Direct Analysis of Triterpenes from High-Salt Fermented Cucumbers Using Infrared Matrix-Assisted Laser Desorption Electrospray Ionization (IR-MALDESI)
High-salt samples present a challenge to mass spectrometry (MS) analysis, particularly when electrospray ionization (ESI) is used, requiring extensive sample preparation steps such as desalting, extraction, and purification. In this study, infrared matrix-assisted laser desorption electrospray ionization (IR-MALDESI) coupled to a Q Exactive Plus mass spectrometer was used to directly analyze 50-μm thick slices of cucumber fermented and stored in 1 M sodium chloride brine. From the several hundred unique substances observed, three triterpenoid lipids produced by cucumbers, β-sitosterol, stigmasterol, and lupeol, were putatively identified based on exact mass and selected for structural analysis. The spatial distribution of the lipids were imaged, and the putative assignments were confirmed by tandem mass spectrometry performed directly on the same cucumber, demonstrating the capacity of the technique to deliver confident identifications from highly complex samples in molar concentrations of salt without the need for sample preparation.
KeywordsFood preservation IR-MALDESI Mass spectrometry imaging Phytosterols Direct analysis HRAM Fermented vegetable composition Q Exactive
For the past two decades, electrospray ionization (ESI) has been used extensively as a soft ionization source particularly suited for analysis of biological specimens because of the ease of coupling liquid chromatography (LC) separation to mass spectrometry (MS) detection. LC-MS has become widely used for qualitative and quantitative analysis of many biomolecules in the food and agriculture sciences. However, there are a number of limitations associated with ESI-MS methods that have yet to be fully overcome. For chemical analysis of complex biological systems, the samples are often subjected to a long and laborious preparation process to extract and purify analytes of interest and remove interfering contaminants [1, 2, 3, 4, 5, 6, 7, 8, 9]. The presence of high levels of low volatility substances such as inorganic salts is particularly destructive to the ESI ionization process, as they inhibit the transfer of ions from the electrospray droplets to the gas phase . Therefore, it is often necessary to desalt samples before analysis by mass spectrometry [4, 6, 11].
One way of circumventing some of these issues is by using ESI for post-ionization by introducing the sample solution into the electrospray plume after its formation, as is done in a number of published methods, including desorption electrospray ionization (DESI) , fused-droplet ESI (FD-ESI) , and matrix-assisted laser desorption electrospray ionization (MALDESI) . The salt tolerance of DESI and FD-ESI has been characterized in detail for pure analytes in solvents with salt content as high as 2 M [13, 15]. It may be hypothesized that compatibility with high salt samples is an inherent property of all ESI post-ionization methods, including MALDESI. Other strategies have been successfully developed to counter the effect of ion suppression by making fundamental changes to the electrospray, including nanoelectrospray ionization (nanoESI)  and probe electrospray ionization (PESI) .
MALDESI is an ambient ionization source where a pulsed laser is used to ablate material from a sample surface that is subsequently captured by an orthogonal electrospray plume where it is ionized . The most current version of MALDESI employs a mid-infrared laser for desorption. The energy of the laser pulse is absorbed by the endogenous water present in the sample with the optional addition of ice as an external matrix. The method has in this specific application been abbreviated as IR-MALDESI . In our laboratory, the instrument is routinely used for mass spectrometry imaging (MSI) of biological tissues without desalting or other chemical processing, and has proven to be useful for analyzing the spatial distributions of metabolites and drugs in tissue specimens [19, 20]. For an estimation of the salt content of the typical MALDESI sample, osmolalities of healthy mammalian fluids range around 0.3 osmoles/kg, roughly equivalent to 0.15 M sodium chloride . Herein, we demonstrate the capacity of IR-MALDESI to directly analyze fermented cucumber tissue stored in molar-level sodium chloride solutions and characterize biologically interesting molecules based on accurate mass as well as tandem MS.
Materials and Methods
LC-MS-grade methanol and water were purchased from Burdick and Jackson (Muskegon, MI, USA). Formic acid (MS-grade) was purchased from Sigma-Aldrich (St. Louis, MO, USA). High purity (99.999%) nitrogen gas for the higher energy collision (HCD) cell was purchased from Arc3 Gases (Raleigh, NC, USA).
Fresh, size 2B (3.5–3.8 cm diameter) pickling cucumbers were obtained from a local processor and packed into 32 oz. (946 mL) jars. A brine equilibrating to 1 M NaCl, 0.025 M acetic acid, and 4 mM potassium sorbate was added to the cucumbers at a 55:45 (w/w) cucumber to brine ratio. The cucumbers were inoculated with a starter culture at 106 CFU/mL Lactobacillus plantarum (LA0445, USDA-ARS, Food Science Research Unit, Raleigh, NC culture collection). L. plantarum starter culture was grown at 30 °C in deMan, Rogosa, and Sharpe (MRS; Becton, Dickinson and Co.) broth to a population of approximately 109 CFU/mL, centrifuged to pellet the cells, and resuspended in saline. After inoculation, the jars were sealed and incubated at 28 °C for 100 d to simulate a typical fermentation and bulk storage time. The cucumbers were fully fermented as evidenced by low residual sugars, a pH of 3.2, and ~120 mM lactic acid. A transverse 6.7 mm slice from the center of each cucumber was stored in the fermentation brine at 4 °C until analyzed.
A single lobe was removed from an axial section of the fermented cucumbers using a precleaned single edge blade, and was further sectioned into 50-μm thick slices using a Leica CM1950 cryostat (Buffalo Grove, IL, USA). This thickness was found to be the lowest practical setting, due to the fragility of the frozen cucumber. The samples were thaw-mounted onto precleaned glass microscope slides (VWR, Radnor, PA, USA) and analyzed without further preparation. Samples for MS/MS analysis were stored mounted on slides at –20 °C until time of analysis.
IR-MALDESI Mass Spectrometry
Direct full MS analysis (MS1) of a whole slice was performed immediately following preparation. A 50:50 mixture of methanol:water with 0.2% formic acid was used as the electrospray solvent. No external matrix was added, as unlike for animal tissue samples, a preliminary experiment had indicated no added benefit of depositing a layer of ice before analysis (data not shown). This is presumably due to the high content of endogenous water in the cucumber itself.
All experiments were conducted at the ambient temperature and pressure of the laboratory. A mid-IR laser (IR-Opolette 2371; Opotek, Carlsbad, CA, USA) at a wavelength of 2.94 μm was used to desorb material from the surface. Two 7-ns pulses at a 20 Hz repetition rate were used to ablate the material for each spectrum acquired. The IR-MALDESI source was coupled to a Q Exactive Plus mass spectrometer (Thermo Scientific, Bremen, Germany) and the m/z range of 200–800 was measured with resolving power of 140,000 (FWHM, m/z 200) in positive ion mode. Due to the pulsed nature of IR-MALDESI, the automatic gain control (AGC) feature was disabled. The ions generated in both ablation events were collected in the C-trap for a fixed injection time of 110 ms, and were subsequently measured in a single Orbitrap acquisition. Electrospray solvent was supplied at a flow rate of 2.0 μL/min. The electrospray driving voltage was set to 4 kV, and the capillary inlet temperature was held at 275 °C.
MS/MS experiments were carried out using the same method as described above, with the instrument in all-ion fragmentation (AIF) mode. Ions of interest were isolated using a 1.0 Th window and fragmented using HCD with normalized collisional energy (NCE) set to 20%. Each target molecule was sampled at around 200 scans/molecule across the tissue so as to include both mesocarp (flesh) and exocarp (skin) tissue. Mass spectra were acquired in the ranges of m/z 50–200 and 100–400.
Mass spectra were analyzed using Xcalibur (Thermo Scientific, Bremen, Germany). For MSI analysis, the raw files were converted to mzml format using MSConvert, part of the ProteoWizard toolkit , and from mzml to imzml format using imzMLConverter .
Ion images were generated using MSiReader  ver. 0.06 with a m/z bin width of 5 ppm (±2.5 ppm). The peak picking function in MSiReader was used to generate a list of masses present in the cucumber tissue, and this list was cross-referenced with the METLIN database ver. c1.1 (beta) . From the resulting putative identifications, the sterol lipids stigmasterol and β-sitosterol as well as the related triterpenoid lupeol were selected for further structural analysis with MS/MS.
Results and Discussion
The initial full imaging scan over a section of cucumber yielded 4500 individual MS1 spectra, each corresponding to the ablation of a 100 × 100 μm area. An unannotated raw spectrum averaged over multiple on-tissue scans is provided as Supplementary Figure S1 in the Electronic Supplementary Information to show the complexity and quality of the data collected. The entire dataset after processing with the routine imaging workflow detailed in the experimental section was interrogated for masses correlating spatially with the tissue, and a list of masses correlated to the on-tissue region was generated using the peak picking tool within MSiReader. The list was used to guide further inquiry, but should not be thought of as a comprehensive survey of unique identities due to lack of filtering for false hits such as may result from selecting multiple isotopes of the same molecule.
Structural Determination with Tandem Mass Spectrometry
Upon examination of fragments stemming from cleavages in the main ring structure of the two sterols, it can be seen that the fragments corresponding to cleavage of the position 9–10 bond combined with either of the 5–6 or 6–7 bonds in the Δ7 sterols are both present, whereas neither of the corresponding fragments for the Δ5 isomer are observed. Sitosterol and stigmasterol can be easily distinguished by the presence of a fragment at m/z 255.211, attributed to the loss of the side chain though a remote hydrogen rearrangement reaction that is possible in stigmasterol but not in β-sitosterol (Figure 2).
The triterpenoids associated with the 409.383 peak are known to have very similar fragmentation patterns [28, 29, 30], and have all been previously observed in cucumbers . However, the lack of a strong peak at m/z 191, as well as the equal abundance of the peaks at 257 and 259, indicate that cycloartenol is not a major constituent [28, 29]. The ion at m/z 219.210 is consistent with either the charge-retained ring cleavage of lupeol as annotated in Figure 2c, or a post-protonated fragment of retro-Diels-Alder decyclization at the 12=13 double bond in amyrin (not shown). The latter would be expected to be accompanied by a signal from the charge-retained fragment of m/z 191.180 (C14H23 ●+), which is not observed. Based on these data, lupeol is assumed to be the principal isomer in the fermented cucumber sample and m/z 409.383 is annotated as such in all figures and diagrams.
Mass Spectrometry Imaging
In Figure 3, it can be seen that sterol content is roughly constant throughout the cucumber, whereas the level of lupeol and/or isomeric pentacyclic triterpenoids is much higher in the exocarp, likely due to their incorporation in the protective layer of cuticular wax, which is known to contain relatively high levels of triterpenoids [39, 40]. Identically performed imaging of raw cucumber (included in the Electronic Supplementary Material as Supplementary Figure S3) shows the same trends in ion distribution, indicating that the presence of 1 M NaCl and the bacterial fermentation process does not significantly alter the ability of this technique to detect and visualize native triterpenoid lipid distribution.
IR-MALDESI is capable of analyzing biologically important molecules directly from complex biological samples treated with molar concentration of salts in the context of an imaging experiment with minimal sample preparation. As such, it has potential use as a complement to conventional chromatography-based analysis of salt-rich foods, such as fermented vegetables, which presently require laborious preparation and desalting before analysis. An added benefit to IR-MALDESI is the inherent suitability for spatial analysis (imaging), which can provide deeper insight into the function of biological systems. The use of high resolution accurate mass (HRAM) spectrometry allows putative peak assignments based on observed mass as well as confident assignments using tandem MS without prior separation.
The authors gratefully acknowledge the financial and material support received from the National Institutes of Health (R01GM087964), Mount Olive Pickle Company, and North Carolina State University.
- 6.Jessome, L.L., Volmer, D.A.: Ion suppression: a major concern in mass spectrometry. LC– GC N. Am. 24, 498–508 (2006)Google Scholar
- 7.Lu, B.Y., Zhang, Y., Wu, X.Q., Shi, H.: Separation and determination of diversiform phytosterols in food materials using supercritical carbon dioxide extraction and ultraperformance liquid chromatography-atmospheric pressure chemical ionization-mass spectrometry. Anal. Chim. Acta 588, 50–63 (2007)CrossRefGoogle Scholar
- 14.Sampson, J.S., Hawkridge, A.M., Muddiman, D.C.: Generation and detection of multiply-charged peptides and proteins by matrix-assisted laser desorption electrospray ionization (MALDESI) Fourier transform ion cyclotron resonance mass spectrometry. J. Am. Soc. Mass Spectrom. 17, 1712–1716 (2006)CrossRefGoogle Scholar
- 27.Sucrow, W., Reimerde, A.: Delta7-sterines from cucurbitaceae. Z Naturforsch B 23, 42–45 (1968)Google Scholar
- 30.Pop, G., Galuscan, A., Peev, C., Militaru, A., Vlase, L., Ardelean, L.: HPLC-MS identification of sterol fractios from vegetable oil. Rev. Chim. (Bucharest) 63(1046–1050) (2012)Google Scholar