Blackwell Science, LtdOxford, UK

FISFisheries Science0919-92682003 Blackwell Science Asia Pty Ltd

692April 2003
637
Thermal stability of TMAOase
M Kimura
et al.

10.1046/j.0919-9268.2002.00637.x
Original Article414420BEES SGML

FISHERIES SCIENCE 2003; 69: 414–420

TMAOase, trimethylamine-N-oxide demethylase, is a

thermostable and active enzyme at 80∞C
Meiko KIMURA,1 Ikuo KIMURA2 AND Nobuo SEKI1*

1
Laboratory of Food Biochemistry, Graduate School of Fisheries Sciences, Hokkaido University,
Hakodate, Hokkaido 041-8611 and 2Central Research Laboratory, Nippon Suisan Kaisha Ltd,
Hachioji, Tokyo 192-0906, Japan

ABSTRACT: Purified trimethylamine-N-oxide demethylase (TMAOase) from walleye pollack muscle

is a thermostable protein that was not inactivated after heating at 80∞C for 30 min. The heated enzyme
was electrophoresed in the same manner as for native enzyme. Circular dichroism (CD) spectra for
purified enzyme changed reversibly in the temperature range of 10–80∞C. As the enzyme was still
active at 80∞C, the CD spectral change did not directly relate to enzyme activity. TMAOase activity
in the myofibrillar fraction decreased sharply above 30∞C, but was extracted and recovered from the
heated myofibrillar fraction, suggesting that the activity seemed to be interrupted and apparently
inactivated due to the thermal alteration of myofibrillar proteins or some unknown factors. The
complicated profile found in dimethylamine (DMA) formation from trimethylamine-N-oxide (TMAO) in
walleye pollack muscle during heating consisted of both enzymic and non-enzymic processes. Most
DMA was produced enzymatically below 40∞C and interrupted above 40∞C. Therefore, DMA and
trimethylamine was formed non-enzymatically at high temperatures regardless of the presence of
native enzyme. A new, simple and easy purification method was proposed based on the thermostable
nature of the enzyme.

KEY WORDS: dimethylamine, formaldehyde, thermal stability, trimethylamine,

trimethylamine-N-oxide (TMAO), trimethylamine-N-oxide demethylase (TMAOase), walleye
pollack.

INTRODUCTION TMAOase activity in walleye pollack muscle

Trimethylamine-N-oxide (TMAO), a natural
osmolyte, is present in most marine species, and
especially accumulated in the various tissues of
marine elasmobranchs, gadoid fishes, a few non-
gadoid fish and molluscs.1–4 Walleye pollack is a
commercially important fish and contains 60–
100 mmol/kg TMAO in the muscle based on our
estimation. It was demethylated to equimolar
formaldehyde (FA) and dimethylamine (DMA) by
catalyzing with an endogenous enzyme, TMAOase
(EC 4.1.2.32), during storage from 0∞C to room
temperature. During frozen storage, however,
TMAO is degraded to FA and DMA or to trimethy-
lamine (TMA) through non-enzymic pathways.5
The FA produced has been suggested to cause a
decrease in solubility and extractability of the myo-
fibrillar proteins and to result in eventual detri-
mental changes in texture and functional
properties, including viscosity, emulsifying ability
and gel-forming ability.
*Corresponding author: Tel: 81-138-40-5568.
Fax: 81-138-40-5568. Email: seki@fish.hokudai.ac.jp
Received 6 August 2002. Accepted 28 October 2002.
homogenate was lost when it was heated at 50∞C
for 30 min.6 A crude enzyme solution was also
inactivated upon heating at 60∞C for 10 min.7 Lall
et al. reported that TMAOase in minced silver hake
muscle was completely inactivated by the heating
at 80∞C.8 Castell et al.9 reported an interesting
result that a large amount of DMA was detected in
the fillets of hake that were cooked at 110∞C for
10 min and then stored in a freezer. They explained
the result that DMA was non-enzymatically pro-
duced from TMAO because of thermal denatur-
ation of TMAOase. We found that TMAOase activity
in walleye pollack myofibrillar fraction decreased
remarkably above 30∞C.10 Based on the many
experimental results, it has been widely accepted
that TMAOase is thermally unstable, like most
enzymes.
We recently purified and isolated TMAOase
from walleye pollack myofibrillar fraction by the
methods of acid extraction, chromatographies and
a naive polyacrylamide gel electrophoreses as
reported elsewhere.11 The enzyme having a molec-
ular weight of 25 000 is an acidic protein and
requires Fe2+ and reducing agents for full activation
at optimum pH 7.0. Concerning the thermal stabil-
Thermal stability of TMAOase FISHERIES SCIENCE
ity of the isolated enzyme, we have obtained an
unexpected result that the purified enzyme is
extremely stable at 80∞C as reported in the present
paper. The result is, therefore, in conflict with the
accepted theory. Thus, we investigated the cause of
the discrepancy between thermal stabilities of the
enzyme in the muscle and myofibril fractions, and
the purified enzyme. Furthermore, the enzymic
and non-enzymic degradations of TMAO were
investigated in order to assess the possible path-
ways at high temperatures.
MATERIALS AND METHODS
Materials
Walleye pollack Theragra chalcogramma that had
been caught off the northern Hakodate coast was
filleted, frozen and stored at - 60∞C. White muscle
fillet was used throughout the experiment. TMAO
and Stains-all (1-ethyl-2-[3-(1-ethylnaphtho[1,2-d]
thiazolin-2-ylidene)-2-methylpropenyl]naphtho-
[1,2-d]thiazolium bromide) were purchased from
Sigma-Aldrich Japan (Tokyo, Japan). Sephacryl
S-300 and molecular weight calibration kits for gel
filtration and sodium dodecylsulfate-polyacryla-
mide gel electrophoresis (SDS-PAGE) were pur-
chased from Amersham Pharmacia (Uppsala,
Sweden). Ultrafilter (UP-20) was obtained from
Advantec (Tokyo, Japan). DE52 was obtained from
Whatman (Maidstone, UK). DMA, FA and other
chemicals were of reagent grade and purchased
from Wako Pure Chemical Industries (Osaka,
Japan).
Preparation of myofibrillar fraction
The thawed walleye pollack fillet was immediately
minced and then homogenized in 5 volumes of
0.1 M NaCl-20 mM Tris-acetate (pH 7.0) with a
homogenizer (Nippon Seiki, Tokyo, Japan) and
passed through a layer of gauze. The homogenate
was centrifuged at 3000 ¥g for 10 min. The precip-
itate was washed three times with the same buffer.
The last centrifugation was performed at 8000 ¥g
for 20 min. The precipitate was used as a myofibril-
lar fraction. Most TMAOase activity was recovered
into the myofibrillar fraction.10
Preparation of purified TMAOase
TMAOase was purified from the myofibrillar frac-
tion by the methods of Kimura et al.11 Briefly,
415
enzyme activity was extracted by means of acid
treatment in which the pH of homogenized and
solubilized myofibrils in 1 M NaCl was adjusted to
4.5 by the addition of 1 M HCl. The supernatant
after centrifugation was neutralized and dialyzed
against 20 mM Tris-acetate (pH 7.0). The enzyme
in the dialytic tube was further purified by suc-
cessive chromatographies on DE52 and Sephacryl
S-300. Finally, the enzyme was purified upon
native PAGE using 7.5% polyacrylamide disk
gels.12
Assay of TMAOase activity
TMAOase activity was measured in terms of deter-
mination of DMA produced. A standard reaction
mixture contained 20 mM TMAO, 2 mM cysteine,
0.2 mM FeCl2 and 20 mM Tris-acetate (pH 7.0) with
an enzyme preparation. In some cases, 2 mM
ascorbate was added to the reaction mixture. The
reaction was started by the addition of enzyme at
25∞C and terminated by the addition of 2.5%
trichloroacetic acid. After centrifugation at 8000 ¥g
for 15 min, the amount of DMA in the supernatant
was determined. The unit of the activity was
defined as the amount of mM DMA produced by
incubation with an enzyme for 1 min. Non-
enzymic decomposition of TMAO was measured in
the same system without enzyme and terminated
by the addition of 2.5% trichloroacetic acid.
TMAOase activity in the myofibrillar fraction
was measured as follows. The assay mixture of
3 g contained 20 mM TMAO, 2 mM cysteine, 2 mM
ascorbate, 0.2 mM FeCl2, 0.1 M NaCl and 20 mM
Tris-acetate (pH 7.0) with 50–80 mg myofibrillar
proteins at a final concentration. After the incuba-
tion at 25∞C for up to 3 h, the mixture was quanti-
tatively transferred to a centrifuge tube containing
1 mL of 10% trichloroacetic acid to terminate
the reaction and then H2O (4 mL) was added to
the tube prior to centrifugation at 8000 ¥g for
15 min. After centrifuging, an aliquot of the termi-
nated reaction mixture was subjected to DMA
determination.
Determination of TMAO, DMA and TMA
An aliquot of the terminated reaction mixture
was subjected to DMA determination by copper-
dithiocarbamate procedure.13 TMA was deter-
mined by a picrate method.14 TMAO was
determined by reduction with Devarda’s alloy,
followed by estimation of produced TMA by the
colorimetric procedure of Dyer et al.14,15
416 FISHERIES SCIENCE
Circular dichroism measurements
Circular dichroism (CD) spectra of TMAOase were
recorded at various temperatures in a Jasco J-725
spectropolarimeter (Nippon Bunkou, Tokyo,
Japan) using a 1.0-mm cuvette at a protein concen-
tration of 3.3 mg/mL in 20 mM Tris-acetate (pH 7.0)
and were corrected for contributions from the
buffer solution.
Protein determination
Protein concentration was determined by the
biuret method using bovine serum albumin (BSA)
as standard.16 In the case of purified TMAOase, a
Micro BCA protein assay kit (Pierce, Rockford, IL,
USA) was used. A variation in color response of the
reagent between TMAOase and BSA is not yet
revised. TMAOase solution containing 1.0 mg/mL
protein measured by Micro BCA assay showed
0.312 absorbance at 230 nm.
RESULTS AND DISCUSSION
Thermal stability of purified enzyme
The purified TMAOase solution was heated at 80∞C
for up to 30 min, and then cooled in an ice bath for
5 min. The solution was still clear and colorless
after heating. The remaining activity was measured
at 25∞C (Fig. 1a). TMAOase activity remained com-
pletely after heating. It was also found that the
Fig. 1 Thermal stability of trimethylamine-N-oxide
demethylase (TMAOase). (a) TMAOase (2 mg/mL) was
heated in 20 mM Tris-acetate (pH 7.0) at 80∞C for up to
30 min and then cooled in an ice bath. The remaining
activity was measured at 25∞C in a standard reaction
mixture. (b) Polyacrylamide gel electrophoresis of native
enzyme (0) and heated enzyme at 80∞C for 30 min (30).
The gels were double stained with Coomassie Brilliant
Blue and Stains-all.
M Kimura et al.
heated enzyme was electrophoresed in the same
manner as that of the native enzyme upon PAGE
(Fig. 1b).
Circular dichroism spectra
Figure 2 shows the CD spectra of TMAOase at var-
ious temperatures in the range of 10–80∞C. The CD
spectra of TMAOase exhibited a minimum at
197 nm and a small negative shoulder between 215
and 225 nm, which was more pronounced at
higher temperatures. The shape of the spectra was
similar to that expected for a rather structureless
conformation. However, the pronounced negative
shoulder in the CD spectrum at 80∞C suggested the
presence of a somewhat ordered structure. After
scanning the spectrum at 80∞C, the sample was
rescanned at 50, 25 and 10∞C. The shape of the
spectrum was superimposable on that at each
scanning temperature. The structural changes
caused by heating were completely reversible.
Therefore, the thermal stability of the enzymic
activity may depend on the conformational revers-
ibility. A question, however, arose as to whether the
enzyme is still active at high temperatures inde-
pendent of the conformational alteration. We
decided to detect the presence of active enzyme at
high temperatures.
Non-enzymic degradation of TMAO
It is well known that TMAO is degraded non-enzy-
matically above 55–60∞C and the thermal degrada-
Fig. 2 Circular dichroism (CD) spectra of
trimethylamine-N-oxide demethylase (TMAOase) from
walleye pollack at various temperatures. CD spectra of
the enzyme (3.3 mg/mL) in 20 mM Tris-acetate (pH 7.0)
were measured at various temperatures.
Thermal stability of TMAOase FISHERIES SCIENCE 417

Fig. 3 Non-enzymic formation

of dimethylamine (DMA) and tri-
methylamine (TMA) from trime-
thylamine-N-oxide (TMAO) at
high temperatures. The forma-
tion of DMA and TMA was mea-
sured in 20 mM TMAO, 0.2 mM
FeCl2, 2 mM cysteine and 10 mM
Tris-acetate (pH 7.0) without
enzyme at various temperatures.

, 40∞C; , 50∞C; , 60∞C;
, 70∞C;
, 80∞C. DMA,
closed symbols; TMA, open
symbols.

tion products are not only DMA and FA, but also
TMA.17,18 To confirm the non-enzymic degradation
of TMAO in our reaction system in the temperature
range of 40–80∞C, the formation of DMA and TMA
was determined as a function of incubation time.
Figure 3 shows that the production of TMA is faster
and greater than that of DMA at an initial reaction
stage at all temperatures. DMA increase showed a
sigmoid curve and was above the later amount of
TMA, suggesting the existence of a TMA demethy-
lating pathway in addition to TMAO reducing and
demethylating routes.

Thermal stability of TMAOase activity

The time course of the formation of DMA from

TMAO in the presence of various amounts of
TMAOase with cofactors (Fe2+ and cysteine) was
measured at 80∞C (Fig. 4). The amount of DMA
increased with an increase in the enzyme concen-
tration and reached a plateau after 15 min incuba-
tion, regardless of the enzyme concentration due
to thermal inactivation or the exhaustion of cofac-
, 0 unit/mL; , 0.14 unit/mL; ,
tors that may be caused by the oxidation of Fe2+
and cysteine during incubation at 80∞C. The non-
enzymic formation of DMA and TMA in the
absence of enzyme was also plotted in this figure
and reached a plateau similar to enzyme reaction.
The formation of TMA was independent of
TMAOase concentration. These results suggested
that the decrease in both reactions was caused by
cofactor exhaustion.
In order to clarify the thermostability and activ-
ity of TMAOase at 80∞C, a solution containing var-
ious amounts of TMAOase was preheated in the
absence of substrate TMAO and cofactors at 80∞C
for up to 30 min and then the substrate and cofac-
tors were added to the solution at the same tem-
perature. After 5 min incubation, the reaction was
terminated with 2.5% trichloroacetic acid. The
Fig. 4 Time course of the formation of dimethylamine
(DMA) and trimethylamine (TMA) in the presence of tri-
methylamine-N-oxide demethylase (TMAOase) at 80∞C.
Assay was carried out at 80∞C in a reaction mixture con-
taining 20 mM TMAO, 0.2 mM FeCl2, 2 mM cysteine and
10 mM Tris-acetate (pH 7.0) with various amounts of
TMAOase as follows:
0.35 unit/mL;
, 0.70 unit/mL. DMA, closed symbols;
TMA, open symbols.
DMA and TMA produced were determined (Fig. 5).
The amount of DMA produced maintained a con-
stant value independent of preheating time of the
enzyme solution at 80∞C, suggesting that the
enzyme was stable and active during preheating at
80∞C for at least 30 min (Fig. 5a). Furthermore,
DMA was linearly increased depending on the
enzyme concentration, while TMA was only
formed non-enzymatically (Fig. 5b).
As it is concluded that TMAOase is active at
80∞C, there is no direct relationship between the
enzyme activity and the reversible conformational
change detected by CD measurement (Fig. 2).
418 FISHERIES SCIENCE M Kimura et al.

Fig. 5 Thermal stability and

activity of trimethylamine-N-
oxide demethylase (TMAOase) at
80∞C. TMAOase in 20 mM Tris-
acetate (pH 7.0) was preheated at
80∞C for various times and then a
substrate solution containing
20 mM TMAO, 0.2 mM FeCl2,
2 mM cysteine and 10 mM Tris-
acetate (pH 7.0) was added and
held at 80∞C for 5 min. The reac-
tion was terminated by the addi-
tion of 2.5% trichloroacetic acid.
(a) Effect of preheating time on
the formation of dimethylamine
(DMA, closed symbols) and trim-
ethylamine (TMA, open sym-
bols). Enzyme concentrations:
, 0 unit/mL; , 0.14 unit/mL; , 0.35 unit/mL;
, 0.70 unit/mL (b) Effect of enzyme
concentration on the formation of , DMA and , TMA .

Thermal stability of TMAOase in

myofibrillar fraction

In a previous report,10 we found that TMAOase

activity in the myofibrillar fraction decreased
above 30∞C and was almost lost above 55∞C after
the fraction was heated at various temperatures for
30 min. Although this result is shown in Fig. 6, it
conflicts with the thermostable nature of purified
TMAOase. To clarify the cause of the incompatibil-
ity, we compared the yield of TMAOase activity
from native and heated myofibrillar fractions by
means of acid treatment (Table 1). Total activity
was reduced from 150 units to 35 units by the heat-
ing procedure at 80∞C for 30 min. Both the enzyme
activities in the acid extracts from native and
heated myofibrillar fractions showed similar val-
ues, 78 and 85 units, respectively. Thus, the yield of Fig. 6 Thermal stability of trimethylamine-N-oxide
total activity from the heated myofibrillar fraction demethylase (TMAOase) activity in myofibrillar fraction.
increased by 243%. The results clearly showed that Myofibrillar fraction (75 mg/mL) was preheated for
the heated myofibrillar fraction contained native 30 min at various temperatures prior to the measure-
enzyme. We supposed, therefore, that TMAOase ment of TMAOase activity.
activity was interrupted and apparently inacti-
vated by the thermal alteration of myofibrillar pro-
teins or some factors in the myofibrillar fraction. DMA re-increased greatly from 55∞C to 70∞C, while
TMA increased steadily from 40∞C to 80∞C. The
ratio of DMA to TMA produced in the temperature
Degradation of TMAO in walleye pollack muscle range of 40–80∞C was very similar to that obtained
in non-enzymic formation from TMAO in a model
The results mentioned above may be a clue to system upon heating for 20 min (Fig. 3). Further-
understanding a complicated DMA formation pro- more, TMAOase cannot catalyze the reaction of
file in walleye pollack muscle during heating as TMA formation. We, therefore, concluded that
shown in Fig. 7. The walleye pollack muscle used DMA formed almost enzymatically below 40∞C and
here contained 60 mmol/kg TMAO and was heated mainly non-enzymatically above 40∞C because the
at various temperatures for 20 min. The amounts enzyme activity was interrupted. There is little
of DMA and TMA produced were measured. DMA direct evidence, however, primarily because the
increased until 40∞C and then decreased at 55∞C. localization of TMAOase and the amounts and
Thermal stability of TMAOase FISHERIES SCIENCE 419

Table 1 Yield of trimethylamine-N-oxide demethylase (TMAOase) activity from native and heated myofibrillar
fractions
Total protein Total activity Yield
Sample* (mg) (unit) (%)
Native myofibrillar fraction 1110 150 100
Extract from native myofibrillar fraction 6.6 78 52
Heated myofibrillar fraction 1110 35 100
Extract from heated myofibrillar fraction 38 85 243

* Myofibrillar fraction was prepared from walleye pollack muscle (15 g) and heated at 80°C for 30 mm. TMAOase activity was
extracted from native and heated myofibrillar fractions by stirring at pH 4.5 in 1 M NaCl. The extract was neutralized and dialyzed
against 20 mM Tris-acetate (pH 7.0).

Fig. 7 Formation of dimethylamine (DMA) and trime-

thylamine (TMA) in walleye pollack muscle at various
temperatures. Minced walleye pollack muscle was
heated for 20 min at various temperatures and then the
reaction was terminated by the addition of 10% trichlo-
roacetic acid in an ice bath. The DMA and TMA in
minced muscle was extracted overnight and determined.
, DMA;
, TMA.
types of reductants in muscle tissue are not well
established.
An improved purification method
As TMAOase was a thermostable protein and an
extractable enzyme from the heated myofibrillar
fraction (Table 1), we attempted to improve a puri-
fication method including the heating procedure.
Walleye pollack muscle was homogenized in 1 M
NaCl and this was followed by the addition of 1.0 M
HCl to adjust the pH to 4.5 while stirring for 30 min.
After the pH was readjusted to 4.5, the homogenate
was centrifuged at 100 000 ¥g for 30 min. The
Fig. 8 A simple method for the purification of trimeth-
ylamine-N-oxide demethylase (TMAOase) from muscle.
The inserted figure shows a native polyacrylamide gel
electrophoresis (PAGE) of the sample concentrated on
an ultrafilter. The gel columns were stained with (a) Coo-
massie Brilliant Blue and (b) Stains-all. The band shown
by an arrow was cut off and eluted with 20 mM Tris-
acetate (pH 7.0) at 0∞C. The eluate having TMAOase
activity was re-electrophoresed and the gels were (c)
double stained with Coomassie Brilliant Blue and
Stains-all.
supernatant obtained was heated at 80∞C for
20 min and then centrifuged. The supernatant was
neutralized and dialyzed against 20 mM Tris-
acetate (pH 7.0) overnight. The dialyzed enzyme
was concentrated by ultrafiltration using an ultra-
420 FISHERIES SCIENCE
filter (MW cut-off 20 000) and further purified
upon native PAGE. Figure 8 shows the native PAGE
of the dialyzed sample that was stained with
Coomassie Brilliant Blue (Fig. 8a) and Stains-all
(Fig. 8b). A major band stained with Stains-all was
cut off and eluted with 20 mM Tris-acetate (pH 7.0)
overnight. The eluted sample was re-electrophore-
sed. The final native PAGE showed a single band
that was double stained with Coomassie Brilliant
Blue and Stains-all (Fig. 8c). Eluted protein from
the band was active and the yield of total activity
increased approximately 10-fold as compared with
the conventional method containing laborious
chromatographies.11
REFERENCES
1. Dyer WJ. Amines in fish muscle. VI. Trimethylamine oxide
content of fish and marine invertebrates. J. Fish. Res. Board
Can. 1952; 8: 314–324.
2. Amano K, Yamada K. Biological formation of formaldehyde
in the muscle tissue of gadoid fish. Nippon Suisan Gakkai-
shi 1964; 30: 430–435.
3. Tokunaga T. Trimethylamine oxide and its decomposition
in the bloody muscle of fish. I. TMAO, TMA, and DMA con-
tents in ordinary and bloody muscles. Nippon Suisan Gak-
kaishi 1970; 36: 502–509.
4. Harada K. Studies on enzyme catalyzing the formation of
formaldehyde and dimethylamine in tissues of fish and
shells. J. Shimonoseki Univ. Fish. 1975; 23: 163–241.
5. Kimura M, Seki N, Kimura I. Enzymic and non-enzymic
cleavage of trimethylamine-N-oxide in vitro at subzero
temperatures. Nippon Suisan Gakkaishi 2002; 68: 85–91.
6. Tokunaga T. Studies on the development of dimethylamine
and formaldehyde in Alaska pollack muscle during frozen
storage. Bull. Hokkaido Res. Fish. Res. Lab. 1964; 29: 108–
122.
M Kimura et al.
7. Phillipy BQ, Hultin HO. Distribution and some characteris-
tics of trimethylamine-N-oxide (TMAO) demethylase activ-
ity of red hake muscle. J. Food Biochem. 1993; 17: 235–250.
8. Lall BS, Manzer AR, Hiltz DF. Preheat treatment for
improvement of frozen storage stability at -10∞C in fillets
and minced flesh of silver hake. J. Fish. Res. Board Can.
1975; 32: 1450–1454.
9. Castell CH, Smith B, Neal W. Production of dimethylamine
in muscle of several species of gadoid fish during frozen
storage, especially in relation to presence of dark muscle. J.
Fish. Res. Board Can. 1971; 28: 1–5.
10. Kimura M, Seki N, Kimura I. Occurrence and some proper-
ties of trimethylamine-N-oxide demethylase in myofibrillar
fraction from walleye pollack muscle. Fish. Sci. 2000; 66:
725–729.
11. Kimura M, Seki N, Kimura I. Purification and characteriza-
tion of trimethylamine-N-oxide demethylase from walleye
pollack muscle. Fish. Sci. 2000; 66: 967–973.
12. Weber K, Osborn M. The reliability of molecular weight
determinations by dodecyl sulfate-polyacrylamide gel elec-
trophoresis. J. Biol. Chem. 1969; 244: 4406–4412.
13. Dyer WJ, Mounsey YA. Amines in fish muscle. II. Develop-
ment of trimethylamine and other amines. J. Fish. Res.
Board Can. 1945; 6: 359–367.
14. Dyer WJ. Amines in fish muscle. I. Colorimetric determina-
tion of trimethylamine as the picrate salt. J. Fish. Res. Board
Can. 1945; 6: 351–358.
15. Dyer WJ, Dyer FE, Snow JM. Amines in fish muscle. V. Trim-
ethylamine oxide estimation. J. Fish. Res. Board Can. 1952;
8: 309–313.
16. Gornell AG, Bardawill CJ, David MM. Determination of
serum proteins by means of biuret reaction. J. Biol. Chem.
1949; 177: 751–766.
17. Tokunaga T. Studies on the quality evaluation of canned
marine products. I. Determination of the ratio of dimethy-
lamine to trimethylamine in canned albacore. Nippon
Suisan Gakkaishi 1975; 41: 547–553.
18. Spinelli J, Koury B. Non-enzymic formation of dimethy-
lamine in dried fishery products. J. Agric. Food Chem. 1979;
27: 1104–1108.