Abstract
The semisynthesis of xanthanolide derivatives is reported from xanthinin and 4-epi-isoxanthanol, two sesquiterpene lactones isolated from the crude chloroformic extract of the leaves of Xanthium macrocarpum DC. (Asteraceae) by liquid/liquid chromatography. In vitro evaluation of their protein farnesyltransferase (PFTase) inhibitory activity has been investigated. In contrast to other biological activities of xanthanolides, PFTase inhibition is not associated with the presence of the potentially toxic α-methylene-γ-lactone function.
Introduction
Xanthium macrocarpum DC. (Asteraceae) is a plant originating from America and widely distributed in the floodplain zone of rivers. It was traditionally used in France for the treatment of several diseases such as catarrh, scrofula, and leprosyCitation1. We earlier reported on the phytochemical study of its aerial parts that revealed the presence of several xanthanolidesCitation2. This subtype of sesquiterpene lactones is already known for its antimicrobialCitation3, antifungalCitation4, antimalarialCitation5, antileishmanialCitation6, cytotoxicCitation7,Citation8, and anti-ulcerogenic activitiesCitation9. A report also pointed out the potential of xanthanolides as protein farnesyltransferase inhibitorsCitation10. The protein farnesyltransferase (PFTase) is an enzyme responsible for the post-translational farnesylation of the oncogenic variants of Ras, and could be considered as a promising target for the treatment of cancerCitation11. However, these xanthanolides (e.g. 8-epi-xanthatin, ) with PFTase inhibitory activity mainly differ in their stereochemistry from the xanthanolides (e.g. xanthatin 2, ) isolated from Xanthium macrocarpum DC. The 8-epi-xanthatin bears a 7,8-cis fused lactone ring with a high degree of conformational flexibility, whereas xanthatin has a 7,8-trans fused ring and is therefore probably much more rigid. Moreover, it was reported that, among helenanolide sesquiterpene lactones, minor differences in the stereochemistry significantly affect the bioactivityCitation12. We therefore decided to evaluate the potential of 7,8-trans fused xanthanolides isolated from a natural source (1–3) and their synthetic derivatives (4–7) as PFTase inhibitors. The biological activities of sesquiterpene lactones are often explained by their ability to react in vivo as Michael acceptors with endogenous nucleophilesCitation13. In fact, we recently reported that in contrast to 1 or 3, reduced compounds such as 8 and 10 were devoid of any antifungal activityCitation4. The alkylating potential of α-methylene-γ-lactone is also related to the toxicity of several plants rich in sesquiterpene lactonesCitation14. Moreover, only artemisinin, an α-methyl sesquiterpene lactone, is nowadays used in therapeuticsCitation15. As far as PFTase activity is concerned, there are only a few reported examples of monomericCitation16 and dimericCitation17 sesquiterpene lactones with such activity in the α-methyl-γ-lactone series. Therefore, with the aim to obtain potentially atoxic compounds, we decided to prepare several 7,8-trans fused derivatives in the α-methyl-γ-lactone series, starting from natural products 1–3, and explore their ability to inhibit protein farnesyltransferase. Beforehand, it was necessary to purify compounds 1–3 present in the crude chloroformic extract from the leaves of Xanthium macrocarpum DC. by centrifugal partition chromatography (CPC), a liquid/liquid chromatography.
Scheme 1. Synthesis of compounds 2, 4, 5, and 7–14. Reagents: (a) SiO2, RT, 2 days, 100%; (b) PCC, CH2Cl2, refluxing, 8 days (4, 5%; 5, 4%); (c) dry air, tetraphenylporphine, CH2Cl2, hυ, 15°C, 5 h, 14%; (d) NaBH4, MeOH, 0°C, 3 h (8, 95%; 14, 95%); (e) AcCl, TEA, CH2Cl2, RT, 24 h, 12%; (f) DDQ, toluene, refluxing, 3h (10, 42%, 11, 2%; 12, 2%); (g) NaBH4, NiCl2, MeOH, 0°C, 3 h, 79%.
Figure 1. Xanthanolides exerting PFTase inhibitory activity.
Materials and methods
General
All reagents and solvents were general purpose grade. The dried leaves from Xanthium macrocarpum were crushed with a Retsch Muhle crushing mill. The leaf powder was extracted in a Soxhlet apparatus (6 L solvent capacity). The purification of 1–3 was performed using a R 1000 rotor (980 mL column) on a bench scale FCPC (fast centrifugal partition chromatograph; Kromaton Technologies) with a KP 100 (100 mL/min, 100 bar) pump (Armen-Kromaton Technologies). Sample was injected with a loop (Kromaton Technologies; total bed volume 50 mL) and fractions were collected with a Gilson FC-204 collector. Column chromatographies were performed on silica gel 60 (Macherey-Nagel; 230–400 mesh) and preparative thin layer chromatography (TLC) with silica gel plates (Macherey-Nagel; SIL G/UV254, 0.25 mm). Eluates were monitored by TLC on silica 60 F254 (Macherey-Nagel; SIL G/UV254, 0.20 mm). The spots were detected under ultraviolet (UV) light at 254 nm or using sulfuric vanillin as revealing agent. Infrared (IR) spectra were determined on a Bruker Fourier transform (FT) IR Vector 22 spectrometer using neat liquid films. 1H-nuclear magnetic resonance (NMR) spectra and 13C-NMR spectra were recorded in CDCl3 with, respectively, CHCl3 at 7.26 ppm and 77.0 ppm as the internal standard on a Bruker Avance DRX 500 (1H: 500 MHz; 13C: 125 MHz) machine or on a Jeol GSX 270 WB (1H: 270 MHz; 13C: 67.5 MHz) spectrometer. High-resolution electron ionization (EI) and electrospray ionization (ESI) mass spectroscopy analyses were done with a JMS-700 (Jeol) double focusing mass spectrometer with reversed geometry or an Orbitrap (Thermoelectron) machine in positive mode. Samples were solubilized with H2O–MeOH (1:1) and injected with a syringe pump.
Plant material
The leaves of Xanthium macrocarpum DC. were collected at Saintes-Gemmes sur Loire (France) in September 2003. A voucher specimen (XM 003) was deposited in the Department of Pharmacognosy, University of Angers, France.
Extraction
The dried and powdered leaves of X. macrocarpum DC. (9.0 kg) were extracted in several portions (~375 g) for 20 h with 6 L of CHCl3 in a Soxhlet apparatus. Evaporation of the solvent under reduced pressure gave 945 g of a crude chloroformic extract as a gum that was kept for 10 months before chromatography.
Isolation
Preparation of the two-phase solvent system
The same volumes of n-heptane, EtOAc, MeOH, and H2O were mixed and thoroughly equilibrated in a separation funnel (4 L) at room temperature until two phases were discernible (30 min). The two phases were separated shortly before use.
Purification procedure on FCPC
The hydrostatic FCPC column was fully filled in ascending mode with the lower phase (stationary phase, 980 mL) using the KP 100 pump at a flow rate of 60 mL/min while the apparatus was run at 500 rpm. Then, 50 mL of sample (120 mg/mL) was injected while the apparatus was run at 1100 rpm. The sample was eluted with 30 mL of stationary phase at 20 mL/min, then the upper phase was pumped into the tail end of the inlet column at a flow rate of 10 mL/min during 5 min, and then at 20 mL/min. Some 390 mL of stationary phase eluted at the head end outlet. Next, 132 fractions of 15 mL were collected. After all desired compounds, monitored by TLC, were eluted, the elution was stopped. Fractions F14–18, F19–28, and F31–40 were respectively combined, concentrated under reduced pressure, and extracted with CH2Cl2. The dichloromethanic extracts were dried over Na2SO4, filtered, and concentrated under reduced pressure. The three fractions were recrystallized in c-hexane to give 385 mg (F14–18) of 1 (0.64% of dried leaves), 285 mg (F19–28) of 2 (0.47%), and 310 mg (F31–40) of 3 (0.51%). The final characterization of compounds 1–3 was carried out by 1H-NMR, 13C-NMR, and comparison with literature data.
Semisynthesis
Xanthatin-1,5-epoxides 4 and 5Citation18 PCC (16.5 mg, 76 µmol) was added to a solution of xanthatin 2 (20 mg, 81 µmol) in dry CH2Cl2 (2 mL). The mixture was heated under reflux for 8 days. The precipitate was filtered on Celite®. CH2Cl2 was removed under reduced pressure and the residue was chromatographed over 5.4 g of silica gel (CH2Cl2–EtOAc, 9:1), yielding the xanthatin-1,5β-epoxide 4 (1 mg, 5% yield) and the xanthatin-1,5α-epoxide 5 (0.8 mg, 4% yield). 4: 1H-NMR (270 MHz, CDCl3): δ 6.84 (1H, d, J = 15.6 Hz, H-2), 6.24 (1H, d, J = 15.6 Hz, H-3), 6.21 (1H, d, J = 2.5 Hz, H-13α), 5.49 (1H, d, J = 2.5 Hz, H-13β), 4.26 (1H, m, H-8), 2.98 (1H, t, J = 4.6 Hz, H-5), 2.85 (1H, m, H-7 and H-10), 2.67 (1H, m, H-6α), 2.27 (3H, s, H-15), 2.01 (2H, m, H-6β and H-9α), 1.81 (1H, m, H-9β), 1.09 (3H, d, J = 7.1 Hz, H-14); 13C-NMR (67.5 MHz, CDCl3): δ 190.4 (C-4), 169.2 (C-12), 144.9 (C-2), 138.8 (C-11), 128.9 (C-3), 118.5 (C-13), 80.1 (C-8), 63.6 (C-1), 63.1 (C-5), 42.9 (C-7), 34.4 (C-9), 31.1 (C-10), 28.3 (C-15), 26.3 (C-6), 16.4 (C-14); MS (ESI+) m/z calcd for C15H19O4 263.1278. Found 263.1278 [M + H]+. 5: 1H-NMR (270 MHz, CDCl3): δ 6.92 (1H, d, J = 15.6 Hz, H-2), 6.30 (1H, d, J = 15.6 Hz, H-3), 6.18 (1H, d, J = 3.2 Hz, H-13α), 5.45 (1H, d, J = 3.2 Hz, H-13β), 3.93 (1H, ddd, J = 2.5, 9.9, 12.4 Hz, H-8), 3.07 (1H, d, J = 5.3 Hz, H-5), 2.88 (1H, m, H-7), 2.70 (1H, m, H-6α), 2.62 (1H, m, H-10), 2.29 (3H, s, H-15), 2.14 (2H, m, H-6β and H-9α), 2.00 (1H, m, H-9β), 1.30 (3H, d, J = 7.4 Hz, H-14); 13C-NMR (67.5 MHz, CDCl3): δ 190.4 (C-4), 169.2 (C-12), 144.8 (C-2), 138.8 (C-11), 128.9 (C-3), 118.5 (C-13), 79.9 (C-8), 63.6 (C-1), 63.1 (C-5), 42.9 (C-7), 34.2 (C-9), 31.1 (C-10), 28.2 (C-15), 26.3 (C-6), 16.4 (C-14).
Acetyl-4-epi-isoxanthanol 6Citation5 Triethylamine (370 µL, 2.6 mmol) and AcCl (190 µL, 2.6 mmol) were added in small portions during 4 days to a solution of 4-epi-isoxanthanol 3 (105 mg, 0.3 mmol) in dry CH2Cl2 (2 mL). The mixture was stirred at room temperature for 5 days and then washed with H2O (3 × 5 mL), dried with Na2SO4, and concentrated under reduced pressure. The residue was chromatographed over 6 g of silica gel (CH2Cl2–EtOAc, 9:1), yielding the acetyl-4-epi-isoxanthanol 6 (12.3 mg, 12% yield). 1H-NMR (270 MHz, CDCl3): δ 6.15 (1H, d, J = 3.2 Hz, H-13α), 5.92 (1H, dd, J = 2.8 and 8.5 Hz, H-5), 5.44 (1H, d, J = 3.2 Hz, H-13β), 5.18 (1H, dd, J = 4.6 Hz, H-2), 4.90 (1H, m, H-4), 4.27 (1H, ddd, J = 2.8, 9.6, 12.4 Hz, H-8), 2.77 (1H, m, H-10), 2.55 (1H, m, H-6α), 2.44 (1H, m, H-7), 2.31 (2H, m, H-6β and H-9α), 2.04 (3H, s, H-17), 2.02 (3H, s, H-19), 1.82 (3H, m, H-3 and H-9β), 1.25 (3H, d, J = 6.0 Hz, H-15), 1.10 (3H, d, J = 7.1 Hz, H-14); 13C-NMR (67.5 MHz, CDCl3): δ 170.5 (C-18), 170.0 (C-16), 169.7 (C-12), 145.6 (C-1), 139.3 (C-11), 126.3 (C-5), 118.5 (C-13), 82.1 (C-8), 67.2 (C-4), 67.1 (C-2), 47.9 (C-7), 39.9 (C-3), 36.8 (C-9), 29.6 (C-10), 25.3 (C-6), 21.2 (C-15), 21.1 (C19), 20.2 (C-17), 19.3 (14-C).
3,5-Oxyxanthatin 7 Dried air was bubbled through the solution of xanthatin 2 (70 mg, 0.28 mmol) and tetraphenylporphine (3 mg, 0.004 mmol) as the photo-sensitizer in CH2Cl2 (30 mL). The reaction mixture was cooled to 15°C and irradiated with a halogen lamp (500 W) for 5 h. CH2Cl2 was removed under reduced pressure and the residue was purified on preparative TLC (EtOAc–CH2Cl2, 4:6) as the eluent, yielding the 3,5-oxyxanthatin (7) (10 mg, 14% yield). 1H-NMR (270 MHz, CDCl3): δ 7.04 (1H, s, H-2), 6.30 (1H, d, J = 3.2 Hz, H-13α), 5.59 (1H, d, J = 3.2 Hz, H-13β), 4.37 (1H, ddd, J = 2.5, 9.2, 11.7 Hz, H-8), 3.45 (1H, m, H-6α), 3.19 (1H, m, H-10), 2.89 (1H, m, H-7), 2.74 (1H, m, H-6β), 2.44 (3H, s, H-15), 2.10 (2H, m, H-9), 1.28 (3H, d, J = 7.2, H-14); 13C NMR (67.5 MHz, CDCl3): δ 186.1 (C-4), 169.1 (C-12), 152.3 (C-3), 150.1 (C-5), 138.6 (C-11), 128.0 (C-1), 120.3 (C-2), 120.2 (C-13), 80.5 (C-8), 44.9 (C-7), 38.5 (C-9), 31.9 (C-6), 29.7 (C-10), 22.6 (C-15), 14.8 (C-14); MS (ESI+) m/z calcd for C15H17O4 261.1121. Found 261.1121 [M + H]+.
2,3-(E)-Dehydro-4-epi-isoxanthanol 9 Triethylamine (140 µL, 1 mmol) and AcCl (75 µL, 1 mmol) were added dropwise during 3 h to a solution of 4,11β,13-tetrahydroxanthatin 8 (58 mg, 230 µmol) in dry CH2Cl2 (2 mL). The mixture was stirred at room temperature for 24 h and was then washed with H2O (3 × 5 mL), dried with Na2SO4, and concentrated under reduced pressure. The residue was purified on preparative TLC (CH2Cl2–EtOAc, 9:1), yielding the 2,3-(E)-dehydro-4-epi-isoxanthanol 9 as an unseparable diastereomeric mixture (8.2 mg, 12% yield). 1H-NMR (270 MHz, CDCl3): δ 6.12 (1H, d, J = 15.6 Hz, H-2), 5.82 (1H, dd, J = 2.8, 8.8 Hz, H-5), 5.61 (1H, m, H-3), 5.43 (1H, m, H-4), 4.29 (1H, ddd, J = 1.8, 9.9, 11.7 Hz, H-8), 3.01 (1H, m, H-10), 2.45 (1H, m, H-6α), 2.30 (3H, m, H-6β, H-7 and H-11), 2.05 (3H, s, H-17), 1.66 (2H, m, H-9), 1.35 (3H, d, J = 6.4 Hz, H-15), 1.23 (3H, d, J = 6.7 Hz, H-13), 15 (3H, d, J = 7.4 Hz, H-14); 13C-NMR (67.5 MHz, CDCl3): δ 178.5 (C-12), 170.3 and 170.2 (C-16), 144.4 (C-1), 136.6 and 136.5 (C-2), 130.3 and 130.2 (C-3), 125.4 (C-5), 81.8 (C-8), 71.3 and 71.1 (C-4), 51.2 and 51.1 (C-7), 41.9 (C-11), 36.3 and 36.2 (C-9), 29.9 and 29.0 (C-10), 27.8 (C-6), 21.4 (C-15), 18.5 (C-14), 12.6 (C-13); MS (ESI+) m/z calcd for C17H25O4 293.1747. Found 293.1749 [M + H]+.
11b,13-Dihydroxanthatin 10 and 11b,13-dihydroxanthatin-1,5-epoxides 11 and 12 DDQ (100 mg, 0.45 mmol) was added to a solution of the alcohol 8 (113 mg, 0.45 mmol) in toluene (10 mL). The mixture was heated under reflux for 3 h. The black precipitate was filtered on Celite®. Toluene was removed under reduced pressure and the residue was pre-purified on preparative TLC (CH2Cl2–EtOAc, 9:1). The interesting fractions were pooled and purified on preparative TLC (CH2Cl2–EtOAc, 7:3) to afford the 11β,13-dihydroxanthatin 10 (47.1 mg, 42% yield), the 11β,13-dihydroxanthatin-1,5α-epoxide 11 (2.8 mg, 2% yield), and the 11β,13-dihydroxanthatin-1,5β-epoxide 12 (2.1 mg, 2% yield). (10): see reference 5. (11): 1H-NMR (270 MHz, CDCl3): δ 6.90 (1H, d, J = 15.6 Hz, H-2), 6.29 (1H, d, J = 15.6 Hz, H-3), 3.94 (1H, ddd, J = 2.8, 6.7, 11.7 Hz, H-8), 3.03 (1H, d, J = 6.0 Hz, H-5), 2.59 (1H, m, H-10), 2.48 (1H, m, H-6α), 2.29 (1H, m, H-11), 2.28 (3H, s, H-15), 2.07 (1H, m, H-9α), 2.04 (1H, m, H-7), 1.91 (1H, m, H-9β), 1.86 (1H, m, H-6β), 1.13 (3H, d, J = 7.4 Hz, H-14), 1.08 (3H, d, J = 7.0 Hz, H-13); 13C-NMR (67.5 MHz, CDCl3): δ 198.5 (C-4), 178.2 (C-12), 144.9 (C-2), 128.9 (C-3), 80.1 (C-8), 63.6 (C-1), 63.1 (C-5), 42.9 (C-7), 41.2 (C-11), 34.4 (C-9), 31.1 (C-10), 29.7 (C-15), 26.3 (C-6), 16.4 (C-14), 12.5 (C-13). (12): 1H-NMR (270 MHz, CDCl3): δ 6.78 (1H, d, J = 15.6 Hz, H-2), 6.23 (1H, d, J = 15.7 Hz, H-3), 4.17 (1H, ddd, J = 4.2, 9.9, 10.3 Hz, H-8), 2.95 (1H, t, J = 6.0 Hz, H-5), 2.75 (1H, m, H-10), 2.48 (1H, m, H-6α), 2.32 (1H, m, H-11), 2.27 (3H, s, H-15), 2.04 (1H, m, H-9α), 1.92 (1H, m, H-7), 1.88 (1H, m, H-9β), 1.78 (1H, m, H-6β), 2.00 (3H, d, J = 6.0 Hz, H-14), 1.88 (3H, d, J = 7.4 Hz, H-13); 13C-NMR (67.5 MHz, CDCl3): δ 198.5 (C-4), 178.2 (C-12), 144.9 (C-2), 128.9 (C-3), 79.8 (C-8), 63.6 (C-1), 63.1 (C-5), 42.9 (C-7), 41.2 (C-11), 34.2 (C-9), 31.1 (C-10), 29.6 (C-15), 26.3 (C-6), 16.4 (C-14), 12.5 (C-13).
11,13-Dihydroxanthinosin 13 NaBH4 (13 mg, 0.343 mmol) was added in small portions during 15 min to a solution of nickel(II) chloride (48.4 mg, 0.375 mmol) and xanthatin 2 (50.3 mg, 0.204 mmol) in MeOH (6 mL) at 0°C. After 3 h, the black precipitate was filtered on Celite®, and the mixture was acidified to pH 4 with 10% HCl. MeOH was removed under reduced pressure and the aqueous solution was extracted with CH2Cl2 (3 × 10 mL). The organic extract was successively washed with a saturated aqueous solution of NaHCO3 (2 × 10 mL), 20% aqueous solution of NH4Cl (2 × 10 mL), brine (2 × 10 mL), and H2O (3 × 20 mL). The dichloromethanic extract was dried with Na2SO4, concentrated under reduced pressure, and purified by column chromatography over silica gel (CH2Cl2–EtOAc, 9:1) to afford the 11,13-dihydroxanthinosin 13 as an unseparable diastereomeric mixture (40.5 mg, 79% yield). 1H-NMR (270 MHz, CDCl3): δ 5.48 (1H, dd, J = 1.8 and 8.5 Hz, H-5), 4.26 (1H, ddd, J = 2.8, 9.6, 12.4 Hz, H-8), 2.53 (2H, m, H-3), 2.46 (1H, m, H-10), 2.27 (4H, m, H-2, H-6α, H-9β, H-11), 2.16 (3H, s, H-15), 1.99 (1H, m, H-6β), 1.60 (2H, m, H-7 and H-9β), 1.21 (3H, d, J = 6.7 Hz, H-13), 1.12 (3H, d, J = 7.1 Hz, H-14); 13C-NMR (67.5 MHz, CDCl3): δ 208.0 (C-4), 178.6 (C-12), 146.2 (C-1), 122.2 (C-5), 81.2 (C-8), 51.6 (C-7), 42.6 (C-3), 41.8 (C-11), 36.5 (C-9), 34.5 (C-15), 34.1 (C-6), 30.0 (C-2), 27.1 (C-10), 18.2 (C-14), 12.6 (C-13); MS (ESI+) m/z calcd for C15H23O3 251.1642. Found 251.1642 [M + H]+.
4,11,13-Tetrahydroxanthinosin 14Citation19 NaBH4 (6 mg, 0.158 mmol) was added to a solution of 11,13-dihydroxanthinosin 13 (20 mg, 0.079 mmol) in MeOH (4 mL) at 0°C. After 3 h, the mixture was acidified to pH 4 with a 10% aqueous solution of HCl. MeOH was removed under reduced pressure and the aqueous solution was extracted with CH2Cl2 (3 × 10 mL). The organic extract was successively washed with saturated aqueous NaHCO3 (2 × 10 mL), 20% aqueous NH4Cl (2 × 10 mL), brine (2 × 10 mL), and H2O (3 × 20 mL). The dichloromethanic extract was dried with Na2SO4, concentrated under reduced pressure, and purified by column chromatography over silica gel (CH2Cl2–EtOAc, 9:1) to afford the 4,11,13-tetrahydroxanthinosin 14 as an unseparable diastereomeric mixture (19 mg, 95% yield). 1H-NMR (270 MHz, CDCl3): δ 5.51 and 5.24 (1H, dd, H-5), 4.22 (1H, m, H-8), 3.8 (1H, m, H-4), 2.50 (1H, m, H-10), 2.22 (4H, m, H-3, H-6α and H-11), 2.02 (3H, m, H-2, H-6β and H-9α), 1.53 (2H, m, H-7 and H-9β), 1.19 (3H, d, J = 6.7 Hz, H-14), 1.17 (3H, d, J = 5.7 Hz, H-15), 1.12 (3H, d, J = 7.4 Hz, H-13); 13C-NMR (67.5 MHz, CDCl3): δ 178.9 and 178.7 (C-12), 147.6, 147.3, 143.1 and 142.6 (C-1), 123.9, 123.7, 121.9 and 121.8 (C-5), 82.1 and 81.7 (C-8), 68.0, 67.8, 67.7 and 67.5 (C-4), 51.7, 51.6 and 48.8 (C-7), 42.4 and 42.3 (C-11), 39.2 and 39.1 (C-3), 38.0 and 37.2 (C-9), 31.2 and 31.1 (C-6), 29.9 and 29.8 (C-10), 27.4 and 27.2 (C-2), 23.7 and 23.0 (C-15), 18.2 and 17.8 (C-14), 12.9 and 12.5 (C-13); MS (ESI+) m/z calcd for C15H25O3 253.1798. Found 253.1798 [M + H]+.
Acetyl-11b,13-dihydro-4-epi-isoxanthanol 15Citation20 NaBH4 (20 mg, 0.172 mmol) was added in small portions to a solution of acetyl-4-epi-isoxanthanol 6 (30 mg, 0.086 mmol) in MeOH (4 mL) at 0°C. After 4 h, the mixture was acidified to pH 4 with 10% HCl. MeOH was removed under reduced pressure and the aqueous solution was extracted with CH2Cl2 (3 × 10 mL). The organic extract was successively washed with saturated aqueous NaHCO3 (2 × 10 mL), 20% aqueous NH4Cl (2 × 10 mL), brine (2 × 10 mL), and H2O (3 × 20 mL). The dichloromethanic extract was dried with Na2SO4, concentrated under reduced pressure, and purified by column chromatography over silica gel (CH2Cl2–EtOAc, 9:1) to afford the acetyl-11β,13-dihydro-4-epi-isoxanthanol 15 (30 mg, 98% yield). 1H-NMR (270 MHz, CDCl3): δ 5.87 (1H, dd, J = 3.2, 9.2 Hz, H-5), 5.18 (1H, dd, J = 4.6, 9.6 Hz, H-2), 4.91 (1H, m, H-4), 4.29 (1H, ddd, J = 3.2, 9.6, 12.7 Hz, H-8), 2.74 (1H, m, H-10), 2.30 (3H, m, H-6α, H-9β and H-11), 2.04 (3H, s, H-17), 2.05 (3H, s, H-19), 1.82 (3H, m, H-3 and H-6β), 1.62 (2H, m, H-7 and H-9β), 1.22 (6H, m, H-13 and H-15), 1.11 (3H, d, H-14); 13C-NMR (67.5 MHz, CDCl3): δ 178.3 (C-12), 170.4 (C-18), 170.0 (C-16), 145.2 (C-1), 126.8 (C-5), 81.9 (C-8), 76.5 (C-4), 67.1 (C-2), 51.2 (C-7), 41.6 (C-11), 39.8 (C-3), 36.3 (C-9), 29.6 (C-10), 26.4 (C-6), 22.7 (C-15), 21.2 (C-19), 20.3 (C-17), 19.1 (C-14), 12.5 (C-13); MS (ESI+) m/z calcd for C21H27O6Na 375.1784. Found 375.1793 [M + Na]+.
11b,13-Dihydro-4-epi-isoxanthanol 16Citation20 NaBH4 (38 mg, 1 mmol) was added in small portions to a solution of 4-epi-isoxanthanol 3 (80 mg, 0.261 mmol) in MeOH (5 mL) at 0°C. After 6 h, the mixture was acidified to pH 4 with 10% HCl. MeOH was removed under reduced pressure and the aqueous solution was extracted with CH2Cl2 (3 × 10 mL). The organic extract was successively washed with saturated aqueous NaHCO3 (2 × 10 mL), 20% aqueous NH4Cl (2 × 10 mL), brine (2 × 10 mL), and H2O (3 × 20 mL). The dichloromethanic extract was dried with Na2SO4, concentrated under reduced pressure, and purified by column chromatography over silica gel (CH2Cl2–EtOAc, 7:3) to afford the 11β,13-dihydro-4-epi-isoxanthanol 16 (30 mg, 0.097 mmol) in 37% yield. 1H-NMR (270 MHz, CDCl3): δ 5.87 (1H, dd, J = 2.8, 8.8 Hz, H-5), 5.35 (1H, dd, J = 3.2, 9.5 Hz, H-2), 4.29 (1H, ddd, J = 2.5, 9.6, 12.0 Hz, H-8), 3.71 (1H, m, H-4), 2.75 (1H, m, H-10), 2.30 (3H, m, H-6α, H-9β and H-11), 2.09 (3H, s, H-17), 1.81 (3H, m, H-3 and H-6β), 1.58 (2H, m, H-7 and H-9α), 1.22 (6H, m, H-13 and H-15), 1.15 (3H, d, J = 7.1 Hz, H-14); 13C-NMR (67.5 MHz, CDCl3): δ 171.2 (C-12), 169.8 (C-16), 145.8 (C-1), 126.2 (C-5), 82.1 (C-8), 77.2 (C-2), 63.9 (C-4), 51.3 (C-7), 48.3 (C-3), 43.4 (C-11), 38.8 (C-9), 29.9 (C-6), 25.3 (C-15), 21.2 (C-10), 19.8 (C-17), 19.3 (C-14), 12.5 (C-13); MS (ESI+) m/z calcd for C17H26O5Na 333.1678. Found 333.1678 [M + Na]+.
Fluorescence-based PFTase inhibition assay
Assays were realized on 96-well plates, prepared with Biomek NKMC and Biomek 3000 from Beckman Coulter and read on a Wallac Victor fluorimeter from PerkinElmer. Per well, 20 µL of farnesyl pyrophosphate (10 µM) was added to 180 µL of a solution containing 2 µL of varied concentrations of tested compounds (dissolved in methanol) and 178 µL of a solution composed by 0.1 mL of partially purified recombinant yeast FTase (2.2mg/mL) and 7.0 mL of Dansyl-GCVLS peptide (in the following buffer: 5.8 mM DTT, 12 mM MgCl2, 12 µM ZnCl2, and 0.09% (w/v) CHAPS, 53 mM Tris/HCl, pH 7.5). Then the fluorescence development was recorded for 15 min (0.7 s per well, 20 repeats) at 30°C with an excitation filter at 340 nm and an emission filter at 486 nm.
Results and discussion
Purification
The liquid/liquid chromatography, CPC, is a method of purification which presents some advantages compared to the traditional solid/liquid chromatography. Indeed, the absence of a solid support avoids irreversible adsorption (no sample loss) or artifact formation. A reduction of solvent consumption can also be expected and the scaling up from analytical to preparative scale is easy and reliable. The centrifugal partition chromatography of the crude chloroformic extract of the leaves of X. macrocarpum DC. was optimized with a 1 L bench scale FCPC®Citation21 and the compounds 1–3 were obtained in, respectively, 0.64%, 0.47%, and 0.51% yield.
Chemistry
Synthesis in the α-methylene-γ-lactone series mainly focused on modification of the side chain realized by oxidation of the dienonic part of xanthatin 2 () and led to a new tricyclic furan 7 and two epoxides 4 and 5 already isolated from Xanthium strumarium L.Citation18.
Thus, a multigram-scale, solvent-free, and quantitative β-elimination of 1 was realized in 2 days on silica to yield the dienone 2Citation22. To achieve the 1,5-oxidized analogs of 2, we initially planed to use a classical oxidant such as magnesium monoperoxyphthalateCitation23. However, under this condition, no evolution was observed with the deactivated dienone 2. We then attempted synthesis of the 1,5-oxidized analogs of 2 by a photooxygenation process in the presence of a photosensitizer, a method already applied to other dienonesCitation24. Instead of the desired epoxide, the furan 7 was obtained, probably via the postulated endoperoxide 17Citation25. A chromium epoxidation of deactivated alkenes (unsaturated acids) was recently reportedCitation26. Therefore, the PCC oxidation of 2 was attempted to finally, slowly (8 days) lead to the desired epoxides 4 and 5, albeit in low yield. Then, to evaluate the influence of the side chain substituents on the PFTase inhibitory activity, the 4-hydroxyl group of 4-epi-isoxanthanol 3 was acetylated to give 6 ()Citation27. In contrast to the xanthanolides in the α-methylene-γ-lactone series, α-methyl-γ-lactone xanthanolides are not so frequently isolated from natural sources. Moreover, their total synthesis is only realized in modest overall yield due to the difficulty in obtaining the fused rings and their asymmetric centersCitation28. We therefore thought that preparation of α-methyl-γ-lactone xanthanolides would be easier through semisynthesis from the α-methylene-γ-lactones 1 and 3 isolated from X. macrocarpum DC. Indeed, a small library of α-methyl-γ-lactones (8–16) was synthezised. Among them, 9, 11, and 12 are original compounds whereas 13–16 were already isolated from Stevia isomeca GrashoffCitation19, Iva dealbata A. GrayCitation20, and Dittrichia graveolens (L.) GreuterCitation29, but never prepared by semisynthesis.
Scheme 2. Synthesis of compounds 6, 15, and 16. Reagents: (a) AcCl, TEA, RT, 12%; (b) NaBH4, RT (16, 37%; 15, 98%).
Reacting with sodium borohydride, only one of the three carbon–carbon double bonds of 2 was reduced, as well as the ketonic carbonyl group, to yield 8. It should be noted that using this achiral reducing agent, the α-methylene-γ-lactone reduction appeared as stereoselective, as already noted in the case of arteannuin BCitation30. Under the same reduction condition, 3 behaved similarly, yielding the stereomer 16 (). On the other hand, such diastereoselectivity was not observed when the NaBH4 reduction of 2 was realized in the presence of nickel chloride. Moreover, with this reagent, a second reduction affected the C2–C3 double bond instead of the ketonic carbonyl, yielding the non-conjugated ketone 13. Therefore, synthesis of alcohol 14 necessitated an extra reductive step (NaBH4). Non-hydroxylic derivatives of 8 were also prepared either by its acetylation to 9 or by its oxidation (DDQ). Under those conditions, the expected dihydroxanthatine 10 was also accompanied by the minor epoxides 11 and 12. A higher structural diversity in the α-methyl-γ-lactone series could be obtained with the synthesis of reduced derivatives of 3 (). Besides the abovementioned alcohol 16, the synthesis of the diacetate 15 was also realized starting from 3.
Farnesyl transferase inhibition
In a first step, the IC50 values of compounds 1–16 were evaluated toward recombinant yeast PFTase ()Citation31. Compounds 1–16 could be sorted into two categories: (i) α-methylene-γ-lactone series (compounds 1–7) and (ii) α-methyl-γ-lactone series (compounds 8–16). Comparison of the PFTase activities of the 7,8-trans lactones 2 (IC50 = 100 µM) and 5 (IC50 = 100 µM) with the literature data for their corresponding cis fused isomers (), respectively, 8-epi-xanthatin (IC50 = 64 µM) and 8-epi-xanthatin-1,5α-epoxide (IC50 = 58 µM), indicated that stereochemistry at C7 and C8 had only a minor impact on the level of activity. Therefore, the influence of the modification of the western part of the xanthanolide skeleton was evaluated in the α-methylene-γ-lactone series and then confirmed in the α-methyl-γ-lactone series. Analysis of group (i) revealed that the compounds bearing a hydroxyl group at the 4-position of the side chain were inactive. The IC50 of alcohol 3 was above 1000 µM. The disfavorable effect of a hydrogen bond donor was also confirmed by comparison of the PFTase inhibitory activity of the 4-hydroxyl derivative 3 with the corresponding 4-oxo derivative 1 (IC50 = 150 µM). Moreover, except for 7, the 4-oxo derivatives 2, 4, and 5 were more active than 3. The lack of activity of the tricyclic compound 7 (IC50 >1000 µM) indicated the disfavorable effect of the backbone-rigidification supported by the furanic ring, in contrast to its deoxy precursor 2 (IC50 = 100 µM). Moreover, the 1,5-oxidized analogs of 2, the tricyclic epoxides 4 and 5, were also more active than the furanic derivative 7. However, the orientation of the ketonic side chain proved to be an important factor, since isomer 5 (IC50 = 100 µM) was more potent than 4 (IC50 = 500 µM). The favorable parameters for PFTase inhibitory activity were then confirmed in the α-methyl-γ-lactone series. Indeed, the 4-hydroxyl derivatives 8, 14, and 16 were less active (IC50 >500 µM) than the 4-oxo derivatives (10–13) (IC50 < 240 µM). Among them, the two oxiranes 11 (IC50 = 90 µM) and 12 (IC50 = 66 µM) were the most potent with the same level of activity. Again, the 4-acetyl derivatives 9 and 15 were more active than the corresponding alcohols, respectively, 8 and 16. Therefore, as in the case of the dimeric arteminolides () and contrary to what is observed for numerous other biological activities, we clearly demonstrated that the unsaturated α-methylene-γ-lactone was not required for the inhibition of PFTase. Moreover, reduction of the α-methylene-γ-lactone does not affect the inhibitory activity as exemplified by compounds 6 and 15, which exhibited the same inhibitory activity (IC50 = 40 µM).
Table 1. Effects of compounds 1–16 on PFTase.
In order to clarify the possible mode of action of this class of inhibitors, compounds 12 and 15, with the most significant PFTase inhibitory activity in the α-methyl-γ-lactone series, were selected for further investigation. Thus, 12 and 15 were subjected to kinetic studies, in a fluorescence-based assay. Analysis of the Lineweaver–Burk plots ( and ) clearly shows that compounds 12 and 15 are non-competitive inhibitors toward both peptide and farnesyl diphosphate (FPP). These results show that these compounds bind neither to the peptide nor to the FPP binding site. However, it has been shown that before leaving the enzyme, the farnesylated peptide binds to an exit groove that might be the binding site of the studied compoundsCitation32. A weak cytotoxic activity was also revealed against human colon tumor cell line HCT 116, with 11.4 and 7.4% for 12 and 12.3 and 7.3% for 15 at, respectively, 50 μM and 10 μM.
Figure 2. Lineweaver–Burk plots for the inhibition of PFTase isolated from yeast by compound 15 in the presence of dansyl-GCVLS peptide and FPP as substrates. Concentrations of inhibitor 15 were 0, 40, and 100 μM with excess FPP (A), or with excess dansyl-GCVLS (B).
Figure 3. Lineweaver–Burk plots for the inhibition of PFTase isolated from yeast by compound 12 in the presence of dansyl-GCVLS peptide and FPP as substrates. Concentrations of inhibitor 12 were 0, 60, and 100 μM with excess FPP (A), or with excess dansyl-GCVLS (B).
Conclusion
Sixteen xanthatin derivatives were investigated as potent inhibitors of protein farnesyltransferase. The results showed that, like the previously reported 7,8-cis fused xanthanolides, 7,8-trans derivatives also inhibit the PFTase. Six compounds showed inhibitory activity with IC50 values between 40 and 100 µM in both the α-methylene-γ-lactone (2, 5, and 6), and the α-methyl-γ-lactone series (11, 12, and 15). Moreover, the structure–activity relationships of the western part of the xanthanolide core revealed several features favorable to PFTase inhibitory activity such as the lack of a polar function or the presence of a 1,5-epoxide. Kinetic experiments illustrated that the two selected α-methyl-γ-lactone derivatives are non-competitive inhibitors of the peptide or FPP. This study clearly indicated that the α-methylene-γ-lactone group, already known to be responsible for side effects such as allerginicity, is not required to insure PFTase inhibitory activity. Although they have to be confirmed by cytotoxic assays, these results suggest that α-methyl-γ-lactone could represent a new lead for the development of non-competitive PFTase inhibitors, which are seldom studied.
Acknowledgements
We are grateful to Kromaton Technologies (Angers, France) for FCPC access, Dr Denis Lesage (LCSOB, UMR 7613, Université Pierre et Marie Curie, Paris VI, France) for HRMS Orbitrap access, David Rondeau (Service commun d’analyses spectroscopiques, Université d’Angers, France) for MS analysis, and Jean-Bernard Créchet (ICSN-CNRS, Gif sur Yvette, France) for the generous gift of the E. Coli strain expressing yeast PFTase and to Dr Jamal Ouazzani and Sylvie Cortial for its production and isolation.
Declaration of interest
This research was supported by a grant from the “Région des Pays de la Loire.” The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.
References
- Bonnier G. In: Orlhac E, ed. Flore Complète Illustrée en Couleur de France, Suisse et Belgique, Vol. 7. Paris: Librairie générale de l’enseignement, 1934:29–30.
- Lavault M, Bruneton J.Xanthium macrocarpum D.C. xanthanolides. Ann Pharm Fr 1979;37:59–63.
- Sato Y, Oketani H, Yamada T, Singyouchi KI, Ohtsubo T, Kihara M, et al. A xanthanolide with potent antibacterial activity against methicillin-resistant Staphylococcus aureus. J. Pharm Pharmacol 199;49:1042–4.
- Pinel B, Landreau A, Larcher G, Bouchara JP, Séraphin D, Richomme P. Synthesis of reduced xanthatin derivatives and in vitro evaluation of their antifungal activity. J Enzyme Inhib Med Chem 2005;20:575–9.
- Joshi SP, Rojatkar SR, Nagasampagi BA. Antimalarial activity of Xanthium strumarium. J Med Aromat Plant Sci 1997;19:366–8.
- Lavault M, Landreau A, Larcher G, Bouchara JP, Pagniez F, Le Pape P, et al. Antileishmanial and antifungal activities of xanthanolides isolated from Xanthium macrocarpum. Fitoterapia 2005;13:363–6.
- Cui scale B, Lee YH, Chai H, Tucker JC, Fairchild CR, Raventos-Suarez C, et al. Cytotoxic sesquiterpenoids from Ratibida columnifera. J Nat Prod 1999;62:1545–50.
- Chao Y, Chengshan Y, Zhongjian J. Xanthanolides, germacranolides, and other constituents from Carpesium longifolium. J Nat Prod 2003;66:1554–7.
- Favier LS, Maria AOM, Wendel GH, Borkowski EJ, Giordano OS, Pelzer L, et al. Anti-ulcerogenic activity of xanthanolide sesquiterpenes from Xanthium cavanillesii in rats. J Ethnopharmacol 2005;100:260–7.
- Kim YS, Kim JS, Park SH, Choi SU, Lee CO, Kim SK, et al. Two cytotoxic sesquiterpene lactones from the leaves of Xanthium strumarium and their in vitro inhibitory activity on farnesyltransferase. Planta Med 2003;69:375–7.
- Adjei AA. Farnesyltransferase inhibitors. Update Cancer Ther 2006;1:17–23.
- Beekman AC, Woerdenbag HJ, Uden WV, Pras N, Konings AWT, Wikström HV, et al. Structure-cytotoxicity relationships of some helenanolide-type sesquiterpene lactones. J Nat Prod 1997;60:252–7.
- Bruneton J. Sesquiterpenoid lactones. In: Pharmacognosy, Phytochemistry, Medicinal Plants, 2nd ed. Paris: Intercept Ltd, 1999:619–36.
- Bruneton J. Asteraceae. In: Toxic Plants Dangerous to Humans and Animals. Paris: Intercept Ltd, Techniques et Documentation, 1999:159–82.
- Haynes RK, Vonwiller SC. From Qinghao, marvelous herb of antiquity, to the antimalarial trioxane qinghaosu and some remarkable new chemistry. Acc Chem Res 1997;30:73–9.
- Oh HM, Kwon BM, Baek NI, Kim SH, Lee JH, Eun JS, et al. Inhibitory activity of 6-O-angeloylprenolin from Centipeda minima on farnesyl protein transferase. Arch Pharmacol Res 2006;29:64–6.
- Lee SH, Kim HK, Seo JM, Kang HM, Kim JH, Son KH, et al. Arteminolides B, C, and D, new inhibitors of farnesyl protein transferase from Artemisia argyi. J Org Chem 2002;67:7670–5.
- Saxena VK, Mishra M. Xanthanolides from Xanthium strumarium. Fitoterapia 1995;66:159–61.
- Bohlmann F, Umemoto K, Jakupovic J. Pseudoguaianolides related to confertin from Stevia isomeca. Phytochemistry 1985;24:1017–19.
- Chikamatsu H, Herz W. Ivalbatin, a new xanthanolide from Iva dealbata. J Org Chem 1973;38:585–90.
- Pinel B, Audo G, Mallet S, Lavault M, Poype FDL, Séraphin D, et al. Multi-grams scale purification of xanthanolides from Xanthium macrocarpum. Centrifugal partition chromatography versus silica gel chromatography. J Chromatogr A 2007;1151:14–19.
- Majewski M, Lazny R. Synthesis of tropane alkaloids via enantioselective deprotonation of tropinone. J Org Chem 1995;60:5825–30.
- Biscoe MR, Breslow R. Oxaziridinium salts as hydrophobic epoxidation reagents: remarkable hydrophobically-directed selectivity in olefin epoxidation. J Am Chem Soc 2005;127:10812–13.
- Nishio T, Mathies P, Job K, Frei B, Jeger O. Photochemical reactions 151st communication. Photo-oxygenation of (E)- and (Z)-7-methyl-β-ionone. Helv Chim Acta 1989;72:943–51.
- Juo RR, Herz W. Photooxygenation of (R)-p-mentha-3,8(9)-diene, and 1-isopropenyl-3,4-dihydronaphthalenes. Preparation of (R)-menthofuran, (R)-evodone, and (±)-chromolaenin. J Org Chem 1985;50:700–3.
- Kumbhat R, Sharma V. Kinetics and mechanism of the oxidation of some unsaturated acids by pyridinium chlorochromate. J Indian Chem Soc 2004;81:745–9.
- Marco JA, Sanz-Cerveza JF, Corral J, Carda M, Jakupovic J. Xanthanolides from Xanthium: absolute configuration of xanthanol, isoxanthanol and their C-4 epimers. Phytochemistry 1993;34:1569–76.
- Kummer DA, Brenneman JB, Martin SF. Domino intramolecular enyne metathesis/cross metathesis approach to the xanthanolides. Enantioselective synthesis of (+)-8-epi-xanthatin. Tetrahedron 2006;62:11437–49.
- Lanzetta R, Lama G, Mauriello G, Parrilli M, Raccioppi R, Sodano G. Ichthyotoxic sesquiterpenes and xanthanolides from Dittrichia graveolens. Phytochemistry 1991;30:1121–4.
- Bhattacharya AK, Pal M, Jain DC, Joshi BS, Roy R, Rychlewska U, et al. Stereoselective reduction of arteannuin B and its chemical transformations. Tetrahedron 2003;59:2871–6.
- Crechet JB, Jacquet E, Bernardi A, Parmeggiani A. Analysis of the role of the hypervariable region of yeast Ras2p and its farnesylation in the interaction with exchange factors and adenylyl cyclase. J Biol Chem 2000;275:17754–61.
- Long SB, Casey PJ, Beese LS. Reaction path of protein farnesyltransferase at atomic resolution. Nature 2002;419:645–50.