Development of a Novel Electrochemical Sensor for Determination of Matrine in Sophora flavescens

A simple and sensitive electrochemical sensor fabricated with graphene nanosheets (GNs) and a hydroxyapatite (HA) nanocomposite–modified glassy carbon electrode (GCE) was developed for the determination of matrine (MT). The as-prepared electrode (GNs/HA/GCE) was verified to outperform bare a GCE and GNs-modified electrode with increased oxidation peak currents and the decreased over-potential in the redox process of MT, indicating the great enhancement of electrocatalytic activity toward the oxidation of MT by the composite of GNs and HA. Under the optimized conditions, the oxidation peak currents were related linearly with the concentration of MT, ranging from 2 μM to 3 mM, and the detection limit (S/N = 3) was 1.2 μM. In addition, the proposed electrochemical sensor can be successfully applied in the quantitative determination of MT in Sophora flavescens extract.


Introduction
Traditional Chinese medicine (TCM), with a long history in China, has gained popularity throughout the world with successful application in clinical treatments for many diseases due to its extensive functions [1][2][3][4]. Matrine (MT), as a main component of Sophora flavescens Root, is a type of quinolizidine alkaloid that possesses extensive anti-inflammatory, antiparasitic, anti-arrhythmia, antibacterial, antitumor and antianaphylactic activities [5][6][7]. The herbs containing MT can be taken orally or used extensively in the synthesis of traditional Chinese medicine as well, and they are effective in the clinical treatment of various diseases such as hepatitis, acute dysentery, acute pharyngolaryngitis, eczema, gastrointestinal hemorrhage and colitis. Considering the significant pharmacological activities and extensive clinical applications of MT, developing a simple, rapid and highly sensitive method for the determination of MT of different contents in herbs is highly demanded for its medicinal preparations.
Several methods including capillary electrophoresis (CE) [8][9][10][11], gas chromatography-mass spectrometry (GC-MS) [12], high-performance liquid chromatography (HPLC) [13][14][15][16], and high-performance liquid chromatography-mass spectrometry (HPLC-MS) [17][18][19][20] have been developed for the detection of matrine and oxymatrine. The GC technique with gas as the mobile phase is suitable for the detection of organics with a low boiling point. As for the HPLC technique, the consumption of bulk solvent and the lack of specificity and selectivity due to the employment of UV detection limits its large-scale application despite its various advantages, such as excellent analytical precision and high sample loading capacity. As far as we know, few reports on the detection of MT by electrochemical techniques have been studied [21][22][23]. For example, Miao [22] investigated the  The voltammetric response of the proposed GNs/HA/GCE was compared with that of other electrodes in order to investigate the effect of the GNs/HA composite. Figure 3A shows CVs of 0.1 mM MT in 0.2 M PBS solution with a pH of 7.5 obtained on bare GCE, GNs/GCE and GNs/HA/GCE, respectively. It was found that MT exhibited electrochemical activity on all electrodes. As shown from the CV obtained on bare GCE, a very small bulge was observed at 1.09 V, indicating the weak electrochemical response of MT. As to the electrochemical response of MT on GNs/GCE, the oxidation peak current (ipa) exhibited a significant increase and the peak potential (Epa) exhibited a negative shift to 0.78 V as well, suggesting that the electrochemical signal was amplified effectively owing to the acceleration of the electron transfer on the electrode surface by GNs which possess a large surface area and remarkable electric conductivity. In contrast, a distinct well-defined and more sensitive anodic peak was observed on GNs/HA/GCE among the three studied electrodes, indicating that the more efficient interface and microenvironment which are beneficial for the electrochemical   The voltammetric response of the proposed GNs/HA/GCE was compared with that of other electrodes in order to investigate the effect of the GNs/HA composite. Figure 3A shows CVs of 0.1 mM MT in 0.2 M PBS solution with a pH of 7.5 obtained on bare GCE, GNs/GCE and GNs/HA/GCE, respectively. It was found that MT exhibited electrochemical activity on all electrodes. As shown from the CV obtained on bare GCE, a very small bulge was observed at 1.09 V, indicating the weak electrochemical response of MT. As to the electrochemical response of MT on GNs/GCE, the oxidation peak current (ipa) exhibited a significant increase and the peak potential (Epa) exhibited a negative shift to 0.78 V as well, suggesting that the electrochemical signal was amplified effectively owing to the acceleration of the electron transfer on the electrode surface by GNs which possess a large surface area and remarkable electric conductivity. In contrast, a distinct well-defined and more sensitive anodic peak was observed on GNs/HA/GCE among the three studied electrodes, indicating that the more efficient interface and microenvironment which are beneficial for the electrochemical The voltammetric response of the proposed GNs/HA/GCE was compared with that of other electrodes in order to investigate the effect of the GNs/HA composite. Figure 3A shows CVs of 0.1 mM MT in 0.2 M PBS solution with a pH of 7.5 obtained on bare GCE, GNs/GCE and GNs/HA/GCE, respectively. It was found that MT exhibited electrochemical activity on all electrodes. As shown from the CV obtained on bare GCE, a very small bulge was observed at 1.09 V, indicating the weak electrochemical response of MT. As to the electrochemical response of MT on GNs/GCE, the oxidation peak current (i pa ) exhibited a significant increase and the peak potential (E pa ) exhibited a negative shift to 0.78 V as well, suggesting that the electrochemical signal was amplified effectively owing to the acceleration of the electron transfer on the electrode surface by GNs which possess a large surface area and remarkable electric conductivity. In contrast, a distinct well-defined and more sensitive anodic peak was observed on GNs/HA/GCE among the three studied electrodes, indicating that the more efficient interface and microenvironment which are beneficial for the electrochemical response of MT could be obtained with the GNs/HA composite. It can be noticed that the oxidation potential of MT on GNs/HA/GCE is slightly higher than that of the GNs/GCE, probably due to the decrease of the electrocatalytic potential of GNs by the surface coverage of HA.
properties of MT. CVs of the background (curve a) and 0.1 mM MT in 0.2 M PBS with a pH of 7.5 on GNs/HA/GCE are shown in Figure 3B, respectively. The anodic peak current was found to decrease obviously in the second scan compared to the first scan, and then gradually decreased with the successive cyclic sweep. The gradual decrease of the anodic peak current might result from the hindrance of the MT to the electrode surface by the formed oxidation product of MT. Moreover, the surface of GNs/HA/GCE could be restored to the initial state by the following method. Firstly, the spent GNs/HA/GCE electrode was kept in PBS with a pH of 7.5 without MT for 3 min under constant stirring, and then CV was carried out in the scan potential range from 0.0 V to 1.2 V until the disappearance of the MT peaks. The voltammogram obtained on the restored electrode was consistent with that obtained at the first cycle. Therefore, the peak current taken from the first cycle was used for the all discussions below. The voltammetric response of MT was greatly influenced by the used types of supporting electrolytes. Among the studied various buffers, including citrate, acetarte, phosphate, borate and Britton-Robinson buffer, phosphate buffer (PBS) was the best with a sharp response and excellent sensitivity. Therefore, PBS solutions with the concentration of 0.1 mM and varying pH values from 5.0 to 8.5 were used in this study ( Figure S1). Epa shifted to lower values with the increasing pH. The relationship between the Epa and pH could be fitted into the regression equation: Epa (V) = 1.208 − 0.057 pH (R = 0.997). It was found that the same electrons and protons are consumed in the electrode reaction as indicated by the slope of the regression equation. Therefore, the oxidation mechanism for matrine can be written as shown in Figure 4, which is in accordance with the report [23,66]. Besides, the ipa varies very little in the range of 7.0-7.5, so the solution pH of 7.5 was chosen in the following experiments, which is close to the human physiological pH.   Figure 3B, respectively. The anodic peak current was found to decrease obviously in the second scan compared to the first scan, and then gradually decreased with the successive cyclic sweep. The gradual decrease of the anodic peak current might result from the hindrance of the MT to the electrode surface by the formed oxidation product of MT. Moreover, the surface of GNs/HA/GCE could be restored to the initial state by the following method. Firstly, the spent GNs/HA/GCE electrode was kept in PBS with a pH of 7.5 without MT for 3 min under constant stirring, and then CV was carried out in the scan potential range from 0.0 V to 1.2 V until the disappearance of the MT peaks. The voltammogram obtained on the restored electrode was consistent with that obtained at the first cycle. Therefore, the peak current taken from the first cycle was used for the all discussions below.
The voltammetric response of MT was greatly influenced by the used types of supporting electrolytes. Among the studied various buffers, including citrate, acetarte, phosphate, borate and Britton-Robinson buffer, phosphate buffer (PBS) was the best with a sharp response and excellent sensitivity. Therefore, PBS solutions with the concentration of 0.1 mM and varying pH values from 5.0 to 8.5 were used in this study ( Figure S1). E pa shifted to lower values with the increasing pH. The relationship between the E pa and pH could be fitted into the regression equation: E pa (V) = 1.208 − 0.057 pH (R = 0.997). It was found that the same electrons and protons are consumed in the electrode reaction as indicated by the slope of the regression equation. Therefore, the oxidation mechanism for matrine can be written as shown in Figure 4, which is in accordance with the report [23,66]. Besides, the i pa varies very little in the range of 7.0-7.5, so the solution pH of 7.5 was chosen in the following experiments, which is close to the human physiological pH. relationship between the Epa and pH could be fitted into the regression equation: Epa (V) = 1.208 − 0.057 pH (R = 0.997). It was found that the same electrons and protons are consumed in the electrode reaction as indicated by the slope of the regression equation. Therefore, the oxidation mechanism for matrine can be written as shown in Figure 4, which is in accordance with the report [23,66]. Besides, the ipa varies very little in the range of 7.0-7.5, so the solution pH of 7.5 was chosen in the following experiments, which is close to the human physiological pH.  Linear sweep anodic stripping voltammetry (LSASV) was employed in the determination of MT for the sake of decreasing the background current. The scan rate of 50 mV/s was selected as the most suitable scan rate in order to enhance the sensitivity. In addition, a preconcentration could be employed to improve the detection sensitivity as to the adsorption-driven electrode reaction. The influence of the accumulation time (t acc ) on the current response of 0.5 µM MT was investigated and the results are shown in Figure 5. Obviously, i pa increased sharply within 240 s and then flattened out. A longer t acc will lead to a better detection limit but a narrower linear range. Thus, 240 s was selected as the accumulation time by considering these factors comprehensively. Moreover, the accumulation was performed under an open circuit owing to the small effect of the accumulation potential on i pa . most suitable scan rate in order to enhance the sensitivity. In addition, a preconcentration could be employed to improve the detection sensitivity as to the adsorption-driven electrode reaction. The influence of the accumulation time (tacc) on the current response of 0.5 μM MT was investigated and the results are shown in Figure 5. Obviously, ipa increased sharply within 240 s and then flattened out. A longer tacc will lead to a better detection limit but a narrower linear range. Thus, 240 s was selected as the accumulation time by considering these factors comprehensively. Moreover, the accumulation was performed under an open circuit owing to the small effect of the accumulation potential on ipa. MT standard solutions with a series of concentrations were determined under the optimized conditions. The response of MT at different concentrations by LSASV is shown in Figure 6. The inset of Figure 6A displays the established linear relationship between ipa and the concentration of MT ranging from 2 μM to 3 mM. The values of the slope and intercept of the calibration curve were 0.00141 and 0.0998, respectively. The detection limit (S/N = 3) was 1.2 μM, indicating that the proposed GNs/HA/GCE was sensitive for the determination of MT. The stability, reproducibility and repeatability of GNs/HA/GCE were also estimated. The modified electrode remained at 92% of its initial current response to MT after two weeks of storage, indicating the electrochemical sensor has good stability. The reproducibility was examined at five modified electrodes prepared in the same conditions and a relative standard deviation (RSD) of 5.3% was obtained. The RSD of the response to 0.1 mM MT was 5.1% for six successive measurements, indicating the electrochemical sensor has the good repeatability. The influence of various potentially interfering substances for the determination of 0.1 Mm MT was studied. The results indicated that 20-fold of Zn 2+ , Cu 2+ SO4 2− , glucose, sucrose and amylum, ascorbic acid and uric acid had almost no influence on the determination, and the tolerance limit was estimated to be less than 7% of the relative error.
The stability, reproducibility and repeatability of GNs/HA/GCE were also evaluated. The current response of GNs/HA/GCE after 14 days of storage in the determination of MT remained at 93.7% compared with that of the initial current response, suggesting the excellent stability of the proposed electrochemical sensor. The responses of five GNs/HA/GCE electrodes prepared at exactly the same conditions to 0.1 mM MT were investigated and the relative standard deviation (RSD) was  Figure 6. The inset of Figure 6 displays the established linear relationship between i pa and the concentration of MT ranging from 2 µM to 3 mM. The values of the slope and intercept of the calibration curve were 0.00141 and 0.0998, respectively. The detection limit (S/N = 3) was 1.2 µM, indicating that the proposed GNs/HA/GCE was sensitive for the determination of MT. The stability, reproducibility and repeatability of GNs/HA/GCE were also estimated. The modified electrode remained at 92% of its initial current response to MT after two weeks of storage, indicating the electrochemical sensor has good stability. The reproducibility was examined at five modified electrodes prepared in the same conditions and a relative standard deviation (RSD) of 5.3% was obtained. The RSD of the response to 0.1 mM MT was 5.1% for six successive measurements, indicating the electrochemical sensor has the good repeatability. The influence of various potentially interfering substances for the determination of 0.1 Mm MT was studied. The results indicated that 20-fold of Zn 2+ , Cu 2+ SO 4 2− , glucose, sucrose and amylum, ascorbic acid and uric acid had almost no influence on the determination, and the tolerance limit was estimated to be less than 7% of the relative error. The proposed electrochemical sensor was employed for the determination of MT in Sophora flavescens extract for the sake of evaluating its validity. In order to operate in the linear range of the proposed method, the Sophora flavescens extract was diluted with the supporting electrolyte. After the evaluation of MT content, a standard MT solution was added into the sample and then the total MT content was determined to calculate the recovery. As shown in Table 1, the contents obtained by the proposed method were compared with that obtained by the HPLC method using the t-test under 95% confidence levels. No significant difference was found between the two methods, suggesting that the proposed electrochemical sensor is reliable for the quantitative determination of MT in herb samples.

Materials and Methods
Matrine was supplied by Aladdin Chemistry Co., Ltd. (Shanghai, China) and used as received., Ca(NO3)2·4H2O, (NH4)2HPO4, graphite and ammonia were obtained from Sinopharm Chemical Reagent Co., Ltd. All other reagents were analytical reagents and used without further purification. Double distilled water was used in all experiments. The standard stock solution (2 × 10 −3 M) of MT was stored at 4 °C darkly before use. Britton-Robinson (B-R) buffer solution with different pH values was employed as supporting electrolytes. Differential pulse voltammetry (DPV) and cyclic voltammetry (CV) measurements were performed at room temperature on a CHI660D electrochemical workstation (CHI Instrument Company, Shanghai) with three electrode cell containing bare or GNs/HA modified GCE (d = 3 mm), platinum (Pt) wire and Ag/AgCl as working, auxiliary and reference electrode, respectively.
The well-known modified Hummers method was employed for the preparation of graphene from purified natural graphite [67,68]. And the wet chemical method was used for the synthesis of hydroxyapatite. Typically, 150 mL of Ca(NO3)2 solution (0.5 M) was added into 20 mL of (NH4)2HPO4 solution (0.5 M) and the mixture was stirred at 37 °C until the clear solution was obtained. Then ammonia water was added to adjust the pH of mixture to 10. Then 70 mL of (NH4)2HPO4 solution (0.5 M) was added dropwise into the mixture under vigorous agitation. It was worth noting that the pH values of the mixture must be maintained at 10 by titration of diluted The stability, reproducibility and repeatability of GNs/HA/GCE were also evaluated. The current response of GNs/HA/GCE after 14 days of storage in the determination of MT remained at 93.7% compared with that of the initial current response, suggesting the excellent stability of the proposed electrochemical sensor. The responses of five GNs/HA/GCE electrodes prepared at exactly the same conditions to 0.1 mM MT were investigated and the relative standard deviation (RSD) was 4.3%, indicating the remarkable reproducibility of the proposed electrode. As to the same electrochemical sensor in the determination of 0.1 mM MT, the RSD of 10 successive measurements was 4.1%, suggesting the outstanding repeatability of the electrochemical sensor. In general, the excellent stability, reproducibility and repeatability of GNs/HA/GCE for the determination of MT were proven.
The proposed electrochemical sensor was employed for the determination of MT in Sophora flavescens extract for the sake of evaluating its validity. In order to operate in the linear range of the proposed method, the Sophora flavescens extract was diluted with the supporting electrolyte. After the evaluation of MT content, a standard MT solution was added into the sample and then the total MT content was determined to calculate the recovery. As shown in Table 1, the contents obtained by the proposed method were compared with that obtained by the HPLC method using the t-test under 95% confidence levels. No significant difference was found between the two methods, suggesting that the proposed electrochemical sensor is reliable for the quantitative determination of MT in herb samples.

Materials and Methods
Matrine was supplied by Aladdin Chemistry Co., Ltd. (Shanghai, China) and used as received., Ca(NO 3 ) 2 ·4H 2 O, (NH 4 ) 2 HPO 4 , graphite and ammonia were obtained from Sinopharm Chemical Reagent Co., Ltd. All other reagents were analytical reagents and used without further purification. Double distilled water was used in all experiments. The standard stock solution (2 × 10 −3 M) of MT was stored at 4 • C darkly before use. Britton-Robinson (B-R) buffer solution with different pH values was employed as supporting electrolytes. Differential pulse voltammetry (DPV) and cyclic voltammetry (CV) measurements were performed at room temperature on a CHI660D electrochemical workstation (CHI Instrument Company, Shanghai) with three electrode cell containing bare or GNs/HA modified GCE (d = 3 mm), platinum (Pt) wire and Ag/AgCl as working, auxiliary and reference electrode, respectively.
The well-known modified Hummers method was employed for the preparation of graphene from purified natural graphite [67,68]. And the wet chemical method was used for the synthesis of hydroxyapatite. Typically, 150 mL of Ca(NO 3 ) 2 solution (0.5 M) was added into 20 mL of (NH 4 ) 2 HPO 4 solution (0.5 M) and the mixture was stirred at 37 • C until the clear solution was obtained. Then ammonia water was added to adjust the pH of mixture to 10. Then 70 mL of (NH 4 ) 2 HPO 4 solution (0.5 M) was added dropwise into the mixture under vigorous agitation. It was worth noting that the pH values of the mixture must be maintained at 10 by titration of diluted ammonia water during the addition of (NH 4 ) 2 HPO 4 solution. Subsequently, the mixtures were stirred for another 2 h and then aged at 37 • C for 24 h. Finally, the resulting mixture was treated with centrifugation and the obtained solid was washed and dried at 90 • C in an oven.
Then 10 mg of as-synthesized graphene was dissolved in 100 mL of deionized water and the mixture was treated with ultrasonic agitation until the homogeneous black suspension was formed. Before use, the glassy carbon electrode (GCE) with the diameter of 3 mm was firstly polished with an abrasive cloth composed of 0.3 and 0.05 µm α-Al 2 O 3 slurry, then washed with ethanol and distilled water thoroughly under ultrasonication and finally dried with high-purity nitrogen gas. Afterwards, 5 µL of graphene solution with the concentration of 0.1 mg/mL was dropped onto the pre-treated GCE, and then the graphene-modified GCE (GNs/GCE) was obtained after being dried under an infrared heat lamp. Finally, the obtained electrode was washed with distilled water. The immobilization of hydroxyapatite was achieved by dropping 5 µL of hydroxyapatite solution (1.2 mg/mL) on GNs/GCE and the obtained GNs/HA/GCE was then dried at room temperature.

Conclusions
In conclusion, a relatively simple voltammetric sensor fabricated with GNs/HA/GCE was proposed for the determination of MT with high sensitivity and selectivity. In comparison with the previously reported method for the detection of MT, GNs/HA/GCE exhibited significant advantages of a wide linear range and a low detection limit. In addition, the proposed electrochemical sensor could be employed for the routine analysis of MT in herb samples owing to the excellent stability, reproducibility and repeatability of GNs/HA/GCE.

Supplementary Materials:
The supplementary material is available online.