A label-free electrochemical immunosensor with a novel signal production and amplification strategy based on three-dimensional pine-like Au–Cu nanodendrites

Abstract

A simple and novel label-free electrochemical immunosensor based on a bimetallic alloy was fabricated in this work. Three-dimensional (3D) pine-like Au–Cu nanodendrites (NDs) electrodeposited on the bare glassy carbon electrode (GCE) were employed as transducing materials to achieve the production and amplification of the electrochemical signal. The electrochemical detection was based on the fact that the immunocomplex can hinder the direct electron transfer in the redox process of Cu. Cu, instead of the traditional redox electron mediators, was exploited as a new signal producer. Au with good electron transfer ability could act as signal amplifier and capture antibodies. In addition, the 3D pine-like morphology of Au–Cu NDs with large specific surface area was beneficial to the immobilization of antibodies and the electron transfer between protein and metal. Under the optimal conditions, by using carbohydrate antigen 724 (CA724) as the target analyte, the label-free electrochemical immunosensor exhibited a wide linear range from 0.5 pg mL⁻¹ to 500 ng mL⁻¹ and a low detection limit of 0.25 pg mL⁻¹. The proposed immunosensor also displayed a satisfying electrochemical performance with good reproducibility, selectivity and stability. In this sense, this novel method will provide a new promising application for the clinical immunoassay of other biomarkers.

---

1. Introduction

In recent years, the electrochemical immunosensor has attracted extensive attention in different fields including environmental monitoring,^1,2 food safety^3–5 and clinical diagnosis^6–8 because it has attractive features including high sensitivity, high selectivity, easy operation and fast analysis.^9,10 As one important branch of electrochemical immunosensors with different fabricated models, the label-free electrochemical immunosensor has obtained wide interest,^11–14 owing to its direct detection of tumor markers. The use of label-free electrochemical immunosensors can not only avoid complicated and time-consuming procedures, but also facilitates the miniaturization of highly portable equipment.¹⁵ In particular, the additional advantage of the label-free electrochemical immunosensor is that it could avoid disturbances of the labels conjugated on the antibodies, which allows the label-free electrochemical immunosensor to be employed in broad biological systems.¹⁶

The electrochemical mechanism of the label-free electrochemical immunosensor is described as follows. Redox electron mediators, such as prussian blue,¹⁷ ferrocenecarboxylic acid,¹⁸ thionine¹⁹ and ferricyanide,²⁰ are employed to produce the electrochemical signal for the label-free electrochemical immunosensor. When the immunoreaction appears on the surface of the modified electrode, the nonconductive immunocomplex will cause the increase of the resistance to hinder the direct electron transfer generating from the redox electron mediator, resulting in the decrease of the electrochemical signal. Therefore, the signal producer is essential in the fabrication of the label-free immunosensor. In addition, another significant issue in fabricating a label-free electrochemical immunosensor is the efficient capture of antibodies on the surface of the electrode.²¹ In conclusion, the transducing materials firmly immobilized on the bare glassy carbon electrode (GCE), with the excellent characteristics of producing the electrochemical signal and capturing the antibodies, will play a very important role in the fabrication of the label-free electrochemical immunosensor.

In our group's previous researches, various kinds of transducing materials have been employed to fabricate the label-free electrochemical immunosensor. Silver hybridized mesoporous ferroferric oxide nanoparticles and thionine mixed graphene sheet were developed in the label-free electrochemical immunosensor for the detection of kanamycin.¹⁹ But it contained a relatively complex process of preparing the transducing materials. Apart from that, a label-free electrochemical immunosensor with simple operation was also proposed based on Pd nanoplates for the detection of alpha fetoprotein (AFP).²² However, the Pd nanoplates were immobilized on the electrode by physical adsorption. In this sense, to improve the stability of the label-free electrochemical immunosensor, the immobilization format of transducing materials should be further improved. Inspired by the electrodeposition work of Wang's group,²³ three-dimensional (3D) pine-like Au–Cu nanodendrites (NDs) were electrodeposited on the GCE and then acted as transducing materials for the fabrication of the label-free electrochemical immunosensor.

There are five advantages of the electrodeposited Au–Cu bimetallic alloy. First, it is prepared on the GCE by the one-step electrodeposition method, which can simplify the experimental procedures. Second, the electrodeposition method shows better stability than the physical adsorption method. Third, Cu can generate redox reaction and produce the electrochemical signal directly. Without introducing any redox electron mediator, Cu also can reduce the manufacturing cost compared with the noble metal like Pd or Pt. Fourth, Au with favorable chemistry characteristics towards the functionalization with biomolecules can be employed as a cross linker to capture antibodies by the bonding of Au–N.²⁴ And the good conductivity of Au can also facilitate the electron transfer. Fifth, the special morphology of 3D pine-like NDs has a larger specific surface area, which can further increase the capture of antibodies and accelerate the electron transfer between the interface of protein and metal. Therefore, the 3D pine-like Au–Cu NDs prepared by a simple electrodeposition method can simultaneously achieve the production and amplification of electrochemical signal.

In this work, a simple and novel label-free electrochemical immunosensor was fabricated by using 3D pine-like Au–Cu NDs as signal producers and amplifiers simultaneously. Carbohydrate antigen 724 (CA724), which was over-expressed in various carcinomas including ovarian, gastric, colorectal, and breast cancers,²⁵ was used as target analyte to be detected by the designed immunosensor. The attractive electrochemical performance of the designed label-free electrochemical immunosensor for the quantitative detection of CA724 will be presented in detail.

2. Materials and methods

2.1. Apparatus and reagents

All electrochemical measurements were performed on a CHI760D electrochemical workstation (Chenhua Instrument Shanghai Co., Ltd, China). Scanning electron microscopy (SEM) images and energy-dispersive X-ray spectroscopy (EDX) analysis were collected using a FEI QUANTA FEG250 coupled with INCA Energy X-MAX-50. A conventional three-electrode system was used for all electrochemical measurements: a GCE (3 mm in diameter) as the working electrode, a saturated calomel electrode (SCE) as the reference electrode, and the platinum wire electrode as the counter electrode.

Human CA724 antigen and antibody to human CA724 were purchased from Shanghai Linc-Bio Science Co., Ltd, China. Chloroauric acid (HAuCl₄·4H₂O), cupric sulfate (CuSO₄) and sodium sulfate (Na₂SO₄) were purchased from Shanghai Aladdin Chemistry Co., Ltd, China. Phosphate buffered saline (PBS) was used as an electrolyte for all electrochemistry measurements. All other reagents were of analytical grade and ultrapure water was used throughout the study.

2.2. Preparation of the 3D pine-like Au–Cu NDs

Typically, the 3D pine-like Au–Cu NDs were directly prepared by a one-step electrodeposition on a GCE in the electrolyte solution containing 5 mM HAuCl₄, 10 mM CuSO₄, 10 mM Na₂SO₄ and 10 mM H₂SO₄. Electrodeposition potential was conducted at −0.8 V for 600 s, followed by thoroughly washed with water and dried at room temperature. For comparison, Au nanoparticles (NPs) were prepared by using the same electrolyte without CuSO₄ under the same electrochemical experiment condition.

2.3. Fabrication of the immunosensor

Fig. 1 shows the schematic diagram of the designed label-free electrochemical immunosensor. A GCE was polished repeatedly with 1.0, 0.3 and 0.05 μm alumina powder and sonicated in ethanol for 30 s. After washed with ultrapure water thoroughly, the clean GCE was used to be electrodeposited in the electrolyte solution by the amperometric i–t method. Then, 6 μL of anti-CA724 solution (10 μg mL⁻¹) was added onto the electrode and dried at 4 °C. After washing, 3 μL of 1 wt% bovine serum albumin (BSA) solution was added to eliminate nonspecific binding sites. After that, the modified electrode was washed and incubated with various concentrations of CA724 for 1 h at room temperature. Finally, the modified electrode was washed thoroughly to remove the unbounded CA724 molecules and ready for measurement with cyclic voltammogram (CV), which was recorded in PBS at pH 6.8 by scanning the potential from −0.8 V to 0.8 V.

Fig. 1 The schematic diagram of the label-free electrochemical immunosensor.

3. Results and discussion

3.1. Characterization

The remarkable 3D pine-like morphology with large specific surface area was confirmed by the SEM images of the electrodeposited Au–Cu NDs. Fig. 2A displays that Au–Cu NDs with high density grow layer-by-layer on the GCE surface. The secondary branches are evenly distributed on both sides of the main branch, just like the layout of pine leaves (Fig. 2B). EDX picture (Fig. 2C) was used to confirm the elements contained in the 3D pine-like NDs. Peaks of Au and Cu are detected without any impurities, suggesting the high purity of the electrodeposited Au–Cu NDs. And the C element is derived from the GCE. In addition, the corresponding EDX mappings display that Au (Fig. 2E) and Cu (Fig. 2F) are uniformly distributed over the entire 3D pine-like NDs (Fig. 2D).

Fig. 2 (A) SEM images of 3D pine-like Au–Cu NDs (A, B and D); EDX spectrum of 3D pine-like Au–Cu NDs (C); EDX mappings of Au (E) and Cu (F) in 3D pine-like Au–Cu NDs.

In order to obtain a remarkable 3D pine-like morphology, different experimental parameters were investigated. First, different molar ratios of HAuCl₄ and CuSO₄ were investigated. When the molar ratio of HAuCl₄ and CuSO₄ is 2 : 1 (Fig. 3A), the shape and size of NPs are irregular. When the molar ratio of HAuCl₄ and CuSO₄ is 1 : 1 (Fig. 3B), the NPs have a short foxtail-like morphology. Therefore, the molar ratio of HAuCl₄ and CuSO₄ can influence the morphology of NPs in the process of electrodeposition. Second, different electrodeposition times were investigated. When the electrodeposition time is 400 s (Fig. 3C), the size of NDs is obviously smaller. When the electrodeposition time is 200 s (Fig. 3D), the 3D pine-like morphology can not even form. Therefore, the electrodeposition time plays an important role in controlling the size of NDs. Third, different electrodeposition potentials were investigated. When the potential is not −0.8 V (Fig. 3E and F), it is much easier to form a film rather than the NDs. Therefore, the electrodeposition time has an important effect in controlling the shape of NDs. Forth, the influences of Na₂SO₄ and H₂SO₄ were investigated. When there is no Na₂SO₄ and H₂SO₄ in the electrolyte solution (Fig. 3G), the morphology of NDs is mixed and disorderly. When there is no Na₂SO₄ in the electrolyte solution (Fig. 3H), the NDs have a disorderly wheat-like morphology. When there is no H₂SO₄ in the electrolyte solution (Fig. 3I), the NDs have an orderly wheat-like morphology. Therefore, Na₂SO₄ and H₂SO₄ are helpful to form the 3D pine-like morphology. In conclusion, the special 3D pine-like morphology can be obtained only at the following experimental parameter: 5 mM HAuCl₄, 10 mM CuSO₄, 10 mM Na₂SO₄ and 10 mM H₂SO₄ at −0.8 V for 600 s.

Fig. 3 SEM images of different experimental parameters: 10 mM HAuCl₄, 5 mM CuSO₄, 10 mM Na₂SO₄ and 10 mM H₂SO₄ at −0.8 V for 600 s (A); 5 mM HAuCl₄, 5 mM CuSO₄, 10 mM Na₂SO₄ and 10 mM H₂SO₄ at −0.8 V for 600 s (B); 5 mM HAuCl₄, 10 mM CuSO₄, 10 mM Na₂SO₄ and 10 mM H₂SO₄ at −0.8 V for 400 s (C) and 200 s (D); 5 mM HAuCl₄, 10 mM CuSO₄, 10 mM Na₂SO₄ and 10 mM H₂SO₄ at −0.4 V (E) and 0 V (F) for 600 s; 5 mM HAuCl₄ and 10 mM CuSO₄ at −0.8 V for 600 s (G); 5 mM HAuCl₄, 10 mM CuSO₄ and 10 mM H₂SO₄ at −0.8 V for 600 s (H); 5 mM HAuCl₄, 10 mM CuSO₄ and 10 mM Na₂SO₄ at −0.8 V for 600 s (I).

3.2. Electrochemical characterization

CV (Fig. 4) was used to characterize the electrochemical signal of Au NPs (curve a) and 3D pine-like Au–Cu NDs (curve b). There is no obvious oxidation or reduction peaks in curve (a), which indicates that Au can not be oxidized or reduced at −0.8 V to 0.8 V in PBS at pH 6.8. However, there are obvious redox peaks in curve (b), which well confirms that the oxidation and reduction peaks are all generating from the Cu in 3D pine-like Au–Cu NDs. It can be observed that the CV curve of 3D pine-like Au–Cu NDs has two reduction peaks. The mechanism of electrochemical signal generation can be expressed as following:

Cu + H₂O − 2e⁻ → CuO + 2H⁺ (1)

2CuO + 2e⁻ + 2H⁺→ Cu₂O + H₂O (2)

Cu₂O + 2e⁻ + 2H⁺ → 2Cu + H₂O. (3)

Fig. 4 CVs of Au NPs (a) and 3D pine-like Au–Cu NDs (b) in PBS at pH 6.8.

Obviously, the electrochemical signal of reduction peak at −0.3 V is the most sensitive, which is employed to characterize the immunosensor.

Chronoamperometry^26–28 was employed to evaluate the electrochemically active surface area of the 3D pine-like Au–Cu nanodendrites modified electrode in 10 mM ferricyanide, with 0.5 M KCl as the supporting electrolyte (Fig. 5A). According to the Cottrell equation: i = nFACD^1/2π^−1/2t^−1/2, where i is the current (A), t is the time (s), F is the Faraday constant (96 485 C mol⁻¹), n is the number of electrons transferred (=1 in this case), D is the diffusion coefficient of ferricyanide (1.6 × 10⁻⁵ cm² s⁻¹), C is the ferricyanide concentration (10 mM), the electrochemically active surface area (A, cm²) can be calculated. A plot of i vs. t^−1/2 transforms the data into a linear relationship with a slope that equals nFACD^1/2π^−1/2 (Fig. 5B). The electrochemically active surface area of the 3D pine-like Au–Cu nanodendrites modified electrode calculated from this slope was 0.541 cm², which was about 7.6 times of the geometric area of GCE (0.071 cm²), indicating that the 3D pine-like Au–Cu nanodendrites has a larger electrochemically active surface area.

Fig. 5 (A) Chronoamperometric curve of the 3D pine-like Au–Cu nanodendrites modified electrode in 10 mM ferricyanide, with 0.5 M KCl as the supporting electrolyte; (B) A plot of i vs. t^−1/2 data obtained from the early stages of the experimental chronoamperometric curve data.

3.3. Optimization of pH

In order to achieve an optimal electrochemical signal, the optimization of pH is necessary. The optimal electrochemical signal response of reduction peak at −0.3 V is achieved at pH 6.8 (Fig. 6). Under the optimal pH, the designed immunosensor can exhibit an optimal electrochemical signal for the quantitative detection of CA724.

Fig. 6 Effect of pH on the reduction peak current responses of the 3D pine-like Au–Cu NDs at −0.3 V. Error bar = RSD (n = 5).

3.4. Characterization of the immunosensor

CV was used to characterize the fabrication process of the label-free electrochemical immunosensor (Fig. 7). It can be observed that the bare GCE exhibits no redox peaks (curve a) at −0.8 V to 0.8 V. After disposition of 3D pine-like Au–Cu NDs, obvious redox peaks (curve b) are observed attributing to the redox process of Cu. When anti-CA724 (curve c), BSA (curve d), and CA724 (curve e) were modified layer-by-layer on the electrode, the gradually decreasing electrochemical signals indicate the successful modification of the nonconductive bioactive substances.

Fig. 7 CVs recorded in PBS at pH 6.8: bare GCE (a), Au–Cu NDs/GCE (b), anti-CA724/Au–Cu NDs/GCE (c), BSA/anti-CA724/Au–Cu NDs/GCE (d), CA724/BSA/anti-CA724/Au–Cu NDs/GCE (e).

3.5. Detection of CA724

Under the optimal conditions, the label-free electrochemical immunosensor based on using the 3D pine-like Au–Cu NDs as transducing materials was employed to detect different concentrations of CA724. Fig. 8A shows the CVs of the proposed immunosensor for the detection of CA724 covering the concentration range from 0.5 pg mL⁻¹ to 500 ng mL⁻¹. A linear relationship between reduction peak current responses at −0.3 V of CV and the logarithmic values of CA724 concentration was obtained (Fig. 8B). When the concentration of CA724 increased at high concentration range, the decrease rate of current response slowed down, due to steric hindrance or saturation of couple antigen molecules.^29,30 Therefore, the current responses have a linear relationship with the logarithmic values of CA724 concentration. And the linear regression equation of the calibration curve was I = 231 − 42.7 log C with correlation coefficient of 0.99. The low detection limit of 0.25 pg mL⁻¹ was obtained, which was ascribed to the simple and novel signal production and amplification strategy of the designed label-free electrochemical immunosensor. The linear range and detection limit of the proposed immunosensor was compared with other reported label-free electrochemical immunosensors for the detection of other tumor markers in Table 1. It could be found that the proposed immunosensor showed a satisfying linear range and detection limit.

Fig. 8 (A) CVs of the proposed immunosensor for the detection of different concentrations of CA724: 0.5 pg mL⁻¹ (curve a), 5 pg mL⁻¹ (curve b), 50 pg mL⁻¹ (curve c), 500 pg mL⁻¹ (curve d), 5 ng mL⁻¹ (curve e), 50 ng mL⁻¹ (curve f) and 500 ng mL⁻¹ (curve g); (B) calibration curve of the label-free electrochemical immunosensor towards different concentrations of CA724. Error bar = RSD (n = 5).

Table 1 Comparison with other reported label-free electrochemical immunosensors for the detection of other tumor markers

Tumor marker	Linear range	Detection limit	Reference
Carcinoembryonic antigen	1.0 pg mL^−1–500 ng mL^−1	0.1 pg mL^−1	31
Human immunoglobulin G	1.0 pg mL^−1–320 ng mL^−1	0.1 pg mL^−1	32
Nuclear matrix protein 22	0.01–15 ng mL^−1	3.33 pg mL^−1	33
Breast cancer susceptibility gene	0.01–15 ng mL^−1	3.97 pg mL^−1	34
AFP	0.01–75 ng mL^−1	4 pg mL^−1	22
AFP	0.01–12 ng mL^−1	5 pg mL^−1	11
Zeranol	0.05–10 ng mL^−1	6 pg mL^−1	35
Tumor necrosis factor-alpha	0.02–34 ng mL^−1	10 pg mL^−1	36
Atrazine	0.05–0.5 ng mL^−1	16 pg mL^−1	37
AFP	0.1–100 ng mL^−1	60 pg mL^−1	18
CA724	0.5 pg mL^−1–500 ng mL^−1	0.25 pg mL^−1	This work

---

3.6. Reproducibility, selectivity and stability

To evaluate the reproducibility of the immunosensor, a series of five electrodes were prepared for the detection of 500 ng mL⁻¹ of CA724 (Fig. 9A). The relative standard deviation (RSD) of the measurements for the five electrodes was less than 5%, suggesting the precision and reproducibility of the proposed immunosensor was quite good.

Fig. 9 (A) Reduction peak current responses of the immunosensor fabricated on five different electrodes for the detection of 500 ng mL⁻¹ CA724; (B) reduction peak current responses of the immunosensor for the detection of 0.5 ng mL⁻¹ CA724 (1), 0.5 ng mL⁻¹ CA724 + 50 ng mL⁻¹ BSA (2), 0.5 ng mL⁻¹ CA724 + 50 ng mL⁻¹ glucose (3), 0.5 ng mL⁻¹ CA724 + 50 ng mL⁻¹ vitamin C (4) and 0.5 ng mL⁻¹ CA724 + 50 ng mL⁻¹ AFP (5). Error bar = RSD (n = 5).

To investigate the specificity of the fabricated immunosensor, interference study was performed by using BSA, glucose, vitamin C and AFP. The 0.5 ng mL⁻¹ of CA724 solution containing 50 ng mL⁻¹ of interfering substance was measured by the designed immunosensor (Fig. 9B). The current variation due to the interfering substance was less than 5% of that without interferences, indicating the selectivity of the immunosensor was acceptable.

To test the stability of the immunosensor, the immunosensor was stored at 4 °C when not in use. After one month, no apparent change of electrochemical current response was found for the detection of the same concentration of CA724. The good stability of the designed immunosensor can be ascribed to the good stability of electrodeposition method and the favorable chemistry characteristics towards the functionalization with biomolecules of Au in 3D pine-like Au–Cu NDs. The reproducibility, selectivity and stability of this immunosensor was acceptable, thus it was suitable for the determination of CA724 in real human serum samples.

3.7. Matrix effect, recovery and real sample analysis

In order to investigate the matrix effect,³⁸ interference experiments were carried out by adding the standard CA724 solution into the human serum samples. When the obtained experimental results were compared with the calibration curve obtained by the proposed immunosensor using samples prepared in PBS, no significant difference was observed. Therefore, the calibration curve shown in the Fig. 8B can be used to determine the recovery values. The low matrix effect could be explained by the specificity of the antibody–antigen interaction.

In order to demonstrate the application value of the proposed immunosensor in the clinical analysis and test the precision and accuracy of the proposed immunosensor, it was used to detect the recoveries after the addition of CA724 in human serum samples by standard addition methods (Table 2). The specific preparation process of the human serum sample was as follow: the human blood was first allowed to undisturbed for 24 h at 4 °C. Blood cells could be observed to sink to the lower layer, and then the upper layer was extracted and placed in a sterile tube. After centrifugation, the supernatant was the human serum sample, which was detected by the proposed immunosensor. The RSD was from 2.4% to 3.3% and the recovery was from 97.3% to 101.5%. Thus, the proposed immunosensor could be effectively applied to the determination of CA724 in human serums.

Table 2 Detection of CA724 in human serum samples with the proposed immunosensor

Initial concentration (ng mL^−1)	Added concentration (ng mL^−1)	Measured concentration (ng mL^−1)	RSD (%, n = 5)	Recovery (%)
5.21	5.00	9.91, 10.1, 10.2, 9.78, 10.4	2.4	97.3
	20.0	24.6, 25.7, 26.2, 24.8, 26.3	3.1	101.5
	45.0	49.2, 48.8, 52.3, 52.3, 50.3	3.3	100.8

---

In order to further validate the proposed immunosensor, a comparison with the commercialized available enzyme-linked immunosorbent assay (ELISA) method is shown in Table 3. The relative error between the two methods was in the range from −2.9% to 4.0%. These data revealed a good agreement between the two analytical methods, further indicating the feasibility of the proposed immunosensor for clinical application.

Table 3 Human serum sample analysis using the proposed method and the ELISA method

Sample	This method^a (ng mL^−1)	ELISA^a (ng mL^−1)	Relative error (%)
a Each value is the average of five measurements.
1	1.05	1.01	4.0
2	5.21	5.33	−2.3
3	10.12	10.42	−2.9

---

4. Conclusions

This work has developed a label-free electrochemical immunosensor, which used the electrodeposited 3D Au–Cu NDs as the signal producers and amplifiers simultaneously, for the quantitative detection of CA724. The designed immunosensor displayed a linear response with a wide range, a low detection limit, acceptable reproducibility, selectivity and stability. Therefore, this novel label-free electrochemical immunosensor might be expanded readily to detect other tumor markers and provide a promising potential in clinical application.

Acknowledgements

This study was supported by the Natural Science Foundation of China (nos 21175057, 21375047 and 21377046), the Science and Technology Development Plan of Shandong Province (no. 2014GSF120004), the Science and Technology Plan Project of Jinan (no. 201307010), Achievements Transformation of Shandong Province (2014ZZCX05101) and QW thanks the Special Foundation for Taishan Scholar Professorship of Shandong Province and UJN (no. ts20130937).

References

1. S. Zhang, B. Du, H. Li, X. Xin, H. Ma, D. Wu, L. Yan and Q. Wei, Biosens. Bioelectron., 2014, 52, 225–231.
2. H. V. Tran, R. Yougnia, S. Reisberg, B. Piro, N. Serradji, T. D. Nguyen, L. D. Tran, C. Z. Dong and M. C. Pham, Biosens. Bioelectron., 2012, 31, 62–68.
3. Z. Cui, Y. Cai, D. Wu, H. Yu, Y. Li, K. Mao, H. Wang, H. Fan, Q. Wei and B. Du, Electrochim. Acta, 2012, 69, 79–85.
4. R. Feng, Y. Zhang, H. Ma, D. Wu, H. Fan, H. Wang, H. Li, B. Du and Q. Wei, Electrochim. Acta, 2013, 97, 105–111.
5. R. C. Alves, F. B. Pimentel, H. P. A. Nouws, R. C. B. Marques, M. B. González-García, M. B. P. P. Oliveira and C. Delerue-Matos, Biosens. Bioelectron., 2015, 64, 19–24.
6. Y. Ding, D. Li, B. Li, K. Zhao, W. Du, J. Zheng and M. Yang, Biosens. Bioelectron., 2013, 48, 281–286.
7. Z. Yang, Y. Chai, R. Yuan, Y. Zhuo, Y. Li, J. Han and N. Liao, Sens. Actuators, B, 2014, 193, 461–466.
8. A. Guo, Y. Li, W. Cao, X. Meng, D. Wu, Q. Wei and B. Du, Biosens. Bioelectron., 2015, 63, 39–46.
9. S. Zang, Y. Liu, M. Lin, J. Kang, Y. Sun and H. Lei, Electrochim. Acta, 2013, 90, 246–253.
10. G. Shen, C. Cai and J. Yang, Electrochim. Acta, 2011, 56, 8272–8277.
11. T. Qi, J. Liao, Y. Li, J. Peng, W. Li, B. Chu, H. Li, Y. Wei and Z. Qian, Biosens. Bioelectron., 2014, 61, 245–250.
12. X. Wang, J. Miao, Q. Xia, K. Yang, X. Huang, W. Zhao and J. Shen, Electrochim. Acta, 2013, 112, 473–479.
13. H. D. Jang, S. K. Kim, H. Chang and J.-W. Choi, Biosens. Bioelectron., 2015, 63, 546–551.
14. W. Lu, J. Ge, L. Tao, X. Cao, J. Dong and W. Qian, Electrochim. Acta, 2014, 130, 335–343.
15. H. Li, J. He, S. Li and A. P. F. Turner, Biosens. Bioelectron., 2013, 43, 25–29.
16. S. Liu, R. Yuan, Y. Chai and H. Su, Sens. Actuators, B, 2011, 156, 388–394.
17. X. Jia, Z. Liu, N. Liu and Z. Ma, Biosens. Bioelectron., 2014, 53, 160–166.
18. J. Lin, Z. Wei, H. Zhang and M. Shao, Biosens. Bioelectron., 2013, 41, 342–347.
19. S. Yu, Q. Wei, B. Du, D. Wu, H. Li, L. Yan, H. Ma and Y. Zhang, Biosens. Bioelectron., 2013, 48, 224–229.
20. L. Yang, H. Zhao, S. Fan, S. Deng, Q. Lv, J. Lin and C.-P. Li, Biosens. Bioelectron., 2014, 57, 199–206.
21. L. Ding, A. M. Bond, J. Zhai and J. Zhang, Anal. Chim. Acta, 2013, 797, 1–12.
22. H. Wang, H. Li, Y. Zhang, Q. Wei, H. Ma, D. Wu, Y. Li, Y. Zhang and B. Du, Biosens. Bioelectron., 2014, 53, 305–309.
23. M. Zhang, J.-J. Lv, F.-F. Li, N. Bao, A.-J. Wang, J.-J. Feng and D.-L. Zhou, Electrochim. Acta, 2014, 123, 227–232.
24. R. Zhang, M. Hummelgård and H. Olin, Langmuir, 2009, 26, 5823–5828.
25. H. Fan, Z. Guo, L. Gao, Y. Zhang, D. Fan, G. Ji, B. Du and Q. Wei, Biosens. Bioelectron., 2015, 64, 51–56.
26. S. Sharma, R. Kamath and M. Madou, J. Anal. Appl. Pyrolysis, 2014, 108, 12–18.
27. P. Ningsih, C. Z. Holdsworth and S. W. Donne, Synth. Met., 2014, 196, 8–19.
28. S. Seghir, N. Stein, C. Boulanger and J. M. Lecuire, Electrochim. Acta, 2011, 56, 2740–2747.
29. K. Wang, J. Liao, X. Yang, M. Zhao, M. Chen, W. Yao, W. Tan and X. Lan, Biosens. Bioelectron., 2015, 63, 172–177.
30. J. Huang, G. Yang, W. Meng, L. Wu, A. Zhu and X. a. Jiao, Biosens. Bioelectron., 2010, 25, 1204–1211.
31. S. He, Q. Wang, Y. Yu, Q. Shi, L. Zhang and Z. Chen, Biosens. Bioelectron., 2015, 68, 462–467.
32. N. Wang, C. Gao, Y. Han, X. Huang, Y. Xu and X. Cao, J. Mater. Chem. B, 2015 10.1039/C4TB01881H.
33. H. Jia, P. Gao, H. Ma, D. Wu, B. Du and Q. Wei, Bioelectrochemistry, 2015, 101, 22–27.
34. H. Fan, Y. Zhang, D. Wu, H. Ma, X. Li, Y. Li, H. Wang, H. Li, B. Du and Q. Wei, Anal. Chim. Acta, 2013, 770, 62–67.
35. X. Xue, D. Wei, R. Feng, H. Wang, Q. Wei and B. Du, Anal. Methods, 2013, 5, 4159–4164.
36. S. Weng, M. Chen, C. Zhao, A. Liu, L. Lin, Q. Liu, J. Lin and X. Lin, Sens. Actuators, B, 2013, 184, 1–7.
37. X. Liu, W.-J. Li, L. Li, Y. Yang, L.-G. Mao and Z. Peng, Sens. Actuators, B, 2014, 191, 408–414.
38. A. Hayat, L. Barthelmebs and J. L. Marty, Anal. Chim. Acta, 2011, 690, 248–252.

---