A R T I C L E S
Nojiri et al.
Scheme 3. Catalytic Asymmetric Mannich-Type Reaction of 2a and
3a Using a Catalyst Prepared by Mixing a Premixed (R)-1/Sc
Solution and a (S)-1/Er Solution
shown in Figure 2. Clear differences in the spectral pattern were
observed between the CD spectra of 1, and the 1/Sc ) 2:1 and
1/Er ) 2:1 solution, suggesting that a chiroptically different
1/RE assembly developed in each solution (Figure 2). Due to
the exquisite combination of factors; (1) structural flexibility
of the chiral ligand 1; (2) high coordination number of RE; and
(3) availability of various coordination patterns of RE, 1
provided chiroptically different asymmetric catalysts in response
to the properties of the RE, consequently affording anti and
syn products with high enantioselectivity under otherwise
identical reaction conditions. An initial rate kinetics study
revealed clear ligand acceleration in both Sc and Er cases (Figure
3).20 Also worth noting is that 1/Er exhibited higher catalytic
activity than the Sc counterparts. This can be attributed to the
different three-dimensional architectures of these catalysts.21 A
significant deviation from linearity between the enantiopurity
of 1 and that of the product in both the Sc and Er cases suggested
the involvement of multiple molecules of 1 around the metal
center at the transition state.22
(Scheme 1b).3,18,19 Determination of the absolute configuration
of syn-4aa by X-ray crystallography revealed that enantiose-
lection of the enolate derived from 2a was inverted. The reversal
of diastereoselectivity was uniformly observed with a broad
range of substrates (Table 1). Particularly noteworthy is that
enhanced catalytic activity was observed with the 1/Er catalyst,
allowing the reaction to reach completion after 1 h of stirring
at 0 °C with 2 mol % of catalyst loading. A comparable level
of stereoselectivity was attained with as little as 1 mol % of
catalyst loading with an extended reaction time (entry 2).
Detrimental effects on the reaction time were detected in the
reaction with 6- and 7-membered R-cyanoketones 2b and 2c,
and lower diastereoselectivity was observed (entries 3,4).
Generally, a high syn-selectivity (anti/syn ) 7/93-4/96) and
enantioselectivity (91-99% ee) were observed in the reaction
with a wide range of aromatic N-Boc imines 3, including those
with heteroaromatic functionality (entries 5-12, 14). Lower
enantioselectivity was observed with imine with a 2-furyl
substituent (entry 13).
The Catalyst in the Presence of Both Sc and Er. With a
suitable catalytic system of 1/RE in hand, in which 1 displayed
different assembly patterns in response to different REs, we
envisioned in situ structural/functional modifications in the
course of the catalytic asymmetric Mannich-type reaction. An
attempted reaction of 2a and 3a with the 1/Sc/Er ) 4:1:1 catalyst
afforded the anti-product 4aa preferentially (anti/syn ) 89/11)
in 88% ee (Scheme 2), comparable to that obtained with the
1/Sc catalyst (Scheme 1a). If both 1/Sc (1/Sc ) 2:1) and 1/Er
(1/Er ) 2:1) catalysts were formed in a comparable quantity
and promoted the reaction independently, the consequence
should have been a syn-selective reaction because the initial
rate kinetics experiment revealed that the reaction rate with the
1/Er (rate constant kEr ) 1.64 × 10-5 mol L-1 s-1) catalyst was
7.6-fold faster than that of the 1/Sc catalyst (kSc ) 2.16 × 10-6
mol L-1 s-1) (Figure 3). At this point, there were two
possibilities: (1) a 1/Sc/Er heteropolymetallic complex was
formed and promoted the anti-selective reaction with enanti-
oselectivity similar to that obtained with 1/Sc; or (2) a 1/Sc
catalyst was preferentially formed over the 1/Er catalyst,
affording a similar reaction output to that of the 1/Sc catalyst.
The CD spectra of the 1/Sc/Er ) 4:1:1, 3:1:1, and 4:2:1 solutions
would represent the formation of the aggregate, which had
chiroptical properties similar to those of the original anti-
selective catalyst 1/Sc ) 2:1 (Figure 4). In particular, the CD
spectral pattern of 1/Sc/Er ) 4:2:1 showed a nearly perfect
match with that of 1/Sc ) 2:1, suggesting that most Er3+ would
be excluded from the coordination site of 1 due to a competitive
coordination of Sc3+ to 1. The formation of the 1/Sc/Er
heteropolymetallic aggregate showing a similar CD pattern
would be less likely. The lability of 1 in the present catalytic
system was examined in the following attempted reaction. When
separate solutions of preformed (R)-1/Sc ) 2:1 and (S)-1/Er )
2:1 solution were mixed together and the resulting solution was
CD and Kinetic Analyses of the 1/Sc and 1/Er Catalyst. Of
prime importance in the present catalytic system is the origin
of diastereoswitching. The paramagnetic nature of Er hampered
a detailed NMR analysis of the 1/Er catalyst. Circular dichroism
(CD) spectra of 1/Sc and 1/Er catalyst provided insight into
the differences in the assembly states of each catalytic system.
The CD spectrum of 1 at the UV-vis region in ethyl acetate is
(16) For recent reviews on direct Mannich reactions using unmodified
pronucleophiles, see: (a) Shibasaki, M.; Matsunaga, S. J. Organomet.
Chem. 2006, 691, 2089. (b) Ting, A.; Schaus, S. E. Eur. J. Org. Chem.
2007, 5797. For general reviews on catalytic asymmetric Mannich
reactions: (c) Kobayashi, S.; Ueno, M. In ComprehensiVe Asymmetric
Catalysis,Supplement 1; Jacobsen, E. N., Pfaltz, A., Yamamoto, H.,
Eds; Springer: Berlin, 2003; Chapter 29.5, p 143. (d) Friestad, G.;
Mathies, A. K. Tetrahderon 2007, 63, 2541.
(17) For selected examples of catalytic asymmetric Mannich reactions with
1,3-dicarbonyl compounds generating quaternary carbon stereocenters,
see: (a) Marigo, M.; Kjærsgaard, A.; Juhl, K.; Gathergood, N.;
Jørgensen, K. A. Chem. Eur. J. 2003, 9, 2359. (b) Hamashima, Y.;
Sasamoto, N.; Hotta, D.; Somei, H.; Umebayashi, N.; Sodeoka, M.
Angew. Chem., Int. Ed. 2005, 44, 1525. (c) Ting, A.; Lou, S.; Schaus,
S. E. Org. Lett. 2006, 8, 2003. (d) Tillman, A. L.; Ye, J.; Dixon, D. J.
Chem. Commun. 2006, 1191. (e) Yamaoka, Y.; Miyabe, H.; Yoshi-
zumi, Y.; Takemoto, Y. Synthesis 2007, 2571. (f) Chen, Z.; Morimoto,
H.; Matsunaga, S.; Shibasaki, M. J. Am. Chem. Soc. 2008, 130, 2170.
For selected examples of catalytic asymmetric Mannich reactions with
1,3-dicarbonyl compounds generating tertiary stereocenters, see: (g)
Uraguchi, D.; Terada, M. J. Am. Chem. Soc. 2004, 126, 5356. (h)
Lou, S.; Taoka, B. M.; Ting, A.; Schaus, S. E. J. Am. Chem. Soc.
2005, 127, 11256. (i) Song, J.; Wang, Y.; Deng, L. J. Am. Chem.
Soc. 2006, 128, 6048. (j) Sasamoto, N.; Dubs, C.; Hamashima, Y.;
Sodeoka, M. J. Am. Chem. Soc. 2006, 128, 14010.
(20) The rate acceleration suggested the involvement of hydrogen-bonding
to promote the reaction and enhance the stereodiscrimination at the
transition state. The hydrogen-bonding in peptide-based catalysts is
well-documented, see: (a) Copeland, G. T.; Miller, S. J. J. Am. Chem.
Soc. 2001, 123, 6496. (b) Sculimbrene, B. R.; Morgan, A. J.; Miller,
S. J. J. Am. Chem. Soc. 2002, 124, 11653. (c) Jakobsche, C. E.; Peris,
G.; Miller, S. J. Angew. Chem., Int. Ed. 2008, 47, 6707, and references
therein.
(18) A general review on enantioselectivity reversal using a single chiral
source, see: Tanaka, T.; Hayashi, M. Synthesis 2008, 3361, and
references therein.
(19) A notable example of an asymmetric catalyst exhibiting sharp enantio-
switching, depending on the assembly state of the catalyst, see: Kato,
N.; Mita, T.; Kanai, M.; Therrien, B.; Kawano, M.; Yamaguchi, K.;
Danjo, H.; Sei, Y.; Sato, A.; Furusho, S.; Shibasaki, M. J. Am. Chem.
Soc. 2006, 128, 6768.
(21) We cannot exclude the possibility that a higher ligand exchange rate
of Er would contribute to enhanced catalyst turnover.
(22) See Supporting Information for details.
9
3782 J. AM. CHEM. SOC. VOL. 131, NO. 10, 2009