Asymmetric supramolecular primary amine catalysis in aqueous buffer: Connections of selective recognition and asymmetric catalysis

Article abstract of DOI:10.1021/ja102819g

A new approach of asymmetric supramolecular catalysis has been developed by combining the supramolecular recognition of β-cyclodextrin (β-CD) and the superior property of a chiral primary amine catalyst. The resulted β-CD enamine catalysts could effectively promote asymmetric direct aldol reactions with excellent enantioselectivity in an aqueous buffer solution (pH = 4.80). The identified optimal catalyst CD-1 shows interesting characteristics of supramolecular catalysis with selective recognition of aldol acceptors and donors. A detailed mechanistic investigation on such supramolecular catalysis was conducted with the aid of NMR, fluorescence, circular dichroism, and ESI-MS analysis. It is revealed that the reaction is initialized first by binding substrates into the cyclodextrin cavity via a synergistic action of hydrophobic interaction and noncovalent interaction with the CD-1 side chain. A rate-limiting enamine forming step is then involved which is followed by the product-generating C-C bond formation. A subsequent product release from the cavity completes the catalytic cycle. The possible connections between molecular recognition and asymmetric catalysis as well as their relevance to enamine catalysis in both natural enzymes and organocatalysts are discussed based on rational analysis.

Full text of DOI:10.1021/ja102819g

Published on Web 04/30/2010
Asymmetric Supramolecular Primary Amine Catalysis in
Aqueous Buffer: Connections of Selective Recognition and
Asymmetric Catalysis
Shenshen Hu, Jiuyuan Li, Junfeng Xiang, Jie Pan, Sanzhong Luo,* and
Jin-Pei Cheng*
Beijing National Laboratory for Molecular Sciences (BNLMS), CAS Key Laboratory of
Molecular Recognition and Function, Institute of Chemistry, Chinese Academy of Sciences,
Beijing, 100190, China
Received April 5, 2010; E-mail: luosz@iccas.ac.cn; chengjp@most.cn
Abstract: A new approach of asymmetric supramolecular catalysis has been developed by combining the
supramolecular recognition of ꢀ-cyclodextrin (ꢀ-CD) and the superior property of a chiral primary amine
catalyst. The resulted ꢀ-CD enamine catalysts could effectively promote asymmetric direct aldol reactions
with excellent enantioselectivity in an aqueous buffer solution (pH ) 4.80). The identified optimal catalyst
CD-1 shows interesting characteristics of supramolecular catalysis with selective recognition of aldol
acceptors and donors. A detailed mechanistic investigation on such supramolecular catalysis was conducted
with the aid of NMR, fluorescence, circular dichroism, and ESI-MS analysis. It is revealed that the reaction
is initialized first by binding substrates into the cyclodextrin cavity via a synergistic action of hydrophobic
interaction and noncovalent interaction with the CD-1 side chain. A rate-limiting enamine forming step is
then involved which is followed by the product-generating C-C bond formation. A subsequent product
release from the cavity completes the catalytic cycle. The possible connections between molecular
recognition and asymmetric catalysis as well as their relevance to enamine catalysis in both natural enzymes
and organocatalysts are discussed based on rational analysis.
Introduction
nature for billions of years, is still poorly understood in the
context of supramolecular systems. To elucidate the possible
Reaction pockets are ubiquitous in enzymes, wherein superb
stereocontrol and catalytic efficiency are achieved by catching
substrates in a beneficial orientation and conformation and by
stabilizing the transition state.1 Inspired by Nature, the construc-
tion of various supramolecular hosts, either synthetic or natural,
to mimic the catalysis of enzymes has been the subject of
intensive research for decades in supramolecular chemistry.2
While remarkable acceleration in reactivity and improvement
of regioselectivity have been achieved in many reactions Via
supramolecular strategy/approaches,3 there are a less number
of reports on enantioselective surpamolecular catalysis, par-
ticularly those related to biomimetic asymmetric catalysts that
could work favorably under normal enzymatic conditions.4 The
reason may be due to the usually nontrivial synthetic efforts in
constructing the chiral supramolecular systems as well as the
frequently encountered failures in converting the exquisite
recognition properties of the deliberately designed host-guest
system into catalytic turnovers.5 Clearly, the delicate balance
between recognition and catalysis, which has been practiced in
connections between recognition and chiral catalysis with a
simple supramolecule-linked asymmetric catalyst will certainly
shed new light on supramolecular catalyst design and strengthen
our understanding of enzymatic catalysis as well.
Herein, we present the first example of asymmetric supramo-
lecular primary amine catalysts that promote direct aldol
reactions with high efficiency and stereoselectivity in aqueous
buffer (pH ) 4.80). With a quite straightforward strategy, a
simple and naturally occurring chiral host, i.e. ꢀ-cyclodextrin,
is endowed with proven organocatalytic motifs6 and such a
marriage resulted in a viable approach that enabled efficient
asymmetric supramolecular catalysis beyond the reach of small
(3) For recent examples of achiral supramolecular host-guest catalysis,
see:(a) Klo¨ck, C.; Dsouza, R. N.; Nau, W. N. Org. Lett. 2009, 12,
2595–2598. (b) Pinacho Criso´stomo, F. R.; Lledo´, A.; Shenoy, S. R.;
Iwasawa, T.; Rebek, J. J. J. Am. Chem. Soc. 2009, 131, 7402–7410.
(c) Yebeutchou, R. M.; Dalcanale, E. J. Am. Chem. Soc. 2009, 131,
2452–2453. (d) Kumar, V. P.; Reddy, V. P.; Sridhar, R.; Srinivas, B.;
Narender, M.; Rao, K. R. J. Org. Chem. 2008, 73, 1646–1648. (e)
Pluth, M. D.; Bergman, R. G.; Raymond, K. N. J. Org. Chem. 2009,
74, 58–63. (f) Hastings, C. J.; Fiedler, D.; Bergman, R. G.; Raymond,
K. N. J. Am. Chem. Soc. 2008, 130, 10977–10983. (g) Pluth, M. D.;
Bergman, R. G.; Raymond, K. N. J. Org. Chem. 2008, 73, 7132–
7136. (h) Das, S.; Brudvig, G. W.; Crabtree, R. H. Chem. Commun.
2008, 413–424. (i) Fenger, T. H.; Marinescu, L. G.; Bols, M. Org.
Biomol. Chem. 2009, 7, 933–943. (j) Pluth, M. D.; Bergman, R. G.;
Raymond, K. N. J. Am. Chem. Soc. 2008, 130, 11423–11429. (k)
Cafeo, G.; Kohnke, F. H.; Valenti, L. Tetrahedron Lett. 2009, 50,
4138–4140. (l) Shenoy, S. R.; Pinacho Criso´stomo, F. R.; Iwasawa,
T.; Rebek, J. J. J. Am. Chem. Soc. 2008, 130, 5658–5659.
(1) (a) Kirby, A. J. Angew. Chem., Int. Ed. Engl. 1996, 35, 706–724. (b)
Cacciapaglia, R.; Di Stefano, S.; Mandolini, L. Acc. Chem. Res. 2004,
37, 113–122. (c) Ringe, D.; Petsko, G. A. Science 2008, 320, 1428–
1429.
(2) (a) van Leeuwen, P. W. N. M., Ed. Supramolecular Catalysis; Wiley-
VCH: Weinheim, 2008. (b) Breslow, R., Ed.; Artificial enzyme; Wiley-
VCH: Weinheim, 2005. (c) Pluth, M. D.; Bergman, R. G.; Raymond,
K. N. Acc. Chem. Res. 2009, 42, 1650–1659. (d) Rebek, J. J. Acc.
Chem. Res. 2009, 42, 1660–1668.
9
7216 J. AM. CHEM. SOC. 2010, 132, 7216–7228
10.1021/ja102819g  2010 American Chemical Society

Selective Recognition and Asymmetric Catalysis
A R T I C L E S
Scheme 1. Chiral Primary Amine Catalyzed Asymmetric Direct
molecules. Moreover, systematic mechanism studies with the
aid of typical host-guest analyzing techniques such as X-ray
crystallography, NMR, UV, CD, and fluorescence spectroscopy
reveal some catalytic features resembling those of the enzyme
catalysis: (1) substrates binding Via cooperative noncovalent
interactions such as hydrophobic effect, hydrogen bonding, or
electrostatic interaction; (2) recognition induced conformation
changes of the host molecule that favor the catalysis; and (3)
stereocontrol Via synergistic interactions of chiral host and side
chains. The details of the mechanism studies as well as the
synthetic scopes and limitations of the current asymmetric
supramolecular catalyst are also presented.
Aldol Reactions
design is based on the considerations that (1) cyclodextrins are
readily available and have been extensively explored as
enzymatic mimics in aqueous buffer;8 (2) in particular, cyclo-
dexdrin derivatives have already been examined as enamine or
enol-based aldolase mimics;9 (3) moreover, native cyclodextrins
together with derivatives have been proven to be feasible
asymmetric catalysts in photocatalytic reactions with moderate
to high enantioselectivity;4f,j,l,17b,c and (4) recently, the use of
cyclodextrins as an immobilizing host in asymmetric aldol
catalysis has also been attempted in a few studies.10 Based on
our previous research on chiral primary amine catalysis, a
cyclohexadiamine skeleton is selected as the primary aminocata-
lytic motif because of its simple structure and high reactivity
as well as selectivity.7a-c,m
Supramolecular chiral diamine CD-1 to CD-6 were readily
synthesized by nucleophilic substitution of mono-(O-6-tosyl)-ꢀ-
cyclodextrin with corresponding diamine in DMF at 80 °C under
an argon atmosphere as previously described (Scheme 2).11 After
precipitation with a large quantity of actone, pure products could
be obtained in gram scale. Unlike native ꢀ-cyclodextrin, all the
compounds are very soluble in water, which makes them suitable
for catalytic applications in aqueous media.
Results and Discussions
1. Design and Synthesis of Catalysts. Simple chiral primary
amines (e.g., 7) have recently been reported to serve as efficient
enamine-type catalysts that mechanistically and functionally
resemble the lysine-based aldolases (Scheme 1).7 It is envisioned
that efficient and biomimetic asymmetric supramolecular ca-
talysis may be evolved by covalently connecting the established
primary aminocatalytic motif with a chiral supramolecular host.
To achieve this, we have chosen the simple natural cyclodextrins
(CD) among several prominent supramolecular hosts. This
(4) For asymmetric supramolecular catalysis with moderate enantiose-
lectivity in aqueous media, see: ref 8g and (a) Fasella, E.; Dong, S. D.;
Breslow, R. Bioorg. Med. Chem. 1999, 7, 709–714. (b) Rousseau, C.;
Christensen, B.; Bols, M. Eur. J. Org. Chem. 2005, 2734–2739. (c)
Reddy, M. A.; Bhanumathi, N.; Rao, K. R. Chem. Commun. 2001,
1974–1975. (d) Suresh, P.; Pitchumani, K. Tetrahedron: Asymmetry
2008, 19, 2037–2044. (e) Brown, C. J.; Bergman, R. G.; Raymond,
K. N. J. Am. Chem. Soc. 2009, 131, 17530–17531. For examples of
supramolecular photochirogenesis in aqueous media, see ref 17b, c
and (f) Luo, L.; Liao, G.; Wu, X.; Lei, L.; Tung, C.-H.; Wu, L.-Z. J.
Org. Chem. 2009, 74, 3506–3515. (g) Wada, T.; Nishijima, M.;
Fujisawa, T.; Sugahara, N.; Mori, T.; Nakamura, A.; Inoue, Y. J. Am.
Chem. Soc. 2003, 125, 7492–7493. (h) Nishijima, M.; Wada, T.; Mori,
T.; Pace, T. C. S.; Bohne, C.; Inoue, Y. J. Am. Chem. Soc. 2007, 129,
3478–3479. (i) Nishijima, M.; Pace, T. C. S.; Nakamura, A.; Mori,
T.; Wada, T.; Bohne, C.; Inoue, Y. J. Org. Chem. 2007, 72, 2707–
2715. (j) Lu, R.; Yang, C.; Cao, Y.; Tong, L.; Jiao, W.; Wada, T.;
Wang, Z.; Mori, T.; Inoue, Y. J. Org. Chem. 2008, 73, 7695–7701.
(k) Nishioka, Y.; Yamaguchi, T.; Kawano, M.; Fujita, M. J. Am. Chem.
Soc. 2008, 130, 8160–8161. (l) Yang, C.; Mori, T.; Inoue, Y. J. Org.
Chem. 2008, 73, 5786–5794.
Crystals of CD-1 were obtained by slow evaporation of an
aqueous solution at room temperature for a few weeks. X-ray
crystallographic analysis proved the molecular structure of CD-1
(Figure 1; see Supporting Information Table S4 for detailed
crystal data). In the solid state, CD-1 assembles into a linear
(8) For research on enzyme mimics based on cyclodextrin, see: (a)
Breslow, R.; Dong, S. D. Chem. ReV. 1998, 98, 1997–2012. (b)
Rekharsky, M. V.; Inoue, Y. Chem. ReV. 1998, 98, 1875–1918. (c)
Iglesias, E. J. Am. Chem. Soc. 1998, 120, 13057–13069. (d) Ortega-
Caballero, F.; Rousseau, C.; Christensen, B.; Petersen, T. E.; Bols,
M. J. Am. Chem. Soc. 2005, 127, 3238–3239. (e) Ortega-Caballero,
F.; Bjerre, J.; Laustsen, L. S.; Bols, M. J. Org. Chem. 2005, 70, 7217–
7226. (f) Dong, Z.; Liu, J.; Mao, S.; Huang, X.; Yang, B.; Ren, X.;
Luo, G.; Shen, J. J. Am. Chem. Soc. 2004, 126, 16395–16404. (g)
Chan, W.-K.; Yu, W.-Y.; Che, C.-M.; Wong, M.-K. J. Org. Chem.
2003, 68, 6576–6582. (h) Yang, J.; Gabriele, B.; Belvedere, S.; Huang,
Y.; Breslow, R. J. Org. Chem. 2002, 67, 5057–5067. (i) Rousseau,
C.; Christensen, B.; Petersen, T. E.; Bols, M. Org. Biomol. Chem.
2004, 2, 3476–3482. (j) Ba´scuas, J.; Garc´ıa-R´ıo, L.; Leis, J. R. Org.
Biomol. Chem. 2004, 2, 1186–1193. (k) Chou, D. T. H.; Zhu, J.;
Huang, X.; Bennet, A. J. J. Chem. Soc., Perkin Trans. 2 2001, 83–89.
(l) Iglesias, E.; Ferna´ndez, A. J. Chem. Soc., Perkin Trans. 2 1998,
1691–1700. (m) Rousseau, C.; Ortega-Caballero, F.; Nordstrøm, L. U.;
Christensen, B.; Petersen, T. E.; Bols, M. Chem.sEur. J. 2005, 11,
5094–5101. (n) Bjerre, J.; Nielsen, E. H.; Bols, M. Eur. J. Org. Chem.
2008, 745–752. (o) Marinescu, L. G.; Bols, M. Angew. Chem., Int.
Ed. 2006, 45, 4590–4593.
(5) For a discussion, see: (a) Diederich, F. Angew. Chem., Int. Ed. 2007,
46, 68–69. For the reviews of construction for supramolecular cavity,
see: (b) Hof, F.; Craig, S. L.; Nuckolls, C.; Rebek, J. J. Angew. Chem.,
Int. Ed. 2002, 41, 1488–1508. (c) Saalfrank, R. W.; Maid, H.; Scheurer,
A. Angew. Chem., Int. Ed. 2008, 47, 8794–8824. (d) Kawase, T.;
Kuruta, H. Chem. ReV. 2006, 106, 5250–5273. (e) Mateos-Timoneda,
M. A.; Grego-Calama, M.; Reinhoudt, D. N. Chem. Soc. ReV. 2004,
33, 363–372.
(6) (a) Mukherjee, S.; Yang, J. W.; Hoffman, S.; List, B. Chem. ReV.
2007, 107, 5471–5569. (b) Berkessel, A.; Groger, H. Asymmetric
Organocatalysis; Wiley-VCH: Weinheim, 2005.
(7) (a) Luo, S. Z.; Xu, H.; Li, J. Y.; Zhang, L.; Cheng, J.-P. J. Am. Chem.
Soc. 2007, 129, 3074–3075. (b) Luo, S. Z.; Xu, H.; Chen, L.; Cheng,
J.-P. Org. Lett. 2008, 10, 1775–1778. (c) Luo, S. Z.; Xu, H.; Zhang,
L.; Li, J.; Cheng, J.-P. Org. Lett. 2008, 10, 653–656. (d) Xu, X.-Y.;
Wang, Y.-Z.; Gong, L.-Z. Org. Lett. 2007, 9, 4247–4249. (e)
Ramasastry, S. S. V.; Albertshofer, K.; Utsumi, N.; Tanaka, F.; Barbas,
C. F., III. Angew. Chem., Int. Ed. 2007, 46, 5572–5575. (f) Li, J.;
Luo, S. Z.; Cheng, J.-P. J. Org. Chem. 2009, 74, 1747–1750. (g)
Ramasastry, S. S. V.; Zhang, H.; Tanaka, F.; Barbas, C. F., III. J. Am.
Chem. Soc. 2009, 129, 288–289. (h) Utsumi, N.; Imai, M.; Tanaka,
F.; Ramasastry, S. S. V.; Barbas, C. F., III. Org. Lett. 2007, 9, 3445–
3448. (i) Ramasastry, S. S. V.; Albertshofer, K.; Utsumi, N.; Barbas,
C. F., III. Org. Lett. 2008, 10, 1621–1624. (j) Zhu, M.-K.; Xu, X.-Y.;
Gong, L.-Z. AdV. Synth. Catal. 2008, 350, 1390–1396. (k) Wu, X.;
Ma, Z.; Ye, Z.; Qian, S.; Zhao, G. AdV. Synth. Catal. 2009, 351, 158–
162. (l) Zheng, B.; Liu, Q.; Guo, C.; Wang, X.; He, L. Org. Biomol.
Chem. 2007, 5, 2913–2915. (m) Luo, S. Z.; Qiao, Y.; Zhang, L.; Li,
J.; Li, X.; Cheng, J.-P. J. Org. Chem. 2009, 74, 9521–9523.
(9) (a) Breslow, R.; Graff, A. J. Am. Chem. Soc. 1993, 115, 10988–10989.
(b) Desper, J. M.; Breslow, R. J. Am. Chem. Soc. 1994, 116, 12081–
12082. (c) Yuan, D.-Q.; Dong, S. D.; Breslow, R. Tetrahedron Lett.
1998, 39, 7673–7676. (d) Tagaki, W.; Yamamoto, H. Tetrahedron
Lett. 1991, 32, 1207–1208. (e) Yuan, D.-Q.; Xie, R. G.; Zhao, H. M.
Chin. Chem. Lett. 1991, 2, 617–620. (f) Watanabe, K.; Yamada, Y.;
Goto, K. Bull. Chem. Soc. Jpn. 1985, 58, 1401–1406.
(10) (a) Liu, K.; Haeussinger, D.; Woggon, W. D. Synlett 2007, 2298–
2300. (b) Shen, Z. X.; Ma, J. M.; Liu, Y. H.; Jiao, C. J.; Li, M.; Zhang,
Y. W. Chirality. 2005, 17, 556–558. (c) Huang, J.; Zhang, X.;
Armstrong, D. W. Angew. Chem., Int. Ed. 2007, 46, 9073–9077.
(11) Petter, R. C.; Salek, J. S.; Sikorski, C. T; Kumaravel, G.; Lin, F. T.
J. Am. Chem. Soc. 1990, 112, 3860–3868.
9
J. AM. CHEM. SOC. VOL. 132, NO. 20, 2010 7217

A R T I C L E S
Hu et al.
Scheme 2a
a (a) Representative synthetic route for supramolecular diamine catalyst. (b) Supramolecular diamine catalysts CD-1 to CD-6 and small molecular analogues
7 to 10.
direct aldol reactions. Previously, significant efforts have been
made to mimic the aldolase I mediated processes using a
supramolecular system. Unfortunately, no enantioselectivity has
been achieved in these studies.9 Even in the golden age of
organocatalysis, where small molecules like L-proline have been
shown to mimic mechanistically aldolase I with spectacular
stereocontrol, the realization of such catalysis under enzymatic
conditions, i.e. buffered aqueous media with reasonable enan-
tioselectivity, remains an elusive goal.13 In this context, it is
delightful to find that the primary-secondary diamine catalyst
CD-1 is indeed a viable aldol catalyst in an aqueous buffer
solution with good enantioselectivity. Using the reaction of
acetone and p-nitrobenzaldehyde as a model reaction, the
pH-activity/stereoselectivity profiles of CD-1 were determined
(Figure 3), and the optimal pH was found to be approximately
(12) (a) Liu, Y.; You, C.-C.; Zhang, M.; Weng, L.-H.; Wada, T.; Inoue,
Y. Org. Lett. 2000, 2, 2761–2763. (b) Hirotsu, K.; Higuchi, T.; Fujita,
K.; Ueda, T.; Shinoda, A.; Imoto, T.; Tabushi, I. J. Org. Chem. 1982,
47, 1143–1144. (c) Kamitori, S.; Hirotsu, K.; Higuchi, T.; Fujita, K.;
Yamamura, H.; Imoto, T.; Tabushi, I. J. Chem. Soc., Perkin Trans. 2
1987, 7–14. (d) Harata, K.; Rao, C. T.; Pitha, J. Carbohydr. Res. 1993,
247, 83–98. (e) Mentzafos, D.; Terzis, A.; Coleman, A. W.; de Rango,
C. Carbohydr. Res. 1996, 282, 125–135. (f) Yoshida, N.; Harata, K.;
Inoue, T.; Ito, N.; Ichikawa, K. Supramol. Chem. 1998, 10, 63–67.
(13) For discussions and reviews on aqueous organocatalysis, see: (a)
Brogan, A. P.; Dickerson, T. J.; Janda, K. D. Angew. Chem., Int. Ed.
2006, 45, 8100–8102. (b) Hayashi, Y. Angew. Chem., Int. Ed. 2006,
45, 8103–8104. (c) Blackmond, D. G.; Armstrong, A.; Coombe, V.;
Wells, A. Angew. Chem., Int. Ed. 2007, 46, 3798–3800. (d) Narayan,
S.; Muldoon, J.; Finn, M. G.; Fokin, W.; Kolb, H. C.; Sharpless, K. B.
Angew. Chem., Int. Ed. 2005, 44, 3275–3279. (e) Mlynarski, J.;
Paradowska, J. Chem. Soc. ReV. 2008, 37, 1502–1511. (f) Paradowska,
J.; Stodulski, M.; Mlynarski, J. Angew. Chem., Int. Ed. 2009, 48, 4288–
4297. For several examples of aldol reaction catalyzed in buffer, see:
(g) Reymond, J. L.; Chen, Y. J. Org. Chem. 1995, 60, 6970–6979.
(h) Dickerson, T. J.; Janda, K. D. J. Am. Chem. Soc. 2002, 124, 3220–
3221. (i) Co´rdova, A.; Notz, W.; Barbas, C. F., III. Chem. Commun.
2002, 3024–3025. (j) Li, J.; Hu, S.; Luo, S. Z.; Cheng, J.-P. Eur. J.
Org. Chem. 2009, 132–140.
Figure 1. Crystal structure of CD-1.
columnar suprastructure, wherein the appended cyclohexadi-
amine group is intermolecularly encapsulated in the hydrophobic
cavity of an adjacent cyclodextrin Via the secondary rim (Figure
2). Similar assembly structures have also been observed in
several other 6-substituted ꢀ-cyclodextrins.12 Consistent with
previous reports, the hydrogen bonding together with hydro-
phobic interactions stabilizes the intermolecular self-assembly
and columnar structure in the crystalline state.
2. Direct Aldol Reaction Catalyzed by Cyclodextrin
Derivatives. With chiral supramolecular primary amine catalysts
in hand, we proceeded to test their catalytic ability in asymmetric
9
7218 J. AM. CHEM. SOC. VOL. 132, NO. 20, 2010

Selective Recognition and Asymmetric Catalysis
A R T I C L E S
Figure 2. Side view of the crystal packing in solid state along R-axis.
from S in catalysis with ent-8 to R with supramolecular CD-1.
Remarkably, the R selectivity was maintained in catalysis with
CD-1-CD-4, regardless of the configurations of the appended
diamines. These results suggest that the cyclodextrin cavity plays
a decisive role in controlling stereoselectivity. In addition, the
catalysis of CD-1 could be inhibited by the addition of
1-bromoadamantane, a known good guest for ꢀ-cyclodextrin
(Table 1, entry 13), further confirming that the cyclodextrin
cavity must have been involved in the catalytic coordinate.
It is observed that CD-1, which differs with CD-2 only on
the absolute configuration of the appended cyclohexadiamine,
demonstrated superior catalytic efficacy (Table 1, entry 1 vs 2,
3 times faster) and enantioselectivity (97% ee vs 68% ee). These
results together with the inefficiency of cis-cyclohexadiamine-
appended CD-3 suggest the match and mismatch of configura-
tions in this series of catalysts. In addition, the observation that
catalysts containing tertiary amine or secondary amine such as
CD-5 and CD-6 were totally inert in aqueous buffer further
proves the advantage of bifunctional primary amine catalysis
in CD-1.
Figure 3. Correlation of pH value with log kobs and ee of the model reaction
catalyzed by CD-1 at 25 °C in 50 mM acetate buffer (1.0 mL) containing
CD-1 (0.2 mM), p-nitrobenzaldehyde (4 mM), acetone 5% (v/v).
4.60-4.80 in terms of both activity and enantioselectivity.14 It
is noteworthy that no apparent dehydration and retro-aldol
products were observed under such conditions and the desired
aldol product was obtained with excellent enantioselectivity
(97% ee), representing the best results that could be achieved
in pure buffered aqueous media for this reaction.
3. Substrate Scope. The catalysis of CD-1 and CD-2 was
next examined with a variety of substrates. The results are
presented in Table 2 where some characteristics of the supramo-
lecular catalysis are evident. (1) The reactions not only tolerate
a range of aromatic aldehydes but also demonstrate interesting
features on recognition over different acceptors. For example,
while the reaction of 2-naphthaldehyde proceeded smoothly,
1-naphthaldehyde resulted in being a quite sluggish substrate
(Table 2, entry 6 vs 5), suggesting that the proper geometry
and orientation of an acceptor in the cyclodextrin cavity are
required for effective catalysis. (2) The catalysis is highly size
sensitive for aldol donors, and among a range of ketone donors
examined, only the small donors such as acetone, butanone, and
cyclopentanone are identified to be effective substrates for CD-
1. Note that the reaction of cyclohexanone is more than 46 times
slower than that of cyclopentanone caused by just one more
CH2 unit (Table 2, entry 11 vs 13). Interestingly, the diaste-
reoselectivity in the reaction of cyclopentanone is switched from
anti to syn by using CD-2 instead of CD-1, whereas the same
level of enantioselectivity is maintained for the major isomer
(Table 2, entry 11 vs 12). (3) The catalysis of CD-1 could be
applied under high concentration (up to 0.25 M) while maintain-
ing similar activity and enantioselectivity (Table 2, entries 1,
2, 6, 10-12; see Supporting Information Table S2 for details).
These features indicate an effective supramolecular approach
to asymmetric organocatalysis that works smoothly in aqueous
buffer with synthetic practicality.13g-j Under the latter condi-
tions, the reaction appears to be heterogeneous and thus vigorous
stirring should be applied to facilitate a smooth reaction.
The kinetics of the catalytic reaction by CD-1 in aqueous
acetate buffer was investigated next. The reaction was found to
follow Michaelis-Menten kinetics at pH 4.80 and 25 °C in the
presenceofexcessacetone(5%v/v),showingaLineweaver-Burk
plot with a Km ) 6.31 mM and a kcat ) 6.07 × 10-3 min-1
(Figure 4). To understand the nature of the catalysis, the
cyclodextrin analogues CD-2 to CD-6 and organocatalysts 7
to 10 were then examined in the model reaction (Table 1).
Though the small molecular catalysts such as primary-tertiary
diamine 7 and primary-secondary diamine 8 were able to
promote the model reactions as well in aqueous buffer, the
reactions generally showed low activity and poor stereoselec-
tivity (Table 1, entries 7-9). No further improvements were
observed when the small molecular diamines are endowed with
long alkyl chains such as in diamines 9 and 10 (Table 1, entry
10, 11) or with the use of ꢀ-cyclodextrin as an additive (Table
1, entry 12). In sharp contrast, covalently connecting a primary-
secondary diamine (e.g., ent-8) with ꢀ-cyclodextrin (i.e., CD-
1) led to a dramatically improved enantioselectivity and an over
6-fold rate acceleration compared with the parent ent-8 (Table
1, entry 1 vs 9). Note also that the stereoselectivity was switched
(14) At pH > 6.0, the reaction rate increases but with significantly reduced
enantioselectivity, indicating the existence of a general base catalyzed
by-pathway under this condition.
9
J. AM. CHEM. SOC. VOL. 132, NO. 20, 2010 7219

A R T I C L E S
Hu et al.
Figure 4. (a) Michaelis-Menten curves and (b) Lineweaver-Burk plot for CD-1 catalyzed model reaction at 25 °C in 50 mM acetate buffer (pH ) 4.80,
1.0 mL) containing CD-1 0.2 mM, acetone 5% (v/v).
Table 1. Asymmetric Supramolecular Catalysis of Direct Aldol
Table 2. Asymmetric Supramolecular Catalysis of Direct Aldol
Reactiona
Reactiona
b
kobs
Entry
R
R1
Cat.
ee (%)
drd
Entry
Cat.
krel
ee (%)
1
2
3
4
5
6
7
8
H
H
H
H
H
H
H
H
H
4-NO2Ph
Ph
4-MePh
CD-1 94.2 (91)c 97
---
1
2
3
4
5
6
7
8
CD-1
CD-2
CD-3
CD-4
CD-5
CD-6
7
6.45
2.06
0.22
0.68
0.054
0.038
0.66
1.00
1.00
0.20
1.23
1.03
0.76
97 (R)
68 (R)
54 (R)
22 (R)
n.d.e
CD-1 10.9 (67)c 91
---
---
---
---
---
---
---
---
CD-1 13.2
92
93
71
4-MeOPh CD-1 4.01
1-Naphth
2-Naphth
4-PhPh
piperonal
2-Naphth
4-NO2Ph
CD-1 0.62
CD-1 32.5 (61)c 97
n.d.e
CD-1 12.1
CD-1 4.02
CD-2 4.39
>99
9 (S)
94
90
8
20 (R)
20 (S)
13 (S)
31 (R)
20 (R)
63 (R)d
9
9
Ent-8
9
10
10
11
12
13
H, Me
CD-1 35.6 (75)c 90/99 53/47
10
11
12b
13c
-(CH2)2- 4-NO2Ph
-(CH2)2- 4-NO2Ph
-(CH2)3- 4-NO2Ph
CD-1 558 (98)c
CD-2 262 (94)c
CD-1 12.0
48/82 25/75
81/55 65/34
23/96 14/86
8
CD-1
a Conditions: catalyst 0.2 mM, aldehyde 4 mM, 25 °C, 50 mM
acetate buffer (pH ) 4.80, 1 mL), donor 10% (v/v). b Expressed in unit
of 10-5 mM/min. c Percent yield with 250 mM aldehyde in 12-48 h.
a Conditions: catalyst 0.2 mM, acetone 5% (v/v), aldehyde 4 mM, 25
°C, 50 mM acetate buffer (pH ) 4.80, 1 mL). b ꢀ-Cyclodextrin (0.2
mM) was added. c 1-Bromoadmantane (8 mM) was added as inhibitor.
d Dehydration product was observed. e The ee of the product is not
determined.
d
1
Syn/anti determined by HPLC or H NMR.
were observed when naphthaldehyde was treated with a solution
of CD catalyst. For example, fluorescence intensity was
decreased when CD-1 was added to a dilute solution of
2-naphthaldehyde solution in acetate buffer (Figure 5a). A Job
plot15 showed that 2-naphthaldeyde and CD-1 formed a 1:1
complex in an acetate buffer solution (Figure 5b).
The association constants of G1 to G9 with native ꢀ-CD and
our CD catalysts were then determined using fluorescence
spectroscopy, and the obtained association constants are sum-
marized in Table 3.16 Several binding features can be sum-
marized based on the data in Table 3: (1) The binding affinities
of CD-1 and CD-2 toward naphthaldehyde analogues such as
naphthenylmethanol G1 and naphthol G2 are generally de-
creased compared to those with native ꢀ-CD (Table 3, entries
4. Mechanism Studies. Detailed mechanism studies were
carried out next to gain more insights into this intriguing
asymmetric supramolecular catalysis. Accordingly, typical
techniques and methods in both supramolecular recognition
studies and in catalytic mechanism studies were utilized in
concert to explore the recognition properties, conformational
changes, and the possible catalytic cycles. CD-1, CD-2, and
(1-, and 2-) naphthaldehydes were selected as the representative
catalysts and substrates, respectively. When necessary, some
other tested substrates as shown in Table 2 were also included
in the study for comparison.
4.1. Recognition Behaviors of CD-1 and CD-2 toward
Substrates. The molecular recognition between catalyst and aldol
acceptor naphthaldehydes was probed with the aid of fluores-
cence and circular dichroism spectroscopy in acetate buffer (pH
) 4.80). A series of naphthalene derivatives such as naphtha-
lenylmethanol, naphthol, and naphthanoic acid were also
incorporated in this study. Pronounced fluorescence changes
(15) Job, P. Ann. Chim. 1928, 9, 113–203.
(16) The measurement of binding constants follows the method presented
in Connor, K. A. Binding Constants: The measurement of Molecular
Complex Stability; John Wiley & Sons: New York, 1987.
9
7220 J. AM. CHEM. SOC. VOL. 132, NO. 20, 2010

Selective Recognition and Asymmetric Catalysis
A R T I C L E S
Figure 5. (a) The family of fluorescence spectra of 0.02 mM of 2-naphthaldehyde aqueous solution at various concentrations of CD-1 (0 to 4.5 mM). (b)
The job’s plot of change in the fluorescence spectrum. [2-Naphthaldehyde] + [CD-1] ) 0.24 mM.
Table 3. Association Constants (Ka) for the Complexion of G1 to
G9 with Cyclodextrin Catalysta
when the hydroxymethylene group is changed to formyl (the
substrate in our catalysis) and carboxyl groups as in the cases
of G3 and G5 (Table 3, entries 7, 8, 13, 14). Compared to native
ꢀ-CD, the bindings of both CD-1 and CD-2 toward these guests
are significantly enhanced. Clearly, additional interactions
between the diamino group side chain and carbonyl group of
substrates, e.g. hydrogen bonding, electrostatic interactions, or
even covalent Schiff base formation, should contribute modestly
to the binding besides the hydrophobic effect. A similar effect
has also been observed in other amino-substituted cyclodextrins
when binding with carbonyl compounds.17 Regarding the
structural impact on binding, (3) it is noted that a stronger
binding affinity is generally observed with CD-1 than that with
CD-2 in recognizing G3 to G4, (4) whereas 2-substituted
naphthalenes are much better guest molecules than 1-substituted
naphthalenes for both CD-1 and CD-2. In both cases, the
observed binding phenomena are consistent with the respective
catalytic behaviors, serving as a reminder that stronger substrate
binding leads to better catalytic outcome. For example, with a
modest enhancement of binding with CD-1 over CD-2 (Table
3, entry 7 vs 8, Ka ) 403 L ·mol-1 vs 339 L ·mol-1), the reaction
of 2-naphthaldehyde (G3) catalyzed by CD-1 is 7 times faster
than that catalyzed by CD-2 (Table 2, entry 6 vs 9). In another
example, CD-1 binds 2-naphthaldehyde 7 times stronger than
1-naphthaldehyde (Table 3, entry 7 vs 10), corresponding to an
over 50-fold rate increase and a significant improvement of ee
(27%) for the reaction of the former over the latter (Table 2,
entry 6 vs 5).
Entry
Host
Guest
Ka (L·mol-1
)
1
2
3
4
5
6
7
8
CD-1
CD-2
ꢀ-CD
CD-1
CD-2
ꢀ-CD
CD-1
CD-2
ꢀ-CD
CD-1
CD-2
ꢀ-CD
CD-1
CD-2
ꢀ-CD
CD-1
CD-1
CD-1
CD-1
G1
G1
G1
G2
G2
G2
G3
G3
G3
G4
G4
G4
G5
G5
G5
G6
G7
G8
G9
45
36
389
148
182
395
403
339
226
60
9
10
11
12
13
14
15
16
17
18
19
n.d.b
n.d.b
1057
741
198
200
212
218
215
The association constants between CD-1 and each of the
enantiomers of aldol products (G6-G9) were also determined
(Table 3, entries 16-19).18 Interestingly, CD-1 demonstrates a
nearly identical binding affinity with association constants Ka
≈ 200 L ·mol-1 for all the four tested aldol products regardless
of their absolute configurations, indicating chiral recognition is
not occurring in our catalytic system. In addition, the lower
binding affinity of product (G6) than substrate (G3) suggests
that product inhibition would not happen under this context,
a Solvent: 50 mM acetate buffer (pH ) 4.80) containing 2% glycol.
b The binding affinities are bellowing the measurement limits of
fluorescence titration. See Figure S3a in the Supporting Information for
the representative fluorescence spectra under this situation.
(17) (a) Kean, S. D.; May, B. L.; Clements, P.; Lincoln, S. F.; Easton,
C. J. J. Chem. Soc., Perkin Trans. 2 1999, 1257–1264. (b) Nakamura,
A.; Inoue, Y. J. Am. Chem. Soc. 2005, 127, 5338–5339. (c) Ke, C.;
Yang, C.; Mori, T.; Wada, T.; Liu, Y.; Inoue, Y. Angew. Chem., Int.
Ed. 2009, 48, 6675–6677.
1, 2 vs 3, entries 4, 5 vs 6), suggesting some interferences of
the appended cyclohexadiamine moiety to binding in these cases.
Remarkably, (2) this interference is reversed to favorable binding
9
J. AM. CHEM. SOC. VOL. 132, NO. 20, 2010 7221

A R T I C L E S
Hu et al.
1
Figure 6. Assignment of H NMR of CD-1.
Figure 7. 1H NMR spectra of CD-1 in D2O with different concentrations. The red dot denotes the protons on cyclohexyl ring.
1
which may serve as the basis for catalytic turnovers in this chiral
supramolecular system.
served in the H NMR spectra of CD-1, particularly regarding
the proton signals on the cyclohexyl ring under different
concentrations (Figure 7). The DOSY spectrum of CD-1 under
2 mM and 20 mM indicated the existence of a dominant single
solution species with a similar diffusion coefficient under both
low and high concentration (see Figure S10 in the Supporting
Information), suggesting intermolecular assembly is not occur-
ring under high concentration. All together, these observations
rule out the possible intermolecular assembly of CD-1 in
solution. Therefore, CD-1 would mainly exist as self-included
conformations in pure water solution, which is consistent with
previous conformational studies on similar 6-substituted
cyclodextrins.19d,20 An ROESY spectrum of CD-1 provides
further support to this conclusion (Figure 8; for the full ROESY
spectrum see Figure S9 in the Supporting Information). As
shown in Figure 8a, all the protons of the cyclohexyl ring
correlate strongly with H5, the inner protons near the primary
rim of the cyclodextrin cavity, while the correlations with H3,
the inner protons near the secondary rim of cyclodextrin, are
relatively weak, illustrating the formation of intramolecular self-
assembly as depicted in the inside diagram (Figure 8a).
4.2. Conformational Analysis of CD Catalysts in Solution
and upon Binding with Substrates. 4.2.1. Conformation of CD-1
and CD-2 in Water and Acetate Buffer (pH 4.8). Unlike their
solid intermolecularly self-assembled structures, 6-monosub-
situted ꢀ-CD derivatives, particularly those with hydrophobic
6-substitutes, tend to form self-included structures in solution
phase as demonstrated by numerous previous studies.19 In our
case, the conformation of CD-1 in water was first analyzed using
1
1D and 2D NMR spectroscopy. The H signals of CD-1 and
CD-2 in D2O and acetate buffer (pH ) 4.80) were fully assigned
by the assistance of COSY, TOCSY, and HSQC spectrum
(Figure 6; for details see Figures S7, S8 in the Supporting
Information). No significant variations (<0.1 ppm) were ob-
(18) The CD-1-product complexes exhibit distinctive fluorescence behavior
(Figure S3b), indicating quite complicated binding processes beyond
simple 1:1 host-guest interactions like those in CD-1-substrate
systems. Apparent association constants (as listed in Table 3) were
determined based on the fluorescence change.
(19) (a) Park, K. K.; Kim, Y. S.; Lee, S. Y.; Song, H. E.; Park, J. W.
J. Chem. Soc., Perkin Trans. 2 2001, 2, 2114–2113. (b) Liu, Y; You,
C.-C.; Kunieda, M.; Nakamura, A.; Wada, T.; Inoue, Y. Supramol.
Chem. 2000, 12, 299–316. (c) Li, D.; Ng, S.-C.; Novak, I. Tetrahedron
Lett. 2002, 43, 1871–1875. (d) Park, J. W.; Lee, S. Y.; Song, H. J.;
Park, K. K. J. Org. Chem. 2005, 70, 9505–9513.
(20) McAlpine, S. R.; Garcia-Garibay, M. A. J. Am. Chem. Soc. 1998,
120, 4269–4275.
9
7222 J. AM. CHEM. SOC. VOL. 132, NO. 20, 2010

Selective Recognition and Asymmetric Catalysis
A R T I C L E S
Figure 8. ROESY spectrum and conformation of (a) CD-1 (0.02 M) in D2O; (b) CD-1 (0.02 M) in acetate buffer (pH ) 4.80); (c) CD-2 (0.02 M) in D2O;
and (d) CD-2 (0.02 M) in acetate buffer (pH ) 4.80).
Under catalytic conditions, i.e. acetate buffer (pH ) 4.80),
CD-1 would become protonated with an estimated pKa (CD-
1-H2+) of ∼5.70.21 As a result, H7-H12 signals on the
protonated (S,S)-cyclohexadiamine ring would be shifted lower
compared with the neutral CD-1. In addition, the protonation
of cyclohexadiamine would decrease its hydrophobicity and thus
cause exclusion of the suspended side chain from the cyclo-
dextrin cavity,17a leading to protons being shifted even lower
as shown in Figure 9. Indeed, the ROESY spectrum of CD-1
in acetate buffer shows only weak correlations between H8, H9,
and H5, a clear indication of the shallowly included (S,S)-
cyclohexadiamine group (Figure 8b).
The solution phase conformation of CD-2, which differs from
CD-1 in the absolute configuration of suspended cyclohexadi-
amine, has also been studied. The ROESY spectrum indicates
a similar conformation change as in the case of CD-1 in which
the suspended (R,R)-cyclohexadiamine of CD-2 is self-included
in the cavity in pure water solution but excluded from the cavity
in acidic acetate buffer (Figure 8c and 8d). Close analysis of
the ROESY spectrum of both CD-1 and CD-2 in acidic acetate
buffer reveals some delicate but noticeable differences. For
comparison, the correlation signals of H12a-H12e are selected
as references for the evaluation of relative intensities of cross
peaks, as these self-correlations of H12 protons would be much
less influenced by side chain configurations of CD-1 and CD-2
and are also relatively condition-independent.22 Accordingly,
there are much stronger correlations between H5, H8, and H9
(21) The pKa (CD-1-H2+) was estimated to be ∼5.70 according to May,
B. L.; Kean, S. D.; Easton, C. J.; Lincoln, S. F. J. Chem. Soc., Perkin
Trans. 1 1997, 3157–3160.
9
J. AM. CHEM. SOC. VOL. 132, NO. 20, 2010 7223

A R T I C L E S
Hu et al.
Figure 9. 1H NMR spectra of (a) CD-1 (0.02 M) in D2O and (b) CD-1 (0.02 M) in acetate buffer (pH ) 4.80).
in CD-1 than those in CD-2 under acidic acetate conditions
(see the area with red square in Figure 8b and 8d).22 These
observations are consistent with the model that the side chain
of CD-1 is shallowly included in the CD cavity, whereas the
side chain of CD-2 is likely suspended from the cavity. A
preliminary molecular modeling study also verifies this conclu-
sion.23 This model well rationalizes the observations that CD-2
exhibits a relatively poorer binding affinity toward guest G3
and G5 than CD-1 (Table 3, entry 8 vs 7, entry 14 vs 13) since
the possible noncovalent interactions between the suspended
diamino moiety and the encapsulated guest, which would
contribute in the binding, are geometrically less favorable in
this context. Previously, an absolute configuration difference
of side chains has been reported to cause a different self-
inclusion depth, which consequently leads to different binding
properties.24
4.2.2. Substrate Binding Mode and Conformation Change
Figure 10. Circular dichroism spectrum of 1-naphthaldehyde (0.04 mM,
black line) and 2-naphthaldehyde (0.04 mM, red line) in 50 mM acetate
buffer (pH ) 4.80) containing 10% acetonitrile and 2.0 mM CD-1.
of CD-1 upon Binding. The substrate binding mode of our
optimal catalyst CD-1 as well as the accompanying conforma-
tional changes was next probed. The hydrophobic interactions
provided by the cavity of cyclodextrin are the major driving
forces for the formation of a catalyst-substrate complex in
aqueous media, which has been extensively researched in the
supramolecular system.25 Moreover, the cavity of cyclodextrin
has a certain degree of rigidity because of the hydrogen-bonding
network among the hydroxyl groups on the rims. Thus, the
match in sizes between guests and hosts is also of great
importance in the course of recognition. With these factors
noted, some initial knowledge regarding the binding mode of
CD-1 is easily conceivable. For example, according to Harata’s
previous study,26 2-substituted naphthalene forms an axial
inclusion in the cyclodextrin cavity while 1-substituted naph-
thalene tends to form equatorial inclusion.26a,27 Therefore, in
our catalytic system 2-naphthaldehyde would axially bind with
CD-1 in a way that the carbonyl group is amenable to
noncovalent interaction with the appended chiral diamino group.
The observed 7-fold larger binding constant of 2-naphthaldehyde
over 1-naphthaldehyde is apparently in line with this binding
mode. The opposite ICD signals between 1-naphthaldehyde and
2-naphthaldehyde in the presence of CD-1 further proved the
different orientation of the naphthaldehydes in the cavity.
Moreover, the intenser ICD signal of 2-naphthaldehyde in CD-1
solution reveals that the 2-naphthaldehyde extended more deeply
via axial inclusion in the cavity of CD-1 than 1-naphthaldehyde,
which is bound in a much shallower equatorial orientation
(Figure 10). The ROESY spectrum was next used to elucidate
the detailed orientation of 2-naphthaldehyde in the cavity.
2-Naphthoic acid was chosen as an analogue for 2-naphthal-
dehyde because of its better solubility in aqueous solution and
similar interactions with the diamine group of CD-1. The intense
cross-peaks between H1′, H3′, and H4′ of 2-naphthoic acid and
cyclodextrin inner protons H3, H5 define a detailed binding
mode as depicted in Figure 11b. By analogue, 2-naphthaldehyde
and CD-1 would also form an axial inclusion with the aromatic
ring partially extending out from the secondary rim of the CD
cavity.
(22) The relative ROESY intensities of CD derivatives can be utilized to
determine conformation change in solution; for examples, see: (a)
Alderfer, J. L.; Eliseev, A. V. J. Org. Chem. 1997, 62, 8225–8226.
(b) Wenz, G.; Strassnig, C.; Thiele, C.; Engelke, A.; Morgenstern,
B.; Hegetschweiler, K. Chem.sEur. J. 2008, 14, 7202–7211. The
ROESY cross-peak between H12a and H12e is set to 100 in both CD-1
and CD-2. The ROESY intensity of selected relevant cross-peaks: CD-
1: 82 (H9a-H9e), 262 (H10-H11), 17 (H8-H5), 14 (H9-H5); CD-2: 88
(H9a-H9e), 266 (H10-H11), 6 (H8-H5), 6 (H9-H5).
(23) The molecular modeling was conducted using the CVFF force field
in the Insight II/Discover program package (Accelrys Inc.) according
to ref 19a. The energy minimum conformations of CD-1 and CD-2
in acidic buffer solution are in accordance with those proposed based
on ROESY spectra. See Figure S5 in the Supporting Information for
detailed data.
Meanwhile, some modest conformation changes of catalyst
CD-1 upon substrate binding have also been observed by
(24) For examples, see: (a) Ikeda, H.; Nakamura, M.; Ise, N.; Oguma, N.;
Nakamura, A.; Ikeda, T.; Toda, F.; Ueno, A. J. Am. Chem. Soc. 1996,
118, 10980–10988. (b) Ikeda, H.; Nakamura, M.; Ise, N.; Toda, F.;
Ueno, A. J. Org. Chem. 1997, 62, 1411–1418.
(26) (a) Harata, K.; Uedaira, H. Bull. Chem. Soc. Jpn. 1975, 48, 375–378.
(b) Shimizu, H.; Kaito, A.; Hatano, M. Bull. Chem. Soc. Jpn. 1979,
52, 2678–2684. (c) Shimizu, H.; Kaito, A.; Hatano, M. Bull. Chem.
Soc. Jpn. 1981, 54, 513–519.
(25) (a) Blokzijl, W.; Engberts, J. B. F. N. Angew. Chem., Int. Ed. 1993,
32, 1545–1579. (b) Otto, S.; Engberts, J. B. F. N. Org. Biomol. Chem.
2003, 1, 2809–2820.
(27) (a) Inoue, Y.; Hakushi, T.; Liu, Y.; Tong, L.-H.; Shen, B.-J.; Jin, D.-
S. J. Am. Chem. Soc. 1993, 115, 475–481. (b) Nishijo, J.; Ushiroda,
Y. Chem. Pharm. Bull. 1998, 46, 1790–1796.
9
7224 J. AM. CHEM. SOC. VOL. 132, NO. 20, 2010

Selective Recognition and Asymmetric Catalysis
A R T I C L E S
Figure 11. ROESY spectra of (a and b) CD-1 (0.02 M) and 2-naphthoic acid (0.02 M) in acetate buffer (pH ) 4.80, containing 2% d6-acetone). (c) CD-1
(0.02 M) and p-nitrobenzoic acid (0.02 M) in acetate buffer (pH ) 4.80, containing 2% d6-acetone).
Scheme 3. Proposed Catalytic Cycle
ROESY spectroscopy (Figure 11; also see Figure S9 in the
Supporting Information for full ROESY spectrum). Upon
treating CD-1 with 2-naphthoic acid in the presence of acetone,
analysis of ROESY spectrum (Figure 11a) suggests that the
cyclohexyl ring is likely pushed out by the incoming naphthalene
molecules as evidenced by much weakened NOE signals. A
similar phenomenon has also been observed when p-nitrobenzoic
acid was used as the binding substrate (Figure 11c). Considering
that both 2-naphthanoic acid and p-nitrobenzoic acid are
competitive inhibitors for the studied reactions,28 the proposed
binding complexes as depicted in Figure 11 may reflect the
ternary structure involved in the reaction coordinates.
proceeds through roughly three stages (Scheme 3): (1) assembly
of aldehyde and acetone donor in the CD-1 cavity and enamine
formation (I); (2) enamine addition to aldehyde, i.e. the C-C
forming step (II); and (3) release of product from the cavity
and regeneration of catalyst (III).
In a manner that closely parallels enzymatic catalysis, the
substrate binding properties of our supramolecular catalyst CD-1
have been unequivocally determined by separate binding studies
with aldol acceptors (Table 3) and the obtained results are in
good consistence with the well-known recognition behaviors
of cyclodextrin derivatives. In our catalytic studies, a connection
between the binding features with the corresponding catalytic
behavior seems self-evident, where substrates with a poor
binding ability normally resulted in low reactivity and stereo-
selectivity (Table 2). The major driving forces for the formation
of a host-substrate complex can be ascribed to the ubiquitous
hydrophobic interactions in cyclodextrin chemistry and nonco-
valent interactions of aldehyde with the appended diamino group
in CD-1 as proven in our substrate binding studies (Table 3).
A control reaction in DMSO has also been examined to prove
4.3. Proposed Catalytic Cycle. On the basis of the data
obtained from the binding studies and conformational analysis,
we proposed the following catalytic cycle for the CD-1 catalyzed
asymmetric direct aldol reaction in aqueous buffer. The reaction
(28) The aldol reactions of acetone and aromatic aldehydes (p-nitroben-
zaldehyde and 2-naphthaldehyde) were ∼2-fold slower when the
corresponding aromatic acids were added to the system, whereas the
enantioselectivity was maintained.
9
J. AM. CHEM. SOC. VOL. 132, NO. 20, 2010 7225

A R T I C L E S
Hu et al.
Figure 12. ESI-MS spectra of reaction mixture after treated with NaBH4.
Table 4. kobs of Deuteration of Aldol Donora
prove that the current supramolecular catalysis highly relies on
the cyclodextrin cavity. In this instance, whether the complex
of two substrates occurs sequentially or simultaneously may
seem to have no bearing on catalysis, but in the deuteration
experiments, we did observe that the deuteration of acetone is
accelerated slightly in the presence of aldehyde (Table 4, entry
2 vs entry 1), suggesting that the binding of aldehyde has a
direct impact on the enamine formation and pre-equilibrium
assembly of substrates in the CD cavity might be allosteric and
synergistic for subsequent catalysis.
Entry
Aldol donor
Cat
kobs (10-5 min-1
)
1
Acetone
Acetone
Acetone
Cyclohexanone
Cyclopentanone
CD-1
CD-1
----
CD-1
CD-1
3.49
5.30
0.28
118
2b
3
4
5
426
a Conditions: catalyst 0.02 M, aldol donor 0.1 M, 25 °C. Entries
1-3: 0.1 M acetate buffer (pH ) 4.80). Entries 4-5: 0.1 M acetate
buffer (pH ) 4.80) containing 10% d6-DSMO. b p-Nitrobenzaldehyde
was added to the system.
Direct experimental evidence for the enamine mechanism was
obtained from ESI-MS studies (Figure 12). The putative enamine
intermediate 4 could be in situ trapped as isopropylated CD-1
(2·H+ in Figure 12) by treating the reaction mixture with
NaBH4. In addition, the background reaction in acetate buffer
solution (pH ) 4.80) without CD-1 has also been examined
and no reaction was observed, excluding the possible enol
pathway.
the involvement of cavity binding in the catalysis as the use of
organic solvent was known to disfavor the encapsulation of a
guest in the cyclodextrin cavity, which probably would lead to
reaction inhibition. Indeed, the reaction was more than 16 times
slower in DMSO than in acetate buffer (kobs) 5.0 × 10-5 mM/
min vs 8.2 × 10-4 mM/min). Additional inhibition experiments
with aldehyde analogues such as p-nitrobenzoic acid and
2-naphthoic acid have also been examined, and only a very
modest inhibition effect was observed in these cases,28 indicating
these acids are competitive inhibitors. Significant inhibition is
only noticed with a much stronger cyclodextrin binder such as
2-bromoadamatane at high concentration (Table 1, entry 13).
These observations together with the generally low binding
constants suggest a highly dynamic and flexible guest binding
and exchange in the current supramolecular catalysis and that
a substrate binding pre-equilibrium precedes the rate-limiting
step of the reaction.
Regarding aldol donor recognition, a dramatic size effect has
been observed among a series of ketone donors examined. Only
small ketones such as acetone, butanone, and cyclopentanone
can be applied in CD-1 catalysis. Cyclohexanone, which is only
one carbon different from cyclopentanone, reacted over 50 times
slower in the presence of CD-1. With the observation that
cyclohexanone demonstrates usually comparable or even better
activity over cyclopentanone in asymmetric direct aldol reactions
catalyzed by small molecular catalysts,29 these results further
To characterize the rate-limiting step, the deuteration rate of
R-protons in acetone, which reflects closely the enamine
formation rate, was next determined by 1H NMR.30,31 In acetate
buffer, CD-1 was found to catalyze the deuteration 12 times
faster over the background exchange (Table 4, entry 1 vs 3).
ThedeuterationcatalyzedbyCD-1alsofollowedMichaelis-Menten
kinetics, showing a Lineweaver-Burk plot with a kcat ) 4.80
× 10-3 min-1. From comparison of the kcat of the overall
reaction (6.07 × 10-3 min-1), the lower deuteration rate suggests
the enamine-forming step might be, or at least partially be, the
rate-limiting step in the current catalysis. Similar slower
deuteration rates have also been observed with cyclohexanone
and cyclopentanone donors. Again, it was noted that the
deuteration of cyclohexanone was much slower (4-fold) than
cyclopentanone,31 further proving the cavity effect and the size
selection in the catalysis of CD-1 (Table 4, entry 4 vs 5).
(30) (a) Hine, J.; Mulders, J.; Houston, J. G.; Idoux, J. P. J. Org. Chem.
1967, 32, 2205–2209. (b) Warkentin, J.; Tee, O. S. J. Am. Chem. Soc.
1966, 88, 5540.
(31) In sharp contrast, the deuteration of cyclohexanone was ∼3 times faster
than cyclopentanone when catalyzed by antibody aldolase according
to Shulman, A.; Sitry, D.; Shulman, H; Keinan, E. Chem.sEur. J.
2002, 8, 229–239.
(29) In the presence of a small molecule catalyst such as 8, the reaction of
cyclohexanone is 3-fold faster than that of cyclopentanone. Details
are presented in Table S3 in the Supporting Information.
9
7226 J. AM. CHEM. SOC. VOL. 132, NO. 20, 2010

Selective Recognition and Asymmetric Catalysis
A R T I C L E S
Figure 13. Proposed transition state of (a) CD-1 and (b) CD-2 catalyzed direct aldol reaction.
Scheme 4. Primary Isotope Effect in the Aldol Reaction of Acetone
and p-Nitrobenzaldehyde
can be ascribed to two factors: (1) the slightly acidic condition,
which is overall beneficial for the current catalysis (Figure 3),
may turn out to retard the enamine formation; note that most
primary aminocatalysis in enzymes such as aldolase or decar-
boxylases occur favorably under neutral pH; (2) the assembly
of two substrates in the CD cavity, i.e. aldol donor and acceptor,
leading to proximity that facilitates C-C bond formation, results
in overwhelming the enamine-forming step.
Upon forming the C-C bond, subsequent hydrolysis and
release from the cavity (III) complete the catalytic cycle. The
release of product from the cavity would be a quite kinetically
facile process since we were unable to trap any product-CD-1
conjugate under the reductive conditions (Figure 11). This
observation, together with the generally low binding affinity of
aldol products (Table 3), suggests that aldol products are not
good guests for CD-1 due to their large sizes and improper
shapes, which well rationalizes the absence of product inhibi-
tions in our supramolecular system.
5. Origin of Stereoselectivity: Cavity and Noncovalent
Interactions. Experimentally, we have observed that (1) ꢀ-cy-
clodextrin supported diamines, regardless of chirality and
absolute configurations of the appended diamines, all gave
R-stereoselective products and (2) CD-1 with the configuration
matched side chain demonstrated superior stereocontrol to its
chirally unmatched analogue CD-2. These observations, com-
bined with the binding features as well as the conformation
analysis and molecular modeling, provide the basis for the
transition state as depicted in Figure 13.
A convincing experiment that supports the enamine-forming
step as the rate-limiting step is obtained from kinetic isotope
experiments (Scheme 4). The primary kinetic hydrogen isotope
effect was found in the aldol reaction of acetone and p-
nitrobenzaldehyde (KIE ) 6.38). Note that the primary KIE
was observed in the reaction with a different aldol acceptor
2-naphthaldehyde (KIE ) 4.79) and that with a different aldol
donor cyclopentanone (KIE ) 5.92). The obtained significant
primary KIE indicates that the rate-limiting step must involve
C-H bond formation or cleavage, which apparently occurs only
in the enamine-forming step in the aldol reactions.32 The
observation of a rate-limiting enamine-forming step stands in
contrast to the well-known primary aminocatalysis in natural
Type I aldolases, antibody aldolases, and organocatalyzed aldol
reaction, where the C-C forming step is assumed as the rate-
limiting step.6a,33 However, rate-limiting enamine formation is
not rare in enamine-based asymmetric catalysis. A rate-limiting
enamine step was recently identified in the classical proline
catalyzed Hajos-Parrish-Eder-Sauer-Wiechert reaction Via
theoretical calculation and 13C KIE studies.34 Accordingly, a
rate-limiting enamine formation followed by product-determin-
ing C-C formation can be formulated in the current catalysis
of CD-1. Presumably, the origin of a rate-limiting enamine step
In this model, the chirality of cyclodextrin is imparted to the
reactive enamine Via a C6-connected diamino moiety, leading
to differentiation of the Re- and Si-faces of the forming enamine,
wherein the Si-face is more accessible in the rigid cyclodextrin
cavity (Figure 13a). As summarized in Table 1, cyclodextrins
appended with chiral diamines such as CD-1, CD-2, and CD-3
all gave better (R)-enantioselectivity than CD-4 endowed with
an achiral diamine, a scenario where conformationally less
flexible diamines induce more bias toward Si-face attack due
to the constraint of the cyclodextrin cavity irrespective of the
absolute configurations of side chains. Note that the stereoin-
ductions with the supramolecular catalysts CD-1 and CD-2 are
in sharp contrast to their small molecular analogues such as
diamine 8, where (R)-8 and (S)-8 gave (R)- and (S)-selective
product, respectively. The uniformly observed (R)-selectivity
as well as the dramatically improved enantioselectivity pinpoints
the critical role of cyclodextrin cavity in channeling the
stereoselective reaction pathway.
(32) In this scenario, the varied reaction rates with different aldehydes (Table
2) suggest that the binding of aldehyde may have a significant impact on
the enamine formation. Our binding studies and the observation of the
accelerated enamine formation in the presence of p-nitrobenzaldehyde
(Table 4, entry 2) are clearly in line with this hypothesis. The elucidation
of this unique feature awaits further studies.
(33) (a) Heine, A.; Desantis, G.; Luz, J. G.; Mitchell, M.; Wong, C.-H.;
Wilson, I. A. Science 2001, 294, 369–374. (b) Bahmanyar, S.; Houk,
K. N. J. Am. Chem. Soc. 2001, 123, 11273–11283.
(34) Zhu, H.; Clemente, F. R.; Houk, K. N.; Meyer, M. P. J. Am. Chem.
Soc. 2009, 131, 1632–1633.
9
J. AM. CHEM. SOC. VOL. 132, NO. 20, 2010 7227

A R T I C L E S
Hu et al.
Figure 14. Size effect of aldehyde acceptors in the aldol reactions of acetone.
The inferior stereocontrol with CD-2 can also be explained
by considering a similar model (Figure 13b). In comparison with
CD-1, the re- and si-faces of the forming enamine becomes
less differentiated in the CD-2 cavity due to the excluded side
chains as well as the unfavorable orientation of the primary
amine moiety and hence both faces are accessible, causing low
enantioselectivity. Based on this model, the use of a larger
substrate would induce more bias toward si-face attack, leading
to better stereoselectivity when catalyzed by CD-2. The
observation of improved enantioselectivity with increasing
acceptor size is clearly in line with this hypothesis (Figure 14).
For example, by switching the acceptor from the smaller
benzaldehyde to the relatively larger 4-toluenaldehyde and
4-phenylbenzaldehyde, the enantioselectivity was increased from
45% to 95% ee (Figure 14). Since the electronic nature is
significantly varied in this series of aldehydes, the size effect
would be the major contributor in this context. Indeed, a good
linear correlation between Charton steric parameters and the
log of e.r. is observed in the CD-2 catalyzed system.35 On
the other hand, the changes in substrates barely influence the
stereoselectivity when it comes to CD-1, suggesting a favorable
cavity setting in balance with other interacting factors such as
noncovalent interactions with the side chain is reached in this
case.36
In addition to the chiral cavity, noncovalent interactions,
particularly those involving the protonated amino group, would
also play significant roles in the asymmetric catalysis by first
assisting in assembling the substrates in the cavity as we have
already discussed above, followed by stabilizing the transition
state, e.g. via hydrogen bonding with the carbonyl group, to
achieve better stereocontrol as shown in Figure 13. In this
regard, the unfavorable noncovalent interactions (Figure 13b)
with the suspended side chain in CD-2 would also account for
the generally low catalytic activity and enantioselectivity.
Though we were unable to directly characterize these nonco-
valent interactions under acetate buffer conditions, the pH
profiles of CD-1 catalysis are clearly in accordance with the
proposal that protonated amino groups are involved in the
noncovalent interactions as well-established in enamine-based
small molecular catalysis using chiral diamine-Bronsted acid
conjugates;7a-c,m,37 note that the stereoselectivity is totally
depleted when the pH was increased over 6.5.
Conclusion
In conclusion, asymmetric supramolecular primary amine
catalysts have been evolved by covalently connecting relevant
organocatalysts with a cyclodextrin host, illustrating a viable
approach for the development of asymmetric supramolecular
catalysts. CD-1 is found to be a remarkable asymmetric
supramolecular aldol catalyst that works effectively in aqueous
acetate buffer solution (pH ) 4.80) with high activity and
stereoselectivity. In a manner closely resembling enzymatic
catalysis, CD-1 selectively binds and situates substrates in the
reactive cavity by the synergistic action of a hydrophobic effect
and noncovalent interactions with a protonated diamino side
chain. The reaction proceeds through a rate-limiting enamine
formation step followed by product-determining C-C formation.
Hydrolytic release of the product from the cavity completes the
cycle, and no product inhibition has been observed. Unlike
enzymatic primary aminocatalysis, CD-1 works much more
favorably under slightly acidic conditions rather than enzymatic
neutral conditions, a distinctive feature that is also characteristic
of small molecular amino catalysts with protonated amino
groups as hydrogen-bonding catalytic motifs.7a-c,m,37 Col-
lectively, CD-1 represents a rare example that merges an
organocatalytic motif with supramolecular principles, leading
to an asymmetric supramolecular catalyst with synthetic ap-
plicability, meanwhile, providing insights into enzymatic catalysis.
Acknowledgment. This work was supported by the Natural
Science Foundation of China (NSFC 20702052 and 20972163), the
Ministry of Science and Technology (MoST) (2008CB617501,
2009ZX09501-018), and the Chinese Academy of Sciences. We
thank Prof. Judith Klinman and Prof. Qiyu Zheng for helpful
discussions.
(35) See Figure S6 in the Supporting Information for details. The value of
Charton steric parameters is according to Charton, M., Motoc, I., Eds.;
Steric Effects in Drug Design; Springer: Berlin, 1983; pp 68-75.
(36) Despite a favorable size effect on enantioselectivity in the catalysis
of CD-2, there are negative impacts on the reaction rate and the
catalytic rate difference between CD-1 and CD-2 for the same reaction
was found to increase linearly with the increasing size of the substrates,
suggesting that other interactions such as noncovalent interactions with
a protonated side chain amonium group should play critical roles for
effective catalysis besides the cavity effect.
Supporting Information Available: Detailed experimental
procedures and product characterization. This material is avail-
able free of charge via the Internet at http://pubs.acs.org.
JA102819G
(37) Saito, S.; Yamamoto, H. Acc. Chem. Res. 2004, 37, 570–579.
9
7228 J. AM. CHEM. SOC. VOL. 132, NO. 20, 2010