Divergent Stereoinduction Mechanisms in Urea-Catalyzed Additions to Imines

Article abstract of DOI:10.1055/s-2003-41503

Catalyst structure/enantioselectivity profiles for the asymmetric Strecker and Mannich reactions were obtained through systematic variation of each modular component of the catalyst. Although the thiourea derivative 1 afforded optimal results in both reactions (97-98% ee), the structural elements responsible for stereoinduction were found to be fundamentally different. Insights gleaned from these studies led to the development of a new generation catalyst for the Mannich reaction that promotes the asymmetric silyl ketene acetal addition to N-Boc benzaldimine in 94% ee. The new catalyst is a simple amino acid derivative possessesing less than half the molecular weight and two fewer stereocenters relative to 1.

Full text of DOI:10.1055/s-2003-41503

CLUSTER
1919
Divergent Stereoinduction Mechanisms in Urea-Catalyzed Additions
to Imines
D
ivergent Stereoin
n
duction Me
c
n
hanisms in
U
a
rea-CatalyzedA
d
G
ditions toImines . Wenzel, Mathieu P. Lalonde, Eric N. Jacobsen*
Department of Chemistry and Chemical Biology, Harvard University, 12 Oxford Street, Cambridge, Massachusetts 02138, U.S.A.
Fax +1(617)4961880; E-mail: jacobsen@chemistry.harvard.edu
Received 16 July 2003
are applicable to the asymmetric Mannich reaction of N-
Boc aldimines with silyl ketene acetals (Scheme 1, Equa-
tion 2).6 Despite the fact that these two reactions are cata-
lyzed with nearly perfect enantioselectivity by the same
catalyst, the N-Boc aldimine substrates utilized in the
Mannich reaction are sterically and electronically differ-
ent from the N-alkyl imines used in the Strecker reaction.
Therefore, the same mechanism of stereoinduction could
not be assumed.7 With this in mind, we undertook an in-
vestigation of the correlation between catalyst structure
and enantioselectivity for both the Strecker and Mannich
reactions, systematically varying the amide, amino acid,
urea, diamine, and salicylaldimine elements in order to
develop a comparative stereoinduction profile.
Abstract: Catalyst structure/enantioselectivity profiles for the
asymmetric Strecker and Mannich reactions were obtained through
systematic variation of each modular component of the catalyst.
Although the thiourea derivative 1 afforded optimal results in both
reactions (97–98% ee), the structural elements responsible for
stereoinduction were found to be fundamentally different. Insights
gleaned from these studies led to the development of a new
generation catalyst for the Mannich reaction that promotes the
asymmetric silyl ketene acetal addition to N-Boc benzaldimine in
94% ee. The new catalyst is a simple amino acid derivative
possessesing less than half the molecular weight and two fewer
stereocenters relative to 1.
Key words: asymmetric catalysis, organocatalysis, nucleophilic
additions, imines, substituent effects
The effect of modifications to the amide bond and urea
moieties of catalyst 1 on the two model reactions is
illustrated in Table 1.8 In general, use of thiourea (Y = S)
rather than urea (Y = O) derivatives resulted in improved
enantioselectivity for both the Mannich and Strecker
reactions, although this effect was more pronounced for
the former.9 For example, thiourea catalyst 4 catalyzed the
Mannich reaction in 97% ee, while urea analog 5 afforded
only 82% ee (vis 98% vs. 97% ee in the Strecker
reaction).10 Enhanced enantioselectivity was also
observed in both reactions with catalysts bearing a
tertiary, rather than secondary, amide.9 Again, this effect
is more pronounced in the case of the Mannich reaction
(Table 1, catalysts 1 vs. 2). The steric properties of the
tertiary amide substituents were important in both
reactions, but with opposing trends. N,N-Dimethyl amide
catalyst 6 proved to be one of the best catalysts for the
Strecker reaction (99% ee), but was only moderately
effective in the case of the Mannich reaction (86% ee).
The more sterically demanding N,N-diisobutyl derivative
8 afforded improved ee in the case of the Mannich
reaction (93% ee), but poorer results in the Strecker
reaction (95% ee). In both reactions, use of an ester
moiety in place of the amide bond resulted in a reduction
in enantioselectivity (10 vs. 3).
Nature is the principal expert practitioner of asymmetric
synthesis, and accordingly serves as both a model and an
inspiration for the design of synthetic chiral catalysts.
While enzymes exploit several types of interactions
between active site and substrate to effect high levels of
stereoinducton, the identification of highly enantio-
selective reactions catalyzed by proline1 and other
remarkably simple organic catalysts2,3 has raised the
question of what constitute the minimal functional and
structural features required for efficient asymmetric
catalysis.4 Herein we report a systematic investigation of
structure–enantioselectivity relationships in an organic
catalyst system for imine addition reactions. This effort
has led to the discovery that a chiral urea catalyst operates
by fundamentally different stereoinduction mechanisms
in the activation of related imine substrates, and to the
identification of an extraordinarily simple and effective
new catalyst for the addition of silyl ketene acetals to N-
Boc-protected aldimines (the Mannich reaction).
In 1998, we described a parallel-library approach to the
discovery of catalysts for the asymmetric hydrocyanation
of imines (the Strecker reaction).5a Through a combina-
tion of empirical and structure-based optimization studies,
thiourea derivative 1 and closely related analogs were
identified as highly general and effective catalysts for the
asymmetric hydrocyanation of N-allyl or N-benzyl
aldimines5a,b,e and ketoimines (Scheme 1, Equation 1).5c
Subsequently, it was discovered that the same catalysts
The identity of the amino acid proved to be critical in the
Mannich reaction (Table 2). Whereas tert-leucine deriva-
tives afforded best results, substitution with valine or ala-
nine led to abrupt decreases in enantioselectivity (1 vs. 11
and 12). Incorporation of the aromatic amino acid phe-
nylglycine (catalyst 13) also resulted in diminished ee’s.
Increasing the steric properties of the amino acid substit-
uent beyond those of tert-leucine, as in catalyst 14, also
led to a reduction in enantioselectivity relative to 1, al-
SYNLETT 2003, No. 12, pp 1919–1922
2
9
.0
9
.2
0
0
3
Advanced online publication: 19.09.2003
DOI: 10.1055/s-2003-41503; Art ID: Y00903ST
© Georg Thieme Verlag Stuttgart · New York

1920
A. G. Wenzel et al.
CLUSTER
though only to a moderate extent. In contrast, the perfor- tivity (42% ee).15 Returning to the original C1-symmetric
mance of the Strecker reaction is remarkably insensitive design of 1, catalyst 25, wherein the diamine/Schiff base
to the identity of the amino acid. Decreasing the size of the moiety was replaced with cyclohexylamine, was prepared
amino acid substituent from tert-butyl to methyl resulted and found to catalyze the Mannich reaction in 90% ee.
in highly enantioselective catalysis (>90% ee, catalyst Aniline derivative 26 afforded even higher levels of ste-
12). Phenylglycine- (13) and 2-amino-3-methyl-3-phe- reoinduction (94% ee and quantitative conversion within
nyl-butyric acid (14)-derived catalysts – while not superi- 48 h).16 Compound 26 represents a remarkable simplifica-
or to 1 – also serve as effective catalysts for the Strecker tion in catalyst design, as it possesses less than half the
reaction.
molecular weight of parent catalyst 1 and two fewer
stereocenters. It can be prepared readily from commer-
cially available starting materials in 95% overall yield,
with only a single chromatographic purification step.
A study of the effect of modifications to the salicylald-
imine portion of the catalyst on enantioselectivity also
revealed dramatic differences between the two reactions
(Table 3). In general, the Strecker reaction proved to be While the Mannich and Strecker reactions are promoted
sensitive to subtle variations in the salicylaldimine sub- with high ee using the same catalyst 1, they in fact obey
stituents. While decreasing the size of the ortho-substitu- quite distinct stereoinduction profiles. This finding is
ent (R2) had a minor effect on ee (15–17), modification of perhaps not completely surprising given the nature of the
the para-substituent (R3) resulted in diminished enanti- imine substrates utilized in each reaction and fundamental
oselectivity (15 vs. 1).11 Methylation of the phenol differences in their respective nucleophiles. Nonetheless,
(R1 = Me, 19 vs. 15), or substitution with 1-hydroxyl-2- it suggests that 1 might belong to the rapidly growing list
naphthaldimine (20), also led to significant reductions in of ‘privileged’ chiral catalysts: systems that promote high
the enantioselectivity of the Strecker reaction. In contrast, enantioselectivity in a range of mechanistically distinct
asymmetric induction in the Mannich reaction was reactions.17 Through mapping of the critical design
virtually insensitive to salicylaldimine modification, with features required for asymmetric induction, a new catalyst
negligible dependence on R2 (15–17) and none on R3 (1 for the activation of N-Boc aldimines has been developed
vs. 15).12 Although lower enantioselectivity was observed that possesses the minimal requirements for asymmetric
in toluene with catalyst 20, this could be attributed to low induction. A catalyst derived from a single a-amino acid
catalyst solubility. Replacing the solvent with THF, which has been shown to impart enzyme-like enantioselectivity
is generally an effective solvent for the Mannich reaction in a synthetically valuable transformation. Whereas
(Table 1, 1), led to complete solubilization of 20 and out- proline and related secondary amine catalysts function as
standing enantioselectivity in the Mannich reaction (98% recyclable auxiliaries by covalently modifying either the
ee).13 As in the case of the Strecker reaction, methylation nucleophile1 or electrophile2 to achieve high levels of
of the phenol (R1 = Me, 19) led to decreased ee in the asymmetric induction, C1-symmetric catalyst 26 induces
Mannich reaction (97% to 82% ee).
high enantioselectivity through simple hydrogen-bonding
interactions. The potential of small molecule hydrogen-
bond donors in asymmetric catalysis appears to be quite
significant,18 and is being explored aggressively in
continuing studies.
The effect of variations in the structure and
stereochemistry of the diamine unit of the catalysts is
summarized in Table 4. Incorporation of (R,R)-diphenyl-
ethylene diamine (21) was well tolerated in the Strecker
reaction (98% ee), but was detrimental to the Mannich
reaction (80% ee). Catalyst prepared from the sterically
demanding (R,R)-di-tert-butyl-ethylene diamine (22) was
inactive in both reactions. Incorporation of mismatched
(S,S)-diamino-cyclohexane into catalyst (23) led to an
unexpected result: while the Strecker reaction was
catalyzed with precipitously lower enantioselectivity and
a reversal in the sense of asymmetric induction relative to
1, the Mannich reaction proceeded in 90% ee with the
same sense of induction.
Me tBu
N
S
Ph
N
H
N
H
O
N
Catalyst 1
HO
tBu
OCOtBu
O
Strecker Reaction
1. 1 (1 mol%),
toluene,
–65 °C, 20 h
N
F3C
N
+
HCN
The remarkable indifference of the Mannich reaction to
variations in the catalyst’s salicylaldimine unit and di-
amine stereochemistry raised the compelling possibility
that the entire diamine/Schiff base moiety might be ex-
cised from the catalyst structure without detrimental ef-
fect on enantioselectivity. To test this hypothesis, C2-
symmetric catalyst 24 (Table 5) was prepared, incorporat-
ing all structural elements identified to be vital to stereoin-
duction within a symmetrical framework.14 However,
catalyst 24 catalyzed the Mannich reaction of N-Boc
benzaldimine with poor conversion and low enantioselec-
(1)
(2)
Ph
H
Ph
CN
2. TFAA
1.25 equiv
98% ee
100% conv.
Mannich Reaction
1. 1 (5 mol%),
toluene,
–40 °C, 48 h
Boc
N
Boc
Ph
OTBS
OiPr
NH
O
+
Ph
H
OiPr
2. TFA, 2 min
2.0 equiv
97% ee
96% conv.
Scheme 1
Synlett 2003, No. 12, 1919–1922 © Thieme Stuttgart · New York

CLUSTER
Divergent Stereoinduction Mechanisms in Urea-Catalyzed Additions to Imines
1921
Table 1 Modifications to the Amide Bond and Urea Moieties
Table 4 Effects of Diamine Structure and Stereochemistry
R2
X
Y
tBu
Me
N
tBu
S
R1
Ph
R2
R1
N
N
N
N
21: R1=Ph, R2=Ph (R,R)
22: R1=tBu, R2=tBu (R,R)
23: R1,R2= -(CH2)4- (S,S)
H
H
H
H
O
N
O
N
HO
tBu
HO
tBu
OCOtBu
OCOtBu
6: R1=Me, R2=Me, X=N, Y=S
7: R1=Me, R2=Me, X=N, Y=O
8: R1=isobutyl, R2=isobutyl, X=N, Y=S
9: R1=cHex, R2=H, X=N, Y=O
1: R1=Bn, R2=Me, X=N, Y=S
2: R1=Bn, R2=H, X=N, Y=S
3: R1=Bn, R2=H, X=N, Y=O
4: R1=Bn, R2=Bn, X=N, Y=S
Catalyst
Strecker ee (%)a
Mannich ee (%)b
21
22
23
98
80
5: R1=Bn, R2=Bn, X=N, Y=O 10: R1=Bn, X=O, Y=O
No reaction
27 (R)
No reaction
90
Catalyst Strecker Mannich Catalyst Strecker Mannich
a Strecker experimental procedure, see ref.6,
b Mannich experimental procedure, see ref.7
ee (%)a
ee (%)b
ee (%)a
ee (%)b
1
2
3
4
5
98
98
97
98
97
97 (97c)
91
6
7
99
98
95
97
80
86
80
93
84
44
Table 5 New Catalyst Designs for the Asymmetric Mannich Reac-
tion
81
8
97
9
Me tBu
N
S
tBu
Me
N
82
10
Ph
Ph
N
H
N
H
a For experimental details on the Strecker reaction, see ref.5b,
b For experimental details on the Mannich reaction, see ref.6,
c Reaction performed in THF.
O
O
24
Me tBu
N
S
Me tBu
N
S
Ph
Ph
Ph
Table 2 Variation of the Amino Acid
N
H
N
H
N
H
N
H
O
O
Me
N
R
S
11: R=iPr (L-Val)
12: R=Me (L-Ala)
13: R=Ph (L-Phg)
14: R=CMe2Ph
25
26
Ph
N
N
H
H
O
N
1. catalyst (5 mol%),
–40 °C, 48 h
Boc
Boc
Ph
OTBS
OiPr
N
NH
O
(D-2-amino-3-methyl
-3-phenylbutyric acid)
HO
tBu
+
Ph
H
OiPr
2. TFA, 2 min
OCOtBu
Catalyst
Conversion (%)a
ee (%)
Catalyst Strecker Mannich Catalyst Strecker Mannich
24b
25b
26c
37
42
90
94
ee (%)
ee (%)
ee (%)
ee (%)
90
11
12
96
91
51
22
13
92
38
100
14a
98 (R)
90 (S)
a Determined via GC relative to dodecane as an internal standard.
b Reaction performed in toluene.
a Enantiomeric catalyst prepared from D-amino acid and (S,S)-diami-
nocyclohexane.
c Reaction performed in THF.
Table 3 Salicylaldimine Modifications
15: R1=H, R2=tBu, R3=tBu
16: R1=H, R2=iPr,R3=tBu
17: R1=H, R2=Me, R3=tBu
18: R1=H, R2=H, R3=H
Me tBu
N
S
Ph
N
N
H
H
O
N
Acknowledgment
19: R1=Me, R2=tBu, R3=tBu
R1O
This work was supported by the NIH (GM-43214), and by
predoctoral fellowships to A.G.W. from the National Science
Foundation and Pfizer and to M.P.L. from NSERC and the Baxter
and Alma Ricard Foundation. We also wish to thank Dr. Takeshi
Kuribayashi for the generous donation of samples of catalysts 5, 9,
and 10 used in this study.
20: 1-hydroxy-2-naphthyl
R2
R3
Catalyst Strecker Mannich Catalyst Strecker Mannich
ee (%)
ee (%)
ee (%)
ee (%)
15
16
17
92
92
91
97
97
96
18
19
20
94
64
67
91
82
94 (98a)
a Reaction performed in THF (see ref.13).
Synlett 2003, No. 12, 1919–1922 © Thieme Stuttgart · New York

1922
A. G. Wenzel et al.
CLUSTER
(7) Kinetic studies carried out on both reactions are consistent
with imine pre-association to catalyst followed by
nucleophile addition to the catalyst-imine complex (ref.5e
and Wenzel, A. G., unpublished results).
(8) The standard screening conditions are depicted in Scheme 1.
Benzaldimines were chosen as substrates for each reaction to
maximize structural and electronic similarity. While
aliphatic N-alkyl imines have been successfully employed in
the Strecker reaction, aliphatic N-Boc aldimines have not
been investigated in the Mannich reaction because no useful
method has been identified for their synthesis.
(9) Although negligible improvement in the Strecker reaction
was observed with the N-allyl benzaldimine substrate
screened in this study, pronounced improvement has been
observed in cases of problematic substrates. (See ref.5e)
(10) In general, conversion was found to correlate with
enantioselectivity, with the more selective catalysts also
proving to be the most reactive.
References
(1) (a) Hajos, Z. G.; Parrish, D. R. German Patent DE 2102623,
1971. (b) Hajos, Z. G.; Parrish, D. R. J. Org. Chem. 1974,
39, 1615. (c) Eder, U.; Sauer, G.; Wiechert, R. Angew.
Chem., Int. Ed. Engl. 1971, 10, 496. (d) Notz, W.; List, B. J.
Am. Chem. Soc. 2000, 122, 7386. (e) List, B. J. Am. Chem.
Soc. 2000, 122, 9336. (f) List, B.; Lerner, R. A.; Barbas, C.
F. III J. Am. Chem. Soc. 2000, 122, 2395. (g) List, B.;
Pojarliev, P.; Castello, C. Org. Lett. 2001, 3, 573. (h) List,
B.; Pojarliev, P.; Martin, H. J. Org. Lett. 2001, 3, 2423.
(i) List, B. Tetrahedron 2002, 58, 5573. (j) List, B.;
Pojarliev, P.; Biller, W. T.; Martin, H. J. J. Am. Chem. Soc.
2002, 124, 827. (k) Córdova, A.; Notz, W.; Zhong, G.;
Betancort, J. M.; Barbas, C. F. III J. Am. Chem. Soc. 2002,
124, 1842. (l) Córdova, A.; Watanabe, S.; Tanaka, F.; Notz,
W.; Barbas, C. F. III J. Am. Chem. Soc. 2002, 124, 1866.
(m) List, B. J. Am. Chem. Soc. 2002, 124, 5656.
(n) Kumaragurubaran, N.; Juhl, K.; Zhuang, W.; Bøgevig,
A.; Jørgensen, K. A. J. Am. Chem. Soc. 2002, 124, 6254.
(o) Northrup, A. B.; MacMillan, D. W. C. J. Am. Chem. Soc.
2002, 124, 6798. (p) Chowdari, N. S.; Ramachary, D. B.;
Barbas, C. F. III Org. Lett. 2003, 5, 1685.
(11) For a comprehensive description of the effect of varying the
R3 substituent on the enantioselectivity of the Strecker
reaction, see ref.5a
(12) An optimization library performed during early-phase
methodological development for the Mannich reaction
revealed that variation of the R3 substituent of the
salicylaldimine (R3 = t-Bu, Me, H, OTIPS, t-BuO, OMe,
OCO-t-Bu, Br, Cl) has no effect on enantioselectivity or
conversion (ref.6).
(13) The Strecker reactions were carried out at lower catalyst
loadings and more dilute conditions {[20] = 0.98 mM in the
Strecker vs. 42 mM in the Mannich}, thereby obviating the
need for a solvent switch with sparingly soluble 20.
(14) For discussions of the advantages of C2-symmetry in
asymmetric catalysts, see: (a) Whitesell, J. K. Chem. Rev.
1989, 89, 1581. (b) Kagan, H. B. In Comprehensive
Asymmetric Catalysis, Vol. 1; Jacobsen, E. N.; Pfaltz, A.;
Yamamoto, H., Eds.; Springer: New York, 1999, Chap. 2.
(15) Analogs of 24 derived from less sterically demanding amino
acids (e.g. valine, alanine) also performed poorly as catalysts
for the Mannich reaction. Catalyst 24 proved almost
completely unreactive in the Strecker reaction.
(16) The model Strecker reaction was catalyzed by 26 in 40% ee
with the opposite sense of stereoinduction relative to 1.
(17) Yoon, T. P.; Jacobsen, E. N. Science 2003, 299, 1691.
(18) The recent report on the use of TADDOL as a Diels–Alder
catalyst appears to represent another, very promising, for
example see: Huang, Y.; Unni, A. K.; Thadani, A. N.;
Rawal, V. H. Nature (London) 2003, 424, 146.
(2) (a) Ahrendt, K. A.; Borths, C. J.; MacMillan, D. W. C. J. Am.
Chem. Soc. 2000, 122, 4243. (b) Jen, W. S.; Wiener, J. J.;
MacMillan, D. W. C. J. Am. Chem. Soc. 2000, 122, 9874.
(c) Betancort, J. M.; Barbas, C. F. III Org. Lett. 2001, 3,
3737. (d) Northrup, A. B.; MacMillan, D. W. C. J. Am.
Chem. Soc. 2002, 124, 2458. (e) Paras, N. A.; MacMillan,
D. W. C. J. Am.Chem. Soc. 2002, 124, 7894. (f) Juhl, K.;
Jørgensen, K. A. Angew. Chem. Int. Ed. 2003, 42, 1498.
(3) For recent reviews on organocatalysis, see: (a) Dalko, P. I.;
Moisan, L. Angew. Chem. Int. Ed. 2001, 40, 3726.
(b) Jarvo, E. R.; Miller, S. J. Tetrahedron 2002, 58, 2481.
(4) Movassaghi, M.; Jacobsen, E. N. Science 2002, 298, 1904.
(5) (a) Sigman, M. S.; Jacobsen, E. N. J. Am. Chem. Soc. 1998,
120, 4901. (b) Sigman, M. S.; Vachal, P.; Jacobsen, E. N.
Angew. Chem. Int. Ed. 2000, 39, 1279. (c) Vachal, P.;
Jacobsen, E. N. Org. Lett. 2000, 2, 867. (d) Su, J. T.;
Vachal, P.; Jacobsen, E. N. Adv. Synth. Catal. 2001, 343,
197. (e) Vachal, P.; Jacobsen, E. N. J. Am. Chem. Soc. 2002,
124, 10012.
(6) Wenzel, A. G.; Jacobsen, E. N. J. Am. Chem. Soc. 2002, 124,
12964.
Synlett 2003, No. 12, 1919–1922 © Thieme Stuttgart · New York