Cytotoxic activity and molecular docking of a novel biflavonoid isolated from Jacaranda acutifolia (Bignoniaceae)

Abstract A novel biflavonoid [kaempferol (6→8″) apigenin] was isolated from the leaves of Jacaranda acutifolia. The structure was elucidated based on chemical evidence, 1D and 2D spectroscopic analyses as well as spectrometric techniques. The compound showed promising cytotoxic activity against breast cancer cell line MCF-7. The anticancer activity was explained via virtual docking of the isolated compound to the main sites in the human cyclin-dependent kinase2 (CDK2) crystal structure.

pharmacological potential and promising activities as antioxidant activities and in the field of tropical diseases, skin problems and venereal infections and as sources of lectins and trypsin inhibitors (Rana et al. 2013;Arcoverde et al. 2014;Mostafa et al. 2015).
Docking studies have become nearly indispensable for study of macromolecular structures and interactions. Mechanical model construction requires heroic patience and endurance to complete a structure which may contain several thousand atoms while computer graphics can build and display in seconds. Macromolecular modelling by docking studies provides most detailed possible view of drug-receptor interaction and has created a new rational approach to drug design where the structure of drug is designed based on its fit to three-dimensional structures of receptor site, rather than by analogy to other active structures or random leads (Bothara et al. 1998).
Herein, we report the isolation and structure determination together with the in vitro cytotoxic activity of a newly isolated biflavonoid (compound 1) from Jacaranda acutifolia Humb. and Bonpl. (Bignoniaceae) to develop a preliminary building block for the construction of a new drug.

Chemistry of the new compound
The methanol soluble fraction of aqueous extract of J. acutifolia was fractionated using Diaion HP-20 using water and methanol in different proportions. Among these, the neat methanol fraction was further separated to afford compound 1 (Figure 1). The compound was isolated as a yellow amorphous powder (12 mg), R f = 0.8 (on TLC; solvent system CH 2 Cl 2 :MeOH of 9:1 with addition of 1 drop acetic acid). It showed a dark purple-coloured spot under uV light, which turned yellow upon exposure to ammonia vapour. It gave brown colour with FeCl 3 reagent. The uV spectrum showed two absorption maxima at λ max (MeOH): 282, 302 (sh), 367 nm, which revealed the presence of 3-OH group (a flavonol moiety). upon addition of NaOMe, a bathochromic shift (33 nm) was obtained in band I indicating that one or both of the 4′ and 4‴ OH groups in ring B of both monomers are free, while the presence of a shoulder at 328 nm indicates that 7 and 7″ OH groups are free.
Further confirmation of the biflavonoid structure was obtained through APT spectrum, which was assigned by 2D experiments (HMBC and HSQC). Two signals at 177.57 and 182.4 confirmed the presence of flavones and flavonol moieties for C-4 and C-4″. Four signals at 128.95, 128.20, 116.63 and 116.44 ppm confirmed the presence of two para-substituted aromatic rings. The signal at 103.12 ppm is for C-3″ of the apigenin moiety. The presence of kaempferol moiety was further confirmed by the presence of signals at δ ppm 135.08 and 177.57 for C-3 and C-4 of kaempferol, respectively, as compared to the recently isolated (8→8″ bikaempferol) (Ye et al. 2012). Signals at 99.42 and 102 ppm were assigned for C-6″ and C-8, respectively, which confirmed the linkage position of the dimer.
The result of COSY spectrum lent a support to the above suggested structure as no correlation (indicated from the absence of cross peaks) between H-6″ and H-8 was observed, this indicated that these two protons are not in the same monomer. To unravel any ambiguity about the structure of the compound, HMBC and HSQC analyses were applied. Concerning HMBC spectrum, the proton signal at δ ppm 6.82 (H-3″ of the apigenin moiety) showed a cross peak with the carbon signal at δ ppm 105.65 (C-10″) which exhibited a long-range coupling with the proton signal at δ ppm 6.35(H-6″), this declared that the free H-6″ belongs to the apigenin moiety and that apigenin is linked to kaempferol through C-8″. A further confirmation was achieved via HSQC spectrum.
The compound showed a protonated molecular ion peak at m/z 555.35 [M + H] + in ESI-MS, positive ion mode, corresponding to the molecular formula C 30 H 18 O 11 . In addition, protonated fragment ion peaks appear at m/z 119, 135 and 303 as a result of retro Diels-Alder fragmentation. Consequently, the compound was identified to be kaempferol (6→8″) apigenin.
To the best of our knowledge, this is the first report of isolation of this compound from nature and the second report of the presence of biflavonoids in family Bignoniaceae.

Cytotoxic activity of the new compound
The compound also showed a potential cytotoxic activity against MCF-7 (breast carcinoma cell) with IC 50 = 18.3 μg/mL, which is equivalent to 33 μM ( Figure 2). Vinblastine was used as a reference drug (IC 50 = 4.6 μg/mL).

Docking study
The binding mode of compound L1 revealed the formation of two hydrogen bonds with the residues Leu 83 and Asp 86 of the of human cyclin-dependent kinase2 (CDK2) (PDB ID1FVT). The isolated compound was able to interact with active site in a similar pattern as the active ligand. The results showed that the isolated compound Kaempferol (6→8″) apigenin exhibited a similar binding mode as the lead compound L1 with higher fitting scores ( Table 1). The binding mode is illustrated in Figures 3 and 4 with fitting score of −61.82, which revealed the formation of extra two π bonds with the two amino acids Lys 129 and Gly 13, beside the formed two hydrogen bonds with two amino acid residues, namely Asp 86 and Lys 129, which indicate more firm binding. Interestingly, the docked compound is characterised by possessing many phenolic groups, which dissociate under physiological conditions resulting in Oions (Wink 2008). The polyphenols can form ionic bonds with positively charged side chains of the aspartate amino acids (Asp 86). Although these bonds are quite weak but because several of them are formed concomitantly, the effect is much stronger leading to more fitting in the binding pocket. Accordingly, the performed molecular modelling studies indicated that the proposed molecules are promising candidates for anticancer activity. Additionally, Inhibition of cyclin-dependent kinase2 (CDK2), a positive regulator of eukaryotic cell cycle progression, may represent a therapeutic strategy for prevention of chemotherapy-induced   (2) −61.82 alopecia (CIA) by arresting the cell cycle and reducing the sensitivity of the epithelium to many cell cycle-active antitumor agents (Davis et al. 2001).

Plant material
Leaves of J. acutifolia Humb. and Bonpl., family Bignoniaceae were collected from the Merryland Botanical Garden, Cairo, Egypt, and were air-dried. They were kindly authenticated by Dr Abd El Salam Mohamed Al-Nowiahi, Professor of Taxonomy, Faculty of Science, Ain Shams university. Voucher specimen of authenticated leaves of J. acutifolia (JAB-2010) was deposited at the Department of Pharmacognosy, Faculty of Pharmacy, Ain Shams university, Abbassia, Cairo, Egypt.

Extraction and isolation
The crushed air-dried leaves of J. acutifolia (2 kg) were boiled in distilled water for 2 h and then filtered while hot. The water extract was lyophilised to fine powder, which was then dissolved and extracted with methanol. The methanol portions were distilled off in rotary evaporator at 55 °C till dryness. The extract was concentrated till constant weight and then kept in vacuum dessicator over anhydrous CaCl 2 . The methanol soluble fraction of aqueous extract of leaves of J. acutifolia (87 g) was dissolved in the least volume of methanol. Then, it was fractionated over a column chromatography (100 L × 4.5 ID cm) packed with ion exchange resin, Diaion HP-20 as a stationary phase. The elution process started with distilled water (F-I) followed by 25% methanol (F-II), then 50% methanol (F-III) and 75% methanol (F-IV) followed by neat methanol (F-V) as an eluent. Finally, the column was eluted with acetone (F-VI) to ensure complete elution. According to the results obtained from TLC and 2D-PC analyses of the eluted fractions, similar fractions were pooled together and dried in vacuo at 45 °C, yielding six main fractions from fraction (I-VI); then, each of the pooled fractions was further subjected to 2D paper chromatographic analysis. Fraction V eluted with neat methanol (4.5 g) was refractionated over a column (30 L × 2.5 ID cm) packed with Sephadex LH-20 as adsorbent. The column was eluted with neat methanol and then acetone for column wash. Subfractions (2 mL each) were collected. Co-chromatograms were made on TLC using solvent system composed of CH 2 Cl 2 :MeOH (9:1) with addition of 1 drop acetic acid. Similar subfractions were pooled together to give one major pure compound, showing a dark purple spot under uV, which turned yellow on exposure to ammonia vapours and gave a brown colour upon spraying with FeCl 3 reagent, this yielded compound 1.

Cytotoxicity and cell viability assay (MTT assay)
Sensitivity of MCF-7 (breast carcinoma cells) to the tested compound was determined using MTT cell viability assay (Mosmann 1983;Van de Loosdrecht et al. 1991;Marks et al. 1992).

Molecular docking study
In silico molecular modelling of kaempferol (6→8″) apigenin on human cyclin-dependant kinase2 (CDK2) was carried out using Discovery Studio 2.5 software (Accelrys Inc., San Diego, CA, uSA) applying C-Docker protocol. The new compound is docked using several enzymes in several trial and error as topoisomerases enzymes. The X-ray crystal structure of human cyclin-dependent kinases (PDB ID1FVT) co-crystallised with its ligand (L1) was downloaded from protein data bank (www.pdb.org). The structure of the enzyme was established using the default protein preparation protocol of Accelry's discovery studio 2.5 (Accelrys ® , Inc., San Diego). First, the protein is prepared by adding hydrogen atoms and cleansing it from any unwanted interactions. Then, determination of the binding site was achieved via detection of the binding mode of bioactive conformation of the reported (L1), co-crystallised with CDK2. Following the binding site determination, the structures of the compound were docked inside the binding site using C-Docker protocol. CHARMm force field was assigned and the binding energies for the selected docking poses were calculated applying the following equation: where ΔG binding : The ligand-enzyme interaction binding energy, E complex : The potential energy for the complex of CDK2 bound with the ligand, E CDK2 : The potential energy of the protein alone and E ligand : The potential energy for the ligand alone.