Rapid quantitation of oil compositions in blended oils by matrix-assisted laser desorption/ionization mass spectrometry

Abstract

Blended oils are becoming popular due to their advantageous physical and chemical properties, better nutritional values and enhanced flavors. However, deliberate mislabeling of oil compositions to mislead consumers is a problem frequently encountered in the market of blended oils. To better control the quality of blended oils and ensure food safety, quantitative labeling of oil compositions has become a trend while there is still a lack of rapid and reliable methods for quantitative analysis of blended oils, particularly for blended oils with multiple compositions. Conventional gas chromatography (GC) method for edible oil analysis needs chemical derivatization and column separation which are laborious and time-consuming. In this study, matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS) was applied to profile triacylglycerols (TAGs) in blended oils with minimal requirement of sample preparation and short analysis time. The MALDI-MS technique was advantageous for high-throughput analysis and provided high quality spectra with good reproducibility for blended oils. Based on the MALDI-MS spectra of blended oils, relationships between the intensity ratio of marker TAG peaks and the concentration of oil compositions have been investigated and employed to establish calibration curves and models for binary and ternary blended oils, respectively, from which quantitative results with good accuracy and precision were obtained after the optimization of marker ions. Quantitative analysis of quaternary blended oils has been demonstrated to be feasible using the intensity ratio of marker ions, but further study is required due to the complexity and challenge of data analysis. The developed intensity ratio-based method could approximate the abundance of oil compositions in blended oils based on the spectral data of pure oils, making this method powerful for rapid semi-quantitative analysis of blended oils. A chemometric approach, partial least squares regression (PLS-R), was employed to establish multivariate calibration models based on the acquired MALDI-MS spectra. It was demonstrated that the PLS-R approach provided good quantitative results for binary, ternary and quaternary blended oils with good limit of detection, i.e., detectability of 1.5% olive oil in sunflower seed oil, and allowed simultaneous quantitation of multiple compositions. Compared with the conventional GC method, the MALDI-MS method showed comparable quantitative performance, while allowed direct analysis of blended oils, analysis of one blended oil sample within minutes, and accurate quantitation of low-abundance oil compositions and blended oils with similar fatty acid (FA) contents. The compositions of several commercial blended oil products were successfully quantified by the developed method and some mislabeled products were determined, indicating the coupling of MALDI-MS and PLS-R was rapid, efficient and powerful for quantitative analysis of blended oils, especially for multi-compositions blended oils. To further simplify the quantitative analysis of blended oils, a framework was constructed to provide semi-quantitation of oil compositions of blended oils using the MALDI-MS spectra of pure oils as reference. Preliminary results demonstrated that the developed framework based on spectral comparison and spectral simulation could determine the compositions of blended oils and then quantify the determined compositions without the establishment of calibration relationships, which would significantly reduce the time required for quantitative analysis and be potential for rapid identification and approximate quantitation of unknown blended oils. However, further improvement and optimization of the framework was needed to decrease the false positive rate of determination and increase the accuracy of quantitation.