Highly enantioselective cyanosilylation of aldehydes catalyzed by a Lewis acid-Lewis base bifunctional catalyst

Article abstract of DOI:10.1016/S0040-4020(00)01039-5

A new bifunctional asymmetric catalyst containing a Lewis acid and a Lewis base (1) was developed and applied to the catalytic asymmetric cyanosilylation of aldehydes. The products were obtained generally with excellent enantiomeric excess. The experiment

Full text of DOI:10.1016/S0040-4020(00)01039-5

TETRAHEDRON
Pergamon
Tetrahedron 57 (2001) 805±814
Highly enantioselective cyanosilylation of aldehydes catalyzed by
a Lewis acid±Lewis base bifunctional catalyst
Yoshitaka Hamashima, Daisuke Sawada, Hiroyuki Nogami, Motomu Kanai and
Masakatsu Shibasakip
Graduate School of Pharmaceutical Sciences, The University of Tokyo, Hongo, Bunkyo-ku, Tokyo 113-0033, Japan
Received 12 May 2000; accepted 7 August 2000
AbstractÐA new bifunctional asymmetric catalyst containing a Lewis acid and a Lewis base (1) was developed and applied to the catalytic
asymmetric cyanosilylation of aldehydes. The products were obtained generally with excellent enantiomeric excess. The experiments using
the control catalyst (5) and the catalyst containing more electron-rich phosphine oxide (6) suggest that the catalyst 1 should promote the
reaction via a dual activation of the aldehyde by the aluminum and TMSCN by the phosphine oxide. This reaction is practical and was
applied to the catalytic asymmetric total synthesis of epothilone A. q 2001 Elsevier Science Ltd. All rights reserved.
1. Introduction
1.1. Design and synthesis of Lewis acid±Lewis base
bifunctional catalyst
It is well-established that the addition of TMSCN to
aldehydes is catalyzed by Lewis acids1 as well as Lewis
bases,2 to afford trimethylsilylated cyanohydrins. Cyano-
hydrins are highly versatile synthetic intermediates, which
can easily be converted to various important building blocks
including a-hydroxy carbonyl derivatives and b-amino
alcohols. In view of their importance, there is currently
considerable focus in developing methods for the asym-
metric synthesis of cyanohydrins, especially using chiral
catalysts.3,4 Although impressive enantiomeric excesses
have been obtained in some cases, it is highly desired to
develop a more general asymmetric catalyst, applicable to a
wide variety of aldehydes. During the course of our study to
develop an asymmetric catalyst from the concept of multi-
functional catalysis,5 we have reported Lewis acid±
Brùnsted base bifunctional asymmetric catalysts, which
have been applied to a wide range of reactions.6 Inspired
by the concept, we expected that a Lewis acid±Lewis base
bifunctional catalyst should simultaneously activate
TMSCN and an aldehyde at de®ned positions in the catalyst,
thus providing a more general asymmetric catalyst for the
cyanosilylation reaction of aldehydes. We report herein that
catalyst 1 is a highly ef®cient catalyst for the cyanosilylation
of aldehydes with broad generality, affording products in
excellent chemical yields and excellent enantioselectivities.7
We selected BINOL as a scaffold for arranging the Lewis
acid and Lewis base moieties as shown in Fig. 1. When the
Lewis acid metal is connected to the two naphthols, it was
anticipated that the Lewis base moieties should be
connected to 3,30-position of BINOL to promote the
reaction ef®ciently via a dual activation pathway. The
following two points should be important for constructing
a successful bifunctional catalyst: (1) The activation ability
of the Lewis acid and Lewis base moieties toward an
aldehyde and TMSCN should be balanced to promote the
reaction via a dual activation pathway. (2) The internal
coordination of the Lewis base to the metal should be
avoided. The design shown in Fig. 1 should be very ¯exible
to optimize these points by changing the metal, the Lewis
base and the linker length connecting the Lewis base and
BINOL. The phosphine oxide ligand 1-L, which ®nally
turned out to be the best ligand, was synthesized in high
yield from MOM-BINOL 7 as shown in Scheme 1.
1.2. Optimization of the catalyst
The combination of the Lewis acid metal (Al, Ga, Ti or Zr)
Keywords: Lewis acid; Lewis base; asymmetric catalyst; cyanosilylation;
epothilones.
* Corresponding author. Tel.: 181-3-5684-0651; fax: 181-3-5684-5206;
e-mail: mshibasa@mol.f.u-tokyo.ac.jp
Figure 1. Design of Lewis acid±Lewis base catalysts.
0040±4020/01/$ - see front matter q 2001 Elsevier Science Ltd. All rights reserved.
PII: S0040-4020(00)01039-5

806
Y. Hamashima et al. / Tetrahedron 57 (2001) 805±814
Scheme 1. Synthesis of 1-L.
and the Lewis base (phosphorus, sulfur or phosphine oxide
oxygen) was ®rst optimized by the reaction of TMSCN with
benzaldehyde, changing these two moieties independently
(Table 1). It was found that the combination of aluminum
chloride as a Lewis acid and a phosphine oxide as a Lewis
base afforded the best catalyst in terms of reactivity as well
as enantioselectivity and the product was obtained with 87%
ee and 91% yield at 2408C for 37 h (entry 1). It is also
important to note that ligands 2-L and 3-L were partially
silylated during the reaction. This indicates that the
chiral catalysts derived from 2-L and 3-L were partially
decomposed under the reaction conditions, thus giving
lower reactivity and lower enantioselectivity. On the other
hand, no silylation of the ligand 1-L was observed under any
combination with Lewis acid metals.
oxide is at the de®ned position close to the activated alde-
hyde. This situation should be important to promote the
reaction via a dual activation by a Lewis acid and a Lewis
base.
Next, we investigated the effect of the relative position of
the Lewis acid (Al) and the Lewis base (the oxygen atom of
the phosphine oxide) by changing the linker length between
the phosphine oxide and the scaffolding BINOL (1 vs 4).
Molecular modeling studies suggested that the coordination
of the Lewis base to the internal aluminum seemed to be
torsionally unfavorable in the case of 1. When considering
4, however, which has an ethylene linker, the internal coor-
dination seemed to be quite stable without strain. Therefore,
in the case of the catalyst 4, strong intramolecular coordi-
nation of the phosphine oxide should reduce the Lewis
acidity of the aluminum, therefore diminishing the catalytic
ef®ciency of 4. In accordance with this expectation, the
reaction of TMSCN with benzaldehyde, catalyzed by 4
(9 mol%), proceeded slowly at 2408C (37 h) and gave the
cyanohydrin in only 4% yield after hydrolysis. This result
stands in contrast to the much higher reactivity of the cata-
lyst 1 (91% yield under the same conditions, Table 1, entry
1). Therefore, in the case of 1, the intramolecular binding of
the phosphine oxide to the aluminum seems to be labile
enough to allow the coordination of the aldehyde to the
aluminum. Consequently, we investigated the best catalyst
To con®rm the activation ability of a phosphine oxide
toward TMSCN, we performed the reaction of hydrocinnam-
aldehyde with TMSCN in the presence or absence of
Bu3P(O), without the Al catalyst 1. Thus, in the presence
of 40 mol% of Bu3P(O), the product was obtained in 81%
yield at ambient temperature for 7.5 h. In the absence of the
phosphine oxide, the yield was only 12% under the same
conditions. At 2408C for 40 h, however, the reaction did not
proceed at all even in the presence of Bu3P(O). These results
appear to suggest that the internal phosphine oxide of 1
would activate TMSCN as a Lewis base if the phosphine
Table 1. Optimization of the combination of the Lewis acid and the Lewis base
Entry
Metal
Ligand
Solvent
Temp. (8C)
Time (h)
Yield (%)
Ee (%)
Conf.
1
2
3
4
5
6
7
8
Me2AlCl
Me3Al
1-L
1-L
1-L
1-L
1-L
2-L
2-L
2-L
3-L
3-L
3-L
3-L
CH2Cl2
Toluene
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
Toluene
CH2Cl2
Toluene
Toluene
CH2Cl2
240
0
0
37
36
60
46
46
71
36
192
110
60
60
216
91
97
100
trace
100
27
92
75
27
87
33
0
±
3
5
2
29
24
±
S
S
±
±
S
S
S
S
R
±
±
R
Ga(OiPr)3
Ti(OiPr)4
Zr(OiPr)4
Et2AlCl
Me3Al
rt
220
240
0
240
220
240
240
220
Ti(OiPr)4
Et2AlCl
Me3Al
9
10
11
12
trace
trace
80
Ti(OiPr)4
Zr(OiPr)4
±
3

Y. Hamashima et al. / Tetrahedron 57 (2001) 805±814
807
Figure 2. Assumption for the improvement of the enantioselectivity by an additive.
1, in which the aluminum acts as a Lewis acid and the
oxygen atom of the phosphine oxide acts as a Lewis base
with the methylene linker to connect the Lewis base and the
scaffolding BINOL.
Surprisingly, aliphatic aldehydes afforded very low ee's:
hydrocinnamaldehyde 11a gave S-12a in 90% yield and in
9% ee; isobutylaldehyde 11c gave S-12c in 80% yield and in
25% ee. We anticipated that there would be a competition
between two reaction pathways in the case of the more
reactive aliphatic aldehydes. The desired pathway involves
the dual interaction between the Lewis acid and the
aldehyde and between the Lewis base and TMSCN, whereas
the undesired pathway involves mono-activation by the
Lewis acid. We assumed that these two pathways could
differ more signi®cantly if the Lewis acidity of the catalyst
would be decreased, and so we investigated the effect of
additives which coordinate to the aluminum to reduce its
Lewis acidity. Moreover, the additive could change the
geometry of the aluminum from tetrahedral to trigonal
bipyramidal,8 which should allow the phosphine oxide to
get into a more favorable position relative to the aldehyde
(Fig. 2).
A preliminary evidence of the bifunctional catalysis by 1
was obtained by the control experiment using catalyst 5,
which contains only a steric bulkiness (diphenylmethyl
group), but not a Lewis base, at the 3,30-position of
BINOL. Thus, 5 (9 mol%) gave the cyanohydrin 12h of
benzaldehyde 11h in 50% yield and with 12% ee (2408C,
37 h). The absolute con®guration of 12h was the opposite
(R) to that obtained by 1. The reversal of the absolute
con®guration of the products (12a and 12c) was generally
observed, using either 1 or 5. In the case of 5, it would be
reasonable to assume that TMSCN attacks the activated
aldehyde, coordinating to the aluminum, from the less
hindered side (the opposite to the diphenylmethyl group).
Then, in the case of 1, TMSCN appears to attack the
aldehyde from the side of the phosphine oxide. This can
be explained if we assume TMSCN, which is activated by
the phosphine oxide would react with the aldehyde.
After several attempts, we found that electron donating
phosphine oxides had a bene®cial effect on the ee. In the
case of hydrocinnamaldehyde 11a, the ee values of 12a
signi®cantly increased from 9% to 41% and 56% by the
addition of 36 mol% of CH3P(O)Ph2 and Bu3P(O) respec-
tively. Further improvement of ee (up to 97%) was achieved
by the slow addition of TMSCN (10 h), via syringe pump, in
the presence of Bu3P(O) (Table 2, entry 6). In the case of
benzaldehyde 11h, however, addition of Bu3P(O) resulted in
1.3. Catalytic asymmetric cyanosilylation of various
aldehydes
Encouraged by the result of benzaldehyde using catalyst 1,
we next investigated the reaction of aliphatic aldehydes.
Table 2. Additive effect
Entry
R
Aldehyde
Product
Additive
none
Ph3P(O)
CH3P(O)(OCH3)2
Ph2P(O)CH3
Bu3P(O)
Time (h)
Yield (%)
Ee (%)
1
2
3
4
PhCH2CH2
PhCH2CH2
PhCH2CH2
PhCH2CH2
PhCH2CH2
PhCH2CH2
PhCH2CH2
(CH3)2CH
(CH3)2CH
(CH3)2CH
Ph
11a
11a
11a
11a
11a
11a
11a
11c
11c
11c
11h
11h
11h
12a
12a
12a
12a
12a
12a
12a
12c
12c
12c
12h
12h
12h
20
36
62
36
36
37
36
36
36
45
36
96
200
90
98
88
78
60
97
67
80
58
96
91
98
trace
9
2
27
41
56
97
40
25
72
90
87
96
±
5
6a
7
Bu3P(O)
(octyl)3P(O)
none
Bu3P(O)
Bu3P(O)
none
8
9
10a
11
12
13
Ph
Ph
Ph2P(O)CH3
Bu3P(O)
a TMSCN was added dropwise over 10 h.

808
Y. Hamashima et al. / Tetrahedron 57 (2001) 805±814
Figure 3. Initial reaction rate in the absence (B) and presence (X) of the additive.
a very sluggish reaction, affording only a trace amount of
the product. However, the reaction proceeded in 98% yield
and in 96% ee in the presence of CH3P(O)Ph2 (entry 12).
Therefore, we used Bu3P(O) as the additive for aliphatic and
a,b-unsaturated aldehydes, and CH3P(O)Ph2 as the additive
for aromatic aldehydes.
pentacoordinated aluminum in the presence of the additive
phosphine oxide. Thus, the additive phosphine oxide should
®nely tune the balance of the Lewis acidity and the Lewis
basicity of the bifunctional catalyst, as well as tuning the
relative position of the activated aldehyde and TMSCN in
the transition state. Since the reaction pathway involving the
activation by the external phosphine oxide is negligible at
2408C, only the internal phosphine oxide can function as an
activator of TMSCN as mentioned above, because the
phosphine oxide exists at the appropriate position close to
TMSCN in the reactive complex.
To con®rm the origin of the bene®cial effect of the additive
phosphine oxide, we measured the initial kinetics of the
reaction of hydrocinnamaldehyde 11a in the absence or
presence of the additive Bu3P(O). The result is shown in
Fig. 3. The initial reaction rate in the presence of Bu3P(O)
was 0.6 times slower than in the absence of Bu3P(O). This
result can be explained from the lower Lewis acidity of the
This catalyst is practical and has a broad generality with
respect to the variety of aldehydes that can be used (Table
Table 3. Catalytic asymmetric cyanosilylation of various aldehydes under the optimized conditions
Entry
R
Aldehyde
Product
Additive
Time (h)
Yield (%)
Ee (%)a
S/R
1
2
3
4
5
6
Ph(CH2)2
CH3(CH2)5
(CH3)2CH
(CH3CH2)2CH
trans-CH3(CH2)3CHvCH2
PhCHvCH
11a
11b
11c
11d
11e
11f
12a
12b
12c
12d
12e
12f
Bu3P(O)
Bu3P(O)
Bu3P(O)
Bu3P(O)
Bu3P(O)
Bu3P(O)
37
58
45
60
58
40
97
100
96
98
94
97
98
90
83
97
98
S
S
S
S
e
±
99
S
S
7b
11g
12g
Bu3P(O)
50
97
99
8c
9
Ph
p-CH3C6H4
11h
11i
12h
12i
CH3P(O)Ph2
CH3P(O)Ph2
96
79
98
87
96
90
S
S
10d
11j
12j
CH3P(O)Ph2
70
86
95
S
a Determined by HPLC after the appropriate conversion. The absolute con®guration was assigned by comparison with optical rotation reported in literature.
b 5 mol% of 1 was used.
c TMSCN was added over 1 min.
d 18 mol% of 1 and 72 mol% of additive were used.
e The absolute con®guration has not been determined yet.

Y. Hamashima et al. / Tetrahedron 57 (2001) 805±814
809
Figure 4. Initial reaction rate using 1 (B) and 6 (X).
3). Speci®cally, the aldehyde 11g was converted to the
cyanohydrin 12g, which is a key intermediate for the
asymmetric synthesis of the antineoplastic natural product
epothilones.9
presence of Bu3P(O) (40 mol%). Furthermore, by the one-
portion addition of TMSCN, catalyst 6 (10 mol%) gave
S-12a in 86% yield and in 68% ee in the presence of
Bu3P(O) (40 mol%) (2408C for 36 h), whereas 1 gave
S-12a in 60% yield and in 56% ee under the same con-
ditions.10 The increased reaction rate and enantioselectivity
by 6 is consistent with the dual activation mechanism of
these catalysts. More electron-rich phosphine oxide should
activate TMSCN more ef®ciently, thus facilitating the
desired dual activation pathway.
1.4. Preliminary mechanistic studies
To elucidate the origin of the excellent enantioselectivity
with wide generality, as a bifunctional catalysis of 1, we
planned kinetic studies comparing the initial reaction rate
using 1 and 6, containing the more electron-rich phosphine
oxide. The initial reaction rate using 6 (10 mol%) was 1.2
times faster than using 1 (10 mol%) (k6/k1ˆ1.2) (Fig. 4),
re¯ecting the higher Lewis basicity of the phosphine
oxide in the reaction of hydrocinnamaldehyde 11a in the
From these mechanistic studies, the enantioselctivity of the
reaction catalyzed by 1 may be explained by the working
model depicted as 14 in Fig. 5, with the external phosphine
oxide coordinating to the aluminum, thus giving a
Figure 5. Proposed catalytic cycle.

810
Y. Hamashima et al. / Tetrahedron 57 (2001) 805±814
Figure 6. Retrosynthetic analysis of epothilones.
pentavalent aluminum. This geometry would allow the alde-
hyde to position itself at the apical site close to the internal
phosphine oxide. TMSCN, interacting with the internal
phosphine oxide, could then transfer cyanide to the alde-
hyde thus giving the observed S-product.11
catalytic asymmetric cyanosilylation of aldehydes. The
products were obtained generally with excellent enantio-
meric excesses. This reaction is the most general catalytic
asymmetric cyanosilylation of aldehydes reported to date.
The absolute con®guration of the products as well as experi-
ments using the control catalyst (5) and the catalyst contain-
ing more electron-rich phosphine oxide (6) suggest that the
catalyst 1 should promote the reaction via a dual activation
of aldehyde by the aluminum and TMSCN by the phosphine
oxide. This reaction is practical and was applied to the cata-
lytic asymmetric total synthesis of epothilone A.
1.5. Application to the catalytic asymmetric total
synthesis of epothilone A
Epothilones (A: 16, B: 17) show a potent antitumor activity
by binding and stabilizing the microtubles in the same way
as taxol, and they are promising drug candidates. Recently,
highly ef®cient total syntheses of epothilones were dis-
closed by Danishefsky, Nicolaou, Schinzer and Shibasaki
et al.12 Our retrosynthesis is shown in Fig. 6. The catalytic
asymmetric cyanosilylation was utilized for the synthesis of
fragment 18. Fine tuning of the standard reaction conditions
was necessary to apply this reaction to a large scale reaction
using the aldehyde 11g. Taking the high functionality of 11g
into account, the cyanosilylation of 11g (70 mg) was initi-
ally performed using 20 mol% of the (S)-1, 80 mol% of
Bu3P(O) and 3 equiv. of TMSCN. Then, the product 12g
was obtained (42 h at 2408C, including 10 h slow addition
of TMSCN) in 95% yield and with 99% ee after hydrolysis
with tri¯uoroacetic acid. However, when 300 mg of 11g was
used, the reaction became much slower and 12g was
obtained in lower yield of 83% (98% ee) after 86 h. Inter-
estingly, when the amount of TMSCN was reduced to
1.8 equiv. from 3 equiv., the reaction proceeded smoothly
again to afford 12g in 97% yield with 99% ee (51 h). It
appeared that keeping the concentration of TMSCN low
enough was important in the case of aldehyde 11g. This
situation is special for the aldehyde 11g (Fig. 6) containing
a thiazole moiety. We anticipate that the excess TMSCN
would be activated by the coordination of the thiazole to Si,
causing deactivation of the catalyst 1 by exchange of the
aluminum chloride to aluminum cyanide.13 Gratifyingly, it
was found that the amount of the catalyst 1 could be reduced
to 5 mol% using 20 mol% of Bu3P(O) and slowly adding
1.2 equiv. of TMSCN for 48 h to give 12g in 97% yield with
99% ee (50 h including the slow addition time). A catalytic
asymmetric total synthesis of epothilone A using the cyano-
silylation as a key step was achieved9 and details will be
reported as a full paper elsewhere.
3. Experimental
3.1. General
NMR spectra were recorded on a JEOL JNM-LA500 spec-
1
trometer, at 500 MHz for H NMR, 125.65 MHz for 13C
NMR, and 202 MHz for 31P NMR. Chemical shifts in
1
CDCl3 were reported down®eld from TMS (ˆ0) for H
NMR. For 13C NMR, chemical shifts were reported in the
scale relative to CHCl3 (77.00 ppm for 13C NMR) as an
internal reference. 31P NMR were carried out with phos-
phinic acid (85%) as an external standard. Optical rotations
were measured on a JASCO P-1010 polarimeter. Column
chromatography were performed with silica gel Merck 60
(230±400 mesh ASTM). The enantiomeric exess (ee) was
determined by HPLC analysis. HPLC was performed on
JASCO HPLC systems consisting of the following: pump,
880-PU or PU-980; detector, 875-UV or UV-970, measured
at 254 nm; column, DAICEL CHIRALPAK AS, AD, or
DAICEL CHIRALCEL OJ, OD; mobile phase, hexane-2-
propanol; ¯ow rate, 0.5±1.0 mL/min. In general, reactions
were carried out in dry solvents under an argon atmosphere,
unless noted otherwise. Tetrahydrofuran (THF) and diethyl
ether were distilled from sodium benzophenone ketyl.
Dichloromethane (CH2Cl2) was distilled from calcium
hydride. Diethylaluminium chloride in hexane (1 M) was
purchased from KANTO CHEMICAL. CO., INC., 2±8,
Nihonbashi, Honcho 3-chome, Chuo-ku, Tokyo, 103-
0023, Japan (fax: 1813-3667-6892). Other reagents were
puri®ed by usual methods.
3.2. Synthesis of (R)-3,30-bis(diphenylphosphinoyl-
methyl)-1,10-binaphthol (1-L)
2. Conclusion
3.2.1. (R)-3,30-Diformyl-2,20-bis(methoxymethy)-1,10-bi-
naphthol (8). To a stirred solution of MOM-protected
(R)-binaphthol 7 (10 g, 26.7 mmol) in Et2O (400 mL) was
A new bifunctional asymmetric catalyst containing a Lewis
acid and a Lewis base (1) was designed and applied to the

Y. Hamashima et al. / Tetrahedron 57 (2001) 805±814
811
added n-BuLi (53.1 mL, 85.5 mmol, 1.61 M in hexane) at
room temperature. The mixture was stirred for 2 h at the
same temperature. After cooling down to 08C, DMF
(7.24 mL, 93.5 mmol) was dropped over 15 min. The
mixture was then warmed up to room temperature and stir-
red for further 2 h. Sat. aq NH4Cl was added to the reaction.
After neutralization, the reaction mixture was extracted with
ethyl acetate several times. The combined organic layer was
washed with water and brine, and then dried over Na2SO4.
Flash column chromatography (hexane/ethyl acetateˆ10:1
to 5:1) was performed carefully after the concentration to
give 8 (9.0 g, 78%) as yellow oil: 1H NMR (CDCl3) d 10.55
(s, 2H), 8.62 (s, 2H), 8.08 (dd, Jˆ8.2, 1.2 Hz, 2H), 7.52
(ddd, Jˆ9.2, 7.05, 1.20 Hz, 2H), 7.43 (ddd, Jˆ8.2, 7.05,
1.50 Hz, 2H), 7.22 (dd, Jˆ9.2, 1.50 Hz, 2H), 4.71 (d,
Jˆ6.1 Hz, 2H), 4.69 (d, Jˆ6.1 Hz, 2H), 2.88 (s, 6H); 13C
NMR (CDCl3) d 190.6, 154.0, 136.7, 132.3, 130.3, 129.6,
128.9, 126.3, 126.1, 125.9, 100.6, 57.0.
3.2.4. (R)-3,30-Bis(diphenylphosphinoylmethyl)-2,20-bis-
(methoxymethyl)-1,10binaphthol (10). To a solution of
diphenylphosphine oxide (2.93 g, 13.3 mmol) in THF
(40 mL) was added NaOtBu (1.4 g, 15 mmol) in THF
(20 mL) at 08C. After stirring for 30 min at room tempera-
ture, the clear solution changed to a white suspension. To
this suspension, 9 (2.5 g, 5.3 mmol) in THF (30 mL) was
added at 2408C and the reaction mixture was gradually
warmed up to room temperature. After the completion of
reaction, sat. aq NH4Cl was added, and the reaction mixture
was concentrated to the half, followed by dilution with ethyl
acetate (80 mL). The organic layer was washed with water
(50 mL£2). Separated aqueous layer was carefully
extracted with ethyl acetate (50 mL£3). The combined
organic layer was washed with brine (50 mL£2). After the
removal of the solvent, further puri®cation was carried out
by ¯ash chromatography (CH2Cl2: MeOHˆ30: 1) to give 10
1
(3.95 g, 93%): H NMR d 8.30 (d, Jˆ2.45 Hz, 2H), 7.92±
7.76 (m, 10H), 7.52±7.41 (m, 12H), 7.35 (m, 2H), 7.15 (m,
2H), 6.85 (d, Jˆ8.55 Hz, 2H), 4.21 (d, Jˆ6.5 Hz, 2H), 4.20
(d, Jˆ6.5 Hz, 2H), 4.15 (dd, Jˆ13.5, 13.5 Hz, 2H), 4.01 (dd,
Jˆ13.5, 13.5, 2H), 2.84 (s, 6H).
3.2.2.
methyl)-1,10-binaphthol. NaBH4 (700 mg, 18.6 mmol)
was added to solution of dialdehyde (4.0 g,
(R)-3,30-Bis(hydroxymethyl)-2,20-bis(methoxy-
a
8
9.29 mmol) in methanol (140 mL) at 08C. After 30 min,
sat. aq NH4Cl was added and the reaction mixture was half-
way concentrated. The residue was diluted with ethyl
acetate (50 mL), and the organic layer was washed with
water. The water layer was extracted with ethyl acetate
(60 mL£4) and combined organic layer was washed with
brine (50 mL£1). The further puri®cation was performed by
¯ash chromatography on silica gel (hexane/acetoneˆ3:1) to
afford the diol in 92.6% as a white solid; IR (KBr) n 3387,
3.2.5. (R)-3,30-Bis(diphenylphosphinoyl)-1,10-binaphthol
(1-L). 10 (3.8 g, 4.73 mmol) was dissolved in the mixture of
methanol/CH2Cl2 (40 mL/40 mL). A catalytic amount of
TsOH (monohydrate) was added and stirred at 408C over
night. Most of the solvent was removed and the residue was
diluted with ethyl acetate. The solvent was washed with
water (30 mL£2) and the aqueous layer was extracted
with ethyl acetate (80 mL£3). The combined organic layer
was washed with brine and dried over Na2SO4. The crude
product was further puri®ed by recrystallization (CH2Cl2±
Et2O) to give 1-L (3.0 g, 89%): IR (KBr) n 3433, 1437,
1
2936, 1622, 1596 cm21; H NMR (CDCl3) d 8.02 (s, 2H),
7.91(m, 2H), 7.43 (ddd, Jˆ 8.25, 6.7, 0.9 Hz, 2H), 7.27(ddd,
Jˆ8.25, 7.0, 1.5 Hz, 2H), 7.15 (m, 2H), 4.98 (d, Jˆ12.8 Hz,
2H), 4.85 (d, Jˆ12.8 Hz, 2H), 4.48 (d, Jˆ6.1 Hz, 2H), 4.45
(d, Jˆ6.1 Hz, 2H), 3.12 (s, 6H); 13C NMR (CDCl3) d 153.1,
134.6, 133.7, 130.9, 129.7, 128.2, 126.8, 125.7, 125.4,
125.2, 99.3, 61.9, 57.1;MS m/z 434 (M1); Anal. Calcd for
C26H26O6: C, 71.87; H, 6.03. Found: C, 71.51; H, 6.20.
1
1160, 746, 724 cm21; H NMR (CDCl3) d 7.80±7.67 (m,
12H), 7.53±7.41 (m, 12H), 7.24 (m, 2H), 7.14 (t, Jˆ
7.65 Hz, 2H), 6.88 (d, Jˆ8.55 Hz, 2H), 4.06 (dd, Jˆ14.4,
14.4 Hz, 2H), 3.92 (dd, Jˆ14.4, 14.4 Hz, 2H); 13C NMR
(CDCl3) d 151.4, 151.3, 133.2, 132.1 (d, Jˆ3.13 Hz),
132.1 (d, Jˆ2.13 Hz), 131.9, 131.7, 131.5, 131.4, 131.1
(d, Jˆ7.13 Hz), 131.0 (d, Jˆ7.13 Hz), 130.9, 129.0 (d, Jˆ
2.0 Hz), 128.7 (d, Jˆ3.13 Hz), 128.6 (d, Jˆ2.0 Hz), 127.7,
126.4, 124.8, 123.6, 121.6, 121.5, 117.1, 34.0 (d, Jˆ
66.8 Hz); 31P NMR (CDCl3) d 37.94; MS m/z 714 (M1).
3.2.3. (R)-3,30-Bis(chloromethyl)-2,20-bis(methoxymethyl)-
1,10-binaphthol (9). A toluene (60 mL) solution of
bis(hydroxymethyl)binaphthol (3.64 g, 8.38 mmol) was
cooled down to 08C. To this solution, MsCl (3.24 mL,
41.9 mmol) and Et3N (8.17 mL, 58.6 mmol) were added
successively. After 1 h, the reaction mixture was treated
with LiCl (1.78 g, 41.9 mmol) and DMF (60 mL) and stirred
at room temperature until the starting material disappeared.
The mixture was washed with water (20 mL£2). The sepa-
rated aqueous layer was extracted with ethyl acetate
(80 mL£3). The combined organic layer was washed with
brine (80 mL), and then dried over Na2SO4. Solvent was
evaporated under reduced pressure, and the residue was
chromatographed (hexane/acetoneˆ4:1) to give 9 (3.76 g,
95%): IR (neat) n 1622, 1597, 752 cm21; 1H NMR (CDCl3)
d 8.11 (s, 2H), 7.90 (d, Jˆ8.25 Hz, 2H), 7.43 (m, 2H), 7.28
(m, 2H), 7.17 (d, Jˆ8.5 Hz, 2H), 4.99 (d, Jˆ11.9 Hz, 2H),
4.94 (d, Jˆ 11.9 Hz, 2H), 4.63, (dd, Jˆ5.5, 0.6 Hz, 2H),
4.52 (dd, Jˆ 5.5, 0.6 Hz, 2H), 2.97 (s, 6H); 13C NMR
(CDCl3) d 152.2, 134.2, 131.2, 130.9, 130.6, 128.1, 127.2,
126.0, 125.5, 99.5, 56.9, 42.3; MS m/z 470 (M1), 472
(M112), 474 (M114); Anal. Calcd for C26H24O4: C,
66.25; H,5.13. Found: C, 65.98; H, 5.28.
31P NMR spectrum is concentration-dependent; [a]23
ˆ
D
1131.58
(cˆ1.0,
CH3OH);
Anal.
Calcd
for
C46H36O4P2´0.67H2O: C, 76.02; H, 5.27. Found: C, 76.02;
H, 5.10.
3.2.6. (R)-3,30-Bis(bis(p-N,N-dimethylaminophenyl)phos-
phinoylmethyl)-2,20-bis(methoxymethyl)-1,10-binaphthol.
To a solution of bis(N,N-dimethylaminophenyl)phosphine
oxide (1.84 g, 6.36 mmol) in THF (120 mL) was added
NaOtBu (0.96 g, 7.14 mmol) in THF (15 mL) at 08C. This
solution was allowed to stir for 30 min at room temperature.
Again, the reaction mixture was cooled down to 08C, and 9
(1.2 g, 2.55 mmol) in THF (4 mL) was added at the same
temperature, followed by stirring at ambient temperature
until the reaction completed. To this reaction mixture, sat.
aq NH4Cl was added to quench. Keeping the mixture pH 8±9
with sat. aq NaHCO3, the same work-up procedure in the case
of 10 was performed. Further puri®cation was carried out by
¯ash chromatography (CH2Cl2/MeOHˆ40:1,20:1) to give

812
Y. Hamashima et al. / Tetrahedron 57 (2001) 805±814
1
the product (1.34 g, 52.4%) as the white powder: H NMR
(CDCl3) d 8.3 (s, 2H), 7.8 (d, Jˆ8.25 Hz, 2H), 7.68 (dd,
Jˆ8.85, 10.7 Hz, 4H), 7.57 (dd, Jˆ8.85, 10.7 Hz, 4H), 7.31
(dt, Jˆ0.95, 6.73 Hz, 2H), 7.10 (t, Jˆ7.35 Hz, 2H), 6.92 (d,
Jˆ8.55 Hz, 2H), 6.70 (dd, Jˆ2.1, 8.85 Hz, 4H), 6.64 (dd,
Jˆ2.1, 8.85 Hz, 4H), 4.25 (d, Jˆ11 Hz, 2H), 4.23 (d,
Jˆ11 Hz, 2H), 4.01 (t, Jˆ15 Hz, 2H), 3.89 (t, Jˆ15 Hz,
2H), 2.96 (s, 24H), 2.78(s, 6H); 13C NMR (CDCl3) d 152.5
(d, Jˆ15 Hz), 152 (d, Jˆ9.3 Hz), 152 (d, Jˆ 9.3 Hz), 133,
132.4 (d, Jˆ10 Hz), 132.2 (d, Jˆ10 Hz), 131.3, 131.2, 130.6,
130.6, 128, 127 (d, Jˆ6.3 Hz), 126 (d, Jˆ4.1 Hz), 125.3,
124.7, 111.3 (d, Jˆ10 Hz), 111.2 (d, Jˆ11 Hz), 40, 40,
31.5 (d, Jˆ68 Hz); 31P NMR (CDCl3) d 35.2 ppm.
mixture was allowed to stir for the time shown in Table 1
at the same temperature. 2N HCl (1.0 mL) was added, and
the mixture was stirred vigorously at room temperature for
1 h to hydrolyze the trimethylsilylether of the product. After
the addition of ethyl acetate (3.0 mL), the mixture was
stirred for further 30 min. The organic layer was separated
and washed with water. The aqueous layer was extracted
with ethyl acetate (20 mL£2). The combined organic layer
was washed with brine and dried over Na2SO4. The crude
product was further puri®ed by ¯ash chromatography
(hexane/ethyl acetateˆ10:1) to give
a cyanohydrin
(12a±12j) in more than 86%. The enantiomeric excesses
of the products were determined after the conversion to
acetylester, benzoylester, p-nitrobenzoylester, ethylcarbo-
nate, and t-butyldimethylsilylether by usual methods.
3.2.7. (R)-3,30-Bis(bis(p-N,N-dimethylaminophenyl)phos-
phinoylmethyl)-1,10-binaphthol (6-L). The MOM protected
6-L (1.34 g, 1.38 mmol) was dissolved in the mixture of
methanol/CH2Cl2 (12 mL/12 mL). A catalytic amount of
TsOH (monohydrate) was added and the mixture was stirred
at 408C over night. Most of the solvent was removed and the
residue was diluted with ethyl acetate (10 mL). The solvent
was washed with sat. aq NaHCO3 (10 mL£2) and the
aqueous layer was extracted with ethyl acetate (10 mL£3).
The combined organic layer was washed with brine and dried
over Na2SO4. The crude product was further puri®ed by
recrystallization to give 6-L (0.8 g, 66%): IR (KBr) n 3518,
3.4.1. (2S)-2-Hydroxy-4-phenylbutanenitrile (12a). This
product was found to be identical with the reported
compound by analytical data: [a]24 ˆ16.68 (cˆ0.68,
D
CHCl3) (97% ee) [lit.14 [a]24 ˆ22.68 (cˆ2.7, CHCl3) for
D
R enantiomer in 40% ee]. The enantiomeric excess was
determined by HPLC after the conversion to TBDMS±
1
ether: IR (neat) n 3086, 3064, 3028, 1255, 1113 cm21; H
NMR (CDCl3) d 7.32±7.29 (m, 2H), 7.24±7.18 (m, 3H),
4.42 (t, Jˆ6.4 Hz, 1H), 2.81 (m, 2H), 2.12 (m, 2H), 0.90 (s,
9H), 0.18 (s, 3H), 0.13 (s, 3H); 13C NMR (CDCl3) d 140.0,
128.7, 128.4, 126.4, 119.9, 61.2, 37.9, 25.6, 18.1, 20.51,
25.32; MS m/z 218 (M12tBu); Anal. Calcd for
C16H25ONSi: C, 69.76; H,9.15; N, 5.08. Found: C, 69.66;
H, 8.98; N, 5.04; HPLC (DAICEL CHIRALCEL OD,
hexane/2-propanol 99/1, 1.0 mL/min) tR 8.4 min and
10.7 min.
1597, 1517, 1364, 1117 cm21; H NMR (CDCl3) d 8.9 (s,
1
2H), 7.59 (d, Jˆ7.9 Hz, 2H), 7.54 (d, Jˆ2.2 Hz, 2H), 7.50
(dd, Jˆ8.8, 8.8 Hz, 2H), 7.48 (dd, Jˆ8.6, 8.6 Hz, 2H), 7.13
(dt, Jˆ0.9, 6.7 Hz, 2H), 7.05 (t, Jˆ7.7 Hz, 2H), 6.93 (d,
Jˆ8.3 Hz, 2H), 6.6 (d, Jˆ8.9 Hz, 8H), 3.91 (dd, Jˆ14.7,
14.7 Hz, 2H), 3.70 (dd, Jˆ14.4, 14.4 Hz, 2H), 2.90 (s,
12H), 2.9 (s, 12H); 13C NMR (CDCl3) d 152.3, 152.0,
152.0, 111.3, 111.2, 133.4, 132.5, 132.4, 130.9, 130.9,
128.3, 127.5, 125.7, 125.0, 123.3, 123.2, 122.9, 118.2, 39.9,
3.4.2. (2S)-2-Hydroxy-n-octanenitrile (12b). 1H NMR
(CDCl3) 4.48 (t, Jˆ6.7 Hz, 1H), 2.31 (bs, 1H), 1.88±1.83
(m, 2H), 1.54±1.47 (m, 2H), 1.39±1.26 (m, 6H), 0.90 (t,
Jˆ7.0 Hz, 3H); 13C NMR (CDCl3) 119.5, 61.5, 35.3, 31.5,
35.8 (d, Jˆ66.9 Hz); 31P NMR (CDCl3)
d
39.8;
[a]21.1 ˆ299.18 (cˆ1.87, CH2Cl2).
D
28.6, 24.4, 22.5, 14.0; [a]21 ˆ213.38 (cˆ1.0, CHCl3) (98%
D
ee) [lit.15 [a]26 ˆ19.18 (cˆ2.82, CHCl3) for R enantiomer
3.3. General procedure for the preparation of the
catalyst
D
in 66% ee]. The enantiomeric excess was determined by
HPLC after the conversion to benzoylester: IR (neat) n
1
1913, 1731 cm21; H NMR (CDCl3) d 8.06 (m, 2H), 7.63
Into a ¯ame dried ¯ask, an achiral phosphine oxide (15 mg,
68.8 mmol) was added and dried at 658C for 2 h under the
reduced pressure. 0.1 mL of dichloromethane was added,
followed by the addition of diethylaluminium chloride
(18 mL, 17.28 mmol, 0.96 M in hexane) under argon atmos-
phere. After stirring for 10 min, the chiral ligand (13 mg,
18.2 mmol) in dichloromethane (0.35 mL) was added at
room temperature. The resulting mixture was stirred at the
same temperature for 1 h to give a clear solution. This
solution can be used directly as the catalyst in catalytic
asymmetric trimethysilylcyanation reactions.
(m, 1H), 7.48 (m, 2H), 5.59 (t, Jˆ6.7 Hz, 1H), 2.05 (m, 2H),
1.62±1.54 (m, 2H), 1.43±1.30 (m, 6H), 0.90 (m, 3H); 13C
NMR (CDCl3) d 164.8, 134.0, 130.0, 128.7, 128.4, 117.0,
61.7, 32.5, 31.5, 28.5, 24.6, 22.5, 14.0; [a]16.3 ˆ253.48
D
(cˆ0.48, CHCl3) (97% ee); Anal. Calcd for C15H19O2N:
C, 73.44; H, 7.81; N, 5.71. Found: C, 73.22; H, 7.96; N,
5.64; HPLC (DAICEL CHIRALPAK AS, hexane/2-propa-
nol 99/1, 0.5 mL/min) tR 13.4 min and 14.8 min.
3.4.3. (2S)-2-Hydroxy-3-methylbutanenitrile (12c). This
product was found to be identical with the reported
compound by analytical data: [a]19 ˆ215.48 (cˆ2.1,
3.4. General procedure for the catalytic asymmetric
trimethylsilylcyanation reactions of aldehydes
D
CHCl3) (90% ee) [lit.14 [a]24 ˆ14.28 (cˆ1.3, CHCl3) for
D
R enantiomer in 34% ee]. The enantiomeric excess was
determined by HPLC after the conversion to benzoylester:
To a stirred solution of the catalyst (0.45 mL, 17.28 mmol,
0.0384 M) was added an aldehyde (0.192 mmol) at 2408C.
After 30 min, TMSCN (46 mL, 0.346 mmol) was slowly
added over 10 h using a syringe pump. (Be careful!
TMSCN should be added dropwise from the top of the
¯ask, where the temperature may be above 158C. Because
the melting point of TMSCN is 11±128C.) The reaction
1
IR (neat) n 2970, 1732, 1258 cm21; H NMR (CDCl3) d
8.08±8.05 (m, 2H), 7.51±7.47 (m, 1H), 7.56±7.61 (m, 2H),
5.45 (d, Jˆ5.75 Hz, 1H), 2.32 (dqq, Jˆ6.7, 6.7, 5.75 Hz,
1H), 1.21 (d, Jˆ6.7 Hz, 3H), 1.18 (d, Jˆ6.7 Hz, 3H); 13C
NMR (CDCl3) d 164.8, 134.0, 130.0, 128.7, 128.5, 116.0,
66.8, 31.4, 17.9, 17.5; MS m/z 203 (M1); [a]22 ˆ248.48
D

Y. Hamashima et al. / Tetrahedron 57 (2001) 805±814
813
(cˆ0.46,CHCl3) (90% ee); HPLC (DAICEL CHIRALCEL
OJ, hexane/2-propanol 9/1, 0.5 mL/min) tR 11.7 min and
13.0 min.
compound by analytical data: [a]27 ˆ243.48 (cˆ2.5,
D
CHCl3) (92% ee) [lit.14 [a]24 ˆ136.88 (cˆ2.0, CHCl3)
D
for R enantiomer in 85% ee]. The enantiomeric excess
was determined by HPLC after conversion to ethylcarbo-
3.4.4. (2S)-2-Hydroxy-3-ethylpentanenitrile (12d). 1H
NMR (CDCl3) d 4.48 (d, Jˆ4.9 Hz, 3H), 1.68±1.41 (m,
6H), 0.98 (t, Jˆ7.3 Hz, 3H), 0.97 (t, Jˆ7.35 Hz, 3H); 13C
NMR (CDCl3) d 119.4, 64.1, 45.7, 21.7, 21.6, 11.2, 11.1;
nate: IR (neat) n 1757, 1252 cm21; H NMR (CDCl3) d
1
7.48±7.37 (m, 5H), 6.19 (s, 1H), 4.27±4.17 (m, 2H), 1.27
(t, Jˆ7.0 Hz, 3H); 13C NMR (CDCl3) d 153.4, 131.2, 130.6,
129.2, 127.9, 115.7, 66.3, 65.6, 14.1; [a]16.4 ˆ213.78
D
[a]21 ˆ217.08 (cˆ1.32, CHCl3) (83% ee) [lit.15 [a]Dˆ
(cˆ2.8, CHCl3) (74.5% ee); Anal. Calcd for C11H11O3N:
C, 64.38; H, 5.40; N, 6.83. Found: C, 64.60; H, 5.51; N,
6.69; HPLC (DAICEL CHIRALCEL OD, hexane/2-propa-
nol 99/1, 1.0 mL/min) tR 10.5 min and 13.0 min.
D
214.28 for S enantiomer in 91% ee]. The enantiomeric
excess was determined by HPLC after the conversion to
p-nitrobenzoylester: IR (neat)
n 1737, 1530, 1348,
1267 cm21; H NMR (CDCl3) d 8.34 (ddd, Jˆ8.85, 2.1,
2.1 Hz, 2H), 8.23 (ddd, Jˆ8.85, 2.1, 2.1 Hz, 2H), 5.68 (d,
Jˆ4.85 Hz, 1H), 1.90 (m, 1H), 1.74±1.57 (m, 4H), 1.05 (dd,
Jˆ7.35, 3.65 Hz, 3H), 1.03 (dd, Jˆ7.35, 3.65 Hz, 3H); 13C
NMR (CDCl3) d 163.1, 151.1, 133.7, 131.1, 123.8, 115.8,
65.2, 43.9, 22.4, 22.3, 11.3, 11.2; MS m/z 276 (M1), 277
1
3.4.8. (2S)-2-Hydroxy-2-(4-methyphenyl)acetonitrile (12i).
This product was found to be identical with the reported
compound by analytical data: [a]25 ˆ246.38 (cˆ1.2,
D
CHCl3) (90% ee) [lit.14 [a]24 ˆ136.48 (cˆ1.3, CHCl3) for
D
R enantiomer in 71% ee]. The enantiomeric excess was deter-
mined by HPLC after the derivatization to ethylcarbonate. 1H
NMR (CDCl3) d 7.43 (d, Jˆ8.25 Hz, 2H), 7.26 (d,
Jˆ8.25 Hz, 2H), 6.20 (s, 1H), 4.28 (m, 2H), 2.39 (s, 3H),
1.33 (t, Jˆ7.3 Hz, 3H); 13C NMR (CDCl3) d 153.5, 140.9,
129.9, 128.4, 127.9, 115.9, 66.3, 65.5, 21.3, 14.1; HPLC
(DAICEL CHIRALCEL OD, hexane/2-propanol 99/1,
0.5 mL/min) tR 17.5 min and 19.6 min.
(M111); [a]22 ˆ238.68 (cˆ0.74, CHCl3) (83% ee); Anal.
D
Calcd for C14H16O4N2: C, 60.86; H, 5.84; N, 10.141. Found:
C, 61.06; H, 5.91; N, 9.86; HPLC (DAICEL CHIRALPAK
AD, hexane/2-propanol 99/1, 1.0 mL/min) tR 22.9 min and
28.5 min.
3.4.5. (3E)-2-Hydroxy-3-octenenitrile (12e). 1H NMR
(CDCl3) d 6.08 (ddt, Jˆ15.3, 6.7, 0.95 Hz, 1H), 5.61 (ddt,
Jˆ15.3, 6.1, 1.6 Hz, 1H), 4.94 (m, 1H), 2.47 (d, Jˆ4.9 Hz,
1H), 2.12 (m, 2H), 1.44±1.30 (m, 4H), 0.91 (t, Jˆ7.3 Hz,
3H); 13C NMR (CDCl3) d 134.1, 128.3, 118.4, 61.9, 31.6,
3.4.9. (2S)-2-(2-Furyl)-2-hydroxyethanenitrile (12j). This
product was found to be identical with the reported
compound by analytical data: [a]24 ˆ239.0 (cˆ0.6,
D
30.6, 22.1, 13.8; [a]23 ˆ119.88 (cˆ1.49, CHCl3) (97% ee)
CHCl3) (95% ee) [lit.16 [a]Dˆ114.08 (cˆ1.89, CHCl3) for
R enantiomer in 79% ee]. The enantiomeric excess was
determined by HPLC after the derivatization to acetylester;
1H NMR (CDCl3) d 7.51 (dt, Jˆ1.8 Hz, 1H), 6.69 (d,
Jˆ3.4 Hz, 1H), 6.48 (s, 1H), 6.47 (dd, Jˆ3.4, 1.8 Hz, 1H),
2.17 (s, 3H); HPLC (DAICEL CHIRALPAK AS, hexane/
2-propanol 1.0 mL/min) tR 8.5 min and 9.5 min.
D
The absolute con®guration of the product has not been
determined yet. The enantiomeric excess was determined
by HPLC after the conversion to benzoylester: IR (neat) n
2958, 2930, 1731, 1601, 1246, 969 cm21; 1H NMR (CDCl3)
d 8.07±8.05 (m, 2H), 7.62 (dt, Jˆ7.4, 1.2 Hz, 1H), 7.47 (m,
2H), 6.24 (dddd, Jˆ15.3, 6.7, 6.7, 0.9 Hz, 1H), 6.07 (dd,
Jˆ6.7, 0.9 Hz, 1H), 5.67 (dddd, Jˆ15.3, 6.7, 1.6 Hz, 1H),
2.16 (m, 2H), 1.46±1.31 (m, 4H), 0.92 (t, Jˆ7.4 Hz, 3H);
13C NMR (CDCl3) d 164.7, 140.9, 134.0, 130.0, 128.6,
128.4, 120.1, 115.9, 62.0, 31.8, 30.5, 22.2, 13.8; Anal.
Calcd for C15H17O2N: C, 74.05; H, 7.04; N, 5.76. Found:
C, 73.91; H, 7.18; N, 5.73; HPLC (DAICEL CHIRALCEL
OJ, hexane/ 2-propanol 9/1, 1.0 mL/min) tR 5.1 min and
5.7 min.
3.4.10. Trimethylsilylcyanation reaction of aldehyde 11g
using 5 mol% catalyst. Into a ¯ame dried ¯ask, 1-L
(64 mg, 0.0895 mmol) was added and dried at 508C for
2 h under the reduced pressure. 3 mL of dichloromethane
was added, followed by the addition of diethylaluminum
chloride (93 mL, 0.0895 mmol, 0.96 M in hexane) under
argon atmosphere. After stirring for 10 min, tributyl-
phosphine oxide (78 mg, 0.358 mmol) in dichloromethane
(1.2 mL) was added at room temperature. The resulting
mixture was stirred at the same temperature for 1 h to
give a clear solution. To this stirred solution of the catalyst
was added aldehyde 11g (300 mg, 1.79 mmol) in dichloro-
methane (1.4 mL) at 2408C. After 30 min, TMSCN
(287 mL, 2.15 mmol) was slowly added over 48 h using a
syringe pump. (Be careful! TMSCN should be added drop-
wise from the top of the ¯ask, where the temperature may be
above 158C. Because the melting point of TMSCN is
11±128C.) The reaction mixture was allowed to stir for
additional 2 h at the same temperature. Tri¯uoroacetic
acid (2.0 mL) was added at 2408C, and the mixture was
stirred vigorously at room temperature for 1 h to hydrolyze
the trimethylsilylether of the product. Adding ethyl acetate
(30 mL), the mixture was stirred for further 30 min. After an
usual workup, the crude product was further puri®ed by
¯ash chromatography (ethyl acetate/hexane 1:3) to give
cyanohydrin 12g in 97% yield with 99% ee.
3.4.6. (2S,3E)-2-Hydroxy-4-phenyl-3-butenenitrile (12f).
This product was found to be identical with the reported
compound by analytical data: [a]28 ˆ224.18 (cˆ0.8,
D
CHCl3) 98% ee) [lit.14 [a]24 ˆ119.28 (cˆ1.9, CHCl3) for
D
R enantiomer in 72% ee]. The enantiomeric excess was
determined by HPLC after the conversion to acetylester:
1
IR (neat) n 1751, 1654, 1213, 1020, 967 cm21; H NMR
(CDCl3) d 7.45±7.30 (m, 5H), 6.98 (d, Jˆ15.9 Hz, 1H),
6.19 (dd, Jˆ15.9, 6.7 Hz, 1H), 6.03 (dd, Jˆ6.7, 0.9 Hz),
2.18 (s, 3H); 13C NMR (CDCl3) d 168.9, 137.9, 134.4,
129.4, 128.9, 127.2, 118.4, 115.5, 61.5, 20.5; MS m/z 201
(M1); [a]22 ˆ154.08 (cˆ0.41, CHCl3) (98% ee); Calcd for
D
C12H11O2N: C, 71.63; H, 5.51; N, 6.96. Found: C, 71.99; H,
5.65; N, 6.70; HPLC (DAICEL CHIRALPAK AS, hexane/
2-propanol 150/1, 1.0 mL/min) tR 22.2 min and 24.1 min.
3.4.7. (2S)-2-Hydroxy-2-phenylacetonitrile (12h). This
product was found to be identical with the reported

814
Y. Hamashima et al. / Tetrahedron 57 (2001) 805±814
Shibasaki, M. Angew. Chem., Int. Ed. Engl. 1999, 38, 3504±
3506. (c) Matsunaga, S.; Das, J.; Roels, J.; Vogl, E. M.;
Yamamoto, N.; Iida, T.; Yamaguchi, K.; Shibasaki, M.
J. Am. Chem. Soc. 2000, 122, 2252±2260.
3.4.11. (2S)-2-Hydroxy-3-methyl-4-(2-methyl-4-thiazo-
lyl)-3-butenenitrile (12g). 1H NMR (CDCl3) d 7.0 (s,
1H), 6.76 (s, 1H), 4.99 (s, 1H), 2.75 (s, 3H), 2.15 (s 1H);
13C NMR (CDCl3) d 166.4, 150.8, 134.6, 121.5, 118.4,
117.5, 66.2, 18.9, 15.0; IR (KBr) 3039, 2821, 2694, 2361,
1508, 1450, 1381, 1277, 1197, 1166, 1091, 980, 901, 813,
743, 452 cm21; EI-MS m/z 194 (M1); EI-HRMS Calcd for
7. Preliminary communication: Hamashima, Y.; Sawada, D.;
Kanai, M.; Shibasaki, M. J. Am. Chem. Soc. 1999, 121,
2641±2642
C9H10N2OS (M1): 194.0514, Found: 194.0513; [a]16.8
ˆ
8. (a) Ooi, T.; Kagoshima, N.; Maruoka, K. J. Am. Chem. Soc.
1997, 119, 5754±5755. (b) Murakata, M.; Jono, T.;
Mizuno, Y.; Hoshino, O. J. Am. Chem. Soc. 1997, 119,
11713±11714.
D
216.58 (cˆ0.7, CHCl3) (99% ee); The enantiomeric excess
was determined by HPLC after conversion to TBDMS
ether: H NMR (CDCl3) d 7.26 (s, 1H), 6.65 (s, 1H), 4.91
1
(s, 1H), 2.70 (s, 3H), 2.20 (s, 3H), 0.93 (s, 9H), 0.21 (s, 3H),
0.12 (s,3H); 13C NMR (CDCl3) d 165, 152, 134, 122, 118,
9. Sawada, D.; Shibasaki, M. Angew. Chem., Int. Ed. Engl. 2000,
39, 209±213; Sawada, D.; Kanai, M.; Shibasaki, M. J. Am.
Chem. Soc. 2000, 43, 10521±10532.
118, 68, 25, 19, 18, 14, 25.27, 25.28; [a]16.7 ˆ217.78
D
(cˆ0.845, CHCl3) (97% ee); EI-MS m/z 308 (M1); EI-
HRMS Calcd for C15H24N2OSiS (M1): 308.1379, Found:
308.1378; HPLC (DAICEL CHIRALPAK AD, hexane/
2-propanol 100/1, 1.0 mL/min) tR 6.5 min and 7.5 min.
10. Modest ee's were attributed to the one-portion addition of
TMSCN in these kinetic studies.
11. When this catalytic asymmetric cyanosilylation was applied to
a larger scale reaction (5 g scale of aldehyde 11h), it was
found that the addition of a catalytic amount of proton source
(5 mol% of MeOH) was necessary to reproduce the results of
the smaller scale reactions shown in Table 3. Although the role
of the proton source is not clear at this moment, we think that
the proton source would facilitate the catalytic cycle by proto-
nating the intermediate 15. The free cyanohydrin 12h was
found to be silylated immediately under the reaction con-
ditions. For a discussion about the role of a proton source in
facilitating the catalytic cycle, see: Takamura, M.;
Hamashima, Y.; Usuda, H.; Kanai, M.; Shibasaki, M.
Angew. Chem., Int. Ed. Engl. 2000, 39, 1650±1652; Idem,
Chem. Pharm. Bull. 2000, 48, 1586±1592.
References
1. (a) Evans, D. A.; Truesdale, L. K.; Carroll, G. L. J. Chem.
Soc., Chem. Commun. 1973, 55±56. (b) Lidy, W.;
Sundermeyer, W. Chem. Ber. 1973, 106, 587±593.
2. (a) Evans, D. A.; Wong, R. Y. J. Org. Chem. 1977, 42, 350±
351. (b) Kobayashi, S.; Tsuchiya, Y.; Mukaiyama, T. Chem.
Lett. 1991, 42, 537±540.
3. (a) Gregory, R. J. H. Chem. Rev. 1999, 99, 3649±3682.
(b) Effenberger, F. Angew. Chem., Int. Ed. Engl. 1994, 33,
1555±1564. (c) North, M. Synlett 1993, 33, 807±820.
4. (a) For recent interesting examples, see: (a) Belokon', Y. N.;
Caveda-Cepas, S.; Green, 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 (b) Hwang, C.-D.;
Hwang, D.-R.; Uang, J. Org. Chem. 1998, 63, 6762±6763.
5. (a) Shibasaki, M. Enantiomer 2000, 4, 513±527.
(b) Shibasaki, M. CHEMTRACTS-Organic Chemistry 1999,
4, 979±998. (c) Shibasaki, M.; Sasai, H.; Arai, T. Angew.
Chem., Int. Ed. Engl. 1997, 36, 1236±1256.
12. Review: Nicolaou, K. C.; Roschangar, F.; Vourloumis, D.;
Angew. Chem., Int. Ed. Engl. 1998, 37, 2014±2045. For
other recent examples, see references cited in Ref. 9.
13. The aluminum cyanide catalyst (1: MtlˆAl±CN) prepared by
the reaction of Et2AlCN and 1-L showed a lower reactivity in
the reaction of 11h.
14. Hayashi, M.; Miyamoto, Y.; Inoue, T.; Oguni, N. J. Org.
Chem. 1993, 58, 1515±1522.
15. Nitta, H.; Yu, D.; Kudo, M.; Mori, A.; Inoue, S. J. Am. Chem.
Soc. 1992, 114, 7969±7975.
16. Matthews, B.; Jackson, W.; Jayatilake, G.; Wilshire, C.;
Jacobs, H. Aust. J. Chem. 1988, 41, 1697±1609.
6. (a) For example: (a) Yoshikawa, N.; Yamada, Y. M. A.; Das,
J.; Sasai, H.; Shibasaki, M. J. Am. Chem. Soc. 1999, 121,
È
4168±4178. (b) Yamada, K.-i.; Harwood, S. J.; Groger, H.;