[A002]

Juan Domingo Sánchez, Carmen Avendaño and J. Carlos Menéndez

Departamento de Química Orgánica y Farmacéutica, Facultad de Farmacia.
Universidad Complutense, 28040 Madrid, Spain.
E-mail:mailto:josecm@eucmax.sim.ucm.es

Heterocyclic quinones (1) are a very important class of compounds from a biological point of view, particularly as antitumour agents (2). We have described an excellent antitumour activity for many 2,5,8-quinolinetriones fused to  a variety of rings, most notably benzene and pyridine (3). Since antitumour quinones normally act on DNA and because of the well-known DNA-intercalant properties of carbazole, we describe here the preparation of compounds containing both substructures. 
 
 

Starting material preparation is summarized in Scheme 1. Compound 2c was initially prepared by nitration of the known (4) carbostyril derivative 2d, but this procedure was abandoned because of difficulties found in scaling up the preparation of this material.  As an alternative, we developed the route shown in Scheme 1, where N-demethyl analogue of 2d (compound 2a) , readily available on a multi-gram scale (5), was regioselectively nitrated at C-6 and then N-methylated under phase-transfer conditions. Use of silver oxide as the base allowed selective O-alkylation of 2b, leading to the isolation of compound 4:
 
 

Our studies on the reactivity of compound 2c towards Grignard reagents were initially aimed at the preparation of quinones derived from the pyrido[f]indole system by application of the Bartoli indole synthesis, which is based on the reaction of vinylmagnesium compounds with nitroarenes (6). To this end, we carried out the reactions summarized in Scheme 2, finding that the substitution pattern of the Grignard reagents employed was decissive in determining the course of the reaction. Thus, compounds with an alkyl group on the same carbon bearing the magnesium gave the expected fused indoles 5 as the sole products, albeit in low yields and with recovery of substantial amounts of the starting material, even when the reaction was prolonged for up to 24 h. On the other hand, when the nucleophilic carbon atom was unsubstituted, the major products were compounds 6, from conjugate addition at the C-5 position followed by elimination of a molecule of methanol. This mode of addition is probably sterically hindered in the previously mentioned reactions using a-substituted Grignard reagents. Finally, the reaction starting from the 2-methoxyquinoline 4 gave a mixture of compound 6, from addition-elimination at the quinoline C-2 position, and the Bartoli product 5e, which was isolated as a 2-quinolinone derivative. This can be rationalized as a consequence of nucleophilic attack of the Grignard reagent at the methyl group of the 2-methoxy unit rather than at the quinoline C-2 position because the latter reaction is hampered because of the replacement of the strongly electron-withdrawing nitro group by an electron-releasing indole nitrogen during the course of the reaction (Scheme 2).
 
 

In view of these results, we decided to study the chemoselectivity of the reactions between compound 2c and arylmagnesium bromides. We were particularly encouraged by the very limited literature precedent on the synthetic use of the reactions between arylmagnesium halides and nitroarenes, consisting of the preparation of a few biaryls by reaction of 1-methoxy-2-nitronaphthalenes with aryl-Grignard reagents (7). Our results on the reactivity of 2c are summarized in Scheme 3, and show that, contary to this precedent, the main reaction products are the N-arylamines 8, which can be explained bearing in mind the steric hindrance of the position conjugated with the nitro group in 2c (i.e., C-5). In agreement with this explanation, the chemoselectivity in favour of N-arylation was higher when m-substituted Grignard reagents were employed, and complete in the only o-substituted example examined. Since, according to the literature (7b), the reaction between 1-methoxy-2-nitronaphthalene and 1-naphthylmagnesium bromide gives exclusively 1,1'-binaphthyl, we studied the equivalent reaction on our substrate, finding that, in our case, the reaction was completely deviated towards N-arylation, giving 8i as the only product. The less hindered 2-naphthylmagnesium bromide, on the other hand, gave a mixture of compounds 8i and 9i.
 
 

Because of our interest in carbazolequinones as potential antitumour agents, we next examined the palladium-catalyzed oxidative cyclization (8) of compounds 8. Their treatment with palladium acetate in refluxing acetic acid afforded the expected fused dimethoxycarbazoles in low yields, the major products being the desired quinones11. One of the experiments also gave a small amount of quinolinequinone 12f (Scheme 4).  All attempts at cyclizing the naphthyl derivatives 4i and 4j were unsuccesful. The observed in situ oxidative demethylations can perhaps be attributed to the harsh reaction conditions employed and the use of acetic acid as solvent, but we have not examined this mechanism in detail yet.
 
 

The o-tolyl derivative 8b showed a different behaviour, as shown in Scheme 5, the major product being the (o-acetoxyanilino)quinolinequinone 12k. This compound was easily cyclized to the corresponding tetracyclic carbazolequinone by treatment with palladium acetate in acetic acid for an additional 16-h period.

Finally, the minor compounds from the Grignard reaction (9) were also transformed into condensed carbazole derivatives 13 via intermediate nitrenes, generated from the 6-nitro groups by treatment with a trialkyl phosphite (9). Although all attempts at cyclization of compounds 9 with refluxing trimethyl phosphite failed, the reaction succeded in the presence of the higher-boiling triethyl phosphite. In the case of the naphthyl derivative 9i, the cyclization was regioselective and took place exclusively at the naphthalene a position to give compound 13i (Scheme 6).
 
 

financial support of this work by CICYT (FEDER project 2FD-1997-2032) is gratefully acknowledged.
 
 

(1) For a review of the chemistry of heterocyclic quinones, see: Middleton, R. W.; Parrick, J. in The Chemistry of Quinonoid Compounds. Patai, S. and Rappoport, Z., Eds. John Wiley & Sons, 1988; Vol. 2, 1019-1066.

(2) For some reviews dealing with recent research in this area, see: a) Krapcho, A. P.; Maresch, M. J.; Hacker, M. P.; Hazelhurst, L.; Menta, E.; Oliva, A.; Spinelli, S.; Beggiolin, G.; Giuliani, F. G.; Pezzoni, G.; Tognella, S., Curr. Med. Chem. 1995, 2, 803-824; b) Skibo, E. B. Curr. Med. Chem. 1996, 3, 900-931; c) Lee, K. H. Med. Res. Rev. 1999, 19, 569-596.

(3) For a review, see: Avendaño, C.; Menéndez, J. C. Recent Res. Devel. in Organic Chem. 1998, 2, 69-86.

(4) Pérez, J. M., Vidal, L.; Grande, M. T.; Menéndez, J. C.; Avendaño, C. Tetrahedron 1994, 50, 7923.

(5) Avendaño, C., de la Cuesta, E., Gesto, C. Synthesis 1991, 727.

(6) a) Bartoli, G.; Palmieri, G.; Bosco, M.; Dalpozzo, R. Tetrahedron Lett. 1989, 30, 2129-2132. b) Bartoli, G.; Bosco, M.; Dalpozzo, R.; Palmieri, G.; Marcantoni, E. J. Chem. Soc. Perkin Trans. 1 1991, 2757-2761. c) Dobbs, A. J. Org. Chem.2001, 66, 638-641 and references therein. d) Pirrung, M. C.; Wedel, M.; Zhao, Y. Synlett 2002, 143-145.

(7) a) Bartoli, G.; Bosco, M.; Cantagalli, G.; Dalpozzo, R. J. Chem. Soc. Perkin Trans. 2, 1985, 773. Hattori, T.; Taj¡keda, A.; Yamabe, O.; Miyano, S. Tetrahedron 2002, 58, 233-238.

(8) Li, J. J., Gribble, G. W. Palladium in Heterocyclic Chemistry, chapter 3. Pergamon, 2000 (Tetrahedron Organic Chemistry Series, volume 20).

(9) For a recent example, see: Moody, C. J.; Rahimtoola, K. F.; Porter, B.; Ross, B. C.  J. Org. Chem. 1992, 57, 2105-2114.