Enantioselective synthesis of dihydropyrans. Catalysis of hetero Diels - Alder reactions by bis(oxazoline) copper(II) complexes

Article abstract of DOI:10.1021/ja992175i

C₂-symmetric bis(oxazoline) - Cu(II) complexes 1 and 2 catalyze the inverse electron demand hetero Diels - Alder reaction of α,β-unsaturated carbonyl compounds (heterodiene) with electron-rich olefins (heterodienophile) in high diastereo- and enantioselectivity, α,β- Unsaturated acyl phosphonates and β,γ-unsaturated α-keto esters and amides are effective heterodienes, while enol ethers and sulfides function as heterodienophiles. A range of substitution patterns is possible on the heterodiene: terminal alkyl, aryl, alkoxy, and thioether substituents are all tolerated. The enantioselective synthesis of dihydropyrans by this method has been shown to be straightforward: cycloadditions may be conducted with as little as 0.2 mol % of the chiral catalyst and are readily run on multigram scale. The reactions exhibit a favorable temperature - enantioselectivity profile, with selectivities exceeding 90% even at room temperature. A simple reaction protocol that employs a solid air-stable catalyst, convenient reaction temperatures, and low catalyst loadings is described. The utility of the derived cycloadducts in the preparation of chiral building blocks is demonstrated. Models for asymmetric induction are discussed considering product stereochemistry, X-ray crystallographic data for the solid catalysts, and mechanistic studies.

Full text of DOI:10.1021/ja992175i

J. Am. Chem. Soc. 2000, 122, 1635-1649
1635
Enantioselective Synthesis of Dihydropyrans. Catalysis of Hetero
Diels-Alder Reactions by Bis(oxazoline) Copper(II) Complexes
David A. Evans,* Jeffrey S. Johnson,1 and Edward J. Olhava1
Contribution from the Department of Chemistry & Chemical Biology, HarVard UniVersity,
Cambridge, Massachusetts 02138
ReceiVed June 24, 1999
Abstract: C2-symmetric bis(oxazoline)-Cu(II) complexes 1 and 2 catalyze the inverse electron demand hetero
Diels-Alder reaction of R,â-unsaturated carbonyl compounds (heterodiene) with electron-rich olefins
(heterodienophile) in high diastereo- and enantioselectivity. R,â-Unsaturated acyl phosphonates and â,γ-
unsaturated R-keto esters and amides are effective heterodienes, while enol ethers and sulfides function as
heterodienophiles. A range of substitution patterns is possible on the heterodiene: terminal alkyl, aryl, alkoxy,
and thioether substituents are all tolerated. The enantioselective synthesis of dihydropyrans by this method
has been shown to be straightforward: cycloadditions may be conducted with as little as 0.2 mol % of the
chiral catalyst and are readily run on multigram scale. The reactions exhibit a favorable temperature-
enantioselectivity profile, with selectivities exceeding 90% even at room temperature. A simple reaction protocol
that employs a solid air-stable catalyst, convenient reaction temperatures, and low catalyst loadings is described.
The utility of the derived cycloadducts in the preparation of chiral building blocks is demonstrated. Models
for asymmetric induction are discussed considering product stereochemistry, X-ray crystallographic data for
the solid catalysts, and mechanistic studies.
Introduction
oxazolidinone moiety participate in highly diastereoselective
Lewis acid promoted hetero Diels-Alder reactions with enol
ethers (eqs 2 and 3).4 A noteworthy feature of the work is a
reversal in diastereofacial bias modulated by judicious selection
of Lewis acid (chelating or nonchelating), enabling the selective
formation of either endo diastereoisomer from the same chiral
auxiliary. Dujardin demonstrated in complementary work that
the heterodienophile may also control the absolute stereochem-
ical course of reaction.5 Enol ethers derived from mandelic acid
react with achiral heterodienes under the influence of catalytic
amounts of Eu(fod)3 to give dihydropyrans of high diastereo-
meric excess (eq 4).6
The hetero Diels-Alder reaction of electron-rich alkenes with
R,â-unsaturated carbonyl compounds is a powerful method for
the preparation of dihydropyrans (eq 1).2 These cycloadducts,
When the present investigation was initiated, one example
of a catalytic, enantioselective intermolecular inverse electron
demand hetero Diels-Alder reaction had been reported in the
literature.7 Wada and co-workers demonstrated that titanium-
TADDOL catalyst B may be employed in cycloadditions of
(3) Representative examples: (a) Monensin: Agtarap, A.; Chamberlin,
J. W.; Pinkerton, M.; Steinrauf, L. J. Am. Chem. Soc. 1967, 89, 5737. (b)
X-537A (Lasalocid A): Westley, J. W.; Evans, R. H.; Williams, T.; Stempel,
A. Chem. Commun. 1970, 71. (c) Bryostatin 1.; Pettit, G. R.; Herald, C.
L.; Doubek, D. L.; Herald, D. L. J. Am. Chem. Soc. 1982, 104, 6846. (d)
Miyakolide: Higa, T.; Tanaka, J.; Komesu, M.; Gravalos, D. C.; Puentes,
J. L. F.; Bernardinelli, G.; Jefford, C. W. J. Am. Chem. Soc. 1992, 114,
7587. (e) Zincophorin: Brooks, H. A.; Gardner, D.; Poyser, J. P.; King, T.
K. J. Antibiot. 1984, 37, 1501. (f) Altohyrtin: Kobayashi, M.; Aoki, S.;
Kitigawa, I. Tetrahedron Lett. 1994, 35, 1243. (g) Phorboxazole: Searle,
P. A.; Molinski, T. F. J. Am. Chem. Soc. 1995, 117, 8126. (h) Palytoxin:
Cha, J. K.; Christ, W. J.; Finan, J. M.; Fujioka, H.; Kishi, Y.; Klein, L. L.;
Ko, S. S.; Leder, J.; McWhorter, W. W.; Pfaff, K.-P.; Yonaga, M.; Uemura,
D.; Hirata, Y. J. Am. Chem. Soc. 1982, 104, 7369.
and their derived tetrahydropyrans, are prevalent structural
subunits in numerous important natural products.3 This article
describes the application of chelating Cu(II) chiral Lewis acid
complexes 1 and 2 to this hetero-cycloaddition process. The
underlying premise for this study was that good levels of
asymmetric induction might be anticipated in this process
through the intervention of the illustrated catalyst-substrate
complex A.
(4) (a) Tietze, L. F.; Schneider, C.; Grote, A. Chem. Eur. J. 1996, 2,
139. (b) Tietze, L. F.; Schulz, G. Liebigs Ann. 1995, 1921. (c) Tietze, L.
F.; Schneider, C.; Montenbruck, A. Angew. Chem., Int. Ed. Engl. 1994,
33, 980. (d) Tietze, L. F.; Montenbruck, A.; Schneider, C. Synlett 1994,
509. (e) Tietze, L. F.; Schneider, C. Synlett 1992, 755.
Important advances have been made in auxiliary-controlled
inverse electron demand hetero Diels-Alder reactions. Tietze
has shown that 1-oxa-1,3-butadienes bearing a chiral acyl
(1) National Science Foundation Predoctoral Fellow.
(2) (a) Tietze, L. F.; Kettschau, G.; Gewart, J. A.; Schuffenhauer, A.
Curr. Org. Chem. 1998, 2, 19. (b) Tietze, L. F.; Kettschau, G. Top. Curr.
Chem. 1997, 189, 1.
(5) (a) Dujardin, G.; Rossignol, S.; Brown, E. Synthesis 1998, 763. (b)
Dujardin, G.; Rossignol, S.; Molato, S.; Brown, E. Tetrahedron 1994, 50,
9037. (c) Dujardin, G.; Molato, S.; Brown, E. Tetrahedron: Asymmetry
1993, 4, 193.
10.1021/ja992175i CCC: $19.00 © 2000 American Chemical Society
Published on Web 02/15/2000

1636 J. Am. Chem. Soc., Vol. 122, No. 8, 2000
EVans et al.
that work, we speculated that 1-oxa-1,3-butadienes possessing
the appropriate activating group at the 2 position could form a
similar type of chelated activated intermediate (e.g., 4 and 5).
It had already been shown in independent investigations by
Boger and Evans that â,γ-unsaturated R-keto esters and R,â-
unsaturated acyl phosphonates participate in hetero Diels-Alder
reactions with electron-rich olefins.10,11 Thus, the phosphonate
and ester groups should fulfill the dual criteria of activating
the substrate for cycloaddition (by lowering the heterodiene
LUMO) and providing a second “docking point” for coordina-
tion to the chiral copper(II) Lewis acid. The chiral environment
afforded by the bis(oxazoline) ligand was expected to confer
reasonable levels of enantioselection in the cycloaddition event
(eq 1). The realization of this prediction has provided a general
entry into the efficient enantioselective construction of dihy-
dropyrans, a conclusion that has also been reached by Jørgensen
and co-workers.12 The full details of our investigation, including
scope, applications, and mechanistic studies, are reported
herein.13
â-oxo-γ,δ-unsaturated sulfones.8 Isopropyl vinyl ether was
found to be a uniquely effective 2π component, facilitating the
synthesis of dihydropyrans in good to excellent enantioselec-
tivity using three terminally substituted heterodienes (eq 5).
We were interested in testing the premise that asymmetric
catalysis could provide a general solution to the enantioselective
synthesis of dihydropyrans. Recent reports from our laboratory
have shown that C2-symmetric bis(oxazoline) (box) copper
complexes 1a and 2a catalyze enantioselective aldol additions
of silyl ketene acetals to pyruvate esters (eq 6) and ene reactions
of unactivated olefins and glyoxylate esters (eq 7).9 Catalyst-
substrate complexes 3a and 3b were implicated as the species
responsible for the observed enantioselection. On the basis of
(6) For other examples of auxiliary-controlled intermolecular inverse
electron demand hetero Diels-Alder reactions, see: (a) Dondoni, A.;
Kniezo, L.; Martinkova, M.; Imrich, J. Chem. Eur. J. 1997, 3, 424. (b)
Degaudenzi, L.; Apparao, S.; Schmidt, R. R. Tetrahedron 1990, 46, 277.
(c) Snider, B. B.; Zhang, Q. J. Org. Chem. 1991, 56, 4908. (d) Sato, M.;
Sunami, S.; Kaneko, C.; Satoh, S. I.; Furuya, T. Tetrahedron: Asymmetry
1994, 5, 1665 and references therein. (e) Lopez, J. C.; Lameignere, E.;
Lukacs, G. J. Chem. Soc., Chem. Commun. 1988, 514. (f) Hayes, P.;
Dujardin, G.; Maignan, C. Tetrahedron Lett. 1996, 37, 3687. (g) Wallace,
T. W.; Wardell, I.; Li, K. D.; Challand, S. R. J. Chem. Soc., Chem. Commun.
1991, 1707. (h) Haagzeino, B.; Schmidt, R. R. Liebigs Ann. Chem. 1990,
1197.
(7) Enantioselective intramolecular inverse electron demand hetero
Diels-Alder reactions catalyzed or promoted by a chiral Lewis acid: (a)
Desimoni, G.; Faita, G.; Righetti, P.; Sardone, N. Tetrahedron 1996, 52,
12019. (b) Narasaka, K.; Hayashi, Y.; Shimada, S.; Yamada, J. Isr. J. Chem.
1991, 31, 261 and references therein. (c) Tietze, L. F.; Saling, P. Synlett
1992, 281.
Acyl Phosphonate Cycloadditions
(8) (a) Wada, E.; Yasuoka, H.; Kanemasa, S. Chem. Lett. 1994, 1637.
See also: (b) Wada, E.; Pei, W.; Yasuoka, H.; Chin, U.; Kanemasa, S.
Tetrahedron 1996, 52, 1205. (c) Wada, E.; Yasuoka, H.; Kanemasa, S.
Chem. Lett. 1994, 145.
The initial phase of this investigation focused on hetero
Diels-Alder reactions of acyl phosphonates with electron-rich
alkenes catalyzed by [Cu(box)]X2 complexes 1 and 2 (eq 9).

EnantioselectiVe Synthesis of Dihydropyrans
J. Am. Chem. Soc., Vol. 122, No. 8, 2000 1637
Table 1. Arbuzov Reactions for Synthesis of Acyl Phosphonates
(Eq 10)
Substrate Synthesis. The Michaelis-Arbuzov reaction pro-
vides a convenient approach to the synthesis of acyl phospho-
nates.14 Treatment of a number of R,â-unsaturated acid chlorides
with trimethyl phosphite (neat, 0f25 °C; CAUTION: chlo-
romethane evolution) afforded unsaturated acyl phosphonates
as yellow liquids in variable yield after low pressure distillation
directly from the reaction pot (Table 1). Since both the starting
material and product are good conjugate acceptors, addition of
a second equivalent of phosphite accounted for low yields in
some cases (entries 1-3).15 The tiglic acid-derived acyl phos-
phonate (entry 5) and the â,â-disubstituted variant (entry 6) were
both formed in reasonably good yield, perhaps as a result of
steric bulk which reduces the intervention of the undesired
reaction pathway. The â,â-disubstituted substrate 6f was
obtained as a mixture of isomeric products (entry 6, E:Z 2.7:1).
In the case of â-alkoxy substrate 6d (entry 4), electronic
deactivation of the â carbon could account for the higher product
yield.
Attempts to employ the Arbuzov reaction to synthesize
cinnamic acid-derived acyl phosphonate 6b were largely unsuc-
cessful under a variety of conditions.16 Recourse was found in
an efficient two-step procedure that entailed base-promoted
addition of dimethyl phosphite to cinnamaldehyde17 and oxida-
tion of resultant allylic alcohol under Parikh-Doering conditions
(eq 11).18 Both methods described were amenable to multigram
preparation of acyl phosphonates.
The product acyl phosphonates could be stored for months
under inert atmosphere at -20 °C and could also be purified
by rapid flash chromatography; however, some hydrolysis to
the corresponding carboxylic acid was always observed. This
contaminant was easily removed with an aqueous bicarbonate
wash to give the pure acyl phosphonate; however, in practice
chromatography was typically eschewed in favor of distillation.
Catalyst Survey. Crotonyl phosphonate 6a was added to a
solution of the appropriate cationic copper(II) catalyst19 (10 mol
%, CH2Cl2) at -78 °C, followed by 3.0 equiv of freshly distilled
ethyl vinyl ether (eq 12). The unpurified cycloadduct was
isolated at the end of the reaction (TLC analysis) by filtration
through silica gel or standard aqueous workup,20 and after flash
chromatography, dihydropyran 7a was obtained in good to
excellent yield. NMR experiments established the Me-OEt cis
relationship arising from an endo transition state. The acetal
proton at C-2 exhibited a small coupling constant (J2-3eq ) 2.3
entry
product
R1
R2
R3
% yield
1
2
3
4
5
6
6a
6b
6c
H
H
H
H
Me
H
Me
Ph
i-Pr
OEt
Me
Et
H
H
H
H
H
Me
27
<10
42
79
84
6d
6e
6fa
71
a E/Z ) 2.7:1.
Hz) with the pseudoequatorial C-3 proton and a larger one with
the pseudoaxial C-3 proton (J2-3ax ) 6.8 Hz); those values are
consistent with other 2,4-substituted cis dihydropyrans.8b The
observation of NOE enhancements between the C-2 and C-4
protons confirmed the syn relationship of the two substituents.
The enantiomeric excess and absolute stereochemistry for 7a
varied with the pendant oxazoline substituent R and counterion
X. High enantioselectivity was observed for [Cu((S,S)-t-Bu-
box)](OTf)2 complex 1a, while changing to the less associating
counterion SbF6 (2a) resulted in a slight diminution in enan-
tiomeric excess. Lower enantioselection was observed when the
oxazoline substituent R was changed from tert-butyl to isopropyl
(2c) or benzyl (2d), but [Cu((S,S)-Ph-box)](X)2 complexes 1b
(X ) OTf) and 2b (X ) SbF6) were both highly selective. Chiral
HPLC analysis indicated that the major enantiomer produced
by [Cu((S,S)-t-Bu-box)](X)2 complexes 1a or 2a was opposite
to the one obtained with all other catalysts. The absolute
stereochemistry was subsequently established by chemical
methods to be that shown in Table 2 (vide infra).
While phenyl- and tert-butyl-substituted catalysts confer
similar levels of enantioselection, a marked difference in their
reactivity has been noted. Specifically, [Cu((S,S)-t-Bu-box)]-
(OTf)2 complex 1a effects complete conversion in 48 h at -78
°C (eq 12), while [Cu((S,S)-Ph-box)](OTf)2 complex 1b requires
only 4 h at the same temperature. The complexes possessing
the hexafluoroantimonate counterion (2) were uniformly more
reactive than those complexes with the triflate anion (1), in
accord with previous findings from our laboratory (e.g., 1a, 48
h; 2a, 22 h).21
(9) (a) Evans, D. A.; Kazlowski, M. C.; Burgey, C. S.; MacMillan, D.
W. C. J. Am. Chem. Soc. 1997, 119, 7893. (b) Evans, D. A.; Burgey, C. S.;
Kozlowski, M. C.; Tregay, S. W. J. Am. Chem. Soc. 1999, 121, 686.
(10) (a) Boger, D. L.; Robarge, K. D. J. Org. Chem. 1988, 53, 3373. (b)
Boger, D. L.; Robarge, K. D. J. Org. Chem. 1988, 53, 5793.
(11) (a) Schuster, T.; Evans, S. A. Phosphorus Sulfur Silicon Relat. Elem.
1995, 103, 259. (b) Telan, L. A.; Poon, C. D.; Evans, S. A. J. Org. Chem.
1996, 61, 7455.
(12) Thorhauge, J.; Johannsen, M.; Jørgensen, K. A. Angew. Chem., Int.
Ed. Engl. 1998, 37, 2404.
(13) A portion of this work has been previously communicated: (b)
Evans, D. A.; Johnson, J. S. J. Am. Chem. Soc. 1998, 120, 4895. (b) Evans,
D. A.; Olhava, E. J.; Johnson, J. S.; Janey, J. M. Angew. Chem., Int. Ed.
Engl. 1998, 37, 3372. (c) Evans, D. A.; Johnson, J. S.; Burgey, C. S.;
Campos, K. R. Tetrahedron Lett. 1999, 40, 2879.
(16) Among those tried: P(OMe)3, PhMe, 0 °C (ref 11a); P(OMe)2-
(OTMS), PhMe, 0 °C (Evans, D. A.; Hurst, K. M.; Takacs, J. M. J. Am.
Chem. Soc. 1978, 100, 3467).
(17) Li, Y.-F.; Hammerschmidt, F. Tetrahedron: Asymmetry 1993, 4,
109.
(18) Parikh, J. R.; Doering, W. v. E. J. Am. Chem. Soc. 1967, 89, 5505.
(19) Cationic copper(II) complexes were prepared as described previ-
ously. See: (a) Reference 9b. (b) Evans, D. A.; Peterson, G. S.; Johnson,
J. S.; Barnes, D. M.; Campos, K. R.; Woerpel, K. A. J. Org. Chem. 1998,
63, 4541.
(20) Washing the organic layer with aqueous NH4OH sequesters the Cu-
(II) and allows for recovery of the ligand. Assays of diastereomeric and
enantiomeric excess were performed prior to flash chromatography.
(21) Evans, D. A.; Murry, J. A.; von Matt, P.; Norcross, R. D.; Miller,
S. J. Angew. Chem., Int. Ed. Engl. 1995, 34, 798.
(14) Karaman, R.; Goldblum, A.; Breuer, E. J. Chem. Soc., Perkin Trans.
1 1989, 765 and references therein.
(15) Szpala, A.; Tebby, J. C.; Griffiths, D. V. J. Chem. Soc., Perkin
Trans. 1 1981, 1363.

1638 J. Am. Chem. Soc., Vol. 122, No. 8, 2000
EVans et al.
Table 3. Cycloadditions of Crotonyl Phosphone 6a with Enols
Table 2. Effect of Ligand and Counterion in the Catalyzed Hetero
Diels-Alder Reaction (Eq 12)
Catalyzed by 1 and 2 (Eq 13)
entry
R
X
endo/exoa
% ee (config.)a
1
2
3
4
5
6
t-Bu (1a)
t-Bu (2a)
i-Pr (2c)
Bn (2d)
Ph (2b)
Ph (1b)
OTf
99:1
69:1
32:1
60:1
>99:1
>99:1
99 (2R,4R)
93 (2R,4R)
39 (2S,4S)
58 (2S,4S)
93 (2S,4S)
94 (2S,4S)
SbF6
SbF6
SbF6
SbF6
OTf
a Diastereomeric and enantiomeric excesses were determined by
chiral HPLC. Product absolute configuration was established by
chemical correlation with a compound of known stereochemistry (vide
infra).
a Determined by capillary GLC or chiral HPLC. b Absolute config-
uration established by X-ray crystallographic analysis of a derivative
(vide infra). c Absolute configuration assigned by analogy. d Reaction
conducted at -40 °C. e Absolute stereochemistry was established by
chemical correlation to compounds of known absolute configuration
(vide infra). f Relative and absolute configuration not assigned. % ee
is for the major diastereomer. g Reaction conducted at -20 °C. h % ee
of minor diastereomer.
To further probe the practicality of the reaction, a systematic
evaluation of the effect of catalyst loading was conducted. The
attenuated reactivity of [Cu((S,S)-t-Bu-box)](SbF6)2 complex 2a
relative to [Cu((S,S)-Ph-box)](SbF6)2 complex 2b was apparent
in these experiments as well, since reducing the amount of 2a
resulted in stalled reactions and incomplete conversion to
product even upon warming to -20 °C. In contrast, 2b was
amenable to substantially reduced catalyst loadings. Dihydro-
pyran 7a was obtained in 95% yield after 16 h at -78 °C using
1 mol % of 2b; no change in enantioselectivity was observed
at this lower catalyst loading. As an added advantage, it was
possible to conduct the reactions at substrate concentrations up
to 1.2 M. At that concentration, complete conversion to the
cycloadduct was realized employing only 0.2 mol % 2b at a
slightly elevated temperature (-50 °C). The yield and enanti-
oselectivity were unchanged (93% yield, 93% ee), but the
protracted reaction time (62 h) signals that this is realistically
the lower limit for catalyst loadings in this reaction. Despite
that limitation, it is noteworthy that 3.6 mg of the chiral ligand
was sufficient to deliver 1.24 g of the enantioenriched product.
Heterodienophile Scope. A representative selection of
electron-rich alkenes was tested in the catalyzed cycloaddition
with crotonyl phosphonate 6a as part of an effort to define the
scope of the reaction (Table 3).
Figure 1. Temperature-enantioselectivity profile for the reaction of
phosphonate 6a and ethyl vinyl ether catalyzed by [Cu((S,S)-t-Bu-box)]-
(OTf)2 complex 1a and [Cu((S,S)-Ph-box)](OTf)2 complex 1b (eq 12).
Reaction Optimization. On the basis of precedent,9 we hoped
that synthetically useful levels of enantioselection might be
maintained at more practical laboratory temperatures. As
evidenced by the data (Figure 1), [Cu((S,S)-t-Bu-box)](OTf)2
(1a) and [Cu((S,S)-Ph-box)](OTf)2 (1b) are capable of delivering
cycloadduct of high enantiomeric excess even at room temper-
ature. Predictably, shorter reaction times are possible: the
reaction of 1a at -20 °C reaches completion in less than 3 h,
while 1b catalyzes the reaction to completion very rapidly (<5
min) at 25 °C. The speed of the reaction coincides with an
exotherm that is noticeable even on fairly small scale (0.3
mmol). Indeed, we noted in our initial report that catalyst
decomposition (brown reaction solution) and low conversion
had been noted for [Cu((S,S)-Ph-box)](OTf)2 complex 1b at 25
°C; this problem has been circumvented simply by introducing
the enol ether slowly (10 min) into the reaction vessel, allowing
for complete conversion to product and retention of the
complex’s characteristic green color. The thermally induced
decomposition described above has not been noted in any other
aspects of this investigation and careful monitoring of the
internal temperature of preparative-scale (ca. 20 mmol) reactions
at low temperature (<-70 °C) reveals that the reaction does
not produce any appreciable exotherm (<3 °C over 10-min
addition).
The synthesis of bicyclic adducts 8a and 9 in high diastero-
meric and enantiomeric excess signals that cyclic enol ethers
are well tolerated. The reduced reactivity of 2,3-dihydropyran
relative to 2,3-dihydrofuran is reflected in the need to conduct

EnantioselectiVe Synthesis of Dihydropyrans
J. Am. Chem. Soc., Vol. 122, No. 8, 2000 1639
Table 4. Catalyzed Reactions of Crotonyl Phosphonate 6a with
Silyl Enol Ethers of Acetophenone (Eq 14)
conferred by 2a and 2b was identical: the same product
enantiomer was formed for both the phenyl and tert-butyl bis-
(oxazoline) Cu(II) catalysts.
Silyl enol phosphonate 16 was obtained as a single geometric
isomer (>20:1 E:Z, 1H NMR) that was assigned the illustrated
stereochemistry by the observation of a characterizic JH-P
coupling of 10.0 Hz for the vinyl proton.24 The observation of
the (E) product geometry implicates a reactive s-cis acyl
phosphonate conformation: accepting a nucleophile from the
s-trans conformation would lead to the formation of the (Z)
isomer. Unfortunately, the (Z) enol was never observed, so it
was not possible to conclude whether 16 was the kinetic or
thermodynamic product (or both).
cat.a
R
product ratio 14:15:16 % ee 14 % ee 15 % ee 16
2a TMS
2b TMS
2a TBS
2b TBS
57:5:38
18:2:80
93:7:0
99
97
99
96
74
-9
82
70
-5
60:40:0
99
The cycloadducts 14/15 were separated from the Michael
adduct 16 by flash chromatography and treated with anhydrous
HCl/MeOH to give (R)-methyl 3-methyl-5-oxo-5-phenylpen-
tanoate (17), a compound of known absolute configuration (eq
15).25 If isomers 14 and 15 possessed the same absolute
configuration at C-4, the enantiomeric excess of the derived
keto ester would be 96% (12:1 isomer ratio f 92% of 99% ee
material and 8% of 74% ee material). The observed enantiose-
lectivity (83%, HPLC) indicated that the two diastereomeric
cycloadducts possessed the opposite configuration at the methyl-
bearing stereocenter. The absolute configuration of Michael
adduct 16 was also assigned by chemical conversion to methyl
ester 17 and was found to be the same as that of the major
cycloadduct 14 (R).
a 2a: [Cu((S,S)-t-Bu-box)](SbF6)2. 2b: [Cu((S,S)-Ph-box)](SbF6)2.
that cycloaddition at higher temperature to achieve complete
conversion.22 The formation of 8a and 9 was also catalyzed by
[Cu((S,S)-Ph-box)](X)2 complexes; however, enantioselectivities
were noticeably lower (for catalyst 1b: ent-8a, 89% ee; ent-9,
52% ee).
The initial results obtained with ethyl vinyl ether were shown
to be optimal for monosubstituted alkenes (Table 3); however,
the dependence of the diastereo- and enantioselectivity on the
size of the enol subsituent provided valuable insight into the
nature of the reaction transition state (vide infra). Methyl vinyl
ether reacted with crotonyl phosphonate in the presence of [Cu-
((S,S)-t-Bu-box)](SbF6)2 complex 2a and [Cu((S,S)-Ph-box)]-
(SbF6)2 complex 2b to yield cycloadduct 10 and ent-10 in
diastereoselectivities that mirror those obtained with ethyl vinyl
ether; however, the enantiomeric excesses were slightly lower.
tert-Butyl vinyl ether exhibited variable selectivity that was
catalyst dependent. A complete collapse of endo selectivity was
found using 2a, while good levels of endo diastereoselection
were maintained using 2b. Enantioselectivities were consider-
ably lower for this heterodienophile relative to methyl and ethyl
vinyl ether. Low selectivity and reactivity were also noted for
tert-butyldimethylsilyl vinyl ether. Ethyl vinyl sulfide induced
no noticeable decomposition of the transition metal catalyst and
was in fact a highly selective participant in the cycloaddition
with 6a to give 13 and ent-13, the latter in especially good yield
and selectivity. Vinyl acetate produced no cycloadduct with 6a
even at 25 °C, while styrene reacted only slowly, producing a
21% yield of the derived cycloadduct after 5 d at 25 °C (2a, dr
4.5:1). The cycloaddition of crotonyl phosphonate 6a and
R-trimethylsiloxy styrene was attempted to assess the viability
of 1,1-disubstituted enol ethers (eq 14, R ) TMS).11b For [Cu-
((S,S)-t-Bu-box)](SbF6)2 complex 2a, the major reaction pathway
was the formal cycloaddition to form both cis and trans
(MeTOTMS) dihydropyrans 14 and 15, while the minor
pathway resulted in the formation of Michael adduct 16.23 The
endo isomer 14 was formed preferentially using either 2a and
2b, with virtually complete π-facial selectivity observed for both
catalysts (99% and 97% ee, respectively). In contrast to those
results obtained with monosubstituted enol ethers, the facial bias
While alteration of the reaction conditions (solvent, temper-
ature) did little to affect the product distribution, effective
channeling of the reaction through the cycloaddition manifold
was realized by changing the oxygen protecting group from
TMS to TBS. Silyl transfer to a putative Cu(II) enolate is
presumably retarded using the bulkier silyl group and the formal
cycloaddition pathway becomes dominant, strongly implicating
a nonconcerted reaction for this heterodienophile (eq 16). Good
enantiocontrol was again realized for both [Cu((S,S)-t-Bu-box)]-
(SbF6)2 complex 2a and [Cu((S,S)-Ph-box)](SbF6)2 complex 2b;
however, the latter catalyst suffered a rather precipitous drop
in diastereoselectivity.26 Despite this fact, the formation of both
diastereomeric products proceeded with high enantioselectivity.
Other ketone-derived enol ethers were tested, but the results
(22) By way of comparison, the reactivity of ethyl vinyl ether is
intermediate to dihydropyran and dihydrofuran. Since these reactions may
reasonably be assumed to proceed through a concerted, but asynchronous
transition state (vide infra), a correlation between heterodienophile reactivity
and enol ether nucleophilicity might be expected. The observed trend
betweeen dihydrofuran and dihydropyran is consistent with the Mayr
nucleophilicity scale, but ethyl vinyl ether was reported to be less
nucleophilic than dihydropyran: Mayr, H.; Patz, M. Angew. Chem., Int.
Ed. Engl. 1994, 33, 938.
(24) Afarinkia, K.; Echenique, J.; Nyburg, S. C. Tetrahedron Lett. 1997,
38, 1663.
(25) (a) Enders, D.; Papadopoulos, K. Tetrahedron Lett. 1983, 24, 4967.
(b) Shi, Y.; Wulff, W. D.; Yap, G. P. A.; Rheingold, A. L. Chem. Commun.
1996, 2601.
(26) In this case, the diastereoisomeric products were separated by
preparative HPLC. Dihydropyran 14 (R ) TBS, 96% ee) was subjected to
acidic MeOH.; the enantiomeric excess of the product keto ester 17 was
also 96%.
(23) The cycloadduct 14/15 and Michael adduct 16 do not interconvert
upon independent resubjection to catalyst 2a.

1640 J. Am. Chem. Soc., Vol. 122, No. 8, 2000
EVans et al.
were uniformally mediocre.27 For reasons that are not known,
the use of the more electron-rich TMS-ketene acetal of tert-
butylthioacetate resulted in low conversion to product. A 30%
yield of a 1.8:1 mixture of cycloadduct and Michael adduct was
obtained with this olefin.
Table 5. Hetero Diels-Alder Reactions of Acyl Phosphonates and
Ethyl Vinyl Ether Catalyzed by 1 and 2 (Eq 18)
The need for heteroatom activation of the 2π component is
not mandatory, as evidenced by the results obtained for the
reaction of cyclopentadiene and 6a (eq 17). Under the influence
product
catalyst
% yield
endo/exoa
% eea
7b: R ) Phb
ent-7b
2a
2b
2a
2b
2a
2b
88
98
78
99
96
49
32:1
>99:1
32:1
>99:1
>99:1
99:1
94
98
93
96
93
77
7c: R ) i-Prb
ent-7c
7d: R ) OEtc
ent-7d
a Determined by capillary GLC or chiral HPLC. b Absolute stereo-
chemistry was established by chemical correlation to compounds of
known absolute configuration (vide infra). c Absolute configuration
assigned by analogy.
of 2a, a separable 1.9:1 mixture of hetero [4+2] and normal
[4+2] adducts was obtained.28 The hetero Diels-Alder product
18 was obtained as a single regio- and diasteroisomer (1H NMR)
in 95% ee. The alternate reaction manifold was less selective,
with the bridged bicyclic product 19 formed in 7:1 endo:exo
selectivity and 84% ee (endo).29 Adducts 18 and 19 are
potentially related to each other by a [3,3] sigmatropic
rearrangement;28b however, a control experiment indicated that
under the conditions of the reaction, no product interconversion
is occurring. A qualitative comparison of the electrophilicity
of unsaturated acyl phosphonates and unsaturated acyl oxazo-
lidinones is now possible. The reaction illustrated in eq 17 was
complete in 1 h at -78 °C, while the corresponding catalytic
Diels-Alder reaction with the crotonyl oxazolidinone requires
8 h at 25 °C.30
of those catalysts relative to [Cu((S,S)-t-Bu-box)](X)2 complexes
1a and 2a. To effect complete conversion to 7d with catalyst
2a, it proved necessary to employ a large excess of ethyl vinyl
ether (10 equiv) due to competing polymerization of that reaction
component, presumably catalyzed by the electrophilic Lewis
acid. Fortunately, this competing reaction is not a general
phenomenon.
To assess the preparative utility of the hetero Diels-Alder
reaction, the catalyzed cycloaddition of 5.17 g of cinnamoyl
phosphonate 6b with ethyl vinyl ether was conducted in the
presence of [Cu((S,S)-Ph-box)](SbF6)2 complex 2b (2.5 mol %),
delivering after 22 h, 6.27 g (98% yield) of ent-7b in 98% ee
and virtually complete diastereoselection (endo/exo > 150:1,
GLC). No special measures are required for the reaction,
indicating the feasibility of producing such enantioenriched
dihydropyrans on a reasonable laboratory scale.
Heterodiene Scope. The other reaction partner in this Diels-
Alder reaction was next considered. In contrast to the 2π moiety,
virtually every permutation of R,â-unsaturated acyl phosphonate
led to a highly selective cycloaddition reaction.
Bicyclic adducts 8b-d were also formed in good yields and
selectivities under the influence of bis(oxazoline) Cu(II) com-
plexes (Table 6). It was feasible in some cases to employ the
less reactive triflate-derived catalysts due to the heightened
reactivity of dihydrofuran relative to ethyl vinyl ether. In contrast
to our carbocyclic Diels-Alder studies, we have found that
triflate catalysts 1a and 1b typically give products whose
enantiomeric excess is 5-10% higher relative to those obtained
using hexafluorantimonate-derived 2a or 2b.
With respect to simple â-substituted R,â-unsaturated acyl
phosphonates, broad substrate generality was enjoyed. In
particular, phenyl-, isopropyl-, and ethoxy-substituted substrates
all participated in high yielding, diastereoselective Diels-Alder
reactions with ethyl vinyl ether to give dihydropyrans 7b-d
(Table 5). The enantioselectivities were consistently excellent
for both 2a and 2b; however, protracted reaction times and
incomplete conversion were observed for the formation of 7b
and 7c using tert-butyl substituted 2a. Our general preference
is to employ [Cu((S,S)-Ph-box)](X)2 complexes 1b (X ) OTf)
or 2b (X ) SbF6) when possible due to the heightened reactivity
The acyl phosphonate derived from tiglic acid (6e) partici-
pated in diastereo- and enantioselective Diels-Alder reactions
with ethyl vinyl ether to give a dihydropyran (20) substituted
at four of the five ring carbon atoms; however, the attenuated
reactivity of this substrate led to low yields as a result of
incomplete conversion when [Cu((S,S)-t-Bu-box)](SbF6)2 com-
(27) Among those tried: 2-methoxypropene (1a, dr 4.5:1, 72 and 37%
ee; 2b, dr 6.9:1, -57 and -38% ee), 2-trimethylsiloxypropene (2a, dr 1.7:
1, 43% ee), 1-cyclohexyl-1-trimethylsiloxyethylene (low conversion),
R-methoxy styrene (2a, dr 2.4:1, 86 and 72% ee; 2b, dr 2.4:1, 36 and 49%
ee). For these enol ethers, only the dihydropyran product was obtained.
(28) Instead of reporting that this was the first enantioselectiVe example
of cyclopentadiene behaving as a dieneophile with an R,â-unsaturated
carbonyl, our initial communication stated only that this was the first
example. We thank Prof. S. Hanessian for alerting us to this error and calling
to our attention this example: (a) Weichert, A.; Hoffmann, H. M. R. J.
Org. Chem. 1991, 4098. In that case the normal Diels-Alder adduct was
preferred (5:1). In our case, we note that when the reaction is conducted
thermally (neat), the normal [4+2] pathway is preferred (2:1). To our
knowledge, the first examples of cyclopentadiene acting as a dieneophile
in hetero Diels-Alder reactions with R,â-unsaturated carbonyls were
reported in 1982: (b) Ismail, Z. M.; Hoffmann, H. M. R. Angew. Chem.,
Int. Ed. Engl. 1982, 21, 859. (c) Dvora´k, D.; Arnold, Z. Tetrahedron Lett.
1982, 23, 4401.
(29) The absolute configuration of 19 was established through chemical
correlation to the corresponding carboxylic acid of known absolute
configuration: (a) Berson, J. A.; Bergman, R. G.; Hammons, J. H.; McRowe,
A. W. J. Am. Chem. Soc. 1967, 89, 2581. (b) Berson, J. A.; Hammons, J.
H.; McRowe, A. W.; Bergman, R. G.; Remanick, A.; Houston, D. J. Am.
Chem. Soc. 1967, 89, 2590.
(30) Evans, D. A.; Miller, S. J.; Lectka, T.; von Matt, P. J. Am. Chem.
Soc. 1999, 121, 7559-7573.

EnantioselectiVe Synthesis of Dihydropyrans
J. Am. Chem. Soc., Vol. 122, No. 8, 2000 1641
Table 6. Hetero Diels-Alder Reactions of Acyl Phosphonates and
2,3-Dihydrofuran Catalyzed by 1 and 2 (Eq 19)
isomers in the starting material was not reflected in the 4.8:1
mixture of diastereomers in the product. A control experiment
in which the unsaturated acyl phosphonate was incubated with
the catalyst in the absence of any heterodienophile demonstrated
that the recovered heterodiene had the same isomeric composi-
tion as the starting material. To account for the observed results,
we propose that the reaction is under Curtin-Hammett control
whereby the chiral Lewis acid induces an isomerization of the
starting material, the activation energy of which is low relative
to the activation energy for the cycloaddition reactions.32 This
(E)/(Z) isomerism presumably proceeds through the interme-
diacy of activated complexes 22 and 23. An endo-selective
cycloaddition of 22 and ethyl vinyl ether accounts for the
formation of the major diastereomer.33 The chiral Cu(II) Lewis
acid is apparently able to make the subtle distinction between
the methyl and ethyl groups. In contrast, the use of TiCl4 as a
promoter gave 21 as a 1.2:1 mixture of diastereoisomers.
product
catalyst
% yield
endo/exoa
% eea
8b: R ) Phb
ent-8b
2a
1b
2a
2b
2a
1b
99
98
79
94
98
(>99:1)
(>99:1)
(49:1)
(>99:1)
(>99:1)
(>99:1)
90
94
90
71
97
84
8c: R ) i-Prb
ent-8c
8d: R ) OEtb
ent-8d
100
a Determined by capillary GLC or chiral HPLC. b Absolute config-
uration assigned by analogy.
r-Keto Ester Cycloadditions
plex 2a was employed (eq 20). This problem was circumvented
by using the more reactive [Cu((S,S)-Ph-box)](SbF6)2 complex
2b (eq 21).
Given the successful application of catalysts 1 and 2 to the
hetero Diels-Alder reactions of acyl phosphonates and enol
ethers, the analogous series of experiments in the R-keto ester
series was investigated (eq 23).
Substrate Synthesis. Substituted â,γ-unsaturated R-keto
esters were accessed by a modified method of Sugimura.34
Boron trifluoride-promoted aldol reactions between the tri-
methylsilyl enol ether of ethyl pyruvate and the appropriate
dimethyl acetals afforded the derived ethers. Treatment of the
unpurified adducts with Florisil (benzene, 25 °C) induced
elimination to provide unsaturated keto esters 24a-c (Scheme
1, eq 24).
γ-Ethoxy substrate 24e could be prepared according to
Tietze’s method through the reaction of ethyl vinyl ether and
ethyl chlorooxalate (neat),35 but was obtained in higher yield
using triethylamine and catalytic amounts of Pd(OAc)2 (Scheme
1, eq 25).5a The γ-methoxy variant 24d could only be obtained
by the latter method since the thermal reaction did not occur at
the reflux temperature of methyl vinyl ether (5-6 °C). The
γ-ethoxy keto ester 24e underwent smooth transetherification
with benzyl mercaptan under mild conditons (Scheme 1, eq 26)
to give γ-thiobenzyl keto ester 24f as a mixture of geometric
isomers that were separated by flash chromatography. The
γ-methyl-â,γ-unsaturated R-keto Weinreb amide 25a was
prepared by monoaddition of 1-propenylmagnesium chloride to
the bis-Weinreb amide of oxalic acid (eq 27),36 while the
The â-ethyl-â-methyl-substituted acyl phosphonate 6f, as
noted in the discussion of substrate preparation, was synthesized
as a 2.7:1 mixture of geometric isomers. When that isomeric
mixture was treated with ethyl vinyl ether in the presence of
[Cu((S,S)-Ph-box)](SbF6)2 2b, an inseparable diasteromeric
mixture of cycloadducts was obtained in 75% yield (eq 22).
(32) For a relevant example, see: (a) Evans, D. A.; Chapman, K. T.;
Bisaha, J. J. Am. Chem. Soc. 1988, 110, 1238. For an example of a catalytic
asymmetric reaction under Curtin-Hammett control, see: (b) Halpern, J.
Science 1982, 217, 401.
(33) Alternatively, the major diastereoisomer could be formed through
an exo selective cycloaddition through the intermediacy of complex 23.
Since the endo transition state is favored for every catalyzed reaction studied
with ethyl vinyl ether, we do not favor this analysis. In theory, employing
a group larger than ethyl should further amplify the proposed Curtin-
Hammett effect; in practice, the â-methyl-â-isopropyl substrate is unreactive.
No reaction was observed with 6f using catalyst 2a.
On the basis of NOE enhancements between the acetal proton
and the Me group at the quaternary carbon, the identity of the
major isomer was established as dihydropyran 21, with the Et
and OEt moieties disposed cis to one another on the six-
membered ring.31 Importantly, the 2.7:1 mixture of geometric
(31) For approaches to the synthesis of quaternary carbon stereocenters,
see: (a) Fuji, K. Chem. ReV. 1993, 93, 2037. (b) Corey, E. J.; Guzman-
Perez, A. Angew. Chem., Int. Ed. Engl. 1998, 37, 388.
(34) Sugimura, H.; Yoshida, K. Bull. Chem. Soc. Jpn. 1992, 65, 3209.
(35) Tietze, L. F.; Meier, H.; Voss, E. Synthesis 1988, 274.
(36) Sibi, M. P.; Marvin, M.; Sharma, R. J. Org. Chem. 1995, 60, 5016.

1642 J. Am. Chem. Soc., Vol. 122, No. 8, 2000
EVans et al.
Scheme 1a
surprisingly low (1.3:1). While the reversal in enantioselectivity
was also noted for this substrate employing [Cu((S,S)-Ph-box)]-
(SbF6)2 complex 2a (-56% ee), the level of asymmetric
induction was not comparable to the tert-butyl variant. In
contrast to the results in the acyl phosphonate cycloadditions,
[Cu((S,S)-i-Pr-box)](OTf)2 complex 1c and [Cu((S,S)-Bn-box)]-
(OTf)2 complex 1d afforded the same product enantiomer as
1a. Initial attempts to expand the scope of the reaction to other
heterodienes were unsuccessful: the reaction of the γ-phenyl-
â,γ-unsaturated-R-oxo ester 24b with ethyl vinyl ether under
identical conditions gave the corresponding cycloadduct in only
78% ee. Examination of other reaction variables appeared
warranted at this juncture.
Reaction Optimization. We first decided to take advantage
of observations made previously in our laboratory during
investigations of the aldol and carbocyclic Diels-Alder reac-
tions.37 A key finding in those studies was that the use of aquo
complex 27a in the presence of 3 Å molecular sieves generated
a catalyst whose activity and selectivity were identical with those
of anhydrous complex 1a. Aquo complexes 27a and 27b were
prepared by the addition of 2.0 equiv of water to a THF solution
of the corresponding anhydro complex (eq 30). Following
solvent removal and trituration with hexane, [Cu((S,S)-t-Bu-
box)(H2O)2(OTf)](OTf) complex 27a and [Cu((S,S)-Ph-box)-
(H2O)2(OTf)2] complex 27b were obtained in high yield as stable
light blue solids characterized by elemental analysis and X-ray
crystallography.
a Reaction conditions: (a) BF3‚OEt2, CH2Cl2, -78 °C; (b) Florisil,
C6H6, 25 °C; (c) Pd(OAc)2 (cat.), Et3N; (d) cat. KHSO4‚H2O, 30 Torr.
Table 7. Effect of Ligand and Counterion in Cycloadditions of
R-Keto Esters (Eq 29)
entry
R
X′
endo/exoa
% ee (config.)a
1
2
3
4
5
t-Bu (1a)
t-Bu (2a)
i-Pr (1c)
Bn (1d)
Ph (1b)
OTf
SbF6
OTf
OTf
OTf
49:1
1.3:1
24:1
24:1
20:1
95 (2R,4R)
95 (2R,4R)
38 (2R,4R)
39 (2R,4R)
56 (2S,4S)
a Diastereomeric and enantiomeric excesses were determined by
chiral HPLC. The absolute configuration was assigned by analogy to
related compounds (vide infra).
The findings described above potentially define a protocol
for streamlining the catalyst preparation, obviating the need to
complex the ligand to the metal salt for every reaction. The
reduction of this concept to practice was straightforward.
Implementation of a protocol wherein the heterodiene and
heterodienophile were added to a mixture of [Cu((S,S)-t-Bu-
box)(H2O)2(OTf)](OTf) complex 27a and molecular sieves gave
results indistinguishable from those obtained using 1a. Subse-
quent control experiments have revealed that 27a is a poor
catalyst for the reaction in the absence of molecular sieves and
that catalyst dehydration does not occur to an appreciable extent
in the absence of the R-keto ester.38
γ-methoxy variant 25b was synthesized through amination of
the appropriate R-keto acid chloride (eq 28, unoptimized).
A survey of reaction solvents employing the operationally
simplified catalyst preparation demonstrated that dichlo-
romethane was not the optimal reaction solvent (Table 8).39
A
(37) (a) Reference 9b. (b) Barnes, D. M. Ph.D. Thesis, Harvard
University, 1997. For other uses of aquo complexes in catalytic asymmetric
reactions, see: (c) Kanemasa, S.; Oderaotoshi, Y.; Yamamoto, H.; Tanaka,
J.; Wada, E.; Curran, D. P. J. Org. Chem. 1997, 62, 6454-6455. (d)
Kanemasa, S.; Oderaotoshi, Y.; Sakaguchi, S.; Yamamoto, H.; Tanaka, J.;
Wada, E.; Curran, D. P. J. Am. Chem. Soc. 1998, 120, 3074-3088. (e)
Ghosh, A. K.; Cho, H.; Cappiello, J. Tetrahedron: Asymmetry 1998, 9,
3687-3691.
The keto esters were purified by flash chromatography and
stored under argon atmosphere at -20 °C. Some decomposition
of the γ-alkoxy compounds was noted over time. Optimal results
with these substrates were obtained when the keto ester was
purified immediately prior to use.
(38) The reaction of E-ethyl-5-methyl-2-oxo-3-hexenoate (24c) and ethyl
vinyl ether was conducted with catalysts (2 mol %) prepared three different
ways: (A) 27a was stirred for 1 h in THF and the reaction was conducted
as normal; (B) 27a was stirred in the presence of 3 Å molecular sieves for
1 h in THF, after which time the mixture was filtered through an oven-
dried 0.2 µm filter into an oven-dried flask. The reaction was conducted as
normal; (C) 27a was stirred in the presence of 3 Å molecular sieves for 1
h in THF and the reaction was conducted as normal. Approximate half-
lives for the reactions (0 °C): A, ≈20 h; B, ≈5 h; and C, <5 min.
Catalyst Survey. The reaction of E-ethyl 2-oxo-3-pentenoate
24a with ethyl vinyl ether in the presence of [Cu((S,S)-t-Bu-
box)](OTf)2 complex 1a (10 mol %, CH2Cl2, -78 °C) led to
the formation of dihydropyran 26a in good yield and excellent
stereoselectivity (Table 7). The enantioselectivity afforded by
the more reactive SbF6-derived 2a was identical with that of
the triflate catalyst (95% ee), but the diastereoselection was

EnantioselectiVe Synthesis of Dihydropyrans
J. Am. Chem. Soc., Vol. 122, No. 8, 2000 1643
Table 8. Catalyzed Hetero Diels-Alder Reactions Employing
Bis(aquo) Complex 27a and 3 Å Molecular Sieves (Eq 31)
A 5 mol % catalyst loading of [Cu((S,S)-t-Bu-box)(H2O)2-
(OTf)](OTf) complex 27a was needed to effect complete
conversion to product in the acyl phosphonate reactions. This
requirement was consistent with a competition experiment
demonstrating that â,γ-unsaturated R-keto esters were somewhat
more reactive that R,â-unsaturated acyl phosphonates in cata-
lyzed cycloadditions (eq 34). A stoichiometric amount of catalyst
was employed in this reaction to avoid potential complications
arising from different substrate‚catalyst binding affinities (eq
8).
entry
solvent
endo/exo
% ee
1
2
3
4
5
CH2Cl2
THF
19:1
24:1
24:1
28:1
24:1
89
97
97
98
93
Et2O
toluene
dioxanea
a This reaction run at room temp.
number of solvents functioned well in the reaction, with THF,
Et2O, and toluene all providing selectivities superior to CH2-
Cl2.
Two observations made simultaneously carried important
consequences for the practical implementation of subsequent
reactions. The first was that nearly complete enantiocontrol
could be maintained even at 0 °C. While it was possible to attain
enantioselectivities exceeding 99% at lower temperature (-40
°C), the favorable temperature-enantioselectivity profile40
meant that synthetically useful levels of enantioselection were
possible at convenient temperatures. The second finding was
that 2 mol % of the Cu(II) complex 27a was sufficient to
catalyze the cycloaddition to completion in less than 1 h. These
advantages led us to pursue this protocol in later reactions.
Reaction Scope. The substrate tolerance for this reaction was
broad, with uniformly high levels of enantiomeric excess found
in the adduct dihydropyrans (Table 9). Alkyl-, aryl-, alkoxy-,
and thiobenzyl-substituted heterodienes reacted with ethyl vinyl
ether to give 26a-f in 96-99% enantioselection using [Cu-
((S,S)-t-Bu-box)(H2O)2(OTf)](OTf) complex 27a. Reaction times
were typically 15-45 min at 0 °C. In contrast to the acyl
phosphonate reactions with tert-butyl-substituted catalysts,
reduced catalyst loadings were possible: 26a was synthesized
in high selectivity (19:1 endo:exo, 96% ee) using 0.5 mol % of
27a. Preparative scale reactions were again straightforward, as
5.0 g of phenyl-substituted keto ester afforded 6.7 g (98%) of
26b (>99:1 endo:exo, 99% ee) using 2 mol % of 27a.
Dihydrofuran and vinyl sulfides functioned efficiently as
heterodienophiles to give 28b,c and 29a,b stereoselectively. We
have further found that the substituted â,γ-unsaturated R-keto
Weinreb amides were also excellent heterodienes, giving 30a,b
with high diastereo- and enantiocontrol.41 The synthetic advan-
tages of compounds incorporating this moiety have been
reviewed.42
The successful use of a solid catalyst precursor prompted us
to consider the possibility of reusing the chiral complex in
multiple reaction cycles.43 When the catalyzed cycloaddition
of 24b and ethyl vinyl ether was conducted with 3 Å molecular
sieves and Florisil in hexanes, a solvent in which [Cu((S,S)-t-
Bu-box)(H2O)2(OTf)](OTf) complex 27a is apparently insoluble,
a rapid reaction ensued.44 The supernatant was decanted by
syringe after keto ester 24b was completely consumed (TLC
analysis), and the resulting solids were rinsed twice with hexane.
Introduction of more reactants led to resumption of the reaction,
a cycle that was repeated five times (Table 11). As evidenced
by the data, reactivity degrades in progressive reaction cycles;
nonetheless, the consistently high yields and selectivities are
noteworthy.
Product Elaboration and Stereochemical Proofs
Upon treatment of the adduct dihydropyrans with dry HCl/
MeOH, rupture of the acetal linkage and solvolysis of the enol
phosphonate ensued to afford the corresponding acetal ester.45
Without purification, the dimethyl acetal was removed in
aqueous acetone with catalytic amounts of pyridinium p-
toluenesulfonate, giving aldehydic esters in 73-86% yield (R1
) Me, Ph, i-Pr) for the two steps (Scheme 2, steps a and b).
Attempts to merge these two steps were generally unsuccessful
and application of the protocol to 4-alkoxy-2,3-dihydropyrans
(7d, R ) OEt) was not feasible due to concomitant elimination
of the alkoxy group. The products of this solvolysis sequence
are Michael adduct equivalents of an aldehyde enolate and
â-substituted acrylate ester.
It appeared likely that the conditions employed in the acyl
phosphonate cycloadditions could be optimized considering the
results obtained with R-keto esters. Accordingly, the improved
reaction conditions were tested for a representative selection of
acyl phosphonates. The results of Table 10 unambiguously show
that superior results were realized with the simplified reaction
protocol.
Reduction of the aldehydic moiety followed by acid-catalyzed
cyclization yielded the corresponding δ-lactones 32a,c, facilitat-
ing unambiguous determination of the absolute configuration
of the original cycloadducts by comparison of optical rotations.46
Those assignments are summarized below.47
(39) The reactions were conducted at 0 °C for the sake of simplicity.
The higher temperature should also accentuate any differences in enanti-
oselectivity.
(40) Reaction of the phenyl-substituted unsaturated keto ester in THF
to give 26b: -40 °C, >99% ee; -20 °C, 98% ee; 0 °C, 97% ee; and 25
°C, 94% ee.
(41) The authors thank Mr. Jacob M. Janey for performing one of these
reactions (30a).
(42) (a) Sibi, M. P. Org. Prep. Proc. Int. 1993, 25, 15-40. (b) Mentzel,
M.; Hoffmann, H. M. R. J. Prakt. Chem. 1997, 339, 517-524.
(43) . For an excellent example, see: Hansen, K. F.; Leighton, J. L.;
Jacobsen, E. N. J. Am. Chem. Soc. 1996, 118, 10924.
(44) The exact role of the absorbent is not clear. Recycling is still possible
in the absence of Florisil, but the dropoff in activity appears greater. In the
presence of SiO2, the ee is 87%.
(45) If the reaction was stopped prior to completion, complete incorpora-
tion of MeOH was observed at the anomeric carbon of the dihydropyran
and the equilibrium cis:trans ratio was 1:3.

1644 J. Am. Chem. Soc., Vol. 122, No. 8, 2000
EVans et al.
Table 9. Catalyzed Hetero Diels-Alder Reactions with [Cu((S,S)-Bu-box)(H2O)2](OTf)2 (27a) and 3 Å Molecular Sieves (Eq 32)
a Determined by capillary GC or chiral HPLC. b Absolute configuration assigned by analogy. c Absolute configuration assigned by chemical
correlation to a compound of known absolute configuration (vide infra). d Reaction conducted with 5 mol % of 27a.
Scheme 2a
Table 10. Hetero Diels-Alder Reactions of Acyl Phosphonates
and Enol Ethers and Sulfides Catalyzed by
[Cu((S,S)-Bu-box)(H2O)2](OTf)2 (27a) (Eq 33)
R1
R2
X
% yield
endo/exo
% eea
Me
Ph
i-Pr
OEt
Me
Me
H
OEt
OEt
OEt
OEt
OEt
SEt
84
95
92
92
98
75
36:1
22:1
22:1
44:1
25:1
16:1
93 (93)
97 (94)
H
H
H
Me
H
95 (93)
97 (93)
g90 (g94)
96 (76)
a Numbers in parentheses refer to enantiomeric excesses obtained
when the reaction was conducted with 10 mol % of Cu(S)-t-Bu-
box)(SbF6)2 (2a), CH2Cl2, -78 °C.
Table 11. Catalyst Recycling in a Hetero Diels-Alder Reaction
with an Apparently Insoluble Catalyst (Eq 35)
a Reaction conditions: (a) HCl/MeOH, reflux; (b) PPTS (cat.),
acetone/H2O (4:1), reflux; (c) NaBH4, EtOH, 25 °C; (d) TFA, CH2Cl2,
25 °C; (e) H2 (300 psi), Pd/C (cat.), EtOAc, 25 °C; (f) H2 (550 psi),
t
Rh/C (cat.), EtOAc, 25 °C; (g) 10 mol % OsO4, NMO, BuOH/H2O
(10:1); (h) RuCl3‚(H2O)3 (7 mol %); NaIO4, CH3CN/EtOAc/H2O (3:
3:1), 0 °C.
cyclea
time (h)
% yield
endo/exo
% ee
1
2
3
4
5
<1.0
1.1
89
98
91
98
92
>99:1
>99:1
>99:1
>99:1
>99:1
96
96
96
93
95
metal catalyzed reduction of the double bond introduced one
or two additional stereogenic centers into the molecule, depend-
ing on the substrate. For the trisubstituted olefin, Pd/C was
sufficient to effect catalytic hydrogenation (Scheme 2, step e,
R1 ) Me, R2 ) H), but for the hindered tetrasubstituted enol
ether (R1 ) R2 ) Me), Rh/C was the required catalyst (step f).
1.5
2.3
4.0
a See text.
Other reactions that preserved the integrity of the cyclic ether
were investigated. Diastereoselection was essentially complete
in all manipulations of the enol double bond due to the influence
exerted by the proximal stereogenic center (dr >20:1). Transition
(46) R ) Me (32a) and Ph (32b): (a) Irwin, A. J.; Jones, J. B. J. Am.
Chem. Soc. 1977, 99, 556. R ) i-Pr: (b) Paquette, L. A.; Dahnke, K.;
Doyon, J.; He, W.; Wyant, K.; Friedrich, D. J. Org. Chem. 1991, 56, 6199.

EnantioselectiVe Synthesis of Dihydropyrans
J. Am. Chem. Soc., Vol. 122, No. 8, 2000 1645
Scheme 3a
a Reaction conditions: (a) 5 mol % Sm(OTf)3, MeOH, reflux; (b)
LiOH, THF/H2O, 25 °C; (c) R-(+)-R-methylbenzylamine, CH2Cl2, 25
Oxidation of the enol bond was feasible: the R-hydroxy
lactone 35 was obtained stereoselectively in 53% yield under
standard dihydroxylation conditions (cat. OsO4, NMO), presum-
ably via the intermediacy of the R,â-dihydroxy phosphonate
36 (step g). Such a species was not observed spectroscopically
(1H NMR), but it was suspected that the equivalent of trialkyl-
amine base generated during the course of the reaction was
responsible for breakdown of the putative diol intermediate.
Oxidation using the conditions of Shing,48 which neither employ
nor generate endogenous base, allowed for isolation of the diol
36, mp 112-114 °C, again with high diastereoselection (step
h). Subjection of 36 to the action of NaHCO3 (MeOH, 25 °C,
5 min) resulted in elimination of dimethyl phosphite, generating
the same lactone obtained under the OsO4/NMO conditions
(where MeOH had been incorporated at the anomeric carbon).
°C
Scheme 4a
Bicyclic enol phosphonates were also amenable to synthetic
manipulation. Catalytic amounts of Sm(OTf)3 were sufficient
to induce the same type of solvolysis reaction performed above
(Scheme 3, step a). Good yields were obtained for the acetal
esters 37 and 38, products that might potentially serve as
protected lactones. In the case of the alkoxy-substituted adduct,
solvolysis further resulted in elimination of MeOH to stereo-
selectively give an R,â-unsaturated ester.
a Reaction conditions: (a) O3, CH2Cl2/MeOH, -78 °C; NaBH4, 25
°C; (b) 1 M aq HCl/THF, 65 °C; (c) TPAP (cat.), NMO, 4 Å mol
sieves, CH2Cl2, 25 °C.
The anomers of the â-methyl-substituted ester 37 were
separated by flash chromatography and the trans isomer was
saponified. The solid ammonium carboxylate 39 was obtained
after acidification and treatment with (R)-(+)-R-methylbenzyl-
amine (steps b and c). Recrystallization gave the pure salt, mp
115-117 °C, in 75% yield for two steps and an X-ray crystal
structure of the product established the configuration of the
carboxylate’s methyl-bearing stereogenic center.49
Figure 2. Secondary orbital interaction (dienophile(HOMO) f diene-
(LUMO)) for the endo transition structure.
ditions of R-keto esters and acyl phosphonates catalyzed by bis-
(oxazoline) Cu(II) complexes.50
Reaction Stereochemistry
The absolute configurations of 26b and 26c were secured in
the following manner (Scheme 4): ozonolysis of the enol double
bond with reductive workup (O3, CH2Cl2/MeOH, -78 °C;
NaBH4), hydrolysis of the resultant lactol ethyl ether (HCl, THF/
H2O, 65 °C), and lactol oxidation (TPAP, NMO, CH2Cl2, 25
°C) afforded lactones 40b and 40c in low yield. A comparison
of optical rotations to those in the literature demonstrated that
the sense of asymmetric induction was the same for cycload-
Mechanistic studies on reactions of R,â-unsaturated carbonyls
with alkyl enol ethers support a concerted asynchronous reaction
pathway.51 With regard to simple diastereoselection, the con-
sistent preference for the cis-2,4-substituted dihydropyran
products is consistent with an endo transition state. A secondary
orbital interaction between the oxadiene C-2 (LUMO) and dien-
ophile OR (HOMO) has been proposed as a stabilizing influence
in these processes (Figure 2).52 Houk has performed DFT
calculations for the reaction of acrolein and methyl vinyl ether
that support the experimental endo transition state preference.53
In addressing the subject of enantiofacial selectivity, we have
noted that [Cu((S,S)-t-Bu-box)](X)2 complexes 1a and 2a exhibit
stereoregular behavior across a number of reaction families. For
substrates that can engage in bidentate coordination to the chiral
Lewis acid, there is evidence supporting a reaction pathway
proceeding via a distorted square-planar Cu(II)‚substrate com-
(47) By verifying the absolute configuration of 7a, it was possible to
chemically correlate cycloadducts 10, 11-endo, and 11-exo. Lewis acid-
catalyzed alcohol exchange (Sm(OTf)3, EtOH) at the anomeric carbon for
10, 12-endo, and 12-exo generated compound 7a as a cis/trans mixture.
Comparison of retention times by GC and chiral HPLC (with co-injection
of authentic samples) allowed stereochemical assignment. The correlations
demonstrate that both heterodienophiles approach 6a from the same
enantioface in the endo transition state. The configuration at the methyl-
bearing stereocenter is opposite for the exo product derived from tert-butyl
vinyl ether (12-exo). For further details, see the Supporting Information.
(48) (a) Shing, T. K. M.; Tai, V. M. F.; Tan, E. K. W. Angew. Chem.,
Int. Ed. Engl. 1994, 33, 2312. (b) Shing, T. K. M.; Tam, E. K. W.; Tai, V.
W. F.; Chung, I. H. F.; Jiang, Q. Chem. Eur. J. 1996, 2, 50.
(50) 40b: (a) Helmchen, G.; Nill, G. Angew. Chem., Int. Ed. Engl. 1979,
18, 66. 40c: (b) King, C.-H. R.; Poulter, C. D. J. Am. Chem. Soc. 1982,
104, 1413.
(49) See the Supporting Information. The authors thank Mr. Kevin R.
Campos for solving this crystal structure.
(51) Desimoni, G.; Bamba, A.; Monticelli, M.; Nicola, M.; Tacconi, G.
J. Am. Chem. Soc. 1976, 98, 2947.

1646 J. Am. Chem. Soc., Vol. 122, No. 8, 2000
EVans et al.
reaction and analogous observations in catalysis of the glyoxyl-
ate ene reaction (eq 39),57 coupled with the dearth of structural
and mechanistic information on this phenomenon, warranted a
more detailed investigation.58
Catalyst Structure. With the goal of understanding the
coordination chemistry of the Cu(II) center in complexes 2a-
c, an X-ray crystallographic study of the hydrated derivatives
of those complexes was undertaken. Crystals of the derived bis-
(aquo) complexes 41a-c were obtained upon exposure of
dichloromethane solutions of 2a-c to atmospheric moisture.59
The structural details of those complexes revealed an interesting
trend that is relevant to the experimental results summarized
above (Figure 4). In tert-butyl-substituted bis(aquo) complex
41a, the Cu(II) center was characterized by a distorted square-
planar geometry. The distortion of the ligated water molecules
was away from the oxazoline substituents an average of +33.3°.
The water ligands of the analogous isopropyl-substituted bis-
(aquo) complex 41c displayed a similar distortion from square
planarity, but the magnitude of the tilt was significantly
smaller: an average of +7.0°. This change suggested that the
origin of distortion observed for 41a was steric, not electronic,
in nature. Such effects are well-precedented in the chemistry
of copper(II).60 Finally, phenyl-substituted bis(aquo) complex
41b also exhibited noticeable distortion from square planarity;
however, in that case the water molecules tilted toward the
oxazoline substituents an average of -9.3°.61
If the bidentate substrates undergoing actiVation adopt
distortions that resemble those obserVed in the X-ray structures,
substantial differences in the local enVironment of the two
prochiral faces might be expected. The enantioselectivity trends
for both the glyoxylate ene and hetero Diels-Alder reaction
placed [Cu((S,S)-i-Pr-box)](X)2 complexes in the middle of the
spectrum, between the [Cu((S,S)-t-Bu-box)](X)2 and [Cu((S,S)-
Ph-box)](X)2 complexes. As evidenced by Figure 4, the
structural distortion for the isopropyl complex was also inter-
mediate to that exhibited by the phenyl or tert-butyl-derived
complexes. While it has not been possible to establish a causal
link between these two phenomena, the fact that the two trends
mirror each other is nonetheless interesting.
Figure 3. Calcualted PM3 structures of [Cu(t-Bu-box)(6a)]2+ and [Cu-
(t-Bu-box)(keto ester)]2+
.
plex. X-ray crystallography, double stereodifferentiating reac-
tions, PM3 semiempirical calculations, and EPR spectroscopy
have supported the viability of such an intermediate.54 Further,
the enantiofacial bias exerted by 1a and 2a is rigidly consistent
in reactions of compounds that may participate in chelation:
N-acyl imides, glyoxylate and pyruvate esters, acyl phospho-
nates, R-keto esters, and alkylidene malonates. The results of
semiempirical calculations (PM3) for catalyst‚substrate com-
plexes illustrated in Figure 3 suggest that the analysis described
above may be reasonably extended to the cycloaddition chem-
istry detailed herein.55
In contrast, the mechanistic details of employing [Cu((S,S)-
Ph-box)](X)2 complexes 1b or 2b in these reactions had not
been scrutinized. The first report of a t-Bu/Ph enantioselectivity
reversal was made by Jørgensen and Johannsen in their studies
of normal demand hetero Diels-Alder reactions between
glyoxylate esters and dienes.56 In a series of publications, the
authors proposed that a tetrahedral copper center was responsible
for the turnover in absolute stereochemistry; the configuration
of the product was the support provided for this hypothesis.
Our results in the inverse electron demand hetero Diels-Alder
(52) For a summary, see: Boger, D. l. In ComprehensiVe Organic
Synthesis; Trost, B. M., Fleming, I., Eds.; Pergamon Press: Oxford, 1991;
Vol. 5, pp 451-512.
Mechanistic Studies. While we have provided calculated
structures of putative intermediates, we recognized the need to
consider not only the nature of the catalyst-substrate complexes
but interactions between that binary complex and the other
reacting partner. To facilitate this analysis, we considered in
the extreme two possible transition structures that differ in metal
center geometry (Figure 5). A change in geometry at copper
from square planar to tetrahedral was expected to expose the
(53) Liu, J.; Niwayama, S.; You, Y.; Houk, K. N. J. Org. Chem. 1998,
63, 1064.
(54) X-ray crystallography: (a) Reference 19b. (b) Evans, D. A.; Rovis,
T.; Kozlowski, M. C.; Tedrow, J. T. J. Am. Chem. Soc. 1999, 121, 1994.
Double stereodifferentiating experiments: (c) Evans, D. A.; Miller, S. J.;
Lectka, T. J. Am. Chem. Soc. 1993, 115, 6460. PM3 and EPR experi-
ments: (d) Reference 9.
(55) Geometry optimizations were performed at the PM3(tm) level using
the Spartan Semiempirical Program 5.0 (Wavefunction Inc., Irvine, CA
92612) on a Silicon Graphics Impact 10000 (195 MHz, 128 M RAM)
running IRIX 6.2. Calculations were performed without counterions or
solvent using these parameters: optcycle ) 5000, maxcycle ) 10000, charge
) 2, multiplicity ) 2. Calculations converged (energy difference between
cycles <0.0005 kcal/mol) in e4 h CPU time.
(56) (a) Johannsen, M.; Jørgensen, K. A. J. Org. Chem. 1995, 60, 5757.
(b) Johannsen, M.; Jørgensen, K. A. Tetrahedron 1996, 52, 7321. (c)
Johannsen, M.; Yao, S.; Jørgensen, K. A. Chem. Commun. 1997, 2169. (d)
Johannsen, M.; Yao, S.; Graven, A.; Jørgensen, K. A. Pure Appl. Chem.
1998, 70, 1117. (e) Yao, S. L.; Johannsen, M.; Audrain, H.; Hazell, R. G.;
Jørgensen, K. A. J. Am. Chem. Soc. 1998, 120, 8599. (f) Reference 11.
(57) Evans, D. A.; Burgey, C. S.; Paras, N. A.; Vojkovsky, T.; Tregay,
S. W. J. Am. Chem. Soc. 1998, 120, 5824.
(58) Other examples of an enantioselectivity turnover from t-Bu to Ph
bis(oxazoline): (a) Ghosh, A. K.; Mathivanan, P.; Cappiello, J.; Krishnan,
K. Tetrahedron: Asymmetry 1996, 7, 2165. (b) Sibi, M. P.; Ji, J.; Wu, J.
H.; Guertler, S.; Porter, N. A. J. Am. Chem. Soc. 1996, 118, 9200-9201.
(c) Yao, S.; Johannsen, M.; Jørgensen, K. A. J. Chem. Soc., Perkin Trans.
1 1997, 2345.
(59) The authors thank David M. Barnes (41a) and Dr. Christopher S.
Burgey (41b,c) for obtaining these crystals and Mr. Kevin R. Campos for
solving the structures.
(60) Hathaway, B. J.; Billing, D. E. Coord. Chem. ReV. 1970, 5, 143.

EnantioselectiVe Synthesis of Dihydropyrans
J. Am. Chem. Soc., Vol. 122, No. 8, 2000 1647
Figure 4. X-ray crystal structures of [Cu(t-Bu-box)(H2O)2](SbF6)2 (41a), [Cu(Ph-box) (H2O)2](SbF6)2 (41b), and [Cu(i-Pr-box)(H2O)2](SbF6)2
(41c) with selected bond angles.
endo selective catalyst with the sterically demanding cyclohex-
ene (endo:exo 95:5, 94% ee).
The manifestation of nonlinear effects in asymmetric catalysis
can have important consequences for the elucidation of reaction
mechanism.64 The enantiomeric excess of the adduct dihydro-
pyran for the hetero Diels-Alder reaction between acyl phos-
phonate 6c and ethyl vinyl ether was monitored as a function
of enantiomeric composition of the catalyst (Figure 6). A linear
relationship between the enantiomeric excess of the catalyst and
Figure 5. Tetrahedral and square planar transition structures for hetero
Diels-Alder reactions of acyl phosphonates with enol ethers.
that of the product was observed, thereby suggesting that neither
catalyst aggregation nor dimer formation was present. Nonlinear
effects were similarly absent in the catalyzed glyoxylate ene
opposite enantioface for reaction, by virtue of the altered spatial
relationship between the pendant oxazoline substituent (R2) and
the heterodienophile π-system.62
reaction; however, the successful isolation of crystals of Cu-
((S,S)-Ph-box)2(OTf)2 demonstrated that catalyst sequestration
by way of 2:1 ligand-metal complexation was feasible, but not
kinetically significant under the reaction conditions.65 The lack
of nonlinear effects in this investigation contrasted with other
systems employing divalent transition metal bis(oxazoline) and
bis(imine) complexes in which substantial asymmetric ampli-
fication was reported.66
Considering the significant role that electronic effects can
play in asymmetric reactions,67 substituted phenyl-bis(oxazo-
line) ligands of varying electronic character were synthesized
and evaluated in catalytic hetero Diels-Alder reactions (Table
12).68 While no significant differences in enantioselectivity were
observed for ligands with electron-poor or -rich aryl moieties,
the contribution of the electronic effects was not discounted.
Ample precedent exists for enantioselective processes in which
The finding that 2b-catalyzed cycloadditions of crotonyl
phosphonate 6a and methyl, ethyl, and tert-butyl vinyl ethers
were all highly endo selective was of particular interest to us.
Houk and co-workers have found experimental and computa-
tional evidence that indicated a preferred s-trans enol ether
conformation in the transition state of inverse electron demand
hetero Diels-Alder reactions.53 In the s-trans conformation,
severe steric interactions between the approaching heterodieno-
phile and activated catalyst-substrate complex would be
expected if the copper(II) center was tetrahedral (R1 T R2).
The absence of such a steric effect, as evidenced by the
uniformly high diastereoselectivities, argued against a change
to tetrahedral geometry.63 Application of the same test to the
glyoxylate ene reaction is also instructive, since 1b is a highly
(62) For experimental verification of this effect in Diels-Alder reactions
of Cu(II) vs Zn(II) Lewis acids, see: Evans, D. A.; Kozlowski, M. C.;
Tedrow, J. S. Tetrahedron Lett. 1996, 37, 7481.
(61) (a) We have considered two plausible sources for this distortion:
electrostatic effects and hydrogen bonding. For hydrogen bonding of water
with benzene, see: Suzuki, S.; Green, P. G.; Bumgarner, R. E.; Dasgupta,
S.; Goddard, W. A.; Blake, G. A. Science 1992, 257, 942. In that case the
separation of the center of mass (C6H6-H2O) was 3.32 Å; the distance
from the water oxygen to the phenyl centroid is ca. 3.9 Å for 41b. (b)
Attempts to model this structure at the PM3 level (omitting the SbF6
counterions) successfully reproduced the direction of distortion, but not the
magnitude (ca. 30°).
(63) The above discussion relies on reactions proceeding through a
concerted, asynchronous transition state, but at this time it is not possible
to exclude a two-step mechanism. Tietze and Evans have both provided
experimental evidence for concerted cycloadditions: (a) Tietze, L. F.; Bratz,
M.; Machinek, R.; Kiedrowski, G. v. J. Org. Chem. 1987, 52, 1638. (b)
Reference 11a. Even if the reaction occurs in two steps, the uniformally
high diastereoselectivities suggest a highly ordered transition state.
(64) Girard, C.; Kagan, H. B. Angew. Chem., Int. Ed. Engl. 1998, 37,
2923.

1648 J. Am. Chem. Soc., Vol. 122, No. 8, 2000
EVans et al.
Figure 7. Eyring plots for the hetero Diels-Alder reaction of 6a and
ethyl vinyl ether catalyzed by [Cu(t-Bu-box)](OTf)2 (1a) and [Cu(Ph-
box)](OTf)2 (1b).
enantiomer) versus reciprocal temperature, demonstrating that
only two diastereomeric transition states were responsible for
the observed enantiomer ratios.70 Consequently, competing metal
geometries were effectively excluded.
Figure 6. Study of nonlinear effects in hetero Diels-Alder reactions
employing catalyst 2b.
Table 12. Electronic Effect of Phenyl Bis(oxazoline) Ligand in the
Catalyzed Hetero Diels-Alder Reaction (Eq 41)
In our previous studies of carbocyclic Diels-Alder and
pyruvate aldol reactions catalyzed by Cu(box) complexes,
mechanistic evidence pointed to the intermediacy of a square-
planar catalyst‚substrate complex that was responsible for the
observed enantioselection. In both of those reactions, and now
with the hetero Diels-Alder reaction catalyzed by 1a and 1b,
Arrhenius behavior was observed, implying a strong resistance
of the metal center to geometric deformation (i.e., square-planar
f tetrahedral interconversion). In contrast, the Cu(pybox)-
catalyzed aldol reaction, while highly enantioselective at -78
°C, did not exhibit Arrhenius behavior. A low energetic barrier
for the interconversion of square-pyramidal and trigonal-
bipyramidal geometries was postulated as the source of the poor
temperature-enantioselectivity profile. The implication of these
results is clear: if the enantioselectivity turnover in the hetero
Diels-Alder reaction is a consequence of a new metal center
geometry, that geometry must react to the complete exclusion
of all others over a 100 °C temperature range. Such a scenario
appears unlikely considering the X-ray crystal structures 41a
and 41b that indicated a reduced propensity of phenyl(box)
copper complexes to distort from square planarity relative to
their tert-butyl counterparts. A similar conclusion was recently
reached by Davies and Deeth in the course of their computa-
tional study of bis(oxazoline) Cu(II) complexes.71
If a square-planar catalys-substrate complex is operative in
reactions mediated by 1b, then approach of the dienophile must
occur syn to the oxazoline substituent. While it is difficult to
rationalize such an occurrence, a model that does fit the
experimental evidence stipulates that the π surface of the phenyl
ring is able to stabilize the partial positive charge that arises
from the asynchronous transition state (Figure 8).72 Such a
transition structure is not possible for 1,1-disubstituted enol
ethers such as R-trialkylsiloxystyrene due to the abutment of
the enol ether substituent with the oxazoline phenyl ring: as
noted previously, those heterodienophiles show the same facial
selectivity using both tert-butyl and phenyl bis(oxazoline)
catalysts.73 This proposal should be construed as speculative;
entry
X
% ee
1
2
3
Cl
89
93
93
H
OMe
dipole-dipole and van der Waals attractions are implicated, but
unaffected by pertubations in the π-donor capability of the
phenyl group.69
As indicated previously, Cu(box) complexes exhibited a
favorable temperature-enantioselectivity profile for catalytic
hetero Diels-Alder reactions of acyl phosphonates and keto
esters, an observation with consequences for the understanding
of the stereochemical outcome of these reactions. Eyring plots
for cycloadditions conducted with 1a and 1b (Figure 7) both
showed a linear dependence of ln(% major enantiomer/% minor
(65) Barnes, D. M. Ph.D. Thesis, Harvard University, 1997.
(66) (a) Evans, D. A.; Lectka, T.; Miller, S. J. Tetrahedron Lett. 1993,
34, 7027. (b) Kanemasa, S.; Oderaotoshi, Y.; Sakaguchi, S.; Yamamoto,
H.; Tanaka, J.; Wada, E.; Curran, D. P. J. Am. Chem. Soc. 1998, 120, 3074.
(c) Evans, D. A.; Kozlowski, M. C.; Murry, J. A.; Burgey, C. S.; Campos,
K. R.; Connell, B. T.; Staples, R. J. J. Am. Chem. Soc. 1999, 121, 669.
(67) Palucki, M.; Finney, N. S.; Pospisil, P. J.; Guler, M. L.; Ishida, T.;
Jacobsen, E. N. J. Am. Chem. Soc. 1998, 120, 948.
(68) The ligands were synthesized according to the method of ref 19b.
For synthessis of the amino alcohols, see: (a) Chang, H. T.; Sharpless, K.
B. Tetrahedron Lett. 1996, 37, 3219. (b) Kawasaki, K.; Katsuki, T.
Tetrahedron 1997, 53, 6337.
(69) (a) Oppolzer, W.; Kurth, M.; Reichlin, D.; Chapuis, C.; Mohnhaupt,
M.; Moffatt, F. HelV. Chim. Acta 1981, 64, 2802. (b) Evans, D. A.;
Chapman, K. T.; Hung, D. T.; Kawaguchi, A. T. Angew. Chem., Int. Ed.
Engl. 1987, 26, 1184. (c) Quan, R. W.; Li, Z.; Jacobsen, E. N. J. Am. Chem.
Soc. 1996, 118, 8156.
(70) Buschmann, H.; Scharf, H.-D.; Hoffmann, H.; Esser, P. Angew.
Chem., Int. Ed. Engl. 1991, 30, 477.
(71) Davies, I. W.; Deeth, R. J.; Larsen, R. D.; Reider, P. J. Tetrahedron
Lett. 1999, 40, 1233.

EnantioselectiVe Synthesis of Dihydropyrans
J. Am. Chem. Soc., Vol. 122, No. 8, 2000 1649
Figure 8. Proposed π-stabilization of an asynchronous hetero Diels-
Alder transition state.
however, the accumulated experimental evidence provides more
support for it than for a tetrahedral copper intermediate.
Figure 9. Qualitative model of the hetero Diels-Alder transition state
showing stereochemical relationships between reactants.
Conclusions
The asymmetric hetero Diels-Alder reaction of R,â-unsatur-
ated carbonyls with electron-rich alkenes catalyzed by C2-
symmetric bis(oxazoline) Cu(II) complexes provides an expe-
dient entry into enantioenriched dihydropyrans. A number of
R,â-unsaturated acyl phosphonates and â,γ-unsaturated R-keto
esters and amides have been successfully employed as hetero-
dienes, while enol ethers and sulfides and certain ketone silyl
enol ethers have functioned well as heterodienophiles. Of
particular note is the development of a simple reaction protocol
that employs a solid air-stable catalyst, low catalyst loadings,
and convenient reaction temperatures, conditions that render
these reactions particularly attractive from a preparative stand-
point.
the diastereoselectivity for all these processes can be accounted
for by an endo transition structure (Figure 9).
The derived cycloadducts may be transformed to useful chiral
building blocks. Under acidic conditions, desymmetrized glutaric
acid derivatives are obtained, while oxidation or reduction leaves
the heterocycle intact and delivers highly functionalized tet-
rahydropyran products. The absolute configurations were de-
termined for nine cycloadducts by chemical or crystallographic
methods; the stereochemical outcome for all reactions catalyzed
by [Cu((S,S)-t-Bu-box)](X)2 complexes 1a and 2a is consistent
with the intervention of a distorted square-planar catalyst-
substrate complex. Structural and mechanistic studies have been
employed to study the source of asymmetric induction for
reactions catalyzed by [Cu((S,S)-Ph-box)](X)2 complexes 1b and
2b; no evidence has been found to support the intervention of
a tetrahedral copper species in reactions catalyzed by phenyl
bis(oxazoline) copper(II) complexes.
The high diastereoselectivity for catalyzed hetero Diels-Alder
reactions is a result of frontier orbital control and/or electrostatic
effects that preferentially place the OR substituent in proximity
with the heterodiene carbonyl carbon (endo orientation). This
stereocontrol element has been implicated in related conjugate
addition reactions to unsaturated imides and azaimides. Our
collective experimental observations suggest that enol amination
reactions (eq 42),74 Mukaiyama Michael reactions (eq 43),73
This investigation has attempted to provide some evidence
to support the notion that cycloadditions catalyzed by chiral
copper complexes may be employed to efficiently and selec-
tively assemble heterocycles. It is hoped that the reactions
detailed herein will be of some use in the enantioselective
construction of pyran rings, prevalent structural units in organic
chemistry.
Acknowledgment. Support for this research was provided
by the NSF. Graduate fellowships from the NSF (J.S.J. and
E.J.O.), Pharmacia and Upjohn (J.S.J.), and Roche (J.S.J.) are
gratefully acknowledged. The NIH BRS Shared Instrumentation
Grant Program 1-S10-RR04870 and the NSF (CHE 88-14019)
are acknowledged for providing NMR facilities. The amino
alcohols and acids utilized in the preparation of the box ligands
were generously provided by NSC Technologies.
Supporting Information Available: Complete experimental
procedures and characterization of new compounds (PDF). This
material is available free of charge via the Internet at
http://pubs.acs.org.
JA992175I
(72) For an excellent discussion on cation π interactions, see: Dougherty,
D. A. Science 1996, 271, 163.
(73) Since experiments suggest that reactions with R-trialkylsiloxystyrene
are not concerted, antiperiplanar approach of the nucleophile should also
be considered; however, reactions with E and Z propionate silyl enol ethers
and ketene acetals in the study of the Mukaiyama Michael reaction have
provided some evidence that these additions proceed via a synclinal
transition state. Evans, D. A.; Willis, M. C.; Johnston, J. N. Org. Lett. 1999,
1, 865.
and hetero Diels-Alder reactions are mechanistically related;
(74) Evans, D. A.; Johnson, D. S. Org. Lett. 1999, 1, 595.