Three Pairs of Novel Enantiomeric 8-O-4′ Type Neolignans from Saussurea medusa and Their Anti-inflammatory Effects In Vitro

Three pairs of novel enantiomeric 8-O-4′ type neolignans (1a/1b–3a/3b), together with seven known analogues (4–10), were isolated from the whole plants of Saussurea medusa. Their structures were elucidated by extensive spectroscopic data analysis and electric circular dichroism (ECD) calculations after chiral separations. All compounds were obtained from S. medusa for the first time, and compounds 1–3 and 5–10 had never been obtained from the genus Saussurea previously. The anti-inflammatory activities of the compounds were evaluated by determining their inhibitory activities on the production of NO and inducible nitric oxide synthase (iNOS) expression in LPS-stimulated RAW 264.7 cells. Compounds (+)-1a, (−)-1b and 5–7 inhibited NO production and had IC50 values ranging from 14.3 ± 1.6 to 41.4 ± 3.1 μM. Compound 7 induced a dose-dependent reduction in the expression of iNOS in LPS-treated RAW 264.7 cells. Molecular docking experiments showed that all active compounds exhibited excellent docking scores (<−7.0 kcal/mol) with iNOS. Therefore, compounds (+)-1a, (−)-1b and 5–7 isolated from the whole plants of S. medusa may have therapeutic potential in inflammatory diseases.


Introduction
The Asteraceae plant, Saussurea medusa Maxim., is a rare perennial medicinal herb that grows in the northwestern part of China (e.g., Tibet, Xinjiang, Qinghai, Gansu, Yunnan and Sichuan provinces) at heights of 3500-5300 m [1]. S. medusa has been commonly used for the treatment of rheumatic arthritis, menoxenia, gynopathy, traumatic bleeding, anthrax, febrile tingling and headache [2,3]. Some phytochemical studies on this plant regarding flavonoids, lignans, terpenoids and chlorophyll constituents [1,[4][5][6] have been previously reported. These compounds exhibit an array of biological activities, such as anti-inflammatory [7], antitumor [8] and immunosuppressive [4] activities. In our previous study, a series of arylnaphthalene lignans were isolated from S. medusa, and some of them exhibited anti-inflammatory activity, which prompted us to continue researching this plant [9].
Enantiomers usually exist in the form of racemates or partial racemates in nature. Although high-performance liquid chromatography (HPLC), electrophoresis dominated chiral separation and other various separation techniques have been developed to separate enantiomers [11], separation still remains a challenge due to their similar chemical and physical properties in a chiral environment [12]. However, it is well known that enantiomers differ qualitatively or quantitatively in their pharmacological, toxicological, and biological activities [13]. Therefore, it is necessary to obtain optically pure compounds and evaluate their pharmacological effects.
Inflammation is a central feature of many pathophysiological conditions in response to tissue injury and host defenses against invading microbes [14]. The pathogenesis of many diseases, including cancer [15], diabetes [16,17], neurodegenerative [18], Parkinson's disease [19], cardiovascular [20,21] and other life-threatening diseases involve inflammation. Usually, the occurrence of inflammation is accompanied by the activation of various immune cells, especially macrophages, which participate in the initiation and spread of the inflammatory response by releasing proinflammatory cytokines and mediators such as endogenous radical nitric oxide (NO) [22]. NO is generated by NO synthases (NOSs) through the oxidation of L-arginine to L-citrulline. Two constitutive isoforms (cNOS) are detected in neuronal tissues (nNOS) and vascular endothelial cells (eNOS), whereas inducible NOS (iNOS) is expressed in various cell types (e.g., macrophages) upon inflammatory stimulation [23]. Inhibiting the activity of iNOS to block excessive NO production has been considered a promising strategy for the treatment of inflammatory diseases [24]. Therefore, the anti-inflammatory activities of the compounds were preliminarily evaluated in vitro by examining their ability to inhibit LPS-induced NO production and iNOS expression in RAW 264.7 macrophage-like cells.
In silico approaches such as molecular docking are considered one of the fundamental elements of drug design and discovery paradigms aimed at elucidating ligand-receptor interaction mechanisms and assisting lead optimization [25]. To preliminarily explore the anti-inflammatory mechanism, molecular docking experiments were performed to examine the interactions between the active compounds and the iNOS protein.

Anti-Inflammatory Effects
All isolated compounds were evaluated for cell viability in RAW 264.7 macrophag- Compound 1 showed weak CE in the ECD spectrum and immeasurable optical rotations, which implied that 1 was likely a racemic mixture. This prediction was confirmed by the presence of two peaks in chiral HPLC analysis using a Daicel IG column. Compounds (+)-1a and (−)-1b were successfully separated in a ratio of approximately 1:1 (see Figure S34). By comparing their calculated ECD and experimental ECD (Figure 3), the calculated ECD curve of the (8S) form matched well with the experimental ECD spectrum of (−)-1b ( Figure 3), which allowed the determination of the absolute configuration of (−)-1b as 8S. Thus, the almost symmetrical ECD curve of its enantiomer (+)-1a was assigned as 8R.
Medusidine B (2) was determined to be C 22 H 30 O 7 from the (+)-HRESIMS ion peak at m/z 429.1896 [M + Na] + (calcd for C 22 H 30 NaO 7 , 429.1884). The 1 H NMR and 13 C NMR spectra (Table 1) of 2 showed two sets of 1,2,4-trisubstituted aromatic rings, five methylene carbons (three oxygenated), two oxygenated methines, two methoxy groups and one upfield methyl group. The aforementioned NMR features suggested its structure to be closely related to that of 1, indicating that they were structural analogs. The differences were the absence of 5/5 -OMe and the existence of an ethyoxy unit at C-7 in 2. This was evident from the upfield chemical shifts of C-5 (δ C 115.6) and C-5 (δ C 119.4), as verified by the 1 H-1 H COSY correlations ( Figure 2) of H-5/H-6 and H-5 /H-6 . Furthermore, the presence of an ethyoxy group at C-7 in 2 was established based on the downfield chemical shifts of C-7 (δ C 81.8) in combination with the HMBC correlation ( Figure 2) from H-7 to C-1 , which was further verified by the 1 H-1 H COSY correlation of H 2 -1 /H 3 -2 .
In terms of the possible staggered conformers with intramolecular hydrogen bonding of the benzylic hydroxy and aryloxy groups, the large and small J values for H-7 and H-8 of 8-O-4 neolignan diastereoisomers correspond to the threo form and erythro form, respectively [27]. In the 1 H NMR spectrum of 2 in CD 3 OD, a large coupling constant (J 7, 8 = 6.4 Hz) was observed. Thus, the relative configuration of C-7 and C-8 was deduced to be in the threo form. Compound 2 showed weak Cotton effects in its ECD spectrum and immeasurable optical rotations, and its partially racemic nature was confirmed by chiral HPLC analysis in the same manner as that of 1. Compounds (−)-2a and (+)-2b were obtained with a variable enantiomeric excess (ee) value of approximately 40% for (+)-2b (see Figure S34). The negative CE at 239 nm in the ECD spectrum of (−)-2a indicated a 7R,8R configuration, while (+)-2b was deduced to be a 7S,8S configuration due to the positive CE at 239 nm [28,29]. Furthermore, ECD calculations also supported their absolute configurations.

Anti-Inflammatory Effects
All isolated compounds were evaluated for cell viability in RAW 264.7 macrophages. The results revealed that none of the compounds displayed cytotoxicity at the measured concentrations. Subsequently, all isolated constituents were screened for their inhibitory effects on NO production in LPS-induced RAW 264.7 macrophages. Quercetin was selected as a positive control with an IC50 value of 15.9 ± 1.2 μM. The results showed ( Table 2) that compound 7 exhibited obvious suppressive activity on the production of NO with an IC50 value of 14.3 ± 1.6 μM, which was comparable to that of the positive control quercetin. Compounds (+)-1a, (−)-1b, 5 and 6 displayed moderate inhibitory activities with IC50 values ranging from 18.5 ± 1.9 to 41.4 ± 3.1 μM.
As far as we know, the most commonly recorded lignans in S. medusa are dibenzylbutyrolactone and tetrahydrofuran lignans [1,3,4]. In our previous study, a series of Medusidine C (3) displayed a quasi-molecular ion peak at m/z 429.1887 (calcd for C 22 H 30 NaO 7 , 429.1884) in the (+)-HRESIMS analysis, corresponding to a molecular formula of C 22 H 30 O 7 . The IR, UV and NMR spectroscopic data of 3 were highly similar to those of 2, suggesting that the planar structure of 3 was the same as that of 2. The small coupling constant (J 7, 8 = 5.9 Hz) in the 1 H NMR spectrum of 3 suggested a relativeerythro configuration.
The ECD and optical rotation data indicated that 3 was also a pair of enantiomers. This deduction was supported by two peaks [ee = 38% for (+)-3a, see Figure S34] observed in the chiral HPLC analysis. Subsequently, chiral resolution was carried out to prepare optically pure (+)-3a and (−)-3b. Meanwhile, the positive CE at 239 nm of (+)-3a justified a 7R,8S configuration and the negative CE at 239 nm of (−)-3b justified a 7S,8R configuration. This was also supported by the ECD calculations.

Anti-Inflammatory Effects
All isolated compounds were evaluated for cell viability in RAW 264.7 macrophages. The results revealed that none of the compounds displayed cytotoxicity at the measured concentrations. Subsequently, all isolated constituents were screened for their inhibitory effects on NO production in LPS-induced RAW 264.7 macrophages. Quercetin was selected as a positive control with an IC 50 value of 15.9 ± 1.2 µM. The results showed ( Table 2) that compound 7 exhibited obvious suppressive activity on the production of NO with an IC 50 value of 14.3 ± 1.6 µM, which was comparable to that of the positive control quercetin. Compounds (+)-1a, (−)-1b, 5 and 6 displayed moderate inhibitory activities with IC 50 values ranging from 18.5 ± 1.9 to 41.4 ± 3.1 µM.
a Data are expressed as the mean ± SD (n = 3). b Positive control.
As far as we know, the most commonly recorded lignans in S. medusa are dibenzylbutyrolactone and tetrahydrofuran lignans [1,3,4]. In our previous study, a series of arylnaphthalene lignans with anti-inflammatory activities were isolated from S. medusa [9]. To date, however, 8-O-4 neolignans have not been reported in S. medusa. As a major class of lignans, 8-O-4 neolignans have been reported to have a wide range of bioactivities, especially in inflammatory responses [33][34][35]. In the present study, 8-O-4 neolignans (compounds 1a, 1b, 5 and 6) and sesquilignan (compound 7) isolated from the whole plants of S. medusa significantly inhibited NO production in LPS-stimulated Raw 264.7 macrophages. To our knowledge, this is the first time that the anti-inflammatory effects of compounds 1-3 and 5 have been reported. In addition, compounds 6 and 7, which had been isolated from the stems of Firmiana simplex [34] and the leaves and twigs of E. alatus [35], respectively, showed comparatively significant inhibitory effects as reported in the previous literature. The results of this study also further confirmed their anti-inflammatory activities.
Some preliminary structure-activity relationships could be drawn. The novel enantiomeric (+)-1a and (−)-1b exhibited similar anti-inflammatory activities. The absence of 5/5 -OMe and the presence of an ethyoxy group at C-7 weakened the inhibitory activities (compounds 2 and 3). The C7 -C8 double bond (compounds 5 and 6) was found to be essential for the observed inhibitory effects. The absence of the C7 -C8 double bond resulted in a loss of activity, as compound 4, which lacked this property, displayed poor inhibitory activity. Compound 7 exhibited high activity mostly due to its planar structure and stereoselectivity. The introduction of a hydroxy group in the ditetrahydrofuran ring led to a loss of activity (compound 10). Compounds 8 and 9 were inactive, likely due to the additional 5 /5 -OMe groups on aromatic rings.

Effect of the Selected Active Compounds on iNOS Expression
The NO inhibition results showed that compound 7 exhibited comparable activity to the positive control quercetin. Considering these results, compound 7, with a relatively high residual amount (11 mg), was selected for assessment of iNOS protein expression by western blot. As demonstrated in Figure 4, iNOS protein expression was significantly increased following stimulation with LPS, and compound 7 caused a dose-dependent reduction in expression of iNOS in LPS-treated RAW 264.7 cells. The results revealed that compound 7 inhibited the production of NO by reducing iNOS protein expression.

Molecular Docking Studies
To explore the possible mechanism of inhibiting NO production and iNOS protein expression, molecular docking studies were performed to investigate the interactions between active compounds and the iNOS protein. The bioactive compounds (+)-1a, (−)-1b, 5-7 and positive control quercetin were selected for molecular docking studies. Table 3 summarizes the binding affinity and binding interactions of active compounds with iNOS. Through careful analysis of the results of NO inhibition and molecular docking experiments, it was discovered that the actual NO inhibition effects of these active compounds corresponded well with the molecular docking results, except in the case of the positive control quercetin. Specifically, the orders of IC50 in the NO inhibition studies are as follows: compound 7 < quercetin < compound 6 < compound (−)-1b < compound (+)-1a < compound 5 ( Table 2). The orders of minimal binding energies in the case of molecular docking studies are as follows: compound 7 < compound 6 < compound (+)-1a < compound (−)-1b = compound 5 < quercetin (Table 3). Compound 7 exhibited the lowest (−9.4 kcal/mol) docking score with iNOS, consistent with its strongest NO inhibition effect. The novel enantiomeric (+)-1a and (−)-1b showed similar docking scores (−7.8 and −7.7 kcal/mol), which was also highly consistent with their similar moderate NO inhibition effects.
The results of the three-dimensional (3D) molecular docking images of the target compounds are shown in Figure 5. The visualization results showed that hydrogen bonds and hydrophobic interactions were formed between target compounds and crucial amino acid residues of the iNOS protein. The docking results (Table 3) showed that all active compounds were anchored to the catalytic site of the iNOS protein through various bonds and exhibited excellent docking scores (< −7.0 kcal/mol) with the iNOS protein.
Among them, compound 7 exhibited the strongest interactions with iNOS and had a free binding energy of −9.4 kcal/mol. The 3D diagram ( Figure 5E) illustrated that compound 7 interacted with amino acid residues TYR367, TYR483, LEU203, PHE363, TRP457 and TRP188 of iNOS. Two hydrogen bonds were found between the hydroxy groups on the two terminal benzene rings of compound 7 and amino acid residues TYR367 and TYR483. These hydrogen bonds strengthened the interactions between compound 7 and iNOS. In addition, we observed π-π interactions between the terminal benzene ring of

Molecular Docking Studies
To explore the possible mechanism of inhibiting NO production and iNOS protein expression, molecular docking studies were performed to investigate the interactions between active compounds and the iNOS protein. The bioactive compounds (+)-1a, (−)-1b, 5-7 and positive control quercetin were selected for molecular docking studies. Table 3 summarizes the binding affinity and binding interactions of active compounds with iNOS. Through careful analysis of the results of NO inhibition and molecular docking experiments, it was discovered that the actual NO inhibition effects of these active compounds corresponded well with the molecular docking results, except in the case of the positive control quercetin. Specifically, the orders of IC 50 in the NO inhibition studies are as follows: compound 7 < quercetin < compound 6 < compound (−)-1b < compound (+)-1a < compound 5 ( Table 2). The orders of minimal binding energies in the case of molecular docking studies are as follows: compound 7 < compound 6 < compound (+)-1a < compound (−)-1b = compound 5 < quercetin (Table 3). Compound 7 exhibited the lowest (−9.4 kcal/mol) docking score with iNOS, consistent with its strongest NO inhibition effect. The novel enantiomeric (+)-1a and (−)-1b showed similar docking scores (−7.8 and −7.7 kcal/mol), which was also highly consistent with their similar moderate NO inhibition effects. The results of the three-dimensional (3D) molecular docking images of the target compounds are shown in Figure 5. The visualization results showed that hydrogen bonds and hydrophobic interactions were formed between target compounds and crucial amino acid residues of the iNOS protein. The docking results (Table 3) showed that all active compounds were anchored to the catalytic site of the iNOS protein through various bonds and exhibited excellent docking scores (<−7.0 kcal/mol) with the iNOS protein. Among them, compound 7 exhibited the strongest interactions with iNOS and had a free binding energy of −9.4 kcal/mol. The 3D diagram ( Figure 5E) illustrated that compound 7 interacted with amino acid residues TYR367, TYR483, LEU203, PHE363, TRP457 and TRP188 of iNOS. Two hydrogen bonds were found between the hydroxy groups on the two terminal benzene rings of compound 7 and amino acid residues TYR367 and TYR483. These hydrogen bonds strengthened the interactions between compound 7 and iNOS. In addition, we observed π-π interactions between the terminal benzene ring of compound 7 and amino acid residue TRP188, suggesting that terminal benzene rings and hydroxy groups on them may play an important role in anti-inflammatory activity.
Int. J. Mol. Sci. 2021, 22, x FOR PEER REVIEW 9 of 14 compound 7 and amino acid residue TRP188, suggesting that terminal benzene rings and hydroxy groups on them may play an important role in anti-inflammatory activity.
The above results provided a potential explanation for the mechanism by which active compounds inhibit NO production and iNOS protein expression. Moreover, the results indicated that molecular docking could be an important tool in drug discovery to accelerate the recognition and investigation of novel drug candidates.

General Experimental Procedures
Optical rotations (Na lamp, 589 nm) were determined in MeOH on a Rudolph Autopol VI automatic polarimeter. UV spectra were measured on a Shimadzu UV-2550 The above results provided a potential explanation for the mechanism by which active compounds inhibit NO production and iNOS protein expression. Moreover, the results indicated that molecular docking could be an important tool in drug discovery to accelerate the recognition and investigation of novel drug candidates.

Plant Material
The whole plants of S. medusa were collected from Yeniu Ditch (altitude 4100 m), Qilian County, Xining City, Qinghai Province, China, in August 2018 and were authenticated by Professor Lijuan Mei from Northwest Institute of Plateau Biology. A voucher specimen (access number: 0341202) was deposited at the Key Laboratory of Tibetan Medicine of the Chinese Academy of Sciences.

ECD Calculations
According to relevant literatures and program packages [36][37][38], the absolute configurations of 1a/1b-3a/3b were determined by TDDFT-ECD calculations. For detailed calculation method, see Experimental Section of Supporting Information.

Determination of NO Production
Measurements of NO production in an activated macrophage-like cell line were performed as described previously [14]. Briefly, RAW 264.7 cells (purchased from Procell Life Science & Technology Co. Ltd., Wuhan, China) were cultured in plastic dishes in Dulbecco's Modified Eagle Medium (DMEM) containing 10% fetal bovine serum (FBS) and supplemented with pyruvate (1.0 mM), glutamine (2.0 mM), streptomycin (10.0 µg/mL) and penicillin (100.0 U/mL). The cell lines were maintained at 37 • C in a humidified atmosphere with 5% CO 2 . The cells were treated with or without LPS (1.0 µg/mL) for 24 h in the presence or absence of the test compounds (3.125, 6.25, 12.5, 25.0 and 50.0 µM). Absorbance was measured at 540 nm after incubating culture supernatant (100 µL/well) with Griess reagent (100 µL) (Sigma-Aldrich, St. Louis, MO, USA) at room temperature, and the absorption coefficient was calibrated using a NaNO 2 solution standard. Cell viability was measured using the MTT-based colorimetric assay according to a previous report [39].

Determination of iNOS Expression
As previously described in the literature [40], the treated cells were washed with PBS and suspended in lysis buffer. Cell debris was then removed after centrifugation. After the protein concentration was determined with BCA reagent, suspensions were boiled in SDS-PAGE loading buffer. The proteins were subjected to gel electrophoresis and electrophoretically transferred onto PVDF membranes (Millipore). The blot was incubated for 2 h with blocking solution at room temperature. After being washed, the membranes were incubated with a 1:1000 dilution of monoclonal anti-iNOS antibody and a 1:5000 dilution of β-actin antibody overnight at 4 • C. Blots were then washed three times with TBST and incubated with a 1:3000 dilution of secondary antibody solution for 1 h at room temperature. Blots were again washed three times with TBST and then detected by using enhanced chemiluminescence reagent and exposed to photographic films. Images were collected, and the related bands were quantitated by densitometric analysis using Gel-Pro analyzer software.

Molecular Docking Study
The specific docking methods and parameters can be seen in our previously published article [9]. Briefly, chemical structures of target compounds were drawn using ChemDraw 14.0 and converted to 3D coordinates in Chem3D. Each of them was subjected to energy minimization by the MM2 method and saved in "pdb" format. The 3D coordinates of the crystal structure of iNOS (PDB ID: 3E6T) were obtained from the RCSB Protein Data Bank (https://www.rcsb.org/pdb, accessed on 6 September 2022) [41] and handled in the Biovia Discovery Studio Visualizer 2020 program to check any missing residues/atoms and delete co-crystallized molecules such as cofactors, inhibitors, and water. The proteins and ligands were processed and converted to "pdbqt" format. A grid box with dimensions of 30, 30 and 30 points in x, y and z directions was built. Molecular docking was performed using AutoDock Vina 1.1.2 with default parameters, and the binding sites were defined within 10 Å around the co-crystallized ligands. Each docking involved nine independent runs. The docked model with the lowest docking energy was selected to represent its most favorable binding pattern.

Conclusions
In summary, three pairs of novel enantiomeric 8-O-4 type neolignans (1a/1b-3a/3b), together with seven known analogues (4-10), were isolated from the whole plants of S. medusa. Their structures were established by spectroscopic data and ECD calculations. Compounds (+)-1a, (−)-1b and 5-7 displayed inhibitory activities on NO production with IC 50 values ranging from 14.3 ± 1.6 to 41.4 ± 3.1 µM. Further iNOS protein expression studies demonstrated that compound 7 induced a dose-dependent reduction in iNOS protein expression. According to molecular docking studies, strong interactions were observed between active compounds and key residues of iNOS. Thus, a preliminary mechanism of inhibiting NO production and iNOS protein expression has been revealed. Overall, these findings not only extended the structural diversity of lignans in S. medusa, but also provided scientific background for the development of S. medusa as a potential medicinal plant to treat inflammatory diseases.