Effect of Variety, Locality and Processing of Coffee Beans on the

INTERNATIONAL JOURNAL OF AGRICULTURE & BIOLOGY
1560–8530/2005/07–1–5–10
http://www.ijab.org
Effect of Variety, Locality and Processing of Coffee Beans on the
Detection and Determination of Aflatoxins
K.M. SOLIMAN
Food Toxicology and Contaminants Department, National Research Center, Dokki, Cairo, Egypt
Corresponding e-mail: Kortja1@ hotmail.com
ABSTRACT
Studies were conducted on the detection and determination of aflatoxins after processing in coffee beans originating from
various parts of the world (Brazil, Ethiopia, Colombia and Indonesia) using standard AOAC methods. The detection of
aflatoxins in green coffee beans could not be successful in overcoming the masked material that covered AFB2, G1 and G2 in
all samples under study. However, AFB1 was the only spot that appeared on the TLC. Due to roasting of coffee beans, the
interfering material with aflatoxins increased and even AFB1 could not be detected by the AOAC methods. An improved
procedure which uses silver nitrate to get rid of caffeine and related compounds, proved useful in removing the interfering
materials due to which the coffee beans matrix and aflatoxins spots were separated clearly. However, this method failed to
remove the interfering materials that masked the area of aflatoxins in roasted coffee beans completely. HPLC was tested as an
alternative technique for the detection of aflatoxins. HPLC chromatogram of green coffee beans using the AOAC method
showed unidentified intensity peaks and low recovery was obtained. HPLC coupled with the improved procedure of extraction
showed that the unidentified peaks due to the matrix of green and decaffeinated coffee beans were eliminated. However, it
failed to overcome the problems of the roasted coffee beans especially in the area of AFG1. An obvious effect of cultivation
locality on the matrix of green coffee beans was observed. This effect also appeared in the recovery percentage of aflatoxins.
Also, this locality effect did not allow reaching a solid conclusion for the effect of variety.
Key Words: Aflatoxins; Coffee beans; Variety; Origin; HPLC
INTRODUCTION
The exact number of species within the genus coffea
(family Rubiaceae), is not known, but is probably around
ninety. They are found in the centre of Africa, from the
Congo basin to the highlands of Ethiopia. Two species are
significant in world trade: coffea arabica and coffea robusta.
Arabica coffee developed in highlands of Ethiopia, the
climate is cool by tropical standards and rainfall fairly high
(Willson, 1999).
Robusta coffee is developed in lowland forest of the
Congo River basin extending up to lake Victoria in Uganda.
Robusta is less aromatic and higher in caffeine and acidity
than Arabica bean. Coffee bean has a complex chemical
composition and as variable as nature and man’s attendance
to its cultivation (Willson, 1999).
Coffee is roasted for about 10 to 20 min at
temperatures ranging from 400 to 425ºF. During roasting,
the chemical make-up of the coffee beans changes giving
much of its aroma and flavor. The caramelize of sugar turns
the coffee bean from green to brown (Sievetz & Desrosier,
1979; Massini et al., 1990; Gutierrez et al., 1993; Pittia et
al., 2001). Large number of volatile compounds in very
small amounts give coffee its unique flavour and aroma
(Clifford & Willson, 1985; Willson, 1999). To date, over
1000 volatile chemicals have been identified in coffee
(Fuster et al., 2000; Clarke & Vitzthum, 2001).
Contamination may occur when either grown crops
(Littehoj & Zuber, 1975) or more badly stored harvests are
infested by moulds (Jckman, 1985). Among the fungal
species, Aspergillus and Penicillium sp. were found to affect
bean color and beverage quality (Batista et al., 2002; Gerrit,
2003). The aflatoxins are potent hepatotoxic and
carcinogenic metabolites produced by the fungi Aspergillus
flavus and A. parasiticus (Diener et al., 1987). Betancourt
and Frank (1983), found that it is necessary to keep the
moisture content (MC) of coffee within a limit of 14.5% to
prevent the mold growth during storage.
The published analytical methods for the detection of
aflatoxins in green coffee beans using TLC technique are
still affected by poor backgrounds and interfering
fluorescent material (Levi & Borker, 1968; Levi, 1969;
Scott, 1968; Nortowicz et al., 1979; Soliman, 2002, 2004).
The aim of this work was to study the different
methods to detect aflatoxins in green coffee beans of
different varieties and localities as well as in different coffee
products.
MATERIALS AND METHODS
Samples. Indonesian and Brazilian coffee beans which
represent the robusta and arabica variety, respectively;
Brazilian, Ethiopian and Colombian coffee beans have the
same variety (arabica) which represent the different origin
SOLIMAN / Int. J. Agri. Biol., Vol. 7, No. 1, 2005
x 4.6 mm) was used for aflatoxins separation. The mobile
phase methanol:water:acetonitrile (25:65:10, v/v/v) was
isocratically used at flow rate of 1 mL/min. The obtained
data were integrated and calculated by Chemstation
software program.
Caffeine analysis. Caffeine was extracted and determined
in green coffee beans samples according to Madison et al.
(1976).
as well as roasted and decaffeinated coffee beans of
Colombian were selected.
Preparation of spiked samples. The stock solutions of AFs
were mixed with samples of coffee beans under study to get
a final concentration of 40 ppb of total aflatoxins (15, 15, 5
and 5 ppb for AFB1, B2, G1 and G2, respectively).
Sample extraction. The extraction and clean up of
aflatoxins were done using the following methods:
A) AOAC for green coffee beans (2000), the finely ground
coffee beans (50 g) were mixed with CHCl3 (250 mL) and
10% H2O and shaken for 30 min, then the extract was
filtered through Whatman No. 1 filter paper. A portion of
filtrate (50 mL) was loaded onto the florisil column and
allowed to flow at a rate 1 drop per second and then the
extract was rinsed with tetrahydrofuran (THF). The
aflatoxins were eluted with 300 mL acetone-methanol (97:3
v/v) and evaporated to dryness on steam bath under
nitrogen.
B) The improved procedure according to Soliman (2004).
The sample (50 g of finely ground coffee beans) was shaken
with 250 mL CHCl3 and 25 mL AgNO3 solution (25g/100
mL H2O) for 30 min and filtered through filter paper No 1.
The filtrate (75 mL) was mixed with 7.5 mL AgNO3 and
shaken for 15 min. The mixture was poured in separating
funnel and 50 mL of CHCl3 layer was collected for clean up
in Sep-Pak florisil (Soliman, 2004). The THF was used to
wash the extract and acetone: methanol (97:3 v/v) was used
to elute the aflatoxins and evaporated to dryness on steam
bath under nitrogen. The residue was quantitatively
transferred to small vial with CHCl3 and evaporated on
steam bath under nitrogen and was reserved for TLC and
HPLC.
Thin layer chromatography. The thin layer
chromatography (TLC) plates (Merck, silica gel G type 60)
without fluorescence indicator was used for aflatoxin
detection.
Dried residues were dissolved in 200 µL of benzene:
acetonitrile (98:2 v/v) and 10 µL of each sample or standard
aflatoxins solution were spotted on the same plate. The
developing system consists of benzene:alcohol:H2O
(46:35:19 v/v/v), two phase solvents system according to
Levi (1969) was used, where the bottom layer was used for
the saturation of the tank and the upper layer for the
developing of the spotted plate. The plate was examined
under long wave of the UV light (365 nm).
High performance liquid chromatography (AOAC,
2000). The final extracts from the above were
derivativatized by trifluoroacetic acid according to the
AOAC (2000). The HPLC used for aflatoxins determination
is an Agilent 1100 system equipped with quaternary pump
model G1311A, fluorescence detector model G1321A set at
360 nm excitation and 440 nm emission wavelengths and
auto-sampler model G 1329A. ODS (Zorbax) column (150
RESULTS AND DISCUSSION
The coffee in world trade comes from two major
species (arabica & robusta). The effect of interferences
compounds on the detection methods of aflatoxins
contamination in green coffee beans and its products needs
to be focused on. Spiked and aflatoxins-free samples were
analyzed by AOAC (2000) method and improved procedure
(Soliman, 2004) for green coffee beans.
Effect of Green Coffee Beans Matrix on the Detection of
Aflatoxins
Effect of green coffee beans verities. To study the effect of
variety matrix on the detection of aflatoxins, samples of
Indonesian (robusta variety) and Brazilian coffee arabica
variety coffee beans were spiked with aflatoxins at a level of
40 ppb and analyzed with above mentioned procedures.
Both the arabica (Brazilian) and robusta (Indonesian)
varieties, when extracted with the AOAC method showed
that AFB1 was the only spot that appear in the TLC plate
(Fig. 1). In Indonesian coffee beans an extraneous
fluorescent spot appeared below the Rf of AFG2. However,
this fluorescent spot did not appear in the extract of
Brazilian coffee beans.
This fluorescent spot appearing in both spiked and
aflatoxins-free coffee beans of the robusta variety may
indicate the relation of variety to each spot. Soliman (2004)
showed that the fluorescent spot is a compound which may
be related to the matrix of coffee beans and is most probably
a caffeine compound. This result matches with the high
level of caffeine in robusta (16.3 mg/g) compared with
arabica (6.9 mg/g) (Table I). In this regard Willson (1999)
reported that the caffeine and acidity content of the robusta
coffee beans are about twice that of arabica.
An improved procedure reported by Soliman (2004),
which uses silver nitrate as the aqueous solution to
overcome the different problems associated with the AOAC
procedure, was successful to remove the interfering
materials in green coffee beans. So a good background was
obtained on TLC and the four types of aflatoxins were
separated clearly (Fig. 2). This result is in agreement with
Scott (1969) and Soliman (2004) who reported that silver
nitrate can be used to get rid of the caffeine and related
compounds.
6
AFLATOXINS IN COFFEE BEANS / Int. J. Agri. Biol., Vol. 7, No. 1, 2005
Table I. The level of caffeine (mg/g) in four types of
green coffee beans
Fig. 2. Effect of green coffee beans variety on
detection of aflatoxins using an improved procedure.
Item
Brazilian Colombian
Caffeine (mg/g) 6.9±0.89 5.4±0.92
1, Aflatoxins-free Indonesian coffee beans; 2, Spiked-Indonesian
coffee beans with aflatoxins; 3, Aflatoxin standards; 4, SpikedBrazilian coffee beans with aflatoxins. 5, Aflatoxins-free Brazilian
coffee beans.
Ethiopian
8.2±1.80
Indonesian
16.3±2.60
Table II.
Comparison of aflatoxins percent
recoveries of green; roasted and decaffeinated coffee
beans using the improved procedure
Matrix
Robusta
Indonesiangreen
Arabica Brazilian-green
Ethiopian-green
Colombian-green
Colombiandecaffeinated
Colombian-roasted
1
2
3
4
5
B1
B2
G1
G2
Total
70.0±4.2 81.5±5.8 67.0±8.3 80.0±4.5 74.6±4.7
89.0±4.2
88.5±7.2
62.3±9.2
81.5±7.9
86.7±6.2
86.5±7.9
78.2±6.1
85.6±8.4
87.8±5.8
62.9±7.5
40.2±7.4
46.4±8.5
91.4±4.5
80.8±6.7
71.7±8.1
75.7±6.8
88.7±5.2
79.7±5.3
63.1±6.7
72.3±7.1
63.9±7.3 78.9±8.7 42.8±5.5 71.4±8.7 64.3±7.6
Fig. 1. Effect of green coffee beans variety on
detection of aflatoxins using the AOAC method. 1,
Aflatoxins-free Indonesian coffee beans; 2, Spiked-Indonesian coffee
beans with aflatoxins; 3, Aflatoxin standards; 4, Spiked-Brazilian
coffee beans with aflatoxins; 5, Aflatoxins-free Brazilian coffee beans
1
2
3
4
Fig. 3. Effect of origin green coffee beans on detection
of aflatoxins using the AOAC method. 1, Aflatoxins-free-
5
Brazilian coffee beans; 2, Spiked Brazilian coffee beans with
aflatoxins; 3, Aflatoxins standard; 4, Aflatoxins-free-Colombian coffee
beans; 5, Spiked Colombian coffee beans with aflatoxins; 6,
Aflatoxins-free-Ethiopian coffee beans; 7, Spiked Ethiopian coffee
beans with aflatoxins.
1
Effect of green coffee beans origin. To study the effect of
coffee beans origin (locality of cultivation) on the detection
of aflatoxins, samples of the same variety (arabica) were
selected to represent different origin of cultivation. These
samples were Brazilian, Ethiopian and Colombian coffee
beans.
Fig. 3 shows that the AOAC procedure could not
overcome the masked material that covered the area of
AFB2, G1 and G2 in all samples under study and AFB1 was
the only spot which appeared in these samples. Slight
fluorescent spot at Rf below AFG2 appeared in Ethiopian
coffee beans which means that higher content of caffeine
may be present in Ethiopian samples. This finding is
confirmed by the higher level of caffeine in Ethiopian
samples that was detected by HPLC in the current study (8.2
2
3
4
5
6
7
mg/g) compared with those of Brazilian and Colombian (6.9
and 5.4 mg/g, respectively) (Table I).
Using the improved procedure of Soliman (2004), the
masking material was not observed in the same samples
(Fig. 4). Four spots of aflatoxins were detected clearly in
Brazilian coffee; however, AFG1 did not appear in
Colombian and Ethiopian coffee beans on TLC.
The percentage recoveries of aflatoxins in Indonesian,
Brazilian, Ethiopian and Colombian green coffee beans are
7
SOLIMAN / Int. J. Agri. Biol., Vol. 7, No. 1, 2005
Fig. 4. Effect of origin green coffee beans on detection
of aflatoxins using an improved procedure. 1,
Fig. 5. Effect of locality and variety variation on
percentage recovery of total aflatoxins.
Aflatoxins-free-Brazilian coffee beans. 2, Spiked Brazilian coffee
beans with aflatoxins; 3, Aflatoxins standard; 4, Aflatoxins-freeColombian coffee beans; 5, Spiked Colombian coffee beans with
aflatoxins; 6, Aflatoxins-free-Ethiopian coffee beans; 7, Spiked
Ethiopian coffee beans with aflatoxins.
1
2
3
4
5
6
100
88.7
79.7
7
74.6
80
P e r c e n ta g e r e c o ve r y ( % )
63.1
60
40
20
0
Indonesian
Brazilian
Ethiopian
Colombian
Green coffee beans
Fig. 6. Effect of processing coffee beans on detection
of aflatoxins using the AOAC method. 1, Aflatoxins-free
Fig. 7. Effect of processing coffee beans on detection
of aflatoxins using an improved procedure. 1, Aflatoxins-
Colombian roasted coffee beans; 2, Spiked-Colombian roasted coffee
beans with aflatoxins; 3, Aflatoxins standard; 4, spiked-Colombian
decaffeinated coffee beans with aflatoxins; 5, Aflatoxins-free
Colombian decaffeinated coffee beans.
free Colombian roasted coffee beans; 2, Spiked-Colombian roasted
coffee beans with aflatoxins; 3, Aflatoxins standard; 4, SpikedColombian decaffeinated coffee beans with aflatoxins; 5, Aflatoxinsfree Colombian decaffeinated coffee beans.
1
2
3
4
5
1
presented in Table II and Fig. 5. No general conclusion of
the variety effect could be inferred due to an apparent effect
of the locality of cultivation, which may mask such varietal
effect. So the higher recovery of Brazilian coffee beans
(88.7%) compared to Indonesian coffee beans (74.6%) may
not be due to the effect of variety variation rather than the
effect of locality which is obviously observed in comparing
the same variety, the Brazilian coffee beans (88.7%) and the
Colombian samples (63.1%).
2
3
4
5
Effect of coffee processing on the detection of aflatoxins.
Processing (roasting and decaffeination) of coffee beans
causes the development of compounds with new flavor and
color that may interfere with the detection of aflatoxins in
the samples of processed coffee beans (Lerici & Nicoli,
1990; Pitta et al., 2001).
Roasting of coffee beans. To study the effect of roasting on
the efficiency of aflatoxins detection procedure, roasted
Colombian samples were used. As expected, the AOAC
procedure failed to separate the aflatoxins spots from the
8
AFLATOXINS IN COFFEE BEANS / Int. J. Agri. Biol., Vol. 7, No. 1, 2005
Fig. 8. Effect of processing on percentage recovery
of total aflatoxins
Fig. 9. HPLC chromatogram of aflatoxins spiked
green coffee bean sample extracted by AOAC
method for green coffee bean.
72.3
80
Unidentified
64.3
63.1
Percentage recovery(%)
60
40
B2
20
Unidentified
B1
G1
0
Green
Decaffinated
G2
Roasting
Colombiancoffee beans
masked material that covered the area of aflatoxins spots
(Fig. 6). The interfering materials that masked the area of
aflatoxins were less intense when the improved procedure
was used to extract the aflatoxins in roasted Colombian
coffee beans. The spots of AFB1 and B2 although appeared
but not clearly (Fig. 7).
The results obtained in the current study confirmed
that the AOAC method (2000) is described only for green
coffee beans and this method is not suitable for the detection
of aflatoxins in roasted coffee beans.
The results further showed that although the improved
procedure succeeded to resolve the problems of the
detection of aflatoxins in green coffee beans, it failed to
overcome the problems of pigmented compounds produced
during roasting process. In this regard, Sievetz and
Desrosier (1979) stated that roasting coffee beans causes
marked chemical, physical, structural and sensorial changes.
Therefore, the need to establish a specific method for the
detection of AFs in roasted coffee beans is highly
recommended.
Decaffeination of coffee beans. Colombian-decaffeinated
coffee beans were used to study the effect of decaffeination
on the detection of aflatoxins using the above mentioned
procedures (AOAC and the improved one). Fig. 6 showed
poor background on TLC plate with AOAC extraction,
however better background was observed when the
improved procedure was used (Fig. 7).
Data in Table II and Fig. 8 showed that the best
recovery was obtained from the decaffeinated samples
(72.3%) compared with the green (63.1%) and roasted
coffee beans (64.3%). The AFB2 recorded the highest
recovery (85.6, 78.9 & 78.2%) for decaffeinated, roasted
and green coffee beans, respectively. On the other hand the
lowest recovery was recorded for AFG1 (46.4, 42.8 and
40.2, for decaffeinated, roasted and green coffee beans,
respectively) compared with the other aflatoxins.
Fig. 10. HPLC spiked green coffee beans
chromatogram of aflatoxins extracted by
improved procedure.
B2
B1
G1
G2
HPLC as an alternative technique for the detection of
aflatoxins. HPLC was tested as an alternative technique to
avoid problem of the interfering compounds on TLC. The
HPLC chromatogram of green coffee beans using the
AOAC method (Fig. 9) showed two unidentified high
intensity peaks (Soliman, 2004). These peaks were well
resolved from the aflatoxins peaks. This showed that HPLC
determination coupled with AOAC as a method of
extraction could resolve the problem of TLC; however, the
low recovery of the AOAC method (46.1%) (Soliman,
2004) may prevent such approach. When the improved
procedure was used for the extraction of green and
decaffeinated coffee beans, the unidentified peaks due to the
matrix of the same above mentioned samples was
9
SOLIMAN / Int. J. Agri. Biol., Vol. 7, No. 1, 2005
Fig. 11. HPLC chromatogram of aflatoxins spiked
decaffeinated
coffee beans extracted by the
improved procedure
B1
TLC technique during the detection of aflatoxins in green
and decaffeinated coffee beans, however, the problem of
roasted coffee beans still exited (Fig. 12). This still needs a
special method for roasted coffee beans.
B2
REFERENCES
AOAC, 2000. Official methods of analysis, Vol. 17. Association of Official
Analytical Chemists. Washington, D.C.
Batista, L.R., S.M. Chalffoun, G. Prado, R.F. Schwan and A.E. Wheals,
2002. Toxigenic fungi associated with processed (green) coffee
beans (Coffea arabica L.). Int. J. Food Microbiol., 62: 1–8.
Betancourt, L.E. and H.K. Frank, 1983. Behavior of ochratoxin A during
green coffee roasting and soluble coffee manufacture. J. Agric. Food
Chem., 46: 673–5
Clarke, R.J. and O.G. Vitzthum, 2001. Coffee: Recent Developments.
Blackwell Science. Oxford and Madden, MA.
Clifford, M.N. and K.C. Willson, 1985. Coffee: Botany Biochemistry and
Production of Beans and Beverage. Chapman and Hall London.
Diener, U.L., R.J. Cole, T.H. Sanders, G.A. Payne, L.S. Lee and M.A.
Klich, 1987. Epidemiology of aflatoxins formation by Aspergillus
flavus. Annual Rev. Phytopathology, 25: 249–70
Fuster, M.D., A.E. Mitchell, H. Ochi and T. Shibamoto, 2000.
Antioxidatives activities of heterocyclic compounds formed in
browed coffee J. Agric. Food Chem., 48: 5600–3
Gerrit, H.G., 2003. Enhancement of coffee quality by mould prevention.
Food Control, 14: 245–9
Gutierrez, C., A.D. Ortola, A. Chiralt and P. Fito, 1993. Analysis for MEB
de la proosidad del café tostado. In: ASIC, Processings 15 Internat.
Scientific Colloquium on Coffee, pp. 661–71. Vol. 2, Montpellier
Jackman, R., 1985. Determination of aflatoxins by enzyme–linked
immunosobent assay with special reference to aflatoxins M1 in raw
milk. J. Sci. Food Agric., 36: 685–98
Levi, C.P. and E. Borker, 1968. Survey of green coffee for potential
aflatoxins contamination. J. AOAC, 51 : 600–2
Levi, C.P., 1969. Collaborative study on method for detection of aflatoxin
B1 in green coffee beans. J. AOAC, 52: 1300–3
Lerici, C.R. and M.C. Nicoli, 1990. II Caffe: aspetti chimici, technologici e
qualitia della bevanda. Rivista della Societa Italtiana di Scienze degli
Alimenti, 19: 1–16
Lillehoj, E.B. and M.S. Zuber, 1975. Aflatoxin problem in corn and
possible solution. Proc. 30th Ann. Corn Sorghum Resc. Conf., Am.
Seed Trade Assoc. Pup., 30: 230–50
Madison, B.L., W.J. Kozarek and C.P. Damo, 1976. High–pressure liquid
chromatography of caffeine in coffee. J. AOAC, 50: 1258–61.
Massini, R., M.C. Nicoli, A. Cassara and C.R. Lerici, 1990. Study on
physical and physicochemical changes of coffee beans during
roasting. J. Food Sci., 2: 123–30.
Nortowicz, V.B., L.R. Buchanan and S. Segall, 1979. Aflatoxins production
in regular and decaffeinated coffee beans. J. Food Sci., 44: 446–8
Pittia, P., D.M. Rosa and C.R. Lerici, 2001. Textural changes of coffee
beans as affected by roasting conditions. Lebensm–Wiss. U–Technol,
34: 168–75
Scott, P.M., 1968. Note on analysis of aflatoxins in green coffee. J. AOAC,
51: 600–3
Scott, P.M., 1969. Analysis of cocoa beans for aflatoxins. J. AOAC, 52: 72–
4
Sievetz, N. and N. Desrosier, 1979. Coffee Technology. Westport, Conn:
AVI Publ. Co.
Soliman, K.M., 2002 Inciden, Level and behavior of aflatoxins during
coffee bean roasting and decaffeination. J. Agric. Food Chem., 50:
7477–81
Soliman, K.M., 2004. Improving method for the detection of aflatoxins in
green coffee beans. J. Egypt Soc. Toxicol., 30: 153–61
Willson, K.C., 1999. CABI publishing. CAB Intrnational, Wallingford,
Oxon OX10 8DE, UK.
G1
G2
Fig. 12. HPLC chromatogram of aflatoxins spiked
roasted coffee beans
extracted by the improved
procedure
B1
B2
Unidentif ied
G1
Unidentified
Unidentif ied
G2
eliminated (Fig. 10 & 11). The HPLC chromatogram of
aflatoxins in roasted coffee bean samples showed a lot of
interference peaks were appeared especially in the area of
AFG1 (Fig. 12). This result show that although the improved
procedure coupled with HPLC resolve the problems of the
(Received 01 November 2004; Accepted 14 December 2004)
10