Analysis of Organophosphate, Pyrethroid, and Methoprene Residues in Wheat End Products and Milling Fractions by Immunoassay

Cereal Chem. 73(5):605-612 Wheat grain was spiked with five levels of three grain protectant mixtures, aged, then milled and further processed into a wide range of end products including seven types of bread and noodles. Enzyme-immunoassay methods for quantitation of residues of three organophosphate (fenitrothion, chlorpyrifos-methyl, and pirimphos-methyl), two synthetic pyrethroid (bioresmethrin, permethrin) grain protectants and an insect growth regulator (methoprene) were applied to the analysis of both the milling fractions and the end products. Three parameters were investigated: 1) potential matrix interferences obtained using a simple methanol extraction protocol; 2) a comparison of data obtained using the immunoassay and conventional instrumental methods (gas-liquid chromatography or high-performance liquid chromatography); and 3) changes in the residue levels during milling and processing. Where matrix interference did occur, it was typically manifested as a decrease in assay sensitivity in the presence of the extract of the sample under study. However, methanol extraction of residues gave relatively few matrix interferences in the case Harvested wheat grain often requires protection during storage from damage by a range of insect species. Cereal grains can be treated with degradable pesticides, including organophosphates, carbamates, synthetic pyrethroids, and insect growth regulators to prevent insect infestation before processing and consumption (Bengston et al 1983, Snelson 1987, Arthur 1992). The identity and amount of pesticide residues in wheat products is important because: 1) small amounts of pesticide or pesticide metabolite may persist in the baked or cooked product; 2) pesticides used on some grains and in some countries may not be allowed in other situations; and 3) an increasing proportion of customers specify "residue-free" grain. Insecticide use brings with it the likelihood of residues appearing in foods derived from these grains. We have developed antibody-based tests for the major pesticides used on stored wheat grain (Hill et al 1991, 1992, 1993; Skerritt et al 1992a,b; Edward et al 1993), and in some cases performed limited assessment of these assays with milling fractions (Skerritt et al 1992b, Edward et al 1993). With wheat grain, good correlations have been obtained between immunoassay data and data obtained using conventional instrumental methods (Skerritt et al 1994). In addition to grain handlers and traders, cereal processing companies and regulatory agencies are potential users of these methods. For processors, analysis of baked or cooked end products and intermediate products such as flours are also important. Therefore, we undertook a large performance trial of the antibody assays with these wheat end products. Three parameters were 'CSIRO, Division of Plant Industry, GPO Box 1600, Canberra ACT 2601 Australia. Fax: + 61 6 246 5351. E-mail: skerrittapican.pi.csiro.au 2 CSIRO, Division of Plant Industry, PO Box 7, North Ryde NSW 2113 Australia. 3 CSIRO, Stored Grains Research Laboratory, GPO Box 1700, Canberra, ACT 2601 Australia. 4 Bread Research Institute of Australia Incorporated, PO Box 7, North Ryde, NSW 2113 Australia. Current address: Cereform, 74-76 Redfern St, Wetherill Park, NSW 2164 Australia. Publication no. C-1996-0822-05R. © 1996 American Association of Cereal Chemists, Inc. of organophosphates, and matrix effects were seen in only some of the pyrethroid assays. The simplest approach to obtaining accurate results, when matrix effects were present, was to prepare the assay standards in an extract of a pesticide-free sample of the matrix under study. Generally, there was a close relationship between residue levels as measured by immunoassay and by instrumental analysis. The extent of residues in different milling fractions and persistence in different products varied with the compound and the product. As the milling extraction rate increased, the levels of residue in the flour, relative to the application rate, were greater. Similarly, baked products prepared from high-extraction-rate flours contained higher levels of pesticide, while white noodles (lowextraction-rate flour) and yellow noodles (alkali treated) contained low levels. Although the application rates used are lower, a greater proportion of pyrethroids, especially permethrin, were retained after milling and subsequent processing, compared with that of the organophosphates. studied: 1) identification of sample matrix interference in the immunoassays using a simple methanol extraction protocol, without clean-up; 2) comparison of immunoassay results with data from gas-liquid chromatography (GC) or high-performance liquid chromatography (HPLC) analyses; and 3) assessment of the breakdown or loss of residues during processing. MATERIALS AND METHODS Grain Treatments and Products Hard wheat (11.5% protein) from Western Australia was treated with five treatment levels of three pesticide mixtures: fenitrothion and bioresmethrin; chlorpyrifos-methyl and methoprene; pirimiphos-methyl and permethrin, representative of the common mixtures used in commercial practice in Australia (Bengston et al 1983), and including several of the major grain protectants used singly in other countries (Snelson 1987). Treatment levels were: fenitrothion, chlorpyrifos-methyl, and pirimiphos-methyl (2.5, 5, 7.5, 10, 20 ppm, mg/kg), bioresmethrin, methoprene, and permethrin (0.25, 0.5, 0.75, 1, 2 ppm). The grain was stored in sealed containers at 20-25C for two months before conditioning and milling. This period, together with losses on spiking, meant that residues in the grain at the time of testing were expected to be 4080% of these original values (Ardley and Desmarchelier 1978; Desmarchelier 1978, 1980; Noble et al 1982; Bengston et al 1983; Papadouolou-Mourkidou and Tomazon 1991). Milling fractions were prepared using a Buhler mill: flour (60, 75, 82% extraction rates), 90:10 flour (90 parts of wholemeal plus 10 parts of 75% extraction rate flour), pollard, and bran. Wholemeal was prepared by milling wheat in a Falling Number mill to pass a 0.8-mm screen. Breads were made according to standard commercial practice (Quail et al 1993). Pan bread formula contained 3% yeast, 2% salt, 2% fat, and 0.5% improver (flour basis). White bread was prepared from 75% extraction rate flour, bran-enriched bread from 90:10 flour and wholemeal bread from wholemeal. Vol. 73, No. 5, 1996 605 Flat (Arabic) bread was prepared using 82% extraction rate flour, with 1% yeast and 1% salt (on a flour basis); baked at 5000C for 30 sec. Chinese steamed bread was prepared from 75% extraction rate flour, with 2% yeast; steamed for 20 min. White (salted) noodles were prepared from 60% extraction rate flour with 3% salt; boiled for 6 min. Yellow (alkaline) noodles were prepared from 75% extraction rate flour with 1% sodium carbonate, and boiled for 7.5 min. Noodle doughs contained 32% water before boiling. Each of these products was prepared from 300 g of flour. Immunoassays Details of the immunoassay methods used have been described elsewhere (Hill et al 1991, 1992, 1993; Skerritt et al 1992a,b; Edward et al 1993; Skerritt et al 1994). Products were analyzed without drying. Samples (10-40 g) were extracted 48 hr in methanol using either 2.5 ml of solvent per gram (wheat, wholemeal, flours, noodles), 4 mug (breads), or 10 mlI/g (pollard, bran) with intermittent shaking. The breads and noodles were first homogenized for 2 min in the methanol with a probe homogenizer (Ystral, Dottingen, Germany). They were each extracted and analyzed in two separate assays; immunoassay data presented are the means of these analyses. Step 1. Microwell plates were precoated with appropriate antibodies by incubating overnight with 1 ,ug of antibody per 100 Vl of 50 mM sodium carbonate buffer, pH 9.6, in each microwell. The fenitrothion and methoprene microwells were coated commercially by Millipore (ImmunoSystems Inc., Scarborough, ME; now EnSys Inc., Durham, NC). Step 2. Methanol extracts of the grain products were allowed to settle for 1 hr after blending, then were diluted 1:5 to 1:50 in 50 mM sodium phosphate, 0.9% NaCl, pH 7.2, 0.05% Tween, 1% bovine serum albumin. For permethrin, 0.005% Tween was used, as the antibody binding is affected by higher concentrations (Hill et al 1991). Predilution of extracts in methanol was performed such that extracts had a final methanol concentration of 5%, except for fenitrothion and bioresmethrin (10%). Step 3. Diluted extracts or pesticide standards (100 Vl) then pesticide-peroxidase conjugates (100 ,ul) diluted in the same diluent used for the grain products were added to each well and incubated for 1 hr at 20°C. Step 4. The plate was washed, 160 VI peroxidase substrate, 3 ,3',5,5'-tetramethylbenzidine chromogen was added (Hill et al 1991) and incubated 30 min at 20°C. Step 5. Stopping reagent (40 ,u of 1.25M sulfuric acid) was added, and absorbance was measured at 450 nm. Immunoassay results were calculated from standard response curves, prepared by spiking pesticide standard into methanol or methanol extracts of pesticide-free grain, milling fractions, and baked products. They were then diluted as for the samples. Instrumental Methods Grain, milling fractions, and end products were stored at -20°C between immunoassay and instrumental analyses to prevent further degradation of pesticide residues. The organophosphate analyses of the milling fractions and all the permethrin and methoprene analyses were performed by the authors, while the bioresmethrin analyses and organophosphate analyses of the end products were performed by a contract laboratory (Academy of Grain Technology, Werribee, Victoria, Australia). In the case of organophosphate analyses, which were performed by both laboratories, the same wheat grain samples were analyzed in both laboratories and data cross-checked for consistency before proceeding to the end products. Grain products were stored by the contract laboratory at 40C until analysis. Methanol extracts of grain fractions containing organophosphates were analyzed by GC with a thermionic detector, after fractionation on 6% SE30: 4% SP2401 on Chromosorb W, 100-120 mesh (Alltech, Deerfield, IL). Permethrin was extracted using methanol, and determined by 606 CEREAL CHEMISTRY RP-HPLC (Novapak C-18 column, Waters, Milford, MA) using elution with 50% acetonitrile in water. Bioresmethrin was determined from hexane extracts using normal-phase HPLC (Micro Porasil column, Waters), eluted at 1 ml/min with a mobile phase of 0.3% propanol in hexane; detection at 225 nm. Methoprene was determined in methanol extracts of grain, milling fractions, and wheat products. Aliquots (1 ml) of each extract were evaporated to dryness under nitrogen and redissolved in 2% tetrahydrofuran (THF) in hexane. Analysis was also performed on a MicroPorasil column, but with elution at 2 ml/min in 2% THF in hexane; detection at 254 nm. Instrumental data not presented for the pyrethroids in end products. Moisture