Synthesis and Comparative Structure–Activity Study of Carbohydrate-Based Phenolic Compounds as α-Glucosidase Inhibitors and Antioxidants

Twenty-one natural and unnatural phenolic compounds containing a carbohydrate moiety were synthesized and their structure–activity relationship (SAR) was evaluated for α-glucosidase inhibition and antioxidative activity. Varying the position of the galloyl unit on the 1,5-anhydro-d-glucitol (1,5-AG) core resulted in changes in the α-glucosidase inhibitory activity and notably, particularly strong activity was demonstrated when the galloyl unit was present at the C-2 position. Furthermore, increasing the number of the galloyl units significantly affected the α-glucosidase inhibition, and 2,3,4,6-tetra-galloyl-1,5-AG (54) and 2,3,4,6-tetra-galloyl-d-glucopyranose (61) exhibited excellent activities, which were more than 13-fold higher than the α-glucosidase inhibitory activity of acertannin (37). Moreover, a comparative structure-activity study suggested that a hemiacetal hydroxyl functionality in the carbohydrate core and a biaryl bond of the 4,6-O-hexahydroxydiphenoyl (HHDP) group, which are components of ellagitannins including tellimagrandin I, are not necessary for the α-glucosidase inhibitory activity. Lastly, the antioxidant activity increased proportionally with the number of galloyl units.


Introduction
Impaired glucose tolerance increases the risk of vascular events such as atherosclerotic coronary artery disease [1,2]. Particularly, postprandial hyperglycemia is a serious risk factor for cardiovascular diseases and is believed to be the cause of oxidative stress that leads to vascular events [3][4][5][6][7]. Thus, controlling postprandial hyperglycemia is an important target to prevent diabetes as well as diabetic complications. In clinical medicine, α-glucosidase inhibitors such as acarbose, miglitol, and voglibose, belong to the class of antidiabetic drugs used for improving postprandial hyperglycemia [8][9][10][11]. Currently, natural products and their derivatives constitute more than half of the drugs in the clinic [12][13][14][15]. Therefore, finding inspiration in nature to develop more efficient and effective medicines has attracted significant interest.
Trees belonging to the Acer species have been used as traditional medicinal plants for many years and are widely known for their sap, which can be concentrated to produce maple syrup [16]. It has been demonstrated that Acer extracts display various bioactivities such as anti-cancer [17,18], antioxidant [19][20][21][22], and antihyperglycemic effects [23,24]. A. Honma et al. identified a compound from Acer saccharum extracts able to suppress hyperglycemia, namely acertannin, and revealed that its effects are a consequence of potent inhibitory activity toward α-glucosidase [25]. The structural components of acertannin include the characteristic 1,5-anhydro-d-glucitol (1,5-AG) sugar moiety, Tellimagrandin I, which belongs to ellagitannins, has also been demonstrated to be an αglucosidase inhibitor and to show antioxidant activity [37,38]. The molecule is characterized by the presence of a hexahydroxydiphenoyl (HHDP) group, two galloyl units and the D-glucose core possesses a hemiacetal hydroxyl functionality ( Figure 2) [39]. The HHDP group provides structural diversity in polyphenols, and the macro-lactone structure is considered to be the element responsible for the pharmacological activity [40]. Nonetheless, to our knowledge, no reports on the evaluation of the synthesis and/or bioactivity of compounds comprising the HHDP functionality on the 1,5-AG core have been reported so far. Furthermore, the hemiacetal hydroxyl group is a fundamental moiety in the carbohydrate chemistry; however, its effects on the bioactivity remain largely unexplored. In the present study, we report the synthesis of a series of 21 carbohydrate-based phenolic compounds to investigate the structure-activity relationship (SAR). α-glucosidase inhibition and antioxidant activity were examined by studying the effects of 1) the position and number of galloyl Tellimagrandin I, which belongs to ellagitannins, has also been demonstrated to be an α-glucosidase inhibitor and to show antioxidant activity [37,38]. The molecule is characterized by the presence of a hexahydroxydiphenoyl (HHDP) group, two galloyl units and the d-glucose core possesses a hemiacetal hydroxyl functionality ( Figure 2) [39]. The HHDP group provides structural diversity in polyphenols, and the macro-lactone structure is considered to be the element responsible for the pharmacological activity [40]. Nonetheless, to our knowledge, no reports on the evaluation of the synthesis and/or bioactivity of compounds comprising the HHDP functionality on the 1,5-AG core have been reported so far. Furthermore, the hemiacetal hydroxyl group is a fundamental moiety in the carbohydrate chemistry; however, its effects on the bioactivity remain largely unexplored. gallic acid functionalities as the phenolic units ( Figure 1) [26]. However, only a few plants belonging to the Acer genus produce the 1,5-AG core containing polyphenols [27,28]. To date, maplexin A-J and ginnalin A-C have been isolated and characterized. The molecules possess varying numbers and positions of the phenol units esterified with the 1,5-AG core [29][30][31][32]. These polyphenols were shown to exhibit different bioactivities such as α-glucosidase inhibition and antioxidant activity. It is noteworthy that different numbers, positions, and types of the phenol units on the 1,5-AG core display non-identical bioactivities [32][33][34][35][36]. Tellimagrandin I, which belongs to ellagitannins, has also been demonstrated to be an αglucosidase inhibitor and to show antioxidant activity [37,38]. The molecule is characterized by the presence of a hexahydroxydiphenoyl (HHDP) group, two galloyl units and the D-glucose core possesses a hemiacetal hydroxyl functionality ( Figure 2) [39]. The HHDP group provides structural diversity in polyphenols, and the macro-lactone structure is considered to be the element responsible for the pharmacological activity [40]. Nonetheless, to our knowledge, no reports on the evaluation of the synthesis and/or bioactivity of compounds comprising the HHDP functionality on the 1,5-AG core have been reported so far. Furthermore, the hemiacetal hydroxyl group is a fundamental moiety in the carbohydrate chemistry; however, its effects on the bioactivity remain largely unexplored. In the present study, we report the synthesis of a series of 21 carbohydrate-based phenolic compounds to investigate the structure-activity relationship (SAR). α-glucosidase inhibition and antioxidant activity were examined by studying the effects of 1) the position and number of galloyl HHDP group Galloyl units Phenolic hydroxyl groups Carbohydrate core In the present study, we report the synthesis of a series of 21 carbohydrate-based phenolic compounds to investigate the structure-activity relationship (SAR). α-glucosidase inhibition and antioxidant activity were examined by studying the effects of (1) the position and number of galloyl units, (2) the type of phenol units, (3) the existence the 4,6-O-HHDP group, and (4) the presence of the hemiacetal hydroxyl group.

Syntheses of Galloylated 1,5-AGs
Recently, A. Kamori et al. reported the synthesis of various natural and unnatural acertannin derivatives and evaluated their SAR against ceramidase and ceramide synthase enzymes [35]. In addition, we have also previously reported a facile method for the preparation of 1,5-anhydroalditol via treatment of per-O-TMS-glycopyranosyl iodide with LiBH 4 [41]. In total, 1,5-AG, which can be easily synthesized from d-glucose on multi-gram scale in three days, possesses four hydroxyl groups; hence, 15 different combinations are possible for mono-, di-, tri-, and tetra-galloylation of 1,5-AG. In the current study, we attempted the synthesis of all of these galloylated compounds starting from 1,5-AG.

Synthesis of a 1,5-AG-Based Tellimagrandin I Analog
To consider the effect of the hemiacetal group on the activity of tellimagrandin I, we also focused on the synthesis of the 1-deoxy analog 67. Deprotection of benzylidene acetal in 21 in a methanol/DCM solvent system using iodine, according to the method reported by Feldman et al. [56], gave diol 62 in a high yield of 93% (Scheme 7). Intermediate 62 was then condensed with the gallic acid derivative 63 [57] to afford the galloylated compound 64. The subsequent removal of the methoxymethyl (MOM) protecting groups provided 65 in 93% yield. The construction of the HHDP group was performed in accordance with the approach previously reported by Yamada et al. [57][58][59][60]. Compound 65 was treated with n-BuNH 2 and CuCl 2 to provide the biaryl derivative 66. Finally, hydrogenation of 66 in MeOH/THF with Pd(OH)2 as the catalyst gave the desired 1,5-AG-based tellimagrandin I analog 67. Scheme 7. Synthesis of the 1,5-AG-based tellimagrandin I analog 67.

The α-Glucosidase Inhibitory Activity
The α-glucosidase inhibitory activity of all samples was assayed utilizing a commercially available FUJIFILM α-glucosidase inhibitory activity assay kit. In total, 25 μL of each sample (25-2000 μg/mL in H 2 O) or acarbose (0.5-8 μg/mL in H 2 O ), 50 μL of 18.5 mM maltose diluted in maleic anhydride buffer (100 mM, pH = 6.0), and 25 μL of the rat α-glucosidase solution were incubated in a micro-tube at 37 °C for 30 min. Subsequently, 400 μL of purified water was added to the solution and the reaction mixture was boiled for 3 min to deactivate α-glucosidase. The generated glucose was The construction of the HHDP group was performed in accordance with the approach previously reported by Yamada et al. [57][58][59][60]. Compound 65 was treated with n − BuNH 2 and CuCl 2 to provide the biaryl derivative 66. Finally, hydrogenation of 66 in MeOH/THF with Pd(OH) 2 as the catalyst gave the desired 1,5-AG-based tellimagrandin I analog 67.

The α-Glucosidase Inhibitory Activity
The α-glucosidase inhibitory activity of all samples was assayed utilizing a commercially available FUJIFILM α-glucosidase inhibitory activity assay kit. In total, 25 µL of each sample (25-2000 µg/mL in H 2 O) or acarbose (0.5-8 µg/mL in H 2 O), 50 µL of 18.5 mM maltose diluted in maleic anhydride buffer (100 mM, pH = 6.0), and 25 µL of the rat α-glucosidase solution were incubated in a micro-tube at 37 • C for 30 min. Subsequently, 400 µL of purified water was added to the solution and the reaction mixture was boiled for 3 min to deactivate α-glucosidase. The generated glucose was measured by LabAssay™ glucose (mutarotase-GOD method). In total, 100 µL of the reaction solution and 150 µL of the coloring solution were incubated in 96-well plates at 37 • C for 10 min, and the absorbance was recorded at 505 nm using a microplate reader (Bio-Rad, Model 680). The inhibition percentage was calculated using the following Equation: Inhibition percentage (%) = 1 − As − Ab Ac × 100 where As is the absorbance of the analyzed sample, Ab is the absorbance of the blank (immediate deactivation), and Ac is the absorbance of the control (without α-glucosidase). Acarbose was used as the positive control.
In addition, correlation of the absorbance on the y-axis and the concentrations on the x-axis resulted in the formation of an approximately straight-line plot. The trolox-equivalent (sample-mol/trolox-mol) was calculated as follows: where Ss is the slope of the sample and St is the slope of trolox.

Discussion
The results of the biological evaluation of the 21 synthesized compounds considering the α-glucosidase inhibitory and antioxidant activities are summarized in Table 1. The comparison of the 1,5-AG-based polyphenol analogs 31-44 and 54 revealed that the α-glucosidase inhibitory activity significantly increased with the number of the galloyl units in the compounds and the highest inhibition was observed for tetra-O-galloyl-1,5-AG (maplexin J) 54 (IC 50 = 2.56 µM). This result is in accordance with the previous reports [32]. In addition, different the position of the galloyl unit on the 1,5-AG core appeared to influence the α-glucosidase inhibitory activity, even for the compounds with the same number of these moieties. Analogous outcomes were noted for 2-galloyl-1,5-AG (31) and methyl gallate (IC 50 = 95.1, 90.5 µM, respectively). Conversely, differing α-glucosidase inhibitory activities were obtained for mono-galloyl analogs 32-34, which are weaker inhibitors than methyl gallate (IC 50 = 127.1-143.9 µM). Furthermore, compound 33 (galloyl unit at the C-4 position) exhibited lowest activity (IC 50 = 143.9 µM). Higher inhibitory activity was detected for di-galloylated analogs 35-37 (IC 50 = 20.5-48.3 µM), which possessed the galloyl unit at the C-2 position than for analogs 38-40 (IC 50 = 26.6-91.4 µM). In particular, compound 35 that contains esterified galloyl units at the C-2 and C-3 positions exhibited three-times higher inhibitory activity than gallic acid. Moreover, analog 38, galloylated at the C-3 and C-4 positions, displayed significantly lower activity (IC 50 = 91.4). Among the tri-galloyllated analogs, compound 42 without a galloyl unit at the C-4 position showed stronger inhibitory activity (IC 50 = 5.34 µM) than analogs 41, 43, and 44 (IC 50 = 6.72, 9.34, and 12.6 µM, respectively). In addition, analog 44 without a galloyl moiety at the C-2 position displayed weak activity. Thus, our results suggested that a galloyl unit at the C-2 position considerably increases the α-glucosidase inhibitory activity, while the presence of this group at the C-4 position causes a decrease in the activity.
Subsequently, we compared maplexin J (54) and its analogs (55 and 56) to elucidate the effect of the phenolic hydroxyl group. The 3 ,4 -di-hydroxybenzoyl analog 55 exhibited good inhibitory activity, whereas the 3 ,5 -di-hydroxybenzoyl analog 56 was a weaker inhibitor of the α-glucosidase enzyme (IC 50 = 3.28, 9.34 µM, respectively). Consequently, these results implied that the presence of two adjacent phenolic hydroxyl groups is essential for the desired activity. We then focused on the evaluation of the influence of the hemiacetal hydroxyl and the HHDP functionalities against the α-glucosidase inhibitory activity. The effect of the hemiacetal hydroxyl moiety can be observed by the comparison of the activity of maplexin J (54) and its analog 61 (IC 50 = 1.68, 2.56 µM, respectively). Furthermore, tellimagrandin I and 67 showed analogous inhibitory activity (IC 50 = 3.37, 3.22 µM, respectively). Therefore, the obtained outcomes suggested that the presence of a hemiacetal hydroxyl group did not have a significant effect on the α-glucosidase inhibitory activity. Lastly, to examine the influence of the HHDP group, the results for maplexin J (54) and the 4,6-O-HHDP analog 67 were compared and it transpired that maplexin J (54) displayed marginally higher activity than its analog 67 (IC 50 = 2.56, 3.22 µM, respectively). Likewise, the assessment of the tellimagrandin I activity and the activity of analog 61 without the HHDP group revealed that 61 was a stronger inhibitor than tellimagrandin I (IC 50 = 3.37, 1.68 µM, respectively). Intriguingly, our results suggested that the 4,6-O-HHDP group has a weakening effect on the α-glucosidase inhibitory activity [38,62].

Conclusions
We synthesized 21 carbohydrate-based phenolic analogs including a series of compounds containing all possible combinations of galloylation on 1,5-AG. The α-glucosidase inhibition and antioxidant activities of these compounds were further studied to evaluate the SAR. Our results suggested that the α-glucosidase inhibitory activity; 1) is significantly enhanced with the increasing number of galloyl units, and changing the position of the galloyl moiety substitution on the 1,5-AG unit tends to affect the activity; particularly, the presence of this functionality at the C-2 position improves the α-glucosidase inhibition, whereas substitution at the C-4 position reduces it, 2) requires two adjacent phenolic hydroxyl groups, 3) is not affected by the presence of the biaryl bond on the 4,6-O-HHDP group, 4) is not influenced by the hemiacetal hydroxyl functionality on the carbohydrate unit. Moreover, the following trends were determined for the antioxidant activity; 1) the activity is dependent on the number of galloyl units; however, it is not affected by their position, 2) the presence of two adjacent phenolic hydroxyl groups is significant, 3) the activity is not affected by the HHDP group or the hemiacetal hydroxyl group. The α-glucosidase inhibitory activity is undoubtedly affected by the position of the galloyl group on 1,5-AG. This outcome indicates that the carbohydrate core is not only a store unit for the galloyl moiety but can also act as a carrier to the biological targets. Thus, derivatives modified at the C-2 or C-4 positions on the carbohydrate unit, which is not d-glucose, have the potential to exhibit stronger antidiabetic activity. The synthesis and profiling of further analogs will be reported in due course.