Chemical composition and antioxidant properties of selected mangrove species from the Sine Saloum Delta (Senegal) using UHPLC-MS/MS and molecular networking Authors: Cheikhouna Gaye, Elnur Garayev, Yoro Tine, Fathi Mabrouki, Myriam Bertolotti, Charifat Saïd Hassane, Sok-Siya Bun-Llopet, Djibril Fall, Anne Bousquet-Mélou, Béatrice Baghdikian Published: 2026 ABSTRACT Mangrove ecosystems are important sources of secondary metabolites with significant pharmacological potential. However, the chemical diversity of West African mangroves remains largely unexplored. This study investigated the phytochemical composition and antioxidant activity of aqueous and hydroethanolic extracts from Avicennia germinans, Laguncularia racemosa, and Conocarpus erectus leaves, harvested from the Sine Saloum Delta Biosphere Reserve (Senegal). Metabolic profiling was performed by UHPLC-MS/MS combined with Feature-Based Molecular Networking (FBMN) for metabolite annotation. Antioxidant activity was assessed by off-line and on-line DPPH and ABTS assays, while the total phenolic content (TPC) was determined by the Folin-Ciocalteu method. A total of 153 metabolites were tentatively annotated, primarily phenolic acids, flavonoids, hydrolysable tannins and terpenoids. L. racemosa and C. erectus exhibited the highest TPC and the strongest antioxidant activities, while A. germinans showed limited free radical scavenging activity despite its high triterpenoid content. On-line UHPLC-DPPH/ABTS-MS/MS analysis enabled the rapid identification of the main antioxidant metabolites such as ellagic acid, myricetin and quercetin glycosides. This metabolomics approach highlights the chemical richness and antioxidant potential of L. racemosa and C. erectus. This study contributes to the chemical characterization and valorization of Senegalese mangroves as promising sources of bioactive natural products.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1038/s41598-026-59275-3. Introduction Mangroves are unique forest ecosystems located at the land-sea interface in tropical and subtropical regions. Mangrove plants are known for their remarkable physiological and biochemical adaptations to stressful environments, such as tidal flooding, salinity, UV exposure, and low oxygen availability1. Plant metabolites are generally classified as primary and secondary metabolites. Primary metabolites, which include carbohydrates, amino acids, proteins and lipids, are directly involved in essential cellular functions such as growth, respiration, and energy production2. In contrast, secondary metabolites are not directly required for basic life processes but play a major role at the ecosystem level and in plant adaptation to their environment, particularly in their defense against biotic and abiotic stresses3. These specialized metabolites include phenolic compounds such as flavonoids or tannins, alkaloids and terpenoids, many of which exhibit important biological activities4. The environmental constraints to which mangroves are subjected stimulate the biosynthesis of these secondary metabolites which contribute to stress resistance and also play an ecological role in defense, reproduction and allelopathy5. Among these metabolites, phenolic compounds, which are ubiquitous in mangrove plants, are of considerable interest due to their antioxidant properties and other potential beneficial effects6. These compounds can neutralize reactive oxygen species (ROS), thus protecting biological systems from oxidative damage, and are therefore of great interest for their pharmacological potential7. In Senegal, mangrove ecosystem represents the northernmost mangrove formation along the West African coast, found mainly in four areas: the Sine-Saloum Delta Biosphere Reserve, Casamance, the mouth of the Senegal River, and the commune of Joal8. Despite their ecological importance, the floristic diversity of Senegalese mangroves is limited, and is represented by six species, belonging to three different botanical families: Rhizophora mangle (Rhizophoraceae), Rhizophora racemosa (Rhizophoraceae), Rhizophora harrisonii (Rhizophoraceae), Avicennia germinans (Acanthaceae), Laguncularia racemosa (Combretaceae), and Conocarpus erectus (Combretaceae)9. Although the antioxidant potential of mangrove has been investigated in Asia and Latin America10, data concerning West African mangroves, particularly those of Senegal, remain scarce. Recent advances in metabolomics, such as ultra-high-performance liquid chromatography coupled with tandem mass spectrometry (UHPLC-MS/MS) and feature-based molecular networking (FBMN), allow for comprehensive characterization of complex plant extracts and facilitate the annotation of structurally related metabolites11. Combined with in vitro antioxidant assays, including on-line coupling approaches, these tools offer powerful strategies for linking chemical profiles with biological activities12. The study aims to evaluate the phytochemical composition and antioxidant potential of three mangrove species: Avicennia germinans, Laguncularia racemosa and Conocarpus erectus from the Sine Saloum Delta Biosphere Reserve (Senegal). The approach combines advanced chemical profiling by UHPLC-MS/MS and molecular networking with in vitro antioxidant assays (DPPH, ABTS), including a high-throughput on-line UHPLC-DPPH/ABTS-MS/MS screening system for the rapid identification of antioxidant compounds. Materials and methods Plant material Plant collection and identification Leaf samples of Avicennia germinans (13.981967° N, 16.764014° W), Conocarpus erectus (13.983009° N, 16.766514° W) and Laguncularia racemosa (13.800833° N, 16.480861° W) were collected on March 31, 2022 in the Saloum Delta Biosphere Reserve in Senegal, under an authorisation issued by the Ministry of the Environment, Sustainable Development and Ecological Transition (authorization no. 00000420-MEDD/DAMPCP). The botanical identification was carried out by Professor Mathieu GUEYE, head of the Botanical Laboratory at IFAN (Institut fondamental d’Afrique noire) of Cheikh Anta Diop University of Dakar (Senegal), and voucher specimens were deposited under the following accession numbers: Avicennia germinans (IFAN62905), Conocarpus erectus (IFAN63024) and Laguncularia racemosa (IFAN62887). All plant samples were collected in accordance with applicable national and international institutional guidelines and legislation. Senegal, a signatory to the Nagoya Protocol, is currently developing a national procedure. The Senegalese National Focal Point (Ministry of the Environment, Sustainable Development, and Ecological Transition) has authorized the collection of plants and the shipment of freeze-dried samples to France (letter No. 0382 dated March 14, 2023). None of these plant species are endangered. They are all classified as “Least Concern” on the IUCN Red List (https://www.iucnredlist.org/) and are not listed in Appendices I, II or III of the CITES Convention. The leaf samples were air-dried at room temperature and then ground using a knife mill (Retsch SM300). The powdered plant material obtained was stored in airtight amber containers, protected from light and moisture, at an ambient temperature of 25 °C until extraction. Extraction procedures Two types of extracts were prepared from powdered leaf samples: a hydroethanolic extract and an aqueous extract. For the hydroethanolic extraction, 500 g of dried plant powder were macerated in an ethanol–water mixture (95:5, v/v) at a material-to-solvent ratio of 1:10 (g/mL) for 18 h, then percolated. This procedure was repeated twice using fresh solvent. The combined extracts were concentrated under reduced pressure, and the remaining aqueous phase was freeze-dried to obtain the final HydroEthanolic Extract (HEE). For the aqueous extraction, a decoction method was used. 10 g of plant powder were mixed with distilled water (1:10 ratio, g/mL), brought to a boil, and stirred gently. After 30 min of boiling, the mixture was cooled, filtered through cotton, frozen, and freeze-dried to obtain the Aqueous Extract (AE). Phytochemical composition determination UHPLC-MS/MS analysis conditions Ultra-High-Performance Liquid Chromatography coupled with tandem mass spectrometry (UHPLC-MS/MS) was used to profile the phytochemical composition of the plant extracts. Analyses were performed using a Thermo Scientific Dionex 3000 UHPLC system connected to a Bruker Impact II quadrupole time-of-flight (Q-TOF) mass spectrometer equipped with an electrospray ionization (ESI) source. Samples were prepared at a concentration of 10 mg/mL in methanol (HEE) or water (AE) and filtered through 0.22 μm PTFE syringe filters (Restek, France) prior to analysis. A 2.1 × 100 mm, 1.8 μm Zorbax Eclipse Plus C18 column (Agilent), maintained at 40 °C was used. The mobile phase consisted of water (A) and acetonitrile (B), both acidified with 0.1% formic acid. The gradient program was as follows: 5% B (0–2 min), 5–30% B (2–17 min), 30–70% B (17–27 min), 70–100% B (27–31 min), held at 100% B for 1 min, followed by re-equilibration at 5% B until 35 min. The flow rate was 0.8 mL/min and the injection volume was 2 µL. Mass spectrometric detection was performed in both positive and negative ionization modes. The MS acquisition parameters were as follows: drying gas N2, drying gas temperature of 250 °C, dry gas flow of 12 L/min, nebulizer pressure of 3.5 bar, capillary voltage of 3500 V in positive mode and 3000 V in negative mode and scan range was set to 50–1500 m/z. A data-dependent acquisition (DDA) method was applied, automatically selecting the three most intense precursor ions per cycle for fragmentation, with stepped collision energies ranging from 20 to 40 eV. For internal calibration, a sodium formate/acetate solution producing cluster ions across the relevant m/z range was automatically injected at the beginning of each analysis. This ensured a mass accuracy of less than 2 ppm across the entire scan range. Data processing and molecular network Raw UHPLC-MS/MS data were calibrated using DataAnalysis software (Bruker, version 5.0). Calibration was performed with a sodium formate/acetate solution, generating cluster ions across the relevant m/z range. This ensured accurate mass correction prior to data conversion. The calibrated files were then converted to ‟.mzXML” format using the MSConvert tool (version 3.0.19330, ProteoWizard Software Foundation, USA). Precursor ion correction was applied as previously describe13 to ensure accurate MS/MS precursor value assignment. Data preprocessing was conducted using MZmine version 3.9.014, including mass detection, chromatogram building, peak deconvolution, isotopic grouping, feature alignment across samples and gap-filling. The resulting aligned feature list, including retention time, m/z, intensity, and MS/MS data, were exported as ‟.csv” and ‟.mgf” files. Parameter settings used for MZmine processing are provided in Table S1 of the supplementary materials. To explore and annotate the chemical composition of the plant extracts, a Feature-Based Molecular Networking (FBMN) workflow was performed via the GNPS platform15,16. This approach integrates feature detection (m/z, retention time, intensity) with MS/MS spectral clustering, enabling visualization of spectral relationships between features with cosine score > 0.70. Data acquired in positive and negative ionization modes were processed separately, resulting in two independent molecular networks. This polarity-based separation improved peak detection and ion-specific annotation. The networks can be accessed through the following public GNPS links: Negative mode (https://gnps.ucsd.edu/ProteoSAFe/status.jsp? task=b22963c380774c7bad578b229bfea01f), accessed on 17 Jun 2025) and Positive mode (https://gnps.ucsd.edu/ProteoSAFe/status.jsp? task=a86e47ced6644879b98d016cd7557007, accessed on 17 Jun 2025). The molecular networks were visualized using Cytoscape software (version 3.7.2)17, where each node represents a distinct feature with corresponding MS/MS spectra and edges represent spectral similarity based on cosine scores. After the metabolites are annotated using Feature-Based Molecular Network (FBMN), the identified compounds are classified according to their molecular structure, chemical properties, and biosynthetic pathways18. Chemioinformatic tools are essential for exploiting structural data. ClassyFire, for example, enables the automated chemical classification of molecular structures and has been used to train other platforms, such as CANOPUS, for the systematic annotation of compound classes from MS/MS data19,20. Heatmaps provide an effective visual representation of the classification results, illustrating the relative abundance of metabolites calculated from the square root of the sum of the areas of their peaks in negative and positive ionization modes. This visualization was generated using a freely available Python script on GitHub21. On the heatmaps, dark shades indicate high metabolite abundance, while light yellow tones correspond to lower levels. Quantification of total phenolic contents The Total Phenolic Content (TPC) was determined according to the Folin-Ciocalteu method22,23 with some modifications adapted from Breaud et al.13. This method is based on the reaction of the extract with the Folin-Ciocalteu reagent in the presence of sodium carbonate (Na₂CO₃), followed by an absorbance measurement, generally around 750–765 nm. Results are expressed as gallic acid equivalents (GAE). Antioxidant activity evaluation Off-line plate-based antioxidant activity assays (DPPH and ABTS) The antioxidant capacity of mangrove extracts was evaluated using two methods: DPPH and ABTS radical scavenging assays. These tests measure the ability of natural compounds to donate electrons or hydrogen atoms to neutralize free radicals. In both cases, decrease in absorbance corresponds to the radical scavenging activity of the sample, and IC₅₀ values (µg/mL) were used to express results. The DPPH assay evaluates the reduction of the purple DPPH radical at 515 nm by antioxidant compounds24. Similarly, the ABTS assay measures the ability of extracts to reduce the ABTS•⁺ radical cation, monitored by a decrease in absorbance at 734 nm25. In this study, tests were performed on 96-well microplates, and extracts were classified based on IC₅₀ thresholds as follows: IC₅₀ ≤ 20 µg/mL (high antioxidant activity), 20 < IC₅₀ ≤ 50 µg/mL (moderate activity) and IC₅₀ > 50 µg/mL (no activity). Preparation of DPPH/ABTS stock solutions, samples, and standard A fresh solution of DPPH (Sigma Aldrich St. Louis, MO, USA) was prepared daily in MeOH (0.060 mg/mL) and kept at room temperature in the dark for 1 h before use. The radical cation ABTS (ABTS•+) was generated from a method adapted from that described by Achard-Baccati et al.26. The radical stock solution was prepared by mixing equal volumes of a 7.0 mM ABTS solution (192.1 mg in 50 mL of distilled water) and a 2.45 mM potassium persulfate solution (66.2 mg in 50 mL of distilled water). The mixture was then stored in the dark at room temperature for 12 to 16 h to allow complete oxidation of the ABTS to its radical. Prior to use, this stock solution was diluted 1:20 with methanol to obtain a working solution of ABTS with an absorbance of 0.700 ± 0.05 at 734 nm. The hydroethanolic and aqueous extracts are solubilized in MeOH and H2O respectively, to obtain a stock solution of 1 mg/mL of each extract. The antioxidant standard, gallic acid (Sigma Aldrich St. Louis, MO, USA), is solubilized in MeOH at a concentration of 250 µg/mL. DPPH and ABTS 96 well plate layout DPPH/ABTS assay was performed using 96-well plates (Sterilin Ltd. Newport. UK) with samples at various concentrations. Stock solutions of the extracts were prepared in methanol and water (for HEE and AE extracts respectively) at 1 mg/mL, then diluted to final concentrations ranging from 2.5 µg/mL to 50 µg/mL. The gallic acid solution was diluted to final concentrations ranging from 0.5 µg/mL to 5.0 µg/mL. The 96-well plate layout was descripted by Breaud et al.13 and Achard-Baccati et al.26. The plates were incubated in a 96-well plate reader (BioTek EON. Providence. RI. USA) at 25 °C for 1 h (DPPH) or 30 min (ABTS). Absorbance was read after incubation at 515 nm (DPPH) and 734 nm (ABTS). Determination of IC50 values The DPPH/ABTS radical scavenging activity was measured in triplicate for each concentration of the tested compounds. The activity is expressed as a percentage of DPPH/ABTS radical scavenging, and the percentage of inhibition was calculated using the following formula (Abs stands for absorbance): \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\% {\text{ }}inhibition{\text{ }} = {\text{ }}\left[ {\left( {Abs{\text{ }}negative{\text{ }}control{\text{ }}{-}{\text{ }}Abs{\text{ }}compound} \right)/Abs{\text{ }}negative{\text{ }}control} \right]{\text{ }} \times {\text{ }}100$$\end{document} where the negative control corresponds to the absorbance of the DPPH/ABTS radical solution at 100%, prior to contact with the compounds to be evaluated. The DPPH and ABTS radical scavenging activity is expressed as IC50, which corresponds to the concentration of the tested compound required to obtain a percentage of inhibition of 50%. On-line RP-UHPLC-DPPH/ABTS-MS/MS assay An on-line reverse-phase UHPLC-DPPH/ABTS-MS/MS workflow was used for the rapid detection and annotation of antioxidant compounds in complex plant extracts, according to the method described by Breaud et al.13, with some modifications. The analytical platform integrated real-time of radical scavenging detection with high-resolution mass spectrometry (Fig. 1). Analyses were performed using an Agilent 1290 Infinity system (Santa Clara, CA, USA) equipped with a binary pump (G4220A), an autosampler (G4226A), a column oven (G1316C), and a diode array detector (DAD, G4212A). A second quaternary pump (Agilent 1200, G1311A) was used to deliver the DPPH (0.2 mM) or ABTS (0.1 mM) solutions, after the column, to the reaction coil in the oven (G1316A). Antioxidant activity was measured by UV absorbance using a second DAD/VIS detector (G1315B). Extracts were prepared at a concentration of 10 mg/mL in methanol (hydroethanolic extracts) or water (aqueous extracts), then filtered through 0.22 μm PTFE syringe filters (Restek, France) before injection. Chromatographic separation was performed on a 2.1 × 100 mm, 1.8 μm Zorbax Eclipse Plus C18 column (Agilent), using ultrapure water (A) and acetonitrile (B), both acidified with 0.1% formic acid (v/v), as mobile phases. The injection volume was 2 µL, the flow rate was 0.2 mL/min, and the column temperature was 43 °C. The following gradient was applied: isocratic at 5% B for 2 min, 5–30% B over 2–17 min, 30–70% B over 17–27 min, 70–100% B over 27–31 min, then isocratic at 100% B for 1 min (31–32 min), followed by a decrease to 5% B in 0.1 min (32–32.1 min), maintained at 5% B over 32.1–35 min (Figure.2). The UHPLC effluent was mixed on-line with the radical reagent in reaction coil (peek tube 23 m × 0.18 mm) maintained at 60 °C (flow rate: 0.2 mL/min). Antioxidant activity was evaluated by monitoring the decrease in absorbance at 515nm for DPPH and 734 nm for ABTS. Mass spectrometry detection was performed in negative electrospray ionization mode using a Synapt G2-Si Q-TOF mass spectrometer equipped with an ESI ionization source. The capillary voltage was 2 kV, the source temperature 100 °C, the sampling cone 40, the source offset 80, the desolvation temperature 150 °C, cone gas flow 100 L/Hr, nebuliser gas flow 5.0 bar and the desolvation gas flow at 600 L/Hr. Data were acquired in DDA acquisition mode with fragmentation of the 3 most abundant precursor ions. The collision energy was set in stepping mode ranging from 20 to 60 eV. A leucine-enkephalin solution was used for lockspray calibration every 15 s to ensure mass accuracy. The initial calibration of the instrument m/z was ensured by a solution of sodium formate acetate injected at the beginning of each run. All MS, UV and VIS data were processed using ChemStation C.01.07 and MassLynx V4.1 software to annotate and identify antioxidant compounds present in the samples. Results and discussion Chemical composition of the extracts Comparative analysis of metabolic profiling in positive and negative ionization modes The metabolomic profiles obtained under the two ionization modes revealed distinct distributions of metabolic pathways (Figure 2). In negative ionization mode, the strongest responses were associated with the shikimate and phenylpropanoid pathways, followed by the terpenoid pathway, indicating a strong accumulation of phenolic and terpenoid derivatives. These results highlight the predominance of phenolic compounds. Conversely, in positive ionization mode, amino acids, peptides, and alkaloids exhibited the strongest signals, reflecting improved detectability of nitrogenous and basic metabolites under these conditions. Although the shikimate/phenylpropanoid and terpenoid pathways remained detectable in both modes, their relative intensities were considerably lower in positive mode, suggesting complementary but differential ionization behavior between phenolic and alkaloid-type metabolites. Overall, the combination of the two ionization modes offers broader chemical coverage, improving the annotation of polar phenolic compounds and nitrogenous secondary metabolites. Comparative analysis of the superclass (Figure S1 of supplementary data) according to the two ionization modes revealed species-specific metabolic orientations among the three mangroves species. In C. erectus and L. racemosa, the negative ionization mode showed a strong enrichment in phenolic acids (C6-C1), coumarins and flavonoids, indicating a predominance of polyphenolic metabolites. This result is consistent with data from the literature. C. erectus contains flavonoids and tannins as its main phenolic constituents27. Flavonoid derivatives were detected and annotated as major compounds in the leaves of L. racemosa28. In A. germinans, the negative mode showed moderate signals for flavonoids. Positive ionization revealed a moderate presence of small peptides, monoterpenoids and flavonoids in all three species. Furthermore, in A. germinans, heatmap profiles revealed the presence of a nicotinic acid alkaloid, phenylpropanoids (C6-C3) and triterpenoids. Triterpenoid compounds have been described in the adventitious roots of A. germinans29. Tentative annotation of major compounds A total of 153 secondary metabolites were pre-annotated, in positive and negative modes (Table S1). The annotated and identified compounds were classified according to the confidence levels described by Schymanski et al.30: level 1 (L1): structure confirmed by the reference standard with MS, MS/MS spectra, and concordance of retention times; level 2a (L2a): probable structure with a library spectrum match or bibliographic data match; level 2b (L2b): structural diagnosis based on MS/MS fragments or ionization behavior, without bibliographic confirmation; level 3 (L3): potential candidates with uncertainties (for example positional isomers). The molecular network was generated in positive and negative mode (Fig. 3) based on cosine similarity between different MS/MS spectra of the predicted metabolites. Phenolic acids Phenolic acids are widely represented in the different mangrove species. The compound with MZmine ID-: 372 (m/z 153.0191), annotated as 2,3-dihydroxybenzoic acid, was consistently detected in all three species and in both aqueous and ethanolic extracts. Shikimic acid (ID-: 48), a key precursor in the biosynthesis of aromatic compounds via the shikimate pathway, was found in L. racemosa in both aqueous and hydroethanolic extracts. The elemental composition of gallic acid (ID-: 181 and ID+: 123) was deduced from the [M-H]- ion signal at m/z 169.0153 and the most intense fragment ion peak at m/z 125.0246 [M-H-CO]-. Other gallic acid derivatives include ethyl gallate (ID-: 1286) and gallic acid trimethyl ether (ID-: 1877 and ID+: 1575), both detected in C. erectus and L. racemosa. This is consistent with previous findings: ethyl acetate extracts of L. racemosa have been shown to contain lowmolecularweight phenolic compounds, highlighting simple phenolic acids as key metabolites31. The compound (ID-: 1517 and ID+: 1275) exhibited high intensity in the aqueous extract of C. erectus at ion peak m/z 300.9983 in negative ionization mode. The MS2 spectrum revealed prominent fragment ions at m/z 300.9983 and 283.9959. The base peak at m/z 300.9983 corresponds to the intact ellagic acid ion [M − H]-, while the ion at m/z 283.9959 results from the neutral loss of a water molecule [M − H−H2O]-. This fragmentation pattern is consistent with previous reports and confirms the identity of ellagic acid32. Previous phytochemical studies on C. erectus leaves have demonstrated the presence of gallic acid33 and ellagic acid34. Hydrolysable tannins Several hydrolysable tannins have been annotated, notably in L. racemosa and C. erectus. Ellagic acid was detected in both species and represents a characteristic hydrolysis product of ellagitannins. Corilagin (ID−: 882), a major ellagitannin, exhibits a characteristic MS/MS fragmentation profile with a prominent product ion at m/z 301.0001, corresponding to deprotonated ellagic acid. This fragment is generally considered a diagnostic marker for ellagitannin-derived structures35–37. Furthemore, 1,2,3,6-tetragalloylglucose (ID-: 1460) and 1,3,6-trigalloylglucose (ID-: 1189) were also observed in L. racemosa and C. erectus, highlighting the richness of these species in galloylated polyphenols and their potential contribution to antioxidant defenses. Hydrolysable tannins constitute a characteristic class of secondary metabolites in the Combretaceae28. It is worth noting that several forms of ellagic acid, including its mono-, di-, and tri-methylated analogs, have already been described in L. racemosa, confirming the recurring presence of these bioactive polyphenols in this species31. Castalagin and vescalagin have been tentatively annotated as major components along with numerous types of hydrolysable tannins, in the methanolic extract of C. erectus38 fruits. Flavonoids Flavonoids constitute the most structurally diverse group among those annotated in negative ionization mode. Numerous flavones, flavonols, and their glycosides were detected in the three mangrove species studied. Apigenin (ID-: 2492), a typical flavone aglycone, was detected in A. germinans and C. erectus, producing an intense molecular ion at m/z 269.0459 and characteristic MS2 fragments at m/z 117.0340 and 151.0031. These results are consistent with previously published data39. Luteolin (ID-: 2257 and ID+:1906) and luteolin 4’-methyl ether (ID-: 2556 and ID+: 2198) showed a similar distribution. Isorhamnetin-3-O-rutinoside (ID-: 1783 and ID+: 1487) eluted at 10.48 min, exhibits a deprotonated molecular ion [M–H]− at m/z 623.1616. The MS/MS spectrum reveals a major fragment ion at m/z 315 ([M–H–308]−). The neutral loss of 308 Da (162 + 146) corresponds to the presence of a hexose and a deoxyhexose moiety attached at the same position on the aglycone40. The aglycone was annotated as isorhamnetin (ID-:2586 and ID+: 2241), which is confirmed by its characteristic product ions at m/z 300, 315 and 151. Isorhamnetin-3-O-glucoside (ID-: 1845 and ID+: 1556) was also widely detected. Among the flavonols, kaempferol-3-O-glucoside (ID-: 1803) and its isomers (ID-: 1796, and 1903) were primarily found in C. erectus. The compound annotated as a kaempferol glucoside had a parent ion at m/z 447.0916 [M-H]-. Its MS/MS spectrum showed a prominent fragment ion at m/z 284.0317 [M-H-162]-, corresponding to the loss of a hexose moiety. This neutral loss is characteristic of O-glycosylated flavonols and confirms the presence of the kaempferol aglycone, thus supporting the annotation of the compound as kaempferol-3-O-glucoside. A recent study reported the presence of a kaempferol glucoside and its derivative in an ethanolic extract of C. erectus leaves41. Other kaempferol derivates have been annotated as kaempferol-3-O-rutinoside, kaempferol-3-O-arabinoside and kaempferol 3-glucuronide, in both negative and positive modes. Quercetin and its derivatives were widely detected in all three species. Extracts from C. erectus and L. racemosa revealed that the quercetin aglycone (ID-: 2258 and ID+: 1912) was eluted at 12.70 min, with a m/z at 603.0757 [2 M-H]- in negative mode, corresponding to the quercetin dimer obtained by ESI ionization. The MS2 spectrum highlighted a prominent fragment at m/z 301.0345 and other fragments at m/z 151.0032 and 178.9980 commonly observed for flavonols such as the quercetin aglycone. Four features with a common fragmentation scheme, similar to that of quercetin and referred to here as quercetin isomers, are eluted at 9.37 min, 10.61 min, 9.64 min and 10.06 min, respectively compounds (ID-: 1518 and ID+: 1230); ID+: 1512, ID+: 1326 and ID+: 1405. Quercitrin (ID-: 1811 and ID+: 1500) with a protonated ion at m/z 447.0938 in negative mode gave fragments at m/z 301.0278 and 271.0252. Quercitrin is present in the AE and HEE extracts of C. erectus, L. racemosa and A. germinans. Other quercetin derivatives, including two isomers of quercetin-3-O-xylose (ID-: 1680 and ID+: 1406) and (ID-: 1751 and ID+: 1450), as well as quercetin-3,4’-O-diglucoside (ID-: 1023 and ID+: 831) have been annotated in extracts of C. erectus and L. racemosa. Quercetin and its glycosylated derivatives have already been described in L. racemosa28 and C. erectus27. The MS2 spectra of compounds 1273 (ID-; m/z 479.0829, Rt 8.37 min), 1300 (ID-; m/z 317.0297, Rt 8.48 min), 1362 (ID-; m/z 479.08536, Rt 8.75 min), 1389 (ID-; m/z 479.0883, Rt 8.85 min), 1521 (ID-; m/z 449.0734, Rt 9.44 min) and 1429 (ID-; m/z 449.0726, Rt 9.04 min), share the characteristic fragmentation pattern common to myricetin derivatives, dominated by fragment signals at m/z 317.0299 and m/z 151.0032 in negative mode. The aglycone myricetin (ID-: 1906 ID+: 1599) and its isomer (ID-:1300 and ID+: 1067) were annotated by their characteristic fragments, as described by Saldanha et al.42. The compound ID-: 1908, was annotated by MS2 mass spectrometry data (m/z 635.0677, Rt 11.02 min), with the [2 M–H]− dimer being generated in dominant intensity by the ESI ion source, as described by Costa et al.28. Previous phytochemical studies have demonstrated the presence of myricetin and its glycosylated derivatives in the leaves of C. erectus43 and L. racemosa28. Triterpenoids Unlike L. racemosa and C. erectus (Combretaceae), A. germinans (Acanthaceae) is characterized by a predominance of triterpenoids, compounds commonly reported in Avicennia species44. Most terpenoids were detected in large quantities in A. germinans extracts, including betulinaldehyde (ID+: 2409, Rt 16.00), uvaol (ID+: 3233, 27.53), erythrodiol (ID+: 2965, Rt 22.32), betunilic acid (ID+: 3180, Rt 26.35), betulin (ID+: 3182, Rt 26.43), asiatic acid (ID+: 2794, Rt 20.18), maslinic acid (ID+: 3034, Rt 23.18) and micromeric acid (ID+: 3165, 25.83) as well as related analogues. Pentacyclic triterpenoids such as betulin and betulinic acid are frequently reported in Avicennia species29,45. The spectrum of betulinic acid in positive ion mode shows a characteristic ion, but the main ion is more of a fragment (m/z 439.3571) with significant water loss [M + H−H₂O]+ than a molecular ion or a protonated molecule. Similar characteristics are observed in the spectrum of betulin (ID+: 3182) with m/z 425.3778 [M + H-H2O]+.These fragmentation characteristics of betulin and betulinic acid have already been described46,47. In positive ionization mode, the neutral loss [M + H − H2O]+ of certain terpenoid compounds was the most positively correlated feature. Moreover, only betulinic acid was also observed in negative mode (ID-: 3652) with the advantage of preserving the deprotonated molecule [M − H]− at m/z 455.3527. Total phenolic content The total phenolic content (TPC) in the mangrove extracts was determined by the Folin-Ciocalteu colorimetric assay, a classic method for estimating the overall content of phenolic compounds48. This method relies on the electron transfer reaction between phenolic hydroxyl groups and the Folin-Ciocalteu reagent, leading to the formation of a blue complex detectable at 760 nm. Although it may also react with other reducing agents49, it remains a widely accepted and reproducible approach for quantifying total antioxidant phenolics compounds. In our study, the observed TPC values (Table 1) varied significantly depending on the species and solvent, with the ethanolic extract of L. racemosa exhibiting the highest phenolic content (12.43 g GAE/100 g extract). Similar high levels of phenolic compounds have already been reported for the leaves of L. racemosa28,31. These values are consistent with the results of UHPLC-MS/MS, which revealed a high abundance of phenolic compounds in L. racemosa, thus confirming the reliability of the Folin-Ciocalteu assay as a complementary analysis in profiling antioxidant potential in mangrove species. Table 1Total phenolic content (TPC) and antioxidant activity (ABTS and DPPH) of Mangrove Leaf Extracts from the Sine Saloum Delta”.SamplesTPCg GAE1/100 g ExtractABTSIC50 µg/mL(Mean ± SD)DPPHIC50 µg/mL(Mean ± SD) Conocarpus erectus EtOH5.56 ± 0.4350.35 ± 0.4186,67 ± 0,14 Conocarpus erectus H203.99 ± 0.1719.33 ± 0.7918,69 ± 0,17 Laguncularia racemosa EtOH5.96 ± 0.4512.17 ± 0.2219,04 ± 0,22 Laguncularia racemosa H2012.43 ± 0.1418.87 ± 0,4315.68 ± 0,18 Avicennia germinans EtOH3.99 ± 0.22> 104.75> 104.75 Avicennia germinans H203.50 ± 0.18> 104.75> 104.751Gallic Acid Equivalent; SD—Standard deviation. Table 1Total phenolic content (TPC) and antioxidant activity (ABTS and DPPH) of Mangrove Leaf Extracts from the Sine Saloum Delta”.SamplesTPCg GAE1/100 g ExtractABTSIC50 µg/mL(Mean ± SD)DPPHIC50 µg/mL(Mean ± SD) Conocarpus erectus EtOH5.56 ± 0.4350.35 ± 0.4186,67 ± 0,14 Conocarpus erectus H203.99 ± 0.1719.33 ± 0.7918,69 ± 0,17 Laguncularia racemosa EtOH5.96 ± 0.4512.17 ± 0.2219,04 ± 0,22 Laguncularia racemosa H2012.43 ± 0.1418.87 ± 0,4315.68 ± 0,18 Avicennia germinans EtOH3.99 ± 0.22> 104.75> 104.75 Avicennia germinans H203.50 ± 0.18> 104.75> 104.751Gallic Acid Equivalent; SD—Standard deviation. Total phenolic content (TPC) and antioxidant activity (ABTS and DPPH) of Mangrove Leaf Extracts from the Sine Saloum Delta”. 1Gallic Acid Equivalent; SD—Standard deviation. Antioxidant assay Off-line antioxidant assay The results showed that three extracts exhibited strong antioxidant activity in both tests. Laguncularia racemosa displayed the highest antioxidant activity (IC₅₀ of 12.17 µg/mL for the HEE extract and 15.68 µg/mL for the AE extracts, for ABTS and DPPH assays, respectively). These results are consistent with previous studies reporting high phenolic compound content and strong antioxidant potential in L. racemosaleaves. For example, Pimentel et al.50, identified several phenolic acids and flavonoids in this species, known for their free radical scavenging capabilities. Costa et al.28 and Shi et al.31 have identified several ellagitannins in this species, compounds widely recognized for their radical scavenging properties. The hydroethanolic extract of Conocarpus erectus showed a weaker response, particularly in the DPPH assay (IC₅₀ = 86.67 µg/mL), classifying it as inactive, despite some activity in the ABTS assay. In contrast, the aqueous extract demonstrated marked antioxidant activity, with IC₅₀ values below 20 µg/mL. These data corroborate the results of Afifi et al.51, who also demonstrated that C. erectus leaf extracts contain numerous polyphenols and flavonoids, such as derivatives of apigenin, luteolin, and kaempferol. This chemical profile is consistent with our molecular network analysis and UHPLC-MS/MS results, which confirmed the presence of several glycosylated flavonoids in this species (Table S3 of supplementary data). Extracts of Avicennia germinans (HEE and AE) exhibited IC₅₀ values greater than 104.75 µg/mL in both tests and were therefore considered inactive. These results suggest a limited presence of ABTS- or DPPH-reactive metabolites in this species, under the tested conditions. To determine which metabolites, contribute most to the observed radical scavenging activity, preliminary tests DPPH and ABTS in 96-well plates were followed by an On-Line RP-UHPLC-DPPH and ABTS assay, allowing direct detection of antioxidant compounds by post-column reaction with a radical solution. On-Line RP-UHPLC-DPPH/ABTS-MS/MS assay The method described by Achard-Baccati et al.26 was applied for rapid detection of the free radical scavenging components. It combined UV (325 nm) and visible DPPH/ABTS bleaching (527/724 nm). The chromatograms of the active extracts are shown in Fig. 4.The extracts which demonstrated antioxidant activity in 96-well plate assays, also showed clear radical scavenging responses in the On-Line DPPH/ABTS assay. Specifically, the aqueous and ethanolic extracts of L. racemosa and C. erectus exhibited well-defined negative peaks at 515 and 734 nm, indicating real-time inhibition of DPPH and ABTS. The On-Line RP-UHPLC-DPPH/ABTS method aims to identify the specific compounds responsible for the antioxidant activities observed in 96-well plates. To this end, the same analytical method was applied for both On-Line detection and profiling by UHPLC-MS/MS. The resulting mass spectrometry data enabled the annotation of bioactive metabolites by comparing their MS/MS spectra. In total, 35 antioxidant compounds were identified in the AE and HEE extracts of L. racemosa et C. erectus leaves (Table 2). Based on their MS/MS spectra, these compounds were tentatively annotated and classified into four major chemical groups: phenolic acids, flavonoids, hydrolysable tannins and unclassified or other compounds. Table 2Identified compounds by On-Line RP-UHPLC-DPPH/ABTS-MS/MS in leaf aqueous and hydroethanolic Extracts of C. erectus and L. racemosa.N°ID-Rt (min)FormulaAnnotation[M-H] -[M-H]- Fragments (relative intensities in %)CLPeak height (mAU)RefLRCEH2OEtOH1430.68C20H18O14HHDP-Hexoside481.0624(+ 0.7)481.0616(100.0); 275.0191(72.1); 299.0190(46.1)L2a33,07911,186 57 2520.69C46H34O30n.i532.0495(−0.1)a249.0410(100.0); 300.9998(65.2); 275.0200(55.6); 487.0352(26.2):1614103731110.86C20H18O14HHDP-Hexoside481.0624(+ 0.3)300.9991(100.0); 481.0617(87.8); 275.0194(69.9);L2a23,68479098525 57 42801.34C46H34O30n.i532.0495(−0.7)a249.0410(100.0); 275.0190(90.6); 487.0382(72.8); 301.0032(50.1):7654436252521.52C34H24O22Di-HHDP-glucose/Pedunculagin783.0686(0.3)783.0683(100.0); 275.0201(73.8); 299.0203(52.1); 300.9995(40.1)L2a82935680196 53,57 62581.53C41H26O26Castalagin/Vescalagin933.0644(−1.6)933.0687(100.0); 300.9991(56.3); 493.0046(54.5); 467.0268(46.1)L2a35282246 38 71811.14C7H6O5Gallic acid169.0142(−5)125.0246(100.0); 97.0296(17.4); 79.0191(14.6)L1109,45624,72045,99043,640 58 83482.01C11H8O66,7-Dihydroxycoumarin-4-acetic acid235.0248(+ 0.4)191.0349(100.0); 135.0449(29.4); 163.0400(29.3):81885418872193722.21C7H6O42,3-Dihydrobenzoic acid153.0193(0.7)109.0292(100.0); 91.0188(8.5); 81.0343(2.8):26,68611,945103942.33C41H26O26Castalagin/Vescalagin933.0640(0.6)631.0583(100.0); 933.0673(87.22); 300.9991(63.3); 425.0160(37.1)L2a39852656 38 114452.74C21H10O13Valoneic acid bilactone469.0049(−0.7)425.0147(100.0); 407.0042(26.7); 299.9910(18.2); 379.0077(13.3)L2a93274243 59 126204.27C16H18O7n.i321.0980(−8.9)96.9597(100.0); 321.1012(58.8); 79.9568(3.0):65,46716,00027,792137885.99C43H30O26b-1-O-Ethylvescalagin961.0965(−1.3)915.0471(100.0); 493.0038(69.0); 300.9971(60.6); 961.1008(48.4)L2a488392 60 148906.63C47H27O29Castacrenin E527.0285(−3.8)a300.9998(100.0); 505.0351(52.6); 275.0208(8.2); 257.0090(6.1)L2a17,6396880 61,62 159366.86C34H32O31Di-HHDP-galloyl-glucose935.0855(+ 4.5)301.0001(100.0); 275.0205(91.7); 935.0856(51.5); 633.0767(45.0)L2a41,056784519,525 63 1610037.22C10H20O14n.i363.0759(0.4)363.0750(100.0); 96.9599(46.0); 241.0025(3.5); 79.9571(2.8):26,72911,2071711757.99C20H32O10n.i431.1923(−0.6)44.9990(100.0); 385.1870(43.9); 153.0923(39.2); 223.1345(17.6):30,39714,7991811767.99C21H18O13Quercetin-3- glucuronide477.0311(+ 1.3)300.9982(100.0); 477.0302(11.6); 229.0133(0.7)L2a95686433 64 1912788.41C28H24O17Myricetin-O-(O-galloyl)-hexoside631.0941(−1.7)316.0234(100.0); 631.0948(93.2); 479.0839(62.9); 317.0294(34.3)L2a11,8585304 65 2013028.48C11H10O16n.i396.9891(+ 0.9)317.0316(100.0); 151.0041(47.4); 178.9992(28.1); 137.0247(25.9):87,46434,82623,5212113598.74C21H18O14Myricetin-3-glucuronide493.0618(+ 1.2)317.0301(100.0); 493.0620(17.3); 178.9987(13.2); 151.0036(11.1)L2a41,77813,94013,912 66 2213898.85C21H20O13Myricetin-3-glucoside479.0831(−0.1)316.0227(100.0); 479.0832(54.8); 271.0247(15.9); 287.0199(9.0)L2a38,38018,979 66 2313988.87n.i303.0928(+ 4.4)96.9598(100.0); 303.0906(32.6); 79.9570(7.8):66,71144,2719122414469.12C19H14O12Ellagic acid pentose433.0412(−0.1)299.9902(100.0); 300.9974(99.7); 433.0401(62.0)L2a25,57582458435 62 2515239.4C28H24O16Quercetin galloyl-O-glucoside615.0992(−2.4)615.1001(100.0); 300.0285(71.2); 463.0891(59.9);L2a19,53940978928 67 2615179.51C14H6O8Ellagic acid300.9986(1.2)300.9983(100.0); 283.9959(8.8); 145.0291(6.5); 245.0088(5.3)L137,31224,808203 58 2715679.63C21H20O12Quercetin-3-O-glucoside463.0882(−1.7)300.0289(100.0); 463.0895(51.9); 271.0257(26.5)L2a58,024972429,143 65 28169610.11C15H10O6Kaempferol285.0400(+ 1.7)284.0401(100.0); 93.0341(4.8); 117.0340(3.9)L15037248629175110.22C20H18O11Quercetin-3-O-xyloside433.0776(+ 1.8)300.0067(100.0); 271.0241(39.6); 433.0769(30.7)L2a13,4516855 68 30181110.59C21H20O11Quercetrin447.0933(−2.9)300.0283(100.0); 447.0942(50.9); 271.0254(30.2); 255.0304(15.3)L2a80,28744,478 69 31190611.01C15H10O8Myricetin317.0303(−3.7)317.0314(100.0); 151.0041(84.3); 137.0248(63.7); 178.9989(54.7)L166,51632,43344,344 70 32196811.31C15H10O7Tricetin301.0357(−2.7)151.0041(100.0); 301.0360(53.7); 178.9988(40.4); 121.0299(34.9)L2b3441988223133133198411.37C12H10O16n.i408.9896(+ 3.2)329.0316(100.0); 314.0079(82.8); 298.9845(75.5); 270.9894(20.7):64,29535,98234230412.95C16H10O83,3’-dimethylellagic acid329.0295(+ 2.3)270.9880(100.0); 314.0065(95.1); 298.9831(59.5); 312.9989(29.8)L2a74714769 32 356044.12C16H14O9n.i349.0602(−3.7)349.0596(100.0); 96.9600(49.8); 241.0029(3.8); 79.9573(2.3):91,42234,975Peak height according to UV Chromatogram (325 nm); Rt: retention time; CL: confidence level (with L1: confirmed structure (standard), L2a: library spectrum match, L2b: diagnostic evidence, L3: tentative candidate); ID-: MZmine ID in negative ionization mode; Ref: references; n.i : not identified; a [M-H]. Table 2Identified compounds by On-Line RP-UHPLC-DPPH/ABTS-MS/MS in leaf aqueous and hydroethanolic Extracts of C. erectus and L. racemosa.N°ID-Rt (min)FormulaAnnotation[M-H] -[M-H]- Fragments (relative intensities in %)CLPeak height (mAU)RefLRCEH2OEtOH1430.68C20H18O14HHDP-Hexoside481.0624(+ 0.7)481.0616(100.0); 275.0191(72.1); 299.0190(46.1)L2a33,07911,186 57 2520.69C46H34O30n.i532.0495(−0.1)a249.0410(100.0); 300.9998(65.2); 275.0200(55.6); 487.0352(26.2):1614103731110.86C20H18O14HHDP-Hexoside481.0624(+ 0.3)300.9991(100.0); 481.0617(87.8); 275.0194(69.9);L2a23,68479098525 57 42801.34C46H34O30n.i532.0495(−0.7)a249.0410(100.0); 275.0190(90.6); 487.0382(72.8); 301.0032(50.1):7654436252521.52C34H24O22Di-HHDP-glucose/Pedunculagin783.0686(0.3)783.0683(100.0); 275.0201(73.8); 299.0203(52.1); 300.9995(40.1)L2a82935680196 53,57 62581.53C41H26O26Castalagin/Vescalagin933.0644(−1.6)933.0687(100.0); 300.9991(56.3); 493.0046(54.5); 467.0268(46.1)L2a35282246 38 71811.14C7H6O5Gallic acid169.0142(−5)125.0246(100.0); 97.0296(17.4); 79.0191(14.6)L1109,45624,72045,99043,640 58 83482.01C11H8O66,7-Dihydroxycoumarin-4-acetic acid235.0248(+ 0.4)191.0349(100.0); 135.0449(29.4); 163.0400(29.3):81885418872193722.21C7H6O42,3-Dihydrobenzoic acid153.0193(0.7)109.0292(100.0); 91.0188(8.5); 81.0343(2.8):26,68611,945103942.33C41H26O26Castalagin/Vescalagin933.0640(0.6)631.0583(100.0); 933.0673(87.22); 300.9991(63.3); 425.0160(37.1)L2a39852656 38 114452.74C21H10O13Valoneic acid bilactone469.0049(−0.7)425.0147(100.0); 407.0042(26.7); 299.9910(18.2); 379.0077(13.3)L2a93274243 59 126204.27C16H18O7n.i321.0980(−8.9)96.9597(100.0); 321.1012(58.8); 79.9568(3.0):65,46716,00027,792137885.99C43H30O26b-1-O-Ethylvescalagin961.0965(−1.3)915.0471(100.0); 493.0038(69.0); 300.9971(60.6); 961.1008(48.4)L2a488392 60 148906.63C47H27O29Castacrenin E527.0285(−3.8)a300.9998(100.0); 505.0351(52.6); 275.0208(8.2); 257.0090(6.1)L2a17,6396880 61,62 159366.86C34H32O31Di-HHDP-galloyl-glucose935.0855(+ 4.5)301.0001(100.0); 275.0205(91.7); 935.0856(51.5); 633.0767(45.0)L2a41,056784519,525 63 1610037.22C10H20O14n.i363.0759(0.4)363.0750(100.0); 96.9599(46.0); 241.0025(3.5); 79.9571(2.8):26,72911,2071711757.99C20H32O10n.i431.1923(−0.6)44.9990(100.0); 385.1870(43.9); 153.0923(39.2); 223.1345(17.6):30,39714,7991811767.99C21H18O13Quercetin-3- glucuronide477.0311(+ 1.3)300.9982(100.0); 477.0302(11.6); 229.0133(0.7)L2a95686433 64 1912788.41C28H24O17Myricetin-O-(O-galloyl)-hexoside631.0941(−1.7)316.0234(100.0); 631.0948(93.2); 479.0839(62.9); 317.0294(34.3)L2a11,8585304 65 2013028.48C11H10O16n.i396.9891(+ 0.9)317.0316(100.0); 151.0041(47.4); 178.9992(28.1); 137.0247(25.9):87,46434,82623,5212113598.74C21H18O14Myricetin-3-glucuronide493.0618(+ 1.2)317.0301(100.0); 493.0620(17.3); 178.9987(13.2); 151.0036(11.1)L2a41,77813,94013,912 66 2213898.85C21H20O13Myricetin-3-glucoside479.0831(−0.1)316.0227(100.0); 479.0832(54.8); 271.0247(15.9); 287.0199(9.0)L2a38,38018,979 66 2313988.87n.i303.0928(+ 4.4)96.9598(100.0); 303.0906(32.6); 79.9570(7.8):66,71144,2719122414469.12C19H14O12Ellagic acid pentose433.0412(−0.1)299.9902(100.0); 300.9974(99.7); 433.0401(62.0)L2a25,57582458435 62 2515239.4C28H24O16Quercetin galloyl-O-glucoside615.0992(−2.4)615.1001(100.0); 300.0285(71.2); 463.0891(59.9);L2a19,53940978928 67 2615179.51C14H6O8Ellagic acid300.9986(1.2)300.9983(100.0); 283.9959(8.8); 145.0291(6.5); 245.0088(5.3)L137,31224,808203 58 2715679.63C21H20O12Quercetin-3-O-glucoside463.0882(−1.7)300.0289(100.0); 463.0895(51.9); 271.0257(26.5)L2a58,024972429,143 65 28169610.11C15H10O6Kaempferol285.0400(+ 1.7)284.0401(100.0); 93.0341(4.8); 117.0340(3.9)L15037248629175110.22C20H18O11Quercetin-3-O-xyloside433.0776(+ 1.8)300.0067(100.0); 271.0241(39.6); 433.0769(30.7)L2a13,4516855 68 30181110.59C21H20O11Quercetrin447.0933(−2.9)300.0283(100.0); 447.0942(50.9); 271.0254(30.2); 255.0304(15.3)L2a80,28744,478 69 31190611.01C15H10O8Myricetin317.0303(−3.7)317.0314(100.0); 151.0041(84.3); 137.0248(63.7); 178.9989(54.7)L166,51632,43344,344 70 32196811.31C15H10O7Tricetin301.0357(−2.7)151.0041(100.0); 301.0360(53.7); 178.9988(40.4); 121.0299(34.9)L2b3441988223133133198411.37C12H10O16n.i408.9896(+ 3.2)329.0316(100.0); 314.0079(82.8); 298.9845(75.5); 270.9894(20.7):64,29535,98234230412.95C16H10O83,3’-dimethylellagic acid329.0295(+ 2.3)270.9880(100.0); 314.0065(95.1); 298.9831(59.5); 312.9989(29.8)L2a74714769 32 356044.12C16H14O9n.i349.0602(−3.7)349.0596(100.0); 96.9600(49.8); 241.0029(3.8); 79.9573(2.3):91,42234,975Peak height according to UV Chromatogram (325 nm); Rt: retention time; CL: confidence level (with L1: confirmed structure (standard), L2a: library spectrum match, L2b: diagnostic evidence, L3: tentative candidate); ID-: MZmine ID in negative ionization mode; Ref: references; n.i : not identified; a [M-H]. Identified compounds by On-Line RP-UHPLC-DPPH/ABTS-MS/MS in leaf aqueous and hydroethanolic Extracts of C. erectus and L. racemosa. Peak height according to UV Chromatogram (325 nm); Rt: retention time; CL: confidence level (with L1: confirmed structure (standard), L2a: library spectrum match, L2b: diagnostic evidence, L3: tentative candidate); ID-: MZmine ID in negative ionization mode; Ref: references; n.i : not identified; a [M-H]. Several ellagitannins were annotated, including HHDP-hexoside (ID-: 43 and 111), di-HHDP-glucose/pedunculagin (ID-: 252), di-HHDP-galloyl-glucose (ID-: 936), valoneic acid bilactone (ID-: 445), and castalagin/vescalagin (ID-: 258 and 394). β−1-O-ethylvescalagin (ID-: 788) was also detected. Hydrolysable tannins are a characteristic class of secondary metabolites in the Combretaceae family28. These metabolites are known for their very high antioxidant potential52, primarily due to their phenolic hydroxyl groups, which stabilize various reactive oxygen species (ROS), throught their ability to donate H• radicals53. Three flavonols exhibit interesting antioxidant activity: kaempferol (ID-: 1696), myricetin (ID-: 1906) and tricetin (ID-: 1968). A diverse group of flavonoid glycosides has been identified, including quercetin-3-glucoside (ID-: 1220), quercetrin (ID-: 1811), quercetin-3-O-glucuronide (ID-: 1176), quercetin galloyl-O-glucoside (ID-: 1523), quercetin-3-O-xyloside (ID-: 1751), myricetin-3-glucoside (ID- 1389), myricetin-3-glucuronide (ID-: 1359) and myricetin-O-(O-galloyl)-hexoside (ID-: 1278). Flavonoids, known for their free radical scavenging properties54, are well represented in aqueous and ethanolic extracts of C. erectus and L. racemosa. Quercetin and its conjugated metabolites have been observed to exhibit numerous antioxidant properties55. Simple but highly active phenolic acids, such as gallic acid (ID-: 181), ellagic acid (ID-: 1517) and ellagic acid pentose (ID-: 1446) were detected in extracts of C. erectus and L. racemosa. These compounds, often released during the hydrolysis of complex tannins, contribute significantly to overall antioxidant capacity due to their strong hydrogen-donating capabilities56. Ellagic acid is a well-known antioxidant polyphenol found in various plant species, including C. erectus43. Conclusions This study presents the first comprehensive metabolomic analysis of antioxidant compounds in Senegalese mangrove species, using an integrated approach combining UHPLC-MS/MS, molecular network analysis, and on-line antioxidant screening. A total of 153 metabolites were tentatively annotated, revealing the remarkable chemical diversity of these mangrove plants. Phenolic acids, flavonoids, hydrolyzable tannins and triterpenoids constitute the predominant chemical classes. Among the species studied, Laguncularia racemosa and Conocarpus erectus exhibited the strongest antioxidant activities and the highest phenolic compound contents, primarily associated with ellagitannins and flavonoid derivatives such as quercetin, myricetin, and kaempferol glycosides. The on-line UHPLC-DPPH/ABTS-MS/MS approach enabled the rapid localization of antioxidant metabolites directly within complex extracts, demonstrating the value of combining metabolomics with bioactivity-guided profiling for the discovery of natural antioxidants. Although Avicennia germinans exhibited limited free radical scavenging activity, this plant proved particularly rich in pentacyclic triterpenoids, notably betulinic acid, ursolic acid, and betulin derivatives, suggesting promising pharmacological potential beyond antioxidant applications. Beyond the characterization of antioxidant metabolites, this work contributes to a better understanding of the chemical diversity of West African mangrove ecosystems and highlights the importance of integrating metabolomics with on-line antioxidant screening for the rapid discovery of bioactive natural products in complex plant matrices. These results support the valorization of Senegalese mangrove species as promising sources of compounds of pharmacological and nutraceutical interest and reinforce their potential for use in traditional medicine. Future studies should focus on the isolation and structural confirmation of the most active metabolites, as well as on evaluating their biological activities in cellular models and in vivo. Furthermore, studying seasonal, environmental, and geographic variations could lead to a better understanding of the ecological regulation of mangrove secondary metabolism. Fig. 1Schematic representation of the on-line UHPLC–DPPH/ABTS–MS/MS system. The analytical setup allows for the coupling of reversed-phase UHPLC separation with on-line antioxidant activity detection and mass spectrometric analysis. After chromatographic separation, the UHPLC effluent is mixed on-line with DPPH or ABTS radical reagents via a temperature-controlled reaction coil. Antioxidant activity is monitored by UV–Vis detection resulting in a decrease in absorbance at 515 nm for DPPH and 734 nm for ABTS. In parallel, the eluent is directed to the mass spectrometer for MS/MS analysis, enabling the simultaneous detection, identification and annotation of antioxidant-active compounds. Representative chromatograms and corresponding radical scavenging profiles are shown on the right. Schematic representation of the on-line UHPLC–DPPH/ABTS–MS/MS system. The analytical setup allows for the coupling of reversed-phase UHPLC separation with on-line antioxidant activity detection and mass spectrometric analysis. After chromatographic separation, the UHPLC effluent is mixed on-line with DPPH or ABTS radical reagents via a temperature-controlled reaction coil. Antioxidant activity is monitored by UV–Vis detection resulting in a decrease in absorbance at 515 nm for DPPH and 734 nm for ABTS. In parallel, the eluent is directed to the mass spectrometer for MS/MS analysis, enabling the simultaneous detection, identification and annotation of antioxidant-active compounds. Representative chromatograms and corresponding radical scavenging profiles are shown on the right. Fig. 2Comparative heatmap analysis of metabolic pathway distribution in mangrove species. Heatmaps illustrating the relative abundance of metabolite classes classified by NPClassifier (NPC#pathway) in three mangrove species (Avicennia germinans, Conocarpus erectus and Laguncularia racemosa). The data represent the square root–transformed peak areas obtained from UHPLC–MS/MS analyses. (A) Negative ionization mode, highlighting a strong enrichment in metabolites derivedfrom the shikimate and phenylpropanoid pathways, particularly in L. racemosa and C. erectus. (B) Positive ionization mode, showing a greater contribution fromamino acids, peptides, alkaloids and terpenoids in all species. Color intensity reflects the relative abundance of metabolites, with darker colors indicating higher values. Comparative heatmap analysis of metabolic pathway distribution in mangrove species. Heatmaps illustrating the relative abundance of metabolite classes classified by NPClassifier (NPC#pathway) in three mangrove species (Avicennia germinans, Conocarpus erectus and Laguncularia racemosa). The data represent the square root–transformed peak areas obtained from UHPLC–MS/MS analyses. (A) Negative ionization mode, highlighting a strong enrichment in metabolites derivedfrom the shikimate and phenylpropanoid pathways, particularly in L. racemosa and C. erectus. (B) Positive ionization mode, showing a greater contribution fromamino acids, peptides, alkaloids and terpenoids in all species. Color intensity reflects the relative abundance of metabolites, with darker colors indicating higher values. Fig. 3Feature-based molecular networking of metabolites detected in negative and positive ionization modes. Feature-based molecular networks were generated from UHPLC–MS/MS data using the GNPS platform and visualized with Cytoscape. These networks represent MS/MS spectral similarity, where nodes correspond to individual molecular features and edges indicates cosine similarity between MS/MS spectra. Node size is proportional to feature intensity, and their color reflects relative abundance in the samples. (Blue) Molecular network obtained in negative ionization mode (NEG), highlighting clusters primarily associated with phenolic acids, flavonoids, hydrolysable tannins and related polyphenolic compounds. (Red) Molecular network obtained in positive ionization mode (POS), showing clusters enriched in alkaloids, peptides, flavonoids and terpenoid-type metabolites, including triterpenes such as betulin. Representative annotated compounds and major clusters are indicated. Structural similarities within clusters reflect shared substructures and fragmentation patterns. Feature-based molecular networking of metabolites detected in negative and positive ionization modes. Feature-based molecular networks were generated from UHPLC–MS/MS data using the GNPS platform and visualized with Cytoscape. These networks represent MS/MS spectral similarity, where nodes correspond to individual molecular features and edges indicates cosine similarity between MS/MS spectra. Node size is proportional to feature intensity, and their color reflects relative abundance in the samples. (Blue) Molecular network obtained in negative ionization mode (NEG), highlighting clusters primarily associated with phenolic acids, flavonoids, hydrolysable tannins and related polyphenolic compounds. (Red) Molecular network obtained in positive ionization mode (POS), showing clusters enriched in alkaloids, peptides, flavonoids and terpenoid-type metabolites, including triterpenes such as betulin. Representative annotated compounds and major clusters are indicated. Structural similarities within clusters reflect shared substructures and fragmentation patterns. Fig. 4On-line UHPLC–DPPH and UHPLC–ABTS radical scavenging profiles of mangrove leaf extracts. Representative on-line antioxidant activity chromatograms obtained after UHPLC separation of leaf extracts from Laguncularia racemosa and Conocarpus erectus. Radical scavenging activity is visualized as negative peaks detected by UV–Vis spectroscopy at 515 nm for DPPH (purple traces) and 734 nm for ABTS (blue traces). (A) L. racemosa hydroethanolic extract, (B) L. racemosa aqueous extract, (C) C. erectus hydroethanolic extract and (D) C. erectus aqueous extract. Differences in peak intensity and profile reflect variations in antioxidant composition among species and extraction solvents. On-line UHPLC–DPPH and UHPLC–ABTS radical scavenging profiles of mangrove leaf extracts. Representative on-line antioxidant activity chromatograms obtained after UHPLC separation of leaf extracts from Laguncularia racemosa and Conocarpus erectus. Radical scavenging activity is visualized as negative peaks detected by UV–Vis spectroscopy at 515 nm for DPPH (purple traces) and 734 nm for ABTS (blue traces). (A) L. racemosa hydroethanolic extract, (B) L. racemosa aqueous extract, (C) C. erectus hydroethanolic extract and (D) C. erectus aqueous extract. Differences in peak intensity and profile reflect variations in antioxidant composition among species and extraction solvents. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1