All the chemical reagents and solvents were obtained from commercial sources and were utilized without further purification. N, N-dimethylglycine (DMG), N, N-dimethylalanine (DMA), N-(dimethylamino) pyridine (DMAP) and 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) were purchased from Sigma-Aldrich (St. Louis, MO, United States). DAG 18:0/18:0/0, DAG 16:0/18:1/0, DAGs 16:0/16:0/0, DAG 14:0/14:0/0, DAG 12:0/12:0/0, DAG 10:0/10:0/0 and DAG 8:0/8:0/0 were purchased from Avanti PolarLipids (Alabaster, AL, United States). DAG 18:0/0/18:0, DAG 16:0/0/18:1, DAG 18:1/18:1/0, DAG 18:1/0/18:1 were purchased from Nu-Check Prep, Inc. (Elysian, MN, United States). Chromatographic grade methanol and acetonitrile were purchased from Merck KGaA (Darmstadt, Germany). Mass spectrometry grade formic acid and other chemical reagents were obtained from Sigma-Aldrich Laboratories, Inc.

Male healthy C57BL/6 mice were provided by China Pharmaceutical University (Jiangsu, China). All mice were housed in an air-conditioned rearing chamber with an ambient temperature of 25 ± 2°C, relative air humidity of 50% ± 10% and the cycle of light and darkness is 12 h. Water and food were allowed to be freely accessible during the entire study. Animal studies have been approved by the Animal Ethics Committee of China Pharmaceutical University. After 1 week of rearing acclimatization, mice (n = 40) were randomly divided into different research groups (n = 10). First group is blank control group, in which mice were administrated with the saline (1 h apart, i.p.). The second group was the acute pancreatitis model group, in which mice were treated with two intraperitoneal injections of L-arginine (3.0 g/kg, 1 h apart, i.p.). The third group was the low-dose resveratrol (30 mg/kg, i.g.) combined with L-arginine group, in which mice were given the previously described L-arginine treatment and were administered resveratrol twice daily for 3 days. The fourth group was the high-dose resveratrol (150 mg/kg, i.g.) combined with L-arginine group, in which mice were treated with both resveratrol and L-arginine as described above. After injecting L-arginine for 1 h, resveratrol was administered by gavage twice daily for three consecutive days. L-arginine was prepared by solubilizing it in 0.9% saline, the pH of this solution was adjusted to 7 with hydrochloric acid. After animal treatment, serum samples were collected from above four groups. Simultaneously, lungs were taken. All samples were stored in a −80°C refrigerator.

DAGs (0.2–200 nM), DMG (0.125 mM), DMAP (0.5 mM) and EDC (0.25 mM), were dissolved in acetonitrile and dichloromethane (ACN: CH₂Cl₂, 1:1). Subsequently, all reagents were respectively added into glass tube and then vortex for 1 min. After centrifuging the tubes at 2,500 rpm for 1 min, the sample in glass vessels reacts at 45°C for 60 min. The derivatization reaction was quenched by introducing 1.5 ml CH₂Cl₂/MeOH (1:1, v/v) mixture and 0.5 ml NH₄OH (25 mM) into the glass tubes. After vortexing the glass tubes for 1 min, the sample was allowed to stand for 5 min. Next, collect the lower CH₂Cl₂ layer and blow dry with nitrogen. A modified Bligh-Dyer method was applied for extraction and desalination of the derived DAGs. Finally, 75 μl mobile phase (50% Isopropanol-50% Acetonitrile) was utilized to re-suspend extractive product for MS analysis. The above charge derivatization procedure was practiced to directly modify DAGs in biological sample.

For the preparation of biological samples, liquid-liquid extraction was performed to remove proteins and concentrate lipids. At first, 400 μl of ice-cold CH₂Cl₂: MeOH (2:1, v/v) and 10 μl of serum were introduced into a glass tube and the mixture was swirled for 2 min. Subsequently, 100 μl of deionized water was introduced and then the CH₂Cl₂ layer was collected. An amount of CH₂Cl₂ extract was shifted to a glass vial and the sample was evaporated until dry under N₂ airflow. All dried samples were derivatized according to the derivatization principle described above.

An AB Sciex Triple Time-of-Flight 5,600 system in positive ion electrospray ionization was utilized for characterization of tagged DAGs. The quantification of DAGs derivatives was performed on Qsight LX50 UHPLC together with a Qsight 220 MS system (PerkinElmer, Waltham, MA, United States). The temperature of the UHPLC autosampler was maintained at 4°C and the sample injection volume was adjusted to 2 μl. Chromatography separation was performed on a Waters Acquity BEH C18 column (1.7 μm, 2.1 mm × 150 mm) at 55°C. The mobile phase consisted of solvent A (40% water-60% MeOH contains 5 mM ammonium acetate) and solvent B (90% isopropanol-10% acetonitrile with 0.1% formic acid). The mobile phase was delivered at a flow rate of 0.2 ml/min. The gradient program used was as follows: 10% B in 0–2 min, 10% B to 30% in 2–5 min, 30% B to 35% in 5–8 min, 35% B to 37% in 8–10 min, 37% B to 37% in 10–12 min, 38% B to 38% in 12–14 min, 38% B to 45% in 14–18 min, 45% B to 65% in 18–20 min, 65% B to 80% in 20–25 min, 80% B to 95% in 25–26 min, 95% B in 26–27 min, 10% B in 27–30 min. The operating parameters of the mass spectrometer were setup as below: dry gas value is 100, atomized gas value is 180, electrospray voltage is 5,500 V, and source temperature is 400°C.

Unlike polar lipid, DAGs possess relatively lower ionization efficiency under ESI-MS, because there is no fixed ionizable group in the structure. Hence, it is advisable to insert a fixed charge-tag to DAGs for enhancing ESI ionization efficiency. Three commercially available N-methylcarboxylic acid derivative reagents, comprising DMG, DMA and N, N, N-trimethylglycine (TMG), were utilized to establish the derivatization procedure (structures were shown in Figure 1A). The foundations for selecting charge derivatization reagents were functional molecules with feasible solubility and easily ionizable under ESI conditions.

Figure 1B shows a route map of derivatization (steglich esterification) reaction. In this reaction, the carboxylic acid (e.g., DMG) is firstly activated by carbodiimide. Traditionally, EDC (a carbodiimide reagent) is used to produce an O-acylisourea intermediate (Jordan et al., 2021). DMAP reacts with the alcohol on the O-acylisourea intermediate in a competitive nucleophilic manner to produce an acyl pyridinium intermediate (Xie et al., 2019). This process minimizes the generation of unproductive acyl migration side products, due to the high reaction efficiency of DMAP. The acyl pyridinium intermediate quickly reacts with hydroxyl of DAGs to reproduce DMAP and generate esterification products.

Using DAG 16:0/18:1/0 as a model compound to verify the designed strategy. The tagged DAGs generated from three individual derivative reagents after reaching equilibrium. Figures 2A–C depict the ESI-MS spectra of DAG 16:0/18:1/0 (1 µg/ml) after the derivatization reactions. For the above derivative reagents, a single reaction product with high ionization intensity was observed, i.e., m/z 680.5824 ([DAG + DMG]⁺), m/z 694.5979 ([DAG + DMA]⁺), m/z 694.5986 ([DAG + TMG]⁺) in positive ion mode. Furthermore, all charge labeling have greatly enhanced the ionization of DAG in ESI-MS. In the case of DMG, the ionic intensity of the derived DAGs was superior to the formation of sodium adducts (Supplementary Figure S1).

The data in Figures 2D–F depict MS² beam-type CID of three individual charge labeled DAGs cations in the positive ion mode. The stepwise cleavage of the DMG labeled DAGs produced a fragmentation peak at m/z 577.5176, resulting in a specific neutrality loss of 103 Da. The characteristic fragmentation signals of m/z 339.2887 and m/z 313.2697 come from the progressive loss of the fatty acyl groups of C18:1 and C16:0, respectively. Figure 2D displays the observed cleavage fragmentation information of DMG-labeled DAG 16:0/18:1/0, whereas the potential cleavage pattern is suggested in Supplementary Figure S2. In spite of the fact that all three charge labels provided better ionization at CID, DMG and DMA were selected to further develop quantitative approaches. The reason for this is that DMG- and DMA-labeled DAGs not only have characteristic fatty acyl information, the neutral loss related to the label favors the exploitation of neutral loss scan (NLS) for the determination and quantification of DAGs in biological samples. Moreover, DMA tag also introduces a stable 117 Da neutral loss. The NLS of DMA is perfectly suited for rectifying varies during detection. Instead, different from beam-type CID of TMG derivatized cation, 118 Da loss (-TMG) from derivatized cation was not observed.

For DAGs with various chain length fatty acyl groups and different esterification positions, we wondered whether charge labeling would occur. The data in Figures 3A–F depict the derivatization spectra of DAG 16:0/16:0/0, DAG 16:0/0/16:0, DAG 18:0/18:0/0, DAG 18:0/0/18:0, DAG 14:0/14:0/0 and DAG 10:0/10:0/0, which were individually derived and performed at the same reaction conditions. The production of DMG tagging products reached a steady state during MS analysis. The fragmentation signal at m/z 557.50 is derived from DAG 16:0/16:0/0 and DAG 16:0/0/16:0, resulting from the sequential loss of the DMG tag (Figures 3A,B). Similarly, the peak at m/z 613.57 is derived from DAG 18:0/18:0/0 and DAG 18:0/0/18:0 (Figures 3C,D). Although the existence of different positional isomers in DAGs, only a single DMG-labeled product was detected. For different fatty acyl chain lengths, both DAG 14:0/14:0/0 (Figure 3E) and DAG 10:0/10:0 (Figure 3F) had DMG-labeled fragments at m/z 501.44 and m/z 445.44, respectively. The above results demonstrate that the presentation of a charge label to DAG species produces a similar ionic response in ESI-MS, no matter the length or positional heterogeneity of the fatty acyl chain. However, DMG may be unable to distinguish positional isomers in the DAG. This is because none any significant and stable diagnostic fragments were observed during ESI-MS analysis.

On the basis of derivatization of DAGs reference standard, we proceeded to explore the applicability of charge derivatization in biological samples. Supplementary Figure S3 summarizes that fragments resulting from CID of DAG 16:0/16:0/0, DAG 18:0/18:0/0 and DAG 18:1/18:1/0 in serum samples. The DAG standards were simultaneously charge labeled by derivatization reaction which resulted in an enhancement of the MS signal. It is agreed with the above findings that polar head group of lipid has a primary effect on the dissociation of DAGs via ESI-MS. Different DGAs species modified with polar charge labels share basically the similar ionization efficiency at low concentrations. The Steglich esterification reaction parameters have been optimized in order to develop a feasible derivatization method that satisfies the demands of less hazardous solvent systems, shorter reaction time and mild reaction temperatures for biological samples. Previous efforts to use common green solvents (e.g., ethyl acetate and acetone) failed to produce a product (Churruca et al., 2012). The reaction is usually carried out using anhydrous chloride solvent in the presence of nitrogen. However, these solvents are undesirable because they can be hazardous to environment and human health. Hence, a combined solvent system, included acetonitrile and dichloromethane (v/v, 1:1), has been tried for derivatization.

Prior researches have pointed out that the initial nucleophilic reaction participated by DMAP is a speed-limiting step for this reaction (Figure 1). To facilitate the reaction as much as possible, the derivatization reaction was carried out with excessive amounts of DMAP (0.5 mM). Because hydroxyl groups are widely present in biological samples, a large amount of charge tags are in favor of complete derivatization. Different amount of charge tags (DMG, range from 1.5 μM to 15,000 μM) were tested to obtain optimal derivatization using DAG 16:0/16:0/0, DAG 18:0/18:0/0, DAG 16:0/18:1/0 (1.5 μM) as mixed standard. Excess DMG promoted the formation of charged derivatives. It was found that when DMG was 1,500 μM, the most DAG derivative products were generated. Interestingly, more DMG (15,000 μM) did not induce higher labeled product, suggesting DAGs have been completely derivatized (shown in Figure 4A).

Different reaction time periods (30 min, 45 min, 60 min, 90 min, and 120 min) and temperatures (45°C, 55°C, 65°C and 80°C) were tested to obtain optimal derivatization. The conditions were selected to be examined in order to keep the solvent system under boiling point and avoid using a reflux condenser, thereby ultimately obtaining a better reaction rate. Based on the MS intensities achieved from the labeled DAGs standards, the intensities of the DAGs at 60 min were higher than those at 90 min and 120 min (Figure 4B). On this basis, the intensity was higher at 80°C than at 65°C. However, we finally opted for 65°C and 60 min based on principles of avoiding unanticipated side reactions in biological system under gentle reaction condition (Figure 4C).

The properties of the 103 Da tag NLS were evaluated in terms of quantification of DAGs under optimal reaction conditions. The mass spectrometry quantitative parameters were optimized in Supplementary Table S1. Multiple reaction monitoring beam-type CID of ^DMG-DAGs and ^DMA-DAGs were shown in Figure 5. These ion pair are m/z 710→m/z 607 for ^DMG-DAG 18:0/18:0/0, m/z 724→m/z 607 for ^DMA-DAG 18:0/18:0/0, m/z 680→m/z 577 for ^DMG-DAG 16:0/18:1/0, m/z 694→m/z 577.1 for ^DMA-DAG 16:0/18:1/0, m/z 654→m/z 551 for ^DMG-DAG 16:0/16:0/0, m/z 668→m/z 551 for ^DMA-DAG 16:0/16:0/0, m/z706→m/z 603 for ^DMG-DAG18:0/18:0/0, m/z720→m/z 603 for ^DMA-DAG 18:0/18:0/0. The chromatograms of DAGs in UPLC-MS analysis were showed in Figure 6. Both ^DMG-DAGs and ^DMA-DAGs possess stable and reproducible characteristic neutral loss fragment.

The charge derivatization introduced a sensitive labeling for quantitative analysis, however, the charge tag may not be able to discriminate the sn-1,2 DAGs from the sn-1,3 DAGs. A gradient elution procedure of mobile phase was performed to separate different isomers of DAGs (in Supplementary Figure S4). As shown in Supplementary Figure S5, the isomers of DAGs were well-separated at retention times of 20.9 min and 21.6 min for ^DMA-DAG 18:0/18:0/0 and ^DMA-DAG 18:0/0/18:0, 20.3 min and 22.9 min for ^DMG-DAG 18:0/18:0/0 and ^DMG-DAG 18:0/0/18:0, 14.9 min and 15.4 min for ^DMG-DAG 16:0/18:1/0 and ^DMG-DAG 16:0/0/18:1, respectively.

Calibration standards, quality control (QC) samples and surrogate matrices excluding DAGs were prepared for the validation of MS quantification method. A series of standard solutions at ten levels of concentration (0.1, 1.0, 5.0, 10, 25, 50, 100, 250, 500 ng/ml) were prepared by dilution of DAG stock solutions with dichloromethane. QC samples were obtained by mixing each biological extract in equal volumes. A modified Bligh and Dyer procedure was applied for preparing surrogate matrices without DAGs. At first, 10 μl of serum was shifted into a 10 ml glass pipe, followed by the addition of 400 μl of cold CH₂Cl₂/MeOH mixed solvent (2:1, v/v). The mixture was swirled for 2 min at 4°C and H₂O (100 μl) was added. The obtained solvent was homogenized and then allowed to stand for 5 min. Finally, the upper layer of organic phase (250 μl) was collected and dried by evaporation to perform derivatization.

Linearity and sensitivity: Solvent calibration curve and surrogate matrix calibration curve were established by adding calibration standards in pure solvent and surrogate matrices without DAGs (in Table 1). Linearity was evaluated by correlation coefficient (r ²). The lower limit of detection (LLOD) and lower limit of quantification (LLOQ) in MS detection were used to assess the sensitivity of this method. The LLOQ should satisfy a signal-to-noise ratio (S/N) greater than 10 and a relative standard deviation (RSD) of the assay results less than or equal to 20% (n = 3). The lower limit of detection (LLOD) requires a signal-to-noise ratio greater than three and an RSD less than or equal to 20% (n = 3).

The availability of alternative matrix curves was evaluated by comparing the slope of the alternative matrix curve and the solvent curve. The concentrations of QCs samples were calculated from both types of curves and the deviations between the profiles were evaluated. After parallelism was confirmed, three standard solutions at low, medium and high concentrations were selected for subsequent validation analysis based on the linear range of the alternative matrix curves.

The matrix effect (ME) is measured as the ratio of three concentrations of standards (neat solution samples) to standards added to the alternative matrix (matrix samples). The ME was calculated by the following equation, where R represents the detection response of the DAGs. M E   ( % ) = R m a t r i x × I S n e a t   s o l u t i o n R n e a t   s o l u t i o n × I S m a t r i x

Extraction recovery efficiency was measured from the concentration of standards before (A) and after extraction (B), where the concentration was calculated from the alternative matrix curve (n = 3, three copies of each concentration were prepared). Recovery (%) = (A/B) × 100%. Matrix effect and recovery were listed in Supplementary Table S2.

Repeatability and stability were examined by analyzing six replicates of the QC samples in parallel. Precision was investigated by intra-day and inter-day RSD, which were measured in 1 day and on three consecutive days (n = 6). The storage stability of the autosampler (4°C) for the derivatized DAG was inspected by testing QC samples over a 48-h period. In addition, the stability of IS at 4°C for 48 h was also assessed. As shown in Supplementary Table S3, the RSD values for all groups were less than 20%.

Several researches have indicated an association with enhanced levels of DAGs in acute pancreatitis. DAGs activate PKC system, which lead to the phosphorylation of cellular calcium channels. Increased intracellular calcium promotes pancreatic cell apoptosis. Recently, Shahram et al. have shed light on potential impact of resveratrol on acute pancreatitis (Agah et al., 2021). Different metabolic pathways involved in triacylglycerol metabolism have been shown to be targets for resveratrol (Churruca et al., 2012). Hence, whether resveratrol can regulate the metabolism of DAGs was explored.

Acute pancreatitis models were invited by two intraperitoneal injections of L-arginine (i.p.). The changes in the histological structure of the pancreas and lungs from the acute pancreatitis model animals were depicted by H&E staining (Supplementary Figure S6). No alterations in the histological structure of the pancreas were observed in normal control mice. Mice with L-arginine-induced acute pancreatitis exhibited obvious alterations in pancreatic histological structures, such as acinar cell atrophy, necrosis and inflammatory cell infiltration, whereas resveratrol treatment significantly diminished pancreatic atrophy and inflammation, especially in the resveratrol high-dose group. In addition, L-arginine exposure has enhanced alveolar thickening and the amount of inflammatory cell aggregation. Whereas, treatment with resveratrol significantly restored the symptoms of alveolar abnormalities. Therefore, the results suggest a potential protective effect of resveratrol on L-arginine-induced acute pancreatitis and related lung injury.

Additionally, we were interested in checking whether the DMG label and 103 Da NLS were available for the analysis of DAGs in the serum of mice with acute pancreatitis. Figure 7A summarizes that the 11 DAG species were quantified in serums from blank control mice, they were DAG 18:0/18:0/0, DAG 18:0/0/18:0, DAG 16:0/18:1/0, DAG 16:0/0/18:1, DAG 18:1/18:1/0, DAG 18:1/0/18:1, DAG 16:0/16:0/0, DAG 14:0/14:0/0, DAG 12:0/12:0/0, DAG 10:0/10:0/0, and DAG 8:0/8:0/0. The chromatograms of DAGs were showed in Figure 6 and the concentrations of DAGs range from 4.14 ng/ml to 145 ng/ml. The DAGs sn-isomers, including DAG 18:0/18:0/0, DAG 18:0/0/18:0, DAG 16:0/18:1/0, DAG 16:0/0/18:1, DAG 18:1/18:1/0, and DAG 18:1/0/18:1 were separately detected. The intensity changes of DAGs from four sets of samples were summarized in Figures 7B,C (normal controls, models, low dose, high dose, N = 6). The content of DAG 18:0/18:0/0, DAG 18:0/0/18:0, DAG 16:0/16:0/0, and DAG 16:0/18:1/0 increased by 3.4 ± 2.1, 3.5 ± 2.0, 3.6 ± 1.5, and 2.3 ± 1.8 times in models as compared to the control, respectively. After resveratrol intervention, the levels of DAGs, including sn-1,2 and sn-1,3-DAGs, both decreased. To investigate the effect of resveratrol on positional isomers of DAGs, the ratio of sn-1,2-DAGs to sn-1,3-DAGs in different groups was analyzed. Figure 7D depict the ratios of sn-1,2-DAGs to sn-1,3-DAGs in four groups. Sn-1,2-DAGs, including DAG 16:0/18:1/0, DAG 18:0/18:0/0, and DAG 18:1/18:1/0, showed an elevated trend in the pancreatitis model. In the resveratrol treatment group, the ratios of sn-1,2-DAGs to sn-1,3-DAGs was decreased. This data suggests that resveratrol may have more influence on sn-1,2-DAGs production when regulating lipolysis. Previous studies have also revealed that resveratrol may regulate certain enzymes involved in the production of DAGs, such as phospholipase C or adipose triglyceride lipase (Yang et al., 2008; Churruca et al., 2012). In future, further studies require more sample numbers of acute pancreatitis. Based on the characteristics of the DMG charge derivatization method for amplifying MS signals, we will proceed to deeply explore the implications of structural alterations of DAGs on diseases, thereby achieving a detailed molecular-level characterization of DAGs and providing precise lipid structure information for disease research.