A New Copper Acetate-Bis(oxazoline)-Catalyzed, Enantioselective Henry Reaction

Article abstract of DOI:10.1021/ja0373871

A highly enantioselective, nitroaldol reaction catalyzed by a chiral Cu(II) bis(oxazoline) complex has been developed. The reaction scope includes both aromatic and aliphatic aldehydes (15 examples) affording products in good yields and enantioselectiviti

Full text of DOI:10.1021/ja0373871

Published on Web 09/27/2003
A New Copper Acetate-Bis(oxazoline)-Catalyzed, Enantioselective Henry
Reaction
David A. Evans,* Daniel Seidel, Magnus Rueping, Hon Wai Lam, Jared T. Shaw, and
C. Wade Downey
Department of Chemistry and Chemical Biology, HarVard UniVersity, Cambridge, Massachusetts 02138
Received July 18, 2003; E-mail: evans@chemistry.harvard.edu
Table 1. Ligand Survey of the Henry Reaction (eq 3)a
The nitroaldol (Henry) reaction is an important carbonyl addition
process that affords products that may be transformed into valuable
building blocks.1 Therefore, it is not surprising that recent efforts
have focused on the development of catalytic enantioselective
reaction variants. The current contributions to this area have been
highlighted by the recent studies of Shibasaki and Trost.2-5 The
purpose of this Communication is to report a new catalyst system
for the nitroaldol reaction (eqs 1,2; M ) Cu, X ) OAc). The basis
for the current study was to identify a weakly Lewis acidic metal
complex bearing moderately basic charged ligands (X) that would
facilitate the deprotonation of nitroalkanes (eq 1) as a prelude to
the aldol addition process (eq 2). It was felt that divalent metal
acetate-ligand complexes of the general structure A might meet
these requirements because acetate has been employed as a Brønsted
base in the racemic nitroaldol reaction.1
a All reactions were performed at room temperature on a 0.33 mmol
scale with 15 mol % of ligand and 13.5 mol % of Cu(OAc)2‚H2O at a 0.1
M concentration using 55 equiv of nitromethane. Reactions were run in a
screw-capped vial and were complete within 24 h. b Enantiomeric excess
was determined by HPLC using a Chiracel OD-H column. c Reaction did
not go to completion within 24 h.
With optimized conditions in hand, the scope of the reaction
was explored (Table 2). In general, high enantiomeric excesses (87-
94% ee) are observed at room temperature for aromatic aldehydes
bearing either electron-withdrawing or electron-donating groups
(entries 1-9).11 Aliphatic branched and unbranched aldehydes are
also acceptable substrates, affording nitro alcohol adducts in good
yields and enantioselectivities (entries 10-15, 90-94% ee).
Reaction enantioselectivity can be further improved by lowering
the temperature at the accompanying expense of increasing the
reaction time. In one instance, the reaction in entry 1 was performed
by mixing the reactants and storing the resulting solution at 0 °C
for 10 days. Subsequent isolation afforded the corresponding aldol
adduct in 81% yield and 96% ee. When the same reaction was
carried out at 40 °C, significant rate acceleration was noted with
an accompanying decrease in the enantioselection to 79% and an
increase in elimination byproduct.11
Large-scale reactions at minimal catalyst loading (1 mol %) were
also evaluated. In one instance, the reaction between 2-methoxy-
benzaldehyde and nitromethane (as shown in entry 3, Table 2) was
performed on a 50 mmol scale with 1 mol % of catalyst loading at
a 1.0 M concentration. Following a reaction time of 56 h, the
resulting product 3c was isolated in 92% yield (94% ee). Using
the same reaction parameters, the nitro alcohol product 3m resulting
from a reaction of isobutyraldehyde (as in entry 13) was isolated
in 93% yield and 91% ee (108 h reaction time). These results
A series of divalent metal acetates in combination with chiral
bidentate ligands were screened as enantioselective catalysts for
the nitroaldol process.6,7 From this survey, bis(oxazoline)8 (box)
complexes derived from Cu(OAc)2 emerged as promising catalyst
candidates. The results from the ligand survey with this metal
acetate are summarized in Table 1. The five box ligands (1a-d,
2) with the illustrated absolute configurations that were evaluated
with Cu(OAc)2‚H2O afforded promising levels of enantioselection
(entries 1-5).8,9 In each instance, the reactions carried out at
ambient temperature were complete within 24 h. From this
comparison, the indabox ligand 2 proved to be the ligand of choice,9
providing the nitro alcohol product in 74% ee (Table 1, entry 5).
With ethanol as the solvent (Table 1, entry 6), the nitro alcohol
product was isolated in 81% ee. Further optimization of this process
showed that the reaction may be performed with lower catalyst
loadings (1-5 mol %), while the use of 10 equiv of nitromethane
was found to be sufficient for the reaction to proceed to completion.
Reaction concentrations could also be increased to as high as 1.0
M with no change in enantioselectivity. Cu(II) carboxylate structure
was also evaluated with ligand 2, and it was concluded that this
catalyst variable is subordinate to ligand architecture.10 In all
instances, the only side reaction observed in these reactions was
the accompanying dehydration product.
9
12692
J. AM. CHEM. SOC. 2003, 125, 12692-12693
10.1021/ja0373871 CCC: $25.00 © 2003 American Chemical Society

C O M M U N I C A T I O N S
Table 2. Henry Reaction of Nitromethane with Various
Aldehydesa
entry
R
product
time (h)
yield (%)b
ee (%)c
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Ph
3a
3b
3c
3d
3e
3f
3g
3h
3i
3j
3k
3l
3m
3n
3o
22
42
27
4
76
72
91
86
88
66
70
73
70
81
86
83
86
87
95
94
93
93
89
91
87
92
90
91
90
92
94
94
93
93
2-MeC6H4
2-MeOC6H4
2-NOC6H4
2-ClC6H4
1-naphthyl
4-FC6H4
4-ClC6H4
4-PhC6H4
PhCH2CH2
i-Bu
15
15
45
21
20
24
48
96
48
48
48
Figure 2. Plausible transition structures for the Henry reaction.
t-Bu
i-Pr
n-Bu
cyclohexyl
lowest reactivity (greatest stability). While transition states B-1
(boat), B-2 (chair), and C-1 (chair) all predict the observed sense
of asymmetric induction, our predisposition is to favor B-1 on the
basis of both steric and electronic considerations.
Further investigations into the mechanism and variants of this
process are currently underway and will be reported in due course.
Acknowledgment. Support has been provided by the NSF and
the NIH (GM 33328-18). D.S., M.R., and H.W.L. gratefully
acknowledge Postdoctoral Fellowships from the Ernst Schering
Research Foundation, the Deutsche Akademischer Austauschdienst
(DAAD), and GlaxoSmithKline, respectively.
a All reactions were performed on a 1 mmol scale with 5 mol % of
Cu(OAc)2‚H2O and 5.5 mol % of ligand 2 at a 0.5 M concentration using
10 equiv of nitromethane in ethanol. Reactions were run at room temperature
in a screw-capped vial for the indicated time. b Values are isolated yields
after chromatographic purification. c Enantiomeric excess was determined
by HPLC using Chiracel OD-H, OJ-H, or AD columns.
Supporting Information Available: Experimental procedures,
spectral data, and stereochemical proofs for all compounds (PDF and
CIF). This material is available free of charge via the Internet at http://
pubs.acs.org.
References
(1) For recent reviews, see: (a) Luzio, F. A. Tetrahedron 2001, 57, 915-
945. (b) Ono, N. The Nitro Group in Organic Synthesis; Wiley-VCH:
New York, 2001. (c) Seebach, D.; Beck, A. K.; Mukhopadhyay, T.;
Thomas, E. HelV. Chim. Acta 1982, 65, 1101-1133.
(2) (a) Sasai, H.; Suzuki, T.; Arai, S.; Shibasaki, M. J. Am. Chem. Soc. 1992,
114, 4418-4420. (b) Shibasaki, M.; Yoshikawa, N. Chem. ReV. 2002,
102, 2187-2209 and references therein.
(3) (a) Trost, B.; Yeh, V. S. C. Angew. Chem., Int. Ed. 2002, 41, 861-863.
(b) Trost, B.; Yeh, V. S. C.; Ito, H.; Bremeyer, N. Org. Lett. 2002, 4,
2621-2623.
(4) For related studies, see: (a) Christensen, C.; Juhl, C.; Jørgensen, K. A.
Chem. Commun. 2001, 2222-2223. (b) Christensen, C.; Juhl, C.; Hazell,
R. G.; Jørgensen, K. A. J. Org. Chem. 2002, 67, 4875-4881. (c) Risgaard,
T.; Gothelf, K. V.; Jørgensen, K. A. Org. Biomol. Chem. 2003, 153-
156.
(5) For enantioselective reactions mediated by chiral quaternary salts, see:
(a) Ooi, T.; Doda, K.; Maruoka, K. J. Am. Chem. Soc. 2003, 125, 2054-
2055. (b) Corey, E. J.; Zhang, F.-Y. Angew. Chem., Int. Ed. 1999, 38,
1931-1934.
(6) Other metal salts were also screened in combination with ligand 2 and
methanol as solvent, but were found to be inferior to Cu(OAc)2‚H2O:
Ni(OAc)2‚4H2O, 30% ee; Co(OAc)2‚4H2O, 20% ee; Mn(OAc)2‚4H2O, 0%
ee; Zn(OAc)2‚4H2O, 0% ee; Mg(OAc)2‚4H2O, 0% ee.
Figure 1. Crystal structure of complex 4.
suggest that catalyst loading in the 1% range might well be below
the practical limit due to the long reaction times.
The X-ray structure of the chiral copper-ligand complex 4
shown in Figure 1 reveals the expected square planar geometry
with the acetate carbonyl moieties oriented toward the vacant apical
positions.
An attempt to rationalize how asymmetric induction is imparted
from complex 4 begins with a statement of the impact of the Jahn-
Teller (JT) effect on Cu(II) coordination.12 As illustrated in Figure
2, JT distortion of an octahedral Cu(II) complex creates four
strongly coordinating and two weakly coordinating sites labeled
red and blue, respectively. Addition of a bidentate ligand L2 affords
a complex positioning the two cis-oriented strongly coordinating
sites in the ligand plane and two trans-oriented weakly coordinating
sites perpendicular to the ligand plane (Figure 2, eq 4). For those
complexes that simultaneously bind both electrophile and nucleo-
phile, the most reactive transition states should position the
nucleophile perpendicular to the ligand plane, while the electrophile,
for maximal activation, should be positioned in one of the more
Lewis acidic equatorial sites in the ligand plane as illustrated for
complex B. By the same argument, complex D should exhibit the
(7) Interestingly, CuF2, when used in place of Cu(OAc)2‚H2O, gave very
similar results (73% ee); however, the reaction rate was significantly lower.
(8) Chiral bis(oxazolines) have frequently been used as ligands in a variety
of asymmetric processes. For recent reviews, see: (a) Ghosh, A. K.;
Mathivanan, P.; Cappiello, J. Tetrahedron: Asymmetry 1998, 9, 1-45.
(b) Johnson, J. S.; Evans, D. A. Acc. Chem. Res. 2000, 33, 325-335.
(9) Davies, I. W.; Gerena, L.; Lu, N.; Larsen, R. D.; Reider, P. J. J. Org.
Chem. 1996, 61, 9629-9630.
(10) The influence of carboxylate variable was made on the reaction of PhCHO
with nitromethane, using ligand 2 and the indicated Cu(II) carboxylate:
Cu(OAc)2‚(H2O) (94% ee); Cu(OCHO)2‚(H2O) (92% ee); Cu(benzoate)2
(87% ee); Cu(pivalate)2 (92% ee); Cu(2,4-dimethoxybenzoate)2 (96% ee);
Cu(4-cyclohexylbutyrate)2 (95% ee).
(11) For some substrates (e.g., Table 2, entries 1, 2, and 6), lower yields are
observed due to subsequent elimination of the nitro alcohol.
(12) (a) Hathaway, B. J. Copper. In ComprehensiVe Coordination Chemistry;
Wilkinson, G., Ed.; Pergamon Press: New York, 1987; Vol. 5, Chapter
53. (b) Hathaway, B. J.; Billing, D. E. Coord. Chem. ReV. 1970, 5,
143-207.
JA0373871
9
J. AM. CHEM. SOC. VOL. 125, NO. 42, 2003 12693