Catalytic enantioselective allylation of ketoimines

Article abstract of DOI:10.1021/ja061510h

A general catalytic allylation of simple ketoimines was developed using 1 mol % of CuF·3PPh₃ as catalyst, 1.5 mol % of La(O ⁱPr)₃ as the cocatalyst, and stable and nontoxic allylboronic acid pinacol ester as the nucleophile. This reaction constituted a good template for developing the first catalytic enantioselective allylation of ketoimines. In this case, using LiOⁱPr as the cocatalyst produced higher enantioselectivity and reactivity than La(OⁱPr)3. Thus, using the CuF-cyclopentyl-DuPHOS complex (10 mol %) and LiOⁱPr (30 mol %) in the presence of BuOH (1 equiv) produced high enantioselectivity up to 93% ee from a range of aromatic ketoimines. Mechanistic studies indicated that LiO ⁱPr accelerates the reaction by increasing the concentration of an active nucleophile, allylcopper.

Full text of DOI:10.1021/ja061510h

Published on Web 05/09/2006
Catalytic Enantioselective Allylation of Ketoimines
Reiko Wada, Tomoyuki Shibuguchi, Sae Makino, Kounosuke Oisaki,
Motomu Kanai,* and Masakatsu Shibasaki*
Contribution from the Graduate School of Pharmaceutical Sciences, The UniVersity of Tokyo,
Tokyo 113-0033, Japan
Received March 4, 2006; E-mail: mshibasa@mol.f.u-tokyo.ac.jp
Abstract: A general catalytic allylation of simple ketoimines was developed using 1 mol % of CuF‚3PPh3
as catalyst, 1.5 mol % of La(OiPr)3 as the cocatalyst, and stable and nontoxic allylboronic acid pinacol
ester as the nucleophile. This reaction constituted a good template for developing the first catalytic
enantioselective allylation of ketoimines. In this case, using LiOiPr as the cocatalyst produced higher
enantioselectivity and reactivity than La(OiPr)3. Thus, using the CuF-cyclopentyl-DuPHOS complex (10
t
mol %) and LiOiPr (30 mol %) in the presence of BuOH (1 equiv) produced high enantioselectivity up to
93% ee from a range of aromatic ketoimines. Mechanistic studies indicated that LiOiPr accelerates the
reaction by increasing the concentration of an active nucleophile, allylcopper.
Introduction
obtained in only moderate yields. Thus, a new concept was
needed to achieve the desired reaction.
The asymmetric allylation of simple ketoimines to afford
enantiomerically enriched R-trisubstituted homoallylamines is
among the most useful transformations in organic synthesis.1
Two main methods have been reported for this type of
reaction: (1) the addition of an allyl Grignard reagent to chiral
N-sulfinyl ketoimines reported by Hua2 and Ellman3 and (2)
the addition of a chiral allylsilane to ketone-derived acyl
hydrazones reported by Leighton.4 Although these reactions are
practical and offer excellent stereoselectivity and substrate
generality, stoichiometric amounts of a chiral controller are
required. The catalytic enantioselective allylation of ketoimines
is an important challenge that has never been achieved so far.5
Indeed, there is not even a general racemic catalytic method
for ketoimine allylation that can be extended to an asymmetric
version because of the low reactivity of ketoimines. Marginally
successful catalytic allylation reactions of ketoimines without
enantiocontrol were reported by our group6 and Yoshida’s
group7 using allylsilanes as nucleophiles; however, both groups
studied only one substrate, and the allylation products were
Meanwhile, we previously developed a catalytic enantiose-
lective allylation of ketones using a CuF-(R,R)-iPr-DuPHOS
(8) complex as the catalyst and allylboronate as the nucleo-
phile.8,9 The addition of La(OiPr)3 as a cocatalyst was essential
for the high reactivity of this system. We proposed that highly
nucleophilic allylcopper, the actual nucleophile, is generated
from allylboronate via transmetalation. Kinetic studies indicated
that La(OiPr)3 facilitates the catalytic cycle by accelerating the
rate-determining catalyst turnover step without affecting the
enantioselectivity; however, the precise mechanism of rate
acceleration by La(OiPr)3 remained unclear.
Taking advantage of the high catalyst activity of CuF in the
allylboration reaction, we launched a project to develop a
catalytic enantioselective allylation of ketoimines. In this paper,
we describe (1) a general catalytic allylation of ketoimines, (2)
an extension to the catalytic enantioselective allylation of
ketoimines, and (3) a proposed mechanism for rate acceleration
by the cocatalyst (LiOiPr in this case) based on NMR studies.
Results and Discussion
(1) For a general review of asymmetric addition of organometallic reagents to
imines, see: (a) Bloch, R. Chem. ReV. 1998, 98, 1407. (b) Enders, D.;
Reinhold, U. Tetrahedron: Asymmetry 1997, 8, 1895.
Catalytic Allylation of Ketoimines. To realize a synthetically
useful catalytic allylation of ketoimines, we first studied the
effect of different protecting groups for the substrate nitrogen
atom. CuF‚3PPh3 was used as the catalyst (10 mol %), and
allylboronate 4 was used as the nucleophile, in the presence of
(2) Hua, D. H.; Miao, S. W.; Chan, J. S.; Iguchi, S. J. Org. Chem. 1991, 56,
4.
(3) (a) Cogan, D. A.; Liu, G.; Ellman, J. Tetrahedron 1999, 55, 8883. (b)
Ellman, J. A.; Owens, T. D.; Tang, T. P. Acc. Chem. Res. 2002, 35, 984.
(4) Berger, R.; Duff, K.; Leighton, J. L. J. Am. Chem. Soc. 2004, 126, 5686.
(5) For examples of catalytic enantioselective allylation of aldimines or
N-acylhydrazones derived from aldehydes, see: (a) Nakamura, H.; Naka-
mura, K.; Yamamoto, Y. J. Am. Chem. Soc. 1998, 120, 4242. (b) Nakamura,
K.; Nakamura, H.; Yamamoto, Y. J. Org. Chem. 1999, 64, 2614. (c)
Fernandes, R. A.; Stimac, A.; Yamamoto, Y. J. Am. Chem. Soc. 2003,
125, 14133. (d) Fernandes, R. A.; Yamamoto, Y. J. Org. Chem. 2004, 69,
735. (e) Gastner, T.; Ishitani, H.; Akiyama, R.; Kobayashi, S. Angew. Chem.,
Int. Ed. 2001, 40, 1896. (f) Hamada, T.; Manabe, K.; Kobayashi, S. Angew.
Chem., Int. Ed. 2003, 42, 3927.
8
15 mol % of La(OiPr)3 (Table 1, entries 1-3). Although
(8) Wada, R.; Oisaki, K.; Kanai, M.; Shibasaki, M. J. Am. Chem. Soc. 2004,
126, 8910.
(9) For the pioneering CuF-catalyzed enantioselective aldol reaction, see: (a)
Kru¨ger, J.; Carreira, E. M. J. Am. Chem. Soc. 1998, 120, 867. (b) Pagenkopf,
B. L.; Kru¨ger, J.; Stojanovic, A.; Carreira, E. M. Angew. Chem., Int. Ed.
1998, 37, 3124. For AgF-catalyzed enantioselective allylation reactions,
see: (c) Yanagisawa, A.; Kageyama, H.; Nakatsuka, Y.; Asakawa, K.;
Matsumoto, Y.; Yamamoto, H. Angew. Chem., Int. Ed. 1999, 38, 3701.
(d) Wadamoto, M.; Yamamoto, H. J. Am. Chem. Soc. 2005, 127, 14556.
(6) Yamasaki, S.; Fujii, K.; Wada, R.; Kanai, M.; Shibasaki, M. J. Am. Chem.
Soc. 2002, 124, 6536.
(7) Kamei, T.; Fujita, K.; Itami, K.; Yoshida, J. Org. Lett. 2005, 7, 4725.
9
10.1021/ja061510h CCC: $33.50 © 2006 American Chemical Society
J. AM. CHEM. SOC. 2006, 128, 7687-7691
7687

A R T I C L E S
Wada et al.
Table 1. Optimization of Reaction Conditions for Catalytic
Table 2. Catalytic Allylation of Ketoimines
Allylation of Ketoimines
entry
substrate
catalyst (x mol %)
additive
time (h)
yield (%)a
1
2
1
2
3a
3a
3a
10
10
10
1
none
none
none
27
4
3
24
2
0
73
85
32
94
3
4b
5b
none
1
tBuOHc
a Isolated yield. b Reactions were performed at 45 °C. c 1.0 equiv of
tBuOH was added.
N-benzoylhydrazone 1 did not afford any allylated product, both
N-phosphinoylimine 2 and N-benzylimine 3a gave the corre-
sponding products in high yield (entries 2 and 3). The less
electrophilic 3a produced a higher reaction rate and a greater
product yield than 2. This apparent contradiction can be partially
understood by analogy to the previous allylation of ketones, in
which regeneration of allylcopper, rather than allyl addition,
was found to be the turnover-limiting step.8 On the basis of
these initial observations, we selected N-benzylketoimine 3 as
the substrate for further optimization.
Next, we attempted to reduce the catalyst loading to 1 mol
%. Under those conditions, however, even increasing the
reaction temperature to 45 °C did not lead to complete
conversion (entry 4). Thus, we sought other means by which
to increase reactivity, particularly through the use of an additive
a Isolated yield. b 5 mol % of CuF‚3PPh3 and 7.5 mol % of La(OiPr)3
were used.
Table 3. Optimization of Catalytic Enantioselective Allylation of
Ketoimine 3a
t
proton source.10 A beneficial effect of added BuOH in the
catalytic allylsilylation of imines was previously recognized by
our group.6 tBuOH is expected to facilitate catalysis through
protonolysis of the intermediate copper amide that is generated
by the addition of allylcopper to an imine (see Scheme 2, 15).
As expected, we observed a significantly increased reaction rate
in the presence of 1 equiv of tBuOH, and product 7 was obtained
in 94% yield after 2 h (entry 5).
The optimized allylation conditions were applied to various
ketoimines (Table 2). Allylated products were obtained in
excellent yields from a variety of substrates, including het-
eroaromatic and enolizable aliphatic ketoimines. Therefore, this
catalytic allylation seemed to be a good template for the further
development of a catalytic enantioselective allylation of ke-
toimines.
Catalytic Enantioselective Allylation of Ketoimines. Hav-
ing established a general racemic catalytic allylation of ke-
toimines, we hoped to develop a corresponding enantioselective
variant through the use of a chiral ligand for copper. The initial
trials were conducted using 3a as a substrate and iPr-DuPHOS
(8) as a chiral ligand (Table 3, entries 1-5).11 Toluene was
a Catalyst was prepared by reducing CuF2‚2H2O with 2 equiv of chiral
phosphine to Cu in situ. See Experimental Section for details. b tBuOH was
slowly added over 2 h. c Isolated yield. d Determined by chiral HPLC.
e Yield and enantioselectivity were not constant in each run. f 10 mol % of
chiral Cu catalyst and 30 mol % of LiOiPr were used.
chosen as a solvent because it gave slightly better reaction rates
than THF. The chiral CuIF complex was reductively generated
in situ by combining CuF2‚2H2O and 8 in a 1:2 ratio.12 Other
catalyst preparation methods, including those employing an
achiral reductant (e.g., PPh3), led to inferior levels of enanti-
oselectivity, probably due to unselective catalysis by achiral
CuIF complexes.13 The rate acceleration effect of La(OiPr)3 was
once again observed in this case (entry 1 vs entry 2).
(10) For examples of beneficial effects of protic additives in facilitation of the
catalytic cycle, see: (a) Kitajima, H.; Ito, K.; Katsuki, T. Tetrahedron 1979,
101, 17015. (b) Evans, D. A.; Rovis, T.; Kozlowski, M. C.; Tedrow, J. S.
J. Am. Chem. Soc. 1999, 121, 1994. (c) Krueger, C. A.; Kuntz, K. W.;
Dzierba, C. D.; Wirschun, W. G.; Gleason, J. D.; Snapper, M. L.; Hoveyda,
A. H. J. Am. Chem. Soc. 1999, 121, 4284. (d) Takamura, M.; Hamashima,
Y.; Usuda, H.; Kanai, M.; Shibasaki, M. Angew. Chem., Int. Ed. 2000, 39,
1650. (e) Majima, K.; Takita, R.; Okada, A.; Ohshima, T.; Shibasaki, M.
J. Am. Chem. Soc. 2003, 125, 15837.
(12) Phosphines can reduce Cu(II)F2 to Cu(I)F-phosphine complexes. See:
Gulliver, D. J.; Levason, W.; Webster, M. Inorg. Chim. Acta 1981, 52,
153.
(11) Other chiral bisphosphine ligands such as BINAP and DTBM-SEGPHOS
gave less satisfactory enantiomeric excesses than iPr-DuPHOS. See
Supporting Information.
(13) Cu(I)F itself is unstable and cannot be isolated in the absence of phosphine
ligands. See: Cotton, F. A.; Wilkinson, G.; Murillo, C. A.; Bochmann, M.
AdVanced Inorganic Chemistry; Wiley-Intersciece: New York, 1999.
9
7688 J. AM. CHEM. SOC. VOL. 128, NO. 23, 2006

Catalytic Enantioselective Allylation of Ketoimines
A R T I C L E S
Table 4. Scope of Catalytic Enantioselective Allylation of
Ketoimines
conducted (Figure 1). Although the NMR experiments were
performed using PPh3 as a ligand, the obtained mechanistic
insight should also apply to the catalytic enantioselective reac-
tion. Consistent with previous observations,8 the 11B NMR
spectrum of a mixture of CuF‚3PPh3 and 4 (1:3) in THF in the
absence of LiOiPr showed two peaks corresponding to copper
fluoroborate 12b (-13.4 ppm) and fluoroboronate 13b (4.2 ppm)
(4/12b/13b ) 56:33:11, Figure 1a).16 The observation of 13b
indicates the formation of allylcopper 14; the peak intensity of
13b should directly correlate with the concentration of 14 (see
eq 1 in Figure 1a).17 In the presence of LiOiPr (CuF‚3PPh3/4/
LiOiPr ) 1:3:1.5), the concentration of allylcopper 14, which
corresponds to the combined peak intensities of boronates 13a
and 13b (overlapping), is increased significantly (4/12b/13a +
13b ) 29:13:58, Figure 1b). These results indicate that LiOiPr
promotes formation of the active nucleophile (14).
The following results suggested that this LiOiPr effect is likely
due to the generation of transient electron-rich copper alkox-
yborate 12c, which apparently has higher transmetalation ability
than fluoroborate 12b. First, when LiOiPr was mixed with 4
(1:1) in the absence of CuF‚3PPh3, the clean formation of
lithium alkoxyborate 12a (-10.1 ppm) was observed (Figure
1c; 4/12a ) 3:97). Similarly, when CuF‚3PPh3 was mixed with
4 (1:1), the Lewis acid/Lewis base adduct (copper fluoroborate
12b: -13.4 ppm) was observed to be the predominant species
in solution (data not shown, similar to Figure 1a). In neither of
these experiments were peaks corresponding to 13a or 13b
observed, indicating that transmetalation to allylmetal species
is not favorable under these conditions.18 On the other hand,
when 12a and 12b were separately generated and then combined
in a 1:1 ratio, the peak of alkoxyborate 12a selectively dis-
appeared, yet no peak assignable to 12c was apparent, presum-
ably because it was converted to 13a and 13b (both observed
at 4.2 ppm) and allylcopper 14. A small peak of fluoroborate
12b (and 12d) also remained (Figure 1d, 12b + 12d/13a +
13b ) 25:75).19 These results suggest that the alkoxyborate 12c
was produced via facile cation exchange between 12a and 12b
and that this reactive precursor (12c), rather than fluoroborate
12b, is the major species that is transformed to allylcopper 14
(eq 4).
a Catalyst was prepared by reducing 10 mol % of CuF2‚H2O with 20
b
mol % of 5. tBuOH was slowly added over 2 h. c Isolated yield.
d Determined by chiral HPLC. e The absolute configuration was determined
to be (R). f THF was used as solvent. g 8 was used as the chiral ligand.
Scheme 1. Removal of the N-Benzyl Group
Unfortunately, yields and enantioselectivities were variable. We
hypothesized that metal salt contaminents present in commercial
La(OiPr)3 might be to blame. Therefore, we examined several
alternate metal alkoxide additives (entries 2-5). LiOiPr exhib-
ited an additive effect similar to that of La(OiPr)3 but gave more
reproducible yields and slightly improved enantiomeric excesses
(84% ee; entry 5).
To further improve the enantioselectivity, extensive tuning
of the DuPHOS structure was conducted next (entries 5-8).
Using a novel ligand (9), having cyclopentyl groups,14 product
7a was obtained with increased enantioselectivity (89% ee; entry
6). The amount of copper catalyst could be reduced to 10 mol
% without affecting the enantioselectivity, provided that 30 mol
% of LiOiPr was used (entry 7).
Having determined the optimum reaction conditions, we then
examined substrate generality. Aromatic ketoimines were
converted to corresponding homoallylamines in good to excel-
lent yield and enantioselectivity (Table 4, entries 1-7). When
aliphatic ketoimines were used as substrates, however, enanti-
oselectivity was not satisfactory (see entry 8 for typical results
using an aliphatic ketoimine). Enantiomerically enriched primary
homoallylamines (e.g., 11) can be obtained by removal of the
N-benzyl group from 7 (Scheme 1). Nicolaou’s IBX chemistry15
proved ideal in this case because benzyl deprotection could be
carried out while leaving the allyl moiety intact. Although there
is still room to improve substrate generality and enantioselec-
tivity, this is the first catalytic enantioselective allylation of
ketoimines.
On the basis of this information, implicating LiOiPr as an
effective generator of active nucleophile 14, we propose the
following catalytic cycle (Scheme 2). The thus-generated
allylcopper 14 reacts with 3 to produce copper amide 15. tBuOH
acts as a proton source to dissociate the product 7 from the
copper catalyst with concomitant production of CuOtBu.
CuOtBu then reacts with 4 to regenerate 14, possibly via a
copper t-butoxyborate corresponding to 12c. Indeed, the catalytic
cycle can be efficiently initiated by CuOtBu; using 10 mol %
(16) The structures of 12 and 13 were assigned by correlation of their 11B NMR
chemical shift values to those of separately synthesized related compounds.
See ref 8 for details.
(17) Allylcopper 14 was not detected by 1H and 13C NMR, possibly because of
the dynamic allylic rearrangement of copper. Similarly, allylsilver is not
detected by NMR (see ref 9c).
Origin of Rate Acceleration by LiOiPr. To gain insight into
the acceleration mechanism of LiOiPr, NMR studies were
(18) Lithium allylborate 12a did not react with ketoimine 3a in the absence of
CuF‚3PPh3.
(19) Theoretically, the relative abundance of 13a + 13b (corresponding to the
concentration of allylcopper 14) cannot exceed 50% under these conditions.
The observed ratio (75%) might be attributed to the existence of side-
reaction pathways generating boron species undetectable by NMR (such
as polyboron aggregates) and/or species giving the same chemical shift as
13a and 13b.
(14) See Supporting Information for the synthesis of 9. Ligand 9 generally
produced higher enantioselectivity than 8 for the substrates shown in Table
4.
(15) Nicolaou, K. C.; Mathison, C. J. N.; Montagnon, T. J. Am. Chem. Soc.
2004, 126, 5192.
9
J. AM. CHEM. SOC. VOL. 128, NO. 23, 2006 7689

A R T I C L E S
Wada et al.
Figure 1. 11B NMR studies for the rate acceleration mechanism of LiOiPr. (a) CuF‚3PPh3 + allylboronate 4 (1:3). (b) CuF‚3PPh3 + 4 + LiOiPr (1:3:1.5).
(c) LiOiPr + 4 (1:1). (d) {LiOiPr + 4 (1:1)} + {CuF‚3PPh3 + 4 (1:1)} (1:1). The intensity of 13a and 13b (the peak in dashed squares) corresponds to the
concentration of the active allylcopper.
Scheme 2. Proposed Catalytic Cycle
This reaction can be applied to a wide range of ketoimines
including heteroaromatic and enolizable ketoimines, using 1 mol
% of catalyst. Second, this catalytic reaction was extended to
an enantioselective variant. In that case, LiOiPr rather than
La(OiPr)3 was the best cocatalyst, leading to significantly
improved reaction rates. Sterically tuned cyclopentyl-DuPHOS
(9) was identified as the optimum chiral ligand, and high
enantioselectivity (up to 93% ee) was produced from aromatic
N-benzylketoimines. Aliphatic ketoimines, however, gave un-
satisfactory enantioselectivity under the present conditions.
Third, the role of LiOiPr was elucidated. Intensive NMR studies
demonstrated that the addition of LiOiPr increases the concen-
tration of the active nucleophile, allylcopper, by generating an
electron-rich precursor, copper alkoxyallylborate 12c. Studies
toward improving the substrate generality of the asymmetric
of CuOtBu, allylation of 3a proceeded in 3 h, giving product
7a in 76% yield even in the absence of LiOiPr.20
reaction are ongoing in our laboratory.
Experimental Section
Conclusions
Typical Procedure for Catalytic Enantioselective Allylation of
Ketoimines. CuF2‚2H2O (4.1 mg, 0.03 mmol) and (R,R)-cyclopentyl-
DuPHOS (31.4 mg, 0.06 mmol) were refluxed in MeOH (1.0 mL) for
2 h. The solvent was evaporated, and the resulting amorphous was
coevaporated with toluene (0.1 mL) twice. LiOiPr (5.9 mg, 0.09
mmol) and ketoimine 3a (62.8 mg, 0.3 mmol) were added. Toluene
(0.3 mL) was added at 0 °C, followed by the addition of allylboro-
In this paper, we describe the first possible solution to a long-
standing problem: a catalytic enantioselective allylation of
ketoimines. First, the general basic methodology for allylation
of ketoimines was developed using a CuIF catalyst combined
with a La(OiPr)3 cocatalyst and allylboronate as a nucleophile.
t
nate 4 (171 µL, 0.9 mmol). BuOH (5.3 M in toluene (tBuOH/tol-
(20) On the basis of the proposed mechanism in Scheme 2, CuOtBu is expected
to be a more atom-economical catalyst for the catalytic enantioselective
allylation of ketoimines. CuOtBu is, however, unstable and difficult to
handle. Therefore, the use of the catalyst combination of CuIF and LiOiPr
is more convenient than the use of CuOtBu.
uene ) 1:1 v/v), 0.3 mmol, 54.6 µL) was slowly added over 2 h using
t
a syringe pump. After the addition of BuOH was completed, the
reaction mixture was stirred for 0.5 h, and then H2O was added to
9
7690 J. AM. CHEM. SOC. VOL. 128, NO. 23, 2006

Catalytic Enantioselective Allylation of Ketoimines
A R T I C L E S
quench the reaction. The product was extracted with AcOEt, and the
combined organic layer was washed with saturated aqueous NaCl. After
drying over Na2SO4, filtration, evaporation, and purification through
silica gel column chromatography gave the allylated product 7a in 92%
yield. The enantiomeric excess of the product was determined by chiral
HPLC analysis. HPLC (DAICEL CHIRALPAK OD-H, hexane/2-
propanol 1000:1, 1.0 mL/min) tR: 12.1 min (major) and 16.1 min
(minor).
Acknowledgment. We thank Dr. Sastry T.V.R.S. for his help
with ligand synthesis. R.W., T.S., and K.O. thank JSPS for
research fellowships.
Supporting Information Available: Experimental procedures
and characterization of the products. This material is available
free of charge via the Internet at http://pubs.acs.org.
JA061510H
9
J. AM. CHEM. SOC. VOL. 128, NO. 23, 2006 7691