Marine Prostanoids with Cytotoxic Activity from Octocoral Clavularia spp.

Octocoral of the genus Clavularia is a kind of marine invertebrate possessing abundant cytotoxic secondary metabolites, such as prostanoids and dolabellanes. In our continuous natural product study of C. spp., two previously undescribed prostanoids [clavulone I-15-one (1) and 12-O-deacetylclavulone I (2)] and eleven known analogs (3–13) were identified. The structures of these new compounds were elucidated based on analysis of their 1D and 2D NMR, HRESIMS, and IR data. Additionally, all tested prostanoids (1 and 3–13) showed potent cytotoxic activities against the human oral cancer cell line (Ca9-22). The major compound 3 showed cytotoxic activity against the Ca9-22 cells with the IC50 value of 2.11 ± 0.03 μg/mL, which echoes the cytotoxic effect of the coral extract. In addition, in silico tools were used to predict the possible effects of isolated compounds on human tumor cell lines and nitric oxide production, as well as the pharmacological potentials.


Introduction
Octocoral of the genus Clavularia (clove polyps) belongs to the family Clavulariidae, with distribution over subtidal reef flats of the Red Sea, Great Barrier Reef, and Indo-Pacific.The polyps of Clavularia show pink, yellow, or green to brownish-gray colors, which are popular in aquariums.Within all Clavularia species, four species, namely C. viridis, C. inflata, C. kentingsnsis, and C. koellikeri, were found in Taiwan.Among those four species, C. viridis and C. inflata were investigated as abundant natural product producers, and prostanoids [1], terpenoids [2,3], and steroids [4,5] were identified.To date, more than 60 marine prostanoids have been identified from the genus Clavularia.Those marine prostanoids present diverse pharmacological effects, including anti-proliferative [6], cytotoxic [7][8][9], and anti-inflammatory activities [10].
Mar. Drugs 2024, 22 The fruit of Arecae catechu Linn (areca nuts), which contains arecoline, is known to be a psychoactive substance.The use of areca nuts as chewing gum is a tradition of people who live in south-central Asia, such as Taiwan and the Philippines.However, this behavior dramatically increases the risk of oral cancer.In 2020, the overall incidence rate of oral cancer was 21.8 per 100,000, which rates fourth among cancer-related deaths of male Taiwanese people.Taiwan has the highest frequency of oral cancer worldwide [11], and there are roughly 3000 fatalities in Taiwan each year.Chemotherapy is one of the typical approaches in the treatment of oral cancer; however, side effects from chemotherapy medications might occur [12].Hence, finding effective and safe anti-oral cancer medications is essential.
In our continuous research for bioactive natural products from marine invertebrates, we recently reported the isolation and detailed structure elucidation of five new eudensamanetype sesquiterpene lactones and three new dolabellanes from Clavularia spp.[3].We now report the isolation and structure elucidation of two new and eleven known prostanoids (1-13) from the same specimen (Figure 1).Moreover, the cytotoxic effects of isolated compounds were evaluated.We also used in silico tools to predict the possible effects of thirteen compounds on human tumor cell lines (for cytotoxicity) and the nitric oxide (NO) production in macrophages, and predict the pharmacological potentials of nine compounds (1-5, 7, 8, 10, and 11) (the other four compounds were not suitable for the in silico tool).
The fruit of Arecae catechu Linn (areca nuts), which contains arecoline, is known to b a psychoactive substance.The use of areca nuts as chewing gum is a tradition of peopl who live in south-central Asia, such as Taiwan and the Philippines.However, this behav ior dramatically increases the risk of oral cancer.In 2020, the overall incidence rate of ora cancer was 21.8 per 100,000, which rates fourth among cancer-related deaths of male Ta wanese people.Taiwan has the highest frequency of oral cancer worldwide [11], and ther are roughly 3000 fatalities in Taiwan each year.Chemotherapy is one of the typical ap proaches in the treatment of oral cancer; however, side effects from chemotherapy med cations might occur [12].Hence, finding effective and safe anti-oral cancer medications i essential.
In our continuous research for bioactive natural products from marine invertebrate we recently reported the isolation and detailed structure elucidation of five new euden samane-type sesquiterpene lactones and three new dolabellanes from Clavularia spp.[3 We now report the isolation and structure elucidation of two new and eleven known pros tanoids (1-13) from the same specimen (Figure 1).Moreover, the cytotoxic effects of iso lated compounds were evaluated.We also used in silico tools to predict the possible effect of thirteen compounds on human tumor cell lines (for cytotoxicity) and the nitric oxid (NO) production in macrophages, and predict the pharmacological potentials of nin compounds (1-5, 7, 8, 10, and 11) (the other four compounds were not suitable for the i silico tool).

Structure Elucidation of Isolated Compounds
Clavulone I-15-one (1) was obtained as a colorless oil and had the molecular formul C25H32O8 (Δ = 9) based on its sodiated mass peak at m/z 483.1987 (calcd. 483.1989).Com pound 1 has UV absorption at 221 and 297 nm, corresponding to the cross-conjugate system of marine prostanoids [9].The IR absorption band was found at 1750 cm −1 , ind cating the presence of ester functional groups.The 1 H NMR spectrum (Table 1

Structure Elucidation of Isolated Compounds
Clavulone I-15-one (1) was obtained as a colorless oil and had the molecular formula C 25 H 32 O 8 (∆ = 9) based on its sodiated mass peak at m/z 483.1987 (calcd.483.1989).Compound 1 has UV absorption at 221 and 297 nm, corresponding to the cross-conjugated system of marine prostanoids [9].The IR absorption band was found at 1750 cm −1 , indicating the presence of ester functional groups.The 1 H NMR spectrum (    , and H-13 to C-12 were used to establish the connection of the ω-side chain and a ketone group located at C-15.An additional acetoxy group (2 ′′ -OAc) was assigned on C-12 due to its downfield-shifted chemical shift and considering the molecular formula.It is noted that the C=C double bond of the ω-side chain was located at C-13/C-14 in 1 instead of C-14/C-15 in clavulone I, which was rare in marine prostanoids.The 5Z and 13E geometries of two disubstituted double bonds were identified by their coupling constants (J 5-6 = 10.9Hz and J 13-14 =16.0 Hz).The 7E geometry was revealed by the de-shielding effect of H-7 caused by the anisotropy of the C-9 carbonyl group.Moreover, the presence of NOESY correlations of H-7/H-4 and H-5/H-6, and the absence of a NOESY correlation between H-13 and H-14 confirmed the above double-bond geometries (Figure 2).So far, all prostanoids isolated from Clavularia with an acetoxy substitution on C-12 were identified as being the S configuration.Thus, the configuration of C-12 in compound 1 was determined to be S by the biogenetic consideration.To verify the absolute configuration of C-4, the computer-generated ECD data (Figure S1) of two isomers (4R,12S-1 and 4S,12S-1) were used for comparison with the experimental data.Unfortunately, none of the calculated ECD curves matched the experimental trend.Thus, the DP4+ probability analysis was selected to determine the configuration of C-4.The conformational search of two isomers was executed by Spartan 20, and the lowest energies of each conformer were calculated using Gaussian 16.The 1 H and 13 C NMR data of major conformers were then averaged according to the Boltzmann distribution.As a result, the isomer 4R,12S-1 had a better correlation coefficient value (R 2 = 0.9989) than that of 4S,12S-1 (R 2 = 0.9986) in comparing the 13 C NMR chemical shifts with experimental data (Figure 3) (Tables S1-S6).Furthermore, the isomer 4R,12S-1 also had a higher probability (all data: 100.00%) in the DP4+ analysis (Table S7).Therefore, the configuration of 1 was determined accordingly.The 5Z and 13E geometries of two disubstituted double bonds were id their coupling constants (J5-6 = 10.9Hz and J13-14 =16.0 Hz).The 7E geometry w by the de-shielding effect of H-7 caused by the anisotropy of the C-9 carbo Moreover, the presence of NOESY correlations of H-7/H-4 and H-5/H-6, and t of a NOESY correlation between H-13 and H-14 confirmed the above double-b etries (Figure 2).So far, all prostanoids isolated from Clavularia with an acetox tion on C-12 were identified as being the S configuration.Thus, the configurat in compound 1 was determined to be S by the biogenetic consideration.To ve solute configuration of C-4, the computer-generated ECD data (Figure S1) of tw (4R,12S-1 and 4S,12S-1) were used for comparison with the experimental dat nately, none of the calculated ECD curves matched the experimental trend.Thu probability analysis was selected to determine the configuration of C-4.The tional search of two isomers was executed by Spartan 20, and the lowest energ conformer were calculated using Gaussian 16.The 1 H and 13 C NMR data of formers were then averaged according to the Boltzmann distribution.As a res mer 4R,12S-1 had a better correlation coefficient value (R 2 = 0.9989) than that (R 2 = 0.9986) in comparing the 13 C NMR chemical shifts with experimental dat (Tables S1-S6).Furthermore, the isomer 4R,12S-1 also had a higher probabilit 100.00%) in the DP4+ analysis (Table S7).Therefore, the configuration of 1 was d accordingly.The 5Z and 13E geometries of two disubstituted double bonds were identified by their coupling constants (J5-6 = 10.9Hz and J13-14 =16.0 Hz).The 7E geometry was revealed by the de-shielding effect of H-7 caused by the anisotropy of the C-9 carbonyl group.Moreover, the presence of NOESY correlations of H-7/H-4 and H-5/H-6, and the absence of a NOESY correlation between H-13 and H-14 confirmed the above double-bond geometries (Figure 2).So far, all prostanoids isolated from Clavularia with an acetoxy substitution on C-12 were identified as being the S configuration.Thus, the configuration of C-12 in compound 1 was determined to be S by the biogenetic consideration.To verify the absolute configuration of C-4, the computer-generated ECD data (Figure S1) of two isomers (4R,12S-1 and 4S,12S-1) were used for comparison with the experimental data.Unfortunately, none of the calculated ECD curves matched the experimental trend.Thus, the DP4+ probability analysis was selected to determine the configuration of C-4.The conformational search of two isomers was executed by Spartan 20, and the lowest energies of each conformer were calculated using Gaussian 16.The 1 H and 13 C NMR data of major conformers were then averaged according to the Boltzmann distribution.As a result, the isomer 4R,12S-1 had a better correlation coefficient value (R 2 = 0.9989) than that of 4S,12S-1 (R 2 = 0.9986) in comparing the 13 C NMR chemical shifts with experimental data (Figure 3) (Tables S1-S6).Furthermore, the isomer 4R,12S-1 also had a higher probability (all data: 100.00%) in the DP4+ analysis (Table S7).Therefore, the configuration of 1 was determined accordingly.).The IR spectrum indicated the presence of hydroxy (3435 cm −1 ) and ester (1717 cm −1 ) functionalities.The 1 H NMR spectrum (Table 1) showed six olefenic methines [δ H 5.57 (H-5), 5.85 (H-6), 5.94 (H-7), 6.17  13 C and DEPT spectra.The 1 H and 13 C NMR data were quite similar to those of 7-acetoxy-7,8dihydroiodovulone I (3) [7], implying that 2 is a congener of 3. The significant difference between them was the pattern of H-11 (doublet in 2 and singlet in 3), and an additional proton (H-10) was found in 2. The above findings suggested that the iodine atom in 3 was replaced by a hydrogen atom in 2. The COSY spectrum (Figure 4) demonstrated the proton sequences of H  4) from H-2 and 1-OMe to C-1 established the α-side chain, while the latter proton sequence constructed the ω-side chain of compound 2. The COSY cross-peaks of H-10/H-11, in combination with the HMBC correlations of H-8/C-9, H-10/C-8, C-9, C-12, H-11/C-12, and H-13/C-12, indicated the presence of a cyclopentenone moiety and its connection to αand ω-side chains.Finally, the acetoxy group attaching at C-4 and the hydroxy group attaching at C-12 were proved by the HMBC correlation from H-4/H-2 ′ to C-1 ′ and the chemical shift of C-12, respectively.The NOESY cross-peaks (Figure 4) of H-5/H-6 and H-14/H-15 supported the 5Z and 14Z geometries.The similar ECD trends (Figure 5) manifested between compounds 2 and 3 shared the same absolute configuration at those chiral centers.To reinforce the accuracy of structure determination, the 1 H and 13 C NMR data for 7S,8R,12R-2 and 7S,8R,12S-2 were calculated by quantum chemical calculations (Tables S8-S13).The DP4+ probability analysis showed isomer 7S,8R,12R-2 had a higher probability (Table S14).Accordingly, the structure of 2 was then determined.
(C-9)], and one non-protonated carbon [δC 79.5 (C-12)] were recogn DEPT spectra.The 1 H and 13 C NMR data were quite similar to those hydroiodovulone I (3) [7], implying that 2 is a congener of 3. The s between them was the pattern of H-11 (doublet in 2 and singlet in 3 proton (H-10) was found in 2. The above findings suggested that the i replaced by a hydrogen atom in 2. The COSY spectrum (Figure 4 4) of H-5/H-6 and H-14 5Z and 14Z geometries.The similar ECD trends (Figure 5) manife pounds 2 and 3 shared the same absolute configuration at those chiral the accuracy of structure determination, the 1 H and 13 C NMR data 7S,8R,12S-2 were calculated by quantum chemical calculations (Table probability analysis showed isomer 7S,8R,12R-2 had a higher probab cordingly, the structure of 2 was then determined.

Cytotoxic Evaluation of Isolated Compounds
Compounds 1, 3-9, and 11-13 were screened for in vitro cytotoxicity against the oral cancer cell line Ca9-22.All tested compounds demonstrated strong cytotoxicity toward the Ca9-22 cell line with IC 50 values less than 13 µM (Table 2).Among them, the major compound 3 displayed a significant cytotoxic effect with an IC 50 value of 3.96 µM.

The Predicted Possible Effects of Thirteen Compounds on Human Tumor Cell Lines for Cytotoxicity
We used the in silico tool CLC-Pred 2.0 to predict the possible cytotoxicity of test compounds for human tumor cell lines.Based on the chemical structure information of the test compound, the tool predicted three kinds of data [13]: the probability for being active (Pa), the probability for being inactive (Pi), and the invariant accuracy of prediction (IAP).Pa means the predicted chance that the test compound could belong to the sub-group of active compounds (the most typical structures of molecules in a sub-set of "actives" in the Prediction of Activity Spectra for Substances (PASS) training set).Pi means the predicted chance that the test compound could belong to the sub-group of inactive compounds (the most typical structures of molecules in a sub-set of "inactive" in the PASS training set).Only predictions with Pa > Pi could be considered possible for the particular test compound.According to the previous studies [13,14], we set the threshold as Pa > 0.7.Most of the compounds were predicted to show cytotoxicity for two human tumor cell lines, A2780cisR and PC-3, but only the in silico predictions of two compounds (2 and 8) were markedly different (Table S15).Compound 2 was predicted to only show cytotoxicity for A2780cisR.Compound 8 was predicted to show cytotoxicity for PC-3, HT-29, and A2780cisR.Additionally, IAP means the average accuracy of prediction obtained for the whole training set through PASS technology in the leave-one-out cross-validation method.The IAP values of compounds were located in a similar range (0.838-0.888).

The Predicted Possible Effects of Thirteen Compounds on NO Production in Macrophages
With the in silico tool InflamNat, we predicted the probability of the test compound with IC 50 (inhibition for NO production) <50 µM in macrophages (Table S16).We selected aminoguanidine [15] and apigenin [16] as the positive controls, predicting the probabilities as 0.657 and 0.621 by the tool InflamNat, respectively.Most of the predicted probabilities of compounds were located in a similar range (0.310-0.383), but only one predicted probability of compound 6 was 0.430.

The Predicted Pharmacological Potentials of Nine Compounds
Four compounds (6, 9, 12, and 13) were not suitable for the in silico tool Swis-sADME, and this kind of situation might be related to the stereochemistry of these compounds.The prediction results using the three models showed that all nine compounds might belong to the moderately soluble compounds or the water-soluble compounds (Table S17).For the pharmacokinetics of the compounds, we performed in silico predictions of gastrointestinal absorption, blood-brain barrier permeation, P-glycoprotein substrate, inhibitors for cytochrome P450, and skin permeation (Table S18).We also used the tool to predict the oral bioavailability of nine compounds (Figure S2).For the drug-likeness of the compounds, we evaluated six indicators, including Lipinski, Ghose, Veber, Egan, Muegge, and the bioavailability score (Table S19).For the medicinal chemistry of the compounds, we evaluated four indicators, including Pan Assay Interference Structures (PAINS), Brenk, lead likeness, and the synthetic accessibility score (Table S20).

The Structure and Activity Relationship of Isolated Prostanoids
Marine prostanoids are found in different genera of soft corals, such as C. viridis, Plexaura homomalla [17], and Telesto riisei [18].Those primitive nonmammalian prostanoids mainly possess antitumor activities.In our cytotoxic assay, although the C=C geometry at C-5 to C-8 and the oxidation/esterification at C-4, C-12, C-15, and C-20 might change the polarity of prostanoids, the cytotoxicity of isolates against oral cancer lines Ca9-22 remains.This finding implied that the skeleton of the prostanoids might demonstrate cytotoxic effects.Our discovery reinforces the cytotoxic potential of coral extracts and underscores the genus Clavularia as a promising source of bioactive marine natural products with anticancer properties.

The Predicted Possible Cytotoxicity of Thirteen Compounds for Human Tumor Cell Lines
Most of the compounds were predicted to show cytotoxicity for two tumor cell lines (A2780cisR and PC-3) (Table 1), but compound 2 was predicted to only show cytotoxicity for A2780cisR, and compound 8 was predicted to show cytotoxicity for PC-3, HT-29, and A2780cisR.A2780cisR is the cisplatin-resistant human ovarian carcinoma [19], PC-3 is the human prostate carcinoma [20], and HT-29 is the human colon adenocarcinoma [21].In addition, the in silico tool CLC-Pred 2.0 provides the prediction of possible cytotoxicity of the test compound against 391 human tumor cell lines [13], but the cell lines do not include the oral cancer cell line.The important role of inflammation is in regard to inducible nitric oxide synthase (iNOS), which yields the pro-inflammatory mediator nitric oxide (NO) from L-arginine [22], and the overexpressed iNOS could contribute to numerous diseases, including cancers [23].NO has been reported to participate in tumor progression and cancer metastasis, and some studies showed the potential of NOS inhibitors for anticancer therapies [24], such as the studies about triple-negative breast cancer [25] and colorectal cancer [26].In macrophages, a NOS inhibitor (aminoguanidine [15]) and a flavonoid derivative (apigenin [16]) were reported to have IC 50 values of 15.06 and 23 µM for inhibition of NO production, respectively.With the in silico tool InflamNat, the predicted probabilities of the positive controls (aminoguanidine and apigenin) with IC 50 (inhibition for NO production) <50 µM in macrophages were in a similar range (0.621-0.657), but most of the predicted probabilities of compounds were below 0.385 (Table S16).Because only one predicted probability of compound 6 was 0.430, we suggested that future investigators might prioritize examining the possible effect of compound 6 on NO production.

The Pharmacological Potentials of Nine Compounds
By reviewing the previous studies about the clinical trials (from 2010 to 2017), Sun et al. summarized the four possible reasons for about 90% of failures of clinical drug development: the poor clinical therapeutic effects (40-50%), the unexpected or unmanageable toxicity (30%), the poor drug-like properties (10-15%), and the low commercial needs or the weak strategic project (10%) [27].For evaluation of the possible pharmacological potentials of the compounds, we performed the in silico predictions of the properties of compounds, including water solubility (Table S17), pharmacokinetics (Table S18), oral bioavailability (Figure S2), drug-likeness (Table S19), and medicinal chemistry property (Table S20).The oral administration of poorly water-soluble molecules usually has two disadvantages [28]: (a) they need high doses for the therapeutic plasma concentrations; and (b) the molecules with slow absorption could lead to inadequate, variable bioavailability, or gastrointestinal mucosal toxicity.In the clinical drug development pipeline, the difficulties encountered by about 40-70% of molecules includes the poor aqueous solubility, absorption, and bioavailability [29].In this study, all nine compounds might belong to the moderately soluble compounds or the water-soluble compounds (Table S17).The gastrointestinal absorption of all nine compounds was predicted to be high (Table S18).The blood-brain barrier (BBB) restricts the access of certain compounds with therapeutic effects on the central nervous system from the peripheral system, and makes them useless [30].Only four compounds (4, 8, 10, and 11) might show blood-brain barrier permeation.Additionally, most of the predicted Log K p values for skin permeation of compounds were located in a similar range (−6.55 to −5.26), but only one predicted Log K p value of compound 1 was −7.04.For the predicted oral bioavailability, the colored zone in Figure S2 could be the suitable physicochemical space of the ideal compound, and six red spots showed the predicted values for the six properties of the test compound.Almost all of the five values of nine compounds were located in the ideal zone, but only one value (size) of compound 3 was not located in the ideal zone.And, only one value (flexibility) of nine compounds was not located in the ideal zone.There are two key roles for pharmacokinetics: P-glycoprotein and cytochrome P450 (CYP) enzymes.P-glycoprotein, the drug transporter, could remove some substances from cells (especially chemotherapeutic drugs) and participate in the development of the resistance of cancer cells to chemotherapeutic drugs [31].All nine compounds might be P-glycoprotein substrates (Table S16).CYP enzymes participate in mediating the metabolism for over 75% of marketed drugs [32], and five (1A2, 2C9, 2C19, 2D6, and 3A4) of the CYPs could contribute to about 95% of the metabolism of drugs via CYPs in humans [33].Most of the compounds were predicted to be inhibitors of CYP (Table S18).For the evaluation of four indicators of drug-likeness (Lipinski, Ghose, Egan, and Muegge), only two compounds (1 and 3) failed to pass the filters of one or two indicators (Table S19).All nine compounds did not pass the Veber filter.All nine compounds  S20).All nine compounds caused alerts for Brenk, and did not pass the filter of lead-likeness.Nevertheless, some scientists provided a reminder that the multifarious definitions of both drug-likeness and lead-likeness might severely reduce the possibility of molecular diversity for drug discovery [34].The clinical development of a natural compound requires a sustainable source compound to be available at a reasonable cost [35], especially marine natural compounds [36].The synthetic accessibility scores of all nine compounds were located in a similar range (4.58 to 5.41; Table S20), which might be at the moderately difficult level for the synthetic accessibility of the compounds.

General
The Varian (Palo Alto, CA, USA) Mercury Plus 400 MHz and VNMRS 600 MHz FT-NMR spectrometers provided the NMR spectra.HRSIMS was performed by using a Bruker APEX II spectrometer (Bremen, Germany).The V-650 spectrophotometer from Jasco (Tokyo, Japan) was utilized to measure the UV data.For measuring infrared data, the Jasco FT/IR-4X spectrophotometer was selected.Circular dichroism data were recorded using a Jasco J-815 CD spectrometer.A Jasco P-2000 polarimeter was used to quantify specific optical rotation.Merck KGaA (Darmstadt, Germany) silica gel 60 (0.015-0.040 mm) was utilized to pack columns.For high-performance liquid chromatography (HPLC), Phenomenex (Torrance, CA, USA) C18, phenyl-hexyl, and biphenyl columns were utilized.The LC-40D solvent delivery module, DGU-405 de-gassing unit, CBM-40 system controller, CTO-40S column oven, SPD-M40 photo diode array detector, and FRC-10A fraction collector made up the Shimadzu (Kyoto, Japan) HPLC apparatus.

Animal Material
The specimens of Clavularia spp.collected in Green Island, Taiwan, were identified by the corresponding author.The collection was stored at the Department of Marine Biotechnology and Resources, National Sun Yat-sen University, and a voucher ID (CI2021) was given.

Parameter of NMR and ECD Calculations
The NMR and ECD calculations were performed using Gaussian 16 software.Initially, potential conformers were generated using Spartan20 software.Conformers exhibiting a probability greater than 1% were optimized using Gaussian 16 at the B3LYP/6-31G(d) level.The electronic and thermal free energies of each conformer were summed, and the Boltzmann distribution ratio was computed.Theoretical calculations of NMR data were carried out using the Gauge-Including Atomic Orbitals (GIAO) method at mPW1PW91/ 6-311G(d,p) in chloroform.Subsequently, regression analyses were conducted to compare calculated and experimental NMR chemical shifts, with linear R 2 values calculated to assess the results.ECD spectra were computed using the parameters of B3LYP/6-311G++(d,p).

In Silico Prediction and Evaluation of Isolated Compounds
With the in silico tool CLC-Pred 2.0 [13] (Department for Bioinformatics, Laboratory for Structure-Function Based Drug Design, Institute of Biomedical Chemistry of the Russian Academy of Medical Sciences, Moscow, Russia), we predicted the possible cytotoxicity of test compounds for human tumor cell lines.By the in silico tool InflamNat [40] (Yunnan University, Kunming, China), we predicted the possible effects of test compounds on NO production in macrophages.The results could be shown as the probability of the test compound with IC 50 (inhibition for NO production) < 50 µM in macrophages without cytotoxicity.In addition, the tool InflamNat was mainly designed for natural products, so it may not be effectively applied to synthetic compounds.Using the in silico tool Swis-sADME [41] (Swiss Institute of Bioinformatics, Lausanne, Switzerland), we predicted the pharmacological potentials of test compounds, including (1) the water solubility (by three kinds of models), pharmacokinetics, (2) drug-likeness, (3) medicinal chemistry property, and (4) oral bioavailability.

Conclusions
The present investigation on octocoral Clavularia spp.resulted in the isolation and characterization of thirteen marine prostanoids, including two new ones.The absolute configuration of new components was determined by the comparison of ECD data and DP4+ probability analysis.In our continuously chromatographic study of Clavularia spp., three types of marine natural products (prostanoids, eudensamanes, and dollabellanes) were identified.Among those isolated compounds, prostanoids show the most potent cytotoxic activity against oral cancer cell line Ca9-22.Our findings lay a foundation for further developing or utilizing marine octocorals and prostanoids as anti-oral cancer agents.

Figure 3 .
Figure 3. (A) Linear correlations between experimental and calculated 13 C NMR data of two possible isomers.(B) Results of the DP4+ probability analysis.
bioavailability score (0.55).For the evaluation of one indicator of medicinal chemistry (Pan Assay Interference Structures (PAINS)), only two compounds (3 and 8) caused alerts (Table
a Positive control.