Catalytic, asymmetric Mannich-type reactions of N-acylimino esters: Reactivity, diastereo- and enantioselectivity, and application to synthesis of N-acylated amino acid derivatives

Article abstract of DOI:10.1021/ja0281840

In the presence of a catalytic amount of Cu(OTf)₂-chiral diamine 3e complex, N-acylimino esters reacted with silyl enol ethers to afford the corresponding Mannich-type adducts in high yields with high enantioselectivities. A wide variety of silyl enol ethers derived from ketones, as well as esters and thioesters, reacted smoothly. In the reactions of α-substituted silyl enol ethers (α-methyl or benzyloxy), the desired syn-adducts were obtained in high yields with high diastereo- and enantioselectivities. Several intermediates for the synthesis of biologically important compounds were prepared using this novel catalytic asymmetric Mannich-type reaction, and at the same time, absolute and relative stereochemical assignments were made. In addition, it has been revealed that alkyl vinyl ethers reacted with N-acylimino esters in the presence of a catalytic amount of the Cu(II) catalyst to give the corresponding Mannich-type adducts in high yields with high enantioselectivities. This is the first example of catalytic asymmetric Mannich-type reactions with alkyl vinyl ethers. The reaction mechanism, structure of chiral catalyst-electrophile complexes, and transition states of these catalytic asymmetric reactions were assumed based on X-ray crystallographic analysis of the Cu(II)-chiral amine complex, PM3 calculations, and FT-IR analyses, etc. Finally, (1R,3R)-N-(3-hydroxy- 1-hydroxymethyl-3-phenylpropyl)dodecanamide (HPA-12, 1), a new inhibitor of ceramide trafficking from endoplasmic reticulum to the site of sphingomyerin (SM) synthesis, has been synthesized efficiently using the present Mannich-type reaction as a key step. The synthesis involved three steps (two-pot), and total yield was 82.9%.

Full text of DOI:10.1021/ja0281840

Published on Web 02/06/2003
Catalytic, Asymmetric Mannich-type Reactions of N-Acylimino
Esters: Reactivity, Diastereo- and Enantioselectivity, and
Application to Synthesis of N-Acylated Amino Acid
Derivatives
Shuj Kobayashi,* Ryosuke Matsubara, Yoshitaka Nakamura,
Hidetoshi Kitagawa, and Masaharu Sugiura
Contribution from the Graduate School of Pharmaceutical Sciences, The UniVersity of Tokyo,
Hongo, Bunkyou-ku, Tokyo, 113-0033, Japan
Received August 17, 2002 ; E-mail: skobayas@mol.f.u-tokyo.ac.jp.
Abstract: In the presence of a catalytic amount of Cu(OTf)2-chiral diamine 3e complex, N-acylimino esters
reacted with silyl enol ethers to afford the corresponding Mannich-type adducts in high yields with high
enantioselectivities. A wide variety of silyl enol ethers derived from ketones, as well as esters and thioesters,
reacted smoothly. In the reactions of R-substituted silyl enol ethers (R-methyl or benzyloxy), the desired
syn-adducts were obtained in high yields with high diastereo- and enantioselectivities. Several intermediates
for the synthesis of biologically important compounds were prepared using this novel catalytic asymmetric
Mannich-type reaction, and at the same time, absolute and relative stereochemical assignments were made.
In addition, it has been revealed that alkyl vinyl ethers reacted with N-acylimino esters in the presence of
a catalytic amount of the Cu(II) catalyst to give the corresponding Mannich-type adducts in high yields with
high enantioselectivities. This is the first example of catalytic asymmetric Mannich-type reactions with alkyl
vinyl ethers. The reaction mechanism, structure of chiral catalyst-electrophile complexes, and transition
states of these catalytic asymmetric reactions were assumed based on X-ray crystallographic analysis of
the Cu(II)-chiral amine complex, PM3 calculations, and FT-IR analyses, etc. Finally, (1R,3R)-N-(3-hydroxy-
1-hydroxymethyl-3-phenylpropyl)dodecanamide (HPA-12, 1), a new inhibitor of ceramide trafficking from
endoplasmic reticulum to the site of sphingomyerin (SM) synthesis, has been synthesized efficiently using
the present Mannich-type reaction as a key step. The synthesis involved three steps (two-pot), and total
yield was 82.9%.
Introduction
this issue, we have recently developed a convenient method for
the preparation of N-acylimino esters using a polymer-supported
N-Acylated amino acid derivatives are often observed in
biologically important compounds such as peptides, ceramide,
etc.1 For the preparation of these compounds, the use of R-imino
esters provides convenient ways.2 According to typical methods,
R-imino esters are treated with nucleophiles, and the protective
groups of the amino moieties are removed and then the amino
groups are acylated (eq 1).3 More conveniently, reactions of
N-acylimino esters would give N-acylated amino acid derivatives
directly (eq 2).4 However, N-acylimino esters are unstable and
their use in organic synthesis has been rather limited. To address
amine.5 Furthermore, a novel chiral copper catalyst has been
developed to achieve enantioselective Mannich-type reactions
of N-acylimino esters.6 In this paper, we present a full account
of this asymmetric reaction including reactivity, selectivity
(diastereo- and enantioselectivity), reaction mechanism, catalyst
structure, etc. Facile synthesis of a new inhibitor of ceramide
trafficking, (1R,3R)-N-(3-hydroxy-1-hydroxymethyl-3-phenyl-
propyl)dodecanamide (HPA-12, Figure 1, 1), using this catalytic
enantioselective reaction is also reported.
Results and Discussion
(1) (a) Helms, G. L.; Moore, R. E.; Niemczura, W. P.; Patterson, G. M. L.;
Tomer, K. B.; Gross, M. L. J. Org. Chem. 1988, 53, 1298. (b) Matsunaga,
S.; Fusetani, N.; Hashimoto, K.; Wa¨lchli, M. J. Am. Chem. Soc. 1989, 111,
2582. (c) Humphrey, J. M.; Chamberlin, A. R. Chem. ReV. 1997, 97, 2243.
(d) von Do¨hren, H.; Keller, U.; Vater, J.; Zocher, R. Chem. ReV. 1997, 97,
2675. (e) Dickson, R. C. Annu. Rec. Biochem. 1998, 67, 27. (f) Kolter, T.;
Sandhoff, K. Angew. Chem. Int. Ed. Engl. 1999, 38, 1532.
(2) (a)Weinreb, S. M.; Scola, P. A. Chem. ReV. 1989, 89, 1525. To our
knowledge, R-imino ester was first used as a precursor to R-amino acid by
Achmatowicz Jr. (b)Achmatowicz Jr., O.; Pietraszkiewicz, M. J. Chem.
Soc., Chem. Commun. 1976, 484.
First, we searched for an effective catalyst for Mannich-type
reactions of N-acylimino esters. Although N-acylimino esters
(4) (a) Kober, R.; Papadopoulos, K.; Miltz, W.; Enders, D.; Steglich, W. Reuter,
H.; Puff, H. Tetrahedron 1985, 41, 1693. (b) Mu¨nster, P.; Steglich, W.
Synthesis 1987, 223. (c) Bretschneider, T.; Miltz, W.; Mu¨nster, P.; Steglich,
W. Tetrahedron 1988, 44, 5403. (d) Saaby, S.; Fang, X.; Gathergood, N.;
Jørgensen, K. A. Angew. Chem., Int. Ed. 2000, 39, 4114. (e) Dudding, T.;
Hafez, A. M.; Taggi, A. E.; Wagerle, T. R.; Lectka, T. Org. Lett. 2002. 4,
387. For a review of N-acyliminium ions and related intermediates, see
Speckamp, W. N.; Moolenaar, M. J. Tetrahedron 2000, 56, 3817.
(3) (a) Fujii, A.; Hagiwara, E.; Sodeoka, M. J. Am. Chem. Soc. 1999, 121,
5450. (b) Ferraris, D.; Dudding, T.; Young, B.; Drury, W. J., III; Lectka,
T. J. Org. Chem. 1999, 64, 2168. (c) Yamada, K.; Harwood, S. J.; Gro¨ger,
H.; Shibasaki, M. Angew. Chem., Int. Ed. Engl. 1999, 38, 3504.
(5) Kobayashi, S.; Kitagawa, H.; Matsubara, R. J. Comb. Chem. 2001, 3, 401.
(6) Kobayashi, S.; Matsubara, R.; Kitagawa, H. Org. Lett. 2002, 4, 143.
9
10.1021/ja0281840 CCC: $25.00 © 2003 American Chemical Society
J. AM. CHEM. SOC. 2003, 125, 2507-2515
2507

A R T I C L E S
Kobayashi et al.
Table 1. Effects of Copper Salts, Chiral Ligands, and Reaction
Conditions
yield
(%)a
Ee
(%)b
entry
catalyst
1
2
3
4
5
6
7
8
CuClO4‚4CH3CN + BINAP
CuClO4‚4CH3CN + xylyl-BINAPc
CuClO4‚4CH3CN + 3a
Cu(OTf)2 + BINAP
Cu(OTf)2 + 3a
44
69
62
71
25
32
24
12
20
32
65
91
92
67
91
63
15
12
8
36
63
63
75
52
80
64
75
85
94
89
0
Cu(OTf)2 + 3b
Cu(OTf)2 + 3c
Cu(OTf)2 + 3d
9
Cu(OTf)2 + 3e
10
11d
12d,e
13d,f
14d,g
15d,f
16d,f
Cu(OTf)2 + 3f
Cu(OTf)2 + 3e
Cu(OTf)2 + 3e
Cu(OTf)2 + 3e
Cu(OTf)2 + 3e
Cu(OTf)‚1/2C6H6 + 3e
Cu(SbF6)2 + 3e
65
Figure 1. Novel inhibitor of ceramide trafficking, HPA-12.
a Isolated yield from the corresponding R-chloroglycinate (imino ester
was generated in situ). b Determined by chiral HPLC analysis. c (S)-(-)-
Bis[bis(3,5-dimethylphenyl)phosphino]-1,1′-binaphthyl. d 2a was added slowly
after the addition of the silyl enol ether to the catalyst. e At -20 °C. f At 0
°C. g At rt.
are known to be unstable and are difficult to prepare in high
yields, we have recently developed a convenient preparation
method of N-acylimino esters from the corresponding R-chloro-
glycine derivatives using a polymer-supported amine. We
selected a model Mannich-type reaction of N-acylimino ester
2a with the silyl enol ether 4a derived from acetophenone, and
several reaction factors such as metals, chiral ligands, conditions,
etc., were examined. It was revealed that a chiral copper(II)
catalyst prepared from Cu(OTf)2 and chiral ligand 3e7 was
promising. The effects of copper salts, chiral ligands, and
reaction conditions are summarized in Table 1. It was already
reported by Lectka et al. that CuClO4‚4CH3CN-BINAP8 was
effective for Mannich-type reactions of N-tosylimino esters with
silyl enol ethers.9-11 In the reaction of 2a, however, low
enantiometic excess was observed (Table 1, entries 1-3). On
the other hand, combination of Cu(OTf)2 and chiral diamine 3
gave promising results; for example, to a mixture of Cu(OTf)2
and ligand 3e (10 mol %), 2a and the silyl enol ether were added
successively to afford the desired adduct in 80% ee, but in low
yield (entry 9). In this case, formation of a dimer of 2a (6) was
observed, presumably due to contamination by water. We
carefully performed the reaction under anhydrous conditions,
and the silyl enol ether was added to the catalyst first and then
2a was slowly charged over 20 min to this mixture. The yield
was dramatically improved, and the desired adduct was obtained
in 92% yield with 94% ee (entry 13). For the copper salt,
Cu(OTf)2 was much better than CuClO4, CuOTf, and even
Cu(SbF6)2.12 In addition, interesting temperature effects on
selectivity were observed in this reaction. When the reactions
(7) (a) Mimoun, H.; Laumer, J. Y. S.; Giannini, L.; Scopelliti, R.; Floriani, C.
J. Am. Chem. Soc. 1999, 121, 6158. (b) Cavallo, L.; Cucciolito, M. E.;
Martino, A. D.; Giordano, F.; Orabona, I.; Vitagliano, A. Chem. Eur. J.
2000, 6, 1127. (c) Kobayashi, S.; Hamada. T.; Manabe, K. J. Am. Chem.
Soc. 2002, 124, 5640, and references therein.
(8) For the preparation of CuClO4‚4CH3CN, see: Kubas, G. J. Inorganic
Synthesis; Shriver, D. F., Ed.; Plenum: New York, 1979; Vol. XIX, p 90.
(9) Ferraris, D.; Young, B.; Cox, C.; Dudding, T.; Drury, W. J., III.; Ryzhkov,
L.; Taggi, A. E.; Lectka, T. J. Am. Chem. Soc. 2002, 124, 67.
(10) For the use of the Cu-BINAP system in enantioselective reactions of
R-imino esters, see: (a) Ferraris, D.; Young, B.; Dudding, T.; Lectka, T.
J. Am. Chem. Soc. 1998, 120, 4548. (b) Drury, W. J., III.; Ferraris, D.;
Cox, C.; Young, B.; Lectka, T. J. Am. Chem. Soc. 1998, 120, 11 006. (c)
Yao, S.; Johannsen, M.; Hazell, R. G.; Jørgensen, K. A. Angew. Chem.,
Int. Ed. Engl. 1998, 37, 3121. (d) Yao, S.; Fang, X.; Jørgensen, K. A.
Chem. Commun. 1998, 2547. (e) Fang, X.; Johannsen, M.; Yao, S.;
Gathergood, N.; Hazell, R. G.; Jørgensen, K. A. J. Org. Chem. 1999, 64,
4844. (f) Juhl, K.; Hazell, R. G.; Jørgensen, K. A. J. Chem. Soc., Perkin
Trans. 1 1999, 2293. (g) Johannsen, M. Chem. Commun. 1999, 2233. (h)
Saaby, S.; Fang, X.; Gathergood, N.; Jørgensen, K. A. Angew. Chem., Int.
Ed. Engl. 2000, 39, 4114. (i) Yao, S.; Saaby, S.; Hazell, R. G.; Jørgensen,
K. A. Chem. Eur. J. 2000, 6, 2435. (j) Knudsen, K. R.; Risgaard, T.;
Nishiwaki, N.; Gothelf, K. V.; Jørgensen, K. A. J. Am. Chem. Soc. 2001,
123, 5843. (k) Nishiwaki, N.; Knudsen, K. R.; Gothelf, K. V.; Jørgensen,
K. A. Angew. Chem., Int. Ed. Engl. 2001, 40, 2992. Cf. (l) Kru¨ger, J.;
Carreira, E. M. J. Am. Chem. Soc. 1998, 120, 837.
(11) For other chiral Cu-catalyzed reactions, see, for example: (a) Wei, C.; Li,
C. J. J. Am. Chem. Soc. 2002, 124, 5638. (b) Evans, D. A.; Tregay, S. W.;
Burgey, C. S.; Paras, N. A.; Vojkovsky, T. J. Am. Chem. Soc. 2000, 122,
7936-7943. (c) Johnson, J. S.; Evans, D. A. Acc. Chem. Res. 2000, 33,
325. (d) Bromidge, S.; Wilson, P. C.; Whiting, A. Tetrahedron Lett. 1998,
39, 8905. (e) Jnoff, E.; Ghosez, L. J. Am. Chem. Soc. 1999, 121, 2617. (f)
Audrain, H.; Jørgensen, K. A. J. Am. Chem. Soc. 2000, 122, 11 543. (g)
Gathergood, N.; Zhuang, W.; Jørgensen, K. A. J. Am. Chem. Soc. 2000,
122, 12 517.
(12) Cu(SbF6)2-diamine complex was prepared according to the reported method.
Evans, D. A.; Scheidt, K. A.; Johnston, J. N. Wills, M. C. J. Am. Chem.
Soc. 2001, 123, 4480.
9
2508 J. AM. CHEM. SOC. VOL. 125, NO. 9, 2003

Catalytic, Asymmetric Mannich-type Reactions
A R T I C L E S
Table 3. Asymmetric Mannich-type Reactions with Silyl Enol
Table 2. Mannich-type Reactions with Silyl Enol Ethers Derived
from Ketones
Ethers Derived from Esters or Thioesters
imine
1
2
entry
R
R
Nu
yield (%)a
Ee(%)b
product
1
2
3
4
5
6
7
8
9
Et
CH3 (2b)
4a
4a
4a
4a
4a
4b
4c
4d
4a
85
88
92
91
91
97
88
85
75
94
94
94
92
92
92
93
94
93
5b
5c
5a
5d
5e
5f
5g
5h
5i
C10H21 (2c)
C11H23 (2a)
C12H25 (2d)
C13H27 (2e)
C11H23 (2a)
C11H23 (2a)
C11H23 (2a)
C11H23 (2f)
Bn
a Isolated yield from the corresponding R-chloroglycinate (imino ester
was generated in situ). b Determined by chiral HPLC analysis.
were carried out at -78 °C, -45 °C, -20 °C, 0 °C, and room
temperature, the best selectivity was obtained at 0 °C. As the
temperature was lowered, the selectivity was decreased. When
the catalyst prepared at room temperature was cooled to -78
°C, a white precipitate was observed and the solution color
changed from green to blue. This observation indicates that the
catalyst structure was changed at lower temperature such as -78
°C. On the other hand, a slight loss of the selectivity was
observed at room temperature. The loss was mainly ascribed
to an uncatalyzed reaction pathway because this reaction
proceeded to some extent without a catalyst to give a racemic
adduct.
Several examples of the Mannich-type reactions of N-
acylimino esters with silyl enol ethers derived from ketones in
the presence of 10 mol % of the copper catalyst are shown in
Table 2. In all cases, the desired adducts were obtained in high
yields with high enantioselectivities.
We then investigated Mannich-type reactions of N-acylimino
esters with silyl enol ethers derived from thioesters and esters
(Table 3). This type of reaction provides one of the most
efficient methods for the preparation of optically active N-acyl
aspartic acid derivatives, which are versatile intermediates for
several kinds of biologically important natural products and
pharmaceutical chemicals.13 Although asymmetric Mannich-type
reactions of N-tosylimino esters or N-p-methoxyphenylimino
esters using chiral copper(I) or palladium(II) catalysts were
a Isolated yield from the corresponding R-chloroglycinate (imino ester
was generated in situ). b Determined by chiral HPLC analysis. c At 0 °C.
d 5 mol % of the catalyst was used.
already reported, silyl enol ethers derived from ketones were
used in most cases and few examples using silyl enol ethers
derived from thioesters and esters are known.14 In the presence
of 10 mol % of the Cu(OTf)2-ligand 3e, N-acylimino ester 2a
was treated with 1-ethylthio-1-trimethylsiloxyethene (7a) in
dichloromethane at -20 °C. The reaction proceeded smoothly
to afford the corresponding adduct in high yield, but in low
selectivity (entry 1). The selectivity was improved when bulky
silyl enol ethers were used. When 1-tert-butylthio-1-tert-
butyldimethylsiloxyethene (7c) was used as a nucleophile and
the reaction was conducted in toluene at 0 °C, the desired
product was obtained in 85% yield with 90% ee (entry 5). Other
solvents such as dichloromethane, xylene, and mesitylene gave
lower selectivities. For R,R-disubstituted enol ethers, the ketene
silyl acetal (7e) derived from the corresponding phenyl ester
gave much better selectivity than the ketene silyl acetal (7d)
derived from the methyl ester. High yield and selectivity (90%
yield and 92% ee) were attained when 1-ethylthio-1-trimethyl-
silyl-2-methyl-1-propene (7f) was used (entry 10). The high
(13) For examples using aspartic acid derivatives as chiral synthons, see: (a)
Yoshida, T.; Takeshita, M.; Orita, H.; Kado, N.; Yasuda, S.; Kato, H.;
Itoh, Y. Chem. Pharm. Bull. 1996, 44, 1128. (b) Humphrey, J. M.; Bridges
R. J.; Hart, J. A.; Chamberlin, A. R. J. Org. Chem. 1994, 59, 2467. (c)
Dunn, P. J.; Ha¨ner, R.; Rapoport, H. J. Org. Chem. 1990, 55, 5017. (d)
McGarvey, G. J.; Williams, J. M.; Hiner, R. N.; Matsubara, Y.; Oh, T. J.
Am. Chem. Soc. 1986, 108, 4943.
(14) The asymmetric Mannich-type reaction with ester-derived ketene acetals
using a Cu(I)-BINAP catalyst was already reported by Lectka et al; however,
the enantioselectivity for simple ester-derived ketene silyl acetals was
moderate (∼72% ee) except for the coumarinone derived ketene silyl
acetals. See ref 9.
9
J. AM. CHEM. SOC. VOL. 125, NO. 9, 2003 2509

A R T I C L E S
Kobayashi et al.
Table 4. Mannich-type Reactions of N-Benzoylimino Esters
Table 5. Diastereo- and Enantioselective Mannich-type Reactions
with R-Monosubstituted Silyl Enol Ethers
a Isolated yield from the corresponding R-bromoglycinate (imino ester
was generated in situ). b Determined by chiral HPLC analysis. c At 0 °C.
d 20 mol % of the catalyst was used.
yield and selectivity were retained even when 5 mol % of the
catalyst was employed (entry 11).
Although wide substrate scope was observed in the present
Mannich-type reactions, lower selectivity was obtained in the
reaction of N-benzoylimino ester 2g. When 2g was treated with
the silyl enol ether derived from acetophenone (4a) in the
presence of 10 mol % of Cu(OTf)2-ligand 3e, the Mannich-
type adduct was produced in 95% yield with only 27% ee (Table
4, entry 1). On the other hand, the yield and selectivity were
improved when CuClO4‚4CH3CN-BINAP systems were used
as the catalysts. In particular, the desired product was obtained
in 79% yield with 97% ee in the presence of CuClO4‚
4CH3CN-(S)-xylyl-BINAP15 (10 mol %) (entry 4). As for
enolate components, silyl enol ethers derived from a ketone as
well as an ester and a thioester worked well to afford the
corresponding adducts in high yields with high ee’s.
Diastereo- and Enantioselective Reactions. Diastereo-
selection in this Mannich-type reaction is of great interest not
only from a synthetic point of view but also from a mechanistic
aspect. We investigated reactions of N-acylimino esters with
R-substituted silyl enol ethers in the presence of 10 mol % of
Cu(OTf)2-ligand 3e (Table 5). When the reaction of N-acylimino
ester 2a with (E)-1-tert-butylthio-1-trimethylsiloxy-1-propene
(9aE) was conducted in dichloromethane at 0 °C, the desired
Mannich-type adduct was obtained in 89% yield with high syn-
selectivity (syn/anti ) 96/4), and the enantiomeric excess of
the syn-adduct was 72% (entry 1). The geometrically isomeric
enolate, (Z)-1-tert-butylthio-1-trimethylsiloxy-1-propene (9aZ),
also reacted smoothly to give syn-adduct (syn/anti ) 96/4), and
the enantioselectivity of the adduct was improved to 90% ee
a Isolated yield from the corresponding R-chloroglycinate (imino ester
was generated in situ). b Determined by 1H NMR. c Ee of the major
diastereomer (determined by chiral HPLC analysis). d E/Z ) 94/6. e 2 equiv
of enolate were used. f E/Z ) 3/97. g E/Z ) 4/96. h E/Z ) 98/2. i Toluene
was used as a solvent. j Isolated yield from the corresponding R-hydroxy-
glycinate (R-halogenoglycinate was not isolated). k E/Z ) <1/>99. l PMP
) p-methoxyphenyl. m E/Z ) 8/92. n E/Z ) 76/24.
(entry 2). It is noted that syn-adducts were obtained from both
(E)- and (Z)-enolates with high selectivities. Although the use
of an S-ethyl group instead of the S-tert-butyl group decreased
both diastereo- and enantioselectivities (entry 3), tert-butyldi-
methylsilyl enol ethers in place of trimethylsilyl enol ethers gave
higher selectivities (entries 4 and 5). Regarding the solvents,
dichloromethane gave better selectivity than toluene in this case
(15) Provided from Takasago Chemical Co. Ltd. For the preparation of xylyl-
BINAP, see: Mashima, K.; Kusano, K.; Sato, N.; Matsumura, Y.; Nozaki,
K.; Kumobayashi, H.; Sayo, N.; Hori, Y.; Ishizaki, T.; Akutagawa, S.;
Takaya, H. J. Org. Chem. 1994, 59, 3064. The use of xylyl-BINAP, see,
e.g.: Ohkuma, T.; Koizumi, M.; Doucet, H.; Pham, T.; Kozawa, M.;
Murata, K.; Katayama, E.; Yokozawa, T.; Ikariya, T.; Noyori, R. J. Am.
Chem. Soc. 1998, 120, 13 529.
9
2510 J. AM. CHEM. SOC. VOL. 125, NO. 9, 2003

Catalytic, Asymmetric Mannich-type Reactions
A R T I C L E S
Scheme 1. Determination of the Absolute Configuration of 8b
Scheme 3. Absolute and Relative Stereochemistry Assignments
of 5a, 8h, 8i, and 10j
Scheme 2. Determination of the Relative Configuration of 10a
(entry 6). We tested other N-acylimino esters and the results
are summarized in entries 7-10. In all cases, the desired adducts
were obtained in high yields with good to high disatereo- and
enantioselectivities. We also examined reactions of silyl enol
ethers derived from ketones. Interestingly, syn-adducts were
obtained stereoselectively in most cases (entries 11-19). It is
noteworthy that 2-benzyloxy-1-phenyl-1-trimethylsiloxyethene
(9e) reacted with N-acylimino ester 2a with high syn-selectivity
and that the enantiomeric excess of the syn-adduct was high
(entry 14). In the reactions of the silyl enol ethers derived from
3-pentanone, the (Z)-tert-butyldimethylsilyl enol ether (9g) gave
the best result (entry 17).
pure 8i was synthesized by the N-acylation of (S)-asparatic acid
diethyl ester, and the absolute configuration of Mannich-adduct
8i was determined to be also R (eq 4). Also, the Mannich adduct
8h was converted to 8i and the absolute configuration was
determined to be R (eq 5).18 The absolute configuration of
Mannich adduct 10j-syn was also determined to be R by
converting optically active (R)-5f to a diastereomer mixture of
10j-syn and 10j-anti (eq 6). Relative stereochemical assignment
of 10j-syn and 10j-anti was made after converting to lactone
14 by NMR analysis (eq 7).19
In addition, we carefully conducted absolute and relative
stereochemistry assignments of Mannich-adduct 10i. It has been
reported that stereochemical courses were sometimes different
between R-methyl substituted and R-benzyloxy substituted
enolates.20 The ketone moiety of Mannich-adduct 10i was
reduced using lithium triethylborohydride (LiBEt3H), followed
by deprotection of the benzyl ether to afford diastereomeric triols
17 and 18 along with diol 19 (Scheme 4, eq 8). These
Absolute and Relative Stereochemistry Assignments. The
Mannich-type adducts obtained in this reaction were successfully
converted to intermediates of biologically important compounds,
and at the same time, absolute and relative stereochemistry
assignments were made. Mannich-type adduct 8b was reduced
to give diol 11 without epimerization. A natural aspartic acid
ester was converted to (S)-11, and comparison of optical
rotations revealed that the absolute configuration of 8b was R
(Scheme 1). On the other hand, Mannich-type adduct 10a was
reduced and the N-substituent was converted to a tert-butoxy-
carbonyl (Boc) group to afford diol 13, whose C-3 epimer was
known to be a promising substituent of quinolone antibacterial
agents.13a The relative stereochemistry was assigned to be syn
1
determined by H NMR analysis (see Scheme 2).16
(17) Ueno, M.; Kitagawa, H.; Ishitani, H.; Yasuda, S.; Nishijima, K.; Hanada,
K.; Kobayashi, S. Tetrahedron Lett. 2001, 42, 7863.
Furthermore, Mannich-type adduct 5a was transferred to
HPA-12 (vide infra) and its absolute configuration was deter-
mined to be R (Scheme 3, eq 3).17 Authentic enantiomerically
(18) Experimental details were shown in the Supporting Information.
(19) The NMR spectra of 14 was compared with one of known compound 15.
Details were shown in Supprting Information. (a) Barluenga, J.; Viado, A.
L.; Aguilar, E.; Fustero, S.; Olano, B. J. Org. Chem. 1993, 58, 5972. (b)
Gair, S.; Jackson, R. F. W.; Brown, P. A. Tetrahedron Lett. 1997, 38, 3059.
(20) (a)Kobayashi, S.; Ueno, M.; Ishitani, H. J. Am. Chem. Soc. 1998, 120,
431. (b)Mukaiyama, T.; Shiina, I.; Uchiro, H.; Kobayashi, S. Bull. Chem.
Soc. Jpn. 1994, 67, 1708.
(16) 1H NMR spectrum data for the anti-isomer of 13: See ref 13(a). 1H NMR
spectrum data for the diastereomeric mixture of 13: Hirabayashi, S.; Ike,
K.; Zanka, A.; Kawakami, T., Ichihara, M. World Intellectual Property
Organization Patent 1992, WO9220652.
9
J. AM. CHEM. SOC. VOL. 125, NO. 9, 2003 2511

A R T I C L E S
Kobayashi et al.
Table 6. Cu(I)-xylyl-BINAP Catalyzed Mannich-Type Reactions
Scheme 4. Absolute and Relative Stereochemistry Assignments
of 10i
entry
imine
Nu
catalyst
yield (%)
syn:antia
Ee(%)b
1
2
3
RdCOPh
9aZc
9aZc
9aEd
9aZc
9aZc
9aEd
A
B
B
A
B
B
76e
86e
94e
73f
83f
90f
89:11
78:22
36:64
20:80
<1:>99
<1:>99
20g
99
94
2
87
98
4
RdTs (2i)
5d
6d
1
a Determined by H NMR. b Ee of the major diastereomer (determined
by chiral HPLC analysis). c E/Z ) 3/97. d E/Z ) 94/6. e Isolated yield from
the corresponding R-bromoglycinate (imino ester was generated in situ).
f Isolated yield from N-Ts-imino ester. g Reverse enationselectivity was
observed.
compounds are the phytosphingosine analogues.21 On the other
hand, authentic sample 20 was synthesized according to
literature,22 and deprotection of the protective group under acidic
conditions and acylation made it possible to confirm the absolute
and relative stereochemistry of 17 (eq 9). Moreover, optically
pure 17 was reduced to afford optically pure authentic sample
19 (eq 10), and the absolute configuration of 19 was determined
by HPLC analysis. It is noted that syn-adducts with the same
absolute configuration were obtained using both R-methyl sub-
stituted and R-benzyloxy substituted enolates in this Mannich-
type reaction using chiral Cu(OTf)2-ligand 3e as the catalyst.
Use of Other Imines and Catalysts. The reaction of
N-benzoylimino ester 2g with (Z)-1-tert-butylthio-1-trimethyl-
siloxy-1-propene (9aZ) using Cu(OTf)2-ligand 3e proceeded in
76% yield with syn/anti ) 89/11, but the enantiomeric excess
of the syn-adduct was low (20% ee) (Table 6, entry 1). When
CuClO4‚4CH3CN-(S)-xylyl-BINAP (10 mol %) was used
instead of the copper (II) catalyst, the desired syn-adduct was
obtained preferentially (syn/anti ) 78/22), and the enantiomeric
excess of the syn-adduct was very high (99% ee) (entry 2).
Interestingly, the geometrically isomeric enolate, (E)-1-tert-
butylthio-1-trimethylsiloxy-1-propene (9aE), gave the anti-
adduct predominantly (syn/anti ) 36/64) with high enantiomeric
excess (entry 3). In the reactions with N-tosylimino ester 2i,
the CuClO4 system gave better results, and anti-adducts were
obtained from both (E)- and (Z)-enolates with high diastereo-
and enantioselectivities (entries 4-6).
reactions are summarized in Table 7, and in all cases, the
Mannich-type reactions proceeded cleanly. It is noteworthy that
this is the first example of a catalytic enantioselective Mannich-
type reaction using alkyl vinyl ethers.23
Mechanism and Catalyst Structure of the Mannich-type
Reactions. In the Mannich-type reaction of N-acylimino esters,
a mixture of the desired product 24 and vinyl ether 23 was
obtained after quenching the reaction with water (Scheme 5).
The vinyl ether 23 was converted to 24 by treatment with
hydrochloric acid. Because it is difficult to monitor this catalytic
reaction by NMR analysis due to the paramagnetic character
of Cu(II), we conducted the NMR study of the reaction without
the Cu catalyst to get information on the reaction mechanism
(R1 ) tBuMe2Si, R2 ) PMP). The reaction was found to proceed
at room temperature without the catalyst, and it was interesting
to find that no Mannich adduct 24 was observed in the 1H NMR
spectra of the reaction mixture. Instead, vinyl ether 23 and an
unidentified product were observed before quenching the
reaction with water. Contrary to the NMR study, TLC analysis
of the reaction mixture showed formation of both 23 and 24,
which were found to be stable on TLC. These experiments
indicate formation of a reactive intermediate, which was easily
transferred to 24 on TLC. We also confirmed that no N-silylated
compounds such as 25 were produced in this reaction.24 From
these experimental results and considerations, we assumed that
the reaction proceeded via [4 + 2] cycloaddition as shown in
Scheme 5.25 The initially formed cycloadduct 22 would be
partially transformed to vinyl ether 23 in situ via hydrogen
transfer, and then the resultant mixture of 22 and 23 would be
converted to the desired Mannich adduct 24 after acidic workup.
X-ray crystallographic analysis of asymmetric catalysts are
sometimes helpful to presume the geometry of a catalyst-
substrate complex as well as to interpret the sense of asymmetric
induction. Although X-ray analysis of the most effective catalyst
in this Mannich-type reaction composed of Cu(OTf)2 and
Vinyl Ethers. In the presence of Cu(OTf)2-ligand 3e (10 mol
%), alkyl vinyl ethers reacted with N-acylimino esters smoothly
to afford the corresponding N-acylated amino acid derivatives
in high yields with high selectivities. Several examples of the
(21) (a)Merrill, A. H., Jr.; Sweeley, C. C. In Biochemistry of Lipids, Lipoproteins,
and Membranes; Vance, D. E., Vance, J., Eds; Elsevier Science B. V.:
Amsterdam, 1996; pp 309-339. (b)Hannun, Y. A. Science 1996, 274, 1855.
See also, (c)Hanada, K.; Nishijima, M.; Kiso, M.; Hasegawa, A.; Fujita,
S.; Ogawa, T.; Akamatsu, Y. J. Biol. Chem. 1992, 267, 23 527. (d)
Kobayashi, S.; Furuta, T.; Hayashi, T.; Nishijima, M.; Hanada, K. J. Am.
Chem. Soc. 1998, 120, 908.
(23) For catalytic enantioselective aldol reactions using alkyl vinyl ethers, see
Carreira, E. M.; Lee, W.; Singer, R. A. J. Am. Chem. Soc. 1995, 117, 3649.
(24) Experimental details were shown in Supporting Information.
(22) Imashiro, R.; Sakurai, O.; Yamashita, T.; Horikawa, H. Tetrahedron 1998,
54, 10 657.
9
2512 J. AM. CHEM. SOC. VOL. 125, NO. 9, 2003

Catalytic, Asymmetric Mannich-type Reactions
A R T I C L E S
Table 7. Mannich-Type Reactions with Alkyl Vinyl Ethers
Scheme 5. Presumed Mechanism for the Mannich-type Reactions
Catalyzed by Cu(II)-Diamine Complex
a Isolated yield from the corresponding R-halogenoglycinate. b Deter-
mined by chiral HPLC analysis. c Isolated yield from the corresponding
R-hydroxyglycinate (R-halogenoglycinate was not isolated). d CuClO4‚
4CH3CN-(S)-xylyl-BINAP catalyst was used at -78 °C. e 3d was used
instead of 3e as a diamine ligand.
addition, the amide carbonyl oxygen may interact with the amine
hydrogen atom of the diamine ligand via hydrogen bonding.
To get more information about the structure of the catalyst-
imine complex, we monitored interaction of the imine with the
diamine-copper catalyst by an FT-IR spectrometer, which made
it possible to conduct the reaction under strictly anhydrous
conditions. As a result, the ester carbonyl stretch shifted as
expected, and also the band of the amide carbonyl group shifted,
which might indicate that the amide oxygen atom also coordi-
nated to the catalyst as the ester carbonyl oxygen atom did.
The CdN absorption was unclear and no information about the
coordination of the imine nitrogen atom was obtained. It is noted
diamine ligand 3e failed, X-ray structure for the complex of
Cu(ClO4)2‚6H2O and diamine 3d was successfully obtained26
after extensive screening of copper salts and diamine ligands.
The X-ray structure obtained is illustrated in Figure 2 (Structure
A),27 and it displays monomeric, bidentate coordination of the
diamine ligand to the copper center in a little distorted square-
planar geometry.28 Two molecules of water also bind to the
metal center in the distorted square-planar face, and two
perchlorate counteranions are arrayed in the apical position. It
is noteworthy that one of the oxygen atoms of the perchlorate
anions interacts with the N-hydrogen atom of the diamine ligand.
On the other hand, PM3 calculations29 suggested the lowest
energy structure of Cu(OTf)2-ligand 3e-N-acylimino ester 2b
complex (Figure 2, structure B).30 In this model, the Cu(II) has
a tetrahedral geometry,31 and both the ester carbonyl oxygen
and the imine nitrogen coordinate to the copper center. In
(30) Although the reliability of PM3 calculations for transition metal complexes
with partially filled d-orbitals is controversial, there are several examples
of PM3 calculations of Cu(II)-complexes. For examples, Evans et al. used
PM3 calculations for Cu(II)-bisoxazoline complexes, and optimized
structures were reported to be almost consistent with X-ray structures. They
also reasonably explained enantioselectivity in several catalytic asymmetric
reactions: Johnson, J. S.; Evans, D. A. Acc. Chem. Res. 2000, 33, 325,
and references therein.
(25) Similar reaction pathways were reported. (a)Gizecki, P.; Dhal, R.; Toupet,
L.; Dujardin, G. Org. Lett. 2000, 2, 585, and ref 2a. See also, (b)Kitajima,
H.; Katsuki, T. Synlett 1997, 568. See also, ref 12.
(31) Although Cu(II) complexes usually adopt a square planar geometry rather
than a tetrahedral geometry, several examples of tetrahedral Cu(II)
complexes have been reported. For recent examples, see: (a) Beheshti,
A.; Clegg, W.; Hyvadi, R.; Hekmat, H. F. Polyhedron 2002, 21, 1547.
Jørgensen et al. also proposed tetrahedral transition state models to explain
enantioselectivity in asymmetric hetero Diels-Alder Reactions and ene
reactions catalyzed by chiral Cu(II) complexes. (b) Johannsen, M.;
Jørgensen, K. A. J. Org. Chem. 1995, 60, 5757. (c) Johannsen, M.; Yao,
S.; Graven, A.; Jørgensen, K. A. Pure Appl. Chem. 1998, 70, 1117.
(26) The blue-color crystal of the complex composed of Cu(ClO4)2‚6H2O and
diamine 3d was grown in a THF-hexane solvent system.
(27) Two water molecules are abbreviated for clarification.
(28) The copper center in a little distorted square planar geometry was also
reported by Evans et al.: See ref 12.
(29) Stewart, J. J. P. J. Comput. Chem. 1989, 10, 209-221.
9
J. AM. CHEM. SOC. VOL. 125, NO. 9, 2003 2513

A R T I C L E S
Kobayashi et al.
Figure 3. Assumed transition state models for the diastereoselective
reactions.
Figure 4. Assumed cyclic transition state models.
syn-adduct, antiperiplanar transition state model TS-2 would
be the most reasonable to minimize the number of repulsive
interactions (Figure 3). On the contrary, TS-1 that affords the
anti-adduct has severe steric interaction between the methyl
group of the nucleophile and the chelated copper catalyst
(repulsion a), and synclinal transition state models are apparently
disfavored by steric repulsions between nucleophiles and the
chelated copper complex or N-acyl groups (COR). In this model
(TS-2), both (E)- and (Z)-enolates would provide the syn-adduct
stereoconvergently. On the other hand, cyclic transition state
models cannot explain the observed high syn-selectivity and the
stereoconvergency (Figure 4). The decrease of the syn-selectivity
for the bulky N-acyl group such as N-COiPr or N-Bz would
be rationalized by the increase of the steric repulsion between
the N-acyl group and the methyl group (repulsion b). In such
bulky N-acylimino esters, enantioselectivity was also decreased,
probably because the bulkiness of N-acylimine inhibits the
coordination to the catalyst. This consideration would be not
inconsistent with IR experiments that revealed almost no shift
of the carbonyl stretches in the mixture of N-Bz-imino ester
and the catalyst.
Facile Synthesis of HPA-12. Recently, our laboratories have
focused on (1R,3R)-N-(3-hydroxy-1-hydroxymethyl-3-phenyl-
propyl)dodecanamide (HPA-12, 1), a new inhibitor of ceramide
trafficking from endoplasmic reticulum to the site of sphingo-
myerin (SM) synthesis. HPA-12 is the first compound of a
specific inhibitor for SM synthesis in mammalian cells, and is
expected as a drug that inhibits intracellular trafficking of
sphingolipids.33 Although we have completed the first asym-
metric synthesis of HPA-12 using zirconium-catalyzed enantio-
selective Mannich-type reactions,34 we planned to develop a
more convenient way using the present Mannich-type reactions
with N-acylimino esters. The Mannich-type adduct (5a) was
Figure 2. X-ray structure of Cu(ClO4)2-ligand 3d (structure A), the assumed
structure of Cu(OTf)2-3e-2b calculated using PM3 (structure B), and a model
for understanding the enantioselectivity (structure C).
that these IR experimental results do not contradict the calcula-
tion model, structure B. In this model, the Si-face of the imine
was shielded by one of the naphthyl rings of the chiral catalyst,
and silyl enol ethers or vinyl ethers attacked the imine from
the opened Re-face (Figure 2, Structure 3).
For diastereoselectivity, the present Mannich-type reactions
using the Cu(II)-diamine complex exhibited high syn-selectivity
regardless of the enolate geometry. From the structure for the
catalyst-imino ester complex, we assume that the reaction
proceeds via a stepwise manner (Scheme 1). The observed syn-
preference can also be explained more reasonably by the
stepwise open transition state model than by the concerted cyclic
model.32 Among the possible transition states that lead to the
(32) Diastereoselections in additions of π-based nucleophiles (silyl enol ethers,
allylic metals, etc.) to aldehydes or ketones have been discussed very well
using open or cyclic transition state models. For aldol reactions, see: (a)
Mahrwald, R. Chem. ReV. 1999, 99, 1095, and references therein. For
allylations, see: (b) Yamamoto, Y.; Asao, N. Chem. ReV. 1993, 93, 2207,
and references therein. For effects of olefin geometry of nucleophiles on
the diastereoselection, see: (c) Keck, G. E.; Savin, K. A.; Cressman, E. N.
K.; Abbott, D. E. J. Org. Chem. 1994, 59, 7889. (d) Roush, W. R.; Ando,
K.; Powers, D. B.; Palkowitz, A. D.; Halterman, R. L. J. Am. Chem. Soc.
1990, 112, 6339.
(33) Yasuda, S.; Kitagawa, H.; Ueno, M.; Ishitani, H.; Fukasawa, M.; Nishijima,
M.; Kobayashi, S.; Hanada, K. J. Biol. Chem. 2001, 276, 43 994.
9
2514 J. AM. CHEM. SOC. VOL. 125, NO. 9, 2003

Catalytic, Asymmetric Mannich-type Reactions
A R T I C L E S
Table 8. Synthesis of HPA-12 (Anti-selective Reduction of
â-Carbamoyl Ketone)
noted that HPA-12 has been synthesized from 2a in three steps
(two-pot), and that the total yield was 82.9%.36 Gram-scale
preparation has also been performed successfully, and the
preparation of many other HPA-12 analogues is feasible
according to this efficient synthetic pathway.
Conclusion
We have developed catalytic asymmetric Mannich-type
reactions of N-acylimino esters with silyl enol ethers or alkyl
vinyl ethers using Cu(OTf)2-chiral diamine complexes as the
catalysts. The reactions proceeded smoothly at 0 °C in most
cases, and high yields and high diastereo- and enantioselectivi-
ties were attained. On the basis of X-ray crystallographic
analysis of the chiral Cu(II) catalyst, PM3 calculations, and FT-
IR analyses, we assumed the reaction mechanism, the most
stable chiral catalyst-N-acylimino ester complex, and the transi-
tion states of these Mannich-type reactions. Moreover, several
useful intermediates for the synthesis of biologically important
compounds were prepared by using these novel Mannich
reactions. In particular, HPA-12 (1), a new inhibitor of ceramide
trafficking from endoplasmic reticulum to the site of sphingo-
myerin (SM) synthesis, has been synthesized in three steps in
82.9% total yield using the present Cu(II)-catalyzed Mannich-
type reaction.
entry
solvent
reagent-1
reagent-2a
yield (%)
syn:anti
1
2
3
4
5
6
7
8
9
THF
THF
THF
THF
THF
THF
THF
THF
9-BBNb
none
none
none
none
0
0
LS-Selectridec,d
LiBEt3He
63
86
92
96h
99h
99
90
90
92
95
55:45
65:35
41:59
53:47
53:47
9:91
44:56
60:40
19:81
16:84
LiAl(OtBu)3Hb
L-Selectridef
k
LiBH4
k
(S)-Alpine-Hydrideb,g LiBH4
(R)-Alpine-Hydrideb,g LiBH4
k
K-Selectrideb
LiBEt3Hl
LiBEt3Hl
LiBEt3Hl
LiBEt3Hl
LiBEt3Hm
b
MeOH NaBH4
10
11
12
Et2O
DME
DME
K-Selectrideb
K-Selectridei
K-Selectridej
a At rt. b At -78 °C for 2 h. c LS-Selectride: lithium triisoamylboro-
hydride. d At -78 °C for 26 h. e At -78 °C for 20 min. f At -78 °C for 1
h. g Alpine-Hydride: lithium B-isopinocampheyl-9-bora-bicyclo[3.3.1]nonyl
hydride. h Containing a small amount of unknown compounds. i At -45
°C for 2 h. j At -45 °C for 15 min. k 2 h. l 1 h. m 45 min.
obtained from N-acylimino ester 2a with 1-trimethylsiloxy-1-
phenylethene (3a) or 1-methoxy-1-phenylethene (21a) using the
chiral Cu(II) catalyst in high yield and selectivity. We then
examined reduction conditions of 5a to afford HPA-12 directly.
Although 9-bora-bicyclo[3.3.1]nonyl hydride (9-BBN) or lithium
triisoamylborohydride (LS-Selectride) was not effective, LiBEt3H
or tri-tert-butoxyaluminum hydride gave the desired product
directly, but diastereomer ratios (syn/anti) were moderate. After
examination of several reducing agents, it was revealed that
combination of K-Selectride and LiBEt3H in THF gave the
best result, and that HPA-12 was obtained in 99% yield with
high diastereoselectivity (syn/anti ) 9/91). Interestingly, the
solvents had significant effects on the selectivity, and moderate
selectivity was obtained in ether (syn/anti ) 60/40). The use of
a bulky reducing agent in a polar solvent prevented a cyclic
intermediate in which both the ketone carbonyl oxygen and the
amide nitrogen coordinated to the reducing agent, leading to
anti-selective reduction of the â-amino ketone.35 It should be
Acknowledgment. This work was partially supported by
CREST and SORST, Japan Science and Technology Corpora-
tion (JST), and a Grant-in-Aid for Scientific Research from
Japan Society of the Promotion of Science. The authors thank
Drs. Kentaro Hanada, Satoshi Yasuda, and Masahiro Nishijima
(National Institute of Infectious Diseases) for their helpful
discussion on biological aspects of HPA-12. The authors are
also grateful to Dr. Kei Manabe and Mr. Tomoaki Hamada (The
University of Tokyo) for their helpful discussion concerning
the copper-diamine catalyst.
Note Added after ASAP Publication: The version pub-
lished on the Web 2/6/2003 contained an error in entry 8 of
Table 5. The final Web version published 2/18/2003 and the
print version are correct.
Supporting Information Available: Experimental section
(PDF) and X-ray data of Figure 2, Structure A. This material is
available free of charge via the Internet at http://pubs.acs.org.
JA0281840
(34) (a) Ueno, M.; Kitagawa, M.; Ishitani, H.; Yasuda, S.; Hanada, K.;
Kobayashi, S. Tetrahedron Lett. 2001, 42, 7863. (b) Ishitani, H.; Ueno,
M.; Kobayashi, S. J. Am. Chem. Soc. 1997, 119, 7153. (c) Ishitani, H.;
Ueno, M.; Kobayashi, S. J. Am. Chem. Soc. 2000, 122, 8180. (d) Kobayashi,
S.; Ishitani, H.; Yamashita, Y.; Ueno, M.; Shimizu, H. Tetrahedron 2001,
57, 861. See also, 20(a).
(35) Pilli, R. A.; Russowsky, D.; Dias, L. C. J. Chem. Soc., Perkin Trans. 1
1990, 1213. Syn-selective reduction of N-acyl-â-amino ketones has been
reported recently. Keck, G. E.; Truong, A. P. Org. Lett. 2002, 4, 3131.
(36) Direct recrystallization gave diastereo- and enantiomerically pure HPA-
12. See Supporting Information.
9
J. AM. CHEM. SOC. VOL. 125, NO. 9, 2003 2515