The network analyzer Agilent 4395 A with module QOM 401 (Q-Sense Open Module) was kindly provided by the laboratory SATIE, University of Cergy - Pontoise, France. QSX 301 crystal sensors with gold coverage, resonated from 5 - 15 MHz were acquired from Biolin, scientific company. Quartz solution was purchased from Helmanex. 1-ethyl-3-[3-(dimethylamino) propyl carbodiimide hydrochloride (EDC) and N-hydroxy sulfoximide (NHS) were purchased from Sigma Aldrich.

Rabbit anti-aflatoxin B1 and aflatoxin B1-BSA conjugate from Aspergillus flavus were obtained from Sigma Co., Ltd. Monoclonal anti-aflatoxin B1, clone AT-B1 produced in mouse, ascites fluid (Product No A9555) and antiaflatoxin B1-peroxidase antibody produced in rabbit (Product No A2681) were acquired from Sigma Co., Ltd. Streptavidin – peroxidase was purchased from ScyTecLaboratories. Ethanolamine and 3-mercaptopropionic acid were obtained from Sigma Aldrich. All the used solvents for the obtained measurements were analytical grade.

QSX 301 crystal sensors with gold coverage, resonated from 5 - 15 MHz, were used. The quartz surface was prepared according to the consecutive procedure, described below. The first step of the procedure was cleaning the quartz surface with “Pr Quartz solution” for 5 min. Then the quartz was rinsed several times with distilled water, followed by washing with isopropanol for 10 min at 50°C. The last step was quartz drying at 50°C. The quartz was placed in thermostat for calibration of the oscilloscope. The dry and tempered quartz was calibrated as a control one in the module cell connected to ultrasound analyzer.

The activated quartz surface was washed with distilled water and phosphate buffer (0.01M, pH 7.4). 10µl of antibody solution (rabbit anti-aflatoxin B1 dissolved in 0.01M phosphate buffer, pH 7.4, containing 15 mM NaN₃, and diluted 1:1000) and 90µl phosphate buffer were pipetted. The change in the signal of piezoelectric in 1 hour time was observed by the quartz resonator. The quartz surface was washed with distilled

Fig. (1). Preparation and activation of the piezocrystal by addition of NHS replacing the activated hydroxyl group of EDC after the formation of oxygen – acyl – carbamide intermediate structures [7].

Fig. (2). Immobilization of antibody - rabbit anti - aflatoxin B1 onto the activated monolayer.

water and phosphate buffer. Two types of blocking agents were used: 100µl ethanolamine (10mM) and 100µl Bovine Serum Albumin (BSA). The blocking solution reacted in the module cell for 1 h and let the blocking of the additional groups onto the quartz surface. The piezoelectric response, as a result of the separated addition of the blocking agents, was followed by the resonator.

The process of immobilization of the rabbit anti - aflatoxin B1 is presented in Fig. (2).

Once the activated monolayer is prepared, several molecules of the antibody can be attached on the surface. The amine groups of the antibodies replace the NHS and that follows to a formation of covalent bonds. This method is useful for the detection of antigen in a sample. The signal from the antigen-antibody reaction is measured and transformed by the resonator (Fig. 2). The SAM technique for immobilization of protein molecules can be used for the preparation of different immunosensors, based on QCM detection techniques analysis.

1.1 g of skimmed milk powder was measured and diluted in 9.990 ml distilled water. The sample was stirred and heated to 90°C for 10 min. Just before the measurement in the module cell, the milk sample was cooled to 25°C [11] and contaminated with a relevant concentration of AFB1-BSA, till the final volume of 10 ml (final concentration of AFB1-BSA, respectively 0.2 µg/L and 2 µg/L). The measurements were done with piezocrystal, which resonated to 15 MHz. The sample was measured in the presence of antibody, immobilized onto the quartz surface without pre–blockage of free groups on the surface.

After activation of quartz surface and immobilization of antibody, the crystal was washed with distilled water and phosphate buffer. 10µL of pre–contaminated sample with a concentration of AFB1-BSA, respectively 0.2 µg/L and 2 µg/L, were pipetted on the quartz in the module cell. Further, it was added 90µl phosphate buffer. The measurements and piezoelectric responses of quartz were monitored in real-time.

After the blockage of non-bonded active groups of activated quartz surface and washing of the quartz in a certain time frame, the detection of antigen - antibody reaction was observed. Concerning the minimal range of concentrations of antigen, we started the experiment with 100 µg/ml stock solution of AFB1-BSA. The addition of 10 µl, 30 µl and 100 µl of stock solution of antigen in phosphate buffer with ratio 10:90, 30:70, was estimated by the quartz resonator. The molecular weight of AFB1 is around 300 g/mol. We detected AFB1, conjugated with BSA by QCM to follow the antigen-antibody reaction and make a comparison with ELISA immunoassay. The QCM detection of AFB1 with unknown concentrations is influenced by pre-determined concentrations of AFB1-BSA. When the AFB1 concentration in the sample is low, then a number of AFB1-BSA molecules may interact with the immobilized antibody on the quartz surface, which will result as a higher signal of the resonator’s response. The reaction antigen - antibody in the module cell of the resonator was monitored for 24 h (Fig. 3-7).

All sensor’s responses were measured in real-time and the results were graphically presented, applying a developed mathematical model for processing the data in the laboratory of SATIE, University of Cergy - Pontoise [12, 13].

According to the presented graphs, we can see that the essential difference in resonator’s responses between the antigen – antibody reaction occurs at concentrations of antigen 10 µg/ml (Figs. 3) and 100 µg/ml (Fig. 5) of AFB1-BSA.

Fig. (3). QCM analysis of the formation of antigen – antibody reaction in the presence of 10 µg/ml AFB1-BSA. G’ and G” characterize the elasticity and the viscosity of the reaction mixture during the Ag-Ab reaction, respectively.

Fig. (4). QCM analysis of formation of antigen – antibody reaction in presence of 30 µg/ml AFB1-BSA. G’ and G” characterize the elasticity and the viscosity of the reaction mixture during the Ag-Ab reaction, respectively.

Fig. (5). QCM analysis of formation of antigen - antibody reaction in the presence of 100 µg/ml AFB1-BSA. G’ and G” characterize the elasticity and the viscosity of the reaction mixture during the Ag-Ab reaction, respectively.

In addition to the detection of antigen – antibody reaction, followed by the piezoelectric response in real-time using oscilloscope Agilent 4395 A, we also observed changes in viscoelastic properties of matter, covering the activated quartz, by measuring different concentrations of AFB1-BSA in pre – contaminated milk samples. Most of the published data about concentrations in food sources show that the accepted limit range for AFB1 content in food samples is in the range from 0,02 µg/L to 0.3 µg/L.

Fig. (6). QCM analysis of detection of 0.2 µg/L AFB1-BSA in contaminated milk sample.

We observed the effect of the immobilized rabbit anti-aflatoxin antibody B1 to the quartz surface in the presence of pre-contamination with AFB1-BSA milk samples. The used concentrations of AFB1-BSA were 0.2 µg/L and 2 µg/L (Fig. 6-8).

Fig. (7). QCM analysis of detection of 2 µg/L AFB1-BSA in the contaminated milk sample.

Sample of pure milk, without AFB1-BSA was prepared as a control sample and the resonator signal was measured. In addition, as a control sample was used the water samples, contaminated with 0.2 µg/L and 2 µg/L AFB1-BSA respectively. Control water samples were used, corresponding to the Newtonian fluids, in which viscosity is only dependent on temperature differences. Those samples have linear raise in stress with increasing shear rates and their viscosity remains constant. Water samples were used to follow the minimal changes in signals on the quartz surface in the presence of contaminated samples with AFB1-BSA. Milk samples were prepared from skimmed milk. Those samples are considered also like Newtonian fluids. Compared to water, skimmed milk has a plenty of protein, so we used the control water samples to see the differences of signals on the quartz resonator.

After processing the obtained data as a result of the experiments, differences in the accumulation module - G ” were observed, concerning the applied concentrations of AFB1-BSA (0.2 µg/L and 2 µg/L). This is a prerequisite for confirming the sensitivity of the used resonator.

Indirect and direct ELISA immunoassays were used as comparative methods for the detection of AFB1.

The protocol for analysis included overnight incubations of different concentrations of AFB1. The mycotoxins’ solutions were prepared in binding buffer (50 mM Na₂CO₃, pH 9.6), with concentrations of AFB1 in the range of 3 g/L to 0.001 g/L. Concentrations lower than 1 mg/L did not show any mathematical relevance between the measured minimal concentrations of AFB1. 50 μl samples were pipetted in each well of the microtiter plate for ELISA. In the next morning, the plate was blocked with 3% BSA in PBS-T (0.05% Tween-20 in PBS).

Fig. (8). Comparison between detection of antigen – antibody reaction in contaminated milk samples with AFB1–BSA concentrations of 0.2µg/L and 2 µg/L. G’ and G” characterize the elasticity and the viscosity of the reaction mixture during the Ab-Ag reaction, respectively.

We used antiaflatoxin B1-peroxidase antibody produced in rabbit for direct ELISA analysis, diluted in ratio 1:1000 in PBS-T. The serum incubation was processed for 2 hours at room temperature. The color reaction was developed with OPD (о-phenylendiamine dihydrochloride) in citrate buffer (pH 5.00). After 30 min the reaction was stopped with 10% solution of 2N H₂SO₄. The absorption was detected on ELISA reader at 450 nm. In the case of indirect ELISA, we used mouse monoclonal antibody against aflatoxin B₁, AT-B1, ascite produced in mouse (dilution ration 1:20000). After overnight incubation of the antibody, we additionally incubated the plate with anti - mouse antibody, conjugate with biotin. We used streptavidin - peroxidase for maximal enhancement of the signal and reaction development. The color reaction was developed respectively to the procedure of direct ELISA (Fig. 9).

The sensitivity of direct and indirect ELISA for detection of AFB1 is very low, which can be seen from very low linearity of the signal for concentrations, lower than 1000mg/L.

Fig. (9). Sensitivity, linear range and detection limits for AFB1 in cases of direct and indirect ELISA, where: A-concentration of AFB1 in the range 1 - 7.6 mg/L; B-concentration of AFB1 in the range 80 - 275 mg/L; C-concentration of AFB1 in the range 400 - 3000 mg/L;