Structurally diverse brefeldin A derivatives as potent and selective acetylcholinesterase inhibitors from an endophytic fungus Penicillium brefeldianum F4a

2.1 General experimental procedures

High performance liquid chromatography (HPLC) was recorded by Dionex UltiMate 3000 system. High-resolution mass spectrometry (HR-MS) data were analyzed through a Thermo Scientific Q Exactive mass spectrometer. Nuclear magnetic resonance (NMR) data were given using a Bruker-AV-600 NMR spectrometer. The Electronic circular dichroism (ECD) spectra were measured using a JASCO J-815 CD spectrometer. A JASCO P-2000 polarimeter was used to afford optical rotations. A Thermo fisher Nicolet 6700 FT-IR spectrometer was used to acquire the IR spectra.

2.2 Fermentation and extraction

The study strain Penicillium brefeldianum F4a was isolated from the Houttuynia cordata Thunb. (Saururaceae), collected from Longkong countryside, Hunan province (26°46′17.4″N, 109°38′27.6″E). The fermentation and extraction of the endophytic Penicillium brefeldianum F4a was described according to our previous study (Bai et al., 2022).

2.3 Isolation and purification

A 35 g extract was subjected to silica gel chromatographic column (CC) and eluted with CH₂Cl₂-MeOH gradient system (100:0–0:100, v/v), resulting in the separation of nine fractions (A-I). Fraction B (4.5 g) underwent Sephadex LH-20 and two subfractions (B₁ and B₂) were afforded. Subfraction B₂ (2.5 g) was fractionated by C₁₈ reversed-phase CC eluted with MeOH-H₂O (30:70-55:45, v/v) to given four subfractions (B_2a-B_2d). Subfraction B_2d was separated by Sephadex LH-20 and then subjected to further purification through semipreparative HPLC with MeOH-H₂O (65:35, v/v) to afford 1 (5.9 mg, t_R: 108.3 min), 8 (3.0 mg, t_R: 32.6 min), and 9 (3.0 mg, t_R: 31.1 min). Fraction C (2.2 g) was divided into two parts (C₁ and C₂) followed by Sephadex LH-20. Subfraction C₂ (1.5 g) underwent C₁₈ reversed-phase column eluting with MeOH-H₂O gradient system (30:70–60:40, v/v) to acquire four subfractions (C_2a-C_2d). Subfractions C_2a and C_2b were fractionated by Sephadex LH-20 and further purified through semipreparative HPLC with MeOH-H₂O (60:40 and 45:55, v/v) to obtain 3 (6.3 mg, t_R: 14.8 min) and 7 (4.7 mg, t_R: 16.2 min), respectively. 2 (5.6 mg, t_R: 19.3 min), 4 (8.2 mg, t_R: 20.8 min), 5 (5.4 mg, t_R: 20.0 min), 6 (13.1 mg, t_R: 18.6 min), and 10 (21.0 mg, t_R: 21.6 min) were isolated from subfraction C_2c, which was separated by Sephadex LH-20 and then subjected to further purification through semipreparative HPLC with MeOH-H₂O (55:45, v/v).

Neobrefeldin (1): colorless crystals, ${[\mathrm{\alpha }]}_{\mathrm{D}}^{25}$  + 10 (c 0.1, MeOH); ECD (MeOH) λ_max (Δε) 202 (+4.72), 208 (+17.16), 226 (+1.64), 240 (+2.78), 293 (−0.16) nm; UV (MeOH) λ_max (log ε) 202 (2.78) nm; IR (KBr) ν_max 3383, 2973, 2945, 2868, 2823, 1463, 1351, 1056, 1022, 1010 cm⁻¹; ¹H and ¹³C NMR data, Table 1; HRESIMS m/z 263.1652 [M–H]⁻ (calcd for C₁₆H₂₃O₃, 263.1647).

Brefeldin H (2): colorless crystals, ${[\mathrm{\alpha }]}_{\mathrm{D}}^{25}$  + 17 (c 0.1, MeOH); ECD (MeOH) λ_max (Δε) 200 (+16.07), 235 (−1.39), 257 (−0.45) nm; UV (MeOH) λ_max (log ε) 202 (2.51) nm; IR (KBr) ν_max 3379, 2973, 2868, 1697, 1462, 1352, 1056, 1022, 1010 cm⁻¹; ¹H and ¹³C NMR data, Table 1; HRESIMS m/z 295.1545 [M–H]⁻ (calcd for C₁₆H₂₃O₅, 295.1546).

Brefeldin I (3): colorless crystals, ${[\mathrm{\alpha }]}_{\mathrm{D}}^{25}$  + 9 (c 0.1, MeOH); ECD (MeOH) λ_max (Δε) 204 (+14.85), 220 (−12.65), 248(+12.51) nm; UV (MeOH) λ_max (log ε) 202 (2.68) nm; IR (KBr) ν_max 3379, 2972, 2868, 1698, 1459, 1352, 1056, 1022, 1010 cm⁻¹; ¹H and ¹³C NMR data, Table 1; HRESIMS m/z 295.1545 [M–H]⁻ (calcd for C₁₆H₂₃O₅, 295.1546).

Brefeldin J (4): colorless crystals, ${[\mathrm{\alpha }]}_{\mathrm{D}}^{25}$  + 16 (c 0.1, MeOH); ECD (MeOH) λ_max (Δε) 200 (+13.46), 221 (+0.57), 238 (+1.63), 288 (−1.99) nm; UV (MeOH) λ_max (log ε) 203 (2.62) nm; IR (KBr) ν_max 3383, 2973, 2868, 1688, 1462, 1352, 1056, 1022, 1010 cm⁻¹; ¹H and ¹³C NMR data, Table 2; HRESIMS m/z 281.1743 [M + H]⁺ (calcd for C₁₆H₂₅O₄, 281.1753);

Brefeldin K (5): colorless crystals, ${[\mathrm{\alpha }]}_{\mathrm{D}}^{25}$ −7 (c 0.1, MeOH); ECD (MeOH) λ_max (Δε) 211 (+16.69), 259 (−0.70) nm; UV (MeOH) λ_max (log ε) 202 (2.67) nm; IR (KBr) ν_max 3382, 2973, 2868, 1690, 1352, 1056, 1022, 1010 cm⁻¹; ¹H and ¹³C NMR data, Table 2; HRESIMS m/z 295.1545 [M–H]⁻ (calcd for C₁₆H₂₃O₅, 295.1546).

seco-4-epi-9-epi-BFA methyl ester (6): colorless and transparent oil, ${[\mathrm{\alpha }]}_{\mathrm{D}}^{25}$  + 23 (c 0.1, MeOH); ECD (MeOH) λ_max (Δε) 211 (+16.86), 252 (−0.71) nm; UV (MeOH) λ_max (log ε) 203 (2.58) nm; IR (KBr) ν_max 3380, 2972, 2868, 1706, 1351, 1056, 1021, 1010 cm⁻¹; ¹H and ¹³C NMR data, Table 2; HRESIMS m/z 313.2006 [M + H]⁺ (calcd for C₁₇H₂₉O₅, 313.2015).

4-epi-9-epi-BFA seco-acid (7): colorless and transparent oil, ${[\mathrm{\alpha }]}_{\mathrm{D}}^{25}$  + 14 (c 0.1, MeOH); ECD (MeOH) λ_max (Δε) 216 (+17.49), 250 (−0.52) nm; UV (MeOH) λ_max (log ε) 202 (2.34) nm; IR (KBr) ν_max 3380, 2972, 2868, 1689, 1351, 1056, 1021, 1010 cm⁻¹; ¹H and ¹³C NMR data, Table 3; HRESIMS m/z 297.1706 [M–H]⁻ (calcd for C₁₆H₂₅O₅, 297.1702).

5-epi-7-dehydrobrefeldin F (8): colorless and transparent oil, ${[\mathrm{\alpha }]}_{\mathrm{D}}^{25}$  + 17 (c 0.1, MeOH); ECD (MeOH) λ_max (Δε) 204 (−10.32), 224 (+7.70), 248 (−3.16) nm; UV (MeOH) λ_max (log ε) 202 (2.37) nm; IR (KBr) ν_max 3380, 2971, 2868, 1747, 1682, 1352, 1206, 1140, 1056, 1022, 1010 cm⁻¹; ¹H and ¹³C NMR data, Table 3; HRESIMS m/z 263.1651 [M–H]⁻ (calcd for C₁₆H₂₃O₃, 263.1647).

9-epi-7-dehydrobrefeldin F (9): colorless and transparent oil, ${[\mathrm{\alpha }]}_{\mathrm{D}}^{25}$ −6 (c 0.1, MeOH); ECD (MeOH) λ_max (Δε) 218 (−16.89), 250 (+1.93) nm; UV (MeOH) λ_max (log ε) 203 (2.40) nm; IR (KBr) ν_max 3380, 2972, 2868, 1748, 1458, 1379, 1056, 1024, 1010 cm⁻¹; ¹H and ¹³C NMR data, Table 3; HRESIMS m/z 263.1652 [M–H]⁻ (calcd for C₁₆H₂₃O₃, 263.1647).

2.4 Preparation of Mosher ester derivatives 6s and 6r

Compound 6 (1.5 mg) was dissolved in dry CH₂Cl₂ (0.5 mL), subjected to a catalytic amount of 4-N, N-dimethylaminopyridine (DMAP) 10 mg and an excess of (S)-(−)-α-methoxy-α-(trifluoromethyl) phenylacetic acid (MTPA) 33.72 mg. The mixture was maintained with stirring for 8 h at room temperature. (S)-MTPA ester 6s (3.2 mg) was isolated from the reaction mixture by semipreparative HPLC after solvent removal. (R)-MTPA ester 6r (3.6 mg) was prepared by (R)-MTPA under the same conditions described above.

2.5 NMR and ECD calculations

First of all, conformational analyses of 1–9 were implemented using Merck Molecular Force Field (MMFF) and the conformers with Boltzmann-population >1 % were selected to consecutively reoptimize. Then, geometry optimizations and frequency analyses were conducted at the B3LYP/6-311G (d, p) level by density functional theory (DFT) and the conformers were checked the consistency with NOESY correlations and ¹H–¹H coupling constants. Subsequently, Gauge Independent Atomic Orbital (GIAO) calculations of ¹³C NMR chemical shifts used DFT at the B3LYP/6-311G (d, p) level with the PCM model in DMSO. The ¹³C NMR chemical shifts for TMS were used as the reference and calculated by the same procedures. The data were evaluated by linear correlation coefficients (R²) and the improved probability DP4+ method (Marcarino et al., 2022). Finally, ECD calculations of the stable conformers were carried out by time-dependent DFT (TDDFT) at the B3LYP/6-311G (d, p) level in methanol solution. ECD spectra were simulated by the software SpecDis 1.71 (Bruhn et al., 2013).

2.6 Cholinesterase enzyme assay in vitro

The inhibitory activities of 1–10 against ChEs were performed by Ellman's method in a 96-well plate with slight modifications (Ellman et al., 1961). AChE (EC3.1.1.7) from Electrophorus electricus (electric eel) and BuChE (EC3.1.1.8) from equine serum were purchased from Macklin. Acetylthiocholine iodide (ATChI) and butyrylthiocholine iodide (BuTChI) as substrates. Stock solutions of the tested compounds and positive control were prepared in DMSO to 2 mM. These solutions were gradually diluted in phosphate buffer saline (PBS) (pH = 7.4, 0.1 M). Briefly, 20 μL enzyme solution (AChE or BuChE, 1.0 U/mL), 20 μL various concentrations compounds, and 130 μL PBS was mixed and incubated at room temperature for 15 min. The reaction was initiated by adding 10 μL 5, 5′-dithiobis-2-nitrobenzoic acid (DTNB) as well as 20 μL substrates, and incubated at 37 °C for 15 min. The OD values were recorded with microplate reader at 412 nm. Galantamine was used as positive control. The IC₅₀ values were determined from triplicate logarithmic concentration inhibition curves. The inhibition rate of ChEs was calculated as the equation: $$\mathrm{I}\mathrm{n}\mathrm{h}\mathrm{i}\mathrm{b}\mathrm{i}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e}(\%)=\left(1-\frac{\text{mean}\text{of}\text{O.D.}\text{of}\text{test}\text{compound}}{\mathrm{m}\mathrm{e}\mathrm{a}\mathrm{n}\mathrm{o}\mathrm{f}\mathrm{O}.\mathrm{D}.\mathrm{o}\mathrm{f}\mathrm{n}\mathrm{e}\mathrm{g}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{e}\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{o}\mathrm{l}}\right)\times 100\%.$$

2.7 Molecular docking

To investigate the potential active sites and putative binding modes of compounds 1, 2, and galantamine, molecular docking studies were performed by the Autodock 4.2 program (https://autodock.scripps.edu/). The X-ray crystal structure of AChE (PDB ID: 4EY6) was downloaded from the Protein Data Bank (PDB). Co-crystallized ligand galantamine, free ions, and water molecules were removed from the receptor using the PyMOL. The receptor AChE was further optimized through the Autodock Tools 1.5.4 (Morris et al., 2009). The ligands (1, 2, and galantamine) were sketched in the Chem 3D and further optimized using the Autodock Tools 1.5.4. The binding pocket was selected as reported cognate ligands previously (Cheung et al., 2012). The poses exhibiting most likely binding mode with the lowest binding free energy were selected and visualized using PyMOL.

3.1 Structure elucidation

Neobrefeldin (1) was obtained as colorless crystals with molecular formula C₁₆H₂₄O₃ by HRESIMS (m/z 263.1652 [M–H]⁻, calcd for C₁₆H₂₃O₃, 263.1647). Based on the analysis of NMR data, 1 was identified as a BFA analogue (Bai et al., 2022) (Table S1). Comparing the ¹H NMR data of 1 with those of BFA revealed that the primary distinctions between them existed in the coupling constants of H-2/H-3 (J = 11.4 Hz) of 1 (Table 1). These changes suggested generation of a cis-arranged double bond between C-2 and C-3. Then, the planar construction was further confirmed by HMBC and ¹H–¹H COSY spectra (Fig. 2). The coupling constant between H-10 and H-11 (J = 15.4 Hz) indicated a trans-arrangement of the double bond. The NOESY correlations between H-4 and H-5, as well as H-5 and H-9, suggested that H-4, H-5, and H-9 were all orientated in the β-configuration (Fig. 3). Based on the literature, when Me-16 is α-oriented, δ_C-16 is 18 ppm. On the contrary, when Me-16 is β-oriented, δ_C-16 is 21 ppm (Zeng et al., 2019). Thus, Me-16 (δ_C 18.3 ppm) of 1 was deduced to be α-oriented (Table 1). The absolute configuration of 1 was confirmed as 1a (2Z, 4R, 5S, 9S, 10E, 15R) by comparing the calculated ECD curve with the experimental ECD spectrum of 1 (Fig. 4). Since the cis-arranged double bond between C-2 and C-3 of 1 have been reported for the first time in BFA analogues with 16-membered lactone, 1 is named as neobrefeldin.

Close

Fig. 2

Key ¹H–¹H COSY and HMBC correlations of compounds 1–9.

Export to PPT

Close

Fig. 3

Key NOESY correlations of compounds 1–9.

Export to PPT

Close

Fig. 4

Calculated ECD spectra and experimental ECD spectra of 1–9.

Export to PPT

Brefeldin H (2), isolated as colorless crystals, was determined to have a molecular formula C₁₆H₂₄O₅ by HRESIMS (m/z 295.1545 [M–H]⁻, calcd for C₁₆H₂₃O₅, 295.1546). Comparison of the 1D NMR data of 2 with 4-epi-15-epi-BFA indicated that their structures were similar (Bai et al., 2022). However, the chemical shift of C-8 (δ_C 79.9 ppm) in 2 was significantly downfield shifted compared to that of C-8 (δ_C 43.0 ppm) in 4-epi-15-epi-BFA (Tables 1 and S2). Considering that 2 had a molecular weight (MW) 16 Daltons higher than 4-epi-15-epi-BFA, it could be inferred that a hydroxyl group was linked at C-8 of 2. This putative structure was further confirmed on the basis of the HMBC correlations of H-8/C-5, 6 and the ¹H–¹H COSY correlations of H-2∼ H-16 (Fig. 2). The small J value H-4/H-5 (J = 2.0 Hz) revealed that H-4 and H-5 were on the same side (Table 1). The NOESY correlations of H-4/H-5 and H-8/H-9, together with the absence of correlations between H-5/H-9 and H-7/H-8 indicated that H-4, H-5, H-7, H-8, and H-9 were α-, α-, α-, β-, and β-orientated, respectively (Fig. 3).

Brefeldin I (3), isolated as colorless crystals, was determined to have a molecular formula of C₁₆H₂₄O₅ on the basis of its HRESIMS (m/z 295.1545 [M–H]⁻, calcd for C₁₆H₂₃O₅, 295.1546). The NMR data of 3 closely resembled those of 4-epi-15-epi-BFA (Bai et al., 2022), with the prime difference being the dramatically shifted downfield chemical shift of C-12 (δ_C 68.8 ppm) in 3, compared to that of C-12 (δ_C 32.0 ppm) in 4-epi-15-epi-brefeldin A (Tables 1 and S2). Due to the MW of 3 was 16 Dalton greater than that of 4-epi-15-epi-BFA, this indicated that 3 had a hydroxyl group linked at C-12. The HMBC correlations of H-10, 14/C-12 and the ¹H–¹H COSY correlations of H-2 ∼ H-16 further confirmed this speculated structure (Fig. 2). The coupling constant of H-4/H-5 (J = 2.3 Hz) was small, suggesting H-4 and H-5 were located on the same side (Table 1). The NOESY correlation of H-4/H-5 and the non-correlation between H-5/H-9 indicated that H-4, H-5, and H-9 were α-, α-, and β-orientated, respectively (Fig. 3). However, the relative configurations of C-7 and C-12 were not identified by NOESY spectrum. To assignment the relative configuration of compound 3, the calculated NMR of four possible conformers (2E, 4S*, 5R*, 7R*, 9S*, 10E, 12R*, 15R*)-3a, (2E, 4S*, 5R*, 7S*, 9S*, 10E, 12S*, 15R*)-3b, (2E, 4S*, 5R*, 7R*, 9S*, 10E, 12S*, 15R*)-3c, and (2E, 4S*, 5R*, 7S*, 9S*, 10E, 12R*, 15R*)-3d were compared with experimental NMR. The linear correlation analysis towards the experimental and the calculated ¹³C NMR and DP4+ probability analysis results were shown in Fig. 5A. With a DP4+ probability of 96.61 %, the relative configuration of compound 3 was assigned as (2E, 4S*, 5R*, 7R*, 9S*, 10E, 12S*, 15R*)-3c.

Close

Fig. 5

NMR calculations with DP4+ analyses. (A) Linear correlations between the scaled calculated and experimental ¹³C NMR chemical shifts for compound 3 and DP4+ probability of ¹³C NMR chemical shifts of 3. (B) Linear correlations between the scaled calculated and experimental ¹³C NMR chemical shifts for compound 4 and DP4+ probability of ¹³C NMR chemical shifts of 4. (C) Linear correlations between the scaled calculated and experimental ¹³C NMR chemical shifts for compound 5 and DP4+ probability of ¹³C NMR chemical shifts of 5. (D) Linear correlations between the scaled calculated and experimental ¹³C NMR chemical shifts for compound 6 and DP4+ probability of ¹³C NMR chemical shifts of 6.

Export to PPT

Brefeldin J (4) was isolated as colorless crystals. Based on the HRESIMS (m/z 281.1743 [M+H]⁺, calcd for C₁₆H₂₅O₄, 281.1753), its molecular formula was assigned as C₁₆H₂₄O₄. The NMR data of 4 closely resembled that of BFA, suggested that they shared a common planar structure and were likely stereoisomers (Table S1 and Fig. 2). The small J value H-4/H-5 (J = 2.7 Hz) revealed that H-4 and H-5 were on the same side (Table 2). The NOESY correlations between H-4 and H-5, H-5 and H-9, suggested that H-4, H-5, and H-9 were β-, β-, and β-orientated, respectively (Fig. 3). However, no effective NOE correlations can confirm the relative configuration of C-7. Thus, a calculated ¹³C NMR method was operated and the ¹³C NMR data of (2E, 4R*, 5S*, 7S*, 9S*, 10E, 15S*)-4a and (2E, 4R*, 5S*, 7R*, 9S*, 10E, 15S*)-4b were calculated. As shown in Fig. 5B, with a DP4+ probability of 100 %, the relative configuration of compound 4 was assigned as (2E, 4R*, 5S*, 7S*, 9S*, 10E, 15S*)-4a.

Brefeldin K (5) was isolated as colorless crystals. Based on the HRESIMS (m/z 295.1545 [M–H]⁻, calcd for C₁₆H₂₃O₅, 295.1546), the molecular formula C₁₆H₂₄O₅ was assigned. Comparing the 1D NMR data of 5 with 4 indicated that their structures were similar (Table 2). However, the chemical shift of C-5 (δ_C 82.5 ppm) in 5 showed a significant downfield shift compared to that of C-5 (δ_C 47.3 ppm) in 4 (Table 2). The MW of 5 was 16 Daltons higher than that of 4, indicating the presence of a hydroxyl group linked at C-5 of 5. The ¹H–¹H COSY correlations and the HMBC correlations from H-4 to C-5 and from H-6 to C-5 further confirmed this putative structure (Fig. 2). The relative configuration of 5 was fixed from NMR calculation and four possible conformers (2E, 4R*, 5S*, 7R*, 9S*, 10E, 15R*)-5a, (2E, 4R*, 5S*, 7S*, 9R*, 10E, 15R*)-5b, (2E, 4R*, 5R*, 7R*, 9S*, 10E, 15R*)-5c, and (2E, 4R*, 5R*, 7S*, 9R*, 10E, 15R*)-5d were calculated. The results fully matched the 5a configuration for compound 5 with 100.00 % probability (Fig. 5C).

The double bond configurations of H-2/H-3 and H-10/H-11 in compounds 2–5 were trans-arranged based on the coupling constants (J = 15–16 Hz). According to the description above, the configuration of Me-16 can be determined by the chemical shift of methyl carbon (Zeng et al., 2019). Thus, the Me-16 of 2 (δ_C 17.7 ppm), 3 (δ_C 17.4 ppm), 4 (δ_C 20.8 ppm), 5 (δ_C 18.0 ppm) was determined to be α, α, β, α-oriented, respectively (Table 1).

seco-4-epi-9-epi-BFA methyl ester (6), obtained as a colorless and transparent oil, its molecular formula was C₁₇H₂₈O₅ based on m/z 313.2006 [M+H]⁺ (calcd for C₁₇H₂₉O₅, 313.2015). The NMR data of 6 closely resembled those of seco-4-epi-7-epi-BFA methyl ester, suggested that they shared a common planar structure and were likely stereoisomers (Bai et al., 2022) (Table 2 and Fig. 2). The trans-arrangement of the double bonds of H-2/H-3 and H-10/H-11 were inferred according to the coupling constants (J = 15.5 and 15.2 Hz). The small J value H-4/H-5 (J = 5.0 Hz) indicated that H-4 and H-5 were on the same side (Table 2). The NOESY correlations between H-4 and H-5, H-5 and H-9 suggested that H-4, H-5, and H-9 were α-, α-, and α-orientated, respectively (Fig. 3). However, no effective NOE correlations can confirm the relative configuration of C-7. Thus, the calculated NMR of two possible conformers (2E, 4S*, 5R*, 7S*, 9R*, 10E, 15S*)-6a and (2E, 4S*, 5R*, 7R*, 9R*, 10E*, 15S*)-6b were calculated. The results fully matched the 6a configuration for compound 6 with 99.97 % probability (Fig. 5D).

4-epi-9-epi-BFA seco-acid (7), isolated as a colorless and transparent oil, was given a molecular formula C₁₆H₂₆O₅ according to the HRESIMS (m/z 297.1706 [M–H]⁻, calcd for C₁₆H₂₅O₅, 297.1702). Comparing the NMR data of 7 with 6 revealed the substitution of the methyl ester from 6 with a methanoic acid in 7 (Table 3 and Fig. 2). Therefore, the planar structure of 7 was the same as the known BFA analogue 4-epi-7-epi-BFA seco-acid (Bai et al., 2022). A combination of ¹H–¹H coupling constants and NOESY spectrum indicated that the relative configuration of 7 was the same as that of 6 (Fig. 3).

5-epi-7-dehydrobrefeldin F (8) was afforded as a colorless and transparent oil. According to the HRESIMS (m/z 263.1651 [M–H]⁻, calcd for C₁₆H₂₃O₃, 263.1647), the molecular formula of C₁₆H₂₄O₃ was assigned. Based on NMR data, 8 was speculated to be a 7 analogue. The comparison of the ¹³C NMR data for 8 and 7 unveiled notable disparities, particularly in the downfield shifted resonances observed at C-1 (δ_C 173.1), C-2 (δ_C 120.7), C-3 (δ_C 158.3), and C-4 (δ_C 84.7) (Table 1). In addition, 8 (Ω = 5) had one more unsaturation than 7 (Ω = 4). The observed changes indicated an α, β-unsaturated γ-lactone ring was formed between C-1 and C-4. Then, the ¹H–¹H COSY and HMBC spectra further confirmed this putative structure (Fig. 2). According to the coupling constants (J = 5.7 and 15.2 Hz), the double bond configurations of H-2/H-3 and H-10/H-11 were confirmed to be cis and trans, respectively. The small J value of H-4/H-5 (J = 6.3 Hz) revealed that H-4 and H-5 were on the same side. The NOESY correlations of H-4/H-5 and H-5/H-9 suggested that H-4, H-5, and H-9 were all orientated in the β-configuration (Fig. 3).

9-epi-7-dehydrobrefeldin F (9) was given as a colorless and transparent oil. According to the HRESIMS (m/z 263.1652 [M–H]⁻, calcd for C₁₆H₂₃O₃, 263.1647), the molecular formula of C₁₆H₂₄O₃ was assigned. Comparing NMR data of 9 with 8 suggested that 9 shared a common planar structure with 8 and they were likely stereoisomers. (Table 3 and Fig. 2). In addition, the large J value H-4/H-5 (J = 9.0 Hz) indicated that H-4 and H-5 oriented to the different direction. The NOESY correlations H-5/H-9 and the absence of correlations between H-4 and H-5 revealed that H-4, H-5, and H-9 were orientated in the β-, α-, and α-configuration, respectively (Fig. 3).

According to the modified Mosher's method, the absolute configuration of C-15 was unambiguously confirmed to be S, based on the Δδ_H values observed between the (S)- and (R)-MTPA esters of 6 (Fig. 6). The comparison of 1D NMR data revealed that the absolute configuration of C-15 in 7 (δ_H-15 3.55, δ_C-15 65.6), 8 (δ_H-15 3.54, δ_C-15 65.6), and 9 (δ_H-15 3.54, δ_C-15 65.6) was the same as that of 6 (δ_H-15 3.54, δ_C-15 65.6).

Close

Fig. 6

Δδ_H values obtained for (S)- and (R)-MTPA esters of 6.

Export to PPT

In order to define the absolute configurations of 2–9, TDDFT ECD calculations were conducted at the PCM/B3LYP/6-311G (d, p) level of theory. The calculated ECD spectra of compounds 2a (2E, 4S, 5R, 7S, 8S, 9S, 10E, 15R), 3c (2E, 4S, 5R, 7R, 9S, 10E, 12S, 15R), 4a (2E, 4R, 5S, 7S, 9S, 10E, 15S), 5a (2E, 4R, 5S, 7R, 9S, 10E, 15R), 6a (2E, 4S, 5R, 7S, 9R, 10E, 15S), 7a (2E, 4S, 5R, 7S, 9R, 10E, 15S), 8a (2Z, 4R, 5S, 9S, 10E, 15S), and 9a (2Z, 4S, 5S, 9S, 10E, 15S) demonstrated excellent consistency with those of the experimental ECD curves (Fig. 4). Hence, the absolute configurations of 1–9 were defined as depicted.

The known compound 10 were identified as 4-epi-brefeldin A by comparing spectroscopic data with literature values (Xiong and Hale, 2016).

3.2 In vitro cholinesterase enzyme assay

The anticholinesterase activity against AChE and BuChE of 1–10 were tested to evaluate the potential anti-AD activity (Table 4). The results suggested that 1–5 and 10 displayed significant inhibitory activity against AChE. It's worth noting that 1 and 2 showed more potent AChE inhibitory activity (IC₅₀ = 0.12 and 0.28 µM) than the positive control galantamine (IC₅₀ = 0.66 µM). Notably, neobrefeldin (1) exhibited the most significant inhibitory activity against AChE, surpassing that of the therapeutic AD drug galantamine by over 5.5-fold. However, besides 1 showed weak inhibitory BuChE activity (IC₅₀ = 175.04 µM), others displayed no inhibitory BuChE activity (IC₅₀ > 200 μM) (Table 4). Comparing the AChE inhibitory activity of 6–9 (the ring opened BFA derivatives) with 1–5 and 10 (Table 4), it was found that the AChE activities of BFA derivatives were significantly influenced by the conformational rigidity of their 16-membered lactone. Due to 80 % of ACh is regulated by AChE in the early stages of AD, 1 and 2 may be as potent and selective AChE inhibitors for developing early anti-AD agents.

3.3 Molecular docking studies

The active site of the AChE lies near the bottom of a deep and narrow gorge, which comprises a peripheral anionic site (PAS) and a catalytic anionic site (CAS) (Sussman et al., 1991). CAS catalytic triad is composed of Ser203, His447, and Glu334. PAS contains amino acids Tyr72, Trp86, Tyr124, and so on (Wang and Bi, 2014). To investigate the putative binding mode of the compounds 1 and 2 interactions within the AChE active site, docking analysis was performed. As Fig. 7 and Table 5 showed compound 1 interacted with Tyr124 through hydrogen bonding (HB), together with Trp86, Phe295, Phe297, Tyr337, and Phe338 through hydrophobic-hydrophobic interaction. Compound 2 exhibited several key HB with Glu202, Ser203, and Tyr337, along with hydrophobic-hydrophobic interactions with Tyr124, Phe297, Tyr337, and Phe338. Galantamine showed key HB with Ser203 and His447, as well as hydrophobic-hydrophobic interactions with Phe297 and Tyr337. Thus, compounds 1 and 2 are dual binding site (CAS and PAS) AChE inhibitors, which are different with single binding site (CAS) AChE inhibitor galantamine.

Close

Fig. 7

Visual representation of compounds 1, 2 and galantamine (yellow) docked with AChE (PDB ID: 4EY6), hydrogen bonds are shown with yellow dotted lines, important amino acid residues involved in the enzyme-ligand interaction in light-blue.

Export to PPT

ACh binds to PAS, which is the first step in the process of catalytic hydrolysis, and then interacts with CAS. Even though compound 2 displays similarly binding site (CAS) compared to galantamine, the higher inhibition potency of compound 2 can be attributed to docking with PAS (Tyr124). Compound 1 shows a better binding affinity with the PAS through the HB interactions with Tyr124 (1.95 Å) than compound 2, docked with PAS through the hydrophobic-hydrophobic interactions with Tyr124 (3.59 Å) at the mouth of the gorge (Fig. 7 and Table 5). Maybe this is the reason compound 1 has the strongest AChE inhibitory activity (IC₅₀ = 0.12 μM). Thus, the AChE inhibitory activity maybe significantly improved by increasing binding affinity with the PAS, which need to be verified by kinetic inhibition tests and enzyme co-crystallization experiments et al.. Most important of all, these findings provide new ideas for the design of AChE inhibitor and the structure optimization of BFA compounds.