Preparative enrichment and separation of isofraxidin from Acanthopanax senticosus with HPD 100 C macroporous resin

Isofraxidin is known as the major pharmacological component in Acanthopanax senticosus. In the present study, in order to screen a suitable resin for the preparative separation and purification of isofraxidin, the adsorption and desorption properties of 26 widely used commercial macroporous resins were evaluated. The preliminary research results indicated that HPD100C and HPD300 have better properties. By comparing static adsorption kinetics of HPD100C and HPD300, HPD100C was selected as the most suitable resin. Langmuir and Freundlich isotherms models were used to describe the interaction between solutes and resins at 25, 30, 35°C, respectively, and the equilibrium experimental data fit better to the Langmuir isotherms. Dynamic adsorption and desorption tests have been carried out to optimize the process parameters. The optimal conditions were as follows: For adsorption, the concentration of isofraxidin in sample solution: 0.0436 mg/ml, processing volume: 24 BV, flow rate: 2 BV/h, temperature: 25°C; for desorption, ethanol–water solution: 60:40 (v/v), eluent volume: 4 BV, flow rate: 3 BV/h. Under the conditions aforementioned, the content of isofraxidin increased 23.92-fold and its recovery was 93.79 ± 0.44%. In conclusion, HPD100C demonstrated the best ability for separating isofraxidin among the 26 tested resins.


INTRODUCTION
Radix Acanthopanax senticosus (RAS), is called "Siberian ginseng" or "Eleutherococcus senticosus", and has a name of Ciwujia in Chinese.It is the roots and rhizomes of A. senticosus that belongs to the Araliaceae family.RAS has been used extensively in China, Russia, Korea and Japan as an adaptogen (Brekhman and Dardymov, 1969;Davydov and Krikorian, 2000).Nowadays, there are some RAS products on the market of many countries, including drugs and healthy food (Davydov and Krikorian, 2000;Weng et al., 2007).In China, it is common for people using RAS to make medicinal liquor or to stew pigs, milk cows and pigeons.In vitro and in vivo studies have demonstrated that RAS possesses many pharmacological effects, such as anti-stress, anti-fatigue, immuno-enhancing effect and anti-depressive effect, etc (Weng et al., 2007;Deyama et al., 2001).Among the active components in RAS and its extracts, isofraxidin (Figure 1) is the remarkable one and has been reported to possess strong biological activity (Zhang et al., 2002;Wang et al., 2003).
According to previous research results, isofraxidin is the major component attribute to the pharmacological effects of RAS.Isofraxidin was reported to exhibit anti-fatigue, anti-stress, and immuno-accomondating effects (Sun et al., 2007).Further investigations indicated that isofraxidin has the ability of diminishing inflammation (Yamazaki et al., 2004) and anbiosis (Liu et al., 2004a).Meanwhile, several studies have suggested that isofraxidin has anti-bacterial (Liu et al., 2004b) and anti-tumor effect  (Yamazaki and Tokiwa, 2010).Due to the fact that natural isofraxidin has low content in RAS and the isolation was difficult, therefore, the development of an efficient separation method turns to be of extreme importance.Although, some methods have been used to separate isofraxidin from RAS, such as liquid-liquid partition (Méndez, 1978;Yang et al., 2010), column chromatographic procedures involving cellulose (Méndez, 1978) and high-speed counter-current chromatography (Xiao et al., 2009) etc.
However, due to their relative low handling capacity in one cycle, these separation methods are inefficient.They are also troubled with various other disadvantages, such as low recovery, large solvent consuming, high labor intensities and operation cost, or even safety problems.Those deficits make them unsuitable for large-scale industrial production (Fu et al., 2006;Jia and Lu, 2008).As a kind of adsorbent, macroporous resin can be used to selectively adsorb constituents from aqueous solution through hydrogen bonding forces and van der Waals forces, etc (Pi et al., 2008;Jiang et al., 2006;Gao et al., 2007).
Moreover, in the previous study (He et al., 2007), 12-1 macroporous resin was tested for the separation of isofraxidin from Sarcandra glabra extract, the resin was not optimized and the adsorption capacity was not satisfactory, the parameters for static adsorption and desorption tests were not optimized.Thus, in this study, we have focused on investigating the adsorption and desorption properties of 26 different resins on separating and purifying isofraxidin form RAS.Then, based on the experimental results, the most suitable resin was selected and used to develop an efficient and simple method of separating and purifying isofraxidin in large scale from RAS.

Chemical and reagents
Radix Acanthopanax senticosus (RAS) was obtained from Sankeshu Medicinal Materials Market (Harbin, China), and was minced into little pieces and sieved through 20 to 40 meshes before use.Isofraxidin standard (95% purity) was purchased from National Institute for the Control of Pharmaceutical and Biological Products (Beijing, China).Acetonitrile of chromatographic grade was purchased from J and K Chemical Limited (Beijing, China).Deionized water was freshly prepared by a Milli-Q water-purification system (Millipore, Bedford, MA, USA) and used in all experiments.Other reagents are analytical grade and were purchased from Beijing Chemical Reagents Co. (Beijing, China).All solvents prepared for HPLC were filtered through 0.45 µm nylon membrane and degassed under ultrasonication before use.

Adsorbents
Macroporous resins tested were purchased from Cangzhou Bon Adsorber Technology Company Limited (Cangzhou, China).Their physical properties are listed in Table 1.In order to remove the monomers and porogenic agents trapped inside the pores of macroporous resin during synthesis process, the adsorbent beads were pretreated with the following procedure: first, the resins were soaked in ethanol for 24 h, and then washed with deionized water by circumfluence until there is no residue of ethanol.The treated resins were stored in a desiccator with deionized water in order to maintain constant moisture content.Prior to use, the resins were wet with ethanol again and then thoroughly replaced with deionized water (Apers et al., 2005;Jung et al., 2001).In order to obtain moisture contents of macroporous resins, the resins were accurately weighed in glass dishes, and dried to a constant weight in a digital blast oven (Shanghai Boxun Industry and Commerce Co., Ltd., Shanghai, China) at 105°C.Their moisture contents were calculated and were also shown in Table 1.

HPLC analysis of isofraxidin
A waters liquid chromatograph (Waters Corporation, Milford, MA, USA), consisting of a Waters 600 Controller equipped with a Waters 717 plus autosampler, and a Waters 2487 UV detector was used to determine isofraxidin.Chromatographic separation was carried out on a Kromasil C18 reversed-phase column (5 µm diameter particles，4.6mm × 250 mm I.D., Kromasil).The mobile phase was acetonitrile-water-formic acid (15:84.9:0.1, v/v/v).The detection wavelength was 205 nm, the flow rate was 1 ml/min, the injection volume was 10 µL, and the column temperature was maintained at 25°C.The regression line for isofraxidin was Y = 20021348X-139160 (R 2 =0.9993, n = 8), where Y was the peak area of isofraxidin and X was the concentration of isofraxidin (mg/ml).

Preparation of crude RAS extracts
The minced RAS powder (2000 g) was extracted with deionized water (12 L) under reflux for 30 min each time, repeated three times.
The extract solution was purified by membrane filtration and then evaporated in rotary vaporization (RAS-52AA, Shanghai Huxi Instrument, and Shanghai, China) to dryness under vacuum condition.The dry extract was stored at 4°C.The contents of isofraxidin in the extract were 0.0533%.

Static adsorption and desorption properties of the resins
The static adsorption tests of RAS extract were performed as follows: 0.5 g resins (dry weight basis) were added into a conical flask with a lid and then 100 ml aqueous sample solutions with known concentrations prepared in section 2.3 were introduced.The flasks were then shaken (100 rpm) for 8 h at 25°C in a constant temperature oscillator (Harbin Donglian Electronic and Technology Development Company Limited Harbin, China).The initial concentration of sample solutions and their concentrations post adsorption were analyzed by HPLC.The static desorption process was carried out as follows: after reaching adsorption equilibrium, the residual solution was removed.The adsorbate-laden resins were first washed in 100 ml deionized water, shaken (100 rpm) for 2 h at 25°C.Then, they were desorbed in 25 ml ethanol-water (95:5, v/v) solution.The flasks were shaken (100 rpm) for 2 h at 25°C.Desorption solutions were also analyzed by HPLC.The suitable resin was selected based on its adsorption capacity and desorption ratio.

Static adsorption kinetics on HPD100C and HPD300
The static adsorption kinetics of isofraxidin on the preliminarily selected resins, HPD100C and HPD300, were also studied using the same procedure as previously described.The respective concentrations of isofraxidin in the sample solutions were monitored by HPLC at predetermined time intervals until equilibrium.

Adsorption isotherms
The adsorption isotherm of isofraxidin on the optimum resin, HPD100C was investigated by contacting 100 ml of sample solutions at different concentrations with pre-weighed resins 0.5 g (dry weight basis) in the shaker bath (100 rpm) for 8 h at 25, 30 and 35°C.The initial and equilibrium concentrations were determined by HPLC.The equilibrium adsorption isotherm on the resin was obtained, and their degrees of fitness to the Langmuir equation and Freundlich equation were evaluated.

Dynamic adsorption and desorption
Dynamic adsorption and desorption experiments were carried out in a glass columns (12 mm × 500 mm) (Tianjin Tianbo Glass Instrument Co., Ltd., Tianjin, China) wet-packed with 5 g (dry weight where mean ± S.D., n=3; "-" indicates that there is no signal in HPLC profiles. basis) HPD100C.The bed volume (BV) and the length of the resin were 25 ml and 10 cm, respectively.In all cases, the sample solutions were flowed downward.Sample solutions were flowed through the glass column at certain flow rate, and the isofraxidin concentration was monitored by HPLC analysis of the effluent liquid collected at 50 ml intervals.Breakthrough point was indicated.Once adsorption reached equilibrium, the loading of the sample solution was stopped.The adsorbate-laden columns were first washed with deionized water and then with different ethanol-water solutions (30:70, 40:60, 50:50, 60:40, 70:30, 80:20, 90:10, v/v), at the same flow rate.The isofraxidin concentration in the effluent solution was determined by HPLC analysis of the effluent liquid collected at 5 ml intervals.The effluent solution were concentrated and dried under vacuum before further analyses.The dynamic adsorption and desorption capacities of HPD100C, the recoveries and the isofraxidin content in the product were calculated.

Dynamic adsorption and desorption capacity, the recovery equation
The following equations were used to quantify the adsorption and desorption capacity and the desorption ratio as well as the recovery.Adsorption evaluation is calculated as: where Qe is the adsorption capacity at adsorption equilibrium (mg/g resin); C0 and Ce are the initial and equilibrium concentration of isofraxidin in the solutions (mg/ml); Vi is the volume of the initial sample solution (ml) and W is the weight of the dry resin (g).Desorption evaluation is calculated as: Where D is the desorption ratio (%); Qd is the desorption capacity after adsorption equilibrium (mg/g resin) ;Cd is the isofraxidin concentration in the desorption effluent solutions(mg/mL); Vd is the volume of the desorption solutions; C0, Ce, Vi and W the same as those defined in Equation (1).The following equation was to calculate the recovery.
where R is the recovery (%), M is the weight of a targeted component laden onto the selected adsorbent; m is weight of the targeted component in product.

Langmuir equation and Freundlich equation
The equilibrium experimental data were fitted to the Langmuir and Freundlich isotherms equations (Fu et al., 2007(Fu et al., , 2008;;Jung et al., 2001) to describe the adsorption behavior between solute and resin.The Langmuir isotherms equation is given as: The aforementioned equation can be rearranged to the following linear form: Where Qmax is the theoretically calculated maximum adsorption capacity (mg/g resin); k and KL are constant; Qe is the adsorption capacity at adsorption equilibrium (mg/g resin) and Ce is the equilibrium concentrations of solutes in the solutions.The Freundlich isotherms equation is given as:

⋅ = (7)
A form of Equation ( 7) can be written as: where KF is constant, an indicator of adsorption capacity; 1/n is an empirical constant related to the magnitude of the adsorption driving force; Qe and Ce are the same as those defined in Equations ( 5) and ( 6).

Adsorption and desorption properties of the resins
As shown in Table 1, the adsorption and desorption properties of different resins are distinct.The adsorption capacities of HPD100C and HPD300 were higher than those of other resins.In general, the adsorption capacity of a resin is correlated with its surface area.HPD100C and HPD300 have larger surface area than other resins, which may account for their higher adsorption capacity for isofraxidin.Therefore, due to their high surface area and similar polarity with isofraxidin, HPD100C and HPD300 owned better adsorption capacity.
On the other hand, HPD100C and HPD300 also showed higher desorption capacity and desorption ratio for isofraxidin than others.It may be due to the fact that the affinity between isofraxidin and resin is mainly the physical force, such as the van der Waals force, which has a low power for holding the isofraxidin on HPD100C and HPD300.Moreover, even though a big average pore diameter may bring in high desorption ratio, they also incur low adsorption and desorption capacity.Thus, due to the affinity and the proper average pore diameter, HPD100C and HPD300 had better desorption capacity and desorption ratio.
In addition, the ranking in Table 1 appears to reflect the different physicochemical properties of these resins.Compared with other resins, HPD100C and HPD300 owned better adsorption and desorption capacity and higher desorption ratio for isofraxidin.Therefore, HPD100C and HPD300 were selected for further study in the following adsorption kinetics experiments.

Static adsorption kinetics on HPD100C and HPD300
The static adsorption and desorption experiments are not enough for evaluating the performance of an adsorbent.A more suitable resin must also have a higher adsorption rate.Therefore, the adsorption kinetics curves were obtained for isofraxidin on HPD100C and HPD300.As can be seen in Figure 2, the adsorption capacity increased with the increasing of the adsorption time, reaching equilibrium at about 3 h on HPD100C and 4 h on HPD300, respectively.In Figure 2, it is apparent that HPD100C has a better adsorption rate than HPD300.Thus, HPD100C was selected for further study in the following tests.

Adsorption isotherms
The adsorption isotherms of HPD100C were investigated with different concentrations of sample solutions at 25, 30, 35°C.The initial concentrations of isofraxidin were 0.0077 0.0144, 0.0190, 0.0297, 0.0436 and 0.0537 mg/ml, respectively.As shown in Figure 3, the adsorption capacities increased with the increasing of the initial solute concentrations, and reached the saturation plateau when the initial concentration of isofraxidin was 0.0436 mg/ml.Equilibrium data gives information about the affinity between solute and adsorbent.The Langmuir isotherm model and the Freundlich isotherm model are the two best-known and the most often-used isotherms model for the adsorption of solute from solution.The Langmuir and Freundlich parameters are summarized in Table 2.As can be seen in Table 2, for Freundlich equation, the adsorption can easily take place when 1/n value is  between 0.1 and 0.5 and tends not to happen when 1/n value is between 0.5 and 1, and is almost unable to occur when 1/n value exceeds 1 (Liu et al., 2008).In isofraxidin's Frendlich equation, all 1/n values are between 0.5 and 1.Thus, even though the R 2 values of isofraxidin's Freundlich equation were rather high, it is hard for the adsorption of isofraxidin on HPD100C to happen.On the other hand, the R 2 values of isofraxidin's Langmuir equation were all above 0.98.Thus, the Langmuir equation can describe the adsorption and desorption behavior of isofraxidin on HPC100C.
Langmuir isotherm model is based on the assumption that the sorbate form is only a single layer (Campbell and Davies, 1995), thus, the results in Table 2 showed that it is the single layer form adsorption between particles of isofraxidin and the surface of HPD100C.As shown in Figure 3, at the same initial concentration, the adsorption capacities decreased with the increasing of the temperature in the investigated temperature range, which indicated that the adsorption was a thermo positive process.Meanwhile, Q max also decreased with increasing temperature for isofraxidin (Table 2).Therefore, 25°C was selected in the following experiments.

Dynamic leakage curves on HPD-100C resin
The initial concentration of isofraxidin in this test was 0.0436 mg/ml, and the flow rates investigated in this test were 2, 3 and 4 BV/h, respectively.The dynamic leakage curves on HPD-100C resin were obtained based on the effluent volume and the concentration of solute in sample solution.As can be seen in Figure 4, the isofraxidin exhibited better adsorption performance at the flow rate of 2 BV/h, this may be due to a better particle diffusion in sample solution.Therefore, 2 BV/h was selected in the following tests.For breakthrough point, when the concentration in leak solution is 10% of the initial concentration, adsorption is presumed to have reached saturation, when the adsorption affinity decreases, or even disappears, and then the solutes leak from the resin.Thus, breakthrough point is usually defined as the time when the leak solution concentration is equal to 10% of the initial concentration.It is important to set up the

Concentration (mg/ml)
Volume (BV) Figure 5. Dynamic desorption curves of isofraxidin on a column packed with HPD100C.
leakage curve.In the present case, the breakthrough point of isofraxidin appeared at a processing volume of sample solution around 24 BV.

Effect of ethanol-water solution on desorption tests
In order to choose proper desorption solution, different ethanol-water solutions (30:70, 40:60, 50:50, 60:40, 70:30, 80:20, 90:10, v/v) were used to perform desorption tests.As can be seen in Table 3, on one hand, with the increase of the ethanol-water solution, the desorption mass of isofraxidin has increased markedly until the ethanol-water solution reaches 60:40 (v/v), and raised slightly passing this point; on the other hand, the desorption contents of isofraxidin gradually increased and reached their peak value at the ethanol-water (60:40, v/v) solution, and then decreased.When ethanol-water solution exceeds 60:40 (v/v), the desorbed impurities also increased.At the ethanol-water (60:40, v/v) solution, desorption mass of isofraxidin was 27.62 mg, and the relative content was the highest comparing with those at other ethanol-water solutions.Thus, ethanol-water (60:40, v/v) solution was selected as the appropriate desorption solution and used in the following dynamic desorption experiments.

Dynamic desorption curve on HPD100C
The ethanol-water (60:40, v/v) solution was used to elute isofraxidin.The flow rates investigated in this test were 2, 3 and 4 BV/h, respectively.The dynamic desorption curves on HPD100C were obtained based on the volume of desorption solution and the concentration of solute.As can be seen in Figure 5, the isofraxidin exhibited better desorption performance at the flow rate of 3 BV/h.At this flow rate, isofraxidin was thoroughly desorbed in 3.5 BV.
The results indicated that the process of 3 BV/h reduces ethanol solutions used and shortens the desorption process.Therefore, 3 BV/h was selected as the proper desorption flow rate considering its lower volume consumption and high efficiency.
The optimum parameters for the preparative separation

Figure 1 .
Figure 1.HPLC chromatograms of sample solution of Acanthopanax senticosus extracts (a) and desorption solution (b) from HPD100C.Inset: The molecular structure of isofraxidin.

Figure 4 .
Figure 4. Dynamic leakage curves of isofraxidin on column packed with HPD100C.

Table 1 .
Physical properties and absorption characteristics of the test macro-porous resins.

Table 3 .
Effects of different ethanol-water solution as desorption solutions on desorption properties of HPD100C for isofraxidin.