JFS: Food Chemistry and Toxicology

Identification of Toxin and Fish

Species in Cooked Fish Liver
Implicated in Food Poisoning
Y.W. H SIEH, Y.C. SHIU, C.A. CHENG , S.K. C HEN , AND D.F. H WANG
Food Chemistry and Toxicology

ABSTRACT: The cooked fish liver retained by the victims was assayed for toxicity and mitochondrial DNA.
Meanwhile, 8 live specimens of puffer Takifugu niphobles were also assayed. The toxicity of cooked fish liver was
280 6 20 mouse units per gram (MU/g). All specimens of T. niphobles showed high toxicity (more than 850 MU/
g) in the liver. The toxin from cooked fish liver and liver of T. niphobles was identified to be tetrodotoxin. The
cooked fish liver and fresh liver of T. niphobles showed the same sequence genotype and the same single restric-
tion site for Bsa I. Therefore, the species of cooked fish liver was suggested as T. niphobles.
Keywords: food poisoning, tetrodotoxin, puffer, species identification

Introduction amplification and analyses of mitochondrial DNA sequence

A FOOD POISONING INCIDENT DUE TO INGESTING FISH OC -

curred in Chunghua County, western Taiwan on January
2, 2000. A total of 5 victims (4 men, 58 to 64 y old and 1 wom-
and restriction enzyme were used to identify the puffer spe-
cies of cooked fish liver.

an, 46 y old) were reported. Symptoms featured paralysis,
coma, nausea, vomiting, ataxia, aphasia, and difficulty of res-
piration. Among these victims, 2 men suffered with more se-
rious symptoms and were treated with intravenous fluids,
mechanical ventilation, and intensive treatment in the hospi-
tal. They were then discharged uneventfully after 1 week of
management. Presently, all patients are normal and healthy.
The fish, which might be a puffer reported by victims, was
collected from the coastal area of Chunghua county. The
small residue (11.02 g) of cooked fish liver was retained by
the victims in this incident.
These symptoms were similar to those of tetrodotoxin
(TTX) poisoning (Hashimoto 1979). TTX was believed to be
specific to puffers and was first isolated from a puffer (Yokoo
1950). In Taiwan, puffer-associated poisoning has been re-
ported sporadically (Hwang and others 1995; DOH 1997).
However, the species implicated in these food poisoning inci-
dents were not identified so far because of the lack of mor-
phological characteristics in the causative residues. To com-
pare with the retained cooked fish liver, live specimens of
puffer Takifugu niphobles were collected from the same
coastal waters and used for assay. Although the puffer T.
niphobles is a strong toxic puffer in Japan (Tani 1951), the
data of anatomical toxicity in Taiwan was not available
(Hwang and others 1992).
Recently, it has been shown that analysis of mitochondrial
DNA was successful in differentiating species of fish (Borgo
and others 1996; DeSalle and Birstein 1996; Jones 1991; Un-
seld and others 1995). There also have several papers de-
scribing polymerase chain reaction (PCR) amplification and
restriction enzyme analysis of the cytochrome b gene as use-
ful for identification of fish species (Ram and others 1996;
Quinteiro and others 1998; Cespedes and others 1998).
Therefore, the present study was undertaken to identify the
responsible toxin in cooked fish liver, along with establishing
the toxicity data of T. niphobles. Then, application of PCR
Materials and Methods
Materials
The remaining uneaten cooked fish liver was weighed
(11.02 g) and kept frozen below –20
C until use. Eight live
specimens of the puffer Takifugu niphobles were collected
from the coastal waters of Chuanghua County, western Tai-
wan in February 2000 and immediately transferred on ice to
the laboratory. The liver of each specimen was removed
and kept frozen below –20
C for use. Each portion of
cooked fish liver and fresh liver of T. niphobles was used
for DNA extraction. Meanwhile, some specimens of both
toxic puffers Lagocephalus lunaris and T. oblongus were
collected from Kaohsiung County, southern Taiwan, and
another toxic puffer T. rubripes collected from Taipei
County, northern Taiwan. The fresh liver of these toxic
puffers was removed for DNA extraction and then used for
the restriction site analysis of the cytochrome b gene. The
liver of nontoxic puffer Lagocephalus gloveri was collected
from Keelung, northern Taiwan and used for control analy-
sis. All chemicals (GR grade) and solvents (LC grade) were
purchased from Sigma (St. Louis, Mo., U.S.A.) and E. Merck
(Darmstadt, Germany).
Assay of toxicity
The livers of toxic puffer T. niphobles and nontoxic puffer
L. gloveri and cooked fish liver were partially thawed, and
examined for toxicity by the tetrodotoxin (TTX) bioassay
(Hwang and Jeng 1991). Toxicity was expressed in mouse
units. One mouse unit (1 MU) being equal to 0.178 g TTX is
defined as the minimum amount of toxin required for killing
an ICR (Institute of Cancer Research) strain male mouse of
20 g weight in 30 min after 1 intraperitoneal injection.
Purification and identification of toxin
Cooked fish liver (5 g) and the liver (10 g) of T. niphobles

948 JOURNAL OF FOOD SCIENCE—Vol. 67, Nr. 3, 2002 © 2002 Institute of Food Technologists

jfsv67n3p0948-0952ms20010106-MO-SR-wcxs.p65
948 5/1/2002, 9:19 AM
 Identification of toxin and fish species

Table 1—Anatonical distribution of toxicity in the puffer Takifugu niphobles collected from Changhua County in Febru-
ary 2000
Sample Body Body Toxicity (MU/g)
No. weight (g) length (cm) Muscle Liver Skin Intestine Bile Ovary Testis
1 74.83 14.90 46  6*1 1810  30 70 5 ND*2 460  30 1090  90 –*3
2 59.52 13.49 35  4 1760  40 65 8 91 510  30 – 11  2
3 50.13 14.45 51  3 850  30 91 9 12  2 210  20 1400  110 –
4 47.43 12.80 28  2 1240  50 180  10 ND 400  30 900  20 –
5 31.48 10.92 29  3 1110  30 160  20 83 940  30 – 72
6 26.03 11.25 54  6 950  50 110  10 ND 990  60 870  10 –
7 28.43 11.00 27  4 840  30 100  10 41 370  40 – 16  1
8 39.21 12.65 52  6 1430  60 160  20 72 580  30 1100  100 –
1 Means  S.D. (n = 2)
2 ND means less than 4 MU/g

Food Chemistry and Toxicology

3 No data

were separately homogenized for 5 min with 3 volumes of
1% acetic acid in methanol and centrifuged (3000  g) for 15
min. The residue was extracted 2 more times in the same
manner. The supernatants were combined, concentrated un-
der reduced pressure, and defatted 3 times with dichlo-
romethane. The aqueous layer (cooked fish liver, 1200 MU; T.
niphobles liver, 17600 MU) was ultrafiltered with Amicon
YM-1 membrane to cut off substances of more than 1000
daltons. The filtrate was examined for toxicity by using TTX
bioassay and freeze-dried.
The lyophilizate was dissolved in 5 ml of 0.03 M acetic
acid and chromatographed on a Bio-Gel P-2 column (2 
98 cm) using 0.03 M acetic acid as eluant. The toxic frac-
tions were monitored by using TTX bioassay, collected,
and combined. Finally, the toxin fractions were lyophilized
to afford partially purified toxins of 1.9 mg (580MU/mg)
for cooked fish liver and 19.1 mg (560MU/mg) for T.
niphobles liver.
Each partially purified toxin was submitted to high per-
formance liquid chromatography (HPLC) as described pre-
viously (Hwang and others 1988). In brief, HPLC was per-
formed on a reversed-phase column (Merck Lichrospher
100 RP-18, 4 mm I.D.  250 mm). The mobile phase was so-
dium 1-heptane sulfonate (2 mM) in methanol (1%)-potas-
sium phosphate buffer (0.05 M, pH 7.0). The flow rate was
0.6 mL/min. The eluate was mixed with an equal volume of
NaOH (3 N), heated in a reaction coil ( Teflon tube, 3 mm
I.D. x 10 m) at 99
C, and then subjected to fluorescent
spectrophotometry (F-1000, Hitachi, Tokyo) and intensity
of fluorescence was measured at 505 nm with 380 nm exci-
tation. Authentic TTX, anhydrotetrodotoxin (anh-TTX),
along with trace 4-epi-TTX were obtained from the liver of
a puffer T. oblongus (Hwang and others, 1988) and used as
reference standards.
DNA extraction
Total cellular DNA was extracted from cooked fish liver
and the liver of puffers including T. niphobles, T. oblongus, T.
rubripes, and L. lunaris essentially according to DeSalle and
Birstein (1996). Briefly, about 1 g puffer muscle was homoge-
nized with extraction buffer (50 mM Tris-HCl, pH 8.0, 0.1 M
EDTA, 1% SDS, 0.2 M NaCl) and 50 L of 5 mg/mL protein-
ase K (Ameresco, Solon, Ohio, U.S.A.) were added. The sam-
ples were incubated overnight at 55
C with shaking. After in-
cubation, tubes were placed on ice for 30 min, centrifuged at
12000  g for 10 min and supernatant transferred to a clean
tube. DNA was extracted once with phenol, twice with phe-
nol-chloroform-isoamylalcohol in a 25:24:1 ratio and once
with chloroform, and then precipitated twice with ethanol at
-20
C. The dried pellets were resuspended in 50 to 100 L
sterile distilled water and the concentration of DNA estimat-
ed by absorbance at 260 nm.
PCR amplification of a fragment of the cytochrome b
gene
The PCR amplification reactions were performed in a to-
tal volume of 100 L. Each reaction mixture contained 1000
ng of extracted template DNA, 0.4 M of each primer, 200
M of each dNTP and 2.5 U of Pro Taq DNA polymerase
(Amresco, Solon, Ohio, U.S.A.) in a reaction buffer contain-
ing 20 mM Tris-HCl, pH 8.0, 15 mM MgCl 2, 1% Triton X-100,
0.1 mM DTT, and 50% glycerol.
The polymerase chain reaction was carried out in a Gene-
Amp PCR System 2400 (Perkin Elmer, Foster City, Calif.,
U.S.A.) programmed to perform a denaturation step of 95
C
for 10 min, followed by 40 cycles consisting of 1 min at 95
C,
1 min at 50
C and 2 min at 72
C. The last extension step was
10 min longer.
The set of primers used for PCR amplification were desig-
nated L14841: 5’AAAAAGCTTCCATCCAACCAACATCTCAG-
CATGATGAAA-3’ and H15149: 5’AAACTGCAGCCCCTCA-
GAATGATATTTGTCCTCA-3’. These primers corresponded
with those described by Kocher and others (1989).
The products of PCR amplification were analyzed by aga-
rose gel electrophoresis. PCR products (5 L) were mixed
with 1 L gel loading solution and loaded in a 1.2% agarose
gel containing 1 g/mL ethidium bromide in TBE buffer.
Electrophoresis separation and DNA fragments were operat-
ed at 100 V for 60 min.
Cleanup and sequencing of the PCR products
PCR product (60 L) was loaded onto a 2% agarose gel
containing 1 g/mL ethidium bromide in TBE buffer and
electrophoresed at 50 V for 120 min. The DNA band was
excised under UV light and melted in 5 volumes Tris EDTA
( TE) buffer at 65
C for 5 min. DNA was extracted twice
with phenol, once with phenol: chloroform: isoamylalco-
hol (24:24:1) and once with chloroform. Finally, the DNA
was precipitated with 1/10 volume 3 M sodium acetate
and 2 volumes ethanol. The dried pellet was resuspended
in 20 L sterile distilled water. The concentration and
quality of the DNA was estimated by agarose gel electro-
phoresis of a 2 L sample.
Purified PCR products from cooked fish liver and the
muscle of T. niphobles were sequenced at Mission Biotech
( Taipei, Taiwan) using the above primers and the ABI
PRISM BigDye Terminator Cycle Sequencing Ready Reac-
tion Kit (Perkin-Elmer/Applied Biosystems Division, Fos-

Vol. 67, Nr. 3, 2002—JOURNAL OF FOOD SCIENCE 949

jfsv67n3p0948-0952ms20010106-MO-SR-wcxs.p65
949 5/1/2002, 9:19 AM
 Identification of toxin and fish species
ter City, Calif., U.S.A.) in a ABI PRISM 377-96 DNA Se-
quencer (Perkin-Elemer/Applied Biosystems Division,
Foster City, Calif., U.S.A.). Two replicate sequences were
obtained from each sample. Sequence analysis was per-
formed using the Wisconsin Package, Version 10 (Genetics
Computer Group, 2000).
Restriction site analysis of PCR products
For the restriction site analysis of the cytochrome b re-
gion, the PCR products were extracted and purified as de-
scribed in the fourth paragraph of the materials and meth-
ods section. The endonuclease Bsa I (Promega, Madison,
Wisc., U.S.A.) was tested for restriction analysis of the am-
Food Chemistry and Toxicology
plified PCR products. Digests were performed in 10 L vol-
umes with 100 to 200 ng amplified DNA, 5U enzyme and
1:10 dilution of the manufacturer recommended 10  di-
gestion buffer and BSA. Digestions were incubated for 2 h
at 50
C. The resulting fragments were separated by electro-
phoresis in a 2.0% agarose gel containing 10 mg/mL ethidi-
um bromide for 1 h at 100 V. The sizes of the resulting DNA
fragment were estimated by comparison with a commercial
100 bp ladder (Protech Technology Enterprise Co., Taipei,
Taiwan).
Statistical analysis
All data were analyzed for the different treatments by
analysis of variance (ANOVA), and Duncan’s new multiple
range test was used to resolve the differences among treat-
ment means (Puri and Mukken 1980). A value of p  0.05 was
used as the criterion for significant differences.
Figure 1—HPLC eluting profile of toxins obtained from
cooked fish liver and the liver of Takifugu niphobles, along
with authentic tetrodotoxin (TTX), 4-epi-TTX, and
anhydrotetrodotoxin (anh-TTX)
950 JOURNAL OF FOOD SCIENCE—Vol. 67, Nr. 3, 2002
jfsv67n3p0948-0952ms20010106-MO-SR-wcxs.p65
950
Results and Discussion
HE TOXICITY OF COOKED FISH LIVER WAS 280  20 MU/ G.
T The anatomical distribution of toxicity in the puffer T.
niphobles was shown in Table 1. All specimens were toxic.
The highest toxicity in the muscle, liver, skin, intestine, bile,
ovary, and testis was 54  6, 1810  30, 180  10, 12  2,
990  60, 1400  110, and 16  1 MU/g, respectively. These
results indicated that the toxicity of T. niphobles was mainly
distributed in the liver, bile, and ovary and was similar to
those of other toxic puffers reported previously (Hwang
and others 1988). Meanwhile, T. niphobles specimens
showed a rather wide individual variation in toxicity, despite
the fact that they were collected from the same place on
the same day.
The toxicity of cooked fish liver retained by the victims was
significantly lower than that of liver in T. niphobles (p  0.05).
However, the minimum lethal dose of TTX for a person on
oral administration is assumed to be 10000 MU (Tani 1945);
the toxicity of cooked fish liver was quite high enough to in-
duce neurotoxicity observed if the patients took over 5 g.
The toxins were extracted and purified from cooked fish
liver and the liver of T. niphobles by the procedures de-
scribed above, resulting in partially purified toxin whose spe-
cific toxicity was 580 and 860 MU/mg for cooked fish liver
and the liver of T. niphobles, respectively. These values are
about 11.6 and 17.2% of specific toxicity (5000 MU/mg), of a
Figure 2—DNA sequence of part of the cytochrome b gene
from cooked fish liver and the liver of Takifugu niphobles,
T. rubripes, T. oblongus, and Lagocephalus lunaris.
Bsa I restriction site is shown with underline. The positions
of primers L14841 and H15149 used for PCR amplifica-
tion are shown with italic capitals.
5/1/2002, 9:19 AM
 Identification of toxin and fish species
pure TTX (Goto and others 1965). Figure 1 shows the HPLC
patterns of 2 toxins and the authentic toxin preparations.
Each toxin gave rise to 3 peaks whose retention times coin-
cided well with those of TTX, 4-epi-TTX and anh-TTX. The re-
sponded fraction from the liver of nontoxic puffer L. gloveri
did not present these peaks. It was suggestive from these re-
sults that the causative cooked fish liver was composed of
TTX and related substances.
The DNA extracted from cooked fish liver and the muscle
of other puffers was tested for amplification using the
L14841/H15149 primers, which should produce a 376-bp
fragment. The electrophoretic analysis of the PCR products
from these samples exhibited the same 376-bp fragment. It
meant that the proteinase K method appeared to be appro-
priate for extracting puffer tissues due to the high lysis pow-
er of proteinase K and SDS. Meanwhile, the 376-bp DNA
fragment was adequate size for the amplication of cyto-
chrome b gene from puffer tissues.
After comparing these DNA sequence in cooked fish liver
and the liver of T. niphobles, it was observed that all of them
were the same (Figure 2). This meant that the sequence of
376-bp region of cytochrome b gene in cooked fish liver re-
tained by the victims was identical with that of T. niphobles.
The restriction map of the sequence was searched for select-
ing endonucleases that could yield 2 fragments. The restric-
tion enzyme Bsa I was found potentially useful for this pur-
pose (Figure 2). Results following digestion of the puffer
products showed that band sizes obtained by electrophoresis
in 2.0% agarose gel were in agreement with the expected siz-
es for the restriction fragments inferred from the sequence
analysis. A single restriction site for Bsa I was found in the
sequence of PCR products from cooked fish liver and the liv-
er of T. niphobles, yielding 2 DNA fragments of 247 and 129
bp (Figure 3). However, no restriction site for Bsa I was
found in the sequence of other toxic puffers including T. ob-
longus, T. rubripes, and L. lunaris, and the undigested PCR
Figure 3—Electrophoretic analysis of PCR products of the
cytochrome b gene digested with Bsa I on 2.0% agarose
gel; M = MW marker, Bio 100 bp DNA Ladder; samples in
lanes are as follows: 1 = cooked fish liver, 2 = liver of
Takifugu niphobles, 3 = liver of T. rubripes, 4 = liver of T.
oblongus, 5 = liver of Lagocephalus lunaris.
jfsv67n3p0948-0952ms20010106-MO-SR-wcxs.p65
951
product was found. Hence, the utilization of restriction en-
donuclease Bsa I in the PCR product of cytochrome b gene
in the puffer T. niphobles is useful for differentiating species
of this puffer from other toxic puffers and identifying the
species of cooked fish liver to be T. niphobles. These facts in-
dicated that the species of cooked fish liver retained by the
victims was T. niphobles.
As described by Unseld and others (1995), the cyto-
chrome b gene as a molecular marker for investigating phy-
logenetic relationships within vertebrates is useful for sev-
eral reasons. First, because of the maternal inheritance of
mitochondria, normally only 1 allele exists per individual
and thus no sequence ambiguities are to be expected from
Food Chemistry and Toxicology
the presence of more than 1 allele. Second, the high abun-
dance of mitochondrial DNA in total cellular nucleic acid
preparations allows more effective PCR amplifications in
comparison to nuclear-encoded, single-copy gene. Third, in
vertebrates the mutation rate of mitochondrial genes is
near 10-fold higher compared to nuclear genes. Thus point
mutations accumulate quickly enough to allow (in most
cases) the discrimination of even closely related species.
Here, we found that direct sequence analysis and restric-
tion enzyme are applicative in identification of causative
species for puffer poisoning. The data of gene base in the
cytochrome b gene of puffer are still few (Brenner and oth-
ers 1993; Crnogorac-Jurcevic and others 1997; Elgar and
others 1996; Genetics Computer Group 2000), so further
study is needed.
References
Borgo R, Souty-Grosset C, Boucho D, Gomot L. 1996. PCR-RFLP analysis of mito-
chondrial DNA for identification of snail meat species. J Food Sci 61:1-4.
Brenner S, Elgar G, Sanford R, Macrae A, Venkatesh B, Aparicio S. 1993. Charac-
terization of the pufferfish (fugu) genome as a compact model vertebrate ge-
nome. Nature 366:265-268.
Cespedes A, Garcia T, Carrera E, Gonzalez I, Sanz B, Hernandz PE, Martin R.
1998. Identification of flatfish species using polymerase chain reaction (PCR)
amplification and restriction analysis of the cytochrome b gene. J Food Sci
63:206-209.
Crnogorac-Jurcevic T, Brown JR, Lehrach H, Schalkwyk LC. 1997. Tetradon fluvi-
atilis, a new puffer fish model for genome studies. Genomics. 41:177-184.
DeSalle R, Birstein VJ. 1996. PCR identification of black caviar. Nature 361:197-
198.
DOH (Dept. of Health, R.O.C.). 1997. Occurrence of food poisoning in Taiwan.
Taipei: Dept. of Health. P 37-38.
Elgar G, Sanford R, Aparicio S, Macrae A, Venkatesh B, Brenner S. 1996. Small is
beautiful: Comparative genomics with the pufferfish (Fugu rubripes). Trends
Genet 12(4):145-150.
Genetics Computer Group. 2000. Program Manual. Wisconsin Sequence Analy-
sis Package, Version 10. Madison, Wisc.: Genetics Computer Group, Inc.
Goto T, Hishi Y, Takahashi S, Hirata Y. 1965. Tetrodotoxin. Tetrahedron 21:2059-
2088.
Hashimoto Y. 1979. Toxins and other bioactive marine metabolites. Tokyo: Japan
Scientific Societies Press. P 70-91.
Hwang DF, Jeng SS. 1991. Bioassay of tetrodotoxin using ICR mouse strain. J
Chinese Biochem Soc. 20:80-86.
Hwang DF, Noguchi T, Arakawa O, Abe T, Hashimoto K. 1988. Toxicological stud-
ies on several species of puffer in Taiwan Bull Jap Soc Scient Fish 54:2001-
2008.
Hwang DF, Kao CY, Yang HC, Jeng SS, Noguchi T, Hashimoto K. 1992. Toxicity of
puffer in Taiwan. Nippon Suisan Gakkaishi 58:1541-1546.
Hwang DF, Cheng CA, Tsai YH, Jeng SS. 1995. Tetrodotoxin associated food poi-
soning due to unknown fish in Taiwan between 1988-1994. J Natural Toxins
4:165-171.
Jones JL. 1991. DNA probes: Applications in the food industry. Trends Food Sci
Technol 2:28-32.
Kocher TD, Thomas WK, Meyer A, Edwards S, Paabo S, Villablanca FX, Wilson AC.
1989. Dynamics of mitochondrial DNA evolution in animals: Amplification
and sequencing with conserved primers. Proc Nat Acad Sci USA 66:6196-6200.
Puri SC, Mukken K. 1980. Multiple comparisons. In: Hall GK, editor. Applied sta-
tistics for food and agricultural scientists. Boston, MA: Medical Publishers. P
146-162.
Quinteiro J, Sotelo CG, Rehbein H, Pryde SE, Medina I, Perez-Martin RI, Rey-Men-
dez M, Mackie IM. 1998. Use of mtDNA direct polymerase chain reaction (PCR)
sequencing and PCR-restriction fragment length polymorphism methodologies
in species identification of canned tuna. J Agric Food Chem 46: 1662-1669.
Ram JL, Ram ML, Baidoun FF. 1996. Authentication of canned tuna and bonito by
sequence and restriction site analysis of polymerase chain reaction products
Vol. 67, Nr. 3, 2002—JOURNAL OF FOOD SCIENCE 951
5/1/2002, 9:19 AM
 Identification of toxin and fish species

of mitochondrial DNA. J Agric Food Chem 44:2460-2467.
Tani I. 1945. A study of the toxicity of Japanese fugu. Tokyo: Teikoku-tosho Co. 103
p.
Unseld M, Beyermann B, Brandt P, Hiesel R. 1995. Identification of the species
origin of highly processed meat products by mitochondrial DNA sequences.
PCR Methods Appl 4:241-243.
Yokoo A. 1950. Study on chemical purification of tetrodotoxin (3)- Purification of
spheroidine. J Chem Soc Japan 71:590-592.
MS 20010106 Submitted 2/16/01, Accepted 6/29/01, Received 7/17/01
We thank Miss Chien-I Ko, Bureau of Health, Chunghua County, for help in getting the caus-
Food Chemistry and Toxicology
ative fish liver and puffer in this food poisoning case. This study was partly supported by
National Science Council (NSC-89-2313-B019-045) and Dept. of Health (DOH 90-FS003), R.O.C.
Authors Hsieh, Shiu, Hwang are with the Dept. of Food Science, National
Taiwan Ocean Univ., Keelung, Taiwan. Author Cheng is with the Dept. of
Food Engineering, National Kaohsiung Univ. of Applied Science, Kinmen
campus, Kinmen, Taiwan. Author Chen is with the Bureau of Food Sanita-
tion, Dept. of Health, Executive Yuan, Taipei, Taiwan (R.O.C). Direct corre-
spondence to author Hwang (E-mail:dfhwang@mail.ntou.edu.tw).

952 JOURNAL OF FOOD SCIENCE—Vol. 67, Nr. 3, 2002

jfsv67n3p0948-0952ms20010106-MO-SR-wcxs.p65
952 5/1/2002, 9:19 AM