ELSEVI ER Journal of Chromatography A, 729 (1996) 381-386

JOURNAL OF
CHROMATOGRAPHY A
Fluorimetric determination of diarrhetic shellfish toxins in
scallops and mussels by high-performance liquid
chromatography
Kazuaki Akasaka, Hiroshi Ohrui*, Hiroshi Meguro, Takeshi Yasumoto
Department of Applied Biological Chemistry, Faculty of Agriculture, Tohoku University, Tsutsumidori-Amamiyamachi I- 1,
Aobaku 981 Sendal Japan
Abstract
The f l uor i met r i c det er mi nat i on of okadai c aci d ( OA) and di nophysi st oxi n-1 ( DTX- 1) , t he pr i nci pal t oxi ns of
di ar r het i c poi soni ng, is r epor t ed. The di gest i ve gl ands of mussel s or scal l ops wer e homogeni zed with 2- pr opanol .
OA and DTX-1 wer e ext r act ed f r om t he homogenat e, wi t h he xa ne - e t hyl acet at e and l abel l ed with 2, 3-(ant h-
r acenedi car boxi mi do) et hyl t r i f l uor omet hanesul f onat e in dr y acet oni t r i l e. Af t er cl eani ng up by passage t hr ough a
shor t silica gel col umn, t he f l uor escent der i vat i ves were det er mi ned by HPLC. The der i vat i ves were at first
s epar at ed on a Devel osi l Ph-5 col umn, and onl y t he t ar get fract i on obt ai ned was i nt r oduced i nt o a Devel osi l ODS
K-5 col umn by a val ve-swi t chi ng devi ce. Bot h t oxi ns wer e det er mi ned in t he r ange 2. 5- 500 pg, and t he det ect i on
l i mi t s wer e 0.8 pg ( OA) and 1.3 pg ( DTX- 1) with a si gnal -t o-noi se r at i o of 3.
Keywords: Di ar r het i c shellfish poi soni ng; Okadai c acid; Di nophys i s t oxi n- l ; Toxi ns
1. I n t r o d u c t i o n
Diarrhetic shellfish poisoning ( DSP) is a
worldwide probl em of public health and in the
shellfish industry [1-4]. Up to the present, about
ten fat-soluble polyethers are known to be DSP
toxins [5-8]. Among these, okadaic acid ( OA)
and dinotphysistoxin-1 (35-methylokadaic acid)
(DTX-1) are the principal toxins, with free
carboxylic acid groups (Fig. 1).
A mouse bioassay met hod has been widely
used for the assay of the toxins [9]. Recently,
some HPLC met hods have been pr oposed for
* Corresponding author.
determining these toxins individually [10-13].
These met hods involved a highly sensitive and
selective det ect i on system as fluorimetry [10-12]
with 9-anthroyldiazomethane and ionspray mass
spect romet ry [13], and made it possible to de-
termine these toxins at nanogram levels. An
immunoassay met hod has also been report ed for
H "'"
Okadal acid : RI =R2=H
OTX-1 : RI =H, R2=-CH~
Fig. 1. Structures of OA and DTX-1.
()t121-9673/96/$15.00 1996 Elsevier Science B.V. All rights reserved
SSDI 0021- 9673( 95) 00890- X
382 K. Ak as ak a et al. / J. Chromat ogr. A 729 (1996) 381- 386
the determination of total of OA and DTX-1
[14]. The detection limit of the method was 5
ng/ml.
We have developed 2,3-(anthracenedicarbox-
imido)ethyl trifluoromethanesulfonate (AE-OTf)
as a highly sensitive fluorescent labelling reagent
for carboxylic acids [15]. Here, we report the
determination of OA and DTX-1 after derivati-
zation with AE-OTf followed by analysis with an
LC- LC system.
2. Experimental
2.1. Materials
Standard OA and DTX-1 were prepared by
the method of Lee et al. [10]. n-Hexane, di-
chloromethane, acetone, ethyl acetate and 2-pro-
panol were used as received. Acetonitrile, which
was distilled under P205 and stored over molecu-
lar sieves 4A, was used as the solvent for both
AE-OTf and tetraethylammonium carbonate
(TEAC) solutions. AE-OTf and TEAC were
prepared by a previously reported method [15].
AE-OTf solution was stored below 0C and used
within 1 h after preparation. Methanol and water
were of HPLC analysis grade. The solvents were
purchased from Wako (Osaka, Japan) or Kanto
Chemical (Tokyo, Japan). LiChrolute Si 60
(Merck, Darmstadt, Germany) was used in a
silica gel column to clean up OA and DTX-1
derivatives before HPLC analysis.
Scallops were purchased from a food market in
Sendai (Japan) and mussels were frozen product
from New Zealand.
screw-cap. To the test-tube, 0.75 ml of n-hexane,
1.0 ml of ethyl acetate and 1.5 ml of 0.32 M
Na2SO 4 solution were added and vortex mixed
for 1 rain, then the mixture was centrifuged at
2000 rpm for 5 min below 5C. After the upper
phase had been removed, 0.5 ml of ethyl acetate
was added for extraction. After centrifugation,
the two upper phases were combined and diluted
to 5 ml with methanol.
2.3. Derivatization and clean-up procedure
An aliquot (0.5 ml) of the extract was placed
in a test-tube and 75 /zl of TEAC solution (1.5
mM) were added. The solvent was removed
under reduced pressure followed by complete
drying under a stream of nitrogen. To the res-
idue, 100/zl of AE-OTf solution (2.25 mM) were
added and vortex mixed for 30 s, and the mixture
was reacted for more than 10 min at room
temperature. After the reaction, the solvent was
removed under reduced pressure and the residue
was dissolved in 200/zl of dichloromethane. The
solution was loaded on to a short column packed
with LiChrolute Si 60 (100-120 mg), which was
washed with 2 ml of dichloromethane before
loading. After loading, the column was washed
with 200/zl of dichloromethane followed by 4 ml
of dichloromethane-acetone (975:25, v/v). The
OA and DTX-1 derivatives were eluted with 2
ml of dichloromethane-acetone-methanol
(95:5:10, v/v/v). After removal of the solvent
under reduced pressure, the residue was dis-
solved in 200/zl of acetonitrile and an aliquot (2
/zl) was injected into the HPLC system.
2.2. Extraction of OA and DTX-1 f rom
digestive glands
2.4. Separation of AE derivatives of OA and
DTX-1
The extraction procedure used was a slightly
modified method of Pereira et al. [12]. One gram
of the digestive gland of a scallop or a mussel
was homogenized with 5 ml of 2-propanol. The
homogenate was centrifuged at 6000 rpm for 5
min below 5C and the precipitate was homogen-
ized again with 5 ml of 2-propanol. The two
supernatants were combined and an aliquot (0.5
ml) was transferred into a test-tube fitted with a
The system consisted of two PU-980 pumps
(JASCO, Tokyo, Japan), a Rheodyne Model
7125 sample injector, an HV-992-01 valve-switch-
ing device and an FP-920 spectrofluorimeter
(JASCO).
First, a sample was injected on to a Develosil
Ph-5 (5 /zm) column (50 mm4. 6 mm I.D.)
(Nomura Chemical, Aichi, Japan) and eluted
with methanol-water (80:20, v/v) at 0.8 ml/min.
K. Ak as ak a et al. / J. Chromat ogr. A 729 (1996) 381- 386 383
The el uat e be t we e n 2 mi n 45 s and 4 mi n 5 s was
i nt r oduced i nt o a Devel osi l ODS K-5 (5 / zm)
col umn (150 mm 4 . 6 mm I. D. ) ( No mu r a
Chemi cal ) , whi ch was el ut ed wi t h me t h a n o l -
wat er (80:20, v/ v) at 0.8 ml / mi n. Dur i ng t he
s epar at i on, bot h col umns wer e kept at 60C in a
wat er - bat h. The f l uor escence i nt ensi t i es of t he
OA and DTX- 1 der i vat i ves wer e moni t or e d at
462 nm ( exci t at i on at 298 nm) and t hei r pe a k
ar eas wer e cal cul at ed wi t h a Ch r o ma t o c o r d e r 12
( Sys t em I ns t r ument , Tokyo, Japan) .
2.5. Recovery test f rom scallop and mussel
homogenate
Vol umes of s t andar d OA and DTX- 1 sol ut i ons
wer e pl aced in a t es t - t ube so as t o be equi val ent
t o 50 or 500 ng of each t oxi n. Af t e r t he sol vent
had be e n r e mo v e d under r educed pr essur e, 0.5
ml of 2- pr opanol h o mo g e n a t e f r om a scal l op or a
mussel was added, f ol l owed by vor t ex mi xi ng for
mor e t han 30 s. OA and DTX- 1 wer e ext r act ed
f r om t he s ampl e wi t h n - h e x a n e - e t h y l acet at e
f ol l owed by l abel l i ng wi t h AE- OTf , cl ean- up and
HPLC anal ysi s as des cr i bed above. The re-
cover i es wer e es t i mat ed usi ng t he s t andar d OA
and DTX- 1 der i vat i ves, whi ch wer e cl eaned up
wi t h a silica gel col umn as des cr i bed above.
3. Re s u l t s ant i d i s c u s s i o n
3.1. Extraction, derivatization and clean-up
The ext r act i on pr oc e dur e r e por t e d by Per ei r a
et al. [12] was modi f i ed in its scale. Wi t h this
pr ocedur e, bot h OA and DTX- 1 wer e quant i t a-
t i vel y ext r act ed. The t oxi ns wer e der i vat i zed wi t h
AE - OT f in t he pr es ence of T E AC as a base.
Whe n OA and DTX- 1 in a scal l op or a mussel
ext r act wer e l abel l ed wi t h AE- OTf , bot h t oxi ns
gave al mos t t he s ame pe a k ar eas wi t h t he use of
mor e t han 2 mM AE- OTf ~ and t he r eact i ons
wer e c ompl e t e d wi t hi n 10 rain in each case. OA
and DTX- 1 wer e added to t he h o mo g e n a t e of a
scal l op or a mussel t o gi ve 100 n g / ml of each
t oxi n. In t he pr opos e d me t hod, bot h t oxi ns in an
ext r act wer e l abel l ed wi t h 2.25 mM AE - OT f for
mor e t han 10 mi n at r o o m t e mpe r a t ur e . Thes e
condi t i ons ma de it possi bl e t o l abel at l east up t o
1 0 / z g / g of t oxi ns in a scal l op and a mussel .
Al t hough t he AE - OT f sol ut i on in dr y acet oni -
trile gave al mos t const ant pe a k ar eas even 6 h
af t er pr e pa r a t i on when it was kept bel ow 0C, it
was used wi t hi n 1 h.
Af t e r t he r eact i on, t he mi xt ur e was cl eaned up
by pas s age t hr ough a shor t silica gel col umn t o
r e move i nt er f er i ng subst ances, such as f at t y acid
der i vat i ves and decompo. sed pr oduct s of t he
r eagent . Tabl e 1 shows t he r ecover i es of bot h
t oxi ns in t he el ut i ng sol vent [2 ml of di chl or ome-
t h a n e - a c e t o n e - me t h a n o l (95:5:10, v / v / v ) ] af t er
ri nsi ng t he col umn wi t h 4 ml of di chl or ome-
t h a n e - a c e t o n e . The r e wer e al mos t no si gni fi cant
di f f er ences in t he r e mova l of i nt er f er i ng sub-
st ances when mor e t han 100 mg of silica gel was
pa c ke d in t he col umn. As nei t her OA nor DTX-
1 der i vat i ves wer e el ut ed by di chl or omet hane, it
was used as t he sol vent t o wash t he col umn and
t o l oad a sampl e. Whe n t he col umn was ri nsed
wi t h 4 ml of d i c h l o r o me t h a n e - a c e t o n e (950:50 or
966:34, v/ v) . s ma l l a mount s of bot h t oxi ns wer e
Ta bl e 1
Ef f ect of wa s hi ng s ol ve nt s on t he r ecover i es of OA and DT X- I der i vat i ves f r om a s hor t silica gel c ol umn
Wa s hi ng s ol vent
d i e h l o r o me t h a n e - a c e t o n e
Re c ove r y ( %)
OA der i vat i ve DTX- 1 der i vat i ve
950:5(I 83. 8 69. 6
966:34 87.1 84.2
975:25 88. 0 89.1
Sa mpl e ( 200/ z l ) was i nj ect ed on t o t he c o l u mn as a d i c h l o r o me t h a n e s ol ut i on a nd r i ns ed wi t h 200 #1 of d i c h l o r o me t h a n e f ol l owed
by r i ns i ng wi t h 4 ml of wa s hi ng s ol vent . Th e aci d der i vat i ves wer e e l ut e d wi t h 2 ml of d i c h l o r o me t h a n e - a c e t o n e - me t h a n o l
(95: 5: t 0).
384 K. Akasaka et al. / J. Chromatogr. A 729 (1996) 381-386
eluted in the eluate, and their recoveries were
lower. On the other hand, by rinsing with 4 ml of
dichloromethane-acetone (975:25, v/v), neither
toxin was eluted in the eluate, and almost 90% of
the toxins were recovered in the eluting solvent.
3.2. HPLC separation
Although most interfering substances were
removed by passage through a silica gel column,
it was impossible to remove interferents com-
pletely because of the trace levels of OA and
DTX-1 and the high sensitivity of the method.
To remove these interfering substances, we em-
ployed a two-column system equipped with a
valve-switching device. First, a sample was
loaded on to a Develosil Ph-5 column, from
which OA and DTX-1 derivatives were eluted
very close together, and most peaks were eluted
within 20 min (Fig. 2A, B and C). The target
fraction 1 (fr. 1) which contained both toxin
derivatives were cut out and loaded by valve
switching on to a Develosil ODS K-5 column,
which gave different separation patterns to those
on the first column. Fig. 3 shows the effects of
the timing of the valve switching on the peak
areas of both toxin derivatives. On cutting out a
wider range than 2 min 45 s and 4 min 5 s, their
peak areas were almost constant.
Fig. 2 also shows typical chromatograms of a
standard mixture of OA and DTX-1 (D) and
toxin-free mussel extract (E) separated by the
LC- LC system. Part (a) in Fig. 2E is magnified
eightfold in sensitivity in Fig. 2F.
The interfering substances in the sample were
effectively removed by the proposed system. As
we used a two-column system which gave differ-
ent separation patterns, this method made it
possible to assign the toxins from their retention
times with higher accuracy than using a single-
column system.
3.3. Quantitativeness, sensitivities,
reproducibilities and recoveries of OA and
DTX- 1
Table 2 gives the data for standard OA and
DTX-1. With the proposed method, the cali-
bration graphs for both toxins show good lineari-
D
A I B
D
2 's
fr.
a
b
i ' s o 15
r a i n
E
F
~
. 12_5 2 5
() 2"5 m i n
Fig. 2. Typical chr omat ogr ams obt ai ned using a Develosil
Ph-5 col umn (A, B and C) and an LC- LC system (D, E and
F). Samples were t he extracts of a scallop (A) and a mussel
(B, E and F) and a st andard mi xt ure of OA and DTX-1 (C
and D). Peaks a and b are derivatives of OA and DTX-1,
respectively. Fr. 1 is a fraction t hat cont ai ned OA and DTX-
1. Part (a) in E is magnified eightfold in sensitivity in F.
ty at least between 2.5 and 500 pg on-column,
and their detection limits were 0.8 pg (OA) and
1.3 pg (DTX-1) at a signal-to-noise ratio of 3.
The determination ranges of the toxins in scallop
and mussel were 0.05-10 /zg/g and their de-
tection limits were 20-30 ng/g. The sensitivities
x lOs i-
m4
0
I,.. A'
Io
m2
0.
0
On; 2min 30sec
Off; 4rain 20see
2min 45sec 3rain 00sec
4min 05sec 3min 50sec
Fig. 3. Effect of valve switching t i mi ng on peak areas of ( 0 )
OA and ( &) DTX-1.
K. Akasaka et al. / J. Chromatogr. A 729 (1996) 381-386 385
Table 2
Quantitativeness, sensitivities and reproducibilities for OA
and DTX-1 derivatives
Parameter OA derivative DTX-1 derivative
r" 0.9997 0.9991
(range) (2. 5-500 pg) ( 2. 5- 500 pg)
Detecti on limit (pg) 0.8 1.3
( S / N=3 )
R.S.D. (%)b (n = 5) 5.0 (50 pg) 4.0 (50 pg)
" r = Correlation coefficient.
b R.S.D. = relative standard deviation of peak area
were more than ten times higher than those
reported for HPLC methods [10-13].
Table 3 gives the recoveries of OA and DTX-1
which were added to a 2-propanol homogenate
of a scallop or a mussel. The relative standard
deviations for OA and DTX-1 were about 1%
( 10/ z g/ g) and 4% (1 ~g/ g) , respectively. Using
the proposed method, OA and DTX-1 were
almost quantitatively recovered from a homoge-
nate of a scallop and a mussel and labelled with
AE-OTf.
3.4. Application of the method
The samples (five scallops and five mussels)
were analysed by the proposed method. Fig. 4
shows typical chromatograms of a scallop (A)
and a mussel (D). Parts (a) and (b) are magnified
two- and eightfold in sensitivity in Fig. 4C and F,
respectively. Fig. 4B and E are chromatograms
C Ix2
i A /,~
12.5 25
0 2 5 r a i n
I L
E !I ) 8
r a i n / 4 ( b ) j . 12.5
0 2 5 m i n
rain
Fig. 4. Typical chromatograms for OA and DTX-1 in a
scallop (A, B and C) and a mussel (D, E and F). A and D are
chromatograms of extracts from a scallop and a mussel,
respectively; B and E are chromatograms of extracts from a
scallop and a mussel to which a mixture of OA and DTX-I
was added (10 ng/ ml of each toxin), respectively; C and F are
chromatograms with parts (a) and (b) magnified by two- and
eightfold in sensitivity, respectively.
for the 2-propanol homogenate of a scallop and a
mussel to which both OA and DTX-1 were
added to give 10 ng/ ml of each toxin. As shown
in Fig. 4A and C, a peak at 23,3 min had the
same retention time as DTX-1, whereas no peak
was detected at 16.3 min where OA should be
eluted. On the other hand, a small peak at 16.3
min was detected in a mussel (Fig. 4D and E).
The data are given in Table 3. Data for other
samples are given in Table 4. Although no or
trace peaks of OA were detected in all the
samples analysed, some samples gave an appar-
Table 3
Recoveries of OA and DTX- I from a homogenat e of a scallop or a mussel
Homogenatecompound Added (p~g/g)
0 1.0
Found (k~g/g) Found (/ xg/ g)
l0
Recovery ( %) Found (kLg/g) Recovery ( %)
Scal l opOA n.d." (n = 5)
DTX-1 0.37 0.03 (n = 5)
Mussel OA <0.05 b (n = 4)
DTX-1 n.d. (n = 4)
0.91 0.06 (n =5) 91
1.53 0.07 (n = 5) 116
1.03 _+ 0.04 (n = 4) 103
1.15 _+ 0.05 (n = 4) 115
10,36 + 0.12 (n = 5) 103.6
11.32 + 0.13 (n =5) 109.5
10.34 0.10 (n - 4) 103.4
10.91 0.14 (n = 4) 109.1
~'n.d. = Not detected.
~<0.05 = peak was detected.
386 K. Ak as ak a et al. / J. Chromat ogr. A 729 (1996) 3 8 1 - 3 8 6
Table 4
OA and DTX-1 determined in scallops and mussels
Sample OA ( / zg/ g) DTX-1 ( / zg/ g)
Scallop A <0.05" 0.10
B <0.05 0.07
C n.d. ~ <0.05
D <0.05 0.06
Mussel E n.d. n.d.
F n.d. 0.07
G n.d. <0.05
H n.d. n.d.
n.d. = Not detected.
b <0.05 = peak was detected.
ent peak of DTX-1. The levels of OA and DTX-
1 detected, however, were too low for other
HPLC methods to detect them and below the
limit defined by the law.
Acknowl edgement
This work was partly supported by a Grant-in-
Aid from the Ministry of Education, Science and
Culture of Japan.
[2] M. EA. Leitao, P. Lassus, P. Maggi, C.L. Baut, J. Chauvin
and P. Truguet, Tray. Int. Peches Marit., 46 (1984) 233.
[3] B. Underdal, M. Yndestad and T. Aune, in D.M. Ander-
son, A.W. White and D.G. Baden (Editors), Toxic
Diniflagellates, Elsevier, New York, 1985, pp. 489-494.
[4] E. Dahl and M. Yndestad, in D.M. Anderson, A.W.
White and D.G. Baden (Editors), Toxic Diniflagellates,
Elsevier, New York, 1985, pp. 495-500.
[5] J.-S. Lee, K. Tangen, E. Dahl, E Hovgaard and T.
Yasumoto, Bull. Jpn. Soc. Sci. Fish., 54 (1988) 1953.
[6] M. Murata, M. Shimatani, H. Sugitani, Y. Oshima and T.
Yasumoto, Bull. Jpn. Soc. Sci. Fish., 48 (1982) 549.
[7] T. Yasumoto, M. Murata, Y. Oshima, G.K. Matsumoto
and J. Clardy, Tetrahedron, 41 (1985) 1019.
[8] M. Murata, M. Sano, T. Iwashita, H. Naoki and T.
Yasumoto, Agric. Biol. Chem., 50 (1986) 2693.
[9] Ministry of Health and Welfare, Japan, Food Sanit. Res.,
31 (1981) 565.
[10] J.-S. Lee, T. Yanagi, R. Kenma and T. Yasurnoto, Agric.
Biol. Chem., 51 (1987) 877.
[11] B. Luckas and R. Meixner, Z. Lebensm.-Unters.-
Forsch., 187 (1988) 421.
[12] A. Pereira, D. Klein, K. Sohet, G. Houvenaghel and J.C.
Braekman, in Abstracts of Sixth International Confer-
ence on Toxic Marine Phytoplankton, 1993, p. 159.
[13] S. Pleasance, M.A. Quilliam, A.S.W. De Freitas, J.C.
Marr and A.D. Cembella, Rapid Commun. Mass Spec-
trom., 4 (1990) 206.
[14] T. Uda, Y. Itoh, M. Nishimura, T. Usagawa, M. Murata
and T. Yasumoto, Bioact. Mol., 10 (1989) 335.
[15] K. Akasaka, H. Ohrui and H. Meguro, Analyst, 118
(1993) 765.
References
[1] T. Yasumoto, Y. Ohshima and M. Yamaguchi, Bull. Jpn.
Soc. Sci. Fish., 44 (1978) 1249.