- Email: [email protected]
Determination of mannitol and three sugars in Ligustrum lucidum Ait. by capillary electrophoresis with electrochemical detection
Analytica Chimica Acta 530 (2005) 15–21
Determination of mannitol and three sugars in Ligustrum lucidum Ait. by capillary electrophoresis with electrochemical detection Gang Chena,∗ , Luyan Zhanga , Xingliang Wua , Jiannong Yeb a
Department of Chemistry, Fudan University, 211 138 Yixueyuan Rd., Shanghai 200032, PR China b Department of Chemistry, East China Normal University, Shanghai 200062, China Received 28 March 2004; received in revised form 23 August 2004; accepted 23 August 2004 Available online 2 December 2004
Abstract A method based on capillary electrophoresis with electrochemical detection has been developed for the separation and determination of mannitol, sucrose, glucose, and fructose in Ligustrum lucidum Ait. for the first time. Effects of several important factors such as the concentration of NaOH, separation voltage, injection time, and detection potential were investigated to acquire the optimum conditions. The detection electrode was a 300 m diameter copper disc electrode at a working potential of +0.65 V (versus saturated calomel electrode (SCE)). The four analytes can be well separated within 13 min in a 40 cm length fused-silica capillary at a separation voltage of 12 kV in a 75 mM NaOH aqueous solution. The relation between peak current and analyte concentration was linear over about three orders of magnitude with detection limits (S/N = 3) ranging from 1 to 2 M for all analytes. The proposed method has been successfully applied to monitor the mannitol and sugar contents in the plant samples at different growth stages with satisfactory assay results. © 2004 Elsevier B.V. All rights reserved. Keywords: Ligustrum lucidum Ait.; Mannitol; Sucrose; Glucose; Fructose; Capillary electrophoresis; Electrochemical detection
1. Introduction In the present, more and more people are becoming interested in plant medicines, because of their low toxicity and good therapeutical performance. Chinese traditional medicine, Fructus Ligustri Lucidi, is the dried ripen fruit of Ligustrum lucidum Ait. As a commonly used herbal medicine, it has the therapeutical functions of nourishing liver and kidney, stimulating heart, enhancing organism immunization, diuresis, laxation, anti-inflammatory, etc. [1,2]. The leaf of L. lucidum Ait. has also been used as a Chinese traditional herbal drug with functions of anti-bacteria, anti-inflammation, and suppressing cough [2]. A variety of physiologically active compounds (such as ligustroside, nuzhenide, oleanolic acid, ursolic acid, betulin, mannitol, glucose, quercetin, etc.) have been found presented in L. lu∗
Corresponding author. Tel.: +86 21 6466 1130; fax: +86 21 6564 1740. E-mail address: [email protected] (G. Chen).
0003-2670/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.aca.2004.08.053
cidum Ait. Different functions of the herbal drugs come from different active constituents. [2]. It has been demonstrated that the diuresis and laxation activities of Fructus Ligustri Lucidi can be attributed to mannitol, a commonly used diuretic or laxative [2,3]. The mannitol content is an important parameter for evaluating the quality of Fructus Ligustri Lucidi [3]. As the primary metabolites, sucrose, glucose, and fructose are found widely presented in plants. Han et al. have found that higher contents of sugars can indicate the better quality of some herbal drugs [4]. Hence, it is interesting to establish some simple, economical and accurate methods for the determination of mannitol, sucrose, glucose, and fructose in L. lucidum Ait. Li and Liu have determined mannitol in Fructus Ligustri Lucidi by spectrophotometry for quality control [3]. Recently, Liu et al. have reported an approach for the simultaneous determination of oleanolic acid and ursolic acid in this plant medicine based on micellar electrokinetic capillary chromatography [5]. In addition, thin layer chromatogra-
16
G. Chen et al. / Analytica Chimica Acta 530 (2005) 15–21
phy (TLC) [6] and high performance liquid chromatography (HPLC) [7,8] have been employed for the determination of oleanolic acid, salidroside, specnuezhenide and other watersoluble active constituents in the fruits of L. lucidum Ait. Capillary electrophoresis (CE) has been widely used in the analysis of sugars [9–13]. The advances in separation and detection approaches of derivatized and underivatized carbohydrates are reviewed by El Rassi [14]. Nowadays, the application of CE for the separation of various active constituents in medicinal plants has become increasingly widespread because of its minimal sample volume requirement, short analysis time and high separation efficiency [15,16]. Electrochemical detection (ED) typically operated in the amperometric mode can be coupled with CE to provide high sensitivity and selectivity for the determination of electroactive substances [17–19]. Direct electrooxidation at platinum or gold electrodes has gained some popularity for the detection of underivatized carbohydrates [20,21]. To reduce the fouling of platinum and gold electrodes, the applied potential for these electrodes must be continuously pulsed based on a complicated procedure. Alternatively, the work on copper electrode by Baldwin and co-workers is of particular interest for the direct detection of carbohydrates in strongly alkaline medium where signal responses are achieved at constant potential without any electrode fouling [12,22]. This detection approach is less complicated, less expensive, but still sensitive. It can be employed to detect not only carbohydrates themselves but also their derivatives such as alditols and sugar acids [12]. In 2000, U.S. Food and Drug Administration (FDA) published a draft of Guidance for Industry Botanical Drug Products. Before a plant drug becomes logically marketed, its spectroscopic or chromatographic fingerprints and chemical assay of characteristic markers are required. CE should find more applications in this area. In this study, CE-ED was employed for the determination of mannitol, sucrose, glucose, and fructose in L. lucidum Ait. without derivatization. It has been found that the mannitol and sugar contents changed at different growth stages. This method is simple, sensitive, reliable, and efficient providing not only a way for evaluating the quality of plant medicines made from L. lucidum Ait. in marketplaces, but also an excellent method for quality control in medicinal factories and constituent investigation of other plants. To our best knowledge, there are no reports published on the determination of mannitol and sugars in L. lucidum Ait. by CE. The optimization, detailed characterization, and advantages of CE-ED approach are reported in the following sections in connection to the measurement of mannitol and three sugars.
received without further purification. The fruits and leaves of L. lucidum Ait. (Glossy Privet) were collected from the east campus of Medical Center of Fudan University (Shanghai, China) and were kindly identified by Professor D. Chen (Department of Pharmacognosy, Medical Center of Fudan University, Shanghai, China). All aqueous solutions were made up in doubly distilled water. Other chemicals were of analytical grade. Stock solutions of mannitol, sucrose, glucose, and fructose (50 mM) were prepared in doubly distilled water and were kept in a 4 ◦ C refrigerator. They were stable for at least 1 month. The electrophoretic separation medium was 75 mM NaOH aqueous solution unless mentioned otherwise. The stock solutions were diluted to desired concentration with the separation medium just prior to use. 2.2. Apparatus The CE-ED system used has been described previously [23]. A ±30 kV high-voltage dc power supply (Shanghai Institute of Nuclear Research, China) provided a separation voltage between the ends of the capillary. The inlet of the capillary was held at a positive potential and the outlet of capillary was maintained at ground. The separations were proceeded in a 40 cm length of 25 m i.d. and 360 m o.d. fused-silica capillary (Polymicro Technologies, Phoenix, AZ, USA). A three-electrode electrochemical cell consisting of a laboratory-made 300 m diameter copper disc working electrode, a platinum auxiliary electrode and a saturated calomel electrode (SCE) as the reference electrode, was used in combination with a BAS LC-4C amperometric detector (Bioanalytical Systems Inc., West Lafayette, IN, USA). The filter of the detector was set at 0.1 Hz. The working electrode was positioned carefully opposite the outlet of the capillary with the aid of a micromanipulator (CORRECT, Tokyo, Japan) and arranged in a wall-jet configuration. The distance between the tip of the working electrode and the capillary outlet was adjusted to ∼25 m by comparison with the bore (25 m) in the capillary while being viewed under a microscope. The electropherograms were recorded using a LKB·REC 1 chart record (Pharmacia, Sweden). A YS 38-1000 220 V alternate constant-voltage power supply (Shanghai Instrumental Transformer Factory, Shanghai, China) was employed to suppress the voltage fluctuation of the power line. The whole system was assembled in a 10 m2 Faraday room that was airconditioned at 20 ◦ C to minimize the effects of external noise sources. 2.3. Sample preparation
2. Experimental 2.1. Reagent and solutions Mannitol, sucrose, glucose, and fructose were all purchased from Sigma (St. Louis, MO, USA) and were used as
Three batches of fruits and leaves of L. lucidum Ait. were collected from the same tree on different days and dried in the air. The batch numbers (031110, 031130, and 031220) were given based on the collection dates. All plant samples were dried at 60 ◦ C for 3 h and then were pulverized. An accurate
G. Chen et al. / Analytica Chimica Acta 530 (2005) 15–21
17
weight amount of the powders (2.5 g) was dispersed in 100 ml doubly distilled water and was kept in a 90 ◦ C water bath for 3 h. The mixture was subsequently sonicated for 30 min and filtered through a filter-paper. The extract was diluted using 75 mM NaOH aqueous solution at a ratio of 10 (1–10) just prior to CE analysis. 2.4. Procedures Before use, the copper disc electrode was successively polished with emery paper and alumina powder, sonicated in doubly distilled water. CE was performed at a separation voltage of 12 kV unless otherwise indicated. The potential applied to the working electrode was +0.65 V (versus SCE). Samples were injected electrokinetically at 12 kV for 6 s. Moreover, sample solutions, standard solutions, and the separation medium were all filtered through a syringe cellulose acetate filter (0.22 m) prior to their use. Peak identification was performed by standard addition method.
3. Results and discussion 3.1. Hydrodynamic voltammograms (HDVs) Alditols and carbohydrates are most normally electroactive at carbon electrodes, the most commonly used working electrode in ED. So, a variety of metal electrodes made of copper, nickel, gold, and platinum have been employed for the electrochemical detection of alditols, carbohydrates, sugar acids, etc. [12,24–27]. Among them, copper and nickel are most widely utilized. Carbohydrates and alditols can be detected by copper electrodes at constant applied potential in strongly alkaline media based on the electrocatalytic oxidation [12,22]. It has been revealed that the minimum structural requirement for facile oxidation is the presence of at least two nearby hydroxyl groups in the compound to be oxidized. All these oxidations were found to consist of many-electron processes with C C bond cleavage involved. Coulometric studies were carried out on glucose as well as its acid derivatives, which suggested the oxidation of glucose on a copper electrode to be a 12-electron process with the sole product appeared to be the formate ion [28]. It has been demonstrated that the oxidation of disaccharides is different from that of monosaccharides on the copper electrode in alkaline medium. The numbers of electrons transferred during the oxidation are 12 for monosaccharides and 3–13 for disaccharides [29]. In the present work, copper disc electrode was employed for the electrochemical detection of mannitol, sucrose, glucose, and fructose at constant applied potentials in 75 mM NaOH aqueous solution. The potential applied to the working electrode directly affects the sensitivity and detection limits of this method and it is necessary to determine the hydrodynamic voltammograms (HDVs) for the analytes to obtain the optimum potential. Fig. 1 depicts the HDVs for the catalytic oxidation
Fig. 1. Hydrodynamic voltammograms (HDVs, A) for 0.5 mM of mannitol, sucrose, glucose, and fructose in CE. Fused-silica capillary: 25 m i.d. × 40 cm length; working electrode: 300 m diameter copper disc electrode; running buffer: 75 mM NaOH; separation and injection voltage: 12 kV; injection time: 6 s.
of mannitol, sucrose, glucose, and fructose using the copper disc electrode detector. The curves were recorded pointwise from +0.4 to +0.7 V (versus SCE) in steps of 50 mV using a separation voltage of 12 kV. All the analytes display similar profiles, with an increase of the response starting at +0.45 V (versus SCE). The current rises rapidly upon raising the potential above +0.50 V (versus SCE). Although an applied potential greater than +0.65 V (versus SCE) results in higher peak currents, both the baseline noise and the background current increase substantially. The high background current leads to an unstable baseline, which is a disadvantage for the sensitive and stable detection. Considering the sensitivity and background current, subsequent amperometric detection work employed a detection potential of +0.65 V (versus SCE) that offered the most favorable signal-to-noise characteristics. The stability of the working electrode was good and the reproducibility was high at the optimum potential. 3.2. Effects of the NaOH concentration Alditols and sugars are both neutral compounds within the physiological pH range. In the present work, strongly alkaline separation medium (12.5–100 mM NaOH) was used to keep all the analytes in anionic form and the activity of the copper electrode for the electrocatalytic detection of mannitol, sucrose, glucose, and fructose. The separation of these analytes by CE is based on their degrees of dissociation [12]. The NaOH concentration in the separation medium affect the zeta-potential (ζ) of the inner wall of the capillary, the viscosity coefficient of the solution, the electroosmotic flow (EOF) as well as the overall charge of the analytes, which determine the migration time, peak height, and the separation of the analytes [30]. The effect of NaOH concentration on the migration time of the analytes is illustrated in Fig. 2A. Upon raising NaOH concentration from 12.5 to 100 mM, the migration time increases rapidly with the resolution improved for all analytes. When NaOH concentration is lower than 25 mM,
18
G. Chen et al. / Analytica Chimica Acta 530 (2005) 15–21
Fig. 2. Effect of (A) NaOH concentration and (B) separation voltage on the migration time of the analytes. Working potential: +0.65 V (vs. SCE); other conditions as in Fig. 1.
the resolution of the four analytes is poor. At the concentration of 75 mM, the four analytes can be well separated within a relatively short time. When the NaOH concentration was above 75 mM, the peak current was low and the peak shape became poor because the electric current in the capillary also increased resulting in Joule heating and peak broadening. In this experiment, 75 mM NaOH aqueous solution was chosen as the separation medium in considering the peak current, resolution, and analytical time. 3.3. Effect of separation voltage and injection time Fig. 2B illustrates the influence of separation voltage on the migration time of the analytes. As expected, increasing the separation voltage from 6 to 18 kV dramatically decreases migration time for all four multiple-hydroxyl analytes, from 18.4 to 3.8 min (mannitol), from 21.2 to 4.2 min (sucrose), from 26.2 to 5.3 min (glucose), and from 28.8 to 5.6 min (fructose). The time for analysis will be reduced markedly upon raising separation voltage. It is found that higher separation voltages are not beneficial to the resolution of the four compounds. The peak of analytes will overlap when the separation voltage is above 15 kV. Moreover, higher separation voltage increases the baseline noise, resulting in poorer detection limits. However, too lower separation voltages will increase the analysis time considerably, which in turn cause peak broadening. Based on experiments, 12 kV was chosen as the optimum voltage to accomplish a good compromise. In this study, samples were introduced into the capillary electrokinetically. Injection time directly affect the amount of sampling, which determines the peak height and peak shape. The effect of the injection time on CE separation was investigated by changing the sampling time from 2 to 10 s in increments of 2 s at an injection voltage of 12 kV. It was found that both the peak current and the peak width increase with increasing sampling time. When injection time exceeds 6 s, the peak current levels off and peak broadening becomes more severe. In this experiment, 6 s (at 12 kV) is selected as the optimum injection time in considering the resolution and sensitivity.
Through the experiments above, the optimum conditions for determining mannitol, sucrose, glucose, and fructose were acquired. The typical electropherogram for a mixture containing 0.5 mM mannitol, sucrose, glucose, and fructose is shown in Fig. 3. Baseline separation for all four analytes can be achieved within 13 min. 3.4. Reproducibility, linearity and detection limits The precision was examined from a series of seven repetitive injections of a sample mixture containing 0.5 mM mannitol, sucrose, glucose, and fructose under the optimum conditions. The time for each run is about 30 min. Reproducible signals were obtained with R.S.D. of 3.2% (mannitol), 2.7% (sucrose), 2.9% (glucose), and 3.2% (fructose) for the peak currents. Such good precision reflects the reduced surface
Fig. 3. Electropherogram for a mixture containing 0.5 mM of mannitol, sucrose, glucose, and fructose. Working potential: +0.65 V (vs. SCE); other conditions as in Fig. 1.
G. Chen et al. / Analytica Chimica Acta 530 (2005) 15–21
19
Table 1 The results of regression analysis on calibration curves and the detection limitsa Compound
Regression equation Y = a + bXb
Correlation coefficient
Linear range (mM)
Detection limit (M)c
Mannitol Sucrose Glucose Fructose
Y = −0.1731 + 127.58X Y = 0.1440 + 85.856X Y = −0.2362 + 114.14X Y = −0.0443 + 108.04X
0.9994 0.9987 0.9997 0.9991
2.5–0.005 2.5–0.005 2.5–0.005 2.5–0.005
1 2 2 2
a b c
Working potential is +0.65 V (vs. SCE). Other conditions are the same as in Fig. 1. Where the Y and X are the peak current (nA) and concentration of the analytes (mM), respectively. The detection limits correspond to concentrations giving a signal-to-noise ratio of 3.
fouling of the copper electrode. These results are a little better in comparison with the work on a cuprous oxide modified sol–gel carbon paste composite electrode coupled with CE (the R.S.D.s are in the range of 3.9–5.0%) [9]. The reproducibility of the method was determined based on seven runs during two consecutive days. The R.S.D.s for the peak current (4.8, 4.1, 3.9, and 5.0 for mannitol, sucrose, glucose, and fructose, respectively) show that this method is both reproducible and rugged. A series of the standard mixture solutions of the four analytes with concentration ranging from 5.0 M to 2.5 mM were tested to determine the linearity at the copper disc electrode in this method. The copper electrode detector offers a well-defined concentration dependence. The results of regression analysis on calibration curves and detection limits are presented in Table 1. The determination limits are eval-
uated on the basis of a single-to-noise ratio of 3. The calibration curves exhibit satisfactory linear behavior over the concentration range of about three orders of magnitude with the detection limits ranging from 1 to 2 M for all analytes. 3.5. Sample analysis and recovery Under the optimum conditions, CE-ED was applied for the determination of mannitol, sucrose, glucose, and fructose in the diluted extracts from the fruits and leaves of L. lucidum Ait. in 75 mM NaOH. Typical electropherograms for the plant samples are shown in Figs. 4 and 5. Peak identification was performed by standard addition method. Assay results were summarized in Table 2. By carefully comparing the electropherograms, it was found that all samples had identical profiles on the basis of relative peak heights and mi-
Fig. 4. Typical electropherograms for the diluted extracts from the fruits of L. lucidum Ait. Working potential: +0.65 V (vs. SCE); other conditions as in Fig. 1. A: fruit 1 (031110); B: fruit 2 (031130); C: fruit 3 (031220). Peak identification: (1) mannitol, (2) sucrose, (3) glucose, (4) fructose.
20
G. Chen et al. / Analytica Chimica Acta 530 (2005) 15–21
Fig. 5. Typical electropherograms for the diluted extracts from the leaves of L. lucidum Ait. Working potential: +0.65 V (vs. SCE); other conditions as in Fig. 1. A: leaves 1 (031110); B: leaves 2 (031130); C: leaves 3 (031220). Peak identification: (1) mannitol, (2) sucrose, (3) glucose, (4) fructose. Table 2 Assay results of the analytes in the plant samples (n = 3, mg/g)a Sample Fruit 1 Fruit 2 Fruit 3 Leaves 1 Leaves 2 Leaves 3
Mannitol (3.5)b
22.19 31.13 (2.4) 48.74 (1.8) 33.22 (1.3) 21.98 (3.4) 31.21 (2.8)
Sucrose
Glucose
Fructose
1.17 (5.6) 1.42 (4.5) 1.67 (4.8) 2.47 (3.2) 1.97 (4.7) 1.91 (4.2)
16.77 (4.0) 27.99 (1.7) 34.09 (1.8) 15.68 (3.1) 15.47 (2.0) 11.72 (2.2)
1.73 (3.8) 4.34 (1.5) 4.60 (1.4) 1.30 (2.6) 0.63 (5.3) 0.34 (5.6)
a Working potential is +0.65 V (vs. SCE). Other conditions are the same as in Fig. 1. b The data in the brackets are the R.S.D.s (%).
gration times. Mannitol, sucrose, glucose, and fructose were found presented in both fruits and leaves, so all the four constituents can be defined as common peaks in the fingerprint of L. lucidum Ait. under the selected conditions. However, the content of individual analytes varied greatly in different samples as shown in Figs. 4 and 5, indicating that growth stages and the parts of plant had great impact on the contents of the constituents investigated. Fig. 4 shows that there were increases in the contents of mannitol and sugars when the
fruits of L. lucidum Ait. became ripe. The mannitol content in the fruits is similar to the previous report [3]. Moreover, this method is an ideal approach for determining mannitol, an important active constituent, in the investigated plant samples because its peak is the highest in the electropherograms. Recovery experiments were performed by adding accurate amounts of mannitol, sucrose, glucose, and fructose to the diluted extract of the plant sample in the running buffer. Subsequently, the standard-spiked sample solution was analyzed under the optimum conditions. The analytical data are summarized in Table 3. The results demonstrated that this method had both a high accuracy and a good precision for the analytes tested. Because the pKa value of mannitol is 13.50 (18 ◦ C) [31], approximately 19% of mannitol is ionized in 75 mM NaOH aqueous solution (pH 12.88). Glycerol, a nearly uncharged alcohol (pKa value ca. 14.15 [32]), did not interfere with the determination of mannitol because mannitol would elute after the neutral compounds. It is well in agreement with a previous report by Baldwin et al. [12]. Under the optimum conditions, glycerol, mannitol, sucrose, and fructose could be separated on baseline (data not shown).
Table 3 Determination results of recovery for this method (n = 3)a Compound
Original amount (mM)
Added amount (mM)
Found amount (mM)
Recovery (%)
R.S.D. (%)
Mannitol Sucrose Glucose Fructose
0.3046 0.0086 0.2327 0.0241
0.25 0.25 0.25 0.25
0.56 0.25 0.47 0.27
102 97 95 98
2.5 2.3 4.1 2.9
a
Working potential is +0.65 V (vs. SCE). Other conditions are the same as in Fig. 1.
G. Chen et al. / Analytica Chimica Acta 530 (2005) 15–21
The separation between mannitol and neutral compound can be improved by increasing the NaOH concentration in the separation medium.
4. Conclusions For the first time, CE-ED was employed for the determination of mannitol, sucrose, glucose, and fructose in L. lucidum Ait. It is characterized by its higher resolution and sensitivity, lower expense of operation, and less amount of sample. The main advantage of CE as an analytical technique for the analysis of plant samples is that capillary is much easier to wash. Because all the four analytes are directly detected on copper electrode, samples do not need derivatization before determination. It is concluded that CE-ED is an efficient approach for the constituent and fingerprint study of plant drugs due to its special attributes.
Acknowledgments The authors are grateful for the financial supports provided by the National Nature Science Foundation of China (Grant No.: 20075008) and the High-tech Research and Development (863) Programme of China (Grant No.: 2004AA639740).
References [1] Committee of National Pharmacopoeia, Pharmacopoeia of, PR China, vol. 1, Press of Chemical Industry, Beijing, 2000, 34 pp. [2] Committee of Chinese Materia Medica, Chinese Materia Medica, vol. 6, Shanghai Science and Technology Press, Shanghai, 1999, pp. 183–189. [3] M.L. Li, M.L. Liu, J. Chin. Med. Mater. 18 (1995) 246.
21
[4] G.R. Han, C.L. Mu, X.F. Cheng, Acta Pharm. Sin. 30 (1995) 266. [5] H. Liu, Y. Shi, D. Wang, G. Yang, A. Yu, H. Zhang, J. Pharm. Biomed. Anal. 32 (2003) 479. [6] Y. Rong, X.C. Ye, Chin. J. Chin. Mater. Med. 14 (1989) 681. [7] L. Shi, Y. Ma, Z. Cai, Biomed. Chromatogr. 12 (1998) 27. [8] L. Shi, Z. Cai, G. Wu, S. Yang, Y. Ma, Chin. J. Chin. Mater. Med. 23 (1998) 77. [9] L. Hua, L.S. Chia, N.K. Goh, S.N. Tan, Electroanalysis 12 (2000) 287. [10] X.J. Huang, W.T. Kok, J. Chromatogr. A 707 (1995) 335. [11] C.G. Fu, L.N. Song, Y.Z. Fang, Anal. Chim. Acta 371 (1998) 81. [12] J.N. Ye, R.P. Baldwin, J. Chromatogr. A 687 (1994) 141. [13] A.Z. Carvalho, J.A.F. da Silva, C.L. do Lago, Electrophoresis 24 (2003) 2138. [14] Z. El Rassi, Electrophoresis 20 (1999) 3134. [15] Y. Li, H. Liu, X. Ji, J.G. Li, Electrophoresis 21 (2000) 3109. [16] G. Chen, H.W. Zhang, J.N. Ye, Anal. Chim. Acta 423 (2000) 69. [17] L.A. Holland, A.M. Leigh, Electrophoresis 23 (2002) 3649. [18] R.P. Baldwin, Electrophoresis 21 (2000) 4017. [19] G. Chen, J.X. Zhang, J.N. Ye, J. Chromatogr. A 923 (2001) 255. [20] D.C. Johnson, W.R. LaCourse, Electroanalysis 4 (1992) 367. [21] T.J. O’Shea, S.M. Lunte, W.R. LaCourse, Anal. Chem. 65 (1993) 948. [22] P. Luo, S.V. Prabhu, R.P. Baldwin, Anal. Chem. 62 (1990) 752. [23] G. Chen, Q.C. Chu, L.Y. Zhang, J.N. Ye, Anal. Chim. Acta 457 (2002) 225. [24] M.C. Chen, H.J. Huang, Anal. Chim. Acta 341 (1997) 83. [25] T.Y. You, O. Niwa, Z. Chen, K. Hayashi, M. Tomita, S. Hirono, Anal. Chem. 75 (2003) 5191. [26] A.Z. Carvalho, J.A.F. da Silva, C.L. do Lago, Electrophoresis 24 (2003) 2138. [27] Q. Hu, T.S. Zhou, L. Zhang, Y.Z. Fang, Analyst 126 (2001) 298. [28] M.Z. Luo, R.P. Baldwin, J. Electroanal. Chem. 387 (1995) 87. [29] N. Torto, T. Ruzgas, L. Gorton, J. Electroanal. Chem. 464 (1999) 252. [30] Y.Z. Deng, J.L. He, High Performance Capillary Electrophoresis, Academic Press, Beijing, 1995, pp. 26–28. [31] S. Budavari, The Merck Index, an Encyclopedia of Chemicals, Drugs and Biologicals, 12th ed., Merck Research Laboratories, Division of MERCK and Co. Inc., Whitehouse Station, NJ, USA, 1996, 979 pp. [32] J.A. Dean, Lange’s Handbook of Chemistry, 15th ed., McGraw-Hill, New York, 1999, 849 pp.
Determination of mannitol and three sugars in Ligustrum lucidum Ait. by capillary electrophoresis with electrochemical detection Gang Chena,∗ , Luyan Zhanga , Xingliang Wua , Jiannong Yeb a
Department of Chemistry, Fudan University, 211 138 Yixueyuan Rd., Shanghai 200032, PR China b Department of Chemistry, East China Normal University, Shanghai 200062, China Received 28 March 2004; received in revised form 23 August 2004; accepted 23 August 2004 Available online 2 December 2004
Abstract A method based on capillary electrophoresis with electrochemical detection has been developed for the separation and determination of mannitol, sucrose, glucose, and fructose in Ligustrum lucidum Ait. for the first time. Effects of several important factors such as the concentration of NaOH, separation voltage, injection time, and detection potential were investigated to acquire the optimum conditions. The detection electrode was a 300 m diameter copper disc electrode at a working potential of +0.65 V (versus saturated calomel electrode (SCE)). The four analytes can be well separated within 13 min in a 40 cm length fused-silica capillary at a separation voltage of 12 kV in a 75 mM NaOH aqueous solution. The relation between peak current and analyte concentration was linear over about three orders of magnitude with detection limits (S/N = 3) ranging from 1 to 2 M for all analytes. The proposed method has been successfully applied to monitor the mannitol and sugar contents in the plant samples at different growth stages with satisfactory assay results. © 2004 Elsevier B.V. All rights reserved. Keywords: Ligustrum lucidum Ait.; Mannitol; Sucrose; Glucose; Fructose; Capillary electrophoresis; Electrochemical detection
1. Introduction In the present, more and more people are becoming interested in plant medicines, because of their low toxicity and good therapeutical performance. Chinese traditional medicine, Fructus Ligustri Lucidi, is the dried ripen fruit of Ligustrum lucidum Ait. As a commonly used herbal medicine, it has the therapeutical functions of nourishing liver and kidney, stimulating heart, enhancing organism immunization, diuresis, laxation, anti-inflammatory, etc. [1,2]. The leaf of L. lucidum Ait. has also been used as a Chinese traditional herbal drug with functions of anti-bacteria, anti-inflammation, and suppressing cough [2]. A variety of physiologically active compounds (such as ligustroside, nuzhenide, oleanolic acid, ursolic acid, betulin, mannitol, glucose, quercetin, etc.) have been found presented in L. lu∗
Corresponding author. Tel.: +86 21 6466 1130; fax: +86 21 6564 1740. E-mail address: [email protected] (G. Chen).
0003-2670/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.aca.2004.08.053
cidum Ait. Different functions of the herbal drugs come from different active constituents. [2]. It has been demonstrated that the diuresis and laxation activities of Fructus Ligustri Lucidi can be attributed to mannitol, a commonly used diuretic or laxative [2,3]. The mannitol content is an important parameter for evaluating the quality of Fructus Ligustri Lucidi [3]. As the primary metabolites, sucrose, glucose, and fructose are found widely presented in plants. Han et al. have found that higher contents of sugars can indicate the better quality of some herbal drugs [4]. Hence, it is interesting to establish some simple, economical and accurate methods for the determination of mannitol, sucrose, glucose, and fructose in L. lucidum Ait. Li and Liu have determined mannitol in Fructus Ligustri Lucidi by spectrophotometry for quality control [3]. Recently, Liu et al. have reported an approach for the simultaneous determination of oleanolic acid and ursolic acid in this plant medicine based on micellar electrokinetic capillary chromatography [5]. In addition, thin layer chromatogra-
16
G. Chen et al. / Analytica Chimica Acta 530 (2005) 15–21
phy (TLC) [6] and high performance liquid chromatography (HPLC) [7,8] have been employed for the determination of oleanolic acid, salidroside, specnuezhenide and other watersoluble active constituents in the fruits of L. lucidum Ait. Capillary electrophoresis (CE) has been widely used in the analysis of sugars [9–13]. The advances in separation and detection approaches of derivatized and underivatized carbohydrates are reviewed by El Rassi [14]. Nowadays, the application of CE for the separation of various active constituents in medicinal plants has become increasingly widespread because of its minimal sample volume requirement, short analysis time and high separation efficiency [15,16]. Electrochemical detection (ED) typically operated in the amperometric mode can be coupled with CE to provide high sensitivity and selectivity for the determination of electroactive substances [17–19]. Direct electrooxidation at platinum or gold electrodes has gained some popularity for the detection of underivatized carbohydrates [20,21]. To reduce the fouling of platinum and gold electrodes, the applied potential for these electrodes must be continuously pulsed based on a complicated procedure. Alternatively, the work on copper electrode by Baldwin and co-workers is of particular interest for the direct detection of carbohydrates in strongly alkaline medium where signal responses are achieved at constant potential without any electrode fouling [12,22]. This detection approach is less complicated, less expensive, but still sensitive. It can be employed to detect not only carbohydrates themselves but also their derivatives such as alditols and sugar acids [12]. In 2000, U.S. Food and Drug Administration (FDA) published a draft of Guidance for Industry Botanical Drug Products. Before a plant drug becomes logically marketed, its spectroscopic or chromatographic fingerprints and chemical assay of characteristic markers are required. CE should find more applications in this area. In this study, CE-ED was employed for the determination of mannitol, sucrose, glucose, and fructose in L. lucidum Ait. without derivatization. It has been found that the mannitol and sugar contents changed at different growth stages. This method is simple, sensitive, reliable, and efficient providing not only a way for evaluating the quality of plant medicines made from L. lucidum Ait. in marketplaces, but also an excellent method for quality control in medicinal factories and constituent investigation of other plants. To our best knowledge, there are no reports published on the determination of mannitol and sugars in L. lucidum Ait. by CE. The optimization, detailed characterization, and advantages of CE-ED approach are reported in the following sections in connection to the measurement of mannitol and three sugars.
received without further purification. The fruits and leaves of L. lucidum Ait. (Glossy Privet) were collected from the east campus of Medical Center of Fudan University (Shanghai, China) and were kindly identified by Professor D. Chen (Department of Pharmacognosy, Medical Center of Fudan University, Shanghai, China). All aqueous solutions were made up in doubly distilled water. Other chemicals were of analytical grade. Stock solutions of mannitol, sucrose, glucose, and fructose (50 mM) were prepared in doubly distilled water and were kept in a 4 ◦ C refrigerator. They were stable for at least 1 month. The electrophoretic separation medium was 75 mM NaOH aqueous solution unless mentioned otherwise. The stock solutions were diluted to desired concentration with the separation medium just prior to use. 2.2. Apparatus The CE-ED system used has been described previously [23]. A ±30 kV high-voltage dc power supply (Shanghai Institute of Nuclear Research, China) provided a separation voltage between the ends of the capillary. The inlet of the capillary was held at a positive potential and the outlet of capillary was maintained at ground. The separations were proceeded in a 40 cm length of 25 m i.d. and 360 m o.d. fused-silica capillary (Polymicro Technologies, Phoenix, AZ, USA). A three-electrode electrochemical cell consisting of a laboratory-made 300 m diameter copper disc working electrode, a platinum auxiliary electrode and a saturated calomel electrode (SCE) as the reference electrode, was used in combination with a BAS LC-4C amperometric detector (Bioanalytical Systems Inc., West Lafayette, IN, USA). The filter of the detector was set at 0.1 Hz. The working electrode was positioned carefully opposite the outlet of the capillary with the aid of a micromanipulator (CORRECT, Tokyo, Japan) and arranged in a wall-jet configuration. The distance between the tip of the working electrode and the capillary outlet was adjusted to ∼25 m by comparison with the bore (25 m) in the capillary while being viewed under a microscope. The electropherograms were recorded using a LKB·REC 1 chart record (Pharmacia, Sweden). A YS 38-1000 220 V alternate constant-voltage power supply (Shanghai Instrumental Transformer Factory, Shanghai, China) was employed to suppress the voltage fluctuation of the power line. The whole system was assembled in a 10 m2 Faraday room that was airconditioned at 20 ◦ C to minimize the effects of external noise sources. 2.3. Sample preparation
2. Experimental 2.1. Reagent and solutions Mannitol, sucrose, glucose, and fructose were all purchased from Sigma (St. Louis, MO, USA) and were used as
Three batches of fruits and leaves of L. lucidum Ait. were collected from the same tree on different days and dried in the air. The batch numbers (031110, 031130, and 031220) were given based on the collection dates. All plant samples were dried at 60 ◦ C for 3 h and then were pulverized. An accurate
G. Chen et al. / Analytica Chimica Acta 530 (2005) 15–21
17
weight amount of the powders (2.5 g) was dispersed in 100 ml doubly distilled water and was kept in a 90 ◦ C water bath for 3 h. The mixture was subsequently sonicated for 30 min and filtered through a filter-paper. The extract was diluted using 75 mM NaOH aqueous solution at a ratio of 10 (1–10) just prior to CE analysis. 2.4. Procedures Before use, the copper disc electrode was successively polished with emery paper and alumina powder, sonicated in doubly distilled water. CE was performed at a separation voltage of 12 kV unless otherwise indicated. The potential applied to the working electrode was +0.65 V (versus SCE). Samples were injected electrokinetically at 12 kV for 6 s. Moreover, sample solutions, standard solutions, and the separation medium were all filtered through a syringe cellulose acetate filter (0.22 m) prior to their use. Peak identification was performed by standard addition method.
3. Results and discussion 3.1. Hydrodynamic voltammograms (HDVs) Alditols and carbohydrates are most normally electroactive at carbon electrodes, the most commonly used working electrode in ED. So, a variety of metal electrodes made of copper, nickel, gold, and platinum have been employed for the electrochemical detection of alditols, carbohydrates, sugar acids, etc. [12,24–27]. Among them, copper and nickel are most widely utilized. Carbohydrates and alditols can be detected by copper electrodes at constant applied potential in strongly alkaline media based on the electrocatalytic oxidation [12,22]. It has been revealed that the minimum structural requirement for facile oxidation is the presence of at least two nearby hydroxyl groups in the compound to be oxidized. All these oxidations were found to consist of many-electron processes with C C bond cleavage involved. Coulometric studies were carried out on glucose as well as its acid derivatives, which suggested the oxidation of glucose on a copper electrode to be a 12-electron process with the sole product appeared to be the formate ion [28]. It has been demonstrated that the oxidation of disaccharides is different from that of monosaccharides on the copper electrode in alkaline medium. The numbers of electrons transferred during the oxidation are 12 for monosaccharides and 3–13 for disaccharides [29]. In the present work, copper disc electrode was employed for the electrochemical detection of mannitol, sucrose, glucose, and fructose at constant applied potentials in 75 mM NaOH aqueous solution. The potential applied to the working electrode directly affects the sensitivity and detection limits of this method and it is necessary to determine the hydrodynamic voltammograms (HDVs) for the analytes to obtain the optimum potential. Fig. 1 depicts the HDVs for the catalytic oxidation
Fig. 1. Hydrodynamic voltammograms (HDVs, A) for 0.5 mM of mannitol, sucrose, glucose, and fructose in CE. Fused-silica capillary: 25 m i.d. × 40 cm length; working electrode: 300 m diameter copper disc electrode; running buffer: 75 mM NaOH; separation and injection voltage: 12 kV; injection time: 6 s.
of mannitol, sucrose, glucose, and fructose using the copper disc electrode detector. The curves were recorded pointwise from +0.4 to +0.7 V (versus SCE) in steps of 50 mV using a separation voltage of 12 kV. All the analytes display similar profiles, with an increase of the response starting at +0.45 V (versus SCE). The current rises rapidly upon raising the potential above +0.50 V (versus SCE). Although an applied potential greater than +0.65 V (versus SCE) results in higher peak currents, both the baseline noise and the background current increase substantially. The high background current leads to an unstable baseline, which is a disadvantage for the sensitive and stable detection. Considering the sensitivity and background current, subsequent amperometric detection work employed a detection potential of +0.65 V (versus SCE) that offered the most favorable signal-to-noise characteristics. The stability of the working electrode was good and the reproducibility was high at the optimum potential. 3.2. Effects of the NaOH concentration Alditols and sugars are both neutral compounds within the physiological pH range. In the present work, strongly alkaline separation medium (12.5–100 mM NaOH) was used to keep all the analytes in anionic form and the activity of the copper electrode for the electrocatalytic detection of mannitol, sucrose, glucose, and fructose. The separation of these analytes by CE is based on their degrees of dissociation [12]. The NaOH concentration in the separation medium affect the zeta-potential (ζ) of the inner wall of the capillary, the viscosity coefficient of the solution, the electroosmotic flow (EOF) as well as the overall charge of the analytes, which determine the migration time, peak height, and the separation of the analytes [30]. The effect of NaOH concentration on the migration time of the analytes is illustrated in Fig. 2A. Upon raising NaOH concentration from 12.5 to 100 mM, the migration time increases rapidly with the resolution improved for all analytes. When NaOH concentration is lower than 25 mM,
18
G. Chen et al. / Analytica Chimica Acta 530 (2005) 15–21
Fig. 2. Effect of (A) NaOH concentration and (B) separation voltage on the migration time of the analytes. Working potential: +0.65 V (vs. SCE); other conditions as in Fig. 1.
the resolution of the four analytes is poor. At the concentration of 75 mM, the four analytes can be well separated within a relatively short time. When the NaOH concentration was above 75 mM, the peak current was low and the peak shape became poor because the electric current in the capillary also increased resulting in Joule heating and peak broadening. In this experiment, 75 mM NaOH aqueous solution was chosen as the separation medium in considering the peak current, resolution, and analytical time. 3.3. Effect of separation voltage and injection time Fig. 2B illustrates the influence of separation voltage on the migration time of the analytes. As expected, increasing the separation voltage from 6 to 18 kV dramatically decreases migration time for all four multiple-hydroxyl analytes, from 18.4 to 3.8 min (mannitol), from 21.2 to 4.2 min (sucrose), from 26.2 to 5.3 min (glucose), and from 28.8 to 5.6 min (fructose). The time for analysis will be reduced markedly upon raising separation voltage. It is found that higher separation voltages are not beneficial to the resolution of the four compounds. The peak of analytes will overlap when the separation voltage is above 15 kV. Moreover, higher separation voltage increases the baseline noise, resulting in poorer detection limits. However, too lower separation voltages will increase the analysis time considerably, which in turn cause peak broadening. Based on experiments, 12 kV was chosen as the optimum voltage to accomplish a good compromise. In this study, samples were introduced into the capillary electrokinetically. Injection time directly affect the amount of sampling, which determines the peak height and peak shape. The effect of the injection time on CE separation was investigated by changing the sampling time from 2 to 10 s in increments of 2 s at an injection voltage of 12 kV. It was found that both the peak current and the peak width increase with increasing sampling time. When injection time exceeds 6 s, the peak current levels off and peak broadening becomes more severe. In this experiment, 6 s (at 12 kV) is selected as the optimum injection time in considering the resolution and sensitivity.
Through the experiments above, the optimum conditions for determining mannitol, sucrose, glucose, and fructose were acquired. The typical electropherogram for a mixture containing 0.5 mM mannitol, sucrose, glucose, and fructose is shown in Fig. 3. Baseline separation for all four analytes can be achieved within 13 min. 3.4. Reproducibility, linearity and detection limits The precision was examined from a series of seven repetitive injections of a sample mixture containing 0.5 mM mannitol, sucrose, glucose, and fructose under the optimum conditions. The time for each run is about 30 min. Reproducible signals were obtained with R.S.D. of 3.2% (mannitol), 2.7% (sucrose), 2.9% (glucose), and 3.2% (fructose) for the peak currents. Such good precision reflects the reduced surface
Fig. 3. Electropherogram for a mixture containing 0.5 mM of mannitol, sucrose, glucose, and fructose. Working potential: +0.65 V (vs. SCE); other conditions as in Fig. 1.
G. Chen et al. / Analytica Chimica Acta 530 (2005) 15–21
19
Table 1 The results of regression analysis on calibration curves and the detection limitsa Compound
Regression equation Y = a + bXb
Correlation coefficient
Linear range (mM)
Detection limit (M)c
Mannitol Sucrose Glucose Fructose
Y = −0.1731 + 127.58X Y = 0.1440 + 85.856X Y = −0.2362 + 114.14X Y = −0.0443 + 108.04X
0.9994 0.9987 0.9997 0.9991
2.5–0.005 2.5–0.005 2.5–0.005 2.5–0.005
1 2 2 2
a b c
Working potential is +0.65 V (vs. SCE). Other conditions are the same as in Fig. 1. Where the Y and X are the peak current (nA) and concentration of the analytes (mM), respectively. The detection limits correspond to concentrations giving a signal-to-noise ratio of 3.
fouling of the copper electrode. These results are a little better in comparison with the work on a cuprous oxide modified sol–gel carbon paste composite electrode coupled with CE (the R.S.D.s are in the range of 3.9–5.0%) [9]. The reproducibility of the method was determined based on seven runs during two consecutive days. The R.S.D.s for the peak current (4.8, 4.1, 3.9, and 5.0 for mannitol, sucrose, glucose, and fructose, respectively) show that this method is both reproducible and rugged. A series of the standard mixture solutions of the four analytes with concentration ranging from 5.0 M to 2.5 mM were tested to determine the linearity at the copper disc electrode in this method. The copper electrode detector offers a well-defined concentration dependence. The results of regression analysis on calibration curves and detection limits are presented in Table 1. The determination limits are eval-
uated on the basis of a single-to-noise ratio of 3. The calibration curves exhibit satisfactory linear behavior over the concentration range of about three orders of magnitude with the detection limits ranging from 1 to 2 M for all analytes. 3.5. Sample analysis and recovery Under the optimum conditions, CE-ED was applied for the determination of mannitol, sucrose, glucose, and fructose in the diluted extracts from the fruits and leaves of L. lucidum Ait. in 75 mM NaOH. Typical electropherograms for the plant samples are shown in Figs. 4 and 5. Peak identification was performed by standard addition method. Assay results were summarized in Table 2. By carefully comparing the electropherograms, it was found that all samples had identical profiles on the basis of relative peak heights and mi-
Fig. 4. Typical electropherograms for the diluted extracts from the fruits of L. lucidum Ait. Working potential: +0.65 V (vs. SCE); other conditions as in Fig. 1. A: fruit 1 (031110); B: fruit 2 (031130); C: fruit 3 (031220). Peak identification: (1) mannitol, (2) sucrose, (3) glucose, (4) fructose.
20
G. Chen et al. / Analytica Chimica Acta 530 (2005) 15–21
Fig. 5. Typical electropherograms for the diluted extracts from the leaves of L. lucidum Ait. Working potential: +0.65 V (vs. SCE); other conditions as in Fig. 1. A: leaves 1 (031110); B: leaves 2 (031130); C: leaves 3 (031220). Peak identification: (1) mannitol, (2) sucrose, (3) glucose, (4) fructose. Table 2 Assay results of the analytes in the plant samples (n = 3, mg/g)a Sample Fruit 1 Fruit 2 Fruit 3 Leaves 1 Leaves 2 Leaves 3
Mannitol (3.5)b
22.19 31.13 (2.4) 48.74 (1.8) 33.22 (1.3) 21.98 (3.4) 31.21 (2.8)
Sucrose
Glucose
Fructose
1.17 (5.6) 1.42 (4.5) 1.67 (4.8) 2.47 (3.2) 1.97 (4.7) 1.91 (4.2)
16.77 (4.0) 27.99 (1.7) 34.09 (1.8) 15.68 (3.1) 15.47 (2.0) 11.72 (2.2)
1.73 (3.8) 4.34 (1.5) 4.60 (1.4) 1.30 (2.6) 0.63 (5.3) 0.34 (5.6)
a Working potential is +0.65 V (vs. SCE). Other conditions are the same as in Fig. 1. b The data in the brackets are the R.S.D.s (%).
gration times. Mannitol, sucrose, glucose, and fructose were found presented in both fruits and leaves, so all the four constituents can be defined as common peaks in the fingerprint of L. lucidum Ait. under the selected conditions. However, the content of individual analytes varied greatly in different samples as shown in Figs. 4 and 5, indicating that growth stages and the parts of plant had great impact on the contents of the constituents investigated. Fig. 4 shows that there were increases in the contents of mannitol and sugars when the
fruits of L. lucidum Ait. became ripe. The mannitol content in the fruits is similar to the previous report [3]. Moreover, this method is an ideal approach for determining mannitol, an important active constituent, in the investigated plant samples because its peak is the highest in the electropherograms. Recovery experiments were performed by adding accurate amounts of mannitol, sucrose, glucose, and fructose to the diluted extract of the plant sample in the running buffer. Subsequently, the standard-spiked sample solution was analyzed under the optimum conditions. The analytical data are summarized in Table 3. The results demonstrated that this method had both a high accuracy and a good precision for the analytes tested. Because the pKa value of mannitol is 13.50 (18 ◦ C) [31], approximately 19% of mannitol is ionized in 75 mM NaOH aqueous solution (pH 12.88). Glycerol, a nearly uncharged alcohol (pKa value ca. 14.15 [32]), did not interfere with the determination of mannitol because mannitol would elute after the neutral compounds. It is well in agreement with a previous report by Baldwin et al. [12]. Under the optimum conditions, glycerol, mannitol, sucrose, and fructose could be separated on baseline (data not shown).
Table 3 Determination results of recovery for this method (n = 3)a Compound
Original amount (mM)
Added amount (mM)
Found amount (mM)
Recovery (%)
R.S.D. (%)
Mannitol Sucrose Glucose Fructose
0.3046 0.0086 0.2327 0.0241
0.25 0.25 0.25 0.25
0.56 0.25 0.47 0.27
102 97 95 98
2.5 2.3 4.1 2.9
a
Working potential is +0.65 V (vs. SCE). Other conditions are the same as in Fig. 1.
G. Chen et al. / Analytica Chimica Acta 530 (2005) 15–21
The separation between mannitol and neutral compound can be improved by increasing the NaOH concentration in the separation medium.
4. Conclusions For the first time, CE-ED was employed for the determination of mannitol, sucrose, glucose, and fructose in L. lucidum Ait. It is characterized by its higher resolution and sensitivity, lower expense of operation, and less amount of sample. The main advantage of CE as an analytical technique for the analysis of plant samples is that capillary is much easier to wash. Because all the four analytes are directly detected on copper electrode, samples do not need derivatization before determination. It is concluded that CE-ED is an efficient approach for the constituent and fingerprint study of plant drugs due to its special attributes.
Acknowledgments The authors are grateful for the financial supports provided by the National Nature Science Foundation of China (Grant No.: 20075008) and the High-tech Research and Development (863) Programme of China (Grant No.: 2004AA639740).
References [1] Committee of National Pharmacopoeia, Pharmacopoeia of, PR China, vol. 1, Press of Chemical Industry, Beijing, 2000, 34 pp. [2] Committee of Chinese Materia Medica, Chinese Materia Medica, vol. 6, Shanghai Science and Technology Press, Shanghai, 1999, pp. 183–189. [3] M.L. Li, M.L. Liu, J. Chin. Med. Mater. 18 (1995) 246.
21
[4] G.R. Han, C.L. Mu, X.F. Cheng, Acta Pharm. Sin. 30 (1995) 266. [5] H. Liu, Y. Shi, D. Wang, G. Yang, A. Yu, H. Zhang, J. Pharm. Biomed. Anal. 32 (2003) 479. [6] Y. Rong, X.C. Ye, Chin. J. Chin. Mater. Med. 14 (1989) 681. [7] L. Shi, Y. Ma, Z. Cai, Biomed. Chromatogr. 12 (1998) 27. [8] L. Shi, Z. Cai, G. Wu, S. Yang, Y. Ma, Chin. J. Chin. Mater. Med. 23 (1998) 77. [9] L. Hua, L.S. Chia, N.K. Goh, S.N. Tan, Electroanalysis 12 (2000) 287. [10] X.J. Huang, W.T. Kok, J. Chromatogr. A 707 (1995) 335. [11] C.G. Fu, L.N. Song, Y.Z. Fang, Anal. Chim. Acta 371 (1998) 81. [12] J.N. Ye, R.P. Baldwin, J. Chromatogr. A 687 (1994) 141. [13] A.Z. Carvalho, J.A.F. da Silva, C.L. do Lago, Electrophoresis 24 (2003) 2138. [14] Z. El Rassi, Electrophoresis 20 (1999) 3134. [15] Y. Li, H. Liu, X. Ji, J.G. Li, Electrophoresis 21 (2000) 3109. [16] G. Chen, H.W. Zhang, J.N. Ye, Anal. Chim. Acta 423 (2000) 69. [17] L.A. Holland, A.M. Leigh, Electrophoresis 23 (2002) 3649. [18] R.P. Baldwin, Electrophoresis 21 (2000) 4017. [19] G. Chen, J.X. Zhang, J.N. Ye, J. Chromatogr. A 923 (2001) 255. [20] D.C. Johnson, W.R. LaCourse, Electroanalysis 4 (1992) 367. [21] T.J. O’Shea, S.M. Lunte, W.R. LaCourse, Anal. Chem. 65 (1993) 948. [22] P. Luo, S.V. Prabhu, R.P. Baldwin, Anal. Chem. 62 (1990) 752. [23] G. Chen, Q.C. Chu, L.Y. Zhang, J.N. Ye, Anal. Chim. Acta 457 (2002) 225. [24] M.C. Chen, H.J. Huang, Anal. Chim. Acta 341 (1997) 83. [25] T.Y. You, O. Niwa, Z. Chen, K. Hayashi, M. Tomita, S. Hirono, Anal. Chem. 75 (2003) 5191. [26] A.Z. Carvalho, J.A.F. da Silva, C.L. do Lago, Electrophoresis 24 (2003) 2138. [27] Q. Hu, T.S. Zhou, L. Zhang, Y.Z. Fang, Analyst 126 (2001) 298. [28] M.Z. Luo, R.P. Baldwin, J. Electroanal. Chem. 387 (1995) 87. [29] N. Torto, T. Ruzgas, L. Gorton, J. Electroanal. Chem. 464 (1999) 252. [30] Y.Z. Deng, J.L. He, High Performance Capillary Electrophoresis, Academic Press, Beijing, 1995, pp. 26–28. [31] S. Budavari, The Merck Index, an Encyclopedia of Chemicals, Drugs and Biologicals, 12th ed., Merck Research Laboratories, Division of MERCK and Co. Inc., Whitehouse Station, NJ, USA, 1996, 979 pp. [32] J.A. Dean, Lange’s Handbook of Chemistry, 15th ed., McGraw-Hill, New York, 1999, 849 pp.