U. S. Food and Drug Administration
Center for Food Safety and Applied Nutrition
June 12, 2001

Dr. Janet Anderson
Director
Biopesticide and Pollution Prevention Division
Environmental Protection Agency
1200 Pennsylvania Avenue
Washington, DC 20460

Dear Dr. Anderson:

Since there is a potential for corn-based foods to contain Cry9C protein, the Food and Drug Administration (FDA) has undertaken studies to evaluate the applicability and usefulness of methods for the determination of Cry9C protein in processed human foods. The accuracy and precision of methods used in food safety regulatory programs administered by FDA are determined through an interlaboratory validation study where feasible. FDA evaluated the EnviroLogix enzyme-linked immunosorbent assay (ELISA) test kit as a potential method for the detection of Cry9C protein in processed corn-based foods. The assay is a double antibody sandwich format and is based on the specific interaction between antibody and antigen.

The performance of the EnviroLogix ELISA test kit for the determination of Cry9C protein in human foods was evaluated using eight types of corn-based foods (starch, refined oil, soft tortillas, tortillas chips, corn flakes, corn puffs, corn muffins, and corn bread) in an interlaboratory study involving seven laboratories in the United States including the Medallion Laboratory, General Mills (Minneapolis, MN), Aventis Crop Science (Research Triangle Park, NC), National Food Processors Association (Washington, DC), EnviroLogix Inc. (Portland, ME), Food Allergy Research and Resource Program, University of Nebraska (Lincoln, NB) and the FDA (Washington, DC and Laurel, MD).

In this study, blind duplicate samples of blank, spiked and incurred (prepared from 100% Cry9C hybrid corn StarLink^TM) were analyzed. Cry9C protein from two different sources was used to spike the food products. Cry9C protein produced and purified from a bacterial host was used to spike test samples at 2.72 and 6.8 ng/g (parts per billion, ppb). Cry9C protein from StarLink^TM corn flour was used to spike test samples at 1.97 ng/g. All control test samples (blank materials) were found to contain <0.5 ng Cry9C protein/g.

A summary of the statistical analysis of the data follows. For samples spiked with StarLink^TM corn protein at 1.97 ng/g, average recoveries ranged from 73 to 122%, the within-laboratory relative standard deviations (RSD_r) ranged from 6 to 22%, and the between-laboratory relative standard deviations (RSD_R) ranged from 16 to 56%. For samples spiked with bacterial Cry9C protein at 2.72 ng/g, the average recoveries were 27 to 96%, the RSD_r ranged from 10 to 35% and the RSD_R ranged from 28 to 84%. The incurred test samples were found to contain Cry9C protein at levels ranging from 0.78 to 3187 ng/g depending on the product. The RSD_r ranged from 5 to 16% and RSD_R from 11 to 71% for the incurred samples.

The estimated limit of quantitation (LOQ) for the method was 1.94 ng/g protein using the American Chemical Society definition of the quantitation level. Using the LOQ coupled with the results from the statistical analysis of the data, it was decided that the method is applicable to the determination of Cry9C protein in eight types of corn-based products at levels ≥2 ng/g (2 ppb).

Dr. Mary W. Trucksess, Chief of the Bioanalytical Chemistry Branch, Division of Natural Products and AOAC International General Referee for Mycotoxins directed this study. The study report will be submitted to the Journal of the AOAC International for publication.

Hypertext updated by dav 2001-JUL-13