Dynamic ligand exchange of the lanthanide complex leading to structural and functional transformation: One-pot sequential catalytic asymmetric epoxidation-regioselective epoxide-opening process

Article abstract of DOI:10.1021/ja043770+

The characteristic property of the lanthanide complex, which easily undergoes a dynamic ligand exchange and alters its structure and function in situ, is described. After the completion of the catalytic asymmetric epoxidation of various α,β-unsaturated amides 2 in the presence of the Sm-(S)-BINOL-Ph₃-As=O (1:1:1) complex 1 (2-10 mol %), the addition of Me₃SiN₃ directly to the reaction mixture led to smooth epoxide-opening at room temperature, affording the corresponding anti-β-azido-α-hydroxyamide 4 in excellent overall yield (up to 99%) with complete regioselectivity and excellent enantiomeric excess (up to >99%). The key to the success of the sequential process was the in situ generation of the highly reactive samarium azide complex through dynamic ligand exchange. In situ IR spectroscopy and other experiments provided strong evidence that the samarium azide complex was generated. In addition, the relatively high Lewis basicity of the amide moiety had a key role in the high reactivity of both the epoxidation and the epoxide-opening reactions. Examinations of other nucleophiles such as sulfur or carbon nucleophiles as well as transformations of epoxide-opened products are also described.

Full text of DOI:10.1021/ja043770+

Published on Web 01/27/2005
Dynamic Ligand Exchange of the Lanthanide Complex
Leading to Structural and Functional Transformation:
One-Pot Sequential Catalytic Asymmetric
Epoxidation-Regioselective Epoxide-Opening Process
Shin-ya Tosaki, Riichiro Tsuji, Takashi Ohshima,* and Masakatsu Shibasaki*
Contribution from the Graduate School of Pharmaceutical Sciences, The UniVersity of Tokyo,
Hongo, Bunkyo-ku, Tokyo 113-0033, Japan
Received October 13, 2004; E-mail: ohshima@mol.f.u-tokyo.ac.jp; mshibasa@mol.f.u-tokyo.ac.jp
Abstract: The characteristic property of the lanthanide complex, which easily undergoes a dynamic ligand
exchange and alters its structure and function in situ, is described. After the completion of the catalytic
asymmetric epoxidation of various R,â-unsaturated amides 2 in the presence of the Sm-(S)-BINOL-Ph3-
AsdO (1:1:1) complex 1 (2-10 mol %), the addition of Me3SiN3 directly to the reaction mixture led to
smooth epoxide-opening at room temperature, affording the corresponding anti-â-azido-R-hydroxyamide
4 in excellent overall yield (up to 99%) with complete regioselectivity and excellent enantiomeric excess
(up to >99%). The key to the success of the sequential process was the in situ generation of the highly
reactive samarium azide complex through dynamic ligand exchange. In situ IR spectroscopy and other
experiments provided strong evidence that the samarium azide complex was generated. In addition, the
relatively high Lewis basicity of the amide moiety had a key role in the high reactivity of both the epoxidation
and the epoxide-opening reactions. Examinations of other nucleophiles such as sulfur or carbon nucleophiles
as well as transformations of epoxide-opened products are also described.
cyanosilylation of ketones4a and Strecker reaction of ketoi-
mines4b-d catalyzed by the chiral gadolinium complex, in which
Introduction
Lanthanide metals are widely utilized for organic reactions.
Among them, the large coordination numbers of lanthanide
metals allow for the development of lanthanide salt-ligand
complexes as Lewis acid catalysts. One role of the ligand is
the improvement of existing catalyses. The reaction rate is often
increased by the addition of a suitable ligand (ligand-accelerating
effect).1 On the other hand, a nucleophilic reagent can be
exchanged for a labile ligand to generate another nucleophilic
complex. The nucleophilic complexes of lanthanide metals
prepared by modification of the ligand work as highly nucleo-
philic catalysts compared with conventional Lewis acid catalysts.
There are only a few examples of the use of this type of catalyst,
however, despite its high reactivity and unique properties.
Utimoto et al. reported that the ytterbium catalyst, prepared by
the reaction of Bu3Yb or Yb(O-i-Pr)3 with Me3SiCN, was highly
effective for the regioselective ring opening of epoxides and
aziridines with cyanide and cyanation of carbonyl compounds.2
Ytterbium cyanide was postulated to be the active cyanating
reagent.3 Our group recently reported the enantioselective
gadolinium cyanide acts as the nucleophile.5
On the basis of the nature of the lanthanide complex, such
as moderate Lewis acidity, multifunctionality, large coordination
numbers, and fast ligand exchange ability, we anticipated that
subsequent addition of other reagents would alter the structure
and function of the lanthanide complex in situ through dynamic
ligand exchange to promote different reactions successively. In
this regard, the lanthanide BINOL complex appeared to be a
suitable catalyst. We previously reported a general and highly
enantioselective epoxidation of R,â-unsaturated carbonyl com-
pounds catalyzed by alkali-metal free lanthanide BINOL
(4) (a) Yabu, K.; Masumoto, S.; Yamasaki, S.; Hamashima, Y.; Kanai, M.;
Du, W.; Curran, D. P.; Shibasaki, M. J. Am. Chem. Soc. 2001, 123, 9908.
(b) Masumoto, S.; Usuda, H.; Suzuki, M.; Kanai, M.; Shibasaki, M. J.
Am. Chem. Soc. 2003, 125, 5634. (c) Kato, N.; Suzuki, M.; Kanai, M.;
Shibasaki, M. Tetrahedron Lett. 2004, 45, 3143. (d) Kato, N.; Suzuki, M.;
Kanai, M.; Shibasaki, M. Tetrahedron Lett. 2004, 45, 3153.
(5) The gadolinium catalyst had higher reactivity than the one prepared from
Ti(O-i-Pr)4, by which Me3SiCN itself works as a nucleophile. See ref 4a
and (a) Hamashima, Y.; Kanai, M.; Shibasaki M. J. Am. Chem. Soc. 2000,
122, 7412. (b) Hamashima, Y.; Kanai, M.; Shibasaki, M. Tetrahedron Lett.
2001, 42, 691.
(6) Epoxidation of R,â-unsaturated ketones: (a) Bougauchi, M.; Watanabe,
T.; Arai, T.; Sasai, H.; Shibasaki, M. J. Am. Chem. Soc. 1997, 119, 2329.
(b) Nemoto, T.; Ohshima, T.; Yamaguchi, K.; Shibasaki, M. J. Am. Chem.
Soc. 2001, 123, 2725. For other R,â-unsaturated carbonyl compounds,
see: (esters) (c) Jacobsen, E. N.; Deng, L.; Furukawa, Y.; Mart´ınez, L. E.
Tetrahedron 1994, 50, 4323. (d) Wu, X.-Y.; She, X.; Shi, Y. J. Am. Chem.
Soc. 2002, 124, 8792. N-Acylimidazolides: (e) Nemoto, T.; Ohshima, T.;
Shibasaki, M. J. Am. Chem. Soc. 2001, 123, 9474. N-Acylpyrroles:
(f) Kinoshita, T.; Okada, S.; Park, S.-R.; Matsunaga, S.; Shibasaki, M.
Angew. Chem., Int. Ed. 2003, 42, 4680. (g) Matsunaga, S.; Kinoshita, T.;
Okada, S.; Harada, S.; Shibasaki, M. J. Am. Chem. Soc. 2004, 126, 7559.
(1) Berrisford, D. J.; Bolm, C.; Sharpless, K. B. Angew. Chem., Int. Ed. Engl.
1995, 34, 1059.
(2) (a) Matsubara, S.; Onishi, H.; Utimoto, K. Tetrahedron Lett. 1990, 31,
6209. (b) Matsubara, S.; Kodama, T.; Utimoto, K. Tetrahedron Lett. 1990,
31, 6379. (c) Matsubara, S.; Takai, T.; Utimoto, K. Chem. Lett. 1991, 1447.
(3) Schaus and Jacobsen suggested the formation of a ytterbium cyanide species
in the asymmetric ring opening of meso-epoxides with Me3SiCN catalyzed
by the (pybox)YbCl3 complex. See: Schaus, S. E.; Jacobsen, E. N. Org.
Lett. 2000, 2, 1001.
9
10.1021/ja043770+ CCC: $30.25 © 2005 American Chemical Society
J. AM. CHEM. SOC. 2005, 127, 2147-2155
2147

A R T I C L E S
Tosaki et al.
Scheme 1. Postulated Catalytic Cycle of Enantioselective Epoxidation (Left Side) and the Following Sequential Epoxide-Opening Reaction
(Right Side) through Dynamic Ligand Exchange of the Lanthanide Complex
complexes.6 The postulated catalytic cycle is illustrated in the
left side of Scheme 1. The lanthanide BINOL complex is
prepared from Ln(O-i-Pr)3 and BINOL in a ratio of 1:1 (step
a). Subsequent addition of alkyl peroxide (ROOH) induces an
exchange of the alkoxide ligand on the lanthanide metal to
generate the active lanthanide peroxide complex (step b).
Coordination of the carbonyl to the lanthanide metal to activate
the electron-deficient olefin (step c), followed by the enantio-
selective 1,4-addition of lanthanide peroxide, afforded lanthanide
enolate (step d). The subsequent epoxide formation, followed
by dissociation from the lanthanide metal, regenerates the
lanthanide BINOL complex (step e). We assumed that the
subsequent addition of another reagent (R′3Si-Nu) after the
enantioselective epoxidation would generate a highly reactive
nucleophilic complex (Y2Ln-Nu) in situ through dynamic ligand
exchange (step f) and that the newly generated complex would
promote the catalytic regioselective epoxide-opening reaction
(steps g and h) in one reaction vessel.
The syntheses would be much more efficient if several bonds
could be formed in one sequence without isolating the inter-
mediates. The increased demand for a highly efficient and
environmentally benign synthetic process requires the develop-
ment of catalysts that promote different reactions in one reaction
vessel.7,8 This type of catalysis would minimize waste (solvents,
reagents, silica gel, etc.) and make waste management unneces-
sary, because compared to stepwise reactions, the amount of
chemicals and energy utilized would substantially decrease.
In this article, we describe the characteristic property of the
lanthanide complex, which easily undergoes dynamic ligand
exchange and alters its structure and function to promote
different reactions. The Sm-BINOL-Ph3AsdO complex, a
catalyst for highly enantioselective epoxidation of R,â-unsatur-
ated amides, is transformed by the subsequent addition of Me3-
SiN3 into a highly reactive azidation reagent for regioselective
ring opening of R,â-epoxy amide. Thus, a mild and efficient
one-pot sequential catalytic asymmetric epoxidation-regio-
selective epoxide-opening process was developed. In situ IR
spectroscopy and other mechanistic studies provide strong
evidence for the generation of the samarium azide complex,
which should work as the active azidation reagent. This is the
first investigation of the physical properties of the lanthanide
azide complex. In addition, we report that the high Lewis
basicity of the amide moiety has a key role in the unexpectedly
high reactivity of both the epoxidation and the epoxide-opening
reactions. Variants of this one-pot sequential process utilizing
other nucleophiles such as sulfur and carbon nucleophiles were
also examined. Finally, anti-â-azido-R-hydroxyamides were
transformed to biologically active compounds.
Results and Discussion
(A) Catalytic Ring-Opening Reaction of r,â-Epoxy Amides
with Azide. We previously reported a general and highly
enantioselective epoxidation of R,â-unsaturated simple amides
2 catalyzed by the Sm-(S)-BINOL-Ph3AsdO complex 1,
prepared from Sm(O-i-Pr)3, (S)-BINOL, and Ph3AsdO in a ratio
of 1:1:1 (Scheme 2).9 The corresponding R,â-epoxy amides 3
were obtained in excellent yield (up to 99%) with excellent
(7) For selected recent examples, see: (a) Evans, P. A.; Robinson, J. E. J.
Am. Chem. Soc. 2001, 123, 4609. (b) Louie, J.; Bielawski, C. W.; Grubbs,
R. H. J. Am. Chem. Soc. 2001, 123, 11312. (c) Morimoto, T.; Fuji, K.;
Tsutsumi, K.; Kakuuchi, K. J. Am. Chem. Soc. 2002, 124, 3806. (d) Sutton,
A. E.; Seigal, B. A.; Finnegan, D. F.; Snapper, M. L. J. Am. Chem. Soc.
2002, 124, 13390. (e) Ajamian, A.; Gleason, J. L. Angew. Chem., Int. Ed.
2004, 43, 3754 and references therein.
(8) For recent examples reported by our group, see: (a) Sakurada, I.; Yamasaki,
S.; Go¨ttlich, R.; Kanai, M.; Shibasaki, M. J. Am. Chem. Soc. 2000, 122,
1245. (b) Sakurada, I.; Yamasaki, S.; Kanai, M.; Shibasaki, M. Tetrahedron
Lett. 2000, 41, 2415. (c) Yamasaki, S.; Kanai, M.; Shibasaki, M. J. Am.
Chem. Soc. 2001, 123, 1256. (d) Tian, J.; Yamagiwa, N.; Matsunaga, S.;
Shibasaki, M. Angew. Chem., Int. Ed. 2002, 41, 3636.
(9) (a) Nemoto, T.; Kakei, H.; Gnanadesikan, V.; Tosaki, S.-y.; Ohshima, T.;
Shibasaki, M. J. Am. Chem. Soc. 2002, 124, 14544. (b) Tosaki, S.-y.;
Nemoto, T.; Ohshima, T.; Shibasaki, M. Org. Lett. 2003, 5, 495.
(c) Ohshima, T.; Nemoto, T.; Tosaki, S.-y.; Kakei, H.; Gnanadesikan, V.;
Shibasaki, M. Tetrahedron 2003, 59, 10485. (d) Kakei, H.; Nemoto, T.;
Ohshima, T.; Shibasaki, M. Angew. Chem., Int. Ed. 2004, 43, 317.
(e) Tosaki, S.-y.; Horiuchi, Y.; Nemoto, T.; Ohshima, T.; Shibasaki, M.
Chem.-Eur. J. 2004, 10, 1527.
(10) Although H8-BINOL can be used as a ligand for the catalytic asymmetric
epoxidation of R,â-unsaturated N-acylpyrroles (refs 6f,g), it was not
effective for R,â-unsaturated simple amides. 3a was obtained from 2a in
10% yield after 48 h when 10 mol % of the Sm-(S)-H8-BINOL-Ph3Asd
O (1:1:1) complex was used.
9
2148 J. AM. CHEM. SOC. VOL. 127, NO. 7, 2005

Ligand Exchange of the Lanthanide Complex
A R T I C L E S
Scheme 2. Catalytic Asymmetric Epoxidation of R,â-Unsaturated
Simple Amides Promoted by the Sm-(S)-BINOL-Ph3AsdO
Complex 1
R-hydroxyamide 4a in 99% yield with complete regioselectivity
(entry 1). Reduced catalyst loading (0.2 mol %) was sufficient
to complete the reaction, affording 4a in 97% yield after 2 h
(entry 2). The use of 5 mol % of the (S)-Sm complex 1 also
completed the reaction with comparable reactivity (entry 3). In
contrast, the reaction with 10 mol % Sm(OTf)3, a much stronger
Lewis acid, proceeded sluggishly to give 4a in 21% yield after
24 h (entry 4).14 Only a trace amount of 4a was obtained in the
absence of the catalyst, even after 48 h.
Table 1. Regioselective Ring Opening of R,â-Epoxy Amide with
Me3SiN3 Using Various Sm Catalysts
(B) One-Pot Sequential Catalytic Asymmetric Epoxi-
dation-Regioselective Epoxide-Opening Process. Encouraged
by the above results, we examined the extension to a one-pot
sequential catalytic asymmetric epoxidation-regioselective ep-
oxide-opening process using the (S)-Sm complex 1. We
hypothesized that both the asymmetric epoxidation and the
following nucleophilic epoxide-opening reaction would be
catalyzed by the Sm complex 1 through a dynamic ligand
exchange. After completion of the catalytic asymmetric epoxi-
dation of R,â-unsaturated amide 2a, 2 equiv of Me3SiN3 were
directly added to the reaction mixture. The epoxide opening
proceeded smoothly at room temperature without significant
adverse effects, and subsequent desilylation afforded 4a in 99%
overall yield with 99% ee (Table 2, entry 1). Table 2 summarizes
the scope and limitations of the sequential process. The present
one-pot sequential process was applicable to various R,â-
unsaturated amides 2 with broad substrate generality, and the
enantioselectivity was generally excellent (96->99%). In the
presence of 2 to 10 mol % of (S)-Sm complex 1, tertiary amides
with various â-aryl substituents (2a-h, 2j, entries 1-10, 12)
underwent the sequential process at room temperature, affording
the corresponding anti-â-azido-R-hydroxyamides 4 in good to
excellent yield (70-99%). In these cases, the epoxide-opening
reaction was completed within 1-3 h, and the overall yield
mainly reflected the conversion of the epoxidation reaction.
Despite the high instability of the corresponding epoxide of 2i
(entry 11), the present sequential process successfully provided
4i, representing an advantageous example. The fact that even
such a challenging substrate gave 45% chemical yield was very
encouraging.15 The asymmetric epoxidation and the epoxide-
opening reaction of R,â,γ,δ-unsaturated amide 2k occurred
regioselectively to afford 4k (entry 13). Both the epoxidation
and epoxide opening of the secondary amide 2l proceeded
slowly compared with that of tertiary amide 2a, giving 4l in
83% overall yield (entry 14). R,â-Unsaturated amides with
â-aliphatic substituents (2m-r, entries 15-20) were also
applicable to the present sequential process. Even in these cases,
the epoxide-opening reactions proceeded smoothly at room
temperature, and it was noteworthy that the products were
obtained with complete regioselectivity. Cyclic substrates (2m,
2n) gave the corresponding tert-alcohols, thereby constructing
stereogenic tetrasubstituted carbon centers (entries 15 and 16).
Acyclic substrates 2o-q afforded 4o-q as the sole detectable
products in good yield (84-92%; entries 17-19). γ-Branched
entry
catalyst
x (mol %)
time (h)
yielda (%)
1b
2b
3b
Sm(O-i-Pr)3
Sm(O-i-Pr)3
Sm-(S)-BINOL-Ph3AsdO
(1:1:1) complex 1
Sm(OTf)3
5
0.2
5
1
2
1
99
97
99
4c
10
24
21
a Isolated yield. b Desilylation was conducted with 1 N HCl aq-MeOH.
c Desilylation was conducted with KF in MeOH.
enantiomeric excess (up to >99%).10 The highly enantioenriched
R,â-epoxy amides, in particular R,â-epoxy morpholinyl amides,9b,e
are versatile intermediates because reductive or nucleophilic
epoxide opening and a modification of the amide moiety provide
efficient access to useful chiral building blocks. There are few
reports on the efficient synthesis of optically active R,â-epoxy
amides11 other than ours, indicating that the ring-opening
reaction of R,â-epoxy amides has not been widely investigated,
despite its utility. The regioselective epoxide-opening reaction
with various nucleophiles is one of the most important trans-
formations in organic synthesis because the epoxide-opened
product is obtained with complete control of the vicinal
stereochemistry. Chong and Sharpless reported a ring-opening
reaction of R,â-epoxy amides with PhSH promoted by 150 mol
% of Ti(O-i-Pr)4.12 Secondary amides showed preference for
ring opening at the â-position, whereas tertiary amides at the
R-position. Aggarwal et al. reported a highly â-selective ring-
opening reaction of tertiary R,â-epoxy amides with either PhSH
or Me3SiN3 promoted by Yb(OTf)3.11b Yamamoto et al. reported
a regioselective epoxide-opening reaction with Me3SiN3 using
catalytic Yb(O-i-Pr)3.13 Although R,â-epoxy carbonyl com-
pounds were not utilized in the epoxide-opening reactions
reported by Yamamoto et al.13 and Utimoto et al.,2 we assumed
that Sm(O-i-Pr)3 would promote the regioselective ring opening
of R,â-epoxy amides to provide â-substituted R-hydroxyamides.
The ring-opening reaction of R,â-epoxy amide with a
lanthanide catalyst was examined prior to the one-pot sequential
reaction. Table 1 summarizes the results of epoxide-opening
reactions with Me3SiN3 in the presence of samarium catalysts.
As we expected, treatment of R,â-epoxy amide 3a with 2 equiv
of Me3SiN3 in the presence of 5 mol % Sm(O-i-Pr)3 led to a
clean epoxide opening within 1 h at room temperature.
Subsequent desilylation gave the corresponding anti-â-azido-
(14) For regioselective epoxide-opening reactions with azide using other Lewis
acid catalysts, see: (a) Denis, J.-N.; Green, A. E.; Serra, A. A.; Luche,
M.-J. J. Org. Chem. 1986, 51, 46. (b) Azzena, F.; Crotti, P.; Favero, L.;
Pineschi, M. Tetrahedron 1995, 48, 13409. (c) Francesco, F.; Pizzo, F.;
Vaccaro, L. Tetrahedron Lett. 2001, 42, 1131. (d) Francesco, F.; Pizzo,
F.; Vaccaro, L. J. Org. Chem. 2001, 66, 3554.
(11) (a) Zhou, Y.-G.; Hou, X.-L.; Dai, L.-X.; Xia, L.-J.; Tang, M.-H. J. Chem.
Soc., Perkin Trans. 1 1999, 77. (b) Aggarwal, V. K.; Hynd, G.; Picoul,
W.; Vasse, J.-L. J. Am. Chem. Soc. 2002, 124, 9964.
(12) Chong, J. M.; Sharpless, K. B. J. Org. Chem. 1985, 50, 1560.
(13) Meguro, M.; Asao, N.; Yamamoto, Y. Chem. Commun. 1995, 1021. The
reaction mechanism was not described in this article.
(15) The syn isomer was contaminated with the anti isomer (anti:syn ) 11:1).
The moderate overall yield was due to the partial decomposition of the
corresponding R,â-epoxy amide during the epoxidation reaction.
9
J. AM. CHEM. SOC. VOL. 127, NO. 7, 2005 2149

A R T I C L E S
Tosaki et al.
Table 2. One-Pot Sequential Catalytic Asymmetric Epoxidation-Regioselective Epoxide-Opening Process with Various R,â-Unsaturated
Amides
substrate
catalyst
time
yieldb
(%)
eec
(%)
entry
R1
R2
NR3R4
(x mol %)
(y/z h)
product
1
2
3
4
5
6
7
8
C6H5
C6H5
C6H5
4-MeOC6H4
4-MeOC6H4
4-MeC6H4
4-FC6H4
1-naphthyl
2-naphthyl
2-naphthyl
2-furyl
3-furyl
(E)-PhCHdCH-
C6H5
-(CH2)3-
-(CH2)4-
Ph(CH2)2-
Ph(CH2)2-
n-propyl
H
H
H
H
H
H
H
H
H
H
H
H
H
H
NMe2
NMe2
morpholinyl
NMe2
morpholinyl
NMe2
NMe2
NMe2
NMe2
NMe2
NMe2
NMe2
NMe2
NHMe
NHBn
NHBn
NMe2
morpholinyl
NMe2
2a
2a
2b
2c
2d
2e
2f
2g
2h
2h
2i
2j
2k
2l
2m
2n
2o
2p
2q
2r
5
2
5
5
5
5
5
10
5
2
12/1
15/2
11/1
13/1
12/1
13/1
11/1
11/3
13/1
16/1
11/0.5
11/1
12/0.5
13/8
12/5
13/1
6/12
5/12
6/12
6/16
99
70
99
95
97
93
98
98
99
71
45f
94
90
83
97
86
84
92
85
75
99
99
4a
4a
4b
4c
4d
4e
4f
4g
4h
4h
4i
4j
4k
4l
4m
4n
4o
4p
4q
4r
99d
>99
99d
99
>99
98
9
>99
98
>99
>99
>99
99
96
99
98
10
11e
12
13
14
15
16
17
18
19
20
10
5
10g
10
10
10
5
H
H
H
H
5
5
10
98
98h
99h
cyclohexyl
NMe2
a MS 4A was used without prior activation (1000 mg/mmol of starting material). b Isolated yield. The regioselectivity was generally below the detection
limit of 500 MHz 1H NMR (>98:2). c Determined by chiral HPLC analysis. d ee was determined after conversion to the corresponding N-Boc amine. e The
corresponding epoxide is decomposed on silica gel. f Isolated yield of the major anti isomer after conversion to TES ether (ref 15). g Gd was used as the
central metal instead of Sm (ref 9d). h ee was determined after conversion to the corresponding benzoate.
substrate 2r exhibited lower reactivity, probably because of
steric hindrance (75% yield; entry 20). While tertiary amides
2o and 2p were successful (entries 17 and 18), the corresponding
N-methylamide 2s (R1 ) Ph(CH2)2-, R2 ) H) did not undergo
the epoxide-opening reaction under the same reaction conditions.
a samarium azide complex was generated in the mixture. We
also performed electrospray ionization mass spectrometry (ESI-
MS) analysis of the mixture of Sm(O-i-Pr)3 and Me3SiN3, but
no clear spectrum that supported the generation of samarium
azide species was obtained, probably because of its oligomeric
nature. When Me3SiN3 was treated with (S)-Sm complex 1
instead of Sm(O-i-Pr)3, on the other hand, two peaks assigned
to [Sm(N3)2(Ph3AsdO)3]+ (MW 1201) and [Sm(N3)2(Ph3Asd
O)4]+ (MW 1524) were observed on the spectrum, indicating
the generation of samarium azide species.17
We also measured the IR spectra of the sequential process to
examine whether the same or related species were generated.
After the catalytic asymmetric epoxidation of 2a reached
completion in the presence of 10 mol % of (S)-Sm complex 1,
Me3SiN3 (20 mol equiv to catalyst) was added to the reaction
mixture, and the IR scan was started. There was a new broad
peak around 2072 cm-1 (Figure 1B). DFT calculation (B3LYP/
LanL2DZ level) indicated that Sm(N3)2(OMe) and SmN3(OMe)2
have absorptions at 2082 and 2083 cm-1, respectively. These
calculations suggested that that samarium azide complex
containing alkoxide ligands has an absorption at a lower
frequency than Sm(N3)3 on the IR spectrum. In fact, there should
be a lot of TBHP and/or t-BuOH present in the reaction mixture
of the sequential process. Thus, we conclude that the broad peak
at 2072 cm-1 corresponds to the samarium azide complex in
which the samarium metal is bound or coordinated by other
(C) Mechanistic Study. Although azide complexes of early
transition metals are reported to be active nucleophiles toward
epoxides,16 the lanthanide metal azide complex has never been
characterized. To gain insight into the structure of the active
species in the epoxide-opening reaction promoted by the Sm
catalyst, we performed several spectroscopic experiments. First,
we performed NMR analysis. When Sm(O-i-Pr)3 (1 mol equiv)
and Me3SiN3 (5 mol equiv) were mixed in THF-d8, the
1
generation of Me3SiO-i-Pr was observed on H and 13C NMR
spectra, suggesting that a ligand exchange occurs from the
isopropoxide to the azide on the samarium metal. Next, we
measured in situ IR spectra to gain more precise information.
When Me3SiN3 (20 mol equiv) was treated with Sm(O-i-Pr)3
(1 mol equiv) in THF, a new peak appeared around 2100 cm-1
on in situ IR spectra (Figure 1A). Treatment of Me3SiN3 with
the (S)-Sm complex 1 instead of Sm(O-i-Pr)3 gave similar
spectra. Density functional theory (DFT) calculation (B3LYP/
LanL2DZ level) indicated that Sm(N3)3 has an absorption at
2093 cm-1 (NdN stretch). These data strongly suggested that
(16) (a) Choukroun, R.; Gervais, D. J. Chem. Soc., Dalton Trans. 1980, 1800.
(b) Blandy, C.; Choukroun, R.; Gervais, D. Tetrahedron Lett. 1983, 24,
4189. (c) Caron, M.; Carlier, P. R.; Sharpless, K. B. J. Org. Chem. 1988,
53, 5185. (d) Nugent, W. A. J. Am. Chem. Soc. 1992, 114, 2768.
(e) McCleland, B. W.; Nugent, W. A.; Finn, M. G. J. Org. Chem. 1998,
63, 6656. (f) Mart´ınez, L. E.; Leighton, J. L.; Carsten, D. H.; Jacobsen, E.
N. J. Am. Chem. Soc. 1995, 117, 5897. (g) Hansen, K. B.; Leighton, J. L.;
Jacobsen, E. N. J. Am. Chem. Soc. 1996, 118, 10924.
(17) The concentration of the monomeric species corresponding to [Sm(N3)2(Ph3-
AsdO)3]+ and [Sm(N3)2(Ph3AsdO)4]+ in the mixture would be very low
because the treatment of Me3SiN3 with either Sm(O-i-Pr)3 or (S)-Sm
complex 1 resulted in similar IR spectra. For details, see the Supporting
Information.
9
2150 J. AM. CHEM. SOC. VOL. 127, NO. 7, 2005

Ligand Exchange of the Lanthanide Complex
A R T I C L E S
Figure 1. In situ IR experiments. (A) Me3SiN3 (20 mol equiv) + Sm(O-i-Pr)3 (1 mol equiv). (B) IR spectra in the one-pot sequential process with (S)-Sm
complex 1.
ligands such as alkoxide.18,19 On the other hand, there were no
peaks observed, other than the ones corresponding to Me3SiN3
(ν ) 2140 cm-1, NdN stretch), when treated with Sm(OTf)3
in THF. As mentioned above, the epoxide-opening reaction with
Sm(O-i-Pr)3 was promoted more smoothly than that with Sm-
(OTf)3 (Table 1). These results suggest that Sm(O-i-Pr)3 or the
(S)-Sm complex 1 works in the epoxide-opening reaction, not
simply as a Lewis acid, but as the active azidation reagent by
formation of the highly reactive samarium azide complex. It is
noteworthy that the ligand exchange to form the samarium azide
complex occurred predominantly even in the presence of other
components such as the substrate, TBHP, and/or t-BuOH,
efficiently promoting the epoxide-opening reaction.
In stark contrast to the success using R,â-epoxy amides that
exhibited high reactivity and complete regioselectivity, R,â-
epoxy ketones and R,â-epoxy esters remained almost unchanged
with Me3SiN3 at room temperature in the presence of 10 mol
% of Sm(O-i-Pr)3.20 These results suggest that the Lewis basicity
of the carbonyl moiety has a key role in the epoxide-opening
reaction promoted by the samarium catalyst. In fact, the catalytic
epoxide-opening reaction of tertiary amides proceeded more
smoothly than that of the secondary amides (Table 2, 2a vs 2l
and 2o vs 2s). To further investigate the origin of the dramatic
difference in reactivity, we performed the ring-opening reaction
of various R,â-epoxy anilides 7a-e, where the Lewis basicity
of the carbonyl moiety was tuned by a para-substituent (X) on
the benzene ring (Scheme 3, step B). As shown in the Hammett
plot in Scheme 3,21 the initial rate (v) of step B increased as
the electron-donating ability of X increased. These results
suggested that more Lewis-basic amide carbonyl coordinates
to the samarium more efficiently and enhances the nucleo-
philicity of the active samarium azide complex, resulting in the
high reactivity of R,â-epoxy amides. These findings prompted
us to examine the effects of Lewis basicity of the amide moiety
in the asymmetric epoxidation. We previously reported that
enhancement of the reactivity appeared to be related to the
decrease in the energy level of the lowest unoccupied molecular
orbital (LUMO) in catalytic asymmetric epoxidation of R,â-
unsaturated carboxylic acid imidazolides. N-Acyl-4-phenyl-
(18) BINOL would not be involved in the samarium azide complex because all
of the BINOL was silylated in the reaction mixture.
(19) The broadened peak at 2072 cm-1 suggested that the samarium azide
complex exists as a mixture of several oligomeric species in equilibrium.
(20) R,â-Epoxy chalcone: no reaction. R,â-Epoxy methyl cinnamate: trace.
(21) For Hammett substituent constants, see: Smith, M. B.; March, J. March’s
AdVanced Organic Chemistry: Reactions, Mechanisms, and Structure, 5th
ed.; John Wiley and Sons: New York, 2002.
9
J. AM. CHEM. SOC. VOL. 127, NO. 7, 2005 2151

A R T I C L E S
Tosaki et al.
Scheme 5 . Proposed Catalytic Cycle of the Epoxide-Opening
Scheme 3 . Effect of Amide Moiety on the Reactivity and the
Hammett Plota
Reaction with Azide
a (a) (S)-Sm complex 1 (10 mol %), TBHP in decane (1.2 equiv), THF,
MS 4A, 25 °C. (b) Sm(O-i-Pr)3 (10 mol %), Me3SiN3 (2 equiv), THF,
25 °C.
Table 3. Molecular Orbital Calculations (B3LYP/6-31G(d) Level)
of R,â-Unsaturated Ester and R,â-Unsaturated Amides
substrate
σp
LUMO energy level (eV)
reactivity would be effective enough to overwhelm the poten-
tially low reactivity that derives from the high LUMO energy
level.22-24
N-methylcinnamamide (2l)
methyl cinnamate
-
-
-0.28
-0.14
0
-1.46
-1.72
-1.64
-1.68
-1.72
-1.78
-1.96
6a
6b
6c
6d
6e
We also performed initial rate kinetic studies of the ring-
opening reaction of R,â-epoxy amide 3l with Me3SiN3 using
catalytic Sm(O-i-Pr)3. The kinetic studies revealed that the
reaction rate had a 1.08 order dependency on the concentration
of Sm(O-i-Pr)3 and a 0.04 order dependency on the concentra-
tion of Me3SiN3 (Scheme 4). On the basis of these mechanistic
studies, the proposed catalytic cycle of the epoxide-opening
reaction with Me3SiN3 is illustrated in Scheme 5. Sm(O-i-Pr)3
or the (S)-Sm complex 1 undergoes ligand exchange by the
addition of Me3SiN3 to generate a highly reactive samarium
azide complex (I). Although the precise structure of the
samarium azide complex is unclear, it might include at least
two samarium metals, and the result of a first-order kinetic
dependency on the catalyst suggests that the epoxide opening
occurs via cooperative reactivity of those metals.25 R,â-Epoxy
amide 3 coordinates to one samarium metal (Sm1) of the
complex in a bidentate manner (II), by which the direction of
0.06
0.53
Scheme 4 . Initial Rate Kinetics of Epoxide-Opening Reaction of 3l
with Azide
imidazolide, which has the lowest LUMO energy level among
the other N-acylimidazolides, gave the best results in terms of
reactivity and enantioselectivity.6e,9c Catalytic asymmetric epox-
idation of R,â-unsaturated simple amides, however, tended to
have relatively high reactivity in contrast to much lower reactive
R,â-unsaturated esters, opposite to the expectation based on the
LUMO energy calculations (Table 3; B3LYP/6-31G(d) level,
N-methylcinnamamide: -1.46 eV, methyl cinnamate:
-1.72 eV).9a,c Thus, we measured the initial rates of the
asymmetric epoxidation of various R,â-unsaturated anilides
6a-e using 10 mol % of (S)-Sm complex 1 (Scheme 4, step
A). As a result, the initial rate (v) of step A increased as the
electron-donating ability of X increased, similar to step B. This
result indicates that the unexpectedly high reactivity of R,â-
unsaturated simple amides for the catalytic asymmetric epoxi-
dation is due to the relatively high Lewis basicity of the amide
carbonyl moiety, which would enhance the nucleophilicity of
the lanthanide peroxide. In addition, the enhancement of the
(22) The catalytic asymmetric epoxidation of R,â-unsaturated N-acylimidazolides
or R,â-unsaturated ketones would proceed easily because of their low
enough LUMO energy level and thus would not require the Lewis base
enhancements.
(23) The initial rate of step A should be increased as the LUMO energy level
decreased (6a f 6e) if the effects of Lewis basicity of carbonyl moiety
were ignored. As observed in the Hammett plot of Scheme 3, the range of
the initial rate of step A was smaller than that of step B, where only the
Lewis basicity defines the reactivity.
(24) For other examples of the superiority of R,â-unsaturated amides to R,â-
unsaturated esters in terms of reactivity, see: (a) Concello´n, Jose´ M.;
Rodr´ıguez-Solla, H.; Go´mez, C. Angew. Chem., Int. Ed. 2002, 41, 1917.
(b) Kamimura, A.; Murakami, N.; Kawahara, F.; Yokota, K.; Omata, Y.;
Matsuura, K.; Oishi, Y.; Morita, R.; Mitsudera, H.; Suzukawa, H.; Kakehi,
A.; Shirai, M.; Okamoto, H. Tetrahedron 2003, 59, 9537.
(25) Azide groups can act as bridging ligands between metal atoms leading to
di- and polynuclear complexes. For selected examples, see: (a) Fehlhammer,
W. P.; Dahl, L. F. J. Am. Chem. Soc. 1972, 94, 3377. (b) DeMunno, G.;
Poerio, T.; Biau, G.; Julve, M.; Lloret, F. Angew. Chem., Int. Ed. Engl.
1997, 36, 1459. (c) Goher, M. A. S.; Al-Salem, N. A.; Mautner, F. A.;
Klepp, K. O. Polyhedron 1997, 16, 825. (d) Clegg, W.; Krischner, H.;
Saracoglu, A. I.; Sheldrick, G. M. Z. Kristallogr. 1982, 161, 307. On the
basis of the large coordination numbers of lanthanide metals, the samarium
azide complex might contain an azido-bridging structure to form the
oligomer.
9
2152 J. AM. CHEM. SOC. VOL. 127, NO. 7, 2005

Ligand Exchange of the Lanthanide Complex
A R T I C L E S
Table 4. Regioselective Epoxide-Opening Reaction by the Sm
Catalyst with PhSSiMe3
(C-â:C-R ) 94:6) (entry 2). Morpholinyl amide 2b was also
applicable (entries 3 and 4), and the electron-donating group
on the aromatic ring increased the regioselectivity (entries 5
and 6). Secondary amide 2l gave the epoxide-opened product
with higher regioselectivity (C-â:C-R > 98:2) than did tertiary
amides. For â-aliphatic substrates, secondary amides gave
successful results in terms of regioselectivity.27 The asymmetric
epoxidation of 2s and the subsequent epoxide opening with
either Me3SiSPh or PhSH proceeded smoothly, giving the
corresponding product 9s with high selectivity (92:8 and 90:
10, respectively; entries 9 and 10). γ-Branched substrate 2t
required a longer reaction time for the epoxide opening
(>72 h), and 9t was obtained with modest regioselectivity
(C-â:C-R ) 85:15, 80:20) (entries 11 and 12). In general,
oxidation of the sulfide by the remaining TBHP (ca. 0.2 equiv)
was not problematic in the present sequential process.28
entry
catalyst
conversiona (%)
ratio (C-â:C-R)
1b
2b
Sm(O-i-Pr)3
100
100
95:5
92:8
Sm-(S)-BINOL-Ph3AsdO
(1:1:1) complex 1
Sm(OTf)3
3c
9
>98:2
1
a Determined by H NMR analysis. b Desilylation was conducted with
1 N HCl aq-MeOH. c Desilylation was conducted with KF in MeOH.
the nucleophile entry is controlled and preferential attack at the
â-position is achieved. The delivery of the bound azide to the
epoxide on a single metal center to afford the epoxide-opened
product with anti stereochemistry is geometrically impossible.16e
Thus, the azide ligand, which is bound to another samarium
metal (Sm2) and activated by coordination of another R,â-epoxy
amide, would be transferred to the epoxide intramolecularly to
give samarium alkoxide(III). The subsequent ligand exchange
with Me3SiN3 regenerates the samarium azide complex (I)
together with the production of silyl ether 5. This step would
proceed very fast so that the rate-determining step of the epoxide
opening should be in a part of (I) to (III) (presumably (II) to
(III)).
On the basis of the postulated catalytic cycle shown in
Scheme 5, a similar reaction mechanism might occur in the
epoxide-opening reaction with PhSSiMe3 or PhSH. The forma-
tion of monosilylated BINOL was detected by ESI-MS analysis
of the mixture of the (S)-Sm complex 1 and PhSSiMe3. This
result, as well as the lower catalyst activity of Sm(OTf)3
compared with that of Sm(O-i-Pr)3 (Table 4, entry 3), suggests
that samarium thiolate (Sm-SPh) is generated through dynamic
ligand exchange and acts as the highly nucleophilic reagent.
We also examined the regioselective ring-opening reaction
of R,â-epoxy amides with Me3SiCN using catalytic Sm(O-i-
Pr)3. On the basis of the reports by Utimoto et al.,2 we expected
that samarium cyanide would be generated and work effectively
as an active cyanating reagent. The results are shown in Table
6. The epoxide-opening reaction of 3a with Me3SiCN in the
presence of 10 mol % of Sm(O-i-Pr)3 proceeded smoothly at
room temperature, giving the corresponding â-cyano-R-tri-
methylsilyloxyamide 10a in 81% yield with good regioselect-
ivity (C-â:C-R ) 88:12) after 24 h (entry 1). The reaction of
secondary amide 3l also proceeded smoothly, and the product
10l was obtained in 80% yield with higher regioselectivity
(C-â:C-R ) 93:7; entry 2) than that obtained with tertiary amide
3a. R,â-Epoxy amide with â-alky substituent (3o) gave the
product 10o in modest yield (57%) with good regioselectivity
(C-â:C-R ) 86:14) at room temperature (entry 3). Warming
the reaction to 50 °C improved the reactivity to afford 10o in
71% yield, without a decrease in regioselectivity (C-â:C-R )
85:15) (entry 4).
(D) Application to Other Nucleophiles. On the basis of the
findings of the regioselective ring-opening reaction of R,â-epoxy
amides with Me3SiN3 using catalytic Sm(O-i-Pr)3, we explored
variant reactions using other nucleophiles, such as sulfur and
carbon nucleophiles.
Table 4 summarizes the results of catalytic epoxide opening
with PhSSiMe3. The epoxide opening of 3a with 2 equiv of
PhSSiMe3 in the presence of 10 mol % of Sm(O-i-Pr)3 resulted
in the clean consumption of 3a within 1 h at room temperature.
After desilylation, the corresponding â-phenylthio-R-hydro-
xyamide 9a was obtained with high regioselectivity (C-â:C-R
) 95:5, entry 1). The (S)-Sm complex 1 gave the product with
similar reactivity and slightly lower regioselectivity (C-â:C-R
) 92:8, entry 2). When 10 mol % of Sm(OTf)3 was used instead
of Sm(O-i-Pr)3 (entry 3), the reaction proceeded with a much
lower reaction rate, giving 9a in 9% conversion after 1 h.26 Only
a trace amount of 9a was obtained in the absence of the catalyst,
even after 48 h. Similar to the epoxide-opening reaction with
Me3SiN3, R,â-epoxy ketones and R,â-epoxy esters were much
less reactive substrates.
A one-pot sequential process with a sulfur nucleophile was
then examined. The results are summarized in Table 5. After
completion of the catalytic asymmetric epoxidation of 2a with
10 mol % of (S)-Sm complex 1, the addition of 2 equiv of
PhSSiMe3 (condition A) directly to the reaction mixture resulted
in smooth epoxide opening within 1 h. Subsequent desilylation
gave the epoxide-opened product in 86% overall yield with 92:8
regioselectivity (entry 1). Alternatively, the addition of 3 equiv
of PhSH (condition B) was also effective, affording the product
with comparable reactivity (93% yield) and regioselectivity
Extension to the one-pot sequential process was also exam-
ined. After completion of the catalytic asymmetric epoxidation
of R,â-unsaturated amide, the addition of Me3SiCN to the
reaction mixture resulted in the sluggish epoxide opening, even
at 50 °C. Improvement of the reactivity to achieve the sequential
process is currently in progress.
(E) Transformations to Biologically Active Compounds.
Regioselective ring opening of the chiral R,â-epoxy amides with
azide provides efficient access to chiral â-amino alcohols, which
(27) Tertiary amides gave the epoxide-opened products with low regioselect-
ivity (ca. 1:1) although the reaction proceeded smoothly under the same
reaction conditions. For the difference of the regioselectivity between
secondary and tertiary amides, see ref 12.
(28) Less than 10% of the oxidized product was formed. When sulfide 9a was
subjected to conditions for the catalytic asymmetric epoxidation [(S)-Sm
complex 1 (10 mol %), TBHP (1.2 equiv), THF, room temperature], a
complex mixture of oxidized products was obtained.
(26) 9a was obtained in 75% yield after 48 h.
9
J. AM. CHEM. SOC. VOL. 127, NO. 7, 2005 2153

A R T I C L E S
Tosaki et al.
.
Table 5. One-Pot Sequential Process with a Sulfur Nucleophile
substrate
time
yieldb
(%)
ratioc
(C- :C-
eed
(%)
entry
R1
NR2R3
condition
x/y (h)
â
R
)
product
1
2
3
4
5
6
7
8
C6H5
C6H5
C6H5
C6H5
4-MeOC6H4
4-MeOC6H4
C6H5
NMe2
NMe2
morpholinyl
morpholinyl
NMe2
NMe2
NHMe
NHMe
NHMe
2a
2a
2b
2b
2c
2c
2l
A
B
A
B
A
B
A
B
A
B
A
B
11/1
11/1
11/1.5
11/1.5
11/1
11/1
15/2
13/4
11/2
6/2
86
93
85
91
83
90
70
74
76
72
75
74
92:8
94:6
95:5
96:4
96:4
99
99
99
9a
9a
9b
9b
9c
9c
9l
99
>99
>99
99
99
99
99
99
99
98:2
>98:2
>98:2
92:8
90:10
85:15
80:20
C6H5
2l
9l
9
Ph(CH2)2-
Ph(CH2)2-
cyclohexyl
cyclohexyl
2s
2s
2t
9s
9s
9t
10
11
12
NHMe
NHMe
NHMe
15/73
14/76
2t
9t
1
a MS 4A was used without prior activation (1000 mg/mmol of starting material). b Isolated yield. c Determined by H NMR analysis. d Determined by
chiral HPLC analysis.
Table 6. Regioselective Epoxide-Opening Reaction by the Sm
Catalyst with Me3SiCN
Scheme 6. Synthesis of Taxol Side Chain and (-)-Cytoxazonea
substrate
temp
C)
time
(h)
yielda
(%)
ratiob
(C- :C-
entry
R1
NR2R3
(
°
â
R)
product
1
2
3
4
C6H5
C6H5
Ph(CH2)2- NMe2
Ph(CH2)2- NMe2
NMe2
NHMe 3l
3a rt
24
39
36
36
81
80
57
71
88:12
93:7
86:14
85:15
10a
10l
10o
10o
rt
3o rt
3o 50
1
a Isolated yield. b Determined by H NMR analysis.
a Conditions: (a) BzCl, Et3N, cat. DMAP, CH2Cl2, room temperature,
98%; (b) cat. Pd-C, H2 (1 atm), EtOAc, room temperature, 96%; (c) SOCl2,
CHCl3, room temperature; (d) 1 N NaOH aq, EtOH, 60 °C, 70% (two steps);
(e) Me3SiCHN2, Et2O-MeOH, room temperature; (f) 1 N HCl aq, MeOH,
reflux, 89% (two steps); (g) cat. Pd-C, Boc2O, H2 (1 atm), EtOAc, room
temperature, 98% from 4c, 99% from 4d; (h) LiAlH4, THF, -40 to 0 °C;
(i) NaBH4, MeOH-THF, room temperature; (j) NaH, THF, room temper-
ature, 58% (three steps); (k) n-BuLi, BH3‚NH3, THF, 0 °C to room
temperature, (39% from 16c, 40% from 16d).
are ubiquitous in various biologically active compounds.29
Having established the above new process for the enantio-
selective synthesis of anti-â-azido-R-hydroxyamides, the utility
was demonstrated by short syntheses of the side chain 11 of
the anticancer drug, taxol,30 and a novel cytokine modulator,
(-)-cytoxazone (12)31 (Scheme 6). For the synthesis of 11, ent-
4a was converted to azido benzoate 13, which was hydrogenated
to produce benzamide 14 via in situ O f N benzoyl transfer.
Treatment of 14 with thionyl chloride to form an oxazoline ring,
followed by hydrolysis of the dimethylamide gave known
carboxylic acid 15.30d Methylation of 15 and the oxazoline ring
opening under acidic conditions afforded the taxol side chain
11. For the synthesis of (-)-cytoxazone (12), morpholinyl amide
4d was converted to N-Boc amine 16d, which was subjected to
stepwise reduction of the morpholinyl amide to primary alcohol.
Finally, subsequent cyclization afforded 12 in 58% yield for
three steps. Alternatively, 12 was also synthesized from 16c or
16d using n-BuLi and BH3‚NH3,32 respectively. In these cases,
the reduction of the tertiary amide to primary alcohol and the
(29) For reviews of asymmetric synthesis of vicinal amino alcohols, see:
(a) Bergmeire, S. C. Tetrahedron 2000, 56, 2561. (b) Kolb, H. C.; Sharpless,
K. B. In Transition Metals for Organic Synthesis; Beller, M., Bolm, C.,
Eds.; Wiley-VCH: Weinheim, Germany, 1998; p 243.
(30) For general review, see: (a) Kingston, D. G. I. Chem. Commun. 2001,
867. (b) Kingston, D. G. I. J. Nat. Prod. 2000, 63, 726. (c) Nicolaou, K.
C.; Dai, W.-M.; Guy, R. K. Angew. Chem., Int. Ed. Engl. 1996, 35, 451.
For synthesis of taxol from baccatin III, see: (d) Kingston, D. G. I.;
Chaudhary, A. G.; Gunatilaka, A. A. L.; Middleton, M. L. Tetrahedron
Lett. 1994, 35, 4483.
(31) Madhan, A.; Kumar, A. R.; Rao, B. V. Tetrahedron: Asymmetry 2001,
12, 2009 and references therein.
(32) Myers, A. G.; Yang, B. H.; Kopecky, D. J. Tetrahedron Lett. 1996, 37,
3623.
9
2154 J. AM. CHEM. SOC. VOL. 127, NO. 7, 2005

Ligand Exchange of the Lanthanide Complex
A R T I C L E S
following cyclization proceeded successively, affording 12 in
39% (from 16c) or 40% (from 16d) yield.
high Lewis basicity of the amide moiety. The developed
sequential process can be extended to other nucleophiles such
as thiols. In addition, the epoxide-opened products are versatile
intermediates, and anti-â-azido-R-hydroxyamides were trans-
formed into biologically active compounds.
Summary
The Sm-BINOL-Ph3AsdO complex, the catalyst for highly
enantioselective epoxidation of R,â-unsaturated amide, under-
goes dynamic ligand exchange by the subsequent addition of
Me3SiN3. The newly generated samarium azide complex works
as a highly reactive reagent for the regioselective ring opening
of R,â-epoxy amides. Thus, we developed a mild and efficient
one-pot sequential asymmetric epoxidation-regioselective ep-
oxide-opening process with various R,â-unsaturated amides,
affording anti-â-azido-R-hydroxyamides in up to 99% yield with
complete regioselectivity and up to greater than 99% ee. Strong
evidence for generation of the samarium azide complex was
obtained by in situ IR spectroscopy and other experiments. This
is the first investigation of the physical property of lanthanide
azide. The relatively high reactivity of both the asymmetric
epoxidation and the epoxide-opening reaction is ascribed to the
Acknowledgment. This work was supported by RFTF, Grant-
in-Aid for Encouragements for Young Scientists (A), and Grant-
in-Aid for Specially Promoted Research from the Japan Society
for the Promotion of Science (JSPS) and Ministry of Education,
Culture, Sports, Science and Technology (MEXT).
Supporting Information Available: Experimental procedures,
characterization of all new compounds with their copies of 1H
and 13C NMR spectra, and other detailed results and discussions
(PDF). This material is available free of charge via the Internet
at http://pubs.acs.org.
JA043770+
9
J. AM. CHEM. SOC. VOL. 127, NO. 7, 2005 2155