An in vitro and in silico antidiabetic approach of GC–MS detected friedelin of Bridelia retusa

2.1 Identification, and authentication

The specimen was collected from Durg, Chhattisgarh, India, in April 2021 (Fig. S1) and authenticated as B. retusa (L.) A.Juss. by BSI, Allahabad, India. The voucher number I (B.S.I/C.R.C. 2020–21/200) was assigned and deposited at BSI, Allahabad.

2.2 Extraction and phytochemical screening

The bark and fruit of B. retusa were cleaned, powdered, and extracted with methanol and ethyl acetate using a Soxhlet extractor at 50–80 °C for 5–10 h. The extracts were filtered and analyzed for bioactive compounds, confirming the presence of alkaloids, cardiac glycosides, flavonoids, saponins, tannins, steroids, and terpenoids.

2.3 Phytochemical analysis by GC–MS

The samples were dissolved in methanol and ethyl acetate and injected into a Shimadzu® GC–MS QP2010 with a SH-I-5Sil MS Capillary column (30 m x 0.25 mm x 0.25 µm) in splitless mode. The oven temperature was set at 45 °C for 2 min, increased to 140 °C at 5 °C/min, then raised to 280 °C and held for 10 min. The injection volume was 2 µL, with helium at 1 mL/min flow rate and 70 eV ionization. The run time was 9.10 to 52.0 min. Compounds were identified by matching mass fragmentation patterns with the NIST 14. L library (Mallard and Linstrom, 2008).

2.4 Antioxidant activity

The DPPH assay, following Karamian et al. (2014), was used to estimate the free radical scavenging ability of B. retusa extracts. Solutions of the extracts (25–250 µg/ml) were mixed with 0.004 % methanolic DPPH, and incubated in the dark at 36 °C for 30 min, and their absorbance was recorded at 517 nm. A blank solution with methanol and DPPH was used as a control, with ascorbic acid as a reference. The IC₅₀ was calculated from the calibration curve using the linear regression equation. $$\%inhibition=\frac{A_{\mathit{con}}-A_{\mathit{test}}}{A_{\mathit{con}}}\times 100$$ where A con– absorption of the control,

A test- absorbance with samples of the extracts.

2.5 Antidiabetic activity

The α-amylase inhibition assay for B. retusa fruit and bark extracts were determined by useing the 3,5-dinitro salicylic acid method, with acarbose as a control. Absorbance was measured at 540 nm. The α-glucosidase inhibition assay was measured followed by Eom et al., (2012), with IC₅₀ calculated, both using acarbose as a standard. $$\%inhibition=\left(\mathit{Ab}{s}_{\mathit{control}}-Ab{s}_{\mathit{sample}}\right)/\left(\mathit{Ab}{s}_{\mathit{control}}\right)x100$$

2.6 Molecular docking

Friedelin, imidazole, and sylvestrene were selected for antidiabetic profiling via molecular docking, based on GC–MS analysis and literature. Their 3D structures were retrieved from PubChem, converted to PDFs using Open Babel, and were used for docking studies with diabetes-related target proteins: DPP-IV (PDB: 4A5S), α-amylase (PDB: 5KEZ), and α-glucosidase (PDB: 1UOK). Protein-ligand interactions were analyzed using UCSF Chimera, with the best binding poses determined by the lowest binding free energy. Drug-likeness was evaluated using SWISSADME, adhering to Lipinski's rule of five (Daina and Zoete, 2016; Kurjogi et al., 2018).

2.7 Molecular dynamics (MD) simulation of protein-ligand complexes

Molecular dynamics (MD) simulation was performed to assess the time-dependent stability of friedelin, imidazole, and sylvestrene complexed with DPP-IV (PDB: 4A5S), α-amylase (PDB: 5KEZ), and α-glucosidase (PDB: 1UOK). Using GROMACS-2022 with the CHARMM36 force field, ligand topology files were generated via the CGenFF server, and protein topology files through pdb2gmx. The complexes were energy-minimized and equilibrated, followed by a 100 ns MD run. Trajectories were analyzed for RMSD, RMSF, Rg, hydrogen bonds, SASA, and PCA, with the Free Energy Landscape (FEL) visualized in 2D and 3D (Hess et al., 1997; Van Der Spoel et al., 2005; Vanommeslaeghe et al., 2010; Huang and MacKerell, 2013).

2.8 ΔG_Bind calculations through MM-PBSA for protein–ligand complex

The binding-free energy (ΔG_Bind) of protein–ligand complexes (DPP-IV, α-amylase, α-glucosidase) with friedelin, imidazole, and sylvestrene were calculated using the MM-PBSA method in GROMACS. This analysis, based on trajectories from the last 10 ns of MD simulation, helps evaluation of interaction energy, conformational changes, and entropy's role in binding affinity (Van Der Spoel et al., 2005). $$\mathrm{\Delta }{G}_{\mathit{Bind}}=\mathrm{\Delta }{G}_{\mathit{Complex}}-\left(\mathrm{\Delta }{G}_{\mathit{Receptor}}+\mathrm{\Delta }{G}_{\mathit{Ligand}}\right)$$ where, ΔG_bind is the total free energy of the protein–ligand complex and ΔG_Receptor and ΔG_Ligand are the total free energies of the separated protein and ligand in a solvent, respectively.

2.9 Statistical analysis

The experiments were conducted in triplicate (n = 3), and the data were presented as the mean ± standard deviation. The IC₅₀ values were determined and calculated using linear regression analysis with the aid of Microsoft Excel 2016 software.

3.1 Phytochemical screening

GC–MS analysis of B. retusa extracts identified 96 peaks, revealing 34 compounds in methanol fruit and 38 in ethyl acetate fruit extracts. Additionally, 12 compounds from methanolic bark and 29 from ethyl acetate bark were recorded. These phytocompounds, detailed in (Table S1 to S6 & Fig. S2–S5) underscore the plant's therapeutic significance.

Tatiya et al., (2017) identified friedelin, β-sitosterol, stigmasterol, and lupeol in B. retusa bark with applications for pain and arthritis. Chitosan flavonoid was noted for analgesic and anti-inflammatory effects. Adhav et al., (2002) reported isoflavone with antiviral properties, and Kumar and Jain (2014) confirmed β-sitosterol, ellagic acid, and gallic acid as antifungal, antibacterial, and antioxidant. Additional compounds such as bisabolene sesquiterpenes, various benzoic acids, and sesamin were identified by Jayasinghe et al., (2003) and Umar et al., (2022), with potential SARS-CoV-2 inhibition. Ogbonnia et al. (2021) described other compounds in Bridelia ferruginea. Present study provides a comprehensive description of 92 compounds from B. retusa, with 88 as newly reported.

3.2 Antioxidant activity

In the DPPH assay, the ability of the extracts to donate hydrogen atoms or electrons was assessed, leading to DPPH% reduction to DPPH-H. The IC₅₀ values were determined from calibration curves and compared to ascorbic acid. The methanolic fruit extract of B. retusa showed the highest DPPH scavenging activity at 83.01 % and an IC₅₀ of 103.03 µg/ml. It was followed by the ethyl acetate fruit extract (IC₅₀: 135.66 µg/ml), ethyl acetate bark extract (IC₅₀: 185.8 µg/ml), and methanolic bark extract (IC₅₀: 206.72 µg/ml). In a similar study, the scavenging activity was reported as follows: ascorbic acid > aqueous extract > ethanol extract > methanol extract > 50 % ethanol extract > 50 % methanol extract > 70 % acetone extract, with IC50 values of 58.78 %, 62.61 %, 62.27 %, 61.94 %, 62.61 %, 70.46 %, and 61.93 %, respectively. (Tatiya et al., 2011). Murukan et al., (2018) reported that purified anthocyanin from B. retusa exhibited significant DPPH scavenging activity with an IC₅₀ of 0.31 mg/ml, ABTS+scavenging activity with an IC₅₀ of 0.55 mg/ml, and a high ferric-reducing capability (414.5 μmol Fe (II)/mg). Sanseera et al., (2016) found that chloroform, hexane, and methanol extracts of B. retusa demonstrated notable DPPH and ABTS scavenging activities. The methanol extracts of leaves, stems, and fruits had IC₅₀ values of 0.52 ± 0.031 mg/mL, 0.12 ± 0.003 mg/mL, and 0.17 ± 0.005 mg/mL for DPPH, and 0.67 ± 0.007 mg/mL, 1.41 ± 0.001 mg/mL, and 5.58 ± 0.009 mg/mL for ABTS. Ogbonnia et al. (2021) noted that various solvent fractions of Bridelia ferruginea showed higher antioxidant activities compared to the crude extract.

Our finding is affirmative to previous findings by some authors. Still, the contribution of our study is GC–MS-based recognition of a wide spectrum of phytocompounds from B. retusa which strongly supports the therapeutic application of the plant by traditional healers worldwide.

3.3 Antidiabetic activity

In the α-amylase inhibition assay, ethyl acetate bark extract showed the highest inhibition at 76.34 %. This was followed by methanolic fruit extract (74.05 % at 130.97 µg/ml), ethyl acetate fruit extract (64.31 % at 153.36 µg/ml), and methanolic bark extract (53.89 % at 181.7 µg/ml) whereas, acarbose, the standard, achieved 84.07 % inhibition (99.52 µg/ml). These results suggest the potential of B. retusa extract for antihyperglycemic activity, making it a candidate for managing hyperglycemia by inhibiting α-amylase, which can reduce post-meal blood glucose spikes (Lordan et al., 2013) (Fig. 1).

Close

Fig. 1

showing antioxidant activity of B. retusa fruit and bark extract through DPPH radical scavenging activity.

Export to PPT

In the α-glucosidase enzyme inhibition assay we found a dose-dependent inhibition activity with standard drug acarbose. The methanolic fruit extract of B. retusa exhibited maximum α-glucosidase inhibition activity with 86.18 % (106.15 μg/ml), followed by the ethyl acetate fruit extract with 78.87 % (122.93 μg/ml), ethyl acetate bark extract 77.04 % (124.01 μg/ml) and methanolic bark extract 65.20 % (151.24 μg/ml). While the standard drug acarbose revealed α-glucosidase inhibition of 88.96 % (89.85 μg/ml) (Fig. 2).

Close

Fig. 2

showing antidiabetic enzymes activities of B. retusa fruit and bark extracts. (a) α-amylase inhibition activity, (b) α-glucosidase inhibition activity.

Export to PPT

Pancreatic α-amylase and α-glucosidase enzymes break down starches and sugars into monosaccharides, crucial for glucose absorption (Ochieng et al., 2017). Inhibiting α-glucosidase can delay glucose absorption and reduce postprandial hyperglycemia, as seen with medications like acarbose and voglibose (Matsui et al., 1996). Our study found that ethyl acetate bark extract and methanolic fruit extract of B. retusa effectively inhibited α-amylase and α-glucosidase, respectively, supporting its traditional use for diabetes management.

3.4 Molecular docking

GC–MS analysis of B. retusa bark and fruit extracts identified 96 bioactive compounds (Table 1). For molecular docking, friedelin, imidazole, and sylvestrene were selected for interaction with three target proteins. Friedelin showed the most favorable binding with DPP-IV (−9.3 kcal/mol), α-amylase (−9.8 kcal/mol), and α-glucosidase (−10.3 kcal/mol), indicating superior binding affinity compared to imidazole and sylvestrene. Imidazole had the lowest affinity with binding energies of −3.6 kcal/mol, −2.7 kcal/mol, and −3.3 kcal/mol for DPP-IV, α-amylase, and α-glucosidase, respectively. Sylvestrene also showed better affinity than imidazole but less than friedelin (Table 2, 3; Figs. 3-5, Fig. S6-S8).

Table 2 Drug-likeness properties of selected ligand molecules of B. retusa.

S. No.	Drug-likeness properties	Friedelin	Sylvestrene	Imidazole
	3D structure
	2D structure
	Molecular formula	C30H50O	C10H16	C3H4N2
	Molecular weight	426.7	136.23	68.08
	LogP	9.8	3.4	−0.1
	H-bond Acceptor	1	0	1
	H-bond Donor	0	0	1
	Rotatable Bond	0	1	0
	Topological Polar Surface Area	17.1\AA2	0\AA2	28.7\AA2
	Heavy atom	31	10	5

Close

Fig. 3

3D representation of best possible pose(s) of ligands- (a) friedelin, (b) imidazole, (c) sylvestrene within the active site of the target molecule (DPP-IV).

Export to PPT

Close

Fig. 4

3D representation of best possible pose(s) of ligands- (a) friedelin, (b) imidazole, (c) sylvestrene within the active site of the target molecule (α-amylase).

Export to PPT

Close

Fig. 5

3D representation of best possible pose(s) of ligands- (a) friedelin, (b) imidazole, (c) sylvestrene within the active site of the target molecule (α-glucosidase).

Export to PPT

Friedelin interacted with 12 residues in DPP-IV (Lys71, Asn74, Ile76, Leu90, Asn92, Phe95, Asp96, Phe98, His100, Ser101, Ile102, Tyr105), 10 in α-amylase (Gln63, Tyr151, Leu162, Thr163, Leu165, Lys200, His201, Glu233, Ile235, Asp300), and 20 in α-glucosidase (Gly141, Ala142, Ala143, Leu162, Phe163, Phe203, Ser222, Gly223, His224, Phe227, Met228, Pro257, Phe281, Met284, Asp285, Lys293, Asp329, Gln330, Glu387, Lys413). Imidazole interacted with 6 residues in DPP-IV (Glu347, Met348, Ser349, Val354, Gly355, Ile375), 6 in α-amylase (Ala198, Ser199, Lys200, His201, Glu233, Val234), and 10 in α-glucosidase (Asp60, Tyr63, Phe163, Arg197, Asp199, Val200, Glu255, His328, Asp329, Arg415). Sylvestrene interacted with 12 residues in DPP-IV (Lys71, Ile76, Glu91, Phe95, Ile102, Tyr105, Ile114, Leu116, Asn75, Leu90, Asn92, Asn74), 9 in α-amylase (Trp58, Trp59, Tyr62, Gln63, His101, Leu165, Asp197, His299, Asp300), and 9 in α-glucosidase (Ala143, Leu162, Phe163, Val200, Phe203, His224, Phe227, Met228, Pro257).

Our investigation shows that the phytocompounds friedelin, imidazole, and sylvestrene from B. retusa significantly inhibit α-glucosidase, α-amylase, and DPP-IV. This suggests extract of B. retusa could aid in glucose control by slowing glucose release and improving blood sugar regulation. Molecular docking highlights friedelin as a promising candidate for developing new antidiabetic inhibitors targeting these enzymes.

Dipeptidyl peptidase IV (DPP-IV) inactivates glucagon-like peptide-1 (GLP-1) and glucose-dependent insulinotropic polypeptide (GIP), key incretin hormones. Inhibiting DPP-IV can help to regulate glucose levels in diabetes. α-Amylase, a calcium metalloenzyme, also contributes to elevated postprandial blood glucose and is a target for diabetes management (Kaur et al., 2021). Smruthi et al., (2016) found that phytocompounds from Syzygium cumini (e.g., friedelin, beta-sitosterol) interact with α-amylase with lower binding energies than acarbose. Similarly, compounds from other plants (e.g., 6 urs-12-en-24-oic acid from P. zeylanica) showed strong binding with DPP-IV and α-amylase, suggesting high potential for managing diabetes (Thamaraiselvi et al., 2021).

For α-glucosidase and α-amylase, compounds like acarbose and epigallocatechin gallate showed varying inhibitory activities, with epigallocatechin gallate as notable (Oboh et al., 2014; Tadera et al., 2006). Docking of amentoflavone, friedelin, and 6-deoxyjacareubin with porcine pancreatic elastase yielded binding energies of −10.94, −7.17, and −6.72 kcal/mol, respectively, with friedelin showing strong interactions, including a Pi-sigma bond with His57 (Ambarwati et al., 2022). Muralikrishna et al. (2019) found that beta-sitosterol and friedelin from Ficus racemose had favorable interactions with hyperglycemic targets, with friedelin forming hydrogen bonds with Lys776. Drug-likeness properties of B. retusa compounds were assessed using SwissADME based on Lipinski’s rule of five (Daina and Zoete, 2016).

Sylvestrene and imidazole met all Lipinski criteria viz. molecular weight (≤500), hydrogen bond acceptors (≤10), donors (≤5), and LogP (<5). However, friedelin exceeded the LogP threshold, indicating high lipophilicity and potential issues with absorption (Bahmani et al., 2017). Despite this, all compounds showed favorable molecular weight and TPSA values, suggesting good absorption and permeability. Friedelin, along with sylvestrene and imidazole, exhibited promising antioxidant and antidiabetic activities, with friedelin standing out as the most effective candidate for diabetes therapy.

3.5 Molecular docking simulation analysis

MD simulations provide insights into protein–ligand interactions and ligand stability. We analyzed nine complexes: DPP-IV-Sylvestrene, DPP-IV-Imidazole, DPP-IV-Friedelin, α-amylase-Sylvestrene, α-amylase-Imidazole, α-amylase-Friedelin, α-glucosidase-Sylvestrene, α-glucosidase-Imidazole, and α-glucosidase-Friedelin (Fig. S9–S11). For stability assessment, RMSD and RMSF were monitored. RMSD analysis (Fig. S9a–c) showed that DPP-IV-Sylvestrene had the most stable RMSD (0.2–0.3 nm), while DPP-IV-Imidazole and DPP-IV-Friedelin had higher deviations. The α-amylase-Friedelin complex was the most stable (0.15–0.25 nm) compared to the other α-amylase complexes. For α-glucosidase, α-glucosidase-Friedelin was the most stable, with RMSD values of 0.15–0.35 nm initially, decreasing to 0.27–0.33 nm later. Higher RMSF values were observed for the imidazole complex, indicating greater fluctuation compared to sylvestrene and friedelin (Fig. S9d–f).

Radius of gyration (Rg) analysis measures the compactness of protein–ligand complexes, reflecting structural stability and integrity. Elevated Rg values suggest decreased compactness and potential structural instability, while lower values indicate a more stable structure. For DPP-IV protein complexes, Rg values ranged from 2.7 to 2.8 nm for all three drugs (Fig. S10a). For α-amylase, Rg values were between 2.32 and 2.4 nm for all drugs (Fig. S10b). For α-glucosidase, Rg values ranged from 2.42 to 2.57 nm, with the α-glucosidase-Friedelin complex showing the lowest Rg values of 2.43 to 2.47 nm (Fig. S10c).

Intermolecular hydrogen bond (H-Bond) analysis was performed for DPP-IV, α-amylase, and α-glucosidase with sylvestrene, imidazole, and friedelin. For DPP-IV, imidazole formed a maximum of 2 hydrogen bonds during 40–50 ns and 80–90 ns, while friedelin formed 1 hydrogen bond from 1 to 93 ns but none later, and sylvestrene formed no hydrogen bonds (Fig. S10d). For α-amylase, no hydrogen bonds were observed with sylvestrene, and friedelin formed a few single bonds around 20–30 ns and 45 ns. The α-amylase-imidazole complex formed 3 hydrogen bonds initially and exhibited intermittent bonding throughout the simulation (Fig. S10e). For α-glucosidase, sylvestrene formed no hydrogen bonds. Friedelin and imidazole formed 1–3 hydrogen bonds intermittently during the simulation (Fig. S10e).

SASA analysis was performed to evaluate the protein surface area exposed to water, indicating the extent of drug interaction. For DPP-IV, SASA values ranged from 320 to 350 nm² (Fig. S11a). For α-amylase, SASA values were lower, between 185 to 225 nm² (Fig. S11b), suggesting a stronger drug interaction. For α-glucosidase, SASA values ranged from 230 to 270 nm² (Fig. S11c). The lower SASA values for α-amylase indicate a more stable drug interaction due to reduced protein surface exposure to water.

PCA analysis was used to assess protein–ligand complex stability by examining eigenvector variances. For DPP-IV, the sylvestrene complex had the least variance, with eigenvector values ranging from −3 to 3 nm (Fig. S11d), compared to −20 to 20 nm for imidazole and −60 to 40 nm for friedelin. For α-amylase, eigenvector values were −27 to 15 nm (sylvestrene), −16 to 18 nm (imidazole), and −10 to 6 nm (friedelin) (Fig. S11e), with the α-amylase-friedelin complex showing the least conformational changes. For α-glucosidase, eigenvector values ranged from −27 to 30 nm (sylvestrene), −18 to 17 nm (imidazole), and −4 to 4 nm (friedelin) (Fig. S11f). The best PCA values were observed for the α-glucosidase-friedelin complex, followed by α-amylase-friedelin and DPP-IV-sylvestrene complexes.

3.6 Binding free energy calculations of top drugs complexed with anti-diabetic target proteins

MM-PBSA free energy calculations were performed on the complexes of sylvestrene, imidazole, and friedelin with DPP-IV, α-amylase, and α-glucosidase, using MD trajectories from the last 10 ns (90–100 ns). The analysis highlighted van der Waals, electrostatic, and polar solvation energies as key stabilizers (Table 4). The maximum ΔG_Bind values were −60.457 ± 11.215 kJ/mol for DPP-IV-sylvestrene, −114.467 ± 8.982 kJ/mol for α-amylase-friedelin, and −102.904 ± 15.329 kJ/mol for α-glucosidase-friedelin. Among these, the α-amylase-friedelin complex exhibited the highest ΔG_Bind.

Molecular dynamics simulations revealed stability and structural dynamics in nine protein–ligand complexes. Friedelin showed the most stable interactions with DPP-IV, α-amylase, and α-glucosidase, indicated by lower RMSD and RMSF values. SASA analysis highlighted strong interactions, particularly with α-amylase. PCA revealed distinct conformational changes for each complex. MM-PBSA calculations highlighted van der Waals, electrostatic, and polar solvation energies as key stabilizers, with the α-amylase-friedelin complex showing the highest binding energy.

Free energy landscape (FEL) analysis for DPP-IV, α-amylase, and α-glucosidase complexes with imidazole, friedelin, and sylvestrene is shown in Figs. 6, 7, and 8. The 2D FEL plots (Fig. 6(a–b), 7(a–b), 8(a–b)) use red for maximum Gibbs free energy and blue for minimum Gibbs free energy, while the 3D FEL plots use red and dark green similarly. For DPP-IV, the least energy conformations occurred at 62.5 ns (imidazole), 44 ns (friedelin), and 42.5 ns (sylvestrene) (Fig. 6(c)). For α-amylase, they were at 70 ns (sylvestrene), 68 ns (imidazole), and 89.3 ns (friedelin) (Fig. 7(c)). For α-glucosidase, the least energy conformations were at 95.6 ns (imidazole), 73.5 ns (friedelin), and at 50 ns and 67 ns (sylvestrene) (Fig. 8(c)).

Close

Fig. 6

showing 2-D, 3-D principal component analysis along with free energy landscape plot for DPP-IV protein complexed with ligands imidazole, friedelin, and sylvestrene.

Export to PPT

Close

Fig. 7

showing 2-D, 3-D principal component analysis along with free energy landscape plot for alpha-amylase protein complexed with ligands imidazole, friedelin, and sylvestrene.

Export to PPT

Close

Fig. 8

showing 2-D, 3-D principal component analysis along with free energy landscape plot for alpha-glucosidase protein complexed with ligands imidazole, friedelin, and sylvestrene.

Export to PPT