The edible seaweed Laminaria japonica contains cholesterol analogues that inhibit lipid peroxidation and cyclooxygenase enzymes

In this study, 5 sterols were isolated and purified from Laminaria japonica, commonly known as edible brown seaweed, and their structures were identified based on detailed chemical methods and spectroscopic analyses. Spectroscopic analyses characterized 5 sterols as 29-Hydroperoxy-stigmasta-5,24(28)-dien-3β-ol, saringosterol (24-vinyl-cholest-5-ene-3β,24-diol), 24-methylenecholesterol, fucosterol (stigmasta-5,24-diene-3β-ol), and 24-Hydroperoxy-24-vinyl-cholesterol. The bioactivities of these sterols were tested using lipid peroxidation (LPO) and cyclooxygenase (COX-1 and -2) enzyme inhibitory assays. Fucosterol exhibited the highest COX-1 and -2 enzyme inhibitory activities at 59 and 47%, respectively. Saringosterol, 24-methylenecholesterol and fucosterol showed higher LPO inhibitory activity at >50% than the other compounds. In addition, the results of molecular docking revealed that the 5 sterols were located in different pocket of COX-1 and -2 and fucosterol with tetracyclic skeletons and olefin methine achieved the highest binding energy (-7.85 and -9.02 kcal/mol) through hydrophobic interactions and hydrogen bond. Our results confirm the presence of 5 sterols in L. japonica and its significant anti-inflammatory and antioxidant activity.

Oxidative stress can activate a variety of transcription factors, which lead to the differential expression of some genes involved in inflammatory pathways [14]. The oxidative stress machinery and inflammatory signaling are not only interrelated, but their impairment can lead to many chronic diseases. Previous phytochemical investigations revealed the hexane extract from L. japonica exerts anti-inflammatory effects [15]. Nevertheless, limited studies on the purification of hexane extracts from L. japonica and structure characterization of the pure compounds have been performed and, to the best of our knowledge, anti-inflammatory and antioxidant activities of these pure compounds have not been reported. Herein, 5 sterols from hexane extracts of L. japonica have been purified by VLC (vacuum liquid chromatography), multi-step MPLC (medium-pressure liquid chromatography), Prep-TLC (preparative thin-layer chromatography) and HPLC (high pressure liquid chromatography), and characterized by high-resolution electrospray ionization mass spectrometry (HRESI-TOFMS), nuclear magnetic resonance (NMR) and gas chromatography-mass spectrometry (GC-MS) analyses. Lipid peroxidation and cyclooxygenase enzyme inhibitory of these sterols were determined by using lipid peroxidation (LPO) and cyclooxygenase enzymes (COX-1 and -2) inhibitory assays according to our previous study [16][17][18][19][20][21]. Moreover, molecular docking analysis was performed to investigate the structure-activity relationships of these sterols.

Plant materials
L. japonica (GS-02-004-2013) used in this study were cultivated at Yantai (Shandong province, China) and harvested in July 2018. Fresh raw L. japonica was dried immediately after harvest under sun light and then shipped to the lab. The raw L. japonica in the lab was first washed with running tap water and deionized water, and then dried in an oven at 50˚C. The dry sample was further pulverized with a blender to obtain the dry L. japonica power through a 50-mesh screen.

Cyclooxygenase enzymes (COX-1 and -2) inhibitory assays
The COX-1 and-2 enzyme inhibitory activity of pure isolates (compounds 1-5) from L. japonica and cholesterol were measured by monitoring the initial rate of O 2 uptake using an Instech micro oxygen chamber and electrode (Instech Laboratories) attached to a YSI model 5300 biological oxygen monitor (Yellow Springs Instrument, Inc.) at 37˚C according to the published procedure. Commercial NSAIDs aspirin, ibuprofen, Celebrex 1 and naproxen were used as positive control groups to verify the sensitivity of the experiment [25-28].

Lipid peroxidation inhibitory assay
The antioxidant activity of all isolates (compounds 1-5) and cholesterol was determined by the LPO inhibitory assay using fluorescence spectroscopy on a Turner model 450 fluorometer (Barnstead/Thermolyne Corp.) according to the reported procedure. Commercial antioxidants BHA, BHT and TBHQ were used as positive control groups to verify the sensitivity of the experiment [17][18][19][20].

Molecular docking
Docking was performed to investigate the molecular interactions between the pure compounds and cyclooxygenase using the AutoDock Vina open-source program (ver. 1.1.2) according to the reported procedure [29,30]. The crystallographic structures of ovine COX-1 (PDB: 3N8Z) and human COX-2 (PDB: 5F1A) were retrieved from the Protein Data Bank in SDF format. The 3D conformations of compounds 2-4 (CID: 14161394, 92113, 5281328) were provided from PubChem (http://pubchem.ncbi.nlm.nih.gov/). The 3D conformations of compounds 1 and 5 were drawn at MolView (https://molview.org/). Virtual screening of all compounds was performed into the active site of the COX-1 and COX-2 domain by using AutoDock Vina. The whole protein structures were targeted for compounds docking [31].

Statistical analysis
The data was presented as mean ± SD (n = 3) and evaluated by one-way analysis of variance (ANOVA) followed by the Duncan's test. All statistical analyses were carried out using R statistical package and SPSS for Windows, Version 17.0 (SPSS Inc., Chicago, IL, USA). According to the spectral data, compound 1 was identified as 29-Hydroperoxy-stigmasta-5,24(28)-dien-3β-ol (Fig 1) [22]. Therefore, according to the spectral data, compound 4 was identified as fucosterol (stigmasta-5,24-diene-3β-ol) (Fig 1) [22,26] Compound 5 was also isolated as a white powder and exhibited a molecular ion peak at m/z 427.3562 [M-H 2 O+H] + in its positive-ion HRESITOFMS consistent with the molecular formula C 29

Cyclooxygenase enzymes (COX-1 and -2) inhibitory assay
Prostaglandins, the inflammation-causing hormones, were converted from arachidonic acid by the catalysis of COX enzymes. Therefore, inhibition of COX enzymes could prevent the production of prostaglandins and reduce inflammation [26,32]. The anti-inflammatory activity of the pure isolates from L. japonica was revealed by their COX-1 and -2 enzyme inhibitions. In this study, compounds (1-5) at 25 μg/mL inhibited COX-1 enzyme by 32, 19, 50, 59, and 26% and COX-2 enzyme by 23,14,33,47, and 9%, respectively (Fig 2). Commercial NSAIDs aspirin, Celebrex 1 , naproxen and ibuprofen were used as positive control at 108, 1, 12, and 15 μg/mL, respectively (Fig 2). Apparently, the COX enzyme inhibitory activity of 24-methylenecholesterol (compound 3) and fucosterol (compound 4) was comparable to the activity of the over-the-counter nonsteroidal anti-inflammatory drugs (NSAIDs) aspirin and ibuprofen. Compound 3 and 4 inhibited the COX-1 enzyme at a higher rate than the COX-2 enzyme, which is similar to aspirin and ibuprofen.

Lipid peroxidation inhibitory assay
Potential antioxidant activity of the pure compounds 1-5 was determined by the LPO assays using the large unilamellar vesicles (LUVs) model system, the results of which are shown in Fig 3. The LPO assay detects compounds that can scavenge free radicals. At 25 μg/mL, the LPO inhibitory activity of sterols 1-5 and cholesterol was 21, 51, 58, 56, 25 and 20%, respectively, as compared to the commercial antioxidants BHA, BHT and TBHQ at 85, 85 and 82%, respectively. The concentrations used to test LPO inhibitory activity for compounds were at 1.8, 2.2 and 1.6 μg/mL, respectively (Fig 3). Compared to compound 1 and 5, saringosterol (compound 2), 24-methylenecholesterol (compound 3) and fucosterol (compound 4) showed the higher LPO inhibitory activity at >50%, at 25 μg/mL concentration.

Molecular docking
Molecular docking was used to predict the interactions between the 5 sterols and the COX-1 and -2 enzyme at the molecular level. Root-Mean-Squared-Deviation (RMSD) is a similarity metric between molecular conformations, the lower the RMSD value is, the more similar the molecular conformations of ligands (before and after the docking) is. As shown in Table 1, the RMSD values of all the detected ligands were below the maximum allowed value of 2 Å, indicating ligands underwent minimal deformation during docking and the docking results were credible [33]. Docking results revealed that the binding energy of compounds 1-5 were -7.64, -6.36, -7.79, -7.85 and -6.44 kcal/mol for COX-1, and -8.49, -8.33, -8.72, -9.02 and -6.97 kcal/ mol for COX-2, which have the similar trends with their COX-1 and -2 enzyme inhibitory activities ( Table 1).
The possible mechanisms beneath the inhibition of compounds 1-5 against COX-1 and -2 was presented through molecular docking (Figs 4-6, Table 1). As shown in Fig 4, compounds  1-5 were predicted to be located in different pockets of COX-1 (pocket A, B and C), as reported in previous studies [34]. The hydrogen bonds and interaction residues were displayed in 3D docked poses of compounds 1-5 binding to COX-1 and COX-2 (Fig 6).  Table 1). Furthermore, the binding sites and interactions between compounds 1-5 and COX-2 are represented in Figs 5 and 6. Compounds 1-4 were located in the hydrophobic pocket A of   (Fig 6). Differ from compounds 1-4, compound 5 bound with COX-2 at another active pocket and showed possible hydrogen bonds with ARG-120 and GLU-524. Moreover, there was a high binding energy between compound 4 and COX-2 (-9.02 kcal/mol), which was comparable to that of the compound 5 (-6.97 kcal/mol).
COX-1 and -2 are bifunctional enzymes that convert arachidonic acid (AA) to prostaglandin G 2 (PGG 2 ) in their cyclooxygenase active site [35]. Our results suggested that fucosterol (compound 4) was located in the hydrophobic pocket of COX-1 and -2 enzyme with the highest binding energy (-7.85 and -9.02 kcal/mol) in molecular docking, compared to the other sterols. The possible mechanism underlying the highest COX-1 and -2 enzyme inhibitory activities of fucosterol might be attributed to the distinct olefin methine presented in its molecular structure, which could occupy the active site of the enzyme and prevent the enzyme from contacting with the substrate. This is in agreement with previous molecular docking results that tetracyclic skeletons and incorporation of an aliphatic chain in sterols were the key structural requirements theoretically for their good COX-1 and COX-2 binding activity [36]. Thus, these findings confirmed that fucosterol with tetracyclic skeletons and olefin methine achieved the good COX-1 and -2 enzyme inhibitory activities through hydrophobic interactions and hydrogen bond.

Conclusions
Purification of hexane extracts from L. japonica afforded 5 sterols, and their structures were identified using spectroscopic and chemical evidence. Spectroscopic methods characterized these compounds as 29-Hydroperoxy-stigmasta-5,24(28)-dien-3β-ol (compound 1), saringosterol (24-vinyl-cholest-5-ene-3β,24-diol) (compound 2), 24-methylenecholesterol (compound 3), fucosterol (stigmasta-5,24-diene-3β-ol) (compound 4), and 24-Hydroperoxy-24-vinyl-cholesterol (compound 5). These pure isolates were tested for antioxidant and anti-inflammatory activities using LPO and COX-1 and -2 enzyme inhibitory assays. Compared to the other compounds, fucosterol (compound 4) exhibited the highest COX-1 and -2 enzyme inhibitory activities at 59 and 47%, respectively. The COX enzyme inhibitory activity of 24-methylenecholesterol (compound 3) and fucosterol (compound 4) was comparable to the activity of the NSAIDs aspirin and ibuprofen. For the LPO assays, saringosterol (compound 2), 24-methylenecholesterol (compound 3) and fucosterol (compound 4) showed higher LPO inhibitory activity at >50% than the other compounds. In addition, the results of molecular docking predicted that 5 sterols were located in different pocket of COX-1 and -2 and confirmed that fucosterol with tetracyclic skeletons and olefin methine achieved the best COX-1 and -2 enzyme inhibitory activities through hydrophobic interactions and hydrogen bond. Our results confirm the presence of 5 sterols in L. japonica and its significant anti-inflammatory and antioxidant activity. Therefore, it appears that L. japonica have the potential to be developed as a dietary supplement, but molecular tests are required to verify the activity of 5 sterols in further study.