Toxins of Pufferfish—Distribution, Accumulation Mechanism, and Physiologic Functions

Home , Atergatis floridus, Kelletia lischkei, Zosimus aeneus

Aqua-BioScience Monographs, Vol. 10, No. 3, pp. 41–80 (2017) www.terrapub.co.jp/onlinemonographs/absm/

Osamu Arakawa1*, Tomohiro Takatani1, Shigeto Taniyama1 and Ryohei Tatsuno2

1Graduate School of Fisheries and Environmental Sciences Nagasaki University 1-14 Bunkyo-machi, Nagasaki, Nagasaki 852-8521, Japan 2Department of Food Science and Technology National Fisheries University, Japan Fisheries Research and Education Agency 2-7-1 Nagatahonmachi, Shimonoseki, Yamaguchi 759-6595, Japan *e-mail: [email protected]

Abstract Received on April 30, 2015 Our many years of studies have provided a lot of information regarding distribution, Accepted on accumulation mechanism, and physiologic functions of the natural toxins harbored by September 12, 2016 Online published on pufferfish. We detected tetrodotoxin (TTX) and/or its derivatives in several species of October 17, 2017 aquatic organisms including marine bacteria. This, along with the fact that pufferfish became non-toxic when they were reared with non-toxic feed, indicated that the toxification Keywords of pufferfish is of exogenous origin. Subsequently, we revealed by various TTX adminis- • tetrodotoxoin tration experiments using non-toxic cultured pufferfish that the TTX administered into • paralytic shellfish poison the muscle or digestive tract rapidly transferred to the liver and skin, and that the toxin • saxitoxin transfer to the gonads was largely different between male and female, suggesting the • palytoxin • pufferfish involvement of maturation in the internal kinetics of TTX. On the other hand, we visual- • Takifugu rubripes ized micro-distribution of TTX in the tissues of various TTX-bearing organisms using an • food poisoning immunohistochemical technique, giving a crucial insight to elucidate physiologic functions of TTX including the function as a defensive or offensive agent. Furthermore, we found that Southeast Asian freshwater pufferfish possess paralytic shellfish poison (PSP) as a main toxin, and that boxfish and Bangladeshi freshwater pufferfish bear a palytoxin (PLTX)-like toxin, and cause a rhabdomyolysis, which overturned the common sense that pufferfish toxin equals TTX.

1. Introduction unit) is defined as the amount of toxin required to kill a 20-g male mouse within 30 min after intraperitoneal Among the food poisonings occurring in Japan, poi- administration], and the minimum lethal dose (MLD) soning due to marine toxins, particularly of pufferfish for humans is estimated to be approximately 10000 MU poisoning due to tetrodotoxin (TTX) is the most fre- (ª2 mg) (Noguchi and Ebesu 2001). quent in terms of the fatalities. TTX is a potent neuro- The main symptoms of human intoxication include toxin of low molecular weight, whose unique struc- numbness of lips, tongue and the limbs, paresthesia, ture was determined by three groups in 1964 (Goto et dysarthria, respiratory distress, and death is caused by al. 1965; Tsuda et al. 1964; Woodward 1964). Various respiratory failure in the most critical cases (Noguchi TTX derivatives have so far been separated from and Ebesu 2001). According to Toda et al. (2012), 651 pufferfish, newts, frogs, and/or some other TTX-bear- incidents of TTX poisoning due to pufferfish have oc- ing organisms (Yotsu-Yamashita 2001; Yotsu- curred in Japan, involving 976 patients and 56 deaths Yamashita et al. 2013) (Fig. 1). TTX inhibits the con- during the 22 years from 1989 to 2010. Many cases duction of action potential by selectively plugging so- occurred during the winter season in coastal prefec- dium channels on the nerve/muscle membrane at ex- tures of the Seto Inland Sea, with “komonfugu” tremely low concentrations (Narahashi 2001). The le- Takifugu poecilonotus, “mafugu” Takifugu porphyreus, thal potency is 5000 to 6000 MU/mg [1 MU (mouse “higanfugu” Takifugu pardalis, “shosaifugu” Takifugu

Fig. 1. Structures of TTX and typical TTX derivatives.

snyderi, “kusafugu” Takifugu niphobles, and “torafugu” tom of eating small necrophagous marine snails, and Takifugu rubripes as the causative species. Recently, TTX poisoning due to the snails has frequently oc- the non-edible pufferfish “dokusabafugu” curred (Arakawa et al. 2010; Hwang and Noguchi Lagocephalus lunaris, which usually inhabits tropical 2007; Noguchi et al. 2011a). At least 28 incidents were or subtropical waters, has been frequently mixed up recorded during 1985–2004 in China, and 9 incidents with edible species in Japanese coastal waters, posing during 1994–2006 in Taiwan, involving 233 patients a serious food hygiene problem. This pufferfish, which and 24 deaths in total. As described later, a poisoning bears a very similar appearance to the almost nontoxic due to the similar marine snail “kinshibai” Nassarius species “shirosabafugu” Lagocephalus spadiceus, also glans also occurred in Kyushu, Japan, in 2007 and possesses high levels of TTX in their muscles, caused 2008, respectively (Taniyama et al. 2009a). In Spain, 5 poisoning incidents with 11 patients due to mistaken the European carnivorous trumpet shell Charonia ingestion in Kyushu and Shikoku Islands during 2008– lampas lampus caused human TTX intoxication in the 2009. same year (2007) (Noguchi et al. 2011a; Radriguez et In Taiwan and China, including Hong Kong, although al. 2008). In New Zealand, 15 dogs were suddenly consumption of pufferfish is officially prohibited, and poisoned at the beaches adjacent to Hauraki Gulf in not eaten as frequently as in Japan, there have also been 2009, all exhibiting similar symptoms, and 5 of them many food poisoning cases due to the ingestion of wild died. McNabb et al. (2010) detected a very high level pufferfish (Noguchi and Arakawa 2008). In countries of TTX in the grey side-gilled sea slug outside of East Asia, people generally do not have a Pleurobranchaea maculate found in tide pools near the custom of eating pufferfish, but poisoning due to acci- beach, and claimed that the dogs were poisoned with dental ingestion of pufferfish occasionally occur all TTX by contact with the sea slugs. over the world, including Australia (Isbister et al. On the other hand, pufferfish poisonings due to other 2002), Brazil (Silva et al. 2010), Thailand (Brillantes toxins than TTX have also occurred. In the United et al. 2003), and Bangladesh. Particularly in Bangla- States, Floridian Sphoeroides pufferfish caused 28 desh, 3 large pufferfish poisoning incidents occurred cases of food poisonings during 2002 to 2004, whose in 2008, involving 141 patients and 17 deaths (Islam main toxic principle was subsequently identified not et al. 2011). Recently, toxic Indo-Pacific pufferfish, as TTX, but as paralytic shellfish poison (PSP) including “senninfugu” Lagocephalus sceleratus, mi- (Landsberg et al. 2006). As described later, poisonings grated from the Red Sea via the Suez Canal to the due to PSP-bearing freshwater pufferfish sometimes Mediterranean Sea, and poisoning by this species have occur in Southeast Asian countries (Kungsuwan et al. occurred in the eastern Mediterranean (Bentur et al. 1997; Ngy et al. 2008; Zaman et al. 1997). Since 1990, 2008). a series of poisonings due to ingesting boxfish, a fam- Pufferfish is not the only cause of TTX poisonings. ily of pufferfish, have occurred in western Japan, in- In China and Taiwan, people have a time-honored cus- volving 13 patients and 1 death (Taniyama et al.

2009b). The boxfish poisoning differs from usual pufferfish poisoning, but is very similar to parrotfish poisoning (a unique variety of food poisoning that has sporadically occurred in Japan) in terms of causing rhabdomyolysis as the main symptom (Arakawa et al. 2010). In Bangladesh, a similar poisoning is frequently caused by freshwater pufferfish. With the change in marine environment, poisonings due to pufferfish or pufferfish toxins seem to be diver- sified, and geographically expanded. Although the Japanese have been eating pufferfish since ancient times, and have created a unique food culture for pufferfish, the fundamental understanding not only on the toxin itself, but also on the organisms that possess it is essential to avoid poisonings, and ensure the safety and reliability for eating pufferfish. Over many years, we have accumulated varied information on the characteristics and distribution of pufferfish toxins, and their accumulation mechanism and physiological functions in bearing organisms. The present article sum- marized the results.

2. Distribution of TTX in aquatic organisms and toxification mechanism of pufferfish

2-1. Detection of TTX or its derivatives in several aquatic organisms

TTX was long believed to be present only in pufferfish. Since Mosher et al. (1965) identified a toxin from the eggs of the California newt Taricha torosa as TTX, however, TTX has been detected in a wide variety of animals, such as the goby “tsumugihaze” Yongeichthys criniger, atelopid frogs (Atelopus spp.), the blue-ringed octopus Hapalochlaena maculosa, the carnivorous gastropod “boshubora” Charonia sauliae, starfish of genus Astropecten, the xanthid crab “subesubemanjugani” Atergatis floridus, and flatworms (Hwang and Noguchi 2007; Miyazawa and Noguchi Fig. 2. ESI/MS spectra of TTX (upper), TTX-U1, and TTX- U2 (lower) separated from A. floridus. From Arakawa O, 2001; Noguchi and Arakawa 2008). Following these Noguchi T, Shida Y, Onoue Y. Occurrence of 11- pioneering studies, we have investigated the distribu- oxotetrodotoxin and 11-nortetrodotoxin-6(R)-ol in a xanthid tion of TTX in aquatic organisms, and detected TTX crab Atergatis floridus collected at Kojima, Ishigaki Island. and/or its derivatives in ribbon worms (Ali et al. 1990; Fish. Sci. 1994; 60: 769–771, Fig. 3. Miyazawa et al. 1988; Noguchi et al. 1991), toxic xanthid crabs (Arakawa et al. 1994a), Bangladeshi tree frogs (Tanu et al. 2001), necrophagous marine snails (Taniyama et al. 2009a), and so forth. In addition, al- 1994a), toxicity and toxin profile of the necrophagous most at the same time when Yasumoto et al. (1986) marine snail N. glans (Taniyama et al. 2009a), and found TTX production by marine bacteria, we detected detection of TTX in cultured cells of marine bacteria a TTX producing ability in marine bacteria including (Noguchi and Arakawa 2008). Vibrio that was separated from several species of TTX- bearing organisms (Hwang et al. 1989; Kungsuwan et 2-1A. Separation of TTX derivatives from A. floridus al. 1988; Noguchi and Arakawa 2008; Noguchi et al. (Arakawa et al. 1994a) 1986a, 1987a). As some examples of a series of such As described later, A. floridus living on reefs of studies, we would like to introduce the separation of Ishigaki Island accumulate PSP at a very high concen- TTX derivatives from A. floridus inhabiting Kojima tration, whereas those inhabiting Kojima, a small islet Islet of Ishigaki Island, Okinawa (Arakawa et al. of Ishigaki Island, as seen in the Pacific coasts of Ja-

Table 1. 1H - NMR spectral data of TTX, TTX-U1, and U2 separated from A. floridus, compared with those of 11-oxoTTX (Khora and Yasumoto 1989), 11-norTTX-6(R)-ol (Endo et al. 1988), and 11-norTTX-6(S)-ol (Yotsu et al. 1992).

H TTX TTX-U1 11-oxoTTX TTX-U2 11-norTTX-6(R)-ol 11-norTTX-6(S)-ol 4 5.49 (9.9) 5.50 (9.5) 5.51 (9.5) 5.52 (8.9) 5.53 (9.1) 5.50 (9.4) 4a 2.34 (9.5) 2.29 (9.5) 2.31 (9.5) 1.82 (9.0) 1.84 (9.2) 2.18 (9.4) 5 4.25 4.37 4.37 4.41 4.42 4.36 6 3.72 3.73 4.37 7 4.08 4.18 4.19 4.22 4.24 4.20 8 4.29 4.26 4.27 4.03 4.05 4.17 9 3.96 3.97 3.98 3.97 3.98 3.98 11 4.02 5.73 5.74 4.04

Spectra were measured in 4%CD3COOD/D2O. Chemical shifts are expressed in ppm (internal standard, t-BuOH = 1.23 ppm). Coupling constants in parentheses are given in Hz. From Arakawa O, Noguchi T, Shida Y, Onoue Y. Occurrence of 11-oxotetrodotoxin and 11-nortetrodotoxin-6(R)-ol in a xanthid crab Atergatis floridus collected at Kojima, Ishigaki Island. Fish. Sci. 1994; 60: 769–771, Table 1.

pan Proper, showed low toxicity, and had TTX as a 6(R)-ol) (Endo et al. 1988), but not with those of 11- major toxin (Noguchi et al. 1986b). We tried to sepa- norTTX-6(S)-ol (Yotsu et al. 1992) mainly in the rate TTX derivatives from 17 A. floridus specimens chemical shifts of H-4a and H-6 (Table 1). collected from Kojima Islet. Whole bodies (300 g) of Thus we concluded that TTX-U1 and U2 were 11- the crab specimens were extracted with 80% aqueous oxoTTX and 11-norTTX-6(R)-ol (Fig. 1). Yasumoto ethanol acidified with HCl, and the toxins extracted and his co-workers first isolated these toxins and an- were purified successively by activated charcoal treat- other TTX derivative, 11-deoxytetrodotoxin, from ment, Bio-Gel P-2 and Bio-Rex 70 column chroma- pufferfish and/or newts as minor components, and es- tography, to afford TTX (0.6 mg, 2300 MU), and two timated that a series of oxidations took place at C11 in unidentified toxins, TTX-U1 (0.9 mg, 3400 MU) and these animals (Endo et al. 1988; Khora and Yasumoto TTX-U2 (0.2 mg, 440 MU). In electrospray ionization 1989; Yasumoto et al. 1988). Because the A. floridus mass (ESI/MS) spectrum of TTX-U1, a set of three specimens mainly contained 11-oxoTTX as well as peaks (m/z 317.8, 336.1, 350.3) corresponding to TTX, the oxidation might be more progressive in this + + + [M+H-H2O] , [M+H] , and [M+H-H2O+MeOH] was crab than in other TTX-bearing animals. recognized while in the spectrum of TTX-U2, a single peak corresponding to [M+H]+ appeared at m/z 290.1 2-1B. Toxicity and toxin profile of N. glans (Taniyama (Fig. 2). In 1H-nuclear magnetic resonance (NMR) et al. 2009a) spectrum of TTX-U1, coupled signals (2.29 and 5.50 In July 2007, a critical food poisoning incident due ppm, J = 9.5 Hz) corresponding to H-4a and H-4 in to consumption of small marine snails occurred in TTX, and four oxymethine signals (3.97–4.37 ppm) Nagasaki City, Nagasaki Prefecture, Japan, in which a were observed. The signals due to protons at C-11 of 60 year-old female temporarily fell into respiratory TTX were absent, instead a singlet newly appeared at arrest after she developed symptoms such as numb- 5.73 ppm. Whole spectral data of TTX-U1 coincided ness of tongue, paralysis in the limbs, and dyspnea. well with those of 11-oxotetrodotoxin (11-oxoTTX) Immediately after the incident, we investigated the (Table 1) (Khora and Yasumoto 1989). Assignments leftover marine snails including N. glans, “akanishi” of the four oxymethine signals (H-5, 7, 8 and 9) were Rapana venosa, “tengunishi” Hemifusus tuba, and supported by 1H-1H-correlation spectroscopy (COSY) “migakibora” Kelletia lischkei, and detected a maxi- spectrum in which cross peaks were observed on H- mum of 4290 MU/g of paralytic toxicity in the cooked 4a/H-5, H-5/H-7, H-7/H-8, and H-4a/H-9. 1H NMR muscles and digestive glands of N. glans. A liquid chro- spectrum of TTX-U2 gave coupled signals at 1.84 and matography-mass spectrometry (LC-MS) analysis 5.33 ppm (H-4a and H-4, J = 9.1–9.2 Hz) and five demonstrated that the main toxic principle was TTX, oxymethine signals at 3.72–4.14 ppm. In the 1H-1H- leading us to conclude that the present case was TTX COSY spectrum, couplings were recognized on H-5/ poisoning due to N. glans. H-6, H-5/H-7, H-6/H-7, H-6/H-8, H-7/H-8, and H-4a/ In connection with this incident, we screened the H-9, in addition to H-4a/H-4. This data was consistent toxicity of 7 species of marine snails by mouse bioassay with those of 11-nortetrodotoxin-6(R)-ol (11-norTTX- (MBA), N. glans (n = 22), H. tuba (n = 4),

1,510

2,410

4,500

Total toxicity*

(MU/individual)

1,570 1,330

3,350 3,590

921 809

(MU/g)

Toxicity*

1,910 1,430

0.28 509

0.33 1,020

0.30 2,450

(g)

Weight

1.5 113 165 1.3 17 22

1.3 16 21

1.0 102 98

1.4 28 38

1.1 61 68

0.8 72 57

1.2 53 62

0.8

1.1 119 133 1.1 41 44

1.4 285 410

0.8 73 57 1.4 36 50

1,670 1.3

1,400 1.1

3,090 1.3

132 1.4 154 216

682

976

725

1,150 1.4 65 94

2,640

1,070 1.6 1,890 3,080

1,540

3,070 1.6 3,850 6,120

1,660 1.5 10,200 15,100

2,860 1.4 1,980 2,830

1,770 1.8 1,880 3,380

1,420 1.6 5,580 9,150

(MU/individual)

1,320

1,310

4,180

3,980

1,130

3,520

9,860 1,300

7,880

6,150 1,540

Total toxicity*

achibana Bay, Nagasaki Prefecture. achibana Bay,

560

451

711 3,770

Muscle Viscera

(MU/g)

288

307

862

542

591

491

Toxicity*

1,360

1,250

1,260

2,370

1,200 3,230 1.1 4,300 4,730 1,970

1,470

10 times higher than that of viscera.

0.28 823

0.43 523

0.80 1,010

(g)

Weight

3.7

3.1

2.2

3.6

4.2

3.1

4.0

2.7

4.8

3.6

4.2

specimens collected at

0.66 3.3

1.1 2.8

1.4 3.6

(g)

N. glans

8.4

7.3

6.1

8.5

7.7

7.3

8.1

6.6

8.8

8.5

7.9

Body weight

0.89 6.5

1.0 7.4

3.4 7.6

Toxicity of Toxicity

(mm)

Shell width

able 2.

4.0 23

1.9 21

2.8 20

(mm)

Shell length

SD 41

SD 38

SD 43

. (2009a).

22 41 22 8.1 3.9 336

20 39 21 7.0 3.3 394

16 36 22 6.0 2.7 416

14 35 20 5.3 3.0 245 15

12 38 21 6.5 2.8 13

10 35 13 5.0 2.2 589

03 40 21 6.7 3.1 494 04

No.

Mean 44 23 8.6 4.2 276

et al

Specimen

Mean

oxicity scores were determined by mouse bioassay

Bold numbers show the specimens in which total toxicity of the muscle was 5.9–1

Jan. 2008 21 46 24 9.1 4.5 216

Nov. 2007 17 38 23 7.0 3.2

Oct. 2007 11 38 20 6.4 2.8

Month of collection

Sep. 2007 01 44 20 9.8 3.9 360

1 2

From Taniyama From Taniyama

Fig. 3. LC-MS chromatograms at m/z 320 (a, c) and 336 (b, d) obtained from muscles of specimen No. 14 (a, b) and No. 5 (c, d) (see Table 2 for specimen numbers). From Taniyama et al. (2009a).

“miyakobora” Bufonaria rana (n = 11), “biwagai” liquid chromatography with postcolumn fluorescence Ficus subintermedia (n = 10), “kinugasagai” Stellaria derivatization (HPLC-FLD). Good positive correla- (Onustus) exutus (n = 8), “yatsushirogai” Tonna tions between ‘TTX amount calculated from LC-MS luteostoma (n = 7), and “urashimagai” Semicassis analysis’ and ‘toxicity determined by MBA’ in the bisulcata persimilis (n = 4), collected from Tachibana muscle and viscera (correlation coefficient, 0.8354 and Bay, Nagasaki Prefecture between September 2007 and 0.9587, respectively), and their regression lines (y = January 2008. Among the 7 species tested, only N. 0.6478x – 93.45 and y = 0.5955x + 147.4, respectively) glans was toxic, and the other 6 species were non-toxic indicated that TTX accounts for about 65% of the total (less than 5 MU/g). In N. glans, toxicity was detected toxicity in the muscle, and for about 60% in the vis- in the muscle and viscera of all specimens, whose mean cera in average (Fig. 4). In LC-MS chromatograms at + toxicity scores were extremely high; 775 ± 615 MU/g m/z 336, a peak estimated to be 11-oxoTTX ([M+H] and 149 ± 2530 MU/g, respectively. Particularly, in the = 336) from its relative elution profile to TTX was N. glans specimens collected soon after the poisoning detected (Fig. 3). When the lethal potency of this in Nagasaki (September 2007), both or either of mus- component was calculated assuming that its relative cle and viscera exceeded 1000 MU/g in 8 of 10 tested toxicity to mice is twofold that of TTX, and ion strength specimens, the highest toxicity scores reaching 2370 per unit amount in the LC-MS analysis is equivalent MU/g in the muscle, and 10200 MU/g in the viscera to TTX, good positive correlations were recognized be- (Table 2). tween ‘sum of the concerned potency and TTX toxic- When the toxin profile of the muscle and viscera was ity’ and ‘toxicity determined by MBA’ in both muscle analyzed by LC-MS, a peak corresponding in reten- and viscera (correlation coefficient, 0.9073 and 0.9763, tion time to the TTX standard ([M+H]+ = 320) was respectively) with the regression lines being y = 1.060x detected in all specimens in the chromatograms at m/z + 75.97 and y = 0.9664x + 176.8, respectively. There- 320 (Fig. 3). No PSP components were detected in any fore, N. glans is judged to have TTX and 11-oxoTTX of 22 N. glans specimens tested by high performance as main components.

radation product of TTX, on alkaline hydrolysis, leading us to conclude that the isolated bacteria produce TTX. Shewanella alga and Alteromonas tetraodonis isolated from a red alga Jania sp. (Yasumoto et al. 1986), Shewanella putrefaciens from the pufferfish T. niphobles (Matsui et al. 1990), and some other marine bacteria isolated from TTX-bearing organisms (Simidu et al. 1987) also produce TTX.

2-2. Exogenous toxification of pufferfish and production of non-toxic pufferfish

The above-mentioned results, with some more de- tailed information such as (1) pufferfish toxicity shows remarkable individual and regional variations, (2) the trumpet shell C. sauliae accumulates TTX by ingesting toxic starfish, and (3) artificially cultured specimens of the pufferfish T. rubripes and T. niphobles with non-toxic diets possess no toxin, but when such pufferfish are orally administered TTX, they accumulate it efficiently, strongly supporting the exogenous toxification of pufferfish; i.e., pufferfish do not syn- thesize TTX, but obtain it from toxic food organisms, which bear bioconcentrated TTX through the food chain starting with marine bacteria (Noguchi and Arakawa 2008). If this is the case, pufferfish should become non-toxic when fed TTX-free diets in an environment in which the invasion of TTX-bearing organisms has been eliminated. To test this hypothesis, a total of more than 5,000 specimens of T. rubripes cultured in such environments (netcages on the sea or land aquaria) for 1–3 years were collected from main pufferfish culturing prefectures (Nagasaki, Saga, Kumamoto, Kagoshima, Ehime, Wakayama and Shizuoka) in Japan during 2001–2004, Fig. 4. Comparison between toxicity scores determined by and toxicity of their livers and some other parts was mouse bioassay and amounts of TTX by LC-MS. From examined by MBA. In addition, typical specimens of Taniyama et al. (2009a). them were submitted to LC-MS analysis (Noguchi et al. 2006). The results showed that all the livers and other parts tested were ‘non-toxic’ in both MBA (less than 2 MU/g) and LC-MS analysis (less than 0.1 MU/ 2-1C. Detection of TTX from marine bacteria g). Thus, it is confirmed that pufferfish is toxified (Noguchi and Arakawa 2008) through the food chain, and non-toxic pufferfish can Intestinal bacteria were separated from various TTX- be successfully produced by the netcage or land cul- bearing organisms, and each bacterial strain was culture. tured with a 500-ml medium. The bacterial cells ob- Although it is currently prohibited to eat pufferfish tained were examined for lethal potency to mice. Vi- liver in Japan, the Saga Prefecture has begun to con- brio alginolyticus from the toxic starfish sider serving the non-toxic cultured pufferfish liver as “togemomijigai” Astropecten polyacanthus, one Vibrio a food following an individual toxicity assay, as the strain from the xanthid crab A. floridus, and another regulation states that the part whose toxicity was con- Vibrio strain from the toxic marine pufferfish T. snyderi firmed to be around less than 10 MU/g by individual showed a paralytic toxicity of 213, 30, and 3 MU per assay is edible. In the plan, a part of a liver will be 500 ml of medium, respectively. The toxic principles used for the assay, and the rest will be sent for cook- produced peaks identical to those of TTX standard and ing, but to make this become effective, toxicity distri- 4,9-anhydroTTX in HPLC-FLD (Fig. 5). In addition, bution in the liver must be clarified in advance. Then gas chromatography-mass spectrometry (GC-MS) we divide the wild T. rubripes livers into 10 parts (L1- analyses revealed that they produce a C9-base, the deg- 5 and R1-5) (Fig. 6), and determined the lethal potency

Fig. 5. HPLC-FLD analysis of TTX fraction from a Vibrio strain (left) and of authentic TTX (right). From Noguchi and Arakawa (2008). of each part by MBA (Taniguchi et al. 2013). Regard- ing the 16 raw and 9 frozen livers, whose parts were all toxic, the relative lethal potency of each part to the mean lethal potency of the individual (8.9–709 MU/g) was calculated (Fig. 7), and subjected to a two-way analysis of variance with 2 factors (left/right and top/ bottom) for each group of livers (raw or frozen). The analysis indicated non-significance among factors and interactions at a significance level of 5% in the frozen livers. However, in the raw livers, although no inter- action between the 2 factors was detected, the right- side and the 4th-portion from the top were significantly higher than the left-side and the other portions, respectively. Therefore, we concluded that individual inspec- tion using R4, which is the region that is below the right-center location of the liver, is suited for toxicity evaluation of liver to secure the safe consumption of pufferfish. We hope to expect a further advanced food culture to be established in the future in Japan. Fig. 6. Sampling scheme of 10 parts from a liver. From Taniguchi et al. (2013). 3. Kinetics of TTX in pufferfish body

3-1. Long-term toxin accumulation and short-term toxin transfer were conducted, in which non-toxic cultured specimens of T. rubripes were reared for 60 days with various As described above, although it has become possi- types of TTX-containing diets (Table 3). The test fish ble to explain the pufferfish toxification through the administered with a crude extract from toxic tissues of food chain starting with bacteria, transfer, accumula- the wild pufferfish “nashifugu” Takifugu vermicularis tion, and elimination mechanisms of TTX taken up into accumulated a small amount of TTX (less than 3 MU/ the pufferfish body through toxic food organisms still g in most cases) mainly in the skin and liver at low remains unclear. To clarify this point, we have con- dose (0.1 MU/1 g body weight/day, 0.1 MU TTX ducted various TTX administration experiments using group), and a large amount (up to 57 MU/g) mostly in non-toxic cultured specimens of pufferfish. liver and ovary at higher doses (0.2 and 1.0 MU/1 g body weight/day, 0.2 MU and 1.0 MU TTX groups, 3-1A. Toxic diet-feeding (TDF) experiment (Honda respectively) (Fig. 8). When a purified sample of TTX et al. 2005a) was administered (0.2 MU purified TTX group), the A total of 5 toxic diet-feeding (TDF) experiments test fish accumulated it up to similar levels as in the

Fig. 7. Distribution of toxicity in raw (upper) and frozen livers (lower). Data are shown as mean (column) and standard deviation (SD, error bar) of each part (n = 16 and 9, respectively). From Taniguchi et al. (2013).

Table 3. Accumulation rate of TTX in the fish of each test group during the feeding experiments.

Feeding Test group Dose of TTX Total amount of Total amount of Accumulation rate experiment No. [MU/(1 g bw◊day)] TTX administered TTX accumulated (%) (MU/fish) (MU/fish) 1 0.1 MU TTX 0.1 480 31 6 0.2 MU TTX 0.2 960 106 11 1.0 MU TTX 1.0 4,800 498 10 0.2 MU TTX initial ad. 0.2 240 38 16 1.0 MU TTX initial ad. 1.0 1,200 208 17 0.2 MU purified TTX 0.2 960 103 11 2 0.1 MU TTX 0.1 220 12 5 3 0.1 MU TTX 0.1 1,050 0 0 4 0.2 MU TTX 0.2 8,400 2,760 33 5 0.1 MU TTX 0.1 1,280 392 31

From Honda S, Arakawa O, Takatani T, Tachibana K, Yagi M, Tanigawa A, Noguchi T. Toxification of cultured puffer fish Takifugu rubripes by feeding on tetrodotoxin-containing diet. Nippon Suisan Gakkaishi 2005; 71: 815–820, Table 1.

crude toxin (Fig. 8), while the fish accumulated gen- lation rate of toxin in 0-year-old fish reared in aquaria erally high concentrations of toxin (up to 80 MU/g) if (experiment 1–3) was calculated to be 0–17%, whereas they were directly fed the minced tissues of T. that in 1-year-old fish reared in netcages (experiment vermicularis (scrap meal and 1/2 scrap meal groups, 4 and 5) to be more than 30% (Table 3). The toxin in which a diet containing 50% and 25% of the minced once accumulated in the pufferfish tissues was retained tissues, respectively, were used) (Figs. 8, 9). No toxin there for at least 45 days without any supplies of toxic (less than 2 MU/g) was accumulated in all parts of the diets as seen in the 0.2 MU and 1.0 MU initial admin- test fish fed with a non-toxic diet, and all muscles and istration groups (Fig. 8), where the test fish were fed testes of the test fish fed with toxic diets. The accumu- toxic diets at the same dose as that of 0.2 MU and 1.0

Fig. 8. Toxin accumulation in the test fish during the TDF experiment-1. A, non-toxic diet group; B, 0.1 MU TTX group; C, 0.2 MU TTX group; D, 1.0 MU TTX group; E, 0.2 MU TTX initial administration group; F, 1.0 MU TTX initial administration group; G, 0.2 MU purified TTX group; H, scrap meal group. From Honda S, Arakawa O, Takatani T, Tachibana K, Yagi M, Tanigawa A, Noguchi T. Toxification of cultured puffer fish Takifugu rubripes by feeding on tetrodotoxin-containing diet. Nippon Suisan Gakkaishi 2005; 71: 815–820, Fig. 1.

MU TTX groups, respectively, for the first 15 days, ‘crude TTX’ extracted and partially purified from toxic and fed a non-toxic diet thereafter. ovaries of T. vermicularis (CTTX group). The groups were then maintained separately in two aerated 90-l 3-1B. Intramuscular administration (IMA) experi- tanks. Each fish was intramuscularly administered 0.1 ment (Ikeda et al. 2009) ml (50 MU) of either purified or crude TTX solution In the TDF experiments, long-term toxin accumula- and immediately returned to the tank. Then, 5 fish from tion was successfully observed, but these experiments each group were randomly collected at 1, 4, 8, 12, 24, are not suitable for tracing short-term inter-tissue toxin 72, 120, and 168 h after toxin administration and the transfer, because it is difficult to accurately adminis- toxin amount of each tissue was quantified by enzyme- ter a single large dose of toxin. To overcome this prob- linked immunosorbent assay (ELISA) for TTX lem, we used intramuscular administration (IMA) (Kawatsu et al. 1997; Ngy et al. 2008). method to introduce TTX into pufferfish body. In both PTTX and CTTX groups, TTX rapidly trans- Non-toxic cultured specimens of T. rubripes (ap- ferred from the muscle to other organs, and the toxin proximately 4 months old; body weight, 13.2 ± 3.4 g; content of the liver and skin exceeded that of muscle body length, 7.1 ± 0.6 cm; n = 80) were divided into within as little as 1 h after administration (Fig. 10). At two groups of 40 individuals; one group was adminis- that time, a high concentration of TTX was present in tered ‘purified TTX’ purchased from Wako (purity > the blood plasma, indicating that TTX transferred 90%) (PTTX group) and the other was administered mainly via the bloodstream. The toxin transfer profiles,

Fig. 9. Toxin accumulation in the test fish of 0.1 MU TTX group (A) and scrap meal group (B) on the 30th and 60th day of TDF experiment-2, and of 0.1 MU TTX group and 1/2 scrap meal group on the 60th day of TDF experiment-3). From Honda S, Arakawa O, Takatani T, Tachibana K, Yagi M, Tanigawa A, Noguchi T. Toxification of cultured puffer fish Takifugu rubripes by feeding on tetrodotoxin-containing diet. Nippon Suisan Gakkaishi 2005; 71: 815–820, Figs. 2, 3.

however, differed between the two groups from 4 to to the temporary storage of a large amount of TTX in a 72 h. In the PTTX group, little TTX was retained in particular organ or tissue other than muscle, liver, and the liver, and most (>96%) of the toxin remaining in skin. the body transferred/accumulated in the skin after 12 h, whereas in the CTTX group, a considerable amount 3-1C. Oral gavage administration (OGA) experiment of toxin (15–23% of the administered toxin or 28–58% (Wang et al. 2012) of the remaining toxin) was transferred/retained in the A single large dose of toxin can be administered us- liver for up to 24 h, despite the fact that 89% of the ing the IMA method to trace short-term inter-tissue remaining toxin transferred/accumulated in the skin at toxin transfer, but this has the disadvantage of intro- the end of rearing period (168 h) (Fig. 11). Matsui et ducing the toxin into the pufferfish body via a route al. (1981) reported that when non-toxic cultured speci- that is not feasible in the natural environment. To es- mens of T. rubripes are fed diets containing crystal- tablish a new method with the combined advantages line TTX or crude extract of toxic pufferfish ovary, of TDF and IMA methods, we used an oral gavage ad- only the test fish fed with the toxic pufferfish ovary ministration (OGA) method, and compared the results accumulate TTX in their liver. The liver tissue of T. with those obtained using the IMA method. rubripes is equipped with a specific TTX uptake Artificially hybridized specimens (designated mechanism (Matsumoto et al. 2005, 2007, 2008a, “torama”; body weight, 213 ± 28 g; body length, 18.3 2008b; Nagashima et al. 2003), suggesting that some ± 0.8 cm; n = 30) of the pufferfish “torafugu” T. substance(s) coexisting in the crude TTX might en- rubripes (female) and “mafugu” T. porphyreus (male) hance the uptake mechanism or change TTX molecules were divided into two groups of 15 individuals each; into a form that is more easily processed by this mecha- OGA group and IMA group. In the OGA group, a Teflon nism, resulting in the above mentioned difference. The tube (outer diameter ¥ length = 3 ¥ 70 mm) connected total amount of toxin remaining in the entire body at 1 to a 5-ml syringe was inserted from the mouth into the to 4 h was approximately 60% of the administered toxin digestive tract of each fish as deeply as possible, and 1 in both groups, which decreased at 8 to 12 h, and then ml (~400 MU) of TTX-containing feed homogenate increased again to approximately 60% to 80% at 24 to was squeezed into the digestive tract. In the IMA group, 168 h (Fig. 11). The temporary decrease may be due each fish was intramuscularly injected with 0.5 ml (400

Fig. 11. Changes in the relative amount of TTX [% of the administered amount (50 MU/individual)] retained in each tissue of the T. rubripes specimens during the rearing period after toxin administration. A: PTTX group; B: CTTX group. Reproduced from Toxicon, 53, Ikeda K, Murakami Y, Emoto Y, Ngy L, Taniyama S, Yagi M, Takatani T, Arakawa O, Transfer profile of intramuscularly administered tetrodotoxin to non-toxic cultured specimens of the pufferfish Fig. 10. Changes in the content of TTX (MU/g or MU/ml) Takifugu rubripes, 99–103, Fig. 3, „ 2009, with permission retained in each tissue of the T. rubripes specimens during from Elsevier. the rearing period after toxin administration. ᮀ: PTTX group; ᭹: CTTX group. The toxin content in blood plasma was determined using a combined sample of 5 individuals for each point. Reproduced from Toxicon, 53, Ikeda K, Murakami Y, decreased and that of the liver increased from 1 to 24 Emoto Y, Ngy L, Taniyama S, Yagi M, Takatani T, Arakawa h after administration (Fig. 12). During this period, O, Transfer profile of intramuscularly administered tetrodo- TTX was also detected in the blood plasma, indicating toxin to non-toxic cultured specimens of the pufferfish that TTX absorbed from the digestive tract was trans- Takifugu rubripes, 99–103, Fig. 2, „ 2009, with permission ported first to the liver through the bloodstream, and from Elsevier. rapidly taken up into the tissue. From 24 to 120 h, the toxin content of the liver decreased gradually, and the toxin appeared in the skin. This indicates that a part of MU) of TTX solution. Immediately after toxin admin- the toxin taken up into the liver was then transferred istration, the fish were returned to a 500-l aerated tank. to the skin within a relatively short period. Subsequently, three specimens per group were ran- On the other hand, intramuscularly administered TTX domly collected at 1, 8, 24, 72, and 120 h after toxin was rapidly transferred to the liver and skin via the administration, and the toxin amount of each tissue was blood, and only a little toxin remained in the muscle quantified by MBA or LC-MS. even at 1 h (Fig. 12). The transfer profile of the toxin In the OGA group, the toxin content (MU/g tissue) to the liver in the IMA group was similar to that in the of the digestive tract (including its contents) rapidly OGA group, but the time until the liver toxin content

Fig. 13. Changes in the relative amount of TTX [% of ad- Fig. 12. Changes in the TTX content (MU/g or MU/ml) re- ministered amount (400 MU/individual)] retained in each tained in each tissue of the “torama” specimens during the tissue of the “torama” specimens during the rearing period rearing period after toxin administration. Symbols and error after toxin administration. Columns and error bars indicate bars indicate the mean and standard deviation (n = 3), re- the mean and standard deviation of total amount per fish (n spectively. From Wang et al. (2012). = 3), respectively. From Wang et al. (2012).

reached it’s maximum was shorter in the IMA group profile to the IMA group was observed when T. rubripes than in the OGA group. In addition, based on the fact were intramuscularly administered TTX (Fig. 10), sug- that the toxin content of the blood plasma at 1 h in the gesting that orally ingested TTX absorbed from the IMA group was much higher than that in the OGA digestive tract is not directly transferred to the skin, group, the toxin transfer rate from the muscle to blood but only after being temporarily taken up into the liver, is much faster than the toxin absorption rate in the di- whereas intramuscularly administered TTX is directly gestive tract and the toxin uptake rate of the liver. On transferred to both the liver and skin via the blood the other hand, the transfer profile of the toxin to the stream. Subsequently, the toxin in the skin temporar- skin in the IMA group differed somewhat from that in ily decreased, but began to increase again due to the the OGA group. Namely, in the OGA group toxin trans- beginning of transfer from the liver to the skin. fer to the skin was first observed at 72 h, whereas in The total amount of toxin remaining in the whole the IMA group toxin transfer to the skin began already body (% of administered toxin) was 31–45% in the at 1 h, peaked at 8 h, temporarily decreased from 8 to OGA group, and 42–74% in the IMA group; the scores 24 h, and increased again thereafter. A similar transfer in the OGA group were generally lower than those in

Fig. 15. Seasonal changes in the toxicity (MU/individual) of Fig. 14. Seasonal changes in the toxicity (MU/individual) of the skin (upper) and testis/liver with GSI (middle), and in the skin (upper) and ovary/liver with GSI (middle), and in the plasma TTX content in male specimens of T. the TTX amount of blood plasma (lower) in the female speci- poecilonotus. Refer to the caption of Fig. 14 for the mean- mens of T. poecilonotus. The sum of free TTX (f-TTX) and ing of b-TTX, f-TTX, and binding ratio, and of white, light TTX binding to high molecular-weight substances (b-TTX) gray, and gray zones). Data are shown by mean of each month was considered as a total TTX amount in plasma (p-TTX), (column or symbol on sequential line). Error bars (SD) for and the percentage of b-TTX in p-TTX was calculated as data other than the binding ratio are omitted to avoid confu- the binding ratio. Data are shown by mean of each month sion. Reproduced from Toxicon, 55, Ikeda K, Emoto Y, (column or symbol on sequential line). Error bars (SD) for Tatsuno R, Wang JJ, Ngy L, Taniyama S, Takatani T, Arakawa data other than the binding ratio are omitted to avoid confu- O, Maturation-associated changes in toxicity of the pufferfish sion. White, light gray, and gray zones indicate ‘ordinary Takifugu poecilonotus, 289–297, Fig. 5, „ 2010, with per- period’, ‘maturation period’, and ‘just after spawning’, re- mission from Elsevier. spectively. Reproduced from Toxicon, 55, Ikeda K, Emoto Y, Tatsuno R, Wang JJ, Ngy L, Taniyama S, Takatani T, Arakawa O, Maturation-associated changes in toxicity of the pufferfish Takifugu poecilonotus, 289–297, Fig. 4, „ 2010, pufferfish body from toxic feed is transferred first to with permission from Elsevier. the liver, and then to the skin via the blood.

3-2. Involvement of sex or sexual maturation in the IMA group (Fig. 13). The difference seems to re- TTX kinetics flect the absorption efficiency of TTX from the digestive tract to some extent. Actually, Matsumoto et al. 3-2A. Annual change in toxicity of wild T. (2008a) reported that the bioavailability of enterally poecilonotus (Ikeda et al. 2010) administered TTX to T. rubripes at comparable doses In marine pufferfish, the liver and ovary usually have (0.25 and 0.50 mg/kg) to that of the present study (about strong toxicity, whereas the muscle and testis are 0.4 mg/kg) was 62 and 84%, respectively. In both OGA weakly toxic or non-toxic (Noguchi and Arakawa and IMA groups, the greatest amount of toxin accu- 2008), indicating sexual differences in pufferfish tox- mulated in the liver (23–52%) after 8 h, followed by icity, and that maturation may affect toxin kinetics in the skin (11–21%) after 72 h (Fig. 13). In conclusion, the pufferfish body. Then we collected wild specimens it was clearly indicated that intramuscularly adminis- of T. poecilonotus (93 females and 45 males) periodi- tered TTX is rapidly transferred not only to the liver cally from the Ariake Sea from October 2006 to De- but also to the skin, while the TTX taken up by the cember 2007, and investigated maturation-associated

Fig. 17. Changes in the relative amount of TTX [% of administered amount (146 MU/individual)] retained in each tissue of the “torakusa” specimens during the rearing period after toxin administration. The values over each column indicate the number of tested specimen for each rearing period. Reproduced from Toxicon, 58, Wang J, Araki T, Tatsuno R, Nina S, Ikeda K, Hamasaki M, Sakakura Y, Takatani T, Arakawa O, Transfer profile of intramuscularly administered tetrodotoxin to artificial hybrid specimens of pufferfish, Takifugu rubripes and Takifugu niphobles, 565–569, Fig. 3, „ 2011, with permission from Elsevier.

Fig. 16. Changes in the TTX (MU/g or MU/ml) content re- changes in tissue toxicity, as well as the amount and tained in each tissue of the “torakusa” specimens during the forms of TTX in the blood plasma. rearing period after toxin administration. The TTX content Based on the seasonal change in gonadosomatic in- was determined for each individual, and the symbols indi- dex (GSI; the ratio of gonad weight to body weight), cate the mean value (refer to Fig. 17 for specimen numbers we considered December–March in females and No- for each rearing period). Reproduced from Toxicon, 58, Wang vember–March in males as the ‘maturation period’, J, Araki T, Tatsuno R, Nina S, Ikeda K, Hamasaki M, April as ‘just after spawning’, and the other months as Sakakura Y, Takatani T, Arakawa O, Transfer profile of the ‘ordinary period’, and used this seasonal classifi- intramuscularly administered tetrodotoxin to artificial hy- cation to investigate the relationship between toxicity brid specimens of pufferfish, Takifugu rubripes and Takifugu and maturation. niphobles, 565–569, Figs. 1, 2, 2011, with permission from „ The seasonal profile of tissue toxicity determined by Elsevier. MBA was markedly different between females and

Fig. 18. Various shapes of TTX-bearing glands in epidermis of the skin section in T. vermicularis under light microscope. The positive stain to TTX antibody results in brown color (A ¥ 100; B ¥ 200; C ¥ 100). Stronger TTX antigen-antibody reaction was recognized at cytoplasm of the glands (arrow heads). TTX was not detected in the negative control section (D ¥ 100). Reproduced from Toxicon, 41, Mahmud Y, Okada K, Takatani T, Kawatsu K, Hamano Y, Arakawa O, Noguchi T, Intra-tissue distribution of tetrodotoxin in two marine puffers Takifugu vermicularis and Chelonodon patoca, 13–18, Fig. 1, „ 2003, with permission from Elsevier.

males. In females, liver toxicity was high during the To cut the binding between TTX and high molecular- ordinary period, and that of ovary was high during the weight substances, 0.1% acetic acid was added to the maturation period (Fig. 14). This finding suggests that high molecular-weight fraction (Yamamori 2002), and ‘turnover of toxins’ occurs between the liver and ovary. then the mixture was submitted to ELISA to quantify Skin toxicity also decreased slightly during matura- the amount of b-TTX. The sum of f-TTX and b-TTX tion period. Therefore, it is presumed that the TTX was considered as the total TTX amount in plasma (p- absorbed from toxic food organisms into the pufferfish TTX), and the percentage of b-TTX in p-TTX (bind- body is transferred mainly to the liver and skin during ing ratio) was calculated. the ordinary period, but is actively transported and In both females and males, the binding ratio of accumulated into the ovary during the maturation pe- plasma TTX was low during the ordinary period, and riod. In males, maturation-associated changes in the high during the maturation period (Figs. 14, 15), sug- toxin distribution in the body were not clearly observed gesting that quantity, species, and/or activity of TTX- (Fig. 15). Unlike ovaries, testes do not actively take binding high molecular-weight substances are in- up TTX. Therefore, even during the maturation period, creased during the maturation period, which might be as well as during the ordinary period, the TTX taken involved in the transportation of TTX from the liver to up into the liver is transferred mainly to the skin, and ovary. only a small portion to the testis. The blood plasma collected from each fish was 3-2B. TTX administration experiment using artificial ultrafiltered through a Microcon YM-50 membrane hybrid pufferfish (Wang et al. 2011) (cut-off 50,000 Da, Amicon). The combined In the previous IMA experiment, we used young and supernatant (low molecular-weight fraction) and the small T. rubripes specimens, which are relatively easy residue (high molecular-weight fraction) contain free to rear and handle, and are most suitable for this type TTX molecules (designated f-TTX) and the TTX mol- of experiment. The maturation of T. rubripes, however, ecules binding to high molecular-weight substances is very slow, and the sexual differences in the toxin (designated b-TTX), respectively (Matsui et al. 2000). transfer/accumulation profile could not be elucidated The low molecular-weight fraction was directly sub- with young specimens. Therefore, we attempted to elu- mitted to ELISA to determine the amount of f-TTX. cidate this point using the hybrid specimens

(‘torakusa’; body weight, 71.5 ± 15.1 g; body length, 12.7 ± 0.6 cm; n = 27) produced by crossbreeding “torafugu” T. rubripes (female) with “kusafugu” T. niphobles (male), which matures earlier than T. rubripes. The GSI of the male ‘torakusa’ specimens (5.96 ± 2.41) was much higher than that of the female specimens (0.40 ± 0.06). In wild T. niphobles and T. poecilonotus, the GSI of both females and males is generally less than 2% prior to maturation, but rapidly rises when maturation begins and reaches up to 10– 20% (Ikeda et al. 2010; Yu and Yu 2002), indicating that maturation had begun to occur, at least in the male specimens. It takes more than 2 years for cultured T. rubripes to mature (unpublished data), whereas T. niphobles usually mature within 1 year (Honma et al. 1980). Therefore, in terms of maturation, the hybrid specimens (10 months old) seem to be closer to T. niphobles. The GSI of females normally begins to increase about 1 month later than that of males (Ikeda et al. 2010; Yu and Yu 2002), suggesting that the female specimens present were in a very early stage of maturation, just before the GSI began to increase. In the test fish administered 146 MU TTX in physiologic saline, TTX rapidly transferred from the muscle via the blood to other organs, as seen in T. rubripes. Toxin transfer to the ovary rapidly increased to 53.5 MU/g tissue at the end of the 72-h test period, whereas the TTX content in the liver and skin was, at most, around 4 to 6 MU/g tissue, and in the testis it was less than 0.01 MU/g tissue (Fig. 16). In the previous study on annual change in T. poecilonotus toxicity, we presumed that the TTX absorbed from toxic Fig. 19. (A) Immunohistological view of a positive stain of food organisms into the female pufferfish body is ac- a representative skin section of T. steindachneri under light tively transported and accumulated in the ovary dur- microscope (¥20). The positive stain to TTX-antibody re- ing the maturation period. The female ‘torakusa’ speci- sults in a brown colour. (B) Magnified view (¥100) of a posi- mens used in the present study seemed to be in the tive stain: (1) undifferentiated basal cell: (2) Malpighian cell; very early stages of maturation, as expected, in which and (3) sacciform cell. (C) Cross-section of the skin treated such a toxin transportation mechanism would have with mouse anti-Streptococcus antibody NUF44 (¥40). TTX antigen does not react with this antibody. Reproduced from begun to function. Toxicon, 40, Tanu MB, Mahmud Y, Takatani T, Kawatsu K, On the other hand, based on the total amount of toxin Hamano Y, Arakawa O, Noguchi T, Localization of tetrodo- per individual (% of the administered toxin), the skin toxin in the skin of a brackish water puffer Tetraodon and the liver contained higher amounts (20–54% and steindachneri on the basis of immunohistological study, 103– 2–24%, respectively), but the amount in the liver rap- 106, Fig. 2, „ 2002, with permission from Elsevier. idly decreased after 8 to 12 h, and fell below the level in the ovary after 48 h (Fig. 17). Although statistically not significant, the sum of the toxin amount in the skin and ovary in female specimens seemed to correspond 4. Physiologic function of TTX to the change in the toxin amount in the skin in the male specimens. In addition, the toxin content of the 4-1. Defense against predators skin remained at the same level in females, but tended to increase in males after 12 h, suggesting that a part While TTX is a fatal toxin to organisms that possess of the TTX that should be, after first being taken up no TTX, including humans, TTX-bearing organisms into the liver, transferred/accumulated into the skin in are assumed to utilize TTX effectively for their sur- male specimens is transferred to the ovary in female vival. Kodama et al. (1985, 1986) found that marine specimens. pufferfish with toxic skin possess unique exocrine

Fig. 20. Light micrographs of representative ovary sections in T. vermicularis (A ¥ 200; B ¥ 400). Four stages of oocytes are observed: a: perinucleolus stage - TTX antigen was not detected, b: late perinucleolus stage - TTX was visualized at nucleus and at yolk vesicle towards cell wall, c: yolk granule stage-I - TTX was visualized at nucleus and at yolk vesicle scatterdly, d: yolk granule stage-II - TTX was recognized at yolk granules (arrows) and yolk vesicles (arrow heads). Reproduced from Toxicon, 41, Mahmud Y, Okada K, Takatani T, Kawatsu K, Hamano Y, Arakawa O, Noguchi T, Intra-tissue distribution of tetrodotoxin in two marine puffers Takifugu vermicularis and Chelonodon patoca, 13–18, Fig. 3, „ 2003, with permission from Elsevier.

glands or gland-like structures in the skin, and release Tetraodon steindachneri and “midorifugu” Tetraodon a large amount of TTX when an electric shock is given. nigroviridis, no gland was observed, but cacciform (se- Saito et al. (1985) also reported that pufferfish release cretory) cells were seen in the skin, in which TTX was the toxin by a mild stimulus such as wiping the body localized (Fig. 19). Namely, it is presumed that when surface with gauze. pufferfish encounter an enemy, they not only threaten On the other hand, we tried to clarify whether TTX it by swelling, but also defend themselves by releas- is localized in the skin gland of pufferfish or not by an ing TTX from the toxic glands or sacciform cells in immunohistochemical technique (Mahmud et al. their skin. The toxin released will be immediately di- 2003a, b; Tanu et al. 2002). Skin sections (3 mm in luted by seawater, and cannot kill the enemy, but would thickness) prepared according to conventional histo- have the ability to repel the predation of pufferfish, as logical procedure were treated successively with 10% general non-toxic fish can sense extremely low levels H2O2 in distilled water and 25% goat serum in 0.01 of TTX with their gustatory receptors (Yamamori et mol/l PBS, and then incubated with monoclonal anti- al. 1988). The Japanese newt “akaharaimori” Cynops TTX antibody (Kawatsu et al. 1997), followed by a pyrrhogaster also possesses TTX-bearing glands in the polymer, Envision+ for 50 min. For negative control, skin, and secretes TTX by external stimuli (Tsuruda et mouse serum was used instead of anti-TTX antibody. al. 2002). After being treated with 0.017% 3,3¢-diaminobenzidine In many marine pufferfish, the highest levels of TTX tetrahydrochloride substrate solution in 0.01 mol/l PBS are found in the ovaries (Noguchi and Arakawa 2008), buffer and 0.01% H2O2 for 1–2 min, the sections were especially in the maturation period (Ikeda et al. 2010). counter-stained with Gill hematoxylin, dipped in am- We investigated microdistribution of TTX in the monia water, and observed under a light microscope. pufferfish ovary by the above-mentioned immunohis- The skin sections of T. vermicularis clearly showed tochemical technique (Mahmud et al. 2003a). Four different shapes of glands in the epidermis layer (Fig. stages of oocytes such as (a) early peri nucleolus stage, 18). TTX was localized as brown color at the glands. (b) late peri nucleolus stage, (c) yolk granule stage-I, Strong TTX antigen-antibody reaction was observed and (d) yolk granule stage-II are found in the at cytoplasm of the glands (arrow heads). An opening immunostained representative sections of ovary in T. extending from the gland towards super epithelial layer vermicularis (Fig. 20). TTX was localized at late peri was visualized. No specific stain of antigen-antibody nucleolus stage, yolk granule stage-I, and yolk gran- reaction was observed in negative control section. In ule stage-II of oocytes. In late peri nucleolus stage and brackish water pufferfish such as “hachinojifugu” yolk granule stage-I TTX was recognized at nucleus

Fig. 21. A representative positive stain section of ovary (¥100) in C. patoca. TTX antigen was detected in connective tissues (arrow heads) and in the nucleus of some perinucleolar oocytes (d). Stronger antigen-antibody complexes (arrows) were scatterdly observed in connective tissues, which seem to be macrophages. Reproduced from Toxicon, 41, Mahmud Y, Okada K, Takatani T, Kawatsu K, Hamano Y, Arakawa O, Noguchi T, Intra-tissue distribution of tetrodotoxin in two marine puffers Takifugu vermicularis and Chelonodon patoca, 13–18, Fig. 4, „ 2003, with permission from Elsevier. and yolk vesicles, while at yolk granule stage-II TTX worms, and horseshoe crabs also possess high concen- was visualized at yolk granules and yolk vesicles. At trations of TTX in their eggs (Miyazawa and Noguchi yolk granule stage-II, toxin from nucleus disappeared. 2001; Noguchi and Arakawa 2008), and TTX may act In contrast, in the positive ovary section of as a repellant, thereby contributing to the survival of “okinawafugu” Chelonodon patoca TTX was visual- the eggs and hatched larvae, like in pufferfish. ized as brown color in the connective tissues (Fig. 21) and in the nucleus of some perinucleolar oocytes. The 4-2. Offense to prey antigen-antibody complexes were, however, faded in the perinucleolar oocytes. As described above, TTX may function defensively Thus, although the microdistribution of TTX in the in many species, but some species are presumed to use oocytes varies depending on the species and matura- it offensively. For example, the blue-ringed octopus tion stage, T. vermicularis accumulate TTX mainly in possesses TTX in its posterior salivary glands, and ar- the yolk granules and yolk vesicles with the progres- row worms in the head near the raptorial apparatus, sion of maturation. In the goby Y. criniger, the amount and they are thought to utilize the TTX to capture food of toxin in the ovary increases during the spawning organisms (Williams 2010). There are several recorded period as in pufferfish (Tatsuno et al. 2013a). In human envenomations by the blue-ringed octopus. A pufferfish, the toxin that accumulates in the eggs is species of the flatworm Planocera harbors TTX in the transferred to the hatched larvae, where it is maintained pharynx, and appears to use it to overcome their much for a certain period after hatching (Nagashima et al. larger gastropod prey (Ritson-Williams et al. 2006). 2010). Itoi et al. (2014) observed juveniles of gener- In 1998, a ribbon worm that accumulates an ex- ally nontoxic fish that ingest pufferfish (T. niphobles tremely high concentration of TTX was found in Hiro- and T. rubripes) larvae and then promptly spit them shima Bay, Hiroshima Prefecture, Japan, and later iden- out, and presumed that the TTX transferred from the tified as “akahanahimomushi” Cephalothrix simula mother works to repel predators, based on the findings (Asakawa et al. 2013). We investigated the intra-tis- that TTX was primarily localized on the body surface sue microdistribution of TTX in this species by the of the larvae and that rainbow trout (Oncorhynchus immunohistochemcal technique (Tanu et al. 2004). In mykiss) and arctic char (Salvelinus alpinus) are able to the nemertean species belonging to the order sense extremely low levels of TTX with gustatory Palaeonemertea, the proboscis is composed of, from receptors (Yamamori et al. 1988). Flatworms, ribbon outside to inside when the organ is everted, an epithe-

Fig. 22. Anti-TTX immunostaining micrographs to show its microdistribution among various tissues and cells in C. simula. (A) Transverse section through the intestinal region to show the proboscis (PR’) in the intestinal lumen (IL); large arrowheads indicate the TTX-containing vesicles basally in the intestinal wall near the blood vessel (BV) and rhynchocoel. PR: proboscis; EP: epidermis; LML: longitudinal muscle layer; IE: intestinal epithelium. Enlargement of the red polygon, which appears in the original figure is omitted. (B) Transverse section through the proboscis. GC: granular cell; IC: interstitial cell; VC: vacuolated cell. (C) Enlargement of the red box on G to show the granular cell (GC) and pseudocnides (PS). Modified from Toxicon, 44, Tanu MB, Mahmud Y, Arakawa O, Takatani T, Kajihara H, Kawatsu K, Hamano Y, Asakawa M, Miyazawa K, Noguchi T, Immunoenzymatic visualization of tetrodotoxin (TTX) in Cephalothrix species (Nemertea: Anopla: Palaeonemertea: Cephalotrichidae) and Planocera reticulata (Platyhelminthes: Turbellaria: Polycladida: Planoceridae), 515–520, Fig. 1, „ 2004, with permission from Elsevier.

lium, an outer circular muscle layer, an inner longitu- on the body surface of the prey, from which TTX-con- dinal muscle layer and a thin endothelium. The pro- taining substances released from the granular cells will boscis epithelium mainly consists of granular cells, be poured. vacuolated cells and interstitial cells. The distal portion of the cytoplasm in the interstitial cell contains 4-3. Other functions pseudocnides. In the transverse sections of the proboscis (PR) in the C. simula, TTX was recognized as strong Saito et al. (2000) found that juveniles of the brown staining of the granular cells (GC) in the epi- pufferfish T. rubripes were attracted by TTX-contain- thelium (Fig. 22). The pseudocnides (PS) themselves ing gelatin. In addition, Okita et al. (2013a) revealed did not appear to possess TTX; they may make wounds using olfactory ablated test fish that T. rubripes juve-

Fig. 23. Agglutinating titer of the puffer sera from each test group against SRBC in the TDF experiment-1. Bars indicate the SD of 3 fish. Asterisks show that the scores are significantly higher than those of non-toxic diet group (p < 0.05). From Honda et al. (2005b). niles detect TTX by their olfactory organ. Interestingly, 1–3 types of mitogens in all the 3 experiments (Fig. general fish that have no TTX may sense TTX by taste 24). These results suggest that TTX or TTX-contain- and repel it, whereas pufferfish smell TTX-bearing ing diets has an immnopotantiating effect on the cul- organisms and preferentially prey on them. Matsumura tured puffer fish T. rubripes, though the mechanism (1995) indicated that mature males of the pufferfish T. remains to be elucidated. niphobles were attracted by a very low concentration Non-toxic hatchery-reared juveniles of T. rubripes, of TTX, and presumed that the TTX released from the in comparison with toxic wild juveniles, are affected ovulated eggs works as a pheromone to attract male by predation more easily (Shimizu et al. 2006, 2007). fish. This could be due to the difference in their swimming T. rubripes are nervous fish though they look the re- behavior and/or presence/absence of TTX. Recently, verse, feel stress easily, and bite each other under the Sakakura et al. (unpublished) found that when TTX culture environment. Therefore, fish-farmers cut their was administered to hatchery-raised nontoxic juveniles, teeth periodically to protect them from damage to the their ecologic behavior became similar to that of wild fins due to the biting. Moreover, they control fish dis- juveniles, making it more difficult for predators to prey eases and parasites by giving a medicated bath, but on them. In addition, Okita et al. (2013b) revealed that massive deaths occur repeatedly without getting ad- when TTX was orally administered to nontoxic T. equate effect. We investigated immune functions of the rubripes juveniles, it passed through the blood-brain test fish used for the above-mentioned TDF experi- barrier to reach the brain, and that TTX was similarly ments, assuming that one of the reasons why cultured localized in the brain in toxic wild juveniles. These T. rubripes easily feel stress and get sick is absence of findings suggest that TTX is involved in controlling TTX, which they should have primarily (Honda et al. information transmission in the central nervous sys- 2005b). tem of pufferfish, and may strongly influence the physi- At the end of experiments 1–4, the test fish were ology and ecology of pufferfish. TTX seems to act not determined for their serum antibody titer against sheep only as a defensive or offensive substance, but also red blood cells (SRBC). The result showed that the test has various physiological functions that are essential fish administered with TTX-containing diets exhibited for the survival of TTX-bearing organisms. slightly or significantly higher agglutinating titer than those of the fish fed with non-toxic diet in all the four 5. Toxin other than TTX bore by pufferfish—para- experiments (Figs. 23, 24). Furthermore, in the ex- lytic shellfish poison (PSP) periments 2–4, the test fish were also examined for their splenocyte proliferation by stimulation with three 5-1. Diversity of PSP components mitogens, Poke weed mitogen (PWM), concanavalin A (ConA), or lipopolysaccharide (LPS). Consequently, PSP is a group of neurotoxins mainly produced by the splenocytes from toxic diets groups were found to toxic dinoflagellates including Alexandrium catenella, show significantly higher proliferation than those from Alexandrium tamarense, Gymnodinium catenatum, etc., non-toxic diet group when they were stimulated with and high-toxicity components have equivalent lethal

Fig. 24. Agglutinating titer of the sera (upper) and proliferation level of the splenocytes (lower) from each test groups in the TDF experiment-2 (left), -3 (center), and -4 (right). The density of splenocytes after incubation with or without PWM, ConA or LPS was determined by measuring the specific absorbance at 570 nm and 600 nm under the presence of Alamer Blue. In the upper graphs, bars indicate the SD of 5–10 fish, and asterisk shows that the score is significantly higher than that of non-toxic diet group (p < 0.05). In the lower graphs, symbols and bars indicate the mean ± SD of triplicate cultures in 4 independent experiments, and asterisks show that the scores are significantly higher than those of non-toxic diet croup (p < 0.05). From Honda et al. (2005b). potencies to TTX (Wiese et al. 2010). In the US and zeteki (Yotsu-Yamashita et al. 2004), and non-toxic Canada, bivalves such as mussels and Alaskan butter derivatives whose C-12 is reduced, from the blue-green clams had been toxified with this toxin, and occasion- alga Lyngbya wollei (Onodera et al. 1997) or the ally caused human poisonings in the past. At first, dinoflagellate Alexandrium minutum (Lim et al. 2007). Schantz et al. (1957), separated a toxic component from the siphons of toxified Alaskan butter clams as the 5-2. Separation of novel PSP components from causative substance, and named saxitoxin (STX) after toxic xanthid crabs the scientific name of the clam, Saxidomus giganteus. The chemical structure was clarified by an X-ray dif- Three crabs of the family Xanthidae, “umoreogigani” fraction analysis in 1975 (Schantz et al. 1975). Subse- Zosimus aeneus, A. floridus, and quently, many derivatives with the same skeleton as “tsubuhiraashiogigani” Platypodia granulosa, inhab- STX, such as 1-N-hydroxy derivative [neosaxitoxin iting reefs of the Southwestern Islands in Japan, pos- (neoSTX)] (Shimizu et al. 1978), 11-sulfate derivatives sess a large amount of PSP, and have caused food poi- [gonyautoxins (GTXs)] (Noguchi et al. 1981; Shimizu soning cases occasionally (Noguchi et al. 2011b). They et al. 1976), N-sulfo-carbamoyl derivatives (Kobayashi have not only large individual and regional variations and Shimizu 1981; Harada et al. 1982a), and in toxicity, but also a marked regional variation in toxin decarbamoyl derivatives (Harada et al. 1983) were composition. A. floridus are also distributed in Chiba successively found in several species of dinoflagellates prefecture or other Pacific coastal prefectures to the and toxified bivalves (Fig. 25). Moreover, south of Chiba in Honshu, as well as Shikoku and deoxydecarbamoyl derivatives, and derivatives whose Kyushu Islands, where they have relatively low toxic- carbamoyl nitrogen or C-11 are modified were sepa- ity with TTX as main toxin, while the three species of rated from G. catenatum (Negri et al. 1998; Oshima et crabs inhabit the reefs accumulate PSP at extremely al. 1993) or the toxic Latin-American frog Atelopus high concentration (Noguchi et al. 2011b).

Fig. 25. Structures of PSP components. STX: saxitoxin; neoSTX: neosaxitoxin; GTX: gonyautoxin; C: C toxin; dc: decarbamoyl; ol: 12b-deoxy; hy: carbamoyl-N-hydroxy; do: deoxy-decarbamoyl; me: carbamoyl-N-methyl; ac: 13-O-acetyl; SEA: saxitoxin ethanoic acid. Blue letters indicate the components we identified.

5-2A. Separation of carbamoyl-N-hydroxy deriva- shifts resembled those of STX cited from the literature tives from Z. aeneus (Arakawa et al. 1994b) (Table 5) (Rogers and Rapoport 1980). The downfield We attempted to separate novel PSP components shift by b effect at C-6 of neoSTX (Shimizu et al. 1978) from Z. aeneus specimens collected on the reefs of was not recognized at C-4 and C-5 of BR-2; thus there Ishigaki Island, Okinawa. Toxins were extracted with was little possibility of 7-N- or 9-N-hydroxy forma- acidic ethanol from the appendages and carapaces (700 tion. Similarly, the chemical shifts of BR-1 agreed with g), defatted, and purified by activated charcoal treat- those of neoSTX. The acid hydrolyzate of BR-2 was ment, and repetitive chromatographies using Bio-Gel indistinguishable from dcSTX in electrophoresis, P-2 and Bio-Rex 70 columns to afford two unknown HPLC-FLD, and ESI/MS analyses, excluding the bind- PSP components, BR-1 (2.4 mg; 1400 MU/mg) and ing of hydroxyl group to the nitrogen molecules of the BR-2 (1.9 mg; 1700 MU/mg). In ESI/MS spectral basic skeleton. Thus the structures of two novel tox- analyses, BR-1 and 2 gave a set of three peaks corre- ins, BR-1 and 2 were concluded to be carbamoyl-N- + + sponding to [M-H2O+H] , [M+H] , and [M- hydroxyneosaxitoxin and carbamoyl-N- + H2O+MeOH+H] , from which their molecular weights hydroxysaxitoxin (Fig. 25), which seem to occur ex- were judged to be 331 and 315, respectively. This in- clusively in xanthid crabs. dicated that BR-1 and 2 might be the respective monohydroxy derivatives of neoSTX and STX. 1H- 5-2B. Separation of carboxymethyl derivative from NMR spectrum of BR-2 showed a close resemblance A. floridus (Arakawa et al. 1995) to that of STX, except for the absence of a H-11 signal We found another new unique PSP component in A. due to deuterium replacement through enolization of floridus collected from Asakawa Bay, Tokushima Pre- the keto group at C12 (Wong et al. 1971) with simpli- fecture. The toxin was extracted and purified by es- fied coupling patterns of the other H-11 and H-10 sig- sentially the same procedures as in the Z. aeneus, and nals. The chemical shifts also consistent with those of an unknown component (ATX, 5.2 mg, 830 MU/mmol) STX excluding the slight downfield shifts of H-13 sig- was obtained. In ESI/MS spectrum analysis, ATX gave nals (Table 4), suggesting the carbamoyl-N-hydroxy a set of three peaks (m/z 340.0, 357.8, 272.0) corre- + + formation in BR-2. The chemical shifts of BR-1 were sponding to [M-H2O+H] , [M+H] , and [M- + compatible with those of neoSTX except for a slight H2O+MeOH+H] , as seen in the carbamoyl-N-hydroxy downfield shift of H-13 signals as described above. In derivatives, indicating that the mol. wt of ATX is 357, the carbon resonance spectrum of BR-2, the chemical i.e. 58 larger than STX. Figure 26 shows 1H-NMR

Table 4. Comparison of 1H NMR data of BR-2 and 1 with those of STX and neoSTX.

Proton BR-2 STX BR-1 neoSTX H-5 4.56 (s) 4.59 (1) 4.74 (s) 4.74 (s) H-6 3.70 (9, 5) 3.69 (9, 5, 1) 3.99 (6, 6) 3.97 (6, 6) H-10 3.65 (10, 9) 3.66 3.68 (10, 10) 3.69 3.44 (10, 9) 3.44 3.49 (10, 10) 3.48 H-11 æ* 2.29 æ* 2.33 2.20 (10, 10) 2.21 2.25 (10, 10) 2.26 H-13 4.24 (11, 9) 4.16 (12, 9) 4.43 (11, 6) 4.33 (11, 6) 3.97 (11, 5) 3.88 (12, 5) 4.22 (11, 6) 4.13 (11, 6)

Chemical shifts are expressed in ppm downfield from DSS. Coupling constants in parentheses are given in Hz. *Unobserved owing to deuterium replacement at C-11. Reproduced from Toxicon, 32, Arakawa O, Noguchi T, Shida Y, Onoue Y, Occurrence of carbamoyl-N-hydroxy derivatives of saxitoxin and neosaxitoxin in a xanthid crab Zosimus aeneus, 175–183, Table 1, © 1994, with permission from Elsevier.

13 spectra of ATX taken in D2O (upper) and 10% D2O/ Table 5. Comparison of C NMR data of BR-2 and 1 with those of STX (Rogers and Rapoport 1980) and neoSTX H2O (lower). Since tripled sets of signals were observed in the spectra, ATX was considered to be a mixture of (Shimizu et al. 1978). three tautomers (designated A, B, and C in order of Carbon BR-2 STX BR-1 neoSTX their signal intensities). In 10% D2O/H2O, the spectrum of tautomer A was essentially analogous to that C-4 83.7 82.8 83.1 82.2 of the hydrate form of STX, except that one of the H- C-5 57.3 57.5 58.0 56.9 11 signals was absent and instead two methylene sig- C-6 53.3 53.5 64.2 64.4 nals (H-15) appeared in upper fields (Table 6). When C-10 42.8 43.3 43.2 43.7 C-11 32.6 33.4 *1 31.9 the spectra were taken in D2O, the signal of H-11 rap- æ idly disappeared owing to deuterium replacement C-12 98.8 99.0 98.8 98.6 through tautomeric enolization of the keto group at C- C-13 64.0 63.6 62.3 61.1 C-2 156.2 156.4 159.5*2 158.8*2 12 (Wong et al. 1971) (Fig. 27) with simplified pat- 2 2 C-8 158.9 158.3 158.3* 158.1* terns of H-10 and H-15 signals. Tautomer B was sup- C-14 159.6 159.3 154.8*2 posed to be the keto form of tautomer A because (1) H-11 signal shifted to a lower field to d 3.10, an ap- Chemical shifts are expressed in ppm (internal standard, di- propriate position for methane proton adjacent to a oxin = 67.4 ppm). carbonyl group, and (2) one of the H-13 signals slightly *1 Unobserved due to deuterium replacement at C-11. shifted to an upper field and the other to a lower field, *2 Assignments may be interchanged. reflecting the changes caused by the replacement of Reproduced from Toxicon, 32, Arakawa O, Noguchi T, Shida Y, Onoue Y, Occurrence of carbamoyl-N-hydroxy derivatives the 12b-quasi-axial hydroxyl group by a ketone (Shimizu et al. 1981). These data, along with the re- of saxitoxin and neosaxitoxin in a xanthid crab Zosimus sult of ESI/MS, strongly suggested that ATX would be aeneus, 175–183, Table 2, © 1994, with permission from Elsevier. the 11-CH2COOH derivative of STX. The configuration of the substituent at C-11 in tautomer A was estimated based on the magnitude of coupling constants involving H-10 and H-11. The di- C-11 (Table 7), and we concluded that ATX is 11- hedral angle between H-10a and H-11b in STX is close saxitoxinethanoic acid (SEA) (Fig. 25). SEA was as- to 90∞, and only a small coupling (2 Hz) is observed sumed to exist as an equilibrium mixture of three between the two protons (Shimizu et al. 1981). Assum- tautomers, the main tautomer being the hydrate form ing that ATX has the same skeletal conformation as of 11b-epimer. The tautomeric transformations could STX, the corresponding coupling of the 11a-epimer take place much faster than in GTXs, as demonstrated must be very small, as seen in GTX2 (11a-epimer) (Hsu by the trailing shoulder of the peak in HPLC and rapid et al. 1979). Since the coupling constants between H- disappearance of H-11 signal in the 1H-NMR spectrum 10 and H-11 in tautomer A (9 and 10 Hz) were both as in D2O. It is interesting that SEA, the first natural large as those (9 and 7 Hz) of GTX3 (11b-epimer), it analgue of STX that has a substituent other than the was determined to be the 11b-epimer. Tautomer C could hydroxylsulfate at C-11, seems to be rather specific to 13 possibly be the 11a-epimer of tautomer A. C-NMR the A. floridus living in some restricted regions. spectral data agreed with the assumed substitution at

1 Fig. 26. H-NMR spectra of ATX measured in D2O (upper) and 10% D2O/H2O (lower). The spectra were measured at 400 MHz using a Bruker ARX 400 spectrometer. Tert-BuOH (1.23 ppm) was used as an internal standard. A water band at 4.76 ppm was removed or diminished by use of the homogate-decoupling technique. The letters (A, B and C) attached to H indicate the corresponding tautomers which gave those signals. Reproduced from Toxicon, 33, Arakawa O, Nishio S, Noguchi T, Shida Y, Onoue Y, A new saxitoxin analogue from a xanthid crab Atergatis floridus, 1577–1584, Fig. 4, „ 1995, with permission from Elsevier.

Table 6. Comparison of 1H-NMR data of ATX with those of STX (Shimizu et al. 1981).

Proton ATX STX Tautomer A Tautomer B Tautomer C Hydrate form Keto form H-5 4.73 (s) 4.79 (s) 4.83 (s) 4.77 (1) 4.53 (s) 1 2 4 H-6 3.80 (9, 5) 3.91* æ* 3.87 (9, 5, 1) 4.01 æ* 1 1 4 H-10a 3.91 (10, 9) 4.20* 3.89* 3.83 (9, 2) 3.80 æ* 3 4 H-10b 3.18 (10, 10) 3.71 (10, 10) 3.40 (10* ) 3.61 (10, 10, 9) 3.69 æ* 2 4 H-11 2.76 (m) 3.10 (m) æ* 2.41 (2H, m) 2.92 æ* 1 2 H-13 4.24 (12, 9) 4.19* æ* 4.32 (11, 9) 4.19 (12, 6) 2 4.03 (12, 5) 4.12 (12, 5) æ* 4.05 (11, 5) 4.11 (12, 5) H-15 2.68 (16, 7) 2.98 (18, 5) 2.65*1 2.51 (16, 6) 2.67*1 2.37 (16*3)

Chemical shifts are expressed in ppm (internal standard, tert-BuOH = 1.23 ppm). Coupling constants in parentheses are given in Hz. *1 Chemical shifts were estimated from the 1H-1H-COSY spectrum. *2 Signals could not be assigned due to overlap with those from the other tautomers. 3 * Coupling constants between germinal protons which were determined from the spectrum measured in D2O. *4 Coupling constants could not be determined due to overlap with signals from hydrate form. Reproduced from Toxicon, 33, Arakawa O, Nishio S, Noguchi T, Shida Y, Onoue Y, A new saxitoxin analogue from a xanthid crab Atergatis floridus, 1577–1584, Table 1, © 1995, with permission from Elsevier.

Table 7. Comparison of 13C-NMR data of ATX with those of STX (Rogers and Rapoport 1980).

Carbon ATX STX C-4 83.5 (C) 82.8 C-5 57.5 (CH) 57.5 C-6 53.5 (CH) 53.5

C-10 48.2 (CH2) 43.3 C-11 39.9 æ* 33.4 C-12 99.3 (C) 99.0

C-13 63.6 (CH2) 63.6

C-15 33.9 (CH2) C-16 180.6 (C) C-2 156.2 (C) 156.4 C-8 158.3 (C) 158.3 C-14 159.3 (C) 159.3

Chemical shifts are expressed in ppm (internal standard, di- oxin = 67.4 ppm). Carbon species given in parentheses were determined by the method of distortionless enhancement by polarization transfer (DEPT). *Could not be assigned owing to deuterium replacement at C-11. Reproduced from Toxicon, 33, Arakawa O, Nishio S, Noguchi T, Shida Y, Onoue Y, A new saxitoxin analogue from a xanthid crab Atergatis floridus, 1577–1584, Table 2, © 1995, with permission from Elsevier.

Fig. 27. Tautomeric transformation in ATX. Reproduced from Toxicon, 33, Arakawa O, Nishio S, Noguchi T, Shida Y, cluding muscles (Table 8). A seasonal variation in tox- Onoue Y, A new saxitoxin analogue from a xanthid crab icity was observed, in which the toxicity in August and Atergatis floridus, 1577–1584, Fig. 5, 1995, with permis- „ October was relatively low. Toxins were extracted from sion from Elsevier. each tissue, ultrafiltered, and submitted to HPLC-FLD analyses, demonstrating that they were composed of STX, neoSTX, and dcSTX, and both TTX-related com- 5-3. PSP of freshwater pufferfish ponents and GTXs were undetectable. Mol% of each component was 92, 0.4, and 7.6 in the eggs, 85, 6, and Small freshwater pufferfish inhabit inland waters in 9 in the liver, and 82, 10, and 8 in the muscle, respec- South/Southeast Asian countries. They are imported tively. and sold as an ornamental fish, but not eaten in Japan. In Bangladesh, Thailand, and Cambodia, however, food 5-3B. Bangladeshi species (Zaman et al. 1997, poisonings due to these pufferfish including fatal cases 1998) occasionally occur. The toxic principle of freshwater Two species of freshwater pufferfish, Tetraodon pufferfish was first reported as TTX (Kodama and cutcutia (n = 264) and Chelonodon patoca (a possibil- Ogata 1984), but later Sato et al. (1997) detected STX ity of species misidentification is suspected; n = 69), as the main toxin in the Thai freshwater pufferfish, were collected from several locations in Bangladesh. Tetraodon fangi. Here, toxicity and toxin profile of They showed lethal potency in mice ranging from 2.0 South/Southeast Asian freshwater pufferfish we inves- to 40.0 MU/g tissue as PSP. Like Thai freshwater spe- tigated are described. cies, toxicity of the skin was higher than the other tissues examined (muscle, liver and ovary) in both spe- 5-3A. Thai species (Kungsuwan et al. 1997) cies, but the toxicity levels were generally lower than From May 1995 to June 1996, we collected two spe- those of the Thai. Water-soluble toxins from T. cutcutia cies of freshwater pufferfish, Tetraodon suvatii and (minced tissues 550 g) were partially purified by acti- Tetraodon leiurus (a total of 69 specimens), and ex- vated charcoal treatment followed by column chroma- amined for their toxicity by MBA. The results showed tographies using Bio-Gel P-2 and Bio-Rex 70 to give that in both species the mean toxicity scores of the skin two fractions, STX fraction (2600 MU) and GTX frac- were generally higher than those of the other parts, and tion (180 MU). In HPLC-FLD analyses, STX and maximum scores exceeded 100 MU/g in all tissues in- dcSTX were detected in the STX fraction, and GTX2,

113)

116)

129)

20)

117)

49)

73)

117 (0

106 (63

123 (24

880)

119) (0 49

26) (11 16

90) (4 67

134) (55 67

174)

11)

60)

64) 41

Liver Egg

220) (36 50

128) (5 8

101) (5 11

230) (7 33

213) (4 63

266) (32 61

99) (19 60

100) (10 16

120) (0 49

Mean toxicity (MU/g)

Skin

5 (0

200 (82

163 (104

105 (64

109 (16

Thai fresh water puf

106)

74)* (33 65 111)

70) (58 93

79) (11 36

191)

58) (5 36

67) (0 20 18)

- -

Muscle

42 (13

17 (5 14 (10

16 (6 36 (0 14 (0 36 (11

51 (24 74 (31

from Udonthani province.

suvatii

(cm)

Mean body length

Occurrence of paralytic shellfish poisons in

(type 1 and 2) and

(g)

31.0 10.0

18.0 8.5 19.0 8.5

93.0 15.0 7 58 0 117

19.5 9.0 0

18.0 9.0 0 10 0 23 26.0 10.0 0 0 0 0

72.0 13.0 18.0 10.0 20.0 9.0 80.0 14.0

21.6 9.0 28.0 9.5

leiurus

Mean body weight

Number of

MBA results of MBA

positive samples

able 8.

41 10

55 97 33 44

77 88

samples tested

, Kungsuwan A, Arakawa O, Promdet M, Onoue A, , Kungsuwan

1 2

T. suvatii

T. leiurus T. leiurus

T. suvatii

T. leiurus

T. leiurus T. leiurus

T. suvatii T. leiurus T. leiurus T. suvatii

T. leiurus T. leiurus

oxicon

June 1995

Feb. 1996

Oct.1995

Aug. 1995

July 1995

Collection date Typeof Number

May 1995

*Range of toxicity; NA, not assayed. Reproduced from with permission from Elsevier doi:10.5047/absm.2017.01003.0041 © 2017 TERRAPUB, Tokyo. All rights reserved. 68 O. Arakawa et al. / Aqua-BioSci. Monogr. 10: 41–80, 2017

Fig. 28. HPLC-FLD analysis of GTX fraction (A), STX fraction (B), and standard toxins (C and D). Reproduced from Toxicon, 35, Zaman L, Arakawa O, Shimosu A, Onoue Y. Occurrence of paralytic shellfish poison in Bangladeshi freshwater puffers, 423–431, Fig. 6, „ 1997, with permission from Elsevier.

Fig. 29. Schematic drawing of electrophoresis of STX and GTX fractions and standard toxins. Rm: Relative mobility against STX. Modified from Toxicon, 35, Zaman L, Arakawa O, Shimosu A, Onoue Y. Occurrence of paralytic shellfish poison in Bangladeshi freshwater puffers, 423–431, Figs. 4, 5, „ 1997, with permission from Elsevier.

3 and dcGTX2, 3 in the GTX fraction (Fig. 28). How- at that time, was conducted, an unknown component ever, only 14% and 22% of mouse toxicity of these (STX-uk) giving a blue fluorescent spot was detected fractions, respectively, could be explained by the toxin in the STX fraction, and 2 similar unknown compo- amount of these components calculated from the peak nents (GTX-uk1 and GTX-uk2), in the GTX fraction area. In addition, no TTX-related components were (Fig. 29). They were presumed to be the main detected, suggesting existence of unknown components component(s) of respective fraction. undetectable with HPLC-FLD. When an electrophore- The STX fraction was further purified by repetitive sis analysis, which had not already been used so much chromatographies of Bio-Rex 70 column, resulting in

doi:10.5047/absm.2017.01003.0041 © 2017 TERRAPUB, Tokyo. All rights reserved. O. Arakawa et al. / Aqua-BioSci. Monogr. 10: 41–80, 2017 69 isolation of STX-uk (0.5 mg; relative toxicity, ca. 1700 The presence of both toxic and non-toxic wild speci- MU/mmol). In the ESI/MS analysis, STX-uk gave peaks mens in the same species indicates that PSP of T. + + corresponding to [M+H] and [M+H-H2O+MeOH] at turgidus is derived from an exogenous origin, and are m/z 313.9 and 344.2, respectively. 1H-NMR spectrum selectively transferred via the blood into the skin, where of STX-uk showed a close resemblance to that of STX, the toxins accumulate. except that a methyl singlet appeared at 3.69 ppm. This, along with the mol wt deduced from ESI/MS spectrum 5-4. PSP of marine pufferfish indicated that STX-uk would be an N-methyl derivative of STX. A possibility of O-methyl hemiketal for- T. pardalis, T. poecilonotus, and T. vermicularis in- mation at C-12 could be ruled out since the diminish- habiting coastal waters of Sanriku possess PSP mainly ment of a H-11 signal owing to deuterium replacement composed of STX as a minor component, in addition through enolization of the keto group at C-12 (Wong to TTX as a major toxin (Kodama et al. 1983; et al. 1971) was observed in the spectrum. The acid Nakamura et al. 1984). Sato et al. (2000) investigated hydrolyzate of STX-uk was indistinguishable from the toxicity of marine pufferfish from the Philippines, dcSTX in electrophoresis and HPLC analysis suggest- and found that many pufferfish of the genus Arothron ing that the methyl group in STX-uk could be located are highly toxified by PSP mainly comprising STX and at carbamoyl moiety (Fig. 25). GTX-uk1 and GTX- neoSTX. As the PSP composition resembled that of uk2 could possibly be carbamoyl-N-methyl derivatives the dinoflagellate Pyrodinium bahamense var. of GTX2,3, since they showed behaviors similar to compressum occurring in the same sea area (Harada et GTX2,3 in column chromatography and electrophore- al. 1982b), they presumed that these pufferfish accu- sis, and gave no peak in HPLC analysis like STX-uk. mulate toxins of the dinoflagellate via the food organ- The methyl group on carbamoyl moiety appears to pre- isms. In Florida, food poisoning due to Sphoeroides vent the toxin molecules from converting into fluores- pufferfish successively occurred during 2002 to 2004, cent purine derivatives (Quilliam and Janecek 1993) and the causative substance was indicated to be PSP on post-column oxidation in HPLC-FLD analysis, al- mainly composed of STX (Landsberg et al. 2006). The though the mechanisms underlying there remain to be toxin was estimated to come from P. bahamense. elucidated. We investigated the toxicity of “hoshifugu” Arothron firmamentum collected from Oita and Iwate Prefectures 5-3C. Mekong pufferfish (Ngy et al. 2008) (Nakashima et al. 2004). MBA showed that only ovary Wild specimens of the Mekong pufferfish Tetraodon and skin of the female specimens were toxic, the tox- turgidus (n = 72) and Tetraodon sp. (n = 15) were col- icity scores being 5–740 as PSP and <5–30 MU/g as lected from several lakes in Kandal and Phnom Penh, TTX, respectively. Interestingly, HPLC-FLD and LC- Cambodia, respectively, from April 2005 to January MS analyses revealed that STX and dcSTX were the 2006. In the MBA for PSP, only T. turgidus showed major toxins in the ovary, while the skin contained only toxicity. The toxicity was localized in the skin (4–37 TTX. A. firmamentum would have a unique PSP accu- MU/g) and ovary (15–27 MU/g), and the other tissues mulation mechanism in their ovary, and above- (muscle, liver, and intestine) were all non-toxic (less menthioned freshwater pufferfish, in their skin. than 2 MU/g). HPLC-FLD analyses demonstrated that the toxic principle was PSP comprising STX and 6. Toxin other than TTX harbored by pufferfish— dcSTX, which accounted for ª85% of the total toxic- palytoxin (PLTX)-like toxin ity. Neither PSP nor TTX were detected in any tissues of 6-1. Parrotfish poisoning the three artificially reared specimens of T. turgidus by HPLC-FLD for PSP or ELISA for TTX. When PSP 6-1A. Outline (dcSTX, 50 MU/individual) was administered The parrotfish “aobudai” Scarus ovifrons have oc- intramuscularly into the same lot of cultured specimens casionally caused human intoxication in Japan, particu- (body weight, 15.0 ± 2.6 g; body length, 6.0 ± 0.5 cm; larly in the western part including Nagasaki, Kochi, n = 15), toxins transferred via the blood from the mus- Hyogo, and Mie Prefectures (Noguchi et al. 1987b). cle into other body tissues, especially the skin (Fig. The statistics showed that since May 1953, at least 28 30). The majority (92.8%) of the toxin remaining in poisoning incidents due to this fish have occurred, in- the body accumulated in the skin within 48 h. When volving a total of 99 patients and 6 deaths. The most the same dosage of TTX was similarly administered, characteristic symptom of these victims is severe mus- all specimens (n = 15) died within 3 to 4 h, suggesting cle pain arising from rhabdomyolysis, which is usu- that this species is not resistant to TTX. Toxin analysis ally accompanied by myoglobin urea and the abnor- in the dead specimens revealed that more than half of mal elevation of serum creatine phosphokinase (CPK) the administered TTX remained in the muscle and a activity (Okano et al. 1998). Although the causative small amount was transferred into the skin (Fig. 31). substance of the poisoning is still unclear, we call it

Fig. 30. Changes in the content (MU/g) (A), and the total amount (MU/individual) (B) of PSP retained in each tissue of the artificially reared specimens of T. turgidus during the rearing period after toxin administration. *Drawn based on the calculated values, assuming that 100% of the administered toxin was retained in the muscle. Reproduced from Toxicon, 51, Ngy L, Tada K, Yu CF, Takatani T, Arakawa O, Occurrence of paralytic shellfish toxins in Cambodian Mekong pufferfish Tetraodon turgidus: selective toxin accumulation in the skin, 280–288, Fig. 3, „ 2008, with permission from Elsevier. palytoxin (PLTX)-like toxin, because its behavior in cles collected from the home of a patient demonstrated solvent partition and chromatography, pharmacologi- that their sarcoplasmic protein band patterns were very cal actions on smooth muscle preparations from guinea similar to that of “kue” Epinephelus bruneus, but not pig and a nerve-muscle unit from lobster, and delayed perfectly identical. Therefore, the causative fish of the haemolytic activity were very similar to PLTX present poisoning, although it is called “kue” locally, (Noguchi et al. 1987b). was estimated to be a different serranid fish from E. bruneus. 6-1B. Parrotfish poisoning-like poisoning (Taniyama A 60-year-old man, who ate scarid fish caught off et al. 2002) the coast of Mie Prefeture several times from January We had previously considered that parrotfish poison- 19 to 20, 2001, was poisoned after 3 h of latency pe- ing was a distinctive poisoning that only S. ovifrons riod. The main symptom was muscle pain cause, and could be prevented by not eating the fish. (rhabdomyolysis) with an abrupt elevation of serum Poisoning cases in which the victims showed similar CPK activity, and it took about 1 week to recover com- symptoms to those of parrotfish poisoning (parrotfish pletely from the symptoms. He told that the causative poisoning-like poisoning), however, occurred due to fish was a scarid fish called “igami” in this region, but serranid fish in 2000, and scarid fish in 2001. The de- the species could not be identified because all the lefto- tails of the two incidents were as follows. vers had been disposed of. The causative fish of the From October 31 to November 5, 2000, thirty-three present poisoning, however, was estimated to be people ate a large serranid fish caught off the coast of “budai” Calotomus japonicus, since both fish samples Kochi Prefecture, and eleven of them were poisoned. of the same lot as the causative fish obtained from a The latency period was 3–43 h (average 19 h and 33 fish store were C. japonicus. min), and all patients indicated muscle pain and myoglobin urea. In a blood analysis, abnormal elevation 6-1C. Origin of parrotfish toxin (Taniyama et al. of serum CPK activity was observed during the first 2003) three days after the onset of disease, and more than 1 After the poisoning incident due to parrotfish oc- month was needed for recovery from the symptoms. curred in Osaka City in September 1997, we have made An isoelectric focusing analysis of the leftover mus- some attempts to clarify the toxification mechanism

Fig. 31. Anatomic distribution of TTX remaining in the dead specimens of T. turgidus. Reproduced from Toxicon, 51, Ngy L, Tada K, Yu CF, Takatani T, Arakawa O, Occurrence of paralytic shellfish toxins in Cambodian Mekong pufferfish Tetraodon turgidus: selective toxin accumulation in the skin, 280–288, Fig. 4, „ 2008, with permission from Elsevier.

ness, nausea, muscle pain, respiratory arrest, paralysis, vomiting

Dizziness, headache, muscle pain, nausea, fatigue, vomiting

Numbness in lips, dizziness, nausea, muscle pain, vomiting Muscle pain, black urine, dyspnea, vomiting

Respiratory arrest, chest tightness, paralysis, muscle pain, headache Numbness in lips, dyspnea, paralysis, blockage of urinary bladder, vomiting Muscle pain, nausea, respiratory attack, vomiting Dyspnea, paralysis, respiratory arrest, muscle pain, black urine, vomiting

Symptoms appeared Dizziness, paralysis, respiratory attack, hypothermia, salivation

Nausea, dyspnea, muscle pain, stomachache, vomiting of S. ovifrons, in addition to investigating the cause of fer poisoning incidents in Bangladesh. poisoning. Specimens of S. ovifrons were collected in the same sea area where the causative fish had been death caught (offshore of Mugi Town, Tokushima Prefecture), Number of and their toxicity was examined mainly by haemolytic

41 activity test, resulting in detection of PLTX-like tox- 44 icity in the muscle of 2, and the liver of 5 of 6 speci- victim mens. Subsequent examination of the gut contents of Number of the toxic specimens revealed that the fish had eaten a plenty of adherent microalgae together with their

* substrative seaweeds. On the other hand, Dr. 50

400

250

300

(g)

500

250

2,000 7 5 Numb Yoshimatsu, Akashio Research Institute of Kagawa 2, Prefecture, provided us with the interesting informa-

Ingested muscle tion that a huge number of algal cells of the dinoflagellate Ostreopsis sp. had suddenly appeared Morbidity and mortality of fresh water puf along the temperate coast of Mugi town in August 1997

able 9.

(one month before the incident); maximum of 150000 T cells were observed to adhere per 1 g of seaweeds. Therefore, the causative parrotfish was presumed to have ingested a large amount of Ostreopsis cells with seaweeds, and involvement of Ostreopsis in parrotfish toxification was strongly suspected. Ostreopsis are the benthic dinoflagellates mainly distributed in tropical and subtropical regions (Fukuyo 1981), and one of the species, O. siamensis produces PLTX analogues . (2000).

et al

anikganj

unshiganj 10-A

etrokona

ajshahi

arisal

handpur (Ukena et al. 2001; Usami et al. 1995). Then, the hathkhira 23-A

Sylhet Ostreopsis sp. separated from the coast of Mugi toxin Chandpur one month after the incident was cultured in a large scale, and the obtained cells were examined for toxicity. The species was found to produce a delayed-active

4C 6Bho 3N

8B 9 5M 7R

Case No.Case Place of incidentoccurrence of Place

10 1S 2M

*Information not available. –4 From Mahmud toxin (1.0 ¥ 10 MU/cell), whose characteristics were

Fig. 32. Delayed haemolysis using mouse red cell suspension, and inhibition of haemolytic activity with anti-PLTX antibody. (A) Ostreopsis sp. toxin, (B) S. ovifrons toxin, and (C) PLTX (formerly abbreviated as PTX) standard. Reproduced from Toxicon, 42, Taniyama S, Arakawa O, Terada M, Nishio S, Takatani T, Mahmud Y, Noguchi T, Ostreopsis sp., a possible origin of palytoxin (PTX) in parrotfish Scarus ovifrons, 29–33, Fig. 4, „ 2003, with permission from Elsevier. very similar to those of PLTX-like toxin; it exhibited a delayed haemolytic activity suppressed by an anti- PLTX antibody (Fig. 32) and ouabain, and a serum CPK-raising activity (Fig. 33). These facts suggest that the S. ovifrons had accumulated PLTX-like toxin by ingesting a large amount of Ostreopsis sp. abnormally growing in temperate sea area, and become toxic to cause human intoxications in Osaka City. Although no massive occurrence of Ostreopsis has been observed, and no parrotfish have been toxified on the coast of Mugi town after 1998, there might be a latent possibility in the sea area where Ostreopsis are distributed that S. ovifrons and some other fish are toxified.

6-2. Bangladeshi freshwater pufferfish poisoning

6-2A. Epidemic survey (Mahmud et al. 2000) Two species of freshwater pufferfish, including previously mentioned T. cutcutia, live in Bangladesh. Because of a lower price, pufferfish (generally muscle only) are consumed by poor people who have little or Fig. 33. Elevated level of CPK in mice after toxin adminis- no knowledge on toxicity of pufferfish, resulting in tration. Reproduced from Toxicon, 42, Taniyama S, Arakawa sporadic occurrence of food poisoning with some fa- O, Terada M, Nishio S, Takatani T, Mahmud Y, Noguchi T, talities. An epidemic investigation revealed that at least Ostreopsis sp., a possible origin of palytoxin (PTX) in 10 food poisoning incidents involving 55 victims and parrotfish Scarus ovifrons, 29–33, Fig. 3, „ 2003, with per- 17 deaths have occurred due to ingestion of the fresh- mission from Elsevier. water pufferfish Tetraodon sp. from 1988 to 1996 (Ta- ble 9). Two typical cases are as follows. A food poisoning incident due to ingestion of 4–5 persisted for more than 7 days. pufferfish (roughly 50 g) occurred in Manikganj dis- In Bhola district, 10 people from a fisherman family trict on March 15, 1994. After ingestion, 5 of 7 con- were intoxicated after ingestion of roughly 2.5 kg of sumers became intoxicated within 3–5 hours, exhibit- freshwater pufferfish on December 7, 1994. Dyspnea, ing the symptoms of severe respiratory troubles, diz- respiratory arrest, muscle pain, vomiting, and discharge ziness muscle pain, discharge of black urine, vomit- of black urine were the main symptoms of the victims. ing, etc. Three of them died at the end of the following Among the patients, a 35-year-old man and a 40-yea- day, while the other two were recovered in a hospital old woman expired two days later due to respiratory within 2–3 days after admission, although muscle pain failure, while the other 8 patients were hospitalized and

Fig. 34. Inhibitory effect of anti-PLTX antibody (left) and ouabain (right) on the haemolysis of mouse and human erythrocytes, respectively, induced by Tetraodon sp. toxin (upper) and PLTX (formerly abbreviated as PTX) standard (lower). Repro- duced from Toxicon, 39, Taniyama S, Mahmud Y, Tanu MB, Takatani T, Arakawa O, Noguchi T, Delayed haemolytic activity by the freshwater puffer Tetraodon sp. toxin, 725–727, Figs. 1, 2, „ 2001, with permission from Elsevier. provided treatment. Their serum CPK value was high ferent types of toxic elements, PSP and PLTX-like (298–430 IU/L) in the blood analysis, and more than 1 toxin. Freshwater pufferfish harbored a little amount week was necessary for their urine color to return to of PSP, and human intoxication cannot fully be ex- normal. plained only with the toxin. The facts that the main In Bangladeshi freshwater pufferfish poisoning, char- symptom of the patient was muscle pain and myoglobin acteristic symptoms of general pufferfish poisoning urea, which is characteristic to the parrotfish poison- (TTX poisoning) or PSP poisoning such as paralysis ing, and that time of death and recovery time of the and respiratory distress were partly observed, but it was patient were much longer that those of PSP poisoning, clearly different from these poisonings in terms that strongly suggested that the main causative substance the main symptom was muscle pain, and that time of of the freshwater pufferfish poisoning was PLTX-like death or time of recovery was longer than those of TTX/ toxin, like in the parrotfish poisoning. PSP poisoning. 6-3. Boxfish poisoning (Taniyama et al. 2009b) 6-2B. Properties of causative toxin (Taniyama et al. 2001) In Goto Islands, Nagasaki Prefecture, “kattoppo-no- The freshwater pufferfish Tetraodon sp. collected miso-yaki” has long been eaten as a local special cui- from Kishorganj district in Bangladesh in May 1995 sine, in which the muscle and viscera of a boxfish were examined for anatomical distribution of toxicity. kneaded with miso, sake, and spring onion are put into In the skin and viscera, 1.7–5.9 MU/g of PSP mainly the abdominal cavity of the fish, and broiled. From composed of STX and dcSTX was detected. After re- 1990 to 2013, 11 food poisoning incidents due to in- moval of PSP by solvent fractionation, 0.5 MU/g of gestion of the dish occurred in Nagasaki, Miyazaki, delayed lethality to mice was recognized in the PLTX- Mie and Kagoshima Prefectures, and a total of 16 per- soluble fraction. The toxin exhibited delayed haemo- sons were poisoned. Their main symptom was severe lytic activity to mouse and human erythrocytes, which muscle pain, myoglobin urea, and abnormal elevation was suppressed by an anti-PLTX antibody or ouabain of serum CPK activity arising from rhabdomyolysis. (Fig. 34), indicating that the pufferfish possess two dif- Fifteen out of the 16 victims recovered in a few days

Table 10. Anatomic distribution of toxicity in toxic specimens of O. immaculatus collected from Western Japan.

No. Place of collection Month of collection Body weight Body length Lethal potency (MU/g) (g) (mm) Muscle Liver Viscera excluding liver

1 01 Arikawa Bay, Nov.-Dec. 2003 563.0 227 0.5* ND ND 1 02 Nagasaki Prefecture ≤ 575.1 206 0.5* ND ND 2 03 ≤ 494.0 239 0.5* ND ND 1 1 1 04 Nov.-Dec. 2004 220.0 153 0.5* 0.5* 1.0* 2 1 1 05 ≤ 115.0 145 0.5* 0.5* 1.0* 1 1 06 ≤ 270.0 160 ND 0.5* 0.5* 1 1 07 ≤ 130.0 152 ND 0.5* 0.5* 1 08 ≤ 358.3 188 0.5* ND ND 2 09 ≤ 634.0 235 0.5* ND ND 1 10 ≤ 205.0 153 ND ND 0.5* 1 11 ≤ 455.0 195 ND ND 0.5* 1 12 ≤ 410.0 187 NE NE 0.5* 2 13 ≤ 279.4 178 ND ND 0.5* ≤ 14 Shimaura Island, May 2004 300.0 133 0.5*1 ND ND 2 15 Miyazaki Prefecture ≤ 191.3 145 ND ND 1.0*

16 Offshore of Mugi, Nov. 2004 267.1 158 ND ND 0.5*1 17 Tokushima Prefecture Jun. 2005 237.4 160 ND ND 1.0*1 1 18 ≤ 259.0 160 ND ND 0.5* 2 19 ≤ 315.0 183 ND ND 0.5* 2 20 ≤ 190.3 146 ND ND 0.5*

21 Offshore of Shimonoseki, Dec. 2004 317.0 168 ND ND 0.5*2 2 22 Yamaguchi Prefecture ≤ 235.3 155 ND ND 0.5* 23 Apr. 2005 603.1 216 ND ND 0.5*2 2 24 ≤ 554.0 202 ND ND 0.5* 25 Jun. 2005 224.0 142 0.5*1 ND 0.5*1 1 2 26 ≤ 627.0 209 0.5* ND 0.5* 2 27 ≤ 395.4 180 ND ND 0.5* 28 Jul. 2005 082.5 101 0.5*1 ND 0.5*1 2 1 29 ≤ 153.0 127 ND 0.5* 0.5* 1 30 ≤ 336.0 172 ND ND 0.5* 2 31 ≤ 117.0 123 ND ND 0.5* 2 32 ≤ 106.2 110 ND ND 0.5* 2 33 ≤ 208.3 142 ND ND 0.5* 34 Aug. 2005 411.0 203 ND ND 0.5*1 2 35 ≤ 405.1 179 ND ND 0.5* 36 Oct. 2005 302.4 159 0.5*2 ND 0.5*1 1 37 ≤ 552.0 204 0.5* ND ND 2 38 ≤ 861.3 215 0.5* ND ND 1 39 ≤ 469.4 187 ND 0.5* ND 2 40 ≤ 288.4 174 ND ND 0.5* 41 Nov. 2005 535.4 192 ND 0.5*1 0.5*1 2 42 ≤ 089.2 110 ND ND 0.5* 2 43 ≤ 777.2 206 ND ND 0.5* 44 Dec. 2005 668.0 198 ND ND 0.5*1 45 May 2006 458.0 178 ND 0.5*1 ND 2 46 ≤ 750.0 208 ND ND 0.5* 2 47 ≤ 458.0 184 ND ND 0.5*

ND: not detected (< 0.5 MU/g); NE: not examined. *1 Delayed lethal potency to mice. *2 Acute lethal potency to mice. From Taniyama et al. (2009b).

Table 11. Anatomic distribution of toxicity in toxic specimens of L. diaphana collected from Western Japan.

No. Place of collection Month of Body weight Body length Lethal potency (MU/g) collection (g) (mm) Muscle Liver Viscera excluding liver 1 Shimaura Island, May 2004 391.0 163 0.5*2 ND ND 2Miyazaki Prefecture 711.2 240 ND ND 0.5*2 3 541.0 237 ND ND 0.5*2 4 272.0 190 ND ND 0.5*1

5Offshore of Mugi, Jun. 2005 439.0 235 ND 0.5*2 2.0*1 6 Tokushima Prefecture Mar. 2006 NE 133 ND ND 1.0*2 7NE178 ND NE 1.0*2

ND: not detected (< 0.5 MU/g); NE: not examined. *1 Delayed lethal potency to mice. *2 Acute lethal potency to mice. From Taniyama et al. (2009b).

to two months, while one died after approximately 2 phokinase (CPK; 8,487 U/L), glutamic oxaloacetic weeks. Since the symptoms were very similar to those transaminase (GOT; 371 U/L), lactate dehydrogenase caused by the parrotfish poisoning, the causative sub- (LDH; 866 U/L), creatine kinase MB fraction (CKMB, stance was estimated to be PLTX-like toxin. Epidemic 81 U/L), and myoglobin (more than 1000 mg/L). The surveys after the incidents in Miyazaki and Nagasaki patient was recovered from these symptoms after 5 revealed that the leftovers were identified “hakofugu” days, and discharged from the hospital. Ostracion immaculatus. During screening tests to In the crayfish poisonings, muscle pain, chest pain, clarify the toxicity of boxfish from western Japan, 47 and dorsal pain with abnormal CPK value appeared in of 129 specimens (36.4%) of O. immaculatus, and 7 of all, numbness of the whole body in 2, muscle rigidity 18 specimens (38.9%) of “umisuzume” Lactoria in 1, respiratory distress in 1 patient, but no nausea, diaphana were found to show acute and/or delayed diarrhea, and bloody urine were recognized. Thus, the lethal activity to mice (0.5–2.0 MU/g) (Tables 10, 11). main symptoms of crayfish poisoning are similar to Among the tissues tested, the frequency of toxicity was those of the parrotfish poisoning, but the research group highest in the viscera excluding liver (28.6% in O. involved in this poisoning concluded it to be Haff dis- cubicus, 33.3% in L. diaphana), followed by muscle ease. (10.9%, 5.6%) and liver (6.2%, 5.6%). From the re- Haff disease is a disease of unknown etiology in sults, we conclude that O. cubicus and L. diaphana which human develop rhabdomyolysis within 24 hours inhabiting the coast of Japan sometimes possess PLTX- after ingesting specific species of fish (Zu jeddeloh like toxin, which cause food poisonings very similar 1939). In the disease, no nervous abnormality, fever, to parrotfish poisoning. megalosplenia, and hepatomegalia were observed in the patients, many of whom stayed alive with no after- 6-4. Haff disease effect, but some were dead. Cases of Haff disease have been known in Sweden, Russia (former Soviet Union), In August 2010, food poisonings due to ingestion of USA, and Brazil since it was first recognized at Haff crayfish occurred successively in Nanjing, China, in beach on the coast of the Baltic Sea in 1924 (Buchholz which 5 patients were poisoned with rhabdomyolysis et al. 2000; dos Santos et al. 2009; Lahgley and Bobbitt as a main symptom (Zhang et al. 2012). A typical case 2007; Zu jeddeloh 1939). Although Haff disease is is as follows. similar to PLTX-like toxin poisoning in terms that it is A woman, who cooked crayfish purchased from a caused by ingesting specific fish, and that the main seafood market and ate it for dinner, showed symp- symptom is rhabdomyolysis, the relationship between toms of diffuse myalgia, chest pain, shortness of breath, both diseases remains to be elucidated. numbness of the whole body, and muscle rigidity after 5 hours, and was admitted to a local hospital. At that 7. Conclusion time, no nausea, vomiting, diarrhea, and abdominal pain were recognized in the patient, and physicochemi- As described above, we have expanded our re- cal examinations indicated no abnormality. Abnormal searches on natural toxins harbored by pufferfish from scores, however, were seen in the serum creatine phos- various viewpoints. Currently, several studies are on-

doi:10.5047/absm.2017.01003.0041 © 2017 TERRAPUB, Tokyo. All rights reserved. 76 O. Arakawa et al. / Aqua-BioSci. Monogr. 10: 41–80, 2017 going to clarify physiologic functions of TTX, rela- Lescano AG, Santos JHA, Alecrim MdGC. Outbreak of tionship between its internal kinetics and growth/matu- Haff disease in the Brazilian Amazon. Rev. Panam. Salud ration, and molecular mechanisms that are involved in Publica 2009; 26: 469–470. pufferfish, toxin accumulation ability of hybrid Endo A, Khora SS, Murata M, Naoki H, Yasumoto T. Isola- pufferfish, and characteristics of PLTX-like toxin, and tion of 11-nortetrodotoxin-6(R)-ol and other tetrodotoxin derivatives from the puffer Fugu niphobles. Terahedron it is gradually being elucidated that 1) TTX may pos- Lett. 1988; 29: 4123–4128. sibly participate in the control of information trans- Fukuyo Y. Taxonomical study on benthic dinoflagellates mission (behavioral regulation) in the central nervous collected in coral reefs. Bull. Jpn. Soc. Sci. Fish. 1981; system of pufferfish (Okita et al. 2013a, b), 2) internal 47: 967–978. kinetics of TTX changes depending on the individual Goto T, Takahashi S, Kishi Y, Hirata Y. Tetrodotoxin. Tetra- growth and/or maturation in TTX-bearing fish (Tatsuno hedron 1965; 21: 2059–2088. et al. 2013a, b), 3) as a candidate for molecules in- Harada T, Oshima Y, Yasumoto T. Structures of two para- volved in the mechanism whereby the change occurs, lytic shellfish toxins, gonyautoxins V and VI, isolated from T. rubripes possess 4 types of isoform genes homolo- a tropical dinoflagellate, Pyrodinium bahamense var. gous with the TTX-binding protein, PSTBP, which compressa. Agric. Biol. Chem. 1982a; 46: 1861–1864. Yotsu-Yamashita et al. (2001) found from T. pardalis Harada T, Oshima Y, Kamiya H, Yasumoto T. Confirmation of paralytic shellfish toxins in the dinoflagellate (Tatsuno et al. 2013c), and 4) in hybrid pufferfish, toxic Pyrodinium bahamense var. compressum and bivalves in parts may be different when mother/father is reversed Palau. Bull. Jpn.. Soc. Sci. Fish. 1982b; 48: 821–825. even if combination of the parent species is the same Harada T, Oshima Y, Yasumoto T. Natural occurrence of (unpublished). We hope for the efforts and develop- decarbamoylsaxitoxin in tropical dinoflagellate and ments of young researchers and students who will bivalves. Agric. Biol. Chem. 1983; 47: 191–193. shoulder these researches that face the future. Honda S, Arakawa O, Takatani T, Tachibana K, Yagi M, Tanigawa A, Noguchi T. Toxification of cultured puffer References fish Takifugu rubripes by feeding on tetrodotoxin-contain- Ali AE, Arakawa O, Noguchi T, Miyazawa K, Shida Y, ing diet. Nippon Suisan Gakkaishi 2005a; 71: 815–820. Hashimoto K. Tetrodotoxin and related substances in a Honda S, Ichibu T, Arakawa O, Takatani T, Tachibana K, ribbon worm Cephalothrix linearis (Nemertean). Toxicon Yagi M, Tanigawa S, Noguchi T. Antibody productivity 1990; 28: 1083–1093. against sheep red blood cells and splenocyte proliferation Arakawa O, Noguchi T, Shida Y, Onoue Y. Occurrence of reaction of tiger puffer (Takifugu rubripes) fed with tetro- 11-oxotetrodotoxin and 11-nortetrodotoxin-6(R)-ol in a dotoxin-containing diets. Aquaculture Sci. 2005b; 53: 205– xanthid crab Atergatis floridus collected at Kojima, 210. Ishigaki Island. Fish. Sci. 1994a; 60: 769–771. Honma Y, Ozawa T, Chiba A. Maturation and spawning Arakawa O, Noguchi T, Shida Y, Onoue Y. Occurrence of behavior of the puffer, Fugu niphobles, occurring on the carbamoyl-N-hydroxy derivatives of saxitoxin and coast of Sado Island in the Sea of Japan (a preliminary neosaxitoxin in a xanthid crab Zosimus aeneus. Toxicon report). Japan. J. Ichthyol. 1980; 27: 129–138. 1994b; 32: 175–183. Hsu CP, Marchand A, Shimizu Y, Sims GG. Paralytic shell- Arakawa O, Nishio S, Noguchi T, Shida Y, Onoue Y. A new fish toxins in sea scallops, Placopecten magellanicus in saxitoxin analogue from a xanthid crab Atergatis floridus. the Bay of Fundy. J. Fish Res. Board Can. 1979; 36: 32. Toxicon 1995; 33: 1577–1584. Hwang DF, Noguchi T. Tetrodotoxin poisoning. Adv. Food Arakawa O, Hwang DF, Taniyama S, Takatani T. Toxins of Nat. Res. 2007; 52: 141–236. pufferfish that cause human intoxications. In: Ishimatsu Hwang DF, Arakawa O, Saito T, Noguchi T, Simidu U, A, Lie HJ (eds). Coastal Environmental and Ecosystem Tsukamoto K, Shida Y, Hashimoto K. Tetrodotoxin-pro- Issues of the East China Sea. Nagasaki University/ ducing bacteria from the blue-ringed octopus Octopus TERRAPUB, Tokyo. 2010; 227–244. maculosus. Mar. Biol. 1989; 100: 327–332. Asakawa M, Ito K, Kajihara H. Highly toxic ribbon worm Ikeda K, Murakami Y, Emoto Y, Ngy L, Taniyama S, Yagi Cephalothrix simula containing tetrodotoxin in Hiroshima M, Takatani T, Arakawa O. Transfer profile of Bay, Hiroshima Prefecture, Japan. Toxins 2013; 5: 376– intramuscularly administered tetrodotoxin to non-toxic 395. cultured specimens of the pufferfish Takifugu rubripes. Bentur Y, Ashkar J, Lurie Y, Levy Y, Azzam ZS, Limanovich Toxicon 2009; 53: 99–103. M, Golik M, Gurevych B, Golani D, Eisenman A. Ikeda K, Emoto Y, Tatsuno R, Wang JJ, Ngy L, Taniyama S, Lessepsian migration and tetrodotoxin poisoning due to Takatani T, Arakawa O. Maturation-associated changes in Lagocephalus sceleratus in the eastern Mediterranean. toxicity of the pufferfish Takifugu poecilonotus. Toxicon Toxicon 2008; 52: 964–968. 2010; 55: 289–297. Brillantes S, Samosorn W, Faknoi S, Oshima Y. Toxicity of Isbister GK, Son J, Wang F, Maclean CJ, Lin CSY, Ujma J, puffers landed and marketed in Thailand. Fish. Sci. 2003; Balit CR, Smith B, Milder DG, Kiernan MC. Pufferfish 69: 1224–1230. poisoning: a potentially life-threatening condition. MJA Buchholz U, Mouzin E, Dickey R, Moolenaar R, Sass N, 2002; 177: 650–653. Mascola L. Haff disease: from the Baltic Sea to the US Islam QT, Razzak MA, Islam MA, Bari MI, Basher A, shore. Emerg. Infect. Dis. 2000; 6: 192–195. Chowdhury FR, Sayeduzzaman ABM, Ahasan HAMN, dos Santos MC, de Albuquerque BC, Pinto RC, Aguiar GP, Faiz MA, Arakawa O, Yotsu-Yamashita M, Kuch U, Mebs

D. Puffer fish poisoning in Bangladesh: clinical and toxi- Matsui T, Hamada S, Konosu S. Difference in accumulation cological results from large outbreaks in 2008. Transact. of puffer fish toxin in crystalline tetrodotoxin in the puffer Roy. Soc. Trop. Med. Hyg. 2011; 105: 74–80. fish, Fugu rubripes rubripes. Bull. Jpn.. Soc. Sci. Fish. Itoi S, Yoshikawa S, Asahina K, Suzuki M, Ishizuka K, 1981; 47: 535–537. Takimoto N, Mitsuoka R, Yokoyama N, Detake A, Matsui T, Taketsugu S, Sato H, Yamamori K, Kodama K, Takayanagi C, Eguchi M, Tatsuno R, Kawane M, Kokubo Ishi A, Hirose H, Shimizu C. Toxification of cultured S, Takanashi S, Miura A, Suitoh K, Takatani T, Arakawa puffer fish by the administration of tetrodotoxin produc- O, Sakakura Y, Sugita H. Larval pufferfish protected by ing bacteria. Nippon Suisan Gakkaishi 1990; 56: 705. maternal tetrodotoxin. Toxicon 2014; 78: 35–40. Matsui T, Yamamori K, Furukawa K, Kono M. Purification Kawatsu K, Hamano Y, Yoda T, Terano Y, Shibata T. Rapid and some properties of a tetrodotoxin binding protein from and highly sensitive enzyme immunoassay for quantita- the blood plasma of kusafugu, Takifugu niphobles. Toxicon tive determination of tetrodotoxin. Jpn. J. Med. Sci. Biol. 2000; 38: 463–468. 1997; 50: 133–150. Matsumoto T, Nagashima Y, Takayama K, Shimakura K, Khora SS, Yasumoto T. Isolation of 11-oxotetrodotoxin from Shiomi K. Difference between tetrodotoxin and saxitoxin the puffer Arothron nigropunctatus. Tetrahedron Lett. in accumulation in puffer fish Takifugu rubripes liver tis- 1989; 30: 4393–4394. sue slices. Fish Physiol. Biochem. 2005; 31: 95–100. Kobayashi M, Shimizu Y. Gonyautoxin VIII, a cryptic pre- Matsumoto T, Nagashima Y, Kusuhara H, Sugiyama Y, cursor of paralytic shellfish poisons. J. Chem. Soc. Chem. Ishizaki S, Shimakura K, Shiomi K. Involvement of car- Commun. 1981; 16: 827–828. rier-mediated transport system in uptake of tetrodotoxin Kodama M, Ogata T. Toxicity of a freshwater puffer into liver tissue slices of puffer fish Takifugu rubripes. Tetraodon leiurus. Bull. Jpn. Soc. Sci. Fish. 1984; 50: Toxicon 2007; 50: 173–179 1949–1951. Matsumoto T, Nagashima Y, Kusuhara H, Ishizaki S, Kodama M, Ogata T, Noguchi T, Maruyama J, Hashimoto Shimakura K, Shiomi K. Pharmacokinetics of tetrodotoxin K. Occurrence of saxitoxin and other toxins in the liver of in puffer fish Takifugu rubripes by a single administra- pufferfish Takifugu pardalis. Toxicon 1983; 21: 897–900. tion technique. Toxicon 2008a; 51; 1051–1059. Kodama M, Ogata T, Sato S. External secretion of tetrodo- Matsumoto T, Nagashima Y, Kusuhara H, Ishizaki S, toxin from puffer fishes stimulated by electric shock. Mar. Shimakura K, Shiomi K. Evaluation of hepatic uptake Biol. 1985; 87: 199–203. clearance of tetrodotoxin in the puffer fish Takifugu Kodama M, Sato S, Ogata T, Suzuki Y, Kaneko T, Aida K. rubripes. Toxicon 2008b; 52: 369–374. Tetrodotoxin secreting glands in the skin of puffer fishes. Matsumura K. Tetrodotoxin as a pheromone. Nature 1995; Toxicon 1986; 24: 819–829. 378: 563–564. Kungsuwan A, Noguchi T, Arakawa O, Simidu U, Tsukamoto McNabb P, Selwood AI, Munday R, Wood SA, Taylor DI, K, Shida Y, Hashimoto K. Tetrodotoxin-producing bacte- MacKenzie LA, Ginkel R, Rhodes LL, Cornelisen C, ria from the horseshoe crab Carcinoscorpius rotundicauda. Heasman K, Holland PT, King C. Detection of tetrodotxin Nippon Suisan Gakkaishi 1988; 54: 1799–1802. from the grey side-gilled sea slug Pleurobranchaea macu- Kungsuwan A, Arakawa O, Promdet M, Onoue Y. Occur- late, and associated dog neurotoxicosis on beaches adja- rence of paralytic shellfish poisons in Thai fresh water cent to the Hauraki Gulf, Auckland, New Zealand. Toxicon puffers. Toxicon 1997; 35: 1341–1346. 2010; 56: 466–473. Lahgley RL, Bobbitt WH. Haff disease after eating salmon. Miyazawa K, Noguchi T. Distribution and origin of tetrodo- South. Med. J. 2007; 100: 1147–1150. toxin. J. Toxicol.-Toxin Rev. 2001; 20: 11–33. Landsberg JH, Hall S, Johannessen JN, White KD, Conrad Miyazawa K, Higashiyama M, Ito K, Noguchi T, Arakawa SM. Saxitoxin puffer fish poisoning in the United States, O, Shida Y, Hashimoto K. Tetrodotoxin in two species of with the first report of Pyrodinium bahamense as the pu- ribbon worm (Nemertini), Lineus fuscoviridis and tative toxin source. Environ. Health Persp. 2006; 114: Tubulanus punctatus. Toxicon 1988; 26: 867–874. 1502–1507. Mosher HS, Fuhrman FA, Buchwald HD, Fischer HG. Lim PT, Sato S, Thuoc CV, Tu PT, Huyen NTM, Takata Y, Tarichatoxin-tetrodotoxin, a potent neurotoxin. Science Yoshida M, Kobiyama A, Koike K, Ogata T. Toxic 1965; 144: 1100–1110. Alexandrium minutum (Dinophyceae) from Vietnam with Nagashima Y, Maho T, Hasobe M, Shimakura K, Shiomi K. new gonyautoxin analogue. Harmful Algae 2007; 6: 321– In vitro accumulation of tetrodotoxin in pufferfish liver 331. tissue slices. Toxicon 2003; 41: 569–574. Mahmud Y, Arakawa O, Noguchi T. An epidemic survey on Nagashima Y, Mataki I, Toyoda M, Nakajima H, Tsumoto freshwater puffer poisoning in Bangladesh. J. Natural K, Shimakura K, Shiomi K. Change in tetrodotoxin con- Toxins 2000; 9: 319–326. tent of puffer fish Takifugu rubripes during seed produc- Mahmud Y, Okada K, Takatani T, Kawatsu K, Hamano Y, tion from fertilized eggs to juveniles. Food Hyg. Saf. Sci. Arakawa O, Noguchi T. Intra-tissue distribution of tetro- 2010; 51: 48–51. dotoxin in two marine puffers Takifugu vermicularis and Nakamura M, Oshima Y, Yasumoto T. Occurrence of saxi- Chelonodon patoca. Toxicon 2003a; 41: 13–18. toxin in puffer fish. Toxicon 1984; 22: 381–385. Mahmud Y, Arakawa O, Ichinose A, Tanu MB, Takatani T, Nakashima K, Arakawa O, Taniyama S, Nonaka M, Takatani Tsuruda K, Kawatsu K, Hamano Y, Noguchi T. Intracellu- T, Yamamori K, Fuchi Y, Noguchi T. Occurrence of lar visualization of tetrodotoxin (TTX) in the skin of a saxitoxins as a major toxin in the ovary of marine puffer puffer Tetraodon nigroviridis by immunoenzymatic tech- Arothron firmamentum. Toxicon 2004; 43: 207–212. nique. Toxicon 2003b; 41: 605–611. Narahashi T. Pharmacology of tetrodotoxin. J. Toxicol.-Toxin

Rev. 2001; 20: 67–84. Okita K, Takatani T, Nakayasu J, Yamazaki H, Sakiyama K, Negri A, Stirling D, Quilliam M, Blackburn S, Bolch C, Ikeda K, Arakawa O, Sakakura Y. Comparison of the lo- Burton I, Eaglesham G, Thomas T, Walter J, Willis R. calization of tetrodotoxin between wild pufferfish Takifugu Three nevel hydroxybenzoate saxitoxin analogues isolated rubripes juveniles and hatchery-reared juveniles with tet- from the dinoflagellate Gymnodinium catenatum. Chem. rodotoxin administration. Toxicon 2013b; 71: 128–133. Res. Toxicol. 2003; 16: 1029–1033. Onodera H, Satake M, Oshima Y, Yasumoto T. New saxi- Ngy L, Tada K, Yu CF, Takatani T, Arakawa O. Occurrence toxin analogues from the freshwater filamentous of paralytic shellfish toxins in Cambodian Mekong cyanobacterium Lyngbya wollei. Natural Toxins 1997; 5: pufferfish Tetraodon turgidus: selective toxin accumula- 146–151. tion in the skin. Toxicon 2008; 51: 280–288. Oshima Y, Blackburn SI, Hallegraeff GM. Comparative study Noguchi T, Arakawa O. Tetrodotoxin—distribution and ac- on paralytic shellfish toxin profiles of the dinoflagellate cumulation in aquatic organisms, and cases of human in- Gymnodinium catenatum from three different countries. toxication. Mar. Drugs 2008; 6: 220–242. Mar. Biol. 1993; 116: 471–476. Noguchi T, Ebesu JSM. Puffer poisoning: epidemiology and Quilliam MA, Janecek M. Characterization of the oxidation treatment. J. Toxicol.-Toxin Rev. 2001; 20: 1–10. products of paralytic shellfish poisoning toxins by liquid Noguchi T, Ueda Y, Hashimoto K, Seto H. Isolation and char- chromatography/mass spectrometry. Rapid Communica- acterization of gonyautoxin-1 from the toxic digestive tions in Mass Spectroscopy 1993; 7: 482–487. gland of scallop Patinopecten yessoensis. Bull. Jpn.. Soc. Radriguez P, Alfonso A, Vale C, Alfonso C, Vale P, Tellez A, Sci. Fish. 1981; 47: 1227–1231. Botana LM. First toxicity report of tetrodotoxin and 5,6,11- Noguchi T, Jeon JK, Arakawa O, Sugita H, Deguchi Y, Shida trideoxy TTX in the trumpet shell Charonia lampas lampas Y, Hashimoto K. Occurrence of tetrodotoxin and in Europe. Anal. Chem. 2008; 80: 5622–5629. anhydrotetrodotoxin in Vibrio sp. isolated from the intes- Ritson-Williams R, Yotsu-Yamashita M, Paul V. Ecological tines of a xanthid crab, Atergatis floridus. J. Biochem. functions of tetrodotoxin in a deadly polyclad flatworm. 1986a; 99: 311–314. Proc. Natl. Acad. Sci. USA 2006; 103: 3176–3179.

Noguchi T, Arakawa O, Daigo K, Hashimoto K. Local dif- Rogers RS, Rapoport H. The pKas of saxitoxin. J. Am. Chem. ferences in toxin composition of a xanthid crab Atergatis Soc. 1980; 102: 7335–7339. floridus inhabiting Ishigaki Island, Okinawa. Toxicon Saito T, Noguchi T, Harada T, Murata O, Hashimoto K. Tet- 1986b; 24: 705–711. rodotoxin as a biological defense agent for puffers. Bull. Noguchi T, Hwang DF, Arakawa O, Sugita H, Deguchi Y, Jpn. Soc. Sci. Fish. 1985; 51: 1175–1180. Shida Y, Hashimoto K. Vibrio alginolyticus, a tetrodotoxin- Saito T, Kageyu K, Goto H, Murakami K, Noguchi T. Tetro- producing bacterium, in the intestines of the fish Fugu dotoxin attracts pufferfish (“torafugu” Takifugu rubripes). vermicularis vermicularis. Mar. Biol. 1987a; 94: 625–630. Bull. Inst. Oceanic Res. Develop. Tokai Univ. 2000; 21: Noguchi T, Hwang DF, Arakawa O, Daigo K, Sato S, Ozaki 93–96. H, Kawai N, Ito M, Hashimoto K. Palytoxin as the causa- Sato S, Kodama M, Ogata T, Saitanu K, Furuya M, Hirayama tive agent in the parrotfish poisoning. In: K, Kamimura K. Saxitoxin as a toxic principle of a fresh- Gopalakrishnakone P, Tan CK (eds). Progress in Venom water puffer, Tetraodon fangi, in Thailand. Toxicon 1997; and Toxin Research. National University of Singapore, 35: 137–140. Singapore. 1987b; 325–335. Sato S, Ogata T, Borja V, Gonzales C, Fukuyo Y, Kodama Noguchi T, Ali AE, Arakawa O, Miyazawa K, Kanoh S, Shida M. Frequent occurrence of paralytic shellfish poisoning Y, Nishio S, Hashimoto K. Tetrodonic acid-like substance; toxins as dominant toxins in marine puffer from tropical a possible precursor of tetrodotoxin. Toxicon 1991; 29: water. Toxicon 2000; 38: 1101–1109. 845–855. Schantz EJ, Mold J, Stanger D, Shavel J, Riel F, Bowden J, Noguchi T, Aarakawa O, Takatani T. Toxicity of the Lynch J, Wyler R, Riegel B, Sommer H. Paralytic shell- pufferfish Takifugu rubripes cultured in netcages at sea or fish poison VI. A procedure for isolation of the poison aquaria on land. CBP Part D 2006; 1: 153–157. from toxic clam and mussel tissues. J. Am. Chem. Soc. Noguchi T, Onuki K, Arakawa O. Tetrodotoxin poisoning 1957; 79: 5230–5235. due to pufferfish and gastropods, and their intoxication Schantz, EJ, Ghazarossian VE, Schnoes HK, Strong FM, mechanism. ISRN Toxicol. 2011a; 2011: Article ID 276939, Springer JP, Pezzanite JO, Clardy JJ. Structure of saxi- 10 pages. toxin. J. Am. Chem. Soc. 1975; 97: 1238–1239. Noguchi T, Miyazawa K, Daigo K, Arakawa O. Paralytic Simidu U, Noguchi T, Hwang DF, Shida Y, Hashimoto K. shellfish poisoning (PSP) toxin- and/or tetrodotoxin-con- Marine bacteria which produce tetrodotoxin. Appl. taminated crabs and food poisoning by them. Toxin Rev. Environ. Microbiol. 1987; 53: 1714–1715. 2011b; 30: 91–102. Shimizu Y, Buckley LJ, Alam M, Oshima Y, Fallon WE, Okano H, Masuoka H, Kamei S, Seko T, Koyabu S, Kasai H, Miura I, Vincent P, Gullo VP, Nakanishi K. Struc- Tsunemoto K, Tamiya T, Ueda K, Nakazawa S, Sugawa tures of gonyautoxin II and III from the east coast toxic M, Suzuki H, Watanabe M, Yatani R, Nakano T. dinoflagellate Gonyaulax tamarensis. J. Am. Chem. Soc. Rhabdomyolysis and myocardial damage induced by 1976; 98: 5414–5416. palytoxin, a toxin of blue humphead parrotfish. Intern. Shimizu Y, Hsu CP, Fallon WE, Oshima Y, Miura I, Med. 1998; 37: 330–333. Nakanishi K. Structure of neosaxitoxin. J. Am. Chem. Soc. Okita K, Yamazaki H, Sakiyama K, Yamane H, Niina S, 1978; 100: 6791–6793. Takatani T, Arakawa O, Sakakura Y. Puffer smells tetro- Shimizu Y, Hsu CP, Genenah A. Structure of saxitoxin in dotoxin. Ichthyol. Res. 2013a; 60: 386–389. solutions and stereochemistry of dihydrosaxitoxins. J. Am.

Chem. Soc. 1981; 103: 605–609. Toxicon 2013a; 63: 64–69. Shimizu D, Sakiyama K, Takahashi Y. Predation of stocked Tatsuno R, Shikina M, Shirai Y, Wang J, Soyano K, Nishihara hatchery-reared juveniles of ocellate puffer Takifugu GN, Takatani T, Arakawa O. Change in the transfer pro- rubripes in salt pond mesocosm. Nippon Suisan Gakkaishi file of orally administered tetrodotoxin to non-toxic cul- 2006; 72: 886–893. tured pufferfish Takifugu rubripes depending of its devel- Shimizu D, Sakiyama K, Sakakura Y, Takatani T, Takahashi opment stage. Toxicon 2013b; 65: 76–80. Y. Predation differences between wild and hatchery-reared Tatsuno R, Yamaguchi K, Takatani T, Arakawa O. RT-PCR- tiger puffer Takifugu rubripes juveniles in a salt pond and MALDI-TOF mass spectrometry-based identification mesocosm. Nippon Suisan Gakkaishi 2007; 73: 461–469. and discrimination of isoforms homologous to pufferfish Silva CCP, Zannin M, Rodrigues DS, Santos CR, Correa IA, saxitoxin- and tetrodotoxin-binding protein in the plasma Junior VH. Clinical and epidemiological study of 27 of non-toxic cultured pufferfish (Takifugu rubripes). poisonings caused by ingesting puffer fish (Tetrodontidae) Biosci. Biotechnol. Biochem. 2013c; 77: 208–212. in the states of Santa Catarina and Bahia, Brazil. Rev. Inst. Toda M, Uneyama C, Toyofugu H, Morikawa K. Trends of Med. Trop. S. Paulo 2010; 52: 8 pages. food poisonings caused by natural toxins in Japan, 1989– Simidu U, Noguchi T, Hwang DF, Shida Y, Hashimoto K. 2011. J. Food Hyg. Soc. Japan 2012; 53: 105–120. Marine bacteria which produce tetrodotoxin. Appl. Tsuda K, Ikuma S, Kawamura M, Tachikawa R, Sakai K, Environ. Microbiol. 1987; 53: 1714–1715. Tamura C, Amakasu D. Tetrodotoxin. VII. On the struc- Taniguchi K, Takao H, Nina S, Yamanaka Y, Okada Y, tures of tetrodotoxin and its derivatives. Chem. Pharm. Nakajima R, Wang J, Tatsuno R, Sakakkura Y, Takatani T, Bull. 1964; 12: 1357–1374. Arakawa O, Noguchi T. Distribution of toxicity in liver of Tsuruda K, Arakawa O, Kawatsu K, Hamano Y, Takatani T, wild pufferfish Takifugu rubripes. J. Food Hyg. Soc. Ja- Noguchi T. Secretory glands of tetrodotoxin in the skin of pan 2013; 54: 277–281. a Japanese newt Cynops pyrrhogaster. Toxicon 2002; 40: Taniyama S, Mahmud Y, Tanu MB, Takatani T, Arakawa O, 131–136. Noguchi T. Delayed haemolytic activity by the freshwa- Ukena T, Satake M, Usami M, Oshima Y, Naoki H, Fujita T, ter puffer Tetraodon sp. toxin. Toxicon 2001; 39: 725–727. Kan Y, Yasumoto T. Structure elucidation of Ostreocin D, Taniyama S, Mahmud Y, Terada M, Takatani T, Arakawa O, a palytoxin analog isolated from the dinoflagellate Noguchi T. Occurrence of a food poisoning incident by Ostreopsis siamensis. Biosci. Biotechnol. Biochem. 2001; palytoxin from a serranid Epinephelus sp. in Japan. J. 65: 2585–2588. Natural Toxins 2002; 11: 277–282. Usami M, Satake M, Ishida S, Inoue A, Kan Y, Yasumoto T. Taniyama S, Arakawa O, Terada M, Nishio S, Takatani T, Palytoxin analogs from the dinoflagellate Ostreopsis Mahmud Y, Noguchi T. Ostreopsis sp., a possible origin siamensis. J. Am. Chem. Soc. 1995; 177: 5389–5390. of palytoxin (PTX) in parrotfish Scarus ovifrons. Toxicon Wang J, Araki T, Tatsuno R, Nina S, Ikeda K, Hamasaki M, 2003; 42: 29–33. Sakakura Y, Takatani T, Arakawa O. Transfer profile of Taniyama S, Isami Y, Matusmoto T, Nagashima Y, Takatani intramuscularly administered tetrodotoxin to artificial T, Arakawa O. Toxicity and toxin profile of tetrodotoxin hybrid specimens of pufferfish, Takifugu rubripes and detected in the scavenging gastropod Nassarius (Alectrion) Takifugu niphobles. Toxicon 2011; 58: 565–569. glans “kinshibai”. J. Food Hyg. Soc. Japan 2009a; 50: Wang J, Araki T, Tatsuno R, Nina S, Ikeda K, Takatani T, 22–28. Arakawa O. Transfer profile of orally and intramuscularly Taniyama S, Sagara T, Nishio S, Kuroki R, Asakawa M, administered tetrodotoxin to artificial hybrid specimens Miyazawa K, Noguchi T, Yamasaki S, Takatani T, Arakawa of the pufferfish Takifugu rubripes and Takifugu O. Survey of food poisoning incidents due to ingestion of porphyreus. Food Hyg. Saf. Sci. 2012; 53: 33–38. marine boxfish, along with their toxicity in Japan. J. Food Wiese M, D’Agostino PM, Mihali TK, Moffitt MC, Neilan Hyg. Soc. Japan 2009b; 50: 270–277. BA. Neurotoxic alkaloids: saxitoxin and its analogs. Mar. Tanu MB, Mahmud Y, Tsuruda K, Arakawa O, Noguchi T. Drugs 2010; 8: 2185–2211. Occurrence of tetrodotoxin in the skin of a rhacophoridid Williams BL. Behavioral and chemical ecology of marine frog Polypedates sp. from Bangladesh. Toxicon 2001; 39: organisms with respect to tetrodotoxin. Mar. Drugs 2010; 937–941. 8: 381–398. Tanu MB, Mahmud Y, Takatani T, Kawatsu K, Hamano Y, Wong JL, Oesterlin R, Rapoport H. The structure of saxi- Arakawa O, Noguchi T. Localization of tetrodotoxin in toxin. J. Am. Chem. Soc. 1971; 93: 7344–7345. the skin of a brackish water puffer Tetraodon steindachneri Woodward RB. The structure of tetrodotoxin. Pure Appl. on the basis of immunohistological study. Toxicon 2002; Chem. 1964; 9: 49–74. 40: 103–106. Yamamori K, Nakamura M, Matsui T, Hara T. Gustatory re- Tanu MB, Mahmud Y, Arakawa O, Takatani T, Kajihara H, sponses to tetrodotoxin and saxitoxin in fish: a possible Kawatsu K, Hamano Y, Asakawa M, Miyazawa K, mechanism for avoiding marine toxins. Can. J. Fish. Aqua. Noguchi T. Immunoenzymatic visualization of tetrodo- Sci. 1988; 45: 2182–2186. toxin (TTX) in Cephalothrix species (Nemertea: Anopla: Yamamori K. Natural forms of tetrodotoxin in puffer fish. Palaeonemertea: Cephalotrichidae) and Planocera Nippon Suisan Gakkaishi 2002; 68: 922–923. reticulata (Platyhelminthes: Turbellaria: Polycladida: Yasumoto T, Yasumura D, Yotsu M, Michishita T, Endo A, Planoceridae). Toxicon 2004; 44: 515–520. Kotaki Y. Bacterial production of tetrodotoxin and Tatsuno R, Shikina M, Soyano K, Takatani T, Arakawa O. anhydrotetrodotoxin. Agric. Biol. Chem. 1986; 50: 793– Maturation-associated changes in the internal distribution 795. of tetrodotoxin in the female goby Yongeichthys criniger. Yasumoto T, Yotsu M, Murata M, Naoki H. New tetrodo-

toxin analogues from the newt Cynops ensicauda. J. Am. VJ, Konoki K, Cho Y, Adachi M, Imazu T, Nishikawa T, Chem. Soc. 1988; 110: 2344–2345. Isobe M. First identification of 5,11-dideoxytetrodotoxin Yotsu M, Hayashi Y, Khora SS, Sato S, Yasumoto T. Isola- in marine animals, and characterization of major fragment tion and structural assignment of 11-nortetrodotoxin-6(S)- ions of tetrodotoxin and its analogs by high resolution ESI- ol from the puffer Arothron nigropunctatus. Biosci. MS/MS. Mar. Drugs 2013; 11: 2799–2813. Biotechnol. Biochem. 1992; 56: 370–371. Yu CF, Yu PF. Are puffer fish more toxic in their spawning Yotsu-Yamashita M. Chemistry of puffer fish toxin. J. seasons? Mar. Biol. 2002; 140: 1053–1057. Toxicol.-Toxin Rev. 2001; 20: 51–66. Zaman L, Arakawa O, Shimosu A, Onoue Y. Occurrence of Yotsu-Yamashita M, Sugimoto A, Terakawa T, Shoji Y, paralytic shellfish poison in Bangladeshi freshwater puff- Miyazawa T, Yasumoto T. Purification, characterization, ers. Toxicon 1997; 35: 423–431. and cDNA cloning of a novel soluble saxitoxin and tetro- Zaman L, Arakawa O, Shimosu A, Shida Y, Onoue Y. Oc- dotoxin binding protein from plasma of the puffer fish, currence of a methyl derivative of saxitoxin in Bangla- Fugu pardalis. Euro. J. Biochem. 2001; 268: 5937–5946. deshi freshwater puffers. Toxicon 1998; 36: 627–630. Yotsu-Yamashita M, Kim YH, Dudley SC, Choudhary G, Zhang B, Yang G, Yu X, Mao H, Xing C, Liu J. Haff disease Pfanul A, Oshima Y, Daly JW. The structure of zetekitoxin after eating crayfish in east China. Intern. Med. 2012; 51: AB, a saxitoxin analog from the Panamanian golden frog 487–489. Atelopus zeteki: a potent sodium channel blocker. Proc. Zu jeddeloh B. Haffkrankheit (Haff disease). Erg. Inn. Med. Natl. Acad. Sci. USA 2004; 101: 4346–4351. 1939; 57: 138–182. Yotsu-Yamashita M, Abe Y, Kudo Y, Ritson-Williams R, Paul