Citation

Xie Peisi, Chen Jing, Xia Yongjun, Cai Zongwei. MALDI and MALDI-2 mass spectrometry imaging contribute to revealing the alternations in lipid metabolism in germinating soybean seeds[J]. Chin. Chem. Lett., 2025, 36(8): 110595  

Peisi Xie^a,1, Jing Chen^a,1, Yongjun Xia^b,c, Zongwei Cai^a,b,*     

Soybean seeds have been cultivated for centuries due to their rich protein and oil content [1,2]. The seed coat, the outermost layer, shields the seed from environmental factors, while the cotyledon, located underneath, is nutrient-rich and encases the embryo. This embryo, vital for plant reproduction, is protected until germination conditions are favorable, initiating with water absorption. This stage is crucial for plant productivity as it involves nutrient utilization, protein synthesis, and root emergence, fundamentally influencing agricultural yields by the biochemical transformations during this phase [3,4].

Lipid metabolism serves as an energy source, aids in membrane synthesis, and participates in signal transduction during different biological processes (e.g., seed germination) [5-8]. In soybeans, water uptake activates enzymes that hydrolyze stored triglycerides (TGs) in the cotyledons into glycerol and free fatty acids [3]. These acids are transformed in mitochondria into acetyl coenzyme A, subsequently generating ATP and forming glycerophospholipids (GPs) for cell membrane stability. One previous work demonstrated that when stored at 15 ℃ for 180 days, the levels of phospholipids, glycerolipids (GLs) and sterols decreased in the cotyledons of germinating soybean seeds [9]. Nevertheless, to the best of our knowledge, no research has been conducted to examine lipid alterations in different structures of soybean seeds during germination.

Technological advancements in matrix-assisted laser desorption/ionization mass spectrometry imaging (MALDI-MSI) have revolutionized the analysis of biomolecules (e.g., lipids, metabolites and proteins) in different biological samples [10-12]. MALDI-MSI enables the in situ examination of these molecules, providing insights into spatial distributions and relative abundances that are crucial for understanding biochemical transformations during seed germination [13-15]. For instance, Bhandari et al. examined the alterations in metabolism during the germination of oilseed rape [16]. They discovered that the metabolite, spermidine conjugate, primarily localized in the hypocotyl-radicle region of mature seeds. However, in germinating seeds, this metabolite was observed to be present in the emerging radicle. Ren et al. used MALDI-MSI to examine variations of lipid metabolism between transgenic and non-transgenic soybean seeds [17]. However, only 184 lipids were detected in seeds and only 18 lipids were found to have significant altered levels between two types of seeds. This may be because the MALDI ion source has a relatively low ionization efficiency, which is estimated to be on the order of 10⁻⁴ or even lower, for various lipid species [18-20]. Innovations like MALDI with laser-induced postionization (MALDI-2) have been developed to address these limitations, enabling more precise analyses without extensive sample preparation [21-23]. However, this novel technology has not been applied into the investigation of differences in lipid metabolism during seed germination.

In this present work, we carried out the first application of MALDI&MALDI-2 MSI to investigate alterations in lipid distribution and metabolism within the embryo and cotyledon of soybean seeds during germination by using a timsTOF flex MALDI-2 mass spectrometer (Bruker Daltonics, Bremen, Germany). Mass spectrometry-based lipidomics were used to examine the changes in levels of various lipid classes in these two structures of soybean seeds between different germinating timings using a Q Exactive Focus Mass Spectrometer (Thermo Scientific, Waltham, USA). The detailed experimental section including materials and reagents, seed germination, lipid extraction, LC-MS/MS instrumental and data analyses, preparation of tissue sections for MALDI-MSI, and MALDI-MSI data acquisition and processing were described in Supporting information.

To investigate the growth of soybean seeds (Hart Co., Ltd., Shenzhen, China) during germination, we cultured the seeds in 24-well plates with 1.5 mL of deionized water in a plant incubator. We observed that from days 0 to 1 in culture (Fig. 1A), the soybean seed underwent hydration and swelling, resulting in an enlargement in size. From days 3 to 6, the seed coat fractured, allowing the emergence of a part of the embryo within the two cotyledons. On day 7, due to the seed coat being shed, the cotyledons became visible. From days 7 to 12, the embryo gradually developed into different structures, such as the radicle and hypocotyl. Different structures except the radicle turned green and initiated photosynthesis, providing energy for seedling growth [24]. On day 14, the plumule penetrated through the cotyledons and gradually elongated upwards. The radicle developed more root hairs, which facilitate a better nutrient absorption [25].

Figure 1

	Download: JPG larger image
Fig. 1. (A) Optical images of soybean seeds at various timings during germination. The (B) physical map, (C) segmentation map, and (D) pLSA score plot.

The segmentation analysis of soybean seeds sections on day 2 during germination confirmed the presence of various internal structures. The results (Fig. 1B and C) showed that three structures including the cotyledon (greenyellow), seed coat (orange) and embryo (teal) were observed in seed sections. As shown in Fig. 1D, clear separations among these three structures were observed from the results of the score plot of the probabilistic latent semantic analysis (pLSA), indicating notable variations in cellular lipid metabolism among these structures.

To explore variations in lipid content within different structures of germinating soybean seeds, we conducted the liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis of lipids in the embryo and cotyledon of seeds on germinating days 2 and 7. The 60 µm-thick sections collected for LC-MS/MS analysis were adjacent to the 10 µm-thick sections used for MSI analysis. These specific time points were chosen for their relevance in capturing lipid dynamics, which refers to observing and analyzing the temporal changes in lipid concentration during the germination process. On day 2, the analysis highlights the initial changes in lipid composition as the seeds start utilizing stored lipids for energy. On day 7, the alterations in lipid profiles become more pronounced, reflecting the increased metabolic activity as germination progresses. These changes indicate ongoing metabolic processes, providing a deeper understanding of lipid metabolism during critical phases of seed germination. Moreover, the selection of these days facilitates the preparation of frozen sections for MSI analysis. Seeds on day 0 present challenges due to their low moisture content, complicating the production of thin frozen sections. Seeds on days 8 and 10 exhibit elongated embryos (Fig. 1A), posing difficulties in preparing complete sections that include both tissues.

The plot of partial least-squares discriminant analysis (pLS-DA) was used to discriminate the differences in lipid metabolism in the cotyledon and embryo between germinating days 2 and 7. As shown in Fig. S1A (Supporting information), the clustered quality control samples confirmed a stable instrumental performance. The distinct separations were found in the two tissues of soybean seeds in both positive and negative ionization modes, demonstrating that the germination process significantly influenced the lipid metabolism in soybean seeds. The volcano plots were applied to distinguish the unaffected and affected lipids that were chosen by the P < 0.05 and fold change < 0.8 (or > 1.2) in two structures during germination. Fig. S1B (Supporting information) illustrated that in positive and negative ionization modes, the embryo had more affected and fewer unaffected lipids compared to the cotyledon.

In the embryo, 504 lipids containing 253 GLs, 178 GPs, 44 sphingolipids (SPs) and 29 sterols were detected. Among them, 129 lipids exhibited significant alterations (Table S1 in Supporting information), meeting the criteria of P < 0.05, variable importance in projection (VIP) (VIP value > 1.0), and fold change < 0.8 (or > 1.2). These lipids included 18 diglycerides (DGs), 1 monoglyceride (MG), 30 TGs, 6 monogalactosyldiacylglerols (MGDGs), 2 lysophosphatidylethanolamines (LPEs), 1 lysophosphatidylcholine (LPC), 14 phosphatidylcholines (PCs), 7 phosphatidylglycerols (PGs), 2 phosphatidic acids (PAs), 4 phosphatidylserines (PSs), 7 phosphatidylinositols (PIs), 15 phosphatidylethanolamines (PEs), 16 ceramides (Cers), 1 sphingomyelin (SM), 2 acylglccampesterol ester (AcHexCmE), 2 acylglcsitosterol ester (AcHexSiE), and 1 acylglcstigmasterol ester (AcHexStE). A total of 434 lipids in the cotyledon including 193 GLs, 27 SPs, 204 GPs, and 10 sterols were detected. Among them, levels of 92 lipids showed significant changes (Table S2 in Supporting information). These lipids contained 11 SPs (8 Cers and 3 SMs), 2 sterols (1 ACHexCmE and 1 AcHexStE), 27 GLs (13 DGs, 4 MGDG and 10 TGs) and 52 GPs (1 LPE, 4 PAs, 22 PCs, 7 PEs, 4 PGs, 9 PIs and 5 PSs).

Previous studies demonstrated that different lipid classes possess diverse biological roles [26-28]. Therefore, we examined the levels of different lipid classes in both the embryo and cotyledon on germinating day 2 and day 7. Significantly changed lipid classes were selected according to the thresholds of fold change (FC, FC < 0.8 or > 1.2) and P value (P < 0.05). As illustrated in Figs. 2A and B, and Tables S3 and S4 (Supporting information), in GLs, decreased levels of MG and MGDG were found only in the embryo, while decreased levels of TG and DG were observed in two tissues. For the GPs in the cotyledon and embryo, the levels of PE, PI, PG and PA decreased, while levels of LPE and PS increased. Besides, a reduced level of PC and an elevated level of LPC were observed exclusively in the embryo. In SPs, an elevation of the Cer level was found solely in the embryo. In sterols, three lipid classes including ACHexCmE, AcHexSiE and AcHexStE with downregulated levels were found in two tissues, while four lipid classes containing zymosteryl ester (ZyE), campesterol ester (CmE), stigmasterol ester (StE), and sitosterol ester (SiE) with downregulated levels were only detected in the embryo. All these data suggested that compared to the cotyledon, the embryo showed faster but similar lipid changes during germination. The embryo is the part of the seed that develops into a new plant. It has high metabolic activity during germination to support its growth [29]. This increased metabolic activity may lead to faster lipid changes compared to the cotyledon. The lipid alterations found in our work were similar to those documented in one previous study that showed downregulated levels of GLs, sterols and GPs in the cotyledon of germinating soybean seeds under a long-time storage [9]. However, our results exhibited a certain degree of disparity with one other former study, which demonstrated a reduced levels of GLs and SPs but increased levels of sterols in the whole mung bean seeds during germination [30]. This could be contributed to the variations in growth and development among various botanical varieties.

Figure 2

	Download: JPG larger image
Fig. 2. Fold changes of different lipid classes classified as GLs, sterols, GPs and SPs in the cotyledon (A) and embryo (B) in soybean seeds on germinating days 2 and 7. Data were expressed as the average ± standard deviation. Six replicates were included in each group. (*P < 0.05, **P < 0.01, ***P < 0.001).

The pathway analysis of lipids in the cotyledon and embryo of soybean seeds was further performed by using one online database (Lipid MAPS Lipidomics Gateway). We found that germination process resulted in alterations of three pathways including biosynthesis and degradation of GPs and GLs (pathway 1), Cer-SM signaling (pathway 2) and sterol synthesis (pathway 3) in soybean seeds (Fig. S2 in Supporting information). In pathway 1, the increased level of LPE and decreased levels of PE, PA, PI and PG in the embryo and cotyledon indicated the lipid composition remodeling to support seedling growth and development. PE, PA, PI and PG are vital constituents of cell membranes [31]. The germination process requires the production of new membranes to establish cellular structures and facilitate cell expansion [32]. Lysophospholipids, such as LPE, are intermediate products in the breakdown of phospholipids. The degradation of phospholipids (e.g., PE) may contribute to the elevated levels of lysophospholipids (e.g., LPE). In our study, while the levels of most phospholipids decreased during germination, the level of PS increased in both the cotyledon and embryo of the seeds (Fig. 2). PS is also known to act as a signaling molecule in stress responses and development [33]. It can interact with specific proteins and enzymes, modulating their activities and influencing cellular processes [34]. The increased level of PS during seed germination may indicate its involvement in regulating specific signaling pathways that are essential for a successful seedling establishment.

TG and DG are both storage forms of energy in plant seeds [13]. The decreased levels of DG and TG in two tissues of soybean seeds indicated their utilization as a source of energy for the early growth of seedlings. Sterols in soybean seeds are produced through a complex series of biochemical pathways (Fig. S2). In general, acetyl-CoA can be transformed into squalene, which undergoes various modifications to produce different sterols (e.g., zymosteryl (Zy)) [35]. By bonding with various fatty acids, sterols can be transformed into ester forms (e.g., ZyE, SiE, CmE) [36]. Sterol esters can be further converted into other sterol forms (e.g., AcHexSiE, AcHexStE) by the chemical reactions of acetylation and hexylation [36]. These sterols are predominantly present in the cellular membranes of seeds and have vital functions in maintaining the flexibility and integrity of membranes. The reduced levels of sterols in both the embryo and cotyledon during germination suggested the membranes of these two tissues are undergoing increased fluidity and adaptability, facilitating cellular growth and expansion. Cer, as a constituent of cellular membranes, participate in diverse cellular activities such as cell apoptosis and signaling [28]. During seed germination, the embryo undergoes significant metabolic changes and cellular activities to initiate its growth. The elevated level of Cer in the embryo suggested a heightened requirement or membrane restructuring and cellular communication in germinating seeds.

LC-MS/MS-based lipidomic analysis has been extensively utilized in different botanical species due to the exceptional sensitivity of electrospray ionization (ESI) source for detecting semi-polar and polar molecules [37,38]. For MALDI source, it provides remarkable sensitivity for analyzing both nonpolar and polar species [39,40]. Therefore, we hypothesized that using both ESI and MALDI sources could expand the detection coverage of lipids in soybean seeds. In order to validate this, we initially compared the type and quantity of lipids detected by using MALDI and MALDI-2. Our previous works demonstrated that, compared to the negative ionization mode, the positive ionization mode in MALDI-2 significantly improved the detection numbers and intensities of lipid species over those observed in traditional MALDI [12,41,42]. Hence, in this study, we performed MSI studies in positive ionization mode using a common organic matrix ((2,5-dihydroxybenzoic acid (DHB)). The instrument, timsTOF flex MALDI-2, was operated at maximum working frequency (10000 Hz for MALDI and 1000 Hz for MALDI-2) with a laser size of 50 µm, a laser power of 90% and a detection range of m/z 100–1050. The optimized laser shots for MALDI-2 and MALDI were 80 and 350, respectively. All ions were assigned to lipid species with the mass-to-charge accuracy better than 5 ppm. Some of these lipids with high signal intensities in soybean seed sections were further identified by using MALDI-MS/MS (Fig. S3 in Supporting information).

As depicted in Fig. 3A, the spectrum of MALDI-2 showed more ion peaks than that of MALDI. A total of 280 and 242 lipid species (Table S5 in Supporting information) were detected in MALDI-2 and MALDI, respectively. These lipid species belonged to various lipid classes including DG, digalactosyldiacylglycerol (DGDG), MG, TG, LPC, lysophosphatidyiglycerol (LPG), lysophosphatidylinositol (LPI), lysophosphatidylserine (LPS), PA, PC, PE, PG, PI, PS, phosphatidylinositol phosphate (PIP) and SM. For lysophospholipids, signal intensities of 17 LPGs (e.g., LPG(14:4)), 4 LPIs (e.g., LPI(O-18:4)) and 12 LPSs (e.g., LPS(14:3)) in MALDI-2 were significantly higher than those in MALDI, while signal intensities of 2 LPGs (e.g., LPG(24:3)), 4 LPCs (e.g., LPC(18:1)) and 3 LPSs (e.g., LPS(22:2)) in MALDI-2 were significantly lower than those in MALDI (Table S5 and Fig. 3B). For phospholipids, 10 PAs (e.g., PA(O-34:3)), 6 PEs (e.g., PE(34:3)), 8 PGs (e.g., PG(34:5)), 2 PIs (e.g., PI(42:9)), 5 PIPs (e.g., PIP(26:1)) and 12 PSs (e.g., PS(44:6)) in MALDI-2 had significant higher intensities than those in MALDI, while 6 PGs (e.g., PG(34:0)), 2 PEs (e.g., PE(22:1)), 60 PCs (e.g., PC(32:0)) and 10 PSs (e.g., PS(32:3)) in MALDI had significant higher intensities than those in MALDI-2 (Table S5 and Fig. 3B). For GLs, signal intensities of 11 DGs ((e.g., DG(35:9)), 4 MGs (e.g., MG(20:3)) and 34 TGs (e.g., TG(54:3)) in MALDI-2 were significantly higher than those in MALDI, while signal intensities of 7 DGs (e.g., DG(40:9)), 4 TGs (e.g., TG(O-53:8)) and 1 DGDG (DGDG(36:4)) in MALDI-2 were significantly lower than those in MALDI (Table S5 and Fig. 3B). A total of 10 SMs (e.g., SM(37:1)) in SPs had higher intensities in MALDI than those in MALDI-2, while 1 SM (SM(35:2)) had lower intensity in MALDI than that in MALDI-2 (Table S5 and Fig. 3B).

Figure 3

Download: JPG larger image

Fig. 3. (A) Average mass spectra acquired by using methods of MALDI-2 (the bottom red line) and MALDI (the upper black line) in sections of soybean seeds on germinating day 2 in positive ionization mode using the DHB matrix. Each method included six soybean sections. (B) Representative ion images of lipids in seed sections using methods of MALDI (left) and MALDI-2 (right). Below the ion images, the average ion spectra for MALDI-2 (the bottom red line) and MALDI (the upper black line) were provided. All ion images and spectra were acquired by using the software of SCiLS Lab MVS 2023b (Bruker, Germany). All scale bars were 1 cm.

Collectively, our data suggested that compared to MALDI, MALDI-2 exhibited improved capabilities for detecting many lipid species in various lipid classes. But for the species in some lipid classes (e.g., SM, LPC and PC), it showed diminished capabilities for their detection. These findings were parallel to those documented in some previous works, indicating that MALDI-2 enhanced the intensities of diverse lipid classes in animal tissues, with the exception of LPC, PC and SM [43-45]. This could be attributed to the extra ion fragmentation during the ionization process in MALDI-2, which may cause a reduction in sensitivity for certain types of biomolecules. Besides, we also found that five lipid classes including LPG, LPI, LPS, PIP and DGDG were detected by the MALDI and MALDI-2 but not the LC-MS/MS, which may due to the differences in detecting various compounds between MALDI and ESI sources. Hence, in order to achieve a wide detection range of lipids in soybean seed sections, we used the combined techniques of MALDI and MALDI-2 in the following MSI experiments.

LC-MS/MS analysis showed that germination process led to changes in the levels of various lipid classes within two tissues of soybean seeds (Fig. 2). To investigate the alterations in spatial locations and abundances of lipid species in those lipid classes in both two tissues, we performed MALDI&MALDI-2 MSI analyses for sections of seeds between germinating days 2 and 7. Each group contained six replicates from three soybean seeds. As shown in Fig. S4, the pLSA score plots indicated distinct separations in the cotyledon and embryo between two different germinating days in both MALDI and MALDI-2, implying significant changes in lipid metabolism within two tissues during seed germination. Similar to the results of lipidomic data (Fig. 2), a majority of significantly changed lipids including 4 PAs (e.g., PA(36:4)), 29 PCs (e.g., PC(36:3)), 8 PGs (e.g., PG(38:6)), 6 PEs (e.g., PE(40:2)), 1 PI (PI(44:11)), 2 MGs (e.g., MG(22:5)), 12 DGs (e.g., DG(35:9)) and 29 TGs (e.g., TG(52:5)) in the embryo and 2 PAs (e.g., PA(O-32:3)), 8 PCs (e.g., PC(32:0)), 3 PGs (e.g., PG(34:0)), 2 PEs (e.g., PE(34:3)), 9 TGs (e.g., TG(52:4)) and 7 DGs (e.g., DG(36:5)) in the cotyledon had decreased abundances (Fig. 4, Tables S6 and S7 in Supporting information). The significantly increased abundances of these lipids classes in the embryo included 4 LPCs (e.g., LPC(16:1)), 1 PC (PC(26:1)), and 1 PE (PE(48:1)) and in the cotyledon included 8 PCs (e.g., PC(38:8)), 1 PG (PG(O-40:4)), 5 DGs (e.g., DG(42:12)) and 8 TGs (e.g., TG(54:3)) (Tables S6 and S7). For PS, the abundances of 22 significantly changed lipid species (e.g., PS(38:4)) were upregulated in the embryo (Table S6). In the cotyledon, a total of 3 PSs (e.g., PS(36:1)) and 1 PSs (PS(O-32:4)) had elevated and reduced intensities, respectively (Table S7). All the above MSI data demonstrated that compared to the embryo, the cotyledon exhibited slower but similar lipid changes during germination.

Figure 4

Download: JPG larger image

Fig. 4. Optical images and representative ion images of sections of soybean seeds on germinating days 2 and 7. In the optical image, the regions outlined by the orange and greenyellow lines represented the regions of the cotyledon and embryo of the seed sections, respectively. All images and spectra were acquired by using the software of SCiLS Lab MVS 2023b (Bruker, Germany). All scale bars were 1 cm.

In this work, a total of 5 lipid classes containing LPG, LPI, LPS, PIP and DGDG were only detected by the MSI but not the LC-MS/MS. For these lipid classes, a total of 11 LPGs (e.g., LPG(18:0)), 4 LPIs (e.g., LPI(O-22:6)), 19 LPSs (e.g., LPS(24:1)) and 4 PIPs (e.g., PIP(38:4)) in the embryo had increased intensities, while 1 DGDG (DGDG(36:4)) in the embryo had reduced intensities (Fig. 4 and Table S6). In the cotyledon, a total of 4 LPGs (e.g., LPG(20:2)), 3 LPIs ((e.g., LPI(O-18:4)), 1 LPS (LPS(14:3)) and 1 PIP (PIP(36:2)) had elevated intensities (Fig. 4 and Table S7). The increased abundances of LPG and LPI could be explained by the degradation of PG and PI, while the elevated abundance of LPS may be attributed to the enhanced synthesis of PS (Fig. 2). PIP is an essential component of cell membranes and is involved in membrane biogenesis [46]. The increase in PIP levels during germination could be a part of the process to support the growth and expansion of membranes in the embryo and cotyledon. All the obtained MSI results suggested that MALDI&MALDI-2 MSI could support and complement the results of the LC-MS/MS analysis. Besides, we should acknowledge that although MALDI and MALDI-2 are powerful tools for lipid analysis, they do have certain limitations. One notable limitation is the potential ion suppression effects, which can impact the detection sensitivity for certain lipid species (e.g., sterols). This limitation can influence the interpretation of our results by potentially underestimating the abundance of these lipids. Despite this challenge, the complementary use of LC-MS/MS helps to mitigate this issue by providing the quantitative data and confirming the presence of these lipids that not detected by MALDI and MALDI-2.

In summary, this work used LC-MS/MS-based lipid analysis combined with MALDI&MALDI-2 MSI to explore the alternations in lipid metabolism in the embryo and cotyledon of germinating soybean seeds. LC-MS/MS data showed that in comparison to the embryo, the cotyledon exhibited slower but similar alterations in lipids during germination. These alternations included decreased levels of GLs, sterols and phospholipids, as well as increased levels of lysophospholipids. Further lipid analysis using MALDI&MALDI-2 MSI showed its capability to enhance the precision and expand the detection scope of LC-MS/MS analysis. Our work may have the potential to advance seed germination research and deepen our comprehension of the biochemical processes during seed growth and development. Besides, the integration of LC-MS/MS and MALDI&MALDI-2 MSI techniques could be broadly applied in seed analysis and other areas of biological research (e.g., crop science and plant physiology), offering new insights into lipid metabolism and its impact on crop improvement and other biological systems.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Peisi Xie: Writing – review & editing, Writing – original draft, Validation, Methodology, Investigation. Jing Chen: Writing – original draft, Resources, Conceptualization. Yongjun Xia: Writing – review & editing, Supervision. Zongwei Cai: Writing – review & editing, Supervision, Project administration, Funding acquisition, Conceptualization.

This work was supported by National Natural Science Foundation of China (No. 22036001).

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2024.110595.