We report a highly efficient substitution reaction of 3-hydroxyoxindoles 1 using TMSN 3 to furnish 3-substituted 3-azidooxindoles as the precursor of quaternary 3-aminooxindoles. Ga(OTf) 3 is found to be the most efficient Lewis acid for this reaction, and catalyst loading could be lowered down to 1.0 mol%. Accordingly, under an nitrogen atmosphere, to a 10 mL Schlenk tube are successively added Ga(OTf) 3 (0.05 mmol, 25.8 mg), 3-hydroxyoxindoles 1 (0.5 mmol) and 2.5 mL of acetonitrile, followed by the addition of TMSN 3 (1.5 mmol, 207 μL). The resulting mixture is stirred at 25 ℃ till the full consumption of the 3-hydroxyoxindoles, monitored by TLC analysis. After the solvent is removed under reduced pressure, the residue is directly subjected to column chromatography, using an eluent of petroleum ether/ethyl acetate (10/1, V/V) as the eluent, to afford the desired products 3. Under this condition, a variety of differently substituted 3-aryl 3-hydroxyoxindoles 1a～1m work well with TMSN 3 to provide the corresponding 3-aryl 3-azidooxindoles 3a～3m in good to excellent yield. 3-Methyl 3-hydroxyoxindoles 1n～1o also react with TMSN 3 to give the desired product 3n～3o in good yield. But 3-allyl 3-hydroxyoxindole 1p reacts with TMSN 3 poorly under this condition, affording product 3p in only 27% yield. These results indicate that this method has certain universality, but the reaction is influenced by the substituents to some extent. The reaction could be run on a gram-scale, as evidenced by the reaction of 3-hydroxyoxindole 1a and TMSN 3 on a 6.0 mmol scale with 1.0 mol% Ga(OTf) 3 at 60 ℃, giving 3-phenyl 3-azidooxindole 3a in 89% yield (1.41 gram). The product 3a can be converted to quaternary 3-aminooxindoles 4 in 60% yield at 50 ℃ by Pd-catalyzed hydrogenation. To understand the reaction mechanism, the optically active 3-allyl 3-hydroxyoxindole 1p (84% ee) is chosen to react with TMSN 3 under the standard condition, which affords product 3p in racemic form. This result provides strong evidence that a prochiral intermediate is involved in the reaction. We tend to believe that the reaction is initiated by the dehydration of 3-hydroxyoxindole 1p to generate a reactive benzylic cation which will react with TMSN 3 to give the desired product 3p.

No abstract