New Metabolite With Inhibitory Activity Against α-Glucosidase and α-Amylase From Endophytic Chaetomium globosum

α-Glucosidase and α-Amylase Inhibition Activity Tracking Liquid-Liquid Extraction and HSCCC Isolation

Two-phase solvent systems composed of 4 solvents were used in liquid-liquid extraction, and thus the components in the treated samples could be fractioned based on different polarities according to their K values. ²⁴ A series of solvent systems based on hexane-ethyl acetate-methanol-water mixed at different volume ratios (9:1:1:9, 7:3:1:9, 5:5:1:9, 3:7:1:9, and 1:9:1:9) was selected for liquid-liquid extraction pretreatment of the crude samples of C. globosum. Fractions A-E were afforded by liquid-liquid extraction, which was followed up with in vitro α-glucosidase and α-amylase inhibition scan. Fraction C was found to be the most efficient α-glucosidase inhibition fraction with an IC₅₀ = 18.67 μM (Supplemental Table S2). This was superior compared to the acarbose control (IC₅₀ = 54.74 µM). Fraction C was analyzed by HPLC and 3 compounds were identified as possible inhibitors of α-glucosidase. The retention times of those compounds ranged from 17 minutes to 21 minutes (Figure 2(A)). HSCCC proved to be a superior method to obtain the bioactive compounds from fraction C compared to traditional HPLC methods. The effectiveness of HSCCC stems from the use of an optimized of the 2-phase solvent system which prompts the separation of the targeted compounds based on their expected K values which range from 0.5 to 2.0. ²⁵ A novel 9 × 9 map-based solvent selection strategy ²⁶ was applied for the 2-phase solvent system setup; hexane-ethyl acetate-methanol-water with different volume ratios were tested for the K values (Table 1). Solvent system hexane-ethyl acetate-methanol-water (3:7:4:6, v/v) gave better K values (1.5869, 0.9664, and 7.5531) for the purification of fraction C. However, the K value of compound 3 (7.5531) is much larger than the ideal range of 0.5 to 2.0, suggesting it would take a longer time to wash compound 3 through the HSCCC (3:7:4:6) system by normal mobile phase (downer phase). Changing mobile phase from downer phase (polar water rich phase) to upper phase (organic phase), after compounds 1 and 2 were washed out helped to overcome this problem and allowed compound 3 to be eluted earlier than initially expected. When HSCCC separation for fraction C was carried out with the 3:7:4:6 solvent system, 3 main compounds were successfully separated (Figure 2(B) and (C)), leading to the isolation of 3 metabolites (1-3). Meanwhile, other fractions were subjected to multiple chromatographic separations leading to isolation of 9 metabolites (4-12).

OPEN IN VIEWER

Figure 2

HPLC chromatograms of fraction C (A), HPLC conditions: column, Hypersil BDS C18 (250 mm × 4.6 mm, 5 μM); mobile phase A and B is acetonitrile and water, gradient elution program: 0–25 minutes, 50% A to 70% A; flow rate, 1.0 mL/min; and detection wavelength, 254 nm; HSCCC chromatograms of fraction C purification (B) at room temperature (column volume: 300 mL; liquid system: hexane : ethyl acetate : methanol : water 3:7:4:6; mobile phase: downer phase, polar water rich phase with a flow rate of 4 mL/min in the head to tail direction from 0 to 200 min, mobile phase was changed into upper phase and flow rate changed into 10 mL/min at 205 minutes; rotation speed of 1200 rpm; detection wavelength, 254 nm. HPLC chromatograms of compounds 1-3 after HSCCC purification (C), HPLC conditions are consistent with fraction C. HPLC, high-performance liquid chromatography; HSCCC, high-speed counter-current chromatography.

OPEN IN VIEWER

Table 1

K ^a of Compounds 1-3 Obtained From Fraction C in Different Solvent Systems.

Solvent system (v/v)		K values
hexane-ethyl acetate-methanol-water	1	2	3
5:5:5:5	0.1455	0.1303	1.1305
4:6:5:5	0.3169	0.2701	1.9863
4:6:4:6	0.7388	0.5356	3.7849
3:7:4:6	1.5869	0.9664	7.5531
3:7:3:7	3.4023	2.3222	16.6439

^a K was calculated as the ratio of the sample concentration in the upper (stationary) to lower (mobile) phases.

Structure Elucidation

Twelve compounds were isolated and purified from C. globosum, including 1 new compound named chaetoglobol acid (1). Eleven known compounds (2-12) were identified by comparing observed and reported NMR data as cyclo( _L -Pro- _L -Trp) (2), ²⁷ (–)-aureonitol (3), ^28,29 chaetomugilins A, Q, I, J (4-7) ^{30 -32} and chaetoviridin J (8), ³³ chaetoviridin B (9), ³⁴ chaetoglobosins B, E (10, 11) ³⁵ and penochalasin B (12). ³⁶

Chaetoglobol acid (1) was isolated as a white powder. The molecular formula of 1 was determined to be C₁₁H₁₄O₄ with 5 degrees of unsaturation by HRESIMS at m/z 233.0778 [M + Na]⁺ (Supplemental Figure S1G). The ¹H NMR spectrum (Table 2) revealed the presence of 7 olefinic protons [δ_H 5.24 (1H, dd, J = 16.5, 2.0 Hz, H-1), δ_H 5.10 (1H, dd, J = 10.0, 2.0 Hz, H-1), δ_H 6.39 (1H, m, H-2), δ_H 6.33 (1H, m, H-3), δ_H 5.73 (1H, dd, J = 14.7, 6.9 Hz, H-4), δ_H 6.87 (1H, dd, J = 15.6, 8.8 Hz, H-8), δ_H 5.92 (1H, dd, J = 15.6, 0.8 Hz, H-9)], 1 methine [δ_H 2.97 (1H, m, H-7)],2 oxymethine [δ_H 4.08 (1H, m, H-5), 3.81 (1H, dd, J = 7.7,6.8 Hz, H-6)], and 2 oxymethylene protons [δ_H 4.11 (1H, dd, J = 3.3, 7.2 Hz, H-11), 3.82 (1H, m, H-11)]. The ¹³C NMR data (Table 2) revealed 11 carbon signals corresponding to 2 methylenes [1 oxygenated at δ_C 69.72 (C-11), 1 olefinic at δ_C 116.79 (C-1)], 8 methines [2 oxygenated at δ_C 84.98 (C-5), δ_C 80.57 (C-6), 5 olefinic at δ_C 136.22 (C-2), 132.92 (C-3), 131.33 (C-4), 146.78 (C-8), and 122.91 (C-9)], and 1 quaternary carbons [one carboxyl at δ_C 168.08 (C-10)]. The structure of 1 was established through a further analysis involving 1D NMR, HMBC, and COSY spectra (Figure 3(A), Supplemental Figure S1(D-F) and Table S1). The ¹H–¹H-COSY and HMBC spectra of 1 allowed the identification of the planar structure of 1. Key COSY correlations, as shown in Figure 3(A), reveal a long chain from C-1 to C-9. The HMBC correlation of H-8/C-10 indicated that the carboxy group (C-10) links to the carbon at C-9. Furthermore, the HMBC correlations of H-6/C-11 and H₂-11/C-5 revealed a tetrahydrofuran moiety in the structure. The geometric configurations (E, E) of double bonds at C3 and C8 in 1 were determined as the same in 3 by the large coupling constants (J ~ ca. 15 Hz).

OPEN IN VIEWER

Table 2

¹H NMR (500 MHz) and ¹³C NMR (125 MHz) Spectroscopic Data for 1 in MeOH-d ₄.

No.	δ C	δ H	(J in Hz)
1	116.79	5.24	(1 H, dd, 16.5, 2)
		5.10	(1 H, dd, 10, 2)
2	136.22	6.39	(1 H, m)
3	132.92	6.33	(1 H, m)
4	131.33	5.73	(1 H, dd, 14.7, 6.9)
5	84.98	4.08	(1 H, m)
6	80.57	3.81	(1 H, dd, 7.7, 6.8)
7	50.59	2.97	(1 H, m)
8	146.78	6.87	(1 H, dd, 15.6, 8.8)
9	122.91	5.92	(1 H, dd, 15.6, 0.8)
10	168.08
11	69.72	4.11	(1 H, dd, 3.3, 7.2)
		3.82	(1 H, m)

OPEN IN VIEWER

Figure 3

¹H-¹H Correlated Spectroscopy (COSY) and key Heteronuclear Multiple-Bond Correlation (HMBC) correlations of compound 1 (A) and CD spectrum of compound 1 and(–)-Aureonitol (3) (B).

(–)-aureonitol (3), a C-₁₃-polyketide metabolite, was reported to be isolated from various Chaetomium species by Abraham, ³⁷ Li, ³⁸ and Saito³⁹ from 1983 to 2018. The absolute (5S,6R,7S) configuration of stereocenters on the tetrahydrofuran ring of 3 had been confirmed by its enantioselective synthesis. ²⁹ As an analog of compound 1, aureonitolic acid was isolated from Chaetomium sp. by Aly (Supplemental Table S3). ³⁹ Proton NMR data of 1 were compared with both aureonitolic acid and (–)-aureonitol 3 (Supplemental Table S3); the chemical shift of their H-5, H-6, H-7 is basically the same, indicating the relative configuration of C-5, C-6, C-7 of 1 is same as 3 and aureonitolic acid. The biosynthesis of 3 was reported by Saito. ⁴⁰ From a biosynthesis pathway, compound 1 shared a polyketide related to both (–)-aureonitol (3) and aureonitolic acid present in the same strain. We next measured CD spectra of 1 and 3. The CD spectra showed the same trend in the range of 200-400 nm (Figure 3(B)). The first negative Cotton effect at 244 nm of 1 could be assigned to Cotton effect at 234 nm of 3. The next positive Cotton effects of 1 were located at 274 nm, which could be assigned to absorption bands at 271 nm of 3. Among the excitations, π-type to π*-type (C=C and/or C=O double bonds) orbital of unsaturated side chains played a dominant role; in addition, 1 had a similar sign of the specific rotation [α]_D ²⁴ −32.4 (c1, CH₃CN) to that of 3, [α]_D ²⁴ −45.5 (c1, CH₃CN). Accordingly, 1 shared the same absolute configuration (5S, 6R, 7S) at the chiralcenters as that of the (–)-aureonitol (3). The structure of 1 was thus established as 3-(5-buta-1,3-dienyl-4-hydroxy-tetrahydro-furan-3-yl)-acrylic acid as shown.

In Vitro α-Glucosidase and α-Amylase Inhibition

α-glucosidase and α-amylase inhibitory activities of compounds 1-12 were investigated. As shown in Table 3, compounds 1 and 9 exhibited promising inhibition of α-glucosidase. Compound 1 showed higher inhibition against α-glucosidase compared to α-amylase. Compounds 1 and 3 shared the same skeleton, differing only at the C-10 position, indicating that the carboxyl group is essential for both α-glucosidase and α-amylase inhibition. Compounds 9 and 4 have a similar skeleton except for only 1 more hydroxy group on 4, indicating that the hydroxy group is detrimental for α-glucosidase inhibition. No apparent inhibitory activity against α-glucosidase and α-amylase was observed for any other investigated compounds, with IC₅₀ values of >50 µM. Overall, compound 1 displayed dual inhibitory activity against α-glucosidase and α-amylase. Compound 1 showed α-glucosidase inhibitory activity (IC₅₀ = 3.04 µM), which was 18-fold higher than that of acarbose (IC₅₀ = 54.74 µM) and displayed moderate inhibitory activity against α-amylase (IC₅₀ = 22.18 µM). Molecular docking analysis was carried out to further research compound 1 inhibitory activity against α-glucosidase. The binding pocket of compound 1 with α-glucosidase crystal structure is shown in Figure 4(A). It was found that Tyr72, Glu277, Gln279, Arg315, His351, Glu411, and Arg442 are the binding residues of the catalytic sites in the active pocket (Figure 4(B)). Among them, the three residues Glu277, His351, and Arg442 are considered to play a vital role in the hydrolysis of glycosidic bond. ⁴¹ This indicates that compound 1 is a potential competitive inhibitor of glycosidase, which suggests new natural product 1 has inhibitory effect against enzymes relevant to the progression of type 2 diabetes, making it a new potential drug candidate for the management of type 2 diabetes.

OPEN IN VIEWER

Figure 4

Molecular docking of compound 1 in α-glycosidase crystal structure (3A4A). (A) The binding pocket of compound 1 docking into α-glycosidase crystal structure. (B) The binding interactions of compound 1 with active site residues of α-glycosidase. The dotted line between compound 1 and the amino acid residues that interact with compound 1 represents the interaction force between them, and the dotted lines with different colors represent different forces.

OPEN IN VIEWER

Table 3

α-Glucosidase and α-Amylase Inhibition.

Compound	IC50 (μM)
	α-glucosidase	α-amylase
Chaetoglobol acid (1)	3.04 ± 0.31	22.18 ± 1.28
(–)-aureonitol (3)	>50	>50
Chaetoviridin B (9)	6.328 ± 0.25	>50
Chaetomugilin A (4)	>50	>50
Genistein	22.07 ± 1.03	>50
Acarbose	54.74 ± 0.16	13.72 ± 0.27

In conclusion, in this study, 12 metabolites (1–12), including 1 new compound, chaetoglobol acid (1), were identified from the cultures of endophytic C. globosum. An efficient bioactive tracking separation strategy based on liquid–liquid extraction and HSCCC was built and applied, leading to the discovery of the new metabolite compound 1 which exerted the most potent inhibitory effects against α-glucosidase activity. Compound 1 also displayed moderate inhibitory activity against α-amylase. These results suggested that compound 1 may serve as a potential new drug lead for the management of type 2 diabetes.

General

NMR spectra were obtained on Bruker Avance 400 MHz and Bruker Avance III 500 spectrometers with tetramethylsilane as an internal standard at room temperature. High resolution (HR) Electrospray Ionization Mass Spectroscopy (ESIMS) were recorded on a Thermo Fisher Scientific Q-TOF mass spectrometer. ESIMS were recorded on a Thermo Fisher LTQ Fleet instrument spectrometer. Electronic Circular dichroism (ECD) spectra were obtained on a Chirascan spectrometer. Silica gel (300, 400 mesh, Qingdao Marine Chemical Ltd., People’s Republic of China) and RP-18 gel (20, 45 µM, Fuji Silysia Chemical Ltd., Japan) were used for column chromatography (CC). Semi-preparative HPLC was performed on a Waters 1525EF liquid chromatography system equipped with Hypersil BDS C18 column (4.6 mm × 250 mm; 10.0 mm × 250 mm). The HSCCC experiment was conducted on a TBE-300B HSCCC system (Tauto Biotech, Shanghai, China), equipped with 3 polytetrafluoroethylene (PTFE) preparative coils (i.d. of the tubing = 1.6 mm, total volume = 300 mL) and a 20 mL sample loop. The revolution radius of the apparatus was 5 cm, and the β values of the multilayer coil ranged from 0.5 (at the internal terminal) to 0.8 (at the external terminal). For the apparatus, β = r/R, where r is the distance from the coil to the holder shaft, and R is the revolution radius or the distance between the holder axis amniocentesis of the centrifuge. Apparatus revolution speed was regulated with a speed controller within the range of 0 to 1200 rpm. The system was equipped with a model TBP5002 constant-flow pump (Tauto Biotech, Shanghai, China), a model UV-500 detector (XUYUKJ Instruments, Hangzhou, China) operating at 254 nm, and a model N2000 workstation (Zhejiang University, Hangzhou, China). A DC-0506 constant temperature-circulating implement (Shanghai Shunyu Hengping Instruments, Shanghai, China) was used to adjust the experiment temperature. Baker’s yeast α-glucosidase, hog pancreas α-amylase, nitrophenyl α-D-glucopyranoside (pNαGP), 2-chloro-4-nitrophenyl α-D-maltotrioside (G3-CNP), acarbose, and genistein were purchased from Sigma-Aldrich (St. Louis, MO).

Fungal Materials and Fermentation

The fungal strain C. globosum TY1, separated from the fresh bark of a medicinal plant Ginkgo biloba, was donated by Prof. J.C. Qin, College of Plant Science, Jilin University. ^12,13 After growing on Potato Dextrose Agar (PDA) medium at 25 °C for 7 days, the mycelium of this strain TY1 was inoculated in PD liquid medium. The pH was adjusted to 6.0 prior to autoclaving. Fermentation was carried out in multiple 500-mL flasks containing 150 mL media on rotary shakers at 220 rpm/min, 25 °C for 14 days.

Measurement of Partition Coefficient (K)

The partition coefficients (K) for target compounds were determined by an HPLC analysis according to the previous study. ²⁴ In brief, a small amount (1 mg) of the crude sample was dissolved into equal volume (0.8 mL) of the aqueous phase (lower phase) and organic phase (upper phase) of the thoroughly equilibrated 2-phase solvent system in a 2-mL test tube. After the equilibration was established, both the upper phase and lower phase were analyzed by HPLC and the peak areas of each compound in the upper phase and lower phase were recorded as A1 and A2, respectively. The partition coefficient (K) was then calculated by the following equation: K = A1/A2.

Extraction and Isolation

The culture broth (42 L) together with 5 L MeOH mycelium extraction was filtered by 8 layers of sterile gauze, and the filtrate was concentrated by a rotary evaporator under reduced pressure at 45 ℃ to give a crude extract (86.0 g). The crude extract was dissolved in the lower phase (500 mL) of hexane-ethyl acetate-methanol-water (9:1:1:9, v/v) in a separating funnel, after which an equal volume of the upper phase (500 mL) of the same solvent system was added and the mixture in the separating funnel was shaken. After equilibrium of the sample distribution between the two phases was achieved, the upper phase was separated. The lower phase was extracted twice by the upper phase. The upper phase was combined and evaporated to dryness as fraction A. Subsequently, the lower phase (500 mL) of hexane-ethyl acetate-methanol-water (7:3:1:9, v/v) was used to dissolve the dried sample from the previous lower phase, and then further extracted twice with the upper phase (500 mL) of hexane-ethyl acetate-methanol-water (7:3:1:9, v/v). The upper phase was combined and evaporated to dryness as fraction B. Fractions C-E were produced using the same method described above with the solvent systems of ethyl acetate–n-butanol–water at the volume ratios of 5:5:1:9, 3:7:1:9, and 1:9:1:9, respectively. All fractions were scanned for α-glucosidase inhibition activity and followed up with purification. Fraction A was subjected to Sephadex LH-20 (MeOH) and further purified by semi-preparative HPLC (80%, MeOH–H₂O, 2 mL/min) to yield 4 (t_R = 16.2 minutes, 5 mg). Fraction B was subjected to Sephadex LH-20 (MeOH), and further purified by semi-preparative HPLC (75%, MeOH–H₂O, 2 mL/min) to yield 5 (t_R = 15 minutes, 2.6 mg). Fraction C was separated by HSCCC with a 2-phase solvent system composed of n-hexane−ethyl acetate−methanol−water (HEMWat, 3:7:4:6 v/v), yielding 1 (t_R = 100 minutes, 98.5 mg), 2 (t_R = 175 minutes, 130.3 mg), and 3 (t_R = 260 minutes, 48.2 mg). Fraction D was further purified by silica gel CC eluting with a gradient of CHCl₃–MeOH (50:1–1:1), to yield 9 fractions (D1-D9). Fraction D2 was subjected to Sephadex LH-20 (MeOH) and further purified by semi-preparative HPLC (30%, MeOH–H₂O, 2 mL/min) to yield 6 (t_R = 25 minutes, 3.8 mg). Fraction D3 was purified by semi-preparative HPLC (60%, MeOH–H₂O, 2 mL/min) to yield 8 (t_R = 11 minutes, 6 mg). Fraction D5 was purified by semi-preparative HPLC (45%, MeOH–H₂O, 2 mL/min) to yield 7 (t_R = 22 minutes, 16 mg). Fraction D6 was purified by semi-preparative HPLC (30%, MeOH–H₂O, 2 mL/min) to yield 10 (t_R = 22 minutes, 7 mg) and 9 (t_R = 24 minutes, 3 mg). Fraction E was subjected to Sephadex LH-20 (MeOH), and further purified by semi-preparative HPLC (30%, MeOH–H₂O, 2 mL/min) to yield 11 (t_R = 29 minutes, 2.9 mg) and 12 (t_R = 18 minutes, 11 mg).

Chaetoglobol acid (1): white powder; [α]_D ²⁴ −32.4 (c1, CH₃CN); CD (c1, CH₃CN) 332 (+0.65), 277 (+0.61), 246 (−3.4), 224 (+1.0), 209 (+0.54), 202 (+0.6); ¹H (500 MHz) and ¹³C NMR (125 MHz) data, see Table 2; HRESIMS m/z 233.0778 [M + Na]⁺ (calcd. for C₁₁H₁₄O₄Na, 233.0790).

Enzyme Inhibition Assays ^42,43

Commercially available α-glucosidase and α-amylase were selected as the target proteins, with pNαGP and G3-CNP as their substrates, respectively. Fractions A-E, compounds 1-12 and acarbose, genistein (positive control) were dissolved in dimethyl sulfoxide. α-lucosidase and pNαGP were dissolved in a potassium phosphate buffer (0.05 M, pH 6.8). α-Amylase and G3-CNP were dissolved in a potassium phosphate buffer (0.1 M, pH 6.0). Enzymatic reaction mixtures were composed of enzymes (20 µL), substrates (30 µL), test compounds (10 µL), and a potassium phosphate buffer (140 µL) and incubated at 37 °C for 30 minutes. Enzymatic activity was detected by spectrophotometry at 405 nm. α-glucosidase and α-amylase were used at final concentrations of 4 and 25 milliunits in a potassium phosphate buffer with pNαGP (0.075 mM) and G3-CNP (0.5 mM) as substrates, respectively. Results are the average of 3 independent experiments performed in duplicating.

Statistical Analysis

All data are expressed as means ± the standard error of the mean (SEM). The statistical analysis was performed with GraphPad Prism 7.0 (GraphPad, La Jolla, CA, USA).

Molecular Docking

The molecular docking experiment of the ligand was performed using the crystal structure of α-glucosidase and maltose co-crystal (yeast isomaltase, PDB code: 3A4A), in the Discovery Studio 2017 R2. The CDOCKER package was used to prepare ligands and receptors. After the docking process, the resulting 3D complexes were shown by Discovery Studio (BIOVIA Software Inc, USA).