Diterpenoids from Blumea balsamifera and Their Anti-Inflammatory Activities

Six new diterpenoids, blusamiferoids A–F (1–6), including four pimarane-type diterpenoids, one rosane-type diterpenoid (3), and one rearranged abietane-type diterpenoid (6), were isolated from the dry aerial parts of Blumea balsamifera. Their structures were characterized by spectroscopic and computational methods. In particular, the structures of 1 and 4 were confirmed by X-ray crystallography. Compounds 5 and 6 were found to dose-dependently inhibit the production of TNF-α, IL-6, and nitrite oxide, and compound 5 also downregulated NF-κB phosphorylation in lipopolysaccharide (LPS)-induced RAW 264.7 cells.


Introduction
Blumea balsamifera (L.) DC. (Asteraceae), also named sambong, is a perennial herbaceous plant and a traditional herb, which is commonly found in Southeast Asia, such as China, Malaysia, Thailand, and the Philippines [1]. As a traditional medicine, the whole plants or leaves of B. balsamifera were widely used to treat a cough, urinary tract infection, gastric ulcer, headache, fever, rheumatism, and menstrual diseases [2,3]. In addition, B. balsamifera is an important plant source of L-borneol, which was designated as the only natural source of Aipian by the Pharmacopoeia of the PR of China [4]. The importance of B. balsamifera in traditional herbs has aroused widespread interest over the past decades. Previous phytochemical investigations revealed that B. balsamifera contains several types of chemicals such as volatile oils, flavonoids, and terpenoids [5]. Pharmacological research has disclosed that the whole plants, crude extracts, and isolated constituents of B. balsamifera contain several biological capacities such as wound healing [6], anti-cancer [7], anti-bacterial [8], anti-inflammatory [9], anti-oxidant [10], and anti-influenza virus activities [11].
As a common folk herb, B. balsamifera is often used to treat rheumatoid arthritis, dermatitis, and colds [12], indicating its anti-inflammatory activity. It was found that the volatile oil of B. balsamifera had a significant anti-inflammatory effect in inflammatory mice [9,13]. It has been reported that non-volatile components of B. balsamifera, such as the ethanol extract and the residue after extraction of the volatile oil, also have a certain inhibitory effect on inflammation, indicating that the non-volatile part of B. balsamifera still has its utilization value [14,15]. However, there are few related reports and it is necessary to further study this part. To further study the non-volatile components of B. balsamifera and their anti-inflammatory activities, we have carried out research on the ethyl acetate fraction of a 95% ethanol extract of B. balsamifera, resulting in the isolation of four new pimarane-type diterpenoids, one rosane-type diterpenoid, and a rearranged abietane-type diterpenoid.  H-5, H-9, and Hb-1 are on the other side. Likewise, ROESY correlations of H3-20/Hb-11, H3-17/Hb-11, H-15/Ha-11 suggest that H3-17 and H3-20 are on the same side. Thus, the relative configuration of 1 was defined. As for the absolute configuration of 1, it was assigned by X-ray diffraction analysis with CuKα radiation. The results show the absolute configuration of 1 as 4S,5R,9S,10R,13S with a calculated Flack parameter of −0.03 (5) (Figure 3). Hence, the structure of 1 was ultimately determined. Blusamiferoid B (2), obtained as white solids, has the molecular formula C20H26O3, as deduced from its HRESIMS, 13 C NMR, and DEPT spectra (eight degrees of unsaturation). The 1 H NMR spectrum of 2 (Table 1) indicates the presence of three methyl signals (δH  , three methines (one sp 2 and two sp 3 ), and seven nonprotonated carbons (including one keto-carbonyl, one ester carbonyl, and two olefinic carbons). Analysis of its 1 H and 13 C NMR data suggests that 2 belongs to a pimarane skeleton. Comparing the NMR data of dabeshanensin B [17] with those of 2 indicates that 2 might be an analogue of dabeshanensin B with a missing double bond at C-5 and C-6, which is confirmed by the 1 H-1 H COSY correlation of H-5/H-6 and HMBC correlations of H3-19/C-5, C-6, H3-20/C-5, H-5/C-4, C-10, C-18 (δC 180.6), and H-6 (δH 4.74)/C-10, C-7 (δC 192.1) ( Figure 2). Therefore, the planar structure of 2 was established. The relative configuration of 2 was determined by analysis of its ROESY spectrum ( Figure 2). The ROESY correlations of H3-19/H-5, H-6; H-5/H-6, Hb-3, indicate they are on the same side, while H3-20 is on the opposite side for the correlation between H3-20 and Ha-3. Meanwhile, the ROESY correlations of H3-20/Hb-11 and H3-17/Hb-11 indicate they are cofacial. Hence, the relative configuration of 2 was assigned. The absolute stereochemistry of 2 was further clarified by comparison of the experimental electronic circular dichroism (ECD) spectrum of 2 with the calculated spectra of (4S,5R,6S,10S,13S)-2 and  that H3-19 and H-7 are at the opposite orientation. On the basis of these results, the absolute configuration of 3 was subsequently assigned by direct ECD calculations of (4S,7R,9S,13R)-3 and (4R,7S,9R,13S)-3. It is evident that the ECD curve of (4S,7R,9S,13R)-3 matches well with the experimental curve (Figure 4), suggesting that the configuration of (4S,7R,9S,13R)-3 is more reasonable. As a result, the absolute configuration of 3 was clarified as 4S,7R,9S,13R.  Blusamiferoid B (2), obtained as white solids, has the molecular formula C 20 H 26 O 3, as deduced from its HRESIMS, 13 C NMR, and DEPT spectra (eight degrees of unsaturation  (Table 1) show 20 signals attributed to three methyls, seven methylenes (one sp 2 and six sp 3 ), three methines (one sp 2 and two sp 3 ), and seven nonprotonated carbons (including one keto-carbonyl, one ester carbonyl, and two olefinic carbons). Analysis of its 1 H and 13 C NMR data suggests that 2 belongs to a pimarane skeleton. Comparing the NMR data of dabeshanensin B [17] with those of 2 indicates that 2 might be an analogue of dabeshanensin B with a missing double bond at C-5 and C-6, which is confirmed by the 1 H-1 H COSY correlation of H-5/H-6 and HMBC correlations of H 3 -19/C-5, C-6, H 3 -20/C-5, H-5/C-4, C-10, C-18 (δ C 180.6), and H-6 (δ H 4.74)/C-10, C-7 (δ C 192.1) ( Figure 2). Therefore, the planar structure of 2 was established.

Figure 3.
Plot of X-ray crystallographic data for compounds 1 (left) and 4 (right). Displace ellipsoids are drawn at the 50% probability level. Blusamiferoid D (4), obtained as colorless small quadrate crystals, has the molecular formula C 19 H 28 O 4 , as deduced from its HRESIMS, 13 C NMR, and DEPT spectra (six degrees of unsaturation). The 1D NMR spectra of 4 exhibits a pattern analogous to that of 1. The differences between 1 and 4 are the presence of three additional hydroxy groups located at C-4 (δ C 74.4), C-5 (δ C 80.2), C-9 (δ C 74.0), and the absence of one carboxylic acid at C-4 in 4 on the basis of the HMBC correlations of H 3 -19/C-5, C-9, H 3 -18/C-4, C-5, and H-7/C-5, C-9. The relative configuration of 4 was assigned by ROESY evidence. The ROESY correlations (Figure 2) of H 3 -19/H 3 -18, Ha-11, Hb-1; 9-OH/Ha-1, and H 3 -17/Ha-11 are observed, indicating that three methyls are on the same side, while 9-OH is on the opposite side. Through analysis of the molecular model, we found that the ROESY correlation of H 3 -19/H 3 -18 can only be observed when H 3 -19 and 5-OH are on the opposite side, thus confirming the relative configuration of 5-OH. This conclusion was also secured by subsequent X-ray diffraction analysis using CuKα radiation, allowing us to assign the absolute configuration of 4 as 4R,5S,9R,10R,13S with a calculated Flack parameter of 0.01 (3) (Figure 3).
Blusamiferoid E (5) was isolated as yellowish gums. Its molecular formula was deduced as C 19 13 C NMR, and DEPT spectra (seven degrees of unsaturation). Through analysis of the 1D and 2D NMR data, it was noted that the presence of a double bond between C-4 and C-5 instead of two hydroxy groups in 5 are the main differences between 4 and 5. In addition to the chemical shifts of C-4 (δ C 149.4) and C-5 (δ C 133.0), the HMBC correlations of H 3 -18/C-4, C-5, H 3 -19/C-5 further confirmed the general structure of 5 ( Figure 1).
The relative configurations at the stereogenic centers in 5 were assigned by analysis of the ROESY spectrum (Figure 2), which shows correlations between H 3 -19/Hb-11, H 3 -17/Hb-11, H 3 -19/H 3 -17 (weak), indicating that H 3 -19, H 3 -17 are located on the same face. Through molecular model analysis, we found that the spacial interaction of H 3 -19/H 3 -17 can only be observed when H 3 -19 and 9-OH are on the opposite side. To confirm our conclusion from the molecular model study, NMR calculations to clarify the relative configuration at C-9 were carried out. The results disclose that 5 is likely the configuration of (9R,10S,13S)-5 based on the DP4+ probability analysis ( Figure S6) and the correlation coefficient (R 2 ) ( Figure S5). Thus, the relative configuration at C-9 was finalized. To assign the absolute configuration of 5, ECD calculations on (9R,10S,13S)-5 and (9S,10R,13R)-5 were conducted. The results show that the ECD spectrum of the former enantiomer agrees well with the experimental spectrum of 5 (Figure 4), showing the absolute configuration of 5 to be 9R,10S,13S.

Biological Evaluation
Based on the traditional medicinal properties of B. balsamifera, we investigated the anti-inflammatory effects of compounds 1-6. Following lipopolysaccharide (LPS) stimulation, we assessed the release of proinflammatory cytokines such as TNF-α and the generation of nitrite oxide pretreated with compounds to study their anti-inflammatory effects. According to the results of an ELISA assay, compounds 1, 3, 4, 5 and 6 significantly suppressed LPS-induced TNF-α secretion, at the same time, compounds 5 and 6 de-creased the production of nitrite oxide induced by bacterial LPS in RAW 264.7 cells (Figure 5A,B). Therefore, we selected compounds 5 and 6 for the follow-up study. Following that, we looked at the drug toxicity of compounds in RAW 264.7 cells. The CCK-8 assay displays that no obvious cytotoxicity of compounds 5 and 6 at 20 µM in RAW 264.7 cells ( Figure 5C). ELISA analysis shows that compounds 5 and 6 could dose-dependently inhibit LPS-induced TNF-α ( Figure 6A,B), IL-6 ( Figure 6C,D), and nitrite oxide generation ( Figure 6E,F). As we know, nuclear factor-κB (NF-κB) plays an important role in the transcriptional regulation of inflammatory cytokines and the development of inflammation. To further study its anti-inflammatory mechanism, we measured the effect of compounds 5 and 6 on the activation of the transcription factor NF-κB pathway. Western blot analysis confirms that compound 5 could dose-dependently down-regulate the expression of COX2 and p-NF-κB, and also significantly down-regulate the expression of iNOS in RAW 264.7 cells induced by LPS ( Figure 7A-D). Whereas, compound 6 could only dose-dependently reduce COX2 expression ( Figure 7E-H), indicating its biological difference from 5. Hence, compound 5 is considered to be a potent anti-inflammatory agent worthy for drug optimization.
tion. Compound 5 also significantly inhibited the production of nitric oxide, which may be due to the presence of 9-OH and Δ 4,5 . In addition, compound 6, a rearranged abietanetype diterpenoid, significantly reduced the production of nitric oxide, showing similar anti-inflammatory activity to the analogue jiadifenoic acid K, reported in the literature [20]. The results suggest that structural diversity leads to different anti-inflammatory activities.    According to the results of the anti-inflammatory activity, compounds 1, 3, 4, 5 and 6 were found to suppress the secretion of inflammatory factor TNF-α, while compounds 5 and 6 also decreased the production of nitric oxide induced by bacterial LPS in RAW 264.7 cells, showing anti-inflammatory activity. Chemically, compounds 1-5 possess a similar chemical skeleton, while compounds 1, 3, 4, and 5 all contain active hydrogen on oxygen atoms. Combined with the results of the anti-inflammatory activity, we speculated that active hydrogen on oxygen atoms may contribute to the reduction of TNF-α generation. Compound 5 also significantly inhibited the production of nitric oxide, which may be due to the presence of 9-OH and ∆ 4,5 . In addition, compound 6, a rearranged abietane-type diterpenoid, significantly reduced the production of nitric oxide, showing similar anti-inflammatory activity to the analogue jiadifenoic acid K, reported in the literature [20]. The results suggest that structural diversity leads to different anti-inflammatory activities.  (E,F) The production of nitrite oxide was measured using the Griess Kit. Data represent mean ± SEM values of three experiments. * p < 0.05, ** p < 0.01, *** p < 0.001 and **** p < 0.0001 compared with LPS alone. ### p < 0.001and #### p < 0.0001 compared with DMSO alone. Dexamethasone (DEX) was used as a positive control.

Plant Material
The dry aerial parts of B. balsamifera were purchased from Baoding Xiande Chinese Medicine Sales Co., Ltd., Guizhou province, China, in December 2019. The material was identified by Professor Bin Qiu at Yunnan University of Traditional Chinese Medicine, and a voucher specimen (CHYX0675) was deposited at the School of Pharmaceutical Sciences, Shenzhen University, China.

Extraction and Isolation
The dry aerial parts of B. balsamifera (50 kg) were soaked with 95% EtOH (300 L × 4 × 24 h) at room temperature. The 95% EtOH extracts were combined and evaporated under reduced pressure to afford a crude extract (2.4 kg), which was suspended in water and partitioned with EtOAc to gain an EtOAc soluble extract (1.6 kg). The EtOAc-soluble part was subjected to silica gel column chromatography, using a gradient of EtOAc in petroleum   Table 2.

Compound Characterization Data
level with the PCM in MeOH. For comparisons of the calculated curves and experimental CD spectra, the program SpecDis 1.62 was used.

NMR Calculations of 5
A conformational search and geometric optimization were adopted using the same method as the ECD calculations in the Gaussian 09 software package [21]. Gauge-Independent Atomic Orbital (GIAO) calculations of NMR chemical shifts were submitted in Gaussian 09 by density functional theory (DFT) with the level of B3LYP/6-31G(d,p) in chloroform with the PCM solvent model. The calculated NMR chemical shifts were analyzed by subtracting the isotopic shifts for TMS calculated with the same methods [22]. Regression analysis of calculated versus experimental 13 C NMR chemical shifts of 5 were carried out. Linear correlation coefficients (R 2 ), mean absolute error (MAE), and corrected mean absolute error (CMAE) were calculated for the evaluation of the results. After Boltzmann weighing of the predicted chemical shift of each isomers, the DP4+ parameters were calculated using the excel file provided by Ariel M. Sarotti [23]. RAW 264.7 cells were pretreated with compounds for 2 h and then stimulated with 1 µg/mL LPS for 12 h. The culture supernatants were collected and centrifuged from the treated cells. The concentrations of TNF-α and IL-6 were measured using the ELISA Kit (Proteintech, Chicago, IL, USA) according to the manufacturer's instructions. Dexamethasone was used as a positive control.

Determination of Nitrite Oxide
RAW 264.7 cells were treated with 1 µg/mL LPS with or without compounds for 24 h, the culture supernatants were collected and centrifuged. The production of nitrite oxide was measured using the Griess Kit (Beyotime, Shanghai, China) according to the manufacturer's instructions. In short, 50 µL of the cell supernatants were mixed with 50 µL Griess reagent I and II, then the absorbance at 560 nm wavelength was measured using a microplate reader (BioTek, Winooski, VT, USA). Dexamethasone was used as a positive control.

Western Blot
RAW 264.7 cells were incubated in different concentrations of compounds for 2 h and then exposed to 1 µg/mL LPS for 12 h. Total protein was extracted from cell lines after LPS treatment using a radioimmunoprecipitation assay (RIPA) buffer (Beyotime, Shanghai, China) including a protease and phosphatase inhibitor cocktail (Roche, Darmstadt, Germany), and protein samples were measured using the BCA assay (Thermo Scientific, Waltham, MA, USA). A 10% SDS-PAGE was used to separate equal quantities of protein extracts (15 µg), which were then transferred to PVDF membranes. The membranes were blocked with 5% BSA, then incubated overnight at 4 • C with the relevant antibodies, followed by a room temperature incubation with a horseradish peroxidase (HRP)-conjugated secondary antibody. The ECL kit (Bio-Rad, Hercules, CA, USA) and analysis system were used to view and quantify the bands (Bio-Rad, Hercules, CA, USA). ImageJ 1.51p software was used to perform densitometry analysis of the immunoblot findings (NIH, Bethesda, MD, USA).

Statistical Analysis
All of the experiments in this study were carried out in triplicate. The data was provided as a mean ± standard error of the mean (SEM). Graphpad Prism 6 (GraphPad Software, San Diego, CA, USA) and Excel (Microsoft) were used to conduct statistical analyses, which included a Student's t-test and a one-way ANOVA test. When * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001, and **** p ≤ 0.0001, differences were judged as significant.

Conclusions
In summary, this study on B. balsamifera afforded six new diterpenoids belonging to pimarane-, rosane-, and abietane-types, which will add new aspects for the chemical profile of B. balsamifera. Compounds 1, 3, 4, 5 and 6 could significantly inhibit LPS-induced TNF-α generation, showing their anti-inflammatory activity. The structure-activity relationship suggested that active hydrogen on oxygen atoms in compounds might be beneficial to inhibit the secretion of TNF-α, while the presence of 9-OH and ∆ 4,5 in compound 5 might contribute to reducing the production of nitrite oxides. In addition, compounds 5 and 6 could dose-dependently inhibit the production of TNF-α, IL-6 and nitrite oxides, and compound 5 significantly inhibits the phosphorylation of NF-κB in LPS-induced RAW 264.7 cells, suggesting that they play a potential role in inflammatory disorders. This finding indicates that the anti-inflammatory effect of B. balsamifera is not only related to volatile components, but also affected by other components in non-volatile parts, which is the result of the joint action of multiple components, and also provides the molecular template for the development of anti-inflammatory drugs.