Indole synthesis via rhodium catalyzed oxidative coupling of acetanilides and internal alkynes.

The prevalence of indoles in bioactive molecules has prompted the development of many useful methods for their preparation. Despite advances, most of these techniques rely heavily on substrate preactivation, and very few employ readily available anilines as starting materials. Transition metal catalyzed processes are illustrative, commonly requiring ortho-halogenated anilines as starting materials thus adding cost and reducing the breadth of readily available starting materials. A major advance in the minimization of substrate preactivation in indole synthesis was recently described by Glorius involving the condensation of simple anilines with 1,3-dicarbonyl compounds followed by Pd(II)-catalyzed oxidative cyclization. Similarly, Gagné, Lloyde-Jones, and Booker-Milburn have described an efficient indoline synthesis via Pd(II) catalyzed coupling of N-arylureas with activated dienes. Herein we describe the realization of a different approach, namely a rhodium catalyzed oxidative coupling of N-acetyl anilines and alkynes, as well as preliminary mechanistic studies. Based on the known ability of N-acetyl anilines to undergo orthometalation, and the ease with which N-acetyl indoles may be deprotected, N-acetylaniline 1a was selected along with 1-phenyl-1propyne 2a for reaction development. Extensive experimentation with a variety of Pd(II) catalysts failed to produce indole 3a, instead giving small amounts of Heck and multiple Heck-type additions. Similarly, Wilkinson’s catalyst, which has been shown to induce similar reactivity with 1,2-diaryldiazines, gave none of the desired product. A low but promising outcome was obtained, however, with [Rh(Cp*)Cl2]2 in conjunction with a stoichiometric Cu(II) oxidant, a catalyst system recently employed by Satoh and Miura and by Jones for other reactions initiated via cyclometalation. Under these conditions, small amounts of 3a (∼3% yield) could be detected by GCMS analysis of the crude reaction mixture (Table 1, entry 1). Continued optimization revealed that the presence/absence of chloride anions and the choice of solvent exert a dramatic impact on reaction outcome. For example, the addition of silver triflate to sequester the chloride ligands increases the yield 5-fold (entry 2). A solvent screen revealed that the yield could be further increased to 55% by using tert-amyl alcohol (1,1-dimethylpropanol) (entry 6), and a second assay of silver salts revealed that, with 10 mol% silver hexafluoroantimonate, a 79% isolated yield of 3a could be obtained as one regioisomer (entry 9). The reaction time could also be significantly shortened since 3a can be isolated in 69% yield even after 5 min of reaction (entry 10). Conversely, if 1 equiv of LiCl is added to the reaction mixture, product formation is completely inhibited, adding additional weight to the observation that chloride anions exert a negative role on catalyst activity (Table 1, entry 11). Both electron-rich (3b, 3h, and 3i) and -deficient (3c and 3d) acetanilides participate in good yield (Table 2). A chloride substituent