Vendor-dependent mobile phase contaminants affect neutral lipid analysis in lipidomics protocols

Lipidomics is a well-established field, enabled by modern liquid chromatography mass spectrometry (LCMS) technology, rapidly generating large amounts of data. Lipid extracts derived from biological samples are complex and most spectral features in LCMS lipidomics datasets remain unidentified, colloquially termed lipidomics “dark matter”. In-depth analyses of triacylglycerol, diacylglycerol, and cholesterol ester species revealed the expected ammoniated and sodiated ions as well as 5 additional higher mass dark matter peaks. These additional peaks were of relatively high intensity and resulted from analyte adduction with alkylated amine contaminants from LCMS-grade methanol and isopropanol. Tandem MS (MS/MS) of adduct peaks yielded no lipid structural information, producing only an intense ion of the adducted contaminant. Analysis of bovine liver extract identified 33 neutral lipids with an additional 73 alkyl amine adducts. Removing alcohols in place for acetonitrile and methyl tert-butyl ether in the mobile phase resulted in a 60% decrease in neutral lipid annotations, but eliminated the formation of alkyl amine adducts. Analysis of LCMS-grade methanol and isopropanol from different vendors revealed alkyl amine adduct formation in one out of three different brands that were tested. Substituting solvents increased lipid annotations by 36.5% or 27.4%, depending on the vendor and resulted in >2.5-fold increases in peak area for neutral lipid species, dramatically affecting their quantification and detection. Using principal component analysis, the same bovine liver sample separated into vendor-based clusters. These findings demonstrate the importance of solvent selection and disclosure during lipidomics protocols and highlight the challenges when comparing data between experiments.