Antioxidant activity of coffee brews

Eur Food Res Technol (2006) 223: 469 474 DOI 10.1007/s00217-005-0226-4 ORIGINAL PAPER Bettina Cämmerer Lothar W. Kroh Antioxidant activity of coffee brews Received: 11 October 2005 / Revised: 29 November 2005 / Accepted: 1 December 2005 / Published online: 1 February 2006 C Springer-Verlag 2005 Abstract The total antioxidant activity of coffee beverages was measured with stabilized radical EPR spectroscopy. Depending on which stabilized radical is used, Fremy s salt (potassium nitrosodisulphonate) or 2,2,6,6- tetramethyl-1-piperidin-1-oxyl (TEMPO) values can differ significantly. For the determination of antioxidant activity of Maillard reaction products in coffee, TEMPO appears to be the better radical marker. Thus the contribution of both main antioxidant active compounds (polyphenols, melanoidins) whose ratio varies with roasting conditions could be estimated. During storage experiments of coffees brews changes in antioxidant action are found to be time dependent. The content of chlorogenic acids increased significantly at higher storage temperatures, probably caused by a release from polymer structures. Additional antioxidant capacity of coffee melanoidins seems to be strongly influenced by atmospheric oxygen. The higher roasted sample is less vulnerable than medium or light roasted coffee. Investigations with model systems showed that among all coffee constituents the carbohydrates are mainly responsible for the formation of oxygen scavenging substances. Keywords Coffee. Antioxidant activity. EPR spectroscopy. Melanoidins Introduction The antioxidant activity of coffee beans and coffee beverages has been known for a long time and has been assessed using different detection methods [1 7]. Several studies indicated that a high content of polyphenols in coffee plays an important role in its strong antioxidant action [3, 7, 8]. The concentration of highly active polyphenols in green beans is B. Cämmerer L. W. Kroh ( ) Technical University Berlin, Institute of Food Chemistry, Gustav-Meyer-Allee 25, 13355 Berlin, Germany e-mail: Lothar.Kroh@TU-Berlin.de Tel.: +49-30-31472701 Fax: +49-30-3147-2585 influenced by the species and its origin, in coffee beverages it depends on the brewing procedure [8]. During roasting phenolic compounds are partially degraded and/or bound to polymer structures depending on roasting conditions. A positive but nonlinear relationship was found for the amount of chlorogenic acids that remained after roasting and antioxidant activity of beans [6]. Melanoidins formed during roasting via Maillard reaction are also known to be antioxidants [2, 9]. Mechanisms of antioxidant action of melanoidins are mainly attributed to reactions involving radicals such as the scavenging of hydroxyl radicals, the ability to break radical chain mechanisms or by oxygen scavenging. Although melanoidins contribute to overall coffee constituents about 25%, their contribution to total antioxidant activity of coffee has been under investigation for a very short time [10]. The aim of this study was to evaluate the total in vitro antioxidant activity of coffee beverages that could be attributed to radical involving reactions. Photometrical methods like the DPPH method, the ABTS + and the DMPD assay [2, 4 6] are widely used methodologies for the investigation of foods, but measurement of radical scavenging properties is also possible by solid phase electron paramagnetic resonance spectroscopy (EPR) [11 13]. In food systems, especially with low water activity fairly stable radicals are present, which can also be detected in solution after extraction with several solvents. Their chemical structures and formation pathways are not always clear. In wine they were attributed to phenolic tannins while in coffee no association with the quinide fraction but with carbohydrates was found [11]. In addition, with the EPR method involving stabilized radicals, the degree of antioxidant effectiveness can be determined by the ability of a solution to scavenge a synthetic free radical species like Fremy s salt (potassium nitrosodisulphonate) or 2,2,6,6-tetramethyl-1- piperidin-1-oxyl (TEMPO). Up to now only the influence of roasting condition on total antioxidant action of coffee has been investigated. Depending on the method used, an increase with increasing roasting degree or an optimum antioxidant action for a medium roasted coffee was found (e.g. [4, 5, 7 9]).

470 Therefore, in this work it was intended to estimate the contribution of the main antioxidant active compounds in coffee polyphenols and melanoidins their ratio varies with roasting conditions because their content is contrary influenced by heat processing. Furthermore, the influence of oxygen influx on antioxidant stability of coffee beverages was determined. Materials and methods Materials Chemicals Chlorogenic acids (CGAs) were from Roth (Karlsruhe, Germany), gallic acid from Serva (Heidelberg, Germany), (+)-arabinoglactans from Fluka (Buchs, Switzerland), TEMPO (2,2,6,6-tetramethyl-1-piperidin-1-oxyl) and Fremy s salt (potassium nitrosodisulphonate) as well as Trolox (6-hydroxy-2,5,7,8-tetramethyl-chroman- 2-carboxylic acid) were purchased from Sigma-Aldrich (Steinheim, Germany). The plant proteins isolated from field beans were provided by the Institute of Nutritional Science, University of Jena. All other reagents and solvents were of analytical grade quality, solvents for HPLC investigation of high-performance liquid chromatography grade. Products Coffee beans were kindly supplied by the Nestlé Research Centre (Lausanne, CH) in three different roasting degrees and the original green beans (blend of 80% Arabica and 20% Robusta): roasting degree light (RD 110, roasting loss 14.5%), medium (RD 85, roasting loss 16.2%) and dark (RD 60, roasting loss 18.9%). Roasting time for all three levels was 6 min and water quenching was 7%. Methods Preparation of coffee brews Coffee brew was made with a commercial available automatic drip coffee maker. Fifty grams of grinded coffee beans and 1,100 ml water were used to produce 800 ml of coffee brew. It was collected in a jug standing on a hot plate, which kept it at temperatures between 68 and 70 C for 8 h. For another investigation the coffee brew was put into a thermos flask in which the brew was cooled down to about 40 Cin8h. Investigation of changes in antioxidant activity on standing Coffee brews were put into a flask and covered with parafilm M. For investigation of oxygen influence the flasks was equipped with a capillary and nitrogen or air was bubbled through the solution during investigation. The brews were allowed to stand on a heating plate for 8 h at 70 C. Every 2 h a sample was taken for HPLC (CGAs) as well as EPR spectroscopic (TEMPO and Fremy s salt) investigations and diluted to an appropriate concentration. Production of coffee model systems The coffee model system consists of plant proteins (1.3%), arabinogalactans (6.7%), sucrose (1.0%), reducing carbohydrates (0.1%), free amino acids (0.1%) and chlorogenic acids (0.6%). They were heated for 4 and 10 min at 180 C and sealed in ampoules. For model systems II, III and IV the relevant component was substituted by the same amount of sand. After heating, the model systems were diluted with water, content of CGA was measured with HPLC and antioxidant activity was determined by EPR spectroscopy. Determination of polyphenol content Polyphenol content was determined by using both the spectrophotometric (Folin-Ciocalteau reagent) and HPLC method. Phenolic compounds in coffee beans, coffee brews and model systems were determined by HPLC-diode array detection (DAD)-electrochemical detection (ECD) methods, according to Rösch et al. [14]. An external standard method was used for quantification within the range of 20 60 mg/l. For CGA s content amount of 3-, 4- and 5-caffeoylquinic acid was summarized. The total polyphenolic compounds were measured according to the method of Singleton [15] and calculated using gallic acid as standard within the range of 100 mg/l. Coffee brews were used as described above, grinded coffee beans were extracted with 80% ethanol. Determination of antioxidant action by ABTS assay The 2,2 -azinobis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS) method on coffee brews was performed according to Rohn et al. [16]. For measurement the coffee brew was diluted with water (1:50). Total antioxidant activity (TAA) was expressed as mmol Trolox/L coffee brew or mmol Trolox/g coffee beans. Determination of antioxidant action by EPR-spectroscopy For solid phase EPR spectroscopy the same amount of ground beans of each coffee was put into a quartz tube (5 mm ID); The sample height was adjusted so as to fit active cavity. For calibration the EMS 104 calibration sample 934201 was used. Results were obtained by an EMS 104 EPR analyser (Bruker, Germany) at 19.9 mw microwave power. Sweep width was 100 G and modulation amplitude was 11.3 G.

471 The EPR spectroscopic measurements with stabilized radicals were performed with TEMPO and Fremy s salt. For investigation with TEMPO the coffee brew was used as prepared. Aliquots (200 µl) were allowed to react with an equal volume of an aqueous 1mmol solution of TEMPO. EPR spectrum was obtained after 30 min, by which the reaction was complete. For measuring the antioxidant activity with Fremy s salt a procedure was used described in [9]. The reaction time was 30 min and the coffee beverages were diluted 100-fold with water. Spectra were measured on a Miniscope MS 100 spectrometer (Magnetech, Berlin, Germany). Microwave power and modulation amplitude were set at 10 db and 1500 mg, respectively, centre field 3397 G. Statistical treatment All of the analyses were performed at least in duplicate. Results and discussion The majority of radicals detected in coffee were reported to be formed during roasting process [12]. Comparing the green and roasted coffee samples via solid phase EPR spectroscopy it was demonstrated that the amount of radicals increases with increasing roasting degree (Fig. 1) which exhibit an involvement of Maillard reaction. It can be assumed that pyrazinumradical cations ( Crosspy ) formed in the early stage of Maillard reaction [17] and especially melanoidins as final products are responsible for the detected radical content of coffee beans. For coffee, Goodman [12] found a single line typical of charred polysaccharides with carbon and oxygen centred free radicals. The content of radicals in ground beans does not give clear information about possible in vitro antioxidant activity of coffee beverages or of their action in other food systems because these free radicals can act prooxidative as well as antioxidative. For this reason total antioxidant activity (TAA) of coffee beverages made from different roasted coffees was evaluated. Two assays were used based on the ability of a substance or food system to scavenge free radicals (ABTS assay and EPR experiments). As different assay conditions result in scavenging of different radical species the values obtained for TAA are not directly comparable, but they are expected to show the same tendency. Results of ABTS assay for light, medium and dark roasted coffee indicate a significant TAA loss of about 40% with increasing roasting degree (Fig. 2). Additionally, in accordance with literature [1] a rise of TAA from the green coffee beans to the light roasting degree was detected. It is due to a release of highly active low molecular weight phenols from the green coffee constituents by moderate heating [15, 18]. During further heat treatment the content of phenolic compounds in coffee decreases from about 12% (RD 110) to about 2% (RD 60) [19]. It is known that during the roasting process, increasing amounts of polyphenols undergo polymerization or autoxidation reactions [20, 21], which probably leads to the formation of antioxidant less active substances. Also involvement of phenols or their degradation products into melanoidin structures or linkage to polymeric material are discussed [10, 18, 22]. Antioxidant action of such complex compounds is still unknown. In coffee model systems there is a positive but nonlinear relationship between antioxidant action and loss of CGAs during roasting [3]. The higher the roasting degree the lower is the content of CGAs. In light roasted coffee samples (RD 110) the amount of CGAs detected contributes to TAA measured by ABTS assay to 15% in contrast to dark samples (RD 60) with only 5% (Fig. 2). Fremy s salt was used as a stabilized radical for EPR spectroscopic investigation of antioxidant action. The results obtained for TAA correspond with that of the ABTS method. Antioxidant action decreases with increasing roasting degree (Fig. 3). Also for EPR measurements a correlation was found for antioxidant action and content of CGAs, but the contribution of these polyphenols to TAA Fig. 1 Radical content of ground coffee beans with different roasting degrees measured with solid phase EPR spectroscopy Fig. 2 Antioxidant activity (ABTS assay) of coffee brews made from coffee at different roasting degrees in comparison with antioxidant activity caused by the amount of CGA detected in these brews

472 Fig. 3 Comparison of antioxidant action in coffee brews at different roasting degrees measured as ability to scavenge Fremy s salt, calculated via a calibration curve from the content of chlorogenic acids detected in the brews (HPLC), expressed as reducing ability (measured by Folin Ciocalteu assay) Fig. 5 Antioxidant activity of coffee model systems determined by EPR-spectroscopy a ability to scavenge TEMPO, b ability to scavenge Fremy s salt; I whole model, II model without CGA, III model without protein, IV model without polysaccharides Fig. 4 Antioxidant activity (expressed as the ability to scavenge the stabilized radical TEMPO) of coffee brews and extracted coffee beans of different roasting degrees calculated via a calibration curve is much higher than was assessed by ABTS method and ranges from about 50 to 65%. Additional we detected the total amount of polyphenols with Folin Ciocalteau method, as in LDL-oxidation reactions the lag time could not always be correlated with a single phenolic compound but is the result of the reducing ability of several polyphenol structures. As expected in coffee the amount of compounds with polyphenol structure detected spectrophotometrically is much higher than the concentration of CGAs alone. According to literature these values obtained by total phenols assay represents antioxidant capacity expressed as reducing capacity. But even these values did not match antioxidant action measured by EPR spectroscopy completely (Fig. 3). Only in the light roasted coffee phenolic compounds are proved to be the main antioxidant active compounds in EPR investigations with Fremy s salt. For medium and dark roasting degree there is a lack of about 15 20% between antioxidant activity measured and calculated (Fig. 3). As Maillard reaction based formation of antioxidant compounds increases with increasing roasting it can be assumed that these deviations indicate the contribution of substances formed during Maillard reaction, and it is known that the antioxidant properties of Maillard reaction products are not limited to electron transfer reactions. In contrast to the results obtained with the stabilized radical Fremy s salt, the antioxidant activity of the coffee beverages measured by EPR spectroscopy using TEMPO as the stabilized radical slightly increased with higher roasting degree (Fig. 4). More significant differences in antioxidant values were detected for a simple extract of appropriate coffee beans. Based on the fact that TEMPO is not or only negligible sensitive to polyphenols, the antioxidant activity measured can therefore not be influenced by content of CGAs present in coffee or coffee brew. Also the known formation of H 2 O 2 from coffee phenols [23, 24] and its possible reaction with the stabilized radical has to be ruled out, as in model systems no degradation of TEMPO radical in the presence of up to 500 µmol H 2 O 2 was detectable. Explanation for the rise of antioxidant action with increasing roasting degree is given by the formation of active compounds during the Maillard reaction. It can be assumed that Maillard reaction products and/or melanoidins contribute to a certain extent to values measured by stabilized radical TEMPO. The reason for the adjustment of antioxidant

473 Fig. 6 Changes in the antioxidant properties and content of CGAs during standing on a heating plate (70 C) in coffee brew of beans with medium roasting degree (RD 85) levels of different coffee brews, in contrast to cold water extracts, can be found in the further moderate heat treatment during coffee making resulting in the formation of antioxidant browning products from intermediate Maillard reaction products formed during roasting processes (Fig. 4). Investigations on simple coffee substituting model systems clearly show that the antioxidant activity detected with the stabilized radical Fremy s salt is mainly due to CGAs (Fig. 5b). Measurements with TEMPO revealed that the model system without CGAs (Fig. 5a-II) possesses the same antioxidant properties similar to the complete model system (Fig. 5a-I). Furthermore it was reported that the polysaccharides play the most important role in the development of antioxidant active compounds during heat treatment, bacause in the model system without polysaccharides (Fig. 5a-IV) a significant lower amount of active compounds were formed. These findings are in good agreement with the investigations of Redgewell and co-workers [25] who suggested a linkage of Maillard reaction products to polysaccharides during roasting. As result of the investigations with EPR spectroscopy, the stabilized radical TEMPO appears to be a better radical than Fremy s salt for the determination of antioxidant active Maillard reaction products in presence of polyphenols. Depending on the storage conditions like temperature and influx of oxygen the antioxidant activity of coffee brews changes with time. Investigations on coffee beverages held for several hours at a temperatures of about 70 C showed a significant increase of free CGAs content (Fig. 6). At lower temperatures (thermos flask, 40 C) only negligible changes were detected (not shown). Probably this finding could be explained as due to the release of CGAs from non covalently linked polymeric skeletons [10]. For the antioxidant action for coffee beverages, treated in this way depending on stabilized radical used, contrary trends were obtained (Fig. 6). The antioxidant activity measured with Fremy s salt increased with standing time because of the higher CGAs concentration, but the antioxidant values calculated from the amount of CGAs released correspond with the detected rise of antioxidant action only for dark roasted coffee with the lowest increase in phenolics. For coffee with roasting degree RD 110 and RD 85 the antioxidant activity measured with Fremy s salt did not increase as much as expected from the released amount of CGAs. It seems that some substances present in fresh coffee brews normally interact with Fremy s salt were inactivated on standing. Taking into consideration that the concentration of CGAs increased during standing it has to be expected that it is a matter of melanoidins. Measurements with TEMPO clarify this assumption. Antioxidant activity measured with TEMPO decreased with increasing storage time (Fig. 6). As TEMPO does not react with phenolic compounds as mentioned before, the released CGAs do not have an influence on the antioxidant activity detected. Preferred reaction partners of TEMPO are Maillard reaction products, and the reduction of antioxidant activity on standing could be explained by further reactions of the highly reactive low molecular weight melanoidin fractions to less active melanoidins. One example could be the reaction of several coffee flavour compounds with Maillard reaction based radicals to covalent bound substances without antioxidant properties [26]. Investigations under influx of atmospheric oxygen show a significant loss of antioxidant activity especially for dark (RD 60) roasted coffee (Fig. 7) in contrast to investigations under nitrogen atmosphere. The higher roasted sample is less vulnerable to oxidation reactions than medium or light roasted coffee (not shown), probably caused by a higher content or more active Maillard reaction products resulting in a higher capacity to scavenge free oxygen molecules. It is evident that antioxidant active structures were consumed by oxidation reactions and it seems that the activity cannot be restored by redox reactions. Investigating the sensitivity of coffee model systems against oxygen it can be proved that among coffee constituents carbohydrates are the most important for formation of compounds responsible for oxygen scavenging properties and its amount increases with increasing roasting time (Fig. 5). CGAs are fairly stable against atmospheric oxygen and their observed release during standing of coffee beverages did not show

474 Fig. 7 Influence of oxygen on antioxidant properties of coffee brew from high roasted coffee (RD 60). Results expressed as the ability to scavenge the stabilized radicals Fremy s salt (a) or TEMPO (b) an influence on oxygen-induced changes of antioxidant activity during standing. Acknowledgments Nestlé Research Centre Lousanne (Switzerland) is thanked for providing coffee samples. 3. Charurin P, Ames JM, Del Castillo MD (2002) J Agric Food Chem 50:3751 3756 4. Del Castillo MD, Ames JM, Gordon MH (2002) J Agric Food Chem 50:3698 3703 5. Steinhart H, Luger A, Piost J (2001) Proceedings of 19th International Scientific Colloquium on Coffee, Trieste 14.-18.5.2001 6. Anese M, Nicoli MC (2003) J Agric Food Chem 51:942 946 7. Richelle M, Tavazzi I, Offord E (2001) J Agric Food Chem 49:3438 3442 8. Sánchez-González I, Jiménez-Escrig A, Saura-Calixto F (2005) Food Chem 90:133 139 9. Nicoli MC, Anese M, Manzocco L, Lerici CR (1997) Lebensm Wiss u Technol 30:292 298 10. Delgado-Andrade C, Morales FJ (2005) J Agric Food Chem 53:1403 1407 11. Gonis J, Hewitt DG, Troup G, Hutton DR, Hunter CR (1995) Free Radic Res 23:393 399 12. Goodman BA, Glidewell SM, Deighton N, Morrice AE (1994) Food Chem 51:399 403 13. Müller J, Rösch D, Kroh LW (2004) Elektronenspinresonanzspektroskopie. In Baltes W, Kroh LW (eds) B. Behrs Verlag, Hamburg, Schnellmethoden zur Beurteilung von Lebensmitteln und ihren Rohstoffen 14. Rösch D, Bergmann M, Knorr D, Kroh LW (2003) J Agric Food Chem 51:4233 4239 15. Singleton VL, Orthofer R, Lamucia-Raventos RM (1999) Methods Enzymol 299:152 178 16. Rohn S, Rawel HM, Kroll J (2004) J Agric Food Chem 52:4725 4729 17. Hofmann T, Bors W, Stettmaier K (1999) J Agric Food Chem 47:391 396 18. Montavon P, Mauron A-F, Duruz E (2003) J Agric Food Chem 51:2335 2343 19. Values provided by Nestlé 20. Nunes FM, Coimbra MA (2001) J Agric Food Chem 49:1773 1782 21. Montavon P, Duruz E, Rumo G, Pratz G (2003) J Agric Food Chem 51:2328 2334 22. Wen X, Enokizo A, Hattori H, Kobayashi S, Murata M, Homma S (2005) J Agric Food Chem 53:2684 2689 23. Akagawa M. Shigemitsu T, Suyama K (2003) Biosci Biotechnol Biochem 67:2632 2640 24. Hiramoto K, Kida T, Kikugawa K (2002) Biol Pharm Bull 25:1467 1471 25. Redgwell RJ, Trovato V, Curti D, Fischer M (2002) Carbohydr Res 337:421 431 26. Hofmann T, Schieberle P (2002) J Agric Food Chem 50:319 326 References 1. Daglia M, Papetti A, Gregotti C, Bertè F, Gazzani G (2000) J Agric Food Chem 48:1449 1454 2. Borelli CR, Visconti A, Mennella C, Anese M, Fogliano V (2002) J Agric Food Chem 50:6527 6533