ЖУРНАЛ АНАЛИТИЧЕСКОЙ ХИМИИ, 2013, том 68, № 6, с. 585-591
ОРИГИНАЛЬНЫЕ СТАТЬИ
УДК 543
OPTIMIZATION OF THE MICELLAR ELECTROKINETIC CAPILLARY CHROMATOGRAPHIC DETERMINATION OF DAURICINE AND DAURISOLINE IN Rhizoma Menispermi AND ITS HERBAL MEDICINE USING EXPERIMENTAL DESIGN AND RADIAL BASIS FUNCTION
NEURAL NETWORK © 2013 г. Wenjuan Lu, Yonglei Chen, Hongli Chen, Xingguo Chen
Department of Chemistry, Lanzhou University
Lanzhou, 730000 P.R. China Reccieved 10.06.2011; in final form 12.04.2012
Orthogonal design has been applied to the optimization of separation and determination of dauricine and daurisoline in Rhizoma Menispermi and its herbal medicine by micellar electrokinetic capillary chromatography. Operational variables, such as the voltage, micelle concentration, buffer concentration and pH were optimized. Their different effects on peak resolution were studied by the experimental design method. Optimized separation conditions were obtained and successfully applied to the separation and determination of dauricine and daurisoline in real samples. The proposed method allows alkaloids in real samples to be determined within 15 min using a buffer system composed of 25 mM HAc, 25 mM NaAc and 2% polyoxyethylene sorbitan monolaurate (Tween-20) (pH 4.5). In addition, a radial basis function neural network with a "4-18-1" structure was developed based on the experimental results of orthogonal design and uniform design, and applied to the prediction of peak resolution of dauricine and daurisoline under the optimum separation conditions given by orthogonal design. The predicted results were in good agreement with the experimental values, indicating that radial basis function neural network may be a potential method for the selection of separation conditions in capillary electrophoresis.
Keywords: micellar electrokinetic capillary chromatography, experimental design, radial basis function neural network, Rhizoma Menispermi, dauricine, daurisoline.
DOI: 10.7868/S0044450213060169
Dauricine and daurisoline are two isoquinoline alkaloids existed in Rhizoma Menispermi, which are the main effective components in Rhizoma Menispermi and have been known to show many pharmacological activities and therapeutic effects [1—4]. Both of them have antiarrhythmic effects and anti-inflammation activity. Furthermore, dauricine can effectively reduce the fibrosis of cardiac hypertrophy and enhance the ability of anti-oxidant in heart tissue; it also shows obvious therapeutic effects on cerebral ischemia-reper-fusion injury in rats. Experimental arrhythmic models show that the antiarrhythmic effect of daurisoline is more potent than that of dauricine. So, it is necessary to establish a suitable analytical method to determine the two active alkaloids in Rhizoma Menispermi and its herbal preparations.
By now, many high-performance liquid chromatography (HPLC) methods have been reported for the determination of dauricine or dauricine and dauriso-line in Rhizoma Menispermi [5—8]. However, HPLC presents low efficiency and is time-consuming. Compared with HPLC methods, capillary electrophoresis
(CE) has the advantages of small sample and reagent consumption, high resolution and high speed of analysis. Up to now, there have been many capillary elec-trophoresis applications in active herbal composition separation [9—12], but no capillary electrophoresis method for the simultaneous determination of the two isoquinoline alkaloids in Rhizoma Menispermi and its herbal preparations has been reported.
Many experimental parameters in CE have to be optimized, such as buffer concentration, pH, concentration of organic modifier, separation voltage, and injection time, etc., which usually affect peak resolution significantly. It is usually difficult to find suitable experimental conditions for a given separation task quickly and efficiently. Therefore, good experimental design becomes vitally important. Fortunately, various chemometrics-based techniques including multivari-ate experimental design and response surface methodology (RSM) have been introduced to aid in optimizing the performance of a system. The importance and theoretical concepts of experimental design and optimization methodology in research have been thor-
dauricine
daurisoline
Fig. 1. Chemical structures of the analytes.
oughly discussed in many publications [13—17]. Orthogonal design, which only focuses on the main effects of the factors, allowing the number of experiments to be reduced evidently, has already shown its effective usefulness in LC [18] and CE [19, 20] separations. Uniform design has also been applied to the optimization of separations in CE [21]. Recently, artificial neural networks (ANNs) have been applied, either separately or in combination with many of the experimental design techniques discussed above, to CE separation optimizations [20, 22—24]. There are some kinds of ANNs such as back-propagation (BP) neural network, a multilayer perceptron (MLP), radial basis function neural networks (RBFNN), etc. Especially, the radial basis function network (RBFNN) has the advantages of small training times and is guaranteed to reach the global minimum of error surface during training. The optimization of its topology and learning parameters are easy to implement. Many problems in chemistry and biology have been successfully solved by RBFNN [25-27].
In this work, a micellar electrokinetic capillary chromatography (MEKC) method was first developed for the separation and determination of dauricine and daurisoline in Rhizoma Menispermi and its herbal medicine. Orthogonal design and uniform design techniques were apllied to the separation, the effects of various factors on peak resolution (R) were discussed with particular emphasis. In addition, a RBFNN model was developed and used to predict R (including R under optimized condition) and satisfactory results were obtained.
EXPERIMENTAL
Chemicals and materials. Standards of dauricine and daurisoline were obtained from Wuhu Delta biotechnology Company, Anhui, China. Their structures are shown in Fig. 1. The crude drug of Rhizoma Menispermi, and Beidougen capsule were purchased from local drug store. All other chemicals were of analytical reagent grade. All solution and buffer were
made in distilled water. Distilled water was used in all through the experiments.
Solutions preparation. Stock standard solutions of dauricine and daurisoline were prepared in distilled water (450 and 400 ^g/mL). Working standard solutions, ranging from 3 to 450 ^g/mL of dauricine and daurisoline, were prepared from stock standard solutions with distilled water. The running buffer contains 25 mM HAc and 25 mM NaAc (pH 4.5) - 2% Tween-20. The running buffer was prepared daily from stock solution of 0.1 M HAc and NaAc.
Sample preparation. The powdered (1.0 g) of Rhi-zoma Menispermi was accurately weighed and dissolved with 10 mL methanol in an ultrasonic bath for 30 min, then the filtrate solution was filtered through 0.45 ^m syringe filters. Same as the former, 0.06 g powder (2 capsules) from mixture of 20 Beidougen capsules was accurately weighed and dissolved with 10 mL methanol in an ultrasonic bath for 30 min, and then the solutions were filtered through 0.45 ^m syringe filters. All the filtrates were directly injected into the CE equipment with appropriate dilution. All solutions were filtered through 0.45 ^m syringe filters before use.
Apparatus. A CL1030 Capillary system (Beijing Cailu Science Instrument Company, Beijing, China) was used. Uncoated silica separation capillaries of 53 cm (44.5 cm effective length) x 75 ^m ID x 375 ^m OD (Yongnian Optical Fiber Factory, Hebei, China) were used throughout the study. UV detection was carried out at 214 nm. The data acquisition was carried out with a HW-2000 Chromatography Workstation. Samples were introduced from the end of the capillary by hydrodynamic injection, where the sample vial was raised by 15.5 cm for 5 s. At the beginning of each working day, the capillary was flushed sequentially with 0.5 M NaOH for 10 min, distilled water for 10 min and the running electrolyte for 10 min. A PHS-10A acidity meter (Xiaoshan Science Instrumentation Factory, Zhejiang, China) was used for the pH measurement.
Table 1. Experimental design chart, four variables, five levels orthogonal design (1—25), uniform design (26—30)
Experiment no. c0a (mmol) Tween-20 (%) Buffer pH Voltage (KV) Resolution (R)
Experimental (n = 3) RSD (%) Predicted Deviation
1 10(1) 0.5(1) 3.0(1) 15(1) 1.06 3.1 1.07 -0.01
2 10(1) 1.0(2) 3.5(2) 18(2) 1.32 2.8 1.28 0.04
3 10(1) 1.5(3) 4.0(3) 20(3) 1.47 2.8 1.32 0.15
4 10(1) 2.0(4) 4.5(4) 22(4) 1.49 2.6 1.43 0.06
5 10(1) 2.5(5) 5.0(5) 25(5) 1.34 3.0 1.36 -0.02
6 15(2) 0.5(1) 3.5(2) 20(3) 0.88 3.5 0.90 -0.02
7 15(2) 1.0(2) 4.0(3) 22(4) 1.20 2.6 1.22 -0.02
8 15(2) 1.5(3) 4.5(4) 25(5) 1.30 2.4 1.33 -0.03
9 15(2) 2.0(4) 5.0(5) 15(1) 1.45 2.7 1.45 0.00
10 15(2) 2.5(5) 3.0(1) 18(2) 1.47 3.2 1.48 -0.01
11 20(3) 0.5(1) 4.0(3) 25(5) 0.68 3.8 0.75 -0.07
12 20(3) 1.0(2) 4.5(4) 15(1) 1.42 2.3 1.41 0.01
13 20(3) 1.5(3) 5.0(5) 18(2) 1.47 2.8 1.50 -0.03
14 20(3) 2.0(4) 3.0(1) 20(3) 1.60 2.4 1.57 0.03
15 20(3) 2.5(5) 3.5(2) 22(4) 1.38 2.3 1.41 -0.03
16 25(4) 0.5(1) 4.5(4) 18(2) 1.28 2.2 1.29 -0.01
17 25(4) 1.0(2) 5.0(5) 20(3) 1.33 3.1 1.29 0.04
18 25(4) 1.5(3) 3.0(1) 22(4) 1.51 2.7 1.40 0.11
19 25(4) 2.0(4) 3.5(2) 25(5) 1.69 2.9 1.68 0.01
20 25(4) 2.5(5) 4.0(3) 15(1) 1.46 3.1 1.47 -0.01
21 30(5) 0.5(1) 5.0(5) 22(4) 1.08 3.1 1.09 -0.01
22 30(5) 1.0(2) 3.0(1) 25(5) 1.17 2.9 1.17 0.00
23 30(5) 1.5(3) 3.5(2) 15(1) 1.40 2.6 1.39 0.01
24 30(5) 2.0(4) 4.0(3) 18(2) 1.57 2.7 1.52 0.05
25 30(5) 2.5(5) 4.5(4) 20(3) 1.32 2.8 1.29 0.03
26 10(1) 1.0(2) 4.5(4) 20(3) 1.31 2.8 1.34 -0.03
27 15(2) 2.0(4) 4.0(3) 15(1) 1.73 2.2 1.46 0.27
28 20(3) 0.5(1) 3.5(2) 22(4) 0.75 3.3 0.73 0.02
29 25(4) 1.5(3) 3.0(1) 18(2) 1.53 2.6 1.56 -0.03
30 30(5) 2.5(5) 5.0(5) 25(5) 1.46 2.7 1.47 -0
Для дальнейшего прочтения статьи необходимо приобрести полный текст. Статьи высылаются в формате PDF на указанную при оплате почту. Время доставки составляет менее 10 минут. Стоимость одной статьи — 150 рублей.