チョコレート中のアフラトキシン分析法の検討

誌名誌名 マイコトキシン

ISSNISSN 02851466

巻/号巻/号 581

掲載ページ掲載ページ p. 7-13

発行年月発行年月 2008年1月

農林水産省 農林水産技術会議事務局筑波産学連携支援センターTsukuba Business-Academia Cooperation Support Center, Agriculture, Forestry and Fisheries Research CouncilSecretariat

Mycotoxins Vol. 58 (1)， 2008 7

A method of analysis for determination of aflatoxins in chocolate

Ayako ITOH *1，3， Yasushi WATANABE本， and Masahiro NAKAJlMA *2

*' Analytical Laboratory Research Institute， FUjIYA Co.， LTD. (228， Soya， Hadano， Kanagawa 257-0031. ]apan) *2 Food Department， Nagoya City Public Health Research Institute (1-11， Hagiyarna-cho， Mizuho-ku， Nagoya 467-8615， japan)

Summary

A method of analysis for determination of aflatoxins (AFs) in chocolate was optimized and

evaluated its within laboratory precision. The method uses acetonitrile-methanol-water

(60+10+40， v/v/v) as extraction solvent and an immunoaffinity column (IAC) for the c¥eanup， and HPLC with f1uorescence detection for the determination of AFs.τbe method was validated

by testing replicate analyses of the finally ground chocolate samples spiked at 0.1 and 10.0

Ilg/kg for each AFB" AFB2， AFG， and AFG2. The average recovery of AFs at 0.1 and 10.0

μg/kg spiking levels ranged from 90 to 97 % for all AFs.τbe relative standard deviation (RSD)

for within day and between day ranged from 1.7 to 3.3 %， and 0 to 4.1 % for all AFs， respectively.

τbe HorRat for the total RSD at each level was within 0.2 for all AFs. These results show that

this method is reliable for the analysis of AFs in chocolate in the range of 0.1 and 10.0μg/kg.

Key words : aflatoxins， cacao， chocolate， immunoaffinity c¥eanup， HPLC，

within laboratory precision

(Received: October 29， 2007， Accepted: December 12， 2007)

Introduction

Aflatoxins (AFs: Aflatoxin B" B2， G， and G2) are mycotoxins produced mainly by Aspergillus

flavus， A. parasiticus and A. nomius， and known for their carcinogenic and hepatoxic effects on various

animal species'). Because of these high toxicity of AFs， the contamination of AFs in agricultural

commodities such as cereals， pulses， nuts， seeds， and spices2) is serious出reatto human health.

Cocoa beans (Theobroma cacao) are cropped in warm and humid areas such as A企ica，Latin

America， and Southeast Asia where AFs contamination in agricultural commodities may be occurred.

They are exported throughout the world and processed into cocoa products including chocolate and

cocoa powder. 1n recent years， consumption of cocoa beans amounts to more than 3 million tons per

year in the world3).τbe cocoa production chain includes fermentation step for 2-12 days to remove

sugary pulp. Cocoa beans are susceptible to spoilage by fungi during and after fermentation， and A.

flavus has been isolated from fermenting cocoa beans4). 1n fact， cocoa beans imported from Venezuela

have been occasionally rejected for contamination of aflatoxin B， (AFB，) above the regulatory limit

(10 ¥J.g/kg) in ]apan since 20055). Since AFs are generally thought to be stable in most food products6l，

*3To whom correspondence should be addressed.

8 Mycotoxins

AFs contaminated in cocoa beans could be carried over to cocoa products such as chocolate. There

are many analytical methods for AFs in various agricultural products乃，however， there are few reports

concerning the analysis for AFs in chocolate until now. These circumstances need developing a

reliable method of analysis for AFs in chocolate.

In recent years， the immunoaffinity column (IAC) cleanup-HPLC with fluorescence detection has

played an important role in AFs analysis7)， because of its simplicity， high specificity， and no use of

carcinogenic solvents such as chloroform and benzene. For the extraction of AFs from a wide variety

of matrices， mixtures of methanol-water and acetonitrile-water have been mainly used in the analytical

methods introducing the IAC8).

The aim of this study was to develop a reliable method for analysis of AFs in chocolate. Dark

chocolate was selected from chocolate samples as a test sample in this study， because dark chocolate

contains the highest cocoa ingredients among chocolate samples. The extraction was performed with

different composition acetonitrile-water and acetonitrile-methanol-water to establish the optimal

extraction solvent for AFs from chocolate. AFs were purified using an IAC and analyzed by HPLC

with fluorescence detection after trifluoroacetic acid (TFA) derivatization.百ledeveloped method was

also evaluated its within laboratory precision

Materia1s and me由ods

Sαmpleprepαrαtion Commercial dark chocolate was frozen at -20 oC， and ground with a mixer

(MX-V100， Matsushita Electric Industrial Co.， Ltd.， Osaka， Japan) to pass through 1000 11m sieve.百le

finely ground sample was transferred into plastic bag， and then mixed well to make a homogeneous

sample. The sample was stored at -20 oC until use. For evaluating recovery， 100 I1L of each AFs

working standard solution (10 ng and 1 I1g/mL for each AF) was added to 10.0 g sample (final

concentration was 0.1 I1g/kg and 10.0 I1g/kg， respectively) and left at room temperature for 1 hour

prior to analysis

Chemicαls AFBl， AFB2， AFG， and AFG2 crystalline standards were purchased from Sigma-Aldrich

Co. (St. Louis， MO， USA). Original standard solutions of each AF were prepared in toluene-

acetonitrile (90+10， v/v) and each concentration was determined according to the Official Method of

Analysis of AOAC International7). The AFs working standard solutions (1 ¥1g， 100 ng， and 10 ng/mL

for each AF) were prepared by dilution with toluene-acetorlItrile (90+ 10， v /v) for spiking and standard

curves. Easi-Extract AFLATOXlN IAC was purchased from R-Biopharm Rhone Ltd. (Glasgow，

Scotland). Phosphate buffered saline (PBS) tablet was purchased from Sigma-Aldrich Co. (St. 1ρuis，

MO， USA). PBS-0.01 % Tween 20 solution was prepared by dissolving 5 PBS tablets and 100 ¥11 of

Tween 20 in 1000 mL of purified water. Acetonitrile and methanol were of HPLC grade. All the other

reagents were of reagent grade.

Extr，αctionαndcleαnup Ten grams of the finely ground sample was homogenized with 40 mL of

extraction solvent at 10，000 rpm for 5 min using the homogenizer (Ace Homogenizer， Nihonseiki

Kaisha Ltd.， Tokyo， Japan). The extract was filtered through glass microfiber filter (Whatrnan 934AH，

Wha回 anInternational Ltd.， Maidstone， England). Four milliliters of filtrate (1 g sample equivalent)

was diluted to 100 mL with PBS-O.01-% Tween 20， and filtered through glass microfiber filter. Fifty

Vol. 58 (1)， 2008 9

milliliters of diluted filtrate (0.5 g sample equivalent) was applied onto an IAC at a flow rate of 1

drop/sec. After washing with PBS-O.01 % Tween 20 (3.3 mL x 4 times) followed by purified water (3.3

mL x 3 times)， the column was dried by passing the air into the column with a syringe. One milliliter

of acetonitrile was applied onto the IAC and kept for 5 min， and then AFs were eluted into a silanized

amber vial (4 mL， Supelco， Bellefonte， PA， USA) by gravity. The above eluting procedure was

repeated another 3 times (total 4 mL elution). The eluate was combined and evaporated to dryness

under a gentle stream of nitrogen gas at 45 oC. The residue was resolved in 0.1 mL of TFA and kept

for 15 min at room temperature， and 0.9 mL of injection solvent (acetonitrile-water， 10+90， v/v) was

added to the mixture.

HPLCαnαlysis The HPLC system consisted of a LC-20AD pump， a SIL-20AC auto sampler， a

CTO-20AC column oven， a RF・10AXLfluorescence detector (excitation 360 nm， emission 450 nm)， a

DGU-20A3 degasser， a CBM-20A communication bus module and a LCsolution data system

(Shimadzu Co.， Kyoto， Japan).官leanalytical column (Inertsil ODS-3V， 4.6 mm x 250 mm， 5μm，GL

Sciences Inc.， Tokyo， ]apan) was kept at 45 oC with a mobile phase of acetonitrile-methanol-0.2 mol/L

ammonium acetate buffer (pH5.0)-purified water (10+30+5+55， v /v /v /v) and a flow rate of 1.0

mLlmin. 100 ¥lL of AFs standard solutions (0.05， 0.1， 1.0， 5.0， 10.0， 20.0 ng/mL for each AF) or test

sample solutions was injected to HPLC system. The standard curves were prepared by plotting the

peak height against the concentration of AF standards. Quantification of each AF in the test sample

solution was performed by measuring peak height at retention time of each AF and comparing them

with the relevant standard curves.

Stαtisticalαnαlysis The within laboratory precision (recovery， relative standard deviations

(RSD) for within day and between day， HorRat) was deduced using the on令 wayanalysis of variance

(ANOVA) according to the AOAC guideline9) and the others¥0-¥2).

Results and discussion

Lineαrity of stαndαrd curves As shown in Fig. 1， the standard curves and the linear

regression equations for each AF in the range of 0.0ι20.0 ng/mL (equivalent to 0.1 to 40μg/kg of

each AF in a sample) showed R2 values greater than 0.9999， indicating good linearity.

Optimizαtion of extraction A preliminary experiment was carried out to establish the optima1

extraction solvent for AFB¥企omspiked chocolate samples. Since acetonitrile-water has been

previously reported to give better mycotoxins recovery than methanol-water in mycotoxins analysis¥3. ¥4)，

acetonitrile-water (60+40， v /v) was firstly selected as extraction solvent for AFB¥ from chocolate

samples. However， this solvent mixture was not applicable for chocolate samples， because the use of

this solvent mixture resulted in phase separation showing the two separated layers by the sa1ting out

effect derived from sample ingredients as shown in Fig. 2A. Our data showed that most AFB， was concentrated in the upper layer (acetonitrile layer) and the recovery rate of AFB， was nearly to 150 %

when the upper layer was used as a sample extract. In conclusion， this solvent mixture seemed to be

inappropriate for analysis of AFs in chocolate samples because AFs accumulation in the upper layer

can give unrealistic high recovery.τbe mixture of acetonitrile-water (90+10， v/v) adopted by the

Japanese official method of analysis for AFB¥'5) was also tested as an extraction solvent， however， the

600000

500000

五4@= 400000

300000 .占S代ζ 200000

100000

。20 25 。

600000

500000

る」@= 4000∞ 300000

」Z腐豆 200000

100000

。20 25 。

10

nunUAUAU《

ununu

nununυnu内

ununu

nunu内

U

《

ununυnu

nununununununυ

nununuAunUAUnu

守

，

FOraaa守内

。

内

4

・E

HZ回-@£d

品伺@乱

y:: 29623x -802.73

R2 :: 0.99992

10 15

AFB， (ng/ml)

500000 r y:: 21812x + 474.44 R2 :: 0.99994

5

400000

nuAU

nunυ

nunυ

nu

《

u

nunu

qd

内

4

4FS回-@F-4mw@且

100000

。。 5 10 15

AFG， (ng!ml)

Mycotoxins

y:: 27804x + 107.52

R2 :: 0.99999

10 15 20 25

AFB2 (ng!ml)

y:: 27450x + 695.65

R2 :: 0.99998

5

5 10 15 20 25

AFG2 (ng/ml)

Fig. 1. Standard curves for aflatoxin B" B2， G， and G2・

A B

Fig.2. Ex廿actsfrom chocolate samples after filtration.

A:百1e extract wi出 acetoni仕il守water(60+40，

v/v). B:τbe extract with acetonitrile-methanol-

water (60+10+40， v/v/v).

recovery rate of AFBI from a chocolate sample was only about 70 % due to hardly mix the sample with

solvent.

Scott et al.16) reported that the problem of phase separation was avoided by using a mixture of

acetonitrile-methanol-water in the fumonisins analysis. According to this information， the extraction

11

was performed with different composition acetonitrile-methanol-water to get optimized recovery of

AFBt from chocolate samples. Optimization was based on the following criteria: 1. Whether the

extract after homogenization showed the homogeneous or the heterogeneous state. 2. Whether the

filtered extract showed a single layer or two separated layers. 3. Whether the recovery rate of AFBt

was acceptable or unacceptable. As shown in Table 1， with acetonitrile-methanol-water (70+10+30，

v/v/v) and (80+10+20， v/v/v)， both extracts after homogenization showed the heterogeneous state，

and both extracts after filtration were separated into two liquid layers. With acetonitrile-methanol-

water (60+10+40， v/v/v) and (70+15+30， v/v/v)， both extracts showed the homogeneous state and a

single layerぐTable1 and Fig. 2B)， however， the recovery rate of the latler ex仕actionmixture

exceeded 100 %. Taking into account these results， acetonitrile-methanol-water (60+10+40， v/v/v) was

adopted as ex仕actionsolvent in the method of analysis for AFs in dark chocolate.

Chromatogrαms of chocolate samples The chromatograms of chocolate samples spiked at

0.1 and 10.0 llg/kg for each AF were shown in Fig. 3. Each AF at 0.1 llg/kg spiking level was fully

detected (Fig. 3A)， and all AFs were well separated under the given conditions， and no interfering

peaks were detected in the chromatograms.

Vol. 58 (1)， 2008

Table 1. Comparison of solvents for ex廿actionof afIatoxin Bl from chocolate spiked at 5.0 ¥Ig/kg

Recovery of afIatoxin B I (%)

τbe number of liquid layer

Condition after homogenization

Composition of solvent

Me出anol

94 116

109

113

11ayer 21ayers 11ayer 21ayers

homogeneous heterogeneous homogeneous heterogeneous

Water A

V

A

U

A

V

A

U

氏U

弓

tη，4

0

0

40

30 30 20

A

U

Aリ

R

d

Aリ

唱

EA'EA

唱

EA

唱

EA

Acetonitrile

A

2-4一一

1-1

4'!!l-

mln

B

22.5 20.0 17.5

主 i ロ: 詑

き ケ " ~ ":"" T:'" 一ーーー.h<:

15.0 12.5 10.0 7.5 5.0

百弓面五瓦両元E面nm

2.5

。ト0.0

郎V

150

100

50

mln

Fig.3. Chromatograms of chocolate samples spiked at 0.1 ¥Ig/kg (A)， and 10.0 ¥Ig/kg (B) for each afIatoxin.

22.5 20.0 17.5 15.0 12.5 10.0 7.5 5.0 2.5

。「

0.0

12 Mycotoxins

Within [，αborαtory precision The optimized method of analysis for AFs in chocolate was

evaluated its within laboratory precision by spiked samples. Chocolate samples were spiked with AFs

working standard solution at 0.1 and 10.0 ¥lg/kg for each AF， and 5 replicates were analyzed for 4

days at each level. Table 2 and 3 show the results of the recovery and the within laboratory precision.

The average recovery of AFs at 0.1 and 10.0 ¥lg/kg spiking levels was 91 % and 91 % (B1)， 94 % and 92

%但ふ 90% and 92 % (G1)， 97 % and 93 % (G2)， respectively.百lewithin day RSD at each level was 2.9

% and 2.4 % (81)， 1.9 % and 2.2 % (82)， 3.3 % and 2.6 % (G1)， 1.7 % and 3.0 % (G2)， respectively. The

between day RSD at each level was 1.2 % and 0.6 %但1)，4.1% and 0.6 % (82)，2.0 % and 1.2 % (G1)， 2.1 %

and 0 % (G2) ， respectively.百leto凶 RSDat each level was 3.1 % and 2.5 %但1)，4.5% and 2.3 %但2)，3.8

% and 2.8 % (G1)， 2.7 % and 3.0 % (G2)， respectively. The HorRat for the total RSD at each level was 0.1

and 0.2 (81)， 0.1 and 0.1但2)，0.1 and 0.2 (G1)， 0.1 and 0.2 (G2)， respectively. In conc1usion， it was

proved that the developed method in this study was successfully applied to加 alysisfor AFs in dark

chocolate in the range of 0.1 and 10.0問 /kg.

Table 2. Recovery and wi白inlaboratory precision at a spiking level of O.l11g/kg (n=20)

Average RSD判(%)

recovery (%) HorRat，

Within day Between day Total

Aflato羽nBI 91 2.9 1.2 3.1 0.1 Aflatoxin B2 94 1.9 4.1 4.5 0.1 Aflatoxin GI 90 3.3 2.0 3.8 0.1 Aflatoxin G2 97 1.7 2.1 2.7 0.1

* 1: Relative standard deviation

Table 3. Recovery and wi出inlaboratory precision at a spiking level of 10.0 llg/kg (n=20)

Average RSD判(%)

recovery (%) HorRat，

Within day Between day Total

Aflatoxin B， 91 2.4 0.6 2.5 0.2 Aflatoxin B2 92 2.2 0.6 2.3 0.1 Aflatoxin GI 92 2.6 1.2 2.8 0.2 Aflatoxin G2 93 3.0 O 3.0 0.2

本1:Relative standard deviation

References

1) IARC:“IARC Monograph on the Evaluation of Carcinogenic Risks to Humans." 82， pp. 171-300 (2002)，

International Agency for Research on Cancer， Lyon， France

2) ]ECFA: AfIatoxins， WHO Food Additives Series， 40， pp. 359-468 (1998)， World Health o rgan ization ，

Geneva， Switzerland

3) h仕p://www.icco.org/statistics/grindings.aspx， International Cocoa Organization (lCCO)， 1ρndon， UK

4) Oyetunji， TO.: A企ican]. Biotech.， 5， 2073-2076 (2006)

5) h均://www.mhlw.go.jp/topics/yunyu/tp0130-1.html. Ministry of Health， Labor and Welfare， ]apan

6) Bullerman， L.B.: ]. Food Prot.， 42， 65-86 (1979)

Vol. 58 (1)， 2008 13

7) Horwitz， W. (edよ“OfficialMethods of Analysis of AOAC International， 18，h edition"， Chapter 49 (2005)，

AOAC International， MD， USA

8) Application Note: Aflatoxin extraction method (2001)， R-Biopharm Rhone， Glasgow， Scot1and

9) AOAC Guidelines for Single Laboratory Validation of Chemical Methods for Dietary Supplements and

Botanicals， http://www.aoac.org/dietsupp6/Dietary-Supplement-webもite/slv...，guidelines.pdf，AOAC In-

ternational， MD， USA (2002)

10) http://www.aoac.org/method_validation/Addenda/.AOAClnternational.MD. USA

11) Thompson， M.， Ellison， S.L.R.， Wood， R.: Pure App1. Chem.， 74，835-855 (2002)

12) Nagata， T， Goto， T， Tanno， K， Yasui， A.， Yukawa， G. (eds.): Handbook of Method Validation in Food

Analysis， (2007)， Scienceforum Inc.， Tokyo， ]apan

13) Gりbel，B.， Lusky， K: J. AOAC Int.， 87， 411-416 (2004)

14) Solfrizzo， M.， Girolamo， A.， Visconti， A.: Food Addit. Contam.， 18，221-235 (2001)

15) Notification No. 326001 (March 26， 2002)， Director of Inspection and Safety Division， Department of

Food Safety， Ministry of Health， Labor and Welfare， ]apan

16) Sco此， P.M.， Lawrence， G.A.， Lombaert， G.A.: Mycotoxin Res.， 15， 50-60 (1999)

チョコレート中のアフラトキシン分析法の検討

伊藤綾子，渡辺康:(株)不二家 (257-0031 秦野市曽屋 228)

中島正博 :名古屋市衛生研究所 (467-8615 名古屋市瑞穂区萩山町 1-11)

チョコレート中のアフラトキシン(アフラトキシン BI，B2， GI， G2 ; AFs)分析法を確立し.室内再現精度

の確認を実施した.チ ョコレ ー ト中の AFsは，アセト ニ トリルーメ タノール 水 (60+10+40，v/v/v)で抽

出し，イムノアフィニティカラムにより精製後，蛍光検出 HPLCにて定量を行った.試料に AFs標準液を

0.1および 10.011g/kgの濃度となるように添加し，日内および日間における繰り返し試験を行った結果，全

ての AFsにおける回収率は90-97 %， 併行再現性の相対標準偏差は1.7-3.3 %，日間再現性の相対標準偏

差は0-4.1 %，巽日分析における室内再現性の相対標準偏差に対するHorRatは， 全て 0.2以内であった.

以上の結果から.チョコレート中の AFsは本法により精度良く分析できる事が確認された.

キーワード:アフラ トキシン，カ カオ ，チョ コレート ，イムノ アフィニティ ークリーンア ップカラム，

高速液体クロマ トグラフィー，室内再現精度