Journal of Chromatography A, 815 (1998) 133–140

Determination of zearalenone in corn by means of immunoaffinity

clean-up and high-performance liquid chromatography with
fluorescence detection
Angelo Visconti*, Michelangelo Pascale
Istituto Tossine e Micotossine da Parassiti Vegetali, C.N.R., Viale L. Einaudi 51, 70125 Bari, Italy

Abstract

A rapid and accurate method to quantify zearalenone in corn is described. The method uses immunoaffinity
chromatography for purification and high-performance liquid chromatography (HPLC) for detection and quantification of the
toxin. Corn samples were extracted with acetonitrile–water (90:10, v / v) and the extract was diluted with water (1:10, v / v)
and applied to a Vicam ZearalaTest immunoaffinity column. The column was washed with water and zearalenone was eluted
with methanol and quantified by reversed-phase HPLC with fluorometric detection ( le x 5274 nm, le m 5440 nm) using
acetonitrile–water–methanol (46:46:8, v / v) as mobile phase. Zearalenone recoveries from the ZearalaTest column were
higher than 95%, and the column can hold a maximum of 4.0 mg of toxin. Average recoveries of zearalenone from corn
spiked at levels of 0.1–10 mg / g ranged from 93 to 99.5%, with relative standard deviations of ,6%. The detection limit was
3 ng / g based on a signal-to-noise ratio of 3:1. Comparative analysis of 14 naturally contaminated samples using this method
and the AOAC official method 985.18 showed a reasonable correlation (r50.87). Advantages of the immunoaffinity method
as compared to the AOAC method are discussed.  1998 Elsevier Science B.V. All rights reserved.

Keywords: Fusarium spp.; Immunoaffinity chromatography; Sample preparation; Food analysis; Zearalenone; Mycotoxins;
Toxins

1. Introduction strong estrogenic and anabolic compound that causes

reproductive problems in farm animals, especially
Zearalenone h6-[(10S)-10-hydroxy-6-oxo-trans-1- swine. Symptoms may include vaginal swelling
undecenyl]-b-resorcylic acid lactonej is one of the (vulvovaginitis) and, in severe cases, vaginal and
most widely distributed mycotoxins produced by rectal prolapses, especially in immature gilts (swine).
Fusarium species, in particular F. graminearum, F. In young male swine it may cause testicular atrophy
culmorum and F. crookwellense, responsible for and enlargement of the mammary glands.
important diseases of various cereals. It is associated Zearalenone ingestion may also cause prolonged
mainly with corn but it occurs also in wheat, barley estrus and reduced sex drive, infertility, fetal mum-
and sorghum among other commodities frequently mification, abortion, stillbirths and reduced litter size
used in human and animal diets [1]. Zearalenone is a [2]. Serious impact on profit returns may be regis-
tered by swine producers and the hog industry as a
*Corresponding author. consequence of reproductive failure related to the

0021-9673 / 98 / $19.00  1998 Elsevier Science B.V. All rights reserved.

PII: S0021-9673( 98 )00296-9
134 A. Visconti, M. Pascale / J. Chromatogr. A 815 (1998) 133 – 140

presence of zearalenone in corn [3]. Although further 2. Experimental

studies are required to confirm whether zearalenone
should be considered a potential human carcinogen, 2.1. Chemicals and materials
positive evidence of carcinogenicity in experimental
animals has been shown [1]. Zearalenone stock solution was prepared by dis-
Presently six countries, namely Austria, Brazil, solving in methanol (1 mg / ml) the solid standard
France, Romania, Russia and Uruguay, have specific purchased from Sigma (St. Louis, MO, USA).
regulations for zearalenone in foods or feeds, with Zearalenone standard solutions for HPLC calibration
maximum tolerated levels ranging from 30 to 1000 or spiking purposes were prepared by dissolving in
ng / g [4]. the mobile phase or methanol, respectively, adequate
Current analytical methods for zearalenone include amounts of the stock solution, previously evaporated
thin-layer chromatography [5–7], gas–liquid chro- to dryness under nitrogen stream. Acetonitrile,
matography [5,8], gas chromatography–mass spec- methanol and water (HPLC grade) and potassium
trometry [8,9], and HPLC [5,10–14]. Although chloride (ACS) were purchased from Mallinckrodt
several of these methods achieve low levels of Baker (Milan, Italy).
detection, in practice they consume large amounts of ZearalaTest immunoaffinity columns were ob-
time and solvent, and require one or more clean-up tained from Vicam L.P. (Watertown, MA, USA).
steps involving liquid–liquid partition or solid-phase Glass microfibre filters (Whatman GF /A) and paper
extraction. Immunological methods, including line filters (Whatman No. 1) were obtained from What-
immunoblot [15], dipstick immunoassay [16] and man (Maidstone, UK).
enzyme-linked immunosorbent assay (ELISA) The blank corn material used for recovery experi-
[17,18], have been used for zearalenone determi- ments was obtained from a retail store in Bari (Italy).
nation in combination with other mycotoxins. The Naturally contaminated corn samples, collected in
ELISA method has also been tested collaboratively northern Italy, were kindly provided by Dr. P. Sarzi
and approved by the AOAC as first action screening Amade´ (Safe Food, Viadana, Italy).
method for zearalenone concentrations higher than
800 ng / g [17]. 2.2. Apparatus
During the past few years, a strong impulse to the
improvement of mycotoxin analysis has been given The HPLC apparatus consisted of a LKB 2150
by the introduction of an important clean-up tech- isocratic pump (LKB, Bromma, Sweden) equipped
nique using antibody-based immunoaffinity columns with a Rheodyne Model 7125 injection valve (Rheo-
(IAC), which provides clean extracts and minimizes dyne, Cotati, CA, USA), a Perkin-Elmer LC 240
the interference of co-extracted compounds from fluorometric detector and a TURBOCHROM 4.0 data
complex matrices, such as cereals or cereal-based system (Perkin-Elmer, Norwalk, CT, USA). The
foods and feeds in which mycotoxins generally analytical column was a C 1 8 reversed-phase Supel-
occur. IACs are available commercially for a number cosil LC-18 (15 cm34.6 mm, 5 mm particles)
of mycotoxins, including aflatoxins, ochratoxin A, (Supelco, Bellefonte, PA, USA) preceded by a
fumonisins, deoxynivalenol and zearalenone [19]. Rheodyne guard filter (0.5 mm).
In this paper we describe the performances of a
method for the determination of zearalenone in corn 2.3. Sample preparation and immunoaffinity clean-
using HPLC with fluorescence detection coupled up
with a clean-up step using commercially available
IAC, the Vicam ZearalaTest. The application of the About 1 kg of corn sample was finely ground by a
method to the analysis of several naturally contami- Model 2A Romer mill (Romer, Union, MO, USA)
nated corn samples as compared to the liquid chro- and homogenized. Subsamples (about 100 g) were
matographic AOAC official method is also de- taken and stored at 2208C until analysis. Ground
scribed. sample (20 g) were weighed into a blender jar, 2 g
 A. Visconti, M. Pascale / J. Chromatogr. A 815 (1998) 133 – 140 135

KCl added and extracted with 50 ml acetonitrile– 236 nm and 418 nm cut-off filter, respectively (see
water (90:10, v / v) by blending at high speed for 2 Section 3), and the mobile phase flow-rate was 1.3
min (Sorvall Omnimixer). The extract was filtered ml / min (instead of 2 ml / min to limit high instru-
through filter paper and 10 ml filtrate was collected ment back-pressure).
and mixed with 90 ml distilled water. The diluted
extract was filtered through Whatman GF /A glass 2.6. Column capacity
microfibre filter and the filtrate collected. A 10-ml
volume of diluted extract (equivalent to 0.4 g The capacity of the ZearalaTest columns was
sample) was passed through the ZearalaTest im- determined by comparing (duplicate measurements)
munoaffinity column at a flow-rate of about 1 drop / the amount of zearalenone added to the immuno-
s, followed by 235 ml distilled water at 1–2 drops / s affinity column with the amount bound. Different
flow-rate. Zearalenone was then eluted with 1.5 ml amounts of zearalenone, from 0.8 to 8.0 mg, were
methanol and collected in a clean vial. The eluted added to the immunoaffinity column by loading 10
extract was evaporated under nitrogen stream at ml (equivalent to 0.4 g matrix) of blank corn extract
|508C and reconstituted with 250 ml of the HPLC spiked with the corresponding amount of
mobile phase. zearalenone.

2.4. HPLC determination and confirmation of 2.7. Recovery experiments

zearalenone
Recovery experiments were performed in quad-
One hundred ml of reconstituted extract (equiva- ruplicate by spiking blank corn samples with
lent to 0.16 g of sample) was injected into the zearalenone at levels of 0.1, 0.5, 1.0, 2.0, 4.0 and
chromatographic apparatus by full loop injection 10.0 mg / g. Spiked samples were left for 1 h, to
system. The mobile phase consisted of a mixture of allow solvent evaporation prior to extraction with
acetonitrile–water–methanol (46:46:8, v / v) eluted at acetonitrile–water.
a flow-rate of 1.0 ml / min. Quantification of
zearalenone was performed by measuring peak areas
at zearalenone retention time, and comparing them 3. Results and discussion
with the relevant calibration curve.
The identity of zearalenone was confirmed in all 3.1. Choice of the excitation and emission
positive samples by injecting sequentially sample wavelengths
extracts using 274 nm and 236 nm excitation wave-
lengths (440 nm emission wavelength) and compar- The UV spectrum of zearalenone in methanol
ing the peak area ratio (236 nm / 274 nm) with that of shows major absorption maxima at 236 nm (e 5
zearalenone standard. 29 200 l mol 21 cm 21 ) and 274 nm (e 513 040
l mol 21 cm 21 ), respectively, whereas the maximum
2.5. Comparison with the AOAC official method of fluorescence emission occurs at 460 nm. Most
publications use 236 nm or 274 nm as excitation
In order to have a comparison with a well wavelength and 418 nm cut-off filter or 440 nm
established method for the determination of emission for fluorescence detection of zearalenone
zearalenone in corn, the naturally contaminated [5,10,11,13,14,20]. Six excitation / emission wave-
samples were analyzed by both the immunoaffinity length combinations (236 and 274 nm excitation and
method described herein and the AOAC official 418, 440 and 460 nm emission) were tested herein to
method 985.18 [20]. This method was applied ac- assess the optimum conditions for zearalenone de-
cording to the published protocol [20], with minor tection. The emission wavelength at 418 nm was
differences, i.e. the fluorescence detector was set at rejected because it generates fluorescence signals of
274 nm (excitation) and 440 nm (emission) instead of low intensity. The highest fluorescence signals were
136 A. Visconti, M. Pascale / J. Chromatogr. A 815 (1998) 133 – 140

observed when using excitation at 236 nm and 4.0 mg of zearalenone. Above this level no increase
emission at 440 nm or 460 nm, but with an undesir- of the fluorescence response was observed, indicat-
able increase of the noise intensity that did not allow ing the saturation of zearalenone binding sites (Fig.
detection of zearalenone at levels ,1 ng (absolute 1). Zearalenone recoveries from the column below
injected amount). The best results in terms of signal- the saturation level were higher than 95%.
to-noise ratio were obtained with the 274 nm To evaluate the immunoaffinity column perform-
excitation / 440 nm emission wavelength combina- ances with respect to matrix interferences, different
tion; under these experimental conditions volumes of diluted blank corn extract, spiked with 1
zearalenone could be detected at levels as low as mg / g zearalenone, were loaded onto the column
0.33 ng (signal-to-noise, 3:1). This wavelength (triplicate measurements). Up to 30 ml of corn
combination was then used in the present study for extract (corresponding to 1.2 g of corn material)
zearalenone determination with both the immuno- were used. Zearalenone recoveries decreased from
affinity method and the AOAC official method. 95 to 80% by increasing the loading volume from
2.5 to 15 ml, although a good precision (R.S.D.,
3.2. Performance of the ZearalaTest 3%) was maintained. Optimal conditions were found
immunoaffinity columns with 10 ml of extract. When 20 ml or more extract
was loaded onto the column, the elution was slowed
The anti-zearalenone antibodies used by the manu- and cloudy extracts were obtained. Therefore, the
facturer to prepare the columns bind zearalenone loading of such high volumes requires vacuum to
analogues, especially a-zearalanol; in particular, they allow elution of the extract and an additional filtra-
showed 100% cross-reactivity to a-zearalanol and tion step before HPLC injection. Serious problems
more than 80% cross-reactivity to the other were encountered when filtration was performed with
zearalenone analogues, i.e. b-zearalanol, zearalanone nylon filters (Micro-Spin Centrifuge Filter, Alltech,
and a- and b-zearalenol [21]. However, this does not Deerfield, IL, USA) which resulted in the loss of
represent a problem when zearalenone determination zearalenone (up to 40%), in agreement with similar
is performed by HPLC which provides a good findings reported elsewhere [12].
separation of these compounds from zearalenone. Cloudy extracts (requiring filtration) were also
The ZearalaTest column capacity was found to be found when a dilution factor 1:5 (v / v with water

Fig. 1. Binding performance of anti-zearalenone antibodies used in the ZearalaTest immunoaffinity columns. Averages of duplicate
measurements (61S.D.) are represented.
 A. Visconti, M. Pascale / J. Chromatogr. A 815 (1998) 133 – 140 137

before loading the extract onto column) was used between 70 and 100% and R.S.D. r (within-laboratory
instead of 1:10 in the attempt to increase the relative standard deviation) ,25% for zearalenone
sensitivity of the method. concentrations higher than 100 ng / g [22].

3.3. Performance of the analytical method 3.4. Analysis of naturally contaminated samples
and comparison with the AOAC official method
Results of the recovery experiments (quadruplicate 958.15
measurements) of the full analytical procedure car-
ried out on spiked corn samples are reported in Table The analysis of fourteen corn samples which
1. Within the spiking range 0.1–10 mg / g the overall originated from northern Italy, 1997 crop, revealed
average recovery was 95.6%, with minimum value at the occurrence of zearalenone in all tested samples,
92.8%, and the average R.S.D. was 2.9%, with as shown in Table 2. The high incidence of
maximum value at 5.4%. These results are better zearalenone contamination of corn is not surprising,
than those obtained from the AOAC collaborative as similar results have been reported in several
study, based on liquid–liquid partition clean-up at countries. In particular, a multiyear monitoring of
different pH values [10], where recoveries of spiked Canadian grains showed 100% incidence (although
samples (0.05 to 4.00 mg / g) averaged 86.3% (from at relatively low levels) of zearalenone contamina-
78.7 to 103%) and the average repeatability was 30% tion in corn produced in Ontario in 4 years and
(ranging from 23 to 41%). The effect of the pH has generally higher than 67% in other years [6]; in some
not been taken into account in the present study due Argentinian regions (over 2200 samples analyzed
to the fact that a protocol of analysis is followed over a period of 10 years) contamination was found
which does not imply any significant change of pH. every year in corn with incidence up to 49% [23];
The limit of detection was 3 ng / g, based on a and 58% of corn samples were found contaminated
signal-to-noise of 3:1. The range of applicability of by zearalenone in a survey involving 19 countries
the method was from 3 to 10 000 ng / g of [24].
zearalenone in corn. For higher levels of contamina- The results obtained with the immunoaffinity
tion, the analysis should be repeated after appropriate method were in good agreement with those obtained
dilution of the acetonitrile–water extract. with the AOAC official method (Table 2). One false
Although no collaborative validation of the meth- negative sample (2) was found with the AOAC
od presented herein has been performed, the re- method and only three samples (7, 8 and 11) showed
covery and repeatability data reported above are R.S.D.s higher than 50% (Table 2). These results
within the criteria approved by CEN, the European obtained with single measurements from two differ-
Committee for Standardization, for the acceptability ent methods are however better than those found in
of analytical methods for zearalenone, i.e. recoveries the AOAC collaborative study, which gave a re-
peatability of 44.4% from duplicate analysis using
the same method.
Table 1 The AOAC method vs. immunoaffinity method
Recoveries (and relevant R.S.D.s) obtained with the immuno-
regression curve (Fig. 2) showed an R.S.D. of 0.87
affinity method from blank corn spiked with zearalenone at
different levels and the slope of 0.75 indicated better recoveries of
the immunoaffinity method. Sample 8 was removed
Spiking level Recovery6S.D.a R.S.D.
(mg / g) (%) (%) from the data set used in Fig. 2 and the relevant
statistical calculation because it proved to be an
0.1 98.665.3 5.4
outlier when the 95% confidence interval was con-
0.5 94.261.8 1.9
1.0 95.463.7 3.9 sidered for the regression curve using all original
2.0 99.562.5 2.5 data. Both the results of the analysis of naturally
4.0 92.862.9 3.1 contaminated samples and the recoveries obtained
10.0 92.860.4 0.4 with spiked materials are indicative of a better
a
S.D.5standard deviation (n54 replicates) accuracy of the immunoaffinity method as compared
138 A. Visconti, M. Pascale / J. Chromatogr. A 815 (1998) 133 – 140

Table 2
Determination of zearalenone in naturally contaminated corn by the immunoaffinity and the AOAC official method 985.18
Sample Immunoaffinity method AOAC official method R.S.D.
(ng / g) (ng / g) (%)
1 45 64 24.7
2 9.5 N.d. 2
3 78 44 39.4
4 3.6 4 7.4
5 18 9.6 43.0
6 60 71 11.9
7 23 10.4 53.3
8 55 140 61.6
9 78 77 0.9
10 150 100 28.3
11 9 39 88.4
12 66 40 34.7
13 37 27 22.1
14 11 8.6 17.3
15 N.d. N.d. 2
N.d.5not detected (,3 ng / g).

to the AOAC one, although a collaborative testing of ing signals at the zearalenone retention time in the
the immunoaffinity method is necessary. immunoaffinity purified extract (right), whereas the
Figs. 3 and 4 show typical chromatograms of sample analyzed by the AOAC method (left) shows
naturally contaminated samples containing levels of an interfering peak. A similar interfering peak was
zearalenone close to the detection limit (Fig. 3) and also observed in sample 4 together with another
to a contamination value commonly found in corn more important peak at longer retention time (Fig. 3,
(Fig. 4). Fig. 4 shows clearly the absence of interfer- left). These peaks are not present in the chromato-

Fig. 2. Regression curve of immunoaffinity method vs. AOAC official method for the determination of zearalenone in a blank and 14
naturally contaminated corn samples.
 A. Visconti, M. Pascale / J. Chromatogr. A 815 (1998) 133 – 140 139

Fig. 3. Chromatograms relevant to a corn extract (sample 4) naturally contaminated at levels close to the detection limit: (left) AOAC
official method and (right) immunoaffinity method. AOAC method. Injected extract: 20 ml (0.4 g sample equivalent); Column: Supelcosil
LC-18 (15034.6 mm I.D., 5 mm); mobile phase: acetonitrile–water–methanol (1.6:2.0:1.0, v / v); flow-rate: 1.3 ml / min; detection:
fluorescence at le x 5274 nm and le m 5440 nm (zearalenone concentration found54 ng / g). Immunoaffinity method. Injected extract: 100 ml
(0.16 g sample equivalent). Column: Supelcosil LC-18 (15034.6 mm I.D., 5 mm); mobile phase: acetonitrile–water–methanol (46:46:8,
v / v); flow-rate: 1.0 ml / min; detection: fluorescence at le x 5274 nm and le m 5440 nm (zearalenone concentration found53.6 ng / g).

Fig. 4. Chromatograms relevant to a naturally contaminated corn extract (sample 9): (left) AOAC official method (zearalenone concentration
found577 ng / g) and (right) immunoaffinity method (zearalenone concentration found578 ng / g). Chromatographic conditions as in Fig. 2.
140 A. Visconti, M. Pascale / J. Chromatogr. A 815 (1998) 133 – 140

gram of the immunoaffinity purified extract (Fig. 3, References

right).
In addition to the above mentioned advantages, i.e. [1] IARC, in: IARC Monographs on the Evaluation of Carcino-
accuracy and specificity, the immunoaffinity method genic Risks to Humans. Vol. 56. Some Naturally Occurring
Substances: Food Items and Constituents, Heterocyclic Aro-
avoids the use of very high amounts (|400 ml) of matic Amines and Mycotoxins, International Agency for
dangerous solvents such as chlorinated organic Research on Cancer, Lyon, 1993, p. 397.
chemicals, and is time saving due to the absence of a [2] D.B. Prelusky, B.A. Rotter, R.G. Rotter, in: J.D. Miller, H.L.
number of liquid–liquid extraction and solvent Trenholm (Editors), Mycotoxins in Grains — Compounds
other than Aflatoxins, Eagan Press, St. Paul, 1994, Ch. 9, p.
evaporation steps which are included in the AOAC
359.
protocol. [3] L.L. Charmley, A. Rosenberg, H.L. Trenholm, in: J.D.
Miller, H.L. Trenholm (Editors), Mycotoxins in grains —
Compounds other than Aflatoxins, Eagan Press, St. Paul,
1994, Ch. 13, p. 471.
[4] Worldwide Regulations for Mycotoxins 1995 — Compen-
4. Conclusions dium, Food and Agriculture Organization of the United
Nations (FAO), Rome, 1997.
The application of state of the art methods for the [5] A. Bottalico, A. Logrieco, A. Visconti, in: J. Chelkowki
determination of zearalenone in foodstuffs and feed- (Editor), Fusarium — Mycotoxins, Taxonomy and Patho-
genicity, Elsevier, Amsterdam, 1989, Ch. 5, p. 85.
stuffs is highly desirable for surveillance and moni- [6] P.M. Scott, Food Addit. Contam. 14 (1997) 333.
toring programs aimed at the prevention of undesired [7] N.M. Quiroga, I. Sola, J. Assoc. Off. Anal. Chem. Int. 77
effects for animal and human health and for animal (1994) 939.
production with consequent serious economical loss- [8] P.M. Scott, S.R. Kanhere, D. Weber, Food Addit. Contam. 10
(1992) 381.
es. The use of immunoaffinity chromatography in the
[9] J.C. Ryu, J.S. Yang, Y.S. Song, O.S. Kwon, J. Park, I.M.
clean-up step provides a number of advantages as Chang, Food Addit. Contam. 13 (1996) 333.
compared to other collaboratively tested methods, [10] G.A. Bennett, O.L. Shotwell, W.F. Kwolek, J. Assoc. Off.
which include: (i) provision of clean extracts due to Anal. Chem. 68 (1985) 958.
the specificity of the antibodies for the single toxin [11] H.L. Chang, J.W. DeVries, J. Assoc. Off. Anal. Chem. 67
(1984) 741.
or a group of related toxins that can be easily [12] N. Chamkasem, W.Y. Cobb, G.W. Latimer, C. Salinas, B.A.
separated by HPLC; (ii) optimal performances in Clement, J. Assoc. Off. Anal. Chem. 72 (1989) 336.
terms of precision and accuracy within a wide range [13] R.W. Bagneris, J.A. Gaul, G.M. Ware, J. Assoc. Off. Anal.
of concentrations which cover the field of practical Chem. 69 (1986) 894.
[14] T. Tanaka, A. Hasegawa, Y. Matsuki, U.S. Lee, Y. Ueno, J.
interest; (iii) rapidity and, finally, (iv) a considerable
Chromatogr. 328 (1985) 271.
reduction in the use of dangerous solvents, highly [15] M.M. Abouzied, J.J. Petska, J. Assoc. Off. Anal. Chem. Int.
desirable and necessary for environmental protection. 77 (1994) 495.
[16] E. Schneider, E. Usleber, E. Maertlbauer, R. Dietrich, G.
Terplan, Food Addit. Contam. 12 (1995) 387.
[17] G.A. Bennett, T.C. Nelson, B.M. Miller, J. Assoc. Anal.
Chem. Int. 77 (1994) 1500.
Acknowledgements [18] T. Tanaka, R. Teshima, H. Ikebuchi, J. Sawada, M. Ichinoe,
J. Agr. Food Chem. 43 (1995) 946.
We thank the staff of Vicam L.P. (Watertown, [19] P.M. Scott, M.W. Trucksess, J. Assoc. Off. Anal. Chem. Int.
MA, USA) and their Italian distributor (Safe Food, 80 (1997) 941.
[20] P. Cunniff (Editor), AOAC Official Methods of Analysis,
Viadana, MN, Italy) for kindly providing the im-
AOAC International, Arlington, VA, 1995, Ch. 49, pp. 45–
munoaffinity columns used in this study and for the 46.
useful discussion about the method applicability. We [21] B. Kohn, S. Kruger, Vicam L.P., Watertown, MA, personal
also thank Dr. P. Sarzi Amade´ (Safe Food) for communication.
providing the contaminated corn samples. This work [22] CEN (European Committee for Standardization), TC275-
WG5, Biotoxins, Doc. n 137, Berlin, 1997.
has been performed within the Special Project
[23] S. Resnick, S. Neira, A. Pacin, E. Martinez, N. Apro, S.
‘‘Quality Improvement in Agri-food Products’’ of the Latreite, Food Addit. Contam. 13 (1996) 115.
Italian National Research Council (CNR), National [24] T. Tanaka, A. Hasehawa, S. Yamamoto, U.S. Lee, Y. Sugiura,
Coordination Institute ‘‘Agro-Industry’’. Y. Ueno, J. Agr. Food Chem. 36 (1988) 979.