Stock solutions (1 mM) of OTA and OTB (Sigma-Aldrich) were prepared in PBS and stored in small aliquots at −20 °C.

The OTA-BSA was obtained from Sigma-Aldrich, whereas the OTB-BSA conjugate was synthesized in our laboratory as earlier described [22]. Blocking reagents were purchased from Fluka (Skim milk powder, #70166), Sigma-Aldrich (Ovalbumin, #A5503; Casein, #009000719) and from Roth (BSA, #8076.2). Goat anti-mouse IgG (Fc-Fragment)-HRP (Dianova, #115-035-008) and rabbit anti-mouse IgG-HRP (Sigma, #A9044) were used for ELISA and Western blotting, respectively. ECL (GE Healthcare, #RPN2132) and TMB liquid substrate (Sigma-Aldrich, #T4444) were used as substrates. Unless otherwise stated, all chemicals were purchased from Sigma-Aldrich, Germany, and were of highest purity available.

The monoclonal mouse anti-OTB antibody (OTB mab), IgG1 with κ-chain, was produced using the hybridoma technique. The production and preliminary characterization of this OTB mab was recently described in detail by Heussner et al. [22].

For purification, the supernatants from the stable hybridoma cell line 2F1.E10 [22] were collected and stored at −20 °C. Purification was performed using protein G sepharose affinity chromatography according to manufacturer's recommendations (HiTrap, GE Healthcare). Briefly, samples were diluted with binding buffer (20 mM NaH₂PO₄, pH 7.0) and applied to the column. After washing, elution was achieved with Glycin-HCl (100 mM, pH 2.7). To avoid globulin damage, pH was titrated to neutral immediately with Tris-HCl (1 M, pH 9.0).

Eluted fractions were collected and protein content was determined via the Bradford assay calibrated with bovine γ-globulin according to manufacturer’s instructions (Roti®-Quant, Roth, Germany). Additionally, the purity of the protein-containing fractions was assayed via SDS-PAGE under reducing and non-reducing conditions on 12% and 7.5% polyacrylamide gels, respectively. Proteins were stained with silver nitrate [23].

Antibody-containing fractions were pooled and calculated protein content was confirmed via Bradford determination. Purified pooled OTB mab was diluted with glycerol (final concentration of glycerol 50% v/v) and stored in small aliquots at −20 °C.

Two different ochratoxin-BSA conjugates were used: OTB-BSA (own synthesis, [22]) and OTA-BSA (Sigma-Aldrich). For both, the stability was unknown and confirmation of protein and ochratoxin content were prerequisites for further use in the Western blot analyses (see below).

Protein concentration was determined via the Bradford micro assay according to the manufacturer’s recommendations (Roti®-Quant, Roth, Germany) using non-linear calibration with BSA. Ochratoxin concentrations were determined by photometric analysis at 360 nm and 380 nm for OTB and OTA, respectively, using non-linear calibration with ochratoxins in Tris buffer. As determination of low ochratoxin concentrations was the ultimate goal, the data were confirmed by additional readings using increasing ochratoxin sample spiking levels (1, 10 and 100 µM).

For semi-quantitative detection, ochratoxin-BSA conjugates were boiled in SDS-PAGE sample buffer (187.5 mM Tris-HCl, pH 8.8, 10% Glycerol, 2% SDS, 20% 2-Mercaptoethanol, 1% Bromophenol blue) and separated on a 10% polyacrylamide gel. Proteins were blotted onto a nitrocellulose membrane (0.2 µm pore size, Roth, Germany), reversibly stained with Ponceau S and blocked with 1% (w/v) BSA in TTBS (100 mM Tris-HCl, pH 7.6, 0.9% (w/v) NaCl, 0.1% (v/v) Tween 20) for 30 min. Thereafter, the OTB mab was used as the primary antibody at a concentration of 10 µg/mL (concentration was optimized in preliminary assays) and incubated for 1 hour at RT. Anti-mouse IgG-HRP (1:80,000, 45 min) and ECL were used for detection. Chemiluminescent signals were detected using a CCD camera system (Fuji LAS-1000).

For quantification, ochratoxin-conjugates in varying concentrations were applied directly on a nitrocellulose membrane (via dot blotting), followed by immunodetection as described above.

The initial ELISA was performed using the same parameters as optimized for the use of the crude hybridoma supernatant (see Table 1) [22]. Briefly, 96 well-plates (MaxiSorp, Nunc, Germany) were coated overnight at 4 °C with OTB-BSA conjugate (16 ng/well) in coating buffer (32 mM Na₂CO₃, 70 mM NaHCO₃, pH 9.6). Plates were washed three times with washing buffer and two times with PBS (10 mM sodium phosphate, 150 mM NaCl, pH 7.4) followed by blocking in blocking solution for one hour at RT. Standards, controls or samples (50 µL) were applied and 50 µL OTB mab were added. After one hour incubation, plates were washed again and 100 µL detection antibody (goat anti-mouse IgG (Fc-Fragment)-HRP, 1:5,000) were incubated for one hour. After washing, 50 µL TMB solution were added to each well and incubated 15 min at RT protected from light. The reaction was stopped by addition of 50 µL 2 M sulfuric acid and absorbance was recorded at 450 nm using an ELISA reader (Tecan, Infinite 200).

For optimization of ELISA conditions, grid experiments, allowing for the combination of variations of the different parameters, were performed. Parameters included coating concentrations (8, 16, 32 and 80 ng OTB-BSA per well), OTB mab concentrations (0.007, 0.02, 0.07, 0.2, 0.7 and 1 µg OTB mab/mL), ionic strength (0.1×, 0.5×, 1× and 1.5× PBS), pH (7.0, 7.4 and 8.0), blocking proteins at different concentrations (Casein (0.5, 1, 2%), BSA (1, 2, 3%), Ovalbumin (0.1, 0.2, 0.5%) and skim milk powder (0.1, 0.2, 0.5%)) and incubation conditions (static at 37 °C vs. shaking at RT). Four concentrations of OTB and OTA (0.01, 0.1, 1, 10 µM) were used for the latter experiments.

Complete competition curves for OTB and OTA were established using, OTA and OTB diluted in PBS (pH 7.0). Concentrations ranged from 10 nM to 6.6 µM and 10 nM to 200 µM for OTB and OTA, respectively. Relative cross-reactivity was determined using Equation 1.

The limit of detection (LOD; reagent blank + 3× SD of reagent blank) for OTB was 3 µg/L (8 nM), whereas the limit of quantification (LOQ; reagent blank + 10× SD of reagent blank) was 3.7 µg/L (10 nM). The LOD for OTA was 129 µg/L (320 nM), whereas the LOQ was 174 µg/L (430 nM).

In order to study matrix-associated effects, a very small basket study with beverages was carried out. Beverages randomly chosen in a local supermarket were coffee, beer (German wheat beer) and red wine (Merlot), known to often contain ochratoxins at low levels. Coffee was brewed using a standard filter coffee maker, whereas the beer and wine were used for analysis as taken from the supermarket.

Sample preparation was adapted from the method by Wang and co-workers [24], who optimized sample preparation for the detection of OTA in beverages and foods via ELISA. Briefly, for wine and coffee, 6 mL of sample were mixed with 24 mL chloroform and vigorously shaken for 30 min at RT. Then, 16 mL PBS (pH 9.0) were added and mixed. For phase separation centrifugation (1,560 × g, 5 min, RT) was applied and the supernatant (water-phase) was aliquoted into fresh tubes and stored at −20 °C until further use.

Beer was used without extraction, but was ultrasonically degassed for approximately 15 min and then stored at −20 °C as small aliquots until further use.

Matrix samples (see Section 2.8) were diluted with PBS (pH 7.0) using dilution factors of 2 to 25 for analysis until the absence of matrix-associated quenching effects was observed. Following establishment of optimal dilutions, diluted matrices were spiked with OTB ranging from 10 nM to 6.6 µM.

Duplicate samples were analyzed in each ELISA plate and run simultaneously with controls for total binding (vehicle, i.e., PBS or diluted sample matrix) and blank (samples without anti-OTB mab). Data were processed first using MS EXCEL 2003. Sample absorbance values were blank-subtracted and then compared to total binding (=100%). Means of the duplicates in each independently run ELISA were used as value of one independent replicate. Means of at least three independent replicates were further processed with GraphPad Prism 4.03 using nonlinear regression (sigmoid dose-response with variable slope) with constraints (0% TOP and 100% BOTTOM) and anchorage points. In addition, the competitor concentration inhibiting 50% binding (IC₅₀) was determined. The degree of competition was calculated using Equation 2.

Spiked beverage samples were compared to OTB standards in PBS using a t-test with p < 0.05.

Antibody purification resulted in a highly purified OTB mab, which was tested via SDS-PAGE with subsequent silver-staining (Figure 2). The gels confirmed the expected distinct banding of immunoglobulin in the absence of any other protein banding. Under reducing conditions, two bands were visible at 53 ± 1 kDa and 31 ± 1 kDa, representing the heavy and light chains, respectively. Under non-reducing conditions, only one protein band at 168 ± 5 kDa was observed.

The analysis of OTA-BSA protein concentration confirmed the 5 mg/mL of OTA-BSA as proclaimed by Sigma-Aldrich, while a concentration of 1.3 ± 0.2 µM OTA (n = 3, mean ± SD) fell short of the expected value. For the in-house synthesized OTB-BSA conjugate, a protein concentration of 16.3 ± 1.7 µg/mL (n = 3, mean ± SD) and a OTB concentration of 27.2 ± 1.2 µM (n = 3, mean ± SD) was established. Above ochratoxin-BSA protein concentrations were used for all ensuing calculations, e.g., dot-blots and western blots.

Above values were also used to calculate the labeling ratios (Ochratoxin: BSA) of 5.3 and 0.36 for OTB-BSA and OTA-BSA, respectively. These values differed from those determined earlier (OTB-BSA, [22]) or provided by the supplier (OTA-BSA, Sigma-Aldrich), suggesting partial loss of ochratoxin-BSA binding during frozen storage.

Western blot analysis enabled the control of the ELISA components, thus confirming the previously described method [22]. In addition, Western blot analysis tested and confirmed the potential usability of the OTB mab for the analysis of OTB-protein complexes. Moreover, Western blot analysis demonstrated that the OTB mab is capable of detecting OTB- and OTA-protein complexes with similar specificity (Figure 3), albeit with different sensitivity (Figure 4).

Whether this approach is applicable to the detection of non-covalently bound, but chemically matrix-linked OTB (e.g., in immunocytochemistry) will be established in future experiments.

Based on the variant combinations of ELISA parameters used in the grid experiments, the most ideal ELISA conditions were identified (Table 1). The comparison of the conditions employed in the initial ELISA with those of the optimized ELISA demonstrated that not only a simpler but also more convenient (RT vs. 37 °C incubation) ELISA was optimal. The most effective modification was the choice of protein for blocking. The use of BSA, rather than casein, ovalbumin or skim milk resulted in highest sensitivity towards OTB and lowest cross-reactivity towards OTA. This may due to the high affinity of OTA to BSA.

These optimized conditions were used to investigate competitive effects of OTB and structurally related molecules e.g., OTA, coumarin and phenylalanine. No cross-reactivity to other structurally related molecules (coumarin and phenylalanine) was observed (data not shown), confirming earlier findings [22]. Detection could be achieved within a range of 0.01 to 10 µM OTB (Figure 5). The OTB concentration causing 50% inhibition (IC₅₀) was 0.15 µM (95% CI: 0.12–0.19, n = 5), suggestive of a very sensitive assay. The limit of detection (LOD; reagent blank + 3× SD of reagent blank) for OTB was 3 µg/L (8 nM), whereas the limit of quantification (LOQ; reagent blank + 10× SD of reagent blank) was 3.7 µg/L (10 nM).

OTA was detected with an IC₅₀ of 5.65 µM (95% CI: 4.18-7.65, n = 6). Based on the IC₅₀’s determined, the cross-reactivity of OTA with the OTB mab was calculated to be very low (2.7%). The LOD for OTA was 129 µg/L (320 nM), whereas the LOQ was 174 µg/L (430 nM).

Reduction of matrix interference is frequently achieved in ELISA by simple dilution with assay buffer, especially when a highly sensitive assay is available. Samples containing high pigment, flavors, tannin or lipids generally interfere with the color development in ELISA. Therefore, chloroform extraction with subsequent alkaline buffer dilution, where the interfering substances are retained in the chloroform layer and toxins (in this case ochratoxins) are easily dissolved in alkaline solution, was chosen.

In a first set of analyses, prepared matrix samples were diluted with PBS pH 7.0 using dilution factors of 2 to 25 for analysis, i.e., until no quenching effects of the matrices were detectable. Based on the latter results, no quenching effects were detected (Figure 6) at dilutions of 1:2, 1:10 and 1:20 for coffee, beer and wine, respectively.

Ten different concentrations of OTB were added (spiking experiment) to diluted matrix samples using the optimized dilutions of the respective sample (matrix) type (see above). The resulting spiked matrix samples were analyzed using the optimized OTB ELISA. The resulting competition curves were compared to that established with pure OTB in PBS (Figure 6). No significantly differences in the curves (form, slope, range) could be detected for all three matrices (coffee, beer, wine), when compared to the standard OTB curve. This t suggests that the optimized OTB ELISA provides for a robust and sensitive OTB detection in these matrices.