Article pubs.acs.org/JAFC
Presence of Undeclared Food Allergens in Cumin: The Need for Multiplex Methods Eric A. E. Garber,*,† Christine H. Parker,† Sara M. Handy,† Chung Y. Cho,† Rakhi Panda,† Mansour Samadpour,‡ Danica H. Reynaud,§ and George C. Ziobro∥ †
Office of Regulatory Science, Center for Food Safety and Applied Nutrition, Food and Drug Administration, College Park, Maryland 20740, United States ‡ IEH Laboratories and Consulting Group, Inc., Lake Forest Park, Washington 98155, United States § AuthenTechnologies LLC, Richmond, California 94806, United States ∥ Office of Food Safety, Center for Food Safety and Applied Nutrition, Food and Drug Administration, College Park, Maryland 20740, United States ABSTRACT: Beginning in the autumn of 2014, millions of dollars of food and over 675 products were recalled in the United States due to the presence of undeclared peanut, attributed to cumin used in the manufacture of the products. Initial analyses also indicated the presence of almond. Subsequent research showed that the presence of peanut and almond did not fully explain the analytical results for the cumin samples. Using a combination of mass spectrometry, DNA-based methods (i.e., PCR and Sanger DNA Sequencing), microscopy, and antibody-based technologies (i.e., ELISA, Western blot analysis, and a novel xMAP multiplex assay) the presence of peanut was confirmed. Screening for secondary sources of adulteration (e.g., tree nuts, mahleb, peach, and cherry) supported the assessment that the cumin contained multiple contaminants. These results demonstrate the limitations of single analyte-specific assays and the need for orthogonal multiplex methods to detect food allergens irrespective of varietal or other differences. KEYWORDS: cumin, peanut, food allergen detection
■
INTRODUCTION Food allergies afflict an estimated 3% of the population and 6% of children. Of food allergic individuals, an estimated 0.6% and 0.5% suffer from peanut and tree nut allergies, respectively.1 Furthermore, the prevalence of food allergies is increasing, with surveys of 25 of 30 countries displaying an increase over a tenyear period.2,3 In addition, the number of people with multiple food allergies increased by 50% between 2001 and 2005.4 Lacking a cure or prophylactic treatment, allergic individuals must avoid consuming allergen-containing foods. To safeguard this population, the Food Allergen Labeling and Consumer Protection Act of 2004 (FALCPA) was enacted.5 In the United States, 90% of food allergies are due to FALCPA-mandated labeled food allergens: egg, milk, peanut, soy, wheat, tree nuts, crustacean seafood, and fish. To help enforce FALCPA and detect inadvertent crosscontact of food allergens, antibody-based assays such as ELISAs and related technologies (e.g., lateral flow devices or dipsticks) are often employed. These methods are robust, being able to detect as little as 1 μg of a food allergen, often as little as 10 ng of a specific antigenic protein, in one gram of a complex mixture of food with minimal fractionation or purification. However, antibody-based methods are not perfect. The specificity of antibody-based methods is limited to the uniqueness of the target epitopes, with sandwich configurations demonstrating greater specificity by requiring the presence of two unique epitopes. To address these needs and limitations associated with the use of antibody-based methods, DNA- and mass spectrometry-based methods are increasingly being © 2016 American Chemical Society
employed due to their ability to provide more specific identification and distinguish between homologous proteins and related foods.6 In late 2014, the Canadian Food Inspection Agency (CFIA), using commercial ELISA test kits, reported the detection of undeclared peanut and almond in a taco spice. Subsequent research by the Food Allergy Research and Resource Program (FARRP) established that the cumin used in the manufacture of the taco spice mix generated positive responses when tested for peanut and almond using ELISA technologies.7 In response to the concerns of undeclared peanut, over 675 food products were ultimately recalled, making this one of the largest food safety recalls. As part of the analyses, the FDA collected cumin samples from multiple sources and conducted ELISA and xMAP multiplex analyses for the presence of undeclared food allergens.8 A significant advantage of xMAP analysis and its ability to simultaneously detect 14 allergens (including 9 tree nuts) plus gluten, is the ability to generate antigenic profiles. This is not easily done using antigen-specific ELISAs. Three cumin samples (designated A, B, and H) and a taco spice mix made using cumin H (subsequently referred to as taco spice mix), displayed a complex antigenic profile uncharacteristic of peanut and almond as the only food allergens present.8 This raised questions requiring additional research before it could be Received: Revised: Accepted: Published: 1202
November 17, 2015 December 27, 2015 January 14, 2016 January 14, 2016 DOI: 10.1021/acs.jafc.5b05497 J. Agric. Food Chem. 2016, 64, 1202−1211
Article
Journal of Agricultural and Food Chemistry
Kjeldahl process measuring total nitrogen content. The protein content of the mixed extraction solutions used to prepare the LCtandem mass spectrometry samples was measured using a Qubit Fluorometer (Invitrogen Life Technologies) as recommended by the manufacturer. Specifically, a dye that fluoresces upon binding to proteins and is compatible with the buffers used is mixed with the sample, and the fluorescence is compared to the fluorescence generated by BSA standards. Immunochemical Methods. Veratox for Almond, Egg, Hazelnut, and Peanut Allergen ELISA test kits (Neogen Corp., Lansing, MI), xMAP Multiplex Food Allergen test kits (Radix BioSolutions, Georgetown, TX), and ELISAs for Almond, Hazelnut, and Peanut (IEH Laboratories & Consulting Group, Inc., Lake Forest Park, WA) were used as recommended by the manufacturers with defined amounts of sample extracted and the responses compared to that of reference materials. The xMAP Multiplex Food Allergen assay involves two extraction protocols and generally employs two antibodies for detection and confirmation of the 14 targeted allergens and gluten. In brief, one of the extraction protocols is based on buffered-detergent employing antibodies developed by IEH Laboratories & Consulting Group, Inc. (IEH) and Elution Technologies, Inc. The second extraction protocol uses SDS/β-mercaptoethanol, relying on the antibodies developed by the Morinaga Institute of Biological Sciences, Inc. (MIoBS) and used in the multilaboratory validated ELISA test kits with excellent recoveries detecting baked foods.12 The built-in confirmation of the xMAP Multiplex Food Allergen assay means that to detect an allergen, all of the antibodies specific for that allergen must generate a positive response, greater than the first nonzero standard, with the relative responses by the antibodies comparable to the reference materials; deviations from the expected relative intensities indicate usually that the reference material was not appropriate, that the analyte was not the suspected allergen, or that the antigenicity had be altered. Protein Electrophoresis and Western Blots. SDS/PAGE protein electrophoresis and Western blot analyses were performed as previously described with slight modifications.13 Specifically, the cumin samples were extracted 1:10 with PBST for 2 h at room temperature followed by centrifugation at 10 000g for 15 min. Once diluted in Laemmli Buffer, the samples were loaded onto Novex 4− 12% Bis-Tris gels (Invitrogen, Carlsbad, CA) at 10 μg total protein per lane. Total protein content was visualized using EZBlue Gel Staining Reagent. Western blot analyses entailed transferring the electrophoretic separated proteins to polyvinylidene difluoride (PVDF) membranes, blocking with 3% BSA, and immunoblotting with a 10% solution of the horseradish peroxidase (HRP) conjugated detector antibodies used in the MIoBS Peanut ELISA Kits (Crystal Chem, Inc., Downers Grove, IL). SuperSignal West Dura Extended Duration Substrate (Thermo Fisher Scientific, Inc.) was added to generate a chemiluminescent signal visualized using a G:Box Chemi XX6 imager (Syngene, A Division of Synoptics Ltd., Frederick, MD). Multiple images were collected for variable durations to enhance the dynamic range of the Western blots. DNA Analysis, Targeted PCR. PCR targeting Ara h 1, Ara h 2, Ara h 3 genes of peanut, the Pru du 1.01 gene of almond, Cor a 1 gene of hazelnut, 5s rRNA intergenic spacer gene of cashews (GenBank AY230649.1), and internal transcribed spacer (ITS) 1 of pistachio (GenBank KJ018021.1) were developed and conducted using protocols established by IEH Laboratories & Consulting Group, Inc. (Lake Forest Park, WA). Additional species-specific PCR tests for peanut, almond, Brazil nut, cashew, hazelnut, macadamia, pistachio, walnut, and the genus Prunus (containing almond, cherry, mahleb, and peach) were developed and conducted according to procedures established by AuthenTechnologies, LLC (Richmond, CA) with all positive amplicons confirmed by Sanger sequencing. A PCR assay specific for the α subunit of the chloroplast ATP synthase CF1 of pine nuts was used.14 DNA was extracted from 25−60 mg of sample using the QIAamp DNA Stool Mini Kit (QIAGEN, Inc., Valencia, CA) per manufacturer’s instructions. Positive (Pinus koraiensis NA71577) and negative controls (extraction and PCR) were included and the samples were amplified using the 129k primer set as previously described.14
definitively concluded whether peanut or another food allergen was present. To address the complex antigenic profiles indicated by the xMAP studies, the cumin samples were analyzed using a combination of other approaches that included DNA-methodologies, mass spectrometry, and microanalytical characterization. The DNA methodologies consisted of a combination of gene-specific methods, such as polymerase chain reaction (PCR) and real time (RT)-PCR, and Sanger sequencing using universal plant primers. Global proteomic screening by mass spectrometry was used to characterize protein content and proteolytic digests of the cumin samples relative to spectral libraries generated from target reference materials. Microanalysis examined the physical and optical activity of the cumin. Together these studies were able to definitively establish the presence of peanut and other food allergens despite varietal differences compared to the reference materials typically employed. The importance of utilizing orthogonal multiplex methodologies to detect food allergen contamination in food and the need for reference materials capable of representing the variability of food products imported is expected to become more commonplace with the growth of the global marketplace.
■
MATERIALS AND METHODS
Reagents and Supplies. Phosphate-buffered saline (PBS, catalog P3813), Tween 20 (catalog P7949), and other reagents, unless specified otherwise, were purchased from Sigma-Aldrich Co. (St. Louis, MO). LC-MS certified clear glass total recovery vials with presplit PTFE/Silicone septa (Waters Corporation, Manchester, UK) were utilized as sample injection vials. To avoid possible cross-contact among samples, disposable polypropylene containers were used as were disposable borosilicate spatulas to aliquot the various cumin samples for the different analyses. Whole seed and ground brown/green cumin (Cuminum cyminum L., family Apiaceae) were purchased from nine suppliers during late 2014 and the first two months of 2015 as imported products from Turkey, India, and Egypt. In addition, two types of black cumin known as Kalo zeera (Bunium persicum (Bioss.) B. Fedtsch. or Bunium bulbocastanum L., family Apiaceae) and Kalonji (Nigella sativa L., family Ranunculaceae) were purchased as representatives of products also commonly referred to as cumin. At least two lots or multiple subsamples of each of the cumin that tested positive were tested using ELISAs to rule out sampling anomalies. Peanuts, tree nuts, and legumes were purchased from the Golden Peanut Company (Alpharetta, GA) and Nuts.com (Cranford, NJ). Peanut and tree nut shells were obtained by purchasing legumes and nuts (in shells) and removing the internal meat. Commercial muffin samples formulated with dark roast peanut flour were prepared in collaboration with General Mills (Minneapolis, MN). Mahleb, produced in Turkey, was purchased from Penzeys Spices (Wauwatosa, WI). Other food samples, fruit, spices, and plant products were acquired from local retail suppliers. Samples were ground using either an IKA A 11 Basic Analytical Mill or an IKA Tube Mill (IKA Works, Inc., Wilmington, NC). Protein Analysis. Protein content of phosphate buffered saline/ 0.05% Tween 20 (PBST) extracts of cumin and other samples, prepared for Western blot analysis, were determined using the DC Protein Assay (Bio-Rad Laboratories, Inc., Hercules, CA), which is a modified version of the Lowry Assay suitable for use with the solvent milieu employed.9 The protein content of the various cumin samples, calibrated relative to bovine serum albumin (BSA), ranged from approximately 40 to 50 mg/g (CV values 0.25%), it did not display cross-reactivity with the Veratox ELISAs for peanut and hazelnut (
