C2-symmetric copper(II) complexes as Chiral Lewis acids. Scope and mechanism of catalytic enantioselective aldol additions of enolsilanes to (benzyloxy)acetaldehyde

J. Am. Chem. Soc. 1999, 121, 669-685

669

C₂-Symmetric Copper(II) Complexes as Chiral Lewis Acids. Scope

and Mechanism of Catalytic Enantioselective Aldol Additions of

Enolsilanes to (Benzyloxy)acetaldehyde

David A. Evans,* Marisa C. Kozlowski, Jerry A. Murry, Christopher S. Burgey,

Kevin R. Campos, Brian T. Connell, and Richard J. Staples

Contribution from the Department of Chemistry and Chemical Biology, HarVard UniVersity,

Cambridge, Massachusetts 02138

ReceiVed August 19, 1998

Abstract: C₂-Symmetric bis(oxazolinyl)pyridine (pybox)-Cu(II) complexes have been shown to catalyze

enantioselective Mukaiyama aldol reactions between (benzyloxy)acetaldehyde and a variety of silylketene acetals.

The aldol products are generated in high yields and in 92-99% enantiomeric excess using as little as 0.5 mol

% of chiral catalyst [Cu((S,S)-Ph-pybox)](SbF₆)₂. With substituted silylketene acetals, syn reaction diastereo-

selection ranging from 95:5 to 97:3 and enantioselectivities g95% are observed. Investigation into the reaction

mechanism utilizing doubly labeled silylketene acetals indicates that the silyl-transfer step is intermolecular.

Further mechanistic studies revealed a significant positive nonlinear effect, proposed to arise from the selective

formation of the [Cu((S,S)-Ph-pybox)((R,R)-Ph-pybox)](SbF₆)₂2:1 ligand:metal complex. A stereochemical

model is presented in which chelation of (benzyloxy)acetaldehyde to the metal center to form a square pyramidal

copper intermediate accounts for the observed sense of induction. Support for this proposal has been obtained

from double stereodifferentiating reactions, EPR spectroscopy, ESI spectrometry, and, ultimately, the X-ray

crystal structure of the aldehyde bound to the catalyst. The C₂-symmetric bis(oxazolinyl)-Cu(II) complex

[Cu((S,S)-tert-Bu-box)](OTf)₂is also an efficient catalyst for the aldol reaction, but the scope with this system

is not as broad.

Introduction

(<10 mol %). In part, the lack of structural data on many of

the relevant catalyst-aldehyde complexes has inhibited further

catalyst refinement. Indeed, some of the fundamental control

elements for these reactions are just being revealed.

The development of a general enantioselective aldol addition

reaction has been an enduring problem in organic chemistry

for nearly 25 years.¹Seminal advances have been realized in

the development of high levels of reaction diastereoselection,

while improvements in chiral auxiliary design have led to the

achievement of absolute stereochemical control for many of

these reactions.¹Nevertheless, the broad extension of high levels

of diastereoselectivity and enantioselectivity to catalytic aldol

reactions has not been trivial. The most notable achievements

that have been made in this area have focused on the addition

of enolsilanes to aldehydes²through catalysis by chiral Lewis

acids (*L_nM⁺, eq 1).³Excellent progress in the development

The realization of high enantioselectivity for the catalyzed

aldol reaction necessarily relies on effective channeling of the

reactants through a transition state that is substantially lower in

energy than competing diastereomeric transition-states. For the

process at hand, a high level of transition-state organization is

required, necessitating control of factors that include (A) mode

of binding (η²vs η¹) of the carbonyl group to the Lewis acid;

(B) the regiochemistry of complexation to the two available

CdO lone pairs; and (C) the establishment of a fixed diaste-

reofacial bias, thereby biasing enol/enolate addition to one of

the two carbonyl π-faces. The discovery of effective strategies

for controlling the conformation of the Lewis acid-bound

aldehyde lies at the heart of current advances in chiral catalyst

design in this area. Recent investigations have sought to

(3) For a comprehensive review of catalytic enantioselective aldol

reactions: Nelson, S. G. Tetrahedron Asymmetry 1998, 9, 357-389.

(4) (a) Carreira, E. M.; Singer, R. A.; Lee, W. J. Am. Chem. Soc. 1994,

117, 8837-8838. (b) Keck, G. E.; Krishnamurthy, D. J. Am. Chem. Soc.

1995, 117, 2363-2364. (c) Mikami, K.; Matsukawa, S. J. J. Am. Chem.

Soc. 1994, 116, 4077-4078. (d) Kobayashi, S.; Uchiro, H.; Shiina, I.;

Mukaiyama, T. Tetrahedron 1993, 49, 1761-1772. (e) Corey E. J.; Cywin,

C. L.; Roper, T. D. Tetrahedron Lett. 1992, 33, 6907-6910. (f) Parmee,

E. R.; Hong, Y.; Tempkin, O.; Masamune, S. Tetrahedron Lett. 1992, 33,

1729-1732. (g) Kiyooka, S.; Kaneko, Y.; Kume, K. Tetrahedron Lett. 1992,

33, 4927-4930. (h) Furuta, K.; Maruyama, T.; Yamamoto, H. J. Am. Chem.

Soc. 1991, 113, 1041-1042. (i) Parmee, E. R.; Tempkin, O.; Masamune,

S.; Abiko, A. J. Am. Chem. Soc. 1991, 113, 9365-9366. (j) Yanagisawa,

A.; Matsumoto, Y.; Nakashima, H.; Asakawa, K.; Yamamoto, H. J. Am.

Chem. Soc. 1997, 119, 9319-9320 and references therein.

of enantioselective “Mukaiyama” aldol variants has been made;⁴

however, no individual catalyst system developed to date

tolerates substantial variation in both the nucleophilic and

electrophilic components while maintaining low catalyst loading

(1) (a) Heathcock, C. H. In Asymmetric Synthesis; Morrison, J. D., Ed.;

Academic Press: San Diego, CA 1984; Vol. 3, Chapter 2. (b) Evans, D.

A. Aldrichim. Acta 1982, 15, 23-32.

(2) (a) Mukaiyama, T.; Narasaka, K.; Banno, K. Chem. Lett. 1973, 1011-

1014. (b) Mukaiyama, T.; Banno, K.; Narasaka, K. J. Am. Chem. Soc. 1974,

96, 7503-7509. (c) Gennari, C. in ComprehensiVe Organic Synthesis; Trost,

B. M., Fleming, I., Eds; Pergamon Press: New York, NY, 1991; Vol. 2,

Chapter 2.4.

Published on Web 01/20/1999

670 J. Am. Chem. Soc., Vol. 121, No. 4, 1999

EVans et al.

incorporate additional stabilizing interactions such as hydrogen

bonding,⁵π-stacking,⁴or chelation, incorporated individually

or in concert, into the catalyst-aldehyde complex to provide a

highly defined carbonyl facial bias (Scheme 1). The focal point

The selectivity observed with both the Cu(II) box and pybox

complexes suggested that these catalysts would also be promis-

ing candidates for the Mukaiyama aldol reaction with aldehyde

substrates that could present the potential for catalyst chelation.

Indeed, accumulated evidence has indicated that chelation is a

critical control element in defining the catalyst-substrate

architecture for the Diels-Alder reaction (Scheme 2, A and B),⁶

and we postulated that extrapolation of these models to include

chelating aldehydes would furnish well-defined catalyst-

substrate complexes (Scheme 3, C and D). In continuing the

of the present investigation has been to discover chiral metal

complexes that exhibit a strong propensity toward substrate

chelation while also meeting the other criteria necessary for their

application to asymmetric catalysis of the aldol reaction.

Previous work from our laboratory has demonstrated that

bidentate bis(oxazolinyl) (box) (1 and 2)-Cu(II) and tridentate

bis(oxazolinyl)pyridine (pybox) (3 and 4)-Cu(II) complexes

can function as effective chiral Lewis acid catalysts in the

Diels-Alder reaction with substrates that can participate in

catalyst chelation (Scheme 2, eq 2).⁶Further studies revealed

theme of developing laboratory analogues of biosynthetic

pathways,⁸(benzyloxy)acetaldehyde⁹was selected as the initial

substrate to probe the utility of these Cu(II) complexes in the

Mukaiyama aldol reaction (eq 3), since this aldehyde may be

regarded as an equivalent to the acetate starter unit in polyacetate

biosynthesis. Moreover, the benzyloxy moiety provides a

convenient point for further elaboration of the resulting aldol

adducts. In this article, we document the use of copper(II)

complexes as effective enantioselective catalysts for the Mu-

kaiyama aldol reaction, where the aldehyde component is

activated through bidentate coordination, an organizational

feature not common to chiral Lewis acids previously reported

for this process;¹⁰furthermore, we provide direct structural

evidence to support the proposed model of stereochemical

induction.

Preliminary Results

Bis(oxazoline) Ligand Survey. The (S,S)-bis(oxazolinyl)

copper complexes 1 and 2 were initially evaluated as catalysts

in the addition of tert-butyl thioacetate trimethylsilylketene

acetal¹¹to (benzyloxy)acetaldehyde (eq 6). The bis(oxazolinyl)

copper complexes 1a-d were prepared by stirring a solution

of the (S,S)-bisoxazoline ligand 5¹²and Cu(OTf)₂(typically 10

mol %, ∼0.03 M in catalyst) in CH₂Cl₂(25 °C, 3 h) as

that these Cu(II) catalysts maintain excellent levels of reactivity

over a range of diene substrates, attesting to the high Lewis

acidity of these complexes.⁷

(8) Evans, D. A.; Clark, J. S.; Novack, V. J.; Sheppard, G. S.; Metternich,

R.; Ng, H. P.; Le, A.; Ennis, M. D.; Kim, A. S.; Ripin, D. H. B. J. Am.

Chem Soc., in press.

(5) (a) Corey, E. J.; Barnes-Seeman, D.; Lee, T. W. Tetrahedron Lett.

1997, 38, 1699-1702. (b) Corey, E. J.; Rohde, J. J. Tetrahedron Lett. 1997,

38, 37-40.

(9) (Benzyloxy)acetaldehyde is commercially available from Aldrich

Chemical Co. Alternatively, an inexpensive two-step synthesis from

2-butene-1,4-diol is amenable to large-scale preparations via bis-benzylation

(Garner, P.; Park, J. M. Synth. Commun. 1987, 17, 189-193), followed by

ozonolysis (Danishefsky, S. J.; DeNinno, M. P. J. Org. Chem. 1986, 51,

2615-2617).

(10) A preliminary account of this work has appeared: Evans, D. A.;

Murry, J. A.; Kozlowski, M. C. J. Am. Chem. Soc. 1996, 118, 5814-5815.

(11) Kuwajima, I.; Kato, M.; Sato, T. J. Chem. Soc., Chem. Commun.

1978, 478-479.

(12) (a) Evans, D. A.; Woerpel, K. A.; Hinman, M. M.; Faul, M. M. J.

Am. Chem. Soc. 1991, 113, 726-728. (b) Lowenthal, R. E.; Abiko, A.;

Masamune, S. Tetrahedron Lett. 1990, 31, 6005-6008. (c) Mu¨ller, D.;

Umbricht, B. W.; Pfaltz, A. HelV. Chim. Acta 1991, 74, 232-240. (d)

Denmark, S. E.; Nakajima, N.; Nicaise, J.-C.; Faucher, A. M.; Edwards, J.

P. J. Org. Chem. 1995, 60, 4884-4892. (e) Evans, D. A.; Peterson, G. S.;

Johnson, J. S.; Barnes, D. M.; Campos, K. R.; Woerpel, K. A. J. Org. Chem.

1998, 63, 4541-4544.

(6) (a) Evans, D. A.; Miller, S. J.; Lectka, T. C. J. Am. Chem. Soc. 1993,

115, 6460-6461. (b) Evans, D. A.; Lectka, T.; Miller, S. J. Tetrahedron

Lett. 1993, 34, 7027-7030. (c) Evans, D. A.; Murry, J. A.; von Matt, P.;

Norcross, R. D.; Miller, S. J. Angew. Chem., Int. Ed. Engl. 1995, 34, 798-

800. (d) Evans, D. A.; Kozlowski, M. C.; Tedrow, J. S. Tetrahedron Lett.

1996, 37, 7481-7484. (e) Evans, D. A.; Barnes, D. M. Tetrahedron Lett.

1997, 38, 57-58. (f) Evans, D. A.; Johnson, J. S. J. Org. Chem. 1997, 62,

786-787. (g) Evans, D. A.; Shaughnessy, E. A.; Barnes, D. M. Tetrahedron

Lett. 1997, 38, 3193-3194.

(7) These copper complexes are also effective enantioselective catalysts

for the hetero Diels-Alder and carbonyl-ene reactions: (a) Evans, D. A.;

Johnson, J. S. J. Am. Chem. Soc. 1998, 120, 4895-4896. Evans, D. A.;

Olhava, E. J.; Johnson, J. S.; Janey, J. M. Angew. Chem. 1998, 24, 3554-

3557. (b) Evans, D. A.; Burgey, C. S.; Paras, N. A.; Vojkovsky, T.; Tregay,

S. W. J. Am. Chem. Soc. 1998, 120, 5824-5825.

Aldol Addition of Enolsilanes to (Benzyloxy)acetaldehyde

J. Am. Chem. Soc., Vol. 121, No. 4, 1999 671

previously described (eq 4).^6aThe cationic hexafluoroantimonate

complex 2a was formed by halide abstraction from the pre-

formed [Cu((S,S)-tert-Bu-box)]Cl₂complex (6)^12ewith AgSbF₆,

followed by filtration through dry Celite (or a PTFE 0.45-µm

filter) to remove the precipitated AgCl (eq 5).^6c,12e

the pybox ligand with CuCl₂in CH₂Cl₂, followed by halide

abstraction with AgSbF₆and filtration¹⁴to remove the precipi-

tated AgCl (eq 8).^6c,10

The [Cu((S,S)-pybox)](SbF₆)₂complexes 4a-d also catalyze

the addition of tert-butyl thioacetate trimethylsilylketene acetal

to (benzyloxy)acetaldehyde (eq 9) with moderate to excellent

enantioselectivity (Table 2, entries 1-5, 62-99% ee). The [Cu-

(Ph-pybox)](SbF₆)₂complex (4c) was the most selective catalyst,

delivering the â-hydroxy ester with excellent enantioselection

within 15 min at -78 °C (entry 5, 99% ee). The analogous

triflate complex 3c was also highly enantioselective, but as

anticipated, the reaction was slower (entry 4, 96% ee).

Addition of (benzyloxy)acetaldehyde to a cooled solution

(-78 °C) of the catalyst, followed by subsequent dropwise

addition of the tert-butyl thioacetate-derived trimethylsilylketene

acetal, afforded the protected â-hydroxy ester (eq 6). Brief

treatment of this silyl ether with 1 N HCl in THF produced the

expected alcohol 7, the enantioselectivity of which was assayed

by chiral HPLC (Daicel OD-H column). The absolute config-

uration of the adduct was established by comparison of the

optical rotation with that in the literature.^4cThis ligand screen

revealed that the phenylglycine (Ph-box, 1c)- and valine (i-Pr-

box, 1b)-derived box complexes were poorly enantioselective

catalysts (Table 1, entries 1 and 2, 9% ee); however, the Cu-

(OTf)₂complexes of both the phenylalanine (Bn-box, 1d) and

tert-leucine (tert-Bu-box, 1a) bisoxazolines delivered the aldol

adduct with high enantioselectivity (entries 3 and 4). Use of

the corresponding hexafluoroantimonate complex [Cu((S,S)-tert-

Bu-box)](SbF₆)₂(2a), the optimal catalyst for the Diels-Alder

reaction, afforded the (S) product with only modest enantiose-

lectivity (Table 1, entry 5, e64% ee).

Reaction Optimization. Due to the superior enantioselec-

tivity exhibited by the [Cu((S,S)-Ph-pybox)](SbF₆)₂complex (4c)

(Table 2, entry 5), a study was initiated to explore the

(benzyloxy)acetaldehyde aldol reaction parameters with this

catalyst system. A survey of permissable solvents for this

reaction was undertaken. Solutions of catalyst 4c in the indicated

solvents were generated using the standard procedure. This

survey (-20 °C, 10 mol % 4c) revealed that employment of

solvents other than CH₂Cl₂led to either a significant decline in

enantioselectivity or no reaction (Table 3). The inferior results

obtained in this solvent screen are likely due to the limited

solubility of the [Cu((S,S)-Ph-pybox)](SbF₆)₂complex (4c) in

media other than CH₂Cl₂. An examination of the temperature

profile of the [Cu(Ph-pybox)](SbF₆)₂-catalyzed aldol reaction

Pyridine(bisoxazoline) Ligand Survey. Dichloromethane

solutions of the (S,S)-pybox ligands 8¹³were complexed with

Cu(OTf)₂to form blue solutions of the chiral triflate complexes

3a-d (eq 7). Preparation of the cationic [Cu((S,S)-pybox)]-

(SbF₆)₂complexes 4a-d was accomplished by precomplexing

(14) Dry Celite, cotton, or a PTFE 0.45-µm filter gave the best results.

Complete removal of the AgCl, as characterized by a clear solution, was

essential for obtaining high enantioselectivity.

(13) Nishiyama, H.; Kondo, M.; Nakamura, T.; Itoh, K. Organometallics

1991, 10, 500-508.

672 J. Am. Chem. Soc., Vol. 121, No. 4, 1999

EVans et al.

demonstrated that an increase in the reaction temperature is

accompanied by a significant decrease in enantioselectivity

(Table 2, entries 6-8). As these reactions are highly exothermic,

careful temperature control is critical to the maintenance of high

selectivity, especially when executing large-scale preparations

(vide infra).

The effect of transient amounts of water on catalyst perfor-

mance was also evaluated. Hydration does not appear to

significantly impact catalyst reactivity, as substitution of CuCl₂‚

2H₂O for CuCl₂during catalyst preparation afforded [Cu((S,S)-

Ph-pybox)](SbF₆)₂solutions with similar activity. Furthermore,

solutions (0.125 M) of the catalyst 4c may be stored without

loss of catalytic activity for up to 1 week at room temperature,

after which time a crystalline solid begins to form. This

precipitate has been characterized by X-ray crystallography to

be the catalytically inactive 2:1 ligand:copper complex (vide

infra). Catalyst loadings as low as 0.5 mol % may be employed

in the aldol reaction (eq 9) using this standard [Cu((S,S)-Ph-

pybox)](SbF₆)₂solution ([BnOCH₂CHO]₀) 2.5 M). Further

experiments demonstrated that the implementation of low

catalyst loadings is feasible with other substrates as well, the

acetate-derived nucleophiles being the most tolerant (vide

infra).

In ongoing efforts to exploit the utility of these acetoacetate

nucleophiles, the investigation of preparative scale versions of

these reactions was undertaken. These studies revealed that the

reactions employing the dioxolinone nucleophile (eq 12, 55

mmol) and Chan’s diene (eq 13, 40 mmol) were both prob-

lematic when conducted on a large scale (80-90% ee, 50-

60% yield). However, subsequent optimization studies on the

related reaction between (benzyloxy)acetaldehyde and 1,3-bis-

(trimethylsiloxy)-1-tert-butoxybuta-1,3-diene (14)¹⁹catalyzed by

[Cu(Ph-pybox)](SbF₆)₂(4c) established that preparative scale

(35.5 mmol) reactions could be performed to deliver the

corresponding product (as a mixture of keto-enol tautomers)

with high selectivity under slightly modified reaction conditions

(eq 14). Due to the exothermic nature of this Cu(II)-mediated

Reaction Scope: [Cu((S,S)-Ph-pybox)](SbF₆)₂(4c). The

scope of the nucleophilic component in the (benzyloxy)-

acetaldehyde aldol reaction with the catalyst 4c was initially

investigated (Table 4). The silylketene acetals derived from tert-

butyl thioacetate, ethyl thioacetate, and ethyl acetate reacted with

(benzyloxy)acetaldehyde in the presence of 0.5 mol % catalyst

to afford the respective â-hydroxy esters with excellent enan-

tioselectivity (entries 1-3, 98-99% ee). In a related series of

reactions, acetoacetate-derived enol derivatives were also evalu-

ated (eqs 12 and 13). The dioxolinone derivative 10¹⁵underwent

facile reaction with (benzyloxy)acetaldehyde, in the presence

of 5 mol % 4c, to provide the corresponding adduct 11 in 92%

ee and 94% yield (eq 12).¹⁶Low catalyst loadings (0.5 mol %)

could be employed with the more reactive Chan’s diene 12¹⁷

as the nucleophile, to afford, after reduction with Me₄NBH-

(OAc)₃,¹⁸the anti diol 13 (15:1 anti:syn) in 97% ee (eq 13).

aldol reaction, the optimal procedure required the slow addi-

tion of (benzyloxy)acetaldehyde to a -90 °C solution of the

catalyst (2 mol %, 0.011 M) and diene (i.e., inverse addition).

Desilylation under nonaqueous²⁰conditions (PPTS/MeOH)

followed by flash chromatography reproducibly afforded good

yields of the desired product (15)²¹with excellent enantiomeric

excess (eq 14, 99% ee, 85% yield). The major competing

reaction under these Lewis acidic conditions is the trimerization

of (benzyloxy)acetaldehyde, which can be kept under 15% using

these conditions. The Ph-pybox ligand routinely recovered

during chromatography (ca. 65%) was subsequently recrystal-

lized (EtOAc) for reuse in these Cu(II) catalyzed aldol reactions.

Diastereoselective anti¹⁸and syn reductions of 15 (eqs 15,

16) permit the efficient production of the synthetically valuable

(15) Sato, M.; Sunami, S.; Sugita, Y.; Kaneko, C. Chem. Pharm. Bull.

1994, 42, 839-845.

(16) The catalytic enantioselective addition of this nucleophile to

aldehydes has also recently been reported: Singer, R. A.; Carreira, E. M.

J. Am. Chem. Soc. 1995, 117, 12360-12361.

(19) (a) Molander, G. A.; Cameron, K. O. J. Am. Chem. Soc. 1993, 115,

5, 830-846. (b) This diene was utilized to minimize overreduction to the

triol which was observed during the syn reduction of the corresponding

methyl ester.

(20) The product is slightly soluble in water.

(21) Beck, G.; Jendralla, H.; Kesseler, K. Synthesis 1995, 1014-1018.

(17) Brownbridge, P.; Chan, T. H.; Brook, M. A.; Kang, G. J. Can. J.

Chem. 1983, 61, 688-693.

(18) Evans, D. A.; Chapman, K. T.; Carreira, E. M. J. Am. Chem. Soc.

1988, 110, 3560-3578.

Aldol Addition of Enolsilanes to (Benzyloxy)acetaldehyde

J. Am. Chem. Soc., Vol. 121, No. 4, 1999 673

polyacetate building blocks 16 and 17.²²Notably, a minor

modification of the conditions developed by Beck et al.²¹(see

Experimental Section) afforded the syn adduct with excellent

diastereoselectivity (>200:1 syn:anti).

Catalyzed Mukaiyama aldol processes are not generally highly

diastereoselective, and reactions employing propionate nucleo-

philes frequently suffer from either low diastereoselectivity or

poor reactivity. In contrast, the [Cu((S,S)-Ph-pybox)](SbF₆)₂

(4c)-catalyzed aldol reaction with these nucleophiles affords the

substituted adducts with high selectivity (Table 5). For example,

the (Z) silylketene acetal of ethyl thiopropionate provided the

syn aldol adduct in excellent diastereo- and enantioselectivity

(97:3 syn:anti, syn 97% ee, Table 5, entry 1). In contrast, the

corresponding (E) propionate silylketene acetal proved to be

an inferior substrate, requiring higher reaction temperature and

giving lower conversion and selectivity (86:14 syn:anti, syn 85%

ee, entry 2). The absolute and realtive stereochemistries were

confirmed by conversion of the known aldol product 18²³to

the syn adduct 19, which exhibited the expected opposite sign

of optical rotation (eq 17).

Attempts to employ non-alkyl-substituted silylketene acetals

as the nucleophilic component (eq 18, R¹* alkyl), such as

phenylacetate and benzyloxythioacetate, were unsuccessful

(Table 5, entries 8-10). On the other hand, 2-(trimethylsilyl-

oxy)furan²⁴underwent facile reaction with (benzyloxy)acetal-

dehyde in the presence of the [Cu(Ph-pybox)](SbF₆)₂catalyst

(4c) to afford the anti aldol adduct 21 in high yield with excellent

diastereo- and enantioselectivity (eq 20).²⁵The densely func-

The exact nature of the silyl component of the silylketene

acetal was not critical, as the (Z)-propionate-derived tert-

butyldimethylsilylketene acetal afforded similar selectivity (94:6

syn:anti, syn 96% ee, entry 3) in comparison to the trimethylsilyl

analogue (97:3 syn:anti, syn 97% ee, entry 1). Thus, additional

protecting group steps are obviated as this manipulation can be

merged with the copper-catalyzed Mukaiyama aldol reaction

to provide â-hydroxy carbonyl products with an intact, syntheti-

cally useful silyl ether protecting group. The use of silylketene

acetals with larger alkyl substituents is also permitted, as

evidenced by the production of the isobutyl-substituted adduct

i

(eq 17, R¹) Bu) with high diastereo- and enantioselectivity

tionalized product of this reaction is a versatile synthon for

natural product synthesis and also may be elaborated to unnatural

hexose derivatives.²⁴Although the inherent stereoconvergency

of this Cu(II)-catalyzed process permits access primarily to the

syn aldol adducts, anti diastereoselectivity can be accessed

through the use of the Sn(II)-box-catalyzed aldol reactions of

glyoxylate esters.²⁶Thus, the copper and tin catalysts together

provide a powerful entry into a diverse array of substituted aldol

products.

and good yield (Table 5, entry 4, 95:5 syn:anti, 95% ee, 85%).

While tert-butyl thioester- or ethyl ester-derived silylketene

acetals could also be employed, the results were not as favorable

as those obtained with ethyl thioester-derived nucleophiles

(Table 5, entry 1 vs entries 5-7). Regardless of the precise

nature of the ester substituent, the best selectivity and reactivity

for the substituted silylketene acetals were obtained when the

alkyl and OTMS moieties were disposed in an anti orientation

about the silylketene acetal double bond. This geometric

requirement was also evident in the analogous reaction of the

butyrolactone silylketene acetal, which afforded a highly selec-

tive syn aldol reaction with good control at both stereogenic

centers (eq 19).

While ester silylketene acetals were successfully utilized in

this reaction, thioester silylketene acetals offer several advan-

(24) Casiraghi, G.; Rassu, G. Synthesis 1995, 607-626.

(25) Anti diastereoselectivity was also observed with this nucleophile

in the Cu(II) catalyzed pyruvate aldol reaction; see the following article in

this issue Evans, D. A.; Burgey, C. S.; Kozlowski, M. C.; Tregay, S. W. J.

Am. Chem. Soc. 1999, 121, 686-699. The absolute and relative stereo-

chemistry of this product were secured by X-ray analysis of the corre-

sponding menthyl carbonate (see Supporting Information).

(22) We have also carried out the same sequence of reactions (eqs 14-

16) employing (4-methoxybenzyloxy)acetaldehyde with similar yields and

selectivities.

(23) Evans, D. A.; Ratz, A. M.; Huff, B. E.; Sheppard, G. S. J. Am.

Chem. Soc. 1995, 117, 3448-3467.

(26) Evans, D. A.; MacMillan, D. W. C.; Campos, K. R. J. Am. Chem.

Soc. 1997, 119, 10859-10860.

674 J. Am. Chem. Soc., Vol. 121, No. 4, 1999

EVans et al.

tages. These latent nucleophiles are easily prepared and more

stable than the corresponding ester derivatives, which undergo

rapid hydrolysis.²⁷In addition, the thioester moiety in the

resultant adducts can be readily converted to an acid, amide, or

ester using either a silver-based reagent²⁸or a bromination/

displacement procedure.²⁹Alternatively, the Fukuyama reduction

procedure (Pd/C, Et₃SiH) can be implemented to reduce

thioesters to aldehydes.³⁰

nucleophiles resulted in the formation of the aldol adducts with

diminished selectivity (entries 3 and 4). These results clearly

indicate that the [Cu((S,S)-Ph-pybox)](SbF₆)₂complex (4c) is

the preferred catalyst for acetate aldol reactions with (benzyl-

oxy)acetaldehyde.

The less reactive ketone enolsilane nucleophiles may also be

utilized in the catalyzed additions to (benzyloxy)acetaldehyde

(Table 6). The reactions of 2-(trimethylsilyloxy)propene and

1-phenyl-1-(trimethylsilyloxy)ethylene with (benzyloxy)acetal-

dehyde proceeded poorly at 10 mol % catalyst loading (e56%

ee); however, high enantioselectivity and yields were realized

upon employment of stoichiometric quantities of the catalyst

(g94% ee; Table 6, entries 1 and 2).

Investigation into the use of propionate nucleophiles with the

[Cu(tert-Bu-box)](OTf)₂catalyst system 1a revealed that the

addition of tert-butyl thiopropionate trimethylsilylketene acetal

to (benzyloxy)acetaldehyde proceeded with anti diastereoselec-

tivity (eq 23, 81:19 anti:syn, 84% anti ee).^4cAlthough the yield

The more nucleophilic³¹substituted enolsilanes were found

to be superior substrates in these reactions as compared to their

unsubstituted counterparts (Table 6, entries 3-5). For example,

the reaction of the (E) enolsilane of 2-methyl-3-pentanone

proceeded to complete conversion in the presence of 10 mol %

of the copper catalyst 4c to afford the corresponding adduct

with 95:5 syn:anti selectivity and 90% ee. As anticipated (vide

supra), the analogous (Z) enolsilane was both less reactive and

less enantioselective (entry 3 vs 4). The successful implementa-

tion of 1-(trimethylsilyloxy)cyclopentene demonstrates that

cyclic enolsilanes are also excellent nucleophiles in the (ben-

zyloxy)acetaldehyde aldol reaction (entry 5, 97:3 syn:anti, 96%

ee).

Reaction Scope: [Cu((S,S)-tert-Bu-box)](OTf)₂(1a). Ex-

amination of the scope of the nucleophilic component in the

additions to (benzyloxy)acetaldehyde catalyzed by 1a was next

undertaken (Table 7). The thioester silylketene acetal (entry 1,

91% ee) afforded significantly higher selectivity than to the ester

derivative (entry 2, 50% ee). Additionally, the use of enolsilane

for this reaction (50%) is not preparatively useful, it serves to

illustrate that diastereoselection is not simply inherent to the

nature of the process but is a consequence of an array of factors,

among the most important of which is the geometry of the

substrate-catalyst complex (vide infra). Further studies are

required in order to define the origin of this reversal in

selectivity.

Reaction Mechanism. The proposed catalytic cycle for the

Cu(II)-catalyzed aldol reaction is outlined in Scheme 4.

Coordination of (benzyloxy)acetaldehyde to the Cu(II) center

produces the substrate-catalyst complex 23, which undergoes

nucleophilic addition to afford the copper aldolate 24. Silylation

to form 25 and subsequent decomplexation yields the product

25a and concomitantly regenerates the [Cu((S,S)-Ph-pybox)]-

(SbF₆)₂catalyst (4c).

(27) These thioester silylketene acetals can be stored at -20 °C for

extended periods of time. For their preparation, see the following. (a)

Silylketene acetals of ethyl and tert-butyl thioacetate: Kobayashi, S.;

Mukaiyama, T. Chem. Lett. 1989, 297-300. (b) (E)- and (Z)-silylketene

acetals of ethyl and tert-butyl thiopropionate: Gennari, C.; Beretta, M. G.;

Bernardi, A.; Moro, G.; Solastico, C.; Todeschini, R. Tetrahedron 1986,

42, 893-909.

(28) Booth, P. M.; Fox, C. M. J.; Ley, S. V. Tetrahedron Lett. 1983, 46,

5143-5146.

(29) (a) Using NBS: Minato, H.; Kodama, H.; Miura, T.; Kobayashi,

M. Chem. Lett. 1977, 413-416. (b) Using Br₂: Minato, H.; Takeda, K.;

Miura, T.; Kobayashi, M. Chem. Lett. 1977, 1095-1098.

(30) Fukuyama T.; Lin S.-C.; Li L. J. Am. Chem. Soc. 1990, 112, 7050-

7051.

(31) Bufeindt, J.; Patz, M.; Mu¨ller, M; Mayr, H. J. Am. Chem. Soc. 1998,

120, 3629-3634.

Aldol Addition of Enolsilanes to (Benzyloxy)acetaldehyde

J. Am. Chem. Soc., Vol. 121, No. 4, 1999 675

Silyl Crossover Experiments. It is evident that silicon

transfer from the initially formed catalyst-Nu-RCHO complex

24 may proceed via an intramolecular or intermolecular process

(Scheme 4, 24 f 25). It has been reported that intermolecular

silyl transfer, for example either to counterion (SbF6^-or OTf^-)

or to unreacted aldehyde, may trigger an achiral catalyzed

process that may compete with the enantioselective variant.³²

The details associated with silicon transfer were investigated

in the present system by employing a mixture of two different

silylketene acetals of comparable reactivities (Scheme 5).^4c

Treatment of 0.5 equiv of each of the depicted silylketene acetals

with 1.0 equiv of (benzyloxy)acetaldehyde and 10 mol % of

[Cu(Ph-pybox)](SbF₆)₂(4c) afforded significant quantities of

the four possible products, as detected by GC/MS analysis.

Deprotection of the silyl ethers and chiral HPLC analysis of

the derived alcohols indicated that both aldol adducts were

essentially enantiomerically pure (99% ee). Although there is

clearly a large intermolecular silyl-transfer component in the

reaction, the transient silyl species³³apparently does not compete

effectively at -78 °C with the biscationic copper catalyst in

this aldol reaction. Control experiments demonstrated that

neither the silylketene acetals nor the silyl ether products are

subject to silyl exchange initiated by the catalyst, indicating that

silyl crossover occurs during the course of the reaction.

Nonlinear Effects. Nonlinear effects (NLE) can provide

useful insight into both the behavior of enantioselective catalyst

systems and the mechanisms of the processes they mediate.^34,35

Consequently, experiments were performed to determine if NLE

were operative in the Cu(II) pybox-catalyzed aldol reaction

under investigation. Indeed, when the aldol reaction was

conducted with the catalyst [Cu(Ph-pybox)](SbF₆)₂(10 mol %)

prepared according to the general procedure with ligand of

reduced enantiomeric excess (eq 24), a strong positive nonlinear

effect was observed (Figure 1). For example, employment of a

catalyst of 25% ee afforded the aldol adduct in 74% ee.

To rationalize this significant NLE, we propose that catalyst

disproportionation is occurring under the conditions for catalyst

preparation (eq 26, CH₂Cl₂, 20 °C, 4 h). The formation of a

stable [Cu((S,S)-Ph-pybox)((R,R)-Ph-pybox)](SbF₆)₂2:1 ligand:

metal complex (26) is postulated, serving as a catalytically

inactive reservoir for the minor (R,R)-Ph-pybox ligand and

consequently enriching the enantiomeric excess of the remaining

[Cu((S,S)-Ph-pybox)](SbF₆)₂catalyst (4c) (eq 26).³⁶As these

reactions were conducted using a 1:1 ratio of ligand:Cu, stoi-

chiometry necessitates the production of an unligated copper

(32) (a) Hollis, T. K.; Bosnich, B. J. Am. Chem. Soc. 1995, 117, 4570-

4581. (b) Carreira, E. M.; Singer, R. A. Tetrahedron Lett. 1994, 35, 4323-

4326. (c) Denmark, S. E.; Chen, C.-T. Tetrahedron Lett. 1994, 35, 4327-

4330.

(33) Potential silylating sources include 24 and Me₃SiSbF₆: Olah, G.

A.; Heiliger, L.; Li, X.-Ya; Prakash, G. K. S. J. Am. Chem. Soc. 1990,

112, 5991-5995.

(34) Girard, C.; Kagan, H. B. Angew. Chem., Int. Ed. 1998, 37, 2923-

2959 and references therein.

(35) A very informative study on NLE in the DBFOX/Ph‚Ni(ClO₄)₂

catalyst system appeared during the preparation of this manuscript:

Kanemasa, S.; Oderaotoshi, Y.; Sakaguchi, S.; Yamamoto, H.; Tanaka, J.;

Wada, E.; Curran, D. P. J. Am. Chem. Soc. 1998, 120, 3074-3088.

676 J. Am. Chem. Soc., Vol. 121, No. 4, 1999

EVans et al.

species. Plausible fates of the released Cu(II) include the

37

formation of Cu(SbF₆)₂or a non-Lewis acidic species such

as CuO or Cu(OH)₂.

Several control experiments were performed to provide

support for the NLE proposal. Independent generation of the

2:1 ligand:copper complex using our standard catalyst prepara-

tion procedure afforded a material that was nearly completely

insoluble in the reaction solvent.³⁸Subjection of this solution

to the (benzyloxy)acetaldehyde aldol reaction revealed that the

2:1 ligand:copper complex was not a catalytically competent

species (-78 °C, 2 d, 9% yield vs 4c, -78 °C, 15 min, 99%

yield). This result, buttressed by the fact that as little as 0.5

mol % of the catalyst 4c is capable of effecting the reaction

within 24 h at -78 °C (Table 4), also serves to demonstrate

that once the 2:1 ligand:metal complex is formed, it is not in

an appreciable equilibrium with catalyst 4c. Furthermore, the

addition of 1 equiv of (R,R)-Ph-pybox ligand to a stock solution

of [Cu((S,S)-Ph-pybox)](SbF₆)₂(4c) caused the immediate (e30

s) precipitation of a pale blue amorphous material (presumably

the 2:1 complex) and a clear colorless solution,³⁹indicating that

the formation of the insoluble 2:1 (S,S)-(R,R) ligand metal-

complex 6 is a facile process at room temperature.

To delineate the course of ligand exchange and the relative

stabilities of the (S,S)-(R,R) and (S,S)-(S,S) 2:1 ligand:metal

complexes, a catalyst preparation was undertaken using excess

ligand relative to copper (eq 27). The use of 30 mol % of 2:1

Figure 2. Crystal structures of the 2:1 Ph-pybox:copper complexes

[Cu((S,S)-Ph-pybox)((R,R)-Ph-pybox)](SbF₆)₂(26-X-ray) and [Cu((S,S)-

Ph-pybox)₂](SbF₆)₂(27-X-ray). Counterions have been omitted for

clarity.

catalyst solution⁴⁰in the reaction of (benzyloxy)acetaldehyde

with tert-butyl thioacetate silylketene acetal provided the aldol

product in 83% ee within 15 min at -78 °C (eq 28). Although

the enantioselectivity of this reaction did not approach the

selectivity obtained in the reaction catalyzed by enantiomerically

pure 4c (99% ee, 15 min, -78 °C), it did surpass the

enantioselectivity observed in the NLE experiment (e.g., 50%

ee ligand f 84% ee product, Figure 1). Thus, this experiment

illustrates both that enrichment of the ligand is being achieved

through the formation of the [Cu((S,S)-Ph-pybox)((R,R)-Ph-

pybox)](SbF₆)₂complex (26) and that this species is favored

relative to the (S,S)-(S,S) 2:1 ligand:metal complex.

Ultimate corroboration for the formation of a catalytically

inactive 2:1 ligand:metal complex was obtained through the

X-ray crystal structural determination of both the [Cu((S,S)-

Ph-pybox)((R,R)-Ph-pybox)](SbF₆)₂(26) and [Cu((S,S)-Ph-

pybox)₂](SbF₆)₂(27) complexes (Figure 2). By inspection, the

(S,S)-(R,R) complex 26 appears favored relative to (S,S)-(S,S)

complex 27, as the ligand phenyl groups project unobstructed

into each of the four quadrants. In comparison, in 27 the phenyl

groups of the ligands protrude into the same two quadrants.

Semiempirical calculations (PM3) qualitatively validated this

analysis, as the heat of formation of the (S,S)-(R,R) complex

26 was found to be 2.9 kcal/mol lower in energy than that of

the (S,S)-(S,S) complex 27.

(S,S):(R,R) ligand (33% ee) and 20 mol % Cu(SbF₆)₂(i.e., 20

mol % CuCl₂and 40 mol % AgSbF₆) should, ideally, produce

10 mol % of the [Cu((S,S)-Ph-pybox)((R,R)-Ph-pybox)](SbF₆)₂

complex (26) and 10 mol % of enantiomerically enriched [Cu-

((S,S)-Ph-pybox)](SbF₆)₂(4c) (catalyst A). Employment of this

(36) We have previously observed positive NLE in the Sm(III)-catalzyed

Meerwein-Pondorf-Verley reduction and the Cu(II)-catalyzed Diels-

Alder reaction: Evans, D. A.; Nelson, S. G.; Gagne´, M.; Muci, A. R. J.

Am. Chem. Soc. 1993, 115, 9800-9801.

(37) To avoid achiral catalysis, the formation of Cu(SbF₆)₂would

apparently necessitate that this complex be either insoluble in CH₂Cl₂or a

noncompetitive catalyst. While the preparation of Cu(SbF₆)₂as a white solid

has been reported, to our knowledge its solubility in common organic

solvents has not been studied: (a) Cader, M. S. R.; Aubke, F. Can. J. Chem.

1989, 67, 1700-1701. (b) Gantar, D.; Leban, I.; Frlec, B.; Holloway, J. H.

J. Chem. Soc., Dalton Trans. 1987, 2379-2382.

Catalyst Characterization and Stereochemical Models:

[Cu(Ph-pybox)](SbF₆)₂. The proposed requirement for chelation

in the [Cu((S,S)-Ph-pybox)](SbF₆)₂(4c)-mediated (benzyloxy)-

(38) A dark blue precipitate (presumably the 2:1 complex) and a very

pale blue solution were observed.

(39) Solutions of 4c are blue and free of precipitate; thus, it appears that

colorless solutions contain very little or no copper species.

(40) A pale blue precipitate (presumably the 2:1 complex) and a blue

solution were observed (eq 28).

Aldol Addition of Enolsilanes to (Benzyloxy)acetaldehyde

J. Am. Chem. Soc., Vol. 121, No. 4, 1999 677

Figure 3. Crystallographic structures of the [Cu(i-Pr-pybox)(L)_n](SbF₆)₂complexes 39 and 40 along with selected bond lengths and angles.

acetaldehyde aldol reaction requires the intermediacy of a five-

coordinate Cu(II) catalyst-substrate complex.^41,42Several five-

coordinate [Cu(pybox)](SbF₆)₂complexes were synthesized with

the intent of obtaining crystal structures that might elucidate

the basic coordination geometry (i.e., trigonal bipyramidal or

square pyramidal)⁴³of these complexes, thus providing a basis

upon which to construct a stereochemical model of the catalyst-

substrate complex. Due to the highly crystalline nature of the

bis(hydrate), [Cu(i-Pr-pybox)(H₂O)₂](SbF₆)₂(28, Figure 3) was

selected as the initial substrate from which to extrapolate

geometrical information on pentacoordinate Cu(II) complexes.

Efforts were also directed toward obtaining crystals of chelated

pentacoordinated [Cu(pybox)]²⁺complexes to model this reac-

tion, in which the catalyst-substrate complex is similarly

organized. In this context, we also obtained an X-ray structure

of the analogous dimethoxyethane (DME) complex 29 (Figure

3). Both the bis(hydrate) and the DME complexes 28 and 29

adopt a square pyramidal geometry, with the SbF₆counterions

fully dissociated from the metal center (Figure 3). A useful

measure of distortion from the ideal square pyramidal geometry

is the N1_(pyridyl)-Cu-O3_(equat)bond angle, which by definition

is 180° for an undistorted complex. For the DME complex 29,

this measure of distortion (N1-Cu1-O3 ) 156.4°) is somewhat

greater than the corresponding measurement in the bis(hydrate)

complex 28 (N1-Cu1-O3 ) 159.0°).

Å), with the more weakly coordinated oxygen ligand positioned

in the apical site of the complexes (28, Cu-OR₂) 2.179 Å;

29, Cu-OR₂) 2.203 Å). From these data, it is reasonable to

conclude that the Cu-O bond lengths in 29 provide a direct

measure of the inherent Lewis acidity of the two nonequivalent

catalyst binding sites.

Based upon the preceding structural data, the trigonal

bipyramidal complex 30 was considered unlikely; furthermore,

this structure predicts the incorrect stereochemical outcome for

the (benzyloxy)acetaldehyde aldol reaction (Scheme 6). For the

As a consequence of the electronic configuration of the Cu-

(II) center (d⁹) and accompanying Jahn-Teller distortion, the

square pyramidal geometry affords a strong coordinating site

in the ligand plane, with a weaker coordination site in the axial

position.⁴²In accord with this expectation, the more tightly

bound oxygen heteroatoms in 28 and 29 are found in the

equatorial plane (28, Cu-OR₂) 1.985 Å; 29, Cu-OR₂) 2.064

square pyramidal copper geometry, two diastereomeric catalyst-

substrate complexes 31a and 31b must be considered in the

analysis of the impact of catalyst structure on reaction stereo-

chemistry. As documented in the [Cu(pybox)]²⁺system by the

X-ray crystal structures 28 and 29, the square pyramidal

geometry affords a strong coordinating site in the ligand plane,

with a weaker coordination site in the axial position. As a

consequence, for maximal carbonyl activation, aldehyde coor-

dination is postulated to occur in the equatorial position, as

illustrated in complex 31a. Accordingly, the catalyst-substrate

complex 31a, successfully predicts the stereochemical outcome

of the (benzyloxy)acetaldehyde aldol reaction, and the diaster-

eomeric square pyramidal complex 31b predicts the wrong

absolute stereochemistry. The high enantioselectivity observed

(41) The weakly coordinating SbF₆counterions are presumed to not

associate directly with the Cu(II) center. Considerable structural data support

this assertion (vide infra).

(42) (a) For a discussion of five-coordinate Cu(II) complexes, see:

Hathaway, B. J. In ComprehensiVe Coordination Chemistry; Wilkinson,

G., Ed.; Pergamon Press: New York, 1987; Vol. 5, Chapter 53. (b)

Cambridge Structural Database survey: 51 square pyramidal structures; 9

trigonal bipyramidal structures (http://sulfur.scs.uiuc.edu/gifs/cuii.htm).

(43) The energy difference between trigonal bipyramidal and square

pyramidal geometries has been calculated to be low; see: (a) Wilcox, D.

E.; Porras, A. G.; Hwang, Y. T.; Lerch, K.; Winkler, M. E.; Solomon, E.

I. J. Am. Chem. Soc. 1985, 107, 4015-4027. (b) Solomon, E. I. Comments

Inorg. Chem. 1984, 3, 227-320.

678 J. Am. Chem. Soc., Vol. 121, No. 4, 1999

EVans et al.

Figure 4. Two perspectives of the X-ray structure of the [Cu(Ph-pybox)(BnOCH₂CHO)](SbF₆)₂complex 41 along with selected bond lengths and

angles.

in these reactions (g95% ee) provides strong support for the

assertion that only one of the two complexes, i.e., 31a, is

catalytically competent.

benzyloxy)acetaldehyde, the enantioselection was restored to

99% ee (eq 29). These experiments suggest that this π-π

interaction plays an important organizational role in the assembly

of the catalyst-substrate complex.⁴⁵

Ultimately, we were successful in obtaining deep blue crystals

of the catalyst-substrate [Cu((S,S)-Ph-pybox)(BnOCH₂CHO)]-

(SbF₆)₂complex (32). The X-ray structure reveals that the

copper geometry is square pyramidal, with the carbonyl oxygen

coordinated to the equatorial site in the ligand plane and the

ether oxygen occupying an apical site (Figure 4). As expected,

the carbonyl oxygen occupies the more Lewis acidic site as

evidenced by the much shorter aldehyde-Cu bond length

(Cu1-O3 ) 1.986 Å) relative to the benzyl ether-Cu bond

length (Cu1-O4 ) 2.328). This complex exhibits a minimal

amount of distortion from an ideal square pyramidal geometry,

as evidenced by the N1_(pyridyl)-Cu-O3_(equat)bond angle of

169.0°. In direct accord with our prediction, it is evident from

this structure that the re face of the aldehyde carbonyl is

completely shielded by the Ph substituent on the ligand.

Further inspection of the [Cu(Ph-pybox)(BnOCH₂CHO)]-

(SbF₆)₂X-ray structure (32) reveals that the phenyl group of

the (benzyloxy)acetaldehyde substrate is oriented under the

pyridine ring of the pybox ligand (∼3.5 Å) in a parallel fashion

(Figure 4). This geometrical arrangement and distance are

consistent with a parallel offset face-to-face π-π interaction

between the phenyl and pyridyl moieties.⁴⁴Experiments were

designed to counter the possibility that the observed orientation

is solely a consequence of crystal packing and of no relevance

to the actual solution behavior. Upon substitution of the benzyl

group of the substrate with an alkyl group, as in (n-butyloxy)-

acetaldehyde, a significant reduction in enantioselectivity was

observed (eq 29, 99 f 88% ee, ∆∆G^q≈ 1 kcal/mol at -78

°C). When the aryl group was reinstalled, as in (4-methoxy-

When the [Cu((S,S)-Ph-pybox)(BnOCH₂CHO)](SbF₆)₂crys-

tals (32) were redissolved in CH₂Cl₂and treated with the

silylketene acetal derived from tert-butyl thioacetate under the

usual reaction conditions (eq 30), the aldol adduct was obtained

in 99% ee (1 h, -78 °C), the same value as obtained for the

catalyzed reaction. This experiment provides strong evidence

that the catalyst-aldehyde complex (32) isolated and character-

ized is also the catalytically relevant species in solution.

Evidence for Chelation. The selection of (benzyloxy)-

acetaldehyde as the aldol reaction substrate was predicated upon

the proposed ability of this aldehyde to engage in bidentate

(45) For examples where π-π interactions in enantioselective catalyst

systems have been characterized by X-ray crystallography, see: (a) Hawkins,

J. M.; Loren, S.; Nambu, M. J. Am. Chem. Soc. 1994, 116, 1657-1660.

(b) Quan, R. W.; Li, Z.; Jacobsen, E. N. J. Am. Chem. Soc. 1996, 118,

8156-8157.

(44) Hunter, C. A.; Sanders, J. K. M. J. Am. Chem. Soc. 1990, 112,

5525-5534

Aldol Addition of Enolsilanes to (Benzyloxy)acetaldehyde

J. Am. Chem. Soc., Vol. 121, No. 4, 1999 679

coordination to the Cu(II)-center. Support for dynamic substrate

chelation was acquired when R-(tert-butyldimethylsilyloxy)-

acetaldehyde, an aldehyde which is expected to be an ineffective

chelator,⁴⁶was implemented in the Cu(II)-catalyzed aldol

reaction: the use of this aldehyde led to a less enantioselective

process (eq 31, 56% ee) as compared to (benzyloxy)acetalde-

hyde (99% ee). Furthermore, hydrocinnamaldehyde, a mono-

dentate substrate incapable of chelation, afforded a racemic

product when employed in this reaction (eq 31). Based upon

these results, catalyst-substrate chelation appears to be an

absolute requirement for obtaining high enantioselectivity in this

process.

Indeed, this has been shown to be the case, as demonstrated by

the low reactivity of (R)-R-(benzyloxy)propionaldehyde. While

the manifestation of matched and mismatched reaction partners

is consistent with all the models, both the diastereomeric square

pyramidal model 34 and the trigonal bipyramidal model 35

would predict the opposite matched and mismatched relation-

ships relative to those observed (Scheme 8); moreover, the

results of these double stereodifferentiating experiments are in

full accord with the proposal that (benzyloxy)acetaldehyde

coordinates to the Cu(II) center in a bidentate fashion.

Solution-State Characterization. To establish a correlation

between the accumulated solid-state structural data and the

solution behavior, the catalyst-substrate species were probed

using a combination of electrospray ionization mass spectrom-

etry (ESI) and electron paramagnetic resonance (EPR) spec-

troscopy. Significantly, the ESI spectrum of [Cu(Ph-pybox)-

(BnOCH₂CHO)](SbF₆)₂clearly affirmed the presence of a

doubly charged catalyst-substrate complex, [Cu(Ph-pybox)-

(BnOCH₂CHO)]²⁺, in solution without any associated coun-

terions (see Supporting Information). Additionally, the EPR

spectra of the [Cu(i-Pr-pybox)(H₂O)₂](SbF₆)₂(28), [Cu(i-Pr-

pybox)](SbF₆)₂(29), and [Cu(Ph-pybox)(BnOCH₂CHO)](SbF₆)₂

(32) complexes exhibited well-defined square pyramidal copper

centers, in direct accord with the corresponding crystal structures

(see Supporting Information).⁴⁸The ratio of g_|/A_|is indicative

of distortion away from square pyramidalization; a value of 126

× 10⁴for 32 is consistent with negligible amounts of distor-

tion.⁴⁹The above solid- and solution-state data together provide

compelling evidence for the presence of complex 31a (Scheme

6) in the reactions of (benzyloxy)acetaldehyde employing the

[Cu(Ph-pybox)](SbF₆)₂catalyst (4c).

Double stereodifferentiating experiments with (R)- and (S)-

R-(benzyloxy)propionaldehyde have also been carried out to

provide support for the square pyramidal catalyst-substrate

model 31a described in Scheme 6. Gennari and Cozzi have

shown that the SnCl₄-mediated addition of the silylketene acetal

derived from tert-butyl thioacetate to R-(benzyloxy)propional-

dehyde provides the chelation-controlled adduct with high

selectivity (98:2).⁴⁷Reaction of (R)-R-(benzyloxy)propionalde-

hyde catalyzed by [Cu(Ph-pybox)](SbF₆)₂afforded an unselec-

tive, slow reaction (Scheme 7, mismatched). This result is

consistent with catalyst-substrate square pyramidal coordina-

tion, where the substrate (Me) and ligand (Ph) substituents mask

opposite aldehyde carbonyl enantiofaces (Scheme 8, 33a). In

the matched case, (S)-R-(benzyloxy)propionaldehyde underwent

a rapid reaction, providing a 98.5:1.5 mixture of diastereomers,

favoring the chelation-controlled product (Scheme 7). In the

square pyramidal complex (Scheme 8, 33b), the R-methyl

substituent of (S)-R-(benzyloxy)propionaldehyde reinforces the

facial bias imposed by the catalyst.

Diastereoselectivity Models. The majority of the diastereo-

selective (benzyloxy)acetaldehyde aldol reactions encountered

in this study afford the syn adducts.⁵⁰This syn selectivity can

be rationalized by attack of the silylketene acetal on the proposed

square pyramidal Cu(II)-aldehyde complex 31a via an open

transition state, which minimizes the number of repulsive gauche

and dipole-dipole interactions (Scheme 9, the shielding Ph-

(ligand) group has been omitted for clarity).^2cOf the three

A corollary to these experiments is that (R)-R-(benzyloxy)-

propionaldehyde would be anticipated to act as a catalyst

inhibitor on the basis of the observation that this enantiomer

ideally complements the catalyst by orienting a Me group in

the only open quadrant available in complex 33a (Scheme 8).

(48) The square pyramidal nature of the [Cu(Ph-pybox)(BnOCH₂CHO)]-

(SbF6)₂spectrum was verified by comparison with the EPR spectra of

compounds known to possess square pyramidal copper centers: (a)

Reference 42a, p 662. (b) Batra, G.; Mathur, P. Transition Met. Chem. 1995,

20, 26-29.

(46) (a) Keck, G. E.; Castellino, S.; Wiley: M. R. J. Org. Chem. 1986,

51, 5480-5482. (b) Kahn, S. D.; Keck, G. E.; Hehre, W. J. Tetrahedron

Lett. 1987, 28, 279-280. (c) Keck, G. E.; Castellino, S. Tetrahedron Lett.

1987, 28, 281-284. (d) Keck, G. E.; Andrus, M. B.; Castellino, S. J. Am.

Chem. Soc. 1989, 111, 8136-8141. For evidence supporting chelation of

an OTBS group, see: (e) Chen, X.; Hortelano, R. R.; Eliel, E. L.; Frye, S.

V. J. Am. Chem. Soc. 1992, 114, 1778-1784.

(49) In addition, the simulated spectrum of [Cu(Ph-pybox)(BnOCH₂-

CHO)](SbF₆)₂closely matched the experimental spectrum (see Supporting

Information) when the following simulation parameters were employed:

g₁) g₂) 1.9368, g₃) 2.3300; A ) 152.92 G, LB ) 100 G.

(50) Reetz has observed syn diastereoselectivity in the TiCl₄and SnCl₄-

mediated aldol reactions of benzyloxyacetaldehyde with propiophenone

enolsilanes: Reetz, M. T.; Kesseler, K.; Jung, A. Tetrahedron 1984, 21,

4327-4336.

(47) Gennari, C.; Cozzi, P. G. Tetrahedron 1988, 44, 5965-5974.

680 J. Am. Chem. Soc., Vol. 121, No. 4, 1999

EVans et al.

possible transition states that lead to the observed syn product,

antiperiplanar transition state 36 has the fewest destabilizing

interactions. By comparison, the antiperiplanar transition state

37, which would afford the anti product, incurs both Me-

(nucleophile) T CH₂(aldehyde) and Me(nucleophile) T catalyst

gauche interactions.

(SbF₆)₂complex 2a is employed but that the opposite sense of

induction is observed when the analogous box-Cu(OTf)₂

complex 1a is employed. We thus conclude that, with the current

The anti diastereoselectivity observed in the aldol addition

of 2-(trimethylsiloxy)furan to (benzyloxy)acetaldehyde (eq 20)

may be rationalized through a similar analysis. Inspection of

each of the acyclic transition states leads to the conclusion that

38 and 39 are preferred on steric grounds (Scheme 10). Further

examination reveals that the synclinal transition state 39 is

favored relative to the antiperiplanar transition state 38 due to

the electrostatic repulsion between the (benzyloxy)acetaldehyde

carbonyl oxygen and the furan oxygen.²⁴

Catalyst Characterization and Stereochemical Models:

[Cu(tert-Bu-box)](OTf)₂. Prior work from this laboratory⁶has

provided the precedent that the [Cu((S,S)-tert-Bu-box)](OTf)₂

and [Cu(tert-Bu-box)](SbF₆)₂complexes 1a and 2a also have

the potential to chelate with (benzyloxy)acetaldehyde. The

relevant complexes of this substrate with 1 and 2 are illustrated

below (Scheme 11). In the absence of counterion participation,

complex 40 affords an unequivocal prediction that the stereo-

chemical outcome of the reaction should afford the illustrated

(S) aldol adduct. If the counterion is an integral part of the

aldehyde-catalyst complex, as in 41, the stereochemical

outcome of the reaction is more ambiguous, since aldehyde

chelation could occur from either equatorial-equatorial or

equatorial-apical (pictured) complexes.

reaction, the triflate counterion remains associated with the metal

complex during the catalytic event. This result stands in contrast

to the analogous aldol reactions with pyruvate esters, where both

1a and 2a afford the same sense of asymmetric induction.⁵¹It

is also noteworthy that the SbF₆-derived complex 2a is less

enantioselective than its triflate counterpart. One might speculate

that 2a is too Lewis acidic for this substrate to allow a highly

selective process to occur.

Methodology Limitations

As demonstrated previously, a chelating substrate is necessary

to achieve high enantioselectivity in the [Cu((S,S)-Ph-pybox)]-

(SbF₆)₂(4c) catalyzed Mukaiyama aldol reaction (eq 33);

moreover, there are strict requirements on the nature of the

chelating substituent (Table 8). Replacement of the benzyloxy

with a benzylthio group, as in (benzylthio)acetaldehyde, resulted

in a decline in enantioselection from 99% to 36% ee. Further-

more, alteration of the tether length can have a dramatic impact,

as evidenced by the complete loss of enantioselectivity when

an additional methylene unit was inserted (â-(benzyloxy)-

propionaldehyde). Simple stereochemical models suggest that

the five-membered chelate with (benzyloxy)acetaldehyde readily

complements the ligand pocket available in [Cu(Ph-pybox)]-

(SbF₆)₂, whereas the six-membered chelate for â-(benzyloxy)-

propionaldehyde adopts a chair or twist-boat conformation,

which undergoes significant steric interactions with the pybox

ligand framework. The complete lack of selectivity obtained

with substrates which, presumably, would attain chelation

geometries similar to that of (benzyloxy)acetaldehyde, such as

ethyl glyoxylate, is not easily rationalized.

The data provided below (eq 32) reveal that complex 40 does

predict the sense of asymmetric induction when the box-Cu-

(51) (a) Evans, D. A.; Kozlowski, M. C.; Burgey, C. S.; MacMillan, D.

W. C. J. Am. Chem. Soc. 1997, 119, 7893-7894. (b) See ref 25.

Aldol Addition of Enolsilanes to (Benzyloxy)acetaldehyde

J. Am. Chem. Soc., Vol. 121, No. 4, 1999 681

General Procedure for the Preparation of (Z)-Ketene Thioacetals.

The Collum procedure⁵⁵was employed for the synthesis of this family

of ketene acetals. As a general precaution, freshly dried/distilled reagents

were used in order to attain the highest levels of stereoselectivity (>95:

5). A chilled (0 °C) slurry of TMP‚HBr (1.3 equiv, 0.09 M in THF)

was treated with n-BuLi (2.4 equiv, 1.6 M in hexanes), stirred for 5

min, and cooled to -78 °C. A solution of the thioester (1.0 equiv, 0.5

M in THF) was cannulated into the light yellow solution and stirred

an additional 30 min. Chlorotrimethylsilane (2 equiv) and triethylamine

(0.5 equiv) were added, and the solution was warmed to 0 °C over a

4-h period before being diluted with pentane and washed with phosphate

buffer (pH ) 7) and 0.5 M aqueous CuSO₄. The organic layer was

dried (Na₂SO₄), concentrated, and distilled under vacuum to furnish

the title compounds.

General Procedure for the Preparation of (E)-Ketene Thioacetals.

The Ireland procedure⁵⁶was employed for the synthesis of this family

of ketene acetals. As a general precaution, freshly dried/distilled reagents

were used in order to attain the highest levels of stereoselectivity (>95:

5). A solution of LDA (1 equiv) and HMPA (23% v/v) was stirred for

10 min prior to cooling (-78 °C) and treatment with the thioester (1.1

equiv). The solution was stirred for 15 min, treated with chlorotrim-

ethylsilane (1 equiv), and warmed slowly to 0 °C. The reaction was

diluted with pentane, washed with phosphate buffer (pH ) 7) and 0.5

M CuSO₄, and dried (Na₂SO₄) prior to concentration under reduced

pressure. The unpurified product was distilled under vacuum to furnish

the title compounds.

Conclusion

In conclusion, efficient catalytic enantioselective Mukaiyama

aldol additions to (benzyloxy)acetaldehyde utilizing the C₂-

symmetric bis(oxazolinyl)pyridine-Cu(II) complex [Cu((S,S)-

Ph-pybox)](SbF₆)₂(4c) have been documented. A wide range

of silylketene acetal and enolsilane nucleophiles can be em-

ployed, utilizing 0.5-10 mol % catalyst loadings, to provide

the aldol products in good yield and with high selectivity (eq

34).

Preparation of [Cu((S,S)-Phenyl-bis(oxazolinyl)pyridine)](SbF₆)₂

(4c). To an oven-dried round-bottom flask containing a magnetic stirring

bar were added, in a nitrogen atmosphere box, (S,S)-bis(phenyloxazoli-

nyl)pyridine (18.5 mg, 0.05 mmol) and CuCl₂(6.7 mg, 0.05 mmol).

To an oven-dried round-bottom flask containing a magnetic stirring

bar was added, in a nitrogen atmosphere box, AgSbF₆(34.4 mg, 0.10

mmol). The flasks were fitted with serum caps and removed from the

nitrogen atmosphere box, and the flask containing the ligand/CuCl₂

mixture was charged with CH₂Cl₂(1.0 mL). The resulting suspension

was stirred rapidly for 1 h to give a fluorescent green suspension.

AgSbF₆(in 0.5 mL CH₂Cl₂) was added via cannula with vigorous

stirring, followed by a 0.5-mL CH₂Cl₂rinse. The resulting mixture was

stirred rapidly for 3 h in the absence of light and filtered through an

oven-dried glass pipet tightly packed with cotton (or alternatively an

oven-dried 0.45-µm PTFE filter) to remove the white AgCl precipitate,

yielding active catalyst [Cu(Ph-pybox)](SbF₆)₂as a clear blue solution.

General Procedure for the Catalyzed Addition of Silylketene

Acetals to Benzyloxyacetaldehyde Using [Cu(Ph-pybox)](SbF₆)₂(4c).

To a -78 °C solution of [Cu(Ph-pybox)](SbF₆)₂in CH₂Cl₂which was

prepared as described above was added benzyloxyacetaldehyde (70.0

µL, 0.50 mmol), followed by a silylketene acetal (0.60 mmol). The

resulting solution was stirred at the indicated temperature (-78 or -50

°C, see text) until the aldehyde was completely consumed (15 min-

48 h), as determined by TLC (30% EtOAc/hexanes). The reaction

mixture was then filtered through a 1.5- × 8-cm plug of silica gel with

Et₂O (50 mL). Concentration of the ether solution gave a clear oil,

which was dissolved in THF (10 mL) and 1 N HCl (2 mL). After

standing at room temperature for 15 min, this solution was poured into

a separatory funnel and diluted with Et₂O (10 mL) and H₂O (10 mL).

After mixing, the aqueous layer was discarded, and the ether layer was

washed with saturated aqueous NaHCO₃(10 mL) and brine (10 mL).

The resulting ether layer was dried over anhydrous MgSO₄, filtered,

and concentrated to provide the hydroxy esters.

Investigation into the reaction mechanism utilizing doubly

labeled silylketene acetals indicated that there is a significant

intermolecular component to silyl transfer; however, any

transient silyl species does not effectively compete with the

chiral copper catalyst 4c at -78 °C. Further mechanistic studies

revealed a significant positive nonlinear effect, proposed to arise

from the selective formation of the stable [Cu((S,S)-Ph-pybox)-

((R,R)-Ph-pybox)](SbF₆)₂2:1 ligand:metal complex (26). A

stereochemical model, 31a, is proposed in which chelation of

(benzyloxy)acetaldehyde to the metal center to form a square

pyramidal copper intermediate accounts for the observed sense

of induction. Support for this proposal has been gained from

double stereodifferentiating reactions, EPR spectroscopy, ESI

spectrometry, and, ultimately, the X-ray crystal structure 32 of

the aldehyde bound to catalyst.⁵²

Experimental Section⁵³

General Procedure for the Preparation of Ketene Acetals. The

thioester (1 equiv) was added to a cold (-78 °C) 0.4 M solution of

lithium diisopropylamide (1.2 equiv) in THF and stirred for 1 h before

the addition of chlorotrimethylsilane (1.1 equiv). The reaction mixture

was warmed to ambient temperature over a 4-h period, diluted with

pentane, washed with phosphate buffer (pH ) 7) and 0.5 M aqueous

CuSO₄, and dried (Na₂SO₄). Removal of the solvent and distillation of

the crude liquid under reduced pressure afforded the desired ketene

thioacetal.⁵⁴

Preparation of (S)-tert-Butyl 4-Benzyloxy-3-hydroxybutanethio-

ate (7, Table 4, Entry 1). Compound 7 was prepared according to the

general procedure using [Cu(Ph-pybox)](SbF₆)₂(200 µL, 2.5 µmol, 0.5

mol %) and the silylketene acetal of tert-butyl thioacetate (122 mg,

0.60 mmol, 153 µL) to provide the pure (S)-hydroxy ester in 100%

(54) Ketene silylthioacetal of tert-butyl thioester: Gerlach, H.; Kunzler,

P. HelV. Chim. Acta 1978, 61, 2503-2509.

(55) Hall, P. L.; Gilchrist, J. H.; Collum, D. B. J. Am. Chem. Soc. 1991,

113, 9571-9574.

(52) Since completion of this study, other chiral chelating Lewis acid

complexes that could well be good catalysts for this family of reactions

have been reported: see ref 35.

(56) Ireland, R. E.; Mueller, R. H.; Willard, A. K. J. Am. Chem. Soc.

1976, 98, 2868-2877. We have also occasionally used the Fukuzumi-

Otera protocol: Otera, J.; Fujita, Y.; Fukuzumi, S. Synlett 1994, 213-214.

(53) General information is provided in the Supporting Information.

682 J. Am. Chem. Soc., Vol. 121, No. 4, 1999

EVans et al.

yield (141 mg, 0.50 mmol). Enantiomeric excess was determined by

HPLC with a Chiralcel OD-H column (94.2:0.8:5.0 hexanes/2-propanol/

EtOAc; 1.0 mL/min; (R) enantiomer t_r) 16.3 min; (S) enantiomer t_r

) 17.9 min) or with a Chiralcel AD column (96:4 hexanes/2-propanol;

1.0 mL/min; (S) enantiomer t_r) 13.1 min; (R) enantiomer t_r) 16.5

min; 99% ee). The analytical data obtained from this material (¹H NMR,

¹³C NMR, IR, and HRMS) were identical to those previously reported:

R_fspot (R_f0.39, 20% EtOAc/hexanes). The reaction mixture was then

filtered through a 2.5- × 8-cm plug of silica gel with Et₂O (200 mL).

Concentration of the ether solution gave a clear oil, which was dissolved

in THF (100 mL) and 1 N HCl (10 mL). After standing at room

temperature for 15 min, this solution was poured into a separatory funnel

and diluted with Et₂O (100 mL). The aqueous layer was discarded and

the organic layer washed with saturated NaHCO₃(50 mL) and brine

(50 mL) and dried over anhydrous MgSO₄. Filtration and concentration

of the resulting solution gave the hydroxy ketoester in 96% yield (1.7

4c

[R]^rt-10.9 (c 3.0, CH₂Cl₂); [R]²⁶(lit.^4c) +10.0 (c 1.0, CHCl₃)

D

96% ee (R).

g, 6.4 mmol): R_f0.50 (50% EtOAc/hexanes); [R]^rt-13.7 (c 3.55,

Preparation of (S)-Ethyl 4-Benzyloxy-3-hydroxybutanethioate

(Table 4, Entry 2). This compound was prepared according to the

general procedure employing [Cu(Ph-pybox)](SbF₆)₂(200 µL, 2.5 µmol,

0.5 mol %) and the silylketene acetal of ethyl thioacetate (106 mg,

0.60 mmol, 132 µL) to provide the pure adol adduct in 95% yield (121

mg, 0.048 mmol) after flash chromatography with 20% EtOAc/hexanes.

Enantiomeric excess was determined by HPLC with a Chiralcel OD-H

column (98:2 hexanes/2-propanol; 1.0 mL/min): (S) enantiomer t_r)

31.6 min; (R) enantiomer t_r) 35.7 min; 98% ee. The analytical data

obtained from this material (¹H NMR, ¹³C NMR, and HRMS) were

D

CHCl₃); IR (neat) 3438, 2864, 1740, 1716 cm^-1; ¹H NMR (400 MHz,

CDCl₃) δ 7.29 (m, 5H), 4.54 (s, 2H), 4.30 (m, 1H), 3.65 (s, 3H), 3.50

(dd, J ) 4.5, 9.6 Hz, 1H), 3.41 (dd, J ) 5.7, 9.6 Hz, 1H), 2.78 (d, J

) 5.9 Hz, 2H); ¹³C NMR (100 MHz, CDCl₃) δ 202.3, 168.6, 137.8,

128.5, 127.9, 127.8, 73.5, 73.1, 66.8, 52.4, 49.7, 46.3; HRMS (CI, NH₃)

exact mass calcd for (C₁₄H₁₈O₅+ NH₄)⁺requires m/z 284.1487, found

m/z 284.1498.

The hydroxy ketoester was subsequently reduced to the anti diol

using tetramethylammonium acetoxyborohydride.⁵⁷A solution of

tetramethylammonium acetoxyborohydride (11.8 g) in acetic acid (60

mL) was added to a -35 °C solution of the ketoester in CH₃CN (100

mL) over 30 min. The resultant milky white solution was stirred at

-35 °C for 18 h and then quenched by the addition of a saturated

solution of Rochelle salts (100 mL) and warming to room temperature.

The resulting mixture was diluted with EtOAc and made basic with a

saturated solution of Na₂CO₃. The aqueous layer was discarded, and

the organic layer was washed with brine (50 mL), dried over MgSO₄,

and concentrated to give the anti diol in 91% yield (1.6 g, 6.1 mmol)

as a white solid. Product ratios were determined by HPLC with a

Chiralcel OD-H column (90:10 hexanes/2-propanol; 1.0 mL/min): syn-

(3R,5S) t_r) 15.9 min; anti-(3R,5R) t_r) 17.9 min; syn-(3S,5R) t_r)

22.8 min; anti-(3S,5S) t_r) 29.5 min; 15:1 anti:syn, 97% anti ee: mp

61 °C; R_f0.40 (50% EtOAc/hexanes); [R]^rt_D+2.2; [R]^rt₅₄₆+6.3 (c 1.8,

CHCl₃); IR (CH₂Cl₂) 3417 (br) 3030, 2949, 2863, 1734 cm^-1; ¹H NMR

(400 MHz, CDCl₃) δ 7.32 (m, 5H), 4.56 (s, 2H), 4.34 (m, 1H), 4.14

(m, 1H), 3.57 (s, 3H), 3.46 (dd, J ) 10.0, 4.0 Hz, 1H), 3.41 (dd, J )

9.7, 7.5 Hz, 1H), 3.01 (s, 1H); 2.51 (d, J ) 5.8 Hz, 1H); 1.63 (m, 2H);

¹³C NMR (100 MHz, CDCl₃) δ 173.0, 138.6, 134.7, 128.5, 127.9, 74.3,

73.4, 68.2, 66.3, 52.1, 41.3, 38.8; HRMS (CI, NH₃) exact mass calcd

for (C₁₄H₂₀O₅+ NH₄)⁺requires m/z 286.1658, found m/z 286.1654.

Large-Scale Preparation of (R)-tert-Butyl 6-Benzyloxy-5-hy-

droxy-3-oxohexanoate and Recovery of Ligand (15, Eq 14). In a

nitrogen atmosphere glovebox, an oven-dried 250-mL round-bottom

flask equipped with a magnetic stirring bar and an oven-dried 25-mL

round-bottom flask were charged with anhydrous CuCl₂(239 mg, 1.8

mmol) and AgSbF₆(1.22 g, 3.6 mmol), respectively. The flasks were

sealed with rubber septa, removed from the glovebox, and charged with

CH₂Cl₂(58 mL and 8 mL, respectively) under a positive pressure of

nitrogen. To the resulting suspension of CuCl₂was added solid (R,R)-

bis(phenyloxazolinyl)pyridine (657 mg, 1.8 mmol) in one portion. The

suspension was rapidly stirred for 2 h, during which time the insoluble

material changed color from brown to dark green to light green. To

this suspension was rapidly added the AgSbF₆solution via syringe.

The suspension was stirred vigorously in the absence of light for 1 h.

The resultant suspension (a blue solution containing a fine white

precipitate) was sequentially filtered open to the atmosphere through a

plug of packed cotton (20 mm × 20 mm) and two oven-dried 0.45-µm

filters (Gelman Acrodisc CR PTFE, 25 mm) directly into a dry 250-

mL round-bottom flask to give a deep-blue catalyst solution which was

used within 2 h.

identical to those previously reported:^4c[R]^rt-10.6 (c 4.2, CH₂Cl₂);

D

[R]²⁶(lit.^4c) +11.4 (c 1.0, CHCl₃), 94% ee (R).

D

Preparation of (S)-Ethyl 4-Benzyloxy-3-hydroxybutanoate (Table

4, Entry 3). The silyl ether was prepared according to the general

procedure employing [Cu(Ph-pybox)](SbF₆)₂(200 µL, 2.5 µmol, 0.5

mol %) and the silylketene acetal of EtOAc (96 mg, 0.60 mmol, 114

µL). Deprotection of the TMS ether using 1 N HCl caused decomposi-

tion to the retroaldol product; thus, a fluoride deprotection procedure

was used instead. The crude silyl ether was dissolved in THF (5 mL)

and cooled to 0 °C. TBAF (1.0 M in THF, 0.60 mmol, 0.60 mL) was

added dropwise. After 15 min, the solution was diluted with Et₂O (10

mL) and saturated NaHCO₃(10 mL) and poured into a separatory

funnel. After mixing, the aqueous layer was discarded and the organic

layer washed with brine (10 mL) and dried over MgSO₄. Filtration

and concentration gave the pure hydroxy ethyl ester in 99% yield (117

mg, 0.49 mmol) after flash chromatography with 20% EtOAc/hexanes.

Enantiomeric excess was determined by HPLC with a Chiralcel OD-H

column (94.2:0.8:5.0 hexanes/2-propanol/EtOAc; 1.0 mL/min): (S)

enantiomer t_r) 24.8 min; (R) enantiomer t_r) 29.3 min; 98% ee; R_f

0.27 (30% EtOAc/hexanes); [R]^rt-8.8 (c 2.1, CH₂Cl₂); IR (CH₂Cl₂)

D

1

3579, 3061, 2905, 1728, cm^-1; H NMR (400 MHz, CDCl₃) δ 7.33

(m, 5H), 4.56 (s, 2H), 4.24 (dq, J ) 4.6, 6.1 Hz, 1H), 4.15 (q, J ) 7.1

Hz, 2H), 3.51 (dd, J ) 4.5, 9.6 Hz, 1H), 3.47 (dd, J ) 6.0, 9.6 Hz,

1H), 2.96 (br s, 1H), 2.54 (d, J ) 6.3 Hz, 2H), 1.25 (t, J ) 7.3 Hz,

3H); ¹³C NMR (100 MHz, CDCl₃) δ 172.0, 137.8, 128.3, 127.7, 127.6,

73.2, 73.0, 67.1, 60.6, 38.2, 14.0; HRMS (EI) exact mass calcd for

+

C₁₃H₁₈O₄requires m/z 238.1205, found m/z 238.1206.

Preparation of (S)-(3-Benzyloxy-2-hydroxypropyl)-2,2-dimethyl-

1,3-dioxin-4-one (11, Eq 12). Compound 11 was prepared according

to the general procedure using [Cu(Ph-pybox)](SbF₆)₂(2 mL, 0.025

mmol, 5 mol %) and the trimethylsilylketene acetal derived from 2,2,6-

trimethyl-1,3-dioxen-4-one¹⁵(126 mg, 0.60 mmol) to provide the pure

adol adduct in 95% yield (165 mg, 0.564 mmol) after flash chroma-

tography with 50% EtOAc/hexanes. Enantiomeric excess was deter-

mined by HPLC with a Chiralcel OD-H column (90:10 hexanes/2-

propanol; 1.0 mL/min): (R) enantiomer t_r) 17.7 min; (S) enantiomer

t_r) 22.6 min; 92% ee; R_f0.20 (50% EtOAc/hexanes); [R]^rt-15.1,

D

[R]^rt₅₄₆-14.9 (c 1.3, CHCl₃); IR (neat) 3438, 3089, 2914, 2863, 1718,

1634 cm^-1; ¹H NMR (400 MHz, CDCl₃) δ 7.33 (m, 5H), 5.31 (s, 1H),

4.56 (s, 2H), 4.10 (m, 1H), 3.51 (dd, J ) 3.6, 9.4 Hz, 1H), 3.40 (dd,

J ) 6.5, 9.4 Hz, 1H), 2.40 (dd, J ) 1.5, 5.4 Hz, 1H), 2.39 (dd, J )

3.1, 5.4 Hz, 1H), 1.66 (s, 3H), 1.65 (s, 3H); ¹³C NMR (100 MHz,

CDCl₃) δ 168.6, 161.1, 137.6, 128.6, 128.5, 128.1, 127.9, 106.7, 95.3,

67.5, 60.4, 37.9, 25.4, 24.7; HRMS (CI, NH₃) exact mass calcd for

(C₁₆H₂₀O₅+ NH₄)⁺requires m/z 310.1664, found m/z 310.1654.

Preparation of (3S,5S)-Methyl 6-Benzyloxy-3,5-dihydroxyhex-

anoate (13, Eq 13). A clear blue solution of [Cu(Ph-pybox)](SbF₆)₂

(2.0 mL, 0.05 mmol, 0.75 mol %) was added over 15 min to a -78 °C

solution of benzyloxyacetaldehyde (1.0 g, 6.7 mmol) and the bis

silylketene acetal of methyl acetoacetate¹⁷(2.1 g, 8.0 mmol) in CH₂-

Cl₂(2 mL). After 2 h at -78 °C, TLC indicated complete consumption

of starting aldehyde (R_f0.13, 20% EtOAc/hexanes) and a new higher

The above catalyst solution (5 mol %) was cooled to -78 °C under

an atmosphere of dry nitrogen. To this solution was added freshly

distilled 1,3-bis(trimethylsiloxy)-1-tert-butoxybuta-1,3-diene^19,58(12.6

g, 39.1 mmol) to give a purple solution. This solution was further cooled

to -93 °C (internal temperature) via a MeOH/liquid nitrogen cold bath,

(57) Evans, D. A.; Gauchet-Prunet, J. A.; Carreira, E. M.; Charette, A.

B. J. Org. Chem. 1991, 56, 741-750.

(58) Careful distillation is required in order to minimize thermal

isomerization to the undesired isomer (bp 50 °C at 0.1 mmHg, bath temp

e 65 °C). See: Anderson, G.; Cameron, D. W.; Feutrill, G. I.; Read, R.

W. Tetrahedron Lett. 1981, 22, 4347-4348.

Aldol Addition of Enolsilanes to (Benzyloxy)acetaldehyde

J. Am. Chem. Soc., Vol. 121, No. 4, 1999 683

at which point the solution turned brown. Freshly distilled benzyloxy-

acetaldehyde (5 mL, 35.5 mmol) was added dropwise via syringe pump

over 15 min (0.33 mL/min). After addition, the internal reaction

temperature had risen to -85 °C. After 5 min at this temperature and

15 min at -78 °C, TLC analysis (30% EtOAc/hexanes) indicated

consumption of the starting aldehyde (R_f0.25). The cold reaction

mixture was poured directly onto a deactivated (5% Et₃N/hexanes) silica

gel plug (5.5 cm × 12 cm) and eluted rapidly with Et₂O (1.5 L). The

filtrate was concentrated in vacuo to yield a yellow oil.

The yellow oil was dissolved in 100 mL of anhydrous MeOH and

treated with pyridinium p-toluenesulfonate (500 mg). When hydrolysis

was complete (1-2 h) by TLC (SM R_f0.51; 30% EtOAc/hexanes),

the volatiles were removed in vacuo, and the yellow oil obtained was

purified by flash chromatography with 20-70% EtOAc/hexanes to

provide a keto-enol tautomeric mixture of tert-butyl (R)-6-benzyloxy-

5-hydroxy-3-oxohexanoate (keto R_f0.19, enol R_f0.11; 30% EtOAc/

hexanes) as a yellow oil in 85% yield (9.37 g, 30.2 mmol).²¹

A small sample (ca. 1-2 mg) of the purified product was converted

to the Mosher ester by the method of Ward and Rhee ((S)-MTPA-Cl,

DMAP, CH₂Cl₂).⁵⁹This material was directly analyzed by HPLC with

a Zorbax SIL column (5% EtOAc/hexanes;, 1.0 mL/min): (S,R)

diastereomer t_r) 22.7 min; (R,R) diastereomer t_r) 26.2 min; g99%

de.

NaHCO₃(200 mL), then water (200 mL), saturated Na₂SO₃(200 mL),

and saturated NaCl (200 mL) before being dried over Na₂SO₄and

concentrated in vacuo to a viscous pale yellow oil. This material is

sufficiently pure for further transformations. Flash chromatography

(40% f 50% EtOAc/hexanes) provided the desired product (>200:1

syn:anti) as a colorless oil (7.82 g, 85%). HPLC Chiralcel AD analysis

(1.0 mL/min; 93:7 hexanes/2-propanol): (3R,5S) t_r) 14.1 min

(desired); (3S,5R) t_R) 18.8 min; (3R,5R) and (3S,3S) t_R) 15.2 and

19.4 min (exact assignment undetermined for anti diols). The analytical

data obtained from this material are consistent with those previously

reported.²¹

Preparation of (2S,3S)-Ethyl 4-Benzyloxy-3-hydroxy-2-methyl-

butanethioate (Table 5, Entry 1). This compound was prepared

according to the general procedure employing [Cu(Ph-pybox)](SbF₆)₂

(3.0 mL, 0.05 mmol, 10 mol %) and the silylketene acetal of ethyl

thiopropionate as a 95:5 mixture of Z:E isomers (114 mg, 0.60 mmol,

130 µL). The product was obtained as a clear oil in 90% yield (121

mg, 0.45 mmol) after flash chromatography with 20% EtOAc/hexanes.

Product ratios were determined by HPLC with a Chiralcel OD-H

column (95.5:1.5:3 hexanes/2-propanol/EtOAc; 1.0 mL/min): syn-

(2S,3S) t_r) 12.6 min; syn-(2R,3R) t_r) 14.3 min; anti-(2R,3S) t_r)

15.2 min; anti-(2S,3R) t_r) 17.5 min; 97:3 syn:anti; 97% syn ee. The

analytical data obtained from this material (¹H NMR, ¹³C NMR, and

HRMS) were identical to those previously reported:^4c[R]^rt+41.1 (c

D

The original silica plug used to remove the copper catalyst was

flushed with 750 mL of 20% concentrated NH₄OH/MeOH and the

filtrate concentrated in vacuo. The residue was partitioned between CH₂-

Cl₂(100 mL) and concentrated NH₄OH (100 mL), and the layers were

separated. The organic layer was washed with concentrated NH₄OH

(2 × 100 mL each), water (100 mL), and brine (100 mL). The organic

extracts were dried over anhydrous Na₂SO₄, filtered, and concentrated

in vacuo to give crude recovered ligand in 62% yield (410 mg, 1.1

mmol) as a waxy yellow solid.

3.6, CH₂Cl₂); [R]²⁶(lit.^4c) -3.3 (c 1.0, CHCl₃); syn:anti 72:28, 90%

D

syn ee (2R,3R).

Preparation of (2S,3S)-Ethyl 4-Benzyloxy-3-tert-butyldimethyl-

siloxy-2-methylbutanethioate (Table 5, Entry 3). The silyl ether-

protected adduct could also be obtained directly, according to the

general procedure employing [Cu(Ph-pybox)](SbF₆)₂(3.0 mL, 0.05

mmol, 10 mol %) and the tert-butyldimethylsilylketene acetal of ethyl

thiopropionate as a 94:6 mixture of Z:E isomers (139 mg, 0.60 mmol,

161 µL). The silyl ether product (after filtration away from the copper

catalyst through SiO₂) was not subjected to the deprotection procedure

in the general procedure but was purified by flash chromatography with

0-30% EtOAc/hexanes. A small amount of the alcohol adduct (<10%,

not isolated in pure form) was obtained, but the major product was the

tert-butyldimethylsilyl ether, which was obtained as a clear oil in 68%

yield (130 mg, 0.34 mmol): [R]^rt_D+12.46 (c 0.92, CH₂Cl₂); IR (CH₂-

tert-Butyl (3R,5R)-6-Benzyloxy-3,5-dihydroxyhexanoate (16, Eq

15). To a cooled (-33 °C) solution of 6.64 g (25.2 mmol, 3.4 equiv)

of Me₄NHB(OAc)₃in 60 mL of 1:1 MeCN/AcOH was added a solution

of tert-butyl (R)-6-benzyloxy-5-hydroxy-3-oxohexanoate in 10 mL of

MeCN via cannula (plus a 5-mL rinse). The reaction was stirred at

-33 °C for 42 h, warmed to 0 °C, stirred for 30 min, and then quenched

with 50 mL of saturated Na/K tartrate. The mixture was stirred at room

temperature for 1 h and then poured into 200 mL of 3:1 EtOAc/H₂O.

The mixture was made basic (to pH 8) with solid Na₂CO₃, and the

phases were separated. The aqueous phase was extracted with EtOAc

(3 × 40 mL), and the organic extracts were combined and washed

with brine (1 × 40 mL) before being dried over MgSO₄and

concentrated in vacuo. The residue was purified via flash chromatoraphy

(40% EtOAc/hexanes) to afford 1.95 g (84% yield) of the desired diol

as a clear, colorless oil which solidified upon storage at -20 °C: [R]_D

1

Cl₂) 2931, 2858, 1680 cm^-1; H NMR (400 MHz, CDCl₃) δ 7.33 (m,

5H), 4.54 (d, J ) 12.0 Hz, 1H), 4.47 (d, J ) 12.0 Hz, 1H), 4.21 (app

q, J ) 5.5 Hz, 1H), 3.41 (d, J ) 5.7 Hz, 2H), 2.88 (dq, J ) 5.2, 6.9

Hz, 1H), 2.84 (q, J ) 7.4 Hz, 2H), 1.22 (t, J ) 7.4 Hz, 3H), 1.17 (d,

J ) 6.9 Hz, 3H), 0.86 (s, 9H), 0.03 (s, 3H), 0.02 (s, 3H); ¹³C NMR

(100 MHz, CDCl₃) δ 202.0, 138.1, 128.3, 127.6, 127.5, 73.2, 72.5,

72.3, 51.6, 25.8, 23.1, 18.1, 14.6, 11.8, -4.4, -5.0; LRMS (FAB/NBA

+ NaI) m/z 405 MNa⁺; HRMS (FAB/NBA + NaI) exact mass calcd

for (C₂₀H₃₄O₃SiS+Na)⁺requires m/z 405.1896, found m/z 405.1891.

Product ratios were determined after conversion to the alcohol with

aqueous HF in MeCN. The combined silyl ether and alcohol products

in MeCN (1.5 mL) were treated with 40% HF/H₂O (0.5 mL). After 30

min, TLC (30% EtOAc/hexanes) indicated complete consumption of

the silyl ether, and the reaction mixture was quenched with saturated

NaHCO₃(5 mL). The resultant mixture was extracted with CH₂Cl₂(3

× 10 mL). The combined organic layers were dried over anhydrous

Na₂SO₄, filtered, and concentrated to provide the hydroxy ester.

Purification by flash chromatography with 10-20% EtOAc/hexanes

provided the title compound as a clear, colorless oil in 70% yield (93

mg, 0.35 mmol). The analytical data obtained from this material (¹H

NMR, ¹³C NMR, and HRMS) were identical to those described above

for the alcohol obtained from the corresponding trimethylsilyl ether.

Product ratios were determined for the alcohol by HPLC with a

Chiralcel OD-H column (95.5:1.5:3 hexanes/2-propanol/EtOAc; 1.0 mL/

min): syn-(2S,3S) t_r) 12.6 min; syn-(2R,3R) t_r) 14.3 min; anti-(2R,3S)

t_r) 15.2 min; anti-(2S,3R) t_r) 17.5 min; 93:7 syn:anti; 96% syn ee.

(2S,3S)-Ethyl 4-Benzyloxy-3-hydroxy-2-isobutylbutanethioate

(Table 5, Entry 4). This compound was prepared according to the

general procedure employing [Cu(Ph-pybox)](SbF₆)₂(3.0 mL, 0.05

mmol, 10 mol %) and the silylketene acetal derived from ethyl

4-methylpentanethioate as a 90:10 mixture of Z:E isomers (174 µL,

0.60 mmol) to provide the pure product as a clear oil in 85% yield

-8.3 (c 1.0, CH₂Cl₂); IR (thin film) 3437 (br), 3030-2863, 1726 cm^-1

;

¹H NMR (500 MHz, CDCl₃) δ 7.36-7.26 (m, 5H), 4.55 (s, 2H), 4.27

(tt, 1H, J ) 6.8, 5.3 Hz), 4.12 (tt, 1H, J ) 7.4, 4.2 Hz,), 3.50 (dd, 1H,

J ) 9.5, 3.9 Hz), 3.40 (dd, 1H, J ) 9.5, 7.3 Hz), 3.17 (br s, 2H),

2.42-2.40 (m, 2H), 1.62-1.56 (m, 2H), 1.45 (s, 9H); ¹³C NMR (100

MHz, CDCl₃) δ 172.2, 138.0, 128.5, 127.8, 127.7, 81.3, 74.4, 73.3,

67.6, 65.4, 42.5, 38.8, 28.1; exact mass calcd for C₁₇H₂₆O₅+ Na

requires m/z 333.1678, found m/z 333.1687 (FAB, m-nitrobenzyl

alcohol, added NaI).

tert-Butyl (3S,5R)-6-Benzyloxy-3,5-dihydroxyhexanoate (17, Eq

16). To a solution of tert-butyl (R)-6-benzyloxy-5-hydroxy-3-oxohex-

anoate (9.20 g, 29.8 mmol) in 300 mL of THF and 60 mL of methanol

was added diethylmethoxyborane (4.7 mL, 36 mmol), and the solution

was stirred for 15 min at room temperature before being cooled to

-78 °C. Sodium borohydride (1.6 g, 42 mmol) is added portionwise

over a 10-min period (moderate gas evolution observed). After the

solution was stirred for 10 h at -78 °C, 120 mL of 30% aqueous

hydrogen peroxide was slowly added to the cold solution via addition

funnel. The mixture was allowed to stir while warming to room

temperature over 10 h and then partioned between EtOAc and water.

The aqueous layer was extracted 3× with EtOAc (200 mL) and then

the combined organics were washed successively 2× with saturated

(59) Ward, D. E.; Rhee, C. K. Tetrahedron Lett. 1991, 32, 7165-7166.

684 J. Am. Chem. Soc., Vol. 121, No. 4, 1999

EVans et al.

(132 mg, 0.43 mmol) after flash chromatography with 10-20% EtOAc/

hexanes. Product ratios were determined by HPLC with a Chiralcel

AD column (99:1 hexanes/2-propanol; 1.0 mL/min): syn-(2R,3R) t_r)

19.7 min; syn-(2S,3S) t_r) 21.1 min; anti-enantiomers t_r) 23.5, 24.4

1H), 4.13 (d, J ) 3.5 Hz, 1H), 4.10 (dt, J ) 3.7, 5.6 Hz, 1H), 3.53 (dd,

J ) 5.4, 9.6 Hz, 1H), 3.48 (dd, J ) 6.0, 9.6 Hz, 1H), 2.90 (m, 2H),

2.44 (br s, 1H), 1.26 (t, J ) 7.4 Hz, 3H); ¹³C NMR (100 MHz, CDCl₃)

δ 202.0, 137.8, 136.7, 128.5, 128.4 (2), 128.3, 127.8, 127.7, 84.0, 74.3,

73.3, 71.5, 70.0, 22.7, 14.4; LRMS (CI/NH₃) m/z 361 (MH)⁺, 378 (M

+ NH₄)⁺; HRMS (CI/NH₃) exact mass calcd for (C₂₀H₂₅O₄S + H)⁺

requires m/z 361.1474, found m/z 361.1476.

min; 95:5 syn:anti, 95% syn ee; [R]^rt+11.48 (c 4.8, CH₂Cl₂). Syn

D

1

isomer: IR (CH₂Cl₂) 3685, 3569, 2960, 2871, 1677, cm^-1; H NMR

(400 MHz, CDCl₃) δ 7.34 (m, 5H, Ph), 4.55 (d, J ) 11.8 Hz, 1H),

4.52 (d, J ) 11.8 Hz, 1H), 3.92 (dt, J ) 3.8, 6.6 Hz, 1H), 3.53 (dd, J

) 3.8, 9.6 Hz, 1H), 3.45 (dd, J ) 6.6, 9.6 Hz, 1H), 2.90 (ddd, J ) 3.6,

6.7, 10.7 Hz, 1H), 2.87 (q, J ) 7.4 Hz, 2H), 2.41 (br s, 1H), 1.73 (ddd,

J ) 4.1, 10.8, 13.3 Hz, 1H), 1.58 (dd septet, J ) 4.1, 10.0, 6.6 Hz,

1H), 1.47 (ddd, J ) 3.6, 9.8, 13.3 Hz, 1H), 1.23 (t, J ) 7.4 Hz, 3H),

0.92 (d, J ) 6.6 Hz, 3H), 0.90 (d, J ) 6.6 Hz, 3H); ¹³C NMR (100

MHz, CDCl₃) δ 202.1, 137.7, 128.4, 127.8(2), 73.4, 71.6, 71.5, 55.1,

37.7, 25.9, 23.7, 23.4, 21.6, 14.5; LRMS (CI/NH₃) m/z 311 (MH)⁺,

328 (M + NH₄)⁺; HRMS (CI/NH₃) exact mass calcd for (C₁₇H₂₆O₃S

+ NH₄)⁺requires m/z 328.1946, found m/z 328.1949.

Preparation of (1′S,4S)-4-(2′-Benzyloxy-1′-hydroxyethyl)-2-ox-

acyclopentan-1-one (20, Eq 19). Compound 20 was prepared according

to the general procedure employing [Cu(Ph-pybox)](SbF₆)₂(3.0 mL,

0.05 mmol, 10 mol %) and the silylketene acetal derived from

γ-butyrolactone (106 µL, 0.60 mmol) to provide the pure product as a

white powder in 95% yield (112 mg, 0.47 mmol) after flash chroma-

tography with 50% EtOAc/hexanes. Product ratios were determined

by HPLC with a Chiralcel OD-H column (90:10 hexanes/2-propanol;

1.0 mL/min): syn-(1′R,4R) t_r) 15.6 min; syn-(1′S,4S) t_r) 18.3 min;

anti enantiomers t_r) 20.9 min, 27.0 min; 96:4 syn:anti; 95% syn ee;

[R]^rt-8.34 (c 4.3, CH₂Cl₂). Syn isomer: mp 47.0-48.0 °C (white

Preparation of (2S,3S)-tert-Butyl 4-Benzyloxy-3-hydroxy-2-meth-

ylbutanethioate (Table 5, Entry 5). This compound was prepared

according to the general procedure employing [Cu(Ph-pybox)](SbF₆)₂

(2.0 mL, 0.05 mmol, 10 mol %) and the silylketene acetal of tert-butyl

thiopropionate as a 95:5 mixture of Z:E isomers (131 mg, 0.60 mmol,

149 µL). The product was obtained as a clear oil in 86% yield (127

mg, 0.43 mmol) after flash chromatography with 10-20% EtOAc/

hexanes. Product ratios were determined by HPLC with a Chiralcel

OD-H column (99:1 hexanes/2-propanol; 1.0 mL/min): syn-(2R,3R) t_r

) 14.7 min; syn-(2S,3S) t_r) 15.9 min; anti-(2S,3R) t_r) 17.3 min;

anti-(2R,3S) t_r) 20.9 min; 85:15 syn:anti; 99% syn ee. The analytical

data obtained from this material (¹H NMR, ¹³C NMR, and HRMS)

were identical to those previously reported:^4c[R]^rt_D+40.9 (c 3.8, CH₂-

D

powder, racemic); IR (CH₂Cl₂) 3672, 3600, 3056, 2923, 2862, 1764

cm^-1; ¹H NMR (400 MHz, CDCl₃) δ 7.32 (m, 5H), 4.57 (d, J ) 11.9

Hz, 1H), 4.52 (d, J ) 11.9 Hz, 1H), 4.35 (dt, J ) 3.2, 8.8 Hz, 1H),

4.31 (m, 1H), 4.17 (dt, J ) 7.2, 9.1 Hz, 1H), 3.53 (d, J ) 5.7 Hz, 2H),

3.51 (br s, 1H), 2.73 (dt, J ) 3.7, 9.6 Hz, 1H), 2.36 (dq, J ) 12.6, 9.3

Hz, 1H), 2.15 (dddd, J ) 3.2, 7.2, 9.4, 12.6 Hz, 1H); ¹³C NMR (100

MHz, CDCl₃) δ 178.1, 137.6, 128.3, 127.7, 127.6, 73.2, 71.9, 68.0,

67.0, 42.4, 22.3; LRMS (EI) m/z 236 (M)⁺; HRMS (EI) exact mass

calcd for (C₁₃H₁₆O₄)⁺requires m/z 236.1049, found m/z 236.1040.

Preparation of (1′R,4S)-4-(2′-Benzyloxy-1′-hydroxyethyl)-cyclo-

pent-2-enone (21, Eq 20). Compound 21 was prepared according to

the general procedure employing [Cu(Ph-pybox)](SbF₆)₂(3.0 mL, 0.05

mmol, 10 mol %) and 2-(trimethylsilyloxy)furan²⁴(101 µL, 0.60 mmol)

to provide the pure product as a clear oil in 90% yield (106 mg, 0.45

mmol) after flash chromatography with 60% EtOAc/hexanes. Product

ratios were determined by HPLC with a Chiralcel OD-H column (90:

10 hexanes/2-propanol; 1.0 mL/min; anti-(1′S,4R) t_r) 16.7 min; syn

enantiomers t_r) 21.8 min, 23.0 min; anti-(1′R,4S) t_r) 26.1 min; 91:9

Cl₂); [R]²⁶(lit.^4c) +10.3° (c 1.0, CHCl₃); syn:anti 8:92; 90% anti ee

D

(2S,3R).

Preparation of (2S,3S)-Ethyl 4-Benzyloxy-3-hydroxy-2-methyl-

butanoate (Table 5, Entry 7). This compound was prepared according

to the general procedure employing [Cu(Ph-pybox)](SbF₆)₂(3.0 mL,

0.05 mmol, 10 mol %) and the silylketene acetal of ethyl propionate

as an 85:15 mixture of E:Z isomers (131 mg, 0.75 mmol, 163 µL),

except that the reaction was performed at -95 °C (liquid N₂/hexanes

bath). The product was obtained as a clear oil in 60% yield (76 mg,

0.30 mmol) after flash chromatography with 10-20% EtOAc/hexanes.

Deprotection of the TMS ether using 1 N HCl also caused retroaldol

reaction, as observed by TLC (30% EtOAc/hexanes). This decomposi-

tion accounts for the low yield, as ¹H NMR indicated 95% conversion.

Product ratios were determined by HPLC with a Chiralcel OD-H

column (90:10 hexanes/EtOAc; 0.5 mL/min): syn-(2R,3R) t_r) 27.8

min; syn-(2S,3S) t_r) 29.8 min; anti-(2S,3R) t_r) 31.1 min; anti-(2R,3S)

t_r) 32.8 min; 84:16 syn:anti; 87% syn ee; [R]^rt_D+7.78° (c 3.6, CH₂-

Cl₂). Syn isomer: IR (CH₂Cl₂) 3684, 3581, 3060, 2932, 2863, 1727

cm^-1; ¹H NMR (400 MHz, CDCl₃) δ 7.34 (m, 5H), 4.56 (d, J ) 12.0

Hz, 1H), 4.53 (d, J ) 12.0 Hz, 1H), 4.12 (q, J ) 7.1 Hz, 2H), 4.07

(app q, J ) 5.6 Hz, 1H), 3.52 (dd, J ) 4.6, 9.5 Hz, 1H), 3.48 (dd, J )

6.2, 9.5 Hz, 1H), 2.75 (br s, 1H), 2.65 (dq, J ) 5.6, 7.2 Hz, 1H), 1.24

(t, J ) 7.2 Hz, 3H), 1.21 (d, J ) 7.2 Hz, 3H); ¹³C NMR (100 MHz,

CDCl₃) δ 175.2, 137.8, 128.4, 127.8, 127.7, 73.4, 71.6, 70.9, 60.6, 42.0,

14.1, 11.9; LRMS (CI/NH₃) m/z 253 (MH)⁺, 270 (M + NH₄)⁺; HRMS

(CI/NH₃) exact mass calcd for (C₁₄H₂₀O₄+ H)⁺requires m/z 253.1440,

found m/z 253.1437.

(2S,3R)-Ethyl 2,4-Bis(benzyloxy)-3-hydroxybutanethioate (Table

5, Entry 9). This compound was prepared according to the general

procedure employing [Cu(Ph-pybox)](SbF₆)₂(2.0 mL, 0.05 mmol, 10

mol %) and the silylketene acetal derived from ethyl benzyloxythio-

acetate as a 90:10 mixture of Z:E isomers (168 µL, 0.60 mmol) to

provide the pure product as a clear oil in 50% yield (93 mg, 0.25 mmol)

after flash chromatography with 30% EtOAc/hexanes. Product ratios

were determined by HPLC with a Chiralcel OD-H column (89.3:0.7:

10 hexanes/2-propanol/EtOAc; 1.0 mL/min): syn-(2S,3R) t_r) 16.2 min;

syn-(2R,3S) t_r) 17.0 min; anti-(2R,3R) t_r) 18.0 min; anti-(2S,3S) t_r

) 23.1 min; 74:26 syn:anti; 76% syn ee; [R]_D^rt-51.0 (c 1.5, CH₂Cl₂).

Syn isomer: IR (CH₂Cl₂) 3569, 3056, 2933, 2872, 1677 cm^-1; ¹H NMR

(400 MHz, CDCl₃) δ 7.36 (m, 10H), 4.84 (d, J ) 11.1 Hz, 1H), 4.51

(d, J ) 11.9 Hz, 1H), 4.46 (d, J ) 11.1 Hz, 1H), 4.45 (d, J ) 11.9 Hz,

anti:syn; 92% anti ee; [R]^rt-77.1 (c 5.1, CH₂Cl₂). Anti isomer: IR

D

(CH₂Cl₂) 3682, 3569, 3067, 2913, 2872, 1790, 1759, 1161, 1090, 1045

1

cm^-1; H NMR (400 MHz, CDCl₃) δ 7.67 (dd, J ) 1.3, 5.7 Hz, 1H),

7.33 (m, 5H), 6.16 (dd, J ) 1.8, 5.7 Hz, 1H), 5.05 (dt, J ) 6.9, 1.7 Hz,

1H), 4.60 (d, J ) 11.9 Hz, 1H), 4.57 (d, J ) 11.8 Hz, 1H), 3.76 (dt,

J ) 6.8, 3.8 Hz, 1H), 3.73 (dd, J ) 3.7, 9.5 Hz, 1H), 3.70 (dd, J )

4.2, 9.5 Hz, 1H), 2.71 (br s, 1H); ¹³C NMR (100 MHz, CDCl₃) δ 172.7,

155.1, 137.3, 128.6, 128.1, 127.9, 122.1, 82.7, 73.7, 71.3, 70.4; LRMS

(CI/NH₃) m/z 235 (MH)⁺, 252 (M + NH₄)⁺; HRMS (CI/NH₃) exact

mass calcd for (C₁₃H₁₄O₄+ NH₄)⁺requires m/z 252.1236, found m/z

252.1238.

Preparation of (S)-5-Benzyloxy-4-hydroxypentan-2-one (Table

6, Entry 1). This compound was prepared according to the general

procedure employing [Cu(Ph-pybox)](SbF₆)₂(4.0 mL, 0.10 mmol, 100

mol %); except that less benzyloxyacetaldehyde (14 µL, 0.10 mmol)

and the enolsilane of acetone (33 µL, 0.20 mmol) were used. The pure

product was obtained as a clear oil in 96% yield (20 mg, 96 µmol)

after flash chromatography with 40% EtOAc/hexanes. Enantiomeric

excess was determined by HPLC with a Chiralcel OJ column (95:5

hexanes/2-propanol; 1.0 mL/min): (R) enantiomer t_r) 30.1 min; (S)

enantiomer t_r) 32.0 min; 98% ee. The analytical data obtained for

this material were identical in all respects to those obtained for the

decarboxylation product of (3S,5S)-methyl 6-benzyloxy-3,5-dihydroxy-

hexanoate (see Supporting Information). Thus, the acetone enolsilane

adduct possesses an (S) configuration: [R]^rt_D-13.8 (c 0.58, CH₂Cl₂);

IR (CH₂Cl₂) 3678, 3572, 3031, 2925, 2862, 1710 cm^-1; ¹H NMR (400

MHz, CDCl₃) δ 7.34 (m, 5H), 4.57 (d, J ) 12.2 Hz, 1H), 4.54 (d, J )

12.2 Hz, 1H), 4.26 (quintet, J ) 5.6 Hz, 1H), 3.49 (dd, J ) 4.6, 9.6

Hz, 1H), 3.44 (dd, J ) 6.0, 9.6 Hz, 1H), 2.99 (br s, 1H), 2.69 (dd, J )

7.3, 17.0 Hz, 1H), 2.63 (dd, J ) 4.4, 17.1 Hz, 1H), 2.18 (s, 3H, Me);

¹³C NMR (100 MHz, CDCl₃) δ 208.6, 137.8, 128.4, 127.8, 127.7, 73.4,

73.1, 66.8, 46.6, 30.8; LRMS (EI) m/z 208 M⁺; HRMS (EI) exact mass

calcd for (C₁₂H₁₆O₃)⁺requires m/z 208.1099, found m/z 208.1109.

Preparation of (S)-4-Benzyloxy-3-hydroxy-1-phenylbutan-1-one

(Table 6, Entry 2). This compound was prepared according to the

Aldol Addition of Enolsilanes to (Benzyloxy)acetaldehyde

J. Am. Chem. Soc., Vol. 121, No. 4, 1999 685

1

general procedure employing [Cu(Ph-pybox)](SbF₆)₂(4.0 mL, 0.10

mmol, 100 mol %), except that less benzyloxyacetaldehyde (14 µL,

0.10 mmol) and the enolsilane of acetophenone (20 µL, 0.11 mmol)

were used. The pure product was obtained as a white crystalline solid

in 80% yield (21 mg, 80 µmol) after flash chromatography with 20%

EtOAc/hexanes. Enantiomeric excess was determined by HPLC with

a Chiralcel AS column (90:10 hexanes/2-propanol; 1.0 mL/min): (S)

enantiomer t_r) 13.1 min; (R) enantiomer t_r) 33.9 min; 94% ee; mp

46.5-47.0 °C; [R]^rt_D-21.4 (c 0.46, CH₂Cl₂); IR (CH₂Cl₂) 3668, 3583,

3021, 2904, 2862, 1678 cm^-1; ¹H NMR (400 MHz, CDCl₃) δ 7.96 (dt,

J ) 8.3, 1.4 Hz, 2H), 7.59 (tt, J ) 1.4, 7.4 Hz, 1H), 7.47 (dt, J ) 1.6,

7.6 Hz, 2H), 7.34 (m, 5H), 4.61 (d, J ) 12.0 Hz, 1H), 4.57 (d, J )

12.0 Hz, 1H), 4.46 (app sextet, J ) 5.2 Hz, 1H), 3.61 (dd, J ) 4.9, 9.7

Hz, 1H), 3.57 (dd, J ) 5.8, 9.7 Hz, 1H), 3.21 (m, 3H); ¹³C NMR (100

MHz, CDCl₃) δ 199.9, 137.9, 136.7, 133.5, 128.6, 128.4, 128.1, 127.8

(2), 73.4, 73.2, 67.1, 41.8; LRMS (EI) m/z 270 M⁺; HRMS (EI) exact

mass calcd for (C₁₇H₁₈O₃)⁺requires m/z 270.1256, found m/z 270.1246.

Preparation of (2S,3S)-1-Benzyloxy-3,5-dimethyl-2-hydroxy-4-

oxohexane (Table 6, Entry 3). This compound was prepared according

to the general procedure employing [Cu(Ph-pybox)](SbF₆)₂(3.0 mL,

0.05 mmol, 10 mol %) and the enolsilane of 2-methyl-3-pentanone as

a 90:10 mixture of E:Z isomers (132 µL, 0.60 mmol) to provide the

pure product as a clear oil in 90% yield (113 mg, 0.45 mmol) after

flash chromatography with 5-10% EtOAc/CH₂Cl₂. Product ratios were

determined by HPLC with a Chiralcel OD-H column (99:1 hexanes/

2-propanol; 1.0 mL/min): anti enantiomers t_r) 18.4, 22.4 min; syn-

(2S,3S) t_r) 19.4 min; syn-(2R,3R) t_r) 23.8 min; 95:5 syn:anti, 90%

syn ee; [R]^rt_D+11.93 (c 2.3, CH₂Cl₂). Syn isomer: IR (CH₂Cl₂) 3686,

Cl₂) 3682, 3600, 3057, 2964, 2882, 1733 cm^-1; H NMR (400 MHz,

CDCl₃) δ 7.32 (m, 5H, Ph), 4.56 (d, J ) 12.1 Hz, 1H), 4.52 (d, J )

12.1 Hz, 1H), 4.25 (m, 1H), 3.54 (dd, J ) 4.2, 9.7 Hz, 1H), 3.49 (dd,

J ) 7.2, 9.6 Hz, 1H), 2.83 (br s, 1H), 2.28 (ddd, J ) 1.3, 8.0, 17.9 Hz,

1H), 2.21 (dddd, J ) 1.0, 3.9, 8.3, 10.7 Hz, 1H), 2.13 (ddd, J ) 8.5,

10.7, 18.5 Hz, 1H), 2.04 (m, 2H), 1.96 (m, 1H), 1.74 (m, 1H); ¹³C

NMR (100 MHz, CDCl₃) δ 222.0, 137.8, 128.4, 127.7, 127.6, 73.3,

72.6, 68.8, 51.1, 38.7, 23.7, 20.7; LRMS (EI) m/z 234 M⁺; HRMS

(EI) exact mass calcd for (C₁₄H₁₈O₃)⁺requires m/z 234.1256, found

m/z 234.1273.

Preparation of (R)-tert-Butyl 4-Benzyloxy-3-hydroxybutaneth-

ioate Using [Cu(tert-Bu-box)](OTf)₂as the Catalyst (7, Table 7,

Entry 1) (General Procedure for [Cu(tert-Bu-box)](OTf)₂in the

Benzyloxyacetaldehyde Aldol Reactions). To a -78 °C solution of

[Cu(tert-Bu-box)](OTf)₂(0.05 mmol, 10 mol %) in CH₂Cl₂(2 mL) was

added benzyloxyacetaldehyde (70.0 µL, 0.50 mmol) followed by the

silylketene acetal of tert-butyl thioacetate (122 mg, 0.60 mmol, 153

µL). The resulting solution was stirred at -78 °C until the aldehyde

was completely consumed (1 h), as determined by TLC (30% EtOAc/

hexanes). The reaction mixture was then filtered through a 1.5- × 8-cm

plug of silica gel with Et₂O (50 mL). Concentration of the ether solution

gave a clear oil, which was dissolved in THF (10 mL) and 1 N HCl (2

mL). After standing at room temperature for 15 min, this solution was

poured into a separatory funnel and diluted with Et₂O (10 mL) and

H₂O (10 mL). After mixing, the aqueous layer was discarded, and the

ether layer was washed with saturated aqueous NaHCO₃(10 mL) and

brine (10 mL). The resulting ether layer was dried over anhydrous

MgSO₄, filtered, and concentrated. Purification by flash chromatography

with 20% EtOAc/hexanes provided the pure (R)-hydroxy ester. Enan-

tiomeric excess was determined by HPLC with a Chiralcel OD-H

column (94.2:0.8:5.0 hexanes/2-propanol/EtOAc; 1.0 mL/min): (R)

enantiomer t_r) 16.3 min; (S) enantiomer t_r) 17.9 min; 91% ee. The

analytical data obtained from this material (¹H NMR, ¹³C NMR, and

HRMS) were identical to those described above, with the exception of

1

3577, 3065, 2974, 2875, 1705 cm^-1; H NMR (400 MHz, CDCl₃) δ

7.32 (m, 5H), 4.52 (s, 2H), 4.02 (dt, J ) 4.8, 5.8 Hz, 1H), 3.47 (dd, J

) 4.7, 9.6 Hz, 1H), 3.41 (dd, J ) 6.2, 9.5 Hz, 1H), 2.96 (dq, J ) 5.7,

7.2 Hz, 1H), 2.75 (septet, J ) 6.9, 1H), 2.50 (br s, 1H), 1.14 (d, J )

7.2, 3H), 1.08 (d, J ) 6.9, 3H), 1.05 (d, J ) 6.9, 3H); ¹³C NMR (100

MHz, CDCl₃) δ 218.4, 137.8, 128.4, 127.8 (2), 73.4, 71.6, 70.8, 45.9,

40.2, 18.1, 17.9, 12.0; LRMS (CI/NH₃) m/z 251 (MH)⁺, 268 (M +

NH₄)⁺; HRMS (CI/NH₃) exact mass calcd for (C₁₅H₂₂O₃+ NH₄)⁺

requires m/z 268.1913, found m/z 268.1920. Anti isomer: IR (CH₂-

the optical rotation, which was of the opposite sign: [R]^rt+10.4 (c

D

2.9, CH₂Cl₂).

Other aldehyde aldol reactions with the [Cu(tert-Bu-box)](OTf)₂

catalyst were performed analogously using the indicated silylketene

acetal and aldehyde.

1

Cl₂) 3687, 3579, 3054, 2974, 2877, 1711, 1604 cm^-1; H NMR (400

MHz, CDCl₃) δ 7.32 (m, 5H), 4.57 (d, J ) 11.9, 1H), 4.51 (d, J )

11.9, 1H), 3.87 (dt, J ) 6.2, 5.2 Hz, 1H), 3.50 (d, J ) 5.1 Hz, 2H),

3.05 (br s, 1H), 3.02 (dq, J ) 6.7, 7.2 Hz, 1H), 2.73 (septet, J ) 6.9,

1H), 1.10 (d, J ) 7.2, 3H), 1.09 (d, J ) 6.9, 3H), 1.06 (d, J ) 6.9,

3H); ¹³C NMR (100 MHz, CDCl₃) δ 219.1, 137.8, 128.4, 127.8 (2),

73.5, 72.9, 72.2, 45.7, 41.0, 17.8 (2), 14.2; LRMS (CI/NH₃) m/z 251

(MH)⁺, 268 (M + NH₄)⁺; HRMS (CI/NH₃) exact mass calcd for

(C₁₅H₂₂O₃+ NH₄)⁺requires m/z 268.1913, found m/z 268.1919.

Preparation of (1′S,2S)-2-(2′-Benzyloxy-1′-hydroxyethyl)cyclo-

pentan-1-one (Table 6, Entry 5). This compound was prepared

according to the general procedure employing [Cu(Ph-pybox)](SbF₆)₂

(3.0 mL, 0.05 mmol, 10 mol %) and the enolsilane derived from

cyclopentanone (107 µL, 0.60 mmol) to provide the pure product as a

white powder in 90% yield (105 mg, 0.45 mmol) after flash chroma-

tography with 30-40% EtOAc/hexanes. Product ratios were determined

by HPLC with a Chiralcel AD column (95:5 hexanes/2-propanol; 1.0

mL/min): anti enantiomers t_r) 17.1, 18.9 min; syn-(1′R,2R) t_r) 20.4

min; syn-(1′S,2S) t_r) 23.3 min; 95:5 syn:anti; 96% syn ee; [R]^rt_D-82.0

(c 4.6, CH₂Cl₂). Syn isomer: [R]^rt_D-90.5 (c 1.3, CHCl₃); >99:1 syn:

anti; 96% syn ee; mp 53-55 °C (white powder, racemic); IR (CH₂-

Acknowledgment. Financial support was provided by the

NSF. Fellowships from the NSF (M.C.K., B.T.C.), NIH

(J.A.M.), and American Cancer Society (C.S.B.) 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

utilized in the preparation of the box and pybox ligands were

generously provided by NSC Technologies.

Supporting Information Available: General experimental

information; absolute and relative stereochemical proofs; minor

diastereomer characterization date; nonlinear effects experi-

ments; silyl crossover experiments; double stereodifferentiating

experiments; EPR spectra; ESI data; and X-ray crystallogrpahic

data (PDF). This material is available free of charge via the

Internet at http://pubs.acs.org.

JA9829822

Article Doi

C2-symmetric copper(II) complexes as Chiral Lewis acids. Scope and mechanism of catalytic enantioselective aldol additions of enolsilanes to (benzyloxy)acetaldehyde

DOI: 10.1021/ja9829822

Source and publish data:

Authors:

Article abstract of DOI:10.1021/ja9829822

Full text of DOI:10.1021/ja9829822

Products guided by the article

R&D Labs maybe for 203515-60-0

Relevant to this article

Hot Product