New natural protein tyrosine phosphatase 1B inhibitors from Gynostemma pentaphyllum

Abstract

Type 2 diabetes mellitus (T2DM) is a chronic metabolic disease mainly caused by insulin resistance, which can lead to a series of complications such as cardiovascular disease, retinopathy, and its typical clinical symptom is hyperglycaemia. Glucosidase inhibitors, including Acarbose, Miglitol, are commonly used in the clinical treatment of hypoglycaemia. In addition, Protein tyrosine phosphatase 1B (PTP1B) is also an important promising target for the treatment of T2DM. Gynostemma pentaphyllum is a well-known oriental traditional medicinal herbal plant, and has many beneficial effects on glucose and lipid metabolism. In the present study, three new and nine known dammarane triterpenoids isolated from G. pentaphyllum, and their structures were elucidated by spectroscopic methods including HR-ESI-MS,¹H and ¹³C NMR and X-ray crystallography. All these compounds were evaluated for inhibitory activity against α-glucosidase, α-amylase and PTP1B. The results suggested that compounds 7∼10 were potential antidiabetic agents with significantly inhibition activity against PTP1B in a dose-dependent manner.

Keywords:

• Gynostemma pentaphyllum
• protein tyrosine phosphatase 1B
• α-glucosidase
• α-amylase
• phytochemistry

Introduction

Gynostemma pentaphyllum (Thunb.) Makino, named “Jiaogulan” in China, is also widely distributed in Korea, Japan, Vietnam and other Asian countries.^1–3 In China, G. pentaphyllum was used as vegetable and functional tea from Ming Dynasty (1368–1644 AD).^4,5 It has also been prescribed as traditional Chinese medicine to treat various diseases, such as non-alcoholic steatohepatitis (NASH) and diabetes.⁶ Previous phytochemical investigations showed that the dammarane triterpenoid saponins, sterols and flavonoids were the major chemical constituents in G. pentaphyllum.^7,8 To date, more than 180 dammarane triterpenoid saponins were reported from G. pentaphyllum with various bioactivities including antioxidant, antidiabetic, anti-inflammatory, hepatoprotective and antitumor effects.^9–13 The structure–activity relationship studies showed that dammarane triterpenoid aglycones, saponins removed sugar moieties, were more effective than the prototypical saponins.^14,15 However, the active ingredients in G. pentaphyllum and their mechanism of pharmacodynamic effects are unclear.

In our previous study,¹⁶ five new dammarane triterpenoid aglycones were isolated from the hydrolyzate of total G. pentaphyllum saponins which showed significant inhibitory activity against protein-tyrosine phosphatase 1B (PTP1B). PTP1B is an important negative regulator and plays a very important role in the process of insulin signalling.¹⁷ Inhibition of PTP1B activity can increase insulin sensitivity of target cells.¹⁸ Therefore, inhibition of PTP1B is considered as an effective treatment for type-2 diabetes mellitus(T2DM) and obesity.^19–22 T2DM is a chronic metabolic disease mainly caused by insulin resistance, which can lead to a series of complications such as cardiovascular disease, retinopathy, and its typical clinical symptom is hyperglycaemia.²³ In addition to PTP1B inhibitors, α-glucosidase and α-amylase inhibitors are also commonly used for hypoglycaemia in clinic, mainly including Acarbose, Miglitol, Voglibose and Etogliflozin. These drugs can reduce postprandial blood glucose and improve insulin levels.²⁴

In the present work, to further investigate the potential hypoglycaemic active constituents in G. pentaphyllum, a hydrolyzate of total G. pentaphyllum saponins was isolated by column chromatography to obtained three undescribed dammarane triterpene derivative 1–3, together with nine known analogues 4–12. Their structures were elucidated by spectroscopic methods including HR-ESI-MS,¹H and ¹³C NMR and X-ray crystallography. In addition, their inhibitory activity against PTP1B, α-glucosidase and α-amylase were also evaluated in vitro.

Materials and methods

General experimental procedures

In the separation process, 100–200 or 300–400 mesh silica-gel (Qingdao Marine Chemical Co., Ltd., China) and Sephadex LH-20 (GE Healthcare, Marlborough, MA) were employed for the column chromatography. The TLC detection was performed by heating silica gel plate (Qingdao Marine Chemical Co. Ltd., China) after spaying with 10% H₂SO₄ in ethanol. HR-ESI-MS data were recorded by a UPLC-Q/TOF-MS system including a 1290 Infinity II UPLC system and an Agilent 6545 Q/TOF-MS system (Agilent, MA, USA). The ¹H (400 MHz) and ¹³C NMR (100 MHz) spectra were detected by Varian INOVA AS 400 (Agilent, Santa Clara, USA). CDCl₃ solvent for NMR analysis was purchased from Sigma Aldrich. The X-ray single crystal diffraction data were obtained by SuperNova, Dual, AtlasS2 diffractometer (Rigaku, Japan). PTP1B (from human, ab51277) was purchased from Abcam Trading Co., Ltd. α-Glucosidase (from yeast, G8821) was purchased from Beijing Solarbio Co., Ltd., and α-amylase (from porcine pancreas, A3176, type VI-B) was purchased from Sigma–Aldrich (Shanghai) Co., Ltd. HepG2 cells was obtained from Cell Bank of the Chinese Academy of Sciences, Shanghai, China. Besides, other analytical grade chemical solvents were purchased from Jinhuada Chemical Co., Ltd., China.

Plant material

The commercialised total saponins extract of G. pentaphyllum was the same batch (No. 20190501) as the sample used in our previous study¹⁶ and purchased from Shaanxi Zhongxin Biotech Co. Ltd. (Unified Social Credit Code: 91610000677914226 R), in 2019. A voucher specimen was authenticated by Professor Daopeng Tan (Pharmacognosy, Zunyi Medical University) and deposited in the Herbarium of Zunyi Medical University.

Acid hydrolysis and isolation

The total G. pentaphyllum saponins (>80%, by UV) was purchased from Shaanxi Zhongxin biotech Co., Ltd. According to the literature,¹⁶ the G. pentaphyllum saponins (500 g) was dissolved in 1000 ml of methanol and then hydrolysed by adding 500 ml of 10% HCl for 8 h under 50 °C.⁵ The dammarane triterpenoid aglycones residue (150 g) was obtained after decompression evaporation and extraction by ethyl acetate (EtOAc). The residue was separated by silica gel column chromatography and gradient eluted with petroleum ether (PE)/EtOAc (50:1–1:1) or CH₂Cl₂/MeOH (50:1–1:1) to afford 10 fractions (Fr. A-J). The Fr. A (5.0 g) was separated again by silica gel column chromatography with n-hexane/EtOAc (50:1) to afford compound 1 (25 mg), 2 (20 mg), 3 (40 mg) and 4 (25 mg). Fr. B (2.5 g) was purified over Sephadex LH-20 column chromatography with CH₂Cl₂/MeOH (1:1) to obtain compound 5 (30 mg), 6 (35 mg) and 7 (40 mg). Fr. C (5.3 g) was separated by silica gel column chromatography with n-hexane/acetone (30:1) to yield compound 8 (35 mg), 9 (20 mg), 10 (15 mg), 11 (15 mg) and 12 (50 mg).

• Gypensapogenin VI (1): White amorphous powder; The ¹H NMR (400 MHz, CDCl₃) and ¹³C NMR (100 MHz, CDCl₃) spectral data was shown in Table 1; HR-ESI-MS: m/z 443.3897 [M + H-H₂O]⁺, (calcd. for 443.3950, for C₃₀H₅₂O₃).

• Gypensapogenin VII (2): White amorphous powder; The ¹H NMR (400 MHz, CDCl₃) and ¹³C NMR (100 MHz, CDCl₃) spectral data was shown in Table 1; HR-ESI-MS: m/z 453.3363 [M + H]⁺, (calcd. for 453.3324, for C₃₀H₄₄O₃).

• Gypensapogenin VIII (3): White amorphous powder; The ¹H NMR (400 MHz, CDCl₃) and ¹³C NMR (100 MHz, CDCl₃) spectral data were shown in Table 1; HR-ESI-MS: m/z 497.3633 [M + H]⁺ (calcd for 497.3586, for C₃₂H₄₈O₄).

Table 1. Primers used for qRT-PCR.

Gene	Forward primer (5′–3′)	Reverse primer (5′–3′)
PTP1B	AGCCAGTGACTTCCCATGTAG	TGTTGAGCATGACGACACCC
GAPDH	CAGCCTCAAGATCATCAGCA	ATGATGTTCTGGAGAGCCC

Molecular docking

The PTP1B, α-glucosidase and α-amylase protein crystal structures were obtained from PDB Database (https://rcsb.org) and modified by Autodock tools.^25,26 In this protocol, missing hydrogens were added, bond orders were assigned, and all water molecules and heteroatoms were deleted except the native ligand. After the optimisation step addressing overlapping hydrogens, a restrained minimisation was applied using the OPLS4 force field.²⁷ Before molecular docking, dammarane triterpenoid aglycones isolated from G. pentaphyllum were selected as ligands for the molecular docking analysis. Their 3D structures were built by ChemBioDraw Ultra14.0 and converted to PDBQT coordinated by AutoDockTools. The rotatable bonds of ligand were assigned using AutoDock Tools, and the molecular docking was performed through the AutoDock Vina. The protein-ligand interactions 2D maps constructed by LigPlot software was used to analyse the interaction forces between proteins and ligands. The protein–ligand interactions 3D maps constructed by PyMOL software was used to show the binding sites between proteins and ligands.

Protein tyrosine phosphatase 1B inhibition assay

The PTP1B inhibitory activities of dammarane triterpenoid aglycones isolated from G. pentaphyllum was determined by the microplate method.¹⁶ Each dammarane triterpenoid aglycones and Na₃VO₄ (positive control) were separately dissolved in dimethyl sulfoxide (DMSO) and then diluted to gradient concentrations by MOPS solution [pH = 7.0, 25 mM MOPS, 1 mM EDTA, 2 mM dl-dithiothreitol (DTT), 0.1 M NaCl]. The reaction system was 100 μL: PTP1B (25 nM), MOPS (25 mM, pH = 7.0), EDTA (1 mM), DTT (2 mM), NaCl (0.1 M) and disodium-4-nitrophenyl phosphate (pNPP, 2 mM). The enzyme-catalyzed reaction was carried out for 30 min under 37 °C and discontinued by the addition of 1.0 M NaOH. The inhibition efficiency of each compound against PTP1B was analysed by detecting the amount of p-nitrophenol (p-NP) at 405 nm by microplate reader. The inhibition activity was expressed as inhibition (%) and calculated as the following Equation (1)(1) $$\text{Inhibition \hspace{0.17em}}\left(\%\right)\hspace{0.17em}=\hspace{0.17em}(1\hspace{0.17em}-\Delta A_{\text{sample}/}\Delta A_{\text{control}})\times 100\%$$(1) : (1) $$\text{Inhibition \hspace{0.17em}}\left(\%\right)\hspace{0.17em}=\hspace{0.17em}(1\hspace{0.17em}-\Delta A_{\text{sample}/}\Delta A_{\text{control}})\times 100\%$$(1)

α-Glucosidase inhibition assay

The α-glucosidase inhibitory activity of dammarane triterpenoid aglycones isolated from G. pentaphyllum was also measured by the microplate method according to the literature.^27–29 In a 96-well plate, 50 μL of chemicals solution, 10 μL of α-glucosidase solution (1 U/mL, pH = 6.8), and 50 μL of buffer solution (PBS, pH = 6.8) were added separately. After mixing well, reaction solution was incubated for 15 min under 37 °C, and then added 40 μL PNPG (5 mmol/L), continue to incubate for 15 min. At last, the reaction was terminated by adding 40 μL of sodium carbonate solution. Before and after incubation, the absorbance was recorded by microplate reader at 405 nm. Acarbose was selected as a positive control. The α-glucosidase inhibition activity was also expressed as Equation (1)(1) $$\text{Inhibition \hspace{0.17em}}\left(\%\right)\hspace{0.17em}=\hspace{0.17em}(1\hspace{0.17em}-\Delta A_{\text{sample}/}\Delta A_{\text{control}})\times 100\%$$(1) .

α-amylase inhibition assay

The α-amylase inhibitory activity of dammarane triterpenoid aglycones isolated from G. pentaphyllum was determined by microplate method, too.²⁸ 50 μL chemicals solution, 50 μL α-amylase (2.5 U/mL) were added in a 96-well plate, after mixing, incubated for 10 min under 37 °C, and then added 100 μL 1.5% soluble starch solution, continue to incubate for 10 min, added 300 μL DNS for colour development, after boiling water bath at 95 °C for 10 min, diluted to 1500 μL by adding 1000 μL PBS. Before and after incubation, the absorbance was recorded on a microplate reader at 405 nm. Acarbose was used as a positive reference. The α -amylase inhibition activity was also expressed as Equation (1)(1) $$\text{Inhibition \hspace{0.17em}}\left(\%\right)\hspace{0.17em}=\hspace{0.17em}(1\hspace{0.17em}-\Delta A_{\text{sample}/}\Delta A_{\text{control}})\times 100\%$$(1) .

Protein tyrosine phosphatase 1B inhibition kinetic assay

The kinetic behaviour of dammarane triterpenoid aglycones isolated from G. pentaphyllum against PTP1B and the corresponding inhibition constants (Ki values) were investigated by Lineweaver-Burk double reciprocal plots and Dixon plots. Different concentrations of p-NPP substrates (1, 2, 5 and 10 mM) were subjected to enzymatic reactions with 100 nM PTP1B and various concentrations of chemicals (0, 0.5, 1.0 and 5.0 μM) in a 96-well plate. The absorbance of the reaction solution was recorded every 3 min by a microplate reader. The enzymatic velocity was calculated based on the time-Δabs plot. All tests were performed in triplicate.

Preparation of Gypensapogenin D-liposome

Gypensapogenin D-Liposome was prepared using the rotary evaporation film ultrasonication method.³⁰ The prescribed amounts of Gypensapogenin D, lecithin and cholesterol were weighed in a dry flask and dissolved by ultrasonication with an appropriate amount of chloroform-methanol (3:1 v/v) mixed solvent. The organic solvent was removed by rotary evaporation under reduced pressure in a constant temperature water bath at 42 °C. After a homogeneous lipid film was formed on the wall of the bottle, appropriate amount of phosphate-buffered saline (PBS) buffer (pH = 7.4) was added, and hydrated in a rotary evaporator at 42 °C under atmospheric pressure for 2 h. Next, the suspension was dispersed through ultrasonication for 10 min under an ice bath and then filtered by a 450 nm micropore filter membrane to obtained Gypensapogenin D-Lip and preserved at 4 °C.

The encapsulation efficiency of Gypensapogenin D-Lip was determined using low-speed centrifugation method.³¹ 1.0 ml of Gypensapogenin D-Lip suspension was taken in a 1.5 ml centrifuge tube and centrifuged at 1000 rpm/min for 10 min. After discarding the supernatant, 1 ml of methanol was added to dissolve it ultrasonically, and filtered through 0.22 μm filter membrane. The free Gypensapogenin D content in the dialysate was tested by HPLC [Chromatographic conditions: Hypersil ODS2-C18 column (4.6 × 250 mm, 5 μm); Column temperature was set as 30 °C; Mobile phase is acetonitrile-water solution (90:10, v/v); Flow rate was 1.0 ml/min; Detection wavelength was set as 203 nm; Injection volume was 20 μL]. The Gypensapogenin D -Lip encapsulation efficiency (EE%) and drug loading (DL%) were calculated according to the following formula (2) and (3), separately. (2) $$\text{EE}\%=(m1-m2)/m1\times 100\%$$(2) (3) $$\text{DL}\%=(m1-m2)/m3\times 100\%$$(3) where m1 is the total amount of Gypensapogenin D added initially, m2 is the amount of unencapsulated Gypensapogenin D in the dialysate, and m3 is the total amount of encapsulated Gypensapogenin D and pharmaceutical excipients.

In vitro glucose consumption assay

HepG2 cells were resuscitated and transferred into culture flasks with DMEM medium containing 10% inactivated foetal bovine serum (FBS), and cultured at 37 °C with 5% CO₂. After the cells were fully adhered to the wall, the medium was discarded and digested with 0.25% trypsin, and the cells were passaged once every 3 d at a ratio of 1:3, and the cells in the logarithmic growth phase were taken and used for experiments. HepG2 cells were inoculated in 96-well plates at 1 × 10⁵ cells/well (100 μL) and cultured for 24 h. The cells were then treated with different concentrations of drug-containing liposomes (1, 10, 100 and 200 μg/mL diluted in complete medium) and incubated for 48 h at 37 °C in humid air with 5% CO₂. The cell viability of HepG2 cells was then detected by adding 10 μL CCK-8 to each cell after treatment for 1 h.

HepG2 cells were inoculated in 6-well plates at a density of 1 × 10⁵ cells/well and cultured for 24 h. The cells were then starved in serum-free medium for 2 h. The cells were then treated with insulin containing 1 × 10⁻⁶ mol/l for 48 h at 37 °C in a 5% CO₂ incubator to induce insulin resistance. Cells were divided into a control group, a model group, a low-moderate-high-dose administration group, and a blank liposome group (i.e. blank liposomes with the same dilution ratio as the drug-containing liposomes), and glucose concentration was measured using a glucose oxidase assay kit after 24 h of treatment. Glucose consumption was calculated by subtracting the glucose concentration in the sample wells from the glucose concentration in the blank wells.

Real-time quantitative PCR analysis

After treatment of insulin resistance HepG2 cells, total RNA was extracted from the cells using Trizol and reverse transcribed into cDNA using the BeyoRT™ III First Strand cDNA Synthesis Kit. qRT-PCR was performed to detect fluorescence using a SYBR Green kit and normalise the mRNA levels to GAPDH expression.

Statistical analysis

The experimental results including PTP1B, α-glucosidase and α-amylase inhibition activity and the effect of Gypensapogenin D in cells assay were analysed using the Origin software (Version 8.0) and expressed as mean value ± standard deviation (n = 3).

Results and discussion

Structure elucidation

Three undescribed dammarane triterpene derivative 1–3, together with nine known analogues 4–12 were isolated from the hydrolyzate product of total G. pentaphyllum saponins. Structures of compounds 1–3 were elucidated based on the HR-ESI-MS, ¹H, ¹³C and 2D NMR spectral data, whereas the structures of other nine known compounds 4–12 were confirmed by comparing with their spectral data in literature.

Gypensapogenin VI (1) was isolated as a white needle crystal (in MeOH) with a molecular formula of C₃₀H₅₂O₃ as determined by the HR-ESI-MS at m/z 443.3897 [M-H₂O + H]⁺, (calcd. for 443.3950) and NMR data, requiring five degrees of unsaturation (Table 2). The ¹H NMR spectrum showed signals due to seven methyl groups (δ: 1.22, 1.19, 0.91 × 2, 0.80, 0.77, 0.71), an olefinic proton (δ: 5.05), and one oxygen-bearing protons (δ: 4.24). The ¹³C NMR spectrum showed 30 carbon resonances and demonstrated the presence of an olefinic carbon by the signals at δ: 154.2, 105.7 and three oxygen-bearing carbon (δ: 80.3, 79.7 and 77.9). So, the compound 1 was deduced to be a triterpenoid aglycone. In the HMBC spectrum, the long-range correlations of δ_H 5.05, 4.76(H-21) with δ_C 42.7(C-17), 80.3(C-23), 154.2(C-20); δ_H 4.24 with δ_C 42.7(C-17), 105.7(C-21), 154.2(C-20); δ_H 2.02 (H-17) with δ_C 30.0(C-22), 46.3(C-13), 105.7(C-21), 154.2(C-20); δ_H 1.19(H-26), 1.22(H-27) with δ_C 37.6(C-24), 79.7(C-25), respectively, indicated the existence of a terminal double bonds at position C-20 with two hydroxylation at position C-23 and C-25 in the side chain. Combining HSQC, ¹H-¹H COSY and NOESY data, compound 1 was elucidated as Figures 1 and 2 and named Gypensapogenin VI.

Figure 1. Chemical structures of compounds 1–12.

Figure 2. HMBC, COSY and NOESY correlations of compounds 1–3.

Table 2. ¹H and ¹³C NMR data of compounds 1–3 (in CDCl₃, J in Hz at 400 MHz and 100 MHz).

NO.	1		2		3
	δH	δC	δH	δC	δH	δC
1	1.67(m), 1.21(m)	37.6	4.41(d, J = 4.7 Hz)	73.9	1.70(m),1.06(m)	38.7
2	1.54(m),	26.4	1.91(m), 1.82(m)	24.0	1.63(m), 1.54(m)	23.7
3	3.13(dd, J = 11.2 Hz, 5.0 Hz)	77.9	3.86(m)	84.0	4.48(dd, J = 10.8, 5.5 Hz)	80.8
4				39.2		37.9
5	0.67(m)	54.8		130.0	0.83(m)	55.9
6	1.43(m), 1.23(m)	20.3	1.93(m), 1.69(m)	20.3	1.40(m),1.23(m)	18.6
7	1.55(m), 1.45(m)	38.0	1.85(m), 1.60(m)	32.0	1.59(m),1.32(m)	35.3
8		39.4		38.3		40.4
9	1.23(m)	49.8	2.25(m)	37.5	1.34(m)	50.6
10		36.1		132.8		37.1
11			1.94(m), 1.57(m)	26.5	1.53(m),1.30(m)	21.3
12	2.03(m), 1.66(m)	37.9	1.64 (m),1.21(m)	25.3	1.51(m), 1.15(m)	24.8
13	1.61(m)	46.3	2.14(m)	45.0	1.81(m)	46.7
14		48.4		48.4		49.8
15	1.57(m), 1.04(m)	30.4	1.59(m), 1.18(m)	37.5	1.74(m), 1.17(m)	31.3
16	2.00(m), 1.64(m)	32.0	1.30(m), 1.23(m)	26.0	1.66(m), 1.56(m)	28.8
17	2.02(dt, J = 11.3, 6.0 Hz)	42.7	2.19(m)	19.0	2.54(td, J = 10.8, 6.7 Hz)	37.7
18	0.77(s)	14.8	0.86(s)	15.1	0.90(s)	15.6
19	0.91(s)	14.6	1.75(m)	32.6	0.87(s)	16.3
20		154.2	3.14(td, J = 11.1, 6.2 Hz)	47.14		137.9
21	5.05(d, J = 1.6 Hz); 4.76(s)	105.7		201.1		173.9
22	1.92(dt, J = 13.6, 8.9 Hz), 1.38(m)	30.0	1.72(m), 1.62(m)	25.3	6.82(m)	145.9
23	4.24 (t, J = 7.2 Hz)	80.3	5.91(t, J = 4.2 Hz)	106.9	5.58(d, J = 9.0 Hz)	78.1
24	2.04(m), 1.61(m)	37.6		150.0	4.90(ddt, J = 9.0, 2.9, 1.4 Hz)	119.0
25		79.7		74.4		140.3
26	1.19(s)	27.1	1.28(s)	31.2	1.78(d, 1.77, J = 1.4 Hz)	25.8
27	1.22(s)	27.7	1.27(s)	32.1	1.80(d, 1.79, J = 1.4 Hz)	18.1
28	0.91(s)	27.0	1.16(s)	27.1	0.86(s)	28.0
29	0.71(s)	14.4	0.85(s)	22.5	0.85(s)	16.5
30	0.80(s)	15.2	0.89(s)	12.6	0.99(s)	15.6
–COO-						171.1
–OCH3					2.05(s)	21.4

Gypensapogenin VII (2) was separated as a white amorphous powder with a molecular formula of C₃₀H₄₄O₃ as detected by the HR-ESI-MS at m/z 453.3363 [M + H]⁺, (calcd. for 453.3324) and NMR data, requiring nine degrees of unsaturation (Table 2). The ¹H NMR spectrum showed signals due to six methyl groups (δ: 1.28, 1.27, 1.16, 0.89, 0.86, 0.85), an olefinic proton (δ: 5.91), and one oxygen-bearing protons (δ: 4.41). The ¹³C NMR spectrum showed 30 carbon resonances and demonstrated the presence of two pairs of olefinic carbon by the signals at δ: 150.0 and106.9; 132.8 and 130.0, three oxygen-bearing carbon (δ: 83.9, 74.4 and 73.9) and a ketone carbonyl group at δ: 201.1. In the HMBC spectrum, the long-range correlations of δ_H 4.41(H-1) with δ_C 24.0(C-2), 83.9(C-3) and 130.0(C-5); δ_H 3.86(H-3) with δ_C 32.6(C-19), 73.9(C-1) and 130.0(C-5), respectively, indicated the existence of a seven-membered ring with an oxygen bridge between C-1 and C-3 and a set of double bonds between C-5 and C-10 in A ring. In addition, the long-range correlations of δ_H 2.19(H-22) with δ_C32.0 (C-16), 74.4 (C-25), 106.9(C-23), 150.0(C-24); δ_H 3.14(H-13) with δ_C 25.9(C-12), 44.9(C-17), 201.0(C-21); δ_H 5.91(H-23) with δ_C 19.0 (C-22), 32.1 (C-20), 150.0(C-24), 201.1(C-21), respectively, indicated the existence of a five-membered cyclic ketone in the side chain. Combining HSQC, H-H COSY and NOESY data, compound 2 was elucidated as Figures 1 and 2 and named Gypensapogenin VII.

Gypensapogenin VIII (3) was isolated as a white needle crystal (in MeOH) with a molecular formula of C₃₂H₄₈O₄ as determined by the HR-ESI-MS at m/z 497.3633 [M + H]⁺ (calcd for 497.3586) and NMR data, requiring eight degrees of unsaturation (Table 2). In the ¹H NMR spectrum, eight methyl group signals at δ: 2.05, 1.80, 1.78, 0.99, 0.90, 0.87, 0.86 and 0.85, two olefinic proton at δ: 5.58 and 4.48, two oxygen bearing proton at δ: 5.58 and 4.48 were observed. In the ¹³C NMR spectrum, 30 carbon resonances signals were showed. Among them, two ester carbonyl carbons (δ_C: 173.9 and 171.1), two pairs of allylic carbons (δ_C: 137.9 and 145.9, 140.3 and 119.0), two oxygen bearing carbons (δ_C: 80.8 and 78.0) were presented. Comparing with their spectral data in literature,^32,33 the compound 3 was very similar to Gypensapogenin D except for an extra acetyl group. Interpretation of HMBC and HSQC spectra revealed that the acetyl group is linked to the hydroxyl group at the C-3 position. So, the structure of compound 3 was identified as Figures 1 and 2 and named Gypensapogenin VIII.

Others previously reported compounds were identified as: gypensapogenin A (4),³⁴ gypensapogenin B (5),³⁴ gypensapogenin C (6),³⁴ gypensapogenin D (7),³⁴ gypensapogenin IV (8),¹⁶ (20S, 23S) 3β, 20β - dihydroxydammar − 22, 24 - dien − 21 - oic acid − 21, 23 – lactone (9),¹⁶ (20 R, 23 R) 3β, 20β - dihydroxydammar − 22, 24 - dien − 21 - oic acid − 21, 23 – lactone (10),¹⁶ (20S, 23 R) 3β, 20β - dihydroxydamma − 24 - dien − 21 - oic acid − 21, 23 – lactone (11),³⁵ 20 R − 21, 24 - cycle − 3, 25 - diol dammar − 23(24) - en − 21 – one (12),¹⁶ by comparative analysis of their spectral data with that reported in the literature. Among them, the chemical structures of Gypensapogenin A and B were finally confirmed by the single crystal X-ray diffraction analysis (Figure 3), and their crystal data were displayed as follows:

Figure 3. Crystal structure of compounds Gypensapogenin A (4) and Gypensapogenin B (5).

Crystal data for Gypensapogenin A, C₃₀H₄₂O₂ (M = 434.63 g/mol): orthorhombic, space group P2₁2₁2₁ (no. 19), a = 7.57800(10) Å, b = 10.57400(10) Å, c = 30.3311(3) Å, V = 2430.42(5) ų, Z = 4, T = 150.00(10) K, μ(Cu Kα) = 0.548 mm⁻¹, Dcalc = 1.188 g/cm³, 30071 reflections measured (5.828° ≤ 2Θ ≤ 148.186°), 4864 unique (R_int = 0.0389, R_sigma = 0.0211) which were used in all calculations. The final R₁ was 0.0327 (I > 2σ(I)), and wR₂ was 0.0872 (all data).

Crystal data for Gypensapogenin B, C₃₀H₄₂O₂ (M = 434.63 g/mol): monoclinic, space group P2₁ (no. 4), a = 7.5829(7) Å, b = 10.4278(6) Å, c = 15.6475(11) Å, β = 103.223(7)°, V = 1204.49(16) ų, Z = 2, T = 150.00(10) K, μ(Cu Kα) = 0.553 mm⁻¹, Dcalc = 1.198 g/cm³, 7422 reflections measured (5.802° ≤ 2Θ ≤ 148.322°), 4295 unique (R_int = 0.0652, R_sigma = 0.0780) which were used in all calculations. The final R₁ was 0.1176 (I > 2σ(I)), and wR₂ was 0.3504 (all data).

Molecular docking analysis

To further predict the PTP1B (PDB ID: 5QDE), α-glucosidase (PDB ID: 3WY4), α-amylase (PDB ID: 1KXV) inhibitory activity of the dammarane triterpenoids, molecular docking was used to simulate the binding modes between PTP1B, α-glucosidase or α-amylase proteins with DPM-1001(positive control, a specific PTP1B inhibitor developed by DepYmed that has entered clinical trials), Acarbose (positive control, a specific α-glucosidase inhibitor) and the 12 dammarane-type triterpenoids from G. pentaphyllum, respectively. The specified size of the grid is 40 × 40 × 40 Å, and the targeted active site of PTP1B was pinpointed at coordinates x: −46.14, y: 16.85, z: 1.30, α-glucosidase was pinpointed at coordinates x: 40.12, y: −21.15, z: −20.15, α-amylase was pinpointed at coordinates x: −9.05, y: −16.93, z: −29.23. Moreover, the exhaustiveness of three proteins is10, and the number of mode runs is 5. After docking, the molecular docking results showed that all of these triterpenoids had good affinity with the PTP1B, α-glucosidase and α-amylase proteins (affinity ≤ −7.0 kcal/mol, Figure 4, Table 3). As shown in Figure 4, the hydroxyl group or oxygen on the heterocycle in most of the ligands had hydrogen bonding interactions with the PTP1B, α-glucosidase or α-amylase protein residues. As examples, the hydroxyl group at the C-3 position or hydrogen bond acceptor (HBA) groups such as –COOH, –C=O in the side chain of compounds 7∼10 were hydrogen bonded to Ser205, Phe280, Ser243, Arg79, Arg238 amino acid residues in PTP1B, respectively. These results suggested that the competitive hydrogen bonding of dammarane-type triterpenoids from G. pentaphyllum with the protein residues might be the key mechanism for their inhibition of PTP1B, α-glucosidase and α-amylase activities.

Figure 4. (A) Docking pose of compounds from G. pentaphyllum with PTP1B target. (B) Docking pose of compounds from G. pentaphyllum with α-glucosidase target. (C) Docking pose of compounds from G. pentaphyllum with α-amylase target. (^aPositive control).

Table 3. The molecular docking results of dammarane-type triterpenoids from G. pentaphyllum against PTP1B, α-glucosidase, α-amylase.

Compound	PTP1B			α-glucosidase			α-amylase
	Binding energy (kcal/mol)	Amino acid involved	H-bond length (Å)	Binding energy (kcal/mol)	Amino acid involved	H-bond length (Å)	Binding energy (kcal/mol)	Amino acid involved	H-bond length (Å)
1	−18.0	Ser205	3.07	−7.4	Ala378	2.85	−7.5	Glu233	3.13
		Glu200	2.81		Glu377	3.19
					Asp379	2.99
2	−22.1	Glu76	3.10, 2.97	−8.6	Leu300	3.08	−8.8	Glu240	2.93
								Lys200	2.39
3	−20.9	Arg238	3.09, 3.26	−8.1	Val380	3.00	−8.0	GIn5	3.06
					Glu231	3.23		Arg252	2.80
4	−23.1	–	–	−8.5	–	–	−8.3	–	–
5	−21.2	–	–	−8.4	–	–	−8.8	Lys200	2.80
6	−20.9	Phe280	2.94,	−8.3	Asp379	2.95	−8.2	Asp197	2.99, 3.10
7	−21.2	–	–	−8.1			−9.3	Asp197	3.01, 3.14
								Lys200	2.93
8	−19.8	Phe280	2.89	−7.5	Val335	3.07	−7.8	Glu240	2.95
		Arg79	3.06, 2.90, 3.00		Asp333	3.10		Lys200	3.06
		Ser205	3.27
9	−23.1	Arg79	3.18	−8.5			−8.7	Lys200	3.22
		Phe280	2.86
10	−21.4	Glu76	2.98-	−7.9			−8.1	Ser289	2.70
								Pro332	3.07
11	−19.4	–	–	−8.6	Val335	3.34	−9.1	Lys200	2.95
								Asp197	2.89
12	−20.0	Ser243	3.04	−8.4	Leu227	3.21	−8.6	Glu240	2.70
		Arg238	2.97, 3.16, 3.27		Leu300	2.94		Lys200	3.14
					Asn301	3.12
DPM-100^a	−17.8	Arg238	3.22
Acarbose^a				−7.6	Thr226, Ala229, Leu227, Asn301, Asp333, Lys398, Glu231, Gly399, Glu377, Gly402, Ala378, Asp379	−7.4	Arg92, Ser226, Thr6, Asp402, Gly334, Arg398

^a Positive control.

Biological evaluation

The PTP1B, α-glucosidase and α-amylase inhibitory activities of these twelve dammarane derivatives isolated from G. pentaphyllum were assayed by the pNPP method. Among them, except compounds 1 ∼ 2 for which no inhibitory activity was detected due to insolubility in DMSO, the other ten compounds (compounds 3 ∼ 12) possessed better inhibitory activity against PTP1B relative to α-glucosidase and α-amylase (Table 4). In particular, compounds 7 ∼ 10 showed better inhibitory activity against PTP1B than the positive control (Na₃VO4) with IC₅₀ values 8.17 ± 0.63 μM, 4.27 ± 0.43 μM, 8.59 ± 2.46 μM and 7.08 ± 0.69 μM, respectively. Comparing their structures with inhibitory activities against PTP1B and combining them with the results of molecular docking (Figure 4), such conformational relationships can be deduced: compound 7 is structurally different from compound 3 in that its hydroxyl group at the C-3 position is substituted, indicating that the hydroxyl group at the C-3 position is a key group in the inhibition of PTP1B activity by dammarane triterpenes from G. pentaphyllum. It was also evident from the molecular docking results that the hydroxyl group at the C-3 position could hydrogen bond with amino acid residues such as Ser205, Phe280 in PTP1B. On the other hand, some hydrogen bond acceptor (HBA) groups present in the side chains of these dammarane-type triterpenoid compounds, such as –COOH, –C=O, –OH are also capable of hydrogen bonding with amino acid residues such as Ser205, Ser243, Arg79, Arg238, Glu76 in PTP1B, which affects the strength of their inhibitory activity against PTP1B. These suggested that the HBA groups in dammarane-type triterpenoid compounds are more important for the inhibitory activity of PTP1B.

Table 4. Inhibitory activity of dammarane-type triterpenoids from G. pentaphyllum against PTP1B, α-glucosidase and α-amylase.

Compound	PTP1B			α-glucosidase		α-amylase
	inhibition (% at 50 μM)^a	IC50^a (μM)	R^2	inhibition (% at 50 μM)^a	IC50^a (μM)	inhibition (% at 250 μM)^a	IC50^a (μM)
1	–	–		–	–	–	–
2	–	–		–	–	–	–
3	51.55 ± 9.16	27.28 ± 3.09	0.9573	53.63 ± 3.64	33.98 ± 3.01	26.52 ± 0.59	–
4	65.88 ± 1.69	19.55 ± 1.48	1.0000	25.80 ± 0.71	–	31.81 ± 0.02	–
5	67.13 ± 1.33	29.43 ± 2.48	0.9539	25.55 ± 1.75	–	–	–
6	78.90 ± 0.87	24.60 ± 2.28	0.9824	24.64 ± 0.65	–	–	–
7	83.49 ± 0.62	7.77 ± 0.85	0.9973	61.94 ± 1.54	27.15 ± 0.17	11.57 ± 1.24	–
8	80.18 ± 0.51	4.26 ± 0.42	0.9974	44.26 ± 3.75	–	26.80 ± 0.44	–
9	83.23 ± 0.21	8.59 ± 2.46	0.9997	7.77 ± 0.80	–	45.49 ± 1.35	–
10	82.70 ± 1.54	7.08 ± 0.69	0.9907	13.72 ± 2.74	–	20.03 ± 0.21	–
11	78.65 ± 1.81	11.74 ± 0.52	0.9922	48.16 ± 8.42	–	35.41 ± 0.29	–
12	74.82 ± 0.91	14.48 ± 0.36	0.9963	26.72 ± 4.13	–	22.79 ± 1.84	–
Na3VO4^b	98.07 ± 0.51	12.65 ± 1.60	0.9958	–	–	–	–
Acarbose^b	–	–		96.51 ± 0.02	0.20 ± 0.05	92.47 ± 0.29	0.35 ± 0.13

^aIC₅₀ values were expressed as means ± SD of three replications.

^bPositive control.

Enzyme kinetic characterisation

The inhibition kinetic behaviour of compounds 7∼10 against PTP1B were further performed to probe the possible inhibition mechanism using Lineweaver − Burk double reciprocal plot and Dixon plot. As shown in Figure 5(A, B), the effect of compounds 7 and 8 on the catalytic rate of PTP1B under different substrate concentrations intersected the Y-axis at the same point, indicating that compounds 7 and 8 are competitive inhibitors against PTP1B. Whereas, the effects of compounds 9 and 10 on the catalytic rate of PTP1B at different substrate concentrations intersected at the same point in the second quadrant, suggesting that compounds 9 and 10 are mixed inhibitors against PTP1B. In the Dixon plot method (Figure 5(E–H)), the K_i values of 7∼10 were determined to be 4.29 μM (R² = 0.9297), 3.46 μM (R² = 0.9335), 4.79 μM (R² = 0.9065) and 7.09 μM (R² = 0.9338), respectively. The relatively lower K_i value of 7∼10 makes the four compounds to be promising competitive PTP1B inhibitors.

Figure 5. Lineweaver-Burk plots for PTP1B inhibition of compounds 7 (A), 8 (B), 9 (C) and 10 (D). Dixon plots for PTP1B inhibition of compounds 7 (E), 8 (F), 9 (G) and 10 (H). Each value was expressed as means ± SD of three replications.

Effect of Gypensapogenin D liposome on glucose consumption in HepG2 cells

Type 2 diabetes mellitus (T2DM) is a metabolic disease that is characterised by the body’s reduced insulin sensitivity and insulin resistance, leading to disturbances in glucose metabolism and increased blood glucose levels. Therefore, the present study utilised a high-glucose, high-insulin-induced insulin resistance cell model to examine the effects of a representative dammarane derivatives from G. pentaphyllum, Gypensapogenin D, on glucose consumption and its PTP1B mRNA expression in insulin-resistant HepG2 cells. During Gypensapogenin D has a poorly solubility, and in order to ameliorate the bioavailability problem caused by the poorly solubility of Gypensapogenin D, liposomes were considered as a delivery system for Gypensapogenin D. The process optimisation was used to screen the Gypensapogenin D liposome was formulated as lecithin-cholesterol-Gypensapogenin D (8 : 1 : 1) and sonicated for 15 min in PBS phosphate buffer. The encapsulation rate of the prepared liposome was 87.32 ± 0.57%; the particle size was 276.44 ± 1.38 nm, the polydispersity index (PDI) was 0.28 ± 0.02, and the zeta potential was −38.63 ± 0.22, which showed a uniform particle size distribution and good stability.

As shown in Figure 6(A, B), the viability of HepG2 cells treated with 1, 10 and 100 μg/mL concentrations of Gypensapogenin D and Gypensapogenin D liposomes was higher than 80%, and therefore, 100 μg/mL was set as the maximum concentration for the subsequent experiments. As shown as Figure 6(C, D), the glucose consumption of HepG2 cells in the model group that developed insulin resistance was significantly lower than that of the control group. Free Gypensapogenin D did not improve the reduction of glucose consumption in insulin-resistant HepG2 cells, whereas Gypensapogenin D liposomes significantly increased glucose consumption in insulin-resistant HepG2 cells in a dose-dependent manner, and there was no significant change in the blank liposome group compared with the model group. Furthermore, it could be observed that the expression of PTP1B mRNA was significantly elevated in the model group and significantly regressed after administration of Gypensapogenin D liposome, whereas similarly there was no decrease in the blank liposome group (Figure 6(E)). These results suggest that Gypensapogenin D, which cannot be taken up by cells due to its poor solubility, once prepared as a liposomal delivery system, improves its bioavailability and is able to exert a hypoglycaemic effect by inhibiting PTP1B.

Figure 6. Gypensapogenin D-Liposome regulated glucose consumption by inhibiting PTP1B. (A-B) Effects of different concentrations of Gypensapogenin D and Gypensapogenin D-Liposome on the viability of HepG2 Cells; (C-D) Effects of different concentrations of Gypensapogenin D and Gypensapogenin D-Liposome on glucose consumption of insulin-resistant HepG2 cells; (E) Effect of Gypensapogenin D-Liposome on PTP1B mRNA expression in insulin-resistant HepG2 cells. Each value was expressed as means ± SD of three replications, *P < .05 vs. Model, **P < .01 vs. Model, ^##P < .01 vs. Control.

Conclusion

PTP1B plays a negative feedback role in insulin signalling,¹⁷ and the PTP1B protein reduces the phosphorylation level of the insulin receptor (IR) and its insulin receptor substrate 1 (IRS-1), inhibiting the transport of type 4 glucose transporter protein and glucose uptake in insulin-sensitive cells.¹⁸ PTP1B inhibitors can effectively regulate insulin sensitivity as well as body weight, making it a potential therapeutic target against diabetes and obesity.^19–22 In order to find more PTP1B inhibitory components from G. pentaphyllum and to explore their effects on other diabetic targets, the triterpenoids in G. pentaphyllum were again intensively investigated in the present study and three new dammarane-type triterpenoids were identified. Subsequently, their inhibitory effects on α-glucosidase, α-amylase and PTP1B were assessed by molecular docking and in vitro experiments. The results showed that these compounds exhibited superior inhibition of PTP1B enzyme as compared to α-glucosidase and α-amylase. To ameliorate the common problem of poor solubility that these compounds generally have, Gypensapogenin D was selected as a representative compound for preparation into a liposomal delivery system and demonstrated a favourable hypoglycaemic effect in insulin resistance model cells. These results laid a good foundation for a more in-depth study of the hypoglycaemic effects and its mechanisms of dammarane-type triterpenoids in G. pentaphyllum in the future.

Authors’ contributions

Xianting Wang and Yidan Deng: wrote the manuscripts and analysed the data. Jianmei Wang: conducted experiments and assisted in the writing of the article. Yimei Du: provided help in the experiment and provided suggestions for the writing of the article. Di Wu: provided help in the data analysing. Lin Qin: provided help in the experiment and provided suggestions for the writing of the article. Xingdong Wu and Qianru Zhang: involved in analysis of the data, reviewed and revised the paper. Jian Xie: provided the design of the experiment. Yuqi He and Daopeng Tan: provided direct design of the experiment and funding acquisition. All authors have approved the final version of the manuscript and agreed to be accountable for all aspects of work.

Abbreviations
NASH	=	non-alcoholic steatohepatitis
T2DM	=	Type 2 diabetes mellitus
PTP1B	=	Protein tyrosine phosphatase 1B
G. pentaphyllum	=	Gynostemma pentaphyllum
HR-ESI-MS	=	High resolution electrospray ionisation mass spectroscopy
NMR	=	Nuclear magnetic resonance spectroscopy
NASH	=	Non-alcoholic steatohepatitis
TLC	=	Thin layer chromatography
UPLC	=	Ultra performance liquid chromatography
Q/TOF-MS	=	Quadrupole-time of flight mass spectrometry
UV	=	Ultraviolet and visible spectrum
EtOAc	=	Ethyl acetate
PE	=	Petroleum ether
DMSO	=	Dimethyl sulfoxide
EDTA	=	Ethylene diamine tetraacetic acid
DTT	=	Dithiothreitol
pNPP	=	Disodium-4-nitrophenyl phosphate
p-NP	=	p-Nitrophenol
PNPG	=	4-Nitrophenyl α-D-glucopyranoside
DNS	=	3,5-Dinitrosalicylic acid
PBS	=	Phosphate buffer solution
HPLC	=	High performance liquid chromatography
EE	=	Encapsulation efficiency
DL	=	Drug loading
FBS	=	Foetal bovine serum
HepG2	=	Human hepatocellular carcinomas
CCK-8	=	Cell counting kit-8
PCR	=	Polymerase chain reaction
RT-qPCR	=	Reverse transcription quantitative real-time polymerase chain reaction
GAPDH	=	Glyceraldehyde-3-phosphate dehydrogenase
HMBC	=	Heteronuclear multiple bond correlation
HSQC	=	Heteronuclear singular quantum correlation
COSY	=	Correlation spectroscopy
NOESY	=	Nuclear overhauser effect spectroscopy
IC50	=	Half maximal inhibitory concentration
PDI	=	Polydispersity index
IR	=	Insulin receptor

Disclosure statement

The authors report no conflicts of interest.

Data availability statement

The data used to support the findings of this study are available from the corresponding author upon request.

Additional information

Funding

This work was financially supported by the National Natural Science Foundation of China (NSFC, Nos. 82260843, 82060649, 82160740), Guizhou Engineering Research Centre of Industrial Key-technology for Dendrobium Nobile (QKJ [2022]048), the Department of Science and Technology of Guizhou Province (QKHPTRC-GCC [2023]075, QKHPTRC-CXTD [2023]024), Zunyi of China (QKHPTRC [2020]-16, ZSKHHZZ [2021]290). Zunyi Medical University Postgraduate Research Fund Project (ZYK184, ZYK197).