Constituents of the Roots of Dichapetalum pallidum and Their Anti-Proliferative Activity

As part of our search for bioactive compounds from the Dichapetalaceae, repeated chromatographic purification of the roots of a hitherto unexamined species, Dichapetalum pallidum, led to the isolation of the newly occurring 7-hydroxydichapetalin P (1) and the known dichapetalins A (2) and X (3). Also isolated were the known compounds friedelin-2,3-lactone (4), friedelan-3-one (6), friedelan-3β-ol (7) and pomolic (8), as well as the dipeptide aurantiamide acetate (5). The compounds were characterized by direct interpretation of their IR, 1D NMR and 2D NMR spectral data and by comparison of their physico-chemical data, including their chromatographic profiles, with the literature and authentic samples in our compound library for the genus Dichapetalum. The compounds were assayed for their anti-proliferative activities against the human T-lymphocytic leukemia (Jurkat), acute promyelocytic leukemia (HL-60) and T-lymphoblast-like leukemia (CEM) cell lines. Overall, dichapetalin X showed the strongest (3.14 μM) and broadest cytotoxic activities against all the leukemic cell lines tested, exhibiting even stronger activities than the standard compound, curcumin.


Introduction
Dichapetalum Thouars (Dichapetalaceae) is a rich source of triterpenoids and several species of the genus have produced a variety of this class of secondary metabolites. Although the genus was initially noted for being lethal to livestock and even humans, recent research has focused on the less toxic plant species and has led to the discovery of a novel class of phenylpyranotriterpenoid compounds, the dichapetalins. This unique class of triterpenoids in which a 2-phenylpyrano moiety is annellated to ring A of a 13,30-cyclodammarane-type triterpenoid is believed to be biogenetically derived from the condensation of the triterpenoid with a C 6 -C 2 moiety, which is probably derived from the shikimic acid pathway. The main distinguishing feature of the dichapetalins is the nature of the side chain on the cyclopentane ring D which may comprise a lactone, methyl ester, spiroketal, lactol, acetyl and furan moieties. Originally known for their cytotoxic activity, dichapetalins with a spiroketal side chain Molecules 2017, 22, 532; doi:10.3390/molecules22040532 www.mdpi.com/journal/molecules have exhibited the most potent antineoplastic properties. Moreover, they have also been reported to have antioxidant, antihelminthic, antifeedant, nematicidal, antifungal, anti-HIV, nitric oxide-and acetylcholinesterase-inhibitory activities [1][2][3][4][5][6][7]. Dichapetalum pallidum (Oliv.) Engl. is distributed in west and central Africa. It has hitherto not been examined, neither for its chemical constituents nor for bioactivity. The main documented ethnobotanical applications are the use of the seeds as food in Togo and the leaves in the treatment of diarrhea [8]. As part of our on-going studies on the less common and less toxic species of the genus, in our search for new and more potent cytotoxic dichapetalins, we report the isolation and characterization of three dichapetalins from the roots of D. pallidum, namely 7-hydroxydichapetalin P (1), dichapetalin A (2) and the recently reported dichapetalin X (3) from D. filicaule [7]. Also isolated were friedelin-2,3-lactone (4), the dipeptide aurantiamide acetate (5), friedelan-3-one (6), friedelan-3β-ol (7) and pomolic acid (8) (Figure 1). Although dichapetalin P has been reported from D. zenkeri [5], its 7-hydroxy derivative (1) has hitherto not been reported as a naturally occurring compound. The dichapetalins exhibited better anti-proliferative activity against the human T-lymphocytic leukemia (Jurkat), acute promyelocytic leukemia (HL-60) and T-lymphoblast-like leukemia (CEM) cell lines than curcumin. The results further substantiate this class of dammarane-type triterpenoids as potential antitumor agents. also been reported to have antioxidant, antihelminthic, antifeedant, nematicidal, antifungal, anti-HIV, nitric oxide-and acetylcholinesterase-inhibitory activities [1][2][3][4][5][6][7]. Dichapetalum pallidum (Oliv.) Engl. is distributed in west and central Africa. It has hitherto not been examined, neither for its chemical constituents nor for bioactivity. The main documented ethnobotanical applications are the use of the seeds as food in Togo and the leaves in the treatment of diarrhea [8]. As part of our on-going studies on the less common and less toxic species of the genus, in our search for new and more potent cytotoxic dichapetalins, we report the isolation and characterization of three dichapetalins from the roots of D. pallidum, namely 7-hydroxydichapetalin P (1), dichapetalin A (2) and the recently reported dichapetalin X (3) from D. filicaule [7]. Also isolated were friedelin-2,3-lactone (4), the dipeptide aurantiamide acetate (5), friedelan-3-one (6), friedelan-3β-ol (7) and pomolic acid (8) (Figure 1). Although dichapetalin P has been reported from D. zenkeri [5], its 7-hydroxy derivative (1) has hitherto not been reported as a naturally occurring compound. The dichapetalins exhibited better anti-proliferative activity against the human T-lymphocytic leukemia (Jurkat), acute promyelocytic leukemia (HL-60) and T-lymphoblast-like leukemia (CEM) cell lines than curcumin. The results further substantiate this class of dammarane-type triterpenoids as potential antitumor agents.

Characterization of Compounds
Repeated silica gel column chromatographic separation of the ethyl acetate extract of the roots of D. pallidum afforded a total of seven triterpene derivatives and a dipeptide, aurantiamide acetate (5). Among the triterpenoids were dammarane-type 7-hydroxydichapetalin P (1), which is reported for the first time, dichapetalin A (2) and the recently reported dichapetalin X (3) from D. filicaule [7].
Compound 1 was obtained as a white crystalline solid. Five methyl protons were identified in the 1 H-NMR spectrum within the range δ H 0.92 to δ H 2.03, the latter being assigned to an acetoxymethyl.
The 1 H-NMR spectrum exhibited signals due to five aromatic protons resonating at δ H 7.24 (m, H-4 ), δ H 7.33 (m, H-3 , 5 ) and δ H 7.35 (m, H-2 , 6 ). The 13 C-NMR spectrum showed 38 signals. Two ester carbonyl functionalities were inferred by chemical shifts at δ C 173.9 and δ C 170.4 and supported by absorptions at 1780 and 1745 cm −1 in the IR spectrum. The intensity of the olefinic/aromatic signals at δ C 125.8 and δ C 128.4 appeared to be twice that of the other signals, indicating the presence of two sets of equivalent carbons. The HSQC data revealed 11 quaternary carbons interspersed within both the sp 2 -and sp 3 -hybridized regions, nine methylene, 13 methine and five methyl carbons. Thus, the expected total number of carbons present in 1 was 40.
Compound 1 was obtained as a white crystalline solid. Five methyl protons were identified in the 1 H-NMR spectrum within the range δH 0.92 to δH 2.03, the latter being assigned to an acetoxymethyl. The 1 H-NMR spectrum exhibited signals due to five aromatic protons resonating at δH 7.24 (m, H-4″), δH 7.33 (m, H-3″, 5″) and δH 7.35 (m, H-2″, 6″). The 13 C-NMR spectrum showed 38 signals. Two ester carbonyl functionalities were inferred by chemical shifts at δC 173.9 and δC 170.4 and supported by absorptions at 1780 and 1745 cm −1 in the IR spectrum. The intensity of the olefinic/aromatic signals at δC 125.8 and δC 128.4 appeared to be twice that of the other signals, indicating the presence of two sets of equivalent carbons. The HSQC data revealed 11 quaternary carbons interspersed within both the sp 2 -and sp 3 -hybridized regions, nine methylene, 13 methine and five methyl carbons. Thus, the expected total number of carbons present in 1 was 40.
Compounds 2 and 3 were identified as dichapetalins A and X, respectively, by comparison of their physico-chemical and spectroscopic data with published results [1,7] as well as by co-TLC with authentic samples in several solvent systems. Similarly, compounds 6-8 were also characterized respectively as friedelan-3-one, friedelan-3β-ol and pomolic acid on the basis of their physic-chemical and spectroscopic properties and comparison of the data with published data [7,9,10].
Compound 5 was obtained as a white amorphous solid. Its IR spectrum exhibited bands typical of an amide (3314 and 1661 cm −1 ), an acetoxy (1726 and 1261 cm −1 ), a monosubstituted phenyl (745 and 695 cm −1 ), and methyl, methylene and methine (2961, 2919, 2858 cm −1 ) groups, respectively. The presence of an acetoxymethyl was confirmed by a 3H singlet at δH 2.03 in the 1 H-NMR spectrum. Signals for two amide NH groups were observed at δH 6.75 and δH 5.98. The 13 C-NMR spectrum showed signals for three carbonyls, that is an acetoxy (δC 170.7), an aliphatic amide (δC 170.2) and an aromatic amide (δC 167.1), two benzylic methylenes and an oxymethylene signals at δC 37.8, δC 38.4 and δC 64.4, respectively, while two methines were found at δC 49.4 and δC 54.9, including the acetoxymethyl at δC 20.8. There were nine aromatic methines and three quaternary carbons within the range δC 126.7-δC 136.7, suggesting the presence of three mono-substituted phenyl rings. This was corroborated by 15 aromatic protons (δH 7.12-7.78) in the 1 H-NMR spectrum. Comparison of the data with literature [12] led to the identification of compound 5 as the dipeptide aurantiamide acetate (Figure 1), previously isolated from red algae and several families of higher plants [13], but reported for the first time in the Dichapetalaceae. Thus, compound 1 was unequivocally identified as 7-hydroxydichapetalin P, the reduced form of dichapetalin P, which has hitherto not been reported as a naturally occurring dichapetalin. Moreover, 1 exhibited peaks in its IR spectrum and characteristic chromatographic profile with respect to its reaction with vanillin reagent as previously reported for the dichapetalins [5]. Based on the close similarity of its NMR data to previously reported congeners, compound 1 is most likely to share a relatively similar stereochemistry as its congeners.
Compounds 2 and 3 were identified as dichapetalins A and X, respectively, by comparison of their physico-chemical and spectroscopic data with published results [1,7] as well as by co-TLC with authentic samples in several solvent systems. Similarly, compounds 6-8 were also characterized respectively as friedelan-3-one, friedelan-3β-ol and pomolic acid on the basis of their physic-chemical and spectroscopic properties and comparison of the data with published data [7,9,10]. Together with the DEPT and HSQC assignments, eight methyls, 11 methylenes, four methines and seven quaternary carbons including a carbonyl (δ C 175.6) were identified, suggesting a triterpenoid skeleton. Further analyses of the NMR data showed a good comparison with the literature [11] for friedelin-2,3-lactone (1), (Figure 1). The HMBC correlations of C-4 (δ C 84.9) to Me-23 (δ H 1.20) and Me-24 (δ H 0.89), and C-10 (δ C 64.9) and C-3 (δ C 175.6) to H-2a (δ H 2.52) and H-2b (δ H 2.63) supported the position of the lactone. Although the related compounds friedelan-3-one and friedelan-3β-ol are replete in the genus, this is the first report of the isolation of the lactone from the genus.
Compound 5 was obtained as a white amorphous solid. Its IR spectrum exhibited bands typical of an amide (3314 and 1661 cm −1 ), an acetoxy (1726 and 1261 cm −1 ), a monosubstituted phenyl (745 and 695 cm −1 ), and methyl, methylene and methine (2961, 2919, 2858 cm −1 ) groups, respectively. The presence of an acetoxymethyl was confirmed by a 3H singlet at δ H 2.03 in the 1 H-NMR spectrum. Signals for two amide NH groups were observed at δ H 6.75 and δ H 5.98. The 13 C-NMR spectrum showed signals for three carbonyls, that is an acetoxy (δ C 170.7), an aliphatic amide (δ C 170.2) and an aromatic amide (δ C 167.1), two benzylic methylenes and an oxymethylene signals at δ C 37.8, δ C 38.4 and δ C 64.4, respectively, while two methines were found at δ C 49.4 and δ C 54.9, including the acetoxymethyl at δ C 20.8. There were nine aromatic methines and three quaternary carbons within the range δ C 126.7-δ C 136.7, suggesting the presence of three mono-substituted phenyl rings. This was corroborated by 15 aromatic protons (δ H 7.12-7.78) in the 1 H-NMR spectrum. Comparison of the data with literature [12] led to the identification of compound 5 as the dipeptide aurantiamide acetate (Figure 1), previously isolated from red algae and several families of higher plants [13], but reported for the first time in the Dichapetalaceae.

Effect of Compounds 1-5 on Cell Proliferation of Leukemia Cancer Cell Lines
Human T-lymphocytic leukemia (Jurkat), acute promyelocytic leukemia (HL-60) and T-lymphoblast-like leukemia (CEM) cell lines were employed in the assay of the anti-proliferative activities of compounds 1-5. Dichapetalins A and X were more active against CEM and HL-60 cells than 7-hydroxydichapetalin P and curcumin.

Effect of Compounds 1-5 on Cell Proliferation of Leukemia Cancer Cell Lines
Human T-lymphocytic leukemia (Jurkat), acute promyelocytic leukemia (HL-60) and T-lymphoblast-like leukemia (CEM) cell lines were employed in the assay of the anti-proliferative activities of compounds 1-5. Dichapetalins A and X were more active against CEM and HL-60 cells than 7-hydroxydichapetalin P and curcumin.
In the initial screening of compounds 1-5 towards Jurkat cell lines, the dichapetalins strongly inhibited the growth of the Jurkat cells. The inhibitory activities (IC50 values) of 7-hydroxydichapetalin P (1), dichapetalin A (2) and dichapetalin X (3) were 2.66, 2.97 and 1.80 μM, respectively, while that for standard curcumin was 2.54 μM. On the other hand, friedelin-2,3-lactone (4) and aurantiamide acetate (5) showed rather weak inhibitory activities (IC50 > 100 μM) (Figure 4). Further testing of the dichapetalins in HL-60 and CEM cells revealed weaker inhibitory potency compared to those in Jurkat cells. In HL-60 cells, measured IC50 values were 29.72, 11.19 and 5.56 μM, respectively, for the dichapetalins and 16.76 μM for curcumin ( Figure 5). This result confirms a previous assay in which the chemosensitivity of dichapetalin A on HL-60 was reported as 11.0 μM 5 . The CEM cells were slightly more sensitive to the dichapetalins than HL-60 cells, with IC50 values of 23.92, 9.62 and 3.14 μM, respectively, for 7-hydroxydichapetalin P, dichapetalin A and dichapetalin X, but more resistant to curcumin (IC50 = 19.96 μM) ( Figure 6).  Further testing of the dichapetalins in HL-60 and CEM cells revealed weaker inhibitory potency compared to those in Jurkat cells. In HL-60 cells, measured IC 50 values were 29.72, 11.19 and 5.56 µM, respectively, for the dichapetalins and 16.76 µM for curcumin ( Figure 5). This result confirms a previous assay in which the chemosensitivity of dichapetalin A on HL-60 was reported as 11.0 µM 5 . The CEM cells were slightly more sensitive to the dichapetalins than HL-60 cells, with IC 50 values of 23.92, 9.62 and 3.14 µM, respectively, for 7-hydroxydichapetalin P, dichapetalin A and dichapetalin X, but more resistant to curcumin (IC 50 = 19.96 µM) ( Figure 6).     Overall, dichapetalin X (3) showed the strongest and broadest cytotoxic activities against all the leukemic cell lines tested, exhibiting even stronger activities than curcumin. Although 7-hydroxydichapetalin P (1) was in all cases a weaker inhibitor than curcumin, their IC 50 values were comparable, indicating that the dichapetalins isolated in this study are cytotoxic to all the leukemia cell lines tested. This suggests that the compounds are potential anticancer agents, corroborating earlier studies that have indicated that the dichapetalins have anticancer properties [2,3,5,6]. So far, dichapetalins M and P, which contain an intact lactone in the side chain and a ketocarbonyl group at C-7, have exhibited the most potent anti-proliferative activities, with dichapetalin M exhibiting nanomolar cytotoxicity towards human colorectal carcinoma (HCT-116) [5]. These two structural features have therefore been implicated in the pharmacological action of this class of compounds. Dichapetalin X and 7-hydroxydichapetalin P have the intact lactone but lack the 7-carbonyl, while both moieties are absent in dichapetalin A. Thus, structurally, 7-hydroxydichapetalin P and dichapetalin P are identical, with the exception of the substituent at the 7 position. Dichapetalin M is 6-hydroxydichapetalin P. Dichapetalin X, on the other hand, has a different side chain where the acetoxy group in the dichapetalin P derivatives is replaced with a butyloxy substituent. These findings suggest the need to carry out further studies to confirm the proposed structure-activity relationship and to determine the mechanisms of action of this intriguing class of triterpenoids.

Conclusions
The identification of dichapetalins in D. pallidum, hitherto unexamined for its chemical constituents, is of chemotaxonomic significance. All the isolated dichapetalins exhibited micromolar anti-proliferative activity towards three leukemia cancer cell lines, namely the Jurkat, HL-60 and CEM cell lines. Dichapetalin X, recently isolated as a novel dichapetalin from D. filicaule with potential antihelminthic activity, but not assayed for its anti-proliferative activity, was the most potent cytotoxic agent against all the leukemic cell lines tested. It exhibited even stronger activity towards these cell lines than the standard compound used, curcumin. The presence of a spiroketal moiety in its side chain further supports the suggested relationship of this structural feature to the potent anti-proliferative activity in this class of compounds.
In order to establish the identity of the pharmacophores, it is important to investigate the mechanisms underlying the observed anti-proliferative properties of the dichapetalins and possibly incorporate these pharmacophores onto smaller scaffolds as a means of synthesizing simple but potent anticancer agents.

Plant Material
Whole roots and stem of Dichapetalum pallidum were collected from Asenanyo Forest Reserve (Ashanti Region, Ghana) in July 2013 and identified by Mr. John Ntim-Gyakari, formerly of the Forestry Commission, Kumasi. A voucher specimen (DPA001) was deposited at the National Herbarium, Botany Department, University of Ghana, Legon.

Cytotoxicity Assay
The cytotoxic activities of the compounds towards Jurkat, HL-60, CEM and MCF-7 cancer cell lines were determined using the MTT assay [14]. The cells were cultured in RPMI supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin. The cells were then seeded in a 96-well plate at 1 × 10 4 cells/well and treated with varying concentrations of the compounds (0-100 µg/mL) and maintained in a humidified incubator at 37 • C in the presence of 5% CO 2 incubated for 72 h. Twenty microliters of 2.5 mg/mL MTT solution in phosphate buffered saline (PBS) was added to each well and incubation was continued for 4 h. Acidified isopropanol containing 1% Triton-X was added to each well, and subsequently the plates were kept in the dark at room temperature to dissolve any formazan crystals formed. Absorbance readings of wells were measured at 570 nm with a microplate reader (Tecan Infinite M200, Grödig, Austria). All experiments were performed in triplicate.