kinetics of accumulation and transformation of paralytic shellfish toxins in the blue mussel mytilus...
TRANSCRIPT
Kinetics of accumulation and transformation of paralytic shellfish
toxins in the blue mussel Mytilus galloprovincialis
Juan Blancoa,*, Ma Isabel Reyerob, Jose Francoc,b
aCentro de Investigacions Marinas, Procesos oceanograficos costeros, Pedras de Coron s/n. Apdo. 13. Vilanova de Arousa 36620, SpainbInstituto Espanol de Oceanografıa, Centro Costero de Vigo. Subida a Radiofaro, San Miguel de Oia, Vigo, Spain
cInstituto de Investigaciones Marinas. Eduardo Cabello s/n. Vigo, Spain
Received 12 June 2003; revised 10 July 2003; accepted 10 October 2003
Abstract
Mussels (Mytilus galloprovincialis) were fed cultures of the Paralytic Shellfish Poisoning agent Alexandrium minutum
(Strain AL1V) for a 15-day period and, for the next 12 days, they were fed the non-toxic species Tetraselmis suecica, in order to
monitor the intoxication/detoxification process. The toxin content in the bivalve was checked daily throughout the experiment.
During the time-course of the experiment, the toxin profile of the bivalves changed substantially, showing increasingly greater
differences from the proportions found in the toxigenic dinoflagellate used as food. The main processes involved in the
accumulation of toxins and in the variation of the toxic profiles were implemented in a series of numerical models and the
usefulness of those models to describe the actual intoxication/detoxification kinetics was assessed. Models that did not include
transformations between toxins were unable to describe the kinetics, even when different detoxification rates were allowed for
the toxins involved. The models including epimerization and reduction provided a good description of the kinetics whether or
not differential detoxification was allowed for the different toxins, suggesting that the differences in detoxification rates between
the toxins are not an important factor in regulating the change of the toxic profile. The implementation of Michaelis–Menten
kinetics to describe the two reductive transformations produced a model that had a poorer fit to the data observed than the model
that included only a first order kinetics. This suggests that, it is very unlikely that any enzymatic reaction is involved in the
reduction of the hydroxycarbamate (OH-GTXs) to carbamate (H-GTXs) gonyautoxins.
q 2003 Elsevier Ltd. All rights reserved.
Keywords: Paralytic shellfish poisoning; Paralytic shellfish toxins; Kinetics; Accumulation; Ddetoxification; Biotransformation; Models
1. Introduction
Paralytic shellfish toxins (PSP toxins) are a group of
substances related to Saxitoxin (Fig. 1) produced, among
others, by some species of phytoplanktonic dinoflagel-
lates. As the bivalves ingest the producer organisms,
these toxins are accumulated—mainly in the digestive
gland—, becoming a serious risk to any mammal,
including man, that consumes them. The toxins contained
in the dinoflagellate cells are taken up by the bivalves,
then initially stored in the digestive gland and later
transferred in part to other organs or tissues through the
bloodstream. The toxins bind to receptors present in the
organs/tissues (Louzao et al., 1992) but a fraction is lost
or degraded at an organ/tissue specific rate (Bricelj and
Cembella, 1995). This occurrence often produces detox-
ification kinetics which is, in appearance, biphasic
(Lassus et al., 1989, 1993) and which must be described
by means of two- or multi-compartmental models (Silvert
and Cembella, 1995; Blanco et al., 1997; Silvert et al.,
1998a). The relative proportions between PSP toxins
inside the bivalves are different from those in the
producer dinoflagellate, which can only be explained by
means of toxin-specific uptake or elimination, or by
0041-0101/$ - see front matter q 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.toxicon.2003.10.007
Toxicon 42 (2003) 777–784
www.elsevier.com/locate/toxicon
* Corresponding author. Tel.: þ34-986-500155; fax: þ34-986-
506788.
E-mail address: [email protected] (J. Blanco).
transformations between toxin types. Differential uptake
has not been found (Bricelj and Shumway, 1998) and
only one experiment with Spisula solidissima suggested a
real differential elimination (Silvert et al., 1998b).
Several kinds of transformations have been documented
for PSP toxins. Epimerization between a and b epimers
appears to take place easily and it is not believed to be
enzymatically mediated. Reduction, acid hydrolysis or
enzymatic hydrolysis have also been demonstrated
(Cembella et al., 1993, 1994; Oshima, 1995; Murakami
et al., 1999a,b), but in some cases at different velocities
in different species, as well as in different organs of the
same organism.
Considering that these processes take place simul-
taneously and that they are difficult or even impossible to
measure directly, it becomes clear that predicting the
accumulation of these toxins is a difficult task which
requires the use of an indirect approach such as dynamic
modelling. This approach has been used with several
species and toxin groups, both taking the transformations
between toxins into account (Fernandez et al., 1998;
Silvert et al., 1998a) and not considering these trans-
formations (Silvert and Cembella, 1995; Blanco et al.,
1995, 1997, 1999).
In this work, we have used the modelling approach to
estimate the rates of all the processes mentioned above,
employing Mytilus galloprovincialis as the bivalve species,
because of its economic and ecological importance.
Alexandrium minutum (strain AL1V) was used as the source
of PSP toxins owing to its economic importance (it
frequently blooms in Galicia and other geographical
areas), and its relatively simple toxin profile (mainly
gonyautoxins, Franco et al., 1994), which greatly simplifies
both the analytical procedures and the models.
2. Material and methods
2.1. Biological material and procedures
Mussels M. galloprovincialis (3–4 cm long) were
collected from the Rıa de Vigo (Galicia, NW Spain), and
checked for the absence of PSP toxins by means of HPLC-
FD analysis. A. minutum (AL1V) was obtained from the
culture collection of the Centro Costero de Vigo of the
Instituto Espanol de Oceanografıa and cultured in K
medium (Keller and Guillard, 1987) in 2 l flasks under
fluorescent lighting. Tetraselmis suecica was cultured in 50 l
polypropylene bags using Walne’s medium (Lavens and
Sorgeloos, 1996) and fluorescent lightning. A new flask was
inoculated every 2 days in order to keep the cultures in a
homogeneous physiological state when needed to feed the
mussels. The mussels were placed in a 5 l tank, which
underwent cleaning and a change of water daily. Over the
course of 15 days, the animals were fed, at regular intervals,
by adding the amount of culture needed to reach an initial
concentration in the tank of roughly 1000 cells of A.
minutum ml21. Small aliquots of the culture used as food
were taken just prior feeding, to quantify the concentration
of Alexandrium and the toxin. After day 15, the Alexan-
drium cells in the diet were replaced with the non-toxic
species T. suecica. Each day three mussels were randomly
taken from the tank, weighed, their soft tissues homogenized
and the toxins they contained extracted following the AOAC
procedure (AOAC, 1990). Small aliquots were taken to
quantify the toxin concentration of the extracts by HPLC.
2.2. PSP toxin quantification
Toxins were identified and quantified by comparing their
retention times and fluorescent response with standard
Fig. 1. Structure and toxic power of saxitoxin and its main analogs (MU ¼ Mouse Unit, the amount of toxin required to kill a mouse weighing
20 g in 15 min upon intra-peritoneal injection).
J. Blanco et al. / Toxicon 42 (2003) 777–784778
PSP1-B solutions obtained from the National Research
Council (Canada). No standard of decarbamoyl-gonyautox-
ins was available. However, in view of the fluorescent
response of these toxins, as compared to the others found the
mussels, they always appeared in low quantities. This fact as
well as the difficulties involved in attempting to obtain an
accurate quantification of these types of toxins in the
absence of standards, prompted us to exclude them from the
analyses. The identification/quantification of the toxins was
carried out following the HPLC-FD technique of Franco and
Fernandez-Vila (1993).
2.3. Models
The models were implemented using Matlab and Matlab
Simulink. The fit was carried out by least square
minimization using the routines in the Matlab Optimization
Toolbox.
All the statistical procedures were performed using the
Minitab 13.1 statistical package, and following the general
methods given in Bates and Watts (1988).
3. Results
The first model implemented (Model 1) assumes the
simplest situation (Table 1): (a) there is no transformation
between toxins; (b) the uptake and detoxification rates are
the same for all toxins and (c) the detoxification follows a
first-order kinetics. In this case, the model can be formulated
as a simple system of differential equations
dGTX1=dt ¼ F½GTX1�mediumAE 2 KGTX1 ð1Þ
dGTX2=dt ¼ F½GTX2�mediumAE 2 KGTX2
dGTX3=dt ¼ F½GTX3�mediumAE 2 KGTX3
dGTX4=dt ¼ F½GTX4�mediumAE 2 KGTX4
where GTXn and ½GTXn� denote the concentration of each
gonyautoxin in mussels and in the maintenance medium,
respectively, F is the clearance rate and K is the
detoxification rate. The equation for each toxin has two
parts: uptake and loss. The uptake portion includes the
amount of toxin cleared by the mussels from the medium,
that is the F½GTX�medium and the fraction of it that is really
absorbed, AE (the absorption efficiency). The loss term is
assumed to be a fixed percentage, K (the detoxification rate)
of the amount of toxin inside the mussel, GTXn:
This model was not able to describe the experimental data
of most toxins (Fig. 2), and only GTX4, the most abundant,
was reasonably well described. Nevertheless, even in this
toxin, the peak values expected from the model were far from
the actual ones and detoxification seemed to proceed too
quickly. The filtration rate, the absorption efficiency and the
detoxification rate, with which the best fit was obtained, were
4.6 l h21, 67% and 0.15 day21, respectively.
The next model implemented (Model 2) assumes the
same filtration rate and absorption efficiency as in the
previous model but allows for a different detoxification rate
for each toxin instead of assuming a common detoxification
rate for all of them. The previous system of Eq. (1) was,
therefore, only modified by the substitution of each K by a
Kn (the detoxification rate for each particular toxin).
Although, the new model fit the data better than the
previous one (Fig. 3), it was totally unsuitable for two
ðGTX2 and GTX3Þ out of the four toxins. The estimated
detoxification rates of the four toxins were very different
from each other, ranging from 0 to 0.28 day1—a wide range
that would seem to be unlikely, considering the structural
similarity between the toxins.
The third model implemented (Model 3) differs from the
first in that some transformations between toxins were
allowed to take place. The transformations allowed were
those that appeared to be most likely, in view of the changes
detected in the proportions between toxins in other
mollusks: (a) epimerization between the a ðGTX1 and
GTX2Þ and b ðGTX3 and GTX4Þ forms, in both directions
and (b) reduction of the two OH-GTXs (4, 1) to their
corresponding H-GTXs (3, 2, respectively). The equations
describing the model are therefore
dGTX1=dt¼F½GTX1�mediumAE2KGTX1 þE4–1GTX4
2R1–2GTX1 ð2Þ
dGTX2=dt ¼ F½GTX2�mediumAE 2 KGTX2 þ E3–2GTX3
þ R1–2GTX1
Table 1
Differences between the models implemented
Model Detoxification rate Epimerization Reduction Compartments
Model 1 Common for all toxins No No 1
Model 2 Different for each toxin No No 1
Model 3 Common for all toxins Yes, first order kinetics Yes, first order kinetics 1
Model 4 Different for each toxin Yes, first order kinetics Yes, first order kinetics 1
Model 5 Common for all toxins Yes, first order kinetics Yes, Michaelis–Menten kinetics 1
Model 6 Common for all toxins Yes, first order kinetics Yes, first order kinetics 2
All models share common filtration rate and absorption efficiency for all toxins.
J. Blanco et al. / Toxicon 42 (2003) 777–784 779
dGTX3=dt ¼ F½GTX3�mediumAE 2 KGTX3 þ E2–3GTX2
þ R4–3GTX2
dGTX4=dt ¼ F½GTX4�mediumAE 2 KGTX4 þ E1–4GTX1
2 R4–3GTX4
where the equations defining first model are complemented
with the different transformations that were assumed to be
proportional to the amount of transformed toxin. Epimer-
izations were described by En–mGTXn and reductions by
Rn–mGTXn; where En–m are the rates of epimerization
between the toxins indicated in the sub-index, and Rn–m; the
reduction rates.
Fig. 2. Output of Model 1 (traces) and observed concentrations (symbols), of the four toxins studied in the experiment. The model assumes the
same incorporation and detoxification rate for the all the toxins; no transformation between toxins; and that the detoxification follows a first
order kinetics.
Fig. 3. Output of Model 2 (traces) and observed concentrations (symbols), of the four toxins studied in the experiment. The model assumes the
same incorporation for the all the toxins; no transformation between toxins; a first order kinetics for detoxification; but a different detoxification
rate for each toxin.
J. Blanco et al. / Toxicon 42 (2003) 777–784780
This model fit the data substantially better that the two
previous ones, adequately describing the main features
observed during the time-course of the experiment (Fig. 4).
In this case, the estimated detoxification rate for all
toxins was 0.06 day21. The estimated values for the
remaining parameters are given in Table 2. The model
output mirrors the general trend of the experimental data,
but obviously with the limitations imposed by their
dispersion. An analysis of the residual deviations of the
model (Fig. 5) showed that it was suitable for the two OH-
GTXs (4, 1), giving normal distributions and no detectable
trend. For the two H-GTXs (3, 2), however, an upward trend
was seen at the beginning of the experiment, while it shifted
downward towards the end. The distributions were not
normal either, indicating that the model did not provide an
accurate description of the kinetics of these two toxins from
a functional point of view.
Modifying the model to allow a different detoxification
rate for each toxin (Model 4) did not improve the fit, as the
variance explained was only marginally larger and the
residuals behaved in the same way as in the previous model.
The estimated values of the parameters are given in Table 2.
The estimated detoxification rates of the four toxins were
similar, ranging from 0.053 day21 (GTX3 and GTX1) to
0.070 day21 (GTX2). Epimerization occurred at a faster rate
between the two H-GTX toxins than between the OH-GTXs
and reduction was also estimated to be faster between the b-
forms than between the a ones.
The use of a Michaelis–Menten (Model 5), instead of
a first-order kinetics model, for the reduction from
the OH-GTXs to the H-GTXs degraded the fit, both
quantitatively and qualitatively. The model implemented
differed from Eq. (2) in which the reduction terms ðR4–3GTX4
and R1–2GTX1Þ were replaced with their corresponding
Table 2
Parameters estimated by fitting three of the models implemented:
Model 2 (differential detoxification without transformations),
Model 3 (common detoxification with transformations) and Model
4 (differential detoxification with transformations)
Models Model 2 Model 3 Model 4
Clearance rate (F; l h21) 4.6 4.6 4.6
Absorption efficiency (AE, %) 67 67 67
Detoxification rates
Common (K; day21) 0.06
GTX1 (K1; day21) 0.00 – 0.053
GTX2 (K2; day21) 0.28 – 0.070
GTX3 (K3; day21) 0.04 – 0.053
GTX4 (K4; day21) 0.13 – 0.063
Epimerization rates
GTX3 ! GTX2 (E3–2; day21) – 0.116 0.061
GTX2 ! GTX3 (E2–3; day21) – 0.008 0.003
GTX4 ! GTX1 (E4–1; day21) – 0.060 0.100
GTX1 ! GTX4 (E1–4; day21) – 0.000 0.014
Reduction rates
GTX1 ! GTX2 (R1–2; day21) – 0.0110 0.0038
GTX4 ! GTX3 (R4–3; day21) – 0.0001 0.0082
Fig. 4. Output of Model 3 (traces) and observed concentrations (symbols), of the four toxins studied in the experiment. The model assumes the
same incorporation and detoxification rate for the all the toxins; a first order kinetics for detoxification; but it also assumes the transformation
between toxins by epimerization and reduction.
J. Blanco et al. / Toxicon 42 (2003) 777–784 781
Michaelis–Menten velocities ðVmax4–3½GTX4�Þ=ðKM4–3½
GTX4� and Vmax4–3½GTX4�Þ=ðKM4–3½GTX4�Þ; were the V
max are the maximal velocities of the reductions and the KM
are the Michaelis–Menten constants.
The inclusion of a second compartment in the model
(Model 6), to take into account the possibility of the
existence of two different pools of toxins with different
detoxification rates, did not effectively improve the fit of the
model to the actual data.
4. Discussion
Paralytic shellfish toxins produced by A. minutum
(AL1V) are transformed in the mussel M. galloprovincia-
lis, as in other bivalves (Shimizu and Yoshioka, 1981;
Sullivan et al., 1983; Cembella et al., 1994; Oshima, 1995;
Bricelj and Shumway, 1998; Silvert et al., 1998b). Two
transformations were identified—the epimerization and
reduction of the N1-OH groups. No evidence of decarba-
moylation or transformation of the GTX toxins to STX or
NeoSTX was found. Epimerization is a spontaneous
transformation that probably takes place in response to
the intracellular environment of the bivalve, which is
different from the environment inside the Alexandrium
cells. The reactions involved in the reductions from
OH-GTXs to H-GTXs are not known but there is no
evidence of their reversibility. The models built using
these considerations accurately described the time-course
of the accumulation/elimination of toxins in M. gallo-
provincialis. The models estimated greater epimerization
rates for the GTX4-1 pair than for GTX3-2, as well as
greater reduction rates between the b epimers than
between the a forms. The reactions involved in the
OH-GTX reductions are not known. Oshima (1995),
suggested that some natural reductants, such as gluta-
thione or cysteine, are the agents responsible for these
reductions, while according to Murakami et al. (1999a),
in the case of Pseudocardium sachalinensis, an enzyme
is involved, and Sakamoto et al. (2000) studying the
reductions produced by glutathione and mercaptoethanol,
suggested that these two compounds are not involved in
the reductions that we have considered in this exper-
iment. The models implementing a first-order reaction
kinetics for reduction slightly overestimated the amounts
of H-GTX during the early part of the experiment, but
the model that implemented a Michaelis–Menten kinetics
(typical of enzymatic reactions) resulted in much greater
overestimations. Although, none of the models is
completely satisfactory, better results were had with
those that used a first-order kinetics than with the one
that used the Michaelis–Menten kinetics, thus supporting
the suggestions made by Oshima (1995), that no enzyme
is involved.
Apart from transformations, the differential detoxifica-
tion is one of the possible mechanisms that may be able to
explain the changes in the proportions between the different
toxins in the bivalves as compared to those found in the algal
cells (Lassus et al., 1993). In this work, the models used
estimate only slight differences between the detoxification
rates of the four GTX toxins. Additionally, the inclusion of
these differences produced only a marginal (statistically
non-significant) increase in the fit, in relation to the
equivalent model that used a common detoxification rate
for all four GTX toxins. It would therefore seem that
differential detoxification does not play a significant role in
the change of the relative contributions of the GTX toxins
over time. (Silvert et al., 1998a), using a similar modelling
approach, found that GTX2 appeared to be depurated at
slower rate than the other toxins from the viscera of the
surfclam S. solidissima. In the mussel Mytilus edulis, GTX2
was found to increase its contribution relative to other toxins
during detoxification, and one possible mechanism to
explain this was reduced detoxification (Lassus et al.,
1993). Our results suggest that GTX2 depurates at the same
rate as the other GTX toxins or, if the minor difference found
in the model with differential detoxification is assumed to be
true, then GTX2 would depurate slightly faster than the
other toxins studied. The same lack of differences has been
found by Ichimi et al. (2001) with M. galloprovincialis and
by Sekiguchi et al. (2001) who analyzed the toxins excreted
Fig. 5. Residual deviations of Model 3 (observed–expected toxin
concentrations) for each of the toxins studied.
J. Blanco et al. / Toxicon 42 (2003) 777–784782
by Patinopecten yessoensis. Consequently, it would seem
that differential detoxification does not cause any major
changes in the toxin profile of bivalves, but the possibility
that this mechanism may affects certain species cannot be
ruled out.
Acknowledgements
This work was funded by the projects ‘Determinacion de
toxinas paralizantes en moluscos y cultivos de dinoflagela-
dos. Dinamica de intoxicacion y detoxificacion en mejil-
lones cultivados’ (CICYT ALI92-011-CO2-01) and
‘Acumulacion de toxinas de tipo paralıtico (PSP) e de tipo
amnesico (ASP) en moluscos bivalvos’ PGIDT99PXI50101
(Xunta de Galicia).
References
AOAC, 1990. Paralytic shellfish poison. Biological method. Final
action. In: Herlich, K., (Ed.), Official Methods of Analysis,
Association of Official Analytical Chemists, Arlington, VA, p.
8812.
Bates, D.M., Watts, D.G., 1988. Nonlinear regression analysis and
its applications. Wiley, New York, 365 pp.
Blanco, J., Fernandez, M., Marino, J., Reguera, B., Mıguez, A.,
Maneiro, J., Cacho, E., Martınez, A., 1995. From Dinophysis
spp. toxicity to DSP outbreaks: a preliminary model of toxin
accumulation in mussels. In: Lassus, P., Arzul, G., Erard, E.,
Gentien, P., Marcaillou, C. (Eds.), Harmful Marine Algal
Blooms, Lavoisier/Intercept, Paris.
Blanco, J., Morono, A., Franco, J., Reyero, M.I., 1997. PSP
detoxification kinetics in the mussel Mytilus galloprovincialis:
one- and two-compartment models and the effect of some
environmental variables. Mar. Ecol. Prog. Ser. 158, 165–175.
Blanco, J., Fernandez, M.L., Mıguez, A., Morono, A., 1999.
Okadaic acid detoxification in the mussel Mytilus galloprovin-
cialis: one- and two-compartment models and the effect of
environmental conditions. Mar. Ecol. Prog. Ser. 176, 153–163.
Bricelj, V.M., Cembella, A.D., 1995. Fate of gonyautoxins in
surfclams, Spisula solidissima, grazing upon toxigenic Alexan-
drium. In: Lassus, P., Arzul, G., Erard, E., Gentien, P.,
Marcaillou, C. (Eds.), Harmful Marine Algal Blooms, Lavoi-
sier/Intercept, Paris, pp. 413–418.
Bricelj, V.M., Shumway, S.E., 1998. Paralytic shellfish toxins in
bivalve molluscs: occurrence transfer kinetics and biotrans-
formation. Rev. Fish. Sci. 6, 315–383.
Cembella, A.D., Shumway, S.E., Lewis, N.I., 1993. Anatomical
distribution and spatio-temporal variation in paralytic shellfish
toxin composition in two bivalve species from the Gulf of
Maine. J. Shellfish Res. 12, 389–403.
Cembella, A.D., Shumway, S.E., Larocque, R., 1994. Sequestering
and putative biotransformation of paralytic shellfish toxins by
the sea scallop Placopecten magellanicus: seasonal and spatial
scales in natural populations. J. Exp. Mar. Biol. Ecol. 180,
1–22.
Fernandez, M.L., Mıguez, A., Morono, A., Cacho, E., Martınez, A.,
Blanco, J., 1998. Detoxification of low polarity toxins (DTX3)
from mussels Mytilus galloprovincialis in Spain. In: Reguera,
B., Blanco, J., Fernandez, M.L., Wyatt, T. (Eds.), Harmful
Algae, Xunta de Galicia and IOC of UNESCO, Santiago de
Compostela, pp. 442–449.
Franco, J.M., Fernandez-Vila, P., 1993. Separation of paralytic
shelfish toxins by reversed phase high performance liquid
chromatography, with postcolumn reaction and fluorimetric
detection. Chromatographia 35, 613–620.
Franco, J.M., Fernandez, P., Reguera, B., 1994. Toxin profiles of
natural populations and cultures of Alexandrium minutum from
Galician (Spain) coastal waters. J. Appl. Phycol. 6, 275–279.
Ichimi, K., Suzuki, T., Yamasaki, M., 2001. Non-selective retention
of PSP toxins by the mussels Mytilus galloprovincialis fed with
the toxic dinoflagellate Alexandrium tamarense. Toxicon,
1917–1921.
Keller, M.D., Guillard, R.R.L., 1987. Media for the culture of
oceanic ultraphytoplankton. J. Phycol. 23, 633–638.
Lassus, P., Fremy, J.M., Ledoux, M.B.M.B.M., 1989. Patterns of
experimental contamination by Protogonyaulax tamarensis in
some French commercial shellfish. Toxicon 27, 1313–1321.
Lassus, P., Ledoux, M., Bardouil, M., Bohec, M., 1993. Influence of
initial toxicity and extraction procedures on paralytic toxin
changes in the mussel. Toxicon 31, 237–242.
Lavens, P., Sorgeloos, P., 1996. Manual on the production and use
of live food for aquaculture, FAO Fisheries Technical Paper
361, http://www.fao.org/DOCREP/003/W3732E/w3732e00.
htm.
Louzao, M.C., Lopez-Corona, P.P., Vieytes, M.R., Botana, L.M.,
1992. Union de saxitoxina a receptores de hepatopancreas de
mejillon mytilus galloprovincialis LmK. Cuadernos de Area de
Ciencias Marinas. Seminario de Estudios Galegos 6, 65–72.(in
Spanish).
Murakami, R., Yamamoto, K., Noguchi, T., 1999a. Difference in
PSP composition among various ports of surf clam. J. Food Hyg.
Soc. Jpn 40, 55–61.
Murakami, R., Yamamoto, K., Noguchi, T., 1999b. Toxicity and
paralytic shellfish poison composition of three species of
bivalves collected in Ibaraki Prefecture, Japan. J. Food Hyg.
Soc. Jpn 40, 46–54.
Oshima, Y., 1995. Chemical and enzymatic transformation of
paralytic shellfish toxins in marine organisms. In: Lassus, P.,
Arzul, G., Erard, E., Gentien, P., Marcaillou, C. (Eds.), Harmful
Marine Algal Blooms, Lavoisier/Intercept, Paris, pp. 475–480.
Sakamoto, S., Sato, S., Ogata, T., Kodama, M., 2000. Formation of
intermediate conjugates in the reductive transformation of
gonyautoxins to saxitoxins by thiol compounds. Fish. Sci. 66,
136–141.
Sekiguchi, K., Ogata, T., Kaga, S., Yoshida, M., Fukuyo, Y.,
Kodama, M., 2001. Accumulation of paralytic shellfish toxins in
the scallop Patinopecten yessoensis caused by the dinoflagellate
Alexandrium catenella in Otsuchi Bay, Iwate Prefecture,
northern Pacific coast of Japan. Fish. Sci. 67, 1157–1162.
Shimizu, Y., Yoshioka, M., 1981. Transformation of paralytic
shellfish toxins as demonstrated in scallop homogenates.
Science 212, 547–549.
Silvert, W.L., Cembella, A.D., 1995. Dynamic modelling of
phycotoxin kinetics in the blue mussel, Mytilus edulis, with
implications for other marine invertebrates. Can. J. Fish. Aquat.
Sci. 52, 521–531.
Silvert, W., Bricelj, M., Cembella, A., 1998a. Dynamic modelling of
PSP toxicity in the surfclam (Spisula solidissima). In: Reguera,
J. Blanco et al. / Toxicon 42 (2003) 777–784 783
B., Blanco, J., Fernandez, M.L., Wyatt, T. (Eds.), Harmful
Algae, Xunta de Galicia and IOC of UNESCO, Santiago de
Compostela, pp. 437–440.
Silvert, W., Bricelj, M., Cembella, A., 1998b. Dynamic modelling
of PSP toxicity in the surfclam (Spisula solidisssima): multi-
compartmental kinetics and biotransformation. In: Reguera, B.,
Blanco, J., Fernandez, M.L., Wyatt, T. (Eds.), Harmful Algae,
Xunta de Galicia and IOC of UNESCO, Santiago de
Compostela, pp. 437–440.
Sullivan, J.J., Iwaoka, W.T., Liston, J., 1983. Enzimatic transform-
ation of PSP toxins in the littleneck clam (Protothaca staminea).
Biochem. Biophys. Res. Commun. 114, 465–472.
J. Blanco et al. / Toxicon 42 (2003) 777–784784