Introduction

Ochratoxin A (OTA) is a mycotoxin produced by Penicillium and Aspergillus moulds and can be found in a wide range of foods, such as cereals, beer, wine, cocoa, coffee, dried vine fruit and spices, and in meat products as a result of contamination of animal feed [1]. OTA is nephrotoxic and carcinogenic and poses a serious threat to the health of both humans and animals [2, 3]. The International Agency for Research on Cancer has classified OTA as a possible carcinogen for humans (group 2B) [4]. The Joint FAO/WHO Expert Committee on Food Additives estimated that cereals and cereal products as food commodities contribute more than 50 % to human OTA exposure [2]. The European Commission has set maximum limits for OTA in raw cereal grains, roasted coffee and coffee products (5 μg kg-1), products derived from cereals (3 μg kg-1), and other food commodities, including dried fruits, grape juice, wines, spices, and liquorice root [5, 6].

Analytical methods for the detection of OTA, alone or in combination with other mycotoxins, have been reviewed [7–10]. Immunoaffinity column (IAC) clean-up in combination with high-performance liquid chromatography (HPLC) with fluorescence detection (FLD) is the most widely used procedure for the determination of OTA in food and feed. IAC offers several advantages, including provision of clean extracts from complex matrices owing to the specificity of the antibody for the toxin. Although antibodies remain the primary mycotoxin-binding element, new and inexpensive synthetic ligands, such as aptamers, with high affinity for the target analyte are increasingly being proposed as alternative materials to antibodies. In addition to the potential for low production costs, the benefits of aptamers include greater capacity, stability during storage and improved tolerance to solvents or extremes of pH or ionic strength [11–13]. Various analytical aptamer-based formats have been exploited, including enzyme-linked oligonucleotide assays, biosensors (called aptasensors), flow cytometry, and capillary electrophoresis/capillary electrochromatography [14]. Aptamers have been produced for several targets, including peptides, proteins, drugs, whole cells [15] and, recently, mycotoxins, e.g. OTA [16, 17], fumonisin B1 [18], aflatoxin B1 and zearalenone [19].

Since its development in 2008, the OTA DNA aptamer has been integrated into several biosensor detection systems, including electrochemical [20–24], electrochemiluminescent [25, 26], colorimetric [27, 28] and fluorescent [29, 30] platforms, enzyme-linked aptamer assays [17] and fluorescent test strips [31–33]. Moreover, affinity column systems based on OTA aptamers have been developed for clean-up of food extracts [16, 34–36]. Although promising as alternative technology in the analysis of OTA in food and beverages, most of these methods are ‘proof-of-concept’ assays measuring OTA in simple solutions such as buffers, without evaluating the potentially interfering compounds present in a real food matrix [27–29]. Moreover, only a few of them were applied to the analysis of OTA in cereal products [16, 21] and lacked the robustness to move the technology to commercial readiness.

A standard procedure for the preparation of aptamer-based affinity columns has recently been applied to the analysis of OTA in wheat extracts prior to HPLC/FLD analysis [36]. The columns showed performance equivalent to that of relevant IACs and were potentially useful to bring the technology to commercial readiness [36]. As a follow-up of these results, optimised DNA-aptamer-affinity columns were developed and a DNA-aptamer-based detection system was developed for the direct (non-HPLC) determination of OTA in wheat.

Here we report the analytical performance of this novel DNA-ligand system on the basis of the measurement of the time-resolved fluorescence (TRF) response due to the interaction between terbium, a fluorescent lanthanide with a long lifetime, OTA and the aptamer to determine OTA in wheat. A comparative study was conducted by analysing 29 wheat samples using the aptamer-affinity column/TRF method or a reference method using IAC/HPLC-FLD analysis.

Materials and methods

Reagents and chemicals

OTA standard, terbium chloride and the rest of the reagents were purchased from Sigma-Aldrich (Milan, Italy). Glass microfibre filters (Whatman GF/A) and paper filters (Whatman no. 4) were obtained from Whatman (Maidstone, UK). All other chemicals and solvents were reagent grade or better and were from Carlo Erba Reagents (Milan, Italy). Ultrapure water was produced by a Milli-Q system (Waters, Milford, MA, USA). OTA-Sense® aptamer-affinity columns, OTA-Sense® buffer solution and the aptamer-probe solution were provided by Neoventures Biotechnology. (London, ON, Canada). The cross-reactivity of the aptamer for 100 nM ochratoxin B was 100-fold less than that for OTA as calculated from equilibrium dialysis experiments [16]. OchraTestTM WB columns were purchased from VICAM. (Milford, MA, USA). Black 96-well low-fluorescence Corning® microplates were from Sigma-Aldrich (Milan, Italy). A stock standard solution of terbium chloride (20 mM) was prepared daily in ultrapure water. The OTA-Sense® detection solution was prepared by adding 350 μL methanol and 10 μL stock terbium chloride solution to 650 μL OTA-Sense® buffer solution. The binding buffer contained 10 mM tris(hydroxymethyl)aminomethane (pH 7.5), 120 mM NaCl, 5 mM KCl and 5 mM MgCl2. The elution buffer consisted of 10 mM tris(hydroxymethyl)aminomethane (pH 7.5), 120 mM NaCl, 5 mM KCl.

Two quality control materials, i.e., BRM 003023 (Biopure, Romer Labs Diagnostic, Tulln, Austria) and FAPAS T1976 (StarEcotronics, Milan, Italy), were used to assess the trueness of the aptamer-based method. The BRM 003023 material had an assigned value of 8.6 μg kg-1 OTA, with an estimated expanded uncertainty of ± 3.6 μg kg-1 OTA. The FAPAS T1976 material had an assigned value of 2.10 μg kg-1 OTA, with a satisfactory range of 1.18-3.03 μg kg-1 OTA. Twenty-nine unprocessed durum wheat samples belonging to different cultivars were obtained from various fields in Italy and Canada. Wheat samples were finely ground with a Tecator Cyclotec 1093 laboratory mill (International PBI, Milan, Italy) equipped with a 500-μm sieve.

OTA standard solutions

An OTA stock solution, with a final concentration of 10.94 μg mL-1 (as determined by spectrophotometric measurement) was prepared by dissolving the solid toxin in methanol. OTA solution (0.547 μg mL-1) for spiking purposes was prepared by diluting the stock solution 1:20 with methanol. Adequate amounts of the stock solution were dried and reconstituted with acetonitrile/water/acetic acid (99:99:2, v/v/v) or with the elution buffer to obtain standard solutions to produce OTA calibration curves by HPLC or by TRF measurement.

Sample extraction

Wheat samples were extracted according to the VICAM instruction manual for OTA analysis in wheat by OchraTest WBTM and HPLC-FLD [37] using acetonitrile/water (60:40, v/v) as the extraction solvent. Briefly, ground samples (50 g) were weighed into a test tube and extracted with 100 mL extraction solvent by vortexing the solution vigorously for 5 min. The extract was filtered through filter paper and diluted with binding buffer in a ratio of 1:20 (v/v). The diluted extract was then filtered through a glass microfibre filter and cleaned up using an aptamer-affinity column. For comparison, an aliquot of the same extract was diluted with phosphate-buffered saline in a ratio of 1:5 (v/v), then filtered through a glass microfibre filter and cleaned up using an IAC.

Aptamer-affinity column clean-up

The aptamer-affinity column was washed with 2 mL binding buffer prior to use. A 3-mL volume of diluted sample extract was cleaned up using the column at a flow rate of about one drop per second. The column was washed with 0.7 mL binding buffer and the eluate was discarded. OTA was eluted from the column by passing 0.5 mL elution buffer into silanised vials at a flow rate of about one drop per second and was analysed by TRF spectroscopy.

IAC clean-up

Ten milliliters of diluted extract was cleaned up using an IAC at a flow-rate of about one drop per second, followed by washing with 10 mL phosphate-buffered saline and 10 mL water at about one or two drops per second. OTA was then eluted with 1.5 mL methanol and collected in a silanised vial. The eluate was evaporated under a nitrogen stream at about 50 °C, reconstituted with 1.5 mL of the HPLC mobile phase and stored at 4 °C until HPLC analysis.

TRF spectroscopy

The TRF of OTA in wheat extracts was measured using a FluoStar OMEGA microplate reader (BMG Labtech, Allmendgruen, Germany) equipped with excitation and emission filters set at 380 and 540 nm, respectively, with a bandwidth of 10 nm. The apparatus parameters were as follows: lag time, 100 μs; integration time, 1.5 ms; flash numbers, 200; maximum gain, 4,095; position delay, 0.2 s. Labtech OMEGA (version 1.2) and OMEGA Mars Data Analysis software programs were used for data acquisition and data processing, respectively. No spacers between the optical system and the microplate were used. Before analysis, the fluorometer was allowed to equilibrate for 5 min. The cross-reactivity of the TRF detection system for 200 nM ochratoxin B was 60-fold less than that for OTA, whereas no cross-reactivity with other compounds structurally similar to OTA, such warfarin, was observed [38].

For OTA detection, a 100-μL aliquot of eluted extract or standard solution was transferred to a clean vial, diluted with 100 μL OTA-Sense® detection solution and, after the solutions had been mixed well by vortexing, 20 μL of the aptamer-probe solution was added. After mixing, 200 μL of this combined solution (equivalent to 14.0 mg matrix equivalent) was transferred to a well of a 96-well microplate and analysed by TRF spectroscopy. Results were obtained in about 3 s. Quantification of OTA was performed by measuring the TRF value of OTA and comparing it with the relevant calibration curve. The test allowed analysis of up to 96 samples simultaneously. The linearity of the analytical response was checked by analysing the calibration standards dissolved in the elution buffer and using seven concentrations over the range 0.08-5.42 ng mL-1 OTA.

HPLC analysis

HPLC analysis of OTA in the eluates from the IAC was done out using an Alliance® high-throughput system (Waters, Milford, MA, USA) equipped with a fluorometric detector (Waters 2475 multiwavelength fluorescence detector) and the Empower™ software systems for instrument control and data processing. The analytical column was a Symmetry C18 column (150 mm × 4.6 mm, 5 μm) (Waters). The flow rate of the mobile phase was 1.0 mL min-1. The oven temperature was set at 35 °C. The mobile phase consisted of a mixture of acetonitrile/water/acetic acid (99:99:2, v/v/v). With this mobile phase, the retention time of OTA was about 7.0 min. The excitation and emission wavelengths of the fluorometric detector were set at 333 and 460 nm, respectively. The injection volume for HPLC analysis was 50 μL. Quantification of OTA was performed by measuring peak areas at the OTA retention time and comparing them with the relevant calibration curve. The linearity of the analytical response was checked by analysing the calibration standards dissolved in the mobile phase and using seven concentrations over the range 0.29-10.00 ng mL-1 OTA.

Recovery experiments

Recovery experiments were performed by spiking blank (OTA-free) wheat samples with OTA spiking solution at concentrations in the range 1–25 μg kg-1. After they had been spiked, the samples were left for 1 h at room temperature to allow solvent to evaporate, and were then extracted with acetonitrile/water (60:40, v/v) and analysed by both the aptamer-affinity column/TRF method and the IAC/HPLC-FLD method. Each recovery experiment was conducted in triplicate.

The limits of quantification (LOQ) of both the aptamer-affinity column/TRF method and the IAC/HPLC-FLD method were calculated as the mean signal of blank (OTA-free) wheat samples (n = 10) plus ten standard deviations of the mean signal.

Evaluation of matrix and terbium effects on TRF response

The optimum emission wavelength for the enhancement of the terbium–OTA complex was determined by measuring the emission spectra of terbium and OTA, alone and in combination, from 450 to 650 nm, with 370-nm excitation wavelength. Scans were performed with a monochromatic fluorometer (Sapphire II, Tecan, Switzerland).

The effect of terbium on the enhancement of the aptamer–terbium–OTA complex was evaluated by measuring the TRF intensity of a standard solution (2.5 ng mL-1 OTA) and of a naturally contaminated (2.2 μg kg-1 OTA, as determined by the IAC/HPLC-FLD method) wheat sample with different terbium concentrations. The eluted sample extract and OTA standard solution were appropriately diluted, then OTA-Sense® detection solutions with different terbium concentrations (i.e. 1, 5, 10 and 20 mM) were added and the mixture was analysed by TRF spectroscopy.

The matrix effect on the TRF response was evaluated in the range 0.16-1.97 ng mL-1 OTA by testing OTA standard solutions that were dried and reconstituted with the eluate from the aptamer-affinity column of a blank (OTA-free) wheat sample. OTA-Sense® detection solution (20 mM terbium stock solution) was added to the spiked extracts and the mixture was analysed by TRF spectroscopy by using different amounts of matrix equivalent, namely 3.5, 7.0 or 14.0 mg. The regression lines obtained were compared with those obtained by using the same amounts of OTA reconstituted with elution buffer.

To evaluate the stability of the aptamer–terbium–OTA complex over time, OTA-Sense® detection solution (20 mM terbium stock solution) was added to OTA standard solutions in the range 0.625-5.00 ng mL-1 and the mixtures were analysed by TRF spectroscopy immediately (t 0) and after incubation for 5 min (t 1), 10 min (t 2), 15 min (t 3), 20 min (t 4) and 30 min (t 5).

Statistical analysis

Comparison of linear regression curves was performed by parallelism statistical tests [39]. The good fit of the regression curves was evaluated by comparing the calculated slope and intercept with the “ideal” values of 1 and 0, respectively [40]. Data were processed using the statistical software package Statistics for Windows (StatSoft, Tulsa, OK, USA).

Results

Recovery experiments and analytical testing range

The results of recovery experiments of OTA from wheat samples extracted with acetonitrile/water (60:40, v/v) and analysed by both the aptamer-based testing system (aptamer-affinity column/TRF) and the IAC/HPLC-FLD method are reported in Table 1. Average recoveries for the aptamer-affinity column/TRF method in the spiking range from 2.5 to 25 μg kg-1 OTA ranged from 72 to 81 % (average 77 %), fulfilling the criteria set by the European Commission for the acceptability of OTA analysis [41]. A lower recovery value was obtained at a spiking level of 1.0 μg kg-1 OTA. Average recoveries for the IAC/HPLC-FLD method ranged from 82 to 91 % (average 86 %). The relative standard deviations of both methods were less than 7 %. The LOQs were 0.5 μg kg-1 for the aptamer-affinity column/TRF method and 0.3 μg kg-1 for the IAC/HPLC-FLD method. The trueness of the method was established by the analysis of two quality control wheat materials, which gave results of 5.89 ± 0.12 μg kg-1 for BRM 003023 and 1.45 ± 0.03 μg kg-1 OTA for FAPAS T1976, both within the range reported by the providers (5.0-12.2 μg kg-1 and 1.18-3.03 μg kg-1 OTA, respectively). However, the recovery values calculated with respect to the reference values were 68.5 % for BRM 003023 and 69 % for FAPAS T1976, slightly lower than the EU prescribed value of 70 %.

Table 1 Average recoveries of ochratoxin A (OTA) from spiked durum wheat analysed with both a DNA-ligand testing system (aptamer-affinity column/ time-resolved fluorescence, TRF) and a reference method (immunoaffinity column/high-performance liquid chromatography with fluorescence detection)
Full size table

Matrix effect and terbium enhancement evaluation

Figure 1 shows the emission spectra of terbium and OTA alone and in combination. No bands were observed for OTA, whereas terbium showed characteristic emission bands at 490, 545, 583 and 622 nm, with the maximum intensity peak at 545 nm. When OTA was added to the terbium solution, its emission was intensified by about 1.5-fold, indicating that a complex was formed and that energy transfer occurred from OTA to terbium. The emission wavelength of 545 nm was selected for TRF measurements of OTA because it gave the highest fluorescence intensity enhancement.

Fig. 1
figure 1

Fluorescence spectra of a terbium solution (3 mM), the ochratoxin A (OTA) standard solution (100 nM) and the terbium–OTA coordination complex (excitation wavelength 370 nm)

Full size image

The TRF of the aptamer–terbium–OTA complex was enhanced by about eightfold at 1 mM terbium, and remained stable up to 20 mM when OTA standard solution was tested (Fig. 2). On the other hand, in the presence of a matrix, the TRF increased with increasing terbium concentrations up to 10 mM, and then remained stable up to 100 mM. We decided to use 20 mM terbium in the final procedure to guarantee that the overall OTA present in the sample testing solution could react with terbium (Fig. 2).

Fig. 2
figure 2

Time-resolved fluorescence (TRF) of the aptamer–terbium–OTA coordination complex of a standard solution containing 2.5 ng mL-1 OTA (filled circles) and a wheat sample naturally contaminated with 2.2 μg kg-1 OTA (empty circles) using different terbium concentrations. The data were corrected by the relevant background TRF signal of the aptamer–terbium complex

Full size image

To evaluate the possible effect of matrix interferences on TRF intensity, matrix-assisted calibration curves in the concentration range 0.16-1.97 ng mL-1 were prepared by using either OTA standard solutions or spiked diluted extract of uncontaminated wheat samples. No significant differences were observed between the slopes (t calc < 2.447, p < 0.05) and positions (t calc < 2.365, p < 0.05) of the regression lines obtained with OTA standard solutions in elution buffer and those obtained in the presence of different amounts of the matrix (Table 2). This indicated the absence of detectable matrix effects that could produce an overestimation of the toxin content in the working conditions. In addition, the measurement of TRF from the aptamer–terbium–OTA complex over time in a solution containing 14 mg of matrix equivalent indicated that the complex was stable for up to 30 min. In particular, no significant differences were observed between the slopes (t calc < 2.447, p < 0.05) and positions (t calc < 2.365, p < 0.05) of the regression line for the TRF at t 0 as compared with those obtained for incubation times of 5–30 min (Table 2).

Table 2 Linear regression equations (y = ax + b) for TRF from the aptamer–terbium–OTA complex (y) versus OTA concentration (x) at different matrix equivalent amounts in the testing solution or at different reaction times. The values of t for the slopes and positions of the regression lines were calculated by a parallelism statistical test
Full size table

Comparison between aptamer-affinity column/TRF and IAC/HPLC-FLD analyses of contaminated wheat samples

The calibration curves obtained for the aptamer-affinity column/TRF and the IAC/HPLC-FLD methods confirmed the linearity (y = 365,001x + 22,794 and y = 42.67x + 0.07, respectively) of the instrumental response in the ranges investigated and showed satisfactory correlation coefficients (r = 0.999 for both methods).

Good correlation (r = 0.985) between OTA concentrations in 29 wheat samples analysed by both the aptamer-affinity column/TRF method and the IAC/HPLC-FLD method was found (Fig. 3). Twenty-four of 29 tested samples were found to be contaminated with OTA at concentrations from 0.4 to 13.5 μg kg−1 (as determined by HPLC). No false-positive results were obtained by the aptamer-based method for uncontaminated samples, whereas only one of the negative samples was found to be contaminated with 0.4 μg kg−1 OTA by HPLC (below the LOQ).

Fig. 3
figure 3

Comparison of OTA contents in naturally contaminated wheat samples obtained by the DNA-aptamer affinity column/TRF method and the immunoaffinity column/high-performance liquid chromatography with fluorescence detection (IAC/HPLC-FLD) method

Full size image

Discussion

Aptamers are DNA ligands able to discriminate target molecules with high affinity and specificity, even in the case of very closely related structures, and present numerous advantages over antibodies. They are produced by an in vitro process which does not require the use of animals and offers a greater degree of control with respect to binding conditions. Moreover, chemical modifications to enhance their stability, specificity, detection and immobilisation can be relatively easy to perform without affecting their affinity for the target [36]. Aptamer-based technologies are becoming a promising and convenient alternative to antibodies for the detection of contaminants in foods.

This work has described for the first time the application of an innovative and easy-to-use commercial DNA-aptamer testing system for high-throughput OTA analysis in wheat. The testing system is composed of two components, a DNA-aptamer-based affinity column for the clean-up of wheat extracts, and a DNA-aptamer-based detection system for the measurement of TRF. Some lanthanide chelates (such as terbium, europium and samarium chelates) have been successfully developed as fluorescence labels for highly sensitive detection of various biological molecules in TRF spectroscopy. These molecules showed great improvement of the fluorescence compared with conventional fluorescence spectroscopy using organic fluorescent labels. This was due to the long lifetime of the fluorescent lanthanide chelates that in combination with time-resolved measurement effectively removes short-lifetime background fluorescence, leading to high signal-to-noise ratios, and enhances the sensitivity remarkably [42]. Quantification by time-resolved fluoroimmunoassay using fluorescent lanthanide chelates has been reported for the determination of some mycotoxins, including OTA [9, 10]. Moreover, a method using TRF spectroscopy after postcolumn enhancement with terbium was previously described for detection of OTA and citrinin in spiked soft cheeses [43].

OTA has been reported to form a coordination complex with cations present in solution and this complex enhances binding to the DNA aptamer, which is highly negatively charged [16]. Terbium appears to act as a cation bridge between the DNA aptamer and OTA, thus representing an advancement in the use of fluorescence for the determination of OTA in wheat [38]. The use of the DNA aptamer reported here represents an improvement as the signal from the terbium in the presence of both the aptamer and OTA was stronger than the signal from terbium bound to OTA. Presumably this was due to the evacuation of water from the physical proximity of the terbium molecule while it is associated with OTA and the aptamer [38]. Terbium can also bind to potential contaminants from the wheat matrix that survive the clean-up as well as to DNA aptamer that is not involved in binding to OTA. Terbium can also fluoresce in these bound states as the binding may result in at least a partial exclusion of water. To eliminate this source of background fluorescence, the system is excited at a wavelength which specifically excites OTA; the energy from OTA is then transferred to the associated terbium atoms, which release it as fluorescence on the basis of the principle of fluorescence resonance energy transfer.

The present DNA-aptamer testing system represents an innovative and rapid technology with analytical performances, in terms of accuracy and precision values, similar to those of the IAC/HPLC-FLD method (in the range 2.5-7.5 μg kg-1). These performances satisfy the EU criteria for acceptance of an analytical method for the determination of OTA [41] in this range, including the EU regulatory level of 5 μg kg-1 [5]. The trueness of the aptamer-based method tested by the analysis of the two quality control materials indicated that the recovery results were slightly lower than the lowest EU prescribed value (70 %). Nevertheless, these results were within the range of acceptability reported by the providers of the quality control materials, thus indicating acceptable performances of the method. A good correlation of the results was obtained by comparative analyses of naturally contaminated wheat samples by the DNA-aptamer testing system and IAC/HPLC-FLD.

Unlike other immunochemical assays, e.g. ELISA [7], the present DNA-ligand testing system did not overestimate the toxin content as demonstrated by the lack of a matrix effect and good stability of the aptamer–terbium–OTA complex over time (up to 30 min). In addition, the smaller clean-up volumes and the absence of a drying step allow a shorter analysis time (about 30 min) as compared with the IAC/HPLC-FLD method (about 90 min). These results, combined with the simplicity and the possibility to process up to 96 samples simultaneously, make this method suitable for high-throughput screening as well as for quantitative determination of OTA in wheat.

Conclusions

The analytical performances of an innovative DNA-aptamer testing system have been tested for the rapid determination of OTA in wheat by affinity column clean-up and TRF energy transfer. TRF spectroscopy has the advantage of low background noise, short operational time and high sensitivity. The test can be performed with no special skills in less than 30 min, including sample preparation. Moreover, the analytical performances fulfil the criteria established by the EU for acceptance of an analytical method.

It is a simple, rapid and sensitive method that can be used to screen large quantities of samples and is suitable for the safety testing of agricultural products.