WO2023229255A1 - Method for separating beta-mangostin with high purity and high efficiency by using smb process and use thereof - Google Patents

Abstract

The present invention relates to a method for separating beta-mangostin with high purity and high-efficiency by using a simulated moving bend (SMB) process and use thereof, wherein the use of a beta-SMB process and a post-treatment process in the present invention enables the collection of high-purity beta-mangostin and allows a continuous process compared with general batch chromatography, thereby improving the productivity of beta-mangostin.

Description

High-purity and high-efficiency beta-mangosteen separation method using the SMB process and its uses

The present invention relates to a method for separating beta-mangosteen with high purity and efficiency from a mixture of mangosteen isomers using an optimized simulated moving bed chromatography process.

This research was conducted with support from the Ministry of Education's National Research Foundation of Korea's Industry-Academy-Research Cooperation Advancement (R&D) Project and the Rural Development Administration's Biogreen-Linked Agricultural Innovation Technology Development Project (Project Number: BRIDGE-2021-B04 and PJ015731).

Mangosteen isomer components (alpha, gamma, beta mangosteen) with high industrial value can be obtained through extraction of the peel of mangosteen ( Garcinia mangostana ), a natural fruit. Among these, alpha-mangosteen is known to be applicable to the manufacture of anti-inflammatory agents, antioxidants, and anticancer agents, while beta-mangosteen is a tyrosinase and TRP-1 (tyrosinase-related protein-1) agent related to melanin production. It is known to induce a skin whitening effect by inhibiting its activity.

Recently, interest in beta-mangosteen has been increasing, and in order to maximize the potential value and industrial usefulness of beta-mangosteen, it is essential to establish a high-purity production system for beta-mangosteen. Furthermore, in order to promote commercialization and secure economic feasibility of the beta-mangosteen production process, in addition to high purity separation conditions, additional conditions for securing high level throughput and high yield for mass production are required.

Most of the processes that have been applied to date to separate mangosteen components from natural raw material extracts are batch chromatography processes. In this process, the mixture is injected into an adsorbent column equipped with one inlet port and one outlet port for a certain period of time, and then the mobile phase solvent is continuously injected until all the mixture is discharged out of the column. Inject. The injected mobile phase solvent passes through the empty space in the column and is discharged through the outlet. In this process, the target component and impurities present in the mobile phase interact differently with the adsorbent, ultimately resulting in differences in the progression speed between each component. Only the target component is separated by using the difference in liver progression speed. In the case of the batch chromatography process, not only is it difficult to simultaneously secure high purity and high yield, but the effectiveness of the adsorbent is very low and the amount of solvent used is very large. In addition, because it is based on batch operation, it is impossible to maintain a steady state, so it is known that there are many restrictions on process operation management and automation system construction. Furthermore, there are bound to be many limitations in improving processing speed due to the inherent limitations of the batch operation mode.

In order to overcome the shortcomings of this batch process, multiple columns and multiple ports are combined to allow the mobile phase and adsorbent phase to cross each other in the opposite direction. Based on this, continuous injection of the mixture and high-purity continuous recovery of the target component are possible. There is a need to adopt a process method that can do this. A representative example of this process method is the simulated moving bed (SMB) chromatography process (Figure 1), and this process is recognized for its value in fields such as the bio, pharmaceutical, and fine chemical industries in the United States and Europe. .

Meanwhile, Korean Patent No. 1743063 discloses 'Method for separation of cresol isomers using pseudo-moving bed adsorption separation method', and Korean Patent No. 1837446 discloses 'Alpha-Mangosteen, Beta-Mango' batch separation method. Although a composition for improving skin wrinkles or moisturizing skin containing a steen, gamma-mangosteen or gartannin compound as an active ingredient is disclosed, the present invention provides a method for separating beta-mangosteen with high purity and efficiency using the SMB process and the same. There is nothing written about ‘usage’.

The present invention was developed in response to the above-mentioned needs, and the present inventors have developed a high-performance and highly efficient "beta-mangosteen" that can continuously separate only beta-mangosteen from a mangosteen peel extract (crude mangosteen mixture) with high purity and high yield. We sought to develop a “separated SMB process.”

To this end, we selected adsorbents and solvents suitable for the separation of target components, determined the dynamic behavior prediction model and related parameters of the raw material (crude extract) components, and designed the optimal operating conditions for the SMB process based on this, and designed the SMB with optimal operating conditions. Process equipment was manufactured. Afterwards, experimental verification was conducted on the optimal design results and the manufactured SMB device, and a post-treatment process for the SMB process product was established to recover the beta-mangosteen component in the raw material crude extract sample with a purity of more than 96%. By confirming that this was possible, the present invention was completed.

In order to solve the above problem, the present invention provides a method of producing beta-mangostin, comprising the step of separating beta-mangosteen by applying mangosteen extract raw materials to simulated moving bed chromatography. Provides a separation method.

Using the beta-SMB process developed in the present invention, not only can beta-mangosteen and other mangosteen components (alpha, gamma) contained in the raw material (crude extract) be separated with high efficiency, but also unknown impurities can be separated. A significant portion can be removed, allowing high purity beta-mangosteen to be recovered. In addition, by using the SMB process, the method of the present invention uses a very small amount of eluent compared to general batch chromatography, enables high-concentration and high-purity separation, and allows continuous processing, increasing the production of beta-mangosteen. It can be improved.

Figure 1 is a schematic diagram of a general four-zone closed loop SMB process structure.

Figure 2 shows a schematic diagram of the development of a beta-SMB process (beta-mangosteen separation SMB process) based on a column model-based process design approach.

Figure 3 is a device assembly design of the beta-SMB process (beta-mangosteen separation SMB process).

Figure 4 shows the results of a pulse injection experiment using an aqueous ethanol solution as a mobile phase (adsorbent: Sepra C8 resin, raw material: single mangosteen reagent 0.1 g/L, raw material injection amount: 200 ml). (a) Mobile phase: 100% ethanol, (b) Mobile phase: 95% ethanol, (c) Mobile phase: 90% ethanol, (d) Mobile phase: 85% ethanol.

Figure 5 shows the results of a pulse injection experiment (adsorbent: Sepra C8 resin, raw material: crude mangosteen sample 30 g/L, raw material injection amount: 15 ml) using an 85% ethanol aqueous solution as a mobile phase and targeting a crude mangosteen mixture sample. am.

Figure 6 is the optimal structure of the beta-SMB process (beta-mangosteen separation SMB process). The front part of β-1 and the back part of β-2 are located in zone III and zone I, respectively, and the back part of α-2 is located in zone II. Meanwhile, since the γ component has the lowest adsorption strength, it is automatically discharged through the raffinate port. Under this situation, beta-mangosteen components (β-1, β-2) are recovered through the extract port, and all other mangosteen components (α-1, α-2, γ) are recovered through the residue port. is removed. At each step, the inlet and outlet ports move the length of one column along the direction of the solvent, enabling continuous injection of the feed mixture and continuous recovery of the beta-mangosteen component.

Figure 7 is a graph comparing the pulse injection experiment results and computer simulation results. Adsorbent: Sepra C18 resin, mobile phase: 85% ethanol, (a) α-1 (based on high concentration: crude sample), (b) α-2 (based on low concentration: single reagent), (c) β-1 (based on high concentration: crude sample), (d) β-2 (low concentration standard: single reagent).

Figure 8 shows simulation results for the periodic steady-state column profile of the optimally designed beta-SMB process. (a) Beginning of a switching period, (b) Middle of a switching period, (c) End of a switching period.

Figure 9 is a photo of the device taken during the beta-SMB process experiment.

Figure 10 shows the results of the beta-SMB process experiment for separating beta-mangosteen. (a) Effluent history for extract port (product port) stream, (b) Effluent history for raffinate port (non-product port) stream.

Figure 11 is raw data of an HPLC concentration analysis chromatogram to prove the beta-SMB process experiment results. (a) raw material solution (crude mangosteen sample solution), (b) SMB product (extract) stream, (c) beta-mangosteen standard reagent.

Figure 12 is a chromatogram of the beta-mangosteen separation process using prep-HPLC.

Figure 13 shows the results of a post-treatment process experiment using prep-HPLC. (a) SMB extract (product) stream, (b) re-HPLC-1, (c) re-HPLC-2 (purity >95%, beta-mangosteen), (d) re-HPLC-3.

In order to achieve the object of the present invention, the present invention includes the step of separating beta-mangosteen by applying mangosteen extract raw material to simulated moving bed chromatography, beta-mangostin (β-mangostin). ) provides a separation method.

The simulated moving bed (SMB) process generally has a structure divided into four zones by two inlets (feed, desorbent) and two outlets (extract, raffinate) (Figure 1). , each zone is connected to one or more columns filled with pre-selected adsorbents. Components with relatively high affinity to the adsorbent (high-affinity solute) move at a slow speed along the flow direction of the mobile phase solvent (eluent), while components with low affinity (low-affinity solute) move at a high speed. At this time, the four ports are moved the length of one column at regular time intervals along the flow direction of the mobile phase, and the moving speed (= column length/port switching time) is set to be greater than the speed of the high affinity component and the low affinity component. Choose to be smaller than the speed of . If these conditions are maintained, components with high affinity are always lag behind the feed port and are recovered from the extract port, while components with low affinity are always ahead of the feed port. beyond), so it is obtained from the raffinate port. As a result, the raw material mixture is always injected into the overlapping region of the high-affinity and low-affinity components, while the residue and extract are always obtained from the separated region. If this condition is maintained continuously, continuous injection of the raw material mixture and continuous recovery of each target component (product) are possible. In addition, it is possible to recover target components with high purity and high yield even under partial-separation situations in which the solute bands of two different components (high-affinity solute and low-affinity solute) are not completely separated within the SMB column but are only partially separated. .

The SMB process according to the present invention is an open-loop system in which the mobile phase is not recirculated. Choosing the open type has the advantage of protecting zone I from contamination and making it easy to adjust the flow rate when problems occur during operation.

In the method for separating beta-mangosteen according to the present invention, the mangosteen extract may contain beta-mangosteen and a mixture of isomers thereof, and the isomers may be alpha-mangosteen and gamma-mangosteen.

In the present invention, simulated moving bed chromatography was used to separate beta-mangosteen with high purity from a mixture of mangosteen isomers, which are difficult to separate.

The method for isolating beta-mangosteen according to one embodiment of the present invention,

Determining the adsorption coefficient and mass transfer coefficient of alpha-mangosteen and beta-mangosteen for the adsorption column applied to simulated moving bed chromatography (step 1);

Calculating the operating conditions of the simulated moving bed chromatography process by applying the adsorption coefficient and mass transfer coefficient determined in step 1 to a simulated moving bed computer simulation program based on the genetic algorithm and column model equation (step 2) ); and

It may include, but is not limited to, supplying mangosteen extract raw materials and eluent to simulated moving bed chromatography and separating beta-mangosteen by applying the operating conditions calculated in step 2 above (step 3).

In the method of the present invention, the adsorption column in step 1 may include C18-coated silica as a stationary phase, but is not limited thereto. In one embodiment of the present invention, the C18-coated silica was confirmed to be the most suitable adsorbent in terms of separation between beta-mangosteen and other mangosteen components and applicability to continuous processes. Furthermore, it was confirmed that for continuous process application, the particle size of the adsorbent must be 50 ㎛ or more and a sufficient level of separation selectivity must be guaranteed even under low pressure.

In addition, the adsorption coefficients (Henry constants) of alpha-mangosteen and beta-mangosteen on an adsorption column containing C18-coated silica as a stationary phase were determined from the pulse injection experiment data (FIGS. 4d and 5) for each component. was determined using the retention time and solute movement theory, and the mass transfer coefficients of alpha-mangosteen and beta-mangosteen for an adsorption column containing C18-coated silica as a stationary phase were determined using the Chung and Wen correlation. and computer simulation results were used to calculate it (Table 2).

In addition, in one implementation of the present invention, the column model equation of step 2 may be a lumped mass-transfer model, and the simulated moving layer computational simulation program may be Aspen Chromatography ^TM , but is not limited thereto. No. The operating conditions of simulated moving bed chromatography calculated in step 2 of the present invention are shown in Table 3.

In the method of the present invention, the eluent in step 3 may be a lower alcohol of C1 to C4, preferably methanol or ethanol, more preferably ethanol, and even more preferably 80 to 95 alcohol. % (v/v) ethanol, most preferably 85% (v/v) ethanol, but is not limited thereto. The ethanol was confirmed to have excellent separation selectivity between beta-mangosteen and its isomers under the conditions of an adsorption column containing C18-coated silica as a stationary phase. Depending on the type of adsorption column, there are different types of ethanol commonly used in the art. It may be variable with the eluent.

In addition, the separation method of beta-mangosteen according to the present invention may additionally perform recycling prep HPLC on the product eluted after the simulated moving bed chromatography in step 3. It is not limited to this.

Existing Preparative HPLC is a system in which the solvent from the pump and the sample injected from the manual injector are mixed and separated on the column, and the separation is confirmed with a detector and then immediately collected. . However, in the case of recycling prep HPLC, the recycling system is controlled by an automatic recycler, so that samples that have not been separated are immediately reinjected into the column, that is, samples that have not been separated are placed on the column until separation is performed. It is a method based on the principle of recycling. In addition, during the circulation process of recycling prep HPLC, a wide volume from the column of the first cycle to the column of the second cycle is maintained at a positive pressure so that the samples are sequentially separated without being remixed due to diffusion or other reasons, so no solvent consumption occurs. This has the advantage of reducing the burden of solvent costs.

In one embodiment of the present invention, the additional recycling prep HPLC may be performed to remove trace residues contained in the product eluted after simulated moving bed chromatography.

The simulated moving bed (SMB) chromatography of the present invention,

Using a three-zone open SMB chromatography with three inlets and two outlets, the desorption solvent (eluent) is injected into the inlet of the first zone, and at the same time, the beta-mangosteen product is released through the outlet of the first zone. ) recovering (step ①);

Simultaneously with step ①, a desorption solvent is injected into the inlet of the second zone so that the alpha-mangosteen and gamma-mangosteen components can be desorbed from the first column of the second zone within one exchange time and proceed toward the third zone. Step (step ②); and

At the same time as step ② above, the mangosteen sample solution is injected into the third zone through the supply fluid inlet, and at the same time, the movement range of the beta-mangosteen component in the mangosteen sample is limited to within the third zone, while alpha-mangosteen and gamma -It can be understood as a three-zone open SMB chromatography that includes the step (step ③) of allowing the mangosteen component to pass through the third zone exit point and be discharged to the outside within one exchange time.

Through the separation method of beta-mangosteen according to the present invention, beta-mangosteen can be separated with a high purity of 70% or more, 80% or more, 90% or more, or 96% or more, and unlike general batch chromatography, production The yield is excellent, showing a yield of 70% or more, 80% or more, 90% or more, or 96% or more, making mass production possible, which shows that the separation method is suitable for commercial use.

Hereinafter, the present invention will be described in detail by examples. However, the following examples only illustrate the present invention, and the content of the present invention is not limited to the following examples.

Materials and Methods

1. Theoretical approach

To develop the beta-SMB process (beta-mangosteen separation SMB process), which is the goal of the present invention, the "column model and parameter-based process design approach" as shown in FIG. 2 was adopted as the overall proceeding method. The process development process of the present invention is largely comprised of solvent selection for the adsorbent and mobile phase, determination of basic parameters of each component (alpha, beta, and gamma mangosteen) in the selected adsorbent, and optimization of the beta-SMB process based on the determined basic parameter information. It is divided into four stages: design and theoretical verification, production of a beta-SMB device that meets the optimal design results, experimental verification based on this, and scale-up. By using a systematic approach based on column models and parameters at each step, reproducibility and reliability of the process design results were maintained.

1-1. Computer simulation based on column model

Column model-based computer simulation plays an important role in various stages of beta-SMB process development, and the key parts are as follows: verification of basic parameter values of major mangosteen components, optimal design of SMB, and prediction of SMB concentration profile.

Simulation is a process of calculating the solution of the so-called column model equation, a mathematical model for the transport behavior of each component in the column. Since the column model equation itself is composed of simultaneous partial differential equations, it is computer-based. The use of numerical computational tools is essential.

Several model equations related to the column model have been introduced in the literature. Among them, the lumped mass-transfer model, which has proven accuracy and computational efficiency, was adopted as the column model of the present invention. The structure of the adopted lumped mass-transfer model equation is as follows (Z. Ma et al., AIChE J. 1997, 43:2488-2508; Benjamin J. Hritzko et al., AIChE J. 2002, 48:2769- 2787; Y.

In the above equation, the subscript i represents the solute, C _b,i and C _i ^* are the solute liquid phase concentration in the inter-particle void (mobile phase) and intra-particle void (pore phase), respectively, and q _i is It refers to the concentration in the adsorbent phase, which is in equilibrium with the liquid phase concentration in the pore phase. The relationship between the liquid phase concentration and the adsorbent phase concentration in equilibrium is expressed by the adsorption model equation. Meanwhile, K _f in the above column model equation is a lumped mass-transfer coefficient, and its value can be calculated in the following way.

In the above equation, d _p represents the diameter of the adsorbent particle, and D _p and k _f refer to intra-particle diffusivity and film mass-transfer coefficient, respectively.

Single column simulation and SMB process simulation based on the column model equation described above were performed using Aspen Chromatography Simulator.

1-2. Optimal design of SMB process

Regarding the optimization of a continuous separation process consisting of multiple adsorbent columns and multiple ports, the fact is that optimization techniques based on stochastic theory such as genetic algorithms are more effective than existing traditional optimization techniques. It has been reported previously. In particular, in the case of the SMB process, many optimization cases based on genetic algorithms have been reported, and their accuracy is known to be excellent (S. Mun et al., J. Chromatogr. A 2012, 1256:46-57; Z. Zhang et al., AIChE J 2002, 48:2800-2816; H.J. Subramani et al., Comput. Chem. Eng. 2003, 27:1883-1901).

Optimization of the SMB process based on genetic algorithms can be broadly divided into two areas. The first area is a column model that can accurately predict the dynamic behaviors of each component in the SMB column and a solution method that can find a solution to this model equation. The second area is to follow a step-by-step procedure based on a genetic algorithm to determine the optimal operation parameters (flow rate, switching time) of SMB and track the solution corresponding to the optimal operation parameters.

The combination of the two areas mentioned above is achieved in the following way. After the SMB operation parameters for each generation calculated from the genetic algorithm stage are substituted into the column model equation, the separation behavior of each component and the purity and yield of SMB can be determined through numerical solutions of this model equation. . This result is again used as key information to track optimal operation parameters at each stage of the genetic algorithm. Through the combination of this column model and genetic algorithm, it is possible to determine the SMB optimal operation parameter solution. A detailed description of these two areas, which are the core of the SMB optimization tool, is summarized below.

In the case of the column model, which is the first area of the SMB optimization tool, the lumped mass-transfer model mentioned above was adopted as the simulation model of the optimization tool. In addition, the Aspen Chromatography simulator based on the biased upwind differencing scheme (BUDS) and gear integration method was used to calculate the numerical solution of the adopted column model equation (Eq. (1)).

The genetic algorithm, the second area of the SMB optimization tool, has undergone several improvements, the most recent version being NSGA-Ⅱ-JG (elitist nondominated sorting genetic algorithm with jumping gene) (R.B. Kasat et al. , Comput. Chem. Eng. 2003, 27:1785-1800; S. Mun et al., 2012). In the present invention, an SMB optimization computer tool based on the NSGA-Ⅱ-JG algorithm was produced using the Visual Basic Application (VBA) programming language. In this process, the operation parameters (flow rate, switching time) of the SMB process were set as targets for the chromosome, and the prediction of SMB performance, which determines the fitness of this chromosome, was performed based on the solution of the column model equation.

2. Experimental approach

2-1. Materials used in the experiment

Crude mangosteen mixture samples and single mangosteen pure reagents (alpha, beta, and gamma mangosteen) used in the beta-SMB process development study of the present invention were provided by the Gyeongsang National University Industry-Academic Cooperation Foundation and Bioclone (Korea), respectively. Acetonitrile and acetic acid, which were used as solvents for the mobile phase during HPLC concentration analysis, were purchased from Daejeong Chemical (Korea) and Sigma-Aldrich Co. (USA), respectively. Silica resin and Sepra C18 resin used in the adsorbent test of the present invention were purchased from Osaka Soda Co. (Japan) and Phenomenex Co. (USA), respectively, and their average particle size was 50 ㎛. The above adsorbent was used by filling a preparative-scale column (2.5 × 22 cm) purchased from Bio-Chem Fluidics Co. (USA).

n-hexane, ethyl acetate, methanol, and ethanol used as mobile phase solvents for normal and reversed phase chromatography columns in the single column experiment (pulse injection experiment) of the present invention were all purchased from Daejeong Chemical. All water used in the experiment was distilled deionized water (DDW), obtained from the Milli-Q system purchased from Millipore Co. (USA) and used in all parts of the experiment.

2-2. Equipment used in the experiment

① Pulse injection experiment device

A Young-Lin HPLC system purchased from Younglin Equipment (Korea) was used for the pulse injection experiment. The Young-Lin HPLC system consists of two Young-Lin SP930D pumps, a Young-Lin UV730D detector, and Autochro-3000 software. The Young-Lin SP930D pump is responsible for smooth transfer of the solvent, and the Young-Lin UV730D detector is responsible for real-time monitoring of the concentration of each component in the column effluent. Autochro-3000 software is responsible for control and data collection of the pump and detector.

② HPLC concentration analysis device

A Waters HPLC system was used as a device to analyze the concentration of samples obtained from pulse injection experiments and SMB experiments targeting crude mangosteen samples. The solvent was transported by a Waters 515 HPLC pump, and concentration analysis of the samples was performed by a Waters 996 PDA detector (wavelength = 254 nm). In addition, an Agilent Eclipse XDB-C18 analytical column (0.46 × 15 cm) was purchased and used as an HPLC analysis column. Injection of the sample was performed through a Rheodyne 7725i injector, and the sample injection volume was 10 μl. The HPLC analysis was operated in gradient mode at room temperature, and of the two types of solvents (referred to as A and B for convenience) used, solvent A is 100% acetonitrile and solvent B is 0.1% acetic acid aqueous solution. The gradient mode operation method based on solvents A and B is shown in Table 1. Throughout the HPLC analysis, the flow rate was maintained at 1 mL/min and control of the Waters HPLC system was performed by Empower 2.0 software.

Gradient mode operation method applied to HPLC analysis

Time (min)	Solvent A (%)	Solvent B (%)	note
45	95	5	Linear gradient
60	100	0	Linear gradient
75	100	0	Washing
90	5	95	Equilibrium

③ SMB process experiment device

The beta-SMB process experimental device of the present invention was self-assembled and manufactured to have a 1-2-2 column configuration structure under a 3-zone open-loop method. The manufactured SMB device (FIG. 3) consisted of 5 rotary valves, 5 columns, and 3 pumps. The rotary valve used in the SMB device is a Select-Trapping (ST) valve purchased from Valco Instrument Co. (USA). This valve connects each column and each port to maintain a flow structure that allows continuous separation. have Control of the rotary valve was performed using Labview 8.0 software. The flow rate of the stream injected into each port of the SMB device was controlled using a Waters 515 HPLC pump purchased from Waters. Meanwhile, the flow rate of the stream discharged to the residue port was determined by mass balance without a separate pump.

2-3. Experimental procedure

① Pulse injection experiment

A preparative-scale column (2.5 × 22 cm) filled with the adsorbent candidate (silica gel, Sepra C18) was installed in the Young-Lin HPLC system, and then each of the single mangosteen reagent components (alpha, beta, and gamma) and the crude mangosteen sample were analyzed. A pulse injection experiment was performed. The injection amount and concentration of the feed injected into the column were set to 200 ㎕ and 0.1 g/L, respectively, for a single mangosteen component, and 15 ㎖ and 30 g/L, respectively, for crude mangosteen samples. Regarding the injection of raw materials into the column, single mangosteen components were injected using a Rheodyne 7725i injector, and crude mangosteen samples were injected using a SP930D pump.

While the pulse injection experiment was in progress, the flow rate was kept constant at 2 mL/min. Effluent history data (concentration profile data of each component in the column effluent over time) for each component following pulse injection was obtained in real time using a UV730D detector for single mangosteen components, and column effluent for crude mangosteen samples. Water was collected at different times and secured through HPLC concentration analysis.

The mobile phase solvent was applied as follows depending on the phase of the adsorbent packed in the column. When silica resin was used as an adsorbent, the experiment was repeated while changing the ratios of n-hexane and ethyl acetate to 1:1, 15:1, and 30:1. On the other hand, when Sepra C18 resin is used as an adsorbent, a mixed solvent of methanol-DDW (methanol content: changed to 100%, 99%, 95%) or ethanol-DDW (ethanol content: changed to 100%, 95%, 90%, 85%) is used. Change) The experiment was repeated while using a mixed solvent as a mobile phase.

② SMB process experiment

The start of the SMB experiment began with the operation of each pump and the start of execution of Labview 8.0 software. At the beginning of the experiment, raw materials and desorbent were continuously injected into the SMB. A crude mangosteen mixture solution with a concentration of 30 g/L was continuously injected into the raw material port, and the concentrations of alpha, beta, and gamma mangosteen in the raw material solution were 4.80 g/L, 0.13 g/L, and 0.70 g/L, respectively. Meanwhile, an 85% ethanol aqueous solution was used as the eluent continuously injected into the desorbent-1 and desorbent-2 ports. The SMB experiment was conducted until a cyclic steady state was sufficiently reached. During the SMB experiment, the accuracy of the flow rate was checked at every switching period, and the concentration of the stream discharged from the extract and residue ports was measured using the Waters HPLC concentration analysis system mentioned above. Real-time analysis was performed.

Example 1. Determination of an adsorbent phase suitable for continuous separation between beta-mangosteen and other mangosteen components in a crude mangosteen sample.

In order to select an adsorbent to be applied to the beta-SMB process, which is the target process of the present invention, the type of adsorbent phase suitable for a continuous separation process such as SMB must be determined. For this purpose, silica resin and C18 series resin, which represent normal phase and reversed phase chromatography adsorbents, were selected and a series of pulse injection experiments were performed.

In the pulse injection experiment based on a silica column (a column filled with silica resin), a mixed solvent of n-hexane and ethyl acetate was used as the mobile phase, and the ratio of n-hexane and ethyl acetate solvents was changed in this process. We conducted experiments as we went. As a result, it was confirmed that the greater the content of ethyl acetate, which corresponds to the polar solvent, the more smoothly the desorption of beta-mangosteen proceeds. This is because normal chromatographic adsorbents are polar. On the other hand, as the content of n-hexane, which corresponds to a non-polar solvent, increases, desorption of beta-mangosteen does not proceed smoothly, resulting in a tailing phenomenon in which the back part of the solute band of beta-mangosteen is elongated. This occurred. However, the breakthrough time of the beta-mangosteen component showed no change or tended to become slightly faster, which was predicted to be due to the non-polar properties of beta-mangosteen.

One of the important considerations to realize the continuous operation mode of the SMB process is that the desorption of all components must proceed smoothly. Additionally, under a chromatography environment where these conditions are met, the adsorbent separation selectivity between target components and non-target components must be above a certain level. As a result of considering these matters, it was confirmed that silica resin cannot meet both requirements of smooth desorption of all components and ensuring an appropriate level of separation selectivity.

To investigate the suitability of reversed-phase chromatography adsorbents for the beta-SMB process, pulse injection experiments were conducted using Sepra C18 resin, which is known to have particularly excellent durability among reversed-phase adsorbents. As a preliminary experiment, 100% methanol was used as a mobile phase, and a single mangosteen reagent (pure reagent as alpha, beta, and gamma mangosteen) was used to quickly understand the experimental results. As a result, it can be confirmed that the Sepra C18 adsorbent can provide a meaningful level of separation selectivity between beta-mangosteen and other major non-target components (alpha- and gamma-mangosteen). In addition, it can be confirmed that the desorption of beta-mangosteen is also proceeding smoothly. However, it was also discovered that the severe tailing of gamma-mangosteen could interfere with securing high purity of beta-mangosteen. For this reason, it was considered necessary to optimally determine the type and composition of solvent in a reversed-phase chromatography environment.

Example 2. Optimal determination of ethanol content in mobile phase under reversed phase chromatography (Sepra C18 adsorbent) environment

There was a need to discover and apply a mobile phase solvent that could further improve the separation selectivity between beta-mangosteen components and other mangosteen components while being environmentally friendly and not harmful to the human body compared to methanol, so ethanol and DDW were used. A mixed solvent was selected as a mobile phase solvent candidate and an experiment was conducted to determine its applicability. In this experiment, a series of pulse injection experiments were conducted on a single mangosteen reagent component while varying the ethanol content in the mixed solvent of ethanol and DDW from 100% to 85% at 5% intervals.

As a result, as shown in Figure 4, as the ethanol content in the mobile phase decreases, the residence time of all mangosteen components in the column tends to increase. What is noteworthy is that there is a large difference between the rate of increase in retention time of beta-mangosteen and the rate of increase in retention time of other mangosteen components as ethanol content decreases. In other words, as the ethanol content decreases, the retention time of the beta-mangosteen component tends to increase rapidly, while the retention time of other mangosteen components shows a pattern of gradually increasing. For this reason, the separation selectivity between beta-mangosteen and other mangosteen components generally tends to improve as the ethanol content decreases. Meanwhile, in the case of the gamma-mangosteen component, a serious tailing phenomenon occurs, but it can be seen that it is greatly alleviated under the condition of 85% ethanol content (Figure 4d). Considering these results, it could be concluded that the optimal content of ethanol in the mobile phase was 85%.

Since the above conclusion is a result of a single mangosteen reagent, a pulse injection experiment was conducted on a crude mangosteen sample to confirm whether it is also valid for a crude mangosteen sample. In this experiment, 85% ethanol solvent, which was confirmed to be the optimal solvent in an experiment targeting a single mangosteen reagent, was used as the mobile phase. As a result of the experiment, as can be seen in Figure 5, a similar level of beta-mangosteen separation selectivity was confirmed in the pulse injection experiment using a crude mangosteen sample as in the experiment using a single mangosteen reagent. Considering the above results, 85% ethanol solvent was determined to be the optimal mobile phase for separation between beta-mangosteen and other mangosteen components under a reversed-phase chromatography environment.

Example 3. Optimal design approach for beta-SMB process

What is noteworthy in the pulse injection experiment of Example 2 is that when injecting the raw material (feed), the injection amount of the crude mangosteen sample was set to be more than 75 times higher than the injection amount of a single mangosteen reagent. This means that the pulse injection experiment conditions for crude mangosteen samples are conditions that maintain the mangosteen concentration in the column higher. Based on this, when analyzing Figures 4d and 5, it was found that the retention time of the remaining mangosteen components (alpha, beta) except gamma-mangosteen becomes faster as the concentration increases, which means that the retention time of the remaining mangosteen components (alpha and beta) increases as the concentration increases. This means that the Steen component follows a Langmuir type nonlinear adsorption tendency.

Accordingly, based on the above results, the following beta-SMB process design approach was adopted. Each mangosteen ingredient is treated by dividing it into two components: high concentration (concentration based on crude sample: equivalent to the concentration level input to the SMB process) and low concentration (concentration based on single reagent component). Determine the Henry constant (linear adsorption coefficient) for For convenience, the high- and low-concentration standard components of beta-mangosteen are denoted as β-1 and β-2, respectively, and alpha-mangosteen and gamma-mangosteen are also denoted as α-1, α-2, and γ, respectively. I decided to do it.

In addition, in the case of the beta-SMB process structure, which is the target of the present invention, the structure is selected to minimize the number of columns and valves to save equipment and management costs. However, in order to improve the separation efficiency between beta-mangosteen and other mangosteen components, more columns were placed in the separation zone (two zones located around the raw material port), and the burden of column regeneration was increased. In order to reduce and improve regeneration efficiency, if possible, two zones, instead of one, were responsible. As a result of reviewing the SMB process structure that can satisfy all of the above-mentioned requirements, the structure presented in Figure 6 was confirmed to be the most suitable.

In the beta-SMB process disclosed in Figure 6, the front part of β-1 and the back part of β-2 are located in zone III and zone I, respectively, and the rear part of α-2 is located in zone II. Meanwhile, since the γ component has the lowest adsorption strength, it has no choice but to be automatically discharged through the raffinate port. Under this situation, the beta-mangosteen target components (β-1, β-2) are recovered through the extract port, and all other mangosteen components (α-1, α-2, γ) are recovered through the residue port. is removed through In addition, at each step, the inlet and outlet ports move the length of one column along the direction of the solvent, enabling continuous injection of the feed mixture and continuous recovery of the beta-mangosteen target ingredient.

Example 4. Determination of basic parameters of alpha, beta, and gamma mangosteen components for optimal design of beta-SMB process

In order to perform the optimal design of the beta-SMB process, it is necessary to secure the basic parameters (Henry constant, mass transfer coefficient) of the alpha and beta mangosteen components (α-1, α-2, β-1, β-2). To this end, the retention time of each component was determined from the results of the pulse injection experiment performed in Example 2 (FIGS. 4d and 5), and the Henry constant (H) of each component was determined using this information and solute movement theory. could be determined (Table 2). The mass transfer coefficient, which is a major basic parameter that must be determined additionally following the Henry's constant, includes the axial dispersion coefficient (E _b ) and the film mass-transfer coefficient (k _f ). Among these, the value of E _b was determined using the Chung and Wen correlation, and the k _f value was determined as the value that could most closely approximate the computer simulation results to the pulse injection experiment data. The mass transfer coefficient values of each determined mangosteen component (α-1, α-2, β-1, and β-2) are shown in Table 2.

Basic parameters of major mangosteen ingredients for beta-SMB process design

	H	Eb ( cm2 /min)	k f (cm/min)
α-1	1.45	Chung & Wen, AIChE J. 1968, 14:857-866	0.01
α-2	2.55		0.25
β-1	4.44		0.04
β-2	5.29		0.95
H: Henry's constant, E b : axial dispersion coefficient, k f : film mass-transfer coefficient.

In order to confirm the appropriateness of the basic parameter values for each component presented in Table 2, a computer simulation was performed based on the parameter values in Table 2 and the column model equation (Eq. (1)). The computer simulation results and experimental data were compared for each component. As a result, it was confirmed that the computer simulation results and the experimental data were in good agreement, as shown in FIG. 7. This means that the basic parameter values presented in Table 2 relatively well express the transport phenomena and separation behavior of each component within the Sepra C18-based reversed-phase column. Therefore, it was expected that the basic parameter values of each component presented in Table 2 could be successfully used in the optimal design process of the beta-SMB process to be carried out in the future.

Example 5. Optimal design of beta-SMB process

A study was conducted on the optimal design (determination of optimal operating conditions) of the beta-SMB process under the determined basic parameters of each mangosteen ingredient (Table 2) and the confirmed SMB process structure (Figure 6). The optimal design was promoted according to the following procedures.

Throughput, that is, with the feed flow rate fixed, the yield of beta-mangosteen components (β-1, β-2) and the removal rate of other mangosteen components (α-1, α-2, γ) are measured. An optimization frame was created to ensure that everything was maintained above 99.9%. An optimization tool matching the above optimization frame was constructed, and the Aspen chromatography simulator based on genetic algorithm and column model equation (Eq. (1)) was used in this process. The optimal operating conditions for the beta-SMB process were determined using the optimization tool configured above, and the results are presented in Table 3.

Optimal operating conditions for beta-SMB process

Zone flow rates (㎖/min)	Q 1	1.95
	Q 2	3.05
	Q 3	3.55
Inlet and outlet flow rates (㎖/min)	Q des-1	1.95
	Q des-2	3.05
	Q feed	0.5
	Q raf	3.55
Switching time (min)	tsw	64.63

To theoretically verify the optimal design results of the beta-SMB process, an SMB process simulation was conducted based on the obtained optimal operating conditions (Table 3) and the column model equation. Through the simulation, a column profile under a cyclic steady state was obtained, and the results are presented in FIG. 8.

As can be seen in the column profile of Figure 8, both the front part of the β-1 solute band and the back part of the β-2 solute band are well-confined within zone III and zone I. Additionally, the rear part of the α-2 band is also well controlled within zone II. Under this situation, all of the remaining mangosteen components (a-1, γ) have no choice but to be automatically discharged through the residue port, so the behavior of these components has no effect on the target component (extract) stream. Considering all these points, the operating conditions (results of optimal design) in Table 3 were determined to be SMB operating conditions that can ensure high purity and high yield continuous separation between beta-mangosteen and other mangosteen components.

Example 6. Experimental verification of the optimally designed beta-SMB process

To experimentally verify the optimally designed beta-SMB process, an SMB process device conforming to the optimal structure of FIG. 6 was assembled and manufactured. Based on this device and the optimal operating conditions in Table 3, continuous separation experiments were performed between beta-mangosteen and other mangosteen components (alpha, gamma) in crude mangosteen samples. Figure 9 is a photo of the SMB device taken during this experiment.

During the SMB experiment, a 30 g/L solution of crude mangosteen sample was continuously injected into the raw material port, and an 85% ethanol aqueous solution was continuously supplied into the desorbent-1 and desorbent-2 ports. The SMB experiment was conducted for a total of 23 stages (25 hours), and step-by-step concentration analysis was performed on the extract port (beta-mangosteen product port) stream and the residue port (non-product port) stream throughout the entire experiment. As a result, is shown in Figure 10. Quantitative analysis of concentration was conducted only on the main components of alpha, beta, and gamma mangosteen. Therefore, in the case of beta-mangosteen purity, the purity considering only the main components (alpha, beta, and gamma mangosteen) was quantitatively investigated, and the purity considering all components was replaced with HPLC analysis chromatogram raw data for the SMB product stream.

As can be seen in the results of Figure 10a, the extract pot confirmed the presence of only beta-mangosteen among the main mangosteen components. On the other hand, only alpha and gamma mangosteen were confirmed to leak into the raffinate port (Figure 10b). The above results meant that, considering only the main components, both purity and yield reached 100% when separating beta-mangosteen through the beta-SMB process.

As evidence for the above results, HPLC concentration analysis chromatogram raw data for each of the crude mangosteen sample solution, extract stream solution, and beta-mangosteen standard sample that were input into the beta-SMB process were obtained and presented in Figure 11. As can be seen in Figure 11b, peaks of alpha and gamma mangosteen components were not found at all in the chromatogram for the extract (product) of the beta-SMB process. Furthermore, it was confirmed that many unknown impurity peaks located in front of the beta-mangosteen peak in Figure 11a (chromatogram for a crude mangosteen sample) were largely removed in Figure 11b. This meant that the beta-SMB process developed in the present invention could not only completely remove alpha and gamma mangosteen components contained in large quantities in crude samples, but also remove a significant portion of other small or trace impurities.

For additional proof of the above experimental results, the beta-SMB process extract (product) chromatogram of FIG. 11b was compared with the beta-mangosteen standard sample chromatogram (FIG. 11c). As a result, the presence of beta-mangosteen components in the extract (product) stream of the beta-SMB process was clearly confirmed, and the presence of impurities around the beta-mangosteen peak was also observed. However, the number and size of the impurity peaks were significantly reduced compared to the number and size of the impurity peaks in the crude mangosteen sample (FIG. 11a). As a result, it is worth noting that beta-mangosteen, which was only a minor component in crude mangosteen samples, became a major component with the largest content in the SMB product stream.

Example 7. Post-treatment process using recycling preparative HPLC

Despite the above results, in the extract (product) stream of the beta-SMB process, several types of minor impurities (including xanthone impurities) exist in addition to the beta-mangosteen component. To further remove these impurities, a post-treatment process for the product of the beta-SMB process was developed according to the following procedure.

JAI NEXT recycling preparative HPLC was used to selectively separate the target substance (beta-mangosteen) in the beta-SMB process product through HPLC chromatogram. The column used was Thermofisher Scientific's Acclaim Polar Advatage Ⅱ (5 ㎛, 250 x 20 mm), and the mobile phase was 100% acetonitrile under isocratic conditions. As a result, as shown in Figure 12, a total of three fractions were secured: re-HPLC-1 from peaks 15 and 56, re-HPLC-2 from peaks 27 and 72, and re-HPLC-3 from peaks 28, 73, and 74. did. As a result of HPLC confirmation of the three fractions, a selective peak of beta-mangostin (t _R = 53 min) was confirmed in the second fraction, re-HPLC-2 (FIG. 13). When using the Recycle-HPLC process, 2.1 mg of beta-mangosteen was secured from an injection volume of 5,000 mg/ml, and the final purity was >96%. This confirms that the beta-mangosteen component in the crude mangosteen sample can be recovered with a purity of 96% or more through the two-step process procedure of “beta-SMB process and recycle-HPLC-based post-treatment process” developed in the present invention. did.

Claims (11)

1. A method of separating beta-mangostin, comprising the step of separating beta-mangostin by subjecting mangosteen extract raw material to simulated moving bed chromatography.
2. The method of claim 1, wherein the mangosteen extract contains beta-mangosteen and a mixture of isomers thereof.
3. The method of claim 2, wherein the isomers are alpha-mangostin and gamma-mangostin.
4. The method of claim 1, wherein the method

   Determining the adsorption coefficient and mass transfer coefficient of alpha-mangosteen and beta-mangosteen for the adsorption column applied to simulated moving bed chromatography (step 1);

   Calculating the operating conditions of the simulated moving bed chromatography process by applying the adsorption coefficient and mass transfer coefficient determined in step 1 to a simulated moving bed computer simulation program based on the genetic algorithm and column model equation (step 2) ); and

   Beta-mangosteen, comprising: supplying mangosteen extract raw materials and eluent to simulated moving bed chromatography and separating beta-mangosteen by applying the operating conditions calculated in step 2 (step 3); Separation method.
5. The method for separating beta-mangosteen according to claim 4, wherein the adsorption column in step 1 includes C18-coated silica as a stationary phase.
6. The method of separating beta-mangosteen according to claim 4, wherein the column model equation of step 2 is a lumped mass-transfer model.
7. The method for separating beta-mangosteen according to claim 4, wherein the eluent in step 3 is a lower alcohol of C1 to C4.
8. The method of claim 7, wherein the lower alcohol is ethanol.
9. The method for separating beta-mangosteen according to claim 4, characterized in that recycling prep HPLC (Recycling Preparative HPLC) is additionally performed on the product eluted after the simulated moving bed chromatography in step 3. .
10. The method of claim 1, wherein the simulated moving bed (SMB) chromatography,

    Using a three-zone open SMB chromatography with three inlets and two outlets, the desorption solvent (eluent) is injected into the inlet of the first zone, and at the same time, the beta-mangosteen product is recovered through the outlet of the first zone. Step (step ①);

    Simultaneously with step ①, a desorption solvent is injected into the inlet of the second zone so that the alpha-mangosteen and gamma-mangosteen components can be desorbed from the first column of the second zone within one exchange time and proceed toward the third zone. Step (step ②); and

    At the same time as step ② above, the mangosteen sample solution is injected into the third zone through the supply fluid inlet, and at the same time, the movement range of the beta-mangosteen component in the mangosteen sample is limited to within the third zone, while alpha-mangosteen and gamma -Separation of beta-mangosteen, characterized in that it is a 3-zone open SMB chromatography, including a step (step ③) of allowing the mangosteen component to be discharged to the outside through the exit point of the 3rd zone within one exchange time. method.
11. The separation method according to claim 1, wherein the beta-mangosteen isolated by the separation method has a purity of 96% or more.

Images