Unravelling chemical mechanisms in microbial interactions by combining thin layer chromatography, ion mobility and MALDI imaging mass spectrometry

Mass spectrometry (MS) is a method of choice in microbiology for untargeted detection and identification of bioactive compounds. Mass Spectrometry Imaging (MSI) has led to a growing interest in the in situ study of biomolecules produced when microorganisms interact with each other. However, in situ identification is still time-consuming and challenging due to the large chemical diversity contained in each pixel of the samples, resulting in very complex average MS spectra.
Here, we propose to exploit the power of combining Kendrick Mass Defect (KMD) analysis and collision cross section (CCS) values using mobility for the characterization of families of related compounds in MSI. The identification of compounds was supported by rapid thin layer chromatography (TLC) separation coupled to MALDI-MS/MS detection. Bacillus velezensis GA1 and Pseudomonas sp. CMR12a were inoculated at different distances (0.5, 1 and 2 cm) on a semi-solid agar-based medium and incubated at 30°C. Regions of interest were cut directly from the Petri dish and transferred to the target ITO plate. This assembly was placed in a vacuum desiccator until completely dry and covered with HCCA matrix (Sunchrom sprayer). Ion mobility in imaging mode was performed using the timsTOF fleX (Bruker Daltonics, Bremen, Germany). TLC separation of ROI extracts is analyzed by MALDI-MS/MS imaging on the rapifleX instrument (Bruker Daltonics) for rapid screening of compounds and on the solariX instrument (Bruker Daltonics) for exact mass and isotopic distribution. The data were processed and integrated with in-house software. The coupling of MALDI-MSI with ion mobility separation brings additional structural features/information. It allows data to be filtered according to a range of CCS values and it improves the confidence level for the identification of detected analytes, in addition to exact mass determination. A relationship between CCS and mass was also performed. In combination with KMD analysis, various families of compounds, such as lipids or lipopeptides could be automatically detected and identified. Through this workflow, the comparison of the three different culture conditions was greatly simplified and highlighted the changes that occur in the metabolism of the bacteria. In particular, we were able to observe a variation within the lipid composition of Pseudomonas sp. CMR12a as a function of distance from the Bacillus velezensis GA1 colony. Finally, TLC separation was successfully optimized for lipopeptides and lipids and validated the identification of detected compounds by simplifying the spectra and allowing image analysis. TLC also enables high throughput, in part due to the parallel imaging of up to 6 traces on a single TLC run. TLC plate matrix coating strategies were compared to optimize the MALDI MS signal. This workflow will also be applied to time-lapse (or time-dependent) experiments, to monitor the bioactive compounds production and migration as a function of the co-culture interaction time.