Extraction and in vitro screening of potential acetylcholinesterase inhibitors from the leaves of Panax japonicus

Accepted Manuscript Title: Extraction and in vitro screening of potential acetylcholinesterase inhibitors from the leaves of Panax japonicus Authors: Sainan Li, Chengyu Liu, Chunming Liu, Yuchi Zhang PII: DOI: Reference:

S1570-0232(17)30984-4 http://dx.doi.org/doi:10.1016/j.jchromb.2017.07.019 CHROMB 20696

To appear in:

Journal of Chromatography B

Received date: Revised date: Accepted date:

31-5-2017 9-7-2017 11-7-2017

Please cite this article as: Sainan Li, Chengyu Liu, Chunming Liu, Yuchi Zhang, Extraction and in vitro screening of potential acetylcholinesterase inhibitors from the leaves of Panax japonicus, Journal of Chromatography Bhttp://dx.doi.org/10.1016/j.jchromb.2017.07.019 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Extraction and in vitro screening of potential acetylcholinesterase inhibitors from the leaves of Panax japonicus Sainan Li a, Chengyu Liu b*, Chunming Liu a**, Yuchi Zhang a a

Central Laboratory, Changchun Normal University, No. 677 North Chang-ji Road,

Changchun 130032, China b

Clinical department of Rehabilitation, College of Acupuncture and Massage,

Changchun University of Traditional Chinese Medicine, Changchun 130117, China

* Corresponding author: tel.: +86 431 8617 2224, fax: +86 431 8617 2224 E-mail address: [email protected], (C. Liu)

** Co-Corresponding author: tel.: +86 431 8616 8777, fax: +86 431 8616 8777 E-mail address: [email protected], (C. Liu)

Highlights 

UFLC-MS is an effective method for screening AChE inhibitors from the leaves of Panax japonicus.



Five saponins with potential anti-Alzheimer's disease activity were identified and isolated.



Continuous MAE-CPC was developed for scaled up production of pure AChE inhibitors.



The MAE-CPC parameters were optimized using a multi-exponential function model.

ABSTRACT Ultrafiltration liquid chromatography-mass spectrometry (UFLC-MS) is an efficient method that can be applied to rapidly screen and identify ligands for acetylcholinesterase (AChE) from the leaves of Panax japonicus. Using this method, we identified 5 major compounds, chikusetsusaponins V, Ib, IV, IVa, and IVa ethyl ester, as potent AChE inhibitors, which were assessed for anti-Alzheimer disease activity using the PC12 cell model. A continuous online method, which consisted of microwave-assisted extraction, a solvent concentration tank, and centrifugal partition chromatography (MAE-SCT-CPC), was newly developed for scaled up production of these compounds with high purity and efficiency. The bioactivities of the compounds separated were assessed by the PC12 cell model. This novel approach of using UFLC-MS coupled with MAE-SCT-CPC and a PC12 cell model could be applied to efficiently screen, extract, and separate AChE inhibitors from complex samples, and could serve as an important platform for the large-scale production of functional food and nutraceutical ingredients.

Keywords: Leaves of Panax japonicus, Alzheimer’s disease, acetylcholinesterase, centrifugal partition chromatography.

1. Introduction Alzheimer's Disease (AD) often results in dementia caused by gradual accumulation of amyloid-beta peptide (Aβ) into microscopic plaques, and the twisting of tau proteins into strands of dead and dying neurons. It is also characterized in early stages with chronic inflammatory response and oxidative stress [1]. Rivastigmine, donepezil, and galantamine are the major therapeutic treatments for AD patients [2]. However, the continuous use of these anti-AD medications can lead to side effects such as headache, constipation, dizziness, nausea, vomiting, and loss of appetite, which severely affect patients’ quality of life [1,3,4]. Thus, new therapeutic approaches to treat AD are urgently needed. In addition, it is becoming increasingly important to screen and identify active compounds from crude extracts. At present, in vivo and in vitro methods are used for screening anti-AD chemical compounds [5]. However, the in vivo screening methods are time-consuming, waste a significant amount of human and material resources, and usually consume large amounts of raw material [6]. As the in vitro evaluation method is rapid and usually requires only a small amount of raw material, it is suitable for selection of active components from a variety of foods and medicinal herbs. In vitro screening methods for AD disease therapeutics include use of the acetylcholinesterase (AChE) inhibitory assay [7] and the PC12 cell model [8]. In conventional experiments, in vitro assays are usually carried out by fractionation, which requires several steps to isolate active compounds and traditional analysis. This method is not efficient as it consumes significant time and effort [9]. To improve the efficiency of drug screening, molecular methods relying on binding selectivity and affinity have been proposed and utilized, thus overcoming the challenges of the in vitro approach [10]. Researchers subsequently demonstrated a combined approach using ultrafiltration liquid chromatography-tandem mass spectrometry (UFLC-MS). This method is very efficient in screening active materials from botanical extracts, as it contributes to the separation of the ligand-receptor complexes, which can be identified later by LC-MS. PC12 cells from rat pheochromocytoma are often used in the study of neurodegenerative disease as they have characteristics similar to midbrain dopamine neurons [11,12]. This cell line provides the advantages of rapid screening and short preparation time, and positive results can be verified during the primary neuronal culture. PC12 screening systems can be applied to cell-based high-throughput

screening assays. Scientists believe that Aβ is influential in the pathogenesis of AD [13]. In PC12 cells, neuronal damage can be induced by Aβ(25-35), making PC12 cells an effective tool to study AD [14]. In this study, an AChE inhibitory assay and Aβ(25-35)-treated PC12 cells were used as in vitro methods for evaluating the anti-AD activities of Panax japonicus leaf extract and its chemical constituents. P. japonicus leaves are often utilized in nutraceutical and traditional medicine in Japan and China. Scientific interest in the leaves of P. japonicus has arisen primarily because of the presence of saponins [15,16], which have previously been demonstrated to possess health benefits [17,18]. In preliminary experiments, in vitro screening showed potent AChE inhibition in P. japonicus leaf extract; however, differentiating active compounds in the extract remained challenging and the degree of AChE inhibition was unknown. In this study, we analyzed the inhibitory effect of the extract on AChE inhibitory activity by extracting and separating the active ingredients based on the preliminary results. The results presented in this paper will contribute to the potential application of these active compounds in the prevention and treatment of AD. Fig. 1 illustrates the structures of these compounds. After evaluating their in vitro activity, pure chemical compounds were extracted and isolated to verify their activities. We developed and tested a new and rapid setup that could be used to continuously extract and isolate AChE inhibitors from raw plant materials. It consisted of microwave-assisted extraction (MAE) coupled with a solvent concentration tank (SCT) and centrifugal partition chromatography (CPC). The many advantages of MAE over conventional extraction techniques include high extraction efficiency, high yield, and large capacity [19]. CPC served as the liquid-liquid partition chromatography technology without any solid matrix, eliminating adsorption effects of the stationary phase material, formation of artifacts, tailing of solute peaks, and contamination [20]. Compared with high-performance liquid chromatography (HPLC), CPC not only improves the recovery of samples but also is compatible with various solvent systems, resulting in a larger maximum capacity. In addition, using CPC, a raw sample can be directly introduced into the hollow column [21]. Thus, CPC was adopted for separation in our research. The process was simplified and highly integrated. The natural medicine compounds were effectively extracted and isolated, indicating that this process has great potential for use in industrial applications. The aim of this study was to use the PC12 cell model and AChE binding to

evaluate the potential anti-AD activities of P. japonicus leaf extract. The target active compounds were extracted and isolated by MAE coupled with SCT and CPC (MAE-SCT-CPC). A multi-exponential function was adopted to establish an optimal CPC solvent system. Using this “extraction-concentration-isolation” process, 5 compounds with AChE binding activities, chikusetsusaponins V, Ib, IV, IVa, and IVa ethyl ester, were extracted and purified. Ultra-performance liquid chromatography (UPLC)-MS and nuclear magnetic resonance (NMR) spectroscopy were used to identify and characterize the compounds. The activities of these compounds were assayed in the PC12 cell screening system. Our results demonstrate the potential of the use of P. japonicus leaf extract in the treatment and prevention of AD and other neurodegenerative diseases.

2. Experimental section 2.1. Instruments The MAE apparatus was built in our laboratory at Changchun Normal University, Changchun, China. The extractor had a volume of 1.0 L. The MAE system included an extraction flask with a discharge nozzle, spiral cooler, blender, and solvent feeder. The SCT was constructed by Changchun Normal University; the mechanism of SCT is similar to that of a rotary evaporator and comprised a cylindrical concentrator (inner diameter: 18 cm, length: 30 cm), spiral agitator (diameter: 10 cm, length: 12 cm, pitch: 3 cm), electric heater, circulating water condenser, solvent recovery bottle, and a vacuum device. An SIC CPC-240 system (System Instruments Co., Ltd., Hachioji, Japan) was refitted and used to carry out CPC. The CPC column was formed by stacking the circular partition disk rotor, which contained 2136 cells and had a volume of ~240 mL. Two high-pressure rotary seals were used to connect the CPC column with the injector and the detector. A Waters Acquity H-class instrument that includes a Waters PDA detector (Milford, CT, USA) was used to perform UPLC. A Q-Exactive orbitrap mass spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) was used to carry out the mass spectrometry (MS).

2.2. Reagents and materials Leaves of P. japonicus were obtained from YuYanFang Medicinal Store (Anguo, China). AChE (E.C.3.2.1.20), acetylthiocholine iodide (AtCHI), dithiobisnitrobenzoic

acid (DTNB), and tacrine were obtained from Sigma–Aldrich Chemical Co. (St. Louis, MO, USA). Microcon YM-100 ultrafiltration chambers with a molecular weight cutoff of 100,000 Da were obtained from Millipore (Bedford, MA, USA). HPLC-grade acetonitrile was obtained from Fisher Scientific (Shanghai, China), and a Milli-Q water purification system was obtained from Millipore.

2.3. Screening procedure using an AChE inhibition assay and UFLC-MS The

screening

experiments

were

conducted

by

the

ultrafiltration

HPLC-DAD-MS method. The preparation of incubation mixtures was carried out by mixing 5 μL of 5 mg/mL sample solution and 10 μL AChE in 185 μL ammonium acetate buffer solution (total volume = 200 μL). The buffer comprised 20 mM ammonium acetate and 10% isopropanol at a pH of 6.8. After incubating for 30 min at 37°C, the mixture was transferred to an ultrafiltration chamber and centrifuged at 13000 × g for 10 min at 25°C. Unbound materials on the filter were removed by centrifuging with 100 μL buffer at room temperature. The bound ligands were released by centrifuging for 10 min with 200 μL methanol at 13000 × g 3 times. The solvent was removed under vacuum and the filtrate was redissolved in 50% methanol for HPLC analysis. We also performed control experiments with denatured enzyme for the screening. The binding degree of the compounds to AChE was based on the following equation: Binding degree (%) = (Ab–Ac) /Aa × 100 Here, Aa refers to the peak area of the compound after addition to the buffer without interactions. Ab and Ac refer to the peak areas of the compound that interacts with the active or denatured AChE. We reduced the combined ultrafiltrate under vacuum and performed LC-MS analysis for the released ligands. Before the screening experiments, we carried out control experiments without AChE. A Waters Acquity BEH C18 column (50 × 2.1 mm, 1.7 μm) was adopted to isolate the saponins. The mobile phase contained water (solvent A) and acetonitrile (solvent B). A gradient program was set as 0–15 min, 95–50% A, with a flow rate of 0.4 mL/min, injection volume of 4 L, and check of the interest peaks with a wavelength of 203 nm. The mass spectrometer was operated in positive ion mode for data collection. The samples were injected using an online UPLC system. The capillary

voltage and spray voltage were 10 V and 4.0 kV, respectively. The capillary temperature was 350°C. Fragment ions were obtained by the in-source collision induced dissociation (CID) method.

2.4. PC12 cell cultures and evaluation of cell viability Rat PC12 cells were provided by American Tissue Culture Collection (Manassas, VA, USA) and kept in DMEM medium containing 100 IU/mL penicillin, 100 μg/mL streptomycin, 5% fetal bovine serum, and 5% horse serum at 37°C with 5% CO2 in humidified conditions. The neuroprotective effect of each compound was investigated by dividing the PC12 cells into 6 groups treated with different conditions: non-treated control, 10 μM Aβ(25-35), 10 μM Aβ(25-35) + donepezil hydrochloride (25 μg/mL), and 10 μM Aβ(25-35) + each single compound (25, 50, and 100 μg/mL), n = 6 per group in MTT assays. Two days after seeding the cells, serum–free medium containing the appropriate drug regime was used to refresh the cultures, which were incubated for a further 48 h. Cell viability was assessed using the MTT assay. The process was as follows: before the end of treatment, PC12 cells were placed in a 96-well plate (2 × 104 cells/well) and 10 μL of 5 mg/mL MTT was added to each well. The medium was removed after incubating for 2 h at 37°C and 100 μL dimethyl sulfoxide was added to the wells to dissolve the formazan. The wells were then analyzed by a microplate reader at 340 nm. The relative cell viability was expressed as the percentage relative to the untreated control cells.

2.5. Measurement of partition coefficient values of the centrifugal partition chromatographic solvents The composition of the two-phase solvent system was tested based on the partition coefficient (K) of the target compound of the raw sample. In each experiment, ~ 2.0 mg raw sample was placed into a test tube, and mixed with the equilibrated two-phase solvent system (1.0 mL each phase), which was based on combinations of n-hexane/ n-butanol /methanol/0.02% aqueous trifluoroacetic acid added. The tube was then shaken to ensure thorough partitioning. When equilibrium was reached, the phases were isolated and evaporated under N2 gas until completely dry. The residue was removed by dissolving in methanol and was characterized by UPLC. The K value was calculated by K = As/Am. The optimal volume ratio was calculated by calibrating

the K value with a nonlinear correlation. 2.6. The extraction parameters of microwave assisted extraction An MAE-prime system in the MAE mode was adopted to extract samples. The raw material was milled to powder and passed through a 60-mesh screen (98% passed). The powdered material was extracted via soxhlet extraction by using water-saturated ethyl acetate for the remove of chlorophyll and flavonoids; the soxhlet was extracted for 30 min at 110°C with a liquid-to-solid ratio of 15 mL/g. Then, the extracted material was placed in a fume hood to dry. The dry sample was placed in a round flask for the MAE extraction. The parameters were as follows: the extraction solvent was 50% aqueous ethanol, extraction time was 20 min at 70°C, microwave power was set to 500 W, and liquid-to-solid ratio was 10 mL/g.

2.7. MAE-SCT-CPC for the separation of AChE inhibitors MAE-SCT-CPC was used to extract and isolate AChE inhibitors from raw material. The instrumental setup for the MAE-SCT-CPC is shown in Fig. 2. The setup comprised a solvent configuration system, solvent separator, MAE, solution concentration system, six-port valve, 30 mL sample loop, CPC equipment, and fraction collector. The setup was connected by loops and pumps. First, 400 g powdered ethyl acetate-extracted raw P. japonicus leaves and 4 L extraction solvent (50% aqueous ethanol) were added to the MAE device, and the microwave heater was activated. After the solution was extracted, it was pumped into the SCT by Pump 1. The six-port valve operated to allow ports 1/6, 2/3, and 4/5 to be linked; then, the heater, circulating water condenser, vacuum, and spiral agitator were turned on, and the parameters of the SCT were set as follows: heater temperature, 80 °C; vacuum degree, -0.09 MPa; spiral stirring speed, 70 rpm. After concentration, the concentrate was then pumped in the sample loop via Pump 2. The CPC solvent systems were adjusted according to the K values. The reagents were pumped into the mixer with various flow rates (Fig. 2, upper middle section). The solvent separator was used to separate the mixed solvents, resulting in 2 layers. The upper organic phase was pushed into the valve connecting port 1 to port 6, and then pumped into the column of the CPC, rotating at 1200 rpm. The T-splitter connected the lower phase and six-port valve. The lower aqueous was pumped at a speed of 1.5 mL/min to the CPC column. After the SCT concentration, the six-way valve connected port 1 and port 2. By switching on Pump 3, the lower phase from the solvent separator was pumped into the

six-port valve to avoid the concentrated solution. After the concentrated solution was pumped into the CPC column, the lower aqueous phase was also pumped into the CPC column at 1.5 mL/min to separate target compounds. The system was operated at 25°C. Peak fractions were calculated by elution profile and evaporation was conducted by a rotary evaporator.

2.8. Purity analyses and identifying CPC fractions by UPLC and MS The MAE extract and CPC fractions (potential AChE inhibitors) of the P. japonicus leaf extract obtained by MAE-CPC were analyzed by UPLC, which showed that the purity of all 5 targeted compounds was > 90%. The retention time and MS/MS and NMR data was used to identify the CPC peaks. MS/MS was performed under positive ion mode with a sheath gas at 50 bar and auxiliary flow rate at 10 bar. The capillary voltage and temperature were set to 10 V and 300°C, respectively. The entrance lens voltage, multipole RF amplitude, and ESI needle voltage was 30 V, 400 V, and 4.5 kV, respectively. Ions ranging from 150–2000 molecular weight units were scanned under positive ion mode. NMR spectra was measured by a Bruker AV500 spectrometer with a 1H frequency of 400 MHz and 13C frequency of 100 MHz (Bruker, Beijing, China) at 25°C. Chemical shifts (δ) were measured by ppm and coupling constants (J) were expressed by Hz.

3. Results and discussion 3.1. Evaluation of AChE inhibitory activity and identification of AChE inhibitors The UF-UPLC-MS method has been used to screen bioactive chemicals in complex materials such as foods and herbs owing to its simplicity, accuracy, and minimal sample preparation [22,23]. Fig. 3 illustrates the ultrafiltration AChE screening data from the P. japonicus leaf extract. The crude extract contained several major peaks (Fig. 3a); however, only 5 peaks, eluted at 3.05, 3.51, 4.24, 5.81, and 6.02 min, were released following methanol disruption and subsequently identified as AChE inhibitors (Fig. 3b). Denatured AChE had no or very weak binding with these compounds (Fig. 3c). The degree of binding (DB) of a ligand to an enzyme is an indicator of its inhibitory effect. The results demonstrate that compounds 3, 4, and 1 possessed the highest DB (0.67, 0.66, and 0.64, respectively), followed by compounds 2 (0.50) and

5 (0.49). A higher DB indicates stronger binding with the receptor, and consequently greater inhibitory effect on the enzyme activity. Variation in DB values suggests that not all compounds inhibit enzyme activity to the same extent. This prompted the development of a large-scale separation and purification method for further evaluation of potential AChE inhibitors. The 5 peaks (Fig. 3) that showed binding capacity to AChE were characterized by LC-ESI-MS, and the retention time (tR), MS, and MS/MS fragmentation ions are shown in Table 1. Peaks 1–5 illustrated the compacted protonated molecular ions [M–H]– at m/z 955.4965, 926.4858, 926.4861, 793.4445, and 808.4628, respectively, which indicated molecular weights of 956, 927, 927, 794, and 809 Da, respectively. The errors between the measured and theoretical values of the compounds were all < 2.00 ppm. The MS/MS data of peaks 1–5 are shown in Table 1. Peak 1 was identified as chikusetsusaponin V, while the fragment ions at m/z 775.2665, 731.4554, 569.5654 correlated with glycosidic cleavage fragments as they can remove CO2 from glucuronic acids. The quasi-molecular ion peaks of peaks 2 and 3 showed [M–H]– at m/z 926.48, the SID survey scans of both base ions were 763.4456 [M–H–glc]–, 701.4954

[M–H–glc–H2O–CO2]–,

595.3123

[M–H–glc–2H2O–ara]–,

551.7474

[M–H–glc–2H2O–ara–CO2]–, and 455.5668 [M–H–2glc–ara]– in MS/MS. Peaks 2 and 3 were identified as chikusetsusaponin Ib and chikusetsusaponin IV, respectively, by comparison with the reference standards. Peak 4 was identified as chikusetsusaponin IVa based on the major fragment ions at m/z 631.3955 [M–Glc–H]–, 613.37277 [M–Glc–H2O–H]–, 587.39532 [M–Glc–CO2–H]–, 569.38595 [M–Glc–H2O–CO2–H]–, and 455.4854 [M–2Glc–H]–. Peak 5 was identified as chikusetsusaponin IVa ethyl ester. This compound produces the ion at m/z 643.4532 by the direct loss of glucosyl. Product ions at m/z 587.6632 were found by cleaving the glucose unit and losing one CO2 unit and one CH2 unit. Subsequent neutral loss of H2O from m/z 587.6632 contributed to the formation of product ion [M–glc–H2O–CO2–CH2–H]– at m/z 569.3965. The MS/MS experiments were conducted with the collision energy set at 25–35%. As shown in Fig. 3, peaks 1–5 were chikusetsusaponins V, Ib, IV, IVa, IVa ethyl ester, respectively. The data shown in Table 1 are consistent with findings of previous studies [24-26].

3.2. Calculation and optimization of a two-phase solvent system for CPC using mathematical functions

The selection of a two-phase solvent system determined the degree of CPC separation, as a proper system can provide proper K values for the targeted compounds that were isolated from P. japonicus leaves. The K values of chikusetsusaponins V, Ib, IV, IVa, and IVa ethyl ester partitioned between the upper phase

and

lower

phase

of

n-hexane–n-butanol–methanol–0.02%

aqueous

trifluoroacetic acid at volume ratios of 0.06:1.00:0.08:0.50, 0.08:1.00:0.10:0.60, 0.10:1.00:0.12:0.70, 0.12:1.00:0.14:0.80, 0.14:1.00:0.16:0.90, 0.16:1.00:0.18:1.00, 0.18:1.00:0.20:1.10, 0.20:1.00:0.22:1.20, 0.22:1.00:0.24:1.30, 0.24:1.00:0.26:1.40, and 0.26:1.00:0.28:1.50 were calculated, as shown in Table 2. The K values were reduced when the n-hexane, methanol, and 0.02% aqueous trifluoroacetic acid ratios were increased. Regression analysis was employed to determine the partition coefficient values as a function of the n-hexane/n-butanol, methanol /n-butanol or 0.02% aqueous trifluoroacetic acid /n-butanol ratios. The volume ratios of n-hexane/ n-butanol, methanol / n-butanol and 0.02% aqueous trifluoroacetic acid /n-butanol were then considered as independent variables, while K values were considered as dependent variables to investigate their relationship. We adopted a multivariate exponential function to study the contributions of the independent and dependent variables to achieve an optimal CPC solvent system. According to this model, the ratio of n-hexane to n-butanol was denoted as x1; the ratio of methanol to n-butanol as x2; and the ratio of 0.02% aqueous trifluoroacetic acid to n-butanol as x3. K values of chikusetsusaponins V, Ib, IV, IVa, and IVa ethyl ester were denoted as k1, k2, k3, k4, and k5, respectively. Based on the nonlinear regression analysis, the relationship of k1, k2, k3, k4, k5, and x1, x2, and x3 can be fitted into the exponential functions c1ea1x1+b1+c2ea2x2+b2+c3ea3x3+b3, which can be confirmed by the least square error. The functions of x1, x2 and x3 are shown below: k1=0.0743e–2.3445x1+0.5544+0.0965e–3.6568x2+0.7157+0.0763e–0.5457x3+0.7445, R2=0.9826; k2=0.1003e–2.3521x1+0.6555+0.1231e–3.7445x2+0.8366+0.0522e–0.3552x3+0.9285, R2=0.9903; k3=0.1873e–3.4554x1+0.9455+0.2068e–3.8695x2+1.1556+0.0665e–0.4844x3+1.7956, R2=0.9817; k4=0.2024e–3.9895x1+1.3545+0.2456e–4.5668x2+1.5246+0.0801e–0.6525x3+2.3551, R2=0.9869; k5=0.3052e–4.8556x1+1.7665+0.3385e–5.6568x2+2.0584+0.1031e–0.8142x3+2.8231, R2=0.9962. The partition coefficients k1, k2, k3, k4, and k5 can be optimized with the above functions. Based on the preliminary results, when x1,1 was > 0.25, the partition coefficients of chikusetsusaponins V and Ib were estimated to be < 0.10, which verifies the separation of the two compounds. When x1 was < 0.05, the partition

coefficient of chikusetsusaponin IVa ethyl ester was estimated to be > 7.11, indicating that longer separation and elution times are required for this compound. Thus, x1 was designated to be in the range of 0.06–0.26. Based on similar reasoning, the range of x2 was designated to be 0.08–0.28. Similarly, the range of x3 was defined as 0.50–1.50. For well-separated target compounds requiring less separation time, the k1 value was set from 0.2 to 0.5, and thus the separation factor, defined as the ratio of the two K values (α=K1/K2, where K1>K2), should be > 1.5 for the CPC unit, and k2–1.5k1≥0, k3–1.5k2≥0 k4–1.5k3≥0 and k5–1.5k4≥0 are required. To reduce the separation time, y5 was set to < 4.0, as a K value > 3.0 means a longer separation time. Optimized commands are as follows: 0.05≤x1≤0.25; 0.08≤x2≤028; 0.50≤x3≤1.50; 0.2≤k1≤0.5; k2–1.5k1≥0; k3–1.5k2≥0; k4–1.5k3≥0; k5–1.5k4≥0; k5≤3.0; Min (k1+k2+k3+k4+k5) The additionally required Min (k1+k2+k3+k4+k5) will lead to one result from the commands. A minimum K value results in the shortest separation time. The results calculated by the MATLAB software were as follows: x1=0.13; x2=0.15; x3=1.12; k1=0.24; k2=0.43; k3=0.73; k4=1.27; k5=2.60 The result indicated that the optimal ratio of n-hexane to n-butanol was 0.13, the ratio of methanol to n-butanol was 0.15, and the optimal ratio of 0.02% aqueous trifluoroacetic acid to n-butanol was 1.12. Under the conditions discussed, K values of chikusetsusaponins V, Ib, IV, IVa, and IVa ethyl ester were calculated to be 0.26, 0.41, 0.73, 1.25, and 2.64, respectively. The optimal volume ratio of n-hexane, n-butanol, methanol,

and

0.02%

aqueous

trifluoroacetic

acid was

calculated

to

be

0.13:1.00:0.15:1.12.

3.3. Separation results and analyses of fractions As shown in the CPC chromatogram illustrated in Fig. 4, 5 fractions were found and collected, and ESI–MS/MS and HPLC/ESI/MS were used to identify their formula structures. The HPLC, MS, and NMR data indicate that the compounds 1–5 are chikusetsusaponins V, Ib, IV, IVa, and IVa ethyl ester, respectively. The target peaks of P. japonicus leaves separated by MAE–SCT–CPC were measured by UPLC–DAD; the UPLC chromatograms of compounds 1–5 is demonstrated in Fig. 5. Based on the conditions adopted, purities of all the compounds exceeded 94.8%. The

formula structures are illustrated in Fig. 1. Isolation rates of chikusetsusaponins V, Ib, IV, IVa, and IVa ethyl ester were 0.55, 0.72, 0.66, 0.28, and 0.33 mg/g, respectively, and the purities were 94.8, 95.8, 96.7, 96.4, and 97.8%, respectively.

3.4. Protective effect of extract and target compounds on Aβ(25-35)-induced PC12 cell damage The protective effects against neuronal damage were assessed by pretreating PC12 cells with the extract and purified compounds from the P. japonicus leaves for 1 h, and treating the cells with Aβ(25-35) for 12, 24, or 48 h. Aβ(25-35) treatment alone reduced PC12 cell viability to 37.45%, 26.37%, and 15.12% of the control at 12, 24, and 48 h, respectively (P < 0.001), and pretreatment with either extract or purified single compounds prevented this decrease in cell viability, most notably when the extract was used. Table 3 shows the results, which indicate that these compounds reduce Aβ(25-35)-induced cell damage. Pretreating with P. japonicus leaf extract at various concentrations (25, 50, and 100 μg/mL) effectively reduced the Aβ(25-35)-induced

toxicity

in

PC12

cells.

Consequently,

concentrations

of

approximately 25, 50, and 100 μg/mL can be considered to demonstrate protective activity. No evident toxic effects were found when PC12 cells were treated with chikusetsusaponins Ib, IVa, or IVa ethyl ester at 100 μg/mL. The cell viability of the extract group (100 μg/mL) was 65.12 %, and those of chikusetsusaponins V, Ib, IV, IVa, and IVa ethyl ester (100 μg/mL) groups were 68.36, 71.56, 69.55, 74.69, and 76.03 %, respectively (P < 0.001). The results were similar to those of the AChE inhibition experiments.

4. Conclusion In this study, a UFLC-MS approach combined with a PC12 cell model was proposed to effectively screen and identify AChE inhibitors from P. japonicus leaves. The P. japonicus leaf extract and single compounds isolated from the extract improved the viability of Aβ(25-35)-treated cells in an in vitro AD model using PC12 cells.

Five

chemical

compounds

were

identified

as

AChE

inhibitors:

chikusetsusaponins V, Ib, IV, IVa, and IVa ethyl ester. A subsequent MAE-CPC approach was developed for semi-preparative separation of the AChE inhibitors. The P. japonicus leaf extract obtained by the optimized MAE was subjected to a CPC

system, which was also optimized using mathematical equations. The optimized MAE-CPC method produced 5 AChE inhibitors with > 90% purity. This unique continuous extraction and online isolation method is highly effective, it can be applied to other bioactive compounds in various food or herbal plants, and has significant potential for industrial applications. Compared with existing methods, the technique developed in this study, UFLC-MS combined with MAE-CPC and a PC12 cell model, provided an effective method for screening, extracting, and separating AChE inhibitors from complex samples. Moreover, this approach may serve as an important platform for the large-scale production of functional food and nutraceutical ingredients.

Acknowledgements This work was supported by the National Natural Science Foundation of China (Nos. 31670358, 31370374, and 31500279), the Natural Science Foundation of Jilin Province (Nos. 20160101336JC, 20170520038JH and 20150520140JH), and the Natural Science Foundation of Changchun Normal University (Nos. [2015]008, [2015]009).

References [1] M.K. Song, D.S. Bischoff, A.M. Song, K. Uyemura, D.T. Yamaguchi, BBA Clinical 7 (2017) 41. [2] M.L. Onor, M. Trevistol, E. Aguglia, Clin. Interv. Aging 2 (2007) 17. [3] Y. Değirmenci, H. Keçeci, Int. J. Gerontol. 10 (2016) 240. [4] L.E. Collins, N.E. Paul, S.F. Abbas, C.E. Leser, S.J. Podurgiel, D.J. Galtieri, J.J. Chrobak, Y. Baqi, C.E. Müller, J.D. Salamone, Pharmacol. Biochem. Be. 99 (2011) 414. [5] M. Leirós, E. Alonso, M.E. Rateb, R. Ebel, M. Jaspars, A. Alfonso, L.M. Botana, Neuroscience 305 (2015) 26. [6] H.P. Song, J. Chen, J.Y. Hong, H.P. Hao, Li.W. Qi, J. Lu, Y. Fu, B. Wu, H. Yang, P. Li, Chem. Commun. 51(2015) 1494. [7] A.H. Abou-Donia, F.A. Darwish, S.M. Toaima, E. Shawky, S.S. Takla, J. Chromatogr. B 955–956 (2014) 50. [8] C. Zheng, T. Geetha, M. Gearing, J.R. Babu, J. Neurochem. 133 (2015) 919. [9] M.N. Alam, N.J. Bristi, M. Rafiquzzaman, Saudi Pharm. J. 21 (2013) 143. [10] X. Ouyang, P.E.G. Leonards, Z. Tousova, J. Slobodnik, J. Boer, M.H. Lamoree, Anal. Chem. 88 (2016) 2353. [11] J. Xu, X. Sun, X. Liu, M. Peng, S. Li, D.Q. Jin, D. Lee, M. Bartlam, Y. Guo, J. Funct. Foods 23 (2016) 565. [12] J.S. Steffan, L. Bodai, J. Pallos, M. Poelman, A. McCampbell, B.L. Apostol, A. Kazantsev, E. Schmidt, Y.Z. Zhu, M. Greenwald, R. Kurokawa, D.E. Housman, G.R. Jackson, J.L. Marsh, L.M. Thompson, Nature, 413 (2001) 739. [13] S. Varadarajan, J. Kanski, M. Aksenova, C. Lauderback, D.A. Butterfield, J. Am. Chem. Soc. 123 (2001) 5625. [14] A. Khan, K. Vaibhav, H. Javed, R. Tabassum, M.E. Ahmed, M.M. Khan, M.B. Khan, P. Shrivastava, F. Islam, M.S. Siddiqui, M.M. Safhi, F. Islam, Neurochem. Res. 39 (2014) 344. [15] K. Yoshizaki, S. Yahara, Nat. Prod. Commun. 7 (2012) 491. [16] K. Yoshizaki, H.P. Devkota, S. Yahara, Chem. Pharm. Bull. 61 (2013) 273. [17] H. He, J. Xu, Y. Xu, C. Zhang, H. Wang, Y. He, T. Wang, D. Yuan, J. Ethnopharmacol. 140 (2012) 73.

[18] L.K. Han, Y.N. Zheng, M. Yoshikawa, H. Okuda, Y. Kimura, BMC Complem. Altern. M. 5 (2005) 1. [19] H. Gu, F. Chen, Q. Zhang, J. Zang, J. Chromatogr. B 1014 (2016) 45. [20] A. Sólyomváry, Á. Alberti, A. Darcsi, R. Könye, Ge. Tóth, B. Noszál, I. Molnár-Perl, L. Lorántfy, J. Dobos, L. Őrfi, S. Béni, I. Boldizsár, J. Chromatogr. B 1052 (2017) 142. [21] A. Abedini, S. Chollet, A. Angelis, N. Borie, J.M. Nuzillard, A.L. Skaltsounis, R. Reynaud, S.C. Gangloff, J.H. Renault, J. Hubert, J. Chromatogr. B 1029–1030 (2016) 121. [22] M. Chen, L. Liu, X. Chen, J. Sep. Sci. 37 (2014) 1546. [23] Z. Wang, S.H. Hwang, B. Huang, S.S. Lim, J. Chromatogr. B 1002 (2015) 319. [24] M. Hattori, Y. Kawata, N. Kakiuchi, K. Matsuura, T. Tomimori, Chem. Pharm. Bull. 36 (1988) 4467. [25] G.C. Kite, M.J.R. Howes, C.J. Leon, M.S.J. Simmonds, Rapid Commun. Mass Spectrom. 17 (2003) 238. [26] Y.J. Li, H.L. Wei, L.W. Qi, J. Chen, M.T. Ren, P. Li, Rapid Commun. Mass Spectrom. 24 (2010) 2975.

Figure captions:

Figure 1. Structures of AChE inhibitors extracted from P. japonicus leaves.

Figure 2. Schematic representation of the instrumental setup of the continuous online extraction and separation system comprising MAE, SCT, and CPC.

Figure 3. UPLC profile of the chemical constituents of P. japonicus leaves obtained by ultrafiltration. (a) P. japonicus leaf extract e(the concentration of the extract was the same as that in the ultrafiltration experiment); (b) Compounds bound to AChE; (c) Compounds bound to denatured AChE. 1: chikusetsusaponin IV; 2: chikusetsusaponin IVa; 3: chikusetsusaponin IVa ethyl ester; 4: chikusetsusaponin V; and 5: chikusetsusaponin Ib.

Figure 4. CPC chromatogram of the MAE extract with a two-phase solvent system of n-hexane / n-butanol / methanol / 0.02% aqueous trifluoroacetic acid at a volume ratio of 0.13:1.00:0.15:1.12. 1: chikusetsusaponin IV; 2: chikusetsusaponin IVa; 3: chikusetsusaponin IVa ethyl ester; 4: chikusetsusaponin V; and 5: chikusetsusaponin Ib.

Figure 5. UPLC of peak fractions eluted by CPC using MAE. 1: chikusetsusaponin IV; 2: chikusetsusaponin IVa; 3: chikusetsusaponin IVa ethyl ester; 4: chikusetsusaponin V; and 5: chikusetsusaponin Ib.

Table 1 – Major AChE inhibitors identified in P. japonicus leaves by LC–MS. [M–H]– Peak

tR/mi

No.

n

UPLC

(m/z)

[M–H]–

measured

(m/z)

Identificat

MS2, m/z

ion

calculated

peak Formula

area ratio 1 a

[M–H–glc]–,

775.2665 1

3.05

956.4965

956.4986

731.4554

[M–H–glc–CO2]–, 569.5654

[M–H–(glc–H2O)

–glc–CO2]– 763.4456

2

3.51

926.4858

926.4880

[M–H–glc]–,

Chikusets

551.7474

usaponin

[M–H–glc]–,

4.24

926.4861

926.4880

C48H76O19

0.64

14.21

C47H74O18

0.50

22.32

C47H74O18

0.67

16.32

C42H66O14

0.66

15.61

C43H68O14

0.49

13.89

Ib

701.4954

[M–H–glc–H2O–CO2]–, [M–H–glc–2H2O–ara]–,

595.3123

ratio 2b

V

595.3123

763.4456

area

701.4954

[M–H–glc–2H2O–ara–CO2]–,

3

usaponin

[M–H–glc–2H2O–ara]–,

455.5668

peak

Chikusets

[M–H–glc–H2O–CO2]–,

[M–H–2glc–ara]–

UPLC

551.7474 [M–H–glc–2H2O–ara–CO2]–,

Chikusets usaponin IV

455.5668 [M–H–2glc–ara]– 631.5448

4

5.81

794.4445

794.4458

[M–glc–H]–,

613.7542

[M–glc–H2O–H]–,

587.6455

Chikusets

[M–glc–CO2

–H]–,

569.6485

usaponin

[M–glc–H2O–CO2–H]–,

455.8754

IVa

[M–2glc–H]– 643.4532 [M–glc–H]–, 5

6.02

808.4618

a

808.4614

Chikusets –H]–,

587.6632 [M–glc–CO2–CH2

usaponin

569.3965

IVa ethyl

[M–glc–H2O–CO2–CH2–H]–

ester

The UPLC peak area ratios were calculated as the UPLC peak area (compound with

AChE) divided by the UPLC peak area of the extract. b

The UPLC peak area ratios were calculated as the UPLC peak area (compound with

AChE) divided by the UPLC peak area (compound with denatured AChE).

Table 2 Partition coefficient values (KUP/LP) of chikusetsusaponins V, Ib, IV, IVa, and IVa ethyl ester from P. japonicus leaves at different volume ratios of n-hexane to n-butanol, methanol to n-butanol, and 0.02% aqueous trifluoroacetic acid to n-butanol. n-hexane–n-butanol–methan No. ol–0.02%

aqueous

trifluoroacetic acid (v:v:v:v)

Values of the independent variables and the dependent variables x1a

x2

x3

k1b

k2

k3

k4

K5

1

0.06:1.00:0.08:0.50

0.06

0.08

0.50

0.69

1.36

2.18

3.25

7.11

2

0.08:1.00:0.10:0.60

0.08

0.10

0.60

0.59

1.03

1.82

2.74

5.69

3

0.10:1.00:0.12:0.70

0.10

0.12

0.70

0.48

0.78

1.50

2.23

4.85

4

0.12:1.00:0.14:0.80

0.12

0.14

0.80

0.41

0.67

1.23

1.89

4.32

5

0.14:1.00:0.16:0.90

0.14

0.16

0.90

0.34

0.51

0.95

1.59

3.09

6

0.16:1.00:0.18:1.00

0.16

0.18

1.00

0.28

0.37

0.82

1.35

2.55

7

0.18:1.00:0.20:1.10

0.18

0.20

1.10

0.25

0.32

0.68

1.01

2.11

8

0.20:1.00:0.22:1.20

0.20

0.22

1.20

0.21

0.23

0.63

0.94

1.74

9

0.22:1.00:0.24:1.30

0.22

0.24

1.30

0.18

0.19

0.51

0.89

1.51

10

0.24:1.00:0.26:1.40

0.24

0.26

1.40

0.15

0.16

0.42

0.66

1.14

11

0.26:1.00:0.28:1.50

0.26

0.28

1.50

0.13

0.13

0.33

0.56

0.89

a

x1: the ratio of n-hexane to n-butanol; x2: methanol to n-butanol; x3: 0.02% aqueous

trifluoroacetic acid to n-butanol. b

k1: K value of chikusetsusaponin V; k2: K value of chikusetsusaponin Ib; k3: K value

of chikusetsusaponin IV; k4: K value of chikusetsusaponin IVa; k5: K value of chikusetsusaponin IVa ethyl ester.

Table 3 Effect of extract and single compounds from P. japonicus leaves on cell viability in Aβ(25-35)-treated PC12 cells. Group

Concentration (μg/mL)

A340

Cell survival rate (%)

Control

0

0.804 ± 0.112

100.0

Aβ(25-35) + donepezil hydrochloride

25

0.711 ± 0.189

88.43

Aβ(25-35) group

–

0.122 ± 0.083

15.12

25

0.383 ± 0.102

47.65

50

0.484 ± 0.118

60.26

100

0.524 ± 0.125

65.12

25

0.410 ± 0.154

50.94

50

0.485 ± 0.186

60.31

100

0.550 ± 0.168

68.36

25

0.393 ± 0.102

48.85

50

0.525 ± 0.144

65.33

100

0.575 ± 0.157

71.56

25

0.374 ± 0.185

46.51

50

0.482 ± 0.156

59.96

100

0.559 ± 0.198

69.55

25

0.421 ± 0.144

52.41

50

0.543 ± 0.158

67.56

100

0.601 ± 0.146

74.69

25

0.432 ± 0.128

53.67

50

0.517 ± 0.147

64.32

100

0.611 ± 0.155

76.03

Aβ(25-35) + extract

Aβ(25-35) + chikusetsusaponin V

Aβ(25-35) + chikusetsusaponin Ib

Aβ(25-35) + chikusetsusaponin IV

Aβ(25-35) + chikusetsusaponin IVa

Aβ(25-35) + chikusetsusaponin IVa ethyl ester