Diastereoselective aldol reaction of an α-alkoxycarbonylamino aldehyde with a silyl enol ether

Article abstract of DOI:10.1016/S0957-4166(99)00324-9

The aldol reaction of optically active α-alkoxycarbonylamino aldehydes with a silyl enol ether in the presence of a Lewis acid afforded γ-amino-β- hydroxyketones diastereoselectively. The effect of the α-amide proton on the diastereoselectivity is discussed.

Full text of DOI:10.1016/S0957-4166(99)00324-9

Pergamon
Tetrahedron: Asymmetry 10 (1999) 3443–3448
Diastereoselective aldol reaction of an α-alkoxycarbonylamino
aldehyde with a silyl enol ether
Yasuki Yamada, Eiji Shirakawa,† Koji Ando, Saizo Shibata and Itsuo Uchida ∗
Central Pharmaceutical Research Institute, Japan Tobacco Inc., 1-1 Murasaki-cho, Takatsuki, Osaka 569-1125, Japan
Received 27 July 1999; accepted 29 July 1999
Abstract
The aldol reaction of optically active α-alkoxycarbonylamino aldehydes with a silyl enol ether in the presence
of a Lewis acid afforded γ-amino-β-hydroxyketones diastereoselectively. The effect of the α-amide proton on the
diastereoselectivity is discussed. © 1999 Elsevier Science Ltd. All rights reserved.
1. Introduction
Optically active β-amino alcohols are biologically interesting compounds which are widely used in
the inhibitors of aspartic proteases such as renin and HIV-proteases.1 We have already reported a series
of potent renin inhibitors which contain a (2S,3S,5S)-2-amino-1-cyclohexyl-6-methyl-3,5-heptanediol
fragment (1) as a transition state mimic in the renin-catalyzed reaction.2
Since aminodiol 1 could be prepared from syn-γ-amino-β-hydroxyketone 4a with a high stereoselec-
tivity by use of the Evans’ reduction method,3 the diastereoselective synthesis of 4a was explored for the
preparation of 1. Thus, we proposed the reaction of optically active α-alkoxycarbonylamino aldehyde
2a with silyl enol ether 3 in the presence of a Lewis acid, which was obtained from the Mukaiyama
stereoselective aldol reaction,4 as shown in Scheme 1. Here we report the stereoselective aldol reaction,
and we also discuss a mechanism to explain the stereoselectivity.
∗
†
Corresponding author. E-mail: itsuo.uchida@ims.jti.co.jp
New address: Graduate School of Materials Science, Japan Advanced Institute of Science and Technology, Asahidai,
Tatsunokuchi, Ishikawa 923-1292, Japan.
0957-4166/99/$ - see front matter © 1999 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(99)00324-9

3444
Y. Yamada et al. / Tetrahedron: Asymmetry 10 (1999) 3443–3448
Scheme 1.
2. Results and discussion
The results of the reaction of N-tert-butoxycarbonyl-L-cyclohexylalaninal 2a5 with O-trimethylsilyl
enolate 36 in the presence of various Lewis acids (1.2 equiv.) are summarized in Table 1. The aldol
reaction of 2a with 3 in dichloromethane at −78°C using titanium(IV) chloride (TiCl4) as the Lewis acid
afforded the desired syn-adduct 4a along with the anti-isomer 5a in a ratio of 80 to 20 (run 1).7 The use
of tin(IV) chloride (SnCl4) also gave similar stereoselectivity as TiCl4 (run 2).8 When boron trifluoride
etherate (BF3·OEt2) was used as a Lewis acid, both the stereoselectivity and yield increased (runs 3 and
4).
Table 1
The aldol reaction of aldehyde 2 with O-trimethylsilyl enol ether 3 in the presence of a Lewis acid
In the aldol reaction of 2a mediated by TiCl4 or SnCl4, the observed syn-selectivity can be rationalized
in terms of the chelation-controlled mechanism (Fig. 1, (A)) as suggested by Terashima et al.8 However,
BF3·OEt2, which is incapable of chelation,9 gave 4a with a higher stereoselectivity. We propose that
this selectivity is explained by the hydrogen bonding interaction between the carbonyl oxygen of the
aldehyde and the α-NH proton (Fig. 1, (B)).10 Thus, nucleophile 3 approaches from the less hindered
side of the hydrogen bonded five-membered ring, giving the syn-adduct 4a. To confirm the effect of
the α-NH proton on the diastereoselectivity, we examined the aldol reaction of N-methylated aldehyde

Y. Yamada et al. / Tetrahedron: Asymmetry 10 (1999) 3443–3448
3445
Figure 1.
2b mediated by BF3·OEt2, and found an inversion of the stereoselectivity (run 5).11 This result can be
rationalized by the Felkin–Anh model (Fig. 1, (C)). Interestingly, in the aldol reaction of 2b with SnCl4,
the ‘Felkin–Anh’ product 5b was obtained in preference to the ‘chelation-controlled’ product 4b (run 6).
To gain more insight into the hydrogen bonding model described above, we observed the conforma-
tional changes in aldehydes 2a and 2b induced by the addition of BF3·OEt2 at −78°C in CD2Cl2 with
1
NMR spectroscopy. The H and 13C NMR data are summarized in Table 2. By addition of 1 equiv.
of BF3·OEt2, the carbamate (NC(_O)O) and the tert-butyl (OC(CH3)3) carbon signals were shifted to
downfield, and the signals derived from noncomplexed ether (14.0, 65.0 ppm in 13C NMR) appeared
in both the NMR spectra of 2a and 2b. These results suggest that 1 equiv. of BF3 is quantitatively
coordinated by the carbamate oxygen of the substrate. From our experimental results, the aldol reaction
of 2a with less than 1 equiv. of BF3·OEt2 did not take place at all: the activation of the aldehyde carbonyl
group by excess BF3 is thought to be required to promote the reaction. However, the downfield shifts of
the aldehyde proton or carbon signals implying the complexation of the aldehyde oxygen of 2a with BF3
were not detected even after addition of 2 equiv. of BF3·OEt2 in the NMR experiment. The life-time of
the transition state, in which the aldehyde oxygen of 2a coordinates with BF3, may be too short to be
detected by NMR. Instead, the different behavior of the aldehyde group was observed between 2a and
2b in the NMR studies. By exposure of N-methylated substrate 2b to BF3·OEt2, the proton and carbon
signals of aldehyde disappeared, and the signals corresponding to an acetal function appeared. This result
indicates that BF3 promoted the conversion of 2b to its trimer, paraldehyde.12 The low yield observed in
the aldol reaction of 2b (Table 1) may be due to this paraldehyde formation. In contrast, no trimerization
was observed in the mixture of 2a and BF3·OEt2. From these results, it is suggested that the aldehyde
carbonyl group of 2a is reluctant to be converted to its trimer since it forms a hydrogen bond with the
α-NH proton.
Aminodiol 1 was prepared as shown in Scheme 2. The reduction of hydroxyketone 4a using tetra-
methylammonium borohydride in propionic acid afforded anti-diol 6 with a high stereoselectivity (84%
de).3 After recrystallization, enantiomerically pure 6 was obtained in 34% yield starting from the crude
mixture of 4a and 5a produced under the conditions of run 3 in Table 1. The Boc group of 6 was removed
by treatment with trifluoroacetic acid to give aminodiol 1.
3. Conclusions
The aldol reaction of optically active α-alkoxycarbonylamino aldehyde 2a with silyl enol ether
in the presence of a Lewis acid afforded syn-γ-amino-β-hydroxyketone 4a diastereoselectively. The

3446
Y. Yamada et al. / Tetrahedron: Asymmetry 10 (1999) 3443–3448
Table 2
1H and 13C NMR data of 2a and 2b and their complexes with BF3·OEt2
Scheme 2.
syn-selectivity with BF3·OEt2 as a Lewis acid can be explained by the transition state model of the
intramolecular hydrogen bonding formed between the α-amide proton and the aldehyde oxygen of 2a.
4. Experimental
4.1. General
Melting points were determined on a Buchi 535 melting point apparatus. Infrared (IR) spectra were
recorded on a Jasco IR700 IR spectrometer or on a Perkin–Elmer 1650 FT-IR spectrometer. Optical
rotations were measured with a Perkin–Elmer 241 polarimeter. Nuclear magnetic resonance (NMR)
spectra were measured on Bruker AMX300 (300 MHz), ARX400 (400 MHz) and Jeol JNMA300 (300
MHz) instruments. Chemical shifts are reported as δ values (parts per million relative to tetramethylsilane
as an internal standard. Fast atom bombardment mass spectra (FABMS) were obtained with a Finnigan
MAT TSQ700 mass spectrometer or with a Jeol SX102 mass spectrometer. Elemental analyses were
performed by the Analytical Group, Central Pharmaceutical Research Institute, Japan Tobacco Inc. Thin-
layer chromatography (TLC) was carried out using Merck precoated silica gel 60 F-254 plates (thickness

Y. Yamada et al. / Tetrahedron: Asymmetry 10 (1999) 3443–3448
3447
0.25 mm). Preparative TLC was carried out using Merck precoated silica gel 60 F-254 plates (thickness
0.5–1.0 mm). Column chromatography was carried out using Merck silica gel 60 (70–230 mesh or
230–400 mesh).
4.2. General procedure for the aldol reactions
A solution of N-Boc-cyclohexylalaninal 2a (1.0 mmol) in CH2Cl2 (10 ml) was treated with a Lewis
acid (1.2 mmol). After being stirred for 10 min, silyl enolate 3 (2.0 mmol) was added and the mixture
was stirred for 1 h. The mixture was poured into a mixture of ice and saturated aqueous NaHCO3, and
extracted with ether. The organic layer was washed with brine, dried over MgSO4 and evaporated. The
residue was separated by silica gel chromatography with a mixture of AcOEt and hexane (1:9) to give
the analytical samples of hydroxyketone 4a and 5a.
4.2.1. (5S,6S)-6-(tert-Butoxycarbonyl)amino-7-cyclohexyl-2-methyl-5-hydroxy-3-heptanone 4a
1
A pale yellow oil. Rf 0.40 (AcOEt:hexane=3:7). H NMR (CDCl3) δ: 0.8–1.9 (m, 13H), 1.11 (d,
6H, J=6.9 Hz, -CH(CH3)2), 1.45 (s, 9H, -C(CH3)3), 2.5–2.8 (m, 3H), 3.40 (d, 1H, J=2.3 Hz), 3.61 (m,
1H), 4.02 (m, 1H), 4.70 (br d, 1H, J=9.9 Hz, -CONH-). FABMS m/z: [M+H]+ calcd for C19H36NO4:
342.2646. Found: 342.2649.
4.2.2. (5R,6S)-6-(tert-Butoxycarbonyl)amino-7-cyclohexyl-2-methyl-5-hydroxy-3-heptanone 5a
White solid. Rf 0.34 (AcOEt:hexane=3:7). Mp 94–96°C. 1H NMR (CDCl3) δ: 0.7–2.0 (m, 13H), 1.10
(d, 6H, J=6.9 Hz, -CH(CH3)2), 1.44 (s, 9H, -C(CH3)3), 2.5–2.8 (m, 3H), 3.40 (br d, 1H), 3.63 (m, 1H),
3.97 (m, 1H), 4.55 (br d, 1H, -CONH-). Anal. calcd for C19H35NO4: C, 66.83; H, 10.33; N, 4.10. Found:
C, 66.96; H, 10.49; N, 4.11.
4.3. (2S,3S,5S)-2-(tert-Butoxycarbonyl)amino-1-cyclohexyl-6-methyl-3,5-heptanediol 6
To propionic acid (200 ml) was added tetramethylammonium borohydride (10 g, 113 mmol) by
portions at 0°C, and the mixture was stirred for 1 h. To the mixture was added dropwise a solution of crude
4a (9.62 g, 28 mmol), which was prepared under the conditions of run 3 in Table 1, in propionic acid (50
ml). After being stirred for 2 h at 0°C, the mixture was poured into a mixture of ice and saturated aqueous
NH4Cl and extracted with CH2Cl2 (500 ml). The organic layer was washed with aqueous 2N NaOH
(200 ml×3) and brine (300 ml×2), dried over MgSO4, then evaporated. The residue was recrystallized
from hexane:AcOEt (10:1) to give 6 (3.3 g, 34%) as white crystals. Rf 0.63 (AcOEt:hexane=1:1). Mp
148–150°C. IR (KBr) cm−1: 1716. [α]D −54.3 (c=1.05, MeOH). H NMR (CDCl3) δ: 0.8–1.9 (m,
16H), 0.90 (d, 3H, J=6.8 Hz, -CH(CH3)2), 0.94 (d, 3H, J=6.7 Hz, -CH(CH3)2), 1.45 (s, 9H, -C(CH3)3),
2.47 (br d, 1H, -OH), 2.68 (br d, 1H, -OH), 3.59 (m, 1H), 3.69 (m, 1H), 3.83 (m, 1H), 4.64 (br d, 1H,
J=9.1 Hz, -CONH-). Anal. calcd for C19H37NO4: C, 66.44; H, 10.86; N, 4.08. Found: C, 66.55; H, 11.09;
N, 4.31.
25
1
4.4. (2S,3S,5S)-2-Amino-1-cyclohexyl-6-methyl-3,5-heptanediol 1
A solution of 6 (2.44 g, 7.1 mmol) in trifluoroacetic acid (50 ml) was stirred for 30 min at room
temperature, then concentrated. Saturated aqueous NaHCO3 (50 ml) was added to the residue, and the
mixture was extracted with CHCl3 (50 ml×3). The combined organic layer was dried over MgSO4 and
evaporated. The residue was purified by silica gel chromatography with a mixture of CHCl3, MeOH

3448
Y. Yamada et al. / Tetrahedron: Asymmetry 10 (1999) 3443–3448
25
and NH4OH (95:5:0.5) to give 1 (1.59 g, 92%) as white crystals. Mp 88–90°C. [α]D −53.3 (c=1.02,
MeOH). 1H NMR (CDCl3) δ: 0.8–1.9 (m, 16H), 0.91 (d, 3H, J=6.8 Hz, -CH(CH3)2), 0.95 (d, 3H, J=6.7
Hz, -CH(CH3)2), 2.75 (ddd, 1H, J=3.3, 6.6, 9.9 Hz, NH2CH<), 3.51 (dt, 1H, J=3.8, 7.0 Hz), 3.61 (ddd,
1H, J=2.8, 6.1, 8.9 Hz). Anal. calcd for C14H29NO2: C, 69.09; H, 12.01; N, 5.75. Found: C, 69.08; H,
11.99; N, 5.96.
References
1. (a) For a review of renin inhibitors, see: Ocain, T. D.; Abou-Gharbia, M. Drugs Fut. 1991, 16, 37. (b) For a review of
HIV-protease inhibitors, see: Babine, R. E.; Bender, S. L. Chem. Rev. 1997, 97, 1359.
2. (a) Yamada, Y.; Ando, K.; Ikemoto, Y.; Tada, H.; Shirakawa, E.; Inagaki, E.; Shibata, S.; Nakamura, I.; Hayashi, Y.;
Ikegami, K.; Uchida, I. Chem. Pharm. Bull. 1997, 45, 1631. (b) Yamada, Y.; Ando, K.; Komiyama, K.; Shibata, S.;
Nakamura, I.; Hayashi, Y.; Ikegami, K.; Uchida, I. Bio. Med. Chem. Lett. 1997, 7, 1863.
3. Evans, D. A.; Chapman, K. T.; Carreria, E. M. J. Am. Chem. Soc. 1988, 110, 3560.
4. Mukaiyama, T. Org. React. 1982, 28, 203.
5. Iizuka, K.; Kamijo, T.; Harada, H.; Akahane, K.; Kubota, T.; Umeyama, H.; Ishida, T.; Kiso, Y. J. Med. Chem. 1990, 33,
2707.
6. Amice, P.; Blanco, L.; Conia, J. M. Synthesis 1976, 196.
7. The configurations of 4a were determined by the procedure described in Ref. 2a.
8. Similar results were reported on the aldol reaction using trimethylsilyl ketene acetal: Takemoto, Y.; Matsumoto, T.; Ito, Y.;
Terashima, S. Chem. Pharm. Bull. 1991, 39, 2425.
9. Reetz, M. T. Angew. Chem., Int. Ed. Engl. 1984, 23, 556.
10. The hydrogen bonding transition state model has been proposed to explain some stereoselective reactions of α-amino
ketones and aldehydes: (a) Hartley, D. Chem. Ind. 1981, 551. (b) Jurczak, J.; Goxebiowski, A.; Raczko, J. Tetrahedron
Lett. 1988, 29, 5975.
11. The configurations of 4b and 5b were determined as follows. The acetylation of the hydroxy group of 4b (5b) gave the
compound identical with that prepared from 4a (5a) by the sequential N-methylation and O-acetylation.
12. Trimerization of an aldehyde induced by BF3·OEt2 has been reported: Denmark, S. E.; Wilson, T.; Willson, T. M. J. Am.
Chem. Soc. 1988, 110, 984.