Preparative Purification of Liriodendrin from Sargentodoxa cuneata by Macroporous Resin

The preparative purification of liriodendrin from Sargentodoxa cuneata using macroporous resin combined with crystallization process was evaluated. The properties of adsorption/desorption of liriodendrin on eight macroporous resins were investigated systematically. X-5 resin was selected as the most suitable medium for liriodendrin purification. The adsorption of liriodendrin on X-5 resin fitted well with the pseudo-second-order kinetic model and Langmuir isotherm model. Dynamic adsorption/desorption tests were performed using a glass column packed with X-5 resin to optimize the separation process of liriodendrin. After one treatment with X-5 resin, the content of liriodendrin in the product was increased 48.73-fold, from 0.85% to 41.42%, with a recovery yield of 88.9%. 97.48% liriodendrin was obtained by further crystallization and determined by HPLC. The purified product possessed strong antioxidant activity. In conclusion, purification of liriodendrin might expend its further pharmacological researches and further applications in pharmacy.

Because of these beneficial pharmacological effects of liriodendrin, large quantities of pure compound are needed as chemical reference standard for further pharmacological researches and further applications in pharmacy. Therefore, many preparative purification methods of liriodendrin are developed, such as high-speed counter-current chromatography [17], HPLC [18], and silica gel column chromatography [10]. But these methods have several disadvantages such as having large amount of organic solvents wastage, being time consuming, having poor yield and recovery, or needing special instruments and are not suitable for large-scale industrial production. Macroporous resins are one kind of durable hydrophilic polymer with advantages of good stability, high adsorption capacity and selectivity, fast adsorption and desorption of mild conditions, easy regeneration, and saving of the cost. The principle of adsorption is based on electrostatic force, hydrogen bonding interaction, and size sieving action between macroporous resins and different molecular from the solution [19]. In recent years, macroporous resins are widely used as medium to enrich and separate the secondary metabolites of traditional Chinese medicine and nature plant, including flavonoids [20,21], coumarins [22], terpenoids [23], lignans [24], and phenolics [25,26]. Liriodendrin after oral administration was in vivo transformation to syringaresinol, which possesses strong antioxidant activity [10]. However, the antioxidant capacity of liriodendrin in vitro has not been investigated recently. As a part of continuing research of expanding its further applications, the antioxidant activity of liriodendrin in vitro was examined.
In this study, a simple method is developed for preparative purification of liriodendrin from the stem of S. cuneata using macroporous resin chromatography combined with crystallization process. By static and dynamic adsorption/desorption tests on resins, some process parameters, such as feed concentration and volume of sample solution, flow rate, ethanol concentration, and volume of desorption solution, are optimized for purification of liriodendrin. After further crystallization, liriodendrin is an obtained pure form and its structure is identified by NMR. Moreover, antioxidant activity of liriodendrin is also evaluated in vitro for the first time.

Materials and Reagents.
Liriodendrin standard was isolated from the stem of S. cuneata in our previous study. The purity was determined to be over 97% by normalization of the peak areas by HPLC. Its structure was elucidated by comparison of the NMR data with the literature [27]. HPLCgrade acetonitrile and formic acid were bought from Concord Technology Co., Ltd. (Tianjin, China). 1,1-Diphenyl-2-picrylhydrazyl (DPPH) and 2,2 -azinobis (3-ethylbenzothiazoline-6-sulphonic acid) diammonium salt (ABTS) were obtained from sigma Chemical Co. (Sigma-Aldrich GmbH, Sternheim, Germany). Other chemicals and reagents were analytical grade. Deionized water was purified by a Milli-Q water purification system (Millipore, Boston, USA).

Adsorbents.
Macroporous resins including D101, AB-8, X-5, ADS-17, and NKA were purchased from Nankai Hecheng S&T Co., Ltd. (Tianjin, China), and HPD100, HPD400, and HPD600 from Cangzhou Bon Adsorber Technology Co., Ltd. (Hebei, China), respectively. Their physical properties were summarized in Table 1. Before use, these resins were soaked in 95% v/v ethanol aqueous solutions for 24 h, then washed with 4% HCl aqueous solution, deionized water, and 4% Adsorption ratio is where is the adsorption capacity at the adsorption equilibrium (mg/g dry resin); is the adsorption ratio (%); is the initial volume of feed solution (mL); 0 and are the initial and equilibrium concentration of liriodendrin in solutions, respectively (mg/mL); and is the weight of the dry resin (g). Desorption ratio is where is the desorption ratio (%); is the concentration of liriodendrin in the desorption solution (mg/mL); is the volume of the desorption solution (mL); , 0 , and are the same as defined above.

Adsorption Kinetics.
The adsorption kinetics of D101 and X-5 resins were studied by mixing 10 mL aqueous solution of S. cuneata extracts (the initial concentration of liriodendrin 2.131 mg/mL) with pretreated resin (1 g dry resin) in each 50 mL Erlenmeyer flask with stopper and shaken (180 rpm) at 25 ∘ C. Liriodendrin concentration in the solution phase was monitored by HPLC at different time intervals until equilibrium. To examine the underlying mechanism of the adsorption kinetic process, equations of pseudo-firstorder model, pseudo-second-order model, and intraparticle diffusion model were used to describe adsorption process [23].
Equation of pseudo-first-order kinetic model is Equation of pseudo-second-order kinetic model is Equation of intraparticle diffusion kinetic model is where and are the adsorption capacity at equilibrium and at any time (mg/g dry resin), respectively; 1 (min −1 ),

Adsorption Isotherms.
The equilibrium adsorption isotherms of liriodendrin on D101 and X-5 resins were studied by the following experiment. D101 or X-5 pretreated resin (1 g dry resin) and 10 mL aqueous solution of S. cuneata extracts (the initial concentration of liriodendrin at 0.213, 0.426, 0.852, 1.279, 1.705, and 2.131 mg/mL) were added into a 50 mL Erlenmeyer flask with stopper, respectively, and shaken (180 rpm) for 8 h at 25 ∘ C. The equilibrium concentration of liriodendrin was analyzed. For further explanation of how molecules or ions of adsorbate interact with adsorbent surface site, the equilibrium adsorption isotherms of D101 and X-5 resins were described by Freundlich and Langmuir equations, respectively. Freundlich equation described the adsorption behavior of the monomolecular layer as well as multimolecular layer and Langmuir equation only described the monolayer adsorption [28,29].
Freundlich equation is Langmuir equation is where is the Freundlich constant, an indicator of adsorption capacity; 1/ is an empirical constant related to magnitude of the adsorption driving force; 0 is the theoretical maximum adsorption capacity (mg/g dry resin); is the Langmuir adsorption equilibrium constant; and are the same parameters as in (1).

Dynamic Adsorption and Desorption Tests.
Dynamic adsorption/desorption tests were performed with a glass column (20 mm × 300 mm) wet-packed of X-5 resin (6 g dry resin). The bed volume (BV) was 30 mL. For the dynamic breakthrough experiment, aqueous solution of S. cuneata extracts was loaded into the glass column at a prescribed flow rate (1, 2, and 4 BV/h), respectively. The liriodendrin concentrations in the effluents that were collected at 10 mL intervals were determined by HPLC.
For dynamic desorption experiment, aqueous solution of S. cuneata extracts was loaded into the resin column mentioned above. After adsorption equilibrium, the resin column was washed first with 2 BV deionized water, and then desorption was conducted with different concentrations of ethanol aqueous solutions (10,20,30,40,50,60,70,80, and 95% v/v) at a flow rate of 1 BV/h successively. The volume of each ethanol aqueous solution was 3 BV and each ethanol desorption solution was analyzed by HPLC. For the quantity of ethanol elution, the test was performed as follows: when the adsorption equilibrium was reached, the resin column was washed first with 2 BV deionized water and then was eluted with 6 BV 40% v/v ethanol aqueous solutions at a flow rate of 1 BV/h. The elution was collected at 1 BV intervals and determined by HPLC. All the 40% v/v ethanol elution was collected and concentrated to dryness under reduced pressure at 60 ∘ C to give refined product.

Crystallization of Liriodendrin.
The refined product was extracted with methyl alcohol-chloroform (1 : 4, v/v) under reflux for 40 min and repeated twice. The extracted solutions were combined and filtered, and the filtrate was concentrated to dryness under reduced pressure at 60 ∘ C. The residue was dissolved in methyl alcohol (1 : 10, w/v). After being kept at 20-30 ∘ C for several days, the solution was centrifuged to obtain crystals of liriodendrin. Then crystals were washed by methyl alcohol.

Antioxidant Activity
2.8.1. DPPH Radical Scavenging Assay. The DPPH radical scavenging activity of samples was evaluated by the method reported [30]. Serial dilutions of the liriodendrin in methanol (1 mL) were added to 4 mL methanol solution of DPPH (0.1 mM/L). After 30 min, the absorbance was measured at 517 nm against a control at room temperature. Vitamin C was used as a positive. The DPPH radical scavenging activity was calculated using the following equation: 2.8.2. ABTS •+ Scavenging Assay. ABTS •+ scavenging assay was performed by the literature method [30]. The ABTS •+ solution was generated by the interaction of ABTS (7 mM/L) and potassium persulfate (2.45 mM/L), stored in the dark at room temperature for 12 h. The ABTS •+ solution was diluted until absorbance was 0.700 ± 0.025 at 734 nm with ethanol. Then, serial dilutions of the liriodendrin in ethanol (1 mL) were added to the ABTS •+ solution (4 mL). After 30 min, the absorbance was measured at 734 nm against a control at room temperature. Vitamin C was used as a positive. The scavenging capability of ABTS •+ was calculated using the following equation:

Resins Screening.
The selection of proper macroporous resins was performed in accordance with the structures and polarities of adsorbates and resins [23]. In the paper, eight macroporous resins of different properties were studied at 25 ∘ C. As shown in Table 2, macroporous resins of nonpolar and weakly polar exhibited better adsorption capacity and desorption ratios than others. This may be caused for special structure of liriodendrin. Although liriodendrin molecule contained polar glucuronyl group, the symmetrical structure of liriodendrin make it possess feature of nonpolar or weakly polar. Therefore, nonpolar and weakly polar resins were applicable to adsorption of liriodendrin. Adsorption capacity and adsorption ratios of D101 and X-5 resins were higher than other resins, and desorption ratios of AB-8, D101, and X-5 resins were similar and higher than other resins. Therefore, D101 and X-5 resins were selected to further evaluate for their properties in the following experiment.

Adsorption Kinetics.
Adsorption kinetics on D101 and X-5 resins were evaluated at 25 ∘ C. Adsorption kinetic curves were shown in Figure 2(a). For the two resins, the adsorption capacities toward liriodendrin increased rapidly in the 30 min, and an asymptotic curve was reached at about 120 min. The fast initial adsorption was likely due to rapid attachment of liriodendrin to the surface of resin and an asymptotic period due to diffusion of liriodendrin into the micropores of resin with high intraparticle mass transfer resistance, which indicated that the behaviors of two resins belonged to the fast adsorption resin type [31]. The parameters of three kinetic models were shown in Table 3(a). According to the calculated correlation coefficient ( 2 ≥ 0.99), the pseudo-second-order kinetic model performed  better to describe the adsorption process on X-5 and D101 resins, which suggested that the adsorption rate was controlled by chemical adsorption mechanism through sharing or exchange of electrons between adsorbate and adsorbent [32]. In terms of the rate constant 2 , the adsorption rate of X-5 and D101 resin was the same.

Adsorption Isotherms.
Adsorption isotherms curves were obtained for liriodendrin on D101 and X-5 resins at 25 ∘ C. As shown in Figure 2(b), adsorption capacities increased with increasing initial concentration, and a turning point was observed when the initial liriodendrin concentration was 1.705 mg/mL. Therefore, this concentration of liriodendrin in sample solution was selected in the following test. Table 3(b) listed the two isotherms equations and relative parameters. The calculated correlation coefficients were shown that the Langmuir model fitted the test data more suitable than the Freundlich model in the studied concentration range, which suggested the monolayer coverage of liriodendrin on the resin [22,28]. The theoretical maximum adsorption capacities 0 on D101 and X-5 resins, determined from the Langmuir equation, were 15.4321 mg/g and 16.7504 mg/g. Therefore, X-5 resin was selected for dynamic adsorption and desorption tests.

Dynamic Breakthrough Curves on X-5 Resin.
Dynamic breakthrough tests were studies based on the feed concentration, the volume of effluent, flow rate, and temperature. The initial feed concentration of liriodendrin was 1.705 mg/mL (from Section 3.3) and the temperature at 25 ∘ C, and then the effect of flow rate was investigated. As shown in Figure 3(a), the different flow rates showed noticeable different breakthrough volume. When the flow rate was 1 BV/h, the best adsorption performance could be obtained. The breakthrough volume of liriodendrin on X-5 resin was 55 mL (about 2 BV) at a flow rate of 1 BV/h, when the liriodendrin concentration in effluents reached 1% of the initial concentration [18]. The adsorption capacity of liriodendrin on X-5 resin was 15.630 mg/g dry resin. The amount of refined sample was obtained from 55 g raw material in the X-5 resin treatment. b The amount of liriodendrin was obtained from 1000 mg refined sample by crystallization.

Dynamic Desorption on X-5 Resin.
In the process of dynamic desorption, the concentration and the volume of desorption solution were investigated. Dynamic desorption was performed with gradient and isocratic elution modes at the flow rate of 1 BV/h, respectively. As shown in Figure 3(b), the desorption ability increased by the increase of ethanol concentration. When the concentration of ethanol reached 40% v/v, the liriodendrin absorbed by X-5 resin could be fully eluted. Therefore, 40% v/v ethanol aqueous solution was selected as the eluent in isocratic elution. The isocratic elution result showed that the concentration of liriodendrin reached maximum values when eluting with 2 BV of 40% v/v ethanol aqueous solutions. When the volume of elution reached 5 BV, liriodendrin could be completely desorbed from X-5 resin (Figure 3(c)). Therefore, the volume of desorption solution was selected to be 5 BV. Under the above optimized condition, all the 40% v/v ethanol elution was collected and concentrated to dryness under reduced pressure at 60 ∘ C to give refined product.

Purity Analysis and Identification of Liriodendrin.
After purification on X-5 resin, the purity of liriodendrin increased from 0.85% of S. cuneata extracts to 41.42%, and the liriodendrin of recovery rate was 88.9%. Such crystallization resulted in a product with 97.48% purity and the recovery rate was 63.7% (Table 4). As shown in Figures 4(a) and 4(b), the HPLC chromatograms of S. cuneata extracts before and after treatment on X-5 resin were compared. Figure 4(c) showed that the purity of liriodendrin crystals was determined to be 97.48% by HPLC. For further chemical structure identification of liriodendrin, NMR analysis was performed and the data was shown as follows: study, the antioxidant activity of liriodendrin isolated from S. cuneata was analysed using DPPH and ABTS. DPPH radical scavenging activity test aims to measure the capacity of sample to scavenge the stable radical DPPH formed in solution by donation of hydrogen atom or an electron [33]. As shown in (Figure 5(a)), the IC 50 value of liriodendrin is 13.61 g/mL, which is less important than that of vitamin C (IC 50 = 6.98 g/mL). ABTS •+ scavenging assay is a widely accepted model to determine the total antioxidant activity. As shown in (Figure 5(b)), the IC 50 values of liriodendrin and vitamin C are 168.42 and 23.29 g/mL, respectively.
The ABTS •+ scavenging assay revealed that liriodendrin possessed weak ABTS •+ scavenging capacities. However, it has been suggested that liriodendrin plays an important role in DPPH inhibitory activities and oxygen radical absorbance capacities. On the other hand, liriodendrin after oral administration was in vivo transformation to syringaresinol, which possesses strong antioxidant activity [10]. The antioxidant activity of liriodendrin in vitro and vivo needed to be further researched.

Conclusions
In the study, highly concentrated liriodendrin was successfully prepared. After S. cuneata extracts were purified on X-5 resin, purity of liriodendrin was 41.42%. Further crystallization resulted in a product with 97.48% purity and the recovery rate was 63.7%. The results indicated that the established method was simple and effective, thus showing potential for industrial scale isolation of liriodendrin from Sargentodoxa cuneata in the future. Also, such production of highly concentrated liriodendrin as an antioxidant agent might expand its further applications both in industrial production and in pharmacy.