6080
L. O. Davis, S. L. Tobey / Tetrahedron Letters 51 (2010) 6078–6081
Base
Mg
Ts
O
NHTs
OEt
O
O
NTs
O
N
H
OEt
R
R
R
O
O
EtO
Br
Br
Mg
O
5
7
2
6
R
O
Mg
Br
R= Me
Ts
O
N
R
R
O
1
Br
CO2Et
EtO
Br
TsHN
7
2
4
Scheme 3.
O
NHTs
CO2Et
O
O
NHTs
CO2Et
O
MgBr2 OEt2
Ts
CH2Cl2
K OtBu
MeO
N
EtO
MeO
Br
6c, 6c'
MeO
0ºC
1c
2
5c
TLC confirmation
Scheme 4.
present on the aryl ring (1c), the reaction proceeded with moderate
yield to generate the terminal vinyl compounds 6c and 6c0 and the
anticipated product 5c was not observed (Table 2, entry 3). The
same observation was noted when the weakly electron-withdraw-
ing group Br (1d) was used (Table 2, entry 4). A nitro group (NO2), a
more strongly electron-withdrawing group, was also employed as
a substituent (1g), and products 6g and 6g0 were found in trace
amounts and the corresponding product 5g was not observed (re-
sults not shown). Taken together these results suggest that elec-
tronic factors do not greatly influence the product distribution.
The ketoallene bearing a cyclohexyl group (1e) was used to inves-
tigate the steric effects upon the generation of the corresponding
product 5e, and it was found that only the products 4e, 6e, and
6e0 were generated (Table 2, entry 5). In an additional variant on
the allene substrate, commercially available ethylester ketoallene
(1f) was also tested and we found that it did not undergo a reaction
at all (Table 2, entry 6).
On the basis of the results above and the isolation of products 6,
it was confirmed that 6 was indeed a precursor to the anticipated
products 5. We explored the possibility that an external base may
be able to facilitate the elimination of 6 to give 5. Our initial at-
tempt to use triethylamine to convert 6b0 into the anticipated 5b
was unsuccessful. However when potassium t-butoxide was used,
products 5b and 5c were isolated from 6b0 and 6c0, respectively, in
good yields (Scheme 2; see Supplementary data).21 The results
confirmed the potential to obtain the masked unnatural amino
acids 5 in high yield, as well as allowed the proposal of a mecha-
nism for our aza-MBH reaction.
Notably, the isolation of products 6 (Table 2) substantiates this
proposed mechanism.22
The base that facilitates elimination is potentially the nitrogen
anion that is generated by the addition of the enolate 7 to the
imine 2; however it is unlikely to occur through an intramolecular
deprotonation because that would lead to an energetically unfa-
vorable four-membered ring transition state. Additionally, the
chelation of the magnesium on the pre-eliminated product (6)
interferes with the access to the a-proton. Since an elimination oc-
curs under Lewis acid reaction conditions with the methyl keto-
allene, we suspect for the phenyl-substituted ketoallenes, a rigid
chelation arrangement precludes the elimination from occurring.
In a control experiment, potassium t-butoxide was added directly
to the reaction containing the Lewis acid after using TLC methods
to confirm the presence of the vinyl bromide product 6c0 (Scheme
4). However, the expected allene 5c was not observed.
In summary, we have presented the first halide-initiated
aza-MBH reaction between allenic ketones and glyoxylate-derived
imines to generate unnatural amino acids. The method reported
herein provides a unique way to generate unnatural amino acids
and introduces a conceptually new application of the aza-MBH
reaction. The products can serve as synthons for synthetic pur-
poses and the vinyl bromine analog has the advantage of being
additionally transformed through cross-coupling methods. Efforts
are in progress to fully understand the mechanism of the reaction
and to expand the scope of the reaction to include other Lewis acid
catalysts and a variety of imine substrates.
On the basis of previous literature11 and our results, the mech-
anism shown in Scheme 3 is proposed. It appears plausible that
both the cationic and anionic components of the Lewis acid
MgBr2ꢀOEt2 are participating in the mechanism. As depicted, the
mechanism proceeds through the nucleophilic attack of the bro-
mide ion onto the central carbon atom of the allene resulting in
enolate 7, a common intermediate that can lead either to product
4 or 5 (Scheme 3). Subsequent addition of 7 to the imine occurs
Acknowledgments
We thank Dr. Marcus Wright for his assistance with NMR
experiments as well as Dr. Paul Jones for his helpful discussions.
In addition, we would like to thank Wake Forest University for
funding this project.
Supplementary data
at either the
sition gives rise to 6 which undergoes elimination to provide the
isolated allene 5. Alkylation at the -position leads to product 4.
a- or c-position of the allene. Alkylation at the a-po-
Supplementary data associated with this article can be found, in
c