Transition Metal/Base-Catalyzed Aldol Reactions of Isocyanoacetic Acid Derivatives with Prochiral Ketones, a Straightforward Approach to Stereochemically Defined β,β-Disubstituted-β-hydroxy-α-amino Acids. 1 Scope and Limitations

Article abstract of DOI:10.1021/jo9623402

A systematic study of the transition metal/base-catalyzed aldol reactions of methyl isocyanoacetate with a wide range of prochiral ketones, giving rise to the 4-(methoxycarbonyl)-5,5-disubstituted-2-oxazolines, has been made. The diastereoselectivity of t

Full text of DOI:10.1021/jo9623402

3470
J . Org. Chem. 1997, 62, 3470-3479
Tr a n sition Meta l/Ba se-Ca ta lyzed Ald ol Rea ction s of Isocya n oa cetic
Acid Der iva tives w ith P r och ir a l Keton es, a Str a igh tfor w a r d
Ap p r oa ch to Ster eoch em ica lly Defin ed
â,â-Disu bstitu ted -â-h yd r oxy-r-a m in o Acid s.1 Scop e a n d Lim ita tion s
Vadim A. Soloshonok,*,† Alexey D. Kacharov,‡ Dimitry V. Avilov,‡ Kohki Ishikawa,§
Nobuya Nagashima§ and Tamio Hayashi*,|
National Industrial Research Institute of Nagoya, Hirate-cho 1-1, Kita-ku, Nagoya City,
Aichi Prefecture 462, J apan, Institute of Bioorganic Chemistry and Petrochemistry,
Ukrainian Academy of Sciences, Kiev 253160, Ukraine, Central Research Laboratories of
Ajinomoto Co., Inc., Suzukicho, Kawasaki 210, J apan, and Department of Chemistry,
Faculty of Science, Kyoto University, Sakyo, Kyoto 606-01, J apan
Received December 13, 1996X
A systematic study of the transition metal/base-catalyzed aldol reactions of methyl isocyanoacetate
with a wide range of prochiral ketones, giving rise to the 4-(methoxycarbonyl)-5,5-disubstituted-
2-oxazolines, has been made. The diastereoselectivity of these reactions was found to be influenced
by a mix of several factors, including steric characteristic of the substituents, nature of the catalyst,
electrostatic, and electron donor-acceptor type interactions. The former factor, a stereochemical
discrimination between the substituents at the ketone carbon, was shown to be the most pronounced
in controlling of the stereochemical outcome, which could be markedly improved with a proper
choice of the catalyst. In particular, for the reactions of methyl isocyanoacetate with polyhaloal-
kyl aryl(alkyl) ketones, high diastereoselectivity (80-98% de) was achieved, thus allowing for
straightforward and generalized access to the corresponding (2R*,3R*)-â,â-disubstituted-â-hydroxy-
R-amino carboxylic acid.
In tr od u ction
reactions. Apart from the methodological challenge, aldol
reactions between the nucleophilic glycine equivalents
and prochiral ketones attract our attention also because
this approach would provide the most straightforward
access to the sterically constrained â,â-disubstituted-â-
hydroxy-R-amino carboxylic acids, which are of exciting
potential interest in the de novo design of peptides and
peptidomimetics with specific conformational and physi-
ological properties.5
Recently we have disclosed that trifluoromethyl ke-
tones react with a Ni(II) complex of the Schiff base of
glycine with (S)-o-[N-(N-benzylprolyl)amino]benzophe-
none to afford the targeted aldol products in a highly
diastereoselective manner (90-98% de) and in respected
chemical yields (56-87%).6 Unfortunately, the scope of
this approach was found to be limited to the reactions of
the trifluoromethyl n-alkyl ketones, while sterically
demanding aryl and sec-alkyl trifluoromethyl ketones and
less electrophilic fluorine-free ketones failed to react with
the Ni(II) complex.6c
Over the past 15 years the art of stereoselective aldol
methodology has been advanced to the state that could
rival enzymatic systems for the synthetic efficiency and
control over relative and absolute stereochemistry of the
products.2 However, the overwhelming majority of the
strategies developed in this area are suitable for the
additions to aldehydes, while the involvement of prochiral
ketones in the stereoselective aldol reactions has yet to
meet with such success.2,3 In particular, highly diaste-
reoselective condensations between glycine R-anion equiva-
lents and prochiral ketones are virtually unknown.4 It
was recognized that poor electrophilicity and steric
shielding of the carbonyl group, as well as greater
difficulty in the stereodiscrimination between the two
carbon-containing substituents in ketones, impose serious
limitations on the control of chemo-, regio-, and stereo-
selective outcomes of the corresponding aldol addition
† National Industrial Research Institute of Nagoya.
‡ Ukranian Academy of Sciences.
(4) For recent comprehensive reviews on asymmetric synthesis of
R-amino acids, see: (a) Williams, R. M. Synthesis of Optically Active
R-Amino Acids. In Organic Chemistry Series; Baldwin, J . E., Magnus,
P. D., Eds.; Pergamon Press: Oxford, 1989; Vol. 7. (b) Duthaler, R. P.
Tetrahedron 1994, 50, 1539.
§ Central Research Laboratories of Ajinomoto Co.
| Kyoto University.
X Abstract published in Advance ACS Abstracts, May 1, 1997.
(1) For preliminary communications, see: (a) Soloshonok, V. A.;
Hayashi, T.; Ishikawa, K.; Nagashima, N. Tetrahedron Lett. 1994, 35,
1055. (b) Soloshonok, V. A.; Kacharov, A. D.; Avilov, D. V.; Hayashi,
T. Tetrahedron Lett. 1996, 37, 7845.
(5) (a) Peptide Chemistry: Design and Synthesis of Peptides, Con-
formational Analysis and Biological Functions; Hruby V. J .; Schwyzer,
R., Eds.; Tetrahedron-Symposia-in-Print, 31; Tetrahedron 1988, 44,
661. (b) Giannis, A.; Kolter, T. Angew. Chem., Int. Ed. Engl. 1993,
32, 1244. (c) Robl, J . A.; Cimarusti, M. P.; Simpkins, L. M.; Weller,
H. N.; Pan, Y. Y.; Malley, M.; DiMarco, J . D. J . Am. Chem. Soc. 1994,
116, 2348. (d) For general review on fluorine-containing amino acids,
see: Fluorine-Containing Amino Acids. Synthesis and Properties;
Kukhar, V. P.; Soloshonok, V. A., Eds.; Wiley: Chichester, 1994.
R-Amino-â-hydroxy carboxylic acids are naturally occurring com-
pounds. Some are proteinogenic and/or essential amino acids are
involved in various physiological processes in living organisms;
therefore, analogs of these amino acids are very interesting as
biomedical tools or potential inhibitors. Other R-amino-â-hydroxy
carboxylic acids are constituents of complex natural products such as
cyclosporin, bouvardin, peptides, and glycopeptides, which usually
possess high antibiotic activity. In addition, these amino acids are
valuable precursors to â-lactam antibiotics.
(2) For reviews of stereoselective aldol addition reactions, see: (a)
Seebach, D. Angew. Chem. Int. Ed. Engl. 1990, 29, 1320. (b) Evans,
D. A. Aldrichimica Acta 1982, 15, 23. (c) Evans, D. A.; Nelson, J . V.;
Taber, T. R. In Topics in Stereochemistry; Eliel, E. L., Allinger, N. L.,
Wilen, S. H., Eds.; Wiley Interscience: New York, 1982; Vol. 13, p 1.
(d) Heathcock, C. H. In Asymmetric Synthesis; Morrison, J . D., Ed.;
Academic Press: New York, 1984; Vol. 3, Chapter 2. (e) Kim, B. M.;
Williams, S. F.; Masamune, S. In Comprehensive Organic Synthesis.
Additions to C-X π Bonds; Heathcock, C. H., Ed.; Pergamon Press:
Oxford; 1991; Vol. 2, Part 2, pp 239-275. (f) Gennari, C. In Compre-
hensive Organic Synthesis. Additions to C-X π Bonds; Heathcock, C.
H., Ed.; Pergamon Press: Oxford, 1991; Vol. 2, Part 2, pp 629-660.
(g) Bach, T. Angew. Chem., Int. Ed. Engl. 1994, 33, 417. (h) Franklin,
A. S.; Paterson, I. Contemp. Org. Synth. 1994, 1, 317-416.
(3) Braun, M. Angew. Chem., Int. Ed. Engl. 1987, 26, 24.
S0022-3263(96)02340-7 CCC: $14.00 © 1997 American Chemical Society

Transition Metal/Base-Catalyzed Aldol Reactions
J . Org. Chem., Vol. 62, No. 11, 1997 3471
Sch em e 1
Our experience in the aldol reactions clearly revealed
that one serious obstacle to the successful condensation
between a nucleophilic glycine equivalent and a ketone
is the reversibility of the reaction, the equilibrium of
which nearly entirely lies on the side of starting com-
pounds. From this point of view, aldol-type reactions
between isocyanoacetic acid derivatives and carbonyl
compounds, one of the useful methods for preparing
â-hydroxy-R-amino acids, hold an apparent advantage,
as these condensations could proceed irreversibly, giving
rise to the corresponding oxazolinecarboxylates.7 More-
over, these reactions were shown to be effectively cata-
lyzed by certain transition metals that could allow for
the rational influence of the regio- and diastereochemical
outcomes of these transformations.8 This study gains
also an additional impetus when considering the fact that
for aldehyde series this method was brought to a high
level of sophistication with the development of catalytic
enantioselective synthesis of â-hydroxy-R-amino acids by
use of a well-designed chiral ferrocenylbisphosphine-
gold(I)9,10 or -silver(I)11 catalyst.
For the reactions of isocyanoacetic acid derivatives with
prochiral ketones, only a handful of examples are ex-
tant.8,12 It was reported that certain alkyl R,â-alkenyl
and R-chloroalkyl alkyl(aryl) ketones in the presence of
Cu(I) (Cu2O, CuCl) smoothly react with the ethyl isocy-
anoacetate to afford a mixture of the corresponding trans/
cis-oxazolines with diastereoselectivity ranging from 0 to
80% de,12 while, under the similar reaction conditions,
the condensation of acetophenone gives a 1 to 1 mixture
of the diastereomeric products.8b
In the work described here, following a systematic
study, we have examined the diastereoselectivity of the
transition metal/base-catalyzed reactions between methyl
isocyanoacetate and a wide range of prochiral ketones
2a -z. Our study indicates that the stereochemical
outcome of these reactions is influenced by a mix of
several factors, including steric characteristic of the
ketone substituents, the nature of the catalyst (metal and
Sch em e 2
(6) (a) Soloshonok, V. A.; Kukhar’, V. P.; Galushko, S. V.; Svistunova,
N. Yu.; Avilov, D. V.; Kuz’mina, N. A.; Raevski, N. I.; Struchkov, Yu.
T.; Pysarevsky A. P.; Belokon’ Yu. N. J . Chem. Soc., Perkin Trans. 1
1993, 3143. (b) Soloshonok, V. A.; Avilov, D. V.; Kukhar’, V. P.
Tetrahedron: Asymmetry 1996, 7, 1547. (c) Soloshonok, V. A.; Avilov,
D. V.; Kukhar’, V. P. Tetrahedron 1996, 52, 12433.
(7) For reviews: (a) Schollkopf, U. Angew. Chem., Int. Ed. Engl.
1977, 16, 339. (b) Matsumoto, K.; Moriya, T.; Suzuki, M. J . Synth.
Org. Chem., J pn. 1985, 43, 764.
(8) (a) Ito, Y.; Matsumura, T.; Saegusa, T. Tetrahedron Lett. 1985,
26, 5781. (b) Saegusa, T.; Ito, Y.; Kinoshita, H.; Tomita, S. J . Org.
Chem. 1971, 36, 3316, and references cited therein.
base) employed, and, in some cases, the electrostatic,
donor-acceptor type interactions. The resultant oxazo-
line diastereomer ratios appear to be determined primar-
ily by the steric influence of the ketone substituents;
however, they were found to be tunable depending on a
proper choice of the catalyst. In particular, for the
reactions of methyl isocyanoacetate 1 with polyhaloalkyl
aryl(alkyl) ketones, a high, synthetically useful level
stereoselectivity (80-98% de) was achieved, allowing
straightforward and generalized access to the corre-
sponding (2R*,3R*)-â,â-disubstituted-â-hydroxy-R-amino
carboxylic acid of biomedicinal importance. Apart from
the synthetic results, the data presented here shed light
on some of the intricacies of these metal/base-catalyzed
reactions, and should allow for the design of more general
and practical catalytic and enantioselective aldol-type
processes in the future.
(9) (a) Soloshonok, V. A.; Kacharov, A. D.; Hayashi, T. Tetrahedron
1996, 52, 245. (b) Soloshonok, V. A.; Hayashi, T. Tetrahedron Lett.
1994, 35, 2713. (c) Soloshonok, V. A.; Hayashi, T. Tetrahedron:
Asymmetry 1994, 5, 1091. (d) Hayashi, T.; Sawamura, M.; Ito, Y.
Tetrahedron 1992, 48, 1999. (e) Ito, Y.; Sawamura, M.; Hamashima,
H.; Emura, T.; Hayashi, T. Tetrahedron Lett. 1989, 30, 4681. (f)
Sawamura, M.; Ito, Y.; Hayashi, T. Tetrahedron Lett. 1989, 30, 2247.
(g) Ito, Y.; Sawamura, M.; Hayashi, T. Tetrahedron Lett. 1988, 29, 239.
(h) Ito, Y.; Sawamura, M.; Kobayashi, M.; Hayashi, T. Tetrahedron
Lett. 1988, 29, 6321. (i) Ito, Y.; Sawamura, M.; Shirakawa, E.;
Hayashizaki, K.; Hayashi, T. Tetrahedron Lett. 1988, 29, 235. (j) Ito,
Y.; Sawamura, M.; Hayashi, T. Tetrahedron Lett. 1987, 28, 6215. (k)
Ito, Y.; Sawamura, M.; Hayashi, T. J . Am. Chem. Soc. 1986, 108, 6405.
(10) (a) Pastor, S. D.; Kesselring, R.; Togni, A. J . Organomet. Chem.
1992, 429, 415. (b) Pastor, S. D.; Togni, A. Helv. Chim. Acta 1991, 74,
905. (c) Togni, A.; Blumer, R. E.; Pregosin, P. S. Helv. Chim. Acta
1991, 74, 1533. (d) Togni, A.; Pastor, S. D. J . Org. Chem. 1990, 55,
1649. (e) Pastor, S. D.; Togni, A. Tetrahedron Lett. 1990, 31, 839. (f)
Togni, A.; Pastor, S. D. Tetrahedron Lett. 1989, 30, 1071. (g) Togni,
A.; Pastor, S. D.; Rihs, G. Helv. Chim. Acta 1989, 72, 1471. (h) Pastor,
S. D.; Togni, A. J . Am. Chem. Soc. 1989, 111, 2333.
Resu lts
(11) (a) Hayashi, T.; Uozumi, Y.; Yamazaki, A.; Sawamura, M.;
Hamashima, H.; Ito, Y. Tetrahedron Lett. 1991, 32, 2799. (b) Sawa-
mura, M.; Hamashima, H.; Ito, Y. J . Org. Chem. 1990, 55, 5935.
(12) Heinzer, F.; Bellus, D. Helv. Chim. Acta 1981, 64, 2279.
We have reacted methyl R-isocyanoacetate (1) (Scheme
1) with prochiral ketones 2a -z to give the corresponding
diastereomeric oxazolines13 3,4a -z. The most represen-

3472 J . Org. Chem., Vol. 62, No. 11, 1997
Ta ble 1. Ca ta lytic Ald ol Rea ction s of Keton es 2a -z w ith Meth yl Isocya n oa ceta te 1a a
ketone
catalyst,b
metal/base
Soloshonok et al.
ratiod
entry
R
R1
mol %
time
yield,c %
trans-3
cis-4
1 (a )
2 (b)
3 (b)
4 (b)
5 (b)
6 (b)
7 (b)
8 (b)
9 (b)
C6H5
C6F5
C6F5
C6F5
C6F5
C6F5
C6F5
C6F5
C6F5
C6F5
C6F5
C6F5
C6F5
C6F5
C6F5
C6F5
C6H5
C6H5
C6H5
C6H5
4-CF3-C6H4
4-CH3O-C6H4
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
Me
CF3
CF3
CF3
CF3
CF3
CF3
CF3
CF3
CF3
CF3
CF3
CH3
CH3
CH3
CH3
CF3
Cu(I)/NEt3
Cu(I)/NEt3
Cu(I)/NEt3
Au(I)/NEt3
Ag(I)/NEt3
Ag(I)/i-Pr2NEt
10
10
10
1
1
1
1
1
1
10
25
24 he
10 min
10 h
4 h
80
93
91
95
92
96
97
99
94
78
60
54
85
83
77
84
92
94
95
93
94
89
96
98
93
92
95
93
94
87
81
74
93
97
94
99
95
89
91
88
91
93
89
95
94
87
91
96
87
83
91
NR
37i
31i
24i
53
91
91
92
94
83
98
64
77
73
95
80
41
37
48
50
86
89
89
88
88
87
93
92
85
>99
94
95
>99
>99
18
56
71
74
79
92
>99
71
99
>99
76
90
86
91
84
35
43
47
9
9
8
6
3.5 h
12 h
3.5 h
45 h
12 h
20 h
50 h
96 h
48 h
50 h
50 h
60 h
5 h
5 h
5 h
7 h
5 h
6 h
5 h
10 h
5 h
10 min
1 h
17
2
36
23
27
5
20
59
63
52
50
14
11
11
12
12
13
7
f
Ag(I)/NMe3
Ag(I)/PyH
Ag(I)/PSg
Cu(I)
10 (b)
11 (b)
12 (b)
13 (c)
14 (c)
15 (c)
16 (c)
17 (d )
18 (d )
19 (d )
20 (d )
21 (e)
22 (f)
23 (g)
24 (g)
25 (g)
26 (g)
27 (g)
28 (h )
29 (h )
30 (i)
31 (j)
32 (k )
33 (l)
34 (m )
35 (m )
36 (n )
37 (o)
38 (p )
39 (q)
40 (r )
41 (s)
42 (s)
43 (s)
44 (t)
45 (t)
46 (u )
47 (u )
48 (v)
49 (w )
50 (w )
51 (w )
52 (w )
53 (x)
54 (y)
55 (z)
NEt3
Ag(I)/NEt3
Cu(I)/NEt3
Au(I)/NEt3
Pd(II)/NEt3
Ag(I)/NEt3
Au(I)/NEt3
Rh(I)/NEt3
Cu(I)/NEt3
Ag(I)/NEt3
Ag(I)/NEt3
Ag(I)/NEt3
Rh(I)/NEt3
Pd(II)/NEt3
Cu(I)/NEt3
NEt3
Ag(I)/NEt3
Cu(I)/NEt3
Cu(I)/NEt3
Cu(I)/NEt3
Cu(I)/NEt3
Cu(I)/NEt3
Cu(I)/NEt3
Ag(I)/NEt3
Cu(I)/NEt3
Cu(I)/NEt3
Cu(I)/NEt3
Cu(I)/NEt3
Cu(I)/NEt3
Cu(I)/NEt3
Ag(I)/NEt3
NEt3
1
1
1
1
1
1
1
10
1
1
CF3
CF3
CF3
CF3
CF3
CClF2
CClF2
CClF2
CClF2
CClF2
C2F5
C2F5
n-C4F9
CN
CH2OMe
CH2F
CH2Cl
CH2Cl
CHCl2
CCl3
CHCl2
CF3
n-C3F7
CF3
CF3
CF3
CF3
CF3
CF3
CF3
CF3
1
1
1
8
15
<1
6
10
200
1
10
10
10
10
10
10
2
10
10
10
10
10
10
2
6 h
5
10 min
10 min
10 min
20 h
0.5 h
0.5 h
0.5 h
0.5 h
0.5 h
0.5 h
0.5 hh
0.5 hh
1 h
<1
<1
82
44
29
26
21
8
<1
29
1
<1
24
10
14
9
16
65
57
<1
38
16
9
C6H5
C6H5
Me
Me
Me
Me
Me
Me
CH2Cl
Me
Me
n-Hep
n-Hep
n-Hep
n-Oct
n-Oct
c-Hex.
c-Hex.
C≡C-Ph
CH2COOEt
CH2COOEt
CH2COOEt
CH2COOEt
Me
1 h
2 h
1 h
2 h
200
5
200
10
5
Ag(I)/NEt3
NEt3
Cu(I)/NEt3
Ag(I)/NEt3
Ag(I)/NEt3
Au(I)/NEt3
Cu(I)/NEt3
Ag(I)/NEt3
NEt3
Cu(I)/NEt3
Cu(I)/NEt3
Cu(I)/NEt3
5 h
10 h
10 minh
0.5 h
1 h
0.5 h
24 h
24 h
24 h
24 h
2
2
>99
62
84
91
CF3
CF3
CF3
CF3
Et
n-Pr
c-Hex
10
2
100
10
10
10
60
62
89
40
38
11
Me
Me
a
All reactions were run in 1,2-dichloroethane at 18-23 °C in the presence of the catalyst indicated in the table. Ratio 1a /2a -z )
1/1.1; 1-5 mmol scale. Reactions were monitored by GLC, and upon completion, the products were isolated by bulb-to-bulb distillation.
b
Cu(I) is CuCl; Ag(I), AgClO4 or AgBF4; Rh(I), [Rh(NBD)2]BF4; Pd(II), PdCl2(MeCN)2; Au(I), [Au(c-HexNC)2]BF4; ratio metal/base 1/1-2.
c Isolated yield. Determined by GLC and 1H and 19F NMR analyses. e At 50 °C. f 1,2-Dichloroethane saturated with NMe3 was used.
d
g
h
i
PS is “Proton Sponge”, 1,8-diaminonaphthalene. Reaction was started at 0 °C and then warmed to room temperature. Less than
50% conversion of the starting materials.
tative results obtained are collected in Table 1. All
reactions were performed under the similar reaction
conditions, in dry 1,2-dichloroethane at ambient temper-
ature in the presence of 1-10 mol % of the transition
metal/base catalyst, except for the condensation of aceto-
phenone (2a ) with 1 which, presumably due to the poor
electrophilicity and steric shielding of the carbonyl in 2a ,
necessitated an elevated temperature to proceed (Table
1, entry 1). In most cases CuCl and triethylamine (TEA),
in a 1/1-2 ratio, was used as the catalyst. To explore
the effect of both the metal and base components of the
catalyst on the stereochemical outcome, we carried out
a series of isocyanoacetate 1 condensations varying the
base, i-Pr2NEt (Table 1, entry 6), NMe3 (entry 7), PyH
(entry 8), 1,8-diaminonaphtalene (entry 9), and the metal
used, Ag(I) (AgClO4 or AgBF4) (entries 5-9, 13, 17, 21-
23, 28, 35, 42, 44, 47, 48, 51), Au(I) {[Au(c-HexNC)2]BF4}
(entries 4, 15, 18, 49), and Rh(I) {[Rh(NBD)2]BF4} (entries
19, 24), and Pd(II) [PdCl2(MeCN)2] (entries 16, 25). For
some ketones we also investigated the diastereoselectivity
of the reactions with 1 catalyzed separately by the metal
(13) The common name, 2-oxazoline, is used throughout. Following
nomenclature rules, the systematic name for 3 and 4 is 4,5-dihydro-
5-(substituent)-5-(substituent)-4-oxazolecarboxylic acid, methyl ester.

Transition Metal/Base-Catalyzed Aldol Reactions
J . Org. Chem., Vol. 62, No. 11, 1997 3473
1
(CuCl, entry 10) and by the base (TEA, entries 11, 27,
43, 45, 52), as well as uncatalyzed condensation of highly
electrophilic octafluoroacetophenone 2b (entry 12). Re-
gardless of the nature of starting ketones and the catalyst
employed, all reactions proceeded irreversibly and were
very clean, as no byproducts detectable by NMR were
observed. In the series of isocyanoacetate 1 reactions
with 2a -z the conversions of the starting materials and
the yields of the targeted oxazolines were generally quite
high, except for the reactions of 1 with ketones 2x-z,
which proceeded with very low reaction rates, apparently
due to the poor electrophilicity of the ketones (Scheme
1, Table 1). The stereochemical outcome of the reactions
is assumed to be kinetically controlled since no time-
dependent alterations in the diastereoselectivity were
detected (entry 2 vs 3). Moreover, additional experiments
have shown that the ratio of diastereomeric oxazolines
3 and 4 does not change while standing in neat TEA
solution at rt, thus allowing us to neglect the possibility
of diastereomer epimerization at the R-position to the
methoxycarbonyl group in 3 and 4.
The ratio of diastereomers 3/4 was determined by GLC
and/or NMR analysis on the crude reaction mixture,
before removal of the solvent and distillation, and by
NMR (1H and 19F, where possible) analysis on the
distilled products. The ratios determined by NMR
analysis were comparable ((1%) with those determined
by GLC analysis. In order to avoid complication of
Scheme 1, the trans/cis definition for 3/4 is used through-
out with respect to the relationship between the sub-
stituent R1 and the methoxycarbonyl group. Thus,
compounds 3, bearing trans-disposed R1 and COOMe
groups, are defined as trans-3 and the diastereomeric
oxazolines 4, with R1 and COOMe groups in cis-disposi-
tion to each other, are cis-4. These assignments are
correct for most of the products 3 and 4 (b,d -z) except
only for 5-aryl-5-methyl derivatives 3,4a ,c, where, ac-
cording to the CIP priority rules,14 diastereomers 3 and
4 are cis and trans, respectively.
The determination of relative configuration for the
products 3,4a -z was based on NMR chemical shifts,
X-ray analysis and comparison with the literature ex-
amples. For the series of 5-phenyl-containing derivatives
3,4a ,d -j relative configuration was unambiguously as-
signed using an apparent difference in the chemical shifts
of the methoxy (COOMe) protons of diastereomers 3,4a,d-
j. As established previously for the relative oxazolines
derived from aromatic aldehydes and 1, the methoxy
protons of the diastereomer with cis-disposed phenyl and
the methoxycarbonyl group, due to the shielding effect
of the aromatic ring,15 are shifted strongly upfield (3.25-
3.28 ppm), as compared with those of the opposite
diastereomer (3.82-3.84 ppm).9 For the derivatives
3,4a ,d -j we observed the same pattern of the chemical
shifts. Thus, in trans-oxazolines 3a ,d -j, bearing cis-
disposed aryl and COOMe groups, the protons of the
latter are substantially shifted upfield (3.16-3.33 ppm)
as compared with the chemical shifts of the methoxy
group in cis-oxazolines 4 (3.86-3.93 ppm). Moreover, in
the series of 5-fluoroalkyl-substituted 3,4d -i, fluorine
resonances of the minor cis-diastereomers 4d -i, with cis-
disposed fluoroalkyl and COOMe groups, presumably due
to the deshielding effect of the carbonyl group,15 are
shifted downfield as compared with 19F NMR chemical
shifts of trans-3d -i. Although the pattern of H NMR
spectra of oxazolines 3,4b,c was quite similar to that of
the 5-phenyl-substituted derivatives 3,4d -i, the difference
between chemical shifts of the methoxy protons of the
diastereomers was much less pronounced. Since we
failed to find in the literature reliable data on the
shielding-deshielding effects of the pentafluorophenyl
ring, we have established configuration of the dominant
diastereomer 3b by X-ray analysis of its crystalline
derivative 7, which was prepared as shown in Scheme 2.
As revealed by X-ray analysis,1a compound 7 has a
(2R*,3R*)-configuration16 that corresponds to the trans
relative configuration of the oxazoline 3b. Accordingly,
oxazolines 4b,c and 3c were assigned as cis and trans
diastereoisomers, respectively. In the aliphatic series,
relative configurations of trans-3m and cis-4m were
determined by comparison of 1H NMR data obtained for
the diastereomers 3m ,4m with those reported in the
literature for trans/cis-4-(ethoxycarbonyl)-5-(chloromethyl)-
5-methyl-2-oxazolines.12 Relative trans-configuration of
the oxazolines 3q,s,t was confirmed via their acidic
hydrolysis to the corresponding amino acid 6q,s,t, the
(2R*,3R*)-configuration of which was established by
comparison with the samples of known stereochemistry.6
To the rest of oxazolines 3k ,l,n -r ,u ,v,x-z trans config-
uration, and consequently cis to 4k ,l,n -r ,u ,v,x-z dia-
stereomers, was assigned on the basis of an apparent
similarity of their NMR spectra to the patterns of
3,4m ,q,s,t. The most characteristic difference in 1H
NMR spectra of the diastereomers 3,4k -v,x-z is the
chemical shifts of the proton at C4. Thus, for the cis
diastereomers 4k -v,x-z, the resonance of this proton
is shifted upfield relative to the corresponding resonance
of the trans-configured products 3,4k -v,x-z. This trend
is also observed in the spectra of aromatic derivatives
3,4a -j and is in an agreement with the data reported
for the oxazolines derived from R-chloroalkyl alkyl(aryl)
ketones and ethyl isocyanoacetate.12 Moreover, in the
series of 5-(trifluoromethyl)-containing compounds 3,4q,s-
v, chemical shifts of the trifluoromethyl groups of cis
diastereomers 4q,s-v are shifted downfield relative to
19F NMR chemical shifts of trans-3q,s-v, repeating the
trend observed for aromatic derivatives 3,4d -i. The
assignment of configuration of 3,4w requires an ad-
ditional comment. The reaction of â-keto ester 2w with
isocyanoacetate 1 follows the usual course and gives the
oxazolines 3,4w with sizable domination of one of the
diastereomers, as disclosed by NMR. In this case, the
resonance of the proton at C4 of the dominant diastere-
omer is shifted upfield (4.94 ppm, d, J ) 2.6 Hz) relative
to the corresponding resonance of the minor diastereomer
(5.43 ppm, d, J ) 2.3 Hz). However, considering the half-
width of the doublets, the one at 4.94 ppm with the half-
width of 1.6 Hz might be ascribed to the trans-3w , since
the signal at 5.43 ppm has the half-width of 3.1 Hz that
could be caused by a small coupling constant between
the trans-disposed R-proton and the trifluoromethyl
group. This trend is observed for all trifluoromethyl-
containing oxazolines 3,4q,s-v: the half-width of the
resonance of the proton at C4 of cis diastereomer 4q,s-v
is sizably bigger than that of trans isomer 3q,s-v.
The results, collected in Table 1, indicate that the
stereochemical outcome of the reaction studied is a
function of the nature of the starting ketone, as well as
the catalyst employed (vide infra). For some cases, the
(14) Cahn, R. S.; Ingold, C.; Prelog, V. Angew. Chem. Int. Ed. Engl.
1966, 5, 385.
(15) Silverstein, R. M.; Bassler, G. C.; Morrill, T. C. Spectrometric
Identification of Organic Compounds, 3rd ed.; Wiley: New York, 1974.
(16) The (2R*,3R*)-configuration, a consequence of the Cahn-
Ingold-Prelog priority (ref 14), is stereochemically equivalent to the
(2R*,3S*)-configuration in the hydrocarbon analogs.

3474 J . Org. Chem., Vol. 62, No. 11, 1997
Soloshonok et al.
R is sterically larger than R1. According to the proposed
transition states A and B, the effects of the catalyst on
the diastereoselectivity should be rather small, as both
the base and metal are distant from the reaction site.
However, they still could influence stereochemical pref-
erences by the interaction with the substituents R
and R1.
Ster ic Effects of Keton e Su bstitu en ts on th e
Ster eoch em ica l Ou tcom e. While the importance of
steric requirements of substituents, among other factors
influencing the stereoselectivity in aldol reactions, is well-
known and rationalized,2 some of the stereochemical
preferences observed in this study were rather unex-
pected and thus worthy of discussion. The stereochem-
ical outcome of the sluggish condensation of acetophenone
(2a ) and 1 suggests that there is no significant stereo-
discrimination between the phenyl and the methyl, and
apparently, the reaction proceeded through both transi-
tion states A and B with comparable reaction rates (entry
1). In sharp contrast, the exothermic condensation of
octafluoroacetophenone (2b) with 1 afforded the product
3,4b with remarkable domination of trans diastereomer
3b (entry 2). The diastereoselectivity in this reaction was
found to be influenced by the catalyst employed (vide
infra) and could be enhanced up to the value of 96% de
(entry 7 vs 2). To get insight in the origin of this
surprising stereocontrol, we investigated the diastereo-
selectivity in the reactions of 1 with acetophenones 2c,2d ,
bearing separately pentafluorophenyl and trifluoromethyl
moieties of 2b, as well as the reaction with ketone 2j,
the CN group of which, being a sterically innocuous
substituent, would mimic the strong electron-withdraw-
ing effect of the trifluoromethyl group.17 The results
shown in entries 13, 17, 20, and 31 strongly suggest that
in the reactions of isocyanoacetate 1 with ketones 2b,d
the trifluoromethyl group plays the role of the stereo-
controlling substituent, as if it were a group larger than
a phenyl. We found next that electron-withdrawing and
electron-releasing substituents on the phenyl of trifluo-
roacetophenone, ketones 2e,f, respectively, do not influ-
ence high trans diastereoselectivity (entries 21, 22 vs 17),
while a substitution of the trifluoromethyl by the bulkier
chlorodifluoromethyl and perfluoroalkyl groups, ketones
2g-i, respectively, enhances further the diastereoselec-
tivity to the level of complete stereocontrol, allowing an
efficient preparation of diastereomerically pure trans
oxazolines 3g-i (entries 26, 28, 29). These data clearly
demonstrate that in the reactions under study trifluo-
romethyl, or more generally the perfluoroalkyl group, acts
as a substituent sterically bulkier than a phenyl or
pentafluorophenyl group. To account for these results,
we can assume that in the transition state A, leading to
trans-oxazolines 3, the phenyl or its substituted analogs,
due to the flat shape of these groups, might more
effectively minimize unfavorable steric interactions than
the trifluoromethyl or perfluoroalkyl group could do,
occupying the sterically unfavorable position in the
transition state B. Apart from purely steric reasons, we
believe that a plausible explanation of the phenomenon
observed might involve also nonbonding, repulsive in-
teractions between the perfluoroalkyl group and the
enolate moiety6 (vide infra).
F igu r e 1. Proposed transition states A and B for the aldol
reactions of isocyanoacetate 1 with prochiral ketones 2a -z.
reactions of 1 with ketones 2b,d -i,n ,m ,q-t,w (Table 1,
entries 5, 20-22, 26, 29, 30, 36, 37, 39, 40, 42, 44, 48,
51), high chemical yields, and appreciable diastereose-
lectivity render this method synthetically useful for
preparing targeted â,â-disubstituted-â-hydroxy-R-amino
carboxylic acid. Hydrolysis of oxazolines of this type to
the corresponding amino acids is a well-established
procedure normally carried out under the mild acidic
conditions.7-12 In this work, we have demonstrated this
transformation with the preparation of certain represen-
tative (2R*,3R*)-â,â-disubstituted-â-hydroxy-R-amino
acids 6d ,g,h ,q,s,t in diastereomerically pure state using
one- and two-stage hydrolysis procedures (Scheme 1).
First, we hydrolyzed oxazolines 3,4d ,g,h ,s (79-99% de
of trans-3) in aqueous methanol in the presence of
catalytic amounts of HCl to afford the corresponding
N-formyl derivatives 5d ,g,h ,s. Products 5d ,g,h ,s are
crystalline compounds and could be easily purified to the
diastereomerically pure form by a conventional crystal-
lization. Further hydrolysis of 5d ,g,h ,s to give targeted
amino acids 6d ,g,h ,s was accomplished by heating of a
solution of 5d ,g,h ,s in 6 N HCl. Direct hydrolyses of
oxazolines are represented by the synthesis of amino
acids 6d ,g,h ,q,s,t from the corresponding oxazolines
3,4d ,g,h ,q,s,t. In this case, resultant amino acids were
crystallized from water to furnish diastereomerically pure
compounds.
Discu ssion
Mech a n istic Ba ck gr ou n d ; Tr a n sition Sta tes in
th e Rea ction s of Isocya n oa ceta te 1 w ith Keton es
2a -z. An examination of Table 1 reveals that the
diastereoselectivity of the reactions under study is de-
termined primarily by the steric properties of the ketone
substituents. On the other hand, the effect of the catalyst
on the ratio of oxazolines 3,4a -z normally is less
pronounced; however, in some cases, the nature of the
catalyst can dramatically influence the stereochemical
result (e.g., entry 5 vs 8, 9; 14 vs 16; 25 vs 26; 41 vs 42;
44 vs 45; 49 vs 51). To account for these general trends,
by analogy with the reactions of R-isocyanoacetic acid
derivatives with aldehydes,7-11 we can propose two
possible transition states, A and B (Figure 1), for the
reactions between methyl isocyanoacetate 1 and ketones
2a -z. In A, the substituent R1 is pointing away from
the enolate moiety and interacts with the sterically
undemanding proton and linear isocyano group, while the
substituent R in A interacts unfavorably with the enolate
methoxy group pointed toward it. It follows that the
transition state A, leading to the formation of trans
oxazolines 3, might be realized in the cases when the
substituent R1 is more sterically demanding relative to
the R. The transition state B, giving rise to cis oxazolines
4, might be favored relative to A when the substituent
Aliphatic amino compounds make up a large body of
naturally occurring and medicinally important com-
pounds. Therefore, it was of particular interest to us to
(17) Hansch, C.; Leo, A. Substituent Constants for Correlation
Analysis in Chemistry and Biology; Wiley: New York, 1979.

Transition Metal/Base-Catalyzed Aldol Reactions
J . Org. Chem., Vol. 62, No. 11, 1997 3475
investigate the condensations of isocyanoacetate 1 with
aliphatic ketones. In the CuCl/NEt3-catalyzed reactions
of 1 with mono-substituted acetones 2k -m low-to-modest
diastereoselectivities were obtained (entries 32-34). It
is interesting to note the sizable difference in the dias-
tereoselectivity of the condensations of 1 with mono-
fluoro- (2l) and methoxyacetones (2k ) (entries 32 vs 33),
as it is stated that “F and OH are chemical isosteres”.18
A remarkable increase in the trans diastereoselectivity
was observed for the condensations of dichloromethyl
(2n ) and trichloromethyl (2o) ketones with 1 (entries 36,
37), while the diastereoselectivity of the reaction between
1 and dichloromethyl chloromethyl ketone (2p ) was found
to be comparable with the stereochemical outcome of the
corresponding condensation of monochloroacetone (2m )
(entry 38 vs 34). The diastereoselectivities obtained in
the series of mono-, di-, and trichloro-substituted acetones
2m -p reactions (entries 34-38) clearly show the steric
influences of the substituents on the starting ketone in
the diastereoselection process in these aldol condensa-
tions.
reactions under basic reaction conditions (entry 52). The
diastereoselectivity of this reaction was found to be highly
dependent on the catalyst employed (entries 49-51) but
comparable in value with that of n-alkyl ketones 2s,t
reactions with 1 (entries 42, 44 vs 51).
Finally, Cu(I)/NEt3-catalyzed reactions of unsubsti-
tuted prochiral hydrocarbon ketones with isocyanoacetate
1 were investigated. We found that the condensations
of ethyl methyl and n-propyl methyl ketones (2x,y) with
1 proceeded with low stereoselection (entries 53, 54),
while cyclohexyl methyl ketone (2z) reacted with 1, giving
rise to trans-oxazoline 3z with markedly higher diaste-
reoselectivity (entry 55). It is worth noting that the
diastereodirecting effect of the cyclohexyl group in the
reaction of 2z with 1 (entry 55) was less pronounced than
that of trifluoromethyl in the condensation of 2q with 1
(entry 39), while the direct competition of these substit-
uents in the reaction of cyclohexyl trifluoromethyl ketone
(2n ) with 1 (entries 46,47) showed a preference for the
cyclohexyl group.
The results discussed here clearly demonstrate the
pivotal role of the steric parameters of substituents on
the starting ketone in the origin of stereocontrol in the
reactions under study. However, steric interactions alone
probably cannot account for the sense and the values of
the diastereoselectivities observed. Thus, while for the
hydrocarbon series (reactions of 1 with 2x-z) a plot of
the Es19 and Es′20 constants for the substituents Et, n-Pr,
c-Hex versus log trans/cis ratios of 3,4x-z gives an
acceptable linear correlation, the diastereoselectivities
obtained in the reactions of 1 with halo-substituted
ketones are much higher to follow this correlation.
With these results in hand and in view of the above-
mentioned data on the condensations of trifluoromethyl-
containing ketones with a Ni(II) complex of the Schiff
base of glycine,6 we anticipated high diastereoselectivity
in the reactions of isocyanoacetate 1 with alkyl trifluo-
romethyl(perfluoroalkyl) ketones. Indeed, the condensa-
tions between 1 and methyl trifluoromethyl (2q) and
methyl n-heptafluoropropyl (2r ) ketones, catalyzed by
Cu(I)/TEA, shown excellent (entry 39) and complete
(entry 40) stereochemical discrimination between the
methyl and the perfluoroalkyl groups, respectively, af-
fording diastereomerically pure trans oxazolines 3q,r in
high yield. Excellent diastereoselectivity was achieved
also in the reaction of R-acetylenic R’-trifluoromethyl
ketone 2v with isocyanoacetate 1. Due to the high
electrophilicity of ketone 2v, its Ag(I)/NEt3-catalyzed
condensation with 1 occurred with a very high reaction
rate and gave the desired trans-oxazoline 3v as an
individual product in high chemical yield (entry 48).
However, Cu(I)/TEA-catalyzed reaction of n-heptyl tri-
fluoromethyl ketone (2s) with 1 gave the targeted oxazo-
lines 3,4s with remarkably reduced trans diastereose-
lectivity (entry 41). Despite the stereoselectivity of this
reaction being to 80% de with an application of Ag(I)/
TEA as the catalyst (entry 42), this value remains still
lower compared with the diastereoselectivity of >98%
reported for the reaction of ketone 2s with a Ni(II)
complex of the chiral Schiff base of glycine.6 A similar
stereochemical result was obtained in the condensation
of n-octyl trifluoromethyl ketone (2t) with 1 (entry 44).
These data suggest once again that, while important,
steric reasons alone fail to rationalize the stereochemical
outcome of the reactions under study (vide infra). In
sharp contrast to the reactions of trifluoromethyl-
containing aryl 2b,d -f, and n-alkyl ketones 2q,s,t, Cu(I)/
NEt3-catalyzed condensation of cyclohexyl trifluorome-
thyl ketone (2u ) with 1 proceeded with the opposite sense
of stereochemical preferences, furnishing cis-oxazoline 4u
as the major diastereomer, albeit with low diastereose-
lectivity (entry 46). Application of Ag(I)/NEt3 as the
catalyst in this reaction gave a mixture of oxazolines 3,4u
with lower diastereoselectivity (entry 47). Condensation
of â-keto ester 2w with 1 was of particular interest, since
highly enolizable 2w normally does not enter aldol
Effect of Electr osta tic In ter a ction s on th e Ster -
eoch em ica l Ou tcom e. An evaluation and use of the
19
20
steric substituent constants Es and Es′ as a measure
of steric requirements is of great importance to the
problems of mechanistic studies, when it is applied to
the substituents of similar nature. However, attempts
to describe the stereochemical behavior of the different
in nature substituents merely by the means of their steric
constants or van der Waals radii21 usually give contro-
versial results. In particular, direct comparison of the
steric characteristic of fluorocarbon and hydrocarbon
substituents seems to be rather inappropriate and meth-
odologically misleading.22 Thus, whereas the Es′ value
of a phenyl group (2.31) is much larger than that of a
trifluoromethyl group (0.78),20 the results reported on
asymmetric reductions of trifluoroacetophenone 2d show
that the sense and value of the asymmetric outcome
heavily depends on the reducing reagent employed.23 For
instance, the reduction of ketone 2d with alpine-borane
afforded the corresponding carbinol (54% ee) of (R)-
configuration, that is the same as compared to the
stereochemistry observed in the reduction of fluorine-free
acetophenone, while the reduction of 2d with DIP-Cl gave
the opposite stereochemical result.23l Since the steric
bulkiness of the substituents could not “depend” on the
reaction condition employed, an explanation for these
anomalous data might involve, apart from steric proper-
(19) Taft, R. W. In Steric Effects in Organic Chemistry; Newman,
M. S., Ed.; Wiley: New York, 1956; p 556.
(20) MacPhee, A.; Panaye, A.; Dubois, J .-E. Tetrahedron 1978, 34,
3553.
(21) Bott, G.; Field, L. D.; Sternhell, S. J . Am. Chem. Soc. 1980,
102, 5618.
(18) Smart, B. E. In Organofluorine Chemistry: Principles and
Commercial Applications; Banks, R. E.; Smart, B. E.; Tatlow, J . C.,
Eds.; Plenum Press: New York, 1994; pp. 57-88.
(22) For instance, such statements as “CF3 is sterically at least as
large as CH(CH3)2”; see ref 18. For asymmetric results, which do not
follow this statement, and the corresponding rationale, see ref 23k.

3476 J . Org. Chem., Vol. 62, No. 11, 1997
Soloshonok et al.
ties, some other features of the substituents, such as the
electronic and electrostatic properties.
addition reactions of PFCHIRAC and fluorine-free
CHIRAC,25 is caused by the electron donor-acceptor type
attractive interaction between the enolate oxygen and the
pentafluorophenyl ring.24
Thus, a dramatic difference between the stereochem-
ical outcomes of the reactions of isocyanoacetate 1 with
acetophenone (2a ) and its trifluoromethyl analogs 2b,d -f
could be reasonably ascribed, as proposed above, to the
larger steric requirements of the spherical trifluoro-
methyl group being larger than that of the flat phenyl
in the transition states B and A (Figure 1), respectively.
However, by analogy with the mechanistic rationale for
the high diastereoselectivity observed in the reactions of
trifluoromethyl ketones with a Ni(II) complex of the
chiral Schiff base of glycine,6 we can assume that the
transition state B could be additionally destabilized by
the repulsive electrostatic interactions between the nega-
tively charged enolate moiety and the trifluoromethyl
group. These effects might take place also in the reac-
tions of isocyanoacetate 1 with per(poly)fluoroalkyl alkyl-
(aryl) ketones 2g-i,q-w and, in general, with other
halogen-containing ketones, influencing the stereoselec-
tivity in these condensations. In particular, the results
obtained allow us to conclude that, in the reactions under
study, a trifluoromethyl group acts as a sterically larger
substituent than a thin and flexible n-alkyl or a flat
phenyl groups, even though the total steric volume of
these groups is much larger than that of the trifluoro-
methyl.
Other stereochemical results which are difficult to
explain on the basis of purely steric reasons are the
higher values of diastereoselectivities observed in the
reactions of 1 with pentafluorophenyl-containing ketones
2b,c as compared with that of acetophenone (2a ) and
trifluoroacetophenone 2d condensations with 1 (e.g.,
entry 1 vs 14, and 5 vs 17). Recently we have shown that
in the reactions of methyl isocyanoacetate 1 with fluo-
robenzaldehydes the number of fluorine atoms on the
phenyl ring of the starting aldehyde controls the stereo-
chemical outcome of the condensation, giving rise in the
case of polyfluorobenzaldehydes reactions to the corre-
sponding cis-oxazolines.9a Our working rationale for this
cis diastereoselectivity involves an electron donor-ac-
ceptor type attractive interaction between the enolate
oxygen and the polyfluorophenyl ring. We can assume
that the same type of the interactions could take place
in the reactions of 1 with ketones 2b,c. Thus, an
attractive interaction between negatively charged enolate
moiety and electron-deficient pentafluorophenyl ring
would stabilized the transition state A, giving some
preference for trans-oxazolines 3b,c formation. This
assumption is strongly supported by the results reported
previously by Ojima and Kwon on the addition reactions
to a pentafluorophenyl-containing chiral iron acyl com-
plex (PFCHIRAC).24 The authors had proven that the
opposite sense of stereodifferentiation, disclosed for the
On the basis of these and literature data,6,23,24,26-28 we
can suggest that the stereochemical outcome of the
reactions, proceeding with an involvement of the fluorine-
containing substituent into the stereoselective step,
might be rationalized as the balance between steric and
electronic/electrostatic effects influencing the pattern of
stereochemical preferences. Under the certain reaction
conditions the former or the latter effects could dominate
and the both could work for the same or different
stereochemical result. Obviously, the stereochemical
behavior of fluorine must await further investigations.29
Ca ta lyst Effect on Rea ction Ra te a n d Ster eose-
lectivity. An important advantage of the transition
metal catalysis in the reactions under study is that the
coordination of a transition metal to the isocyano moiety
increases substantially the C-H acidity of the methylene
group of isocyanoacetate, allowing the reaction to proceed
under more gentle conditions and to enjoy wider func-
tional group tolerance as compared with those of the old-
fashioned protocol relying on strongly basic conditions.7,8
Thus, Ito has shown that while CuCl alone is capable of
catalyzing the condensation of ethyl isocyanoacetate with
carbonyl compounds, addition of TEA to the CuCl causes
a remarkable acceleration of the reaction rate.8a How-
ever, no data on the relevance of the catalyst to the
diastereoselectivity of the reactions have been reported.
Given the example of the reaction between isocyano-
acetate 1 with octafluoroacetophenone (2b) (entries 2-12),
we investigated the effect of both the metal and base
components of the catalyst on the stereochemical out-
come, as, due to high reactivity of the ketone 2b, it was
possible to explore also the diastereoselectivity of the
reactions catalyzed separately by the metal (entry 10)
and the base (entry 11), as well as the uncatalyzed
condensation (entry 12). The first conclusion which can
be drawn from the results obtained is the dramatic
influence of the base in the catalyst on the rate and the
diastereoselectivity of the reactions. In the series of
Ag(I)/trialkylamine-catalyzed reactions of ketone 2b with
1 (entries 4-9), the remarkable influence on the diaste-
reoselectivity of the nature, particularly the steric bulk,
of the amine was revealed. Thus, Ag(I)/i-Pr2NEt-
catalyzed reaction (entry 6) afforded a mixture of oxazo-
lines 3,4b with modest diastereoselectivity, while the
condensation catalyzed with Ag(I)/NMe3 (entry 7) fur-
nished trans-oxazoline 3b in excellent de (96%). The
(25) CHIRAC is an abbreviation for (Ph3P)(CO)CpFeCOMe; see for
examples: (a) Davies, S. G.; Walker, J . C. J . Chem. Soc., Chem.
Commun. 1985, 209. (b) Liebeskind, L. S.; Welker, M. E.; Fengle, R.
W. J . Am. Chem. Soc. 1986, 108, 6328.
(26) (a) (Enantiocontrolled Synthesis of Fluoro-Organic Compounds;
Tetrahedron: Asymmetry, Special Issue; Hayashi, T.; Soloshonok, V.
A., Eds.) (b) Tetrahedron 1996, 52, No 1 (Fluoroorganic Chemistry:
Synthetic Challenges and Biomedical Rewards; Resnati, G.; Solos-
honok, V. A. Eds.; Tetrahedron Symposium-in-Print No. 58). (c) EPC-
Synthesis of Fluoro-Organic Compounds: Stereochemical Challenges
and Biomedicinal Targets; Soloshonok, V. A. Ed.; Wiley: Chichester,
scheduled to appear in 1997.
(27) Soloshonok, V. A. In Biomedical Frontiers of Fluorine Chem-
istry; Ojima, I.; McCarthy, J . R.; Welch, J . T. Eds.; ACS Symposium
Series 639; American Chemical Society: Washington, D. C., 1996, pp
26-41.
(28) Schlosser, M.; Michel, D. Tetrahedron 1996, 52, 99.
(29) The state of controversy in this area was best exemplified
recently by Schlosser: “All in all, the more results one sees, the more
one feels frustrated. Whenever the question about the effective size
of fluorine is asked again, a different answer is obtained.” See ref.
28.
(23) (a) Aaron, C.; Dull, D.; Schmiegel, J . L.; J aeger, D.; Ohashi, Y.
O.; Mosher, H. S. J . Org. Chem. 1967, 32, 2797. (b) Nasipuri, D.;
Bhattacharya, P. K. J . Chem. Soc. Perkin Trans 1 1977, 576. (c)
Midland, M. M.; Kazubski, A. J . Org. Chem. 1982, 47, 2495. (d)
Bucciarelli, M.; Forni, A.; Moretti, I.; Torre, G. Synthesis 1983, 897.
(e) Brown, H. C.; Pai, G. G. J . Org. Chem. 1985, 50, 1384. (f) Brown,
H. C.; Cho, B. T.; Park, W. S. J . Org. Chem. 1988, 53, 1231. (g) Corey,
E. J .; Bakshi, R. K. Tetrahedron Lett. 1990, 31, 611. (h) Chong, J . M.;
Mar, E. K. J . Org. Chem. 1991, 56, 893. (i) Corey, E. J .; Link, J . O.;
Sarshar, S.; Shao, Y. Tetrahedron Lett. 1992, 33, 7103. (j) Corey, E.
J .; Link, J . O.; Bakshi, R. K. Tetrahedron Lett. 1992, 33, 7107. (k)
Ramachandran, P. V.; Teodorovic, A. V.; Brown, H. C. Tetrahedron
1993, 49, 1725. (l) Ramachandran, P. V.; Teodorovic, A. V.; Gong, B.;
Brown, H. C. Tetrahedron: Asymmetry 1994, 5, 1075.
(24) PFCHIRAC is an abbreviation for [(C6F5)Ph2P](CO)CpFeCOMe
(Cp ) η5-cyclopentadienyl); Ojima, I.; Kwon, H. B. J . Am. Chem. Soc.
1988, 110, 5617.

Transition Metal/Base-Catalyzed Aldol Reactions
J . Org. Chem., Vol. 62, No. 11, 1997 3477
reaction catalyzed by Ag(I)/TEA showned lower stereo-
selectivity (entry 5). In contrast, the condensation of 2b
with isocyanoacetate 1, catalyzed by 1 mol % of Ag(I)/
PyH or Ag(I)/PS, proceeded with low reaction rates and
gave a mixture of oxazolines 3 and 4 with poor diaste-
reoselectivities (entries 8,9). The effect of the metal used
for catalyzing the reaction of 1 with 2b was less pro-
nounced (entries 3-5). Surprisingly, the reactions cata-
lyzed separately by CuCl and TEA showed quite different
values of the diastereoselectivity (entry 10 vs 11). Con-
sidering also the stereochemical outcome of the uncata-
lyzed condensation (entry 12), we can assume that the
use of a sterically undemanding base is essential for high
reaction rate and diastereoselectivity to be achieved.
Presumably, the presence of the base is critical for
organization of the transition states A and B and their
thermodynamic preferences are influenced by the ster-
eochemical properties of the base. Despite the higher
diastereoselectivity in the reactions of 1 with ketone 2b
that was obtained with the use of trimethylamine (entry
7 vs 5), for the rest of the study we chose to apply TEA,
since the latter is much more easy to handle and its
accurate concentration can be controlled.
The use of the metal in the catalyst and its nature were
also found to be important for high diastereoselectivity
to be achieved. Thus, in all condensations catalyzed by
TEA alone, the stereoselectivities obtained were sizably
lower compared with that of transition metal/TEA-
catalyzed reactions (entry 27 vs 26, 43 vs 42, 45 vs 44).
The reaction of 1 with keto ester 2w is particularly
noticeable, since it did not proceed at all if the transition
metal is not present (entry 52 vs 49-51). Selection of
the proper metal for each ketone condensation allowed
us to improve the diastereoselectivity in many cases. We
have found that the stereochemical outcome of the
chlorodifluoroacetophenone 2g reaction with isocyano-
acetate 1 is very sensitive to the metal used, varying from
70-84% de (entries 23-25) to virtually stereospecific
formation of trans-oxazoline 3g, if catalyzed by CuCl
(entry 26). The same catalyst, CuCl/NEt3, was shown
to be the best for the reactions of aryl perfluoroalkyl
ketones 2h ,i, giving rise to the targeted trans-3h ,i as the
single reaction products with excellent chemical yields
(entries 29, 30). A similar but even more pronounced
change in the diastereoselectivity, depending on the
metal used, was observed also in the reactions of 1 with
trifluoromethyl alkyl ketones. Here Cu(I) or Au(I)/TEA
catalysts gave a mixture of oxazolines 3,4s,t,w with
remarkably lower diastereoselectivity (entries 41, 46, 49,
50) as compared with the synthetically useful trans
stereoselectivity (90% de and over) obtained in the
reactions catalyzed by Ag(I)/TEA (entries 42, 44, 48, 51).
Interestingly, even in the reaction of 1 with ketone 2u ,
where the trifluoromethyl does not play the role of a
diastereodirecting group, an application of Ag(I)/TEA as
the catalyst increased the ratio of trans-oxazoline 3u in
a mixture with the dominant 4u .
Su m m a r y. The results presented here indicate that
the stereochemical outcome of the transition metal/base-
catalyzed aldol type addition reaction between methyl
isocyanoacetate 1 and prochiral ketones 2a -z is subject
to a number of substrate and catalyst influences, includ-
ing the steric and electronic/electrostatic properties of the
substituents on the starting ketone and the nature of
both metal and base in the catalyst. The results suggest
that some of the factors, in essence, could be manipulated
to afford synthetically useful diastereoselectivity in these
reactions.
Exp er im en ta l Section
Gen er a l. All reagents, unless otherwise stated, are com-
mercially available and were used as received. p-Methoxy-
phenyl trifluoromethyl ketone (2f),30 1,1,1-trifluoro-4-phenyl-
3-butyn-2-one (2v),31 n-heptyl trifluoromethyl ketone (2s),23k
n-octyl trifluoromethyl ketone (2t ),23k cyclohexyl trifluoro-
methyl ketone (2u ),23k and Au(c-HexNC)2]BF432 were prepared
according to the corresponding literature procedures. Melting
points were taken in open capillaries and are uncorrected.
Monitoring of the reactions by GLC was performed using a
fused silica capillary column. 1H and 19F NMR spectra were
measured at 299.95 and 282.24 MHz, respectively. Unless
indicated, NMR spectra were taken in CDCl3 solutions using
tetramethylsilane (TMS) and CFCl3 as the internal standards.
X-ray analysis of the compound 7 was performed on an
Rigaku AFC-5S diffractometer. A total of 1386 unique reflec-
tions were considered and used in the analysis. The structure
was solved by the TEXSAN method. The R and Rw factors
were of 0.056 and 0.052. A number of heavy atoms were
refined (31) (C, N, O, F). Crystals of 7 were grown from diethyl
ether and pentane. Crystal data for 7: C18H11F8NO4, mono-
clinic, space group C2/c. Radiation: graphite monochroma-
tized Cu KR radiation. Crystal size: 0.8 × 0.2 × 0.1 mm3.
Unit cell: a ) 29.111(6), b ) 6.995(1), c ) 19.583(4) Å, â )
110.01(2)°, V ) 3747 Å3, Z ) 8, Dx ) 1.621 Mg cm-3, µ for Cu
33
KR ) 14.7 cm-1
.
Unless otherwise stated, yields refer to isolated yields of
products of greater than 95% purity as estimated by capillary
1
GC and/or H and 19F NMR spectrometry. All new compounds
were characterized by 1H and 19F NMR and elemental analysis.
Ald ol Rea ction s of r-Meth yl Isocya n oa ceta te (1) w ith
Keton es 2a -z. Gen er a l P r oced u r e. To a magnetically
stirred solution of the corresponding transition metal/base (1-
10 mol %, as indicated in Table 1) and methyl isocyanoacetate
(typically 1 mmol) in 2 mL of freshly distilled 1,2-dichloro-
ethane was added an appropriate ketone 2 (1.1 mmol) under
dry nitrogen atmosphere at room temperature (20-22 °C).
Stirring was continued until complete consumption of the
starting 1, or the reaction mixture was stirred for a longer
time (Table 1) to attain a suitable conversion. The progress
of the reaction was monitored by GLC. The solvent was
removed in vacuo and aldol products were isolated via bulb-
to-bulb distillation. The ratio of diastereomers 3/4 was
determined by GLC analysis on the crude reaction mixture,
before removal of the solvent and distillation, and by NMR
(1H and 19F, where possible) analysis on the distilled products.
The ratios determined by NMR analysis were comparable
((1%) with those determined by GLC analysis. NMR spectra
and microanalytical data for the oxazolines 3,4a -z are listed
below.
Given the diversity of the coordination numbers of the
transition metal used, as well as the ligands they are
bearing, it is unlikely that a single mechanistic rationale
for their influence in the stereodiscrimination process
through the transition states A and B could be proposed.
However, the results obtained strongly suggest that Cu(I)
and Ag(I) are rather superior over the rest of the metals
explored and could be recommended, in a complex with
TEA, as the catalysts of choice for this type of diastereo-
selective aldol reaction of isocyanoacetate 1 with prochiral
ketones.
4-(Me t h o x y c a r b o n y l)-5-m e t h y l-5-p h e n y l-2-o x a zo -
lin es (3,4a ). trans-3a : 1H NMR δ 1.86 (br s, 3 H), 3.16 (s, 3
H), 4.65 (d, J ) 1.8 Hz, 1 H), 7.21 (d, J ) 1.8 Hz, 1 H), 7.26-
7.57 (m, 5 H). cis-4a : 1H NMR δ 1.61 (br s, 3 H), 3.86 (s, 3
(30) Creary, X. J . Org. Chem. 1987, 52, 5026.
(31) Ramachandran, P. V.; Gong, B.; Teodorovic, A. V.; Brown, H.
C. Tetrahedron: Asymmetry 1994, 5, 1061.
(32) Bonati, F.; Minghetti, G. G. Gazz. Chim. Ital. 1973, 103, 373.
(33) Atomic coordinates for the structure (9)have been deposited
with the Cambridge Crystallographic Data Centre. The coordinates
can be obtained, on request from The Director, Crystallographic Data
Centre, 12 Union Road, Cambridge, CB2 1EW, UK.

3478 J . Org. Chem., Vol. 62, No. 11, 1997
Soloshonok et al.
H), 4.84 (d, J ) 2.0 Hz, 1 H), 7.11 (d, J ) 2.0 Hz, 1 H), 7.26-
7.57 (m, 5 H). Anal. Calcd for C12H13NO3: C, 65.74; H, 5.98;
N, 6.39. Found: C, 65.88; H, 6.09; N, 6.17.
4-(Meth oxyca r bon yl)-5-(tr iflu or om eth yl)-5-(p en ta flu o-
r op h en yl)-2-oxa zolin es (3,4b). trans-3b, 1H NMR δ 3,71 (s,
3 H), 5.32 (m, 1 H), 7.15 (d, J ) 1.8 Hz, 1 H). cis-4b: 1H NMR
δ 3.86 (s, 3 H), 5.31 (m, 1 H), 7.14 (d, J ) 1.8 Hz, 1 H). Anal.
Calcd for C12H5F8NO3: C, 39.69; H, 1.39; N, 3.86; F, 41.85.
Found: C, 39.83; H, 1.34; N, 4.02; F, 41.64.
4-(Meth oxyca r bon yl)-5-m eth yl-5-(p en ta flu or op h en yl)-
2-oxa zolin es (3,4c). trans-3c: 1H NMR δ 1.83 (t, J ) 1.8
Hz, 3 H), 3.54 (s, 3 H), 4.81 (d, J ) 1.6 Hz, 1 H), 7.09 (d, J )
1.6 Hz, 1 H). cis-4c: 1H NMR δ 1.78 (t, J ) 2.5 Hz, 3 H), 3.82
(s, 3 H), 4.91 (d, J ) 2.0 Hz, 1 H), 7.04 (d, J ) 2.0 Hz, 1 H).
Anal. Calcd for C12H8F5NO3: C, 46.62; H, 2.61; N, 4.53; F,
30.72. Found: C, 46.43; H, 2.67; N, 4.61; F, 31.00.
C8H13NO4: C, 51.33; H, 7.00; N, 7.48. Found: C, 51.41; H,
6.94; N, 7.50.
4-(Meth oxyca r bon yl)-5-(flu or om eth yl)-5-m eth yl-2-ox-
a zolin es (3,4l). trans-3l: 1H NMR δ 1.31 (d, J ) 2.0 Hz, 3
H), 3.80 (s, 3 H), 4.39, 4.51 (ABX, J AB ) 9.9 Hz, J AX ) 47.6
Hz, J BX ) 46.7 Hz, 2 H), 4.66 (d, J ) 2.0 Hz, 1 H), 6.96 (d, J
) 2.0 Hz, 1 H). cis-4l: 1H NMR δ 1.55 (d, J ) 2.0 Hz, 3 H),
3.79 (s, 3 H), 4.44, 4.53 (ABX, J AB ) 9.9 Hz, J AX ) 46.8 Hz,
J BX ) 45.2 Hz, 2 H), 4.48 (d, J ) 2.0 Hz, 1 H), 6.97 (d, J ) 2.0
Hz, 1 H). Anal. Calcd for C7H10FNO3: C, 48.00; H, 5.75; N,
8.00; F, 10.85. Found: C, 47.94; H, 5.77; N, 7.89; F, 10.76.
4-(Meth oxyca r bon yl)-5-(ch lor om eth yl)-5-m eth yl-2-ox-
a zolin es (3,4m ). trans-3m : 1H NMR δ 1.41 (s, 3 H), 3.65,
3.73 (AB, J AB ) 11.9 Hz, 2 H), 3.80 (s, 3 H), 4.73 (d, J ) 2.0
Hz, 1 H), 6.96 (d, J ) 2.0 Hz, 1 H). cis-4m : 1H NMR δ 1.56
(s, 3 H), 3.67, 3.75 (AB, J AB ) 11.9 Hz, 2 H), 3.79 (s, 3 H), 4.50
(d, J ) 2.0 Hz, 1 H), 6.97 (d, J ) 2.0 Hz, 1 H). Anal. Calcd
for C7H10ClNO3: C, 43.88; H, 5.26; N, 7.31; Cl, 18.50. Found:
C, 43.92; H, 5.27; N, 7.44; Cl, 18.39.
4-(Met h oxyca r b on yl)-5-(d ich lor om et h yl)-5-m et h yl-2-
oxa zolin es (3,4n ). trans-3n : 1H NMR δ 1.56 (s, 3 H), 3.81
(s, 3 H), 4.91 (d, J ) 2.0 Hz, 1 H), 5.73 (s, 1 H), 6.98 (d, J )
2.0 Hz, 1 H). cis-4n : 1H NMR δ 1.76 (s, 3 H), 3.80 (s, 3 H),
4.56 (m, 1 H), 6.20 (s, 1 H), one resonance is obscured. Anal.
Calcd for C7H9Cl2NO3: C, 37.19; H, 4.01; N, 6.20; Cl, 31.37.
Found: C, 37.24; H, 4.05; N, 6.27; Cl, 31.22.
tr a n s-4-(Meth oxycar bon yl)-5-(tr ich lor om eth yl)-5-m eth -
yl-2-oxa zolin e (3o): 1H NMR δ 1.75 (s, 3 H), 3.83 (s, 3 H),
5.07 (d, J ) 2.0 Hz, 1 H), 7.04 (d, J ) 2.0 Hz, 1 H). Anal.
Calcd for C7H8Cl3NO3: C, 32.28; H, 3.09; N, 5.38; Cl, 40.83.
Found: C, 32.34; H, 3.14; N, 5.41; Cl, 40.61.
4-(Me t h oxyca r b on yl)-5-(d ich lor om e t h yl)-5-(ch lor o-
m eth yl)-2-oxa zolin es (3,4p ). trans-3p : 1H NMR δ 3.81 (s,
3 H), 3.98 (AB, J ) 12.2 Hz, 2 H), 5.08 (d, J ) 2.0 Hz, 1 H),
6.21 (s, 1 H), 7.02 (br d, J ) 2.0 Hz, 1 H). cis-4p : 1H NMR δ
3.81 (s, 3 H), 4.06 (AB, J ) 11.9 Hz, 2 H), 5.04 (d, J ) 2.0 Hz,
1 H), 6.36 (s, 1 H), 7.02 (br d, J ) 2.0 Hz, 1 H). Anal. Calcd
for C7H8Cl3NO3: C, 32.28; H, 3.09; N, 5.38. Found: C, 32.37;
H, 3.12; N, 5.39.
4-(Met h oxyca r b on yl)-5-(t r iflu or om et h yl)-5-p h en yl-2-
oxa zolin es (3,4d ). trans-3d : 1H NMR δ 3.27 (s, 3 H), 5.24
(d, J ) 2.0 Hz, 1 H), 7.24 (d, J ) 2.0 Hz, 1 H), 7.33-7.41 (m,
3 H), 7.42-7.48 (m, 2 H); 19F NMR δ -80.05 (s). cis-4d : 1H
NMR δ 3.91 (s, 3 H), 5.14 (d, J ) 2.2 Hz, 1 H), 7.16 (d, J ) 2.2
Hz, 1 H), 7.33-7.41 (m, 3 H), 7.42-7.48 (m, 2 H); 19F NMR δ
-75.97 (s). Anal. Calcd for C12H10F3NO3: C, 52.76; H, 3.69;
N, 5.13; F, 20.86. Found: C, 52.65; H, 3.71; N, 5.17; F, 20.94.
4-(Meth oxyca r bon yl)-5-(tr iflu or om eth yl)-5-[4-(tr iflu o-
r om eth yl)p h en yl]-2-oxa zolin es (3,4e). trans-3e: 1H NMR
δ 3.27 (s, 3 H), 5.26 (d, J ) 2.0 Hz, 1 H), 7.26 (d, J ) 2.0 Hz,
1 H), 7.55-7.66 (m, 4 H). cis-4e: 1H NMR δ 3.93 (s, 3 H),
5.10 (d, J ) 2.2 Hz, 1 H), 7.19 (d, J ) 2.2 Hz, 1 H), 7.70-7.90
(m, 4 H). Anal. Calcd for C13H9F6NO3: C, 45.76; H, 2.66; N,
4.11; F, 33.41. Found: C, 45.69; H, 2.61; N, 4.09; F, 33.34.
4-(Met h oxyca r b on yl)-5-(t r iflu or om et h yl)-5-(4-m et h -
oxyp h en yl)-2-oxa zolin es (3,4f). trans-3f: 1H NMR δ 3.33
(s, 3 H), 3.81 (s, 3 H), 5.21 (d, J ) 2.0 Hz, 1 H), 6.86 (d, J )
2.0 Hz, 1 H), 7.20-7.42 (m, 4 H); 19F NMR δ -80.07 (s). cis-
4f: 1H NMR δ 3.91 (s, 3 H), 3.82 (s, 3 H), 5.16 (d, J ) 2.0 Hz,
1 H), 6.84 (d, J ) 2.0 Hz, 1 H), 7.20-7.42 (m, 4 H); 19F NMR δ
-75.93 (s). Anal. Calcd for C13H12F3NO4: C, 51.49; H, 3.99;
N, 4.62; F, 18.80. Found: C, 51.53; H, 3.94; N, 4.67; F, 18.71.
4-(Meth oxycar bon yl)-5-(ch lor odiflu or om eth yl)-5-ph en -
yl-2-oxa zolin es (3,4g). trans-3g: 1H NMR δ 3.28 (s, 3 H),
5.28 (d, J ) 2.0 Hz, 1 H), 7.24 (d, J ) 2.0 Hz, 1 H), 7.34-7.41
(m, 3 H), 7.42-7.48 (m, 2 H); 19F NMR δ -66.00, -67.10 (AB,
J ) 173.4 Hz). cis-4g: 1H NMR δ 3.93 (s, 3 H), 5.16 (d, J )
2.3 Hz, 1 H), 7.16 (d, J ) 2.3 Hz, 1 H), 7.34-7.41 (m, 3 H),
7.42-7.48 (m, 2 H); 19F NMR δ -62.75, -63.60 (AB, J ) 173.0
Hz). Anal. Calcd for C12H10ClF2NO3: C, 49.76; H, 3.48; N,
4.84; F, 13.12. Found: C, 49.77; H, 3.50; N, 4.85; F, 13.11.
4-(Meth oxyca r bon yl)-5-(p en ta flu or oeth yl)-5-p h en yl-2-
oxa zolin es (3,4h ). trans-3h : 1H NMR δ 3.27 (s, 3 H), 5.41
(d, J ) 2.0 Hz, 1 H), 7.21 (d, J ) 2.0 Hz, 1 H), 7.36-7.41 (m,
3 H), 7.43-7.46 (m, 2 H); 19F NMR δ -79.02 (br s, 3F),
-123.01, -125.16 (AB, J ) 302.3 Hz, 2 F). cis-4h : 1H NMR
δ 3.91 (s, 3 H), 5.07 (d, J ) 2.2 Hz, 1 H), 7.18 (d, J ) 2.2 Hz,
1 H), 7.36-7.41 (m, 3 H), 7.43-7.46 (m, 2 H). Anal. Calcd
for C13H10F5NO3: C, 48.31; H, 3.12; N, 4.33; F, 29.39. Found:
C, 48.45; H, 3.21; N, 4.44; F, 29.17.
tr a n s-4-(Met h oxyca r b on yl)-5-(h ep t a flu or op r op yl)-5-
p h en yl-2-oxa zolin e: 1H NMR δ 3.29 (s, 3 H), 5.41 (d, J )
2.2 Hz, 1 H), 7.20 (d, J ) 2.2 Hz, 1 H), 7.36-7.42 (m, 3 H),
7.43-7.47 (m, 2 H); 19F NMR δ -81.60 (m, 3F), -118.81,
-121.86 (AB, J ) 288.2 Hz, 2 F), -120.70, -125.82 (ABX, J
) 289.1 Hz, J ) 11.5 Hz, 2 F). Anal. Calcd for C14H10F7NO3:
C, 45.05; H, 2.70; N, 3.75; F, 35.63. Found: C, 45.12; H, 2.74;
N, 3.81; F, 35.49.
4-(Met h oxyca r b on yl)-5-cya n o-5-p h en yl-2-oxa zolin es
(3,4j). trans-3j: 1H NMR δ 3.20 (s, 3 H), 5.41 (d, J ) 2.0 Hz,
1 H), 7.29 (d, J ) 2.0 Hz, 1 H), 7.42-7.74 (m, 5 H). cis-3j: 1H
NMR δ 3.90 (s, 3 H), 5.00 (d, J ) 2.3 Hz, 1 H), 7.19 (d, J ) 2.3
Hz, 1 H), 7.42-7.74 (m, 5 H). Anal. Calcd for C12H10N2O3:
C, 62.61; H, 4.38; N, 12.17. Found: C, 62.79; H, 4.42; N, 12.19.
4-(Met h oxyca r b on yl)-5-(m et h oxym et h yl)-5-m et h yl-2-
oxa zolin es (3,4k ). trans-3k : 1H NMR δ 1.23 (s, 3 H), 3.38
(s, 3 H), 3.41-3.44 (m, 2H), 3.73 (s, 3 H), 4.62 (d, J ) 2.0 Hz,
1 H), 6.89 (d, J ) 2.0 Hz, 1 H). cis-4k : 1H NMR δ 1.47 (s, 3
H), 3.23 (s, 3 H), 3.48-3.52 (m, 2H), 3.72 (s, 3 H), 4.33 (d, J )
2.0 Hz, 1 H), 6.90 (d, J ) 2.0 Hz, 1 H). Anal. Calcd for
4-(Met h oxyca r b on yl)-5-(t r iflu or om et h yl)-5-m et h yl-2-
oxa zolin es (3,4q). trans-3q: 1H NMR δ 1.51 (s, 3 H), 3.82
(s, 3 H), 4.90 (d, J ) 2.0 Hz, 1 H), 6.98 (d, J ) 2.0 Hz, 1 H); 19
F
NMR δ -80.78 (s). cis-4q: 1.75 (s, 3 H), 3.81 (s, 3 H), 4.64 (d,
J ) 2.0 Hz, 1 H), 7.01 (d, J ) 2.0 Hz, 1 H); 19F NMR δ -76.57
(s). Anal. Calcd for C7H8F3NO3: C, 39.82; H, 3.82; N, 6.64;
F, 27.00. Found: C, 39.85; H, 3.84; N, 6.69; F, 26.93.
tr a n s-4-(Met h oxyca r b on yl)-5-(h ep t a flu or op r op yl)-5-
m eth yl-2-oxa zolin e (3r ): 1H NMR δ 1.53 (s, 3 H), 3.84 (s, 3
H), 4.95 (d, J ) 2.0 Hz, 1 H), 6.99 (d, J ) 2.0 Hz, 1 H). Anal.
Calcd for C9H8F7NO3: C, 34.74; H, 2.59; N, 4.50; F, 42.74.
Found: C, 34.85; H, 2.62; N, 4.57; F, 42.51.
4-(Meth oxyca r bon yl)-5-(tr iflu or om eth yl)-5-n -h ep tyl-2-
oxa zolin es (3,4s). trans-3s: 1H NMR δ 0.87 (t, J ) 6.5 Hz,
3 H), 1.24-1.40 (m, 10 H), 1.82-1.97 (m, 2 H), 3.81 (s, 3 H),
4.90 (d, J ) 2.0 Hz, 1 H), 6.98 (d, J ) 2.0 Hz, 1 H); 19F NMR
δ -80.81 (s). cis-4s: 1H NMR δ 0.88 (t, J ) 6.6 Hz, 3 H), 1.24-
1.40 (m, 10 H), 1.82-2.20 (m, 2 H), 3.80 (s, 3 H), 4.72 (d, J )
2.0 Hz, 1 H), 6.99 (d, J ) 2.0 Hz, 1 H); 19F NMR δ -76.53 (s).
Anal. Calcd for C13H20F3NO3: C, 52.87; H, 6.83; N, 4.74; F,
19.30. Found: C, 52.94; H, 6.89; N, 4.67; F, 19.25.
4-(Met h oxyca r b on yl)-5-(t r iflu or om et h yl)-5-n -oct yl-2-
oxa zolin es (3,4t). trans-3t: 1H NMR δ 0.87 (t, J ) 6.5 Hz, 3
H), 1.25-1.37 (m, 12 H), 1.82-1.98 (m, 2 H), 3.82 (s, 3 H),
4.91 (d, J ) 2.3 Hz, 1 H), 6.99 (d, J ) 2.3 Hz, 1 H); 19F NMR
δ -80.80 (s). cis-4t: 1H NMR δ 0.88 (t, J ) 6.5 Hz, 3 H), 1.25-
1.41 (m, 12 H), 1.82-2.20 (m, 2 H), 3.80 (s, 3 H), 4.72 (d, J )
2.3 Hz, 1 H), 7.00 (d, J ) 2.3 Hz, 1 H); 19F NMR δ -76.55 (s).
Anal. Calcd for C14H22F3NO3: C, 54.36; H, 7.17; N, 4.53; F,
18.43. Found: C, 54.44; H, 7.21; N, 4.50; F, 18.32.
4-(Meth oxyca r bon yl)-5-(tr iflu or om eth yl)-5-cycloh exyl-
2-oxa zolin es (3,4u ). trans-3u : 1H NMR δ 0.99-2.20 (m, 11
H), 3.82 (s, 3 H), 4.94 (d, J ) 2.0 Hz, 1 H), 7.00 (d, J ) 2.0 Hz,
1 H); 19F NMR δ -80.82 (s). cis-4u : 0.93-2.20 (m, 11 H), 3.80
(s, 3 H), 4.80 (d, J ) 2.0 Hz, 1 H), 7.00 (d, J ) 2.0 Hz, 1 H); 19
F
NMR δ -76.53 (s). Anal. Calcd for C12H16F3NO3: C, 51.61;
H, 5.78; N, 5.02; F, 20.41. Found: C, 51.63; H, 5.77; N, 5.06;
F, 20.38.

Transition Metal/Base-Catalyzed Aldol Reactions
J . Org. Chem., Vol. 62, No. 11, 1997 3479
tr a n s-4-(Met h oxyca r b on yl)-5-(t r iflu or om et h yl)-5-(2-
p h en yleth yn yl)-2-oxa zolin e (3v): 1H NMR δ, 3.81 (s, 3 H),
5.10 (d, J ) 2.1 Hz, 1 H), 7.08 (d, J ) 2.1 Hz, 1 H), 7.26-7.48
1 H); 19F NMR δ -78.88 (s, 3 F), -120.88 (m, 2 F). Anal. Calcd
for C13H12F5NO4: C, 45.76; H, 3.54; N, 4.11; F, 27.84. Found:
C, 45.83; H, 3.60; N, 4.05; F, 27.67.
(m,
5 δ -80.99 (s). Anal. Calcd for
H); 19F NMR
(2R*,3R*)-Met h yl N-for m yl-3-(t r iflu or om et h yl)-3-n -
h ep t ylser in a t e (5s): yield 63% (from 3,4s); mp 55 °C; 1H
NMR (CD3COCD3) δ 0.89 (t, J ) 7.3 Hz, 3 H), 1.30 (m, 8 H),
1.52 (m, 2 H), 1.80 (m, 2 H), 3.76 (s, 3 H), 5.02 (br s, 1 H), 8.24
(s, 1 H). Anal. Calcd for C13H22F3NO4: C, 49.83; H, 7.08; N,
4.47; F, 18.19. Found: C, 49.88; H, 7.12; N, 4.43; F, 18.02.
(2R*,3R*)-Met h yl N-for m yl-3-(t r iflu or om et h yl)-3-(2-
p h en yleth yn yl)ser in a te (5v): yield 97% (from 3v); mp 112-
113 °C; 1H NMR δ 3.83 (s, 3 H), 4.27 (s, 1 H), 5.31 (d, J ) 8.0
Hz, 1 H), 6.56 (d, J ) 8.0 Hz, 1 H), 7.38-7.47 (m, 5 H), 8.28
(s, 1 H). Anal. Calcd for C14H12F3NO4: C, 53.34; H, 3.84; N,
4.44; F, 18.08. Found: C, 53.42; H, 3.83; N, 4.49; F, 17.94.
Hyd r olysis of t h e Oxa zolin es tr a n s-3d ,g,h ,q ,s,t or
(2R*,3R*)-Meth yl N-F or m yl-3,3-d isu bstitu ted -ser in a tes
5d,g,h ,s to (2R*,3R*)-3,3-Disu bstitu ted-ser in es 6d,g,h ,q,s,t.
The hydrolysis was accomplished according to the general
procedure described in ref 9a. The targeted amino acids were
purified by recrystallization from water. NMR spectra, melt-
ing points and microanalytical data for 6d ,g,h ,q,s,t are listed
below.
C14H10F3NO3: C, 56.57; H, 3.39; N, 4.71; F, 19.18. Found: C,
56.55; H, 3.42; N, 4.78; F, 19.09.
4-(Meth oxyca r bon yl)-5-(tr iflu or om eth yl)-5-[2-(eth oxy-
ca r bon yl)eth yl]-2-oxa zolin es (3,4w ). trans-3w : 1H NMR
δ 1.24 (t, J ) 7.3 Hz, 3H), 3.13, 3.42 (AB, J ) 17.5 Hz, 2H),
3.79 (s, 3H), 4.11 (m, J ) 3.3 and 7.3 Hz, 2H), 4.94 (d, J ) 2.6
Hz, 1H), 6.94 (d, J ) 2.6 Hz, 1H); 19F NMR δ -80.85 (s). cis-
4w : 1H NMR δ 1.28 (t, J ) 7.0 Hz, 3H), 3.03, 3.16 (AB, J )
16.0 Hz, 2H), 3.81 (s, 3H), 4.20 (q, J ) 7.0 Hz, 2H), 5.43 (d, J
) 2.3 Hz, 1H), 6.96 (d, J ) 2.3 Hz, 1H); 19F NMR δ -76.58 (s).
Anal. Calcd for C10H12F3NO5: C, 42.41; H, 4.27; N, 4.95; F,
20.13. Found: C, 42.53; H, 4.25; N, 4.99; F, 20.08.
4-(Met h oxyca r b on yl)-5-et h yl-5-m et h yl-2-oxa zolin es
(3,4x). trans-3x: 1H NMR δ 0.97 (t, J ) 7.6 Hz, 3 H), 1.27 (s,
3 H), 1.79 (q d, J ) 7.6 Hz, J ) 1.8 Hz, 2H), 3.76 (s, 3 H), 4.42
(d, J ) 2.0 Hz, 1 H), 6.92 (m, 1 H). cis-4x: 1H NMR δ 0.94 (t,
J ) 7.6 Hz, 3 H), 1.48 (s, 3 H), 1.55, 1.65 (ABX, J ) 7.6 Hz, J
) 24.5 Hz, 2H), 3.75 (s, 3 H), 4.40 (d, J ) 2.2 Hz, 1 H), 6.93
(m, 1 H). By 1H NMR the product contains about 5% of
starting methyl R-isocyanoacetate (1): 1H NMR δ 3.82 (s, 3H),
4.23 (s, 2H).
(2R*,3R*)-3-(Tr iflu or om eth yl)-3-ph en ylser in e (6d): yield
62% from 3,4d , 89% from (2R*,3R*)-5d ; mp 204-207 °C (dec);
1H NMR (CD3SOCD3) δ 3.90 (s, 1 H), 7.39 (m, 3 H), 7.60 (m,
2 H). Anal. Calcd for C10H10F3NO3: C, 48.20; H, 4.05; N, 5.62.
Found: C, 47.93; H, 4.00; N, 5.41.
4-(Me t h oxyca r b on yl)-5-n -p r op yl-5-m e t h yl-2-oxa zo-
lin es 3,4y). trans-3y: 1H NMR δ 0.96 (t, J ) 7.3 Hz, 3 H),
1.27 (s, 3 H), 1.34-1.48 (m, 2H), 1.71-1.78 (m, 2H), 3.76 (s, 3
H), 4.44 (d, J ) 1.8 Hz, 1 H), 6.93 (m, 1 H). cis-4y: 1H NMR
δ 0.90 (t, J ) 7.6 Hz, 3 H), 1.49 (s, 3 H), 1.55, 1.65 (ABX, J )
7.6 Hz, J ) 24.5 Hz, 2H), 3.75 (s, 3 H), 4.40 (d, J ) 2.0 Hz, 1
(2R*,3R*)-3-(Ch lor odiflu or om eth yl)-3-ph en ylser in e (6g)
(ch lor oh yd r a te): yield 74% from 3g, 79% from (2R*,3R*)-
5g; mp 172-173 °C (dec); 1H NMR (CD3SOCD3) δ 4.63 (s, 1
1
H), 6.93 (m, 1 H). By H NMR the product contains about 8%
H), 7.42 (m,
3 H), 7.61 (m, 2 H). Anal. Calcd for
of starting methyl R-isocyanoacetate (1).
C10H10ClF2NO3‚HCl: C, 39.76; H, 3.67; Cl, 23.47; N, 4.64.
4-(Met h oxyca r b on yl)-5-cycloh exyl-5-m et h yl-2-oxa zo-
lin es (3,4z). trans-3z: 1H NMR δ 0.95-1.97 (m, 11 H), 1.24
(s, 3 H), 3.76 (s, 3 H), 4.52 (d, J ) 2.0 Hz, 1 H), 6.92 (d, J )
2.0 Hz, 1 H). cis-4z: 1H NMR δ 1.35 (s, 3H), 3.75 (s, 3 H),
4.35 (d, J ) 1.9 Hz, 1 H), 6.96 (d, J ) 1.9 Hz, 1 H), c-Hex
resonances are obscured by those of the major isomer. Anal.
Calcd for C12H19NO3: C, 63.98; H, 8.50; N, 6.22. Found: C,
64.05; H, 8.53; N, 6.17.
Found: C, 39.67; H, 3.61; Cl, 23.69; N, 4.55.
(2R*,3R*)-3-(Tr iflu or om et h yl)-3-m et h ylser in e (6q ):6
yield 78% from 3q; mp 226-229 °C (dec); 1H NMR (CD3COCD3)
δ 1.22 (s, 3H), 4.22 (s, 1H); 19F NMR (CD3SOCD3) δ -77.10
(s).
(2S,3S)-â-(Tr iflu or om eth yl)-â-n -h ep tylser in e (6s) (h y-
d r och lor id e):6c yield 65% from 3,4s, 81% from 5s; mp 178-
179 °C; 1H NMR (CD3COCD3) δ 0.89 (t, J ) 7.1 Hz,
3H), 1.17-1,73 (m, 12H), 4.32 (s, 1H); 19F NMR (CD3SOCD3)
-75.13 (s).
Gen er a l Meth od of Hyd r olysis of th e Oxa zolin es
tr a n s-3d ,g,h ,s,v to (2R*,3R*)-Meth yl N-F or m yl-3,3-d isu b-
st it u t ed -ser in a t es 5d ,g,h ,s,v. The starting oxazoline
5d ,g,h ,s,v (typically 2 mmol) was dissolved in 3 mL of MeOH
and 2 mL of water, and one drop of 2 N HCl was added under
stirring at ambient temperature. The progress of the hydroly-
sis was monitored by TLC or GLC, and upon completion MeOH
was evaporated in vacuo. The products 5d ,g,h ,s,v were
extracted by ethyl acetate and purified to a diastereomerically
pure state by recrystallization from ethyl acetate/n-hexane 2/1.
NMR spectra, melting point and microanalytical data for
5d ,g,h ,s,v are listed below.
(2S,3S)-â-(Tr iflu or om et h yl)-â-n -oct ylser in e (6t ) (h y-
d r och lor id e):6c yield 59% from 3,4t; mp 166-169 °C; 1H
NMR (CD3COCD3) δ 0.86 (t, J ) 7.1 Hz, 3H), 1.18-1.78 (m,
14H), 4.31 (s, 1H); 19F NMR (CD3SOCD3) δ -75.11 (s).
(2R*,3R*)-Meth yl N-Ben zoyl-3-(tr iflu or om eth yl)-3-(pen -
ta flu or op h en yl)ser in a te (7). To a solution of 3,4b (diaste-
reoisomers ratio 9:1) (240 mg) in MeOH (4 mL) was added
concd HCl (0.7 mL). The mixture was stirred for 2 h at 50 °C
and then evaporated in vacuo to yield white solid. To the
crystalline residue were added CH2Cl2 (6 mL), NEt3 (0.4 mL),
and PhCOCl (0.16 mL), and the mixture was stirred for 2 h at
ambient temperature. After the usual workup, the crude
product was purified by preparative TLC on silica gel using
the mixture hexane/ethyl acetate (3:1) as an eluent: 1H NMR
δ 3.72 (s, 3 H), 4.77 (t, J ) 4.1 Hz, 1 H), 6.13 (d, J ) 10.1 Hz,
1 H), 7.02 (br d, J ) 10.1 Hz, 1 H), 7.43-7.60 (m, 3 H), 7.76-
7.83 (m, 2 H). Anal. Calcd for C18H11F8NO4: C, 47.28; H, 2.43;
N, 3.06; F, 33.24. Found: C, 47.34; H, 2.41; N, 3.11; F, 33.01.
(2R*,3R*)-Meth yl N-for m yl-3-(tr iflu or om eth yl)-3-p h en -
1
ylser in a te (5d ): yield 84% (from 3,4d ); mp 124-125 °C; H
NMR δ 3.47 (s, 3 H), 4.71 (s, 1 H), 5.59 (d, J ) 8.3 Hz, 1 H),
6.83 (d, J ) 8.3 Hz, 1 H), 7.41 (m, 3 H), 7.60 (m, 2 H), 8.25 (s,
1 H); 19F NMR δ -75.91 (s). NMR data for (2R*,3S*)-
diastereomer: 1H NMR δ 3.87 (s, 3 H), 4.76 (s, 1 H), 5.45 (d,
J ) 8.2 Hz, 1 H), 6.19 (d, J ) 8.2 Hz, 1 H), 7.41 (m, 3 H), 7.60
(m, 2 H), 7.90 (s, 1 H); 19F NMR δ -77.34 (s). Anal. Calcd for
C12H12F3NO4: C, 49.49; H, 4.15; N, 4.81; F, 19.57. Found: C,
49.51; H, 4.17; N, 4.78; F, 19.44.
Ack n ow led gm en t. This research was supported by
grants from the Ministry of Education and Uehara
Memorial Foundation, J apan, the International Science
Foundation (U6M000, U6M200), USA and INTAS (Net-
work 93-799), and UC. V.A.S thanks the J apanese
Society for Promotion of Science (J SPS) and the Science
and Technology Agency of J apan (STA) for the award
of the corresponding Fellowships.
(2R*,3R*)-Meth yl N-for m yl-3-(ch lor od iflu or om eth yl)-
3-p h en ylser in a te (5g): yield 91% (from 3g); mp 132-133 °C;
1H NMR δ 3.44 (s, 3 H), 4.86 (s, 1 H), 5.67 (d, J ) 8.3 Hz, 1
H), 6.95 (d, J ) 8.3 Hz, 1 H), 7.41 (m, 3 H), 7.61 (m, 2 H), 8.23
(s, 1 H). Anal. Calcd for C12H12ClF2NO4: C, 46.84; H, 3.93;
N, 4.55. Found: C, 46.93; H, 3.91; N, 4.47.
(2R*,3R*)-Meth yl N-for m yl-3-(pen taflu or oeth yl)-3-ph en -
ylser in a te (5h ): yield 89% (from 3h ); mp 121-125 °C; 1H
NMR δ 3.45 (s, 3 H), 4.70 (s, 1 H), 5.63 (d, J ) 8.6 Hz, 1 H),
6.63 (d, J ) 8.6 Hz, 1 H), 7.41 (m, 3 H), 7.56 (m, 2 H), 8.25 (s,
J O9623402