Six- versus five-membered ring formation in radical cyclizations of 7-bromo-substituted hexahydroindolinones

Article abstract of DOI:10.1021/ol036347z

(Equation presented) Radical cyclization of N-allyl-7-bromo-3a-methyl- hexahydroindol-2-one affords a six-membered ring product that prevails over the isomeric five-membered compound. The former product is generated through two reaction pathways: (a) 6-en

Full text of DOI:10.1021/ol036347z

ORGANIC
LETTERS
2004
Vol. 6, No. 6
917-920
Six- versus Five-Membered Ring
Formation in Radical Cyclizations of
7-Bromo-Substituted
Hexahydroindolinones
Paitoon Rashatasakhon, Ayse Daut Ozdemir, Jerremey Willis, and Albert Padwa*
Department of Chemistry, Emory UniVersity, Atlanta, Georgia 30322
chemap@emory.edu
Received December 2, 2003
ABSTRACT
Radical cyclization of N-allyl-7-bromo-3a-methyl-hexahydroindol-2-one affords a six-membered ring product that prevails over the isomeric
five-membered compound. The former product is generated through two reaction pathways: (a) 6-endo-trig ring closure and (b) rearrangement
of an intermediate methylenecyclopentyl radical obtained by 5-exo-trig cyclization.
The aspidosperma indole alkaloid family constitutes an
important class of naturally occurring compounds that has
attracted the attention of synthetic chemists due to the
interesting biological properties of some of its members.1
These alkaloids share as part of their structure the [6.5.6.5]-
ABCE ring system found in aspidospermidine (1) and
spegazzinidine (2). The presence of the sterically congested
to address this problem.3 Our interest in this area led us to
consider an approach to the ABCE tetracyclic core 4, wherein
an amido-substituted furan undergoes an intramolecular
Diels-Alder reaction across a tethered indole π-bond. In an
earlier report we demonstrated that the [4 + 2]-cycloaddition/
rearrangement sequence was remarkably efficient given that
two aromatic rings were compromised in the reaction
(Scheme 1).4 To apply this strategy to the synthesis of the
Scheme 1
C(7) quaternary carbon center represents a particular chal-
lenge toward the synthesis of this family of natural products.2
A number of creative strategies have emerged over the years
indole alkaloid spegazzinidine (2), we needed to address the
problem of assembling the final D-ring of the pentacyclic
skeleton. In this communication we report an efficient general
strategy for construction of the remaining six-membered
D-ring that is based on a cyclohexenyl radical cyclization
of a hexahydroindolinone intermediate.
(1) (a) Saxton, J. E. Nat. Prod. Rep. 194, 11, 493. (b) The Alkaloids;
Brossi, A., Suffness, M., Eds.; Academic Press: San Diego, 1990; Vol.
37, pp 145-204.
(2) Overman, L. E.; Sworin, M. In Alkaloids: Chemical and Biological
PerspectiVes; Pelletier, S. W., Ed.; Wiley-Interscience: New York, 1985;
Vol. 3, pp 275-307.
10.1021/ol036347z CCC: $27.50 © 2004 American Chemical Society
Published on Web 02/14/2004

a heterocyclic array are much less common. Seminal studies
by Bechwith9 and Stork’s groups10 have shown that, under
tin hydride mediated reaction conditions, vinyl radical
cyclization gives a mixture of both 5-exo and 6-endo
products. The kinetic work by Beckwith revealed that
formation of the six-membered ring is not solely due to a
6-endo-trig cyclization but is the result of a rapid rearrange-
ment of the methylene cyclopentyl radical, via a reversible
3-exo-trig cyclization.11 It was envisioned that by keeping
the hydride concentration low, rearrangement of the kineti-
cally formed radical 9 derived from bromide 7 to the
thermodynamically more stable radical 10 would occur,
leading to product 12. Comparison of the strain energies of
9 and 10, as well as radical stability, supports this idea.
Indeed, when 7 (0.01 M) was treated with tributyltin hydride
and a catalytic amount of AlBN, six-membered ring product
12 was the major product formed in 89% yield. In contrast,
when bromide 7 was treated with Bu3SnH at a concentration
of 0.1 M, a significant quantity (20%) of the 5-exo cyclization
product 13 (3:1 mixture of diastereomers) was obtained along
with the 6-endo cyclization product 12 in a ratio of 1:3,
together with the simple reduction product 6 (19%). These
results clearly indicate that the vinyl radical rearrangement
pathway is responsible, to a considerable extent, for the
regiochemical outcome of the reaction.
The cyclization method was next extended to the N-benzyl-
substituted hexahydroindolinones 14 and 15. Exposure of
bromo-enamide 14 to Bu3SnH under standard radical forming
conditions furnished pyrrolo[3.2.1-de]phenanthridinone 16
in 68% yield together with 27% of the reduced hexahydroin-
dolinone 17. In this case, selective 6-endo cyclization took
place on the aromatic ring. Interestingly, the closely related
o-bromobenzyl-substituted hexahydroindolinone 15 failed to
cyclize but instead gave largely the reduction product 17 in
75% yield. The different behavior observed with these two
systems is presumably reflective of the slower rate of addition
to the enamido π-bond.12 The successful cyclization of 14
encouraged us to also apply the reaction to the simpler
8-bromo-hexahydro-1H-quinolinone system 18. Gratifyingly,
subjecting a sample of 18 to the standard radical conditions
furnished the cyclized pyrido[3.2.1-jk]carbazolonone 19 in
81% yield, thereby demonstrating the facility of the 5-exo-
trig cyclization pathway.13 At this juncture, we decided to
extend our studies toward the homologous N-butenyl hexahy-
droindolinone 20. A review of the literature revealed,
somewhat surprisingly, that simple 1,6-heptadienyl radicals
Scheme 2
Scheme 2 depicts the basic features of our strategy directed
toward aspidospermidine construction. The first step, selec-
tive bromination of the enamido π-bond, should be extremely
rapid and efficient since analogous examples are known.5
We hoped that generation of a cyclohexenyl radical (i.e., 8)
from 7 would initiate a 6-endo-trig cyclization ultimately
leading to 12 after abstraction of hydrogen from tributyltin
hydride. A model study designed to test the feasibility of
this concept began by the condensation of allylamine with
1-methyl-(2-oxocyclohexyl) acetic acid to give the desired
bicyclic lactam 6 in 96% yield.6 Hexahydroindolinone 6 was
subsequently treated with bromine in CH2Cl2 followed by
reaction with NEt3 to deliver the cyclization precursor 7 in
95% yield. Exposure of 7 to several radical cyclization
conditions led to various mixtures of the 6-endo- and 5-exo-
trig cyclization products 12 and 13, with the best yield and
product ratio obtained using n-Bu3SnH/AIBN in refluxing
benzene under slow addition conditions.
Since their introduction in 1982,7 vinyl radical cyclizations
have been widely used in organic synthesis,8 although
preparative sequences incorporating vinyl radicals as part of
(3) (a) Woodward, R. B.; Cava, M.; Ollis, W. D.; Hunger, A.; Daeniker,
H. U.; Schenker, K. J. Am. Chem. Soc. 1954, 76, 4749. (b) Wenkert, E. J.
Am. Chem. Soc. 1962, 84, 98. (c) Magnus, P.; Gallagher, T.; Brown, P.;
Pappalardo, P. Acc. Chem. Res. 1984, 17, 35. (d) Stork, G.; Dolfini, J. E.
J. Am. Chem. Soc. 1963, 85, 2872. (e) Marino, J. P.; Rubio, M. B.; Cao,
G.; Dios, de, A. G. J. Am. Chem. Soc. 2002, 124, 13398. (f) Kuehne, M.
E.; Bornmann, W. G.; Earley, W. G.; Marko, I. J. Org. Chem. 1986, 51,
2913. (g) Sumi, S.; Matsumoto, K.; Tokuyama, H.; Fukuyama, T. Org.
Lett. 2003, 5, 1891.
(4) Lynch, S. M.; Bur, S. K.; Padwa, A. Org. Lett. 2002, 4, 4643.
(5) (a) Wei, L. L.; Mulder, J. A.; Xiong, H.; Zificsak, C. A.; Douglas,
C. J.; Hsung, R. P. Tetrahedron 2001, 57, 459. (b) Goffin, E.; Legrand, Y.;
Viehe, H. G. J. Chem. Res., Synop. 1977, 105.
(7) Stork, G.; Baine, N. H. J. Am. Chem. Soc. 1982, 104, 2321.
(8) (a) Giese, B.; Kopping, B.; Go¨bel, T.; Dickhaut, J.; Thoma, G.;
Kulicke, K. J.; Trach, F. Org. React. 1996, 48, 301. (b) Motherwell, W.
B.; Crich, D. Free Radical Reactions in Organic Synthesis; Academic
Press: London, 1992. (c) Curran, D. P. Synthesis 1988, 4117. (d) Curran,
D. P. Synthesis 1988, 489. (e) Jasperse, C. P.; Curran, D. P.; Fevig, T. L.
Chem. ReV. 1991, 91, 1, 1237.
(9) Beckwith, A. L. J.; O’Shea, D. M. Tetrahedron Lett. 1986, 27, 4525.
(10) Stork, G.; Mook, R., Jr. Tetrahedron Lett. 1986, 27, 4529.
(11) More recently, Crich’s group reported preferential formation of the
5-exo product when the reaction was conducted in a rapid radical quenching
environment (i.e., PhSeSePh/Bu3SnH), reconfirming that the five-membered
ring closure is the kinetically favored process; see Crich, D.; Hwang, J.-T.;
Liu, H. Tetrahedron Lett. 1996, 37, 3105.
(6) (a) Ragan, J. A.; Claffey, M. C. Heterocycles 1995, 41, 57. (b) Ennis,
M. D.; Hoffman, R. L.; Ghazal, N. B.; Old, D. W.; Mooney, P. A. J. Org.
Chem. 1996, 61, 5813.
(12) Vazquez, A. N.; Garcia, A.; Dominguez, D. J. Org. Chem. 2002,
67, 3213.
918
Org. Lett., Vol. 6, No. 6, 2004

preferred route now corresponds to 7-endo-trig cyclization
leading to 22 in 72% yield together with minor quantities of
the 6-exo cyclized product 21 (18%). The presence of a
3-bromo substituent on the N-but-3-enyl π-bond sufficiently
retards the 6-exo-trig cyclization, so that 7-endo closure now
becomes the predominant path. Formation of azepine[3.2.1-
hi]indolone 22 from the initially generated 7-endo cyclized
radical would involve hydrogen atom abstraction from Bu3-
SnH followed by further reaction of the resulting secondary
bromide with tributyltin radical in the traditional manner.
The isolation of 21 as the minor product from this reaction
is also consistent with the suggestion that the 6-exo cyclized
radical (i.e., 25) is rapidly quenched by hydrogen abstraction
to give tertiary bromide 26, which is then reduced under the
reaction conditions. Another possible explanation is that
Scheme 3
Scheme 5
have not been thoroughly investigated.7 This state of affairs
is most likely attributable to the poorer prospects for synthetic
utility of this higher homologue of the 1,5-hexadienyl radical.
On the basis of studies using the parent 6-heptenyl radical
as a model,14 the rate of 6-exo-trig cyclization is expected
to be an order of magnitude slower than the 1,5-hexadienyl
cyclization, its ring closure should be considerably less
regioselective, and it is also possible that a [1,5]-hydrogen
atom transfer could occur to produce a 1-butenylallyl radical.
A solution of 20 in benzene was treated with Bu3SnH/AIBN
Scheme 4
radical 25 ejects the adjacent bromine atom to give diene
27, which in turn is reduced to 21.17 Although this pathway
seemed less likely, we decided to prepare a sample of 27 in
order to probe the likelihood of this possibility. We found
(vide infra) that when diene 27 was subjected to the standard
Bu3SnH/AIBN conditions, it could be recovered unchanged,
thereby eliminating this mechanistic possibility.
at 80 °C for 12 h. Workup and chromatography led to a 7:2
mixture of the 6-exo and 7-endo cyclization products.15 The
regiochemical preference in cyclization of vinyl radicals is
known to parallel that of the alkyl analogues.16 Although
6-exo-trig and 7-endo-trig modes of cyclization are possible
with hexahydroindolinone 20, six-membered ring formation
predominates. Interestingly, when N-(3-bromo-but-3-enyl)
hexahydroindolinone 23 was reacted with Bu3SnH/AIBN
under similar conditions, the same two products were formed
but in a strikingly different ratio. With this system, the
Our approach toward the synthesis of diene 27 was based
on an intramolecular Stille cross-coupling reaction18 of
hexahydroindolinone 28. The intramolecular Stille reaction
was performed with PdCl2(PPh3)2 (5% mol) as catalyst using
a microwave reactor at 100 °C. To our surprise, the
palladium-catalyzed reaction of 28 gave the seven-membered
cyclized diene 29 as the major product in 48% yield together
with lesser quantities of the six-ring diene 27 (16%). The
formation of 29 presumably results from a preferential
7-endo-trig cyclization, and the regiochemical outcome is
similar to that encountered with the radical cyclization of
(13) The cyclization of 18 to 19 is somewhat related to systems studied
by Ishibashi and co-workers. For leading references, see: Ishibashi, H.;
Ishita, A.; Tamura, O. Tetrahedron Lett. 2002, 43, 473.
(14) (a) Beckwith, A. L. J.; Moad, G. J. Chem. Soc., Chem. Commun.
1974, 472. (b) Chatgilialoglu, C.; Ingold, K. U.; Scaiano, J. C. J. Am. Chem.
Soc. 1981, 103, 7739. (c) Newcomb, M. Tetrahedron 1993, 49, 1151.
(15) Pyrrolo[3.2.1-ij]quinolin-2-one 21 was obtained as a 2.3:1 mixture
of diastereomers.
(16) Beckwith, A. L. J.; Ingold, K. U. In Rearrangements in Ground
and Excited States; de Mayo, P., Ed.; Academic Press: New York, 1980.
(17) Still another possibility is the occurrence of a 1,2-bromo group
migration followed by quenching of the tertiary radical and eventual
reduction of the resulting primary bromide.
(18) (a) Marsault, E.; Deslongchamps, P. Org. Lett. 2000, 2, 3317. (b)
Bru¨ckner, S.; Abraham, E.; Klotz, P.; Suffert, J. Org. Lett. 2002, 4, 3391.
(c) Zhang, H. X.; Guibe, F.; Balavoine, G. J. Org. Chem. 1990, 55, 1857.
Org. Lett., Vol. 6, No. 6, 2004
919

ring systems. Further work on these cyclizations and the
synthesis of spegazzinidine and its analogues is currently
underway and will be reported in due course.
Scheme 6
Acknowledgment. This research was supported by the
National Science Foundation (grant CHE-0132651). A.D.O.
thanks Istanbul Technical University for a Visiting Scholar-
ship.
Supporting Information Available: Complete descrip-
tion of the synthesis and characterization of all compounds
prepared in this study. This material is available free of
charge via the Internet at http://pubs.acs.org.
dibromide 23. In conclusion, our studies have shown that
the D-ring of the aspidospermidine skeleton can be efficiently
achieved by a free radical cyclization of a N-alkenyl-7-bromo
hexahydroindolinone derivative. The present methodology
should be useful for the construction of other heterocyclic
OL036347Z
920
Org. Lett., Vol. 6, No. 6, 2004