Time-of-flight secondary ion mass spectrometry (TOF-SIMS) enables chemically imaging the distributions of various lipid species in model membranes. However, discriminating the TOF-SIMS data of structurally similar lipids is very difficult because the high intensity, low mass fragment ions needed to achieve submicrometer lateral resolution are common to multiple lipid species. Here, we demonstrate that principal component analysis (PCA) can discriminate the TOF-SIMS spectra of four unlabeled saturated phosphatidylcholine species, 1,2-dilauroyl-sn-glycero-3-phosphocholine (DLPC), 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC), 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC), and 1,2-distearoyl-sn-glycero-3-phosphocholine (DSPC) according to variations in the intensities of their low mass fragment ions (m/z ≤ 200). PCA of TOF-SIMS images of phase-separated DSPC/DLPC and DPPC/DLPC membranes enabled visualizing the distributions of each phosphatidylcholine species with higher contrast and specificity than that of individual TOF-SIMS ion images. Comparison of the principal component (PC) scores images to atomic force microscopy (AFM) images acquired at the same membrane location before TOF-SIMS analysis confirmed that the PC scores images reveal the phase-separated membrane domains. The lipid composition within these domains was identified by projection of their TOF-SIMS spectra onto PC models developed using pure lipid standards. This approach may enable the identification and chemical imaging of structurally similar lipid species within more complex membranes.

Principal component analysis (PCA) of time-of-flight secondary ion mass spectrometry (TOF-SIMS) data enables differentiating structurally similar molecules according to linear combinations of multiple peaks in their spectra. However, in order to use PCA to correctly identify variations in lipid composition between samples, the discrimination achieved must be based on chemical differences that are related to the lipid species, and not sample-associated contamination. Here, we identify the positive-ion TOF-SIMS peaks that are related to phosphatidylcholine lipid headgroups and tail groups by PCA of spectra acquired from lipid isotopologs. We demonstrate that restricting PCA to a contaminant-free lipid-related peak set reduces the variability in the spectra acquired from lipid samples that is due to contaminants, which enhanced differentiating different lipid standards, but adversely affected the contrast in PC scores images of phase-separated lipid membranes. We also show that PCA of a restricted data set consisting of the peaks related to lipids and amino acids increases the likelihood that the discrimination of TOF-SIMS data acquired from intact cells is based on differences in the lipids and proteins on the cell surface, and not sample-specific contamination without compromising sample discrimination. We expect that the lipid-related peak database established herein will facilitate interpreting the TOF-SIMS data and PCA results from studies of both model and cellular membranes, and enhance identifying the origins of the peaks that contribute to discriminating different types of cells.