1696

Journal of Food Protection, Vol. 68, No. 8, 2005, Pages 1696±1701 Copyright ᮊ, International Association for Food Protection

Identi®cation of Tetrodotoxin in a Marine Gastropod ( glans) Responsible for Human Morbidity and Mortality in Taiwan

PAI-AN HWANG,1 YUNG-HSIANG TSAI,2 JOU-FANG DENG,3 CHAO-AN CHENG,4 PING-HO HO,5 AND DENG-FWU HWANG1*

1Department of Food Science, National Taiwan Ocean University, Keelung, Taiwan; 2Department of Food Science and Technology, Tajen Institute of Technology, Pingtung, Taiwan; 3Clinical Toxicology, Department of Medicine, Veterans Division of General Hospital, Taipei, Taiwan; 4Department of Food Technology, National Kinmen Institute of Technology, Kinmen, Taiwan; and 5National Museum of Marine Biology and Aquarium, Pingtung, Taiwan Downloaded from http://meridian.allenpress.com/jfp/article-pdf/68/8/1696/1680424/0362-028x-68_8_1696.pdf by guest on 02 October 2021

MS 04-515: Received 15 November 2004/Accepted 28 March 2005

ABSTRACT

The toxicity of the gastropod Nassarius glans was investigated. This gastropod was implicated in an incident of food paralytic poisoning on Tungsa Island, Taiwan, in April 2004. Six victims consumed both digestive glands and muscle. These tissues contained high concentrations of toxin; their highest toxicity scores were 2,048 and 2,992 MU/g, respectively, based on the tetrodotoxin (TTX) bioassay. The toxin was puri®ed from these gastropods and analyzed by high-performance liquid chromatography, which revealed TTX and related compounds 4-epi TTX and anhydro-TTX; paralytic shell®sh poisons were not found. The urine and blood samples from patients were cleansed using a C18 Sep-Pak cartridge column and 3,000 molecular weight cutoff Ultrafree microcentrifuge ®lters, and the eluate was ®ltered and analyzed by liquid chromatography and mass spectrometry. The detection limit for TTX was 1 ng/ml. The standard curves were linear in the range 30 to 600 ng/ml for urine and 1 to 30 ng/ml for blood. TTX was detected in all urine samples but in only three of four blood samples tested. Thus, the causative agent of gastropod food poisoning was identi®ed as TTX.

During April 2004, six people (from 22 to 48 years Zeuxius samiplicutus (8, 32), and several in the old) presented with symptoms of toxicity (paralysis, coma, family Olividae such as Oliva miniacea, Oliva mustelina, nausea, vomiting, ataxia, and aphasia) after they ate entire and Oliva hirasei (17). All these gastropods contain only specimens of the gastropod Nassarius glans () TTX and/or its related components, except for members of on Tungsa Island, Taiwan. Unfortunately, two of the six the Naticidae and Nassariidae collected from South Tai- individuals died 30 min after ingestion because of intrac- wan, which contain TTX and paralytic shell®sh poison table bradycardia and multiple organ failure; the other four (PSP) (2, 7, 8). people recovered after medical intervention. The symptoms TTX has been recognized as a causative agent of food exhibited by these individuals were similar to those of te- poisoning for more than 30 years. Symptoms appear within trodotoxin (TTX) poisoning. 10 to 45 min of exposure and are usually classi®ed into The TTX and related toxins are some of the most po- four main levels of severity, depending on the amount of tent marine toxins and are known to inhibit cell function toxin ingested (25). In Asia, gastropods are widely used as by blocking voltage-dependent sodium channels (22). TTX a traditional food, and poisoning incidents associated with has been widely used as an important biochemical tool in nassariid gastropods have been reported occasionally in Tai- neurophysiological studies (6). TTX was originally isolated wan and Mainland China since 1994 (8, 16, 31). The re- from puffer ®sh and has since been detected in a wide range sponsible toxins were determined to be TTX and PSP. We of including amphibians, ®sh, echinoderms, arthro- examined the toxicity of the gastropod N. glans and iden- pods, mollusks, nemerteans, and platyhelminths (20). TTX ti®ed the responsible toxin. and/or related compounds have been reported in many gas- tropod mollusks, including the trumpet shell Charonia sau- MATERIALS AND METHODS liae (24), the frog shell Tutufa lissostoma (26), the Japanese Materials. Urine (70 ml) and blood (30 ml) samples were ivory shell Babylonia japonica (27), the rock shells Rapana collected from patients when they arrived at the hospital (about rapiformis and Rapana venosa venosa (10), several species 15 h after ingestion of toxic gastropods). Twenty specimens of N. in the family Naticidae such as the lined moon shell Natica glans from the coastal area of Tungsa Island, Taiwan, were pro- lineata, banded moon shell Natica vitellus, and bladder vided by patients (Fig. 1). The samples were frozen at Ϫ20ЊC moon shell Polinices didyma (9, 11, 13), several species in until the assay was carried out. The edible part from 20 specimens the family Nassariidae such as the basket shells Niotha was removed from the shells and separated into digestive gland clathrata, Zeuxius scalaris, Zeuxius siquijorensis, and and muscle (including salivary gland, brain, and mouth organs). The ICR mouse strain from the Institute of Cancer Research was * Author for correspondence. Tel: ϩ886-2-24622192, Ext 5103; Fax: purchased from National Laboratory Breeding and Re- ϩ886-2-24626602; E-mail:[email protected]. search Center (Taipei, Republic of China). Healthy mice weighing J. Food Prot., Vol. 68, No. 8 TETRODOTOXIN IN A MARINE GASTROPOD IN TAIWAN 1697 Downloaded from http://meridian.allenpress.com/jfp/article-pdf/68/8/1696/1680424/0362-028x-68_8_1696.pdf by guest on 02 October 2021

FIGURE 1. Location of poisoning outbreak and photograph of FIGURE 2. HPLC of Nassarius glans toxins, authentic TTX, 4-epi Nassarius glans. TTX, and anh-TTX. between 18 and 20 g were used. Authentic TTX, 4-epi TTX, and Assay of toxicity. The dissected tissue was weighed, ho- anhydrotetrodotoxin (anh-TTX), which were obtained from the mogenized with 10 volumes of 1% acetic acid in methanol for 5 liver of the puffer Takifugu oblongus, were used as the reference min, and centrifuged (2,000 ϫ g for 20 min). This sequence was standards (4, 14). Authentic gonyautoxins 1 through 4 and saxi- then repeated. The supernatants from both processes were com- toxins (STXs) obtained from the purple clam Soletellina diphos bined, concentrated under reduced pressure at 45ЊC, and examined (15) and the xanthid crab Zosimus aeneus (3), respectively, was for toxicity by the mouse assay for TTX (12, 35). Toxicity was also used as reference standards. expressed in mouse units (MU). One mouse unit is de®ned as the

TABLE 1. Toxicity of Nassarius glans specimens collected from Tungsa Island Digestive gland Muscle Body Body Total toxicity Specimen no. weight (g) length (cm) Weight (g) Toxicity (MU/g) Weight (g) Toxicity (MU/g) (MU/specimen)

1 9.13 4.47 1.03 255 3.18 926 3,208 2 6.37 4.20 0.64 317 2.30 547 1,461 3 10.04 4.61 1.19 371 3.89 2,992 12,079 4 11.90 4.57 2.23 185 4.06 1,531 3,628 5 8.76 4.53 1.25 261 2.63 1,396 3,941 6 9.54 4.27 1.06 2,040 3.66 1,156 6,393 7 8.31 4.24 1.34 426 3.10 2,248 7,539 8 8.80 4.41 0.66 411 3.62 732 2,896 9 9.14 4.24 1.00 240 3.19 1,424 4,782 10 8.67 4.26 1.02 457 3.33 1,640 5,927 11 9.30 4.36 0.96 2,048 3.45 1,840 8,314 12 8.81 4.13 0.69 98 3.79 496 1,948 13 10.68 4.46 1.70 1,680 3.73 636 5,228 14 9.11 4.45 1.25 231 3.88 914 3,834 15 11.80 4.49 1.51 196 5.17 472 3,744 16 9.13 4.25 0.94 312 4.06 880 3,866 17 12.05 4.64 1.58 174 5.01 1,368 7,129 18 9.10 4.30 1.00 219 3.49 1,000 3,709 19 9.46 4.57 1.72 402 3.65 2,120 7,545 20 8.85 4.18 1.15 452 3.61 1,472 5,833 Mean Ϯ SD 9.45 Ϯ 1.33 4.38 Ϯ 0.16 1.20 Ϯ 0.39 538 Ϯ 608 3.92 Ϯ 0.61 1,167 Ϯ 557 5,188 Ϯ 1,959 1698 HWANG ET AL. J. Food Prot., Vol. 68, No. 8 Downloaded from http://meridian.allenpress.com/jfp/article-pdf/68/8/1696/1680424/0362-028x-68_8_1696.pdf by guest on 02 October 2021

FIGURE 3. Standard curve for TTX in urine as detected by LC- FIGURE 4. Standard curve for TTX in blood as detected by LC- MS. MS.

amount of toxin required to kill a 20-g male ICR mouse in 30 A (1.8 N KOH:2.5 M ammonium formate:formanide, 1:4:5) and min after a single intraperitoneal injection (12), in this case 0.178 solution B (0.08 M periodic acid solution). The ¯uorogenic re- ␮ g of TTX. action was performed at 65ЊC for 0.7 min. The intensity of the Puri®cation of toxin from gastropods. After bioassay, the ¯uorescence was measured at 388 nm excitation (21). extract of muscle and digestive glands of 20 specimens was Instrumentation: LC-MS. For this analysis, combined LC- mixed. The extract (110,000 MU) was defatted with dichloro- MS was performed using a model 1100 series LC/MSD trap sys- methane. The aqueous layer was concentrated and ®ltered through tem (Agilent, Palo Alto, Calif.) coupled to a mass spectrometer a Dia¯o YM-1 membrane (Amicon, Beverly, Mass.) to remove with a positive ion electrospray ionization interface. The drying substance larger than 1,000 Da. The ®ltrate was applied to a Bio- Њ gas temperature was set at 350 C with N2, the fragmentor was set Gel P-2 column (2 by 94 cm; Bio-Rad, Hercules, Calif.) and elut- at 150 V, the nebulizer pressure was set at 60 psi, and the Vcap ed with 0.03 M acetic acid (9, 14). Toxic fractions were combined, was set at 3,500 V. The HPLC system was equipped with a Zorbax freeze dried, dissolved in a small amount of water, and analyzed 300SB-C3 column (4.6 mm inside diameter by 150 mm) at room by high-performance liquid chromatography (HPLC) and liquid temperature. The injection volume was 5 ␮l. The mobile phase chromatography with mass spectrometry (LC-MS). for TTX analysis was 1% acetonitrile, 1 mM trimethylamine, 10 Extraction of toxin from urine and blood. Urine (50 ml) mM ammonium formate (pH 4.0 and ¯ow rate of 0.4 ml/min) and blood (20 ml) samples from each of the four surviving pa- (34). tients were thawed and mixed with 3 ml of 0.5 M acetic acid and Normal serum and urine were analyzed to control for inter- then centrifuged at 10,000 ϫ g for 10 min. The supernatant was ference from endogenous substances. Readings consisting of three freeze dried and quantitatively diluted to 5 ml with deionized wa- replicates were subjected to linear regression analysis to generate ter. Samples were loaded onto the cartridge column (C18 Sep-Pak a standard curve. cartridges, Millipore, Waters, Mass.) that had been regenerated RESULTS with 10 ml of methanol and equilibrated with 10 ml of water. Collection of toxin began immediately, and the toxin was eluted Toxicity and toxins of gastropods. Twenty specimens with 10 ml of methanol, resulting in the pigments being retained of N. glans were analyzed for toxicity using the TTX bio- in the column. The eluates were freeze dried, dissolved in 2 ml assay. All of the specimens were determined to be toxic. of 0.3% acetic acid, and ®ltered with a 3,000 molecular weight The mean (ϮSD) toxicity was 5,188 Ϯ 1,959 MU per spec- cutoff Ultrafree microcentrifuge ®lter (Micron YM-3, Millpore). imen (Table 1). The gastropods contained high amounts of The ®ltrate was freeze dried, dissolved in a small amount (1 ml) of water, and subjected to LC-MS analysis. TABLE 2. TTX concentration in the urine and blood of poisoning Instrumentation: HPLC. HPLC (L-2100, Hitachi Ltd., To- victims kyo, Japan) was performed on an ODS reversed-phase column (YMC-pack ODS-AQ, YMC, 6.0 mm inside diameter by 300 mm) TTX (ng/ml) by LC-MS with a ¯uorescence detector (F-1000, Hitachi Ltd., Tokyo, Japan). Patient no. Sex/age (yr) Urine Blood The mobile phase for TTX and gonyautoxin analysis was sodium 1-heptane sulfonate (2 mM) in methanol (1%)±potassium phos- 1 M/44 169 Ͻ1 phate buffer (0.05 M, pH 7.0). For STX analysis, the mobile phase 2 M/35 201 5 was 1-heptane sulfonate (2 mM) in methanol (20%)±potassium 3 M/34 274 3 phosphate buffer (0.05 M, pH 7.0). The TTX was detected by 4 M/22 325 8 mixing the eluate with 3 N NaOH at a ratio of 1:1 followed by 5a M/48 Њ heating at 99 C for 0.4 min, and the ¯uorescence was monitored 6a M/33 at 505 nm with 381 nm excitation. For gonyautoxin and STX analyses, the eluates were mixed with an equal volume of perio- a Patients 5 and 6 died, and samples of urine and blood were not date reagent, which was composed of equal volumes of solution collected. J. Food Prot., Vol. 68, No. 8 TETRODOTOXIN IN A MARINE GASTROPOD IN TAIWAN 1699

FIGURE 5. LC-MS chromatograms of au- thentic TTX in urine (A), blood (B), urine sample (C), and blood sample (D) of pa- tient 4. The retention time for TTX was 4.18 min. Downloaded from http://meridian.allenpress.com/jfp/article-pdf/68/8/1696/1680424/0362-028x-68_8_1696.pdf by guest on 02 October 2021

toxin in muscle (1,167 Ϯ 557 MU/g) and digestive gland retention time (4.18 min) (Fig. 5). The LC-MS mass spec- (538 Ϯ 608 MU/g). The highest toxicity scores were 2,992 trum of the urine and blood samples of patient 4 revealed MU/g for muscle, 2,048 MU/g for digestive gland, and authentic TTX as an ion peak at m/z ϭ 320.1 at the same 12,079 MU total for an individual specimen. Toxicity of retention time (4.18 min) (Fig. 6). These data indicated that individual specimens was variable, ranging from 1,461 to the causative agent of this food poisoning incident was 12,079 MU. After puri®cation, 58.9 mg of the toxin was TTX. obtained at a toxicity of 1,860 MU/mg. The examination DISCUSSION of extracts from the gastropods by HPLC revealed three major peaks (Fig. 2), which had the same retention times In this study, we ®rst determined the toxicity of the N. (12.9, 14.6, and 17.0 min) as those of TTX, 4-epi TTX, glans specimens that caused the serious cases of food poi- and anh-TTX, respectively. soning. The gastropods contained high amounts of TTX in their digestive glands and muscle. The toxicity of N. glans Toxicity and toxins of urine and blood. Areas of the is higher than that of Z. samiplicutus, Zeuxius suf¯atus, N. TTX peak from urine were linear between 30 and 600 ng/ clathrata, and Z. scalaris, which have caused food poison- ml (Fig. 3) and consistent with the TTX peak in blood ing episodes in Taiwan (8, 16, 32). The minimum lethal between 1 and 30 ng/ml (Fig. 4). The TTX detection limit dose of TTX for a human by oral administration is esti- was 1 ng/ml, and the coef®cient of variation was less than mated to be 10,000 MU (33). The average toxicity of in- 3.0%. TTX concentrations in the urine and blood samples dividual N. glans specimens in this study was 5,188 Ϯ of patients 1 through 4 are listed in Table 2. TTX was 1,959 MU; thus, patients may have died because they con- detected in the urine samples of all four patients, but TTX sumed more than two specimens. Some members of the concentration was below the detection limit in the blood family Nassariidae, including N. clathrata, Z. samiplicutus, sample of patient 1. The concentration of TTX in the urine Z. suf¯atus, Z. siquijorensis, Z. scalaris, and Z. scalaris- and blood ranged from 169 to 325 ng/ml and from Ͻ1to like species, have been reported to contain TTX and/or PSP 8 ng/ml, respectively (Table 2). The LC-MS chromatograms (8, 16, 18, 23, 32). These gastropods are carnivorous and of the urine and blood samples of patient 4 revealed au- prefer to inhabit coastal waters, so they might accumulate thentic TTX as an ion peak at m/z ϭ 320.1 at the expected TTX and/or PSP through the food chain (30). 1700 HWANG ET AL. J. Food Prot., Vol. 68, No. 8

FIGURE 6. LC-MS mass spectra of au- thentic TTX in the urine (A), blood (B), urine sample (C), and blood sample (D) of patient 4. The retention time for TTX was 4.18 min, and for the protonated molecular ion [M ϩ H]ϩ of TTX, m/z ϭ 320.1. Downloaded from http://meridian.allenpress.com/jfp/article-pdf/68/8/1696/1680424/0362-028x-68_8_1696.pdf by guest on 02 October 2021

Hines and coworkers (5) used [3H] saxitoxinol to study soning cases usually have been identi®ed based on the pres- the STX distribution in whole animals. They found that STX ence of the toxin in the suspect food. However, examination analogue was eliminated mainly by urine; none was found of urine and blood for TTX by LC-MS also is important in feces. We also found that the concentration of TTX in the because the results can con®rm the toxin identity. urine was signi®cantly higher than that in the blood, indi- Recently, immunoaf®nity chromatography has been cating that TTX in serum appears to be rapidly eliminated used as a speci®c puri®cation method prior to HPLC anal- in the urine (29). The concentration of TTX in the blood was ysis of urine for TTX (19), but monoclonal antibodies lower than that in the urine. We speculated that most of toxin against TTX are dif®cult to obtain and very expensive. The had already been eliminated through urine by the time the combination of a C18 Sep-Pak cartridge column, ultra®l- patients arrived at the hospital 15 h after ingesting the gas- tration with a microcentrifuge ®lter, and LC-MS is very tropods. However, Oda et al. (28) demonstrated that TTX useful for detecting TTX in urine and blood samples of elimination from the body took about 5 days. Thus, it is patients for diagnosis of TTX-associated food poisoning. important to collect urine and blood from patients as soon This combined method also has been reported to be useful as possible after they become ill. TTX-associated food poi- for detecting PSP in cats (1). J. Food Prot., Vol. 68, No. 8 TETRODOTOXIN IN A MARINE GASTROPOD IN TAIWAN 1701

ACKNOWLEDGMENT 18. Jeon, J. K., H. Narita, M. Nara, T. Noguchi, J. Maruyama, and K. Hashimoto. 1984. Occurrence of tetrodotoxin in a gastropod mol- This study was partly supported by the National Science Council lusk, ``araregai'' Niotha clathrata. Nippon Suisan Gakkaishi 50: (grant no. NSC-93-2120-M-019-001), Republic of China. 2099±2102. 19. Kawatsu, K., T. Shibata, and Y. Hamano. 1999. Application of im- REFERENCES munoaf®nity chromatography for detection of tetrodotoxin from urine samples of poisoned patients. Toxicon 37:325±333. 1. Andrinolo, D., L. F. Michea, and N. Lagos. 1999. Toxic effects, 20. Miyazawa, K., and T. Noguchi. 2001. Distribution and origin of te- pharmacokinetics and clearance of saxitoxin, a component of para- trodotoxin. J. Toxicol. Toxin Rev. 20:11±33. lytic shell®sh poison (PSP), in cats. Toxicon 37:447±464. 21. Nagashima, T., J. Maruyama, T. Noguchi, and K. Hashimoto. 1987. 2. Cheng, C. A., S. J. Lin, and D. F. Hwang. 1996. Paralytic toxins of Analysis of paralytic shell®sh poison and tetrodotoxin by ion-pairing the gastropod Natica lineata in Pingtung Prefecture. Food Sci. 23: high performance liquid chromatography. Nippon Suisan Gakkaishi 845±853. 53:819±823. 3. Daigo, K., A. Uzu, O. Arakawa, T. Noguchi, H. Seto, and K. Hash- 22. Narahashi, T., J. M. Moore, and J. Scott. 1964. Tetrodotoxin block- imoto. 1985. Isolation and some properties of neosaxitoxin from a age of sodium conductance increase in lobster giant axons. Gen.

xanthid crab Zosimus aeneus. Nippon Suisan Gakkaishi 51:309±313. Physiol. 47:965±974. Downloaded from http://meridian.allenpress.com/jfp/article-pdf/68/8/1696/1680424/0362-028x-68_8_1696.pdf by guest on 02 October 2021 4. Goto, T., Y. Hishi, S. Takahashi, and Y. Hirata. 1965. Tetrodotoxin. 23. Narita, H., T. Noguchi, J. Maruyama, M. Nara, and K. Hashimoto. Tetrahedron 21:2059±2088. 1984. Occurrence of a tetrodotoxin-associated substance in a gastro- 3 5. Hines, H. B., S. M. Naseem, and R. W. Wannemacher, Jr. 1993. H- pod, hanamushirogai, Zeuxius siquijorensis. Nippon Suisan Gakka- saxitoxinol metabolism and elimination in the rat. Toxicon 31:905± ishi 50:85±88. 908. 24. Narita, H., T. Noguchi, J. Maruyama, Y. Ueda, K. Hashimoto, Y. 6. Hucho, F. 1995. Toxins as tools in neurochemistry. Angew. Chem. Watanabe, and K. Hida. 1981. Occurrence of tetrodotoxin in a trum- Int. Ed. Engl. 34:39±50. pet shell ``boshubora'' Charonia sauliae. Bull. Jpn. Soc. Sci. Fish. 7. Hwang, D. F., C. A. Cheng, and S. S. Jeng. 1994. Gonyautoxin 3 as 47:935±941. minor toxin in the gastropod Niotha clathrata in Taiwan. Toxicon 25. Noguchi, T., and J. S. M. Ebesu. 2001. Puffer poisoning: epidemi- 32:1573±1579. ology and treatment. J. Toxicol. Toxin Rev. 20:1±10. 8. Hwang, D. F., C. A. Cheng, H. T. Tsai, D. Y. Shih, H. C. Ko, R. Z. 26. Noguchi, T., J. Maruyama, H. Narita, and K. Hashimoto. 1984. Oc- Yang, and S. S. Jeng. 1995. Identi®cation of tetrodotoxin and para- currence of tetrodotoxin in the gastropod mollusk Tutufa lissostoma lytic shell®sh toxins in marine gastropods implicated in food poi- (frog shell). Toxicon 22:219±226. soning. Fish. Sci. 61:675±679. 27. Noguchi, T., J. Maruyama, Y. Ueda, K. Hashimoto, and T. Harada. 9. Hwang, D. F., C. H. Chueh, and S. S. Jeng. 1990. Occurrence of 1981. Occurrence of tetrodotoxin in the Japanese ivory shell Baby- tetrodotoxin in the gastropod mollusk Natica lineata (lined moon lonia japonica. Nippon Suisan Gakkaishi 47:909±913. shell). Toxicon 28:21±27. 28. Oda, K., K. Araki, T. Totoki, and H. Shibasaki. 1989. Nerve con- 10. Hwang, D. F., C. H. Chueh, and S. S. Jeng. 1991. Occurrence of duction study of human tetrodotoxication. Neurology 39:743±745. tetrodotoxin in the gastropods Rapana rapiformis and R. venosa ven- 29. O'Leary, M. A., J. J. Schneider, and G. K. Isbister. 2004. Use of osa. Mar. Biol. 111:65±69. 11. Hwang, D. F., C. H. Chueh, K. P. Tai, and S. S. Jeng. 1991. Tetro- high performance liquid chromatography to measure tetrodotoxin in dotoxin and derivatives in several species of the gastropod Naticidae. serum and urine of poisoned patients. Toxicon 44:549±553. Toxicon 29:1019±1024. 30. Shiomi, K., H. Tamanaka, and T. Kikuchi. 1984. Accumulation of 12. Hwang, D. F., and S. S. Jeng. 1991. Bioassay of tetrodotoxin using tetrodotoxin by marine gastropod. Nippon Suisan Gakkaishi 50: ICR mouse strain. J. Chin. Biochem. Soc. 20:80±86. 1269±1273. 13. Hwang, D. F., L. C. Lin, and S. S. Jeng. 1992. Occurrence of tetro- 31. Shui, L. M., K. Chen, J. Y. Wang, H. Z. Mei, A. Z. Wang, Y. H. dotoxin related toxins in the gastropod mollusk Niotha clathrata Lu, and D. F. Hwang. 2003. Tetrodotoxin-associated snail poisoning from Taiwan. Nippon Suisan Gakkaishi 55:63±67. in Zhoushan: a 25-year retrospective analysis. J. Food Prot. 66:110± 14. Hwang, D. F., T. Noguchi, O. Arakawa, T. Abe, and K. Hashimoto. 114. 1988. Toxicological studies on several species of puffer in Taiwan. 32. Sui, L. M., K. Chen, P. A. Hwang, and D. F. Hwang. 2002. Identi- Nippon Suisan Gakkaishi 54:2001±2006. ®cation of tetrodotoxin in marine gastropods implicated in food poi- 15. Hwang, D. F., T. Noguchi, Y. Nagashima, I. C. Liao, and K. Hash- soning. J. Nat. Toxins 11:213±220. imoto. 1987. Occurrence of paralytic shell®sh poison in the purple 33. Tani, I. 1945. Study of the toxicity of Japanese fugu. Teikoku Tosh- clam Soletellina diphos (bivalve). Nippon Suisan Gakkaishi 53:623± oun Co., Tokyo. 626. 34. Tanu, M. B., and T. Noguchi. 1999. Tetrodotoxin as a toxic principle 16. Hwang, D. F., Y. C. Shiu, P. A. Hwang, and Y. H. Lu. 2002. Tetro- in the horseshoe crab Carcinoscorpius rotundicauda collected from dotoxin in gastropods (snails) implicated in food poisoning in north- Bangladesh. J. Food Hyg. Soc. Jpn. 40:426±430. ern Taiwan. J. Food Prot. 65:1341±1344. 35. Yasumoto, T. 1991. Tetrodotoxin, p. 296±300. In Environmental 17. Hwang, P. A., Y. H. Tsai, Y. H. Lu, and D. F. Hwang. 2003. Paralytic Health Bureau, Ministry of Health and Welfare (ed.), Standard meth- toxins in three new gastropod (Olividae) species implicated in food ods of analysis in food safety regulationÐchemistry. Japan Food poisoning in southern Taiwan. Toxicon 41:529±533. Hygiene Association, Tokyo.