New Cytotoxic Natural Products from the Red Sea Sponge Stylissa carteri

Bioactivity-guided isolation supported by LC-HRESIMS metabolic profiling led to the isolation of two new compounds, a ceramide, stylissamide A (1), and a cerebroside, stylissoside A (2), from the methanol extract of the Red Sea sponge Stylissa carteri. Structure elucidation was achieved using spectroscopic techniques, including 1D and 2D NMR and HRMS. The bioactive extract’s metabolomic profiling showed the existence of various secondary metabolites, mainly oleanane-type saponins, phenolic diterpenes, and lupane triterpenes. The in vitro cytotoxic activity of the isolated compounds was tested against two human cancer cell lines, MCF-7 and HepG2. Both compounds, 1 and 2, displayed strong cytotoxicity against the MCF-7 cell line, with IC50 values at 21.1 ± 0.17 µM and 27.5 ± 0.18 µM, respectively. They likewise showed a promising activity against HepG2 with IC50 at 36.8 ± 0.16 µM for 1 and IC50 30.5 ± 0.23 µM for 2 compared to the standard drug cisplatin. Molecular docking experiments showed that 1 and 2 displayed high affinity to the SET protein and to inhibitor 2 of protein phosphatase 2A (I2PP2A), which could be a possible mechanism for their cytotoxic activity. This paper spreads light on the role of these metabolites in holding fouling organisms away from the outer surface of the sponge, and the potential use of these defensive molecules in the production of novel anticancer agents.


Introduction
Marine environments have proven to be an important source of unique chemical entities with a wide range of biological activities [1][2][3][4][5][6]. Owing to its biodiversity and seasonal variations in air and water temperatures, the Red Sea is one of the most important areas for marine research. Marine sponges are soft-bodied, sessile organisms which belong to the Porifera phylum. In both salt and freshwater ecosystems, more than 8000 species of sponges have been described. The sessile nature of marine sponges has led to the development of mechanisms for chemical defense to deter marine predators such as sharks, tortoises, and invertebrates [7]. Marine sponges are therefore an extremely rich source of secondary metabolites possessing various biological activities. Investigation of Red Sea marine sponges permitted detection of a wide range of secondary metabolites, including terpenes, alkaloids, sterols, steroidal glycosides, and ceramides [8][9][10]. Ceramides are bioactive lipids, which have been found in many marine invertebrate organisms. These compounds are involved in a number of physiological functions including apoptosis, arrest of cell growth, and cell-senescence [11]. They have also been reported to be precursors of complex sphingolipids (SLs). Ceramides possessing cytotoxic activity have previously been isolated from marine sponges [12]. The marine sponge Stylissa carteri is abundant in coastal Red Sea reefs, typically at depths between 5 and 15 m. Numerous secondary metabolites have already been isolated from S. carteri, including alkaloids [13], cyclic heptapeptides [14], and the pyrrole-2-aminoimidazoles stylissazoles A-C [15]. Alkaloids isolated from S. carteri were suggested to be prospective supports for human immunodeficiency virus (HIV) inhibition [16]. In addition, Stylissa carteri-associated bacteria were reported to have anti-plasmodial activity [17]. In this study, a bioactivity-guided fractionation was performed assisted by LC-HRESIMS investigation of the Red Sea sponge Stylissa carteri methanolic extract, which led to the isolation of two new compounds, stylissamide A (1) and stylissoside A (2). The potential cytotoxic activities of the isolated compounds are also reported, in addition to the investigation of a possible mechanism of cytotoxic activity through molecular docking simulation studies.
Compound 2 ( Figure 1) was obtained as a white amorphous powder, and its molecular formula was determined to be C46H91NNaO10 by HRESIMS at m/z 840.6547 [M + Na] + (calcd for m/z 840.6541), representing one degree of unsaturation (Supplementary Materials, Figure S11). The 1 H and 13 C NMR spectral data of compound 2 are listed in Table 1   Compound 2 ( Figure 1) was obtained as a white amorphous powder, and its molecular formula was determined to be C 46 H 91 NNaO 10 by HRESIMS at m/z 840.6547 [M + Na] + (calcd for m/z 840.6541), representing one degree of unsaturation (Supplementary Materials, Figure S11). The 1 H and 13 C NMR spectral data of compound 2 are listed in Table 1 4). In addition, there was a resonance at δ C 14.2 assigned for the two terminal methyl groups (C-19 and C-21 ) and at δ C 175.0 attributed to the amide carbonyl (C-1 ). The 13 C NMR spectrum revealed the presence of an anomeric carbon δ C 101.2 together with the characteristic signals at δ C 70.2, 71.6, 71.0, 73.1, and 62.6 indicating the presence of a sugar moiety. The structure of compound 2 was characterized by comparison of its 13 C NMR spectral data with those of the known cerebroside, agelasphin, possessing a 2-hydroxy fatty acid group [21,22]. The 1 H NMR signal at δ H 5.61 (d, J = 3.4 Hz) clearly indicated that the galactose had an α-linkage [21]. Analysis of the 1  In a similar way as for compound 1, the chain length of the fatty acid was determined based on the results of its methanolysis followed by detecting peaks from HRMS. The HRMS showed one molecular ion peak at 379.3188 [M + Na] + (calcd for C22H44NaO3: 379.3188) indicating the presence of a C21 fatty acid methyl ester for compound 2. As for compound 1, The cerebroside's relative configuration was suggested to be (2S, 3S, 4R, 2'R), since the optical rotation +17.40 (c 1.00, MeOH), In a similar way as for compound 1, the chain length of the fatty acid was determined based on the results of its methanolysis followed by detecting peaks from HRMS. The HRMS showed one molecular ion peak at 379.3188 [M + Na] + (calcd for C 22 H 44 NaO 3 : 379.3188) indicating the presence of a C 21 fatty acid methyl ester for compound 2. As for compound 1, The cerebroside's relative configuration was suggested to be (2S, 3S, 4R, 2 R), since the optical rotation +17.40 (c 1.00, MeOH), the afore mentioned 1 H NMR (H-2, H-3, H-4, H-2 ) and 13 C NMR signals (C-1, C-2, C-3, C-4, C-2 ) were in good agreement with those of phytosphingosine-type cerebroside molecular species possessing (2S,3S,4R,2 R)-configuration [23,24]. So, the full structure of 2 was determined to be (R)-N-[(2S,3S,4R)-3,4-dihydroxy-1-{[(2R,3R,4S,5S,6R)-3,4,5-trihydroxy-6-(hydroxymethyl)tetrahydro-2H-pyran-2-yl)oxy]nonadecan-2-yl}-2 -hydroxyhenicosanamide (stylissoside A), which, to the best of our knowledge, is a new compound.

Evaluation of the Antitumor Activity in Vitro
The potential cytotoxicity of compounds 1 and 2 isolated from Stylissa carteri was measured by the sulphorhodamine B (SRB) assay adopting the method of Skehan et al. [37] following the protocol described by Vichai and Kirtikara [38] on breast (MCF-7) and liver (HepG2) cancer cell lines. Color intensity was measured on an ELISA reader and the respective IC50 (concentration of

Evaluation of the Antitumor Activity In Vitro
The potential cytotoxicity of compounds 1 and 2 isolated from Stylissa carteri was measured by the sulphorhodamine B (SRB) assay adopting the method of Skehan et al. [37] following the protocol described by Vichai and Kirtikara [38] on breast (MCF-7) and liver (HepG2) cancer cell lines. Color intensity was measured on an ELISA reader and the respective IC 50 (concentration of the compound which reduces survival of cancer cells to 50%) values were calculated. As presented in Table 3, both compounds resulted in promising anticancer activities against breast (MCF-7) and hepatic (HEPG2) cancer cell lines. Compound 1 exhibited stronger cytotoxicity against MCF-7 with an IC 50 value of 21.1 ± 0.17 µM, while compound 2 displayed a slightly lower cytotoxicity, with an IC 50 value of 27.5 ± 0.18 µM. The case was reversed in HepG2 cancer cells, where compound 2 was more active (IC 50 30.5 ± 0.23 µM) than 1 (IC 50 36.8 ± 0.16 µM). The inhibitory properties of these compounds were compared with those of the standard drug cisplatin. Each data point represents the mean ± SD of four independent experiments (significant differences at p < 0.05).

In Silico Studies
Modeling and docking simulations of the newly isolated compounds (1 and 2) were performed using the Molecular Operating Environment (MOE) software [39] and the crystal structure of Su(var)3-9, enhancer of Zeste, Trithorax (SET)/inhibitor 2 of protein phosphatase 2A (I2PP2A) oncoprotein; (PDB: 2E50) [40]. D-erythro(e)-C 18 ceramide was chosen as a reference compound in the simulation studies due to its higher affinity for binding to SET protein compared to other endogenous ceramides or sphingosine [41]. PP2A has been reported to exhibit a tumor suppressive function by inducing apoptosis or programmed cell death in various cancerous cells [42]. Sphingolipid ceramide has long been described to activate PP2A through a direct interaction with the PP2AC catalytic domain [43], but its mechanistic details have so far remained unknown. Alternatively, another mechanism of ceramide-induced activation of PP2A has been recently found to proceed through the direct binding of ceramide with SET oncoprotein, which functions as an inhibitor of tumor suppressor PP2A [41]. Because the crystal structure of the ceramide-SET complex was not available in the protein data bank, induced fit docking was performed using the crystal structure of the SET protein and C 18 ceramide as a flexed ligand. The top generated pose (S = −4.5237 kcal/mol) of the initial flexible docking revealed that the C 18 ceramide chain of 1 interacts at the hydrophobic binding pocket along the space between the helix of the dimerization domain and the β sheets within the protein structure.
The pocket is predominately lipophilic with some regions of hydrophilicity ( Figure 5). Furthermore, our simulation results showed that the ceramide molecule adopts an "extended" conformation inside the SET protein, in which the sphingosine and lipid chains are placed in opposite directions. This finding was further confirmed by the NMR model suggested previously by De Palma et al. [41]. The docking results of compound 1 revealed that the ligand exhibited a similar extended orientation and comparable binding interactions within the pocket (S = −6.6821 kcal/mol). The lipid chain of compound 1 forms hydrophobic interactions with the amino acid residues Phe-68, Tyr-127, Pro-214, and Trp-213. Additionally, along the N-terminal region, the 1,4-dihydroxy array of the dihydrosphingosine chain donates two H-bonds to Glu-111 (2.843 and 2.849 Å, respectively), while the 3-hydroxy group accepts an H-bond from Gln-65 with 2.138 Å ( Figure 6A). Previous data had indicated that the amino acid residues 36-124 of the N-terminal region are involved in the inhibition of PP2A as well as the binding with ceramide [41]. On the other hand, compound 2 was found to exhibit a different compact conformation within the active site, where the dihydrosphingosine backbone runs perpendicular to both the fatty acyl chain and the sugar moiety. Therefore, compound 2 showed a weaker binding pattern within the pocket (S = −9.2672 kcal/mol) ( Figure 6B

General Experimental Procedures
1 H NMR (400 MHz), 13 C NMR (100 MHz), DEPT-135 and 2D NMR spectra were registered on a Varian AS 400 (Varian Inc., Palo Alto, CA, USA) using the residual solvent signal as an internal standard. High-resolution mass spectra were recorded using a Bruker BioApex (Bruker Corporation) machine. Pre-coated silica gel G-25 UV254 plates were used for thin layer

General Experimental Procedures
1 H NMR (400 MHz), 13 C NMR (100 MHz), DEPT-135 and 2D NMR spectra were registered on a Varian AS 400 (Varian Inc., Palo Alto, CA, USA) using the residual solvent signal as an internal standard. High-resolution mass spectra were recorded using a Bruker BioApex (Bruker Corporation) machine. Pre-coated silica gel G-25 UV254 plates were used for thin layer

Sponge Material
The sponge Stylissa carteri was collected from Sharm El Sheikh in the Egyptian Red Sea. The sponge material was air-dried and kept at low temperature (−24 • C) before further processing. Identification of the sponge was performed by Dr. Tarek Temraz, Marine Science Department, Faculty of Science, Suez Canal University, Ismailia, Egypt. Voucher specimens were deposited under registration number SAA-46 in the herbarium section of Pharmacognosy Department, Faculty of Pharmacy, Suez Canal University, Ismailia, Egypt.

Ceramide Hydrolysis
An aliquot of 5 mg of 1 and 2 were heated with 5% HCl/MeOH (0.5 mL) at 70 • C for 8 h. The mixture was extracted with n-hexane and concentrated in vacuo to afford the corresponding hydroxy fatty acid methyl esters. The HRMS of 1 showed a molecular ion peak at m/z 295.2249 [M + Na] + (calcd for C 16

Identification of the Sugar Moiety in Compound 2
Compound 2 (5 mg) was heated at 70 • C with 5% HCl/MeOH (0.5 mL) for 8 h then the reaction mixture (including the esterified fatty acid) was extracted with chloroform. The methanolic layer (containing the sugar moiety) was neutralized with Ag 2 CO 3 . The resulting precipitates were filtered off and the filtrate was concentrated in vacuo then injected on HPLC (Cosmosil Sugar-D, 4.6 ID × 250 mm, 1 mL/min, RI detector, 95% aqueous acetonitrile). HPLC runs of standard glucose and galactose were carried out in parallel and the respective retention times were compared with that of the unknown sugar in 2. Galactose showed a retention time of 4.76 min, similar to that of the sugar in compound 2, however, glucose displayed a retention time of 4.62 min, which was eluted earlier from the sugar in 2. Therefore, the sugar in the new compound 2 was identified as galactose.

Determination of the Configuration of the Sugar Moiety in 2
Compound 2 (2 mg) was hydrolyzed by heating in 0.5 M aqueous HCl (0.1 mL) and then neutralized with Amberlite IRA400. After drying in vacuo, the residue was dissolved in pyridine (0.1 mL) containing L-cysteine methyl ester hydrochloride (0.5 mg) and heated at 60 • C for 1 h in 0.1 mL solution of o-tolyl isothiocyanate (0.5 mg) in pyridine. The reaction mixture was directly analyzed by reversed-phase HPLC (Cosmosil-5-C 18 -AR-II, 4.6 ID × 250 mm, 0.8 mL/min, λ = 250 nm, 25% acetonitrile in 50 mM H 3 PO 4 ). The peaks at 19.58 min and 17.52 min coincided with those of the respective derivatives of D-galactose. The same derivatization procedures were performed on 5 mg of the standard compounds, D-galactose (t R = 18.6 min) and L-galactose (t R = 19.3 min).

Metabolomic Profiling
The metabolomics profiling was performed according to a method reported by Elsayed et al. [44]. These files were imported to the data mining software MZmine 2.10 for peak picking, deconvolution, deisotoping, alignment, and formula prediction. Dereplication of compounds was carried out by comparison with those in the Marinlit database and the Dictionary of Natural Products (DNP) 2015.

Cytotoxicity Assays
The cytotoxicity of compounds 1 and 2 was measured by the sulphorhodamine B (SRB) assay as described by Skehan [37], following the protocol described by Vichai and Kirtikara [38] on breast (MCF-7) and liver (HepG2) cancer cell lines. At an initial concentration of 3 × 10 3 cell/well in a 150 µL fresh medium, cells were seeded in 96-well microtiter plates, and left for 24 h to adhere to the plates. Different concentrations of the drug were added 0, 5, 12.5, 25, and 50 µg/mL respectively.
The plates were incubated for 48 h. The cells were fixed at 4 • C with 50 µL cold trichloroacetic acid (10% final concentration) for 1 h. The plates were washed with distilled water (automatic washer Tecan, Germany) and stained with 50 µL 0.4% SRB dissolved in 1% acetic acid at room temperature for 30 minutes. With 1% acetic acid, the plates were washed and air-dried. 100 µL/well of 10 M tris base (pH 10.5) was used to solubilize the dye. Using an ELISA microplate reader (Sunrise Tecan reader, Germany), the optical density of each well was measured at 570 nm spectrophotometrically. The mean background absorbance was automatically subtracted, and mean values were determined for each concentration of drugs. The experiment was repeated three times, then the IC 50 values were calculated.

In Silico Studies
All the molecular modeling calculations and docking simulation studies were performed using the Molecular Operating Environment (MOE 2014.0901, 2014; Chemical Computing Group, Canada) software. All MOE minimizations were performed until an RMSD gradient of 0.01 kcal/mol/Å with the force field (Amber10:ETH) and gas phase solvation to calculate the partial charges automatically. Before simulations, the protein was protonated using the LigX function and the monomer was identified. Induced fit docking simulation was performed initially to predict the active site using C 18 ceramide as a flexed ligand. Triangle matching with London dG scoring was chosen for initial placement, then the top 30 poses were refined using force field (Amber10:ETH) and Affinity DG scoring. The top pose from this simulation was analyzed and further used as a reference for the docking simulation of the newly isolated compounds 1 and 2. The output database dock file was created with different poses for each ligand and arranged according to the final score function (S), which is the score of the last stage that was not set to zero.

Conclusions
The present research highlighted the efficacy of LCMS profiling when combined with bioassay-guided drug discovery from marine invertebrates to accelerate the conventionally long processes of identifying an active metabolite by successive isolation from crude extracts. Dereplication experiments focused on the chemotaxonomic sorting helped in the identification of putative active metabolites, while structural assignment of the isolated compounds, using both HRMS and NMR, supported the identified hits. Metabolomic profiling of the bioactive extract displayed the existence of numerous secondary metabolites, mainly oleanane saponins and phenolic and lupane di-and triterpenes. Therefore, bio-guided isolation combined with LC-MS metabolomic profiling of the Red Sea sponge Stylissa carteri crude extract was performed, leading to the characterization of two new compounds, stylissamide A (1) and stylissoside A (2), which, to the best of our knowledge, have not been isolated from any natural source before. Both compounds, showed promising cytotoxic activity against breast (MCF-7) and liver (HepG2) cancer cell lines. Compound 1 exhibited stronger cytotoxicity against breast (MCF-7) cancer cells, while compound 2 exhibited higher activity against the HepG2 cancer cell line compared to cisplatin as the standard. Moreover, a docking study was performed to investigate the possible mechanism(s) of the cytotoxic activity of 1 and 2. The newly isolated metabolites were docked into the crystal structure of the SET oncoprotein and inhibitor 2 of protein phosphatase 2A (I2PP2A). We believe that the in vivo biological investigations of these metabolites will be of value for future anti-cancer drug development.