The rearrangement of vinyl cyclopropyl ketone 1a (X )
Me; R1 ) R2 ) R3 ) H) to dihydrofuran 2a may be catalyzed
under forcing conditions (g100 °C) with Cu(I) or Rh(I).5,6
The success of these preliminary studies and our continuing
interest in heterocycle synthesis from three-membered rings7
led us to consider methods that would effect this transforma-
tion under mild conditions. Initially, we chose to examine
several mechanistically distinct approaches (Scheme 2).
In the event, all three strategies proved adequate for
isomerization to dihydrofuran. Lewis acid catalysis was
observed with several metal salts; Cu(OTf)2 facilitated
complete conversion of cyclopropane 1a to dihydrofuran 2a
at ambient temperature with a catalyst loading of 10 mol %
(4 h).10 Pd(0) and Ni(0) catalysts were also effective, with
the latter typically providing >95% conversion in less than
1 h, while Pd(0) catalysts required several hours to reach
completion. The combination of BINOL/PPh3 provided 2a
but showed incomplete conversion even over extended
reaction times (74% conversion, THF, 25 °C, 17 h). Of the
methods examined, the use of Ni(0) was most efficient in
terms of catalyst loading and reaction time and was selected
for further exploration.
Scheme 2. Strategies for 1-Acyl-2-vinylcyclopropane
Rearrangement (See Text)
The application of the LnNi system to a group of diverse
vinyl cyclopropyl ketones resulted in an efficient rearrange-
ment protocol (Table 1). Cyclopropanes made from deriva-
tives of alkyl acetoacetates (entries 1-3) underwent smooth
isomerization at ambient temperature. The reactions were
complete in 10 min and required only 2 mol % of Ni(COD)2.
2,2’-Bipyridyl proved the most effective ligand for these
substrates. Low conversion was observed in the absence of
a ligand. As reported previously,6 the substrate containing
the geminal ester moiety (entry 4) does not isomerize to 2
even with a higher catalyst loading and longer reaction time.
Substitution at the keto-arm was successful as well (entries
5-8), with substrates containing alkene, aryl, cyano, and
acetal functional groups undergoing efficient rearrangement.
We were able to reduce catalyst loading to 1 mol %
Ni(COD)2 by changing the supporting ligand from 2,2’-
bipyridyl to triphenylphosphine. This extended the reaction
time to 1 h; however, dihydrofurans 2e-h were all isolated
in >95% yield. We observed incomplete conversion using
(Ph3P)2Ni(COD) for substrates 1-3. The two metal-ligand
systems could be interchanged based on the exact (to date
undefined) requirements of the substrate.
Structural changes to the alkene were also tolerated.
Disubstitution of the vinyl group was tolerated employing 2
mol % (Ph3P)2Ni(COD) (entry 9). While trisubstitution of
the olefin was initially unsuccessful using the (Ph3P)2Ni-
(COD) catalyst system, switching to 2,2’-bipyridyl as the
ligand led to higher yields. It was necessary to increase the
catalyst loading to 5 mol % and the reaction time to 4.5 h
(entry 10); however, this protocol successfully afforded
dihydrofuran 2j in 80% yield.
Under Lewis acid catalysis (A), chelation to the geminal
acyl groups would act to weaken the cyclopropane bond
leading to ring opening.8 In the limiting case, this would
result in an allyl cation/enolate zwitterion that could undergo
internal quenching to give product. Our second option (B)
was to use transition metals capable of triggering π-allyl
chemistry. In this scenario, association of the alkene followed
by π-allyl formation by ring opening of the cyclopropane
would result in an (allyl)metal species with a pendant
enolate.4a,d A third option (C) was use of a Lewis base in
conjunction with a Brønsted acid. Under these conditions
the Brønsted acid would activate the pendant carbonyl in
the same manner as a Lewis acid, then attack of the Lewis
base would open the cyclopropane in a “pull-push” fashion.9
Attempted asymmetric catalysis of this rearrangement
using scalemic catalysts in no case led to product enantio-
enrichment. The retention of chirality through an inversion-
inversion mechanism is the normal reaction mode for
stabilized anions11,12 and we hypothesized that this mecha-
nism was operative.
(3) For example: Zuo, G.; Louie, J. Angew. Chem., Int. Ed. 2004, 43,
2277-2279.
(4) (a) Morizawa, Y.; Oshima, K.; Nozaki, H. Tetrahedron Lett. 1982,
23, 2871-2874. (b) Burgess, K. Tetrahedron Lett. 1985, 26, 3049-3052.
(c) Shimizu, I.; Ohashi, Y.; Tsuji, J. Tetrahedron Lett. 1985, 26, 3825-
3828. (d) Yamamoto, K.; Ishida, T.; Tsuji, J. Chem. Lett. 1987, 1157-
1158. (e) Hirio, K.; Arinaga, Y.; Ogino, T. Chem. Lett. 1992, 2329-2332.
(5) Vardapetyan, A. A.; Khachatryan, D. S.; Panosyan, G. A.; Morlyan,
N. M. J. Org. Chem. USSR 1986, 22, 2034-2037.
(6) Dzhemilev, U. M.; Khusnutdinov, R. I.; Shchadneva, N. A.; Tolstikov,
G. A. Bull. Acad. Sci. USSR, DiV. Chem. Sci. 1990, 39, 1447-1450.
(7) (a) Pohlhaus, P. D.; Bowman, R. K.; Johnson, J. S. J. Am. Chem.
Soc. 2004, 126, 2294-2295. (b) Pohlhaus, P. D.; Johnson, J. S. J. Org.
Chem. 2005, 70, 1057-1059.
(10) Other promoters tested, THF, 25 °C (GC yield, time): TiCl4(THF)2
(21%, 4 h), SnCl2 (<3%, 4 h), AlCl3 (<3%, 4 h), ZnCl2 (<3%, 4 h), MgCl2
(<3%, 4 h), Sn(OTf)2 (55%, 4 h), Mg(OTf)2 (<3%, 4 h), Zn(OTf)2 (<3%,
4 h), La(OTf)3 (24%, 4 h), TiF4 (<3%, 4 h), ZrCl4 (48%, 4 h), DMAP/
BINOL (<3%, 17 h), P(OPh)3/BINOL (<3%, 17 h), PPh3 (19%, 18 h),
quinuclidine (14%, 18 h).
(11) Consiglio, G.; Waymouth, R. M. Chem. ReV. 1989, 89, 257-276.
(12) Hayashi, T.; Yamane, M.; Ohno, A. J. Org. Chem. 1997, 62, 204-
207.
(8) Young, I. S.; Kerr, M. A. Angew. Chem., Int. Ed. 2003, 42, 3023-
3026.
(9) McDougal, N. T.; Schaus, S. E. J. Am. Chem. Soc. 2003, 125, 12094-
12095.
574
Org. Lett., Vol. 8, No. 4, 2006