Multiplex Lateral Flow Immunoassay for Mycotoxin Determination

E-mail: [email protected]. Cite this:Anal. Chem. 86, 10 ... Citation data is made available by participants in Crossref's Cited-by Linking service. Fo...
0 downloads 0 Views 1MB Size
Download PDF
Article pubs.acs.org/ac

Multiplex Lateral Flow Immunoassay for Mycotoxin Determination Suquan Song,†,∥ Na Liu,†,∥ Zhiyong Zhao,† Emmanuel Njumbe Ediage,‡ Songling Wu,§ Changpo Sun,§ Sarah De Saeger,‡ and Aibo Wu*,† †

Institute for Agro-food Standards and Testing Technology, Laboratory of Quality & Safety Risk Assessment for Agro-products (Shanghai), Ministry of Agriculture, Shanghai Academy of Agricultural Sciences, 1000 Jinqi Road, Shanghai 201403, China ‡ Laboratory of Food Analysis, Ghent University, Harelbekestraat 72, 9000 Ghent, Belgium § Academy of State Administration of Grain P.R.C, No. 11 Baiwanzhuang Avenue, Xicheng District, Beijing 100037, China S Supporting Information *

ABSTRACT: A new lateral flow immunoassay (LFA) is proposed for qualitative and/or semiquantitative determination of aflatoxin B1 (AFB1), zearalenone (ZEA), deoxynivalenol (DON), and their analogues (AFs, ZEAs, DONs) in cereal samples. Each of the mycotoxin specific antibody was class specific and there was no cross reactivity to other groups of compounds. The visual limits of detection (vLOD) of the strip were 0.03, 1.6, and 10 μg/kg for AFB1, ZEA and DON, respectively. The calculated limits of detection (cLOD) were 0.05, 1, and 3 μg/kg, respectively. Meanwhile the cutoff values were achieved at 1, 50, and 60 μg/kg for AFB1, ZEA and DON, respectively. Recoveries ranged from 80% to 122% and RSD from 5% to 20%. Both the vLOD and cLOD for the three mycotoxins were lower than the EU maximum levels. Analysis of naturally contaminated maize samples resulted in a good agreement between the multiplex LFA and LC−MS/MS (100% for DONs and AFs, and 81% for ZEAs). Careful analysis of the results further explained the general overestimation of LFA compared to chromatographic methods for quantification of mycotoxins.

M

immune function.10 The AFs are of great concern due to their highest toxicity and are potent cancer-promoting agents, causing liver cirrhosis or primary liver carcinomas.11 Because of the serious threat posed by these mycotoxins and to guarantee food safety, most countries have established regulations to control mycotoxins contamination in food and feed. Typically, the European Commission has set the maximum levels (MLs) for the sum of AFs (AFB1, AFB2, AFG1, and AFG2) at 4 μg/kg and that for AFB1 at 2 μg/kg in cereals and products derived from cereals. Likewise, the MLs for DON and ZEA in unprocessed maize are set at 1750 μg/kg and 200 μg/kg, respectively.12,13 The increasing awareness about mycotoxins as well as the intensifying legislative framework worldwide has aroused the need for fast and efficient analytical methods for monitoring mycotoxins in food and feed.14,15 Chromatographic-based methods such as liquid chromatography tandem mass spectrometry (LC−MS/MS) have been extensively used as a confirmatory method in recent years.4,16−18 In spite of the numerous advantages that such techniques offer, they are timeconsuming, expensive, and labor-intensive with extensive sample preparation. They require skilled technicians for operation and are unsuitable for field applications.19 Immunochemical techniques such as enzyme-linked immunosorbent

ycotoxins are secondary metabolites produced by many invading species of filamentous fungi, which contaminate various agricultural commodities under favorable temperature and humidity conditions.1 Cereals, such as corn, wheat, and rice, are examples of plant-derived products especially susceptible to fungal infestation. About 25% of the world’s food crops are contaminated by mycotoxins, resulting in an estimated annual loss of 1 billion metric tons of food products equivalent to about 5 billion dollars per year.2 Besides, mycotoxins are carcinogenic, nephrotoxic, hepatotoxic, neurotoxic, mutagenic, estrogenic, and immunosuppressive agents. They may usually pose great threat to human and livestock when they enter the food chain through contaminated cereals or feedstuffs.1 Owning to their wide contamination with high frequency, mycotoxins have become a major concern in global food safety issues.3 Among the mycotoxins that have been found in cereals and derived products, aflatoxins (AFs), zearalenone (ZEA), and related compounds (ZEAs) as well as deoxynivalenol (DON) and related compounds (DONs) have been reported as the most frequently occurring toxins.4−7 AFs are commonly produced by Aspergillus molds while ZEAs and DONs are mainly produced by Fusarium species. ZEA and its analogues are suspected to be triggering factors for central precocious puberty observed in adolescent females in the United States8 and have also been reported to be carcinogenic with symptoms of reproductive toxicity.9 Ingestion of high dose of DON is reported to cause acute gastroenteritis with vomiting effects, while low dose impairs growth as well as an altered © 2014 American Chemical Society

Received: February 8, 2014 Accepted: April 18, 2014 Published: April 18, 2014 4995

dx.doi.org/10.1021/ac500540z | Anal. Chem. 2014, 86, 4995−5001

Analytical Chemistry

Article

Figure 1. (A) Scheme of the multiplex lateral flow immunoassay (LFA) for multiplex myxotoxins. (B1) Semiquantitative analysis platform for LFA. (B2) Qualitative analysis platform for LFA. Strips 1 to 9 are the schematic illustrations of detection results. (1) AFs (−), ZEAs (−), DONs (−); (2) AFs (+), ZEAs (−), DONs (−); (3) AFs (−), ZEAs (+), DONs (−); (4) AFs (−), ZEAs (−), DONs (+); (5) AFs (−), ZEAs (+), DONs (+); (6) AFs (+), ZEAs (−), DONs (+); (7) AFs (+), ZEAs (+), DONs (−); (8) AFs (+), ZEAs (+), DONs (+); (9) invalid result.

is the first report on multiclass analysis of mycotoxins and their analogues by LFA with three monoclonal antibodies.

assay (ELISA) have been widely used to screen mycotoxins.20,21 However, conventional ELISA methods require some laboratory operations which are time-consuming. Lateral flow immunoassay (LFA) has been widely developed for rapid detection (qualitative) of single or multiple analytes,3,22−24 with the possibility to inaccurate results due to individual subjectivity. Recently quantitative immuno-chromatographic assays have been developed for single analyte detection.24−27 However, considering co-occurrence of mycotoxins in cereal grains efforts to design simultaneous detection of multimycotoxins were made in a pioneering study.28,29 In spite of these developments, there are still some theoretical and technological gaps, unanswered questions, and continuous demands from end-users which call for more research with the existing technology especially in the area of multiplex mycotoxin determination. Plant metabolites can conjugate with mycotoxins to form the masked mycotoxins, which stay in the plant vacuoles and cell wall and escape routine monitoring. Several studies have highlighted their potential threat to human and animal health as some of these masked forms can undergo partial or total cleavage reverting to the native compound and hence are able to exert the same toxic effect as the native compounds.30,31 Realistically the coexistence of the parent mycotoxins and their masked forms can always lead to underestimation of the total mycotoxin content in the sample and therefore underestimation of the exposure of consumers at doses exceeding the MLs.32 A possible remedy is to broaden the current analytical methods for mycotoxins by integrating conjugates and other mycotoxin metabolites in the routine monitoring process. With the availability of three class specific monoclonal antibodies, a rapid immunoassay platform for simultaneous multiclass qualification or semiquantification of deoxynivalenol, zearalenone, and aflatoxins and their main conjugates or analogues was successfully established. The developed assay was applied for the determination of mycotoxins in spiked and naturally contaminated cereal samples. To our knowledge, this

■

EXPERIMENTAL SECTION Experimental Section. The mycotoxin standards aflatoxin B1 (AFB1), aflatoxin B2 (AFB2), aflatoxin G1 (AFG1), aflatoxin G2 (AFG2), zearalenone (ZEA), zearalanone (ZAN), α-zearalenol (α-ZOL), β-zearalanol (β-ZOL), deoxynivalenol (DON), 3-acetyldeoxynivalenol (3-AC-DON), 15acetyldeoxynivalenol (15-AC-DON), deoxynivalenol-3-glucoside (D3G), T-2, fumonisin B1 (FB1), fumonisin B2 (FB2), and ochratoxin A (OTA) were provided by Sigma-Aldrich (St. Louis). The monoclonal antibodies (mAbs) against AFB1, ZEA, and DON together with the conjugates AFB1-bovine serum albumin (BSA), ZEA-BSA, as well as DON-BSA were developed in our lab. Hydrogen tetrachloroaurate (III) hydrate (HAuCl4·3H2O), sodium citrate (C6H5Na3O7·2H2O), and goat antimouse immunoglobulin (IgG) were purchased from SigmaAldrich (St. Louis). Sample pad (spun bonded polyester, 6613), gold conjugate pad (borosilicate glass fiber with PVA binder, 8964), as well as nitrocellulose (NC) membranes (vivid 170) were purchased from PALL Corporation (NY). Absorbing pad was purchased from Shanghai Goldbio Tech Co., Ltd. (Shanghai, China). Water was purified with a Milli-Q system from Millipore (Bedford, MA). All the organic solvents in the study were of analytical reagent grade. Apparatus. HGS510 dispenser and sprayer platform and HGS201 cutter (Hangzhou Autokun Technology Co., Ltd., Hangzhou, China) were used to prepare test strips. The CHR100 strip reader was purchased from KAIWOOD Technology Co., Ltd. (Taiwan, China). A Shimadzu LC/MS8030 triple quadrupole mass spectrometer (LC/MS/MS) equipped with an electrospray ionization interface (Shimadzu, Kyoto, Japan) was used in the study. Samples. Wheat and maize samples were from local markets in China, Shanghai province. Because certified blank samples were not available, samples with undetected levels of the analytes after LC−MS/MS screening were chosen as 4996

dx.doi.org/10.1021/ac500540z | Anal. Chem. 2014, 86, 4995−5001

Analytical Chemistry

Article

“blank” and used in spiking and recovery experiments. For sample preparation, 5 g of samples was extracted with 25 mL of methanol/water (70:30, v/v) for 15 min. After centrifugation at 4000g for 10 min, 5 mL of supernatant was dried and the residue was reconstituted with the same (5 mL) volume of PBS (0.01 M, pH 7.4). Preparation of LFA Strip. Colloidal gold (CG) with a mean diameter of 25 nm and anti-AFB1-CG conjugates were prepared by the method described by Liu et al.33 Furthermore, the anti-DON-CG and anti-ZEA-CG conjugates were prepared with the same procedure. Before preparing the LFA strips, parameters such as the pH value, the proportion of different constituents in the blocking buffer, the ratio of CG labeled mAb, the speed of the dispensing platform, and the volume of liquid were optimized. The LFA strips consisted of four parts as follows: sample pad, conjugate pad, NC membrane, and absorbent pad (Figure 1). The absorbent pads were employed without any pretreatment. The sample pads and conjugate pads were first blocked with blocking buffer (0.1 M PBS (pH 7.4), containing 10% (w/v) sucrose, 1% (w/v) BSA, and 0.5% (w/v) trehalose) followed by overnight drying at 37 °C. For preparing the conjugate pads, the three CG labeled mAbs (CG-mAbs) were mixed at equal ratio and then diluted 10-fold with 0.1 M PBS. Afterward, the mixture was dispensed onto the glass fiber at a speed of 2 μL/cm and dried. For preparing the test zones, the three capture reagents (DON-BSA, ZEA-BSA, and AFB1BSA) and goat antimouse immunoglobulin (0.25 mg/mL) were separately spotted onto the NC membrane in turn at a jetting rate of 0.7 μL/cm to generate three test lines and one control line. The four lines were positioned at a 3.8 mm interval. Finally the sample pad, conjugated pad, NC membrane, and absorbent pad were laminated onto a plastic backing and divided into strips, which were installed in the shell and stored with desiccators at room temperature until use. Qualitative or Semiquantitative Immunoassay Procedure. For qualitative assay, 60 μL of mycotoxin standard or extracted sample solution were loaded onto the sample pad of the LFA strip. Driven by capillary forces, the liquid migrated to the absorbent pad. The CG-mAbs, immobilized on the conjugate pad, were then redissolved in the solution and reacted with the mycotoxins (if present) while the whole complex migrated along the membrane. Upon reaching the test line, CG-mAbs were captured by the corresponding capture reagents (mycotoxin-BSA) resulting in the appearance of a pink line. The color intensity (CI) of the test line was inversely correlated with the mycotoxin concentration in the sample. In the absence of target mycotoxin, the largest amounts of CGmAbs were trapped by the competitive antigen and the most intensive red color band developed on the corresponding test line. However, if there were enough target analytes in the sample, all the CG-mAbs were occupied. Therefore, no mAbs could react with the mycotoxin-BSA reagents immobilized on the NC membrane and no visible band appeared in the test line. The control line should always be visible because of the reaction between CG-mAbs and goat antimouse IgG, which was considered to be an indicator of the good functionality of the test. For semiquantitative analysis, the intensity of the test lines was determined after 15 min using the strip reader and the data were expressed as relative optical density (ROD). ROD is the ratio of the optical densities of the positive (B) to the negative (B0) sample. The concentration of the three analytes in the samples were quantified from a calibration curve (B/B0 × 100%

versus the concentration of each analyte), which was run simultaneously. Calibration, Sensitivity, and Specificity of the LFA. The standard curve for each analyte was constructed in matrix by spiking at different concentrations for the different analytes. The concentration range was 0.1 to 100 μg/kg for DON and ZEA and 0.025 to 10 μg/kg for AFB1 (Figure 2). The standard curve was fitted using the four parameter logistic equation by SigmaPlot (version 12.0)

Figure 2. Calibration curves for multiplex mycotoxins (AFB1, ZEA, and DON) detection with the developed LFA. B/B0 is the ratio of the optical densities of the positive sample to the negative sample. The inset shows the visual detection limits of three mycotoxin standards. AFB1/ZEA/DON concentrations were as follows (from left to right): 0/0/0, 0.01/0.4/2.5, 0.02/0.8/5, 0.03/1.6/10, 0.06/3/20, 0.12/6/30, 0.3/12/40, 0.5/25/50, 1/50/60 μg/kg. Error bars represent standard error with n = 6.

The assay sensitivity was evaluated by analyzing a series of concentrations of the mycotoxin mixture. The visual limit of detection (vLOD) for qualitative evaluation was defined as the minimum concentration that gave very weak color intensity in the test line visibly different from that of the negative control line (very intense coloration). The cutoff value was the concentration that gave a complete disappearance of the visible band. For semiquantitative evaluation, the calculated limit of detection (cLOD) was defined as the concentration at which B/B0 equals to 80% (thus 20% inhibition of the signal, IC20), and the IC80 was used to evaluate the maximum detection ability of the LFA in this study. The specificity, expressed as cross reactivity (CR), was evaluated by assessing the recognition of the specific analytemAb toward other mycotoxins or analogues. The CR was expressed as the percentage of the concentration at 50% inhibition (IC50) of the analogues to the corresponding target analyte. LC−MS/MS Analysis. The samples were first extracted with 25 mL of acetonitrile/water (84:16, v/v) for 60 min. Before analysis, 1 mL of supernatant was diluted by an equal volume of a methanol/water mixture (20:80, v/v). The analytical column was an Agilent poroshell 120 EC-C18 (100 mm × 3 mm, 2.7 μm). A mobile phase consisting of ammonium acetate (5 mM) was used at a flow rate of 0.3 mL/min. The gradient elution program applied was as follows: 0−1 min, 20% methanol; 1−2 min, 20−90% methanol; 2−6 min, 90% methanol; 6−6.2 min, 4997

dx.doi.org/10.1021/ac500540z | Anal. Chem. 2014, 86, 4995−5001

Analytical Chemistry

Article

90−20% methanol; 6.2−8 min, 20% methanol. The injection volume was 5.0 μL. The MS conditions were thoroughly optimized for each mycotoxin. The optimal parameters were listed in Table S-1 in the Supporting Information. Quantification of mycotoxins was performed by measuring peak areas in the MRM mode. Safety Precautions. AFB1 is a known liver carcinogen, therefore direct exposure and laboratory contamination should be avoided. All experiments were carried out in the fume hood, and researchers should wear laboratory coat, safety glasses, gloves, mouth-muffle, and face mask.

Table 1. Cross Reactivity (CR) of Analytes with Monoclonal Antibody Detected by LFAa class AFs

ZEAs

■

DONs

RESULTS AND DISCUSSION Optimization of the LFA Strips. A variety of labels have been used for signal generation in immunoassays, such as enzyme, quantum dots, dye-loaded liposomes, carbon nanoparticles, or magnetic beads.34 Because of its ease of conjugation with antibodies, CG has been widely used in LFA. A 25 nm CG was chosen since it has been reported35,36 to show good stability and sensitivity compared to other commercially available sizes. Hence, a 25 nm CG labeled antibody was adopted as a signal molecule. For convenience, three different conjugates of CG-mAbs were mixed together and dispensed onto a single conjugate pad (Figure 1) instead of coating them on three overlapping conjugate pads. With the use of the single conjugate pad, the pink colored lines developed within 5 min. However, 15 min was recommended for the development of a more stable intensity with the reason that, after this time point, the B/B0 ratio of the samples exhibited the lowest value which implied good assay sensitivity. Furthermore, 60 μL of sample extract was determined sufficient to dissolve the CG-mAbs and develop a satisfactory consequence in the end-results. The performance of different types of NC membrane (PALL vivid, Whatman-AE99 and Sartorius CN 140) were also evaluated. PALL vivid 170 was chosen for its good functionality when compared with the other NC membranes (Whatman-AE99 and Sartorius CN140). Other parameters also optimized included the concentration of the coated antigens (0.5 mg/mL) and goat antimouse IgG (0.25 mg/mL), while the antibody concentration was 10 μg/mL for the CG conjugate of anti-DON, antiZEA, and anti-AFB1. Determination of Limits of Detection and Inhibition Concentrations for the Different Test Lines. The inhibition curves obtained with the strip reader for the different analytes are shown in Figure 2. The calculated LOD (cLOD) for AFB1, ZEA, and DON were 0.05, 1, and 3 μg/kg. The visual cutoff was 1 μg/kg, 50 μg/kg, and 60 μg/kg for AFB1, ZEA, and DON, which were far below the MLs established in the EU.12,13 The IC50 values were calculated to be 0.2 μg/kg, 4 μg/ kg, and 10 μg/kg for AFB1, ZEA, and DON, respectively (Figure 2 and Table 1). The vLOD of the LFA for qualitative analysis was determined with different mycotoxin concentrations spiked in a blank sample. As shown in the inset of Figure 2, the intensity of the test lines decreased with increasing mycotoxin concentration. At the following concentrations, 0.03 μg/kg for AFB1, 1.6 μg/kg for ZEA, and 10 μg/kg for DON, the intensity of the test line showed an obvious difference from the control line. Hence these concentrations were set as the vLODs. By performing background subtraction and normalization, the CI was within uniform standard, which significantly improved the detection capability and greatly reduced the uncertainty of measurements near the cutoff value.

a

analytes

antibody

IC50 (μg/kg)

CR (%)

AFB1 AFB2 AFG1 AFG2 ZEA ZAN α-ZOL β-ZOL DON 3-AC-DON 15-AC-DON D3G

Anti-AFB1 mAb

0.2 0.2 0.3 0.2 4 1 2 2 10 8 200 20

100 124 66 96 100 251 193 153 100 127 5 49

Anti-ZEA mAb

Anti-DON mAb

The analysis was performed in standard solution (n = 4).

Cross Reactivity. For the simultaneous detection of three mycotoxins by the LFA, the corresponding capture antigens (AFB1-BSA, DON-BSA, or ZEA-BSA) were immobilized at different sites on the strip constituting different test zones. It was shown that, when a single CG-mAb specific (anti-AFB1 mAb or anti-ZEA mAb or anti-DON mAb) was loaded directly onto the NC membrane, only the corresponding test line appeared (AFB1-BSA/ZEA-BSA/DON-BSA) with no pink line for the other mycotoxins. This implies each of the three antibodies were each specific to the corresponding mycotoxin(s) and no cross reactivity with any other of the two mycotoxin test zones. The cross reactivity toward other mycotoxins was also evaluated. The results are illustrated in Table 1. Most of the antibodies exhibited high cross reactivity to the class specific analogues. For ZEA and its analogues, high CRs were observed with ZAN (251%), α-ZOL (193%), and β-ZOL (153%). For DON and its analogues, high CRs were also observed with 3AC-DON (127%) and D3G (49%), with the exception of 15AC-DON (5%). For the AFs, high CRs for AFB2 (124%), AFG1 (66%), and AFG2 (96%) were observed. For the mycotoxins not belonging to the same class, the cross reactivity was below 0.1% with IC50 higher than 1000 μg/kg. Hence the antibodies used in the development of the LFA were class specific. With the LFA developed in this work, the masked mycotoxins could also be determined to some extent (D3G). Hence this multicomponent LFA could also be used for multimycotoxin determination. Optimization of the Sample Preparation and Loading Step. Two approaches were evaluated. In the first approach, the extract was diluted 3.5 times in PBS after extraction and then loaded onto the LFA. While in the second approach, the sample extract was preconcentrated (through evaporation) and reconstituted in injection solvent. Results are shown in Table S2 in the Supporting Information. The IC80 and IC50 for DON and AFB1, respectively, were far higher than their respective ML of 1750 μg/kg and 9 μg/kg. Faced with these problems, we opted for a second approach where the sample extract was first evaporated to dryness and then equal volumes of PBS were used to reconstitute the sample before loading onto the strip. With the second approach, the sensitivity of the LFA toward the three mycotoxins improved significantly. Hence the IC50 for all these analytes were far below the MLs. Validation with Cereal Samples. The LFA was validated for maize and wheat samples. The parameters were obtained by 4998

dx.doi.org/10.1021/ac500540z | Anal. Chem. 2014, 86, 4995−5001

Analytical Chemistry

Article

Table 2. Recoveries (R) of AFB1, ZEA, and DON in Spiked Maize and Wheat (n = 6) LFA sample

analytes

spiked concentration (μg/kg)

AFB1

0.5 1 2 10 25 50 40 80 150 0.5 1 2 10 25 50 40 80 150

ZEA maize DON

AFB1

ZEA wheat DON

LC−MS/MS

mean ± SD (μg/kg)

R (%)

RSD (%)

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

106 98 87 96 88 87 80 82 112 122 95 85 92 120 103 88 88 86

17 19 13 5 7 7 19 12 11 5 18 8 13 20 17 7 7 7

0.5 1 2 10 22 44 32 66 169 0.6 1 2 9 30 52 35 70 129

0.1 0.2 0.2 0.5 1 3 6 8 19 0.1 0.2 0.1 1.3 5 9 2 5 9

mean ± SD (μg/kg)

R (%)

RSD (%)

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

128 86 83 80 102 118 75 88 72 70 93 97 86 106 115 82 106 78

19 17 4 21 13 3 16 9 3 7 6 10 18 4 5 8 13 8

0.6 0.9 1.7 8.0 25.6 59.1 30.0 70.2 107.5 0.4 0.9 1.9 8.6 26.6 57.3 32.6 84.7 116.9

0.1 0.1 0.1 1.6 3.4 2.0 4.7 6.5 2.7 0.1 0.2 0.2 1.5 1.0 2.9 2.5 10.9 9.9

Table 3. Comparison of the Analysis Results for AFs, ZEAs, and DONs in Naturally Contaminated Samples by the Developed LFA and LC−MS/MS (n = 4)a LFA

LC−MS/MS

maize sample

AFB1 (μg/kg)

ZEA (ZAN, α-ZOL, β-ZOL)b (μg/kg)

DON (3-Ac-DON, D3G)c (μg/kg)

AFB1 (μg/kg)

ZEA (μg/kg)

DON (μg/kg)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

n/d n/d n/d n/d n/d n/d n/d n/d n/d n/d n/d n/d n/d n/d n/d n/d n/d n/d n/d n/d n/d

113 (45, 59, 74) P n/d 29 (11, 15, 19) 116 (46, 60, 75) 61 (25, 32, 40) 16 (7, 8, 11) P P P P 128 (51, 66, 84) 32 (13, 16, 21) P n/d 20 (8, 10, 13) 66 (26, 34, 43) n/d n/d 25 (10, 13, 16) n/d

P P 104 (82, 215) 69 (54, 141) P P 169 (132, 347) P P P P P P 237 (186, 487) 66 (52,135) 206 (162, 425) 403 (317, 830) 193 (152, 397) 119 (93, 244) 358 (281, 736) 124 (97, 255)

n/d n/d n/d n/d n/d n/d n/d n/d n/d n/d n/d n/d n/d n/d n/d n/d n/d n/d n/d n/d n/d

81.2 103.8 6.7 17.6 82.3 49.8 18.5 447.7 381.4 149.7 999.3 85.5 20.0 276.4 7.2 11.6 46.3 8.6 7.0 17.5 7.5

1730.5 1267.9 83.8 255.5 645.5 1256.7 151.7 6149.0 4607.7 960.5 9821.7 1829.0 574.0 260.2 113.9 258.2 520.9 221.2 131.2 391.5 177.7

a

n/d = not detected. P = above cutoff. bThe calculated concentration of each mycotoxin compound according to the concentration of ZEA. cThe calculated concentration of each mycotoxin compound according to the concentration of DON.

be applied for simultaneous quantification of mycotoxins in real samples. To evaluate the recoveries of the sample preparation, spiked blank maize and wheat were investigated. The parameters were obtained by spiking the cereal samples at three concentration levels. The spiked samples were also analyzed in-parallel by LC−MS/MS. The results (Table 2) demonstrated good agreement between the measured values and the fortified concentration in both LFA and LC−MS/MS, with recoveries ranging from 80% to 122% for LFA and from 70% to 128% for

spiking the samples at six concentration levels and quantified by use of matrix-matched calibration curves. The results are shown in Table S-3 in the Supporting Information. The coefficients of determination (R2) for the different analytes were higher than 0.97, which indicated good linearity of the analytical ranges. The parameters of cLOD, IC50, cutoff value, vLOD, and working range of the qualitative and semiquantitative LFA are also listed in Table S-3 in the Supporting Information. Both the vLOD and cLOD for three mycotoxins in the two cereal matrixes were lower than the MLs; therefore, the method could 4999

dx.doi.org/10.1021/ac500540z | Anal. Chem. 2014, 86, 4995−5001

Analytical Chemistry

■

ACKNOWLEDGMENTS The authors thank the funds from the National Basic Research Program of China (Grant 2013CB127801), Pujiang Plan of Shanghai (Grant No. 13PJ1407200), Shanghai Municipal Commission for Science and Technology (Grant 13231202800), and the Chinese-Belgian Joint Project of BELSPO, MOST, China (Grant S2012GR0 016) and Belgium (Grant BL/02/C58).

LC−MS/MS. Moreover, acceptable RSDs ranging from 5% to 20% were also achieved with LFA. Results of Naturally Contaminated Samples. Analysis of 21 naturally contaminated maize samples was performed using the developed LFA followed by LC−MS/MS for confirmation (Table 3). With the LC−MS/MS analysis, all 21 samples were found positive with DON and ZEA and no AFB1. Of the 21 samples, 4 were contaminated above the MLs of 1750 μg/kg for DON and 200 μg/kg for ZEA. Using the LFA, all 21 samples were contaminated with DONs (100% agreement) whereas only 17/21 (81%) samples were found contaminated with ZEAs. As listed in Table S-3 in the Supporting Information, the cLOD of the LFA for ZEA was 9 μg/kg, which was higher than the concentrations confirmed by LC−MS/MS in the selected five samples. For concentrations above the cutoff of the LFA, a value “P” (positive) was marked, which meant the concentration was above the calculated range of the LFA. However, for samples 2 and 10, the concentration of ZEA confirmed by LC−MS/MS was 104 μg/kg and 150 μg/kg, which was within the detection range of the LFA (9 μg/kg to 186 μg/kg for ZEA), but the data was still marked as “P”. Such overestimation of ZEA could be the result of an actual contamination (cocontamination) of “masked-ZEAs” in the sample. Also for sample 13, an overestimation for DON was detected. This result explains the general overestimation of LFA compared to chromatographic methods for quantification of mycotoxins.37,38

■

CONCLUSION In the present study, a multiplex LFA platform was constructed and validated which could simultaneously qualify or semiquantify AFs, ZEAs, and DONs within a total time of 15 min. The technical platform offers multiple advantages for simplicity, rapidity, sensitivity, cost-effectiveness, and time-efficiency. The validation results of the method in the spiked samples showed that the developed approach was reliable and could be employed for multiplex mycotoxin detection in cereals. Because the monoclonal antibodies adopted were class specific, the LFA strip could simultaneously detect three groups of mycotoxins in a single assay. In real sample analysis, the vLOD and cLOD for three mycotoxins were lower than the EU maximum levels. The proposed method was successfully applied to naturally contaminated maize samples. The results were in good agreement with those obtained using confirmatory LC−MS/ MS. The overestimation of some mycotoxins was related to the recognition capability of the class specific antibodies utilized in the LFA. ASSOCIATED CONTENT

S Supporting Information *

Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Phone: +86-21-37196975. Fax: +86-21-62203612. E-mail: [email protected]. Author Contributions ∥

REFERENCES

(1) Bennett, J. W.; Klich, M. Clin. Microbiol. Rev. 2003, 16, 497−516. (2) Al-Taher, F.; Banaszewski, K.; Jackson, L.; Zweigenbaum, J.; Ryu, D.; Cappozzo, J. J. Agric. Food Chem. 2013, 61, 2378−2384. (3) Li, X.; Li, P.; Zhang, Q.; Li, R.; Zhang, W.; Zhang, Z.; Ding, X.; Tang, X. Biosens. Bioelectron. 2013, 49, 426−432. (4) Song, S.; Ediage, E. N.; Wu, A.; De Saeger, S. J. Chromatogr., A 2013, 1292, 111−120. (5) Cetin, Y.; Bullerman, L. B. J. Agric. Food Chem. 2006, 54, 1949− 1955. (6) Zhang, D.; Li, P.; Yang, Y.; Zhang, Q.; Zhang, W.; Xiao, Z.; Ding, X. Talanta 2011, 85, 736−742. (7) Shim, W. B.; Kim, K. Y.; Chung, D. H. J. Agric. Food Chem. 2009, 57, 4035−4041. (8) Massart, F.; Meucci, V.; Saggese, G.; Soldani, G. J. Pediatr. 2008, 152 (690−5), 695 e1. (9) Kim, I. H.; Son, H. Y.; Cho, S. W.; Ha, C. S.; Kang, B. H. Toxicol. Lett. 2003, 138, 185−192. (10) Pestka, J. J. Animal Feed Sci. Technol. 2007, 137, 283−298. (11) IARC. Some naturally occurring substances: food items and constituents, heterocyclic aromatic amines and mycotoxins monographs of the Evaluation of the carcinogenic Risk of Chemical to Human, Lyon, France; 1993, Vol. 56, pp 1−151. (12) European Community Council Regulation No. 1881/2006, Official Journal of the European Union 2006, L 364/5. (13) European Community Council Regulation No. 1126/2007, Official Journal of the European Union 2007, L 255/14−17. (14) Deng, G.; Xu, K.; Sun, Y.; Chen, Y.; Zheng, T.; Li, J. Anal. Chem. 2013, 85, 2833−40. (15) Mak, A. C.; Osterfeld, S. J.; Yu, H.; Wang, S. X.; Davis, R. W.; Jejelowo, O. A.; Pourmand, N. Biosens. Bioelectron. 2010, 25, 1635− 1639. (16) Kong, W.; Wei, R.; Logrieco, A. F.; Wei, J.; Wen, J.; Xiao, X.; Yang, M. Food Chem. 2014, 146, 320−326. (17) Anfossi, L.; Baggiani, C.; Giovannoli, C.; D’Arco, G.; Giraudi, G. Anal. Bioanal. Chem. 2013, 405, 467−480. (18) Huang, L. C.; Zheng, N.; Zheng, B. Q.; Wen, F.; Cheng, J. B.; Han, R. W.; Xu, X. M.; Li, S. L.; Wang, J. Q. Food Chem. 2014, 146, 242−249. (19) Grebe, S. K.; Singh, R. J. Clin. Biochem. Rev. 2011, 32, 5−31. (20) Turner, N. W.; Subrahmanyam, S.; Piletsky, S. A. Anal. Chim. Acta 2009, 632, 168−180. (21) Tang, X.; Li, X.; Li, P.; Zhang, Q.; Li, R.; Zhang, W.; Ding, X.; Lei, J.; Zhang, Z. PLoS One 2014, 9, e85606. (22) Kolosova, A. Y.; De Saeger, S.; Eremin, S. A.; Van Peteghem, C. Anal. Bioanal. Chem. 2007, 387, 1095−1104. (23) Tang, D.; Zhong, Z.; Niessner, R.; Knopp, D. Analyst 2009, 134, 1554−1560. (24) Laura, A.; Gilda, D. A.; Claudio, B.; Cristina, G.; Gianfranco, G. Food Control 2011, 22, 1965−1970. (25) Xu, Y.; Huang, Z.-B.; He, Q.-H.; Deng, S.-Z.; Li, L.-S.; Li, Y.-P. Food Chem. 2010, 119, 834−839. (26) Molinelli, A.; Grossalber, K.; Krska, R. Anal. Bioanal. Chem. 2009, 395, 1309−1316. (27) Anfossi, L.; D’Arco, G.; Calderara, M.; Baggiani, C.; Giovannoli, C.; Giraudi, G. Food Addit. Contam., Part A: Chem. Anal. Control Expo. Risk Assess. 2011, 28, 226−234. (28) Wang, Y. K.; Yan, Y. X.; Ji, W. H.; Wang, H. A.; Li, S. Q.; Zou, Q.; Sun, J. H. J. Agric. Food Chem. 2013, 61, 5031−5036.

■

■

Article

S.S. and N.L. contributed equally to this work.

Notes

The authors declare no competing financial interest. 5000

dx.doi.org/10.1021/ac500540z | Anal. Chem. 2014, 86, 4995−5001

Analytical Chemistry

Article

(29) Nivarlet, N.; Lattanzio, V.; Lippolis, V.; Della Gatta, St.; Huet, A. C.; Visconti, A.; Delahaut, P.; Granier, B. Multiplex lateral flow immunoassay for Fusarium toxins” CONffIDENCE: Contaminants in food and feed: Inexpensive detection for control of exposure, Ghent, Belgium, 2011; http://www.conffidence.eu/img/enewsletter/enews_ poster_mycosensor_GhentSymposium2011.pdf. (30) De Boevre, M.; Jacxsens, L.; Lachat, C.; Eeckhout, M.; Di Mavungu, J. D.; Audenaert, K.; Maene, P.; Haesaert, G.; Kolsteren, P.; De Meulenaer, B.; De Saeger, S. Toxicol. Lett. 2013, 218, 281−292. (31) Dall’Erta, A.; Cirlini, M.; Dall’Asta, M.; Del Rio, D.; Galaverna, G.; Dall’Asta, C. Chem. Res. Toxicol. 2013, 26, 305−312. (32) Berthiller, F.; Crews, C.; Dall’Asta, C.; Saeger, S. D.; Haesaert, G.; Karlovsky, P.; Oswald, I. P.; Seefelder, W.; Speijers, G.; Stroka, J. Mol. Nutr. Food Res. 2013, 57, 165−186. (33) Liu, G.; Han, Z.; Nie, D.; Yang, J.; Zhao, Z.; Zhang, J.; Li, H.; Liao, Y.; Song, S.; De Saeger, S.; Wu, A. Food Control 2012, 27, 200− 205. (34) Chan, C. P.; Cheung, Y. C.; Renneberg, R.; Seydack, M. Adv. Biochem. Eng. Biotechnol. 2008, 109, 123−154. (35) Wang, Y.-K.; Shi, Y.-B.; Zou, Q.; Sun, J.-H.; Chen, Z.-F.; Wang, H.-a.; Li, S.-Q.; Yan, Y.-X. Food Control 2013, 31, 180−188. (36) Guo, Y.; Ngom, B.; Le, T.; Jin, X.; Wang, L.; Shi, D.; Wang, X.; Bi, D. Anal. Chem. 2010, 82, 7550−7555. (37) Dall’Asta, C.; Galaverna, G.; Aureli, G.; Dossena, A.; Marchelli, R. World Mycotoxin J. 2008, 1, 237−246. (38) Ruprich, J.; Ostry, V. Cent. Eur. J. Public Health 2008, 16, 34− 37.

5001

dx.doi.org/10.1021/ac500540z | Anal. Chem. 2014, 86, 4995−5001