Introduction

Sea dragon (Syngnathus) belongs to a group of small marine fishes in the family of Syngnathidae, which is both an aquatic product and an animal-derived traditional Chinese medicine of great medicinal value, and is often applied in the preparation of medicinal wines and medicinal diets in food1,2. As an aquatic economic animal, sea dragon is known as “animal ginseng”. As a natural nutritional treasure trove, sea dragon as a kind of marine animal is rich in high-quality proteins, minerals, polyunsaturated fatty acids (PUFAs) (especially n-3 PUFAs and n-6 PUFAs) and other biologically active substances. Several studies have shown that its extracts exhibit potential activities such as anti-inflammatory3, inhibition of tumour cell proliferation4, and alleviation of exercise fatigue5. Among them, n-3 PUFAs have been shown to improve lipid metabolism disorders by lowering serum triglyceride levels6. In traditional medicine, sea dragon is frequently employed as an adjunctive therapy for symptoms associated with kidney yang deficiency, such as sexual dysfunction. However, the precise mechanisms underlying its therapeutic effects remain to be elucidated3. Currently, sea dragons are mainly derived from captive breeding due to the scarcity of its wild resources7, and their main species are Solenognathus hardwickii (SH), Syngnathus acus Linnaeus (SL), Syngnathoides biaculeatus (SB), Hexagonal Solenognathus hardwickii (HSH), Quadrangular Syngnathoides biaculeatus (QSB), and Syngnathus exilis (SE), etc. However, as an important marine resource, the phenotypes of sea dragons are very similar among species, which makes them difficult to be distinguished8. Moreover, the material composition of sea dragons and the differences in composition between species are still unknown today.

With the rapid development of modern identification technology, omics technology has shown amazingly high precision and high throughput in biomarker discovery, clinical research and toxicology compared with traditional identification technology9. Common omics techniques include genomics, transcriptomics, proteomics, metabolomics and lipidomics, etc., which have become effective tools for identification of food components, clarification of efficacy, and excavation of mechanisms10. Both metabolomics and lipidomics are omics techniques for studying low molecular compounds (typically <2000 Da)11, including sugars, amino acids, and fatty acids, etc., which provide powerful tools for quality control, nutritional value assessment, food safety monitoring, and personalized nutrition of modern foods12,13. Nowadays, untargeted metabolomics has been widely used in food safety control and quality analysis mainly by focusing on hydrophilic small molecule metabolites14. Lipidomics is a branch of metabolomics, an independent grouping based on lipid complexity that aims to identify and quantify large numbers of lipids simultaneously15. The integration of metabolomics and lipidomics is becoming an emerging research strategy aimed at providing a more complete fingerprint profile16, which has been applied in the food field to authentication of camellia oil17, pork18, fresh tea leaves19, minced pork20, and fermented goat milk21. However, there are still fewer relevant reports on the composition of sea dragons. Among them, comprehensive small molecule metabolite and lipid molecule composition related studies have not been reported yet. Consequently, strategies for integrating untargeted metabolomics and lipidomics have not been used to analyze the construction of sea dragon fingerprints and species identification.

Sepsis, as a systemic inflammatory response syndrome triggered by infection, is often accompanied by an imbalance in immune homeostasis, which may lead to severe organ damage, hypotension, multi-organ failure, and even death22. Food safety and nutritional interventions are one of the important directions in the prevention and treatment of sepsis in the field of food science. Active ingredients in foods, such as polyphenols, antioxidants, fatty acids and amino acids, had the potential to modulate immune responses, reduce inflammation and improve intestinal barrier function, which was pivotal in the prevention and treatment of sepsis23. Network pharmacology as a systems biology approach to the study of functional factors in food can reveal the multi-target mechanism of action of active ingredients in food in sepsis by constructing and analyzing food ingredient-target-disease networks. This approach enables the identification of potential therapeutic targets of food ingredients for sepsis, the optimization of food formulations, and the development of functional foods with anti-inflammatory and immunomodulatory functions24,25. Thus, network pharmacology can contribute to unraveling the complexity of sepsis, providing new strategies for sepsis prevention and treatment, and advancing the development of personalized nutrition and precision medicine.

In this study, the comprehensive fingerprinting of six species of sea dragons was performed using UPLC-Q exactive orbitrap MS (UPLC-QEO/MS)-based untargeted metabolomics and lipidomics, to screen the major differential metabolites and lipids. Network pharmacology and molecular docking were also utilized to establish the correlation between differential metabolites and lipid profiles and their anti-sepsis activity. The research findings will provide theoretical support for further revealing its intrinsic nutritional value and anti-sepsis potential for the development of new food and health products.

Results

Comprehensive analysis of metabolites in sea dragons

For metabolomics, the total ion flow chromatograms (TIC) in positive and negative ion modes are shown in Fig. S1, respectively. As shown in Figs. 1a and 2b, a total of 2264 metabolites were detected in all samples, including 1266 in positive ion mode and 998 in negative ion mode. To further discuss the metabolic profiles in sea dragons, the peak areas of the detected features were normalised and statistically analysed. These metabolites were categorized into 18 classes, mainly including fatty acids (37.97%), lipids and lipid-like molecules (25.77%), organoheterocyclic compounds (11.26%), and organic nitrogen compounds (5.86%). Overall, fatty acids (108), lipids and lipid-like molecules (259), and organoheterocyclic compounds (616) were the most abundant metabolites. Metabolic profiling of the six sea dragon species (Solenognathus hardwickii (SH), Syngnathus acus Linnaeus (SL), Syngnathoides biaculeatus (SB), Hexagonal Solenognathus hardwickii (HSH), Quadrangular Syngnathoides biaculeatus (QSB), and Syngnathus exilis (SE)) revealed that fatty acids and lipid-like molecules constituted the predominant compound classes (A-R: Alkaloids - Others) across all species (Fig. 1c). Notably, these components exhibited particularly high proportions in three species: SE, QSB, and HSH. The clustered heatmap (Fig. 1d), generated by normalizing metabolite peak areas, illustrated distinct expression patterns among species, which red indicated that the substance is expressed with a high relative content in the group, and blue indicated that the substance is expressed with a low relative content in the group. Specific metabolite categories dominated in each species: alkaloids and derivatives in SH, organosulfur compounds in SB, organoheterocyclic compounds in HSH, amino acids and peptides in SE, Organic nitrogen compounds in SL, and terpenoids in QSB. Collectively, these findings demonstrated that the six species of sea dragons showed significant interspecies divergence in metabolic profiles.

Fig. 1: Metabolomic analysis of six sea dragon species.
figure 1

a The number of metabolites in two ionization modes; b Metabolite classes and proportions; c Chordal plots of metabolites of the six sea dragon species; d The heatmap of metabolite species of the six sea dragon species. Note: The letters in Fig. 1c correspond to the compounds in the legend, including A: Alkaloids, B: Alkaloids and derivatives, C: Amino acids and Peptides, D: Benzenoids, E: Carbohydrates, F: Fatty acids, G: Lipids and lipid-like molecules, H: Nucleosides, nucleotides, and analogues, I: Organic acids and derivatives, J: Organic nitrogen compounds, K: Organic oxygen compounds, L: Organoheterocyclic compounds, M: Organosulfur compounds, N: Phenylpropanoids and polyketides, O: Polyketides, P: Shikimates and Phenylpropanoids, Q: Terpenoids, R: Others.

Full size image
Fig. 2: Analysis of differential metabolites in positive ion mode.
figure 2

a Number of differential metabolites in positive and negative ion modes; b Top 10 representative differential metabolites with VIP values; c The heatmap of 30 differential metabolites in positive ion mode.

Full size image

Comparison of metabolic differences between sea dragons

To further analyze the differences in metabolite levels among the six sea dragon species, significant metabolites were screened. Principal component analysis (PCA) and partial least squares discriminant analysis (PLS-DA) was used to analyze the metabolites of sea dragons, and the results were shown in Fig. S2. The PCA results revealed partial overlap between the metabolite profiles of samples SH and SL in the positive ion mode (PC1 and PC2 accounted for 70.3% and 12.5% of the total variance, respectively), indicating substantial chemical similarity. In the negative ion mode, projections of HSH and QSB clustered more closely (PC1 and PC2 explained 56.4% and 23.4% of the variance, respectively), suggesting convergence in their polar metabolite distributions. A partial least squares discriminant analysis (PLS-DA) model was further constructed based on the metabolite data of different sea dragons obtained in the two ionization modes. Variable importance for the projection (VIP) value > 1 was considered to be a differentiating component26. Subsequently, the PLS-DA model was validated using leave-one-out cross validation (LOOCV) and 1000 permutations test, and the results (p < 0.001) proved that the model was well predicted and not overfitted (Fig. S3)27. This result indicated that the PLS-DA model has good classification efficacy and could accurately identify the six species of sea dragons.

A total of 47 differential metabolites were screened based on the VIP value, of which nine classes of metabolites were screened in the positive ion model; six classes of metabolites were screened in the negative ion model (Fig. 2a). Top ten differential metabolites identified in the positive ion mode were shown in Fig. 2b, including Trimethylamine N-oxide, 4-Fluoro-7-azaindole, LPC(16:0), Glycerophosphocholine, 1-O-Hexadecyl-sn-glycero-3-phosphocholine (LPC(O-16:0/0:0)), Betaine, PC(P-16:0/0:0), 4-Chloro-5,6,7,8-tetrahydroquinazolin-2-amine, LPC(18:1), and 6-Fluoroindole. Bubble plots demonstrated the differences in metabolites and intensities between the two modes for six samples, including 30 differential metabolites in the positive ion mode (Fig. 2c, Table S1) and 17 metabolites in the negative ion mode (Fig. S4, Table S2). It is worth noting that the most abundant species of differential metabolites in the positive ion mode was lipids and lipid-like molecules, such as LPC(16:0), glycerophosphocholine, LPC(O-16:0/0:0), PC(P-16:0/0:0), and LPC(18:1). Top ten differential metabolites in the negative ion mode (Fig. S4) were palmitic acid, 10-thiastearic acid, nitrofurazone, docosahexaenoic acid (DHA), stearic acid, trans-vaccenic acid, lactate, oleic acid (OA), N-(2,3-dihydro-1H-inden-5-yl) methanesulfonamide, and cis-9-palmitoleic acid. Among them, the most abundant differential metabolite species in the negative ion mode was fatty acids. The results indicated that the 47 features were effective in distinguishing six species of sea dragons, among which lipids were expected to be able to serve as biomarkers for the authentication of sea dragons.

In-depth analysis of the primary metabolites in sea dragons - lipids

Since the extraction and detection methods employed in traditional metabolomics are not fully suitable for lipid molecules, traditional metabolomics has obvious limitations in detecting lipid molecules. In contrast, lipidomics, an analytical technique specifically designed for lipid molecule detection, enables the comprehensive characterization of metabolites. To deeply explore the lipidomic differences among six species of sea dragons, this study adopted a systematic lipidomic strategy. This strategy was based on untargeted metabolomics and further integrated with lipidomic technology. It allows for the characterization of lipid composition at the molecular level and addresses key issues in previous metabolomics - based lipid analyses, such as the lack of methodological specificity and limited lipid coverage. As a result, a robust and comprehensive lipid analysis platform was established. Using this platform, we systematically analyzed the molecular diversity of lipids in the six sea dragon species, providing high-quality data support for the discovery of functional lipids14,15,16.

The lipid composition of the sea dragon was comprehensively detected in both positive and negative ion modes, with TIC plots shown in Fig. 3a. The lipids in the sea dragon consisted of six major classes and 44 subclasses, and a total of 2078 lipid molecules were detected, including 185 fatty acyls (FA), 9 saccharolipids (SL), 9 sterol lipids (ST), 596 sphingolipids (SP), 737 glycerolipids (GL) and 492 glycerophospholipids (GP). The percentage of each subclass of lipids to the total lipid content was shown in Fig. 3b, where subclass of lipids with less than 0.1% in SP were combined in the others group. Among these lipid classifications, GL (52.28%) were the most abundant, followed by GP (21.78%), SP (16.53%), and FA (8.21%). The content of ST (0.89%) and SL (0.31%) were both less than 1%. In positive ion mode, a total of 1090 lipid molecules were detected from 5 major classes and 21 subclasses; 988 lipid molecules from 5 major classes and 36 subclasses were detected in the negative ion mode (Fig. 3c).

Fig. 3: Lipidomic analysis of six sea dragon species.
figure 3

a TIC plots in positive ion mode (up) and negative ion (down); b Classes and proportions of lipids; c Number of lipids in the two ionization modes; d Sankey plots of lipid species of six sea dragon species.

Full size image

To further explore the differences in the lipid composition of the six sea dragon species, the Sankey diagram illustrated the distribution of different lipid classes in sea dragon (Fig. 3d). Six major classes of lipids (GL, GP, SPs, FA, ST, and SL) were detected in all sea dragon species, which demonstrated their richness and uniqueness in terms of lipids. The results of lipidomics indicated that there were significant differences in the most abundant lipid subclasses, namely TAG, DAG, and PC, among the six species of sea dragons. This finding was consistent with our previous research, which showed that a relatively large number of components such as TAG, DAG, and PC were detected in seafood28.

Analysis of differential lipids between sea dragons at molecular level

PCA and PLS-DA results showed significant differences in all six samples, which revealed that there is variability in the lipid composition in the six species of sea dragons, confirming that the lipid composition would be useful in identifying the species of sea dragons (Fig. S5). The results of PCA showed that the cumulative variance contribution rates of the first two principal components exceeded 60% for both modes (PC1 = 47.3%, PC2 = 26.7% in positive mode; PC1 = 44.6%, PC2 = 21.1% in negative mode), indicating that the datasets contained substantial biological variation information (Fig. S5a, d). PLS-DA further validated this discrepancy (Fig. S5b, e). LOOCV and a 1000 permutation test approved that the PLS-DA model was highly predictive without overfitting (Fig. S6).

To improve the reliability of the data, VIP values from the PLS-DA model were used to identify representative lipids as markers of species separation. Lipids with VIP > 1 was selected as differential lipids in both the positive and negative ion models (Table S3, Table S4). 76 differential lipids were screened in the positive ion mode, including FA, ST, SP, GL, and GP (Fig. S7). 62 differential lipids were screened in the negative ion mode, mainly including FA, SP, and GP (Fig. S8). Notably, PLS-DA results indicated that the positive ion mode was mainly enriched for neutral lipids (such as triacylglycerol (TAG), phosphatidylcholine(PC)), while the negative ion mode was more sensitive to acidic lipids (such as Free fatty acid (FA), Fatty acid ester of hydroxyl fatty acid (FAHFA)), and that the two modes were complementary, highlighting the efficacy of two-dimensional analysis in resolving the lipidomic features of sea dragon. TAG and FA were the dominant components in the above differential lipids, including TAG (16:1/18:1/22:5), TAG (18:1/20:4/20:5), TAG (12:0/12:0/22:6), TAG (16:0/18:2/20:4), FA (18:3), FAHFA (20:4/20:3) and FA (18:2), etc. It contains a variety of unsaturated fatty acids such as OA (18:1), palmitoleic acid (PA, 16:1), α-Linolenic acid (ALA, 18:3), eicosapentaenoic acid (EPA, 20:5), DHA (22:6), linoleic acid (LA, 18:2) and arachidonic acid (ARA, 20:4). This study found that sea dragon, as a typical representative of a marine fish enriched with functional lipids, demonstrated its richness and functionality at the lipid level.

Functional PUFAs - a feature of sea dragon lipids

By summarizing the differential components of metabolomics and lipidomics, we were surprised to find a total of 25 PUFAs, including 13 n-3 PUFAs and 12 n-6 PUFAs, and FAHFA (20:5/18:2) existed in both series, and the specific molecular information was summarized in Table S5. Six species of sea dragons showed significant differences in n-3 PUFAs (Fig. 4a) and n-6 PUFAs (Fig. 4b). Therefore, we focused on the distribution of n-3 and n-6 PUFAs in the six species of sea dragon. As shown in Fig. 4c, the peak intensities of these five polyunsaturated fatty acids showed significant differences, with FAHFA (18:2/20:4), FAHFA (20:5/18:2), and TAG (16:0/18:2/20:4) containing multiple long-chain PUFA chains at the same time. The peak intensities of ALA, EPA, DHA, LA and ARA were shown in Fig. 4d, where QSB contained a high percentage of each type PUFAs, 3.59%, 5.21%, 5.10%, 4.48%, and 4.98%, respectively. This may be attributed to the dietary characteristics of the sea dragon, which as a marine fish often feeds on small algae, marine fish and shrimp, and is therefore rich in functional ingredients such as EPA and DHA29.

Fig. 4: Analysis of PUFAs in all differential metabolites and lipids in six species of sea dragon.
figure 4

a The heatmap of n-3 PUFA; b The heatmap of n-6 PUFAs; c Total peak area of ALA, EPA, DHA, LA, and AA; d ALA, EPA, DHA, LA, and AA as a percentage of total lipid peak area; e Differential PUFAs in QSB as a percentage of each content of ALA, EPA, DHA, LA, and AA.

Full size image

Based on this, QSB was considered to possibly possess better bioactive functions. The relative abundances of the five fatty acids in QSB were calculated using the peak area method, as shown in Fig. 4e. The highest abundances of ALA were found in TAG (18:1/18:3/19:2) and FA (18:3), which accounted for 10.15% and 8.03% of the total ALA, respectively. The highest abundances of EPA were found in TAG (18:1 /20:4/20:5) and DAG (16:0/20:5), which accounted for 3.30% and 3.25% of the total EPA, respectively. The abundance of LPC (22:6) in DHA was as high as 5.78%. LA and ARA were mainly concentrated in TAGs and FAHFAs, which had a total abundance of up to 25% or more. It is noteworthy that PUFAs in QSB were mainly composed of TAGs and FAHFAs. TAGs were detected only in the positive ion mode, which had a total of 76 differentiated lipids, of which TAGs accounted for 34. FAHFAs were detected only in the negative ion mode, and 13 out of the 62 differentiated lipids in the negative ion mode were FAHFAs. Therefore, TAG and FAHFA were important for effective differentiation among the six species of sea dragons, such as TAG(18:1/18:3/19:2), TAG(18:1/20:4/20:5), TAG(16:0/18:2/20:4), FAHFA(20:4/20:3), and FAHFA(18:2/20:4).

Network pharmacology and molecular docking analysis - anti-sepsis

The SwissTargetPrediction database was utilized to finally obtain 572 relevant targets for 34 active components from sea dragon (Fig. 5a). Briefly, 1933 genes were screened in five databases (GeneCards, TTD, OMIM, Drugbank, and GAD), using “Sepsis” as the keyword and the median correlation score as the threshold. The above genes were deduplicated and corrected using the Uniprot database, resulting in a total of 1628 disease targets. Venn diagram analysis comparing the predicted targets of sea dragon’s differential components with known sepsis-related targets identified 167 shared potential anti-sepsis targets (Fig. 5b). Using the STRING database and Cytoscape 3.9.1 for interaction network analysis, we screened 56 core targets and 374 protein-protein interaction (PPI) edges. The results highlighted IL-6, STAT3, JAK, AKT1, BCL2, MAPK1, and MAPK14 as key targets mediating the anti-sepsis effects of sea dragon (Fig. 5c). Functional enrichment analysis of Gene Ontology (GO) terms and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways was conducted on the core target proteins through the Metascape database, with the results visualized in Fig. 5d. Briefly, GO enriched a total of 211 terms, and the top ten terms were visualized based on the p-value. Among them, biological process (BP) totaled 57 terms, involving biological processes such aspositive regulation of cell migration, cellular response to cytokine stimulus, and regulation of MAPK cascade; cellular component (CC) totaled 64 terms, including side of membrane and receptor complex; there were 90 molecular function (MF) terms, mainly enriched in kinase binding, protein kinase activity, etc. The results of KEGG pathway enrichment analysis yielded 168 terms for the differentially active components, and the main pathways involved chemokine signaling pathway, NF-κB signaling pathway, JAK-STAT signaling pathway, inflammatory mediator regulation of TRP channels, and p53 signaling pathway. 25 components, 56 targets and 168 pathways were visualized using Cytoscape 3.9.1 (Fig. 5e). The core anti-sepsis components of the differential components in sea dragons were summarized in Table S6, including palmitic acid (Comp 31), DHA (Comp 34) and AA (Comp 43), which have all been proved to be effective in ameliorating sepsis30,31.

Fig. 5: Network pharmacologic analysis of differential components in sea dragons against sepsis.
figure 5

a Core targets of differential composition in six sea dragon species; b Venn diagramt of core targets of differential components in sea dragon species and disease targets in sepsis, which total 167 intersecting targets; c PPI network of the targets of differential components in six sea dragon species against sepsis, which total 56 core targets. Larger and darker circles represent higher degree values and stronger significance; d GO functional enrichment analysis (top 10) and KEGG signaling pathway enrichment analysis (top 20) of differential components in sea dragon species against sepsis; e “sea dragon-component-target-pathway” interaction network of differential components in sea dragon species against sepsis. f Molecular docking results of differential components in six sea dragon species with core anti-sepsis targets.

Full size image

To further elucidate the pharmacodynamic mechanism of sea dragon in sepsis, the 56 core targets and their corresponding 25 key components were validated by molecular docking technique and the results of stable binding (binding energy < −5 kJ·moL−1) were summarized in Table S7. It was found that the binding energies of the lipid molecules (especially DHA and AA) with the targets of JAK2, BCL2, and MAPK14 were able to bind effectively and stably (Fig. 5f and S9). Among them, Hydrogen bonds were observed between FA(18:2) and STAT (PDB id: 6NJS) at active sites LYS-581; DHA and BCL2 (PDB id: 6GL8) at active sites ARG-146; FA(18:2) and BCL2 at active sites ARG-146; DHA and MAPK14 (PDB id: 9CJ1) at active sites ASP-68 and LYS-53; AA and MAPK14 at active sites ALA-51, LEU-104 and THR-106; FA(18:2) and MAPK14 at active sites THR-106. Therefore, the molecular docking results indicated that various active components in sea dragons exhibited favorable binding affinities with the key targets relevant to anti-sepsis. This suggested that these components may significantly contribute to combating sepsis through their interactions with these crucial targets.

Discussion

This study systematically analyzed the bioactive components and their action networks of six sea dragon species by integrating metabolomics and lipidomics techniques, and constructed the molecular fingerprints of the functional components of sea dragon species for the first time. Compared with the traditional single-omics study, the multi-omics technology breaks through the technical bottleneck in the identification of the active components of the sea dragon, and accurately localized 47 differential metabolites and 138 differential lipids.

Based on the untargeted metabolomics analysis, lipids accounted for 63.74% of the total metabolites in the sea dragon, constituting its core metabolite group. Among them, as a classical functional lipid, glycerophospholipids were the dominant component in the differential metabolites, the above results were consistent with previous studies that large amounts of LPC and PC were detected in atlantic salmon32. It is noteworthy that among all metabolites, fatty acids, lipids and lipid-like molecules were the dominant components in six sea dragon species (especially in HSH, QSB, and SE), which was consistent with the metabolomic profiling of the comprehensive analysis of fish and shrimp33. Among which, fatty acids have been broadly reported to play an indispensable role in the physiological functions of organisms, including energy supply, cell membrane structure and function, Signal transmission process; In addition, fatty acids are capable of regulating gene expression; they can interact with transcription factors to influence the transcription and translation processes of genes, which in turn regulate cellular metabolism and function. In particular, fatty acids are important in anti-inflammatory and cardiovascular protective effects34,35. The lipids and lipid-like molecules detected in this study mainly include sphingolipids, steroids, steroid derivatives, and glycerophospholipids. Sphingolipids are essential components of biofilm to maintain cell structure and participate in various physiological processes, and they are also involved in the regulation of cell proliferation, differentiation, apoptosis and other physiological processes, which is of great significance to the development and homeostasis maintenance of organisms36,37. Steroids and steroid derivatives, with their unique chemical structures and biological activities, have a wide range of applications in food additives, nutritional supplementation, functional food development and food safety and quality control. Among them, some biologically active steroid compounds have been used to develop foods with specific health functions, such as regulating blood lipids and enhancing immunity38. It is worth noting that glycerophospholipids not only have a variety of physiological functions such as participation in cell signaling, lipid metabolism and energy metabolism, but also can be widely used as emulsifiers and stabilizers in dairy products, confectionery and beverages to improve the texture and stability of food6,39. Overall, this study revealed the abundance levels of metabolites in sea dragons for the first time.

Notably, untarget metabolomics detected the inclusion of multiple functional lipids in the sea dragon in both positive and negative ion modes, such as OA and DHA. As a naturally occurring fatty acid, OA is the main component of many oils (such as olive oil and peanut oil). Studies have shown that OA can regulate intracellular signalling pathways and inhibit the release of inflammatory factors, such as tumour necrosis factor-alpha (TNF-a) and interleukin-6 (IL-6), thereby reducing the inflammatory response. In terms of lipid lowering, OA can promote the oxidative metabolism of fatty acids, reduce the levels of triglycerides and cholesterol in the blood, and improve the lipid profile40. Studies have proven that although monounsaturated fatty acids (such as OA and PA) possess anti-inflammatory and signaling properties, but excessive intake may increase the risk of cardiovascular disease. Therefore, the rational use of OA can better meet the human body’s demand for fatty acids and promote health and nutritional balance41. DHA, as a typical n-3 PUFAs, plays an important role in brain development, cardiovascular health and immune function, therefore has been widely used in infant formulas, functional foods, and medical and healthcare foods42. However, metabolomics has some limitations in the detection of lipids16, so it is expected that a more comprehensive lipidomics can be used to mine and analyze the lipid components for more accurate marker analysis.

Based on this, this study innovatively constructed an untargeted lipidomic analysis platform based on UPLC-QEO MS to systematically analyze its lipid molecular diversity. Among them, TAGs (43.30%) and PCs (16.75%) constituted the core components of the lipidome of the sea dragon (VIP > 1). TAGs were predominantly characterized by highly unsaturated structures (average number of double bonds ≥ 3) such as C16:0/C18:1/C20:5, whereas PCs were enriched in the acylated isoforms of n-3 PUFAs, such as PC 16:0/20:5. The research revealed that TAG and PC were proved to be predictive parameters for dietary nutrition-induced lipid metabolism abnormalities in fish and involved in the regulation of glycerophospholipid metabolism. Glycerophospholipid metabolism, as an important component of lipid metabolism in fish, is crucial to the maintenance of the structure and function of cell membranes, cell signalling and energy homeostasis, etc. Changes in TAG and PC levels can sensitively reflect changes in lipid metabolism due to differences in dietary intake in fish, and therefore can be used as an effective predictive indicator to help understand the occurrence and progression of abnormalities in lipid metabolism in fish. Mechanisms and developmental processes of lipid metabolism in fish43. In addition, several unsaturated fatty acids have been detected in differential lipids. Among them, PUFAs, as an essential fatty acid that cannot be synthesised by the human body itself and must be obtained from food, have shown excellent benefits in human health. Numerous studies have shown that the intake of PUFAs is closely related to the reduction of cardiovascular disease morbidity and mortality, and its mechanism of action may include regulating blood lipid levels, lowering blood viscosity, inhibiting inflammation, and improving vascular endothelial function. Meanwhile, PUFAs also have a positive effect on neurodevelopment, especially during the development of the brain and retina in foetuses and infants, and an adequate intake of PUFAs is essential for the normal development and functional maintenance of the nervous system34,42. Sea fish, in particular, is widely recognised as a rich source of essential fatty acids (EFAs) due to its unique living environment and dietary habits. For example, common species of marine fish such as salmon, cod and tuna are rich in PUFAs with important physiological functions such as EPA and DHA, etc28,42. Therefore, increasing the intake of marine fish has become one of the healthy dietary strategies recommended by many nutritionists and health experts. Overall, the six species of sea dragons showed significant differences at lipid molecular level, while lipidomic analysis will further contribute to the exploration of the potential mechanisms of multiple functional lipid components. The in-depth study of fatty acid metabolism and physiological functions of sea dragons will not only provide a better understanding of the biological significance of fatty acids, but also provide valuable references for human health and nutritional studies.

PUFAs as a class of fatty acids that are essential for human health, they contain two or more double bonds and mainly include the n-3 and n-6 PUFAs30. These PUFAs cannot be synthesized on their own in the body and must be obtained through the diet44, which play important roles in regulating blood lipids, preventing cardiovascular diseases, and promoting brain development45,46. n-3 PUFAs mainly includes ALA, EPA, and DHA. It has been confirmed that ALA can also be converted to EPA and DHA in vivo, but the conversion efficiency is too low, so it is mainly ingested from deep-sea fish and algae47. Sea dragon, as a marine fish, may present a better advantage in EPA and DHA content. n-6 PUFAs is also used as an essential fatty acid, which mainly consists of LA and ARA. LA is an important component of phospholipids that make up biological membranes, is essential for maintaining normal cell membrane function, and plays an important role in regulating cholesterol homeostasis48. ARA plays a key role in the maintenance of the nervous system, the regulation of pancreatic islet function, and the amelioration of cardiovascular diseases49. Catalyzed by different enzyme systems in living organisms, ARA can produce important active substances such as prostaglandins, thromboxane, prostacyclin, and leukotrienes47. Nevertheless, researches have shown that excess n-6 PUFAs and high n-6/n-3 PUFAs ratios contribute to the pathogenesis of many diseases, including cardiovascular disease, cancer, inflammatory and autoimmune disorders. Therefore, an appropriate n-6/n-3 PUFAs ratio is essential for good health. However, modern dietary n-6/n-3 ratios generally exceed 10:1, and excess ARA metabolites (e.g., leukotriene B4) can exacerbate inflammatory responses through activation of the NF-κB pathway50, underscoring the need for the development of n-3-dominant marine lipid resources.

It is noteworthy that the bioactivity potential of sea dragon is not only due to its PUFA characteristics, but also reflected in the level of lipid molecular diversity. In this study, it was found that TAGs accounted for 43.30% of the total lipids in sea dragons, and specific molecules such as TAG(18:1/18:2/22:5) could be used as species identification markers. Our previous study similarly found that TAGs, as a major component of GL, can serve as markers for identifying fish fraudulent behaviors, such as TAG (16:0/16:0/18:1) and TAG (16:1/18:1/18:1)28. Rocchetti et al. showed that n-3 diets rich in TAGs were effective in preventing adipogenesis and inflammatory responses51. In addition, FAHFAs were detected for the first time in sea dragons, such as FAHFA(18:2/20:4), FAHFA(18:2/22:5). FAHFA has been reported to have important physiological functions such as improving glucose tolerance, enhancing insulin sensitivity, maintaining glucose homeostasis and anti-inflammatory52. Fish, especially deep-sea fish, has been studied as a contemporary high-quality dietary source due to its rich variety of PUFA. In summary, as a new type of marine lipid pool, sea dragon (especially QSB) was rich in a variety of PUFAs (such as ALA, EPA, DHA, LA, and ARA), and the synergistic effect of its PUFA profile and characteristic lipid molecules provides a multi-targeted action basis for the development of anti-inflammatory and metabolism-regulating functional food products, and it is expected to break through the technical bottlenecks of the current development of homogenization of deep-sea fish oils.

This study revealed the multi-target regulatory potential of sea dragon in the treatment of sepsis, and its functional substances (such as alkaloids, carbohydrates and n-3 PUFAs) may interfere with the key pathological links of sepsis through synergistic effects, which have demonstrated to be effective in reducing sepsis-induced inflammatory damage or immune imbalance53. Körner et al.54 revealed that dietary n-3 PUFAs improved inflammatory resolution and sepsis survival54. In contrast, the high level of PUFAs in sea dragon found in the present study may be its core anti-sepsis component, suggesting its potential for synergistic application with conventional anti-inflammatory drugs. Combined with network pharmacology and molecular docking analyses, it was shown that the active ingredients of sea dragon may exert multidimensional regulatory effects by targeting key signaling nodes such as IL6/STAT3, PI3K/AKT1, BCL2, and MAPK. For example, the IL6-STAT3 axis serves as a core driver of the inflammatory storm in sepsis55, and the components of sea dragon may attenuate the systemic inflammatory response by inhibiting the over-activation of this pathway; at the same time, the modulation of AKT1 and BCL2 balances apoptosis and survival, whereas the moderate modulation of MAPK14/MAPK1 may alleviate inflammatory organ damage56,57. Based on GO and KEGG enrichment analyses, the anti-sepsis effect of sea dragon may be realized by synergistically modulating multidimensional pathological processes such as inflammatory responses (such as IL-6/STAT3, MAPK pathway), oxidative stress (Nrf2/HO-1 pathway), apoptosis (BCL2/AKT1-mediated mitochondrial pathway), and lipid metabolism disorders (PPARγ/PUFA pathway). This mechanism was further supported by molecular docking validation, in which key active ingredients (such as DHA, EPA) in sea dragon showed strong binding affinity with core targets such as STAT3 and AKT1. This cross-pathway and multi-link intervention not only confirms the “multi-target-multi-disease module” association predicted by network pharmacology, but also reveals the direct interactions between the components of sea dragon and the key targets of sepsis through molecular docking experiments, which highlights its unique advantage in counteracting the complex pathological features of sepsis by systematically regulating the immune-metabolism-apoptosis network. The results of the study were summarized from the “computational prediction” to the “computational prediction”. The results provide a theoretical basis for the development of natural product-based multi-target therapeutics for sepsis from the dual dimensions of computational prediction and molecular interactions.

In summary, the substance composition of six sea dragon species was investigated for the first time using untargeted metabolomics and untargeted lipidomics. Metabolomics screened a total of 47 differential metabolites based on VIP values (VIP > 1) in both positive and negative ion modes, and fatty acids and lipids and lipid-like molecules predominated. Untargeted lipidomics further elaborated the lipid composition in the six sea dragon samples. 76 and 62 differential lipids were screened in positive and negative ion mode, respectively. In addition, sea dragons were shown to contain high levels of polyunsaturated fatty acids for the first time such as TAG (18:1/18:3/19:2), FA (18:3), TAG (18:1/20:4/20:5), DAG (16:0/20:5), LPC (22:6), TAG (16:0/18:2/20:4), FAHFA (20:4/ 20:3) and FAHFA (18:2/20:4), etc., which can be effectively used as potential markers to identify the six sea dragon species. Based on metabolomics combined with lipidomics, the compositional profiles of sea dragon were refined for the first time, and the effective identification of the six species of sea dragon was achieved. In addition, a comprehensive analysis using a combination of network pharmacology and molecular docking technology was utilized to clarify the mechanism of action of the differential components in sea dragon against sepsis, providing new ideas for the targeted treatment of sepsis. These findings not only underscore the medicinal value of Syngnathus but also pave the way for the development of novel therapeutic strategies against sepsis, a condition that remains a formidable challenge in modern medicine. Future studies can further optimize the extraction process of the active ingredients of sea dragon and verify its multi-target synergistic effect through animal models to promote its application in the precision treatment of sepsis.

Methods

Reagents

Methanol (MeOH), acetonitrile (ACN), methyl tert-butyl ether (MTBE), ammonium formate, dichloromethane, isopropanol was chromatographic grade and purchased from CNW Technologies (Dusseldorf, Germany). Other chemicals and reagents were purchased from Merck Life Science (Darmstadt, Germany).

Sample preparation

Six species of sea dragons were SH, SL, SB, HSH, QSB, and SE, respectively. Sea dragon samples were pulverized into powder and stored in 4 °C for further analysis. Metabolites extraction: the sea dragon powder (50 mg) was added to 1 mL of the extraction solution (MeOH:ACN:H2O = 2:2:1 (v/v/v)), and then homogenized (35 Hz, 4 min). The sample homogenate was vortex-mixed and then sonicated in an ice-water bath for 5 min, and the above steps were repeated three times. The sample homogenate was allowed to stand at −40 °C for 1 h, followed by centrifugation (4 °C, 13800 g) for 15 min, and the supernatant was taken for further analysis.

Lipids extraction: the sea dragon powder was accurately weighed 25 mg into a centrifuge tube, and 200 μL of water was added to prepare a homogenized solution of the samples. Each group consisted of 6 parallel samples. About 2.5 times volume of lipid extraction solution (MTBE:MeOH = 5:1, v/v) was added to the homogenate and mixed well. Then, the above solution was homogenized (35 Hz, 4 min) and then sonicated in an ice bath for 5 min, which was repeated three times and then the supernatant solution was taken and vacuum dried. After drying, the sample was reconstituted with 200 μL of MeOH, sonicated for 10 min in an ice-water bath. The MeOH solution was centrifuged (4 °C, 16200 g) at low temperature for 15 min. 100 μL of the supernatant solution was used for subsequent analysis.

UPLC-QEO/MS conditions for metabolomics

The metabolites in sea dragons were separated using a VanquishTM UPLC system (Thermo Fisher Scientific, US) equipped with a Waters Acquity UPLC BEH Amide column (2.1 mm × 50 mm, 1.7 μm). The mobile phase A was an aqueous phase containing ammonium acetate (25 mmol·L−1) and ammonia (25 mmol·L−1, and the mobile phase B was ACN. The gradient elution profile was as follows: 0–0.25 min, 95% B; 0.25–3.5 min, 95–65% B; 3.5–4.0 min, 65–40% B; 4.0–4.5 min, 40% B; 4.5–4.55 min, 40–95% B; 4.55–6.0 min, 95% B. The column temperature was maintained at 30 °C, and the flow rate was set to 0.5 mL·min−1. The sample plate temperature was maintained at 4 °C, and the injection volume was 2 μL. MS analysis was performed using an Orbitrap Exploris 120 mass spectrometer in both positive and negative ion modes for metabolite detection, with data acquisition controlled by Xcalibur 4. 4 software (Thermo, US)58. The detailed instrumental parameters were set as follows: sheath gas flow rate as 50 Arb, Aux gas flow rate as 15 Arb, capillary temperature 320 °C, full MS resolution at 60000, MS/MS resolution at 15,000, collision energy at SNCE 20/30/40, and spray voltage at 3.8 kV (positive) or −3.4 kV (negative), respectively.

UPLC-QEO/MS conditions for lipidomics

The instrument used for lipidomics was the same as that used for the metabolomics described above, with the exception of the column. For untargeted lipid separation, a Phenomenex Kinetex C18 column (2.1 mm × 100 mm, 2.6 μm) was used, and the instrument parameters were adjusted accordingly. The mobile phase A consisted of an acetonitrile-water (6:4, v/v) solution containing 10 mmol·L−1 ammonium formate, and the mobile phase B was an isopropanol-water (9:1, v/v) solution containing 0.5 mmol·L−1 ammonium formate. The gradient elution profile was as follows: 0–1.0 min, 40% B; 1.0–6.3 min, 40–85% B; 6.3–8.6 min, 85% B; 8.6–8.7 min, 85–100% B; 8.7–9.3 min, 100% B; 9.3–9.4 min, 100–40% B; 9.4–12.0 min, 40% B. The column was thermostatted at 55 °C, while the auto-sampler operated at 4 °C with a 2 μL injection volume, and the flow rate was set to 0.3 mL·min−1. The detailed parameters of the instrument were consistent with those used in the metabolomics experiments.

Network pharmacology

The steps for target acquisition were as follows: SwissADME website (http://www.swissadme.ch/) was used to validate the differential composition of six species sea dragons consistent with high pharmacokinetic absorption (GI absorption) and good drug likeness (2 or more “YES”). The Swiss Target Prediction database (http://swisstargetprediction.ch/) was utilized to screen for relevant target proteins (Probability > 0) of the active ingredients. Using “Sepsis” as the keyword, relevant targets were collected in the following five databases: GeneCards (https://www.genecards.org/), TTD (https://db.idrblab.net/ttd/), OMIM (https://www.omim.org/), Drugbank (https://go.drugbank.com/), and GAD (https://maayanlab.cloud/Harmonizome/resource/Genetic+Association+Database). The above target information was standardized and named by Uniprot database. Then the targets of five databases were integrated and the duplicates were removed to obtain the comprehensive sepsis targets.

The specific methods for Screening of core targets and construction of PPI networks were as follows: The intersecting targets of active components and disease targets were obtained through Venny 2.2.0, which were the potential core targets of the active components of sea dragon against sepsis. The core targets were imported into the STRING database (https://cn.string-db.org/) to analyze the protein interaction network, and the protein species was set to Homo sapiens and the minimum interaction threshold was set to highest confidence >0.7, while the other parameters remained unchanged. The PPI network was constructed by Cytoscape 3.9.1, and the threshold was set to be the median of three important topological parameters (degree, betweenness centrality, and closeness centrality). The above eligible targets were taken as the core targets for the anti-sepsis of the differential components of six sea dragon species.

The detailed steps for GO and KEGG analysis were as follows: GO functional enrichment analysis and KEGG pathway enrichment analysis of the core targets of anti-sepsis of the differential components of six sea dragon species were performed through the Metascape database (https://metascape.org/). P < 0.05 obtained from KEGG and GO analyses indicated statistical significance. Data visualization was performed via online bioinformatics tools.

The step for Construction of “Sea dragon-component-target-pathway” networks was as follows: Cytoscape 3.9.1 software was used to construct a “component-target-pathway” network for the anti-sepsis analysis of effective differential components of sea dragon.

Molecular docking

The PubChem (https://pubchem.ncbi.nlm.nih.gov) database was utilized to obtain the 2D structures of the potential active components and converted to 3D structures using Chem3D 20.0 software. Pymol 2.2.0 software and AutoDock Tool 1.5.7 software were exploited to perform ligand removal, desolvation, hydrogenation, and charge addition operations were performed on the crystal structures, which were saved as pdbqt format files to further define the binding site grid box positions and sizes. Molecular docking was performed by AutoDock vina 1.1.2 software to assess the affinity of the components to the target proteins, and the results were visualized and analyzed using Pymol 2.2.0 software.

Data processing and analysis

The raw data files of untargeted metabolomics and lipidomics were analyzed after conversion to mzXML format using the “msconvert” program from ProteoWizard software. Retention time correction, peak identification, peak extraction, peak integration, and peak alignment were performed on the mass spectrometry data using CentWave algorithm in XCMS software with minfrac set to 0.5 and cutoff set to 0.3. The metabolites of untarget metabolomics were identified through an in-house R package and annotated with the HMDB (https://hmdb.ca/) and BiotreeDB (V3.0) database59. The lipidomics data were identified using the lipidblast database, where lipid (sub) class, MS2.score. accurate m/z, and RT (retention time) were calculated. MarkerView software (Sciex) was employed for peak alignment and normalization. Additionally, the lipid network was constructed using BioPAN.

Six replicate samples were set up for each sea dragon from extraction to detection, totaling 36 data sets. Multivariate statistical analyses were performed using MetaboAnalyst 6.0 (https://www.metaboanalyst.ca/MetaboAnalyst/), including PCA and PLS-DA. Data screening excluded features with standard deviation (SD) values higher than 20%. Data scaling was set to Mean centering only. VIP values were obtained from PLS-DA analyses, and VIP > 1 was considered to be a contributing lipid component. To avoid overfitting, PLS-DA performed a replacement test (1000 replacements). Origin 2022 software, GraphPad Prism 9 software and OmicStudio tools (https://www.omicstudio.cn/tool) were mainly used to generate graphs.