Lipid-Based Discrimination of Foodborne Pathogens using Liquid Chromatography Ion Mobility Tandem Mass Spectrometry

Abstract

Foodborne pathogens present a significant public health concern, contributing substantial morbidity, a significant economic burden, and a threat to human food safety. Rapid and accurate identification of bacterial pathogens is essential for effective disease management and reducing the number of unexpected deaths. The limitations of conventional detection methods, including long analysis times, high costs, and limited discriminatory power to bacterial serotypes, highlight the need for rapid, accurate, and highly selective analytical techniques capable of addressing complex biological systems. Therefore, we have applied a multidimensional analytical platform based on liquid chromatography- ion mobility- mass spectrometry (LC-IM-MS/MS) for the detection and discrimination of foodborne bacterial pathogens using lipidomic profiling. Our method was able to detect common phospholipids, such as phosphatidylglycerols (PGs), phosphatidylethanolamines (PEs), and cardiolipins (CLs) from bacterial lipid extracts, and mass spectral results were able to discriminate between Listeria monocytogenes and Escherichia coli or Salmonella enterica species. However, distinguishing between serotypes of Escherichia coli or Salmonella enterica based on mass spectra was challenging. The integration of LC and IM separations enabled the resolution of lipid isomers that share the same m/z ratio, and tandem mass spectra allowed for structural elucidation, supporting accurate identification and characterization. Multivariate statistical analyses, including principal component analysis (PCA) and linear discriminant analysis (LDA), were applied to evaluate the robustness of our datasets and showed clear clustering and discrimination among nine bacterial samples. In addition to the single bacterial analyses, we also explored bacterial mixtures by co-culturing to mimic real- world conditions of co-existing multiple bacterial species. Mass spectral observation confirmed the capture of identical peaks from each individual sample and differences in lipid expression profiles were observed between individual and mixed culture. Future work will expand this approach to multi-omics, including proteomics and metabolomics, and incorporation of machine learning strategies to enable robust and accurate discrimination from complex datasets.