Introduction

Mycotoxins are the most important naturally occurring contaminants relevant to public health and trade, occurring frequently in foods, feedstuffs, and traditional Chinese medicines (TCMs) [1, 2] to cause illness or human and animal death, which has been a universal yet serious issue all over the world. AFB1, the secondary metabolite mainly produced by Aspergillus fungi, is one of the most toxic mycotoxins because of its highly hepatotoxic, oncogenic, teratogenic, mutagenic, and immunotoxic properties in human and domestic animals [1, 3,4,5,6]. It has been classified as a group 1A carcinogen to humans [7], and can be found in a wide range of matrices at trace levels [8,9,10]. Many countries have stipulated maximum admissible levels of AFB1 at μg/kg or ppb level [11, 12]. Therefore, reliable methods for sensitive determination of AFB1 are in urgent demand to meet safety requirements, monitor exposure risk, and minimize regulatory and trade losses.

Apart from traditional analytical tools for mycotoxin detection [13,14,15], development of a simple, easy-to-operate, and cost-effective biosensing strategy for sensitive AFB1 detection of mycotoxins is highly desirable. Recent biosensor technologies exhibit exceptional performance capabilities [16,17,18,19] like simplicity, high specificity and sensitivity, low cost, compact size, and continuous real-time analysis. Shao et al. developed a sensitive graphene oxide and detachable nanoladders–based fluorescence turn-off biosensor, showing a limit of detection (LOD) of 0.1 nM for ochratoxin A (OTA) [20]. Similarly, Zhu et al. developed a two-way colorimetric biosensor based on unmodified gold nanoparticles and a switchable double-stranded DNA concatemer for OTA detection [21]. In addition to conventional colorimetric immunosensors, electrochemiluminescence (ECL) immunosensors as an alternative and promising solution has received special attention in many fields owing to its wide dynamic concentration response range and spatial control for real-time and green measurements in miniaturized formats [22, 23]. Nevertheless, no studies have reported label-free ECL immunosensors for sensitive detection of AFB1 in complex matrices like those of TCMs. Through the measurement of electrical signals arising from a light emission process in a redox reaction of electrogenerated reactants between the electrode and electrolyte, label-free ECL immunosensors allow for the development of simple, cheap, and disposable devices with no labels, attracting considerable interest [24, 25]. Taking these advantages into account, we have developed a label-free ECL immunosensor for AFB1 detection in lotus seeds.

Unquestionably, new luminescent materials for signal amplification platform are in continuously increasing need for ECL immunoassays. In recent decades, various promising nanomaterials with outstanding physicochemical properties or electrical conductivity have been introduced as the luminophores for signal amplification to decrease the detection limit of immunosensors [26, 27]. Among them, semiconductor quantum dots (QDs) have been the predominant candidate of interest as they exhibit excellent tunable photoluminescence, superior biocompatibility, high ECL performance, and physical/chemical stability along with size-dependent emission spectra and narrow emission peaks [28,29,30,31,32]. Various kinds of QDs have been developed for the fabrication of ECL immunosensors, but, pure QDs are found to produce lower ECL signals than conventional luminescent materials [33]. To resolve this problem, core/shell heterostructures could be used to improve photostability and facilitate charge separation [34]. In addition, the use of capping agents will improve the surface state of QDs and enhance ECL emission. As a preferred capping agent, wide band gap ZnS can effectively eliminate the effect of surface defects on the luminescence efficiency of QDs. Thus, core-shell ZnCdS@ZnS QDs were chosen as the signal sources to obtain strong, stable ECL intensity for detection of AFB1.

In order to construct a stable and sensitive ECL immunosensor, self-assembled monolayers are an easy and effective way to steadily assemble the core-shell QDs on the surface of the Au electrode. As a perfluorosulfonated polymer with outstanding film-forming properties, Nafion is commonly used to assemble inorganic or biological materials to improve the mechanical stability of the modified electrode [35, 36] and has been used successfully to assemble some conventional luminescent materials [36, 37] for sensitive target detection by ECL methods. Therefore, Nafion was used to attach core-shell ZnCdS@ZnS QDs on the desired surface of the Au electrode.

Here, a simple and easy-to-construct ZnCdS@ZnS QDs–based label-free ECL immunosensor with fewer steps of fabrication and higher sensitivity than previously reported sensors was proposed for the specific determination of AFB1 in complex TCMs through an antigen-antibody (Ag–Ab) reaction, offering a preferable alternative to heavy and large instrument–based methods. Testing in real lotus seed samples confirmed the superiority of the newly developed ECL immunosensor for rapid, highly efficient, and sensitive detection of mycotoxins at trace levels in complex matrices, providing a powerful and universal analytical platform for quantitative detection of a variety of analytes in food and TCM safety testing, environmental pollution monitoring, and disease diagnostics.

Experimental section

Materials and chemicals

N-Hydroxysuccinimide (NHS), 1-ethyl-3-(3-dimethylamino-prophl) carbodiimide hydrochloride (EDC), potassium ferricyanide (K3Fe(CN)6), sulfuric acid, and K2S2O8 were all purchased from Shanghai Macklin Biochemical Technology Co., Ltd. (www.macklin.bioon.com, Shanghai, China). ZnCdS@ZnS QDs was obtained from Xingzi New Material Technology Development Co., Ltd. (www.xznanomat.com, Shanghai, China). 5% Nafion was purchased from Shanghai Hesen Electric Co., Ltd. (www.hesen.cn, Shanghai, China). Twenty-five micrograms per milliliter of aflatoxin B1 (AFB1) standard solution was supplied by Pribolab Company (www.pribolab.cn, Singapore) and kept at − 20 °C in the dark. Murine monoclonal antibodies of AFB1 (anti-AFB1 Abs, 1.0 mg/mL, stored in PBS solution) was acquired from Shandong Lvdu Bio—Science & Technology Co., Ltd. (www.lvdu.net, Shandong, China). A 0.1 M phosphate-buffered saline (PBS) (pH 7.4) was prepared by mixing solutions of 0.1 M Na2HPO4 and 0.1 M NaH2PO4 and then adjusting the pH with 0.1 M NaOH. The 0.01 M PBST (pH 7.4) obtained from supplementing 0.01 M PBS with 0.05% Tween-20 as the washing buffer and 0.01 M PBS was purchased from Solarbio Science & Technology Co., Ltd. (www.solarbio.com, Beijing, China). All the chemicals were of analytical grade. All solutions were prepared with ultra-pure water from a Milli-Q purification system.

Preparation of standard and sample solutions

The AFB1 standard solution (25 μg/mL) was diluted with PBS (0.01 M, pH 7.4) to obtain different concentrations of working solutions, which were stored in the dark at − 20 °C.

Lotus seeds were crushed and sifted through 40 meshes, and 1 g powder was carefully weighed and added to 5 mL of 80% methanol solution with vortexing for 5 min. The mixed solution was then extracted on a mixing oscillator for 30 min, followed by centrifugation at 6000 rpm for 10 min. Afterwards, the supernatant was collected and diluted with 0.01 M phosphate-buffered saline (pH 7.4) solution for analysis.

Fabrication and detection principle of the ECL immunosensor

The fabrication procedure of the ECL immunosensor is listed in Scheme 1A. First, an Au electrode with a 2-mm diameter was serially polished with 1.0, 0.3, and 0.05 μm alumina slurries to obtain a mirror-like surface, then rinsed with deionized water, methanol, and deionized water, respectively. Before modification, the bare Au electrode was scanned in 5 mmol/L [Fe(CN)6]3−/4− solution containing 0.04 M KNO3 between − 0.1 and + 0.5 V, then again in 0.5 M H2SO4 between + 0.3 and + 1.5 V, respectively, until reproducible signals were obtained. Then, 4 μL of ZnCdS@ZnS QDs (0.5 mg/mL) solution was dropped onto the surface of the Au electrode and allowed to dry at room temperature. Next, 3 μL of 1% Nafion solution was dropped onto the modified electrode and dried at room temperature for 30 min to form a stable film. The modified Au/QDs/Nafion electrode was then immersed in 50 μL of 0.2 M EDC + 0.1 M NHS solution as cross-linker for 20 min at 37 °C. After rinsing with PBST (pH 7.4), the electrode was incubated in 50 μL of 25 μg/mL of anti-AFB1 antibody for 15 min at 25 °C. Finally, after rinsing with PBST (pH 7.4), the resulting electrode was used to fabricate the ECL signal probe and incubated in 50 μL of AFB1 (Ag) standard solution with different concentrations for 30 min at 25 °C for ECL immunosensor detection. The prepared electrode was washed thoroughly with PBST (pH 7.4).

Scheme 1
scheme 1

(A) Fabrication and (B) detection principle of the proposed QDs-based ECL immunosensor for detection of AFB1

Full size image

ECL detection was performed in a detector cell with 10 mL 0.1 M PBS (pH 7.4,) solution containing 0.1 M K2S2O8 and 0.1 M KCl, and the detection principle is shown in Scheme 1B. The applied potential varied from − 0.7 to − 1.6 V at a scanning rate of 0.1 V/s and the voltage of the photomultiplier tube (PMT) was set to 800 V. K2S2O8 was used as a co-reactant in the electrolyte. In the three-electrode system, when S2O82− and QDs both obtained electrons from the working electrode, QDs could generate QDs* and produce ECL emission via electron transfer. The conductivity of the modified electrode decreased as AFB1 concentration increased due to steric resistance and nonconductivity of AFB1, thus decreasing ECL intensity.

Results and discussion

Characterization of ZnCdS@ZnS QDs

The microstructure and morphology of the ZnCdS@ZnS QDs were characterized first by transmission electron microscopy (TEM), photoluminescence (PL), and UV-Vis investigation. The TEM images in Fig. 1A illustrate that the ZnCdS@ZnS QDs exhibited smooth and uniform surfaces with good dispersibility. The average particle size of ZnCdS@ZnS QDs was approximately 8.5 nm; a promising surface area capable of hosting large amounts of antibodies to obtain ultra-sensitive ECL immunosensor probes. The PL (Fig. 1B(a)) spectrum of the ZnCdS@ZnS QDs showed a good profile along with a strong emission peak at 465 nm and the UV-Vis (Fig. 1B(b)) spectrum indicated a wide absorption band, implying their promising fluorescence properties.

Fig. 1
figure 1

(A) TEM image, and (B) PL ((a) PL spectrum of the ZnCdS@ZnS QDs showed a good profile along with a strong emission peak at 465 nm) and UV-Vis (inset, (b) UV-Vis spectrum indicated a wide absorption band, implying their promising fluorescence properties) spectra of ZnCdS@ZnS QDs

Full size image

Electrochemical and ECL behavior of ZnCdS@ZnS QDs

The electrochemical and ECL behaviors of ZnCdS@ZnS QDs were characterized next. The cyclic voltammogram of QDs in Fig. 2a showed that when the Au/QDs electrode was immersed in 0.1 M PBS buffer solution containing 10.1 M K2S2O8 and 0.1 M KCl for scanning at potentials from 0 to − 1.6 V, two independent reduction peaks occurred around − 0.68 and − 1.17 V, results of the reduction of S2O82− and QDs, respectively. These indicate that both S2O82− and QDs could be reduced without interference for the highly efficient generation of strong ECL signal from QDs. After continuous scanning for > 5 cycles over more than 90 s, the ZnCdS@ZnS QDs on the surface of Au electrode remained stable and capable of extremely strong ECL signal (about 18,000 a.u.) under voltage of − 0.7 to − 1.6 V in Fig. 2b, demonstrating the promising ECL performance and outstanding stability of the ZnCdS@ZnS QDs for real-world use.

Fig. 2
figure 2

(a) Cyclic voltammogram and (b) ECL signal of ZnCdS@ZnS QDs in 0.1 M PBS buffer solution containing 0.1 M K2S2O8 and 0.1 M KCl

Full size image

In this work, during the negative potential scan, the ECL behavior of the modified Au/QDs electrode was generated by the reduction reaction between S2O82− in the electrolyte and QDs on the electrode. The reduction of co-reactant S2O82− produced a strong oxidant SO4•-, while the QDs assembled on the electrode were reduced to QDs•-. Then, SO4•- could react with electron-injected QDs•- to produce the excited state QDs* through electron transfer. Light was generated when the excited state QDs* come back to the ground-state QDs, and the ECL emission of ZnCdS@ZnS QDs was measured and listed as follows:

$$ {\displaystyle \begin{array}{l}{\mathrm{S}}_2{{\mathrm{O}}_8}^{2-}+\mathrm{ne}-\to {{\mathrm{S}\mathrm{O}}_4}^{2-}+{\mathrm{S}\mathrm{O}}_4\bullet -\\ {}\mathrm{QDs}+\mathrm{ne}-\to n\mathrm{QDs}\bullet -\\ {}{\mathrm{S}\mathrm{O}}_4\bullet -+\mathrm{QDs}\bullet -\to \mathrm{QDs}\ast +{{\mathrm{S}\mathrm{O}}_4}^{2-}\\ {}\mathrm{QDs}\ast \to \mathrm{QDs}+\mathrm{hv}\end{array}} $$

ECL characteristics of the developed immunosensor

To verify the feasibility of the current immunosensor, the ECL behaviors of the sensor after each modification step were characterized. As shown in Fig. 3, weak or no ECL signal (curve a) was observed from the bare Au electrode. The assembly of ZnCdS@ZnS QDs onto the surface of the bare Au electrode resulted in significant enhancement of the ECL signal (curve b). After modification of the Au/QDs electrode with 1% Nafion to form an Au/QDs/Nafion composite film, the ECL signal exhibited a slight change (curve c).

Fig. 3
figure 3

ECL profile of (a) bare Au, (b) Au/QDs, (c) Au/QDs/Nafion, (d) Au/QDs/Nafion/anti-AFB1 Ab, and (e) Au/QDs/Nafion/anti-AFB1 Ab/Ag electrodes

Full size image

However, the conjunction of anti-AFB1 Ab led to a noteworthy decrease in ECL intensity (curve d), suggesting that the anti-AFB1 Ab could act as an electron transfer blocking layer to prevent diffusion of S2O82− toward the electrode surface. Finally, when the Ag or target AFB1 in tested samples was assembled on the modified Au/QDs/Nafion/anti-AFB1 Ab electrode, the ECL signal trended toward a decrease again, and the resulting ECL signal was suitable for sensitive quantitation of AFB1. These findings are in agreement with the EIS observations of the electrode across stepwise modification, both of which indicate that the ECL immunosensor was indeed fabricated as expected.

Analytical performance and sensitivity of the ECL immunosensor

Under the optimal experimental conditions, the ECL intensity of the described label-free ECL immunosensor response to each concentration (c) of AFB1 was measured, and the corresponding ΔECL-c standard calibration curve following the bonding of various amounts of AFB1 (Ag) onto the Au/QDs/Nafion/anti-AFB1 Ab electrode is demonstrated in Fig. 4.

Fig. 4
figure 4

Calibration ΔECL-c curve of the proposed ECL immunosensor. Error bars were calculated of relative standard deviation from triple parallel experiments (n = 3)

Full size image

The ΔECL values increased linearly accordingly with increasing AFB1 concentrations in the range of 0.05 to 100 ng/mL. The linear regression equation is ΔIECL = 48.817c + 1750.9, with a correlation coefficient of 0.9975, where ΔIECL represents the difference of ECL intensity between the Au/QDs/Nafion/anti-AFB1 Ab electrode (cAFB1 = 0) and the Au/QDs/Nafion/anti-AFB1 Ab/Ag electrode bonded with different concentrations of AFB1. The limit of detection (LOD) was 0.01 ng/mL (S/N = 3), which is much lower than other reported type immunosensors for AFB1 detection (Table 1). In addition, compared with other reported immunosensors used in simple foods or agricultural products [38,39,40,41], the detectable range of the present ECL immunosensor is wide enough with low LOD for complex TCM matrices like lotus seeds.

Table 1 Comparison of the newly-proposed ECL immunosensor with other immunosensors for the detection of AFB1
Full size table

The above results indicate that the newly developed ZnCdS@ZnS QDs–based label-free ECL immunosensor exhibits superior performance as a powerful tool for accurate and sensitive detection of AFB1 in different kinds of matrices.

Selectivity, stability, and reproducibility of the ECL immunosensor

The introduction of a highly specific anti-AFB1 Ab will improve the selectivity and specificity of any AFB1 immunoassay. To evaluate this property of the newly developed ECL immunosensor, ochratoxin A (OTA) and zearalenone (ZEA) were selected as interference mycotoxins to be tested under the same experimental conditions. As shown in Fig. 5a, the ECL signal was similar between 10 ng/mL of OTA or ZEA and the blank (without AFB1). In contrast, the presence of 0.1 ng/mL of AFB1 led to a similar decrease in ECL intensity compared with the addition of 10 ng/mL of OTA or ZEA, suggesting excellent selectivity and specificity of the antibody to AFB1.

Fig. 5
figure 5

(a) Selectivity investigation of the ECL immunosensor for AFB1 detection against two interference mycotoxin: (1) 0 ng/mL AFB1, (2) 0.1 ng/mL AFB1, (3) 0.1 ng/mL AFB1 + 10 ng/mL OTA, (4) 0.1 ng/mL AFB1 + 10 ng/mL ZEA, (5) 10 ng/mL OTA, (6) 10 ng/mL ZEA; and (b) the ECL stability of the proposed ECL immunosensor under continuous cyclic voltammetry scan with 0.1 ng/mL AFB1. Error bars were calculated of relative standard deviation from triple parallel experiments (n = 3)

Full size image

The ECL intensity of the sensor in the presence of 0.1 ng/mL of AFB1 under continuous scans for > 5 cycles over more than 80 s was recorded (Fig. 5b). High ECL intensities (approximately 8000 a.u.) and stable signals with relative standard deviation (RSD) < 1.29% were observed, illustrating good stability of the ECL immunosensor. To evaluate the batch-to-batch variations, five batches of ECL immunosensors were fabricated in parallel for the detection of 0.1 ng/mL AFB1 under the same optimum conditions. The RSD was < 2.88% across these batches, indicating outstanding reproducibility of the constructed ECL immunosensor.

Real sample analysis for AFB1

The feasibility and applicability of the proposed ECL immunosensor to real-world analyses was investigated by analyzing real samples of complex matrix. Medicinal and edible lotus seed samples were spiked with three concentrations of AFB1 (5, 10, and 20 ng/mL), followed by extraction using the preparation procedure described in the methods above. After the newly developed ECL immunosensor was incubated with lotus seed extractions, it was immersed into the electrolyte of the three-electrode system for detection of ECL signals. Decreasing ECL intensities and the ΔIECL-c regression equation were used to determine the amount of AFB1 in the spiked lotus seed samples under the same experimental conditions. The average recovery values (n = 5) in Table 2 were 99.2%, 100.7%, and 101% with RSDs < 3.80%, respectively. Therefore, the proposed ECL immunosensor exhibited excellent advantages and great potential for trace detection of mycotoxins in real complex samples.

Table 2 Recovery (n = 5) of the proposed ECL immunosensor for AFB1 in real lotus seed samples
Full size table

Conclusion

In this study, a simple, easy-to-fabricate, and cost-effective ZnCdS@ZnS QDs–based label-free ECL immunosensor was successfully developed and achieved selective and sensitive detection of AFB1 in real complex samples. The superiority and advantages of this new immunosensor can be recapitulated as follows: (1) This method introduced amplification strategies based on stable assembly of QDs through the addition of Nafion solution. In addition, the ZnCdS QDs capped with ZnS is more environmentally friendly and the core-shell QDs enhanced the ECL emission to obtain strong ECL signals compared with pure QDs. (2) The use of Nafion as the film-forming solution led to the stable conjunction of an abundance of QDs, as well as anti-AFB1 Ab and AFB1 onto the surface of the working electrode to improve sensitivity. (3) The development of a label-free ECL immunosensor omits the need for an additional “label” step and for the use of secondary antibodies as in traditional immunoassay, which largely lowers the time and cost of analysis. (4) This method bridges sensitive ECL technology and small molecule determination at trace levels through specific antibodies for target capturing fragments and core-shell QDs for signal amplification. (5) It provides a practical platform for trace detection of multiple targets in complex TCM matrices without interference from their complicated components. In summary, the developed label-free ECL immunosensor with high sensitivity is suitable to broad applications and real-time high-throughput analysis in food and TCM safety testing, environmental pollution monitoring, and disease diagnostics.