Chemical Transformations of Eremanthine . Synthesis of Micheliolide and 1 ( R ) , 10 ( R )-Dihydromicheliolide

Eremantina (1), uma substância natural abundante, foi transformada em quatro etapas no diol 5. Após hidrogenólise de 5 (55 psi de H 2 , Pd/C, 30 min) obteve-se 7. Hidrogenação de 5 usando-se uma baixa pressão de hidrogênio (5 psi) e um menor tempo de reação (15 min) forneceu uma mistura de 6 e 7 (3:1). Os compostos 6 e 7 foram, a seguir, transformados nas respectivas α-metileno-γlactonas miqueliolido (9) e 1(R),10(R)-diidromiqueliolido (8), após eliminação de metanol.


Introduction
Sesquiterpene lactone is an important class of naturally occurring substances generally found in Compositae family. 1 Many of these compounds are endowed with an impressively rich spectrum of biological activity 2 as antileishmanial, 3 antifungal, 4 cytotoxic 5 and herbicide, 6 among others. 7This diverse bioactivity of sesquiterpene lactones along with their structural complexity makes these compounds important targets for synthetic purposes. 8urthermore, with rare exceptions, their availability from natural sources is very limited.Therefore, it is important to sinthesize these compounds from easily available starting materials.
Eremanthine (1), a sesquiterpene lactone isolated from Brazilian compositae Eremanthus elaeagnus 9 and Vanillosmopsis erythropappa 10 is an inhibitor against infections caused by cercariae of Schistosoma mansoni.The abundance of this substance turned possible the obtention of others potentially active derivatives, as well as the synthesis of less abundant natural lactones, through chemical modifications of 1. 11 Continuing the research programme of chemical transformations of eremanthine (1), this compound was converted to the diol 5, 12 a potential precursor for the synthesis of micheliolide (9), an anticancer sesquiterpene lactone isolated from Michelia compressa 13 and Michelia champaca. 14n this paper we report the obtention of diol 5 and its transformation into micheliolide (9) and 1(R),10(R)dihydromicheliolide (8).

Results and Discussion
Initially, eremanthine (1) was transformed into diol 5 as outlined in Scheme 1 (conditions i -iv). 12he α-methylene-γ-lactone of eremanthine (1) was protected as methanol adduct 2. The choice was due to the stability of this group and relative facility to be removed.15 Thus, the reaction of eremanthine (1) with methanol catalysed by sodium methoxide furnished adduct 2 in nearly quantitative yield.Epoxidation of compound 2 with excess of peracetic acid solution in CH 2 Cl 2 furnished diepoxide 3 16 and crude product was submitted to ring opening through treatment with glacial acetic acid and equimolar amount of potassium iodide, in reflux of acetone.The use of equimolar amount of KI provided the chemoselective opening of the more reactive 4,15-αepoxide through the nucleophilic attack of iodide at C 15 .
On the other hand, the protonation of 9,10-α-epoxide contributed for the generation of a cationic intermediary at C 10 where elimination of H + at C 1 furnished the compound 4 17 in 66% yield after purification by flash chromatography.Hydrogenolysis of 4 with hydrogen catalysed by palladium on charcoal and mixture of sodium acetate and ethanol gave diol 5. 12 At this point we decided to investigate this reaction more carefully in order to carry out the hydrogenolysis of C 15 -I and C 9 -OH in one step.
The classic literature of Organic Chemistry reports that hydrogenation of allylic alcohol with hydrogen and catalyst, for example palladium on charcoal, proceeds initially with hydrogenolysis of C-OH followed by reduction of double bond C-C. 18In our case, the hydrogenation of tetrasubstituted double bond C 1 -C 10 at allylic alcohol 4 seemed to be an unfavourable reaction since tetrasubstituted olefins are more resistent and require higher temperatures and pressures. 19This resistence is usually a function of increasing substitution and is presumably caused by steric factors.
As we had observed before, the use of 45 psi of hydrogen pressure didn't cause any hydrogenolysis of C 9 -OH at 4. We planned to use a higher hydrogen pressure during several hours in order to convert 4 to 6 and an experiment was performed in which we used the highest recommended pressure for the Parr hydrogenation apparatus (condition v -Scheme 1).The reaction course was examined by TLC in regular times of one hour and after 48 h, two spots (Rf 0.29 and 0.6, EtOAc as eluent) were observed.The sloweluting spot (R f 0.29) corresponded to diol 5 by comparison with an authentic sample of this compound.On the other hand, the fast-eluting spot (R f 0.6) seemed to correspond to the target molecule 6 due to its lower polarity.Because of the high polarity of compound 5, we decided to extract the reaction products using two solvents of different polarities (ethyl ether and ethyl acetate) in order to separate the two fractions of R f 0.29 and 0.6 by extraction.The crude product was partitioned first with ethyl ether and then exhaustively with ethyl acetate.After the usual aqueous work up and evaporation of the solvents, it was obtained two residues which were submitted to TLC.The spot of the ethereal residue corresponded to the product with R f 0.6 and the residue of ethyl acetate to diol 5 (R f 0.29).The ratio of diol 5 to product of R f 0.6 was 5:1 and ) was observed at δ 1.64.With these spectral data we concluded that spot of R f 0.6 corresponded to two substances, where one of them was the compound 6.A new TLC analysis of the ethereal residue (R f 0.6) was performed using 35% EtOAc/hexane as eluent (elution repeated three times).After staining, it was observed two spots of very similar R f .These results suggested that substance 6 had been formed and then transformed in part to another product in the reaction medium, maybe compound 7.
However the H-14 doublet of 7 was masked in the 1 H NMR spectrum of this mixture.To confirm the in situ conversion of 6 to 7, we decided to perform the hydrogenolysis reaction using as starting material diol 5 (condition i -Scheme 2) in order to get 7 as a single product.
After reaction time the TLC revealed that diol 5 had been transformed to a single product of R f 0.6 (eluent: 3 x 35% EtOAc/hexane).The 1 H NMR spectrum of product indicated hydrogenolysis of the C 9 -OH bond besides reduction of double bond C 1 -C 10 at 5. A doublet at δ 0.95 (3H, J 7.2 Hz) was attributed to H-14.One singlet at δ 3.32 (3H, H-16) and one triplet at δ 3.99 (1H, J 10.3 Hz, H-6) confirmed that 7 was the single product of this reaction.In the 13 C NMR spectrum was only observed one signal of sp 2 carbon (δ 175.9;C=O) confirming thus the hydrogenation of double bond C 1 -C 10 .The stereochemistry of the stereogenic carbons C 1 and C 10 was determined by NOE experiment. 20The trans junction between the five and seven-membered-rings of the hidroazulene system was confirmed by axial-axial coupling constants between H-5 and H-1 (J 11.2 Hz).The fast hydrogenation reaction of tetrasubstituted double bond C 1 -C 10 was an unexpected result since compound 2, which has a trisubstituted double bond C 9 -C 10 , hydrogenated slowly [H 2 (60 psi); 10% Pd-C (0.1 equiv.);EtOH (r.t.-4h)]. 21ur attention was focused, at this stage, to examine the means for effecting only hydrogenolysis of the C 9 -OH bond in 5, in order to get compound 6, the immediate precursor of micheliolide (9).After many experiments, we found that the best condition to carry out this reaction was the use of a low hydrogen pressure (5 psi) and a short reaction time (maximum of 15 minutes) (condition ii -Scheme 2).With this condition, the major product obtained was the compound 6, as a mixture with 7 (94% yield, 3:1 ratio by 1 H NMR). The separation of 6 and 7 by column chromatography proved to be troublesome (practically identical R fs ) and to our delight, compound 6 could be crystallized from hexane.
Finally, restoration of α-methylene-γ-lactone function of compounds 6 and 7 was achieved using basic conditions. 15The reactions were quenched with aqueous HCl in order to consume the excess of NaOH and lactonize the hydroxy acids formed in this stage.Micheliolide (9)  and 1(R),10(R)-dihydromicheliolide (8) were obtained in 80% and 85% yield, respectively.Micheliolide (9) has  ( 1 : 3 ) already been synthesized before by BF 3 -mediated rearrangement of parthenolide. 22n conclusion, we have developed an efficient and straightforward synthesis (six steps) of micheliolide (9) (31% overall yield) and a new compound 1(R),10(R)dihydromicheliolide (8) (45% overall yield) from the abundant natural product eremanthine (1), using inexpensive and easily available reagents.We expect that the synthesis outlined herein to be useful for the chemistry of sesquiterpene α-methylene-γ-lactones and related systems.

Experimental
Infrared spectra were recorded on a Perkin-Elmer 1420 spectrophotometer using either thin films on NaCl plates (film) or KBr discs (KBr).Nuclear magnetic resonance (NMR) spectra were recorded on a Bruker AC-200 ( 1 H : 200 MHz and 13 C : 50.3 MHz) spectrometer.CDCl 3 was used as the solvent, with Me 4 Si (TMS) as internal standard.Coupling constants (J) are reported in Hz.Multiplicities are indicated as s (singlet), d (doublet), t (triplet), dd (doublet of a doublet), dt (double triplet), m (multiplet), bs (broad singlet), bd (broad doublet). 13C multiplicities were assigned using a DEPT sequence.Mass spectra were obtained at 70 eV on a VG AutoSpecQ mass spectrometer.Chromatographic purifications were carried out with 230-400 mesh silica gel (flash chromatography).The eluent mixtures, used in the chromatographic separations, were prepared volume to volume (v/v) and are expressed in percentage (%).Thin layer chromatography was performed on aluminium sheets coated with 60 F 254 silica.The TLC were revealed spraying with 2% Ce(SO 4 ) 2 in 2 mol L -1 H 2 SO 4 and followed by heating.The melting points were taken on a Kofler apparatus and are uncorrected.Hydrogenations were carried out using a Parr apparatus.
Epoxidation of 2. Adduct 2 (1.000 g , 3.811 mmol) was dissolved in a solution of AcO 2 H/CH 2 Cl 2 (60 mL), prepared as described above.The resulting solution was kept in the dark and stirred at room temperature for 2 days.The solution was washed with water (2 x 40 mL), aqueous 5% NaHCO 3 (2 x 40 mL) and again with water (2 x 40 mL).The organic layer was separated and the aqueous phases were extracted with CH 2 Cl 2 (3 x 40 mL).The combined organic layers were dried (Na 2 SO 4 ), filtered under activated charcoal and the solvent removed under reduced pressure to furnish diepoxide 3 as a colourless crystalline residue (1.077 g, 96% yield).R f 0.29 (50% EtOAc/hexane).IR (KBr) ν max /cm -