ASSESSING THE QUANTIFICATION OF SOY PROTEIN IN INCURRED MATRICES USING TARGETED LC-MS/MS

Soy-derived ingredients are commonly added as auxiliary components to a diverse range of food products. The versatile end-applications of commercially processed soy ingredients is concerning for the safety of allergic consumers. Immunological detection of soy proteins in food matrices has some drawbacks, including loss of epitope binding and matrix masking. Therefore, the aim of this study was to assess naturally incurred matrices with an existing liquid-chromatography tandem mass spectrometry (LC-MS/MS) workflow to determine the recovery of total soy protein. The existing LC-MS/MS workflow was time intensive. The peptide responses of five soy protein targets were compared between two dilution methods and two trypsin digestion conditions. The tryptic digestion was shortened by 3 hours from the addition of subsequent additions of trypsin at 1:100 (trypsin:protein) to a single addition at 1:50. External standards were prepared with an optimized serial dilution method after reconstitution (AR). This workflow requires less time and reagents. Three peptides, LSA, VFD, and NIL, had reproducible peptide responses among all conditions and thus were chosen as the final quantifying peptides. Absolute quantification of total soy protein was achieved with a combination of internal and external standards. Several replicates of external standards were prepared for the optimized serial dilution (AR). The variation of the external standard peptide responses was insignificant among replicate standard curves whether prepared on the same day or different days. Therefore, all replicate curves were able to be pooled into an average, or “master” standard curve. The total soy protein in incurred bread and frankfurter matrices was then estimated from the master curve. Several factors were assessed for the effect on protein recovery: soy ingredient type and concentration, heat treatment, and matrix. The effect of the matrix was determined to have the most appreciable impact on protein recovery. The lowest percent protein recoveries, less than 50%, were calculated for all uncooked matrices. The cooked matrices had percent recoveries between 50-150 % for both matrices. Soy ingredient type was also an issue for texturized vegetable protein (TVP), where total soy protein recoveries were low and variable among replicate extracts. With the LC/MS-MS method, detection of TVP was low but above the limit of detection. Further work is needed on the total soy protein recoveries of these matrices with immunochemical methods. Advisors: Melanie L. Downs and Joseph L. Baumer