Aza-[1,2]-Wittig rearrangements of N-benzyl glycine methyl esters. A new approach to the synthesis of N-aryl phenylalanine derivatives

Article abstract of DOI:10.1016/j.tetlet.2015.01.037

Treatment of N-(arylmethyl)-N-aryl glycine methyl ester derivatives with Bu₂BOTf and ⁱPr₂NEt effects an aza-[1,2]-Wittig rearrangement that provides N-aryl phenylalanine methyl esters in good yields. Analogous substrates bearing N-carbonyl groups are converted to 1,4,2-oxazaborole derivatives under similar conditions.

Full text of DOI:10.1016/j.tetlet.2015.01.037

Tetrahedron Letters xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Aza-[1,2]-Wittig rearrangements of N-benzyl glycine methyl esters.
A new approach to the synthesis of N-aryl phenylalanine derivatives
⇑
Renata K. Everett, John P. Wolfe
Department of Chemistry, University of Michigan, 930 N. University Ave., Ann Arbor, MI 48109-1055, USA
a r t i c l e i n f o
a b s t r a c t
i
Article history:
2 2
Treatment of N-(arylmethyl)-N-aryl glycine methyl ester derivatives with Bu BOTf and Pr NEt effects an
Received 3 December 2014
Revised 5 January 2015
Accepted 6 January 2015
Available online xxxx
aza-[1,2]-Wittig rearrangement that provides N-aryl phenylalanine methyl esters in good yields.
Analogous substrates bearing N-carbonyl groups are converted to 1,4,2-oxazaborole derivatives under
similar conditions.
Ó 2015 Elsevier Ltd. All rights reserved.
Keywords:
Boron
Wittig rearrangement
Enolate
Heterocycle
In recent years, our group has reported the successful pairing of
enolate [1,2]-Wittig rearrangements with aldol reactions, which
afford substituted
a,b-dihydroxy esters in good yields and high
1
diastereoselectivity (Scheme 1). Asymmetric variants of these
reactions using 2-phenylcyclohexanol as a chiral auxiliary provide
the diol products in up to 95% ee2 and analogous tandem Wittig
rearrangement/Mannich reactions provide access to the corre-
3
sponding amino alcohols. Given the synthetic utility of these reac-
tions, we sought to further expand the scope of enolate Wittig
rearrangements and we hypothesized that a tertiary amine could
undergo the corresponding enolate aza-[1,2]-Wittig rearrange-
ment under similar conditions. The rearrangement itself would
yield substituted phenylalanine derivatives and if coupled with a
subsequent aldol reaction this sequence could produce biologically
Scheme 1. Tandem Wittig rearrangement/aldol reaction.
examples of aza-[1,2]-Wittig rearrangements of O-methyl-N-ben-
zyl glycine derivatives (via their corresponding boron enolates),
which afford substituted phenylalanine methyl ester products.
Given the paucity of examples of successful aza-[1,2]-Wittig
rearrangements, our initial studies were focused solely on the rear-
rangement (rather than the tandem rearrangement/aldol
sequence). We first elected to examine the rearrangement of
N-benzyl-N-boc-glycine methyl ester, as prior studies on
aza-[2,3]-Wittig rearrangements have shown that electron-
withdrawing N-substituents improved reactivity. However, as
illustrated in Eq. 1, when 6a was subjected to the reaction condi-
tions we have previously employed in [1,2]-Wittig rearrangements
of 1, no rearrangement occurred. Instead, 1,4,2-oxazaborole
interesting b-hydroxy-
However, although [1,2]- and [2,3]-Wittig rearrangements of
alkoxy carbanions have been well documented over the past 70
a-amino acid derivatives.
a-
4
,5
years, the corresponding aza-Wittig rearrangements have been
6
less explored. Moreover, although aza-[2,3]-Wittig rearrange-
6
ments are not uncommon, aza-[1,2]-Wittig rearrangements have
7
rarely been observed. In most instances these have been reported
as side reactions in aza-[2,3]-Wittig rearrangements;7b,c the migra-
tion of benzyl groups in synthetically useful yields (>60%) has only
been reported on a single occasion.8 Herein we describe the first
⇑
Corresponding author. Tel.: +1 734 763 3432; fax: +1 734 615 3790.
E-mail address: jpwolfe@umich.edu (J.P. Wolfe).
9
derivative 8a was formed.
http://dx.doi.org/10.1016/j.tetlet.2015.01.037
040-4039/Ó 2015 Elsevier Ltd. All rights reserved.
0
Please cite this article in press as: Everett, R. K.; Wolfe, J. P. Tetrahedron Lett. (2015), http://dx.doi.org/10.1016/j.tetlet.2015.01.037

2
R. K. Everett, J. P. Wolfe / Tetrahedron Letters xxx (2015) xxx–xxx
Table 2
a
Aza-[1,2]-Wittig rearrangement
ð1Þ
Entry
R
Ar
Yieldb
1
2
3
4
5
6
7
8
9
H (9a)
p-Br (9b)
Ph
Ph
Ph
Ph
64% (10a)
54% (10b)
65% (10c)
24% (10d)
67% (10e)
53% (10f)
68% (10g)
66% (10h)
54% (10i)
A brief survey of the reactivity of other glycine methyl ester
derivatives bearing N-pivaloyl or N-acetyl groups gave similar
results (Table 1), as did the use of 9-BBNOTf in place of Bu BOTf
Table 1, entry 4). The structure of 8d was confirmed via X-ray
crystallography, which clearly revealed the presence of a boron–
oxygen bond. Unfortunately, efforts to force these compounds to
undergo the desired rearrangement by increasing temperature
and reaction time were unsuccessful. Moreover these compounds
were also unreactive towards acids, bases and oxidants such as
alkaline hydrogen peroxide.
p-OMe (9c)
p-CF (9d)
3
H (9e)
H (9f)
H (9g)
H (9h)
H (9i)
p-BrPh
o-BrPh
2-Furyl
2-Thiophenyl
N-Ts-2-pyrrolyl
2
(
a
i
Conditions: (i) 1.0 equiv of 9, 3.2 equiv of Bu
.25 M, 0–40 °C. (ii) H , pH 7 buffer, MeOH.
Isolated yield, average of two or more experiments.
2 2 2 2
BOTf, 4.0 equiv of Pr NEt, CH Cl ,
0
2 2
O
b
Given the lack of rearrangement of substrates 6, we elected to
modify the glycine N-substituent in hopes that changing the steric
or electronic properties of the nitrogen atom would facilitate the
desired transformation. Efforts to effect rearrangements of N-ben-
zyl glycine methyl esters bearing N-alkyl, N-phosphoryl and
N-tosyl groups were also unsuccessful. However, we were gratified
to find that a substrate bearing an N-phenyl group did undergo
diastereoselectivity of this reaction was low (Eq. 2). Efforts to
improve the diastereoselectivity through use of 9-BBN-OTF in
place of Bu BOTf were unsuccessful; a similar mixture of diastereo-
2
mers was obtained under these conditions. Substrate 13 proved to
be unreactive under our standard conditions (Eq. 3), which may be
due to difficulty in generating the requisite boron enolate from the
more sterically encumbered substrate. This hypothesis is consis-
[
1,2] rearrangement. Further experimentation revealed that by
extending reaction time and increasing temperature, the desired
tent with our observation that treatment of 9a with excess Bu2-
amine could be isolated in 65% yield (Table 2, entry 1).10
i
BOTf/ Pr
2
NEt followed by addition of an aldehyde to the reaction
With the optimized reaction conditions in hand, additional sub-
strates were examined in order to determine the scope of the
transformation. The requisite substrates were prepared in three
steps from substituted anilines and benzaldehyde derivatives via
mixture (Eq. 4) did not afford an aldol product as was previously
observed in reactions of 1 (Scheme 1).
ð2Þ
ð3Þ
imine formation, reduction and N-alkylation with a-bromo methyl
acetate. As shown in Table 2, substitution on both the N-aryl group
and benzyl group was tolerated. Additionally, substrates bearing
heteroaromatic groups underwent the [1,2] rearrangement in good
yields (Table 2, entries 7–9).11 However, the rearrangement of 9d,
which contains an N-p-trifluoromethylphenyl group proceeded in
poor yield due to a combination of slow reaction rate and product
decomposition as a result of the extended reaction time.
To further explore reaction scope and elements of stereocontrol
we prepared substrates 13 and 11, which bear a methyl group
adjacent to the ester or at the benzylic position, respectively.
Substitution at the benzylic position was tolerated as the
rearrangement of 11–12 proceeded in good yield. However, the
ð4Þ
Finally, we have conducted preliminary studies on asymmetric
aza-[1,2]-Wittig rearrangements. We previously have illustrated
that 2-phenylcyclohexanol provides good results in asymmetric
Wittig rearrangement/aldol reactions of glycolate esters. As such,
substrate 14 bearing this chiral auxiliary was synthesized and sub-
jected to the standard reaction conditions. Unfortunately, although
the yield of this transformation was good, the diastereoselectivity
was modest (Eq. 5). Nonetheless, this experiment indicates the
possibility of achieving asymmetric induction, although further
optimization is clearly needed.
Table 1
a
Formation of 1,4,2-Oxazaborole derivatives
Entry
R
1
R
2
BOTf
Yieldb
ð5Þ
t
1
2
3
4
O Bu (6a)
Bu
Bu
Bu
2
2
2
BOTf
BOTf
BOTf
40% (8a)
87% (8b)
48% (8c)
90% (8d)
t
Bu (6b)
CH
3
(6c)
t
Bu (6b)
9-BBNOTf
In conclusion, we have developed a new aza-[1,2]-Wittig rear-
rangement of N-aryl-N-benzyl glycine methyl esters. These
transformations constitute rare examples of benzyl group migra-
tion in aza-Wittig rearrangements and provide a concise four-step
a
i
Conditions: (i) 1.0 equiv of 6, 3.2 equiv of R
.25 M, 15 min, 0 °C to rt. (ii) H , pH 7 buffer, MeOH, 1 h, 0 °C to rt.
Isolated yield, average of two or more experiments.
2 2 2 2
BOTf, 4.0 equiv of Pr NEt, CH Cl ,
0
2 2
O
b
Please cite this article in press as: Everett, R. K.; Wolfe, J. P. Tetrahedron Lett. (2015), http://dx.doi.org/10.1016/j.tetlet.2015.01.037

R. K. Everett, J. P. Wolfe / Tetrahedron Letters xxx (2015) xxx–xxx
3
approach to the construction of substituted N-arylphenylalanine
derivatives. In contrast to prior examples of aza-[1,2-]-Wittig rear-
rangements, these reactions do not require the use of the strong
reducing agents or strong bases and proceed under relatively mild
conditions. Future studies will be directed towards improvements
and new synthetic applications of this method.
5. For selected reviews, see: Tomooka, K.; Yamamoto, H.; Nakai, T. Liebigs Ann.
Recueil 1997, 1275–1281; (b) Nakai, T.; Mikami, K. Chem. Rev. 1986, 86, 885.
6.
For a review on the aza-Wittig rearrangement, see: Vogel, C. Synthesis 1997, 5,
97.
4
7. (a) Coldham, I. J. Chem. Soc., Perkin Trans. 1 1993, 1275; (b) Tomoyasu, T.;
Tomooka, K. Synlett 2004, 1925–1928; (c) Gawley, R. E.; Zhang, Q.; Campagna,
S. J. Am. Chem. Soc. 1995, 117, 11817; (d) Murata, Y.; Nakai, T. Chem. Lett. 1990,
2
069.
Katritzky, A. R.; Luo, Z.; Fang, Y.; Feng, D.; Ghiviriga, I. J. Chem. Soc., Perkin Trans.
2000, 1375.
For other examples of related 1,4,2-oxazaborole derivatives see: (a) Ishikura,
M.; Kato, H. Tetrahedron 2002, 58, 9827; (b) Pelter, A.; Smith, K.; Hutchings, M.
G.; Rowe, K. J. Chem. Soc., Perkin Trans. 1 1975, 129; (c) Pelter, A.; Hutchings, M.
G.; Rowe, K.; Smith, K. J. Chem. Soc., Perkin Trans. 1 1975, 138; (d) Denniel, V.;
Bauchat, P.; Toupet, L.; Carboni, B.; Danion, D.; Danion-Bougot, R. Tetrahedron
Lett. 1995, 36, 3507; (e) Denniel, V.; Bauchat, P.; Carboni, B.; Danion, D.;
Danion-Bougot, R. Tetrahedron Lett. 1995, 36, 6875.
8
9
.
.
2
Acknowledgment
The authors thank the NSF (CHE 1111218) for financial support
of this work.
Supplementary data
10. When a single equivalent of dibutylboron triflate was used the rearrangement
reaction proceeded to 74% conversion (26% unreacted starting material) as
Supplementary data associated with this article (experimental
procedures, characterization data and 1H and C NMR spectra
for all new compounds, and crystallographic structural data for
compound 8d), in the online version, at http://dx.doi.org/10.
judged by 1
H NMR analysis. This experiment suggests that the reaction
13
proceeds through a boron enolate that contains a single equivalent of an
organoboron species, rather than an intermediate that bears two dialkylboron
units (one bound to oxygen and one bound to nitrogen). We have elected to
classify this as a Wittig-type rearrangement as it does not appear that an
ammonium ion resulting from protonation of the amino group (as in a Stevens-
type rearrangement) is required. However, there likely is a Lewis acid/base
1
016/j.tetlet.2015.01.037.
2
chelate interaction between the O-BR group and the amino group, so in this
case there may not be a clean division between classification as a Wittig
rearrangement versus a Stevens rearrangement.
References and notes
1
2
3
4
.
.
.
.
Bertrand, M. B.; Wolfe, J. P. Org. Lett. 2006, 8, 4661.
11. We were unable to examine the reactivity of the corresponding NH or
N-alkylpyrrole derivatives as these compounds were unstable and rapidly
decomposed upon isolation.
Giampietro, N. C.; Kampf, J. W.; Wolfe, J. P. J. Am. Chem. Soc. 2009, 131, 12556.
Giampietro, N. C.; Wolfe, J. P. Angew. Chem., Int. Ed. 2010, 49, 2922.
Wittig, G.; Lohmann, L. Liebigs Ann. 1942, 550, 260.
Please cite this article in press as: Everett, R. K.; Wolfe, J. P. Tetrahedron Lett. (2015), http://dx.doi.org/10.1016/j.tetlet.2015.01.037