Palladium-catalyzed asymmetric allylic alkylation of ketone enolates

Article abstract of DOI:10.1002/chem.200400666

Palladium-catalyzed asymmetric allylic alkylation of nonstabilized ketone enolates to generate quaternary centers has been achieved in excellent yield and enantioselectivity. Optimized conditions consist of performing the reaction in the presence of two e

Full text of DOI:10.1002/chem.200400666

Palladium-Catalyzed Asymmetric Allylic Alkylation of Ketone Enolates
Barry M. Trost* and Gretchen M. Schroeder[a]
Abstract: Palladium-catalyzed asym-
metric allylic alkylation of nonstabi-
lized ketone enolates to generate qua-
ternary centers has been achieved in
excellent yield and enantioselectivity.
Optimized conditions consist of per-
forming the reaction in the presence of
two equivalents of LDA as base, one
equivalent of trimethytin chloride as a
Lewis acid, 1,2-dimethoxyethane as the
solvent, and a catalytic amount of a
chiral palladium complex formed from
p-allyl palladium chloride dimer 3 and
cyclohexyldiamine derived chiral ligand
4. Linearly substituted, acyclic 1,3-di-
alkyl substituted, and unsubstituted al-
lylic carbonates function well as elec-
trophiles. A variety of a-tetralones, cy-
clohexanones, and cyclopentanones can
be employed as nucleophiles. The abso-
lute configuration generated is consis-
tent with the current model in which
steric factors control stereofacial differ-
entiation. The quaternary substituted
products available by this method are
versatile substrates for further elabora-
tion.
Keywords: allylation · asymmetric
catalysis · asymmetric synthesis ·
palladium
·
quaternary
stereocenters
Introduction
of the allylic leaving group to generate a p-allyl transition
metal complex, alkylation by the nucleophile to generate a
new transition metal olefin complex, and finally decomplex-
ation which gives the product and returns the transition
metal so that it can re-enter the cycle. When this reaction is
performed in the presence of chiral ligands, asymmetric in-
duction can potentially be achieved at the electrophile and/
or at the nucleophile. Enantioselectivity at the electrophile
has been extensively studied, however, enantioselectivity at
the nucleophile has received far less attention. In order for
chiral ligands to effect stereochemical control in a Pd-cata-
lyzed reaction, they must influence bond-making and bond-
breaking events occurring outside the coordination sphere
of the metal. Discrimination of the enantiotopic faces of the
nucleophile is especially difficult because the nucleophile is
segregated from the chiral environment by the p-allyl
moiety (Figure 1).
The first example in which a prochiral nucleophile was
employed in the transition metal catalyzed AAA was report-
ed in 1978 by Kagan.[6] For the reasons stated above, the
enantioselectivity was rather low. Since this initial report,
only a handful of examples have been subsequently dis-
closed with similar poor results.[7] Chiral palladium com-
plexes have been used almost exclusively to effect this trans-
formation.[8,9] While the progress toward achieving transition
metal catalyzed AAA using prochiral enolates is impressive,
the method is somewhat limited as “soft”, stabilized carban-
ions are typically employed. The extension of this reaction
to include simple ketone enolates such as that of cyclohexa-
Alkylation of ketone enolates is a venerable reaction in or-
ganic synthesis.[1] As this transformation has proven invalua-
ble, the development of asymmetric methods has received
considerable attention.[2] Asymmetric alkylation to generate
quaternary centers, that is carbon centers with four different
non-hydrogen substituents, has proven particularly challeng-
ing. While significant advances toward achieving this goal
have been made by the use of stoichiometric chiral auxilia-
ries,[3] a more attractive strategy would involve using a cata-
lytic amount of the chirality inducing agent.[4] Catalytic
asymmetric transformations benefit in terms of atom econo-
my and require fewer chemical transformations as the chiral
auxiliary is commonly installed and removed in distinct
steps.[5]
Transition metal catalyzed asymmetric allylic alkylation
(AAA reaction) has been shown to be an effective method
for the synthesis of quaternary substituted carbon centers.
The catalytic cycle consists of four steps: coordination of the
transition metal to the olefin of the electrophile, ionization
[a] Prof. B. M. Trost, Dr. G. M. Schroeder
Department of Chemistry
Stanford University
Stanford, CA 94305-5080 (USA)
Fax : (+1)650-725-0002
E-mail: bmtrost@stanford.edu
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
174
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/chem.200400666
Chem. Eur. J. 2005, 11, 174 – 184

FULL PAPER
Table 1. Selected optimization studies in variation of the Lewis acid for
Equation (1).[a]
Entry
Lewis acid
Yield 5 [%][b]
ee 5 [%][c]
1
2
3
4
5
6
7
8
9
Bu3SnCl
53
65
NR
trace
21
28
78
NR
65
65
69
–
Me3SnCl
Bu2Sn(OAc)2
Bu3SnOAc
Bu3SnOTf
Bu3SnI
Bu3SnF
Bu2BOTf
B(OMe)3
–
32
35
À25
–
58
[a] All reactions were performed using 1 equiv LDA, 1.05 equiv Lewis
acid, 1.1 equiv allyl acetate (2), 2.5% 3, and 5% 4 at room temperature.
[b] Isolated yield. [c] Determined by chiral HPLC.
Figure 1. Transition metal catalyzed AAA of prochiral nucleophiles.
none would represent a significant achievement. The com-
patibility of such hard nucleophiles with palladium in the
AAA reaction has since been explored. We recently dis-
closed in a preliminary communication our efforts toward
affecting the palladium-catalyzed AAA of nonstabilized
ketone enolates to generate quaternary centers.[10] Herein,
we describe in detail its successful application.
reaction of tetralone 1 with acetate 2. For example, trime-
thyltin chloride gave allylated product 5 in 69% ee whereas
tributyltin chloride gave 5 in 65% ee (entry 1 vs 2). Another
interesting trend was observed; the yield and enantioselec-
tivity of the reaction was found to correlate with the leaving
group ability of the Lewis acid (entries 1, 3–7). Lewis acids
with poor leaving groups gave better results than those with
good leaving groups. For example, tributyltin chloride
proved superior to the corresponding acetate, triflate, and
iodide. The reversal in the sense of enantioselectivity using
tributyltin fluoride is a remarkable and curious result
(entry 7). Several boranes and borates were also found to be
competent; however, stannanes gave superior enantioselec-
tivity than did the boron derived Lewis acids. For example,
trimethyltin chloride gave better ee (69%) than did trime-
thylborate (58% ee, entry 9). A variety of aluminum,
indium, titanium, and cerium enolates were found to be in-
effective resulting in either little or no reaction. Given these
results, trimethyltin chloride became the Lewis acid of
choice.
Results
Reaction optimization: 2-Methyl-1-tetralone (1) was chosen
for initial examination as the nucleophile in the presence of
allyl acetate (2) as the electrophile, LDA as the base, 1,2-di-
methoxyethane (DME) as the solvent, and a catalytic
amount of a chiral palladium complex formed from p-allyl
palladium chloride dimer 3 and cyclohexyldiamine derived
chiral ligand 4 [Eq. (1)]. A Lewis acid, tributyltin chloride,
was also added with the thought that the Lewis acid would
“soften” the lithium enolate by transmetallation to form the
tin derivative. Gratifyingly, the reaction was found to pro-
ceed to give the desired alkylated product 5 in 53% yield
and moderate enantiomeric excess (65% ee). Given this ini-
tial success the Lewis acid additive, base, solvent, and palla-
dium source were systematically varied.
A dramatic effect on the reaction yield was observed in
variation of the base (Table 2). Enolates generated from
lithium amide bases were found to react readily; whereas,
those generated from sodium and potassium bases resulted
in recovery of the starting ma-
terial (entries 1, 2 vs 3). Fur-
thermore, the reaction was
found to be sensitive to the
amount of base used. For exam-
ple, on increasing the equiva-
lents of LDA from 1 to 1.25,
1.5, and 2, the enantioselectivity
gradually increased (entries 3–
7). With three equivalents of
LDA, a slight decrease in enan-
While the addition of a Lewis acid was thought to help
stabilize the very reactive ketone enolate, it also provided
an additional variable for altering the enolate structure.
Thus, a variety of Lewis acids were examined as additives
(Table 1). One trend observed in variation of the Lewis acid
was a correlation between Lewis acid size and ee. In general,
the smaller the Lewis acid, the higher the ee obtained in the
tioselectivity was observed, thus two equivalents of base
gave the best results and allowed for isolation of allylated
tetralone 5 in excellent yield (99%) and ee (88%). Impor-
tantly, when employing two equivalents of base, the reaction
could be run in the absence of the tin additive and gave the
allylated product in 96% yield and slightly diminished ee
(85%, entry 8). Other lithium amide bases (LiHMDS and
Chem. Eur. J. 2005, 11, 174 – 184
www.chemeurj.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
175

B. M. Trost and G. M. Schroeder
Table 2. Variation of the base for Equation (1).[a]
Entry Base
Equivalents Additive Yield 5 [%][b] ee 5 [%][c]
Table 3. Variation of the ligand for Equation (1).[a]
Entry Ligand
Yield 5
ee 5
[%][b]
[%][c]
1
2
3
4
5
6
7
KHMDS
NaH
LDA
LDA
LDA
LDA
LDA
1
1
1
1.25
1.5
2
Me3SnCl NR
Me3SnCl NR
Me3SnCl 65
Me3SnCl 78
Me3SnCl 99
Me3SnCl 99
Me3SnCl 61
–
–
1
2
4
99
88
69
78
80
88
84
85
61
71
63
86
81
86
3
8
LDA
2
None
96
9
LiHMDS
LiHMDS
LTMP
LTMP
1
2
1
2
Me3SnCl 58
Me3SnCl 94
Me3SnCl 74
Me3SnCl 99
10
11
12
3
90
86
NR=No Reaction. [a] All reactions were performed using 1.1 equiv allyl
acetate (2), 2.5% 3, and 5% 4 at room temperature unless otherwise
noted. [b] Isolated yield. [c] Determined by chiral HPLC.
[a] All reactions were performed with 1.05 equiv allyl acetate, 1 equiv tri-
methyltin chloride, 2 equiv LDA, 2.5% 3, and 5% (S,S)-ligand in DME
at room temperature. [b] Isolated yield. [c] Determined by chiral HPLC.
lithium tetramethylpiperide) also followed this trend with
excess base consistently giving higher levels of enantioselec-
tivity (entries 9–12). The amine of the base, a stoichiometric
by-product in enolization, affected the enantioselectivity of
the reaction as hexamethyldisilazide gave inferior results
(entry 10). On the other hand, changing the amide to piperi-
dide had little effect (entry 12).
alents of LDA as the base, palladium dimer 3, and chiral
ligand 4 in DME as solvent. A slightly higher ee in the pres-
ence of trimethyltin chloride led us to adopt its addition as
part of our standard protocol. Under these conditions, the
allylated product 5 was isolated in 99% yield and 88% ee.
Variation of the reaction solvent was examined. Moderate
enantioselectivity was achieved in all solvents examined in-
cluding THF, a 10% HMPA/THF solvent mixture, DME, di-
chloromethane, and toluene. The highest enantiomeric
excess and shortest reaction times were achieved in DME,
thus it became the solvent of choice.
Two sources of palladium were tried, p-allylpalladium
chloride dimer 3 and dibenzylideneacetone complex
[Pd2dba3]·CHCl3 (6). p-Allylpalladium chloride dimer 3 was
found to give slightly higher ee than dba complex 6. This
may be an indication that the achiral dba competes to some
extent with the chiral ligand for palladium. Importantly, use
of the opposite enantiomer of the ligand gave the opposite
enantiomer as the major product as indicated by chiral
HPLC and the sign of the optical rotation. The control ex-
periment in which no palladium was added gave no reaction
and indicates minimal or no background reaction.
X-ray crystal structures of a related family of ligands sug-
gest that they are not C2-symmetric as one might suppose.[11]
Two non C2-symmetric ligands were tried in an attempt to
ascertain whether a more pronounced deviation from C2
symmetry impacted the reaction and could possibly result in
enhanced enantioselectivity. As shown in Table 3, all reac-
tions gave similar yields and enantioselectivities. Reactions
with naphthyl-phenyl ligand 7 were notably slower than
those with the standard ligand 4. Likewise, reactions using
2-isopropyl-1-tetralone as the nucleophile gave the corre-
sponding allylated product in similar ee (35% with 4 and
28% with 7). Thus non C2-symmetric ligands did not pro-
vide any advantage over standard ligand 4.
Reaction scope—Variation of the electrophile: With opti-
mized conditions in hand, a variety of allylating agents was
explored (Table 4). Crotyl methyl carbonate gave the alkyl-
ated product 9 in good yield (84%) and outstanding ee
(90%) (entry 2). Linearly substituted allyl systems starting
from either E or Z isomers gave products of only E geome-
try in good ee (entry 3). Thus, p–s–p equilibration was
faster than nucleophilic attack. Unfortunately, cinnamyl
methyl carbonate gave very poor conversion (entry 4). The
poor conversion likely results from the formation of a more
stable and therefore less reactive palladium p-allyl complex.
A 1,3-dialkyl allylic carbonate, gave the alkylated product
12 in excellent ee and de, but in disappointing yield (17%).
The yield could be improved on switching to the more reac-
tive allylic phosphate and increasing the amount of electro-
phile added (entry 7). At the end of the reaction, no electro-
phile remained by TLC. Presumably, the yield of this reac-
tion could be further improved by adding additional electro-
phile or by increasing the concentration of the reaction. The
ability to control the stereoselectivity at the nucleophile as
well as at the electrophile is remarkable. A cyclic analogue
gave the alkylated product 13 in good yield, but poor ee
(entry 8). A 1,1-dialkyl system and a branched allyl system
gave disappointing results (entries 9–10). To summarize, lin-
early substituted, acyclic 1,3-dialkyl substituted, and unsub-
stituted p-allyls (allyl acetate) gave the best results.
Reaction scope—Variation of the nucleophile: A variety of
a-tetralone derived nucleophiles were examined in the pal-
ladium-catalyzed AAA with allyl acetate (2) (Table 5).[12,13]
Substitution in the 2-position of a-tetralone was tolerated
when the substituent was not too sterically demanding. The
methyl, ethyl, benzyl, and allyl groups all gave comparable
To summarize, optimization showed that the best condi-
tions for alkylation of 2-methyl-1-tetralone (1) with allyl
acetate (2) consisted of running the reaction with two equiv-
176
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 174 – 184

Asymmetric Allylic Alkylation
FULL PAPER
Table 4. Reaction scope: Variation of the electrophile.[a]
standard reaction conditions to
this substrate gave modest ee
(45%).[14]
Entry
Electrophile
Product
Yield [%][b]
99
ee [%] (de)[c]
The
palladium-catalyzed
1
allyl acetate (2)
88
AAA was applied to some a’-
blocked, a-alkyl cyclohexa-
nones. Benzylidene,[15] furanyl-
idene,[16] and ketene dithioace-
tal[17] substituted cyclohexa-
nones were examined in their
reaction with allyl acetate
(Scheme 2). The benzylidene
cyclohexanone gave allylated
product 24[18] in nearly quantita-
tive yield (98%) and 82% ee at
room temperature. In the case
of the furanylidene derivative, a
temperature effect was ex-
plored. At room temperature,
the product 25 was produced in
only 79% ee (89% yield),
whereas at 08C the enantiomer-
ic excess increased to 92%
(95% yield). A similar effect
was observed with the ketene
dithioacetal substituted cyclo-
hexanone. In this case the reac-
tion at room temperature gave
ketone 26 of 70% ee (64%
yield). At 08C, an improved
79% ee (51% yield) was ob-
tained. Finally, at À108C the al-
lylated product was isolated in
67% yield and 82% ee. Two
equivalents of allyl acetate
were used in the experiment at
À108C to help maintain the
rate of reaction.
5
9
2
3
4
5
84
72
11
16
90
82
–
10
11
>95 (91)
12
12
6
17
41
>95 (>95)
>95 (>95)
7[d]
12
8[e]
87
64
82
37
13
47
13
14
15
9
10
[a] All reactions were performed with 1.05–1.1 equiv electrophile, 1 equiv trimethyltin chloride, 2 equiv LDA,
2.5% 3, and 5% (S,S)-4 in DME at room temperature unless otherwise indicated. [b] Isolated yield. [c] Deter-
mined by chiral HPLC or chiral GC. [d] Two equivalents of electrophile were used. [e] The enantioselectivity
at the nucleophile was determined after hydrogenation of the olefin.
Table 5. Reaction scope: Variation of the nucleophile.[a]
results with ee values varying from 73–88% (entries 1, 2, 4,
5). In the case of allyl substituted a-tetralone, crotyl methyl
carbonate was used as the electrophile (entry 5). Unfortu-
nately, a more bulky isopropyl group gave significantly di-
minished ee (35%, entry 3). A methoxy substituted tetra-
lone also gave good yield (83%) and ee (85%) (entry 6).
The ring size of the nucleophile was varied and the five-
and seven-membered ring analogues were examined
(Scheme 1). Commercially available 2-methyl-1-indanone
gave the allylated product 21 in a disappointing 38% ee
(79% yield) with allyl acetate as the electrophile. Likewise
methyl substituted benzosuberone[13] gave ketone 22 in only
6% ee. Thus, these reaction conditions appear to be limited
to six-membered ring cycloalkanones. A furan analogue of
tetralone 1 was also tried, unfortunately, application of the
Entry
R1
R2
Product, yield [%][b]
ee [%][c]
1d
2
3
CH3
C2H5
CH(CH3)2
CH2Ph
CH2CH=CH2
CH3
H
H
H
H
H
5, 99
16, 96
17, 99
18, 98
19, 71
20, 83
88
80
35
73
85
85
4
5e
6
OCH3
[A] All reactions were performed with 1.1 equiv allyl acetate, 1 equiv tri-
methyltin chloride, 2 equiv LDA, 2.5% 3, and 5% (R,R)-4 in DME at
room temperature unless otherwise indicated. [b] Isolated yield. [c] De-
termined by chiral HPLC. [d] (S,S)-4 was used. [e] Crotyl methyl carbon-
ate was used as the electrophile.
Chem. Eur. J. 2005, 11, 174 – 184
www.chemeurj.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
177

B. M. Trost and G. M. Schroeder
Table 6. Synthesis of ketene dithioacetal 27.[a]
Entry
Additive
T[8C]
X
Yield 27 [%][b]
ee 27 [%][c]
1
Me3SnCl
Me3SnCl
Me3SnCl
Me3SnCl
None
rt
0
À15
À20
0
OAc
OAc
Cl
Cl
OAc
35
64
53
trace
40
52
64
70
–
2[d]
3[d]
4[d]
5[d]
58
[a] All reactions were performed with 1.1 equiv electrophile, 1 equiv ad-
ditive, 2 equiv LDA, 2.5% 3, and 5% (S,S)-4 in DME unless otherwise
noted. [b] Isolated yield. [c] Determined by chiral HPLC. [d] Two equiva-
lents of electrophile were used.
Scheme 1. Variation of the ring size.
cetal 27, the thioalkyl group was varied. The ethyl derivative
was tried next with the thought that increasing the steric
bulk of the ketene dithioacetal would allow for more effec-
tive facial discrimination of the approaching nucleophile by
the chiral ligand [Eq. (2)]. Disappointingly, palladium cata-
lyzed AAA with allyl acetate at 08C gave the allylated prod-
uct in poor yield (42%) and ee (49%).
A more rigid cyclic ketene dithioacetal was also examined
[Eq. (3)]. Reaction under the standard conditions gave the
allylated product 29 in good yield (76%) and modest ee
(57%) at room temperature. The improved yield stems from
increased stability of the cyclic ketene dithioacetal to the re-
action conditions as noted by the fact that the reactions
were noticeably cleaner. Like the previous examples, the ee
could be increased to 63% by lowering the reaction temper-
ature to 08C. Unfortunately, further decreases in the reac-
tion temperature gave diminished yield.
Scheme 2. AAA of cyclohexanones.
Due to the disappointing results with ketene dithioacetals,
we turned our attention to enol ether nucleophiles. Reaction
of a tert-butyl enol ether with allyl acetate under the stan-
dard reaction conditions was examined to give allylated
product 30 [Eq. (4), Table 7)]. Surprisingly, the enantioselec-
tivity of the reaction varied widely with different bottles of
n-butyllithium. It was thought that perhaps the difference in
behavior was due to the presence of varying amounts of alk-
oxides in the bottles, thus tert-butanol was added to the re-
action (entries 1–4). A clear dependence of the enantioselec-
tivity of the reaction on the amount of tert-butanol added
was observed. A gradual increase in ee was observed as the
Given our success with the six-membered ring ketene di-
thioacetal and the versatility of the ketene dithioacetal
moiety, five membered-ring analogues were tried
(Table 6).[16] Application of the standard reaction conditions
to a methyl ketene dithioacetal derivative gave the allylated
product 27 in low yield (35%) and modest ee (52%)
(entry 1). Decreasing the reaction temperature did give im-
proved yield (64%) and ee (64%, entry 2), however further
lowering of the reaction temperature led to poor conversion.
Switching to a more reactive electrophile, allyl chloride, al-
lowed the reaction to proceed at À158C and gave 27 in
70% ee (53% yield, entry 3).
Once again, further reductions
in the reaction temperature
failed to give improved results
and resulted in recovery of the
starting material (entry 4). Per-
forming the reaction in the ab-
sence of trimethyltin chloride
had a slight negative effect on
the ee (58%, entry 5).
Due to the modest yield and
ee in generating ketene dithioa-
178
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 174 – 184

Asymmetric Allylic Alkylation
FULL PAPER
Table 7. Synthesis of tert-butyl enol ether 30.[a]
as the electrophile and gave the desired product in similar
yield and ee (entry 5). To conclude, alkylated ketone 31
could be obtained in 70–80% yield and 81–88% ee under a
variety of conditions.
The AAA reaction of a nitrogen analogue was briefly ex-
amined [Eq. (5)].[19] The traditional conditions without tert-
butanol were tried and a pronounced temperature effect
was observed. Thus, on decreasing the reaction temperature
from room temperature to 0 to À408C, the ee of product 32
increased from 50 to 62 to 75% without affecting the yield
of the reaction (54, 62, and 55%, respectively). No further
attempts were made to increase the yield and enantioselec-
tivity of this reaction.
Entry
tBuOH (equiv)
Yield 30 [%][b]
ee 30 [%][c]
1
2
3
1
3
5
7
7
85
86
80
87
83
15
79
89
91
95
4
5[d]
[a] All reactions were performed with 1.1 equiv allyl acetate, 1 equiv tri-
methyltin chloride, 2 equiv LDA, 2.5% 3, and 5% (S,S)-4 in DME at
room temperature unless otherwise noted. [b] Isolated yield. [c] Deter-
mined by chiral HPLC. [d] 2.2 Equiv electrophile and 1 mol% catalyst
were used.
amount of tert-butanol added was increased from one to
three to five and finally to seven equivalents. The ee seemed
to plateau at 7 equiv tert-butanol, thus these became the
conditions of choice and allowed for isolation of 30 reliably
The allylated products generated in this reaction are ver-
satile substrates for further transformations. For example,
1,3-carbonyl transposition[20] of keto ketene dithioacetal 26
gave a,b-unsaturated thiol ester 33, a useful substrate for an-
nulation protocols[21] [Eq. (6)].
The use of mercuric chloride to
effect dehydration of the inter-
mediate alcohol was critical as
use of 10% HCl/MeOH or
acidic silica gel gave significant
in 87% yield and 91% ee
(entry 4). Gratifyingly, the cata-
lyst loading could be reduced to
1 mol% without significant loss
of chemical yield and with an
improved 95% ee (entry 5).
The six-membered ring ana-
logue was also tried in this reac-
tion. Surprisingly, the cyclohexanone derivative was found
to give the product in very low ee in the presence and ab-
sence of tert-butanol (5 and 1%, respectively). The reason
for this dramatic difference in reaction is not clear.
With optimized conditions in hand, the reaction was tried
with crotyl derived electro-
amounts of side products resulting from capture of the
stable allylic carbocation intermediate by methanethiol.
Likewise, tert-butyl enol ether 31 proved a versatile com-
pound for further elaboration (Scheme 3). Treatment of 31
with methyllithium gave aldehyde 34 in 81% yield. The in-
philes [Eq. (4), Table 8]. In the
presence of crotyl methyl car-
bonate, the alkylated product
31 was isolated in 71% yield
and 81% ee (entry 1). The reac-
tion was somewhat slow and re-
quired several hours, thus the
more reactive crotyl chloride
was tried. At 08C, the reaction
gave 31 in 71% yield (84% ee,
Table 8. Synthesis of tert-butyl enol ether 31.[a]
Entry
Electrophile
(equiv)
T [8C]
Yield 31 [%][b]
ee 31 [%][c]
entry 2).
The enantiomeric
1
(1.1)
rt
71
81
excess increased to 87% when
the reaction was performed at
2
3
crotyl chloride
crotyl chloride
crotyl chloride
(2.2)
(2.2)
(2.2)
0
À20
0
71
61
74
84
87
84
À208C (entry 3). Also, the cata- 4[d]
lyst loading could be decreased
to 2 mol% without affecting
the yield or ee (entry 4). A
5
(2.2)
0
73
82
[a] All reactions were performed with 1 equiv trimethyltin chloride, 2 equiv LDA, 2.5% 3, and 5% (S,S)-4 in
DME unless otherwise noted. [b] Isolated yield. [c] Determined by chiral HPLC. [d] Two mol% catalyst was
crotyl phosphate was also tried used.
Chem. Eur. J. 2005, 11, 174 – 184
www.chemeurj.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
179

B. M. Trost and G. M. Schroeder
formed. Strong NOE values
(2.9–4.7%) were observed be-
tween the benzylic hydrogen
and the methyl group or alkyl
benzoate. With knowledge of
the relative stereochemistry in
hand, the characteristic shield-
ing effect of the mandelate
phenyl group allowed for de-
termination of the absolute
stereochemistry as shown in
Scheme 3. Versatility of enol ether 31.
termediate alcohol was never isolated as it eliminated tert-
butanol upon work-up. Remarkably, this reaction accom-
plished a 1,3-carbonyl transposition while forming a new
Figure 2.[23]
1
In the H NMR spectra of ester 40, the signal for the aro-
matic proton (Hd) was shifted to higher field than the corre-
sponding proton in ester 41. As illustrated in the Newman
projection of 40, Hd is shielded by the O-methylmandelate
esterꢁs phenyl ring only if the absolute stereochemistry of
the quaternary center is (R). By analogy, in the 1H NMR
spectra of ester 41 the signals for the methyl group (Ha) as
well as Hb and Hc were shifted to higher field than the cor-
responding protons in ester 40. Again, the Newman projec-
tion 41 illustrates the shielding by the O-methylmandelate
esterꢁs phenyl ring which can only occur if the quaternary
center is (R).
À
carbon carbon bond in a single step. On the other hand,
treatment of enol ether 31 with two equivalents of lithium
dimethylcuprate gave ketone 35 in 86% yield as a 1:1 mix-
ture of diastereomers. Thus, addition of one equivalent of
lithium dimethylcuprate was followed by elimination of tert-
butanol and addition of a second equivalent of lithium di-
methylcuprate to give the desired product. The intermediate
enone was never observed as the second addition was much
more facile than the first.
To conclude, a variety of a-tetralone, cyclohexanone, and
cyclopentanone nucleophiles can be employed in the palla-
dium-catalyzed AAA with a range of electrophiles to give
the corresponding alkylated products in good yield and ee.
For cyclohexanones, the benzylidene, furanylidene, or
ketene dithioacetal groups are compatible with the reaction.
For cyclopentanones, a tert-butyl enol ether gave excellent
results.
Discussion
In palladium-catalyzed AAA of prochiral nucleophiles, the
structure of the nucleophile plays a critical role in discrimi-
Determination of the absolute
stereochemistry: The absolute
configuration of allylated tetra-
lone
5
was determined as
shown in Scheme 4. The termi-
nal olefin of 5 was hydroborat-
ed and oxidized to give the cor-
responding primary alcohol 36
in 88% yield.[22] Reduction of
the ketone with sodium borohy-
dride gave a diasteromeric mix-
ture of diols 37. Chemoselective
protection of the primary alco-
hol as the benzoate allowed for
separation of the two diaster-
eomers by column chromatog-
raphy. Each diastereomer was
then reacted with (S)- and (R)-
(O-methoxy)mandelic acid to
give
enantiomerically
pure
mandelate esters 40–43. To de-
termine the relative configura-
tion between the benzylic alco-
hol and the adjacent quaternary
center, NOE studies were per- Scheme 4. Assignment of the absolute configuration.
180
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 174 – 184

Asymmetric Allylic Alkylation
FULL PAPER
the allylated product was ob-
tained in only 70% yield and
67% ee.
To address the possibility
that the amine by-product from
the lithium amide base was af-
fecting the reaction, the silyl
enol ether of 2-methyl-1-tetra-
lone (1) was prepared and
treated with allyl acetate
[Eq. (7)]. Alkylation was ach-
ieved by cleaving the silyl enol
ether with nBuLi, generating a
reaction mixture which was free
from the stoichiometric amine
by-product. The enantiomeric
excess of this reaction was de-
termined to be much lower
(67%) than that in the presence
Figure 2. Extended Newman projections of mandalate esters 40–43.
of the amine by-product (88%).
nation of the enantiotopic facial approaches by the chiral
catalyst. The importance of enolate structure is evidenced
by the strong dependence of reactivity and enantioselectivity
on the choice of base, Lewis acid additive, and solvent.
Choice of base proved critical as only lithium enolates were
competent in the reaction. One possible explanation for this
observed reactivity is that the more oxaphilic lithium atom
is better able to stabilize the very reactive ketone enolate.
The harder sodium and potassium enolates may react with
the palladium catalyst rendering it catalytically inactive. As
an alternative explanation, the effect of lithium might result
from the difference in size of the enolate counterions with
lithium being the smallest and therefore least sterically de-
manding counterion, allowing the nucleophile to access the
chiral pocket created by the palladium and ligand. One un-
known issue is the effect of aggregation of the various eno-
lates on both reactivity and enantioselectivity.
Several explanations for the dependence of the ee on the
amount of base can be proposed: one, the excess base could
deprotonate the ligand, thereby altering the nature of the
chiral environment; two, the excess base could form an ag-
gregate with the nucleophile, changing the nature of the nu-
cleophile; three, the excess base could hydrogen bond with
the amine by-product generated during enolization, thereby
preventing the amine from associating with the nucleophile;
last, the excess base could react with the Lewis acid addi-
tive, trimethyltin chloride, forming the corresponding amino
stannane.
This experiment clearly indi-
cates a role for the amine in the reaction. Further evidence
for this assumption can be found in Table 2 which demon-
strates a dependence of enantioselectivity on the choice of
lithium amide base with LDA and LTMP giving higher ee
than LiHMDS.
To investigate the possibility that the aminostannane is
formed from trimethyltin chloride and excess LDA, (diiso-
propylamino)trimethyltin was independently synthesized[24]
and tried in the AAA of the silyl enol ether of 1 [Eq. (8)].
After cleaving the silyl enol ether with one equivalent of
nBuLi and treating the resulting lithium enolate with one
equivalent of the aminostannane, the allylated product 5
was isolated in quantitative yield and 82% ee. Since it was
shown that the analogous reaction using trimethyltin chlo-
ride generated the product in just 63% yield and 67% ee
[Eq. (7)], these results support the idea that the amino stan-
nane is formed and may be responsible at least in part for
the enhanced enantioselectivity when two equivalents of
base are employed. Presumably, the only difference between
the reaction conditions in Equation (8) and the standard
conditions is the presence of lithium chloride. The reaction
depicted in Equation (8) was performed in the presence of
one equivalent of lithium chloride to see if the ee would rise
to 88%. Tetralone 5 was isolated in quantitative yield and
83% ee, therefore the addition of lithium chloride failed to
further enhance the enantioselectivity. It should be noted
To explore the hypothesis that excess LDA was exerting
its beneficial effect by deprotonation of the amide hydro-
gens of the ligand, the ligand was intentionally deprotonated
prior to addition to the reaction mixture. The palladium cat-
alyst, prepared from palladium dimer 3 and ligand 4, was
deprotonated with trityllithium to presumably generate the
dianion of the ligand and then the dianion was used in the
usual manner in the AAA, but using just one equivalent of
base to generate the enolate. Contrary to this hypothesis,
Chem. Eur. J. 2005, 11, 174 – 184
www.chemeurj.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
181

B. M. Trost and G. M. Schroeder
that the low solubility of lithium chloride in DME compli-
cates the interpretation of this result.
trophile may lie in the fact that the p-allyl is canted such
that the substituent in the 2-position is pointed toward the
ligand and results in an unfavorable steric interaction. The
poor results obtained with a cyclic electrophile can also be
explained in terms of the structure of the p-allyl palladium
complex. Acyclic electrophiles can form a variety of palladi-
um p-allyl complexes with the syn–syn form being the
lowest energy (Figure 4). Cyclic electrophiles can only exist
in the anti–anti conformation thus the approaching nucleo-
phile likely encounters a very different environment.
Given the dependence of enantioselectivity on the amine
portion of the base and the hypothesis that an aminostan-
nane is formed in the reaction, one would expect a depend-
ence of ee on the amine portion of the aminostannane. This
was indeed found to be the case. Replacement of one of the
isopropyl groups with tert-butyl gave the allylated product in
similar yield and ee. However, when the reaction was run in
the presence of (dimethylamino)trimethyltin the yield and
ee dropped to 48 and 73%, respectively. Also, replacement
of the amine group by methoxy resulted in lower ee (75%,
60% yield).
Figure 4. Palladium p-allyl complexes.
The absolute configuration generated from (S,S)-4 is con-
sistent with the current model (Figure 5). In this model, the
wall above palladium represents the cyclohexane ring of the
ligand and the flaps on either side of palladium represent
the phenyl groups on phosphorus. The (S,S)-chiral ligand
generates (R)-allylated product 5 in the allylation of 2-
methyl-1-tetralone (1). As illustrated by the model, the
facial approach leading to (R)-5 allows the nucleophile to
approach the p-allyl palladium complex in the open space
created by a raised flap. In the alternative approach, the nu-
cleophile encounters unfavorable steric interactions with
one of the flaps or phenyl groups of the chiral ligand. Thus,
the trajectory of approach for the (R)-product is less steri-
cally hindered than the approach which generates (S)-5. The
dependence of the enantioselectivity of the reaction on the
substituent in the 2-position of a-tetralone can be rational-
ized with this model. The decrease in ee with larger substitu-
ents results as the substituent is positioned such that it may
interfere with the lowered flap in the trajectory that gives
the (R)-product. The alternative facial approach allows for
the large substituent to be directed toward the open space
created by the raised flap and places the aryl portion of the
tetralone towards to lowered flap. This competition between
the large substituent and the aryl portion of tetralone for
the open space results in the lowered enantioselectivity.
While the cartoon in Figure 5 depicts a C2-symmetrical
complex, X-ray crystal structures suggest that this may not
be a true representation of the catalyst. The unsymmetrical
nature of the chiral complex may lead to a memory effect
where its equilibration is required for good enantioselectivi-
ty. Thus, fast reactions can lead to low enantioselectivity and
slowing the reaction, for example by lowering the reaction
temperature, can result in increased ee.
An alternative explanation for the role of the Lewis acid
additive could be that an ate complex rather than a simple
trialkylstannyl ether may more accurately describe the nu-
cleophile (Figure 3). The reactivity and
enantioselectivity correlated with the
leaving group ability on the tin (Table 1).
Lewis acids with poor leaving groups
gave better results than those with good
leaving groups. The amine by-product
could then be exerting its effect on this
species by displacement of the chloride or
by affecting the state of aggregation.
Further evidence for the importance of
Figure 3. Tin–Ate
complex.
the state of aggregation comes from the
dependence of the enantioselectivity on the choice of sol-
vent. The ee increased on going from toluene to THF and
then to DME. The ability of solvent to affect the state of ag-
gregation of the enolate may be the source of this selectivity.
Of the solvents examined, DME is the solvent most capable
of breaking up aggregates. The lithium enolate of a-tetra-
lone has been shown to be predominately a monomer–tet-
ramer equilibrium in THF, but with the addition of a 2-iso-
propyl group the enolate is predominately a dimer.[25] Stud-
ies of 2-phenyl-1-tetralone show that the lithium enolate
exists as a monomer–dimer mixture and that the equilibrium
lies in favor of the dimer (>90%) at concentrations usually
used in synthesis (>0.1m).[26] However, in the case of 2-
phenyl-1-tetralone, the dimer was shown to be much less re-
active in alkylation reactions than the monomer. Clearly,
the exact nature of the nucleophile is a complicated picture
with a pronounced effect on this reaction.
Conclusion
Linearly substituted, 1,3-disubstituted, and unsubstituted
(allyl acetate) electrophiles functioned well, however, substi-
tution in the two position of the p-allyl was not tolerated.
The poor enantioselectivity obtained with a 2-methallyl elec-
The alkylation of simple ketone enolates of cycloalkanones
can now by achieved asymmetrically in a catalytic fashion.
The optimum conditions consist of performing the reaction
182
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 174 – 184

Asymmetric Allylic Alkylation
FULL PAPER
0.18m. The cooling bath was removed
and the reaction was allowed to stir at
room temperature until the reaction
was complete. The product was puri-
fied directly by flash chromatography
on silica gel.
A typical example is given as follows:
nBuLi (370 mL, 0.594 mmol) was
added at À788C to a solution of fresh-
ly distilled diisopropylamine (83 mL,
0.594 mmol) in DME (0.4 mL). After
stirring at À788C for 15 min, a so-
lution of 2-methyl-1-tetralone (1)
(48 mg, 0.297 mmol) in DME (0.4 mL)
was added. After stirring at 08C for
15 min, a solution of trimethyltin chlo-
ride (59 mg, 0.297 mmol) in DME
Figure 5. Rationale for chiral recognition.
(0.4 mL) was added. The enolate so-
lution was stirred at 08C for 15 min, cooled to À788C, and charged with
with two equivalents of LDA and one equivalent of trime-
thyltin chloride in the presence of chiral ligand 4 using
DME as the solvent. Quaternary centers are being formed
with a high degree of absolute stereochemical control. The
allylated products available by this method are quite versa-
tile for further elaboration.
a prestirred solution (rt, 5 min) of allyl acetate (2) (35 mL, 0.327 mmol),
palladium dimer
3 (2.7 mg, 0.00742 mmol), and ligand 4 (10.2 mg,
0.0149 mmol) in DME (0.6 mL). The cooling bath was removed and the
reaction was allowed to stir at room temperature for 30 min. The product
was purified directly by flash chromatography on silica gel eluting with
5% ethyl acetate/petroleum ether to give tetralone 5 (59 mg, 99%).
2-Allyl-2-methyl-1-tetralone (5): Prepared using (S,S)-ligand 4; Rf =0.53
(10% ethyl acetate/petroleum ether); determination of enantiomeric
excess: HPLC (Chiralcel OD column, 99.9:0.1 heptane/isopropanol,
flow=0.70 mLminÀ1), tR (major) enantiomer=19.14 min, tR (minor)
The importance of the cations associated with the enolate
is illustrated by their effect on both reactivity and enantiose-
lectivity. This strongly suggests that the actual structure of
the nucleophile is not simple and is likely an aggregate. In
employing lithium amide bases such as LDA and LHMDS,
there is the possibility of forming LDA-lithium enolate
mixed aggregates.[27] As the optimum conditions for AAA
call for two equivalents of LDA, one should be especially
cognizant of this here. Such aggregates have been crystal-
lized and can themselves form dimers and trimers.[28] Fur-
thermore, the addition of trimethyltin chloride and the pro-
duction of lithium salts during the reaction of a lithium eno-
late with an allylic carbonate may cause the nucleophile to
possess a variety of structural states during the course of the
reaction.
enantiomer=17.95 min; [a]D =+13.98 (c
= 1.18, 23.88C, dichlorome-
thane, 88% ee); IR (thin film): n˜ = 3074, 2930, 1682, 1602, 1455, 1221,
1
916, 741 cmÀ1; H NMR (CDCl3, 300 MHz): d = 8.02 (d, J=6.9 Hz, 1H),
7.44 (dt, J=7.5, 1.2 Hz, 1H), 7.31–7.19 (m, 2H), 5.81–5.70 (m, 1H), 5.08–
5.03 (m, 2H), 2.96 (t, J=6.0 Hz, 2H), 2.45 (dd, J=13.8, 7.2 Hz, 1H), 2.26
(dd, J=13.8, 7.5 Hz, 1H), 2.15–2.02 (m, 1H), 1.97–1.84 (m, 1H), 1.17 (s,
3H); 13C NMR (CDCl3, 75 MHz): d = 202.3, 143.4, 134.1, 133.2, 128.8,
128.1, 126.7, 118.3, 44.6, 41.1, 33.3, 25.3, 21.8; HRMS: m/z: calcd for
C14H16O: 200.1202; found: 200.1197.
Palladium-catalyzed AAA—General procedure B: nBuLi (2 equiv) was
added at À788C to
a solution of freshly distilled diisopropylamine
(2 equiv) in DME. After stirring at À788C for 15 min, a solution of the
nucleophile (1 equiv) in DME was added. After stirring at 08C for
15 min, the enolate solution was charged with tBuOH (7 equiv) followed
by a solution of trimethyltin chloride (1 equiv) in DME. The enolate so-
lution was stirred at 08C for 5 min, cooled to À788C, and charged with a
prestirred solution (rt, 5 min) of electrophile (1.05–1.1 equiv), palladium
dimer 3 (2.5%), and ligand 4 (5%) in DME. After the final addition, the
concentration of the reaction in nucleophile was 0.18m. The cooling bath
was removed and the reaction was allowed to stir at room temperature
until the reaction was complete. The product was purified directly by
flash chromatography on silica gel.
The compatibility of the types of ligands employed, which
contain secondary amides, to such strong bases raises ques-
tions about whether the amides are deprotonated under the
reaction conditions. The success of less stabilized nucleo-
philes such as simple enolates provides impetus for explor-
ing a much broader range of nucleophiles.
A typical example is given as follows: nBuLi (200 mL, 0.312 mmol) was
added at À788C to
(44 mL, 0.312 mmol) in DME (0.2 mL). After stirring at À788C for
15 min, solution of 2-tert-butoxymethylene-5-methylcyclopentanone
a solution of freshly distilled diisopropylamine
Experimental Section
a
(28 mg, 0.156 mmol) in DME (0.2 mL) was added. After stirring at 08C
for 15 min, tBuOH (100 mL, 1.05 mmol) was added followed by a solution
of trimethyltin chloride (31 mg, 0.156 mmol) in DME (0.2 mL) was
added. The enolate solution was stirred at 08C for 5 min, cooled to
À788C, and charged with a prestirred solution (rt, 5 min) of allyl acetate
(2) (19 mL, 0.172 mmol), palladium dimer 3 (1.4 mg, 0.0039 mmol), and
ligand 4 (5.4 mg, 0.0078 mmol) in DME (0.3 mL). The cooling bath was
removed and the reaction was allowed to stir at room temperature for
1 h. The product was purified directly by flash chromatography on silica
gel eluting with 10% ethyl acetate/petroleum ether to give ether 30
(30.2 mg, 87%).
All palladium reactions were performed under an atmosphere of dry ni-
trogen in flame-dried glassware. Solvents were distilled under an atmos-
phere of nitrogen before use and transferred via an oven-dried syringe.
nBuLi was titrated prior to the preparation of LDA.[29]
Palladium-catalyzed AAA—General procedure A: nBuLi (2 equiv) was
added at À788C to
a solution of freshly distilled diisopropylamine
(2 equiv) in DME. After stirring at À788C for 15 min, a solution of the
nucleophile (1 equiv) in DME was added. After stirring at 08C for
15 min, a solution of trimethyltin chloride (1 equiv) in DME was added.
The enolate solution was stirred at 08C for 15 min, cooled to À788C, and
charged with a prestirred solution (rt, 5 min) of electrophile (1.05–
1.1 equiv), palladium dimer 3 (2.5%), and ligand 4 (5%) in DME. After
the final addition, the concentration of the reaction in nucleophile was
2-tert-Butoxymethylene-5-methyl-5-(2-propenyl)cyclopentanone
(30):
Prepared using (S,S)-ligand 4. Rf =0.48 (20% ethyl acetate/petroleum
Chem. Eur. J. 2005, 11, 174 – 184
www.chemeurj.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
183

B. M. Trost and G. M. Schroeder
ether). Determination of enantiomeric excess: chiral GC (cyclosil B, iso-
therm 1208C) tR(major)=74.825 min, tR(minor)=74.169 min; [a]D =
+38.38 (c = 1.37, 23.78C, dichloromethane, 96% ee); IR (thin film): n˜ =
[8] For reviews of the palladium-catalyzed asymmetric allylic alkylation
see: a) B. M. Trost, Chem. Pharm. Bull. 2002, 50, 1–14; b) B. M.
Trost, D. L. Van Vranken, Chem. Rev. 1996, 96, 395–422; c) B. M.
Trost, Acc. Chem. Res. 1996, 29, 355–364; d) A. Heumann, M. Re-
glier, Tetrahedron 1995, 51, 975–1015; e) T. Hayashi, in Catalytic
Asymmetric Synthesis (Ed.: I. Ojima), VCH, New York, 1993; f) M.
Sawamura, Y. Ito, Chem. Rev. 1992, 92, 857–871; g) J. C. Fiaud, in
Metal-Promoted Selectivity in Organic Synthesis (Eds.: M. Graziani,
A. J. Hubert, A. F. Noels), Kluwer, Dordrecht, 1991; h) G. Consiglio,
R. M. Waymouth, Chem. Rev. 1989, 89, 257–276.
2977, 2869, 1708, 1630, 1458, 1371, 1265, 1156, 980, 945 cmÀ1 1H NMR
;
(CDCl3, 500 MHz): d = 7.50 (t, J=2.5 Hz, 1H), 5.76–5.68 (m, 1H), 5.02
(d, J=14 Hz, 2H), 2.44–2.41 (m, 2H), 2.15–2.13 (m, 2H), 1.87–1.81 (m,
1H), 1.60–1.55 (m, 1H), 1.34 (s, 9H), 1.00 (s, 3H); 13C NMR (CDCl3,
125 MHz): d = 210.9, 149.0, 134.6, 117.6, 115.2, 79.8, 49.5, 41.3, 32.5,
28.3, 22.0, 21.3; elemental analysis calcd (%) for C14H22O2: C 75.63, H
9.97; found: C 75.84, H 10.18.
[9] An example using molybdenum catalysis was recently reported:
B. M. Trost, K. Dogra, J. Am. Chem. Soc. 2002, 124, 7256–7257.
[10] B. M. Trost, G. M. Schroeder, J. Am. Chem. Soc. 1999, 121, 6759–
6760.
[11] B. M. Trost, B. Breit, S. Pleukert, J. Zambrano, J. W. Ziller, Angew.
Chem. 1995, 107, 2577–2579; Angew. Chem. Int. Ed. Engl. 1995, 34,
2386–2388.
[12] The required nucleophiles were generally synthesis via their corre-
sponding hydrazones.F. Henin, A. MꢁBoungou-MꢁPassi, J. Muzart, J.
Pete, Tetrahedron 1994, 50, 2949–2954; precursors to tetralones 19
and 20 were synthesized by direct alklylation as described in the lit-
erature: K. J. Shea, G. J. Stoddard, W. P. England, C. D. Haffner, J.
Am. Chem. Soc. 1992, 114, 2635–2643; C. Harde, F. Bohlmann, Tet-
rahedron 1988, 44, 81–90.
[13] D. L. Boger, R. J. Mathvink, J. Org. Chem. 1992, 57, 1429–1433.
[14] The starting material for furan analogue 23 was synthesized from
1,3-cyclohexanedione: R. A. Zambias, C. G. Caldwell, I. E. Kopka,
M. L. Hammond, J. Org. Chem. 1988, 53, 4135–4137.
[15] A. T. Rowland, M. W. Brechbiel, A. S. Gerelus, J. Chem. Educ. 1985,
62, 908–910.
[16] A. M. Islam, M. T. Zemaity, J. Am. Chem. Soc. 1957, 79, 6023–6024.
[17] a) R. K. Dieter, Y. J. Lin, J. W. Dieter, J. Org. Chem. 1984, 49, 3183–
3195; b) R. K. Dieter, J. Org. Chem. 1981, 46, 5031–5035.
[18] R. Ann Cornubert, Chim. (Paris) 1921, 16, 141–220.
[19] P. F. Schuda, C. B. Ebner, T. M. Morgan, Tetrahedron Lett. 1986, 27,
2567–2570.
[20] See ref. [16] and B. M. Trost, L. J. Stanton, J. Am. Chem. Soc. 1975,
97, 4018–4025.
[21] V. K. Aggarwal, A. Thomas, S. Schade, Tetrahedron 1997, 53,
16213–16228.
Acknowledgement
We thank the National Institutes of Health (GM-13598) and the National
Science Foundation for their generous support of our programs. G.M.S.
was a NSF and ARCS Graduate Fellow. Mass spectra were provided by
the Mass Spectrometry Regional Center of the University of California—
San Francisco supported by the NIH Division of Research Resources.
[1] a) F. A. Carey, R. J. Sundberg, Advanced Organic Chemistry Part B:
Reactions and Organic Synthesis, Plenum, New York, 1990; b) D.
Caine, Comprehensive Organic Synthesis, Vol. 3 (Eds.: B. M. Trost, I.
Fleming), Pergamon, Oxford, 1991, Chapter 1.1; c) H. B. Mekelbur-
ger, C. S. Wilcox, Comprehensive Organic Synthesis, Vol. 2 (Eds.:
B. M.Trost, I. Fleming), Pergamon, Oxford, 1991, Chapter 1.4.
[2] For some recent reviews see: a) J. Christoffers, A. Mann, Angew.
Chem. 2001, 113, 4725–4732; Angew. Chem. Int. Ed. 2001, 40, 4591–
4597; b) E. J. Corey, A. Guzman-Perez, Angew. Chem. 1998, 110,
402–415; Angew. Chem. Int. Ed. 1998, 37, 388–401; c) R. Noyori,
Asymmetric Catalysts in Organic Synthesis, Wiley, New York, 1994;
d) H. Fuji, Chem. Rev. 1993, 93, 2037–2066.
[3] J. Seyden-Penne, Chiral Auxiliaries and Ligands in Asymmetric Syn-
thesis, Wiley, New York, 1995.
[4] Catalytic Asymmetric Synthesis (Ed.:I. Ojima), 2nd ed., Wiley-VCH,
Weinheim, 2000.
[5] For discussions of atom economy see: a) B. M. Trost, Acc. Chem.
Res. 2002, 35, 695–705; b) B. M. Trost, Angew. Chem. 1995, 107,
285–307; Angew. Chem. Int. Ed. Engl. 1995, 34, 259–281; c) B. M.
Trost, Science 1991, 254, 1471–1477.
[6] J. C. Fiaud, A. Hibon deGournay, M. Larcheveque, H. B. Kagan, J.
Organomet. Chem. 1978, 154, 175–185.
[7] a) T. Hayashi, K. Kanehiri, R. Hagihara, M. Kumada, J. Org. Chem.
1988, 53, 113–120; b) M. Sawamura, Y. Nakayama, W.-M. Tang, Y.
Ito, J. Org. Chem. 1996, 61, 9090–9096; c) M. Sawamura, M. Sudoh,
Y. Ito, J. Am. Chem. Soc. 1996, 118, 3309–3310; d) B. M. Trost, R.
Radinov, E. M. Grenzer, J. Am. Chem. Soc. 1997, 119, 7879–7880;
e) B. M. Trost, X. Ariza, Angew. Chem. 1997, 109, 2749–2751;
Angew. Chem. Int. Ed. Engl. 1997, 36, 2635–2637; f) H. Brunner, I.
Deml, W. Dirnberger, B. Nuber, W. Reisser, Eur. J. Inorg. Chem.
1998, 43–54; g) R. Kuwano, Y. Ito, J. Am. Chem. Soc. 1999, 121,
3236–3237; h) S.-L. You, X.-L. Hou, L.-X. Dai, B.-X. Cao, J. Sun,
Chem. Commun. 2000, 1933–1934; i) X.-C. He, B. Wang, G. Yu, D.
Bai, Tetrahedron: Asymmetry 2001, 12, 3213–3216.
[22] J. A. Soderquist, A. Negron, Org. Synth. 1992, 70, 169–176.
[23] B. M. Trost, J. L. Belletire, S. Godleski, P. G. McDougal, J. M. Balko-
vec, J. Org. Chem. 1986, 51, 2370–2374.
[24] K. Jones, M. F. Lappert, J. Chem. Soc. 1965, 1944–1951.
[25] L. M. Jackman, J. Bortiatynski, Adv. Carbanion Chem. 1992, 1, 45–
87.
[26] D. Z. Wang, Y.-J. Kim, A. Streitwieser, J. Am. Chem. Soc. 2000, 122,
10754–10760.
[27] X. Sun, D. B. Collum, J. Am. Chem. Soc. 2000, 122, 2459–2463.
[28] L. M. Jackman, E. F. Rakiewicz, J. Am. Chem. Soc. 1991, 113, 1202–
1210.
[29] The use of commercially available LDA is not recommended.
Received: June 30, 2004
Published online: October 28, 2004
184
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 174 – 184