Angewandte
Chemie
allenoates (path C). The latter possibility is a well-known
enantioselectivity dropped dramatically when the l-pipe-
process.[9] Herein, we report the first asymmetric C H
colic-acid-derived L2 was used (entry 6). In spite of an
improvement in yield when either the l-proline-derived L3 or
l-ramipril-derived guanidine L4 was used, the enantioselec-
tivity decreased (versus L1) and the by-product 4a was
detected in higher quantities (entries 7 and 8). Previous
reports showed that copper(I) acetylide could be formed from
terminal alkynes and copper salts.[13] These reactions gener-
ally required base, which also acted as a ligand to assist the
process. With this in mind, we wondered whether in situ
protonation of the imine unit of the guanidine ligand could
affect the outcome of the reaction. Accordingly, we sought to
exploit the guanidinium hydrochloride salts L·HCl as pro-
spective ligands. Fortuitously, L1·HCl improved the yield
slightly while the chemo- and enantioselectivity remained
unchanged (entry 9). Varying the amide unit of guanidinium
salts in L5·HCl or L6·HCl gave no improved outcomes in
terms of both yield and enantioselectivity (entries 10 and 11).
CuBr exhibited significantly increased reactivity, and even
decreasing the amount of copper salt to 60 mol% provided an
almost quantitative yield at the cost of a slight deterioration in
enantioselectivity (entry 12 versus entry 5). The ligand
L1·HBr performed just as well as the neutral ligand
(entry 13). Considering the poor solubility of CuBr in organic
solvents, the soluble copper salt CuBr·Me2S was tested (see
the Supporting Information). In this case, the use of the
copper(I) salt could be reduced to 15 mol% in the presence of
5 mol% of L1·HBr, thus resulting in a comparable yield and
enantioselectivity (entry 14). After an extensive screen of the
reaction conditions, our optimal reaction conditions involved
5 mol% of L1·HBr, either 50 mol% of CuBr or 15 mol%
CuBr·Me2S, and 1.2 equivalence of a-alkyl-a-diazoesters
(conditions A; see the Supporting Information), or 5 mol%
of L1·HCl, 90 mol% of CuCl, and 1.2 equivalents of a-aryl-a-
diazoesters (conditions B; see the Supporting Information).
The applicability of this process to a broad range of
terminal alkynes with 2a is shown in Table 2. Straight chain
alkyl substituted 1-alkynes gave slightly lower e.r. values
compared to cyclohexyl- and benzyl-substituted ones (3aa–
ad). Pleasingly, important functional groups, such as esters,
ethers, amides, and halogens, were tolerated, thus providing
the corresponding functionalized allenoates in up to 92–99%
yield and 89:11–97:3 e.r. (3af–aj). Moderate enantioselectiv-
ity was obtained with ethynyl benzene (1k), thus representing
a limitation of the current reaction conditions. Finally, it is
interesting to note that a gram-scale asymmetric synthesis of
3a could also be accommodated with either 15 mol% of
CuBr·Me2S or 50 mol% of CuBr. The absolute configuration
of 3a was established to be Sa by X-ray diffraction study of the
corresponding lactone derivative[10d,14] (see the Supporting
Information).
À
insertion between terminal alkynes and a-diazoesters in the
presence of cost-efficient copper(I) salts and chiral guanidi-
nium salts. Various trisubstituted chiral allenoates were
afforded directly in high yield and enantioselectivity under
mild reaction conditions. Some experiments were carried out
to probe into the reaction mechanism.
Initial studies focused on the reaction between the 1-
alkyne 1a and ethyl 2-diazopropanoate (2a) in CH2Cl2 at
308C with 20 mol% of a copper salt and 10 mol% of the
chiral guanidine L1, derived from (S)-tetrahydroisoquinoline-
3-carboxylic acid (Table 1). The 2,3-allenoate 3a was obtained
Table 1: Optimization of the reaction conditions.[a]
Entry
L
Copper salt
(x mol%)
Yield [%][b]
e.r. [%][d]
3a
(3a/4a)[c]
1
2
3
4
5
6
7
8
L1
L1
L1
L1
L1
L2
L3
L4
L1·HCl
L5·HCl
L6·HCl
L1
L1·HBr
L1·HBr
CuCl2 (20)
CuCl (20)
CuI (20)
<5 (97:3)
<5 (96:4)
18 (90:10)
30 (>20:1)
43 (>20:1)
6 (91:9)
68 (90:10)
73 (41:59)
55 (>20:1)
23 (>20:1)
62 (>20:1)
99 (>20:1)
99 (>20:1)
99 (>20:1)
94:6
93:7
88:12
94:6
98:2
78:22
90:10
68:32
97:3
CuCl2 (50)
CuCl (100)
CuCl (100)
CuCl (100)
CuCl (100)
CuCl (100)
CuCl (100)
CuCl (100)
CuBr (60)
CuBr (50)
CuBr·Me2S
(15)
9
10
11
12
13[e]
14[e]
91:9
94:6
96:4
96:4
97:3
[a] Unless otherwise noted, all reactions were carried out with L
(10 mol%), copper salt, 1a (0.10 mmol), and 2a (0.10 mmol) in CH2Cl2
(0.5 mL) at 308C for 2.5 h. [b] Yield of isolated product. [c] Determined
by 1H NMR spectroscopy. [d] Determined by HPLC, and the predominant
enantiomer was pure. [e] L·HBr (5 mol%) and 2a (1.2 equiv).
predominately, with the alkynoate 4a also detected as a minor
product, when using either CuCl2, CuCl, or CuI as the metal
source (entries 1–3). Although promising enantioselectivity
for 3a was given, the total yield was extremely low. To our
delight, when the amount of the metal salt was increased, the
yield, the ratio of 3a to 4a, and the enantioselectivity all
improved (entries 4 and 5). The reaction conditions led to the
isolation of 3a as a single product. CuCl performed better
than CuCl2, thus providing 3a in 43% yield and 98:2 e.r.
(entry 5). Encouraged by these results, we next identified the
best structure for the chiral guanidine ligand. The yield and
The scope of such an enantioselective strategy was then
explored by conducting the reaction with various a-diazo-
esters. To facilitate the determination of e.r. values by HPLC,
1l was selected as the representative 1-alkyne (Table 3). A
series of 2-diazopropanoates reacted smoothly with 1l, thus
delivering the desired allenoates in excellent yield and
enantioselectivity regardless of the steric hindrance of the
ester substituent (entries 1–5). The a-alkyl-diazoesters 2 f–h
Angew. Chem. Int. Ed. 2015, 54, 9512 –9516
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
9513