Synthesis, antioxidant activity, and density functional theory study of catechin derivatives

Key Laboratory of Pu-er Tea Science, Min Yunnan Agricultural University, Kunming 6 126.com; wangxuanjun@gmail.com; shengj Tel: +86-159-1257-9655 State Key Laboratory for Conservation and Kunming 650201, PR China State Key Laboratory of Phytochemistry and Institute of Botany, Chinese Academy of Sci † Electronic supplementary information computational study, antioxidant activity ‡ These authors contributed equally to th Cite this: RSC Adv., 2017, 7, 54136

Reactive oxygen species (ROS) are continuously produced as byproducts of mitochondrial electron transport during cellular respiration in the body. ROS include the superoxide anion (O 2 À ), hydrogen peroxide (H 2 O 2 ), the hydroxyl radical (HOc) and singlet oxygen ( 1 O 2 ). 1 These species are unstable and react readily with other molecules to achieve stability. "Oxidative stress" is the pro-oxidative state when the ROS level exceeds the capability of defense mechanisms. This can pose a threat to cells, causing damage to DNA, lipids, and proteins, and potentially leading to conditions such as cell ageing, cardiovascular diseases and cancerous cell growth. 2,3 Antioxidants provide protection against oxidative damage by scavenging free radicals and reducing ROS. For this reason, many compounds have been widely studied for their antioxidant activities, using various methodologies. 4,5 Catechins are plant polyphenolic compounds. The most important and rich natural source of catechins has been iden-tied in green tea (Camellia sinensis Theaceae), 6,7 which is one of the most widely consumed beverages in the world. The major green tea catechins are (À)-epicatechin (EC), (À)-epicatechin-3gallate (ECG), (À)-epigallocatechin (EGC), (À)-epigallocatechin-3-gallate (EGCG) (Fig. 1), which is the most important active compound. These are responsible for the benecial effects of green tea in the prevention or treatment of various conditions such as cancer, heart diseases, diabetes, and neurodegenerative diseases. These active compounds have also been reported to possess antioxidant, anti-inammatory, anti-microbial and antiviral effects. 8,9 However, catechins were also known to have negative effects such as pro-oxidative, cytotoxic, and phytotoxic activities, and attempts to use catechins in the treatment of human diseases have been mostly unsuccessful. [9][10][11][12] It is well known that structural features, namely the number of galloyl and hydroxyl (OH) groups in catechin molecules, play an important part in their biological activities, particularly their antioxidant properties. 13,14 There have been very few structure-function studies of catechins with respect to their molecular mechanisms of action. Guo et al. 15 reported that the free radical scavenging activity of EGCG was stronger than that of its non-galloylated counterparts (that do not possess a galloyl D-ring, EGC). However, Valcic et al. 16 provided convincing evidence that the gallate B-ring of EGCG, and not the galloyl D-ring, is the primary site for antioxidant reactions. Hence, further studies of catechins are required to give clearer insights into the structure-function relationship of the various phenolic hydroxyl groups in the A-, B-, and D-rings of catechins, and their chemopreventive mechanism.
The aim of this study was to synthesize a series of catechin derivatives, determine their radical scavenging capacities, reveal the character of interaction and evaluate the structureactivity relationships within the group of catechins i.e. the effect of the galloyl moiety on the C-ring, the number of hydroxyl groups in the B-ring, and the spatial arrangement of the substituents on the C-ring. The synthesized compounds were characterized by 1 H-NMR, 13 C-NMR, and mass spectrometry (MS) analyses. The antioxidant activities were experimentally veried using 2,2-diphenyl-1-picrylhydrazyl (DPPH) assays, and further rationalized based on computational studies using density functional theory (DFT).
To obtain catechin derivatives 5-8, EC (1), ECG (2), EGC (3) and EGCG (4), respectively were permethylated with dimethylsulfate and potassium carbonate in anhydrous acetone under reux (overall yield 90%). [17][18][19][20] Trimethylated EGCG (3M-EGCG, 11) and pentamethylated EGCG (5M-EGCG, 12) were prepared starting from EGCG (Scheme 1), as previously described. 21,22 Molecular geometries obtained through theoretical methods are useful to explain the three-dimensional structures of compounds. Optimization of all compounds was carried out at the B3LYP/6-311++G (d, p) level of DFT. 23 Optimized geometries of compounds 1-4, 11, and 12 are shown in Fig. 2, and important bonds lengths are listed in Table 1. Among all the compounds, only compound 1 has an X-ray structure reported in the literature. 24 The differences between X-ray and calculated bond lengths were found in the range 0.001-0.017Å for 1, and the maximum deviation for C(2)-C(11) in 1, was 0.017Å. Molecular electrostatic potential (MEP) mapping through a computer-aided method is a very useful approach to explore the reactivity of compounds. The nucleophilic and electrophilic sites in any compound are expressed in term of different color codes; a deep red color indicates an electron-rich site, whereas deep blue indicates an electron-decient site. MEP mapping results for compounds 1-4, 11, and 12 are shown in Fig. 3. It is clear that electronic density in the compounds is concentrated in the oxygen of the 7-OH in the A-ring, which is directly attached to the benzene ring, and that protons attached to the A-ring are electron-rich sites. A similar trend was observed for 2, 4, and 11, with O-dihydroxy substitution in the B-ring, but 12 had some extra localization of electronic density on methoxy groups in B-ring. In 2, 4, and 12, the electronic density was more dispersed and was concentrated on the A-ring and B-ring, but not on the oxygen of 5 00 -OH in D-ring for 2 and 4, or on 4 00 -OH in D-ring for 12, while the protons attached to the D-ring were electron-decient sites.
Bond dissociation enthalpies (BDEs) have been calculated at the B3LYP theory level for the respective hydrogen atom elimination paths. Lower BDEs are associated with higher antioxidant activity. 25 Table 2 lists the calculated BDE values for the studied compounds. As shown in Table 2, the BDE values are lower in the B-ring (3 0 -OH and 4 0 -OH) than A-ring (5-OH and 7-OH) and D-ring (3 00 -OH, 4 00 -OH, and 5 00 -OH), except in the case of compound 12. In compound 2, the BDE (379.73 kJ mol À1 ) of the 3 0 -OH in B-ring is the lowest, and the BDE (414.89 kJ mol À1 ) of 7-OH in the A-ring is the highest. A similar trend was observed for compounds 1, 3, and 4, with the lowest BDE values found for 3 0 -OH (381.04 kJ mol À1 , 385.80 kJ mol À1 , and 384.98 kJ mol À1 , respectively), and the highest BDE values found for 7-OH (406.50 kJ mol À1 , 413.06 kJ mol À1 , and 416.47 kJ mol À1 , respectively). The lowest BDE of compound 11 was 381.83 kJ mol À1 for 4 0 -OH, while the lowest BDE of compound 12 was 390.48 kJ mol À1 for 5 00 -OH. The BDE in the B-ring increased owing to the involvement of hydrogen in an intramolecular hydrogen bond with a phenolic hydroxyl, which occurs when the hydrogenation reaction produces free radicals.
Frontier molecular orbital analysis by computational methods is a useful way to understand the reactivity and electronic transitions within molecules. 26 Frontier orbitals (HOMO and LUMO) are the main participants in electronic transitions, and their energy gap depicts the reactivity. 27 The HOMO and LUMO of compounds 1-4, 11, and 12 were explored at the B3LYP/6-311++G (d, p) level of DFT.
The distribution pattern of frontier molecular orbitals (HOMOs/LUMOs along with corresponding energies) of compounds 1-4, 11, and 12 at the ground states are shown in Fig. 4. The p-cloud in the LUMOs of compounds 1 and 3 is distributed on the A-and C-rings, but the p-cloud in HOMOs of compounds 1 is distributed on the B-ring, while that of 3 is distributed on the entire skeleton (A-, B-, and C-rings). The pcloud in the HOMOs of compounds 2, 4, 11, and 12 is distributed on the A-, B-, and C-rings, and the p-cloud in the LUMOs of compounds 2, 4, 11, and 12 is distributed on the D-ring. Introducing different groups on the benzene ring does not have much effect on the electronic cloud. The free radicals  scavenging reaction of catechins is an electronic process, and the results show that the free radical reaction occurred primarily on the B-ring. Detailed HOMO and LUMO energies of compounds 1-4, 11, and 12, along with their gaps, are listed in Table 3. Among all compounds, 2 showed the lowest HOMO-LUMO energy gap, i.e., 0.162 eV, while 12 showed the largest energy gap (0.206 eV).
Reactivity indices such as chemical hardness (h), electrophilicity (u), electronic chemical potential (m), electron acceptor power (u + ), and electron donor power (u À ) are excellent tools to describe the hardness, reactivity and stability of compounds. [28][29][30][31][32] The reactivity indices of compounds 1-4, 11, and 12 were calculated and are given in Table 4. It is clear that among all compounds, 2 has the lowest value of h, i.e., 0.081 eV, whereas 12 has the highest value (0.103 eV). These ndings are consistent with the LUMO-HOMO band gaps of all compounds. Compound 2 has the highest electronic chemical potential (m) value (0.142 eV) among compounds 1-4, 11, and 12, while 12 has the lowest chemical potential value (0.113 eV). The results indicate that 12 has the lowest electrophilicity index value and is nucleophilic in nature, whereas 2 has the highest value u, i.e., 0.124 eV, and is strongly electrophilic in nature. In addition, among the set of compounds, 2 has the highest electron   Compounds 1-8, 11, and 12 were tested against DPPH at different concentrations, and the results were compared with those of a standard natural antioxidant, ascorbic acid. All results are presented as IC 50 values in Table 1. Almost all the permethylation derivatives (compounds [5][6][7][8] showed no antioxidant activity, and the highest scavenger activity was observed in compound 2, whose IC 50 value was 136.637 mM. Among all the synthesized compounds, 11 and 12 with IC 50 1254.901 mM and 4119.869 mM, respectively, showed antioxidant activity ( Table 5).
The radical scavenging capacities of catechin derivatives were interacted with structure-activity relationships, which the B-ring of catechins is a stronger electron donor than other rings. In addition, the acid-base equilibrium may affect the antioxidant activity of these compounds. Participations of hydroxyl group in acid-base equilibrium is phenolic groups, including dipoledipole interaction and hydrogen bonging interaction could be simultaneously offered stability on the reactive oxygen species.

Conclusions
In conclusion, this study reports the synthesis of catechin derivatives, starting from green tea catechins EC, ECG, EGC, and EGCG. All compounds were screened for antioxidant activity: the highest antioxidant activity was observed in compound 2, whose IC 50 value was 136.637 mM; permethylated derivatives (compounds 5-8) showed no antioxidant activity, partially methylated derivatives (compounds 11 and 12) showed weak activity. A theoretical calculation based on DFT was performed to understand the antioxidant activity of catechins and their derivatives in relation to molecular structure. On the basis of the geometries, positive charge of hydrogen, MEP, EDE, the HOMO and LUMO, and reactivity indices (h, m, u, u + , and u À ), we interpreted the free radical scavenging capacity of the catechins and their derivatives in terms of their structures. We also found that the B-ring is a stronger electron donor than the A-or D-ring, and that there is a good relationship between the BDEs. These theoretical results will be helpful in the future development of antioxidant compounds.

Author contributions
Cheng-Ting Zi, Xuan-Jun Wang, and Jun Sheng conceived and designed the experiments. Han Tang and Pan Zhang synthesized the compounds. Jing Wang and Qi Wang measured antioxidant activity. Bo Hou and Bang-Lei Zhang did the theoretical calculations. Yei-Wei Huang, Ya Wang, and Ze-Min Xiang did the analysis of the data. Cheng-Ting Zi and Xuan-Jun Wang wrote the manuscript.

Conflicts of interest
The authors declare no competing nancial interests.