2.4 Occurrence and accumulation in seafood

2.4.1 Uptake and elimination of PSP toxins in aquatic organisms

During the process of filtration the dinoflagellate cells and cysts are transported to the oesophagus and the stomach of the bivalve molluscs. The digestion takes place in the stomach and the diverticulae whereby the PSP toxins are released and enter the digestive organs. The particular toxin mixture retained in soft tissues of the shellfish varies in concentration and over time, and is determined by the species and strains of the dinoflagellates and shellfish as well as by other factors like environmental conditions. In mussels, it was found that the viscera, which constitute only 30 percent of the total tissue weight, contribute 96 percent of total toxicity. In clams the toxins rapidly concentrate in the viscera and gradually decrease afterwards. After a lag period of four or more weeks, the toxins are mainly detected in the siphon. The composition is not consistent but varies with the time and location in the animals (Mons et al., 1998).

Various authors have reported on the toxicity of various scallop tissues and a number of generalities have emerged (Shumway et al., 1988):

• The adductor muscle does not accumulate toxins and has in fact been shown to inactivate the toxins when present. One exception was the purple-hinged scallop, Hinnites giganteus, where toxin levels reached 2000 mg/100 g of tissue.

• Digestive gland, mantle, gonad and gill tissues all retain the toxins although the levels vary between tissues and between species.

• There are seasonal variations in toxicity level of the various tissues.

After uptake and distribution, the toxins may undergo transformation. In feeding experiments non-toxic butter clams were fed A. catenella containing GNTX 1-4 and neoSTX but no STX. After a period of 83 days STX was also detected, leading the authors to conclude that some type of synthesis or biotransformation of GNTX 1-4 and/or neoSTX to STX occurs in vivo. Similar findings were reported by other authors (Mons et al., 1998).

One common transformation, termed epimerization, occurs when a portion of the original STX molecule rearranges. Scallop and mussel, for example, can perform epimerization of STX they receive from the toxic algae when the H and OSO₃^- switch locations on the number 11 position of the STX molecule. Such a transformation can decrease toxicity eleven-fold. On the contrary, there are also transformations that increase toxicity. For example, a six-fold increase in toxicity occurs when the SO₃^- group is separated from position 21 on the STX molecule by acid hydrolysis (Mons et al., 1998).

The butter clam has a distinctive ability to chemically bind the highly toxic STX in its siphon tissue and can retain PSP toxins for up to two years after initial ingestion. The littleneck clam, Prothotaca staminea, can also become toxic but less so than the butter clam. The lower toxicity of the littleneck clam is partially due to its ability to perform transformations that change highly toxic STXs to the moderately toxic forms. The combined effect of the littleneck clam’s capability to transform STXs to less toxic forms, and the ability of butter clams to concentrate and retain highly toxic forms, can result in a wide difference in toxicity between these two species. This toxicity difference is particularly significant since butter clams and littleneck clams can coexist on the same beach and are, to the unskilled harvester, similar in appearance (Mons et al., 1998). MacKenzie et al. (1996) noted the changes in PSP-toxin profiles in the surfclam tuatua (Paphies subtriangulata) inhabiting the beaches in the Bay of Plenty, New Zealand, during the contamination phase (peak levels (412 µg STX eq/100 g) in January 1993 and over a six-month period one year later when low toxin levels (40 µg/100 g) persisted. Toxin profiles during peak contamination consisted of various levels of carbamate derivatives GNTX 1-4, neoSTX and STX with some traces of the decarbamoyl derivative dc-STX. These profiles resembled those produced by the dinoflagellate A. minutum, which caused the PSP incident. One year later, only traces of derivatives other than STX remained and almost all of this toxin was sequestered within the siphon.

Andrinolo et al. (1999a) demonstrated that natural depuration from PSP toxins by Aulacomya ater, a native South American filter-feeder bivalve, occurs in the form of an exponential decay of the first order (one-compartment model). Depending on their detoxification kinetics, bivalves have been classified into two major groups: slow detoxifiers (e.g. Saxidomus giganteus, Spisula solidissima, Placopecten magellanicus, Patinopecten yessoensis) and rapid-to-moderate detoxifiers (e.g. Mytilus edulis and Mya arenaria) (Androlino et al., 1999a). A biphasic, two-compartment model best describes detoxification kinetics in some species. During toxification, the viscera typically attain toxicities two to five times higher than whole tissues, whereas locomotor tissues (foot and adductor muscle) are least toxic. However, the viscera detoxify faster than other tissues, leading to a steady decline in their contribution to total toxin burden during detoxification. Biotransformation of toxins in tissues is most pronounced in a few clam species capable of enzymatic decarbamoylation (e.g. Protothaca staminea), and more limited in others such as Mya arenaria and Mytilus edulis. Overall, changes in toxin profile are greatest when ingested dinoflagellates are rich in low potency, N-sulfocarbamoyl toxins (Bricelj and Shumway, 1998).

Some bivalves can avoid ingesting toxic dinoflagellates such as the northern quahaug (Mercenaria mercenaria) which retracts its siphon and closes its valves in the presence of Alexandrium sp. (Mons et al., 1998).

Blanco et al. (1997) studied detoxification kinetics in the mussel Mytilus galloprovincialis previously exposed to a bloom of the PSP producing dinoflagellate G. catenatum. The toxin profile observed in the mussels was very similar to that of G. catenatum, showing that biotransformation had little or no importance in this case. Detoxification took place in two phases:

i. a fast one, which took place during the early detoxification period (only a small amount of the toxin, relative to the initial amount, remains in the bivalves after the first few days of detoxification); and

ii. a slow one, lasting from the end of the first phase to the end of detoxification. Environmental conditions (salinity, temperature and light transmission) and body weight affected detoxification especially during the fast first phase.

When Pacific oysters (Crassostrea gigas) were fed toxic or non-toxic A. tamarensis and A. fundyense, a stop/start clearance behaviour (filter pump switched off/on) of the oysters was observed suggesting that PSP toxins were not directly involved in inhibiting the initial feeding response. When control oysters were fed a reference microalga, Isochrysis sp., known to support their growth, this behaviour was not seen. When Pacific oysters, which were acclimated to Isochrysis sp., were fed mixtures of Alexandrium/Isochrysis, further evidence of stop/start clearance behaviour was seen (Wildish et al., 1998).

Adult Pacific oysters (Crassostrea gigas) experimentally contaminated with PSP toxins (by exposure to A. minutum) up to concentrations of 150-300 mg STX eq/100 g, were fed diets based on non-toxic dinoflagellates or diatoms in order to study detoxification. Despite the large individual variations in toxin levels, a detoxification time of three to four days was measured for reaching the safety threshold of 80 mg/100 g in the oysters. Detoxification rates did not differ significantly when oysters were fed Isochrysis galbana, Tetraselmis suesica, Thalassiosira weissflogii or Skeletonema costatum. GNTX2/GNTX3 were the major compounds found in the oysters during depuration, whereas C toxins were quite low and STX and neoSTX undetectable. The toxin profile was the same as in A. minutum suggesting no biotransformation in the oyster (Lassus et al., 2000).

The Chinese scallop, Chlamys farreri, has a high ability to accumulate PSP toxins. After exposure for 48 hours to toxic A. minutum 5000 mg STX eq/100 g were found in the viscera of these scallops and the rate of detoxification was slow. The viscera accounted for 97 percent of the total toxin content. The ratio of different PSP toxins has changed during the experimental period, for example, the ratio of GNTX1 and GNTX4 to total toxins decreased while that of GNTX2 and GNTX3 increased. The toxin profile in the scallops was different from that in the algae. Toxin profile in the scallop faeces matched well with that in the early stage of A. minutum in batch culture (Zou et al., 2001).

In large containers (20 litres), the adult pelagic harpacticoid copepod, Euterpina acutifrons, was incubated with a high toxic strain of A. minutum (1 000 or 10 000 cells/ml) for up to five days. Only trace levels of PSP-toxins were found in the extracts analysed by LC. With a low and a high toxic strain of A. minutum (1 000 and 10 000 cells/ml), 10 to 15 percent of copepods were inactive after one to two days. It is suggested that E. acutifrons avoids feeding on the dinoflagellates after tasting a few cells (Bagøien et al., 1996).

Purple clams (Hiatula diphos) were contaminated with PSP toxins by feeding them with cells of A. minutum and then fed to maculated ivory shells (Babylonia areolata), which are carnivorous gastropods. The toxin composition in the clams, gastropods and dinoflagellates were similar but the profile differed in the gastropods. There was a notable degradation of GNTX1 in the gastropod compared to the clam and the dinoflagellate that resulted in a decrease in toxicity while the total amount of toxins was accumulatively increasing. The transmitted GNTX1-4 of A. minutum could only be found in the viscera of these shellfish species (Chen and Chou, 1998). In a later study, Chou and Chen (2001a) studied accumulation, distribution and elimination of PSP toxins in purple clams (Hialuta rostrata) after feeding a toxic strain of A. minutum. The high toxicity of the digestive gland was confirmed. Depuration efficiency between toxic clams fed non-toxic algae and those put in starvation was similar. Toxin profile of the clams was similar to that of A. minutum at the end of the feeding period (GNTX4 and GNTX1 were dominant). However, at the end of the elimination period GNTX3 and GNTX2 were dominant indicating inconsistent removal rates of different toxins or transformation of toxins. No PSP toxins other than GNTX1-4 were found. The non-visceral tissues were also toxic after feeding with toxic algae, however, the toxicity was low and the profile was also similar to that of the toxic algae.

2.4.2 Shellfish containing PSP toxins

Although most filter-feeders are relatively insensitive to the STXs, there are differences among the various species of bivalves in the way they deal with and respond to the STXs. Mussels, for instance, appear in general to accumulate much higher levels of PSP toxins than oysters under similar circumstances. Subsequent laboratory feeding studies showed that mussels readily consumed concentrations of Alexandrium equal to or greater than those that caused oysters to cease pumping and close up. Electro-physiological investigations of isolated nerves from Atlantic coast bivalves demonstrated that those from oysters were sensitive to the toxins, while those from the mussels were relatively insensitive (Mons et al., 1998).

The group of shellfish identified in cases of PSP consists mostly of bivalve molluscs. This group includes mussels, clams and, to a lesser extent, oysters, scallops and cockles in temperate zones. An extensive list of shellfish found to contain PSP toxins is given in Table 2.1.

In April 1991, the ormer Haliotis (Eurotis) tuberculata from the Galician coast of Spain was found to contain PSP toxins. In October 1993, the market for this mollusc was closed. Samples from December 1995 were contaminated with 252 ± 25 mg STX eq/100 g of meat by mouse bioassay analysis and 454 ± 86 mg STX eq (sum of STX and dcSTX converted to STX eq by conversion factors of 1.9 and 1.14, respectively)/100 g of meat by LC. No value below 140 mg STX eq/100 g of meat was detected by the mouse bioassay. The major component was dcSTX (83 to 100 percent) with STX in much smaller proportion. The epithelium carried 2.6 times more toxin than the muscle. Attempts at natural detoxification, keeping ormers under controlled laboratory conditions for three months, did not work. The elimination of epithelium and gut would result in around 75 percent less toxicity (Bravo et al., 1999).

Chlamys nobilis from the waters of the Hong Kong Special Administrative Region, China contained 320 mg STX eq/100 g soft tissue. Following the red tide from March to April 1998, high levels of PSP toxins were detected in Perna viridis from waters of Hong Kong Special Administrative Region, China (Zhou et al., 1999). In 5 percent of samples of shellfish caught along the Chinese coast from north to south, PSP toxins were found. Although the PSP toxin levels were low (only two samples exceeded the regulatory threshold limit), it indicated that PSP toxin producers existed in this area (Zhou et al., 1999).

Table 2.1 Shellfish found to contain PSP toxins

Source: Mons et al., 1998, except as indicated

* Takatani et al., 1997; ** Lagos, 1998; *** Shumway et al., 1988; ^# MacKenzie et al., 1996; ^## Todd (1997)

2.4.3 Other aquatic organisms containing PSP toxins

The grazing habits of two abundant copepod species from the Gulf of Maine, Acartia tonsa and Eurytemora herdmani, were compared using cultured isolates of Alexandrium spp., which differed in toxicity per cell and toxin composition and a non-toxic dinoflagellate Lingulodinium polyedrum. Toxicity of the dinoflagellates had no effect on the grazing efficiencies of the two copepod species. Neither species showed strong evidence of incapacitation or adverse effects from ingested toxins. E. herdmani accumulated higher levels of PSP toxins but also had fuller guts. The experiments with mixed dinoflagellates suggested that both copepod species have the ability to choose their prey items based on palatability and not on toxicity. Although the toxin retention efficiencies of copepods tested were low overall (5 percent), high levels of PSP toxins were accumulated in copepod grazers, supporting evidence that zooplankton may serve as PSP toxin vector to higher trophic levels (Teegarden and Cembella, 1996).

Three species of marine copepods (Acartia tonsa, Centropagus hamatus, Eurytemora herdmani), commonly co-occurring with toxic Alexandrium spp., appeared to be able to discriminate between toxic and non-toxic Alexandrium spp. cells by chemosensory means, suggesting that selective behaviour rather than physiological effects governs the grazing response of copepods. Feeding behaviour varied among copepod species, suggesting that grazing pressure on toxic Alexandrium spp. is not uniform throughout the zooplankton community (Teegarden, 1999).

In large volumes (20 litres), the adult pelagic harpacticoid copepod Euterpina acutifrons, incubated with a high toxic strain of A minutum (1 000 or 10 000 cells/ml) for up to five days, revealed only trace levels of PSP-toxins in the extracts analysed by LC. With both a low and a high toxic strain of A. minutum (1 000 and 10 000 cells/ml), 10 to 15 percent of copepods were inactive after one to two days. It is suggested that E. acutifrons may avoid feeding on the dinoflagellates after tasting a few cells (Bagøien et al., 1996).

Of the crabs involved in human PSP in Japan and Fiji, most are xanthid crabs (Lophozozymus pictor), though some other species are also involved (horseshoe crab and marine snail). These species share the common feature of living in coral reefs and feeding by surface grazing (Mons et al, 1998; Sato et al., 2000).

Out of 459 specimens of xanthid crabs collected in Taiwan Province of China during October 1992 and May 1996 and analysed for tetrodotoxin and PSP toxins, five specimens (Zosimus aeneus, Lophozozymus pictor, Atregatopsis germaini, Atergatis floridus, Demania reynaudi) were found to contain PSP toxins besides tetrodotoxin. The percentages of PSP toxins varied from 11 to 97 percent (the remaining 89 to 3 percent was tetrodotoxin). The toxin profile of the PSP toxins varied within the different species. The source of the PSP toxins was A. minutum (Hwang and Tsai, 1999).

Algal toxins can also cause mortalities in fish as they move through the marine food web. Some years ago, tons of herring died in the Bay of Fundy after consuming small planktonic snails that had been feeding on Alexandrium. From the human health point of view, it is fortunate that herring, cod, salmon and other commercial fish are sensitive to PSP toxins and, unlike shellfish, die before the toxins reach dangerous levels in their flesh. Some toxins, however, accumulate in the liver and other organs of the fish, and so animals such as other fish, marine mammals and birds that consume whole fish, including the viscera, are at risk. In 1987, 14 humpback whales died suddenly from exposure to a bloom of A. tamarensis in Cape Cod Bay (Massachusetts). Researchers later learned that the whales had eaten mackerel whose organs contained high concentrations of STX (Mons et al., 1998).

In May 1996, star fish Asterias amurensis collected in the estuary of the Nikoh River (Kure Bay, Hiroshima Prefecture, Japan) appeared to contain PSP toxins (in the mouse bioassay 8.0 MU/g for whole body and 28.7 MU/g for viscera). The PSP toxins were supposed to come via the food chain from toxic bivalves living in the same area. The starfish toxin comprised of highly toxic components. GNTX1, GNTX2, GNTX3, GNTX4, dc-GNTX3 and dc-STX were the major components and accounted for approximately 77 mole%, along with N-sulfocarbamoyl derivatives C1-C4. GNTX1 was present in the largest amount (37.4 mole%) (Asakawa et al., 1997).

Atlantic mackerel (Scomber scombrus) are lethal vectors of PSP toxins to predators. Mackerel appeared to retain PSP toxins (STX [96 percent], GNTX2 and GNTX3 [4 percent]) year-round. The toxin content of the liver (determined by LC) increased significantly with fish age, suggesting that mackerel progressively accumulate the toxins during their life. The toxin content of the liver also increased significantly during the summer feeding sojourn in the Gulf of St. Lawrence, Canada. Zooplankton was the likely source of the PSP toxins in mackerel. Mean toxin content was 17.4 nmol/liver and the mean toxicity was 112.4 mg STX eq/100 g liver wet weight (Castonguay et al., 1997).

Marine puffers (Arothron mappa, A. manillensis, A. nigropunctatus, A. hispidus, A. stellatus, A. reticularis) in waters of the Philippines contained considerable amounts of PSP toxins (major component STX) besides tetrodotoxin (TTX), another potent marine toxin present in finfish. The toxicity was detected in liver, intestine, muscle and skin (Sato et al., 2000). Freshwater puffers (Tetraodon leiurus complex, Tetraodon suvatii), collected from the northeastern province of Thailand appeared to contain PSP-toxins. Toxicity was highest in liver and varied with location and season of fish catch. Toxin profiles from eggs, liver, skin and muscle showed the presence of STX, neoSTX and dcSTX (Kungsuwan et al., 1997). Sato et al. (1997) identified STX in the freshwater puffer Tetraodon fangi, which caused food poisoning in Thailand. Tetrodotoxin was not detected in this species. Two species of freshwater puffers (Tetraodon cutcutia, Chelonodon patoca), collected from several locations in Bangladesh, showed lethality in the mouse bioassay (at 2.0 to 40.0 MU/g tissue as PSP). Toxicity of skin was generally higher than other tissues examined (muscle, liver, ovary). Analyses of T. cutcutia revealed the presence of STX, dcSTX, GNTX2 and GNTX3, dcGNTX2 and three unidentified components possibly related to PSP. No tetrodotoxin or related substances were detected (Zaman et al., 1997a).