Copper-catalyzed enantioselective allylic alkylation of terminal alkyne pronucleophiles

Article abstract of DOI:10.1021/ja5084333

The copper-catalyzed enantioselective allylic alkylation of terminal alkynes with primary allylic phosphates was developed by the use of a new chiral N-heterocyclic carbene ligand bearing a phenolic hydroxy group at the ortho position of one of the two N-aryl groups. This reaction occurred with excellent γ-branch regioselectivity and high enantioselectivity, forming a controlled stereogenic center at the allylic/propargylic position. Various terminal alkynes, including silyl, aliphatic, and aromatic alkynes, could be used directly without premetalation of the C(sp)-H bond. On the basis of the results of experiments using an isomeric secondary allylic phosphate, which gave a branched product through an α-selective substitution reaction with retention of configuration, a reaction pathway involving 1,3-allylic migration of Cu in a ([σ + π]-allyl)copper(III) species is proposed.

Full text of DOI:10.1021/ja5084333

Article
pubs.acs.org/JACS
Copper-Catalyzed Enantioselective Allylic Alkylation of Terminal
Alkyne Pronucleophiles
Ayumi Harada, Yusuke Makida, Tatsunori Sato, Hirohisa Ohmiya,* and Masaya Sawamura*
Department of Chemistry, Faculty of Science, Hokkaido University, Sapporo 060-0810, Japan
*
S Supporting Information
ABSTRACT: The copper-catalyzed enantioselective allylic alkylation of
terminal alkynes with primary allylic phosphates was developed by the
use of a new chiral N-heterocyclic carbene ligand bearing a phenolic
hydroxy group at the ortho position of one of the two N-aryl groups.
This reaction occurred with excellent γ-branch regioselectivity and high
enantioselectivity, forming a controlled stereogenic center at the allylic/
propargylic position. Various terminal alkynes, including silyl, aliphatic,
and aromatic alkynes, could be used directly without premetalation of
the C(sp)−H bond. On the basis of the results of experiments using an isomeric secondary allylic phosphate, which gave a
branched product through an α-selective substitution reaction with retention of configuration, a reaction pathway involving 1,3-
allylic migration of Cu in a ([σ + π]-allyl)copper(III) species is proposed.
INTRODUCTION
RESULTS AND DISCUSSION
Optimization of Copper-Catalyzed Enantioselective
Allylic Alkylations of Terminal Alkynes. Earlier, we
reported that allylic alkylation of terminal alkynes with
enantioenriched chiral secondary (Z)-allylic phosphates
■
■
Catalytic enantioselective allylic alkylation of alkynyl nucleo-
philes is a powerful strategy for asymmetric organic synthesis.
Readily available precursors are easily transformed into chiral
,4-enynes (skipped enynes), which are versatile synthetic
intermediates with two different handles for further trans-
1
proceeded with excellent S 2′-type regioselectivity and 1,3-
N
anti stereospecificity under the influence of a catalytic amount
1
−5
t
formations connected to the stereogenic carbon center.
To
of Cu(I) salt/1,10-phenanthroline and a stoichiometric LiO Bu
8
t
date, this has been achieved by using metal acetylide reagents.
base (Scheme 1). The use of LiO Bu was important for
2
a,b
2c
Hoveyda
and Carreira used alkynylaluminum and
8
alkynylboron reagents with copper and iridium catalyst systems,
respectively. However, the strongly basic and nucleophilic
natures of the metalating reagents and the metal acetylides
reduce the functional group compatibility of these methods.
Accordingly, a catalytic enantioselective allylic alkylation using
terminal alkynes directly as pronucleophile substrates is highly
desirable, due to its potential for the convergent synthesis of
Scheme 1. Previous Work
6
complex molecules.
Here, we report copper-catalyzed enantioselective allylic
alkylation using terminal alkynes and primary allylic phosphates
as the substrate couple in the presence of a stoichiometric
t
7
+
LiO Bu base. This reaction occurred with excellent γ-branch
regioselectivity and high enantioselectivity, forming a controlled
stereogenic center at the allylic/propargylic position. This was
facilitated by the catalysis of a copper(I) complex with a chiral
N-heterocyclic carbene (NHC) ligand bearing a phenolic
hydroxy group at the ortho position of one of the two N-aryl
groups. Various terminal alkynes, including silyl, aliphatic, and
aromatic alkynes, could be used directly without premetalation
of the C(sp)−H bond. Thus, base-sensitive functional groups
such as primary alkyl tosylate and p-nitrobenzoate were
compatible with this enantioselective reaction.
promotion of the reaction. Therefore, we inferred that a Li ion
activated the phosphate leaving group as a Lewis acid and that
the ancillary effect of 1,10-phenanthroline was due to binding
+
to the Li ion.
With this knowledge, we initiated a program to develop
catalytic enantioselective allylic alkylations using terminal
alkynes and achiral primary allylic phosphates as the substrate
couple. Initial screening of leaving groups was conducted for
the reaction of (Z)-5-phenyl-2-pentenol derivatives 2 with
Received: August 18, 2014
©
XXXX American Chemical Society
A
dx.doi.org/10.1021/ja5084333 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
a
Table 1. Copper-Catalyzed Enantioselective Allylic Alkylations of Terminal Alkyne 1a with (Z)-Allylic Substrate 2
t
b
c
d
entry
leaving group X
ligand
amt of LiO Bu (equiv)
temp (°C)
yield (%)
b/l (3aa:4aa)
ee (%)
1
2
3
4
5
6
7
8
9
Cl
Br
none
none
none
none
Phen
PPh3
DPPE
SIMes
IMes
L1
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
2.6
1.3
1.3
1.3
1.3
1.3
2.6
2.6
1.3
1.3
1.3
40
40
0
0
OCO Me
40
0
2
OP(O)(OEt)2
OP(O)(OEt)2
OP(O)(OEt)2
OP(O)(OEt)2
OP(O)(OEt)2
OP(O)(OEt)2
OP(O)(OEt)2
OP(O)(OEt)2
OP(O)(OEt)2
OP(O)(OEt)2
OP(O)(OEt)2
OP(O)(OEt)2
OP(O)(OEt)2
OP(O)(OEt)2
OP(O)(OEt)2
OP(O)(OEt)2
OP(O)(OEt)2
OP(O)(OEt)2
OP(O)(OEt)2
40
53
68
22
41
48
16
64
98
0
17:83
18:82
18:82
20:80
96:4
40
40
40
40
40
>99:1
91:9
10
11
12
13
14
15
16
17
18
19
20
21
22
40
43
53
L2
40
83:17
e
,
f
L2
−50
40
L3
88
96
99
81
40
64
66
94
92
93
87:13
>99:1
>99:1
>99:1
>99:1
>99:1
>99:1
>99:1
>99:1
>99:1
7
65
71
75
85
85
92
62
28
38
L4
40
e
e
e
e
e
L4
0
L4
−20
−50
−50
−50
40
L4
L4
,
f
L4
L5
L6
40
L7
40
a
The reaction was carried out with 1a (0.18 mmol), 2 (0.15 mmol), CuCl (10 mol %), ligand (12 mol %), and LiOtBu (1.3 or 2.6 equiv to (Z)-2a)
b
c
1
in toluene (0.6 mL) for 24 h unless otherwise noted. The yield of isolated product. The ratio was determined by H NMR analysis of the crude
product. The enantiomeric excess was determined by HPLC analysis. The reaction was carried out for 48 h. The reaction solvent was toluene/
d
e
f
DCM 4/1.
triisopropylsilylacetylene (1a) in the presence of CuCl (10 mol
t
%
) and LiO Bu (1.3 equiv) in toluene at 40 °C, and we
confirmed that a phosphate was the only effective leaving group
among those examined (Table 1, entries 1−4). In fact, the
reaction between 1a and (Z)-allylic phosphate 2a afforded the
linear, achiral product 4aa with (Z)-alkene geometry as the
major isomer (branched/linear (b/l) 17:83) (Table 1, entry
9
4).
Next, various achiral ligands were screened for the selective
formation of the racemic, branched product (3aa) in the
reaction between 1a and (Z)-2a in the presence of CuCl (10
t
mol %) and LiO Bu in toluene at 40 °C (Table 1, entries 5−
9
9
). Even with 1,10-phenanthroline (Phen), which gave
excellent S 2′-type selectivity in the previously reported
N
8
reaction with secondary allylic phosphates, the linear product
Figure 1. Achiral and chiral NHC ligands (shown in forms of HX
salts). The HX salts of L2−L7 were newly synthesized in this study.
4aa was obtained as the major isomer (entry 5). Monodentate
and bidentate phosphine ligands such as Ph P and 1,2-
3
diphenylphosphinoethane (DPPE) showed similar regioselec-
tivities with less efficient substrate conversions (entries 6 and
On the basis of these results, we investigated various chiral
NHC ligands (L1−L7; Figure 1). The results obtained under
the conditions used for the screening of the achiral ligands
9
7
1
). However, a copper N-heterocyclic carbene (SIMes; Figure
) complex prepared in situ from 1,3-bis(2,4,6-
(toluene, 40 °C, 24 h) are shown in Table 1, entries 10, 11, 13,
9
−11
trimethylphenyl)imidazolinium chloride (SIMes·HCl), CuCl,
14, and 20−22.
The ring-saturated C -symmetric NHC
2
t
12
and LiO Bu showed favorable branch selectivity (b/l 96:4) with
ligand (S,S)-L1, which has two stereogenic carbon centers in
the imidazolidine ring with two mesityl groups at both nitrogen
atoms, produced the branched 1,4-enyne isomer (S)-3aa with
moderate enantioselectivity (43% ee) and 91% branch
10
moderate chemical yield (entry 8). The use of 1,3-bis(2,4,6-
trimethylphenyl)imidazolium chloride (IMes·HCl; Figure 1)
resulted in exclusive branch selectivity (b/l > 99:1), but the
yield was low (16%) (entry 9).
13
regioselectivity (entry 10). A similar chiral NHC ligand,
B
dx.doi.org/10.1021/ja5084333 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
S,S)-L2, bearing a 2-methoxyphenyl group instead of one of
Article
t
(
KO Bu and LiCl (1/1) resulted in no reaction (entry 4). When
LiO Bu was changed to the sterically less demanding base
LiOEt, the chemical yield was low and enantioselectivity
decreased to 74% ee, but the excellent branch regioselectivity
was retained (entry 5). The even smaller alkoxide base LiOMe
induced no reaction (entry 6). The use of Li CO resulted in
t
the mesityl groups in L1 gave better chemical yield and
enantioselectivity (53% ee) but decreased branch selectivity (b/
l 83:17) (entry 11). Next, we investigated the effect of
incorporating the function of an alkoxide base within the chiral
14
NHC ligand. The N-hydroxyalkyl NHC ligand L3, having a
stereogenic center only in the N-alkyl side arm, gave nearly
racemic 3aa (entry 13). On the other hand, the phenol type
ligand (S,S)-L4 with a diphenylethylenediamine-derived NHC
core imparted an enantioselectivity (65% ee) better than those
of the nonhydroxylated NHC ligands (S,S)-L1 and L2 with
exceptional branch selectivity (b/l > 99:1) and excellent yield
2
3
no reaction (entry 7).
Substrate Scopes. Various allylic phosphates with different
steric demands at the γ position were then subjected to the
allylic alkylation of 1a with the Cu-(S,S)-L4 catalyst system in
9
,13
toluene/DCM (4/1) at −50 or −30 °C (Table 3).
The
allylic phosphates with methyl, ethyl, or octyl groups at the
position γ to the leaving group were converted to the
corresponding products with excellent branch selectivities (b/
l > 99:1) and with high enantioselectivities (entries 1−3). A
sterically more demanding γ substituent such as a cyclohexyl
group was also tolerated, with a high level of enantioselectivity
(91% ee) being retained (entry 4).
Due to the mildness of the reaction conditions, various
functional groups were tolerated (Table 3, entries 5−14). For
example, allylic phosphates (2) bearing silyl ether (2f), pivalate
(2g), THP ether (2h), and p-toluenesulfonate (2i) groups as
the aliphatic γ substituent reacted at −30 °C to afford the
corresponding enyne products in good yields with high
enantioselectivities (90−95% ees) (entries 5−8). Notably,
even the base-sensitive p-nitrobenzoate moiety survived under
the reaction conditions (entry 9). Bromo, cyano, ketone, and
dimethylamino substituents in the aromatic ring of the
benzoate groups were also tolerated (entries 10−13). A
carbamate group was compatible with this enantioselective
reaction (entry 14).
(
96%) (entry 14). This result suggests a functional role of the
phenolic hydroxy group in (S,S)-L4. A change of the mesityl
group of (S,S)-L4 to other aromatic groups ((S,S)-L5−L7) did
not lead to an improvement in enantioselectivity; however, the
excellent branch regioselectivity was maintained (entries 20−
2
2).
The enantioselectivity with the Cu-(S,S)-L4 catalyst was
improved to 85% ee by lowering the reaction temperature to
50 °C with a reduction of the yield (40%) (Table 1, entries
−
9
1
5−17). A higher yield (64%) was obtained with an increased
t
amount of LiO Bu, with the enantioselectivity unchanged
(
entry 18). Thereafter, we found that the mixed solvent system
toluene/DCM (4/1) gave an enantioselectivity as high as 92%
ee (entry 19). In contrast to the phenol-containing ligand (S,S)-
L4, the anisol-substituted ligand (S,S)-L2 induced no reaction
at −50 °C (entry 12). Thus, (S,S)-L4 produced much more
active catalyst than (S,S)-L2. This result strongly suggests
cooperative participation of an ionized OH group in the
catalysis.
The nature of the base had a strong impact on the yield,
branch regioselectivity, and enantioselection, as in the case for
the copper-catalyzed alkynylation of secondary allylic phos-
The effects of the structure of the alkyne substrate on the
reactivity, regioselectivity, and enantioselectivity were examined
9
,13
in the reaction with (Z)-2a (Table 4).
The silylacetylenes
8
,9
t
phates. For example, replacing LiO Bu in the reaction of 1a
1b−d, with silyl groups other than the triisopropylsilyl group,
also served as substrates to produce the corresponding 1,4-
enynes in good to high yields, with a tendency toward increased
regioselectivities and enantioselectivities with the increasing
steric demands of the silyl substituents (entries 1−3).
t
and (Z)-2a with NaO Bu under thus far optimized conditions
(
Table 1, entry 19 and Table 2, entry 1) caused a significant
decrease in chemical yield (53%), branch selectivity (b/l
8:12), and enantioselectivity (73% ee) (Table 2, entry 2). No
8
t
reaction occurred with KO Bu (entry 3). The combined use of
Various aliphatic alkynes 1e−g) with different degrees and
patterns of branching reacted with (Z)-2a with excellent branch
selectivities (b/l > 99:1) and with reasonably high enantiose-
lectivities, although the enantioselectivity in the reaction of the
linear aliphatic alkyne 1e was relatively low (73% ee) (Table 4,
a
Table 2. Base Effect
9
,13
entries 4−7).
The diastereoselective reactions with O-
protected chiral propargylic alcohols (S)-1h and (S)-1i
produced controlled stereogenic centers at both the propargylic
positions with diastereocontrols of 91:9 and 89:11, respectively
entries 8, 9, and 11). On the other hand, the reaction with
the R isomer of 1h gave a slightly decreased diastereoselectivity
87:13), showing a minor match/mismatch effect (entry 10).
15
(
b
c
d
entry
base
yield (%) b/l (3aa:4aa) ee (%)
(
t
1
2
3
4
5
6
7
LiO Bu (Table 1, entry 19)
66
53
0
>99:1
88:12
92
73
The use of tertiary propargylic alcohol 1j also gave a reasonably
high enantioselectivitiy (entry 12). The N,N-dibenzylpropargyl-
amine derivative (1k) reacted with high enantioselectivity
t
NaO Bu
t
KO Bu
t
KO Bu/LiCl (l/1)
0
(
entry 13). Phenylacetylene (1l) reacted with an enantiose-
LiOEt
20
0
>99:1
74
lectivity (65% ee) lower than those for the aliphatic alkynes
LiOMe
15
(
3
entry 14). The 1,3-enyne derivative 1m afforded dienyne
15
Li CO3
0
2
ma with moderate enantiocontrol (70% ee) (entry 15).
a
Effect of Alkene Geometry. The alkene geometry of the
The reaction was carried out with 1a (0.18 mmol), (Z)-2a (0.15
substrate is important for enantioselectivity and/or reactivity.
The reaction of (E)-2a under the optimized conditions (the
conditions for Table 1, entry 19) afforded (R)-3aa, the
antipode of the product derived from (Z)-2a, with exceptional
mmol), CuCl (10 mol %), (S,S)-L4·HBF (12 mol %), and base (0.39
mmol) in toluene/DCM (4/1, 0.6 mL) for 48 h. The yield of isolated
product. The ratio was determined by H NMR analysis of the crude
product. The enantiomeric excess was determined by HPLC analysis.
4
b
c
1
d
C
dx.doi.org/10.1021/ja5084333 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
a
Table 3. Scope of Enantioselective Allylic Alkylation of 1a
Table 4. Enantioselective Allylic Alkylation of Various
a
Alkynes 1 with 2a
a
The reaction was carried out with 1 (0.18 mmol), (Z)-2a (0.15
t
mmol), CuCl (10 mol %), (S,S)-L4·HBF (12 mol %), and LiO Bu
4
b
(0.39 mmol) in toluene/DCM (4/1, 0.6 mL) for 48 h. The yield of
c
1
isolated product. The ratio was determined by H NMR analysis of
the crude product. The enantiomeric excess was determined by
HPLC analysis. Diastereomeric ratio. (S)-1h (1.02 mmol) and (Z)-
2
d
e
f
g
a (0.85 mmol) were used. The reaction was in toluene.
a
The reaction was carried out with 1a (0.18 mmol), 2 (0.15 mmol),
t
CuCl (10 mol %), (S,S)-L4·HBF (12 mol %), and LiO Bu (0.39
mmol) in toluene/DCM (4/1, 0.6 mL) for 48 h. The yield of isolated
product. Constitutional isomer ratio b/l > 99:1 (determined by H
NMR analysis of the crude product). The enantiomeric excess was
determined by HPLC analysis. Diastereomeric ratio 1:1. Constitu-
tional isomer ratio b/l 90:10. LiO Bu (0.19 mmol) was used.
4
b
branch selectivity (>99:1), but in low product yield (11%) with
9
c
1
merely 28% ee (eq 1). The reactions of other E substrates
d
resulted in similarly low yields and enantioselectivities (see the
Supporting Information), showing that further investigations
are required for the reactions of these substrates.
e
f
g
t
D
dx.doi.org/10.1021/ja5084333 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
configuration. Even with (R,R)-L4, the configuration was
retained efficiently (entry 2).
This interesting regiochemical convergence observed in the
reactions with isomeric substrates 2a and (S)-2a′ suggests
strongly that 1,3-allylic migration of Cu in the allylcopper(III)
species was involved and that (S)-3aa was produced from the
16
most stable species. A possible reaction pathway for the α-
branch-selective alkynylation of the enantioenriched secondary
allylic phosphate 2′ with Cu-(S,S)-L4 that can explain the
retention of configuration is illustrated in Figure 2. The chiral
NHC ligand coordinates to Cu as an anionic C,O-bidentate
Copper-Catalyzed Allylic Alkynylations with an
Enantioenriched Secondary Allylic Phosphate. To see
whether or not the branch regioselectivity of the present
copper-catalyzed allylic alkylation was due to an S 2′-type
N
reaction pathway, which is common in most copper-catalyzed/-
2
1
,16
ligand. The initial interaction of alkyne 1 to Cu is η
mediated allylic substitution reactions,
we conducted
t
coordination to form A. Next, A recruits LiO Bu through a
comparative experiments with the enantioenriched secondary
+
C(sp)−H···O hydrogen bond and O···Li ···O ionic interactions
allylic phosphate (S)-2a′ (87% ee), a constitutional isomer of
to form B. Deprotonation−cupration of the C(sp)−H bond
2a, as a substrate for the reaction with 1a (Table 5). Contrary
+
affords lithium phenoxo(alkynyl)cuprate C. The Li ion bridges
the anionic acetylide carbon and phenoxide oxygen atoms in C
Table 5. Copper-Catalyzed Allylic Alkylations of Terminal
Alkyne 1a with the Enantioenriched Secondary Allylic
Phosphate (S)-2a′
(
see Proposed Models for Enantiodiscrimination for the
+
a
importance of the location of the Li ion for enantioselection).
Next, the secondary allylic phosphate 2′ binds to Cu
1
,3
(
formation of D), adopting a conformation avoiding A strain.
The copper center attacks the terminal carbon in a syn-S 2′
N
+
manner with the intramolecular assistance of the Li ion as a
Lewis acid to activate the phosphate leaving group, resulting in
b
c
entry
ligand
yield (%)
ee (%)
allylic oxidative addition to form the (π-en-σ-yl)copper(III)
3
1
2
(S,S)-L4
(R,R)-L4
87
86
87
77
complex E1. In this complex, the σ-bonded sp carbon atom
2
and the η -coordinated alkene moiety of the allyl ligand are
a
located at the positions pseudo trans and cis to the acetylide
carbon, respectively. Due to this geometry, this copper(III)
complex E1 may be resistant to reductive elimination to form
the linear skipped enyne 4 and undergo 1,3-allylic migration of
Cu, leading to the more stable isomer E2. Reductive
elimination of this complex (E2) affords the branched isomer
3 and the copper(I) complex A. The observed complete
retention of configuration indicates that the π-face inversion of
the alkene moiety in the allyl ligand in E1 is slower than the
1,3-migration−reductive elimination sequence.
The reaction was carried out with 1a (0.18 mmol), (S)-2a′ (0.15
mmol), CuCl (10 mol %), and ligand (12 mol %) in toluene (0.6 mL)
b
c
at 40 °C for 24 h. The yield of isolated product. The enantiomeric
excess was determined by HPLC analysis.
to our expectations based on the knowledge on the related
copper catalysis, the reaction with the Cu-(S,S)-L4 catalyst
produced the branched product (S)-3aa exclusively (b/l >
9
9:1) (entry 1). The enantiomeric purity of (S)-2a′ was
completely preserved with retention of the stereochemical
Figure 2. Proposed reaction pathways for allylic alkylations of terminal alkynes 1 catalyzed by the Cu-(S,S)-L4 system.
E
dx.doi.org/10.1021/ja5084333 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Proposed Reaction Pathway for the Allylic Alkylation
of Terminal Alkynes with Primary Allylic Phosphates.
The above discussion on the reaction pathway of the α-selective
alkynylation of (S)-2a′ may provide insight into the mechanism
of the enantioselective reaction of the primary allylic
phosphates (Z)-2. A proposed reaction pathway giving the
major enantiomer of the skipped enyne 3 is also shown in
Figure 2. In this reaction pathway, diastereoselective π
3.6. These results are in accord with the above assumption and
thus indicate that C,O-bidentate chelation by the phenolic
chiral ligand L4 allows robust ligand coordination toward
catalytic copper species, which is important for constructing a
chiral environment favorable for enantioselection.
Similarly, we investigated the effects of 1,10-phenanthroline/
Cu molar ratios on the regioselectivity of the reaction between
1a and (Z)-2a (Figure 4). When the 1,10-phenanthroline/Cu
coordination of the alkene moiety toward the lithium
+
(
alkynyl)(aryloxo)cuprate C forms F. Li -assisted allylic
oxidative addition gives a (π-en-σ-yl)copper(III) complex
3
(
E3) with a secondary sp carbon atom bound to Cu. Due to
steric repulsions between the γ substituent R′ and the chiral
NHC ligand L4, this intermediate is relatively unstable and
isomerizes to the more stable isomer E2 with retention of the
3
configuration at the secondary sp carbon atom. This
intermediate (E2) is common to the reaction pathway with
2′. The isomerization is accomplished through a flip of the allyl
ligand associated with dissociation of the π coordination of the
alkene moiety, probably by way of some isomeric intermediates.
The A1 strain in the allyl ligand may also contribute to the
lability of the π-en-σ-yl coordination. Finally, reductive
elimination of E2 gives the major enantiomer of the skipped
enyne 3.
,3
Evaluation of Cu−Ligand Coordination: Effect of the
Phenolic Hydroxy Group of L4. The NHC ligands lacking a
phenolic hydroxyl group (IMes, L1−3), except for SIMes, gave
reduced regioselectivities in comparison with the phenolic
chiral ligand L4 (Table 1, entries 8−11, 13, and 14). This may
be due to the partial dissociation of these monodentate ligands
Figure 4. Effects of 1,10-phenanthroline/Cu molar ratios and the
regioselectivity of the reaction between 1a (0.18 mmol) and (Z)-2a
t
(
0.15 mmol). Conditions: CuCl, 10 mol %; LiO Bu; toluene, 0.6 mL;
4
0 °C; 24 h.
molar ratio was increased in the range from 0.6 to 2.4, the low
branch regioselectivity (18%) was constant. Although 0.6 and
(
IMes, L1−3): NHC-ligated catalytic copper species promote
excellent branch selectivity, while NHC-free reaction pathways
produce the linear isomer 4aa preferentially, as is the case with
no added ligand shown in Table 1, entry 4 (b/l 17:83). To test
this assumption, we investigated the effects of (S,S)-L2/Cu
molar ratios on the regio- and enantioselectivities of the
reaction between 1a and (Z)-2a. As shown in Figure 3, the
branch regioselectivity (74−97%) was markedly improved with
the enantioselectivity (43−57% ees) nearly unchanged as the
1
.2 equiv of phenanthroline slightly increased the product yield,
further loading of this additive caused inhibition of the reaction
as a function of the loading. These results suggest that
phenanthroline−copper coordination produced inactive species
and that 1 molar equiv of phenanthroline favored the copper
catalysis. Importantly, this should not be through chelation to
the Cu atom but more probably through interaction with the
+
Li ion, because the bidentate coordination of phenanthroline
(S,S)-L2/copper ratio was increased in the range from 0.6 to
to an (alkoxo)(alkynyl)cuprate produces an 18e complex,
which does not accept the coordination of the allylic substrate.
This is consistent with the mechanism we proposed previously
for the stereospecific reaction between terminal alkynes and
8
chiral secondary allylic phosphates (Scheme 1).
Proposed Models for Enantiodiscrimination. We
propose the enantioselection models depicted in Figure 5 on
the basis of the fact that the reaction of the Z isomer of 2a
showed enantioselectivity higher than that of the E isomer
(
Table 1, entry 19, vs eq 1). In the π complex F (see Figure 2),
the copper adopts a tetrahedral coordination geometry and the
+
Li ion bridges the anionic acetylide carbon, phenoxide oxygen,
and phosphate leaving group. This well-defined location of the
+
Li ion is important. According to these assumptions, F1 has
less steric strain than F2, because the mesityl substituent is
proximal to the γ substituent (R′) of 2 in F2. These
considerations are consistent with the experimental observa-
tions that the enantioselectivity is significantly influenced by the
steric nature of the nonhydroxylated N-aryl group in the phenol
type NHC ligands (L4−L7, Table 1, entries 14 and 20−22).
The ortho,ortho disubstitution in L4 and L5 was critical for the
high enantioselectivity. The ligands L6 and L7, with 2,5-
disubstituted N-aryl groups, were much less effective.
Figure 3. Effects of (S,S)-L2/Cu molar ratios and the regio- and
enantioselectivity of the reaction between 1a (0.18 mmol) and (Z)-2a
t
(
0.15 mmol). Conditions: CuCl, 10 mol %; LiO Bu; toluene, 0.6 mL;
40 °C; 24 h. All reactions produced the skipped enynes 3aa + 4aa in
98% total yield.
F
dx.doi.org/10.1021/ja5084333 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
AUTHOR INFORMATION
Corresponding Authors
E-mail for H.O.: ohmiya@sci.hokudai.ac.jp.
E-mail for M.S.: sawamura@sci.hokudai.ac.jp.
■
*
*
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
This work was supported by Grants-in-Aid for Young Scientists
A) and Challenging Exploratory Research, JSPS, to H.O. and
■
(
by CREST and ACT-C, JST, to M.S. Y.M. thanks the JSPS for
scholarship support.
REFERENCES
■
(
1) For reviews on Cu-catalyzed enantioselective allylic substitutions
with organometallic reagents, see: (a) Falciola, C. A.; Alexakis, A. Eur.
Figure 5. Models for enantiodiscrimination.
J. Org. Chem. 2008, 3765. (b) Alexakis, A.; Bac
̈
kvall, J. E.; Krause, N.;
Pamies, O.; Dieguez, M. Chem. Rev. 2008, 108, 2796. (c) Harutyunyan,
̀
́
S. R.; den Hartog, T.; Geurts, K.; Minnaard, A. J.; Feringa, B. L. Chem.
Rev. 2008, 108, 2824.
In the case of the reaction of (E)-2, the mesityl substituent
causes steric repulsions toward the R′ substituent in both F3
and F4. Thus, the energy difference between these
intermediates is smaller than that between F1 and F2. These
considerations explain the more efficient enantioselection in the
reaction with the allylic substrates with a Z configuration.
(2) (a) Dabrowski, J. A.; Gao, F.; Hoveyda, A. H. J. Am. Chem. Soc.
2
011, 133, 4778. (b) Dabrowski, J. A.; Haeffner, F.; Hoveyda, A. H.
Angew. Chem., Int. Ed. 2013, 52, 7694. (c) Hamilton, J. Y.; Sarlah, D.;
Carreira, E. M. Angew. Chem., Int. Ed. 2013, 52, 7532.
(3) For selected examples of transformations of 1,4-enynes, see:
(a) Heffron, T. P.; Jamison, T. F. Org. Lett. 2003, 5, 2339. (b) Heffron,
T. P.; Trenkle, J. D.; Jamison, T. F. Tetrahedron 2003, 59, 8913.
(c) Shi, X.; Gorin, D. J.; Toste, F. D. J. Am. Chem. Soc. 2005, 127,
5802. (d) Prasad, B. A. B.; Yoshimoto, F. K.; Sarpong, R. J. Am. Chem.
Soc. 2005, 127, 12468. (e) Buzas, A.; Gagosz, F. J. Am. Chem. Soc.
CONCLUSION
■
The Cu-catalyzed γ-branch-selective, enantioselective allylic
alkylation of terminal alkynes using primary allylic phosphates
as electrophiles was developed by the use of a new chiral NHC
ligand bearing a phenolic hydroxy group at the ortho position of
one of the two N-aryl groups. This protocol produces
enantioenriched chiral skipped enynes with a tertiary stereo-
genic center at the allylic/propargylic position. Various terminal
alkynes, including silyl, aliphatic, and aromatic alkynes, can be
used directly without premetalation of the C(sp)−H bond. A
reaction pathway involving 1,3-allylic migration of Cu in a
2
2
006, 128, 12614. (f) Vasu, D.; Das, A.; Liu, R.-S. Chem. Commun.
010, 46, 4115. (g) Sato, T.; Onuma, T.; Nakamura, I.; Terada, M.
Org. Lett. 2011, 13, 4992.
(4) For representative examples of an important intermediate bearing
a 1,4-enyne moiety in total synthesis, see: (a) Jacobson, M.; Redfern,
R. E.; Jones, W. A.; Aldridge, M. H. Science 1970, 170, 542. (b) Stork,
G.; Isobe, M. J. Am. Chem. Soc. 1975, 97, 4745. (c) Iida, K.; Hirama,
M. J. Am. Chem. Soc. 1994, 116, 10310. (d) Hickmann, V.; Alcarazo,
̈
M.; Furstner, A. J. Am. Chem. Soc. 2010, 132, 11042.
(
[σ+π]-allyl)copper(III) species was proposed. Mechanistic
investigations aided by theoretical calculations are currently
ongoing in our laboratory.
(5) For the synthesis of chiral 1,4-enynes through Cu-catalyzed
enantioselective allylic alkylation of chlorinated enynes, see: Li, H.;
Alexakis, A. Angew. Chem., Int. Ed. 2012, 51, 1055.
(6) For Cu-catalyzed enantioselective conjugate additions of terminal
alkyne pronucleophiles, see: (a) Knopfel, T. F.; Zarotti, P.; Ichikawa,
T.; Carreira, E. M. J. Am. Chem. Soc. 2005, 127, 9682. (b) Yazaki, R.;
Kumagai, N.; Shibasaki, M. J. Am. Chem. Soc. 2010, 132, 10275.
̈
EXPERIMENTAL SECTION
■
Procedure for the Copper-Catalyzed Allylic Alkylation
(
Table 4, Entry 9). CuCl (8.4 mg, 0.085 mmol), L4 (53 mg, 0.102
t
(c) Sanz-Marco, A.; García-Ortiz, A.; Blay, G.; Fernan
R. Chem. Eur. J. 2014, 20, 668.
7) The possibility of stoichiometric formation of alkynyllithium
́
dez, I.; Pedro, J.
mmol), and LiO Bu (176 mg, 2.21 mmol) were placed in a vial
containing a magnetic stirring bar. The vial was sealed with a Teflon-
coated silicon rubber septum, and then the vial was evacuated and
filled with argon. Toluene (2.7 mL) was placed in the vial, and then
the mixture was stirred at room temperature for 5 min. Next, the chiral
propargylic alcohol derivative (S)-1h (230 mg, 1.02 mmol) and
CH Cl (0.7 mL) were added. Finally, allylic phosphate (Z)-2a (253
(
prior to the transmetalation with Cu(I) salt to form the
alkynylcopper(I) species was ruled out because the alkynes are not
t
acidic enough for deprotonation with LiO Bu.
(
8) Makida, Y.; Takayama, Y.; Ohmiya, H.; Sawamura, M. Angew.
Chem., Int. Ed. 2013, 52, 5350.
9) For Tables 1−4 and eq 1 unreacted allylic phosphate (2) was
detected in the crude materials after removal of the catalyst.
10) For reviews on N-heterocyclic carbenes (NHCs), see: (a) N-
2
2
mg, 0.85 mmol) was added at −20 °C. After 48 h of stirring at −20 °C,
the reaction mixture was diluted and extracted with diethyl ether (5
mL × 3). The combined organic layers were filtered through a short
plug of silica gel with diethyl ether as an eluent. After the volatiles were
removed under reduced pressure, flash column chromatography on
silica gel (0−3% EtOAc/hexane) gave 3ha (188 mg, 0.5 mmol) in 60%
yield.
(
(
Heterocyclic Carbenes in Transition Metal Catalysis; Glorius, F., Ed.;
Springer: Heidelberg, Germany, 2007; Topics in Organometallic
Chemistry Vol. 21. (b) N-Heterocyclic Carbenes in Synthesis; Nolan, S.
P., Ed.; Wiley-VCH: Weinheim, Germany, 2006. (c) Egbert, J. D.;
Cazin, C. S. J.; Nolan, S. P. Catal. Sci. Technol. 2013, 3, 912.
ASSOCIATED CONTENT
Supporting Information
Text and figures giving experimental details and character-
ization data for all new compounds. This material is available
free of charge via the Internet at http://pubs.acs.org.
■
(11) For recent references on enantioselective allylic substitutions
*
S
with NHC-Cu(I) alkoxide systems, see: (a) Guzman-Martinez, A.;
Hoveyda, A. H. J. Am. Chem. Soc. 2010, 132, 10634. (b) Gao, F.;
McGrath, K. P.; Lee, Y.; Hoveyda, A. H. J. Am. Chem. Soc. 2010, 132,
14315. (c) Gao, F.; Lee, Y.; Mandai, K.; Hoveyda, A. H. Angew. Chem.,
G
dx.doi.org/10.1021/ja5084333 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Int. Ed. 2010, 49, 8370. (d) Shintani, R.; Takatsu, K.; Takeda, M.;
Hayashi, T. Angew. Chem., Int. Ed. 2011, 50, 8656. (e) Park, J. K.;
Lackey, H. H.; Ondrusek, B. A.; McQuade, D. T. J. Am. Chem. Soc.
2
011, 133, 2410. (f) Jung, B.; Hoveyda, A. H. J. Am. Chem. Soc. 2012,
134, 1490. (g) Gao, F.; Carr, J. L.; Hoveyda, A. H. Angew. Chem., Int.
Ed. 2012, 51, 6613. (h) Vieira, E. M.; Haeffner, F.; Snapper, M. L.;
Hoveyda, A. H. Angew. Chem., Int. Ed. 2012, 51, 6618. (i) Delvos, L.
B.; Vyas, D. J.; Oestreich, M. Angew. Chem., Int. Ed. 2013, 52, 4650.
(
(
j) Takeda, M.; Shintani, R.; Hayashi, T. J. Org. Chem. 2013, 78, 5007.
k) Takeda, M.; Takatsu, K.; Shintani, R.; Hayashi, T. J. Org. Chem.
2
(
1
(
014, 79, 2354.
12) Scholl, M.; Ding, S.; Lee, C. W.; Grubbs, R. H. Org. Lett. 1999,
, 953.
13) The absolute configuration of 3ea was determined by a
comparison of its specific rotation with the value reported previously
ref 5). The absolute configurations of 3aa and the other products
(
given in Tables 3 and 4 were assigned by consideration of the
stereochemical pathway.
(
14) (a) Clavier, H.; Coutable, L.; Guillemin, J.-C.; Mauduit, M.
Tetrahedron: Asymmetry 2005, 16, 921. (b) Clavier, H.; Coutable, L.;
Toupet, L.; Guillemin, J.-C.; Mauduit, M. J. Organomet. Chem. 2005,
690, 5237.
(
15) Evaluation of the product enantiomeric excess values for the
reactions of terminal alkynes 1h,l,m with different reaction times and
conversions confirmed that no racemization of 3ha,la,ma occurred
under the reaction conditions.
(
16) For mechanistic studies based on DFT calculations, see:
a) Yoshikai, N.; Zhang, S.-L.; Nakamura, E. J. Am. Chem. Soc. 2008,
30, 12862. (b) Yamanaka, M.; Kato, S.; Nakamura, E. J. Am. Chem.
(
1
Soc. 2004, 126, 6287. For a review, see: (c) Yoshikai, N.; Nakamura,
E. Chem. Rev. 2012, 112, 2339.
H
dx.doi.org/10.1021/ja5084333 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX