Sinularones A–I, New Cyclopentenone and Butenolide Derivatives from a Marine Soft Coral Sinularia sp. and Their Antifouling Activity

Nine new compounds, namely sinularones A–I (1–9), characterized as cyclopentenone and butenolide-type analogues, were isolated from a soft coral Sinularia sp., together with a known butenolide (10). Their structures were elucidated by means of spectroscopic (IR, MS, 1D and 2D NMR, CD) analysis. The absolute configurations were determined on the basis of CD and specific rotation data in association with the computed electronic circular dichroism (ECD) by time dependent density functional theory (TD DFT) at 6-31+G(d,p)//DFT B3LYP/6-31+G(d,p) level. Compounds 1–2 and 7–10 showed potent antifouling activities against the barnacle Balanus amphitrite.


Introduction
The genus Sinularia is a dominant biomass with about 100 known species widely distributed in the tropical reef environment [1][2][3][4], presenting a rich array of chemical diversity involving sesquiterpenes, diterpenes, polyhydroxylated steroids, and polyamine compounds. Several metabolites derived from Sinularia play an active role in chemical defense [5,6], and display a range of biological activities, such as antimicrobial [7], anti-inflammatory [8], and cytotoxic [9,10]. In the course of our search for bioactive natural products from marine invertebrates inhabited in the South China Sea, an octocoral Sinularia sp. was collected off Hainan Island. Chemical examination of the EtOAc soluble fraction of this specimen resulted in the isolation of the new cyclopentenone derivatives (1)(2)(3)(4)(5)(6) and furanones (7)(8)(9), along with the known butenolide (10) (Figure 1). This paper reports the structural elucidation of the new compounds and the evaluation of their antifouling activities.
The NMR spectroscopic data of sinularone C (3) were closely related to those of 2, except for the absence of the signals for an ethoxy group. Analysis of 1D and 2D NMR (COSY, HMQC and HMBC) disclosed a 2,3-dimethylcyclopent-2-enone nucleus, and a n-pentyl group being linked to the nucleus at C-4, showing the partial structure to be the same as that of 2. The COSY cross-peaks between H-5 (δ H 3.11) and H 2 -6 (δ H 2.49, 3.08) and the HMBC interactions of these protons with a carbonyl carbon C-7 (δ C 174.9) and the ketone C-1 (δ C 204.9) allowed to link C-6 of an acetyl unit to cyclopent-2-enone nucleus at C-5. In addition, a quaternary carbon observed at δ C 92.4 was assigned to C-4 on the basis of the HMBC interactions of C-4 to H-5 and H 2 -6. The molecular formula (C 14 H 20 O 3 ) (HRESIMS m/z 259.1315 [M + Na] + ) of 3 showing a C 2 H 6 O unit less than that of 2 and requiring 5° unsaturation, in association with the obvious downfield shifted C-4 in comparison with that of 2, allowed to connect C-4 and C-7 to form a γ-lactone, which was fused to the cyclopentenone ring across C-4 and C-5. The observed NOE interactions from H-5 to H 2 -8 (δ H 1.79, 1.97) and H 2 -9 (δ H 1.10, 1.16) indicated H-5 to be oriented in the same face as n-pentyl group. This assignment was supported by the irradiation of H 2 -8 causing the NOE enhancement of H-5 (1.7%, 1.2%). Compound 3 is likely a precursor to generate 2 by ethoxylation. Thus, the absolute configuration of C-5 in 3 is suggested to be the same as that of 2. Accordingly, the chiral centers of 3 were assumed to be 4R and 5S.   The molecular formula (C 14 H 20 O 3 ) of sinularone D (4) was determined to be the same as that of 3 on the basis of the HRESIMS data (m/z 259.1313 [M + Na] + , calcd. 259.1310) and NMR data. The IR absorptions at 1778, 1710 and 1657 cm −1 were characteristic of lactone, ketone, and olefinic groups. Comparison of the NMR data revealed the resonances of 4 in respect to those of a 2,3-dimethylcyclopent-2-enone nucleus (Tables 1 and 2) compatible to the data of 3. However, the COSY and HMBC relationships revealed a n-butyl group to replace a n-pentyl group of 3. This unit was determined to be linked to C-5 (δ C 54.5) as evident from the COSY correlation between H 2 -6 (δ H 1.45, 1.67) and H-5 (δ H 2.55, t, J = 6.5 Hz) in association with the HMBC interactions from H 2 -6 to C-1 (δ C 203.4). The remaining resonances included two methylenes and a carbonyl carbon (δ C 176. 1, C-12). The COSY correlation between H 2 -10 (δ H 2. 16, 2.23) and H 2 -11 (δ H 2.72, 2.78) and the HMBC interactions from H 2 -10 to C-12 and from H 2 -11 to C-4 (δ C 91.5, C), in association with the molecular formula requiring 5° unsaturation, conducted the quaternary carbon C-4 to be located by a γ-lactone. The NOE interaction between H 2 -6 and H 2 -10 reflected the same orientation of H-5 and the heterocyclic atom of lactone. In regard to the absolute configuration of the spiro carbon C-4, we extended the CD method originally used for the chiral center of allylic amines [14]. The negative Cotton effect at 222 nm belonging to the π-π* charge-transfer transition polarized by the lactone carbonyl group and the enone unit, correlated with anti-clockwise screw ( Figure 5). Therefore, the sign of the Cotton effect reflected a 4R configuration. In combination with NOE data, C-5 was thus assigned to S configuration. The NMR spectroscopic data of 5 and 6 are very similar, while HRESIMS data revealed both compounds shared the same molecular formula (C 15 H 24 O 3 ), indicating 4° unsaturation. APT spectrum of 5 displayed 15 carbon resonances, involving one ketone (δ C 209.4, C-1), two olefinic carbons (δ C 138. 8, C-5; 170.8, C-4), two oxymethines at δ C 85.2 (C-8) and 75.3 (C-11) and four methyls (δ C 21.3, 21.9, 28.4, 28.6). The observation of HMBC interactions from the isolated methylene protons at δ H 2.18 (2H, s, H 2 -2) to C-1, C-3 (δ C 38.9, C), C-4, and C-5 and from H-4 (δ H 7.35, s) to C-1, C-2, and C-3, in addition to the methyl singlets at δ H 1. 15 (6H, s, H 3 -13 and H 3 -14) to C-2, C-3, and C-4, ascertained the presence of 5-substituted 3,3-dimethylcyclopent-4-enone. In regard to the remaining resonances, the COSY correlations extended a moiety from C-8 to C-12, in which C-8 and C-11 were oxygenated. Additionally, the HMBC interactions from H 3 -15 (δ H 0.87, s) to the methylene carbon C-6 (δ C 33.5), quaternary carbon C-7 (δ C 72.5) and C-8 and from H-8 (δ H 3.57, t, J = 7.2 Hz) to C-11 indicated the formation of an ether bridge across C-8 and C-11, while C-7 was co-positioned by a methyl and a hydroxy groups. The side chain was linked to C-5 as evident from the HMBC relationships of H 2 -6 (δ H 2.09, 2.27) with C-1, C-5, and C-4. Based on the NOE interaction between H-8 and H-11, a cis-geometry of the 8,11-epoxy bonding was assigned. Thus, the relative configurations were suggested to be 8S* and 11S*.
Analysis of 2D NMR data revealed sinularone F (6) to be a stereoisomer of 5. Both compounds showed the NOE interaction between H-8 and H-11, indicating them to be oriented in the same face toward tetrahydrofuran ring. The major difference was found by the chemical shifts at C-6 and C-15 (Table 2), implying 6 to be a C-7 epimer of 5 rather than an enantiomer. Since the calculated ECD data could not provide confidential evidence to judge the configurations, the calculation for specific rotation and 13 C NMR data was performed. The conformations with relative energies from 0 to 2.5 kcal/mol were used in optical rotation computations at B3LYP/6-311+G(2d,p) level, while Boltzmann statistics were used for rotation computations of all conformations. The computed specific rotation [15] for 7S,8R,11R-isomer is −16.9 and −27.5 for 7R,8R,11R-isomer. These data are in opposite sign to the experimental data that of 5 ([α] D 23 +18.0) and 6 ([α] D 23 +22.6), indicating 5 and 6 to be the enantiomers of the calculated isomers. These assignments were also supported by the relative shift errors of the experimental 13 C NMR data of 6 and 5 that were in accordance with the error distribution calculated at B3LYP/6-311+G(2d,p) level (Table 3) [16]. Thus, the absolute configurations of 5 were in agreement with 7R,8S,11S, whereas those of 6 were assigned to 7S,8S,11S.  3 -9) and a methyl triplet (δ H 1.27, H 3 -1′), while 13 C NMR involved two carbonyl carbons at δ C 171.8 (C-1) and 174.8 (C-7), two olefinic carbons at δ C 124.7 (C-2) and 158.2 (C-3), and an acetal carbon at δ C 105. 5 (C-4). The HMBC interactions of H 3 -8 to C-1, C-2, and C-3 and from H 3 -9 to C-2, C-3, and C-4 disclosed a α,β-unsaturated 2,3-dimethyl-γ-lactone, the same as that of a known butenolide [17]. In addition, an ethylpropanoate was recognized by the COSY relationships between two vicinal methylenes along with a methyl triplet H 3 -1′ coupled to the oxymethylene at δ H 4.17 (H 2 -2′), in combination with the HMBC interactions from C-7 to the protons of H 2 -6 (δ H 2.49, 2.78), H 2 -5 (δ H 1.91, 2.32), and the oxymethylene (δ C 61.5). The positive specific rotation ([α] D +4.03°) and the similar Cotton effect in comparison with those of a known butenolide (11) supposed C-4 to be 4S configuration [11,17].
The NMR data of sinularone H (8) were similar to those of 7 with the exception of the presence of an additional methoxy group. Examination of the HMBC cross-peaks afforded the interactions between the carbonyl carbon (δ C 172.9, C-7) and the methoxy protons (δ H 3.58, s) and between the acetal carbon (δ C 109.1, C-4) and the oxymethylene (δ H 3.13, 3.24), requiring the formation of a methyl ester, while the ethoxy group was substituted to C-4. The absolute configuration of C-4 was supposed to be the same as that of 7 on the basis of similar specific rotation and CD data.
The NMR data of sinularone I (9) were mostly identical to those of a known butenolide (11) [17]. The distinction was attributed to the presence of an ethyl ester to replace a methyl ester of the known analogue, as evident from the molecular weight of 9 (C 21 H 36 O 5 ) to be 14 amu more than that of the latter, and the presence of an ethoxy group in its NMR spectra. The absolute configuration was determined to be 4S on the basis of the similar specific rotation and Cotton effect as those of 8 and the known analogues [17].
All compounds were tested for their cytotoxicity against a panel of tumor cell lines including human ovarian carcinoma A2780, human lung adenocarcinoma A549, human gastric carcinoma BGC823, human hepatoma Bel7402, and human colonic carcinoma HCT-8. However, they showed weak inhibitory activity with IC 50 > 10 μg/mL. In order to detect whether these compounds play a role for ecological functions, the test for antifouling activity against the larvae of the barnacle Balanus amphitrite was performed [18,19]. The bioassay results revealed compounds 1-2, 7-10 showed potent inhibition with the EC 50 values (Table 4) lower than the standard requirement (EC 50 < 25 μg/mL) in regard to the efficacy level of natural antifouling agents as established by the US navy program [20]. However, compounds 3-6 showed weak inhibition with EC 50 > 50 μg/mL. In addition, the bioactive compounds (1)(2)(7)(8)(9)(10) showed weak toxicity against the barnacle with LC 50 > 50 μg/mL. A primary discussion of structure-activity relationship implied that α,β-unsaturated 2,3-dimethyl-γ-lactone is a functional unit for anti-barnacle. Among the active compounds, 10 is the most active, suggesting it to be a promising candidate as a nontoxic natural antifouling agent.

General
Optical rotations were measured on a Perkin-Elmer 243B polarimeter. IR spectra were determined on a Thermo Nicolet Nexus 470 FTIR spectrometer. CD spectra were measured using J-810-150s spectropolarimeter (Jasco, Darmstadt, Germany). 1 H and 13 C NMR and 2D NMR spectra were recorded on a Bruker Avance 600 MHz and Bruker Avance 500 MHz using TMS as an internal standard. HRESIMS data were obtained on a LTQ Orbitrap XL instrument. HPLC was performed with a C 18 packed column (250 × 10 nm) and using a DAD detector.

Animal Material
The soft coral Sinularia sp. was collected from the inner coral reef at a depth of around 8 m in Hainan Island of China, in May 2004, and the samples were frozen immediately after collection. The specimen was identified by Leen van Ofwegen (National Museum of National History Naturalis). The voucher specimens  are deposited at the State Key Laboratory of Natural and Biomimetic Drugs, Peking University, China.

Computational Calculation
The computational ECD, specific rotation, and 13 C NMR calculations were performed by the B3LYP functional and a generic basis set, employ the 6-311+G(d,p) basis set [21,22]. This generic basis set has been shown to be effective, both efficient and reliable, in predicting structural and reactivity properties for homogeneous systems. All calculations are performed with the Gaussian 03 package with tight self-consistent field convergence and ultrafine integration grids.

Larval Settlement Bioassays
Adults of the barnacle Balanus amphitrite Darwin were exposed to air for more than 6 h, and then were placed in a container filled with fresh 0.22 µm filtered sea water (FSW) to release nauplii. The collected nauplii were reared to cyprid stage according to the method described by Thiyagarajan et al. [23]. When kept at 26-28 °C and fed with Chaetoceros gracilis, larvae developed to cyprids within four days. Fresh cyprids were used in the tests.

Conclusions
The Sinularia genus commonly produces cembranoids, which are considered to be the biomarkers for chemotaxonomy. The structural patterns such as cyclopentenone/cyclopentanone and unsaturated γ-lactone are unusual natural products found from the soft coral genus Sinularia. Present work provided a group of new cyclopentanone/cyclopanone and butenolide-type analogues to enrich the chemical diversity of Sinularia corals. The unprecedented 2,3-dimethylbutenolide based natural products were originally found from marine red algae [24] and brittle star [17], implying that dimethylbutenolides may be the signal molecules of the soft coral specimen to maintain coexistence with other benthos. This is the first report to indicate derivatives of unsaturated γ-lactone possessing potent antifouling activity. The metabolites containing a cyclopentenone or cyclopentanone ring are rarely discovered in marine soft corals, but they are often obtained from marine microorganisms. In addition, dimethylbutenolides showed potent antifouling activities, indicating that they contribute to a chemical ecological function. The literature survey revealed that the cyclopentenone group is relevant for cytotoxicity [25], thus cyclopentenone containing metabolites are regarded as the cytotoxins involved in chemical defense against predators. Whether the butenolides or cyclopentenones originated from a coral host or derived from other benthos serving as a food chain are uncertain.
The ethoxylated compounds such as 2, 7-9 are probably artifacts derived during the extraction process, whereas the naturally occurring forms are suggested to be free of the ethyl group.