Neuroprotective Potential of Xanthoceras sorbifolium Seed Oil: GC-MS Profiling and Fatty Acid-Binding Protein 7-Targeted Computational Modeling

Abstract

This study investigated the neuroprotective potential of Xanthoceras sorbifolium Bunge (XSB) seed oil through fatty acid profiling, antioxidant assays, and in silico targeting of FABP7. Among the solvent-to-solid ratios tested, 1:20 (w/v) gave the highest oil yield (72.91%) and the strongest ABTS, DPPH, and FRAP activities. GC-MS identified 17 fatty acids from the 1:20 (w/v) oil extract, with linoleic acid (38.93%) and oleic acid (31.3%) as the major constituents. Following GC-MS fatty acid profiling, lipid structural characterization was performed using ¹H NMR and FT-IR. ADME/T prediction and BOILED-EGG analysis suggested favorable pharmacokinetic properties and BBB permeability for the selected fatty acids. Molecular docking and simulation revealed strong and stable interactions of five compounds with FABP7: nervonic acid (−6.1 kcal/mol), erucic acid (−6.002 kcal/mol), eicosadienoic acid (−6.08 kcal/mol), oleic acid (−6.03 kcal/mol), and linoleic acid (−6.00 kcal/mol), outperforming the native ligand, oleic acid (−5.8 kcal/mol). These findings indicate that XSB seed oil contains bioactive lipids with promising FABP7-targeted neuroprotective potential and warrant further investigation as therapeutic leads for neurodegenerative diseases.

1. Introduction

Neurodegenerative disorders comprise a diverse group of chronic neurological diseases marked by progressive dysfunction and loss of neurons in the central and peripheral nervous systems [1]. Among these conditions, Alzheimer’s disease (AD) is the most prevalent and is a leading contributor to the global rise in dementia, severely affecting cognition, behavior, and quality of life [2]. Despite available symptomatic therapies, current drug resources offer limited efficacy and may cause significant side effects, highlighting an urgent need for safer disease-modifying interventions. A major obstacle in the development of therapeutics for neurodegenerative diseases is the blood–brain barrier (BBB), a highly specialized and selective endothelial interface that restricts the entry of over 98% of small molecules and nearly all large molecules into the central nervous system [3,4], thus necessitating the rational design or identification of compounds with precise physicochemical attributes conducive to efficient BBB permeation.

Growing evidence indicates that plant-derived compounds, possessing antioxidant, anti-inflammatory, and neuroprotective actions, may provide promising therapeutic avenues for AD and other neurodegenerative disorders. Xanthoceras sorbifolium Bunge (XSB), a medicinal plant native to China, is noteworthy for its traditional use in cognitive health and age-associated neurological conditions [5]. Its seed oil is rich in polyunsaturated fatty acids, notably nervonic and oleic acids, as well as triterpenoid saponins [6], all of which modulate oxidative stress, synaptic function, and neuroinflammation [7]. Previous studies show that extracts from X. sorbifolium confer protection against oxidative and inflammatory damage, improving memory performance in in vivo AD models [8]. Building on this background, the present study aimed to characterize XSB seed oil and to prioritize its fatty-acid constituents according to their potential neuroprotective relevance, with particular emphasis on predicted BBB-related properties and interactions with FABP7.

FABP7 is a brain-enriched fatty-acid-binding protein that has been associated with brain lipid handling and neurological disease mechanisms [9,10]. In this study, FABP7 was selected as a docking target because of its established roles in fatty-acid handling within the central nervous system and its emerging association with neurological diseases [11]. Its lipid-binding characteristics make it a relevant candidate target for investigating potential interactions with fatty acids identified in XSB oil. The mechanisms through which individual XSB-derived fatty acids may relate to Alzheimer’s disease remain insufficiently defined. Therefore, the present study aimed to prioritize XSB oil fatty acids with potential neuroprotective relevance, with particular emphasis on their predicted BBB-related properties and potential interactions with FABP7.

Recent studies have further supported the anti-Alzheimer’s potential of Xanthoceras sorbifolium oil. Du et al. [12] reported that an integrated network pharmacology–metabolomics approach identified active oil constituents linked to AD-related targets and validated their biological relevance in vitro and in vivo. More recently, a 5×FAD mouse study showed that XSBO improved cognitive performance, reduced Aβ deposition, alleviated neuroinflammation, and modulated gut microbiota and unsaturated fatty acid metabolism [13]. Building on this growing literature, the present study focuses on a complementary question: which XSBO-derived fatty acids are most likely to be BBB-permeable and to interact with FABP7. We combined extraction optimization, chemical profiling, antioxidant assays, ADME/T prediction, BOILED-EGG analysis, FABP7 docking, molecular dynamics simulation, and network pharmacology to prioritize candidate lipids for further studies.

2. Materials and Methods

2.1. Seed Preparation and Oil Extraction

Mature seeds of XSB were purchased from Agricultural Corporation Hyundai Forestry Bio Industry Co., Ltd., located in Seosan-si, Chungcheongnam-do, South Korea. The seed shells were carefully removed to retain the kernels intact. The dried seed kernels were then stored at 4 °C until further use.

2.2. Preparation of Seed Oil

Seeds of XSB were prepared by weighing 12 g per sample and roasting at 60 °C in a dry oven overnight, as a mild pre-treatment to improve oil release while limiting thermal degradation. The roasted kernels were finely ground and distributed into beakers containing 36 mL, 60 mL, 120 mL, or 240 mL of solvent to achieve solvent-to-sample ratios of (1:3, 1:5, 1:10, and 1:20 w/v), respectively, adapted from Cui, C., et al. [14]. All other chemicals used throughout the analysis were of analytical grade and were purchased from Duksan Chemicals (Ansan, South Korea). Each grounded sample was mixed with hexane at 37 °C in a shaking incubator (SI-600R, Jeio Tech, Daejeon, Korea) set at 100 rpm for 16 h. Hexane was used as the extraction solvent due to its common use for nonpolar lipid recovery; solvent ratios were selected to evaluate mass-transfer effects on extraction efficiency and antioxidant co-extraction. The extract was centrifuged using a Laborgene 1248R centrifuge (Labogene Co., Daejeon, Korea) at 4000× g for 10 min, and the resulting supernatant was filtered through a 0.45 µm nylon microfiltration membrane (Millipore, Billerica, MA, USA) to further remove impurities. The filtrate was dried under a gentle stream of nitrogen gas to remove solvents and moisture, thereby preserving thermolabile constituents and preventing oxidative degradation [15]. The oil yield was calculated using the following formula [16]:

$$Yield\left(\%\right)={\displaystyle \frac{m_1}{m_0}}\times 100$$

(1)

Here, m₁ and m₀ are the extracted oil (g) and the XSB seed (g), respectively; the oil obtained was stored at −20 °C for further use.

2.3. Antioxidant Properties

2.3.1. DPPH Radical-Scavenging Assay

The DPPH radical-scavenging activity was evaluated according to the method described by [17]. Solutions of all four samples were prepared in dimethyl sulfoxide (DMSO) at a single concentration (100 mg/mL, n = 3). The DPPH solution was prepared by dissolving 3.9 mg of 2,2-diphenyl-1-picrylhydrazyl in 100 mL of 95% ethanol and stirring continuously for 2 h. Subsequently, 0.1 mL of each sample solution was mixed with 0.9 mL of the DPPH solution and incubated in the dark at room temperature for 30 min. Absorbance was measured at 517 nm using a UV spectrophotometer (UV-2075 Plus, JASCO International Co., Ltd., Tokyo, Japan). The percentage of DPPH radical scavenging was calculated using the following equation:

$$DPPH\left(\%\right)=\left(1-{\displaystyle \frac{A_1}{A_0}}\right)\times 100$$

(2)

Here, A₁ is the absorbance of the sample, and A₀ is the value of the control.

2.3.2. ABTS Radical-Scavenging Assay

The ABTS radical-scavenging activity was determined as previously described [17]. ABTS reagents were generated by reacting 7 mM ABTS with 2.45 mM potassium persulfate in distilled water and incubated in the dark under continuous stirring for 16 h. All four samples of XSB oil were prepared at a single concentration (100 mg/mL, n = 3) in Dimethyl Sulfoxide (DMSO). An aliquot of 0.05 mL from each sample was mixed with 0.95 mL of the freshly prepared ABTS solution and incubated in the dark at room temperature for 30 min. Absorbance was measured at 734 nm using a UV spectrophotometer (UV-2075 Plus, JASCO International Co., Ltd., Tokyo, Japan). The percentage of ABTS radical scavenging was calculated using the following formula:

$$ABTS\left(\%\right)=\left(1-{\displaystyle \frac{A_1}{A_0}}\right)\times 100$$

(3)

Here, A₁ is the absorbance of the sample, and A₀ is the value of the control.

2.3.3. FRAP Assay

The ferric reducing antioxidant power (FRAP) of XSB oil was determined as described in [18]. The FRAP reagent was freshly prepared by combining 300 mmol/L sodium acetate buffer (pH 3.5), 10 mmol/L 2,4,6-tris(2-pyridyl)-s-triazine (TPTZ) dissolved in 40 mmol/L HCl, and 20 mmol/L ferric chloride hexahydrate (FeCl₃·6H₂O) in a volumetric ratio of 10:1:1, respectively. Subsequently, 0.1 mL of XSB solution at a single concentration (100 mg/mL; n = 3) was mixed with 0.9 mL of the FRAP reagent for all four samples and incubated in the dark for 30 min. The absorbance was measured using a UV spectrophotometer (UV-2075 Plus, JASCO International Co., Ltd., Tokyo, Japan). at 593 nm, with ascorbic acid serving as the calibration standard. Results were quantified and expressed as milligrams of ascorbic acid equivalents per gram (mg AAE/g).

2.3.4. TBARS Assay

The thiobarbituric acid reactive substances (TBARS) value was determined using a modified method as described in [19]. For XSB oil samples, 0.5 mL of each sample was mixed with 5 mL of solution (Butanol:Methanol, 1:1) in a 15 mL tube and vortexed for 2 min, then centrifuged (Laborgene 1248R, Labogene Co., Ltd., Daejeon, Republic of Korea) at 5000× g for 10 min. Subsequently, 2 mL of the oil layer was transferred to a new tube and combined with 2 mL of TBA reagent, consisting of 3.75 g/L thiobarbituric acid, 150 g/L trichloroacetic acid (TCA), and 0.25 mol/L hydrochloric acid (HCL). The mixture was heated in a water bath at 90 °C for 15 min, then cooled to room temperature. Absorbance was measured at 532 nm using a UV spectrophotometer (UV-2075 Plus, JASCO International Co., Ltd., Tokyo, Japan). Malondialdehyde (MDA) was employed as the standard, and the TBA value was expressed as milligrams of MDA equivalents per kg of oil (mg MDA equiv./kg).

2.4. Fatty Acid Composition Analysis

2.4.1. Fatty Acid Methyl Ester FAME Preparation

Fatty acid methyl esters (FAMEs) were prepared following a previously reported method [15]. The method involved transesterifying the XSB oil with 5 mL of 6% sulfuric acid (H₂SO₄) in methanol, followed by incubation at 90 °C for 90 min in a water bath. After the reaction, 2 mL of n-hexane was added, and the mixture was vortexed thoroughly, then centrifuged (Laborgene 1248R, Labogene Co., Ltd., Daejeon, Republic of Korea) at 2000 rpm for 10 min at room temperature. The upper organic phase was carefully collected, and the solvent was evaporated under a gentle stream of nitrogen gas. The resulting residue was reconstituted in 1 mL of n-hexane, and the FAMEs were subsequently analyzed by gas chromatography–mass spectrometry (GC-MS).

2.4.2. GC-MS Analysis

The fatty acid composition of XSB oil extract was analyzed by gas chromatography coupled to mass spectrometry (GC-MS) on an Agilent 7890A system (Agilent Technologies, Santa Cruz, CA, USA) equipped with an Agilent 5975C mass-selective detector. Separation was achieved on an Agilent J&W DB-5ms fused silica capillary column (60 m × 0.25 mm, 0.25 µm film thickness) using helium as the carrier gas at a constant flow rate of 1 mL/min. The oven temperature was initially held at 50 °C for 3 min, then ramped to 310 °C at 5 °C/min and maintained at 310 °C for 25 min. The injector temperature was set to 280 °C. Mass spectra were acquired in full-scan mode over the range 35–600 m/z with an ionization energy of 70 eV and a scan rate of 0.132 s/scan. Identification of compounds was based on comparison with spectra from the Agilent MassHunter Library, incorporating data from both the National Institute of Standards and Technology (NIST) and Wiley Libraries.

2.4.3. FT-IR

The infrared spectra of XSB oil (50 mg) were recorded using a Fourier-transform infrared (FTIR) spectrophotometer (Frontier, PerkinElmer, Waltham, MA, USA). Spectral data were collected over the range 400–4000 cm⁻¹ with a spectral resolution of 4 cm⁻¹. Each spectrum was accumulated from 32 scans to improve signal quality.

2.4.4. ¹H-NMR Spectroscopy

An accurately weighed 50 mg portion of the oil sample was dissolved in 600 µL of deuterated chloroform (CDCl₃) for proton nuclear magnetic resonance (¹H NMR) analysis. Spectra were acquired on a Bruker Avance ARX 400 spectrometer (Burker, Billerica, MA, USA) operating at a proton frequency of 400 MHz, using 64 scans to ensure an adequate signal-to-noise ratio. This protocol consists of established methods for lipid and fatty acid profiling by ¹H NMR, enabling reliable identification and quantification of characteristic proton environments in complex oil matrices.

2.5. In Silico Analysis

2.5.1. ADME/T Analysis

ADME/T (Absorption, Distribution, Metabolism, Excretion, and Toxicity) properties of the identified fatty acids from XSB oil were predicted using the Molsoft ICM-Pro v3.9-4. software (2025; https://molsoft.com/mprop/; accessed on 7 July 2025). For each compound, critical pharmacokinetic parameters, such as the number of hydrogen bond acceptors (HBAs), hydrogen bond donors (HBDs), molecular logP (MolLogP), aqueous solubility (MolLogS), topological polar surface area (MolPSA), blood–brain barrier (BBB) score, and drug-likeness model score, were computationally calculated [20,21]. All input structures corresponded to fatty acids identified from the sample, and results were systematically tabulated. The analysis provided a comprehensive profile of each compound’s physicochemical properties and potential CNS penetration, with drug-likeness assessed using established predictive models. This in silico assessment aided in prioritizing compounds for further biological investigation based on their predicted pharmacokinetic and CNS-active properties.

2.5.2. BOILED-EGG Model Prediction

The blood–brain barrier (BBB) penetration potential of fatty acids identified in the sample was assessed using the BOILED-EGG predictive model implemented in SwissADME (2025; http://www.swissadme.ch/; accessed on 9 July 2025). For each compound, the canonical SMILES code was entered into the SwissADME interface to calculate key molecular descriptors, including topological polar surface area (TPSA) and multiple consensus lipophilicity indices (iLOGP, XLOGP3, WLOGP, MLOGP). The BOILED-EGG model integrates TPSA and lipophilicity to graphically predict the likelihood of passive gastrointestinal absorption and BBB penetration. Compounds falling within the BOILED-EGG’s “yolk” region were considered to possess a high probability of crossing the BBB, reflecting favorable physicochemical properties for central nervous system (CNS) pharmacokinetics. This approach facilitated rapid in silico screening of candidate fatty acids for BBB permeability profiles, which is essential for prioritizing bioactive compounds with CNS-targeted potential. This model was used for comparative analysis with the BBB score using the MolSoft software.

2.5.3. Network Pharmacology

Prediction of Potential Targets for Compounds in XSB Oil

Network pharmacology analysis was conducted by first obtaining the SMILES notation of all fatty acids present in the XSB oil from the PubChem database (2025; https://pubchem.ncbi.nlm.nih.gov/; accessed on 10 July 2025). Potential molecular targets for these compounds were identified using the SuperPred database (2025; https://prediction.charite.de/; accessed on 10 July 2025), with target gene names standardized to “Homo sapiens” using UniProtKB (2025; https://www.uniprot.org/; accessed on 10 July 2025).

Identification of Alzheimer’s Disease-Related Target Genes

To identify targets for Alzheimer’s disease, relevant genes were retrieved from the GeneCards database (2025; http://ctdbase.org/; accessed on 12 July 2025). Duplicate entries were removed to obtain a refined gene list. Common targets shared by XSB oil fatty acids (FAs) and Alzheimer’s disease genes were identified using the Venny 2.1 diagram tool (https://bioinfogp.cnb.csic.es/tools/venny/; accessed on 14 July 2025), revealing overlapping genes that may mediate the interaction between XSB oil fatty acids and these biological processes.

Identification of Common Targets and Construction of Compound–Target Network

The compound–target interaction network was constructed in Cytoscape 3.9.1 and analyzed with the Network Analyzer plugin to identify the most active compounds. To further elucidate key molecular interactions, a compound–disease pathway interaction network was developed using STRING with an interaction confidence score threshold of 0.4 for Homo sapiens proteins. This systematic approach integrates fatty acid profiling and bioinformatics to elucidate potential molecular mechanisms of XSB oil fatty acids in relation to Alzheimer’s disease pathways using network pharmacology.

2.5.4. Preparation of FABP7 Protein (Receptor) and Ligands

In preparation for molecular docking, the ligands and receptors were retrieved and cleaned before analysis. To begin, the 3D structure of the target protein, FABP7 (PDB ID: 1FE3), was obtained from the RCSB Protein Data Bank (2025; https://www.rcsb.org; accessed on 20 July 2025). Afterward, both protein structures were cleaned by removing water molecules, adding hydrogen atoms, and minimizing them before being used in the computational experiment using the Swiss-PdbViewer software (2025; https://spdbv.unil.ch/; accessed on 22 July 2025). Meanwhile, the 3D structures of the ligands were retrieved from the PubChem database (2025; https://pubchem.ncbi.nlm.nih.gov/; accessed on 22 July 2025). Using UCSF Chimera (ver. 1.13.1; Resource for Biocomputing, Visualization, and Informatics (RBVI), University of California, San Francisco, CA, USA), the canonical SMILES code was used to generate the ligand structure, which was then minimized in preparation for molecular docking.

2.5.5. Molecular Docking Analysis

The AutoDock Vina tool was run in the same software as USCF Chimera, and the prepared protein and ligand PDB files were opened to perform the computational docking simulations. The binding site in the protein was defined with the aid of BIOVIA Discovery Studio; the grid box for FABP7 (PDB: 1FE3) was centered at x: 95.9143, y: 74.4811, and z: −40.8689, with dimensions of 14 × 13 × 16 Å around the active site. These conditions facilitated the optimal orientation of the ligands to interact with the most significant residues within the active sites of the target receptors. After running the files through the AutoDock Vina (Version 1.2.7; Center for Computational Structural Biology, The Scripps Research Institute, La Jolla, CA, USA), the interactions between the ligands and the targeted amino acids were visualized by generating 3D and 2D diagrams using PyMOL (Version 3.1.8; Schrödinger, LLC, New York, NY, USA) and BIOVIA Discovery Studio (Version 2025; Dassault Systèmes, Vélizy-Villacoublay, France).

2.5.6. Molecular Dynamics Simulation Analysis

Based on previous assessments, molecular dynamics (MD) simulations were performed using the NAMD 2.14 package [22] with the AMBER ff14SB force field [23] for the protein and the General Amber Force Field (GAFF) for the ligand [24]. The ligand topology and parameters were generated using the Antechamber and LEaP modules of AmberTools, with partial atomic charges derived using the AMI-BCC method. Using the LEaP module of AmberTools, the docked protein–ligand complex was prepared by merging the protein coordinates with the GAFF-based ligand parameter and topology files generated through Antechamber [25].

The complex system was solved in a rectangular periodic box filled with TIP3P explicit water molecules, with a 12 Å buffer extending beyond all sides of the solute [26]. Counterions (Na⁺ and Cl⁻) were added to neutralize the overall charge of the system [27]. The final system topology and coordinate files were used for NAMD simulations.

During MD simulations, energy minimization was performed for 10,000 steps to remove unfavorable contacts and optimize the geometry of the solvated system. Subsequently, the system was gradually heated from 0 K to 310 K under constant-volume (NVT) conditions using a Langevin thermostat with a damping coefficient of 1 ps⁻¹ [28]. After heating, equilibration was performed at 310 K and 1 atm under constant pressure and temperature, using the Langevin Piston Nose–Hoover method for pressure control [29].

The production MD simulation was conducted at 310 K over a 100 ns timescale using a 2-fs integration timestep. All bonds involving hydrogen atoms were constrained using the SHAKE algorithm [30]. Non-bonded interactions were calculated with a 14 Å cutoff. Long-range electrostatic interactions were computed using the Particle Mesh Ewald (PME) method [31]. Coordinates were calculated every 50 steps, and trajectories were retrieved at 1000-step intervals and analyzed using root mean square deviation (RMSD), root mean square fluctuation (RMSF), and radius of gyration (Rg) [32].

2.6. Statistical Analysis

All experiments, except FT-IR, ¹H-NMR, and GC-MS, were performed in triplicate. The data were analyzed simultaneously using ANOVA with Duncan’s multiple range test (p < 0.05) using the IBM SPSS Statistics 25 program (SPSS Inc., Chicago, IL, USA) and GraphPad Prism (Version 10.3.1, GraphPad Software LLC, Boston, MA, USA).

3. Results and Discussion

3.1. Oil Extraction

Effects of Sample-to-Solvent Ratios on Oil Yield

The solvent-to-sample ratio is a critical parameter that influences the efficiency of oil extraction. In this study, the effect of the solvent-to-sample ratio on oil yield was examined at a constant temperature of 37 °C (Table 1). Ratios of 1:3, 1:5, 1:10, and 1:20 (w/v) were tested while holding the other extraction conditions constant. Oil yield progressively increased with the solvent volume, reaching 14.5% (1.75 g) at 1:3, 27.9% (3.35 g) at 1:5, 50.75% (6.09 g) at 1:10, and a maximum of 72.91% (8.75 g) at 1:20. Increasing the solvent-to-sample ratio within the tested range enhanced the mass transfer of lipids from XSB seeds, resulting in higher oil yield, consistent with previous reports on yellow horn seed oil extraction [33]. Similar trends of enhanced extraction efficiency with increasing solvent volume have been documented, with the microwave-assisted aqueous saline process achieving a 60.43% yield at a 4:1 (v/w) ratio at 70 °C [33,34]. However, beyond a certain solvent threshold, further increases might not proportionally enhance extraction efficiency due to potential dilution effects and reduced phase separation [34].

Table 1. Effects of varying solvent-to-sample ratios on the oil extraction yield.

Sample	Solvent: Sample Ratio (w/v)	Sample (g)	Oil Yield (%)
1	1:3	12 g	14.5 ± 0.63% d
2	1:5	12 g	27.9 ± 0.80% b
3	1:10	12 g	50.75 ± 0.73% c
4	1:20	12 g	72.91 ± 0.97% a

Different letters (^a,b,c,d) indicate significant differences between samples (p < 0.05), while the same letters indicate no significant difference. Values represent mean ± SD (n = 3).

The optimized n-hexane solvent-to-sample ratio thus provides an efficient basis for extracting XSB oil, and the present study further evaluates its antioxidant capacity and fatty acid composition. Given its fatty acid profile, XSB oil may be considered for use as a functional oil ingredient (e.g., dietary supplement formulations, fortified foods) or as a source of specific long-chain unsaturated fatty acids for nutraceutical development. However, large-scale application requires complete solvent removal, regulatory safety testing (toxicology and contaminant screening), and processing optimization to ensure compliance with food-grade standards.

3.2. Antioxidant Properties

The effect of different solvent-to-sample ratios on ABTS radical-scavenging activity was evaluated, and the results are summarized in Figure 1b. Oil extracts obtained at the ratios of 1:3, 1:5, 1:10, and 1:20 (w/v) exhibited ABTS radical-scavenging activities of approximately 36.5%, 38.4%, 39.7%, and 41.3% (p < 0.05), respectively. Notably, the highest ABTS activity was observed at extraction ratios of 1:10 and 1:20, which were statistically similar, whereas the 1:3 and 1:5 extracts exhibited slightly lower scavenging activity.

Figure 1. Antioxidant activity of oil extracts obtained at different solvent-to-sample ratios: (a) FRAP activity, expressed as µg AAE/g oil; (b) ABTS radical-scavenging activity, expressed as %; (c) DPPH radical-scavenging activity, expressed as %. Different letters indicate statistically significant differences (p < 0.05), determined by Duncan’s multiple range test.

In the DPPH assay, as shown in Figure 1c, a significant enhancement in DPPH radical scavenging was observed at higher solvent-to-sample ratios. Specifically, the DPPH radical-scavenging activity was 13.9% for the 1:3 ratio, 14.6% for the 1:5 ratio, and 16.96% for the 1:10 ratio, with a statistically significant increase to 27.08% at the 1:20 ratio (p < 0.05). These findings indicate that increasing the solvent ratio effectively improves the extracts’ capacity to donate hydrogen atoms, as reflected by the higher DPPH scavenging observed under the 1:20 extraction condition. Similarly, FRAP activity (Figure 1a) corresponded to 17.7 μg AAE/g at the 1:3 ratio, 20.8 μg AAE/g at 1:5, 22.1 μg AAE/g at 1:10, and increased significantly to 28.9 μg AAE/g at the 1:20 ratio (p < 0.05). Overall, this upward trend at higher extraction ratios suggests greater concentration and availability of the reducing agents in extracts prepared with increased solvent volumes. At the 1:20 solvent-to-sample ratio, the extraction was more exhaustive, yielding oil enriched in unsaturated fatty acids and minor antioxidant components (tocopherols, phytosterols, phenolics). This compositional enrichment likely underlies the consistently higher DPPH, ABTS, and FRAP activities observed for this oil sample [35,36] Based on these results, the 1:20 sample was selected for further fatty acid profiling through GC-MS, FT-IR, and H-NMR.

3.3. TBARS Analysis

The TBARS test was performed to determine whether optimizing the solvent-to-sample ratio not only increases oil yield and antioxidant capacity but also improves XSB oil’s resistance to oxidative deterioration during storage, which is critical to its potential for food and nutraceutical applications [37]. The evolution of TBARS values, expressed as MDA concentration, for all oil samples during storage is depicted in Figure 2. Regardless of the solvent-to-sample ratio used in the extraction (1:3, 1:5, 1:10, 1:20), TBARS values increased steadily and progressively over the 10-day storage period at RT, indicating ongoing lipid peroxidation. By day 10, MDA levels ranged from approximately 4.0 to 4.3 ± 0.03 μM across all treatments; no substantial reduction in TBARS was observed as the solvent ratio increased under these storage conditions. These results demonstrate that solvent variation influenced the initial oil recovery and antioxidant potential; it exerted minimal impact on retarding oxidative deterioration during prolonged storage, as assessed by the TBARS assay. This behavior is likely attributable to the high content of unsaturated fatty acids in XSB oil, as double bonds are particularly susceptible to autoxidation and the formation of secondary oxidation products during storage [38]. Therefore, improving oxidative stability may require additional strategies, such as protection from light and oxygen, low-temperature storage, incorporation of natural antioxidants, or encapsulation.

Figure 2. TBARS values (MDA concentration) of oil samples extracted at different solvent-to-sample ratios (1:03, 1:05, 1:10, 1:20) over 10 days of storage. Error bars represent the standard deviation from three independent measurements.

3.4. Fatty Acid Composition in XSB Oil

3.4.1. GC-MS Results

The analysis of the GC-MS chromatogram and peak area data revealed 17 fatty acid methyl esters, as listed in Table 2. Among those, XSB seed oil is predominantly composed of unsaturated fatty acids, with linoleic acid (9,12-octadecadienoic acid, methyl ester) and oleic acid (9-octadecenoic acid, methyl ester) accounting for 38.93% and 31.03% of the total peak area, respectively. These two fatty acids represent the major constituents of the oil, indicating a high proportion of polyunsaturated and monounsaturated fatty acids. Other significant components include 13-docosenoic acid methyl ester (7.49%), cis-13-eicosenoic acid methyl ester (6.7%), hexadecanoic acid methyl ester (6.21%), and 15-tetracosenoic acid methyl ester (nervonic acid) (2.32%), all of which contribute to the unsaturated fatty acid content. Minor components, such as myristic acid methyl ester, (Z)-methyl heptadec-9-enoate, and 10-octadecenoic acid methyl ester, were detected at low abundance (<1%). The chromatogram in Supplementary Figure S1 demonstrates a clear separation of these constituents by retention time, confirming the reliability of identification and quantification.

Table 2. Fatty acid composition of XSB oil (1:20 w/v).

No.	Compound	RT	Peak Area (%)	MF	MW	PubChem ID
1	Myristic acid (C14:0)	35.992	0.04	C15H30O2	242.40 g/mol	31284
2	Palmitoleic acid (C16:1, Δ9)	39.728	0.05	C17H32O2	268.4 g/mol	14258
3	Palmitoleic acid (C16:1, Δ7)	39.817	0.04	C17H32O2	268.4 g/mol	543305
4	Palmitic acid (C16:0)	40.232	6.21	C17H34O2	270.5 g/mol	8181
5	Heptadecenoic acid (C17:1, Δ9)	41.76	0.04	C18H34O2	282.5 g/mol	10902087
6	14-Methylpalmitic acid	42.182	0.06	C18H36O2	284.5 g/mol	520159
7	Linoleic acid (C18:2, ω-6)	43.572	38.93	C19H34O2	294.5 g/mol	5284421
8	Oleic acid (C18:1, Δ9)	43.69	31.03	C19H36O2	296.5 g/mol	5364509
9	Oleic acid (C18:1, Δ10)	43.743	0.5	C19H36O2	296.5 g/mol	25642
10	Stearic acid (C18:0)	44.072	1.99	C19H38O2	298.5 g/mol	8201
11	Eicosadienoic acid (C20:2, ω-6)	47.091	0.46	C21H38O2	322.5 g/mol	6430995
12	Gondoic acid (C20:1, Δ13)	47.195	6.7	C21H40O2	324.5 g/mol	14122961
13	Arachidic acid (C20:0)	47.584	0.22	C21H42O2	326.6 g/mol	14259
14	Erucic acid (C22:1, Δ13)	50.482	7.49	C23H44O2	352.6 g/mol	5363109
15	Behenic acid (C22:0)	50.826	0.42	C23H46O2	354.6 g/mol	13584
16	Nervonic acid (C24:1, Δ15)	53.514	2.32	C25H48O2	380.6 g/mol	5364424
17	Lignoceric acid (C24:0)	53.827	0.26	C25H50O2	382.7 g/mol	75546

RT: Retention Time; MW: Molecular Weight; MF: Molecular Formula.

The predominance of linoleic and oleic acids, together with other long-chain unsaturated fatty acids such as docosenoic, eicosenoic, and nervonic acids, provides abundant double bonds that are susceptible to free-radical-initiated peroxidation, which is consistent with the progressive increase in TBARS values during storage [37]. The identified unsaturated fatty-acid profile may help explain the progressive increase in TBARS values observed during storage, because unsaturated fatty acids are susceptible to lipid peroxidation.

This fatty acid distribution is consistent with previous studies showing that XSB oil is rich in bioactive unsaturated fatty acids, which may underpin its nutritional value and potential neuroprotective or cardiometabolic benefits [7]. The high proportion of long-chain and very long-chain unsaturated fatty acids, including nervonic acid, also suggests promising exploratory potential for industrial uses and health-oriented applications. However, their susceptibility to oxidation highlights the need for appropriate stabilization strategies during processing and storage [14].

3.4.2. FT-IR

FTIR spectroscopy was used to characterize the major functional groups present in the seed oil and to verify that extraction conditions did not alter its fundamental molecular structure [15]. Variations in the number and types of fatty acids within triacylglycerol molecules can cause shifts in peak positions and shapes, reflecting differences in molecular structure [15,39]. The FTIR spectrum of the (1:20) sample across the 4000–400 cm⁻¹ region confirms the presence of typical functional groups and molecular structures representative of triglycerides and fatty acids in the seed oil, as also explained in Table S1.

In Figure 3, a broad peak at 3477 cm⁻¹ corresponds to O–H and N–H stretching vibrations, which can be attributed to hydroxyl or amine groups likely arising from free fatty acids. The prominent peaks at 2921 cm⁻¹ and 2851 cm⁻¹ are assigned to asymmetric and symmetric C–H stretching vibrations, representing methylene groups within the fatty acid chains. The sharp absorption at 1743 cm⁻¹ is indicative of the C=O stretching vibration associated with ester carbonyl groups typical of triglycerides [15,39]. A band at 1462 cm⁻¹ reflects C–H bending (scissoring) vibrations linked to aliphatic fatty acid chains. Additional characteristic ester C–O stretching peaks appear at 1238 cm⁻¹ and 1161 cm⁻¹, providing further confirmation of the triglyceride structure. Lastly, the band at 721 cm⁻¹ is attributed to rocking vibrations of long-chain hydrocarbons, consistent with the presence of saturated and unsaturated fatty acid chains within the oil matrix. These spectral features align with previously reported FTIR spectra of plant seed oils and XSB oil [15].

Figure 3. FTIR spectrum of XSBO at the 1:20 sample-to-solvent ratio.

3.4.3. ¹H-NMR Analysis

Figure 4 shows the NMR spectrum of XSB oil (1:20), which comprehensively characterizes the proton environments in a triglyceride-rich oil sample from a seed source. The chemical shift region around 5.2–5.4 ppm (H–I) corresponds to vinylic protons (–CH=CH–) arising from unsaturated fatty acid chains, in agreement with the high proportions of linoleic and oleic acids identified by GC-MS. Meanwhile, the multiplet at 4.1–4.3 ppm (G) is assigned to the methylene protons of the esterified glycerol, confirming the presence of triglycerides.

Figure 4. ¹H-NMR spectrum of XSB oil at the 1:20 (w/v) sample-to-solvent ratio.

The signal at ~2.7 ppm (F) reflects bis-allylic protons, typical for polyunsaturated fatty acids (e.g., linoleic acid), and the resonances at 2.0–2.3 ppm (D–E) correspond to allylic methylene protons adjacent to double bonds. The strong multiplet at ~1.5 ppm (C) indicates bulk methylene hydrogens in the long alkyl chains, consistent with the dominance of C16–C24 fatty acids, and the peak at ~1.2 ppm (B) denotes the saturated –(CH₂)_n– backbone of these acyl groups.

Finally, the triplet at ~0.9 ppm (A) is characteristic of terminal methyl groups from the ω-ends of the identified fatty acids, including linoleic, oleic, docosenoic, eicosenoic, and nervonic acids, providing NMR data consistent with the GC-MS composition. Together, these assignments confirm a complex mixture dominated by long-chain acylglycerols with both saturated and unsaturated fatty acids, consistent with the expected structure of natural triglyceride oils. The data support the successful extraction and identification of triglyceride-rich lipids, with the NMR chemical shifts and patterns aligning with the literature on edible plant oils [33].

However, the ¹H-NMR spectrum does not independently confirm the identity or relative abundance of individual fatty acids, such as linoleic, oleic, docosenoic, eicosenoic, or nervonic acid. The identification and relative quantification of these compounds are based on the GC-MS analysis presented in Table 2, whereas the NMR data provide complementary structural information on the overall lipid matrix, as explained in Table S2.

3.5. In Silico Analysis

The following data were presented stepwise: ADMET screening to exclude compounds with poor drug-likeness or CNS properties; network pharmacology to map possible downstream targets and pathway relevance; BOILED-EGG for passive BBB-penetration likelihood; docking against FABP7 to identify likely binders; and MD simulations on top candidates to assess complex stability.

3.5.1. ADME/T Analysis

The data in Table 3 present the predicted ADME properties of a series of methyl ester derivatives of fatty acids, as assessed by MolSoft in silico models. All compounds exhibit two hydrogen-bond acceptors and no hydrogen-bond donors, a pattern typical of long-chain fatty acid esters, which contributes to their overall high lipophilicity. This is reflected in MolLogP values exceeding 5 for all compounds, indicating a pronounced hydrophobic character that, while beneficial for membrane permeability, may be unfavorable for aqueous solubility and oral bioavailability according to traditional drug-likeness rules. MolLogS values range from 0.48 to 1.36 mg/L, confirming low aqueous solubility across the series, a trait that warrants consideration in formulation strategies. Topological polar surface area (MolPSA) remains constant at approximately 21 Ų, consistent with high passive permeability and likely blood–brain barrier (BBB) penetration, as indicated by uniformly high BBB scores (4.31–4.53). However, the drug-likeness model scores are consistently negative (ranging from −1.04 to −1.31), which aligns with known trends for highly lipophilic, non-aromatic natural product derivatives and suggests that these compounds may diverge from the chemical space of conventional orally active drugs. Overall, these results highlight the prominence of hydrophobic interactions and potential CNS bioavailability for the tested methyl esters, while also pointing to solubility and drug-likeness constraints that would need to be addressed for therapeutic development.

Table 3. Predicted pharmacokinetics (ADME) parameters of the screened compounds (Molsoft).

No.	Compound	HBA	HBD	MolLogP	MolLogS	MolPSA	BBB Score	DLS
1	Myristic acid (C14:0)	2	0	6.34 (>5)	1.36 (in mg/L)	21.09 A2	4.51	−1.31
2	Palmitoleic acid (C16:1, Δ9)	2	0	6.81 (>5)	0.73 (in mg/L)	21.09 A2	4.53	−1.04
3	Palmitoleic acid (C16:1, Δ7)	2	0	6.81 (>5)	0.73 (in mg/L)	21.09 A2	4.53	−1.04
4	Palmitic acid (C16:0)	2	0	7.36 (>5)	0.57 (in mg/L)	21.09 A2	4.53	−1.31
5	Heptadecenoic acid (C17:1, Δ9)	2	0	7.32 (>5)	0.68 (in mg/L)	21.09 A2	4.52	−1.04
6	14-Methylpalmitic acid	2	0	7.31 (>5)	0.68 (in mg/L)	21.09 A2	4.52	−1.15
7	Linoleic acid (C18:2, ω-6)	2	0	7.32 (>5)	0.71 (in mg/L)	21.09 A2	4.51	−1.04
8	Oleic acid (C18:1, Δ9)	2	0	7.82 (>5)	0.72 (in mg/L)	21.09 A2	4.51	−1.04
9	Oleic acid (C18:1, Δ10)	2	0	7.73 (>5)	0.65 (in mg/L)	21.09 A2	4.51	−1.04
10	Stearic acid (C18:0)	2	0	8.37 (>5)	0.48 (in mg/L)	21.09 A2	4.51	−1.31
11	Eicosadienoic acid (C20:2, ω-6)	2	0	8.33 (>5)	0.60 (in mg/L)	21.09 A2	4.46	−1.04
12	Gondoic acid (C20:1, Δ13)	2	0	8.84 (>5)	0.60 (in mg/L)	21.09 A2	4.46	−1.04
13	Arachidic acid (C20:0)	2	0	8.84 (>5)	0.60 (in mg/L)	21.09 A2	4.46	−1.04
14	Erucic acid (C22:1, Δ13)	2	0	9.85 (>5)	0.65 (in mg/L)	21.09 A2	4.40	−1.04
15	Behenic acid (C22:0)	2	0	10.39 (>5)	0.57 (in mg/L)	21.09 A2	4.40	−1.31
16	Nervonic acid (C24:1, Δ15)	2	0	10.86 (>5)	0.71 (in mg/L)	21.09 A2	4.31	−1.04
17	Lignoceric acid (C24:0)	2	0	11.40 (>5)	0.61 (in mg/L)	21.09 A2	4.31	−1.31

HBA: hydrogen bond acceptor; HBD: hydrogen bond donor; MolLogP: log of lipophilicity; MolLogS: log of aqueous solubility; MolPSA: molecular polar surface area; BBB: blood–brain barrier; DLS: drug-likeness score.

3.5.2. BOILED-EGG Model Prediction

The SwissADME “BOILED-EGG” model provides a graphical prediction of passive gastrointestinal absorption (HIA) and blood–brain barrier (BBB) permeation for small molecules, plotting logWLOGP (lipophilicity) against topological polar surface area (TPSA) [21].

In this model, the yellow region (“yolk”) in Figure 5 represents compounds predicted to be well absorbed in the gut, while the surrounding white region indicates molecules likely to penetrate the BBB; molecules outside these boundaries are generally expected to have poor oral bioavailability or CNS access [21]. Analysis of the 17 tested fatty acids in Table 4 shows that all compounds cluster at low TPSA (~26.3 Ų), suggesting high passive permeability. However, their lipophilicity (WLOGP values between 4.86 and 8.76) places several compounds at or beyond the traditional cutoff for optimal intestinal absorption, with those above WLOGP 5 possibly experiencing solubility or efflux liability [21].

Figure 5. SwissADME BOILED-EGG-predicted blood–brain barrier (BBB) permeation of the detected fatty acids (TPSA: topological polar surface area; WLOGP: Wildman–Crippen lipophilicity).

Table 4. Key descriptors used in the BOILED-EGG model.

No.	Compound	TPSA	iLOGP	XLOGP3	WLOGP	MLOGP
1	Myristic acid (C14:0)	26.3	3.88	6.41	4.86	3.94
2	Palmitoleic acid (C16:1, Δ9)	26.3	4.44	6.37	5.42	4.33
3	Palmitoleic acid (C16:1, Δ7)	26.3	4.32	6.37	5.42	4.33
4	Palmitic acid (C16:0)	26.3	4.41	7.38	5.64	4.44
5	Heptadecenoic acid (C17:1, Δ9)	26.3	4.76	6.91	5.81	4.57
6	14-Methylpalmitic acid	26.3	4.55	7.75	5.89	4.67
7	Linoleic acid (C18:2, ω-6)	26.3	4.55	7.75	5.97	4.67
8	Oleic acid (C18:1, Δ9)	26.3	4.63	7.45	6.2	4.8
9	Oleic acid (C18:1, Δ10)	26.3	4.5	7.45	6.2	4.8
10	Stearic acid (C18:0)	26.3	4.81	8.35	6.42	4.91
11	Eicosadienoic acid (C20:2, ω-6)	26.3	4.81	8.35	6.75	4.91
12	Gondoic acid (C20:1, Δ13)	26.3	5.16	8.24	6.98	5.25
13	Arachidic acid (C20:0)	26.3	5.35	9.3	7.2	5.36
14	Erucic acid (C22:1, Δ13)	26.3	5.65	9.32	7.76	5.68
15	Behenic acid (C22:0)	26.3	6.13	10.2	7.98	5.79
16	Nervonic acid (C24:1, Δ15)	26.3	6.06	10.4	8.54	6.1
17	Lignoceric acid (C24:0)	26.3	6.32	11.28	8.76	6.21

TPSA: topological polar surface area; iLOGP: in silico LogP; XLOGP3: XLogP version 3; WLOGP: Wildman–Crippen LogP; MLOGP: Moriguchi LogP.

Notably, shorter and moderately unsaturated fatty acids, such as myristic, palmitoleic, palmitic, and heptadecenoic acids, fall within the yellow region, supporting predictions of favorable oral bioavailability. In contrast, longer-chain saturated and highly unsaturated acids, such as nervonic and lignoceric, exhibit WLOGP values above 8.5, moving away from the optimal absorption zone and posing challenges for oral delivery. These results highlight how fatty acid chain length and saturation modulate ADME properties, with shorter and monounsaturated species aligning more closely with classical drug-likeness, while longer, highly lipophilic molecules may demand alternative formulation or delivery strategies. The BOILED-EGG plot, therefore, facilitates rapid assessment of oral absorption potential alongside BBB permeation, directly informing the selection and development paths for fatty acid-based drug leads.

3.5.3. Network Pharmacology Analysis

Target prediction analysis was conducted to evaluate the molecular mechanisms by which XSB oil fatty acids may influence Alzheimer’s disease. As shown in Figure 6A, network construction revealed that predicted fatty acid targets overlapped with numerous Alzheimer-associated proteins, spanning pathways implicated in apoptosis, synaptic plasticity, and neurotrophic signaling. The Venn diagram in Figure 6B quantifies this relationship, showing 82 shared protein targets between XSB oil fatty acids and Alzheimer’s disease genes, highlighting significant mechanistic convergence. These results support the hypothesis that XSB oil fatty acids engage multiple AD-relevant targets, offering pleiotropic neuroprotective potential and providing a basis for further functional validation.

Figure 6. Compound–target and Venn diagram visualization of network pharmacology analysis. (A) Predicted interactions of XSBO fatty acid targets with Alzheimer’s disease-related proteins, highlighting their engagement in neurodegenerative pathways. (B) Venn diagram indicating 82 shared protein targets between XSBO fatty acids and the Alzheimer’s disease gene set.

While network pharmacology provides systems hypotheses, we emphasize that in vitro binding assays (e.g., fluorescence displacement or isothermal titration calorimetry for FABP7), cell-based assays assessing neuroprotective endpoints, and ultimately in vivo models will be necessary to validate therapeutic potential. The computational predictions reported here serve as an initial prioritization and must be confirmed by experimental binding and functional assays to evaluate neuroprotective efficacy and safety before considering therapeutic application.

3.5.4. Molecular Docking Analysis

The three-dimensional structure of the target protein FABP7 was retrieved from the Protein Data Bank (PDB) and is depicted in Supplementary Figure S2. A total of 17 fatty acids, along with the native ligand of the protein (control), were selected as ligands for molecular docking studies. To validate the docking protocol, co-crystallized ligands were extracted from the PDB complexes, redocked into the protein, and the resulting poses were superimposed on the original structures, yielding RMSD values below 2 Å, thereby confirming the reliability of the docking approach. Based on the binding affinities in Figure 7 and RMSD values, the top-scoring ligands were identified. A comparative analysis of docking scores revealed that several fatty acids demonstrated stronger predicted binding affinities than the control, oleic acid (–5.80 kcal/mol). Notably, all top-scoring ligands with docking scores at or above –6.00 kcal/mol, including linoleic acid (C18:2, ω-6; –6.00 kcal/mol), oleic acid (C18:1, Δ9; –6.03 kcal/mol), eicosadienoic acid (C20:2, ω-6; –6.08 kcal/mol), erucic acid (C22:1, Δ13; –6.002 kcal/mol), and nervonic acid (C24:1, Δ15; –6.13 kcal/mol), outperformed the control (Table 5). These compounds vary in chain length and degree of unsaturation but commonly possess one or more double bonds, which may enhance their ability to interact with the target protein. The consistently lower docking scores suggest that these unsaturated fatty acids could form more stable or favorable interactions within the active site compared to saturated or shorter-chain analogs. This observation supports their potential as promising candidates for further investigation, as their enhanced binding affinity may translate into greater biological efficacy relative to the control compound.

Figure 7. Interaction analysis of FABP7 and ligands, including complex and residue interactions and three- and two-dimensional docked analysis. (A) Control: Oleic Acid; (B) Linoleic Acid (C18:2, ω-6); (C) Oleic Acid (C18:1, Δ9); (D) Eicosadienoic Acid (C20:2, ω-6); (E) Erucic Acid (C22:1, Δ13); (F) Nervonic Acid (C24:1, Δ15).

Table 5. Docking scores of 17 fatty acids against FABP7.

Ligands	SMILES	Docking Score (kcal/mol)
Control (oleic acid)	-	−5.80 ± 0.10
Myristic acid (C14:0)	CCCCCCCCCCCCCC(=O)OC	−5.21 ± 0.12
Palmitoleic acid (C16:1, Δ9)	CCCCCC/C=C\CCCCCCCC(=O)OC	−5.53 ± 0.15
Palmitoleic acid (C16:1, Δ7)	CCCCCCCCC=CCCCCCC(=O)OC	−5.63 ± 0.11
Palmitic acid (C16:0)	CCCCCCCCCCCCCCCC(=O)OC	−5.40 ± 0.12
Heptadecenoic acid (C17:1, Δ9)	CCCCCCC/C=C\CCCCCCCC(=O)OC	−5.70 ± 0.11
14-Methylpalmitic acid	CCC(C)CCCCCCCCCCCCC(=O)OC	−5.491 ± 0.25
Linoleic acid (C18:2, ω-6)	CCCCC/C=C\C/C=C\CCCCCCCC(=O)OC	−6.00 ± 0.05
Oleic acid (C18:1, Δ9)	CCCCCCCC/C=C\CCCCCCCC(=O)OC	−6.03 ± 0.21
Oleic acid (C18:1, Δ10)	CCCC/C=C\CCCCCCCCCCCC(=O)OC	−5.81 ± 0.30
Stearic acid (C18:0)	CCCCCCCCCCCCCCCCCC(=O)OC	−5.55 ± 0.18
Eicosadienoic acid (C20:2, ω-6)	CCCCC/C=C\C/C=C\CCCCCCCCCC(=O)OC	−6.08 ± 0.22
Gondoic acid (C20:1, Δ13)	CCCCCC/C=C\CCCCCCCCCCCC(=O)OC	−5.97 ± 0.15
Arachidic acid (C20:0)	CCCCCCCCCCCCCCCCCCCC(=O)OC	−5.74 ± 0.194
Erucic acid (C22:1, Δ13)	CCCCCCCC/C=C/CCCCCCCCCCCC(=O)OC	−6.002 ± 0.20
Behenic acid (C22:0)	CCCCCCCCCCCCCCCCCCCCCC(=O)OC	−5.89 ± 0.10
Nervonic acid (C24:1, Δ15)	CCCCCCCC/C=C/CCCCCCCCCCCCCC(=O)OC	−6.13 ± 0.05
Lignoceric acid (C24:0)	CCCCCCCCCCCCCCCCCCCCCCCC(=O)OC	−5.86 ± 0.11

Values represent mean ± SD (n = 3).

We then analyzed the top-performing compounds, those with docking scores above–6.0 kcal/mol, in detail using BIOVIA Discovery Studio to visualize and interpret their binding interactions with the target protein. BIOVIA Discovery Studio is a widely used tool that automatically generates clear 2D diagrams of protein–ligand complexes, highlighting key hydrogen bonds and hydrophobic contacts between the ligand and active site residues. This analysis will allow us to compare the interaction profiles of the top ligands, providing insight into the molecular basis for their enhanced binding affinities. By examining the specific amino acid contacts and interaction types, such as hydrogen bonds, pi–pi stacking, and hydrophobic interactions, we can better understand the structural features that contribute to strong binding and prioritize candidates for further study. The docked pose of the native ligand, oleic acid, with the FABP7 protein (Figure 7A) reveals key intermolecular interactions that contribute to binding affinity. The ligand forms conventional hydrogen bonds with THR A:60 and LYS A:58, while several hydrophobic contacts and pi–alkyl interactions are observed with ALA A:75, VAL A:25, MET A:20, LEU A:23, LEU A:117, and multiple PHE residues (A:16, A:57, A:104). These contacts collectively stabilize the ligand in the binding pocket.

As shown in Figure 7B, linoleic acid (C18:2, ω-6) in the protein-binding site forms a conventional hydrogen bond with THR A:60, serving as a key stabilizing interaction. Multiple hydrophobic and pi–alkyl contacts are present with residues such as ALA A:75, MET A:20, VAL A:25, LEU A:23, PHE A:16, PHE A:57, PHE A:104, and LEU A:117. These interactions collectively help anchor the ligand within the hydrophobic pocket. The oleic acid (C18:1 Δ9 ω-6) docking pose reveals the formation of conventional hydrogen bonds with THR A:60, THR A:53, and ARG A:106, which contribute to strong binding specificity (Figure 7C).

Additional stabilizing hydrophobic and pi–alkyl interactions occur with ALA A:75, MET A:20, VAL A:25, LEU A:23, PHE A:16, PHE A:57, PHE A:104, LEU A:117, and PRO A:38. These multiple contacts collectively anchor the ligand within the protein’s binding site. The docking interaction of eicosadienoic acid (C20:2) in the target protein binding site shows conventional hydrogen bonding with ARG A:126, anchoring the ligand to the pocket, as shown in Figure 7D. Additional hydrophobic and pi–alkyl contacts are formed with ALA A:75, MET A:20, VAL A:25, LEU A:23, LEU A:117, TYR A:19, PHE A:16, and PHE A:104, further stabilizing the ligand conformation. Erucic acid (C22:1 Δ13) (Figure 7E) forms conventional hydrogen bonds with ARG A:78 and GLN A:95, anchoring the ligand to the binding site. Additionally, extensive hydrophobic and pi–alkyl interactions are observed with ALA A:75, MET A:20, VAL A:25, LEU A:23, LEU A:117, LYS A:58, TYR A:19, PRO A:38, PHE A:16, PHE A:57, and PHE A:104. These multiple interaction types cooperatively stabilize erucic acid within the protein binding pocket, highlighting the molecular basis for its strong binding affinity. Lastly, nervonic acid (C24:1 Δ15) forms conventional hydrogen bonds with ARG A:78 and GLN A:95, providing key polar interactions that anchor the ligand to the active site.

Hydrophobic and pi–alkyl contacts are established with multiple residues, including LEU A:23, LEU A:117, MET A:20, VAL A:25, TYR A:19, LYS A:58, PRO A:38, ALA A:75, PHE A:16, PHE A:57, and PHE A:104, effectively stabilizing nervonic acid within the binding pocket. These diverse intermolecular forces suggest robust and specific protein–ligand interactions, as visualized in Figure 7.

In conclusion, the top-performing XSB oil fatty acids demonstrated robust binding interactions with key residues of FABP7, involving both hydrogen bonding and hydrophobic contacts. These results suggest that FABP7 can effectively bind and transport these fatty acids, facilitating their passage across the brain and potentially enhancing their neuroprotective effects within the central nervous system.

3.5.5. Comparative Interpretation of Docking and Molecular Dynamics Results

Comparative interpretation of the docking and molecular dynamics results showed that several long-chain unsaturated fatty acids, including nervonic acid, erucic acid, eicosadienoic acid, oleic acid, and linoleic acid, exhibited favorable predicted interactions with FABP7. The predicted binding poses indicated that these ligands could be accommodated within the hydrophobic lipid-binding cavity through interactions involving their hydrocarbon chains. However, the present computational results do not establish a direct causal relationship between carbon-chain length, degree of unsaturation, and FABP7 affinity. Nervonic acid, erucic acid, and eicosadienoic acid showed among the most favorable docking scores and stable molecular-dynamics profiles in the present analysis, whereas other evaluated fatty acids displayed relatively less favorable predicted interactions or stability profiles. These differences may reflect variations in ligand size, conformational flexibility, and hydrophobicity, although this interpretation remains hypothetical because experimental binding measurements and a systematic quantitative structure–activity relationship analysis were not performed. Accordingly, the docking and molecular-dynamics results should be regarded as an initial prioritization of XSB oil fatty acids for FABP7-binding and functional studies, rather than as evidence that these structural features directly enhance neuroprotective activity. Fatty acids are involved in several processes relevant to neurodegenerative diseases, including neuronal membrane homeostasis, neuroinflammation, oxidative stress, mitochondrial dysfunction, lipid peroxidation, and blood–brain barrier regulation. Nevertheless, the biological effects of individual fatty acids may vary according to fatty-acid type, concentration, disease context, and metabolic state [40].

3.5.6. Molecular Dynamics Simulation Analysis

The molecular dynamics (MD) simulations were thoroughly analyzed to evaluate the conformational stability, flexibility, and compactness of the protein–ligand complexes over the 100 ns trajectory. The structural parameters, including the root mean square deviation (RMSD; Figure 8a), root mean square fluctuation (RMSF; Figure 8b), and radius of gyration (Rg; Figure 8c), were computed to assess the stability and dynamic behavior of the protein backbone upon ligand binding. These analyses provide valuable insight into how top fatty acid ligands influence the conformational equilibrium and internal motions of the protein throughout the simulation period [32].

Figure 8. (a) Root mean square deviation (RMSD), (b) root mean square fluctuation (RMSF), and (c) radius of gyration (Rg) profiles of FABP7–ligand complexes during 100 ns molecular dynamics simulations for the top five XSB oil fatty acids.

The RMSD profiles revealed that all complexes reached equilibrium within the first nanoseconds and remained stable throughout the simulation, with average deviations ranging from 1.0 to 2.5 Å. Such low fluctuations indicate that the protein–ligand systems retained overall structural integrity without significant conformational drift. Among the studied complexes, those containing erucic acid (C22:1 Δ13) and nervonic acid (C24:1 Δ15) exhibited the most stable trajectories with minimal fluctuations, suggesting tight binding and enhanced structural rigidity. In contrast, the linoleic acid and eicosadienoic acid (C20:2 ω-6) complexes displayed slightly higher RMSD values, indicating moderate conformational flexibility within the binding region. The oleic acid (C18:1 Δ9) complex showed consistent stability, maintaining steady RMSD values close to 1.5 Å across the entire simulation window. These results collectively confirm that all simulated systems were dynamically equilibrated and structurally well maintained under the applied physiological conditions.

Analysis of residue-level fluctuations through RMSF further elucidated the local flexibility of the protein backbone. Most residues exhibited fluctuations of less than 2.0 Å, indicating well-preserved secondary structure throughout the simulation. Minor peaks observed in the flexible loop regions and terminal residues reflect normal backbone motions typically seen in solvent-exposed areas. The erucic acid (C22:1 Δ13) and nervonic acid (C24:1 Δ15) complexes exhibited notably lower RMSF values, indicating greater stabilization of protein residues upon ligand binding.

Conversely, the linoleic acid (C18:2 ω-6) and eicosadienoic acid (C20:2 ω-6) complexes exhibited higher fluctuations in the loop regions, suggesting local mobility induced by weaker hydrophobic or van der Waals interactions. The oleic acid (C18:1 Δ9) complex displayed intermediate behavior, maintaining moderate flexibility consistent with stable ligand association. These findings imply that the chain length and degree of unsaturation of fatty acids influence local residue mobility and interaction strength within the binding pocket. The radius of gyration (Rg) values were used to analyze the compactness of the protein–ligand complexes. All systems maintained consistent Rg values throughout the trajectory, confirming the absence of large-scale structural expansion or collapse. The erucic acid (C22:1 Δ13), nervonic acid (C24:1 Δ15), and oleic acid (C18:1 Δ9) complexes remained tightly packed, with average Rg values of 14–15 Å, indicative of stable tertiary structure. The linoleic acid complex showed mild fluctuations between 15 and 18 Å, while transient increases observed in the eicosadienoic acid (C20:2 ω-6) system may reflect short-lived conformational breathing motions of surface residues. The overall constancy of Rg values supports the RMSD and RMSF analyses, indicating that the protein retained its compact, stable conformation during ligand association.

Collectively, these MD simulation results demonstrate that all fatty acid-protein complexes are structurally stable and dynamically equilibrated under physiological conditions. The higher stability and reduced flexibility observed in complexes with longer monounsaturated fatty acids, particularly erucic and nervonic acids, suggest stronger binding affinity and more persistent protein–ligand interactions compared to polyunsaturated counterparts. These observations highlight the influence of fatty acid chain length and degree of unsaturation on the conformational dynamics and stability of the protein–ligand complexes.

4. Conclusions

Together with recent in vivo and systems-level studies, the present work reinforces the case for continued investigation of Xanthoceras sorbifolium Bunge seed oil in the context of Alzheimer’s disease. Our findings indicate that XSB oil is a promising source of unsaturated fatty acids with predicted BBB permeability and strong in silico affinity for FABP7, supporting its potential as a reservoir of FABP7-targeted lead compounds for neurodegenerative diseases. By integrating extraction optimization, antioxidant assessment, lipid profiling, ADME/T and BOILED-EGG analyses, FABP7-focused docking, molecular dynamics simulations, and network pharmacology, this study provides a coherent framework for prioritizing individual fatty acids for further investigation. Future work should experimentally validate these computational and chemical predictions in appropriate biological models and refine the pharmacological properties of the most promising candidates to advance their therapeutic potential.