Biotransformation of the Phenolic Constituents from Licorice and Cytotoxicity Evaluation of Their Metabolites

Biotransformation of four bioactive phenolic constituents from licorice, namely licoisoflavanone (1), glycyrrhisoflavone (2), echinatin (3), and isobavachalcone (4), was performed by the selected fungal strain Aspergillus niger KCCM 60332, leading to the isolation of seventeen metabolites (5–21). Structures of the isolated compounds were determined on the basis of extensive spectroscopic methods, twelve of which (5–7, 10–17 and 19) have been previously undescribed. A series of reactions including hydroxylation, hydrogenation, epoxidation, hydrolysis, reduction, cyclization, and alkylation was observed in the biotransformation process. All compounds were tested for their cytotoxic activities against three different human cancer cell lines including A375P, MCF-7, and HT-29. Compounds 1 and 12 exhibited most considerable cytotoxic activities against all the cell lines investigated, while compounds 2 and 4 were moderately cytotoxic. These findings will contribute to expanding the chemical diversity of phenolic compounds, and compounds 1 and 12 may serve as leads for the development of potential cancer chemopreventive agents.


Introduction
Biotransformation can be defined as a specific modification (or modifications) of a chemical compound to a product with structural similarity by means of biological catalysts. A biological catalyst can be a whole microorganism or its enzyme, or other organisms [1][2][3][4]. Microorganisms can catalyze various reactions including hydroxylation, dehydrogenation, methylation, etc., to modify the chemical structure of a bioactive substrate resulting in the formation of metabolites which maintain the core structure of the substrate [5][6][7]. Biotransformation using microorganisms has advantages over conventional chemical synthesis due to its environmental acceptability, stereo-and regio-selectivity, and mild conditions [8]. It could be used as an alternative to chemical synthesis for generation and optimization of lead compounds in drug discovery and development [9]. Moreover, microorganisms possess the capability to mimic mammalian metabolism as their expression of enzymes that are homologous to phase I and II xenobiotic-metabolizing enzymes such as cytochrome P450 monooxygenases, UDP-glucuronosyltransferases, aryl sulfotransferases, and glutathione S-transferases [10][11][12].
Licorice is one of the most popular medicinal plants that has been used as a remedy for cough, gastric ulcer, inflammation, abdominal pain, cardiovascular diseases, and cancer since ancient time [13]. The major constituent glycyrrhizin and its aglycone glycyrrhetinic acid are also used in modern medicine. On the other hand, the phenolic constituents of licorice have been implicated in health-beneficial effects, particularly in stomach ulcers. It has been reported that the phenolic constituents of licorice have pharmacological effects for health [14]. However, little is known on the biotransformation and cytotoxicity of these health promoting phenolic compounds.
Compound 6 was obtained as a yellow amorphous powder. HRESIMS of 6 showed an [M + Na] + peak at m/z 411.1060 (calcd for C 20 H 20 O 8 Na, 411.1056) which established its molecular formula as C 20 H 20 O 8 which corresponds to a dihydroxylated metabolite of 2. The 1 H-and 13 C-NMR spectral data revealed twenty carbon signals consisting of two methyl groups, one methylene, six methines, ten quaternary carbons, and one carbonyl quaternary carbon. There were no olefinic proton signals of the prenyl moiety in the 1 H-NMR spectrum of 6. Moreover, the HMBC spectrum showed correlations between the two methyl protons (δ H 1.25) and the two oxygen-bearing carbons at δ C 79.2 and 72.5 indicating the dihydroxylation took place at the olefinic double bond of the prenyl group. On the basis of these evidences, structure of compound 6 was elucidated as 5,7,3 ,4 -tetrahydroxy-5 -(2,3-dihydroxy-3-methylbutyl)isoflavone. It showed 1 H-and 13 C-NMR spectral features closely related to those of 6. Compared with the NMR spectral data of compound 2, compound 7 exhibited the absence of olefinic proton signals belonging to the prenyl group. Meanwhile, presence of one oxygen-bearing methine carbon signal at δc 69.2 corresponding with the proton signal at δ H 3.79 and one oxygen-bearing quaternary carbon signal at δc 77.4 were observed based on the HSQC and HMBC correlations. These results suggested that the olefinic double bond of the prenyl group was epoxidized. Thus, the structure of 7 was elucidated as 5,7,3 ,4 -tetrahydroxy-5 -(2-epoxy-3-methylbutyl)isoflavone.
Compound 12 was obtained as a pale yellow amorphous powder. HRESIMS analysis showed the [M + Na] + peak at m/z 395.1833 which was in accord with the molecular formula C 22 H 28 O 5 . By comparing the NMR data of 12 with those of 10, it was found that the NMR data of 12 were identical to those of 10 except for the additional ethoxyl proton signals at δ H 3.56 (2H, q, J = 7.0 Hz) and 1.29 (3H, t, J = 7.0 Hz) ( Table 2). It was suggested that the methoxyl group belonging to the 3-methoxy-3-methylbutyl substituent in 10 was replaced by an ethoxyl group in 12. This deduction was consistent with the difference in molecular ion masses [∆m/z = 14.0155 mmu (CH 2 )]. HMBC correlations between the protons at δ H 3.56 (H-1 ) and 1.29 (H-2 ) and the carbon at δc 76.3 (C-3 ) confirmed the attachment of the ethoxyl group at C-3 ( Figure 3). Therefore, compound 12 was assigned 4,2 ,4 -trihydroxy-3 -(3-O-ethyl-3-methylbutyl)dihydrochalcone.
Compound 13 was obtained as a pale yellow amorphous powder. The HRESIMS of 13 displayed an [M + Na] + peak at m/z 349.1416 which was consistent with the molecular formula C 20 H 22 O 4 , indicating 10 indices of hydrogen deficiency. UV spectrum showed absorption maxima at 220 and 286 nm. Comparison of its NMR spectroscopic data with 6 of 16 those of 10 indicated that 13 have a similar structure but with a 2,2-dimethyldihydropyran ring in the case of 13 (Table 2). On the basis of the HMBC correlation from H-5 to C-3 together with the presence of the intramolecular H-bonded hydroxyl proton signal at δ H 13.17 (2 -OH), it was confirmed that the additional 2,2-dimethyldihydropyran ring was fused to ring A via C-3 and C-4 positions. Compound 13 was therefore characterized as 4,2 -dihydroxy-(2,2-dimethyl-3,4-dihydropyran)-(5",6":3 ,4 )dihydrochalcone.
Compound 14, isolated as a pale yellow amorphous powder, showed a sodium adduct molecular ion peak at m/z 349.1415 in the HRESIMS corresponding to the molecular formula C 20 H 22 O 4 , which was the same as that of 13. The overall NMR data of 14 showed analogous structural features to those of 13 except for the absence of an H-bonded hydroxyl proton resonance in the lower field (Table 3). These data suggested that the 2,2dimethyldihydropyran ring was fused to C-2 and C-3 positions of 14. Supportive evidence for this deduction was provided by the up-field shifted carbon resonance at δ C 155.2 (C-2 ) after combined analysis of its HSQC and HMBC data. Moreover, the NMR data of 6 were in good accordance with those of deoxydihydroxanthoangelol H in which a methoxyl group was attached at C-4 instead of a hydroxyl group in 6 [20]. Thus, the structure of 14 was assigned 4,4 -dihydroxy-(2,2-dimethyl-3,4-dihydropyran)-(5",6":3 ,2 )dihydrochalcone.   (Table 3) in conjunction with HSQC and HMBC experiments delineated the presence of twenty carbon atoms consisting of the following functional groups: two methyl, three methylene, seven methine, and seven quaternary and a carbonyl carbons (Table 3). In the 1 H-NMR spectrum, resemblance of the resonance signals between compounds 15 and 10 suggested that both have the same skeleton of α,β-dihydrochalcone. Meanwhile, the 1 H-NMR spectrum of 15 obviously showed characteristic signals for a 2-(1-methyl-1-hydroxyethyl)dihydrofuran ring fused to an aromatic ring at δ  (Tables 2 and 3). The HMBC correlations from H-1 (δ H 2.57) and H-2 (δ H 1.50) to C-3 (δ C 115.8) confirmed the 3-hydroxy-3-methylbutyl group to be attached at C-3 on the skeleton of 4,2 ,4 -trihydroxychalcone. Thus, compound 17 was characterized as 4,2 ,4 -trihydroxy-3 -(3-hydroxy-3-methylbutyl)chalcone.
Structures of three other known compounds were identified as brosimacutin M (18) [25], brosimacutin H (20) [26], and bavachromanol (21) [27,28] by comparing their spectral data with those reported in the literatures (Figures S74-S76). However, absolute configuration of their hydroxyl groups remained undetermined due to the limited quantities of the isolates. Further study may be necessary to determine the absolute configuration in compounds 18, 20, and 21.

Proposed Metabolic Pathways of Isobavachalcone (4) Catalyzed by A. niger KCCM 60332
Biotransformation of isobavachalcone (4) by the selected fungal strain A. niger produced metabolites 10-21 through hydrogenation, epoxidation, hydrolysis, reduction, cyclization, and alkylation ( Figure 4). The prenyl substituent and α,β-double bond were the major sites for biotransformation by A. niger. data with those reported in the literatures (Figures S74-S76). However, absolute configuration of their hydroxyl groups remained undetermined due to the limited quantities of the isolates. Further study may be necessary to determine the absolute configuration in compounds 18, 20, and 21.

Cytotoxicity Evaluation
The parent compounds 1-4 and all isolated metabolites 5-21 were evaluated for in vitro cytotoxic potential against human cancer cell lines A375P, HT-29, and MCF-7 using

Cytotoxicity Evaluation
The parent compounds 1-4 and all isolated metabolites 5-21 were evaluated for in vitro cytotoxic potential against human cancer cell lines A375P, HT-29, and MCF-7 using modified MTT method [29]. The results are presented on Table 4. Noteworthily, compounds 1 and 12 showed the strongest cytotoxic activities against human cancer cell lines A375P, A549, and MCF-7 with IC 50 values ranging from 4.4 to 10.1 µM, while compounds 2 and 4 were moderately cytotoxic.

Plant Material
The roots and rhizomes of licorice (Glycyrrhiza inflata) were purchased from the herbal market Sehwadang (Gwangju, Korea) in March 2018, which were identified by Dae Hyo Pharmacy Co., Ltd. (Suwon, Korea).

Extraction and Isolation of Substrates 1 and 2
The dried plant material (2.5 kg) was powdered and extracted with 95% EtOH (7.5 L × 3) and was dispersed in water and successively extracted with n-hexane, CH 2 Cl 2 , EtOAc and n-BuOH. The CH 2 Cl 2 extract (50 g) was separated by column chromatography eluted with CHCl 3 : MeOH to obtain fractions C1-C20. Fraction C11 was further separated by HPLC to yield compound 1 (30 mg), and fraction C15 was further separated by HPLC to yield compound 2 (45 mg). The structures of 1 and 2 were confirmed by comparison of their 1 H-NMR data with those previously reported [15,30].

Microorganisms and Screening for Biostransformation
Biotransformation was carried out according to the two-stage procedure [35]. In the screening studies, the actively growing microbial cultures were incubated in 250 mL flasks containing 50 mL of media with gentle agitation (200 rpm) at 25 • C in a temperaturecontrolled shaking incubator. Ethanol solution (20 mg/mL, 50 µL) of the substrate 1, 2, 3, or 4 was added to each flask 24 h after inoculation. And further incubation was performed under the same condition for six days. Two controls were used for biotransformation in this study, i.e., culture controls consisting of microorganisms growing in the culture media without substrates, and substrate controls consisting of culture media and substrates incubated without microorganisms. General sampling and TLC monitoring were performed on Merck silica gel F 254 -precoated glass plates. A. niger was identified as the most potent strain to metabolize 1-4 and therefore selected for scale-up fermentation.

Scale-up Fermentation, Extraction, and Isolation of Metabolites 5-21
For scale-up fermentation, A. niger was incubated in 500 mL Erlenmeyer flasks containing 150 mL of media. After a further 24 h incubation, the ethanol solution (20 mg/mL, 150 µL) of each substrate (1, 2, 3, or 4) was evenly distributed to each flask containing stage II cultures (Table 5). After incubation, the liquid cultures of 1, 2, 3, or 4 were extracted three times with equal volumes of EtOAc, and the organic layer was collected and concentrated at reduced pressure.
The organic extract of 1 incubated with A. niger was subject to HPLC with a gradient solvent system of 45% MeOH to 67% MeOH to afford 5 (4.5 mg, t R = 50.3 min) at a flow rate of 2.0 mL/min.
The organic extract of 2 incubated with A. niger was subject to HPLC with a gradient solvent system of 55% MeOH to 69% MeOH to afford 6 (3.2 mg, t R = 14.2 min) and 7 (2.0 mg, t R = 16.2 min) at a flow rate of 2.0 mL/min.
The organic extract of 3 incubated with A. niger was subject to HPLC with a gradient solvent system of 45% MeOH to 72% MeOH to afford 8 (5.0 mg, t R = 27.1 min), 9 (2.5 mg, t R = 31.2 min).
The organic extract of 4 incubated with A. niger was subject to HPLC with a gradient solvent system of 50% MeOH to 90% MeOH to furnish nine fractions (Fr. A-I) and 13 (1.9 mg, t R = 50.3 min) at a flow rate of 2.0 mL/min. Fr. G, H, or I was further separated on HPLC eluting with a gradient solvent system from 60% MeOH to 84% MeOH to yield 10 (1.7 mg, t R = 19.3 min), 11 (2.8 mg, t R = 14.6 min), and 12 (3.6 mg, t R = 16.6 min), respectively. Fr. D was purified by HPLC with an isocratic solvent system of 65% MeOH to yield 14 (3.3 mg, t R = 20.0 min). Fr. E was further purified by HPLC with an isocratic solvent system of 65% MeOH to yield 15 (3.6 mg, t R = 18.6 min) and 16 (3.7 mg, t R = 19.4 min). Fr. F was further purified on HPLC eluting with an isocratic solvent system of 58% MeOH to yield 17 (2.8 mg, t R = 16.5 min). Fr. C was subject to HPLC with a gradient solvent system from 54% MeOH to 60% MeOH to yield 18 (2.2 mg, t R = 16.7 min) and 19 (2.7 mg, t R = 20.1 min). Fr. A was purified by HPLC with an isocratic solvent system of 55% MeOH to yield 20 (2.1 mg, t R = 7.9 min). Fr. B was purified by HPLC with an isocratic solvent system of 62% MeOH to yield 21 (1.9 mg, t R = 14.0 min).  Table 2).

Cytotoxicity Assay
Tested compound solutions were prepared in DMSO and stored as stock solution at 4 • C. Upon dilution into culture medium, the final DMSO concentration was below 1% (v/v), a concentration without effect on cell replication. The human cancer cell lines consisted of human melanoma (A375P), human colorectal adenocarcinoma (HT-29), and breast adenocarcinoma (MCF-7). All cell lines were cultured in Dulbecco's modified Eagle's medium (DMEM) with 5% fetal bovine serum (FBS), 100 U/mL penicillin and 100 µg/mL streptomycin in a humidified incubator at 37 • C with 5% CO 2 . The cells were plated into 96-well plates at approximately 5000 cells per well suspended in 100 µL medium. After being cultivated for 24 h, the culture medium was removed, and serial dilutions of the test compounds were treated into each well containing cells in duplicates. After being cultivated for 48 h, the culture medium was removed and 100 µL of MTT solution (0.5 mg/mL) was added to each well and incubated for another 4 h. Following dissolving the MTT formazan crystals, absorbance of the plates was read on a microplate reader at 490 nm for measuring the reduction of the tetrazolium salt MTT (3-(4,5-dimethylthiazol-2yl)-2,5-diphenyltetrazolium bromide) by metabolically active cells. Demethylzeylasteral (DZ) was used as a positive control. IC 50 values were calculated and are presented in the Table 4.
A. niger is a filamentous ascomycete fungus that is ubiquitous in soils, plants, animals, and even in marine environments [36]. Investigations focused on microbial biotransformation of bioactive compounds revealed that A. niger has been considered as a potential biocatalyst for the modification of chemicals to identify undescribed derivatives or chemical intermediates [37,38]. In this study, A. niger demonstrated its ability to catalyze various reactions for isoflavonoids and chalcones including hydroxylation, hydrogenation, epoxidation, hydrolysis, reduction, cyclization, and alkylation reactions. It's worth noting that the metabolic routes were affected by the presence or absence of a linear prenyl group in the substrates. In the presence of a linear prenyl group in substrates 2 and 4, metabolism preferentially took place on the prenyl group by A. niger. Conversely, metabolism took place on ring A or α,β-double bond in substrates 1 and 3 which lack linear prenyl groups. It is hypothesized that presence of the linear prenyl group may be given a higher priority in the regioselectivity rendered by A. niger.
In traditional herbal medicine and oriental clinical practice, licorice has been used as a potential anti-cancer or cancer chemopreventive natural agent [39]. Biological investigations have revealed that licorice extracts show different cytotoxic activities [40][41][42][43]. However, most studies on the effective constituents responsible for these bioactivities are focused on the major compounds such as glycyrrhizin, isoangustone A, glabridin, liquiritigenin, isoliquiritigenin, and licochalcone A [44][45][46][47]. Little is known on the biological effects of the phenolic compounds that have been isolated from licorice. In this study, comparative evaluation on the cytotoxicity of the licorice constituents (1-4) and their metabolites (5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20)(21) has been conducted to investigate structure-cytotoxic activity relationship using three human cancer cell lines A375P, HT-29 and MCF-7. Compound 1 showed potent cytotoxic activities, with IC 50 values ranging from 7.5 to 9.2 µM against the three cancer cell lines tested. However, its metabolite 5 was inactive, indicating that introduction of the hydroxyl group at C-8 of licoisoflavanone could decrease its cytotoxic activity. Meanwhile, compound 2 showed moderate cytotoxic activity whereas its metabolites 6 and 7 were inactive, suggesting that the prenyl group at C-5 position could improve the cytotoxic activities instead of the 2,3-dihydroxy-3-methylbutyl or 2,3-epoxy-3-methylbutyl groups. On the other hand, metabolite 8 showed improved cytotoxic activities compared with its parent compound 3, indicating the importance of the hydroxyl group at C-3 position for retrochalcone. Noteworthily, metabolite 12 showed more potent cytotoxic activities than its parent compound 4 against A375P, HT-29 and MCF-7 cancer cell lines with IC 50 values ranging from 4.4 to 10.1 µM. Whereas other metabolites (10, 11, and 13-21) exhibited reduced cytotoxic activities compared with 4 against the three cell lines tested.
These results generate new ideas for the investigation of cytotoxic constituents from licorice and provide a potential value for the development of more potent inhibitors of tumor promotion.