Sage Journals: Discover world-class research

Abstract

Erythrina variegata L., a member of the Fabaceae family, is a valuable medicinal plant with a rich history of traditional medicinal uses. However, its potential as an antidiabetic agent has remained largely unexplored. In this study, three isoflavonoid compounds, namely eryvarin M (1), eryvarin H (2), and neobavaisoflavone (3), were isolated from E. variegata. These compounds displayed significant inhibitory effects on both protein tyrosine phosphatase 1B and α-glucosidase enzymes. The specific attachment position of the prenyl group to the isoflavonoid skeleton plays a pivotal role in determining the variation in their activity. Moreover, the results obtained from docking simulations corroborate the experimental findings, confirming the robust binding affinities of these metabolites toward the target proteins. The docking analysis further highlights that these compounds exhibit highly negative values of binding-free energies against proteins, indicative of their favorable binding affinities. In addition, the formation of hydrogen bonds in the binding region contributes to their binding interactions. These findings significantly enhance our understanding of the therapeutic potential of both E. variegata and its isoflavonoids as potential inhibitors for diabetes and related metabolic disorders, shedding light on their promising role in the field of diabetes research.

Keywords

α-glucosidase inhibitor isoflavonoid molecular docking protein tyrosine phosphatase 1B Vong nem

Introduction

Erythrina variegata L. var. orientalis (L.) Merr., also known by synonyms, such as Erythrina indica Lamk, Erythrina spathacea DC, and Erythrina orientalis (L.) Murr, belongs to the Fabaceae family.^1,2 The name “Erythrina” is derived from the Greek word “erythros” which means “red” referring to the striking red color of this plant species flowers.³ E. variegata is indigenous to the coastal regions of East Africa and is distributed across Africa, tropical regions of South Asia, Southeast Asia, Pacific Islands, Australia, and Southern China. In Vietnam, E. variegata grows naturally and can be found scattered throughout lowland mountainous provinces, the central region, and the delta. It is commonly cultivated as a hedge plant in fields and home gardens. E. variegata tree holds various medicinal uses in traditional medicine. It is regarded as a nerve sedative, an ingredient in ophthalmic eye drops, an anti-asthmatic agent, an anticonvulsant, an antiseptic, and an astringent.^3–5 Traditional literature suggests that E. variegata leaves possess drying and calming properties and are associated with the heart and spleen meridians. They are known for their soothing and antiseptic effects.⁶ In Vietnam, people frequently consume E. variegata leaves and apply heated leaves to treat hemorrhoids. The leaves are also used for treating insomnia and chronic colitis. A combination of E. variegata leaves and lotus leaves is employed to address nosebleeds and bloody stools, and to treat infantile diarrhea.⁷

E. variegata has been found to contain more than 100 compounds, including alkaloids, flavonoids, pterocarpan, triterpenes, steroids, alkyl trans-ferulate, proteins, and lecithin. Alkaloids and flavonoids are the two major class of compounds present in different parts of the plant.^6,8–14 Currently, many studies have isolated flavonoids from E. variegata with isoflavonoids reported as the primary constituents among prenyl isoflavones, flavan-3-ols, flavonols with sugar moieties, and triterpenoids.^15–32 Given its rich chemical composition, several studies have investigated the plant’s biological activities, such as antibacterial effects,^22,33–35 antioxidant activity,^5,36 pain relief, anti-inflammatory effects,^36–39 cardiovascular effects,⁴⁰ anticonvulsant effects, and anxiolytic effects.^{3,5,8,41–44} However, there is limited research on the antidiabetic effects of this plant. In a study by Santhiya et al.⁴⁵ in 2016, the E. variegata L. bark methanolic extract showed in vitro stimulatory effects on glucose uptake assay by isolated rat hemidiaphragm and glucose uptake by Yeast cells, and inhibitory effects on α-amylase and α-glucosidase enzymes as well. In addition, methanol extract of E. variegata leaf significantly and dose dependently reduced and normalized blood glucose levels in streptozotocin (STZ)-induced diabetic Wistar rats.⁴⁶

Binding of insulin to its receptors leads to the phosphorylation of insulin-sensing substrates, which triggers the activation of a series of intracellular signaling pathways that lead to a cascade of biochemical reactions including glucose transport into cells and glycogen synthesis.⁴⁷ Protein tyrosine phosphatase (PTP) together with protein tyrosine kinases are involved in the regulation of a variety of cellular functions, including proliferation, differentiation, and apoptosis. PTPs dephosphorylate tyrosine residues of proteins and are considered as negative regulators (back correctors) of insulin signaling. Among the various PTPs, protein tyrosine phosphatase 1B (PTP1B) plays an important role in the regulation of body weight and glucose homeostasis by acting as a negative regulator of the insulin and leptin signaling pathway, thus has attracted considerable attention from scientists as a potential target for research in the treatment of diabetes and metabolic disorders.⁴⁸ Overexpression of PTP1B enzyme leads to the inhibition of the insulin receptors signaling cascade, and increased expression of this protein leads to insulin resistance,⁴⁹ while in a PTP1B, knockdown mice showed increased insulin sensitivity and enhanced resistance to obesity.⁵⁰ Therefore, active substances capable of inhibiting PTP1B enzyme activity not only have great potential in researching and finding drugs to treat type 2 diabetes but also effective agents in the treatment of obesity.^51–55

In our study, the extraction and isolation of isoflavonoids (1–3) with their potential to inhibit PTP1B and α-glucosidase enzymes using the bioguided extraction and isolation method are described. In addition, computational modeling methods were employed to analyze the binding interactions and binding energies of these compounds with targeted proteins.

Results and discussion

Chemistry

The ¹H nuclear magnetic resonance (NMR) spectrum of compounds 1 and 3 displayed characteristic signals attributed to aromatic protons of two benzene rings, along with those corresponding to hydroxyl groups (C-7 and C-4′), and methoxy groups. The ¹³C NMR spectrum revealed signals from oxygen-substituted carbons and two quaternary carbons (C-10 and C-1′). Compound 1 exhibited the presence of methylene and methine groups [δ_H 4.56 (1H, dd, J = 10.8, 12.0 Hz, H-2a), 4.17 (1H, dd, J = 5.4, 12.0 Hz, H-2b)/δ_C 71.7 (C-2) and δ_H 4.42 (1H, dd, J = 5.4, 10.8 Hz, H-3)/δ_C 48.2 (C-3)], while compound 3 contained an olefinic group [δ_H 8.10 (1H, s, H-2)/δ_C 153.1 (C-2), and δ_C 125.4 (C-3)], and a ketone carbon [δ_C 191.7 (C-4) for 1 and δ_C 175.4 (C-4) for 3] in both the ¹H NMR and ¹³C NMR spectra, indicating that both compounds are isoflavonoids.

In-depth analysis of the ¹H NMR and ¹³C NMR spectra of compound 1 revealed the presence of five aromatic proton signals [δ_H 7.75 (1H, d, J = 8.4 Hz, H-5), 6.54 (1H, dd, J = 2.1, 8.4 Hz, H-6), 6.40 (1H, d, J = 2.1 Hz, H-8), 6.56 (1H, s, H-3′), 6.78 (1H, s, H-6′)], two methoxy groups [δ_H 3.70 (3H, s, 2′-OCH₃), 3.72 (3H, s, 5′-OCH₃)/ δ_C 56.4 (2′-OCH₃), 57.1 (5′-OCH₃)], and five oxygenated carbons (C-7, C-9, C-2′, C-4′, and C-5′). Compound 3 featured six aromatic protons [δ_H 8.04 (1H, d, J = 8.7 Hz, H-5), 6.98 (1H, dd, J = 3.0, 8.7 Hz, H-6), 6.88 (1H, d, J = 3.3 Hz, H-8), 7.34 (1H, d, J = 2.4 Hz, H-2′), 6.87 (1H, d, J = 8.4 Hz, H-5′), 7.27 (1H, dd, J = 2.4, 8.4 Hz, H-6′)], but displayed three oxygenated carbons (C-7, C-9, and C-4′), along with a prenyl group [δ_H 3.25 (1H, d, J = 7.5 Hz, H-1′′), 5.36 (1H, m, H-2′′), 1.72 (3H, s, 4′′-CH₃), 1.70 (3H, s, 5′′-CH₃)/ δ_H 29.0 (C-1′′), 118.6 (C-2′′), 132.3 (C-3′′), 25.9 (C-4′′), 17.8 (C-5′′)]. Comparison of the ¹H NMR and ¹³C NMR data of these compounds with literature references led to the identification of the structures of 1 as eryvarin M and 3 as neobavaisoflavone.^56,57

Compound 2 was obtained in the form of a light-yellow powder. Its ¹H NMR spectrum exhibited signals corresponding to two trisubstituted benzene rings [δ_H 6.88 (1H, d, J = 8.0 Hz, H-5), 6.34 (1H, dd, J = 2.0, 8.0 Hz, H-6), 6.26 (1H, d, J = 2.0 Hz, H-8), 6.85 (1H, s, H-6′), 6.50 (1H, s, H-3′)] and two methoxy groups [δ_H 3.81 (3H, s, 5′-OCH₃), 3.72 (3H, s, 2′-OCH₃)] (Figure 1). In the ¹³C NMR spectrum, 17 carbon signals were observed, attributed to 12 sp² carbons from the 2 benzene rings, and 2 methoxy carbons [δ_C 57.5 (5′-OCH₃), 56.6 (2′-OCH₃)]. The presence of an oxygenated methylene group [δ_H 4.87 (2H, s, H-2)/ δ_C 69.6 (C-2)] and an olefinic group [δ_H 6.47 (1H, s, H-4)/ δ_C 130.3 (C-3), 122.0 (C-4)] in both the ¹H NMR and ¹³C NMR spectra confirmed the identity of compound 2 as an isoflav-3-ene (Figure 1). Based on this evidence, compound 2 was identified as eryvarin H.⁵⁸

Figure 1.

Chemical structure of compounds (1–3) isolated from E. variegata.

Biology

Compounds (1–3) underwent in vitro testing to evaluate their inhibitory effects on PTP1B enzyme activity. The results presented in Table 1 demonstrate that all tested compounds exhibited potential inhibition against this enzyme, with IC₅₀ values of 20.31 ± 0.51, 8.03 ± 0.16, and 9.23 ± 0.29 µM, respectively. Ursolic acid (UA), used as a positive control, displayed an IC₅₀ value of 3.31 ± 0.19 µM in this assay. Among these isoflavonoids, eryvarin M (1) and eryvarin H (2) contain two methoxy groups at positions C-2′ and C-5′ (Figure 1) while neobavaisoflavone (3) only bears a prenyl moiety at C-3′. Compound 2, an isoflav-3-ene type, exhibited the most potent inhibition against PTP1B with an IC₅₀ of 8.03 ± 0.16 µM, while compound 1, a basic isoflavanone skeleton, displayed weaker inhibitory effect with an IC₅₀ value of 20.31 ± 0.51 µM. As a prenylated isoflavone, compound 3 showed significant inhibitory activity with an IC₅₀ value of 9.23 ± 0.29 µM. This observation might suggest that the isoflav-3-ene type may exert stronger inhibitory activity toward PTP1B enzyme compared to the isoflavanone type.

Table 1.

In vitro inhibitory activity of compounds (1–3) on PTP1B and α-glucosidase enzymes.

Compounds	PTP1B (IC₅₀, μM)^a	α-glucosidase (IC₅₀, μM)^a
Eryvarin M (1)	20.31 ± 0.51	142.50 ± 0.30
Eryvarin H (2)	8.03 ± 0.16	68.96 ± 0.74
Neobavaisoflavone (3)	9.23 ± 0.29	72.67 ± 0.92
Ursolic acid^b	3.31 ± 0.19	–
Acarbose^b	–	123.54 ± 0.29

Results are expressed as IC₅₀ values (µM), determined by regression analysis and expressed as the means ± SD of three replicates.

Positive control.

A study by Bae et al.⁵⁹ in 2006 reported that the both prenylated isoflavanones, orientanol E and 2,3-Dihydroauriculatin, exhibited strong PTP1B inhibitory activities with IC₅₀ values of 10.1 ± 0.3 and 2.6 ± 0.5 μM, respectively. In contrast, eryvarin M (1), which lacks a prenyl moiety, showed lower activity compared to the aforementioned compounds. In addition, warangalone, an isoflavone with a prenyl moiety, demonstrated moderate activity against PTP1B, with a reported IC₅₀ value of 42.5 ± 2.8 μM, whereas senegalensis and warangalone 4′-methyl ether did not exhibit inhibitory activity (IC₅₀ > 150 μM). Neobavaisoflavone (3), an isoflavone with a prenyl moiety at C-3′, displayed significant inhibitory activity (IC₅₀ value of 9.23 ± 0.29 μM) when compared to these reported compounds. Furthermore, several non-prenylated isoflavones showed moderate inhibitory activity against the PTP1B enzyme. Alpinumisoflavone and parvisoflavone B had inhibitory IC₅₀ values of 41.5 ± 2.4 and 42.6 ± 2.4 μM on PTP1B, respectively.^60,61 Glisoflavone exhibited a remarkable inhibitory effect with an IC₅₀ value of 27.9 ± 1.4 μM, while glycyrrhisiflavone and cycloglyrrhisoflavone did not show significant activity (IC₅₀ > 50 μM). A report in 2012 noted that most isoflavones and isoflavanones bearing a prenyl group displayed good inhibitory activity against PTP1B, except for erythraddison I which showed no activity (IC₅₀ > 50 μM). This lack of activity was attributed to the decomposition of the prenyl moiety due to the hydroxylation at C-1 and C-2, and the methoxylation at the C-3 position as well.⁶² In 2015, Guo et al.⁶³ reported several non-prenylated isoflavones, including genistein, formononetin, and glabrone, did not exhibit in vitro inhibitory effects on the PTP1B enzyme (IC₅₀ > 100 μM). However, some isoflavones, such as auriculasin, 6,8-Diprenylorobol, 8-γ,γ-Dimethylallylwighteone, and osajin, which contained one or two prenyl moieties in their structure, were reported to have potential inhibitory activity, their IC₅₀ values ranging from 2.4 to 17.4 μM.⁶⁴ Compound 3 represents the fundamental structural skeleton of an isoflavone, featuring a prenyl moiety at the C-3′ position, exhibits significant inhibition on PTP1B enzyme. Based on both experimental bioassay results and the above analysis data, it is strongly suggested that the presence of a prenyl moiety plays a pivotal role in enhancing the inhibitory activity of isoflavones against the PTP1B enzyme. However, the specific attachment position of the prenyl group is a critical determinant of this activity variation. Notably, this study also marks the first report of the inhibitory activity of eryvarin H (2), an isoflav-3-ene type compound, against both α-glucosidase and PTP1B enzymes.

Molecular docking study

PTP1B contains two different sections, the catalytic site and the allosteric site, both of which are crucial to the enzyme’s operation and control. The core domain of the enzyme occupies PTP1B’s catalytic site, which dephosphorylates tyrosine residues on the target protein. The allosteric site is situated on the C-terminal domain, about 20 residues from the catalytic site, and is surrounded by helices 3, 6, and 7 (Figure 2(a)). The allosteric site is significantly less polar and poorly conserved among PTPs, in contrast to the active site of PTP1B.⁶⁵ That means, to potentially achieve PTP specificity, targeting PTP1B through its allosteric site, which is distinct from other PTPs, is a viable strategy. Therefore, focusing on the allosteric site may present a viable strategy for creating PTP1B inhibitors that have higher bioavailability and selectivity. In this study, three isoflavonoids (1–3) isolated from E. variegata were subjected to in silico studies to investigate the binding modes at both catalytic site and allosteric site of PTP1B. The docking score is shown in Table 2 and binding modes are shown in Figures 3 and 4.

Figure 2.

Docking site of protein. (a = upper) Two crystal proteins of PTP1B (1BZC and 1T4J) was aligned with two crystal ligands at catalytic site and allosteric site. (b = under) α-glucosidase with crystal ligand.

Table 2.

Molecular interaction of the protein PTP1B.

Compound	Binding affinities (kcal mol⁻¹)		Catalytic site		Allosteric site
Compound	Catalytic site	Allosteric site	H-bond interacting residues	Hydrophobic interacting residues	H-bond interacting residues	Hydrophobic interacting residues
Eryvarin M (1)	−6.484	−5.410	Arg47, Lys120	Tyr46, Ala217, Asp48, Arg221	Arg199, Gln288, Ile281	Phe196
Eryvarin H (2)	−7.437	−6.065	Arg221, Cys215, Lys120	Tyr46, Ala217	Gln288, Arg79	Phe196
Neobavaisoflavone (3)	−7.811	−6.091	Lys120, Arg221	Tyr46, Ala217	Arg199	Gln288, Ile281, Phe196
Reference at catalytic site	−8.982	−	Gly218, Gly220, Ser216, Ala217, Arg47, Asp48, Cys215, Ile219	Arg221, Phe182, Tyr46	–	–
Reference at allosteric site	–	−6.822	–	–	Arg79, Ser285, Ile281	Glu200, Asp236, Ile281, Arg199

Figure 3.

2D ligand interaction diagram and binding modes for the PTP1B at catalytic site. (a) Eryvarin M. (b) Eryvarin H. (c) Neobavaisoflavone.

Figure 4.

2D ligand interaction diagram and binding modes for the PTP1B at allosteric site. (a) Eryvarin M. (b) Eryvarin H. (c) Neobavaisoflavone.

Herein, the co-ligand in PTP1B (1BZC), 4-carbamoyl-4-{[6-(difluoro-phosphono-methyl)-naphthalene-2-carbonyl]-amino}-butyric acid, showed eight hydrogen bonds at Gly218, Gly220, Ser216, Ala217, Arg47, Asp48, Cys215, and Ile219 and three hydrophobic interactions with Arg221, Phe182, and Tyr46. Figure 3 shows the bidding pose and dynamics of co-ligand 1BZC. The phosphate group in this compound was found to form hydrogen bond with Cys215, the key residues in the activity of PTP1B.^66,67 Compared to our three compounds, eryvarin M (1) displayed interaction with two hydrogen bonds and four hydrophobic interactions (−6.484 kcal mol⁻¹). Meanwhile, eryvarin H (2) and neobavaisoflavone (3) showed stronger binding affinity (−7.437 and −7.811 kcal mol⁻¹, respectively). Notably, eryvarin H (2) was found to create hydrogen bond with Cys215 residue. Neobavaisoflavone (3) also showed interactions like reference ligand, at Lys120, Arg221, Tyr46, and Ala217.

To compare the interaction between compounds and PTP1B at allosteric site, co-ligand in PTP1B (1T4J), 3-(3,5-dibromo-4-hydroxy-benzoyl)-2-ethyl-benzofuran-6-sulfonic acid [4-(thiazol-2-ylsulfamoyl)-phenyl]-amide, was docked. This reference compound showed three hydrogen bonds at Arg79, Ser285, and Ile281 and four hydrophobic interactions at Glu200, Asp236, Ile281, and Arg199. Figure 4 shows that eryvarin M (1) displayed the weakest binding affinity (−5.410 kcal mol⁻¹) and the lowest potency among three isolated compounds with three hydrogen bonds and four hydrophobic interactions. Eryvarin H (2) and neobavaisoflavone (3) showed stronger binding affinity (−6.065 and −6.091 kcal mol⁻¹, respectively). From these results, docking scores are consistent with our in vitro results (Tables 1 and 2). These also indicated that the isoflavonoids (1–3) are likely to bind to catalytic site of PTP1B.

The deep gap that separates the two subunits where α-glucosidase’s active site is located has numerous conserved residues that are involved in substrate binding and catalysis. In this study, the crystal ligand, acarbose, was selected as reference compound. Acarbose showed four hydrogen bonds between residues Ser676, Ser679, Asp282, and Asp616 and sugar moieties of acarbose (Table 3). Compared to the crystal ligand, eryvarin M (1) and eryvarin H (2) had similar binding affinities, while neobavaisoflavone (3) displayed the strongest binding energy (−7.756 kcal mol⁻¹). At the allosteric site, neobavaisoflavone (3) was found to interact with α-glucosidase with two hydrogen bonds at Asp404 and Arg600 and five hydrophobic interactions at Phe525, Trp481, Asp282, Asp616, and Phe649 (Figure 5C). Like neobavaisoflavone (3), eryvarin M (1) also showed two hydrogen bonds at these sites (Figure 5A). Meanwhile, eryvarin H (2) has only one hydrogen bond with Arg600 residue (Table 3 and Figure 5B).

Table 3.

Molecular interaction of the protein α-glucosidase.

Compound	Binding affinities (kcal mol⁻¹)	H-bond interacting residues	Hydrophobic interacting residues
Eryvarin M (1)	−6.698	Asp404, Arg600	Asp282, Phe525, Phe649, Asp616, Leu650
Eryvarin H (2)	−6.685	Arg600	Trp376
Neobavaisoflavone (3)	−7.756	Asp404, Arg600	Phe525, Trp481, Asp282, Asp616, Phe649
Reference at catalytic site	−6.738	Ser676, Ser679, Asp282, Asp616

Figure 5.

2D ligand interaction diagram and binding modes for the α-glucosidase. (a) Eryvarin M. (b) Eryvarin H. (c) Neobavaisoflavone.

Structurally, the hydroxyl groups and heterocyclic oxygen in flavonoid framework play an important role in strong interactions with protein. In this case study, compared to eryvarin M (1), the removing carbonyl group or adding a functional group at the C-3′ position in two compounds eryvarin H (2) and neobavaisoflavone (3) indicated the better binding affinities and inhibitory activity. Based on this evidence, further studies are needed to optimize the structure of flavonoids to promote their ability to inhibit the activity of enzymes PTP1B and α-glucosidase, which play a role in the treatment of diabetes.

For accessing the druggability of potential medicinal compounds, the “Lipinski’s rule of five” has long been used. According to this rule, there are factors taken into account, such as molecular weight, the number of hydrogen bond donors and acceptors, and log P. In this study, the three isolated isoflavonoids including eryvarin M (1), eryvarin H (2), and neobavaisoflavone (3) from E. variegata have their molecular weights smaller than 500 Daltons. The number of hydrogen bond donors and hydrogen bond acceptors of these compounds are all smaller than 5 and 10, respectively. The log P values are also smaller than 5. Hence, there was no violation to the classic Lipinski’s rule,⁶⁸ indicating favorable drug-like properties of the studied compounds. In addition, the number of rotatable bonds, total polar surface area (TPSA), and aqueous solubility (log S) were also calculated as an assessment of physicochemical parameters. For good oral bioavailability and intestinal absorption, the number of rotatable bonds and the TPSA value should not exceed 10 and 140 Å², respectively.⁶⁹ In case of investigated compounds, their number of rotatable bonds and TPSA values fit within the favorable ranges, suggesting good physicochemical properties. The log S values of (1) and (2) are −3.44 and −3.58, respectively, suggesting that they are soluble while the remaining compound is moderately soluble with the log S value of −4.91. The physicochemical properties of the studied compounds are comprehensively presented in Table 4.

Table 4.

Physicochemical properties of isoflavonoids (1–3).

Compound	MW(g mol⁻¹)	Log P	nHBD	nHBA	TPSA	MR	Log S	nRotB
Eryvarin M (1)	316.31	2.08	2	6	85.22	82.53	−3.44	3
Eryvarin H (2)	300.31	2.57	2	5	68.15	83.22	−3.58	3
Neobavaisoflavone (3)	322.35	3.74	2	4	70.67	95.69	−4.91	3

MW: molecular weight; log P: log of octanol/water partition coefficient; nHBD: number of hydrogen bond donor(s); nHBA: number of hydrogen bond acceptor(s); TPSA: total polar surface area; MR: molar refractivity; log S: log of solubility; nRotB: number of rotatable bond(s).

In silico ADME predictions of the three studied compounds are exhibited in Table 5. The log of skin permeability (log K_p) values ranges from −5.14 to −6.60. These isoflavonoids (1–3) were all predicted to be well absorbed in the gastrointestinal and permeating the blood–brain barrier. Several cytochrome-P enzymes play a vital role in drug biotransformation, including CYP1A2, CYP2C19, CYP2C9, CYP2D6, and CYP3A4. All three compounds were predicted to inhibit CYP1A2 and CYP3A4, but not to inhibit CYP2C19. They were also suggested to be a substrate of P-glycoprotein (P-gp). DL-AOT Prediction Server was used to calculate the LD₅₀ values (mg kg⁻¹) of eryvarin M (1), eryvarin H (2), and neobavaisoflavone (3), which are 2.90, 3.29, and 3.24, respectively. Based on the predicted results (Table 6), eryvarin H (2) and neobavaisoflavone (3) were classified into “None required” and “Caution” group, while eryvarin M (1) might bring harm to the body since this compound was ranked “Warning.”

Table 5.

ADME prediction.

Compound	Log K_p (cm s⁻¹)	GI Abs	BBB per	Inhibitor interaction
				P-gp substrate	CYP1A2 inhibitor	CYP2C19 inhibitor	CYP2C9 inhibitor	CYP2D6 inhibitor	CYP3A4 inhibitor
Eryvarin M (1)	−6.60	High	No	Yes	Yes	No	No	No	Yes
Eryvarin H (2)	−6.24	High	Yes	Yes	Yes	No	No	Yes	Yes
Neobavaisoflavone (3)	−5.14	High	Yes	No	Yes	No	Yes	No	Yes

ADME: Absorption, Distribution, Metabolism, and Excretion; Log K_p: log of skin permeability; GI Abs: Gastrointestinal absorption; BBB Per: blood–brain barrier permeability; P-gp: P-glycoprotein; CYP: cytochrome-P.

Table 6.

Toxicity of isoflavonoids (1–3) from E. variegata predicted by DL-AOT prediction server.

Compound	LD₅₀ (mg kg⁻¹)	Toxicity
Eryvarin M (1)	2.90	Warning
Eryvarin H (2)	3.29	None required
Neobavaisoflavone (3)	3.24	Caution

Conclusion

In conclusion, isoflavanone (1), isoflav-3-ene (2), and neobavaisoflavone (3) were identified for the first time in the E. variegata. These compounds (1–3) exhibited notable inhibitory effects on both PTP1B and α-glucosidase enzymes, as evidenced by their respective IC₅₀ values of 20.31 ± 0.51, 8.03 ± 0.16, and 9.23 ± 0.29 μM against PTP1B, and 142.50 ± 0.30, 68.96 ± 0.74, and 72.67 ± 0.92 μM against α-glucosidase enzyme. This experimental bioassay data, coupled with a thorough analysis of existing literature, underscore the significant role of the prenyl moiety in enhancing the inhibitory activity of isoflavones against the PTP1B enzyme. However, it is crucial to note that the specific attachment position of the prenyl group to the isoflavonoid skeleton is a key determinant of this activity variation. Notably, eryvarin H (2), representing the isoflav-3-ene type compound, has been reported to possess inhibitory activity against both α-glucosidase and PTP1B enzymes, marking an important discovery in this study.

Moreover, the results from docking simulations align well with the experimental findings, confirming the strong binding affinities of isoflavanone (1), isoflav-3-ene (2), and neobavaisoflavone (3) toward the target proteins. The docking analysis further reveals that these compounds exhibit highly negative values of binding-free energies against proteins, signifying their favorable binding affinities. Especially, the formation of hydrogen bonds, particularly with the amino acid Leu204 in the Site 3 binding region, is suggested to play a crucial role in the development of potential compounds targeting PTP1B. This study provides valuable insights into the inhibitory properties of isoflavanone (1), isoflav-3-ene (2), and neobavaisoflavone (3) against PTP1B and α-glucosidase enzymes, using a combination of experimental and computational approaches. These findings contribute significantly to our understanding of the therapeutic potential of these isoflavones as inhibitors for diabetes and related metabolic disorders.

Experimental

Chemistry

The stem bark of E. variegata was gathered in Hanoi, Vietnam. Dr Nguyen Quoc Binh conducted the botanical authentication of the sample, and a voucher specimen (VN01HN) was formally deposited at the Department of Chemical Analysis, Institute of Natural Products Chemistry, Vietnam Academy of Science and Technology (VAST).

General procedure for the isolation of compounds

The isolation process begins with the stems of E. variegata. Impurities are meticulously removed, and the stems are finely chopped, thoroughly washed, and then dried at a temperature of approximately 50 °C until the moisture content reaches around 10%. Subsequently, the dried sample is ground into a powder with a particle size of approximately 0.2 mm, yielding 2 kg of E. variegata stem powder. This powder is then subjected to three consecutive extractions using n-hexane solvent at a ratio of 1/3 (weight/volume) with the assistance of an ultrasonic apparatus, conducted at 35 °C for a duration of 90 min. Following this process, the solvent is separated, and the remaining powder residue is obtained. The powder residue undergoes further extraction with ethyl acetate (EtOAc) solvent at a ratio of 1/5 (weight/volume) using the same ultrasonic apparatus, operated at 35 °C for a duration of 120 min. The resulting extract is then filtered, and the solvent is removed via vacuum distillation, ultimately yielding a total EtOAc extract (115.6 g). This total EtOAc extract is subsequently loaded onto a normal-phase silica gel chromatography column, featuring particle sizes ranging from 63 to 200 μm. Elution is carried out using a solvent system consisting of n-hexane/acetone. Initially, elution is performed with a polarity ratio of 10/1, which is gradually increased to 1/10 (volume/volume). This multi-step process yields 15 fractions, which are denoted as EVF.1–EVF.15. Fraction EVF.12 is further subjected to chromatography on an RP-C18 column (4.0 × 60 cm; 150 μm particle size). A stepwise gradient of MeOH–H₂O (ranging from 1:4 to 4:1) is employed, resulting in 10 subfractions (F.12-1–F.12-10). Subfraction F.12-3 is then subjected to further purification using semi-preparative Agilent HPLC. An isocratic solvent system of 30% MeOH in H₂O is used over a period of 42 min. This process yields compounds 1 (17.1 mg, tR = 35.5 min) and 2 (10.5 mg, t_R = 38.7 min). Similarly, fraction EVF.10 is chromatographed on an RP-C18 column (4.0 × 60 cm; 150 μm particle size) and subjected to a stepwise gradient of MeOH–H₂O (ranging from 1:2 to 6:1), resulting in 10 subfractions (F.10-1–F.10-10). Subfraction F.10-6 undergoes further purification through semi-preparative Agilent HPLC, employing an isocratic solvent system of 45% MeCN in H₂O over an 80-min period. This process yields compound 3 (7.9 mg, t_R = 54.5 min).

Eryvarin M ( 1 ): Dark yellow powder; ¹H NMR (300 MHz, acetone-d₆) δ_H (ppm): 4.56 (1H, dd, J = 10.8, 12.0 Hz, H-2a), 4.17 (1H, dd, J = 5.4, 12,0 Hz, H-2b), 4.42 (1H, dd, J = 5.4, 10.8 Hz H-3), 7.75 (1H, d, J = 8.4 Hz, H-5), 6.54 (1H, dd, J = 2.1, 8.4 Hz, H-6), 6.40 (1H, d, J = 2.1 Hz, H-8), 6.56 (1H, s, H-3′), 6.78 (1H, s, H-6′), 3.70 (3H, s, 2′-OCH₃), 3.72 (3H, s, 5′-OCH₃); ¹³C NMR (75 MHz, acetone-d₆) δ_C (ppm): 71.7 (C-2), 48.2 (C-3), 191.7 (C-4), 129.9 (C-5), 111.0 (C-6), 162.6 (C-7), 103.3 (C-8), 163.8 (C-9), 116.1 (C-10), 114.7 (C-1′), 153.1 (C-2′), 101.2 (C-3′), 147.2 (C-4′), 142.3 (C-5′), 115.6 (C-6′), 56.4 (2′-OCH₃), 57.1 (5′-OCH₃).

Eryvarin H ( 2 ): Light yellow powder; ¹H NMR (500 MHz, methanol-d₄) δ_H (ppm): 4.87 (2H, s, H-2), 6.47 (1H, s, H-4), 6.88 (1H, d, J = 8.0 Hz, H-5), 6.34 (1H, dd, J = 2.0, 8.0 Hz, H-6), 6.26 (1H, d, J = 2.0 Hz, H-8), 6.85 (1H, s, H-6′), 6.50 (1H, s, H-3′), 3.81 (3H, s, 5′-OCH₃), 3.72 (3H, s, 2′-OCH₃); ¹³C NMR (125 MHz, methanol-d₄) δ_C (ppm): 69.6 (C-2), 130.3 (C-3), 122.0 (C-4), 128.7 (C-5), 109.7 (C-6), 159.3 (C-7), 103.7 (C-8), 156.2 (C-9), 117.7 (C-10), 120.0 (C-1′), 153.6 (C-2′), 101.8 (C-3′), 148.6 (C-4′), 143.1 (C-5′), 114.0 (C-6′), 57.5 (5′-OCH₃), 56.6 (2′-OCH₃).

Neobavaisoflavone ( 3 ): Light yellow powder; ¹H NMR (300 MHz, acetone-d₆) δ_H (ppm): 8.10 (1H, s, H-2), 8.04 (1H, d, J = 8.7 Hz, H-5), 6.98 (1H, dd, J = 3.0, 8.7 Hz, H-6), 6.88 (1H, d, J = 3.3 Hz, H-8), 7.34 (1H, d, J = 2.4 Hz, H-2′), 6.87 (1H, d, J = 8.4 Hz, H-5′), 7.27 (1H, dd, J = 2.4, 8.4 Hz, H-6′), 3.25 (1H, d, J = 7.5 Hz, H-1′′), 5.36 (1H, m, H-2′′), 1.72 (3H, s, 4′′-CH₃), 1.70 (3H, s, 5′-CH₃); ¹³C NMR (75 MHz, acetone-d₆) δ_C (ppm): 153.1 (C-2), 125.4 (C-3), 175.7 (C-4), 128.4 (C-5), 115.5 (C-6), 163.1 (C-7), 103.1 (C-8), 155.6 (C-9), 115.4 (C-10), 124.5 (C-1′), 125.4 (C-2′), 128.5 (C-3′), 158.8 (C-4′), 115.6 (C-5′), 131.2 (C-6′), 29.0 (C-1′′), 118.6 (C-2″), 132.3 (C-3′′), 25.9 (C-4″), 17.8 (C-5″).

Biology

PTP1B assay

To determine the inhibitory activity of compounds isolated from the stems and branches of the E. variegata against the enzyme PTP1B, an in vitro enzymatic assay was conducted following the method developed by To and colleagues in 2021.⁷⁰ The principle of this method involves the hydrolysis of p-nitrophenyl phosphate (p-NPP) by the enzyme PTP1B, resulting in the production of p-nitrophenol. The p-nitrophenol product generated after the phosphatase reaction has a yellow color. The amount of this color compound formed is directly proportional to the activity of the enzyme PTP1B. The enzyme activity is assessed by measuring the absorbance of the sample at 405 nm. Each well (final volume 110 μL) is loaded with 2 mM p-nitrophenyl phosphate (p-NPP) and 0.05–0.1 µg of PTP1B enzyme in a buffer solution containing 50 mM citrate (pH 6.0), 0.1M NaCl, 1 mM EDTA, and 1 mM dithiothreitol (DTT), with or without the test compound. The mixture is then incubated at 37 °C for 10 min, followed by the addition of 50 μL of p-NPP in a buffer solution. After incubating at 20 °C for 20 min, the reaction is stopped by adding 10M NaOH. The amount of p-nitrophenol produced by the enzyme through the phosphatase reaction is determined by measuring the absorbance at a wavelength of 405 nm using a spectrophotometer. The hydrolysis of p-NPP without the presence of PTP1B enzyme is corrected by measuring the increase in absorbance at 405 nm in the absence of PTP1B. All experiments are repeated three times.

α-glucosidase assay

The method for determining the inhibitory activity of α-glucosidase enzyme is performed using a 96-well microplate, with a reaction volume of 200 µL per well. The reaction components include 2.5 mM p-nitrophenyl α-D-glucopyranoside, 100 mM phosphate buffer (pH 6.8), 0.2 U mL⁻¹ α-glucosidase, and test samples (10 µL in a 200 µL reaction mixture) at different concentrations. In negative control samples, the test sample is replaced with 100 mM phosphate buffer, and acarbose is used as a positive control. The enzyme reaction mixture is incubated at 37 °C. After 30 min, the reaction is stopped with 0.1M Na₂CO₃. The absorbance of the reaction is determined at a wavelength of 405 nm using an Infinite F50 spectrophotometer (Tecan, Switzerland). The percentage inhibition of α-glucosidase activity (%I) is calculated using the following formula:

% I = \frac{A_{C} - A_{T}}{A_{C} - A_{0}} x 100

where %I is the percentage of α-glucosidase enzyme activity inhibited; AC is the absorbance of the control sample (without test sample); AT is the absorbance of the test sample; and A0 is the absorbance of the blank sample (containing only buffer, no α-glucosidase enzyme).

All tests are repeated three times. The IC₅₀ values, representing the concentration of the test samples that inhibits 50% of α-glucosidase enzyme activity, are calculated based on the log (concentration of test sample) and percentage inhibition curve.

Molecular docking study

Protein preparation

In this study, protein structures were retrieved from the Protein Data Bank (RCSB). The structure of human PTP1B with a catalytic inhibitor (PDB ID: 1BZC) was downloaded for the molecular docking process at the catalytic site. However, since the crystal protein (PDB ID: 1BZC) lacked an allosteric inhibitor, we also selected the co-crystal ligand at human PTP1B (PDB ID: 1T4J) to validate the docking process at the allosteric site. To assess similarity and identify both the catalytic and allosteric sites, the two crystal structures of PTP1B were aligned. Subsequently, PDB ID: 1BZC was chosen for further processing based on factors, such as resolution and the presence of an α-helix chain (residues Gly283–Asp298) compared to PDB ID: 1T4J. For human α-glucosidase, the protein crystal structure was also obtained from the RCSB database (PDB ID: 5NN8). The protein preparation followed established protocols as described in previous publications (please provide the appropriate citations). The selected crystals for the simulation process underwent the removal of water molecules and co-ligands. Subsequently, energy minimization was performed using Swiss PDB Viewer with standard optimization parameters (please provide the relevant citation). To prepare for molecular docking, polar hydrogen molecules were added to both protein structures using AutoDock Tool 1.5.6, and the structures were exported in a dockable pdbqt format.

Ligand preparation

The 3D structures of the compounds were prepared using MarvinSketch software (ChemAxon, USA). Subsequently, the compounds were optimized and set up with the MMFF94 force field using Avogadro (please provide the relevant citation). The co-crystal ligands of the proteins (PDB ID: 1BZC, 1T4J, and 5NN8) were also extracted and subjected to the same protocol for reference. Finally, AutoDock Tool 1.5.6 was employed to automatically establish torsion bonds for all ligands and convert them to the pdbqt format, which is suitable for screening.

Molecular docking simulation

The software AutoDock Vina 1.1.2 was used to conduct the molecular docking process. The positions of the crystal ligands were employed to define the grid box encompassing the active site of the protein. Docking results were expressed in kcal mol⁻¹. The Discovery Studio Visualizer 2020 program was employed to visualize the interactions between chemical compounds and protein targets.

ADMET predictions

Three isoflavonoids (1–3) isolated from E. variegata were screened based on the classic Lipinski’s rule of five.⁶⁸ During this process, SwissADME⁷¹ was used for in silico calculating physicochemical properties of these compounds. Their acute oral toxicity prediction was performed by DL-AOT Prediction Server.⁷²

Footnotes

Acknowledgements

The authors thank Dr Nguyen Quoc Binh for his help in the identification of the plant materials.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was financially supported by the Vietnam Academy of Science and Technology under the grant no. UDPTCN 03/21-23.

ORCID iD

Phi-Hung Nguyen

References

Bich

BH.

Medicinal plants and medicinal animals in Vietnam. Vol. 2. Shanghai, China: Scientific and Technical Publishing House, 2006.

Loi

DT.

Medicinal plants and remedies of Vietnam. Ha Noi, Vietnam: Medicinal Publishing House, 2006.

Martins

Brijesh

J Ethnopharmacol 2000; 248: 112280.

Anwar

The pharmacognostic and pharmacological studies on medicinal valued herbal drugs, Erythrina variegate var. orientalis, Matricaria chamommilla, Psoralea corylifolia and Chenopodium album. PhD thesis, Faculty of Pharmacy, University of Karachi, Karachi, Pakistan, 2006.

Kumar

Lingadurai

Jain

, et al. Pharmacogn Rev 2010; 4(8): 147.

Ministry of Health. Vietnam pharmacopoeia V. Vol. 2. Ha Noi, Vietnam: Medical Publishing House, 2017.

Nadkarni

AK.

Dr. KM ’Nadkarni’s Indian materia medica: with Ayurvedic, Unani-tibbi, Siddha, allopathic, homeopathic, naturopathic & home remedies, appendices & indexes. Vol. 1. Mumbai, India: Popular Prakashan, 2007.

Ghosal

Dutta

Bhattacharya

SK.

J Pharm Sci 1972; 61(8): 1274–1277.

Ito

Yakugaku Zasshi 1999; 119(5): 340–356.

10.

Zhang

Bao

, et al. Nat Prod Bioprospect 2020; 10: 57–66.

11.

Mohindra

Sharma

SR.

Fitoterapia 1993; 64: 15–15.

12.

Tang

Bao

, et al. Phytochemistry 2022; 198: 113160.

13.

Tang

Bao

, et al. Fitoterapia 2023; 166: 105408.

14.

Tang

, et al. J Org Chem 2021; 86(19): 13381–13387.

15.

Deshpande

Pendse

Indian J Chem 2007; 15B: 205–207.

16.

Tellikepalli

Gollapudi

Shokri

, et al. Phytochemistry 1990; 29: 2005–2007.

17.

Tanaka

Doi

Etoh

, et al. J Nat Prod 2001; 64: 1336–1340.

18.

Tanaka

Sato

Fujiwara

, et al. Lett Appl Microbiol 2002; 35: 494–498.

19.

Hegde

Dai

Patel

, et al. J Nat Prod 1997; 60: 537–539.

20.

Tanaka

Hirata

Etoh

, et al. J Nat Prod 2002; 65: 1933–1935.

21.

Tanaka

Hirata

Etoh

, et al. Phytochemistry 2003; 62: 1243–1246.

22.

Sato

Tanaka

Fujiwara

, et al. Phytomedicine 2003; 10: 427–433.

23.

Tanaka

Etoh

Phytochemistry 1997; 45: 205–207.

24.

Tanaka

Etoh

Phytochemistry 1998; 47: 475–477.

25.

Tanaka

Hosoya

, et al. Phytochemistry 1998; 48: 355–357.

26.

Xiaoli

Naili

, et al. Chem Pharm Bull 2006; 54: 570–573.

27.

Zhang

Lai

, et al. J Ethnopharmacol 2007; 109: 165–169.

28.

Zhang

Yao

, et al. Arch Pharm Res 2008; 31: 1534–1539.

29.

Muthukrishnan

Palanisamy

Subramanian

, et al. J Acupunct Meridian Stud 2016; 9(4): 207–212.

30.

Rahman

Kaisar

, et al. Turk J Pharm Sci 2010; 7(1): 21–28.

31.

Alam

Rana

Islam

, et al. Food Chemistry 2020; 320: 126646.

32.

Alam

Akhtaruzzaman

Handbook of phytonutrients in indigenous fruits and vegetables. CABI Digital Library, 2022, https://www.cabidigitallibrary.org/doi/book/10.1079/9781789248067.0000

33.

Tanaka

Atsumi

Shirota

, et al. Chem Biodivers 2011; 8(3): 476–482.

34.

Fatema

Doctoral dissertation, University of Dhaka, Dhaka, Bangladesh, 2021.

35.

Rahman

Sultana

Faruquee

, et al. Saudi Pharm J 2007; 15(2): 140.

36.

Gupta

(ed.). The wealth of India: a dictionary of Indian raw materials and industrial product. 3rd ed. New Delhi, India: National Institute of Science Communication and Council of Industrial and Scientific Research, 2002.

37.

Haque

Ali

Saha

, et al. Dhaka Univ J Pharm Sci 2006; 5(1): 77–79.

38.

Thongmee

Itharat

J Med Assoc Thai 2016; 99: S166–S171.

39.

Uddin

MMN

Emran

Mahib

MMR

, et al. Bioinformation 2014; 10(10): 630.

40.

Chatterjee

Burman

Nag

, et al. Indian J Pharm 1981; 13(2): 153.

41.

Bhattacharya

Debnath

Sanyal

, et al. J Res Indian Med 1971; 6: 235–241.

42.

Patil

Meena

Harish

DR.

In Silico Pharmacol 2021; 9(1): 51.

43.

Chu

Tan

, et al. Afr Health Sci 2019; 19(3): 2526–2536.

44.

Biradar

Patil

Joshi

, et al. Adv Tradit Med 2020; 26: 1–14.

45.

Santhiya

Priyanga

Hemmalakshmi

, et al. J App Pharm Sci 2016; 6(7): 147–155.

46.

Kumar

Lingadurai

Shrivastava

, et al. Pharm Biol 2011; 49(6): 577–582.

47.

Moller

DE.

Nature 2001; 414: 821–827.

48.

Johnson

Ermolieff

Jirousek

MR.

Nat Rev Drug Disc 2002; 1: 696–709.

49.

Ahmad

Azevedo

Cortright

, et al. J Clin Invest 1997; 100: 449–458.

50.

Elchebly

Payette

Michaliszyn

, et al. Science 1999; 283: 1544–1548.

51.

Asante-Appiah

Kennedy

BP.

Am J Physiol Endocrinol Metab 2003; 284: E663–E670.

52.

(a) Cheng

, et al. Chem Pharm Bull 2008; 56: 982–984. (b) Kirste B, Harrer W, Kurreck H, et al. J Am Chem Soc 1981; 103: 6280–6286.

53.

(a) Xu

Zou

Tan

, et al. J Asian Nat Prod Res 2011; 13: 93–96. (b) Cao Y, Chen J, Tan NH, et al. Bioorg Med Chem Lett 2010; 20: 2456–2460. (d) Cao Y, Chen JJ, Tan NH, et al. Magn Reson Chem 2010; 48: 656–659.

54.

Croteau

Ketchum

Long

, et al. Phytochem Rev 2006; 5: 75–97.

55.

Ramachandran

Kennedy

BP.

Curr Top Med Chem 2003; 3: 749−757.

56.

Kim

Chang

Ham

, et al. Molecules 2020; 25: 2683.

57.

Yaseen

Yang

Zhang

, et al. Nat Prod Res 2021; 35(23): 5364−5368.

58.

Yenesew

Derese

Irungu

, et al. Planta Med. 2003; 69: 658−661.

59.

Bae

Njamen

, et al. Planta Medica 2006; 72(10): 945–948.

60.

Jang

Njamen

, et al. J Nat Prod 2006; 69(11): 1572–1576.

61.

Jang

Thuong

, et al. Chem Pharm Bull 2008; 56(1): 85–88.

62.

Nguyen

Sharma

Dao

, et al. Bioor Med Chem 2012; 20(21): 6459–6464.

63.

Guo

Niu

Xiao

, et al. J Funct Foods 2015; 14: 324–336.

64.

Wang

Yuk

Kim

, et al. Bioor Med Chem Lett 2016; 26(2): 318–321.

65.

Johnson

Ermolieff

Jirousek

MR.

Nat Rev Drug Discov 2002; 1: 696–709.

66.

Yang

Tang

, et al. Free Radic Biol Med 2023; 194: 147–159.

67.

Tanchuk

Tanin

Vovk

AI.

Chem Biol Drug Des 2012; 80:121–128.

68.

Lipinski

CA.

J Pharmacol Toxicol Methods 2000; 44(1): 235–249.

69.

Jagannathan

ACS Omega 2019; 4(3): 5402–5411.

70.

Bui

Nhung

NTA

, et al. Molecules 2021; 26: 3691–3708.

71.

Daina

Michielin

Zoete

Sci Rep 2017; 7: 42717.

72.

Mukherjee

Rajan

J Chem Inf Model 2021; 61(5): 2187–2197.

Potential protein tyrosine phosphatase 1B and α-glucosidase inhibitory flavonoids from Erythrina variegata : Experimental and computational results

Abstract

Keywords

Introduction

Results and discussion

Chemistry

Biology

Molecular docking study

Conclusion

Experimental

Chemistry

General procedure for the isolation of compounds

Biology

PTP1B assay

α-glucosidase assay

Molecular docking study

Protein preparation

Ligand preparation

Molecular docking simulation

ADMET predictions

Footnotes

Acknowledgements

Declaration of conflicting interests

Funding

ORCID iD

References