Supporting Lipidomics data for: Marine-derived alkaloids circumvent resistance mechanisms by targeting bacterial phospholipids

Liquid Chromatography Tandem Mass Spectrometry (LC-MS/MS) Based Lipidomics Method Development The E. coli lipidomics method utilized here was developed in conjunction with the Protein Biogenesis Section of NHLBI. Details of that method are included in a concurrent report in preparation and a summary is included here for reference. 22 For all LC-MS/MS methods, LC-MS grade water, methanol, isopropanol and acetic acid were purchased through Fisher Scientific. All lipidomics methodology for E. coli utilized a conserved liquid chromatography method and a LD40 X3 UHPLC (Shimadzu Co., Kyoto, Japan). Via this LC method lipids were separated by headgroup class with a Water XBridge Amide column (2.5 μm, 3 mm X 100 mm, Milford, MA) and a 9-minute binary gradient from 5% 12 mM ammonium acetate in water pH 7.5 in acetonitrile to 50% 12 mM ammonium acetate in water pH 7.5 in acetonitrile. All lipids for were detected under negative polarity using a 7500 QTrap mass spectrometer (AB Sciex Pte. Ltd., Redwood City, CA). E. coli lipidomics methodology was developed by establishing list of theoretical multiple reaction monitoring pairs (MRMs) for phosphatidylethanolamine (PE), phosphatidylglycerol (PG), cardiolipin (CL), and associated lysophospholipids (LPE and LPG) based on the previous profiles of the fatty acid content of E. coli. This theoretical list was then reduced to only include lipids that were confidently detected in lipid extract from a E. coli.Lipidomics Sample Preparation and LC-MS/MS Analysis An overnight culture of E. coli (ATCC 25922) cultured in LB at 37 °C was diluted 1:50 in fresh LB to reach an OD600 of approximately 0.05. The diluted cultures were then incubated, shaking at 37 °C for 90 minutes to obtain early log phase cells, 2.5 hours for late log phase cells, and 6.5 hours for stationary phase cells. For each compound concentration and growth phase tested, 1 mL of culture was added to 1.5 mL microcentrifuge tubes along with the indicated concentration of sceptrin, polymyxin B, sodium dodecyl sulfate (SDS), or DMSO vehicle control. Sample tubes were incubated on a rocker at 37 °C for 30 minutes. Following this incubation, samples were centrifuged to remove media, and pellets were washed once with 1 mL of sterile 0.9% NaCl in water (w/v%) at room temperature. After the NaCl solution was removed by centrifugation, 0.4 mL of ice-cold LC-MS grade methanol (Fisher Scientific, catalog no. A456-4) was added to the pellet. These microcentrifuge tubes were immediately placed on ice and incubated for 5 min. After which, 0.4 mL of ice-cold LC-MS grade water (Fisher Scientific, W6-4) was added. The resulting cell suspensions were transferred to new vials (Fisher Scientific, catalog no. NC0418367) and stored at -80 C until processing for LC-MS analysis. Lipids were extracted from bacterial preps by adding 0.4 mL of chloroform to the above water:methanol mixtures. Samples were shaken for 30 min under refrigeration and centrifuged at 16,000 x g for 20 min at 4o C to establish layering. A portion of the bottom (organic) layer (~300 µL) was collected and dried down under vacuum. Samples were resuspended in an equivalent volume of 5 µg•mL-1 butylated hydroxytoluene in 6:1 isopropanol:methanol. Prior to LC-MS/MS injection, 50 uL of the initial buffer (5% 12 mM ammonium acetate in water pH 7.5 in acetonitrile) was added to each sample to equilibrate pH and solvent conditions and improve peak shape. LC conditions were maintained as described above. All lipid target details including source conditions are included in Supplementary Dataset 2. Lipidomics LC-MS/MS Data Processing and Bioinformatics All signals were integrated using SciexOS (Version 3.1, AB Sciex Pte. Ltd., Redwood City, CA). Signals with greater than 50% missing values were discarded and remaining missing values were replaced with the lowest registered signal value. Signals with an average intensity, after total sum normalization, across the dataset below 30,000 counts were excluded. Signals with a Quality Control (QC) coefficient of variance greater than 30 % were discarded. Statistical analyses and data visualization were carried out using in-house developed R code (version 4.4.1; 2024-06-14). Principle Component Analysis (PCA) was performed on z-scaled lipidomic profiles using the prcomp() function. Significant differences in lipid levels between DMSO and sceptrin treatment conditions were measured using the t_test() function from the Rstatix package. The matrix barplots presented in this paper were generated using the publicly available mbplot R package which was developed in-house.113 To reduce Type I error (false positive) from multiple comparison, resulting p-values were adjusted, where indicated, using the rstatix adjust_pvalue function with the Benjamini-Hochberg method.