Dactylospenes A–E, Sesterterpenes from the Marine Sponge Dactylospongia elegans

Chemical investigation on a marine sponge, Dactylospongia elegans, yielded five new γ-oxygenated butenolide sesterterpene derivatives, dactylospenes A–E (1–5), as well as two known biosynthetically related compounds, luffariellolide (6) and furospinosulin B (7). The structures of these compounds were elucidated on the basis of their spectroscopic data, experimental and calculated electronic circular dichroism (ECD) analysis, as well as comparison of the NMR data with those of known analogs. These metabolites are the first γ-oxygenated butenolide sesterterpenes to be reported from this genus. These compounds were evaluated in antimicrobial, anti-inflammatory, and cytotoxic assays. Only compounds 1, 3, and 6 exhibited moderate cytotoxicity against DU145, SW1990, Huh7, and PANC-1 cancer cell lines with IC50 values in the range of 2.11–13.35 μM. Furthermore, compound 2, without cytotoxicity, exhibited significant inhibitory effects (inhibitory rate 77.5%) on nitric oxide production induced by lipopolysaccharide at 10 μM.


Introduction
Sesterterpenoids, mainly found in marine organisms, are a rare class of terpene that are structurally diverse and have a wide spectrum of biological activities [1,2]. Marine sponges represent an abundant source of bioactive sesterterpenoids [3,4]. A literature survey revealed that only three sesterterpenoids and no γ-oxygenated butenolide sesterterpenes or related derivatives have been isolated from sponges of the genus Dactylospongia [5,6]. Aside from a few sesterterpenoids, sponges of the genus Dactylospongia, in particular D. elegans, have been widely investigated as a rich source of sesquiterpene quinones/quinols, sesquiterpene acids, sesterterpene lactones, macrolides, and steroids [7]. These metabolites showed a spectrum of bioactivities, such as cytotoxic [8], anti-inflammatory [9], antibacterial [10], and protein kinase inhibitory activities [11].
As part of the ongoing bioactive natural product discovery from the organic extract of a marine sponge D. elegans, one fraction showed different LC-DAD-MS profiles to those of previously reported As part of the ongoing bioactive natural product discovery from the organic extract of a marine sponge D. elegans, one fraction showed different LC-DAD-MS profiles to those of previously reported sesquiterpene quinones/hydroquinones based upon LC-MS analysis [6]. Further chemical investigation of this fraction led to the isolation of γ-oxygenated butenolide sesterterpene derivatives, dactylospenes A-E (1-5), and two known compounds (6)(7), shown in Figure 1. Herein, we report the details of the isolation, structure elucidation, and biological activity evaluation of these metabolites.

Results
Dactylospene A (1) was obtained as a light red oil, and its molecular formula was determined as C25H38O3 according to the HRESIMS peak at m/z 404.3166 [M + NH4] + , requiring seven degrees of unsaturation. The IR spectrum showed absorption bands for hydroxy (3342 cm −1 ) and ester carbonyl (1760 cm −1 ) groups. Comparison of the 1 H and 13 C NMR data of 1 (Table 1) with those of the known compound 7 showed that they shared a linear sesterterpene skeleton [12], which was further confirmed by the HMBC and COSY correlations shown in Figure 2. The observation of a downfield shift in the C-21, from δC 73.1 in 7 to δC 99.1 in 1, confirmed that the hydroxyl group connected at C-21. The relative configuration of the double bonds in 1 was inferred to be the same as compound 7 and further established by the NOESY experiments ( Figure 3). Strong NOESY correlations of overlapped proton signals at δH 2.00 and δH 5.11 suggested that these three groups of double-bonds Δ 6,7 , Δ 10,11 , and Δ 14,15 were all E-geometry. Furthermore, the absolute stereochemistry of C-21 can be determined by following the determination method of O,O-dimethyllingshuiolide A [13]. The characteristic positive Cotton effect at 253 nm in the CD spectrum of 1 was virtually identical to that of the simplified models (detailed in the supplementary information S1), (S)-5-hydroxy-4methylfuran-2(5H)-one (4S-8) ( Figure 4). Consequently, absolute configuration of 1 was unassigned as 21S. Dactylospene B (2) was also purified as a light red oil and exhibited a pseudomolecular ion [M + NH4] + peak in the positive HRESIMS at m/z 418.3308, consistent with the molecular formula C26H40O3, which was supported by the 1 H and 13 C NMR data. The molecular formula indicated seven degrees of unsaturation. The IR spectrum of 2 showed strong absorption at 3465 and 1766 cm −1 , assignable to hydroxy and ester carbonyl functionalities, respectively. The 13 C NMR and DEPT spectra ( Table 2) of 2 revealed 26 carbon signals, including one carbonyl carbon (δC 170.3), three sp 2 quaternary carbons (δC 168. 6, 144.4, and 130.9), three sp 2 methine carbons (δC 124. 6, 117.9, and 116.9),

Results
Dactylospene A (1) was obtained as a light red oil, and its molecular formula was determined as C 25 H 38 O 3 according to the HRESIMS peak at m/z 404.3166 [M + NH 4 ] + , requiring seven degrees of unsaturation. The IR spectrum showed absorption bands for hydroxy (3342 cm −1 ) and ester carbonyl (1760 cm −1 ) groups. Comparison of the 1 H and 13 C NMR data of 1 (Table 1) with those of the known compound 7 showed that they shared a linear sesterterpene skeleton [12], which was further confirmed by the HMBC and COSY correlations shown in Figure 2. The observation of a downfield shift in the C-21, from δ C 73.1 in 7 to δ C 99.1 in 1, confirmed that the hydroxyl group connected at C-21. The relative configuration of the double bonds in 1 was inferred to be the same as compound 7 and further established by the NOESY experiments ( Figure 3). Strong NOESY correlations of overlapped proton signals at δ H 2.00 and δ H 5.11 suggested that these three groups of double-bonds ∆ 6,7 , ∆ 10,11 , and ∆ 14,15 were all E-geometry. Furthermore, the absolute stereochemistry of C-21 can be determined by following the determination method of O,O-dimethyllingshuiolide A [13]. The characteristic positive Cotton effect at 253 nm in the CD spectrum of 1 was virtually identical to that of the simplified models (detailed in the Supplementary Information S1), (S)-5-hydroxy-4-methylfuran-2(5H)-one (4S-8) ( Figure 4). Consequently, absolute configuration of 1 was unassigned as 21S. Dactylospene B (2) was also purified as a light red oil and exhibited a pseudomolecular ion [M + NH 4 ] + peak in the positive HRESIMS at m/z 418.3308, consistent with the molecular formula C 26 H 40 O 3 , which was supported by the 1 H and 13 C NMR data. The molecular formula indicated seven degrees of unsaturation. The IR spectrum of 2 showed strong absorption at 3465 and 1766 cm −1 , assignable to hydroxy and ester carbonyl functionalities, respectively. The 13 C NMR and DEPT spectra ( Table 2) of 2 revealed 26 carbon signals, including one carbonyl carbon (δ C 170.3), three sp 2 quaternary carbons (δ C 168. 6, 144.4, and 130.9), three sp 2 methine carbons (δ C 124. 6, 117.9, and 116.9), one sp 3 oxymethine (δ C 103.9), one methoxy (δ C 56.1), two sp 3 quaternary carbons (δ C 42.0 and 33.2), two methine carbons (δ C 43.9 and 41.8), eight methylene carbons (δ C 38. 6, 30.8, 29.4, 28.7, 27.5, 22.4, Mar. Drugs 2020, 18, 491 3 of 11 21.9, and 21.7), and five methyl carbons (δ C 25. 2, 23.5, 22.4, 17.2, and 15.9), which accounted for four degrees of unsaturation. The remaining three degrees of unsaturation were caused by the presence of three rings in the molecule.  6, 30.8, 29.4, 28.7, 27.5, 22.4, 21.9, and 21.7), and five methyl carbons (δC 25. 2, 23.5, 22.4, 17.2, and 15.9), which accounted for four degrees of unsaturation. The remaining three degrees of unsaturation were caused by the presence of three rings in the molecule.   and methyl group, whereas the NOESY correlations of H-8/H3-19 indicated that these protons were oriented in the other direction. Therefore, we established the relative stereochemistry of 2 as 4R*,5S*,8R*,9R*. The molecular formula for dactylospene C (3) was also deduced as C26H40O3 by HRESIMS. Analysis of the 1D and 2D NMR data ( Table 2) for 3 and 2 showed that they shared the same planar structure. The similar NOESY correlations of H-5/H-20a, H-5/H3-18, and H-8/H3-19 revealed the same relative configurations of 3 as those of 2 at C-4, C-5, C-8, and C-9. However, the different but not mirrored CD spectra ( Figure   To establish the absolute configuration of 2 and 3, we compared the ECD spectra of 2 and 3 with calculated ECD spectra of simplified models, 4R-8 and 4R,5S,8R,9R-9 ( Figure 4). The negative cotton effect around 250 nm in 2 clearly indicated that the absolute configuration at C-16 is R. Meanwhile, the positive cotton effect around 250 nm in 3 allowed us to conclude that the absolute configuration at C-16 in 3 is S. Then, we determined the absolute configuration of the decaline substructure. As described above, 2 and 3 are a pair of diastereomers, and the only difference is the absolute configuration at C-16. Therefore, 2 and 3 have the same absolute configuration in the decaline moiety.  The molecular formula for dactylospene C (3) was also deduced as C26H40O3 by HRESIMS. Analysis of the 1D and 2D NMR data ( Table 2) for 3 and 2 showed that they shared the same planar structure. The similar NOESY correlations of H-5/H-20a, H-5/H3-18, and H-8/H3-19 revealed the same relative configurations of 3 as those of 2 at C-4, C-5, C-8, and C-9. However, the different but not mirrored CD spectra (  To establish the absolute configuration of 2 and 3, we compared the ECD spectra of 2 and 3 with calculated ECD spectra of simplified models, 4R-8 and 4R,5S,8R,9R-9 ( Figure 4). The negative cotton effect around 250 nm in 2 clearly indicated that the absolute configuration at C-16 is R. Meanwhile, the positive cotton effect around 250 nm in 3 allowed us to conclude that the absolute configuration at C-16 in 3 is S. Then, we determined the absolute configuration of the decaline substructure. As described above, 2 and 3 are a pair of diastereomers, and the only difference is the absolute configuration at C-16. Therefore, 2 and 3 have the same absolute configuration in the decaline moiety. Interpretation of the 2D NMR data, including COSY, HMQC, and HMBC spectra, led to the construction of the planar structure of 2. The COSY spectrum of 2 suggested the presence of four individual spin systems: , and C-20−C-21−C-22 (d), which were accomplished with the assistance of the HMBC experiment ( Figure 1). HMBC correlations from H 3 -24 to C-22, C-23, and C-25, from H 3 -25 to C-22, C-23, and C-24, and from H 2 -22 to C-24 and C-25 determined the existence of dimethylallyl moiety (C20-C25). Moreover, HMBC correlations from H 2 -3 in fragment a, and H-5 and H 2 -6 in fragment b to the same carbon C-4, as well as HMBC correlations from H 3 -17 to C-3, C-4, and C-5, allowed the linkage of fragments a and b via C-4 and the assignment of the methyl group CH 3 -17 at C-4. The dimethylhomoallyl group was attached to C-4 based on the HMBC correlations from H 2 -20 to C-3, C-4, and C-5. Further HMBC correlations from H 3 -18 to C-9, and from H 3 -19 to C-8, C-9, and C-10, revealed the connectivity of C-8 and C-10 through C-9 and placed the methyl group H 3 -19 at C-9 as well. Another two groups of HMBC cross-peaks from H-1 to C-5 and C-9, from H 2 -2, H-5 and H-6b to C-10, supported the linkage of C-1 and C-5 via C-10. Moreover, a suite of resonances at δ C 170.3 (C-15), 168.6 (C-13), 116.9 (CH-14), 103.9 (CH-16), and 56.1 , could be assigned to an α,β-unsaturated-γ-methoxy-γ-lactone moiety, which was further supported by HMBC correlations from H-16 to C-13, C-14, and C-15, from H-14 to C-13, C-15, and C-16, and from 16-OCH 3 to C-16. This moiety was further linked to C-9 Mar. Drugs 2020, 18, 491 5 of 11 through fragment C-11-C-13, based on the HMBC correlations from H-14 to C-12 and from H-11 to C-9, C-10, and C-13. Thus, the planar structure of 2 was determined as depicted. A NOESY experiment was also performed to determine the relative configuration of 2. The NOESY cross-peaks of H-5/H-20a and H-5/H 3 -18 indicated the cofacial orientation of these protons and methyl group, whereas the NOESY correlations of H-8/H 3 -19 indicated that these protons were oriented in the other direction. Therefore, we established the relative stereochemistry of 2 as 4R*,5S*,8R*,9R*.
The molecular formula for dactylospene C (3) was also deduced as C 26 H 40 O 3 by HRESIMS. Analysis of the 1D and 2D NMR data ( Table 2) for 3 and 2 showed that they shared the same planar structure. The similar NOESY correlations of H-5/H-20a, H-5/H 3 -18, and H-8/H 3 -19 revealed the same relative configurations of 3 as those of 2 at C-4, C-5, C-8, and C-9. However, the different but not mirrored CD spectra ( Figure 5) and specific rotation values ([α] 25 D 13.9 for 2, [α] 25 D 56.2 for 3) of 2 and 3 indicated that these two compounds are a pair of diastereomers.
To establish the absolute configuration of 2 and 3, we compared the ECD spectra of 2 and 3 with calculated ECD spectra of simplified models, 4R-8 and 4R,5S,8R,9R-9 ( Figure 4). The negative cotton effect around 250 nm in 2 clearly indicated that the absolute configuration at C-16 is R. Meanwhile, the positive cotton effect around 250 nm in 3 allowed us to conclude that the absolute configuration at C-16 in 3 is S. Then, we determined the absolute configuration of the decaline substructure. As described above, 2 and 3 are a pair of diastereomers, and the only difference is the absolute configuration at C-16. Therefore, 2 and 3 have the same absolute configuration in the decaline moiety. According to the calculated ECD spectra of 9, if the compound has 4R,5S,8R,9R configuration, it shows a positive cotton effect around 205 nm. In the experimental ECD spectra, the positive cotton effect around 205 nm in 3 is more emphasized than the negative cotton effect around 205 nm in 2. This indicated that the decaline moiety in 2 and 3 has positive cotton effect around 205 nm. Therefore, we clarified the absolute configuration at C-4, C-5, C-8, and C-9 in 2 and 3 to be 4R,5S,8R,9R.  Dactylospene E (5) was also obtained as a light yellow oil. The molecular formula of C27H44O4 was deduced from its HRESIMS data (m/z 450.3578 [M + NH4] + ). Compound 5 showed nearly the same chemical shifts as those of compound 4. Correlations from the 2D NMR spectra confirmed the same planar structure between 5 and 4 and the same configuration of epi-halimane core in 5 and 4. A comparison with CD spectra and specific rotation values obtained for 5 and 3 unambiguously assigned absolute configuration as 4R,5S,8R,9R,16R.
In addition to the five new compounds 1-5, two known compounds, identified as luffariellolide (6) and furospinosulin B (7), were also isolated as metabolites of D. elegans. These compounds were identified by comparing their spectral data with the spectroscopic data reported in the corresponding literature [12,14]. Moreover, the C-25 stereocenter in the γ-oxygenated butenolide unit could be assigned as an S configuration by comparison the specific rotation data of 6 ([α] 25 D −35.3, MeCN, c 1.0) with that of 1. Compounds 4 and 5 could possibly be formed by reaction with MeOH from compounds 2 and 3 during the isolation. However, when compounds 2 and 3 were stirred with silica and ODS in MeOH for 48 h, neither 4 nor 5 were detected by HPLC-UV analysis.
All the isolated compounds were tested for antimicrobial activity against two strains of hospitalacquired, methicillin-resistant Staphylococcus aureus (MRSA H0556 and MRSAH0117) and cytotoxic activity against DU145, SW1990, Huh7, and PANC-1 cancer cell lines. Unfortunately, the compounds tested exhibited no activity against the above strains. Only compounds 1, 3, and 6 exhibited moderate cytotoxicity against the above four cancer cell lines, with IC50 values in the range of 2.11-13.35 μM, while the other isolates were inactive (IC50 values > 50 μM) (Table 4). Moreover, compounds 1-5 were subjected to an evaluation of their anti-inflammatory activity. Compound 2 exhibited significantly greater inhibitory effects than 3 (inhibitory rate 77.5% for 2 and 30.4% for 3) on nitric oxide (NO) production induced by lipopolysaccharide (LPS) treatment of RAW 264.7 cells at 10 μM. In addition, the proliferation rate of RAW 264.7 cells was 151.2% with the treatment of 2 at 10 μM, which indicated that the anti-inflammatory effect of 2 was not achieved by its cytotoxicity. The biological evaluation indicated that R-γ-methoxy butenolide moiety positively affected the activity. Dactylospene D (4) was isolated as a yellow oil and assigned the molecular formula of C 27 H 44 O 4 , based on HRESIMS data for the [M + NH 4 ] + ion at m/z 450.3581, which is consistent with six degrees of unsaturation. The 1 H and 13 C NMR signal patterns (Table 3) suggested the same structure core as compounds 2 and 3 with a α,β-unsaturated-γ-methoxy-γ-lactone moiety linked to the epi-halimane-type diterpenoid. Key HMBC and COSY correlations shown in Figure 2 confirmed this proposed structure. Through analysis of the remaining signals in the 13 C NMR, two methyls (δ C 24.9 and 25.0), three sp 3 methylenes (δ C 41.1, 39.8, and 17.8), one methoxy (δ C 49.1), and one sp 3 oxygenated quaternary carbon (δ C 74.6) were deduced. The attachment at C-4 was a 2-methoxy-2-methylpentane unit, instead of the dimethylhomoallyl moiety in 2 and 3, supported by the COSY correlations of H-20a/H-21a and H-21b/H 2 -22, in addition to the HMBC correlations from H-20a to C-3, from H 3 -24 and H 3 -25 to C-22 and C-23, and from 23-OCH 3 to C-23. Moreover, NOESY correlations of H-5/H-20a, H-5/H 3 -18, and H-8/H 3 -19 revealed the relative configurations of 4, which were the same as compounds 2 and 3. Finally, the similarity of their CD spectra between 2 and 4 suggested that compound 4 shared the same absolute configurations as those of 2 ( Figure 5). Dactylospene E (5) was also obtained as a light yellow oil. The molecular formula of C 27 H 44 O 4 was deduced from its HRESIMS data (m/z 450.3578 [M + NH 4 ] + ). Compound 5 showed nearly the same chemical shifts as those of compound 4. Correlations from the 2D NMR spectra confirmed the same planar structure between 5 and 4 and the same configuration of epi-halimane core in 5 and 4. A comparison with CD spectra and specific rotation values obtained for 5 and 3 unambiguously assigned absolute configuration as 4R,5S,8R,9R,16R.
In addition to the five new compounds 1-5, two known compounds, identified as luffariellolide (6) and furospinosulin B (7), were also isolated as metabolites of D. elegans. These compounds were identified by comparing their spectral data with the spectroscopic data reported in the corresponding literature [12,14]. Moreover, the C-25 stereocenter in the γ-oxygenated butenolide unit could be assigned as an S configuration by comparison the specific rotation data of 6 ([α] 25 D − 35.3, MeCN, c 1.0) with that of 1. Compounds 4 and 5 could possibly be formed by reaction with MeOH from compounds 2 and 3 during the isolation. However, when compounds 2 and 3 were stirred with silica and ODS in MeOH for 48 h, neither 4 nor 5 were detected by HPLC-UV analysis. All the isolated compounds were tested for antimicrobial activity against two strains of hospital-acquired, methicillin-resistant Staphylococcus aureus (MRSA H0556 and MRSAH0117) and cytotoxic activity against DU145, SW1990, Huh7, and PANC-1 cancer cell lines. Unfortunately, the compounds tested exhibited no activity against the above strains. Only compounds 1, 3, and 6 exhibited moderate cytotoxicity against the above four cancer cell lines, with IC 50 values in the range of 2. 11-13.35 µM, while the other isolates were inactive (IC 50 values > 50 µM) (Table 4). Moreover, compounds 1-5 were subjected to an evaluation of their anti-inflammatory activity. Compound 2 exhibited significantly greater inhibitory effects than 3 (inhibitory rate 77.5% for 2 and 30.4% for 3) on nitric oxide (NO) production induced by lipopolysaccharide (LPS) treatment of RAW 264.7 cells at 10 µM. In addition, the proliferation rate of RAW 264.7 cells was 151.2% with the treatment of 2 at 10 µM, which indicated that the anti-inflammatory effect of 2 was not achieved by its cytotoxicity. The biological evaluation indicated that R-γ-methoxy butenolide moiety positively affected the activity.