Structure elucidation of active compounds from Coffea canephora Pierre ex A.Froehner cascara and their potential as anticancer against breast cancer cells

Context : One approach to cancer therapy medication is exploring medicinal plants that contain one or more compounds specifically targeting cancer cells with fewer side effects. Cascara from coffee fruit ( Coffea canephora Pierre ex A.Froehner) is a waste rarely processed but has various chemical contents that can be used in medicine. Aims : To evaluate the in silico and in vitro activity of compounds isolated from ethanolic extract of C. canephora cascara against HeLa and MCF-7 breast cancer cells. Methods : Isolation of the compounds by radial chromatography and thin layer chromatography techniques, and the chemical structures were elucidated by infrared radiation, ultraviolet, nuclear magnetic resonance spectroscopy, and mass spectrometry. In silico study about the compounds binding with the receptor responsibility to cancer (caspases 3 and 9). In vitro study by examining the cytotoxicity of HeLa and MCF-7 cells of the isolated compounds from C. canephora . Results : Four known bioactive compounds, lupeol (1), stigmasterol (2), ursolic acid (3), and caffeic acid (4), were isolated from the ethanol extract of C. canephora cascara. Based on the ESI-MS results, the m/z value for lupeol was 427.50 [M+H] + , stigmasterol was 454.48 [M+ACN+H] + , ursolic acid was 456.51 [M+H] + , and caffeic acid was 179 [M-H]. In silico and in vitro data show that the ursolic acid compound has activity against HeLa and MCF-7 cancer cells with IC 50 values of 25.98 ± 0.01 µg/mL and 12.70 ± 0.11 µg/mL, respectively. Conclusions : All isolated compounds from C. canephora cascara have a promising ability to interact with caspases 3 and 9, particularly ursolic acid, which has the smallest IC 50 value against HeLa and MCF-7 breast cancer cells.


INTRODUCTION
Cancer is one of the diseases caused by perpetual genetic mutations until the alteration of cell function due to the abnormality of some genes (Schegoleva et al., 2022).Mutation of cancer cells is caused by some factors, for instance, radiation of X-rays, UV rays, high exposure to radioactive substances, carcinogenic substances, and viruses (Melosky et al., 2022).Based on WHO data in 2020, cancer is one of the biggest mortality rates caused by many factors (Giaquinto et al., 2022).
There are some medications related to the manifestation of cancer therapy: surgical therapy, radiation, chemotherapy, immunotherapy, and targeted medicines.Recently, chemotherapy and radiation have been generally applied, yet there were some side effects and resistance to these therapies (Li et al., 2022;Love et al., 1989).As time goes by, researchers develop other approaches to cancer therapy medication.One is by exploring medicinal plants that contain one or more compounds specifically targeting cancer cells with fewer side effects (Li et al., 2022).Therefore, the challenge of continually searching for new cancer drugs with diverse chemical structures that can effectively combat cancer cells and have low side effects remains essential, including through the exploration of natural substances, primarily from plants (botanical raw materials) (Li et al., 2022).Botanical raw materials that have been proven to have anticancer properties include species of coffee plants (Coffea sp) (Gallardo-Ignacio et al., 2022).
Coffea canephora Pierre ex A.Froehner (family Rubiaceae) has anticancer properties (Gallardo-Ignacio et al., 2022).In coffee cultivation, around 50-60% of cascara waste is generated (Rios et al., 2020).Recently, processed cascara has been produced as food and supplement products because it contains proteins, polysaccharides, and active compounds (Klingel et al., 2020).This will be a promising approach to developing cancer therapy that can be targeted specifically to cancer cells without affecting the normal cells in the body.According to Yashin et al. (2017), some compound contents found in coffee beans are also present in the coffee cascara, making it an opportunity for initial testing and development of the potential anticancer properties of coffee cascara, particularly from Robusta coffee varieties (Duangjai et al., 2016;Durán-Aranguren et al., 2021).
The process of designing, discovering, and optimizing bioactive compounds in the development of a new drug can be facilitated through in silico methods.In silico represents a fast and cost-effective way to identify new drugs.In silico is an effective method to identify the activity of many compounds isolated from plants to the targeted receptor using software (Hardjono et al., 2016;Nur et al., 2023).This method is useful for screening secondary metabolites isolated from coffee as a preliminary approach for further experiments.This study aimed to evaluate the in silico and in vitro activity of compounds isolated from ethanolic extract of C. canephora cascara against HeLa and MCF-7 breast cancer cells.

Chemical and reagents
Chemicals such as CeSO4, ethanol, n-hexane, ethyl acetate, and EDTA were purchased from Merck, USA.DMEM media, FBS, antibiotic penicillin-streptomycin, trypsin, trypan blue from Gibco, USA.HeLa and MCF-7 cell lines from ATCC, PrestoBlue cell viability reagent from Invitrogen, USA, and cisplatin and DMSO were purchased from Sigma-Aldrich, USA.

Preventive parading
C. canephora fruits were cleaned with water and cut into small pieces, then next were air-dried to give 5.0 kg of dry sample.The dried sample was ground and extracted with 70% ethanol (EtOH).After filtration, the extract was dried in a vacuum at 45ºC to give a crude extract.Fractionation of the extract was carried out with vacuum liquid chromatography (VLC) using a column (80 mm id × 500 mm) of silica gel (7734).The mobile phase was n-hexane/ethyl acetate (EtOAc) with increased polarity.Radial chromatography (RC) was utilized for the purification process, using 95:5 n-hexane-EtOAc in 10% polarity increment to weight compounds 1 (25 mg), 2 (14 mg), 3 (18 mg) and 4 (32 mg) (Rosandy et al., 2018).

Culture cell conditions and cytotoxic assay
The cytotoxicity of compounds 1-4 was determined with a cell viability test using PrestoBlue® assay.The cells were maintained in a Roswell Park Memorial Institute (RPMI) medium with 10% (v/v) Fetal Bovine Serum (FBS) and 1 µL/1 mL antibiotics (1% penicillin-streptomycin).Cultures were incubated at 37°C in a humidified atmosphere of 5% CO2 (Aisyah et al., 2022;Nur et al., 2021).MCF-7 cells were plated in 96 multiwell culture plates at a density of 1.7 × 10 4 cells/well.After twenty-four hours, the medium was discarded, and fresh medium containing samples (C.canephora extracts, fractions, and isolates) with different concentrations of 7.81, 15.63, 31.25, 62.50, 125.00, 250.00, 500.00, 1000.00 µg/mL and with the positive control cisplatin.After incubation with the sample for 24 h, PrestoBlue® reagent (resazurin dye) was added.The PrestoBlue® assay results were read using a multimode reader at 570 nm (Tecan Infinite M200 PRO, Switzerland).The IC50 value was determined by linear regression using Microsoft Excel software.Then, the IC50 value was calculated as the extract concentration that inhibited cell viability by 50% after the exposure time.The selectivity index (SI) was defined as the ratio of the IC50 value observed against cancer cell lines to the IC50 value observed in non-malignant cells (Céspedes et al., 2023).

Preparation of ligand structure
To create ligand structures, including lupeol, stigmasterol, ursolic acid, caffeic acid, and cisplatin as reference compounds, began with designing the 2D form by using the ChemDraw Ultra 8.0 program inside the package ChemOffice v8.0 program.Then, to 3D form by using Chem3D v8.0 on the same package.The file was saved in *mol format.The 3D structure was optimized geometrically by using the Hyper-Chem Release v8.07 program.The optimization of the tested compounds was carried out by the software Chimera for the addition of hydrogen and charges, and then the file was saved as *.mol2 format for further docking process by using the DOCK 6 program (Nursamsiar et al., 2022).

Preparation of molecules
The 3D structures of caspases 3 and 9 target proteins were displayed in the package Discovery Studio Visualizer v4.5 program.Chain B (caspase 3) and chain A (caspase 9), as the receptor, were separated from water molecules and other of their native ligands and then saved as *.pdb format.By operating the Chimera program, hydrogen, and charges were added to the protein structure, then the file was saved as *.mol2 format for further docking process by using the DOCK 6 program (Nur et al., 2023;Nursamsiar et al., 2020).

Validation of docking method
Validation was conducted by re-docking its native ligands into the active side of the receptors/proteins.The Relative Mean Standard Deviation (RMSD) determined the method's validity.If the score is less than 2, the docking method is valid and ready to be attached to the tested compounds (Nur et al., 2023).

Docking simulation
The docking simulation of tested compounds was performed using the DOCK 6 program and visualized by Chimera (UCSF, 2023).Both caspase 3 and 9 were prepared without hydrogen atoms to determine the spheres around the protein surface.The identification of spheres is based on the location of the natural ligands by the RMSD, about 8 Å each atom.After sphere identification, the grid box area and the size of the natural ligand binding site were determined in the margin of 5 Å.Energy minimization was applied to the optimized ligand compounds from C. canephora and cisplatin as reference compounds before the docking process, while the docking process was next performed by the rigid-docking method.The in silico evaluation parameters regarding the ligand structure's orientation to the protein target, hydrophobic interaction, the formation of hydrogen bonds, and the grid score of docking processes from each ligand were evaluated.

Statistical analysis
Data were expressed as the mean ± standard deviation of at least two experiments and analyzed using the GraphPad Prism 5 ( 2007) program (GraphPad Software, Inc.).Normality was tested using the Kolmo Gorov-Smirnov test, followed by the Bartlett or maximum F test to assess homogeneity.Duncan's multiple-range assay was used to analyze the kinetic characterization data of the culture medium.When significant differences were found, comparisons of means between treatments were performed using simple categorical analysis of variance (ANOVA) plus Dunnett's parametric post hoc test.P<0.05 was indicated as statistically significant.

NMR data of isolated compounds
Compound 1 was isolated as a white amorphous solid.The molecular formula is C30H51O, and it was generated using ESI-MS [M+H]+ at m/z 427.50.The UV spectrum showed only one absorbance peak at 203.53 nm, indicating the absence of a conjugated system of double bonds.The FTIR spectrum showed a broad absorbance peak at ῡ of 3367 cm -1 , indicating the stretching vibration of the O-H bond, and the vibration band of C-O bending was observed at 1040 cm -1 .The absorbance peak at 761 cm -1 indicated the out-of-plane bending vibration of O-H.The Csp3-H stretching bands appeared at 2943 and 2870 cm -1 .
Bending vibrations of Csp3-H of the methyl groups gave absorbance at 1457 (asymmetrical) and 1382 cm -1 (symmetrical), while the methylene rocking bending vibration band appeared at 720 cm -1 .The olefinic group was detected in this compound, shown by an absorbance peak of symmetrical C=C stretching at 1641 cm -1 .The absorption band at 3072 cm -1 indicated the vibration of the C sp2-H stretch, and the peak at 882 cm -1 explained that this compound contains vinyltype double bond 1 and exhibited 30 carbons, which were identified using 13 C NMR.It was revealed that the compound was formed by seven methyls, 11 methylenes, six methines, and six quarternary carbons.The Double Bond Equivalent (DBE) value of six indicated that this compound was a pentacyclic triterpene with only one double bond.A carbon at δC 78.9 (C-3) showed the presence of an oxygenattaching carbon, while the sole double bond in the molecule was shown by a pair of carbons at δC 150.9 (C-20) and 109.4 (C-29).Seven singlet methyls were detected in this compound, suggesting that all the methyl groups were attached to quarternary carbons.From the 1 H NMR, it was observed that there was a proton at δH 3.19 (dd, J = 10.8 and 5.4, H-3) integrating for one proton, indicating a proton attached to an oxygen-attaching carbon, which in this case was C-3.In addition, there were no carbon and proton signals with chemical shifts indicating a methoxy group.Thus, it can be concluded that C-3 attached a hydroxyl (-OH) instead of a methoxy group.Two highly deshielded protons at δH 4.70 and 4.57 (H-29) were attached to the same carbon (C-29).This explains that the highly deshielded carbon was a methylene group forming a double bond with a quarternary carbon (C-20).When a double bond in a triterpene molecule is formed by methylene and a quarternary carbon, it is confirmed outside the ring system.This skeletal characteristic belongs to the lupane-type triterpene lupeol (Rosandy et al., 2021).See Fig. S1.
Compound 2 was obtained as a colorless needle crystal.Only one absorbance peak was observed in the UV spectrum at 204.03 nm.This indicated the absence of a conjugated double bond in the molecule.The FTIR spectral data showed the stretching vibration band of O-H at 3431 cm -1 , C-O bends at 1052 cm -1 , and the out-of-plane bending vibration of O-H at 803 cm -1 .The peaks of stretching Csp3-H appeared at 2958, 2937, and 2869 cm -1 , while the bending vibrations of Csp3-H of the methyl groups gave absorption at 1464 (asymmetrical) and 1380 cm -1 (symmetrical).An absorbance peak of symmetrical stretching C=C at 1645 cm -1 indicated the existence of the olefinic group in this compound.The structure of 2 was further determined using 13 C and 1 H NMR spectral data.The carbon experiment showed that this compound contained 29 carbons of six methyls, nine methylenes, 11 methines, and three quarternary carbons.A methine carbon at δC 71.5 (C-3) indicated a carbon attached to an oxygen atom (oxymethine), revealing a hydroxyl group's existence.The presence of two pairs of highly deshielded carbons at δC 140.5 (C-5), 121.5 (C-6), 138.1 (C-22), and 129.0 (C-23) suggested that there were two double bonds in the molecule.The DBE value of six suggests that this compound was a tetracyclic compound containing two double bonds.The 1 H NMR experiment of 2 (Fig. 1) displayed information on the presence of two singlet methyl signals at δH 0.68 (H-18) and 0.99 (H-19), indicating that they were bound directly to the ring system.Three doublet methyls at δH 0.91 (H-21), 0.81 (H-26), and 0.79 (H-27), and one triplet methyl at δH 0.84 (H-29) were located at the side chain outside the rings of the main skeleton.A proton at δH 3.51 (H-3) was attached to the oxygen-binding carbon (C-3).The occurrence of three highly deshielded protons at δH 5.34 (H-6), 4.99 (H-22), and 5.14 (H-23) indicated the presence of two pairs of double bonds.The typical signal for the olefinic H-6 of the steroidal skeleton was evident from a proton at δH 5.34 integrating for one proton.The similar coupling constants (J = 15.2Hz) of the other olefinic protons, H-22 and H-23, revealed that they were neighbors and were in transposition.Based on the discussion above and data comparison with the literature, compound 2 was identified as stigmasterol (Astuti et al., 2022).See Fig. S2.
Compound 3 was obtained as a colorless needle crystal.Only one absorbance peak was observed in the UV spectrum at 204.03 nm.This indicated the absence of a conjugated double bond in the molecule.The FTIR spectral data showed the stretching vibration band of O-H at 3431 cm -1 , C-O bends at 1052 cm -1 , and the out-of-plane bending vibration of O-H at 803 cm -1 .The peaks of stretching Csp3-H appeared at 2958, 2937, and 2869 cm -1 , while the bending vibrations of Csp3-H of the methyl groups gave absorption at 1464 (asymmetrical) and 1380 cm -1 (symmetrical).An absorbance peak of symmetrical stretching C=C at 1645 cm -1 indicated the existence of the olefinic group in this compound.The structure of compound 3 was further determined using 13 C and 1 H NMR spectral data.The carbon experiment showed that this compound contained 29 carbons of six methyls, nine methylenes, 11 methines, and three quarternary carbons.A methine carbon at δC 71.5 (C-3) indicated a carbon attached to an oxygen atom (oxymethine), revealing a hydroxyl group's existence.The presence of two pairs of highly deshielded carbons at δC 140.5 (C-5), 121.5 (C-6), 138.1 (C-22), and 129.0 (C-23) suggested that there were two double bonds in the molecule.The DBE value of six suggested that this compound was a tetracyclic compound containing two double bonds.The 1 H NMR experiment of 3 (Fig. 1 displayed information on the presence of two singlet methyl signals at δH 0.68 (H-18) and 0.99 (H-19), indicating that they were bound directly to the ring system.Three doublet methyls at δH 0.91 (H-21), 0.81 (H-26), and 0.79 (H-27), and one triplet methyl at δH 0.84 (H-29) were located at the side chain outside the rings of the main skeleton.A proton at δH 3.51 (H-3) was attached to the oxygen-binding carbon (C-3).The occurrence of three highly deshielded protons at δH 5.34 (H-6), 4.99 (H-22), and 5.14 (H-23) indicated the presence of two pairs of double bonds.The typical signal for the olefinic H-6 of the steroidal skeleton was evident from a proton at δH 5.34 integrating for one proton.The similar coupling constants (J = 15.2Hz) of the other olefinic protons, H-22 and H-23, revealed that they were neighbors and were in transposition.Based on the discussion above and data comparison with the literature, compound 3 was identified as ursolic acid (Tan et al., 2019).See Fig. S3.

In silico test of active compounds
Redocking results over receptors of caspases 3 and 9 using DOCK showed RMSD scores of 0.87 Å and 0.64 Å, respectively.RMSD score is one validation reference to assess the accuracy of the docked geometric poses with their crystallographic conformations (Adianingsih et al., 2022;Nur et al., 2023).The result showed valid by RMSD <2 Å (Ramírez and Caballero, 2018).The redocking result of natural ligands caspases 3 and 9 is shown in Fig. 2. It showed that the natural ligand poses due to redocking coincide (shown in gray color) with their crystallographic poses (red).The function score used in the DOCK 6 program is the grid score, and the attachment score is shown in Table 1.The figure of redocking of the natural ligands of caspases 3 and 9 are shown in Fig. 3.
Hydrogen bond spacing had a role to determine the inhibitory activity of a ligand.A strong hydrogen bond might be proportional to a very weak covalent bond.The hydrogen bond with the gap of donoracceptor 1.2 -1.5 Å was categorized as a strong bond, whereas the gap of 1.5 -2.2 Å was categorized as a moderate bond, and 2.2 -3.2 Å was categorized as a weak bond (Adianingsih et al., 2022;Riwu et al., 2022).The interaction and bond distance of each compound with its receptor are shown in Tables 2  and 3. Visualization of molecular attachment to caspases 3 and 9 receptors is shown in Figs. 3 and 4.
A prediction study of compound interaction with caspase-3 (PDB ID: 1PAU) and 9 (PDB ID: 1JXQ) was conducted by molecular attachment study by running the protocol DOCK 6 with the method of flexible ligand docking.The ligand is freely moved in this method, yet the receptor remains rigid.This method is widely used due to its accuracy, which is more than the rigid docking method, where the flexible ligand docking method gives more ligand confirmation of interaction.DOCK molecular attachment was validated based on root mean square deviation (RMSD) from the redocking result of the natural ligand of caspases 3 and 9.

Cytotoxic activities
By the cytotoxicity results of the compounds isolated from C. canephora cascara ethanolic extract, ursolic acid showed the smallest value of IC50 among the other compounds, even smaller than the positive control (cisplatin) on MCF-7 cells (Table 4).In vitro studies of ursolic acid have also been reported to inhibit the progression of breast cancer cells targeting the NF-B pathway (Fitriana et al., 2022;Liu et al., 2015).Non-malignant cells were extracted to calculate the selectivity index (SI) with the IC50 value until a concentration of 1000 µg/mL, which was the maximum concentration evaluated.Based on the results, ursolic acid showed the highest selectivity against Hela cells and MCF-7 cells.Table 4 shows the IC50 level of some compounds derived from C. canephora cascara to HeLa and MCF-7 breast cancer cells.
According to an in vitro cytotoxicity assay of the four compounds isolated from C. canephora cascara, ursolic acid exhibited the smallest IC50 in both HeLa and MCF-7 cells.Ursolic acid is widely known to have activities against diseases, including antiinflammatory, antimicrobial, antidiabetic, and anticancer.Some studies regarding its activity towards breast cancer have been reported, for example, by molecularly suppressing glycolytic metabolism via activation of Sp1/caveolin-1 signaling (Khochapong et al., 2021;Luzak et al., 2022).Since ursolic acid also interacts with caspases 3 and 9 in the present study.It could be understood that the small value of IC50 obtained from ursolic acid is based on its ability to perform cell apoptosis in cancer, particularly breast cancer.The selectivity index (SI) was calculated by dividing the IC50 value of normal cells by the IC50 value of cancer cells.The SI value represents the sample selectivity of the cell lines tested.

CONCLUSION
The IC50 value of the compounds isolated from C. canephora cascara against HeLa and MCF-7 breast cancer and the correlation with in silico results towards interaction to caspases 3 and 9.It can be concluded that all of the compounds have potential for cancer therapy, particularly ursolic acid, which showed the smallest IC50 value against HeLa and MCF-7 breast cancer.More advanced approaches to getting a specific molecular target in cancer cells are needed.

Figure 1 .
Figure 1.Chemical structure of isolated compounds from C. canephora cascara.

Table 1 .
Attachment score of all compounds to caspases 3 and 9.

Table 4 .
IC50 level of some compounds derived from cascara to HeLa and MCF-7 breast cancer cells.