New Antiproliferative Triflavanone from Thymelaea hirsuta—Isolation, Structure Elucidation and Molecular Docking Studies

In this study isolates from Thymelaea hirsuta, a wild plant from the Sinai Peninsula of Egypt, were identified and their selective cytotoxicity levels were evaluated. Phytochemical examination of the ethyl acetate (EtOAc) fraction of the methanolic (MeOH) extract of the plant led to the isolation of a new triflavanone compound (1), in addition to the isolation of nine previously reported compounds. These included five dicoumarinyl ethers found in Thymelaea: daphnoretin methyl ether (2), rutamontine (3), neodaphnoretin (4), acetyldaphnoretin (5), and edgeworthin (6); two flavonoids: genkwanin (7) and trans-tiliroside (8); p-hydroxy benzoic acid (9) and β sitosterol glucoside (10). Eight of the isolated compounds were tested for in vitro cytotoxicity against Vero and HepG2 cell lines using a sulforhodamine-B (SRB) assay. Compounds 1, 2 and 5 exhibited remarkable cytotoxic activities against HepG2 cells, with IC50 values of 8.6, 12.3 and 9.4 μM, respectively, yet these compounds exhibited non-toxic activities against the Vero cells. Additionally, compound 1 further exhibited promising cytotoxic activity against both MCF-7 and HCT-116 cells, with IC50 values of 4.26 and 9.6 μM, respectively. Compound 1 significantly stimulated apoptotic breast cancer cell death, resulting in a 14.97-fold increase and arresting 40.57% of the cell population at the Pre-G1 stage of the cell cycle. Finally, its apoptosis-inducing activity was further validated through activation of BAX and caspase-9, and inhibition of BCL2 levels. In silico molecular docking experiments revealed a good binding mode profile of the isolates towards Ras activation/pathway mitogen-activated protein kinase (Ras/MAPK); a common molecular pathway in the development and progression of liver tumors.


Introduction
Cancer is a leading cause of morbidity and mortality. Years of research attempting to find new and effective anticancer agents has not only improved survival and the quality of life of patients, but has also significantly reduced the need for surgical procedures [1]. Natural products derived from plants have served as an important source of pharmacologic agents for several diseases [2], including cancer [3]. An example is the Thymelaeaceae family of plants, that have been shown to exhibit many pharmacological activities [4][5][6][7][8][9][10][11][12][13][14][15][16][17][18]. For example, extracts of Daphne giraldii [4] and Wikstroemia indica [5] from the Thymelaeaceae family of plants are known to contain various pharmacologically active constituents, including: flavonoids [6], coumarins [7], diterpenoids [8], lignans, volatile oils and polysaccharides [9] that have been used for the tretament of several ailments. The plants Lasiosiphon eriocephalus [10], Daphne mezereum [11], and Daphne acutiloba [12] have also been shown to have cytotoxic activities against different types of cancer. Another important member of the Thymelaeaceae family is T. hirsute, a native plant of North Africa. T. hirsute contains flavonoids, coumarins, tannins, and saponins [13], and has been demonstarted to have antiviral activity against HIV-1 [14], antioxidant activity [15], antidiabetic and antihypertensive activities [16]. Other studies have also shown isolates from T. hirsuta to have significant cytotoxicity against hepatocellular carcinoma (HCC) [17], which is one of the leading causes of cancer deaths worldwide, with a rapidly increasing incidence rate in Egypt [18].
In a continuing effort to discover bioactive metabolites from Egyptian plants [17,19], our group focused on the isolation of active constituents from the EtOAc fraction of the total MeOH extract from the whole T. hirsute plant. Different spectroscopic techniques were then used to identify the isolated compounds. The cytotoxic activities of eight of the isolated compounds were tested against cancer and normal cell lines using a sulforhodamine-B (SRB) assay, following the method reported by Vichai and Kirtikara [20]. Molecular docking experiments were also conducted to gain insight into the molecular mechanism of the isolates.

Identification of Compounds 1-10
Compound 1 was isolated as a yellow powder and its molecular formula was determined as C 45 Figure S1) also showed a fragment ion m/z of 541.1134, which represented a neochamaejasmin B fragment ion, in addition to a fragment ion m/z of 415.0932 that represented the loss of a phloroglucinol unit (C 6 H 6 O 3 ) from the neochamaejasmin B fragment ion [21], indicating that neochamaejasmin B is part of compound 1.  The 13 C-NMR data (Table 1) showed the presence of three carbonyls (δ C 198. 9, 196.6, 197.3). The structure of compound 1 was therefore determined to consist of three flavanone units. By comparing the spectroscopic data: 1 H-NMR, 13 C-NMR and HMBC spectra ( Figures S2-S7), with the published data of mithnin [17], neochamaejasmin B [22], and 3,3"-biflavanones [23][24][25], compound 1 was found to be a triflavanone. By comparing the J values (H-2, J = 4.5 Hz and H-2" J = 9.0 Hz) with those of mithnin [17], and 3,3"biflavanones [23][24][25], C-2/C-3 and C-2"/C-3" of the first and second flavanone regions of 1, were found to have a cis-trans geometry.

Cytotoxic Activities
Two of the isolates, Genkwanin (7) and β sitosterol glucoside (10), had already been tested for their cytotoxic activity against HepG2 cancer cell line [31,36], therefore, we only screened for the cytotoxicity of the other eight isolated compounds 1-6, 8 and 9 against liver HepG2 and normal Vero cells (Table 2, Figure 3).

Cytotoxic Activities
Two of the isolates, Genkwanin (7) and β sitosterol glucoside (10), had already been tested for their cytotoxic activity against HepG2 cancer cell line [31,36], therefore, we only screened for the cytotoxicity of the other eight isolated compounds 1-6, 8 and 9 against liver HepG2 and normal Vero cells (Table 2, Figure 3).

Annexin V/PI Staining
The promising anti-cytotoxic activity of Compound 1 prompted us to further investigate this compound for its apoptosis-inducing activity in MCF-7 cells. After treatment, cells were subjected to flow cytometric analysis of annexin V/PI staining with cell cycle analysis to determine the cell population in different cell cycle phases. As seen in Figure 4, compound 1 significantly stimulated apoptotic breast cancer cell death, demonstrating a 14.97-fold cell death increase. It also increased the apoptosis ratio by 40.57%, compared to 2.71% for the control. It increased the induction of early apoptosis by 2.44% compared to 0.61% for the control, and late apoptosis by 22.52% compared to 0.23% for the control. Our results are in accordance with our previous studies [37,38] of apoptosis induction with various isolates from medicinal plants. As shown in Figure 3, compounds 1, 2 and 5 exhibited remarkable cytotoxic activities against HepG2 cells with IC 50 values of 8.6, 12.3 and 9.4 µM, respectively. However, they exhibited non-toxic activities against the Vero cells (IC 50 > 50 µM). Following, the three compounds were further tested against the cancer cell lines, MCF-7 and HCT-116. Compounds 1, 2 and 5 showed cytotoxic activities against MCF-7 cells, with IC 50 values of 4.26, 66.14 and 46.38 µM, respectively, and against HCT-116 cells, with IC 50 values of 9.60, 100.90 and 24.35 µM, respectively. Compound 1 clearly showed very promising cytotoxic activity against all five tested cancer cell lines, and was consequently chosen for further investigation to determine its apoptosis induction ability in MCF-7 cells.

Annexin V/PI Staining
The promising anti-cytotoxic activity of Compound 1 prompted us to further investigate this compound for its apoptosis-inducing activity in MCF-7 cells. After treatment, cells were subjected to flow cytometric analysis of annexin V/PI staining with cell cycle analysis to determine the cell population in different cell cycle phases. As seen in Figure 4, compound 1 significantly stimulated apoptotic breast cancer cell death, demonstrating a 14.97-fold cell death increase. It also increased the apoptosis ratio by 40.57%, compared to 2.71% for the control. It increased the induction of early apoptosis by 2.44% compared to 0.61% for the control, and late apoptosis by 22.52% compared to 0.23% for the control.
Our results are in accordance with our previous studies [37,38] of apoptosis induction with various isolates from medicinal plants.

Cell Cycle Analysis
Cell cycle analysis is an important test that determines the percentages of the cell population that are in each cell phase.. In this studyMCF-7 cancer cells were treated with compound 1, and then subjected to DNA flow cytometry to determine the stage where cell proliferation was arrested in the cell cycle. As seen in Figure 5, compound treatment significantly increased the Pre-G1 population by 40.57%, compared to 2.71% for the control. In contrast, there was no significant change in the cell population in the G0/G1, S or G2/M cell phase.

Cell Cycle Analysis
Cell cycle analysis is an important test that determines the percentages of the cell population that are in each cell phase.. In this studyMCF-7 cancer cells were treated with compound 1, and then subjected to DNA flow cytometry to determine the stage where cell proliferation was arrested in the cell cycle. As seen in Figure 5, compound treatment significantly increased the Pre-G1 population by 40.57%, compared to 2.71% for the control. In contrast, there was no significant change in the cell population in the G0/G1, S or G2/M cell phase.

Enzymatic Assays for Apoptotic Markers
For further investigation of compound 1 apoptotic behavior, treated and untreated MCF-7 cells with compound 1 were analyzed for the activities of pro-apoptotic BAX and caspase-9, and the anti-apoptotic BCL2. As shown in Figure 6, treatment with compound 1 significantly increased BAX level (362 pg/mL) compared to the control (43.6 pg/mL), as well as caspase-9 level (21.14 pg/mL) compared to the control (5.16 pg/mL), while it decreased BCL2 level to 2.85 ng/mL compared to the control (7.59 ng/mL). Hence, the enzymatic activity of both BAX and caspase-9 and the inhibition of BCL2 represent significant evidence of the induction of apoptosis in the MCF-7 cells.

Enzymatic Assays for Apoptotic Markers
For further investigation of compound 1 apoptotic behavior, treated and untreated MCF-7 cells with compound 1 were analyzed for the activities of pro-apoptotic BAX and caspase-9, and the anti-apoptotic BCL2. As shown in Figure 6, treatment with compound 1 significantly increased BAX level (362 pg/mL) compared to the control (43.6 pg/mL), as well as caspase-9 level (21.14 pg/mL) compared to the control (5.16 pg/mL), while it decreased BCL2 level to 2.85 ng/mL compared to the control (7.59 ng/mL). Hence, the enzymatic activity of both BAX and caspase-9 and the inhibition of BCL2 represent significant evidence of the induction of apoptosis in the MCF-7 cells.

Molecular Docking
Molecular docking simulation is an important approach for predicting the coupling of a substrate with its receptor. The Ras activation/pathway mitogen-activated protein kinase (Ras/MAPK) pathway has been identified as being capable of inducing tumor initiation, development, and progression in HCC and thus, it represents a promising target for cancer drug development. As noted above, several substituted coumarins have been shown to have potent anti-hepatocellular carcinoma activities through their targeting of the Ras/MAPK pathway [39][40][41]. With the exception of two of the compounds (Compounds 9 and 10), the rest of the identified compounds are coumarins, which prompted us to evaluate the binding modes of the compounds with MAPK.
Docking experiments were used to map the binding pose of isolates with MAPK and to predict their binding affinities. The docking scores and binding free energies of the lowest energy pose of the compounds in the MAPK active site-using the Schrodinger Glide module-are shown in Table 3. MAPK is an ATP-dependent protein kinase, in which the hydrophobic pocket (the ATP-binding pocket) is rich in alanine, valine, isoleucine, and leucine amino acids.

Molecular Docking
Molecular docking simulation is an important approach for predicting the coupling of a substrate with its receptor. The Ras activation/pathway mitogen-activated protein kinase (Ras/MAPK) pathway has been identified as being capable of inducing tumor initiation, development, and progression in HCC and thus, it represents a promising target for cancer drug development. As noted above, several substituted coumarins have been shown to have potent anti-hepatocellular carcinoma activities through their targeting of the Ras/MAPK pathway [39][40][41]. With the exception of two of the compounds (Compounds 9 and 10), the rest of the identified compounds are coumarins, which prompted us to evaluate the binding modes of the compounds with MAPK.
Docking experiments were used to map the binding pose of isolates with MAPK and to predict their binding affinities. The docking scores and binding free energies of the lowest energy pose of the compounds in the MAPK active site-using the Schrodinger Glide module-are shown in Table 3. MAPK is an ATP-dependent protein kinase, in which the hydrophobic pocket (the ATP-binding pocket) is rich in alanine, valine, isoleucine, and leucine amino acids. The docking study suggested that the isolates interact with several main amino acid residues at the MAPK active site through hydrophobic contacts (orange lines), H-bonding (yellow dotted lines), and electrostatic interactions (cyan color) (Figure 7).  The docking study suggested that the isolates interact with several main amino acid residues at the MAPK active site through hydrophobic contacts (orange lines), H-bonding (yellow dotted lines), and electrostatic interactions (cyan color) (Figure 7). The MAPK protein is capable of forming hydrogen-bond and hydrophobic interactions with ligands at the active site as shown in Figure S8. Our modeling suggests that nearly all of the isolates could form hydrophobic and π-π stack interactions with PHE169. The hydroxyl and ketonic groups of the isolates also should contribute to hydrogen-bond interactions with several amino acids, including TYR36, LYS54, LYS116, and ASP113. These interactions are expected to allow for tight binding of the isolates to MAPK ( Figures  S9-S11). Hydroxylated-chromones mimic the adenine ring of ATP and occupied the hydrophobic pocket in the hinge region, forming hydrogen-bond interactions with MET 110, and LYS116.
To choose the best docking pose with the lowest energy, the docking procedure was repeated several times. The results of the docking protocols were similar, suggesting the The MAPK protein is capable of forming hydrogen-bond and hydrophobic interactions with ligands at the active site as shown in Figure S8. Our modeling suggests that nearly all of the isolates could form hydrophobic and π-π stack interactions with PHE169. The hydroxyl and ketonic groups of the isolates also should contribute to hydrogen-bond interactions with several amino acids, including TYR36, LYS54, LYS116, and ASP113. These interactions are expected to allow for tight binding of the isolates to MAPK (Figures S9-S11). Hydroxylated-chromones mimic the adenine ring of ATP and occupied the hydrophobic pocket in the hinge region, forming hydrogen-bond interactions with MET 110, and LYS116.
To choose the best docking pose with the lowest energy, the docking procedure was repeated several times. The results of the docking protocols were similar, suggesting the glide docking technique was highly reproducible. The isolates' extra precise glide docking with MAPK showed a good docking score of −8.55 and −8.34 kJ/mol with a glide Emodel value of −58.78 and −100.70 kcal mol −1 , especially for compounds 7 and 8, respectively. The reliability of the planned docking protocol was evaluated by assessing the interactions of the most widely recognized flavonoid of modulation activity, rutin, with MAPK [42]. The isolates showed a higher binding affinity and binding energy compared to rutin, which showed the same pose ( Figure S12) as the isolates, achieving encouraging docking ratings of −7.32 kJ/mol and a glide Emodel value of −69.18 kcal mol −1 . Furthermore, the isolates were compared to sorafenib, an oral Ras/MAPK inhibitor that is used as a therapy for advanced liver carcinoma [43]. Sorafenib occupies the active site of MAPK ( Figure S13) in the same pose as the isolates and shows hydrogen-bond interactions with GLU72, LYS116, and ASP169, achieving a comparable docking score (−8.37 kJ/mol) and binding energy (−81.1 kcal mol −1 ) to those of our isolates.

Plant Material
The whole plant of T. hirsuta was collected from the Sinai Peninsula of Egypt and was authenticated by the Botany department, faculty of Science, Suez Canal University, Egypt. The voucher specimen (No. SAA-162) was stored at the Pharmacognosy Herbarium, Suez Canal University, Ismailia, Egypt.

Extraction and Isolation
The plant (10 kg) was air dried, grounded, and the resulting powder was macerated using methanol at room temperature to yield 250 g of dry extract. Fractionation using vacuum liquid chromatography (VLC) with gradient elution led to the separation of different fractions. Using a hexane, ethyl acetate, and methanol gradient elution, the EtOAc fraction was subjected to column chromatography of silica gel type. Five fractions were obtained and subjected to different chromatographic separation techniques, including silica gel and sephadex LH-20 column chromatography.
For the fraction of 10% methanol in ethyl acetate, 9 g was purified using sephadex LH-20 with a mixture of methanol and chloroform (1:1) as the isocratic mobile phase to yield 1 (23 mg). For the fraction of 60% ethyl acetate in hexane, 7 g was purified using silica gel column chromatography and isocratic elution to yield 2 (17 mg), 3 (25 mg), and 4 (23 mg). For the fraction of 70% ethyl acetate in hexane, 5 g was purified using silica gel column chromatography and isocratic elution to yield 5 (27 mg) and 6 (14 mg). For the fraction of 40% ethyl acetate in hexane, 5 g was purified using sephadex LH-20 with a mixture of methanol and chloroform (1:1) as the isocratic mobile phase to yield 7 (15 mg) and 9 (10 mg). For the fraction of 100% ethyl acetate, 8 g was purified using sephadex LH-20 with mixture of methanol and chloroform (1:1) as the isocratic mobile phase to yield 8 (30 mg) and 10 (35 mg).

Cell Lines
The cell lines used in this study were purchased from the National Cancer Institute, Cairo, Egypt and maintained in Dulbecco's Modified Eagle Medium/F-12 (DMEM⁄F12, Sigma-Aldrich, USA), supplemented with 2 mM L-glutamine (Lonza, Belgium) and 10% fetal bovine serum (FBS, Sigma-Aldrich, MO, USA), 1% penicillin-streptomycin (Lonza, Belgium). All cells were cultured following routine tissue culture work, and treated with serial concentrations of the compounds for 48 h. Absorbance was subsequently measured (at 570 nm) using ELISA microplate reader (BIO-RAD, model iMark, Japan). The viability was calculated relative to a control and the IC 50 values were determined using the nonlinear regression curve fit, as previously reported in [44]

Procedure
The cytotoxic activities of the isolated compounds against HepG2, VERO E6, MCF-7 and HCT-116 cells were tested using the sulforhodamine-B (SRB) assay [20] and the results are represented in figure e and Table 2

Molecular Modeling
A molecular modeling analysis was conducted using the Glide docking method implemented in the molecular modeling software Schrodinger-10.1. The X-ray crystal structure of the MAPK enzyme's catalytic domain inhibitor complex (PDB ID: 5EKN; 2.6 Å) was obtained from the Protein Data Bank (PDB) (http://www.rcsb.org/pdb). Using the protein preparation wizard and the OPLS-2005 force field, the MAPK-ligand complex was optimized for the Glide docking calculations. Crystallographic water was extracted, if present, and hydrogen was added to the structure corresponding to pH 7.0; the most likely positions of hydroxyl and thiol hydrogen atoms, taking into account the correct ionization conditions for both the basic and acidic amino acid residues of the protein. The protein charge and protonation state were then modified by the protein assignment script, and the protein-inhibitor complex was subjected to energy minimization until the non-hydrogen atom average root mean square deviation (RMSD) exceeded 0.3 Å to trigger steric clashes using the OPLS-2005 force field.
The 3D structure of the N58 was constructed using the ligand preparation wizard and optimized with the Maestro build stand. For each input molecule the ligand preparation feature produces many low-energy 3D structures with different ionization states, tautomers, stereo-chemistries, and ring conformations. Partial atomic charges were attributed to N58 using the force-field of OPLS-2005, and potential ionization states were created at a pH of 7. The van der Waal radii of the receptor atoms were multiplied by 0.8 with a partial atomic charge of 0.15 to minimize the potential for non-polar parts of the receiver. At the center of the active site, a grid box with coordinates X = 10, Y = 10 and Z = 10 was generated. The resulting ligand structures were further optimized by energy minimization until the RMSD limit of 0.01 Å was reached. Having ensured that the MAPK enzyme and inhibitor molecules were in the appropriate form, the properties and shape and the active site of MAPK were characterized using the "grid generation row" in Glide. In the final step, using the optimized protein-ligand geometries, the separated compounds were docked inside the active site of MAPK. The extra precision (XP) glide scoring function, which flexibly docks ligands, was applied to rank the docking poses and to evaluate the binding affinities between the protein-ligand. Maestro's Pose Viewer versatility was used to envision and evaluate the key elements of the ligand-receptor interaction. Using a glide score function, the final best-docked structure with the lowest energy was selected for further experiments. The inhibitor was removed from the MAPK enzyme crystal structure and then re-docked using the above step to evaluate the accuracy and precision of the docking protocol established [48,49].

Data Availability Statement:
No new data were created or analyzed in this study. Data sharing is not applicable to this article. The data presented in this study are available in Supplementary Materials.