Anti-Inflammatory Polyketide Derivatives from the Sponge-Derived Fungus Pestalotiopsis sp. SWMU-WZ04-2

Five undescribed polyketide derivatives, pestaloketides A–E (1–5), along with eleven known analogues (6–16), were isolated from the sponge-derived fungus Pestalotiopsis sp. Their structures, including absolute configurations, were elucidated by analyses of NMR spectroscopic HRESIMS data and electronic circular dichroism (ECD) calculations. Compounds 5, 6, 9, and 14 exhibited weak cytotoxicities against four human cancer cell lines, with IC50 values ranging from 22.1 to 100 μM. Pestaloketide A (1) is an unusual polyketide, featuring a rare 5/10/5-fused ring system. Pestaloketides A (1) and B (2) exhibited moderately inhibited LPS-induced NO production activity, with IC50 values of 23.6 and 14.5 μM, respectively, without cytotoxicity observed. Preliminary bioactivity evaluations and molecular docking analysis indicated that pestaloketides A (1) and B (2) had the potential to be developed into anti-inflammatory activity drug leads.


Introduction
Marine sponge-derived fungi have been proven to be a large and promising source of novel drug leads [1]. Among them, the pestalotiopsis species isolated from specific habitats are especially recognized as important producers of structurally varied, biologically active metabolites [2][3][4]. Since the discovery of taxol from the fungal Pestalotiopsis microspora [5], many novel secondary metabolites with potential pharmaceutical properties have been reported from this genus, including anti-inflammatory, cytotoxic, antiviral, antioxidant, and antimicrobial activities [6][7][8]. Accordingly, these findings have inspired many researchers to investigate the bioactive metabolites produced by Pestalotiopsis species.
The pestalotiopsis species, mainly distributed in both terrestrial and marine habitats, can produce many secondary metabolites. Polyketides possessing a rearranged or a modified different carbon nucleus have been reported from Pestalotiopsis species [9,10]. However, a novel tricyclic 5/10/5 skeleton has not been declared.
Compound 3 was detected as yellow oil, giving the molecular formula of C 12 H 16 O 5 from the analysis of their 13 C NMR data and HRESIMS (Table 2). Carefully, analysis of the 1D NMR data, in combination with the HSQC spectrum, revealed characteristic signals corresponding to one olefinic proton (δ H 5.83 (1H, s, H-5)) and three methyls (δ H 1.08, s, δ H 1.24, s, δ H 2.03, s). The 13 C NMR and HSQC spectroscopic data of 3 exhibited resonances for one lactone carbon (δ C 164.8 (C-6)) and three oxygenated carbons (δ C 65.9, 75.7, 76.4). Detailed analysis of these above data of 3 revealed that compound 3 was very similar to those of 12, 3-methyl-2-penten-5 [13], except for the presence of the lactonic ring groups at C-2 in 3. The aforementioned conclusion was supported again by the key correlations from H-7 to C-3/C-5/C-4, from H-8 to C-9/C-12/C-13, and from H-2 to C-3/C-4/C-9/C-6.
The key NOESY correlations of H-2 with H 3 -12 and H-3β suggested that these protons were cofacial; thus, the relative configuration of compound 3 was deduced as 2S, 8S. The absolute configuration of C-2 and C-8 in 3 were elucidated by comparing the calculated ECD spectrum of the 2S, 8S-model and the experimental ECD curve of 3 ( Figure 4). Thus, compound 3 was assigned ( Figure 1). Compound 4 was yielded as a yellow oil, gave the molecular formula of C 10 H 14 O 5 from the HRESIMS ion at m/z 237.0747 [M+Na] + and 13 C NMR data. Analysis of the 1D NMR spectroscopic data between 4 and 3 indicated both compounds to be structurally similar ( Table 2). The major difference was that the lactonic ring groups at C-2 in 3 were replaced by one acetate at C-2 in 4. The aforementioned results were supported by the key HMBC correlations from H-5 to C-6/C-3/C-7 (δ C 22.9), from H-2 to C-3/C-4/C-6, and from H-3 to C-7/ C-2/C-5/C-8, and the COSY correlations H-2 and H-3. According to the above evidence, by the biosynthetic pathway, similar chemical shift, and specific rotation (4, [α] 25 D -52 (c 0.2, MeOH, 3[α] 25 D -58 (c 0.20, MeOH)) data comparison, the relative configurations of 3 and 4 were concluded to be the same for C-2 and C-8. This assignment was proved by the ECD spectrum, the result of which showed good accordance with 3 ( Figure 4). Thus, the structure of 4 was assigned and named pestaloketide C.
Compounds 6-16 were determined to be the known 11-keto-9(E),12(E)-octadecadienoic acid (6) (Table 3). Anti-inflammatory activities were performed for compounds 1-4, 7-8, 10-13, and 15-16 with NO production inhibitory activity. Pestalolactones A (1) and B (2) showed moderate inhibitory of NO production with IC 50 values of 23.6 and 14.5 µM, respectively, without cytotoxicity observed. Others were inactive (100 µM). The result showed that pestalolactones A (1) and B (2) had the potential to be developed into anti-inflammatory activity drug leads (Table 4).  Interestingly, pestaloketide A (1) is reported to be the first tricyclic 5/10/5 skeleton polyketide from the sponge-derived fungus Pestalotiopsis sp. In addition, pestaloketides A (1) and B (2) exhibited moderately inhibited LPS-induced NO production activity. To further investigate the anti-inflammatory mechanism of pestaloketides A (1) and B (2), molecular docking of 1 and 2 with inducible NO oxidase (iNOS) as target was employed, and dexamethasone was used for redocking ( Figure 5). Docking results display that the docking pose of dexamethasone (Figure 5a, green) fit well with its original pose (Figure 5a, purple) in cocrystal, and compounds 1-2 exhibited good interactions with the INOS target in its pocket. Pestaloketide A (1) had a hydrogen bond interaction with Q257, and had nonpolar interactions with residues V346, Y367, and R382 and the cofactor heme. For pestaloketide B (2), hydrogen bond interactions were formed with residues Q257, Y341, N348 and Y367, and nonpolar interactions were formed with residues V346, R382, and W457 and the cofactor heme. These results indicate that 2 has a stronger association with the INOS protein than 1, which is consistent with our in vitro biological activity experiment results. Therefore, both pestalolactones A (1) and B (2) had the potential to be developed into anti-inflammatory activity drug leads.
Mar. Drugs 2022, 20, x FOR PEER REVIEW 7 of 11 nonpolar interactions with residues V346, Y367, and R382 and the cofactor heme. For pestaloketide B (2), hydrogen bond interactions were formed with residues Q257, Y341, N348 and Y367, and nonpolar interactions were formed with residues V346, R382, and W457 and the cofactor heme. These results indicate that 2 has a stronger association with the INOS protein than 1, which is consistent with our in vitro biological activity experiment results. Therefore, both pestalolactones A (1) and B (2) had the potential to be developed into anti-inflammatory activity drug leads.

Computational Section
The calculations were applied by the Spartan'14, Gaussian 09 software, and Merck Molecular Force Field (MMFF), respectively. The conformers of 1-4 were chosen at the B3LYP/6-311+G(d,p) level. The overall calculation of the ECD was performed using the TDDFT method for the stable conformers of new compounds. The spectra were obtained by SpecDis 1.6.

Cytotoxicity Assay
The method for the assay of cytotoxicity activity of 1-16 was conducted according to the one described previously [11]. Positive control (Adriamycin).

Inhibition of NO Production Assays
The activity of compounds 1-16 were examined by inhibited NO production in LPSstimulated RAW. The detailed process of the assay is described in the previously published paper [9]. Positive control (dexamethasone).

Molecular Docking
The three-dimensional structure of INOS (PDB ID:3E6T) was acquired from the Protein Data Bank (http://www.rcsb.org, accessed on 30 October 2022) [25,26], for which the resolution was 2.5 Å. Using the Chain A of the INOS structure as the receptor, pestaloketides A (1) and B (2) were docked using Autodock vina [27] and AutoDockTools-1.5.6 [28]. The geometrical restraints for 1 and 2 were generated by Grade Web Server (http://grade. globalphasing.org, accessed on 29 October 2022). A grid box of a 48.02 Å × 42.58 Å × 33.75 Å size was centered on the catalytic site. All docking parameters were set to default values. The docking results were further analyzed and presented using PyMOL (http://www.pymol.org, accessed on 29 October 2022) and LigPlot + [29].
Institutional Review Board Statement: Not applicable.

Data Availability Statement:
The data presented in this study are available in the main text and the supplementary materials of this article.

Conflicts of Interest:
The authors declare no conflict of interest.