Studies into the synthesis of forskolin intermediates by electrocyclisation of allene-sulfoxides

Ionone has been converted into bicyclic triene-sulfoxides ( 10 ), ( 16 ) and ( 18 ) by elctrocyclisation of the corresponding divinyl allenesulfoxides ( 9 ), ( 15 ) and ( 17 ) respectively. Reduction of bicycle ( 18 ) by lithium aluminium hydride gave alcohol ( 19 ) by a novel fragmentation.


Introduction
Forskolin (1) is a highly oxygenated diterpene isolated from methanolic extract of the Indian plant Coleus forskohlii by de Souza et al. 1 The molecule displayed very interesting physiological properties: thus it exhibits potent positive inotropic activity and blood pressure lowering properties.In clinical studies, forskolin has shown considerable potential for the treatment of glaucoma, congestive heart failure and bronchial asthma. 2,3Its potential as a drug coupled with its structural complexity have made forskolin the subject of much synthetic effort. 4-16 The use of allene sulfoxide electrocyclisation reactions in synthesis has been developed and exploited by Okamura and co-workers.17Thus it was shown that the dienol (2) could be smoothly converted into the bicyclic compound (3) by treatment with phenylsulfenyl chloride (Fig. 1).

ISSN 1551-7004
Page 855  ARKAT USA, Inc Further transformations carried out on the cyclisation product suggest that such a protocol could be used for an approach to oxygenated terpenoids such as forskolin.

Figure 1
Thus it was speculated that a suitable ionone derivative with the requisite oxygenfunctionality in the ring destined to become the ring A of the natural product could be used for the Okamura cyclisation to generate the required bicyclic substance which would then be transformed into a forskolin precursor (Fig. 2).

Results and Discussion
With this in mind α-ionone (4) was epoxidised regioselectively using m-chloroperbenzoic acid and the resulting epoxide (5) was isomerised to the corresponding hydroxyionone derivative (6) by treatment with potassium carbonate in hot methanol.The hydroxyl group was then protected as the t-butyldimethylsilyl ether (7) according to the Corey-Venkateswarlu procedure. 18The ketone (7) was then transformed into the required acetylene-alcohol ( 8) by simply treating it with lithium acetylideethylene diamine complex.This furnished the product in rather modest yield (<50%) after purification by silica gel chromatography.As this was a very slow reaction and a lot of unconsumed starting material was isolated, a modified protocol was investigated.Thus when the ketone (7) was treated with lithium trimethylsilylacetylide at -78 o C and the trimethylsilylacetylene moiety protodesilylated with catalytic potassium carbonate in methanol the desired product was now obtained in 90% overall yield.The olefinic protons of (8) appeared as doublets at 5.60 and 6.40 ppm (J=16 Hz).The compound (8) was then photoisomerised in the presence of acetophenone as sensitiser to give the corresponding cis isomer (9) in nearly quantitative yield as a 1:1 mixture of diastereoisomers.The 1 H NMR spectrum of the product clearly showed that isomerisation had occurred as the olefinic protons now resonated at 5.86 ppm as a pair of very close doublets (J=12 Hz), appearing as a multiplet for one diastereoisomer and two separate doublets at 5.54 and 5.56 respectively (J=12 Hz) for the other.When this mixture of alcohols was treated with phenylsulfenyl chloride in the presence of triethylamine (TEA) at -78 o C, a rapid reaction ensued as evidenced by the almost instantaneous disappearance of the orange colour of the reagent and a mixture of isomers (10) was obtained in 75-80% yield (Fig. 3).
No attempt was made to determine the composition of the mixture.Nevertheless upon purification by chromatography, one of these isomers was isolated in the pure form as a crystalline solid and exhibited olefinic signals at 5.98 (singlet, 1 H) and 6.08 (broad singlet, 2H) ppm.The methine proton of the silyl ether resonated at 3.92 ppm as a broad doublet (J= 4.0 Hz) indicating an axial silyloxy group in this isomer.
Reaction of the sulfoxide mixture (10) with lithium aluminium hydride at ambient temperature was very sluggish, so refluxing dimethoxyethane was used as medium.Under these conditions a product mixture containing the required diene sulphide (11) as a 1:1 mixture of diastereomers, was obtained.However, it was also evident that substantial sulphide cleavage had occurred on account of the strong odour of thiophenol during work up.The allylic methylene in the NMR spectrum of (11) appeared as two distinct AB quartets: one at 4.15 ppm and the other at 3.65 ppm.The olefinic hydrogens appeared as multiplets at 5.72 ppm and 5.90 ppm and the carbinolic proton as a broad singlet at 4.06 ppm for the axial silyloxy group and as a doublet of doublet at 4.38 ppm (J= 12 Hz and 6 Hz for axial-axial and axial-equatorial couplings respectively) for the equatorial isomer.

Figure 3
At this stage it was intended to convert the sulphide into a sulfoxide (12)  a Pummerer rearrangement 19 to aldehyde (13).Unfortunately upon treatment of the sulphide mixture with m-chloroperbenzoic acid, only a 40% yield of the required sulfoxide (12) was obtained and when the latter was treated with acetic anhydride to induce a Pummerer 19 rearrangement none of the desired product was obtained (Fig 4 ).

Figure 4
At this stage it was felt that the silyloxy group was possibly the source of our problem in view of its steric bulk.Therefore it was decided to revert to a sterically less demanding protecting group and the benzyloxymethyl group was judged to be a suitable candidate.The alcohol (9) was deprotected using TBAF (Fig 5) to furnish the diol (14) and this was selectively protected as the benzyloxymethyl ether ( 15) by treatment with benzyloxymethyl chloride in the presence of diisopropylethylamine.Exposure to phenylsulfenyl chloride gave the expected mixture of sulfoxides ( 16) in 70% yield.However, little progress was made, as transformations on this mixture also proved problematic.
Next, the sequence was repeated with the monoacetate (17) of alcohol (14), this was prepared in 76% yield by exposure to a mixture of acetic anhydride, pyridine and DMAP at 0 o C. Upon treatment with phenylsulfenyl chloride in the usual manner, the desired sulfoxide mixture (18) was obtained in 70% yield (Fig 6 ).In conclusion it can be said that the allenesulfoxide electrocyclisation reaction provides a good method to assemble a cyclohexadiene system compatible with the skeleton of natural products like forskolin.More work is however required for the further transformation of the cyclised products into suitable synthetic precursors.

Experimental Section
General Procedures.Infrared spectra were run as thin films.The NMR spectra were run at 300 MHz for 1 H NMR in CDCl3 solution using tetramethylsilane (TMS) as internal standard.Mass spectra and high-resolution mass spectra (HRMS) were measured using the electron impact (EI) technique.Chromatography refers to the flash column technique over Merck Kiesel gel 60H (230-400 mesh) and thin layer chromatography was carried out using plates precoated with Merck silica 60F254.Ether and petrol (fraction boiling between 40 o and 60 o C) were distilled prior to use.Tetrahydrofuran was dried by refluxing over sodium and benzophenone under nitrogen and then distillation.Benzene, toluene, dichloromethane and acetonitrile were dried by refluxing over calcium hydride and then distillation.

Cyclisation of compound (9).
A 50-ml two-neck flame dried flask was charged successively with dry methylene chloride (30 ml); alcohol (9) (1.500 g, 4.310 mmol) and dry triethylamine (1.300 ml, 9.482 mmol).The solution was cooled to -78 o C under nitrogen.Phenylsulfenyl chloride (0.747 g, 5.172 mmol) in dry methylene chloride (2 ml) was added dropwise whereupon its orange colour was spontaneously discharged.When addition was completed, the pale yellow solution was stirred for 30 min at -78 o C and then allowed to warm up to ambient temperature over 30 min.The reaction was quenched by addition of aqueous sodium bicarbonate solution (15 ml).The layers were partitioned and the aqueous phase was extracted with a further portion of methylene chloride (20 ml).The combined organics were washed (brine) and dried (Na2SO4).Evaporation and purification of the crude product by chromatography (20% petrol/ ether -> 80% ether/ petrol) gave a mixture of diastereomers (1.344 g, yellow syrup) and a pure sample of the most polar diastereomer (0.215 g, 79% combined yield).The latter compound had: δH (CDCl3  The silylether (9) (400 mg, 1.117 mmol) was dissolved in dry THF (8 ml) and the solution was treated with a solution of TBAF in THF (1 M, 1.3 ml) at room temperature.The solution was stirred at room temperature for 30 min and then was diluted with water (20 ml).The mixture was extracted with ether (2 x 25 ml), the combined organics were washed with brine (25 ml) and dried (Na2SO4  14) (150 mg, 0.641 mmol) was dissolved in dry methylene chloride (5 ml) and the solution was cooled to 0 o C followed by addition of acetic anhydride (0.5 ml, 5.304 mmol), pyridine (0.5 ml, 6.113 mmol) and DMAP (10 mg, 0.082 mmol).The solution was stirred with slow warming up to room temperature over about 4 h.The reaction mixture was then diluted with methylene chloride (25 ml) and was washed with sodium bicarbonate solution (20 ml), water (20 ml), 2 M HCl (20 ml) and finally brine (20 ml).The solution was dried (Na 2 SO 4 ) and then evaporated to yield the product (17), as a pale yellow oil (167 mg, 95%).The compound had: νmax/cm