Xenoacremones D–H, Bioactive Tyrosine-decahydrofluorene Analogues from the Plant-Derived Fungus Xenoacremonium sinensis

Five novel tyrosine-decahydrofluorene analogues, xenoacremones D–H (1–5), each bearing a fused 6/5/6 tricarbocyclic core and a 13-membered para-cyclophane ring system, were isolated from the endophytic fungus Xenoacremonium sinensis. Compound 1 was a novel polyketide synthase–nonribosomal peptide synthetase (PKS–NRPS) tyrosine-decahydrofluorene hybrid containing a 6/5/6/6/5 ring system. Their structures were elucidated from comprehensive spectroscopic analysis and electronic circular dichroism (ECD) calculations. All compounds were evaluated for their inhibitory activities on LPS-induced NO production in macrophages and their cytotoxicities against the NB4 and U937 cell lines. Compounds 3 and 5 exhibited potent anti-inflammatory activities in vitro. Compounds 1 and 3–5 displayed significant antiproliferative activity against the tumor cell lines (IC50 < 20 µM).

Analysis of its 2D NMR data confirmed the whole structure, which had a tyrosine-decahydrofluorene skeleton and resembled that of hirsutellone B and pyrrospirone A [9,13]. The proton spin systems from H-1 to H-15 observed in the 1 H-1 H COSY spectrum, as well as the HMBC correlations from H-7 to C-6, C-11 and C-13 and from H-14 to C-3, C-5, C-6 and C-15, indicated the presence of a decahydrofluorene moiety. Furthermore, the 1 H-1 H COSY cross-peaks of H-1/H-2/H-3 and the HMBC correlations from H-1 to C-3, C-16, C-17 and C-18 and from H-1' to C-1, C-16 and C-17 revealed the presence of a methylene (C-1), an oxygenated methine (C-2) and a quaternary carbon (C-17) in 1, which were different from those of hirsutellone B (Figure 2). The HMBC correlations and the degrees of unsaturation indicated that the methylene at C-1 was linked at C-17 to form a cyclohexane moiety, and C-17 was the connectivity of a spiro center between the cyclohexane and γ-lactam ring. Additional HMBC correlations from H-1' and H-3' to a quaternary carbon C-2' (δC 83.5) led to the assignment of C-2' for the γ-position of the lactam ring, and its up-field shift revealed the attachment of a hydroxyl group. In addi-
Analysis of its 2D NMR data confirmed the whole structure, which had a tyrosinedecahydrofluorene skeleton and resembled that of hirsutellone B and pyrrospirone A [9,13]. The proton spin systems from H-1 to H-15 observed in the 1 H-1 H COSY spectrum, as well as the HMBC correlations from H-7 to C-6, C-11 and C-13 and from H-14 to C-3, C-5, C-6 and C-15, indicated the presence of a decahydrofluorene moiety. Furthermore, the 1 H-1 H COSY cross-peaks of H-1/H-2/H-3 and the HMBC correlations from H-1 to C-3, C-16, C-17 and C-18 and from H-1' to C-1, C-16 and C-17 revealed the presence of a methylene (C-1), an oxygenated methine (C-2) and a quaternary carbon (C-17) in 1, which were different from those of hirsutellone B (Figure 2). The HMBC correlations and the degrees of unsaturation indicated that the methylene at C-1 was linked at C-17 to form a cyclohexane moiety, and C-17 was the connectivity of a spiro center between the cyclohexane and γ-lactam ring. Additional HMBC correlations from H-1' and H-3' to a quaternary carbon C-2' (δ C 83.5) led to the assignment of C-2' for the γ-position of the lactam ring, and its up-field shift revealed the attachment of a hydroxyl group. In addition, the HMBC correlations from H-1' to C-3', C-16 and C-18 and from H-3' to C-5' and C-9' completed the linkages of the phenyl and 6/5/6/6/5 pentacarbocyclic moieties to form the 13-membered macrocyclic ether of 1. Consequently, its planar structure containing a spiro-ring system was determined. tion, the HMBC correlations from H-1' to C-3', C-16 and C-18 and from H-3' to C-5' and C-9' completed the linkages of the phenyl and 6/5/6/6/5 pentacarbocyclic moieties to form the 13-membered macrocyclic ether of 1. Consequently, its planar structure containing a spiro-ring system was determined. Hz) between H-12 and H-13 implied their trans-configuration, and H-13 was axial. Moreover, the ∆ 4,5 geometry was assigned as Z by its coupling constant (J = 9.1 Hz). To determine the absolute stereochemistry of 1, the theoretically calculated electronic circular dichroism (ECD) spectra were obtained by time-dependent density functional theory (TDDFT). The conformations of 1a and 1b (1b was the enantiomer of 1a) were compared using ECD calculations at the B3LYP level. The Cotton effects were identical with the calculated curve of the enantiomer 1a (Figure 4), which confirmed its absolute configuration as 2R,3S,6S,7S,9R,11S,12R,13S,14S,15S,17S,2'R.
The relative configuration of 1 was ascertained by the NOESY experiment. The NOE cross-peaks of H-7/H-11, H-7/H-9, H-12/H-8a and H-12/H-10a indicated these protons were axial positions. Additional NOE cross-peaks of H-7/H-14, and H-13/H-6' placed these protons in the β-orientation (Figure 3). However, the NOE cross-peaks of H-6/H-12 and H-15/H-3 indicated that they were co-facial. Further NOE cross-peaks of H-1/H-1'β and H-1α/H-3 revealed that the methylene at C-1 was β-oriented and C-18 of the γ-lactam was α-oriented. The vicinal coupling constant (J12,13 = 8.3 Hz) between H-12 and H-13 implied their trans-configuration, and H-13 was axial. Moreover, the Δ 4,5 geometry was assigned as Z by its coupling constant (J = 9.1 Hz). To determine the absolute stereochemistry of 1, the theoretically calculated electronic circular dichroism (ECD) spectra were obtained by time-dependent density functional theory (TDDFT). The conformations of 1a and 1b (1b was the enantiomer of 1a) were compared using ECD calculations at the B3LYP level. The Cotton effects were identical with the calculated curve of the enantiomer 1a (Figure 4), which confirmed its absolute configuration as 2R,3S,6S,7S,9R,11S,12R,13S,14S,15S,17S,2'R.   Xenoacremone E (2) was shown via HRESIMS to have a molecular formula of C29H33NO6. The 1 H and 13 C NMR data (Table 1) of 2 were also similar to those of hirsutellone B. The position of the double bond in 2 was deduced by the 1 Figure 2). Further HMBC correlations from H-3 to C-4 and C-5 and from H-2 to C-4, as well as the chemical shifts of H-4 (δH 3.84, dd, J = 4.2, 1.2 Hz) and C-4 (δC 67.3), indicated the presence of a hydroxyl group at C-4. Additional HMBC correlations from H-3' to C-1' (δC 63.5) and from H-1' (δH 3.6, s) to the quaternary carbons C-17 (δC 58.7), C-18 and C-2', combined with the degree of unsaturation of 2, indicated the presence of an epoxide moiety at C-1' and C-17 in the γ-lactam ring. The NOE correlations (Figure 3) of H-14 to H-3/H-7, H-15 to H-1' and H-1' to H-9' indicated that these hydrogens were co-facial and β-oriented, while the correlations of H-13 to H-6', H-10α to H3-19/H3-20 hinted that these hydrogens were α-oriented. The ECD spectrum of 2 was determined and the Cotton effects were identical with the calculated curve of the enantiomer 2a ( Figure 4). Thus, the absolute configuration of 2 was assigned (3S,4S,7R,9R,11S,12R,13R,14R,15R,17S,1'S,2'R) by comparison of the experimental and calculated ECD spectra (Figure 4). Xenoacremone F (3) was determined to be C30H37NO5 on the basis of its HR ESIMS and NMR spectra ( Table 1). Comparison of its NMR data with those of 2 revealed that 3 had one more methoxyl group (δC 49.7, δH 3.24) and one less epoxide group than 2. The HMBC correlations from -OCH3 to C-2' confirmed the assignment of the methoxyl group Xenoacremone E (2) was shown via HRESIMS to have a molecular formula of C 29 H 33 NO 6 . The 1 H and 13 C NMR data (Table 1) of 2 were also similar to those of hirsutellone B. The position of the double bond in 2 was deduced by the 1 Figure 2). Further HMBC correlations from H-3 to C-4 and C-5 and from H-2 to C-4, as well as the chemical shifts of H-4 (δ H 3.84, dd, J = 4.2, 1.2 Hz) and C-4 (δ C 67.3), indicated the presence of a hydroxyl group at C-4. Additional HMBC correlations from H-3' to C-1' (δ C 63.5) and from H-1' (δ H 3.6, s) to the quaternary carbons C-17 (δ C 58.7), C-18 and C-2', combined with the degree of unsaturation of 2, indicated the presence of an epoxide moiety at C-1' and C-17 in the γ-lactam ring. The NOE correlations (Figure 3) of H-14 to H-3/H-7, H-15 to H-1' and H-1' to H-9' indicated that these hydrogens were co-facial and β-oriented, while the correlations of H-13 to H-6', H-10α to H 3 -19/H 3 -20 hinted that these hydrogens were α-oriented. The ECD spectrum of 2 was determined and the Cotton effects were identical with the calculated curve of the enantiomer 2a (Figure 4). Thus, the absolute configuration of 2 was assigned (3S,4S,7R,9R,11S,12R,13R,14R,15R,17S,1'S,2'R) by comparison of the experimental and calculated ECD spectra (Figure 4). Xenoacremone F (3) was determined to be C 30 H 37 NO 5 on the basis of its HR ESIMS and NMR spectra ( Table 1). Comparison of its NMR data with those of 2 revealed that 3 had one more methoxyl group (δ C 49.7, δ H 3.24) and one less epoxide group than 2. The HMBC correlations from -OCH 3 to C-2' confirmed the assignment of the methoxyl group at C-2' (Figure 2). Further HMBC correlations from H-1' to C-17, C-18 and C-2' indicated that the hydroxyl group was located at C-17 in the γ-lactam ring. The NOE correlations ( Figure S3) of H-6 to H-7/H-14, H-7 to H-9/H-11/H-14, H-15 to H-1' and H-3'β to H-9' indicated that these hydrogens were in the β-orientation, similar to those of 2. The NOE correlations of H-13 to H-6', H-9' to H-1'α and H-3'α, as well as H-10α to H 3 -19/H 3 -20/H-12 indicated their α-orientation. The ECD spectrum was determined to elucidate the absolute configuration, which was compared to the experimental ECD curve. The ECD spectrum of 3 generated for the tyrosine-decahydrofluorene rings resembled that of compound 2, which was consistent with the experimental data of 3a ( Figure S4). Therefore, the absolute configuration of 3 was established as depicted in Figure 1.

H-1 H COSY cross-peaks between H-3 and H-5 and key HMBC correlations from H-7, H-14 and H-5 to C-6 and from H-7, H-14 and H-3 to C-5 (
Xenoacremone G (4) has the molecular formula C 29 H 33 NO 4 with 14 degrees of unsaturation, as determined by HR ESIMS data. The 1 H and 13 C NMR spectroscopic data of 4 ( Table 1) were similar to those of 3, and the differences were the absence of the methoxyl group at C-2' and the presence of an extra double bond (δ C 153.7, δ H 6.44; δ C 134.8) in 4. Further analysis of the 2D NMR data, particularly the HMBC correlations ( Figure 2) from H-1' to C-17 and C-2' and from H-3' to C-1', confirmed the location of the double bond at C-1' and C-17. The relative configuration of 4 was assigned as depicted by key NOE correlations ( Figure S3) of H-6 to H-14, H-14 to H-3, H-3'β to H-9', H-13 to H-6' and H-15 to H-1'. The absolute configuration of 4 was identified by comparing its experimental and calculated ECD data ( Figure S3). Therefore, 4 was confirmed as 3R,6R,7S,9R,11S,12R,13R,14S,15R,2'R, and it had ECD effects similar to those of 2 ( Figure S4).
The HR ESIMS data of xenoacremone H (5) suggested that it had the molecular formula of C 29 H 35 NO 4 . Comprehensive analysis of NMR data for the two compounds indicated that 5 possessed the similar planar structure as 4, where a pair of double bond in γ-lactam ring disappeared ( Table 1). The NOESY correlations ( Figure S3) of H-12 to H-6 and H-15 to H-6 placed these hydrogens in the α-orientation, while the correlations of H-7 to H-14, H-13 to H-11/H-8' and H-3 to H-14/H-17 indicated that these hydrogens were in the β-orientation, indicating that 5 had a similar stereochemistry than 1. The vicinal coupling constant (J 12,13 = 7.7 Hz) between H-12 and H-13 implied their trans-configuration, which was consistent with that of compound 1. The absolute configuration of 5 was determined by ECD calculation ( Figure S4). Its ECD curve was similar to that of 1 (Figure 4). The experimental ECD spectrum of 5 was in accordance with the calculated ECD spectrum for 5a. Therefore, the absolute configuration of 5 was established as 3R,6S,7S,9R,11S,12R,13S,14S,15S,17S,2'S.

Biological Assay
Compounds 1-5 were evaluated for their anti-inflammatory activities based on the inhibition of nitric oxide (NO) production in LPS-induced RAW264.7 macrophages. The results showed that compounds 3 and 5 exhibited significant inhibitory activities against the production of NO, with IC 50 values in the vicinity of 12.8 and 6.7 µM, respectively ( Table 2). Compound 5 showed the most potent inhibitory activity, which was stronger than the positive control resveratrol and compounds such as caffeic acid phenthyl ester (9.3 µM) and aspirin (43.2 µM). All the compounds were also examined for their cytotoxicity against the NB4 and U937 cell lines. Compounds 1 and 3-5 displayed cytotoxicities with IC 50 values less than 20 µM (Table 3). Table 2. Inhibitory effects of compounds 1-5 on NO production in LPS-induced RAW264.7 cells a .

Fungal Material
The endophytic fungus ML-31 was isolated from twigs of the mangrove plant Ceriops tagal collected in Hainan Province, China, in July 2013. The plant species was identified by Yi Sun, and the fungus was identified as X. sinensis on the basis of its rRNA gene sequence. The accession number for the biosynthetic cluster (xen) is MT876600 in the GenBank database at NCBI. The strain was deposited at the Institute of Chinese Materia Medica, China Academy of Chinese Medical Sciences.

Computational of ECD
A conformational search was carried out in the MMFF94 molecular mechanics force field using the MOE (Molecular Operating Environment) software package [27], and all the conformers within an energy window of 10 kcal/mol were regarded as the initial conformations. Monte Carlo protocols were used in the experiment. The geometry optimization and frequency calculations were performed with Gaussian16 RevB.01 [28], using the ωB97XD or B3LYP functional at the 6-311G (d,p) level of theory to verify the stability and obtain the energies at 298.15 K and 1 atm pressure. The Boltzmann distribution was calculated according to their Gibbs free energies. ECD calculations were conducted by using the Cam-B3LYP functional at the TZVP level of theory. The solvation model based on density (SMD) was used as the solvation model [29]. The Boltzmann-averaged ECD spectra were obtained by using SpecDis 1.71 software [30]. Methanol was used for structural optimization.

Cytotoxicity Assays
The cytotoxicities of 1-5 against human carcinoma cells of lines U937 and NB4 were tested using the CCK-8 method. The cells were sustained in RPMI-1640 supplemented with 10% (v/v) fetal bovine serum (FBS) and 0.5% (v/v) penicillin-streptomycin solution (10,000 units/mL penicillin and 10,000 µg/mL streptomycin, 100×) in a humidified atmosphere containing 5% CO 2 at 37 • C. The cells were digested by trypsinization and then diluted to a concentration of 1 × 10 4 cells/mL. The diluted cell suspensions were then placed into 96-well microtiter plates and incubated with the test samples for 72 h. The control contained 2 µL of MeOH. After incubation, CCK-8 solution was added to each well, and the plates were incubated for 4 h. The absorption was measured at a wavelength of 450 nm.

Assay of the Inhibition of NO Production in RAW264.7 Murine Macrophages
The nitrite concentration in the medium was measured as an indicator of NO production according to the Griess reaction. RAW 264.7 macrophages were seeded in three replicates at a density of 1 × 10 5 cells/well and incubated overnight at 37 • C with 5% CO 2 .