Pentacyclic Triterpenes from Branches of Maytenus robusta and in vitro Cytotoxic Property Against 4 T 1 Cancer Cells

Dois novos friedelanos, 1 e 2, e cinco triterpenos pentacíclicos conhecidos foram isolados dos galhos de Maytenus robusta. Suas estruturas químicas foram identificadas como 3,16-dioxo-29hidroxifriedelano (1), 3-oxo-16b,29-di-hidroxifriedelano (2), 3-oxofriedelano (3), 3b-friedelinol (4), 3,16-dioxofriedelano (5), 3-oxo-29-hidroxifriedelano (6) e 3,16-dioxo-12a-hidroxifriedelano (7). A estrutura e estereoquímica dos triterpenos 1 e 2 foram estabelecidas por infravermelho (IR), ressonância magnética nuclear (NMR) 1D/2D, espectrometria de massas de alta resolução com ionização química à pressão atmosférica (HR-APCIMS) e difração de raios X de pó. A atividade citotóxica in vitro dos triterpenos 1 a 6 foi avaliada frente a células de câncer de mama murino. Os triterpenos 1 e 2 apresentaram atividade citotóxica contra células 4T1 em baixa concentração.


Experimental
All solvents used in this work were of analytical grade.All other chemicals were available commercially at a reagent grade and were used without further purification.

General experimental procedures
Optical rotations were measured on an ADP220 Bellinghan + Stanley Ltd polarimeter.Thermal analyses were carried out in a simultaneous thermogravimetric (TG) and differential thermal analyzer (DTA) Shimadzu DTG-60.The infrared (IR) spectra were recorded on a Perkin Elmer-Spectrum One (attenuated total reflectance, ATR) spectrometer.The 1 H and 13 C nuclear magnetic resonance (NMR) spectra were recorded at 400.129 and 100.613MHz, respectively, as well as correlation spectroscopy (COSY), heteronuclear single quantum coherence (HSQC), heteronuclear multiple bond correlation (HMBC) and nuclear Overhauser effect spectroscopy (NOESY) experiments were performed on a Bruker DRX400 Avance spectrometer, using CDCl 3 or CDCl 3 with two drops of pyridine-d 5 as solvent, with direct or inverse probes and a field gradient.The chemical shifts were registered in ppm (d) using tetramethylsilane (TMS) as the internal standard.The coupling constants (J) were registered in hertz (Hz).High-resolution atmospheric pressure chemical ionization mass spectrometry (HR-APCIMS) spectra were acquired on a Shimadzu LCMS-IT-TOF system.The mass spectrometer was equipped with an atmospheric pressure chemical ionization (APCI) source operating in both the negative and positive modes.Detector voltage: 1.70 kV; interface temperature: 400 o C; CDL temperature: 250 o C; block heater temperature: 200 o C; nebulizing gas flow (N 2 ): 2.5 L min -1 .Analyses were carried out using manual injection.The samples were dissolved in CHCl 3 and then diluted with MeOH.Column chromatography (CC) processes were carried out using silica gel 60 (70-230 mesh).Thin layer chromatography (TLC) was carried out using precoated silica gel plates.
The powder X-ray diffraction (XRD) data were collected on a Shimadzu XRD-7000 under 40 kV, 30 mA, using Cu Ka (1.54056 Å) coupled with a polycapillary unit, graphite monochromator, scanned over an angular range of 3-35° (2q) with a step size of 0.01° (2q) and a time constant of 3 s step -1 .The sample holder was submitted to a spinning speed of 30 cycles per minute to reduce any preferred orientation and to minimize rugosity effects.The final matrix was averaged over 3 independent scans.The powder indexing tool used for peak identification, indexing and automatic space group determination was DICVOL91. 14

Plant material
Branches of M. robusta were collected in June 2010 at Parque Estadual do Itacolomi, Ouro Preto City, Minas Gerais, Brazil.After botanical identification, a voucher specimen of M. robusta was deposited in the Herbário Professor José Badini, and identified by Dr Maria Cristina Teixeira Braga Messias, a botanist at the Departamento de Biodiversidade, Evolução e Meio Ambiente, Universidade Federal de Ouro Preto, under the code OUPR: 25559.

MTT cytotoxicity assay
Dulbecco's modified eagle medium (DMEM) culture medium Gibco ® was used to maintain the 4T1 murine breast cancer cells (obtained from American Type Culture Collection).MilliQ ® water (Simplicity 185, Millipore, Bedford, USA) was used throughout the experiments.The 4T1 cells were employed in the in vitro cytotoxic evaluation.These cells were thawed, submitted to replication, and after having reached an adequate confluence, they were placed on three 96-well microtiter plates (500 per well), totaling 120,000 cells.Each sample was dissolved and diluted using DMSO and hot ethanol (approx.55 °C) 1:9 solution, to reach a concentration ranging from 1 × 10 -3 to 1 × 10 -7 mol L -1 , in a volume of 100 μL.The cells were treated with the solution for 48 hours in DMEM supplemented with sodium bicarbonate (3.7 g L -1 ), penicillin (5,000 units mL -1 ), and 10% m/m sterile fetal bovine serum.The culture was maintained at 37 °C in a humidified atmosphere containing 5% CO 2 .Five replicates were used for an appropriate statistical evaluation.By means of [3-(4,5-dimethyl-2-thiazolyl-2,5-diphenyl)-2H-tetrazolium bromide] MTT assay, metabolic feasibility assays were performed to evaluate the efficacy of the treatment employed.MTT was purchased from Sigma ® .The results were plotted based on percentage of cell death vs. molar concentration of the sample, using the software Prism 6.0.A non-linear regression model (sigmoidal dose-response option) was used to determine the IC 50 values.4T1 cells treated with 2% Triton X-100 were used as positive damaged control 4T1.For negative control, the cells were treated with MTT stock solution added to 100 μL of medium alone.The experiments, in quintuplicate, were performed immediately.The IC 50 (μM) were determined for compounds 1 (9.80), 2 (1.23), 3 (98.10), 4 (58.10), 5 (23.10) and 6 (11.00).

Results and Discussion
Compound 1 was isolated as white solid powder, and its molecular formula, C 30 H 48 O 3 , was established by HR-APCIMS (m/z: 439.3526 [M+H−H 2 O] + , calcd.439.3576).In the IR spectrum of 1 two intense absorptions bands, at 1718 and 1671 cm -1 , and a band at 3470 cm -1 were observed, which were attributed to carbonyl groups and hydroxyl group, respectively. 15he 13 C NMR spectral data of 1 were closely related to those of 3-oxo-29-hydroxyfriedelane and 3,16-dioxofriedelane reported by Mahato and Kundu. 16his information enables to identify 1 as a friedelane derivative with one hydroxyl bonded to C-29 and two oxo  1).The stereochemistry of triterpene 1 is in agreement with the powder X-ray diffractometry results (Figure 2) and it is possible to establish that, in solid state, the ring A of triterpene 1 is in an intermediate state between boat and chair-like conformation.
Compound 2 was isolated as a white solid powder, and its molecular formula, C 30 H 50 O 3, was established by HR-APCIMS (m/z: 441.3722 [M+H−H 2 O] + , calcd.441.3733).In the IR spectrum of 2 an absorption band at 3390 cm -1 , characteristic of hydroxyl group, and at 1712 cm -1 , corresponding to a carbonyl group, were observed. 15he 13 C NMR spectral data of 2 were closely related to those reported for celasdin A by Kuo and Kuo. 17 This information confirms triterpene 2 as a friedelane derivative with an oxo group bonded to C-3 and two hydroxyl groups bonded to C-16 and C-29.In the 1 H NMR spectrum, the six singlet signals (d H 0.70, 0.84, 1.02, 1.11, 1.13 and 1.32) and a doublet (d H 0.88, J 6.6 Hz) observed were attributed to seven methyl groups.In this spectrum, signals at d H 3.36 (m, 2H) and at d H 4.11 (t, 1H, J 8.8 Hz) were also detected.In the 13   Compounds 3 to 7 were isolated as white, amorphous solid materials.The chemical structures of these compounds were confirmed by comparison on thin-layer chromatography (TLC) with authentic samples and through the similarity with 13 C NMR data previously published. 16,18,19he final structure solution of triterpenes 1 and 2 (Figure 2) was obtained by the Rietveld refinement using the lowest energy profile fitting option.This technique has been recently employed to analyze the chemical structure of 3-oxofriedalane. 20The peaks were searched and fitted with a modified Pseudo-Voigt peak profile, performing both modified Pawley and Rietveld refinement to optimize powder diffraction parameters and crystal structure so that the best possible agreement between simulated and experimental powder patterns was achieved.Initially, the pattern was processed for background contribution; peaks were searched and manually inspected, when necessary, to prevent background fluctuations from being considered reflections.The molecule was sketched using ChemSketch®. 21The hydrogen atoms were introduced based on the NMR spectra and the molecule allowed to have its energy minimized by the Monte Carlo method before using it as a model.The crystal structure solution was first obtained without any preferred orientation correction and close-contact penalty.The results were compared with the fit using Marck-Dollese preferred orientation and close-contact penalty algorithm.The result that combined the lowest final energy obtained by Monte Carlo optimization was used as the solution.The corresponding structures were deposited at the CCDC under the numbers 974658 for unit cell parameters: a = 11.5986,b = 22.6631 and c = 5.3911 for space group P212121 (1) and 974657 for unit cell parameters: a = 11.7751,b = 23.9910 and c = 5.4078 for space group P212121 (2).
Even using few micromoles, all tested samples (triterpenes 1 to 6) presented a relevant in vitro cytotoxicity activity against 4T1 cells (Figure 3).In a study developed by Lu et al., 22 4T1 cells were treated with doxorubicin, one of the most effective antitumor drugs, and the observed result was an IC 50 of 3.27 μM.In the present work, using the same cell line, triterpene 2 induced an IC 50 of 1.23 μM.This value is 2.64 times smaller than the IC 50 induced by doxorubicin in 4T1 cells.These results suggest compound 2 as a potential antitumor agent.Compound 2, which presented the highest cytotoxic activity, has an oxo group at carbon C-3 and two hydroxyl groups bonded to C-16 and C-29 respectively.The cytotoxic property of triterpene 1 was slightly lower than compound 2. The difference between them was associated with C-16, which is connected to a hydroxyl group in triterpene 2 and to an oxo group in compound 1.Compounds 2 and 6 differ only by the substitution of the hydroxyl group attached to C-16 for one hydrogen.Compound 6 is about ten times less active against 4T1 cells than triterpene 2 that has hydroxyl group attached to C-16.The cytotoxic activity of the triterpenes subjected to assays againt 4T1 cells seems to be also related to C-29.Compound 5 differs from triterpene 1 by presenting a hydrogen atom instead of a hydroxyl group attached to C-29 and is about ten times less active against 4T1 cells than triterpene 1. Triterpene 3 differs from compound 2 by presenting only hydrogen atoms bonded to C-16 and C-29.This configuration resulted in a reduction of the cytotoxic activity of triterpene 3 about 100 times.Compound 4 differs from compound 3 by presenting a hydroxyl group bonded to C-3 instead of oxo group and 4 was about two times more potent than 3.
According to the results obtained for compounds 1 to 6, it is possible to establish that the presence of hydroxyl groups bonded to C-16 and C-29 may contribute in the increase of cytotoxicity of friedelane-type triterpenes against 4T1 cells.Moreover, hydroxyl group bonded to C-3 also contributed to increase the cytotoxicity.However, more detailed studies can reveal if this behavior is exclusive of hydroxyl group or if other organic groups can also act the same way.Detailed studies regarding the structure-activity relationship, including 3D-QSAR models, can help to evaluate the role of these groups in the cytotoxicity activity, since the mechanism of antitumor activity of friedelanetype triterpenes was not found in the literature.The results obtained in this work along with more detailed studies of the mechanisms of cell death (necrosis, apoptosis) may contribute to a better assessment of the potential of PCTTs in the treatment of cancer.
C NMR spectra, signals at d C 74.40, 74.45 and at 212.78 were visualized and associated with two carbinolic groups and one carbonyl group respectively.By the analysis of HSQC contour map, the correlation between the signal at d H 3.36 with the signal of C-29 (d H 74.45) was observed.Correlations among the signal of H-29 (d H 3.36) with the signals of C-19 (d C 30.36), C-20 (d C 33.14), C-21 (d C 27.51) and C-30 (d C 25.71) were observed in HMBC contour map.In this map, correlations between the signal of H-16 at d H 4.11 with the signals of C-15 (d C 44.28), C-17 (d C 36.45), C-18 (d C 44.09), C-22 (d C 36.45) and C-28 (d C 25.44) were also visualized.In the NOESY contour map, correlations of H-16 (d H 4.11) with the signal of H-27 (d H 1.02) and H-7 (eq, d H 1.46) were detected.These correlations enable to confirm that H-16 has a-configuration and the hydroxyl group linked to C-16, has a b-orientation.In the HSQC contour map, the correlation between hydrogen signal at d H 0.88 (d, J 6.6 Hz, H-23) with the carbon signal of C-23 at d C 6.88 was also verified.In the HMBC contour map, correlations of H-23 with the signals of C-3 (d C 212.78), C-4 (d C 58.04) and C-5 (d C 42.14) were detected.These correlations enabled to confirm C-3 as carbonyl group.The analysis of the NOESY contour map enables to establish that rings D and E of triterpene 2 have a boat-like conformation.The conformation of these rings was confirmed mainly by the correlation between H-18 and H-30.Furthermore, correlations between H-16 and H-28 and among H-29 and H-27 were not observed in the NOESY contour map.The complete chemical shift assignment of triterpene 2, in Table 2, was established by detailed analysis of HSQC, HMBC, COSY and NOESY contour maps.The spatial position of atoms, obtained by Rietveld refinement for compound 2, is consistent with the information obtained after analysis of the 1D/2D NMR spectra.It is possible to determine that, in solid state, the hydroxyl attached to C-16 is in beta position in plan A. In low energy reaching, in solid state, ring A is in an intermediate state between boat and chair-like conformation.

Figure 2 .
Figure 2. Rietveld fitted X-ray diffraction for compound 1 (above) and compound 2 (below).The chemical structures derived from the fittings are presented as insets on the diffractogram.

Figure 3 .
Figure 3. Representative curves of surviving 4T1 cells vs. the log of compound molar concentration.The IC 50 value was evaluated through non-linear regression (n = 5).

Table 2 .
NMR spectral data a of 3-oxo-16b,29-dihydroxyfriedelane(2) C 43.24 (C-18), d C 31.14 (C-22) and d C 218.38 (C-16).These NMR data confirmed the location of a second oxo group attached to C-16.A boat-like conformation of ring D and E was established mainly based on the correlation between H-18 and H-30 observed in the NOESY contour map.By detailed analysis of the HSQC, HMBC and NOESY contour maps, it was possible to establish all chemical shift assignments of compound 1 (Table