New Bioactive Sesquiterpeniods From the Plant Endophytic Fungus Pestalotiopsis theae

Three new secondary metabolites pestalothenins A–C (1–3), including two new humulane-derived sesquiterpeniods (1 and 2) and one new caryophyllene-derived sesquiterpeniod (3), together with five known compounds (4–8) were isolated from the crude extract of the plant endophytic fungus Pestalotiopsis theae (N635). Their structures were elucidated by the extensive analyses of HRESIMS and NMR spectroscopic data. The absolute configurations of 1–3 were determined by comparison of experimental and calculated electronic circular dichroism (ECD) spectra. The cytotoxic effects of these compounds were evaluated in vitro. Compound 6 showed moderate cytotoxicity against T24 and MCF7 cell lines. In addition, compounds 1–8 were also evaluated for antibacterial activity.


General Experimental Procedures
IR data were recorded using a Nicolet IS5 FT-IR spectrophotometer. NMR spectra were recorded on a Bruker Avance spectrometer operating at 400, 500, or 600 MHz with tetramethylsilane as an internal standard. HRESIMS data was obtained using an Agilent Accurate-Mass-Q-TOF LC/MS 6520 instrument. Optical rotation was measured by an Anton Qingdao Ocean Chemical Co.,Ltd.,China) and Sephadex LH-20 (Amersham Biosciences) were used for column chromatography (CC). HPLC analysis was performed with the Waters 2489 HPLC system using an octadecylsilyl (ODS) column (Pack, ReproSil-Pur Basic, 250 × 4.6 mm, 5 µm) with a flow rate of 1.0 mL/min. HPLC separation was performed on an Agilent HPLC instrument equipped with a variable-wavelength UV detector using an ODS column (C18, 250 × 9.4 mm, 5 µm) with a flow rate of 2.0 mL/min.

Fungal Material
The fungus P. theae has been previously described (Guo et al., 2020a).

Computation Section
Systematic conformational analyses were performed via the Molecular Operating Environment (MOE) ver. 2009.10.
(Chemical Computing Group, Canada) software package using the MMFF94 molecular mechanics force field calculation. The MMFF94 conformational analyses were further optimized using DFT at the B3LYP/6-311G(2d,p) basis set level. The stationary points have been checked as the true minima of the potential energy surface by verifying they do not exhibit vibrational imaginary frequencies. The 80 lowest electronic transitions were calculated at the B3LYP/6-311G(2d,p) level, and the rotational strengths of each electronic excitation were given using both dipole length and dipole velocity representations. ECD spectra were stimulated using a Gaussian function with a half-bandwidth of 0.3 eV. Equilibrium populations of conformers at 298.15 K were calculated from their relative free energies ( G) using Boltzmann statistics. The overall ECD spectra were then generated according to Boltzmann weighting of each conformer. The systematic errors in the prediction of the wavelength and excited-state energies are compensated for by employing UV correction.
In a 96-well plate, each well was plated with (2-5) × 10 3 cells (depending on the cell multiplication rate). After cell attachment overnight, the medium was removed, and each well was treated with 100 µL of medium containing 0.1% DMSO, or appropriate concentrations of the test compounds and the positive control cisplatin (100 mM as stock solution of a compound in DMSO and serial dilutions; the test compounds showed good solubility in DMSO and did not precipitate when added to the cells). The plate was incubated at 37 • C for 48 h in a humidified, 5% CO 2 atmosphere. Proliferation was assessed by adding 20 µL of MTS (Promega) to each well in the dark, followed by incubation at 37 • C for 90 min. The assay plate was read at 490 nm using a microplate reader. The assay was run in triplicate.

Antibacterial Assay
Antibacterial activities of compounds 1-8 were evaluated in replicate as per National Center for Clinical Laboratory Standards recommendations using broth micro dilution method to determine the MIC values with some modifications. In brief, the bacteria were grown in a LB medium (0.5% yeast extract, 1% peptone, 0.5% NaCl in deionized H 2 O). The assay was carried out in flat bottom 96-well microtiter plates. Microorganisms were pre-incubated at 37 • C for 24 h in medium. Compounds 1-8 were dissolved in DMSO at an initial concentration at 25 mg/mL, and then 1 µL was added to 149 µL medium in 96-well microtiter plates. Then the microorganism solution was added into the 96well plate (100 µL per well). The densities of the cells were approximately 1.0 × 10 6 CFU/mL. Positive control drugs were ampicillin (Sigma, purity > 900 µg/mg) for S. aureus and S. pneumoniae, gentamicin (Sigma, purity ≥ 99%) for E. coli and B. subtilis. After 24 h incubation, the absorbance was determined at 600 nm by a microplate reader. The MIC value was determined as the lowest concentration inhibiting microbial growth.  Figures 1, 2) and HSQC NMR (Supplementary Figure 4) spectroscopic data of 1 revealed the presence of four singlet methyl groups, three methylenes, one oxymethine, one sp 3 quarternary carbon, two trisubstituted olefin units, one carboxylic carbon (δ C 170.5), one aldehyde group (δ C 193.2; δ H 9.47) and two ketone carbons (δ C 203.3 and 209.5, respectively), which accounted for 6 • of unsaturation. Therefore, the remaining one unsaturation unit required that compound 1 possessed a monocyclic ring system. In the 1 H-1 H COSY spectrum (Figure 2 and Supplementary  Figure 3), homonuclear vicinal coupling correlations between H 2 -1 and H-2 and between H 2 -8 and H-9 confirmed the structural fragments of C-1-C-2 and C-8-C-9 in 1. In the HMBC spectrum (Figure 2 and Supplementary Figure 5), correlations from H-4 to C-2, C-3 and C-14 and from H-14 to C-2, C-3 and C-4 indicated that both C-4 and the aldehyde carbon C-14 (δ C 193.2) were directly connected to the C-2/C-3 olefin at C-3. The HMBC crosspeaks (Figure 2 and Supplementary Figure 5) from H 2 -1 to C-11, C-12 and C-13, and from the geminal methyl groups H 3 -12 and H 3 -13 to C-1, the ketone carbon C-10 (δ C 209.5) and the sp 3 quarternary carbon C-11 (δ C 47.3) indicated that C-1, C-10, C-12 and C-13 were all attached to C-11. Futhermore, HMBC correlations (Figure 2 and Supplementary Figure 5) from H 2 -8 to C-10 and from H-9 to C-10 and C-11 implied that the ketone carbon C-10 was located between C-9 and C-11. While the cross-peaks from H-9 and H 3 -17 to the carboxylic carbon C-16 (δ C 170.5) established the location of the acetyl group at C-9. Other HMBC correlations (Figure 2 and Supplementary Figure 5) from the olefinic proton H-6 to C-5, C-7, C-8 and C-15, from H 2 -8 to C-6, C-7 and C-15, and from H 3 -15 to C-6, C-7 and C-8 led to the completion of C-5-C-6-C-7-C-8 subunit with the methyl group C-15 attached to the   spectrum, the cross peak between H-2 and H-14 assigned the C-2/C-3 olefin as E-geometry. While the C-6/C-7 olefin of 1 was assigned as Z-geometry on the basis of NOESY (Figure 3 and Supplementary Figure 6) correlation of H-6 with H 3 -15.

Structure Elucidation
To determine the absolute configuration of 1, a comparison between the experimental and the simulated electronic circular dichroism (ECD) spectra (Figure 4) generated by the timedependent density functional theory (TDDFT) (Zhu et al., 2019) for two enantiomers (9S)-1 (1a) and (9R)-1 (1b) was performed (Figure 4). The MMFF94 conformational search and DFT re-optimization at the B3LYP/6-311G(2d,p) level yielded two lowest energy conformers for 1a (Supplementary Figure 19). The experimental ECD curve of 1 was nearly identical to the calculated ECD spectrum of 1a, suggested the 9S absolute configuration for 1. The molecular formula of pestalothenin B (2) was deduced as C 19 H 28 O 6 (6 • of unsaturation) by analysis of HRESIMS spectrum at m/z 375.1882 [M + Na] + (calcd for C 19 H 28 O 6 Na, 375.1886) and NMR data. Analysis of its NMR spectroscopic data ( Table 1 and Supplementary Figures 7, 8, 10) revealed the presence of six methyl groups (two methoxy groups), two methylenes, three oxymethines, one sp 3 quaternary carbon, two trisubstituted olefin units, one carboxylic carbon (δ C 170.0), one aldehyde group (δ C 194.0; δ H 9.37) and one ketone carbon (δ C 208.3). These data were similar to those of 1, indicating that compound 2 was also a humulane-type sesquiterpene. The main differences were that the methylene group at C-4 (δ C/H 39.4/2.85, 3.83) and the ketone carbon C-5 (δ C 203.3) in 1 were replaced by two oxymethines (δ C/H 79.2/4.71; 78.9/4.14) and two methoxy groups (δ C/H 57.2/3.26; 56.7/3.24) in 2. These assignments were further confirmed by 2D NMR data analysis. In the 1 H-1 H COSY spectrum (Figure 2 and Supplementary Figure 9), the cross-peaks of H-4 with H-5 and H-5 with H-6 indicated a C-4-C-5-C-6 subunit of 2. HMBC correlations (Figure 2 and Supplementary Figure 11) from H-4 to C-2, C-3, C-14 and C-15, from H-5 to C-3, C-7 and C-16, from H 3 -15 to C-4 and from H 3 -16 to C-5 linked C-4 to C-3 and located two methoxy groups at C-4 and C-5, respectively. Therefore, the planar structure of compound 2 was established as shown. The relative configuration of 2 was assigned by analysis of NOESY data. NOESY correlations (Figure 3 and Supplementary Figure 12) of H-2 with H-14 assigned the C-2/C-3 olefin as E-geometry. The C-6/C-7 olefin of 2 was also assigned as E-geometry based on NOESY correlations of H-6 with H-8 and H-5 with H 3 -17. NOESY correlation (Figure 3 and Supplementary Figure 12) of H-5 with H 3 -15 indicated the assignment of the α-configuration of H-5 and H 3 -15. While NOESY correlation (Figure 3 and Supplementary  Figure 12) of H-4 with H 3 -16 indicated that H-4 and H 3 -16 were in the β-orientation, thus establishing the relative configuration of 2. The absolute configuration of C-9 in 2 was deduced as 9S on the basis of biosynthetic considerations and by analogy to 1, 4, and 5 (Liao et al., 2013;Kapustina et al., 2020). The absolute configurations of C-4 and C-5 in 2 were also deduced by comparison of the experimental and calculated ECD spectra for the two stereoisomers, (4R, 5R, 9S)-2 (2a) and (4S, 5S, 9S)-2 (2b) (Figure 5). The MMFF94 conformational search and DFT re-optimization at the B3LYP/6-311G(2d,p) level yielded 3 lowest energy conformers for 2a (Supplementary Figure 19). The overall calculated ECD spectra of 2a and 2b were then generated by Boltzmann weighting of the conformers (Figure 5). The experimental CD curve of 2 matched well with the calculated ECD spectrum of 2a, suggesting that compound 2 has the absolute configuration of 4R, 5R, 9S.
Pestalothenin (3) Figures 13, 14, 16) of 1 revealed the presence of two methyls, five methylenes (two oxygenated), two methines, one sp 3 quaternary carbon, two trisubstituted olefin units, and one α,β-unsaturated ketone carbon (δ C 196.3). These data accounted for all 1 H and 13 C resonances except for two exchangeable protons, and suggested that 3 was a bicyclic compound. Analysis of the 1 H-1 H COSY spectrum (Figure 2 and Supplementary Figure 15) of 3 showed one isolated spin-system of C-3-C-2-C-1-C-9-C-10, as shown by the bold bonds in Figure 2. HMBC correlations (Figure 2 and Supplementary Figure 17) from H 2 -12 and H 3 -13 to C-1, C-10, and C-11, from H-1 and H 2 -10 to C-11, C-12 and C-13 allowed the construction of the cyclobutane ring. Further HMBC crosspeaks (Figure 2 and Supplementary Figure 17) from H 2 -3 to C-4, C-5, and C-14, from the olefinic proton H-5 to C-3, C-4 and C-14, and from H 3 -14 to C-3, C-4 and C-5 indicated that both C-3 and C-14 were directly connected to the C-4/C-5 olefin at C-4. Other correlations (Figure 2 and Supplementary  Figure 17) from the olefinic proton H-7 to C-8, C-9 and C-15, from H-9 to C-7, C-8 and C-15, and from H 2 -15 to C-7, C-8 and C-9 completed the C-7-C-8-C-9 subunit with C-15 attached to the C-7/C-8 olefin at C-8. Finally, HMBC correlations (Figure 2 and Supplementary Figure 17) from H-5 to C-7, from H-7 to C-5, and from H 2 -15 and H-3 to the α,β-unsaturated ketone carbon C-6 (δ C 196.3) revealed that C-6 was located between C-5 and C-7 to form the cyclononene ring, which was fused to the cyclobutane ring at C-1/C-9 to complete the bicyclo[7.2.0]undeca-2,5-dien-4-one core structure of 3. The two exchangeable protons were located at C-12 and C-15, respectively, by default, which partially supported by the chemical shift values for C-12 (δ C 70.6) and C-15 (δ C 63.5). Thus, the gross structure of 3 was established as a caryophyllene-type sesquiterpenoid (Figure 1). The relative configuration of 3 was established on the basis of the NOESY data (Figure 4). NOESY correlations (Figure 3 and Supplementary  Figure 18) of H-5 with H 3 -14, and of H-7 with H 2 -15 defined the Z-geometry and E-geometry for C-4/C-5 and C-7/C-8 olefins, respectively. NOESY correlations (Figure 3 and Supplementary  Figure 18) of H-1 with H-10, H 2 -12 and H 2 -15 indicated that these protons were on the same side of the ring system, whereas those of H-9 with H-10 and H 3 -13 placed these protons on the opposite side of the molecule, thus establishing the relative configuration of 3. To establish the absolute configuration of 3, ECD spectrum of 3 was recorded in MeOH and compared with the DFT-calculated spectra of two enantiomers 1R, 9S, 11R and 1S, 9R, 11S at the B3LYP/6-311 + G(2d,p) level. The MMFF94 conformational search and DFT re-optimization at the B3LYP/6-311G(2d,p) level yielded 3 lowest energy conformers for 3a (Supplementary Figure 19). The calculated ECD spectrum of 3a showed a good agreement with the experimental curve (Figure 6), which supported the absolute configuration being 1R, 9S, 11R. Thus, the completed structure of 3 was elucidated as depicted in Figure 1.

CONCLUSION
In summary, eight sesquiterpeniods including three new ones, pestalothenins A-C (1-3) were identified from the fermentation of the plant endophytic fungus P. theae (N635). The structures of the new compounds were elucidated via analyses of their MS, NMR, and ECD spectroscopic data. Pestalothenin A (1) differs from the known fungal metabolite 9,15-Dihydroxy-2,6humuladiene-5,10-dione (5) (Pulici et al., 1996a) by different configuration of C-2/C-3 and C-6/C-7 olefins and by having aldehyde and acetyl groups instead of hydroxymethyl and the hydroxy groups at C-3 and C-9, respectively. Pestalothenin B (2) is structurally related to pestalothenin A, but differs in having E-geometry for C-6/C-7 olefin and having two oxymethines and two methoxy groups rather than the methylene and ketone groups at C-4 and C-5, respectively. Biogenetically, the humulane-type sesquiterpeniods (1, 2, 4, and 5) could be derived from humulene which was formed from farnesyl pyrophosphate, first via oxidation, reduction and dehydration, and then followed by a series of methylation and acetylation reactions. While pestalothenin C (3) differs from the known humifusane A (Tian et al., 2011) by having oxymethylene group instead of methyl group at C-8. Biogenetically, humulene could be the biosynthetic intermediator of β-caryophyllene which acted as a key precursor in nature to form diverse tricyclic sesquiterpenes by transannular cyclizations. Starting from β-caryophyllene, caryophyllene-type sesquiterpeniods (3 and 6-8) could be generated via a series of reactions including transannular cyclization, oxidation, reduction, methylation and acetylation reactions. Compound 6 showed cytotoxic against T24 and MCF-7 cell lines. Our findings not only expand the chemical space of humulane-type and caryophyllene-type sesquiterpeniods, but also suggest that the fungal genus Pestalotiopsis might be a rich source of bioactive secondary metabolites.

DATA AVAILABILITY STATEMENT
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary Material.