602 Phakpoom (Optimization and Efficien).cdr OPTIMIZATION AND EFFICIENCY IMPROVEMENT OF IN HOUSE IMMUNOAFFINITY COLUMN KU-AF02 FOR - AFLATOXIN DETECTION PHAKPOOM KOOPRASERTYING , THANAPOOM MANEEBOON , PANRAPEE 1,5 2 IAMTAWEEJAROEN , RATCHANEE HONGPRAYOON1,5 3 and WARAPA MAHAKARNCHANAKUL 4* 1The Graduate School, Kasetsart University, Bangkok 10900, Thailand 2 Scientific Equipment and Research Division, Kasetsart University Research and Development Institute, Kasetsart University, Bangkok 10900, Thailand 3 at Kamphaeng SaenDepartment of Plant Pathology, Faculty of Agriculture , Kasetsart University, Nakhon Pathom 73140, Thailand 4 Department of Food Science and Technology, Faculty of Agro-Industry, Kasetsart University, Bangkok 10900,Thailand 5 Center for Advanced Studies for Agriculture and Food (CASAF), Kasetsart University Bangkok 10900, Thailand, Received 4 January 2016/Accepted 26 April 2017 ABSTRACT In-house immunoaffinity column for aflatoxins (AFs) detection is firstly developed by Kasetsart University since 2011 i.e. the first IAC prototype, KU-AF01. Detection of aflatoxin B (AFB ) up to 200 ng is considered significantly 1 1 efficient, but the production cost is considered high. The high cost was due to the cost of monoclonal antibody specific to aflatoxin and its supporting materials i.e. CNBr-activated sepharose 4B. Therefore, this study was aimed to improve the efficiency of in-house immunoaffinity colums by replacing CNBr-activated sepharose 4B with other supporting materials and to lower the ratio of antibody to supporting materials. Response surface methodology (RSM) was applied to determine the optimum alternative supporting materials and the ratio. Results revealed that when compared with CNBr-activated sepharose 4B, all materials tested could recover higher than 80% AFB , except the 1 activated CH sepharose 4B. Results also indicated that the optimum ratio was 0.4 mg anti-aflatoxin monoclonal antibody to 204 µL CNBr-activated sepharose 4B. CNBr-activated sepharose 4B was considered to provide the best precision in recovering AFB . KU-AF02 increased the ability to detect AFB to 500 ppb. The recovery of AFs in the 1 1 reference materials using KU-AF02 was more than 96% successful, with HorRat value range of 0.34 – 0.76. The reference materials used were 2 levels of AFs in peanuts, i.e. 47.68 and 72.14 ng/g. In terms of quality control in IAC production, there were no significant differences among the 5 batches produced (p > 0.05). KU-AF02 demonstrated ostable and constant percent recovery at 4 C for up to 12 months (tested with standard 200 ppb AFB ). This study 1 indicated that KU-AF02 could lower the ratio of supporting materials to the antibody and that KU-AF02 has high stability and has good ability for cleaning up AFs. Therefore, KU-AF02 is recommended to be used as an in-house immunoaffinity column for aflatoxin detection in Thailand. Keywords: Aflatoxins, immunoaffinity column, in-house method, mycotoxin detection * Corresponding author: fagiwpm@ku.ac.th 114 BIOTROPIA 4 2 7 114 126 Vol. 2 No. , 201 : - DOI: 10.11598/btb.201 .2 . .7 4 2 602 INTRODUCTION Aflatoxins (AFs) are secondary metabolites called difuranocoumarins, produced by Aspergillus flavus A. parasiticus and , commonly found in food and feed (Calvo 2002). AFs are toxic and may et al. have been the most investigated mycotoxin. AFs have been associated with aflatoxicosis in livestock, domestic animals and humans worldwide. At present, the maximum limit of AFs in food is set at 20 ppb by the United States Food and Drug Administration (US FDA) (Anukul . et al 2013), whereas only 4 ppb is allowed in food by the European Union (European Union 2006). Countries worldwide have established their respective limit of AFs. Usually, the contamination of AFs has high prevalence, but low concentration at the level of part per billion (ppb). In the last couple of years, several Thai exported products were rejected due to AFs contamination; for example 40 ppb of AFs in rice (Chinaphuti & Aukkasarakul 2009) and 31 ppb of AFs in chili pouch packed together with instant flavoured noodles (Rapid Alert System for Food and Feed 2012). A s analysis in food is normally carried out by F extracting food samples with organic solvents such as methanol, followed by AFs detection. The food extract solvent, however, may contain Afs together with other impurities which consequently will interfere with the results of analyses. To increase the reliability of analysis results, the impurities have to be removed from the food extract solvent prior to analysis, using immunoaffinity column (IAC). At present, IAC is the common requirement in mycotoxin detection and quantification. Unfortunately, in Thailand most of IAC are imported which resulted to high analysis cost. In 2007, the first prototype of IAC developed from polyclonal antibody was conducted at Kasetsart University and was succeeded in recovering 20 ppb AFB (Wongsuttichot 2007). 1 This prototype was initially meant to be used as an in-house IAC to reduce analysis cost. Later, the polyclonal antibody of this prototype IAC was replaced with monoclonal antibody to improve the efficiency, but the production cost was still considered costly (Mahakarnchanakul 2011). et al. Since antibodies and supporting material contribute as a major part of the production cost, these components have to be optimized to reduce the production cost of IAC for cleaning up AFs. Lower cost production may be encouraged using KU-AF02 as in-house IAC. It is expected that the increasing success in detecting aflatoxins will result to safer food in Thailand. MATERIALS AND METHODS Materials and Chemicals Cyanogen bromide-activated sepharose 4B (CNBr-activated sepharose 4B) as supporting material used for antibody immobilization was purchased from GE Healthcare (Sweden). The apparatus for IAC production such as solid phase extraction (SPE) empty tube 1 mL, 20 m loose μ frits, SPE outlet cap and SPE inlet cap (Vertical Chromatography, Thailand) were used for packing the coupled medium. Chrome grade acetonitrile and methanol were purchased from Mallinckrodt (USA). The monoclonal antibody (MAb) specific to AFs was kindly provided by Department of Plant Pathology, Faculty of Agriculture at Kamphaeng Saen, Kasetsart University, Thailand. Acetic acid, sodium chloride and sodium bicarbonate were purchased from Merck (Germany). Standard AFB and sodium 1 acetate were purchased from Sigma (Spain). Tris hydrochloride was purchased from Vivatis (USA). Ground peanuts containing aflatoxins (as the reference material) were prepared by Mycotoxin Laboratory, at Scientific Equipment and Research Division, Kasetsart University Research and Development Institute (KURDI), Kasetsart University, Thailand. Preparation of Solution According to CNBr-activated sepharose 4B manual (GE Healthcare 2009), coupling buffer solution (0.1M NaHCO containing 0.5M NaCl, 3 pH 8.3) was prepared by dissolving 8.40 g NaHCO and 29.20 g NaCl in 1 L filtered 3 deionized (DI) water. Then, the blocking buffer solution (0.1M Tris-HCl, pH 8.0) was freshly prepared by dissolving 15.76 g Tris hydrochloride in DI water. Before adjusting volume, pH was adjusted to 8.0. This step was followed by preparing the washing buffer solution (0.1 M Acetic acid/Sodium acetate, pH 4.0 containing 0.5 M NaCl and 0.1 M Tris-HCl, pH 8.0 containing 0.5 M NaCl) which was prepared by mixing 90 mL 0.2M NaAc with 410 mL of 0.2M Acetic acid, added with 29.20 g NaCl followed by adjusting the total volume to 1,000 mL. Phosphate buffer saline 10x (PAA, Austria) was diluted with DI water to achieve 1x PBS (pH 7.4) to re-suspend the medium. PBS with 0.02% NaN 3 was prepared by dissolving 0.02g NaN in 100 mL 3 PBS. All solutions were filtered through 0.45 µm nylon filter before usage. Karnovsky solution, as fixative for electron microscopy, was prepared by mixing 2 g of paraformaldehyde with 25 mL of distilled water. The solution was heated in 60 °C with stirrer plate. Then the solution was added with 1M Optimization and efficiency improvement of immunoaffinity column for aflatoxin detection – Kooprasertying et al. 115 The medium was then collected using sintered filter flask, washed from the excess antibody using ±50 mL coupling buffer and transferred into ±50 mL blocking buffer (0.1M Tris-HCl, pH 8.0). The medium in blocking buffer was shaken using orbital shaker for another 2 hours, washed with washing buffer (0.1 M Acetic acid/Sodium acetate, pH 4.0 containing 0.5 M NaCl and 0.1 M Tris-HCl, pH 8.0 containing 0.5 M NaCl). This was conducted for 3 cycles. Each cycle used ±50 mL washing buffer. The medium was then resuspended in 50 mL PBS. Finally, 0.2 mL medium was packed in each SPE empty tube with polyethylene frits and the column was filled up with PBS containing ±0.6 mL 0.02% NaN in 3 each column. (modified from GE Healthcare 2009) and stored at 4 °C until usage. Optimization of Antibody and CNBr– Activated Sepharose 4B To develop IAC, the major cost of IAC production depended on the amount of antibody a n d C N B r – a c t i v a t e d s e p h a r o s e 4 B (Mahakarnchanakul et al. 2011). Thus, the amount of these 2 components has to be optimized. Response surface methodology was applied to determine the proper ratio. The experiment was designed as central composite design and the data obtained were analyzed using MINITAB (version 14). The alpha value was face centered. The experimental treatments are summarized and shown in Table 1. The recovery percentage of AFB , standard 1 deviation (SD) and relative standard deviation (RSD) were the parameters for response surface determination to optimize volume of antibody and CNBr–activated sepharose 4B. The seven columns from each treatment (Table 1) were tested with 15 mL 30% methanol (containing 200 ng of AFB , which was prepared 1 from stock of 10 ppm AFB ). The solution was 1 passed through IAC at a flow rate of 1 - 2 drops per second (dps). After the whole solution had been passed through the IAC, the column was washed twice with 10 mL deionized water (DI water) at the same flow rate. The captured AFs were eluted with 1 mL methanol at 1 dps, and then filtered (nylon syringe filter, Whatman) and subjected to high performance l iquid chromatography (HPLC) evaluation. NaOH until the solution was clear. Finally, 5 mL 50% Glutaraldehyde and 20 mL 0.2M Cacodylate buffer (pH 7.2) were added into the solution. To prepare 0.2M Cacodylate buffer (pHv7.2), 8.56 g Sodium cacodylte, 25 g Calcium chloride and 2.5 mL 0.2N HCl were dissolved with 200 mL of DI water. Amplification of Monoclonal Antibody Specific to Aflatoxins Monoclonal antibody (MAb) specific to aflatoxins from hybridoma clone C10 was produced using Integra Celline flasks (CL1000) based on the manual from the manufacturer (INTEGRA Biosciences AG CH-7205 Zizers, Switzerland). Briefly, the hybridoma was maintained between the space of membrane 6(initial cell amount was 2 x 10 cells/mL which had > 80% of viable cell) in the Hybridoma Serum Free Medium (HSFM). The flask was filled with HSFM up to 250 mL. At certain times, cells were 7collected, leaving 3 x 10 cells/mL to be maintained in the flask. The flask was then re- filled with fresh HSFM up to 250 mL. The collected medium was then centrifuged (2,000 rpm for 5 minutes). The supernatant was purified by AKTAprime™ system (GE Healthcare, Sweden) using Hitrap protein G column (GE Healthcare, USA). The purified antibody concentration was determined by measuring OD 280 nm and calculated as the following equation. ○Then, MAb was lyophilized and stored at 4 C until usage. IAC Preparation To prepare 50 columns of IAC, purified 50 mg MAb was dissolved in coupling buffer solution (0.1M NaHCO containing 0.5 M NaCl, pH 8.3) 3 (1 mg/mL) then coupled with 10.2 mL CNBr- activated sepharose 4B. The mixture was then shaken using orbital shaker (Boekel scientific, ○USA) at 150 rpm, 4 C, for 18 hours. At the end of the incubation period, the mixture was left at 26 °C to separate the medium (CNBr-activated sepharose 4B coupling with Mab). The supernatant was then measured at OD 280 nm, in order to determine the excess of MAb using the following formula: 116 BIOTROPIA Vol. 24 No. 2, 2017 Concentration of antibody (mg.mL) = OD 280 nm value x dilution factor 1.4 The results were transformed to percent recovery and the optimum ratio of antibody and CNBr-activated sepharose 4B volumes was determined. Percent recovery was calculated using the following formula: After calculating the recovery percentage, the Standard Deviation (SD) and Relative Standard Deviation (RSD) were determined. RSD was calculated to describe the precision and the repeatability of each treatment. RSD was calculated using formula: The recovery percentage, SD and RSD were then used as suitable variables in the response surface analysis, performed by MINITAB version 14. High Performance Liquid Chromatography Condition To determine the amount of AFB , AFB , 1 2 AFG and AFG , the HPLC with fluorescent 1 2 detector (Separation module e2695 with Fluorescence detector 2475, Waters, USA) was used. A symmetry C18 column (5 m, 4.6 x 150 μ mm) with C18 guard column (Water Corporation, USA) was used to separate the aflatoxins. Photochemical derivatization (PHRED, AURA Industries, USA) was used to enhance the AFB 1 and AFG detection by fluorescence detector. 1 The column was equilibrated with a mixture of water:acetonitrile:methanol [60:20:20 (v/v)] which was also the mobile phase. Flow rate at 1 mL/minute was used and the column oven temperature was 35 C. The excitation and ○ emission wavelength of the fluorescence detector were set at 365 and 445 nm, respectively. Improvement of the Efficiency of In-house IAC The CNBr-activated sepharose 4B was replaced by CH-activated sepharose 4B, HiTrap NHS-activated sepharose 4B or NHS-activated sepharose 4B having spacer arm length of 8, 10 and 14 atoms, respectively. The IAC efficiency were compared by passing 200 ng AFB in 15 mL 1 30% methanol through the IAC and then evaluated by HPLC. Beside the recovery percentage, the distributions of antibody on the supporting material were observed by transmission electron microscopy. The coupled supporting material was then attached with gold labeled anti-mouse IgG (H+L) system (KPL, USA). The wet gel was dehydrated with rising concentrations of ethanol (30, 50, 70, 90, and l00%, v/v), absolute acetone and embedded in activated Micropal. After sedimentation of the gel particles occurred, the 117 Table 1 Coupling percentage, actual amount of antibody and recovery percentage of 200 ng AFB from IAC containing 1 different volumes of antibody and supporting material (CNBr-activated sepharose 4B) Column Experimental design Results Volume of Antibody to supporting material ratio (mg/mL) Coupling percentage (%) Actual antibody (mg) AFB1 recovery*, a (%) Expected antibody (mg) CNBr-activated sepharose 4B (µL) 1 0.3 153 1.96 97.26 0.31 79.66 ± 2.03f 2 0.4 153 2.61 90.02 0.37 82.70 ± 2.63e 3 0.5 153 3.26 90.95 0.47 85.82 ± 0.83d 4 0.3 204 1.47 93.29 0.29 86.64 ± 2.51d 5 0.4 204 1.96 97.26 0.42 103.36 ± 2.31b 6 0.5 204 2.45 90.95 0.47 88.62 ± 3.70d 7 0.3 255 1.17 92.51 0.29 87.83 ± 2.13d 8 0.4 255 1.46 93.14 0.38 94.80 ± 4.23c 9 0.5 255 1.96 97.26 0.52 107.13 ± 1.25a Note: * = Values in the same column followed by different letters are significantly different at p < 0.05 a = Mean ± SD from 7 replicates Percent recovery = concentration of AFB1 after passing through the IAC column concentration of AFB1 before passing through the IAC coloum x 100 Relative Standard Deviation (%) = standard deviation mean x 100 Optimization and efficiency improvement of immunoaffinity column for aflatoxin detection – Kooprasertying et al. 118 BIOTROPIA Vol. 24 No. 2, 2017 mixture was cooled down to 4 °C. The lower part of the stiffened gelatin containing the sepharose beads was fixed at 4 °C in Karnovsky solution for 30 minutes, washed in cacodylate buffer (pH 7.2, 0.2 M), dehydrated and embedded in activated Micropal. Ultrathin sections were then made with the LKB Ultratome I11 (LKB-Produkter, Stockholm). The sections were viewed with electron microscope (HT 7700, Hitachi, Japan) at an accelerating voltage of 80 kV with 10,000x magnification. IAC Capacity Testing The capacity of the optimized IAC was tested by passing various amounts of standard AFB (1, 1 10, 100, 200, 300, 500 and 1,000 ng) through the IAC column. Total AFs recovery was also conducted at 50 ng for each of AFB , AFB , AFG 1 2 1 and AFG (total AFs amount was 200 ng). The 2 results were then transformed into recovery percentage. RSD and Horwitz ratio (HorRat) value were then determined. The HorRat value was calculated using formula: The predicted RSD was calculated from Horwitz ratio (HorRat) using formula: (-0.15)Predicted RSD = 2C , where: C = concentration found or added, expressed as dimension less mass fraction (Thompson 2007). Ground peanuts as reference material containing AFs at 47.68 ng/g (AFB 38.25 and 1 AFB 9.75 ng/g) and 72.14 ng/g (AFB 56.28 and 2 1 AFB 15.86 ng/g) were extracted and passed 2 through the optimized and Aflatest IAC columns based on the Aflatest HPLC instruction manual (Vicam 2009). Briefly, 25 g reference material with 5 g methanol were blended at high speed (for 2 minutes at 26 °C) with 125 mL methanol:water (70:30), then filtered through filter paper No.1 (Whatman). Fifteen milliliters of the filtered extract was diluted with 30 mL DI water, filtered through glass microfiber filter (Whatman) and 15 mL was passed through the IAC column. Efficiency Comparison between In-house IAC and Commercial IAC The AFB and AFs concentration resulted 1 from reference material clean-up by the in-house IAC and commercial IAC were compared with statistical analysis including correlation (SPSS version 17), test (SPSS version 17) and Bland-t- Altman plot (Sigma plot version 12). The Stability of IAC Production and Shelf- life Testing Five batches of optimized IAC were produced in order to check the quality of producing different batches. Seven columns from each batch were tested with 200 ng AFB . The results were 1 analyzed to obtain data on the quality difference among batches productions using SPSS software version 17. For shelf-life testing, IAC was produced and stored at 4 °C. The IAC was tested within batches with 200 ng AFB after 1, 3, 6, 9 1 and 12 months of storage duration. The results were analyzed to obtain data on the difference among storage duration and acceptable range of recovery percentage using SPSS software version 17. RESULTS AND DISCUSSION Optimization of Materials using IAC Production After MAb was coupled with CNBr-activated sepharose 4B, the excess amount of the antibody was measured to determine the coupling percentage. All treatments gave excellent coupling percentage (> 90%), ranging from 90.02 to 97.26% (Table 1). The amount of bound antibodies did not different from the expected amount of antibodies in each column. The highest coupling percentage (97%) was obtained when the volume ratio of the antibody to the supporting material was 300 µL:153 µL, 400 µL:204 µL, 500 µL:255 µL. Under these conditions, ratio of antibody and supporting material were much lower than the ratio recommended by the operating procedure (GE Healthcare 2009) i.e. at least 2.5 - 5 times. The operating procedure recommended 5 - 10 mg of HorRat = RSD from the results RSD predicted from Horwitz equation 119 protein per mL of CNBr-activated sepharose 4B, which is quite large amount of antibody to be used. When the amount of antibody was higher or lower than 2 mg of antibody per 1 mL of CNBr- activated sepharose 4B, the coupling percentage decreased as shown in Table 1. The supplier's recommended ratio of MAb and supporting material might not be suitable for specific purpose or antibody. For instance, in developing IAC for zearalenone and ochratoxin clean-up, some research found that 1 mg antibody coupled with 1 mL of CNBr-activated sepharose 4B, as supporting material, provide 40 – 50% stability toward high concentration of acetonitrile and methanol (Uchigashima et al. 2009; Uchigashima et al. 2012). The ratio used was obviously lower than the recommended ratio, but still the recovery percentage were satisfactorily ranging from 80 to 100%. In some cases, however, the supplier's recommended ratio is useful. For example, in developing IAC for bisphenol A in serum, some research found that the ratio of the polyclonal antibody and supporting material (14 mg to 2 mL) provided coupling percentage of 86±3% (Zhao et al . 2003). In the case of IAC for quinolone and sulfonamide clean-up, it was found that the ratio of 8 mg antibody to 1 mL supporting material provided coupling percentage of 91.1% (Li et al. 2008), which was lower compared to coupling percentage provided by our study (97%). High coupling percentage, however, is not the guarantee for the IAC efficiency. Therefore, the efficiency should be evaluated as recovery percentage. According to The Codex Alimentarius Commission (1995), the acceptable recovery percentage of mycotoxin analysis is 80 - 110%. 1All conditions could recover AFB in the acceptable range, except the ratio of 300 µL antibody to 153 µL CNBr-activated sepharose 4B (Table 1). Although our study provided high coupling percentage (97%), the efficiency to recover AFB was not acceptable (79%).1 In terms of extraction step, basically 70 – 80% methanol or acetonitrile is sufficient to be used as the solvent for AFs extraction. To allow antibody completely capture the toxin, the organic solvent in extraction process should be diluted to 23 – 26%. The highest organic solvent concentration to be passed through IAC column was 30% (Scott & Trucksess 1997). In our study, the specificity test was conducted by passing through 200 ng AFB1 diluted in 15 mL 30% methanol in the IAC column. The concentration of organic solvent before being passed through the column was slightly higher than the recommendation provided by Vicam manual (23 - 26%). When the amount of supporting materials asw varied toward a constant 0.3 mg of antibody, the results showed significantly different recovery percentage ( < 0.05). Recovery percentage p increased from 79 to 87% as a result of increasing the amount of CNBr-activated sepharose 4B from 153 to 255 µL (Table 1), which might have been caused by the steric hindrance between the tightly packed of antibody on the surface of each bead in the column. Other possible cause was that higher amount of supporting material (at a constant amount of antibody) would provide higher surface and extended reaction time between the antibody and the toxin. Similar results were obtained when the amount of antibody was increased to 0.4 and 0.5 mg. On the contrary, equal amount of supporting materials toward varied amount of antibody would increase recovery percentage. Increasing the amount of antibody did not increase the coupling percentage, which indicated that the excess of antibody would lose during the coupling process. Also, excess amount of antibody might cause the occurrence of steric hindrance which resulted to the less specificity to toxin. Therefore, ratio of antibody to the supporting material is the critical factor to develop the IAC. The ratio should be balanced to achieve the best efficiency in toxin detection. ANOVA results in our study showed that the amount of antibody and CNBr-activated sepharose 4B as supporting material had significant effect on the recovery percentage ( < p 0.05), in which our study used 200 ng AFB for 1 testing the efficiency of IAC column. Therefore, our study chose the 400 µL:204 µL ratio of the antibody to the supporting material or 2 mg antibody per 1 mL supporting material as the expected optimum condition to provide high coupling percentage. To confirm the ratio of these two coupling substances, the response surface methodology was used to determine the optimum coupling condition. Before the optimization process was Optimization and efficiency improvement of immunoaffinity column for aflatoxin detection – Kooprasertying et al. 120 BIOTROPIA Vol. 24 No. 2, 2017 conducted, the recovery percentage, standard deviation and relative standard deviation variables were analyzed to select suitable variables for plotting the response surface. Suitable variables were chosen based on the R value > 85% and 2 p value < 0.05. The selected suitable variables had R 2 of 88.2% and values of 0.033 for antibody and p 0.049 for supporting materials. The value of recovery percentage was determined to be the optimum condition of coupling step based on contour plot of the response surface methodology (Fig. 1). Based on setting criterion of 100%, recovery percentage was selected between the areas of 400 µL of antibody and 204 µL of supporting material. Improving the Efficiency of the In-house IAC This experiment was aimed to improve the efficiency of IAC by varying the spacer arm length between the CNBr and sepharose 4B. Within the affinity of chromatography system, normally the binding site of the target toxin may be located at the surface. Some binding site, however, may be located deep within the molecule hindering the antibody to access the target toxin. Increasing spacer arm length may overcome the hindrance by interposing antibody and supporting material. Length of the spacer arm may provide better affinity, but the length increase must be cautiously done. Previously, the experiment was designed to vary the spacer arm length between the sepharose 4B and CNBr, due to the limited availability of cyanogen bromide in local market, under the permission from the Thailand Ministry of Defence (Royal Thai Government Gazette 2008). This situation made it impossible to vary the length of spacer arm within the same supporting material. Therefore, we used four supporting materials available in local market i.e. CNBr- activated sepharose 4B, CH-activated sepharose 4B, Hitrap NHS-activated sepharose 4B and NHS-activated sepharose 4B. Recovery percentage of the coupling step with different supporting materials are shown in Table 2. Coupling monoclonal antibody with three supporting materials (CNBr-activated sepharose 4B, Hitrap NHS-activated sepharose 4B and NHS-activated sepharose 4B), resulted in successful recovery of AFB (96.6 - 101.1%, 88.9 - 1 98.0% and 95.9 - 103.9%, respectively), except the CH-activated sepharose 4B which provided much lower recovery (59.70 - 84.47%). Low amount of AFB was possible to be recovered using the CH-1 activated sepharose 4B as supporting material, which might be due to the limited length of spacer arm. Based on the specification of each supporting material, the recovery percentage may decrease depending on the number of spacer arms. The CH-activated sepharose 4B used in our Figure 1 Contour plot of recovery percentage determined by antibody and supporting material (Gel) 121 study contained 10 atoms spacer arm and provided the lowest recovery percentage. Among the other three supporting materials, CNBr- activated sepharose 4B exhibited the lowest SD value (Table 2). Therefore, CNBr-activated sepharose 4B was selected as material for IAC production. Table 2 shows similar values of coupling percentage and quite different values of recovery percentage. CNBr-activated Sepharose 4B, Hitrap NHS-activated Sepharose 4B and NHS-activated as supporting materials had moderate active role on the surface to capture the antibody and the target toxin, resulting to high recovery percentage (93 – 101%). The developed IAC was named as KU-AF02 and was further tested for capacity, reproducibility and stability. The distribution of antibody on each supporting materials by Transmission Electron Microscope (TEM) was conducted to investigate the relationship between the affinity of antibody on supporting material to the coupling percentage. Unfortunately, the illustrations presented in this manuscript did not exhibit the relevance of the affinity of antibody and supporting material to the coupling percentage Table 2 Recovery percentage of AFB recovered by IAC containing different supporting material1 Supporting material Spacer arm (n of atom) Coupling percentage (%) Recovery percentagea (%) CNBr-actiavated 4B - (0) 98.21 98.67 ± 1.80 CH-activated Sepharose 4B 6-aminohexanoic acid (8) 97.56 71.83 ± 9.78 Hitrap NHS-activated 4B 6-aminocaproic acid (10) 98.40 93.85 ± 3.63 NHS-activated Sepharose 4B 6-aminocaproic acid (14) 96.08 100.59 ± 3.06 aNote: = Mean and SD from 5 replicates (a) (b) (c) (d) (e) Figure 2 The transmission microscopy illustration of (a) CNBr-activated sepharose 4B and monoclonal antibody distribution on (b) CNBr-activated sepharose 4B © CH-activated sepharose 4B (d) Hitrap NHS-activated 4B and (e) NHS activated sepharose 4B at 10,000x magnification Optimization and efficiency improvement of immunoaffinity column for aflatoxin detection – Kooprasertying et al. 122 BIOTROPIA Vol. 24 No. 2, 2017 (Fig. 2). The dark color indicates the uniform distribution of antibody (Fig. 2b), which was the coupling of antibody with CNBr-activated sepharose 4B. The condensed color was distributed from the outer surface of supporting material to the inner surface. The antibody which should have been shown as dark color was not found at the inner or the outer layer of sepharose (Fig. 2a), although the sepharose was stained with gold label. The dark color shown in Figure 2c was the coupling of antibody with CH-activated sepharose 4B, which was located at the inner or outer surface; the color was slightly toned down from the outer to the inner surface of the supporting material. The dark color shown on Figure 2c, which was the coupling of antibody with activated CH sepharose 4B, was located at the surface or outer and the color slightly dropped from the outer to the inner of supporting material. In case of Hitrap NHS-activated sepharose 4B and NHS-activated sepharose 4B (Fig. 2d and 2e), the antibody seemed to be equally distributed to the inner surface, but the color intensity was less than that in Figure 2a. The coupling percentage of antibody on four supporting materials was shown to be similar (> 90%) (Table 2), which might be presented in Figure 2c. The density of antibody on the surface of CH-activated sepharose 4B seemed to be higher than that on other supporting materials. It was expected that CH-activated sepharose 4B had potential to enhance the affinity of antibody to the supporting material. However, the steric hindrance of the condensed antibody on CH- activated sepharose 4B, caused the lower capture of the tested toxin. The steric hindrance might interfere the affinity of antibody resulting to the lowest recovery percentage for CH-activated sepharose 4B. The Capacity of IAC Recovery percentages of various amount of AFB (1 to 1,000 ng) in 15 mL 30% methanol 1 obtained using the developed IAC KU-AF02 (coupling condition: 0.4 mg of antibody with 204 µL of CNBr-activated sepharose 4B) are shown in Table 3. The capacity of KU-AF02 achieved the standard of analysis when 1 to 500 ng was tested. The recovery percentages were all above 80%. Particularly at 500 ng AFB , 1 the recovery percentage was 87% with HorRat value less than 2 (Table 3). As expected, the recovery value slightly decreased when the amount of AFB increased. Recovery percentage 1 was lower than 80% when 1,000 ng of AFB was 1 loaded. At this amount of AFB , although the 1 HorRat value was less than 2, the recovery percentage was considered not satisfying for the capacity of IAC. Thus, the developed IAC KU-AF02 was successful in increasing the recovery of AFB up to 500 ng compared to the 1 previous IAC, KUAF-01 (Mahakarnchanakul et al. 2011). KU-AF02, was also successful in recovering other derivatives of AFs i.e. AFB , AFG and 2 1 AFG . The experiment was carried out by mixing 2 50 ng of each AFB , AFB , AFG and AFG in 1 2 1 2 15 mL 30% methanol and passing them through the KU-AF02. The IAC column of KU-AF02 was capable to recover 45.37 ± 1.87 ng of AFB1, 40.64 ± 2.63 ng of AFB 43.41 ± 1.93 ng of AFG2, 1 and 33.75 ± 2.20 ng of AFG , with total aflatoxin 2 recovery percentage of 81.6%. Table 3 Recovery percentage at different amount of AFB recovered by optimized IAC1 Amount of AFB1 (ng) Recovery percentagea (%) Calculated RSD Predicted RSD HorRat value 1 102.86 ± 6.42 6.25 22.39 0.28 10 103.81 ± 1.93 1.86 15.85 0.12 100 95.41 ± 2.27 2.38 11.22 0.21 200 95.51 ± 1.71 1.79 10.11 0.18 300 90.07 ± 2.19 2.43 9.52 0.25 500 87.06 ± 3.57 4.10 8.81 0.47 1000 75.35 ± 6.04 8.02 7.94 1.01 aNote: = Mean and SD from 7 replicates 123 The results showed that IAC KU-AF02 recovered AFB , AFB and AFG higher than 1 2 1 80%, except for AFG the recovery percentage 2 was only 67%. Although preliminary test showed that our monoclonal antibody had specificity with single toxin test when loading AFB , AFB and 1 2 AFG and AFG separately in IAC, the recovery 1 2 percentage of each test obtained was > 98% and was not significantly different (p > 0.05) among those recovery percentages. After mixed AFs were loaded, IAC KU-AF02 could recover the parent form of B and G aflatoxins (B , 90.74% and G , 1 1 86.81%) better than the others (B , 81.28% and G2 2 67.50%). Similar to the results when tested with commercial IAC, higher recovery percentage of AFB , AFB and AFG (79.18%, 72.68%, 78.39%) 1 2 1 over AFG (27.25%) were observed.2 The remaining AFG in test samples 2 apparently reduced performance of total AFs detection by IAC, either in the developed IAC or the commercial IAC. Therefore, samples contaminated with high amount of AFG should 2 be carefully handled because the AFG2 may decrease the recovery of other forms aflatoxins. AFG is the highest contaminant in unpolished 2 rice (5.5 ppb) (Jankhaikhot 2005). Further tests were done using reference materials i.e. two-level contaminant contained in defatted ground peanuts. The reference materials were extracted and passed through optimized IAC. The results showed that the recovery percentage of AFB , AFB and AFs obtained 1 2 from reference material no. 1, representing medium contamination (47.68 ng/g) and reference material no. 2, representing high contamination (72.14 ng/g), were between 80 - 110% with HorRat value of 0.8 and ± 0.4, respectively. These results also showed the capacity of KU-AF02 in detecting contaminated food using peanuts as the representative. Similar to the test in pure toxin solution, developed IAC KU-AF02 recovered parent forms (B and G) of aflatoxins from the reference material better than those from the derivative forms, particularly in mixed forms. Efficiency Comparison of In-house and Commercial IAC Seven replicates of medium and high contaminants aflatoxins in peanut (reference materials no.1 and no.2) were quantified using HPLC to determine the amount of AFB and Afs. 1 The correlation between recovery percentage of AFB and AFs using KU-AF02 IAC column and 1 recovery percentage using commercial IAC column were determined (Fig. 3). Close correlation (R) between the obtained amount of AFB and the AFs using the developed 1 and commercial IAC columns were 0.961 and 0.962, respectively. The R values representing 2 the confidence values of X and Y variables were Figure 3 Correlation (linear-regression) between clean-up samples with commercial IAC column (x-axis) and optimized IAC column (y-axis) for AFB ( ) and AFs ( ) detection in peanut reference materials1 80 75 70 65 60 55 50 45 40 35 30 30 40 50 60 70 80 Concentration from commercial IAC (ppb) y = 0.8854x + 4.6804 2R = 0.9237 y = 0.9451x + 4.1188 2R = 0.926 AFB1 AFs AFs AFB1 Concentration from optimized IAC (ppb) Optimization and efficiency improvement of immunoaffinity column for aflatoxin detection – Kooprasertying et al. 0.923 and 0.926, respectively. Close correlations of the linear regression equations indicated similar trend of analysis, but might not be directly compared the efficiency of these tested two IAC columns. tThe paired samples -test and Bland-Altman plot were used to compare the results. The -value t of AFB data (1.283) was lower than critical two- t 1 tailed value (2.160), which meant no difference between two data sets resulting from the two tested column ( > 0.05). A similar result was p found in case of Afs. Apart from -test, the Bland–Altman plot was t used to determine the agreement between two tested methods, for example between modified and standard methods (Bunce 2009; Eken 2009; Hanneman 2008). The use of this method has been reported in the application of medical instrument (Zaki . 2013), computed et al tomography and ultrasound prostate volume measurements (Gloi . 2008). In 2013, Tansakul et al et al. (2013) compared between ELISA and LC- MS/MS for fumonisin detection in corn samples using Bland-Altman analysis and found no difference between these two methods at a range of 0 - 20 mg/kg. 1In our study, the amount of AFB and AFs obtained from the developed IAC KU-AF02 and commercial IAC columns were compared using Bland–Altman plot (Fig. 4a and 4b). None of these data were located outside the limit line (±1.96 SD). Thus, the clean-up of contaminated peanut samples using the developed IAC KU- AF02 or commercial IAC column provided similar results, ranging from 35 to 60 ppb for AFB or ranging from 40 to 80 ppb for AFs (Fig. 1 4). The Stability of Batch Production and Shelf- life of Developed IAC The stability of IAC production was determined in 5 batches. The results showed that all batches could recover AFB higher than 95% 1 (98.32, 99.68, 101.04, 101.27 and 102.38) and there were no significant difference among the batches produced ( >0.05 . Thus, the quality of p ) each batch had no effect on the IAC efficiency. Stability of recovery percentage of KU-AF02 stored for 1 - 12 months, are shown in Table 4. Along the storage duration, recovery percentage decreased from 95 to 85%, but still greater than 80%. There were significant differences of recovery percentage among storage durations (Table 4). The HorRat values for all storage duration were less than 2 (Table 4). Based on the result, the developed IAC KU-AF02 had shelf-life up to 12 months when stored at 4 °C. The long stability will encourage the use of the developed IAC KU- (a) (b) Figure 4 Bland–Altman plots of AFB (a) and AFs (b) concentration after being cleaned-up using the developed IAC KU-1 AF02 and commercial IAC columns 124 BIOTROPIA Vol. 24 No. 2, 2017 Average concentration of AFB 1 from developed and commercial IAC (ppb) D if fe re n ce C o m m er ci al a n d D ev el o p ed I A C Mean + 1.96SD Mean Mean - 1.96SD 6 4 2 0 -2 -4 -6 -8 30 35 40 45 50 55 60 65 8 6 4 2 -2 -4 -6 -8 0 40 45 50 55 60 65 70 75 80 Average concentration of AFs from developed and commercial IAC (ppb) Mean Mean + 1.96S D Mean - 1.96S D D if fe re n ce C o m m er ci al a n d D ev el o p ed I A C AF02, as in-house method. This developed IAC KU-AF02 may also be distributed to other laboratories. CONCLUSIONS The production cost of the developed IAC KU-AF02 was successfully reduced compared to the previous IAC (KU-AF01) by reducing the ratio of antibody to CNBr-activated sepharose 4B. The optimized coupling condition was determined by Response Surface Methodology (RSM) and the range of recovery percentage was acceptable. CNBr-activated sepharose 4B was determined to be the suitable supporting material in this study. The capacity of KU-AF02 was higher than that of the commercial and previous IAC (KU-AF01). KU-AF02 was successful to recover AFs in two reference materials. No significant difference was observed when five batches of KU-AF02 were produced. Optimized IAC had a shelf-life of 12 months when produced and stored at 4 °C. There were no significant difference in using the developed IAC KU-AF02 and commercial IAC to clean-up AFB and AFs from ground peanuts reference 1 materials. ACKNOWLEDGEMENTS The authors acknowledge the financial support received from the Center for Advanced Studies for Agriculture and Food (CASAF), Kasetsart University. We are greatfully thankful to the Scientific Equipment and Research Division, Kasetsart University Research and Development Table 4 Recovery percentage, RSD and HorRat value of 200 ng AFB obtained from recovery process using the developed 1 IAC KU-AF02 stored at different storage duration Storage duration (months) Recovery percentage*, a (%) RSD HorRat value 1 95.01 ± 2.62a 2.75 0.27 3 94.57 ± 3.73a 3.95 0.39 6 93.18 ± 1.19a 1.27 0.13 9 90.52 ± 0.87ab 0.96 0.10 12 85.61 ± 5.18c 6.06 0.60 Note: * = Values in the same column followed by different letters are significantly different at p < 0.05 a = Mean and SD from 7 replicates Institute, Kasetsart University, Thailand for their kind support of the facilities. Thanks are also extended to the Department of Plant Pathology, Faculty of Agriculture at Kamphaeng Saen Kasetsart University for providing monoclonal antibody specific to aflatoxins. REFERENCES Anukul N, Vangnai K, Mahakarnchanakul W. 2013. Significance of regulation limits in mycotoxin contamination in Asia and risk management programs at the national level. J Food Drug Anal 21(3):227-41. Bunce C. 2009. Correlation, agreement, and Bland-Altman analysis: statistical analysis of method comparison studies. Am J Ophthalmol 148(1):4-6. Calvo AM, Wilson RA, Bok JW, Keller NP. 2002. Relationship between secondary metabolism and fungal development. Microbiol Mol Biol Rev 66(3):447-59. Chinaphuti A, Aukkasarakul S. 2009. Contamination of aflatoxin B1 and ochratoxin A in brown rice and white rice. [Abstract]. Thai Rice Research Journal 3:57-65. Eken C. 2009. Bland-Altman analysis for determining agreement between two methods. J Emerg Med 36(3):307. European Union. 2006. Commission Regulation (EC) No 1881/2006 of 19 December 2006 setting maximum levels for certain contaminants in foodstuffs. http://eur-lex.europa.eu/LexUriSer v/LexUriServ. do?uri=OJ:L:2006:364:0005:0024:EN:PDF . [Retrieved on 20 June 2015]. GE Healthcare. 2009. CNBr-activated sepharose 4B. Instruction Affinity Media 1-12. Gloi A, McCourt S, Zuge C, Goettler A, Schlise S, Cooley G. 2008. A Bland-Altman analysis of the bias between 125 Optimization and efficiency improvement of immunoaffinity column for aflatoxin detection – Kooprasertying et al. computed tomography and ultrasound prostate volume measurements. Med Dosim 33(3):234-8. Hanneman SK. 2008. Design, analysis, and interpretation of method-comparison studies. AACN Adv Crit Care 19(2):223-34. Jankhaikhot N. 2005. Aflatoxin in unpolished rice and exposure and estimation to aflatoxin in unpolished rice consumption in Bangkok. MS thesis. Salaya (TH): Mahidol University. Li C, Wang Z, Cao X, Beier RC, Zhang S, Ding S, Li X, Shen J. 2008. Development of an immunoaffinity column method using broad-specificity monoclonal antibodies for simultaneous extraction and cleanup of quinolone and sulfonamide antibiotics in animal muscle tissues. J Chromatogr A 1209(1-2):1-9. Mahakarnchanakul W, Hongprayoon R, Kladpun S, Puthpongsiriporn T, Malichan S, Maneeboon T, Yangyune P, Sinprasopchai L, Saijit P. 2011. Development of Immunoaffinity Column to Reduce Costing of Aflatoxin Determination. Research Report. Bangkok (TH). Rapid Alert System for Food and Feed [Internet]. 2012. RASFF Portal Reference 2012.BZA, 2012. h t t p s : / / w e b g a t e . e c . e u r o p a . e u / r a s f f - window/portal. [Retrieved on 15 May 2015]. Royal Thai Government Gazette [Internet]. 2008. Ministry of Defense announcement 1987: Prescribed chemical licensed under the Arms Control Act. http://www.chemtrack.org/Law/Mod-2551.pdf. [Retrieved on 31 August 2015]. Scott PM, Trucksess MW. 1997. Application of immunoaffinity columns to mycotoxin analysis. J AOAC Int 80(5):941-9. Tansakul N, Jala P, Laopiem S, Tangmunkhong P, Limsuwan S. 2013. Co-occurrence of five toxins in Fusarium corn-dried distiller's grains with solubles in Thailand and comparison of ELISA and LC-MS/MS for fumonisin analysis. Mycotoxin Res 29(4):255-60. The Codex Alimentarius Commission [Internet]. 1995. Codex General Standard for Contaminant and Toxins in Food and Feed. http://www.fao.org/ fileadmin/user_upload/agns/pdf/CXS_193e.pdf. [Retrieved on 20 June 2015]. Thompson M. 2007. Limitations of the application of the Horwitz Equation: A rebuttal. Trends Analyt Chem 26(7):659-61. Uchigashima M, Saigusa M, Yamashita H, Miyake S, Fujita K, Nakajima M, Nishijima M. 2009. Development of a novel immunoaffinity column for aflatoxin analysis using an organic solvent-tolerant monoclonal antibody. J Agric Food Chem 57(19):8728-34. Uchigashima M, Yamaguchi Murakami Y, Narita H, Nakajima M, Miyake S. 2012. Development of an immuno-affinity column for ochratoxin analysis using an organic solvent-tolerant monoclonal antibody. Methods 56(2):180-5. Vicam. 2009. Aflatest: Instruction Manual for HPLC use. Watertown (US): VICAM. Wongsuttichot P. 2007. Development of an immunoaffinity column from polyclonal antibodies specific to aflatoxin B1 in peanut determination. MS thesis. Bangkok (TH): Kasetsart University. Zaki R, Bulgiba A, Ismail R, Ismail NA. 2013. A systematic review of statistical methods used to test for reliability of medical instruments measuring continuous variables. Iran J Basic Med Sci 16(6): 803–7. Zhao M, Liu Y, Li Y, Zhang X, Chang W. 2003. Development and characterization of an immunoaffinity column for the selective extraction of bisphenol A from serum samples. J Chromatogr B Analyt Technol Biomed Life Sci. 783(2):401-10. 126 BIOTROPIA Vol. 24 No. 2, 2017 Page 1 Page 2 Page 3 Page 4 Page 5 Page 6 Page 7 Page 8 Page 9 Page 10 Page 11 Page 12 Page 13