Synthesis and determination of the absolute configuration of Fudecalone

We have synthesized the proposed structure of Fudecalone, but its NMR spectral data were not identical with those of the natural compound. Further investigations on the conformational isomerization and synthesis of a diastereomer show natural fudecalone to be a trans -fused octalone. The absolute configuration has been determined following the synthesis of both enantiomers


Introduction
In 1995, Omura and his coworkers isolated and identified a new drimane sesquiterpene, fudecalone, 1, from a culture broth of Penicillium sp.FO-2030. 1 It exhibited anticoccidial activity against monensin-resistant Eimeria tenella at concentrations of more than 16 µM.The structure was elucidated mainly by NMR, and the conformation was reported to be 1a, as shown in Fig. 1.As outlined in our preliminary communications, 2,3 synthetic 1 was not identical with natural fudecalone, and we have found that the correct structure is 2. In this paper, we describe details of the synthesis of 1 and 2. We also record the synthesis of both enantiomers of 2, which enables us to clarify the absolute configuration of fudecalone.

Results and Discussion
First, to synthesize the proposed cis-fused structure of fudecalone, we planned the retrosynthetic strategy shown in Scheme 1.We thought fudecalone 1 would be obtained by reduction of a keto lactone 3, which was prepared by the cationic cyclization of an enol ether 4. The compound 4 might be obtained by Birch reduction of the known phthalide 6 4 followed by alkylation of the resulting lactone enolate with the homoprenyl halide 5. We started our synthesis from the known phthalide 6. 4 However, the reported procedure was lengthy and complicated, and we developed a simpler method for multi-gram preparation of 6, as shown in Scheme 2. Diels-Alder reaction of the known pyrone 7 with dimethyl acetylenedicarboxylate followed by decarboxylation gave the aromatic diester 8a. 5 For the regioselective reduction of one of the two ester carbonyls, 8a was converted into the half-esters, 8b and 8c.However, disappointingly, all our attempts at reduction (LiBH 4 or DIBAL for 8b, BH 3 for 8c were unsuccessful, probably owing to the electron-donating effect of the methoxy group at the para position.We then decided to take advantage of the electronic effect of this group.Both ester groups of 8a were reduced to the diol 9, whose treatment with conc.HCl in ether at 0 °C gave the chloro-alcohol 10 regioselectively.Jones' oxidation and then base treatment caused lactone formation to afford 6 in 63% yield over five steps from 7.

Scheme 2
Birch reduction of 6 with potassium in liquid ammonia, followed by alkylation of the resulting enolate with the iodide 11 6 gave 12 in 62% yield (Scheme 3).Acid hydrolysis of the methyl enol ether and isomerization of the double bond afforded the conjugate enone 13 (46%), which was converted into an enol acetate 14 with LDA and acetic anhydride.When the compound 14 was treated with BF gas in wet CH 2 Cl 2 , 7 an axial attack of a cationic side chain to enol acetate took place and the desired cis-octalone 3 was obtained as the sole product, in excellent yield (90%).On the other hand, direct cyclization of 12 under various conditions was unsuccessful.The stereochemistry, including the conformation, was confirmed by X-ray analysis.For converting 2 into fudecalone, an unsaturated ketone and lactone carbonyl were reduced with DIBAL (4 equiv.)followed by selective oxidation of the resulting allylic alcohol 15 with MnO 2 to give 16 as an inseparable diastereomeric mixture.Disappointingly, neither diastereomer had 1 H NMR identical with the natural product, although the patterns of the peaks were very similar.Further NOESY experiments of 15 revealed the stereochemistry to be 1b, which is the conformational isomer of the reported structure 1a.

Scheme 3 ISSN 1551-7012
According to MM3 calculations, the steric energy for 1b is more stable by 2.1 kcal/mol than for 1a, and we thought it interesting that the unstable conformer might be produced in the biogenetic process and can exist as a natural product because of the extra-high energy barrier between 1a and 1b.We therefore investigated the conformational isomerization from 1b to 1a.
We have reported previously, 3 that the intermediate 2 was converted in seven steps into the selenide 17, which had the same conformation as 1a (Scheme 4).However, when the selenide was eliminated by oxidation with m-CPBA, the conformation of the obtained enone 18 has changed to the previous 1b type again, which was confirmed by X-ray analysis.From these results, we concluded that the energy barrier between 1a and 1b was not so high and that the proposed relative stereochemistry was incorrect, and so we attempted to synthesize the diastereomers.

Scheme 4
We supposed two types of possible diastereomers; the trans-octalone 19 and the trans-lactol 20 (Figure 2).As most drimane sesquiterpenoids have a trans-fused decalin skeleton, we preferred the trans-octalone 19 as the next target.The previous intermediate 3 was treated with various bases to isomerize it from the cis-fused octalone to the trans-isomer (Scheme 5).However, treatment with organic bases resulted only in a recovery or a decomposition of 3, and aq.NaOH gave an eliminated product 22.To avoid elimination, we then investigated the epimerization, after hydrogenation of the double bond.

Scheme 5
The enone 3 was hydrogenated with Pd-C to afford a ketone 23 as almost a single isomer (Scheme 6).This direct hydrogenation made the methyl group of 23 α-oriented, because the αface of the olefin is sterically congested by the axial 7-methyl group.Treatment of 23 with aq.LiOH and re-lactonization of the resulting hydroxy acid, by refluxing with PPTS in benzene for 4 hours gave 26 in 69 % yield over two steps.Interestingly, when the trans-ketone 26 was treated with PPTS for a longer time, it was completely isomerized into the cis-ketone 23.On the other hand, when cis-ketone 23 was treated under the same conditions, or with DBU under reflux in THF, it was not converted into the trans-ketone 26.The lactone-opening was therefore essential for the isomerization from cis-23 to trans-26 via 24 and 25.These results were supported by MM3 calculations.The steric energy of cis-23 is 1.7 kcal/mol more stable than that of trans-26, and trans-25 was estimated to be 1.9 kcal/mol more stable than cis-24.

Scheme 6
The enolate derived from 26 by LDA treatment was trapped with TMSCl to give the silyl enol ether 27.Treatment of 27 with NBS gave the bromide 28, then dehydrobromination with DBU afforded an enone 21 in 79% overall yield, and the stereochemistry of 21 was confirmed by X-ray analysis (Scheme 7).The unsaturated ketone and the lactone carbonyl of 21 were reduced with DIBAL (4 eq.), and the resultant allylic alcohol was re-oxidized with MnO 2 to furnish (±)-2 in 56% yield over two steps.In this case, the lactol was obtained as a single isomer and the stereochemistry of its OH group was determined to be S* by NOE experiment.The 1 H and 13 C NMR spectroscopic data of (±)-2 showed complete accordance with those of natural fudecalone, except for the NOE data. 3

Scheme 7
The NOESY spectrum of our (±)-2 showed slight differences from the reported data, as illustrated in Figure 3. NOEs of (±)-2 between H-6a and each of H-3a and H-10 ax indicated that C-6a had the S* configuration, and NOEs between H-1 and each of H-9 ax and H-10 eq indicated that C-1 also had the S* configuration.On the other hand, NOEs between H-6a and each of H-1 and H 3 -13 of 1a indicated by dashed arrows, which were the decisive factor in determining fudecalone to be 1a, 1 were hardly observed.From these points of view, the relative configuration of fudecalone was determined to be 1S*, 3aS To determine the absolute configuration, we next started a synthesis of optically active fudecalone.The racemic enone of 21 was reduced with NaBH 4 and CeCl 3 to give the allylic alcohol 29, which was converted into a diastereomeric mixture of camphanates.Two diastereomers were easily separated by silica-gel column chromatography to afford the (+)-and (−)-camphanates, 30 and 31, in 49% and 46% yields, respectively.Methanolysis of 30 gave (−)-29 and the absolute configuration was determined at this stage by a modified Mosher's method 8 after converting (−)-29 into the corresponding (R)-and (S)-MTPA esters 33 and comparing their chemical shifts in 1 H NMR (Scheme 8).The lactone of (−)-29 was then reduced with DIBAL and selective oxidation of the allylic alcohol (+)-32 with MnO 2 afforded (+)-fudecalone [(+)-2].Similarly, 31 was converted into (−)-2 in 72% yield over three steps.Although the optical rotations and melting points of our synthetic (+)-and (−)-2 did not agree with those reported, we concluded that the natural fudecalone was (+)-2 from the positive rotation.
In conclusion, we have synthesized the proposed structure of fudecalone 1b, which was found to be a conformational isomer of the proposed structure 1a.Although our investigations of the conformational isomerization from 1b to 1a were fruitless, we synthesized the trans-fused octalone 2 as a racemate and found that it showed identical NMR spectroscopic data with that of the natural fudecalone.By further NOESY experiments, we determined the correct relative configuration of fudecalone.Moreover, by optical resolution, we synthesized both enantiomers and determined the absolute configuration of the natural compound to be 1S, 3aS, 6aS, 10aS.
Bioassays of both enantiomers and some intermediates are now in progress, and the details will be reported in due course.General Procedures.All boiling points (bp) and melting points (mp) were uncorrected.RT denotes room temperature.Infrared spectra (IR) were measured on a Jasco FT/IR-230 spectrometer.Proton and carbon-13 magnetic resonance spectra ( 1 H-NMR and 13 C-NMR) were recorded on a JEOL JNM-AL300 or a JEOL JMN α-500 spectrometer.Chemical shifts are reported in ppm (δ) relative to internal chloroform (δ 7.26 for 1 H and δ 77.0 for 13 C).HR-FAB-MS spectra were recorded on a JEOL JMS-HX110 mass spectrometer.Optical rotations were measured on a Jasco DIP 1000 polarimeter.Melting points were measured on a Yanagimoto micro melting apparatus.Analytical thin-layer chromatography (TLC) was carried out using 0.25 mm Merck silica gel 60 F 254 precoated glass-backed plates.Compounds were visualized by ultra violet light (254 nm), iodine vapor, or phosphomolybdic acid spray reagent.Column chromatography was performed on Merck silica gel 60 or Kanto Chemical 60 N (spherical, neutral).All solvents were reagent grade.Tetrahydrofuran (THF) and diethyl ether were freshly distilled from sodium/benzophenone under argon.Dichloromethane, benzene, and hexamethylphosphoric triamide (HMPA) were distilled from calcium hydride and stored over 4A-molecular sieves.Methanol and absolute ethanol were used without purification.

5-Methoxy-3-methylphthalic acid 2-methyl ester (8b).
To a solution of the dimethyl ester 8a, 3.32 g, 13.9 mmol) in THF (40 ml) and MeOH ( To a solution of the residue in THF (750 ml) aq.NaOH (42 g, 996 mmol, 200 ml) was added dropwise at 0 ºC and the mixture was stirred for 18 h at RT.The mixture was acidified to pH 2 by the addition of conc.HCl and the aqueous layer saturated with (NH 4 ) 2 SO 4 , and extracted with EtOAc.The organic extracts were washed with brine, dried over MgSO 4 , and concentrated in vacuo.The residue was purified by column chromatography (hexane:EtOAc=3:1 Potassium (7.00 g, 0.17 g atom) was dissolved in liquid NH 3 (100ml) at -78ºC under argon atmosphere.To this was added a solution of phthalide (12.6 g, 71 mmol) and t-BuOH (7.7 ml, 78 mmol) in THF (100 ml) and the mixture was stirred for 10 min.After the addition of dried LiBr (17.5 g, 201 mmol) followed by stirring for 1 h, NH 3 was evaporated off below -30 ºC under slightly reduced pressure and THF (100 ml) was added to the residue.A solution of the iodide 11 (19.4 g, 92 mmol) in HMPA (37 ml) was added to the mixture at -78 ºC.After stirring for 2 h, the reaction mixture was poured into water and extracted with Et