Maillard reaction is a nonenzymatic reaction between reducing sugars and free amino acid moieties, which is known as one of the most important modifications in food science. It is essential to characterize the structure of Amadori rearrangement products (ARPs) formed in the early stage of Maillard reaction. In the present study, the Nα-acetyl-lysine-glucose model had been successfully set up to produce ARP, Nα-acetyl-lysine-glucose. After HPLC purification, ARP had been identified by ESI-MS with intense [M+H]+ ion at 351 m/z and the purity of ARP was confirmed to be over 90% by the relative intensity of [M+H]+ ion. Further structural characterization of the ARP was accomplished by using nuclear magnetic resonance (NMR) spectroscopy, including 1D 1H NMR and 13C NMR, the distortionless enhancement by polarization transfer (DEPT-135) and 2D 1H-1H and 13C-1H correlation spectroscopy (COSY) and 2D nuclear overhauser enhancement spectroscopy (NOESY). The complexity of 1D 1H NMR and 13C NMR was observed due to the presence of isomers in glucose moiety of ARP. However, DEPT-135 and 2D NMR techniques provided more structural information to assign the 1H and 13C resonances of ARP. 2D NOESY had successfully confirmed the glycosylated site between 10-N in Nα-acetyl-lysine and 7′-C in glucose.

1. Introduction

Maillard reaction, also called nonenzymatic reaction, occurs between reducing sugars and amino acids, peptides, or proteins. The reaction had been studied by many researchers after it was first observed by Maillard [14]. As one of the most important reactions in food science, Maillard reaction produces color and flavor compounds during food process. There are different mechanisms of Maillard reaction to produce various final products via the formation of complex intermediates [3, 5, 6]. To simplify the process of Maillard reaction, the mechanisms are described as early stage, intermediate stage, and final stage [7, 8]. For example, the reversible condensation between the aldehyde group of glucose and the amino group of protein produces a Schiff base; then an essentially irreversible rearrangement changes the Schiff base to colorless Amadori rearrangement products (ARPs) [8, 9] (Figure 1). This ARP intermediate undergoes cycles of condensations with additional amines, dehydrations, and oxidative fragmentations to yield final heterogeneous chemical compounds as advanced glycation end-products (AGEs) [10, 11]. Nowadays, it is still a great challenge to control the reaction in food quality, nutrition assessment, and medicinal aspects due to its unclear mechanisms. Modern analytical techniques have been applied in structural characterization of ARPs, such as circular dichroism, infrared spectroscopy, fluorescence spectroscopy, mass spectrometry, and nuclear magnetic resonance (NMR) spectroscopy [12]. It has been found that the presence of tautomers in the ketose moiety of ARPs limits the structure determination by 1D 1H and 13C NMR due to the complexity of the spectra [13, 14]. However, 2D NMR techniques can provide more information for the structural characterization of ARPs.

In the present study, the Nα-acetyl-lysine-glucose model had been set up for a detailed illustration of structural information of ARP (Nα-acetyl-lysine-glucose). 1D 1H NMR and 13C NMR, the distortionless enhancement by polarization transfer (DEPT-135) spectrum, and 2D 1H-1H and 13C-1H correlation spectroscopy (COSY) were used to assign correlations between the signals in the 1H and 13C NMR spectra. To confirm the connection between the amino acid and glucose, 2D nuclear overhauser enhancement spectroscopy (NOESY) experiments were performed.

2. Experimental

D-glucose and Nα-acetyl-lysine were purchased from Sigma (St. Louis, MO, USA) with the highest purity. Water was purified by Milli Q distilled water (Millipore, Bedford, MA, USA). All other reagents and solvents were of analytical grade and were purchased from Sigma-Aldrich (St. Louis, MO, USA).

To set the reaction model, 0.1 M of Nα-acetyl-lysine was dissolved in 1 M D-glucose solution with the ratio of 1 : 10 by molecular weight. Mixture solution was freeze-dried by SC250DDA Speedvac Plus (Thermo Electron Corporation, Waltham, MA) and heated in sealed vials for 1 hour at 90°C. For structural analysis of Amadori rearrangement product, glucosylated Nα-acetyl-lysine, the reaction mixture was purified using high-performance liquid chromatography (HPLC) with a 250 × 5.0 mm i.d. C18 reverse-phase semipreparative column and ARP-containing fractions were collected and lyophilized. Before NMR analysis, purified glucosylated Nα-acetyl-lysine was identified by electrospray ionization mass spectrometry (ESI-MS) with intense [M+H]+ ion in the spectrum from m/z 100 to 500.

After HPLC purification, freeze-dried glucosylated Nα-acetyl-lysine was dissolved in dimethyl sulfoxide- (DMSO-, 500 μL) and transferred to a 5 mm NMR tube for NMR analysis. All NMR experiments were performed on a Bruker DRX 500 spectrometer (Bruker BioSpin, Germany) equipped with a 5 mm TBO probe and operated at 25°C (298 K) with a proton frequency of 500.13 MHz. The chemical shifts (δ values), given in parts per million (ppm), were referenced to the signals of the residual protons (δ 2.50 ppm) and carbon atom (δC 39.5 ppm) in DMSO-. All 1D (1H, 13C, and DEPT-135) and 2D (1H-1H COSY, 13C-1H COSY, and NOESY) NMR measurements were performed using standard Bruker pulse sequences. Sweep widths of 5000 and 25000 Hz were used in 1H and 13C NMR, respectively. 2D 1H-1H COSY and 13C-1H COSY spectra were collected in quadrature with 1024 points in and 256 points in , and the sweep widths were 5000 and 15000 Hz of 1H and 13C dimensions, respectively. 2D NOESY was recorded with mixing time of 400 ms and 256 increments containing 16 transients of 2048 complex data points. 2D NMR data were applied with a 90° phase-shifted, squared sine-bell window function in both dimensions prior to Fourier transformation.

3. Results and Discussion

[M+H]+ ion at 351.13 m/z had been found in ESI-MS spectrum, and the purity of ARP had been determined as over 90% by the relative intensity of the [M+H]+ ion.

Structural characterization of purified ARP, glucosylated Nα-acetyl-lysine (Figure 2), was carried out by performing a series of 1D and 2D NMR experiments, and the resonance assignments of 1H and 13C had been accomplished.

The signals with the chemical shifts from 1.0 to 4.0 ppm in 1H NMR spectrum had been grouped and assigned to aliphatic protons in glucose moiety (δ 2.80 to 3.90 ppm) and Nα-acetyl-lysine moiety (δ 1.0 to 2.80 ppm) of glucosylated Nα-acetyl-lysine, respectively (Figure 3). The complexity of 1H NMR spectrum in glucose moiety suggested that different conformations existed in the glucose moiety. Signals in 13C NMR spectrum also presented as groups of two moieties: glucose moiety (δ 50 to 100 ppm) and Nα-acetyl-lysine moiety (δ 20 to 60 ppm) (Figure 4(a)). The finding of multiple chemical shifts of 13C signals in glucose moiety as 1H signals had confirmed the presence of isomers with open-chain and closed-chain structures in the moiety (Figure 2) [15]. In 13C NMR spectrum, two carbonyl signals had been observed at 169 ppm and 175 ppm that were assigned to 5-C and 2-C in ARP structures.

The DEPT-135 spectrum was a spectral editing sequence, which could verify –CH, –CH2, and –CH3 carbons by the phase of signals, and signals from quaternary carbons and other carbons with no attached protons were always absent. In DEPT-135 spectrum of ARP, three –CH and two –CH2 in glucose moiety and one –CH, four –CH2, and one –CH3 in Nα-acetyl-lysine moiety had been found (Figure 4(b)). The signal at ~83 ppm was verified as –CH group and assigned to 2′-C in the open-chain structure while the signal at 102 ppm was assigned to 1′-C in the closed-chain structure. The 13C signal at 22 ppm was also identified as the overlapping signal of two carbons –CH3 and –CH that were assigned to 1-C and 7-C, respectively.

The 1H signal at 7.7 ppm suggested an amide proton (3-NH) of lysine moiety, and this proton was strongly correlated with the proton at 4.0 ppm according to the 2D 1H-1H COSY spectrum in Figure 5. Meanwhile, the correlated carbon of the proton at 4.0 ppm was the signal at 51 ppm in 13C-1H COSY spectrum in Figure 6, which was verified as a –CH group in DEPT-135 spectrum. Based on this information, the 1H signal at 4.0 ppm was assigned to α-H (4-H) of lysine moiety. The 1H-1H COSY spectrum showed the other correlations between the protons at 4.0 ppm and 1.5 ppm, 1.5 ppm and 1.3 ppm, and 1.5 ppm and 2.7 ppm, indicating the connection of lysine aliphatic chain from 4-H to 9-H in 1H-1H COSY spectrum (Figure 5). The 1H signal at 2.7 ppm was assigned to 9-H as the connection to 10-NH in lysine chain. According to the integral of the signals, the overlapping of 6-H and 8-H at 1.5 ppm had been found. The correlations between the protons at 1.6 ppm and 1.5 ppm and 1.5 ppm and 2.7 ppm suggested that the 1H signals at 1.5 ppm and 1.6 ppm were assigned to the two protons at 6-H, and 1H signals at 1.5 ppm were also assigned to 8-H. Consequently, signal at 1.3 ppm was assigned to 7-H and the assignment of 9-H was also confirmed. The carbon resonances in lysine aliphatic chain were further assigned by the correlations in 13C-1H COSY spectrum with their corresponding proton assignments (Table 1). The 13C-1H COSY spectrum suggested that the protons at position 6 had split at 1.6 ppm and 1.5 ppm, which was also presented by the strong correlation in 1H-1H COSY spectrum. The 13C assignments in Nα-acetyl-lysine moiety were also confirmed by 1D 13C NMR and DEPT-135 spectra.

The multiple correlations in the glucose moiety of 1H-1H COSY spectrum also suggested the presence of isomers in the moiety. The strong correlations in 1H-1H COSY spectrum between 3.4 ppm and 3.5 ppm and 3.5 ppm and 3.8 ppm had been observed; the two sets of 1H signals were correlated to the –CH2 at 61 ppm and 64 ppm in the 13C-1H COSY spectrum, respectively. Consequently, 1H signals at 3.4 ppm and 3.5 ppm were assigned to 5′-H in the open-chain structure, while 1H signals at 3.5 ppm and 3.8 ppm were assigned to 5′-H of glucose moiety in the closed-chain structure. The –CH2 at ~50 ppm in glucose moiety was assigned to 7′-C, and the correlated proton signals had been found at ~2.9 ppm in 13C-1H COSY spectrum. The assignment of 7′-H was essential to the confirmation of the ARP that contained the chemical bonding between 10-N and 7′-C. The –CH signals at 78 ppm and 70 ppm were assigned to 3′-C and 4′-C in the open-chain structure, and the –CH signals at 82 ppm, 76 ppm, and 69 ppm were assigned to 2′-C, 3′-C, and 4′-C in the closed-chain structure, respectively. The signals of aliphatic protons in glucose moiety were consequently assigned according to the correlations in 13C-1H COSY spectrum (Table 1).

2D NOESY experiments were carried out to confirm the crosslink between the glucose and Nα-acetyl-lysine moieties (Figure 7). The correlations between 2.9 ppm and 2.7 ppm in NOESY spectrum suggested that the 7′-H had an NOE effect with proton at 9-H that could only be found when ARP formed in the reaction model. In NOESY spectrum, there was a correlation between 2.7 ppm and 2.6 ppm that suggested that 1H signal at 2.6 ppm had an NOE effect with 9-H, and the signal was then assigned as 10-NH. The 1H signal of 10-NH shifted from ~8 ppm in free lysine to high magnetic field also suggested that the glucose moiety was chemically bonded to 10-NH.

In summary, structural characterization of ARP glycosylated Nα-acetyl-lysine in Maillard reaction had been accomplished by serials of NMR experiments. The complexity of 1D 1H NMR and 13C NMR was observed due to the presence of isomers in glucose moiety of ARP with the open-chain structure and the closed-chain structure. However, DEPT-135 and 2D NMR techniques provided more structural information of ARP, and the assignment of 1H and 13C resonances had been accomplished. 2D NOESY had successfully confirmed the glycosylated site between 10-N in Nα-acetyl-lysine and 7′-C in glucose.

Conflict of Interests

There are no competing interests amongst the authors. The authors declare that there is no conflict of interests.

Authors’ Contribution

Hui Wang and Chuanjiang Li contributed to the paper as the first authors.


The Ph.D. scholarship for Dr. Eric Dongliang Ruan from the University of Hong Kong is gratefully acknowledged.