Advanced lipidomic techniques for high-throughput profiling of complex sphingolipids in plant tissues
Toshiki Ishikawa
TL;DR
This paper introduces a new method for quickly and efficiently analyzing complex sphingolipids in plant tissues using advanced lipidomic techniques.
Contribution
A novel single-tube extraction protocol and targeted mass spectrometry method for high-throughput profiling of plant sphingolipids is developed.
Findings
A single-tube extraction protocol enables rapid preparation of all sphingolipid classes with improved recovery.
Over 900 molecular species were separated and quantified in a single run using targeted mass spectrometry.
The method allows large-scale profiling of sphingolipid distribution across diverse plant materials.
Abstract
Sphingolipids are essential membrane components that regulate various cellular functions by forming ordered domains. The structural core of sphingolipids comprises a long-chain sphingoid base and its N-acylated ceramide, modified with diverse head groups to form complex sphingolipids. In plants, sphingolipid structures exhibit unique structural diversity in both the ceramide moiety and head group. The complicated sphingolipid structures suggest unique functions, but complicate lipidomic studies. A new methodology was established for simple sample preparation and targeted mass spectrometry, to advance high-throughput techniques for plant sphingolipidomics. A single-tube extraction protocol, consisting of the sequential procedures of enzyme inactivation, alkaline saponification and acidic 1-butanol phase partitioning, was optimized to enable rapid (~ 2.5 h) preparation of all sphingolipid…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5- —Saitama University
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsSphingolipid Metabolism and Signaling · Metabolomics and Mass Spectrometry Studies · Lipid Membrane Structure and Behavior
Introduction
Sphingolipids are characterized by long-chain bases (LCB) bound to fatty acids (FA) to form the unique amide lipid ceramide (Cer; Fig. 1). Cer is further modified with glycan- and/or phosphate-containing hydrophilic moiety at the C-1 hydroxy group, resulting in amphipathic complex sphingolipids, which are key components of eukaryotic cellular membranes (Mamode Cassim et al. 2019; Michaelson et al. 2016). The Cer moiety self-assembles to form highly ordered lipid compartments in membranes, the so-called membrane nanodomains, which serve as platforms for various membrane functions. While the fundamental functionality of sphingolipids is conserved in eukaryotes, their chemical structures vary widely across organisms, reflecting species-specific functions shaped by the genetic evolution of metabolic enzymes. Comprehensive qualitative and quantitative analytical techniques, collectively termed sphingolipidomics, are crucial for understanding the biology of this unique lipid class.Fig. 1. Structures of Cer and major complex sphingolipids in plants. Representative chemical structures of the three major sphingolipid classes commonly found in plant tissues are as follows: Cer, t18:1(8E)-h24:0; GlcCer, d18:2(4E, 8Z)-h16:0; and GIPC, t18:1(8E)-h24:0, with mannose at the 2nd residue of glycan. Hex, hexose; HexN, hexosamine; HexNAc, N-acetyl hexosamine; GlcA, glucuronic acid; Inos, myo-inositol; P, phosphate
Glucosylceramides (GlcCer) and glycosylinositol phosphoceramides (GIPC) are major classes of complex sphingolipids found in plants and fungi, with distinct glycan structures in each group. Plant GIPCs consist of a conserved core, Cer-phosphoinositol-glucuronate (GlcA), and extensions of various sugar chains (Mortimer and Scheller 2020; Fig. 1). Land plant GIPCs include two subclasses based on the second sugar residue of the glycan head: hexose (Hex) and (N-acetyl) hexosamine (HexN(Ac)). Two glycosyltransferases, GIPC Mannosyl Transferase 1 (GMT1) and Glucosamine Inositolphosphoceramide Transferase 1 (GINT1), are responsible for the specific sugar residue of the GIPC subclasses, which were first identified in Arabidopsis thaliana (Arabidopsis) and later in Oryza sativa (rice) and Medicago truncatula (Fang et al. 2016; Ishikawa et al. 2018; Lin et al. 2023; Moore et al. 2021). Both enzymes, belonging to the Glycosyltransferase Family 64 (GT64), are well-conserved in land plants (Ishikawa et al. 2018). The expression patterns of GMT1 and GINT1 determine the tissue distribution of the two GIPC subclasses. In Arabidopsis, AtGMT1 is ubiquitously expressed in whole tissues, whereas AtGINT1 is restricted to the pollen and seeds, reflecting tissue-specific distributions of Hex- and HexN(Ac)-GIPCs. Although the molecular functions of the different sugar residues remain elusive, genetic analyses have suggested subclass-specific functions for GIPCs. In Arabidopsis, the loss of predominant Hex-GIPCs in Atgmt1 mutants causes severe growth retardation and seedling lethality, which are never complemented by the ectopic expression of AtGINT1 (Fang et al. 2016; Ishikawa et al. 2018). Genetic disruption of AtGINT1 resulted in the loss of HexN(Ac)-GIPCs in flowers and seeds (Ishikawa et al. 2018). Atgint1 mutant seeds showed enlarged size and reduced sensitivity to salt and abscisic acid upon germination. A GINT1 ortholog in M. truncatula is expressed during root nodule symbiosis, and HexN(Ac)-GIPCs are required for nodule development (Moore et al. 2021). Rice contains HexN(Ac)-GIPCs with two or three sugar residues in whole wild-type tissues, and the ectopic overexpression of OsGMT1 affects immunity and flowering time (Lin et al. 2023). These findings indicated that the metabolic diversity of GIPC subclasses is closely related to the agricultural traits of plants. However, knowledge of the distribution of GIPC subclasses across a wide variety of plant species, including practical crops, is limited. Therefore, an improved method for high-throughput sphingolipidomics profiling is required.
Two factors make comprehensive analysis of plant sphingolipids difficult: poor solubility and molecular diversity. Plant GIPCs are only slightly soluble in some organic solvents commonly used for lipid extraction, such as chloroform. Therefore, specific solvent systems are necessary for extraction and LC–MS/MS analysis (Liu et al. 2024). Markham et al. (2006) established an efficient solvent system for GIPCs: the lower phase of a 2-propanol/hexane/water mixture. This is the standard method used in studies of plant GIPC (Blaas and Humpf 2013; Ishikawa et al. 2016; Markham and Jaworski 2007; Tarazona et al. 2015; Tellier et al. 2014). However, this process involves multiple sample preparation steps and requires considerable time for the evaporating large volumes of water-rich solvents. Another option for GIPC extraction is the batch method, which uses hot ethanol extraction and cold precipitation followed by column chromatography (Mamode Cassim et al. 2021). This method is suitable for the preparation of large amounts of GIPC for biochemical assays but is less quantitative and cannot cover Cer and GlcCer simultaneously. In the present study, we established a high-throughput sample preparation procedure suitable for comprehensive sphingolipidomics in plants. The optimized single-tube extraction with two-phase partitioning improved the recovery of GIPCs, time performance, volume of organic solvents consumed, and user-friendly handling during sample preparation.
The molecular diversity of GIPCs presents another technical challenge due to the complexity of the ceramide backbone. This diversity arises from over 100 or more species with various modifications with hydroxylation and unsaturation, further multiplied by head group diversity, complicating MS analysis. In this study, multiple reaction monitoring (MRM) for targeted detection was scheduled for transient monitoring only around the retention time of each sphingolipid species. Scheduled LC–MS/MS enabled the targeting of all complex sphingolipids in a single run, thereby facilitating high-throughput profiling of plant sphingolipids for large-scale omics surveys. Finally, this method was applied to profile the tissue-specific distribution of GIPC subclasses across various plant species, demonstrating its advantages in sphingolipidomics.
Materials and methods
Reagents
The solvents used for the mobile phases were HPLC or LC–MS grade from Wako Pure Chemical (Osaka, Japan). Internal standards (GlcCer d18:1-c12:0, Cer d18:1-h12:0 and d18:1-c12:0, ganglioside GM1) were purchased from Avanti Polar Lipids (Alabaster, AL, USA).
Plant materials
The plant species used in this study are listed in Table S1. Arabidopsis and rice seeds were harvested in the laboratory, while seeds from other plants were purchased from garden stores. Seeds were sown on soil and cultivated under a 12 h light/dark (60–100 µmol photons m^−2^ s^−1^) for 2−3 weeks at 23–25 °C. Whole parts of the shoots were collected, frozen in liquid N_2_, and lyophilized before lipid extraction.
Preparation of total sphingolipids
Total sphingolipids were prepared by the single-tube extraction method established in this study. The experimental scale was adjusted based on sample size (Table 1). Here, we show the minimal-scale procedure using a 2-ml disposable plastic tube and up to 5 mg of dried plant material. Although glassware has been commonly used for organic solvent-based lipid extraction, we confirmed that use of plastic labware did not affect the recovery of sphingolipids and the background of the targeted MS/MS analysis. Dry seeds or lyophilized shoots were immersed in 270 µL 1-butanol/methanol (2:1, v/v) and homogenized using stainless beads by vigorous shaking at 1200 rpm for 3 min for shoots or 1500 rpm for 5 min for seeds using a Shake Master Neo (Biomedical Science, Tokyo, Japan). An internal standard mixture containing 5 nmol/mL C12-Cer, 5 nmol/mL C12-hCer, 50 nmol/mL C12-GlcCer, and 100 nmol/mL GM1 ganglioside in methanol/2-propanol/water (2:2:1 by volume) was added to the samples before homogenization. The homogenate was heated for 10 min at 80 °C to inactivate the lipases and facilitate GIPC solubilization. After cooling, 180 µL of 1 N KOH was added, and the mixture was incubated for 20 min at 50 °C to hydrolyze the glycerolipids. Under these conditions, abundant glycerolipids (e.g., galactolipids in shoots and triacylglycerol in seeds) were completely removed, as confirmed by LC–MS/MS analysis (data not shown). The mixture was then acidified and phase-partitioned by mixing with 900 µL of 0.4 N HCl and 600 µL of 1-butanol. After centrifugation, the upper butanol-rich layer was collected in a new tube and evaporated to dryness at 40–50 °C by using a centrifugal concentrator connected to a diaphragm pump. The residue was resuspended in tetrahydrofuran (THF)/methanol/water (2:1:2, v/v) containing 0.1% formic acid (Markham and Jaworski 2007).Table 1. Experimental scales of solvent volumes based on sample sizeSteps/reagents/materialsSample size/solvent volumesLyophilized sample ~ 10 mg ~ 20 mg ~ 50 mgTube size2 mL5 mL15 mLIS mix^a^5 µL5–10 µL10–30 µLExtraction: 1-butanol/methanol (2:1)270 µL450 µL1.8 mLAlkaline hydrolysis: 1 N KOH180 µL300 µL1.2 mLAcidification: 0.4 N HCl900 µL1.5 mL6 mLPhase partitioning: 1-butanol600 µL1 mL4 mLResuspension solvent^b^100–200 µL200–500 µL250–1000 µL^a^Internal standard (IS) mixture: 5 nmol/mL C12-Cer, 5 nmol/mL C12-hCer, 50 nmol/mL C12-GlcCer, 100 nmol/mL GM1^b^THF/methanol/water = 2:1:2 supplemented with 0.1% formic acid. Final volume is adjustable to lipid contents in analytes
LC–MS/MS analysis
Sphingolipidomics analysis was performed using previously established methods (Ishikawa et al. 2016; Markham and Jaworski 2007) for targeted and high-throughput profiling. Sphingolipids were analyzed via MRM-based targeted detection using LCMS-8030/8040 (Shimadzu GLC, Kyoto, Japan) equipped with a Shim-pack Scepter HD-C18-80 column (3 µm, 2.1 mm × 75 mm) and a guard cartridge (3 µm, 2.1 mm × 10 mm) held at 40 °C. Gradient elution was performed using solvent A (THF/methanol/5 mM ammonium formate = 3:2:5 containing 0.1% formic acid) and solvent B (THF/methanol/5 mM ammonium formate = 7:2:1 containing 0.1% formic acid) under the following gradient conditions: 0 min, 10%A maintained for 2.5 min; 4 min, 25%B; 35 min, 100%B maintained for 2 min; 37.1 min, 10%B maintained for 3 min of preconditioning. Cer, GlcCer and GIPC species were detected by the structure-dependent MRM settings (Table 2). MRM scheduling was used for the comprehensive detection of whole sphingolipidomes in a single run. The MRM windows were opened for ± 1 min of retention time. A complete list of MRMs is provided in Table S2. Other LC–MS/MS conditions were as follows: capillary voltage, 4.5 kV; desolvation N_2_ gas flow, 50 L/min; nebulizer N_2_ gas flow, 1.5 L/min; conversion dynode voltage, 6 kV; source temperature, 250 °C; and collision gas (Ar), 230 kPa. The peak area was analyzed using LabSolutions software (Shimadzu, Kyoto, Japan). The calculation factors to compensate for the different MS sensitivities between the targets and internal standards were obtained using purified sphingolipids, as described previously (Ishikawa et al. 2016; Markham and Jaworski 2007).Table 2. Sphingolipid species targeted by LC–MS/MS analysisClassLCB^b^MS1MS2^c^#Species#MRMsCerd18:0d18:1d18:2t18:0t18:1[M + H]^+^[LCB-nH_2_O + H]^+^140140GlcCerd18:0d18:1d18:2t18:0t18:1[M + H]^+^[LCB-nH_2_O + H]^+^140140GIPC^a^d18:0d18:1t18:0t18:1[Cer-H_2_O + H]^+^672396^d^^a^GIPC contains 6 subclasses with different head groups (see Fig. 1)^b^LCBs are specific but FAs are commonly targeted in the lipid classes: α-OH and normal species of 16:0, 18:0, 20:0, 20:1, 22:0, 22:1, 23:0, 23:1, 24:0, 24:1, 25:0, 25:1, 26:0 and 26:1^c^“n” means the number of H_2_O loss; n = 1 for t18 Cer; n = 2 for d18 Cer and d18 GlcCer; n = 3 for t18 GlcCer^d^MRMs of GIPCs contain m/z overlapping in several species with isomeric ceramide moieties, resulting in 66 of total number of MRMs for each subclass including 112 targeted species
Assessment of sphingolipid recovery
To compare the extraction efficiency of the present method with previously established protocols, total lipids were prepared using the standard Bligh and Dyer method (1959; CHCl_3_/MeOH/H_2_O), methyl-tert-butyl methyl ether (MTBE) phase partitioning with methanol and water (Matyash et al. 2008; MTBE/MeOH/H_2_O), and 2-propanol/hexane/water extraction without partitioning (Markham et al. 2006; 2-PrOH/hexane/H_2_O). A schematic comparison of these methods is shown in Fig. S1. An additional procedure employing 1-butanol phase partitioning without KOH treatment or HCl acidification (1-BtOH/MeOH/H_2_O) was also evaluated. Total lipid extracts obtained from these methods were treated with 33% methylamine in 70% ethanol at 50 °C for 1 h to hydrolyze ester lipids, followed by evaporation to dryness under a stream of N_2_ gas. For the chloroform and MTBE extractions, the internal standard mixture was added after phase partitioning because of the low recovery of GM1 into the organic layers. Quantitative values obtained by the complete 1-butanol extraction method (1-BtOH/MeOH/HCl) as described above were used as the 100% reference to evaluate the relative recovery of the other methods.
Results and discussion
Optimization of a single-tube extraction method for plant sphingolipids
We developed a high-throughput extraction method for plant sphingolipidomic analyses. The extraction scheme is illustrated in Fig. 2 (see Fig. S1 for schematic comparison with previous methods described below). The procedure included several pretreatments: (1) heating in a water-free alcoholic solvent to inactivate lipases and relax tightly packed membrane lipid molecules; (2) alkaline hydrolysis of glycerolipids; and (3) acidic 1-butanol/water phase partitioning for higher yields, removal of debris and impurities, and reducing solvent volume. All pretreatment steps were designed to be conducted in a single disposable tube by sequentially adding solvents, resulting in a time-saving and simple handling. An acidification step before phase partitioning is necessary for better recovery of GIPC. Figure 3 compares sphingolipid yields obtained using different extraction methods. GlcCer was well recovered by all the methods tested, but the yield of GIPC was highly different between the solvent systems. Phase separation using chloroform/methanol/water (the Bligh and Dyer method) and MTBE/methanol/water solvent systems are commonly used for lipidomic experiments and show good recovery of GlcCer and Cer. However, these methods show significantly lower GIPC recovery rates from plant leaves. 2-Propanol/hexane/water mixture has been used as the standard for GIPC extraction (Markham et al. 2006). However, the solvent system showed a slightly lower recovery rate for Cer compared to the chloroform- or MTBE-based extraction and for GIPC compared to the 1-butanol/methanol/water system reported here. 1-Butanol/water partitioning under strongly acidic conditions improved the GIPC yield, as estimated by the removal of GIPC-interacting polysaccharides, as reported by Voxeur and Fry (2014). The acidic butanol partitioning containing plant tissue debris yielded all the major sphingolipid classes into the organic phase by single extraction, whereas the 2-propanol/hexane/water extraction required three repeated extractions for recovery, resulting in increasing the volume of extracts. The smaller solvent volume reduces evaporation time and is suitable for high-throughput processing using smaller tubes with a centrifugal concentrator. In addition, phase separation effectively removed insoluble materials as pellets in the interphase, resulting in clear extracts that were easily injected into the LC–MS. The procedure was completed in ~ 2.5 h (mostly incubation and evaporation) and was suitable for the parallel preparation of a large set of samples for LC–MS/MS-based sphingolipidomic analysis (we usually process 24–48 samples in a single extraction batch, and several batches can be processed per day).Fig. 2. Scheme of the quick extraction of total sphingolipids. The plant tissue homogenate was sequentially mixed with reagents and treated to hydrolyze ester lipids, remove insoluble debris, and obtain total sphingolipid extracts enriched in the 1-butanol layer of the acidic phase. See Materials and Methods and Table 1 for the detailed recipes and experimental procedures. IS, internal standard; 1-BtOH, 1-butanol; MeOH, methanolFig. 3Recovery rates of sphingolipids from plant leaves using various solvent systems. Fine powders (5 mg) from dried Arabidopsis and rice leaves were extracted using (1) chloroform/methanol/water, (2) MTBE/methanol/water, (3) 2-propanol/hexane/water, (4) 1-butanol/methanol/water, or (5) 1-butanol/methanol/HCl. The extracts were mixed with internal standards and quantified using liquid chromatography-tandem mass spectrometry. The recovery rate is expressed as a relative value, with the result of (5) set to 100%. Data are presented as mean ± SD from three independent extractions. Different letters indicate statistical significance according to Tukey’s honest significant difference test
Comprehensive sphingolipidomic analysis by scheduled monitoring of whole targets
The structural diversity of the Cer moiety of plant sphingolipids exceeds 100 species owing to the combination of various LCB and fatty acid components, further multiplied by the number of distinct hydrophilic head groups. Furthermore, molecular composition varies among plant species and tissues. Therefore, technical developments in sphingolipidomics are required to cover the entire plant sphingolipidome. To achieve the simultaneous quantitative profiling of plant sphingolipids, we established a scheduled MRM system. During reverse phase separation, sphingolipids were eluted in the following order: head group hydrophilicity (GIPC > GlcCer > Cer), carbon chain length of fatty acids, and hydroxylation rates of LCB and fatty acyl moieties. The retention time of each species was determined experimentally and theoretically, and the elution profile was based on the linear relationship between the number of carbon and hydroxyl modifications in each class (Ishikawa et al. 2016). The theoretical library of plant sphingolipids comprising experimentally determined and hypothetically estimated MRM settings and retention times is shown in Table S2. Each MRM window was opened with a retention time of 1 min. To reduce overlapping MRM schedules, we intentionally prolonged the gradient elution for 35 min, whereas a shorter elution time (10–20 min) was sufficient for the separation of sphingolipid species. As a result, the number of the overlapping MRMs was restricted to no more than 100, yielding 15 or more data points per peak. This approach enables the simultaneous measurement of more than 700 MRM transitions within a single analysis, resulting in a substantial improvement in time efficiency compared with non-scheduled measurements performed separately for a similar number of MRMs. Figure 4 shows a representative total ion chromatogram obtained using the scheduled MRM system. Under the elution conditions, most pigments, the major impurities of the total sphingolipid extracts, were eluted earlier than the sphingolipids within the first 4 min. Therefore, the use of a flow switching system, if available, is a useful option for alleviating the decreased MS sensitivity caused by ion source/mass analyzer pollution from impurities during successive analyses.Fig. 4. Simultaneous profiling of plant sphingolipids by using scheduled MRM. A total of 732 MRMs (396 GIPCs, 168 GlcCers, and 168 Cers) were simultaneously targeted using the scheduled MRM system in a single LC–MS/MS analysis. In the upper box, the green (GIPC), blue (GlcCer), and pink (Cer) lines indicate the open window for each MRM in ± 1 min retention time. The lower panel shows a representative total ion chromatogram (black line) and absorbance at 254 nm used to monitor impurities (red line). The arrow indicates the timing of the flow switch of the eluent from the waste to the MS
Application of the new method to the large-scale profiling of GIPC subclass distributions
We applied a high-throughput sphingolipidomic method to obtain a large-scale profile of the quantitative distribution of GIPC subclasses, namely Hex- or HexN(Ac)-GIPCs, classified according to the variable 2nd sugar residue (Fig. 1). The experiment was conducted in triplicate using two tissue types (young shoots and mature dry seeds) from 61 angiosperm species (see Table S1 for the plant list). Although plant seeds are rich in polysaccharides, proteins, and neutral storage lipids, the single-tube extraction method successfully eliminates these impurities. The composition of all GIPC species are shown in Tables S3 and S4, and the molar ratios of the Hex- and HexN(Ac)-GIPCs are summarized with a heatmap in Fig. 5. As reported previously, Arabidopsis has only Hex-GIPCs in shoots but both Hex- and HexN(Ac)-GIPCs in seeds. Closely related Brassicaceae plants showed a similar distribution of GIPC subclasses, except for Eruca sativa (rocket) and Raphanus sativus (radish), whose young shoots predominantly contained HexNAc-GIPCs instead of Hex-GIPCs, while maintaining similar ceramide backbones (Table S3). In a wide variety of eudicots, Hex-GIPCs are preferred in the shoots; however, both subclasses are ubiquitously present in the seeds. In contrast, several eudicot species, such as Solanaceae (e.g., Nicotiana benthamiana and Solanum lycopersicum) and Cucurbitaceae (e.g., Begonia semperflorens and Luffa cylindrica), and all monocot species (e.g., Allium ampeloprasum and Poaceae crops) tested in this study preferred HexN(Ac)-GIPCs in both shoots and seeds. Some eudicot species including Fabaceae (e.g., Glycine max and Lotus japonicus) and Asteraceae (e.g., Helianthus annuus) plants contain both subclasses in their shoots. Overall, the distribution of GIPC subclasses seems variable and independent of the taxonomic relationships of the plant species, probably due to tissue-specific transcriptional regulation of GMT1 and GINT1 genes. The functional aspects of the diverse GIPC composition will be addressed in future studies, particularly with respect to their different tissue distributions during shoot and seed development.Fig. 5. Distribution of GIPC subclasses in the shoots and seeds of angiosperms. The sphingolipidomic profiles of the shoots and seeds of various plants were obtained, and the ratio (mol%) of Hex- to HexN(Ac)-GIPCs was obtained using a heat map. See Tables S3 and S4 for the quantitative dataset of each GIPC species
Concluding remarks
This study provides technical advances in high-throughput sample preparation and simultaneous quantification of the comprehensive sphingolipids in plants. This method is well-suited for large-scale comparative sphingolipidomics of genetic mutants, chemical/environmental treatment experiments, and natural lipid products in a wide variety of plant species and tissues, and will facilitate further elucidation of the biological aspects of complex sphingolipids in plants.
Supplementary Information
Below is the link to the electronic supplementary material.Supplementary file1