Cembrane-type compounds have been found to be the important diterpenoidal constituents in marine coelenterates [1,2,3,4,5,6,7,8]. In the investigation of the bioactive metabolites from soft corals of Taiwanese waters, many bioactive cembranoids have been isolated from octocorals (Alcyonaceae) belonging to the genera Sinularia [9,10,11,12], Lobophytum [13,14], Sarcophyton [15,16,17,18,19,20] and Pachyclavularia [21]. Some of these metabolites have been shown to exhibit cytotoxic activity against the growth of various cancer cell lines [9,11,14,15,16,17,18,19,20,21], and/or anti-inflammatory activity [10,14]. Our recent study of the chemical constituents of the Dongsha Atoll soft coral Sarcophyton crassocaule [22,23] has yielded cembranoids sarcocrassocolides A–L, of which some were found to exhibit significant cytotoxic and anti-inflammatory activities. Our continuing chemical investigation on the same collection of this organism, with the aim of discovering other biologically active natural products, again led to the isolation of three new cembranoids, sarcrocrassocolides M–O (1–3) (Chart 1). The structures of 1–3 were established by extensive spectroscopic analysis, including careful examination of 2D NMR (¹H–¹H COSY, HMQC, HMBC and NOESY) correlations. The cytotoxicity of compounds 1–3 against human breast carcinoma (MCF-7), human colon carcinoma (WiDr), human laryngeal carcinoma (HEp-2) and human medulloblastoma (Daoy) cell lines was studied, and the ability of 1–3 to inhibit the up-regulation of pro-inflammatory iNOS (inducible nitric oxide synthase) and COX-2 (cyclooxygenase-2) proteins in LPS (lipopolysaccharide)-stimulated RAW264.7 macrophage cells was also examined. It was found that compounds 1–3 were cytotoxic towards the above cancer cells; 2 being the most cytotoxic. Compounds 1–3 were found to significantly inhibit the expression of iNOS protein.

The HRESIMS (m/z 429.1892 [M + Na]⁺) of sarcrocrassocolide M (1) established the molecular formula C₂₂H₃₀O₇, appropriate for eight degrees of unsaturation, and the IR spectrum revealed the presence of lactonic carbonyl (1757 cm⁻¹) group. The ¹³C NMR and DEPT (Table 1) spectroscopic data showed signals of three methyls (including one acetate methyl), five sp³ methylenes, two sp² methylenes, five sp³ methines (including four oxymethines), one sp² methines, one sp³ and five sp² quaternary carbons (including two ester carbonyls). The NMR signals (Table 1) observed at δ_C 169.1 (C), 139.3 (C), 121.6 (CH₂), 81.4 (CH), and 37.4 (CH), and δ_H 6.28, 5.63 (each, 1H, d, J = 2.0 Hz), 4.61 (1H, t, J = 2.5 Hz), and 3.07 (1H, dt, J = 11.5, 2.5 Hz) showed the presence of an α-methylene-γ-lactonic group by comparing with similar NMR data of known cembranoids with the same five-membered lactone ring [22,23]. Signals resonating at δ_C 60.1 (C), 60.0 (CH) and δ_H 2.53 (1H, dd, J = 7.0, 4.0 Hz) revealed the presence of a trisubstituted epoxide. One trisubstituted and one 1,1-disubstituted double bond were also identified from NMR signals appearing at δ_C 129.4 (C), 128.1 (CH), and δ_H 5.46 (1H, dd, J = 7.0, 5.5 Hz), and at δ_C 113.5 (CH₂), 146.6 (C), δ_H 5.16 and 5.12 (1H, s, each), respectively. In the ¹H–¹H COSY spectrum, it was possible to identify three different structural units, which were assembled with the assistance of an HMBC experiment. Key HMBC correlations of H₃-18 to C-3, C-4 and C-5; H₂-19 to C-7, C-8 and C-9; H₃-20 to C-11, C-12 and C-13 and H₂-17 to C-1, C-15 and C-16 permitted the establishment of the carbon skeleton (Figure 1). Furthermore, the acetoxy group positioned at C-13 was confirmed from the HMBC correlations of the methyl protons of an acetate (δ_H 2.02) to the ester carbonyl carbon at δ_C 169.2 and the oxymethine signal at 77.4 (C-13, CH). The ¹³C NMR signals at δ_C 87.1 (CH) and HRESIMS showed the presence of a hydroperoxy group at a methine carbon C-7 [9]. On the basis of the above analysis, the planar structure of 1 was established unambiguously. The relative structure of 1 was elucidated by the analysis of NOE correlations, as shown in Figure 2. The NOE interaction of H-1 (δ 3.07) with H-2β (δ 1.71), H-3 (δ 2.57) and H-11 (δ 5.46) revealed the β-orientation of H-1 and H-3. H-3 did not exhibit NOE correlation with H₃-18 (δ 1.32, s) instead it correlated with one proton of H₂-5, reflecting the trans stereochemistry of 3,4-epoxide. The proton H-7 (δ 4.30) showed NOE interactions with H-3, and both H-7 and H-11 had NOE correlations with one proton of H₂-9 (δ 2.14). Thus, H-7 was placed on the β-face. The E geometry of the trisubstituted double bond at C-11 and C-12 was assigned from the NOE correlation of H₃-20 (δ 1.76) with one proton of H₂-10 (δ 2.40), but not with the olefinic proton H-11, in addition to the upper field chemical shift of C-20 (δ 14.8). H-14 (δ 4.61) exhibited NOE correlations with both H-13 (δ 5.40) and H₃-20, but not with H-1, indicating the α-orientation of both H-13 and H-14. These results, together with other detailed NOE correlations of 1 (Figure 2), unambiguously established the structure of sarcocrassocolide M, as shown in formula 1. Therefore, the relative structure of compound 1 was determined.

Compound 2 possessed the same molecular formula (C₂₂H₃₀O₇) as that of 1, as revealed from HRESIMS. Furthermore, it was found that the NMR spectroscopic data of 2 (Table 1) were similar to those of 1. Analysis of the 2D NMR (¹H–¹H COSY, HMQC, and HMBC) correlations revealed that compound 2 possesses the same planar structure as that of 1. From the NOESY spectrum, it was found that H-7 (δ 4.38) showed a weak NOE interaction with H₃-20 (1.78), but not with H-11 (δ 5.41), revealing the α-orientation of H-7. Further analysis of other NOE interactions revealed that 2 possessed the same relative configurations at C-1, C-3, C-4, C-13 and C-14, as those of 1 (Figure 2). Therefore, 2 was found to be the C-7 epimer of 1.

Compound 3 was shown by HRESIMS to possess the molecular formula C₂₀H₂₈O₅ (m/z 371.1835 [M + Na]⁺). The IR spectrum of 3 also revealed the presence of lactonic carbonyl (1752 cm⁻¹) group. Comparison of the ¹H and ¹³C NMR data (Table 1) of compounds 1 and 3 showed that the structure of 3 should be very close to that of 1, with the exception of signals assigned to C-13, where an acetoxymethine (δ_H 5.40, 1H, s; δ_C 77.4) in 1 was replaced by a methylene (δ_H 2.58, 1H, brd, J = 14.4 Hz, δ_H 2.32, 1H, dd, J = 14.4, 8.0 Hz; δ_C 44.9) in 3. The planar structure of 3 was elucidated by analyzing the ¹H–¹H COSY and HMBC correlations (Figure 1). The relative stereochemistry of 3 was confirmed from the key NOESY correlations (Figure 3), and the structure of sarcocrassocolide O, as shown in formula 3, was established unambiguously. Thus, 3 is the 13-deacetoxy derivative of 1.

Similar to sarcocrassocolides F–L [22], 1–3 should be the oxidized products of the related 3α,4α-epoxycembranolides with 7,8-olefinic group, although we have not yet discovered the similar oxidation from cembranolides possessing 7,8-double bond and 3β,4β-epoxide, such as sarcocrassolide (4), sinularolide E (5), and 13-acetoxysarcocrassolide (6) (Chart 2), which were isolated by our previous study [24,25].

The cytotoxicity of compounds 1–3 against the proliferation of a limited panel of cancer cell lines, including Daoy, HEp-2, MCF-7 and WiDr carcinoma cell lines was evaluated. The results (Table 2) showed that all compounds 1–3 were found to exhibit cytotoxicity against all or part of the above carcinoma cell lines. In this assay, the in vitro anti-inflammatory effects of compounds 1–3 were also tested. The inhibition of LPS-induced up-regulation of pro-inflammatory proteins, iNOS and COX-2 in RAW264.7 macrophage cells was measured by immunoblot analysis (Figure 4). At a concentration of 10 μM, compounds 1–3 were found to significantly reduce the levels of iNOS protein to 4.2 ± 1.6%, 52.9 ± 12.8%, and 22.7 ± 2.8%, respectively, relative to the control cells stimulated with LPS only. At the same concentration metabolites 2 and 3 did not show activity in inhibiting the expression of the pro-inflammatory COX-2 expression with LPS treatment, but compound 1 could reduce the expression of COX-2 to 62.8 ± 22.4%. Thus, compounds 1–3 might be useful anti-inflammatory agents, while 1 is a promising anti-inflammatory lead compound. Compound 1 could inhibit the expression of both iNOS and COX-2 which might be arisen from the presence of β-hydroperoxy group at C-7 by comparison with compound 2.

Our investigation demonstrated that the soft coral, S. crassocaule, could be a good source of bioactive substances. The isolated compounds 1–3, in particular 1, are potentially anti-inflammatory and may become lead compounds in the future drug development. Also, it is noteworthy to mention that cembranoids 1–3 possessing an α-methylene-γ-lactonic group with a rarely found 1,1-disubstituted double bond at C-19/C-8 and containing a hydroperoxy group at C-7, were discovered for the first time from corals of this species.