A label-free electrochemical impedance immunosensor for the sensitive detection of aflatoxin B1. 2014(Autosaved).pdf

Analytical Methods View Article Online Published on 28 October 2014. Downloaded by Universidad Nacional Agraria La Molina on 10/06/2016 16:48:59. PAPER Cite this: Anal. Methods, 2015, 7, 2354 View Journal | View Issue A label-free electrochemical impedance immunosensor for the sensitive detection of aflatoxin B1 Liguo Chen,ab Jianhui Jiang,a Guoli Shena and Ruqin Yu*a Because of the potential health impact of aflatoxin B1 (AFB1), it is essential to monitor the level of this mycotoxin in a variety of foods and agricultural products. In this paper, a novel immunosensor for the rapid detection of AFB1 based on label-free electrochemical impedance spectroscopy (EIS) monitoring was achieved. The immunosensor was fabricated by stepwise immobilization of 1,6-hexanedithiol, colloidal Au, and aflatoxin B1–bovine serum albumin conjugate (AFB1–BSA) on a gold electrode via selfassembling technique. The interfacial properties of the modified electrodes were evaluated using the Fe(CN)63 Received 22nd August 2014 Accepted 6th October 2014 /4 redox couple as a probe via cyclic voltammetry (CV) and EIS. An equivalent circuit model with a constant phase element was used to interpret the obtained impedance spectra. The impedance via the specific immuno-interaction at the sensor surface was utilized to detect AFB1 in samples. Under the optimized conditions, the impedance increment was linearly related to the AFB1 concentration in the DOI: 10.1039/c4ay01981d range of 0.08 to 100 ng mL www.rsc.org/methods coefficient of 0.9919. 1 1 Introduction The aatoxins, primarily produced by Aspergillus avus or Aspergillus parasiticus, are widely found in food crops including grains, cereals, peanut products, sorghum, and soy seeds.1 In particular, aatoxin B1 (AFB1) (Fig. 1), a commonly occurring aatoxin, is one of the most potent carcinogens with potential hazards to human and animal health.2 Owing to the risk of aatoxins, the current maximum tolerable levels for aatoxins issued by the European Commission are 2 ng g 1 for AFB1 and 4 ng g 1 for total aatoxins (AFB1, AFB2, AFG1 and AFG2) in corn, groundnuts, nuts, dried fruit and cereals.3,4 For such a low content, developing simple, selective and sensitive analytical methods to detect the trace amount of AFB1 in foodstuffs and feeds is an extremely important issue. During the past two decades, various analytical techniques have attracted increasing attention for the determination of aatoxins such as thin layer chromatography (TLC),5 gas chromatography coupled with mass spectrometry (GC/MS) and highperformance liquid chromatography (HPLC),6–11 capillary electrophoresis (CE),12 and a variety of immunoassay methods.13–16 These methods are highly sensitive and specic, but are oen complicated and time-consuming, need some secondary antibodies, or rely on expensive instruments or skillful operators. with a detection limit of 0.05 ng mL 1 (S/N ¼ 3) and a correlation For example, among the conventional immunoassay methods, the enzyme-linked immunosorbent assay (ELISA)16 is undoubtedly the most frequently applied one. Nevertheless, there are some limitations of ELISA such as depending on some complicated enzyme labeling procedure and specic reagents of synthesized redox-labelled detection probes, consequently preventing its widespread use in common practice. Thus, there is an urgent demand for ultra sensitive methods of immunoassay. Label-free electrochemical immunosensors17–24 can offer advantages over chromatographic procedures or ELISA. They have been reported to be a very attractive approach for detecting affinity interactions by monitoring changes of electronic or interfacial properties generated by the immunocomplex formation on the electrode surface. Sun et al.22 described a simple, highly sensitive, label-free, and a State Key Laboratory of Chem/Biosensing and Chemometrics, Chemistry and Chemical Engineering College, Hunan University, Changsha 410082, China. E-mail: rqyu@hnu.edu.cn; Fax: +86-731-8882 2782 b Zhuzhou Office of Food Safety Committee, Zhuzhou, Hunan 412007, China 2354 | Anal. Methods, 2015, 7, 2354–2359 Fig. 1 Chemical structure of AFB1. This journal is © The Royal Society of Chemistry 2015 View Article Online Published on 28 October 2014. Downloaded by Universidad Nacional Agraria La Molina on 10/06/2016 16:48:59. Paper Schematic illustration of the process of immobilization of immunoantigen onto the gold electrode and the immunoreaction procedure. Scheme 1 Analytical Methods and low-cost characteristics. However, this method has seldom been applied in detection of AFB1 up to now. The current work focused on development of a label-free impedimentary immunosensor to determine AFB1 level. The results revealed that the sensitivity could be substantially improved via the self-assembled colloid. Under the optimized conditions, there was a good linear relationship between impedance increment and the concentration of AFB1 in the range of 0.08 and 100 ng mL 1. The detection limit can reach as low as 0.05 ng mL 1. Our method is a very promising alternative in the determination of AFB1. 2 Experimental section 2.1 Fig. 2 Cyclic voltammograms of 5 mM K4[Fe(CN)6]/K3[Fe(CN)6] (1 : 1) in PBS (0.1 M KCl, pH 7.4) after different steps of modification; (a) bare gold electrode; (b) 1,6 hexanedithiol/gold electrode; (c) Au-colloid/1,6 hexanedithiol/gold electrode; (d) AFB1–BSA (10 mg mL 1) immobilized gold electrode. Scan rate: 100 mV s 1 vs. SCE. Aatoxin B1 (AFB1), aatoxin B2 (AFB2), aatoxin M1 (AFM1), aatoxin M2 (AFM2), AFB1–bovine serum albumin (BSA) conjugate (AFB1–BSA) and the monoclonal anti-aatoxin B1 (from mouse) were provided by Sigma-Aldrich Co. (St. Louis, MO, USA). Bovine serum albumin (BSA) was provided by Beijing Zhongshan Golden Bridge Biotechnology Co., Ltd (Beijing, China). 1,6-Hexanedithiol was obtained from Fluka. HAuCl4$3H2O and trisodium citrate were obtained from Shanghai Chemical Reagents (Shanghai, China). The phosphate buffer solution (PBS) used was a 10 mmol L 1 Na2HPO4– KH2PO4 solution of pH 7.4. The K3[Fe(CN)6]/K4[Fe(CN)6] mixture solution (1 : 1, 5 mmol L 1) used in electrochemical measurements was prepared by using 10 mmol L 1 PBS (containing 0.1 mol L 1 KCl, pH 7.4). All other chemicals were of analytical grade and used as received. Doubly distilled water was used throughout all the experiments. 2.2 Equivalent circuit to fit the impedance spectroscopy in the presence of redox couples. Rs, resistance of the electrolyte. solution; Ret, electron-transfer resistance; Zw, Warburg impedance; Cdl, doublelayer capacitance. Scheme 2 mediatorless approach for fabrication of an impedimetric immunosensor based on immobilization of anti-MCLR on gold electrode. Qiu et al.23 designed a label-free amperometric immunosensor based on chitosan-branched ferrocene and gold nanoparticles for detecting hepatitis B surface antigen. Radi and co-workers24 prepared a novel label-free electrochemical impedimentary immunosensor for sensitive detection of ochratoxin A. It is widely accepted that impedance sensing is a very good choice for biological binding events with fast, simple This journal is © The Royal Society of Chemistry 2015 Chemicals Apparatus The cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) measurements were carried out on a CHI 660a working station (CH Instruments, Shanghai, China) in conjunction with a computer. Throughout all the experiments, a three-electrode system was used, with a fabricated electrode (a gold disk electrode, 4 mm diameter) as a working electrode, a saturated calomel electrode (SCE) as a reference electrode and a platinum electrode as an auxiliary electrode. Impedance measurements were performed in the frequency range from 0.1 Hz to 10 000 Hz. All potentials were measured and reported versus the SCE, and all measurements were carried out in a 25 mL cell at room temperature. 2.3 Synthesis of gold nanoparticles All glassware used for gold nanoparticle synthesis were thoroughly cleaned in freshly prepared aqua Regia (three parts HCl plus one part HNO3), followed by extensive rinsing with doubly distilled water. Gold nanoparticles were prepared according to the literature25 with a slight modication. Briey, 1 mL of 1.0% trisodium citrate was added into 100 mL of 0.01% HAuCl4; the mixed solution was reuxed under stirring, and then the mixture was kept boiling for another 10 min. While the solution color turned to deep wine red, the heating continued for 15 minutes and the heater was Anal. Methods, 2015, 7, 2354–2359 | 2355 View Article Online Published on 28 October 2014. Downloaded by Universidad Nacional Agraria La Molina on 10/06/2016 16:48:59. Analytical Methods Fig. 3 Nyquist diagram of electrochemical impedance spectroscopy for the immunosensor electrode at a potential +0.24 V vs. SCE. (a) Bare Au electrode; (b) Au-colloid/1,6-hexanedithiol/gold electrode; (c) AFB1–BSA immobilized electrode; (d) after being immersed in BSA; (e) the resulting electrode after immunoreaction at 37  C. All the measurements were carried out in 10 mM PBS (containing 0.1 M KCl, 5 mM K4[Fe(CN)6]/K3[Fe(CN)6], pH 7.4) and the data were recorded in the frequency range from 0.1 Hz to 10 kHz. Paper between 1.0 and +1.55 V in 0.1 mol L 1 H2SO4 for 10 min. Aer that, the electrode was washed with doubly distilled water and dried in a nitrogen stream at room temperature.26 The prepared bare electrode was placed in a freshly made 2.5 mmol L 1 1,6hexanedithiol solution in ethanol. Aer adsorption for 4 h, the electrode was thoroughly washed with ethanol and water to remove physically adsorbed 1,6-hexanedithiol and dried at room temperature. Subsequently, the electrode was immersed in the Au nanoparticle solution overnight in a refrigerator. Aer sufficiently rinsing, the electrode was incubated in an AFB1–BSA solution for 4 h at room temperature in order to adsorb enough immuno-antigen on the electrode, and then it was rinsed thoroughly with 10 m mol L 1 PBS to remove the unstable AFB1–BSA on the electrode surface. To block the possible remaining active sites and avoid the non-specic adsorption on the electrode surface, the electrode was dipped in a 5.0% BSA solution for 30 min at room temperature. Finally, the immunosensor obtained was rinsed with water and PBS (pH 7.4) to remove the physically adsorbed molecules and stored in PBS at 4  C when not in use. For the measurement, antigen and antibody with different concentrations were dissolved in 10 mmol L 1 PBS (pH 7.4). Aer 50 min incubation at 37  C, the resulting electrode was washed with doubly distilled water and PBS, separately. The impedance changes of the immunosensors were measured in a solution of 5 mmol L 1 K3[Fe(CN)6]/K4[Fe(CN)6] (1 : 1, containing 0.1 mol L 1 KCl, pH 7.4) mixture solution (Scheme 1). 3 Results and discussion 3.1 Fig. 4 Effect of pH for the immobilization of AFB1–BSA on the impedance change. For optimization, in the experiment, the AFB1 antibody dilution ratio at 1 : 4000 and 0.5 ng mL 1 AFB1 was applied and the incubation time was 40 min (n ¼ 3, error bars represent R.S.D.). removed. The mixture was stirred continuously until it was cooled to room temperature. Colloidal Au solution thus prepared was stored in a brown glass bottle at 4  C for use. 2.4 Fabrication of immunosensor and measurement procedure First, the gold electrode was polished with a 0.05 mm alumina powder and successively soaked in water, ethanol and water for 5 min each under ultrasonic agitation. Then, the gold electrode was dipped in “piranha” solution (H2SO4/H2O2, 7 : 3 by volume) (Warning: Piranha solution reacts violently with almost all organic materials and must be handled with extreme caution!) for 30 min and electrochemically treated by cycling the potential 2356 | Anal. Methods, 2015, 7, 2354–2359 Electrochemical characteristics of the modifying process The electron-transfer behaviors for the AFB1 immunosensor by self-assemble technique in preparation processes were performed by cyclic voltammetry. Fig. 2 shows the typical CV of the fabricated AFB1 immunosensor in the presence of 5 mmol L 1 ferrocyanide/ferricyanide mixture (1 : 1) solution containing 10 mmol L 1 PBS (pH 7.4) and 0.1 mol L 1 KCl. As can be seen, Fig. 2a depicts an intact pair of redox peaks for the CV of the bare Au electrode. The immobilization of 1,6-hexanedithiol on the gold electrode leads to a signicant decrease in peak current of the redox probe (Fig. 2b). The formation of an insulating lm partially blocked the electron-transfer of [Fe(CN)6]3 /4 to the modied electrode. When the Au nanoparticles were modied onto the electrode surface, a slight increase of the current response can be found in Fig. 2c. It indicates that the Au nanoparticles can enhance the efficiency and rate of electron transfer at the electrode surface. As expected, the subsequent chemical binding of the AFB1–BSA induces an obvious reduction of current response in the electrochemical signal (Fig. 2d), exhibiting that the immobilization of the AFB1–BSA builds up an almost insulating layer on the Au electrode and hinders the interfacial electron transfer. 3.2 Electrochemical impedance characteristics of the modifying process Electrochemical impedance spectroscopy is one of the most effective methods to investigate the interface features of This journal is © The Royal Society of Chemistry 2015 View Article Online Published on 28 October 2014. Downloaded by Universidad Nacional Agraria La Molina on 10/06/2016 16:48:59. Paper Effect of the incubation time of the immunoreaction on the resulting electrode. Incubation time from 20 min to 80 min. In this experiment, the AFB1 antibody dilution ratio at 1 : 4000 and 0.5 ng mL 1 AFB1 was utilized (n ¼ 3, error bars represent R.S.D.). Fig. 5 Analytical Methods Fig. 3 shows the electrochemical impedance spectroscopy for each step to reect the procedure of the immunoassay. As shown in Fig. 3, the faradic impedance spectroscopy on a bare gold electrode reveals a straight line (curve 3a), implying that the electron-transfer process is not a limiting step of the electrochemical process. In the Au-colloid/1,6-hexanedithiol/gold electrode, curve 3b, the semicircle diameter increases obviously indicating a lower electron-transfer resistance at the electrode interface. With the immunocomplex formation on the electrode (curves 3c and 3d), the semicircle diameter increases apparently compare with curve 3b, illustrating that the AFB1–BSA and BSA lm was successfully immobilized on the electrode surface. Aer the immunoreactions, here was a remarkable increase in the semicircle diameter of the impedance spectrum (curve 3e), implying that the interfacial electron-transfer rate on electrode increases remarkably. The increase of the diameter of the semicircle conrms that the immunoreactions were successfully completed on the electrode surface. When AFB1 exists in solution, the concentration of free antibody decreases, which leads to the reduction of the antibody volume combined on the AFB1–BSA surface and the corresponding electrochemical impedance of the immunosensor. All the impedance measurements were performed in the presence of a redox probe K4[Fe(CN)6]/K3[Fe(CN)6] at the scanning frequencies from 0.1 to 10 kHz. 3.3 Fig. 6 Relationship between the dilution ratio of antibodies and the impedance changes caused by the immunoreaction (n ¼ 3, error bars represent R.S.D.). surface-modied electrodes and has been extensively developed in the eld of immunosensors.27–29 A typical shape of a faradic impedance spectroscopy usually includes a semicircle portion and a straight line one. The semicircle portion is found at higher frequencies corresponding to the electron transfer limiting process, and the linear portion is found at the low frequencies resulting from the diffusion limiting step of the electrochemical process. As shown in Scheme 2, the impedance spectroscopy was represented as an equivalent circuit.30 The equivalent circuit includes the ohmic resistance of the solution resistance (Rs), the Warburg impedance element (Zw), the double layer capacitance (Cdl) and the electron transfer resistance (Ret). The two components, Rs and Zw, represent bulk properties of the electrolyte solution and diffusion feature of the redox probe in solution, respectively. Accordingly, these parameters are not affected by the electrochemical reaction occurring at the electrode surface. Cdl and Ret depend on the dielectric and insulating features at the electrode/electrolyte interface, respectively. This journal is © The Royal Society of Chemistry 2015 Optimization of the experimental conditions 3.3.1 Inuence of pH on the AFB1–BSA adsorption onto the colloidal Au-modied electrode. The congurations and electrostatic states of proteins usually depend on the medium pH, so that the pH value should be of great importance in the protein immobilization process. Fig. 4 shows the effects of pH of adsorbing AFB1–BSA onto the colloidal Au-modied electrode was tested over a pH range from 4.5 to 10.0. As manifested in Fig. 4, an optimum relationship between the impedance change of the resulting electrode and the pH of the PBS was observed at around pH 7.4, which was used in all experiments. 3.3.2 Effect of incubation time on the antibody immobilization. As seen in Fig. 5, the incubation time for the immunoassay was investigated as a factor inuencing the transducer performance. When the incubation time was over 50 min, the impedance changes did not increase further, indicating that the immunoreaction on the electrode was almost completed. Therefore, the immunochemical incubation time of 50 min was selected to evaluate the analytical performance of the sensor in all the subsequent assays. 3.3.3 Effect of the AFB1 monoclonal antibody concentration. The effect of the antibody concentration of the immunosensor in the immunoreaction was also investigated. The relation between the impedance changes caused by the immunoreaction and the dilution ratio of antibodies is shown in Fig. 6. One can observe that the impedance changes decrease with the AFB1 monoclonal antibody titer up to 1 : 20 000 and then decrease distinctly at higher dilution ratios. Because the number of adsorptive sites is limited, saturated binding of AFB1 could be reached by increasing the AFB1 antibody titer. Higher Anal. Methods, 2015, 7, 2354–2359 | 2357 View Article Online Analytical Methods Paper detection limit (three times the signal to noise ratio) of the immunosensor was about 0.05 ng mL 1. Published on 28 October 2014. Downloaded by Universidad Nacional Agraria La Molina on 10/06/2016 16:48:59. 3.5 Impedance changes (DRet) vs. the logarithm of the AFB1 concentration. All the measurements were carried out in 10 mM PBS (containing 0.1 M KCl, 5 mM K4[Fe(CN)6]/K3[Fe(CN)6], pH 7.4) and the data were recorded in the frequency range from 0.1 Hz to 10 kHz. Inset: linear relationship between the impedance changes (DRet) and the logarithm of the different concentration of the AFB1. Error bars are the standard deviation of the mean n ¼ 3. r2 ¼ 0.9919 (n ¼ 3, error bars represent R.S.D.). Fig. 7 titers of antibodies might presumably result in increased disorder in the alignment of adsorbed AFB1 antibodies and spatial hindrance of the immunoreaction. Thus, the AFB1 antibody concentration of 1 : 20 000 titer was recommended for all the subsequent experiments. 3.4 Detection of AFB1 The electrochemical impedance immunosensor was applied to the detection of AFB1 in samples to test its performance. Fig. 7 illuminates the impedance change (DRet) responses to the specic immunointeraction on the sensing interface plotted versus the logarithm value of concentration of AFB1. The Ret decreased clearly with the increase of the concentration of AFB1, revealing that the interactions between antibodies and antigens took place corresponding linearly to AFB1. A linear relation between the relative responses of the electron-transfer resistance and the logarithmic value of concentrations was observed in the range from 0.08 ng mL 1 to 100.0 ng mL 1 and the response equation was shown as DRet (kU cm2) ¼ 4.09–0.332 log C (ng mL 1) and a correlation coefficient was 0.9919 (n ¼ 6) (S/N ¼ 3) (inset in Fig. 7). The Table 1 Reproducibility and recovery of the modied electrodes To investigate the immunoassay reproducibility of the sensor fabricated with the label-free method, the various concentrations of AFB1 in the linear detection range were determined with the same electrode or different electrodes. The maximum relative standard deviations were 9.4% (n ¼ 6) for intra-assay and 10.8% (n ¼ 6) for inter-assay. To demonstrate the applicability and reliability of the immunosensor prepared using the present procedure, a recovery experiment of different AFB1 concentration was performed. As seen in Table 1, recovery in the range of 88.0–112% with a relative standard derivation of 5.2–9.6% was achieved with all the measurements carried out four times. 3.6 Control experiments To evaluate the selectivity of the proposed immunoassay, control experiments were performed using the other aatoxins (AFB2, AFM1, AFM2) for comparison with AFB1. These drugs belonged to the aatoxins class and were dissolved in PBS (pH 7.4). These different immunosensors were used to detect 50 ng mL 1 AFB1 separately, the alternation antigen using the same method. The results of the study showed that the Ab was relatively specic for AFB1 with cross-reactivity of 11% for AFM1, of 10% for AFM2, and of 40% for AFB2. The experimental results obtained indicate that the proposed sensor has reasonable specicity to AFB1 as applied to the selective determination of the target analytes. 3.7 Agricultural sample analysis The feasibility of the newly developed label-free immunoassay for possible applications was investigated by analyzing agricultural samples, such as mouldy rice, mouldy corn and groundnut samples. The samples were collected and the extraction procedure was performed as described by the literature.14 Table 2 lists the AFB1 level determination results and the relative deviations of the proposed label-free immunoassay method as compared with the conventional AFB1 analysis method in foods, ELISA. As can be seen in Table 2, the results of both methods were in reasonable agreement, suggesting that there is no difference of signicance between the two methods and using the label-free immunosensor for the determination Recovery studies of rice samples using the immunoassay for AFB1 Sample no. Added AFB1 (ng mL 1) Found AFB1a (ng mL 1) Recoverya (%) R.S.D.a (%) 1 2 3 4 5 0 2.5 12.5 50 100 0 2.2 11.4 52.2 99.0 — 88.0 91.2 104 99.0 — 5.9 9.0 7.5 8.8 a Each of the samples was evaluated in triplicate tests, and each of the results in Table 1 is the average of three parallel experiments. 2358 | Anal. Methods, 2015, 7, 2354–2359 This journal is © The Royal Society of Chemistry 2015 View Article Online Published on 28 October 2014. Downloaded by Universidad Nacional Agraria La Molina on 10/06/2016 16:48:59. Paper Analytical Methods Table 2 The results of AFB1 level determination for two methods, the studied method and the ELISA methoda Samples Mouldy rice extract Mouldy corn extract Groundnut extract 25.1 52.3 8.60 24.3 2.29 51.9 5.43 9.10 3.99 By the proposed method (ng mL 1) By ELISA method (ng mL 1) Relative deviation (%) a Each of the samples was evaluated in triplicate tests, and each of the results in Table 2 is the average of three parallel experiments. of AFB1 concentration in foods and agricultural products is feasible. 4 Conclusions In this work, a novel label-free impedance immunosensor for the direct detection of AFB1 in foodstuffs is described. Compared with the conventional detection methods, the immunoassay was simple to perform and had high selectivity in the detection of AFB1, without any redox probe or enzyme labeling. The reliability and applicability of directly detecting the antibody–antigen interaction by EIS measurement were also demonstrated. The results of the immunoassay for AFB1 in foods were satisfactory. As expected, the proposed immunoassay has a great potentiality to be a new alternative method for the detection of AFB1 in foods and agricultural products. Acknowledgements This work was supported by National Grand (2009ZX10004-312), NSFC (21025521, 21035001), CSIRT program, NSF from Hunan Province (10JJ7002) and General Administration of Quality Supervision, Inspection and Quarantine of the People's Republic of China (no. 2007GYB146). Notes and references 1 J. H. Owino, O. A. Arotiba, N. Hendricks, E. A. Songa, N. Jahed, T. T. Waryo, R. F. Ngece, P. G. Baker and E. I. Iwuoha, Sensors, 2008, 8, 8262. 2 M. S. Nasir and M. E. Jolley, J. Agric. Food Chem., 2002, 50, 3116. 3 J. Gilbert, Nat. Toxins, 1999, 7, 347. 4 J. Gilbert and E. Anklam, TrAC, Trends Anal. Chem., 2002, 21, 468. 5 L. Lin, J. Zhang, P. Wang, Y. Wang and J. Chen, J. Chromatogr. A, 1998, 815, 3. 6 Y.-Y. Hu, P. Zheng, Z.-X. Zhang and Y.-Z. He, J. Agric. Food Chem., 2006, 54, 4126. 7 I. A. Ahmed and R. K. Robinson, J. Agric. Food Chem., 1998, 46, 580. 8 M. Ventura, A. Gomez, I. Anaya, J. Dı́az, F. Broto, M. Agut and L. Comellas, J. Chromatogr. A, 2004, 1048, 25. This journal is © The Royal Society of Chemistry 2015 9 L. E. Edinboro and H. T. Karnes, J. Chromatogr. A, 2005, 1083, 127. 10 H. Reinhard and B. Zimmerli, J. Chromatogr. A, 1999, 862, 147. 11 C. Cavaliere, P. Foglia, C. Guarino, M. Nazzari, R. Samperi and A. Laganà, Anal. Chim. Acta, 2007, 596, 141. 12 C. M. Maragos and J. I. Greer, J. Agric. Food Chem., 1997, 45, 4337. 13 H. Cao, Food Control, 2011, 22, 43. 14 T. Yun, C. Xia, S. GuoLi and Y. RuQin, Anal. Biochem., 2009, 387, 82. 15 A. Korde, U. Pandey, S. Banerjee, H. Sarma, S. Hajare, M. Venkatesh, A. Sharma and M. Pillai, J. Agric. Food Chem., 2003, 51, 843. 16 S. Lipigorngoson, P. Limtrakul, M. Suttajit and T. Yoshizawa, Food Addit. Contam., 2003, 20, 838. 17 M. Zayats, Y. Huang, R. Gill, C.-a. Ma and I. Willner, J. Am. Chem. Soc., 2006, 128, 13666. 18 G. S. Bang, S. Cho and B.-G. Kim, Biosens. Bioelectron., 2005, 21, 863. 19 A.-N. Kawde, M. C. Rodriguez, T. M. Lee and J. Wang, Electrochem. Commun., 2005, 7, 537. 20 S. Zhang, F. Huang, B. Liu, J. Ding, X. Xu and J. Kong, Talanta, 2007, 71, 874. 21 S. Grant, F. Davis, K. A. Law, A. C. Barton, S. D. Collyer, S. P. Higson and T. D. Gibson, Anal. Chim. Acta, 2005, 537, 163. 22 X. Sun, H. Shi, H. Wang, L. Xiao and L. Li, Anal. Lett., 2010, 43, 533. 23 J.-D. Qiu, R.-P. Liang, R. Wang, L.-X. Fan, Y.-W. Chen and X.-H. Xia, Biosens. Bioelectron., 2009, 25, 852. 24 A.-E. Radi, X. Munoz-Berbel, V. Lates and J.-L. Marty, Biosens. Bioelectron., 2009, 24, 1888. 25 K. C. Grabar, R. G. Freeman, M. B. Hommer and M. J. Natan, Anal. Chem., 1995, 67, 735. 26 Z.-S. Wu, M.-M. Guo, S.-B. Zhang, C.-R. Chen, J.-H. Jiang, G.-L. Shen and R.-Q. Yu, Anal. Chem., 2007, 79, 2933. 27 L. Yang, Y. Li and G. F. Erf, Anal. Chem., 2004, 76, 1107. 28 Z. Chen, J. Jiang, G. Shen and R. Yu, Anal. Chim. Acta, 2005, 553, 190. 29 M. Dijksma, B. Kamp, J. Hoogvliet and W. Van Bennekom, Anal. Chem., 2001, 73, 901. 30 J. Randles, Discuss. Faraday Soc., 1947, 1, 1. Anal. Methods, 2015, 7, 2354–2359 | 2359