Novel caffeoylquinic acid derivatives from Lonicera japonica Thunb. flower buds exert pronounced anti-HBV activities

Lonicera japonica Thunb., possesses antiviral and hepatoprotective activities, and is widely used as a health food and in cosmetics. However, its major constituents, caffeoylquinic acid derivatives, and their anti-HBV activity were lacking systematic research. In this study, four novel caffeoylquinic acids, five simple caffeic acids and fourteen known caffeoylquinic acids are isolated and identified from L. japonica. Most caffeoylquinic acids inhibited HBsAg and HBeAg secretion, and HBV DNA replication. In particular, 100 μg ml−1 monocaffeoylquinic acid 9 inhibits HBsAg and HBeAg secretion, and HBV DNA replication by 83.82, 70.76 and 39.36% compared to the control. Unfortunately, 50 μg ml−1 tricaffeoylquinic acid 23 promotes HBsAg and HBeAg secretion, and HBV DNA replication by 172.39, 9.92 and 55.40%. Finally, structure–activity relationships reveal that caffeoylquinic acids containing a caffeoyl group have better inhibitory activities. The results indicate that caffeoylquinic acids from L. japonica could serve as anti-HBV agents for functional food or medicinal use.


Introduction
The hepatitis B virus (HBV) is a major global health problem, due to its worldwide distribution and potential adverse consequences. Current therapeutic strategies for HBV, however, are far from satisfactory. 1 Traditional Chinese medicines, with multiple components and diverse activities, are important sources for novel anti-HBV drug discoveries. Caffeoylquinic acids, which exist widely in the plant kingdom, especially in Asteraceae, 2 Umbelliferae, 3 and Caprifoliaceae, 4 are a type of polyphenol compound formed by esterication and condensation of a molecule of quinic acid and a multimolecule caffeic acid. Caffeoylquinic acids and their derivatives are reported to possess anti-HBV activities. 5,6 Lonicera japonica Thunb. (Caprifoliaceae) is native to East Asia. In China, L. japonica is widely cultivated in Shandong and Henan Provinces. Jin Yin Hua, also known as Ren Dong, the ower buds of L. japonica, are famous in traditional Chinese medicine. 7 During the recent decades, they have been used for treating inuenza, cold, fever, and infections. 8 Moreover, they have also been used in indigenous beverages, such as tea, in Korea and China for many years. 9 Previous phytochemical investigations have been reported for different types of chemical constituents, including avonoids, caffeoylquinic acids, iridoids, saponins, and other compounds. [9][10][11][12] Furthermore, caffeoylquinic acids are major constituents of L. japonica. 13,14 However, the caffeoylquinic acid derivatives form L. japonica and their anti-HBV activities are lacking in systematic research.
Province, People's Republic of China. The plants were identied by Dr XB Zeng of Shenzhen People's Hospital and a voucher specimen (no. 20170930) was deposited at Center Lab of Longhua Branch, Shenzhen People's Hospital, Second Clinical Medical College of Jinan University, Shenzhen, China.

Ester hydrolysates and HPLC analysis
Ester hydrolysates of compounds 1-4 were carried out according to the following method. In brief, compounds 1-4 (each 2.0 mg) were hydrolysed by reaction in 1 N NaOH (4 ml) and tetrahydrofuran (4 ml) at 0 C for 4 h. The reaction mixture was extracted by adjusting pH to 4.5. The extraction product was directly analysed by reversed-phase HPLC using a Diamonsil C 18 analytical column (5 mm, 4.6 Â 250 mm), which was eluted isocratically with 25% acetonitrile in water (containing 0.1% formic acid) at a ow rate of 1 ml min À1 . The temperature of the column oven was 25 C, and 20 ml was injected into the system each time. The UV spectra were collected at 327 nm. Peaks of the ester hydrolysate of compounds 1-4 were identied by comparing the retention times of authentic samples of 4-Ocaffeoylquinic acid methyl ester (compound 14, t R ¼ 10.411 min) aer simultaneous treatment under the same conditions.

Cytotoxic activity assay
The HepG 2 cells and HepG 2.2.15 cells were from the American Type Culture Collection (ATCC, Manassas, USA). All the cells were maintained in DMEM containing 10% FBS (foetal bovine serum, HyClone, Logan, UT) and cultured at 37 C (5% CO 2 , 95% relative humidity). The cytotoxicity assay was performed, according to the MTT method in 96-well microplates. Briey, 200 ml of adherent cells was seeded into each well of the 96-well cell culture plates and allowed to adhere for 24 hours before drug addiction with an initial density of 1.0 Â 10 5 cells per ml. Each tumour cell line was exposed to the test compounds at concentrations of 100, 50, and 20 mg ml À1 (DMEM with 0.1% DMSO) for three times within 48 hours. The control group was treated in the same solvent. Aer 48 hours, the media and test samples were replaced with MTT (5 mg ml À1 ) and incubated in the dark for 4 h. The formazan crystals formed were dissolved in DMSO. Absorbance was measured using a microplate reader at 520 nm. Cell viability (%) was calculated as A sample /A control Â 100%.
Anti-HBV activity assay HepG 2.2.15 cells were maintained in DMEM containing 10% FBS (foetal bovine serum, HyClone, Logan, UT) and cultured at 37 C (5% CO 2 , 95% relative humidity). The anti-HBV assay was performed according to previous research. 15 Briey, 500 ml of adherent cells was seeded into each well of the 24-well cell culture plates and allowed to adhere for 24 hours before drug addiction with an initial density of 3 Â 10 5 cells per mL. The cell supernatants were collected every three days and fresh cell culture medium contain corresponding test compounds were added. The experiment lasted nine days. The collected cell supernatants were used to analyse the levels of HBsAg, HBeAg and HBV DNA with their corresponding reagent kit. The levels of HBsAg and HBeAg in the supernatants were measured with an ELISA method. Real-time PCR assay was used to detect the HBV DNA. The relative level was calculated using the equation: relative level (%) ¼ [(A test À A control )/A control ] Â 100 (where A is the levels of secreted HBsAg or HBeAg or HBV DNA), and the subscript words 'test' and 'control' represent that the test sample and control group, respectively.

Statistical analysis
All data is expressed as mean AE SD. At least three independent experiments were performed, each in quintuplicate. The data is analysed using a one-way ANOVA. Statistically signicant effects were analysed, and the means are also compared using least-signicant difference (LSD) test. Statistical signicance is determined at p < 0.05.   (14), which was identied by HPLC analysis. Therefore, we concluded that the trans-caffeoyl moiety was attached to C-4 of quinic acid methyl ester. Moreover, the other part of the ester hydrolysates 3-hydroxy-3-  16 Therefore, the C-7 00 conguration of compound 1 was identied as 7 00 S.  (Table 1) were consistent with that of 1 but with a difference of one more methoxyl group at d H 3.07 (3H, s). The HMBC correlation from d H 3.07 to C-7 00 (d C 79.24) indicated that the extra methoxyl group was attached to C-7 00 of the propionyl moiety (Fig. 3). The 10 ppm down-eld shi of C-7 00 in the 13 C NMR spectrum further indicated the attachment of the extra methoxyl group to C-7 00 . In a similar manner, the optical rotation of the ester hydrolysates of compound 2 was measured with value [a] 22 D ¼ À4.7 , which was nearly equal to [a] 22 D ¼ À4.6 of compound 1. Therefore, the C-7 00 conguration of compound 2 was also identied as 7 00 S. Based on the above spectral evidences, the structure of compound 2 was elucidated as 4-O-trans-caffeoyl-5-O-[3S-methoxy-3-(3,4-dihydroxyphenyl)-propionyl] quinic acid methyl ester.

Results and discussion
Compounds 3 and 4 were isolated as yellow amorphous powder. The molecular formula of compounds 3 and 4 were determined to be C 26  was different. This suggested compound 3 was also a 4,5disubstituted quinic acid methyl ester but contains a transcaffeoyl group and a cis-caffeoyl group. At the same time, a comparison of the 1 H NMR and 13 C NMR spectra of compound 4 ( Table 1) (14), which were identied by HPLC analysis. The above results conrmed that the cis-caffeoyl group was linked to C-5 of compound 3 and C-3 of compound 4, respectively. The key HMBC and NOESY (H 4 H, blue) correlations of compounds 3 and 4 were shown in Fig. 3. Therefore, the structures of compounds 3 and 4 were identied as 4-O-trans-caffeoyl-5-O-cis-caffeoyl quinic acid methyl ester and 3-O-cis-caffeoyl-4-O-trans-caffeoyl quinic acid methyl ester, respectively.

Cytotoxicity
All the compounds (1-23) isolated were tested for their potential cytotoxicity against HepG 2 and HepG 2.2.15 cells. The purpose of this design was to determine their appropriate drug concentration in anti-HBV assay to eliminate the cytotoxicity interference. As shown in Fig. 4A and B, it was found that the cytotoxic results against HepG 2 were consistent with the experimental results of cytotoxicity against HepG 2.2.15 cells. This suggests that the effect of these caffeoylquinic acids on the growth of both two cell lines is consistent, although HepG 2.2.15 cells contain a complete HBV genome. The cytotoxicity assay results showed that, most compounds (1-5, 9-10, 12-16, 19-22) at a concentration range of 20-100 mg ml À1 showed no signicant cytotoxicity against HepG 2 and HepG 2.2.15 cells. Therefore, the above compounds were selected 100 mg ml À1 for test drug concentration in the subsequent anti-HBV assay. In addition, compounds 11, 18 and 23 exhibit weak cytotoxicity against HepG 2 and HepG 2.2.15 cells at concentration of 100 mg ml À1 and selected 50 mg ml À1 for the next anti-HBV activities. At the same time, compounds 6-8 and 17 were selected 20 mg ml À1 were tested for 20 mg ml À1 , compounds 11, 18 and 23 were tested for 50 mg ml À1 , and the other compounds (1-5, 9-10, 12-16, 19-22) were tested for 100 mg ml À1 . Results are expressed as the mean AE SD (n ¼ 3). *p < 0.05 compared with the control group, **p < 0.01 compared with the control group, ***p < 0.001 compared with the control group. due to their signicantly cytotoxicity against HepG 2 and HepG 2.2.15 cells at 50 mg ml À1 concentration.

Anti-HBV activities and their structure-activity relationship
To evaluate their anti-HBV activities, namely the inhibiting the secretion of Hepatitis B surface antigen (HBsAg) and Hepatitis B e antigen (HBeAg), as well as HBV DNA replication, compounds 1-23 were assayed on HepG 2.2.15 cell line stably transfected with the HBV genome in vitro, as reported previously. 15 The experiment results are summarized in Fig. 5, and most caffeoylquinic acid derivatives showed excellent anti-HBV activities on HepG 2.2.15 cell. HBsAg is the outer protein of HBV, which is a sign of HBV infection. Its level is related to the numbers of HBV. 31,32 As shown in Fig. 5A (Chen, et al., 2006). As seen in Fig. 5C, most compounds exhibited remarkable decrease in the relative HBV DNA levels (p < 0.05). Compound 8 showed the most signicant inhibitory activity and its relative HBV DNA level was decreased by 83.11 AE 2.75%. However, several compounds (4,11,12,16,(20)(21)(22) showed no signicantly inhibitory activity and some compounds (5,18,19,23) showed strangely promoting activities. Unfortunately, we found that compound 23 promoted HBsAg and HBeAg secretion, and HBV DNA replication by 172.39 AE 2.59, 9.92 AE 6.91, and 55.40 AE 17.81%. Contrary to the study on anti-HBV activity of caffeoylquinic acids, there is presently no study that reports the HBV promoting activity of caffeoylquinic acids. Structure-activity relationship analysis of these caffeoylquinic acids in anti-HBV activity suggested that the presence of caffeoyl groups in the molecule resulted in stronger anti-HBV activity. This conclusion resembles previous study on the structure-activity relationship of these compounds. 34 The two adjacent phenolic structure and a, b-conjugated unsaturated ester structure of caffeoyl groups increased benzene ring plane conjugation and resulted in the promotion of anti-HBV activity. For simple caffeic acid alone, compounds 6-8, 17 have signicantly anti-HBV activity (p < 0.001). However, compound 5 expressed inferior anti-HBV activity when compared to the other simple caffeic acid because its relative HBV DNA level (68.35 AE 15.26%) was signicantly increased compared to the control group (p < 0.001). This is because compound 5 (isovanillic acid, 3-hydroxy-4-methoxybenzoic acid) could not form a benzene ring plane conjugation without a two adjacent phenolic structure, so it exhibited poor activity. For monocaffeoylquinic acid, only compound 12 (3-O-caffeoylquinic acid) showed excellent anti-HBV activities (p < 0.001). Other isolated monocaffeoylquinic acids, namely 13 (1-O-caffeoylquinic acid methyl ester), 14 (4-O-caffeoylquinic acid methyl ester), 15 (3-O-caffeoylquinic acid methyl ester), and 18 (5-O-caffeoylquinic acid methyl ester), did not signicantly inhibited the secretion of HBsAg, HBeAg, and HBV DNA replication. These results suggest that carboxyl group in quinic acid was important for the accelerating effect on anti-HBV activities in monocaffeoylquinic acids. However, the caffeoylquinic acid ethyl ester derivatives (9) showed the best anti-HBV activity (p < 0.001). Perhaps further study on the structure-activity relationship of monocaffeoylquinic acids is worth discussing. For dicaffeoylquinic acids, compared to their structural analog (20), new compounds 1 and 2 have only one caffeoyl group in the molecule, so they exhibited weaker activities than that of compound 20. Moreover, the anti-HBV activity of new compound 3 was far better than that of compound 20, which could be due to the cis-caffeoyl group in compound 3. The above results may suggest that the existence and conguration of the double bond moiety in caffeoyl group is important for anti-HBV activity. Tricaffeoylquinic acid, compound 23 (3,4,5-tri-O-caffeoylquinic acid methyl ester), had a steric hindrance because it contained three caffeoyl groups. The increased steric hindrance effect inhibited the anti-HBV activity of compound 23. Unfortunately, the relative HBsAg level (172.39 AE 2.59%) and HBV DNA level (55.40 AE 17.81%) of compound 23 was signicantly promoted (p < 0.001). Therefore, more studies are necessary to further understand the anti-HBV action mechanism of caffeoylquinic acid derivatives.

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