Aptasensor for the simple detection of ochratoxin A based on side-by-side assembly of gold nanorods

Abstract

A new optical aptasensor is described based on the side-by-side assembly of gold nanorods (GNRs) for the one-step determination of ochratoxin A (OTA), which is a mycotoxin identified as a contaminant in grains and wine throughout the world. A DNA aptamer against OTA exhibiting high affinity and binding specificity has already been used as a biorecognition element to improve detection. Here, thiol-modified DNA was decorated on the side sites of GNRs acting as probes. A linker DNA containing aptamer sequences against OTA hybridized with the DNA decorated on the GNRs. It was shown that the GNRs assembled side-by-side through DNA hybridization in the absence of OTA, exhibiting strongly enhanced optical properties. However, in the presence of OTA, the GNRs dispersed as a result of specific aptamer–OTA recognition and conformational changes in the aptamer. The resulting changes in the absorption spectra of the GNRs were used for sensing. A linear correlation was demonstrated between the absorbance of the GNRs at 708 nm and the concentration of OTA over the range 0.5–20 ng mL⁻¹. The limit of detection was 0.22 ng mL⁻¹ (3σ), which meets the demand for detection of the allowed concentrations of OTA in foods according to European Commission Regulations (2 ng g⁻¹). The selectivity and feasibility of the developed OTA aptasensor for the detection of OTA in red wine samples was also evaluated and it was shown to be a simple, efficient and cost-effective one-step detection method. This method is a promising tool in toxin screening to guarantee food safety and to minimize potential risks to human health.

---

Introduction

Ochratoxin A (OTA), a mycotoxin identified as a food contaminant throughout the world, exists in a wide variety of products, such as cereals, dried fruits, nuts, coffee beans, beer and wines. Among the identified ochratoxins, OTA is the most prevalent and has the highest toxicity. It is considered to be a potential carcinogen (group 2B) by the International Agency for Research on Cancer.¹ The presence of OTA in agricultural products and foods may result in serious toxic effects, even at very low concentrations.² Therefore it is necessary to develop reliable approaches for the determination of OTA to guarantee food safety and to minimize potential risks to human and environmental health. Common analysis techniques – such as thin-layer chromatography, high-performance liquid chromatography and enzyme-linked immunosorbent assay – are well established and have been widely used over the past few years.³ However, more simple and cost-effective methods are still required.

Antibodies are widely used as affinity-based recognition elements for OTA screening, but they are both expensive and unstable.⁴ Aptamers have been adopted as an alternative approach to overcome the limitation of antibodies and have the advantages of low cost, high chemical stability, specificity and recognition affinity.⁵ DNA-based aptamers have been demonstrated as a promising tool for the purification and determination of OTA and a high binding affinity has been demonstrated.^6,7 Different types of aptamer-based biosensors (aptasensors) have been constructed for OTA detection.^8–13 Among these, a colorimetric aptasensor based on the salt-induced aggregation of gold nanoparticles is one of the simplest strategies for the rapid determination of OTA.¹⁴ OTA was directly detected using a one-step procedure. However, the reported detection limit (20 nM) was limited and needs further enhancement.

Gold nanorods (GNRs) have received much attention as a result of their geometric anisotropy and unique optical properties.^15,16 GNRs exhibit tunable longitudinal surface plasmon resonance (SPR) depending on their aspect ratio.¹⁷ Biosensors based on the aggregation or competitive dispersion of GNRs have been developed in previous research.^18,19 However, unordered aggregates of anisotropic GNRs have limitations with respect to repeatability when designed for practical sensing applications. Fortunately, GNRs can be assembled into superstructures in solution with controllable and reproducible end-to-end or side-by-side oriented configurations.¹⁵ These assemblies present strongly enhanced SPR responses in the visible and near-infrared regions compared with individual nanorods and unordered aggregates, induced by simultaneous contributions from collective interparticle coupling.¹⁷ The SPR spectra change in peak wavelength and intensity depending on the number and shape of the assemblies at fixed orientations and distances.^20,21 Therefore controllable GNR assemblies achieved through specific interactions, including biotin–streptavidin binding,^22,23 DNA hybridization,^24–26 cysteine–lead,²⁷ oligonucleotide–mercury,^28,29 and antibody–antigen recognition,^21,30 offer great potential in nanoscale optical sensing for research and practical purposes.

There have been several biosensor studies based on end-to-end or side-by-side GNR assemblies involving aptamer-specific recognition events.^31,32 DNA aptamers against thrombin and adenosine have been reported. Aptamers immobilized on nanorod surfaces may, however, lead to an insufficient binding affinity when used in thrombin detection. Small amounts of end-to-end assemblies are formed at low concentrations of thrombin, limiting further sensing applications. In the adenosine example, side-by-side assemblies resulting from electrostatic interactions were affected by matrix interferences and their application to real complex samples may be limited. Although aptamer-target recognition induced GNR assemblies are potentially promising for use as aptasensors, it remains challenging to design an appropriate sensor construction to give an excellent sensing performance.

In this work, taking advantage of specific OTA aptamers and controllable GNR assemblies, we developed a new optical aptasensor based on side-by-side GNR assembly for the one-step simple determination of OTA. Free aptamers were designed as linker DNA dispersed in aqueous solution. GNRs decorated with complementary DNA on the side sites acted as probes. A simple one-step detection procedure was achieved for rapid OTA screening. To our knowledge, this is the first aptasensor for the detection of OTA based on GNR oriented assemblies. The absorption intensity of the GNRs at a fixed wavelength was used as a sensing indicator, which could be easily obtained in the laboratory without sophisticated and expensive equipment. The analytical performance of the established approach (sensitivity and feasibility) was evaluated. Compared with other methods, this is simple, easy to operate and cost-effective. It is a promising technique for practical use in mycotoxin screening of agricultural products and food.

Experimental

Reagents and materials

Hydrogen tetrachloroaurate(III) hydrate (HAuCl₄·3H₂O), sodium borohydride (NaBH₄), silver nitrate (AgNO₃), hexadecyltrimethylammonium bromide (CTAB), L-ascorbic acid, dithiothreitol (DTT), thiol-functionalized polyethylene glycol (PEG-thiol), tris(2-carboxyethyl)phosphine (TCEP), OTA, ochratoxin B (OTB) and aflatoxin B₁/M₁(AFB₁/AFM₁) were obtained from Sigma-Aldrich. T-2 toxin, deoxynivalenol (DON) and zearalenone (ZON) were purchased from Fermentek (Jerusalem, Israel). All other chemical reagents used were of analytical grade. Ultrapure water with a resistivity of 18.2 MΩ cm was used in the experiments (Millipore, Billericay, USA).

All the DNA samples used in this work were obtained from Sangon Biotech (Shanghai, China) Co. Ltd; detailed DNA sequences and modifications are listed in Table 1. The sequence of OTA aptamer (underlined) was designed according to previously reported work by J. A. Cruz-Aguado and G. Penner.^6,7 A random DNA sequence (without the OTA aptamer) was designed as a negative control oligonucleotide.

Table 1 DNA sequences and modifications used in aptasensor fabrication

DNA	Sequence
Linker DNA	5′-GTCTCACTGTGACTCTCG̲A̲T̲C̲G̲G̲G̲T̲G̲T̲G̲G̲G̲T̲G̲G̲C̲G̲T̲A̲A̲A̲G̲G̲G̲A̲G̲C̲A̲T̲C̲G̲G̲A̲C̲A̲-3′
Random DNA	5′-CTAGCCCACACCCACCGCATTTCC-3′
3′-Thiolated DNA	5′-AGTCACAGTGAGACTTTTTTT-C(6)-SH-3′
5′-Thiolated DNA	5′-SH-C(6)-TTTTTTTCACCCGATCGAG-3′

---

Instruments and measurements

Transmission electron microscopy (TEM) images of GNRs were obtained using a JEM-1230 transmission electron microscope (JEOL, Tokyo, Japan). A 10 μL GNR droplet was evaporated on a 300 mesh copper grid with a carbon film at room temperature. At least 200 particles in the TEM photos were examined to characterize the size of the nanorods. UV-vis absorption spectra were recorded using a SpectraMax M5 microplate reader with Softmax Pro Software (Molecular Devices, Sunnyvale, USA). The scanning wavelength of the absorption spectra was set from 400 to 1000 nm. The GNRs were diluted to a volume of 100 μL and then added to each well to obtain the absorption spectra. Dynamic light scattering (DLS) tests were performed using a Zetasizer Nano ZS90 instrument (Malvern Instruments, Malvern, UK) with a 633 nm laser source and a detector angle of 25 °C. The sample was diluted using ultrapure water before the DLS measurements and measured an average of three times.

GNR synthesis

The GNRs were synthesized chemically using a seed-mediated growth method according to previous reports.^33–35 The preparations of gold seeds were obtained by quickly mixing 5 mL of 0.5 mM HAuCl₄ and 5 mL of 0.2 M CTAB with 0.6 mL of freshly prepared 10 mM NaBH₄, stirring vigorously for 3 min and then aging at 25 °C. To grow the GNRs, 70 μL of 0.079 M ascorbic acid were mixed with 5 mL of 1 mM HAuCl₄, 5 mL of 0.2 M CTAB and 0.15 mL of 4 mM AgNO₃, then 12 μL of gold seeds were added and left at 30 °C for 2 h. The synthesized GNRs were purified by centrifugation (10 000 rpm, 15 min) at room temperature and dissolved in 2 mL of 5 mM CTAB. The concentration of synthesized GNRs was calculated according to previously reported work.¹⁸

Fabrication of GNR–DNA probes

Thiolated DNA was decorated on the side surfaces of the GNRs to fabricate the GNR probes based on previously reported procedures.²⁴ It was achieved based on one fact – that is, the CTAB attached to the end sites of the GNRs was much easier to replace than the CTAB on the side sites.³⁶ A 0.5 mL aliquot of concentrated GNRs was modified with DTT on the end sites for 8 h at a molar ratio of 10 : 1 and stabilized using PEG-thiol at a molar ratio of 100 : 1 for 3 h. After centrifugation and purification, TCEP-treated 3′ and 5′-thiolated DNA (see ESI†) was added and reacted for at least 24 h at a molar ratio of 1000 : 1 at room temperature. The GNR-3′-DNA and GNR-5′-DNA probes were centrifuged (5000 rpm, 10 min) and stored in 5 mM CTAB.

Optimization of linker DNA concentration

To obtain side-by-side GNR assemblies and to study the optimum concentration of linker DNA, 25 μL of GNR-5′-DNA probes and 25 μL of GNR-3′-DNA-probes were mixed with 4 μL of linker DNA (for pretreatment, see ESI†) at different concentrations (resulting in final concentrations of 0, 2, 5, 10, 20, 50, 100, 200, 300 and 400 nM) in ultrapure water by continuous oscillation. Then the mixture was added to 50 μL of 1 mM Tris–HCl buffer (pH 8.5) containing 10 mM MgCl₂, 100 mM NaCl and 0.02% sodium dodecylsulfate.³⁷ The well-mixed reaction solution was then incubated at 45 °C for 1 h before using for the TEM, UV-vis absorption spectrometry and DLS measurements. As a control, random DNA samples with different concentrations were also used in this test, replacing the linker DNA.

OTA detection

To detect OTA in Tris–HCl buffer, 25 μL of GNR-5′-DNA probes and GNR-3′-DNA-probes were mixed with 50 μL of OTA standard solutions of different concentrations (0, 0.5, 1, 2, 5, 10 and 20 ng mL⁻¹) and then added to ultrapure water with 4 μL of linker DNA, resulting in a final concentration of 100 nM. The reaction solution was stirred and incubated at 45 °C for 1 h and then used for measurements.

To evaluate the selectivity of the aptasensor, GNR-5′–DNA, GNR-3′–DNA probes and linker DNA were mixed and reacted in the presence of other toxic mycotoxins common in agricultural products and food, including OTB, AFB₁, AFM₁, T-2 toxin, DON and ZON. The concentration of mycotoxin utilized was 10 ng mL⁻¹ in Tris–HCl buffer (pH 8.5).

Pretreatment of wine samples

The red wine samples used in our experiment were purchased from local supermarkets. The wine samples were prepared according to previously reported procedures.^38,39 The red wine was initially spiked with the OTA stock solution to obtain a final concentration of 100 ng mL⁻¹. Several dilutions (concentrations of 2, 5, 10 and 20 ng mL⁻¹) were then prepared with non-spiked red wine, followed by a 2 h equilibration before extraction. The spiked red wine samples were then extracted via a two-step procedure using toluene and chloroform and mixed with the same volume of Tris–HCl buffer (pH 8.5). After complete phase separation, a measured amount was taken out from the top phase and used for detection.

Results and discussion

Illustration of OTA aptasensor

A schematic illustration of OTA detection is shown in Fig. 1. In the absence of OTA, GNR assemblies were observed after reaction because the two types of DNA decorated on the side sites of the GNRs could hybridize with the complementary linker DNA. However, in the presence of the OTA target, the GNRs on the opposite side dispersed or disassembled in solution. The conformational change of the OTA aptamer at the 3′ end of the linker DNA for specific recognition of the OTA target will result in difficulties in DNA hybridization, consequently leading to difficulties in GNR assembly.

Fig. 1 Schematic illustration of OTA detection using the aptasensor based on side-by-side assembly of GNRs.

The OTA-sensitive GNR assemblies consist of three components: two DNA-decorated GNR probes and a linker DNA. It was inspired by work on aptamer-linked nanoparticle aggregates and core–shell nanorod dimmers.^40,41 The 3′-thiolated and 5′-thiolated DNA is attached to the side sites of GNRs at the 3′ and 5′ ends, respectively. Thymine at the 3′ or 5′ end of the thiolated DNA is designed to reduce the effects of steric hindrance and to improve the flexibility for oligonucleotide hybridization.⁴² The linker DNA can be divided into three segments. The first and second segments hybridized with 3′-thiolated DNA and three nucleotides of 5′-thiolated DNA, respectively. The third segment of the OTA aptamer hybridized with the other nine nucleotides of 5′-thiolated DNA. In the presence of the OTA molecule, the third segment of the linker DNA changed in conformation for specific recognition. Thus the three-DNA base pairs were not stable and were also not strong enough to hold the GNR-5′-DNA probe, leading to dispersed GNRs.

Characterization of GNR–DNA probes

The geometrical and optical properties of the GNR probes were characterized by TEM, UV-vis absorption spectrometry and DLS. The synthesized GNRs were well dispersed in solution, had an aspect ratio of 3.1 (average length 62 nm and average diameter 20 nm) based on the TEM image in Fig. 2. The calculated concentration of the GNRs was 1.25 nM, which was significant for the sensor design and optimization of the GNR assembly.

Fig. 2 TEM micrograph of synthesized GNRs and UV-vis absorption spectra of GNRs before and after DNA decoration.

The UV-vis absorption spectra of the GNRs and GNR–DNA probes in 5 mM CTAB are shown in Fig. 2. A blue shift of 4 nm in the longitudinal peak (from 712 to 708 nm) and a small red shift of 1 nm (from 510 to 511 nm) in the transverse peak were observed after DNA decoration. This indicated a decrease in the aspect ratio of the nanorods after DNA modification,⁴³ indicating the successful decoration of thiolated DNA on the side sites of the GNRs. The spectral intensity and peak shape of the GNR–DNA probes showed little change after DNA functionalization, demonstrating that dispersed GNR probes were obtained without aggregation.⁴⁴ The average hydrodynamic size of the GNRs determined by DLS increased from 65.8 to 78.8 and 76.5 nm after DNA modification, illustrating the successful construction of GNR–3′-DNA and GNR–5′-DNA probes. The average number of DNA loading per nanorod was calculated as 251 with 6-carboxyfluorescein and thiol dual-labelled DNA following the same fabrication procedure (see ESI†). Thus a high overall yield of GNR–DNA probes was demonstrated.

Optimization of GNR assembly

The decorated 3′-thiolated DNA and 5′-thiolated DNA on the side sites of the GNR probes were supposed to hybridize with linker DNA under the right conditions according to our design, resulting in the side-by-side assembly of nanorods. Thus adding the correct amount of linker DNA was critical for hybridization and assembly. Different final concentrations of linker DNA from 0 to 400 nM were initially used to obtain the assemblies and were analysed to determine the optimum conditions. The absorption spectra of Fig. 3 show that a red shift in the transverse peak and a blue shift in the longitudinal peak of the GNRs were observed when a large amount of linker DNA (400 nM) was added, accompanied by a significant decrease in the absorption intensity of the longitudinal peak. In contrast, there was no significant change in the absorption intensity of the UV-vis spectra when different amounts of random DNA were added. It was demonstrated that the GNR assemblies were fabricated in Tris buffer induced by the hybridization of linker DNA and thiolated DNA on the GNR probes because the absorption spectra were in accordance with previous reports.^21,27

Fig. 3 UV-vis absorption spectra of GNR-DNA probes after reacting with 0 and 400 nM linker DNA; UV-vis absorption intensity of GNRs at 708 nm after reaction with different amounts of linker DNA and random DNA.

The absorption intensity of the GNRs at 708 nm (longitudinal peak wavelength of the GNR–DNA probes) decreased initially (0–100 nM) and then increased slowly (>100 nM) with increasing concentrations of linker DNA (Fig. 3). This indicated that the configuration and yield of the GNR side-by-side assemblies significantly depended on the concentration of linker DNA. As reported EI-Sayed and co-workers,⁴³ there is a linear relationship both experimentally and theoretically between the absorption maximum of the longitudinal SPR peak and the mean aspect ratio of the GNRs. The particle concentration in the buffer also influences the absorption intensity.⁴⁴ Both these variables contribute to the observed change in the absorption spectra. With increasing concentrations of linker DNA (0–100 nM), the number and size of the GNR assemblies increased and reached saturation gradually. When abundant linker DNA was added (>100 nM) and simultaneously mixed with the GNR probes, excess linker DNA hybridized with the 3′ and 5′-thiolated DNA on the side sites of the GNR probes. It is more difficult for the linker DNA to hybridize with the two types of GNR probes at the same time, which resulted in a slight increase in the number of dispersed GNRs. The mean aspect ratio of the GNRs and the particle concentration in the buffer initially decreased as a result of the increasing side-by-side GNR assembly, and then increased as a result of the increasingly dispersed GNRs. Thus linker DNA with an optimum concentration of 100 nM was used for further quantitative analysis.

The GNR side-by-side assemblies could also be directly characterized using TEM images. Fig. 4 shows the variation in structure of the side-by-side assembly of nanorods reacting with 5 and 100 nM linker DNA. Large GNR ladders were observed when adding to 100 nM linker DNA. The hydrodynamic diameters determined by DLS measurements also significantly changed after adding and reacting with different amounts of linker DNA. Fig. S1† shows that the average diameter of the GNR assemblies induced by 100 nM of linker DNA was 342 nm. However, the average hydrodynamic diameter of the assemblies induced by 5 nM linker DNA was only 106 nm based on TEM and spectral analysis.

Fig. 4 Representative TEM images of GNR side-by-side assemblies added to 5 and 100 nM linker DNA.

Analytical performance of proposed aptasensor

Fig. 5 shows the representative TEM images of GNR probes after reaction with increasing concentrations of OTA (0, 2, 10 and 20 ng mL⁻¹) and 100 nM of linker DNA. In the presence of OTA molecules (2 ng mL⁻¹), a small number of GNR ladders or dispersed GNRs were observed in solution. With increasing amounts of OTA (10 and 20 ng mL⁻¹), more dispersed GNRs were obtained. The hydrodynamic diameters of the GNR probes after reaction and the number of nanorods in the assemblies determined by statistical analysis are shown in Fig. S2 and S3,† respectively. It was demonstrated that the OTA molecules were specifically recognized by the aptamer in Tris–HCl buffer (pH 8.5). The linker DNA was not strong enough to hold the GNR–5′-DNA probe, leading to a decreased number of GNR ladders. More GNRs thus dispersed in solution, with increased recognition of the aptamer–OTA and decreased interparticle DNA hybridization.

Fig. 5 Representative TEM images of GNRs after reaction with different concentrations of OTA (a–d: 0, 2, 10 and 20 ng mL⁻¹) and 100 nM linker DNA.

Fig. 6 shows that the absorption spectra of the GNRs varied with various OTA concentrations in the range 0–20 ng mL⁻¹ as a result of the changes in configuration and amount of GNR assemblies. It is clearly shown that, with decreasing OTA concentration, the longitudinal peak of the GNRs first blue-shifted and then red-shifted, accompanied by a decrease in absorption intensity. A significant broadening of the longitudinal peak and another SPR peak at wavelengths >900 nm was also observed when the OTA concentration was <5 ng mL⁻¹. This result showed that larger side-by-side assemblies were fabricated with lower concentrations of OTA. The absorption intensity, especially the fixed wavelength at 708 nm, was used for further quantitative analysis.

Fig. 6 UV-vis absorption spectra and calibration graph of GNR assemblies with different concentrations of OTA standard solution (0, 0.5, 1, 2, 5, 10 and 20 ng mL⁻¹ or ppb).

The sensitivity of the developed optical aptasensor based on GNR side-by-side assemblies was determined based on the 3σ rule. The data were linearly correlated (Abs = 0.025C_OTA + 0.37) with a correlation coefficient of 0.98 in the concentration range 0.5–20 ng mL⁻¹ (Fig. 6). The limit of detection of this approach was calculated as 0.22 ng mL⁻¹ (0.5 nM) at the subnanomolar level, which meets the need for the detection of the amounts of OTA allowed in foods for retail sale based on European Commission standards (2 ng g⁻¹)^45,46 and standards in China (5 ng g⁻¹).⁴⁷

To evaluate the selectivity of the established OTA aptasensor, other common toxic mycotoxins in agricultural products and food, including OTB, AFB₁, AFM₁, T-2 toxin, DON and ZON, were tested using proposed aptasensor at concentrations of 10 ng mL⁻¹. The change in absorption intensity of the GNRs at 708 nm between other the mycotoxins and a negative sample (Tris–HCl buffer) after reaction was much smaller than that for OTA (≤9.7% of the OTA response) (Fig. 7). Thus it was demonstrated that the optical aptasensor based on GNR side-by-side assemblies had a good selectivity towards OTA detection.

Fig. 7 Selectivity evaluation of fabricated aptasensor with other mycotoxins.

Detection of OTA in red wine samples

The proposed aptasensor was applied to the determination of OTA in red wine samples to evaluate its feasibility. Wine samples were purchased from a local supermarket and spiked with different concentration of OTA (2, 5, 10 and 20 ng mL⁻¹) and then extracted using a two-step procedure. A non-artificial contaminated red wine sample was used as a blank. According to the results in Table 2, the recoveries of OTA in red wine samples were in the range 92.0–118.2%. This evaluation of the developed aptasensor indicated that it has promise in practical use for the detection of OTA in real wine samples.

Table 2 Determination of OTA in red wine samples^a

Sample	Spiked (ng mL^−1)	Mean ± SD detected (ng mL^−1)	Recovery (%)
a SD = standard deviation.
Sample 1	2.0	2.364 ± 0.304	118.2
Sample 2	5.0	5.205 ± 0.532	104.1
Sample 3	10.0	9.204 ± 0.876	92.0
Sample 4	20.0	18.644 ± 1.130	93.2

---

Conclusions

GNRs were successfully assembled side-by-side to develop a new optical aptasensor for the determination of OTA. Linker DNA containing aptamer sequences specifically recognized the OTA molecules. Otherwise the linker DNA hybridized with compromised DNA on the side sites of the nanorods, leading to an oriented GNR assembly. OTA could be detected at low concentrations in the linear range 0.5–20 ng mL⁻¹ based on the absorption intensity of the UV-vis spectra induced by conformational and number changes in the GNR assemblies. The obtained subnanomolar LOD was lower than the maximum food safety limit and was achieved without the need for expensive and sophisticated equipment or highly qualified operators. The proposed aptasensor based on GNR assemblies is simple, efficient, selective and cost-effective with a one-step detection procedure; it holds great potential for the determination of OTA in red wine. It is a promising alternative toxin screening strategy to guarantee food safety.

Acknowledgements

The authors thank the 985-Institute of Agrobiology and Environmental Sciences at Zhejiang University for the use of their facilities and equipment. This work was financially supported by the National Nature Science Foundation of China (No. 31401572), the China Postdoctoral Science Foundation (Grant No. 2014M551750 and 2015T80617) and the Public Welfare Technology Applied Research Project of Zhejiang Province (No. 2014C32002).

References

1. A. Pfohl-Leszkowicz and R. A. Manderville, Mol. Nutr. Food Res., 2007, 51, 61–99.
2. E. Petzinger and K. Ziegler, J. Vet. Pharmacol. Ther., 2000, 23, 91–98.
3. C. M. Maragos, Toxin Rev., 2004, 23, 317–344.
4. A. Rhouati, C. Yang, A. Hayat and J.-L. Marty, Toxins, 2013, 5, 1988–2008.
5. S. Song, L. Wang, J. Li, C. Fan and J. Zhao, TrAC, Trends Anal. Chem., 2008, 27, 108–117.
6. J. A. Cruz-Aguado and G. Penner, J. Am. Chem. Soc., 2008, 56, 10456–10461.
7. J. A. Cruz-Aguado and G. Penner, Anal. Chem., 2008, 80, 8853–8855.
8. C. Yang, V. Lates, B. Prieto-Simón, J.-L. Marty and X. Yang, Talanta, 2013, 116, 520–526.
9. L. Jiang, J. Qian, X. Yang, Y. Yan, Q. Liu, K. Wang and K. Wang, Anal. Chim. Acta, 2014, 806, 128–135.
10. T. H. Ha, Toxins, 2015, 7, 5276–5300.
11. J. H. Soh, Y. Lin, S. Rana, J. Y. Ying and M. M. Stevens, Anal. Chem., 2015, 87, 7644–7652.
12. D. Afzali, F. Fathirad and S. Ghaseminezhad, J. Food Sci. Technol., 2016, 53, 909–914.
13. E.-J. Jo, H. Mun, S.-J. Kim, W.-B. Shim and M.-G. Kim, Food Chem., 2016, 194, 1102–1107.
14. C. Yang, Y. Wang, J.-L. Marty and X. Yang, Biosens. Bioelectron., 2011, 26, 2724–2727.
15. L. Vigderman, B. P. Khanal and E. R. Zubarev, Adv. Mater., 2012, 24, 4811–4841.
16. M. Grzelczak, J. Perez-Juste, P. Mulvaney and L. M. Liz-Marzan, Chem. Soc. Rev., 2008, 37, 1783–1791.
17. H. Chen, L. Shao, Q. Li and J. Wang, Chem. Soc. Rev., 2013, 42, 2679–2724.
18. X. Xu, X. Liu, Y. Li and Y. Ying, Biosens. Bioelectron., 2013, 47, 361–367.
19. X. Xu, Y. Ying and Y. Li, Sens. Actuators, B, 2012, 175, 194–200.
20. L. Wang, L. Xu, H. Kuang, C. Xu and N. A. Kotov, Acc. Chem. Res., 2012, 45, 1916–1926.
21. L. Wang, Y. Zhu, L. Xu, W. Chen, H. Kuang, L. Liu, A. Agarwal, C. Xu and N. A. Kotov, Angew. Chem., Int. Ed., 2010, 49, 5472–5475.
22. K. Caswell, J. N. Wilson, U. H. Bunz and C. J. Murphy, J. Am. Chem. Soc., 2003, 125, 13914–13915.
23. K. G. Thomas, S. Barazzouk, B. I. Ipe, S. S. Joseph and P. V. Kamat, J. Phys. Chem. B, 2004, 108, 13066–13068.
24. W. Ma, H. Kuang, L. Xu, L. Ding, C. Xu, L. Wang and N. A. Kotov, Nat. Commun., 2013, 4, 2689.
25. S. Pal, Z. Deng, H. Wang, S. Zou, Y. Liu and H. Yan, J. Am. Chem. Soc., 2011, 133, 17606–17609.
26. Z. Li, Z. Zhu, W. Liu, Y. Zhou, B. Han, Y. Gao and Z. Tang, J. Am. Chem. Soc., 2012, 134, 3322–3325.
27. H.-H. Cai, D. Lin, J. Wang, P.-H. Yang and J. Cai, Sens. Actuators, B, 2014, 196, 252–259.
28. Y. Wang, Y. F. Li, J. Wang, Y. Sang and C. Z. Huang, Chem. Commun., 2010, 46, 1332–1334.
29. Y. Zhu, L. Xu, W. Ma, Z. Xu, H. Kuang, L. Wang and C. Xu, Chem. Commun., 2012, 48, 11889–11891.
30. J.-Y. Chang, H. Wu, H. Chen, Y.-C. Ling and W. Tan, Chem. Commun., 2005, 1092–1094.
31. S. J. Zhen, C. Z. Huang, J. Wang and Y. F. Li, J. Phys. Chem. C, 2009, 113, 21543–21547.
32. J. Wang, P. Zhang, J. Y. Li, L. Q. Chen, C. Z. Huang and Y. F. Li, Analyst, 2010, 135, 2826–2831.
33. M. Liu and P. Guyot-Sionnest, J. Phys. Chem. B, 2005, 109, 22192–22200.
34. B. Nikoobakht and M. A. El-Sayed, Chem. Mater., 2003, 15, 1957–1962.
35. T. K. Sau and C. J. Murphy, J. Am. Chem. Soc., 2004, 126, 8648–8649.
36. C. J. Murphy, T. K. San, A. M. Gole, C. J. Orendorff, J. X. Gao, L. Gou, S. E. Hunyadi and T. Li, J. Phys. Chem. B, 2005, 109, 13857–13870.
37. X. Geng, D. Zhang, H. Wang and Q. Zhao, Anal. Bioanal. Chem., 2013, 405, 2443–2449.
38. Z. Zhu, M. Feng, L. Zuo, Z. Zhu, F. Wang, L. Chen, J. Li, G. Shan and S.-Z. Luo, Biosens. Bioelectron., 2015, 65, 320–326.
39. C. Yang, V. Lates, B. Prieto-Simón, J.-L. Marty and X. Yang, Biosens. Bioelectron., 2012, 32, 208–212.
40. J. Liu and Y. Lu, Nat. Protoc., 2006, 1, 246–252.
41. L. Tang, S. Li, F. Han, L. Liu, L. Xu, W. Ma, H. Kuang, A. Li, L. Wang and C. Xu, Biosens. Bioelectron., 2015, 71, 7–12.
42. L. Xu, H. Kuang, C. Xu, W. Ma, L. Wang and N. A. Kotov, J. Am. Chem. Soc., 2012, 134, 1699–1709.
43. S. Link, M. Mohamed and M. El-Sayed, J. Phys. Chem. B, 1999, 103, 3073–3077.
44. C. Yu and J. Irudayaraj, Anal. Chem., 2007, 79, 572–579.
45. Commission of the European Communities, Off. J. Eur. Communities: Legis., 2006, 364, 5–24.
46. European Commission, Off. J. Eur. Communities: Legis., 2012, 176, 43–45.
47. J. N. Selvaraj, Y. Wang, L. Zhou, Y. Zhao, F. Xing, X. Dai and Y. Liu, Food Addit. Contam., Part A, 2015, 32, 440–452.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra04439e

---