Copaifera multijuga, Copaifera pubiflora and Copaifera trapezifolia Oleoresins: Chemical Characterization and in vitro Cytotoxic Potential against Tumoral Cell Lines

Copaifera species (Fabaceae) comprises approximately 70 species of large trees, from which 16 can be found in Brazil. The oleoresins obtained from their trunk are widely used in Brazilian folk medicine, which display important antitumoral potential. Chemically, these oleoresins are mainly composed of a mixture of sesquiterpenes and diterpenes. In this paper we are describing the isolation and identification of 12 already known terpenes from oleoresins obtained from three different Copaifera species (C. multijuga, C. pubiflora and C. trapezifolia) and 2 novel diterpenes (ent-16-hidroxy-3,13 clerodadien-15,18-dioic acid and ent-labda-5,13-dien-15-oic acid) from C. trapezifolia. Both new compounds were identified by nuclear magnetic resonance (NMR) spectroscopic (H and C NMR, correlation H-H (COSY), heteronuclear multiple quantum coherence (HMQC) and heteronuclear multiple bond correlation (HMBC)) and by high-resolution electrospray ionization mass spectrometry (HR-ESIMS) analyses. The cytotoxic potential of these oleoresins, their main non-volatile compounds and their volatile compound fractions were evaluated against a panel of tumoral (MCF-7, ACP01, A549, HeLa) and normal cell lines (MCF-10A, GM07492-A) through XTT (tetrazolium salt) and SRB (sulforhodamine B) assays. The novel diterpene ent-labda-5,13-dien-15-oic acid displayed relevant cytotoxic effect against most of the cancer cell lines with mean inhibitory concentration (IC50) values ranging from 3.57 ± 1.12 to 22.56 ± 1.03 μg mL, and a high selectivity level in both assays.


Introduction
The Copaifera L. genus occur throughout South America, Africa and Asia. It belongs to the Fabaceae family and comprises approximately 70 species of large trees, from which 16 can be found in Brazil. 1 The oleoresins obtained from their trunk are widely used in Brazilian folk medicine, and many folk uses were corroborated by researchers, including: anti-inflammatory, anticancer, wound healing, antiparasitic, and antimicrobial properties, among others. [1][2][3] Chemically, these oleoresins are composed of a diversified mixture of terpenoids, mainly sesquiterpenes and diterpenes, which are the major constituents from their volatile and non-volatile fractions, respectively. 4 Among all ethnopharmacological applications that have been described for Copaifera species oleoresins, their antitumor effect stands out since its efficacy has been proven. Lima et al. 5 evaluated the anticancer activity of Copaifera multijuga oleoresin against melanoma cells and its inhibition of lung metastasis. The results of this study showed that this oleoresin and its fractions display tumoricidal activity in the melanoma cell line, once oral treatment of 2.0 g kg -1 reduced tumor growth and its weight by 58 and 76%, respectively. Gomes et al. 6 investigated the antineoplasic activity of Copaifera multijuga oleoresin and its hexanic and chloroformic fractions against ascitic and solid Ehrlich tumor, demonstrating that it promoted inhibition of the solid tumor on paws after three days of oral treatment (150 mg kg -1 ), which was similar to the control group (vincristine 0.5 mg kg -1 ).
Despite of the fact that some studies pointed out the antitumoral efficacy of these natural products, most related scientific research were only performed with crude oleoresins of Copaifera multijuga and its fractions. Moreover, it can also be observed that only a limited number of studies were carried out to evaluate the cytotoxic potential of their constituents, thus denoting the need for further investigations aiming to better establish which compounds are related with the cytotoxic and antitumor potential previously reported for the Copaifera oleoresins. 1 In this regard, and as part of ongoing efforts to explore the chemical and biological properties of Brazilian Copaifera, we are reporting the chemical characterization of the oleoresins obtained from three different species of two different regions of Brazil: C. multijuga, C. pubiflora (from the north) and C. trapezifolia (from southeast), as well as the evaluation of the in vitro cytotoxic potential of diterpenes and sequiterpenes isolated from these oleoresins against a panel of cancer cell lines.

Maintenance and cell culture
The cell lines used were MCF-7 (breast adenocarcinoma), MCF-10A (normal mammary gland), ACP-01 (gastric carcinoma), A549 (lung adenocarcinoma), HeLa (human cervical cancer) and GM07492-A (normal human fibroblast). The cells were stored in liquid nitrogen (-196 °C) in aliquots of 1 × 10 6 cells mL -1 in a freezing solution composed of 90% fetal bovine serum and 10% dimethyl sulfoxide (DMSO). The cells were grown as monolayer cultures in 5 mL of DMEM + HAM-F10 (1:1, v/v) culture medium supplemented with 10% fetal bovine serum, 1% stabilized solution of antibiotics penicillin/streptomycin and 0.2% of antibiotic kanamycin solution in 25 cm 2 disposable flasks, and kept at 37 °C in an atmosphere containing 5% CO 2 . The cells were subcultured every two or three days, washed using phosphate buffered saline (PBS 1×) and detached from the inner surface of the culture flask using trypsin. Approximately 1.0 mL of complete culture medium was then added to the flask for trypsin inactivation, and between 50 and 100 µL of the resulting cell suspension were cultured in new vials containing 5 mL of complete culture medium and incubated at 37 °C.

Cell culture and treatment solutions
The cultured cells were trypsinized and plated in 96-well microplates at a concentration of 1 × 10 4 cell per well in DMEM + HAM-F10 medium (1:1: v/v) supplemented with 10% fetal bovine serum and antibiotics. After 24 h of incubation at 37 °C in an 5% CO 2 incubator, the cell cultures were treated with concentrations of 3.9 to 500.0 µg mL -1 of oleoresins and 7.8 to 1000.0 µM of isolated compounds. The oleoresins and compounds were solubilized in DMSO just prior to use. DMSO at 1% in culture medium was the vehicle control, and doxorubicin was used as positive control (PC). The negative control received no treatment. Each experiment was performed in triplicate.

XTT assay
The XTT assay was performed using the Cell Proliferation Kit II 24 h after the treatments. The culture medium was removed from the plates and the wells were washed with PBS (1×). Subsequently, 100 µL of Dulbecco's modified eagle medium (DMEM) medium without red phenol containing 10% of the XTT solution (tetrazolium salt solution and electron coupling solution in the ratio 50: 1 (v/v)) were added to each well and the cells were incubated for 4 h at 37 °C in a 5% CO 2 incubator. After incubation, the absorbance was read in a 96-well plate reader at 492 nm (reference 690 nm).

Sulforhodamine B (SRB) assay
The cell medium was completely discarded for the SRB assay and the wells were extensively washed with PBS (1×). The cells were then fixed with 25 µL of 50% (m/v) trichloroacetic acid for 1 h at 4 °C and the wells were then washed four times with 100 µL distilled water and allowed to dry at room temperature. Subsequently, the cellular proteins were stained by adding 50 µL of 0.4% SRB (m/v) solution in 1% acetic acid (v/v) solution to each well for 15 min at room temperature. The excess dye was removed using 1% acetic acid solution and the plates were left to dry at room temperature. Finally, the protein-bound dye was dissolved in 150 µL of 10 mM Tris-HCl buffer (pH 10.5) by stirring and quantified in a spectrophotometer at 550 nm with reference to 650 nm. As in the XTT assay, the cell viability percentage was calculated considering the negative control with 100% viability.

Results analysis
The mean inhibitory concentration (IC 50 ) was determined by calculating the non-linear regression using the GraphPad Prism 6.07 program. 7 The selectivity index (IS) was also used for data analysis and indicates selectivity of the compound between a neoplastic and a normal cell line, as well as its potential use in clinical trials. Thus, in this study, IS corresponds to the ratio between the IC 50 values of the compound in the normal cells (MCF-10A or GM07492-A) and cancer cells, following the equations: IS = IC 50 MCF-10A / IC 50 MCF-7 and IS = IC 50 GM07492-A / IC 50 of all other tumoral cell lines.

Computational details
The geometry optimizations, vibrational frequencies and orbital molecular calculations were performed using the Gaussian 16 package 8 at the B3LYP/6-311++G(2d,p) theory level. [9][10][11][12] The nature of the stationary point was determined by inspecting the eigenvalues obtained through the Hessian matrix. The molecular volume and surface area were obtained using the HyperChem Professional 8.0 software. 13 The lipophilicity (expressed as logP octanol/water value) was predicted using the free SwissADME web tool. 14-16 Molecular superimposition was carried out using the Pymol 2.3 software. 17
Compound Ct1 was isolated as a yellow solid. The HRESIMS analysis of Ct1 gave an [M + H] + ion at  m/z 351.2183 (calcd. 351.2171), which matched with the molecular formula C 20 H 32 O 5 . The 1 H and 13 C NMR data acquired for Ct1 were similar to those previously reported for ent-hardwickiic acid (Cp1), 26 a clerodanetype diterpene, commonly found in Copaifera species oleoresins. 30 Therefore, the NMR data reported for Cp1 were then used to propose its chemical structure.
The 1 H and 13 C NMR data of Ct1 denoted the typical chemical shifts concerning the trans-decalin ring of Cp1, which correspond to the three methyl protons H-17 (d H 0.84, d, J 6.5 Hz), H-19 (d H 1.26, s) and H-20 (d H 0.78, s), the H-3 vinylic proton signal at d H 6.87 (dd, J 4.2, 3.0 Hz) and the carboxylic acid moiety at C-18 (d C 172.3). However, in the 1 H NMR spectra of Ct1 it was not possible to observe the presence of the chemical shifts related to the furan group of the ent-hardwickiic acid, thus denoting the main chemical difference between these diterpenes.
In addition to the signals described above, the 1 H NMR spectra also evidenced the presence of a second vinylic proton in the chemical structure of Ct1 resonance at d  were attributed by HMQC spectrum analysis to C-7, C-8 and C-10, thus confirming a double bond between C-5 and C-6. Compound Ct2 was therefore identified as ent-5,13-labdadiene-15-oic acid, an isomer of ent-copalic acid, which has not been previously reported in the scientific literature. 1 H, 13   [α] D 25 +79.8 o (c 0.0157, CH 3 OH). Figure 3 represents the main HMBC correlations of Ct1 and Ct2.
Several authors have considered the diterpene ent-copalic acid (Cm4) as the chemical marker of the Copaifera genus, once this metabolite has been found in oleoresins of all species of this genus. 31 Despite of that, we have isolated a positional isomer of this diterpene from C. duckei oleoresin, which has never been isolated from Copaifera species before, and in which the exocyclic double bond between C-8/C-17 of Cm4 changed for C-7/C-8. 3 Additionally, another positional isomer, in which the double bound is located between C-5/C-6 (Ct2), was isolated and identified. Considering that the main analytical techniques commonly used to characterize Copaifera oleoresins are not able to distinguish these isomers, 4 the discovery of yet another positional isomer of Cm4 reinforces the need to establish novel diterpenes as chemical markers of Copaifera oleoresins for further application in the quality control of these important biologically active natural resources.
Overall, the results depicted in Tables 1 and 2 show that all three oleoresins displayed cytotoxic activity against most of the tumoral cell lines, thus collaborating with their ethnopharmacological application for the treatment of human cancer. 1 In-depth analysis of these results allowed to point out that OCP displayed promising IC 50 values 32 against breast adenocarcinoma cells (MCF-7; IC 50 values of 10.27 ± 1.11/41.85 ± 1.07 µg mL -1 ) and tumoral gastric cancer cells (ACP01; 21.17 ± 1.05/28.75 ± 1.06 µg mL -1 ) Considering the results displayed by the oleoresins against the tumoral cell lines, 32 it became relevant to individually evaluate the cytotoxic potential of their chemical constituents, aiming to select the main metabolites responsible for this activity. The fractions containing the volatile terpenoids (OCm1, OCp1 and OCt1) were investigated for the first time and as shown in Tables 1 and 2, these fractions displayed IC 50 values that can be considered promising according to Suffness and Pezzuto. 32 However, it was noted that these volatile terpenoids are not selective considering their selective indexes lower than 2, once OCm1, OCp1 and OCt1 promoted in vitro antiproliferative effects against tumoral and normal cells viability with very close IC 50 values.
Regarding the 14 non-volatile terpenoids isolated and identified in this study, most IC 50 values displayed by these compounds were above the criteria to be considered as a lead compounds in the discovery of new anticancer agents (IC 50 values higher than 10.0 µg mL -1 , 32 except for the new diterpene Ct2, which displayed relevant cytotoxic effect against most of the tumoral cell lines (IC 50 values ranging from 3.57 ± 1.12 to 22.56 ± 1.03 µg mL -1 ; Tables 1 and 2) and a high selectivity level in both XTT and SRB assays.
As described above, the phytochemical study performed with these oleoresins allowed to isolate and identify three different double bond position isomers (Cm4, Cp2 and Ct2). It is interesting to point out that the diterpenes Cm4 and Cp2 displayed moderate cytotoxic activity and very Computational calculations were used to perform a qualitative comparative analysis for compounds Ct2, Cp2 and Cm4, in order to correlate possible structural differences between the compounds and their biological activities. Preliminary quantitative structure-activity relationship (QSAR) approaches developed by our research group and involving diterpenes indicated that in vitro cytotoxicity of ent-kaurenoic acid derivatives against human breast carcinoma cell line may be related to its logP (lipophilicity), as well as to electronic parameters (HOMO and HOMO-1 molecular orbital energies), thus, suggesting that the interaction between these derivatives and the cell involves charge displacement and can occur by any kind of intermolecular interaction. 33 Figure S13 and Table S1, respectively). Interestingly, for the three compounds, the HOMO orbital is mainly distributed over the double bond in the trans-decalin ring, while LUMO and HOMO-1 are localized more in the side chain containing the carboxylic acid group. Although the distribution of molecular orbitals is similar, the DE gap indicates that Ct2 is more polarizable than Cp2 and Cm4, a factor that may usually be associated with a high chemical reactivity and low kinetic stability. 36 Regarding lipophilicity, no significant differences in logP values were observed between the three compounds. Spatial characteristics can also be important to explain the great difference between the IC 50 values shown by these compounds. In this sense, Ct2 presented smaller surface area, molecular volume and dipole moment than Cp2 and Cm4 (Table S1, SI section). A molecular superimposition evaluation considering the equilibrium geometries of these compounds indicated that Cp2 and Cm4 presented a very close spatial conformation, which is distinct from that observed for Ct2 (Figure 4). This factor can also be relevant to explain the higher activity of Ct2 in comparison with the results shown by Cp2 and Cm4.
Finally, it is important to report that the apparently better results obtained in the SRB assay (Table 2) in comparison with those of the XTT assays may be due to the fact that SRB stains the total protein content of the cell and does not depend on the cellular metabolism. Therefore, lower IC 50 values might be observed in the SRB assay in comparison with the XTT assay. 37 Moreover, the XTT assay have the disadvantage of being more susceptible to variations in cellular levels of nicotinamide adenine dinucleotide (NADH), glucose and other factors than the SRB assay, which is one of the reasons why it was adapted by the National Cancer Institute for its screening programme. 38

Conclusions
The present study lead to the isolation and identification of 14 diterpenes from three different Copaifera oleoresins (C. multijuga, C. pubiflora, and C. trapezifolia). Compounds ent-16-hidroxy-3,13 clerodadiene-15,18-dioic acid (Ct1) and ent-labda-5,13-dien-15-oic acid (Ct2) have not been previously reported in the scientific literature. Our findings also revealed the existence of a third position isomer of the metabolite that is considered by the literature as the chemical marker of the Copaifera genus (copalic acid; Cm4), thus denoting the need to establish additional diterpenes as chemical markers of Copaifera oleoresins for further application in the quality control of these important biologically active natural resources. It was possible to conclude through the cytotoxic studies that the most active compound was the new diterpene Ct2, which displayed relevant cytotoxic effect against most of the tumoral cell lines and a high selectivity level in both XTT and SRB assays.

Supplementary Information
Supplementary data (NMR and computational details) are available free of charge at http://jbcs.sbq.org.br as PDF file.

Author Contributions
LJC and MFCS performed the isolation and identification of 14 terpenoids from oleoresins obtained from three different Copaifera species (C. multijuga, C. pubiflora and C. trapezifolia); LJC, TOT and MOG were responsible for the cytotoxicity assays; JJMS and AEMC performed the identification of the main volatile compounds by GC-MS; RPO and RLTP were responsible for the computational investigations; SRA, RCSV, JKB and RAS designed the study, supervised the laboratory work and contributed to critical reading of the manuscript.