The Use of Fatty Acid Profile as a Potential Marker for Brazilian Coffee ( Coffea arabica L . ) for Corn Adulteration

A composição de ésteres metílicos de ácidos graxos (EMAGs) das variedades de café (Coffea arabica L.) Catuai, Catucaí, Bourbom, Mundo Novo, Rubí e Topázio conhecidas por produzirem bebidas de qualidade intermediária, ótima, ótima, intermediária, intermediária e inferior, respectivamente, foi determinada pela primeira vez. A (%) de área média dos seguintes ácidos das seis variedades foi: palmítico (38,2), esteárico (8,3), oléico (8,7), linoléico (38,5), linolênico (1,6) e araquidônico (3,6), respectivamente. O método é bastante rápido, com a completa caracterização das amostras estudadas (99%) ocorrendo em menos de 6 minutos. Enquanto esses valores podem fornecer informação para a avaliação da qualidade do café, nenhum efeito significativo (p < 0,05) da variedade de café foi encontrado na (%) de área das EMAGs. Além disso, foram comparados os EMAGs de: seis amostras de milho, seis marcas de café comerciais e uma amostra de café comercial intencionalmente adulterada com três quantidades de milho. Apesar da razão ácido linoléico/ácido esteárico encontrada nos EMAGs do café e do milho serem significativamente diferentes, não foi possível utilizar esse marcador para detectar adulteração do milho em cafés comerciais.


Introduction
Coffee (Coffea arabica, L.) quality may be lowered through hydrolysis of triacylglycerols (TAGs) resulting in release of free fatty acids (FAs) which are oxidized to produce off-flavor. 1 However, no studies have been conducted relating FAs to coffee quality despite the fact that they are known to be important flavor components. 2 Coffee quality can also be lowered by adulteration with cereals, coffee twigs, caramel, etc.This practice may be widespread in Brazil, with corn being considered the most widely used adulterant, due to its significantly lower cost.Few studies have been reported detecting corn in coffee [3][4][5] but only one study has detected corn commercial coffees using γ-tocopherols as the probe. 6ur coffee program studies the role of fatty acids in coffee flavors and also develops chromatographic methods to detect corn as an adulterant in coffee.
In this study, we report for the first time the fatty acid methyl ester (FAME) composition of the coffee (Coffea arabica L.) varieties Catuai, Catucaí, Bourbom, Mundo Novo, Rubí and Topázio known to produce beverage of intermediate, excellent, intermediate, intermediate and poor quality, respectively, was determined for the first time.The coffee FA composition found in this study was significantly Vol. 19, No. 8, 2008   different from that of corn. 7This difference opened the possibility of using FAs as indicators of coffee adulteration by corn as in the case with olive oil FAs. 8 Experimental Chemicals HPLC-grade hexane and methanol, FAME standards were purchased from Sigma-Aldrich, Inc. (St. Louis, MO).

Coffee and corn samples
The roasting, grinding, and extraction processes described below were performed in triplicate for each sample and/or variety.

Six coffee varieties
About 1 kg of six green coffee (C.arabica) bean varieties (Catuai, Catucaí, Bourbom, Mundo Novo, Rubí and Topázio) were supplied by Incofex Inc. (Viçosa, MG, Brazil).The samples were divided into roughly three equal parts.Each part was roasted at 180 o C for 10 min in a coffee roaster with three burners, ground in a coffee grinder.

Commercial coffee brands
One kg of six roasted and ground coffee (C.arabica) popular local brands was purchased from a local supermarket in Viçosa, MG, Brazil.All samples were divided into roughly three equal portions.

Intentionally adulterated coffee (C. arabica) sample
A commercial coffee brand was mixed with 5, 10 and 20% of ground roasted corn (m:m, corn:coffee).

Commercial corn
One kg of six commercial corn samples was randomly selected from a local supermarket in Viçosa, MG, Brazil.The samples were divided into roughly three equal portions, roasted and ground under the conditions the same conditions utilized for coffee.

Oil extraction
About 10 g of all the samples were extracted overnight separately with hexane in a Soxhlet extractor.After cooling, the solution was dried over anhydrous sodium sulphate, and evaporated under reduced pressure at 35 ºC to obtain oils, which were sealed under N 2 , and stored in a freezer (-5 ºC) until analysis.

Instrument reproducibility
The reproducibility of the gas chromatographs was verified by analyzing 10 times the FAMES of a randomly chosen sample.The standard deviation (SD) obtained in the area % of FAMEs was calculated.

FAMEs
Fatty acid methyl esters were analyzed by gas chromatography (GC) by placing 10 mg of oil in a 2 mL crimp cap vial then adding 1.0 mL of 0.25 mol L -1 sodium methoxide solution.The vial was sealed with a crimp cap and placed in a heating block that was maintained at 60 o C.After 30 minutes the vial was removed from the heating block allowed to cool to room temperature and 1.0 mL of 0.25 mol L -1 hydrochloric acid and 2 mL of hexane were added.The contents of the vial were agitated thoroughly for 2 min on a Vortex mixer and then allowed to separate.The top organic layer was drawn off and placed in a 2 mL GC vial and analyzed by GC under the conditions described below.
Fatty acid esters were analyzed by GC with a fused silica capillary column (30 m × 0.25 mm ID × 0.25 μm) coated with the SP 2380 stationary phase.The oven temperature was programmed ramped from 180 to 210 °C at 7 °C min -1 , from 210 to 265 °C at 30 °C min -1 and held 3 min at 265 °C.The injector was maintained at 250 °C and 1 μL injections were made with a split ratio of 100:1.Column flow rate of 1.4 mL min -1 with a split flow of 200 mL min -1 (split ratio of 142:1) and a septum purge of 4 mL min -1 .Analytes were detected by flame ionization with a detector temperature of 250 °C.Helium was used as the carrier gas at linear velocity of 35 cm s -1 .Analytes were identified by comparison to known standards.Samples were analyzed in duplicate and results averaged.Area % of the FAMEs were determined by summing the peak areas corresponding to the eight fatty acids identified, dividing each area by the total area of the FAMEs peaks and multiplying by 100.

Statistical analyses
Single-factor, mixed model analyses of variance (ANOVAs) were conducted comparing relative percentage of FAME types and FAME ratios for coffee, corn, and coffee/corn compositions.If a F-test statistic was significant at p ≤ 0.05, the FAME types and ratios were declared statistically different.
A two-factor, mixed model ANOVA was used to examine FAME ratio and coffee/corn composition interaction.If a significant F-test statistic was obtained at p ≤ 0.05 for the interaction, no FAME ratio or coffee/corn composition main effects were examined.
Differences of least squares means at p ≤ 0.05 was used as the pairwise multiple comparison procedure when a significant F-statistic was obtained from any ANOVA.A Bonferroni adjustment to the p-value was made when the number of comparisons was large.
Coffee varieties and corn brands in the FAME singlefactor ANOVA analyses were considered to be random effects whose contribution to the residual experimental error were either negligible or zero, indicating no significant differences between the coffee varieties or corn brands.
Levene's homogeneity of variance test was performed on all data to determine if transformations were needed before any ANOVAs were conducted.All analyses were performed on transformed data where necessary, but raw data means and standard deviations are presented for ease of interpretation.All analyses were performed using SAS ® PC Windows Version 9.1.3software.

FAMEs in six coffee varieties
Typical chromatogram obtained on GC analysis of coffee and corn FAMEs are presented in Figure 1.As it can be observed the GC run time was very short, i.e. 6 min.A smooth base line and good reproducibility was obtained with all the samples.The instrument functioned adequately as a small SD in the TAG area percentages was obtained when a randomly chosen coffee oil was injected ten times successively.For the FAMEs present in minor (ca.0.2%) to moderate amounts (ca.10%) the SD was about 0.1 while for the major FAMEs it was about 0.6.With the coffee samples, the SDs tended to be slightly larger but still very low and acceptable.Good reproducibility, precision and low SD made this method suitable to quantify FAs.

Potential of FAMEs as adulterant markers
3][14][15][16][17][18][19] The only chemical method able to detect corn in coffee, based on γ-tocopherol was described by us recently. 6In that study, one out of six commercial coffees examined was adulterated with 8.9% corn.Hence, using the same coffee and corn samples, we could quickly evaluate the potential of any chemical marker to detect corn in coffee.Since large differences in coffee FAMEs and corn 8 were noted, we decided to evaluate the potential of FAMEs as adulterant markers using the same samples and procedures described in our previous study. 6Several ratios were evaluated but the most consistent results were obtained with C18:1/C18:0 (linoleic acid/stearic acid).
The following steps were conducted to evaluate the potential of FAMEs as adulterants: i) determine C18:1/ C18:0 ratios in six unadulterated coffee varieties utilizing their averages as base line values; ii) determine C18:1/ C18:0 ratios in six roasted unadulterated corn utilizing their averages as base line values; iii) determine if C18:1/ C18:0 ratios increased linearly in an unadulterated roasted coffee intentionally adulterated with three corn levels (5, 10 and 20%); iv) determine C18:1/C18:0 in six commercial coffee brands, knowing that only one of the samples (sample 3) was contaminated with 8.9% corn 6 and v) quantify corn in commercial coffee samples if C18:1/C18:0 ratios differed significantly from those of unadulterated coffees.
FAMEs area % and the C18:1/C18:0 ratios for six varieties of roasted Brazilian coffee (Coffea arabica L.), six roasted Brazilian corn samples, three intentionally adulterated Brazilian coffee (Coffea arabica L.) samples and six roasted Brazilian coffee (Coffea arabica L.) brands are presented in Table 2.The average area % of C16:0, C18:0, C18:1, C18:2, C20:0 and C18:3 for six coffee varieties was: 38.2, 8.3, 8.6, 38.5, 1.6 and 3.6, respectively; while that for corn brands was 14.6, 2.2, 33.3, 47.2, 0.6 and 1.6, respectively.Similarly the C18:1/C18:0 ratios in coffee and corn were 1.0 and 15.7, respectively.Since large differences in C18:1/ C18:0 ratios were obtained between coffee and corn, the potential of using this ratio as a marker of adulteration was investigated.Hence we first evaluated if C18:1/C18:0 ratios could detect corn in an intentionally adulterated coffee adulterated with 5, 10 and 20% roasted corn.The % contamination was arbitrarily chosen and based on the supposition that it had to be sufficiently high for a financial return but at the same time could not be too high as to significantly alter coffee flavor.
The results with the intentionally adulterated samples are presented in Table 2.As expected, a linear increase in C18:1/C18:0 with % roasted corn added was obtained that could be described by the equation Y = 0.945 + 0.01214X, where Y is the C18:1/C18:0 and X is the corn % contamination level.The standard error for the intercept and concentration were 0.00982 and 0.000742, respectively, and the correlation coefficient (R 2 ) was 0.996.Hence, it appeared that C18:1/C18:0 presented potential as a marker to detect corn adulteration.We next analyzed six commercial coffee samples, knowing that only brand 3 was contaminated with 8.9% corn based on our new method based on γ-tocopherol as an adulterant marker 6 .The C18:1/ C18:0 ratios in adulterated samples were not higher than in the pure coffees.In all the cases, the ratios were lower than that of the average area % of the six coffee varieties.The lower C18:1/C18:0 ratios in commercial samples were most likely related to storage, processing, etc.
The discrepancy between the FAMEs and γ-tocopherol as an adulterant marker results can be possibly explained in the differences in the % composition of the two markers used in this study.Average percentages of α-, β-, γ-, and δ-tocopherols in six unadulterated coffee varieties were 29.0, 61.7, 3.3 and 6.0, respectively, while average α-, γand δ-tocopherol in six unadulterated corn samples were 3.6, 91.3 and 5.1, respectively.Thus, γ-tocopherol concentration values in coffee and corn were very different, i.e., 91.3 and 3.3, respectively.On the other hand, the average C18:1/C18:0 ratios in unadulterated corn vs. coffees were only 1.0 and 15.7, respectively.Although the ratios were quite large, they were not sufficiently different to detect adulteration.Based on these data, it can be concluded that γ-tocopherol is a much better marker than FAMEs to detect coffee adulterated with corn.

Conclusions
We have thus determined, for the first time, the TAG and FA compositions of six coffee varieties used in our Brazilian coffee breeding program.The method was much quicker than the methods reported in the literature with complete characterization of our samples being possible in less than 6 min.This quick method will be useful to screen a large number of samples in our coffee program.These values should provide us insights while evaluating coffee quality.In addition, we have shown that FAMEs cannot be used as adulteration markers of corn in coffee.Also, based on our FAME and tocopherol adulteration results, it appears that for a compound to serve as a marker of adulteration in coffee, a very large difference (at least by a factor of ca.30) should exist between its concentrations in the pure and adulterated samples.These conclusions should serve as guidelines for future development of coffee adulteration methods.C18:1/ C18:0 1.0 ± 0.05 15. ± .031.0 ± 0.01 1.1 ± 0.01 1.2 ± 0.01 1.0 ± 0.05 1.0 ± 0.04 0.9 ± 0.04 0.9 ± 0.01 0.9 ± 0.03 0.9 ± 0.02 *Values are means ± SD (n = 6); **Values are means ±SD (n = 3); C16:0, C18:0, C18:1, 18:2, C 20:0 and C18:3 are palmitic, stearic, oleic, linoleic, arachidic and linolenic acids, respectively; 1 Mean commercial coffee brand % within a row followed by the same letter was not significantly different based on least squares means at p < 0.05; Less than 1% of myristic and behenic acids were also detected.

Figure 1 .
Figure 1.Typical gas chromatogram obtained on analysis of FAMEs from coffee (A) and corn (B) using a fused silica capillary column (30 m × 0.25 mm ID × 0.25 μm) coated with the SP 2380 stationary phase.The oven temperature was programmed ramped from 180 to 210 °C at 7 °C min -1 , from 210 to 265 °C at 30 °C min -1 and held 3 min at 265 °C.The injector was maintained at 250 °C and 1 μL injections were made with a split ratio of 100:1.Column flow rate of 1.4 mL min -1 with a split flow of 200 mL min -1 (split ratio of 142:1) and a septum purge of 4 mL min -1 were used to separate the fatty ester isomers.Analytes were detected by flame ionization with a detector temperature of 250 °C.Helium carrier gas was used with a linear velocity of 35 cm s -1 .Peak identification a: palimitic acid (C16:0), b: stearic acid (C18:0), c: oleic (C18:1), d: linoleic acid (C18:2), e: arachidic (C20:0) and f: linolenic (C18:3) acids, respectively.