Highly sensitive ECL immunosensor based on multi-labeling of luminol via a dendrimer on Fe₃O₄ nanoparticles

Abstract

In this study, a dendrimer was used to multiply immobilize luminol for constructing highly sensitive microalbuminuria (mAlb) electrochemiluminescence (ECL) immunosensor. The composite dendrimer was linked to Fe₃O₄ magnetic nanoparticles by layer-by-layer self-assembly to fabricate a magnetically controlled mAlb immunosensor. The determination of mAlb concentration was realized by competition reactions between mAlb and glucose oxidase-labeled mAlb with anti-mAlb on the dendrimer. MAlb concentrations were determined between 2 and 322 pg mL⁻¹, and the detection limit was 0.084 pg mL⁻¹. The immunosensor proposed in this work was successfully applied to mAlb assay in urine samples.

---

Introduction

ECL immunosensors have gained considerable research interest because of the specific selectivity of immunoreactions, the high sensitivity of these types of sensors, and the convenience of ECL measurements. Luminescence reagent immobilization is a primary consideration during immunosensor fabrication because the compounds immobilized in the surface of the sensor could directly react with the analytes in solution. The strategies adopted for immobilization include chemical reaction, electrostatic adsorption, embedding and others.¹ ECL immunoassays have largely focused on improvements in sensitivity.² Several strategies have been reported to improve the sensitivity of ECL immunosensors, among which increasing the loading amount of luminescence reagents is a highly effective strategy.³ Nanoparticles, such as nano-Au,⁴ nano-Ag,⁵ carbon nanotube,⁶ and silica nanohybrids,⁷ may also be used to load luminescence reagents. Unfortunately, the loading amount on the immunosensor is limited by the binding sites and areas occupied by these particles. Another type of nanoparticles, magnetic nanoparticles (MNPs), have been developed to prepare sensitive and renewable immunosensors; these nanoparticles exhibit superparamagnetism as well as large specific surface areas and coupling volumes.⁸ Previous studies have reported the application of Fe₃O₄ magnetic nanoparticles (MNPs) in ECL immunosensors;⁹ however, the amount of immobilized molecules of luminescent probes and the intensity of the ECL signal remain limited.

Dendrimers, also known as hyperbranched molecules, are a class of artificial macromolecules synthesized by repeated stepwise reaction of branching units. The molecules have highly branched and 3D architectures with very low polydispersity and high degrees of surface functionality and versatility. The structure of dendrimers is built around a central multi-functional core molecule with branches and end groups.¹⁰ Dendrimers function as carriers of detection probes because of their unique architecture and macromolecular characteristics.¹¹

In this paper, a poly(diethylenetriaminepentaacetic acid-glycol ester) (PDTPA) dendrimer was used to multiply immobilized ECL reagents for immunosensor fabrication. MAlb was selected as an analyte because it is a biomarker for diabetic nephropathy and hypertensive renal damage. This biomarker is significant for the early diagnosis of renal complications from diabetes and pregnancy-induced high-pressure kidney damage. Trace levels of mAlb are observed in urine.¹² Thus, developing a sensitive method to determine anti-mAlb in fluids is important. To obtain a renewable sensitive film, a luminol–PDTPA–anti-mAlb composite was modified on Fe₃O₄ NPs. Composite NPs were then attached to a magnetically controlled glassy carbon electrode (GCE) by a magnet. The analyte mAlb could replace glucose oxidase (GOD) -labeled mAlb in competitive immunoreaction. Luminol was electrochemically oxidized into luminol radicals and reacted with H₂O₂ produced by the GOD left after the competitive reaction, and ECL signals produced could correspond with the mAlb concentrations in the samples. One diethylenetriaminepentaacetic acid (DTPA) molecule contains five carboxyl groups. The PDTPA dendrimer formed by the reaction between DTPA and ethylene glycol likewise contains carboxyl groups on its surface, which can connect with luminol molecules. The loading amount of luminol on the electrode surface greatly increased, thereby markedly enhancing ECL signals. Fig. 1 shows the structure of ECL immunoassay.

Fig. 1 Schematic diagram of the ECL biosensor.

Experiments

Apparatus and Reagents

Cyclic voltammetry (CV) and ECL measurements were carried out on a Model MPI-E ECL analyzer systems (Xi'an Remex Instrument Co. Ltd, China) using a three-electrode system consisted of a platinum wire as an auxiliary electrode, an Ag/AgCl electrode as reference electrode and the magnate immunosensor as working electrode. A pHS-2C model pH meter (Shanghai Leici Instruments, China) and a DK-8B Electrothermal Constant Temperature Incubator (Shanghai Jinghong Instruments, China) were also used. The X-ray diffraction experiment (XRD) was carried out on a X'Pert PRO X-ray diffractometer (PANalytical, Netherland). The infrared spectrometry experiment (IR) was performed on a Nicolet iS10 (Thermo Fisher scientific, USA).

MAlb (0.7 mg mL⁻¹) and anti-mAlb (1 mg mL⁻¹) were purchased from URIT Co., Ltd. (Guilin, China). Bovine serum albumin (BSA) was obtained from Weike biochemical reagent Co. Ltd. (Shanghai, China). 3(Aminopropyl)triethoxysilane (APS, 98%) was purchased from the Johnson Matthey Company (Alfa Aesar, USA). GOD (120 U mg⁻¹) was purchased from Sigma-Aldrich. FeCl₃·6H₂O, FeCl₂·4H₂O, thiourea, diethylenetriamminepentaacetic acid, 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) and N-hydroxysulfosuccinimide (Sulfo-NHS) were all obtained from Shanghai Convachem Co., Ltd and were used without further purification. 30% H₂O₂ was purchased from Beijing Chemical Reagent (Beijing, China), and a fresh solution of H₂O₂ was prepared daily before the experiments.

0.01 mol L⁻¹ luminol stock solution was prepared by dissolving 0.0886 g luminol (>98%, Fluka) in 0.1 mol L⁻¹ NaOH. 0.05 mol L⁻¹ tris(hydroxymethyl)aminomethane and 0.1 mol L⁻¹ HCl were used for the preparation of tris(hydroxymethyl)aminomethane (Tris)–HCl buffer solution. Phosphate buffered solution (PBS, pH 7.4) was prepared from 0.1 mol L⁻¹ Na₂HPO₄, 0.1 mol L⁻¹ NaH₂PO₄ and 0.1 mol L⁻¹ KCl.

Other reagents were of analytical reagent grade. All solutions were prepared with double distilled water (18.2 MΩ cm⁻¹).

Preparation of immunosensor

Preparation and modification of the Fe₃O₄ MNP. Fe₃O₄ MNP were synthesized according to our previous work.^11b Firstly, Fe₃O₄ was prepared by chemical coprecipitation of Fe(II) and Fe(III) chlorides (molar ratio 2 : 1) in water at the temperature of 35 °C. 1 mol L⁻¹ NaOH was added to the solution to maintain the pH value about 10 under vigorous agitation at 50 °C. Then the solution was heated at 80 °C for 1 h in N₂ atmosphere. Finally the resulting precipitate was separated by magnetic decantation and washed with double-deionized water. The Fe₃O₄ MNPs product was vacuum-dried in furnace at 60 °C for 6 h, then stored protected by N₂ atmosphere for further use. The MNPs was characterized by XRD, and the size was around 20 nm.

24 mg Fe₃O₄ MNP was taken and dispersed in 10 mL methanol under ultrasonic processing, then 200 μL APS was added into the mixture solution and stirred vigorously on a shaking bed for 2 h in N₂ atmosphere. After the removal of excess APS by centrifugation, the precipitate was washed by methanol and double distilled water till neutral. The amino group modified Fe₃O₄ MNP was obtained.

Synthesis of PDTPA and luminol–PDTPA–anti-mAlb. PDTPA was prepared according to our previous work.^11b 1.5 g DTPA, 2 mL ethylene glycol, 20 mL dimethylsulfoxide and 2 mL pyridine were mixed together. Then 1 mL 4.6 mg mL⁻¹ SnCl₂ solution was added under vigorously stirring. The reaction was carried out at 140–150 °C for 5 h. Finally, volatiles were removed under reduced pressure at 100 °C. All processes were performed under N₂ protection.

Luminol–PDTPA and luminol–PDTPA–anti-mAlb was synthesized by the method revised in the ref. 13. 0.3 g PDTPA, 500 μL EDC (20 mg mL⁻¹), 500 μL Sulfo-NHS (10 mg mL⁻¹) were mixed and reacted for 5 h. Then 400 μL luminol (0.01 mmol L⁻¹) was added in the solution and reacted for 12 h in dark to form luminol–PDTPA. The resulting mixture was centrifuged to eliminate the supernate. The remained luminol–PDTPA precipitate was mixed with 200 μL mAlb (0.1 mg mL⁻¹) and reacted for 8 h at 50 °C. The solution was then dialyzed against 0.01 mol L⁻¹ PBS (pH 7.4) at 4 °C for another 8 h to produce the luminol–PDTPA–anti-mAlb. Then the product was diluted to 2 mL with 0.01 mol L⁻¹ PBS (pH 7.4) for further use.

Synthesis of the immuno-MNPs. 0.15 g Fe₃O₄ NPs was put at the bottom of a 10 mL baker by strong magnet, 2 mL PBS containing 10 mmol L⁻¹ thiourea was added and reacted for 4 h, then luminol–PDTPA–anti-mAlb solution for 3 h. The multiple dendrimer layers were immobilized on the Fe₃O₄ NPs by layer-by-layer self-assembly method. The product immuno-MNPs were store at 4 °C for further use.

Synthesis of GOD labeled mAlb (GOD-mAlb). The labeling of mAlb by GOD also used the sulfo-NHS and EDC. The operational steps and dosage were same with the method mentioned above. The product was diluted to 2 mL with 0.01 mol L⁻¹ PBS (pH 7.4) before further use.

The fabrication of immunosensor. 2.5 mg immuno-MNPs was immobilized on the GCE by strong magnet, then the magnetic GCE was put into 0.01 mol L⁻¹ PBS (pH 7.4) solution containing 5 μg mL⁻¹ GOD-mAlb. The sensor was incubated in 1% BSA solution at 37 °C to block the active sites remained.

During the assay, the immunosensor was immersed in mAlb solution ranged from 2 to 322 pg mL⁻¹ for 30 min for the competition between the mAlb in the solution and the GOD-mAlb on MNPs. After each assay, the immunosensor should be stored in refrigerator at 4 °C.

The ECL detection. The ECL detection was carried out in 0.05 mol L⁻¹ Tris–HCl buffer solution (pH 7.5) containing 3 mmol L⁻¹ glucose with the scan range of −0.3–0.6 V (vs. SCE) and the scan rate of 100 mV s⁻¹. The high voltage of photomultiplier tube was set at 800 V. The measuring time was 90 s. The ECL signal–time curve under continuous potential scanning was performed for 5 cycles with the magnification of 4. The immunosensor was stored at 4 °C when not in use.

Results and discussion

XRD studies of the Fe₃O₄ magnetic nanoparticle

Fig. 2 shows the XRD pattern of the Fe₃O₄ magnetic nanoparticle, which is quite identical to pure magnetite and can be indexed to (220), (311), (400), (422), and (511) planes of a cubic unit cell, which corresponds well with the magnetite structure (JCPDS card no. 65-3107) and reference reported.¹⁵ The mean crystal size determined by Debye–Scherrer equation has been found to be about 20 nm. Also, it can be seen that no characteristic peaks of impurities were observed.

Fig. 2 The XRD study of the Fe₃O₄ magnetic nanoparticle.

The IR studies of the synthesized materials

In order to characterize the materials synthesized, IR measurement was carried out. As shown in Fig. 3, the PDTPA synthesized has wide peak in the range of 3100 cm⁻¹ to 3500 cm⁻¹, which belongs to the carboxylic acid group of the dendrimer. In PDTPA–luminol and luminol–PDTPA–anti-mAlb, the peak around 1640 cm⁻¹ belongs to the amide group formed. Compared with the luminol–PDTPA–anti-mAlb, the fine structure in the fingerprint area disappeared, partly due to the steric hindrance and the complicated structure of the anti-mAlb. The modified Fe₃O₄ MNP has wide peak around 3300 cm⁻¹ to 3500 cm⁻¹, which belongs to the amine group of the APS. The results correspond well with ref. 16. The final product of Fe₃O₄–luminol–PDTPA–anti-mAlb also has the amide group peak at 1640 cm⁻¹ These could all partly verify the formation of the target materials.

Fig. 3 The IR study of the synthesized materials.

The effect of the layer number on the ECL signals

As shown in Fig. 4, with the increase of the number of the self-assembly dendrimer layers, the luminous intensity of the immunosensor increased. When the number reached 5, it reached the maximum value. For the more, the layer would not block the diffusion of the reactants and the products when the layer number was 5. So 5 layers was chosen for the all assay.

Fig. 4 The effect of layer number on ECL response a–f: self-assembled layers (n) = 1, 2, 3, 4, 5, 6.

The CV and ECL characterization of the immunosensor

There have been many reports about combing usage of the GOD and luminol in the fabrication of ECL sensors.¹⁴ The mechanism is that the electrochemical catalyzation of the GOD on glucose could form H₂O₂, and then H₂O₂ could react with the luminol labelled on the electrode and thus the ECL reaction could be carried out. Due to the highly catalyzing efficiency of GOD, after the incubation step, the smaller the concentration of mAlb in the solution, the larger the amount of GOD-mAlb that remains on the electrode. The ECL intensity of the test results would also be bigger. CV and ECL response of the immunosensor were also studied (Fig. 5). The bare luminol–PDTPA–anti-mAlb@Fe₃O₄ electrode had the highest current response (a′) and no ECL response (a). After the specific combination of the GOD-mAlb, the current response dropped (b′) with the ECL response increased largely (b). The decrease of the current could be attributed to the steric hindrance of the GOD-mAlb, which block the electron transfer from the electrode to the substrate solution. After the addition of mAlb, the current response dropped slightly further (c′), but the ECL response dramatically decreased (c). The current only changed slightly because the mAlb and GOD-mAlb are both macromolecules, which had same blocking effect, although the concentrations of them were different, which could lead to small changes of sensor, so was the current of the assay. The demonstration proved that there were competition reaction between the mAlb and the GOD-mAlb.

Fig. 5 CV and ECL response of the immunosensor a (a′): luminol–PDTPA–anti-mAlb@Fe₃O₄; b (b′): luminol–PDTPA–anti-mAlb–(mAlb-GOD)@Fe₃O₄; c (c′): luminol–PDTPA–anti-mAlb–(mAlb-GOD)@Fe₃O₄ + 300 pg mL⁻¹ mAlb. Measurement solution: 3 mmol L⁻¹ glucose, 0.05 mol L⁻¹ Tris–HCl, 35 ± 1.0 °C.

The optimization of the of the buffer solution and its pH

The catalytic activity of the enzyme and the reaction between antibody and antigen were affected by the buffer solutions. The intensities of the ECL signal of the immunosensor were tested in 0.1 mol L⁻¹ borate buffer solution, 0.05 mol L⁻¹ Tris–HCl and 0.1 mol L⁻¹ PBS. Then the 0.05 mol L⁻¹ Tris–HCl buffer solution was chosen.

The effects of pH on the ECL response were also evaluated in the range of 6.0–9.0. As shown in Fig. 6, the experiment results showed that the ECL intensity got a highest value at pH 7.5. With the increase of the pH away from 7.5, the ECL intensities of the system decreased. This might partly owing to the slow decomposition of the PDTPA in alkaline solution. So pH 7.5 was used for all the assay.

Fig. 6 The effect of pH on ECL response in 0.05 mol L⁻¹ Tris–HCl buffer solution containing 3 mmol L⁻¹ glucose at 35 ± 1.0 °C, with incubation time of 30 min.

The optimization of the immunoreaction time and temperature

The effects of specific immunoreaction time and temperature on the ECL intensity were also evaluated. The reaction time was tested between 5–60 min. As shown in Fig. 7a, the ECL response reached the highest value at 35 min and remained almost the same after that, thus 35 min was chosen as the immunoreactions time for the subsequent assay. The temperature on the assay was also tested. With the increase of the temperature from 5 to 35 °C, the ECL intensity increased and reached the maximum value at 35 °C (Fig. 7b). When the temperature continued to increase, the ECL intensity decreased, partly due to the gradual losing of the activities of the antigen/antibody and enzyme.

Fig. 7 Effect of incubation time (a) and incubation temperature (b) on the ECL intensity. All the other conditions were the same to Fig. 6.

ECL response of the immunosensor to mAlb concentrations

As shown in Fig. 8, under the optimized conditions, the intensity ratio of I₀/I (I₀ was blank value) has a linear relationship with the concentration of the mAlb in the range of 2–322 pg mL⁻¹ with the equation of I₀/I = 1.1014 + 0.0105C (pg mL⁻¹) and a correlation coefficient of r = 0.9953. The DL was 0.084 pg mL⁻¹, with the signal-to-noise ratio of 3.

Fig. 8 Responses and calibration curve (inset) of immunosensor. The mAlb concentrations are 0, 2, 12, 22, 42, 82, 162, 322 pg mL⁻¹ (a–h) respectively.

In order to evaluate the sensitivity of the immunosensor, some parameters of reported immunoassays for mAlb were compared and the results are shown in Table 1. The detection limit of the developed immunosensor is the lowest one. The result suggests that the immunosensor is highly sensitive and has great potential for accurate detection of mAlb.

Table 1 Comparison of some permanents of various immunosensors for mAlb

Detection methods	System	Linear range	DL	Ref.
Fluorescence spectrometry	Albumin/creatinine ratio	5–220 mg mL^−1	0.02 mg mL^−1	17
Fluorescence immunoassay	Non-immunological dye binding assay	0–600 mg L^−1	—	18
Fluorescence immunoassay	Anti-rabbit IgG/albumin/fluorophore	4.65–600 mg L^−1	1.6 mg mL^−1	19
Sequential injection analysis system based on fluorescent	Albumin-dye blue 580	1–100 mg L^−1	0.3 mg L^−1	20
Quartz crystal microbalancer	Trimethylolpropane trimethacrylate–mAlb	60–150 ppm	—	21
Electrochemiluminescence	Luminol–PDTPA–anti-mAlb@Fe3O4	2–322 pg mL^−1	0.084 pg mL^−1	This work

---

Selectivity of the immunosensor

Selectivity is a very important parameter to evaluate the validation of a sensor. Thus in order to investigate the selectivity of the immunosensor, the influence of possible interfering substances in urine samples such as inorganic ions, amino acids and others on the intensity of the ECL signals was tested. The results showed that when the mAlb concentration was 50 pg mL⁻¹ and the relative error was less than ±5%, substances like K⁺, NO₃⁻ (35 μmol L⁻¹), Na⁺, Ca²⁺ (100 μmol L⁻¹), Zn²⁺, Cu²⁺ (10 μmol L⁻¹), NH₄⁺, SO₄²⁻ (10 μmol L⁻¹), L-tyrosine (30 μmol L⁻¹), L-tryptophan (10 μmol L⁻¹), L-hydroxyproline (30 μmol L⁻¹), ascorbic acid (20 μmol L⁻¹), xanthine (83 μmol L⁻¹).

The reproducibility and stability of the immunosensor

The reproducibility and stability of the immunosensor were also examined to evaluate the performance of the immunosensor. When the mAlb concentration was 50 pg mL⁻¹, sensors prepared by five different NPs in a batch and five NPs in different batches were tested respectively. The relative standard deviations (RSD) of the ECL signals were 2.24% and 4.52% respectively, showing that the immunosensor had good reproducibility. When the mAlb concentration was 50 pg mL⁻¹, the ECL signals had no apparent change after 2 weeks. The ECL intensity dropped by 4.4% after 3 weeks and 8.2% after 6 weeks. These demonstrated that the immuno-NPs had good stability. It also demonstrates the layers are stable, probably due to the amide group formed during the synthesize, which make strong connection between the modified Fe₃O₄ MNP and PDTPA dendrimer.

Application of the immunosensor to urine samples

The practical application of the immunosensor was demonstrated by analyzing five fresh human urine samples. Before the test, the samples were centrifuged at 3000 rpm for 10 min. The supernatant was diluted appropriately step by step by PBS (pH 7.4) in order to obtain responses in the linear range of the proposed method. The analytical results were shown in Table 2. The results showed satisfied recoveries in the range of 92.1–105.9% which indicated that the developed immunoassay methodology might be a precise tool for the detection of mAlb. This could also verify the validation of the immunosensor.

Table 2 The determination of mAlb in human urine

Samples	DL (mg L^−1)	RSD % (n = 5)	Added (mg L^−1)	Detected (mg L^−1)	RSD % (n = 5)	Recoveries (%)
1	3.73	4.3	10	12.65	3.6	92.1
2	5.39	2.8	10	15.97	2.2	103.8
3	7.60	1.9	10	17.36	2.7	98.6
4	9.17	2.0	10	20.05	1.5	105.0
5	10.42	2.7	10	21.63	2.3	105.9

---

Conclusion

The present study demonstrated the fabrication of an ECL immunosensor based on the self-assembly of multiple luminol-labeled dendrimers. ECL signals were dramatically amplified by the multiple labeling strategy. The results obtained indicate a sensitive method for determining mAlb. The proposed immunosensor could be easily renewed, conveniently use, and rapidly operated (within 40 min). Dendrimers with a 3D network structures were used for luminol labeling, which produces signal amplification in immunoassays.

References

1. X. Pei, B. Zhang, J. Tang, B. Liu, W. Lai and D. Tang, Anal. Chim. Acta, 2012, 758, 1.
2. (a) C. Wang, J. Wu, C. Zong, J. Xu and H. X. Ju, Chin. J. Anal. Chem., 2012, 40, 3; (b) A. Roda and M. Guardigli, Anal. Bioanal. Chem., 2012, 402, 1.
3. (a) S. Bi, Y. Yan, X. Yang and S. Zhang, Chem.–Eur. J., 2009, 15, 4704; (b) G.-F. Jie, P. Liu and S.-S. Zhang, Chem. Commun., 2010, 46, 1323.
4. (a) Z.-P. Li, Y.-C. Wang, C.-H. Liu and Y.-K. Li, Anal. Chim. Acta, 2005, 551, 85; (b) Y. Li, L. Qi, Y. Shen and H. Ma, Anal. Chim. Acta, 2014, 811, 36.
5. X. He, C. Hu, H. Liu, G. Du, Y. Xi and Y. Jiang, Sens. Actuators, B, 2010, 144, 289.
6. Y. Tao, Z.-J. Lin, X.-M. Chen, X.-L. Huang, M. Oyama, X. Chen and X.-R. Wang, Sens. Actuators, B, 2008, 129, 758.
7. X. Wang, M. Zhou, Y. Zhu, J. Miao, C. Mao and J. Shen, J. Mater. Chem. B, 2013, 1, 2132.
8. (a) A. H. Lu, E. e. L. Salabas and F. Schüth, Angew. Chem., Int. Ed., 2007, 46, 1222; (b) G. Reiss and A. Hütten, Nat. Mater., 2005, 4, 725.
9. (a) E. Han, L. Ding, S. Jin and H. Ju, Biosens Bioelectron, 2011, 26, 2500–2505; (b) G. Guan, L. Yang, Q. Mei, K. Zhang, Z. Zhang and M.-Y. Han, Anal. Chem., 2012, 84, 9492.
10. D. A. Tomalia, Prog. Polym. Sci., 2005, 30, 294.
11. (a) F. Vögtle, G. Richardt, N. Werner, Dendrimer chemistry, Wiley–VCH, Datenschutz, 2009; (b) J. Li, Q. Xu, C. Fu and Y. Zhang, Sens. Actuators, B, 2013, 185, 146; (c) C. Fu and J.-P. Li, Chin. J. Anal. Chem., 2013, 41, 1762.
12. S. Basu, M. Bhattacharya, A. Majumdar, T. Chatterjee and S. Todi, J. Crit. Care, 2009, 13, 380.
13. X. Yang, Y. Guo and A. Wang, Anal. Chim. Acta, 2010, 666, 91.
14. X. Chena, B. Suc, X. Songa, Q. Chend, X. Chena and X. Wanga, TrAC, Trends Anal. Chem., 2011, 30, 665.
15. A. Kaushik, P. R. Solanki, K. K. Kaneto, C. G. Kim, S. Ahmad and B. D. Malhotra, Electroanalysis, 2010, 22, 1045.
16. A. Kaushik, P. R. Solanki, A. A. Ansari, S. Ahmad and B. D. Malhotra, Electrochem. Commun., 2008, 10, 1364.
17. C.-C. Lin, J.-L. Hsu, C.-C. Tseng and G.-B. Lee, Microfluid. Nanofluid., 2011, 10, 1055.
18. S. Choi, E. Y. Choi, H. S. Kim and S. W. Oh, Clin. Chem., 2004, 50, 1052.
19. W. Laiwattanapaisal, U. Kunanuvat, W. Intharachuti, C. Chinvongamorn, S. Hannongbua and O. Chailapakul, Talanta, 2009, 79, 1104.
20. K. Omidfar, A. Dehdast, H. Zarei, B. K. Sourkohi and B. Larijani, Biosens. Bioelectron., 2011, 26, 4177.
21. T.-Y. Lin, C.-H. Hu and T.-C. Chou, Biosens. Bioelectron., 2004, 20, 75.

---