SUPPORTING INFORMATION Amorfrutins are potent antidiabetic dietary natural products Christopher Weidner1,2, Jens C. de Groot3, Aman Prasad4, Anja Freiwald1, Claudia Quedenau1, Magdalena Kliem1, Annabell Witzke1, Vitam Kodelja1, Chung-Ting Han1, Sascha Giegold5, Matthias Baumann5, Bert Klebl5, Karsten Siems6, Lutz Müller-Kuhrt6, Annette Schürmann7, Rita Schüler7, Andreas F. H. Pfeiffer7, Frank C. Schroeder4, Konrad Büssow3 and Sascha Sauer1. SUPPORTING MATERIAL AND METHODS Compounds and natural products library Compounds were purchased from the following sources: rosiglitazone (Cayman, Biozol, Eching, Germany), nTZDpa (Tocris, Biozol, Eching, Germany), pioglitazone (Sigma Aldrich, Taufkirchen, Germany), telmisartan, troglitazone, GW0742, GW7647, T0901317 (all from Sigma-Aldrich, Taufkirchen, Germany), amorfrutin 1 (NP-003520), amorfrutin 2 (NP-003521), amorfrutin 3 (NP- 006430), amorfrutin 4 (NP-009525), natural product library (all available from Analyticon Discovery, Potsdam, Germany). We used a diverse library of natural products from Analyticon Discovery. This library consisted of approximately 8,000 natural products derived from plants and microorganisms. The inclusion of pure plant derivatives and microbial metabolites represented a variety of different substance classes and structures. Purity of natural compounds was determined by high performance liquid chromatography (HPLC) or nuclear magnetic resonance (NMR) spectroscopy and on average 95 % was achieved. Structural elucidation was performed by NMR and liquid chromatography coupled to mass spectrometry (LC/MS). The amorfrutins were isolated from roots of Glycyrrhiza foetida (approximately 3.5g per kg plant material) and alternatively from fruits of Amorpha fruticosa (approximately 500mg per kg plant material) using organic extraction and iterative HPLC separation of organic fractions. Amorfrutin 1 was additionally synthesized in-house as described below. Protein expression for screening For bacterial expression we transformed the hPPARγ plasmid (pET28a vector (Novagen); kindly provided by Krister Bamberg (19)) in BL21 (DE3) cells. The clone was grown to an OD600 of 1.5 in SB medium (12 g/l Bacto-tryptone, 24 g/l yeast extract, 0.4% (v/v) glycerol, 17 mM KH2PO4, 72 mM K2HPO4), supplemented with 20 µg/ml thiamine and 100 µg/ml ampicillin in shaker flasks. Protein expression was induced with 0.1 mM isopropyl-beta-D-thiogalactopyranoside (IPTG) for 20 hours at 18 °C. All further steps were performed at 4-8 °C. Cells were pelleted by centrifugation and resuspended in a 3-fold volumes of buffer A (20 mM Tris, pH 8, 150 mM NaCl, 1 mM TCEP, 10 % glycerol), supplied with a protease inhibitor cocktail (Roche, Mannheim, Germany) and 500 units Benzonase (Merck, Darmstadt, Germany). Cells were lysed by sonification, followed by centrifugation (23,000 g for 1 45 min) and filtration through a 0.45 µm syringe filter. The filtrate was applied to an 8 ml Co2+ charged TALON Superflow Metal Affinity column (Clontech, Heidelberg, Germany). After washing with 10 column volumes of buffer A the protein was eluted with a gradient of imidazole (10-300) mM over 10 column volumes. After 10-fold dilution with buffer A the protein was applied to a 1.6 ml HQ-POROS20 column (Applied Biosystems, Darmstadt, Germany). Pure hPPARγ was collected in the flow-through and concentrated with Biomax concentrators (10 kDa cut off, Millipore, Schwalbach, Germany) to a final concentration of 2 mg/ml. The purified and concentrated protein was checked for monomeric state by dynamic light scattering (Laser Scatter201, RiNA GmbH, Berlin, Germany). For subsequent screening experiments the buffer was changed to 10 mM ammonia acetate using 10K Nanosep centrifugal devices (Pall Life Sciences, Dreieich, Germany). Briefly, the Nanosep device was equilibrated in 500 µl of the corresponding buffer and centrifuged at 11,000 g for 30 min at 4 °C. The PPARγ-ligand binding domain (LBD) sample was transferred to the filter and centrifuged again. The remaining residual volume was recovered in the new buffer. This procedure was repeated four more times. Mass spectrometry screening of large libraries of natural products Dried natural products were dissolved in 100 % DMSO to a concentration of 10 mM and stored at -20 °C. For high-throughput screening the compounds were diluted 1:10 in water and subsequently 1.5 nmol were added to 40 µg of PPARγ diluted in a final volume of 100 µl 10 mM ammonium acetate buffer. Incubation was performed in a Thermowell 96-well PCR plate (Corning Life Science, Fisher Scientific, Schwerte, Germany) at room temperature for 30 min. Curcumin and troglitazone were used as positive controls. As a negative control experiment, natural products were incubated in the absence of protein to check for unspecific binding. After incubation the samples were transferred from 96-well plates to 96- well MultiScreen Ultracel filter plates (Millipore) and centrifuged at 2,250 g for 60 min at room temperature. The flow-through was discarded. The filters were washed with 120 µl water and centrifuged again. The elution of protein-binding natural products was carried out using 120 µl 50 % acetonitrile solution. The eluate was collected in U96 MicroWell plates (Nunc, Wiesbaden, Germany) during centrifugation. Subsequently, the samples were dried and frozen at -20 °C for storage or directly diluted in 10 µl 50 % acetonitrile solution and shaken at room temperature for 30 min prior to electrospray ionisation mass spectrometry (ESI-MS) measurement. The U96 MicroWell plate was covered by heat sealing. Four µl of the 50 % acetonitrile sample solution were subjected to electrospray ionisation mass spectrometry, in which a high capacity ion trap plus mass spectrometer HCTultra (Bruker Daltonics, Bremen, Germany) was coupled to an HPLC-Workstation (Agilent, Böblingen, Germany). The flow rate was set to 50 µl/min followed by scanning in alternate ion mode from 200 to 1500 m/z. Data were analysed with Data Analysis 3.4 software (Bruker Daltonics). Signal height differences of probe signals and negative controls of at least one order of magnitude difference were considered as potential hits. Spectra were evaluated using scripts for automatic data analysis that were implemented in the Data Analysis software. 2 Synthesis of amorfrutin 1 For in vivo testing, a method for the synthesis of multi-gram quantities of pure amorfrutin 1 had to be developed. Existing methods for the synthesis of related bibenzyl derivatives were deemed unsuitable because of the difficulty to purify Wittig salts and the yields were relatively low (40). Therefore, we investigated alternative approaches and hypothesized that the triflate 2 could serve as a precursor for large-scale syntheses of amorfrutin 1 and other amorfrutins (41). Recent work by Kamisuki et al. had shown that the triflate 2 can be obtained in excellent yields from readily available 2,4,6- trihydroxybenzoic acid (1) (42). Using Sonogashira conditions, we reacted triflate 2 with phenylacetylene to furnish alkyne 3 (41), which could easily be converted into methyl ester 5. This very high-yielding sequence allowed for facile preparation of 5; however, introduction of the prenyl side-chain proved unexpectedly difficult. Whereas Friedel-Crafts alkylation of the 4-hydroxy derivative of 5 has been reported to furnish the desired 3-prenylated product in good yields (40), Friedel-Crafts alkylation of the 4- methoxy phenol 5 using a variety of conditions resulted primarily in undesired prenylation in position 5 of the ring. Similarly, Claisen rearrangement of a reverse prenyl ether derivative of 5 produced modest yields of mixtures of the 3- and 5-prenylated products, in agreement with literature precedent (43). Therefore, we investigated whether 5 could be alkylated under basic conditions. Phenoxide ortho- alkylation is a well-established method for prenylating electron-rich phenols (44); however, the presence of the deactivating carbomethoxy group in 5 suggested that conditions must be carefully controlled. We observed that deprotonation of 5 with potassium hydride in toluene and subsequent treatment with prenyl chloride at elevated temperatures produced the desired 7 in good yields. In addition to the desired 7, this procedure also produced smaller amounts of the corresponding O-prenylated compound 6, which could be converted back to the unprenylated starting material 5 and was thus reclaimed (45). Finally, methyl ester 7 was hydrolyzed to produce amorfrutin 1 of greater than 99% purity as determined by analysis via HPLC-MS and NMR spectroscopy. Starting materials were synthesized purchased from Sigma-Aldrich or Acros Organics and used without further purification. Anhydrous solvents were prepared using 4 Å molecular sieves. Silica gel flash column chromatography was performed on Combiflash RF (Teledyne ISCO) chromatography systems using ethyl acetate/hexanes or methanol/dichloromethane mixtures as solvents. 1H and 13C NMR spectra were obtained on 400, 500, and 600 MHz Varian NMR spectrometers using CDCl and acetone-d (Cambridge Isotope Laboratories) as solvents. For amorfrutin 1, protons and 3 6 carbons were assigned based on (1H,1H)-dqfCOSY, (1H,1H)-NOESY, (1H,13C)-HSQC, and (1H,13C)- HMBC spectra. 3 O OH O O O O HO OH TfO O O a b 1 OH 2 OCH 3 OCH 3 3 c O OCH O O 3 1 2 OH O 6 d 5 4 4 OCH OCH 3 3 f e O OCH O OCH 3 3 + O OH 6 7 OCH OCH 3 3 g O OH OH OCH aammoorrfrfurutitnin A 11 3 Multi-gram scale synthesis of amorfrutin 1. a, Kamisuki et al. (42). b, Sonogashira coupling (41), 91% yield. c, H2/Pd(OH)2, 96% yield. d, Lithium methoxide/methanol, 89% yield. e, Potassium hydride/toluene, prenyl chloride, 58% yield, in addition to 28% of the corresponding O-prenylated compound 6. f, CeCl3, NaI, acetonitrile, 85% yield (45). g, Potassium hydroxide/methanol/water, 92% yield. 5,7-Dihydroxy-2,2-dimethyl-4H-1,3-benzodioxin-4-one (8) O OH O O HO OH HO O 1 OH 8 OH Compound 1 (10.3 g, 55 mmol, monohydrate) was dissolved in 100 mL of anhydrous 1,2- dimethoxyethane (yielding a reddish solution) and cooled to 0 °C. To this solution was added Fischer HPLC grade acetone (8.6 g, 150 mmol) and N,N-dimethylaminopyridine (500 mg, 4.1 mmol). After 4 stirring the resulting suspension for 5 min at 0 °C, thionyl chloride (16 ml, 26.3 g, 220 mmol) was added drop-wise via a separatory funnel. After 30 min at 0 °C, the mixture was brought to 20 °C and then stirred for another hour, checking for formation of the less polar product by thin layer chromatography (TLC, 5:1 hexane-ethyl acetate). After completion of the reaction, the mixture was poured slowly into a 4 L Erlenmeyer flask containing a vigorously stirred mixture of ice-cold water (300 mL), 300 ml of ethyl acetate, 100 ml of hexane, and excess of NaHCO (40 g). Care should be taken as the reaction mixture 3 will foam vigorously upon addition to the NaHCO . The resulting salmon-coloured mixture was stirred 3 for 15 to 20 min, the organic phase was separated, and the aqueous phase extracted with two 100-ml portions of diethyl ether. The combined organic solutions were washed with brine, dried over Na SO , 2 4 and evaporated in vacuo. The oily residue was purified via chromatography on silica gel using a hexane- ethyl acetate gradient, yielding 8 (13.9 g, 33.8 mmol, 61%) as powdery white crystals. The product consistently eluted out at approximately 50% ethyl acetate. Spectroscopic data were identical to those reported previously (42). 5-Hydroxy-7-methoxy-2,2-dimethyl-4H-1,3-benzodioxin-4-one (9) O O O O HO O HO O 8 OH 9 OCH 3 To a stirred solution of white crystalline 8 (13.9 mg, 66.1 mmol) in tetrahydrofuran (160 mL) triphenylphosphine (19.1 g, 73 mmol) and methanol (2.33 g, 2.94 ml, 73 mmol) were added. The well- stirred colourless solution was cooled to 0 °C, and diisopropyl azodicarboxylate (14.7 g, 14.4 ml, 73 mmol) was added drop-wise, resulting in a light orange solution. The resulting slightly turbid mixture was stirred at 22 °C for 2 hours. Reaction progress was monitored by TLC (1:1 hexane-ethyl acetate). After 2 hours, the mixture was diluted with diethyl ether (100 mL) and poured into stirred saturated aqueous NaHCO solution (200 ml). The organic phase was separated, the aqueous phase extracted with two 100- 3 ml portions of ether, and the combined ether phases dried over Na SO . Recrystallization from hexane- 2 4 ethyl acetate yielded 9 (12.0 g, 53.5 mmol, 81%) as white, sand-like crystals. Spectroscopic data of 9 were identical to those reported previously (42). 7-Methoxy-2,2-dimethyl-5-[(trifluoromethyl)-sulfonyl]-4H-1,3-benzodioxin-4-one (2) O O O O HO O TfO O 9 OCH 2 OCH 3 3 5 To a stirred solution of 9 (12.0 g, 53.5 mmol) in pyridine (200 mL) at -10 °C under argon triflic anhydride (16.6 g, 9.9 ml, 58.9 mmol) was added drop-wise. The orange mixture was stirred at 0 °C for 4 hours. Reaction progress was monitored by TLC (5:1 hexane-ethyl acetate). The reaction mixture was poured into vigorously stirred ice-cold aqueous NaHCO solution (500 mL), which was extracted with three 200- 3 ml portions of a 1:1 mixture of hexane and diethyl ether. To remove most of the pyridine, the combined extracts were evaporated in vacuo. The residue was re-dissolved in 200 ml of a 1:1-mixture of diethyl ether and hexane, washed with four 100-ml portions of saturated aqueous NaHCO solution, and dried 3 over Na SO . After concentrating in vacuo, the product crystallized as light yellow needles was 2 4 recrystallized from ethyl acetate/hexane mixtures to afford pure 2 as a white powder (17.0 g, 47.7 mmol, 89%). The compound consistently eluted at approximately 30% ethyl acetate. Spectroscopic data of 2 were identical to those reported previously (42). 7-Methoxy-2,2-dimethyl-5-phenethynyl-4H-1,3-benzodioxin-4-one (3) (41) O O O O TfO O O 2 OCH 3 OCH 3 3 To a stirred solution of triflate 2 (16.9 g, 47.4 mmol) in a 5:1 mixture of dimethylformamide and triethylamine (150 ml) at 0 °C bis(triphenylphosphine)palladium(II) chloride (0.78 g, 0.025 eq.) and phenylacetylene (7.15 g, 7.7 ml, 71.1 mmol, 1.5 eq.) were added. The vessel was sealed tightly, and the light orange suspension was heated to 85 °C for 6 hours, during which the reaction mixture turned black. Reaction progress was monitored by TLC (5:1 hexane-ethyl acetate). The mixture was cooled, diluted with diethyl ether (200 mL), washed twice with saturated aqueous NaHCO solution, dried over MgSO 3 4 and concentrated in vacuo. Chromatography over silica gel using a hexane-ethyl acetate solvent gradient yielded 3 as a white solid (13.6 g, 44.1 mmol, 91%). 1H NMR (600 MHz, acetone-d ): δ 7.60– 7.63 (m, 2H, ArH), 7.41– 7.46 (m, 3H, ArH), 6.92 (d, J = 2.5 6 Hz, 1H, 5-H), 6.60 (d, J = 2.5 Hz, 1H, 3-H), 3.93 (s, 3H, OCH ), 1.71 (s, 6H) ppm. 3 13C NMR (126 MHz, acetone-d ): δ 165.6 (C-4), 159.3 (C-2), 158.4 (C=O), 132.5 (2C-H, phenyl), 129.7 6 (C-H, phenyl), 129.3 (2C-H, phenyl), 126.9 (C-6), 123.8 (phenyl), 116.0 (C-5), 107.5 (C-1), 106.2 (C(CH ) ), 102.9 (C-3), 95.7 (alkyne), 88.4 (alkyne), 25.7 (2C, C(CH ) ) ppm. 3 2 3 2 HR-ESI+MS: m/z 331.0970 (calcd for C H NaO [M+Na]+ 331.0946) 19 16 4 6 1H-NMR spectrum (acetone-d, 600 MHz) of 3. 6 (1H,13C)-HMQC spectrum (acetone-d, 600 MHz) of 3. (1H,13C)-HMBC spectrum (acetone-d, 600 MHz) of 3. 6 6 7-Methoxy-2,2-dimethyl-5-phenethyl-4H-1,3-benzodioxin-4-one (4) O O O O O O 3 OCH 4 OCH 3 3 To a stirred solution of 3 (19.2 g, 63.1 mmol) in ethanol (200 ml) under argon in a Schlenck flask connected to H and argon gas lines, Pd(OH) catalyst (4.2 g of 15% Pd(OH) on activated carbon) was 2 2 2 added. The flask was purged with argon and then H . The mixture was stirred vigorously for 1 hour at 2 room temperature whilst maintaining a weak stream of H gas through the reaction vessel. Reaction 2 progress was monitored by TLC in 5:1 hexane-ethyl acetate, and a noticeably less polar spot indicating the formation of 4. The catalyst was filtered off using a vacuum frit. The filtrate and combined washings were concentrated and purified by recrystallization from hexane-ethyl acetate mixtures, yielding 4 as slightly off-white crystals (18.6 g, 60.3 mmol, 96%). 1H NMR (400 MHz, acetone-d ): δ 7.24-7.29 (m, 4H, phenyl), 7.14-7.20 (m, 1H, phenyl), 6.58 (d, J = 6 2.6 Hz, 1H), 6.43 (d, J = 2.6 Hz, 1H), 3.85 (s, 3H, OCH ), 3.30-3.35 (m, 2H, CH -CH -phenyl), 2.85-2.90 3 2 2 (m, 2H, CH -CH -phenyl), 1.67 (s, 6H, C(CH ) ) ppm. 2 2 3 2 7 2-Hydroxy-4-methoxy-6-phenethylbenzoic acid methyl ester (5) O O O OCH 3 O OH 4 OCH 5 OCH 3 3 A solution of n-butyl lithium (1.6N in hexane, 15.0 ml) was added drop-wise to stirred methanol (50 ml) at 0 °C under argon. Subsequently, a solution of 4 (7.8 g, 25 mmol) in tetrahydrofuran (THF, 100 ml) was added in one portion, and the resulting mixture was heated to 60 °C. After 1.5 hours at 60 °C, the mixture was cooled to 20 °C and concentrated in vacuo to a volume of 30 ml. This material was diluted with diethyl ether (200 ml) and 1 N aqueous hydrochloric acid was added with stirring until the aqueous layer was slightly acidic (pH 3). The organic layer was separated, the aqueous layer extracted with two 100-ml portions of ether, and the combined organic solutions were dried over Na SO and concentrated in vacuo. 2 4 The residue was chromatographed over silica gel using a hexane/ethyl acetate solvent gradient, yielding 5 as a white crystalline solid (6.42 g, 22.4 mmol, 89%). 1H NMR (600 MHz, acetone-d ): δ 11.61 (s, 1H, OH), 7.24-7.30 (m, 4H, phenyl), 7.17-7.20 (m, 1H, 6 phenyl), 6.41 (d, J = 2.6 Hz, 1H), 6.36 (d, J = 2.6 Hz, 1H), 4.00 (s, 3H, CO CH ), 3.81 (s, 3H, OCH ), 2 3 3 3.16-3.19 (m, 2H, CH -CH -phenyl), 2.80-2.82 (m, 2H, CH -CH -phenyl) ppm. 2 2 2 2 2-Hydroxy-4-methoxy-3-(3-methylbut-2-en-1-yl)-6-phenethylbenzoic acid methyl ester (7) O OCH O OCH O OCH 3 3 3 OH OH O + 5 OCH 7 OCH 6 OCH 3 3 3 To a stirred solution of methyl ester 5 (3.30 g, 11.6 mmol) in toluene (200 ml) under argon potassium hydride (0.481 g, 12.0 mmol) was added. The solution was stirred for 20 min at 20 °C and then heated to 70 °C. After 20 min at 70 °C the clear reaction mixture was cooled to 20 °C and prenyl chloride (1.40 g, 1.51 mL, 13.4 mmol) was added. The mixture was then heated to 75 °C. After 2 hours, the reaction was cooled to 20 °C, diluted with diethyl ether (200 ml), washed with three 100-ml portions of saturated aqueous NaHCO , and concentrated in vacuo. The residue was chromatographed over silica gel using a 3 hexane/ethyl acetate solvent gradient, yielding 2.4 g (6.8 mmol, 58%) of 7 as a white crystalline solid, in addition to 1.15 g (3.2 mmol, 28%) of the O-prenylated derivative 6. 2-Hydroxy-4-methoxy-3-(3-methylbut-2-en-1-yl)-6-phenethylbenzoic acid methyl ester (7): 1H NMR (600 MHz, acetone-d ): δ 11.79 (s, 1H, OH), 7.25-7.31 (m, 4H, phenyl), 7.17-7.21 (m, 1H, phenyl), 6.52 6 (s, 1H, 5-H), 5.18 (triplet of septets, J = 7.2, 1.4 Hz, 1H, CH=C(CH ) ), 4.02 (s, 3H, CO CH ), 3.85 (s, 3 2 2 3 3H, OCH ), 3.30 (d, J = 7.2 Hz, CH -CH=C(CH ) ), 3.19-3.23 (m, 2H, CH -CH -phenyl), 2.85-2.89 (m, 3 2 3 2 2 2 2H, CH -CH -phenyl) 1.75 (br.s, 3H, CH ), 1.62 (br.s, 3H, CH ) ppm. 2 2 3 3 8 2-(3-methylbut-2-enyloxy)-4-methoxy-6-phenethylbenzoic acid methyl ester (6): 1H NMR (600 MHz, acetone-d ): δ 7.26-7.30 (m, 2H, phenyl), 7.20-7.23 (m, 2H, phenyl), 7.16-7.20 (m, 1H, phenyl), 6 6.48 (d, J = 2.3 Hz, 1H), 6.44 (d, J = 2.3 Hz, 1H), 5.18 (triplet of septets, J = 6.4, 1.4 Hz, 1H, CH=C(CH ) ), 4.57 (d, J = 6.4 Hz, CH -CH=C(CH ) ), 3.82 (s, 3H, OCH ), 3.77 (s, 3H, OCH ), 2.85- 3 2 2 3 2 3 3 2.89 (m, 2H, CH -CH -phenyl), 2.78-2.83 (m, 2H, CH -CH -phenyl) 1.76 (br.s, 3H, CH ), 1.74 (br.s, 3H, 2 2 2 2 3 CH ) ppm. 3 2-Hydroxy-4-methoxy-3-(3-methylbut-2-en-1-yl)-6-phenethylbenzoic acid (amorfrutin A1) O OCH O OH 3 OH OH Amorfrutin A1 7 OCH OCH 3 3 A solution of 7 (4.9 g, 13.8 mmol) in methanol (100 ml) was added to a solution of potassium hydroxide (43 g) in a mixture of methanol (350 ml) and water (50 ml). The mixture was heated under reflux and progress was monitored by TLC. After 7 hours, the mixture was cooled to 20 °C and concentrated in vacuo to a volume of 150 ml. The concentrated mixture was diluted with 200 ml of water, cooled at 0 °C and acidified (to pH 3) with 2 N aqueous hydrochloric acid. The acidic suspension was extracted with three 100-ml portions of ether. The combined ether extracts were washed with brine, dried over MgSO , 4 and evaporated to dryness. The residue was chromatographed over silica gel using a hexane/ethyl acetate solvent gradient, yielding 4.5 g of crude product, which was recrystallized from hexane/ethyl acetate, furnishing the pure amorfrutin A1 (4.3 g, 12.6 mmol, 92%) as a white crystalline powder. 1H NMR (600 MHz, acetone-d ): δ 7.24-7.29 (m, 4H, phenyl), 7.16-7.19 (m, 1H, phenyl), 6.49 (s, 1H, 6 5-H), 5.15 (triplet of septets, J = 7.2, 1.4 Hz, 1H, CH=C(CH ) ), 3.84 (s, 3H, OCH ), 3.30 (d, J = 7.2 Hz, 3 2 3 2H, CH -CH=C(CH ) ), 3.27-3.30 (m, 2H, CH -CH -phenyl), 2.90-2.93 (m, 2H, CH -CH -phenyl), 1.75 2 3 2 2 2 2 2 (br.s, 3H, CH ), 1.62 (br.s, 3H, CH ) ppm. 3 3 13C NMR (126 MHz, acetone-d ): δ 174.0 (1-CO H), 145.7 (C-6), 162.9 (C-2), 162.0 (C-4), 145.7 (C-5), 6 2 131.1 (CH -CH=C(CH ) ), 129.1 (2C-H, phenyl), 128.8 (2C-H, phenyl), 126.4 (C-H, phenyl), 123.3 2 3 2 (CH -CH=C(CH ) ), 115.1 (C-3), 106.7 (C-5), 105.2 (C-1), 55.8 (OCH ), 39.8 (CH -CH -phenyl), 38.9 2 3 2 3 2 2 (CH -CH -phenyl), 25.7 (CH -CH=C(CH ) ), 22.4 (CH -CH=C(CH ) ), 17.6 (CH -CH=C(CH ) ) ppm. 2 2 2 3 2 2 3 2 2 3 2 HR-ESI+MS: m/z 363.1602 (calcd for C H NaO [M+Na]+ 363.1572). 21 24 4 9 7.4 7.0 6.6 6.2 5.8 5.4 5.0 4.6 4.2 3.8 3.4 3.0 2.6 2.2 1.8 1.4 f1 (ppm) 1H-NMR spectrum (acetone-d, 600 MHz) of A1. 6 (1H,13C)-HSQC spectrum (acetone-d, 600 MHz) of A1. (1H,13C)-HMBC spectrum (acetone-d, 600 MHz) of A1. 6 6 Nuclear receptor binding assays Identified PPARγ ligands were further validated and characterized by a time-resolved fluorescence resonance energy transfer (TR-FRET)-based competitive binding assay according to the manufacturer’s protocol (Lanthascreen, Invitrogen, Darmstadt, Germany). Increasing the concentration of potential ligands results in a displacement of the labelled PPARγ-ligands and hence a decrease of the TR-FRET signal. Recruitment of transcriptional cofactors was determined by the use of Lanthascreen coactivator assays (Invitrogen) according to the user’s manual. Fluorescence was measured with the POLARstar Omega (BMG LABTECH, Offenburg, Germany). Data were fitted using GraphPad Prism 5.0 according equation: Y=Bottom + (Top-Bottom)/(1+10^((LogIC50- X)*HillSlope)) with variable Hill slope. IC50 values were converted to general Ki values according to Cheng and Prusoff (46). Activation of CAR and PXR was determined in cofactor recruitment assays by Cerep, Inc. (Celle l'Evescault, France). Reporter-gene assay Cellular activation of PPARγ was assessed in a reporter gene assay according to the manufacturer’s protocol (GeneBLAzer PPARγ DA Assay, Invitrogen). In brief, HEK 293H cells were stably expressing a GAL4-PPARγ-LBD fusion protein and an UAS-beta-lactamase reporter gene. Cells were incubated with indicated concentrations of compounds. Fluorescence was measured with POLARstar Omega. Obtained 10