Quassinoids from the Roots of Eurycoma longifolia and Their Anti-Proliferation Activities

A phytochemical investigation on the roots of medicinal plant Eurycoma longifolia resulted in the isolation of 10 new highly oxygenated C20 quassinoids longifolactones G‒P (1–10), along with four known ones (11–14). Their chemical structures and absolute configurations were unambiguously elucidated on the basis of comprehensive spectroscopic analysis and X-ray crystallographic data. Notably, compound 1 is a rare pentacyclic C20 quassinoid featuring a densely functionalized 2,5-dioxatricyclo[5.2.2.04,8]undecane core. Compound 4 represents the first example of quassinoids containing a 14,15-epoxy functionality, and 7 features an unusual α-oriented hydroxyl group at C-14. All isolated compounds were evaluated for their anti-proliferation activities on human leukemia cells. Among the isolates, compounds 5, 12, 13, and 14 potently inhibited the in vitro proliferation of K562 and HL-60 cells with IC50 values ranging from 2.90 to 8.20 μM.


Introduction
Quassinoids are a class of highly oxygenated degraded triterpenoids mainly distributed in plant family Simaroubaceae [1]. Based on the number of carbon atoms involving the construction of their basic scaffolds, quassinoids are commonly categorized into six distinct groups: C 26 , C 25 , C 22 , C 20 , C 19 , and C 18 types [2]. Quassinoids have been reported to display a wide range of biological activities, including antitumor, antimalarial, anti-inflammatory, antiviral, neuroprotective, and antifeedant activities [2,3]. Especially since the discovery of bruceantin, a C 20 quassinoid isolated from Brucea antidysenteria (Simaroubaceae) in the early 1970s that showed remarkable antileukemic activity, the antitumor activities of quassinoids have attracted extensive attention from both chemical and biological communities [4][5][6][7].
Eurycoma longifolia Jack (Simaroubaceae), commonly known as "Tongkat Ali", is a flowering shrub plant that widely distributed in Southeast Asia [8]. The roots of E. longifolia were traditionally used by local people for the treatment of malaria, dysentery, glandular swelling, persistent fever, aches, and sexual insufficiency [8]. Besides, the antitumor activities of the crude extract of E. longifolia roots were reported in 2005 [9]. Previous phytochemical investigations on the roots of E. longifolia have afforded a wide variety of chemical components, including quassinoids, canthin-6-one alkaloids, β-carboline alkaloids, tirucallane-type triterpenes, squalene derivatives, and biphenyl neolignans [10]. Among them, quassinoids are the most characteristic chemical constituents of this plant [11][12][13].
Previously, our group had reported the isolation and characterization of six novelquassinoids (longifolactones A-F) with unprecedented C 26 or C 20 scaffolds from the petroleum ether-soluble fraction of the ethanol extract of E. longifolia roots [14]. Among them, longifolactione F is the first example of quassinoids containing an unprecedented densely functionalized 2,5-dioxatricyclo [5.2.2.0 4,8 ]undecane ring system. In our continuing studies on searching structurally unique and biologically interesting metabolites from medicinal plants, the ethyl acetate-soluble fraction of the title plant was further investigated. As a result, longifolactones G-P (1)(2)(3)(4)(5)(6)(7)(8)(9)(10), 10 new C 20 quassinoids, together with four known ones were isolated. Their structures and absolute configurations were unambiguously established by extensive spectroscopic data analysis and single-crystal X-ray diffraction experiment. Notably, compound 1 is the second member of the rare class of quassinoids featuring densely functionalized 2,5-dioxatricyclo [5.2.2.0 4,8 ]undecane core. Besides, compound 4 represents the first example of quassinoids containing a 14,15-epoxy functionality, and compound 7 features an unusual 14α-OH substituent that makes 7 the second member of this rare class of quassinoids so far. Herein, we reported the isolation and structure elucidation of these new quassinoids. In addition, the in vitro anti-proliferation activities of all isolates on two human leukemia cell lines (K562 and HL-60 cells) were also described.
The above NMR spectroscopic data of 1 were closely similar to those of longifolactone F [14], suggesting the structural similarity of these two compounds. Different from longifolactone F, the NMR signals corresponding to a methylene group (CH 2 -3) and a methine group  were replaced by resonances of a double bond [δ H 6.07 (1H, br s); δ C 163.3 and 125.6] in 1. In the HMBC spectrum, key correlations between H 3 -29 and C-3, H-5 and C-3, H 2 -6 and C-4 were observed, suggesting the presence of an α,β-unsaturated ketone motif in ring A of 1, which was also confirmed by the characteristic chemical shift values of C-2-C- 4 (δ C 199.7, 125.6, and 163.3). After a comprehensive interpretation of its 1 H-1 H COSY and HMBC spectra, the gross structure of 1 was established as a C 20 quassinoid with a rare 2,5-dioxatricyclo [5.2.2.0 4,8 ]undecane core ( Figure 2). In the NOESY spectrum of 1, key NOE correlations between H-1 and H-5/H-11, H-5 and H-9 were observed, indicating that these protons had the same orientation. Meanwhile, the NOE correlations between  were observed, suggesting that these protons were located on the same face of the molecule ( Figure 3). Finally, suitable crystals of 1 were acquired. The following X-ray diffraction analysis with Cu Kα radiation resulted in an excellent Flack parameter of 0.06 (4), which not only allowed the verification of the planar structure of 1 but also led to the establishment of its absolute configuration as 1S,5S,7R,8R,9R,10S,11R,12R,13S,14R,15R ( Figure 4).  The molecular formula of 2 was deduced as C 21 H 30 O 8 on the basis of its HR-ESI-MS data (m/z 433.1832 [M + Na] + ; calcd for C 21 H 30 O 8 Na, 433.1833), and the 13 C NMR data analysis. The 1 H and 13 C NMR spectroscopic data of 2 (Tables S1 and S2) highly resembled those of 6-dehydroxylongilactone [19], except for the presence of additional signals due to a hemiketal carbon (δ C 102.5) and an oxygenated methyl group [δ H 3.83 (3H, s); δ C 52.7] in 2. Subsequently, detailed analysis of its 1 H-1 H COSY and HMBC spectra allowed the establishment of a 6/6/6/5 ring system for 2 that was identical to 6-dehydroxylongilactone. Besides, the HMBC correlation between 1 -OCH 3 and C-16 indicated the presence of an extra methoxycarbonyl group in 2. Based on the molecular formula information, as well as the obvious down-field shift of C-15 (δ C 102.5), the remaining methoxycarbonyl and hydroxyl groups were both assigned to attach to C-15, which was also confirmed by the HMBC correlation between H-14 and C-16 ( Figure 2). Thus, the planar structure of 2 was established. Finally, the structure of 2 was fully resolved by an X-ray diffraction experiment.
The molecular formula of 3 was determined to be C 20 H 28 O 5 on the basis of its sodiated molecular ion peak at m/z 371.1831 [M + Na] + (calcd for C 20 H 28 O 5 Na, 371.1829) and 13 C NMR data. The 1 H and 13 C NMR spectral data of 3 were similar to those of the co-isolated known compound chaparrolide (11), which indicated that 3 was also a C 20 quassinoid. Compared with those of 11, the 1 H and 13 C NMR spectra of 3 showed additional signals for a cis-disubstituted double bond [δ H 6.31 (1H, d, J = 9.6 Hz) and 5.76 (1H, d, J = 9.6 Hz); δ C 132.7 and 130.2] and an exo-olefin group [δ H 5.00 (1H, s) and 4.84 (1H, s); δ C 145. 8 and 111.3], while the signals corresponding to a ketone carbonyl, a methylene group, a methine group, and a methyl group were absent. In the 1 H-1 H COSY spectrum of 3, the correlation between H-11 and H-12 indicated that the C-11 in 3 was an oxygenated-substituted methine instead of the ketone carbonyl in 11 ( Figure 2). Moreover, the observed HMBC cross-peaks between H 2 -29 and C-3/C-5, H-3 and C-5, H-3 and C-1 indicated the presence of two conjugated double bonds in ring A of 3 ( Figure 2). Similarly, an X-ray diffraction experiment using Cu Kα radiation was performed, which led to the full assignment of planar structure and absolute configuration for 3 (1R,5S,7R,8S,9R,10S,11S,12R,13R,14S, Figure 4).
The molecular formula of 4 was assigned as C 20 H 26 O 7 based on its HR-ESI-MS data (m/z 401.1574 [M + Na] + ; calcd for C 20 H 26 O 7 Na, 401.1571) and 13 C NMR spectroscopic data, 18 mass units less than the co-isolated known C 20 quassinoid, 14,15β-dihydroxyklaineanone (13). The NMR spectra of 4 showed characteristic signals similar to those of 13, except for the presence of one oxygen-bearing methine [δ H 3.36 (1H, s); δ C 52.9] and one oxygenated bearing quaternary carbon (δ C 67.9). Considering the molecular formula information, the two hydroxyl groups at C-14 and C-15 in 13 were replaced by an epoxide ring in 4. Furthermore, the NOE correlation between H-15 and H 3 -18 in the NOESY spectrum suggested that the epoxide ring had the β-orientation ( Figure 3). Similar to 1-3, the structure with absolute configuration (1S,5S,7R,8S,9R,10S,11R,12R,13S,14R,15R) of 4 was definitively assigned by an X-ray diffraction experiment ( Figure 4).
The HR-ESI-MS of compound 8 displayed a sodiated molecular ion peak at m/z 431.1311 [M + Na] + (calcd for C 20 H 24 O 9 Na, 431.1313), allowing the determination of a molecular formula of C 20 H 24 O 9 that was identical to the known C 20 quassinoid 13-epi-eurycomadilactone [21]. The 1 H and 13 C NMR spectral data of 8 (Tables S1 and S3) closely resembled those of 13epi-eurycomadilactone, combined with its molecular formula information, suggesting that 8 was a stereoisomer of the known compound. Further analysis of the 2D NMR data of 8 confirmed that 8 had the same planar structure as 13-epi-eurycomadilactone. Different from 13-epi-eurycomadilactone, the NOESY spectrum of 8 showed the correlation between H-13 and H 2 -30, indicating the α-orientation for the H 3 -18 in 8 ( Figure 3). The structure with absolute configuration (1S,5S,7R,8R,9R,10S,11S,13R,14R,15R) of 8 was finally determined on the basis of an X-ray crystallography study by using the anomalous dispersion of Cu Kα radiation ( Figure 4).
Compound 9 was assigned to possess a molecular formula of C 20 H 26 O 9 by the HR-ESI-MS ion peak at m/z 433.1472 [M + Na] + (calcd for C 20 H 26 O 9 Na, 433.1469) and 1D NMR spectral data analysis, which was two mass units more than that of 8. The 1 H and 13 C NMR spectra of 9 exhibited similar signals to those of 8 (Tables S1 and S3), except for the signal assigned to a ketone carbonyl (δ C 197.0, C-2 in 8) was replaced by the signals of an oxygenated methine [δ H 4.55 (1H, overlapped); δ C 72.5] in 9. Thus, compound 9 was assumed to be a C-2 hydroxylated derivative of 8. This deduction was further verified by the spin system from H-1 to H-3 in the 1 H-1 H COSY spectrum of 9 ( Figure 2). Furthermore, the α-orientation of the 2-OH was determined on the basis of key NOE correlation between H-2 and H 3 -19 ( Figure 3). A further crystallographic analysis led to the unambiguous establishment of the structure and absolute configuration (1S,2S,5S,7R,8R,9R,10S,11S,13R,14R,15R) of 9 ( Figure 4).
The molecular formula of 10 was deduced to be identical to that of 9 on the basis of its HR-ESI-MS data (m/z 433.1470 [M + Na] + ; calcd for C 20 H 26 O 9 Na, 433.1469) and 13 C NMR data. Comparison of the NMR data of 10 with those of 9 (Tables S1 and S3) indicated that 10 possessed the identical gross structure to 9. The main differences of the NMR spectral data between 10 and 9 were the obvious down-field shifts of C-5 (∆δ +5.1) and C-6 (∆δ +5.2) in 10, suggesting that 10 might be a C-5 epimer of 9. Further analysis of its 2D NMR spectroscopic data verified that 10 possessed the identical planar structure to 9. In the NOESY spectrum, NOE correlation between H-5 and H 3 -19 was observed, suggesting the β-orientation for H-5 in 10 ( Figure 3). Similar to 1-9, the single-crystal X-ray diffraction study (Cu Kα) allowed the assignment of the complete stereochemistry of 10. As a result, the absolute configuration of 10 was definitively assigned to be 1S,2S,5R,7R,8R,9R,10S,11S,13R,14R,15R (Figure 4).

Anti-proliferation Activities of Isolated Quassinoids
The isolated compounds were tested for their anti-proliferation activities on two human leukemia cell lines, K562 and HL-60. As shown in Table S5, compounds 5, 12, 13, and 14 exhibited potent inhibitory effects on the proliferation of both K562 and HL-60 cells with IC 50 values ranging from 2.90 to 8.20 µM.

General Methods
Melting points were measured on an X-5 melting point instrument (Fukai, Beijing, China) without correction. Optical rotations were determined in MeOH on a P-1020 polarimeter (JASCO, Tokyo, Japan) with a 1 cm cell at room temperature. UV spectra were acquired on a JASCO V-500 UV/vis spectrometer. IR spectra were obtained with a JASCO FT/IR-480 plus infrared spectrometer using KBr pellets. HR-ESI-MS data were collected using an Agilent 6210 TOF-MS spectrometer (Agilent Technologies, Santa Clara, CA, USA). Other experimental procedures were performed as described previously [14]. The human leukemia cell lines, HL-60 and K562, were purchased from the American Type Culture Collection (ATCC) and cultured in RPMI-1640 medium supplemented with 10% fetal bovine serum (FBS) and 2 mM L-glutamine.

Plant Material
The roots of Eurycoma longifolia were collected from Malacca, Malaysia, in June 2014 and authenticated by Prof. Guang-Xiong Zhou (College of Pharmacy, Jinan University). A voucher specimen (No. 20140501) was deposited in the Center for Bioactive Natural Molecules and Innovative Drugs Research, College of Pharmacy, Jinan University.

Extraction and Isolation
The air-dried and powdered roots of E. longifolia (10 kg) were extracted with 95% (v/v) EtOH five times at room temperature. The combined EtOH extract was concentrated under vacuum to yield a crude extract (270 g), which was suspended in water and then partitioned successively with petroleum ether, ethyl acetate, and n-BuOH.

Cell Proliferation Assay
HL-60 and K562 cells were cultured in 96-well plates and incubated at 37 • C, 5% CO 2 incubator. After incubation for 24 h, the cell supernatants were discarded and supplemented with cell culture medium containing compounds at different concentrations. At 48 h after incubation, the cell supernatants in each well were removed and replaced with 100 µL culture medium containing 10 µL of cell counting kit-8 (Sigma-Aldrich, St. Louis, MO, USA), followed by incubation at 37 • C, 5% CO 2 for 2 h. The absorbance at 450 nm of the cells was measured using a Plate Reader. Cell proliferation was calculated according to the OD 450 value in cells that were treated with or without compounds.

Conclusions
In summary, a further phytochemical study on the roots of the medicinal plant Eurycoma longifolia resulted in the isolation and characterization of 14 highly oxygenated C 20 quassinoids, including 10 new ones (longifolactones G-P, 1-10). Structurally, compound 1 is the second member of a rare class of quassinoids featuring an unusual 2,5dioxatricyclo [5.2.2.0 4,8 ]undecane ring system. Compound 4 possesses a 14,15-epoxy functionality that is unprecedented in quassinoids, and compound 7 features an unusual α-oriented hydroxyl group at C-14. In addition, compounds 5, 12, 13, and 14 showed potent anti-proliferation activities on two human leukemia cell lines, K562 and HL-60.
Supplementary Materials: The following are available online, Detailed UV, IR, HR-ESI-MS, and NMR spectra of compounds 1-14, as well as crystallographic data of compounds 1-10 are available as Supplementary Materials. Data Availability Statement: All data supporting this study is available in the manuscript and the Supplementary Materials.