Preparation of an Immunoaffinity Column Based on Bispecific Monoclonal Antibody for Aflatoxin B1 and Ochratoxin A Detection Combined with ic-ELISA

A novel and efficient immunoaffinity column (IAC) based on bispecific monoclonal antibody (BsMAb) recognizing aflatoxin B1 (AFB1) and ochratoxin A (OTA) was prepared and applied in simultaneous extraction of AFB1 and OTA from food samples and detection of AFB1/OTA combined with ic-ELISA (indirect competitive ELISA). Two deficient cell lines, hypoxanthine guanine phosphoribosyl-transferase (HGPRT) deficient anti-AFB1 hybridoma cell line and thymidine kinase (TK) deficient anti-OTA hybridoma cell line, were fused to generate a hybrid-hybridoma producing BsMAb against AFB1 and OTA. The subtype of the BsMAb was IgG1 via mouse antibody isotyping kit test. The purity and molecular weight of BsMAb were confirmed by SDS-PAGE method. The cross-reaction rate with AFB2 was 37%, with AFG1 15%, with AFM1 48%, with AFM2 10%, and with OTB 36%. Negligible cross-reaction was observed with other tested compounds. The affinity constant (Ka) was determined by ELISA. The Ka (AFB1) and Ka (OTA) was 2.43 × 108 L/mol and 1.57 × 108 L/mol, respectively. Then the anti-AFB1/OTA BsMAb was coupled with CNBr-Sepharose, and an AFB1/OTA IAC was prepared. The coupling time and elution conditions of IAC were optimized. The coupling time was 1 h with 90% coupling rate, the eluent was methanol–water (60:40, v:v, pH 2.3) containing 1 mol/L NaCl, and the eluent volume was 4 mL. The column capacities of AFB1 and OTA were 165.0 ng and 171.3 ng, respectively. After seven times of repeated use, the preservation rates of column capacity for AFB1 and OTA were 69.3% and 68.0%, respectively. The ic-ELISA for AFB1 and OTA were applied combined with IAC. The IC50 (50% inhibiting concentration) of AFB1 was 0.027 ng/mL, the limit of detection (LOD) was 0.004 ng/mL (0.032 µg/kg), and the linear range was 0.006 ng/mL~0.119 ng/mL. The IC50 of OTA was 0.878 ng/mL, the LOD was 0.126 ng/mL (1.008 µg/kg), and the linear range was 0.259 ng/mL~6.178 ng/mL. Under optimum conditions, corn and wheat samples were pretreated with AFB1-OTA IAC. The recovery rates of AFB1 and OTA were 95.4%~105.0% with ic-ELISA, and the correlations between the detection results and LC-MS were above 0.9. The developed IAC combined with ic-ELISA is reliable and could be applied to the detection of AFB1 and OTA in grains.


Introduction
Mycotoxins are natural secondary metabolites produced by filamentous fungi under suitable conditions, among which aflatoxin B 1 (AFB 1 ) and ochratoxin A (OTA) are the most toxic and exist widely in grains [1][2][3][4]. In 2012, aflatoxins were listed in the group 1 classification (carcinogenic to humans) by the International Agency for Research on Cancer (IARC), and OTA was classified as group 2B (possibly carcinogenic to humans) by IARC in

Apparatus
ELISA plates were washed using a Multiskan MK2 microplate washer (Thermo Fisher Scientific Inc., Waltham, MA, USA). Absorbances of ELISA were measured on a Multiskan MK3 microplate reader (Thermo Fisher Scientific Inc., Waltham, MA, USA). Absorbances of antibody concentrations were measured on a NanoDrop 2000 c Ultraviolet spectrometer (Thermo Fisher Scientific Inc., Waltham, MA, USA).The LC-MS assay was conducted using an Agilent 6400 series LC system and an ECLIPS PLUS C 18 (2.1 × 100 mm, 1.8 µm) (Agilent Technologies, Santa Clara, CA, USA) with a Triple Quadrupole Mass Spectrometer.

Mutagenesis of HGPRT and TK Deficient Hybridoma Cell Lines
The chemical mutagen scheme was performed as follows [28,45,46]: AFB 1 hybridoma cell line E 4 was cultured in HAT medium for 2 d before mutagenic treatment in a humidified 5% CO 2 incubator at 37 • C. The cells were washed twice with PBS and cultured in HT medium for 2 d to restore the cells to normal morphology and were cultured in HAT medium for 24 h. After washing the cells with PBS twice, the cells were mutagenized in different concentrations of MNNG (5.0 µg/mL, 10.0 µg/mL, 20.0 µg/mL, and 40.0 µg/mL). The mutagenic treatment time was 2 h, 4 h, 6 h and 10 h. At the same time, the control group was cultured in 1‰ DMSO complete medium (RPMI-1640 culture media containing 20% fetal bovine serum) without MNNG. The cells were washed twice with PBS to finish the mutagenesis and cultured in complete medium. Finally, the cells were cultured for 3 d and the survival rate of the cells was determined by trypan blue staining, and the combination of MNNG concentration and mutagenic treatment time resulting in a survival rate of about 70% were selected as the optimal mutation parameters. OTA hybridoma cell line B 3 was mutated with MMS (1.0 µg/mL, 5.0 µg/mL, 10.0 µg/mL, and 20.0 µg/mL) and for 2 h, 4 h, 6 h and 10 h by the same steps as above.

Screening of HGPRT and TK Deficient Hybridoma Cell Lines
After mutagenic treatment with MNNG, AFB 1 hybridoma cells E 4 were collected and incubated in a 6-well culture plate in selective semi-solid media containing 6-TG (50 µg/mL). After 10 d, the visible white clonal cell clusters were placed into a 96-well culture plate and cultured with 200 µL complete media containing 6-TG (50 µg/mL). The cell supernatants were analyzed by ELISA to screen positive clones which could recognize and bound to AFB 1 . Then the cells were tested by HAT media to confirm their HAT sensitivity. If the cell line could not grow in HAT media, it was HGPRT-deficient, and was named E 4 -HGPRT − . Via the same steps, the TK-deficient OTA hybridoma cell line named B 3 -TK − was screened with 5-BrdU (30 µg/mL).
2.5. Generation and Characterization of BsMAb 2.5.1. Generation of BsMAb E 4 -HGPRT − and B 3 -TK − were fused under the effect of PEG 4000 [47]. The fused cells were cultured in 96-well culture plates with HAT media for 10 d. The cell supernatants were tested by ic-ELISA to confirm the presence of BsMAb which could recognize and bound to AFB 1 and OTA simultaneously.
The selected tetra-doma cell lines were cloned by 4 rounds of limiting dilution assays, and then inoculated into female Balb/c mice that had been primed with 500 µL of sterile liquid paraffin. Ascites was collected from the Balb/c mice and purified with Protein G to obtain BsMAb. The concentration of BsMAb was measured by a NanoDrop 2000 c ultraviolet spectrometer. The subtype of BsMAb was determined by Pierce Rapid Isotyping Kits-Mouse. The purity and molecular weight of BsMAb were estimated by the SDS-PAGE method.

Antibody Specificity Determination
The cross-reactivity (CR) could be used as an index to evaluate the specificity of the anti-AFB 1 /OTA BsMAb. Inhibition curves were fitted through ic-ELISA data to determine the IC 50 (50% inhibiting concentration) of each common mycotoxin contaminant (aflatoxin B 1 , aflatoxin B 2 , aflatoxin G 1 , aflatoxin G 2 , aflatoxin M 1 , aflatoxin M 2 , ochratoxin A, ochratoxin B, ochratoxin C, zearalenone, deoxynivalenol, T-2 toxin and fumonisin B 1 ). The CR was calculated using the following equation: CR = IC 50 of AFB 1 or OTA IC 50 of common mycotoxin contaminant ×100% (1)

Preparation of BsMAb Based IAC
According to the instruction, the procedures of IAC preparation was as follow: 0.5 g of CNBr-Sepharose 4B powder was swelled with 3 mL of HCl (1 mM) and washed with 20 mL of coupling buffer (0.1 M NaHCO 3 , pH 8.3) 5 times. Then, the Sepharose gel was mixed with 2 mL anti-AFB 1 /OTA BsMAb solution (9.61 mg in coupling buffer) and gently stirred at 25 • C. The concentration of unconjugated BsMAb in supernatant was determined by NanoDrop 2000 c ultraviolet spectrometer every 0.5 h to calculate the coupling efficiency. The optimal coupling time was determined when the coupling efficiency was over 90%.
The coupling efficiency was calculated as follows: The agarose gel was redissolved in 50 mL blocking buffer (0.1 M Tris-HCl, pH 8.0) to block the free active sites at 25 • C for 2 h, and washed with 20 mL HAc-NaAc buffer (0.1 M, pH 4.0) and 20 mL Tris-HCl buffer (0.1 M Tris-HCl, pH 8.0) alternately for 4 cycles. Finally, 1 mL of immune gel prepared above was transferred to a polyethylene column and balanced with 20 mL loading buffer (0.1 M PBS, pH 7.4). The IAC prepared was stored with PBS containing 0.01% Merthiolate sodium (v/w) at 4 • C until use.

Optimization of Elution Conditions and Evaluation of the IAC Capacity
Mixed standard solution containing 50 ng AFB 1 and 50 ng OTA was diluted in 15 mL loading buffer and drawn through the IAC. Then 20 mL of loading buffer was applied to wash the IAC. Finally, elution buffer was used to elute the analyte bound to the column. The collected eluent diluted twice with PBS was detected by ic-ELISA and recovery rates were evaluated.
Under the optimal conditions, 15 mL of loading buffer containing excessive mixed standard of AFB 1 (300 ng) and OTA (300 ng) was passed through the IAC. The eluent diluted with PBS was detected by ic-ELISA. The column was used repeatedly for several cycles, and the column capacity and preservation rates were calculated and compared. The column capacity/preservation rate was expressed as follows: Column Capacity = Concentration of AFB 1 or OTA × Loading Amount Immune Gel Volume Preservation Rate = Column Capacity for Several Cycles Column Capacity for the First Treatment (4)

ic-ELISA Combined with IAC
The collected eluent was detected by the ic-ELISA developed and optimized in our laboratory. The parameters of ic-ELISA are shown in Table 1 and the procedures are as follows. Coating antigen solution was coated in a 96-well polystyrene microtiter plate with 100 µL/well at 4 • C for 14 h, then washed with 300 µL/well wash solution (PBS containing 0.05% Tween-20) twice. 120 µL/well blocking buffer was used to block uncoated sites at 37 • C for 3 h, and then the plate was dried in a draught drying cabinet at 37 • C. Then, 50 µL MAb solution and 50 µL AFB 1 /OTA standard solution of gradient concentrations or the collected eluent diluted twice with PBS were added to each well and incubated, and the plate was washed 4 times. 100 µL/well HPR-IgG was added for incubation, and then washed 4 times. After washing, 100 µL/well of TMB chromogenic solution was added into each well and incubated for 10 min at 37 • C, and 50 µL/well of 2 M H 2 SO 4 was added to terminate the chromogenic reaction. The absorbance at 450 nm (A 450nm ) was measured. The A 450nm of the well without standard solution was as B 0 and with standard solution was as B. The inhibition curve was fitted by Origin 8.5 software adopting B/B 0 versus the concentration of AFB 1 or OTA, and the IC 50 was estimated ( Figure 1). The limit of detection (LOD) was defined as the IC 10 value. The linear range was taken as IC 20 to IC 80 . The IC 50 of AFB 1 was 0.027 ng/mL, the LOD was 0.004 ng/mL (0.032 µg/kg), and the linear range was 0.006 ng/mL~0.119 ng/mL. The IC 50 of OTA was 0.878 ng/mL, the LOD was 0.126 ng/mL (1.008 µg/kg), and the linear range was 0.259 ng/mL~6.178 ng/mL. The LODs of the ic-ELISA to AFB 1 and OTA were obviously lower than the maximum levels. Therefore, this method was suitable for quantitative detection of AFB 1 and OTA. measured. The A450nm of the well without standard solution was as B0 and with standard solution was as B. The inhibition curve was fitted by Origin 8.5 software adopting B/B0 versus the concentration of AFB1 or OTA, and the IC50 was estimated ( Figure 1). The limit of detection (LOD) was defined as the IC10 value. The linear range was taken as IC20 to IC80. The IC50 of AFB1 was 0.027 ng/mL, the LOD was 0.004 ng/mL (0.032 µg/kg), and the linear range was 0.006 ng/mL~0.119 ng/mL. The IC50 of OTA was 0.878 ng/mL, the LOD was 0.126 ng/mL (1.008 µg/kg), and the linear range was 0.259 ng/mL~6.178 ng/mL. The LODs of the ic-ELISA to AFB1 and OTA were obviously lower than the maximum levels. Therefore, this method was suitable for quantitative detection of AFB1 and OTA.

Analysis of Corn and Wheat Samples
The pretreatment method for AFB1 and OTA analysis was carried out as follows [43]. Corn and wheat samples were evenly crushed and passed through a 40-mesh sieve. 5.0 g of sample was put into a 50 mL polypropylene centrifuge tube, and 25 mL of methanolwater (70:30, v:v, containing 2.4% NaCl) was added, and shaken periodically for 5 min. After centrifuging at 4000 rpm for 10 min, the supernatant was collected through a 0.45 µm PTEE membrane filter to obtain the sample extract. Then 5 mL of the sample extract was diluted with 10 mL loading buffer and drawn through the AFB1/OTA-IAC. Nonspecifically adsorbed impurities were washed off using 20 mL loading buffer. Then, elution buffer optimized in 2.7 was used to elute, and diluted twice with PBS before detection by ic-ELISA. Concurrently, the samples were detected with LC-MS.

Analysis of Corn and Wheat Samples
The pretreatment method for AFB 1 and OTA analysis was carried out as follows [43]. Corn and wheat samples were evenly crushed and passed through a 40-mesh sieve. 5.0 g of sample was put into a 50 mL polypropylene centrifuge tube, and 25 mL of methanol-water (70:30, v:v, containing 2.4% NaCl) was added, and shaken periodically for 5 min. After centrifuging at 4000 rpm for 10 min, the supernatant was collected through a 0.45 µm PTEE membrane filter to obtain the sample extract. Then 5 mL of the sample extract was diluted with 10 mL loading buffer and drawn through the AFB 1 /OTA-IAC. Nonspecifically adsorbed impurities were washed off using 20 mL loading buffer. Then, elution buffer optimized in 2.7 was used to elute, and diluted twice with PBS before detection by ic-ELISA. Concurrently, the samples were detected with LC-MS.
LC-MS conditions were modified from published methods [49] as follows: an Agilent LC system coupled with an Agilent 6400 series triple Quadrupole mass spectrometer was used for the confirmatory analysis. An Agilent ECLIPS PLUS C 18 (2.1 × 100 mm, 1.8 µm) with solvent consisted of 0.1% formic acid (mobile phase A) and acetonitrile (mobile phase B, 0-2 min, 30%; 2-5 min, 90%; 5-10 min, 5%) at a flow rate of 0.30 mL/min. The temperature was 40 • C and the injection volume was 5 µL. The analysis was performed using positive-ion electrospray interface (ESI) with a multiple reaction monitoring (MRM) mode. The retention times of AFB 1 and OTA were 3.339 min and 3.94 min, respectively. Blank corn and wheat samples without AFB 1 and OTA, having been confirmed by LC-MS, were used to spike recovery experiments. AFB 1 and OTA of different concentrations (4 µg/kg, 10 µg/kg, 20 µg/kg, 50 µg/kg, 100 µg/kg) were spiked into the blank samples. Eight grain samples were randomly selected from a local market, including four corn samples and four wheat samples. After pretreatment, the sample eluents were detected by ic-ELISA, and the results were compared with those of LC-MS.

Mutagenesis and Screening of HGPRT and TK Deficient Hybridoma Cell Lines
The AFB 1 hybridoma cell line E 4 and OTA hybridoma cell line B 3 were applied for construction of a HGPRT and TK deficient mutant, respectively. When the concentration of MNNG was 10 µg/mL and the treatment time was 6 h, the cell survival percentage was 69.9% (Figure 2a), which was determined as the optimal mutation condition of E 4 cells. When the concentration of MMS was 5 µg/mL and the treatment time was 4 h, the cell survival percentage was 70.2% (Figure 2b), and the optimal mutation condition for B 3 cells was determined. As mutagens, MNNG and MMS have strong mutagenic effect on cells. Over-high concentration and overlong treating time will reduce the activity of cells, or even result in apoptosis.
LC system coupled with an Agilent 6400 series triple Quadrupole mass spectrometer was used for the confirmatory analysis. An Agilent ECLIPS PLUS C18 (2.1 × 100 mm, 1.8 µm) with solvent consisted of 0.1% formic acid (mobile phase A) and acetonitrile (mobile phase B, 0-2 min, 30%; 2-5 min, 90%; 5-10 min, 5%) at a flow rate of 0.30 mL/min. The temperature was 40 °C and the injection volume was 5 µL. The analysis was performed using positive-ion electrospray interface (ESI) with a multiple reaction monitoring (MRM) mode. The retention times of AFB1 and OTA were 3.339 min and 3.94 min, respectively.
Blank corn and wheat samples without AFB1 and OTA, having been confirmed by LC-MS, were used to spike recovery experiments. AFB1 and OTA of different concentrations (4 µg/kg, 10 µg/kg, 20 µg/kg, 50 µg/kg, 100 µg/kg) were spiked into the blank samples. Eight grain samples were randomly selected from a local market, including four corn samples and four wheat samples. After pretreatment, the sample eluents were detected by ic-ELISA, and the results were compared with those of LC-MS.

Mutagenesis and Screening of HGPRT and TK Deficient Hybridoma Cell Lines
The AFB1 hybridoma cell line E4 and OTA hybridoma cell line B3 were applied for construction of a HGPRT and TK deficient mutant, respectively. When the concentration of MNNG was 10 µg/mL and the treatment time was 6 h, the cell survival percentage was 69.9% (Figure 2a), which was determined as the optimal mutation condition of E4 cells. When the concentration of MMS was 5 µg/mL and the treatment time was 4 h, the cell survival percentage was 70.2% (Figure 2b), and the optimal mutation condition for B3 cells was determined. As mutagens, MNNG and MMS have strong mutagenic effect on cells. Over-high concentration and overlong treating time will reduce the activity of cells, or even result in apoptosis. After MNNG mutation treatment, the E4 cells were inoculated in 6-TG medium for selective screening. The screened E4-HGPRTcell line gradually formed colonies with the ability to fission and grow. Finally, stable inherited deficient cells E4-HGPRTwere obtained. Similarly, stable inherited deficient cells B3-TKwere screened in 5-BrdU. Figure 3 depict the growth curves of E4-HGPRTand B3-TKin screening reagent (6-TG and 5-BrdU), complete media and HAT media after mutagenic treatment. The unmutated cells could not survive in screening reagent (Figure 3a,d). In complete medium culture, both normal cell line and mutant could grow normally (Figure 3b,e). The cell apoptosis in HAT medium was adopted as an indication of HGPRT or TK deficiency. The growth curve After MNNG mutation treatment, the E 4 cells were inoculated in 6-TG medium for selective screening. The screened E 4 -HGPRT − cell line gradually formed colonies with the ability to fission and grow. Finally, stable inherited deficient cells E 4 -HGPRT − were obtained. Similarly, stable inherited deficient cells B 3 -TK − were screened in 5-BrdU. Figure 3 depict the growth curves of E 4 -HGPRT − and B 3 -TK − in screening reagent (6-TG and 5-BrdU), complete media and HAT media after mutagenic treatment. The unmutated cells could not survive in screening reagent (Figure 3a,d). In complete medium culture, both normal cell line and mutant could grow normally (Figure 3b,e). The cell apoptosis in HAT medium was adopted as an indication of HGPRT or TK deficiency. The growth curve (Figure 3c,f) shows that E 4 -HGPRT − and B 3 -TK − were unable to divide in HAT media, which indicates that E 4 -HGPRT − and B 3 -TK − were sensitive to HAT media. ( Figure 3c,f) shows that E4-HGPRTand B3-TKwere unable to divide in HAT media, which indicates that E4-HGPRTand B3-TKwere sensitive to HAT media.

Generation of BsMAb
E4-HGPRTand B3-TKwere fused by hybrid-hybridoma technology. A total of 83 tetradoma colonies were generated, and the cell supernatant was identified by ELISA. After 3-4 rounds of subcloning, a positive tetradoma T26 was chosen and was used for production of anti-AFB1/OTA BsMAb by induction in vivo. Clear bands of heavy chain (about 55 kDa) and light chain (about 25 kDa) from BsMAb were observed in the electropherogram ( Figure 4). The concentration of BsMAb was 9.61 mg/mL measured by NanoDrop ultraviolet spectrophotometer.

Antibody Specificity Determination
The cross-reactivity of the BsMAb with other common mycotoxins was detected by ELISA. As shown in Table 2, five mycotoxins containing similar structures displayed evident cross-reactivities. The cross-reaction rate with AFB2 was 37%, with AFG1 15%, with

Antibody Specificity Determination
The cross-reactivity of the BsMAb with other common mycotoxins was detected by ELISA. As shown in Table 2, five mycotoxins containing similar structures displayed evident cross-reactivities. The cross-reaction rate with AFB 2 was 37%, with AFG 1 15%, with AFM 1 48%, with AFM 2 10%, and with OTB 36%. Negligible cross-reaction was observed with other tested compounds.

Antibody Specificity Determination
The cross-reactivity of the BsMAb with other common mycotoxins was detected by ELISA. As shown in Table 2, five mycotoxins containing similar structures displayed evident cross-reactivities. The cross-reaction rate with AFB2 was 37%, with AFG1 15%, with AFM1 48%, with AFM2 10%, and with OTB 36%. Negligible cross-reaction was observed with other tested compounds.

Antibody Specificity Determination
The cross-reactivity of the BsMAb with other common mycotoxins was detected by ELISA. As shown in Table 2, five mycotoxins containing similar structures displayed evident cross-reactivities. The cross-reaction rate with AFB2 was 37%, with AFG1 15%, with AFM1 48%, with AFM2 10%, and with OTB 36%. Negligible cross-reaction was observed with other tested compounds.

Antibody Specificity Determination
The cross-reactivity of the BsMAb with other common mycotoxins was detected by ELISA. As shown in Table 2, five mycotoxins containing similar structures displayed evident cross-reactivities. The cross-reaction rate with AFB2 was 37%, with AFG1 15%, with AFM1 48%, with AFM2 10%, and with OTB 36%. Negligible cross-reaction was observed with other tested compounds.

Antibody Specificity Determination
The cross-reactivity of the BsMAb with other common mycotoxins was detected by ELISA. As shown in Table 2, five mycotoxins containing similar structures displayed evident cross-reactivities. The cross-reaction rate with AFB2 was 37%, with AFG1 15%, with AFM1 48%, with AFM2 10%, and with OTB 36%. Negligible cross-reaction was observed with other tested compounds.

Antibody Specificity Determination
The cross-reactivity of the BsMAb with other common mycotoxins was detected by ELISA. As shown in Table 2, five mycotoxins containing similar structures displayed evident cross-reactivities. The cross-reaction rate with AFB2 was 37%, with AFG1 15%, with AFM1 48%, with AFM2 10%, and with OTB 36%. Negligible cross-reaction was observed with other tested compounds.

Antibody Specificity Determination
The cross-reactivity of the BsMAb with other common mycotoxins was detected by ELISA. As shown in Table 2, five mycotoxins containing similar structures displayed evident cross-reactivities. The cross-reaction rate with AFB2 was 37%, with AFG1 15%, with AFM1 48%, with AFM2 10%, and with OTB 36%. Negligible cross-reaction was observed with other tested compounds. It may be speculated from the cross-reactivities of aflatoxins that the specific recognition site of the BsMAb to aflatoxins was mainly coumarin plus cyclopentenone, followed by the difuran ring. The double bond on the furan ring was more conducive to the recog- It may be speculated from the cross-reactivities of aflatoxins that the specific recognition site of the BsMAb to aflatoxins was mainly coumarin plus cyclopentenone, followed by the difuran ring. The double bond on the furan ring was more conducive to the recognition of the antibody than the single bond (the cross-reaction rate of AFG1 was higher >1000 <0.1 It may be speculated from the cross-reactivities of aflatoxins that the specific recognition site of the BsMAb to aflatoxins was mainly coumarin plus cyclopentenone, followed by the difuran ring. The double bond on the furan ring was more conducive to the recognition of the antibody than the single bond (the cross-reaction rate of AFG 1 was higher than that of AFG 2 , and the same trend exists between AFM 1 and AFM 2 ). Very often anti-AFB 1 antibodies have high cross-reactivity for other aflatoxins, and a positive result for AFB 1 is a sufficient motive for detailed analysis of aflatoxins in the sample by other techniques [43]. IAC was applied as pretreatment method in this study, and a certain amount of cross-reactivity of the antibody would be more acceptable than that used in specific quantitative analysis.

Antibody Affinity Determination
The affinity constant (Ka) at different coating concentrations was confirmed by ELISA and shown in Figure 5. The average Ka of the BsMAb to AFB 1 and OTA were 2.43 × 10 8 L/mol, and 1.57 × 10 8 L/mol, respectively. Previous studies indicated that the affinity constant (Ka) of high affinity antibodies is between 10 7 L/mol-10 12 L/mol [48], suggesting that the BsMAb prepared in this study is a high affinity antibody.

FB1
>1000 <0.1 T-2 >1000 <0.1 It may be speculated from the cross-reactivities of aflatoxins that the specific recognition site of the BsMAb to aflatoxins was mainly coumarin plus cyclopentenone, followed by the difuran ring. The double bond on the furan ring was more conducive to the recognition of the antibody than the single bond (the cross-reaction rate of AFG1 was higher than that of AFG2, and the same trend exists between AFM1 and AFM2). Very often anti-AFB1 antibodies have high cross-reactivity for other aflatoxins, and a positive result for AFB1 is a sufficient motive for detailed analysis of aflatoxins in the sample by other techniques [43]. IAC was applied as pretreatment method in this study, and a certain amount of cross-reactivity of the antibody would be more acceptable than that used in specific quantitative analysis.

Antibody Affinity Determination
The affinity constant (Ka) at different coating concentrations was confirmed by ELISA and shown in Figure 5. The average Ka of the BsMAb to AFB1 and OTA were 2.43 × 10 8 L/mol, and 1.57 × 10 8 L/mol, respectively. Previous studies indicated that the affinity constant (Ka) of high affinity antibodies is between 10 7 L/mol-10 12 L/mol [48], suggesting that the BsMAb prepared in this study is a high affinity antibody.

Preparation of BsMAb Based IAC
While anti-AFB1/OTA BsMAb was coupling with CNBr-Sepharose 4B, the concentration of antibody in the supernatant of the coupling solution was determined every 0.5 h, and the coupling efficiency was calculated to determine the optimal coupling time. Figure

Preparation of BsMAb Based IAC
While anti-AFB 1 /OTA BsMAb was coupling with CNBr-Sepharose 4B, the concentration of antibody in the supernatant of the coupling solution was determined every 0.5 h, and the coupling efficiency was calculated to determine the optimal coupling time. Figure 6 shows that the coupling rate rapidly increased to 90.1% for 1 h and the increase in speed over 1 h became slower. Thus 1 h was selected as the optimal coupling time.

Preparation of BsMAb Based IAC
While anti-AFB1/OTA BsMAb was coupling with CNBr-Sepharose 4B, the concentration of antibody in the supernatant of the coupling solution was determined every 0.5 h, and the coupling efficiency was calculated to determine the optimal coupling time. Figure  6 shows that the coupling rate rapidly increased to 90.1% for 1 h and the increase in speed over 1 h became slower. Thus 1 h was selected as the optimal coupling time.

Optimization of Elution Conditions and Evaluating of the IAC Capacity
To achieve high extraction efficiency, three affecting conditions were optimized and were evaluated by the recovery of AFB1 and OTA.
Firstly, four kinds of elution buffer (A, B, C, D) were tested. The recoveries are shown in Figure 7a. Solution A achieved the lowest recovery, containing no methanol and ionic

Optimization of Elution Conditions and Evaluating of the IAC Capacity
To achieve high extraction efficiency, three affecting conditions were optimized and were evaluated by the recovery of AFB 1 and OTA.
Firstly, four kinds of elution buffer (A, B, C, D) were tested. The recoveries are shown in Figure 7a. Solution A achieved the lowest recovery, containing no methanol and ionic compound. When the column was eluted with solution D, the recovery of AFB 1 and OTA was up to 90%. The concentration of methanol, pH and ion concentration could all influence the recovery. The methanol concentration has significant effect, and ion concentration and low pH also improved the elution recovery. It can be seen from Figure 7a that the recoveries of solution B and D are all around 80%. However, high methanol concentration may affect the activity of the antibody [43], so solution D containing NaCl and medium concentration of methanol is preferred. Under the optimized elution conditions (4 mL of solution D with 60% methanol and 1 M NaCl used as elution buffer), the IAC was used repeatedly for seven cycles and the capacity was detected at every cycle. As shown in Table 3, the maximum binding capacities of the IAC for AFB1 and OTA were both over 165 ng. As the column is used repeatedly, the capacity decreases gradually. After seven cycles of use, the preservation rates of column capacity for AFB1 and OTA were 69.3% and 68.0%, respectively (Table 3), which is equivalent to 114.4 µg/kg and 116.3 µg/kg of AFB1 and OTA in samples, respectively, by conversion. According to some investigation results of mycotoxin contamination in agro--products and food samples [50][51][52][53], the capacity of the IAC could meet the needs of practical applications.

Analysis of Corn and Wheat Samples
AFB1 and OTA in spiked corn and wheat samples were detected by IAC combined with ELISA, and confirmed by LC-MS. The recoveries of IAC-ELISA ranged from 95.4% to 105.0%, and the coefficients of variation (CV) were less than 10% ( Table 4). The results indicated that the assay performed well with appropriate recovery and accuracy. This Secondly, to further investigate the effect of methanol concentration, elution solutions containing different concentrations of methanol (30%, 40%, 50% and 60%, v/v, 1 M NaCl, pH 2.3) were evaluated. The recovery rose along with the increase of methanol concentration, and the highest recovery was obtained when the concentration of methanol was 60% ( Figure 7b).
Thirdly, under the conditions optimized above, using solution containing 60% methanol, 1 M NaCl, pH 2.3 as elution solution, the effect of elution solution volumes (1, 2, 3, 4, 5 mL) on recoveries were also investigated. As shown in Figure 7c, recoveries increased from 87.2% to 96.2% for AFB 1 and 76.7% to 96.6% for OTA as the volume increased from 1 mL to 4 mL. The highest recovery was acquired when the volume was above 4 mL, and further increase of the volume did not visibly improve the recovery, thus 4 mL was selected as the optimized volume.
Under the optimized elution conditions (4 mL of solution D with 60% methanol and 1 M NaCl used as elution buffer), the IAC was used repeatedly for seven cycles and the capacity was detected at every cycle. As shown in Table 3, the maximum binding capacities of the IAC for AFB 1 and OTA were both over 165 ng. As the column is used repeatedly, the capacity decreases gradually. After seven cycles of use, the preservation rates of column capacity for AFB 1 and OTA were 69.3% and 68.0%, respectively (Table 3), which is equivalent to 114.4 µg/kg and 116.3 µg/kg of AFB 1 and OTA in samples, respectively, by conversion. According to some investigation results of mycotoxin contamination in agro-products and food samples [50][51][52][53], the capacity of the IAC could meet the needs of practical applications.

Analysis of Corn and Wheat Samples
AFB 1 and OTA in spiked corn and wheat samples were detected by IAC combined with ELISA, and confirmed by LC-MS. The recoveries of IAC-ELISA ranged from 95.4% to 105.0%, and the coefficients of variation (CV) were less than 10% ( Table 4). The results indicated that the assay performed well with appropriate recovery and accuracy. This may be ascribed to IAC pretreatment, which could relieve matrix interference effectively and ameliorate the recovery and accuracy of ELISA for AFB 1 and OTA detection in food matrices. The results were confirmed by LC-MS to further prove the reliability of this assay. Eight real samples of corn and wheat were simultaneously analyzed by the established IAC-ELISA and LC-MS. The results are shown in Figure 8. AFB 1 and OTA both tested positive in the eight samples, but the levels were all below the maximum levels. The correlations between the detection results and LC-MS were both above 0.9. Thus it can be seen that the established IAC-ELISA could be applied to the detection of AFB 1 and OTA simultaneously in real corn and wheat samples with high accuracy.  Figure 8. AFB1 and OTA both tested positive in the eight samples, but the levels were all below the maximum levels.
The correlations between the detection results and LC-MS were both above 0.9. Thus it can be seen that the established IAC-ELISA could be applied to the detection of AFB1 and OTA simultaneously in real corn and wheat samples with high accuracy.

Conclusions
In this study, we develop a BsMAb based IAC which could bind AFB1 and OTA simultaneously, and the binding sites were 1:1 with the two mycotoxins. The coupling rate

Conclusions
In this study, we develop a BsMAb based IAC which could bind AFB 1 and OTA simultaneously, and the binding sites were 1:1 with the two mycotoxins. The coupling rate and the binding site ratio of AFB 1 and OTA are easy to adjust and measure, and mutual interference could be avoided because of the homogeneity of the antibody structure. It is more efficient, convenient, and economical than the IACs based on single-specific antibodies. With satisfactory matrix effect elimination effect and recovery rate, it could be applied to detect AFB 1 and OTA rapidly and effectively, combined with ic-ELISA. Besides, the application of this anti-AFB 1 /OTA BsMAb in IAC could provide reference for the application of other BsMAb in IAC in the future.
The BsMAb prepared in this study demonstrated different cross-reactivities to five biotoxins with similar structures. It is necessary to further improve the hapten structures and the screening strategy to prepare a more specific BsMAb if it were to be applied in the field with higher requirements for specificity. As a mature ELISA detection mode that could quantitatively determine two analytes simultaneously is unavailable at present, the ic-ELISA used in this study can just detect AFB 1 and OTA respectively. If simultaneous detection of two toxins in ELISA can be achieved, this study will have more practical application value.