXirts: StruCtureSanCI BiOSyrt thesis

Y. Shimizu', A. V. Krishna Prasad,H-N. Chou, G. Wrensfordand M, Alam~,

t Departmentof Pharrnacognosyand Environmental Health Sciences, The University of Rhodeisland, Kingston 0288I, U.S.A. ~ Departmentof Pharmacology, University of Houston,TX.

ABSTRACT Several novel toxins have been isolated from the dinoflagellate Gymnodiniumbreve Davis Ptychodiscusbrevis!, and their structures have been determined. All of them are characterizedby the presenceof large 7-mem- bered, 8-memberedand 9-memberedrings. NfvlR and X-ray analysis of the conformation of the moleculeshave revealed the enormous flexibilit of this linear polycyclic system, which may be a crucial factor in the interaction of these toxins with membranes, The feeding studies of the biosynthesis of the brevetoxins using stable isotope precursors have shown the unusual nature of the dinoflagellate polyketides. A new seriesof toxins, hemibrevetoxinshas been isolated from G, breve,They are an abridged form of brevetoxins,and their structuresare expected to provide information on the structure-activity relationship and biosynthesisof brevetoxins.

INTRODUCTION Chemical studies of the secondary rnetabolitesof the Florida red tide causing organism, Gymrtodiniumbreeze Davis have shown that it produces a number of compounds with neurotoxic properties, Chemically these neuro- toxins areclassified as polyetherswhich have previously beenencountered m fungi and bacteriaof various species.The polyether toxins of Gymnodinium breveare, however,unique in the sensethat they havea largering system including8- and9-rnernbered rings andthus, inherently have greater confor- mationalflexibility, Mechanistically, these toxins have been shown to interact with sodium channels and cause a depolarization of the excitable membranes'.The conforrnationalflexibility, in combinationwith the linear ring system,may play an important rolein the penetrationof thesemolecules into the membraneand binding. Thus, the elucidation of the toxin conforma- tion will be essentialto our understandingof the molecularmechanism of the Y. Shimizu, et al. actionof theseimportant toxms. Recent chemical investigation of ciguatoxins hasrevealed that they also have a polyethercyclic system closely resembling brevetoxins,which implies a similarity in molecular properties, mechanisms of actionand biosynthetic pathwaysof thesetwo classesof toxins.

MATERIALS AND METHODS

G. breve was cultured in NH-15 medium either in Fernbach flasks I each! or in 5 gallonglass bottles 2 I! under4000-5000 Lux illumination2 hr. DL cycle!for 3 weeks.The cells were extracted with methylene chloride, and the extract was fractionated between petroleum ether and 10 % aqueous methanol. The residue from the methanolic extract was purified by flash chromatographyon SiO2 firstwith methylenechloride: benzene: methanol 0: 5: 1!, and then with methylenechloride: ethyl acetate:methanol 0: 30: 1!, High pressureliquid chromatography [normal phase, Si02, isooctane: isopropanol:I!] of the residuefrom the desiredfraction resulted in the purificationof brevetoxins and hemibrevetoxins. NMR spectra were recorded on Brucker 300 and 500 MHz instruments.

RESULTSAND DISCUSSION Although a number of chemical characteristicsof G. brevetoxin were reportedbefore 1972 the first paperdealing with an insightinto the chemical structure NMR and MS dataof brevetoxin!was reportedin 1974by Shimizu, etat. at theFood and Drugs from theSea Conference held in PuertoRico3,4, In that report thefull NMR, IR, and MS spectraof GB-2toxin brevetoxin 8>were presented,and its polyetherstructure was speculated>, Subsequently, Baden etat. and Risk et al. alsoreported the isolation of a toxinidentical with GB-2s 7. Thestructure of brevetoxin8 wasestablished by X-raycrystallography in 1981".The unprecedentedpolyether structurearoused enormous interest in thestructure of othertoxins, especiallybrevetoxin A GB-1toxin!, which was known to be the most toxicof a8 brevetoxinsand to have a considerably different structure from that of brevetoxin B. Thus, after the structure of brevetoxin 8 had been clarified, our efforts were focused on the structure elucidation of brevetoxin A. ExtensiveNMR studies, including COSYand numerousdecoupling ex- perimentsled to partialstructures Figure la!. However,both protonand carbonsignals for certainportions of thestructure were undetectable or badly defused,and we were unable to complete the structurefrom the NMR data, Meanwhile,Nakanishi's group proposed a strucure Figure lb! for brevetoxin A on the bases of NMR spectra and mass fragmentation patternss~". Thestructure differed significantly from subsequent analyses but presentedthe correctstructures of the 8, D andE-rings. The connectivities BrevetoxinStructure and Synthesis Mee 0 CHO

CHO

Teepestaet al 1984

Figure 1. Structure of Brevetoxin A. a! Structural fragmentsof brevetoxin A as determined by NMR data. b! Structure of brevetoxin A. around the E, F and G-rings were uncertain, F- and G- rings were apparently switched and the angular methyl group misplaced. The locations of double bonds, which were also deduced by the mass fragmentation pattern of the rnethylated NaBH4 reduction product of an ozonolysis product, were also misplaced. Later, the NMR and massspectral data were re-interpretedinto a new structure by the samegroup11. BrevetoxinA was obtainedas crystals evenin an early stageof the isola- tion work. However, the silky needles obtained from methanol in the refrigeratorwere unfit for X-raycrystallography. Later, it wasdiscovered that brevetoxinA formed different crystalsat a highertemperature -60'C!. These crystals,which were grown to beautiful prisms of ca,2 mmwidth, gavea good X-raydiffraction map, however,the computeranalysis of the data was ini- tially unsuccessfuLSuccessful analysis of the structure

35 Y, Shimizu, ef af. derivativeof brevetoxinA'3. Surprisingly,both X-raydetermined structures showedthat the G ringexisted in a boat-chairform, and the molecules were twisted90 at the ring. Assignmentof NMR signalsto the new structure showedthat some of the missing NMR signals belonged tothe G ring,raising thepossibility that the ring existed inslowly interconverting conformational isomersin solution.The crystallographic structure of brevetoxin A from crys- tallizationat 60 C waseventually solved>>. The unit cellwas packed with two conformersdiffering at ring E andthe aldehyde side chain Figure 3!. The NMRsignals were not clearly observed in theE ring, Mys

X O. GB-1 brev CH X X H,OH:GB-7

CH=X l7

0 0 ~ 0 H H H

X-0: GB-2 ~ox' e! XW, OH:GB-3 XmO,37&Ac: GB-5 X 0, 27,2Bit-ePoxide:GB-S Figure2: Structure of Brevetoxins.

In the revisedNMR assignmentof the structureof brevetoxinA, Vakarushi'sgroup reported that measurementofNMR spectrum in benzene madeit possibletoassign the missing part of the signals Tempesta, private communication!>'.The experiment was repeated by our group, but improve- mentof thespectrum was limited, Chemical shift andcoupling pattern differencesdue to solventand temperature changes made it difficult for us to comparethese results with thedata collected in CDzCIq and CDCI3. In an

36 BrevetoxinStructure and Synthesis

Brevetoxin A - Conformatian 2

Figure 3: Two Different Conforrners of BrevetoxinA Crystals.

effort to resolvethis problemotherwise, we decidedto look at the low tem- perature I%MRof brevetoxin A in CD~C12.The experimentwas prompted by the report that low temperatureNMR of ciguatoxinhad overcomea similar problem whichwas due to conformational perturbations2, Normally, a measurementat low ternpertureis apt to causea freezein conformationsor slower conforrnationalinterconversions, resulting in more complexor broadenedNMR signal. In fact, spectra measuredat -30"C and -50'C showedconsiderable broadenings of most signals as well as chemical shift changes, thereby making the assignment of the signals impos- sible. However, the COSY spectrum at 253 K exhibited some significant differencesto the COSYspectrum at roomtemperature Figures 4 and5!. A 0 numberof crosspeaks that were not observedat roomtemperature 98 K!, now appearedin the spectrumat 253 K. Thereis a newset of crosspeaks at 6 3.45,3.16 pprn, which seemsto be connectedto thecross peak that represents protons 26 and 27. Thiscross peak i is,in turn,related to crosspeak v, whichis connectedto cross peak vi. Cross

37 y. Shimizu, et al, peaki is also related toii and iii, and ii is connected toiv, and iii is also related tojv, Cross peaks ii to vi areall new cross peaks, Since they all seem to be relatedto thecross peak that represents protons 26 and 27, this pattern may representoneof the missing connectivities inthe structure ofbrevetoxin A in solution,namely the protons on C-27 to C-31.of ringG Figure4!. Thenew cross peak vii, seems tohave a connection with cross peak ii and to beconnected to thesystem that contains ii andiv, Theposition of cross

i D

2.'5

3.S

r s 3 0 2.0PPV jo Figure4: NMRSpectrum of Brevetoxin A. C+Clp at 253'K.0 to 4.5pp|n region!.

3g HrevetoxinStructure and Synthesis peakvif onthe COSY spectrum suggests that it maybe related to theolefinic portionof ringE, thatis representedeither by a crosspeak between protons 17and 15 or protons19 and 20 Figure5!. As mentionedearlier, E is a moiety in which some form of conformational interchange might be taking place.Resonances had been assigned to theolefinic protons of thering in a

20 ~

s r

oP

5!! s'5 'S2 c,'5 c 0 3% 3i p s 2 7 !.'0 !.0

Figure5: NMR Spectrum of BrevetoxinA. CD2Clpat 298'K.0-6.5 ppm region!.

39 Y. Shimizu, et aL previousstudy, but it wasdifficult to assign signals in thespin system con- nectedto the olefin.The subsequent solution of the X-raystructure of brevetoxinA indicated that there were two conforinations that the double bondin the E ringcould adoptis. In conformationI, the olefin is up to the ring, whilein conformation2, it is twisteddown as shownin Figure3. In the crystalsformed at a hightemperature, theunit cell is packedwith these two conformers.Thus, it maybe reasonableto assumethat theseconformers undergoa slowinterchange in solution. Thepeak at viii seeinsto beconnected to thepeak at vr'i.Because of the chemicalshifts of peaksviii andvii, oneinay make a preliminaryconclusion thatthe two cross peaks are connected to theolefinic portion of ring F..An examinationof the spectrum, however, will clearlyindicate that some of the protoncouplings ofthat connectivity arestill missing, The proton spectra of brevetoxinA atvarious low temperatures shows that it existsin morethan one conformationin severalportions of the structurethat are slowly inter- converting,It has been shown that the molecule in the solid state consists of twoperpendicularly linkedpolycyclic sheets linked at ring G, Rings A-F and H-jmake up the two sheets. In solution, ring G slowlyinterconverts fromthe boat-chairform to the crown form. The appearance of the low temperature 'H NMRspectra clearly lerids support to thisspeculation. Molecular mechanics calculationshave indicated that the boat-chairform is the moststable of the twoconformations, but onlyby -2.9Kcai/mol. Thus one should expect rela- tivelyeasy conversion between the two forms, and as a resultit wouldbe very difficultto seevery sharply defined resonances of certain protons, especially closearound ring G whichis directlyinvolved in the conformational change.Despite these factors, some assignments of the resonances around ringsE andG havebeen made. Our NMR experiments indicate that some resonancesaround these two rings are still missingand it is our contention thatmore detailed NMK experiments are required before the assigninents can be madewith any degreeof certainity, In addition to brevetoxins,G. brevealso produces severalsmaller corn- pounds,which cause characteristic rounding of themouse neuroblastoma cells,similar to brevetoxins,as well as showingcytotoxicity at 5 pmoles level'4.These compoundshave been recently re-named as hemibrevet>.Like brevetoxins, hemibrevetoxin must be biosynthesizedbya cyclization cascade initiated from the right hand of the chainby the opening ofthe cis-epoxide, followed by a hydrideion transfer and consecutivetrans-epoxide openings. The presenceof alkenetails in hemibrevetoxinsnotonly affirms the polyene origin of brevetoxins,but also showsits relationshipwith ciguatoxins,which have been shown to havea

40 BrevetoxinStructure and Synthesis

OH

1 CHO

25 Me 26 OH

CHO

Figure 6. Hemebrevetoxins:Structure and Biosynthesis. a! Hemebrevetoxin B. b! Cyclization cascadeinvolved in the biosynthesisof hernebrevetoxinB,

diene or diol side chain. Work on the structures of brevetoxin A and C is in progress. Considerableprogress has been madein the understandingof the biosyn- thesisof brevetoxins.Results of the feedingexperiments, utilizing I-i C, 2-l C and 1,2-'-C acetates,have shown that the biosynthesisof brevetoxinsdoes not follow the ordinary polyketide pathway known in fungi and bacteria 's "'s. In an earlier paper,we proposeda new biosynthetic mechanismwhich involved the condensation of dicarboxylic acids such as succinate, hydroxylmethylglutarate,and ct-ketoglutarate'~.Later, Lee et at. concurred with our hypothesis,however, differed in certain proposedprecursors' . Our recentanalysis of various feeding studies seemsto implicate a deepinvolve- ment of amino acid metabolism in the formation of the precursorsof the brevetoxins Figure7!. Acknowledgements The work written in this paper was supported by NIH grantsGM 2S754 and GM 24425,which is greatly appreciated.

41 Y. Shimizu, et ai.

OH

S: suocittatc HMG' hydroxytmethy lgltttarate KG: a-lteoglutarate P: propiottato A: acetate

Figure7. Biosynthesis ot Brevetoxin A,

LITERATURECITF.D Steindinger,K.A. and D.G. Baden. 1984. Toxic Marine Dinoflagellates. ln:Dinoflagellates D.L.Spector, ed.!. Academic Press, New York, pp. 201-261, 2, Murata,M., A. M. Legrand,Y. Ishibashi and T. Yasumoto.1989. Struc- turesof ciguatoxin and its congener. J.Am. Chetn. Soc. 111; 8929 8931. 3. Alam,M. 1975,Marine biotoxins. Texas Rep. Biol. Med, 33: 183-199, 4. Shitnizu,Y., M. Alarn and W. E. Fallon. 1974. Red Tide Toxins. In: Food andDrugs from the Sea Proceedings H.H. Weber and G. D. Ruggieri, eds.!.Marine Technology Society, Washington, D,C. pp,238-251 Shimizu,Y. 1978,Dinoflagellate Toxins, In: MarineNatural Products, Chetnicaland Biochemical Perpectives. Vol 1. P.J, Scheuer, ed.!, Aca- dernic Press,New York. pp. 1-42. Baden,D, G.,T. J.Mende and R. E.Block. 1979. Two Similar Toxins Isolatedfrotn Gyrrtttodittiuttt breve. In: Toxic Dinoflagellate Blooms, D. L. Taylorand H. Seliger, eds.!. Elsevier, New York. pp. 327 - 334. Risk,M., Y. Y. Lin, R. D. McFarlane, M. V. S.Ramnujam, L. L. Smithand N. M. Trieft.1979. Purification and Chemical Studies on a MajorToxin I'romGytrtuodittium breve. In; Toxic Dinoflagellate Blooms D. L, Taylor andH. H. Seliger,ed.!. Elsevier, New York. pp. 335-344. Lin,Y, Y., M. Risk, S. M. Ray, D, Van Engen, J,Clardy, J. Golik, J. C. James andK. Nakanishi. 1981. Isolation and structure of brevetoxinB fromthe "RedTide" dinoflagellate Ptychodisctts brevis Gytttttodtttium breve!. J. Am, Chem, Soc. 103: 6773-6775.

42 BrevetoxinStructure and Synthesis

Tempesta,M., J. Golik, J, C, James,K. Nakanishi, J.Pawlak, T, Iwashit M. L. Gross and K. B. Torner. 1984.The structure elucidation c~f brevetoxinA Gyrnnodiniumbreve!. The International Chemical Congress of Pacific Basin Societies. Honolulu, HI. Abs. 10E45. 10. Nakanishi, K. 1985,The chemistry of brevetoxins:A review. Toxic<>rt 23: 473-479, 11. Pawlak, JM. S. Tempesta,J. Golik, M, G, Zagroski, M. S. Lee, J< Nakanishi, T. Iwashita, M. L, Gross and K. B. Tomer. 1987. Structure ~f brevetoxin A as constructed from NMR and MS data. J, Am. Chem. Soc 109: 1144 - 1150. 12. Shimizu, Y., H-N, Chou, H. Bando, G. Van Duyne and J. Clardy. 1986.Structure of brevetoxin A G B-l !, the most potent toxin In the Rorida red tide organism Gymnodinfumbreve Pfychodiscusbrevis!. J, Am, Chem. Soc, 108: 514 - 515 13. Van Duyne, G. D. 1990. X-RayCrystallographic Studies of Marine Tox- ins. In: Marine Toxins S. Hall and G. Strichartz, eds.!. American ChemicalSociety, Washington, D. C. pp. 144-163. 14. Shirnizu, Y. 1982. Recent progress in marine toxin research. Pure % Appl, Chem, 54: 1973- 1980. 15. Prasad, A. V. K, and Y. Shimizu. 1989. The structureofhemibrevetoxirt- B: A new type of toxinin the Gulf of Mexicored tide organism.J, Am. Chem. Soc. 111: 6476 - 6477. 16. Lee,M. SD. S.Repeta, K, Nakanishi and M. G, Zagroski, 1986. Biosyn- thetic origin and assignmentsof C NMR peaksof brevetoxin B. j. Arn. Chem. Soc. 108: 7855 - 7856. 17, Chou, H. N. and Y. Shimizu. 1987.Biosynthesis of brevetoxins. Evi- denceof the mixed origin of the backbonecarbon chain and the possible involvement of dicarboxylic acids,J. Am. Chem. Soc,109: 2184- 2185. 18, Shimizu, Y., S. Gupta and A. V. K. I'rasad. 1990.Biosynthesis of Di- noflagellateToxins. In: ToxicMarine Phytoplankton E. Graneli,B. G. Sundstrom,L. Elderand D, M, Anderson,eds.!. Elsevier, New York. pp, 62-73 19, Lee, M. S., G-W. Qin, K. Nakanishi and M. GL Zagroski. 1989. Biosyntheticstudies of brevetoxtns,potent neurotoxins produced by the dinoflagellate Gymnodiniumbreve. J. Am. Chem Soc. 111: 6234 - 6241

43 Pharmacology QAd Toxicology Effectsof Toxinson the Voltage-GatedNa Channel

M. T, Tosteson HarvardMedical School, Laboratory for MembraneTransport, Department of Cellularand Molecular Physiology, 240 Longwood Av. Boston, MA 021]5.

ABSTRACT Thevoltage-gated sodium channel is a transmernbraneglycoprotein whichpresents a varietyof bindingsites for a largenumber of neurotoxins whichmodify its normal behavior, In thisintroductory paper I wish to review theeffects of varioustoxins, including ciguatera fish toxin,which seems to act at a siteon sodiumchannels unique to thisclass of polyethers.

INTRODUCTION Reportsof tishpoisoning by coastal marine fish resembling ciguatera fish poisomngdate back to theearly 15th century t, However, despite the long historyof ciguatera,we are just beginning to learn about the chemical charac- teristicsand structure of theciguatera toxins. It is known,however, that there areat leastthe~ toxiccomponents in ciguatera: the principaltoxin, called ciguatoxin,whose structure has just become available; maitotoxin, which is morepolar than ciguatoxin and has a molecularweight of around3300, with no apparentchemicaf similarities between rnaitotoxin and ciguatoxin; and third,okadaic acid, a polyetherfatty acid derivative which has been isolated from Prorocentrumlima, a dinoflagellate included in the benthic ciguatera community2+, During the past 5 yearsor so it hasbeen shown that someof the toxic componentsof theciguatoxins act as neurotoxins on theNa+ channel, whereas othersact on Ca~+ channels. This property will helpas an assay in thefurther physico-chemicalcharacterization of the toxins, while at the same time making thesetoxins potentialtools in the study of electricallypropagated impulses Na+ channels! as well as in the study of signal transduction Ca2+ channels!.As an introductionto thepharmacology of ciguatoxin,which will bediscussed in the papersthat follow, I willreview briefly the current knowl- edgeof the structureof thevoltage-gated Na+ ion channel,I will thenshow how toxinsmodify these properties, trying to indicate,whenever possible the potentialsites in the moleculewhere particular toxins might act. Formore in- depthreviews on each of thesesubjects the reader is referred to thebook by B. Hilleas well as the reviewsby C, Armstrong,W. Catterall and G. Strichartzs s. M, T. Tosteson

GENERAL ASPECTSOF CHANNELS Excitationand electricalsignaling in the nervoussystem involve the rnove- ment of ions through ionic channelswhich are specifically responsiveto some stimulus:membrane potential change, neurotransrnitter or other chemical stimulus,mechanical deformation, etc. The proteins which underlie these responsesmight be thought of as enzymes, in thatthey catalyze the movement of ions acrossotherwise impermeablemembranes". These polypeptides have beenshown to havespecific binding sitesfor moleculessuch as the marine toxin tetrodotoxin,which modify their properties.The use of this specific bindinggreatly facilitated the isolation,purification and characterizationof thevoltage-activated Na' channel. Sodiumchannels are large, 200-300 kilodalton kD! glycopeptideswhich are anchoredin the membraneby hydrophobic,electrostatic and covalent bonding.The subunitstructure of the Na' channelwas foundto vary accord- ing to the tissue from which it had been isolated. Thus, the Na+ channel isolatedfrom marnrnalianbrain wasfound to be a complexof three subunits: the a subunit 260kD, 8, 6 kD! and 62 3 kD!. In skeletal muscle, the 82 subunitis absentand in the channelpurified from electriceel only the o- subunit was found~. Theseresults imply that the large a subunit is the main functional componentof the sodiumchannel. The primary structure of these channels was deduced by cloning and sequencingthe DNA complementaryto its mRNAii, lt wasfirst notedby Noda that there are four homologous domains in the electroplax sodium channel,with six regions of probablea helical structurelong enough to be membrane-spanningsegments, designated Sl throughS6". Sl-S3 and S5-S6 are mainly hydrophobic, whereas S4 is amphipathic, highly charged and thought to be involvedin voltage-gating.It has been found recentlythat a syntheticpeptide with the sequenceof S4incorporates spontaneously into lipid bilayersand promotesthe formation of voltage-gated,cationic chan- nels'~.

STRUCTUREAND FUNCTION OF SODIUM CHANNELS When the protein channel!is stimulated,there is a structuralrearrange- ment, thedetails of whichcurrently escapeus, whichleads to the flow of ions down their electrochemicalpotential gradient. This responseof the channel is calledgating, an openingor closingof the pore.When the pore is open,it allows passivetlow of a restrictedclass of ionsdown their electrochemical potentialgradient. These ionic currents can also produce the opening of other voltage-sensitivechannels in adjacentmembranes, such as the voltage-sensi- tive Ca~' channelsin the sarcoplasrnicreticulum, an eventwhich produces Ca~+ release in the sarcornere of muscle cells, Toxins and Na Channels

Figure1 showsthe electricalcurrent which flows throughXa+ channels when themembrane potential is clampedat different values. The tracesshow the salient features ot the voltage-gatedNa' channel:the activation of the currentto a peak followedby a decreasein the current at a Axedvoltage!, or inactivation. The simplest kinetic schemewhich reflects thesefeatures calls for threeconformations or statesof the Na' Ca2'!channel: a nonconducting, resting state,a conducting,open state and a different non-conducting,inacti- vated state. The true channel kinetics are in all probabilities more complicated. ln fact, in order to fit the data, it is necessaryto postulate the existenceof intermediate nonconducting statesbetween the resting and the f tnA! +75 mV lo

C 0 V!

020l2 3

Time ms!

Figure 1. The Voltage-GatedNa Channel. The electricalcurrent flow left panel! and conductance right pane!!of Na channels at different fixed! rnetnbranepotentials, open states.ln additionmore than one conducting stateand multiple path- waysto the inactivatedstate havebeen proposed'3, From Figure1, it is possibleto seethat the current-voltagerelation for thepeak Na current indicatesthat the channelsare largely closed at rest negativepotentials! and that theyopen upon depolarization ci'. right panel!. Withthe development of thepatch clamp techruque, it has been possible to expandour knowledge of the workingsof the channelprotein through mea- surementsof the propertiesof onesingle channel out of 15-2NNchannel~ pm2!rather than those of manyindividual channels atone point in time.This, in turn, hasresulted in a morein-depth knowledge of thefunction of theNa

49 M, T. Tosteson channeland of themode of actionof theantagonists of theprotein, such as the neurotoxins. In the pastfew yearsthere has been a rapid advancein the elucidationof the mechanismof actionof naturallyoccurring neurotoxins, The quickening pacehas been driven by aninterest in thetoxic molecules produced by marine organismsand has been aidedby the applicationof highly sophisticated biochemicaland pharmacologicaltechniques. These studies have led to an increasedunderstanding of the variable statesof the sodium channel as it goes throughthe modifications which eventually lead to the actionpotential. Fur- thermore,it has been possibleto determinethe heterogeneityof channel structurethrough differential effectsof toxinsand, finally, it hasbeen possible toestablish the interaction between the different toxin binding sites, which has helpedin the determinationof the rolethat the different parts of the Na' channelprotein play in channelactivation.

TOXINS AND THEIR ACTIONS In thelight of thesead vances, I wish to outline the modeof action of toxins whichare known to act at variousstates of the Na' channel,with particular emphasison those states which interact with the toxins which we are con- cerned with in this Conference. Table I shows a list of the toxins I will consider, The first group listed is composedof those toxins which block the sodiumchannel of manyexcitable membranes with high-affinity Kz-10 9M!

Table ]. Na Channel Inhibitors: Binding Siteand Action.

Binding Ligand Action Site

F.xterna 1 organic cations: TTX STX inhibition of Na Peptides: ir-conotoxins transport, reversible binding

Within lipid soluble: persistentactivation, Channel BTX vera trid inc irreversible binding, TTX sensitive

External polypeptide toxins: inhibit activation, sea anemone toxin reversiblebmding a-scorpion toxin

External $-scorpiontoxins modify activation

50 Toxins and Na Channels

and selectivity, Tetrodotoxin TTX! was first used as a label to count the number of Na' channelson membranesand later to follow the progressof the channel protein during purificationt4. The mechanism of action of TTX and saxitoxin STX! involves a guanidiniutn ion present in both compounds, This group blocksthe open channel,and this blockage is antagonizedby Na+,Ca~+ and H+ ions. Thesetoxins acton! y on the sideof the channelprotein accessible from the external surface of the cell membrane'+'". Thus, the studies with TI'X, STX and the It-conotoxins has shown that Na' channels differ in structure between their internal and external surfaces,that they have binding sites for mono and divalent cations and that their functioning can be ascribed to conformationai changeswithin the molecule, The characteristicof the next group of toxins, the alkaloids, is that they can depolarize the membranesof nerve and muscle cells and increasethe resting Na' permeability of excitable cells'9, Batrachotoxin BTX!is the most studied compound in this class. The binding of this classof toxins leadsto the stabilization of the channelsin the open state without affecting the activation process. Studieswith thesetoxins reveal that their effect, which is elicited when they are added either to the inside or to the outside of the membrane,is to increasethe population of open channels over a broad range of membrane potentials. This results in slow depolarization of the membraneleading to a transientperiod of rapid, sponta- neous impulse firing. Binding to closedchannels is a slow process,which has been interpreted to indicate that the toxins bind to a site in a hydrophobic environmentz".Although the propertiesof the alkaloid-modifiedNa+ chan- nels are qualitatively the same, differences between the various toxins are apparent, indicating that they may not alter the same region in the channel molecule z'zz, The next group of toxins, act only when added to the outside of cells. These toxins prolong the duration of the action potential by slowing down inactivation. The binding of ot-toxin to the channel seemsto be voltage dependent and the dissociationof the toxin from its binding site is relatively fast at large potentialsz3z4. The polypeptide toxins which act when added to the outside of the cell groups 3 and 4 in Table I! are divided in two classesdue to the differencein theirelectrophysiological actions. The ones in group3, sea anemone toxin and the tx-scorpiontoxins slow down the inactivation of Na channelsin a great variety of excitablecells24~. The interaction of B-toxins with Na+ channels resultsin a morerapid activation and a slowingdown of bothdeactivation and inactivation,The end result is a setof conductingchannels that close slowly at rest and show repetitive itnpulse firing in responseto minimal stimulationz" '. A group of lipid soluble toxins not listed in the table but which havebeen shownto depolarizenerve and muscle membrane in a dose-dependent,TTX- antagonizedmanner are ciguatoxin CTX! and the brevetoxins PbTX! which seemto be part of a separategroup of toxins.The depolarizationof nerve

51 M, T, Tosteson terminals producedby theselipophilic toxins causesmassive release of trans- mitter,which accounts for theiractions on a varietyof organsystems2s >9, The actionof thesetoxins is just beginningto be explored in detail, as the availabil- ity of purifiedmaterial increases. When PbTXor ciguatoxinarc applied to muscleor to nerve, the peak sodiumcurrent is reduced, but the sodium current can be activated at potentials when the channel is normally closed~-~'. It has also been found that; a! activation is 1000xslower, b! sodium channelinactivation is inhibited and c! of the two toxins, ciguatoxin is about 50 times tnore potent than brevetoxin. Studieson the binding competition with other toxins known to act on the Na' channelprotein have revealed that both toxins enhancethe binding PbTx to its own receptor site and that both can displacethe binding of labelledbrevetoxin to its receptor. Other competition experimentsof this sort have shown that neither CTX nor PbTx can displace toxins which bind to sites 1-4 see Table 1! in the Na+ channel protein3~>3. Thus these experiments suggest that PbTx and CTX share a 5th receptor site on the Na+ channel protein, and that it is possible that this receptorsite is locatedon a regionof the Na+ channelinvolved in voltage- dependentgating. ln conclusion, toxins alter the state of the Na+ channel in several ways. Different toxins interact with one another, further indicating that the structural changeswhich underlie the workings of the Na' channel protein move through vast domains of the protein during normal physiological gating. Clearly more work remains to be done both structurally and physi- ologicallyin order to understandthe mode of action of ciguatoxin and brevetoxins,

LITERATURECITED Anderson, D.M. and P,S,Lobel. 1987, The continuing enigma of ciguatera. Biol.Bull. 172:89-107. Murata, M., A.M. Legrand, Y. Ishibashi and T. Yasumoto. 1989.Structure of ciguatoxinand its congener,J, Am. Chem, Soc, 111: 8929-8931, Yasurnoto,T. 1978. Large scaletoxin production by Diolopsalissp.-bio- chemicalstudy of maitotoxin and ciguatoxin. So.Pac. Comm. Rep. 785: 7-8, Murakami, Y., Y, Oshima and T. Yasumoto. 1982. Identification of okadaicacid as a toxiccomponent of a marined inoflagellateProrocertt rum lima. Bull. Jpn. Soc.Sci. Fish. 48: 69-72. Hille, B, 'I984. Ionic Channels of Excitable Membranes. Sinauer Associ- ates inc., Sunderland, MA. Armstrong, C.M 1981.Sodium channelsand gating currents, Physiol. Rev, 61: 644-683.

52 Toxins and Na Channels

Catterall, W.A. 1986. Molecular properties of voltage-sensitivesodium channels. Ann.Rev.Biochem. 55: 953-985. Strichartz,G. 1989.The molecular pharmacology of toxinsthat modify voltage-gatedsodium channels. In: NaturalToxins. Characterization pharmacologyand therapeutics, C,L.Ownby and G,V.Odell eds!. PergamonPress, New York, pp, 43-53. Moczydlowski,E. 1986,Single channel enzymology. In: ion channel reconstitution, C.Miller, ed!. PlenumPress, New Yorkand London. pp. 75-113. 10. Catterall, W.A., J.W. Schmidt, D.J. Messner and D.J FeIler. 1986, Structure and biosynthesisof neuronal sodium channels, Ann. NY. Acad. Sci. 479:186-203, 11. Noda, M., S. Shirnizu, T, Tanabe, T, Takai, T, Kayano, T, Ikeda, H, Takahashi, H. Nakayama,Y. Kanaoka,N, Minarnino,K. Kangawa,H. Matsuo,M.A. Raftery,T. Hirose, S, lnayama, H. Hayashida,T. Miyata and S, Nurna, 19S4.Primary structure of Electrophoruselectricus sodium channel deduced from cDNA sequence,Nature 312:121-127, 12. Tosteson,M.T., D.S.Auld and D.C.Tosteson.1989. Voltage gated chan- nels formed in lipid bilayers by a positively chargedsegment of the Na-channelpolypeptide. Proc. Natl. Acad.Sci. 86: 707-710. 13. Armstrong, C,M. and Wm.F. Gilly. 1979. Fast and slow steps in the activation of sodium channels.J. Gen,Physiol, 74: 691-711, 14. Haffemann, D.R. 1972, Binding of radioactive tetrodotoxinto nerve membranepreparations. Biochim. Biophys.Acta. 266:548-556. 15. Narahashi,T., J.W.Moore and W,R,Scott, 1964.TTX blockage of sodium conductanceincrease in lobster giant axons.J. Gen. Physiol.47: 964-974, 16. Henderson, R., J.M. Ritchie and G.R. Strichartz. 1974. Evidence that TTX and STXact at a metalcation binding sitein the sodiumchannels of nerve membranes. Proc. Natl. Acad. Sci. 71: 3936-3940. 17. Kao,C.Y. 19S6.Structure-activity relationsof TTX, STXand analogues. Ann. NY. Acad, Sci, 479: 52-67. 18. Weiss,R.E. and R. Horn. 1986. Single-channelstudies of TTX-sensitive and TI'X-resistant sodium channels in developing rat muscle reveal different openchannel properties. Ann, NY. Acad,Sci. 479: 152-161. 19. Albuquerque,E.X. and J.W. Daly. 1976,Batrachotoxin, a selective probe for channelsmodulating sodium conductancesin electrogenic mern- branes.In: The specificityand actionof animal,bacterial and plant toxins, P.Cuatrecasas,ed!. Chapman and Hall, I.ondon.pp. 297-338, 20. Angelides,D., S, Rerakawaand G.B.Brown. 1986.Spatial relations of the neurotoxinbinding sites on the sodiumchannel. Ann. NY. Acad.Sci 479: 221-237. 21, Khodorov,B.l. and S.V.Revenko, 1979, Further analyses of themecha- nisrnof action of batrachotoxinon the membraneof myelinatednerve-

53 M. T. Tosteson

N euroscience 4: 1315-1330. 22, Leibowitz,M,DJ.B. Sutroand B. Hille. 1986.Voltage-dependent gat- ing in veratridine-modified Na+ channels.J. Gen. Physiol. S7:25-46. 23. Wang, G,K. and G. Strichartz,1982, Simultaneous modifications of so- dium channelgating by two scorpion toxins. Biophys,J. 40: 175-179. 24. Catterall,W.A. 1977.Membrane potential-dependent binding of scor- pion toxin to the action potential Na+ ionophore. Studies with a toxin derivative prepared by lactoperoxidase-catalyzediodination, J, Biol, Chem, 252: S660-8668, 25, Frelin,C, P.Vigne, 1 l.Schweitz and M, Lazdunski.1984. The interaction of seaanemone and scorpionneurotoxins with tetrodotoxin-resistantNa channelsin rat myoblasts.Molec. Pharmacol.26: 70-74. 26. Wang,G.K. and G.R.Strichartz. 1985.The actions of an isolated B-toxin on Na channelactivation, Biophys. J. 47: 438a. 27, Wheeler, K.P., D.D. Watt and M, Lazdunski. 1983. Classification of Na channel receptors specific for various scorpion toxins, Pflugers. Arch, 397: 164-165. 28, Gallagher, J.P. and P.Shinnick-Gallagher,1985, Effects of crude brevetoxinon membranepotential and spontaneousor evoked end-plate potentialsin rat hemidiaphragm.Toxicon 23: 489-496. Ohizumi, Y, 1987. Pharmacological actions of the marine toxins ciguatoxinand maitotoxinisolated from poisonousfr'sh. BioL Bull. 172: 132-136 30. Huang,J.M.C., C.H. Wu and D,G. Baden. 1984. Depolarizing action of a red-tide dinoflagellate brevetoxin on axonalmembranes. J, Pharrnacol, Exp. Therap.229: 615-621, 31. Benoit,E., A.M. Legrandand J.M. Dubois. 1986. Effectsof ciguatoxin on current and voltage clamped frog rnyelinatednerve fibre. Toxicon24: 357-364. 32. Poli, M,AT.J, Mende and D.G. Baden. 1986, Brevetoxins,unique acti- vatorsof voltage-sensitive sodium channels, bind to specificsites in rat brain synaptosomes.J, Am. Soc.Pharmacoh Fxptl. Ther, 3S:129-135. 33. Lombet, AJ-N. Bidard and M. Lazdunski. 19S7.Ciguatoxin and brevetoxinsshare a commonreceptor site on the neuronalvoltage-de- pendentNa channel,FEBS Lett. 219: 355-359.

54

K. Terao, et aL smooth muscle cells. The mode of action of CTX has also been studied by a number of researchers,and it has been confirmed that CTX causesan increased Na+ permeability acrossthe tetrodotoxin-sensitiveNa+ channels of smooth and/or cardiacmuscle cells'2-'s. In contrastto thosepharmacological studies, there hasbeen only limited information about the in vivo pathological changes induced by either MTX or CTX. In the present study we demonstrate the pathomorphologicalchanges induced by MTX or thoseof CTXand compare thedifferences between the two experimentalpoisonings,

MATERIALS AND METHODS The MTX used in the present studies was extracted from a cultured dinoflagellate,Garnbierkscus toxicus, and the CTX was isolated mainly from reefsnappers Lutjanus bohar collected in Micronesiaand Okinawa, Japan. Both phycotoxinswere purified as previously described'~'". For the experimental animals,4-week old male ICR mice weighing20 to 23 g were obtainedfrom Charles River Japan,Inc. Specimensfor morphologicalexaminations were preparedas previously described' t. For electron microscopy, specimens were observedwith a HitachiH 700H transmissionelectron microscope.

RESULTS TheEffects of MTX andCTX on theHeart One of the target organs of both MTX and CTX is the heart. Figure 1 showsthe ultrastructureof the heart of a mouse30 min after receiving 200ng/ kg of MTX, A markedswelling of theendothelial cells lining the capillariesin thecardiac tissuewas often seen. The lumenof the capillarieswas extremely narrow or evenclosed, Cardiacmuscle cells around the capillaries became degeneratedand necrotic.The cells were condensed and the arrangementof almostall of the rnitochondrialcristae was irregular. Various doses of CTX resulted in multiple single cell necrosesin the myocardiumot theleft andright ventricles and the septum of theheart Figure 2a!. Marked swelling of the cardiac musclecells was seenwith the electron microscope Figure 2b!, The mitochondria became rounded and most or- ganelles including myofibrils, sarcoplasmicreticulum and T-systems were separated from each other. Myofibrils in severalcells had partially disap- peared.Occasionally, an accumulationof blood plateletswas seenin the capillaries among the cardiac muscle cells. Marked effusion of a serum-like substance was found around the capillaries and interstitial space of the heart. Usually a severe dyspnoea and lung edema then followed. These changesin thesurviving mice persisted for over3 days. Almostall mice given over I itg/kg CTX died within severalhours after the injection. About I month later,the surviving mice initially given 0.7 ltg/kg I'athomorphologicalStudies of Ciguatoxins

Figure 1. Effect of MTX on Mouse Myocardium. Electronmicrograph of the myocardiuznof a mouseg>ven 200 ng/kg of MTX30 minutesbefore sacrifice. IUotemarked narrowing of lumenof the capillary arrow heads!. M! myocardium. x5,100

seemedto be completelywell clinically. At autopsy,the heartsof theseani- mals were enlarged,The degeneratedmicroorganelles as seenwith the electronmicroscope had regeneratedcompletely. However, the tubularin- vaginationof the sarcolemmaand sarcoplasmicreticulum were unusually dilated and the effusionwas replacedby thickbundles of collagen fibers.Fibrosis of theheart was seen even 5 monthsafter a singleinjection of 0.5itg/kg CTX Figure2c!, With a singledose of 0.3itg/kg or less,no visible changeswere detected in thecardiac tissue, However, 15 repeated injections of CTXO.l p.g/kg resultedin markeddilation of the right and left ventricles andalso showed severe histopathological changes similar to a singledose of 0,7 p.g/kg Figure 2d!. Manyantagonistic drugs for cholinergicand adrenergic autonerves such as atropine,guanethidine and 5-hydroxydopamine had no effecton the severityof CTXinduced lesions in theheart, whereas reserpine even aggra- vatedthe pathological findings. It is of interestthat the cardiacinjuries inducedby CTX were not altered by pretreattnent with atropine, although this agentprevented the CTX induced diarrhea completely. From these findings it may be assumedthat CTX attackscardiac tnuscle cellsdirectly, whtle MTX preferentially attacks endothelial lining cei|s of the capillaries in the cardiactissue. K. Terao, ef a!.

Figure2. Effectof CTXon Mouse Cardiac Tissue. Micrographsof the cardiac tissue of a mousegiven 0.7 pg/kg of CTX.Up@'r panela! Lightmicrograph of the left ventricle. There are many single cell necroses.x 60 f~uer Panelb! Electronmicrograph of the cardiaccells of a mouse5I3 min. after the injection.Swelling of a cardiaccell is prominent.An accumulationof blood platelets P! is seen in thecapillary, Around the capillary, effusionis presentin theinterstitial space I!. N! Necrotizingcell. x 2,600

58 PathomorphologicalStud res of Ciguatox!ns

Figure 2. Effectsof CTX on MouseCardiac Tissue. Micrographsof the cardiac tissue of a mousegiven 0.7 !rg/kg of CTX.Upper Panelc! Electronmicrograph of the cardiac cells about 4 weeksafter a single injectionof CTX.Around the capillary there are bundles of collagenfibers I! withmany macrophages. M! Cardiac muscle. x 2,600Lower pane! d! Electron micrographofthe heart of a mousegiven repeatedinjections of 0.1 Itg/kg of CTX.Almost all muscle cells are swollen. Intermuscular spaces I! are irnpreg- natedby a denseserum-! ike substance. N! necrotizedcardiac cell. x 2,600

59 K. Terao,et ah

TheEffects ofMTX and CTX on Digestive Organs Diarrheaalways occurred shortly after the administration ofCTX but was notencountered with MTX. Interestingly, the pathomorphological findings showedno discernible changes in the mucosa of theintestine in CTXtreated animals.Only swelling of theunmyelinated nerves and synapses in the smoothmuscle of theintestine was observed Figures 3a and b!, In contrastto theCTX poisoning, tnaitotoxicosis wasalways accompanied bymultiple ero- sionsand ulcers in thestomach. No abnormalchanges were seen in the unmyelinatednervesof the gastrointestinal organsafter the administration of MTX.

Figure3, Effectof CTX on Mouse Intestine. Electronmicrographs of nerve fibers N! in Auerbach'splexus of theintestine. Leftpanel a! Controlx 14,X|0.Right panel b! Sixhoursafter theadministra- tion of 0.7Itg/kg of CTX, Nervefibers and synapsesare markedly swollen.Almost all synapticvesicles disappeared. x 14,000

TheEffects ofA4TX and CTX on Adrenal Glands BothMTX and CTX injured the adrenal glands. However, the targetof MTXwas the zona fasciculata of theadrenal cortex and that of CTXwere cells in theadrenal medulla. Histological examinations revealed the presence of many vacuolesin the adrenocortexof the MTX treated ani- rnals.Ultrastructurally, these vacuoles were identified as autophagosomes containingremnants of microorganelles Figure 4a!. It is well knownthat MTX is a potentactivator of calciumchannels in PathomorphologicalStudies of Ciguatoxins

Figure 4. I'.ft'ectof MTX on MouseAdrenal Glands. a! Electronmicrographs of the adrenal glands of a mousegiven 400 ng/kg of MTX.Cells in zonafasciculata 24hr afterinjection Ixlumerous autophagosomes V! are seen in the cells, x2600. variouscell types,and it hasbeen reported that this toxin resultsin Caz+- dependent releaseof certain hormones. Therefore,we determined the total calciumcontent of theadrenal glands and compared that of thethymus and spleenof MTX-injectedmice pretreated with CoCI> Table I!, Homogenatesof theseorgans were extracted with 0.2N perchloric acid and analyzed with the Zeemanatomic spectrophototneter. In mice injected with MTX only, the total calciumof adrenalglands was increased significantly when compared with

Table1. Effectof M'I'X on OrganCalcium

Treatment Spleen Adrenal Glands

Control 1.7110.20 0.95 t0.09 238+0.09

MTX 1.33+0 25 I. 00+0. 10 3.09+ 0.15t

lviTX + CoC12 1,3I+0.31 1,08i0.12 2.2&0. I I

Totalcalcium content Itg/mg protein! was measured 1 hr after i.p. injections of 200ng/kg of maitotoxin.t Effect of thetreatment issignificant p<0.05!. K Terao et ttI thatof thec, showed no significantincrease after the injectionof MTX.Moreover, a histochemicalanalysis of calciumdistribution demon- stratedthat the density of calcium antimonate inmitochondria and SER of the micegiven MTX was far tnore intensive than that of control mice

Figure4. Eftectol MTXand CTX on Mouse Adrenal Glands. Leftpam I b! Electronmicrographs of the adrenal glands of a mousegiven 400 ng/kgof MTX.Histochemical detection of calcium in theadrenocortical cells. Controlanimal. x45,000 Right panel c! Histochetnicaldetection of calciumin the adrenocorticalcells. A mousetreated with MTX. Densereaction products locatedon themitochondria M! andsmooth endoplasmic reticulutn S!.x45,000

62 Pathotnorpholof;icalStudies of Ciguatoxins

Figure 4, Effect of MTX and CTX on Mouse AdrenalGIands. d!Electronmicrographs of the adrenal glands of a mousegiven 0.7 ltg/kg of CTX,Norepinephrin secreting cells in themedulla of theadrenal gland after treatment with CTX. Vacuoles A! were seenin thesecells. x2,600

5G EUl 4G 0O OE20

Time hr! Figure S. Effectof MTX on PlasmaCortisol in Mice. Timecourse of effectsof i.p.injections of MTX00 ng/kg!of plasmacortisol in mice".

63 K. Terao, ef aL hoursprior to sacrifice.Almost all lymphocytesin thecortical layer of the thymusbecame necrotic. Numerous macrophages with lymphocytic debris wereobserved in thecortex. However, with a singledose of only 50ng/kg or less,no discerniblechanges occurred. Yet, when injections of 45 ng/kg of MTX were repeated 13 times, marked atrophy of the thymus was seen.Histologically, lymphocytes in the cortexdisappeared and corticaltis- suewas replaced by manyfibroblasts and collagen fibers Figures6b and c!. Thespleen also decreased markedly in volumein MTX treatedani- mals, Histologically,periarterial lymphoid sheathsin white pulp were incompletelyformed Figure 6d!. Electronmicroscopy revealed that the end o- thelialcells lining the venoussinus increased in heightand containedmany electron-densegranules Figure 6e!. The volume of thesinuses decreased and thenumber of bloodcells and macrophages present was diminished.T-cell dependentareas in variouslymphoid tissues, including lymph nodes and Peyer'spatches in thesmall intestine, became atrophic. Lymphocytes in the circulatingblood also decreased in number from 65 % beforeinjection> to 'l3%,13 days after the administration of MTX Figure7!. Micepretreated with CoC12a calciumchannel inhibitor, and bilateraladrenalectomized mice

Figure6. Effectot' MTX on MouseLymphoid Tissue. Micrographsof the lymphoidtissues. Left pane! a! A photomicrographof the thymusof a mousetreated with 400ng/kg of MTX 24 hr after the injection. Marked lymphocytic debris was seenin the cortex of the organ. x Rightpane b! A photomicrographof thethymus of a mouseafter 13 repeated injectionsof 45 ng/kg MTX. At the corticallayer C! no lymphocyteswere present. M! Medulla. x30 pathomorpho!ogicalStudies of Ciguatoxirt~

Figure6. Effectof MTXon MouseLymphoid Tissue. Micrographsofthe lymphoid tissues. Upper panel c! Anelectron micrograph of thesame animal Figure 6b, opposite page!. Numerous fibroblasts F! and bundlesofcol lagen fiber were seen in thecortical layer. x 3,000 Lour paneld! A photomicrographof the spleen of a mousetreated with MTX.Note marked atrophy of T-cell dependent areas. x SO K, Terao, et af.

Figure6. Effectof MTXon Mousel.yrnphoid Tissues. ie! Micrographsof the lymphoidtissues, An electronmicrograph of the same spleen Figure 6d, previous pape!. Fndothelial lining cells E! of venous sinuses increasedin heightand contained inany electron dense granules. Venous sinuses S! were narrowed. x 2,60 !

100

30

50

0 0 10 Grays Figure7. Effectof MTXon MouseBlood Lymphocytes. Thettme course tif "' of lymphocytesin themouse blood, large solid circles; % of lymphocyte,large open circles; % ofpolymorph leukocytes, small circles; 8ody weights solid, experimental;open, control!. PathomorphologicalStudies ofCiguatoxins

Figure 8. Effect of CTX on Mouse Penis. Phototnicrographsof the penises.Upper pane! 8a!Control flaccid state, x 7 Lowerpane 8b! Continuous erected penis from a mousetreated with CTX. x 7

showedno discernible changes in lymphoidtissues. TheEffects ofCTX on Penis and Vas Defererts Oneof themost peculiar phenomena observed after the administration of CTXwas the continuous erection of thepenis. About half of thesurviving micegiven 0.7 ltg/kg showedcontinuous erection of thepenis even after death.Histologically, the corpus cavernosurn of the penis was heavily en- gorged,with the length of thecorpus cavernosum being about twice that of the normalflaccid state Figure Sa and b!, Theenlarged corpus compressed the penilebone, so that the gla~s penis was erected and the length of the prepuce shortened,There were many small thrombi in theengorged cavernosutn. A number of investigators have studied the effects of MTX on the vas deferenselectrophysiological ly, They confirmed the Ca2+-dependent contrac- tion of smoothmuscle cells of the vas deferens.When viewed with the electron microscope,synapses and unmyelinatednerve fiberswere markedly swollen.However, no pronunentchanges were seen in thesmooth muscle cells themsel ves.

DlSCUSsrON Table2 summarizesthe pathornorphologicalfindings of MTX and CTX poisoningsin varioustarget organs. Maitotoxicosis apparently differed from

67 K. Terao, cf al.

Table2. Comparisonof Pathologyof Maitotoxicosisand Ciguatoxicosis in M

TargetOrgans Maitotoxicosis Cigua toxicosis

Heart Singlecell necrose~ Marked swelling of Narrowing of ca rdiac muscle cells capillaries Effusion into intermuscularspace

Stomachand Intestine %multiplegastric No discernible erosions and ulcers changes in mucosa ~welling of autonerve fibers and synapsesin smooth muscle

Adrena I Glands Multiple Vacuoles in the autophagosomesin medulla cells of zona fasciculata

Lymphoid Tissues Massive necroses of lymphocytesin the thymic cortex Atrophy of T-ceil dependentareas in the spleenand lymph nodes

Penis Continuous erection Thrombosis

ciguatoxicosisin the pathomorphologicalresponses of mice,The trigger of maitotoxicosis was the MTX-induced increasein the permeability of the plasmamembrane of cells in thezona fasciculata of the adrenal glands to Ca~ ', The stimulated Ca2' influx caused the release of cortisol into the blood. This excessamount of cortisol produced acute i~volution of the thymic lymphocytes.Reduced formation of T-celldependent areas in variouslym- phoidtissues followed. Thisview was also strongly supported by the findings thatpretreatment with CoCl>,a potentinhibitor of theCa + channel,as well as unilateral or bilateral adrenalectomyprevented the lymphoid lesions. Gastric ulcers that occurredshortly after the administration of MTX may be closely related to the excessof cortisol in the blood,

6g PathomorphologicalStudies of Ciguatoxins

Incontrast tothe MTX-poisoning, CTX-intoxication maybe mainly caused bya directeffect of thistoxin on various target cells. CTX stimulated an increasedNa+ permeability across the tetrodotoxin-sensitive Na' channels of cardiacmuscle cells and unmyelinated nerve cefls in thesmall intestine and vasdeferens, Pretreatment with atropine or bilateraladrenalectomy hadno particulareffect on theheart injuries induced by CTX. Therefore, whether or not norepinephrineplayed an importantrole in the cardiaclesions is unclear Resultsof therepeated injection experi ments indicated that both MTX and CTXaccumulated in the bodyof mice,for bothtoxins resulted in severe injuriesof the targetorgans when administrated at successivelow doses, whereasa singlelow dose produced no significant changes.

Literature Cited Takahashi,M., Y, Ohizumi and T. Yasumoto.1982. Maitotoxin, a Caz+ channelactivator candidate. J. Biol,Chem. 257, 7287-7289. 2, Takahashi,M., M. Tatsumi,Y. Ohizumiand T, Yasurnoto.1983. Ca + channelactivating function of maitotoxin,the most potent marine toxin known,in clonalrat pheochromocytoma cells, J, Biol. Chem. 258: 10,944- 10,949. 3. Ohizumi,Y. and T. Yasumoto. 1983a. Contractile response oftherabbit aortato maitotoxin,the most potent marine toxin. j. Physiol.London, 337; 71 1-721. 4, Ohizumi,Y andT. Yasurnoto.1983b, Contraction and increase in tissue calciumcontent induced by maitotoxin,the most potent known marine toxin, in intestinal smooth muscle, Br. J. Pharmac. 79:3-5. 5. Legrand,A-M. and R. Bagnis.1984a. Effects of ciguatoxinand maitotoxin on isolated rat and rabbit duodenum. Toxicon22: 471-475. 6. Legrand,A-M, and R. Bagnis.1984b. Effects of highlypurified maitotoxinextracted from dinoflageIlateGambierdiscus toxicus on action potential of isolated rat heart. J. Molec.Cell Cardiol. 16:663-666. 7. Kobayashi, M., S, Kondo, T. Yasumoto and Y Ohizumi.1986. Cardiotoxic effects of maitotoxin,a principaltoxin of seafoodpoisoning on guineapig andrat cardiacmuscle. J. Pharmac, Exp.Ther. 238:1077-1083. 8. Gusovsky,F., T. Yasumotoand J. W. Daly. 1987, Maitotoxin stimulates phosphoinositidebreakdown in neuroblastomahybrid NCS-20 cells, Cell. Molec. Neurobiol. 7: 317-322. 9. Gusovsky,FJ.W. Daly, T. Yasumotoand E. Rojas.1988. Differential effectsof maitotoxinon ATPsecretion and on phosphoinositide break- down in rat pheochromocytomacells. FEBS Let. 233:139-142. 10. Terao,K., E, Ito, Y, Sakamaki,K. Igarashi,A. Yokoyamaand T- Yasumoto.1988, Histopatho]ogical studies of experimentalmarine

69 K. Terao, et at,

toxin poisoning.II. Theacute effects of maitotoxinon the stomach, heartand lymphoid tissues in miceand rats. Toxicon 26: 395-402. 11, Terao,K., E. Ito, Y. Kakinuma,K. Igarashi,M. Kobayashi,Y. Ohizumi andT, Yasumoto.1989, Histopathological studies on experimentalrna- rine toxin poisoning.4, Pathogenesisof experimentalmaitotoxin poisoning, Toxicon 27:979-988. 12. Rayner,M. D. andT.l. Kosaki. 1970, Ciguatoxin; Effects on Na+fluxes in frog muscle.Fed. Proc. 29; 548. 13. Setliff,P. J., M.D. Rayner and S.K.Hong. 1971.Effect of ciguatoxinon sodiumtransport across the frog skin. Tox. Appl. Pharmacol.18: 676- 6S4. 14. Ohizumi, Y., S. Shibataand K. Tachibana. 1981. Mode of the excitatory andinhibitory actions of ciguatoxinin theguinea-pig vas deferens. J. Pharmacol.Exp. Therap, 217:475-480. 15, Ohizumi, YY. Ishida and S, Shibata. 19S2,Mode of the ciguatoxin- inducedsupersensitivity in the guinea-pig vas deferens. J. Pharmacol. Exp. Therap. 221:745-752. 16. Bidard,J-N., H.P.M. Vijverberg, C. Frelin,E. Chungue, A-M. Legrand,R. Bagnisand M. Lazdunski.1984, Ciguatoxin is a noveltype of Na+ channeltoxin. J. Biol. Chem. 259:8353-S357. 17. Yasumoto,T. 198'5.Recent progress in the chemistryof dinoflagellate toxins.In: Toxic Dinoflagellates, D.M. Anderson, A. W. Whiteand D,G, Baden,eds.!. Elsevier, New York, pp 259-262. 18. Legrand,A-MM. Litaudon, J.N. Genthon,R. Bagnisand T. Yasumoto.1989. Isolation and someproperties of ciguatoxin.J. Appl. Phycol. 1: 183-188, 19. Aguas,A.P. and P.A. Nickerson. 1981, Subcellular distribution of cal- ciurn in zona fasciculata cells of the rat adrenal cortex. Tissue and Cell. 13: 491-499.

70 Mechanismsof Pharmacological andToxicological Action of Ciguatoxio and Maitotoxin

Y. Ohizurni Facultyof PharmaceuticalScience Tohoku University, Aoba, Aobaku, Sendai980, Japan

ABSTRACT Ciguatoxin CTX! caused an increasein thecontractile force of isolated guineapig atria.Tetrodotoxin markedly inhibited theinotropic action of CTX.CTX shifted the current-voltagecurve of Na+ inwardcurrents in a negativedirection. These results suggest that CTX activates Na' channels by modifyingthe voltage dependence of channel activation to increaseNa' in- ward currents,thus producingcardiotonic actions. Furthermore, CTX at concentrationsabove 3 pg ml-' led to cardiacarrest and mitochondrial swell- ing. Maitotoxin MTX! caused a sustained inward Car+ current in singlecar- diac cells. TheI-V relationof MTX-activatedcurrent was linear. Themean opentime was about 10 msec. These data suggest that MTX activates Ca2 channelswith novel properties, MTX at10 ng ml ' ormore produced a gradual risein diastolictension, Myocardial cells were characterized ultrastructurally byseverely overcontracted sarcomeres and swollen mitochondria, These mor- phologicalchanges were abolished in the Ca~+ free solution, suggesting that MTX increasesthe Ca~+influx througha newclass of Car'channels, to create theCaz' overloaded state and causing the observed cardiotoxic effects,

INTRODUCTION Ciguaterais a diseasecaused primarily by theingestion of a varietyof fishesinhabiting tropical and subtropicalseas. It is well knownthat the principaltoxin in ciguaterafish is ciguatoxin CTX!, a highlytoxic substance ED' in mice,0.45 Itg/kg!. CTXhas been reported to inducea depolarization of the cell membraneof the pedalganglion, muscle and neuroblastoma NIE115'-3-'. CTX causesa marked releaseof neurotransmitter from auto- nomic nerve endings and an inotropic action4s-".Maitotoxin MTX!, a water-solublesubstance, has been isolated from the viscera of a surgeonfish, Ctenochaetusstriatus, and from the toxic dinoflagellateGambierdiscus toxfcus.MTX inducesa Caz+-dependentrelease of noradrenalinefrom PC12 cells and Ca~'-dependentcontraction of smooth muscle"~' . MTX has been Y, Qhizumi shownto causecardiotonic and cardiotoxic effects on heartmuscle'4-ts. Since Ca2'plays an important role in regulationof cellularfunctions, MTX has been widelyused as a quantitativeprobe by numerousinvestigators' . Kccently, MTX has beenshown to increaseformation of inositol phosphatcsfrom phosphoinositides,and cause the release ofarachidonic acid from phospholip- ids in PCI2 cells~42s.

MATERIALSAND METHODS

Mechanical response Guineapig left atria werc excised and mounted vertically in anorgan bath containinga Krebs-Ringerbicarbonate solution. Left atriawere electrically stimulatedat a frequencyof 2 Hz by rectangularpulses of 5 msecat supra- maxima! intensities. Isometric contractions were measured with a force-displacementtransducer. Guinea pig myocardialcells were prepared andthe beating activity of singlecells measured asreported previously'" " . Eletrophysiotogical experi ments. Guineapig myocardial cells were superfused at37'C with Tyrode's solu- tion. Electricalproperties of the myocyteswere examined by patch-clamp methodsusing firepolished pipettes~r. The current was measured by means of a patch-clampamplifier, data stored on tape through a digitalPCM data recordingsystem, and analyzed with a computer. Cytoplasmic-freeCa2+ concentration. Isolatedrat cardiacrnyocytes were prepared as describedabove, The [Ca2+];in rat myocardialcells was measured according to themethod of Powellet al, withslight modification>s. The cells were suspended in Tyrode's solutioncontaining 0.5 mMCaClz and incubated at 30'Cfor 1 hr in the presenceof 50p.M Quin2 acetoxymethyl ester. The Quin2-loaded cells were thenwashed twice, suspended in Tyrode's solution containing I mM CaClz and kept at room temperature. Electronmicrograph experiments After treatment of the atria with MTX or CTX, the atria were rinsed quicklywith a fixativesolution 0'C! containing2.5% glutaraldehyde in each respectiveincubation media, followed by immersio~in thesame fixative solutionfor a furtherhour at roomtemperature. Subsequently, tissues were rinsedrepeatedly in 100 mM phosphate buffer pH 7.4!and postfixed with buffered1% osmic acid for 1 hr. Followingstandard dehydrating procedures usingethanol, the tissues were embedded in Eponand sliced thinly with a diamondknife, Thesesections were doubly stainedwith uranyl acetateand leadcitrate, and then observed in a ]EM100-8 electron microscope>7.

72 Actionof Ciguatoxinand Maitotoxin

RESULTS

Mecharticalresponse CTXcaused a concentration-dependent inotropiceffect on the guinea pig isolatedleft atria in therange of 0.1 to 10ngrnl', Theinotropic effect of CTX wasmarkedly inhibited by treatment wtth tetrodotoxin TTX, 5 x '107 !Vl!,and at concentrationsof 3 x10 5 M,TTX abolished the positive inotropic action of CTX.The inotropic action of CTX at lower concentrations .1-0,3 ng ml >! was completelyinhibited by practolol 0 s M!or reserpine mg kg' daily,For 3 days!,whereas athigher concentrations .6-10ng ml'> CTX action was only partiallyinhibited by both d rugs Table 1!. At concentrationsabove 3 pgml-', CTXcaused a biphasicinotropic change and finally led to cardiac arrest.

Tablel. Antagonistsofthe lnotropic Effect of CTX on Guinea Pig Atria.

increase in Drug Dose Before After tension of CTX "f~ CTX Vo CTX 'k n /mI!

None ! 0 100.0+ 22.4 100,0+ 22.4 0.5 i 0.2

TTX! 5 x 10' M -47.1+ 5.5 21,5+0.1' 68.6+5.5' 8,9f 3.8'

Practolol! ]0 ' M -18.4+ 5.9 22.4+ 5.7' 40.8+5.0' 0.6+0.1

Reserpine! 2 mgkg' day' 0 80.1 k 1.5 80.1+ 1.5 0.6+ 0.1

Effect~of tetrodotoxin TIX!, practolol and reserpine onthe positive inotropic effectof ciguatoxin CTX! in theguinea-pig left atria. The maximum response to CTX in control is expressedas 0%, Eachvalue indicates the mean+ S.E.mean andnumber~ in parenthesesare the number of experiment~.'Significantly dif- ferentfrom the maximumresponse in theabsence of CTXat P <0.05.

ln the isolatedguinea pig left atria,MTX .1 to 4 ng ml'! causeda concentration-dependentinotropic effect. The MTX-induced inotropic effect wasnearly abolished by Co ' mM!, but waslittle affected by propranolol V'M!, reserpinerng kg-',twice!, or TTX x 10 M!, MTXat concentrations above5 ngml-' causeda biphasic inotropic change and a gradualrise in diastolictension of theatria, and finally led to cardiacarrest. The increase in theresting tension with concentrations in the range of 5 to10 ng ml t occurred ina concentrationdependent manner. The MTX ngml'!-induced increase in tensionwas abolished by C

73 Y. Ohizumi

In the isolatedatrial cells,CTX ng ml-'! induceda time-dependent increasein thedegree and rate of longitudinalcontractions, After theapplica- tion of CTX,the degreeof contractionswas increased by 110%.The cellular motionwas changed into irregularbeating and myocytesthen gradually shortened.MTX .1-10 ngml-'! causedan increase in thedegree and therate of contractionsof myocytes. The cellular motion was changed into irregular beatingand then shortened with a fibrillatorymovement. The proportion of myocytesshowing arrhythmic motions increased with time,attained a maxi- rnum, and then slowly declinedwith a concomitantincrease in the percentage of damagedround ce!ls. In thepresence of verapamil0 s M!,the change of normallybeating rnyocytes into cells showingarrhythmic movements was delayed,and regularly beating cells still remainedin the preparation. Electrophysiologicalexperiments In thewhole-cell patch-clamp experiments, CTX shiftedthe voltage de- pendenceof Na+channel activation toward more hyperpolarized potentials without affectingthe time courseof inactivationand the peakcurrent ampli- tude. In thepresence of CTX0 ng ml '! Na' currentswere inactivated after the depolarizingpulse and then a small sustainedinward current was observed,The holding current at -80mV shiftedin theinward direction in the presenceofMTX .3 ngml-1!. This inward current was enhanced by adrenalin andabolished by Cd~+.The I-V relationof MTX-activatedcurrents was linear and the reversalpotential was -23 inV. In the cell-attachedpatch-clamp ex- periments,MTX induced long bursts of channelopening events. The unitary conductancewas 12pS andthe meanopen time of MTX-activatedchannels was 10 ms. CytoplasmicCa-'+ cottcentration In Quin2loaded cardiac myocytes, the [Ca2+];increased from a resting level of 1161 9nM to a maximum of 750+ 46 nM after treatment with MTX ng ml-' or more!.

Ultrastructnre In the guineapig left atriatreated with CTXat 3 !ig ml-' for 60 min, the mitochondria were swollen and the sarcoplasmcontained few glycogen gran- ules, Markedultrastructural changes were observed 30 min after application of MTX0 ng inl '!, The sarcomereswere severely contracted. In somecases, a localizedovercontraction proceeded to the extentthat the stretchedareas betweencluinped myofibrils were partly emptyof contractilematerial, and brokensegments of myofibrilswere scat tered between the condensed contrac- tion clumps,Mitochondria were heavily swollen and showedan increasein the matrixspace with disorganizedcristae. The chromatin of the nucleuswas aggregatedperiphera!!y and the sarcoplasmwas clear and containedfew

74 Actionof Ciguatoxinand Maitotoxin glycogengranules, InCa2'-free medium no apparent ultrastructural change wasobserved, suggesting that these morphological changes were dependent on external Ca~'.

DISCUSSION In the guineapig left atria, CTX causeda cardiotonicaction'44". This effectwas markedly inhibited by Tl'X. Theinotropic effect of CTXat lower c'oncentrationswas abolished by practololand reserpine, while that of CTXat highconcentrations waspartially inhibited by boththe drugs, In thesingle atrialcells CTX produced an increase in thedegree of longitudinalcontraction, suggestingdirect action on thecell membrane.Furthermore, CTX shifted the voltage-dependenceof Na channelactivation to morenegative membrane potentials,suggesting that CTX modified the activation process of fastNa+ currents, CTXcaused prolongation of theaction potential duration and this effectwas also reversed by TTX,These results suggest that CTX-induced prolongationof the actionpotential duration was attributed to an increasein theNa' current.It is alsosuggested that CTX activated Na' channelsby modifying the voltage-dependenceof channel activation to increaseNa+ in- wardcurrents, thus producing an increasedCa~' availability in thecardiac musclecell9. It is possiblethat the sustained inward current induced by CTX at high concentrationscontributed to depolarization,thus producing cardiotoxic actions, MTX produceda positiveinotropic effect on cardiacmuscle'4-'r. MTX at high concentrationsinduced a risein diastolictension followed by atrial arrest,In myocardialcells MTX caused an increase in thedegree of contrac- tion and subsequentarrhythrnogenic actions. These cardiotonic and cardiotoxiceffects were inhibited by Caz'antagonists and were abolished by Caz+-freemedium, The intracellular free Ca~ concentration of myocytes was increasedmarkedly by MTX. Furthermore,the MTX-induced ultrastructural changeswere not observed in Caz+-freemedium, These results suggest that MTX increasesthe Caz' permeabilityof someCa + channelsto elevatethe intracellularCaz+ concentration, thus producing cardiotonic and cardiotoxic effects'7. ln orderto clarifythe mechanismof Ca~+-dependentcardiotoxic and cardiotonicactions of MTX,patch-clamp techniques were used to analyse eletrophysiologicalproperties of a sustainedinward current inducedbv MTX>s.This current was predominantly carried by Ca2' or Ba+, Quiteim- portantly,the current-voltagerelationship and unitaryconductance of MTX-activatedchannels were clearly different from those of voltage-depen- dentCaz' channels. MTX-activated channels had a meanopen time that was ten times longerthan that of voltage-dependentCa2 channels.These data Y. Ohizumi suggestthat MTX activates a new class of voltage-independent Cazchannels to increasethe Ca ' permeability,which may account for themechanism of Caz' - dependenteffects's.

Acknourledgements I amgrateful to ProfessorTakeshi Yasumoto of TohokuUniversity for generouslysupplying ciguatoxin and maitotoxin.

LITERATURECITED 1. Boyarsky,LL. andM,D. Rayner. 1970. The effect of ciguateratoxin on Apiysianeurons, Proc. Soc, Exp. Biol. Med. 134: 332-335. 2. Rayner,M,D. and T,I, Kosaki. 1970. Ciguatoxin: Effects on Na+ fluxes in frog muscle.Fed. Proc. 29: 548, 3. Rayner,M,D. 1972.Mode of actionof ciguatoxin.Fed, Proc, 31: 1139- 1145. 4. Lewis,R,J. and R. Endean. 1986. Direct and indirect effectsof ciguatoxin on the guinea-pigatria and papillarymuscles. Naunyn-Schrniedebergs Arch. Phamacol. 334: 313-322. 5. Bidard,J,N., I.P.M, Vijverberg,C. Frelin, E, Chungue, A.M. Legrand, R, Bagnisand M. Lazdunski.1984. Ciguatoxin is a novel type of Na+ channel toxin. J. Biol. Chem. 259: 8353-8357, 6. Ohizumi, YS, Shibataand K. Tachibana.1981. Mode of the excitatory and inhibitory actions of ciguatoxin in the guinea-pig vas deferens.J, Pharmacol.Exp. 217:475-4SO, 7. Ohizumi, Y., Y. Ishida and S. Shibata. 19S2.Mode of ciguatoxin-in- ducedsupersensitivity in the guinea-pigvas deferens. J. Pharmacol. Exp. Ther, 221:748-752. 8. Lewis,R.J. and R. Endean,1984. Modeof actionof ciguatoxin from the Spanishmackerel, Scomberomorus commersoni, on the guinea-pig ileum andvas deferens. J. Pharmacol. Exp. Ther. 228: 756-760. 9. Seino, A., M. Kobayashi, K. Mornose, T. Yasumoto and Y. Ohizumi.1988. The mode inotropic action of ciguatoxinon guinea-pig cardiac muscle. Br, J. Pharmacol. 95: 876-S82. 10. Takahashi,M., Y. Ohizumiand T. Yasumoto.1982, Maitotoxin, a Ca ' channelactivator candidate. J. Biol. Chem, 257; 7287-7289. 11. Takahashi, MM. Tatsumi, Y. Ohizumi and T. Yasumoto. 1983. Ca ' channelactivating function of maitotoxin,the most potent marine toxin known,in clonalrat pheochronocytomacells. J. Biol.Chem. 25S: 10944- 10949. 12. Ohizumi,Y., A. Kajiwaraand T. Yasumoto.1983. Excitatory effects of themost potent marine toxin, maitotoxin, on the guinea-pig vas defer- ens.J. Pharrnacol.Exp, Ther. 227:199-204.

76 Actionof Ciguatoxinand Maitotoxin 13. Ohizumi,Y.and T. Yasurnoto, 1983, Contractile response ofthe rabbit aortato maitotoxin, themost potent marine toxin. J. PhysioL 337: 711-721. 14. Legrand,A.M. and R. Bagnis, 1984. Effects ofciguatoxin and maitotoxin onisolated rat atriaand rabbit duodenum, Toxicon 22: 471-475, 15. Kobayashi,MG. Miyakoda, T. Nakamura and Y Ohizumi.1985a. Caz+-dependent arrythrnogenic effects of maitotoxin, themost potent marine toxin known, on isotatedrat cardiacmuscles. Eu r. J. Pharmacol. 111; 121-123. Kobayashi,M.,Y. Ohizumi and T. Yasumoto. 1985b. The mechanism of actionofmaitotoxin inrelation to Caz' movements inguinea-pig and rat cardiacmuscles, Br. J, Pharmacol.86: 385-391. 17, Kobayashi, M., S. Kondo, T. Yasumoto and Y. Ohizumi.1986. Cardiotoxic effects of maitotoxin,a principaltoxin of seafoodpoisoning, on guinea-pig and rat cardiac muscle, J.Pharrnacol. Exp. Ther. 238:1077-1083. 18, KobayashiM., R. Ochiand Y. Ohizumi.1987. Maitotoxin-activated singlecalcium channels in guinea-pigcardiac cells. Br. J. Pharmacol.92: 665-671. 19 Freedman,S.B., R.J, Miller, D.M. Miller and D.R. Tindall. 1984.lnterac- tionsof maitotoxinwith voltage-sensitivecalcium channels in cultured neuronal cells. Proc. Natl, Acad. Sci. 81:4582-4585. 20, Schettini,GK. Koike,I.S. Login, A.M. Judd, M.J. Cronin, T. Yasumoto and R.M, Macleod, 1984.Maitotoxin stimulates hormonal release and calciumflux in rat anteriorpituitary cells in vitro.Am. J. Physiol. 247:E520-E525. 21, Kim,Y,l., LS Login and T. Yasumoto. 1985. Maitotoxin activates quantal transmitterrelease at theneuromuscular junction: evidence for elevated intraterminalCaz+ in themotor nerve terminal. Brain Res, 346: 357-362, 22, Ueda,H., S. Tamura, N. Fukushimaand H. Takagi.1986. Pertussis toxin AP! enhancesmaitotoxin a putativeCaz+ channel agonist!-induced Caz'entry into synaptosomes. Eur. J. Pharmacol. 122: 379-380. 23. Niki, l., T. Tamagawa, H. Niki, A. Niki, T. Koide and N Sakamoto,1986. Stimulation of insulinrelease by maitotoxin, an activa- tor of voltage-dependentcalcium channels, Biomed, Res. 7: 107-112. 24. Gusovsky,F., T, Yasumotoand J.W.Daly. 1989,Calcium-dependent effectsof maitotoxinon phosphoinositidebreakdown and cyclic AMP accumulation in PC12 and NCB-20 cells. Mol. Pharmacol. 36: 44-53. 25. Choi,O.HW, L. Padgett,Y. Nishizawa,F. Gusovsky, T. Yasumotoand J,W. Daly. 1990. Maitotoxin: Et'fects on calcium channels, phosphoinositide breakdown, and arachidonate release in pheochromocytomaPC12 cells. MoL Pharmacol,37: 222-230. 26- Ohizumi,Y., M. Kobayashi,A. Muroyama,H. Nakamuraand J. Kobayashi.1988. The mechanism of theinotropic action of striatoxin,a Y. Ohizumi

novelpolypeptide toxin from a marinesnail, in isolatedcardiac muscle Br.J. Pharrnacol,95: 867-875. 27. Hamill, O,P., A, Marty, E, Neher, B. Sahmann a nd 1;J Sigworth.1981, fmproved patch-clamp techniques for high-resolution currentrecording from cellsand cell-freemembrane patches, Pflugers Arch. 391: 85-100. 28. Powell,TP,E.R. Twistand V.W. Twist. 1984. Cytoplasmic free calcium measuredby Quin2fluoresence in isolatedmyocytes at restand during potassium-depolarization.Biochem. Biophys. Res. Com. 122:1012-1020 Inhibitionof Skeletal Muscle Response toAcetyl- cholineby Dinoflagellateand Ciguatoxic Fish Extracts

G.Escalona de Motta, J.A, Mercado', T.K. Tosteson-' and D.L.Ballantine2 UniversityofPuerto Rico, l Medical Sciences Campus, Institute of Neurobiology,203 Blvd. del Valle, San Juan, I'.R. 00901-02, and2MayagCiez CampusDepartment of Marine Sciences, P.O. Box 5000, Mayagiicz, PR 00681

ABSTRACT Theefiect of toxicdinoflagellate and ciguatoxicfish extractson the con- tractilityof fastskeletal muscle fibers was studied recording the amplitude andtime course of twitcheselicited in isolatedfrog sartorius muscles by either bath applied acetylcholine ACh! or high concentrationsof extracellular potassium,Toxic extractswere obtainedby methanolextraction, From the dinoflagellateOstreopsis lenticularis and from ciguatoxic barracuda Sphyraetta barrttcuda!,harvested along the southwest coast of PuertoRico. In thepres- enceof 20Itg/mL of theciguatoxic barracuda extract, the amplitude of the twitchesproduced by either ACh or potassiumwas reduced to approximately 10to 20%of the control, A similarinhibitory effect was observed after addi- tionof 500pg/ml of O. !euticttlarisextracts, Veratridine increased by approximately 0% the amplitude and duration of ACh-induced muscular twitches.These potentiated responses were reversiblyinhibited by the ciguatoxicfish extracts.Muscle twitches elicited by potassiumwere also blockedby the AChreceptor blocker alpha bungarotoxin and by low calcium, high magnesiumsolution, Taken together these pharmacological observa- tions suggestthat the inhibitory componentpresent in thesetoxic extractsis an antagonistof thesubsynaptic nicotinic cholinergic receptor.

IN TRODUCTION Ciguaterais a complexhuman toxic syndrome caused by theingestion of fishassociated with tropicalcoral reefs. It is manifestedby early and generally severegastrointestinal symptoms followed by sensoryand motor abnormali- ties, that appear somewhat later in the toxic episode of longer duration'--' Besidesnausea, vomiting and diarrhea, long lasting paresthesias of the lower extremitiesand generalized motor weaknessare the most freciuentlyreported clinical manifestations4 s. G. Escalona de Motta, et al.

Toxinsproduced by benthicdinoflagellates, in particular Garnfnergiscus toxicus,its ecological associate Ostreopsis lenticular'>s, have been implicated as ciguateracausingtoxinss r, Basedon their natural abundance and toxicity the dinoflagellatesProrocentrunt lima, P. concave>um andO. siamensis are also likely sourcesof ciguaterarelated toxinss. Recently, the structure of theciguatoxin isolatedfrom the morayeel Gyrnnothoraxjatunicus and of one of the toxins presentin G,toxicus were elucidated". These authors proposed that theG toxieustoxin was a precursorof themoray eel ciguatoxin. However, the total numberand structures of ciguatera toxins remain unclear. Amongthefish associated withciguatera syndrome, barracuda Spltyraena barracuda!is one of the tnostfrequently reported in the Caribbeans.Tosteson et al. found that extractsobtained from the tissues,particularly the head and viscera,of one third of the barracudaspecimens caught along the southwest coastof PuertoRico were ciguatoxic as determinedby the mousebioassay method'"", Toxicity of thesebarracuda tissues exhibited seasonalvariations, a characteristicthat may reflect yearly fluctuations in the abundanceof toxic, benthicdinoflagellates. Ballantine etal, observed seasonal fluctuations in tox- icityand population density in naturalpopulations of O.lenticularis harvested in this coastal area'~. Methanol extracts of clonal cultures of both G. toxicus and O. lenticularis isolated from the coastal waters of southwest Puerto Rico are toxic to mice'~ In this work, we found that toxic methanol extracts from ciguatoxic barracuda and 0. lenticularisinhibited contraction»of frog sartorious muscleselicited by appliedand nerve-releasedacetylcholine, These results suggestthat a compo- nent among those present in these crude, complex extracts, acts as an antagonistof the nicotinic cholinergicreceptor.

MATERIALSAND METHODS

Biotogicatpreparation Small"! specimensof the frog Ranapipiens were sacrificedby rapid destructionof thebrain and spinal cord and dissected under the microscopeto isolateboth sartoriusmuscles joined by the symphysisof the innominate bones.The isolated muscle pairs were kept in normalfrog Ringer's with the followingionic composition: 120 mM NaCI,2.1 mM KCI,and I.87 mM CaCI~ bufferedto pH 7.2 with 5 mM of N-Tris hydroxymethyl!methyl-2- aminoethanesulfonic acid and NaOH. Neostigmine-6 ItM! wa»routinely addedto thissolution to inhibitmuscle acetylcholinesterase activity. Recordingof musclecontraction In a contractionchamber, muscles were fastened by the joined pelvic ends to a fixedsupport and by thetibial ligaments to a forcetransducer Gould StathamVCB! which was on linewith a pre-amplifiercoupled to a chart Toxininhibition oF Skeletal Muscle Response recorder Beckman R511A! operating ata sensitivity of20 mm/g of force and a chartvelocity oF2.5 mm/sec. Muscles were stretched slowly to a resting tensionof 250-500 mg. Application ofa standardconcentration seebelow! of acetylcholine ACh! to the paired sartorius muscles inthe bath under these partiallyisometric conditions, resulted in a seriesof summedmuscular twitchesorfasciculations thatdecayed with time as the muscle fatigued, The amplitudeof thesetwitches was recorded aspen displacements in the chart andmeasured in mm!every four seconds. These amplitude values were averagedand plotted using the Sigma Plot computer program jandel Scien- tific, CA.! Reagentsand drugs Musclecontractions were activated insome experiments bydepolarizing thepreparations with AChas the chloride salt SigmaChemical Co.!, dis- solvedin the bath at a finalconcentration of5.5 x IO" M.In other experiments muscleswere depolarized adding KC1 to the bath solution in quantitiessuffi- cient to increasethe extracellularK' concentrationto 10 mM. In some experimentsveratridine and alpha-bungarotoxin obtained from Sigma ChemicalCo.! were added to thefrog Ringer's. Toxicextracts of thedinoflagellate Ostreopsrs fenficularis and of ciguatoxic barracuda Sphyraena harracl/da!, prepared as described by Tosteson etal. were

HR

Fish0 min!+ Veralvkane0 sec!/W

~ 000 4 Fish0 min!/W/Vorahhana0 aas!/W

4 S.Sx10'll ACh

Figure1. Fffectof DinoflageIlateExtracts on ACh Induced Muscle Contractions. Representativerecords of the inhibitory effectof O. /en/jcafarisextracts on twitchesinduced by applyingACh to frog sartoriusmuscles. Control responses to AChwere obtained in normalfrog Ringer's NR! before top record!and after bottomrecord! exposure of the musclesto %0 itg/ml of the extract middle record!,Filled arrow heads indicate application of AChwhile open arrow heads signalbeginning of washout periodwith NR. G. EscaIona de Motta,et al. dissolvedin smallvolumes of distilledwater 0 mg/ml of O. Iertffcuiarfs extracts,2 mg/ml of 5.barracuda residue! for their additionto thesolution bathingthe muscle '" ' . Theoily ciguatoxic barracuda extract was previously solubilizedin p.lvolumes of purifiedfortnamide Fisher Scientific, pA.!

RESULTS Applicationof ACh ,5 x 10 t' M! tomuscles previously exposed for 10 minto frogRinger's contairting 20 Iig/mL of theciguatoxic barracuda extract or 500Itg/mL of theO. lenticularis extracts, resulted only in verysmall and feeblemuscle twitches Figure 1!. This inhibition of theACh-induced muscle twitchesin thepresence ofeither extract was reversed by washing the prepara- tion severaltimes with normal Ringer'ssolution. A five fold increasein the frog Ringer'spotassium concentration to 10 mM!produced forceful and multiple muscle twitches which decayed very L"

ta M a Burx IO min! 9 x 10' M Ce' + B.tt a 10' M MB"

tomM ltt:I Wash HR 10mM KCt.

Figure2. Effectof IncreasedConcentrations of ExtracellularKt on IViuscle Contraction. Recordsillustrating the responseof two sartorius muscle pairs to an increasein IK+}pto IO nM. Top recordsshow the rapidly decaying fast contractions pro- duced by high K+ tn normal frog Ringer's.Treatment of the preparations with eitherthe ACh receptor blocker alpha bungarotoxin bottom left! or a lowCa2+, highMg~+ Ringer's solution bottomright! abolishedthe fastcomponents of ihe responsesand resulted m slow, tonic contractures. Filled arrow heads signal applicationof high externalpotassium and open arrow headsindicate beginning of wash out period with NR. rapidly seeFigure 2, top records!. TheseK' mduced twitches were inhibited by blockingthe motor end plate nicotiniccholinergic receptor with alpha- bungarotoxin,or inhibiting the release of AChfrom the motor nerve termmal by equilibratingthe preparationin a solutionof low Caz' and high Mg ' concentrations,Under these conditions only thesesustained or tonic muscle Toxininhibition of SkeletalMuscle Response

contractureswereobserved see Figure 2,bottom records!, demonstrating that the fastmuscle twitches were elicited by ACh releasedtrom motornerve terminalsdepolarized byhigh extracellular K'. Asseen in Figure 3,these fast contractionswere inhibited in thepresence of fishand dinoflagellate extracts.! lowever, upon return to normalRinger's, recovery of thefast re- sponsetohigh lK la was not complete in these preparattons reflecting, most probably,a decreasein the amount of AChreleased from nerve terminals with repeateddepolarizations. increasing lKlu to 40nM caused only slow, tonic

0,035

0,028

0.021

0.014

0,007

0.000

Figure3, Effectof CiguatoxicBarracuda and DinoftagetlateExtracts on the Amplitude of K+ induced Muscle Contraction. Amplitudeof musctetwitches in frogsartorius muscles depolarized by increas- ing [K+}a to 10 mlvl. The area under the curve was measuredto determine magnitudeof the response expressed in arbitraryunits! bothin the presencenf 20lig/ml of theciguatoxic barracuda extract right diagonals, n=2! and 500 mg/ mlof dinoflagetlate extract left diagonals, n=2! and in theabsence ofany of these extracts,Solid bar indicates muscles response in nortnalfrog Ringer's n=7! while openbar depictspartial recoveryof the responseafter washingout the toxic extractswith normal Ringer's n=6!. Vertica! lines indicate f S.E. musclecontractures that werenot affectedby theciguatoxic fish extract. The observedinhibitory effect of thesetoxic extractson tnuscle contrac- tionsinduced by eitherexternally applied ACh or by K' stimulationof ACh releasefrom nerve terminals suggested that a componentacting as an antago- nist of the nicotiniccholinergic receptor was present. flowever, using chromatographicallypure morayeel ciguatoxin,Molgo rt al, demonstrated thatthis toxin exerts an excitatory action on amphibian muscle, mediated by an increasein the membraneNa+ permeability>4, To explorea possiblerela- ttonshipbetween the action of thesecrude extracts and the musclemembrane Na+conductance, we exposed the muscles to veratridinc. Thisplant alkaloid

g3 C;.Escalona de Motta, ef al. isknown to shift the activation ofthe voltage-gated Na'cha~nels tomore negativemembrane potentials andtoblock channel inactivation'>. Thesetwo actionsresult inprolonged activation ofNa' channels atresting membrane potentialsandwould, thus,amplify anydirect depolarizing actionbysimilar componentspresent inthe extracts tested here. Short0 sec!pulses of10+ M veratridineapplied tomuscles in normal frogRinger's hada directexcitatory action. Treatment withveratridine did enhance,byapproximately 100%,the amplitude andduration ofthe response

20

'l5 E E Ct1054 120 18 8

TOE{sec.!

Figure4. Effect of Veratridine onACh Induced Muscle Contractions. potentiatingeffect of a 30sec pulse of 10 s M veratridine onthe fast contraction elicitedapplying 5.5x EO*M AChtofrog sartorius muscles solid circles, n=5!. Theresponse to5.5 x 1%'M ACh obtained innortnal frog Ringer's

of thesemuscles to externallyapplied ACh seeFigure 4!. No excitatory responsewasobserved inmuscles treated with veratridine inthe presence of ciguatoxicbarracuda extract and addition ofthis alkaloid did not antagonize the inhibitoryeffect of thefish extracton ACh-inducedmuscle twitches,Similar to controlpreparations a veratridine pulse applied to a musclewashed with normal Ringer's after treatment with the barracuda extract,did resultin anenhanced ACh response Figure 5!. Asseen inFigure 6,muscles whose response toACh had been potentiated byveratridine were still sensitive to the inhibitory action of theciguatoxic barracudaextract. This inhibition was easily reversed by washingthe prepa- rationwith normal Ringer's. Incontrast, thepotentiating effect of veratridme ToxinInhibition of SkeletalMuscle Response

NR

Oino Wild 0 min! Lo.sg j cue

HR 0 min!

$.5x10r II ACh Waeh HR Figure5. Effectof CiguatoxicBarracuda Extracts on Veratridinepotentiation of ACh Induced Muscle Contractions. Effectof a 30sec pulse of I 0~M veratridineonthe fast contractions eltcited by 5.5 x I0.< M ACh to frog sartoriusmuscles. Middle recordshows the musclere- sponseto ACh during a 10min incubationperiod with ciguatoxicbarracuda extract0 Itg/m !. Bottomrecord was obtained after washing out theprepara- tionwith normalFrog Ringer's NR!. Solid arrow heads indicate application of AChwhile open arrows heads signal beginning oF wash out period with NR. wasvery long lasting and was observed again when ACh was applied to musclesthat had been washed with normalRinger's. These results further indicatedthat inhibition of musclecontractility by theseciguatoxt'c extracts was mediatedby an antagonismof ACh action.

DISC VSSION Resultsreported here provideevidence that toxicmethanol extracts ob- tainedfrom tissues of ciguatoxic barracuda and from the benthic dinoflagellate O.lerrficrtlaris contain a component that antagonized thestimulatory effect of AChon the amphibian skeletal muscle fiber. Pharmacologically thiscompo- nentappears to be actingas a reversibleblocker of the nicotiniccholinergic receptor.Fast contractions or twitchesof frogsartorius muscle fibers acti- G. Escalonade Motta, ei al.

30 E 25

20 D 15 CL1004 0 128 1820 24 28 32 36 40 44 TIME sec. !

Figure6. Effect of Ciguatoxic Barracuda Extracts on the Amplitude of Veratridine Potentiated ACh Induced Muscle Contractions. Potentiationof the contractionsproduced applying 5.5 x 10s M ACh to frog sartoriusmuscles by 0+>veratridine solid triangles, n=2! was reversibly abol- ishedin the presenceof 20pg/m!ciguatoxic barracuda extract opentriangles, n=2!. When washed with normal frog Itinger's, the amplitude of the ACh- elicitedtwitches recovered to previous,potentiated levels solid trianglesand circles, n=2!. vated by ACh, either externally applied or released from the motor nerve terminalsby a K+depolarization, were similarly inhibitedby barracudaand dinoflagellateextracts. Musclesexposed for a short time to a low concentra- tion on veratridineexhibited a long lasting,nearly irreversible,enhanced responseto ACh. Theseveratridine-potentiated responses were inhibited in the presenceof ciguatoxicbarracuda extract but recoveredfully washingthe preparationwith normal Ringer's, further suggesting an antagonismof ACh receptorsby theseextracts, ln contrastwith our observations,recent reports concerning the action of chromatographicallypure tnoray eel ciguatoxinon frog neuromuscular trans- mission, demonstrate that this toxin facilitates ACh release from motor nerve terminalsand inducespostsynaptic membrane depolarizations that activate spontaneouscontractile activity in cutaneouspectoris muscle fibers . How- ever, it is not yet clear that this ciguatoxinis the single causativeagent involvedin ciguateraintoxication. The broad spectrum of clinicalsymptoms experiencedby ciguateravictims as well as the large number and variety of toxinsfound in thevarious dinofiagellate species implicated in this poisoning Toxininhibition of SkeletalMuscle Response

suggestthat more than one component isinvolved inthis toxic syndrome ". Ciguateraintoxication produces a paralysisof thelower extremities in somepatients. This has been ascribed to a curarizingneurornuscular block as it isalleviated by the administration of acetylcholinesterase blocking agents'7,This clinical observation suggests thatat least one toxic component associatedwith the ciguatera syndrome may be an antagontst of the nicotinic cholinergicreceptor. Chromatographic purification of theciguatoxic barra- cudaextracts used in thepresent study has produced several different toxic fractions'".Further experiments are needed to determineif one of thesefrac- tionscontains the proposed cholinergic blocker.

LITERATURECITED 1. Bagnis,RT, Kuberskiand S. Laugier.1979. Clinical observations on 3,009cases ofciguatera fish poisoning! inthe South Pacific. Am. J. Trop. Med. Hyg. 28: 1067-1073, 2. Lawrence, D.N., B.M, Enrique, R.M. Lumisk and A. Maceo,1980. Ciguatera fish poisoining in Miami.J. Arn. Med. Assoc, 244:254-280. 3, Casanova,M.F. 1982.Ciguatera poisoning: Medical letter. Arch. Ncurol, 39:387. 4. Maharaj,S.R., P. Kulkarniand lVl.A. Pinnock,1986. Ciguatera fish poisoningin a Jamaicanfamily. W.I. Med. J. 35: 321-323, 5, Escalonade Motta, G., J. Feliu and A. Izquierdo.1986. identification and epidemiologicalanalysis of Ciguateracases in PuertoRico, Mar. Fish. Rev, 48; 14-18. 6. Yasumoto,T., I. Nakajima,R. Bagnis and R. Adachi, 1977. Finding of a dinoflagellateas a likelyculprit of ciguatera.Bull. Jpn. Soc, Sci, Fish. 43: 1021-1026, 7. Yasumoto,T., I. Nakajima,Y.Oshima and R. Bagnis.1979. A newtoxic dinoflagellatefound in associationwith ciguatera, In: Toxic Dinoflagel- late Blooms, F. J. R,Taylor and H. Seliger,eds,!, Elsevier,New York, pp, 65-70. 8, Anderson, D.M. and P.S, Lobel, 1987.The continuing enigmaof ciguatera.Biol. Bull. 172:89-107. 9. Murata, M., A,M. Legrand, Y. Ishibashi, M. Fukui and T. Yasumoto.1990. Structures and configurationsof ciguatoxinfrom rno- ray eel Gymnnthoraxjavarticus and its likely precursor from the dinoflagellate Garnbferdiscustnxicus. J. Am. Chem.Soc. 112: 4380-4386. 10. Tosteson,T.R., D.L. Ballantine and H.D. Durst. 1988 Seasonal frequency of ciguatoxic barracudain southwest PuertoRico. Toxicon, 26: 795-801. ll. Hoffrnan, P.A., H.R. Granade and J.R, McMillan, 1983.The mouse ciguatoxin bioassay:a dose-responsec~rve and symptomatologyanaly-

87 G. Escalona de Motta, et af.

sis, Toxicon. 21: 363-369. 12. Ballantine,D.L., T.R. Tosteson and A.T, Bardales.1988. Population dy namicsand toxicity of naturalpopulations of benthicdinoflagellates,n southwesternPuerto Rico, J, Exp.Mar. Biol. Ecol. 119; 20'1-212. 13. Tosteson,T,R., D,L. Ballantine, C.G. Tosteson, V. Hensley and A. T Bardales.1989, Associated bacterial flora, growth and toxicity of cul tured benthicdinoflagellates Ostreopsis lenticularis and Gambierdiscus toxicus.Appl. Environ,Microbiol, 55: 137-141. 14, Molgo,J,, J.X. Comella, A.M. Legrand. 1990, Ciguatoxin enhances quan- tal transmitterrelease from frog motor nerve terminals. Br. J. Pharmacol 99: 695-700. 15. Catterall,W.A. 1985.The voltage sensitivesodium channel: a receptor for multipleneurotoxins. fn: Toxic Dinoflagellates, D. M. Anderson,A, W, White, D, G, Baden,eds.!. Elsevier,New York. pp. 329-342. 16. Tindall,D,R., D.M. Miller, and J.W. Somber. 1989, Culture and toxicity of dinoflagellatesfrom ciguateraendemic regions of the world. Toxicon. 27: 83, 17. Donnet, A., M, Habib, R,A. Vigouroux, A. Sourgeade and R, Khalil. 1987.Manifestations myastheniforrnes persistantes apres intoxi- cation ciguaterique:Effet favorabledes anticholinesterasiques. La Presse Medicale. 16: 734. 18. Tosteson, T.R., R.A. Edwards and D.G. Baden. 1992. Caribbean ciguatera and polyether dinofiagellate toxins: Correlation of barracuda ciguatoxinswith standardtoxins, In: Proceedingsof the Third Interna- tional Conference on Ciguatera Fish I'oisoning T.R. Tosteson ed.!. PolysciencePublishers, Quebec, Canada. Seepp. 101-116. CaribbeanCiguatera and Polyether Dinoflagellate TOxins:COrrelatiOn OfBarracuda Ciguatoxins With Standard TOxins.

T. R.Tosteson', R. A. Edward'-and D. G. Baden"- 1DepartmentofMarine Sciences, University ofPuerto Rico, Mayagiiez Campus,Mayaguez, Puerto Rico 00681, and Department ofBiology and LivingResources, Rosenstiel School of Marineand Atmospheric Science, UniversityofMiami, 46%! Rickenbacker Causeway, Miami, Florida 33I 49- l098.

ABSTRACT Tissuesamples ofciguatoxic barracuda Sphyrae~ra barracuda! caught along thesouth coast of Puerto Rico were extracted and the crude extracts chromato- graphicallypurified. Thesepartially purified toxic fractionswere further fractionatedusing reversephase high performanceliquid chromato- graphy.Ciguatoxic barracuda fractions showed different retention times on C-18columns and were consistently more hydrophobic than the polyether dinoflagellatetoxins, brevetoxin and okadaic acid, Oral and i,p. mouseas- sayswere analyzed to identifyaguatoxic fractions. Immunoassays and scxfium channel receptor assays were employed to compare HPLC fractions of thebarracuda ciguatoxins with brevetoxin, Based on relativetoxicity and comparativeassays with brevetoxinthere appear to be four distincttoxins in th»ciguatoxic barracuda extracts. The two most hydrophobic, toxic fractions weredistinct in thatone showed high i.p. toxicityand negligible similarity to brevetoxinPbTx-3, while the otherwas of highoral toxicityand showed comparableactivity to PbTx-3in bothirnrnuno and sodium channel binding assays.These data suggest that diverse, multiple toxins cause ciguatera fish poisoning in the barracuda.

INTRODUCTION Ciguaterafish poisoning is a humanhealth problem that affects the people of PuertoRico, and all personsliving neartropical seas for whommarine fish representa significantsource of food>~. This typeof poisoninghas been recogmzedas a public health problem in the United Statesand its terri- toriess7s.Ciguatera traditionally has been limited to tropicalregions; however,increased commercialization of tropicalreef fishes and tourism have increasedthe frequency of this typeof fishpoisoning among persons living in temperate portions of the United Statesand Canada4". T. R. Tosteson, el al.

Barracudaisa frequentlyimplicated fish species in ciguaterafish poison ing in theCaribbean3's, Ciguatoxicity of barracuda Sphyraenu barracuda! head,viscera and flesh tissues has been determined in specimenscaught a long thesouthwest coast of PuertoRico", Monthlyfrequencies of ciguatoxicbarra- cudashowed an apparent seasonal variability, suggesting that the toxins, at least in their activeform, were not accumulatedin barracudatissues for extendedperiods of time.Variability in barracudaciguatoxicity may reflect fluctuationsin the toxicityof smallerreef fish prey, seasonalfluctuations in toxic vectors benthicdinoflagellates and bacteria!and/or changesin the ability of the barracudato detoxify ingested poisons or their precursors.Benthic bacterial and/or algal micro-organisms were originally suggestedaspossible vectors of ciguateratoxin s!by Randall'. Toxinsassoci- atedwith thesemicrobes eaten by herbivorousfish would be passedthrough the "food chain" when these animals are consumed by larger carnivorous fish'3. The toxic benthic dinoflagellate Gambierdfscustoxfcus, initially discov- eredin FrenchPolynesia Gambier Islands!, appears to be directly relatedto abundanceand occurrence of ciguatericfish in the Pacific'4-~s.I'he structure ofciguatoxin isolated from the moray eel in the Pacific has been reported, and appearsto be closelyrelated to one of the toxic componentsisolated from Cambierdiscus torfcus'7's Toxic benthicdinoflagellates Ostreopsis lenticularis and Gambierdiscus toxicus have been isolated from coastal waters of southwest Puerto Rico'~ ~'. The abundance and toxicity of the more predominant benthic dinoflagellate,Ostreopsis lenticularis, showed seasonal variation2", Recent reportssuggest that bacteriamay play a rolein toxic dinoflagellate blooms, the toxicity of 0, lenticularisgrown in laboratory culture, and ciguatera fish poi- soning~ ~s.The preciseorigi~ of the toxins responsiblefor ciguatera has not beenresolved, The chemical and pharmacologicalrelationship between fish "ciguatoxins" and those toxins r'solatedfrom presuznedbenthic microalgal and/or bacterialvectors has not been conclusively resolved. Ciguatoxin iso- lated from the moray eel in the Pacific and one of the toxins isolated from Cambierdiscustoxicus have different histopathological and pharmacological actions-" ~" ~~. The diversity and complexity in pharmacological action attributed to ciguatoxic fish extracts "ciguatoxin s!"! no doubt reflect the variable sources of thesematerials and the probability that suchextracts contain more than one pharmacologicallyactive component~s~". The study reportedhere wascon- ducted to assessthe diversity of ciguatoxinsfound in the Caribbeanbarracuda andtheir relationship to theimmunological and pharmacological specificity of the standard polyether dinoflagellate toxins, brevetoxin I'bTx-3 and okadaic acid. BarracudaCiguatoxins in the Caribbean

MATERIALSAND METHODS

Taxi!1$ Ciguatoxicbarracuda tissues 80.4 kg! collected along the southwest coastofPuerto Rico were extracted inaqueous methanol MeOH!, followed by back-extractionof residual water with ethyl acetate EtAC!. IOHO Itgtoxic extract/gfish tissue extracted were recovered, fora totalof 195 gm. Toxic residueswere purified initially with Bio-Sil A and subsequently repeatedly fractionatedon SephadexLH20, alternately with MeOH/EtAC 0:50! and ethylalcohol 95%!, Peak toxicity fractions eluted from the LH-20 column at V,/V,ratios between 0,56 - 0.58. Pooled fractions within this range were used in the work reported here Brevetoxinswere purified from laboratory cultures ofPtychodisctts brn~is, andokadaic acid was obtained from Dr. Robert Dickey atthe FDA Dauphin IslandLaboratory. Synthetic tritiated PbTx-3 was produced from PbTx-2 by chemicalreduction employingcerium chloride and sodium boro- tritiide.Crude PbTx-3 was purified using reverse phase high performance liquidchromatography HPLC!. HPLC-purified toxinhad demonstrated spe- cific activities of 10-15 Ci/mM. HighPerformattce Liquid Chromatography Samplescontaining 20-50 mouse units of partially purified fish ciguatoxin weredissolved in 100% methanol and centrifuged 5,000 x g,10'! to remove particulatematerial. Clear supernatant solutions were subjected toreverse phase C-18! HPLC using 85% isocratic aqueous methanol as themobile phase.Fractions were collected at tenminute intervals, regardless of the numberofpeaks generated within a giveninterval. This procedure was un- dertakento preventloss of anyfractions generated, and allow for the reproductionofseparated fractions forfurther analyses. Following collection ofsix fractions 0 minutes!,themobile phase was changed to 100% methanol and a single '10minute fraction was collected. ToxicityAssays Methanolwas removed from the eluted fractions byflash evaporation and theresidues taken up induplicates of 1.0ml Wessonoil or 0.9%saline contain- ing 0.1%Emuflor EL-620. Fractions taken up in 0.9%saline-0.1% Emuflor, correspondingto 2 MU equivalentsof originalextract were injected i.p. into duplicatewhite Swiss mice female, 20-25 gm!, and corresponding fractions takenup in Wessonoil wereadministered by gavageto similarsets of animals.Arumals were observed for periodsof 48 hours.Individual bioas- sayswere terminated when the animals were unable to rightthemselves when placedon their backs, or when they were in obvious respiratory distress.

91 T, R. Tosteson,et al,

Immunoassay Theradioimmunoassay RIA! wascarried out accordingto the methodof Bigazziet aI,"'. This assay utilized specific antibody made against brevetoxin PbTx-3.In the work reportedhere, keyhole limpet hemocyanin KLH! was covalentlylinked to brevetoxinand the KLH- PbTx-3!conjugate used as the completeantigen. Purified PbTx-3 was dissolved in minimalvolumes of redistilledpyridine and succinylated with 10-fold molar excess succinic anhy- dride. Followingseparation of unreactedPbTx-3 and succinic anhydride from toxin-succinateusing thin layer chromatography0/30 ethyl acetate/light petroleum!,the free carboxyl function on the conjugate was covalently coup! ed to the e-amino function of lysyl residues on the KLH using standard procedures>',Following coupling, the mixturewas dialyzed against PBS pH 7.4!overnight, and the protein concentration adjusted to yield "toxin equivalents"of I mg/ml. A singlegoat wasimmunized with the KLH-PbTx-3conjugate on alternate weeks,with blood samplestaken during the interval weeksfor assessmentof toxin antibody titers. Bloodsamples were allowed to clot and the serum separatedby centrifugation,Stirred aliquots of antiserawere treated with 0,5 ml volumesof saturatedammonium sulfate and allowedto precipitate over- night at 4'C. The mixture was then centrifuged 000 g!, the supernatant solution decantedand saved. The solution was then brought to 50% satura- tion by the addition of more 0.5 ml volumes of saturated ammonium sulfate, and the precipitate allowed to form overnight once again. The precipitate from this secondcentrifugation was redissolvedin 0.3 ml volumes of original serum and dialyzed against PBScontaining 0.01%sodium azide. For long term storage,the antibody solution was dialized against distilled water, and aliquots approximately 25 ml! lyophilized.These may be reconstitutedin PBS pH 7.4!containing azide as needed,3' SynaptosomeBinding Assay Synaptosomeswere prepared in multipleruns, generally using material from 4-6rat brains.Frozen brains were purchased in multiplesof 30-50and storedat -80'C until used.Synaptosomes prepared according to themethod of Dodd et al. and storedas a pooledfraction until sufficientsynaptosornes generally from 20 brains! were prepared32.Protein was measuredon resuspendedintact or lysedsynaptosomes just prior to bindingexperiments usingthe technique of Bradford>>.Synaptosornes were stable and theiractiv- ity in bindingexperiments reproducible for a periodof 2-3months. Bindingof tritiatedbrevetoxin PbTx-3! was measured using the rapid centrifugationtechnique"4. Binding assays were performed in a bindingme- diumconsisting of 50mM HEPES pH 7.4!,130 mM choline chloride, 5.5 rnM glucose,0.8 mM magnesium chloride, 5.4 mM potassium chloride, 1 mg/ml BSA,and 0.0I% Emuflor EL 620 as an emulsifier for the toxin. Synaptosomes

92 BarracudaCiguatoxins in theCaribbean 0-80ltg total protein!, suspended in 0.1 ml binding medium minus BSA, wereadded to a reactionmixture containing tritiated I'bTx-3 and other effec- torsin 0.9% ml binding medium in1,5 rnl polypropylene micnxentrifuge tubes.After mixing and incubating at4 'C for 1 hour, samples were centri- fuged5,000 g, 2'!. Supernatant solutions were aspirated from each tube and thepellets rapidly washed with several drops oF a washmedium~". Pellets werethen transferred toliquid scintillation vials and bound radioactivity measured.Non-specific binding was measured in the presence of 10 AM PbTx-3and was substracted from total binding to yielda calculatedmeasure ofspecific binding, Free tritiated probe was measured bycounting analiquot ofsupernatant solution prior to aspiration.

RESULTS

HPLC Analyses Figure1 showsa representative elution pattern of thepartially purified barracudaciguatoxin, monitored ata wavelengthof215 nm, Following an initialpeak of material,the most significant quantities detected were in frac- tions5 and 7. Fraction7, the most hydrophobicmaterial found in the barracudaextracts, was recovered from theHPLC column after the 60' elution

k-k U-U 5-8 V-7 V-T O-O

Figure1. HPLC Fractionation of BarracudaCiguatoxin. HPLCelution pattern of ciguatoxicbarracuda fractions; C-18 reverse phase column,4.6 mm x 25 cm; flow rate 1.4ml/min.; mobilephase.' isocratic 859o aqueousmethanol, Detector at a wavelength of 215 nm. Fraction symbols noted at the bottom of each fraction eluted.

93 T. R, Tosteson, et al. periodby washingthe column with 100% methanol. Ciguatoxic barracuda fractionsshowed different retention times on C-18columns and wereconsis tentlymore hydrophobic than the polyether dinoflagellate toxins, brevetoxin and okadaicacid. The brevetoxinsand okadaicacid would elute off the columnin fractions1 and2~, Therelative quantities mg! of materialrecov- ered in the fractions reflected those detected by the UV monitor, with the exceptionthat only 22% of the total material recovered in the seven fractions was found in fractions5 7%! and 7 %!, while the secondlargest quantity 3%! of recoveredmass was found in fraction 3, undetectedby the UV monitorat 215 nm. Thelargest fraction of recoveredmaterial 6%! wasfound in fraction1. Eachfraction recoveredis identifiedby an individualsymbol, as indicatedin Figure'l. Thesesymbols are used in subsequentillustrations of the resultsobtained in the radioirnmuno and synaptosomebinding assays.

Tnidcity Assays Resultsof thetoxicity bioassays were numerically ranked as follows, if the animals died within 24 hours the result was scored as a 6, if the animal survivedbut wasclearly sick, it wasscored as a 1, and if the treatedanimal neitherdied nor appeared to be sick,as a 0.All animalsused in thebioassays eitherdied or appearedto be sick. Thus,all sampleswere scored either 6 or 1. Peakoral toxicity was found in fractions1, 2 and7, while peaki.p. toxicity was seenin fractions2 and5. Fraction2 wasdistinct in that it was maximally toxic for both oral and i.p.routes of administration, whereasfractions 5 and7 were distinguished by maximum toxicities in i,p, and oral assays respectively.The pattern of toxicity oral > i,p. route! of the material recov- ered in fraction I was similar to that seen in fraction 7. Fractions 3,4 and 6 did not show marked toxicity in either assay,however, test animals were clearly and reproducibly sick after the administration of the materials collected in each of these fractions. Radioimiir unoassay Activities of the respectiveI IPLC fractions in the RIA and synaptosome binding assayswere scoredfrom 6, for those fractions showing the greatest degreeof specificdisplaceinent, to 'I for thosedisplaying the leasteffect. For purposesof analysis,scores of 6 and 5 were taken as indicative of significant activity, scoresof 4 and3 ofmoderate activity and scoresof 2 and1, of minimal or negligible activity. Theresults of the RIAdisplacement assay are illustrated in Figure2. The % maximumbinding abscissa!refers to the labeledbrevetoxin probe, and representshere the % ofbrevetoxin probe remaining bound in the presenceof variableconcentrations of the test materials. Increasingamounts of each fraction ordinate, log scale! resulted inthe displacement of the labeled probe asindicated, Thus, in thepresence of increasing amounts of pureunlabeled BarracudaCiguatoxins in the Caribbean

100

75

50 6

0 1 10 100

Relative Sample Voluzne Figure2, RIA Displacernent Assay of BarracudaCiguatoxin HPLC Fractions. Maxi'mumbinding % referstothe displacernent ofthe labeled brevetoxin probe bythe test fractions, and represents the% ofbrevetoxin probe remaining bound in thepresence of given amounts of thetest materials. Relative Sample Volume indicatesthe amount of testmaterial added log scale>. Fraction symbols same as notedin Figure1. Additional materials tested were; crude barracuda ciguatoxm prior to HPLCfractionation solid circles! and pure 'bTx-3 opencircles!

PbTx-3! thelabeled probe was most effectively displaced, whereas crude ciguatoxinunfractionated on HPLC solid circles! and fraction6 solid tri- angles,point down! were least effective in displacingthe labeled PbTx-3 probe fromits antibody.Among the HPLCfractions, 7 wasthe most effective com- petitorof thePbTx-3 probe, followed closely by fractions1 and4. Fractions3, 5, 2 and6 respectively,showed decreasing abilities to displacethe PbTx-3 probefrom itsantibody. This assay reflects the presence of similar structural componentsin the antigenicsites epitopes!of the test moleculesand the PbTx-3probe, Thus, the moresimilar the structuralcomponents of the epitopes,the more effectively will the testmo!ecule compete with thebinding of the PbTx-3probe to its antibody.The epitopesof fractions7, and to a somewhatlesser degree fractions 1 and 4, contain structures very similar to the epitopeof PbTx-3.Okadaic acid on the otherhand did notdisplace tritiated PbTx-3in thesecompetition studies Figure3!. Basedon the resultsof this assay,okadaic acid does not appear to bea componentof the HPLCfraction- atedbarracuda ciguatoxins.

95 T. R. Tosteson, et al, o 125

~ 0 ~ ~ ~ ~ ~ 100 ~ OX 0 QO75 0 o 50 0 25 0 %0S- 0 o 0 1K-2 1E- 1 1. 10 100 1000 1E4 Competitor Cortcetttrettiort nM!

Figure 3, RIA DisplacementAssay of OkadaicAcid. Maximum binding % refersto the displacementof the labeledbrevetoxin probe, and representsthe % of brevetoxin probe remainmg bound >nthe presenceof given amountsof okadaic acid solid points! and unlabeled pure PbTx-3 open drcles!. Relative SampleVolume indicates the amount of material added log scale!.

SytMptosorrreAssay The resultsof the synaptosorneassay are illustratedin Figure 4. In this assaytest fractionscompete with tritiated PbTx-3for binding siteson rat brain synaptosornes.These assays reflect the similarity in moleculartopography of toxinsfor binding sitesthat arenot necessarilythe sameas those recognized by antibodies.The degree to whichthe testfractions displace the binding of the PbTx-3probe to thesynaptosomes sodium channel receptors! is a measureof their compositepotency relative to PbTx-3, The % maximumbinding abscissa!once again refersto the PbTx-3probe as noted above, and the ordinate indicates the amount of each fraction added to competewith the bindingof the PbTx-3probe with synaptosomesodium channel components.Pure PbTx-3 ! displacedthe probe most effectively, while fraction7 wasthe mosteffective competitor of the probeamong the HPLCfractions of thebarracuda ciguatoxin. The maximum specific displace- ment activity of fraction 7 opendiamond! was closelyfollowed by fractions 4 and 2, with tractions I and 3 showing somewhat less activity respectively, Thus,fraction 7, the mosthydrophobic of the HPLC fractionsof the barracudaciguatoxin had the highest compositepotency comparedto PbTx-3itself. Hydrophobicity has been correlatedwith increasedpotency among the brevetoxins~-

96 BarracudaCiguatoxins in the Caribbean

100

75 685500

10 100

Relative Sample Volume Figure4, SynaptosomeDisplacement Assay of BarracudaCiguatoxin HPLC Fractions. Maximumbinding Va refers to the displacement ofthe labeled brevetoxin probe bythe test fractions, and represents the% ofbrevetoxin probe remaining bound in thepresence of given amounts of thetest materials, Relative Soap/e Volume indicatesthe amount of testmaterial added log scale!. Fraction symbols same as notedin Figure'l. Additionalmaterials tested were; crude barracuda ciguatoxin priorto HPLCfractionation solid circles! and pure PbTx-3 open circles!.

Disc VSSION Theresults of toxicity,radioirnrnuno andsynaptosome binding assays are summarizedin Figure5. In this figurethe Relative Maximal Response ab- scissa!,numerically scored from 6 maximumactivity! to 0 no activity!is shownfor eachHPLC fraction of thepartially purified barracuda ciguatoxin ordinate!.HPLC fractions of thcbarracuda ciguatoxin differed widely in theirtoxicity, depending onwhether they were administered orally or by i.p. inoculation.Oral toxicity was found in fractions1, 2, and 7, while only frac- tions2 and5 showedi,p, toxicity. Only fraction 2 wastoxic when administered byboth routes. Oral toxicity may Indicate materials that are truly "ciguatoxic," Resultsof thebinding cotnpetition studies showed that fractions 1,4 and 7 hadsignificant acti vity scored6 and 5! in both the radioimmuno and synapto- somebinding assays. Fraction 2 is particularly interesting because of its high oraland i,p. toxicityand the factthat it showedsignificant activity in the synaptosomeassay, coupled with low RIA activity, This highly toxic material did notcontain the antibody binding portion epitope!of thePbTx-3 antigen

97 T. R. Tosteson, et al.

~ > o C4 a> 4 at61 Cl

0 0 2 3 4 5 6 7 6 C]RIA, K9SYNAPTOSOME,&LP. B1OASSAY.UUORAL BIOASSAY

Rgure5. Assay Summary; Harracuda Ciguatoxin HPLC Fractions. RelativeMaxitnal Response of the barracudaciguatoxin Hf'LC fractionsin the RIA clearbar!, Synaptosome Assay hatched bar!, I.P. Toxicity Assay solid bar! andOral Toxicity Assay crossedbar!. See text for explanation, whichwas used to elicitantibody production. Thus, the toxicmaterial in this fractionof thebarracuda cigua toxin may either be a non-polyethertype toxin ora polyethertoxin that is not epitopic with the brevetoxin employed in these studies. The resolutionof this questionwill require further study. The RlA and synaptosomeassays showed roughly the samedegree of activityin eachof fractions3,4 and6. Thesefractions, which were not toxicto the testanimals, showed moderate, significant and negligible activities respec- tively in bothbinding assays. Fraction 4 is interestingin that it contained materialthat was epitopic with PbTx-3,with a similarcomposite potency in the synaptosomeassay, and yet wasnot toxicto the testanimals. Fractions 4 and 7 have well correlated,significant activities in both binding assays,yet differ markedlyin their toxicities. Therelationship between this presumed polyether,PbTx-3 like materialand fraction 7 remainsto bedetermined. All assaysconsidered together suggest that peakoral toxicity wasassoci- atedwith fractions1, 2 and7, Fractions2 and7 showedsignificant activity score5 or6! in thesynaptosome assays, closely followed by fraction1 witha moderatelevel of activityin this assay.Significant activities were seen in the RIA assayin fractions1 and7. Thus,the overall pattern of relative activity w» nearlyidentical in thesefractions despite marked differences in hydrophobic- ity elution times!. Fraction 2 was distinct from fractions 1 and 7 in that it BarracudaCiguatoxins ln theCaribbean showedhigh i,p. toxicity, and negligible immunoassay activity, as noted above.Fractions 3 and 4 displayednegligible toxicities, however, they showedmoderate andsignificant activity inboth the RIA and synaptosome assays.Fraction 4 may be a polyether,VbTx-3 like matenal that is not toxic to mice.Peak i.p. toxicities were found in fractions2 and 5, however,these fractionsdisplayed widely different activities inoral toxicity and synaptosome binding.Fraction 5 was distinct from all other fractions inthat it showedhigh i.p. toxicity,only moderate tonegligible activities in thebinding assays, and nooral toxicity. Fraction 6 showed little activity in anyof thefour assays, however,this materialreproducibly made animals sick, Onthe basis of presentresults, there appear to beat least five chromato- graphicallydifferent components in the crude ciguatoxic barracuda extracts, fractions1, 2, 3-4, 5 and7, Fourof thesecomponents, fractions I, 2,5, and 7 weretoxic, Whetherthese represent different toxins or metabolitesof the sametoxin remainsto be resolved.The crude barracuda ciguatoxin had multipletoxins, the more hydrophobic ofwhich fractions 5 and7! appeared to be distinguishedon theone hand fraction5! by componentsof high i.p. toxicitywith only negligiblesimilarity to brevetoxinPbTx-3, and on the other fraction 7!, components of high oral toxicity, similar in activityto PbTx- 3. Furtherwork is in progressto resolvethe nature and relationship between thesecomponents. Acknowledgements This work wassupported in part by researchgrants from the SeaGrant Programof theUniversity of PuertoRico, the NIH-MBRS Program, University of PuertoRico at Mayagiiez TRT!, and the U,S,Army Med.Res, and Dev. Command,contract ¹DAMD 17-88-C-8148at the University of Miami DGB!.Opinions, interpretations,conclusions and recommendationsare thoseof theauthors and are not necessarilyendorsed by theU.S. Army.

LITERATURECITED Craig, C.P. !980. Editorial: It's always the big ones that should get away. JAMA 244: 272-273. Lawrence,D.N., M.B. Enriquez,R.M. Lumish and A Maceo. 1980,Fdi- torial: Ciguaterafish poisoning.JAMA 244:254-258. 3. Olsen,D.A., D.W. Nellis and R.S.Wood. 1984.Ciguatera in the eastern Caribbean. Mar. Fish.Rev. 46: 13-18, Ragelis,E. I', 1984. Ciguateraseafood poisonirtg: Overview. In: Seafood Toxins, E.P.Ragelis,ed,!. ACS Symposium Series ¹262, American Chemical Society,Washington, D.C, pp, 25-36 Anderson,Donald M, and Phillip S.Lobel. 1987.The continuing enigma of ciguatera.Biol. Bull., 172:89-107. T, R. Tosteson, et nl.

McMillan,],P., H,R. Granade and P. Hoffman. 1980, Ciguatera fish poi- soningin theUnited States: Preliminary studies, J, ColLVirgin Is. 6- S4-107. Tacket,C, 1981.Studies of epidemiologicaland clinicalaspects of ciguatera.Paper read at FirstConference on Ciguatera, Caribbean FisheriesManagement Council, San Juan, Puerto Rico. Escalonade Motta, C.,J.F. Feliu and A. Izquierdo. 1986. Identification andepidemiological analysis of ciguateracases in PuertoRico. Nat. Mar Fish, Rev. 48: 14-18. Todd,E.C.D. 1992. How ciguatera affects Canadians. In: Proceedings of the Third InternationalConference on Ciguatera Fish Poisoning T,R. Tostesoned.!. Polyscience Publishers, Quebec, Canada. See pp, 199-214, 10. de Sylva,Donald P. 1963,Systematics and Life History of theGreat Barracuda.University of MiamiPress, Coral Gables, Florida. 11, Tosteson,T,RD.L Ballantineand H.D. Durst. 1988.Ciguatoxic barra- cudafrequency in southwestPuerto Rico. Toxicon 46:795-S01, 12, Randall,J,E. 1958. A Reviewof ciguateratropical fish poisoning, with a tentativeexplanation of its cause,BulL Mar, Sci. Gulf Carib. 8: 236-267. 13, Helfrich, P. and A.H. Banner. 1963,Experimental induction of ciguatera toxicityin fish throughdiet. Nature 197:1025-1026, 14. Yasumoto,T., I. Nakajima,R. Bagnisand R. Adachi. 1977.Finding of a dinoflagellateas a likelyculprit of ciguatera.Bull. Jpn. Soc. Sci. Fish., 43: 1021-1026. 15. Yasumoto,T., A. Inoue,R. Bagnisand M. Garson.1979. Ecological sur- vey on a dinoflagellatepossibly responsiblefor the induction of ciguatera,Bull, Jap,Soc. Sci. Fish. 45:395-399. 16, Adachi, R. and Y. Fukuyo. 1979.The thecalstructure of a marinetoxic dinoflagellate Gnmbferdiscustoxicus gen, et sp, nov. collected in a ciguateraendemic area. Bull. Jpn.Soc. Sci. Fish. 45:67-71. 'I 7. Murata, M., A.M. I.egrand, Y. Ishibashi and T. Yasumoto. 1989. Structures of ciguatoxin and its congener. J, Am. Chem, Soc, 111:S929-8931. 18. Yasumoto, T., M. Murata, Y. Ishibashi, M. Fukui and A. M. Legrand.1992. Structure determination of ciguatoxinof morayeels and Gnmbierdiscustoxicus . In: Proceedingsof the Third International Con- ferenceon Ciguatera Fish Poisoning T.R. Tostesoned.!. Polyscience Publishers,Quebec, Canada. See pp, 3-12. Ballantine,D.L., A.T, Bardales, T.R. Tostesonand H.D, Durst. 1985, Sea- sonalabundance of Gnmbierdiscustoxicus and Ostreopsissp. in coastal waters of southwest PuertoRico, Proc. Fifth Inter. Coral Reef Cong. 4:417-422. 20. Ballantine,D.L, T.R.Tosteson, and A.T. Bardales. 198S. Population dy- BarracudaCiguatoxins in theCaribbean namicsand toxicity ofnatural populations ofbenthic dinoflagellates in southwestPuerto Rico. J. Exp, Mar. Biol. Ecol. 119:201-212. Tosteson,T.R., D,L. Ballantine, C,G. Tosteson, A.T. Bardales H.D. Durst andT.B. Higerd. 1986, Comparative toxicity of Caa>b>'rrdiscastuxicas, Qstreopsiscf. lertticu arisand associated microbial flora. Mar. Fish. Rcv. 48:57-59. Tosteson,T.R., D.L. Ballantine, C,G. Tosteson, V. Henslcy and A. Bardales,I'989. Associated bacterial flora, growth and toxicity of cul- turedbenthic dinoflagellates Osfreiipsis lenticularis and Camb erdiscus foxicus. Appl. Environ, Microbiol, 55:137-141. 23. Buck,J.D. and R.H, Pierce. 1989. Bacteriological aspects of Floridared tides;A revisitand newerobservations. Estuarine, Coastal and Shelf Science 29: 317-326. 24, Gonzalez,I.,C.G, Tosteson, V. Hensley and T.R, Tosteson. 1991. Role of AssociatedBacteria in Growth and Toxicity of BenthicDinoflagellates. ASMAnnual Meeting. Dallas, Texas, May 5-9, Abstr.¹Q 123. 25. Kodama,Arthur M., Y.Hokama, T, Yasumoto,M. Fukui,S.J. Manea and N. Sutherland.1989. Clinical and laboratoryfindings implicating palytoxinas cause of ciguaterapoisoning due to Decapterusmacrosiaria Mackerel!. Toxicon 27:1051-1053. 26. Terao,K. and T. Yasumoto.1992, Pathomorphological studies on ex- perimentalmaitotoxicosis and ciguatoxicosis in mice. In: Proceedings of theThird International Conference onCiguatera Fish Poisoning T.R. Tostesoncd.!. PolysciencePublishers, Quebec, Canada. See pp. 61-78. 27. Ohizumi,Y. 1992.Mechanisms of pharmacological andtoxicological ac- tion of ciguatoxin and maitotoxin. In: Proceedingsof the Third InternationalConference on CiguateraFish Poisoning T.R. Tosteson ed.!. PolysciencePublishers, Quebec, Canada. See pp. 79-88. 28. Vernoux, J.P. and A. El. Andaloussi. 1986.Heterogeneite des ciguatoxines extraites de poissons peches aux Antilles francaises. Biochimie 68: 287-291. 29. Vernoux,J.P. and F, Talha, 1989,Fractionation and purification of some muscular and visceralciguatoxins extracted from Caribbeanfish, Comp. Biochem.Physiol. 94B:499-504- 30. Bigazzi, PK, Wicker, E- Goryski, R. Zcschke,J. Puleo,G. Andres and S. Gutcho. 1973. Procedures using labeled antibodies or antigens. In: Methods in Immunodiagnosis, N-R. Roseand P,E,Bigazzi eds.!. John Wiley and Sons,New York. p, 107. 31. Abraham, G,E, and P.K.Grover. 1971,Covalent linkage of steroid hor- monesto proteincarriers for use in radioimmunoassay.In: Competitive Protein BindingAssays, W, Odell and W. Daughaday,eds.!. J. P. Lippincott, Philadelphia.p.140.

101 T. R. Tosteson, et aL

32. Dodd,P.R. J,A, Hardy, A.E, Oakley, J.A. Richardson, E.K. Perry and J.P. Delaunoy.1981. A rapidmethod for preparingsynaptosomes: Com- parisonwith alternativeprocedures. Brain Res. 226:107. 33, Bradford,M.M. 1976.A rapidand sensitive method for thequantitation of microgramquantities of protein.Analyt. Biochem. 72:248. 34, Trainer, V.LR.A. Edwards,AM, Szmant,A.M. Stuart, T.J. Mende and D.G. Baden,'l990. Brevetoxins; unique activators of voltage-sensitive sodiumchannels, In: MarineToxin: Origin, Structureand Molecular Pharmacology, S, Hall andG. Strichartz,eds.!. ACSSymposium Series 418,Washington D.C. pp 165-177. 35. Baden,D.G., T.J. Mende, J. Walling and D.R. Schultz,1984. Specific antibodiesdirected against toxins of Ptychodiscusbrevis Florida's redtide dinoflagellate!.Toxicon 22:783. 36. Baden, D.G., Kathleen S. Rein, !Vlasao Kinoshita and Robert E, Gawley.1991. Computational modeling of the polyetherladder toxins brevetoxinand ciguatoxin.In: Proceedingsof the Thirdinternational Conferenceon Ciguatera Fish Poisoning T.R. Tosteson ed.!. Polyscience Pubiishers,Quebec, Canada, Seepp. 117-130,

102 Computational Modeling of the PolyetherLadder Toxins Brevetoxinand Ciguatoxin

Daniel G. Baden'Kathleen S. Rein2,Masao Kinoshita'~ and RobertE. Gawley~ UniversityofMiami 1Rosenstiel School otMarine and Atmospheric Science,Department of Chemistry,and Schoolof MedicineDepartment of Biochemistryand Molecular Biology.

ABSTRACT MonteCarlo and standard computational methods were used to generate conformationalmodels of thebrevetoxins and ciguatoxin. These were used to constructa modelof polyetherladder toxin interactionwith thevoltage- sensitivcsodium channel which explainson a molecularlevel how brevetoxins andciguatoxin might maintain voltage-sensitive channels in anopen configu- ration.

INTRODUCTION Sodiumchannel isolated from rat brainconsists of threeseparate and distinctprotein subunits: the a-subunit,and the 61 and 82-subunits,which togethercomprise the channel in a 1:1:1stoichiometry'. The ct-subunit is a glycoproteinof approximately260 kilodaltons kD!, andis a transmembrane proteinwhich binds neurotoxins at specificloci or topographicsites. The two 8-subunitsare smaller molecular weight peptides each of about30 kD! and areintegral membrane subunit glycoproteins. Schematically, thechannel has beenillustrated as shown in Figure1.

Figurel. SubunitStructure of the VoltageSensitive Sodium Channel, Schematicsubunit structure of the voltage-sensitivesodium channel showing disulfidebonds S-S!, phosphorylation sites P!, and glvcosidation sites pitch- forks!' D, G, Baden, et aL

Using primary structuredata, Catterall developeda modelfor how the a- subunit inserts itself into the membrane of excitable cell types2. Each a-subunitconsists of four homologousdomains; each homologous domain being composedof six transmembranepeptide sequencesS1-S6, with the highly positively chargedS4 regions shadedin the figure!being most highly conserved~.These S4 domains have been postulated to completely transverse the membrane,and aHfour S4regions in concertcontribute tn the ion shuttling ability of the a-subunit Figure 2 .

Figure 2, FunctionalMap of the Sodium Channelet-Subunit. Catterall's functional map of the a-subunit of the voltage-sensitivesodium channel TheS4 regions shown shaded!are thought to be responsiblefor ion transport character of the transrnembrane-spanningprotein.

The polyether laddertoxins brevetoxinand ciguatoxin representa novel classof marinetoxins first characterizedin 1981 Figure3!s 7. Ciguatoxin is the mostrecent toxin in the group~,These three classesof compounds have in common a number of structural features,including: trans/syn stereochemis- try along theentire backbone, with oxygenatoms alternatmg between the top and bottom;each ether oxygen constitutes a oneatom bridge betweenadjacent rings; eachtoxin has regionsof rigid, semirigid, and flexible character;the two brevetoxin backbones have lactone functions on one end, with an enone function on the other; the structureof ciguatoxin has not beenfully elucidated, but the vicinalaHyBc diol maybe oxidizedirt situ by an induciblecytochrome P450oxidase or constitutivealcohol dehydrogenase to an enone videi nfre!, and the spirocyclic acetal on the other end may be opened to a dihydroxyketoneor a hemiacetal.Thus, there appears to be electrophilicfunc- tionalgroups on eitherend of eachpolyether ladder. These common structural features are consistent with the observation that all three toxin types bind to the sameunique site of the voltage-sensitive ComputationalModeling of PolyetherToxins

sodiumchannel", Additionally, thefact that the reactive functional grpups are ateither end, and the finding that all three have or in the case of ciguatpxin mightbe enzymaticaIIy activated to have!an enonefunction as thempst reactiveFunctionality, ishighly suggestive ofa common mode ofbinding and activity.Pharmacologically, the least toxic of the threetypes is pbT2 brevetoxin8!, followed by PbTx-1 brevetoxin A!,and the most pptent ciguatoxin.The present study seeks to examine the computer-generated populationofstructural conformers which fall between the energy constraints onewould expect under normal biological conditions in living systems. Based onthe apparent correlation between flexibility and potency, wesought tp voltagedevelopsensitivea mecharusticsodiummodelchannel for polyetherVSSC!, Cieesralsamanrladder toxin interaction with the

Rhgsm s-9~ lhahr OrRis~- ! 013

Figure3, The Polyether Ladder Toxins PbTx-2 Brevetoxin-B!, PbTx-1

MATERIALSAND METHODS Themost efficient method for randomly sampling conformational space makesuse of internal coordinate Monte-Carlo subroutine techniques approachhas recently been compared to othercomputational algorithms tp locatethe 256 low energy minima of cycloheptaJccane 6 rotatablebond»

10S D. G. Baden, et al. andwas found to befar superiorto itscompetitors> o. The Monte Carlo method isincorporated into themolecular modeling software Macromodel, developed byProf. Clark Still at Columbia University. Because the Monte Carlo method hasnot beenexplicitly tested on asymmetricstructures like theseladder toxins,we have approachedthe problemof multiple ring flexibility with caution, ForPbTx-2 type toxins, there are only two areaswhere the molecule might flex:the two 7-memberedrings D and E, and the 8-memberedring H, All otherrings in themolecule are held rigid by virtue of theirtrans-fused decalin- likegeometries, Thus, these hinge points were modeled independently of one anotherusing Monte Carlo methodson the smaller models.The starting structureswere randomly varied usingthe standardvariables of Batchntin version2.7 theMacromodel batch-mode program!. All bondsnot specifically lockedby theirinclusion in a rigidring werevaried, Thus,for example,the ringfusion between rings C andD ofPbTx-2 is rigidwhereas the ring fusion betweenrings D and E isnot rigid. Synaptosomebinding protocols and inhibition constantcalculations, and HPLC analysesfor toxin hydrophobicitywere carried out as described previously".

RESULTS Muchof thepharmacological data available on thebrevetoxins surrounds PbTx-3,the reducedaldehyde-reduced primary alcoholderivative ot PbTx- 2. Some correlations in the literature show that within a structural class of brevetoxin,i.e. PbTx-2type or PbTx-1type, minor substituentsplay an insignificant role in affecting toxicity or inhibition constant binding character".However, there was shown to be a statistically significant differ- encebetween the two brevetoxinclasses, By using radioactive derivatives of Table1. Correlation of Potencywith BindingConstants and Iiydrophobicity.

Synaptosome HPLC Relative Toxin Mouse LD~ Binding Hydrophobicity Ko ma~ i

PbTx-2Type 200 2.6 6.80 2-3 least

PbTx-1Type 100 0,8-1.0 6.30 0.3-0.7 middle

Ciguatoxin 0.45 .10! .80! 0.14 most ComputationalModeling of PolyetherToxins

eitherPbTx-1 orPbTx-2, specific binding parameters suchas binding affinity expressedasKD, the dissociation constant! and number ofsites B~, binding maximumexpressed as pmol/mgsynaptosomal protein!, it is possibleto comparedirectly the relative binding afBnities and maxima. Moreover, de rived valuesfor ciguatoxindissociation constants and binding maxima can be calculated,there being no radioactively labeled material for directbinding studies, Purifiednatural toxins within eitherthe PbTx-2 or PbTx-1toxin class used as competitive inhibitors of radioactive toxin of the sameclass failed to show any statisticalcorrelation, i,e, all classPbTx-2 toxins exhibited inhibition con- stantsequivalent to one another.PbTx-I toxins wereuniformly and statisticallymore efficient at displacingPbTx-2 type toxinfrotn site5 of the VSSCthan were PbTx-2 type toxins. Inhibitionconstant studies indicate to us thatPbTx-I brevetoxins bind with a higheraffinity to thespecific binding site thando PbTx-2type toxins.Ciguatoxin, by our studiesand by literature accounts,competitively inhibits brevetoxin binding at muchlower concentra- tionsthan either of thebrevetoxin classes. Inhibition constants for ciguatoxin displacingPbTx-2 type brevetoxin from its specificsite are in the0.!-0.2 nM concentrationrange. Thus,the dissociation constants calculated for each of thethree toxin types progressfrom the least potent PbTx-2class at 2.6 nM, to the PbTx-1class at 0.8- 1.0nM, and culnunate with ciguatoxin,displaying a calculateddissociation constantof 0.1nM. Mouseintraperitonea! bioassay progresses in exactly the same manner, although the magnitude of the derived values exceedsthe differencesin specific binding data. As expected, the inhibition constant data parallelsthe dissociationconstant and mousebioassay data. High perfor- manceliquid chromatographyretention on C-18reverse phase columns indicatesthat hydrophobicity increases from PbTx-2 type to PbTx-1type to ciguatoxin,These data correlations seem to indicatethat toxicity is relatedto bindingaffinity, which is in turn relatedto relativehydrophobicity. That bindingmaximum does not vary from toxin type to toxintype is reassuring, for it indicatesto us thatspecific site 5 behaveslike a homogeneousbinding domainwhen exposed to thesetoxins, and iurther, that we cannot distinguish sub-populationsof high affinity binding sites Table 1!. It isthe pharmacologi- cal and bioassay data that we seek to correlate with the structural characteristicsof each toxin type as determined by computer-aided molecular tnechanics,As descrtbed below, we believewe havedeveloped a working modelwhich explains the variableonset of symptomsand curious a!coho!- relatedrecurrence of ciguaterasymptoms in humans,and which also lends credibilityto theobservations ofreduced reversibility and rapid onset times in toxinderivatives which possess conjugated aldehyde systems versus u, g unsaturatedprimary alcohol systems.

107 D. G. Baden, et a!.

Themolecular topography involved in bindingand expression of activity by thepolyether ladder toxins can be modeledas activity A, triangles!and binding B, squares! loci on thetoxins, as shown in Figure4. Theinherent flexibilityof eachtoxin permits either more c! or less b, thena! effective simultaneousinteraction of the A and B domains with specific sites on the VSSC.A cursoryexamination of the structures in Figure3 revealsthat PbTx-2 has 16rotatable bonds, PbTx-1 has 31, and ciguatoxin hasover 40.

PbTx-2 Type

0! PbTx-I Type

c! Ciguatoxin

Domain

Figure4. SchematicModel of PolyetherLatter Toxin Interaction with VSSC. Toxin flexibility increasesfrom a! to b! to c!, and correlateswith increasing potency. It is proposedthat increasedflexibility allows for more favorable inter- action of binding B! and activity A! centers located on distal ends of each polyetherladder toxin.

Whileour long-termgoal is to evaluatethe conformational space of all threepolyether ladder toxin types, we havebegun our effortswith themost rigidof thethree types. Examination of the structuressuggested that PbTx-2 would bemost rigid. Indeed,Lin etal. originallydescribed the structureas a "rigid ladder-likestructure" . The 8-memberedring of PbTx-2diol- dibenzoatewas calculated to preferthe crown conformation by several kcal/ mols. internalcoordinate Monte Carlo modeling of the G-H-Irings, in whichall torsionsof the 8-membered ring are varied except for the double bond in the tworing fusion bonds, reveals only two conformers kcal/mol above global minimum,These conformations differ by 4.06 kcal/mol which corresponds to greaterthan 99.9% of thelower energy conformer, By contrast, modeling « theC-D-E-F rings shows a numberofconformers and suggests that the toxin is not rigid at all. Stereoviews of theseven lowest energy isomers and their

108 ComputationalModeling ofPolyether Toxins

7.E,.8.288e !18s;

6. E i 869kcal21'i.i

5. E~i 3.88kcal .084'%%dl

e. E,.1.60 k: .80 i !

3 E~ ~ 073 ~~ ir.2'%%dr

2 E, ~ 0.63 v:~! 0 0'%%d!

1. Glebe'Minimum 8.9'%%d!

FigUre5 ConformationalAnalysis of the 6-7-7WRing Region Rings C-F! of f~revetoxinH. Stereoviewsof the 7 lowest energy conformersof the model compounds,with relativeenergies and the Bo!tzmanndistribution of eachat 298'K. D. G. Baden, et al.

Boltzmannpopulations are shown in Figure5. Notethat the four lowest energystructures, which differ by only1.6 kcal/mol comprise greater than 99%of the totalpopulation. Note also that structures2 and4 possessan approximate90' bendat theD- ring. With thesepreliminary results in hand,we usedthe internal coordinate MonteCarlo technique to probethe conformations of PbTx-2, Although the completedetails of thesecalculations will bereported elsewhere, the method did locatea numberof conformations,including all four of the low energyC- D-E-Fconformations corresponding to structures1-4 of Figure 5. interestingly,the relative positions of thethree lowest energy conformations werescrambled. For example,the lowestenergy conformation of the toxin containsthe D-E configurationcorresponding to conformation3 in Figure 5, Conformation1 of figure 5 appearsnext, at 0 26kcal/mol above the global minimum. Thencomes conformer 2, at 0.57kca1/mol aboveglobal minimum and conformer4 at 1,82kcal/mol above the global mini~urn. Figure 6 illus- tratesthe superimpositionof the global minimum and the 0.57 kcal/mol

Figure6, SuperimposedConformers of PbTx-2,The two structurescorrespond to the global instant minimum extended! and the +0.57 kcal/mmol bent! struc- ture. The K ring is locatedon the right side. structures,and illustrates the dramatic effect of D-ring flexibility on the posi- tion of the A-ring relative to the stationaryK-ring. The A-ring sweepsover 12 angstromsby this singlering flip. These4 structurescomprise greater than 98% of the total population and we would expect that within this composite population lies a conformationor conformationsthat are "active" with respect to VSSC interaction. For ciguatoxin,the computationalchemistry was more difficult, there beingmore than two hingeregions. Manual manipulation of the9-membered F ring hasidentified several structures witht'n 5 kcal/mol of instant global minimum, By performinga similarsuperimposition of two of the generated structures, AE = 0.8kcal/mol! we see that the M ringin ciguatoxinmay sweep a distancegreater than 35 angstroms Figure 7!.

110 C ornputattonalMode!tng of Polyether Toxtns

Figure7. Superimposed Conformers of ciguatoxin.By f!exing only the F;F-G region,and differing in energyby 0.8 kcat/mote, a 35angstrom spread between thelocation of theM ring.The mostly rigid A-E ring is located on the right of the figure.

As neurotoxins,the polyetherladder toxins interact with the VSSCbased primarilyon theirbinding to a singlespecific site called site 5, located on thea- subunitof thechanneL Unpublished photoaffinity probe results indicate that brevetoxinbinds in a 1:1stoichiometry to DomainIV of the channel.ln all likelihood,site 5 is locatedon the extracellularloop between polypeptide segmentsS5 and S6 Ph.D. dissertation work of V.L. Trainer, University of Miami!. We propose that brevetoxins interact with the u-subunit of the voltage- sensitive sodium channel as illustrated in Figure g. We suggestthat free movementof the sliding helix S4region! is hamperedby the polyetherladder toxins so that the channel remainsin a conductingstate as follows: I! the e- ~minogroup of a lysineon or nearthe S4 helix adds to thetoxin enal ineither Schiffbase or michaelfashion! to form a reversibleS4-toxin conjugate; ! the toxin flexesto wrap aroundor againstthe extracellular loop between S5 and S6,and is held in placeby a cotnbinationof hydrophobic interactions, hydro- genbonding, or reactionwith a secondelectrophilic site on the other end of the toxin such asthe lactone of the brevetoxins,or the maskedcarbonyl at the l.-M ring fusionof ciguatoxin!. D. G. Baden,et af,

Figure8. Model of Toxin Interaction With VSSC. Thetransmembrane S4helix is proposedto beheld in placedue to ion pairingby a spiralpath of "+" chargedarginine residues wrapped around the helix. Ion pairingis perturbedby membranedepolarization, which results in a screw-like motionthrough a 60'rotation and a 5angstrom outward movement>. Toxin couldbe attacked by anexposed amino group forming a Schiff'sbase or Michael addition, therebyperturbing normal kinetics.

Thesetwo steps correspond to activityand binding,illustrated schernati- cally in Figure4, although availabledata do not distinguish which is which. We believeboth stepsare required for the full expressionof polyether ladder toxin action. This model is consistent with several intriguing facts; squid or crayfishaxons exposed to toxin exhibit depolarization which can be reversedby washing, but the reversibility is dependenton the length of exposure;upon bindingto voltageclamped neurons,closed channelswere openedand channelkinetics were modulatedto exbibit slow inactivation kinetics";toxins administered to VSSCin rat brain synaptosornescan be reversiblybound if washoutoccurs within 20 min to 1 hr of application;and all threetoxin types exhibit competitivedisplacement equilibrium binding con- stants in radioactivebrevetoxin Rosenthal Analysess,interestingly, the reversibility of toxin actionis lessenedfor the more flexibletoxins. Our hypothesisis thatthe increased flexibility is reflectedin the second "wraparound" step describedabove. Note also that the calculated differ- encesin energyfor the toxinconformers allow significant populations of the bent isomerseven in the absenceof external influences. Moreover, differences of 2-3 kcal/mol could be easilyovercome by hydrophobicinteractions or hydrogenbonding, which might significantly stabilize a higherenergy con- formerrelative to the calculated global minimum. Thect, 8 unsaturatedenal system required for theproposed action is presentin thebrevetoxins and may be produced enzymatically in ciguatoxin seeFigure 9! by the constituitiveenzyme alcohol dehydrogenase, but more likelyby aninducible cytochrorne P450 oxidase. It is interestingto notethat personsafflicted with ciguatera often relate recurrent ciguatera symptoms with theconsumption of alcohol.It hasalways been considered merely

1]2 ComputationalModeling ofpolyether Toxins

CgTx CgT OH OH CHO Figure9. Activation of Ciguatoxinby Cytochrome P450. proposedmechanism bywhich ciguatoxin could be activated viaa cytochrome p450oxidase to producean a, 8 unsaturatedenalsystem like the brevetoxins. Suchan enal may be required forthe activity expressed byany of the polyether ladder toxins. anecdotal,but perhaps there is a biochemicalbasis for ciguatntoxication, since alcoholdetoxification induces cytochrorne P450 oxidase activity. ln summary,this working model describes the computational, pharmaco- logicaland binding results thus far generated, andconsolidates ourthoughts on the specificinteraction between these toxins and the VSSC,Additional workin progress includes chemical stabilization ofSchiff's base by reduction following incubation with radioactive toxin, and isolation and characteriza- tion of theadduct. Synthetic polyether toxins are also being developed, utilizingthe information gathered concerning hinge regions and activity and bindingcenters. We anticipate fully developingquantitative structure-activ- ity relationships QSARl for polyetherladder toxins which will aid in the completedescription of voltage-sensitivesodium channel topography and toxinology. Acknowledgements Thiswork wassupported in partby theUS Army MedicalResearch and DevelopmentCommand, DAMDl7-88-C-8148, Opinions, interpretations, conclusions,and recommendationsare thoseof the authors and are not neces- sarilyendorsed by the USArmy.

Li TURATUREClTED Catteral,WA. 1989.Molecular properties of voltage-sensitivesodium channels.lCSU Short Reports. 9:14. Catterall,WA. 1986 lvfolecularproperties of voltage-sensitivesodium channels, Ann. Rev. Biochem. 55:953, Noda,M., T. Ikeda,T. Kayano,H. Suzuki,H. Takashirna,M, Kurasaki,H. Takahashi and S. Numa. 1986. Existence of distinct sodium channel messengerRNA's in rat brain. Nature.320:188, Tanabe,T., H. Takashima, A. Mikami, V. Elodkerzi, H. Takahashi,K.

113 D. G. Baden, ef al.

Kangawa,M. Kojina, J. Matsuo,T. Hirose and S. Numa. 1987, Primary structure of the receptor for calcium channel blockers from skeletal muscle. Nature, 328:313. 5. Lin, Y., M. Risk,S. Ray,D. VanEngen,J, Clardy,J. Golik, J, Jamesand K, Nakanishi. 1981. isolation and structure of brevetoxin B from the "red tide" dinollageBatePtychodiscus brevis Gy>»»oui»iam breve! J. Am. Chem. Soc. 103:6773, 6. Shirnizu, Y., H, Chou, H. Bando, G. Van Ouyne a nd Clardy, 1986. Structureofbrevetoxin-A GB-1!,the mostpotent toxinin the Florida red tide organismGym»odi»ium breve Ptychodiscusbrevis! J, Arn, Chem. Soc. 108: 5924. 7, Murata, MA-M. Legrand, Y, lshibashi and T. Yasumoto. 1989.Structure of ciguatoxinand its cogeners.J. Arn. Chem. Soc. 111:8929. 8. Trainer, V.L., R.A. Edwards, AM. Szmant, AM. Stuart, T.J. Mende and D.G. Baden. 1990.Brevetoxins: Unique activators of voltage-sensitive sodium channels,ACS Symposium Series 1990 418: 166. 9. Chang,G., W. Guida and W,C,Still, 1989, An internal coordinateMonte Carlo Method for searchingconformational space.J. Am. Chem. Soc, 1]1:4379. 10. Saunders,MK, Houk, Y. Wu, W.C.Still, W, M. Lipton, G, Chang and W. Guida. '1990.Conformations of cycloheptadecane:A comparisonof methodsfor conformationalsearching J. Arn. Chem.Soc. 112:1419. 11. Huang,J.M.C., C.H. Wu and D.G. Baden. 1984. Depolarizing action of a redtide brevetoxinin axonalmembranes. J. Pharrn,Exp. Ther. 229:615.

114 Antibody Production and Developmentof a Radio irnmunoaSSayfor the PbTx-2-TypeBrevetOXirIS.

M. A. Poli and J. F. Hewetson PathophysiologyDivision, U,S,Army Med ical ResearchInstitute of InfectiousDiseases, Ft. Detrick, Frederick, Maryland 21701-501 I,

ABSTRACT A goatpolyclona1 antiserum specific for brevetoxins containing the PbTx- 2 backbonestructure was raisedagainst PbTx-3 conjugated to bovineserum albumin.This antiserum was used to developa sensitive,specific radio- immunoassay.The minimum detection limits for PbTx-2,PbTx-3, and PbTx-9 wereapproximately 300pg. Cross-reactivitywith PbTx-1occurred at 100-fold higherconcentrations. No cross-reactivitywas detected against saxitoxin, palytoxin,or thediarrhoeic shellfish toxins okadaic acid, pectenotoxin-6, or yessotoxin, Thesame detection limits were achieved when measuring brevetoxins in human urine and in extracts of clams Mercenariamercenaria!, The successful developmentof a radioirnmunoassayfor brevetoxindetection in shellfish extractsis an important advancein the monitoringof seafoodstocks, The abilityto detecttoxins in humanurine suggests potential for the diagnosis of neurotoxicshellfish poisoning and for monitoringor confirming human expo- sure.

INTRODUCTION Thebrevetoxins are a groupof polyetherneurotoxins produced by the dinoflagellatePtychodiscus brevis . Thesetoxins are extremely potent to both mammalsand fish'. During bloomsof P. brevis,known as Florida red tides, brevetoxinscan be concentrated to levels dangerous to manby filter-feeding molluscs2.Human intoxications from neurotoxic shellfish poison of contami- natedshellfish and massive hsh kills can result from the presence of P. brevis in the nearshore marine environments 4. Nine structurally distinct brevetoxins havebeen described to dates. Each is a derivativeof oneof thetwo polyetherbackbone structures illustrated in FigureI. In thiswork, we usethe PbTx nomenclature system and the terms PbTx-I-typeand PbTx-2-typeto differentiatebetween the backbonestruc- tures, M, A. Poli, et al.

OH

0

Rz

Figure 1. Structures of the Brevetoxins

A! PbTx-!-typebrevetoxins: PbTx- 1: R=CHzC CHz ! CHO PbTx-7 R=CHzC CHz! CHzOH PBTx-10: R=CHzCH CHs! CHzOH

B! PbTx-2-typebrevetoxins: PbTx-2: RI=H, Rz=CHzC WHz! CHO PbTx-3: R>=H, Rz=CHzC =CHz! CHzOH PbTx-5: Rq=Ac Rz=CHzC =CHz! CHO PbTx-6: Rz=H Rz=CHzC =CHz! Cl O 7, 28 epoxide! PbTx-8: Ri=H Rz=CHzCOCHzC1 PbTx-9: Rz=H Rz=CHzCH CHs! CHzOH

Forthe most part, detection of thebrevetoxins has been by meansof itr vive bioassaysbased upon the mouse bioassay ofMcFarren or thefish bioassay of Badenet al.7>. While useful in determiningthe general toxicity of a sample theseassays possess neither the specificity nor thesensitivity desired in a researchsetting. Further, today's climate of reducedacceptance by both the

116 Developmentofa Brevetoxin lmmunoassay

generalpublic and the scientific community ofthe use of animals inresearch, coupledwith increasingly stringent regulation ofsuch research, hasmade the developmentofsensitive, specific in vitro assays even more desirable. ln aninitial attempt to producea radioimmunoassay, Badenetal. dc~ scribeda polyclonalantiserum against the brevetoxins". This serum bound PbTx-2andPbTx-3 with equal affinity, and showed about 50% cross-reactivity withbrevetoxins containing the PbTx-1 backbone structure'", Competition curvesdemonstrated a significant decrease inantibody binding to [sHIPbTx-3 in the rangeof 2-5 ng unlabeledPbTx-3 in phosphate-bufferedsa- line.However, no datawere presented regarding detection limits in the presence ofbiological fluids or tissue extracts, andcross-reactivity withother marinetoxins was not addressed. This group is now developing anenzyme- linkedimmunosorbent assay ELISA! using this antiserum, withencouraging results Baden,personal communication!u, Wereport here the production of specific antibodies and development ofa radioimmunoassayfor the brevetoxins. The assay is specificfor thosetoxins containing the PbTx-2-type polyether backbonestructure and little cross- reactivityoccurred with PbTx-1.No cross-reactivitywas observed with other marine toxins known to accumulatein shellfish.We further demonstratethat the assayis equallyeffective in assayingcrude shellfish extracts and human urine. Thesuccessful development of a radioimmunoassayfor the brevetoxins in biologicalsamples is an importantadvance in dinoflagellatetoxinology, We feelit alsoholds great potential for monitoring shellfish stocks and the diagno- sisand epidemiology of neurotoxicshellfish poisoning.

MATERlALSAND METHODS

Tozms Unlabeled brevetoxins were supplied by Dr. D.G, Baden University of Miami, FL!. [3 i]PbTx-3 9-14 Ci/mmol! was preparedand purified by Dr. Badenas previously described". Each was in excessof 99%pure and exhibited a singlepeak of UV absorbanceat 208 nm. Radiolabeled.toxm was kept as a stocksolution of 100pg/ml in ethanolat -20'C.Unlabeled brevetoxins were storedin methanol00 ltg/ml! at -20'C. Potencywas confirmedby the Gambusiabioassays. Okadaicacid, pectenotoxin-6,and yessotoxinwere a generousgift from Dr, TakcshiYasumoto TohokuUniversity, Sendai, Japan!. Saxitoxin was 'upphedby Dr.Sherwood Hall U.S.Food and Drug Administration, Wash- '"gton,D.C.!, and palytoxin was received from the HawaiiBiotechnology institute Aeia, Hl!.

117 M. A. Poli, et al.

Anfigen Preparationand Immunization PbTx-3was conjugated to bovine serum albumin BSA! according to a modifiedversion of the procedureof Badenet a!.". PbTx-3.5 mg plus ap- proximately 0.01pCi [sH]pbTx-3as a tracerfor the reaction! was dissolved in 0.1 ml freshly distilled pyridine containing a 10-fold molar excessof succinic anhydride. The reactionvessel was sealedand the solution stirred in a water bath at 85'C tor 2 hr. The reaction mixture was then dried under a stream of nitrogen and streakedonto a 10x 20 cm silica gel high performance thin-layer chromatographyplate Whatrnan High PerformanceDiamond System, IIP- KF, 200 Irm layer, Whatman Chemical Separations,Inc.!. Development in a solvent system of chloroform/methanol/trifluoroacetic acid 00:10:I! re- solved two major bands. The predominant broad band at Rf=0.27-0.35 contained85-90% of the tracer radioactivity and correspondedto the PbTx-3 hemisuccinatederivative. This fraction was scraped, eluted with methanol, and dried in a reaction vial. The free carboxyl group on the succinylatedPbTx-3 derivative was cova- lently coupled to the e-amino groups on the lysine residues of BSA by a standard carbodiimide condensation'z.Dry succinate derivative was dis- solved in 50% aqueous pyridine at 50 mg/ml. Carbodiimide reagent I-ethyl-3--dimethyl aminopropyl! carbodiimide, 100 mg/ml in 50% aque- ous pyridine was added in 10-fold molar excessand the mixture was stirred gently at room temperature for I hr. Bovine serum albumin 0 mg/ml in distilled water>was thenadded dropwise in an equimolar ratio of toxin/lysine residues and the reaction mixture stirred gently overnight at room temperature.The resulting cloudy solution was dialyzed for 24 hours against several changesof 0.9% NaCl and the antigen concentration adjusted with normal salineto 0.5-1mg PbTx-3/ml, An adult femalegoat was administeredtoxin/BSA conjugate containing 0.25-0.5rng conjugatedPbTx-3 via intramuscular injection in Freund's com- plete adjuvant. Injectionswere given in I ml total volume, half admimstered tnto eachhind leg, After the initial dose, subsequentboosts were given bi- weekly in Freund'sincomplete adjuvant. Blood sampleswere taken beginning 8 weekspost-immunization and analyzedfor PbTx-3binding activity. Characterizationof Anti-PbTx-3Binding Acfivity Serumwas diluted I:20,000with phosphate-bufferedsaline PBS!,pH 7.4, contaimng 0.0'l%emulsifier Fmulphor EL-620,GAF Corp.!. Triplicate tubes containing antiserumand increasingconcentrations of [sH}PbTx-3 were incu- bated2-4 hours at 4'C. Nonspecificbinding of toxin to serum constituentswas measuredby using parallel tubes containing pre-immune serum matched for total protem concentration,After incubation,0.5 ml of a suspensionof 2 mg/ ml dextran-coated charcoal was added. The tubes were mixed, incubated 15 min at 4'C, and centrifuged 15 min at 1500x g. Aliquots of 0,75rnl of the clear

118 Developmentof a BrevetoxinJmmunoassay

supernatantwere transferredto scintillation vials, acidified with 50ItL IN 1 IG andcounted in a Beckmanliquid scintillation counter with a countingeffj. ciencyof 45%,At eachconcentration of radiolabeled PbTx-3, specific binding was calculated as the difference between total and nonspecific binding. Rosenthalplots wereconstructed by plotting bound/freevs bound DPMat eachlabel concentration, and bindingaffinity KD! and binding capacity B,! were estimatedfrom thesedata. Competition curves,including standardcurves for PbTx-l,PbTx-2, PbTx-3 and PbTx-9,were constructed by incubatingantiserum with increasing amounts of unlabeled competitor in the presenceof a constant [sH]I'bTx-3 concentration.The amountof radiolabeledtoxin bound in the supernatant was quantifiedas aboveand plotted versuscompetitor concentration. Each point was plotted as the meanof triplicate determinations. Variation from the meanwas typically less than 3%. Toxinconcentrations in unknow~samples were determinedby comparison with standardcurves. Solid-phaseExtraction of Urine Urine sampleswere processedover a solid-phaseextraction SPE!column prior to analysis, Up to 5 ml of urine was loadedonto a 6mI SPEcolumn C1S BondElut,Analytichem International, Harbor City, CA! and pushed through with a syringeat a rateof approximately3 ml/min. Thecolumn was washed oncewith 3 mlof water,again with 3 mlof 70%methanol, and the sample elutedwith 3 mlof 100%methanol. The methanol fraction was then evapo- ratedto drynessand redissolved in assaybuffer to theoriginal sample volume, Preparationof Shellf'sh Extracts Clams Mercenariarnercenaria! were purchased fresh at a localgrocery store.The clam was removed whole from the shell, weighed and homogenized to a thickslurry with a Brinkmanntissue homogenizer in a minimumvolume of distilledwater. Thisslurry wasextracted three times with two volumesof acetoneby stirringfor 15-20min at roomtemperature followed by filtration throughWhatman ¹1 filter paper.The filtrates were pooled and evaporated to drynessunder vacuum. The residue was dissolved in 90%methanol ml/ g originaltissue weight! and liquid-liquid partitioned three times with equal volumes of light petroleum ether BP 30-60'C!. The ether fraction was dis- cardedand the 90%methanol fraction again evaporated to dryness.The residuewas taken up in 100%methanol at I ml/g originaltissue weight, sealed in serum vials and stored at -20"C. DataAnafysis All radioimmunoassaydata, including the generation of best-fitstandard curves,were analyzed with the IBM PC Radioimmunoassay DataReduction System,version 4.113.

119 M. A. Poli, et af.

1@SULTS Theconjugation of PbTx-3to BSAas describedhere resulted in a highly immunogenic product, The overall conjugation efficiency of the two-step processranged from 37-49%,and the molar ratio of toxin/BSA in the final product ranged from 7.6-l3.9 n=4!. Higher conjugation efficiencieswere achievedwhen we usedfreshly recrystallizedsuccinic anhydride in the succinylation step. Anti-PbTx-3anhbodies appeared rapidly in the goat serum after the initial immunization.By thefirst bleedat week8, the bindingcapacity had reached 0.5 nmoles PbTx-3/ml serum, where it remained constant for several weeks.After cessationof boostsat weekIO, the binding capacityremained constantfor 6 weeksbefore rapidly diminishing to verylow levels. Resuming boostsat this time resultedin a rapidincrease to approximatelytwice the

0.8

0.7

5 0.$

9 ~ ~0.3

0.2

0.1

0 'IO 20 >HjpbTx-3 nM! eOUtio DPM x 1000!

Figure 2, Characterizationof PbTx-3Binding Activity in Goat Antiserum. Serum was diluted I:20,000with PBSand binding at increasing iaHJPbTx-3 concentrations deternuned as described in Materials and Methods. Plotted points are means of triplicate determinations.Variation from the mean was typically 3% or less. A! Saturability of lsIt]pbTx-3 binding, B! Rosenthal analysisof binding data. In this representativeexperiment, Ko= 0.8 nM, B~= 8 Itg/mb

l20 Developmentofa Brevetoxin lmmunoassay

previouslevels. A second"resting" of theanimal for severalmonths resulted in an evengreater rebound effect. Estimates of the bindingcapacity of the laterbleeds by Rosenthalanalysis of binding datawere 6.7-8.9 nmoles/mi serum Figure 2B!. AntiserumAffi'nity and Specificity Bindingcurves Figure 2A! demonstrated the saturability of I~HIpbTx-3 binding.Rosenthal analysis Figure 2B! indicated a single class of binding

100

80

0 Ql 60 K403 40

20

0 0.01 0.1 1 10 100 1,000 COMPETITORCONCENTRATION rtg/tube! Figure3.Standard Competition Curves for PbTx-1 and PbTx-2-Type Hrevetoxins. Bindingof I3HJPbTx-3was determined as described in Materialsand Methods in thepresence of increasing amounts of competitortoxins. Plotted points are meansof triplicatedeterminabons. Variation of themean was typically 3% or less;error bars have been omitted for clarity. Lines are computer generated best- fit curves.PbTx-2-type toxins:

121 M. A. Poli, ef al, sites with an apparent affinity constant KD! of I nM .5-2.1 nM, n=I 'I!, Standardcompetition curves for PbTx-2,PbTx-3, and PbTx-9 Figure 3! showedthese toxins to be nearly equipotentin their ability to inhibit the bindingof labeledtoxin by theantiserum. The linear portion of the standard curveslay between0,3-10 ng unlabeledtoxin/assay tube. The competitorconcentrations necessary to inhibit 50%of [aHIPbTx-3 binding ICss!were calculated from the computer-generatedbest-fit curves Figure 3!. The IC~ values calculated for PbTx-2, PbTx-3,and PbTx-9 were 1.81.08!, 2,13 +.11!,and 1.084.04! ng/tube, respectively.In contrast,the ICsovalue calculated for competitionby PbTx-Iwas 200 +13!ng/tube. The linearportion of all curveshad a slopeof I +,09!,

125

100

75 Q3

Ql

25

0 0.1 1 10 100 1,000 10,000 COMPETITORCONCENTRATION ng!

Figure4. Lackof Cross-Reactivityof Anti-PbTxSerum with Other Marine Toxins. Binding of [3HIPbTx-3 was determined as described in Materialsand Methods in the presenceof increasingamounts of othermarine toxins. Plottedpoints are meansof tnplicatedeterminations. Variation from themean was typically 3% or less;error bars were omitted for clarity. +! okadaicacid, a ! pectenotoxin-6,! yessotoxin, ~ !saxitoxin, ! palytoxin, ! PbTx-3control. Developmentof a Brevetoxinlmrnunoassay

Competitionexperiments with the diarrhoeicshellfish toxins, okadaic acid,pcctenotoxin-6, and yessotoxin, the paralytic shellfish toxin saxitoxin, and the zoanthidtoxin palytoxin Figure4! demonstratedno inhibitionof [ H]I'bTx-3binding by anyof thesetoxins over the concentration ranges tested.

Radioimmnnoassayof Urirte Greaterthan 5% urine in theassay tubes interfered significantly with the assay data not shown!. Processingthe urine over a CtsSPE column was necessaryto elimtnatethis interference.Control experimentswith urine spikedwith traceamounts of radiolabeledPbTx-3 demonstrated negligible lossof radiolabelin thewater and 70% methanol wash steps, and >99% of the administeredradiolabel eluted with 100%methanol. Interfering contaminants

12

75 7 I .O

25

0 25 50 75 1GO 0 0.1 '1 10 100 X PROCESSEDUR NE COLtPETfTOR ng/tube!

Figure5.Effect of Processed Urine on Brevetoxin Radioimmunoassay Sensihvity. Humanurine was processed through a Cpssolid-phase exhaction column as describedin Materials and Methods toremove interfering substances. A!The effectof increasingamounts of processedurine on I3HIPbTx-3 binding by the antiserum. B! Standard competition curves for unlabeled PbTx-3 in thepres- enceof C!PBS slope = 1.08,ED~ -2.4 ng! or O!50% processed urine slope = 1.07,EDsa= 2.4 ng!. Plottedpoints are means of triplicatedeterminations. Varia- tionfrom the mean was typically 3% or less; error bars were omitted for clarity. Standardcurves are computer-fitted.

123 M. A, Volt, ef al. wereeffectively removed by this process. No significant binding interference wasdetected in assayscontaining up to 90%processed urine Figure5A!. A standardcompetition curve generated in the presence of 50% processed urine Figure5B! demonstrated nochange in thecalculated ICso when compared a controlcurve generated in PBS .44 ngvs. 2.41 ng, respectively!, The urine of a nortnal human volunteer M,P,! was used to assess the overallaccuracy of theprocedure for urineanalysis. Urine was spikedwith PbTx-2at 5, 2,5,and 1.25ng/ml prior to SPEprocessing, The valuesdeter- rninedby radioimrnunoassaywere 4.3 0.2!, 2.3 +0.2!,and 1.4 +0.1!ng/ml, respectively Table I!,

Table l. Brevetoxin lmmunoassay: Analyses of Spiked Samples in Urine and Shellfish Extracts.

Matrix Toxin Concentration RIA Result

Unne PbTx-2 5 ng.ml 4.3 +0.2!

2.5 ng/ml 2.3 +0 2!

1.25ng/ml 1.4 R0.1!

Clam Extract PbTx-3 10 ltg/g 10.8 S0.6!

5 lrg/g 5,6 i0.5!

Radioimmunoassayof Shellfish Extracts Crudeshellfish extract was used directly in the assaywithout SPE treat- ment,No significant interference occurred atdilutions greater than 1:30 3 mg tissueequivalents/assay tube! Figure6A!. Standard curves generated in PBS, SPE-processed,andcrude extract 0 mg tissueequivalents/assay tube! had identicalslopes and ICstt values Figure 6B!. Analysis of crudeclam extract spikedwith 5 and10 p.g PbTx-3/g tissue yielded toxin valuesof 5.6 +0,5!and 10.8 * 0,6!ng/rnl, respectively Table 1 !.

DISCUSSION Saturationcurves and Rosenthal analysis of bindingdata indicated a singleclass of antibodieswith an apparent affinity constant KD! of approxi- mately1nM, Asthe antiserum ispolyclonal, this suggests the predorninan«

124 Developmentofa Brevetoxin lmmunoassay

7$

25

10,000 T,OOO 100 10 0.1 t '10 100 DLUTTONFACTOR CO%ETtTOR n9/tube!

Figure6. Effectof Shellfish Estract on Brevetoxin Radioimmunoassay Sensitivity, Crudeclam Mercertaria mercenaria! extracts were prepared asdescribed in Materi- alsand Methods and dissolved in methanol I g originaltissue/ml!. A! The effectof crude clam extract e! or SPE-processed clamextract ! onf H]pbTx-3 bindingby theantiserum. B! Standard curves for unlabeledPbTx-3 in thepres- enceof D!PBS slope = 1.05,EC, = 3.33ng!, ! crudeclam extract slope = 1.02, ECM=3,55 ngl, or o! SPE-processedextract slope = 0.99,ECM= 3.1O ng!. Final extractdilutions were l.lOO0 rng originaltissue equivalents/tube!. Plotted points are means of triplicate determinations. Variation from the mean was typically3% or less;error bars were omitted for clarity.Curves are computer- fitted.

of a single classor the simultaneousoccurrence of multiple classeswith affinity constantssimilar enough that theycannot be resolvedby Rosenthal analysis,The binding capacity B ! derivedfrom the Rosentha!analysis reached6.7-8.9 nmoles/inl serum -8 pg/ml serum!in thelatest bleeds, This allowedthe assayto be run at antiserumdilutions of up to 1:25,000.Varying thedilution from 1:10,000to l:25,000did not affectassay sensitivity data not shown!,Typically, assays were run at 1:10,000dilution to minimizesample counting times. Theminimum detectionlimits, as determinedby thelinear portionsof the standardcompetition curves for PbTx-2,pbTx-3 and pbTx-9, were approxi- mately300 pg/assay tube. Whi!eother PbTx-2-type brevetoxins have yet to M. A. I'oli, et al. be evaluatedin this system,significant cross-reactivity with PbTx-5,PbTx-6, and PbTx-8is expected,as thedifferences between aB of thesetoxins lie at or near the site of conjugation to the protein carrier", Conversely, PbTx-I was boundto a muchlesser degree minimum detection limit approximately30 ng/ml!, probablydue to the differentbackbone structure. We hypothesize similar results with PbTx-7 and PbTx-10. Theassay described here appears to be specificfor the brevetoxins.No othermarine toxin testedcross-reacted with this serum.Especially notable in this regardwere the diarrhoeic shellfish toxins pectenotoxin-6, yessotoxin, and okadaicacid, each of which possessessome cyclic polyether character,and are known to accumulatein shellfish.Lack of cross-reactivityby saxitoxinand palytoxinwas expected because neither toxin sharesstructural similarity to the brevetoxins.Still to be testedare the ciguateratoxins, including the ciguatoxinsand maitotoxins,as well as the new and as yet only partially characterizedtoxins of Ostreopsislenticularis'4. These compounds will be in- vestigatedas they becomeavailable. We believethis assayholds great promisefor the detectionof brevetoxins in shellfishstocks. Experiments with clamsartificially exposedto P. brevis culturesin the water columnunder laboratoryconditions are underway. Further,the ability to detectthese toxins in humanurine suggests potential for thediagnosis of neurotoxicshellfish poisoning and for monitoringor confirm- ing human exposureto the brevetoxins.

LITERATURECITED Poli,M.A. 1988,Laboratory procedures for detoxificationof equipment and wastecontaminated with brevetoxinsPbTx-2 and PbTx-3.J. Assoc. Off. Ana!. Chem, 71:1000-1002. Tester, P.A., R.P. Stumfp and P.K. Fowler. 1988.Red Tide, the first occurrence in North Carolina waters. an overview. In: Proc. Oceans '88 Conf. Instituteof Electricaland Electronics Engineers, and Marine Tech- nologicalSociety, Sponsors!, IEEE, New York. pp. 808-811, Music,S.I., J.T. Howell and C.L. Brumback,1973, Red tide: its public health itnplications.J. Ra. Med, Assoc,60:27-29. 4. Steidinger,K.A. and E,A, Joyce, Jr. 1973.Florida red tides. Florida Dept. Nat.Resources Ed. Ser. No. 17,Fla, Dept. Nat. Res.,St, Petersburg. Baden,Daniel G, 1989,Brevetoxins: unique polyether dinoflagellate tox- ins. FASEB J. 3:1807-1817. Poli,M A.,T.J. Vende and D.G.Baden. 1986. Brevetoxins, unique acti- vators of voltage-sensitivesodium channels,bind to specific sites in rat brain synaptosomes.Mol. Pharmacol.30:129-135. McFarren,E,FH, Tanabe,F.J, Silva, W.B. Wilson, J.E. Campbell and K.H. Lewis. 1965.The occurrence of a ciguatera-likepoison in oysters,

126 Developmentof a BrevetoxinImmunoassay

clams,and Gymnodirtiumbreve cultures, Toxicon 3:123, 8. Baden,D.GT.J. Mende and R.E.Block. I979, Two similar toxins iso lated from Gymrtodirtiumbreve. In Toxic DinoflagellateBlooms D.L, Taylor,and I LH. Seliger,eds.!. Elsevier New York.pp. 327-334. 9. Baden,I?.G., T.j. Mende,J. Wallingand D.R. Schultz. 1984. Specific a ntibodiesd irectedagainst toxins of Ptychodiscusbrevis Florida's red tide dinoflagellate!. Toxicon 22:783-789, 10. Baden,D,G., T.J Mende, A,M, Szmant,V,L. Trainer, R.A. Edwardsand L,E.Roszell. 1988. Brevetoxin binding: molecular pharmacology versus immunoassay.'I'oxicon 26:97-103. Il, Trainer,V.L. and D,G, Baden.1989. Enzyme immunoassays of brevetoxins,ln: ToxicMarine Phytoplankton, F.. Granelli, B. Sundstrom L. Elderand D.M. Anderson, eds.!. Elsevier, New York. pp 430-435, 12, Abraham,C.E. and P.K. Grover. 1971. Covalent linkage of steroidhor- monesto proteincarriers for usein radioimmunoassay.In: Competitive ProteinBindinq Assays, W.O..Odell and W.H. Daughaday, eds.!. J,B. Lippincott, Philadelphia. pp. 140-I48. I3. Rodbard,D., P.j. Munsonand A. De Lean. 1978,Improved curve-fit- ting,parallelism testing, characterization ot' sensitivity and specificity, validation,and optimizationof radioligandassays. In: Radioirnmunoas- say and Related Proceduresin Medicine, Vol.l. International Atomic Energy Agency, Vienna. pp. 469-504. 14. Tindall,D,R., DM. Millerand P.M. Tindall. 1990. Toxicity of Ostreopsis lettticuIarisfrom the British and UnitedStates Virgin Islands, In: Toxic Marine Phytoplankton, E. Granelli, B.G,G. Sundstrorn. L. Elder and D.M.Anderson, eds.!, Elsevier, New York. pp- 424-429. Ecology and Fisheries Excretionof Ciguatoxinfrom Moray Eels Muraenidae! of the Central Pacific

R.J.Lewis, M. Sellin, R, Street, Ivl.J, I iolmes and N.C. Cillespie Southern FisheriesResearch Centre, I'O Box76, Deception Bay QLD 4008, Australia

ABSTRACT Ciguaterais a diseasein humanscaused by consumptionof otherwise ediblefishes that havebecome contaminated with ciguatoxinthrough the marinefood chain. A long-standingdogma in ciguatera research is thatonce contaminated,fishes retain the original level of toxicityfor manyyears. In thisstudy we analyzed the pattern of change in toxicityof visceraof 218 moray eels,Lycodontisjavanicus Bleeker! collected from Teaoraereke,Tarawa, in the Republicof Kiribati over a 500day period. Our resultsindicated that the concentrationof ciguatoxin per g ofviscera! declined exponentially over the 500day period of this study p.001!. All 47 samplesfrom Teaoraereke testedtoxic to mice,indicating that recruitmentof non-toxiceels could not explainthis decreasein toxicity. Thehalf-life for ciguatoxinefflux from the visceraof moray eelswas determinedto be 264days. Over the 500-day period, the size of the visceradid not vary significantly. Interestingly,no relationshipcould be establishedbetween the toxicityof visceraand viscera weight. The half-lifedetermined in our study assumesthat ciguatoxin input into eelsstopped prior to the start of our collections, The calculatedefflux rate will underestimatethe actual efflux rate if the input of ciguatoxin did not stop at this time but was merely reduced. The efflux rate of ciguatoxin is likely to vary betweenfish species and perhapseven between individuals of a species! and may in part determine the ciguatera risk of a species. The decline in viscera toxicity paralleled a reduction in the number of ciguatera casesin Tarawa, Ciguatoxin was separatedfrom a minor, lesspolar form of ciguatoxin by high performanceliquid chromatography. The implicationsof theseresults are discussedand a model describing the ciguatera risk of a fish speciesis proposed,

INTRODUCTION Humansin tropicaland sub-tropicalareas that eat fish are at risk from ciguatera. This risk stemsfrom the ability of otherwiseedible fishes to be- comecontaminated with ciguatoxinthrough the marinefood chain'-~. The R. J. Lewis,et aL structureofciguatoxin inmoray eels has been elucidated . Ciguatoxin ap- parentlyarises from a less-polarform of ciguatoxinproduced by wild Gambierdiscustoxicus which is oxidized three times! to ciguatoxinas it passes throughthe marine food chain3. Manyspecies ofherbivorous, detritivorous, omnivorous andcarnivorous fisheshave been implicated in ciguatera<. However,certain fishes, i.e, Lutjanus bohar Forss&1, Lycodonfisj avanicus Bleeker, Scomberomorus commerson Lacepede,S phyraena spp.and Epi nephetus spp.,are most likely to cause cigua tera inparticular areas. Outbreaks ofciguatera are often biphasic in nature with aninitial upsurge followed by a gradualdecline in theincidence of ciguatera inan areas'o. In thispaper we report the potential for the excretion decay! of ciguatoxininfishes and propose a model that can account for increasesor decreasesinciguatera risk within and between fish species. This conclusion stemsfrom an analysisof visceratoxicity in a demersalflsh, the moray eel Lycodontisjavanr'cus! collected from Tarawa, Republic of Kiribati,an area whereciguatera has been endemic for several decades Tebano and McCarthy, unpublishedresults!.

MATERIALSAND METHODS

Eel collection Morayeels Lycodontis javanicus! were captured in fish-baitedcage traps setat variouslocations on oceanreefs adjacent to the islandof Tarawa Figure 1!. Theviscera including liver! of eacheel wasremoved and stored frozen priorto air dispatchto Brisbane,Australia. On arrivalin Brisbaneviscera werestill in eithera chilled or frozen state. Viscerawere pooled n=l-'IS! to a convenientsample weight for extraction.3 to 1,0kg!. Eachsample of visceracontained only viscera of similarsize to allow the relationship between sizeand toxicity to bedetermined. The date of arrival of viscerain Brisbane wasrecorded as the collection date and represented the collection of eelsup to onemonth prior tothat date. Duringthis study, eels captured from theocean reefadjacent to thevillages of Teaoraereke, Bikenibeu, Bariki and Betio Figure 1! werefound to be similarlytoxic. However,a pooledsample of viscera fromfive eels from the ocean reef adjacent to Tanea did not containdetectable ciguatoxin.The fleshof oneeel from Teaoraereketested nontoxic, This studyreports the analysis of visceratoxicity of 2l7 eels7 pooledsamples! collectedfrom theocean reef adjacent to Teaoraereke in the center of thetoxic zone on Tarawa. Eels were obtained from mne collection dates over a 500- day period. Extracfr'onoj ciguatoxin Viscerasamples nW7! werethawed and cooked in a plasticcooking bag andthen refrozen prior to extraction.Cooking denatured proteins that oth-

132 ExcretionofCiguatoxin from Moray Eels

Figure 1. Eel Collection Site. Mapof the southern half of Tarawa, Republic ofKiribati. The dashing indicates theouter barrier reef. The areas reported toxic in 1983 are indicated bythe dotted line adaptedfrom Tebanoand McCarthy,unpubhshed!. Eels were collected initiallyfroin outer reefs adjacent to thefive villages indicated. Only results of toxicity testingof eelscollected on outerreefs adjacent to Teaoraerekefrom September1987 to january1989 were used in thisstudy.

erwisehampered homogenization in acetone. Frozen samples were chopped intosmall pieces and minced in a hand-operatedmeat mincer to yielda fine slurry, Ciguatoxin was extractedfrom the viscerawith acetonefor 15 min- utesusing an air powered homogenizer Ystral! fitted with 20 T shaft, Each viscerasample was extracted twice withacetone at roomtemperature. A third extractionwith acetone yielded <0.15% of totaltoxicity", Partialpttrifi cation of ciguafoxirt Twoliquid-liquid partitioningschemes were compared to determinethe most efficient methodfor isolatingciguatoxin, Bothschemes first extracteda 90%aqueous methanol phase with hexane Figure 2!. MethodI then used the standarddiethyl ether-waterparhtioning, while method II Figure2! useda modificationof this method5% ethanol addedto the aqueousphase! sug- gestedby Vernouxef al.'2. Forthis purposea mincedsample of viscerawas dividedinto two equalportions for extraction.After four diethyl ether ex- «actionsusing method 1, a totalof 0.9 g oflipid and1,160 mouse units m.u.!of ciguatoxinwere recovered. Four additional diethyl ether extractions yielded

133 R. J, l.ewis,et al. sozap ael viecaza aod liver eaaplae D.3 - l.a ka!

0'tevopor acatoaa practica ecetooaioaolobla aatezial

aa oo ~ rotavapoz

aetbaool fraatioo hexane-voluble f zectioo

~ Charfractioo voter-aolobla t'zactloo

Figure 2, Extractionof Eel Ciguatoxin. I'rocedureused for the extractionancl partial purification of ciguatoxin. Fractions marked by an asteriskwere testedfor toxicity in mice. 0.3 g and930 m.u. of ciguatoxinand an additional extractionof the aqueous phaseusing method II yielded 2,1g and 145m,u, of ciguatoxin. Extractionof the other portion by method II alone tlu'eediethyl ether extractions! yielded 7.0 g and 2,750m,u. ot ciguatoxin. Consequently,method 11 Figure 2! was used for all assaysreported here. This method has the additional advantage of a rapid separationof the diethyl ether-25%aqueous ethanol phasescom- pared with an often difficult separation of the diethyl ether-water phases method 1!. Much of the high yield of lipid impurities obtainedusing Method Il could be removed with -20'C acetoneprecipitation acetone precipitate contained- 5% of lethality!. The hexaneand water phasescontained no detectablecigua toxin at 1 g/ kg dose! although the high yield of impurities in these fractionsprecluded a sensitiveassay. Characterizationvf toxin contentof viscera A portion of the diethyl ether fraction was purified to homogeneityby low pressureand high pressureliquid chromatography HPLC!. Fast atom bom- bardment FAB! massspectral measuretnents were performed on two toxic fractions separatedby HPLC. Bioassay A portion -5 mg! of eachdiethyl ether Fractionwas suspendedin 0.5 ml Tween 60 saline and assayedin duplicate by i.p. injection into 20 + 2 g Quackenbushmice of either sex. For eachmouse the signs and the time to Excretionof Ciguatoxin from Moray Eels deathwere recorded. The relationship between dose and time to deathwas usedto quantifyeach fraction and is approximated by:

log m.u,=2.3 log I + T'! where,m.u. = numberof mouseunits of ciguatoxininjected and T =time to death in hours". Onem,u. is theLD50 dose of ciguatoxinfor a 20g mouse.Based on the dataof Tachibanaone rn.u, = 9ng ciguatoxin ". Hexaneand 25% aqueous ethanolsoluble fractions were intermittently assayed inmice at doses up to 1 g/i g Statisticalattalt/sis Dataare expressed as the mean 11 population standard deviation, Linear regressionanalyses were performed on unweighted data or ondata weighted with the numberof fishper sample,

600

400 0

0 I 0 I o 200

0 0

100 200 300 400 500 Days

Figure 3. l.el VisceraWeights vs. Time. Averageweight of moray eel visceracollected over a SK day period. representthe average viscera weight sampleweight/number of fishper sample! for eachot 47samples obtained during ninecollections. Viscera weight did not changesignificantly over tiine. Eel collections commenced in September, l'%7. R, J. Lewis,et al.

Moray eelcollection Wholeviscera from 217moray eels, collected over a 500-dayperiod from oneciguatera-prone siteon Tarawa were pooled into 47 samples- The aver- ageweight of whole individual eels from a subsample of38 eels was 3.6+2,3 kgwith individual weights ranging from 0.6 kg to 10 kg. A totalof 35.9 kg of pooledeel viscera were collected from the 217 eels. Figure 3 showsthe averageviscera weight for each of the47 pooledsamples. No significant changeoccurred in the average weight of visceracollected over the period of thisstudy.

U 0 I

0 0 200 400 600 Viscera weight g!

Figure4, EelViscera Toxicity vs. VisceraWeight, Sampletoxicity mouseunits per g viscera!vs. averageviscera weight. Oata representthe toxicity of the 47 samplescollected during the study period,

Visceratoxicity A total of 99,200m.u. of ciguatoxin was extracted from 35.9 kg of eel viscera. Theaverage toxicity was 2.43+ 1.69m.u. per g viscera n=47!and rangedfrom 0.59to 7.3 m,u.per g. No significantrelation was foundbe- tweentoxicity and average viscera weight Figure 4!. Thetoxicity of viscera n=47!was found to declinesignificantly over the 500-dayperiod of the collections Figure 5!. Thisdecline was significant for both weighted using numberof fish per sample!or unweighteddata. The weightednegative

136 Excretionof Ciguatoxinfrom Moray Eels

10

Oe 0 0 is 0

O 0

s 0 INIll

100 200 300 400 500

Days

Figure 5. Eel VisceraToxicity vs. Time. Toxicity mouseunits per g viscera!of eelviscera over a %0day period, Note log scalefor y axis.Numbers adlacent to eachdata point indicate the number of fish pooledfor that sample,Samples of 3 or lessare unmarked for clarity. Toxicity declinedsignificantly over the S00day period n=47,p<0.001!. Eel collections cotnmencedin September1987 linearregression of log toxicity vs time is approximatedby: log y = 0.62- 0,00114x p<0.001! wherey = toxicity mouseunits/g viscera!and x = days from startof collec- tion, Theequation for the weighteddata was chosen as themost appropriateas it takesinto accountthe high variabilityin visceratoxicity. The exponential relationship for efflux decay!was chosenas the simplestmodel to explain the observeddecrease in toxicity. Theslope of this regressionestimated that the half-lifefor the effiux decay!of ciguatoxinstored in the viscerawas 264 days. Averagingthe toxicity for eachcollection date n=9!estimated the half- life for the decay of ciguatoxinwas 316 days. All 47 satnples contained detectablelevels of ciguatoxinand no seasonalf uctuationsin toxicity were evident. Charactert'zationof the tox'rncontent of viscera Two toxinswere isolatedby reversephase HPLC Figure6!. Themore-

137 R. J. Lewis, ef al.

polartoxin was the major component representing approximately 90% of total toxicity. Thistoxin chromatographed similar to ciguatoxinfrom the Spanish mackerel,Scorrtberomorus commersort and was regarded as ciguatoxints. approximateMH+ m/z of thistoxin was determined to be1111.3, similar to 1111.584reported for ciguatoxinby Muratacf al.s. Theless-polar toxin in ducedsigns in micesimilar to ciguatoxinand had a MH+rn/z of 10955 indicatingit differedfrom ciguatoxinby the absenceof one oxygen atom. Presumably,one hydroxyl group on ciguatoxinwas missingin this

10 20 30

LU o V Z

Cl cv O 0 o NCO O

MINUTES

Figure6. HPLCAnalysis of EelCiguatoxin. HPLCelution profile of a semi-purifiedfraction containing 120mouse units m.u.! of ciguatoxin CTX!. Chromatography was performed on a Hamilton PRP-1 reverse phase column im, 150 x 4.1mm! eluted with acetonitrile-water :1! at 0.5 ml,'min and monitored at 206 nm. Ciguatoxinwas eluted at 8.4 ininutes as a homogeneouspeak indicated by the dark bar. A less-polar toxin elutedat 27 minutesas a broadhomogeneous peak. TheAH+ rn/x for ciguatoxin and the less-polartoxin was 1111.3and 1095.5,respectively. The approximate number of mouse units mu! is indicated for both toxins.

less polar toxin,

DISCUSSION Theconcentration of ciguatoxinin the visceraof a populationof moray eels collected from one site within the toxic area of Tarawa was monitored overa 500-dayperiod 't987-1989!, using a modifiedextraction procedure and themouse bioassay. In thisperiod there was a significantexponential decline in the concentrationof ciguatoxin in the eel viscera. The half-life for this declinein toxicitywas estimated to be264 days, Wepropose that this loss stemsfrom excretion and/or decay of ciguatoxin.Excretion comprises the lossfrom eels of storednative ciguatoxin and decay comprises conversion of ciguatoxinto a less-toxicmoeity within eels. Thecalculated half-life assumes

138 Excretionot Ciguatoxinfrom Moray Eels

that the concentrationof ciguatoxinin the diet of eelsreduced to zerobefore thecommencement of eel collections. The actual half-life will beshorter if the concentrationof ciguatoxin inthe diet of eelsreduced only partially prior to thesecollections. Parrotfishesand surgeonfishestaken frotn reefs near Teaoraerekeeight months after the last eel collectioncontained ciguatoxin.As these fishes can be part of thediet of morayeels it is likely the actualhalf-life for the excretion decay! ofciguatoxin from eels is considerably shorterthan 264 days7. Case history data on fishpoisoning including ciguatera!collected by theSouth Pacific Epidemioh1gical and Health informa- tion Servicefrom 1973to 1989 Figure 7! indicatea generalincrease in the 1000 900 800 700 800 500 100 300 200 100

19731975 1977 1979 1981 1983 1985 1987 1989

Figure 7. Annual Incidenceof Fish Poisoningin the Republicot' Kiribati. Data973 to 1989!provided by theSouth Pacific Commission Epidemiological and Health Information Services,and includesmostly casesof ciguateraas well as other forms of fish poisoningincluding histaminepoisoning. number of ciguatera casesover this period. The upsurge in ciguatera in 1986/87 may reflect a transientincrease in ciguatoxin production that could explain the high levelsof ciguatoxin found in eelsat the start of the collection period reported here. At least two other tnechanisms could contribute to the observed reduction in eel toxiCity. Firat, 1'mmigrationof nOn-toxiceels may have OCurredaS a result of our collectionsdepleting eel stocks. This is consideredunlikely for; a! all 47 samplescontained significant levels of ciguatoxin, b! the collection site is in the centerof a large toxic zone and, c! our collections did not influence eel size over the collection period assumingviscera weight is pro- portional to whole weightas found for other fish species4. Eel growth would contribute to a reduction in the concentrationof ciguatoxin in their tissuesof eels. Quantifyingthe effectof growth wasnot possible,as dataon the growthrate of morayeels were not available. Toexplain the reduction in concentrationof ciguatoxinin eel visceraby growthalone, these eels would

139 R. J. Lewis, et al. havehad to doublein sizein 264days. Thisstudy revealed that the concentration of ciguatoxin in the viscera of eelsfrom Tarawa did notcorrelate with eel visceraweight. Lack of corre}a tionpresumably extends to wholeeelweight, as typically viscera weight correlateswith whole weight for fishest4. Vernouxfound a correlationbe- tweensize and toxicity for CarartxFatus Agassiz but not for theclosely related Carattxbarfholomaei Curier captured in thesame area's. L. boharwas found to

ADJACENT NON-CIGUATOXIN PRODUCING AREAS

Figure8. CiguateraRisk: A Model A model describing the ciguatera risk of herbivorous including detntivorous and oinnivorous fishes!and carnivorous Fishesin an area produc- ing ciguatoxin. The risk from ciguatericfishes is driven by the rate of ciguatoxin CTX! production. This rate is increased +! or decreased -! by the influence of both genetic and environmental factors, The horizontal arrows indicate the transfer of ciguatoxin through the food chain from the benthos to herbivorous and carnivorous species, Factors influencing the concentration of ciguatoxin within and between hsh speciesare indicated; c = concentrationof CTX in the total diet of a parhcuiarspecies oF fish; a = first passeFficiency of assimilation of CTX from the diet into the tissueof a particularFish species; o,i = rate of emmigration including mortality! of parficular toxic fish speciesand immigra- tion of non-toxicFish species, respectively; e =rate of excretion decay! of stored ciguatoxinfroin a tissueof a particularfish species leading to a reducedCTX concentration!,g = growth of individualfish species leading to a dilutionof CTX in fishes!. havea correlationbetween size and toxicity irrespective of location' ' . A modelincorporating the influence of keyvariables excrehon, growth, first-passassimilation efficiency, concentration of ciguatoxin in diet, immigra- tionand emigration! is proposedto explainthe ciguatera risk posed by fish speciesin a ciguatera-endeinicarea Figure8!. Therate of productionof ciguatoxin and related compounds! bybenthic dinoflageUate species incl«- ingGtrmbierdiscus Foxicus determines the overall level ot ciguatoxinin the systemand consequently theoverall level of ciguaterarisk3, Genetic and environmentalfactors are proposed asthe key factors influencing ciguatoxin

140 Excretionof Ciguatoxinfrom Moray Eels production,9,17,18 Thisinodel assumes that fishes assiinilate ciguatoxin from theirdiet, as indicatedby feedingstudies with Acanrkurus xantkopterus Valenciennes's,This is a basicrequirement of the fotxt chain hypothesis originallyproposed byRandall', However,noassessment of the efficiency of assimilationof ciguatoxin from the diet hasbeen undertaken for any fish species.The potential to excrete ciguatoxin has been investigated fora few species,L, bakermay not excrete ciguatoxin, while Spkyrarnabarracuda Walbaum!may excreteciguatoxin more rapidly than eels>" 2'. Between- speciesvariability in excretion, assimilation and the dietary concentration of ciguatoxinare likely to explainmost of the variabilityin ciguaterarisk be- tweenspecies. Foreach species the magnitudes ofthe variables presented in the modelmay specifically determine the relationshipbetween fish sizeand toxicity. Within-speciesvariability in the~ factorsmay explain why soine individuals are more toxic than others. Thismodel can explain the different general patterns of ciguatera that are observed10,Both genetic and environmental factors influencing theperiodic- ity of ciguaterarisk and the importof new geneticmaterial perhaps a ciguatoxin-producingstrain ot G.toxicus! may explain the first appearance of ciguaterain an area. The rate of declineof ciguaterarisk will be determined by factorsincluding excretion, growth and mortality. This proposed model may be usefulin the developmentof managementstrategies that may be implementedto reducethe ciguatera risk in an area. Until proceduresto detect toxicfishes become available that are suitable for large scalescreening, the use of a bio-indicator speciesis desirableto measurethe level of ciguatera risk in an area. Moray eelsare a sensitive indicatorof ciguateraand appear useful for the long-term monitoringof ciguatera. Herbivorousspecies are likely to be moreuseful bio-indicators if detailsof short-term changesin ciguatoxinproduction are required. The eels from Tarawa contained at least two toxins. The major toxin is similarin molecular weightand chromatography to the majorciguatoxin from eels from FrenchPolynesia and is similarin chromatographyto ciguatoxin from SpanishmackereL~', A lesspolar form of ciguatoxinhas also been isolated from Tarawa eels. This toxin differs in molecularweight from ciguatoxinby the lossof oneoxygen presumablyby the lossof onehydroxyl group!. This toxinmay be the sameas the lesspolar toxins previously reportedto occurin eel viscera+~.The relationship between "ciguatoxins" fromeels and the "ciguatoxin"found in the fleshand visceraof otherfish speciesremains to be established»24.2r. Species specific differences in as- similationand conversionof the lesspolar ciguatoxin produced by G. toricus mayexplain the different composition of "ciguatoxins" present in different fish species3.The presence in fishesof a classof ciguatoxinswhich have no detectabletoxicity should not be discounted.

141 R. J, Lewis, et ak

Ackitowfedgemerits Thisproject was supported by theFishing Industry Research and Devel- opmentCouncil, Australia. We thank Mr. T, Tebanofor assistancewith morayeel collection, Dr.J. Mcleod for inassspectral measurements, Dr.D, Die for statisticaladvice and Mr. K. Jones Waters Associates! for theuse of a photo diodearray detector. We thank also Miss S. Brice and Mrs. H. Smythfor generalassistance, Mr,G. Smith for photographyand Mrs. E. Donovanfor typing the manuscript.

Literature Cited 1. Randall,J.E. 1958,A review of ciguatera,tropical fish poisoning, with a tentativeexplanation of its cause.Bull. Mar. Sci,Gulf Carribean8: 236- 267. 2. Scheuer, P.J., W. Takahasi, J. Tsutsumi and T. Yoshida. 1967.Ciguatoxin: Isolation and cheinical nature. Science155: 1267-1268, 3. Muiata, M., A,M, Legrand,Y. Ishibashiand T. Yasumoto, 1989, Struc- tures ofciguatoxin and its congener.J. Am. Chem.Soc. 111: 8929-8931. 4. Cooper,M,J. 1964.Ciguatera and other marine poisoningsin the Gilbert Islands. Pac. Sci. 18: 411-440. 5, Halstead, B.W. 1967. I'oisonous and Venomous Marine Animals of the World. Vol. 2. Vertebrates.U,S, Goverment Printing Office, Washington, D.C., 330pp 6. Helfrich, P., T, Piyakarnchanaand P,S,Miles, 1968.Ciguatera fish poi- soning: 1. The ecology of ciguateric reef fishes in the Line Islands. BerniceP. Bishop Museum,Occasional Papers. 23: 305-369. 7. RandaII, J,E. 1979.A survey of ciguatera at Enewetak and Bikini, Marshall Islands, with notes on the systematics and food habits of ciguatoxictishes, Fish.Bull. 78: 201-249, 8, Gillespie, N.C., R.J. Lewis, J,H. Pearn, A.T.C. Bourke, M.J. Holmes, J.B, Bourkeand W.J.Shields. 1986. Ciguaterain Australia. Occurence,clini- cal features,pathophysiology and management. Med. J. Australia145: 584-590. 9. Bagnis,R., J. Bennett,M. Barsinas,J.H. Drollet, G. Jacquet,A.M, Legrand, P.H.Cruchet and H, Pascal.1988. Correlation between ciguateric fish anddamage to reefsin theGambier Islands French Polynesia!. Proc. 6th Inter, Coral ReefSyrnp. 2: 195-200. 10. Helfrich,P. and A3 L Banner.1968. Ciguatera fish poisoning: II. General patternsof developmentin the Pacific. BerniceP. BishopMuseum, OccasionalPapers 23: 371-382. 11. Lewis,R.J. and R. Endean. 1984a.Ciguatoxin from the flcsh and viscera of the barracuda,5 phyraenaj clio. Toxicon 22: 805-810.

142 Excretionof Ciguatoxinfrom Moray Eels

]2. Vernoux,J.P., N. Lahlou,S,A. El Andaloussi,N. Riyecheand L. Ph. Magras.19S5, A studyof thedistribution of ciguatoxinin individual Caribbeanfish. Acta Tropica 42: 225-233. 13. Lewis,R.J. and R, Endean. 1984b. Mode oF action of ciguatoxinfrom the Spanishmackerel, Scomberomorus n>funu rid>ui, on theyuinea-pig ileum andvas deferens, J, Pharmacol.Exp. Ther. 228: 756-760. 14. Weatherley,A.H. and H.S. Gill, 1987. The Biology of Fish Growth. Aca- demicPress, London, 443p. 15. Vernoux,J.-P, 1988. La ciyuateradans I'lie deSaint-Barthelemy: As- pectsepidemiologiques, toxicologklues et preventifs.Oceanol. Acta 11: 37-46, 16. Banner,A.H. 1974.The biological origin and transmission ofciguatoxin. In: Bioactive Compoundsfroin the Sea, H.J. Hurnmand C.E.Lane, eds.!. Marcel Dekker, New York. pp.15-36, 17. Gillespie,N., R. Lewis,J. Burkeand M. Holmes.1985. The siy,nificance of the absenceof ciguatoxinin a wild populationof C. tnxirus.Proc. 5th Inter Coral Reef Cong.,4: 437-441, 18 Lewis,R,JN,C. Gillespic, M.J. Holmes, J,B. Burke, A,B, Keys, A.F. Fifoot and R, Street, 19S8.Toxicity of lipid-soluble extracts from demersal fishes at Flinders Reef, southern Queensland. Proc. 6th Inter Coral Reef Symp., 3: 61-65. 19 Helfrich,P. and A.H. Banner. 1963. Experimental induction of ciguatera toxicity in fish through diet. Nature 197: 1025-1026. 20 Ban~er, A.II., P. Helfrich and T. Piyakarnchana.1966. Retention of ciguateratoxin by the red snapper,Lutj anus bo/urr. Copeia, 2: 297-301. 21 Tosteson, T.R., D,L. Ballantine and H.D. Hurst. 1988. Seasonal fre- quency of ciguatoxic barracudain southwest Puerto Rico. Toxicon 26: 795-801. 22 Legrand, A.M., M. Litaudon, J.M. Genthon, R. Bagnis and T. Yasumoto. 19S9.Isolation and some propertiesof ciguatoxin. J. Appl. Phycol,I: 183-188. 23. Tachibana, K. 1980, Structuralstudieson marinetoxins. Ph.D. Disserta- tion, Univ. of Hawaii. 142 p. 24. Nukina, M., L.M. Koyanagi and P.J.Scheuer, ]971, Two interchange- able forms of ciguatoxin, Toxicon.22: 169-176. 25. Chungue,E., R. Bagnis,N, Fusetaniand Y. Hashimoto. 1977,Isolation of two toxins froin a parrot fish Scarusgibbus. Toxicon 15: 89-93, 26. Vernoux, J.-P. and A, El. Andaloussi. 1986, Heterogeneite des ciguatoxinesextraites de poissonspciches aux Antilles franqaises. Biochimie 68: 287-291. 27. Vernoux J.-P.and F. Talha. 1989.Fractionation and purification of some muscularand visceral cr'guatoxins extracted from Caribbean fish. Comp. Biochem.Physiol. 94B: 499-504, seasonalAbundance and Toxicity of Gambierdiscus toxicus Adachi et Fukuyo from 0'ahu, Hawai'i

E.J.McCaffreyi, M.M.K, Shimizu', P.J.Scheuer' and J.T. Miyahara~ ~DepartmentofChemistry, University of Hawaii at Manoa, Honolulu, Hl 96822.2Departrnent of Pharmacology, John A. BurnsSchool of Medicine,University of Hawaii at Manoa,Honolulu, HI 96822

ABSTRACT

Inshore reef environments of the island of O'ahu were monitored for the presenceof the benthicdinoflagellate Gambierdiscus toxicus from September, 1988to December,]989. Low population levels ranging from 15 cells/g alga to 225 cells/g alga were found in associationwith the macroalgaeSpyridia filamentosa,Acanthophera spicifera, and Dicfyotasandvicerisis. S. fitamentowwas the preferredsubstrate, Severalclonal and batch culturesof G. toxicuswere isolated and grown in thelaboratory. Lethality of crudeacetone and methanol extracts were assayed by intraperitoneal i.p.! injection into mice, The nature of the toxins was assessedby their effecton the isolated guinea-pigatrium.

INTRODUCTION With thesuccessful elucidation of themolecular structure of ciguatoxinby Yasumotoand co-workers,the focus of ciguateraresearch has onceagain turnedtoward ecological aspects'. Historically, ciguatera has priinarilypre- sentedproblems in ecologyand therapy, The ecologicalconcerns centered on the identificationof the primarytoxin producer,the factorsgoverning its geographicaland temporalfluctuations, details of the toxin transmissionfrom dinoflagellatevia microalgae,herbivorous fish, and carnivorousflsh to man,The discovery of Gambierdiscustoxicus Adachi et Fukuyoas a principal toxinproducer laid thefoundation for meaningfulresearch into the factors which influencethe genesisof a criticalconcentration of toxic dinoflagellates thatis likely to causedetectable fish toxicityand hence human intoxication2 s. Ciguateraintoxication has notbeen a majorpublic healthhazard in Ha- waii, but it exists,occasionally makes the headlinesin the newspapersand representsa latentthreat to commercialand sport fisheries4.Although G. foxicuspopulahons had been quantitatively related to macroalgal substrates at a singlesite, no systemahc evaluation of G.toxicus seasonal distribution along E. J. McCaffrey,rt ul. the shorelineof the islandof 0'ahu had beencarried outs. We now report sucha survey,which includes toxicity assays.

METHODS Collectionof Algae Macroalgaewere collected at eightsites Figurc I! on theisland of CYahu by wadingor snorkeling. Algal samples were removed from the substratum

Figure 1. Dinoftage! late Collection Sites, A map of the island of 0'ahu showing the location of collection sites: I! Lua ualei BeachPark, ! The reef runway, Keehi I.agoonside, ! Black Votnt, ! KahalaBeach, ! Hawaii Kai, ! Kailua BeachPark, ! Kaneohe Bay, 8! Coconut Island.

and sealedwithin plasticbags with a minimum of disturbance.Additional specimensof eachalga collected were preserved in 5'7i~formalin in seawater for lateridentification and for an assessmentof the substratespecificity o«he dinoflagellatespecies. At thetime of collection,note was madeof theambient watertemperature, depth and substratum type for eachalgal species sampled Abundanceand Toxic

147 E, J. McCaffrey,et al, inoculatedwith seed cultures of thedinoflagellate approx. 200 ml; 200 cells/ inl!. The cellswei'e maintained for 3 - 4 weeksas describedpreviously, at whichtime the cell density had reached 1 -1.5x106 cells/ml. Thecells were harvested by filtrationonto glass fiber filters Whatman GF/A!, washedwith deionizeddistilled water to reinovesalts and lyophilized,The washed, dried cells were sonicated in acetone and extracted in acetoneover 48 hours.The extract suspension was filteredand the acetone- insolublematerial extracted in inethanolover 48 hours, Solventwas removed from both the acetoneand methanol extracts under vacuum Buchi, Rotavapor!and the residues taken to drynessunder a streamof nitrogen. The crude extracts were stored at -20'C, Male and female mice Swisswhite; 18-23g!were injected i.p. with doses of thedinoflagellate extracts prepared as fine suspensions in 1%Tween 80 in physiologicalsaline. Signs expressed inthe mice following administration of theextracts were identified according to thecriteria detailed in Hoffmanet al, andrecorded at 1 hrand then at intervals up to 48 hourss. The mice were used and sacrificed in accordancewith ethical standards. The cardiotoxiceffects of selectedextracts preparedfrom cultured G. toxicttswere evaluated as follows. Male guineapigs 00-500g! were sacrified, thehearts excised, and placed in oxygenatedKrebs-bicarbonate solution. The leftand right atria were dissected free and used separately to study the effects of the G. toxicusextracts upon the electrically-stimulatedand spontaneously elicitedcontraction, respectively, The left atriumwas stimulated by rectangu- lar pulses<1,5 threshold voltage, 4 msecduration and at 1.5 Hz! delivered througha GrassSD9 stimulator, The isometric responses of both the left and right atria were measuredusing force transducersand recordedon a Grass polygraph,The crude extracts were dissolved in 10%methanol and 100!tl aliquotsof thesolutions added to the organbath to give a final concentration of 10 ill/ml.

RESULTS The results of the geographicalsurvey are summarizedin Table 1. The Kabalabeach site No. 4! appearedto provide a suitablelocation for seasonal monitoring. Two speciesof dinoflagellates,G, toxicusand Ostreopsissiarttettsis wereassessed monthly fora periodof 16months. As seenin Figure2, thecell densityshowed a markedsteady increase during that time. Lethalityof batch and clonalcultures wasdetermined by i,p. injectioni«o mice, Those extracts A for acetone, M for methanol! from batch and clonal cultureswhich elicited toxic syinptoms in mice are indicatedby +! in Table 2. Thoseextracts which caused death are denoted by L!. No acetoneextract provedto belethal. Cultures of Prorocerttrutttlima also yielded lethal methanol extracts. Abundanceand Toxicity of C. toricusin Ifawaii Table I. SpeciesMacroalgae from Sites on the Island of 0'ahu.

A! gal Species Site Substrate G. toxicus O. sia mensis Chlorophyta Hali medaopuntia sand

H. discoidea sand f.'nteromurphasp mud Phacop hyta Turbinaria ornata

Sargassumpotyphyltum rock

S. echirtocarpum rock Padinasp. 1

Padinasp. 2 rock Di ctyotad rchotoma

D, sadvtcertsia sand

D, acutitoba sand Rhodophyta Spyridia fitamentosa rock

3 sand/rubble

4 sand/rubble

sand/mud

sand

mud

sand Acanthophoraspicifera mud

P/ocamium sadvicense rock

Laureucia obtusa rock

L. succisa Grateloupia filici no Unidentified Algae Brown filamentous sand

Brown filamentous

Green Filamentous mud

l49 E j McCaffrey,etal.

2 kg

Egmoo

SOO Cho0 IVS D JF HI It AII J S AIV D0 lttIOIV7 H

Figure 2. DinoflagellateDensities vs. Time. Seasonal variation ot G. Toxicirs! and Ostreapst's~iamensis ~ !.

BOth aCetOneand methanOleXtraCtS Of Hawaiian GambierdfSCIIStnrdCIIS Batch g, tox.KB 1 S/I! havebeen found to be cardiotoxic when tested upon the isolatedguinea pig atrium.Representative traces obtained in the courseof theseexperiments are presented in Figure 3. Theacetone extract A, 10Itg/ml! of C:. toxicIIsexerted a positiveinotropic effect upon both the electrically stirrtulated Figure3, panel 1! and spontaneously beating atria Figure3, panel 3!, This extractalso elicited a positivechronotropic effect upon thespontane- ously beatingatrium Figure3, panel3!. Thepositive inotropic effects of the acetnne extractsof G. Ioxicuswere blocked in the presenceof the adrenergic bluckerspropanolol Pro! and phentolamine Phe! and the sodium ion channel blncking agenttetrodotoxin TTX!. Theeffects of the G.toxicus acetone extract upon the isolatedguinea-pig atria closely resembled those described for ciguatoxinand for extracts of certainspecies of toxic fish . Theeffects of the rrtethanolextracts M; 10pg/ml! of G,toxicus upon the electrically stimulated guirtea-pigatrium are shown in Figure3, panel2. Theextract induced a pcs,itiveinotropic response similar to that recorded previously for maitotoxin,

DISCUSS!ON Irt view of thewidely quoted but asyet unproven hypothesis that ciguatera outbreaksfollow thecreation of newsurfaces from naturalor man-made causes,we had expected a highG. Ioxicus concentration atStation 2, where the roe f had beendisturbed for theconstruction of a runway for the Honolulu >inert', Thiswas not the case. Kahala beach on theSE shore of 0'ahu provedto be the collection area of choice. Again as in an earlier survey the red Abundanceand Toxicity of G. toxicusin Hawaii

Table 2. Cell Extract Yields, Toxicity to Mice ot' Crude Extracts from Cultured Dinoflagellates.

Clonal Yield of Cells Yield of Extract Toxicity to Mtce Cultures tng! rng!

Batch Culture

GT; KB 1/18

A! 23.4 130 240 + + L!

B! 35.6 1.6 5.0 + + L!

TC. tox.; Stn 4 22.2 ] 3.0 13.4 + i L!

Clona1 Cultures

GT 174 242.2 22.0 14,0

GT 177 269.9 120 160 + + L!

PL 4 A! 41.3 7.9 + L!

VL 4 B! 108.3 12.1 6.4 - + L!

GT - Gamhierrtiscustoxicus PL Prorocrrttrum li ma KB Kaha la Beach alga Spyridiafitarttentosa Coulfen! Harvey Cerarrtiaceae,Gigartinaies, Florideae, Rhodophyta!was the preferred host of G. toxicus. Cell density throughout the 16 monthsof the survey was very modest, no doubta reflectionof the factthat normally ciguatera poisonings on 0'ahu are infrequent and often are causedby fish caught outside l lawai'i. Thecultures, as elsewhere yielded two distincttoxic fractions, which by solubility,symptomatology m mice, and their effect on the isolated guinea-pig atrium were related to, or identical with, maitotoxin and ciguatoxin'1, The relationship and biogenesisof thesetwo toxins continue to be unresolved. Ackrtowfedgtrtents Thisresearch was supported by the U.S. Army Medical Research Acquisi- tionActivity and by an M.B.R.S. MH! scholarshipto M.M.K.S.

151 E.j.McCaffrey, eta!. t27

10esn

ux w

k 1rK W

Figure3. Effect of G. toxt'cnsExtracts on IsolatedGuinea Pig Atrium. The inotropicand chronotropic effects of the acetoneand methanolextracts of Cambierdrscastoricnrc Panels1 and3 showthe effectsof the acetone A, 10pg/ml! extracton the electrically-stimulated atrium panel1! and spontaneously-beating atrium panel3! of the guinea-pig.The effects of the additionof tetrodotoxin TTX,5x10 r lVI!and propranolol and phentolamine Pro, lx10+ M; andPhe, 5x!0 r M!on the extract-induced responses of theatria are shown. panel 2 shows the effectsof the methanolextract of G. foxicus !VI,10 !rg/ml! on the electrically- stirnulatedguinea-pig atrium,

LITERATURECITED Murata,M., A.M. Legrand,Y. Ishibashiand T. Yasumoto.1989. Struc- turesof ciguatoxin and its congener. J.Am. Chem, Soc. 111:8929-8931. Yasumoto,T I. Nakajima,R. Bagnisand R. Adachi.1977. Finding of a dinoflagellateas a likelyculprit of ciguatera,Bull. Jpn. Soc. Sci. Fish. 43:1021-1026. Adachi,R. andY. Fukuyo.1979, The thecal structure of a marinetoxic dinoflagellateGambierdiscus toricus gen, et sp. nov. collectedin a ciguatera-endemicarea.Buli. Jpn. Soc. Sci. Fish. 45:67-71, Anderson,B,SJ.K. Sirns, N. Wiebengaand M. Sugi.1983. The epidemi- ologyof ciguaterafish poisonmg in Hawaii,1975-1981. Haw. Med. J. 42;326-334. Abundanceand Toxicity of G. toxicusin Hawaii

5. Shimizu,YH, Shimizu,P. J. Scheuer, Y. Ilokama,M. Oyamaand J,T, Miyahara.1982. Gambierdfscus toxicus, a ciguatera-causingdinoflagel- late from Hawaii. Bull. Jpn, Soc.Sci. Fish. 48:811-813. 6. Steidinger,R.A. 1983. A re-evaluationof toxic dinoflagellate biology and ecology.In: Progressin PhycologicalResearch, Vol, 2, F.E.Round andV.J. Chapman, eds,h Elsevier, New York. pp 147-188 7, Carlson,R,K, 1984.Distribution, periodicity and culture of benthic/ epiphyticdinoflagellates in a ciguatera-endemic region of theCaribbean. Ph.D.dissertation, Southern Illinois University, 305 p. S. Hoffman, I'.AH.R. Granade and J.P. McMillan. 1983. The mouse ciguatoxinbioassay; Adose-response curve and symptomatology analy- sis. Toxicon 21:363-369, 9. Miyahara,J.T., C.K. Kambayashi and Y.Hokama. 1989, Pharmacological characterizationof the toxins in ciguatericfishes. ln: Mycotoxinsand Phycotoxins88 S. Natori, K. Hashimotoand Y, Ueno,eds,k Elsevier, New York. pp, 399-406. 10. Randall,J,E. 1958. A reviewof ciguatera,tropical fish poisoning, with a tentative explanationof its cause. Bull, Mar. Sci,Gulf Carib. 8:236-267. 11. Durand-Clement,lvl. 1987. Study of productionand toxicity of cultured Gambierdr'scus toxicus. Biol Bull. 172:108-121.

153 Societal Impact Public Health, Epidemiologicaland Socioeconomic Patternsof Ciguaterain Tahiti

R, Bagnis,A. Spiegel,L. Nguyen and R. Plichart Institut Territorialde RecherchesMedicales Louis Malardts, BP 30, PapeeteTahiti, French Polynesia

ABSTRACT The study deals with 30 year follow-up of the number of human cases, clinicalfeatures and evolution of thedisease, the names of theciguateric fishes, fishingareas and the socialand economicinfluence of ciguateraon Tahitian life. The evaluation of ciguateric morbidity was carried out by the censusof officially recorded casessince 1960,by the anatysisof standard clinical and epidemiological questionnaires for each patient, filled out by medical and paramedical staff, and by direct inquiries in the population, A significant increaseof the recorded casesoccurred in 1965,the yearly number growing suddenly from 64 to 305. Then the incidencerate per 1,000inhabitants ranged from 4 to 7, with an average of about 480 caseseach year, from 1965 to 1989. During the period of the study, nearly 80species, from 27 fish families, involving various trophic levels were incriminated. They include mainly groupers, snappers, emperor-fish, jacks, surgeonfish, parrotfish, mullets, wrassesand trigger-fish. A great number of the toxic fishes camefrom other islandsthan Tahiti, mainly from the Tuamotu archipelago and they were sold on the marketplace of Papeete.It hasbeen estimated that eachyear on an average,ciguatera results in the lossof about4,000 days of work and 3,000tons of reef fish, that are bannedfrom saleon the market-placein the Tahitian community.

INTRODUCTION Thefirst report on toxicfish fromTahiti appeared in thejournal of a boatswain'smate of the Bountyfrom 1788to 1791<.For the lasttwo centuries thepresently most populated island of FrenchPolynesia 35,000 inhabitants! hasbeen concerned with ciguatera fish poisoning. Studies of overallepide- miologicaland clinical featureshave already been published, however, no long-termsurvey of thespecific human and economic patterns of ciguaterain thebig volcanic island has ever been carried out2~- Theaim of thepresent work is to reportthe follow-up of variousparam- R. Bagnis,et ai. eterssuch as reported cases of ciguatera from 1960!, poisonous fishes, toxic fishingareas, clinical features from 1965!, monthly cases from 1974! socioeco- nomicinfluence and the relationship between symptomatology and ingested fishcaught around the island of Tahiti from1987!.

MATERlALSAND METHODS Theevaluation of patientstook place as part of an epiderniologicalre- searchprogram in progresssince 1965.A standardizedone page questionnairewasused to obtain some infortnation about the clinical features, thenames of thepoisonous fishes and the area of capture,The information wasgathered from patients whose statements andclinical findings were highly suggestiveof ciguatera. Most of thequestionnaires were completed by an examiningphysician, and in a fewcases nurses filled in therequested data. A roughestimation of lost workingdays and of thecost of treatmentassociated with attacksof ciguaterawas made based on thelength of theconfinement eitherat homeor in hospital,without any normal professional activity, for all thepatients examined and treated at the Luis Malarde Institute Clinic over the pastthreeyears, Q FreneltPalynasieL n=29182 ta00

1000 600eoez04aesa70 72 74 7lj 70eo ez 04 ee ee Yesr Figure 1. Ciguaterain Tahiti and FrenchPolynesia. Ciguateracases in Tahiti and in FrenchPolynesia 960 1989!.

158 Cigua tera in Tahiti

A comparisonof the frequencies ot symptoms associated with thetoxic fishand their food habits herbiv

RESULTS Iiicideuce<>f cigt FromJanuary 1, 1960 t<> December 31, 1989, 12,4! 3 casesof ciguaterafish poisoning werc recorded in Tahiti, out of a total of 29,182 in French I'olynesia.The yearly distribution is indicatedin Figurel. lt mustbe ac- knowledgedthat reporting procedureshave changedduring this period.From 1960 the<>ugh 1964, only the severecases of c>guaterawere rep<>rtedby physicianst

55

50 Mbl 45 0 l aer 40V 8e g

0 20 40 60 60 100 120 140 160 160 200 1974 1979 Morlt,hly range >I.r.! 1989

Figure 2. Ciguaterain Tahih: GeneralTrend. Seculartrend of ciguateracases in Tahiti 974 - 1988!.

159 R. Bagnis,et al,

OOOO

1OO

BOO 0~ uBM

BOO

30

1OJuu,Febr. sarahape. Ray Juuu July Bua.Supt. DahNue. Ouu. u-B140 Pertud 191<-1989 muuthu

Figure 3. Monthly Frequencyof Ciguaterain Tahiti. Monthly ciguateracases in Tahiti 974 - 1989!. ing thesefive yearswere anamnestic and were taken from the various consult- ing books of the Tahiti hospital, The yearly averageof lessthan 100cases is very low. From 1965 to 1973,more caseswere reported, thanks to a new program of epidemiologicalmonitoring of ciguateraand the 3,842cases for the ntneyear period gives ana verageof 428cases per year. From 1974to 1989,the reporting was even better after ichtyosarcotoxemiahad been classified as a monthly reporteddisease by the SouthPacific Commission. During this third period,8,140 patients complainedof ciguateraattacks in Tahiti, with an aver- age of 508cases per year. A very light increasein the number casesmay be observedduring this period, Taking themonth of January1974 as a reference, the generaltrend of thedisease may bedescribed by the formula Figure 2!:

Y=0,051x monthly range + 36.9722

Seasonalpatterns were aLsoapparent, with periodic significant changesin incidence, decreasingin March-Apriland increasing in August-September Figure 3!. Geographicaldistributiorr of toxicfi'sh This study coversthe period 1965-1989.For the 11,984cases in which the toxic fish were identified, less than 50% came from Tahitian waters, more than 30% werc caught in peripheral islands and nearly 20% had an unknown Cit,uaterajrt g 45

Out of Tahi 31.80

Tahiti 48.80%

Non ideograph 1 den tif ted 19,80%

a=1 1994 Parlfyft f965- t 999

Figure 4. Ciguaterain Tahiti: GeneralOrigin of Toxic Fish, Ciguatera casesin Tahiti and rough origin of toxic fish.

Ofjfpl Polyffrafa 4 amBIM 47 5 7 0 344 a m li

Aualrff 15% Marff,1183% 151V 4.254

IOV rahnN 4ff6 af 35419

Mar

rkafPfraa 0O5 9 R~4 a Torallyuakvorrrf 9~ <> Foodsforec I 5954 Raafaoraotaff4 5 R-519bl Period 19155-1999

Figure 5. Ciguaterain Tahiti. Detailed Origin of Toxic Fish. Ciguateracases in Tahiti and detailed origin of toxic fish.

161 R, Bagnis,et af. geographicalorigin Figures4 and5!. In Tahiti,no district was safe from endemicciguatera. More than 2/3 of the cases8.01%! were due to fish caughton the west coast. Nearly half the toxic fishes 9%! camefrom the urbanarea consisting of Mahina,Arue, Pirae, Papeete, Faaa, Punaauia, Paea districts Figures 6 and7!, TheTaiarapu Peninsula Tahiti iti! providedless than 27% of the local toxic fishes. Toxicfishes caught in theperipheral island waters responsible for 31,6% of thetotal attacks of ciguaterain Tahiti Figure5!, came from all thearchipela- goes.However, inTahiti the most important supplier of poisonous fishes was theTuamotu archipelago /3 of thecases! leaving less than 30% coming from otherSociety Island archipelagoes, Toxic fishes of unkowngeographical ori- gin weremainly 7%! distributedthrough the Papeete inarket place, 15% purchasedfrom roadside vendors, 8% in foodstores and only 2% were sup- pliedby restaurants,Close to 2/3 of thepoisonous fishes bought on the marketplace, based on available data, came from the Tuamoto atolls. Disfrfbutioriofthe toxic fishes according toichfhyological family Ciguaterafishes were identified in 11,669cases during the 1965-1989 period Figure 8!. Considering all geographical origins, nearly 80 species of

Figure6, Ciguaterain Tahiti:Verified Cases and Distribution of ToxicFish. Distribuhonof officialcases of ciguaterainduced by fishescaught in thevarious districts of Tahiti tram 1965 to 1989.

162 Ciguaterain Tahiti

Manna 041% 9 ill 4 !rarahrhr

0 d Alfddhiha S919 t edda 1079 '4 Paarr 2.499

I Tdahra s 0.28'1 Panda e.S0 0 O

t apara MahdiaaPaprdrr 274 " 410S 0 arr91

n-5824 Pertad 1965-1989

Figure 7. Ciguaterain Tahiti. CasesCaused by Tahitian Fish. Detailed district origin of the toxic fish caught in Tahiti waters. fish belonging to 27 families from the various trophic levels were implicated.The mostoften poisonous fishesfor the patientsof Tahiti were surgeon-fish Acanthurids! and groupers Serranids!,each fatnily involved in nearly 20% of the cases.Jacks Carangids! and snappers Lutjanids, Sparids, Etelids! were eachpoisonous in nearly 12% of the cases,while emperor-fish Lethrinids!, parrot-fish Scarids!, wrasses Labrids! were eachpathogenic in 10 to 7.5%of the cases,and Mugilids in less than5% of the cases.Triggerfish Balistids!,moray eels Murenids!and barracudas Sphyraenids! accounted for 2,5 to 1% of the cases, Less than 1/3 of aH these toxic fishes are herbivores. Thispattern is not valid if we considerthe specificdistribution of onlythe Tahitian toxic fish, which was estimatedaccurately for the last three years Figure9!. The herbivorous fishes,Acanthurids and Mugilids wereboth re- sponsiblefor morethan half the cases,with onlytwo speciesimplicated, the surgeonfish,Ctertochaetus striatus 5%! and the mullet,Crenimugtl crenilabis 5%!. The chief toxic carnivorous fishes were jacks especiallyCltrltnr tttellttttpygus!involved in nearly 12% of the cases,groupers mainly Cephalopholfsargus and Epinephelustauvitta! and snappers mamly Lutjattus gibbusand L. monostfgtttus!with 10% of the casesfor each family, The em- perOr-fiSh LethrittuS 77ti ttiatuS! waS irnpliCated in ClOSetO 7% Of the CaSeS. R. Bagnis,et al.

Various P6riod 1965 1989 7.40Fi n 11 669

tish Giant

Jacks 11.905 Parrot fish e.9OX

Mullets Emperor tish 4,40K 9.70%

11.80" 19.40-

Figure8. Ciguaterain Tahiti:Distribution of CausitiveFish. Casesof ciguaterain Tahitiand toxic fish all geographicalorigins included!.

Relationshiphetweett ctittica! features and toxic species A study was carried out on 629 patients examinedat the Louis Malarde instituteClinic, to establish a correlation between significant differences in someclinical featuresof ciguateraand the food habits oFthe Tahitian toxic fish.Four syndromes were characterized: digestive takingmainly intoac- countvomiting and diarrhea!, neurosensory mainly tingling and numbness, sensoryreversal to cold, muscle and joint pains, chilliness, dizziness, itching!, neuromotor mainly asthenia,and lack of coordination!, and cardiovascular cardiacarrhythmia and decreased blood pressure! Figure 10!. Digestivesigns diarrhea!were significantly more frequent after ingestion of' carnivorous fish S3.55%!than after ingestion of herbivorous fish 1.07%, p<0.001!.No differenceswere observedin the frequencyof occunenceof neurosensorysymptoms. Neurornotor signs generalweakness and lack of coordination! were more often present p.02! in consumersof carnivorous fishes 2.80% and 16.11%,respectively! than in consumersof herbivorous fishes 2% and 9.53%!. The mostsignificant differences were observed for the cardiovascular signs.Bradycardia and the fall of theblood pressure were more frequent p <

164 Ciguaterain Tahiti

10s! in the consumersof carnivores respectively 21.7% and 19.07%!than in the consumersof herbivores .92% and 4.61%!. In the four families more especially studied, groupers, jacks,snappers and emperor-fishes,the lowest and the highestfrequencies of occurrenceof cardiovasculardisturbances were observedin jacks and snappers,respectively.

Others P6riod 1967 lg69 7.20 ~a 689

Giant

Jack ll 90K Surgeon--fish 35.10%

Emperor Ash 5,70%

Snappors 9.10 arrot-fish OFi

roupers MuHets lO 30ea 15,307

Figure 9 Ciguatera >nTahiti: Distribution of CausativeTahitian Fish. Ciguateracases involving toxic fish fromTahiti only.

Socioeconomicfeafures A rough evaluation of the social impact of ciguatera was attempted througha surveycarried out from 1987to 1989among the patients at the Louis Malarde Institute Clinic. The main criteria of evaluation were the length of the restat home generally!or in the hospital in a fewcases!, together with the inability to resume normal activity. Results indicated that about I/3 of the patients who experienceda ciguateraattack had to be confined to bed, The incapacityto work usually ranged from two to sevendays, occasionally lasting threeor four weeks.For the 629 patients surveyed during the period, ciguatera resultedin the loss of about 4,000working days.The averagecost of the nonproductive days,on the basisof the minimum monthly wage,was closeto $100,000 US!.

165 R. Bagnis,et aL

ltypnte»sion

poise stovr

in coordination

Weakness

i'I ching

Chilliness

Bizzi»ess

M»seta/joint pains

Caid reversai

Ti» giing f »umhness

tri a I eh e a n.thtt Vomitingaa7. 4O~ no",. tin;; + SifNIFifhNI oil'I i titNCF

Figure10, Ciguatera Symptoms Correlated with Consumptionof Carnivorousas Comparedwith Herbivorous Tahitian Fish. Frequencyof somecommon or typicalciguatera symptoms by main groupsof fish caught in Tahitian waters 987 - 1989!,

Due to thelack of a simple,fast, reliableand cheapassay for the detection of ciguatoxicity,several species of reef fishesare bannedfrom saleon the marketplace of Papeete.These are the groupers Pfectropomus feopardus tonu!, Epirtephefusmicrodorr hapuu!, the snappers Lutjanus bohar haamea!, L. morrostigmus taivaiva!, L. rinulatus haputu!, the giant wrasseChcilirrus undufafus mara!, the great barracuda Sphyraerrabarracuda ono!, the surgeon- fish Cteuochaefusstriafus mito!, the moray-eels and the trigger-fish generally.This law protectingthe consumerresults in a lossof earningin the order of about 1 million dollars US! for the native fishermen.

DlSCVSSiON The light increasein thenumber of casesin Tahitiduring the 1974-1989 periodmay be associatedwith the demographicgrowth. The ciguaterainci- dencerate remained the same at about 4 per 11,000inhabitants for the period. Ciguaterain Tahiti

Seasonalvariations were observed in theincidence of ciguatera,The period of lowestincidence corresponded with the end of thehot and rainy season, while theperiod of highestincidence was found in thecold and dry seasons.The onlysignificant correlation with a biologicalparameter that appeared con cernedthe macroalgalpopulation. In highislands especially, the alg~l coveringis thelowest in March-Apriland the highest in August-September Payri,personal communication!, A large algal covering is necessaryfor Gambierdiscustoxicus to developin abundance. Concerningthe geographical origin of toxicfish, the relatively minor role of Tahitianfish in theciguatera morbidity may be associated with the decrease of reef-fishstock by overfishing. The higher occurrence of toxic fishes along thewest coast of themain island might be associated with the larger number o f bothfishermen and fish. Overthe last three years, the urban reef area of Tahiti suppliedhalf the toxic fishes. This pattern might be related to thegreater reef damage caused by human activity. On the other hand, in 1987,2/3 of the patientsof the urbanarea, where 100,000 people lived 5% of thetotal island population!, were poisonedby fish from the Tuamotu Islands. Availabledata on the relationshipbetween the clirucalfeatures and toxic fishesconsumed conhrm the higher richness and severity of theclinical pic- tureinduced by carnivorous fish, as previously suggesteds This study also revealeda significantlyhigher frequency of diarrheaafter poisoning by car- nivorousfish, It had previously been reportedthat digestiveand neurosensitivedisorders predominated in patientspoisoned by herbivorous fish . However,the samples of surveyedpatients in theprevious study were smallerand dealt more with systemicsyndrornes than isolated symptoms or signs. Regardingthe strong cardiovascular toxicity of thesnappers, it mustbe notedthat in Tahiti,most of the peopletreated in the hospitalfor a ciguatera attackwere poisoned by snappers.Such a patterncould be associatedeither with a higheramount of ciguatoxinin theflesh or theoccurrence, apart from the ciguatoxin,of oneor severalother toxins which are very activeon the cardiacmuscle and blood pressure. Present studies on thevarious ciguatoxin- like substancesor other toxins presentin ciguatoxicfish are very likely to providean answerto thesequestions in thefuture. Acknoreledgements This study was jointly supportedby the Governmentand the Territory Assemblyof FrenchPolynesia, and by the FrenchMinistry of Research.,to whom we expressour gratitude. We also thank the officials of the local Publtc Health Department and the staH of the various medical centersof Tahiti for their cooperationin thecollection of information,as well asMr. J.Lagardere for revising our English,

167 R. Bagnis,el al.

LITERATURECITED Morrison,J. 1972.The journal of a boatswain'smate of theBounty. London Editions, 1935.pp, 129-13I3.. Bagnis,R, 1968. Clinical aspects ofciguatera fish poisoning! in French Polynesia.Haw. Med. J. 28:25-28 3, Bagnis,RJ. Bennet,M. Barsinas,M. Chebret,G. Jacquet, 1.Lechat, Y. Mitermite,Ph. Pkrolat and S. Rongeras. 1985. Epidemiology of ciguatera in FrenchPolynesia from 1960 to 1984.Proc. 5th Inter, Coral Reef Cong. 4:475-482. Bagnis,R.J. and A, M. Legrand. 1987. Clinical features on 12,980 cases of ciguatera fish poisoning!in FrenchPolynesia. In: Progress in Venom and Toxin Research. P, Gopalakrishnakoneand D.K.Tan, eds.!. Na- tionalUniversity of Singaporeand International Society of Toxinology, Asia-PacificSection. pp. 372-377 Bagnis,R., A.M. Legrand, Ph. Cruchet, F. de Dekker, J.N. Genthon and H. Pascal.1988. Human intoxications by ciguatericphycotoxins in French Polynesia:incidence, clinical and epidemiological features in 1987.In: Mycotoxinsand Phycotoxins. S, Natori, K. Hashimotoand Y. Ueno, eds.! Elsevier, Amsterdam. Cignatera:Clinical, Epidemiological and AnthropologicalAspects

P.Alvarez', C. Duval',T, Col6ni,G. Gonzhlves', E. Martinez',J. G6mez', L. Compres', N, PiAa and P,L.Castellanos2. CiguateraProject, Instituto Dominicano de Tecnologia Industrial INDOTEC!.Banco Central de la RepublicaDominicana. Apartado Postal 329-2,Santo Domingo, Dominican Republic. 2 Asesor, Organizaci6n Panamericana de la Salud OPS!.

ABSTRACT This work presentsthe preliminary results of an interdisciplinary study comprising clinical/ epidemiological,anthropological and laboratoryaspects of ciguatera conducted in Santo Domingo and somefishing communitiesof the Dominican Republic. The epidemiologicalsurveillance, utilizing sentinel centers,allowed the study of thirteen outbreaksand six isolatedcases in which 81 cases confirmed by laboratory analysis! out of 127 underwent intoxication, Information regarding "popularknowledge" about ciguatera, collected by the anthropological team from five communities belonging to different areasof the country is alsosummarized. A thorough analysisof the therapeuticand preventive measures employed as "home medicine" in these communitiesis alsoreported. The advantagesof a multi-disciplinaryteam andthe effortof buildingan interdisciplinary framework are discussed.

INTRODUCTION Ainong the diseasesrelated to consumptionof fresh marinespecies, ciguaterais perhapsthe one having the highestincidence in theDominican Republic.Even though this disease formed a part of our popular culture and generateda numberof beliefs,it had not yet beenstudied locally. Existent data camefrom researchdone in otherCaribbean areas, particularly the Virgin Islandsand Puerto Rico,as well asthe tropical Pacific. Afterthe occurrence of an important ciguatera outbreak in theDominican Republic,aninter-disciplinary andinter-institutional group was organized at the DoniinicanInstitute of IndustrialTechnology INDOTEC!. This group hadas objectives the study of clinical,epidemiological, fisheries and cultural aspectsofciguatera intoxication in this country. The first stage of thestudy programwas liinited to theidentification and characterization ofcases occur- ringin thecity of SantoDomingo and surrounding areas, It alsoaimed at the P. Alvarez, et al. rectificationof popularbeliefs in fourimportant fishing communities located in differentparts of thecountry. AII theseactivities were performed as part of threesubprograrns: clinical/epidemiologicaI, anthropological and laboratory studies. The latter, besidesits researchactivities would assistthe other two concerningoutbreaks and confirmation of cases. Thereis a law in our country banningthe commercialr'zationof four differentspecies of fishes. However, many indications suggested that other specieswere also involved and that the forbiddenspecies were still being consumed. This work summarizes how each subprogram targeted the ciguateraproblem. The results point out the advantages of a multi-disciplin- arystudy and the value of thisgroup in buildinga systemaimed at creatinga betterknowledge of theproblem in our country.This effort will promote developmentof a moreintegrated plan regarding the future prevention of cigua tera.

MATERIALS AND METHODS In orderto get a basicdescription of the intoxication,the clinical-epide- miologicatgroup establisheda surveillancenetwork utilizing five sentinel centerslocated at themost important public and private medical care facilities, includingthe medicalstaffs of the top ratedhotels in SantoDomingo. The medicalpersonnel were trained and orientedby interviews,meetings and printeddata, including a postershowing the clinicalsigns and symptomsof ciguatera.In orderto gatherdata a notificationcard was devised to recordthe personalrecord ot eachpatient. Each case was subsequentlysurveyed by questionnaire.This information was computer processed, incorporating clini- caland epidemiological data given by thepatient and his physician. Using all this informationthe ciguateracases were diagnosed.The diffusion media newspapers,community informants! and the sentinelcenters contributed to caseidentification, However, evolution of the recorded casesafter diagnosis was not followni, Wheneverpossible, samples of the suspectedfish were sent to the labora- tory. Once there, they were identified and maintained at 10'C until processing.Toxin wasextracted by using the McMillan et af. methoddevel- oped in l980'. Toxicity tests were done following methods developed by Sawyer et al., using mice of both sexes,weighing 19-21grams2. Degreeof toxicity wasassessed following the methoddeveloped by McMillan, modified by our laboratory staff~. In this method a range from 0 to 5 is established, where 0 meansno toxicityand 5 maximumtoxicity deathof experimental mice before six hours!>. Popular beliefs related to ciguaterawere explored by direct and indirect observationplus non-structured, targeted interviews, supplemented by pho- tographic and audio records. Indirectly, information was provided by key

0 Ciguaterain the DominicanRepublic informers,members of thecommuruty with enoughexperience torecognize and communicatepertinent information concerning this topic.The four ft'shingcommunities under research were Puerto I'lata, l.a Romana,I'alenque and Pedernales Figure 'I!. Thesecommunities were chosenbrause of their fishing production3,

Figure l. CommunitiesStudied by the Anthropology and EpidemiologySubpro- gram. l! Santo Domingo EpidemiologicalSurveillance! 2!La Romana.3! San Cristobal 4! Pedernales 5> Puerto I'lata

RESVLT5 A total of thirteen outbreaks and six t'solated cases of intoxication were studied in which there were 81 3.8%! casesof intoxication out of 127, Forty fishsamples 82%! out of a totalof fortynine suspected samples were found to have varying degrees of ciguatoxicity after being analyzed by the laboratory.In only oneof theseoutbreaks could the presenceof ciguatoxin not be confirmed. Table 1 summarizes this information. In 74,6% of the cases the fishes were consumed at home while the rest were eaten at restaurants.The conlirmedspecies were "meros" Serranidaespp.!, "jureI" and"cojinua" Cararrx spp.!, "chillo" Lufjanidarayas!, "puerco" Balistes spp.! and "loro" Scarussp.! Figure2>. It wasimpossible to identifythe species in two of the outbreaks. The toxicities of the 49 analyzed samples were as follows:20.5% were grade 0, 10.2%grade 1, 2.0%grade 3, 18.4%grade 4 and 26,5%were grade 5. Noorganoleptic characteristics appearance, flavor, odor! were detectedas indicative of ciguatoxicity.

171 I - rn CJ'

0 P PP' Q Q P U 91999m293 9 RR Q 5 OQ OQ 00 00 & B QQ OQ 0% Q Q Q Q 0 Q Q Q Q Q Q Q Q Q C Q

o7 p

F Rl x x x gz x x QC Q Q Q DJ Q Q Q 2 2 2 C g Cv Cll rt> x pp> Q I 2 Pl

Pc g K'-

+ + + + + + + l + + + + + + + + + + + Ciguaterain thc DominicanRepublic

Interviewsof 67patients out of the81 involved in theciguatera cases were collectedfor clinicaland epidemiological characterization of the disease. From theinterview data, 38 7%! of thepatients were males and the average age «f all patientswas 33 years range 6-77; st. dev.15,5!. Nine of thesepatients declaredthey had other medical disorders prior to theintoxication. In two «f thesecases the disease was of a chronicnature, such as arterial hypertension anddiabetes mellitus. Alcohol drinking combined with fish consumption was reportedfor 15patients 3.1%!. However, synergistic or antagonisticaction could not be evaluated in such cases.

Amount ot Samples

ie 10MeroChllloo.l. POOrOOCorlieI.OrOJvrel SOOOrOIO OOIIPOOSOrrOOOOO ~ hoh Ioxlo ~ olooolox io

Cltfuatera Project: INDOTEC/SESPAS/OPS o,l~ nonloootlll ~ 4

Figure2. SuspectedCiguatoxic Fish Samples Analyzed by the Laboratory,Apzil- November, 1989,Dominican Republic.

The averageincubation period of thepoisoning was 6 hours range0.5-27; st. dev. 4.4!. Castrointestinajsylnptoms were present,nausea 2.8%!, diar- rhea 4.6%! and voiniting0.1%! being their first manifestations.Nervous system disorders were pain in the joints 4,1%!, myalgia 6.7%!, pruritus and/or paresthesis7.8%!, and mainly temperaturesensibility inversion 3%!. Therapyutilized was tairly diverseincluding steroids, antihistaminics, analgesicsand vitamins, accompanied in some cases with fluids andelectro- lytesby parenteraland oral administration. The fishermenas well as the communitiesunder investigation differed greatlyfrom a socioeconomicstandpoint. Puerto Plata has the highest fisher- iesproduction rate of the country, exploiting not only Dominican coasts, but alsonearby extraterritorial waters. This zoneis the centerof the biggest

173 P. Alvarez, et a!.

Amount of Samples 100 21 S

Toxicity Grade Clguetera Project; INDOTEC-SESPAS-OPS

Figure3. Toxicityof AnalyzedFish Samples. Seetext for explanationof ToxicityGrade. fishingcompanies of the country, owning ships and equipment dedicated to deepsea fishing. On thecontrary, La Romana,I'alenque and Pedernales exploitjust thecoastal platform. In the first twolocations the fishermen, workingmostly on a parttime basis, combine their activities with other related efforts suchas tourism and agriculture.Even though Pedernales has the highestfisheries production rate among the three towns, these fishermen wereat thelowest socioeconomic level. Theywork on a fuji timebasis, doing differentkinds of fishing,predominantly using SCUBA equipinent. Somereticence in thefishing communities to talking about ciguatera was evident,even though the visits of theanthropology team were welcomed. This reactionprompted an investigationin orderto findthe hiddencauses of this negativeattitude. The adherence of fishermenand commercial concerns to thelaw of l 975that banned the sale of thirteenspecies of fish consideredto be ciguatoxicwas investigated, Although the numberof bannedspecies was reducedto four in 1986,"barracuda" Sphyraenabarracuda!, "picua" Sphyraena picudijla!, "medregal" Serioja risejiarta! y "pejerey" Alectiscrinitus!, this law wasfrequently broken among fishermen. The apparent complicity between fishermen and fish marketspermitted the latter to buy bannedfishes at re- ducedprices, and subsequently sell themto consumersat the regularprices establishedby thecornrnercial classification. These transgressions of the law didnot bring about frequent confiscations by the authorities, Lack of enforce- mentmay havebeen be due to:

174 Ciguaterain the DominicanRepublic

1. Fishinginspectors, paid to enforce the law, shawM cornrnercial interests with fishermenin that they are ex-flshermen orpart-time fishermen; 2. Amongfishing inspectors, theinterest infishing activities and commer- cializationovercame the ciguatera banning regulations; 3, Lowincome encouraged fishermen to accept bribes; 4. Tradershave develops strategies to catch the high risk species. Ciguaterawas widely known in thefishing communities, because of its highincidence. Inall, the disease was considered tobe simple and as common asa cold,Originally, fishermen believed that fishescaught the ciguatoxin wheningesting the copper left bysunken boats. I~ter, the informationtrans- mittedto themby theFisheries Department of the Agriculture Ministry, made themtranspose the copper concept to "a smallherb" instead, obviously mak- ingreference to the microscopic algae d inoflagellate! responsible for the toxin production.In Pedernales, fishermen also believed that ciguatera was caused by the "aguasvivas" jellyfishes! floating on the sea.Informants believed ciguatoxicitywas characteristic not only of speciesliving in deepwater, but alsofound in shallow species, and that this condition could occur at any time of theyear. Temperature increases and lack of foodwere believed to ob!igate fishesto consumethe small herb producing the toxin. The following specieswere consideredciguatoxic by the fishermen: "albacora" Thunnusafafunga! "barracuda"

Accordingto thefishermen, clinical manifestations of intoxication were: swollenmouth, redness, headache, dizziness, aching joints, tremors, diarrhea, vomiting,lack of motorcoordination, hair loss and blurred vision. The avail- abletherapy in thefour communities was targeted for two different situations: a! whena personhad just ingested a suspectedciguatoxic fish, and b! when first symptomsappeared. In the first situation,the affected person was to drinkmilk, lemon juice, sugar in highamounts, sweetened water, warm water, a mixtureof urineand ash to provokevomiting! or cookingoil to preventthe toxinfrom entering the bloodstream!. In the second situation, people were to drink extractsobtained from red mangrove Rhizophoramangle!, "uva de playa" Coccoloba uvifera!, black cane nodes Sacchnrum officinaru! or sargassum Sargassuesp,!. It isimportant to pointout the high amount of sugar utilized forprophylactic ortherapeutic measu res. The use of sugar issupported bythe factthat according toour informants ciguatera never attacked diabetic people.

DISCVSSION Themost important product of thisresearch is theincreased knowledge aboutciguatera now available for health care facilities, since this poisoning is a seriousproblem for hotels, restaurant services, and the Dominican population ingeneral. Clinical and epidemiological characteristics of the disease were the same as those cited in the literature4-". Gastrointestinal and neurological syrnptorns,specially paresthesia, were the basis for clinical diagnosis in all the studiedcases. Our clinical findings may rein force the belief that cold tempera- turesensibility reversal is a verydistinctive attribute of ciguatera,This feature maybe eventuallyuseful for differentialdiagnosis of this disorder'~,It is importantto notethat 18 9S%!out of 19episodes single cases and outbreaks! wereconfirmed by ciguatoxindetection tests on fish samples.This factillus- tratesthe importanceof a closelink betweenlaboratory, clinical and epidemiologicaI investigators working together to givea promptcase determi- nation, allowing for quick localization of affectedpeople, accurate identificationof causativefish samplesand preservationof fish samplesfor subsequentprocessing. Early determination and localization of caseshas also madeit possibleto startan investigationabout the useof mannitolfor treat- ment ot early stagesof the intoxication, Resultsof anthropologicaI and epidemiological studies suggested the need to examine carefully certain aspects of the clinical treatment of ciguatera.These results suggested we investigatesome of the prophylactic and therapeuticmeasures found in thecommunities for their possibleuse in future prevention programs, Some of the epiderniological characteristics found in this descriptivestudy, suchas concurrentalcohol consumption, previouschronic illness, previous medication use and fish cooking style, could beconsidered as factors for a hypothesisconcermng protection or risk. They

176 Ciguaterain the OnminicanRepublic areactually being evaluated by our research team by using analytical epide- miologicaltechniques, Theseresults will lead to studies of the correlation betweenseverity of clinicalsymptoms and degree of toxicity. Thisinforma- tion will be very usefulin the future. Theanthropological study revealed an unexpected problem. Legal action banningsuspicious fish species from sale influenced negatively the attitude of fishermen,generating a defensivebehavior when asked about ciguatera. This may be due to their desireto protecttheir sourcesof income.Thus, they diminishedthe importanceof ciguateraor hid its existence.Nevertheless, this defensivebehavior isin contracliction with practices they developed to recog- nize ciguatoxicfish and their preventiveand therapeuticmeasures. These practicesare not just "beliefs" assupposed at the beginning!but "popular knowledge",following the conceptsof homemedicine" i". A "belief" sup- poses the absenceof a logical principle, being siinply an adherenceto a statement considered true due to confidence in the source from which it came. "Popular knowledge" is on the other hand a responseof common sense,originating in the observationof reality and experiencethat attributes eventsto cause-effectprinciples and not magic or religious beliefs, If as we postulatethe therapeuticpractices found are part of homeinedi- cine,then they are not opposed to scientific medicine,but rather support its existence.Some of the commentsof our informantssupported this contention: I. Cause-effectrelationship. The conceptof food chain in fishesas responsiblefor ciguatera. Particularfish features smell, eye color, lossof weight,ease of scale detachment!as ciguateraindicators. 2. Susceptibility concept, The belief that strong people will not get ciguatera. 3. Scientific medicine elements. Useof substancesto induce vomiting. The useof sugarand itsassociation with diabetes. 4. Home medicine elements. Use of animal parts,like fish spines,and plant infusions,like the ones comingfrom mangrove, black cane and sargassum, as therapeuticagents. It is importantto establish a clear distinction between "belief" and "popu- lar knowledge."An adequateeducational process should be enough to modify popularknowledge. Since there are not magicalor religiouselements in the treatmentof ciguatera, it is easyto replace empirical medication with a more scientific one. However, it is difficult to modify a "belief"for it requires knowledgeof related cultural aspects. Practices utilized to avoidlish intoxica- tionare related to prevention.These facts should be taken into consideration for the implementation of sanitary-educationalprograms concerning ciguatera. P. Alvarez, et aL Thefindings reported here indicate that the ciguatera problem in the DominicanRepublic requires an in-depth interdisciplinary approach. This is necessaryin order to definethe complex relationship between fishermen, intermediaryfish traders, people and facilities involved with processingand retailingof fish, These relationships should be taken into consideration forthe successfuldevelopment of control and prevention programs, This research hasreinforced the belief that only by closeinteractions between sociologists, clinicalphysicians, epiderniologists, biologists, chemists and other laboratory specialists,will it bepossible todevelop better techniques that will leadin the future to a reducedrisk of ciguateraamong the population,and an inprovement in patient care,

Acknotoledgements Theauthors would like to expresstheir sincerethanks to the PanAmeri- canHealth Organization for its sponsorship and collaboration. Thanks also to theDominican Ministry of Healthfor its collaboration,A greatpart of the laboratorywork wouldnot havebeen possible without the assistance of RafaelinaTejada, Luis Sanchez and Juan Bonifacio. Rafael Alba, a veryfine artist,was the designer of ourinformation poster and brochure. Thanks. We arevery indebted to FrankRichardson for his help with the English version of themanuscript, graphics and pictures, and to FabianTello and Fabiola Oviedo for theircomputing work.

LITERATURECITED McMillan, J,PH.R. Granadeand P.A. Hoffman. 1980.Ciguatera fish poisoningin the UnitedStates Virgin Islands:Preliminary studies. J. Coll, Virgin Island. 6:84-107. Sawyer,P., D, Jollow,P. Scheuer,R.York, J. McMillan,N, Withers,H, Fudenbergand T. Higerd, 1984,Effects of ciguatera-associatedtoxins onbody temperature in mice. In:Seafood Toxins, E. Ragelis, ed.!. ACS SymposiumSeries ¹262, American Chemical Society, Washington, D, C. pp, 321-329. DesarrolloPesquero en la RepublicaDominicana. 1980. Fisheries De- velopmentLimited and InstitutoDominicano de TecnologiaIndustrial INDOTEC!, 435 pp. Bagnis,R., T. Kuberki and S, Lauyer, 1979. Clinical observationon 3,009 casesof ciguatera fish poisoning! in theSouth Pacific, Am. J.Trop. Med. Hyg. 26: 1067-1073. Bagnis,R. 1968.Clinical aspects of ciguatera fish poisoning! in French Polynesia.Haw, Med, J. 28: 25-28, Bruce, S., J. Anderson, J.K. Sims, N.H, Wiebenga and S. Mitsuto, 1983,The epidemiology of ciguaterafish poisoning in Hawaii, 1975-1981. Haw. Med. J. 42: 326-332.

]78 Ciguaterain theDominican Republic

7. Engleberg, N.C., J,G.Morris, J. Lewis,J.P. McMillan, R.A. Pollardand I'.A. Blake.1983. Ciguatera fish poisoning:A major commonsource outbreak in the U. S, Virgin Islands. Am. Intern. Med, 98:336-337, 8, FrenetteCh., J,O. MacLean and T.W. Gyorkos, 1980. A largecommon- sourceoutbreak of ciguaterafish poisoning. J, Infect.Ois. 58: 9, Menendez,E. 1981.Poder, Estratificacion y Salud. Ediciones de laCasa Chata.Mexico, 590 pigs, 10. Nader, F. F. 1951.Fundamentos de AntropologiaSocial. Fondode CulturaEconomica. Mexico, 461 pags,

179 How CiguateraAffects Canadians

E. C, D. Todd

Bureau of Microbial Hazards, Food Directorate, Health Protection Branch,Health and Welfare Canada, Sir Frederick G. BantingResearch Centre,Tunney's Pasture, Ottawa, Ontario KI A OL2,Canada

INTRODUCTION Ciguateraaffects Canadians in two ways, those who are tourists to tropical or subtropicalcoastal regions and thosewho consume imported tropical fishes,Many hundreds of Canadians have been exposed to ciguatoxic fishes andmany have developed syinptoms. Unfortunately, these are rarely prop- erlydiagnosed in the country of origin and only those who persist in seeking medicalhelp on their return to Canada are likely to know the cause of their chronicsymptoms. There inay be 500 persons returning to Canada each year with ciguatera,costing $2,950,000 in medical treatment and lossof work time. Becauseof this concern,the Health ProtectionBranch has alerted Cana- diantour operators of therisks of locallyconsumed seafood and asked them to warn their customers,since law suitshave been initiated by ill touristswho werenot previously aware of therisks. The Branch has also worked with authoritiesin the DominicanRepublic and the PanAmerican Health Organi- zationto establish a National Cigua tera Commission to alert fisherinen, market owners,restaurant and hoteloperators, and tourist agencies of thebest ways of limitingexposure to toxicfishes, and the most effective treatment for the poisoned. Immigrantsto Canadahave consumed toxic fishes in Canadathrough giftsfrom overseas friends or throughthe local fish markets. Many of these immigrantsare unwilling to irutiatean investigation and their illnesses remam undetected,Tropical fishes are now beingserved in restaurantsmore fre- quentlythan in thepast and, therefore, more consumers are being exposed to potentiallyhazardous fishes. Control is limited,however, because the fishes arenot alwayscorrectly labeled as to species or country of origin. Thus, in this categorythere may be as many as 100 cases of ciguateraa yearin Canada,at a cost of $247,000to the national economy.To reduce the number of cases occurringin the future,reconunendations are proposedhere for medical authorities,tourists, tour operatorsand fishsupphers. E.C.D. Todd, et aI,

CiguateraArising From Fishes Eaten In EndemicRegions Canadahas no partof its coastlinein subtropicalor tropicalwaters yet someCanadians are at risk from ciguaterapoisoning associated with tropical fishes. These include the increasing number of tourists going to the Caribbean. An estimated 99,000tourists mostly Canadians! visited Puerto Platain the DominicanRepublic in 1985,compared with 22,000in 1981,In 19SS,153,000 Canadians visited the Dominican Republic as a whole . Reports of individual casesof ciguaterawere first madein the 1960'sand 1970'sby physicianswho had treated patients on their return to Canada,but littlewas publisheduntil the1980'ss. With the development of group tours to specific resorts, severaloutbreaks occurred that could be properly documented on theirreturn. Thelargest and most recent of thesetook place in a resortarea on the southwest coast of Cuba at the end of March, 19874. A fish casserole was served4 hoursbefore a groupdeparted for Montreal,and 57 of the 61eating this dish became ill 93% attack rate! during the flight or shortly after arrival, Once local health authorities became aware of a problem, the travel agencybooking the tour wasnotified and all passengerswere advised to contactthe McGill UniversityCentre for TropicalDiseases, Montreal General Hospital.Symptoms ranged from verysevere to mild, with someof the neurologicalproblems lasting 25 weeks,Ninety-one percent 91%!of cases hadneurological and gastrointestinal symptoms, 5% had neurological symp- tomsalone and 4% had gastrointestinal symptoms alone. No cardiovascular symptomswere observed in any of the patients.Twelve cases were seenand examinedat the Centre for Tropical Diseases,and two of thesehad persistent diarrheaapparently caused by GiardiaIambIia which was found in their stools, This could have resulted from an independent infection in Cuba or from germinationof cystsin thegut throughthe gastroenteritis brought on by theciguatera poisoning. Most cases 1%! consulteda physician mean of 1.7 visits!, one personwas hospitalized and 26% were off work for a meanof 7,8 days, There was no correlation betweenthose who vomited extensively and those with lesssevere illness, indicating that vomiting did not have a protec- tive effect gastrointestinalillness began 1-24hours, mean of 5.7 hours, after consumingthe fishes!. Presumablythe ciguatoxin had already beenabsorbed or was in the intestineby the time vomiting commenced.Cold sensitivity was the most distinctive symptom, manifestedby a burning sensationon touching cold surfaces,disagreeable feeling in cold air, and an oral burning sensatio~ when consuming cold foods. The symptoms are generally similar to those describedin other outbreaksexcept for the extentof insomnia 8% of patients, lasting up to 150days, mean45 days!. Without the follow-up made on their return to Canada,it is likely that very few caseswould have been reported as ciguatera poisoning victims. A control group of tourists who did not eat the fishes were also questioned, but none had distinctive ciguatera-like

182 Ciguatera in Canada symptoms. However, sevenother membersof the group had typical ciguatera illness three days precedingthe outbreak. The sourceof fishesmay havebeen the sameas for the casserole,but thehistory of thetype of fish servedat the resort was not known, as the outbreakwas not investigatedby Cubanauthori- ties, although ciguatera was previously known trom this country . It is possiblesome of the milder casesate a mixtureof toxic andnon-toxic fishes in the casserole. A total of at least ll incidents of illness affecting C.anadianshave been reported, ranging from a few cases to the above episode uf 57 4" '". Freudenthal also records 13 incidents from 1978to 1987 in Nassau County, New York". Sheconsidered that only a small percentageof tourists visiting theCaribbean, who werevictims of ciguaterapoisoning, were investigated by New York health personnel,mainly becausecases did not seekassistance from the appropriateauthorities. Outbreaks and cases in the UnitedStates and its territoriesincluding areas in tropicator subtropicallocalities where ciguatera fishesare locally caught Florida,Puerto Rico, U.S. Virgin Islands,Guam, Hawaii!show two peaksfor incidents,one in 1979and the other in 1985 Table 1!. Sinceunderreporting is widespreadand detailsof the outbreaksare not given, the use of such data to determineoverall trends may be misleadingt~t~.However, it is apparentthat morepeople are ill from fishes consumed at horne than from fishes served in food service establishments.

Table 1. Ciguatera in the United States

a! Outbreaks and Cases

Year 1979 1980 1981 1982 1983 1984 1985 l986 1987

Outbreaks 18 15 15 8 13 18 27 18 11

Cases 85 52 152 37 43 78 106 70 35

b! Outbreaksand Siteof Toxic Fish Consumption

Year 1979 1980 1981 1982 1983 1984 1985 1986 1987

Homes 12 14 3 3 9 t3 20 16 11 Foodservice4 0 2 5 2 0 2 0 0 Unknown 2 1 10 0 2 5 5 2 0

183 E.C.D, Todd, et al.

Table2. CtguateraPoisonings in Canada

Onset t mvtnce Location Hospital Food Countryof Date time Ortgin

January, Toronto, home overnight, dried Jamatca 19fLT Ontario recovered barracuda quickly

April Toronto, horne 0 day,1 grouper Florida 26-27,1984 Ontario 1 day,7 5 days,1

'November Toronto, restaurant a few days groupel 5, 1986 Ontario

April Toronto, home mongol 24-26,1987 Ontario snapper

October Hull, home overnight maht-mahi 24, 1989 Quebec

June7, Kitchener, home overnight grouper Spain 1989 Ontario

February Vancouver, home child, grouper Fiji 26, 1990 British hospital Fiji red Columbia for 6 days sahnon!

t N=nausea, C=abdominal cramps, H=headaches,V=vomiting, D=diarrhea g By ttnmunoassayusing anttbody to partiagypurified ciguatoxin Y. Hokarna, Universityof Hawaii, personalcommunication!. CigtrateraArising FromFish Eaten in Canada ln recent years, the demand in Canada for tropical fishes has increasiM.This is partly becausethere are moreimmigrants particularly Vietnameserefugees! from tropical countrieswhere such fishesare tradition- ally eaten,and also becauseCanadians, travelling more widely to the Caribbeanand pacific, are developing a tastefor such fishes. The demand for ethnicfood as a wholein Canadais expectedto rise 8.1%each year until 1996.For instance, in 1985/86,625 lots of snapper53,676 kg!, 4S8lots ot grouper38,126 kg! and 2S2lots of kingfish2,307 kg! were importedto Ciguatera in Canada

Cases Incubation Duration symptomst ouse at risk! period hrs! toiticity his! rating,

2 ! 1.5 D., weakness,numbnr~s iif tongu~ and lips

9 9! 1 day-16 N.V.C.D.,weakness, tingling or 1-2! $ months numbnessof tongue.law, lace, hands,feet

1 ! days hypertension,brad yea rdia prurit is, not done reversedhot-cold sensation. Case claimedhe hadciguatera through hisknowledge of thedisease, but no sampleswere available for moose bi

7 ! no detailsavailable notdone

1 ! 1.5 pruritisof t

2 ! 5.5 12 hi.V.D., weakness,numbness, in 1-2 arms,legs tongue

4 ! up to 45 N.V.C.H., soregiints, itching 4-5 days tongue

Canadawith a totalof 10lots rejected, none for ciguatoxinpresencel4. This increaseddemand may explain the documentation ofapparent ciguatera inci- dentsoccurring inCanada. These have been reported since 1983 with an averageofone incident and 3.7 cases a year Table 2!. In the seven incidents, all fisheswere eaten at home,except one which was eaten in a high-class restaurant,Five incidents were in Ontario, and one each in Queh~ and Britis ''h Columbia.Thetypes of fishes were grouper four incidents!, snapper, mahi- mahiand barracuda one incident each!. The origin of the fishes was known in onlyfour of the incidents 7.]%!: Florida, Jamaica, Spain and Fiji. The latest E,C,D. Todd, et rtl. episodeoccurred in February,1990, caused by a grouperimported from Fiji. There werefour cases, two of which weremild and two more severe 5, Table3 indicatesthat the dosewas clearly a factorin the severity of the illness. The 5- year old boy was the most seriouslyill. The mousebioassay modified McMillan extractionprocedure followed by intraperitonealinjection of the extractinto mice to give a toxicity rating of 0 to 5 basedmainly on rectal temperature!was performed on the samplesassociated with illnessthat were available'6,Toxicity ratings ranged from 1 to5 Table3!, Thesenumbers do not necessarilyreflect the amountof ciguatoxinpresent, for it is know~that toxins other than ciguatoxinwill causea fall in temperatureand deathin fnice. However,whether these toxins, mostlyundefined, will causehuman intoxicationby ingestion,as opposedto mouseintoxication by intraperitoneal injection,is notknown. Ideally,these results should be confirmed by chemical or immunologicalprocedures that arespecific for ciguatoxin. A grouper caus- ing 9 casesin Torontoin 1984was the only samplethat had someform of verificatt'on.This hada mousetoxicity rating of between1 and2, but also had

Table 3. Doseand Severityof Symptomst

Sex Age lkidy Dose1 Dose Symptoms Weight pieceseaten> g/kg! lkg!

61.25 5 curried 18.4 severefor 3 d,off <1125g! work for 2 wk, rashlasted months

F 13 56.75 l pan fied 4.0 mild, some 25 g! itchinglasting months

M 12 1 pan fried 5.8 mild, no 25 g! continuing symptoms

18.15 2 panfried 24.8 hospitalizedfor 6 f450g! d, ill for 2 tnonths, stiff and sorejoints after excersicefor months

t A family outbreak. Information from A. Hamade,Environmental Health Divi- sion, Health Department, Richmond, British Columbia. g Fishcut into 15approximately ecjual pieces, either pan tried or curried.Each piece weighedabout 225 g.

186 Ciguaterain Canada a ratingof between4 and5 foran ELISA test with polycyclic ether antiserum and sheepanti-ciguatoxin, both for the water-solublefraction in butanoland the ether-solublefraction in methanol Y. Hokama,University of Hawaii, personalcommunication!. IIowever, in 1984the ELISA methodology had not beenperfected and theciguatoxin was not purr

Estimates of Ciguatera Casesand Their Cost in Canada Thereis approximatelyone incident per year of ciguatera poisoning, with an averageof three casesfrom personseating toxic fishes in Canada Table 2!. It has been estimated that most foodborne illnesses in Canada and the United Statesare underreportedby a factorof 350to 1'z'", Thiswould mean that thereare approximately 1,000cases each year in Canada. This figure may seem high, but it is known that most of the victims are immigrants,some ot whom are reluctant to contacta healthagency because of languageproblems or fear of an authority figure. Until surveysare doneof certainsegments of the population, this figure will be difficult to determinewith accuracy,However, becausemost Canadians do not eat tropical fishes,it might be wiser to be more conservativeand selecta figure of 100 casesa year. Costs have beendeter- minedfor one episodein 1990dollars Table4! with an averagecost per caseof $2,470.Hospitalization expenseswere the main factorin this figure. The total annual costsof ciguateraarising from fisheseaten within the country, there- fore, is $247,000. The numberof touristsreturning ill to Canadais moredifficult toestiinate, but certainly exceedsthose at home. In 1988,there were 570,000visits to Bermudaand the Caribbeanfor an averageof 10.0days, 244,000visits to Hawaii for an averageof 14.8days, and 141,000visits to Oceania Australia, New Zealandand the Pacificislands! for an averageof 24,0days~. Therefore, there are about 1,000,000Canadians visiting areaswhere ciguatera is endemic, for anaverage time of ] 0-24days. If it assumedthat 10% of theseeat tropical fishesduring their stay and only one in 200of thesehas some kind of illness fromciguatera, then the number ill wouldbe 500. Costs for theseare higher thanfor residentcases Table 4! with anaverage cost per case of $4,875-$6,920 mean $5,900!.This results in a cost of $2,950,000for tourist-associated ciguatera.Time off workor lost vacation time seems to be a majorcomponent of thesecosts because of the long-term debilitating effects of thedisease. Total costfor both types of ciguateraisabout $3,200,000.

Treatment For Ciguatera Untilrecently, effective treatment had eluded medical researchers, al- thoughthe use of a varietyofdrugs had been attempted in the 1970s and 1980s.In 1988,Palafax et al. published data that indicated intravenous manni-

187 E,C,D. Todd, et aL

Table4. Costof CiguateraPoisoning to Canadiansf

Location Jamaica Dominican Republic Toronto Year 1983' 1985' 1984 Cases Actual Percent Actual Percent Actual Percent Cost Total Cost Total Cost Total

Medical Costs 410 4.2 550 2.7 1,380 6,2

Hospitalizetion NA NA NA 17,610 79.2 investigation NA NA NA NA 1,150 5.2 Value of time off work or vacation 9,340 95.8 20,210 97.3 2,060 9.3 Loss to fish retader 0 0 0 30 Total Costs 9,750 100 20,760 100 22,230 100

Costsper case 4,875 6,920 2,470 Mean costper casefor incidentsinvolving tourists in Jarnaincaand the Dominican Republic 5,900

tActual costgiven in l990 Canadiandollars. tol, when given I g/kg over 30 min, had successfully lessened symptoms'9.All neurologicand neurosensorymanifestations were resolved completely,usually within 48 hours, although improvementswere noticed in minutes. astrointestinal symptoms disappearedmore slowly. All 24 pa- tientsfrom the Marshall islandstreated gave noticeable improvement. Pearn et al. confirmed the efficacy of intravenous mannitol in a study of 12 Queenslandcases, although some responded better than othersso. They used the sameconcentration of mannitol,after rehydration therapy, over a 45 min period. Fortwo casesthis was repeatedwithin 24 hours. A lower doseof 0.5 g/kg infused over 1 hour did not seem to be effective, suggesting that the rapid increasein osmolalityat higherdosages may be important, The mecha- nismof the therapeuticaction is not yet known. ln summary,intravenous mannitol infusion seems to bea mostpromising form of treatmentfor casesin theacute stages if givenat a concentrationof 1 g/ kg over30-45 min. Improvements were rapid even for onepatient 8 daysafter ingestionof toxicfishes. It is a cheap,simple procedure that canbe usedin areaswith limitedmedical facilities. Previous experience with intravenous manrutolfor therapy in otherclinical syndromes hascaused few problems, but patientsshould be properly hydrated aftervomiting and diarrhea! before the Ciguatera in Canada inannitolis given.Pearn ef al.are planning a single-blind clinical trial and otherresearchers should be attempting the same before a whole-heartedrec- ommendationcanbe made . If thisstyle of therapy is justified, then patients whoare still ill ontheir return to non-endemic areas, could also be given this treatment.Therefore, centers for tropical disease should be aware of this type oftherapy and advice given to those who have the symptoms toseek this form of treatmentas soonas possible.

CiguateraControl in the DominicanRepublic As a resultof complaintsby touristsill in theDominican Republic in the mid 1980s,I was askedto work with representativesof the Pan American HealthOrganization and the DominicanRepublic to determinethe extent of theproblem and recommendpossible control measures. The following was the position in the Dominican Republicin 1987con- cerning cigua tera, 1. Tourists aregenerally unaware of ciguateraand are more likelyto be familiar with danger of non-potablewater and infectionsfrom shellfish. In Mexico, in contrast, tourists are usually we/1 aware of the potential for gastroenteritisand may even expectto be ill, but it doesnot stop thein from going there. 2. The demandfor fishesby touristsand localresidents and for exportis increasing,and is beingencouraged by the government,with someretailers andprocessors expanding rapidly, Fish costs, however, are relativelyhigh and this is one reasonlocal peopledo not eat as much fish as in other Caribbeancountries. Also the fearof ciguateraor other formsof illnessis recognized.The high price also limits exports to manycountries. Freshwater Tilapiafarming is beingencouraged by thegovernment asanother source of fish protein. 3. Speciesconsidered to be potentiallytoxic are greatbarracuda Sphyraeria barracuda!, southern sennet Sphyraeria picudifla!, Atlantic moonfish Selertesetapiririis!, Atlantic lookdown Selertr vomer!, alrnaco jack Seriola rivoJiana!,greater amberjack Serio!a durrirrili!, yellow jack Caranx barfhofomaei!, Africanpompano Alectis ciliaris!, tiger grouper Mycteroperca tigris!, black grouper Mycteroperca boriaci!, yellow An grouper Mucteroperca vetrettosa!, greenmoray eel Gymriothorazfatiebris! and frigate tuna Auxi s thazard!. 4. Knowledgeof ciguatera illness is limited because a!there is no report- ingsystem for monitoring ciguatera cases, b! theinethods used by fishermen to recognizetoxic fishes have not been scientiFically conArmed. In addition, somemarkets are selling potentially toxic Ashes included in theold list of prohibitedfishes but dropped froin the new, revised list!. Thisnew ban prevents year-round catching of 4 species ofnon-com- mercialAshes Sphyraena barracuda, S, picudiVa, Seriofa rivoliana, Afectis E.C.D. Todd, et aL ciliaris!.It replacedone that prevented the harvesting of manymore species but only from Mayto October,The rationale for this changewas not ex- plained,but it is apparentthat somecommercially important fishes are allowedto besold that were previously considered to beseasonally toxic. SuggestedControl Strategy I. Immediate Action -6 months! 1. Touristsshould be advised by tour operatorsof thepotential of illness from eafingtropical fishes,and resortoperators should be responsiblefor obtainingfishes least likely to be toxic,i.e., small individuals of speciesnot knownto be locallytoxic, and importedfishes and shellfishfrom temperate seas. An educationalleaflet or notice could be designedlocally, explaining the problemand that every precaution has been taken to minimizeillness, 2. Physiciansattending resorts, hotels and in hospitalsin touristregions should becomefamiliar with ciguatera and other seafoodtoxin problems to giveadvice for careand treatment.They should also counsel cases that most of them will recoverwithin a few days to weeks, althoughweakness may persistand recurrencesmay be triggeredby exercise,alcohol or eatingnon- toxic fishes. Physicians should also report incidents to the appropriate governmentauthority ies!and fish samplesshould be retrieved for testing. 3, Designatedlaboratories should beable to test fishesfor the presenceof ciguatoxin, and they should be familiar with the mousebioassay with positive and negativesamples. 4, Literatureon ciguatoxin and other toxins should becollected and filed for reference. 5. A National Ciguatera Commission should be establishedto coordi- nate these activities. ll. Short-rangeAction -12 months! An educationalstrategy should bedeveloped not only for tourists but alsofor fishermen,retail shopand restaurantoperators. 2, Epidemiological teams should be set up to link with government departments to obtain a better understanding of the true nature of the ciguateraproblem in the DominicanRepublic. Bureaucratic responsibilities mayhave to be alteredto allow oneperson or group of persons!to monitor and summarizethe information.A phonesurvey could be attemptedand emergencyroom hospital recordsreviewed for the nature and seasonalityof illnessesinvolving consumption of fishes,as hasbeen done in the Virgin islands. 3, A limitedsurvey should be conducted to testrepresentative samples of the main economicspecies of fishes oneseaten most often!. If fishes, however,are identified as toxic by the mousebioassay no immediateban is suggestedunless large numbers of thesame species are involved or fishesare associatedwith huma~ illness.It is alsoimportant to makesure that the

190 Ciguaterain Canada laboratorymethodology is accurate and at leastone other laboratory should confirmthe toxicityif enoughsample remains, Extracts could also be sent to laboratoriesin othercountries for confirmationby thebest current methodol- ogy,Trimmings such as tails, heads, intestines, as well as whole fishes obtained by inspectorscould be used as samples. Also, fishes considered tobe toxic bv fishermenshould be testedto seeif their localknowledge is valid. These fishesmay turn out to benonwiguatoxic but maystill beinappropriate to eat becauseof disease,spawning conditions or spoilage.! 4, The laboratoriesshould also learn to detectscombroid and possibly other fish toxins, as well asciguatera. 5. Contacts with other Caribbean countries should be established or strengthenedto help determinethe most useful strategiesfor dealingwith the ciguatera problem llfi Long-term Action -5 years! 1. The monitoring of ciguateracases in tourists and residentsshould continue to help determine what fishes are involved, locationsapparently toxic for specificspecies, seasonal changes and yearly changes. The supply and toxicityof fishesobtained from beyondDominican waters need to be carefullynoted so that there is no confusionwith localsources. 2. More extensivesurveys based on the limitedsurveys done earlier shouldbe attempted with resources obtained from various agencies. Samples takenfrom specific locations at certai~ times of theyear, based on the previous surveyand epidemiological data,may define more precisely where ciguatoxic fishesare likely to occur. Additional sampling should also be done following hurricanes,shipwrecks, construction, dredging, dumping, where coral reef damageis expectedor newrock surfaces are available for dinofiagellate colonization.A scientificproject to determinethe presence of Gamb~erdiscus toxicus,Prorocentrum spp. and other potentially toxic dinoflagellate species, andassociated symbiotic algae, may be attempted butthis should be done with biologistshaving research experience university and institution teams, includingexperts from other countries!. 4. Theuse of a siinplified"stick" iinmunoassay testshould be available withinthis year, This could be used by inspectorsat lishing communities, sportsfishermen, processors, foodservice operators andlaboratory analysts. 5, TheNational Ciguatera Commission should convene atregular inter- valsto controlthe whole process and should be vigilant concerning changes becauseciguatera toxicity seems to be a variablephenoinenon. Action Takenin theDominican Republic and Canada A NationalCiguatera Commission wasestablished inthe Dominican Republicwhich initiated project Ciguatera tomake the local public and tour- istsaware of the ciguatera problem. The Commission: E.C.D, Todd, et a!.

producedposters and pamphlets in Spaiiishand English; alertedphysicians and nurses of thesymptoms and supplied appro- priateillness investigation I'orms; established Instituto Dominicano de Tecnologia industrial INDOTEC!,responsible for testing and tracing suspected fishes; informedvisitors that fish caughtbetween November and April may be toxic,even though fishermen consider this the leasttoxic period; publicizedthe banon barracuda, southern sennet, jacks and African pompano; organizedscientific conferenceson ciguatera. More details of the activities carried out by the Dominican Republic are publishedin theseproceedings, The I lealth Protection Branch in Canada has also become more involved in the following areas. 1. The Health ProtectionBranch is cooperating with the Pan American HealthOrganization in attemptsto developcontrol strategiesin the Carib- bean. 2. A Health Advisory notice hasbeen issued describing the diseaseand its originsand hasbeen sent to tourorganizers. It concludesthat "ciguatera poisoning is not new and hasbeen recognized in returning travellers for inany years. It is likely that many undiagnosedcases have occurred in the past in tourists to the Caribbean. Caseshave also occurred after eating toxic Ashes importedinto Canada.Tour organizersand resort operatorsshould be aware of this problem,Steps should be taken to ensurepotentially toxic species of Ashesare excluded from menus servedto vacationers,especially when little choiceof food is offered in a packagevacation. Travellers should be advised to avoideating potentially toxicfishes, as definedby local authoritiesfor the area of travel. Further inforination is available from Dr. E. Todd 13! 957-0887, Health Protection Branch, Food Directorate," It is interesting to notethat thisadvisory notice was translatedinto Span- ish and circulated in the DominicanRepublic where it was drawn to the attentionof some United Statestourists who wereaffected by ciguatera. 3. There is interestin pursuing methods permitting the detection of ciguaterain individualfish before they are marketed.

Other Recommendations As a resultof the occurrenceof a largenumber of ciguateracases 03! on thesmall West Indian island of Antiguain September-October,1980, specific recommendationswere made l'. Andrews, Centers for Disease Control, At- lanta, GA, USA, personal communication!.These included improving surveillance,especially after storms and otherfactors disturbing coral reefs, educationof consumers,fishermen and exporters, and exchange of informa-

192 Ciguatera in Canada tionby countriesthat have a ciguateraproblem. lt wasnoted that the t'i»his causingthe illness had been bought from the fish market in themain town one monthafter a hurricanehad passed through the area. Also, at least some of thefishes had been caught off RedondaRock, an area known to contain toxic fishes.One reason for fishingin areasthat are generally rcwognized as. per pound. Becauseof ciguateraoccurring in touristsfrom the New York area, various recommendationswere suggested by Freundenthal".These included better epidemiologicinformation, holding of fishcarcasses fortwo day» to allow for testing,preferably by the "stick" iminunoassaytest, education of consumers and tourists, counseling of victiins, and overall coordination of the various groups involved. Toimprove awareness and ultimately reduce the prevalence of ciguatera it is recommended to: 1, Makephysicians and nurses more aware of ciguatera,particularly in areaswhere the disease is mostlikely to beencountered. These include tropi- cal communities where seafoodis eaten,and hospitalswith tropical disease centersin largecities where tourists returning froin overseasvacations may seek help. In addition, medical personnel associated with ethnic commu nities where tropical fishes are imported, and pharmacistswho may be askedfor drugs to help lessenthe syrnptoins,should alsobe alerted. 2. Promotethe use of any proven therapeuticmeasures, including the useof intravenous mannitol, at leastin the acutestages of illness, 3. Inform the public of the possibledangers of eating potentially toxic fishes through postersand warning noticesat l'ishingcoinmunities, fish inar- kets, travel agencies,and tourist informationbcioths, Multilingual notices maybe required in certainareas. It mayalso be desirable to alertspecifically thoseat highestrisk young children,pregnant and nursingwomen, and possiblythose with underlyingdiseases!. 4, Acquirebetter epidemiologic data from acuteand chronic cases, e.g., typeand duration of symptoms,time of year,fish species, and amount of fish eaten.Hotlines can be set up for ill personsto contactphysicians for counsel- ing as well as reporting. 5, Publicizeany bans on specificfishes that have been appropriately derived,including information onspecific seasons, if any, to avoid consuming potentially toxic fishes. Encouragecontact between countries which have an endemic ciguatera problein, through internationalorganizations like CAREC.Information on outbreaks,areas where toxic fishes occur, seasonal- ity,and adoption of specific control measures should beshared. Standardized formsfor assisting the investigation of outbreaks and follow up of cases

193 E,C.D. Todd, et al. should be agreedupon. Althoughthese should be ascomplete as possible, they alsomust be practicalfor investigatorsto fill out. Investigationsof out- breaksand single casesshould be reported to a national agency and made available to public health officials of other countries. 7. Correctly label fishesat marketsand for export so the speciescan be accuratelydetermined if it is involvedin an illness.Consider the possibility of storing frozen fish parts heads,tail viscera!for at leasttwo days at restaurants after the remainderof the fish hasbeen prepared into a mealand served. This would allow the possibility for toxin detectionif illness occurs. S. Make thebest methodologyavailable to analystswishing to test fishes for ciguatoxin or other fish toxins. Existing bioassay techniques shou.ldbe replaced by specific chemical or immunological proceduresonce they are validated and standards are available.

LITERATURECITED

1, Anon. 1986. Turismo en cifras, 1985. Secretaria de Estado de Turismo, RepublicaDominicana. 2. Statistics Canada. 19S9. 19SS International travel. Travel between Canada and other countries. 66-201, Ottawa, Canada, 3, Todd, E. 1985.Ciguatera poisoning in Canada, In: Toxic Dinoflagellates D,M. Anderson, A,W, White, and D.G. Baden, eds.!. Elsevier, New York, pp. 505-510. 4. Frenette.,C., J.D. MacLeanand T,W, Gyorkos. 1988.A large common- sourceoutbreak of ciguaterafish poisoning. J. Infect. Dis. 158:1128-1130. 5. Munoz, E.V. 19SO.Toxicidad y alimentaci6n de algunos peces sospechososde provocar la ciguatera. Informe Cientifico-TecnicoNo, 123. Academia de Ciencias de Cuba, La Habana. 6. BourgaultA.-M., and E. Todd. 1985.Ciguatera poisoning acquired in the DominicanRepublic - Quebec.Can. Dis.Weekly Rep,11:151-2. 7. Cohen, L., J.D. MacLean and R. Lalonde. 1986,Ciguatera fish toxin poisoning: An outbreakin Canadiantourists in the Dominican Republic, Can. Dis. Weekly Rep. 12:73-75. 8, Heimbecker,R,O, 1979,Ciguatera poisoning - snowbirdsbeware. Can. Med. Assoc. J. 120:635. 9. Philpott, J.P. I987. Ciguatera poisoning in travellers to the Dominican Republic. Can. Dis. Weekly Rep. 13:71. 10. Wittes, R.C.,and J.D.MacLean. 1984.Ciguatera poisoning with neuro- logic findings: four more probable cases, Can. Dis. Weekly Rep. 10:202-204. I 1. Freudenthal,A. 1990.Public health aspectsof ciguaterapoisoning con- tractedon tropicalvacations by North Americantourists. In: Toxic Marine Phytoplankton E. Graneli, B. Sundstrom, L. Elder, and D.M,

'194 Ciguaterain Ca> ~

Anderson, eds.!, Elsevier,New York, pp. 45S-463, 12. MacDonald,K., and P.M.Griffin, 1986.Foodborne disease outbreak annualsummary, 1982. Morbid. Mortal, Weekly Rep. 35 SS-1!:7-16 13. Bean,N.H., P.M.Griffin, J.S.Goulding, and C.B. lvey. 1990.Foodbo~ diseaseoutbreaks,5-year summary 1983-1987. Morbid. Mortal. We kl Rep. 39 SS-I!:15-57. 14. Fisheriesand OceansCanada -19S7. CanadianFisheries Annual Stat;~t; cal Review. 1986. Vol, 19. Ottawa, Canada. 15, Garry,J.D., K, Higoand A. Harnade.l990, Ciguatera fish poisoning British Columbia Disease Surveillance, ln press. 16. Hoffman, P.A., H,R, Granade and J.P. McMi!lan, 1983. The mouse ciguatoxinbioassay: a dose-response curve and syrnptornology analysis. Toxicon 21:363-369. 17. Todd,E. 19S9a.Preliminary estimates of costsof foodbornedisease in Canada and costs to reduce salmonellosis. J, Food Protect. 52:586-594 18. Todd,E. 19S9b.Preliminary estimates of costsof foodbornedisease in the United States. J. Food Protect. 52:595-601. 19. Palafax,N,A., L.G.Jain, A,Z, Pinaro,T.M. Gulick, R.K. Williams, and 1,J, Schatz. 19SS.Successful heatment of ciguaterafish poisoning with in- travenous mannitol. J.A.M.A, 259:2740-2742. 20, Pearn,J.H., R.J.Lewis, T. Ruff,M. Tait, J. Quinn, W. Murtha, G. King, A, Ivlallettand N,C. Gillespie.1989. Ciguatera and mannitol:experience with a new treatment regimen. Med. J. Austral. 151:77-80.

K. Morris

Georgetown Hospital 0:30 p,m.! most of the symptoms associatedwith the poisonwere already very much evident. Thetime of appearanceof first symptomsafter consumption varied from half an hour to forty-eight hours. It must alsobe mentionedthat many victims sought medical attention out of caution! even though they experiencedno symptomsat all. Whenone considers the time, distanceand transportation problemsinvolved it wasvery prudent of thevictims not to sit and wait for the first symptoms to appear,hoping that they would never appear, From all accountsthe fish was fresh and not preserved in any way, and preparationwas done in the normalways, Thefish hadbeen speared by a thirty-seven year old fisherman,with sometwenty years spearfishing experi- ence,from Fancy a nearby village to the west! who was on the government payroll andwho should have been on thejob at that time. Threeother barra- cudas were caughtby the samemethod on the sameday by the sameman and at the samelocation Quashie Head!, one being larger than the ciguatoxic one. The fish wasestimated to be about thirty pounds 0 lbs! in weight bv the fisherman himself and by those who saw the whole fish. lt was caught at about 7:00a.m. and by 8:30a.m. was on salein Owia by the fisherman himself. Sellingprice was two dollarsand fifty cents $EC! per pound. The fisherman confirmed that the fish looked no different from any other barracuda he had evercaught, or he would not havesold it, I le did acknowledgehearing that barracuda were sometimespoisonous. The predominant syrnptornsexperienced by the victims were vomiting and diarrhoea but most of the classicalsymptoms were present. Later symp- toms appearingone week after consumption! included the shedding of pubic hairs in femalesand a burning sensation.during micturition in both malesand females Personal communication with medical staff and patients at the Georgetown Hospital!, Of the more than one hundred and five people who sought medical attention 52.38%were fernale,47.62% were males and agesranged from 1 I/2 years to 78 yearsof age. The age distribution of the patients was; 0- S years 12.38% 6 9 years 14.28% 10-1S years 17.14% 16+ years 56.20 % The averageweight of fish consumed per person was 1.41 ozs. Some individuals had no fish but drank only the 'sauce' in which the fish was boiled,The type of fish preparationmade no difference with respectto the time of appearanceor featureof the symptomsobserved. Most people had the fish boiled,and somehad morethan one'serving' of differentfish prepara- tions, The breakdownof patientswith respectto the typeof fishpreparation theyconsumed was as follows;

198 Ciguat ra in St.Vincent and thepro>

Boiled 70,48 % Stewed 2.85 % Sauceonly I 8.09% Fried 7.63 % Roasted 0.95 %

Manyof the victimsreported experiencing their first mild! syrr Pt<>rr>t'pt about one to two hours after consumption. At about six ! hours aft' 'r c<><'< sumptionmost of thevictinis had already experienced the sympton~s t<>ti1I of fifteenpeople ate the fleshof the fish and/or drankthe 'sauceyet encedno symptoms.Out of precautionhowever they sought medical at t~>1ti'r> ri at the GeorgetownHospital, OF this group,40% were in the l.5 5 yo>r ~>lg range, 27%between 6 - 9 years,6"~< between l0 - 15years and 27'il sixtr>s~~~ ing the remainsof foodthat was thrown away by their respective owners,haft~.r they had determined that the food was the causeof their poisoning domestic animals were cats, dogs and chickens,the last of v hich ar< utilized as a sourceof meat protein. In the casesof the dogs and cats, ~ I<~rc appearedon the skin and they becamedehydrated from diarrhc><.aar>cf vomiting.They alsolost their appetite.In somecases there was rr<»+c-~I>r weaknessas their liinb muscleswere unable to support their bua due to frequentdefecation. The feces was black due to internalhaemorrf>aging ir> the largeintenstine and rectum. I'ersonalcominunication with Chief Vet< ri- naryOfficer, Dr. N.C. Raninga!. Byfar the most lamentable and regrettable incident to occurin the whc>lo episodeof thisincidence of fishpoisoning was the attitude of oneof the ha~ driverstransporting the victims to theGeorgetown I lospital. On his arri' al at SandyBay the driver stopped the bus, ordered all passengersto his< rr>»rk. saidthe fare was one dollar $EC!per person and enquired of their a combinedand unrelenting assault of vomiting,diarrhoea and ten>per>t~rc reversalsin the dark of' night nowhere near to any toilet facilities, sorr><- weakthey had to belifted to andfrom the bus, some not knowing~c>w awaythey were from the end of their lives, being asked topay tor wf>> "! c>u>I cl havebeen their lastbus ride. Manyhad no moneywith them.All pa>se~g~ ti> from this bus were therefore transferred to a bus based at Sandy completethe journey to the GeorgetownHospital. Thatthirty pounds offish could be shared among some one huncl«~ cl ar>d twentyindividuals isan indication ofsome of the socio-ecomomic pr»l ~rr>~c>f thearea. If thefish were marketed in Kingstown for example, where f~~a rr

199 K. Morris sitting. Restaurantsin Kingstownserve a middayfish meal $EC9,00! consist- ing of a sliceof stewedfish of about4 ozs,in weightand barracuda is not excludedfrom the menu. Assuminga directrelationship betweenseverity of symptomsand quantity of fishconsumed, with theconcentration of the toxin beingconstant, one can easilyconclude that had the incidenttaken place in Kingstownthe data coHected would have been much different.

Notes ' A villageon theextreme north east coast of St.Vincent and the Grenadines, population985! 1,035,507 females and 528 males. There were 22 births and 7 deaths recorded for the year. Mr. Kerwin Morris, Chief Fisheries Officer of St, Vincent and the Grenadines, attendedthe Conferencerepresenting Mr. DavonJoseph, Development Offi- cer,Organization of EasternCaribbean States, Fisheries Unit, Kingstown, St. Vincent and the Grenadines, Mr, Morris submitted this report summarizing theevents surrounding an incidenceof ciguaterafish poisoning in St. Vincent and the Grenadines. LegalAspects of CiguateraFish Poisoning in Puerto Rico

J.V. Biaggi Law Offices,Biaggi Busquets k lVfariRocca, P,O. Box ! 589, Mayagiiez,Puerto Rico, Ot81

INTRODUCTION In PuertoRico for many yearsany person who becameintoxicated with ciguateraby eating fish in a restaurantwas certain of obtaining compensation for damages,The amountof compensationwas dependent on thesevereness of the intoxication, duration of symptomsand hospitalization.All the p!aintiff had to prove was thathe or shehad eaten fish in the defendant'srestaurant, and that asa consequencedeveloped the symptoms associated with ciguatera fish poisoning. In orderto receivecompensation in PuertoRico's system of law there are three basicelements that the plaintiff must prove: 1! A wrongful or negligentact or omission; 2! That the plaintiff suffereddamages; 3! A causal relationbetween the damages and the wrongful or negligent act or omission. TheSupreme Court of theCommonwealth of Puerto Rico has applied the anglo-saxontheory of "impliedwarranty" or "absoluteresponsibility" in casesof foodpoisonings which involved establishments that manufactured food for humanconsumption on the precedentsestablished in the casesof Castrovs. Payco, Inc., 75 DPR63 953!, and Mendoza vs. Cerveceria Corona, Inc., 97 DPR499969!, This meantthat the plaintiff did not haveto provea negligentact or omision on the part of thedefendant inorder to win his or her case. Thesystem of law in PuertoRico is ratherdifferent from that in the continentalUnited States, particularly in tort actions.In thecommon law systemused in theUnited States, if a plaintiffclaims damages fora tortaction andthe defendant can prove that the plaintiff contributed in any way to the causeof hisor herdamages, the plaintiff is barredfrom receiving damages. Thebasis of thisjudgement is that the plaintiff showed "contributory negli- gence." TheSupreme Court of PuertoRico had ruled that the lower courts were boundto judge tort actions inthe two cases mentioned above under the anglo- saxondoctrine of"implied warranty", however, thecourts could not use the doctrineof "contributory negligence" to dismiss these cases. The Supreme J. V. Biaggi

Court directed the lower courts to use the doctrine of "comparative negli- gence."Under this doctrine the plaintiff received compensation, however, the percentageof negligencewhich he or sheincurred was deducted from thetotal compensationawarded. Distinct from the Commonwealth of Puerto Rico, under federal or state laws,if thedefendant could prove that at thetime ot eatingthe culprit fish the plaintiff knew that by eatingfish he or shecould becomeintoxicated with ciguatera,this constituted an assumption of risk, therefore"contributory neg- ligence"was evident and the casewould be dismissed,ln the U.S.Virgin Islandswhere a vastmajority ot the peoplehave had or knewof encounters with ciguatera,most of the legalcases filed in stateor federalcourts were dismissed see Bronson vs. Club Comanche,Inc. 2S6 F. Supp, 21, 1968!. However, in Puerto Rico even if the defendantcould prove "contributory negligence",al! hecould expect would bea reductionin thecompensation he would have to pay under the doctrine of "comparativenegligence." As a result,most if not all legalcases claiming intoxicationby ciguateracaused by fish consumption in public establishmentsbrought to court in Puerto Rico weresettled out of court. Thiswas a rather unjust and unfair situation.

RECENT LEGAL PROGRESS At the time this paperwas presentedat the Third InternationalConfer- enceon Ciguatera Fish Poisoningin May of 1990,this was the legal status in Puerto Rico of casesinvolving ciguatera fish poisoning, In this presentationI indicated that there were two test casesbefore the courts that could changethe legalbases on which ciguaterafish poisoning cases were being judged. Sucha changehas in fact occurred. In the first of these two test casesthe defendant won by means of a Summary Sentence. This means that the case before the court presented no controversy of factand that the court had only to resohe a matterof law. The defense accepted the fact that the plaintiff consumed fish at the defendant's restaurant,and that as a consequencethe plaintiff developedsymptoms asso- ciated with ciguateraintoxication and suffered damages. All the court had to decidewas whetherthe "impliedwarranty" or "absoluteresponsibility" doc- trine applied in casesinvolving ciguatera fish poisoning. Someot the many scientific and legal argumentswhich were used to convincethe court that this doctrine could not apply in ciguateracases were: 1. Ciguatera fish poisoning is a unique and exceptionalkind of intoxica- tion where human hands mishandling of food! play no part. The neurotoxin thataffects humans that consume ciguatoxic fish is accumulatedin the fish by meansof a naturalprocess, through the food chain; 2. Theother cases resolved by theSupreme Court Castrovs. Payco Inc., 1953, and Mendoza vs. Cerveceria Corona, Inc., 1969, cited above! dealt with

202 Ciguatera in Puerto Rico humanillness caused by theconsumption of manufactured foodproducts. Fishesare not manufacturedfood products; 3. Thereis no way at this time to prevent the damages that a personwho eatstoxic fish might suffer; 4. Thereis no possibleway in which the defendantscould have been negligent,since human hands mishandling of food! play no part in ciguatera fish poisoning; 5. lf therewas anyone responsible for thedamages caused by ciguatera fish poisoning, it was the GOOD LORD who createdthe oceansand all the organismsthat thrivein it. Asa resultwe would have had to bringHIM into the caseas a thirdparty defendant, however, it wouldbe impossibleto find anyonewho couldor would deliver the subpoena. Thedefendant won the secondtest case in trial. Theplaintiff asked the SupremeCourt to reversethe decisionof the SuperiorCourt, which had judged in favor of the defendant, the restaurantthat served the fish from which the plaintiff contracted ciguatera. The SupremeCourt denied the request and confirmed the decisionof the lower court Mendez Corradavs. Ladi's Place, 90 JTS 125, 1990!. In deciding this issuethe SupremeCourt found that the evidencebefore the court incontrovertibly showed that: 1. Ciguatera is a seriousproblem around the world, speciallyin tropical areas. This intoxication is distinguishableFrom other typesof fish poisonings whichare causedby the growth of bacterialcontaminants due to inadequate handling and processingpractices; Scientificinvestigations indicate that ciguatera is causedby smallmicro- organisms dinoflagellates! which are eaten by smallfishes, which in turn are eatenby largerpredators. The toxin is transmittedthrough the food chain to thelarger fishes which are consumed by humanbeings; 2. The toxin or toxins are not destroyedby conventionalmeans of cooking suchas: frying, boiling or broiling,Neither are they eliminated by conven- tionalmeans of processingand handling such as: drying, salting, freezing or smoking; 3. Thetoxin is not detectableby theappearance, odor or tasteof the fish. At theprese~t time there is noadequate method of detectingthe toxin. The originand identity of thetoxin or toxinsinvolved in ciguaterafish poisoning, theorganisms that generate and transmit it, aswell as the ecological factors involvedare not completelyknown or understood; 4. Morethan four hundred 00! speciesof fishare implicated in thistype ofpoisoning. The outbreaks ofciguatera aresporadic andunpredictable, both in timeand geographic distribution. Not all members of the same species capturedin thesame place, at thesame time are toxic; 5. Medicaltreatment issymptomatic and there is noknown antidote;

203 J. V, Biaggi

6. Thereis no known practicalmethod to detecta toxicfish when it is captured,nor before or afterits commercialdistribution. Oneknows that the fishis toxiconly after having eaten it. Theserealities, said the Supreme Court, can only support one conclusion: "Thereisno way to prevent the damage that a consumermight suffer. The onlyway to prevent the damages would be total abstinence.The doctrine of 'absoluteresponsibility' doesnot apply in casesinvolving ciguatera fish poi- soning,this doctrine developed only to protectthe consumerfrom careless manufacturers." n summary,ciguatera cases represent fortuitous events which do not generateresponsibility. As long as the product fish!is well handledand preserved,fishermen, wholesalers, distributors, restaurant owners and all personsinvolved in thefishing industry can rest assured that they will not be heldlegally responsible iftheir product happens tohave eaten the wrong kind of dinoflageflate.