Two new spongian diterpene derivatives from the aquaculture sponge Spongia officinalis Linnaeus, 1759

Abstract 2β,3α,19-Triacetoxy-17-hydroxyspongia-13(16),14-diene (1), a novel acetoxy diterpenoid, and 18-nor-2,17-hydroxyspongia-1,4,13(16),14-quaien-3-one (2), belonging to the rare 18-nor-spongian carbon skeleton, together with six related known metabolites (3‒8), were isolated from the aquaculture Spongia officinalis Linnaeus, 1759. Their structures were elucidated by comprehensive spectroscopic analysis, quantum chemical calculation of NMR parameters, and electronic circular dichroism (ECD). Compounds 3, 4, and 5 exhibited moderate inhibition against STAT3/NF-κB, HIF-1, Wnt signalling pathways. Compounds 1, 3, and 5 showed cytotoxicity activities against K562 cell line with IC50 values of 7.3, 3.5, and 6.4 μM, respectively. Graphical Abstract


Introduction
Marine sponges have been a rich source of natural products with high structural diversity and a broad spectrum of biological activities (Carroll et al. 2020). However, the difficulty in collecting wild sponge resources and the low content of active substances have severely restricted the pharmaceutical-related research of the sponge-derived active natural product. Therefore, large-scale domestication and cultivation of medicinal sponges (Ou et al. 2020) have considerable significance. There was only one report regarding the chemical constituents of aquaculture sponge in Xiamen, China, with only several known sterols isolated (Gang et al. 2018). As a result, intensive research on the chemical composition of aquaculture sponges is becoming increasingly important.
The relative configuration of compound 2 was determined by analysing the NOE spectrum. In the 1 D NOE difference experiment, the Me-19 proton at d H 1.26 was enhanced when the proton H-17a at d H 4.10 was irradiated. Correspondingly, irradiation of the proton H-17b at d H 3.68 enhanced the Me-19 at d H 1.26 and the H 2 -11a at d H 1.97. In the NOESY spectrum, the clear NOE cross peaks between H 2 -17 and H 2 -11, H 2 -17 and Me-19, H 2 -11 and Me-19, were observed. All data indicated the cis configuration of H 2 -11, Me-19, and H 2 -17. Given the above evidence, the relative configuration of compound 2 was determined as 8S Ã ,9R Ã ,10S Ã ( Figure S21).
The stereochemical analysis showed the relative configuration of compound 2 possessed four possible configurations due to the presence of three stereocenters (C-8, C-9, C-10). The 1 H and 13 C NMR chemical shifts were predicted using the GIAO method with chloroform as the solvent by DFT calculations at the B3LYP/6-31g(d) level of theory. The calculated 13 C NMR chemical shifts of (8S Ã ,9R Ã ,10S Ã )-2 showed the best agreement with the experimental values with the highest correlation coefficient (R 2 : 0.9987 and others: 0.9944-0.9973, Table S6 and Figure S3). Also, statistical analyses using the improved DP4þ method further confirmed that the most reliable structure was (8S Ã ,9R Ã ,10S Ã )-2 with a probability of 100% based on 1 H NMR, 100% based on 13 C NMR, and 100% on both carbon and proton chemical shifts (Tables S6 and S7).
All isolated compounds were evaluated for cytotoxicity against several human cancer lines K562, H69AR, ASPC-1, MDA-MB-231 in vitro, with adriamycin as the positive control. Among them, compounds 1, 3, and 5 showed cytotoxicity activities against K562 with IC 50 values 7.3, 3.5, and 6.4 lM, respectively. Compound 3 showed cytotoxicity activities against H69AR with IC 50 values 9.5 lM. Other compounds exhibited no cytotoxic activities at a concentration of 30 lM.
The signalling inhibition of HIF-1, Wnt, STAT3/NF-jB, and activation of PPARc, p53 of all isolated compounds were measured through high-throughput screening. At the same concentration of 20 lM, the inhibitory rate of 3 (88.5%) and 5 (85.2%) against Wnt signalling pathways were stronger than that of the positive control LF3 (70.9%). The inhibition rate of 3 (87.9%) and 5 (84.7%) against the HIF1 target was close to that of the positive control drug KC7F2 (96.0%). Other compound test results were shown in Table S10.

General experimental procedures
Optical rotation were measured on a Jasco P-1020 digital polarimeter. The UV spectrum were recorded on a Beckman DU640 spectrophotometer. CD spectrum were obtained on a Jasco J-810 spectropolarimeter. IR spectrum were taken on a Nicolet NEXUS 470 spectrophotometer in KBr discs. NMR spectra were measured by Agilent 600 MHz and JEOL JNMECP 600 spectrometers. The 7.26 ppm and 77.16 ppm resonances of CDCl 3 were used as internal references for 1 H and 13 C NMR spectra, respectively. HRESIMS spectrum were measured on Micromass Q-Tof Ultima GLOBAL GAA076LC mass spectrometers. Semi-preparative HPLC were performed using a Waters 1525 pump equipped with a 2998 photodiode array detector and a YMC C18 column (YMC, 10 Â 250 mm, 5 lm). Preparative HPLC were performed using a Shimadzu LC-20AR pump equipped with an LC-20A array detector and a SilGreen C18 column (SilGreen, 20 Â 250 mm, 5 lm). MCI gel CHP 20P (75-150 lm, Mitsubishi Chemical Industries, Tokyo, Japan), Silica gel [(200-300 mesh, 300-400 mesh silica gel G and silica gel H), Qingdao, China] were used for column chromatography, and precoated Silica gel plates (GF 254 , Qingdao, China) were used for TLC, and spots visualized by heating SiO 2 plates sprayed with 10% H 2 SO 4 in EtOH.

Animal material
In March, the season when the sponges began to breed and grow, sponge animal transplants were put into the intertidal waters (109 24 0 E, 19 52 0 N) of Shui Niupo Village, Danzhou City, Hainan Province, China. We use the ceramic tiles model as the attachment base for submerged farming ( Figure S2). The tile model is fixed by directly tying the sponge tissue block to the tile with a cable tie and then using a string to fix the tile on the main rope sinking on the seabed. When the tide is high, the water depth is 2-3 meters, and when the tide is low, the water depth is kept at 0.5 meters. The breeding scale is 50,000 square meters, and each sponge is cultivated at an interval of about 0.5 meters. After 18 months of culture, the sponge was collected and then frozen immediately in a -20 C refrigerator to prevent deterioration. The specimen was identified by Huilong, Ou, Department of aquaculture, College of Marine Sciences, Hainan University. The voucher specimen (No. HNYZ-2018-01) was deposited at State Key Laboratory of Marine Drugs, Ocean University of China, P. R. China.

Extraction and isolation
A frozen specimen of S. officinalis Linnaeus, 1759 (1.5 kg, dry weight) was homogenized and then extracted with CH 3 OH five times (2 days each time) at RT. The combined solutions were concentrated in vacuo and subsequently desalted by redissolving CH 3 OH to yield a residue (165 g). The extract was dispersed in 1.5 L of water and extracted five times with an equal volume of ethyl acetate. The ethyl acetate soluble portion was concentrated under reduced pressure to give a yellow residue (28 g), which was subjected to silica gel vacuum column chromatography eluted with a gradient of petroleum/acetone (v/v 200:1-1:1) and subsequently CH 2 Cl 2 /MeOH (v/v 20:1-1:1) to obtain nine fractions (Frs.1-Frs.9). Nine fractions were subjected to silica gel column chromatography, respectively, eluting with a gradient of petroleum/acetone (v/v 200:1-1:1) and subsequently CH 2 Cl 2 /MeOH (v/v 20:1-1:1) to obtain sub-fractions. Each fraction was detected by TLC to yield fifteen fractions ).

Cytotoxic activity assay
These compounds were tested against H69AR, ASPC-1, and MDA-MB-231 cell lines by using the standard method for the SRB proliferation assay, and K562 cell lines by using the standard method for the MTT proliferation assay. The adriamycin was a positive control. All the cell lines were purchased from the Shanghai Institute of Cell Biology (Shanghai, China). Cell cultures and detection methods were performed according to the procedures described in the previous article .

Signalling pathway activity assay
Positive control drugs LF3, KC7F2, SAR405838, Brevilin A, and rosiglitazone were purchased from Sigma company. The human colon cell lines (based on STAT3/NF-jB dual signalling A549, based on Wnt signalling HEK293S GNTI, based on HIF1 signalling MDA-MB-231, based on p53 signal HCT116 and based on the PPARc signalling Chang Liver) were all provided and preserved by Qingdao Marine Pharmaceutical and Biological Products Functional Laboratory. All five cloned cells have corresponding signal expression activity. The cloned cells constructed by the corresponding reporter system vectors are all constructed by referring to the literature for detailed construction methods (Sun J et al. 2020). In the biological screening process, the compounds with a fluorescence inhibition rate of more than 50% were selected for repeated detection. Simultaneously, to eliminate the fluorescence inhibition or activation caused by the cytotoxic effect of the compound itself, cell viability was measured using an MTT assay. The concentration of tested compounds was 20 lM. Compounds with a growth inhibition rate of more than 50% were selected and retested with a concentration of 5 lM. Each experiment has been repeatedly verified, and the significance of the data is compared by one-way analysis of variance. The relative cell survival rate or relative fluorescence activity is expressed as the ratio of the treatment group data and the control group data.

Computational details
The quantum chemical calculations were carried out by Gaussian 16 software (Frisch et al. 2016) using the density functional theory (DFT). In brief, the input geometries were built based on distance constraints from key correlations observed in the NOESY spectrum by the Chemdraw Pro 14.1 software with an MMFF94 force field, and then GFN2-xTB method conformer search were carried out by means of the xtb software (Grimme et al. 2017) using the molecular dynamics (MD) method. Conformational sampling procedures were conducted with xtb and Molclus program . The lowest energy conformers within 10 kcal/mol were subjected to further DFT calculations at the B3LYP-D3(BJ)/6-31g(d) level in the gas phase, and all minima displayed no imaginary frequencies by vibrational frequency analysis at the same level. Thermal corrections to Gibbs energies were obtained from frequency calculations at 298 K. Single-point energies were calculated at the (U)xB97X-D/def2-QZVPP level, in conjunction with the SMD solvation model of Truhlar and co-workers (Marenich et al. 2009) using water as the solvent by the ORCA program system. (Neese 2012) The population of each conformer was calculated by Boltzmann distribution based on Gibbs free energy with the Shermo (Lu and Chen 2020). Finally, a free energy correction þ7.91 kJ mol À1 was applied to all free energies to consider the conversion from the gas phase (1 atm) to the liquid phase (1 M). The structures were visualized with the VMD program (Humphrey et al. 1996).
Then, conformers with distributions higher than 5% were chosen for the TDDFT calculations at the PBE1PBE/TZVP level in methanol, using the SMD solvation model in agreement with the experiment condition including in all cases 60 excited states. The overall calculated ECD spectra were then generated by Boltzmann-weighting and simulated experimental spectra by the Multiwfn software (Lu and Chen 2012). Electronic transitions were expanded as Gaussian curves, with an FWHM (full width at half maximum) for each peak set to 0.4 eV.
GIAO calculations of NMR shielding constants with spin-spin interactions were accomplished for all stable conformers using the DFT method at the PCM/B3LYP/6-31g(d) level in chloroform, using the optimal method for DP4þ analysis (Grimblat et al. 2015). The shielding constants (including C and H) obtained were directly statistically analysed with experimental chemical shifts.