Simulated Sunlight Selectively Modifies Maillard Reaction Products in a Wide Array of Chemical Reactions

Abstract The photochemical transformation of Maillard reaction products (MRPs) under simulated sunlight into mostly unexplored photoproducts is reported herein. Non‐enzymatic glycation of amino acids leads to a heterogeneous class of intermediates with extreme chemical diversity, which is of particular relevance in processed and stored food products as well as in diabetic and age‐related protein damage. Here, three amino acids (lysine, arginine, and histidine) were reacted with ribose at 100 °C in water for ten hours. Exposing these model systems to simulated sunlight led to a fast decay of MRPs. The photodegradation of MRPs and the formation of new compounds have been studied by fluorescence spectroscopy and nontargeted (ultra)high‐resolution mass spectrometry. Photoreactions showed strong selectivity towards the degradation of electron‐rich aromatic heterocycles, such as pyrroles and pyrimidines. The data show that oxidative cleavage mechanisms dominate the formation of photoproducts. The photochemical transformations differed fundamentally from “traditional” thermal Maillard reactions and indicated a high amino acid specificity.


Figure S1
Principal component analysis of ribose-histidine FT-ICR-MS raw data. Samples were irradiated for four and eight hours in a suntester system. Additionally, control samples, which were kept under the same conditions, however, protected from light exposure, as well as freshly prepared model systems were analyzed. All experiments were carried out in two independent experiments. Each sample was further injected in triplicate measurements (total number of samples per treatment = 6). Samples and replicate injections were measured in randomized order.

Figure S2
Effect of solar irradiation on elemental compositions of ribose-lysine MRPs. Model systems were irradiated for eight hours and compared to unirradiated control samples. Irradiation experiments were performed in duplicate. Each sample then was analyzed by FT-ICR-MS in triplicate injections (N = 2 × 3). Peak intensities of all features found in irradiated samples were compared to the same features in the unirradiated control samples by Student's t-Test (n = 3): Features, which showed a significant decrease in peak intensities in both independent irradiation experiments are colored in blue. Features, which showed a significant increase or were newly formed upon irradiation are highlighted in red, respectively. (a) Volcano plot. (b) Number of molecular formulae showing significant changes in peak intensities. (c) Van Krevelen diagram of all significantly affected molecular formulae. Pie charts illustrate the reduced occurrence of nitrogenfree (CHO) MRPs in photochemical reactions. Black pie chart represents elemental compositions, which did not show a significant change in peak intensities upon irradiation.

Figure S3
Effect of solar irradiation on elemental compositions of ribose-arginine MRPs. Model systems were irradiated for eight hours and compared to unirradiated control samples. Irradiation experiments were performed in duplicate. Each sample then was analyzed by FT-ICR-MS in triplicate injections (N = 2 × 3). Peak intensities of all features found in irradiated samples were compared to the same features in the unirradiated control samples by Student's t-Test (n = 3): Features, which showed a significant decrease in peak intensities in both independent irradiation experiments are colored in blue. Features, which showed a significant increase or were newly formed upon irradiation are highlighted in red, respectively. (a) Volcano plot. (b) Number of molecular formulae showing significant changes in peak intensities. (c) Van Krevelen diagram of all significantly affected molecular formulae. Pie charts illustrate the reduced occurrence of nitrogenfree (CHO) MRPs in photochemical reactions. Black pie chart represents elemental compositions, which did not show a significant change in peak intensities upon irradiation.

Figure S4
Quantification of (a) urea and (b) asparagine in ribosehistidine model systems. After lyophilization, the model systems were reconstituted in 2% acetonitrile solution to achieve a dilution factor of 1:10 (v/v) with respect to the original model system. Calibration curves were computed from analyzed standard solutions as shown below.
Calibration standards used for quantification of urea and asparagine in ribose-histidine model systems.

Figure S10
Overview of compositional descriptors retrieved for the ribose-lysine model system after molecular formulae computation from FT-ICR-MS data. Bar charts are grouped into features, which showed a significant decrease (blue; log2FC < -1 and p < 0.01, Student's t-Test (n = 3)) and significant increase (red; log2FC > 1 and p < 0.01, Student's t-Test (n = 3)) in peak intensities in both independent irradiation experiments, respectively. Features that did not show a significant change in peak intensities after an irradiation time of eight hours are colored in black. Represented descriptors are (a) number of carbon atoms per formula, (b) measured m/z-values, (c) number of oxygen atoms per formula, (d) number of nitrogen atoms per formula, (e) number of double bond equivalents per carbon atom, and (f) average carbon oxidation state.

Figure S11
Overview of compositional descriptors retrieved for the ribose-arginine model system after molecular formulae computation from FT-ICR-MS data. Bar charts are grouped into features, which showed a significant decrease (blue; log2FC < -1 and p < 0.01, Student's t-Test (n = 3)) and significant increase (red; log2FC > 1 and p < 0.01, Student's t-Test (n = 3)) in peak intensities in both independent irradiation experiments, respectively. Features that did not show a significant change in peak intensities after an irradiation time of eight hours are colored in black. Represented descriptors are (a) number of carbon atoms per formula, (b) measured m/z-values, (c) number of oxygen atoms per formula, (d) number of nitrogen atoms per formula, (e) number of double bond equivalents per carbon atom, and (f) average carbon oxidation state.