Highly sensitive detection of sulfadimidine in water and dairy products by means of an evanescent wave optical biosensor

Abstract

During recent years, there has been an increasing demand for the development of rapid biosensing technologies that could be performed outside the laboratory, for example on farms, near rivers, in food collection stations and store houses or in food production plants. Therefore, cost-effective and automatic detection methods are promising for onsite residual analysis in food and environmental monitoring. In this work, we propose an automatic, rapid, highly sensitive and reusable planar waveguide evanescent wave immunosensor (PWEI) for onsite determination of sulfadimidine (SM2) in water and dairy products. The PWEI is based on an indirect inhibition immunoassay that takes place at an optical transducer chip chemically modified with an analyte derivative. Fluorescence produced by labeled antibodies bound to the transducer is excited by the evanescent wave formed on the transducer surface and detected by photodiodes through a lock-in amplifier, which is inversely correlated with the analyte concentration. Each test cycle is completed automatically in 15 min. The optical transducer chip of the PWEI modified with the analyte derivative is robust and features high reusability, which allows for regeneration over 300 times without sensitivity loss. Under the optimized conditions, the dose–response curve established for SM2 shows a low detection limit of 0.06 μg L⁻¹. The 50% inhibition concentration is 1.39 ± 0.08 μg L⁻¹ with a linear working range from 0.19 μg L⁻¹ to 10.10 μg L⁻¹. The cross-reactivity towards organic compounds structurally similar to SM2 is negligible. The recoveries of SM2 in a variety of dairy products and natural water range from 80% to 107%. The PWEI features portable dimensions of 42 cm × 50 cm × 24 cm (length × width × height) and shows great prospects for the onsite measurement of SM2, when used in combination with the appropriate pretreatment.

---

1. Introduction

Sulfonamides (SAs) are commonly used for therapeutic and prophylactic purposes in animals,¹ and also as additives in animal feed, due to their low cost and effectiveness as growth promoters.² Due to the toxicity of SAs and their abuse in practice, strict maximum residue levels (MRL) have been established, e.g. 25 μg L⁻¹ in milk set by the Codex Committee of FAO/WHO³ and 100 μg L⁻¹ set by the European Council.^4–6 Sulfadimidine (SM2), also known as sulfamethazine, is an SA which has been widely used for the treatment and control of inflammation associated with Bordetella bronchiseptica infection in animals by feed medication at a high level.⁷ However, SM2 is also the most common antimicrobial contaminant in animal feed, generating potentially serious problems in human health, such as allergic or toxic reactions.¹ Therefore, methods to detect its abuse and residual levels in various water samples^8–10 and animal food products^11,12 have been increasingly in demand in recent years.

Traditional methods for the determination of SM2 in food and environmental samples include high-performance liquid chromatography (HPLC),^13,14 gas chromatography (GC)¹⁵ and liquid chromatography-mass spectrometry (LC/MS).^16,17 These methods are accurate, sensitive and specific; however, they are also labor-intensive and expensive, and need sophisticated instrumentation. During recent years, there has been an increasing demand for the development of rapid biosensing technologies that could be performed outside the laboratory, for example on farms, near rivers, in food collection stations and store houses or in food production plants. Therefore, cost-effective and automatic detection methods are promising for onsite abuse and residual analysis in food and environmental monitoring. Biosensors show remarkable advantages, including high specificity even in complex matrices and the potential to become cost-effective, portable and easy-to-use test devices.¹⁸ For example, surface plasmon resonance (SPR)¹⁹ and the fluorescence polarization immunoassay (FPIA)²⁰ have been reported as alternatives to traditional methods for determination of SAs with advantages including simple detection procedures, quick response and real-time monitoring. However, the sensitivities of SPR and FPIA are not high enough for trace SM2 detection in aqueous samples. Moreover, SPR in real applications is limited to the regeneration of the receptor surface, which must at least be washed or even entirely replaced between analyses of different samples.²¹ In this work, we propose an automatic and compact planar waveguide evanescent wave immunosensor (PWEI) to realize the rapid, highly sensitive, reusable and onsite determination of SM2 in food and environmental monitoring. The principle of the PWEI is based on an immunoassay and evanescent wave formed on the surface of the planar waveguide due to the total internal reflection propagation of incident light. As far as we know, this is the first work reporting the application of a PWEI to determine SM2. Based on the developed PWEI platform, in this work the method for SM2 detection is optimized and fully validated in terms of linearity, accuracy, precision, recovery and specificity.

2. Materials and methods

2.1 Materials

Potassium dihydrogen phosphate, sodium phosphate dibasic, sodium chloride, potassium chloride, hydrochloric acid, sodium dodecyl sulphate (SDS), toluene and ethyl alcohol were purchased from Sinopharm Chemical Reagent Co. Ltd. 3-Mercaptopropyl-trimethoxysilane (MTS), N-(4-maleimidobutyryloxy) succinimide (GMBS), bovine serum albumin (BSA), sulfadimidine (SM2), sulfadiazine (SDZ), sulfamerazine (SM1) and sulfamethoxazole (SMX) were purchased from Sigma-Aldrich and stored at 4 °C. Cy5.5 and N-hydroxysuccinimide (NHS) ester were obtained from GE Healthcare Life Sciences. Monoclonal anti-sulfadimidine antibody and the analyte derivative–hapten conjugate of BSA-SM2 were purchased from Shijiazhuang Solarpex Biotechnology Co. Ltd. Labeling of the SM2-antibody with Cy5.5 was performed according to the method described by Mujumdar.²² All chemicals were analytical grade if not specified and used as received without further purification.

1000 mg L⁻¹ SM2 stock solution was prepared using methanol and stored at 4 °C before use. Phosphate buffered saline (10 mM PBS, pH 7.4) was prepared using DI water (18.2 MΩ cm). The SM2 stock solution was diluted to a series of concentration levels using the 10 mM PBS buffer solution.

2.2 PWEI platform

The planar waveguide evanescent wave immunosensor used in this study has been described in detail in our previous report²³ and is presented in Fig. 1. Briefly, the pulse laser beam from a 635 nm pulse diode laser was directly coupled into one beveled edge face of a planar waveguide transducer and propagated along the transducer via total internal reflection. The evanescent wave generated at the surface of the waveguide interacted with the surface-bound fluorescently labeled target conjugate, and caused excitation of the fluorophores. The emitted fluorescence was collected by the high-numerical-aperture polymer fibers (NA = 0.46) located beneath the waveguide and opposite to the chemically modified surface, and subsequently filtered by means of a band pass filter and detected by photodiodes through a lock-in amplifier.

Fig. 1 (A) Schematic set-up of planar waveguide evanescent wave immunosensor (PWEI) platform and (B) instrument photograph of PWEI.

BK7 glass was adopted as the planar waveguide transducer with a refractive index of 1.515 and a size of 60 mm × 15 mm in area and 1.5 mm in depth. One end face of the BK7 glass was polished and beveled to 45° for light coupling. A thin layer of SiO₂ film (60 nm thickness) was coated on the waveguide chip surface by the chemical vapour deposition method (Foxconn Nanotechnology Research Center, Beijing). The planar waveguide was embedded into a rectangular Teflon flow cell with a size of 42 mm × 2.0 mm in area and 50 μm in depth. All reagents were delivered by a flow delivery system operated with one peristaltic injection pump, a six-way valve and three one-way valves for the liquid switch. A 1 mL regeneration solution loop was designed to store the regeneration solution. A 1 mL pre-incubation loop was kept at 37 °C for incubating the mixture of the test sample and the Cy5.5-labeled antibody solution. The fluid delivery system, data acquisition and processing were automatically controlled by the built-in computer. As shown in Fig. 1B, the instrument size of the PWEI was 42 cm × 50 cm × 24 cm (length × width × height).

2.3 Surface chemical modification of the waveguide chip

To make the waveguide chip biosensitive to the target, the SM2 derivative (SM2–BSA conjugate) was immobilized covalently on the chip surface by the following steps. The waveguide chip was first cleaned with detergent and subsequently immersed in piranha solution (H₂SO₄ : H₂O₂ = 3 : 1, v/v) for one hour, then rinsed with DI water and dried in nitrogen gas. Fig. 2 shows the procedure for the chemical modification of SM2–BSA onto the chip surface. Silanization of the chip surface was achieved by immersion in a 2% (v/v) MTS toluene solution for 2 h at room temperature. The silanized surface was rinsed with toluene and dried using nitrogen gas. To immobilize SM2–BSA onto the surface of the thiol-silanized chip, the waveguide chip was first immersed in 2 mM GMBS (in ethanol) solution for 1 h at room temperature, and subsequently washed thoroughly with ethanol and DI water. Then, the SM2–BSA conjugate was applied on the specific binding site of the chip surface overnight at 4 °C. Finally, the surface was shielded from non-specific binding by immersing the chip in BSA (2 mg mL⁻¹) for 1 h. The modified chip was stored at 4 °C before use.

Fig. 2 Schematic diagram of the immobilization of the hapten conjugate SM2–BSA onto the waveguide chip surface.

2.4 Immunoassay procedures

All the measurements were performed using a binding inhibition test format. The SM2 derivative (SM2–BSA) was immobilized on the chip surface and competed with the SM2 analyte in the test samples to bind to the Cy5.5-labeled antibody. To optimize the Cy5.5-labeled antibody concentration, the binding inhibition immunoassay was carried out at antibody concentrations of 0.4 μg mL⁻¹, 0.6 μg mL⁻¹ and 0.8 μg mL⁻¹, respectively. Firstly, 800 μL analyte standard solution was mixed with 200 μL Cy5.5-labeled monoclonal antibody solution and pre-incubated in the loop at 37 °C for 4 min. After the binding reaction reached an equilibrium state, the mixture was pumped into the flow cell at a constant flow rate of 200 μL min⁻¹. Only the unbound antibodies were able to bind to the analyte derivative covalently bound on the waveguide surface. The binding process lasted 5 min, after which the excitation laser was turned off to avoid strong photobleaching of the Cy5.5 dyes. After that, the waveguide chip was rinsed with 10 mM PBS buffer to remove the residual unbound Cy5.5-labeled antibodies on the surface. Then, the laser was turned on to generate an evanescent wave on the chip surface, which was used to excite the fluorescence of the bound Cy5.5 dyes. The fluorescence signal was recorded and was found to be inversely related to the analyte concentration because the number of free antibodies able to bind to the surface was reduced. Finally, a regeneration process was performed by rinsing the chip surface with SDS solution (0.5%, pH 1.9) for 5 min, leaving it ready for a new test cycle.

All of the immunoassay processes were conducted automatically by the control system of the PWEI embedded into a built-in computer. The total time for one test cycle was less than 15 min.

2.5 Data analysis

All sample fluorescence signals were normalized with the signal corresponding to a blank, i.e. the signal obtained in the absence of analyte. The standard curve for SM2 detection was plotted against the logarithm of the concentration of SM2 ranging from 0.001–1000 μg L⁻¹ through a five-parameter logistic model as follows:^18,24
where [Ag] was the SM2 concentration; SS was the signal strength of the optical immunosensor; A₁ and A₂ were the maximum (blank signal, x → 0) and minimum signal (background signal, x → ∞) of the titration curve, respectively; [Ag₀] was the SM2 concentration at the midpoint or inflection point (IC₅₀); and p was the slope of the tangent at the inflection point. Error bars indicated in the curve represent the relative standard deviations for three individual experiments.

2.6 Selectivity and recovery

In the selectivity experiment, three structural analogues, SDZ, SM1 and SMX, were chosen instead of SM2. The interference chemicals were processed in the same way as the SM2 standard solutions. The relative 50% inhibition values of the cross-reactivity (CR) were used to judge the selectivity of the sensing system via the following formula:²⁵

CR(%) = [IC₅₀(SM2)/IC₅₀(structural analogue)] × 100%

The dairy samples including liquid milk, yoghurt and baby formula milk were bought from the local supermarket. For solid baby formula milk, 4 g of sample was firstly dissolved in 20 mL 10 mM PBS buffer at 80 °C for 5 min. 2 mL dissolved formula milk (or raw liquid milk/yoghurt), 3 mL acetonitrile and 15 mL 10% trichloroacetic acid were added into a centrifuge tube and then spiked with SM2 standard solutions to give final concentrations at two levels (1 μg L⁻¹ and 5 μg L⁻¹). The mixture was centrifuged at 12 000 rpm for 5 min to precipitate proteins and dissolve organic substances. Subsequently, 200 μL of the supernatant sample was diluted to 10 mL with 10 mM PBS buffer for PWEI detection. Three types of water sample, including bottled water, lake water and wastewater, were also chosen for the recovery experiments. No pretreatment of the bottled water was needed. Lake water and wastewater were filtered through a 0.22 μm membrane filter and subsequently stored at 4 °C for testing. Samples were measured within 24 h of collection to avoid biodegradation. The calibration curve for the buffer solution was adopted to calculate the spiked SM2 concentrations in the recovery experiment, because interference caused by the pretreated milk matrix was negligible.

3. Results and discussion

3.1 Performance of PWEI for SM2 detection

The temporal fluorescence responses for different concentrations of SM2 in a test cycle are recorded in Fig. 3. An observable decrease of the PWEI fluorescence response to SM2 was observed, even when the SM2 concentration was decreased to 0.01 μg L⁻¹. When the SM2 concentration was increased to 1000 μg L⁻¹, a slight fluorescence signal was observed, indicating that non-specific adsorption on the chip surface was negligible due to the immobilized BSA acting as a shielding agent. This result was further confirmed by the negligible fluorescence signal due to the input of 10 μg mL⁻¹ Cy5.5-labeled BSA solution.

Fig. 3 Temporal fluorescence responses for different concentrations of SM2 at 0.4 μg mL⁻¹ Cy5.5-labeled SM2–antibody and for Cy5.5-labeled BSA without SM2 in a test cycle.

The antibody concentration is an important factor in immunoassays as it strongly affects the detection limits and working ranges.^24,26,27 Therefore, the concentration of Cy5.5-labeled SM2–antibody was optimized to confirm the calibration curve with the highest sensitivity. Immunoassay determinations of SM2 were carried out using three Cy5.5-labeled SM2-antibody concentration levels of 0.4 μg mL⁻¹, 0.6 μg mL⁻¹ and 0.8 μg mL⁻¹. The detection limit (LOD) of the PWEI was calculated from the calibration curve as being the analyte concentration that provided a 10% decrease of the blank signal.²⁸ The dynamic detection range of the test was defined by the analyte concentrations causing 20% and 80% inhibition of the maximum fluorescence signal.

The comparison between the standard curves for the three antibody concentrations is shown in Fig. 4. The linearly quantitative SM2 detection ranges of the PWEI were 0.19–10.10 μg L⁻¹ at 0.4 μg mL⁻¹ Cy5.5-labeled antibody, 0.28–6.42 μg L⁻¹ at 0.6 μg mL⁻¹ Cy5.5-labeled antibody and 0.48–7.98 μg L⁻¹ at 0.8 μg mL⁻¹ Cy5.5-labeled antibody, as described by inhibitory concentrations of 20–80%. The LOD values were determined to be 0.06 μg L⁻¹ at 0.4 μg mL⁻¹ Cy5.5-labeled antibody, 0.13 μg L⁻¹ at 0.6 μg mL⁻¹ Cy5.5-labeled antibody and 0.22 μg L⁻¹ at 0.8 μg mL⁻¹ Cy5.5-labeled antibody. The Cy5.5-labeled antibody concentration of 0.4 μg mL⁻¹ gave the highest detection sensitivity, and also resulted in reduced reagent costs; therefore, it was adopted for the subsequent experiments. Compared with other methods reported previously, such as LC-MS (8.2 μg L⁻¹),¹⁷ ELISA (4.3 μg L⁻¹),²⁹ SPR (0.28 μg L⁻¹),¹⁹ and FPIA (10 μg L⁻¹),²⁰ the proposed method showed a superior sensitivity for SM2 detection with the lowest corresponding LOD, which also meets the strict MRL of 25 μg L⁻¹ SAs in milk samples set by the FAO/WHO.⁶

Fig. 4 Standard curves at three different concentrations of Cy5.5-labeled antibody of 0.4 μg mL⁻¹, 0.6 μg mL⁻¹ and 0.8 μg mL⁻¹. Error bars represent the standard deviations for three individual experiments.

3.2 Regeneration and stability

The regeneration and stability of the biosensing surface element are important factors that may limit the reliability of an immunosensor.⁵ Strong acid or alkaline buffer solution can be used to break the binding between the antibody and the analyte derivative covalently immobilized on the waveguide chip surface.³⁰ In our system, 0.5% SDS solution (pH 1.9) was pumped into the flow cell as the regeneration agent at a flow rate of 1 mL min⁻¹. The robustness of the PWEI method was checked by monitoring the fluorescence response of the same waveguide chip towards the same test sample over 50 test cycles (Fig. 5). The results showed that 50 repeated test cycles gave a relative standard deviation of 3.0%, indicating that a satisfactory regeneration performance was achieved by using 0.5% SDS solution (pH 1.9). Moreover, the waveguide chip was reused up to six times during one month, running approximately 300 test cycles without significant sensitivity loss.

Fig. 5 Regeneration of fluorescence signals towards 0.4 μg mL⁻¹ Cy5.5-labeled SM2–antibody on the PWEI platform.

3.3 Specificity

Specificity studies were performed by investigating the dose–response curves with three structural analogues, SDZ, SM1 and SMX, as shown in Fig. 6. Comparing the IC₅₀ values for the target and the cross-reacting compounds potentially present in the sample is a typical method to estimate the cross-reactivity of a technique.³¹ The IC₅₀ values were calculated to be 1.39 μg L⁻¹, >10 000 μg L⁻¹, >10 000 μg L⁻¹ and 48.21 μg L⁻¹ for SM2, SMX, SDZ and SM1, respectively, as shown in Table 1, which corresponded to the CR values of 100%, <0.01%, <0.01% and 2.74%, respectively.

Fig. 6 Standard curves and cross-reactivity of immunoassay on the PWEI platform towards SM2, SMX, SDZ and SM1 in 10 mM PBS buffer solution.

Table 1 Structures and cross-reactivity of PWEI towards SM2 and other structural analogues

Compounds	Common structure
	R =	IC50 (μg L^−1)	CR (%)
Sulfadimidine (SM2)		1.39	100
Sulfamethoxazole (SMX)		>10 000	<0.01
Sulfadiazine (SDZ)		>10 000	<0.01
Sulfamerazine (SM1)		48.21	2.74

---

Molecular modeling studies on the structures of SAs have provided valuable insight into the principles of the cross-reactivity characteristics of anti-SM2 antibodies. As reported in previous studies,^32,33 the molecules of SA antibiotics have a characteristic bend around the tetrahedral –SO₂– grouping (see the common structure of SA antibiotics in Table 1), which is the part responsible for the recognition of SA antibiotics by group-specific antibodies. However, the cross-reactivity of anti-SM2 antibodies towards SMX and SDZ was negligible with CR values of <0.01%, which were in agreement with previous studies.^11,34,35 The results indicated that the characteristic bend around the tetrahedral –SO₂– grouping was not the binding site of the antibody adopted on the PWEI platform. Moreover, a slight cross-reactivity towards SM1 was observed with a CR value of 2.74%. A possible reason for this was that the adopted antibody specifically recognized the similar R-group in the structures of SM2 and SM1, i.e., there was only a methyl group difference between the two molecules. However, the interference from SM1 should be negligible in most cases because the interfering response signal towards SM1 is equal to the target signal only when the concentration of SM1 present is 35-fold higher than that of SM2. In summary, the CR values show that the selectivity of the PWEI platform is convincing even in a complicated matrix containing other SA antibiotics.

3.4 Application to real samples

In order to investigate the accuracy of the proposed method, the PWEI was applied to the measurement of SM2 in real test samples, including dairy products and water samples. In the recovery experiment, the fluorescence signals were almost the same as that responding to the buffer solution before the test samples were spiked with SM2. The detection performance of the PWEI system for the samples spiked with SM2 is presented in Table 2. The results showed that the recoveries by means of the PWEI platform ranged from 80–107%. Satisfactory variations were also demonstrated with a relative standard deviation of less than 17%. These results confirm that the proposed PWEI has the ability to be applied to the detection of SM2 in real aqueous samples.

Table 2 Recovery of PWEI towards SM2 in dairy and water samples (n = 3, mean ± SD)

Samples	Spiked (μg L^−1)	Found (μg L^−1)	Recovery (%)	RSD (%)
Liquid milk	1.00	0.93 ± 0.09	93	10
	5.00	4.71 ± 0.47	94	10
Yoghurt	1.00	1.02 ± 0.08	102	8
	5.00	4.15 ± 0.08	83	9
Baby formula	1.00	1.01 ± 0.10	101	10
	5.00	4.92 ± 0.20	98	4
Bottled water	1.00	1.07 ± 0.05	107	5
	5.00	5.19 ± 0.87	104	17
Lake water	1.00	0.80 ± 0.03	80	4
	5.00	4.40 ± 0.62	88	14
Wastewater	1.00	1.03 ± 0.04	103	4
	5.00	4.19 ± 0.91	84	2

---

4. Conclusion and outlook

In this work, we proposed an automatic, rapid, highly sensitive and reusable immunosensor (i.e. PWEI) to determine SM2 in food and environmental monitoring. The analytical performance of the PWEI was confirmed by the highly sensitive detection of SM2 with an IC₅₀ of 1.39 ± 0.08 μg L⁻¹ at a Cy5.5-labeled antibody concentration of 0.4 μg mL⁻¹. The detection limit of 0.06 μg L⁻¹ satisfied the requirements for monitoring of trace SM2 in dairy products and a variety of water samples. The whole test cycle was automatically completed in 15 min. The cross-reactivity towards organic compounds structurally similar to SM2 was negligible in most cases. Results for real spiked dairy and water samples showed satisfactory recovery ratios ranging from 80% to 107%. Moreover, the PWEI features portable dimensions of 42 cm × 50 cm × 24 cm (length × width × height). All the above-mentioned results show great prospects for the application of the PWEI for highly-sensitive, rapid and onsite detection of SM2, once the appropriate pretreatment can be supplied. Through adopting other antibodies, the proposed PWEI is a universal platform for the detection of other pollutants in both environmental aqueous samples and milk products. It also offers the ability to simultaneously detect multiple trace pollutants in test samples, such as by covalently immobilizing variable antigen derivatives at the different total reflection points on the waveguide chip surface, or by using different sized quantum dots to label different antibodies for recognizing different trace pollutants.

Acknowledgements

This research is supported by the Major Scientific Equipment Development Project of China (2012YQ030111), Beijing Nova Program (Z141109001814078) and China Postdoctoral Science Foundation (2013M541868).

References

1. S. Wang, H. Y. Zhang, L. Wang, Z. J. Duan and I. Kennedy, Food Addit. Contam., 2006, 23, 362–384.
2. A. Gentili, D. Perret and S. Marchese, TrAC, Trends Anal. Chem., 2005, 24, 704–733.
3. N. F. Fresenius, Fresenius’ J. Anal. Chem., 1999, 364, 270–272.
4. Codex Alimentarius, Bulletin of Ministry of Agriculture of the Slovak Republic, 1996, XXVII: part 14, pp. 271–295.
5. Council Regulation (EEC), Community procedure for the establishment of maximum residue limits of veterinary medicinal products in food stuffs of animal origin, 1990, vol. 224, pp. 1–8.
6. FAO/WHO, A second meeting of the joint FAO/WHO expert committee on food additives, Code of Federal Regulations, 1992, vol. 21: p. 365.
7. Analytical Methods Committee, Royal Society of Chemistry, Analyst, 1992, 117, 817.
8. R. Hirsch, T. Ternes, K. Haberer and K. L. Kratz, Sci. Total Environ., 1999, 225, 109–118.
9. K. G. Karthikeyan and M. T. Meyer, Sci. Total Environ., 2006, 361, 196–207.
10. R. H. Lindberg, P. Wennberg, M. I. Johansson, M. Tysklind and B. A. V. Andersson, Environ. Sci. Technol., 2005, 39, 3421–3429.
11. Y. Xu, J. Ding, H. Y. Chen, Q. Zhao, J. Hou, J. Yan, H. Wang, L. Ding and N. Q. Ren, Food Chem., 2013, 140, 83–90.
12. A. A. Mehtabuddin, T. Ahmad, S. Nadeem, Z. I. Tanveer and J. Arshad, J. Anim. Plant Sci., 2012, 22, 473–478.
13. N. Furuaswa, Biomed. Chromatogr., 2001, 15, 235–239.
14. K. Pietruszka and A. Posyniak, Bull. Vet. Inst. Pulawy, 2010, 393–396.
15. A. Cannavan, S. A. Hewitt, W. J. Blanchflower and D. G. Kenned, Analyst, 1996, 121, 1457–1461.
16. G. Bartolucci, G. Pieraccini, F. Villanelli, G. Moneti and A. Triolo, Rapid Commun. Mass Spectrom., 2000, 14, 967–973.
17. D. H. Kim, J. O. Choi, J. Kim and D. W. Lee, J. Liq. Chromatogr. Relat. Technol., 2003, 26, 1149–1159.
18. C. Barzen, A. Brecht and G. Gauglitz, Biosens. Bioelectron., 2002, 17, 289–295.
19. A. Sternesjo, C. Mellgren and L. Bjorck, Anal. Biochem., 1995, 226, 175–181.
20. S. X. Zhang, Z. H. Wang, I. S. Nesterenko, S. A. Eremin and J. Z. Shen, Int. J. Food Sci. Technol., 2007, 42, 36–44.
21. S. X. Zhang, Z. H. Wang, I. S. Nesterenko, S. A. Eremin and J. Z. Shen, Int. J. Food Sci. Technol., 2007, 42, 36–44.
22. S. R. Mujumdar, R. B. Mujumdar, C. M. Grant and A. S. Waggoner, Bioconjugate Chem., 1996, 7, 356–362.
23. X. H. Zhou, B. D. Song, H. C. Shi, L. H. Liu, H. L. Guo and M. He, Sens. Actuators, B, 2014, 198, 150–156.
24. J. W. Sheng, M. He, H. C. Shi and Y. Qian, Anal. Chim. Acta, 2006, 572, 309–315.
25. K. M. Maxey, K. R. Maddipati and J. Birkmeier, J. Clin. Immunoassay, 1992, 15, 116–120.
26. P. Hua, J. P. Hole and J. S. Wilkinson, Opt. Express, 2005, 13, 1124–1130.
27. E. Mauriz, A. Calle, L. M. Lechuga, J. Quintana, A. Montoya and J. J. Manclus, Anal. Chim. Acta, 2006, 561, 40–47.
28. F. Long, H. C. Shi, M. He, J. W. Sheng and J. F. Wang, Anal. Chim. Acta, 2009, 649, 123–127.
29. I. S. Nesterenko, M. A. Nokel and S. A. Eremin, J. Anal. Chem., 2009, 64, 435–462.
30. F. Long, H. C. Shi, M. He and A. N. Zhu, Biosens. Bioelectron., 2008, 23, 1361–1366.
31. M. Franek, V. Kolar, A. Deng and S. Crooks, Food Agric. Immunol., 1999, 11, 339–349.
32. P. Cliquet, E. Cox, W. Haasnoot, E. Schacht and B. M. Goddeeris, J. Agric. Food Chem., 2003, 51, 5835–5842.
33. C. A. Spinks, G. M. Wyatt, H. A. Lee and M. R. A. Morgan, Bioconjugate Chem., 1999, 10, 583–588.
34. X. W. Li, G. Zhang, Q. T. Liu and C. H. Feng, et. al., Food Addit. Contam., Part A, 2009, 26, 314–325.
35. P. Bjurling, G. A. Baxter, M. Caselunghe, C. Jonson, M. O’Connor, B. Persson and C. T. Elliott, Analyst, 2000, 125, 1771–1774.

---