Application of Centrifugal Partition Chromatography for Bioactivity-Guided Purification of Antioxidant-Response-Element-Inducing Constituents from Atractylodis Rhizoma Alba

Activity-guided separation of antioxidant response element (ARE)-inducing constituents from the rhizomes of Atractylodis Rhizoma Alba was performed by the combination of centrifugal partition chromatography (CPC) and an ARE luciferase reporter assay. From 3 g of the active n-hexane fraction, one polyacetylene, (6E,12E)-tetradeca-6,12-dien-8,10-diyne-1,3-diyl diacetate (47.3 mg), and two sesquiterpenes, atractylenolide I (40.9 mg), and selina-4(14),7(11)-dien-8-one (6.0 mg) were successfully isolated by CPC with n-hexane–ethyl acetate–methanol–water (8:2:8:2, v/v). The chemical structures of the isolated compounds were determined by 1H- and 13C-NMR and ESI-MS. Among the isolated compounds, (6E,12E)-tetradeca-6,12-diene-8,10-diyne-1,3-diol diacetate and selina-4(14),7(11)-dien-8-one increased ARE activity 32.9-fold and 16.6-fold, respectively, without significant cytotoxicity, when 5 µM sulforaphane enhanced ARE activity 27.1-fold. However, atractylenolide I did not increase ARE activity at 100 µM, and showed cytotoxicity at concentrations over 10 µM.


Introduction
Oxidative stress is involved in various pathological conditions including ageing, cancer, inflammation, and chronic diseases [1]. Reactive oxygen species can be scavenged by the use of antioxidants including vitamins, carotenoids, and polyphenolic compounds [2]. Induction of antioxidant signaling pathway via upregulation of endogenous antioxidant/phase II detoxifying enzymes may be an alternative defensive approach. The activation of nuclear factor erythroid 2-related factor 2 (Nrf2)-antioxidant response element (ARE) pathway leads to the induction of antioxidant/phase II detoxifying enzymes including heme oxygenase 1, NAD(P)H: quinine oxidoreductase 1, and glutathione S-transferase. There is growing interest in molecules that activate the Nrf2-ARE signaling pathway [3].
Atractylodis Rhizoma Alba (the rhizome of Atractylodes macrocephala Koidz. or A. japonica Koidz. ex Kitam.) is a traditional medicine in Korea, Japan, and China. In Korea, the aerial parts of Atractylodis Rhizoma Alba are eaten as vegetables, while the rhizomes are used as medicinal herbs. According to According to Korean traditional medicine publications, Atractylodis Rhizoma Alba has warm properties and a bitter taste; it strengthens the stomach and spleen, resolves dampness, and stops sweating [4,5].
The separation and purification of the bioactive molecules from natural products requires repeated chromatography, and in some cases the active ingredient does not elute by absorption to the solid phase [13]. To overcome these problems, centrifugal partition chromatography (CPC) is a viable alternative method. CPC is a preparative, support-free, liquid-liquid chromatographic technique, which is an efficient tool for bioactivity-guided isolation of natural ingredients that allows increased sample loading and the absence of sample loss due to irreversible sample absorption to a solid column [14][15][16][17][18].

ARE Luciferase Assay and HPLC Analysis of the Crude Extract and Sub-Fractions
Based on the ARE luciferase assay, the n-hexane fraction was the most active, followed by the ethyl acetate fraction. At a concentration of 30 µg/mL, the n-hexane and ethyl acetate fractions enhanced ARE activity 27.1 and 17.1-fold, respectively. Sulforaphane, the positive control, increased ARE activity 49.9-fold at a concentration of 5 µM (Figure 2A).
HPLC analysis of crude samples revealed that compounds 1-3 were major constituents of A. macrocephala in the n-hexane fraction ( Figure 2B). Based on its ARE-enhancing activity and HPLC chromatogram, the n-hexane fraction was selected to purify active molecules by CPC.

ARE Luciferase Assay and HPLC Analysis of the Crude Extract and Sub-Fractions
Based on the ARE luciferase assay, the n-hexane fraction was the most active, followed by the ethyl acetate fraction. At a concentration of 30 µg/mL, the n-hexane and ethyl acetate fractions enhanced ARE activity 27.1 and 17.1-fold, respectively. Sulforaphane, the positive control, increased ARE activity 49.9-fold at a concentration of 5 µM (Figure 2A).
HPLC analysis of crude samples revealed that compounds 1-3 were major constituents of A. macrocephala in the n-hexane fraction ( Figure 2B). Based on its ARE-enhancing activity and HPLC chromatogram, the n-hexane fraction was selected to purify active molecules by CPC.

Optimization of the Two-Phase Solvent System and CPC Operation
A two-phase solvent system with petroleum (60-90 °C)-ethyl acetate-ethanol-water (4:1:4:1, v/v) has previously been reported for the separation of actractylon and atractylenolide III from A. macrocephala by high-speed counter-current chromatography [19]. Based on this result, several twophase solvent systems consisting of n-hexane-ethyl acetate-methanol-water at different ratios were tested to calculate K-values for major constituents of the n-hexane fraction (Table 1). CPC was performed with a solvent ratio of 8:2:8:2 (v/v), although the K-value was better at a ratio of 9:1:9:1 (v/v) for the three major compounds 1-3 in the n-hexane fraction. The peak at 28.6 min, a major component in ethyl acetate fraction, was also expected to exert ARE-inducing activity ( Figure  1). Therefore, this peak is more polar than compounds 1-3, which is why the two-phase solvent system with 8:2:8:2 was more suitable for polar conditions than the 9:1:9:1 system.
The retention time of each compound 1-3 in CPC was calculated using a general chromatographic retention equation (VR = VM + KCVS, VR: the retention volume of the solute, VM: the mobile phase volume, KC: the partition coefficient, VS: the stationary volume) [20]. In this experiment, SCPC-1000 was used (column volume: 1000 mL, flow rate: 10 mL/min, mobile phase retention: 54%). Therefore, the calculated retention time for each peak was 102.16 min (peak A), 140.24 min (peak B), and 251.12 min (peak C).

Optimization of the Two-Phase Solvent System and CPC Operation
A two-phase solvent system with petroleum (60-90 • C)-ethyl acetate-ethanol-water (4:1:4:1, v/v) has previously been reported for the separation of actractylon and atractylenolide III from A. macrocephala by high-speed counter-current chromatography [19]. Based on this result, several two-phase solvent systems consisting of n-hexane-ethyl acetate-methanol-water at different ratios were tested to calculate K-values for major constituents of the n-hexane fraction (Table 1). CPC was performed with a solvent ratio of 8:2:8:2 (v/v), although the K-value was better at a ratio of 9:1:9:1 (v/v) for the three major compounds 1-3 in the n-hexane fraction. The peak at 28.6 min, a major component in ethyl acetate fraction, was also expected to exert ARE-inducing activity ( Figure 1). Therefore, this peak is more polar than compounds 1-3, which is why the two-phase solvent system with 8:2:8:2 was more suitable for polar conditions than the 9:1:9:1 system.
The retention time of each compound 1-3 in CPC was calculated using a general chromatographic retention equation (V R = V M + K C V S , V R : the retention volume of the solute, V M : the mobile phase volume, K C : the partition coefficient, V S : the stationary volume) [20]. In this experiment, SCPC-1000 was used (column volume: 1000 mL, flow rate: 10 mL/min, mobile phase retention: 54%). Therefore, the calculated retention time for each peak was 102.16 min (peak A), 140.24 min (peak B), and 251.12 min (peak C).
As shown in Figure 3, the n-hexane fraction (3 g) separated excellently into three major response CPC peak fractions: A (90-110 min), B (120-146 min), and C (178-227 min) ( Table 2). The first two peaks (A and B) eluted as expected, but the third peak (C) eluted earlier than the calculated time.
This difference is due to the gradual outflow of the stationary phase, which can be observed by a reduction in the operating pressure ( Figure 3A).
peaks (A and B) eluted as expected, but the third peak (C) eluted earlier than the calculated time. This difference is due to the gradual outflow of the stationary phase, which can be observed by a reduction in the operating pressure ( Figure 3A).  Each CPC peak fraction from A to C was concentrated to yield 47.3 mg of peak A, 40.9 mg of peak B and 6.0 mg of peak C. Afterward, the fractions obtained from the CPC chromatogram were analyzed via analytical HPLC. Fraction peaks A-C from the CPC with a descending mode corresponded to the peak order of 2, 1 and 3 in the HPLC chromatogram in Figure 3. The purities of atractylenolide I (1), (6E,12E)-tetradeca-6,12-dien-8,10-diyne-1,3-diyl diacetate (2) and selina-4(14),7(11)-dien-8-one (3) were demonstrated to be >99%, 95% and 97%, respectively, according to the area of the HPLC peaks at 254 nm ( Figure 3B).

Apparatus and Materials
CPC was performed on an Armen fully integrated SCPC-100+1000 CPC spot instrument (Armen Instruments, St-Ave, France). This instrument is a fully automated system consisting of a CPC column compartment (1000 mL rotor made of 21 stacked disks with a total of 1512 twin cells), a pump, an injector, a UV/vis detector, a fraction collector, a digital screen flat PC, and Armen Glider CPC software. The HPLC analysis was performed by an Agilent 1260 HPLC system (Agilent Technologies, Palo Alto, CA, USA): G1312C binary pump, a G1329B autosampler, a G1315D DAD detector, a G1316A column oven and ChemStation software. HPLC-grade solvents were purchased from Fisher Scientific (Pittsburgh, PA, USA). All other organic solvents used for extraction and CPC operation were analytical grade and obtained from Daejung (Gyonggi-do, Korea).

Apparatus and Materials
CPC was performed on an Armen fully integrated SCPC-100+1000 CPC spot instrument (Armen Instruments, St-Ave, France). This instrument is a fully automated system consisting of a CPC column compartment (1000 mL rotor made of 21 stacked disks with a total of 1512 twin cells), a pump, an injector, a UV/vis detector, a fraction collector, a digital screen flat PC, and Armen Glider CPC software. The HPLC analysis was performed by an Agilent 1260 HPLC system (Agilent Technologies, Palo Alto, CA, USA): G1312C binary pump, a G1329B autosampler, a G1315D DAD detector, a G1316A column oven and ChemStation software. HPLC-grade solvents were purchased from Fisher Scientific (Pittsburgh, PA, USA). All other organic solvents used for extraction and CPC operation were analytical grade and obtained from Daejung (Gyonggi-do, Korea).

Plant Material and Preparation of Crude Extracts
The dried rhizomes of A. macrocephala were purchased from the Kyungdong oriental herbal market, Seoul, Korea, in October 2015, and identified by one of author (Dr. CY Kim). A voucher specimen was deposited in the Herbarium of the College of Pharmacy, Hanyang University (HYUP-AM-001). The dried rhizomes (309 g) were ground and extracted in 4 L methanol three times for 3 h under reflux. The methanol solution was evaporated by rotary evaporator to obtain 46 g of extract. The extract was suspended in water and successively fractionated by n-hexane, ethyl acetate, and n-butanol. The fractions obtained were 9.78 g of n-hexane, 6.08 g of ethyl acetate, and 2.92 g of n-butanol.

Selection of the Two-Phase Solvent System
The two-phase solvent system was selected according to the K-values of the target compounds as a series of solvent systems of n-hexane-ethyl acetate-methanol-water. Briefly, approximately 2 mg of the sample was added to each test tube, 2 mL of each phase of a pre-equilibrated two-phase solvent system was added, and the contents were thoroughly mixed. After equilibration, 10 µL of the upper and lower phases were analyzed by HPLC at 254 nm. The K value was calculated as the peak area of each compound in the upper stationary phase divided by that of the lower mobile phase.

CPC Separation Procedure
According to the partition coefficient, n-hexane-ethyl acetate-methanol-water (8:2:8:2, v/v) was used as the two-phase solvent system for CPC separation. The two-phase system was prepared in a 4000-mL separation funnel by adding 1200 mL of n-hexane, 300 mL of ethyl acetate, 1200 mL of methanol, and 300 mL of water (volume ratio 8:2:8:2). The separation funnel was shaken vigorously and the solvent mixture was thoroughly equilibrated at room temperature. The upper and lower phases were separated and degassed by sonication for 1 h before use.
First, the rotor was filled with the stationary phase (upper phase), and then the rotor was set at 1100 rpm and the lower, more hydrophilic phase was pumped into the channel at a flow rate of 10 mL/min in descending mode. When the stationary phase and mobile phase in the rotor were equilibrated (stationary phase retention: 54%), 3 g of the n-hexane fraction dissolved in the 10-mL mixture of stationary phase and mobile phase (1:1, v/v) was injected. The effluent was monitored at 254 nm and in scan mode (240-400 nm). Each peak fraction was collected according to the chromatogram to obtain three CPC peak fractions A-C. The fractions were concentrated and analyzed by HPLC.

HPLC Analysis and Identification of the CPC Peak Fractions
The crude extract and each peak fraction obtained by CPC were analyzed by HPLC with an INNO C18 analytical column (4.6 × 250 mm, 5 µm). The mobile phase was composed of acetonitrile containing 0.1% formic acid (A) and water containing 0.1% formic acid (B). The gradient elution conditions were as follows: initial 0

Structural Identification
Identification of the CPC purified compounds was performed by ESI-MS with an Advion compact mass (Advion, Ithaca, NY, USA) and NMR. The ESI-MS spectra conditions were as follows: positive ion mode; mass range, m/z 100-1200; capillary temperature, 200 • C; capillary voltage, 150 V; source voltage offset, 30; source voltage span, 10; source gas temperature, 150 • C; and ESI voltage, 3500 V. 1 H-NMR (400 MHz) and 13 C-NMR (100 MHz) were measured on a Bruker model digital Avance III 400 NMR in CDCl 3 . The NMR spectra were processed by the MestReNova 9.0 software (Mestrelab Research, Santiago de Compostela, Spain).

Cell Culture and Cell Viability
Human hepatoma (HepG2) cells were obtained from American Type Culture Collection (ATCC, Manassas, VA, USA) and cultured in a DMEM medium supplemented with 10% FBS and 1% penicillin-streptomycin in a humidified atmosphere containing 5% CO 2 in air at 37 • C. For determination of cell viability, 4 × 10 4 cells/well were seeded in 96-well plates and incubated overnight. Cells were treated for 12 h with various concentrations of the isolated compounds 1-3. Samples were dissolved in DMSO and the final DMSO concentration did not exceed 0.1%, which did not affect cell cytotoxicity. Cells were then exposed to 0.5 mg/mL MTT for an additional 1 h at 37 • C. The media was removed and 100 µL of dimethylsulfoxide was added to dissolve the formazan crystals produced in each well. Absorbance was measured at 570 nm using an Infinite M200 PRO microplate reader (Tecan, Salzburg, Austria).

ARE Luciferase Reporter Assay
Construction of the HepG2-ARE cells (transfected Pgl4.37 [luc2P/ARE/ Hygro] (Promega)) was reported previously [23]. HepG2-ARE cells were cultured in DMEM high glucose media supplemented with 10% FBS, 1% penicillinstreptomycin and 1% hygromycin B. HepG2-ARE cells were plated at a density of 1 × 10 5 cells/well in 24-well plates and allowed to reach 70-80% confluence. After 12 h of serum starvation, cells were incubated with the extract, sub-fractions or isolated compounds 1-3 for 12 h. After treatment, cells were washed twice with ice-cold phosphate-buffered saline, and 120 µL of passive lysis buffer (Promega, Madison, WI, USA) was added to each well and reacted at 4 • C for 30 min. The supernatant was transferred to an e-tube and centrifuged at 13,000 rpm for 15 min at 4 • C, and 60 µL of the supernatant was incubated with 30 µL of the luciferase assay substrate (Promega, Madison, WI, USA) in a 96-well plate. Luminescence was measured by an EnSpire multimode plate reader (PerkinElmer, Waltham, MA, USA). Sulforaphane (5 µM) (Calbiochem, Darmstadt, Germany) was used as a positive control.

Statistical Analysis
All data are reported as means ± S.E. The statistical significance of differences between treatments was assessed using the Student's t-test. Probability values less than 0.01 were considered significant.

Conflicts of Interest:
The authors declare no conflict of interest.