Comprehensive Characterization of Lignans from Forsythia viridissima by UHPLC-ESI-QTOF-MS, and Their NO Inhibitory Effects on RAW 264.7 Cells

Lignans are known to be an important class of phenylpropanoid secondary metabolites. In the course of our studies on the chemodiversity of lignans, the necessity arose to develop a method for the fast detection and identification of bioactive lignan subclasses. In this study, we detected 10 lignan derivatives of different extracts of F. viridissima by UHPLC-ESI-QTOF-MS. Lignan glycosides (1 and 2), lignans (3 and 4), and lignan dimers (5–10) were identified by analysis of their exact masses and MSe spectra along with the characteristic mass fragmentation patterns and molecular formulas. We further investigated NO inhibitory effects of F. viridissima fractions and their major lignan derivatives to evaluate those anti-inflammatory effects. The methylene chloride fraction of F. viridissima as well as compounds 8 and 10 showed potent dose-dependent NO inhibitory effects on RAW 264.7 cells. Corresponding to the NO inhibition by compounds 8 and 10, lipopolysaccharide (LPS)-induced inducible nitric oxide synthase (iNOS) expression was notably reduced by both compounds. Our combined data with the bioactive results and the component analysis by UHPLC-ESI-QTOF-MS suggest that the methylene chloride fraction of F. viridissima roots could be potential anti-inflammatory agents and these are related to major lignans including dimeric dibenzylbutyrolactone lignans.


Introduction
Lipopolysaccharide (LPS), the major cell wall component of gram-negative bacteria, induces inflammatory responses when administered to cells or animals. It induces the production of inflammatory mediators, including nitric oxide (NO), prostaglandin E 2 , and proinflammatory cytokines [1][2][3][4]. Although enhanced production of inflammatory mediators is important for host defense against external stimuli including LPS, excess production of inflammatory mediators causes severe inflammatory diseases, including septic shock, rheumatoid arthritis, systemic lupus erythematosus (SLE), and inflammatory bowel disease (IBD) [5][6][7]. Therefore, an agent that alleviates excess amounts of inflammatory mediators could be applied to treat various inflammatory diseases. Although various anti-inflammatory drugs, such as non-steroidal anti-inflammatory drugs (NSAIDs), have been

Results and Discussion
Dried roots of F. viridissima (2.7 kg) were extracted with 80% aqueous MeOH (3 times × 4 L, 90 min, 25 • C) by ultrasonication, and the crude extracts were diluted in H 2 O and partitioned successively with n-hexane, CH 2 Cl 2 , and n-BuOH. Compounds 1-10 for in vitro assay were isolated from CH 2 Cl 2 fraction of the roots of F. viridissima by using chromatographic methods including HPLC, MPLC over C 18 RP column as previously described [16]. The specific lignans of CH 2 Cl 2 and n-BuOH fractions were identified by UHPLC-ESI-QTOF-MS e analysis in both positive and negative ion modes. Lignan glycosides (1 and 2), lignans (3 and 4), and lignan dimers (5)(6)(7)(8)(9)(10) were identified in total, CH 2 Cl 2 , and n-BuOH fractions in both positive and negative ion modes. A total of 10 compounds was detected, and their structures are shown in Figure 1, and their retention time (Rt), error (ppm), molecular ions, and molecular formulas are shown in Table 1.
Molecules 2019, 24, x FOR PEER REVIEW 2 of 10 evaluation for the development of new anti-inflammatory drugs due to the severe adverse effects of NSAIDs [8].
'Yeon-kyo' is the fruit of Forsythia viridissima (Oleaceae) and F. suspensa, listed in the 11 th edition of the Korean Pharmacopoeia (KP11). Arctigenin is known as an indicator component for the quantification of F. viridissima. Lignan is one of the representative secondary metabolite class in F. viridissima. Other lignans, such as matairesinol, arctiin, and matairesinoside [9], have been reported as constituents of F. viridissima along with phenylethanoid glycosides, flavonoids, and triterpenoids [10,11]. Traditionally, 'Yeon-kyo' has been used for antiviral, anti-inflammatory, diuretic, antimicrobial, detoxification, and antipyretic activities. Among the different biological activities, the anti-inflammatory activity as a representative effect has been shown by numerous in vivo and in vitro studies with F. suspensa and F. viridissima [12,13]. Arctigenin, a phytochemical marker of F. viridissima [14], and other lignans [15] are also known to exert anti-inflammatory activities. Previously, we reported six new dimeric lignans and one new lignan glycoside along with nine known lignans from the roots of Forsythia viridissima [16]. Based on our previous research, this study aimed to investigate the profiles of the lignan dimers, lignans, and lignan glycosides with the help of a mass spectrometric technique and to assess the anti-inflammatory activities of isolated compounds of roots of F. viridissima.

Results and Discussion
Dried roots of F. viridissima (2.7 kg) were extracted with 80% aqueous MeOH (3 times × 4 L, 90 min, 25 °C) by ultrasonication, and the crude extracts were diluted in H2O and partitioned successively with n-hexane, CH2Cl2, and n-BuOH. Compounds 1-10 for in vitro assay were isolated from CH2Cl2 fraction of the roots of F. viridissima by using chromatographic methods including HPLC, MPLC over C18 RP column as previously described [16]. The specific lignans of CH2Cl2 and n-BuOH fractions were identified by UHPLC-ESI-QTOF-MS e analysis in both positive and negative ion modes. Lignan glycosides (1 and 2), lignans (3 and 4), and lignan dimers (5-10) were identified in total, CH2Cl2, and n-BuOH fractions in both positive and negative ion modes. A total of 10 compounds was detected, and their structures are shown in Figure 1, and their retention time (Rt), error (ppm), molecular ions, and molecular formulas are shown in Table 1.

HR-MS Characterization of Lignan and Lignan Glycosides
The lignans present in F. viridissima were identified by calculation of the molecular formula from exact mass and MS e spectra with the characteristic mass fragmentation patterns. The lignan glycosides, matairesinoside (1) and arctiin (2) were detected mainly in the n-BuOH fractions of F. viridissima ( Figure S1). Matairesinoside and arctiin (m/z, 543.1844 and 557.2019 [M + Na] + ) ( Table 1) Figure S5) was also observed in agreement with [17,18].
Both viridissimaol A (6) and B (7) detected in the dichloromethane fraction were lignan dimers with matairesinol and arctigenin moieties, and therefore, showed similar fragment ion peaks. Viridissimaol A (6), the protonated ion peak and the addition of the sodium ion were detected at m/z 729.2905 and 751.2716, respectively, and loss of one or two water molecules corresponded at m/z 711.2792 or 693.2686 ( Figure 2 and Figure   To elucidate whether F. viridissima roots exert anti-inflammatory effects through the inhibition of proinflammatory responses, we performed a cell viability assay and an NO assay using single components from F. viridissima roots in LPS-unstimulated and LPS-stimulated RAW 264.7 macrophage, a murine macrophage cell line which secretes inflammatory mediators through activation with TLR ligands such as LPS. Since the anti-inflammatory effects of the compounds are important when these effects are significant at cytotoxic concentration, we attempted to validate the non-cytotoxic concentrations of the compounds in RAW 264.7 macrophages. As shown in Figure 4, only compounds 3 (at 50 and 100 µ M) and 9 (at 100 µ M) showed significant cytotoxicity, whereas the other compounds did not show cytotoxicity at the highest dose of each compound. Based on the cell viability, the alleviation of NO production by the compounds in LPS-stimulated macrophages was estimated at non-cytotoxic concentrations. Of all compounds, compounds 3, 6, 7, 8, and 10 notably inhibited LPS-mediated production of NO in a dose-dependent manner. Among them, dibenzylbutyrolactone lignans, we selected two compounds which have slightly showed more inhibitory activity and have a different part of molecules. Compounds 8, which have 7, 8 unsaturated double-bond, and compound 10, which have C-O-C bond linkage in dimer structure, represent antiinflammatory potential among lignan subclass from the methylenechloride fraction of F. viridissima roots. To verify that the reduced production of NO by compounds 8 and 10 was due to the transcriptional regulation of iNOS, an enzyme responsible for the production of NO, the iNOS protein expression level was measured by immunoblotting. Corresponding to the NO inhibition by To elucidate whether F. viridissima roots exert anti-inflammatory effects through the inhibition of proinflammatory responses, we performed a cell viability assay and an NO assay using single components from F. viridissima roots in LPS-unstimulated and LPS-stimulated RAW 264.7 macrophage, a murine macrophage cell line which secretes inflammatory mediators through activation with TLR ligands such as LPS. Since the anti-inflammatory effects of the compounds are important when these effects are significant at cytotoxic concentration, we attempted to validate the non-cytotoxic concentrations of the compounds in RAW 264.7 macrophages. As shown in Figure 4, only compounds 3 (at 50 and 100 µM) and 9 (at 100 µM) showed significant cytotoxicity, whereas the other compounds did not show cytotoxicity at the highest dose of each compound. Based on the cell viability, the alleviation of NO production by the compounds in LPS-stimulated macrophages was estimated at non-cytotoxic concentrations. Of all compounds, compounds 3, 6, 7, 8, and 10 notably inhibited LPS-mediated production of NO in a dose-dependent manner. Among them, dibenzylbutyrolactone lignans, we selected two compounds which have slightly showed more inhibitory activity and have a different part of molecules. Compounds 8, which have 7, 8 unsaturated double-bond, and compound 10, which have C-O-C bond linkage in dimer structure, represent anti-inflammatory potential among lignan subclass from the methylenechloride fraction of F. viridissima roots. To verify that the reduced production of NO by compounds 8 and 10 was due to the transcriptional regulation of iNOS, an enzyme responsible for the production of NO, the iNOS protein expression level was measured by immunoblotting. Corresponding to the NO inhibition by compounds 8 and 10, the LPS-induced iNOS expression was notably reduced by these compounds ( Figure 5). These data prompted us to estimate the effects of the methylene chloride fraction on NO production in LPS-stimulated RAW 264.7 macrophages since compounds 8 and 10 were prepared in the methylene chloride fraction. As expected, the methylene chloride fraction notably inhibited the LPS-stimulated NO production without showing any cytotoxicity ( Figure 6).
LPS-stimulated NO production without showing any cytotoxicity ( Figure 6).
To date, to the best of our knowledge, there are few reports aimed at assessing the in vitro biological effects or identifying the complex bioactive lignans from the root parts of F. viridissima. In our present study, we narrow it down to study chemical analysis including dibenzylbutyrolactone dimer lignans by using UHPLC-ESI-QTOF-MS method, since the methylene chloride fraction of F. viridissima roots showed potent NO inhibitory effects in our experimental systems. Taken together, combined with the bioactive results and the component analysis by UHPLC-ESI-QTOF-MS, we suggest that the roots of F. viridissima-mediated anti-inflammatory effects are related to major lignans including dimeric dibenzylbutyrolactone lignans.  6, 7, 8, 9, and 10). The cell viability data were expressed as relative values to the untreated control group. NO levels were calculated according to a standard curve plotted using nitrite standard solution. Data represent the mean ± SD. * p < 0.05 relative to the control group (untreated group for cell viability and lipopolysaccharide (LPS)-treated group for nitrite level). Each number above graph is the compound number used for experiments. White, black, and gray bar represent Untreated, LPS-treated, and compound-treated groups, respectively. Experiments performed three times independently in three replicates for each experiment.  6, 7, 8, 9, and 10). The cell viability data were expressed as relative values to the untreated control group. NO levels were calculated according to a standard curve plotted using nitrite standard solution. Data represent the mean ± SD. * p < 0.05 relative to the control group (untreated group for cell viability and lipopolysaccharide (LPS)-treated group for nitrite level). Each number above graph is the compound number used for experiments. White, black, and gray bar represent Untreated, LPS-treated, and compound-treated groups, respectively. Experiments performed three times independently in three replicates for each experiment.

Plant Materials
Forsythia viridissima roots were collected in June 2015 from the Medical Herb Garden, College of Pharmacy, Seoul National University, Goyang-si, Gyeonggi-do, Korea. F. viridissima was identified by S. I. Han (Medical Herb Garden, College of Pharmacy, Seoul National University).

Chemicals and Reagents
All tested compounds were isolated from CH2Cl2 fraction of F. viridissima as described previously [16]. All isolates were lyophilized to remove any solvent that may be present. After dissolution in dimethyl sulfoxide (DMSO) for use in cell culture, it was diluted to a suitable concentration. LPS and DMSO were purchased from Sigma-Aldrich. LPS was dissolved in phosphate-buffered saline (PBS). Mouse anti-glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Mouse anti-iNOS was purchased from BD Biosciences (Franklin Lakes, NJ, USA). To date, to the best of our knowledge, there are few reports aimed at assessing the in vitro biological effects or identifying the complex bioactive lignans from the root parts of F. viridissima. In our present study, we narrow it down to study chemical analysis including dibenzylbutyrolactone dimer lignans by using UHPLC-ESI-QTOF-MS method, since the methylene chloride fraction of F. viridissima roots showed potent NO inhibitory effects in our experimental systems. Taken together, combined with the bioactive results and the component analysis by UHPLC-ESI-QTOF-MS, we suggest that the roots of F. viridissima-mediated anti-inflammatory effects are related to major lignans including dimeric dibenzylbutyrolactone lignans.

Plant Materials
Forsythia viridissima roots were collected in June 2015 from the Medical Herb Garden, College of Pharmacy, Seoul National University, Goyang-si, Gyeonggi-do, Korea. F. viridissima was identified by S. I. Han (Medical Herb Garden, College of Pharmacy, Seoul National University).

Chemicals and Reagents
All tested compounds were isolated from CH 2 Cl 2 fraction of F. viridissima as described previously [16]. All isolates were lyophilized to remove any solvent that may be present. After dissolution in dimethyl sulfoxide (DMSO) for use in cell culture, it was diluted to a suitable concentration. LPS and DMSO were purchased from Sigma-Aldrich. LPS was dissolved in phosphate-buffered saline (PBS). Mouse anti-glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Mouse anti-iNOS was purchased from BD Biosciences (Franklin Lakes, NJ, USA).

Chromatographic Profiling of the Compounds Present in F. viridissima Subfractions
The analysis of the dichloromethane subfractions was performed on a Waters Xevo G2 Q-TOF mass spectrometer (Waters MS Technologies, Manchester, UK), which was equipped with an electrospray ionization interface with Waters ACQUITY UHPLC system (Waters, Co., Milford, MA, USA). The UHPLC-MS data were obtained by MassLynx 4.1 software (Waters, UK). The separation of the compounds was carried out on ACQUITY UHPLC ® BEH C18 column (100 × 2.1 mm, 1.7 µm, Waters Co.). The mobile phase was composed of 0.1% (w/v) formic acid in water (solvent A) and acetonitrile

Cell Viability Assay
RAW 264.7 cells (6.0 × 10 4 cells/well) were seeded into a 96-well plate. After overnight incubation, cell culture medium was removed, various concentrations of compound diluents were applied to the cells, and then incubated for 24 h. Same volume of DMSO was treated to compound-untreated group to exclude the effect of DMSO on cell viability. After incubation, the cells were treated with EZ-Cytox solution (DAEIL lab, Seoul, Korea; 1/10 of the culture medium) for additional 1 h at 37 • C. Cell viability was determined by measuring the absorbance of the resulting formazan product at 450 nm using a VersaMax Microplate Reader (Molecular Devices, LLC, Silicon Valley, CA, USA).

Nitrite Assay
Nitrite assay is based on the fact that measurement of nitrite (NO 2 − ), which is one of two primary, stable and nonvolatile breakdown products of NO, reflects the quantity of NO in the supernatant [19]. RAW 264.7 cells (6.0 × 10 4 cells/well) were seeded into a 96-well plate and incubated overnight for cell adhesion. After adhesion, compounds were applied to the cells with indicated concentrations, LPS (1 µg/mL) was subsequently treated to the cells, and cells were then incubated for 24 h. After 100 µL cultured media was transferred to a new 96-well plate, 100 µL Griess reagent (a mixture of 1% sulfanilamide, 2.5% phosphoric acid (H 3 PO 4 ), and 0.1% N-(1-naphthyl) ethylenediamine in distilled water) was added to each well [19]. Sodium nitrite was serially diluted from 64 µM to 1 µM and a standard curve was generated by measuring absorbance after application of Griess reagent. The absorbance at 540 nm was determined by a VersaMax Microplate Reader.

Statistical Analysis and Experimental Replicates
Cell viability assay, nitrite assay, and immunoblotting were repeated three times. Each result was represented as mean ± standard deviation (SD). Differences between experimental conditions were assessed by Student's t-test, performed by Prism 3.0 (Graph Pad Software, San Diego, CA, USA), and * p < 0.05 was considered statistically significant.