Amino acid catalyzed direct asymmetric aldol reactions: A bioorganic approach to catalytic asymmetric carbon-carbon bond-forming reactions

Article abstract of DOI:10.1021/ja010037z

Direct asymmetric catalytic aldol reactions have been successfully performed using aldehydes and unmodified ketones together with commercially available chiral cyclic secondary amines as catalysts. Structure-based catalyst screening identified L-proline a

Full text of DOI:10.1021/ja010037z

5260
J. Am. Chem. Soc. 2001, 123, 5260-5267
Amino Acid Catalyzed Direct Asymmetric Aldol Reactions: A
Bioorganic Approach to Catalytic Asymmetric Carbon-Carbon
Bond-Forming Reactions
Kandasamy Sakthivel, Wolfgang Notz, Tommy Bui, and Carlos F. Barbas III*
Contribution from The Skaggs Institute for Chemical Biology and the Department of Molecular Biology,
The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, California 92037
ReceiVed January 3, 2001
Abstract: Direct asymmetric catalytic aldol reactions have been successfully performed using aldehydes and
unmodified ketones together with commercially available chiral cyclic secondary amines as catalysts. Structure-
based catalyst screening identified L-proline and 5,5-dimethyl thiazolidinium-4-carboxylate (DMTC) as the
most powerful amino acid catalysts for the reaction of both acyclic and cyclic ketones as aldol donors with
aromatic and aliphatic aldehydes to afford the corresponding aldol products with high regio-, diastereo-, and
enantioselectivities. Reactions employing hydroxyacetone as an aldol donor provide anti-1,2-diols as the major
product with ee values up to >99%. The reactions are assumed to proceed via a metal-free Zimmerman-
Traxler-type transition state and involve an enamine intermediate. The observed stereochemistry of the products
is in accordance with the proposed transition state. Further supporting evidence is provided by the lack of
nonlinear effects. The reactions tolerate a small amount of water (<4 vol %), do not require inert reaction
conditions and preformed enolate equivalents, and can be conveniently performed at room temperature in
various solvents. In addition, reaction conditions that facilitate catalyst recovery as well as immobilization are
described. Finally, mechanistically related addition reactions such as ketone additions to imines (Mannich-
type reactions) and to nitro-olefins and R,â-unsaturated diesters (Michael-type reactions) have also been
developed.
Introduction
perturbed pKa that is essential to the enamine mechanism of
these antibody aldolases and their natural counterparts, the class
I aldolase enzymes.6a,d,j During the course of mechanistic
characterization of aldolase antibodies we have studied the
efficacy of small peptides, amines, and amino acids in imine/
enamine based decarboxylation and aldol reactions.6b,d,e,k While
designed7 and evolved6k peptides can function with some
efficiency in aqueous buffer by maintenance of a perturbed
amino group pKa through electrostatic effects, amino acids and
amines are much more effective catalysts in nonaqueous solvents
where the amine functionality can be maintained in its reactive
unprotonated state.6b,e Indeed, a hydrophobic active site appears
to be key to the functioning of aldolase antibodies as well.6d
Screening of commercially available amino acids and chiral
amines in nonaqueous solvents for their ability to catalyze a
retro-aldol reaction using a UV sensitive reporter aldol based
on 4-dimethylaminocinnamaldehyde generation revealed L-
proline as a promising catalyst.6f This result immediately
suggested the application of this catalyst to intermolecular aldol
addition reactions because this reaction is characteristically
reversible and varying the concentrations of the reactant with
respect to the equilibrium constant allows the reaction to be
driven to completion in either direction. L-Proline is also the
well-known catalyst of the intramolecular Hajos-Eder-Sauer-
Wiechert reaction, an enantiogroup-differentiating aldol cyclo-
dehydration reaction.8 Evidence suggests that this intramolecular
reaction proceeds via an enamine reaction mechanism.9 Again,
mechanistic parallels were observed with antibody aldolases that
were demonstrated to catalyze similar aldol cyclodehydration
reactions albeit with enhanced ee values.6c,e,h
The aldol reaction is widely regarded to be one of the most
important carbon-carbon bond-forming reactions utilized in
organic synthesis. The development of asymmetric methodolo-
gies for this type of reaction has not only broadened its scope
and applicability but had also provided insight into fundamental
stereochemical aspects important to other carbon-carbon bond
forming reactions such as Diels-Alder and Michael reactions.
In general, most methodologies available for the asymmetric
aldol reaction fall in one of the following categories: (a) the
chiral auxiliary-assisted aldol reaction based on the use of
stoichiometric quantities of the chiral appendage;1 (b) chiral
Lewis acid-catalyzed Mukaiyama-type2 and chiral Lewis base-
catalyzed3 aldol reactions; (c) heterobimetallic bifunctional
Lewis acid/Brønsted base-catalyzed direct aldol reactions;4 and
(d) Aldol reactions catalyzed by aldolase enzymes5 and antibod-
ies.6 A significant characteristic of the latter two methodologies
is the employment of unmodified carbonyl compounds as aldol
donor substrates, whereas the first two methodologies require
some degree of preactivation of the substrates involved.
For the past several years, we spent considerable effort on
the development of catalytic asymmetric aldol and related
reactions that are facilitated by amine-based mechanisms within
antibodies.6 Catalytic antibodies prepared using diketone haptens
and reactive immunization have provided aldolase antibodies
with broad scope and excellent enantioselectivities in a wide
range of direct asymmetric aldol reactions. Key to the function
of these catalysts is an active site lysine residue with a highly
* To whom correspondence should be addressed. Fax: +1-858-784-
2583. E-mail: carlos@scripps.edu.
10.1021/ja010037z CCC: $20.00 © 2001 American Chemical Society
Published on Web 05/09/2001

Direct Asymmetric Catalytic Aldol Reactions
J. Am. Chem. Soc., Vol. 123, No. 22, 2001 5261
Table 1. Exploration of Various Amino Acids and Commercially
Available Derivatives as Catalysts of the Direct Asymmetric Aldol
Addition Reaction of Acetone and 4-Nitrobenzaldehyde
Here we describe (i) structure/activity relationships of proline
and proline-like amino acid catalysts of direct asymmetric aldol
addition reactions, (ii) the synthetic scope of L-proline and 5,5-
dimethyl thiazolidinium-4-carboxylate (DMTC) catalysis in
aldol addition reactions, (iii) studies concerning the reaction
mechanism, and (iv) catalysis of related enamine-based Mannich
and Michael-type addition reactions.
Results and Discussion
Structure/Activity Relationships of Amino Acid Catalysts.
In our initial experiments,10 we studied the model aldol addition
reaction of acetone with 4-nitrobenzaldehyde wherein acetone
was a component of the solvent system anhydrous DMSO/
acetone (4:1). The large excess of acetone was provided to
enforce an equilibrium favoring the aldol addition product. Since
acetone is also an inexpensive and commonly used solvent, use
of a large excess can be justified. Several natural amino acids
identified initially by their catalysis of retro-aldol reactions were
studied (Table 1, entries 1 and 2). We found that, after stirring
the homogeneous mixture for 4 h at room temperature,11 only
the reaction catalyzed by L-proline (Table 1, entry 2) resulted
in significant quantities of a new product, 68% isolated yield,
characterized to be â-hydroxyketone 1 resulting from the cross-
aldolization reaction. Moreover, chiral-phase HPLC analysis
revealed that 1 was formed in 76% ee. The failure of both
N-methyl valine (Table 1, entry 3) bearing an acyclic secondary
amine functionality and 2-pyrrolidine carboxamide (Table 1,
entry 6) to yield significant amounts of the desired aldol product
in our model reaction clearly demonstrated that a cyclic
secondary amine moiety as well as an acidic proton in
appropriate spatial proximity is essential for efficient catalysis
to occur. Interestingly, this observation can be correlated to our
amine-based antibody catalysis wherein the amino functionality
of lysine and the phenolic group of tyrosine are believed to be
involved in the enamine-based catalytic cycle.6d,e Further,
comparison of 2-azetidinecarboxylic acid and pipecolic acid
a Isolated yields after column chromatography. b The ee was deter-
mined by chiral-phase HPLC analysis. c Not determined. d Yield was
estimated from HPLC analysis. e Opposite enantiomer.
(Table 1, entries 4 and 5) with L-proline showed that the five-
membered pyrrolidine ring is best suited as the secondary cyclic
amine moiety. This result is in accord with known structure/
activity relationships between amines and their enamine reactiv-
ity.12 Provided with this structure/activity relationship we
performed a structure-based catalyst screen for the direct
asymmetric aldol reaction. We focused our screen on various
commercially available or readily prepared chiral amines
structurally and chemically related to L-proline (Table 1, entries
7-16) by reacting 4-nitrobenzaldehyde in DMSO/acetone
(1) (a) Evans, D. A.; Nelson, J. V.; Taber, T. R. Stereoselective Aldol
Condensations. In Topics in Stereochemistry; Eliel, E. L., Wilen, S. H.,
Eds.; Wiley-Interscience: New York, 1982; Vol. 13, pp 1-116. (b)
Heathcock, C. H. The Aldol Addition Reaction. In Asymmetric Synthesis;
Morrison, J. D., Ed.; Academic Press: New York, 1984; Vol. 3, Chapter
2, pp 111-212. (c) Braun, M. Angew. Chem., Int. Ed. Engl. 1987, 26, 24-
37. (d) Kim, B. M.; Williams, S. F.; Masamune, S. The Aldol Reaction:
Group III Enolates. In ComprehesiVe Organic Synthesis: Additions to C-X
π-Bonds, Part 2; Heathcock, C. H., Ed.; Pergamon Press: Oxford, 1991;
Chapter 1.7, pp 239-276. (e) Paterson, I. The Aldol Reaction: Transition
Metal Enolates. In ComprehesiVe Organic Synthesis: Additions to C-X
π-Bonds, Part 2; Heathcock, C. H., Ed.; Pergamon Press: Oxford, 1991;
Chapter 1.9, pp 301-320. (f) Gennari, C. Asymmetric Synthesis with Enol
Ethers. In ComprehesiVe Organic Synthesis: Additions to C-X π-Bonds,
Part 2; Heathcock, C. H., Ed.; Pergamon Press: Oxford, 1991; Chapter
2.4, pp 629-660. (g) Mukaiyama, T.; Kobayashi, S. Tin(II) Enolates in
the Aldol, Michael, and Related Reactions. Org. React. 1994, 46, 1-103.
(h) Cowden, C. J.; Paterson, I. Asymmetric Aldol Reactions Using Boron
Enolates. Org. React. 1997, 51, 1-200. (i) Braun, M. Simple Diastereo-
selection and Transition State Models. In StereoselectiVe Synthesis, Methods
of Organic Chemistry (Houben-Weyl), E21 ed.; Helmchen, G., Hoffman,
R., Mulzer, J., Schaumann, E., Eds.; Thieme: Stuttgart, 1996; Vol. 3, pp
1603-1666. (j) Seyden-Penne, J. Chiral Auxiliaries and Ligands in
Asymmetric Synthesis; Wiley: New York, 1995; pp 306-340. (k) Sankhava-
si, W.; Yamamoto, M.; Kohmoto, S.; Yamada, K. Bull. Chem. Soc. Jpn.
1991, 64, 1425. (l) Hintermann, T.; Seebach, D. HelV. Chim. Acta 1998,
81, 2093. (m) Murer, P.; Rheiner, B.; Juarasti, E.; Seebach, D. Hetereocycles
1994, 39, 319. (n) Juarasti, E.; Beck, A. K.; Hansen, J.; Matt, T.;
Mukhopadhyay, T.; Simson, M.; Seebach, D. Synthesis 1993, 1271. (o)
Seebach, D.; Hoffmann, M. Eur. J. Org. Chem. 1998, 1337, 7. (p) Enders,
D.; Oberborsch, S.; Adam, J. Synlett 2000, 644. (q) Enders, D.; Teschner,
P.; Raabe, G. Synlett 2000, 637. (r) Enders, D.; Wortmann, L.; Peters, R.
Acc. Chem. Res. 2000, 33, 157. (s) Enders, D.; Bettray, W. Pure Appl.
Chem. 1996, 68, 569. (t) Hanessian, S.; Gomtsyan, A.; Malek, N. J. Org.
Chem. 2000, 65, 5623. (u) Bennani, Y. L.; Hanessian, S. Chem. ReV. 1997,
97, 3161. (v) Katsuki, T.; Yamaguchi, M. Tetrahedron Lett. 1985, 26, 5807.
(2) For excellent reviews, see: (a) Carreira, E. M. Mukaiyama Aldol
Reaction. In ComprehensiVe Asymmetric Catalysis; Jacobsen, E. N., Pfaltz,
A., Yamamoto, H., Eds.; Springer-Verlag: Heidelberg, 1999; Vol. III,
Chapter 29.1. (b) Braun, M. Addition of Achiral Enolates to Achiral
Carbonyl Compunds in the Presence of Chiral Additives and Catalysts. In
StereoselectiVe Synthesis, Methods of Organic Chemistry (Houben-Weyl),
E21 ed.; Helmchen, G., Hoffman, R., Mulzer, J., Schaumann, E., Eds.;
Thieme: Stuttgart, 1996; Vol. 3, pp 1730-1736. (c) Nelson, S. G. Catalyzed
Enantioselective Aldol Additions of Latent Enolate Equivalents. In Tetra-
hedron: Asymmetry 1998, 9, 357-389.
(3) (a) Denmark, S. E.; Wong, K.-T.; Stavenger, R. A. J. Am. Chem.
Soc. 1997, 119, 2333-2334. (b) Denmark, S. E.; Stavenger, R. A.; Wong,
K. T.; Su, X. J. Am. Chem. Soc. 1999, 121, 4982-4991. (c) Denmark, S.
E.; Stavenger, R. A.; Su, X.; Wong, K.-T.; Nishigaichi, Y. Pure Appl. Chem.
1998, 70, 1469-1476. (d) Denmark, S. E.; Stavenger, R. A.; Wong, K.-T.
J. Org. Chem. 1998, 63, 918-919. (e) Denmark, S. E.; Stavenger, R. A. J.
Org. Chem. 1998, 63, 9524-9527. (f) Denmark, S. E.; Stavenger, R. A.
Acc. Chem. Res. 2000, 33, 432-440. (g) Denmark, S. E.; Stavenger, R. A.
J. Am. Chem. Soc. 2000, 122, 8837-8847.
(4) (a) Yamada, Y. M. A.; Yoshikawa, N.; Sasai, H.; Shibasaki, M.
Angew. Chem., Int. Ed. Engl. 1997, 36, 1871. (b) Yamada, Y. M. A.;
Shibasaki, M. Tetrahedron Lett. 1998, 39, 5561-5564. (c) Yoshikawa, N.;
Yamada, Y. M. A.; Das, J.; Sasai, H.; Shibasaki, M. J. Am. Chem. Soc.
1999, 121, 4168-4178 and references therein. (d) Trost, B. M.; Ito, H. J.
Am. Chem. Soc. 2000, 122, 12003-12004.
(5) For excellent reviews on the use of natural aldolase enzymes see:
(a) Gijsen, H. J. M.; Qiao, L.; Fitz, W.; Wong, C.-H. Chem. ReV. 1996, 96,
443-473. (b) Wong, C.-H.; Halcomb, R. L.; Ichikawa, Y.; Kajimoto, T.
Angew. Chem., Int. Ed. Engl. 1995, 34, 412-432. (c) Wong, C.-H.;
Whitesides, G. M. Enzymes in Synthetic Organic Chemistry; Pergamon
Press: Oxford, 1994. (d) Bednarski, M. D. In ComprehensiVe Organic
Synthesis; Trost, B. M., Ed.; Pergamon Press: Oxford, 1991; Vol. 2, p 455.
(e) Machajewski, T. D.; Wong, C.-H. Angew. Chem., Int. Ed. 2000, 39,
1352-1374. (f) Wymer, N.; Toone, E. J. Curr. Opin. Chem. Biol. 2000, 4,
110-119.

5262 J. Am. Chem. Soc., Vol. 123, No. 22, 2001
SakthiVel et al.
Table 2. Direct Asymmetric Aldol Reactions Catalyzed by
(4:1) with 20 mol % catalyst at room temperature for 4-24 h.
In each case, the aldol product was purified by column
chromatography and analyzed by chiral-phase HPLC techniques.
The results of this screening are summarized in Table 1. We
observed that most of the amino acids studied catalyzed the
aldol reaction albeit with varying yields. In these studies catalyst
loading was reduced from 30 to 40 mol % used in the
preliminary screen to 20 mol %. Whereas the commercially
available L-proline derivatives (Table 1, entries 8 and 9) as well
as L-thiaproline (Table 1, entry 7) showed approximately the
same enantioselectivities as L-proline itself, catalysis by 5,5-
dimethyl thiazolidinium-4-carboxylate (DMTC, Table 1, entry
10) provided aldol product 1 with 86% ee and 66% yield, which
represents a 10% increase in enantiomeric excess compared to
the L-proline-catalyzed reaction. In contrast to this, catalysts
bearing substituents at the 2-position of the thiazolidinium-4-
carboxylate scaffold (Table 1, entries 12-14) provided aldol
product 1 in dramatically reduced yield (<10%). In these cases,
even higher temperatures (37 and 50 °C) did not increase the
yield of the aldol product. Instead, the formation of the aldol
condensation product was observed. Similar effects were
observed with R-methylproline (Table 1, entry 15) and a diamine
salt13 (Table 1, entry 11) as catalysts where the major product
of the reaction was the condensation product. This result can
be rationalized by the fact that the assumed enamine formation
is disturbed by steric factors together with a change in the pKa
value of the amine functionality due to additional substitution
at the centers adjacent to the amino group. We questioned if an
amino acid-catalyzed enantioselective dehydration reaction
might influence the ee values of aldol products where the
condensation product is also found. We found that the ee of
the addition products is not affected by enantioselective
dehydration, since no resolution of racemic 1 was observed upon
treatment with L-proline or DMTC under the same reaction
L-Proline and DMTCe
a Isolated yields after column chromatography. b The ee was deter-
mined by chiral-phase HPLC analysis. c The reaction was performed
in neat acetone. d Chloroform was used as solvent. e Typically, the
reactions were performed in DMSO with the corresponding ketone
donor substrate (20 vol %), aldehyde acceptor substrate (0.1 M), and
catalyst (20 mol %) for 24-48 h at room temperature. Following
aqueous workup with saturated ammonium chloride solution and
extraction with ethyl acetate, the organic layer was dried (MgSO4),
filtered, and concentrated and the residue purified by column chroma-
tography (silica, hexanes/ethyl acetate) to afford the corresponding aldol
product.
(6) (a) Wagner, J.; Lerner, R. A.; Barbas, C. F., III Science 1995, 270,
1797. (b) Bjo¨rnestedt, R.; Zhong, G.; Lerner, R. A.; Barbas, C. F., III J.
Am. Chem. Soc. 1996, 118, 11720. (c) Zhong, G.; Hoffmann, T.; Lerner,
R. A.; Danishefsky, S.; Barbas, C. F., III J. Am. Chem. Soc. 1997, 119,
8131. (d) Barbas, C. F., III; Heine, A.; Zhong, G.; Hoffmann, T.;
Gramatikova, S.; Bjo¨rnestedt, R.; List, B.; Anderson, J.; Stura, E. A.; Wilson,
I. A.; Lerner, R. A. Science 1997, 278, 2085. (e) Hoffmann, T.; Zhong, G.;
List, B.; Shabat, D.; Anderson, J.; Gramatikova, S.; Lerner, R. A.; Barbas,
C. F., III J. Am. Chem. Soc. 1998, 120, 2768. (f) Zhong, G.; Shabat, D.;
List, B.; Anderson, J.; Sinha, S. C.; Lerner, R. A.; Barbas, C. F., III Angew.
Chem., Int. Ed. Engl. 1998, 37, 2481. (g) Sinha, S. C.; Barbas, C. F., III;
Lerner, R. A. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 14603. (h) List, B.;
Lerner, R. A.; Barbas, C. F., III Org. Lett. 1999, 1, 59. (i) List, B.; Shabat,
D.; Zhong, G.; Turner, J. M.; Li, A.; Bui, T.; Anderson, J.; Lerner, R. A.;
Barbas, C. F., III J. Am. Chem. Soc. 1999, 121, 7283. (j) Zhong, G.; Lerner,
R. A.; Barbas, C. F., III Angew. Chem., Int. Ed. 1999, 38, 3738. (k) Tanaka,
F.; Barbas, C. F., III Chem. Commun. 2001, 8, 769.
(7) For examples regarding peptide catalysis in other organic transforma-
tions, see: Sigman, M. S.; Jacobsen, E. N. J. Am. Chem. Soc. 1998, 120,
4901. (b) Iyer, M. S.; Gigstad, K. M.; Namdev, N. D.; Lipton, M. J. Am.
Chem. Soc. 1996, 118, 4910. (c) Horstmann, T. E.; Guerin, D. J.; Miller,
S. J. Angew. Chem., Int. Ed. 2000, 39, 3635.
(8) (a) Hajos, Z. G.; Parrish, D. R. J. Org. Chem. 1974, 39, 1615. (b)
Eder, U.; Sauer, G.; Wiechert, R. Angew. Chem., Int. Ed. Engl. 1971, 10,
496. (c) Agami, C.; Platzer, N.; Sevestre, H. Bull. Soc. Chim. Fr. 1987, 2,
358-360.
conditions for 3 days. On the basis of this screening, we
conclude that substitution at the 4-position of L-proline and of
the 5-position of L-thiaproline does not abolish the catalytic
activity, but increases the enantioselectivity of the reaction.
Synthetic Scope of L-Proline and 5,5-Dimethylthiazoli-
dinium-4-carboxylate (DMTC) Catalyzed Aldol Addition
Reactions. On the basis of the results of these screenings, we
characterized L-proline and DMTC as catalysts of aldol addition
reactions involving a variety of ketone donors and a number of
aromatic and aliphatic aldehyde acceptor substrates (Table 2).
In all other reactions of acetone with aromatic aldehydes, the
corresponding products 2-6 (Table 2) were formed in yields
and ee values similar to the model reaction with 4-nitrobenzal-
dehyde. Interestingly, catalysis by DMTC generally provided
increased ee values for the products derived from aromatic
aldehydes compared to catalysis with L-proline whereas both
L-proline and DMTC afforded all other products 7-14 with
comparable ee values. Remarkably, in the case of benzaldehyde
as aldol acceptor, an increase of 29% in the ee of 4 was obtained
by catalysis with DMTC, whereas both catalysts afforded the
corresponding aldol products 8 and 9 from the reaction of
acetone with isobutyraldehyde and cyclohexanecarboxaldehyde,
respectively, in excellent ee’s. These results suggest that DMTC
is the preferred catalyst of aldol addition reactions involving
aromatic aldehyde acceptor substrates.
(9) (a) Agami, C.; Puchot, C.; Sevestre, H. Tetrahedron Lett. 1986, 27,
1501. (b) Agami, C. Bull. Soc. Chim. Fr. 1988, 3, 499. (c) For an excellent
review on nonlinear effects in asymmetric synthesis see: Girard, C.; Kagan,
H. B. Angew. Chem., Int. Ed. 1998, 37, 2922.
(10) Preliminary results of this screen have been reported. See: List,
B.; Lerner, R. A.; Barbas, C. F., III J. Am. Chem. Soc. 2000, 122, 2395.
(11) In this context, room temperature refers to a temperature range of
24-26 °C.
(12) For an excellent review concerning enamine chemistry see: Hick-
mott, P. W. Tetrahedron 1982, 38, 1975.
However, although DMTC demonstrates improved enanti-
oselectivities in reactions with aromatic aldehydes and compa-
rable enantioselectivities in all the other examples studied, the
chemical yields of the reactions employing DMTC as catalyst
(13) The diastereomeric salt from the less expensive enantiomeric (1S)-
camphorsulfonic acid has also been employed in this study and (R)-1 was
formed in 45% yield and 79% ee.

Direct Asymmetric Catalytic Aldol Reactions
J. Am. Chem. Soc., Vol. 123, No. 22, 2001 5263
are somewhat lower than in the case of catalysis by L-proline.
This is mainly due to a decreased rate of formation of the desired
aldol product, as determined in HPLC studies.
antibodies to kinetic resolution of tertiary aldols,6i we attempted
several L-proline and DMTC catalyzed resolutions of tertiary
aldols in DMSO. We noted significant catalysis of the corre-
sponding retro-aldol reaction; however, we were unable to detect
enantiomeric enrichment in the unreacted tertiary aldols.
To extend the scope of donors used in our amino acid-
catalyzed direct catalytic asymmetric aldol reactions, we
investigated cyclopentanone and cyclohexanone as potential
cyclic ketones, and found that both of these substrates undergo
the desired cross-aldolization to afford syn/anti-mixtures of the
corresponding aldol products 7 and 14. In these two cases,
DMTC and L-proline were comparable with respect to enanti-
oselectivity, whereas L-proline gave a slightly better syn/anti-
ratio as well as chemical yields. The diastereoselectivity of the
reactions in favor of the corresponding anti-products was modest
for both catalysts. Nonetheless, anti-14 could be isolated in 90%
ee from the reaction with DMTC as catalyst.
Recovery and Reuse of the Catalyst. We have studied two
routes to catalyst recycling. The first route involves direct
recovery of the catalyst from the reaction mixture. We noted
that while reactions performed in DMSO provide a homoge-
neous reaction, L-proline and DMTC dissolve, reactions in
chloroform are facile; however, the catalysts do not dissolve to
any significant extent. When the reaction was performed in
chloroform,14 1 was obtained in 61% ee. While this marks a
reduction in ee as compared to the reaction performed in DMSO,
we could quantitatively recover L-proline by simple filtration
from the reaction mixture. Further, reuse of the catalyst in a
second reaction indicated that there was no loss in activity
suggesting that many rounds of catalyst use and recovery should
be possible. As a second route to catalyst recovery and reuse,
we studied catalyst immobilization. In this study, L-proline (20
mol %) was immobilized on a silica gel column, cyclohexane-
carboxaldehyde in DMSO/acetone (4:1) was introduced into the
column and incubated for 48 h at room temperature, and then
the column was thoroughly washed with ethyl acetate. After
purification aldol product 9 was subjected to chiral-phase HPLC
analysis and shown to be formed in 53% ee and 63% yield.
The recovered silica gel column was then reused for a second
aldol reaction where isobutyraldehyde in DMSO/acetone (4:1)
was allowed to react within the column. In this case, aldol
product 8 was isolated in 56% yield and 63% ee. While silica
gel immobilization and recovery of L-proline is facile, the
reduced optical yields that are obtained may not justify this
approach.
Reaction Mechanism. As proposed for the Hajos-Eder-
Sauer-Wiechert reaction, we assume that the key intermediate
of the direct intermolecular asymmetric aldol reactions described
is an enamine formed between L-proline and the corresponding
ketone donor substrate.8,9 In analogy to the mechanism of
aldolase antibody 38C2, this enamine attacks the carbonyl group
of the aldehyde acceptor with high enantiofacial selectivity
which is imposed by a highly organized tricyclic hydrogen
bonded framework resembling a metal-free Zimmerman-
Traxler type transition state (Scheme 1).
This mechanism reflects our observation that both a base and
an acidic proton are required for effective catalysis to occur.
Further support for hydrogen bonding as an essential feature of
the transition state comes from our findings that the addition
of water severely compromises the enantioselectivity and
furthermore decreases the rate of formation of the aldol product.
Yet, it is interesting that the reaction tolerates a small amount
of water (<4 vol %) without affecting the enantiomeric excess
of aldol product 1 (Figure 1).
Next, we turned our attention to 2-butanone as an aldol donor
substrate. This is of particular interest, since this substrate could
form two regioisomeric products, one of which gives rise to
possible syn/anti isomers. However, upon stirring a solution of
the catalyst in 2-butanone/DMSO (1:4) together with isobu-
tyraldehyde or 4-nitrobenzaldehyde, only products 10 and 11,
respectively (Table 2), were formed, resulting from the nucleo-
philic attack of the methyl group of the ketone. Since the
inherent allylic strain in the more substituted enamine is also
present in the regioisomeric enamine involved in the transition
state of this reaction mechanism (Vide infra), it is most likely
that, in this case, sterics are responsible for the observed
selectivity. Therefore, we propose that regioselective formation
of the enamine intermediate is merely a kinetic phenomenum.
Again, catalysis by DMTC afforded higher enantioselectivities
(74% ee vs 59% ee) for the aromatic aldehyde as compared to
the aliphatic isobutyraldehyde where only L-proline was an
efficient catalyst, providing the corresponding aldol product 10
in 60% yield and 80% ee.
In our preliminary communication,10 with the exception of
isobutyraldehyde, only aromatic aldehydes were used as ac-
ceptors. For these studies, anhydrous DMSO was found to be
the best solvent for the model reaction providing 1 (Vide infra).
However, aliphatic R-unsubstituted aldehydes, e.g. pentanal, did
not yield any cross-aldol products under these conditions, but
reacted to provide mainly self-aldolization products. We have
found, however, that use of chloroform as solvent affords the
desired crossed-aldol product 12 in 56% yield and 68% ee. In
this solvent, L-proline was insoluble and the heterogeneous
mixture was simply stirred at room temperature for 24 h. Both
the chemical and the optical yield could be further improved
by performing the reaction in neat acetone which afforded 12
in 75% yield and 73% ee. Moreover, 2-butanone served as an
aldol donor in the reaction with pentanal as well, when
chloroform was used as solvent and L-proline as catalyst thereby
affording the crossed-aldol product 13 in 58% ee and 65% yield.
In addition to the donors described, various other donors, e.g.
3-pentanone, acetylcyclohexane, isopropylmethyl ketone, 3-meth-
yl-2-butanone, and cyclopropylmethyl ketone, were studied with
4-nitrobenzaldehyde as acceptor but did not yield significant
amounts of the corresponding aldol products with either of the
catalysts. The absolute configurations of aldols 1-14 were
assigned based on chiral-phase HPLC analysis and comparison
of optical rotations with previously reported compounds.
Since, in principle, aldol reactions are reversible, another
question to be addressed in this context is whether the ee varies
as a function of time. This is an important feature to be
determined and we found that the ee of aldol product (R)-1 in
our model reaction does not vary significantly (ee ) 75 ( 2%)
when monitored over a period of 24 h.
(14) There is, however, a marked dependence of the ee on the solvent
used. Study of the solvent dependence of the model aldol addition reaction
of acetone to 4-nitrobenzaldehyde to provide aldol 1 resulted in ee values
of 76% (DMSO), 76% (DMF), 67% (acetone), 61% (CHCl3), 60%
(THF),and 56% (CH3CN). While chloroform was not the optimal solvent
for this aldol reaction, it has provided improvements over DMSO in cases
involving aliphatic acceptor substrates (see Table 2, entries 12 and 13).
Tertiary Aldols. Tertiary aldols present significant challenges
for the design of enantioselective small molecule catalysts. This
challenge becomes apparent upon examination of Zimmerman-
Traxler transition state models that we have proposed previously
for this reaction.6i Given our success in applying aldolase

5264 J. Am. Chem. Soc., Vol. 123, No. 22, 2001
SakthiVel et al.
Scheme 1. Enamine Mechanism of the Direct Catalytic
Asymmetric Aldol Reaction Catalyzed by L-Proline
Figure 1. Effect of water on the enantiomeric excess of aldol product
1.
To gain further insight into the mechanism of this reaction
we were motivated to use potential nonlinear effects as a
mechanistic probe.9c Previous studies of the intramolecular
Hajos-Eder-Sauer-Wiechert reaction had reported a nonlinear
relationship between the ee of L-proline and the ee of the
cyclodehydration product.9a,b This effect had been interpreted
to suggest that two molecules of L-proline were involved in the
transition state of this ketone-ketone intramolecular addition.
In contrast to this report, however, we observed a linear effect
for the reaction of acetone and 4-nitrobenzaldehyde in DMSO
with L-proline as a catalyst (Figure 2). This result is consistent
with the transition state proposed above, involving a single
molecule of catalyst at the carbon-carbon bond-forming step.
Two potential explanations for the differences between these
two studies are (1) that the ketone-aldehyde addition reactions
studied here are more facile and do not require a second
molecule of L-proline or (2) that the intramolecular Hajos-
Eder-Sauer-Wiechert reaction is sterically constrained such
that the transition state we propose for our reactions is not
accessible.
According to our reaction mechanism, formation of the
enamine occurs following the formation of an iminium ion
intermediate which can, in principle, partition between two
different enamines. One pathway is depicted in Scheme 1 and
leads to the acetone enamine that serves as C-nucleophile in
the carbon-carbon bond-forming step. Another possible enam-
ine is the one that results from abstraction of the R-hydrogen
of L-proline thereby eliminating the chiral content of the
molecule which upon hydrolysis leads to racemization of the
catalyst. This potential side reaction might ultimately be
responsible for decreased enantioselectivities. To rule out this
possibility, we recovered L-proline from the reaction of acetone
and 4-nitrobenzaldehyde in chloroform, determined its optical
rotation, and compared it with the optical rotation of the catalyst
prior to the reaction. We found that the value of the optical
rotation determined in water for the recovered catalyst ([R]D )
-68.6) was essentially the same as unreacted L-proline ([R]D
) -68.9). We interpret this as strong evidence that racemization
of the catalyst during the course of the reaction does not occur.
Aldol Reactions with Hydroxyacetone as an Aldol Donor
Substrate. In addition to simple aliphatic ketones, we have
investigated whether L-proline15 and DMTC are capable of
catalyzing aldol addition reactions using unprotected hydroxy-
acetone as a particularly intriguing member of the R-heteroatom-
substituted family of ketone donors. In our studies with aldolase
antibodies, we demonstrated for the first time that the hydroxy-
Figure 2. Linear effect in the L-proline catalyzed aldol reaction of
acetone with 4-nitrobenzaldehyde in DMSO. The line fits the equation
y ) 0.69x - 0.47, R2 ) 0.995.
acetone aldol reaction is particularly valuable since a 1,2-diol
unit is formed concurrently with carbon-carbon bond formation.6d
A number of natural aldolase enzymes use dihydroxyacetone
phosphate as substrates wherein the phosphate group is an
essential component of the substrate.5 Like 2-butanone, a
hydroxyacetone donor may provide three different regio- and
diastereomeric products and their corresponding enantiomers,
and to date, only hydroxyl protected glyoxalate esters have been
employed in the Mukaiyama-type catalytic asymmetric aldol
reaction with aldehydes.16
We found that the aldol reaction between hydroxyacetone
and cyclohexanecarboxaldehyde is catalyzed by both L-proline15
and DMTC providing the corresponding anti-diols in 60% and
45% yield, respectively. In contrast, aldolase catalytic antibodies
selectively provide the syn-diastereomer.6d,e We noted that
DMTC-catalyzed hydroxyacetone reactions are significantly
slower than the corresponding L-proline-catalyzed reactions at
room temperature, therefore we performed these reactions at
37 °C. Most significantly, aldol product 15 (Table 3) was formed
with excellent diastereoselectivity (dr >20:1) and enantiose-
lectivity (ee >99%), and furthermore, no regioisomeric aldol
products were detected by HPLC analysis.
For these cases, the diastereoselectivities and enantioselec-
tivities were determined by comparison with the syn-enantiomers
obtained from aldehydes via the Horner-Wadsworth-Emmons
reaction followed by Sharpless asymmetric dihydroxylations
using either AD-mix-R or AD-mix-â.17 The proline-catalyzed
reactions were performed with both D- and L-proline rendering
all four stereoisomers available and the establishment of
(16) Mukaiyama, T.; Shiina, I.; Uchiro, H.; Kobayashi, S. Bull. Chem.
Soc. Jpn. 1994, 67, 1708.
(15) Notz, W.; List, B. J. Am. Chem. Soc. 2000, 122, 7386.
(17) Walsh, P. J.; Sharpless, K. B. Synlett 1993, 605.

Direct Asymmetric Catalytic Aldol Reactions
J. Am. Chem. Soc., Vol. 123, No. 22, 2001 5265
Table 3. Direct Asymmetric Aldol Reaction of Hydroxyacetone
Catalyzed by L-Proline and DMTC as a Route to Anti-Diols
Figure 3. Potential transition states for the aldol reaction of hydroxy-
acetone with aldehydes.
diastereoselectivities can be obtained whereas aromatic as well
as R-unsubstituted aldehydes and protected D-glyceraldehyde
show only low diastereoselectivities. Again, catalysis by DMTC
afforded significantly higher enantiomeric excesses for aromatic
aldols 18, 19, and 20 with comparable diastereoselectivities
albeit with reduced chemical yield as compared to L-proline.
At this point, it can only be assumed that the additional
heteroatom substitution within the ring system and the dimethyl
substitution pattern of DMTC affects the conformation of the
five-membered ring, thereby providing a different steric envi-
ronment that ultimately leads to enhanced enantiomeric excesses.
However, we have no evidence suggesting a mechanism by
which the 5,5-dimethyl substitution improves the enantioselec-
tivity for aromatic aldols.
The following factors contribute to the excellent stereose-
lectivities observed for the aldol reaction of unmodified hy-
droxyacetone with aldehydes. In contrast to 2-butanone, the
regiochemistry of enamine formation is controlled by the
π-donating OH group, which interacts with the π*-orbital of
the CdC bond thereby stabilizing the hydroxyl enamine,6e,18
and furthermore, the enamine double bond is presumed to
possess a (E)-configuration due to minimization of 1,3-allylic
strain.19 However, it should be noted that the geometry of
enamine formation may not necessarily be reflected in the
diastereoselectivity of the products obtained. Thus, upon re-
facial attack of the aldehyde by the si-face of the hydroxyacetone
enamine, the anti-product is formed via a six-membered chairlike
transition state I whereas the formation of the syn-products can
be rationalized by boatlike transition state II, where the facial
selectivity of the hydroxyacetone enamine is reversed (Figure
3).
We assume that due to decreased eclipsing interactions with
sterically less hindered aldehydes or hydrogen bonding with
R-oxygenated aldehydes, other transition states such as the
boatlike transition state II may become increasingly important.
Similar transition states have been proposed for other stereo-
selective reactions.20
a dr ) anti:syn. The ratio was determined by 1H NMR spectroscopy
or weighing of the separated diastereomers. b Isolated yields after
column chromatography. c ee of the anti isomer (ee of the syn isomer).
Ee values were determined by chiral-phase HPLC analysis. d Identical
ee and dr values for 17a and 17b. e Diastereomers could not be
separated. f Combined yield of separated diastereomers. g Opposite
absolute configuration of the â-position (Sharpless AD-â-product).
h From optical rotation.
Amino Acid-Catalyzed Mannich- and Michael-Type Re-
actions. To further expand the scope of amino acid based
catalysis, we have examined other important carbon-carbon
bond-forming reactions (Scheme 2). We recently reported the
utility of this approach in catalysis of Mannich-type reactions.21
In this study, L-proline and DMTC were also determined to be
the most promising catalysts. In a typical reaction, the catalyst
(20 to 30 mol %) was allowed to react with a variety of either
preformed or in situ generated imines (using anisidine and
various aldehydes) in DMSO/acetone (4:1) at room temperature
for 24-48 h and the â-aminoketones of structural types 23a-b
and 24a-b were isolated in up to 89% ee as products. Provided
this success, here we have extended our amino acid based
catalysis approach to Michael-type addition reactions of acetone
conditions for the separation of all four stereoisomers by
chiral-phase HPLC techniques permitted an unambiguous as-
signment of the anti-configuration shown in Table 3 to aldol
product diol 15, which is further supported by its crystal
structure.15
Similarly, high enantioselectivities were obtained for the anti-
diols obtained from the reaction of hydroxyacetone with
isobutyraldehyde, rac-2-phenylpropionaldehyde, 3,3-dimethyl
butyraldehyde, and protected D-glyceraldehyde to afford aldol
products 16, 17a and 17b, 21, and 22, respectively (Table 3).
In these cases, however, DMTC did not yield substantial
amounts of any of the possible aldol products. The anti- and
syn-1-deoxyhexose aldol products 22 provide an indication of
the utility of these types of asymmetric aldol reactions as applied
to carbohydrate chemistry.
(18) Lin, J.-F.; Wu, C.-C.; Lien, M.-H. J. Phys. Chem. 1995, 99, 16903.
(19) Hoffmann, R. W. Chem. ReV. 1989, 89, 1841.
(20) For example, see: Corey, E. J.; Helal, C. J. Angew. Chem., Int. Ed.
Engl. 1998, 37, 1986.
(21) Notz, W.; Sakthivel, K.; Bui, T.; Zhong, G.; Barbas, C. F., III
Tetrahedron Lett. 2001, 42, 199.
In the case of R-substituted aliphatic aldehydes, excellent

5266 J. Am. Chem. Soc., Vol. 123, No. 22, 2001
SakthiVel et al.
Scheme 2. Amino Acid Catalyzed Mannich- and
Michael-Type Reactions
regio-, diastereo-, and enantioselectivity to form anti-1,2-diols.
Notably, even linear R-unsubstituted aliphatic aldehydes, e.g.
pentanal, can be successfully employed in these direct aldol
reactions. In comparison to L-proline, DMTC is the preferred
catalyst of aldol reactions involving aromatic acceptor aldehydes.
We assume that these novel reactions proceed via an enamine
mechanism and a highly organized metal-free Zimmerman-
Traxler-type transition state, which is supported by the observa-
tion of a linear effect for our model reaction and the loss of
enantioselectivity upon addition of water to the reaction medium.
The catalysts remain configurationally stable during the course
of the reaction and are readily recovered for reuse under certain
reaction conditions. Further, these catalysts are functional in
related ketone addition reactions such as Mannich and Michael-
type reactions.23 The catalysts L-proline and DMTC are envi-
ronmentally safe and available in both enantiomeric forms. The
reactions do not require an inert atmosphere or heavy metals
and can be performed at room temperature without preactivation
of the donor substrates. Given the capacity of amines to act as
catalysts via both nucleophilic (enamine-based) and electrophilic
(iminium-based)24 activation, their potential in catalytic asym-
metric synthesis remains to be fully tapped.
Experimental Section
General. Chemicals and solvents were either purchased puriss p.A.
from commercial suppliers or purified by standard techniques. For thin-
layer chromatography (TLC), silica gel plates Merck 60 F254 were
used and compounds were visualized by irradiation with UV light and/
or by treatment with a solution of phosphomolybdic acid (25 g), Ce-
(SO4)2‚H2O (10 g), concentrated H2SO4 (60 mL), and H2O (940 mL)
followed by heating or by treatment with a solution of p-anisaldehyde
(23 mL), concentrated H2SO4 (35 mL), acetic acid (10 mL), and ethanol
(900 mL) followed by heating. Flash chromatography was performed
by using silica gel Merck 60 (particle size 0.040-0.063 mm), 1H NMR
and 13C NMR spectra were recorded either on a Bruker AMX 300 or
a Bruker AMX 250. Chemical shifts are given in δ relative to
tetramethylsilane (TMS); the coupling constants J are given in Hz. The
spectra were recorded in CDCl3 as solvent at room temperature; TMS
served as internal standard (δ ) 0 ppm) for 1H NMR, and CDCl3 was
used as internal standard (δ ) 77.0 ppm) for 13C NMR. HPLC was
carried out using a Hitachi organizer consisting of a D-2500 Chromato-
Integrator, a L-4000 UV-Detector, and a L-6200A Intelligent Pump.
Optical rotations were recorded on a Perkin Elemer 241 Polarimeter
(λ ) 589 nm, 1 dm cell). High-resolution mass spectra were recorded
on an IonSpec FTMS mass spectrometer with a DHB matrix. Gas
chromatography mass spectrometry (GCMS) experiments were per-
formed on a Hewlett-Packard 5890 gas chromatograph and a 5971A
mass selective detector. Electrospray ionization (ESI) mass spectrometry
experiments were performed on an API 100 Perkin-Elmer SCIEX single
quadrupole mass spectrometer.
to various Michael acceptors.22 When L-proline (20 mol %) in
DMSO/acetone (4:1) was reacted with alkylidene malonates 25
and 26 at room temperature for 12 h the Michael adducts 27
and 28 were isolated in 65% and 69% yield, respectively. These
reactions, however, were not enantioselective. Further, as seen
in our aldol and Mannich studies, a variety of ketones could be
used as donor substrates. Similar results were obtained when
â-nitro-olefins 29 and 30 were reacted with acetone in DMSO
with L-proline catalysis. The acetone addition products 31 and
32 were isolated as the sole products in 80% yield each, but
again the products were racemic. Nonetheless, these are the first
catalytic acetone addition reactions to Michael acceptors and
these mild reaction conditions may be of use in the construction
of nonchiral compounds, and further studies concerning catalyst
structure and reaction conditions might provide useful asym-
metric variants of these reactions.23
Conclusions
In summary, we have demonstrated the scope and utility of
the first amine catalysts of the direct asymmetric aldol reaction.
Structure-based catalyst screening of available chiral amines
revealed L-proline and DMTC as powerful catalysts of aldol
addition reactions of unmodified ketones such as acetone,
2-butanone, and cyclic ketones with a variety of aldehydes. Both
aromatic and aliphatic aldehydes react with acetone to provide
the corresponding aldols in moderate to excellent ee values.
2-Butanone reacts with high regioselectivity, and in the case of
hydroxyacetone, the aldol reaction proceeds with excellent
General Procedure for the Preparation of Aldol Products. To a
mixture of anhydrous DMSO (4 mL) and ketone donor (1 mL) was
added the corresponding aldehyde (0.5 mmol) followed by L-proline
or DMTC (20-30 mol %) and the resulting mixture was stirred at
room temperature for 4-72 h. The reaction mixture was treated with
saturated ammonium chloride solution, the layers were separated, and
the aqueous layer was extracted several times with ethyl acetate, dried
with anhydrous MgSO4, and evaporated. The pure aldol products were
obtained by flash column chromatography (silica gel, mixture of
hexanes/ethyl acetate).
(4R)-(4-Nitrophenyl)-4-hydroxy-2-butanone (1): 1H NMR (300
MHz, CDCl3) δ 2.21 (s, 3H), 2.83 (m, 2H), 3.56 (d, J ) 3.2 Hz, 1H),
5.25 (m, 1H), 7.52 (d, J ) 7 Hz, 2H), 8.20 (d, J ) 7.0 Hz, 2H); HPLC
(Daicel Chiralpak AD-RH, CH3CN/H2O ) 30:70, flow rate 0.5 mL/
min, λ ) 254 nm) tR(major) ) 18.76 min, tR(minor) ) 22.56 min;
(22) We have recently reported amine catalyzed sequential Michael-
Aldol reactions to effect an asymmetric Robinson annulation reaction, see:
Bui, T.; Barbas, C. F., III Tetrahedron Lett. 2000, 41, 6951.
(23) We have recently described enantioselective direct Michael additions
of ketones to alkylidene malonates and â-nitro-olefins using (S)-1-(2-
pyrrolidinylmethyl)-pyrrolidine as a catalyst. Further, asymmetric three-
component one-pot Knovenagel-Michael reactions are also catalyzed.
Betancort, J. M.; Sakthivel, K.; Thayumanavan, R.; Barbas, C. F., III
Tetrahedron Lett. 2001 ASAP.
(24) (a) Jen, W.; Wiener, J. M. J.; MacMillan, D. W. C. J. Am. Chem.
Soc. 2000, 121, 9874. (b) Ahrendt, K. A.; Borths, C. J.; MacMillan, D. W.
C. J. Am. Chem. Soc. 2000, 122, 4243.

Direct Asymmetric Catalytic Aldol Reactions
J. Am. Chem. Soc., Vol. 123, No. 22, 2001 5267
[R]D +46.2 (c 1, CHCl3); HR-MS C10H11NO4Na+ 232.1411 (calcd.
232.0586), C10H11NO4 209.20.
(S)-4-(2-Methoxyphenylamino)-4-(1-naphthyl)butan-2-one (23b):
1H NMR δ 2.16 (s, 3H), 2.98 (dd, J ) 8.3, 16.2 Hz), 3.09 (dd, J )
4.0, 16.6 Hz), 3.88 (s, 3H), 4.96 (b, 1H), 5.70 (dd, J ) 4.0, 8.3 Hz,
1H), 6.29 (m, 1H), 6.59-6.62 (m, 2H), 6.76 (m, 1H), 7.37 (m, 1H),
7.49-7.61 (m, 3H), 7.74 (m, 1H), 7.90 (m, 1H), 8.17 (m, 1H); 13C
NMR δ 30.6, 49.9, 50.5, 55.4, 109.3, 111.1, 116.9, 121.0, 122.1, 123.0,
125.5, 125.7, 126.4, 127.8, 129.2, 130.4, 134.0, 136.4, 137.3, 206.8;
HPLC (Daicel Chiralpak AS, hexanes/i-PrOH ) 90:10, flow rate 1.0
mL/min, λ ) 254 nm) tR(major) ) 11.78 min, tR(minor) ) 8.61 min;
(4R)-(4-Acetamidophenyl)-4-hydroxy-2-butanone (2): 1H NMR
(300 MHz, DMSO-d6) δ 2.01 (s, 3H), 2.65 (m, 2H), 4.91 (m, 1H),
5.25 (m, 1H), 7.24 (d, J ) 8.5 Hz, 2H), 7.49 (d, J ) 8.5 Hz, 2H), 9.88
(s, 1H); HPLC (Daicel Chiralpak AD, hexanes/i-PrOH ) 85:15, flow
rate 0.8 mL/min, λ ) 254 nm) tR(major) ) 23.25 min, tR(minor) )
20.12 min; [R]D +12.5 (c 1, CHCl3); HR-MS C12H15NO3Na+ 244.0955
(calcd. 244.0949), C12H15NO3 244.26.
(4R)-4-(1-Naphthyl)-4-hydroxy-2-butanone (3): 1H NMR (300
MHz, CDCl3) δ 2.21 (s, 3H), 2.95 (m, 2H), 3.44 (bs, 1H), 5.34 (m,
1H), 7.45 (m, 3H), 7.82 (m, 4H); HPLC (Daicel Chiralpak AD, hexanes/
i-PrOH ) 92.5:7.5, flow rate 1 mL/min, λ ) 254 nm) tR(major) )
17.86 min, tR(minor) ) 14.64 min; [R]D +36.6 (c 0.5, CHCl3); HR-
MS C14H14O2Na+ 237.2604 (calcd. 237.2549), C14H14O2 214.26.
(4R)-4-(Cyclohexyl)-4-hydroxy-2-butanone (9): 1H NMR (400
MHz, CDCl3) δ 0.97-1.25 (m, 6H), 1.62-1.77 (m, 5H), 2.18 (s, 3H),
2.55 (m, 2H), 2.89 (bs, 1H), 3.80 (m 1H); HPLC (Daicel Chiralpak
AD, hexane/i-PrOH ) 97:3, flow rate 1 mL/min, λ ) 280 nm) tR-
(major) ) 11.54 min, tR(minor) ) 8.46 min; [R]D +41.5 (c 1, CHCl3);
HR-MS C10H18O2Na+ 193.2561 (calcd. 193.2524), C10H18O2 170.254.
(5R)-6-Methyl-5-hydroxy-3-heptanone (10): 1H NMR (300 MHz,
CDCl3) δ 0.9 (t, 6H), 1.14 (t, 3H), 1.75 (m, 1H), 2.52 (m, 4H), 3.02
(bd, 1H), 3.81 (m, 1H); HPLC (Daicel Chiralpak AD, hexanes/i-PrOH
) 99.4:0.6, flow rate 1 mL/min, λ ) 280 nm) tR(major) ) 25.80 min,
tR(minor) ) 19.98 min; [R]D +49.8 (c 1, CHCl3); ESI-MS C8H16O2-
144.1152 (calcd. 144.1152), C8H16O2 (144.121).
[R]D +115.6 (c 1, CHCl3); HR-MS C21H20NO+ 302.1526 (MH+
-
H2O: calcd. 302.1539), C21H21NO2 319.40.
Typical Experimental Procedure for the Catalytic Asymmetric
Three-Component One-Pot Mannich-Type Reaction of Acetone,
p-Anisidine, and Aldehydes. To a mixture of anhydrous DMSO (4
mL) and acetone (1 mL) was added p-anisidine (0.5 mmol) and the
corresponding aldehyde (0.5 mmol) followed by L-proline or DMTC
(20 mol %) and the resulting homogeneous reaction mixture was stirred
at room temperature for 24-48 h. The reaction mixture was further
processed as described above.
(R)-4-Cyclohexyl-4-(4-methoxyphenylamino)butan-2-one (24a):
1H NMR δ 0.96-1.25 (m, 5H), 1.45-1.87 (m, 6H), 2.13 (s, 3H), 2.58
(m, 2H), 3.62 (m, 1H), 3.74 (s, 3H, OMe), 6.56 (d, 2H, J ) 8.8 Hz),
6.75 (d, 2H, J ) 8.8 Hz); 13C NMR δ 26.3, 26.4, 29.3, 29.5, 30.6,
41.7, 45.5, 55.7, 114.7, 114.9, 141.7, 151.9, 208.6; HPLC (Daicel
Chiralpak AS, hexanes/i-PrOH ) 90:10, flow rate 1.0 mL/min, λ )
254 nm) tR(major) ) 8.23 min, tR(minor) ) 14.88 min; [R]D +3.1 (c
1, CHCl3); HR-MS C17H26NO2 275.1885 (MH+: calcd. 276.1970),
+
(5R)-5-(4-Nitrophenyl)-5-hydroxy-3-pentanone (11): 1H NMR
(300 MHz, CDCl3) δ 1.11 (t, 3H), 2.5 (q, 2H), 2.82 (m, 2H), 3.7 (bd,
1H), 4.08 (m, 1H), 5.28 (m, 1H), 7.53 (d, J ) 7 Hz, 2H), 8.20 (d, J )
7 Hz, 2H); 13C NMR δ 7.4, 13.6, 19.1, 30.5, 36.8, 50.2, 68.9, 123.7,
126.3, 150.1, 211.3; HPLC (Daicel Chiralpak AS, hexanes/i-PrOH )
90:10, flow rate 1 mL/min, λ ) 280 nm) tR(major) ) 17.96 min, tR-
(minor) ) 14.72 min; [R]D +43.7 (c 1, CHCl3); ESI-MS C11H13NO4-
Na+ 246 (calcd. 246.0737).
C17H25NO2 275.39.
(R)-4-(4-Methoxyphenylamino)-5-methylhexan-2-one (24b): 1H
NMR δ 0.89 (d, J ) 7.0 Hz, 3H), 0.96 (d, J ) 7.0 Hz, 3H), 1.91 (m,
1H), 2.14 (s, 3H), 2.55 (m, 2H), 3.64 (m, 1H), 3.73 (s, 3H), 6.57 (d,J
) 9.2 Hz, 2H), 6.75 (d, J ) 9.2 Hz, 2H); 13C NMR δ 18.4, 18.8, 30.6,
31.3, 45.2, 55.7, 56.2, 115.0, 141.6, 152.1, 208.4; HPLC (Daicel
Chiralpak AS, hexanes/i-PrOH ) 97:3, flow rate 1.0 mL/min, λ ) 254
nm) tR(major) ) 15.35 min, tR(minor) ) 11.90 min; [R]D +1.5 (c 1,
CHCl3); GC-MS C14H21NO2 235.32.
1
(4R)-4-Hydroxy-2-octanone (12): H NMR (300 MHz, CDCl3) δ
0.92 (t, 3H), 1.4 (m, 6H), 2.18 (s, 3H), 2.57 (m, 2H), 3.01 (bs, 1H),
4.04 (bs, 1H); 13C NMR δ 13.9, 22.5, 27.5, 30.7, 36.0, 49.8, 67.4, 210.1;
HPLC (Daicel Chiralpak AS, hexane/i-PrOH ) 99.4:0.6; flow rate 1
mL/min, λ ) 280 nm) tR(major) ) 24.04 min, tR(minor) ) 21.18 min;
General Procedure for the Preparation of Michael Adducts 27,
28, 31, and 32. L-Proline (20 mol %) is added to a solution of â-nitro-
olefine or alkylidene malonate (1 mmol) in DMSO/acetone (4:1, 10
mL) and the mixture is stirred for 24 h at room temperature. The
reaction mixture was treated with saturated ammonium chloride solution
and the product was extracted with diethyl ether, dried over MgSO4,
and evaporated. Purification by flash column chromatography (silica
gel, mixture of hexanes/ethyl acetate) afforded the corresponding
Michael adducts. See also ref 23.
+
[R]D +37.4 (c 1, CHCl3); GC-MS C8H16O2 144 (calcd. 144.11).
1
(5R)-5-Hydroxynonane-3-one (13): H NMR (300 MHz, CDCl3)
δ 0.92 (t, 3H), 1.16 (t, 3H), 1.39 (m, 6H), 2.54 (m, 4H), 3.07 (bs, 1H),
4.11 (bs, 1H); HPLC (Daicel Chiralpak AS, hexanes/i-PrOH )
99.4:0.6; flow rate 1 mL/min, λ ) 280 nm) tR(major) ) 42.96 min,
+
1
tR(minor) ) 28.50; [R]D +31.6 (c 1, CHCl3); GC-MS C9H18O2 158
Ethyl 2-carboethoxy-5-oxo-3-phenylhexanoate (27): H NMR δ
(calcd.158.13).
1.01 (t, 3H), 1.26 (d, 3H), 2.03 (s, 3H), 2.94 (m, 2H), 3.69 (d, 1H),
3.94 (m, 3H), 4.19 (q, 2H), 7.24 (m, 5H); HR-MS C17H22O5+ 306.3521
(MH+: calcd. 306.3576), C17H22O5 306.25.
General Procedure for the Catalytic Asymmetric Mannich-Type
Reaction of Acetone and Preformed Aldimines. To a mixture of
anhydrous DMSO (4 mL) and acetone (1 mL) was added the
corresponding aldimine (0.5 mmol) followed by L-proline or DMTC
(20 mol %) and the resulting homogeneous reaction mixture was stirred
at room temperature for 24-48 h. Then, half-saturated NH4Cl solution
and ethyl acetate were added with vigorous stirring, the layers were
separated, and the aqueous phase was extracted thoroughly with ethyl
acetate. The combined organic phases were dried (MgSO4), concen-
trated, and purified by flash column chromatography (silica gel,
mixtures of hexanes/ethyl acetate) to afford the desired â-amino ketones.
The enantiomeric excesses of the products were determined by HPLC
analysis using chiral stationary phases.
Ethyl 2-carboethoxy-3-(1-naphthyl)-5-oxohexanoate (28): 1H
NMR δ 0.87 (t, 3H), 1.21 (t, 3H), 2.02 (s, 3H), 3.14 (m, 2H), 3.86 (m,
3H), 4.17 (q, 2H), 7.36-7.84 (m, 6H), 7.84 (bd, 1H); HR-MS
+
C21H24O5 356.4123 (MH+: calcd. 356.4192), C21H24O5 356.41.
1-Nitro-2-phenylpentan-4-one (31): 1H NMR δ 2.12 (t, 3H), 2.91
(d, J ) 8 Hz, 2H), 4.00 (q, 1H), 4.66 (m, 2H), 7.20-7.30 (m, 5H);
+
HR-MS C11H13NO3 207.2312 (MH+: calcd. 207.2325), C11H13NO3
207.23.
1
5-Methyl-4-nitromethyl-2-hexan-2-one (32): H NMR δ 0.92 (t,
6H), 1.82 (m, 1H), 2.16 (s, 3H), 2.43-2.63 (m, 3H), 4.41 (d, J ) 5.8
Hz, 2H); GC-MS C8H15NO3 173 (calcd. 173.21).
(S)-4-(2-Methoxyphenylamino)-4-(4-nitrophenyl)butan-2-one (23a):
1H NMR δ 2.15 (s, 3H), 2.99 (m, 2H), 3.87 (s, 3H), 4.97 (m, 2H),
6.29 (m, 1H), 6.64-6.79 (m, 3H), 7.55 (d, J ) 10.5 Hz, 2H), 8.15 (d,
2H, J ) 10.5 Hz); 13C NMR δ 30.5, 50.9, 53.3, 55.4, 109.5, 111.1,
117.6, 120.9, 123.9, 127.3, 135.8, 146.9, 150.6, 205.4; HPLC (Daicel
Chiralpak AD-RH, H2O/MeCN ) 90:10, 0.1% TFA, flow rate 0.5 mL/
Acknowledgment. Dedicated to Professor K. Barry Sharp-
less, a pioneer in asymmetric catalysis, on the occasion of his
60th birthday. This study was supported in part by the NIH
(CA27489) and The Skaggs Institute for Chemical Biology. We
thank James Turner for assistance and Ben List for early
contributions.
min, λ ) 254 nm) tR(minor) ) 49.63 min, tR(major) ) 45.14 min;
[R]D +11.7 (c 1, CHCl3); HR-MS C17H17N2O3 297.1242 (MH+
-
+
H2O: calcd. 297.1231), C17H18N2O4 (314.34).
JA010037Z