Oxidized Juncuenin B Analogues with Increased Antiproliferative Activity on Human Adherent Cell Lines: Semisynthesis and Biological Evaluation

Phenanthrenes have become the subject of intensive research during the past decades because of their structural diversity and wide range of pharmacological activities. Earlier studies demonstrated that semisynthetic derivatization of these natural compounds could result in more active agents, and oxidative transformations are particularly promising in this regard. In our work, a natural phenanthrene, juncuenin B, was transformed by hypervalent iodine(III) reagents using a diversity-oriented approach. Eleven racemic semisynthetic compounds were produced, the majority containing an alkyl substituted p-quinol ring. Purification of the compounds was carried out by chromatographic techniques, and their structures were elucidated by 1D and 2D NMR spectroscopic methods. Stereoisomers of the bioactive derivatives were separated by chiral-phase HPLC and the absolute configurations of the active compounds, 2,6-dioxo-1,8a-dimethoxy-1,7-dimethyl-8-vinyl-9,10-dihydrophenanthrenes (1a–d), and 8a-ethoxy-1,7-dimethyl-6-oxo-8-vinyl-9,10-dihydrophenanthrene-2-ols (7a,b) were determined by ECD measurements and TDDFT-ECD calculations. The antiproliferative activities of the compounds were tested on different (MCF-7, T47D, HeLa, SiHa, C33A, A2780) human gynecological cancer cell lines. Compounds 1a–d, 4a, 6a, and 7a possessed higher activity than juncuenin B on several tumor cell lines. The structure–activity relationship studies suggested that the p-quinol (2,5-cyclohexadien-4-hydroxy-1-one) moiety has a considerable effect on the antiproliferative properties, and substantial differences could be identified in the activities of the stereoisomers.

P henanthrenes are a rather uncommon class of aromatic secondary metabolites and have a limited occurrence in nature. They are presumably formed by oxidative coupling of the aromatic rings of stilbene precursors. 1 Hitherto, more than 460 phenanthrenes, among them mono-, di-, and triphenanthrenes, were identified from different plant families, e.g., Orchidaceae, Juncaceae, Combretaceae, and Dioscoreaceae. 2 A large number of phenanthrenes with different substitution patterns have been reported with various pharmacological activities, such as antimicrobial, anti-inflammatory, cytotoxic, antiproliferative, spasmolytic, antiallergic, and anxiolytic effects. 2,3 The most thoroughly investigated denbinobin, a phenanthrenequinone, showed promising antiproliferative activity. 4−6 The mechanism of action of denbinobin involves caspase-dependent and caspase-independent apoptosis of colon cancer (COLO 205) cells. 4 Moreover, by enhancing the synthesis of reactive oxygen species (ROS), denbinobin induces apoptosis in Jurkat leukemia and human lung adenocarcinoma (A549) cells. 5,7 A structure−activity relationship study showed that the quinone substructure is required for enhancement of mitogen-activated protein kinase (MAPK)-independent ROS production activity. 5 Another phenanthrenequinone, calanquinone A, differing from denbinobin by the presence of an additional methoxy group, and synthetic phenanthrenequinones also possessed cytotoxic activity on different cell lines (HepG2, Hep3B, Ca9-22, A549, MCF-7, MDA-MB-231, . Structure−activity relationship studies showed that intramolecular hydrogen bonding between the C-4 carbonyl and HO-5 in 3-methoxyphenanthrene-1,4-diones may be necessary for the cytotoxicity. 6 The phenanthrene dehydroeffusol showed activity against two metastatic cancer cell lines, SGC-7901 (human gastric carcinoma) and AGS (human caucasian gastric adenocarcinoma) in a dose-dependent manner (IC 50 35.9 and 32.9 μM for SGC-7901 and AGS cells, respectively). Gastric cancer cell mediated vasculogenic mimicry on SGC-7901 cells were also inhibited by this compound using the tube formation assay. 8 Dehydroeffusol effectively inhibited the gastric cancer cell invasion and migration without exerting significant acute toxicity. 9 Juncusol was observed to be cytotoxic on HeLa cells; treatment for 24 h increased the cell population in the G2/M and sub-G1 phases. 10 Juncuenin B possessed significant antiproliferative activity against different human tumor cell lines such as MDA-MB-231 (breast cancer), HeLa (cervical cancer), and A2780 (ovarian cancer) with IC 50 values 9.4, 2.9, and 7.3 μM, respectively. 10 The aim of this work was the semisynthetic modification of juncuenin B (1,7-dimethyl-8-vinyl-9,10-dihydrophenanthrene-2,6-diol) to prepare a variety of structurally related compounds with higher antiproliferative activities. The experimental approach was inspired by a 3-fold of reasons. First, the free radical scavenging-related, oxidized metabolites of phenolic compounds represent a chemical space particularly rich in bioactive compounds. 11 Second, we have recently demonstrated that the hypervalent iodine(III) reagent [bis-(trifluoroacetoxy)iodo]benzene (PIFA) may be efficiently used to obtain product mixtures rich in such oxidized scaffolds. 12 Third, the p-quinol moiety, anticipated from a PIFA-catalyzed oxidation of p-phenols, is the pharmacophore of several potent inhibitors of ATR-dependent signaling, 12,13 a crucial mechanism of DNA damage response that is an emerging antitumor target and a subject of many ongoing clinical studies. 14 Therefore, we followed a diversity-oriented semisynthetic approach using two hypervalent iodine reagents, PIFA, and (diacetoxyiodo)benzene (PIDA) to obtain oxidized derivatives of juncuenin B. Furthermore, various alcohols (MeOH, EtOH, and n-BuOH) were used as solvents to ensure formation of a set of differently substituted derivatives. Eleven chiral oxidation products (1−11), representing a wide structural diversity, were obtained in this way as racemates or stereoisomeric mixtures. Stereoisomers of the bioactive derivatives were resolved by chiral-phase HPLC and absolute configurations were assigned by ECD measurements and TDDFT-ECD calculations. The antiproliferative activity of the racemic compounds and separated optically active stereoisomers was investigated on MCF-7, T47D, HeLa, SiHa, C33A, and A2780 cell lines, which enabled us to establish both structure−activity and stereochemistry−activity relationships. In order to gain insight into their tumor selectivity, the compounds have also been tested on the NIH/3T3 cell line.

■ RESULTS AND DISCUSSION
Eleven racemic and enantiomerically pure chiral semisynthetic derivatives (1−11) (Scheme 1 and Figure 1) have been prepared from juncuenin B, a naturally occurring vinylsubstituted achiral dihydrophenanthrene, in five transformations (I−V) by hypervalent iodine(III) reagents using a diversity-oriented approach. PIFA and PIDA were used as oxidants (PIFA in processes I−IV, PIDA in process V), under different conditions, in MeCN−MeOH (process I), MeOH (II), EtOH (III), n-BuOH (IV, V). PIFA and PIDA are wellknown reagents for the oxidation of phenols leading to the formation of quinone-type products. This takes place through an aryloxyiodonium(III) intermediate forming a phenoxenium ion that undergoes a nucleophilic attack, 15 but PIFA and PIDA may also oxidize aromatic compounds via a single-electron transfer mechanism. 16 Following the oxidation, the mixture of products were subjected to solid-phase extraction on silica to remove the remaining oxidizing agent and the oxidation sideproducts. The purification process was followed by MPLC and HPLC.
During the reaction processes I−V, racemic mixtures of the compounds were formed, all of them bearing p-or o-quinol rings, substituted with methoxy-, ethoxy-, and n-butoxy groups based on the solvent used. In the case of reaction IV, only compound 10 could be isolated. It did not show structural analogy with the aforementioned compounds; therefore, the reaction was performed with PIDA under the same conditions (process V). This reaction yielded 11, analogous to the compounds derived from 4 and 7. As a result of chromatographic purifications, compounds with one stereogenic center were obtained as racemates (3, 4, 7−11), while compounds with two stereogenic centers were isolated as diastereomeric mixtures of two racemates (1a+1d, 1b+1c, 2a, 2b, 6a, 6b). Separations of stereoisomers were performed using chiralphase HPLC when the mixtures of stereoisomers or the racemates (1a−d, 4a,b, 7a,b) were found effective in the antiproliferative assay. This was done in order to compare the antiproliferative activity of the pure enantiomers and identify not only 2D structure−activity but also stereochemistry− activity relationships. The separation of the stereoisomers also enabled the characterization and determination of the absolute configurations of the bioactive derivatives by comparison of the experimental and TDDFT-calculated ECD spectra.
The structure determination was carried out by extensive spectroscopic analysis, using 1D and 2D NMR ( 1 H− 1 H COSY, HSQC, and HMBC) spectroscopy and ECD data analysis.
In the NMR spectra of 1a−1d, two sets of signals could be identified that were attributed to two racemic diastereomers 1a +1d and 1b+1c (ratio 1:1). The 1 H NMR spectrum of the 1a− 1d mixture showed the presence of a pair of isomeric compounds (1a+1d and 1b+1c) as some signals were duplicated. Since these compounds are derived from juncuenin B, similar signals were found in the 1 (Figure 2). The location of the methoxy groups was concluded from the HMBC spectra, as proton signal at δ H 3.07 (CH 3 O-1) showed correlations with δ C 83.8 (C-1), whereas the signal at δ H 2.92/2.93 (CH 3 O-8a) correlated with δ C 73.2 (C-8a) ( Figure 2). Thus, the C-2 and C-6 hydroxy groups of juncuenin B were oxidized to carbonyl groups (δ C 202.6 and 185.9, respectively) and methoxysubstituted dienone moieties were formed. All of the above evidence confirmed the 2D structures of the 1a−1d isomers ( Figure 2).
The NMR spectra of the mixture of 1b+1c were similar to those of 1a+1d, only small differences were observed for the CH 3 O-1, C-10, and C-11 resonances of 1a+1d and 1b+1c, which suggested that these compounds are structurally related to each other. Some 1 H and 13 C NMR signals were duplicated in the spectra of 1b+1c, revealing the presence of a diastereomeric pair as for 1a+1d.
In order to elucidate the absolute configurations of 1a−d, the solution TDDFT-ECD method was applied to (1R,8aR)-1 and (1R,8aS)-1. 17 The initial Merck Molecular Force Field (MMFF) conformational search resulted in nine conformers for each in a 21 kJ/mol energy window, respectively. These geometries were reoptimized at the ωB97X/TZVP PCM/ MeCN level 18 resulting in three conformers for both stereoisomers ( Figures S8 and S9, Supporting Information). ECD spectra were computed at various levels and compared to the experimental spectra of 1a−1d. There were only minor differences in the experimental ECD spectra of 1a versus 1b and 1c versus 1d, and the computed ECDs are mainly influenced by one of the two chirality centers. Boltzmannaveraged ECD data of both (1R,8aR)-1 and (1R,8aS)-1 gave acceptable to good agreement with the experimental data of 1a and 1b with the C-1 stereogenic center definitively influencing the ECD profile ( Figure 4). Thus, the absolute configuration of this stereocenter could be determined as (1R) for the firstand second-eluted stereoisomers 1a and 1b and (1S) for the thirdand fourth-eluting 1c and 1d. The shape of the B3LYP/TZVP PCM/MeCN and PBE0/TZVP PCM/MeCN ECD spectra of (1R,8aR)-1 reproduced the experimental ECD curve of 1a better, while those of (1R,8aS)-1 were similar to the experimental spectrum of 1b ( Figure 5). The better agreements allowed tentative assignment (8aR) for 1a and (8aS) for 1b. DFT-NMR calculations of the B3LYP/6-31+G(d,p) conformers of (1R,8aR)-1 and (1R,8aS)-1 at the mPW1PW91/6-311+G(2d,p) level 19 could not differentiate Journal of Natural Products pubs.acs.org/jnp Article          and C-7 (δ C 134.1), respectively. Methoxy groups attached at C-4a and C-8a were positioned by means of the HMBC interactions between δ H 3.04 (CH 3 O-4a) and δ C 73.1 (C-4a), and between δ H 2.72 (CH 3 O-8a) and δ C 75.1 (C-8a). The vinylic system was connected to C-8 via cross peaks between δ H 5.63, dd, 5.78, dd and δ C 150.2.
All NMR characteristics of compound 2b were similar to compound 2a, indicating that they are diastereomers. The main differences between 2b and 2a were found between the 13 C NMR chemical shifts of C-4a and 8a and adjacent carbons, suggesting that 2b and 2a differ in the configurations of C-4a and 8a.
In     Journal of Natural Products pubs.acs.org/jnp Article presence of an aromatic ring B, the oxidation of the phenolic group at C-2 to a carbonyl moiety (δ C 202.4), and the addition of methoxy groups at C-1 and C-5. The 1 H NMR spectrum of the racemic mixture of 4a and 4b differ from juncuenin B in the presence of a methoxy group (δ H 2.85) ( Table 1) and the carbonyl group at ring C. In the JMOD spectrum 19 carbon signals were observed ( Table 2). The methoxy group was located at C-8a, as the aforementioned CH 3 O proton signal showed HMBC correlation with the carbon signal resonating at δ C 73.8. Following the NMR studies the enantiomers 4a and 4b were separated on a Lux amylose-1 chiral-phase HPLC column [t R = 5.5 min (4a) and 10.3 min (4b)].
Compounds 6a and 6b were obtained as racemic diastereomers. The difference between compounds 1a−d and compounds 6a and 6b involved the replacement of the C-1 and C-8a methoxy groups with ethoxy substituents (Tables 1  and 2).
For the configurational assignments of 7a and 7b, the abovementioned TDDFT-ECD protocol was applied to (R)-7 and also on the truncated methoxy model compound (R)-4.
DFT reoptimization of the initial six MMFF conformers of both (R)-4 and (R)-7 yielded six conformers above 1% population indicating that the ethoxy group does not have considerable freedom ( Figure S64, Supporting Information). The Boltzmann-weighted ECD spectra of both sets of conformers reproduced well the experimental ECD spectrum of the first-eluting enantiomer 7a (Figure 7 and Figure S37, Supporting Information). Consequently, the absolute configurations of 7a and 4a, which were separated under the same conditions, could be elucidated as (R) and that of the secondeluting 7b and 4b as (S).
In the case of compound 10, it was apparent from the 1D NMR spectra that three n-butoxy groups are connected to the phenanthrene skeleton. Both the phenolic substituents of juncuenin B were oxidized to carbonyl groups, and a Δ 9,10 double bond was also formed. The location of the methyl, vinyl, and n-butoxy groups were determined by their diagnostic HMBC correlations. Three-bond heteronuclear interactions of C-1 (δ C 81.9) with the oxymethylene protons at δ H 2.92, m, and of the deshielded C-5 (δ C 94.6) with two additional oxymethylenes at δ H 3.14, m, and 3.54, m confirmed the positions of the n-butoxy groups at C-1 and C-5, respectively.
In the second step, the enantiomerically pure compounds (1a, 1b, 1c, 1d, 4a, 4b, and 7b) were evaluated for their antiproliferative activity against the same cell lines (Table 4). Among these compounds, the highest activities were recorded for 1c, 4a, and 7a, reaching or exceeding that of the positive control cisplatin against HeLa and/or T47D cell lines. (R)-7a was found to be the most promising compound with substantial antiproliferative effects against all tested cell lines except for SiHa, but it was inactive against the nontumoral NIH/3T3 cells.
Compounds 4a, 4b, 7a, 7b, and 11 contain a nonoxidized A ring and a C-ring comprising a p-quinol alkyl ether moiety. Moreover, 4a and 4b are substituted with a methoxy-, 7a, and 7b with an ethoxy-, and 11 with an n-butoxy group at C-8a, respectively. In the case of these compounds, the effect of the length of the ether chain at C-8a is undefined: on the T47D, HeLa, C33A, and A2780 cell lines, compound 7a with an ethoxy group, showed the highest inhibition, while its enantiomer (7b) had no activity. On the HeLa cell line, the antiproliferative activity was inversely proportional to the chain length of the alkoxy groups (IC 50 values 11 > 7a > 4a). Compounds 4a and 7a showed an antiproliferative effect on HeLa, C33A, MCF-7, and A2780 cells, which was comparable to or stronger than that of juncuenin B, while their stereoisomers 4b and 7b had no relevant influence on the cell proliferation, indicating that (8aR) is the preferred absolute configuration of these compounds. Regarding the pure enantiomers 1a−d, the (1S,8aR)-configuration (1c) was the most beneficial for the antiproliferative effect. The activities of (1S,8aR)-1c, (R)-4a, and (R)-7a suggest the preference of the (R) configuration for the antiproliferative activity. Since 1a is less active than 1c, the (1S) configuration seems to be also essential in the case of quinoidal compounds.
It is worth mentioning that several of the semisynthetic compounds demonstrated an altered cell line specificity as compared to their parent compound, juncuenin B. For example, while juncuenin B exerted approximately 2.5-fold stronger activity on MCF-7 than on T47D cells, a similar rate of selectivity in the opposite direction was observed for the enantiopure compound 7a. While both of these cell lines are estrogen receptor α positive and therefore estrogen-dependent, they show critical differences in their bioenergetic properties, 21 as well as in apoptotic events including the activation of different caspases and mitochondrial changes. 22 This selectivity inversion suggests that compounds obtained from juncuenin B Determinations were performed using an MTT assay by treating the cells with compounds (0.1−30 μM) for 72 h. Data are based on two independent experiments. b IC 50 value from literature. 10 Journal of Natural Products pubs.acs.org/jnp Article were not only more potent than the starting material, but also these compounds likely act on different biochemical targets. Further studies are necessary for the clarification of these targets, but chemical analogies from our previous work suggest the involvement of ATR-dependent signaling at least to some extent. 12,13 In conclusion, oxidation of juncuenin B with hypervalent iodine reagents under different conditions resulted in the identification of phenanthrene derivatives substituted with hydroxy, methyl, methoxy, vinyl, ethoxy and n-butoxy groups. The majority of the components contain a substituted p-quinol ring. All compounds are reported here for the first time. Several compounds showed promising antiproliferative activities, and based on their structural analogy, structure−activity relationships were deduced. No significant activities were observed either for compounds without a substituted p-quinol ring (3, 8, 9, and 11), or for those containing two such rings (2a, 2b, and 5). However, derivatives containing one substituted p-quinol ring showed promising antiproliferative activities against different human cancer cell lines, and this was observed for compounds with an o-alkoxy-quinone moiety (1a−d, 6a) and for those with an intact A ring (4a, 7a, and 11). Clarifying the mechanism of action for these compounds requires further studies.
In the current work, similar structural elements were introduced in juncuenin B as present in denbinobin and calanquinone. The aromatic ring oxidations of juncuenin B afforded quinols which were found to be bioactive to the same degree as the lead compounds denbinobin and calanquinone. The starting compound juncuenin B can be readily isolated in a good yield (0.043% w/w) from Juncus inflexus root. There is a strong rationale for further development of oxidized derivatives of juncuenin B in the future.

■ EXPERIMENTAL SECTION
General Experimental Procedures. NMR spectra were recorded in methanol-d 4 and CDCl 3 on a Bruker Avance DRX 500 spectrometer at 500 MHz ( 1 H) and 125 MHz ( 13 C). The signals of the deuterated solvents were used as references. The chemical shift values (δ) are given in ppm and coupling constants in Hz. 2D experiments were performed with standard Bruker software. In the 1 H− 1 H COSY, HSQC, and HMBC experiments, gradient-enhanced versions were applied. The HRMS data were acquired on a Thermo Scientific Q-Exactive Plus Orbitrap mass spectrometer equipped with ESI ion source in positive ionization mode. The data were acquired and processed with MassLynx software. ECD and UV spectra were recorded on a J-810 spectropolarimeter.
In the reaction processes [bis(trifluoroacetoxy)iodo]benzene (PIFA) and (diacetoxyiodo)benzene (PIDA) were used as oxidizing agents (Sigma-Aldrich, Stockholm, Sweden). To reaction progress and separation steps were monitored by TLC on Kieselgel 60F 254 silica plates obtained from Merck (Merck, Darmstadt, Germany), and examined under UV light at 254 and 366 nm. Solid phase extraction (SPE) was carried out on silica gel (Kieselgel 60 GF 254 , 15 μm, Merck). MPLC was performed with a Combi Flash Rf + Lumen instrument (Teledyne Isco). All solvents used for MPLC were of at least analytical grade (VWR Ltd., Szeged, Hungary). Both the reaction mixtures and processed fractions were investigated by HPLC on a Jasco instrument equipped with an MD-2010 Plus PDA detector (Jasco Analytical Instruments, Japan) in a detection range of 210−400 nm, with an Agilent Eclipse XDB C 8 column (4.6 × 150 mm, 5 μm) (Agilent Technologies, Inc., USA). MeCN−H 2 O solvent system was used as eluent. The final purification steps were carried out by using HPLC with an Agilent Eclipse XDB C 8 (9.4 × 250 mm, 5 μm) (Agilent Technologies, Inc., USA) equipment, and MeCN−H 2 O mixture as the mobile phase. Enantiomers were separated by using a Lux amylose-1 column (250 × 21.2 mm) (Phenomenex, USA), using a cyclohexane−isopropanol solvent system as mobile phase. The HPLC system for semipreparative separations comprised of a Waters 600 controller, Waters 600 pump, and Waters 2998 photodiode array detector. The data were acquired and processed with the Empower software.
The antiproliferative activity and tumor selectivity of the compounds were measured by standard MTT assays. Tumor cell lines MCF-7, T47D, HeLa, A2780, and noncancerous NIH/3T3 were provided by European Collection of Authenticated Cell Cultures (ECACC, Salisbury, UK); cell lines SiHa and C33A by LGC Standards GmbH (Wesel, Germany). All chemicals and materials used for the antiproliferative assays were purchased from Sigma Aldrich Ltd. (Budapest, Hungary).
Juncuenin B. Juncuenin B was isolated previously from the MeOH extract of Juncus inf lexus roots by chromatographic methods (VLC on silica gel, and gel chromatography on Sephadex LH-20). J. inf lexus is a good source of this compound as approximately 1.5 g was isolated (0.043% w/w) in crystalline form from 3.5 kg dried plant material. 23 Synthesis and Purification Process. In reaction I, 50 mg juncuenin B was dissolved in MeCN−MeOH (9:1) at a concentration of 1 mg/mL, and stirred with 2 equiv of PIFA (161.2 mg) for 30 min at 0°C. After drying under nitrogen gas, SPE was applied to absorb the remaining oxidizing agent and the possibly decomposed compounds. Thereafter, the reaction mixture was fractionated by MPLC with n-hexane−EtOAc gradient solvent system to obtain five fractions (I/1−5). From fraction I/1 compounds 1a−d (1.00 mg, 2% each) were isolated by RP-HPLC using MeCN−H 2 O (1:1) as eluent (3 mL/min flow rate). Fraction I/2 was purified by RP-HPLC, MeCN−H 2 O (3:2) (3 mL/min flow rate) was applied as eluent to gain compounds 2a (0.88 mg) and 2b (0.75 mg).
For process reaction II, 100 mg juncuenin B was dissolved in anhydrous MeOH at a concentration of 1 mg/mL, and stirred with 2 equiv of PIFA (322.32 mg) for 30 min at room temperature. After evaporating the solvent under reduced pressure, SPE was applied to absorb the remaining oxidizing agents and the possibly decomposed compounds. The mixture was fractionated by MPLC with n-hexane− EtOAc gradient solvent system to obtain four fractions. Fraction II/1, after being chromatographed by RP-HPLC using MeCN−H 2 O (1:1) (flow rate 3 mL/min) as mobile phase, compounds 1a−d (2.00 mg) and 5 (3.8 mg) were obtained. Fractions II/2 and II/3 were separated under the same conditions as II/1. From fraction II/2 1a−d (1.00 mg) and 3 (1.8 mg); from II/3 compounds 1a−d (3.00 mg) and 4a,b (2.00 mg) were isolated.
In the case of reaction III, 100 mg juncuenin B was dissolved in anhydrous EtOH at a concentration of 1 mg/mL, and stirred with 2 equiv of PIFA (322.32 mg) for 30 min at room temperature. After evaporating the solvent under reduced pressure, SPE was applied to remove the remaining oxidizing agents and decomposed compounds. The mixture was fractionated by MPLC with a gradient solvent system of n-hexane−EtOAc to obtain 10 fractions (III/1−10). Fraction III/1 afforded compound 5 (1.6 mg) after purification by HPLC with MeCN−H 2 O (1:1) as mobile phase (flow rate 3 mL/ min). Fraction III/2 was rechromatographed by RP-HPLC, using an MeCN−H 2 O gradient system (flow rate 3 mL/min) as eluent to yield compounds 6a (3.0 mg) and 6b (1.85 mg). Fraction III/4 was found to be pure and gave compounds 7a,b (13.0 mg). Fraction III/6 was subjected to HPLC using MeCN−H 2 O (1:1) (flow rate 3 mL/min) as mobile phase and afforded compounds 7a,b (4.4 mg, 4.4%) and 8 (2.1 mg). Fraction III/7 was purified by RP-HPLC with MeCN−H 2 O (1:1) as eluent to give compound 9 (1.80 mg).
In reaction IV, 50 mg juncuenin B was dissolved in n-BuOH at a concentration of 1 mg/mL, and stirred with 2 equiv of PIFA (161,15 mg) for 30 min at room temperature. After evaporating the solvent under reduced pressure, prepurification with SPE was performed and the resulting mixture was fractionated by MPLC with an n-hexane− EtOAc gradient system to afford six fractions (IV/1−6). From fraction IV/1 compound 10 (1.80 mg) was isolated.
Journal of Natural Products pubs.acs.org/jnp Article For reaction V, 100 mg juncuenin B was dissolved in n-BuOH at a concentration of 1 mg/mL, and stirred with 2 equiv of PIDA (242 mg) for 30 min at room temperature. The solvent was evaporated under reduced pressure, and then SPE was used to absorb the remaining oxidizing agents and other impurities. The mixture was fractionated by MPLC with an n-hexane−EtOAc gradient solvent system to afford four fractions (V/1−4). Fraction V/2 was subjected to HPLC, eluting with MeCN−H 2 O (3:2) (flow rate 3 mL/min) as mobile phase to give compound 11 (5.5 mg).
Separation of Enantiomers. Racemic mixtures, which had high antiproliferative activity, were separated on a Lux amylose-1 chiralphase column in order to compare the antiproliferative activity of the pure enantiomers. For the separation of compounds 1a−d, cyclohexane−isopropanol (95:5) and for compounds 4a+b and 7a+b cyclohexane−isopropanol (8:2) solvent systems (15 mL/min) were used as eluent.
Computational Methods. Mixed torsional/low-mode conformational searches were carried out by means of the Macromodel 10.8.011 software using the MMFF with an implicit solvent model for CHCl 3 applying a 21 kJ/mol energy window. 24 Geometry reoptimizations of the resultant conformers (ωB97X/TZVP with PCM solvent model for MeCN) and TDDFT ECD calculations were performed with Gaussian 09. 25 For ECD using various functionals (B3LYP, BH&HLYP, CAM-B3LYP, PBE0) and the TZVP basis set with the same solvent model as in the preceding DFT optimization step. ECD spectra were generated as the sum of Gaussians with 3000 cm −1 half-height widths, using dipole-velocity computed rotational strengths. 26 Boltzmann distributions were estimated from the ωB97X energies. The MOLEKEL program was used for visualization of the results. 27 Antiproliferative (MTT) Assay. Growth-inhibition properties of racemates and pure enantiomer compounds were determined by standard MTT assays on six human malignant gynecological cell lines (MCF-7, T47D, HeLa, SiHa, C33A, A2780). To investigate tumor selectivity, the MTT assay was repeated on mouse embryonic fibroblast cells (NIH/3T3) under the same conditions. All cell lines were maintained in minimal essential medium (MEM) supplemented with 10% fetal bovine serum, 1% nonessential amino acids, and 1% penicillin−streptomycin−amphotericin B mixture, and were stored in humidified air containing 5% CO 2 at 37°C. All cell types were seeded into 96−well plates at a density of 5000 with the exception of C33a, which was seeded at a density of 10 000 and treated by increasing concentrations (0.1−30 μM) of the compounds for 72 h under cell culturing conditions. After the incubation, 5 mg/mL MTT [3-(4,5dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] solution was added to samples for 4 h and precipitated blue formazan crystals were dissolved in DMSO. Absorbance values of the samples were measured at 545 nm using a microplate reader (Stat Fax-2100, Awareness Technologies Inc., Palm City, FL, USA) and untreated cells were used as control. Normalized sigmoidal dose−response curves were fitted to the determined data and the IC 50 values were calculated by GraphPad Prism 5.01 (GraphPad Software, San Diego, CA, USA). Cisplatin (Ebewe Pharma GmbH, Unterach, Austria) was used as a reference agent in the same concentration range.