The first preparation of the unstable 1-hydroxy-2,3-dimethylindole, and structural determination of its air-oxidized product, 3-hydroxy-2,3-dimethyl-3 H -indole N -oxide 1

1-Hydroxy-2,3-dimethylindole ( 1 ) has been prepared for the first time. Under atmospheric oxygen, 1 was converted rapidly into 3-hydroxy-2,3-dimethyl-3 H -indole N -oxide ( 2 ). The structure was deduced, based on its products obtained by the reaction with Ac 2 O in pyridine and confirmed by X-ray single crystallographic analysis


Introduction
In 1989, 2 we devised a simple and general synthetic method for 1-hydroxyindoles 3 consisting of the following two steps: (1) reduction of indoles to 2,3-dihydroindoles, (2) their oxidation with 30% H 2 O 2 in the presence of a catalytic amount of sodium tungstate or phosphotungstate.
Application of our 1-hydroxyindole synthetic reaction to 4 showed the major formation of an unknown product (A), monitored on TLC.Upon standing in the reaction mixture at room temperature, this product (A) changed rapidly into another new, stable product (B).
Attempts to isolate A in pure state were not successful because it polymerized during column-chromatography owing to its instability and sensitivity to air.Therefore, the in situ methylation with Me 2 SO 4 /K 2 CO 3 of the A in the reaction mixture obtained immediately after the oxidation of 4 was examined.As a result, the stable 1-methoxy-2,3-dimethylindole (5a) was obtained in 22% overall yield from 4. A similar acetylation of A with Ac 2 O in pyridine afforded the stable 1-acetoxy-2,3-dimethylindole (5b) in 63% yield.These facts prove that the product (A) is 1-hydroxy-2,3-dimethylindole (1).Finally, we succeeded in the isolation of 1 in 67% yield by the alkaline hydrolysis of 5b, and its spectroscopic data were taken.The half-life of 1 in the pure state is found to be about 24 h.A half-life of 1 in CHCl 3 at room temperature under atmospheric oxygen is shown to be about 4 h.

Scheme 1
On the basis of the above observations, the compound (B) was obtained in 64% overall yield from 4 by the sequence of reactions: (1) oxidation of 4 to 1 and work-up, and (2) allowing a CHCl 3 solution of crude 1 to stand at room temperature under atmospheric oxygen for 22 h, a sufficient time for transforming 1 completely into B.
In an attempt to determine the structure of B chemically, it was reacted with Ac 2 O in pyridine at room temperature, providing a 56% yield of 3-methylindole-2-carboxaldehyde (8) together with 5-(6a) and 7-acetoxy-2-acetoxymethyl-3-methylindoles (7a) in 2 and 3% yields, respectively.The yields of these three products were not affected by reaction temperature or time as can be seen from Table 1.The reaction of B with trifluoromethanesulfonic anhydride did not produce 6b, 7b, and 8 at all, (in contrast to Ac 2 O).With trifluoroacetic anhydride and benzoic anhydride, the expected formation of 6c,d and 7c,d was not observed, although the aldehyde (8) was obtained in 8-24% yield.The aldehyde (8) was identical with the authentic 3-methylindole-2-carboxaldehyde, prepared by Vilsmeier-Haack reaction of 3-methylindole (9) by the procedure of Chatterjee et al. 6 The structures of 6a and 7a were determined by the following observations.Compounds 6a and 7a were allowed to react with Boc 2 O in CH 2 Cl 2 in the presence of N,Ndimethylaminopyridine (DMAP) 7 to provide 5-acetoxy-(10) and 7-acetoxy-2-acetoxymethyl-1tert-butoxycarbonyl-3-methylindole (11) in 96 and 82% yields, respectively.Comparison of the 1 H NMR spectra of 10 with 6a demonstrates an anisotropy effect of the Boc group on the orthocoupled C(7)-proton (δ 8.15, d, J = 9.0 Hz) by ca.0.9 ppm, proving these are 5-substituted compounds.In the case of 11, however, the protons did not shift toward lower magnetic field compared with those of 7a.This fact confirms that 7a and 11 are 7-substituted indoles.
There are still other possibilities that 6a and 7a are the 3-acetoxymethyl-2-methylindoles, 12a and 12b, respectively, as shown in Scheme 2. In order to eliminate these structures, we applied the following series of reactions to 7a.First, 7a was converted into 7-hydroxy-3methylindole-2-methanol (13) in 73% yield by the hydrolysis with LiOH in MeOH.Then, selective methylation of the phenolic group of 13 with CH 2 N 2 provided an 87% yield of 7methoxy-3-methylindole-2-methanol 8 (14).Subsequent oxidation of 14 with active MnO 2 in CHCl 3 afforded a formyl compound (15b) in 44% yield.By these reactions, 12b should have been transformed into 15a.At this stage, comparison of the 1 H NMR spectrum of 14 with that of 15b was made, and no anisotropy effect of the formyl group on the C(4)-proton was observed.
Consequently, the formyl group should not be present at the 3-position of the indole nucleus and 15b is proved to be 7-methoxy-3-methylindole-2-carboxaldehyde. 8 The structure of 6a was determined by comparing the differences between the 1 H NMR chemical shifts of the diacetoxy compounds (6a, 7a) and the 1-acyl compounds (10, 11).As shown in Table 2, the ∆δs observed between protons of 10 and 6a were almost the same with those observed between protons of 11 and 7a.These findings confirm that 6a has a structure similar to 7a.Although the structure (B) was deduced to be 2 on the basis of its spectroscopic data and chemical behavior upon reaction with Ac 2 O as stated above, we still needed conclusive evidence.Finally, we luckily found a suitable recrystallizing solvent for B, providing prisms for X-ray single crystallographic analysis.The results shown in Figure 1 clearly show a novel 3-hydroxy-2,3-dimethyl-3H-indole N-oxide (2) structure.The atomic parameters are listed in Table 3.
The mechanism for the formation of 6a, 7a, and 8 may be explained as shown in Scheme 3. In the reaction of 2 with Ac 2 O, the diacetate ( 16) is formed initially.The following [3,3]sigmatropic rearrangement of either the 3-(route a) or 1-acetoxy groups (route b) to the methylene carbon at the 2-position affords 17 or 18, respectively.In the intermediate (17), subsequent [3,5] or [3,3]-sigmatropic rearrangement of the 1-acetoxy group toward the benzene ring results in the formation of 6a or 7a (route c).On the other hand, 8 would be formed either through the vinylic acetate 19, a tautomer of 18, or through the vinylic acetate 20 originating from 17 (route d).19 20 [3,3] sigmatropic rearrangement [3,3] sigmatropic rearrangement In conclusion, we have succeeded in the first preparation of 1-hydroxy-2,3-dimethylindole (1), which was an unstable compound and oxidized rapidly into a novel compound, 3-hydroxy-2,3-dimethyl-3H-indole N-oxide (2), on standing at room temperature under atmospheric oxygen.An interesting chemical behavior of 2 upon reaction with Ac 2 O was also demonstrated.

Experimental Section
General Procedures.Melting points were determined on a Yanagimoto micro melting point apparatus and are uncorrected.IR spectra were determined with a HORIBA FT-720 spectrophotometer, and 1 H NMR spectra with a JEOL GSX-500 spectrometer, with tetramethylsilane as internal standard.MS spectra at 70 eV using electron impact mode were recorded on a JEOL SX-102A or JEOL JMS-GC mate mass spectrometer.Column chromatography was performed on silica gel 60N (SiO 2 , 70-230 mesh, from Kanto Chemical Co.Inc.) throughout the present study.
Positional parameters and B (eq) for 2 are summarized in Table 3.

Table 1 .
The reaction of 2 with Ac 2 O