New Ophiobolin Derivatives from the Marine Fungus Aspergillus flocculosus and Their Cytotoxicities against Cancer Cells

Five new sesterterpenes, 14,15-dehydro-6-epi-ophiobolin K (1), 14,15-dehydro- ophiobolin K (2), 14,15-dehydro-6-epi-ophiobolin G (3), 14,15-dehydro-ophiobolin G (4) and 14,15-dehydro-(Z)-14-ophiobolin G (5), together with four known ophiobolins (6–9) were isolated from the marine fungus Aspergillus flocculosus derived from the seaweed Padina sp. collected in Vietnam. The five new ophiobolins were first isolated as ophiobolin derivatives consisting of a fully unsaturated side chain. Their structures were elucidated via spectroscopic methods including 1D, 2D NMR and HR-ESIMS. The absolute configurations were determined by the comparison of chemical shifts and optical rotation values with those of known ophiobolins. All compounds (1–9) were then evaluated for their cytotoxicity against six cancer cell lines, HCT-15, NUGC-3, NCI-H23, ACHN, PC-3 and MDA-MB-231. All the compounds showed potent cytotoxicity with GI50 values ranging from 0.14 to 2.01 μM.


Introduction
The marine environment is an enormous reservoir of novel sources of biologically active metabolites, many of which display unique structural skeletons that can be used as lead structures for the development of new drugs [1,2]. To adapt and live in an environment that is significantly different from terrestrial organisms, marine organisms frequently produce structurally unique chemical compounds [3,4]. Specifically, secondary metabolites from marine microorganisms are recognized as a novel chemical source for drug discovery and development. Among marine-derived microbes, marine fungi produce a wide range of promising biologically active compounds [5]. Numerous novel compounds from marine fungi have displayed a wide range of bioactivities such as antiviral, antibacterial, anticancer, antiplasmodial and anti-inflammatory [6][7][8]. In the marine context, Aspergillus [9,10]. The genus Aspergillus is known as a major contributor of pharmacologically bioactive compounds, including anticancer asperazine, antibacterial varixanthone and antifungal amphotericin B [11,12].
Ophiobolins are a group of sesterpenoids with an unusual tricyclic 5-8-5 ring system. They show a broad range of inhibitory activities against nematodes, fungi, bacteria and cytotoxic activity against cancer cells [13,14]. They are produced by the fungal genus Bipolaris, Aspergillus, Sarocladium and Drechslera [15]. The first ophiobolin, ophiobolin A, was isolated from Biolaris spp. and displays inhibitory activity against calmodulin-activated cyclic nucleotide phosphodiesterase [16]. These findings made the compound a useful calmodulin probe for research purposes and implied an application in anti-cancer therapy [17]. Interestingly, more than half of the 49 ophiobolins identified between 1999 and 2016 exhibit cytotoxic activities against human cancer cell lines [18]. Although their biological properties have been well exploited in recent years, their structure-activity relationship remains unestablished [18]. Consequently, this study focused on the discovery of bioactive natural products from marine fungi. During our ongoing investigation for new bioactive compounds from marine microorganisms, the fungal 168ST-16.1 strain was isolated from the seaweed Padina sp. collected at Da Nang, Vietnam, and, based on its 28S rRNA gene sequence, it was identified as Aspergillus flocculosus. Subsequent chemical investigations on an EtOAc extract of the fungal culture broth using reversed-phase HPLC led to the isolation of the five new ophiobolins, named, 14,15dehydro-6-epi-ophiobolins K and G (1 and 3), 14,15-dehydro-ophiobolins K and G (2 and 4) and 14,15dehydro-(Z)-14-ophiobolin G (5), together with four known ophiobolins, 6-epi-ophiobolins C and N (6 and 8) and ophiobolins C and N (7 and 9) [19][20][21] (Figure 1). Herein, details of the structure elucidation and biological activity of these compounds are described.

Results and Discussion
Compound 1 was obtained as an amorphous powder. The molecular formula of 1 was determined to be C25H34O3 based on HRESIMS. The 1 H NMR spectroscopic data of 1 displayed resonances for an aldehyde proton (δH 9.23), four olefinic protons (δH 6.84, 6.42, 6 Table 2). Spin systems and their partial structures were confirmed and assembled by combined analysis of COSY and HMBC correlations ( Figure 2). Three spin systems, H2-1/H-2/H-6, H-
Compound 2 had the same molecular formula C 25 H 32 O 3 as 1. Its 1 H and 13 C NMR data were similar to those of 1, differing only by slightly shifted proton and carbon signals. It has been reported that H-2 of the 6-epi isomer having H-6α (A/B-trans) is upfield-shifted in comparison with the A/B-cis ophiobolin [13] (Figure 3b). The H-2 proton (δ H 2.57) of 2 is downfield-shifted than that (δ H 2.19) of 1, indicating that 2 has an A/B-cis ring structure. This study also revealed that the chemical shifts of the geminal proton H 2 -4 are closer to each other when the A/B ring junction is cis than when it is trans (Figure 3b). The key NOE correlations of H-2/H-6, H-2/H 3 -20 and H-2/H 3 -22 suggested that 2 has an A/B-cis ring structure and is a stereoisomer of 1 (Figure 3a). Based on these results, the structure of 2 was determined and named 14,15-dehydro-ophiobolin K. Compound 3 was obtained as an amorphous powder with the molecular formula of C25H32O2 based on HRESIMS. The molecular formula of 3 has one less CH2 and one less oxygen compared to that of 1. The 1 H and 13 C NMR data of 3 were quite similar to those of 1, displaying one additional singlet olefin signal (δH 6.01) and a sp 2 quaternary carbon at C-3 (δC 179.9), while lacking a methylene and sp3 quaternary carbon signal. The HMBC correlations from H3-20 (δH 2.12) to C-2 (δC 49.5), C-3 (δC 179.9) and C-4 (δC 129.4) revealed that a double bond existed between C-3 and C-4 by the dehydroxylation of the tertiary alcohol at C-3 in 1 (Figure 2). NOESY correlations from H-6/H-10 and H-2/H3-20/H3-22, the lack of NOESY correlation of H-2/H-6 and, comparison of the NMR spectral data and spectral properties of 3 with those of 6-epi-ophiobolin N (8), suggested that 3 has the same ring system as the A/B-trans ophiobolin [13,19] (Figure 4). On the basis of detailed data analysis, the structure of 3 was elucidated and named 14,15-dehydro-6-epi-ophiobolin G.
Compound 4 was isolated as an amorphous powder with the molecular formula of C25H32O2 as determined by HRESIMS. Its 1 H NMR data were similar to those of 3, differing only by slightly shifted signals. In contrast to the data for 3, the NOESY correlation of H-2/H-6 indicated that 4 has an A/B-cis ring structure (Figure 4). Thus, compound 4 was identified as a stereoisomer of 3 and named 14,15-dehydro-ophiobolin G.
Compound 5 was isolated as an amorphous powder with the same molecular formula C25H32O2 as compound 4, as determined by HRESIMS. The 1 H NMR data of 5 and 4 were nearly identical except for the H-16 proton which was slightly downfield-shifted than that of 4. By comprehensive analysis of its 1D and 2D NMR data, the planar structure of 5 was elucidated to be the same as that of 4, differing only in the orientation of H3-23. The NOESY correlations of H-16/H-10 and H-23/H2-13 suggested the relative configuration of Δ 14,15 in 5 was Z conformation, which is different from that of compound 4 (Figure 4). Therefore, the structure of 5 was elucidated to be as shown in Figure 1, and named 14, 15-dehydro-(Z)-14-ophiobolin G. Compound 3 was obtained as an amorphous powder with the molecular formula of C 25 H 32 O 2 based on HRESIMS. The molecular formula of 3 has one less CH 2 and one less oxygen compared to that of 1. The 1 H and 13 C NMR data of 3 were quite similar to those of 1, displaying one additional singlet olefin signal (δ H 6.01) and a sp 2 quaternary carbon at C-3 (δ C 179.9), while lacking a methylene and sp3 quaternary carbon signal. The HMBC correlations from H 3 -20 (δ H 2.12) to C-2 (δ C 49.5), C-3 (δ C 179.9) and C-4 (δ C 129.4) revealed that a double bond existed between C-3 and C-4 by the dehydroxylation of the tertiary alcohol at C-3 in 1 (Figure 2). NOESY correlations from H-6/H-10 and H-2/H 3 -20/H 3 -22, the lack of NOESY correlation of H-2/H-6 and, comparison of the NMR spectral data and spectral properties of 3 with those of 6-epi-ophiobolin N (8), suggested that 3 has the same ring system as the A/B-trans ophiobolin [13,19] (Figure 4). On the basis of detailed data analysis, the structure of 3 was elucidated and named 14,15-dehydro-6-epi-ophiobolin G.
Compound 4 was isolated as an amorphous powder with the molecular formula of C 25 H 32 O 2 as determined by HRESIMS. Its 1 H NMR data were similar to those of 3, differing only by slightly shifted signals. In contrast to the data for 3, the NOESY correlation of H-2/H-6 indicated that 4 has an A/B-cis ring structure (Figure 4). Thus, compound 4 was identified as a stereoisomer of 3 and named 14,15-dehydro-ophiobolin G.
Compound 5 was isolated as an amorphous powder with the same molecular formula C 25 H 32 O 2 as compound 4, as determined by HRESIMS. The 1 H NMR data of 5 and 4 were nearly identical except for the H-16 proton which was slightly downfield-shifted than that of 4. By comprehensive analysis of its 1D and 2D NMR data, the planar structure of 5 was elucidated to be the same as that of 4, differing only in the orientation of H 3 -23. The NOESY correlations of H-16/H-10 and H-23/H 2 -13 suggested the relative configuration of ∆ 14,15 in 5 was Z conformation, which is different from that of compound 4 ( Figure 4). Therefore, the structure of 5 was elucidated to be as shown in Figure 1, and named 14, 15-dehydro-(Z)-14-ophiobolin G. The structures of the four known compounds were determined as 6-epi-ophiobolin C (6), ophiobolin C (7), 6-epi-ophiobolin N (8) and ophiobolin N (9) by comparing of their 1 H, 13  The cytotoxicity of all the isolated compounds (1-9) against cancer cell lines, such as HCT-15, NUGC-3, NCI-H23, ACHN, PC-3 and MDA-MB-231, was investigated using the sulforhodamine B (SRB) assay, with adriamycin as a positive control. The results showed that all compounds were strongly active against 6 cancer cell lines with GI50 values in the range of 0.14 to 2.01 μM (Table 3). Compound 1 displayed the strongest cytotoxicity against the HCT-15, NUGC-3 and MDA-MB-231 cell lines with GI50 values of 0.21, 0.19 and 0.14 μM, respectively. Based on the cytotoxicity results, the analogs with one double bond (6)(7)(8)(9) in the side chain seemed to be slightly more active than those with three double bonds (1)(2)(3)(4)(5). 5 was least active against all cell lines, even with GI50 values ranging from 1.53 to2.01 μM, indicating that the geometry of C-14/C-15 might appear to have a slight influence on their activities. In addition, results for all the strongly active compounds indicated that the stereochemistry of C-6 and the hydroxyl group at C-3 might not noticeably affect the cytotoxicity.  The structures of the four known compounds were determined as 6-epi-ophiobolin C (6), ophiobolin C (7), 6-epi-ophiobolin N (8) and ophiobolin N (9) by comparing of their 1 H, 13  The cytotoxicity of all the isolated compounds (1-9) against cancer cell lines, such as HCT-15, NUGC-3, NCI-H23, ACHN, PC-3 and MDA-MB-231, was investigated using the sulforhodamine B (SRB) assay, with adriamycin as a positive control. The results showed that all compounds were strongly active against 6 cancer cell lines with GI 50 values in the range of 0.14 to 2.01 µM (Table 3). Compound 1 displayed the strongest cytotoxicity against the HCT-15, NUGC-3 and MDA-MB-231 cell lines with GI 50 values of 0.21, 0.19 and 0.14 µM, respectively. Based on the cytotoxicity results, the analogs with one double bond (6)(7)(8)(9) in the side chain seemed to be slightly more active than those with three double bonds (1)(2)(3)(4)(5). 5 was least active against all cell lines, even with GI 50 values ranging from 1.53 to2.01 µM, indicating that the geometry of C-14/C-15 might appear to have a slight influence on their activities. In addition, results for all the strongly active compounds indicated that the stereochemistry of C-6 and the hydroxyl group at C-3 might not noticeably affect the cytotoxicity.

Fungal Material and Fermentation
The fungus 168ST-16.1 was isolated from the algae Padina sp., collected at a depth of 10 m in Son Tra peninsular, Da Nang, Vietnam (16 • 09 97.8" N, 108 • 29 96.1" E), in August 2016. The fungal strain was identified as Aspergillus flocculosus (GenBank accession number MG920345) by DNA amplification and ITS region sequencing and named Aspergillus flocculosus 168ST-16.1.