Angewandte
Chemie
Table 2: Exploring the substituent effects on yield and selectivity with
2,3-diketoesters.[a]
was an effective catalyst for highly enantioselective carbonyl–
ene reactions. As the success of this process was proposed to
result from the rigidity of the coordinated substrate as a result
of two-point binding, we thought this catalyst system would
also be suitable for high enantiocontrol in reactions with 2,3-
diketoesters. However, reaction of 3a with a-methylstyrene
(4a), catalyzed by scandium(III)[7e] triflate ligated with pybox
(L1), generated the product 5a with only 18% ee (Table 1,
Entry
3
5
mol%
R
R1
Yield
[%][c]
ee
Catalyst[b]
[%][d]
1
2
3
4
5
6
7
8
3a
3b
3c
3d
3e
3 f
3g
3g
3g
3g
3g
3g
5a
5b
5c
5d
5e
5 f
5g
5g
5g
5g
5g
5g
10
10
10
10
10
10
10
5
1
5
5
5
Me
Me
Me
Me
Me
Me
Et
Et
Et
Et
Et
4-MeC6H4
4-ClC6H4
Ph
Bn
Me
tBu
Ph
Ph
Ph
92
91
90
90
64
58
91
91
90
91
88
87
80
83
82
82
85
86
92
92
92
94
94
91
Table 1: Catalyst screening and optimization of reaction conditions.[a]
9[e]
10[f]
11[f,g]
12[h]
Ph
Ph
Ph
Et
[a] Reactions were carried out on a 0.25 mmol scale of 3 (keto form) with
3 equiv of a-methylstyrene (4a) in 2.0 mL of solvent, unless noted
otherwise. [b] X mol% as listed; Y=1.2 times X. [c] Yield of product after
chromatographic purification. [d] Determined by chiral-stationary-phase
HPLC analysis. [e] Reaction was stirred for 3 h. [f] Reaction was
performed at À788C then slowly warmed to 08C. [g] Reaction was carried
out on a 5.0 mmol scale of 3g. [h] The hydrate form of 3g was used, and
the reaction was run for 24 h.
Entry
Lewis acid
Ligand
Solvent
Yield [%][b]
ee [%][c]
1
2
3
4
5
6
7
8
9
10
Sc(OTf)3
Cu(OTf)2
Cu(SbF6)2
Cu(SbF6)2
Cu(SbF6)2
Cu(SbF6)2
Cu(SbF6)2
Cu(SbF6)2
Cu(SbF6)2
Cu(SbF6)2
L1
L5
L5
L2
L3
L4
L5
L5
L5
L5
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
THF
CH3CN
DCE
toluene
91
90
92
68
84
18
70
75
58
65
28
–
–
74
65
89
trace
trace
87
entries 1–6). However, product yields were significantly
impacted with R1 = Me and tBu. However, modification of
the 3-keto unit from acetyl (3c) to propanoyl (3 f) increased
the enantioselectivity from 82% to 92% (entry 7) without
diminishing the product yields. Catalyst loading could be
decreased to 1 mol% using longer reaction times, without
effecting either the yield or enantioselectivity (entries 8 and
9). Enantioselectivities were further improved to 94% when
the reaction was performed at À788C and then slowly
warmed to 08C (entry 10). To test the reproducibility of this
process, a gram-scale reaction was performed with 3g, and the
a-hydroxy-b-ketoester 5g was obtained in 88% yield with
94% ee (entry 11). Remarkably, the readily accessible
hydrate form of 3g could also employed in the catalytic
asymmetric carbonyl–ene process, thus producing 5g with
similar yield and ee value upon isolation after a 24 h reaction
time, as compared to a 1 hour reaction time needed for the
keto form 3g (entry 12 versus entry 10). Use of the hydrate
rather than the keto form is obviously only a limitation in
reaction time.
Reactions between the structurally diverse 2,3-diketoest-
ers 3 and various alkenes 4 were examined under optimized
reaction conditions (Table 3). The alkyl (R) substituents of 3,
including methyl, ethyl, isopropyl, benzyl, and cyclohexyl,
reacted smoothly with a-methylstyrene to generate 5c and
5g–j in high yield and enantioselectivity. However, enantio-
meric excess fell to 68% when R = phenyl (5k). To further
expand the reaction scope of the 2,3-diketoesters 3, additional
functionalities (R) were explored. Reactions of 3 (R = styryl
derivatives) with a-methylstyrene provided 5l and 5m,
respectively, in excellent yield and enantiomeric access. The
84
[a] Reactions were carried out on a 0.25 mmol scale of 3a (keto form)
with 3.0 equiv of a-methylstyrene (4a) in 2.0 mL of solvent at room
temperature. [b] Yield of product isolated after after column chroma-
tography. [c] Determined by chiral-stationary-phase HPLC analysis.
DCE=1,2-dichloroethane, M.S.=molecular sieves, THF=tetrahydro-
furan, Tf=trifluoromethanesulfonyl.
entry 1). However, this outcome was surprising in view of
prior reports of the slow conversions of keto analogues of
glyoxylates,[5a] as the reaction was complete within 1 hour at
room temperature (entry 1). Encouraged by this result, we
examined reactions with other Lewis acid/chiral ligand
catalysts and discovered that bidentate chiral copper(II)
bis(oxazoline) (box) complexes, which had been previously
employed in asymmetric carbonyl–ene reactions[3d,5a] with
glyoxylates, were optimum. The best results were obtained
using Cu(OTf)2 or Cu(SbF6)2 ligated with L5 (entries 2
and 3). With the chiral box ligands L2–L4 and Cu(SbF6)2,
comparable activity was found, but the carbonyl–ene product
was formed with lower ee values (entries 4–6). Changing the
solvent to THF and CH3CN resulted in only trace amounts of
product, whereas DCE and toluene provided similar results to
those obtained with CH2Cl2 (entries 7–10). Since Cu(SbF6)2
ligated with L5 provided the highest yield and ee value, this
catalytic system was selected for further optimization.
[3d]
[5a]
A variety of aryl and alkyl ester derivatives (3) were
examined to determine the influence of structure on enantio-
selectivity, but they had only minimal impact (Table 2,
Angew. Chem. Int. Ed. 2014, 53, 6468 –6472
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim