Paternò-Büchi functionalization of phospholipids in tissue sections for double-bond position isomer resolved MALDI mass spectrometry imaging

Abstract

This work describes the development and implementation of novel matrices and sample preparation protocols for matrix-assisted laser desorption/ionization (MALDI) tandem mass spectrometry imaging (MS2I) of C=C double bond (DB) position isomers (DBPI) of lipids. The methodology is based on MALDI-laser-triggered Paternò-Büchi (PB) functionalization of the lipid DBs with PB-reactive MALDI matrices. Collisional activation of the PB-functionalized lipids leads to DB fragmentation, and the corresponding neutral losses are used to pinpoint DB positions in their fatty acid residues. In the first part of the project, liquid PB reagents were employed for on-tissue lipid functionalization before solid compounds were considered to serve as reagent and MALDI matrix. Benzophenone (BPh) was identified as a PB-reactive matrix and matrix application protocols were developed for pneumatic spraying or sublimation of BPh onto samples. Comparing the MS images acquired with the new matrix BPh to imaging results of a common matrix demonstrated BPh’s benefit for MALDI-MSI. Apart from working as a matrix for common MSI, BPh allowed to localize DB positions in various lipid standards and to resolve lateral DBPI-distributions with high resolution in mouse cerebellum (15 µm pixel size) and in a parasite worm (20 µm pixel size). To deposit the PB-reactive matrix, no special equipment and no additional sample pretreatment were required. The second project part targeted to find PB-reactive matrices, providing more efficient lipid ionization and functionalization for higher lateral resolutions and the DBPI analysis of more lipid classes. All twelve screened BPh derivatives could be used as MALDI matrices and enhanced the ion intensity (up to 50-fold) but only five compounds yielded PB products. The highest PB reactivity was found for 2-benzoylpyridine (BzPy). With BzPy, DBPI analysis was enabled for more lipid classes and for first examples in the negative-ion mode. Also, the lateral resolution of MALDI-MSI experiments was improved (7 µm) compared to BPh (15 µm). The annotation lists of mouse cerebellum tissue lipids contained species that were exclusively detected with BzPy attachment. Most likely due to its basic nitrogen, BzPy boosts the intensity of protonated ions. Compared to 15 µm pixel size in DBPI-resolved MS2I with BPh, BzPy allowed a reduced pixel size of 10 µm, making possible the resolution of lateral DBPI distributions in β-cells of the murine pancreatic islets. The DBP isomerism was elucidated for six lipid signals. Imaging the DBPI-diagnostic fragment ion signals showed distinct distributions for four DBPIs. Applying the established method to disease-model tissue was the third part of the project. Namely, it was used for MSI and DBPI analysis of diabetes- and cardiovascular-disease-model mouse tissue. Following up the findings in murine pancreatic islets, the DBPI ratios in diabetic and healthy pancreas tissue were compared. Significant variations suggested changing roles of the DBPIs. Concerning cardiovascular-diseased mice, DBPI distributions in lung tissue were studied. Besides resolving bronchi and blood vessels in MS images, an elevation of a phosphatidylcholine DBPI level was detected in the epithelium of the bronchi. These results exemplify the applicability of the developed method to biochemical and medical studies. In future, the method should be optimized for higher ion- and PB-yields to facilitate BPI-resolved MS2I of more lipid classes and low-abundance lipids. With matrices fulfilling this requirement, the PB-reactive matrix approach could be implemented as a routine method in medical research and diagnosis. As such, it would assist the investigation of localized lipid metabolism, e.g. desaturase and elongase activities and specificities.