Catalytic silicon-mediated carbon-carbon bond-forming reactions of unactivated amides

Article abstract of DOI:10.1021/ja108764d

In the presence of catalytic amounts of trialkylsilyl triflate and triethylamine, unactivated amides react with imines to afford the corresponding Mannich-type adducts in high yields with high anti selectivities. While silicon enolates have been widely us

Full text of DOI:10.1021/ja108764d

COMMUNICATION
pubs.acs.org/JACS
Catalytic Silicon-Mediated Carbon-Carbon Bond-Forming Reactions
of Unactivated Amides
Shu Kobayashi,* Hiroshi Kiyohara, and Miyuki Yamaguchi
h
Department of Chemistry, School of Science, and Graduate School of Pharmaceutical Sciences, The University of Tokyo,
and The HFRE Division, ERATO, Japan Science Technology Agency (JST), Hongo, Bunkyo-ku, Tokyo 113-0033, Japan
S Supporting Information
b
Scheme 1. (a) Conventional Mannich Reaction of Silicon
Enolates (Stoichiometric) and (b) Ideal Catalytic Silicon
Cycle
ABSTRACT: In the presence of catalytic amounts of
trialkylsilyl triflate and triethylamine, unactivated amides
react with imines to afford the corresponding Mannich-type
adducts in high yields with high anti selectivities. While
silicon enolates have been widely used in organic synthesis
for four decades, this is the first example of the catalytic use
of the silicon species, to the best of our knowledge. More-
over, it is noteworthy that unactivated simple amides bear-
ing R-protons that are less acidic than those of ketones and
aldehydes can be successfully used in catalytic direct-type
addition reactions. Finally, a preliminary trial of an asym-
metric catalytic version was conducted and showed promis-
ing enantioselectivity of the desired product.
presence of catalytic amounts of trimethylsilyl triflate (Me3SiOTf)
and pyridine as catalysts (Table 1).
ilicon enolates are among the most useful enolates in modern
When a p-methoxyphenyl (PMP) iminoester was employed,
the desired adduct was obtained in 31% yield, which was almost
the same as the catalyst loading (30 mol %, entry 1). This result
may indicate that the product from acetophenone and the PMP
iminoester did not release the Si species and thus that catalyst
turnover did not occur. A benzyloxycarbonyl (Cbz) iminoester
was not productive under these conditions (entry 2), where the
imine was assumed to decompose. Interestingly, a p-toluenesul-
fonyl (Ts) iminoester afforded the desired adduct in high yield
(84%) in comparison with the catalyst loading (30 mol %) (entry 3).
Involvement of the active silicon species is essential; pyridine
or pyridinium triflate did not catalyze the desired Mannich
reaction at all (entries 4 and 5). We further examined other
nucleophiles, such as S-ethyl thioacetate, ethyl acetate, N-methy-
lacetanilide, and N,N-dimethylacetamide. Unexpectedly, amides
were found to be productive substrates (entries 8 and 9), although
more acidic thioesters and esters were not (entries 6 and 7).5
This is inconsistent with the general reactivity tendency of carbonyl
compounds, where more acidic carbonyl compounds can
provide nucleophilic species (enolates) more easily. Such a
mode of catalytic activation of amides is not only mechanisti-
cally novel but also synthetically valuable.6 Encouraged by these
results, we further optimized the reaction conditions, and it was
found that the imine derived from benzaldehyde also reacted
with the amide using the catalyst system (entry 10) and that
S
organic chemistry.1 A variety of silicon enolates have been
reported to date,2 and they serve as convenient carbonyl
equivalent donors through nucleophilic additions to aldehydes,
imines, R,β-unsaturated carbonyl compounds, etc., in the pre-
sence of a substoichiometric amount of a Lewis acid catalyst
(i.e., Mukaiyama-type reactions).3 Asymmetric variants of
these reactions have also been intensively studied over the past
two decades,3b,3c,3f,3i and silicon enolates have played a vital role
in the development of many useful chiral Lewis acids.
Despite their versatile utility, there are some drawbacks in
silicon enolate chemistry, mainly from the viewpoint of atom
economy. While most silicon enolates are isolable and storable,
they are prepared from the corresponding carbonyl compounds
using stoichiometric amounts of bases such as lithium diisopro-
pylamide (LDA) or tertiary amines. Moreover, adducts of silicon
enolates have Si-heteroatom bonds, which are cleaved in many
cases by treatment with an acid or a fluoride anion to provide the
desired products (Scheme 1a). If catalytic generation of silicon
enolates could be attained, these drawbacks would be overcome.
An ideal catalytic cycle including in situ generation of silicon
enolates using a substoichiometric amount of a silicon Lewis acid
and a base4 is shown in Scheme 1b. However, this catalytic cycle
has long been thought to be unlikely because silicon species
might be incorporated into products through the formation of a
covalent bond between silicon and a heteroatom of the products.
We began with an investigation of the catalytic silicon system
in Mannich-type reactions, which provide potentially useful building
blocks for nitrogen-containing compounds. First, we chose aceto-
phenone as a nucleophile and treated it with an imino ester in the
Received: September 29, 2010
Published: December 20, 2010
r
2010 American Chemical Society
708
dx.doi.org/10.1021/ja108764d J. Am. Chem. Soc. 2011, 133, 708–711
|

Journal of the American Chemical Society
COMMUNICATION
Table 1. Investigation of Catalyst Turnover and Optimiza-
tion of Reaction Conditions
Table 2. Substrate Scope of Catalytic Direct-Type Addition
of Amides to Imines
entry
R1
R2
X
Lewis acid
t (h)
yield (%)
1
CO2Et
CO2Et
CO2Et
CO2Et
CO2Et
CO2Et
CO2Et
CO2Et
CO2Et
Ph
PMP
Cbz
Ts
Ph
Ph
Ph
Ph
Ph
Me3SiOTf
Me3SiOTf
Me3SiOTf
-
30
30
2
31
trace
84
0
entry
R1
amide x temp. (°C) time (h) yield (%) anti/syn
2
3
1
2a
3
4
5
6
7
8
9
Ph
Ph
Ph
1a
1a
1b
1a
2
rt
rt
rt
rt
rt
rt
rt
rt
rt
rt
rt
rt
rt
60
rt
15
0
24
12
12
12
12
12
12
12
12
12
12
12
16
16
16
24
24
24
48
quant
quant
quant
quant
92
-
-
4
Ts
48
48
24
24
24
24
24
12
5
5
5
5
5
5
5
5
5
5
5
5
Ts
TfOH
0
-
6a,b
7a,b
8a,b
9a,b
10a,c
11a,d
Ts
SEt
Me3SiOTf
Me3SiOTf
Me3SiOTf
Me3SiOTf
Me3SiOTf
iPr3SiOTf
0
4-ClC6H4
-
Ts
OEt
0
4-MeOC6H4 1a
-
Ts
NMePh
NMe2
NMe2
NMe2
77
71
51
quant
3-NO2C6H4
1-Nap
1a
1a
1a
1a
quant
98
-
Ts
-
Ts
2-Nap
quant
97
-
Ph
Ts
2-thienyl
-
a NEt3 was used instead of pyridine. b In THF/CH2Cl2. c Using 20 mol
% Me3SiOTf and NEt3 without solvent. d Using 5 mol % iPr3SiOTf and
NEt3 in toluene.
10 (E)-cinnamyl 1a
81
-
11 tBu
12 cyclopropyl
13 CO2Et
14 Ph
15b Ph
16b,c Ph
17b,c Ph
18b,c Ph
19b,c Ph
1a
1a
quant
quant
quant
0
-
-
1c 10
1c 10
1c 10
1c 10
1d 10
1e 10
1f 10
1.5/1
-
a bulkier silicon Lewis acid showed high catalytic activity
(entry 11).
quant
81
6/1
27/1
22/1
16/1
4.2/1
Under the optimal conditions (iPr3SiOTf, NEt3 in toluene),
the scope of imines was investigated (Table 2). The desired
products were obtained in high yields using a wide range
of imines, including aromatic, R,β-unsaturated, and aliphatic
imines (entries 1-12).7 It should be noted that simple amides
aswell asa Weinreb amide reacted smoothly (entry3). Additionof
R-alkyl-substituted amides led to the construction of two stereo-
genic centers, which constitute the basic structures of a variety of
biologically active compounds. When N-methylpyrrolidinone
(NMP) was employed as a nucleophile, the imine derived from
benzaldehyde did not react at all, even at 60 °C (entry 14);
however, the imino ester reacted quantitatively (entry 13). These
results indicate that (1) formation of an active silicon species
occurred even when NMP is employed as a nucleophile and (2)
it is presumably because of poor electrophilicity that the imine
derived from benzaldehyde did not afford the desired adduct. A
series of Lewis acids were then screened as cocatalysts to
activate the imines, and finally it was found that CuOTf
emerged as the most active cocatalyst in Mannich-type reac-
tions (entry 15).8 After optimization of the reaction condi-
tions, 2,4,6-mesitylenesulfonyl imine (2) and dichloromethane
proved to be the electrophile and solvent of choice, respec-
tively, providing the Mannich-type adduct with high anti
selectivity (anti/syn = 27/1) (entry 16). Other cyclic amides
(1d and 1e) as well as an acyclic amide (1f) also reacted
smoothly under the reaction conditions to afford the desired
products in high yields with high anti selectivities in most cases
(entries 17-19).
quant
30
15
0
74
a 2-Nitrobenzenesulfonyl was used instead of Ts. b CuOTf (5 mol %)
was used. c 2,4,6-Mesitylenesulfonyl imine (2) was used instead of Ts in
CH2Cl2.
aza-Diels-Alder product in moderate yield (eq 1).9 More-
over, this new methodology could also be extended to asym-
metric catalysis (eq 2). Thus, the enantioselective Mannich-type
reaction of 1a with imine 3 proceeded smoothly in the presence
of iPr3SiOTf (10 mol %), NEt3 (10 mol %), and CuOTf (R)-p-
3
Tol-BINAP10 (5 mol %) to afford the desired adduct in
moderate yield with good enantiomeric excess (ee). It should
be noted that in both cases, the active silicon species that were
formed catalytically reacted smoothly.
The catalytic silicon system could be applied to other reac-
tions. In the presence of iPr3SiOTf (10 mol %), NEt3 (10 mol %),
and CuOTf (5 mol %), 4-aminobut-3-en-2-one (1g) reacted
with the Ts imine derived from benzaldehyde to afford
the Mannich-type adduct in high yield. The adduct was
further treated with trifluoroacetic acid to give the formal
We then conducted several experiments to clarify the me-
chanism of the catalytic silicon system. Because we hypothe-
sized the catalytic cycle shown in Scheme 1b, we first tried
709
dx.doi.org/10.1021/ja108764d |J. Am. Chem. Soc. 2011, 133, 708–711

Journal of the American Chemical Society
COMMUNICATION
Scheme 2. Formation of Amide/R3SiOTf Complexes (1)
reaction proceeded smoothly with anti selectivity when NEt3
was added to this system (eq 7).
Scheme 3. Formation of Amide/R3SiOTf Complexes (2)
to detect a silicon enolate. Amides 1a and 1c were treated with
Me3SiOTf and Et3N in CH2Cl2 (or CD2Cl2) or toluene (or
toluene-d8) independently at 0 °C, room temperature, and
60 °C. However, no clear signals showing formation of silicon
enolates were observed by NMR analyses. It was recently
reported that amides 1a and 1c react with Me3SiOTf at
room temperature to form amide/Me3SiOTf complexes.11 We
Through these experiments, it was shown that silicon enolates were
not detected by NMR analyses and that amide/R3SiOTf complexes
such as 4 and 7 were observed as major silicon species. On the other
hand, complex 7 did not react with imine 2 even in the presence of
CuOTf. For carbon-carbon bond formation, deprotonation of 7 is
needed, and indeed, the reaction with imine 2 proceeded in the
presence of NEt3. A possibility is that 7 and NEt3 generated a very
small amount of silicon enolate 6, whichreactedwithimine2to afford
Mannich adduct 8. The high anti selectivity may be explained by
assuming acyclic transition states, which are often observed in
reactions of silicon enolates with imines.14 When a large amount of
silicon enolate exists, another reaction pathway (such as a
silicon-cation-mediated reaction)15 may be dominant (rapid
reaction but low diastereoselectivity). Another possibility is
that a Cu enolate may be formed from 7, NEt3, and CuOTf.
While this possibility cannot be denied at this stage, amides
1a-c reacted with imines without CuOTf, and it may be
difficult to explain the experimental results that the diastereos-
electivities were not dependent on metal triflates.8
In summary, we have developed a catalytic silicon system that
enables catalytic direct-type addition of unactivated amides to
imines. While silicon enolates have been widely used in organic
synthesis for four decades, this is the first example of the catalytic
use of the silicon species, to the best of our knowledge. In addition,
this is also the first successful use of unactivated simple amides in
catalytic direct-type addition reactions. Although several trials to
detect silicon enolates were unsuccessful, the synthetic advantages
of the present catalytic silicon system over the use of stoichio-
metric silicon enolates are obvious: (1) catalytic use of silicon
species; (2) high diastereoselectivities; (3) silicon enolates of
amides are unstable and difficult to prepare in many cases. Further
investigations to clarify the more detailed mechanism as well as to
apply this process to other reactions, including asymmetric catal-
ysis, are currently underway in our laboratory.
i
treated 1c with Pr3SiOTf (TIPSOTf) at room temperature
and obtained 4 as a white solid. Solid 4 was then combined
with Et3N in toluene or CH2Cl2 at the same temperature.
However, silicon enolate 5 was not detected by NMR ana-
lysis, and independent signals of 4 and Et3N were observed
(Scheme 2).
Because silicon enolates were not detected, we then endea-
vored to prepare silicon enolates by other methods. According
to the literature on formation of silicon enolates of amides,
although they can be prepared from the corresponding lithium
enolates,12 the synthesis is strongly dependent on the amide
structure. Silicon enolates of amides are unstable, and a mixture
of O-enolate and C-enolate is often formed. Indeed, we tried to
prepare the TIPS enolate 5 from the corresponding lithium
enolate but failed despite several trials. Instead, we could prepare
tert-butyldimethylsilyl (TBS) enolate 6 from amide 1d according
to a literature method (LDA, TBSCl).12,13 It was found that
amide/TBSOTf complex 7 was formed from 1e and TBSOTf
(Scheme 3). When 7 and Et3N were combined, no silicon
enolate 6 was detected by NMR analysis, and independent
signals of complex 7 and Et3N were observed. On the other
hand, when silicon enolate 6 was treated with Et3NH OTf,
3
complex 7 was formed immediately. This result shows that 6 is
more basic than NEt3.
We further conducted several control experiments.
Imine 2 reacted with amide 1d in the presence of TBSOTf
(10 mol %), NEt3 (10 mol %), and CuOTf (5 mol %) in
dichloromethane at 0 °C for 48 h to afford the desired
Mannich adduct 8 in 65% yield with high anti selectivity
(anti/syn = 20/1) (eq 3). On the other hand, the reaction of 2
with silicon enolate 6 proceeded rapidly in the presence of
CuOTf (5 mol %), and after 30 min, product 8 was obtained
quantitatively but with low diastereoselectivity (anti/syn =
1/1) (eq 4). When the same reaction was conducted in the
’ ASSOCIATED CONTENT
S
presence of Et3NH OTf, the reaction proceeded gradually,
Supporting Information. Experimental procedures, product
b
3
and the anti product was obtained with high selectivity (eq 5).
Furthermore, no reaction occurred between 2 and complex 7
in the presence of CuOTf (5 mol %) (eq 6); however, the
characterization, and detailed mechanistic studies. This ma-
terial is available free of charge via the Internet at http://pubs.
acs.org.
710
dx.doi.org/10.1021/ja108764d |J. Am. Chem. Soc. 2011, 133, 708–711

Journal of the American Chemical Society
COMMUNICATION
’ AUTHOR INFORMATION
Corresponding Author
shu_kobayashi@chem.s.u-tokyo.ac.jp
’ ACKNOWLEDGMENT
This work was partially supported by a Grant-in-Aid for Science
Research from the Japan Society for the Promotion of Science
(JSPS) and the Global COE Program, The University of Tokyo,
MEXT, Japan.
’ REFERENCES
(1) Dilman, A. D.; Ioffe, S. A. Chem. Rev. 2003, 103, 733.
(2) (a) Gilman, H.; Clark, R. N. J. Am. Chem. Soc. 1947, 69, 967.
(b) Kr€uger, C. R.; Rochow, E. G. J. Organomet. Chem. 1964, 476.
(3) For aldol reactions, see: (a) Mukaiyama, T.; Banno, K.; Narasaka, K.
J. Am. Chem. Soc. 1974, 96, 7503. (b) Kobayashi, S.; Mukaiyama, T. Chem.
Lett. 1989, 297. (c) Kobayashi, S.; Uchiro, T.; Fujishita, Y.; Shiina, I.;
Mukaiyama, T. J. Am. Chem. Soc. 1991, 113, 4247. For Mannich-type
reactions, see: (d) Ikeda, K.; Achiwa, K.; Sekiya, M. Tetrahedron Lett. 1983,
24, 4707. (e) Mukaiyama, T.; Kashiwagi, K.; Matsui, S. Chem. Lett. 1989,
1397. (f) Ishitani, H.; Ueno, M.; Kobayashi, S. J. Am. Chem. Soc. 1997, 119,
7153. For 1,4-addition, see: (g) Narasaka, K.; Soai, K.; Mukaiyama, T.
Chem. Lett. 1974, 1223. (h) Narasaka, K.; Soai, K.; Aikawa, Y.; Mukaiyama,
T. Bull. Chem. Soc. Jpn. 1976, 49, 779. (i) Kitajima, H.; Katsuki, T. Synlett
1997, 568. Also see the following review: (j) Kobayashi, S.; Manabe, K.;
Ishitani, H.; Matsuo, J.-i. Sci. Synth. 2002, 4, 317–369.
(4) House, H. O.; Czuba, L. J.; Gall, M.; Olmstead, H. D. J. Org.
Chem. 1969, 34, 2324.
(5) For examples of catalytic direct-type addition of ester-equivalent
donors, see: (a) Bunlaksananusorn, T.; Rodriguez, A. L.; Knochel, P.
Chem. Commun. 2001, 745. (b) Kumagai, N.; Matsunaga, S.; Shibasaki,
M. J. Am. Chem. Soc. 2004, 126, 13632.
(6) (a) Saito, S.; Kobayashi, S. J. Am. Chem. Soc. 2006, 128, 8704.
(b) Morimoto, H.; Lu, G.; Aoyama, N.; Matsunaga, S.; Shibasaki, M.
J. Am. Chem. Soc. 2007, 129, 9588. (c) Matsubara, R.; Berthiol, F.;
Nguyen, H. V.; Kobayashi, S. Bull. Chem. Soc. Jpn. 2009, 82, 1083.
(d) Saito, S.; Tsubogo, T.; Kobayashi, S. Chem. Commun. 2007, 1236.
(7) The reaction of the Ts imine derived from 3-phenylpropanal
with 1a was not successful.
(8) Ni(OTf)2, AgOTf, and Sc(OTf)3 also worked efficiently to give
the desired product in over 90% yield. Zn(OTf)2, Cu(OTf)2, and
Yb(OTf)3 gave lower yields (50-75%). The anti/syn ratio was ∼6/1
in all cases.
(9) Mancheno, O. G.; Arrayꢀas, R. G.; Carretero, J. C. J. Am. Chem.
Soc. 2004, 126, 456.
(10) Ferraris, D.; Young, B.; Dudding, T.; Lectka, T. J. Am. Chem.
Soc. 1998, 120, 4548.
(11) (a) Nagao, Y.; Miyamoto, S.; Miyamoto, M.; Takeshige, H.;
Hayashi, K.; Sano, S.; Shiro, M.; Yamaguchi, K.; Sei, Y. J. Am. Chem. Soc.
2006, 128, 9722. Also see: (b) Djuri, S. W. J. Org. Chem. 1984, 49, 1311.
(12) (a) Woodbury, R.; Rathke, M. W. J. Org. Chem. 1978, 43, 881.
(b) Frick, U.; Simchen, G. Liebigs Ann. 1987, 839.
(13) The preparation of the corresponding TIPS enolate failed
[the C-enolate was formed predominantly (LDA/TIPSCl)].
(14) For example, see: Kobayashi, S.; Kobayashi, J.; Ishitani, H.;
Ueno, M. Chem.;Eur. J. 2002, 8, 4185.
(15) Carreira, E. M. In Comprehensive Asymmetric Catalysis; Jacobsen,
E. N., Pfaltz, A., Yamamoto, H., Eds.; Springer-Verlag: Berlin, 1999;
p 997.
711
dx.doi.org/10.1021/ja108764d |J. Am. Chem. Soc. 2011, 133, 708–711