Ion mobility mass spectrometry for the study of mycobacterial mycolic acids

Lipids are highly structurally diverse molecules involved in a wide variety of biological processes. The involvement of lipids is even more pronounced in mycobacteria, including the human pathogen Mycobacterium tuberculosis, which produces a highly complex and diverse set of lipids in the cell envelope. These lipids include mycolic acids, which are among the longest fatty acids in nature and can contain up to 90 carbon atoms. Mycolic acids are ubiquitously found in mycobacteria and are alpha branched and beta hydroxylated lipids. Discrete modifications, such as alpha, alpha’, epoxy, methoxy, keto, and carboxy, characterize mycolic acids at the species level. Here, we used high precision ion mobility-mass spectrometry to build a database including 206 mass-resolved collision cross sections (CCSs) of mycolic acids originating from the strict human pathogen M. tuberculosis, the opportunistic strains M. abscessus, M. marinum and M. avium, and the nonpathogenic strain M. smegmatis. Primary differences between the mycolic acid profiles could be observed between mycobacterial species. Acyl tail length and modifications were the primary structural descriptors determining CCS magnitude. As a resource for researchers, this work provides a detailed catalogue of the mass-resolved collision cross sections for mycolic acids along with a workflow to generate and analyse the dataset generated.

www.nature.com/scientificreports/ However, none of these studies have investigated the use of ion mobility coupled to mass spectrometry to gain deeper insight into mycolic acids. Ion mobility-mass spectrometry (IM-MS) is a technique used to achieve molecular ion separation based on ionic mass, charge, and conformations in the gas phase [23][24][25] . In contrast to conventional mass spectrometry, the additional information on conformation, which is measured in terms of collisional cross section (CCS), improves the accuracy of identification of molecules with similar structures. Indeed, through ion mobility, the rotationally averaged CCS of an ion can be measured. CCS is an important distinguishing characteristic of an ion in the gas phase. It is related to its chemical structure and three-dimensional conformation.
This method has been previously used to identify lipid isomers, which are lipids that exhibit the same mass but different structures (hence, different CCS values), and classify them according to lipid head group, chain length, and degree of unsaturation [26][27][28][29] .
In this study, we focus on the relationship between lipid structure and gas-phase conformation via IM-MS analysis to generate an atlas of mycolic acid methyl ester (MAME) isomers in relevant mycobacterial species.

Results and discussion
Data processing workflow. A subset database from online sources (https:// www. lipid maps. org/) and the literature was created in comma-separated values (CSV) format 15,[30][31][32] . Analysis of fragmentation data from the pooled quality control samples performed in Mass Hunter Qualitative Analysis (Agilent Technologies) was used to extract compounds with corresponding spectral information. Data were placed into Sirius 4 (Lehrstuhl Bioinformatik Jena) for molecular formula assignment. Fragment formulas were added to the database 33,34 . Finally, all compounds in the database were annotated with CCS using Mass Profiler (Agilent Technologies).
A screening workflow was set up in Skyline (MacCoss Lab), in which all mycolic acids were extracted from IM-QToF 35 . All Ions MS/MS data were collected with accurate mass, retention time and CCS (Fig. 1). Data were filtered to exclude matches with an isotope dot product score lower than 0.8 and a mass error of > 10 ppm.
Mycolic acid distribution and ion forms. MAMEs were identified and assigned primarily based on exact mass measurements and MS/MS profiles. The CCS measurements and corresponding mobility-mass cor-  www.nature.com/scientificreports/ relations were utilized to provide additional confidence in the lipid assignment in conjunction with the mass measurement accuracy (± 10 ppm observed in this work). This work represents 206 CCS values for uniquely identified mycolic acids from 5 mycobacterial species and 6 MAME families analysed in the positive ion mode and obtained on a drift tube instrument operated with nitrogen drift gas ( DT CCS N2 ). Using the solvents required for chromatography, across all 5 mycobacterial species, ~ 76% and ~ 24% of the total ion abundance of corresponded to mycolic acid methyl ester species with proton [M + H] + and sodium [M + Na] + adducts, respectively ( Fig. 2A). In previous literature, most MAMEs are reported as sodium or ammonium adducts 36 . In contrast, we observe that the large majority of adducts are [M + H] + , which can be explained by the preconditioning of the LC system by phosphoric acid; this was performed to reduce the presence of ions, such as Na + and K + .
In terms of mycolic acids categories, carboxy; methoxy; keto; epoxy/w1-methoxy; alpha' and alpha were annotated. More particularly, in Mtb, alpha mycolic acid methyl esters represent 55%, keto mycolic acid methyl esters represent 39% and methoxy mycolic acid methyl esters represent 6% of the total MAME ion abundance measured. In M. abscessus, alpha mycolic acid methyl esters represent 55%, alpha' mycolic acid methyl esters represent 32%, epoxy/w1 methoxy mycolic acid methyl esters represent 9% and carboxy mycolic acid methyl esters represent 5% of the total MAME ion abundance measured. In M. avium, alpha mycolic acid methyl esters represent 44%, epoxy/w1 methoxy mycolic acid methyl esters represent 6% and carboxy mycolic acid methyl esters represent 50% of the total MAME ion abundance measured. In M. marinum, alpha mycolic acid methyl esters represent 40%, keto mycolic acid methyl esters represent 52%, methoxy mycolic acid methyl esters represent 9% and carboxy mycolic acid methyl esters represent 8% of the total MAME ion abundance measured. In M. smegmatis, alpha mycolic acid methyl esters represent 58%, alpha' mycolic acid methyl esters represent 14%, and epoxy/w1-methoxy mycolic acid methyl esters represent 28% of the total MAME ion abundance measured. These observations are summarized in the histogram in Fig. 2B and are in agreement with the thin layer chromatography from growing mycobacteria (Figs. S1 and S2). These distributions correlate well with those found in the literature using conventional methods 14,37-41 .

IM-MS correlation.
In this work, all MAME classes exhibit a positive mobility-mass correlation in conformational space analyses, as typically observed with lipid classes in other studies. Within the MAME lipid trendline, unique families could be further differentiated by their respective DT CCS N2 information, as each family exhibited an average CCS increase in 0.1-0.2 Å 2 per mass unit within the investigated mass range of 850-1400 m/z (Fig. 3 www.nature.com/scientificreports/ size for each of the mycolic acid methyl ester families with a corresponding slope, DT CCS N2 vs. mass, ranging from 0.1-0.2 Å 2 per mass unit. The CCS/mass relationship presented here describes MAMEs differing only in chain length, decorations, or adduct type and can be used to identify lipid features that are ambiguous to characterize by mass alone or conformations.

, Supplementary Table 1). A closer examination of the IM-MS data shows a regular increase in
For protonated adducts, the same mass m/z can correlate to more than one CCS. These findings can also be observed for the alpha MAMEs in M. marinum at m/z 1152.202 correlated to CCS at 388. 17  As in M. avium, these data indicate that for some carboxy MAMEs, a m/z can correlate to more than one CCS, such as for m/z 1001.987 with CCS at 340.01 Å 2 and CCS at 348.89 Å 2 (Fig. 4), suggesting that this lipid can adopt a different conformation in the gas phase. This information may indicate that the carboxy position is different in the meromycolic chain. A similar conclusion can be drawn for alpha' mycolic acids, in which the double bond found in the meromycolic acid chain could be in different positions. As a result, the conformation www.nature.com/scientificreports/ of the meromycolic chain and therefore the CCS changes; however, the m/z remains the same, giving new insights into the complexity of mycolic acids. In this study, 206 DT CC N2 measurements are presented for the MAMEs for 5 mycobacterial species representing 6 classes of decorations. This correlation m/z to CCS is unique to MAMEs and provides resources for future investigations. In addition, using this approach, we observed that MAMEs can adopt multiple CCS values, which could result from different conformations in the gas phase. The majority of isomeric mycolic acids belong to the alpha, keto and carboxy classes. Compared to the acyl chain length, the degree of unsaturation is four times more impactful on CCS 26,28 ; thus, some MAMEs are more likely to form isomers, resulting in the same m/z but different CCS values. In conclusion, this is the first study reporting on a workflow and the use of IM-MS on MAMEs from 5 different mycobacterial lipids, providing a resource for researchers.
Overall, this study demonstrates that IM-MS improves separation of mycolic acids, especially between the different isomeric forms that share the same molecular weight. Considering the important roles of mycolic acids in mycobacterial virulence and their differential distribution across mycobacterial species, accurate separation and characterisation of mycolic acids would greatly help the research in this field. As demonstrated in this study, IM-MS could be a powerful tool for this purpose based on its accuracy and specificity compared with conventional mass spectrometry or ion mobility spectrometry alone. Specifically, in LC-IM-MS studies CCS values are being increasingly leveraged as molecular identifiers that provide an orthogonal dimension of information to mass and retention time, and many software workflows such as the screening workflow shown here are being adapted to accommodate filtering on CCS values for increased specificity for both targeted compound detection and unknown compound identification. To enable the mycobacterial research community, this study presents a database of mycolic acids from various mycobacterial species and an example workflow of data analysis for such purposes. We hope to encourage future studies on the species-specific mycolic acid profiles and their functions in mycobacterial species. However, validation of CCS values is restricted to the limited number of lipid standards that are commercially available for mycolic acids, making one of the limitations for this study.

Material and methods
Bacterial strains and growth conditions. Mycobacterium tuberculosis reference strain H37Rv, M. Preparation of mycolic acid methyl esters. Mycolic acid methyl esters (MAMEs) were prepared as described by Bottai et al. 42 . Briefly, heat-killed mycobacteria were first washed 3 times with PBS. The pellets were then delipidated by CHCl 3 /MeOH 1:2 (v/v) extraction for 12 h at room temperature followed by one CHCl 3 / MeOH 1:1 (v/v) extraction and one CHCl 3 /MeOH 2:1 (v/v) extraction for 3 h at room temperature. To analyse www.nature.com/scientificreports/ the cell wall mycolic acids, the delipidated bacterial pellets were mixed with 2 ml 15% tetrabutyl ammonium hydroxide (TBAH) solution and incubated at 100 °C for 5 h. After cooling to room temperature for 1 h, the samples were incubated with 2 ml of CHCl 3 and 100 μl of iodomethane for 1 h for methylation of the released mycolic acids. The organic lower layer after centrifugation at 2000×g at 4 °C for 10 min was collected and dried under a stream of N 2 . Dried methylated mycolic acids were resuspended in 3 ml diethyl ether and mixed thoroughly. All supernatants were collected after centrifugation at 2000×g at 4 °C for 10 min, transferred into a new glass tube and dried under a stream of nitrogen. Pellets containing MAMEs were resuspended in 50 μl of CHCl 3 and 5 ml of MeOH and left at 4 °C overnight for precipitation. MAMEs were collected after centrifugation at 2000×g at 4 °C for 10 min and dried under a stream of nitrogen. The purified MAMEs were suspended in CHCl 3 at a final concentration of 10 mg/ml and analysed by IM-MS in the positive ion mode.
Instrumentation. The instrumentation and experimental procedures are based on the work from Leaptrot et al. 26 . Briefly, a low-field drift-tube-based 6560 Ion Mobility LC/Q-TOF mass spectrometer (Agilent Technologies) was utilized to acquire accurate mass and CCS measurements from MAME samples. The instrument consists of an orthogonal electrospray ionization (ESI) source (Agilent Jet Stream) with a heated sheath gas nebulizer for desolvation and focusing of ions at atmospheric pressure. A single bore, resistively coated, glass capillary was used to transfer ions into the vacuum system. As ions exit the transfer capillary, they are directed by a high-pressure ion funnel into a trapping funnel with two wire grids for ion trapping and gating. Ions accumulate in the trapping funnel and are subsequently pulsed into the uniform field drift tube, which is approximately 78 cm in length. An additional rear ion funnel refocuses ions as they exit the drift tube and pass to a lower pressure region via a hexapole ion guide. Ions passed through a quadrupole mass filter and collision cell before mass measurement was performed in an orthogonal time-of-flight (TOF) mass spectrometer.
Experimental parameters. Data were acquired across all samples using UHPLC-IM-QToF. Chromatographic separation was performed using an Agilent 1290 Infinity II system (Agilent Technologies). Reversedphase chromatography was performed using an InfinityLab RRHD Eclipse Plus C18 column, 2.1 × 50 mm, 1.8 µm (Agilent Technologies). The LC system was first preconditioned with phosphoric acid before sample acquisition. For this procedure, the LC was disconnected from the mass spectrometer, and the LC system and column were washed overnight with 0.5% (v/v) phosphoric acid at 0.1 mL/min in water/acetonitrile 20:80 (v/v), followed by a wash with LC/MS grade water for 4 h until the pH was neutral and finally reconnected to the MS system. The column temperature was set at 65 °C. Mobile phase A consisted of water/acetonitrile 40/60 + 0.1% formic acid + 10 mM ammonium formate, and mobile phase B consisted of isopropanol/acetonitrile 90/10 + 0.1% formic acid + 10 mM ammonium formate. The following gradient was applied at a flow rate of 0. For all experiments, source conditions remained constant to ensure that the single field calibration could be used to measure CCS across all acquired data 43 . The following parameters were used: sheath gas temperature, 300 °C; nebulizer pressure, 35 psig; sheath gas flow, 11 l min −1 ; capillary voltage, 3500 V; nozzle voltage, 1000 V; and fragmentor voltage, 400 V. The data were collected in profile 2 GHz (extended dynamic range) mode.

IM-QToF All Ions MS/MS data acquisition.
With the chromatographic conditions described fragmentation data were acquired using an "All Ions MS/MS" data independent acquisition approach with two alternating experiments. The first experiment was used to acquire full spectrum ion mobility data of precursors using no collision energy. The second experiment was then employed to collect fragmentation information across the same mass range as the previous experiment by applying a collision energy of 42 eV. Ion mobility acquisition settings were as follows: mass range 800-1400 m/z, frame rate 2.5 frames per second, maximum drift time 100 ms, trap full time 20,000 ms and trap release time 150 ms.
QToF iterative MS/MS. MS/MS was performed only on a pooled sample to acquire data-dependent fragmentation data. Fragmentation was performed in the collision cell following quadrupole isolation of precursor ions with a preference for mycolic acids using a preferred inclusion list. Alternate scans were performed first in TOF-only mode to collect and analyse full-spectrum MS precursor data and for mass axis calibration, a second scan with a collision energy of 20 eV and a third scan with an increased collision energy of 42 eV. Data acquisition speed was optimized for collision-induced dissociation (CID) experiments based on chromatographic peak and number of potential mycolic acid coeluting in the chromatographic space. Thus, 10 spectra/second was used to ensure sufficient MS/MS scans were acquired. In this mode, the pooled sample was injected multiple times, thereby significantly improving the coverage of precursors being selected for fragmentation.

Data availability
The datasets used and analysed during the current study are available from the corresponding author on reasonable request.