*2.2. Effects of ATE and ABE Subfractions on Protein Glycation*

ATE and ABE were further consecutively partitioned into H2O, 1-butanol (BuOH), ethyl acetate (Ea), and *n*-hexane (Hex) fractions. As shown in Figure 3, ATE-BuOH and ATE-Ea exhibited greater antiglycation properties than the other fractions (Figure 3a), and the same findings were found for the ABE fractions (Figure 3b). The ATE-BuOH fractions were obtained and chromatographically isolated to subfractions (a–f), whereas the ATE-Ea fractions were isolated to subfractions (a–h). Inhibitory percentages of ATE-BuOH subfractions -a to -f against protein glycation were as follows: −7%, 0%, 88%, 85%, 56%, and 53%, respectively, at a concentration of 250 μg/mL (Figure 4a). Similarly, the

inhibitory potency of the ATE-Ea subfractions (a to h) was 15% to 37% (Figure 4b). In addition, the effects of ATE-BuOH-C and ATE-BuOH-D on the middle stage of glycation were determined by the BSA-MGO assay. MGO belongs to the group of RCS, which is a critical precursor in the formation of AGEs. Figure 5 shows that ATE-BuOH-c and ATE-BuOH-d exhibited significant inhibition of 26% and 30%, respectively, at a concentration of 250 μg/mL. These data suggested the RCS-trapping capacities of ATE-BuOH-c and -d and the main antiglycation components that possibly existed in the ATE-BuOH subfractions.

**Figure 3.** Effects of different solvent fractions of ATE (**a**) and ABE (**b**) on glucose-mediated development of fluorescence of AGEs (BSA-glucose assay). Groups with different letter superscripts are significantly different (*p* < 0.05). a–c, treated with 125 μg/mL; A–C, treated with 250 μg/mL.

**Figure 4.** Effects of ATE-BuOH (**a**) and ATE-Ea (**b**) subfractions on glucose-mediated development of fluorescence of AGEs (BSA-glucose assay). Results are means ± SD for *n* = 3. Groups with different letter superscripts are significantly different (*p* < 0.05). a–h, treated with 125 μg/mL; A–F, treated with 250 μg/mL.

#### *2.3. Effect of ATE-BuOH-Containing Phenolics on the Individual Stage of Protein Glycation*

According to our previous study, phenolic acids were major components in ATE-BuOH [22]. In addition, medium–high polarity ethanol extracts of the hull, testa, branty subfractions, eluted from ATE-BuOH below 50% methanol with Diaion HP-20 resin, possessed more significant antiglycation (Figure 4a) as compared with ATE-EA, especially ATE-BuOH-c and –d (Figure 5).

**Figure 5.** Effects of ATE-BuOH subfractions C and D on MGO-mediated development of fluorescence of AGEs (BSA-MGO assay). Results are means ± SD for *n* = 3. AG, aminoguanidine. C, the subfraction C of ATE-BuOH. D, the subfraction D of ATE-BuOH.

A high-performance liquid chromatography (HPLC) analysis was carried out to investigate the chemical composition of ATE-BuOH. As shown in Figure 6, chlorogenic acid, caffeic acid, *p*-coumaric acid, and ferulic acid were identified in ATE-BuOH with contents of 1.01 ± 0.03, 1.32 ± 0.04, 9.51 ± 0.94, and 2.54 ± 0.68 mg/g, respectively. However, gallic acid was not detectable in ATE-BuOH.

**Figure 6.** High−performance liquid chromatography (HPLC) chromatographic profiles recorded at 280 nm of (**a**) phenolic compound mixed standards (each 0.1 mg/mL) and (**b**) the ATE-BuOH fraction (10 m(g/mL) (1: gallic acid, 2: chlorogenic acid, 3: caffeic acid, 4: p-coumaric acid, and 5: ferulic acid).

In the BSA-glucose assay, chlorogenic acid and ferulic acid exhibited 20% and 28% inhibitory activity, respectively (Figure 7), indicating that these phenolics aid in reducing glucose-mediated protein modification. The BSA-MGO assay also determined the inhibition of MGO-mediated protein glycation by these phenolics. The results showed that chlorogenic acid and ferulic acid exhibited significant inhibition of 24% and 15%, respectively. Meanwhile, *p*-coumaric acid had a slight impact on antiglycation (Figure 7). In contrast, the other phenolic compounds identified in ATE-BuOH, such as caffeic acid and 6-methoxy-2-benzoxazolinone [22], showed no antiglycation effect at a concentration of 100 μM (data not shown).

**Figure 7.** Chemical structures and the effects of chlorogenic acid (**a**), ferulic acid (**b**), and *p*-coumaric acid (**c**) identified in ATE-BuOH fractions on protein glycation and crosslinking. Dose responses of glycation inhibition (%) on individual stages of protein glycation were determined by model systems of the BSA-glucose assay (upper panel), the BSA-MGO assay (middle panel), and the G.K.-ribose assay (bottom panel). Data are the means ± SD for *n* = 6. Groups with different letters (a–d) are significantly different from each other in individual assays (*p* < 0.05).

G.K. peptide-ribose assay was used to generate peptides with advanced Maillard reaction products with dimerization through lysine-lysine crosslinking [23]. Rahbar et al. [23] and our previous study [8] demonstrated that incubation of G.K. peptides with ribose resulted in late glycation product formation. Therefore, the present study utilized this model system to evaluate the inhibitory effect of phenolics on protein crosslinking. As shown in Figure 7 (lower panel), chlorogenic acid and ferulic acid exhibited substantial anti-crosslinking activities (47% and 43%, respectively, at a concentration of 100 μM).

#### **3. Discussion**

Protein glycation and subsequent AGE formation in the body have been evidenced as a risk factor in the development of diabetic macrovascular and microvascular complications and age-related diseases [24,25]. Clinical observation has revealed that patients with complicated diabetes have 40 to 100% higher AGE levels than healthy subjects [26]. Therefore, investigating AGE inhibitors, especially natural anti-glycation agents with fewer adverse effects, may be a beneficial approach to preventing diabetic complications.

The main question addressed by this study was whether adlay could inhibit protein glycation. Adlay seeds have long been consumed as a food supplement and herbal medicine in traditional Chinese medicine [10,12]. Although the health-promoting effects and therapeutic potential of adlay have been reported [10,12], studies on the potential of adlay to act against protein glycation and AGE formation are limited. This study used a classic in vitro glycation assay to evaluate the effect of individual parts of adlay (hull, testa, bran, and polished adlay) on glucose-mediated BSA glycation. Ethanol was chosen as the initial extraction solvent based on safety considerations for human consumption [27]. Moreover, most active ingredients, including polyphenols, phytosterols, and coixol, found in adlay have been extracted by ethanol with a high yield in previous studies [22,28,29]. For the first time, this study demonstrated that ATE and ABE exhibited better glycation

inhibitory effects (Figure 2), suggesting antiglycating agents are possibly present in the bran and testa of adlay.

Using a solvent extraction technique, two ATE-BuOH subfractions, ATE-BuOH-C and ATE-BuOH-D, exhibited superior antiglycation activities (Figures 4a and 5). Phenolic compounds in adlay seed were analyzed by HPLC in our previous studies [13,22,28,29]. Caffeic acid, chlorogenic acid, *p*-coumaric acid, ferulic acid, gallic acid, *p*-hydroxybenzoic acid, syringic acid, and vanillic acid were identified in ATE-BuOH and/or ATE-Ea fractions. In the present investigation, *p*-coumaric acid (9.51 ± 0.94 mg/g), along with chlorogenic acid (1.01 ± 0.03 mg/g), caffeic acid (1.32 ± 0.04 mg/g), and ferulic acid (2.54 ± 0.68 mg/g) were identified in ATE-BuOH through HPLC analysis (Figure 6). Notably, chlorogenic acid and ferulic acid from ATE-BuOH exhibited significant potent inhibition of the individual stage of protein glycation, especially in the reduction in protein crosslinking (Figure 7). After partially purified by column chromatography, the ATE-BuOH-c and –d (eluent from 50 to 25% methanol) showed superior inhibitory effects on antiglycation (Figures 3a, 4a and 5). A previous study has demonstrated that the major components in subfractions eluted from 50 to 25% methanol from ATE-BuOH were caffeic acid (9.02 mg/g subfractions), chlorogenic acid (30.30 mg/g subfractions), and ferulic acid (0.05 mg/g subfractions) [22]. These results suggest that the antiglycation properties of ATE were at least partly related to its phenolic acid content.

Research has demonstrated that antioxidative polyphenols show potent antiglycation activities [1,6,8]. Several studies have drawn attention to the positive correlation between the free radical scavenging activity and antiglycation capacity, which may be due to the interruption of ROS formation during glycation [6,8]. In addition, the antioxidant ability of phenolic acids, such as caffeic acid and chlorogenic acid, depends on the number of hydroxyl groups in the molecule that would be strengthened by steric hindrance [30]. The present study showed that chlorogenic acid and ferulic acid are potent inhibitors that act at the individual stage of glycation, including post-Amadori glycation, as evidenced by the decreased development of AGE-related fluorescence in BSA-glucose assay (Figure 7, upper panel). Notably, these phenolic acids exhibited MGO-trapping activity and anticrosslinking action, as evidenced by the MGO-BSA assay (Figure 7, middle panel) and G.K. peptide-ribose assay (Figure 7, bottom panel). However, certain compounds, such as caffeic acid, had no inhibitory effect on glycation (data not shown), which was consistent with a former study [31].

Since MGO is a critical intermediate and precursor during AGE formation, the MGO-BSA assay was utilized to determine the middle stage of protein glycation. MGO served as the glycating agent, and BSA provided the amine source targeted by the glycating agents. Albumin was used as the protein because it is the most abundant protein in serum. A clinical study noted that the serum levels of glycated albumins are two to three times higher in diabetic patients than in healthy people [32]. It is known that MGO can readily react with proteins that contain lysine residues and crosslink with proteins, along with ROS production [33]. MGO-glycated BSA was detected by specific fluorescence formation. This study showed that ATE-BuOH subfractions and their active phenolics (chlorogenic acid and ferulic acid) achieved nearly 30% inhibition of MGO-mediated protein modification (Figures 5 and 7), indicating the possible MGO-trapping potential of adlay.

Nagaraj and his colleagues [34] indicated that protein crosslinking is the major end result of the Maillard reaction. In this study, a synthetic G.K. peptide that contained lysine residue was incubated with ribose. This chemical model was designed to generate peptides with AGEs that dimerize through lysine-lysine crosslinking and increase the late glycation product formation, as determined by intrinsic fluorescence formation [23]. Chlorogenic acid and ferulic acid were found to be anti-crosslinking agents (Figure 7, lower panel). Taken together, the data suggested that the presence of methoxy groups at the C3 position of the aromatic ring in ferulic acid and the quinic acid groups in chlorogenic acid may contribute to the antiglycation actions.
