Chapter 2
Use of Galvanic Cell Voltages To Clock the Progress of Maillard Reactions in Real Time Downloaded by COLUMBIA UNIV on December 12, 2016 | http://pubs.acs.org Publication Date (Web): November 18, 2016 | doi: 10.1021/bk-2016-1237.ch002
George P. Rizzi* Department of Chemistry, Miami University (Middletown Campus), Middletown, Ohio 45042 *E-mail:
[email protected]
Electrochemical aspects of the Maillard reaction provide a means to clock its progress. Reaction intermediates define a galvanic cell whose varying potential measured between Pt and reference electrodes closely parallel the time course of the reaction and provide a useful non-invasive way to follow its progress.
Introduction For many years, the Maillard reaction (MR) has been extensively studied because of its relevance to food flavor and the biochemistry of cell ageing. In its basic form, the “reaction” begins with a simple amine/carbonyl/minus water condensation between reducing sugars and amino acids or proteins. These initially formed imines (Schiff bases in equilibrium with cyclic N-substituted glycosylamines) do not accumulate in foods or in model systems but instead undergo facile isomerization to the relatively more stable and in some cases isolable Amadori compounds, typically N-substituted 1-amino-1-deoxy-2-ketoses. At elevated temperatures, Amadori compounds undergo acid catalyzed β-elimination of amine substituents to form a pair of α-dicarbonyl sugar derivatives with intact carbon skeletons known as deoxyosones. The highly reactive deoxyosone intermediates are referred to individually as 1-deoxyosones (1-DEO) and 3-deoxyosones (3-DEO). Historically the deoxyosones were collectively called “reductones” and recognized as the cause of chemical reducing properties observed in model Maillard reactions. In fact, only 1-DEO belongs to the reductone class defined by its containing of the readily oxidizable-CO-CH(OH)=C(OH)- moiety. Early Maillard literature © 2016 American Chemical Society Granvogl et al.; Browned Flavors: Analysis, Formation, and Physiology ACS Symposium Series; American Chemical Society: Washington, DC, 2016.
Downloaded by COLUMBIA UNIV on December 12, 2016 | http://pubs.acs.org Publication Date (Web): November 18, 2016 | doi: 10.1021/bk-2016-1237.ch002
claimed that the oxidized form of 1-DEO, i.e. “dehydroreductones” or tricarbonyls like -CO-CO-CO- were solely responsible for Strecker degradation of amino acids (1). Now it is well recognized that various 1,2-dicarbonyls including 3-DEOs can produce this reaction (2). In spite of many related investigations, true dehydroreductones have not yet been isolated from foods or model reactions. By analogy with the hydroquinone/quinone systems, the redox behavior of ene-diol reductones is naturally associated with production of an electrochemical potential. Thus, oxidation potentials (Eo vs. normal H electrode for RH2 Þ R + 2H+ + 2e-) have been observed for ascorbic acid, reductic acid and triosereductone [the simplest reductone] at -330, -420 and -220 mV, respectively (3). Electrochemical activity was first reported in Maillard reactions in 2003 (4). Negative potentials were observed in model sugar/amino acid reactions indicative of chemically reducing species. At that time, it was concluded that the reducing species observed were reductones based on similar electrochemical behavior of an authentic Amadori compound known to be the penultimate precursor of reductones. In the 2003 study, samples were removed from heated reaction mixtures at various times and rapidly cooled to room temperature before electrochemical measurements could be made. In the following study, a less-invasive approach was used employing heat resistant electrodes, which permitted continuous voltage measurements in situ without disturbing reactions in progress.
Material and Methods Chemicals β-Alanine, sugars, ascorbic acid and inorganic reagents were high purity analytical grade materials obtained from commercial suppliers. The Amadori compound N-(1-deoxy-D-fructos-1-yl)piperidine was synthesized in two steps via a published procedure (5). The phosphate buffer was prepared by adding 0.1M aqueous KH2PO4 to 0.1M K2HPO4 to obtain a nominal 0.1M pH 7 solution. Galvanic Cell Construction and Operation A two electrode cell consisted of a platinum wire (anode) and a silver/silver chloride (cathode) reference electrode. The electrodes were situated ca. 3 cm apart in a 4-necked 100 mL round-bottomed flask equipped with a thermometer, a reflux condenser and provided with external magnetic stirring. Heating was provided by an external thermostatically controlled oil bath. The reference electrode employed was an Accumet Model 13-620-53 Ag/AgCl (saturated KCl also saturated with AgCl) electrode designed to operate up to 110 °C [Fisher Scientific Co., New York, NY]. Cell voltage was measured via the high impedance mV input of a Corning Model 240 pH meter and time dependant voltage changes were followed on a Kipp and Zonen type BD 41 strip chart recorder (Kipp & Zonen, Delft, Holland), full scale set at 500 mV. Recorder was calibrated with a laboratory grade potentiometer and a high impedance voltmeter. In this system, test substances 16 Granvogl et al.; Browned Flavors: Analysis, Formation, and Physiology ACS Symposium Series; American Chemical Society: Washington, DC, 2016.
Downloaded by COLUMBIA UNIV on December 12, 2016 | http://pubs.acs.org Publication Date (Web): November 18, 2016 | doi: 10.1021/bk-2016-1237.ch002
capable of reducing silver ions registered negative voltages, for example – 34 mV for 0.0014 M ascorbic acid in pH 7.12 phosphate buffer. Conversely, an oxidizing medium like 0.01 M K4Fe(CN)6 plus 0.05 M K3Fe(CN)6 at ambient pH in water produced a positive voltage, i.e. + 300 mV. Maillard reactions were generally conducted by dissolving β-alanine and sugars in 40 mL of 0.1 M pH 7 phosphate buffer to afford initial concentrations of 0.03 M and 0.09 M, respectively. βAlanine was employed instead of α-alanine to minimize competitive reactions resulting from the Strecker degradation. Following preliminary trials to establish convenient reaction rates, final experiments were done at either ca. 70 °C or at reflux temperature, ca. 100 °C. For 70 °C reactions, stirred aqueous mixtures initially at ca. 22 °C were immersed at t=0 min into an oil bath maintained at 70±5 °C. By t=15 min, the reaction temperature had risen and remained constant at ca. 70±2 °C and voltage measurements were started. A control experiment with buffer alone indicated complete reference voltage stabilization after t=15 min, [temp. coefficient for the Ag/AgCl electrode was ca. – 1.01 mV/°C]. For reactions run at reflux temperature, an oil bath at 110 °C was used and voltage measurements were also commenced at t=15 min.
Results and Discussion Galvanic electrode potentials developed during Maillard reactions were measured to follow the progress of model reactions versus time. Voltages produced in situ between platinum and Ag/AgCl reference electrodes provided non-invasive direct evidence for redox activity during the reaction. The use of a special high temperature, reference electrode permitted continuous voltage measurements at 70-100 °C. Model reactions generally consisted of β-alanine at ca. 0.03 M and various reducing sugars at 0.09 M in 0.1 M potassium phosphate buffer at pH 7. Reactions were run to estimate initial performances of different sugars and terminated short of completion at 60 min. Also, voltage measurements were delayed 15 min from the start of each reaction to permit complete thermal equilibration of the reference electrode. In all cases, smooth increases in voltage were observed in the 15-60 min intervals and for each sugar the data from three independent runs were averaged and presented with standard error in Figures 1-3. Reasons for variability in data between runs are not clear, but relatively slow reference electrode accommodation may have been the cause. Clear evidence was obtained for redox activity in reactions of pentoses with β-alanine at 72 °C. For D-xylose (Figure 1), the voltage reached - 80 mV after 60 min compared to virtually no voltage change in buffer alone. Some redox activity was also seen with D-xylose alone, but this was expected based on the known catalytic activity of phosphate ion on reducing sugars (6). Phosphate ions and other polyatomic anions are believed to facilitate the dehydration of aldehydo sugars by an addition/elimination mechanism leading directly to deoxyosones. Similar qualitative results were obtained with D-ribose and D-arabinose at 72 °C (Figures 2 and 3), however, differences in reactivity were 17 Granvogl et al.; Browned Flavors: Analysis, Formation, and Physiology ACS Symposium Series; American Chemical Society: Washington, DC, 2016.
Downloaded by COLUMBIA UNIV on December 12, 2016 | http://pubs.acs.org Publication Date (Web): November 18, 2016 | doi: 10.1021/bk-2016-1237.ch002
apparent in terms of voltage. In Maillard reactions with β-alanine both D-ribose and D-arabinose exhibited relatively higher redox potentials of – 140 mV and 170 mV, respectively, after 60 min possibly a result of differing reaction rates and concentrations of reaction intermediates. The reaction rate of β-alanine, and D-glucose proved to be too slow to measure at 72 °C. However, at 100 °C a smooth increase in voltage took place reaching a value of ca. – 180 mV in 60 min. Since Amadori compounds are well-known penultimate precursors of deoxyosones, i.e. reductones, it was of interest to examine the reaction of a glucose based Amadori compound for redox activity. Accordingly, the behavior of N(1-deoxy-D-fructos-1-yl)piperidine was investigated in 0.1 M pH 7 phosphate buffer at 100 °C. Preliminary experiments indicated a much greater activity for the Amadori compound compared with mixtures of β-alanine and D-glucose. Ultimately, a relatively low initial Amadori concentration of 0.011 M was employed to demonstrate redox activity (Figure 4). After 60 min at 100 °C, a large negative voltage ( ca. - 300 mV) was observed, which was consistent with significant 1-deoxyosone production.
Figure 1. Redox potentials for D-xylose/β-alanine reactions in phosphate buffer at 72 °C.
18 Granvogl et al.; Browned Flavors: Analysis, Formation, and Physiology ACS Symposium Series; American Chemical Society: Washington, DC, 2016.
Downloaded by COLUMBIA UNIV on December 12, 2016 | http://pubs.acs.org Publication Date (Web): November 18, 2016 | doi: 10.1021/bk-2016-1237.ch002
Figure 2. Redox potentials for D-ribose/β-alanine reactions in phosphate buffer at 72 °C.
The existence of a measurable electrical potential during the Maillard reaction provides evidence for redox equilibria likely resulting from the presence of reductones like 1-DEOs. Apart from electrochemical inferences, the highly reactive 1-DEOs have yet to be isolated and/or directly quantified in Maillard reactions. The reactive nature of 1-DEOs have been established by the behavior of separately synthesized examples and rationalized in terms of well defined fragmentation products (7). Less is known of the tricarbonyl redox partner of 1-DEO. Simple tricarbonyls like alloxan, ninhydrin, and dehydroascorbic acid are well-known and recognized as active Strecker degradation reactants. In the carbohydrate series, tricarbonyls have been mentioned occasionally as possible reaction intermediates. In one instance, the reactive species was reportedly synthesized and characterized. In 1963, Hodge et al. synthesized the hexotriulose derived from a 6-deoxy-2,3-hexodiulose via oxidation and characterized the product via its tris-p-nitrophenylhydrazone and quinoxaline derivatives (8).
19 Granvogl et al.; Browned Flavors: Analysis, Formation, and Physiology ACS Symposium Series; American Chemical Society: Washington, DC, 2016.
Downloaded by COLUMBIA UNIV on December 12, 2016 | http://pubs.acs.org Publication Date (Web): November 18, 2016 | doi: 10.1021/bk-2016-1237.ch002
Figure 3. Redox potentials for D-arabinose/β-alanine reactions in phosphate buffer at 72 °C
Figure 4. Redox potential for a reaction of N(1-deoxy-D-fructos-1-yl)piperidine in phosphate buffer at 100 °C. 20 Granvogl et al.; Browned Flavors: Analysis, Formation, and Physiology ACS Symposium Series; American Chemical Society: Washington, DC, 2016.
Downloaded by COLUMBIA UNIV on December 12, 2016 | http://pubs.acs.org Publication Date (Web): November 18, 2016 | doi: 10.1021/bk-2016-1237.ch002
Being chemically similar to ninhydrin, the triulose was obtained in the form of its stable, crystalline monohydrate. Detection of electrical activity in the MR suggests the presence of tricarbonyls in equilibrium (with 1-DEOs), although the increasing negative potential with time is consistent with diminishing concentration of the oxidized species (3). At this point, it is hypothesized that the low levels of oxidized species are due to their relatively higher reactivity compared to reduced counterparts and the fact that 1-DEOs are being continually generated from sugars via a non-oxidative process. In summary, Galvanic potential measurements are shown to be a useful non-invasive tool for measuring the progress of Maillard reactions in real time. The simple, inexpensive method could find application in quality control of fruit extracts or liquid beverages made from coffee or cocoa. More fundamentally, the results provided indirect evidence for the dehydroreductones once believed to be key Maillard reaction intermediates. If indeed the dehydroreductones are pivotal Maillard reactants, then future research might advantageously be directed toward their control or prevention.
References 1. 2. 3. 4. 5.
6. 7. 8.
Hodge, J. E. Chemistry of browning reactions in model systems. J. Agric. Food Chem. 1953, 1, 928–943. Rizzi, G. P. The Strecker degradation of amino acids: newer avenues for flavor formation. Food Rev. Int. 2008, 24, 416–435. Marthaler, M.; Schellenberg, M. U.S. Patent 3,620,744A, 1971. Rizzi, G. P. Electrochemical study of the Maillard reaction. J. Agric. Food Chem. 2003, 51, 1728–1731. Hodge, J. E.; Rist, C. E. The Amadori rearrangement under new conditions and its significance for non-enzymatic browning. J. Am. Chem. Soc. 1953, 75, 316–322. Rizzi, G. P. Role of phosphate and carboxylate ions in Maillard browning. J. Agric. Food Chem. 2004, 52, 953–957. Smuda, M.; Glomb, M. A. Fragmentation pathways during Maillard-induced carbohydrate degradation. J. Agric. Food Chem. 2013, 61, 10198–10208. Reynolds, T. M. Chemistry of non-enzymic browning II. Adv. Food Res. 1965, 14, 182.
21 Granvogl et al.; Browned Flavors: Analysis, Formation, and Physiology ACS Symposium Series; American Chemical Society: Washington, DC, 2016.
