An unexpected and useful E-ring oxidative cleavage in furostanes

Treatment of a acetylated furostanols derived from steroid sapogenins with a wet mixture of KMnO 4 and Fe 2 (SO 4 ) 3 .nH 2 O in CH 2 Cl 2 / tert -butyl alcohol produces the unexpected E-ring oxidative cleavage to afford the corresponding cholestan-22,16-diones. Based in this new reaction, facile syntheses of an OSW-1 precursor and kryptogenin acetate have been designed.


Introduction
The stereoselective β-epoxidation of ∆ 5 -steroids using biphasic systems that involve potassium permanganate and metal salts has received considerable attention. 1In particular, we have reported the transformation of 3β-hydroxy-∆ 5 -steroids into the corresponding 3β-hydroxy-6-oxosteroids [i.e cholesterol acetate into 3β-hydroxy-cholestan-6-one and diosgenin acetate into laxogenin], using a protocol that involves β-epoxidation, HBr induced oxirane ring opening, oxidation the resulting bromohydrin to a bromoketone and reduction to the parent ketone by treatment with Zn dust in refluxing acetic acid, (see Scheme 1). 2

Results and Discussion
Attempt at β-epoxidation of the acetylated furostanol 1a using wet KMnO 4 /Fe 2 (SO 4 ) 3 in CH 2 Cl 2 /tert-butyl alcohol afforded the unchanged starting material and two compounds that after separation were characterized as the desired epoxidated furostanol 2a (16.6%) and the unexpected epoxidated diketone 3a (43.3%), (see Scheme 2 and Table 1, entry 1).In an experiment conducted without the ferric salt the unchanged starting material was recovered after 24 hours.

Scheme 2
Some reports have shown that the E-ring of furostanes undergoes cleavage in the presence of Lewis acids.For instance, Ac 2 O/BF 3 .Et 2 O treatment of furostanes has been reported to produce E-ring cleavage followed by 17→16 hydride shift and Wagner-Meerwein rearrangement to afford ∆ 13 -17β-methyl-18-norsteroids (Scheme 3). 3 Activation of acetic anhydride by BF 3 .Et 2 O resulted in acetylation of O-16 that triggers the observed rearrangement.The absence of a nucleophile capable to produce the nucleophilic displacement of the acetylated oxygen, justifies the observed course.
In this new reaction of furostanic compounds with wet KMnO 4 /Fe 2 (SO 4 ) 3 , the fact that in absence of the ferric salt no reaction was observed suggests a mechanism in which, after coordination to the ferric cation, the tetrahydrofuranic oxygen attached to both C-16 and C-22 can be displaced by water present in the reaction media leading to a diol which is rapidly oxidized to the observed 16,22-diketone, (Scheme 4).The distribution of products indicates that the epoxidation of the double bond is faster than the E-ring cleavage and prompted us to study the reaction conditions for the total conversion of 1a into 3a.We were pleased to find that the simple increase on the amounts of the reagents and reaction time resulted in the conversion of the starting material 1a into 3a, (see Table 1, entry 2).It has been reported that treatment of 1b with K 2 Cr 2 O 7 in acetic acid led to the 16,22diketone 4b in poor yields.4a In an alternative approach, oxidation with Oxone ® of the acetylated i-steroid 5, derived from diosgenin, produced the hemiketal 6 and the desired 16,22-diketone 7. Four times recirculation of the mixture through the same oxidation procedure afforded 7 in 80% that was isomerised to 4b in a long sequence to the synthesis of the potent antitumor agent OSW- After those facts we decided to explore the application of this new reaction to the synthesis of kriptogenin acetate (4a), 5 and 4b that has been employed as synthetic precursor of OSW-1.4b Application of the new reaction conditions to 1a or to our previously reported 1b 6d led to the corresponding epoxidated diketones 3a and 3b which were converted into the desired 5,6 unsaturated derivatives 4a (kryptogenin acetate) and 4b by consecutive treatments with HBr in CH 2 Cl 2 and Zn dust in refluxing acetic acid.

Table 1 .
Results of of epoxidation-oxidative E-ring opening of compound 1a

Acetoxy-5β,6β-epoxycholestan-16,22-dione (3b). Obtained
Wet KMnO 4 /Fe 2 (SO 4 ) 3 .nH 2 O in CH 2 Cl 2 /tert-butanol, ii) HBr/CH 2 Cl 2 iii) Zn(dust)/acetic acid reflux Reactions were monitored by TLC on ALUGRAM ® SIL G/UV 254 5 cm x 2.5 cm plates from MACHEREY-NAGEL.Chromatographic plates were sprayed with a 1% solution of vainillin in 50% HClO 4 and heated until colour developed.NMR spectra were recorded in on Varian Unity INOVA (300 or 400 MHz) spectrometers using TMS for 1 H or the solvent signal (CDCl 3 ) for 13 C as reference.Mass spectra were recorded on a Jeol SX-102-A spectrometer.Melting points were measured on Melt-Temp II equipment and are uncorrected.KMnO 4 and Fe 2 (SO 4 ) 3 .nH 2 O were finely grounded in a mortar, H 2 O was added and the mixture was placed in a round bottom flask containing CH 2 Cl 2 (10 ml) (see Table1for amounts of KMnO 4 , Fe 2 (SO 4 ) was stirred at room temperature for the indicated time.Ethyl ether (25 ml) and celite (2 g) were added and the mixture was stirred for 15 min before filtering through a small pad of silicagel and elution with ethyl ether (25 ml) and ethyl acetate (2 x 30 ml).The filtrate was washed with H 2 O (5 x 30 ml), saturated aqueous NaCl solution (2 x 30 ml), dried (anh.Na 2 SO 4 ) and evaporated to afford the desired compound.