Assessment of the Optimum Linker Tethering Site of Alternariol Haptens for Antibody Generation and Immunoassay Development

Immunochemical methods for mycotoxin analysis require antigens with well-defined structures and antibodies with outstanding binding properties. Immunoreagents for the mycotoxins alternariol and/or alternariol monomethyl ether have typically been obtained with chemically uncharacterized haptens, and antigen conjugates have most likely been prepared with mixtures of functionalized molecules. For the first time, total synthesis was performed, in the present study, to obtain two haptens with opposite linker attachment locations. The functionalized synthetic haptens were purified and deeply characterized by different spectrometric methods, allowing the preparation of bioconjugates with unequivocal structures. Direct and indirect competitive enzyme-linked immunosorbent assays, using homologous and heterologous conjugates, were employed to extensively evaluate the generated immunoreagents. Antibodies with high affinity were raised from conjugates of both haptens, and a structure-activity relationship between the synthetic haptens and the specificity of the generated antibodies could be established. These results pave the way for the development of novel highly sensitive immunoassays selective of one or two of these Alternaria mycotoxins.


Introduction
Alternaria sp. fungi, particularly A. alternata, are ubiquitous plant pathogens and saprophytes that infect economically relevant crops such as cereals, vegetables, oilseeds, and fruits. Moreover, these microorganisms can contaminate these commodities after harvest even under refrigeration conditions. They are known to produce a wide variety of toxic secondary metabolites [1], and some of them have been identified by the EFSA Panel on Contaminants in the Food Chain (CONTAM) as a potential risk to human and animal health due to their toxicity and occurrence in food and feed. Surprisingly, there are no specific international regulations for any of the Alternaria mycotoxins, and the available data on toxicity, occurrence, and dietary exposure are still limited. In 2011, EFSA carried out the first assessment of the risk of these mycotoxins to human and animal health, based on government and published data [2]. More recently, EFSA conducted a survey on the dietary exposure of European consumers to Alternaria toxins [3]. This study found that 8% of these mycotoxins are present in food, with infants and other children being the most exposed population group, and fruit and fruit-based products contributing most to dietary

Hapten Design and Synthesis
The generation of antibodies to AOH has so far been based on the preparation of the required immunogens from AOH itself, which does not allow fine control over the specific position of the mycotoxin framework where the functionalized linker is introduced. In this study, we have synthesized two regioisomeric haptens of AOH from scratch. One of them, hapten ALa, incorporates a five-atom carboxylated aliphatic spacer arm through the hydroxyl group at C-9, whereas the other one, hapten ALb, incorporates the same linker via the hydroxyl group at C-3 ( Figure 1). In contrast to previous strategies, these two haptens allowed the preparation of bioconjugates with well-defined compositions.
position of the mycotoxin framework where the functionalized linker is introd this study, we have synthesized two regioisomeric haptens of AOH from scratch them, hapten ALa, incorporates a five-atom carboxylated aliphatic spacer arm the hydroxyl group at C-9, whereas the other one, hapten ALb, incorporates t linker via the hydroxyl group at C-3 ( Figure 1). In contrast to previous strategie two haptens allowed the preparation of bioconjugates with well-defined compos The synthetic strategy for preparing hapten ALa was based on a convergen odology previously used by several research groups to synthesize AOH and oth turally related molecules [17][18][19][20]. A key step in this synthesis is a Pd(0)-catalyze coupling reaction between an aryl triflate (4), which already contained the spa and an appropriately functionalized arylboronic acid (6) (Scheme 1). The aryl t was prepared in two steps from the readily available 1,3-benzodioxinone 1 [21,2 an O-alkylation reaction with methyl 5-bromovalerate was performed under s Williamson ether synthesis conditions. The alkylation process produced a 6:1 mi di-and mono-O-alkylation products, 2 and 3, respectively, which were easily se by column chromatography to provide the product resulting from the selective O tion of the less hindered hydroxyl group, i.e., 3, with a 75% yield. The free hydroxy of 3 was then converted to the required triflate group by reaction with triflic anhy pyridine, giving the triflate 4 in 91% yield. The additional required coupling react boronic acid 6, was prepared from the orcinol-derived bromide 5 [23] by haloge exchange using butyllithium and reaction of the resulting lithiated derivative wit propyl borate.
The subsequent palladium-catalyzed Suzuki-Miyaura coupling between the flate 4 and the labile boronic acid 6 gave the biaryl 7 in 75% yield. Hydrolysis of th oxymethyl ether (MOM) groups by treatment with methanolic HCl, followed by i lecular transesterification promoted by trifluoroacetic acid, completed the synthes tricyclic benzochromenone backbone and afforded the methyl ester of hapten AL pound 8, in 97% yield. To complete the synthesis of hapten ALa, only the hydr the methyl ester moiety of 8 was required, which was initially carried out und conditions (LiOH in THF-H2O at room temperature, rt). However, under thes tions, the central lactone group of the benzochromenone core was partially ope quiring acid treatment of the reaction crude to reconstruct the tricyclic ring sy proved more convenient to carry out this transformation using enzymatic hydro a lipase from Candida antarctica immobilized on an acrylic resin was used to hydro methyl ester group, providing hapten ALa in practically quantitative yield. The synthetic strategy for preparing hapten ALa was based on a convergent methodology previously used by several research groups to synthesize AOH and other structurally related molecules [17][18][19][20]. A key step in this synthesis is a Pd(0)-catalyzed cross-coupling reaction between an aryl triflate (4), which already contained the spacer arm, and an appropriately functionalized arylboronic acid (6) (Scheme 1). The aryl triflate 4 was prepared in two steps from the readily available 1,3-benzodioxinone 1 [21,22]. First, an O-alkylation reaction with methyl 5-bromovalerate was performed under standard Williamson ether synthesis conditions. The alkylation process produced a 6:1 mixture of di-and mono-O-alkylation products, 2 and 3, respectively, which were easily separated by column chromatography to provide the product resulting from the selective O-alkylation of the less hindered hydroxyl group, i.e., 3, with a 75% yield. The free hydroxyl group of 3 was then converted to the required triflate group by reaction with triflic anhydride in pyridine, giving the triflate 4 in 91% yield. The additional required coupling reactant, aryl boronic acid 6, was prepared from the orcinol-derived bromide 5 [23] by halogen-metal exchange using butyllithium and reaction of the resulting lithiated derivative with triisopropyl borate.
The subsequent palladium-catalyzed Suzuki-Miyaura coupling between the aryl triflate 4 and the labile boronic acid 6 gave the biaryl 7 in 75% yield. Hydrolysis of the methoxymethyl ether (MOM) groups by treatment with methanolic HCl, followed by intramolecular transesterification promoted by trifluoroacetic acid, completed the synthesis of the tricyclic benzochromenone backbone and afforded the methyl ester of hapten ALa, compound 8, in 97% yield. To complete the synthesis of hapten ALa, only the hydrolysis of the methyl ester moiety of 8 was required, which was initially carried out under basic conditions (LiOH in THF-H 2 O at room temperature, rt). However, under these conditions, the central lactone group of the benzochromenone core was partially opened, requiring acid treatment of the reaction crude to reconstruct the tricyclic ring system. It proved more convenient to carry out this transformation using enzymatic hydrolysis, so a lipase from Candida antarctica immobilized on an acrylic resin was used to hydrolyze the methyl ester group, providing hapten ALa in practically quantitative yield. Upon completion of the synthesis of hapten ALa, its carboxylic group was activated by forming the corresponding N-hydroxysuccinimidyl ester. This transformation was carried out under conventional activation conditions, with N-(3-dimethylaminopropyl)-Nethylcarbodiimide hydrochloride (EDC·HCl) and N-hydroxisuccinimide (NHS) in N,Ndimethylformamide (DMF) at rt, yielding the corresponding N-hydroxysuccinimidyl ester, ALa-NHS, in good yield. The activated hapten was extracted essentially pure from the reaction as judged by 1 H NMR, so it was further used without additional purification by column chromatography. NMR spectra of all of the intermediates and the hapten can be found in the Supplementary Materials file.
Hapten ALb was synthesized following a similar procedure as hapten ALa, except that in this case the tricyclic benzochromenone core was built first, with the hydroxyl groups appropriately protected to allow subsequent incorporation of the spacer arm at the required C-3 position. As shown in Scheme 2, the synthesis of the benzochromenone ring system began with the palladium-catalyzed cross-coupling reaction between the aryl boronic acid 6 and the previously reported bromobenzaldehyde 9 [24,25]. This coupling was carried out under conditions similar to those previously used for the conversion of 4 and 6 into 7, obtaining the biphenyl-2-carbaldehyde 10 in 77% yield. The formyl group was further oxidized to the carboxylic group under Pinnick oxidation conditions to afford the biphenyl-2-carboxylic acid 11, which was then treated with methanolic HCl at 55 • C to promote deprotection of the MOM groups and further intramolecular esterification, thus completing the formation of the tricyclic benzochromenone system. Under these conditions, both sequential processes worked extremely well, affording 12 in practically quantitative yield. O-alkylation of the phenol-like hydroxyl group at the C-3 position of 12 with methyl 5-bromovalerate, using Cs 2 CO 3 in DMF as base, gave the O-alkylated derivative 13 in 94% yield. The methyl ester of 13 was further converted to the corresponding carboxylic group under enzymatic hydrolytic conditions, yielding 14 also in high yield. The hapten ALb was first obtained by hydrogenolysis of both benzyl ether groups of 14 using 5% Pd on activated carbon as catalyst. With hapten ALb in hand, we activated the carboxylic group using the carbodiimide-NHS procedure as was done for hapten ALa. However, the overall yield from these two processes was low, most likely motivated by an intermolecular esterification reaction between a hydroxyl group and the aliphatic active ester that resulted in the spontaneous formation of a transparent thin film, a polyester polymer, on the flask walls. By reversing the order of these steps, i.e., by activating the carboxylic group first and then releasing the hydroxyl groups, a much better result was obtained. Thus, treatment of carboxylic acid 14 with EDC and NHS as before, followed by hydrogenolysis of the benzyl ether moieties of the resulting N-hydroxysuccinimidyl ester 15 with 5% Pd on activated carbon in acetone, gave the desired N-hydroxysuccinimidyl ester of hapten ALb, ALb-NHS ester, with a very high overall yield. As in the case of the active ester of hapten ALa, the ALb-NHS ester was extracted essentially pure from the reaction as judged by 1 H NMR, so it was further used without additional purification by column chromatography. NMR spectra of all of the intermediates and the hapten can be found in the Supplementary Materials file.

Bioconjugate Preparation
Bioconjugates of haptens ALa and ALb were prepared by the active ester method. The activated haptens were dissolved in dimethyl sulfoxide (DMSO) instead of DMF to improve the solubility. Moreover, the number of hapten equivalents required for efficiently labelling the studied proteins was higher than usual. Commonly, 20-fold hapten-to-protein molar excess for bovine serum albumin (BSA), and 10-fold excess for ovalbumin (OVA) and horseradish peroxidase (HRP) are usually employed in our laboratory. For these haptens, 40-fold and 15-fold excess was used for BSA and HRP conjugates, respectively. Moreover, extremely slow addition of the hapten over the protein solution was required. These concentrations and procedures were necessary, probably due to low hapten solubility in buffer and potential intermolecular polymerization reactions that inactivate the hapten. The obtained bioconjugates were purified by size-exclusion chromatography and characterized Toxins 2021, 13, 883 6 of 16 by matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF/MS) analysis. The two BSA conjugates had similar hapten densities, with haptento-protein molar ratios of 15.2 and 18.6 for BSA-ALa and BSA-ALb, respectively, which is considered optimal for immunogens-excessive molar ratios could lead to low protein solubility, and higher or lower hapten densities could be counter-productive for highaffinity antibody generation. Regarding ovalbumin (OVA) conjugates, molar ratios were lower than those of BSA conjugates-around 3 for both haptens -, as it is desirable for coating conjugates to enhance the competitive reaction with the target analyte. Finally, the hapten densities of the enzyme tracers were estimated to be 2.0 and 2.2 for haptens ALa and ALb, respectively, which is within the expected range for HRP conjugates. The MALDI spectra of the prepared bioconjugates can be seen in Figure 2.
EER REVIEW 6 of 16 could lead to low protein solubility, and higher or lower hapten densities could be counter-productive for high-affinity antibody generation. Regarding ovalbumin (OVA) conjugates, molar ratios were lower than those of BSA conjugates-around 3 for both haptens -, as it is desirable for coating conjugates to enhance the competitive reaction with the target analyte. Finally, the hapten densities of the enzyme tracers were estimated to be 2.0 and 2.2 for haptens ALa and ALb, respectively, which is within the expected range for HRP conjugates. The MALDI spectra of the prepared bioconjugates can be seen in Figure  2.

Assessment of the Immune Response
Four polyclonal antibodies were generated in this study, two from each BSA-hapten conjugate. To evaluate the immune response to the prepared synthetic haptens, binding of the antibodies to the homologous conjugate-the conjugate with the same hapten that was used to generate the corresponding antibody-was studied by checkerboard competitive ELISA, using the direct and the indirect assay formats.
Concerning direct assays, the IC50 values for AOH of the obtained antibodies were in the low nanomolar range (Table 1). ALa-type antibodies showed equal or similar IC50 values for AOH and AME. In particular, antibody ALa#1 showed very high affinity-IC50 values were 2.2 nM-for both mycotoxins, and the cross-reactivity (CR) values of antibodies ALa#1 and ALa#2 for AME were 100% and 199%, respectively. These are the first reported polyclonal antibodies with equivalent recognition to both Alternaria toxins. To date, only one monoclonal antibody with such specificity has been published [15]. In con-

Assessment of the Immune Response
Four polyclonal antibodies were generated in this study, two from each BSA-hapten conjugate. To evaluate the immune response to the prepared synthetic haptens, binding of the antibodies to the homologous conjugate-the conjugate with the same hapten that was used to generate the corresponding antibody-was studied by checkerboard competitive ELISA, using the direct and the indirect assay formats.
Concerning direct assays, the IC 50 values for AOH of the obtained antibodies were in the low nanomolar range (Table 1). ALa-type antibodies showed equal or similar IC 50 values for AOH and AME. In particular, antibody ALa#1 showed very high affinity-IC 50 values were 2.2 nM-for both mycotoxins, and the cross-reactivity (CR) values of antibodies ALa#1 and ALa#2 for AME were 100% and 199%, respectively. These are the first reported polyclonal antibodies with equivalent recognition to both Alternaria toxins. To date, only one monoclonal antibody with such specificity has been published [15]. In contrast, ALb-type antibodies bound AOH with high affinity, but their recognition for AME was negligible-CR values were below 1% ( Table 1). The IC 50 values to AOH of these specific antibodies were 1.2 nM, an affinity comparable to that of previously published polyclonal antibodies [11,13,16]. The position of the spacer arm in hapten ALa provided a closer mimic of the alkylated hydroxyl group of AME (C-9 position), whereas in hapten ALb the hydroxyl groups at C-7 and C-9 were unsubstituted, as in the molecule of AOH ( Figure 1). Therefore, display of the hydroxyl group at C-9 was maximized in hapten ALb, which explains the much lower affinity of ALb-type antibodies for AME compared to AOH. Regarding the indirect assay format, the four antibodies bound the corresponding homologous coating conjugate. As observed with the direct format, the ALa-derived antibodies recognized AOH and AME, whereas the ALb-derived antibodies were more specific to AOH ( Table 1). The IC 50 values were consistent with previously published results for indirect competitive ELISA with polyclonal antibodies [10,11,16]. Our strategy to prepare immunizing haptens with opposite linker tethering sites clearly demonstrated that the linker position strongly determines the specificity of antibodies to these Alternaria mycotoxins.

Assessment of Heterologous Conjugates
Heterologous conjugates constitute a well-known strategy for improving the sensitivity of immunoassays. To further characterize the generated antibodies, competitive assays were carried out using the heterologous conjugate, i.e., assay conjugates of haptens ALa and ALb for ALb-and ALa-type antibodies, respectively. In the direct assay format, low binding to the heterologous tracer-with the linker on the opposite side of the AOH molecule compared to the immunizing conjugate-was observed (A max values were below 0.6). In contrast, the change in the linker attachment site was not detrimental to hapten recognition in the indirect format, as the four antibodies bound the corresponding heterologous coating conjugate (Table 2). Reasonably, higher antibody and/or conjugate concentrations were required with the heterologous conjugates to reach sufficient signal. The obtained IC 50 values using the heterologous coating conjugate were mostly lower than those obtained with the homologous assays. Anyway, CR values did not significantly change with heterologous conjugates.

Conclusions
In this study, two de novo synthesized and purified AOH haptens were comprehensively characterized by spectrometric methods, and bioconjugates with unique structure and composition were prepared for the first time. In this perspective, it is worth noting the challenges of obtaining stable enzyme tracers with high activity. This matter was most likely caused by the chemical characteristics of Alternaria toxins and their haptens, which could explain why no direct competitive immunoassays for these mycotoxins have been reported up to now. Once this issue was overcome, the resultant immunoreagents were thoroughly investigated utilizing both direct and indirect competitive ELISA, as well as homologous and heterologous conjugates. Remarkably, antibodies capable of binding AOH and AME with affinities in the low nanomolar range were eventually generated from both haptens. Given that the levels of these mycotoxins are not yet regulated, both specific and generic antibodies are relevant. Our findings showed that hapten ALa, with the linker at the methylated hydroxyl group in AME (C-9 position), was particularly well-suited for producing antibodies that recognized similarly both toxins, whereas antibodies generated from hapten ALb, with the spacer arm at the hydroxyl group in C-3 position, primarily bound AOH. In contrast to previous one-pot hapten synthesis and bioconjugation procedures, the strategy described here for producing AOH haptens with alternative linker tethering sites not only enabled high-affinity antibodies with different specificities, but it may also help to improve the sensitivity of immunoassays to Alternaria mycotoxins by using site heterologous haptens.

Preparation of methyl 5-((2,2-dimethyl-4-oxo-5-(((trifluoromethyl)sulfonyl)oxy)-4Hbenzo[d][1,3]dioxin-7-yl)oxy)pentanoate (4)
Triflic anhydride (230 µL, 1.369 mmol, 1.5 equiv) was added to a solution of phenol 3 (296 mg, 0.913 mmol) in anhydrous pyridine (4.5 mL) at 0 • C under nitrogen. The reaction mixture was allowed to warm to rt and stirred for 20 h, then cooled down to 0 • C and treated with a saturated aqueous solution of NaHCO 3 , stirred for a few minutes at rt and then extracted with Et 2 O. The organic layers were washed with water, a 1% (w/v) aqueous solution of CuSO 4 and brine, dried over anhydrous MgSO 4 and concentrated at reduced pressure. The obtained residue was chromatographed on silica gel, using hexane-EtOAc mixtures from 9:1 to 8:2 as eluent, to give aryl triflate 4 (378.8 mg, 91%) as a white semisolid. IR (ATR) ν max (cm  (7) (i) Preparation of boronic acid 6. A solution of n-BuLi in hexane (1.3 M, 336 µL, 0.436 mmol, 1.05 equiv) was dropwise added to a solution of aryl bromide 5 (122.3 mg, 0.420 mmol) in anhydrous THF (2.5 mL) at −78 • C under nitrogen. The reaction mixture was stirred at this temperature for 40 min, B(O i Pr) 3 (322 µL, 1.386 mmol, 3.3 equiv) was then added and the mixture stirred for 1.5 h. After this time, the dry ice bath was replaced by an ice bath and the mixture treated with an aqueous saturated solution of NH 4 Cl (0.7 mL), then diluted with water and extracted with Et 2 O. The organic layers were washed with brine, dried over anhydrous Na 2 SO 4 and concentrated under reduced pressure to give boronic acid 6 (100.0 mg, 93%) as a thick oil that was immediately used in the next reaction without further purification since it is relatively prone to protodeboronation [26]. 1  (ii) Coupling reaction between aryl triflate 4 and boronic acid 6. A mixture of the above obtained boronic acid 6 (47.4 mg, 0.185 mmol), aryl triflate 4 (41.6 mg, 0.091 mmol), powdered K 2 CO 3 (43.2 mg, 0.312 mmol) and Pd(PPh 3 ) 4 (11.4 mg, 9.9 µmol) under nitrogen was dissolved in anhydrous DMF (1.2 mL), previously degassed by three freeze-vacuumthaw cycles. The mixture was heated at 93 • C and stirred at this temperature for 24 h. The mixture was cooled to rt, quenched with water and extracted with EtOAc. The combined organic layers were successively washed with water, a 1.5% (w/v) aqueous solution of LiCl and brine, and dried over anhydrous MgSO 4 . The obtained residue after evaporation of the solvent was chromatographed on silica gel, using hexane-EtOAc 8:2 as eluent, to afford biaryl compound 7 (35.2 mg, 75%) as a yellowish oil. 1  Lipase from Candida antarctica immobilized on acrylic resin (23 mg) was added to a solution of methyl ester 8 (16.6 mg, 0.0446 mmol) in a 4:1 mixture of 100 mM sodium phosphate buffer (pH 7.4) and THF (1.5 mL) at 30 • C. The resulting heterogeneous mixture was smoothly stirred for 24 h at rt and then filtered to separate the enzyme. The filtrated and washing THF phases were combined, diluted with EtOAc, washed with brine, dried over anhydrous MgSO 4 , and concentrated in vacuo to afford hapten ALa (14.9 mg, 93%) as a white amorphous solid. 1

Preparation of 3,5-bis(benzyloxy)-2 ,4 -bis(methoxymethoxy)-6 -methyl-[1,1 -biphenyl]-2-carbaldehyde (10)
An ampoule containing a mixture of freshly prepared aryl boronic acid 6 (104.5 mg, 0.408 mmol, 2 equiv), 2,4-bis(benzyloxy)-6-bromobenzaldehyde 9 (80.9 mg, 0.204 mmol), K 2 CO 3 (63.6 mg, 0.460 mmol, 2.2 equiv) and Pd(PPh 3 ) 4 (26.6 mg, 0.023 mmol, 0.1 equiv) in anhydrous DMF (2 mL) was exhaustively degassed by freeze-thaw cycles. The ampoule was closed under vacuum and heated at 95 • C for 19 h. After cooling, the ampoule was opened and the reaction mixture was poured onto water and extracted with EtOAc. The combined organic extracts were washed with water, a 1.5% (w/v) aqueous solution of LiCl and brine, dried under anhydrous MgSO 4 and concentrated under vacuum. The resulting crude reaction mixture was chromatographed on silica gel to give biaryl-2-carbaldehyde 10 (93.6 mg, 77%) as a viscous yellowish oil. 1  The mixture was thermostated at 55 • C in an oil bath and stirred at this temperature for 24 h. After this time, the mixture was cooled to rt, diluted with a concentrated aqueous solution of NaHCO 3 and extracted with Et 2 O. The organic phase was washed with brine, dried over anhydrous MgSO 4 and concentrated under vacuum to give 7,9-bis(benzyloxy)alternariol 12 (54.7 mg, 98%) as an amorphous whitish solid. The crude reaction product thus obtained was sufficiently pure, as judged by its NMR spectroscopic data, to be used in the next step without further purification. 1 (13) Methyl bromovalerate (29.5 mg, ca. 22 µL, 0.151 mmol, 1.1 equiv) was added via syringe to a stirred suspension of Cs 2 CO 3 (57.8 mg, 0.177 mmol, 1.3 equiv) and phenol 12 (60.1 mg, 0.137 mmol) in anhydrous DMF (2 mL) at rt under nitrogen and the mixture was stirred for 19 h. The resulting pale yellowish reaction mixture was diluted with water and extracted with EtOAc. The combined organic extracts were washed successively with water, a 1.5% (w/v) aqueous solution of LiCl and brine, dried over anhydrous MgSO 4 and concentrated under reduced pressure. The crude reaction product was purified by chromatography on silica gel, using CHCl 3 as eluent, to afford the O-alkylated product 13 (71.4 mg, 94%) as a pale yellowish semi-solid. 1  buffer, pH 7.4, as eluent. Fractions containing BSA or OVA conjugates were pooled and diluted with elution buffer to a final concentration of 1 mg/mL. BSA conjugate solutions were passed through 0.45 µm sterile filters. BSA and OVA conjugate solutions were stored at −20 • C. HRP conjugate solutions were 1:1 diluted with PBS containing 1% BSA (w/v) and 0.02% (w/v) thimerosal and stored at 4 • C. The hapten-to-protein molar ratio of the prepared conjugates was determined by MALDI-TOF-MS and running BSA, OVA, and HRP for reference in the same plate, as previously described [27].
Animal manipulation was performed according to Spanish laws (RD1201/2005 and law 32/2007) and the European Directive 2010/63EU regarding the protection of experimental animals. Polyclonal antibodies to AOH and AME were obtained from the sera of immunized animals. Briefly, two female New Zealand white rabbits-weighing 2 kg at the beginning of the experiment-were immunized by four periodic subcutaneous injections of a 1:1 water-in-oil emulsion containing 300 µg of the BSA-hapten conjugate. The inoculum was prepared with complete Freund's adjuvant for the first injection and with incomplete Freund's adjuvant for subsequent injections. Boosts were applied with 21-day intervals. Animals were exsanguinated by intracardiac puncture 10 days after the last injection, and the blood was left overnight in the refrigerator at 4 • C for coagulation. Sera were separated from cells by centrifugation (3000× g, 20 min). Finally, immunoglobulins were partially purified by precipitation twice with one volume of cold saturated (3.9 M) ammonium sulphate solution. Antibodies were stored at 4 • C as precipitates.

Competitive ELISA Procedures
Immunoassays were carried out by competitive ELISA using the capture antibodycoated direct format and the conjugate-coated indirect format. After each incubation step, plates were washed three times with a 150 mM NaCl solution containing 0.05% (v/v) Tween 20. For direct assays, microplates were coated by overnight incubation at 4 • C with 100 µL per well of GAR solution (1 µg/mL) in 50 mM carbonate-bicarbonate buffer, pH 9.6. For indirect assays, microwells were coated with 100 µL per well of OVA-hapten conjugate solution in the same coating buffer, and overnight incubation at rt. The competitive reaction was performed by mixing in each well 50 µL of analyte solution in PBS with 50 µL of antibody dilution or enzyme tracer solution in PBS containing 0.05% (v/v) of Tween-20, and incubating 1 h at rt. For indirect assays, 100 µL per well of GAR-HRP diluted 1/10,000 in PBS containing 0.05% (v/v) of Tween-20 and 10% (v/v) of adult bovine serum was added. Signal was obtained using 100 µL per well of TMB as the chromogenic enzyme substrate and incubation at rt during 10 min. Finally, 100 µL of 1 M H 2 SO 4 was added and the absorbance was read at 450 nm using 650 nm as reference wavelength.
Standard mycotoxin solutions were obtained by serially diluting in buffer the most concentrated standard solution, which was prepared from a concentrated stock solution in DMF. Eight-point standard curves were built using those solutions and a blank sample.