Catalytic enantioselective cyanosilylation of ketones: Improvement of enantioselectivity and catalyst turn-over by ligand tuning

Article abstract of DOI:10.1016/S0040-4039(00)02039-6

Sterical and electronical tuning of the bifunctional catalyst 1 afforded the improved catalyst 2, which promoted the cyanosilylation of ketones with higher enantioselectivity as well as with improved catalyst turn-over with a factor of up to 10. Thus, chiral quaternary α-hydroxynitriles were obtained with excellent ee (up to 94% ee) using 1 mol% of 2 in the case of aryl ketones and 2.5 mol% of 2 in the case of aliphatic ketones.

Full text of DOI:10.1016/S0040-4039(00)02039-6

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 42 (2001) 691–694
Catalytic enantioselective cyanosilylation of ketones:
improvement of enantioselectivity and catalyst turn-over by
ligand tuning
Yoshitaka Hamashima, Motomu Kanai and Masakatsu Shibasaki*
Graduate School of Pharmaceutical Sciences, The University of Tokyo, Hongo, Bunkyo-ku, Tokyo 113-0033, Japan
Received 10 October 2000; revised 7 November 2000; accepted 10 November 2000
Abstract—Sterical and electronical tuning of the bifunctional catalyst 1 afforded the improved catalyst 2, which promoted the
cyanosilylation of ketones with higher enantioselectivity as well as with improved catalyst turn-over with a factor of up to 10.
Thus, chiral quaternary a-hydroxynitriles were obtained with excellent ee (up to 94% ee) using 1 mol% of 2 in the case of aryl
ketones and 2.5 mol% of 2 in the case of aliphatic ketones. © 2001 Elsevier Science Ltd. All rights reserved.
We have recently disclosed the first general catalytic
enantioselective cyanosilylation of ketones promoted by
the novel bifunctional catalyst 1.1,2 Chiral quaternary
a-hydroxynitriles were obtained with excellent selectivi-
ties from a wide variety of ketones. From mechanistic
studies, we proposed the working model for the transi-
tion state as shown in Fig. 1: the Lewis acid (Ti) and
the Lewis base (phosphine oxide) activate the ketone
and TMSCN, respectively, at defined positions.3 How-
ever, there still remained ‘difficult substrates’ that gave
only good to moderate ee. For example, 2-heptanone
8a and 1-indanone 8b gave the corresponding products
with 76 and 69% ee, respectively. Relatively high cata-
lyst loading (10 mol%) was also necessary to promote
the reaction efficiently. During the course of our pursuit
for truly practical asymmetric catalysts, we planned to
improve these points by tuning the ligand structure.
Herein, we report the new catalyst 2 containing a
benzoyl substituent on the catechol moiety, which is a
greatly improved catalyst in terms of enantioselectivity,
substrate scope and catalyst turn-over.
For the highly enantioselective dual activation pathway
to take place, it is essential that the ketone coordinates
to Ti at the position syn to the phosphine oxide (site A,
Fig. 1). The catechol moiety at C-3 should differentiate
the coordination sites A and B, sterically shielding site
B. We expected that introducing a bulky group to this
catechol moiety should hinder site B more effectively,
making the desired coordination of the ketone at site A
more favored. Furthermore, if the substituent X is an
electron-withdrawing group, the Tiꢀphenoxide bond
should become stronger, thus giving rise to the more
stable catalyst complex, which should be advantageous
CN
Me3
Si
RL
Ph
Ph
O
Ph
Ph
(site A)
P
O
Rs
O
O
P
3
O
O
O
O
TMSO CN
RL
1 or 2
iPrO
O
+ TMSCN
O
Ti
Ti
RS
RS
RL
X
THF
O
NC
O
O
X
CN
(site B)
1: X = H
2: X = COPh
Figure 1. Working transition state model.
Keywords: bifunctional asymmetric catalyst; Lewis acid; Lewis base; cyanosilylation; ketone.
* Corresponding author.
0040-4039/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(00)02039-6

692
Y. Hamashima et al. / Tetrahedron Letters 42 (2001) 691–694
OTBS
O
O
O
O
F
4
O
O
MeO
Ph
O
TBSO
Ph
O
O
b, c, d, e
Cr(CO)3
O
O
Ph
O
a
OH
6
MeO
3
5
MeO
Ph
Ph
O
O
TsO
HO
P
O
O
f, g
h, i, j
O
O
H
O
O
7
MeO
2-L
HO
,
Scheme 1. Reagents and conditions: (a) 4, NaH, THF; I2, 84%; (b) TBAF, THF, 99%; (c) PDC, 4 A MS, CH2Cl2, 94%; (d)
,
PhMgBr, THF, 95%; (e) PDC, 4 A MS, CH2Cl2, 88%; (f) TsOH·H2O, MeOH, 97%; (g) TsCl, py, 83%; (h) Ph2PK (2.2 equiv.),
THF; (i) H2O2, MeOH–H2O, 40% (two steps); (j) LiI, DMF, 150°C, 80%.
for a minimized catalyst loading. From these steric and
electronic reasons, we decided to introduce a benzoyl
group at the para-position of the phenoxide (2-L). The
new ligand 2-L was synthesized as shown in Scheme 1.4
catalyst (Scheme 2, method B).6 Thus, even when 1
mol% of 2 was used, the reaction proceeded at −20°C
for 88 h to give the product in 92% yield with 94% ee
(entry 11). Under the same conditions, 1 mol% of
catalyst 1 gave the product only in 31% yield with 84%
ee even after 130 h (entry 6). Therefore, the new
catalyst 2 has shown an improved enantioselectivity, as
well as an improved catalyst turn-over with a factor of
10.
We first assessed the ability of 2 by the reaction of
2-heptanone 8a, 1-indanone 8b and acetophenone 8c,
using 10 mol% of 2 (Table 1). As we expected, the
enantioselectivity for all the ketones was significantly
improved, and for 8c an ee of 97% could be obtained
(entry 7). Next, we investigated the use of a reduced
With the optimized reaction conditions in hand, we
investigated the generality of this improved catalytic
asymmetric cyanosilylation of ketones. As shown in
Table 2, the cyanohydrins were obtained in good to
excellent yields with excellent ee using 1 mol% of 2 in
the case of aryl ketones and 2.5 mol% of 2 in the case
of aliphatic ketones.7 The absolute configuration of the
products can be explained by the working model
depicted in Fig. 1: Ti and the phosphine oxide activate
the ketone and TMSCN, respectively, which defines the
catalyst loading. When the new catalyst
2 was
employed, the catalyst loading could be reduced to 2.5
mol% without affecting the reaction time, chemical
yield and enantioselectivity (entries 8 and 9).5 However,
when 1 mol% of 2 was used, the reaction became slower
and the product was obtained in only 52% yield after
even 88 h (entry 10). It was found that the reactivity of
the catalyst was improved to a synthetically acceptable
range, slightly modifying the preparation method of the
Table 1. Catalytic asymmetric cyanosilylation of ketones: comparison of 1 and 2
catalyst (mol %) (methoda)
ee (%)
entry
ketone
temp (°C)
yield (%)
time (h)
-50
-50
36
44
88
71
76
86
1
2
O
1 (10) (A)
2 (10) (A)
8a
8b
O
3
4
1 (10) (A)
2 (10) (A)
-40
-40
96
96
72
90
69
84
5
6
1 (10) (A)
1 (1) (B)
2 (10) (A)
2 (5) (A)
2 (2.5) (A)
2 (1) (A)
2 (1) (B)
-30
-20
-30
-30
-30
-20
-20
36
130
44
85
31
76
84
84
52
92
92
84
97
95
96
94
94
O
7
8c
Ph
CH3
8
44
9
48
10
11
88
88
a For the preparation method of the catalyst, see Scheme 2.

Y. Hamashima et al. / Tetrahedron Letters 42 (2001) 691–694
693
method A
TMSCN
(2 eq)
+ Ti(OiPr)4
rt
THF
evaporation
0 °C
ligand
catalyst
catalyst
in toluene
70 °C, 1 h
rt, 1 h
rt, 2 h
method B
TMSCN
(2 eq)
+ Ti(OiPr)4
rt
evaporation
THF
0 °C
ligand
in toluene
70 °C, 1 h
rt, 1 h
rt, 1 h
Scheme 2. Two methods for catalyst preparation.
Table 2. Catalytic enantioselective cyanosilylation of various ketones by 2
2 (mol %) (methoda)
temp/°C
ee/%
entry
yield/%
ketone
time/h
O
R = H
R = Cl
1
2
8c
8d
1 (B)
1 (B)
-20
-25
88
92
92
72
94
90
CH3
R
O
O
3
4
5
8e
8f
1 (B)
2.5 (A)
2.5 (A)
-10
-30
-30
92
70
92
90
91
72
92
93
90
CH3
O
8g
CH3
O
6
8a
2.5 (A)
-45
92
80
82
a For the preparation method of the catalyst, see Scheme 2.
sense of entry of TMSCN to the ketone. The beneficial
effect of the benzoyl substituent of 2 could be
attributed to steric, as well as electronic, factors. Shield-
ing the coordination site B sterically, the position of the
ketone should be defined at site A (Fig. 1). Further-
more, electronically stabilizing the Ti complex, the cata-
lyst 2 should be more stable than 1, thus making it
possible to use lower catalyst loading.
Acknowledgements
Financial support was provided by CREST, The Japan
Science and Technology Corporation (JST), and RFTF
of Japan Society for the Promotion of Science. Y.H.
thanks JSPS Research Fellowships for Young
Scientists.
In conclusion, we have succeeded in improving the
enantioselectivity, generality of the substrates and cata-
lyst turn-over, developing the new tuned catalyst 2
containing a benzoyl group at the catechol moiety. This
contribution clearly demonstrated that a general and
practical asymmetric catalyst for the cyanosilylation of
ketones should be accessible by thorough tuning of 2.
Studies toward this goal including the determination of
the catalyst structure, as well as the application to the
catalytic asymmetric synthesis of lead compounds for
drugs are in progress.
References
1. Hamashima, Y.; Kanai, M.; Shibasaki, M. J. Am. Chem.
Soc. 2000, 122, 7412–7413.
2. For other examples using chemical catalyst, see; Belokon’,
Y. N.; Green, B.; Ikonnikov, N. S.; North, M.; Tararov,
V. I. Tetrahedron Lett. 1999, 40, 8147–8150 and references
cited therein.
3. Enantioselective Lewis acid–Lewis base bifunctional cata-
lysts have been applied to
a variety of reactions:

694
Y. Hamashima et al. / Tetrahedron Letters 42 (2001) 691–694
Hamashima, Y.; Sawada, D.; Kanai, M.; Shibasaki, M. J.
the same conditions as entry 2 of Table 1. Elucidation of
the structural differences between the catalysts depending
on the preparation method is currently under
investigation.
Am. Chem. Soc. 1999, 121, 2641–2642; Kanai, M.;
Hamashima, Y.; Shibasaki, M. Tetrahedron Lett. 2000, 41,
2405–2409; Takamura, M.; Hamashima, Y.; Usuda, H.;
Kanai, M.; Shibasaki, M. Angew. Chem. Int. Ed. 2000, 39,
1650–1652; Takamura, M.; Funabashi, K.; Kanai, M.
7. Representative procedure. (a) Using catalyst preparation
method A: Ti(OiPr)4 (11 mL, 0.0372 mmol) and 2-L (20
mg, 0.0378 mmol) were mixed in toluene (0.3 mL), and the
mixture was stirred at 70°C for 1 h. After removing
toluene under reduced pressure, the resulting yellow solid
was dried in vacuo for 1 h. The residue was dissolved in
THF (0.2 mL) and treated with TMSCN (10 mL, 0.0744
mmol) at rt for 1 h. To this solution cyclohexylmethylke-
tone 8f (200 mL, 1.49 mmol) was added at −30°C, followed
by the addition of TMSCN (300 mL, 2.23 mmol). After 70
h, pyridine (0.1 mL) and water (1.0 mL) were successively
added for quenching. Usual workup and purification by
silica gel column chromatography afforded the corre-
sponding product (90%, 93% ee) as a colorless oil; (b)
using catalyst preparation method B: Ti(OiPr)4 (11 mL,
0.0372 mmol) and 2-L (20 mg, 0.0378 mmol) were mixed
in toluene (0.3 mL), and the mixture was stirred at 70°C
for 1 h. Then, TMSCN (10 mL, 0.0744 mmol) was added
at 0°C. After stirring at rt for 1 h, toluene was evaporated
under reduced pressure. The residue was further dried in
vacuo for 1 h. The resulting yellow solid was dissolved in
THF (0.2 mL). To this solution, acetophenone 8c (434 mL,
3.72 mmol) and TMSCN (740 mL, 5.58 mmol) were succes-
sively added at −20°C. After 88 h, the reaction was
quenched. Further purification was performed by silica gel
column chromatography to give the product (92%, 94% ee)
as a colorless oil. For determination of ee and the absolute
configuration of the products, see Ref. 1.
Shibasaki, M. J. Am. Chem. Soc. 2000, 122, 6327–6328.
1
4. Spectroscopic data for 2-L: H NMR (500 MHz, CDCl3);
l 1.94 (m, 1H), 2.16 (m, 1H), 2.72 (ddd, J=9.45, 15.0, 15.0
Hz, 1H), 2.84 (ddd, J=3.35, 9.45, 15.3 Hz, 1H), 3.23 (m,
1H), 3.38 (ddd, J=3.05, 9.15, 16.5 Hz, 1H), 3.60 (ddd,
J=5.20, 8.90, 11.3 Hz, 1H), 3.74 (dd, J=8.90, 9.20 Hz,
1H), 3.89 (m, 1H), 6.98 (d, J=8.25 Hz, 1H), 7.43–7.80 (m,
18H), 9.73 (s, 1H): 13C NMR (125.65 MHz, CDCl3); l
31.5, 36.0 (d, J=68.2 Hz), 65.3, 74.7, 75.8, 84.8, 116.5,
123.8, 128.0, 128.7, 128.8, 128.8, 128.9, 129.0, 129.0, 129.6,
130.0, 130.5, 130.6, 130.8, 130.9, 131.0, 131.2, 131.7, 132.1,
132.3, 138.3, 145.7, 154.8, 195.1: 31P NMR (202.35 MHz,
CDCl3); l 34.5: [h]2D7 +13.2 (c=2.34, CHCl3).
5. The concentration of the catalyst was adjusted to 0.3 M
when 2.5–10 mol% of the catalyst was used. When 1 mol%
of catalyst was used, however, it was necessary to reduce
the concentration of the catalyst to 0.2 M, since TMSCN
(mp 11–12°C) froze out from THF solution. Lower con-
centration of the catalyst and reagents should be the
reason for the slower reaction rate when 1 mol% of
catalyst was used (Table 1, entries 6, 10 and 11).
6. However, the catalyst prepared by method B showed
higher reaction rate only in the case of aryl ketones. In the
case of aliphatic ketones, the catalyst prepared by method
A was more reactive. For example, in the case of 2-hep-
tanone 8a, catalyst 2 (10 mol%) prepared by method B
afforded the product in only 14% yield with 81% ee under
.
.