Iso-seco-tanapartholides: Isolation, Synthesis and Biological Evaluation

The isolation, identification and total synthesis of two plant-derived inhibitors of the NF-κB signaling pathway from the iso-seco-tanapartholide family of natural products is described. A key step in the efficient reaction sequence is a late-stage oxidative cleavage reaction that was carried out in the absence of protecting groups to give the natural products directly. A detailed comparison of the synthetic material with samples of the natural products proved informative. Biological studies on synthetic material confirmed that these compounds act late in the NF-κB signaling pathway. (© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim, Germany, 2009)

. Activity of the fractions isolated from the initial crude purification in the NF-κB dependent-luciferase reporter gene assay. Fraction 8 was selected for further purification.
A dose response curve was generated for fraction 8 prior to further purification ( Figure S2). The clear dependence of the biological activity on the concentration of the sample encouraged us to carry out further purification of fraction 8. different concentrations of the extract were incubated with HeLa 57A cells for 2 hours, prior to stimulation with 50 ng/ml PMA for 4 hours. The cells were lysed and luciferase activity was measured in a plate reader luminometer. Results were compared relative to the control sample, where similar amounts of DMSO were used and then represented as the mean ± SD of 2 similar experiments.
At this stage, whilst we had sufficient material in hand, an approximate calculation for the IC 50 of fraction 8 was also carried out and gave a value of 110 µM (using a MW = 278 Da (for isoseco-tanapartholide) and a concentration of 3 x 10 -2 µg/µl for the concentration of fraction 8 required to reduce the maximal observed biological readout by 50%). Due to the very limited amount of purified natural product that was eventually isolated this IC 50 calculation was not repeated on the final material. However, a pure sample of 1 would be expected to have an IC 50 value that was significantly less than that observed for fraction 8 (as was found to be the case for synthetic 1). At this stage it was also confirmed using a standard in vitro luciferase assay that fraction 8 did not inhibit the activity of the reporter protein directly (data not shown).
Purification stage 2: Fraction 8 from the fractionated crude extract was repurified using Reverse-Phase High Performance Liquid Chromatography (RP-HPLC). A preparative C18 silica column was used with acetonitrile (ACN), water (H 2 O) and 0.1% (v/v) trifluoroacetic acid (TFA) as the mobile phase. 1g of material was purified to produce 53 new fractions of 10ml volume, using a linear gradient from 5% to 75% ACN in H 2 O with 0.1% TFA at a flow rate of 10 ml/min. Detection wavelengths of 220 nm, 250 nm and 270 nm were used. Fractions from this run were screened using the HeLa 57A assay (see Figure S3) and fractions 34-41 were selected, concentrated in vacuo and prioritized for further purification. Purification stage 3: In the final round of purification, the combined fractions 34-41 from stage 2 were repurified using a preparative C18 silica column as follows: a first gradient step from 5 % to 20 % ACN (2.0 CV) / a segment step at 20% ACN (1.5 CV) / a gradient step from 20% to 22.5% ACN (1.0 CV) / a segment step at 22.5% ACN (1.5 CV) / a gradient segment from 22.5% to 25% ACN (1.0 CV) and a segment step at 25% ACN (1.5 CV) was used. The flow rate was 1.5 ml/min and 50 new fractions of 4ml volume were collected. Fractions from stage 3 were screened as described previously with fractions 25-36 ( Figure S4) all showing inhibitory activity in the HeLa 57A assay. Fraction 26 was selected for structure identification studies as it was judged to contain significant amounts of material and to be the cleanest. Fractions Relative Activity Figure S4. Activity of the fractions isolated from purification stage 3.
The TOCSY spectrum ( Figure S7) clearly showed that protons H-13, H-6, H-7, H-9 and H8 were part of one extended spin-system whereas the COSY spectrum ( Figure S6) indicated that proton H-7 coupled to all the other protons in the spin-system except H9. According to the HSQC spectrum ( Figure S8), both protons H-13 were bonded to the same carbon at 122.1 ppm which suggested that they were geminal protons of a C=C double bond. In the HMBC spectrum ( Figure S9), long-range correlations were observed from protons H-13 and H-6 to C-12 (δ 171.4) and from H-13 and H-8 to C-11 (δ 140.9). All the above mentioned correlations indicated the presence of a sesquiterpene lactone ring with an aliphatic chain bonded to the carbon C7 adjacent to exocyclic double bond. The HMBC spectrum ( Figure S9) also showed two ketone carbonyls, C-10 and C-1 (δ 209.0 and 204.9, respectively). Since the carbonyl C-10 correlates in the HMBC spectrum with protons of CH 2 -8, CH 2 -9 and CH 3 -14 it was concluded that the aliphatic chain ends with an acetyl group. On the other hand, the carbonyl C-1 correlated with both protons of CH 2 -2 and proton H-6. According to the COSY spectrum ( Figure S6), protons of CH 2 -2 and proton H-3 form a separate ABX spin-system. Moreover, the HMBC spectrum ( Figure S9) showed long-range correlations from H3 to C-4 and C-5 (δ 175.6 and 138.0, respectively). Both carbons C-4 and C-5 also correlate with protons of CH 3 -15, CH 2 -2 and H6. The observation of these correlations enabled us to propose that C-6 of the sesquiterpene lactone ring is substituted by a 3-hydroxy-2-methyl-5-oxocyclopent-1-enyl ring. The overall structure ( Figure S10) with a molecular formula of C 15 H 18 O 5 is also in good accordance with the mass spectroscopic analysis ( Figure S11) where peaks were detected at m/z (M + H + ) = 279, (M + Na + ) = 301 and (M + Na + + CH 3 CN) = 345.
The relatively small coupling observed between H-6 and H-7 (J = 5.7 Hz) did not allow us to draw an unambiguous conclusion concerning the stereochemistry of the lactone ring. Therefore a selective 1D gs-NOESY experiment was carried out ( Figure S12). nOe enhancements observed for protons H-9 and H8 after selective inversion of H-6 clearly confirmed the trans-configuration which has been reported for all natural products isolated plants of the genus Artemisia to date. However, closer inspection of H-2a, H-3 and H-13 resonances in the 1 H NMR spectrum revealed that RFMF_26 consisted of two diastereomers ( Figure S13). This conclusion was also apparent from the H-3, C-3 cross-peak in the HSQC spectrum ( Figure S14) and the H-3,H-2a and the H-3,H-2b crosspeaks in the COSY spectrum ( Figure S15). Since the biggest differences in chemical shifts are observed for H-3 and H2 resonances it seemed very likely that RFMF_26 fraction contained both C-3 epimers. Attempts to assign the stereochemistry at C3 for the major and minor epimers proved unsuccessful due to the relatively remote nature of this stereocentre and the relative flexibility of this structure.    Coupling constants shown in parentheses in column 3. For the numbering system used see Figure S10.

Table S2
Assignment of 13 C NMR spectrum of the active compound in RFMF_26. For C3 signal value in parentheses corresponding to the signal assigned to the minor C3-epimer

S1
To a solution of 8 (668 mg, 2.71 mmol) in DCM (15 ml) at room temperature was added m-CPBA (493 mg, 2.85 mmol) and reaction followed by TLC. Reaction was complete after 6 hr. Saturated aqueous NaHCO 3 solution (20 ml) was added and the mixture extracted with DCM (2 x 70 ml). Combined organic phases were washed with water (30 ml), dried and concentrated to give a crude mixture which was purified by flash column chromatography (SiO 4 , 6:4 hexanes: ethyl acetate) giving the β-epoxide S1 as a white solid (711 mg, 85 % yield). X-ray quality crystals of S1 were obtained by slow evaporation of hexanes/DCM solution of pure S1 (data submitted to CCDC  3, 177.3, 159.5, 141.5, 82.2, 68.5, 66.4, 50.1, 41.4, 40.8, 33.5, 25.7, 24.5, 12.6, 9.26 Evidence to support the assigned stereochemistry of compounds 21 and 23, prepared by selenylation of 19 and 20 respectively came from the use of nOe experiments. The observed enhancements were consistent with the assigned stereochemistry ( Figure S12 and S13) Figure S16 Experiment used to support the stereochemical outcome in the synthesis of 21.

S15
Figure S17 Experiment used to support the stereochemical outcome in the synthesis of 23.

S16
Comparison of data for synthetic 1 and 2 with previous literature reports.
Upon completion of the synthesis of compounds 1and 2 we compared our analytical data with that reported in literature. References S3-S11 list the previous reports of the isolation of isoseco-tanapartholide. Huneck and co-workers were the first to name and assign structure 1 to iso-seco-tanapartholide isolated from a plant of the genus Artemisia. S3 They assigned the stereochemistry of the hydroxyl group at C3 as being β (S) based on "observed couplings" in the 1 H NMR spectrum (400MHz, CDCl 3 ) although details of how this was achieved are missing in their report. S3 Comparison of the 1 H NMR spectra we obtained for our synthetic samples of 1 and 2 with the signals reported by Huneck S3 (Table S3) showed how similar the spectra of the two isomers are to each other and highlighted the difficulties of doing this comparison with Huneck's report. S3 Despite the samples being run under analogous conditions, it is not possible to conclude based on the 1 H NMR spectrum alone whether Huneck had isolated 1 or its isomer 2, although it is clear that the basic structure of iso-seco-tanapartholide is correct. To date we have been unable to obtain an authentic sample of the material reported by Huneck. An analogous conclusion is also reached when a comparison of the 13 C NMR data reported for iso-seco-tanapartholide by Marco S4 (also isolated from a plant of the genus Artemisia), is carried out with our data for synthetic 1 and 2 (Table S4) Interestingly, the compound isolated and reported by Todorova S5 in 1985 (structure S1 below). has very similar spectroscopic analysis to our synthetic 1 and 2 and to the compounds reported by Huneck S3 and Marco S4 . It is therefore likely that the compound isolated by Todorova in 1985 is either 1 or 2 and not the structure reported in that paper. It has not been possible to confirm this due to the absence of authentic material. No further reports concerning a compound with the structure reported by Todorova S12  Professor Ryu kindly provided us with a 1 H NMR spectrum for "compound 2" reported in reference S10 ( Figure S18). This sample referred to as "compound 2" was isolated from Artemisia iwayomogi. Whilst it is interesting to note that this sample is present as a single epimer, in the absence of isolated material and optical rotation data it is not possible to ascertain whether this material corresponds to iso-seco-tanapartholide 1 or epi-iso-secotanapartholide 2. Figure S18. 1 H NMR of the named "compound 2" reported by Ryu et al. (CDCl 3 , 600MHz) S10

Comparison of data for synthetic 1 and 2 with our sample isolated from Tanacetum parthenium (Fraction 26 from purification of extract #2335)
In order to confirm our initial conclusion that the material we had isolated was a mixture of diastereoisomers differing only in the configuration at C3, direct comparison of 1 H NMR spectra for our isolated material with our synthetic 1 and 2 was made ( Figure S19). The 1 H NMR spectra for synthetic 1 and 2 were rerun in CD 3 CN to enable the comparison and as can be seen from Figure S19 all the relevant signals are very closely matched, with the only easily detectable difference in chemical shifts between them being those observed in the 1 H NMR for the C3 proton. Superimposing all three 1 H NMR spectra (isolated material, synthetic 1 and synthetic 2) exhibited an excellent match and revealed 2 to be the major component in our isolated sample. Unfortunately, due to the very limited amount of material isolated in a pure form, we were unable to obtain an optical rotation.

Figure S19
Overlay of 1 H NMR spectra (500MHz, CD 3 CN) of a) synthetic compound 2 b) synthetic compound 1 c) our isolated material from T.parthenium.

S20
Comparison of data for synthetic 1 and 2 with a sample isolated from Achillea. S11 Professor Todorova kindly provided us with a sample of "iso-seco-tanapartholide" from a plant of the genus Achillea. We carried out a detailed comparison of this material with our synthetic 1 and 2. This comparison clearly showed ( Figure S20) that this sample was a mixture of two epimers with the same relative stereochemistry as 1 and 2. It also showed that the major isomer present had the same relative stereochemistry as 2, whilst the minor isomer had the same relative stereochemistry as 1. ([α] 20 D = -6.5 (c = 0.002 in CHCl 3 ) suggested that the major isomer present in Todorova's sample had the same relative and absolute configuration as our synthetic 2 and therefore that 2 is a natural product, which we have named epi-iso-seco-tanapartholide.
We thank Professor Saliba for forwarding us the 1 H NMR spectra (CDCl 3 , 300MHz) of "compounds 3a and 3b" from this paper. Comparison of these NMR spectra with our authentic samples confirmed that compound 3a has the same relative configuration as epi-iso-secotanapartholide 2 and that compound 3b is a mixture of the two epimers, as stated, with epi-isoseco-tanapartholide 2 being the major one present. (-)-α-santonin (8.01 g, 32.5 mmol) was dissolved in glacial acetic acid (80 mL) and irradiated using a 125W mercury lamp in a photochemical reactor vessel for 14 hr. The mixture was concentrated and the residual thick brown oil was dissolved in hot methanol (7 mL) and left in a freezer overnight to crystallize. The white crystalline solid product (2.53 g, 25% yield) was collected by filtration and washed with cold methanol (7 mL).  9, 177.1, 170.3, 160.9, 143.2, 85.5, 81.2, 48.2, 47.3, 41.3, 37.9, 36.8, 25.3, 22.3, 20.0, 12.4, 9.47 [α]  To conc. sulfuric acid (20 mL) at 0 o C was added O-Acetylisophotosantonic lactone 10 (1.34 g, 4.37 mmol) portionwise over 10 min then stirred for 10 min at 0 o C before the ice-bath was removed and the mixture allowed to warm up to room temperature. Stirring was continued for 50 min. Then the resulting brown solution was poured into ice/water mixture and left to warm up to room temperature before extracting with dichloromethane (3 x 60ml). The organic extracts were combined and washed with 5% aqueous sodium hydroxide solution (20 mL), then water (20 mL) and dried (Na 2 SO 4 ) and finally concentrated giving the desired product (1.07 g) as a white solid in quantitative yield. This material was used in subsequent step without further purification. Material should be stored in the dark as it decomposes easily in ordinary light.  204.4, 177.6, 160.8, 139.2, 133.1, 129.5, 80.3, 47.5, 42.3, 40.3, 32.7, 26.9, 24.4, 12.9, 9.81 [α]  To a solution of alkene 8 (707 mg, 2.87 mmol) and 4-methylmorpholine N-oxide (NMO, 681 mg, 5.81 mmol) in 9:1 THF/water (10 mL) at room temperature was added a 2.5 wt % solution of osmium tetroxide in tert-butanol (2 mL, 0.160 mmol). The mixture was stirred for 6 hrs and quenched by addition of saturated aqueous sodium sulfite solution (10 ml) and the mixture stirred for 2 hrs before extraction with ethyl acetate (3 x 50 mL) and the combined organic phases were washed with 10% aqueous sodium sulfite solution (10 mL) followed by water (10 mL). After drying, the solvent was removed in vaccuo and the residue was purified by flash column chromatography (SiO2, 4:6 hexanes: ethyl acetate) to give the desired product as a white solid (710 mg, 87%) and 3:1 mixture of inseparable diastereomers. On one occasion, chromatography led to an analytically pure sample of major isomer 9 being isolated.

S32
To a solution of fully silyl-protected triol 24 (752 mg, 1.40 mmol) in dry THF (10 mL) at 0 o C was added TBAF (10.0 mL, 1.0 M in THF, 10.0 mmol). The resulting solution was stirred at 0 o C for 1.5 hrs then allowed to warm up to room temperature and stirred for a further 1.5 hrs at which point saturated aqueous ammonium chloride solution (20 mL) was added and mixture extracted with ethyl acetate (2 x 100 mL). Combined organic layers were washed with brine (40 mL), dried (NaSO 4 ) and concentrated under vacuum. The residue was purified by flash column chromatography (SiO 2 , ethyl acetate) to provide the desired product (360 mg) as a white solid in 92% yield.  1, 148.0, 141.1, 134.4, 117.4, 90.7, 81.7, 76.8, 76.4, 47.0, 43.0, 38.8, 23.9, 22.1, 13.3 [α]  To a solution of triol 25 (167 mg, 0.596 mmol) in 20 mL of dry DCM/Acetone (3:1) at 0 o C was added lead tetraacetate (326 mg, 0.735 mmol) and stirred for 25 min at 0 o C. Then a small amount of silica was added into the flask and solvent was removed under vacuum. The residue was dry loaded onto column and purified by flash column chromatography (SiO 2 , ethyl acetate) to provide the desired product (163 mg) as a thick oil in 98% yield.  207.9, 203.4, 173.3, 170.2, 138.4, 137.7, 123.0, 76.1, 71.7, 44.4, 42.9, 39.6, 30.1, 27.4, 14.2 [α]  Biology experimental procedures a) IC 50 Determination in Hela 57A assay (TNFα-stimulated production of a luciferase reporter gene). Hela 57As were seeded at 3000 cells per well in each well of a 96 well plate (Greiner) in D-Mem containing 10% foetal calf serum and standard antibiotics. The cells were left to grow at 37 0 C for 3 days after which time compound was added as a solution in DMSO. For each dose response curve generated, cells were treated with compound at a final concentration of 100, 50, 25, 12.5, 6.25, 3.12, 1.56 and 0 µM for 2 hrs. For each compound six dose response curves were generated per analysis. After the two hours preincubation with compound was complete, a solution of TNFα in D-Mem (10ng/ml) was added to each well and the plate incubated at 37 0 C for a further 5 hrs. The media was then carefully removed from each well and the cells washed with PBS (2 x 50µl). The final wash was removed and luciferase lysis buffer (100µl/well) was then added to each well and the plate incubated at room temperature for 20 minutes. A solution of 100 µl luciferase assay buffer (containing luciferin) was then added and the plate read on plate luminometer giving a readout in Relative Light Units(RLU) for each well. IC 50 values were determined using SigmaPlot  software. Data for each compound was plotted and a curve of best fit was applied to each data set (R 2 = at least 0.98). The equation for each best fit curve was used to calculate the IC 50 (the reported values are the average of at least 3 calculated values). b) Western blot analysis: Experiments were carried out according to the protocol previously reported by Arenzana-Seisdedos, Hay et al. S15 c) Band shift assays: Experiments were carried out according to the protocol previously reported by Matthews, Hay et al. S16 As shown in Figures S21 and S22, a clear dose dependent response was observed when these experiments were carried out in the presence of either 1 or 2. As was the case for the IC 50 determinations, 1 and 2 are essentially equipotent in this assay. As discussed, further evidence to support the mode of action of our natural product extract was carried out using immunofluorescence based experiments that probed the location of the p65-subunit of NF-κB. Again, due to limitations on material, this experiment was carried out using fraction 8 of the extract #2335. The experiment was carried out with nasal polyps fibroblast cells as follows: Eosinophils (2.5 x 10 6 /ml) were cultured at 37ºC in Iscove´s DMEM containing 5% FCS for 2 hours. The cells were centrifuged and dried on air for 10 minutes. The cells were then fixed with 4% (w/v) p-formaldehyde/PBS for 10 minutes and washed 3 times with PBS. The cells were permeabilized and non-specific binding was blocked in buffer containing 0.2% (w/v) Triton X-100 in DAKO Protein block Serum Free buffer at room temperature for 30 minutes. Rabbit polyclonal p65 antibody diluted in DAKO Antibody Diluent with 0.2% (v/v) Triton X-100 was added to the cells for 1 hour. The cells were washed 3 times with the same buffer and incubated with anti-mouse IgG FITC antibody, diluted in DAKO Antibody Diluent with 0.2% (v/v) Triton X-100, for 1 hour. Finally, cells were washed 3 times in the same buffer, glass coverslip were applied and cells examined by fluorescent microscope.  Figure S23. Immunofluorescence using nasal polyps fibroblasts cells, stimulated with TNF-α. A, cells were not pre-treated and not stimulated. B, cells were not pre-treated but stimulated with 50ng/ml TNF-α. C, cells were pre-treated with 20,0x10 -2 mg/ml of 2335 F8 extract and stimulated with 50ng/ml TNF-α after 2 hours incubation. D, cells were pre-treated with 20,0x10 -2 mg/ml of 2335 F8 extract, but not stimulated. Coverslips were fixed and immunoblotted with the rabbit polyclonal p65 antibody.