Mechanism and scope of salen bifunctional catalysts in asymmetric aldehyde and α-ketoester alkylation

Article abstract of DOI:10.1016/j.tet.2005.03.117

Metal complexes of C₂-symmetric Lewis acid/Lewis base salen ligands provide bifunctional activation resulting in rapid rates in the enantioselective addition of diethylzinc to aldehydes (up to 92% ee). Further experiments probed the reactivity of the individual Lewis acid and Lewis base components of the catalyst and established that both moieties are essential for asymmetric catalysis. These catalysts are also effective in the asymmetric addition of diethylzinc to α-ketoesters. This finding is significant because α-ketoesters alone serve as their own ligands to accelerate racemic 1,2-carbonyl addition of Et₂Zn and racemic carbonyl reduction. The latter proceeds via a metalloene pathway, and often accounts for the predominant product. Singular Lewis acid catalysts do not accelerate enantioselective 1,2-addition over these two competing paths. The bifunctional amino salen catalysts, however, rapidly provide enantioenriched 1,2-addition products in excellent yield, complete chemoselectivity, and good enantioselectivity (up to 88% ee). A library of the bifunctional amino salens was synthesized and evaluated in this reaction. The utility of the α-ketoester method has been demonstrated in the synthesis of an opiate antagonist.

Full text of DOI:10.1016/j.tet.2005.03.117

Tetrahedron 61 (2005) 6249–6265
Mechanism and scope of salen bifunctional catalysts in
asymmetric aldehyde and a-ketoester alkylation
Michael W. Fennie, Erin F. DiMauro, Erin M. O’Brien, Venkatachalam Annamalai
*
and Marisa C. Kozlowski
Department of Chemistry, Roy and Diana Vagelos Laboratories, University of Pennsylvania, Philadelphia, PA 19104-6323, USA
Received 8 July 2004; revised 28 February 2005; accepted 4 March 2005
Available online 10 May 2005
Abstract—Metal complexes of C2-symmetric Lewis acid/Lewis base salen ligands provide bifunctional activation resulting in rapid rates in
the enantioselective addition of diethylzinc to aldehydes (up to 92% ee). Further experiments probed the reactivity of the individual Lewis
acid and Lewis base components of the catalyst and established that both moieties are essential for asymmetric catalysis. These catalysts are
also effective in the asymmetric addition of diethylzinc to a-ketoesters. This finding is significant because a-ketoesters alone serve as their
own ligands to accelerate racemic 1,2-carbonyl addition of Et2Zn and racemic carbonyl reduction. The latter proceeds via a metalloene
pathway, and often accounts for the predominant product. Singular Lewis acid catalysts do not accelerate enantioselective 1,2-addition over
these two competing paths. The bifunctional amino salen catalysts, however, rapidly provide enantioenriched 1,2-addition products in
excellent yield, complete chemoselectivity, and good enantioselectivity (up to 88% ee). A library of the bifunctional amino salens was
synthesized and evaluated in this reaction. The utility of the a-ketoester method has been demonstrated in the synthesis of an opiate
antagonist.
q 2005 Elsevier Ltd. All rights reserved.
1. Introduction
Catalysts 1 and 2 are examples of interdependent catalysts
since the Lewis acid is directly coordinated to the Lewis
base moiety. While the metal activates the electrophile as a
Lewis acid, the heteroatom-containing moiety activates the
nucleophile as a Lewis base. A bifunctional Lewis acid/base
catalyst with electronically decoupled sites, however, would
allow for separate activation of electrophilic and nucleo-
philic substrates. It would, therefore, be possible to tune
each activation site independently. In particular, new
catalysts can be envisioned with functional groups specifi-
cally tailored to activate the nucleophile and electrophile of
a given reaction with the optimal spacing and orientation
between groups. Such systems may provide better
selectivity and turnover for reactions that have not been
successful with either monofunctional or interdependent
bifunctional catalysts.
Recent studies have shown that bifunctional ligands possess
useful characteristics for simultaneous substrate activation.1
For example, ligands have been reported in which one
portion engages a Lewis acid moiety that coordinates an
electrophilic substrate while a separate basic portion
coordinates to the nucleophilic substrate. The positioning
of the two reactants in close proximity and with the correct
relative geometry facilitates a reaction in a manner similar
to that of some enzyme catalysts.2 Seminal Lewis acid/
Lewis base catalysts are the Corey oxazaborolidine 1 for
ketone reduction3 and the Noyori b-amino alcohol catalyst 2
for the addition of diethylzinc to aldehydes4 (Fig. 1).
The salen motif is one of several privileged ligands5,6 for
asymmetric catalysis that impart high levels of enantio-
selectivity in many diverse transformations.7 In addition,
salen–metal complexes have been demonstrated to function
as chiral Lewis acid catalysts.8 Salen ligands are readily
available from inexpensive salicylaldehyde or phenol
precursors and are typically air-stable crystalline solids.
We therefore directed efforts toward the construction of
catalyst systems with general structure 3 (Fig. 2), which
Figure 1. Interdependent Lewis acid/Lewis base bifunctional catalysts.
Keywords: Bifunctional catalysis; a-Ketoester; Asymmetric alkylation.
*
Corresponding author. Tel.: C1 215 898 3048; fax: C1 215 573 7165;
e-mail: marisa@sas.upenn.edu
0040–4020/$ - see front matter q 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2005.03.117

6250
M. W. Fennie et al. / Tetrahedron 61 (2005) 6249–6265
Figure 2. Bifunctional salen scaffold.
incorporates a basic moiety into a structurally well-defined
and rigid salen complex.9 Bifunctional salen complexes 3
would provide an accessible Lewis acid center for
electrophile activation and a basic functional group for
activation of the nucleophile. A survey of the chemical
literature revealed few reports of chiral salens with
secondary functional groups.10,11
Scheme 1. Synthesis of quinoline and pyridyl salens.
Two expedient routes to chiral methylene-amine salens
were devised. Method A, outlined in Scheme 2, allows for
the rapid construction of amino salens in three simple steps
from phenol 7. A Mannich reaction of 8 with a secondary
amine provides aldehyde 12, which is then condensed with
(R,R)-cyclohexanediamine to give methylene-tethered salen
13. Although this route provides access to a number of
amino salicylaldehydes, it is not without limitations. The
aromatic Mannich reaction13 to install the methylene-amine
group is only effective with simple, unbranched secondary
amines. It was, therefore, necessary to explore alternative
synthetic routes to a-branched amino salicylaldehydes
(Scheme 3).
An important structural consideration in the design of these
bifunctional scaffolds was the nature of the tether connect-
ing the salen to the basic functional group. This tether must
be sufficiently short and/or rigid to prevent internal
complexation of the Lewis base to the Lewis acid.
Compounds 4 and 5 are inherently rigid by virtue of the
biaryl bonds, whereas the short tether in compound 6
prevents self-complexation (Fig. 3). In spite of the
simplified structure and ease of preparation of such ligands,
chiral variants had not been explored as ligands for
asymmetric catalysis previously.
Figure 3. Bifunctional salen ligands.
In previous work, we showed that these amino-salen
bifunctional catalysts catalyzed the rapid addition of
diethylzinc to aldehydes.12a Having demonstrated the
premise behind these catalysts in this well-understood,
privileged reaction, they were also examined in the addition
of diethylzinc to a-ketoesters, an unexplored reaction with a
fast background rate and multiple reaction products.12b,c In
this report, we describe the full structural and mechanistic
details of these catalysts and their scope in these reactions.
2. Results and discussion
2.1. Ligand synthesis
Scheme 2. Synthesis of methylene tethered salens (Method A).
This alternate route began with acetonide 15 which was
prepared in three steps from tert-butyl phenol 7 (Scheme 3).
Treatment of this acetonide with anhydrous HBr, generated
in situ, effected deprotection along with formation of benzyl
bromide 16. Direct addition of secondary amines afforded
a-branched amino salicylaldehydes 13m–r. The described
methods provide a highly modular synthesis and easy access
to numerous derivatives.
Formylation of phenol 7, followed by bromination of 8,
Stille coupling between 9 and 10, and subsequent
condensation with (R,R)-cyclohexanediamine provided
quinoline salen 4 in four steps (Scheme 1). In a similar
fashion, pyridinyl salen 5 was obtained using stannane 11
(Scheme 1).

M. W. Fennie et al. / Tetrahedron 61 (2005) 6249–6265
6251
have been reported to catalyze the addition of diethylzinc to
aldehydes with excellent enantioselectivity, these catalysts
are generally less reactive than the titanium-based
systems.14a
When a b-aminoalkoxy zinc catalyst is employed, the
reaction proceeds by interdependent dual activation of the
aldehyde electrophile and the diethylzinc nucleophile
(Fig. 4, 17).16,17 With an independent bifunctional salen
framework, an apical coordination site on the Lewis acidic
metal center could activate the aldehyde18 while the
tethered base could independently activate the diethylzinc
nucleophile (Fig. 4, 18).19
The zinc complex of quinoline salen 4 proved to be very
reactive in the aldehyde alkylation (Eq. 1, Table 1, entry 1)
as 93% conversion was observed after 2 h at room
temperature. In pyridyl salen 5, the Lewis basic nitrogen
is disposed closer to the zinc center. This arrangement
allows for fast catalysis even at lower temperatures along
with substantial enantiocontrol (Table 1, entry 2).
Scheme 3. Synthesis of amino salens (Method B). aSalen derived from
(1S,2S)-cyclohexanediamine.
(1)
Pyridine salen derivatives, such as a para-dimethylamino
pyridine salen, were pursued, but synthesis of such
compounds was non-trivial. At this juncture pyridine salens
were abandoned in order to pursue a class of amino-salen
ligands that could be prepared and tested in a more rapid and
efficient manner. Morpholine salen 13b was selected
(Scheme 2), on the basis of its highly crystalline nature, as
Figure 4. Transition state models for diethylzinc addition.
Table 1. Biaryl salen ligands in the addition of Et2Zn to PhCHO (Eq. 1)a
Entry
Catalyst
Solvent
T (8C)
Time (h)
Conv. (%)
ee (%)b
1
2
Zn-4
Zn-5
Toluene
Toluene
25
K50
2
20
93
97
10 (R)
78 (S)
a Addition of 2.0 equiv of Et2Zn with 10 mol% catalyst.
b Determined by chiral GC (Cyclodexb). Absolute configuration determined by comparison to literature.
2.2. The addition of diethylzinc to aldehydes
the initial probe for the addition of diethylzinc to
benzaldehyde. The screening process began with an
evaluation of several Lewis acidic metal complexes of
13b (Table 2, Eq. 1). While the Ni, Cu, Mg and Ti(Oi-Pr)2
complexes20 with 13b were effective catalysts, the Zn
complex proved the most versatile giving high selectivities
at lower temperatures (Table 2, entries 5–8).
The addition of diethylzinc to aldehydes is a well-studied
reaction and is used as a benchmark to measure the efficacy
of new ligands. Titanium complexes of chiral diols and
bis(sulfonamide) ligands have proven the most reactive and
enantioselective to date.14 One drawback of these systems,
however, is the requirement for a stoichiometric amount of
Ti(Oi-Pr)4.15 Although various chiral aminoalcohol ligands
The effect of the chiral diamine backbone on the geometry
and reactivity of the Zn-morpholine salen complexes was
briefly examined. A significant loss of enantioselectivity
was observed with the zinc complexes derived from (R)-
BINAM (21) and (S,S)-1,2-diphenylethylenediamine (22)
(Fig. 5, Table 2, entries 9–10).
In order to survey the role of the Lewis base component,
zinc complexes of select amino salens and methyl ether
salen 23 were examined (Table 3, Eq. 1). Reactions were
conducted at K30 8C and quenched after 6 h for a direct
comparison of catalyst reactivity.
Figure 5. Salens comprised of other diamine backbones were inferior.

6252
M. W. Fennie et al. / Tetrahedron 61 (2005) 6249–6265
Table 2. Lewis acid complexes of morpholine salens in the addition of Et2Zn to PhCHO (Eq. 1)a
Entry
Catalyst
T (8C)
Time (h)
Conv. (%)b
ee (%)b
1
2
3
4
5
6
7
8
9
10
Ni-13b
25
25
25
25
2
2
2
2
2
12
15
24
2
99
100
99
100
100
98
98
46
95
84
32 (R)
10 (S)
3 (S)
55 (S)
54 (S)
80 (S)
87 (S)
83 (S)
19 (S)
23 (S)
Cu-13b
Mg-13b
Ti(Oi-Pr)2-13b
Zn-13b
Zn-13b
Zn-13b
Zn-13b
Zn-21
25
K30
K35
K70
25
Zn-22
25
2
a Reactions performed in toluene with 2.0 equiv Et2Zn and 10 mol% catalyst.
b Determined by GC (Cyclodexb). Absolute configuration assigned by comparison to the literature.
While the highest enantioselectivity was observed with
morpholine salen 13b (87% ee at K35 8C), an even more
active catalyst was observed with the slightly more basic
piperidyl salen 13c, resulting in 90% conversion after 6 h at
K30 8C (Table 3, entry 3). Surprisingly the less basic
pyridyl derivative 5 provided the most reactive catalyst
(Table 3, entries 5–6). Catalysts derived from salens 13m–n,
containing a-branched amines, were less reactive (Table 3,
entries 7–10). With the methyl ether derivative 23, a much
less reactive complex was formed which was not catalytic at
K30 8C (Table 3, entry 11). Indeed, significant conversion
was only observed at higher temperatures (i.e., room
temperature Table 3, entry 12).
A substantial loss of catalytic activity was observed with
salens lacking secondary Lewis bases. For example, the
catalyst derived from tert-butyl salen 24 provided only 2%
conversion after 6 h at K30 8C (Table 3, entry 13) along
with significant amounts of the reduced benzyl alcohol
byproduct (15%). Zn-24 requires 18 h at room temperature
to achieve high conversion (90%). In order to rule out any
steric effect from the larger tert-butyl groups in 24, the zinc
Table 3. Zn-salen complexes in the addition of Et2Zn to PhCHO (Eq. 1)a
Entry
Salen
R
T (8C)
Time (h)
Conv. (%)b,c
ee (%)b,d
wpKa (BHC e
)
1
2
13b
13b
K30
K35
6
15
72
98
77 (S)
87 (S)
6
6
3
4
13c
13c
K30
K40
6
15
90
82
76 (S)
85 (S)
9
9
5
6
5
5
K30
K50
6
20
99
97
51 (S)
78 (S)
5
5
7
8
13m
13m
K30
25
6
6
5 (11)
90 (5)
52 (S)
48 (S)
9
9
9
10
13n
13n
K30
K30
6
30
33
85
73 (S)
82 (S)
9
9
11
12
23
23
K30
25
6
3
8 (3)
95
18 (R)
18 (R)
K3
K3
13
14
24
25
K30
K30
6
6
2 (15)
3 (7)
57 (S)
21 (S)
NA
NA
a Reactions conducted in toluene with 10 mol% salen and 2.1 equiv Et2Zn.
b Determined by chiral GC (Cyclodex-b).
c Values in parentheses refer to amount of benzyl alcohol side product.
d Absolute configuration assigned by comparison to the literature.
e Estimated from the pKa(H2O) values of N-methylmorpholine, N-methylpiperidine, benzylamine, methylamine, pyridine and dimethyl ether.

M. W. Fennie et al. / Tetrahedron 61 (2005) 6249–6265
6253
complex of cyclohexane salen 25, which is an isosteric
analog of 13b–c, was examined. As expected, nocatalysis was
observed after 6 h at K30 8C with Zn-25 (Table 3, entry 14).
2.3. Reaction mechanism
From the results collected in Table 3, a good Lewis base is
required for an active catalyst. Furthermore, the electron
pair donor/zinc-binding ability of this Lewis base can be
correlated to catalyst reactivity. For the methylene-tethered
Lewis base salens, a similar steric environment is expected
in the vicinity of the sp3 base. As such, the observed
correlation between the conjugate acid (BHC) pKa values
and reactivity of Zn-13b, Zn-13c, and Zn-23 most likely
reflects a difference in reactivity due to the inherent donor
ability.21 On the other hand, the most reactive pyridyl
derivative Zn-5 indicates that the less hindered pyridine
with softer Lewis base character possesses superior donor/
zinc-binding ability. The results with a-branched amine
derivatives Zn-13m and Zn-13n, indicate that zinc-binding
ability decreases with increasing steric hindrance.22
Figure 7. Non-Arrhenius behavior of catalyst Zn-13b.
failed to provide catalysis. The results of this study support a
bifunctional activation model in which the Lewis acid and
Lewis base portions in the same structure work together to
promote a rapid asymmetric reaction.
Reaction rates for the reaction in Eq. 1 (1 equiv PhCHOC
10 equiv Et2Zn) at 25 8C were also measured using a
REACT-IR spectrometer. Rates were obtained at
4–20 mol% catalyst by monitoring the rate of disappearance
of the aldehyde carbonyl stretching frequency over time.
Using this data, the reaction was found to be first order in
catalyst and first order in benzaldehyde, further supporting
the proposed model 18 (Fig. 4).
The above data support a direct role for the Lewis base in
the course of the reaction. In addition, the enantiomeric
excess of the reaction in Eq. 1 with the Zn-13b did not vary
significantly as a function of conversion (Fig. 6) indicating
that the composition of the catalytic species does not change
over the course of the reaction. These results are in accord
with our proposed transition state model (Fig. 4, 18).
2.4. Substrate scope
To demonstrate that the new catalyst system can accept
different substrates, a number of other aldehydes were
examined in the diethylzinc addition with Zn-13b (Eq. 2,
Table 4). Similar levels of conversion and enantiomeric
excess were obtained for aromatic aldehydes as well as
aliphatic aldehydes.
(2)
Table 4. Zn-13b in the addition of Et2Zn to RCHO (Eq. 2)a
Figure 6. Conversion versus ee for Et2Zn C PhCHO (Eq. 1) with Zn-13b.
Entry
R
Time (h)
Conv. (%)b ee (%)b
1
2
3
4
5
6
Ph
12
36
10
36
24
36
98
97
92
99
78
84
87 (S)
89 (S)
83 (S)
69 (S)
75 (S)
91 (S)
To provide further evidence for this model, the reaction in
Eq. 1 was examined for a nonlinear effect using the zinc
complex derived from 13b. A linear correlation was found
between the product enantiomeric excess and the enantio-
meric excess of 13b (see Supporting information) consistent
with a monomeric catalyst model 18.
p-MeO-C6H4
p-Cl-C6H4
o-Me-C6H4
Cy
(CH3CH2)2CH
a Reactions were in PhCH3 using 10 mol% ligand and 2.1 equiv Et2Zn.
b Determined by chiral GC (Cyclodexb). Absolute configuration assigned
by comparison to the literature for entry 1. For all others, absolute
configuration assigned by analogy.
Interestingly, at temperatures above K20 8C the catalyst
begins to deviate from Arrhenius behavior (Fig. 7),
indicating the concurrent operation of different catalytic
mechanisms.
Metal complexes of morpholine salen 13b and piperidine
salen 13c were also examined in the addition of other
dialkylzinc nucleophiles to benzaldehyde (Eq. 3, Table 5).
At K10 8C the addition of dimethylzinc to benzaldehyde
proceeds to completion with modest enantioselectivity
(Table 5, entry 2). The addition of phenylacetylide to
benzaldehyde is very rapid at 0 8C with the zinc and
titanium complexes of 13b–c (Table 5, entries 5–6). When
A cooperative effects study was performed, using tert-butyl
salen 24 in combination with N-methylmorpholine as an
external Lewis base additive.23 Neither Zn-24 or N-methyl-
morpholine alone catalyzed the reaction. A combination of
the Zn-24 salen (10%) and N-methylmorpholine (20%) also

6254
M. W. Fennie et al. / Tetrahedron 61 (2005) 6249–6265
Table 5. Addition of RZnMe to PhCHO (Eq. 3)a
Entry
R
Catalyst
T (8C)
Time (h)
Conv. (%)b,c
ee (%)b
1
2
3
4
5
6
7
Me
Me
Me
CCPh
CCPh
CCPh
CCPh
Zn-13b
Zn-13b
Zn-13b
Zn-13b
Zn-13b
Ti(Oi-Pr)2-13c
Ti(Oi-Pr)2-13c
25
K10
K30
K30
0
2.5
10
20
12
1
99
99
42
95 (88)
99 (96)
99
44
50
50
66 (S)
63 (S)
62 (S)
73 (S)
0
K20
1
20
100
a Reactions were in PhCH3 using 10 mol% ligand and 2.1 equiv MeZnR.
b Determined by 1H NMR and chiral GC (Cyclodexb). Absolute configuration determined by comparison to the literature.
c Isolated yields in parentheses.
the temperature is lowered to K20 8C, the catalytic activity
of Ti(Oi-Pr)2-13c is maintained and the enantioselectivity
improves to 73% ee (Table 5, entry 7).
reactivity profile intermediate to that of aldehydes and
ketones, we soon discovered that a-ketoesters react
spontaneously and often uncontrollably with Grignard
reagents, dialkylzinc reagents and organozinc halides.
Furthermore, suppression of reactivity is minimal at low
temperatures. Noyori and coworkers described similar
enhanced reactivity with a-functionalized aldehydes and
pyruvates, hypothesizing that bidentate coordination to the
nucleophile facilitates the uncatalyzed achiral pathway.33
(3)
2.5. Addition of diethylzinc to a-ketoesters
The a-hydroxyester product of the addition of an organo-
metallic to an a-ketoester contains a chiral quaternary
center24 and is a versatile synthetic precursor.25 While a
handful of methods are available for the catalytic asym-
metric addition of soft nucleophiles to a-ketoesters,26
relatively few methods exist for the enantioselective
addition of hard nucleophiles.
The development of a selective a-ketoester alkylation is
further complicated due to competing reaction pathways.
Two main products (addition and reduction) were
encountered in the addition of ethylmagnesium bromide to
a-ketoesters at K10 8C (Eq. 5, Table 6). We speculate that
the reduction product arises via a metalloene pathway as
illustrated in Scheme 4. Reduction is therefore possible with
any organometallic that contains a b-hydrogen.34
(4)
Diastereoselective additions of organometallics to chiral
a-ketoesters have been studied extensively and many
successful reports have been published.27 Examples of the
corresponding enantioselective addition are rare and most
require the use of stoichiometric or excess chiral ligand.28,29
For example, Seebach and Weber29a reported the enantio-
selective (68% ee) addition of octyl Grignard to methyl
pyruvate in low yield (10%) in the presence of 1.1 equiv of
TADDOL. Using excess (1.25–1.9 equiv) sparteine as a
chiral mediator, Noyori et al.29b added EtMgBr and n-BuLi
to ethylbenzoylformate with low selectivity (18%). In a
third report, the asymmetric addition of a butyl group to
methyl benzoylformate was accomplished in up to 43% ee,
using a stoichiometric amount of a chiral lithium alkoxy-
tributylaluminate.29c Recently, Shibasaki and co-workers
have reported an enantioselective addition of dimethylzinc
to a-ketoesters (59–96% ee) using a catalytic hydroxy-
proline derivative, but this method requires extensive
reaction times (42 h) and does not tolerate more reactive
alkylzinc reagents such as diethylzinc.30
(5)
Table 6. Addition of EtMgBr to R1COCO2Et (Eq. 5)
Entry
R1
Addition conv.
(%)a
Reduction
conv. (%)a
1
2
3
4
5
6
7
8
9
p-MeO-C6H4
p-Br-C6H4
Ph
o-Me-C6H4
b-Naphthyl
Me
Cy
i-Pr
t-Bu
81
69
60
38
37
38
31
21
21
16
22
19
19
30
52
54
48
63
a Determined by 1H NMR spectroscopy
Ketones and aldehydes are suitable electrophiles for
Grignard reagents, however, they are generally inert toward
dialkylzinc reagents and organozinc halides in the absence
of catalyst. While many catalysts give rise to substantial
ligand accelerated catalysis in31 the asymmetric addition of
alkyl zinc reagents to aldehydes, the corresponding ketone
additions have been more difficult due to the greater steric
hindrance of ketone carbonyls.32 Assuming that the steric
and electronic properties of a-ketoesters would confer a
Scheme 4. Possible mechanism for reduction of a-ketoesters by EtMgBr.
Thus, several factors must be considered in developing
enantioselective catalysts for the reaction in Eq. 6. First, the
catalyst must accelerate asymmetric addition faster than the
uncatalyzed, racemic addition. Second, the catalyst must
accelerate addition to a greater degree than reduction. Prior

M. W. Fennie et al. / Tetrahedron 61 (2005) 6249–6265
Table 7. M-13c in the addition of Et2Zn to PhC(O)CO2Et (Eq. 6, R1ZPh, R2ZEt)a
6255
Entry
Catalyst
T (8C)
Time (h)
Redn (%)b
Addn (%)b
Addn ee (%)b
1
2
3
4
5
6
7
8c
None
None
Ni-13c
Cu-13c
Zn-13c
Zn-13c
Mg-13c
Ti(Oi-Pr)2-13c
0
K40
0
0
0
K30
K40
K30
24
2
24
24
24
2
86
45
62
12
6
48
0
8
11
23
35
75
93
36
99
68
-
-
14 (S)
18 (R)
20 (R)
0
34 (R)
52 (R)
2
48
a Reactions were in toluene with 10 mol% catalyst and 1.2 equiv Et2Zn.
b Determined by chiral GC (Cyclodexb column) and 1H NMR spectroscopy. Absolute configuration determined by comparison to the literature.
c 1.32 equiv Et2Zn.
to examining the salen catalysts in the additions of
dialkylzincs to a-ketoesters, two control reactions were
performed (Table 7, entries 1 and 2). Treatment of ethyl
oxo(phenyl)acetate with Et2Zn (Eq. 6, R1ZPh, R2ZEt)
revealed that reduction was the predominant pathway (86%
at 0 8C, 45% at K40 8C) with lesser amounts of 1,2-addition
(6)
As expected, the reaction is optimal in nonpolar, non-
coordinating solvents such as toluene, hexanes and pentane
(Table 8). Further studies were undertaken with toluene due
to incomplete catalyst solubility in hexanes or pentane.
(11% at 0 8C, 23% at K40 8C) occurring. From this result, it
was clear that the rate of the 1,2-addition pathway was slow
enough that the development of an asymmetric version by
means of ligand-accelerated catalysis would be possible.
The uncatalyzed reactions also revealed that suppression of
the reduction pathway would be a major challenge in the
development of this reaction. Recent related work of
Shibasaki avoids this problem by using dimethylzinc; the
lack of b-hydrogens precludes metalloene reduction.30
Table 8. Solvent screening with Ti(Oi-Pr)2-13c in the addition of Et2Zn to
PhC(O)CO2Et (Eq. 6, R1ZPh, R2ZEt)a
Entry
Solvent
Reduction
Conv. (%)b
Addition
Addition ee
Conv. (%)b (%)b
1
2
3
4
5
6
Toluene
Hexanes
Pentane
CH2Cl2
TBME
THF
8
0
0
9
4
7
68
88
88
77
66
12
52 (R)
62 (R)
50 (R)
47 (R)
30 (R)
17 (R)
Our salen catalysts were found to be much more reactive in
aldehyde alkylation compared to the Noyori type catalysts
(Fig. 4). Believing that this higher reactivity profile would
be useful, our investigation commenced with the piperidine
salen metal complexes M-13c. In the absence of a catalyst,
reduction is the principal pathway (86%) at 0 8C in the
reaction of Et2Zn with ethyl oxo(phenyl)acetate (Table 7,
entry 1). In contrast, Zn-13c at 0 8C, affords 93% addition
with very little reduction (Table 7, entry 5). This represents
a 120-fold change in selectivity with respect to the
uncatalyzed reaction, definitively indicating that the salen
catalysts accelerate addition to a far greater degree than
reduction. The Mg complex proved even more reactive,
providing complete conversion to the addition product
within 2 h at K40 8C (Table 7, entry 7). When the Ti(Oi-
Pr)2 complex was examined, promising enantioselectivity
(52%) was also observed at K30 8C (Table 7, entry 8).
Encouraged by the improved enantioselectivity, we under-
took a thorough examination of the reaction conditions with
Ti(Oi-Pr)2-13c.
a Reactions were at K30 8C with 10 mol% catalyst and 1.32 equiv Et2Zn.
b Determined by chiral GC (Cyclodexb) and 1H NMR spectroscopy.
Absolute configuration determined by comparison to the literature.
For the reactions in Tables 7 and 8 the Ti catalyst was
generated in situ by treatment of the salen with Ti(Oi-Pr)4.
The resultant solution was then treated with 1.32 equiv
Et2Zn followed by the substrate. As a result, the i-PrOH
released upon metal complexation was converted into
EtZn(Oi-Pr) which may act as a catalyst or an inhibitor.
To remove the EtZn(Oi-Pr) from our analysis, the Ti
complex was isolated by removal of all of the volatiles in
vacuo. Under these conditions, the catalyst is freely soluble
in PhCH3 and was found to be much more reactive (99%
conversion after 2 h vs 68% conversion after 48 h), more
chemoselective (0% vs 8% reduction), and more
Table 9. M(Oi-Pr)2-13c catalysts in the addition of Et2Zn to PhC(O)CO2Et (Eq. 6, R1ZPh, R2ZEt)a
Entry
Catalyst
T (8C)
Time (h)
Redn (%)b
Addn (%)b
Addn ee (%)b
1
2
Ti(Oi-Pr)4
K40
K40
K40
0
2
2
2
24
2
46
0
0
7
3
30
99
92
36
45
–
Ti(Oi-Pr)2-13c
Al(Oi-Pr)-13c
V(O)(Oi-Pr)-13c
Zr(Oi-Pr)2-13c
62 (R)
21 (R)
17 (R)
24 (S)
3c
4c
5
K40
a Reactions were in toluene with 10 mol% catalyst and 1.2 equiv Et2Zn.
1
b Determined by chiral GC (Cyclodexb) and H NMR spectroscopy. Absolute configuration determined by comparison to the literature.
c 1.32 equiv Et2Zn.

6256
M. W. Fennie et al. / Tetrahedron 61 (2005) 6249–6265
enantioselective (62% ee vs 52% ee). Even at low
temperature this catalyst retained high reactivity (2 h,
K40 8C, 99% conversion, 0% reduction, 62% ee). This
improved protocol was used to screen further M(Oi-Pr)n
complexes for catalytic activity (Table 9).
R1ZCO2R, BZN-piperidine). In line with this hypothesis,
the structurally similar complexes Ti(Oi-Pr)2-13a–d
exhibited comparable reactivity and selectivity (Table 10,
entries 1–4). More hindered amines are less effective for
coordination of metal species (i.e., Et2Zn) which is
consistent with the decreased reactivity of 2,6-dimethyl-
piperidinyl Ti(Oi-Pr)2-13n (Table 10, entry 7). If the amine
base is distant from the Ti center, bifunctional activation is
less likely. This conjecture is supported by the lesser
reactivity and selectivity of quinolinyl catalyst, Ti(Oi-Pr)2-4
(Table 10, entry 12) compared to pyridinyl catalyst, Ti(Oi-
Pr)2-5 (Table 10, entry 6). Replacement of the amine with
nonbasic groups should change the catalytic activity if this
Lewis base causes activation. The lesser reactivity of the
tert-butyl salen complex Ti(Oi-Pr)2-24 (Table 10, entry 15)
is consistent with this proposal. The use of chelating
diamines (Table 10, entry 5), chiral diamines (Table 10,
entries 8–11), or other activating groups (Table 10, entries
13–14) was not successful.
Ti(Oi-Pr)2-13c outperformed all other M(Oi-Pr)n com-
plexes. Encouraged by the remarkable ligand acceleration
(Table 9, entry 2 vs entry 1), we examined a number of other
Lewis base functionalized salen ligands from our library as
their titanium complexes (Table 10). This ligand screening
confirmed the identification of Ti(Oi-Pr)2-13c as an optimal
lead catalyst for further study and provided insight into the
mechanism of catalysis.
2.6. Reaction mechanism
We propose that the piperidine group plays a critical role in
this catalyst as a Lewis base activating group (Fig. 4, 18,
Table 10. Ti(Oi-Pr)2-salen in the addition of Et2Zn to PhC(O)CO2Et (Eq. 6, R1ZPh, R2ZEt)a
Entry
1
Salen
Lewis base
Redn (%)b
Addn (%)b
91
Addn ee (%)b
13a
0
44 (R)
2
3
4
5
13b
13c
13d
13e
9
0
72
99
94
71
54 (R)
62 (R)
54 (R)
0
3
12
6
7
5
0
91
84
57 (R)
20 (S)
13n
10
8
13o
13pc
13q
13rc
4
3
6
0
1
5
76
66
98
98
57
12 (R)
3 (S)
9
10
11
12
15 (R)
50 (S)
7 (S)
13
26
2
66
5 (S)
14
15
27
24
15
20
63
56
0
4 (S)
a Reactions were in toluene with 10 mol% catalyst and 1.2 equiv Et2Zn for 2 h.
b Determined by chiral GC (Cyclodexb) and 1H NMR spectroscopy. Absolute configuration determined by comparison to the literature.
c Salen derived from (1S,2S)-cyclohexanediamine.

M. W. Fennie et al. / Tetrahedron 61 (2005) 6249–6265
6257
Although the tert-butyl salen is not a useful catalyst for this
reaction alone, we investigated if asymmetric catalysis with
Ti(Oi-Pr)2-24 was possible under double activation con-
ditions. N-Methylmorpholine (NMM)23 was selected as a
suitable Lewis base additive to be used in combination with
this complex. The results of this experiment are collected in
Table 11. In the absence of catalyst, reduction is the major
reaction pathway (entry 1). The course of the reaction is not
influenced by the presence of 20 mol% Lewis basic
N-methylmorpholine (entry 2). Although some catalysis is
observed with Lewis acidic Ti(Oi-Pr)2-24 (Table 11, entry
3), the chemoselectivity and enantioselectivity are poor.
When 2 equiv of N-methylmorpholine are added for every
equiv of Ti(Oi-Pr)2-24 (entry 4), similar results are observed
as for Ti(Oi-Pr)2-24 alone. Excess N-methylmorpholine
relative to substrate with Ti(Oi-Pr)2-24 diminishes
reactivity and selectivity, showing that the amine is not
simply deaggregrating the Et2Zn (entry 5). The tethered
Lewis acid and Lewis base clearly act in a cooperative
manner as evidenced by the higher reactivity and enantio-
selectivity found in entry 6.
Figure 8. Proposed transition structure.
Table 11. Addition of Et2Zn to PhCOCO2Et (Eq. 6, R1ZPh, R2ZEt):
Cooperative effects studiesa
Entry
Catalyst
Redn.
(%)
Addn.
(%)b
Addn. ee
(%)b
1
2
3
4
5
6
None
NMM
Ti(Oi-Pr)2-24
Ti(Oi-Pr)2-24C NMMc
Ti(Oi-Pr)2-24C NMMd
Ti(Oi-Pr)2-13c
45
30
20
15
17
0
23
14
56
56
12
99
–
–
4 (S)
2 (S)
0
Figure 9. Linear relationship between catalyst ee and product ee (Et2ZnC
PhCOCO2Me using 10 mol% Ti(Oi-Pr)2-13c at K78 8C for 2 h).
62 (R)
Also investigated was the effect of chiral product on
enantioselectivity. When enantiopure 2-hydroxy-2-phenyl-
butyric acid ethyl ester was used as an apical ligand instead
of i-PrO in 28 no significant enantiomeric excess was
observed (7% ee (R)). This result suggests that the product,
which can coordinate to titanium as an apical ligand, does
not have a significant effect on the course of the reaction.
a Addition of 1.2 equiv Et2Zn with 10 mol% catalyst at K40 8C for 2 h.
b Determined by chiral GC (Cyclodexb).
c 2.0 equiv (relative to Ti(Oi-Pr)2-24) of N-methylmorpholine added.
d 12 equiv (relative to Ti(Oi-Pr)2-13c) of N-methylmorpholine added.
Despite their success in aldehyde alkylation, we found that
DAIB-type catalysts such as EtZn-MIB35 performed poorly
with a-ketoesters36 (Eq. 7). This result serves to demon-
strate the advantage of independent bifunctional activation.
In the presence of 10 mol% Ti(Oi-Pr)2-13c, the reaction of
methyl oxo(phenyl)acetate and diethylzinc is much faster
than we originally estimated as judged by React-IR
(reaction complete within 10 min at K40 8C). Although
good catalytic activity is maintained at temperatures as
low as K78 8C (reaction timeZ2 h), moderate enantio-
selectivity persists (74% ee). Unfortunately, this remarkable
ligand-accelerated catalysis complicates kinetic
measurements.
(7)
2.7. Diversity oriented synthesis of a ligand library
In an effort to increase enantioselectivity, we sought to
extend our library of methylene tethered amino-salen
ligands and their titanium complexes. Four segments of
the amino-salen titanium complexes (Ti(L)2-salen) can be
altered easily. Modification of the diamine linker changes
the stereochemical environment of the apical position or
even the general shape of the salen structure. For example,
BINAM allows the formation of salen metal complexes in
which the metal ligands are not necessarily coplanar.38 The
Lewis acidity of the salen metal center can be altered by
introduction of electron withdrawing or donating
The data support a mechanism that involves ionization of an
alkoxide group to provide a five-coordinate cationic Ti
species (Fig. 8).37 Unfortunately, attempts to promote the
proposed ionization of isopropoxide with the addition of
TMSOTf (5 mol% with 10 mol% Ti(Oi-Pr)2-13c) resulted
in loss of catalytic activity (46% conv. addition, 0% ee; 15%
conv. reduction after 2 h at K40 8C).
The lack of a nonlinear effect with enantio-impure 13c in
this reaction (Fig. 9) is also consistent with this proposed
mechanism.

6258
M. W. Fennie et al. / Tetrahedron 61 (2005) 6249–6265
Table 12. Addition of Et2Zn to PhCOCO2Me (Eq. 6, R1ZPh, R2ZMe): Screening the apical ligands (L) [Ti(L)2-salen, R2Zt-Bu, NR32Zpiperidine]a
Entry
Catalyst
L
Redn.
Conv.
(%)b
Addn.
Conv.
(%)b
Addn.
ee (%)b
1
Ti(Oi-Pr)2-13c
Ti(Ot-Bu)2-13c
TiCl2-13c
Oi-Pr
Ot-Bu
Cl
0
0
99
99
72 (R)
72 (R)
2c
3
0
98
70 (R)
4
28
9
32
63
0
0
5d
6
7
8
9
Ti(OAr)2-13c
Ti(OR*)2-13c
Ti(OR*)2-ent13c
Ti(OR*)2-13s
0
0
0
0
92
92
74
60
81 (R)
57 (R)
78 (S)
18 (S)
a Reactions were in PhCH3 at K40 8C for 2 h using 10 mol% catalyst.
b Determined by chiral GC (Cyclodex b). Absolute configuration determined by comparison to the literature.
c
˚
With 4 A MS.
d In CH2Cl2.
C5-substituents (R2) to the phenol ring. The basicity,
hard/soft character, and the immediate stereochemical
environment of the pendant amine base (NR32) are also
readily changed. Finally, the apical ligands (L) on the
titanium center can be directly modified to provide
complexes with different reactivity and substrate-binding
profiles.
availability or ease of preparation. The elements
examined provide a range of electronic and steric
properties. Similarly, the various C5-substituted salens
were easily prepared using the aromatic Mannich
reaction (Scheme 2 and Eq. 8). The aminosalicylalde-
hyde precursors are either commercially available or
prepared in one step.39 In our early examination of the
Lewis base moiety we discovered that the best catalysts
contain the cyclic amines piperidine, morpholine and
pyrrolidine (Table 10, entries 2–4). Since a-substituted
piperidine derivatives are not useful (Table 10, entry 7)
a number of commercially available substituted piper-
azines were investigated. Using a parallel synthesis
strategy and a systematic screening approach, 30 of the
720 possible catalysts were prepared and tested in a
rapid and efficient manner.
(8)
For each of the four variable segments, our selections
were based on several factors, including commercial
Beginning at the titanium center, various apical ligands
were examined (Table 12). If our proposed mechanism
Table 13. Addition of Et2Zn to PhCOCO2R (Eq. 6, R1ZPh): Screening the diamine [Ti(L)2-salen, LZOi-Pr, R2Zt-Bu, NR32Zpiperidine]a
Entry
Catalyst
Redn. (%)b
Addn. (%)b
Addn. ee (%)b
1
2
Ti(Oi-Pr)2-13c
Ti(Oi-Pr)2-34
RZMe
RZEt
0
0
99
99
72
62
3
RZEt
18
47
0
4
5
Ti(Oi-Pr)2-35
RZMe
RZEt
0
0
94
92
81
65
a Reactions were in PhCH3 at K40 8C for 2 h using 10 mol% catalyst.
b Determined by chiral GC (Cyclodex b).

M. W. Fennie et al. / Tetrahedron 61 (2005) 6249–6265
6259
Table 14. Addition of Et2Zn to PhCOCO2R (Eq. 6, R1ZPh): Screening the amine Lewis base (NR23) [Ti(L)2-salen, LZOi-Pr, R2Zt-Bu, diamine linkerZ
(R,R)-1,2-cyclohexanediamine]a
Entry
Catalyst
NR32
Addn. (%)b
ee (%)b
1
2
3
Ti(Oi-Pr)2-13c
RZMe
RZEt
RZMe
99
99
96
72
62
51
Ti(Oi-Pr)2-13f
Ti(Oi-Pr)2-13g
4
5
6
7
8
RZMe
RZEt
RZMe
RZEt
89
78
97
96
97
81
72
76
64
70
Ti(Oi-Pr)2-13h
Ti(Oi-Pr)2-13i
Ti(Oi-Pr)2-13j
Ti(Oi-Pr)2-13k
Ti(Oi-Pr)2-13l
RZMe
9
RZMe
RZMe
RZMe
97
90
98
76
80
79
10
11
a Reactions were in PhCH3 at K40 8C for 2 h using 10 mol% catalyst.
b Determined by chiral GC (Cyclodex b). Absolute configuration (R) determined by comparison to the literature. No reduction observed.
(Figure 8, 28) is correct then the nature of the alkoxide
ligands should not dramatically affect the selectivity as the
remaining alkoxide ligand in 28 is distal from the site of
stereochemical induction.40 Indeed, similar results were
obtained when Ti(Ot-Bu)4 was used in place of Ti(Oi-Pr)4
(Table 12, entry 1 vs 3). The TiCl2-complex is partially
insoluble and unreactive in toluene. Although soluble in
CH2Cl2, TiCl2-13c is a slow, nonselective catalyst
(Table 12, entries 4–5).41 When a chiral Ti-alkoxide is
used to generate the salen complex, the influence of the
alkoxide configuration on the enantioselectivity is relatively
small (Table 12, entries 7–9 vs entry 1). An improvement in
enantioselectivity was observed for the catalyst derived
from p-(t-Bu)C6H4OH (entry 6).
Finally we examined catalysts derived from various para-
substituted amino salens (Table 15). The steric (Table 15,
entries 6–7) and electronic (Table 15, entries 3, 8) of the
t-Bu substituent were optimal.42 Thus, only the para-
adamantyl catalyst afforded comparable enantioselectivity
to the t-Bu catalyst (Table 15, entries 4–5 vs entries 1–2).
New catalysts derived from various combinations of
optimal structural elements were then examined
(Table 16). Unfortunately, the ultimate combination of
structural elements is elusive as their respective beneficial
effects do not appear to be additive. To date we have
identified seven highly reactive complexes that catalyze the
addition of diethylzinc to methyl oxo(phenyl)acetate (Eq. 6,
R1ZPh, R2ZMe) in O90% conversion and O80% ee:
Table 12, entry 6; Table 13, entry 4; Table 14, entries 4, 10;
Table 16, entries 2, 3 and 7.
We next examined various diamine linkers. The catalyst
derived from (R,R)-1,2-diphenylethylenediamine is slightly
more selective than that derived from cyclohexanediamine
for both the methyl and ethyl oxo(phenyl)acetate (Table 13,
entries 4–5 vs entries 1–2).
2.8. Substrate scope
The scope of this reaction was examined with the easily
prepared Ti(Oi-Pr)2-13c since it affords a good combination
of reactivity and selectivity. In all cases, the catalyst
greatly accelerates the addition pathway (Table 17) and
provides a moderate degree of stereochemical control.
Interestingly, we discovered a direct correlation between
enantioselectivity and the steric size of the ester substituent
The catalysts derived from substituted piperazine salens are
generally more enantioselective than the lead catalyst derived
from piperidine salen 13c (Table 14). A slight increase in
enantioselectivity was observed for both the ethyl and methyl
oxo(phenyl)acetate with the catalyst derived from N-methyl-
piperazine (Table 14, entries 4–5 vs entries 1–2).
Table 15. Addition of Et2Zn to PhCOCO2R (Eq. 6, R1ZPh): Screening the para-substituent (R2) [Ti(L)2-salen, LZOi-Pr, NR32Zpiperidine, diamine linkerZ
(R,R)-1,2-cyclohexanediamine]a
Entry
Catalyst
R2
Addn. (%)b
Ee (%)b
1
2
3
4
5
6
7
8
Ti(Oi-Pr)2-13c
t-Bu
RZMe
RZEt
RZEt
RZMe
RZEt
RZEt
RZEt
RZEt
99
99
94
98
92
81
87
74
72
62
46
73
61
25
39
16
Ti(Oi-Pr)2-36
Ti(Oi-Pr)2-37
NMe2
1-adamantyl
Ti(Oi-Pr)2-38
Ti(Oi-Pr)2-39
Ti(Oi-Pr)2-40
Me
H
Cl
a Reactions were in PhCH3 at K40 8C for 2 h using 10 mol% catalyst.
b Determined by chiral GC (Cyclodex b). Absolute configuration (R) determined by comparison to the literature. No reduction observed.

6260
M. W. Fennie et al. / Tetrahedron 61 (2005) 6249–6265
Table 16. Addition of Et2Zn to PhCOCO2Me (Eq. 6, R1ZPh, R2ZMe): Combining optimal catalyst elementsa
Entry
Catalyst
L
R2
Addn. (%)b
ee (%)b
1
2
Ti(OAr)2-13g
Ti(OAr)2-41
t-Bu
t-Bu
Me
Me
77
97
43
80
3
4
5
6
7
Ti(Oi-Pr)2-41
Ti(Oi-Pr)2-42
Ti(Oi-Pr)2-43
Ti(Oi-Pr)2-44
Ti(Oi-Pr)2-45
Oi-Pr
Oi-Pr
Oi-Pr
Oi-Pr
Oi-Pr
t-Bu
t-Bu
t-Bu
t-Bu
t-Bu
Me
Et
Boc
Ph
93
98
66
99
88
83
57
60
68
81
2-pyridine
8
Ti(Oi-Pr)2-46
Oi-Pr
Me
Me
91
61
9
Ti(Oi-Pr)2-46
Oi-Pr
91
74
a Reactions were in PhCH3 at K40 8C for 2 h using 10 mol% catalyst.
b Determined by chiral GC (Cyclodex b). Absolute configuration (R) determined by comparison to the literature. No reduction observed.
(Eq. 6, R2). For the benzoylformate ketoesters (R1ZPh),
higher enantioselectivity was observed with smaller ester
substituents (Table 17, entries 1–3). This phenomenon is
apparently general since the enantioselectivity is greater
with the methyl ester (R2ZMe) than with the ethyl ester
(R2ZEt) for a variety of a-ketoesters. One exception to this
trend was observed with the pyruvate ketoesters (R1ZMe).
With this series we observed higher enantioselectivity with
larger ester substituents (Table 17, entries 4–5). The
importance of steric differentiation about the keto carbonyl
is interesting as it provides evidence for catalyst binding at
the keto-carbonyl.
Table 17. Addition of Et2Zn to a-ketoesters (Eq. 6) using (R,R)-Ti(Oi-Pr)2-
13ca
Entry
R1
R2
Yield (%)b
ee (%)b
1
2
3
4
5
Ph
Ph
Ph
Me
Me
Me
Et
Bn
Me
i-Pr
89
78
100
90
72 (R)
62 (R)
50 (R)
16 (S)
65 (S)
98
a Reactions were in PhCH3 at K40 8C for 2 h using 10 mol% catalyst.
b Determined by chiral GC (Cyclodex b). Absolute configuration assigned
by comparison to the literature for entries 1 and 2. For all others, absolute
configuration assigned by analogy. No reduction observed.
In a recent report by Wang and co-workers, it was shown
that small variations in the ligand to Ti(Oi-Pr)4 ratio in the
addition of zinc acetylides to aldehydes had a major impact
on the enantioselectivity of the products.43 Examining
different amounts of piperidine salen and Ti(Oi-Pr)4 showed
that an optimal ratio of 1–7 (ligand to Ti(Oi-Pr)4) increased
the enantioselectivity in the addition of diethylzinc to
methyl benzoylformate from 72 to 88% ee. Using these
conditions, many substrates displayed marked increases
(Table 18).
Table 18. Addition of Et2Zn to a-ketoesters (Eq. 6) using (R,R)-Ti(Oi-Pr)2-
13c prepared with different amounts of Ti(Oi-Pr)4
a
Entry
R1
R2
10%
Ti(Oi-Pr)4
ee (%) b
70%
Ti(Oi-Pr)4
ee (%)b,c
1
2
3
4
5
6
7
8
Ph
Ph
Ph
Fur
p-MeO
p-Br
o-MeO
p-NO2
p-CF3
PhCC
PhCH]CH
PhCCl]CH
Cy
i-Pr
Me
Me
Me
Et
Bn
72 (R)
62 (R)
50 (R)
72 (R)
52 (R)
42 (R)
72 (R)
39 (R)
73 (R)
53 (R)
20 (R)
53 (R)
24 (R)
30 (R)
16 (S)
65 (S)
88 (R)
68 (R)
50 (R)
72 (R)
73 (R)
70 (R)
85 (R)
65 (R)
83 (R)
80 (R)
5 (R)
68 (R)
75 (R)
72 (R)
10 (S)
50 (S)
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
i-Pr
In order to demonstrate the utility of this method, the
addition of diethylzinc to methyl oxo(phenyl)acetate
(Table 18, entry 1) was performed on a 5 mmol scale
using 5 mol% Ti(Oi-Pr)2-13c without additional Ti(Oi-Pr)4
(Scheme 5). The a-hydroxyester was isolated in 96% yield
(99% conv.) with 78% ee. With added Ti(Oi-Pr)4 accord-
ingly higher selectivity levels were obtained. After ester
9
10
11
12
13
14
15
16
a Reactions were in PhCH3 at K40 8C for 2 h using 10 mol% catalyst. No
reduction observed.
b Determined by chiral GC (Cyclodex).
c Determined by chiral HPLC (Chialpak AD). Absolute configuration
assigned by comparison to the literature for entries 1 and 2. For all others,
absolute configuration assigned by analogy.
Scheme 5. Large scale synthesis and enrichment. (a) Et2Zn, Ti(Oi-Pr)2-13c,
PhCH3, K40 8C, 2 h. (b) (i) EtOH, KOH. (ii) Recrystallize.

M. W. Fennie et al. / Tetrahedron 61 (2005) 6249–6265
6261
hydrolysis, the corresponding a-hydroxy acid 48 could be
readily enriched to 98% ee by recrystallization. For this two-
step sequence, a 70% yield of a-hydroxy acid 48 with 98%
ee is obtained.
2.9. Comparison to known methods
We have developed a number of efficient catalysts for the
enantioselective addition of diethylzinc to a-ketoesters. The
catalysts are rapid, chemoselective and general. This
method is superior to other established methods for the
stereoselective synthesis of chiral quaternary hydroxy acids.
One of the most practical diastereoselective alkylations of
a-ketoesters utilizes a large excess of ethyl zinc chloride
(5 equiv) prepared in situ from ethyl Grignard and ZnCl2
(Scheme 6). Utilizing this method, 2-hydroxy-2-phenyl-
butyric acid 48 can be prepared in 97% ee and 60% overall
yield from oxo(phenyl)acetic acid.44 In comparison, our
method provides the product in greater overall yield.
Scheme 8. Synthesis of a Wyeth pharmaceutical agent. Conditions: (a) 3-
MeOPhMgBr. (b) KOH, MeOH. (c) MeI, DBU (74%, three steps). (d)
Et2Zn, Ti(Oi-Pr)2-13c. (e) KOH, MeOH, recrystalization (99% ee, 78%,
two steps). (f) NH2Me, SOCl2 (82%). (g) BH3S(CH3)2 (93%). (h) 2-Chloro-
pentanoyl chloride, Et3N. (i) KOH, i-PrOH. (j) NaH, three recycles (43%,
three steps). (k) LiAlH4 (90%).
addition of 3-methoxyphenyl magnesium bromide to ethyl
oxalate, followed by saponification and methylation, to
provide ketoester substrate 51 (Scheme 8). Ti(Oi-Pr)2-13c
catalyzed alkylation, followed by saponification and
recrystallization provided a-hydroxy acid 52b in 99% ee
and 78% yield. Conversion of the acid to the amide and
subsequent reduction yielded amine 53, which was cyclized
by treatment with 2-chloro-pentanoyl chloride and base to
provide heterocycles 54 and 55 as a 1:1 mixture of
diastereomers which were assigned based on NOE and
ROESY experiments. Conditions to convert diastereomer
54 to the required 55 were explored. Thermodynamic
conditions (tert-butoxide in refluxing THF) gave a 5:1 ratio
of 54 to 55, whereas kinetic conditions (NaH in refluxing
PhCH3) gave a 1.6:1 ratio of 54 to 55. Reduction of 55 with
LiAlH4 provided Wyeth compound 56.
Scheme 6. Diastereoselective synthesis of quaternary a-hydroxyesters. (a)
(1R, 2R)-R*OH, p-TsOH, benzene, 6. (b) 5 equiv EtZnCl, K78 8C (3 h) to
rt (1 h). (c) KOH, EtOH.
The best catalyst for the asymmetric addition of TMSCN to
ketones45 is the titanium complex of ligand 49 (Scheme 7).46
However, the synthesis of ligand 49 is long (O9 steps in the
longest linear sequence) and catalyst preparation is
complex. While catalyst loadings between 1 and 2.5 mol%
are generally sufficient, the reactions require 2–4 d between
K10 and K50 8C. In contrast, the synthesis of chiral ligand
13c requires only three steps (Schemes 1 and 2) and reaction
of the a-ketoester with Ti(Oi-Pr)2-13c requires only 2 h at
K40 8C (Scheme 5).
3. Conclusions
These readily constructed bifunctional salen catalysts are
among the most rapid for the addition of diethyl zinc to
aldehydes, providing high enantioselectivities for both
aromatic and aliphatic substrates. In addition, amino-salen
titanium complexes rapidly alkylate a-ketoesters with
excellent chemoselectivity and good enantioselectivity.
Mechanistic studies have shown that a bifunctional
mechanism is plausible for both of these reactions.
Furthermore, an opiate antagonist has been synthesized
enantioselectively using this methodology.
Scheme 7. Enantioselective synthesis of quaternary a-hydroxyesters.
2.10. Synthesis of a Wyeth pharmaceutical agent
4. Experimental
4.1. General
The utility of this methodology has been highlighted in the
enantioselective synthesis of a Wyeth pharmaceutical agent,
3-[(2R, 6R)-2-ethyl-4-methyl-6-n-propyl-2-morpholinyl]
phenol, which possesses potent opiate antagonistic activity.
The initial report of this compound described a racemic
synthesis.47 The synthesis commenced with Grignard
Full details of the characterization of all new compounds
and the mechanistic experiments can be found in Supporting
information.

6262
M. W. Fennie et al. / Tetrahedron 61 (2005) 6249–6265
4.1.1. 5-tert-Butyl-2-hydroxy-3-piperidin-1-ylmethyl-
benzaldehyde (12c). Piperidine (1.1 equiv) and paraform-
aldehyde (1.1 equiv) were combined in EtOH (0.3 M). The
slurry was heated at 60 8C for 1–2 h (a clear solution usually
results). Salicylaldehyde 8 (1.0 equiv) was added in one
portion. After heating at reflux for 3 d, the EtOH was
removed in vacuo. Chromatography afforded compound 12c
as a pale yellow solid. 1H NMR (500 MHz, CDCl3) d 10.40
(s, 1H), 7.65 (s, 1H), 7.26 (s, 1H), 3.71 (s, 2H), 2.03 (br s,
4H), 1.66 (m, 4H), 1.50 (br s, 2H), 1.28 (s, 9H); 13C NMR
(125 MHz, CDCl3) d 191.1, 159.9, 141.5, 132.3, 123.9,
122.8, 122.4, 61.3, 53.9, 34.1, 31.3, 25.7, 23.8. Compound
12c has previously been synthesized by an alternative
method.48
of 1 mL water, allowed to warm to room temperature, and
then extracted with pentane or EtOAc. The extracts were
dried over Na2SO4, and concentrated. Chromatography
(10% EtOAc/hexanes) afforded pure a-hydroxy ester.
4.1.6. 2-Hydroxy-2-(3-methoxy-phenyl)-butyric acid
methyl ester (52a). (93%): [a]2D0 K9.9 (c 0.015, CH2Cl2,
85% ee); IR (film) 3513, 2957, 2883, 2837, 1729, 1602 cmK1
;
1H NMR (500 MHz, CDCl3) d 0.88 (t, JZ10 Hz, 3H), 1.91–
1.99 (m, 1H), 2.12–2.19 (m, 1H), 3.73 (s, 3H), 3.75 (s, 3H),
6.85–6.87 (m, 1H), 7.17–7.21 (m, 2H), 7.28–7.30 (m, 1H);
13C NMR (500 MHz, CDCl3) dZ8.5, 33.1, 53.6, 55.6, 79.2,
111.9, 113.4, 118.4, 129.6, 143.9, 160.0, 176.1; HRMS
(ESI) calcd for C12H16O4 (MNaC) 247.0946, found
247.0930.
4.1.2. (K)-(R,R)-Cyclohexanediamine-(p-t-Bu)-(piper-
idin-1-ylmethyl)-salen (13c). Salicylaldehyde 12c
(2.0 equiv) in EtOH (1 mL/100 mg salicylaldehyde) was
treated with the diamine (1.0 equiv) and was allowed to stir
at room temperature for 24 h. After removal of the volatiles,
13c was crystallized from EtOH as yellow clusters (90%
yield): mp 88–94 8C; [a]2D0 K315 (c 0.5, CHCl3); IR (film)
4.1.7. 2-Hydroxy-2-(3-methoxy-phenyl)-butyric acid
(52b). The a-hydroxy ester (168 mg, 0.75 mmol) was
stirred 5% KOH in EtOH (5 mL). After 1.5 h, the reaction
was diluted with water and brought to pH 2 with HCl (10%).
The mixture was extracted with EtOAc 2X, washed with
brine, and dried (Na2SO4). Crystallization from CCl4
yielded 52b as a white solid (78% yield): mp 118–122 8C;
[a]2D0 K26 (c 0.03, CH2Cl2, O99% ee); IR (film) 3428,
1
2933, 2857, 2791, 2756, 1629 (s), 1467, 1271 cmK1; H
NMR (500 MHz, CDCl3) d 13.35 (br s, 2H), 8.29 (s, 2H)
(HC]N), 7.36 (s, 2H), 7.06 (s, 2H), 3.58 (d, JZ13.4 Hz,
2H), 3.50 (d, JZ13.4 Hz, 2H) [AB], 3.29 (m, 2H), 2.43 (br
s, 8H), 1.87 (m, 4H), 1.70– 1.36 (m, 16H), 1.24 (s, 18H); 13C
NMR (CDCl3) d 164.8, 157.2, 140.4, 130.5, 126.4, 125.0,
117.9, 72.7, 57.0, 54.5, 33.8, 32.3, 31.4, 26.0, 24.4, 24.3;
MS (ES) m/z 629 (MHC); HRMS (ES) m/z calcd for
C40H61N4O2 (MHC) 629.4794, found 629.4786.
2972, 2941, 2841, 2629, 1718, 1602 cmK1 1H NMR
;
(CDCl3) 0.97 (t, JZ12 Hz, 3H), 2.02–2.14 (m, 1H), 2.23–
2.34 (m, 1H), 3.83 (s, 3H), 6.84–6.88 (m, 1H), 7.19–7.22
(m, 2H), 7.27–7.32 (m, 1H); 13C NMR (500 MHz, CDCl3) d
8.4, 33.1, 55.7, 79.1, 112.0, 113.7, 118.4, 129.7, 143.1,
160.0, 179.6; HRMS (ESI) calcd for C11H14O4 (MC)
210.0892, found 210.0884.
4.1.3. General procedure for addition of Et2Zn and
Me2Zn to aldehydes. The salen (0.025 mmol) was
introduced into a dry Schlenk flask, and the system was
purged with N2. After dissolution in PhCH3 (1 mL), Et2Zn
(25 mL, 1.0 M in PhCH3, 0.025 mmol) or Me2Zn (12.5 mL,
2.0 M in PhCH3, 0.025 mmol) was added. After stirring at rt
for 1 h, the reaction was cooled to the indicated reaction
temperature. Et2Zn (500 mL, 1.0 M in PhCH3, 0.500 mmol)
or Me2Zn (250 mL, 2.0 M in PhCH3, 0.500 mmol) was
added slowly. After 5 min, the aldehyde (0.25 mmol)
was added dropwise. At the specified time, the reaction
was quenched with 1 N HCl (aromatic aldehydes) or
saturated NH4Cl (aliphatic aldehydes), and then extracted
with pentane. Spectral data for the addition of Et2Zn to
benzaldehyde is identical to that reported in literature.49
4.1.8.
2-Hydroxy-2-(3-methoxy-phenyl)-N-methyl-
butyramide. To a solution of 52b (394 mg, 1.9 mmol) in
PhCH3 and THF (1:1) at K5 8C was added of SOCl2
(164 mL, 2.3 mmol) followed by dropwise addition of
methylamine (8 mL, 2 M in THF, 19 mmol). After com-
pletion, the mixture was washed with brine and back
extracted with PhCH3. The combined organic extracts were
washed with brine and dried (Na2SO4). Chromatography
(10% MeOH/CH2Cl2) afforded pure a-hydroxy amide as a
white solid (82%); mp 98–104 8C; [a]2D0 K11.8 (c 0.0093,
CH2Cl2, O99% ee); IR (film) 3381, 2968, 2883, 1652,
1602, 1536 cmK1; H NMR (500 MHz, CDCl3) d 0.93 (t,
1
JZ12.3 Hz, 3H), 2.07–2.12 (m, 1H), 2.20–2.30 (m, 1H),
2.80 (d, JZ4.9 Hz, 3H), 3.82 (s, 3H), 6.82–6.86 (m, 1H),
7.14–7.16 (m, 2H), 7.29 (m, 1H); 13C NMR (500 MHz,
CDCl3) d 8.0, 26.6, 32.6, 55.5, 79.3, 111.9, 113.3, 118.0,
129.7, 144.4, 160.0, 174.7; HRMS (ESI) calcd for
C12H17NO3 (MC) 223.1202, found 223.1202.
4.1.4. General procedure for formation of the metal
salen complexes. The salen (0.030 mmol) was introduced
into a dry Schlenk flask, and the system purged with N2.
After dissolution in PhCH3 (1 mL), Ti(Oi-Pr)4 (18 mL,
1.4 M in hexanes, 0.025 mmol) was added. After stirring at
room temperature for 1 h the PhCH3, the volatiles were
removed in vacuo. Fresh PhCH3 (1 mL) was added to
provide the catalyst solution.
4.1.9. 2-(3-Methoxy-phenyl)-1-methylamino-butan-2-ol
(53). BH3$SMe2 (182 mL, 1.68 mmol) was added dropwise
under N2 to a stirred suspension of the a-hydroxy amide
(150 mg, 0.67 mmol) in PhCH3 at 0 8C. After heating at
105 8C for 2.5 h, the reaction was quenched with 10%
aqueous Na2CO3 and extracted with PhCH3. The combined
organic extracts were dried (Na2SO4) and concentrated.
Kugelrohr distillation provided pure amine 53 as a yellow
oil (93%); [a]2D0 C35 (c 0.033, CH2Cl2); IR (film) 3324,
3078, 2834, 2797, 2366, 2268, 2082, 1600, 1582 cmK1; 1H
NMR (500 MHz, CDCl3) d 0.70 (t, JZ7.4 Hz, 3H), 1.65–
1.70 (m, 2H), 2.28 (s, 3H), 2.58 (d, JZ11.7 Hz, 1H), 2.95
4.1.5. General procedure for the addition of diethylzinc
to a-ketoesters. After the catalyst solution (see above) was
cooled to K40 8C, Et2Zn (300 mL, 1.0 M in PhCH3,
0.30 mmol) was added slowly. After 30 min, the a-ketoester
(0.25 mmol) was added dropwise neat or as a 1 M solution
in PhCH3. After 2 h, the reaction was quenched by injection

M. W. Fennie et al. / Tetrahedron 61 (2005) 6249–6265
6263
(d, JZ11.6 Hz, 1H), 3.74 (s, 3H), 6.79–6.81 (m, 1H), 6.98–
6.99 (m, 1H), 7.04–7.05 (m, 1H), 7.38–7.40 (m, 1H); 13C
NMR (500 MHz, CDCl3) d 7.4, 33.4, 36.5, 55.0, 61.5, 75.1,
111.4, 111.6, 111.7, 128.9, 158.0, 169.0; HRMS (ESI) calcd
for C12H19NO2 (MHC) 210.1494, found 210.1504.
1H), 2.96 (d, JZ10.9 Hz, 1H), 3.82 (s, 3H), 3.78–3.85 (m,
1H), 6.77–6.79 (dd, JZ2.6, 8.12 Hz, 1H), 6.94–6.97 (m,
1H), 7.04–7.05 (m, 1H), 7.23–7.28 (m, 1H); 13C NMR
(500 MHz, CDCl3) d 7.1, 14.6, 19.0, 26.8, 30.0, 36.5, 46.6,
55.4, 60.7, 64.7, 68.7, 111.5, 111.8, 117.5, 117.5, 129.1,
159.7; HRMS (ESI) calcd for (MHC) 278.2120, found
278.2119.
4.1.10.
6-Ethyl-6-(3-methoxy-phenyl)-4-methyl-2-
propyl-morpholin-3-one (54). To a solution of of 53
(45 mg, 0.215 mmol) in CH2Cl2 at 0 8C under N2 was added
2-chloro-pentanoyl chloride (37 mg, 1 M in CH2Cl2,
0.258 mmol). After 6 h, the reaction mixture was washed
with 1 M HCl, dried (MgSO4) and concentrated. After
dissolution in i-PrOH (10 mL), aqueous NaOH (2 mL,
10 M) was added dropwise at 0 8C and the mixture warmed
to rt and stirred overnight. After removal of the volatiles and
dilution with H2O, the reaction mixture was neutralized
with 10% HCl, extracted with CH2Cl2, and dried (MgSO4).
Chromatography (60% EtOAc/hexanes) afforded a 1:1
mixture of diastereomers 54 and 55 as colorless oils.
NOE experiments on diastereomer 55 found a strong NOE
between H6 and H10a (absence of NOE for H10b). A NOE
was also present between H6 and protons on C11.
Acknowledgements
We thank the NIH (GM59945) and the University of
Pennsylvania for support of this work. Fellowships from the
Sloan Foundation (M.C.K., Research Fellow), Bristol-
Meyers Squibb (E.F.D.), the GAANN program (M.W.F.),
and Eli Lilly (E.M.O.) are gratefully acknowledged. We
thank Professor Patrick Walsh for a sample of titanium
tetra[(S)-1-(4-methylphenyl)propoxide] and for helpful
discussions.
Supplementary data
Diastereomer 54 was converted to 55 by refluxing in NaH
(20 equiv) and THF. The reaction was quenched with H2O
and extracted with ether to yield diastereomer 55 after three
recycles (43% yield, three steps): diastereomer 54; [a]2D0C
46 (c 0.0065, CH2Cl2); IR (film) 2961, 2870, 1656, 1582,
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/j.tet.2005.03.
117
1
1488 cmK1; H NMR (500 MHz, CDCl3) d 0.66 (t, JZ
7.6 Hz, 3H), 0.88 (t, JZ7.4 Hz, 3H), 1.20 (m, 2H), 1.73–
1.79 (m, 4H), 2.94 (s, 3H), 3.58 (d, JZ12.8 Hz, 1H), 3.68
(d, JZ12.7 Hz, 1H), 3.75 (s, 3H), 3.83–3.86 (dd, JZ3.62,
11.02 Hz, 1H), 6.74–6.69 (m, 2H), 7.19–7.23 (m, 2H); 13C
NMR (500 MHz, CDCl3) d 8.2, 14.2, 18.7, 34.9, 35.7, 53.9,
55.5, 73.1, 77.1, 112.7, 113.2, 118.9, 129.9, 141.7, 160.2,
170.0; HRMS (ESI) calcd for C17H25NO3 292.1913 (MHC),
found 292.1925. Diastereomer 55; [a]2D0K28 (c 0.005,
CH2Cl2); IR (film) 2961, 2873, 2358, 1651, 1584,
References and notes
1. For reviews on bifunctional catalysis, see: (a) Hayashi, T. In
Organic synthesis: an interdisciplinary challenge; Streith, J.,
Prinzbach, H., Schill, G., Eds.; Blackwell Scientific: Boston,
1985; p 35. (b) Hayashi, T.; Kumada, M. Acc. Chem. Res.
1982, 15, 395–401. (c) Sawamura, M.; Ito, Y. Chem. Rev.
1992, 92, 857–871. (d) van den Beuken, E. K.; Feringa, B. L.
Tetrahedron 1998, 54, 12985–13011. (e) Rowlands, G. J.
Tetrahedron 2001, 57, 1865–1882. (f) Shibasaki, M.; Yoshi-
kawa, N. Chem. Rev. 2002, 102, 2187–2210. (g) Cahard, D.;
Ma, J. A. Angew. Chem. Int. Ed. 2004, 43, 4566.
1
1488 cmK1; H NMR (500 MHz, CDCl3) d 0.62 (t, JZ
7.3 Hz, 3H), 0.93 (t, JZ7.4 Hz, 3H), 1.52 (sextet, JZ
7.5 Hz, 2H), 1.65–1.70 (m, 1H), 1.75–1.80 (m, 1H), 1.95–
19.8 (m, 1H), 2.22 (sextet, JZ7.3 Hz, 1H), 3.24–3.26 (d,
JZ12.3 Hz, 1H), 3.45–3.51 (d, JZ12.3 Hz, 1H), 1.84 (s,
3H), 4.14–4.16 (dd, JZ3.3, 11.3 Hz, 1H), 6.75–6.78 (m,
1H), 6.86–6.88 (m, 1H), 6.95–6.96 (m, 1H), 7.20–7.24 (m,
1H); 13C NMR (500 MHz, CDCl3) d 7.3, 14.2, 18.9, 26.9,
35.1, 35.1, 55.5, 58.5, 72.0, 111.8, 112.1, 117.2, 129.6,
144.6, 159.8, 169.9; HRMS (ESI) calcd for C17H25NO3
(MHC) 292.1915, found 292.1920.
2. (a) Dreyer, M. K.; Schulz, G. E. J. Mol. Biol. 1996, 259,
458–466. (b) Sugimoto, T.; Kaiser, E. T. J. Am. Chem. Soc.
1978, 100, 7750–7751. (c) Scrutton, M. C. In Eichhorn, G. L.,
Ed.; Inorganic biochemistry; Elsevier: Amsterdam, 1973; Vol.
1.
3. Corey, E. J.; Bakshi, R. K.; Shibata, S. J. Am. Chem. Soc. 1987,
109, 5551–5553.
4.1.11. 2-Ethyl-2-(3-methoxy-phenyl)-4-methyl-6-pro-
pyl-morpholine (56). After heating a suspension of
LiAlH4 (39 mg, 1.02 mmol) and 55 (14.8 mg,
0.051 mmol) in PhCH3 (2 mL) at reflux for 2 h, the reaction
mixture was quenched with H2O, followed by NaOH, and
H2O. The mixture was extracted with PhCH3, washed with
brine, and dried (Na2SO4). Chromatography (10% MeOH/
CH2Cl2) afforded 57 as a single diastereomer (90% yield).
[a]2D0 K14.5 (c 0.0047, CH2Cl2); IR (film) 2934, 2873,
4. (a) Noyori, R.; Kitamura, M. Angew. Chem., Int. Ed. Engl.
1991, 30, 49–69. (b) Kitamura, M.; Suga, S.; Niwa, M.;
Noyori, R. J. Am. Chem. Soc. 1995, 117, 4832–4842.
5. Yoon, T. P.; Jacobsen, E. N. Science 2003, 299, 1691–1693.
6. Bandini, M.; Cozzi, P. G.; Umani-Ronchi, A. Pure Appl.
Chem. 2001, 73, 325–329.
7. For reviews, see: (a) Katsuki, T. Synlett 2003, 281–297. (b)
Katsuki, T. Adv. Synth. Catal. 2003, 344, 133–147. (c) Ito,
Y. N.; Katsuki, T. Bull. Chem. Soc. Jpn. 1999, 72, 603–619. (d)
Canali, L.; Sherrington, D. C. Chem. Soc. Rev. 1999, 28,
85–93.
1
2792, 1610, 1583 cmK1; H NMR (500 MHz, CDCl3) d
0.62 (t, JZ7.4 Hz, 3H), 0.98 (t, JZ7.2, Hz, 3H), 1.43–1.46
(m, 2H), 1.46–1.67 (m, 5H), 1.94 (d, JZ11.1 Hz, 1H), 2.28
(s, 3H), 2.66 (sextet, JZ7.4 Hz, 1H), 2.78 (d, JZ9.2 Hz,
8. For selected examples see: Al: (a) Sammis, G. M.; Jacobsen,

6264
M. W. Fennie et al. / Tetrahedron 61 (2005) 6249–6265
E. N. J. Am. Chem. Soc. 2003, 125, 4442–4443. Cr: (b) Schaus,
Ref. Kitajima, H.; Ito, K.; Aoki, Y.; Katsuki, T. Bull. Chem.
Soc. Jpn. 1997, 207–217.
S. E.; Branalt, J.; Jacobsen, E. N. J. Org. Chem. 1998, 63,
403–405. Mg: (c) Corazza, F.; Floriani, C.; Chiesi-Villa, A.;
Guastini, C. J. Chem. Soc., Dalton Trans. 1988, 2341–2345.
Ti: (d) Belokon, Y.; Caveda-Cepas, S.; Geen, B.; Ikonnikov,
N. S.; Khrustalev, V. N.; Larichev, V. S.; Moscalenko, M. A.;
North, M.; Orizu, C.; Tararov, V. I.; Tasinazzo, M.;
Timofeeva, G. I.; Yashkina, L. V. J. Am. Chem. Soc. 1999,
121, 3968–3973. V: (e) Belokon, Y. N.; North, M.; Parsons, T.
Org. Lett. 2000, 1617–1619. Zn: (f) Singer, A. L.; Atwood,
D. A. Inorg. Chim. Acta 1998, 157–162.
20. Ni and Cu complexes of 13b were prepared from the
corresponding metal acetates while the Mg and Ti(Oi-Pr)2
complexes were generated in situ by treatment of 13b with
1.0 equiv of MgBu2 or Ti(Oi-Pr)4 for 1 h.
21. For an argument which correlates thermodynamic basicity
(conjugate acid pKa) and Lewis basicity, see: Aggarwal, V. K.;
Emme, I.; Fulford, S. Y. J. Org. Chem. 2003, 68, 692–700.
22. For qualitative measurements of the electron-pair donating
properties of solvent molecules toward different types of
solutes and the deduced relative scales for Lewis basicity, see:
Sandstroem, M.; Persson, I.; Persson, P. Acta Chem. Scand.
1990, 44, 653–675.
23. N-methylmorpholine (pKa(HBC)w7.5) was used as a Lewis
base additive in all cooperative effects studies, since it is
expected to exhibit electron-pair donor ability midway
between morpholine salen 13b (pKa(HBC)w6) and piperidine
salen 13c (pKa(HBC)w9).
9. Dimauro, E. F.; Mamai, A.; Kozlowski, M. C. Organo-
metallics 2003, 22, 850–855.
10. Lam, F.; Xu, J. X.; Chan, K. S. J. Org. Chem. 1996, 61,
8414–8418.
11. For selected examples see: (a) Hu, Y.; Hu, H. Synthesis 1991,
325–326. (b) Karunakaran, S.; Kandaswamy, M. J. Chem.
Soc., Dalton Trans. 1995, 1851–1855. (c) White, D. J.; Laing,
N.; Miller, H.; Parsons, S.; Coles, S.; Tasker, P. A. Chem.
Commun. 1999, 2077–2078. (d) Miller, H. A.; Laing, N.;
Parsons, S.; Parkin, A.; Tasker, P. A.; White, D. J. J. Chem.
Soc., Dalton Trans. 2000, 3773–3782. (e) Black, D.; Blake,
A. J.; Finn, R. L.; Lindoy, L. F.; Nezhadali, A.; Rougnaghi, G.;
Tasker, P. A.; Schroeder, M. Chem. Commun. 2002, 4,
340–341. (f) Coxall, R. A.; Lindoy, L. F.; Miller, H. A.;
Parkin, A.; Parsons, S.; Tasker, P. A.; White, D. J. J. Chem.
Soc., Dalton Trans. 2003, 55–64.
24. For reviews of the enantioselective construction of quaternary
centers see: (a) Martin, S. F. Tetrahedron 1980, 36, 419–460.
(b) Corey, E. J.; Guzman-Peresz, A. Angew. Chem., Int. Ed.
1998, 37, 388–401. (c) Christoffers, J.; Mann, A. Angew.
Chem., Int. Ed 2001, 40, 4591–4597. (d) Betancort, J. M.;
´
Garcıa, C.; Walsh, P. J. Synlett 2004, 749–760.
25. Coppola, G. M.; Schuster, H. F. a-Hydroxy acids in
enantioselective synthesis; VCH: Weinhein, 1997.
12. (a) DiMauro, E. F.; Kozlowski, M. C. Org. Lett. 2001, 3,
3053–3056. (b) DiMauro, E. F.; Kozlowski, M. C. Org. Lett.
2002, 4, 3781–3784. (c) DiMauro, E. F.; Kozlowski, M. C.
J. Am. Chem. Soc. 2002, 124, 12668–12669.
26. Mukaiyama aldol: (a) Evans, D. A.; Burgey, C. S.; Kozlowski,
M. C.; Tregay, S. W. J. Am. Chem. Soc. 1999, 121, 686–699.
Isocyanoacetates: (b) Ito, Y.; Sawamura, M.; Hamashima, H.;
Emura, T.; Hayashi, T. Tetrahedron Lett. 1989, 30,
4681–4684. Self-aldol: (c) Juhl, K.; Gathergood, N.;
Jørgensen, K. A. Chem. Commun. 2000, 2211–2212.
13. For a review, see: Henry, H. In Trost, B. M., Ed.;
Comprehensive organic synthesis; Pergamon: Oxford, 1991;
Vol. 2, p 953.
27. For selected examples see: (a) Whitesell, J. K.; Deyo, D.;
Bhattacharya, A. Chem. Commun. 1983, 802. (b) Basavaiah,
D.; Bharathi, T. K. Tetrahedron Lett. 1991, 32, 3417–3420. (c)
Tamai, Y.; Nakano, T.; Miyano, S. Chem. Lett. 1992, 807–808.
(d) Tamai, Y.; Nakano, T.; Miyano, S. J. Chem. Soc., Perkin
Trans. 1 1994, 439–445. (e) Basavaiah, D.; Pandiaraju, S.;
Bakthadoss, M.; Mathukumaran, K. Tetrahedron: Asymmetry
1996, 7, 997–1000. (f) Loupy, A.; Monteux, D. A. Tetrahedron
2002, 58, 1541–1549 and references cited therein. (g) Wipf, P.;
Stephenson, C. R. J. Org. Lett. 2003, 5, 2449–2452.
14. (a) Schmidt, B.; Seebach, D. Angew. Chem., Int. Ed. Engl.
1991, 30, 1321–1323. (b) Seebach, D.; Beck, A. K.; Schmidt,
B.; Wang, Y. M. Tetrahedron 1994, 50, 4363–4384. (c)
Takahashi, H.; Kawakita, T.; Yoshioka, M.; Kobayashi, S.;
Ohno, M. Tetrahedron 1998, 30, 7095–7098. (d) Takahashi,
H.; Kawakita, T.; Ohno, M.; Yoshioka, M.; Kobayashi, S.
Tetrahedron 1992, 48, 5691–5700. (e) Pritchett, S.;
Woodmansee, D. H.; Gantzel, P.; Walsh, P. J. J. Am. Chem.
Soc. 1998, 120, 6423–6424. (f) Seebach, D.; Beck, A. K.;
Heckel, A. Angew. Chem., Int. Ed. 2001, 40, 92–138.
15. The method of Ohno et al. Takahashi, H.; Kawakita, T.;
Yoshioka, M.; Kobayashi, S.; Ohno, M. Tetrahedron 1998, 30,
7095–7098. Takahashi, H.; Kawakita, T.; Ohno, M.;
Yoshioka, M.; Kobayashi, S. Tetrahedron 1992, 48,
5691–5700 makes use of a catalytic (30–60 mol%) amount
of titanium..
28. (a) Noyori, R.; Suga, S.; Kawai, K.; Okada, S.; Kitamura, M.
Pure Appl. Chem. 1988, 60, 1597–1606. (b) Zadel, G.;
Breitmaier, E. Chem. Ber. 1994, 127, 1323–1326. (c) Tan,
L.; Chen, C.; Tillyer, R. D.; Grabowski, E. J. J.; Reider, P. J.
Angew. Chem., Int. Ed. 1999, 711–713.
29. See (a) Weber, B.; Seebach, D. Tetrahedron 1994, 50,
6117–6128. (b) Nozaki, H.; Aratani, T.; Toraya, T.; Noyori,
R. Tetrahedron 1971, 27, 905–913. (c) Abenhaim, D.;
Boireau, G.; Sabourault, B. Tetrahedron Lett. 1980, 21,
3043–3046.
16. (a) Noyori, R.; Kitamura, M. Angew. Chem., Int. Ed. Engl.
1991, 30, 49–69. (b) Kitamura, M.; Suga, S.; Niwa, M.;
Noyori, R. J. Am. Chem. Soc. 1995, 117, 4832–4842.
17. (a) Pu, L.; Yu, H. Chem. Rev. 2001, 101, 757–824. (b) Soai, K.;
Shibata, T. In Comprehensive asymmetric catalysis; Jacobsen,
E. N., Pfaltz, A., Yamamoto, H., Eds.; Springer: Berlin, 1999;
pp 911–922.
30. Funabashi, K.; Jachmann, M.; Kanai, M.; Shibasaki, M.
Angew. Chem. Int., Ed. 2003, 42, 5489–5492.
31. Berrisford, D. J.; Bolm, C.; Sharpless, K. B. Angew. Chem.,
Int. Ed. Engl. 1995, 34, 1059–1070.
18. For previous applications of salen ligands as catalysts in the
diethylzinc addition to aldehydes see Keller, F.; Rippert, A. J.
Helv. Chim. Acta 1999, 82, 125–137. (b) Cozzi, P. G.; Papa,
A.; Umani-Ronchi, A. Tetrahedron Lett. 1996, 37, 4613–4616.
19. Bifunctional catalysts have been proposed in the diethylzinc
32. (a) Ph2Zn: (a) Dosa, P. I.; Fu, G. C. J. Am. Chem. Soc. 1998,
120, 445–446. (b) Garcia, C.; Walsh, P. J. Org. Lett. 2003, 5,
3641–3644. (c) Prieto, O.; Ramon, D. J.; Yus, M. Tetrahedron:
Asymmetry 2003, 14, 1955–1957. (d) Li, H. M.; Garcia, C.;
Walsh, P. Proc. Natl. Acad. Sci. 2004, 101, 5425–5427.
additions
with
binaphthyl-3,30-dicarboxamides.
See
´
dialkylzincs: (e) Ramon, D. J.; Yus, M. Tetrahedron Lett.

M. W. Fennie et al. / Tetrahedron 61 (2005) 6249–6265
6265
´
1998, 39, 1239–1242. (f) Ramon, D. J.; Yus, M. Tetrahedron
was prepared from 2-hydroxy-5-nitro-benzaldehyde by the
established method: Ando, M.; Emoto, S. Bull. Chem. Soc.
Jpn. 1978, 51, 2433–2434.2-Hydroxy-5-tricyclo[3.3.1.10.0]
dec-1-yl-benzaldehyde was prepared from 4-adamantyl phenol
in a manner analogous to the preparation of 8.
1998, 54, 5651–5666. (g) Garcia, C.; LaRochelle, L. K.;
Walsh, P. J. J. Am. Chem. Soc. 2002, 124, 10970–10971. (h)
Yus, M.; Ramon, D. J.; Prieto, O. Tetrahedron: Asymmetry
2002, 13, 2291–2293. (i) Jeon, S. J.; Walsh, P. J. J. Am. Chem.
´
Soc. 2003, 125, 9544–9545. (j) Yus, M.; Ramon, D. J.; Prieto,
40. Jiang, Y.; Gong, L.; Feng, X.; Hu, W.; Pan, W.; Li, Z.; Mi, A.
Tetrahedron 1997, 53, 14327–14338.
O. Tetrahedron: Asymmetry 2003, 14, 1103–1114. (k) Jeon,
S. J.; Li, H. M.; Garcia, C.; LaRochelle, L. K.; Walsh, P. J.
J. Org. Chem. 2005, 70, 448–455. vinyl zincs: (l) Li, H. M.;
Walsh, P. J. J. Am. Chem. Soc. 2004, 126, 6538–6539. alkyl,
aryl, vinyl zincs: (m) Betancort, J. M.; Garcia, C.; Walsh, P. J.
Synlett 2004, 749–760. (n) de Parrodi, C. A.; Walsh, P. J.
Synlett 2004, 2417–2420.
41. The relative rates of ligand exchange from TiL2-salen parallels
the relative pKa values of LH: see Chen, H.; White, P. S.;
´
Gagne, M. R. Organometallics 1998, 17, 5358–5366. A
similar lack of reactivity was observed by Belokon and North
in their studies of the addition of TMSCN to RCHO with a
TiCl2-salen catalyst prepared in situ from Ti(Oi-Pr)2Cl2 and
salen: see Belokon, Y.; Flego, M.; Ikonnikov, N.; Margarita,
M.; North, M.; Orizu, C.; Tararov, V.; Tasinazzo, M. J. Chem.
Soc., Perkin Trans. 1 1997, 1293–1295.
33. Noyori, R.; Kitamura, M. Angew. Chem., Int. Ed. Engl. 1991,
30, 49–69.
34. For other examples of this problem, see: Yong, K. H.; Chong,
J. M. Org. Lett. 2002, 4, 3553–3556. Diastereoselective
additions to a-ketoesters are often limited to MeMgX and
ArMgX which do not contain b-hydrogens.
35. Nugent, W. Chem. Commun. 1999, 1369–1370.
36. No enantioselection was seen in a prior DAIB a-ketoester
alkylation: Noyori, R.; Suga, S.; Kawai, K.; Okada, S.;
Kitamura, M.; Oguni, N.; Hayashi, M.; Kaneko, T.; Matsuda,
Y. J. Organomet. Chem. 1990, 382, 19–37.
42. These results are similar to those observed in the Ti(Oi-Pr)2-
salen catalyzed addition of trimethylsilylcyanide to aldehydes:
see Jiang, Y.; Gong, L.; Feng, X.; Hu, W.; Pan, W.; Li, Z.; Mi,
A. Tetrahedron 1997, 53, 14327–14338. and the MnCl-salen
catalyzed epoxidation: (a) Jacobsen, E. N.; Zhang, W.; Gular,
M. L. J. Am. Chem. Soc. 1991, 113, 6703–6704.
43. Xu, Z.; Wang, R.; Xu, J.; Da, C.; Yan, W.; Chen, C. Angew.
Chem. Int., Ed. 2003, 42, 5747–5749.
37. (a) Jiang, Y.; Gong, L.; Feng, X.; Hu, W.; Pan, W.; Li, Z.; Mi,
A. Tetrahedron 1997, 53, 14327–14338. (b) Seebach, D.;
Marti, R. E.; Hintermann, T. Helv. Chim. Acta 1996, 79,
1710–1740.
44. Basavaiah, D. Rama Krishna. P. Tetrahedron 1995, 51,
12169–12178.
45. For a review, see: North, M. Tetrahedron: Asymmetry 2003,
147–176.
38. Zhou, X-G. ; Che, C-M. J. Chem. Soc., Dalton Trans. 1999,
3303–3309.
46. Hamashima, Y.; Kanai, M.; Shibasaki, M. Tetrahedron Lett.
2001, 42, 691–694.
39. 5-tert-Butyl-2-hydroxy-benzaldehyde (8) was prepared as
illustrated in Scheme 1 Casiraghi, G.; Casnati, G.; Puglia,
G.; Sartori, G.; Terenghi, G. J. Chem. Soc., Perkin Trans. 1
1980, 1862–1865. 5-Dimethylamino-2-hydroxy-benzaldehyde
47. White, A. C. (Wyeth) EP. 8405143, 8319372, 8061816.
48. Miller, H. A.; Laing, N.; Parsons, S.; Parkin, A.; Tasker, P. A.;
White, D. J. J. Chem. Soc., Dalton Trans. 2000, 3773–3782.
49. Lutz, C.; Knochel, P. J. Org. Chem. 1997, 62, 7895–7898.