Experimental and theoretical studies on the hydrogen-bond-promoted enantioselective hetero-Diels-Alder reaction of Danishefsky's diene with benzaldehyde

Article abstract of DOI:10.1021/jo060129c

The enantioselective hetero-Diels-Alder (HDA) reaction of Danishefsky's diene with benzaldehyde has been achieved catalytically by a series of α,α,α′,α′-tetraaryl-1,3-dioxolane-4, 5-dimethanol (TADDOL) derivatives through hydrogen-bonding activation, affo

Full text of DOI:10.1021/jo060129c

Experimental and Theoretical Studies on the
Hydrogen-Bond-Promoted Enantioselective Hetero-Diels-Alder
Reaction of Danishefsky’s Diene with Benzaldehyde
Xue Zhang,† Haifeng Du,† Zheng Wang,† Yun-Dong Wu,*,‡ and Kuiling Ding*,†
State Key Laboratory of Organometallic Chemistry, Shanghai Institute of Organic Chemistry, Chinese
Academy of Sciences, 354 Fenglin Road, Shanghai 200032, China, and Department of Chemistry, The
Hong Kong UniVersity of Science and Technology, Clear Water Bay, Kowloon, Hong Kong, China
chydwu@ust.hk; kding@mail.sioc.ac.cn
ReceiVed January 20, 2006
The enantioselective hetero-Diels-Alder (HDA) reaction of Danishefsky’s diene with benzaldehyde has
been achieved catalytically by a series of R,R,R′,R′-tetraaryl-1,3-dioxolane-4,5-dimethanol (TADDOL)
derivatives through hydrogen-bonding activation, affording 2-phenyl-2,3-dihydro-4H-pyran-4-one after
trifluoroacetic acid workup in moderate yield and good enantioselectivity. The R,R′-aryl substituents in
the TADDOL molecules are found to exert a significant impact on both the activity and the
enantioselectivity of the catalysis. In combination with the experimental investigations, the mechanism
of the present reaction has also been studied theoretically using the ONIOM (B3LYP/6-31G*:PM3) method
with trans-1,3-dimethoxy-1,3-butadiene as the model for Danishefsky’s diene. In agreement with the
experimental findings, the calculation results indicate that this TADDOL-catalyzed HDA reaction proceeds
via a concerted mechanism through an asynchronous and zwitterionic transition structure (TS). The
carbonyl group of benzaldehyde is activated by forming an intermolecular hydrogen bond with one of
the hydroxy groups of TADDOL. Meanwhile, the intramolecular hydrogen bond between the two hydroxy
groups in TADDOL is found to facilitate the intermolecular hydrogen bonding with benzaldehyde. The
sense of asymmetric induction is well-rationalized by the analysis of the relative energies of TADDOL-
complexed TSs, while the different stereocontrol capabilities exhibited by TADDOLs in the reaction can
be qualitatively established on the basis of the structural features of their corresponding TSs.
Introduction
acid catalysts have been developed for the catalysis of this type
of reaction.2 Since the pioneering work by Rawal et al. in 2003,3
the enantioselective HDA reaction promoted by a hydrogen bond
using diol compounds as the chiral Bronsted acids has become
a topic of interest in asymmetric organocatalysis.4-6 In these
catalytic systems, the hydrogen-bonding interaction between the
hydroxyl group of the diol catalyst and the carbonyl group of
the substrate is considered to be the main driving force of the
catalysis. Here, the orientation of the R-aryl substituents in the
catalysts, such as TADDOL7a,8 (R,R,R′,R′-tetraaryl-1,3-dioxolan-
4,5-dimethanol), provides a chiral environment for the enantio-
selective control of the reaction. An interesting phenomenon
The enantioselective hetero-Diels-Alder (HDA) reaction of
carbonyl compounds with 1,3-dienes represents one of the most
important methods for the construction of optically active six-
membered oxo heterocycles with extensive applications in
natural or unnatural product synthesis.1 Various chiral Lewis
† Shanghai Institute of Organic Chemistry.
‡ The Hong Kong University of Science and Technology.
(1) (a) Jacobsen, E. N., Pfaltz, A., Yamamoto, H., Eds. ComprehensiVe
Asymmetric Catalysis; Springer: Berlin, 1999, Vol. I-III. (b) Yamamoto,
H., Ed. Lewis Acids in Organic Synthesis; Wiley-VCH: New York, 2001.
(c) Tietze, L. F.; Kettschau, G. In StereoselectiVe Heterocyclic Synthesis;
Metz, I. P., Ed.; Springer-Verlag: Berlin, 1997, Vol. 189.
10.1021/jo060129c CCC: $33.50 © 2006 American Chemical Society
Published on Web 03/08/2006
2862
J. Org. Chem. 2006, 71, 2862-2869

Hetero-Diels-Alder Reaction of Danishefsky’s Diene
is that naphth-1-yl-TADDOL (4b) exhibits remarkably superior
performance compared to that of its analogues, such as phenyl-
TADDOL (4a) or naphth-2-yl-TADDOL (4c), in several HDA
or DA reactions in terms of both activity and enantioselectivity
(eq 1-3). For example, Rawal demonstrated that TADDOL
(R,R)-4b is the best catalyst for the HDA reaction of diene 1a
with benzaldehyde (2a), affording (S)-3a in 97% yield with 96%
ee (eq 1).3 In the catalysis of the all-carbon Diels-Alder reaction
of diene 1a with methacrolein (2b), TADDOL (R,R)-4b (91%
ee, 83% yield) is much more efficient than (R,R)-4a (31% ee,
30% yield) or (R,R)-4c (45% ee, 33% yield) (eq 2).4c We found
that the HDA reaction of Brassard’s diene (1b) with 2a under
the catalysis of (R,R)-4b afforded (S)-3c in 67% yield with 83%
ee (eq 3).5b On the contrary, the use of (R,R)-4a or (R,R)-4c as
the catalyst for the same reaction only resulted in a low yield
of product 3c in almost racemic form.
In contrast to the overwhelming majority of effort devoted
to the synthetic aspects of HDA reactions of carbonyl com-
pounds with dienes, the number of theoretical and mechanistic
studies is limited.9 In a study, McCarrick et al. used ab initio
methods to model the reaction between formaldehyde and 1,3-
butadiene.9b-c While the uncatalyzed reaction was found to
(2) For reviews, see: (a) Danishefsky, S. J.; Deninno, M. P. Angew.
Chem., Int. Ed. Engl. 1987, 26, 15. (b) Jørgensen, K. A. Angew. Chem.,
Int. Ed. 2000, 39, 3558. (c) Maruoka, K. In Catalysis Asymmetric Synthesis,
2nd ed.; Ojima, I., Ed.; Wiley-VCH: New York, 2000, Chapter 8A. (d)
Jørgensen, K. A. In Cycloaddition Reactions in Organic Synthesis;
Kobayashi, S., Jørgensen, K. A., Eds.; Wiley-VCH: Weinheim, Germany,
2002, pp 151-185. For more recent examples, see: (e) Fan, Q.; Lin, L.;
Liu, J.; Huang, Y.; Feng, X.; Zhang, G. Org. Lett. 2004, 6, 2185. (f) Fu,
Z.; Gao, B.; Yu, Z.; Yu, L.; Huang, Y.; Feng, X.; Zhang, G. Synlett 2004,
1772. (g) Yamashita, Y.; Saito, S.; Ishitani, H.; Kobayashi, S. J. Am. Chem.
Soc. 2003, 125, 3793. (h) Du, H.; Long, J.; Hu, J.; Li, X.; Ding, K. Org.
Lett. 2002, 4, 4349. (i) Wang, B.; Feng, X.; Huang, Y.; Liu, H.; Cui, X.;
Jiang, Y. J. Org. Chem. 2002, 67, 2175. (j) Long, J.; Hu, J.; Shen, X.; Ji,
B.; Ding, K. J. Am. Chem. Soc. 2002, 124, 10. (k) Kii, S.; Hashimoto, T.;
Maruoka, K. Synlett 2002, 931. (l) Kezuka, S.; Mita, T.; Ohtsuki, N.; Ikeno,
T.; Yamada, T. Bull. Chem. Soc. Jpn. 2001, 74, 1333. (m) Liu, P.; Jacobsen,
E. N. J. Am. Chem. Soc. 2001, 123, 10772. (n) Thompson, C. F.; Jamison,
T. F.; Jacobsen, E. N. J. Am. Chem. Soc. 2001, 123, 9974. (o) Doyle, M.
P.; Phillips, I. M.; Hu, W. J. Am. Chem. Soc. 2001, 123, 5366-5367. (p)
Kezuka, S.; Mita, T.; Ohtsuki, N.; Ikeno, T.; Yamada, T. Chem. Lett. 2000,
824. (q) Evans, D. A.; Johnson, J. S.; Olhava, E. J. J. Am. Chem. Soc.
2000, 122, 1635. (r) Dossetter, A. G.; Jamison, T. F.; Jacobsen, E. N. Angew.
Chem., Int. Ed. 1999, 38, 2398. (s) Wang, B.; Feng, X.; Cui, X.; Liu, H.;
Jiang, Y. Chem. Commun. 2000, 1605. (t) Du, H.; Ding, K. Org. Lett. 2003,
5, 1091. (u) Du, H.; Zhang, X.; Wang, Z.; Ding, K. Tetrahedron 2005, 61,
9465. (v) Yuan, Y.; Long, J.; Sun, J.; Ding, K. Chem.sEur. J. 2002, 8,
5033. (w) Yuan, Y.; Li, X.; Sun, J.; Ding, K. J. Am. Chem. Soc. 2002, 124,
14866. (x) Doyle, M. P.; Valenzuela, M.; Huang, P. Proc. Natl. Acad. Sci.
U.S.A. 2004, 101, 5391. (y) Gademann, K.; Chavez, D. E.; Jacobsen, E. N.
Angew. Chem., Int. Ed. 2002, 41, 3059.
proceed via a concerted pathway with an unsymmetrical
transition structure (TS), coordination of the Lewis acid BH3
to formaldehyde not only makes it much more reactive, but also
increases the asynchronicity and charge separation (zwitterionic
character) in the TS, making a stepwise mechanism favored in
solution. Later, Jørgensen and co-workers reported a semiem-
pirical AM1 study on the mechanism of the HDA reaction of
(6) For reviews on organocatalysts, see: (a) Dalko, P. I.; Moisan, L.
Angew. Chem., Int. Ed. 2001, 40, 3726. (b) Dalko, P. I.; Moisan, L. Angew.
Chem., Int. Ed. 2004, 43, 5138. (c) Houk, K. N., List, B., Eds. Acc. Chem.
Res. 2004, 37, 487. (d) List, B., Bolm, C., Eds. AdV. Synth. Catal. 2004,
346, 1021-1022 (special issue on organocatalysis). (e) Seayad, J.; List, B.
Org. Biomol. Chem. 2005, 3, 719. (f) Berkessel, A., Groger, H., Eds.
Asymmetric Organocatalysis; Wiley-VCH: Weinheim, 2005. For more
recent representative papers, see: (g) Wenzel, A. G.; Jacobsen, E. N. J.
Am. Chem. Soc. 2002, 124, 12964. (h) McDougal, N. T.; Schaus, S. E. J.
Am. Chem. Soc. 2003, 125, 12094. (i) Okino, T.; Hoashi, Y.; Takemoto,
Y. J. Am. Chem. Soc. 2003, 125, 12672. (j) Nugent, B. M.; Yoder, R. A.;
Johnston, J. N. J. Am. Chem. Soc. 2004, 126, 3418. (k) McDougal, N. T.;
Trevellini, W. L.; Rodgen, S. A.; Kliman, L. T.; Schaus, S. E. AdV. Synth.
Catal. 2004, 346, 1231. (l) Uraguchi, D.; Terada, M. J. Am. Chem. Soc.
2004, 126, 5356. (m) Uraguchi, D.; Sorimachi, K.; Terada, M. J. Am. Chem.
Soc. 2004, 126, 11804. (n) Okino, T.; Hoashi, Y.; Furukawa, T.; Xu, X.;
Takemooto, Y. J. Am. Chem. Soc. 2005, 127, 119. (o) Shi, M.; Chen, L.
H.; Li, C. Q. J. Am. Chem. Soc. 2005, 127, 3790. (p) Fuerst, D. E.; Jacobsen,
E. N. J. Am. Chem. Soc. 2005, 127, 8964. (q) Rueping, M.; Sugiono, E.;
Azap, C.; Theissmann, T.; Bolte, M. Org. Lett. 2005, 7, 3781. (r) Akiyama,
T.; Itoh, J.; Yokota, K.; Fuchibe, K. Angew. Chem., Int. Ed. 2004, 43, 1566.
(s) Joly, G. D.; Jacobsen, E. N. J. Am. Chem. Soc. 2004, 126, 4102. (t)
Taylor, M. S.; Jacobsen, E. N. J. Am. Chem. Soc. 2004, 126, 10558. (u)
Yoon, T. P.; Jacobsen, E. N. Angew. Chem., Int. Ed. 2005, 44, 466. (v)
Berkessel, A.; Cleemann, F.; Mukherjee, S.; Muller, T. N.; Lex, J. Angew.
Chem., Int. Ed. 2005, 44, 807. (w) Taylor, M. S.; Tokunaga, N.; Jacobsen,
E. N. Angew. Chem., Int. Ed. 2005, 44, 6700. (x) Akiyama, T.; Saitoh, Y.;
Morita, H.; Fuchibe, K. AdV. Synth. Catal. 2005, 347, 1523. (y) Hoffmann,
S.; Seayad, A. M.; List, B. Angew. Chem., Int. Ed. 2005, 44, 7424-7427.
(7) For a comprehensive review on TADDOL chemistry, see: (a)
Seebach, D.; Beck, A. K.; Heckel, A. Angew. Chem., Int. Ed. 2001, 40, 92.
For the preparation of TADDOL derivatives, see: (b) Beck, A. K.; Bastani,
B.; Plattner, D. A.; Petter, W.; Seebach, D.; Braunschweiger, H.; Gysi, P.;
La Vecchia, L. Chimia 1991, 45, 238.
(3) Huang, Y.; Unni, A. K.; Thadani, A. N.; Rawal, V. H. Nature 2003,
424, 146.
(4) For reviews on Brønsted acid catalysis, see: (a) Schreiner, P. R.
Chem. Soc. ReV. 2003, 32, 289. (b) Bolm, C.; Rantanen, T.; Schiffers, I.;
Zani, L. Angew. Chem., Int. Ed. 2005, 44, 1758. For recent examples, see:
(c) Thadani, A. N.; Stankovic, A. R.; Rawal, V. H. Proc. Natl. Acad. Sci.
U.S.A. 2004, 101, 5846. (d) Momiyama, N.; Yamamoto, H. J. Am. Chem.
Soc. 2005, 127, 1080. (e) Uraguchi, D.; Sorimachi, K.; Terada, M. J. Am.
Chem. Soc. 2005, 127, 9360. (f) Akiyama, T.; Morita, H.; Itoh, J.; Fuchibe,
K. Org. Lett. 2005, 7, 2583. (g) Tang, Z.; Yang, Z.-H.; Chen, X.-H.; Cun,
L.-F.; Mi, A.-Q.; Jiang, Y.-Z.; Gong, L.-Z. J. Am. Chem. Soc. 2005, 127,
9285. (h) Tang, Z.; Jiang, F.; Yu, L.-T.; Cui, X.; Gong, L.-Z.; Mi, A.-Q.;
Jiang, Y.-Z.; Wu, Y.-D. J. Am. Chem. Soc. 2003, 125, 5262. (i) Tang, Z.;
Jiang, F.; Cui, X.; Gong, L.-Z.; Mi, A.-Q.; Jiang, Y.-Z.; Wu, Y.-D. Proc.
Natl. Acad. Sci. U.S.A. 2004, 101, 5755.
(5) For a highlight on the activation of carbonyl compounds by double-
hydrogen-bonding catalysts, see: (a) Pihko, P. M. Angew. Chem., Int. Ed.
2004, 43, 2062. For recent examples of chiral hydrogen-bonding catalyst-
mediated HDA reactions between 1,3-diene and carbonyl compounds, see:
(b) Du, H.; Zhao, D.; Ding, K. Chem.sEur. J. 2004, 10, 5964. (c) Unni,
A. K.; Takenaka, N.; Yamamoto, H.; Rawal, V. H. J. Am. Chem. Soc. 2005,
127, 1336. (d) Rajaram, S.; Sigman, M. S. Org. Lett. 2005, 7, 5473.
(8) For discussions on solid-state photoreactions of TADDOL-included
carbonyl compounds, see: (a) Toda, F. Pure Appl. Chem. 2001, 73, 1137.
(b) Tanaka, K.; Toda, F. Chem. ReV. 2000, 100, 1025.
J. Org. Chem, Vol. 71, No. 7, 2006 2863

Zhang et al.
SCHEME 1. Enantioselective HDA Reaction of
Danishefsky’s Diene (1c) with Benzaldehyde (2a) Catalyzed
by TADDOL Derivatives
TADDOL-catalyst loading of 20 mol %, followed by workup
with trifluoroacetic acid (TFA). As shown in Scheme 1, under
otherwise identical conditions, the HDA product 2-phenyl-2,3-
dihydro-4H-pyran-4-one (3a) was obtained in varying yields
and ee values, depending on the structures of the TADDOLs
used. The reaction of Danishefsky’s diene (1c, 1.0 mmol) with
benzaldehyde (2a, 0.5 mmol) in the presence of 20 mol % of
4a proceeded to give 2-phenyl-2,3-dihydro-4H-pyran-4-one in
18% yield with 12.1% ee. Remarkably, all the TADDOL
derivatives with 1-naphthyl substituents were superior to those
having 2-naphthyl or phenyl rings. Moreover, the change of
the R group in the backbone of TADDOL from methyl (4b) to
1,4-butylene (4d) or 1,5-pentylene (4f) affected the enantio-
selectivity slightly. Among the three 1-naphthyl-substituted
catalysts 4b, 4d, and 4f, 4b turns out to be the best one in terms
of both enantioselectivity and reactivity.
benzaldehyde with Danishefsky’s diene catalyzed by aluminum
complexes.9e The results indicated that a concerted pathway is
most likely in the absence of a catalyst, while a Makaiyama-
aldol two-step pathway operates the reaction catalyzed by
(MeO)2AlMe. Intrigued by Rawal’s experimental observation
that hydrogen-bonding solvents such as chloroform or alcohols
can greatly enhance the HDA reaction between 1a and ketonic
substrates, Domingo and Andres recently studied the model
HDA reaction between N,N-dimethyl-1-amino-3-methoxy-1,3-
butadiene and acetone promoted by hydrogen-bonding inter-
action(s) with the solvent (chloroform) molecule(s) using density
functional theory (DFT) at the B3LYP/6-31G* level.9g The
results suggested that the feasibility of Rawal’s cycloadditions
is not only a consequence of the increase of the electrophilicity
of the ketone by the explicit solvation, but also attributable to
the strong nucleophilic character of the substituted butadiene
1a.
To clarify the effect of the hydrogen-bonding interaction and
the impact of aryl groups in the TADDOL-catalyzed enanti-
oselective HDA reaction of diene with carbonyl compounds and
to gain a deeper understanding of the mechanistic aspects of
the catalysis, herein we report our results on the experimental
and computational studies of the hydrogen-bond-promoted
enantioselective HDA reaction of Danishefsky’s diene (1c) with
benzaldehyde (2a; Scheme 1). Although this reaction represents
one of the most important C-C bond-forming reactions in
asymmetric catalysis,2 it has not yet been furnished by hydrogen-
bonding activation.
Two possible pathways for the enantioselective addition of
Danishefsky’s diene to aldehydes have been reported: Mu-
kaiyama aldol condensation versus concerted [4+2] cyclo-
addition.2d Presently, the isolation and characterization of the
reaction intermediate would provide useful information in this
respect (Scheme 2). The 1H NMR spectral identification of the
isolated primary product formed by the reaction of 1c with 2a
in the presence of 4b prior to CF3COOH treatment suggests
that the [4+2] cycloadduct is formed exclusively (see Figure
S1 in Supporting Information), supporting a concerted cyclo-
addition mechanism (Scheme 2).
Computational Methods: For the present system, a complete
DFT10 analysis would be impractical, owing to the very bulky
structural elements, such as the R,R,R′,R′-tetraaryl groups of
the TADDOL catalysts. On the other hand, the use of semiem-
pirical approaches would very likely be flawed by an imperfect
description of the critical parameters in the TSs, which may
result in a poor estimation of the effect of structural elements
in the catalyst on the activity and enantioselectivity. The
ONIOM (B3LYP/6-31G*:PM3) method11 appears to be the best
choice for the present case: the critical parts of the reacting
system can be relatively accurately described using DFT
calculation at the B3LYP/6-31G*12,13 level, while the effects
of the different aryl substituents of the TADDOLs can be
accounted for by a semiempirical PM314 treatment. To further
reduce the computational cost, we utilized the trans-1,3-
dimethoxy-1,3-butadiene (5; Figure 1) as the working model
of Danishefsky’s diene (1c). The energies of all stationary points
(including the B3LYP/6-31G*:PM3 optimized geometries) were
calculated by DFT at the B3LYP/6-31G* level.
Results and Discussion
Experimental Studies on the TADDOL-Catalyzed Enan-
tioselective HDA Reaction of Danishefsky’s Diene with
Benzaldehyde. A variety of R,R,R′,R′-tetraaryl-substituted
TADDOL derivatives (4a-g, Scheme 1), synthesized according
to the published procedure,5b were tested for their catalytic
efficiency in the HDA reaction of Danishefsky’s diene (1c) with
benzaldehyde (2a). Typically, the reactions were run at room
temperature for 72 h under solvent-free conditions with a
The calculations were performed using the Gaussian 98
program.15 The two geometry layers of the ONIOM calculation
are denoted as the high layer and the low layer, as depicted in
Figure 1, where the high layer is the level on the core layer and
the low layer is the level on the whole system. The core layer
contains the reactive part of the catalytic system: benzaldehyde,
the model diene (5) and trans-4,5-bis(hydroxymethyl)-1,3-
(10) (a) Parr, R. G.; Pearson, R. G. J. Am. Chem. Soc. 1983, 105, 7512.
(b) Parr, R. G.; Yang, W. Density Functional Theory of Atoms and
Molecules; Oxford University Press: New York, 1989.
(11) (a) Maseras, F.; Morokuma, K. J. Comput. Chem. 1995, 16, 1170.
(b) Svensson, M.; Humbel, S.; Froese, R. D. J.; Matsubara, T.; Sieber, S.;
Morokuma, K. J. Phys. Chem. 1996, 100, 19357.
(9) For examples of theoretical studies on HDA reactions, see: (a) Tietze,
L. E.; Fennen, J.; Anders, E. Angew. Chem., Int. Ed. 1989, 28, 1371. (b)
McCarrick, M. A.; Wu, Y. D.; Houk, K. N. J. Am. Chem. Soc. 1992, 114,
1499. (c) McCarrick, M. A.; Wu, Y. D.; Houk, K. N. J. Org. Chem. 1993,
58, 3330. (d) Yao, S.; Johannsen, M.; Jørgensen, K. A. J. Chem. Soc., Perkin
Trans. 1 1997, 16, 2345. (e) Roberson, M. I.; Jepsen, A. S.; Jørgensen, K.
A. Tetrahedron 2001, 57, 907. (f) Ujaque, G.; Lee, P. S.; Houk, K. N.;
Hentemann, M. F.; Danishefsky, S. J. Chem.sEur. J. 2002, 8, 3423. (g)
Domingo, L. R.; Andres, J. J. Org. Chem. 2003, 68, 8662. (h) Iwakura, I.;
Ikeno, T.; Yamada, T. Angew. Chem., Int. Ed. 2005, 44, 2524.
(12) Becke, A. D. Phys. ReV. A 1988, 38, 3098.
(13) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. A: At., Mol., Opt. Phys.
1988, 37, 785.
(14) (a) Stewart, J. J. P. J. Comput. Chem. 1989, 10, 209. (b) Stewart,
J. J. P. J. Comput. Chem. 1989, 10, 221.
2864 J. Org. Chem., Vol. 71, No. 7, 2006

Hetero-Diels-Alder Reaction of Danishefsky’s Diene
SCHEME 2. Mukaiyama Aldol vs Concerted [4+2] Cycloaddition Pathways
dioxacyclopentane (with hydrogen atoms replacing four aryl
rings and two methyl groups of the TADDOL molecule). For
this high layer, we used DFT with Becke’s three-parameter
hybrid functional (B3),12 and the Lee, Yang, and Parr correlation
functional (LYP).13 The low layer is modeled at the semiem-
pirical PM314 level. In the calculations, the TSs were character-
ized by vibration frequency calculations that showed one and
only one imaginary frequency corresponding to motions along
the reaction coordinate under consideration. Single-point cal-
culations at the B3LYP/6-31G* level were done on the
geometries located at the ONIOM level (or, in the cases of
uncatalyzed reactions, pure DFT). All atomic charges were from
the natural population analysis (NPA).16
well-known structural features of TADDOL molecules.5b,7a,8
Accordingly, coordination pattern C can be safely excluded.
For coordination modes A and B, it was found that structure
optimization calculations starting from the initial structure A
or B always end up with the mode A configuration, thus only
the trans mode A is feasible. Remarkably, A is also consistent
with many observed crystal structures of analogous TADDOL-
carbonyl compound adducts.5b,7a,8
FIGURE 2. Possible intermolecular hydrogen-bonding patterns be-
tween benzaldehyde and TADDOL.
The X-ray crystal structures18 of TADDOLs 4a, 4b, and 4c
were taken as the starting geometries for all the calculations
involving these compounds, respectively, and the hydrogen-
bonded molecular complex between 4a and DMF5b was taken
as the model of the interaction pattern between catalyst and
substrate by replacing DMF with benzaldehyde. For comparison,
four computational models were constructed, as shown in
Scheme 3. Model I corresponds to the uncatalyzed HDA reaction
between diene 5 and benzaldehyde 2a. The influence of the
hydrogen bonding on the reactions in the presence of different
TADDOL catalysts, including 4a, 4b, or 4c, were studied by
means of models II, III, and IV, respectively.
Uncatalyzed Reaction: To rationalize the enantioselectivity
of the TADDOL-catalyzed HDA reaction between Danishef-
sky’s diene with benzaldehyde, eight possible diastereomeric
TSs resulting from different regio-, exo-, endo-, and enantio-
selectivities should in principle be considered for extensive
analysis. For the uncatalyzed reaction, no enantioselectivity
needs to be considered, and the cycloaddition reaction between
the model diene 5 and the benzaldehyde 2a can take place along
two regio-isomeric channels: the ortho and the meta (Model I
shown in Scheme 3). The ortho channel corresponds to the C1-
FIGURE 1. The two geometry layers of the reaction system in the
ONIOM calculation.
Because the carbonyl group of benzaldehyde possesses two
possible hydrogen-bond-acceptor sites (lone pair cis or trans to
the phenyl group) and each TADDOL molecule has two
potential hydrogen-bond-donor sites, the benzaldehyde may
interact with either one or both of its lone pairs to the
TADDOL’s hydroxyl proton through single or double hydrogen
bonding (A, B, or C, respectively, in Figure 2). Although
double-hydrogen-bonding activation has been used as an ef-
ficient tool for the catalysis of the reactions of carbonyl
compounds,5a,17 the existence of an intramolecular hydrogen
bond between the two hydroxyl groups has been one of the
(15) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.;
Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A.
D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi,
M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.;
Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick,
D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.;
Ortiz, J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi,
I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.;
Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M.
W.; Johnson, B. G.; Chen, W.; Wong, M. W.; Andres, J. L.; Head-Gordon,
M.; Replogle, E. S.; Pople, J. A. Gaussian 98, revision A.9; Gaussian,
Inc.: Pittsburgh, PA, 1998.
(17) For examples of the activation of carbonyl compounds by double
hydrogen bonding, see: (a) Hine, J.; Linden, S.-M.; Kanagasabapathy, V.
M. J. Org. Chem. 1985, 50, 5096. (b) Kelly, T. R.; Mechani, P.; Ekkundi,
V. S. Tetrahedron Lett. 1990, 31, 3381.
(18) The crystal structures corresponding to CCDC-236876, CSD refer-
ence codes YONVEC and YONVAY, are used for the starting geometries
of 4a, 4b, and 4c, respectively.
(16) Reed, A. E.; Weinstock, R. B.; Weinhold, F. J. Chem. Phys. 1985,
83, 735.
J. Org. Chem, Vol. 71, No. 7, 2006 2865

Zhang et al.
FIGURE 3. B3LYP/6-31G* optimized TSs and the activation energies (∆E, kcal/mol) corresponding to the two channels (ortho and meta) for the
uncatalyzed HDA reaction between diene 5 and benzaldehyde 2a. The bond lengths involved in the TSs are given in angstroms.
SCHEME 3. Computational Models Used for
TADDOL-Catalyzed HDA Reactions between the Model
Butadiene 5 and Benzaldehyde 2a
Mukaiyama aldol type pathway could be located. The energy
of the located TS for the meta route is 14 kcal/mol higher than
that of the ortho channel, which is consistent with frontier
molecular orbital theoretical expectation.19 Therefore, the meta
channels associated with the cycloaddition in models II, III, and
IV can be excluded. For the ortho channel of Model I, the TS-
endo is calculated to be about 0.8 kcal/mol more stable than
that of the TS-exo.20 For the slightly favorable TS, TS-endo,
the lengths of the forming C-C and C-O bonds are 1.884 and
2.232 Å, respectively, indicating the asynchronous nature of
the bond-forming process.
TADDOL-Catalyzed Reactions: For the reactions between
diene 5 and benzaldehyde-TADDOL adducts 2a‚‚‚4a-c (Mod-
els II, III, and IV), the attempts to find the zwitterionic aldol-
like intermediates again proved to be unsuccessful, thus
excluding the Mukaiyama stepwise pathway for the present
reaction. This is in agreement with experimental observations.
The TSs for a concerted reaction pathway leading to the HDA
adducts 6‚‚‚4a-c were located. Their corresponding TSs are
denoted as TS-(re)-4a/TS-(si)-4a (Model II), TS-(re)-4b/TS-
(si)-4b (Model III), and TS-(re)-4c/TS-(si)-4c (Model IV),
respectively, depending on the direction of the suprafacial attack
of diene 5 to the re or si face of the hydrogen-bonded
benzaldehyde (2a‚‚‚4a-c; Scheme 3 and Figure 4).
O6 and C4-C5 bond-forming process, whereas the meta
channel corresponds to the C1-C5 and C4-O6 bond forma-
tions. For the two regioisomers, each has an exo and endo
approach. Therefore, four potentially viable TSs (TS-endo and
TS-exo for the ortho channel; TS′-endo and TS′-exo for the
meta channel) need to be considered for Model I.
Consistent with the calculation results reported previously
by Houk et al.,9b,c Jørgensen et al.,9e and Domingo and Andres9g
on analogous model reactions, DFT calculations on Model I at
the B3LYP/6-31G* level indicated that the uncatalyzed reaction
proceeds via a concerted but asynchronous pathway (Figure 3),
and no zwitterionic intermediate or TS corresponding to the
Unlike the uncatalyzed counterpart, the TADDOL 4b-
catalyzed HDA reaction between 5 and 2a exhibited a clear
preference for the endo over the exo approach. The calculated
relative energies for the exo TSs TS-exo-(si)-4b/TS-exo-(re)-
4b (16.3 and 13.5 kcal/mol, respectively) are significantly
(19) Houk, K. N.; Strozier, R. W. J. Am. Chem. Soc. 1973, 95, 4094.
(20) It is difficult to predict the endo/exo preference of the reaction on
the basis of such a small energy difference.
2866 J. Org. Chem., Vol. 71, No. 7, 2006

Hetero-Diels-Alder Reaction of Danishefsky’s Diene
comparison with that of the uncatalyzed reaction (20.2 kcal/
mol, see Figure 3) in terms of their preferential TSs (endo
approach, si-face attack). Although the reductions in the
activation energy by the hydrogen-bonding activation of the
benzaldehyde may be somewhat overestimated due to the
negligence of solvent effects, it can be qualitatively concluded
that the reactions are accelerated by the catalysts and the
catalytic efficiency is in the order 4b > 4c > 4a. In each case,
the si-TS is found to be more stable than the competing re-TS,
thus, correctly accounting for the observed (S) sense of
asymmetric induction in the experiments. A comparison among
the values of the si-TS and re-TS energy differences for TS-4b
(2.6 kcal/mol), TS-4a (1.8 kcal/mol), and TS-4c (1.6 kcal/mol)
can predict the relative trend of the stereocontrol capability of
TADDOL 4a, 4b, and 4c; that is, 4b should be superior to both
4a and 4c, while the latter two remain at essentially the same
level. Indeed, as shown in Scheme 1, the reaction of benzal-
dehyde with Danishefsky’s diene in the presence of 20 mol %
of 4b afforded the cycloadduct 6 in 77% yield with 76.3% ee
(S), while using 4a (18% yield, 12% ee (S)) or 4c (36.5% yield,
19% ee (S)) yields modest results. However, it should be noted
that these energy differences should not be over-interpreted in
elucidating the stereochemical outcome of the experiment.
Although the energy difference between the si-TS and the re-
TS of Models II and IV when 4a and 4c are utilized as the
catalysts (1.8 and 1.6 kcal/mol, respectively) are high enough
to expect good enantioselective control of the catalysis; the
enantiomeric excesses of the products observed in the actual
reaction systems are very low. This might be attributed to the
various possible sources of errors associated with the calculation,
such as approximations of the model substrate, inaccuracies of
methods, and complex reaction conditions different from those
of calculation. Moreover, low reactivity of the catalysis is
expected for both TS-(si)-4a and TS-(si)-4c as a result of their
relatively high activation energy barriers. Under such circum-
stance, the background reaction may become more competitive
with the catalytic process. Indeed, the reaction of 1c with 2 in
the absence of any catalyst under otherwise identical reaction
conditions gave racemic product 3a in 10% yield. Therefore,
these calculation results of the activation energies of the
reactions are well-consistent with the dramatic differences of
the catalytic activities and the enantioselectivities of catalysts
4a-c observed in the experiments despite difficulty in quantita-
tive predictions.
FIGURE 4. TSs involved in the reactions (endo mode) of model diene
5 with 2a in the presence of 4a, 4b, or 4c and their activation energies
(∆E, kcal/mol).
higher than those of the endo ones TS-endo-(si)-4b/TS-endo-
(re)-4b (10.0 and 12.6 kcal/mol, respectively; see Supporting
Information), presumably caused by the more severe steric
repulsions present in the exo TSs.21 Therefore, the contribution
of the exo approach of the diene in the catalytic system can be
negligible. In the following discussion, only the endo TSs are
considered.
Figure 4 shows the TSs (endo approach) of the HDA reactions
of model diene (5) with benzaldehyde (2) catalyzed by 4a-c
(Model II-IV) and their corresponding activation energies,
obtained from B3LYP/6-31G*//B3LYP/6-31G*:PM3 calcula-
tions on re- or si-face attack of the model diene (5) to 4a-c-
activated benzaldehyde. The values of the activation energies,
∆E, were calculated on the basis of the total energies of the
stationary points. As can be seen from the figure, the activation
energies of the reactions in the presence of 4a, 4b, and 4c were
reduced by 4.2, 10.2, and 7.7 kcal/mol, respectively, in
The NPA16 indicates that there is a considerable charge
transfer from the donor diene 5 to acceptor 2a, complex 2a‚‚‚
4a, 2a‚‚‚4b, or 2a‚‚‚4c in the TSs for all the model reactions.
The values of the charge transfer from 5 to the benzaldehyde
system are 0.27 e (endo TS), 0.47 e (TS-(si)-4a), 0.49e (TS-
(si)-4b), and 0.47e (TS-(si)-4c), respectively. These data show
that these TSs have some zwitterionic character, and the
formation of the hydrogen bond at the carbonyl oxygen atom
facilitates the charge transfer in the reaction process. The fact
that the hydrogen-bond interaction results in larger stabilizations
in the TSs than for reactants can be explained by the large
stabilization of the negative charge that is developed at the
carbonyl O6 oxygen atom during nucleophilic attack.
A comparison of the TS geometries can provide clues for
understanding the differences in the catalytic activity of TAD-
DOLs 4a-c in the title reaction. Table 1 shows the calculated
intra- and intermolecular O-H‚‚‚O hydrogen bond distances
in free TADDOLs 4a-c, TADDOL-benzadehyde complexes
(21) Similar preference and interpretation for an endo orientation can
also be found in the Lewis-acid-catalyzed HDA reaction between Dan-
ishefsky’s diene and aldehydes; for examples, see: (a) Danishefsky, S. J.;
Larson, E.; Ashkin, D.; Kato, N. J. Am. Chem. Soc. 1985, 107, 1246. (b)
Simonsen, K. B.; Svenstrup, N.; Roberson, M.; Jorgensen, K. A. Chem.s
Eur. J. 2000, 6, 123.
J. Org. Chem, Vol. 71, No. 7, 2006 2867

Zhang et al.
TABLE 1. Comparison of Hydrogen-Bonding Interactions between
Catalysts 4a-c and 2a, as Well as In Preferential TSs
The shortening of d1 and d2 in TS-(si)-4b are also somewhat
more significant than those in TS-(si)-4a and TS-(si)-4c.
Therefore, 4b can be considered more acidic than 4a and 4c,
and this causes a larger stabilization of the TS, as discussed
above.
Then how can we account for the energetic preference of
TS-(si)-4 over TS-(re)-4? Figure 5a shows a quadrant diagram
perspective of the representative TADDOL derivative 4a, while
Figure 5b,c are the overlaid TSs of the si or re approach for all
three TADDOLs 4a-c, where the overall geometrical features
are very similar. In Figure 5a, the phenyl groups in quadrants
II and IV extend forward, so these two locations are the hindered
quadrants. On the other hand, the other two phenyl rings in
quadrants I and III extend backward, providing less-hindered
quadrants. A comparison of the two TSs (TS-(si)-4a-c and TS-
(re)-4a-c) shows that the substrates lie in the less-hindered
quadrants I and III in TS-(si)-4a-c, thus forming a preferential
transition state (Figure 5b). In contrast, in the TS-(re)-4a-c,
both substrates are situated at the hindered quadrants II and IV,
resulting in the less-stable TSs (Figure 5c). This quadrant map
illustrates the steric interactions that occur between the substrates
and the catalysts in the TSs and well explains the sense of
asymmetric induction observed in the catalysis.
H-bond length (Å)
structure
4a
4b
4c
2a‚‚‚4a
2a‚‚‚4b
2a‚‚‚4c
TS-(si)-4a
TS-(si)-4b
TS-(si)-4c
d1
d2
1.910
1.853
1.900
1.756
1.725
1.752
1.678
1.642
1.675
1.842
1.825
1.831
1.634
1.612
1.622
2a‚‚‚4a-c, as well as for the favorable TSs of TS-(si)-4a, TS-
(si)-4b, and TS-(si)-4c. Upon complexation between 2a and 4a-
c, the length of d1 in 2a‚‚‚4a-c is significantly reduced by about
0.13-0.15 Å compared to those in the free 4a-c. In going from
the complexes to the TSs, the d2 distance is shortened by about
0.21 Å, while d1 is also shortened by about 0.08 Å. Thus, both
the intra- and intermolecular hydrogen bonds can reinforce each
other in the TSs, which can be considered as Bronsted-acid-
assisted Bronsted acid (BBA) catalyst systems.4d A closer
examination indicates that 4b has a shorter d1 than 4a and 4c.
A further question is why catalyst 4b causes a higher
enantioselectivity than either 4a or 4c? As mentioned above,
the spatial orientation of the aryl groups distributed in the
quadrant map is the key point for the enantioselective control
of the reaction process. The impact of the aryl groups in
TADDOL catalysts on the level of enantioselectivity of the
reaction can be addressed by a careful comparison of the
different steric hindrances developed in the TSs using the
quadrant diagram. Although the superposition of the favorable
TSs (TS-(si)-4) of reactions in the presence of catalysts 4a-c
FIGURE 5. Quadrant diagram perspective of the TADDOL molecule: (a) superposition TSs of the si approach (b) and re approach (c) of the
diene, where the TS with the green line represents TS-4a, the blue line represents TS-4b, and the gray line represents TS-4c.
2868 J. Org. Chem., Vol. 71, No. 7, 2006

Hetero-Diels-Alder Reaction of Danishefsky’s Diene
Syntheses of TADDOLs 4a-g. TADDOL derivatives 4a-d and
4f-g were prepared as described in the literature.5b,7 TADDOLs
4e were prepared by following a similar procedure.
adopt very similar overall structures, the 1-naphthyl groups in
TS-(re)-4b extend more significantly forward in quadrants II
and IV than the aryl groups in TS-(re)-4a or TS-(re)-4c. Thus,
TS-(re)-4b suffers from a larger steric interaction than TS-(re)-
4a or TS-(re)-4c and as a result, develops a higher energy barrier
for the re-face approach. This explains why the catalysis by 4b
results in higher enantioselective induction than those by 4a
and 4c.
4e. Yield: 72%; mp 209-211 °C; [R]20 +107.3° (c ) 1.0,
D
CHCl3). IR (KBr): νmax 3411, 3217, 3055, 2967, 2899, 1631, 1599,
1504, 1433, 1335, 1274, 1123, 1106, 1019, 902, 817, 792, 758,
745 cm-1. 1H NMR (300 MHz, DMSO-d6 ): δ 8.27 (s, 2H), 8.03-
7.93 (m, 8H), 7.84-7.74 (m, 6H), 7.59-7.48 (m, 12H), 7.32-
7.29 (d, J ) 11.7 Hz, 2H), 4.96 (s, 2H), 1.75-1.50 (m, 8H). MS
(MALDI): 715.3 (M + Na+). HRMS (MALDI): calcd for
C49H40O4Na, 715.2489; found, 715.2819. Anal. Calcd for
C49H40O4: C, 84.94; H, 5.82. Found: C, 84.67; H, 5.80.
Conclusions
In summary, an enantioselective HDA reaction of Danishef-
sky’s diene with benzaldehyde has been achieved under the
catalysis of a variety of R,R,R′,R′-tetraaryl-1,3-dioxolane-4,5-
dimethanol (TADDOL) derivatives through hydrogen-bonding
activation, affording 2-phenyl-2,3-dihydro-4H-pyran-4-one after
TFA workup in moderate yield and good enantioselectivity. It
was found that the R,R′-aryl substituents of catalysts significantly
impact both the activity and the enantioselectivity of the
TADDOL catalysts. A 1H NMR spectrum of the isolated
primary adduct demonstrates that the TADDOL-catalyzed
reaction proceeds via a concerted pathway. In combination with
the experimental investigations, the mechanism of this reaction
has also been studied theoretically using the ONIOM method.
In agreement with the experimental findings, the calculation
results indicate that the carbonyl group of benzaldehyde is
activated by forming an intermolecular hydrogen bond with one
of the hydroxyl groups of the TADDOL molecule. Meanwhile,
the intramolecular hydrogen bond in the catalyst can facilitate
the intermolecular hydrogen bonding with the benzaldehyde
substrate. The sense of asymmetric induction could be well-
rationalized by the analysis of the relative energies of the
TADDOL-complexed TSs. The impact of aryl substituents in
TADDOL catalysts on the level of enantioselective control of
the reaction could be addressed on the basis of the structural
features of their corresponding TSs. The information attained
from this research will impact the design of a new generation
of chiral diol catalysts for asymmetric reactions of carbonyl
compounds.
Typical Procedure for TADDOL-Catalyzed Asymmetric
Hetero-Diels-Alder Reaction of Danishefsky’s Diene and Al-
dehyde. A 1.5-mL polypropylene microtube was charged with
catalyst 4b (66.6 mg, 0.1 mmol) and freshly distilled benzaldehyde
2a (265 mg, 2.5 mmol), and the mixture was agitated for 30 min
at 15 °C before Danishefsky’s diene 1c (86 mg, 0.5 mmol) was
quickly added. The reaction mixture was kept at 15 °C for 72 h,
followed by dilution with 5.0 mL of ether and quenching with 10
drops of TFA. The mixture was neutralized with saturated NaHCO3
solution (0.8 mL), and the aqueous layer was extracted with Et2O
(3 × 15 mL). The combined organic phase was dried over Na2-
SO4, concentrated, and purified by silica gel column chromatog-
raphy with hexanes/ethyl acetate (4:1) as eluent to afford 2-phenyl-
2,3-dihydro-4H-pyran-4-one (3a)2i as a colorless liquid (67 mg, 77%
yield). IR (liquid film): νmax 3064, 1676, 1596, 1402, 1272, 1228,
1210, 1040, 990, 934, 864, 826, 796, 760, 732, 720, 640, 612 cm-1
.
1H NMR (300 MHz, CDCl3): δ 7.46 (d, J ) 6.6 Hz, 1H), 7.42-
7.36 (m, 5H), 5.51 (dd, J ) 5.7, 1.2 Hz, 1H), 5.41 (dd, J ) 14.18,
3.4 Hz, 1H), 2.95-2.84 (m, 1H), 2.68-2.60 (m, 1H). The
enantiomeric excess (ee) was determined to be 76.3% in favor of
(S) by HPLC on Chiralcel OD column, hexane/2-propanol ) 90:
10, flow rate ) 1.0 mL/min, tR1 ) 14.0 min (S, major), tR2 ) 18.7
min (R, minor).
The primary cycloadduct generated in the 4b-catalyzed HDA
reaction of 1c with 2a was isolated prior to CF3COOH treatment.
The reaction mixture was stirred at 15 °C for 72 h, followed by
the evaporation of the solvent in vacuo. The residue was purified
by silica gel column chromatography with EtOAc/n-hexane ) 1:8
as eluent to afford the intermediate, which was identified as the
structure of the [2+4] adduct by 1H NMR (See Figure S1 in
Supporting Information).
Experimental Section
General Methods. 1H NMR spectra were recorded in CDCl3 or
DMSO-d6 on a 300-MHz spectrometer at 25 °C. Chemical shifts
were expressed in ppm with TMS as an internal standard (δ ) 0
ppm) for 1H NMR. Coupling constants, J, are listed in hertz. Optical
rotation was measured with a 341 automatic polarimeter. HRMS
was taken on a 4.7-tesla FTMS instrument. IR spectra were taken
on a FTIR instrument and were reported in cm-1. The enantiomeric
excesses were determined by HPLC on a chiral column. All the
experiments that were sensitive to moisture or air were carried out
under an argon atmosphere using standard Schlenk techniques.
Commercial reagents were used as received without further
purification unless otherwise noted. Toluene, tetrahydrofuran, and
hexane were freshly distilled from sodium benzophenone ketyl and
degassed before use.
Acknowledgment. Financial support from the NSFC, CAS,
and the Major Basic Research Development Program of China
(Grant No. G2000077506) and Committee of Science and
Technology of Shanghai Municipality is gratefully acknowl-
edged. We also thank Dr. Yuxue Li and Dr. Christian A.
Sandoval for helpful discussions.
Supporting Information Available: 1H NMR spectrum of the
primary cycloadduct, global electrophilicity analysis of the dienes,
and Cartesian coordinates of all of the structures with their computed
total energies (16 pages). This material is available free of charge
via the Internet at http://pubs.acs.org.
JO060129C
J. Org. Chem, Vol. 71, No. 7, 2006 2869