Synthetic studies towards 1α-hydroxysolasodine from diosgenin and the unexpected tetrahydrofuran ring opening in the Birch reduction process.

The synthesis of 1α-hydroxysolasodine from diosgenin was attempted. The Pd/C catalyst mediated dehydrogenation of diosgenin generated the 1,4,6-trien-3-one (3), which was reacted with Ac2O in pyridine in the presence of a catalytic amount of POCl3 followed by hydrolysis to give the 22-hydroxyfurostan (4) in 65% yield. Conversion of the primary 26-OH group into the azide and simultaneously 22-OH dehydration were achieved in one step by Mitsunobu reaction. Treatment of the (25R)-26-azidofurosta-1,4,6,20(22)-tetraen-3-one (5) with chlorotrimethylsilane (TMSCl)/NaI/MeCN and cyclisation in situ provided the (22R,25R)-spirosola-1,4,6-trien-3-one (6) in good yield. Stereoselective and regioselective epoxidation of trienone (6) with 30% H2O2 and 5% NaOH in methanol gave the 1α,2α-epoxy-(22R,25R)-spirosola-4,6-dien-3-one (7). Birch reduction of the epoxide (7) with Li/NH3 in THF followed by the treatment with NH4Cl, however, failed to generate the expected 1α-hydroxysolasodine, but provided a tetrahydrofuran ring opening product, (22S,25R)-1α,16β-dihydroxy-22,26-epiminocholest-4-en-3-one (8). Compounds 3 and 5-8 as well as solasodine were evaluated for their cell growth inhibitory activities against human prostate cancer PC3, human cervical carcinoma Hela, and human hepatoma HepG2 cells. At the concentration of 10 μM, only epoxide 7 displayed moderate inhibitory rates towards these cells (40-54%).