Synthesis of and chemical model reaction studies with 3-deoxyandrogens: evidence supporting a 2, 3-enolization hypothesis in human placental aromatase catalysis

Abstract

A number of hitherto undescribed Δ²- and Δ³-3-deoxy-5α-androgen derivatives (17β,19-diols 12 and 13, 19-hydroxy-17-ketones 16 and 17, 19-oxo-17-ketones 18 and 19) were synthesized in good yield. The lithium-ammonia reduction of 19-(tetrahydropyran-2-yloxy)androst-4-ene-3, 17-dione 7 followed by Shapiro reaction allowed easy construction of both Δ²- and Δ³-3-deoxy-5α-steroid systems. An improved synthesis of the known Δ⁴-3-deoxyandrogen derivatives (28, 30, 31) was accomplished in high yield. Masking of the 19-hydroxy group was necessary in order to generate the Δ⁴-3-deoxyandrogen system in good yield, in contrast to the account of an earlier synthesis of compounds 28 and 30 by Numazawa. Chemical model reactions of the third oxidative process in aromatase action were carried out with Δ²-, Δ³-, Δ⁴- and Δ^2,4-3-deoxy-19-oxo-17-keto steroids (18, 19, 31 and 5, respectively). The findings illustrate the need for both Δ²- and Δ⁴-unsaturation in order to generate Δ¹⁽¹⁰⁾-unsaturation under the model reaction conditions. This study supports our 2, 3-enolization hypothesis in the proposed catalytic mechanism for human placental aromatase. The importance of the stereoelectronic nature of substrate 1β-H in the enzyme-catalysed third oxidative process is discussed.