Immunoassay for the Detection of Animal Central ... - ACS Publications

Apr 25, 2016 - (22) The optimized icELISA was then used to detect the presence of spiked CNT (bovine brain) in the processed meat and feed samples...
0 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/JAFC

Immunoassay for the Detection of Animal Central Nervous Tissue in Processed Meat and Feed Products Qinchun Rao,† Juergen A. Richt,‡ and Yun-Hwa Peggy Hsieh*,† †

Department of Nutrition, Food and Exercise Sciences, Florida State University, Tallahassee, Florida 32306-1493, United States Diagnostic Medicine/Pathobiology, College of Veterinary Medicine, Kansas State University, Manhattan, Kansas 66506, United States

‡

S Supporting Information *

ABSTRACT: An indirect competitive enzyme-linked immunosorbent assay (icELISA) based on the detection of the thermalstable central nervous tissue (CNT) marker protein, myelin basic protein (MBP), was developed to detect animal CNT in processed meat and feedstuffs. Two meat samples (cooked at 100 °C for 30 min and autoclaved at 133 °C for 20 min) of bovine brain in beef and two feed samples (bovine brain meal in beef meal and in soybean meal) were prepared at levels of 0.0008, 0.0031, 0.0063, 0.0125, 0.025, 0.05, 0.1, 0.2, 0.4, 0.8, and 1.6%. An anti-MBP monoclonal antibody (mAb3E3) was produced using the hybridoma technique and characterized using Western blot. The optimized icELISA was CNT-specific without crossreactivity with either meat (beef and pork) or soybean meal samples and had low intra-assay (%CV ≤ 3.5) and interassay variability (%CV ≤ 3.3), with low detection limits for bovine MBP (6.4 ppb) and bovine CNT spiked in both meat (0.05%) and feed (0.0125%) samples. This assay is therefore suitable for the quantitative detection of trace amounts of contaminated animal CNT in processed food and feed products. KEYWORDS: bovine spongiform encephalopathy, central nervous tissue, myelin basic protein, immunoassay, ELISA, specified risk material

■

INTRODUCTION

of animal feed in the U.S. has been recalled due to omitting the cautionary BSE statement on the labels or MBM contamination during processing.12 The major route of BSE transmission to humans, causing vCJD, is likely to be the dietary consumption of BSE-infected beef products.13 During animal slaughtering and meat processing, meat products can easily become contaminated by CNT-based SRM.14 More than 65,713 t of SRM-contaminated cattle products have been recalled since 2003 in the U.S. alone.15 For regulation enforcement, many methods for the detection of bovine CNT contamination in meat products have been developed. Overall, these detection methods can be classified into four types: chromatography,16 spectroscopy,17 polymerase chain reaction (PCR),18 and immunochemical techniques.19 However, no reported method can detect CNT contamination in animal feed products that have been severely heat-treated, for example, under conditions such as the 133 °C/3 bar/20 min required for processing MBM in the EU.20 According to a previous study,21 bovine 18.5-kDa myelin basic protein (MBP), the major central nervous system (CNS) myelin protein, is a very promising candidate for the immunodetection of CNT due to its high thermal stability. In this study, we produced an anti-MBP monoclonal antibody (mAb) and then went on to develop and validate an MBP-specific indirect competitive enzyme-linked immunosorbent assay (icELISA) to further

Transmissible spongiform encephalopathies (TSE) are a group of rare degenerative brain diseases characterized by tiny holes that give the brain a “spongy” appearance1 and include two serious foodborne diseases: variant Creutzfeldt−Jakob disease (vCJD) in humans and bovine spongiform encephalopathy (BSE). There is strong epidemiologic and biochemical evidence suggesting that vCJD and BSE are caused by the same infectious agent, namely, the abnormal prion protein (PrPSc).2 PrPSc is the pathologic isoform of the cellular prion protein PrPc; PrPSc is a proteinaceous infectious particle that lacks nucleic acid and is extremely resistant to heat and proteases.3 PrPSc is mainly found in specified risk material (SRM) such as the central nervous tissue (CNT) of infected animals.4 About 90% of the BSE infectivity is associated with bovine CNT-based SRM from infected cattle.5 It is generally believed that the recycling of animal byproducts from BSE-infected animals in the feed chain in the form of meat and bone meal (MBM) is responsible for the spread of BSE in cattle6 and the BSE contamination of human food products for vCJD in humans.2 To prevent PrPSc contaminated animal products from being sold as food and BSE-infected animal byproducts from being used as feed, many countries and areas have imposed regulations banning their use. In 2001, the European Union (EU) prohibited the feeding of (1) proteins derived from animals to ruminants and (2) processed animal protein to farmed animals.7 In addition, SRM is barred from entering the food and feed chain in Europe.8 Similar to the EU, the U.S. has prohibited (1) the use of ruminant proteins in ruminant feed,9 (2) cattle materials in animal feed,10 and (3) the use of SRM for human food.11 Between 2006 and 2014, more than 41,818 t © XXXX American Chemical Society

Received: February 2, 2016 Revised: April 14, 2016 Accepted: April 24, 2016

A

DOI: 10.1021/acs.jafc.6b00572 J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Article

Journal of Agricultural and Food Chemistry

temperature with properly diluted mAb3E3 (20 ng/mL in the antibody buffer (PBS containing 1% BSA [g/mL] and 0.05% Tween 20 [mL/mL], pH 7.2)). The membrane was then incubated for 2 h with goat antimouse IgG (Fc specific) horseradish peroxidase (antiIgG-HRP) conjugate (Sigma-Aldrich) diluted 1:5,000 in the antibody buffer. The bound antibody was detected with chemiluminescence using the Immun-Star WesternC Chemiluminescent Kit (Bio-Rad) following the manufacturer’s instructions. The image was analyzed with a ChemiDoc XRS system (Bio-Rad) and Quantity One 1-D analysis software (Bio-Rad). Precision Plus Protein WesternC standards (Bio-Rad) were used for molecular weight estimation and purified bovine 18.5-kDa MBP as the positive control. Preparation of Spiked Meat Samples. Two types of spiked meat samples, cooked (100 °C for 30 min) and autoclaved (133 °C for 20 min) bovine brain in beef, were prepared using a previously developed extraction method22 with modifications. Briefly, 1 g of mashed bovine brain was mixed with 9 g of ground beef to make 10% (g/g) adulterated samples. After heat treatment, 50 mL of the extraction buffer (20 mM Tris-HCl containing 600 mM MgCl2, pH 7.4) was added to each of the heated 10% spiked beef samples. These samples were then homogenized (1 min at 13,000 rpm) using the ULTRA-TURRAX T25 basic homogenizer (IKA Works, Inc., Wilmington, NC) and rotated end-over-end at room temperature for 1 h on the rocker (Labnet International, Inc., Woodbridge, NJ) at 45 rpm, followed by centrifuging at 16,000g for 20 min at 4 °C to separate the pellet and supernatant, after which the supernatant was recentrifuged at the same speed and temperature for 20 min. Nonspiked cooked and autoclaved beef extracts were prepared in the same manner as the spiked samples. The samples spiked with lower levels of CNT contaminant were prepared by diluting the 10% adulterated sample extracts with nonspiked beef extract at levels of 0.0008, 0.0031, 0.0063, 0.0125, 0.025, 0.05, 0.1, 0.2, 0.4, 0.8, and 1.6%. Preparation of Spiked Feed Samples. Before preparing the spiked feed samples, nonspiked beef meat meal and bovine brain meal were produced as follows. Ground beef and mashed bovine brain were placed separately in beakers covered loosely with aluminum foil. Both were autoclaved at 133 °C for 30 min, after which the slurries were dried on sheets of aluminum foil in a convection oven at 79.4 °C for 24 h. The dry weights of the beef meat and bovine brain samples were 26% and 20% of the original materials, respectively. These dried samples were individually ground into fine granules with a Toastmaster coffee grinder (Salton, Inc., Miramar, FL) and then packed in sealed plastic bags and stored at −80 °C until use. Two types of spiked feed samples, bovine brain meal in either beef meal or soybean meal, were subsequently prepared using the same extraction method22 with modifications. Briefly, 1 g of bovine brain was mixed with 9 g of beef meal or soybean meal to make 10% (g/g) spiked samples of each. Both spiked samples were extracted by adding 20 mL of the extraction buffer and mixing well. After heating at 100 °C for 30 min, an additional 30 mL of the extraction buffer was added to each of the sample mixtures. Extracts of nonspiked samples (beef meal and soybean meal) were prepared in the same manner as the spiked feed samples. All samples were then homogenized and centrifuged as for the meat samples described previously. The feed samples spiked with lower levels of CNT contaminant were prepared by diluting the 10% spiked sample extracts with nonspiked beef meal extract or soybean meal extract at the same levels as the equivalent spiked meat samples. All meat and feed sample extracts were stored at −80 °C until use. Indirect Competitive ELISA (icELISA). First, the assay was finetuned to select the optimum concentrations of the immobilized antigen (bovine 18.5-kDa MBP) and mAb3E3 using inELISA with two-dimensional titration (antigen and antibody). The detailed procedure used for the inELISA is as described in a previous study.22 The optimized icELISA was then used to detect the presence of spiked CNT (bovine brain) in the processed meat and feed samples. The optimized icELISA procedure is as follows. One hundred nanograms of purified bovine MBP diluted in 100 μL of 50 mM carbonate−bicarbonate buffer (pH 9.6) was coated onto each well of a 96-well polystyrene high bind microplate (Corning Inc., Corning, NY)

improve the detectability of animal CNT in processed meat and feedstuffs.

■

MATERIALS AND METHODS

Materials. Bovine and porcine CNT (brain and spinal cord) were obtained from the College of Veterinary Medicine, Kansas State University (Manhattan, KS) and the Meat Processing Center at the University of Florida (Gainesville, FL), respectively. Horse brain and spinal cord were obtained from ELISA Technologies, Inc. (Gainesville, FL). Deer and elk brains were purchased from Grande Premium Meats (Del Norte, CO). Bovine dorsal root ganglion and brains from different species (canine, chicken, rabbit, and rat) were purchased from Pel-Freez, LLC. (Rogers, AR). Goat brain was obtained from a local farm in Quincy, FL. Soybean meal was obtained from Say Best/Grain States Soya, Inc. (West Point, NE). Beef and pork loin were purchased from local supermarkets in Tallahassee, FL. All tissues were stored at −80 °C until use. All general chemicals and reagents were of analytical grade. All solutions were prepared using distilled−deionized water from a NANOpure Diamond ultrapure water system (Barnstead International, Dubuque, IA). Animal Immunization. The immunization and antibody production procedures were conducted according to the approved animal protocol in compliance with the University’s Animal Welfare guidelines. Bovine 18.5-kDa MBP, was purified according to a previous study.21 The purified bovine MBP dissolved in 10 mM phosphate buffered saline (PBS, pH 7.2) with 0.3% of sodium dodecyl sulfate (SDS, g/mL) was heated in boiling water for 30 min and used as the immunogen. Three female BALB/c mice (6−8 weeks old) were immunized either subcutaneously or intraperitoneally with 100 μg of the immunogen mixed 1:1 (mL/mL) with Freund’s complete adjuvant, followed by two booster injections at 4-week intervals with 100 μg/mouse of bovine MBP mixed 1:1 (mL/mL) with Freund’s incomplete adjuvant. Test sera were collected by tail bleeding 10 days after each injection; the titers of the sera were determined by indirect noncompetitive ELISA (inELISA), as described in a previous study.22 The mouse exhibiting the highest serum titer to the immunogen received a final boost of 100 μg of the immunogen in PBS before the fusion. Antibody Production. Using the hybridoma technique,23 spleen cells from the immunized mouse were fused with the myeloma cell line (P3 × 63.Ag8.653., ATCC CRL 1580) at a ratio of 5:1 in the presence of polyethylene glycol. The cells were diluted to an appropriate density and cultured in hypoxanthine−aminopterin−thymidine (HAT) media (Sigma-Aldrich, St. Louis, MO). The supernatants were used for the initial screening against the immunogen using inELISA. For the secondary selection, positive cells from the initial screening were expanded to test for cross-reactivity with CNT extracts from other species. The cell lines selected were cloned twice by a limiting dilution method. Finally, the mAb produced, mAb3E3, was selected based on its desired affinity and selectivity, and purified according to the method of Ey et al.24 The isotype of mAb3E3 was determined using the mouse mAb isotyping reagents (Sigma-Aldrich) following the manufacturer’s instructions. Western Blot (WB). WB using mAb3E3 was performed to reveal its antigenic protein in the CNS protein extracts. First, protein extracts of 13 heated (100 °C for 30 min) CNT samples from 10 species (brain from bovine, canine, chicken, deer, elk, goat, horse, porcine, rabbit, and rat; and spinal cord from bovine, horse, and porcine) were prepared according to the method described previously22 and then separated by SDS−PAGE according to the method of Laemmli.25 Second, the separated protein bands in the gel were transferred to a nitrocellulose membrane (Bio-Rad) according to the method of Towbin et al.26 After transferring, the membrane was stained with a staining solution (0.1% Ponceau S [g/mL] and 5.0% acetic acid [mL/ mL]), and the image was scanned with an Epson Perfection 1660 PHOTO scanner (Seiko Epson Corp., Nagano, Japan). After washing with PBS, the membrane was incubated in the blocking buffer (PBS containing 3% bovine serum albumin (BSA) [g/mL], pH 7.2) for 2 h at room temperature, then overnight at 4 °C, and 2 h at room B

DOI: 10.1021/acs.jafc.6b00572 J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Article

Journal of Agricultural and Food Chemistry

Figure 1. Immunoblotting analysis of heated (100 °C for 30 min) CNT samples using mAb3E3. (A) Ponceau S staining of the membrane after protein transfer. (B) Chemiluminescent detection. The amount of proteins loaded to each lane on the 12% gel was 1.6 μg. The 18.5-kDa MBP is indicated by the arrow. Precision Plus Protein Western C standards were used to indicate the molecular weights. Cab, canine brain; Chb, chicken brain; Hb, horse brain; Hs, horse spinal cord; Pb, porcine brain; Ps, porcine spinal cord; Rbb, rabbit brain; Rtb, rat brain; Bb, bovine brain; Bs, bovine spinal cord; Db, deer brain; Eb, elk brain; Gb, goat brain, BMBP, purified bovine 18.5-kDa MBP. and incubated overnight at 4 °C and then at 37 °C for 1 h. The microplate was washed three times with 250 μL/well of the washing buffer (PBS containing 0.05% Tween 20 [mL/mL], pH 7.2) and then blocked by adding 250 μL/well of blocking buffer (PBS containing 1% BSA [g/mL], pH 7.2). The microplate was incubated for 1 h at 37 °C and washed twice with the washing buffer. Meanwhile, the purified antibody diluted in antibody buffer (59 ng/mL) was 1:1 (mL/mL) mixed with different spiked samples in separate 1.5 mL centrifuge tubes. After shaking for 30 min at room temperature, these mixtures (100 μL/well) were each loaded onto the microplate and incubated at 37 °C for 2 h. After another three washes, 100 μL/well of anti-IgGHRP conjugate (Sigma-Aldrich) diluted 1:3,000 (mL/mL) in antibody buffer was added to the plate, which was then incubated at 37 °C for a further 1 h. Following four washing steps, the color was developed by adding 100 μL/well of 0.4 mM ABTS (2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) diammonium salt) substrate solution. The reaction was stopped by adding 100 μL/well of 1% (g/mL) SDS solution after incubating for 30 min at room temperature followed by 30 min at 37 °C. The absorbance was read using a PowerWave XS microplate reader (BioTek Instruments, Inc. Winooski, VT) at 415 nm. Assay Validation. The optimized icELISA was validated according to the U.S. Food and Drug Administration (FDA)’s guidance for industry bioanalytical method validation.27 First, the assay precision was examined. Intra-assay and interassay CVs of the assay were computed from four sets of spiked heated samples. The intra-assay variability was measured by analyzing at least three replicates of each sample in the assay within the same plate. The interassay variability was calculated from the analysis of at least three replicates of each sample carried out on five different days. Second, the assay sensitivity, i.e., the limit of detection (LOD), was determined. The LOD of the assay is the lowest spiking level that produces a significant difference between the baseline (non-CNT) reading and the reading for a spiked sample (P < 0.05). Third, the recovery rate of the assay was studied. Five spiked samples, consisting of 10% bovine brain in beef and 0.1

and 1.6% bovine brain meal in beef meal or soybean meal (g/g), were analyzed. Statistical Analysis. Each tested sample was at least duplicated, and each experiment was measured at least twice. One-way ANOVA with Dunnett’s post-test was performed; P < 0.05 was considered to be statistically significant.

■

RESULTS AND DISCUSSION Antibody Characterization. The antibody, mAb3E3 (isotype: IgG1), was selected from the hybridomas because of its strong immunoreactivity with bovine 18.5-kDa MBP. After purification, the IgG concentration of mAb3E3 was 59 μg/mL at 280 nm. Next, proteins in heated (100 °C for 30 min) CNT extracts were examined with WB using mAb3E3 (Figure 1). The Ponceau S stained membrane (Figure 1A) confirms that the proteins from each CNT sample were successfully transferred onto the membrane similar to the protein profile in the stained gel described in a previous study.22 Using WB (Figure 1B), the mAb3E3 antibody detected a set of antigenic proteins in the CNT extracts from different animal species; mAb3E3 successfully detected the MBP from eight animal species (bovine, chicken, deer, elk, goat, horse, porcine, and rat). However, the 18.5-kDa MBP from porcine and ruminant (bovine, deer, elk, and goat) species exhibited higher color intensity than those for other species (chicken, horse, and rat). Additionally, the color intensity of MBP from two spinal cord samples (bovine and horse) was higher than that of their corresponding brain extracts. Besides the 18.5-kDa MBP, two other proteins also reacted with mAb3E3 (Figure 1B). One is a 21.5-kDa protein, which was observed with low color intensity in the samples from 7 animal species (bovine, deer, elk, goat, porcine, rabbit, and rat). The other protein of ∼18 kDa in size reacted only in samples from canine and rabbit C

DOI: 10.1021/acs.jafc.6b00572 J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Article

Journal of Agricultural and Food Chemistry

Figure 2. Two-dimensional titration using indirect noncompetitive ELISA. A five-parameter logistical model was used to establish the curves. The selected concentration of mAb3E3 for the indirect competitive ELISA is indicated by a dotted line (59 ng/mL). The value r2 quantifies the goodness of fit. Results are expressed as the mean of A415 ± SD.

quantity of immobilized antigen, and this coated antigen and the antigen that is present in the unknown sample compete for the limited number of binding sites on the antibody. If the unknown sample contains the target antigen that is capable of binding to the antibody, this prevents this antibody from binding to the immobilized antigen on the plate. The maximum expected optical density (OD) is measured in the assay with no competitor in the unknown samples. The OD is reduced in the presence of the antigen competitor in the unknown sample, and the degree of the inhibition of the reaction is proportional to the relative amount of the antigen competitor present. During assay development, both the antigen (bovine 18.5kDa MBP) coating concentration and the antibody (mAb3E3) concentration for icELISA were optimized using an inELISA. In this study, two antigen coating concentrations, 50 and 100 ng/ well of bovine MBP, were compared. The IgG concentration of mAb3E3 was a 2-fold dilution from 0.4 to 1180 ng/mL. From the inELISA results (Figure 2), the absorbance of the 100-ng coating (solid line) was higher than that of the 50-ng coating (semidotted line) for the same antibody concentration. The absorbance of 100-ng MBP reacted with 59 ng/mL mAb3E3 was about 1.2, six times higher than that of 50-ng MBP (dotted line). In a competitive ELISA, the concentration of the antianalyte antibody should be limited (unsaturated) in order to achieve desirable assay sensitivity.33 From the curve of the 100-ng MBP (solid line, Figure 2), 59 ng/mL of mAb3E3 was significantly below the saturated concentration (>590 ng/mL) of the antibody. Although a higher but still unsaturated 3E3 antibody concentration could be chosen to obtain a higher absorbance value, this increase in antibody concentration could substantially increase the cost of the diagnostic test. Therefore, taking into account these various factors, 100 ng/well bovine MBP and 59 ng/mL mAb 3E3, respectively, were selected as the concentrations of antigen coating and antibody for the novel icELISA. Using a competitive ELISA, the amount of antigen present in an unknown sample can be quantitatively identified with the aid of a standard inhibition curve generated from a set of standard

brain. There was no 18.5 kDa band in canine and rabbits, and the color intensity of the 18-kDa protein for rabbit was much higher than that for the canine samples and was similar to that for the 18.5-kDa MBP from ruminant species. This cross-reaction of mAb3E3 seems to be due to two major factors: the concentration of target analyte (MBP) in the samples and the similarity of the antigenic epitope in different animal species or tissues. Both factors affect the color intensity shown on the WB membrane. On the basis of the analysis of the amino acid sequence of MBP from different species, bovine 18.5-kDa MBP (primary accession number: P02687) shares about 89−93% identity with the MBP from rat (P02688−2), rabbit (P25274), horse (P83487), and pig (P81558).28 Although MBP is the major protein of CNS myelin (∼30%), in the peripheral nervous system (PNS) myelin MBP varies from about 5 to 18% of the total PNS protein.29 In addition, the MBP family of proteins has many isoforms with different molecular weights, with different animal species having different isoforms.29 For example, it has been reported that both canine and rabbit CNT contain a major MBP isoform with a smaller molecular weight (canine: 18 kDa; rabbit: 18.2 kDa).30,31 It should also be noted that the 21.5-kDa MBP isoform is one of the major MBP isoforms in calves.32 On the basis of the results of the WB (Figure 1B), both the 21.5-kDa and 18-kDa antigenic proteins detected by the mAb 3E3 are likely to be MBP isoforms that contain the same epitope as the 18.5-kDa MBP. In general, the 3E3 antibody reacts strongly with the 18.5kDa MBP from porcine and ruminant species and the 18-kDa MBP from rabbit. However, in practice, except for porcine CNT, the contamination of food and feed products with CNT from rabbit is extremely rare. Assay Development. In general, a competitive ELISA measures the inhibition of the immobilized (coated on the microplate) antigen in an optimized system using an antibody specific for the immobilized antigen, where the degree of inhibition reflects the quantity of the antigen in an unknown sample added to the system. This assay uses a predetermined D

DOI: 10.1021/acs.jafc.6b00572 J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Article

Journal of Agricultural and Food Chemistry

Figure 3. Mean standard curves of the indirect competitive ELISA using mAb3E3 for (A) spiked meat (solid line) and feed (dashed line) samples and (B) bovine 18.5-kDa MBP (n = 4). Spiked meat samples: cooked (100 °C for 30 min) and autoclaved (133 °C for 20 min) bovine brain in beef (n = 11); spiked feed samples, heated (100 °C for 30 min) bovine brain meal in beef meal and soybean meal (n = 14). A five-parameter logistical model was used to establish the standard curves. The corresponding values of IC50 are indicated by a dotted line. The value r2 quantifies the goodness of fit. Data are reported as the mean of the % maximum antibody binding ± SEM.

thus extremely important to choose the correct standard curve model for that assay.35 For most immunoassays, the four- or five-parameter logistic model is more appropriate than the linear, quadratic, or log−log linear models.36 The five-

samples containing known amounts of the target antigen. Generally, a significant source of variability in the standard curves comes from the choice of the statistical model used for the standard curve.34 To improve the accuracy of an assay, it is E

DOI: 10.1021/acs.jafc.6b00572 J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Article

Journal of Agricultural and Food Chemistry

3B). The IC50 value of MBP was determined to be 11.8 ppm. On the basis of this curve, the concentration of MBP in an unknown sample extract can be determined. If two different standard curves (one using meat or feed standards and the other using MBP standards) are established in the same experiment, this approach can indicate not only the adulteration level of bovine CNT but also the amount of bovine MBP in the unknown samples tested. Assay Validation. Once the assay had been developed, the precision, sensitivity, and recovery of the optimized competitive ELISA were validated according to the FDA guidance for industry bioanalytical method validation.27 Precision is the closeness of agreement (degree of scatter) between a series of measurements obtained from multiple sampling of the same homogeneous sample under the prescribed conditions.27 Precision can also be regarded as reproducibility, offering a statistical measure of the variation between replicate determinations in the same assay (intra-assay variability) and in different assays (interassay variability), and is generally represented by the coefficient of variation (CV).39 The CV is defined as the ratio of the standard deviation to the mean. The intra-assay and interassay CVs of the icELISA developed in this study were computed from four sets of differently spiked samples using the equations described in a previous study.40 The results are listed in Table 1. Compared with the maximum FDA acceptable CV (15%),27 the assay exhibited low intraassay variability (≤3.5%) and interassay variability (≤3.3%, Table 1).

parameter logistic (5PL) equation was described in a previous study22 and is equivalent to the four-parameter logistic (4PL) equation with an additional parameter added for asymmetry. Asymmetry denotes the degree of asymmetry in the shape of the sigmoidal curve with respect to IC50 (the half maximal inhibitory concentration). A value of one indicates perfect symmetry, which would then correspond to the 4PL model. This additional parameter provides a better fit when the response curve is not symmetrical.37 The best model for the standard curve can be chosen based on the validation of standard recovery.38 Standard recovery was performed by calculating the concentration of each standard and then comparing it to the actual concentration.22 The closer the recovery is to 100%, the less difference is between the calculated concentration and the true concentration. In this study, the standard recovery for two statistical models, fourand five-parameter logistics, were compared (Table S1 in Supporting Information). Overall, in each set of spiked samples, the quality of the curve fit from the 5PL model was better than that from the 4PL model, indicating that the 5PL model better reflects the real situation in the immunoreaction of the optimized assay. Therefore, the 5PL model was selected to establish the standard curve for this icELISA. Using the 5PL model, a set of mean standard curves for 4 types of spiked samples was derived from different relevant curves measured on different days. In meat samples, two different sample treatments were performed: cooked (100 °C for 30 min) and autoclaved (133 °C for 20 min). There was no significant difference between the results for the same spiking amounts for these two sets of meat samples (P > 0.05, Figure S1A in Supporting Information), indicating no treatment effect on the assay being observed. In the feed samples, the sample matrix is different, with one consisting of beef meal and the other of soybean meal. Similar to the spiked meat samples, the assay exhibited no significant differences in the results obtained for the two sets of feed samples for the same spiking amounts (P > 0.05, Figure S1B in Supporting Information). This means that the matrix effect on the assay was also not significant. These results made it possible to establish a typical meat standard curve using the 5PL model for both, the cooked and autoclaved adulterated meat samples (solid line, Figure 3A); in addition, it allowed us to plot a typical feed standard curve derived from both, beef meal and soybean meal samples, in the same manner (dashed line, Figure 3A). Interestingly, the data in Figure 3A show that although the IC50 value for meat samples (4.8%) was about 7 times higher than that for feed samples (0.7%), the slopes of both the standard curves were the same (−0.6). The difference in the IC50 values between meat and feed samples indicates that the concentration of target analyte (MBP) is much higher in feed than in meat samples, thus making this assay more sensitive to feed than to meat samples. This is expected because of the difference in the nature of these samples, with the feed samples being prepared on a dry weight basis and the meat samples on a wet weight basis. Given that the dry weight of bovine brain was about 20% (g/g) of the wet material, under the same extraction conditions, the concentration of target analyte (MBP) will be much lower in meat than in feed samples. Using either of the standard curves (meat or feed) of the assay (Figure 3A) enables the spiking level of CNT in meat or feed to be calculated accordingly. To determine the amount of MBP in the spiked samples, a typical standard curve for bovine MBP was therefore established using the 5PL model (Figure

Table 1. Coefficient of Variation (CV) of the Indirect Competitive ELISA Using mAb3E3 intra-assay coefficient of variation (CV, %) feed

meat

0.025% bovine brain meal in beef meal 0.025% bovine brain meal in soybean meal 0.05% bovine brain in beef (cooked) 0.05% bovine brain in beef (autoclaved)

day 1 day 2

day 3 day 4

day 5 interassay

0.4

0.7

1.9

1.0

0.9

1.7

1.3

0.1

0.1

1.2

0.7

1.7

2.7

1.2

1.6

1.9

3.5

3.3

0.1

2.8

0.8

0.6

1.0

1.8

The performance of an assay can also be evaluated by profiling the precision measured for different sample (analyte) concentrations.41 These precision profiles are obtained by plotting the values of CV against the concentration of a measured analyte,39 providing useful information about the reproducibility for different sample concentrations (Figure S2 in Supporting Information). The precision profiles for three standard curves (meat, feed, and bovine MBP) were estimated by a second-order polynomial curve fit (curves not shown) and a CV cutoff of 10% defined as the maximum acceptable precision.39 The assay precision range for the results reported here is 0−1.7% (g/g) for the meat samples, 0−0.34% (g/g) for the feed samples, and 0−30.8 ppm for the bovine 18.5-kDa MBP. Sensitivity, namely, the limit of detection (LOD), is an assay’s capacity to measure the smallest amount of target analyte under the standard conditions defined.39 In this study, F

DOI: 10.1021/acs.jafc.6b00572 J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Article

Journal of Agricultural and Food Chemistry

samples. This icELISA is suitable for the quantitative detection of low levels of undeclared animal CNT, especially bovine CNT, in processed food and feed products. Given that CNTbased SRM has been banned in both food and feed worldwide, this new assay has the potential to become a useful analytical tool not only for improving the safety of animal feed but also for enforcing food labeling laws.

the LOD was defined as the smallest quantity of the analyte that could be significantly differentiated from the background (0 ppm of bovine 18.5-kDa MBP or 0% adulteration level) in the assay using one-way ANOVA (Dunnett’s post-test, P < 0.05). The LODs of the icELISA were 0.05% for cooked and autoclaved bovine brain in beef, 0.0125% for heated bovine brain meal in feed (beef meal or soybean meal), and 6.4 ppb for bovine MBP. Comparing these with the LODs of other immunoassays (from 1% to 10%) for the detection of CNT in meat using MBP as the marker,42−45 this icELISA has better LODs not only for meat but also for feed samples. Additionally, the assay working range, which consists of the difference between the sensitivity and the uppermost precision range of the standard curve, is 0.05−1.7% (g/g) for bovine brain in beef, 0.0125−0.34% (g/g) for bovine brain meal in feed, and 0.0064−30.8 ppm for bovine MBP. Recovery is the extraction efficiency of an analytical process, reported as a percentage of the known amount of an analyte carried through the sample extraction and processing steps of the method.27 Recovery experiments were carried out on blank samples (beef, beef meal, and soybean meal) that were artificially contaminated with bovine brain proteins. Five adulterated samples (bovine brain in beef [10%, g/g] and bovine brain meal in beef meal or soybean meal [0.1 and 1.6%, g/g]), along with their relevant standards, were simultaneously analyzed with the optimized assay. The recovery was calculated after obtaining the adulteration level of these spiked samples extrapolated from their respective standard curves. From Table 2, the mean recovery of feed samples containing bovine brain

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jafc.6b00572. Comparison of the standard recovery using two standard curve models for the indirect competitive ELISA; mean standard curves of the indirect competitive ELISA using Ab3E3 for four spiked meat and feed samples; and precision profiles of CV values vs sample concentrations estimated by a second-order polynomial curve fit (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*Tel: +1 850-644-1744. Fax: +1 850-645-5000. E-mail: [email protected]. Funding

This research was funded by the USDA-NRI program of Animal Protection (project no. 2006-35204-17407) and NIAID-NIH PO1 (AI 77774-01). Notes

The authors declare no competing financial interest.

Table 2. Recovery of Bovine Brain Protein from Meat and Feed Samples That Were Artificially Contaminated with Bovine Brain Tissue spiked sample feed

meat a

bovine brain meal in beef meal (n = 5) bovine brain meal in soybean meal (n = 4) bovine brain in beef (autoclaved, n = 2)

true value (%, g/g)

calculated value (%, g/g)a

recovery (%)a

± ± ± ±

129.8 ± 5.2 88.3 ± 5.5 145.0 ± 12.0 103.3 ± 2.0

0.1 1.6 0.1 1.6

0.13 1.41 0.15 1.65

0.01 0.09 0.01 0.03

10

9.4 ± 0.1

■

REFERENCES

(1) U.S. NIH. What Are Transmissible Spongiform Encephalopathies? http://www.ninds.nih.gov/disorders/tse/tse.htm#What_is. (accessed Oct 19, 2015). (2) Bruce, M. E.; Will, R. G.; Ironside, J. W.; McConnell, I.; Drummond, D.; Suttie, A.; McCardle, L.; Chree, A.; Hope, J.; Birkett, C.; Cousens, S.; Fraser, H.; Bostock, C. J. Transmissions to mice indicate that ’new variant’ CJD is caused by the BSE agent. Nature 1997, 389, 498−501. (3) Prusiner, S. B. Prion diseases and the BSE crisis. Science 1997, 278, 245−251. (4) Safar, J.; Roller, P. P.; Gajdusek, D. C.; Gibbs, C. J. Thermalstability and conformational transitions of scrapie amyloid (prion) protein correlate with infectivity. Protein Sci. 1993, 2, 2206−2216. (5) European Commission’s Scientific Steering Committee. Opinion of the Scientific Steering Committee on the Human Exposure Risk (HER) via Food with Respect to BSE. http://ec.europa.eu/food/fs/ sc/ssc/out67_en.pdf (accessed Oct 19, 2015). (6) Wilesmith, J. W.; Ryan, J. B. M.; Atkinson, M. J. Bovine spongiform encephalopathy - epidemiologic studies on the origin. Vet. Rec. 1991, 128, 199−203. (7) European Commission. Regulation (EC) No 999/2001 of the European Parliament and of the Council of 22 May 2001 laying down rules for the prevention, control and eradication of certain transmissible spongiform encephalopathies. Off. J. Eur. Commun. 2001, L147, 1−40. (8) European Commission. Commission Decision of 29 June 2000 regulating the use of material presenting risks as regards transmissible spongiform encephalopathies and amending Decision 94/474/EC. Off. J. Eur. Commun. 2000, L158, 76−82. (9) U.S. FDA. 21 CFR 589.2000 - Animal Proteins Prohibited in Ruminant Feed. http://www.gpo.gov/fdsys/granule/CFR-2012title21-vol6/CFR-2012-title21-vol6-sec589-2000 (accessed Oct 19, 2015).

93.9 ± 0.7

Results are expressed as the mean ± SEM.

meal was determined to be 116.6%, which indicates that the results calculated for both the 0.1% adulterated feed samples (bovine brain meal in either beef meal or soybean meal) overestimated the amount of bovine brain proteins present in the sample. This is likely because the adulteration level was so small (0.1%) that a deviation in the calculation of as little as 0.13% (calculated for 0.1% bovine brain meal in beef meal, Table 2) could lead to an unrealistically high recovery rate (∼130%). For the 10% adulterated meat sample, the recovery rate was 93.9%. In summary, the icELISA reported here for the detection of CNT is based on the detection of MBP, a thermo-stable marker protein that can be recognized in severely heated CNT samples. This assay is tissue-selective, distinguishing between CNT and other tissues such as muscle, and does not react with soy protein commonly used in animal feed products. The assay exhibits good performance characteristics, including low intraassay and interassay CVs, and low LODs in meat and feed G

DOI: 10.1021/acs.jafc.6b00572 J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Article

Journal of Agricultural and Food Chemistry (10) U.S. FDA. 21 CFR 589.2001 - Cattle Materials Prohibited in Animal Food or Feed to Prevent the Transmission of Bovine Spongiform Encephalopathy. http://www.gpo.gov/fdsys/granule/ CFR-2012-title21-vol6/CFR-2012-title21-vol6-sec589-2001 (accessed Oct 19, 2015). (11) USDA. 9 CFR Parts 309, 310 and 318 Prohibition of the use of specified risk materials for human food and requirements for the disposition of non-ambulatory disabled cattle; prohibition of the use of certain stunning devices used to immobilize cattle during slaughter; rule. Fed. Regist. 2007, 72, 38699−38730. (12) U.S. FDA. Animal and Veterinary Recalls Archive. http://www. fda.gov/AnimalVeterinary/SafetyHealth/RecallsWithdrawals/ ucm393160.htm (accessed Oct 19, 2015). (13) Scott, M. R.; Will, R.; Ironside, J.; Nguyen, H. O. B.; Tremblay, P.; DeArmond, S. J.; Prusiner, S. B. Compelling transgenetic evidence for transmission of bovine spongiform encephalopathy prions to humans. Proc. Natl. Acad. Sci. U. S. A. 1999, 96, 15137−15142. (14) Anil, M. H.; Love, S.; Helps, C. R.; Harbour, D. A. Potential for carcass contamination with brain tissue following stunning and slaughter in cattle and sheep. Food Control 2002, 13, 431−436. (15) USDA. Recall Case Archive. http://www.fsis.usda.gov/wps/ portal/fsis/topics/recalls-and-public-health-alerts/recall-case-archive (accessed Oct 19, 2015). (16) Lucker, E.; Biedermann, W.; Lachhab, S.; Truyen, U.; Hensel, A. GC-MS detection of central nervous tissues as TSE risk material in meat products: analytical quality and strategy. Anal. Bioanal. Chem. 2004, 380, 866−870. (17) Gangidi, R. R.; Proctor, A.; Pohlman, F. W.; Meullenet, J. F. Rapid determination of spinal cord content in ground beef by nearinfrared spectroscopy. J. Food Sci. 2005, 70, C397−C400. (18) Nowak, B.; von Mueffling, T.; Kuefen, A.; Ganseforth, K.; Seyboldt, C. Detection of bovine central nervous system tissue in liver sausages using a reverse transcriptase PCR technique and a commercial enzyme-linked immunosorbent assay. J. Food Prot. 2005, 68, 2178−2183. (19) Agazzi, M. E.; Moreno, J. B.; von Holst, C.; Lucker, E.; Anklam, E. Quantitative analysis of tissues of the central nervous system in food products by GFAP-ELISA test kit. Results of an interlaboratory study. Food Control 2004, 15, 297−301. (20) European Commission. 97/199/EC: Commission Decision of 25 March 1997 laying down the animal health requirements and the veterinary certification for the import of petfood in hermetically sealed containers from certain third countries which use alternative heat treatment systems and amending Decision 94/309/EC. Off. J. Eur. Commun. 1997, L84, 44−48. (21) Rao, Q. C.; Hsieh, Y.-H. P. Effect of pH, temperature and storage time on the stability of bovine myelin basic protein. Food Control 2015, 50, 166−172. (22) Rao, Q. C.; Hsieh, Y.-H. P. Enhanced immunodetection of bovine central nervous tissue using an improved extraction method. Food Control 2014, 46, 282−290. (23) Kohler, G.; Milstein, C. Continuous cultures of fused cells secreting antibody of predefined specificity. Nature 1975, 256, 495− 497. (24) Ey, P. L.; Prowse, S. J.; Jenkin, C. R. Isolation of pure Igg1, Igg2a and Igg2b immunoglobulins from mouse serum using protein asepharose. Immunochemistry 1978, 15, 429−436. (25) Laemmli, U. K. Cleavage of structural proteins during assembly of head of bacteriophage-T4. Nature 1970, 227, 680−685. (26) Towbin, H.; Staehelin, T.; Gordon, J. Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets - procedure and some applications. Proc. Natl. Acad. Sci. U. S. A. 1979, 76, 4350− 4354. (27) U.S. FDA. Guidance for Industry Bioanalytical Method Validation. http://www.fda.gov/downloads/drugs/ guidancecomplianceregulatoryinformation/guidances/ucm368107.pdf (accessed Oct 19, 2015).

(28) Gasteiger, E.; Gattiker, A.; Hoogland, C.; Ivanyi, I.; Appel, R. D.; Bairoch, A. ExPASy: the proteomics server for in-depth protein knowledge and analysis. Nucleic Acids Res. 2003, 31, 3784−3788. (29) Quarles, R. H.; Macklin, W. B.; Morell, P. Myelin Formation, Structure and Biochemistry. In Basic Neurochemistry: Molecular, Cellular and Medical Aspects, 7th ed.; Siegel, G. J., Albers, R. W., Brady, S., Price, D., Eds.; Lippincott Williams & Wilkins: Philadelphia, PA, 2006; pp 51−72. (30) Oji, T.; Kamishina, H.; Cheeseman, J. A.; Clemmons, R. M. Measurement of myelin basic protein in the cerebrospinal fluid of dogs with degenerative myelopathy. Vet. Clin. Pathol. 2007, 36, 281−284. (31) Martenson, R. E.; Law, M. J.; Deibler, G. E. Identification of multiple in vivo phosphorylation sites in rabbit myelin basic-protein. J. Biol. Chem. 1983, 258, 930−937. (32) De Derosbo, N. K.; Carnegie, P. R.; Bernard, C. C. A.; Linthicum, D. S. Detection of various forms of brain myelin basic protein in vertebrates by electroimmunoblotting. Neurochem. Res. 1984, 9, 1359−1369. (33) Diamandis, E. P.; Christopoulos, T. K. Immunoassay Configurations. In Immunoassay; Diamandis, E. P., Christopoulos, T. K., Eds. Academic Press: San Diego, CA, 1996; pp 227−236. (34) Herman, R. A.; Scherer, P. N.; Shan, G. Evaluation of logistic and polynomial models for fitting sandwich-ELISA calibration curves. J. Immunol. Methods 2008, 339, 245−258. (35) Sittampalam, G. S.; Gal-Edd, N.; Arkin, M.; Auld, D.; Austin, C.; Bejcek, B.; Glicksman, M.; Inglese, J.; Lemmon, V.; Li, Z.; McGee, J.; McManus, O.; Minor, L.; Napper, A.; Riss, T.; Trask, O. J., Jr.; Weidner, J. Assay Guidance Manual. http://www.ncbi.nlm.nih.gov/ books/NBK53196/ (accessed Oct 19, 2015). (36) Findlay, J. W. A.; Dillard, R. F. Appropriate calibration curve fitting in ligand binding assays. AAPS J. 2007, 9, E260−E267. (37) Gottschalk, P. G.; Dunn, J. R. The five-parameter logistic: A characterization and comparison with the four-parameter logistic. Anal. Biochem. 2005, 343, 54−65. (38) Nix, B.; Wild, D. Calibration Curve-Fitting. In The Immunoassay Handbook, 2nd ed.; Nature Pub. Group: London, 2001; pp 198−210. (39) Crowther, J. R. Validation of Diagnostic Tests for Infectious Diseases. In The ELISA Guidebook; Crowther, J. R., Ed.; Humana Press: Totowa, NJ, 2000; Vol. 149, pp 301−345. (40) Rao, Q. C.; Hsieh, Y.-H. P. Competitive enzyme-linked immunosorbent assay for quantitative detection of bovine blood in heat-processed meat and feed. J. Food Prot. 2008, 71, 1000−1006. (41) Holzhauser, T.; Vieths, S. Indirect competitive ELISA for determination of traces of peanut (Arachis hypogaea L.) protein in complex food matrices. J. Agric. Food Chem. 1999, 47, 603−611. (42) Tersteeg, M. H. G.; Koolmees, P. A.; van Knapen, F. Immunohistochemical detection of brain tissue in heated meat products. Meat Sci. 2002, 61, 67−72. (43) Sultan, K. R.; Tersteeg, M. H. G.; Koolmees, P. A.; de Baaij, J. A.; Bergwerff, A. A.; Haagsman, H. P. Western blot detection of brain material in heated meat products using myelin basic protein and neuron-specific enolase as biomarkers. Anal. Chim. Acta 2004, 520, 183−192. (44) Herde, K.; Bergmann, M.; Lang, C.; Leiser, R.; Wenisch, S. Glial fibrillary acidic protein and myelin basic protein as markers for the immunochemical detection of bovine central nervous tissue in heattreated meat products. J. Food Prot. 2005, 68, 823−827. (45) Rencova, E. Comparison of commercially available antibodies for the detection of central nervous system tissue in meat products by enzyme-linked immunosorbent assay. J. Food Prot. 2005, 68, 630−632.

H

DOI: 10.1021/acs.jafc.6b00572 J. Agric. Food Chem. XXXX, XXX, XXX−XXX