Compound 1 was obtained as yellow oil, and its molecular composition of C₄₀H₅₈O₂ was established by HR-ESI-MS (m/z 571.4751 [M + H+]⁺, C₄₀H₅₈O₂⁺). The ESI-MS of 1 showed quasimolecular ion peaks at m/z 571.3 [M + 1]⁺ and 592.7 [M + Na]⁺, respectively. The IR absorptions at 1724 1650, 1237, 1001 and 798 cm^–1 showed the presence of carbalkoxy and double groups. Analyses of the ¹H-NMR, ¹³C-NMR, DEPT and HMQC spectra revealed not only the presence of 7 methyls, 15 methylenes, 7 methines and 11 quaternary carbon (including one carbalkoxy group at δ_C 177.0, d), but also the presence of one tetrasubstituted double bond, one tetrasubstituted double bond and two terminal double bonds (Tables 1). Four methyl singlets (δ_C 24.7, q; 33.7, q; 22.1, q; 15.4, q) and one trisubstituted double bond (δ_C 137.7, s; 128.5, d) suggesting that a pimarane-type diterpenoid structure might be presented [9,18–21]. Meanwhile, three methyl singlets (δ_C 23.0, q; 33.1, q; 11.8, q) and two terminal double bonds (δ_C 139.5, s; 105.9, t; 151.0, d; 109.0, t) might indicate the structure of a dolabradiene-type diterpenoid [22–25]. Therefore, Compound 1 seemed to be a dimeric diterpenoid composed of two moieties, one being a pimarane-type diterpenoid (part A) and the other a dolabradiene-type diterpenoid (part B).

The HMBC correlations of Me18 (δ_H 0.89, s) with C-5 (δ_C 55.0, d), C-3 (δ_C 42.1, t), C-4 (δ_C 33.3, s) and C-19 (δ_C 22.1, q); Me19 (δ_H 0.85, s) with C-5 (δ_C 55.0, d), C-3 (δ_C 42.1, t), C-4 (δ_C 33.3, s) and C-18 (δ_C 33.7, q); Me20 (δ_H 0.82, s) with C-9 (δ_C 49.8, d), C-1 (δ_C 39.2, t); Me17 (δ_H 0.99, s) with C-14 (δ_C 128.5, d), C-15 (δ_C 56.2, d), C-13 (δ_C 38.0, s) and C-12 (δ_C 32.4, t); H-5 (δ_H 1.06, m) with C-20 (δ_C 15.4, q), C-6 (δ_C 22.7, t); H-9 (δ_H 1.74, m) with C-8 (δ_C 137.7, s), C-14 (δ_C 128.5, d), C-20 (δ_C 15.4, q), C-11 (δ_C 18.6, t)and C-10 (δ_C 38.87, s); H-14 (δ_H 5.55, s) with C-9 (δ_C 49.8, d), C-13 (δ_C 38.0, s), C-7 (δ_C 36.23, t), C-12 (δ_C 32.4, t) and C-17 (δ_C 24.7, q) confirmed that part A belonged to a pimarane-type diterpenoid. The HMBC correlations of Me-20′ (δ_H 0.57, s) with C-10′ (δ_C 53.3, d), C-8′ (δ_C 43.0, d), C-9′ (δ_C 38.1, s ) and C-11′ (δ_C 35.2, t); Me-19′ (δ_H 1.14, s) with C-4′ (δ_C 139.5, s), C-10′ (δ_C 53.3, d), C-5′ (δ_C 39.1, s ) and C-6′ (δ_C 36.8, t); Me-17′ (δ_H 1.01, s) with C-15′ (δ_C 151.0, d), C-13′ (δ_C 36.18, s) and C-12′ (δ_C 32.0, t); H-15′ (δ_H 5.81, dd, J = 12.6, 21 Hz) with C-14′ (δ_C 38.93, t), C-13′ (δ_C 36.18, s), C-12′ (δ_C 32.0, t) and C-17′ (δ_C 23.0, q); H-16′ (δ_H 4.82, d, J = 13.2 Hz) with C-15′ (δ_C 151.0, d) and C-13′ (δ_C 36.18, s); H-1′ (δ_H 2.28, m) with C-3′ (δ_C 148.6, s), C-2′ (δ_C 113.6, s), C-10′ (δ_C 53.3, d), C-9′ (δ_C 38.1, s) and C-5′ (δ_C 39.1, s); H-1′ (δ_H 2.56, dd, J = 7.8, 14.4 Hz) with C-3′ (δ_C 148.6, s), C-2′ (δ_C 113.6, s), C-10′ (δ_C 53.3, d) and C-9′ (δ_C 38.1, s); H-10′ (δ_H 1.38, m) with C-5′ (δ_C 39.1, s), C-9′ (δ_C 38.1, s), C-11′ (δ_C 35.2, t), C-19′ (δ_C 33.1, q) and C-1′ (δ_C 22.2, t); H-18′ (δ_H 5.26, s) with C-5′ (δ_C 39.1, s), C-3′ (δ_C 148.6, s) and C-4′ (δ_C 139.5, s); H-18′ (δ_H 5.08, s) with C-5′ (δ_C 39.1, s), C-3′ (δ_C 148.6, s) and C-4′ (δ_C 139.5, s) confirmed that part B belonged to a dolabradiene-type diterpenoid. Then the chemical shift of δ_C 56.2 (d, C-15), δ_C 177.0 (s, C-16), δ_C 148.6 (s, C-3′) were compatible with a bond between C-3′ and the lactone O-atom, which was further confirmed by the HR-ESI-MS spectrum. The HMBC correlations from H-15 (δ_H 3.20, s) to C-3′ (δ_C 148.6, s), C-14 (δ_C 128.5, d), C-2′ (δ_C 113.6, s), C-13 (δ_C 38.0, s), C-12 (δ_C 32.4, t), C-17 (δ_C 24.7, q) and C-16 (δ_C 177.0, s) established the connection between part A and part B at the site of C-15, C-2′ and C-3′ (Figure 2). Furthermore, other HMBC correlations along with the comparison of its NMR data and the known pimarane-type diterpenoid and dolabradiene-type diterpenoid comprehensively confirmed that the planar structure of 1 was assigned as a novel dimeric diterpenoid [18–26].

The relative stereochemistry of 1 was established by NOE spectrum. Comparison of the NMR data of 1 with the pimarane-type diterpenoid and dolabradiene-type diterpenoid showed that the Me-17 (δ_C 24.7, q), Me-18 (δ_C 33.7, q) and Me-20′ (δ_C 11.8, q) were in a relative α-orientation, while Me-19 (δ_C 22.1, q), Me-20 (δ_C 15.4, q), Me-19′ (δ_C 33.1, q) and Me-17′ (δ_C 23.0, q) were in a relative β-orientation [18–26]. NOE correlations between Me-18/H-5 and H-5/H-9 indicated that H-5 and H-9 were on the β-side of 1. NOE correlations between H-10′/Me-19′ and Me-17′/H-8′ indicated that they were oriented on another side of the nucleus (Figure 3).

Five known terpenoids, Tagalsin C (2), Tagalsin I (3), lup-20(29)-ene-3β,28-diol (4), 3-oxolup- 20(29)-en-28-oic acid (5) and 28-hydroxylup-20(29)-en-3-one (6), were identified by comparison of their physical and spectral data with those in the literature [23,27–31].

The effects of terpenoids on settlement and mortality of B. albicostatus cyprids are shown in Figure 4 and their EC₅₀ and LC₅₀ values are summarized in Table 2. Capsaicin was used as a positive standard with EC₅₀ value of 1.32 μg/cm² (Figure 4 and Table 2). Filtered seawater (FSW) was used as a negative control. Comparison of means using ANOVA and Dunnet’s test showed that all of the terpenoids significantly inhibited settlement compared with the negative control (P < 0.001) without significant toxicity (P > 0.05). From Dunnet’s test, capsaicin significantly reduced larval settlement at 5 μg/cm² (P < 0.001), and Compounds 1–6 significantly reduced larval settlement at 0.5–15 μg/cm² (P < 0.001). Moreover, the non-toxicity compounds were defined as those that do not directly kill fouling organisms at or near the levels at which they deter fouling [32]. In the present study, the LC₅₀ values of these terpenoids were all above 250 μg/cm². The lower EC₅₀/LC₅₀ ratio indicated that all these terpenoids had the ability to inhibit larval settlement in a non-toxic way. In addition, the stability test showed that Compounds 1 and 3 were degraded while Compounds 2, 4, 5 and 6 remained stable over 2-month exposure under FSW.

As shown in Table 2, Compounds 1–6 exhibited antifouling activity against cyprid larvae of the barnacle B. albicostatus, with EC₅₀ values ranging from 0.65 to 13.9 μg/cm². Compound 2 was more active than capsaicin, while Compounds 1, 3, 4, 5 and 6 showed antifouling activity weaker than capsaicin. Among these terpenoids, diterpenoid (Compound 2) showed the best antifouling activity against cyprid larvae of the barnacle B. albicostatus, followed by triterpenoid (Compounds 4, 5 and 6), and then dimeric diterpenoid (Compounds 1 and 3). In addition, the antifouling activity of Compound 3 was stronger than Compound 1, which suggested that the diterpenoid dimmer with two identical diterpenoid subunits might display more potent antifouling activity than the one with two different diterpenoid subunits. By comparing the EC₅₀ value of Compound 5 and Compound 6 and considering that the only difference between the structures of these two compounds is the difference group on the C-28 of lupane triterpenoid, it is evident that the carboxyl group at the C-28 of lupane triterpenoid might have a more potent antifouling activity than the oxygenated methylene group. However, when carbonyl group at the C-3 position in lupane triterpenoid (Compound 6) is substituted by hydroxyl group, the resulting compound (Compound 4) showed almost the same antifouling activity as Compound 6. This result indicated that the hydroxyl group might be another important functional group expressing potent antifouling activity in lupane triterpenoid.

¹H-NMR spectra were measured with Varian UNITY PLUS 500 spectrometer (USA). ¹H-, ¹³C- and 2D-NMR spectra were measured with Varian INOVA 600 spectrometer (USA). High-resolution ESI mass spectra data were measured with a Waters Q-TOF MicroTM spectrometer (USA) and were provided by Zhengzhou University, China. Low-resolution ESI mass spectra data were recorded on an AB 3200Q TRAP spectrometer (USA). IR spectra were measured using a Nicolet 380 FT-IR spectrophotometer (USA). The optical rotation data were obtained on a Rudolph Autopol IV polarimeter (USA). 200–300 mesh Silica gel (Qingdao Marine Chemical Factory, Qingdao, China) were used for column chromatography. Compounds were monitored by TLC on GF 254 silica gel (Qingdao Marine Chemical Factory, Qingdao, China). ODS and Sephadex LH-20 were obtained from Pharmacia Co. Capsaicin was purchased from Sigma Chemical Co., USA, with purity above 95%.

The roots of C. tagal were collected from Hainan Province, China in July 2005. The air-dried and powdered material (4.1 kg) was extracted with 95% EtOH at room temperature. The extract was concentrated and partitioned sequentially with petroleum ether, ethyl acetate and H₂O. The petroleum ether extract (11.5 g) was subjected to a silica gel column, eluted with a petroleum ether–ethyl acetate gradient to obtain 13 fractions (A–M). Fraction C (0.135 g) was chromatographed over silica gel and Sephadex LH-20 repeatedly, and the major fraction obtained was then purified by reversed-phase semi-preparation HPLC (elution with 1:1 hexane-CHCl₃) to yield Compound 1 (50 mg). Fraction K (1.059 g) was chromatographed over silica gel and Sephadex LH-20 repeatedly (elution with 30:1 petroleum ether–acetone) to yield Compound 4 (5 mg) and Compound 5 (10 mg). Fraction J (5.3 g) was chromatographed over silica gel repeatedly (elution with 30:1 petroleum ether–acetone) to yield Compound 6 (6 mg). Moreover, the ethyl acetate extract (73.6 g) was subjected to a silica gel column, eluted with a CHCl₃–MeOH gradient to obtain 12 fractions (A2-L2). Fraction B2 (8.32 g) was chromatographed over silica gel and Sephadex LH-20 repeatedly. Compound 2 (10 mg) was further separated by silica gel (elution with 40:1 petroleum ether–ethyl acetate) and Compound 3 (53 mg) was purified by silica gel (elution with 2.5:1 hexane-CHCl₃).

8(14)-enyl-pimar-2′(3′)-en-4′(18′)-en-15′(16′)-en-dolabr-16,15,2′,3′-oxoan-16-one (1): C₄₀H₅₈O₂; yellow oil; [α]_D²⁵ = +68.25 (c = 0.50, CHCl₃); UV (CHCl₃): 214 (2.5905), 224 (3.0367), 234 (2.929), 244 (2.7263). IR ν_max (KBr): 1724,1650, 1237, 1001, 798 cm^–1; ESI-MS, m/z 571.3 ([M + 1]⁺), 592.7 [M + Na]⁺, HR-ESI-MS at m/z 571.4751 ([M + H+]⁺, C₄₀H₅₈O₂⁺calc. 571.4741 ); ¹H-NMR and ¹³C-NMR data for 1 are listed in Table 1.

Antifouling efficacies of the six terpenoids isolated from C. tagal were evaluated by the settlement inhibition assay with cyprid larvae of B. albicostatus. B. albicostatus adults were collected from intertidal rocks in Xiamen, China. Adults released the nauplius I and nauplius II stages upon immersion in seawater after dried for 12 h [8]. Naupilar larvae actively swimming towards the light were collected using a pipette. They were cultured in filtered seawater (FSW, 0.22 μm, salinity 30‰, and temperature 25 °C) by feeding with the diatom Chaetoceros muelleri at a concentration of 2.0 × 10⁵ cells·mL⁻¹. The FSW and diet were renewed everyday. After 5~6 days, most of the nauplii had metamorphosed to cyprids, and then the cyprids were collected within 24 h. Antifouling efficacies of the six terpenoids were investigated according to the method of Hellio et al. with a minor modification [33]. The tested compounds were introduced to the glass dishes using CH₂Cl₂ as carrier solvent. After evaporation of the solvent at room temperature, 30 cyprids and 10 mL FSW were added to each glass dishe. There were three replicates for FSW negative control and each concentration of the tested compounds. Dishes were incubated in dark at 25 °C for 48 h. After that, numbers of larvae which settled, died or did not settle in each replicate were enumerated under a stereomicroscope. Cyprids that did not move and did not respond to a touch with a metal probe were regarded as dead [34]. Cyprids that permanently attached and metamorphosed were counted as settled [33,35]. In addition, the stability of terpenoids was tested using the method of thin layer chromatography (TLC), which were identified by comparison of the spots and R_f values of these terpenoids before and after 2-month exposure in FSW control [17]. Three replicates were set up for each of FSW control. The tested terpenoids were dissolved in CH₂Cl₂ and evaporated after 2-month exposure in FSW (three replicates). Compounds that showed the same spots and R_f values with before were regarded as stability. Compounds that showed the different spots and R_f values with before were regarded as degradation. Moreover, statistical calculations were performed with the SPSS software package. One-way ANOVA followed by a Dunnett's test for multiple comparisons of treatment means with a negative control was used for the settlement or mortality percentages analysis. The significance level was defined as p < 0.05. Both EC₅₀ (the concentration that reduces the settlement rate by 50% relative to the negative control) and LC₅₀ (the concentration that results in 50% mortality relative to the negative control) were estimated using the Spearman-Karber method with a 95% confidence interval [36,37].