Facile synthesis of azocino[4,3-b]indoles by ring-closing metathesis

The azocino[4,3-b]indole system, tricyclic substructure of the indole alkaloids apparicine and ervaticine, is efficiently assembled by ring-closing metathesis of 2-allyl-3-(allylaminomethyl)indoles. The metathesis sites are introduced into the indole nucleus by reductive amination of a 3-formyl derivative with allylamine, followed by a-lithiation with subsequent electrophilic trapping with acrolein. 2006 Elsevier Ltd. All rights reserved.


Introduction
Ruthenium-catalyzed ring-closing metathesis (RCM) 1 has emerged as a powerful tool for the construction of a great variety of carbo-and heterocycles from acyclic precursors. 2 In particular, the RCM methodology has turned out to be very useful for the synthesis of medium-sized rings, 3 which is generally problematic due to disfavored entropic factors and transannular interactions. Our interest in the development of indole annulation methodologies led us to consider RCM reactions of indole-containing dienes 4,5 for the efficient construction of medium-sized indolo 2,3-fused carboand azacycles, which are common structural arrangements in many natural and synthetic bioactive compounds. 6 In this paper we report a direct synthetic approach to the azocino [4,3-b]indole system I by RCM of appropriate 2,3-dialkenylindoles incorporating a nitrogen atom in the tether linking the two double bonds (Scheme 1). It should be noted that I constitutes the tricyclic substructure of apparicine, 7 an indole alkaloid known for 40 years but still awaiting its first total synthesis, 8 and also the unprecedented 2-acylindole analogue ervaticine. 9

Results and discussion
We set out to explore the feasibility of the protocol using model RCM precursors unfunctionalized at the benzylic indole a-position, such as 2-allyl-3-(allylaminomethyl)indoles 3. 10 For the preparation of these substrates, a fast formylation-reductive amination sequence starting from the known 2-allylindole 1 11 was envisaged (Scheme 2). A strong electron-withdrawing benzenesulfonyl group was placed at the indole nitrogen to guarantee the stability of the proposed gramine-type intermediates. Thus, Friedel-Crafts reaction of 1 with Cl 2 CHOMe in the presence of TiCl 4 gave the aldehyde 2 (90%), which was subjected to reductive amination with allylamine, followed by reaction of the resulting secondary amine with (t-BuOCO) 2 [4,3-b]indole 4a in 60% yield. The N-tosyl derivative 3b proved to be a better substrate as it led to 4b in a higher yield (89%).
With model azocino [4,3-b]indoles in hand, we sought to elaborate C-6 functionalized derivatives simply by extending the chemistry outlined above to an O-protected 2-(1hydroxyallyl)indole. To this end, we selected silyl ether 6, which was easily prepared from aldehyde 5, 12 by reaction with vinylmagnesium bromide followed by protection of the resulting alcohol with tert-butyldimethylsilyl chloride (63% overall yield, Scheme 3). Disappointingly, we were not able to introduce the formyl group needed for the reductive amination step since 6 gave only a complex mixture upon subjection to the above Friedel-Crafts protocol. This unsuccessful result prompted us to change the order of the synthetic steps. Functionalization at the 2-position of a properly 3-substituted indole by a-metalation followed by electrophilic trapping seemed to be the logical solution.
With this aim, we focused our attention on 3-(aminomethyl)indoles 8 and 9, which were available from indole-3-carbaldehyde 7 13 through reductive amination techniques, using tosylamine or, as above, allylamine followed by acylation (Scheme 4). Unfortunately, treatment of these substrates with either LDA, sec-BuLi or tert-BuLi in THF under a variety of experimental conditions, followed by addition of DMF, HCOOMe, or acrolein led to the recovery of the starting product.  We reasoned that the replacement of the indole protecting group by a methoxymethyl (MOM) group could facilitate the a-lithiation, despite a probable reduction in stability of some synthetic intermediates due to the lower electronwithdrawing character of this moiety. Thus, we turned to aldehyde 10 14 (Scheme 5), which was converted into the allylaminomethyl derivatives 11 under the usual conditions. As the tosyl compound 11b partially decomposed under chromatographic purification, we decided to continue the synthesis only with the more stable N-Boc derivative 11a, which could be isolated in a reproducible 72% yield.  We were pleased to find that the desired a-lithiation did take place from 11a upon treatment with tert-BuLi in THF at À78 C. After quenching with acrolein, the unstable alcohol 12 was isolated (79%) and immediately protected as the tert-butyldimethylsilyl ether 13 (64%) or, alternatively, oxidized with MnO 2 (64%) to the ketone 14. Satisfactorily, when 13 was subjected to the previously used RCM conditions (first generation Grubbs catalyst in refluxing dichloromethane) the expected tricyclic compound 15 was obtained in good yield (85%). However, no cyclization was observed from ketone 14 under the above protocol, probably due to the presence of an electron-poor double bond, and only dimeric products coming from intermolecular metathesis reactions were formed. This problem was circumvented simply by using the more efficient second generation Grubbs catalyst at room temperature, leading to tricyclic ketone 16 in 86% yield. Finally, the saturated forms of the eight-membered heterocycles 17 and 18 were obtained by catalytic hydrogenation over Pd/C.

Conclusion
We combined with the easy preparation of the dienic precursors from simple indolic derivatives make this strategy attractive for the construction of medium-sized indolo 2,3-fused carbo-and azacycles, which are scaffolds found in many bioactive compounds.
3 mmol) in CH 2 Cl 2 (6 mL) was added to a solution of TiCl 4 (0.51 mL, 4.7 mmol) and Cl 2 CHOCH 3 (0.4 mL, 4.7 mmol) in CH 2 Cl 2 (6 mL) at À78 C, and the resulting mixture was stirred at À78 C for 2 h. The reaction mixture was diluted with H 2 O, basified with 10% aqueous Na 2 CO 3 , and extracted with CH 2 Cl 2 . The organic extracts were dried and concentrated and the residue was purified by flash chromatography (9:1 hexanes-AcOEt) to give aldehyde 2: 0.69 g (90%); 1 H NMR . Aldehyde 2 (0.33 g, 1.0 mmol) was allowed to react as above with allylamine and the resulting secondary amine (0.32 g) was dissolved in CH 2 Cl 2 (10 mL) and treated with tosyl chloride (0.19 g, 1.0 mmol) and Et 3 N (0.14 mL, 1.0 mmol) at rt overnight. The reaction mixture was diluted with CH 2 Cl 2 and washed with 1 N HCl and brine. The organic extracts were dried and concentrated and the residue was chromatographed (flash, CH 2 Cl 2 ) to give 3b:

1-(Phenylsulfonyl)-3-(tosylaminomethyl)indole (8).
A solution of aldehyde 7 13 (0.5 g, 1.75 mmol) and tosylamine (0.9 g, 5.25 mmol) in dry toluene (15 mL) was heated at reflux (Dean-Stark) for 24 h. The solvent was removed and the residue was dissolved in MeOH (10 mL) and treated with NaBH 4 (66 mg, 1.75 mmol) at rt for 24 h. The solvent was removed and the residue was diluted with H 2 O and extracted with Et 2 O. The organic extracts were dried and concentrated and the resulting residue was purified by flash chromatography (6:4 hexanes-AcOEt) to give tosylamine 8: