Enantioselective Ag-catalyzed allylation of aldimines

Article abstract of DOI:10.1002/ejoc.200900669

A highly enantioselective synthesis of homoallylic amines, using allyltrimethoxysilane under Ag¹ catalytic conditions, has been developed. Among the chiral ligands investigated, a remarkable difference in the resulting Ag¹ complexes

Full text of DOI:10.1002/ejoc.200900669

SHORT COMMUNICATION
DOI: 10.1002/ejoc.200900669
Enantioselective Ag-Catalyzed Allylation of Aldimines
Marina Naodovic,[a] Manabu Wadamoto,[a] and Hisashi Yamamoto*[a]
Dedicated to Professor Alain Krief
Keywords: Aldimines / Silver / Allylsilane / Asymmetric synthesis
A highly enantioselective synthesis of homoallylic amines,
using allyltrimethoxysilane under AgI catalytic conditions,
has been developed. Among the chiral ligands investigated,
a remarkable difference in the resulting AgI complexes was
observed. Under mild conditions and low catalyst loadings,
homoallylamines were produced in high ee values (up to
80%) and good yields. The methodology can be further ex-
tended to a diastere- and enantioselective crotylation of ald-
imines.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2009)
hydes.[5] Recently we have demonstrated that these condi-
tions can be successfully applied to a highly diastereo- and
enantioselective allylation of ketones.[6] Encouraged by the
excellent reactivity of allyltrimethoxysilane-AgI-phosphane
system for the asymmetric allylation reaction of aldehydes
and ketones, we have focused our attention to their aza-
analogues. In contrast to carbonyl electrophiles, imines rep-
resent more challenging substrates since Lewis acid cata-
lyzed reactions with imine electrophiles usually suffer from
inhibition of the catalyst by the amine product. The devel-
opment of an efficient protocol for asymmetric allylation of
imines would therefore be highly a beneficial addition to
the limited number of published methodologies.[2a]
For our initial studies we have adopted the reaction con-
ditions described for the allylation reaction of ketones.[6] As
in the case of carbonyls, the reaction with aldimines re-
quired the presence of a proton source to liberate the amine
product. Imines derived from electron-rich amines were
shown to be more desirable substrates for the reaction since
electron-poor imines provided trace or no product. Solvent
screening revealed THF to be the best choice of solvent
providing the products in highest yields and enantio-
selectivities. It should be noted that in the early stage of our
studies crotylsilane was used, fearing that simple allytrime-
thoxylsilane would be insufficiently nucleophilic reagent.
One of the remarkable observations made during the initial
investigation of the reaction conditions was regarding the
outstanding difference of the reactivity of monophosphane-
and diphosphane-AgI catalysts (Scheme 1). Namely, when
the reaction was carried out in the presence of monophos-
phane ligands, the products were obtained in quantitative
yields. Diphosphane ligands, on the other hand, appear to
form inactive complexes.
Introduction
Nitrogen functionalities represent a crucial architectural
detail in many biologically active natural products and
amino acids. The most efficient and convenient synthetic
methods for preparation of these groups of compounds are
based on the utility of imines as electrophilic reagents.[1]
Many of these methods use various organometallic nucleo-
philes, some of which are highly toxic (such as organostann-
anes) or are nonselective (regio- and/or stereoselective).[2]
Even though several reports of asymmetric allylation of ald-
imines and ketimines are present in the literature, the exam-
ples of practical and efficient methods are scarce.
Allylic trialkoxysilanes have long been recognized as
highly reactive species, upon formation of pentacoordinate
intermediates in the presence of nucleophiles.[3] Beside their
excellent reactivity, another advantage of allylic trialkoxysil-
anes is low toxicity, as opposed to allylic stannanes which
are other commonly used allyl-transfer reagents. During
our ongoing research program regarding Ag-catalyzed reac-
tions, we have demonstrated that allyltrimethoxysilane acts
as a highly reactive allyl-transfer reagent under AgI-phos-
phane catalytic conditions.[4] Herein we would like to report
the development of enantioselective allylation reaction of
aldimines using trimethoxyallylsilanes in the presence of
AgI-phosphane ligand.
Results and Discussion
Our group reported the first example of asymmetric sil-
ver-catalyzed addition of allyltrimethoxysilanes to alde-
[a] Department of Chemistry, The University of Chicago,
5735 South Ellis Avenue, Chicago, Illinois 60637, USA
Fax: +1-773-702-5059
E-mail: yamamoto@uchicago.edu
Eur. J. Org. Chem. 2009, 5129–5131
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5129

M. Naodovic, M. Wadamoto, H. Yamamoto
SHORT COMMUNICATION
phosphane ligands, however the ligand/AgF ratio had to be
increased to 1:2 [Scheme 1, (R)-Difluorophos:AgF ratio is
1:1].[7] Unfortunately, the yields of homoallylamines were
generally low and in some cases indicated that the catalyst
was inhibited by the amine product (Table 2). Despite the
low reactivity generally observed in Table 2, diflurophos-
AgF catalyst emerged as a promising lead for improvement
(entry 5, Table 2).
Table 2. Enantioselective allylation of aldimine using AgF-diphos-
phane catalysts.[a]
Scheme 1. Crotylation of aldimines using different Ag-phosphane
catalysts.
In our previous research, we have observed formation
of three different diphosphane-Ag complexes which exhibit
remarkably different reactivity and selectivity. Ratio of
these complexes can be regulated by the nature of silver salt
or varying the ratio of diphosphane ligand and silver salt.[7]
We believe that using (R)-Diflurophos and AgF in 1:1 ratio
generates predominantly inactive complex (Scheme 1).
Encouraged by the high reactivity of AgI-monophos-
phane catalyst, several monophosphane and MOP-type li-
gands were prepared and screened for the allylation reac-
tion. Some of the results are shown in Table 1. Although
many of the ligands exhibited high levels of reactivity, the
enantioselectivities observed were unsatisfactory.
[a] Reactions were carried out using 1.0 equiv. of aldimine,
2.0 equiv. of allyltrimethoxysilane and 2.5 mol-% of (R)-Difluo-
rophos or 5 mol-% of (R)-Mop at –78 °C to –20 °C for 24 h. [b]
Isolated products yields. [c] Enantiomeric excess was determined
by HPLC analysis (OD-H, Chiralec column).
Table 1. Asymmetric crotylation of aldimines using Ag-monophos-
phane catalysts.[a]
With these new results, modifications of the reaction con-
ditions were performed to affect the reactivity of the Ag-
diphosphane catalyst. Various additives (alcohols and
amines) used instead of MeOH, and different solvents or
AgI-salts provided no improvement in the reactivity and
caused a decrease of the enantioselectivity and/or reactivity.
On the other hand, modification of the aldimine substrates
Table 3. Enantioselective allylation of aldimines using AgF-phos-
phane catalysts.[a]
[a] The reactions were carried out using 1.0 equiv. of 4 and
2.0 equiv. of 2 at –78 °C for 18 h. [b] Isolated products yield. [c]
Diastereomeric ratio determined by 1H NMR spectroscopy. [d] En-
antiomeric excess determined by HPLC analysis (OD-H Chiracel
column).
[a] Reactions were carried out using 1.0 equiv. of aldimine,
2.0 equiv. of allyltrimethoxysilane and 2.5 mol-% of (R)-Difluo-
rophos or 5 mol-% of (R)-Mop at –78 °C to –20 °C for 24 h. [b]
Isolated products yields. [c] Enantiomeric excess was determined
Contrary to our expectations of poor reactivity of simple
allyltrimethoxysilane, we were pleased to find it is suffi-
ciently reactive under the reaction conditions. Furthermore,
the allylation reaction does proceed in the presence of di- by HPLC analysis (OD-H, Chiralec column).
5130 Eur. J. Org. Chem. 2009, 5129–5131
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Enantioselective Ag-Catalyzed Allylation of Aldimines
crude reaction mixture was purified by column chromatography
(hexanes/EtOAc).
had a more drastic effect on the performance of the cata-
lytic system. The imine substrate bearing 2-PhOC6H4 group
yielded a homoallylamine in synthetically useful yield and
selectivity. With these optimized conditions in hand, enan-
tioselective synthesis of homoallylamines using a wide
range of imine substrates can be carried out (Table 3).
Acknowledgments
We thank National Institutes of Health (NIH) and National Sci-
ence Foundation (NSF) for the financial support.
Conclusions
[1] a) E. F. Kleinman, in: Comprehensive Organic Synthesis (Eds.:
B. M. Trost, I. Fleming), Pergamon Press, Oxford, 1991, vol.
2, pp. 893; b) Comprehensive Organic Synthesis (Eds.: B. M.
Trost, I. Fleming), Pergamon Press, Oxford, 1991, vol 8, pp. 1;
c) For a recent review of enantioselective reactions with imines,
see: S. Kobayashi, H. Ishitani, Chem. Rev. 1999, 99, 1069.
[2] a) H. Ding, G. F. Friestad, Synthesis 2005, 2815; b) R. Bloch,
Chem. Rev. 1998, 98, 1407; c) D. Enders, U. Reinhold, Tetrahe-
dron: Asymmetry 1997, 8, 1895; d) S. E. Denmark, J. C. Nica-
ise, Chem. Commun. 1996, 999.
In summary, new method for the enantioselective synthe-
sis of homoallylamines has been developed. Using simple
allyltrimethoxysilane in combination with commercially
available ligands and AgF, products can be obtained in syn-
thetically useful yields and high enantioselectivity. We have
also shown that the methodology can be efficiently ex-
panded to crotylsilane substrate to afford products in a
highly diastereoselective fashion. Another advantageous
feature of this protocol is the excellent level of regioselecti-
vity observed.
[3] a) K. Sato, M. Kira, H. Sakurai, J. Am. Chem. Soc. 1989, 111,
6429; b) G. Cerveau, C. Chuit, R. J. P. Coriu, C. Reye, J. Or-
ganomet. Chem. 1987, 328, C17; c) A. Hosomi, S. Kohra, Y.
Tominga, J. Chem. Soc., Chem. Commun. 1987, 1517.
[4] a) M. Wadamoto, H. Yamamoto, Chem. Asian J. 2007, 2, 692;
b) M. Naodovic, H. Yamamoto, Chem. Rev. 2008, 108, 3132.
[5] A. Yanangisawa, H. Kageyama, Y. Nakatsuka, K. Asakawa,
Y. Matsumoto, Angew. Chem. 1999, 111, 3916; Angew. Chem.
Int. Ed. 1999, 38, 3701.
Experimental Section
General Experimental Procedure for Allylation of Aldimines: A
flame-dried test tube was charged with AgF (3.1 mg, 0.024 mmol)
and (R)-Difluorophos (8.2 mg, 0.012 mmol). To this mixture was
added dry THF (2.5 mL) and MeOH (0.24 mmol) and the resulting
mixture was stirred for 15 min. After 15 min, THF and MeOH
were removed in vacuo. After all the volatiles were removed, the
solid residue was dissolved in THF (2.5 mL) and cooled to –78 °C.
After the reaction mixture was cooled, aldimine (0.24 mmol) and
allyltrimethoxysilane (0.48 mmol) were added. Upon reaction com-
pletion, as determined by TLC analysis, the reaction mixture was
filtered through a short silica plug and concetrated in vacuo. The
[6] M. Wadamoto, H. Yamamoto, J. Am. Chem. Soc. 2005, 127,
14556.
[7] We have recently reported that depending on the ligand/AgX
ratio, complexes with remarkably different reactivity can be
formed. Distribution of reactive and unreactive complexes can
be affected by the ration and the nature of the counteranion.
See: N. Momiyama, H. Yamamoto, J. Am. Chem. Soc. 2004,
126, 5360.
Received: June 17, 2009
Published Online: August 26, 2009
Eur. J. Org. Chem. 2009, 5129–5131
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5131