Gold(I)-catalyzed Cycloisomerization of the Mixed N,S-acetals Generated from Homopropargylic Amines; Mechanistic Implication for the Formal Alkyne Prins Reaction

Abstract

1 10 AgSbF6 (10) rt 20 h 40% 2 10 AgBF4 (10) rt 22 h 40% 3 10 AgOTf (10) rt 4 h 55% 4 10 AgOTf (10) 0 C 10 h 81% isolated yield after two-step conversion. 1d was isolated in 15% yield. 1d was isolated in 7% yield. Developing new methodologies for the synthesis of piperidines have been attracting the interest of synthetic chemists. Recently, we reported gold(I)-catalyzed cycloisomerization of N,O-mixed acetals generated from homopropargylic amines, which led to a unique approach to highly substitutued piperidin-4-ones. A key aspect of this reaction is the chemoselective activation of alkynes (path A) over the activation of ether groups (path B), as shown in Scheme 1. In an effort to extend the scope of this reaction, we envisioned gold(I)-catalyzed cycloisomerization of the corresponding N,S-mixed acetals. Because the thioether groups are known to attack alkynes in the presence of gold catalysts, the main goal of this study was to compare the reactivity between N,O-acetals and N,S-acetals for the cycloisomerization reaction. Moreover, the cycloisomerization of mixed N,S-acetal would be highly efficient for the preparation of 2-alkylpiperidine alkaloids including coniine and coniceine, because the thioether moiety can be easily removed with the concomitant hydrogenation of the olefin in single step. At the outset of the study, we expected that the cycloisomerization reaction of the N,S-acetals should be faster than N,O-acetal, based on the relative nucleophilicity. However, a preliminary study using 5 mol % of Au[P(C6F5)3]SbF6 showed only poor conversion. Increasing the catalyst loading to 10 mol % provided a mixture of the cycloisomerization product 1b and the piperidin-4-one 1c in ~45% combined yield, although the starting material was completely consumed. Our extensive efforts to isolate the cycloisomerization product in good yields were fruitless, due to the extensive hydration of 1b to form ketone 1c. Thus, we sought to obtain the ketone 1c by treating the crude mixture generated from the cycloisomerization reaction with catalytic p-TsOH (20 mol %). Under this two-step protocol, the ketone 1c was obtained in 40% yield (entry 1, Table 1). Then, we explored the effect of various counteranions. Although using AgBF4 had no effect (entry 2), changing to AgOTf increased the yield to 55% (entry 3).