New chiral bis(diphenylphospholane) ligands: Design, synthesis, and application to catalytic enantioselective aldol reaction to ketones

Article abstract of DOI:10.1016/j.tetlet.2005.04.084

New chiral bidentate diphenylphospholanes were designed targeting a catalytic enantioselective aldol reaction to ketones. Ligands 5l and 5m having cis-2-butenyl and cyclopropyl groups at the linker part, respectively, were identified as effective chiral ligands for a CuF-catalyzed enantioselective aldol reaction to ketones. Catalysts prepared from CuF?3PPh ₃?2EtOH and these ligands produced ketone aldol products with up to 66% ee, which is promising particularly for this extremely difficult and important catalytic enantioselective carbon-carbon bond forming reaction. The enantioselectivity was strongly dependent on the linker structure. Construction of a deep chiral pocket around the copper metal with stable bidentate chelation is the key to meaningful enantioinduction.

Full text of DOI:10.1016/j.tetlet.2005.04.084

Tetrahedron
Letters
Tetrahedron Letters 46 (2005) 4325–4329
New chiral bis(diphenylphospholane) ligands:
design, synthesis, and application to catalytic enantioselective
aldol reaction to ketones
Kounosuke Oisaki,a Dongbo Zhao,a Yutaka Suto,a Motomu Kanaia,b,
*
and Masakatsu Shibasakia,*
aGraduate School of Pharmaceutical Sciences, The University of Tokyo, Tokyo 113–0033, Japan
bPRESTO, Japan Science and Technology Corporation, Japan
Received 21 March 2005; revised 22 April 2005; accepted 25 April 2005
Available online 10 May 2005
Abstract—New chiral bidentate diphenylphospholanes were designed targeting a catalytic enantioselective aldol reaction to ketones.
Ligands 5l and 5m having cis-2-butenyl and cyclopropyl groups at the linker part, respectively, were identified as effective chiral
ligands for a CuF-catalyzed enantioselective aldol reaction to ketones. Catalysts prepared from CuFÆ3PPh3Æ2EtOH and these ligands
produced ketone aldol products with up to 66% ee, which is promising particularly for this extremely difficult and important cata-
lytic enantioselective carbon–carbon bond forming reaction. The enantioselectivity was strongly dependent on the linker structure.
Construction of a deep chiral pocket around the copper metal with stable bidentate chelation is the key to meaningful
enantioinduction.
Ó 2005 Elsevier Ltd. All rights reserved.
Catalytic enantioselective construction of chiral tertiary
alcohols through carbon–carbon bond formation is one
of the most important and challenging research targets
in current organic synthesis.1 Although there are many
biologically active naturally occurring compounds and
artificial pharmaceuticals containing chiral tertiary alco-
hols, there are only few reliable and practical synthetic
methods for these chiral building blocks. Catalytic enan-
tioselective cyanosilylation,2 organozinc addition,3 and
allylation4 of simple ketones are recent entries in this
category. One extremely important piece, however, is
missing; that is, the catalytic enantioselective aldol reac-
tion to ketones. The first and only example of a catalytic
enantioselective aldol reaction to simple ketones was re-
ported by Denmark and Fan.5 The use of very unstable
trichlorosilyl enolate as a nucleophile, narrow substrate
generality in terms of both nucleophiles (only acetate-
derived enolate can be used) and electrophiles, and mod-
erate enantioselectivity are the main drawbacks of this
pioneering reaction. The difficulty in developing this
reaction is partly due to the attenuated reactivity of ke-
tones and the intrinsic facile reversibility of an aldol
reaction to ketones. Thus, development of a catalytic
enantioselective aldol reaction to ketones is among the
most difficult tasks for producing a practical range of
enantioselectivity and substrate generality.
We reported a new general method for catalytic aldol
reaction to ketones using CuF as a catalyst and rela-
tively stable and commonly used ketene trimethylsilyl
acetals as nucleophiles.6,7 In this method, the addition
of a stoichiometric amount of (EtO)3SiF was critical.
Mechanistic studies indicated that a copper enolate gen-
erated through transmetalation between silicon and cop-
per atoms is the actual nucleophile, and (EtO)3SiF
facilitates the rate-determining catalyst turnover step.8
This method has almost completely overcome the reac-
tivity problem of ketones in the aldol reaction, produc-
ing a high chemical yield and reaction rate from a wide
range of ketones and silyl enolates. Thus, this reaction
can be a valuable base for developing a general and
practical catalytic enantioselective aldol reaction to
ketones. In this letter, we describe our preliminary
results of this venue by developing a series of new
chiral bidentate phosphine ligands.
Keywords: Asymmetric catalysis; Aldol reaction; Ketones; Chiral
tertiary alcohol; Chiral phospholane.
*
Corresponding authors. Tel.: +81 358414830; fax: +81 356845206
(M.S.); e-mail: mshibasa@mol.f.u-tokyo.ac.jp
0040-4039/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2005.04.084

4326
K. Oisaki et al. / Tetrahedron Letters 46 (2005) 4325–4329
Various available chiral diphosphine ligands were first
screened in the reaction between acetophenone (1a)
and ketene silyl acetal 2 in the presence of 2.5 mol % Cu-
FÆ3PPh3Æ2EtOH and 5 mol % chiral ligands (Scheme 1).9
Although high chemical yields were obtained in each
case, the maximum enantiomeric excess was only 33%,
even when using privileged chiral bidentate phosphines.
Unexpectedly low enantioselectivity was partly ration-
alized from the reaction mechanism of this CuF-cata-
lyzed aldol reaction. Previous studies suggested that
the reaction proceeds through a linear transition state
from a copper enolate (Fig. 1).10 Thus, a substrate ke-
tone should approach the copper enolate from an out-
side space far from the chiral environment constructed
by chiral bidentate phosphine ligands. As a result,
shielding one specific enantioface of the ketone by steric
hindrance of chiral ligands might be inefficient.
R2
R1
P
O
*
O
Cu
P
MeO
steric bulkiness
Figure 1. Proposed transition state model.
Ph
Ph
P
Ph
P
linker
H
1) BuLi, THF
2)
P
linker
BH3
H3B
BH3
L
L
Ph
4
Ph
Ph
(L = OTs, OMs, Br)
Ph
Ph
P
These considerations led us to design new chiral biden-
tate phosphines containing a deeper chiral pocket, which
can control the orientation of the substrate ketone at a
position far from the metal center in the transition state.
Molecular modeling studies suggested that a diphenyl-
phospholane is a suitable chiral module to construct chi-
ral bidentate phosphines for this purpose. The fact that
low but meaningful enantioselectivity (24% ee) was ob-
tained using Ph-BPE in the preliminary ligand screening
(Scheme 1) supports this idea. We anticipated that if an
appropriate linker to connect the two chiral diphenyl-
phospholanes was identified, the phenyl groups of the
phospholane would act far from the metal center to
where a ketone exists in the transition state, and enantio-
selectivity could be expected to improve. The diphenyl-
linker
P
DABCO
toluene
Ph
Ph
5
Scheme 2. General scheme of bis(diphenylphospholane) synthesis.
The enantioselectivity of these ligands was assessed in
the catalytic aldol reaction of ketone 1a. As expected,
the linker had a profound effect on enantioselectivity
(Table 1). Ligand 5b, which has a propyl linker, pro-
duced higher enantioselectivity than 5a and 5d which
have either an ethyl or a butyl linker (entries 1–4). Con-
straining the propyl linker flexibility by introducing a
methyl substituent (ligand 5c) improved the enantiose-
lectivity (entry 2 vs entry 3). Molecular modeling studies
suggested that ligands 5b and 5c would produce a more
obtuse bite angle than Ph-BPE (5a) when the chelate was
formed with copper metal. Therefore, the phenyl group
of the phospholane should work more efficiently to pro-
duce steric bulkiness at the far position from the copper
center, thus giving higher enantioselectivity using 5b and
5c than 5a. On the other hand, if the linker is longer
phospholane module can be synthesized on
a
multigram-scale according to the reported procedure.11
Linking the two phospholanes was accomplished
through a substitution reaction by the BH3-protected
phospholane followed by deprotection using DABCO
(Scheme 2). Following this general scheme, we synthe-
sized 12 new bidentate chiral phosphines as shown in
Table 1.
1) CuF•3Ph3P•2EtOH (2.5 mol %)
chiral phosphine (5 mol %)
(EtO)3SiF (120 mol %)
THF, 4 ˚C
OH
O
O
OSiMe3
+
Ph
Me
OMe
2) 3HF•Et3N
Ph
Me
OMe
1a
2
3a
iPr
iPr
iPr-DuPHOS
95% (3 h)
15% ee
tol-BINAP
100% (1.5 h)
29% ee
P(p-tol)2
P
P
P(p-tol)2
iPr iPr
Me
O
Ph
P
Ph
Ph
O
O
DM-SEGPHOS
87% (5.5 h)
33% ee
Ph-BPE
92% (1.2 h)
24% ee
P
Me
Me
2
2
P
P
Ph
O
Me
Scheme 1. Preliminary screening of privileged bidentate phosphines.

K. Oisaki et al. / Tetrahedron Letters 46 (2005) 4325–4329
Table 1. Catalytic enantioselective aldol reaction to ketonesa
4327
1) CuF•3Ph3P•2EtOH (2.5 mol %)
chiral phosphine (5 mol %)
(EtO)3SiF (120 mol %)
THF, 4 ºC
OH
O
O
OSiMe3
+
Me
OMe
2) 3HF•Et3N
R
Me
OMe
R
1a: R = Ph
1b: R = c–Hex
2
(S)–3a
Entry
1
Chiral ligand
Time (h)
1.2
Yield (%)b
ee (%)c
R/S
Ph
P
Ph
P
92
24
R
Ph
Ph
5a: Ph-BPE
Ph
P
Ph
P
2
3
24
24
86
53
41
50
S
S
5b
5c
Ph
Ph
Ph
Ph
P
P
Ph
Ph
Ph
Ph
P
P
4
18
24
24
24
86
65
73
83
29
S
R
S
R
Ph
5d
Ph
Ph
Ph
Ph
P
N
5
P
8
Ph
Ph
5e
Ph
Ph
P
6d
4
P
Ph
Ph
Ph
5f
Ph
P
P
7d
3
Fe
Ph
5g
Ph
Ph
P
PPh2
8
9
20
13
87
86
0
0
—
—
Ph
5h
5i
Ph
P
PPh2
Ph
O
O
O
Ph
Ph
P
10
11
18
18
100
95
18
14
S
P
5j
Ph
Ph
O
Ph
Ph
P
R
P
5k
Ph
Ph
(continued on next page)

4328
K. Oisaki et al. / Tetrahedron Letters 46 (2005) 4325–4329
Table 1 (continued)
Entry
Chiral ligand
Time (h)
Yield (%)b
ee (%)c
R/S
Ph
P
Ph
12d
13d,e
8
43
95
78
62
66
S
—
P
5l
Ph
Ph
Ph
P
Ph
P
14d
13
91
60
S
5m
Ph
Ph
a Substrate 1a was used unless otherwise mentioned.
b Isolated yield.
c Determined by chiral HPLC after conversion to the corresponding 3,5-dinitrobenzoate.
d DME was used as solvent. Reactions in DME generally gave 3–4% higher ee than in THF.
e Substrate 1b was used.
than propyl without any constraining group (ligand 5d),
the chelation should become less stable due to entropy
factor. As a result, a less enantioselective pathway
catalyzed by monodentate coordinated copper should
compete with the desired pathway catalyzed by the
bidentate coordinated copper, thus giving the product
with lower enantioselectivity in the case of 5d than
5b.12 Consistent with these considerations, ligands with
more flexible linkers (entries 5–7) produced only very
low enantioselectivity. In addition, no enantioselectivity
was observed using C1-symmetric bidentate phosphines
(entries 8 and 9), which indicated that both of the chiral
phospholanes are essential for enantioinduction.
To summarize, we established a new ligand design con-
cept for a CuF-catalyzed enantioselective aldol reaction
to ketones. A modular approach was used to identify
new chiral bidentate diphenylphospholane ligands 5l
and 5m, which produced up to 66% ee. The linker moi-
ety connecting the two diphenylphospholane modules
had a profound effect on enantioselectivity. This linker
effect was rationalized by the generation of a deep chiral
pocket around the catalytically active copper metal, as
well as the formation of a stable chelation to copper.
These factors are essential for constructing an effective
chiral environment to control the orientation of the sub-
strate ketone in the linear transition state of this cata-
lytic enantioselective aldol reaction. Although
enantioselectivity is still moderate, the applicability to
both aromatic and aliphatic ketones and the use of a rela-
tively stable ketene trimethylsilyl acetal are significant
advantages to the previous example, considering the for-
midable difficulty and high importance of the target
reaction.17 The results described here demonstrated that
a rational design of chiral bidentate phosphine ligands is
possible toward a general and practical catalytic enan-
tioselective aldol reaction to ketones. Studies toward
this goal are ongoing. In addition, the unique chiral
environment constructed by the newly developed ligands
5l and 5m should be useful for other transition metal-
catalyzed organic reactions.
The rational design concept based on these investigations
is to construct a C2-symmetric bidentate chiral phos-
phine with a wide bite angle and stable chelation ability.
Thus, ligands 5j–m were synthesized and their enantio-
selectivity was assessed (Table 1, entries 10–14). Gratify-
ingly, the best enantioselectivity (62% ee) was obtained
using ligand 5l having cis-2-butene linker (entry
12).13,14 A comparable result was also obtained using li-
gand 5m having cyclopropyl linker (entry 14). These link-
ers should define the positions of the steric hindrance (the
phenyl groups of the phospholane) far from the metal
center while maintaining stable bidentate coordination
to the copper atom. The predominant effect of the linker
structure on the enantioselectivity was also demon-
strated using ligands 5j and 5k. Although enantioselec-
tivity was much lower than when using 5l and 5m, the
absolute configuration of the aldol product depended
on the chirality at the linker part, and not on the chirality
of the phospholane (entries 10 and 11).
Acknowledgements
Financial support was provided by PRESTO of Japan
Science and Technology Agency (JST) and Grant-in-
Aid for Specially Promoted Research of MEXT. K.O.
and Y.S. thank JSPS for research fellowships.
Finally, the best ligand 5l also produced appreciable
enantioselectivity (66% ee) from an aliphatic ketone 1b
(Table 1, entry 13).15 Although enantioselectivity is still
in the moderate range, this result is significant if com-
pared with the results obtained under the following rep-
resentative conditions; the aldol product was obtained
with only 32% ee under Denmarkꢀs conditions,5 and
with 12% ee using tol-BINAP as a ligand under CuF-
catalyzed reactions. Thus, this is the most enantioselec-
tive catalytic aldol reaction reported to date using ali-
phatic ketones.16
References and notes
´
1. (a) Ramon, D. J.; Yus, M. Angew. Chem., Int. Ed. 2004,
43, 284–287; (b) Corey, E. J.; Guzman-Perez, A. Angew.
´
´
Chem., Int. Ed. 1998, 37, 388–415.
2. For practical examples, see: (a) Hamashima, Y.; Kanai,
M.; Shibasaki, M. J. Am. Chem. Soc. 2000, 122, 7412–
7413; (b) Yabu, K.; Masumoto, S.; Yamasaki, S.; Hama-

K. Oisaki et al. / Tetrahedron Letters 46 (2005) 4325–4329
4329
shima, Y.; Kanai, M.; Du, W.; Curran, D. P.; Shibasaki,
M. J. Am. Chem. Soc. 2001, 123, 9908–9909; (c) Tian,
S.-K.; Deng, L. J. Am. Chem. Soc. 2001, 123, 6195–6196;
(d) Tian, S.-K.; Deng, L. J. Am. Chem. Soc. 2003, 125,
9900–9901; (e) Deng, H.; Isler, M. P.; Snapper, M. L.;
Hoveyda, A. H. Angew. Chem., Int. Ed. 2002, 41, 1009–
1012. For other examples, see references cited therein.
11. (a) Guillen, F.; Fiaud, J.-C. Tetrahedron Lett. 1999, 40,
2939–2942; (b) Guillen, F.; Rivard, M.; Toffano, M.;
Legros, J.-Y.; Daran, J.-C.; Fiaud, J.-C. Tetrahedron 2002,
58, 5895–5904; (c) Pilkington, C. J.; Zanotti-Gerosa, A.
Org. Lett. 2003, 5, 1273–1275.
12. There was no significant difference in reactivity between
catalysts coordinated by a monodentate phosphine and a
bidentate phosphine.
´
3. For practical examples, (a) see: Garcıa, C.; LaRochelle, L.
K.; Walsh, P. J. J. Am. Chem. Soc. 2002, 124, 10970–
10971; (b) Yus, M.; Ramon, D. J.; Prieto, O. Tetrahedron:
Asymmetry 2003, 14, 1103–1114; (c) Jeon, S.-J.; Walsh, P.
J. J. Am. Chem. Soc. 2003, 125, 9544–9545; (d) Li, H.;
Walsh, P. J. J. Am. Chem. Soc. 2004, 126, 6538–6539; (e)
13. A triphenylphosphine-free catalyst generated via in situ
reduction of CuF2Æ2H2O with the chiral bidentate phos-
phine (see Ref. 4) produced the same enantioselectivity as
the catalyst prepared by simply mixing CuFÆ3PPh3Æ2EtOH
and 5l. Therefore, the competitive coordination of triphen-
ylphosphine to generate achiral CuF was negligible when
5l was used as a chiral ligand.
´
Betancort, J. M.; Garcıa, C.; Walsh, P. J. Synlett 2004,
749–760. For other examples, see references cited therein.
4. For a practical example, see: Wada, R.; Oisaki, K.; Kanai,
M.; Shibasaki, M. J. Am. Chem. Soc. 2004, 126, 8910–
8911. For other examples, see references cited therein.
5. Denmark, S. E.; Fan, Y. J. Am. Chem. Soc. 2002, 124,
4233–4235.
6. Oisaki, K.; Suto, Y.; Kanai, M.; Shibasaki, M. J. Am.
Chem. Soc. 2003, 125, 5644–5645.
7. For metal fluoride-catalyzed asymmetric aldol reactions to
14. Representative spectroscopic data of 5l: 1H NMR
(500 MHz, CDCl3) d: 1.46–1.54 (m, 2H), 1.85–1.95 (m,
2H), 2.00–2.08 (m, 2H), 2.18–2.28 (m, 2H), 2.35–2.45
(m, 2H), 2.50–2.60 (m, 2H), 3.23–3.31 (m, 2H), 3.65–3.75
(m, 2H), 5.15–5.21 (m, 2H), 7.14–7.34 (m, 20H); 13C NMR
(126 MHz, CDCl3) d: 23.93, 24.13, 31.63, 37.50, 46.39,
46.49, 48.46, 48.57, 125.55, 125.63, 125.67, 127.27, 127.81,
127.85, 127.88, 128.30, 128.32, 138.88, 144.83; 31P NMR
(202 MHz, CDCl3) d: 12.0 (s).
15. General procedure for catalytic enantioselective aldol
reaction to ketones: A solution of CuFÆ3PPh3Æ2EtOH
(9.6 mg, 0.010 mmol, 2.5 mol %) and chiral phosphine
ligand (0.020 mmol, 5 mol %) in degassed DME (0.6 mL)
was stirred at room temperature for 30 min. To this
solution, ketone (0.40 mmol) and (EtO)3SiF (0.48 mmol,
1.2 equiv) were added at room temperature and stirred for
15 min. Ketene silyl acetal 2 (0.80 mmol, 2 equiv) was then
added at 0–4 °C, and the mixture was stirred for the
indicated period shown in Table 1. 3HFÆEt3N (0.3 mL)
was added to quench the reaction, and the mixture was
stirred at room temperature for 30 min. After the addition
of satd NaCl, the product was extracted with AcOEt, and
the combined organic layer was washed with brine. Drying
with Na2SO4, filtration, concentration, and purification by
silica gel column chromatography (AcOEt/hexane) gave
the aldol product.
aldehydes, see: (a) Kruger, J.; Carreira, E. M. J. Am.
¨
Chem. Soc. 1998, 120, 867–868; (b) Pagenkopf, B. L.;
Kruger, J.; Stojanovic, A.; Carreira, E. M. Angew. Chem.,
¨
Int. Ed. 1998, 37, 3124–3126; (c) Yanagisawa, A.; Nakat-
suka, Y.; Asakawa, K.; Kageyama, H.; Yamamoto, H.
Synlett 2001, 69–72; (d) Yanagisawa, A.; Nakatsuka, Y.;
Asakawa, K.; Wadamoto, M.; Kageyama, H.; Yama-
moto, H. Bull. Chem. Soc. Jpn. 2001, 74, 1477–1484; (e)
Wadamoto, M.; Ozasa, N.; Yanagisawa, A.; Yamamoto,
H. J. Org. Chem. 2003, 68, 5593–5601.
8. The following experimental results support these ideas. (1)
The enantioselectivity was independent of the alkoxide
substituents of the silicon atom. Both ketene triethoxysilyl
acetal and ketene trimethoxysilyl acetal produced the same
enantioselectivity, which suggested that the silicon is not
relevant to the enantio-differentiation step (aldol addition
to substrate ketones). (2) The order dependencies of the
initial reaction rate on [ketone], [catalyst], and [silyl
enolate] were determined to be 0, 1.5, and –0.8, respec-
tively, which indicated that the addition step is not the
rate-determining step (see Ref. 6).
16. The previous best result for aliphatic ketones was
only 35% ee from 4-phenylbutan-2-one in Denmarkꢀs
catalysis.
9. Chiral amine ligands such as pybox and salen, and
monophosphines such as MOP and Feringaꢀs phospho-
ramidite produced only very low reactivity and/or no
enantioselectivity.
10. Both E- and Z-silyl enolates produced the syn-isomer with
low diastereoselectivity (syn:anti = 1.6:1), which suggested
a linear transition state (see Ref. 6).
17. Other ketene silyl acetals than 2 can be utilized for the
present catalytic enantioselective aldol reaction to ketones.
The enantio-induction at the a-position was much more
efficient (up to 90% ee was obtained using tol-BINAP;
unpublished result) than the tertiary alcohol construction
at the b-position. See Ref. 6 for preliminary results of this
type of asymmetric reaction.