M. L. Rosenberg et al. / Tetrahedron Letters 50 (2009) 6506–6508
6507
O
O
O
O
d
O
O
O
O
O
OEt
OEt
OEt
HO
Br
X
X
Br
X
Br
a
b
c
e
N
R
N
R
N
H
N
N
N
R
R
R
10
14
18
R=H, X= N2
11 R=Bz, X=N2
12
R=H, X= N2
15 R= Bz, X=N2
16
R=H, X= N2
19 R=Bz, X= N2
7
R=Bz
5
6
R= Bz
R= Tos
4
8 R= Tos
9 R= H
20
H2
R=Tos, X = H 2
13 R=Tos, X=N 2
R=Tos, X = H 2
17 R=Tos, X=N 2
R=Tos, X =
21 R=Tos, X=N 2
22
R=H, X=H2
Scheme 2. Reagents and conditions: (a) 5: BzCl, Et3N, DMAP, 84%; 6: NaH, TosCl, 95%; (b) 7: CuBr2, 77%; 8: CuBr2, 81%; 9 from 7: HCl, MeOH, 90%; (c) 10: ethyl
2-diazoacetoacetate, Et3N, TiCl4, Ti(Oi-Pr)4, À20 °C, 73%; 11: ethyl acetoacetate, NaH, n-BuLi, Ti(Oi-Pr)4, À78 °C, 81%; 12: ethyl 2-diazoacetoacetate, Et3N, TiCl4, Ti(Oi-Pr)4,
À20 °C; (d) 14: cat. TsOH; and (e) LiBr, DMF, 85 °C.
conditions. Exposing N-tosyl-protected ketone 6 to the bromina-
tion conditions with CuBr2 gave bromo ketone 8 in 81% yield.
An aldol reaction between bromo ketone 8 and ethyl 2-diazo-ace-
toacetate gave alcohols 13. Once again, the isolated yield of prod-
ucts from the aldol reaction was very dependent on the reaction
temperature and equivalents of diazo ester used. At room temper-
ature, only a complex product mixture was obtained. The isolated
yields of alcohols 13 increased from 24% with 1 equiv to 58% when
5 equiv of the diazoester was used. We only detected one of the
two possible diastereomers of alcohol 13 each time the reaction
took place. However, which specific diastereomer that formed
seemed to change with reaction conditions, and we were unable
to control this transformation. The modest 58% isolated yield of
the aldol reaction products turned out to be due to a partial dehy-
dration of diazo alcohols 13 under the reaction conditions to give
compounds 17. The rate of dehydration was much slower com-
pared to alcohols 10 and 11, such that alcohols 13 could be isolated
after silica gel flash chromatography. The overall combined yield
for the two transformations was 91%. The E and Z isomeric alkenes
17 could be isolated and characterized. Both diastereomers of alco-
hols 13 gave only one isomer of 17 when the dehydration took
place during the aldol reaction. However, if either of the diastereo-
mers of alcohols 13 was isolated, they both eliminated water
slowly to give what appeared to be the thermodynamically more
stable isomer of 17. The diastereomer formed under the reaction
conditions was the least thermodynamically stable of the two iso-
mers as it slowly rearranged to the other isomer upon standing. It
was difficult to assign the structure of the two isomers of 17 based
on NMR data, thus we performed quantum theory calculations on
the two isomers (Fig. 1). The isomer labeled B with Z-geometry was
predicted to be the thermodynamically more stable isomer by
2.7 kcal/mol. The reaction of 17 with LiBr-promoted aromatization
to give diazo compound 21 in 64% yield, independent of the alkene
stereochemistry in compound 17. We attempted an intramolecular
C–H insertion reaction on compound 21 to obtain the tosyl-pro-
tected analogue of the tricyclic building block 1. Unfortunately,
all attempts failed to give the ring closure at the indole 3-position
in compound 21. It became apparent that it was necessary to re-
move the tosyl-protecting group before the critical ring closure
step, and we therefore focused on procedures for detosylation
reactions. Treating compound 21 with a base did not yield the de-
sired diazo compound 18, but induced facile pyrazole formation.
Pyrazole 23 was formed as the major product in all our deprotec-
tion experiments independent of the base and reaction tempera-
ture. Compound 23 was obtained almost quantitatively within
five minutes even when 21 was exposed to a mild base such as
TBAF. Formation of pyrazoles is a well known transformation from
vinyl diazoacetates,5 but has not been reported to be a major path-
way with keto diazo esters.6 A plausible reaction mechanism is
outlined in Scheme 3.
Abstraction of an a-hydrogen in compound 21 produced an eno-
late which underwent an intramolecular 1,3-dipolar cycloaddition
reaction.5,7 The resulting five-membered ring rearranged through a
1,3-hydrogen shift led to the completion of the pyrazole ring and
the formation of 23. Detosylation of 21 using reducing reagents
was also attempted, but the diazo functionality was incompatible
with the reaction conditions and gave compound 20 as the major
product. The instability of the diazo functionality in some of the
intermediates, and the extremely facile base-induced pyrazole for-
mation forced us to go back and introduce the diazo functional group
as late as possible in the synthetic sequence.
The ethyl acetoacetate side chain was introduced using dianion
chemistry.3 The dianion of ethyl acetoacetate was added to a
À78 °C solution of bromide 8 and Ti(Oi-Pr)4 in THF. The reaction
wasrathersluggishand itwasnecessaryto use5 equivof thedianion
to achievean 81%yieldof 12. Itwas foundcrucialto keepthereaction
temperature low. A complex product mixture was obtained when
the reaction was performed at 0 °C. Elimination of water from
compound 12 was achieved in high yield with a catalytic amount
of p-TsOH. The elimination products 16 were obtained as an insepa-
rable mixture of E and Z isomers. The stereochemistry, however, was
lost in the next step with LiBr-promoted aromatization of 16 to
N+
CO2Et
O
O
O
O
CO2Et
OH
CO2Et
O-
N
N
HN
N
-
N
OEt
EtO
OH
N2
N
Br
N2
Br
1,3-H
shift
N
Tos
N
N
N
Tos
Tos
Tos
Tos
A: 2.7 kcal/mol
B: 0.0 kcal/mol
23
Figure 1. Calculated energies for the E and Z isomers of 17.
Scheme 3.