Rhodium(I)-catalyzed ene-allene carbocyclization strategy for the formation of azepines and oxepines

Article abstract of DOI:10.1021/ol049390a

A novel strategy for the preparation of seven-membered heterocyclic compounds has been realized. Treatment of ene-allene 1 with a catalytic quantity of rhodium biscarbonyl chloride dimer affords the cyclization product 2 in moderate to high yields. The sc

Full text of DOI:10.1021/ol049390a

ORGANIC
LETTERS
2004
Vol. 6, No. 13
2161-2163
Rhodium(I)-Catalyzed Ene-Allene
Carbocyclization Strategy for the
Formation of Azepines and Oxepines
Kay M. Brummond,* Hongfeng Chen, Branko Mitasev, and Anthony D. Casarez
Department of Chemistry, UniVersity of Pittsburgh, Pittsburgh, PennsylVania 15260
kbrummon@pitt.edu
Received April 2, 2004
ABSTRACT
A novel strategy for the preparation of seven-membered heterocyclic compounds has been realized. Treatment of ene-allene 1 with a catalytic
quantity of rhodium biscarbonyl chloride dimer affords the cyclization product 2 in moderate to high yields. The scope and limitations of this
new method are currently under investigation, and the results obtained to date are discussed within.
Our laboratory has been extensively active in exploring
allenes and transition metal catalyzed carbocyclization reac-
tions.1 During our studies involving the Pauson-Khand
reaction, we discovered a Rh(I)/Ir(I)-catalyzed allenic Alder-
ene reaction that affords convenient formation of cyclized
cross-conjugated trienes.2 Given that these trienes are typi-
cally formed in minutes and in high yields, we were very
interested in testing the extreme limits of this reaction. One
of these extreme cases involved the replacement of the alkyne
with an alkene group. Treatment of ene-allene 3 to 5 mol %
[Rh(CO)2Cl]2 in toluene at 90 °C for 2.5 h resulted in the
complete consumption of starting material. To our surprise,
the only isolable product from this reaction was the tetrahy-
droazepine 4, obtained as a single stereoisomer in 42% yield.3
Although the yield of this reaction was somewhat low, the
fortuitous formation of azepine 4 provides another example
of a novel carbon-carbon bond forming process involving
allenes.4 In addition, azepine formation in this manner should
be useful in the preparation of natural products and natural
product-like compounds.5 For these reasons, further inves-
tigations into the scope and synthetic utility of this reaction
were enticing.
(3) For ruthenium-catalyzed cycloisomerization of 1,6-enynes to form
seven-membered carbocycles initiated by C-H activation, see: Trost, B.
M.; Toste, F. D. J. Am. Chem. Soc. 1999, 121, 9728. The initial disclosure
of this finding was published in the abstracts for the 226th ACS National
Meeting, New York, September 7, 2003. After this submission, Itoh reported
a similar result involving the formation of a seven-membered carbocycle:
Makino, T.; Itoh, K. Tetrahedron Lett. 2003, 44, 6335. In a full account a
single azepine was assembled using this carbocyclization method: Makino,
T.; Itoh, K. J. Org. Chem. 2004, 69, 395.
(4) For a palladium-catalyzed carbocyclization to form five-membered
rings involving ene-allenes initiated by C-H activation, see: Franzen, J.;
Backvall, J.-E. J. Am. Chem. Soc. 2003, 6056. For a review on palladium-
catalyzed reactions of allenes, see: Zimmer, R.; Dinesh, C. U.; Nandanan,
E.; Khan, F. A. Chem. ReV. 2000, 100, 3067. For a carbocyclization
involving an allene and a π-allyl palladium intermediate generated from
an allylic acetate, see: Doi, T.; Yangisawa, A.; Nakanishi, S.; Yamamoto,
K.; Takahashi, T. J. Org. Chem. 1996, 61, 2602.
(1) Brummond, K. M.; Sill, P. C.; Chen, H. Org. Lett. 2004, 6, 149.
Brummond, K. M.; Gao, D. Org. Lett. 2003, 5, 3491. Brummond, K. M.;
Chen, H.; Fisher, K. D.; Kerekes, A. D.; Rickards, B.; Sill, P. C.; Geib, S.
J. Org. Lett. 2002, 4, 1931. Brummond, K. M.; Sill, P. C.; Rickards, B.;
Geib, S. J. Tetrahedron Lett. 2002, 43, 3735. Brummond, K. M.; Lu, J.;
Petersen, J. L. J. Am. Chem. Soc. 2000, 122, 4915. Brummond, K. M.; Lu,
J. J. Am. Chem. Soc. 1999, 121, 5087. Brummond, K. M. An Allenic [2 +
2 + 1] Cycloaddition. In AdVances in Cycloaddition; Harmata, M., Ed.;
JAI Press Inc.: Stamford, CT, 1999; Vol. 6, p 211. Brummond, K. M.;
Wan, H.; Kent, J. L. J. Org. Chem. 1998, 63, 6535. Brummond, K. M.;
Wan, H. Tetrahedron Lett. 1998, 39, 931. Kent, J. L.; Wan, H.; Brummond,
K. M. Tetrahedron Lett. 1995, 36, 2407.
(2) Brummond, K. M.; Chen, H.; Sill, P.; You, L. J. Am. Chem. Soc.
2002, 124, 15186. Shibata subsequently reported a similar finding: Shibata,
T.; Takesue, Y.; Kadowaki, S.; Takagi, K. Synlett 2003, 268.
10.1021/ol049390a CCC: $27.50 © 2004 American Chemical Society
Published on Web 05/25/2004

Scheme 1. Deuterium Labeling Study
Table 1. Scope of Cyclization Strategy to Form Heterocycles
the starting material, which showed only broadened peaks
1
and a trace amount of the azepine by H NMR (entry 4).
Replacement of the (E)-alkene in entries 1 and 2 with a (Z)-
alkene gave none of the azepine product. We attribute this
to steric congestion in the initial oxidative addition step. We
were surprised to find that simply increasing the length of
the alkyl chain on the alkene from a methyl to a propyl group
resulted in a substantial decrease in the yield of azepine 6
(28%, entry 5). Likewise, a phenyl group on the terminus
of the alkene also gave only a trace of the azepine (entry 6).
An unsubstituted allyl group gave good yields of the azepines
for both di- and trisubstituted allenes (entries 7 and 8, Table
1). In the case of entry 7 a 1.3:1 mixture of E:Z isomers
was observed. A silicon group on the terminus of the alkene
gave excellent yields of the azepines (entries 9 and 10). It
should be noted that for both entries 9 and 10, there is a
tert-butyl group on the terminus of the allene. An allene and
alkene tethered together by the ester of an amino acid also
gave good yields of the azepines as is demonstrated by entries
11-15. A benzamide (Bz) and carboxybenzyl (CBz) group
on the nitrogen proved to be compatible with the reaction
conditions. Although the CBz-protected amines showed
evidence of decomposition during flash chromatography even
when the silica gel was pretreated with triethylamine.
a Toluene was used as the solvent; DCE was used for all other entries.
b Diastereomeric ratio ) 1.3:1. c Diastereomeric ratio ) 2.5:1. d Diastereo-
meric ratio ) 4.5:1. e Difficult to purify by flash chromatography. f Yield
1
in parentheses was determined by H NMR.
Next, bis allyl ethers 9 were subjected to the same reaction
conditions (entries 16 and 17). The corresponding oxepines
10 were obtained in a somewhat low but promising yield. It
would appear that these conditions are not optimal for the
ether substrates since an NMR study revealed the formation
of multiple byproducts. Finally, attempts to use an all carbon
tether resulted in either decomposition or the isolation of
minor amounts of unidentifiable products. Thus, it would
appear that a heteroatom in the tether is necessary for the
carbocyclization process under these reaction conditions.
The results obtained to date, with regard to the scope of
this azepine-forming process, are summarized below and in
Table 1. In an effort to increase the yield of this transforma-
tion, the catalyst loading was increased from 5 to 10 mol %
of Rh(CO)2Cl]2. This led to a marginal increase in the overall
yield as is evidenced by comparing eq 1 to entry 1, Table 1.
An additional alkyl group on the terminus of the allene 5
gives a much higher yield of the azepine 6 (entry 2, Table
1). Continuing along this line of reasoning, placement of a
tert-butyl group on the terminus of the allene gives one of
the highest yielding conversions observed when using only
5 mol % of the catalyst (entry 3). On the basis of these results
it was not too surprising that substitution of the terminus of
the allene with two hydrogens gave rapid decomposition of
A working mechanism for this reaction has been postulated
on the basis of the results shown in Table 1 and a deuterium
labeling study that was performed, which is shown in Scheme
1. Initially, oxidative addition of the rhodium(I) catalyst into
the allyl fragment proximal to the heteroatom of 13 gives a
π-allyl rhodium deuteride species 14. The conformation
necessary for the oxidative addition most likely prohibits
reaction of the Z-alkene isomer and the terminally disubsti-
tuted alkene. Next, carbocyclization of the rhodium deuteride
species onto the proximal double bond of the allene gives
azepine 15. Finally, reductive elimination affords the product
(5) For a Rh(I)-catalyzed [5 + 2] cycloaddition strategy to form
dihydroazepines, see: Wender, P. A.; Pedersen, T. M.; Scanio, M. J. C. J.
Am. Chem. Soc. 2002, 124, 15154. For azepine-containing natural products,
see: Evans, P. A.; Holmes, A. B. Tetrahedron 1991, 47, 9131; O’Hagan,
D. Nat. Prod. Rep. 1997, 637.
2162
Org. Lett., Vol. 6, No. 13, 2004

16 and the Rh(I) catalyst. The conversion of 13 to 16 occurs
at each C-2 and C-6 position of the azepine ring had
occurred.
in 1 h and gives an 89% yield when using 10 mol % of
1
rhodium biscarbonyl chloride dimer. Moreover, H NMR
In conclusion, treatment of ene-allenes with a rhodium(I)
catalyst results in the formation of tetrahydroazepines and
oxepines.6 Depending upon the substitution pattern of the
substrates the yields can vary from excellent to complete
decomposition of the substrates. We are continuing our
investigations into the scope and limitations of this new ring-
closing process to form seven-membered heterocycles.
data showed that complete incorporation of one deuterium
(6) Representative Experimental Procedure. To a flame-dried test tube
equipped with a magnetic stirring bar were added N-[(E)-2-butenyl]-N-5,5-
dimethyl-2,3-hexadienyl-4-methylbenzenesulfonamde (5, entry 3, Table 1,
29.9 mg, 0.009 mmol) and dichloroethane (DCE, 0.5 mL). The test tube
was evacuated and charged with argon three times, and then [Rh(CO)2Cl]2
(1.7 mg, 0.0045 mmol) was added in DCE. The mixture was heated at 90
°C for 1.5 h and monitored by GC until the starting material was consumed.
The solvent was removed in vacuo, and the residue was purified by flash
chromatography (SiO2, hexanes/ethyl acetate ) 9:1) to afford 6, entry 3,
Table 1, as a colorless oil (28.3 mg, 95%): 1H NMR (300 MHz, CDCl3)
δ 7.72-7.68 (m, 2H), 7.32-7.29 (m, 2H), 6.35 (dd, J ) 1.5 and 9.3 Hz,
1H), 5.21 (d, J ) 1.0 Hz, 1H), 4.83 (dd, J ) 5.5 and 9.3 Hz, 1H), 3.62
(ddd, J ) 3.8, 7.7, 13.5 Hz, 1H), 3.52 (ddd, J ) 3.8, 7.7, 13.5 Hz, 1H),
2.98 (dtq, J ) 1.1, 7.0, 7.0, 1H), 2.73 (dddd, J ) 1.1, 3.8, 7.7, 15.5, 1H),
2.55 (dddd, J ) 1.1, 3.8, 7.7, 15.5, 1H), 2.42 (s, 3H), 1.07 (d, J ) 7.2 Hz,
3H), 1.04 (s, 9H); 13C NMR (75 MHz, CDCl3) δ 143.5, 137.8, 136.4, 135.3,
129.7, 126.9, 125.7, 119.1, 47.8, 41.0, 31.9, 31.2, 29.3, 21.5, 21.1; IR (neat)
ν 2958, 1647, 1350, 1164 cm-1; EI-HRMS calcd for C19H27NO2S [M+]
m/z 333.1763, found 333.1777. The stereochemistry of the double bond
was confirmed by NOE.
Acknowledgment. We gratefully acknowledge the fi-
nancial support provided by the National Science Foundation
and Johnson & Johnson.
Supporting Information Available: Complete experi-
mental details and full data for all ene-allene precursors and
cycloisomerization products. This material is available free
of charge via the Internet at http://pubs.acs.org.
OL049390A
Org. Lett., Vol. 6, No. 13, 2004
2163