Pennycress seeds from the SPRING32 germline were used (Nottingham Arabidopsis Stock Center, NASC, UK). Conditions for germination and growth were reported elsewhere [6]. Seeds from five developmental stages were selected for lipidomic analysis, which correspond to green (G, 12 days after flowering, DAF), green-yellow (GY, 19 DAF), yellow-green (YG, 26 DAF), yellow (Y, 33DAF), and mature (M, 45 DAF), according to a previous work [6]. Seeds corresponding to these different maturation stages were harvested, frozen in liquid nitrogen, and stored at −80 °C. A total of two samples per maturation stage (G, GY, YG, Y, and M) were analyzed.

Notation adopted for the identity of a molecular lipid subclass was either (C:n), where x is the total carbon number of the molecule and y corresponds to the total number of double bonds; or (ca:na/cb:nb/cc:nc) where ca, cb, and cc are the carbon number of each fatty acyl in the corresponding TAG subclass (with ca + cb + cc = C); na, nb, nc correspond to the number of double bonds of each fatty acyl (na + nb + nc = n). Phosphatidylethanolamine (PE, 18:0/18:0, 1069–79-0 Chemical Abstracts Service, CAS) and diacylglycerol (DAG, 18:0/22:6, 65,886–80-8 CAS) were used as internal standards in LC‒MS and purchased from Avanti Polar Lipids (Alabaster, AL, USA). Triheptadecanoin (C17:0/C17:0/C17:0, ≥ 99%, 2438–40-6 CAS) was used as internal standard for gas chromatography with flame-ionization detection (GC‒FID) analysis and purchased from Sigma-Aldrich (Madrid, Spain). The following standards were used in HPTLC analysis: 1-oleoyl-rac-glycerol (C18:1, ≥ 99%, 111–03-5 CAS); 1,2-dioleoyl-rac-glycerol (C18:1/C18:1, ≥ 97%, 2442–61-7 CAS); glyceryl trioleate (C18:1/C18:1/C18:1, > 99%, 22–32-7 CAS); trierucin (C22:1/C22:1/C22:1, > 99%, 2752–99-0 CAS); oleic acid (C18:1), ≥ 99%, 112–80-1 CAS); linoleic acid (C18:2, ≥ 99%, 60–33-3 CAS); erucic acid (C22:1, ≥ 99%, 112–86-7 CAS); methyl oleate (99%, 112–62-9 CAS); cholesteryl oleate (≥ 98%, 303–43-5 CAS); and cholesteryl stearate (96%, 35,602–69-8 CAS), which were purchased from Sigma-Aldrich (Madrid, Spain).

High-performance liquid chromatography (HPLC)-grade n-hexane (C6, 99%), n-heptane (C7, 99.9%), tert-butyl methyl ether (MTBE, 99.9%), glacial acetic acid (AcH, 99.8%), methanol (MeOH, 99.9%), isopropanol (99.9%), dichloromethane (DCM, 99.9%), chloroform (CHCl3, 99%), and tetrahydrofuran (THF, 99.9%, without stabilizer) were purchased from PanReac (Barcelona, Spain). Ammonium acetate for HPLC was purchased from Merck. LC/MS-grade MeOH was purchased from Fisher Scientific (Loughborough, UK).

Total lipids were extracted from pennycress seeds (0.1 g) with CHCl3‒MeOH (6 mL, 2:1, V/V) by the “Bligh and Dyer” method [7]. For total fatty acid quantification, a method based on a previous work [8] was performed through direct transmethylation using triheptadecanoin (30‒35 mg) as internal standard. Fatty acid methyl esters of total lipids were analyzed by GC‒FID, as previously described [6].

Glycerolipids in pennycress seed extracts were separated by LC in mode HPLC and quantified by MS/MS using multiple reaction monitoring (MRM) on the LIPANG platform using a general method previously described [9].

The lipid extracts corresponding to 25 nmol of total fatty acids were dissolved in 100 μL of CHCl3‒MeOH (2:1, V/V) containing 125 pmol of each internal standard: PE 18:0/18:0, DAG 18:0/22:6, and sulfoquinovosyl diacylglycerol (SQDG) 16:0/18:0. The latter was obtained as previously described, from spinach thylakoid and subsequent hydrogenation [10].

The LC separation method was adapted from a previous publication [11], using an Agilent 1260 Infinity II HPLC system (Santa Clara, CA, USA). The volume injected was 20 μL. The separation was performed on a diol column (150 mm length × 3 mm i.d., 5 μm particle size) (Macherey–Nagel, Düren, Germany), at 40 °C. The mobile phases were: C6‒isopropanol‒water‒ammonium acetate (1 M), pH 5.3 (625:350:24:1, V/V) (A), and isopropanol‒water‒ammonium acetate 1 M, pH 5.3 (850:149:1, V/V) (B). The flow rate of the mobile phase was 200 μL/min.

The distinct glycerolipid classes were eluted successively as a function of the polar head group. Internal standards were useful to correct the three different response zones of the chromatogram: DAG 18:0/22:6 was used to calibrate the responses of DAG, digalactosyldiacylglycerols (DGDG), monogalactosyldiacylglycerols (MGDG), and TAG; PE 18:0/18:0 to PE response. SQDG 16:0/18:0 was used to calibrate to cardiolipins (DPG), phosphatidic acids (PA), phosphatidylcholines (PC), phosphatidylglycerols (PG), phosphatidylinositols (PI), phosphatidylserines (PS), and SQDG responses.

Mass spectrometric analysis was done on a 6470 triple quadrupole (QqQ) mass spectrometer (Agilent) equipped with a Jet stream electrospray ion source under the following settings: drying gas heater: 260 °C; drying gas flow: 13 L/min; sheath gas heater: 300 °C; sheath gas flow: 11 L/min; nebulizer pressure: 25 psi; capillary voltage: ± 5000 V; and nozzle voltage: ± 1000 V. Nitrogen was used as the collision gas. The quadrupoles Q1 and Q3 were operated at widest and unit resolution, respectively.

ESI‒MS conditions for glycerolipid classes are shown in Supplementary Table 1S (Supplementary Information). Quantification was done by MRM using 30 ms as dwell time for all classes except CL (50 ms). MRM transitions of TAG were adapted from a previous publication [12], and they are detailed in Supplementary Table 2S.

Mass spectra were processed by MassHunter Workstation software (Agilent) for the identification and quantification of lipids. In addition to lipid amounts (pmol) being corrected for response differences between internal standards and endogenous lipids, instrument responses were also corrected by comparison with a external standard-quality control (QC) that corresponds to a known lipid extract from an Arabidopsis cell culture that was previously qualified and quantified by thin-layer chromatography (TLC) and GC‒FID [9].

HPTLC silica gel 60 plates (20 × 10 cm) without fluorescence indicators were employed, from Merck (Darmstadt, Germany). Before being used, plates were immersed in THF (5 min), dried at 70° C and in a vacuum (50 mbar) for 15 min, and predeveloped up to 90 mm using C7‒MTBE‒AcH (70:30:1, V/V).

The application of solutions of standards and samples as 4-mm bands was carried out with the Automatic TLC Sampler system (ATS4) from CAMAG (Muttenz, Switzerland), using the ATS4 filling quality method. Solutions of individual standards (0.33‒2 mg mL‒1), per standard in DCM‒MeOH (1:1, V/V), were applied in triplicate on the same plate (3 µL/band). Each sample (4 µL) was applied in triplicate on three different plates. Samples were dissolved (3‒4 mg mL‒1) in DCM‒MeOH (1:1, V/V). The minimal distance between tracks was 6 mm and distance from the lateral and lower plate edges was 10 mm. One or more tracks were left empty, as blanks for chromatography and MS analysis. The plates were developed in a horizontal developing chamber (20 × 10 cm, CAMAG) up to 70 mm using C7‒MTBE‒AcH (70:30:1, V/V). Detection was carried out using a TLC Scanner 3 (CAMAG) at 190 nm in absorption mode. winCATS software (v 1.4.3.6336, CAMAG) was used to control and process data from sample application and densitometry.

The TLC‒MS Interface 2 (CAMAG) was used for an extraction of each TAG band directly from the plate. Details of elution-based interface description and operation can be found elsewhere [13]. The interface was equipped with an oval, 4 × 2-mm extraction head that was positioned on the corresponding TAG-band center, whose x, y coordinates were provided by densitometry and winCATS software. Then, MeOH was delivered for band extraction at 0.2 mL/min by using a PU-2080 HPLC pump (Jasco, Tokyo, Japan). The eluate was directed through a 2-mm stainless steel frit to remove silica gel and then sent to a mass spectrometer, an ion trap Amazon Speed Spectrometer (Bruker Daltonics, Bremen, Germany). Electrospray ionization in positive ion mode (ESI(+)‒MS) was selected for TAG analysis. Electrospray ionization was conducted with capillary and endplate offset voltages of −4500 and −500 V, 36 psi as the pressure of the nebulizer gas (N2), 6.0 L/min as the flow rate of the drying gas (N2), and 120 °C as the drying gas temperature. Spectra were acquired in the m/z 70–1500 range at the ultra-scan mode. Bruker Daltonics Trap Control software packages v 8.0 and Data Analysis v 5.2 were used to control the mass spectrometer and process data. For each TAG peak, several HPTLC‒ESI(+)‒MS experiments were performed from replicate bands, and confirmation of identity was carried out by MS/MS (see Results section). These experiments were performed from different plates.

Data from LC‒MS are expressed as averages ± standard deviation (SD), with three replicates in each experimental group. The statistical comparisons among the different developmental stages during the seed maturation of pennycress were made using one-way analysis of variance (ANOVA), and averages were compared using Duncan’s multiple range test (P < 0.05). When data showed nonnormality, log or reciprocal transformations were made, and ANOVA was conducted with the transformed data.

To evaluate the correlation between the results from LC‒MS and HPTLC‒MS, Pearson and Spearman correlation coefficients were calculated from R Posit Cloud (RStudio Pro 2023.12.1 Build 402.pro1 version, Posit Software, PBC). Scatter graphs were done using ggplot function. The Shapiro‒Wilk test was used to evaluate whether sample data fit to a normal distribution.