Anti-Inflammatory Lobane and Prenyleudesmane Diterpenoids from the Soft Coral Lobophytum varium

New lobane-based diterpenoids lobovarols A–D (1–4) and a prenyleudesmane-type diterpenoid lobovarol E (5) along with seven known related diterpenoids (6–12) were isolated from the ethyl acetate extract of a Taiwanese soft coral Lobophytum varium. Their structures were identified on the basis of multiple spectroscopic analyses and spectral comparison. The absolute configuration at C-16 of the known compound 11 is reported herein for the first time. The anti-inflammatory activities of compounds 1–12 were assessed by measuring their inhibitory effect on N-formyl-methionyl-leucyl-phenyl-alanine/cytochalasin B (fMLP/CB)-induced superoxide anion generation and elastase release in human neutrophils. Metabolites 2, 5, and 11 were found to show moderate inhibitory activity on the generation of superoxide anion, while compounds 5, 8, 11, and 12 could effectively suppress elastase release in fMLP/CB-stimulated human neutrophil cells at 10 μM. All of the isolated diterpenoids did not exhibit cytotoxic activity (IC50 > 50 μM) towards a limited panel of cancer cell lines.

Lobovarol B (2) was also isolated as a colorless oil with a hydroxy group (IR νmax 3445 cm −1 ). The NMR data (Tables 1 and 2) showed the characteristic signals of lobane-type diterpenoids as in 1. Its HRESIMS m/z 373.2350 [M + Na] + and NMR data deduced a molecular formula C21H34O4 with a 14 mass unit difference from compound 1. Comparison of NMR data of compounds 2 and 1 revealed that compound 2 is the methyl ether of 1 due to the appearance of the methoxy signals (δC 55.6, CH3; δH 3.47, 3H, s). The HMBC correlation observed from the methoxy protons to the dioxymethine carbon (δC 97.7, CH, C-14) designated the C-14 position of the methoxyl. Therefore, compound 2 was identified as the methyl acetal arising from methylation of 14-OH of 1. The structure of 2 was further The relative configuration at the seven chiral centers of 1 was determined by the analysis of nuclear Overhauser effect (NOE) correlations along with molecular modeling using MM2 force field calculations ( Figure 4). The nuclear Overhauser effect spectroscopy (NOESY) spectrum of compound 1 was remeasured in C 6 D 6 for better resolution since the proton signals of H 2 -3 and H-4 in CDCl 3 were overlapped (3H, 1.56, m). In C 6 D 6 , NOE interactions were observed for H- Furthermore, the similar δ C values of C-1, C-2, C-7, and C-8 to C-12 of the previously-reported β-elemene and lobane-type diterpenoids [12,[32][33][34], isolated from the same genus Lobophytum or prepared by enantiocontrolled synthesis [37], suggested the 1R,2R,4S-configuration in compound 1. Moreover, the R configuration established for C-16 in the related lobane diterpenoids 11 (latter discussed) also implied the absolute configuration of chiral centers of the prenyleudesmane 5 and hence the lobanes 1-4. The NOE correlations observed for both H-3α (δ H 1.50, m) and H-5α (δ H 1.22, m) with H-14 (δ H 5.22, s), and for H-4 with H-15 (δ H 3.03, dd, J = 2.0, 1.6 Hz) indicated that the protons at C-14 and C-15 of the pyran ring should be syn to each other and were assigned arbitrarily as α-oriented. In turn, H-15 exhibited NOE interactions with both H 2 -16 protons (δ H 1.58, ddd, J = 13.2, 11.2, 2.0 Hz, H-16α and 1.73, m, H-16β) while H-17, which has an axial-axial coupling with H-16α (J = 11.2 Hz), displayed a significant NOE correlation with H-16β. Therefore, H-17 should be β-configured. This was also suggested by the absence of NOE response of H-17 with H-14. The above-mentioned NOEs found for H-14 with H-3α and H-5α, and for H-15 with H-4 revealed that the pyran ring should be perpendicular to the β-elemene ring system. To further prove the β-position of the epoxide ring, a conformation analysis using Chem3D, molecular mechanics calculations (MM2) and dihedral driving calculation were carried out [38,39]. The most stable (the lowest-energy) conformations for compound 1 and its 13,15-epimer 1a which possesses an α-epoxide are represented in Figures 4 and 5, respectively. In this perspective, we focused on the calculated distances between the diagnostic proton pairs having key NOE correlations in conformer 1, which were found shorter than 3.0 Å, in comparison with those calculated for 1a ( Table 3). The results demonstrated that the β-configuration of the epoxide ring could only fulfill all described NOE correlations mentioned above. On the basis of the above findings, the (1R, 2R, 4S, 13R, 14R, 15S, 17R)-configuration of 1 was, thus, established.
Lobovarol B (2) was also isolated as a colorless oil with a hydroxy group (IR ν max 3445 cm −1 ). The NMR data (Tables 1 and 2) showed the characteristic signals of lobane-type diterpenoids as in 1. Its HRESIMS m/z 373.2350 [M + Na] + and NMR data deduced a molecular formula C 21 H 34 O 4 with a 14 mass unit difference from compound 1. Comparison of NMR data of compounds 2 and 1 revealed that compound 2 is the methyl ether of 1 due to the appearance of the methoxy signals (δ C 55.6, CH 3 ; δ H 3.47, 3H, s). The HMBC correlation observed from the methoxy protons to the dioxymethine carbon (δ C 97.7, CH, C-14) designated the C-14 position of the methoxyl. Therefore, compound 2 was identified as the methyl acetal arising from methylation of 14-OH of 1. The structure of 2 was further confirmed by the analysis of COSY and HMBC correlations ( Figure 3). Moreover, compound 2 displayed analogous NOE correlations and possessed the same sign of optical rotation ([α] 25 D −34.7) as those of 1, implying the same absolute configuration for both 1 and 2.          which has one oxygen atom less than that of 2. The IR absorption band at ν max of 3450 cm −1 again indicated the presence of a hydroxy functionality in the molecule. Again, careful inspection of the 1 H and 13 C NMR spectroscopic data (Tables 1 and 2) of 3 showed resonances and coupling constants identical to those of the β-elemene ring system, as verified in compounds 1 and 2 and other known lobane-type diterpenoids. Comparison of the 21 carbon signals of 3 with those of 2 showed that the trisubstituted epoxy signals in 2 was replaced by those of a trisubstituted double bond (δ C/ δ H 140.6, C; 121.1, CH/5.73, br d, J = 6.0 Hz) in 3. The planar structure of 3 was further established by analyzing its COSY and HMBC correlations ( Lobovarol E (5) was obtained as a white powder. The molecular formula was deduced to be C 20 H 32 O 2 as indicated by the HRESIMS (m/z 327.2292 [M + Na] + ) and NMR data (Tables 1 and 2), implying five degrees of unsaturation. Its IR absorption band at 3422 cm −1 revealed the presence of a hydroxy functionality, which was further supported by the NMR signals at δ C 67.9 and δ H 4.25. The NMR data (Tables 1 and 2) showed the presence of one 1,1-disubstitued (δ C 150.9, C and 105.4 CH 2 ; δ H 4.72 and 4.43, each 1H, s) and a trisubstituted (δ C 146.1, C and 120.8, CH; δ H 5.33, 1H, d, J = 8.5 Hz) olefinic bonds, a trisubstituted epoxide (δ C 59.8, C; 67.5, CH; δ H 2.82, 1H, d, J = 8.0 Hz), and a hydroxyl-bearing methine (δ C 67.9, CH; δ H 4.25, 1H, dd, J = 8.5, 8.0 Hz). One olfeinic methyl (δ H 1.72, 3H, s), and three tertiary methyls (δ H 1.33, 1.32, and 0.73, each 3H, s), were also identified. Therefore, the compound was suggested to have a bicyclic structure to fulfill the five degrees of unsaturation. The bicyclic structure of 5 was found to be the same as that of one eudesmene from the nearly the same NMR data of positions 1 to 10, 16, and 17 of 5 with the corresponding sesquiterpene (14) [40]. From the COSY correlations of 5 (Figure 3), three partial structures consecutive proton systems extended from H 2 -1 to H 2 -3, H-5 to H 2 -9, and H-12 to H-14 were established. Analysis of HMBC correlations of 5 led to the establishment of its planar structure. It was also found that the key HMBC correlations observed from both H 3 -19 and H 3 -20 to the epoxide carbons C-14 (δ C 67.5, CH) and C-15 (δ C 59.8, C) and from the hydroxymethine H-13 (δ H 4.25, dd, J = 8.5, 8.0 Hz) to C-11 (δ C 146.1, C) and C-14 demonstrated the positions of the epoxide and the hydroxyl to be at C-14/C-15 and C-13, respectively. This was further proved by the matched chemical shifts of 1 H and 13 C atoms of the side chain of 5 with the correspondent atoms of the known compound 17,18-epoxyloba-8,10,13(15)-trien-16-ol (11) [13] which was also isolated in this study. Therefore, the prenyleudesmane molecular structure of 5 was established as illustrated in Figure 3.
The relative configuration of 5 was determined by analyzing the NOE correlations in the NOESY spectrum, as well as a lowest energy stable conformation generated using MM2 calculation (Figure 4). The NOE interactions of H-5 with H-7, but not with H 3 -17, reflected the 5R*, 7S*, 10S*-configuration. The NOE correlations displayed for the β-oriented H-7 with the olefinic proton H-12, but not with H 3 -18, disclosed the E geometry of the 11,12-double bond. The α-orientation of the hydroxyl at C-13 was suggested by the NOE correlations of H-12/H-13 and H-12/H-7, as shown in a molecular model in Figure 4. The NOE correlations of H-12/H-7 and H 3 -18/H-13 proved the E-geometry of C-11/C-12 double bond. The above finding and other detailed NOE correlations (Figure 4) established the relative stereochemistry of 5. The relative configuration at chiral carbons C-13 was further suggested by that correspondent to C-16 of the known biogenetically related metabolite 11 which has been also isolated from the same organism in this study. Fortunately, the larger quantity of compound 11 enabled us to determine the absolute configuration of 11 and hence that of 5, through the esterification of 16-hydroxy group in 11 by Mosher's method [41,42]. Analysis of the calculated ∆δ H (δ S − δ R ) values of protons neighboring C-16 of the prepared (S)-and (R)-2-methoxy-2-(trifluoromethyl)-2-phenylacetic (MTPA) esters (11a and 11b, Figure 6) led to the assignment of the R configuration at C-16 in 11 and consequently the correspondent 13R configuration in 5. On the basis of the above findings, the absolute configuration of 5 was established as 5R, 7S, 10S, 13R. However, the stereochemistry at C-14 remained undetermined in spite of the NOE correlation of H-14/H-12.
Mar. Drugs 2017, 15, 300 8 of 14 [13] which was also isolated in this study. Therefore, the prenyleudesmane molecular structure of 5 was established as illustrated in Figure 3. The relative configuration of 5 was determined by analyzing the NOE correlations in the NOESY spectrum, as well as a lowest energy stable conformation generated using MM2 calculation ( Figure  4). The NOE interactions of H-5 with H-7, but not with H3-17, reflected the 5R*, 7S*, 10S*configuration. The NOE correlations displayed for the β-oriented H-7 with the olefinic proton H-12, but not with H3-18, disclosed the E geometry of the 11,12-double bond. The α-orientation of the hydroxyl at C-13 was suggested by the NOE correlations of H-12/H-13 and H-12/H-7, as shown in a molecular model in Figure 4. The NOE correlations of H-12/H-7 and H3-18/H-13 proved the Egeometry of C-11/C-12 double bond. The above finding and other detailed NOE correlations ( Figure  4) established the relative stereochemistry of 5. The relative configuration at chiral carbons C-13 was further suggested by that correspondent to C-16 of the known biogenetically related metabolite 11 which has been also isolated from the same organism in this study. Fortunately, the larger quantity of compound 11 enabled us to determine the absolute configuration of 11 and hence that of 5, through the esterification of 16-hydroxy group in 11 by Mosher's method [41,42]. Analysis of the calculated ΔδH (δS − δR) values of protons neighboring C-16 of the prepared (S)-and (R)-2-methoxy-2-(trifluoromethyl)-2-phenylacetic (MTPA) esters (11a and 11b, Figure 6) led to the assignment of the R configuration at C-16 in 11 and consequently the correspondent 13R configuration in 5. On the basis of the above findings, the absolute configuration of 5 was established as 5R, 7S, 10S, 13R. However, the stereochemistry at C-14 remained undetermined in spite of the NOE correlation of H-14/H-12. The cytotoxic activity of the isolated compounds (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12) were screened against human lung adenocarcinoma (A549), human prostatic carcinoma (LN-cap), and human colon adenocarcinoma (DLD-1) cell lines using the Alamar Blue assay. The results showed that these compounds are not cytotoxic toward the three cancer cell lines.
Since many lobane diterpenoids were reported to exhibit anti-inflammatory activity through different mechanisms [25][26][27], the isolated metabolites in this study were evaluated for their antiinflammatory potential through measuring their ability to suppress fMLP/CB-induced superoxide anion generation and elastase release in human neutrophils. The results (Figure 7) demonstrated that compounds 2, 5, and 11 expressed a moderate inhibitory effect (22.08 ± 4.71, 20.59 ± 2.15, and 28.16 ± 5.06%, respectively) at 10 μM against superoxide anion generation in fMLP/CB-stimulated cells. Moreover, compounds 5, 8, 11, and 12 were found to be more active in inhibiting the elastase release (33.94 ± 5.85 to 45.34 ± 4.08%) than compounds 2, 4, 9, and 10 which exhibited a moderate activity (23.07 ± 6.55 to 28.44 ± 5.28%) at 10 μM. The weak inhibition against elastase release was exerted by compounds 3, 6, and 7 (11.40 ± 1.28 to 15.14 ± 2.52%). It is noteworthy to mention that although compounds 5 and 11 possessed the same side chain, it seems that the ring system of β-elemene in 11 has a role in increasing the anti-inflammatory effect relative to β-selinene moiety. Moreover, except for compound 2, other lobane diterpenoids possessing a pyran ring in their side chain (1, 3, 6, and 7) showed weaker activity against elastase release in the fMLP/CB-stimulated neutrophils. The cytotoxic activity of the isolated compounds (1-12) were screened against human lung adenocarcinoma (A549), human prostatic carcinoma (LN-cap), and human colon adenocarcinoma (DLD-1) cell lines using the Alamar Blue assay. The results showed that these compounds are not cytotoxic toward the three cancer cell lines.
Since many lobane diterpenoids were reported to exhibit anti-inflammatory activity through different mechanisms [25][26][27], the isolated metabolites in this study were evaluated for their anti-inflammatory potential through measuring their ability to suppress fMLP/CB-induced superoxide anion generation and elastase release in human neutrophils. The results (Figure 7) demonstrated that compounds 2, 5, and 11 expressed a moderate inhibitory effect (22.08 ± 4.71, 20.59 ± 2.15, and 28.16 ± 5.06%, respectively) at 10 µM against superoxide anion generation in fMLP/CB-stimulated cells. Moreover, compounds 5, 8, 11, and 12 were found to be more active in inhibiting the elastase release (33.94 ± 5.85 to 45.34 ± 4.08%) than compounds 2, 4, 9, and 10 which exhibited a moderate activity (23.07 ± 6.55 to 28.44 ± 5.28%) at 10 µM. The weak inhibition against elastase release was exerted by compounds 3, 6, and 7 (11.40 ± 1.28 to 15.14 ± 2.52%). It is noteworthy to mention that although compounds 5 and 11 possessed the same side chain, it seems that the ring system of β-elemene in 11 has a role in increasing the anti-inflammatory effect relative to β-selinene moiety.

General Procedures
Optical rotations were measured on a JASCO P-1020 polarimeter (Jasco Corporation, Tokyo, Japan). IR spectra were recorded on a JASCO FT/IR-4100 spectrophotometer (Jasco). ESIMS and HRESIMS data were performed on a BRUKER APEX II mass (Bruker, Bremen, Germany) spectrometers. The NMR spectra were recorded on a Varian Unity INOVA 500 FT-NMR (Varian Inc., Palo Alto, CA, USA) at 500 MHz for 1 H and 125 MHz for 13 C or on a Varian 400 FT-NMR (Varian Inc.) at 400 MHz for 1 H and 100 MHz for 13 C in CDCl3 or C6D6 using TMS as internal standard (δ in ppm, J in Hz). Silica gel 60 (230-400 mesh, Merck, Darmstadt, Germany) pre-coated silica gel plates (Merck, Kieselgel 60 F254, 0.2 mm) were used for open CC and analytical TLC analysis, respectively. Isolation by HPLC was performed by a Hitachi L-2455 instrument (Hitachi Ltd., Tokyo, Japan) equipped with a reversed-phase (RP-18) column (ODS-3, 5 μm, 250 × 20 mm, Sciences Inc., Tokyo, Japan).

General Procedures
Optical rotations were measured on a JASCO P-1020 polarimeter (Jasco Corporation, Tokyo, Japan). IR spectra were recorded on a JASCO FT/IR-4100 spectrophotometer (Jasco). ESIMS and HRESIMS data were performed on a BRUKER APEX II mass (Bruker, Bremen, Germany) spectrometers. The NMR spectra were recorded on a Varian Unity INOVA 500 FT-NMR (Varian Inc., Palo Alto, CA, USA) at 500 MHz for 1 H and 125 MHz for 13 C or on a Varian 400 FT-NMR (Varian Inc.) at 400 MHz for 1 H and 100 MHz for 13 C in CDCl 3 or C 6 D 6 using TMS as internal standard (δ in ppm, J in Hz). Silica gel 60 (230-400 mesh, Merck, Darmstadt, Germany) pre-coated silica gel plates (Merck, Kieselgel 60 F254, 0.2 mm) were used for open CC and analytical TLC analysis, respectively. Isolation by HPLC was performed by a Hitachi L-2455 instrument (Hitachi Ltd., Tokyo, Japan) equipped with a reversed-phase (RP-18) column (ODS-3, 5 µm, 250 × 20 mm, Sciences Inc., Tokyo, Japan).

Cytotoxicity Assay
Cancer cell (A549, LN-cap, and DLD-1) lines were purchased from the American Type Culture Collection (ATCC). Alamar Blue assay [44,45] protocol was used to evaluate the cytotoxicity for the isolated metabolites from L. varium.

In Vitro Anti-Inflammatory Assay
Human neutrophils were obtained from whole blood using dextran sedimentation and Ficoll centrifugation. Purified neutrophils were resuspended in a Ca 2+ -free HBSS buffer (pH 7.4) at 4 • C prior to use.

Measurement of Superoxide Anion Generation
The production of superoxide anion was assayed by the method based on the superoxide oxide dismutase inhibitable reduction of ferricytochrome c [46,47]. Briefly, neutrophils incubated with ferricytochrome c (0.5 mg/mL) and Ca 2+ (1 mM) were equilibrated at 37 • C for 2 min and then treated with different concentrations of the tested compounds for 5 min. Cells were activated by 100 nM fMLP for 10 min in the pretreatment of cytochalasin B (CB, 1 µg/mL) for 3 min (fMLP/CB).
The evaluation of anti-inflammatory activity showed that diterpenoids 2, 5, and 11 possess moderate inhibitory activity on the generation of superoxide anion, while 5, 8, 11, and 12 could effectively suppress elastase released after stimulation of human neutrophils by fMLP/CB. The active metabolites might be considered as promising leads in the development of anti-inflammatory drugs.