Food Chemistry 75 (2001) 255–259

www.elsevier.com/locate/foodchem

Analytical, Nutritional and Clinical Methods Section

A comparison of the TLC-densitometry and HPLC method for the

determination of biogenic amines in fish and fishery products
R. Jeya Shakila*, T.S. Vasundhara, K.V. Kumudavally
Defence Food Research Laboratory, Mysore 570011, India

Received 26 February 2000; received in revised form 3 March 2001; accepted 3 March 2001

Abstract
A comparative study on the suitability of chromatographic techniques such as thin layer chromatography (TLC)-densitometry
and high performance liquid chromatography (HPLC) for the analysis of biogenic amines in fish and fishery products was carried
out. The HPLC method was found to offer a good linearity (5–100 ng), sensitivity and repeatability ( < 3%), but required sophis-
ticated instrumentation, technical skill and high operational cost and time. On the other hand, with a little loss in linearity (20–300
ng) and repeatability ( < 8%), TLC-densitometry was found to be rapid and less expensive. In addition, this method is suitable for
rapid and simultaneous screening of several samples at a time. Therefore, the TLC-densitometric method can be effectively used in
the fish industry to detect biogenic amines, especially the toxic histamine, and putrescine and cadaverine, which can potentiate
histamine toxicity in fish and fishery products. # 2001 Elsevier Science Ltd. All rights reserved.

1. Introduction has become possible due to the advancements in the

field of stationary phases for TLC. Precoated plates are
Biogenic amines are formed in foods by the bacterial commonly available and they can yield excellent
decarboxylation of free amino acids (Taylor & Sumner, separation in few minutes, and hence, a comparison was
1987). They are of importance from the point of food made between precoated TLC-densitometry and HPLC
intoxication and also as chemical indicators of spoilage. for the separation of biogenic amines in fish and fishery
They are heat stable and therefore suitable for assessing products. The factors taken into consideration are sen-
the quality of heat-processed foods (Ababouch, Alaoui, sitivity, linearity, rapidity, repeatability, operational
& Busta, 1986; Ienistea, 1971). Detection of these requirements, and efficiency.
amines in foods is often complicated because of the lack
of a rapid detection method.
Several methods have been reported for the analysis 2. Materials and methods
of histamine and other amines, which include fluori-
metric, enzymatic and chromatographic techniques. 2.1. Biogenic amine analysis
Among these, only the chromatographic techniques
have the capacity to separate the different biogenic 2.1.1. Preparation of standard amines
amines. Chromatographic techniques such as thin layer Standard putrescine, cadaverine, histamine, spermi-
chromatography (TLC) (Chin & Koehler, 1983; Flei- dine and tyramine were obtained from Sigma Chemicals
scher, 1979; Shalaby, 1994), gas liquid chromatography Co., USA. A stock standard solution was prepared by
GLC; Starusckiewicz & Bond, 1981) and high perfor- diluting accurately 0.20–0.25 g of each compound in 10
mance liquid chromatography (HPLC) (Mietz & Karmas ml of 5% trichloroacetic acid (TCA) solution. A 10-fold
1977; Yen & Hsieh, 1991) have been used for the analysis dilution of this solution formed the working standard.
of biogenic amines. In the last few years, high resolution
2.1.2. Derivatization of the amines
The amines were derivatized following the method of
* Corresponding author. Present address: Fisheries College &
Research Institute, Thoothukudi 628 008, Tamil Nadu, India. Tel.:
Rosier and Petegham (1988). One millilitre of the standard
+91-461-322354; fax: +91-461-340574. was taken in a 5-ml screw cap test tube, to which, 1.0 ml
E-mail address: jerosh99@md5.vsnl.net.in (R. Jeya Shakila). of phosphate buffer (pH 9.0, undiluted, E.Merck,
0308-8146/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0308-8146(01)00173-X
256 R. Jeya Shakila et al. / Food Chemistry 75 (2001) 255–259

India), a drop of 4 N sodium hydroxide solution and 2.0 amines to study the nature and quantities of different
ml of dansyl reagent (50 mg of dansyl chloride (Sigma amines. Fresh and canned fish (10 g) or salt-cured fish (5
grade) in 10 ml of acetone) were added. After thorough g) were homogenized with 30 ml of 5% hot (80–90 C)
mixing, tubes were covered with aluminium foil and left TCA solution for 2 min and then, centrifuged at 3000
in the incubator at 55 C for 1 h for dansylation. Tubes rpm for 10 min. Supernatant solution was filtered
were cooled and stored at 5oC until further use. through Whatman No. 41 filter paper and 1 ml of the
filtrate was used for derivatization.
2.2. TLC method

2.2.1. Fractionation, detection and quantification of 3. Results and discussion

amines
The procedure of Fleischer (1979) was followed with Extraction of amines from fish samples is an impor-
slight modification. Amines were fractionated on pre- tant step prior to the separation of biogenic amines.
coated silica gel GF254 TLC plate (0.25 mm thickness, Extraction with 5% TCA solution offered rapid extrac-
520 and 2020, E. Merck, India). Ten microlitres of tion (< 10 min) of amines. Earlier, Rosier and Peteg-
dansylated standard was applied on the plate and ham (1988) reported that an extraction time of 2 min
developed using chloroform:triethylamine (100:25) seemed sufficient to reach a plateau. Using the TCA
mixture. The plate was sprayed with iso- solution, a good recovery (98%) of the amines was also
propanol:triethanolamine (8:2) to enhance the fluores- recorded by Luten et al. (1992). As amines are highly
cence. Fractionated fluorescent amine spots were
detected under the UV light at 365 nm. Amines were
quantified by a computerized scanning densitometer
(Model CS-930, Shimadzu Corporation, Kyoto, Japan)
operated under the fluorescence mode (365 nm).

2.2.2. HPLC method

Quantitative analysis of amines was also carried out
on Waters HPLC system (Waters Associates, USA).
Separation was achieved using a 3003.9 mm, C-18 m-
Bondapak RP-column (Shandon HPLC, England) and
a UV detector (Model-440) set at 254 nm. Solvents used
were high purity HPLC grade methanol and water
(vacuum filtered and degassed prior to use). The linear
gradient elution programme described by Rosier and
Petegham (1988) was followed. Dansylated amine stan-
dard (5–10 ml) was injected on to the column, after the
equilibration with methanol/water (70/30) for 30 min.
Methanol concentration was increased gradually to
75% after 3.5 min, 80% after 7.0 min, and finally, 100%
after 10 min. A total separation time of 20 min was
required for each sample and then the column was
equilibrated. The instrument condition was adjusted to
give a full-scale response by keeping AUFS at 0.1. Data Figure 1. TLC separation of standard dansyl amines and ammonia on
were processed with a mircoprocessor-based integrator. a precoated silica gel GF 254 plate: NH3 — Ammonia; PUT —
Putrescine; Cad — Cadaverine; SPD — Spermidine; HIS — Histamine;
TYR — Tyramine.
2.3. Statistical analysis

Linear regression coefficient and coefficient of varia- Table 1

tion between the volume applied and the area print-out Rf and Rt values of the different standard biogenic amines (dansyl
were calculated following the standard procedures derivatives) separated by TLC and HPLC techniques
(Snedecor & Cochran, 1967). Amines TLC (Rf values) HPLC (Rt values)

Putrescine 0.617 10.56

2.4. Analysis of fish and fishery products for amines Cadaverine 0.724 11.24
Spermidine 0.824 16.23
Fresh, salt-cured and canned fish of a few varieties Histamine 0.870 13.56
sold in the retail markets were analyzed for biogenic Tyramine 0.911 15.54
 R. Jeya Shakila et al. / Food Chemistry 75 (2001) 255–259 257

reactive substances, dansyl chlorides are used for deri- level was determined by the application of varied con-
vatization of amines. The dansyl derivatives could be centrations of different amines on the TLC plate.
easily detected at a very low concentration under UV Detectable fluorescence spots and their response in the
light due to their fluorescent characteristics. Dansyl densitometer were observed; and the minimum detec-
chloride is also a non-specific reagent and was found to tion level was found to be in the range of 15–20 ng.
react with all the amino compounds such as amines, Fig. 2 shows the densitometric-scanning pattern of the
ammonia and free amino acids (Fleischer, 1979). standard amines and ammonia separated on the TLC
Precoated silica gel 60 GF plates were found to offer a plate. The order of separation of amines on TLC took
neat and reproducible resolution of different amines. place according to the increase in molecular weight. The
The order of separation of amines took place according distinctly resolved amine spots on the TLC plate were
to the increase in the molecular weight. Fig. 1 shows the neatly scanned in the following order of separation viz.,
TLC separation of different biogenic amines and Table 1 dansyl ammonia, putrescine, cadaverine, spermidine,
gives the Rf values. Dansyl amino acids and ammonia histamine and tyramine as clear peaks.
showed lower mobility, remained near the origin, and Biogenic amines were neatly resolved using methanol
did not interfere with the separation of amines. and water under the gradient elution mode on HPLC
Detection of amines under the long wave UV light (Fig. 3). The order of separation was different from that
showed bright coloured fluorescent spots. Histamine noticed in TLC. It was neither in increasing order of
appeared as yellowish, tyramine as green, and other molecular weight nor basic nature. Table 1 gives the
amines as greenish blue spots. The minimum detection retention time (Rt) of the individual amines. The bio-
genic amines eluted as distinct peaks at different reten-
tion times. The interfering substances like ammonia and
amino acids eluted much earlier than amines, as noticed
in TLC. Amines could be detected at 5–10 ng levels by
HPLC.
On TLC, it was possible to fractionate 10–12 samples
simultaneously on one plate at a time by inserting two
plates simultaneously and for development, 24 samples
could be fractionated in 30–40 min. In the densitometer,
time of scanning one sample was 10 min, whereas
HPLC analysis required about 30 min (including equi-
libration time) for each sample and only one sample can
be analysed at a time. Solvents required were of analar
grade for TLC analysis and about 100 ml was sufficient
for developing 24 samples. HPLC analysis required 30–

Fig. 2. TLC densitometric scanning pattern of standard dansyl amines Fig. 3. HPLC separation of the standard dansyl amines and ammonia in
and ammonia: NH3 — Ammonia; PUT — Putrescine; Cad — a gradient elution programme: NH3 — Ammonia; PUT — Putrescine;
Cadaverine; SPD — Spermidine; HIS — Histamine; TYR — Cad — Cadaverine; SPD — Spermidine; HIS — Histamine; TYR —
Tyramine. Tyramine.
258 R. Jeya Shakila et al. / Food Chemistry 75 (2001) 255–259

40 ml of high purity, particulate free special grade sol- TLC. The sensitivity (fluorescent units per ng dansy-
vent for analysis of each sample, which is quite expen- lated amine) of the amines on TLC was 4.94, 3.50, 0.84
sive. Therefore, TLC method of amine analysis was and 0.37 for putrescine, cadaverine, histamine and tyr-
rapid and relatively inexpensive compared with the amine, respectively. They were 12.30, 8.15, 9.35 and 5.80
HPLC method, which required sophisticated instru- for the HPLC method, in the same order. The linear
mentation, careful maintenance, expensive solvents, regression coefficient ranged from 0.998–0.999 for
accessories and high operational skill. HPLC and from 0.996–0.998 for TLC method, indicat-
HPLC method is extensively used for the determina- ing only marginal differences between the two methods.
tion of biogenic amines in developed countries and it is The responses for amines were different in both meth-
reported to be more efficient, sensitive and reproducible ods because of the differences in the absorption char-
compared to TLC. The linearity, sensitivity and repeat- acteristics and sensitivity. However, the quantities of
ability of these methods were statistically examined and amines determined by the two methods correlated well.
the results are given in Table 2. The response was linear To test the repeatability, five different concentrations
over a range of 5–100 ng for HPLC and 20–300 ng for of each amine were analysed at least six times separately

Table 2
Linearitya and coefficient of variation for different amines analyzed by TLC and HPLC

Amines Regression Intercept (A) Slope (B) Linear equation Coefficient of

coefficient (r) variation (n=6)

Putrescine
TLC 0.998 0.69 4.94 Y= 0.69+4.94 X 7.37
HPLC 0.998 4.27 12.30 Y=4.27+12.30 X 1.43
Cadaverine
TLC 0.999 3.46 3.50 Y= 3.46+3.50 X 6.11
HPLC 0.997 2.96 8.15 Y=2.97+8.15 X 2.06
Histamine
TLC 0.996 0.66 0.84 Y= 0.66+0.84 X 5.55
HPLC 0.999 0.27 9.35 Y= 0.27+9.35 X 2.87
Tyramine
TLC 0.996 0.22 0.37 Y= 0.22+0.37 X 7.29
HPLC 0.998 1.90 5.80 Y= 1.90+5.80 X 2.80
a
Fluorescence intensity on TLC and UV absorption intensity for HPLC with respect to increase in amine concentration. Y, Area print out by the
Detector; X, ng of dansyl amines on TLC.

Table 3
Biogenic amine profile (in mg%) of some commercially important fresh canned and salt dried fish analyzed by TLC-densitometry and HPLC
methods*

Species TLC-densitometry HPLC

Put Cad His Tyr Put Cad His Tyr

Fresh fish
Mackerel 0.95 2.80 2.07 2.58 0.82 3.23 1.95 2.74
Sardine 0.33 2.74 NDa 1.62 0.54 2.29 ND 1.18
Seerfish 1.42 1.77 ND 0.94 1.56 2.12 ND 1.07
Shrimp 1.14 ND ND 0.88 1.34 0.41 ND 1.26
Canned fish
Mackerel in brine ND 0.58 ND ND 0.21 0.67 ND ND
Sardine in oil ND 0.17 ND ND ND 0.23 ND ND
Tuna in oil ND 0.18 ND ND 0.11 0.19 ND 0.12
Salt-dried fish
Mackerel 26.44 95.59 32.10 39.84 30.18 97.62 35.26 41.38
Sardine 26.85 152.81 61.22 16.95 28.56 163.22 59.61 17.81
Seerfish 54.68 112.80 ND 15.42 56.65 124.80 ND 15.41
Shrimp 112.71 52.80 ND 70.47 110.55 55.32 ND 69.32
a
ND, not detectable.
*Mean values of three determinations.
 R. Jeya Shakila et al. / Food Chemistry 75 (2001) 255–259 259

on TLC and HPLC. The good repeatability with a Research Laboratory, Mysore for their encouragement
coefficient of variation of less than 8% was obtained for and support during the period of study. First author
TLC, as observed earlier by some workers (Fleischer, acknowledges the CSIR, New Delhi for the financial
1979; Shalaby, 1994). However, the HPLC method assistance extended for the study.
offered a very good repeatability with a coefficient of
variation of less than 3%.
References
To study the application of TLC densitometry and
HPLC methods, the biogenic amine contents of fish and Ababouch, L. (1991). Histamine food poisoning: an update. Fish Tech.
fishery products were determined separately by these New., 11, 3–5, 9.
Ababouch, L., Alaoui, M. M., & Busta, F. F. (1986). Histamine levels
methods, and the results are presented in Table 3. The
in commercial processed fish in Morocco. J. Food Prot., 49, 904–908.
quantities obtained by the two methods were more or Chakraborti, R. (1991). Histamine content in dried fish products from
less similar. The TLC densitometry method yielded Kakinada coast. Fishery Technology, 28, 59–81.
slightly lower values for amines when compared to the Chin, K. D. H., & Koehler, P. E. (1983). Identification and estimation
HPLC method. However, the differences were not very of histamine, tryptamine, phenylethylamine and tyramine in soy
sauce by thin-layer chromatography of dansyl derivatives. Journal
significant.
of Food Science, 48, 1826–1828.
The major amines detected in fish are putrescine, Fleischer, H. R. (1979). Diamines and polyamines. In J. C. Touchstone,
cadaverine, histamine and tyramine. In fresh and canned & J. Sharma (Eds.), Densitometry in thin layer chromatography:
fish, the concentrations of these amines were relatively practice and applications (pp. 241–250). New York: John Wiley.
low and in a very few samples, histamine was detected Ienistea, C. (1971). Bacterial production and destruction of histamine
in foods and food poisoning caused by histamine. Nahrung., 15,
in small amounts. The histamine concentration was far 109–113.
below the defect action level (DAL) of 10 mg% (Aba- Luten, J. B., Bouguet, W., Seuren, L. A. J., Burggraaf, M. M., Rick-
bouch, 1991) and unlikely to cause toxicity problems. wel-Booy, G., Durand, P., Etienne, M., Gouyou, J. P., Landrein,
On the other hand, salt-cured fish contained relatively A., Ritchie, A., Lecerg, M., & Guinet, R. (1992). Biogenic amines in
high amounts of biogenic amines, particularly in mack- fishery products: standardization methods within EC. In H. H. Huss,
M. Jakabsen, & J. Liston (Eds.), Quality assurance in the fish indus-
erel and sardine; histamine level was above the DAL.
try (pp. 427–439). B.V., Amsterdam, The Netherlands: Elsevier
Earlier some workers have also recorded such high Science Publishers.
levels of histamine (Chakraborti, 1991; Subburaj, Kar- Mietz, J. L., & Karmas, E. (1977). Chemical quality index of canned
unasagar, & Karunasagar, 1984). The results elicit that tuna as determined by high pressure liquid chromatography. Journal
despite a mariginal loss in linearity, repeatability and of Food Science, 42, 155–158.
Rosier, J., & Petegham, C. V. (1988). A screening method for the
sensitivity, the TLC-densitometric method was also
simultaneous determination of putrescine, cadaverine, histamine,
found to be an effective screening method for the deter- spermidine and spermine in fish by means of high pressure liquid
mination of biogenic amines in fish and fishery products chromatography of their 5-dimethyl amino napthalene 1-sulphonyl
similar to that of HPLC. derivatives. Zeit. Lebens. Unters., Forsch., 186, 25–28.
In conclusion, for routine analysis of fish and fishery Shalaby, A. R. (1994). Separation, identification and estimation of
biogenic amines in foods by thin-layer chromatography. Food
products, the TLC method using precoated silica plates
Chemistry, 49, 305–310.
coupled with densitometry can be used as a quick Snedacor, G. D., & Cochran, W. G. (1967). Statistical methods. Ames:
screening method to assess the presence of histamine and The Iowa State University Press.
other potentiating amines. This method can be easily Starusckiewicz Jr., W. F., & Bond, J. F. (1981). Gas chromatographic
adopted by regulatory agencies and food industries as a determination of cadaverine, putrescine and histamine in food. J.
Assoc. Off. Anal. Chem., 64, 584–591.
qualitative tool for the simultaneous analysis of the
Subburaj, M., Karunasagar, I., & Karunasagar, I. (1984). Incidence of
quality as well as histamine toxicity hazards of seafoods. histidine decarboxylating bacteria in fish and market environs. Food
Microbiology, 1, 263–267.
Taylor, S. L., & Sumner, S. S. (1987). Determination of histamine,
Acknowledgements putrescine and cadaverine. In D. E. Kramer, & J. Liston (Eds.),
Seafood quality determination (pp. 235–246). B.V., Amsterdam, The
Netherlands: Elsevier Science Publishers.
The authors are thankful to Dr. D. Vijaya Rao, Head, Yen, G., & Hsieh, C. (1991). Simultaneous analysis of biogenic
Department of Freeze Drying and Animal Products amines in canned fish by HPLC. Journal of Food Science, 58, 158–
Technology and Dr. S.S. Arya, Director, Defence Food 180.