HETEROCYCLES

Communication | Special issue | Vol. 80, No. 2, 2010, pp. 879-885
Received, 19th August, 2009, Accepted, 11th September, 2009, Published online, 16th September, 2009.
DOI: 10.3987/COM-09-S(S)115

■ Novel Radical Cyclization Method Accompanied by Elimination of Hydrazyl Radical

Shoji Kobayashi, Hidefumi Hirao, Tatsuro Kawauchi, and Ilhyong Ryu*

Department of Chemistry, Graduate School of Science, University of Osaka Prefecture, Sakai, Osaka 593-8531, Japan

Abstract

The potential of hydrazine group as the new radical leaving group was studied. Radical cyclization of (E)-1,1-dialkyl-2-(1-alkyloct-2-en-7-ynyl)hydrazines with n-Bu3SnH/AIBN, followed by protodestannylation, gave 1-alkenyl-2-methylenecyclopentanes, which arose by 5-exo cyclization and subsequent β-elimination of hydrazyl radical.

Radical cyclization is a powerful method for the synthesis of cyclic molecules including natural products, bioactive agents, and various functional molecules.1 A wide variety of leaving groups such as halogens, selenides, sulfides, sulfoxides, sulfones, stannanes, and xanthates are generally used, if addition/elimination sequence is involved in the process.2 These functional groups, however, sometimes cause problems, especially in the stage of their introduction, owing to insufficient functional group compatibility, inherent toxicity of reactants, or other unforeseen difficulties encountered.

In the meantime, during the course of our studies to develop new radical carbonylation reactions for lactam ring synthesis,3,4 we discovered that cyclization of the acyl radicals proceeded regioselectively onto the nitrogen of imines (and oxazoline) C­–N bonds, providing the 6-endo cyclization products (Scheme 1, Eq. 1).4f To expand this methodology, we have examined carbonylative radical cyclization of alkynyl hydrazone 4 in the presence of Bu3SnH and V-40 as a radical initiator (Scheme 1, Eq. 2). In contrast with 6-endo cyclization onto imine nitrogen, 6-endo cyclization onto hydrazone nitrogen was not smooth, giving 5 in low yield, together with a small amount of 2-((tributylstannyl)methylene)cyclopentanone 6, which does not contain a hydrazine unit.5

It can be rationalized that δ-lactam 5 was generated by 6-endo cyclization of α,β-unsaturated acyl radical 8 (Scheme 2).6 On the other hand, it was supposed that cyclopentanone 6 was generated through hydrazine 9, which was initially produced by 5-exo cyclization of 8, by expelling the hydrazyl radical from cyclopentane ring with the aid of tributyltin radical addition to carbonyl oxygen. To the best of our knowledge, there are no published studies demonstrating that such a hydrazine moiety acts as a radical leaving group. Moreover, only a few examples have been reported for the cleavage of N-C bond to generate the nitrogen-centered radicals.7,8 We therefore considered that evaluating the potential of hydrazine functionality as the radical leaving group would provide not only a conceptually new method for radical reactions, but a new entry into the generation of nitrogen-centered radical for further usage. Herein, we wish to report a novel radical cyclization method involving elimination of the hydrazyl radical, which used vinyl radical cyclization as a model.

We have designed a reaction system consisting of 5-exo cyclization of vinyl radical 12 due to exceedingly facile cyclization mode.7a,9 Incorporation of carbon or heteroatom as X into the backbone would produce the corresponding 1,4-diene with carbocycle or heterocycles. The fragmentation of intermediate alkyl radical 13 to afford diene 14 and hydrazyl radical 16 would be the key to the success of this approach. Since the nitrogen-carbon bond is generally strong, it seems of importance to investigate the substituents (R2 and R3) that can stabilize the N-centered radical as much as possible to facilitate β-scission.10

We initially envisaged the use of inexpensive N,N-dimethylhydrazine as the leaving group and decided to study the behavior of (E)-1,1-dimethyl-2-(oct-2-en-7-ynyl)hydrazine 21a as a model (Scheme 4). To prepare this compound, we embarked on the sequence shown in Scheme 4. Thus, commercially available 5-hexyn-1-ol 17 was converted to aldehyde 19 in a conventional manner. Condensation of 19 with N,N-dimethylhydrazine in the presence of molecular sieves 3Å proceeded smoothly at 80 °C to give the conjugated hydrazone 20. However, we were unable to obtain allylhydrazine 21a by reduction of 20 with LiAlH4. The parallel study indicated that allylhydrazine lacking α-substituent was unstable and it was immediately auto-oxidized to the corresponding hydrazone during isolation. Therefore, n-butyl group was introduced through 1,2-addition to suppress undesirable auto-oxidation, thereby affording the chemically stable hydrazine 21b in 94% isolated yield.

With allylhydrazine 21b in hand, we then examined the radical cyclization under thermal conditions (Scheme 5). A mixture of 21b (0.05 M), AIBN (30 mol%), and Bu3SnH (1.2 equiv) in benzene was refluxed for 3 h. After usual work-up, the crude product was purified by silica gel column chromatography. As expected, cyclic diene 15b which does not contain the hydrazine moiety was obtained in 16% yield, along with cyclic hydrazine 22 (37%) and acyclic hydrazine 23 (22%).11 This result suggests that dimethylhydrazine can serve as a leaving group in the tin radical-promoted 5-exo cyclization, albeit with low efficiency. We suspected that competitive hydrogen abstraction of the intermediate radical (cf. 12 and 13) from Bu3SnH leading to 23 and 22 would be reduced by decreasing the concentration of the reaction system. Indeed, the proportion of cyclic diene 15b was marginally increased (15b:22:23 = 30:45:25) when the reaction was carried out at a concentration of 0.015 M of 21b.12 However, the influence of concentration on the product selectivity was less than we expected.13

Next, we explored the effect of substituents (R2 and R3) on the hydrazine moiety (Table 1). Each substrate was readily prepared from the common intermediate 19 in a similar manner to that described in Scheme 4. When one of the substituents on the nitrogen was replaced by m-tolyl group, the yield of 15b was increased up to 31% (entry 2). Much improvement was realized by introducing two phenyl groups on the hydrazine moiety (entry 4). In contrast, benzyl substituents showed negative effects and no cyclic diene 15b was obtained, while cyclic hydrazine akin to 22 was formed in a small amount (entry 3).14 Encouraged by the positive results achieved by introducing the aromatic hydrazines, we then explored the substituents at α-position (R1). The radical cyclization of methyl-substituted allylhydrazine 21f provided diene 15c in 22% yield along with the cyclic hydrazine akin to 22 in 61% yield (entry 5). It is likely that the reduced steric hindrance at the position α to radical provokes smooth hydrogen abstraction from tin hydride, thereby decreasing the yield of 15c. In contrast, the reaction with phenyl-substituted hydrazine 21g gave rise to the desired diene 15d in 60% yield with an excellent E/Z ratio (entry 6).15

In summary, we have developed a new radical cyclization method employing hydrazine as the leaving group. The fragmentation of intermediate alkyl radical 13 proceeds efficiently to give 1,4-diene 14 even though it involves scission of a rather strong carbon-nitrogen bond. Among the substrates tested, diphenylhydrazine gave the best result. From a synthetic standpoint, a variety of hydrazines are readily accessible from aldehydes or ketones by simple operations. These preliminary investigations represent an important guideline for developing a new type of radical cyclization with nitrogen-based leaving groups, which could be useful for the construction of both carbocycles and heterocycles. Further investigations to improve the leaving-group ability, as well as the use of other starting radicals are currently underway.16

ACKNOWLEDGEMENTS
This work was supported by a Grant-in-Aid for Scientific Research from MEXT and JSPS, Japan.

References

1. For reviews on free radical reaction used in the synthesis of natural products and related molecules, see: (a) C. P. Jasperse, D. P. Curran, and T. L. Fevig, Chem. Rev., 1991, 91, 1237; CrossRef (b) P. J. Parsons, C. S. Penkett, and A. J. Shell, Chem. Rev., 1996, 96, 195; CrossRef (c) M. Malacria, Chem. Rev., 1996, 96, 289; CrossRef (d) T. Naito, Heterocycles, 1999, 50, 505; CrossRef (e) H. Ishibashi, T. Sato, and M. Ikeda, Synthesis, 2002, 695; CrossRef (f) W. R. Bowman, A. J. Fletcher, and G. B. S. Potts, J. Chem. Soc., Perkin Trans. 1, 2002, 2747. CrossRef
2. ‘Radicals in Organic Synthesis,’ Vol. 1 and Vol. 2, ed. by P. Renaud and M. P. Sibi, Wiley-VCH, Weinheim, 2001.
3. For reviews on radical carbonylations, see: (a) I. Ryu and N. Sonoda, Angew. Chem., Int. Ed. Engl., 1996, 35, 1051; CrossRef (b) I. Ryu, N. Sonoda, and D. P. Curran, Chem. Rev., 1996, 96, 177; CrossRef (c) C. Chatgilialoglu, D. Crich, M. Komatsu, and I. Ryu, Chem. Rev., 1999, 99, 1991; CrossRef (d) I. Ryu, Chem. Soc. Rev., 2001, 30, 16. CrossRef
4. For our recent work, see: (a) I. Ryu, K. Matsu, S. Minakata, and M. Komatsu, J. Am. Chem. Soc., 1998, 120, 5838; CrossRef (b) T. Fukuyama, Y. Uenoyama, S. Oguri, N. Otsuka, and I. Ryu, Chem. Lett., 2004, 33, 854; CrossRef (c) M. Tojino, Y. Uenoyama, T. Fukuyama, and I. Ryu, Chem. Commun., 2004, 2482; CrossRef (d) Y. Uenoyama, T. Fukuyama, O. Nobuta, H. Matsubara, and I. Ryu, Angew. Chem. Int. Ed., 2005, 44, 1075; CrossRef (e) I. Ryu, Y. Uenoyama, and H. Matsubara, Bull. Chem. Soc. Jpn., 2006, 79, 1476; CrossRef (f) M. Tojino, N. Otsuka, T. Fukuyama, H. Matsubara, and I. Ryu, J. Am. Chem. Soc., 2006, 128, 7712; CrossRef (g) Y. Uenoyama, T. Fukuyama, and I. Ryu, Org. Lett., 2007, 9, 935. CrossRef
5. For acyl radical cyclization onto hydrazone carbon, see: I. M. Brinza and A. G. Fallis, J. Org. Chem., 1996, 61, 3580. CrossRef
6. We have not precluded fully an alternative mechanism that includes 5-exo cyclization of acyl radical 8 followed by aziridination to produce azabicyclo[3.1.0]hexane ring that rearranges to δ-lactam to release its strain. For computational studies, see: (a) H. Matsubara, C. T. Falzon, I. Ryu, and C. H. Schiesser, Org. Biomol. Chem., 2006, 4, 1920; CrossRef (b) C. H. Schiesser, U. Wille, H. Matsubara, and I. Ryu, Acc. Chem. Res., 2007, 40, 303. CrossRef
7. For reviews on the generation and use of nitrogen-centered radicals: (a) A. G. Fallis and I. M. Brinza, Tetrahedron, 1997, 53, 17543; CrossRef (b) L. Stella, ‘Radicals in Organic Synthesis,’ Vol. 2, ed. by P. Renaud and M. P. Sibi, Wiley-VCH, Weinheim, 2001, pp. 407-426; CrossRef (c) S. Z. Zard, Chem. Soc. Rev., 2008, 37, 1603. CrossRef
8. (a) O. Kitagawa, S. Miyaji, Y. Yamada, H. Fujiwara, and T. Taguchi, J. Org. Chem., 2003, 68, 3184; CrossRef (b) J. Kemper and A. Studer, Angew. Chem. Int. Ed., 2005, 44, 4914; CrossRef (c) J. Guin, S. Mück-Lichtenfeld, S. Grimme, and A. Studer, J. Am. Chem. Soc., 2007, 129, 4498. CrossRef
9. M. Newcomb, ‘Radicals in Organic Synthesis,’ Vol. 1, ed. by P. Renaud and M. P. Sibi, Wiley-VCH, Weinheim, 2001, pp. 318-336.
10. For investigation of substituent effects on the stability of N-centered radicals of aniline derivatives, see: Z. Li and J.-P. Cheng, J. Org. Chem., 2003, 68, 7350. CrossRef
11. The crude 1H NMR spectroscopy indicated that the Bu3Sn group on exomethylene was present after the reaction, but was removed during silica gel column chromatography.
12. In this case, decrease of reaction rate and incomplete conversion of starting material were observed.
13. It is assumed that 1,5-hydrogen shift in alkyl radical 13 may be a competitive reaction course, thereby unavoidably affording 22 even under the diluted conditions.
14. This observation supports the possibility of the intermediate alkyl radical undergoing 1,5-hydrogen shift to provide the stable benzyl radical (see, footnote 13).
15. Typical procedure: A magnetic stirring bar, 21g (70.1 mg, 0.191 mmol), AIBN (15.2 mg, 0.0926 mmol), Bu3SnH (62.0 µL, 0.230 mmol) and benzene (4.5 mL) were placed in a 20 mL 2-neck round-bottomed flask and the mixture was brought to reflux for 3 h under argon. The solvent was removed under reduced pressure. The residue was purified by flash chromatography on silica gel (hexane) to give 15d (21.1 mg, 0.115 mmol, 60%). Compound 15d: 1H-NMR (500 MHz, CDCl3) δ 1.50-1.66 (m, 2H), 1.78-1.82 (m, 1H), 1.98-2.03 (m, 1H), 2.33-2.40 (m, 1H), 2.43-2.48 (m, 1H), 3.13 (q, 1H, J = 8.3 Hz), 4.82 (s, 1H), 4.95 (s, 1H), 6.09 (dd, 1H, J = 15.6, 8.3 Hz), 6.40 (d, 1H, J = 16.0 Hz), 7.20 (t, 1H, J = 7.4 Hz), 7.30 (t, 2H, J = 7.8 Hz), 7.37 (d, 2H, J = 7.4 Hz); 13C-NMR (125 MHz, CDCl3) δ 24.63, 32.68, 34.26, 48.75, 106.57, 126.04, 126.89, 128.46, 130.00, 133.04, 137.67, 155.24. ; IR (neat) ν 3060, 3026, 2954, 2867, 1651, 1495, 1448 cm-1; HRMS (EI): m/z [M]+ calcd for C14H16: 184.1252, found 184.1228.
16. Recently substituted allyldiphenylphosphine oxides were developed as radical allylating agents with group 15 elements, see: G. Ouvry, B. Quiclet-Sire, and S. Z. Zard, Angew. Chem. Int. Ed., 2006, 45, 5002. CrossRef