Unusual Secondary Metabolites of the Aerial Parts of Dionysia diapensifolia Bioss. (Primulaceae) and Their Anti-Inflammatory Activity.

The genus Dionysia, belonging to the Primulaceae family, encompasses more than 50 species worldwide with a center of diversity located in the arid Irano-Turanian mountains. In this study, a phytochemical investigation of the aerial parts of D. diapensifolia Bioss. led to the isolation of 24 phenolic compounds 1–7 and 9–25, and one sesquiterpenoid 8. Compound 1 was identified as new natural product, while isolation of 2 and 3, already known as synthetic products, from a natural source is reported for the first time in the present study. Isolation of compound 8 from a Dionysia species and indeed the whole Primulaceae family is reported for the first time too. Structure elucidation was performed by extensive spectroscopic analyses (1D-, 2D-NMR, and MS), and by comparison with reported literature data. Furthermore, DP4+ chemical shift probability calculations were performed to establish the relative configuration of compound 1. Additionally, subfractions obtained by liquid-liquid extraction of the methanolic extract of the plant, and subsequently the isolated new and selected known compounds 1–4, 6, 8–11 obtained from the diethyl ether subfraction were investigated for their inhibitory effect on NO release and iNOS and COX-2 expression in J774A.1 murine macrophages. The results showed a potential anti-inflammatory activity of the obtained subfractions, of which the diethyl ether subfraction was the most active one in inhibiting NO release and COX-2 expression (p < 0.001). Among the investigated isolated compounds, compound 4 significantly (p < 0.001) inhibited NO release and iNOS and COX-2 expression in a comparable manner like the used positive controls (L-NAME and indomethacin, respectively). Moreover, other isolated substances displayed moderate to high inhibitory activities, illustrating the potential anti-inflammatory activity of Dionysia diapensifolia.


Introduction
Inflammation is a pathophysiological process, identified by redness, edema, pain and other symptoms, which plays a pivotal role in protecting the body from injury and infection [1]. However, prolonged inflammatory responses are the most important cause of chronic inflammatory diseases

General Experimental Procedures
UV/VIS and ECD spectra measurements were performed at a J-8000 spectropolarimeter (JASCO, Tokio, Japan). Optical rotations were measured at a JASCO P-2000 polarimeter. IR spectra were recorded on an ALPHA FT-IR apparatus (Bruker, Ettlingen, Germany) equipped with a Platinum ATR module. One-and two-dimensional NMR experiments were performed at a Bruker Avance II 600 spectrometer operating at 600. 19 MHz ( 1 H) and 150.92 MHz ( 13 C) at 300 K, with deuterated chloroform (chloroform-d), acetone (acetone-d6), methanol (methanol-d4) or DMSO (DMSO-d6) as solvents, containing 0.03% TMS. These solvents were purchased from Euriso-top SAS (Saint-Aubin Cedex, France). The chemical shifts were recorded as δ values referenced to the TMS signal. HPLC analysis was carried out at an LC-20AD XR system (Shimadzu, Düsseldorf, Germany) equipped with DAD detector, auto-sampler, and column thermostat. LC-and LC-ESI-MS parameters (general method): stationary phase: Phenomenex Aqua C18 5 µm, 250 × 4.6 mm; mobile phase: A = H2O + 0.02% TFA, B = acetonitrile; gradient: 0 min: B = 2%; 20 min: B = 50%; 40 min: B = 98%; 50 min B = 98%; T: 35 °C; flow: 1 mL/min; sample concentration and injection volume: 2-8 mg/mL, 5-15 µL in ACN, THF or MeOH. HPLC-ESIMS experiments were carried out at a HP1100 system (Agilent, Waldbronn, Germany) hyphenated to an Esquire 3000 plus ion trap (Bruker Daltonics, Bremen, Germany), using electrospray ionization (ESI) in alternating mode and using the following MS parameters: 1:5 split The HMBC spectrum showed a correlation from H-5 (δH 6.74) to a carbonyl group at δC 187.5 and from H-6 (δH 7.00) to C-4a (δC 121.6), resulting in the identification of a carbonyl group at position C-4 ( Figure 2). Furthermore, a polar aliphatic methine group (H-3) at δH 4.22 displayed a correlation to C-4, C-2 and C-1′, which later established the position of a phenyl group at C-2. The HMBC correlation from H-2′ to C-2 and the lack of correlation from H-2′ to C-3, confirmed the position of C-2 and C-3. Due to the down field shift of the resonances of C-3 and C-2 and considering the molecular formula, the remaining coordination sites of the mentioned carbons (C-2 and C-3) were filled by an oxygen-bridge, introducing two chiral centers at C-2 and C-3. However, optical rotation measurement resulted in "zero rotation value" for this molecule, suggesting the presence of a racemate. As the NOESY spectrum measured for 1 was not conclusive to determine its relative configuration, DP4+ calculation was performed. The results indicated the presence of a Z-epoxy ring with probability of greater than 99.9% as correct isomer ( Figure S24, Supplementary Materials). Therefore, the structure of 1 was thus deduced as a racemate of 2R,3S-and 2S,3R-epoxyflavone.

Plant Material and Chemicals
Dionysia diapensifolia aerial parts were collected at Arsanjan, Islam Abad village, Shiraz, Fars province, Iran. The plant material was dried at room temperature in the dark and stored in paper bags until extraction. Identification of plant material was performed by Asso. Prof. Dr. Hossein Batooli (Kashan Botanical Garden) and a voucher specimen (KBG: 3496) has been deposited in the Research Institute of Forests and Rangelands; Agricultural Research, Education and Extension Organization (AREEO, Tehran, Iran).

Measurement of NO Release
NO levels were measured as nitrite NO2 − , as index of NO released by J774A.1 macrophages culture medium 24 h after LPS treatment by Griess reaction, as previously reported [14]. In brief, after adhesion, the cellular medium of J774A.1 macrophages (5 × 10 4 /well; 96-well plate) was replaced with fresh medium alone or containing serial dilutions of the methanolic extract and subfractions of D. diapensifolia (5-0.5 µg/mL) or isolated compounds (50-5 µM) and incubated for 1 h and then co-exposed to LPS (1 µg/mL) for further 24 h. Thereafter, 100 µl of cell culture medium were mixed with 100 µl of Griess reagent (equal volumes of 1% (w/v) sulphanilamide in 5% (v/v) phosphoric acid and 0.1% (w/v) N-(1-napthyl)ethylenediamine dihydrochloride in water and incubated at room temperature for 10 min. Subsequently, the absorbance was measured at 550 nm in a microplate reader Titertek (Dasit, Cornaredo, Milan, Italy). The amount of NO 2 − measured in the samples is expressed as µM concentration, calculated by using a sodium NO 2 − standard curve.

iNOS and COX-2 Determination by Cytofluorimetry
After treatment with extracts, fractions or compounds and LPS as described for NO evaluation, J774A.1 macrophages were collected, washed with phosphate buffered saline (PBS) and incubated in fixing solution (containing PBS, 2% FBS and 4% formaldehyde) at 4 • C for 20 min and subsequently in fix perm solution (containing PBS, 2% FBS, 4% formaldehyde and 0.1% Triton X) at 4 • C for 30 min. Anti-iNOS (BD Laboratories, Milan, Italy) or anti-COX-2 antibodies (BD Laboratories) were then added for 30 min. The secondary antibody, in fix perm solution, was added and then macrophages were evaluated using a fluorescence-activated cell sorting (FACSscan; Becton Dickinson, Milan, Italy) and elaborated with Cell Quest software as previously reported [15].

Data Analysis
Data are reported as mean ± standard error of the mean (S.E.M.) values of at least three independent experiments, each triplicate. Statistical analysis was performed by analysis of variance test, and multiple comparisons were made by Bonferroni's test by using Prism 5 (GraphPad Software, San Diego, CA, USA). p values lower than 0.05 were considered as significant. The HMBC spectrum showed a correlation from H-5 (δ H 6.74) to a carbonyl group at δ C 187.5 and from H-6 (δ H 7.00) to C-4a (δ C 121.6), resulting in the identification of a carbonyl group at position C-4 ( Figure 2). Furthermore, a polar aliphatic methine group (H-3) at δ H 4.22 displayed a correlation to C-4, C-2 and C-1 , which later established the position of a phenyl group at C-2. The HMBC correlation from H-2 to C-2 and the lack of correlation from H-2 to C-3, confirmed the position of C-2 and C-3. Due to the down field shift of the resonances of C-3 and C-2 and considering the molecular formula, the remaining coordination sites of the mentioned carbons (C-2 and C-3) were filled by an oxygen-bridge, introducing two chiral centers at C-2 and C-3. However, optical rotation measurement resulted in "zero rotation value" for this molecule, suggesting the presence of a racemate.

Investigation of Anti-Inflammatory Activity of Subfractions and Isolated Compounds on NO Release, iNOS Expression and COX-2 Expression Inhibition on J774.A.1 Macrophages
Many chronic inflammations display elevated levels of inflammation mediators or their producing enzymes such as NO, iNOS and COX-2 [24]. NO production can be modulated by various inducers such as lipopolysaccharide (LPS), tumor necrosis factor alpha (TNF-α), and interleukin-1β (IL-1β). However, the transcription of iNOS is the major regulatory step of NO production [25]. It has also been shown that NO production modulates the expression of COX-2 through the activation of an inflammation cascade, leading to an up-regulation of COX-2 and increasing levels of PGE2 [26,27]. Considering the correlation between these mediators and their critical role in inflammation, inhibition of them seems a promising therapeutic strategy for treatment of many chronic inflammatory diseases.
The crude methanolic extract of D. diapensifolia, along with subfractions obtained by partitioning with solvents of different polarity, were tested for their cytotoxic activity on a J774A.1 macrophage cell line prior further pharmacological investigation. As shown in Table 2, the results indicated that all subfractions were not toxic in the applied concentration range, while the MeOH extract of D. diapensifolia was slightly toxic on J774A.1 macrophages at higher concentrations (Table 2).

Investigation of Anti-Inflammatory Activity of Subfractions and Isolated Compounds on NO Release, iNOS Expression and COX-2 Expression Inhibition on J774.A.1 Macrophages
Many chronic inflammations display elevated levels of inflammation mediators or their producing enzymes such as NO, iNOS and COX-2 [24]. NO production can be modulated by various inducers such as lipopolysaccharide (LPS), tumor necrosis factor alpha (TNF-α), and interleukin-1β (IL-1β). However, the transcription of iNOS is the major regulatory step of NO production [25]. It has also been shown that NO production modulates the expression of COX-2 through the activation of an inflammation cascade, leading to an up-regulation of COX-2 and increasing levels of PGE2 [26,27]. Considering the correlation between these mediators and their critical role in inflammation, inhibition of them seems a promising therapeutic strategy for treatment of many chronic inflammatory diseases.
The crude methanolic extract of D. diapensifolia, along with subfractions obtained by partitioning with solvents of different polarity, were tested for their cytotoxic activity on a J774A.1 macrophage cell line prior further pharmacological investigation. As shown in Table 2, the results indicated that  Subsequent inhibitory activity assessment of the subfractions on NO production revealed the diethyl ether subfraction as most active subfraction, which could reduce the NO release by more than 70% in the applied concentration range (p < 0.001 vs. LPS alone, Table 3). Furthermore, evaluation of the iNOS inhibitory activity indicated that all subfractions significantly reduced iNOS release in a concentration dependent manner (p < 0.001 vs. LPS alone, Table 3). However, the diethyl ether subfraction displayed a lower activity than other tested fractions. As shown in Table 3, the ability to inhibit the COX-2 expression was only moderate with a maximum of 31.8% for the diethyl ether fraction (concentration of 5 µg/mL, p < 0.01 vs. LPS). Table 3. Effect of tested of the subfractions of D. diapensifolia. on NO release, iNOS and COX-2 expression. Data are expressed mean ± S.E.M of the percentage of inhibition vs. LPS (n = 3). ***, **, * denote p < 0.001, p < 0.01, p < 0.05 vs. LPS. In the next step, the isolated new compounds (1-3) along with compounds 4, 6, 8, 9, 10 and 11 were selected for further investigation. The selection of those known compounds was based on the lack of pharmacological activity in inflammation or in the case of 9-11, due to their contribution as major constituents of the subfraction. First, the cytotoxic activity of selected isolated compounds was assessed. Results illustrated that only compounds 6 and 10 at concentrations above 50 µM were growth-inhibitory (Table S1, Supplementary Materials). Subsequently, all the selected compounds were investigated for their effect on NO release, as well as on iNOS and COX-2 expression levels. As shown in Figure 4, the results indicated that compound 4 was the most active compound in NO release inhibition at all tested concentrations (5-50 µM) in a concentration-dependent manner. However, other investigated compounds displayed moderate activity in this assay.  Further investigation on the effect on iNOS expression levels showed that compounds 4, 6, 8, 11 caused a comparable reduction of iNOS expression like L-NAME in a concentration dependent manner ( Figure 5). In this regard, compounds were assessed further for their impact on COX-2 Further investigation on the effect on iNOS expression levels showed that compounds 4, 6, 8, 11 caused a comparable reduction of iNOS expression like L-NAME in a concentration dependent manner ( Figure 5). In this regard, compounds were assessed further for their impact on COX-2 expression. As shown in Figure 6, almost all tested compounds (except 1, 2, and 10) revealed a significant COX-2 expression reduction (p < 0.001 vs. LPS) in comparison to indomethacin (positive control) and in all tested concentrations (5-50 µM). In comparison, compound 1 and 2 moderately inhibited the COX-2 expression. Further investigation on the effect on iNOS expression levels showed that compounds 4, 6, 8, 11 caused a comparable reduction of iNOS expression like L-NAME in a concentration dependent manner ( Figure 5). In this regard, compounds were assessed further for their impact on COX-2 expression. As shown in Figure 6, almost all tested compounds (except 1, 2, and 10) revealed a significant COX-2 expression reduction (p < 0.001 vs. LPS) in comparison to indomethacin (positive control) and in all tested concentrations (5-50 µM). In comparison, compound 1 and 2 moderately inhibited the COX-2 expression.  Dionysia species are mainly known for their exudate flavonoids, covering a broad range of diverse structural features [8,16]. Previous studies showed the presence of not only regular flavonoids, but also irregular and aberrant flavonoids with unusual substitution patterns such as 2'hydroxyflavone, 3',4'-dihydroxyflavone and 8,2'-dihydroxyflavone [8]. Investigations of exudate flavonoids of D. diapensifolia led to the identification of 2',β-dihydroxychalcone (6), flavone (9), 2'hydroxyflavone (10), 2'-methoxyflavone (11), 3'-methoxy-4',5'-methylendioxyflavone, 5-hydroxy-2'methoxyflavone, apigenin, apigenin-7-methylether, naringenin (19), naringenin-7-methyl ether, Dionysia species are mainly known for their exudate flavonoids, covering a broad range of diverse structural features [8,16]. Previous studies showed the presence of not only regular flavonoids, but also irregular and aberrant flavonoids with unusual substitution patterns such as 2'-hydroxyflavone, 3',4'-dihydroxyflavone and 8,2'-dihydroxyflavone [8]. Investigations of exudate flavonoids of D. diapensifolia led to the identification of 2',β-dihydroxychalcone (6), flavone (9), 2'-hydroxyflavone (10), 2'-methoxyflavone (11), 3'-methoxy-4',5'-methylendioxyflavone, 5-hydroxy-2'-methoxyflavone, apigenin, apigenin-7-methylether, naringenin (19), naringenin-7-methyl ether, kaempferol, kaempferol-3-methylether, kaempferol-7-methylether and the isolation of an unusual dihydrochalcone (R)-(+)-3-acetoxy-3-phenyl-propiophenone (4) for the first time from a Dionysia species [10,16]. However, in this study, our phytochemical investigation led to the isolation of a new natural product 1 and two compounds 2, 3 so far only known as synthetic products, which can be classified as unusual flavonoids and phenolic compounds. Moreover, in the present study, the sesquiterpenoid 8 is reported for the first time from the genus Dionysia and the Primulaceae family.

% Inhibition of NO
Since the HPLC analysis of the crude methanolic extract of D. diapensifolia showed not only two major peaks for compound 9 and 11 (co-eluting with 13, Figure 7) but also a large peak at R T 10.8 min, which was mainly concentrated in the obtained n-butanol fraction ( Figure S22 Most of the known isolated compounds were described previously for their anti-inflammatory activity. Wang et al. showed that 6-hydroxyflavone (21) and its derivatives could potentially inhibit the downstream iNOS expression in kidney mesangial cells [28]. Luteolin (20), a natural flavone known as potent anti-inflammatory agent, found in many plant species, indicated to reduce TNF-α and proinflammatory cytokines expression, including that of TNF-α, IL-6 and IL-1β in RAW264.7 [29]. Apigenin (19) is a multitarget compound with promising anti-inflammatory activity. It has been reported that apigenin could suppress LPS induced nitric oxide (NO) and cyclooxygenase-2 (COX-2) expression in RAW 264.7 macrophage cells. In addition, it can attenuate acute lung injury (ALI) through inhibition of COX-2 and NF-kB activation pathways [30]. Furthermore, apigenin could significantly inhibit production of IL-1β and down regulated iNOS and COX-2 expression in a murine DSS colitis model [30]. Monohydroxy dibenzoylmethane (3), as curcumin structure analogue, displayed high inhibitory activity on chemical induced tumor promotion and inflammation on mouse skin. Furthermore, it induced apoptosis in human colorectal carcinoma cells through sequential activation of caspase cascades [31]. Considering previously published data along with our findings presented in this study qualify Dionysia diapensifolia and related species as valuable sources of flavonoids and chalcones with high anti-inflammatory capacity. Most of the known isolated compounds were described previously for their anti-inflammatory activity. Wang et al. showed that 6-hydroxyflavone (21) and its derivatives could potentially inhibit the downstream iNOS expression in kidney mesangial cells [28]. Luteolin (20), a natural flavone known as potent anti-inflammatory agent, found in many plant species, indicated to reduce TNF-α and proinflammatory cytokines expression, including that of TNF-α, IL-6 and IL-1β in RAW264.7 [29]. Apigenin (19) is a multitarget compound with promising anti-inflammatory activity. It has been reported that apigenin could suppress LPS induced nitric oxide (NO) and cyclooxygenase-2 (COX-2) expression in RAW 264.7 macrophage cells. In addition, it can attenuate acute lung injury (ALI) through inhibition of COX-2 and NF-kB activation pathways [30]. Furthermore, apigenin could significantly inhibit production of IL-1β and down regulated iNOS and COX-2 expression in a murine DSS colitis model [30]. Monohydroxy dibenzoylmethane (3), as curcumin structure analogue, displayed high inhibitory activity on chemical induced tumor promotion and inflammation on mouse skin. Furthermore, it induced apoptosis in human colorectal carcinoma cells through sequential activation of caspase cascades [31]. Considering previously published data along with our findings presented in this study qualify Dionysia diapensifolia and related species as valuable sources of flavonoids and chalcones with high anti-inflammatory capacity.

Conclusions
In this study, 25 compounds (19 flavonoids, five phenolic compounds and a sesquiterpenoid) were isolated from the aerial part of D. diapensifolia, of which three (compounds 1-3) were new natural products and one (compound 8) was found to be a new constituent of the Dionysia genus and the Primulaceae family. Among the known isolates, compounds 6, 9, 10, 11, 12 and 19 were already identified from exudates of D. diapensifolia, while the remaining compounds are reported for the mentioned species for the first time in this study. Since NOESY correlations were not conclusive, determination of the relative configuration of compound 1 was performed by DP4+ chemical shift calculations to establish the relative configuration of chiral centers C-2 and C-3. Furthermore, our results indicated that D. diapensifolia subfractions display remarkable anti-inflammatory activities through suppression of the release of inflammatory mediators such as NO and the reduction of the iNOS and COX-2 expression. Although a traditional use of this plant has not been reported clearly in literature, it might be used as a rich and cheap source of anti-inflammatory compounds, which can be used in the medicine and cosmetic industries.