The development of scalemic multidentate niobium complexes as catalysts for the highly stereoselective ring opening of meso-epoxides and meso-aziridines

Article abstract of DOI:10.1021/ja0708666

The discovery and development of a new Lewis acid system based on a complex formed from niobium(V) methoxide and (R)-3,3′-bis(2-hydroxy-3- isopropylbenzyl)-1,1′-binaphthalene-2,2′-diol, a novel tetradentate BINOL derivative, is presented. The system was s

Full text of DOI:10.1021/ja0708666

Published on Web 06/13/2007
The Development of Scalemic Multidentate Niobium
Complexes as Catalysts for the Highly Stereoselective Ring
Opening of meso-Epoxides and meso-Aziridines
Kenzo Arai, Simone Lucarini, Matthew M. Salter, Kentaro Ohta,
Yasuhiro Yamashita, and Sh uj Kobayashi*
Contribution from the Graduate School of Pharmaceutical Sciences, The UniVersity of Tokyo,
The HFRE DiVision, ERATO, Japan Science Technology Agency (JST), Hongo, Bunkyo-ku,
Tokyo 113-0033, Japan
Received February 6, 2007; E-mail: skobayas@mol.f.u-tokyo.ac.jp
Abstract: The discovery and development of a new Lewis acid system based on a complex formed from
niobium(V) methoxide and (R)-3,3′-bis(2-hydroxy-3-isopropylbenzyl)-1,1′-binaphthalene-2,2′-diol, a novel
tetradentate BINOL derivative, is presented. The system was shown to be extremely effective in promoting
the desymmetrative ring opening of linear and cyclic meso-epoxides using anilines as nucelophiles, delivering
the corresponding (R,R) anti-amino alcohols in good to excellent yields (up to quantitative) and excellent
enantioselectivity (up to 96% ee). Furthermore, the catalyst system displays a remarkable sensitivity to
steric bulk at the â-carbon of the epoxide, selectively facilitating ring opening of smaller epoxides in the
presence of more sterically hindered epoxides. This property was confirmed by a series of competition
reactions using a mixture of meso-2-butene oxide and another aliphatic meso-epoxide, with the result that
the former, less encumbered epoxide reacted preferentially with up to 98% chemical selectivity. While it
was found to be most convenient to conduct the reactions with 10 mol % catalyst loading at 0.16 M, at
higher overall concentration the reaction still proceeded efficiently with as little as 0.25 mol % catalyst to
give the desired products with no significant reduction in yields or enantioselectivities. In addition, the current
catalyst system was also found to mediate the asymmetric ring opening of nonsymmetrical cis-2-alkene
oxides with anilines to give preferentially the corresponding (2R,3R)-2-amino-3-ols arising from ring opening
at the methyl terminus, in excellent yields (up to quantitative) and good to excellent regio- and
enantioselectivities (up to 18:1 and >99% ee, respectively). Intriguingly, it was discovered that the same
catalyst system also promoted the ring-opening desymmetrization of aziridines with aniline nucleophiles to
give the corresponding (S,S) vicinal diamines in good to excellent yields and enantioselectivity (up to 95%
and 84% ee [>99% ee following a single recrystallization]). Catalyst systems that promote closely related
reactions with opposite stereochemical outcomes in high selectivity such as the current niobium system
are extremely unusual. To the best of our knowledge, this report constitutes not only the first example of
the catalytic desymmetrization of both meso-epoxides and meso-aziridines but also a rare example of
such complementary stereoselectivity in a catalytic reaction.
Introduction
requirements placed on such systems are becoming ever more
stringent. For example, in addition to promoting reactions in
The development of new chiral catalysts for the promotion
of asymmetric reactions leading to the generation of enantiopure
products is one of the most important tasks in synthetic
high yields and with high enantioselectivities, new chiral
catalysts are increasingly required to possess the ability to
distinguish between substrates that are closely related structural
analogues. In this context, a nonenzymatic catalyst which is
able to recognize the difference between methyl and ethyl groups
in a substrate would be an ideal catalyst in this field.
1
chemistry. As a result, this area has attracted a great deal of
attention over nearly three decades, and the pace of development
has shown no sign of abating in recent years. The majority of
protocols disclosed to date involve the use of complexes of
homochiral ligands and transition metals; the central role played
by these reagents has led to the development of Lewis acids on
the basis of almost every amenable metal in the periodic table.
As a result of the attention devoted to the development of
modern catalyst systems by the chemical community, the
Epoxides and aziridines are extremely useful and versatile
3
functional elements in organic synthesis. In addition to their
2
(
2) For reviews of Lewis acid catalyzed asymmetric reactions see (a) Bach, T.
Angew. Chem., Int. Ed. 1994, 33, 417. (b) Nelson, S. G. Tetrahedron:
Asymmetry 1998, 9, 357. (c) Gr o¨ ger, H.; Vogl, E. M.; Shibasaki, M. Chem.
Eur. J. 1998, 4, 1137. (d) Mahrwald, R. Chem. ReV. 1999, 99. 1095. (e)
Kobayashi, S.; Manabe, K. Acc. Chem. Res. 2002, 35, 209. (f) Dilman, A.
D.; Ioffe, S. L. Chem. ReV. 2003, 103, 733. (g) Shibasaki, M.; Matsunaga,
S. J. Organomet. Chem. 2006, 691, 2089. (h) Kobayashi, S.; Ogawa, C.
Chem. Eur. J. 2006, 5954.
(
1) For reviews see (a) ComprehensiVe Organic Synthesis; Trost, B. M., Ed.;
Pergamon Press: Oxford, U.K., 1991; Vol. 2, Chapter 1. (b) Carreira, E.
M. In ComprehensiVe Asymmetric Catalysis; Jacobsen, E. N., Pfaltz, A.,
Yamamoto, H., Eds.; Springer: Heidelberg, Germany, 1999; Vol. 3, p 998.
10.1021/ja0708666 CCC: $37.00 © 2007 American Chemical Society
J. AM. CHEM. SOC. 2007, 129, 8103-8111
9
8103

A R T I C L E S
Arai et al.
anilines,11 RLi,12 and other carbon-centered nucleophiles have
been described. As for aziridines, while several protocols for
the nonstereoselective ring-opening reaction have been reported,
the catalytic enantioselective version of the reaction has not been
well explored. In particular, use of meso-aziridines as substrates
has been little investigated. Although examples of ring-opening
3e
Scheme 1. Desymmetrization of meso-Epoxides and Aziridines by
Ring Opening under the Influence of a Chiral Catalyst
1
3
14
reactions of aziridines using TMSCN, MeMgBr, and
1
5
TMSN3 have been described, to the best of our knowledge,
no methods for desymmetrisation of meso-aziridines using less
reactive nucleophiles such as anilines have been reported.
Needless to say, the introduction of such a catalyst system would
be of great synthetic interest. Furthermore, the invention of a
catalyst system which promotes not only the efficient and highly
stereoselective ring opening of meso-epoxides but also of meso-
aziridines, with aniline nucleophiles, would enhance yet further
the utility of Lewis acid catalysis.
presence in a diverse variety of natural products and resin
precursors, their inherent reactivity toward ring opening with a
wide range of nucleophiles to give products with, in the case
of nonterminal epoxides and aziridines, contiguous chiral centers
makes them very valuable synthetic intermediates. The higher
reactivity of epoxides in particular has led to their wide use in
synthesis and has been fueled by the fact that the ring-opening
reaction of meso-epoxides proceeds smoothly with a range of
nucleophiles such as with a range of mild Lewis acid catalysts
to give chiral products.4 Of greater synthetic interest is the
use of a chiral catalyst in the ring-opening reaction which leads
to the formation of chiral nonracemic products via desymme-
trization of the epoxide or aziridine (Scheme 1).
In recent years, our group has reported the use of complexes
of transition metals and 2,2′-binapthol (BINOL) derivatives as
catalysts for a variety of transformations such as enantioselective
-6
1
6
17
aldol reactions, hetero-Diels-Alder reactions, Mannich type
reactions, Strecker-type reactions, the allylation of imines,20
and [3+2] cycloaddition reactions. In this context, and as part
18
19
In the case of epoxides, several catalyst systems that mediate
the desymmetrative process using a broad array of metals and
21
22
of our ongoing interest in meso-epoxide chemistry, we recently
7
8
9
10
nucleophiles including TMSCN, thiols, TMSN3, alcohols,
reported the invention of the first highly enantioselective Lewis
acid catalyst system which shows remarkable selectivity in the
(
3) For recent monographs on the chemistry of expoxides and aziridines see
(a) Rao, A. S. ComprehensiVe Organic Synthesis; Trost, B. M., Fleming,
I., Eds.; Pergamon Press: Oxford, U.K., 1991; Vol. 7, p 357. (b) Jacobsen,
E. N.; Wu, M. H. ComprehensiVe Asymmetric Catalysis; Jacobsen, E. N.,
Pfaltz, A., Yamamoto, H., Eds.; Springer-Verlag: Heidelberg, Germany,
(11) (a) Hou, X. -L.; Wu, J.; Dai, L. -X.; Xia, L. -J.; Tang M. -H. Tetrahedron:
Assymmetry 1998, 9, 1747. (b) Sekine, A.; Ohshima, T.; Shibasaki, M.
Tetrahedron 2002, 58, 75. (c) Schneider, C.; Sreekanth, A. R.; Mai, E.
Angew. Chem., Int. Ed. 2004, 43, 5691. (d) Carr e´ e, F.; Gil, R.; Collin, J.
Tetrahedron Lett. 2004, 45, 7749. (e) Carr e´ e, F.; Gil, R.; Collin, J. Org.
Lett. 2005, 7, 1023. (f) Kureshy, R. I.; Singh, S.; Khan, N. H.; Abdi, S. H.
R.; Suresh, E.; Jasra, R. V. Eur. J. Org. Chem. 2006, 1303. (g) Kureshy,
R. I.; Singh, S.; Khan, N. H.; Abdi, S. H. R.; Agrawal, S.; Mayani, V. J.;
Jasra, R. V. Tetrahedron Lett. 2006, 47, 5277. (h) Mai, E. Schneider, C.
Chem. Eur. J. 2007, 2729.
1
1
999; Vol. III, Chapter 35. (c) McCoull, W.; Davis, F. A. Synthesis 2000,
347. (d) Sweeny, J. B. Chem. Soc. ReV. 2002, 31, 247. (e) Pineschi, M.
Eur. J. Org. Chem. 2006, 4979.
(
4) (a) Chini, M.; Crotti, P.; Favero, L.; Macchia, F.; Pineschi, M. Tetrahedron
Lett. 1994, 35, 433. (b) Meguro, M.; Asao, N.; Yamamoto, Y. J. Chem.
Soc., Chem. Commun. 1994, 2597. (c) Hou, X. -L.; Wu, J.; Dai, L. -J.;
Tang, M.-H. Tetrahedron: Asymmetry 1998, 9, 1747. (d) Sekar, G.; Singh,
V. K. J. Org. Chem. 1999, 64, 287. (e) Sagawa, S.; Abe, H.; Hase, Y.;
Inabe, T. J. Org. Chem. 1999, 64, 4962.
(12) (a) Alexakis, A.; Vranken, E.; Mangeney, P. Synlett 1998, 11165. (b)
Vranken, E.; Alexakis, A.; Mangeney, P. Eur. J. Org. Chem. 2005, 1354.
(13) Mita, T.; Fujimori, I.; Wada, R.; Wen, J.; Kanai, M.; Shibasaki, M. J. Am.
Chem. Soc. 2005, 127, 11252.
(
5) (a) Das, U.; Crousse, B.; Kesavan, V.; Bonnet-Dolpon, D.; B e´ gu e´ , J. P. J.
Org. Chem. 2000, 65, 6749. (b) Chandrasekhar, S.; Ramachandar, T.; Jaya
Prakesh, S. Synthesis 2000, 1817. (c) Rajendder Reddy, L.; Arjun Reddy,
M.; Bhanumanti, N.; Rama Rao, K. New J. Chem. 2001, 25, 221. (d)
Rajendder Reddy, L.; Arjun Reddy, M.; Bhanumanti, N.; Rama Rao, K.
Synthesis 2001, 831. (e) Raghavendra Swamy, N.; Kondaji, G.; Nagaiah,
K. Synth. Commun. 2002, 32, 2307. (f) Ollevier, T.; Lavie-Compin, G.
Tetrahedron Lett. 2002, 43, 7891. (g) Cossy, J.; Bellosta, V.; Hamoir, C.;
Desmurs, J. R. Tetrahedron Lett. 2002, 43, 7083. (h) Sekine, A.; Oshima,
T.; Shibasaki, M. Tetrahedron 2002, 58. 75. (i) Dur a´ n Pach o´ n, L.; Gamez,
P.; van Brussel, J. J. M.; Reedijk, J. Tetrahedron Lett. 2003, 44, 6025. (j)
Chakraborti, A. K.; Kondaskar, A. Tetrahedron Lett. 2003, 44, 8315. (k)
Zhao, P. Q.; Xu, L. W.; Xia, C. G. Synlett 2004, 846. (l) Ollevier, T.;
Lavie-Compin, G. Tetrahedron Lett. 2004, 45, 49. (m) Placzek, A. T.;
Donelson, J. L.; Trivedi, R.; Gibbs, R. A.; De, S. K. Tetrahedron Lett.
(14) Muller, P.; Nury, P. Org. Lett. 1999, 1, 439.
(15) (a) Li, Z.; Fernandez, M.; Jacobsen, E. N. Org. Lett. 1999, 1, 1611. (b)
Fukuta, Y.; Mita, T.; Fukuda, N.; Kanai, M.; Shibasaki, M. J. Am. Chem.
Soc. 2006, 128, 6312.
(16) (a) Ishitani, H.; Yamashita, Y.; Shimizu, H.; Kobayashi, S. J. Am. Chem.
Soc. 2000, 122, 5403. (b) Yamashita, Y.; Ishitani, H.; Shimizu, H.;
Kobayashi, S. J. Am. Chem. Soc. 2002, 124, 3292. (c) Kobayashi, S.; Saito,
S.; Ueno, M.; Yamashita, Y. Chem. Commun. 2003, 2016. (d) Kobayashi,
J.; Nakamura, M.; Mori, Y.; Yamashita, Y.; Kobayashi, S. J. Am. Chem.
Soc. 2004, 126, 9192.
(17) (a) Kobayashi, S.; Komiyama, S.; Ishitani, H. Angew. Chem., Int. Ed. 1998,
37, 979. (b) Kobayashi, S.; Kusakabe, K. -I.; Komiyama, S.; Ishitani, H. J.
Org. Chem. 1999, 64, 4220. (c) Kobayashi, S.; Kusakabe, K. -I.; Ishitani,
H. Org. Lett. 2000, 2, 1225. (d) Yamashita, Y.; Saito, S.; Ishitani, H.;
Kobayashi, S. Org. Lett. 2002, 4, 1221. (e) Kobayashi, S.; Shimizu, H.;
Yamashita, Y.; Ishitani, H.; Kobayashi, J. J. Am. Chem. Soc. 2002, 124,
13678. (f) Yamashita, Y.; Saito, S.; Ishitani, H.; Kobayashi, S. J. Am. Chem.
Soc. 2003, 125, 3793. (g) Yamashita, Y.; Kobayashi, S. J. Am. Chem. Soc.
2004, 126, 11279. (h) Kobayashi, S.; Ueno, M.; Saito, S.; Mizuki, Y.;
Ishitani, H.; Yamashita, Y. Proc. Natl. Acad. Sci. 2004, 101, 5476. (i)
Yamashita, Y.; Mizuki, Y.; Kobayashi, S. Tetrahedron Lett. 2005, 46, 1803.
(18) (a) Ishitani, H.; Ueno, M.; Kobayashi, S. J. Am. Chem. Soc. 1997, 119,
7153. (b) Ishitani, H.; Komiyama, S.; Kobayashi, S. Angew. Chem., Int.
Ed. 1998, 37, 3186. (c) Kobayashi, S.; Ishitani, H.; Ueno, M. J. Am. Chem.
Soc. 1998, 120, 431. (d) Kobayashi, S.; Hasegawa, Y.; Ishitani, H. Chem.
Lett. 1998, 1131. (e) Ishitani, H.; Kitazawa, T.; Kobayashi, S. Tetrahedron
Lett. 1999, 40, 2161. (f) Ishitani, H.; Ueno, M. Kobayashi, S. J. Am. Chem.
Soc. 2000, 122, 8180. (g) Kobayashi, S.; Kobayashi, J.; Ishitani, H.; Ueno,
M. Chem. Eur. J. 2002, 8, 4185.
(19) Ishitani, H.; Komiyama, S.; Hasegawa, Y.; Kobayashi, S. J. Am. Chem.
Soc. 2000, 122, 762.
(20) Gastner, T.; Ishitani, H.; Akiyama, R.; Kobayashi, S. Angew. Chem., Int.
Ed. 2001, 40, 1896.
(21) (a) Kobayashi, S.; Shimizu, H.; Yamashita, Y.; Ishitani, H.; Kobayashi, J.
J. Am. Chem. Soc. 2002, 124, 13678. (b) Yamashita, Y.; Kobayashi, S. J.
Am. Chem. Soc. 2004, 126, 11279.
(22) (a) Azoulay, S.; Manabe, K.; Kobayashi, S. Org. Lett. 2005, 7, 4593. (b)
Ogawa, C.; Azoulay, S.; Kobayashi, S. Heterocycles 2005, 66, 201. (c)
Boudou, M.; Ogawa, C.; Kobayashi, S. AdV. Synth. Catal. 2006, 348, 2585.
2
005, 46, 9029. (n) Kamal, A.; Ramu, R.; Azhar, M. A.; Khanna, G. B. R.
Tetrahedron Lett. 2005, 46, 2675. (o) Wu, J.; Xia, H. -G. Green Chem.
005, 7, 708.
6) (a) Das, B.; Krishnaiah, M.; Venkateswarlu, K. Tetrahedron Lett. 2006,
7, 6027. (b) Huang, J.; O’Brien, P. Synthesis 2006, 425. (c) Bonollo, S.;
2
(
4
Fringuelli, F.; Pizzo, F.; Vaccaro, L. Green Chem. 2006, 8, 960. (d)
Schneider, C. Synthesis 2006, 3919. (t) Torregrosa, R.; Pastor, I. M.; Yus,
M. Tetrahedron 2006, 63, 469.
(
7) (a) Cole, B. M.; Shimizu, K. D.; Krueger, C. A.; Harrity, J. P.; Snapper,
M. L.; Hoveyda, A. H. Angew. Chem., Int. Ed. Engl. 1996, 35, 1668. (b)
Shimizu, K. D.; Cole, B. M.; Krueger, C. A.; Kuntz, K. W.; Snapper, M.
L.; Hoveyda, A. H. Angew.Chem., Int. Ed. Engl. 1997, 36, 1704. (c) Schaus,
S. E.; Jacobsen, E. N. Org. Lett. 2000, 2, 1001. (d) Yamasaki, S.; Kanai,
M.; Shibasaki, M. J. Am. Chem. Soc. 2001, 123, 1256. (e) Zhu, C.; Yuan,
F.; Gu, W.; Pan, Y. Chem. Commun. 2003, 692. (f) Pakulski, Z.;
Pietrusiewiscz, K. M. Tetrahedron: Asymmetry 2004, 15, 41.
(
8) Iida, T.; Yamamoto, N.; Sasai, H.; Shibasaki, M. J. Am. Chem. Soc. 1997,
1
19, 4783.
(
9) Mart ´ı nez, L. E.; Leighton, J. L.; Carsten, D. H.; Jacobsen, E. N. J. Am.
Chem. Soc. 1995, 117, 5897. (b) Leighton, J. L.; Jacobsen, E. N. J. Org.
Chem. 1996, 61, 389.
(
10) (a) Iida, T.; Yamamoto, N.; Matsunaga, S.; Woo, H. -G.; Shibasaki, M.
Angew. Chem., Int. Ed. 1998, 37, 2223. (b) Matsunaga, S.; Das, J.; Roels,
J.; Vogl, E. M.; Yamamoto, N.; Iida, T.; Yamaguchi, K.; Shibasaki, M. J.
Am. Chem. Soc. 2000, 122, 2252. (c) Ready, J. M.; Jacobsen, E. N. J. Am.
Chem. Soc. 2001, 123, 2687.
8104 J. AM. CHEM. SOC.
9
VOL. 129, NO. 26, 2007

Scalemic Multidentate Niobium Complexes
A R T I C L E S
Scheme 2. Mannich-Type Reaction Catalyzed by a Nb(V)/
desymmetrization reaction of closely structurally related meso-
2
3
BINOL-Derived Catalyst System
epoxides by ring opening with aniline nucleophiles. The
system in question is constructed from scalemic multidentate
complexes of niobium(V) and BINOL derivatives and was
discovered as part of our program of research into new, high-
specification Lewis acid catalysts on the basis of less-exploited
2
4,25
metals.
Herein, we describe in detail the development of
this catalyst system and report the results of investigations into
the substrate scope and selectivity of the meso-epoxide ring-
opening reaction mediated by the same catalyst. In addition, to
the best of our knowledge, the first ever desymmetrization of
meso-aziridines with aniline nucleophiles is also described.
Results and Discussion
Development of the Multidentate-Nb(V) Catalyst System
and Initial Substrate Screening. The starting point for our
investigation was the Nb(V) system which we have developed
as a catalyst for the Mannich-type reactions of imines with
silicon enolates (Scheme 2).26 This species, prepared from a
mixture of a Nb(V) salt and novel tridentate BINOL-derived
ligand 1, was shown to catalyze the Mannich reactions in
high yields and with excellent enantioselectivities. NMR and
X-ray crystallographic studies suggested structure 2, in which
two niobium atoms are straddled by two equivalents of the
ligand, as the most plausible catalyst species. This arrangement,
in which the metal centers are held in a unique spatially well-
defined binuclear array by firm but flexible ligation, facilitates
Table 1. Nb(V) Catalyzed Ring Opening of meso-Cyclohexene
Oxide with Aniline
yield (%)a
ee (%)b
entry
Nb source
auxiliary ligand
additive
1c
Nb(OMe)5 none
NMId
18
21
29
30
34
2
4
e
e
e
e
NMId
2
3
4
5
Nb(OMe)5 (R)-BINOL
Nb(OMe)5 (S)-BINOL
Nb(OMe)5 (S)-BINOL
Nb(OMe)5 2,2′-biphenol
NMId
2,6-lutidine
2,6-lutidine
33
27
48
(
(
(
23) For a preliminary description of our early investigations see Arai, K.; Salter,
M. M.; Yamashita, Y.; Kobayashi, S. Angew. Chem., Int. Ed. 2007, 46,
9
55.
6
7
8
9
0
Nb(OMe)5 2,2′-biphenol
2,6-lutidine
2,6-lutidine
2,6-lutidine
2,6-lutidine
2,6-lutidine
49
48
49
55
49
38
34
41
48
47
24) The racemic ring opening of epoxides catalyzed by NiCl
been reported: Constantino, M. G.; Lacerda, V., Jr.; Arag a˜ o, V. Molecules
001, 6, 770.
5
(NifNb) has
Nb(OEt)5
2,2′-biphenol
2,2′-biphenol
2,2′-biphenol
n
2
Nb(O Pr)5
25) For some other applications of niobium salts in synthesis see (a) Suzuki,
K.; Hashimoto, T.; Maeta, H.; Matsumoto, T. Synlett 1992, 125. (b)
Hashimoto, T.; Maeta, H.; Matsumoto, T.; Morooka, M.; Ohba, S.; Suzuki,
K. Synlett 1992, 340. (c) Colletti, S. L.; Halterman, R. L. J. Organomet.
Chem. 1993, 455, 99. (d) Maeta, H.; Nagasawa, T.; Handa, Y.; Takei, T.;
Osamura, Y.; Suzuki, K. Tetrahedron Lett. 1995, 36, 899. (e) Howarth, J.;
Gillespie, K. Tetrahedron Lett. 1996, 37, 6011. (f) Yamamoto, M.;
Nakazawa, M.; Kishikawa, K.; Kohmoto, S. Chem. Commun. 1996, 2353.
i
Nb(O Pr)5
t
1
Nb(O Bu)5 2,2′-biphenol
Nb(OiPr)5
11
2,2′-biphenol
2,6-lutidine
67
60
44
49
63
58
62
56
33
41
31
40
38
33
48
32
i
1
1
1
2
3
4
Nb(O Pr)5
2-phenylphenol 2,6-lutidine
3-phenylphenol 2,6-lutidine
4-phenylphenol 2,6-lutidine
i
Nb(O Pr)5
i
Nb(O Pr)5
i
(g) Howarth, J.; Gillespie, K. Molecules 2000, 5, 993. (h) Andrade, C. K.
15
16
Nb(O Pr)5
phenol
2,6-lutidine
2,6-lutidine
2,6-lutidine
2,6-lutidine
i
Z.; Azevedo, N. R. Tetrahedron Lett. 2001, 42, 6473. (i) Andrade, C. K.
Z.; Oliveira, G. R. Tetrahedron Lett. 2002, 43, 1935. (j) Miyazaki, T.;
Katsuki, T. Synthesis 2003, 1046. (k) Arai, S.; Sudo, Y.; Nishida, A. Synlett
Nb(O Pr)
4-fluorophenol
5
i
t
1
1
7
8
Nb(O Pr)5
4- butylphenol
i
Nb(O Pr)5
2-naphthol
2
004, 1104. (l) Arai, S.; Takita, S.; Nishida, A. Eur. J. Org. Chem. 2005,
5
262. (m) Yadav, J. S.; Subba Reddy, B. V.; Eeshwaraiah, B.; Reddy, P.
a
Isolated yields. b Determined by chiral high-performance liquid chro-
N. Tetrahedron 2005, 61, 875. (n) Arai, S.; Sudo, Y.; Nishida, A.
Tetrahedron 2005, 61, 4639. (o) Hotha, S.; Tripathi, A. Tetrahedron Lett.
matography (HPLC). c 12 mol % 1. d N-Methylimidazole. e In CH Cl /
2
2
2
2
005, 46, 4555. (p) Sudo, Y.; Arai, S.; Nishida, A. Eur. J. Org. Chem.
006, 752. (q) Yadav, Y. S.; Narsaiah, A. V.; Basak, A. K.; Goud, P. R.;
toluene (1:1).
Sreenu, D.; Nagaiah, K. J. Mol. Catal. A: Chem. 2006, 255, 78.
26) Kobayashi, S.; Arai, K.; Shimizu, H.; Ihori, Y.; Ishitani, H.; Yamashita,
Y. Angew. Chem., Int. Ed. 2005, 44, 761.
(
(
highly substrate selective reactions. In view of the fact that the
niobium(V) atoms at the heart of our catalyst system are already
heavily coordinated, we judged that monodentate species would
function most efficiently as substrates. In view of the known
high oxophilicity of niobum salts, we selected epoxides as
suitable substrates and decided to focus on the asymmetric ring
opening of meso-epoxides with anilines.
We selected the ring-opening reaction of meso-cyclohexene
oxide 3 using aniline as a nucleophile and the catalyst
established for the Mannich type reaction described above as
our model system. The reaction was conducted under standard
reaction conditions to afford the corresponding 1,2-amino
alcohol 4 in only low yield and with extremely low enantiose-
lectivity (Table 1, entry 1). Our working model regarding the
structure of the catalytic species envisaged a niobium metal
center coordinated to the three alkoxide groups of the ligand
27) For examples of racemic ring-opening reactions of aziridines see (a) Tanner,
D. Angew. Chem., Int. Ed. Engl. 1994, 33, 599. (b) Meguro, M.; Asao, N.;
Yamamoto, Y. Tetrahedron Lett. 1994, 35, 7395. (c) Leung, W. -H.; Yu,
M.-T.; Wu, M.-C.; Yeung, L.-L. Tetrahedron Lett. 1996, 37, 891. (d)
Osborn, H. M. I.; Sweeney, J. Tetrahedron: Asymmetry 1997, 8, 1693. (e)
Lucet, D.; Le Gall, T.; Mioskowski, C. Angew. Chem., Int. Ed. 1998, 37,
2
580. (f) Sekar, G.; Singh, V. K. J. Org. Chem. 1999, 64, 2537. (g)
Chandrasekhar, M.; Sekhar, G.; Singh, V. K. Tetrahedron Lett. 2000, 41,
0079. (h) Wu, J.; Hou, X.-L.; Dai, L.-X. J. Org. Chem. 2000, 65, 1344.
1
(
i) Sweeney, J. B. Chem. Soc. ReV. 2002, 31, 247. (j) Cossy, J.; Bellosta,
V.; Alauze, V.; Desmurs, J. -R. Synthesis 2002, 2211. (k) Yadav, J. S.;
Subba Reddy, B. V.; Vishweshwar Rao, K.; Saritha Raj, K.; Prasad, A. R.
Synthesis 2002, 1061. (l) Watson, I. D. G.; Yudin, A. K. J. Org. Chem.
2
003, 68, 5160. (m) Bisai, A.; Pandey, G.; Pandey, M. K.; Singh, V. K.
Tetrahedron Lett. 2003, 44, 5839. (n) Yadav, J. S.; Subba Reddy, B. V.;
Baishya. G.; Venkatram Reddy, P.; Harshavardhan, S. J. Synthesis 2004,
1
854. (o) Fan, R.-H.; Zhou, Y.-G.; Zhang, W.-X.; Hou, X.-L.; Dai, L.-X.
J. Org. Chem. 2004, 69, 335. (p) Yadav, J. S.; Subba Reddy, B. V.;
Jyothirmai, B.; Murty, M. S. R. Tetrahedron Lett. 2005, 46, 6385. (q)
Minakata, S.; Okada, Y.; Oderaotoshi, Y.; Mokatsu, M. Org. Lett. 2005,
7
, 3509. (r) Wu, J.; Sun, X.; Ye, S.; Sun, W. Tetrahedron Lett. 2006, 47,
813.
4
J. AM. CHEM. SOC.
9
VOL. 129, NO. 26, 2007 8105

A R T I C L E S
Arai et al.
Scheme 3. Synthesis of Tetradentate Ligands 10a-e
Figure 1. Postulated structures of the catalyst in the presence of auxiliary
ligands.
with the two remaining coordination sites being occupied by
two equivalents of the alkoxide species present in the starting
Nb(OR)5 salt (Figure 1). This would imply a certain lack of
rigidity in the coordination sphere of the catalyst and subsequent
loss of enantioselectivity in the reaction. With this in mind, the
affect of adding a bidentate auxiliary ligand to the reaction with
a view to creating a still more rigid structure such as 5a or 5b
was investigated (entries 2-6).
Accordingly, the reaction was repeated in the presence of 11
mol % of (R)-BINOL, however, little improvement in either
yield or enantioselectivity was observed (entry 2). Interestingly,
use of (S)-BINOL led to a considerable improvement in
enantioselectivity, but the absolute selectivity and yield were
still low (entry 3). Use of 2,6-lutidine as an amine additive had
little effect (entry 4), whereas changing the auxiliary ligand from
catalyst combinations observed (entry 17), the yield and the
selectivity were not yet satisfactory.
(S)-BINOL to the racemic 2,2′-biphenol gave the desired product
Development of a Novel Tetradentate BINOL-Derived
Ligand System. Taken together, these results clearly indicated
that the most effective catalyst system involved coordination
of the metal center with a total of four phenoxides per niobium
atom along with addition stabilization with a more weakly bound
nitrogenous ligand. We then designed a new type of tetradentate
ligand with all four phenoxides. Thus, BINOL protected as its
bis-methoxymethyl ether 6 was treated with 2 equiv of sec-
butyl lithium, and the resulting dianion was quenched at
in 48% ee (entry 5) indicating that the observed increase in
enantioselectivity does not require a symbiotic relationship
between the chirality of ligand 1 and the auxiliary ligand. At
this point, we were able to establish the absolute configuration
of the major isomer unambiguously as 1S,2S by single-crystal
X-ray analysis of the camphenyl ester of 4. A survey of common
Nb(V) sources (entries 6-10) revealed that those constituted
from more bulky alkoxides (entries 9 and 10) gave superior
results, with the iso-propoxide salt providing the product in the
-78 °C with a range of 2-methoxymethoxybenzaldehydes 7a-e
i
best yield and ee. Thus, Nb(O Pr)5 was adopted as the niobium
to afford the resulting diol 8a-e. Treatment of the latter with
acidic methanol in CH2Cl2 resulted in cleavage of the MOM
groups and the conversion of the alcohols into the corresponding
methyl ethers giving tetraols 9a-e which were reduced with
Et3SiH/BF3•OEt2 to afford the desired tetradentate BINOL de-
rivatives 10a-e in good yields over three steps (see Scheme 3).
Application of these new ligands to the ring-opening reaction
of meso-cyclohexene oxide with aniline gave interesting results
(Table 2). It was discovered that while the catalyst derived from
10a with Nb(OMe)5 gave the expected 1S,2S-amino alcohol 4
in acceptable chemical yield but low enantioselectivity (entry
1), use of 10b afforded the 1R,2R isomer as major product of
the reaction (entry 2) albeit with low overall selectivity.
source of choice for the continued development of the catalyst
system.
At this stage, while the efficacy of adding an auxiliary ligand
to the reaction mixture had been demonstrated, the exact role
of the added species and indeed its mode of coordination (mono-
13
or bidentate) remained unclear. C NMR studies of the catalyst
prepared from a stoichiometric amount of Nb(OEt)5 with 1 (1.1
equiv) in the presence of 2,2′-biphenol (1.1 equiv) and 2,6-
lutidine (1.2 equiv) in CD2Cl2 showed, in addition to a peak at
1
60 ppm corresponding to the phenoxy aromatic carbon of
coordinated 2,2′-biphenol, a peak at 150 ppm which was
assigned to the corresponding carbon atom of an unbound
phenolic group. The result suggested that the 2,2′-biphenol was
in fact binding to the metal center in monodentate fashion, an
idea which was reinforced by the observation that not only did
use of the monomethyl ether of 2,2′-biphenol (entry 11) as
opposed to the parent diol (entry 9) in the reaction still permit
the production of the ring-opened product but that it did so in
markedly better yield than previously and with comparable
enantioselectivity. This observation prompted us to screen many
phenolic derivatives as auxiliary ligands with a view to
adumbrate further the role of the latter in the reaction (entries
i
Furthermore, it was found that switching to the Pr-substituted
ligand 10c gave the 1R,2R product ent-4 in essentially quantita-
tive yield and with comparatively high enantioselectivity (entry
3). However, increasing the steric bulk of the ligand still further
t
by the introduction of Bu groups (10d, entry 4) led to a
substantial drop-off in both yield and selectivity, whereas use
of an aryl-substituted ligand (10e, entry 5) gave the ring-opened
product in very high yield, but with the 1S,2S enantiomer as
the major product. In contrast, the 6,6′-disubstituted BINOL
derived tetradentate ligands 10f and g again gave the 1R,2R
enantiomer in very high yield but with low selectivity (entries
6 and 7). Interestingly, ligand 11 in which one of the phenoxyl
groups has been converted to a methyl ether still afforded the
1
2-18). Although the use of 4-tert-butylphenol provided the
ring-opening product in a respectable yield and with an
enantioselectivity equal to the best level observed with other
8106 J. AM. CHEM. SOC.
9
VOL. 129, NO. 26, 2007

Scalemic Multidentate Niobium Complexes
A R T I C L E S
Table 2. Ring-Opening Reaction of meso-Cyclohexene Oxide with Aniline Catalyzed by Complexes of Nb(V) and Tetradentate Ligands
entry
1a
2
Nb source
ligand
10a
10b
10c
10d
10e
10f
R1
R2
yield (%)
64
ee (%)
-35
25
Nb(OMe)5
Nb(OMe)5
Nb(OMe)5
Nb(OMe)5
Nb(OMe)5
Nb(OMe)5
Nb(OMe)5
Nb(OMe)5
H
H
H
H
H
H
Me
I
Me
iPr
tBu
Ph
iPr
iPr
iPr
76
3
quant.
63
70
4
32
5a
6
96
-18
58
93
7
10g
11
89
52
8
H
83
37
9b
Nb(O Pr)5
i
10c
10c
10c
H
H
H
iPr
iPr
iPr
85
64
90
11
66
68
i
1
1
0c
1
Nb(O Pr)5
i
Nb(O Pr)5
a
1S,2S) compound formed. b Reaction run in CH2Cl2. c Reaction run in toluene-CH2Cl2 (1:1).
(
same 1R,2R-amino alcohol as the majority of the tetradentate
ligands examined although with somewhat reduced selectivity
such outstandingly high levels of selectivity in an epoxide ring-
opening reaction are, to the best of our knowledge, unprec-
edented.
(
entry 8).
i
The marked preference for the Pr substituted tetradentate
The high reactivity and enantioselectivity of the ring-opening
was generally maintained even at lower catalyst loading although
the efficiency of the reaction was found to be dependent on
concentration especially at very low loadings (e1 mol %).
Accordingly, it was found that even at catalyst loadings as low
as 0.5 or 0.25 mol % running the reaction at high concentration
over an extended period delivered the ring-opened product in
very good yield and with high enantioselectivity.
ligand 10c was in line with that observed for the tridentate
i
system 1, however, whereas in the latter case Nb(O Pr)5 had
been the optimum niobium(V) source (using CH2Cl2 as solvent),
i
the combination of Nb(O Pr)5 and 10c in CH2Cl2 gave the 1R,2R
product in significantly lower enantiomeric excess although in
much higher chemical yield (entry 9). In mixed toluene-CH2-
Cl2 solvent systems, however, approximately the same level of
enantioselectivity was observed for the catalysts derived from
Scope of the Aniline Nucleophile. With these results in hand,
we turned our attention to the scope of the aniline in the reaction.
Accordingly, cis-but-2-ene oxide 13a was allowed to react with
a range of differently substituted anilines 12a-i in the presence
of 10 mol % of the catalyst under the conditions described
above, affording the corresponding 1,2-amino alcohols 14aa-
i
Nb(O Pr)5 as for that formed from Nb(OMe)5 (entries 10 and
1
1), but as yields and selectivities were marginally better with
the latter, Nb(OMe)5 was adopted as the niobium source of
choice.
Scope of the Desymmetrization Reaction of meso-Ep-
oxides. With these results in hand, we turned our attention to
the scope of the reaction. Accordingly, we conducted the
reaction with a range of linear and cyclic epoxides with striking
results (Table 3). These investigations revealed that the reaction
with cis-but-2-ene oxide 13a (entry 1) and aniline 12a proceeded
very smoothly to give the corresponding 1,2-hydroxylamine
14ai (Table 4). These experiments showed that catalyst activity
was general for a wide range of different anilines with both
electron-rich and electron-poor examples functioning well in
the reaction, although the ring-opened product was obtained in
slightly lower yield in the case of ortho-substituted anilines.
This generality was in striking contrast to the chemoselectivity
observed with respect to the epoxide component in the reaction
1
4aa in quantitative yield and very high enantioselectivity;
however, when the reaction was conducted with the closely
related aliphatic meso-epoxides 13b-d and the aromatically
substituted cis-stilbene oxide 13e, the reaction proceeded very
sluggishly, affording the corresponding ring-opened products
only in very low yields and low enantioselectivities (entries 2,
(
vide supra). In addition, we investigated the reactivity of cyclic
meso-epoxides 13g and 13h with electron-poor aniline 12g
entries 10-13). Gratifyingly, it was found that these substrates
(
also underwent ring opening smoothly to give the anticipated
products in good to excellent yields and acceptable enantiose-
lectivity (entries 10 and 12). Lowering the temperature still
further (-40 °C, entry 11; -30 °C, entry 13) and carrying out
the reaction under more dilute conditions (0.08 M) afforded the
ring-opened products with improved ee and in essentially the
same yield.
4-5) except for cis-oct-4-ene 13c (entry 3, 84% ee). In contrast,
on switching to cyclic meso-epoxide substrates, the correspond-
ing ring-opened products were obtained in very high yields and
generally high enantioselectivities (entries 6-15). The high
chemoselectivity of the reaction, favoring cis-but-2-ene 13a and
cyclic meso-epoxides over other linear meso-epoxides is striking;
J. AM. CHEM. SOC.
9
VOL. 129, NO. 26, 2007 8107

A R T I C L E S
Arai et al.
Table 3. Investigation of Substrate Scope in the Nb(V) Catalyzed
Ring-Opening Reaction
Table 4. Scope of the Aniline Nucleophile in the Epoxide
Ring-Opening Reaction
entry
epoxide
Ar
temp (
°
C) yield (%)a
ee (%)
1
2
3
4
5
6
7
8
9
Me
13a Ph
12a
12b
12c
12d
12e
12f
-15
-15
-15
0
-10
0
-15
-15
-15
quant.
82
92
90
95
99
89
96
96
14aa
94
84
93
91
90
90
96
96
95
Me
Me
Me
Me
Me
Me
Me
Me
13a (2-Me)Ph
13a (3-Me)Ph
13a (4-Me)Ph
13a (2-OMe)Ph
13a (4-OMe)Ph
13a (3-CF3)Ph
14ab
14ac
14ad
14ae
14af
14ag
14ah
14ai
12g
13a (3,5-(CF3)2)Ph 12h
13a (4-Br)Ph
12i
1
0
13g (3-CF3)Ph
12g
-15
98
14gg
83
1
1
1b,c
2
12g (3-CF3)Ph
13h (3-CF3)Ph
12g
12g
-40
-15
94
88
14gg
14hg
90
82
1
3b,c
13h (3-CF3)Ph
12g
-30
89
14hg
89
a
Isolated yields. b Forty-two hours. c Total concentration 0.08 M.
Table 5. Competition Experiments in the Ring Opening of
Epoxide Mixtures
a
Isolated yields unless stated otherwise. b Catalyst loading: 1 mol %:
chemical
excess
(%)c
8
2% yield, 92% ee; 0.5 mol %: 90% yield, 90% ee; 0.25 mol %: 86%
1
4aa
14
c
1
yield, 88% ee. Yield determined by N NMR vs naphthalene as internal
standard.
entry
R
yield (%)a ee (%) yield (%)b ee (%)
1
2
3
4
Et
Pr
Bu
Ph
13b 86.2
13c 83.1
13d 76.0
13e 87.3
93
93
93
93
92
1.0
2.4
0.8
1.4
1.5
56 14ba
81 14ca
57 14da
74 14ea
78 14ka
98
94
98
97
98
Competition Reactions between Mixtures of meso-Ep-
oxides. Having established parameters with regard to the scope
of both the epoxide and aniline components, we returned to the
question of the remarkable chemoselectivity of the Nb(OMe)5-
d
5
-CH2OCH2- 13k 95.7
a
Isolated yield. b Yield determined by H NMR after isolation of minor
1
1
0c catalyst system. To probe this feature of our system more
c
product, using naphthalene as internal standard. Chemical excess ) [(14aa
-
rigorously, we conducted a series of competition reactions in
which the parent aniline 12a was allowed to react with an
equimolar mixture of cis-but-2-ene oxide 13a and another,
bulkier epoxide 13b-e, 13j in the presence of the catalyst (Table
d
14)/(14aa + 14)] × 100%. 10 mol % of catalyst used.
the corresponding 1-(N-phenyl)amino-2-ols 16aa-ca, arising
from ring opening at the least hindered position of the epoxide
selectively in good yields and very high enantiomeric excesses
(Table 6, entries 1-4). The regioselectivity of the reaction was
moderate to very good (16/17 ratio up to 12:1) and in all cases
the regioisomeric 2-(N-phenyl)amino-1-ols 17aa-ca were
obtained in considerably lower optical purity.
In the case of cis-â-methylstyrene oxide 15d however, the
regioselectivity was reversed and 1-phenyl-1-(phenylamino)-
propan-2-ol 17da was isolated as the major product in good
yield but with only moderate enantioselectivity (49% ee),
whereas minor product 16da was obtained in much lower yield
but in much higher optical purity (85% ee). This was not entirely
unexpected and is indicative of a competing Lewis acid assisted
SN1-type background reaction that proceeds via an intermediate
benzylic cation which is trapped by the aniline in a largely
nonstereoselective manner. It is reasonable to speculate that the
5). To our delight, the competition reactions proceeded smoothly
and with high chemical selectivity to give predominantly 14aa,
resulting from ring opening of 13a with aniline, in high yield
and very high enantioselectivity. Only traces of the products of
addition of the nucleophile to the other epoxide 13 were formed.
In all cases except for the competition between 13a and 13c
(entry 2), the ratio of 4aa to the product derived from the other
epoxide exceeded 60:1, and in the competition reaction with
cis-dec-5-ene oxide 13d (entry 3) the selectivity was 98%.
Ring-Opening Reactions of Unsymmetrically Disubstituted
cis-Epoxides. Interestingly, the high levels of selectivity
observed in the competitive desymmetrization reactions were
maintained in the ring opening of unsymmetrical cis-epoxides.
Exposure of cis-epoxides 15a-c to aniline 12a in the presence
of the Nb(V)-10c catalyst system (10 mol %) and MS 4A gave
8108 J. AM. CHEM. SOC.
9
VOL. 129, NO. 26, 2007

Scalemic Multidentate Niobium Complexes
A R T I C L E S
Table 6. Nb-Catalyzed Ring Opening of Unsymmetrical cis-Epoxides
entry
15 (R)
13 (R1)
yield (%)a
16 (ee)b
17
16/17
1
2
3
4
5
Et
Pr
Pr
15a
15b
15b
15c
15d
H
12a
12a
12a
12a
12a
89
88
83
82
92
16aa (95)
16ba (98)
16ba (98)
16ca (97)
16da (85)e
17aa
17ba
17ba
17ca
17da
9.1/1
3.4/1
3.7/1
12.0/1
1/5.6
H
H
H
H
d
PhCH2CH2
Ph
6
7
8
9
0
1
2
Et
Et
Pr
Pr
15a
15a
15b
15b
15c
15c
15c
2-OMe
2-Me
2-OMe
2-Me
2-OMe
2,6-Me2
2-Cl
12e
12b
12e
12b
12e
12j
67
63
79
66
95
16ae (95)
16ab (88)
16be (98)
16bb (85)
16ce (98)
16cj (94)
16ck (94)
17ae
17ab
17be
17bb
17ce
17cj
8.1/1
10.4/1
4.3/1
7.2/1
8.5/1
7.0/1
11.4/1
1
1
1
PhCH2CH2
PhCH2CH2
PhCH2CH2
quant.
quant.
12k
17ck
1
1
3
4
PhCH2CH2
PhCH2CH2
15c
15c
3-Cl
3-CF3
12l
12g
82
90
16cl (99)
16cg (>99)
17cl
17cg
15.0/1
18.0/1
15
16
17
18
PhCH2CH2
PhCH2CH2
PhCH2CH2
PhCH2CH2
15c
15c
15c
15c
4-Me
4-F
4-Cl
4-Br
12d
12m
12n
12i
83
87
83
83
16cd (86)
16cm (97)
16cn (98)
16ci (94)
17cd
17cm
17cn
17ci
5.6/1
13.0/1
14.0/1
12.0/1
a
Combined yield of 16 + 17. Determined by Chiralpak HPLC. (ee of major regioisomer only). c Ratio calculated from H NMR. d Reaction run at
b
1
e
0
°C. Major isomer 17da obtained in 49% ee.
minor product is generated by a different, more stereoselective
catalyst-mediated SN2-type pathway. Next, we examined the
effect of varying the substituent on the aniline ring in the
reaction with cis-epoxides 15a-c and were pleased to find that
ring opening occurred smoothly in all cases. In the cases of the
all-aliphatic epoxides 15a and 15b (entries 6-9), the reactions
with o-anisidine 12e and o-toluene 12b proceeded in moderate
combined yield and moderate to good regioselectivity, affording
the major products 16ae-ab with good to excellent enatiose-
lectivity. In line with the previously observed trend, except for
sponding enantioenriched products have been reported, to the
best of our knowledge, no methods for desymmetrization of
meso-aziridines using anilines as nucleophiles have been
reported.
We selected aziridines 18a-e and aniline 12a as model
substrates for our initial studies and screened many aziridines
using catalyst systems derived from Nb(OMe)5 and either
BINOL ligands 1 or 10c. Although the catalysts derived from
Nb(V) and tridentate ligand 1 showed high activity in the ring-
opening reaction, the diamine products were generated in
racemic or near-racemic forms. By contrast, application of the
catalyst formed with tetradentate ligand 10c to the reaction of
N-phenyl aziridine 18a with p-toluidine 12d at room temperature
obtained the corresponding ring-opened product in relatively
good yield albeit with low enantioselectivity (Table 7, entry
1). We next examined the effect of adding molecular sieves to
the reaction mixture and were pleased to discover that when
carried out in the presence of either 4 Å or 3 Å MS, the reactions
proceeded with both impoved yields and enatioselectivities
(entries 2 and 3), although less inspiring results were obtained
with 5 Å MS. It was further found that N-benzyl aziridine 18b
and N-diphenylmethylene aziridine 18c showed moderate enan-
tioselectivity (entries 5 and 6) although longer reaction time or
higher temperatures failed to provide the products in higher
yields. Interestingly, when activated aziridines (18d and 18e)
bearing electron-withdrawing groups on the nitrogen were used,
the ring-opening reactions proceeded in very poor yield or not
at all (entries 7 and 8). These results imply that the efficiency
of the reaction is governed by the availability of the aziridine
nitrogen lone pair for coordination to the catalyst rather than
by the ability of the nitrogen center to act as a leaving group in
the ring opening.
1
7ae (entry 6), the minor regioisomer was formed in substan-
tially lower ee than the corresponding major regioisomer. On
switching to oxirane 15c (entries 10-18), the yield of the
reaction improved dramatically and the corresponding 1,2-amino
alcohols were obtained in up to quantitative yields with good
to excellent regioselectivity (16/17 up to 18:1, entry 14) and
very high enantioselectivity with anilines bearing both electron-
donating (entries 10-11, 14) and electron-withdrawing substit-
uents (entries 12-13, 15-18) in the 2-, 3-, or 4-position.
Desymmetrization of meso-Aziridines Using Aniline Nu-
cleophiles. Having demonstrated the effectiveness of our catalyst
system in mediating the desymmetrization of meso-epoxides,
we sought further to broaden the applicability of the catalyst
system to include other monodentate electrophiles. Accordingly,
we turned our attention to the ring-opening reaction of meso-
aziridines as they and their ring-opened derivatives are useful
intermediates for the synthesis of versatile nitrogen-containing
compounds. While several protocols for the nonstereoselective
ring opening of aziridines have been described, catalytic
enantioselective versions have not been well explored, and in
particular, use of meso-aziridines as substrates has been little
3
b
investigated. Although as described above examples of ring-
opening reactions of aziridines using reactive nucleophiles such
We reasoned that the relatively modest levels of enantiose-
lectivity observed in these trials might be due to competition
1
3
14
15
as TMSCN, MeMgBr, and TMSN3 to afford the corre-
J. AM. CHEM. SOC.
9
VOL. 129, NO. 26, 2007 8109

A R T I C L E S
Arai et al.
Table 7. Investigation of Conditions in the Nb-Catalyzed
Ring-Opening Reaction of Cyclohexene-Derived Aziridines with
Anilines 12a,d
Table 9. Investigation of Niobium Source and Temperature in the
Ring Opening of 18i and 12a
additive time (h) yield (%)a
ee (%)b
entry
Nb source (R)
temp (
°
C)
yield (%)a
ee (%)b
entry
aziridine (R)
aniline 13
1
2
3
4
5
6
7
8
9
Me
Et
22
22
22
22
22
15
5
0
89
85
93
91
81
92
90
81
74
79
81
74
66
81
53
79
82
84 (>99)
82
1
2
3
18a Ph
18a Ph
18a Ph
18a Ph
18b Bn
12d (4-Me)Ph none
41 73
19ad 21
12d (4-Me)Ph MS 4Å 44 92
12d (4-Me)Ph MS 3Å 43 96
12d (4-Me)Ph MS 5Å 96 32
19ad 48
19ad 53
19ad 57
19ba 40
19cd 40
c
i
Pr
Bu
Bu
Pr
Pr
Pr
Pr
Pr
t
c
4
i
5
6
7
8
12a Ph
MS 4Å 24 25
i
18c CHPh2 12d (4-Me)Ph MS 4Å 44 80
18d Ts
18e Bz
i
d
12a Ph
12a Ph
MS 4Å 24 10
MS 4Å 48 n.r.
19da
0
i
d
19ea n.d.e
i
-5
-10
-15
80
80
80
i
1
1
0
1
a
1
Determined by H NMR using dibenzylether as an internal standard.
i
Pr
b
c
Determined by chiral HPLC analysis. The reaction was performed at
d
e
-
10 °C. n.r. ) no reaction. n.d. ) not determined.
a
Isolated yields. b Determined by chiral HPLC analysis. c Reaction run
for 24 h. d ee after single recrystallization.
Table 8. Effect of Changing N-Phenyl Ring Substitution
moiety (entries 2 and 3) gave much less satisfactory results in
terms of both reactivity and enantioselectivity than those
examples in which the aryl ring bears electron-releasing groups
(entries 4-10). In particular, N-(p-NO2)Ph aziridine 18g (entry
3
) showed very low reactivity presumably because of low
entry
aziridine (Ar)
time (h)
yield (%)a
ee (%)b
electron density on the nitrogen atom, leading to poorer
coordination of the aziridine to the catalyst metal center (vide
supra). On the other hand, introduction of an electron-donating
OMe group in the para-position of the N-aryl substituent
activated the aziridine 18h toward ring opening giving the
corresponding diamine 19ha in excellent yield with improved
enantioselectivity (entry 4). Closer examination showed that
o-OMe-susbstituted aziridine 18i gave the best result in terms
of enantioselectivity (entry 5), although the m-OMe system 18j
afforded the product in slightly improved yield with lower
enantioselectivity (entry 6). Aziridines bearing other ortho-
electron-donating substituents gave comparable results (18l and
1
18a (Ph)
40
86
19aa
62
2
3
18f (4-F)Ph
18g (4-NO2)Ph
62
72
77
9
19fa
19ga
59
53
4
5
6
7
8
9
18h (4-MeO)Ph
18i (2-MeO)Ph
18j (3-MeO)Ph
18k (2-HO)Ph
18l (2-EtO)Ph
62
46
78
40
78
72
62
92
89
93
86
96
89
92
19ha
19ia
19ja
19ka
19la
70
74
56
11
72
70
69
18m (2-PhO)Ph
18n (2,4-(MeO)2Ph
19ma
19na
10
a
1
Determined by H NMR using dibenzylether as an internal standard.
b
Determined by chiral HPLC analysis.
1
8m, entries 8 and 9) as did N-(2,4-dimethoxyphenyl) substi-
with a nonstereoselective background reaction catalyzed by the
molecular sieves in the reaction mixture. The existence of such
a competition reaction was confirmed by carrying out the
corresponding control reaction in which aziridine 18a and aniline
tuted aziridine 18n (entry 10). While the importance of the
presence of electron-donating substituents is not yet completely
understood, it may be possible that the alkoxy group acts as a
supplementary binding point for a niobium center in the catalyst
and helps to fix the aziridine in a favorable orientation in the
catalyst chiral pocket. Furthermore, derivatization of an enan-
tioenriched sample of 19aa obtained under these conditions
allowed us to confirm the absolute stereochemistry of the major
enantiomer as 1S,2S by comparison of its physical data with
that of the known system. This absolute configuration is the
opposite to that obtained in the analogous ring opening of meso-
epoxides (vide supra) and further illustrates the complementarity
and synthetic utility of our system.
Having established o-methoxyophenyl (OMP) as the optimum
protecting group for the aziridine nitrogen, we re-examined the
effect of different niobium sources on the yield and selectivity
of the reaction. A comprehensive screen of niobium alkoxides
in the benchmark reaction of o-Me-substituted aziridine 18i with
aniline 12a revealed that the ring-opening reaction proceeded
1
2a were stirred together in toluene in the presence of MS 3 Å
without the catalyst for 24 h affording the corresponding racemic
ring-opened product in 60% yield. This was in direct contrast
to the reactions involving meso-epoxides in which no ring
opening mediated by molecular sieves was observed at the low
temperatures (-15 °C) used for those reactions. We therefore
sought to re-engineer our catalyst preparation conditions and
developed a new procedure in which the catalyst was formed
in the presence of molecular sieves which were then filtered
off before the addition of the substrate and the nucleophile.
Gratifyingly, application of this new protocol to the benchmark
ring-opening reaction of 18a with 12a gave the desired product
in improved chemical yield and enantioselectivity (86% and
2
8
6
2% ee, respectively) (Table 8, entry 1). Using the new
procedure, we then investigated the effect of placing substituents
on the aromatic ring of a series of N-aryl aziridines 18f-n on
reactivity and selectivity (Table 8).
i
most efficiently with Nb(O Pr)5 as the metal source, affording
Examination of these data clearly showed that N-aryl aziri-
dines having para- electron-withdrawing substituents on the aryl
(
28) Aoyama, H.; Tokunaga, M.; Kiyosu, J.; Iwasawa, T.; Obora, Y.; Tsuji, Y.
J. Am. Chem. Soc. 2005, 127, 10474.
8110 J. AM. CHEM. SOC.
9
VOL. 129, NO. 26, 2007

Scalemic Multidentate Niobium Complexes
A R T I C L E S
Table 10. Investigation of Scope of the Aniline Nucleophile
satisfactory results in terms of both reactivity and selectivity
were obtained. The best results were obtained using anilines
with electron-withdrawing groups such as CF3 or a halogen atom
at either the meta or para position. We also examined a range
of monocyclic aziridines 20a, 20b, and bicyclic azridines 20c-
2
0e. In general, the substrates showed good to high reactivity,
but selectivities were lower than those observed in the analogous
ring-opening reaction of the benchmark [4.1.0] system 18i.
Within the range of substrates examined, the bicyclic aziridines
20c-20e were found to be more reactive than their monocyclic
counterparts 20a and 20b, presumably because of the effects
of greater release of ring strain in the reaction of the bicyclic
systems. In addition, all the final 1,2-diamine products were
obtained as solids and in the great majority of cases could be
isolated as essentially single enantiomers after a single recrys-
tallization. This straightforward operation further enhances the
synthetic utility of the current methodology and provides a new
protocol for the synthesis of enantiopure or highly enantioen-
riched C2-symmetric and non-C2-symmetric 1,2-diamines.
Conclusion
In summary, we have discovered and developed a Lewis acid
system on the basis of niobium alkoxides and a tetradentate
BINOL derivative which catalyzes the desymmetrative ring
opening of both meso-epoxides and meso-aziridines with anilines
giving the corresponding (R,R)-1,2-amino alcohol and (S,S)-
1
,2-diamine products in good to excellent yields and very high
to excellent enantioselectivity. Furthermore, the catalyst displays
a remarkable ability to distinguish between different meso-
epoxides stemming from its sensitivity to steric bulk at the
â-carbon of epoxides. In the ring-opening reactions of both
epoxides and aziridines, formation of the catalyst in the presence
of molecular sieves was found to be important for the realization
of high yields and selectivity. The synthetic utility of the system
is further enhanced by its ability to promote the asymmetric
ring opening of nonsymmetric cis-epoxides with anilines with
good to excellent regio- and enantioselectivity. To the best of
our knowledge, the protocol described herein constitutes the
first report not only of the catalytic enantioselective desymme-
trization of both meso-epoxides and meso-aziridines with
anilines but also of a nonenzymic catalyst system that provides
stereochemically complementary products for two different but
closely related reactions, and as such we believe that it will be
of significant interest to the chemical community.
a
Isolated yields. Determined by chiral HPLC analysis. c ee after a single
b
d
e
recrystallization. Reaction run at room temperature (rt). 20 mol %
catalyst. Reaction run at 0 °C.
f
the desired diamine product 19ia at room temperature in 93%
yield and 81% ee (Table 9, entry 3). Furthermore, it was
discovered that lowering the reaction temperature still further
to 5 °C gave the product with higher enantioselectivity and with
no deleterious effect on yield (entry 7). However, no advantage
was found in going to even lower temperatures, as under such
conditions both yield and enantioselectivity were gradually
eroded.
Scope of the Aniline Nucleophile in the Desymmetrization
of meso-Aziridines. Having identified the optimal catalyst and
reaction conditions (5 mol % Nb(O Pr)5, 5.5 mol % 10c, MS 3
Å removed by filtration, toluene, 5 °C), we proceeded to probe
the substrate scope of the reaction (Table 10). Screening several
readily available anilines revealed that the reaction had broad
generality for this class of nucleophile, giving the desired 1,2-
diaryl amines in good to high yields with moderate to good
enantioselectivity in the majority of cases.
i
Acknowledgment. This work was partially supported by a
Grant-in-Aid for Scientific Research from Japan Society of the
Promotion of Sciences (JSPS).
Supporting Information Available: Full experimental pro-
cedures and spectral data for all compounds used in the study.
This material is available free of charge via the Internet at
http://pubs.acs.org.
In the case of ortho-substituted aniline 12b (entry 2) and those
with an electron-donating substituent (12f, entry 5), less
JA0708666
J. AM. CHEM. SOC.
9
VOL. 129, NO. 26, 2007 8111