Cooperative catalysis of metal and O-H...O/sp3-C-H...O two-point hydrogen bonds in alcoholic solvents: Cu-catalyzed enantioselective direct alkynylation of aldehydes with terminal alkynes

Article abstract of DOI:10.1002/chem.201301280

Catalyst-substrate hydrogen bonds in artificial catalysts usually occur in aprotic solvents, but not in protic solvents, in contrast to enzymatic catalysis. We report a case in which ligand-substrate hydrogen-bonding interactions cooperate with a transition-metal center in alcoholic solvents for enantioselective catalysis. Copper(I) complexes with prolinol-based hydroxy amino phosphane chiral ligands catalytically promoted the direct alkynylation of aldehydes with terminal alkynes in alcoholic solvents to afford nonracemic secondary propargylic alcohols with high enantioselectivities. Quantum-mechanical calculations of enantiodiscriminating transition states show the occurrence of a nonclassical sp³-C-H...O hydrogen bond as a secondary interaction between the ligand and substrate, which results in highly directional catalyst-substrate two-point hydrogen bonding. Efficient enantioselective direct carbonyl addition of terminal alkynes is achieved through ligand-substrate two-point hydrogen bonds consisting of O-H...O and sp³-C-H...O interactions that cooperate with copper in alcoholic solvents (see picture). Copyright

Full text of DOI:10.1002/chem.201301280

FULL PAPER
DOI: 10.1002/chem.201301280
3
À
À
Cooperative Catalysis of Metal and O H···O/sp -C H···O Two-Point
Hydrogen Bonds in Alcoholic Solvents: Cu-Catalyzed Enantioselective
Direct Alkynylation of Aldehydes with Terminal Alkynes
Takaoki Ishii,[a] Ryo Watanabe,[b] Toshimitsu Moriya,[a] Hirohisa Ohmiya,[a]
Seiji Mori,*[b] and Masaya Sawamura*[a]
Dedicated to Professor Teruaki Mukaiyama in celebration of the 40th anniversary of the development of the
Mukaiyama aldol reaction
Abstract: Catalyst–substrate hydrogen
bonds in artificial catalysts usually
occur in aprotic solvents, but not in
protic solvents, in contrast to enzymatic
catalysis. We report a case in which
ligand–substrate hydrogen-bonding in-
teractions cooperate with a transition-
metal center in alcoholic solvents for
enantioselective catalysis. Copper(I)
complexes with prolinol-based hydroxy
amino phosphane chiral ligands catalyt-
ically promoted the direct alkynylation
of aldehydes with terminal alkynes in
alcoholic solvents to afford nonracemic
secondary propargylic alcohols with
high enantioselectivities. Quantum-me-
chanical calculations of enantiodiscri-
minating transition states show the oc-
3
À
currence of a nonclassical sp -C H···O
hydrogen bond as a secondary interac-
tion between the ligand and substrate,
which results in highly directional cata-
Keywords: asymmetric catalysis
cooperative catalysis copper
·
·
·
·
density functional calculations
hydrogen bonds
lyst–substrate
bonding.
two-point
hydrogen
Introduction
vents. This cooperative catalysis provides an efficient
method to access a wide range of enantioenriched secondary
propargylic alcohols that are versatile synthetic intermedi-
ates for more complex nonracemic compounds.[17,18] A hy-
droxy group of the ligand plays a critical role in promoting
the reaction, and protic solvents promote a favorable reac-
tion rate and enantioselectivity. Guided by quantum-me-
Directional hydrogen bonds are the central element of enzy-
matic catalysis by stabilizing a transition state in a chemo-
and stereocontrolled manner.[1] Chemists can utilize this in-
formation to design hydrogen bonds in an effort to develop
artificial catalysts. In fact, various enantioselective artificial
catalysts reminiscent of enzymes have been developed to
date.[2–5] However, catalyst–substrate hydrogen bonds in arti-
ficial catalysts usually occur in aprotic solvents[6–8] or in or-
ganic/aqueous biphasic systems[9,10] in contrast to enzymatic
catalysis in aqueous media.[11]
chanical calculations, we propose the existence of a non-
3
À
classical sp -C H···O hydrogen bond as a secondary interac-
tion between the ligand and substrate, which results in
highly directional catalyst–substrate two-point hydrogen-
bonding.
Herein, we report the enantioselective direct alkynylation
of aldehydes with terminal alkynes[12–16] catalyzed by cop-
per(I) complexes with prolinol-based hydroxyamino phos-
phane chiral ligands and present experimental and computa-
tional evidence for directional hydrogen-bonding interac-
tions between chiral catalysts and substrates in alcoholic sol-
The catalytic enantioselective addition of terminal alkynes
À
to aldehydes is an important C C bond-formation reaction
to prepare enantioenriched secondary propargylic alcohols
that are versatile synthetic intermediates. To date, the Zn-
(OTf)2/N-methylephedrine/Et3N,[13]
InIII/1,1’-bi-2-naphthol
(BINOL)/Cy2NMe (Cy=cyclohexyl),[14] and RuII/bis(oxazo-
linyl)arene[16] systems have been developed by Carreira, Shi-
basaki, and Nishiyama. These systems, however, have prob-
lems, such as limited substrate scope or use of precious
metals and organic bases in large quantities. In contrast, our
protocol uses a base metal copper and a natural amino acid-
derived chiral ligand and is applicable for both aliphatic and
aromatic aldehydes in combination with various types of al-
kynes, thus eliminating the problems of the reported sys-
tems.[19]
[a] T. Ishii, T. Moriya, Prof. Dr. H. Ohmiya, Prof. Dr. M. Sawamura
Department of Chemistry
Faculty of Science, Hokkaido University
Sapporo 060-0810 (Japan)
Fax : (+81)11-706-3749
E-mail: sawamura@sci.hokudai.ac.jp
[b] R. Watanabe, Prof. Dr. S. Mori
Faculty of Science, Ibaraki University
Mito 310-8512 (Japan)
E-mail: smori@mx.ibaraki.ac.jp
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201301280.
Chem. Eur. J. 2013, 00, 0 – 0
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
1
&
ÞÞ
These are not the final page numbers!

Results and Discussion
Design of hydrogen-bonding-based catalysts: An earlier in-
vestigation on the CuI-catalyzed enantioselective addition of
terminal alkynes to aldehydes identified the TRAP[20] class
of chiral bisphosphanes as effective ligands to render a cata-
lytically active CuI complex that promotes alkynylation
under mild conditions[15] (Figure 1a). Unfortunately, the sub-
strate scope was limited to aromatic aldehydes and phenyla-
cetylene derivatives, and the obtained enantioselectivities
were only moderate.
The design of a second-generation catalyst system was
based on a previous observation that the use of alcoholic
solvents gave better reaction rates and enantioselectivities
than aprotic solvents. This finding implies the cooperation
of an alcoholic solvent molecule through a six-centered tran-
sition state (Figure 1a) and prompted us to develop a Cu-
based chiral catalyst that incorporates an alcoholic
hydrogen-bonding site and a phosphane-based bi-
Figure 1. Cooperative transition state hypotheses: a) First and b) second
generation.
Table 1. The effect of solvents and chiral ligands on the copper-catalyzed alkynylation
dentate coordination site within a chiral organic
scaffold to promote well-defined ligand–substrate
hydrogen bonding in the enantiodiscriminating
transition state (Figure 1b). By having an additional
oxygen-based coordination site, such a hydroxylated
phosphane ligand may be able to function as a tri-
dentate ligand for the Cu center. We expected that
the resulting tridentate coordination would enable
monomerization of a CuI/acetylide complex without
introducing a large-bite-angle chelating moiety as in
the TRAP ligands.
of 1a with 2a.[a]
Another important design concept was to place a
pyramidal sp3-hybridized atom rather than a planer
sp2-hybridized atom at the pivotal position that con-
nects the phosphane and alcoholic sites, so that the
chiral tridentate ligand fits to a tetrahedral geome-
try of the Cu center. These considerations identified
a class of prolinol-based hydroxyamino phosphane
L to be ideal (Figure 1b).
Entry Ligand
Solvent
T
[8C]
Conversion 3a
of 1a [%][a] Yield e.r.[c]
[%][b]
1
2
3
4
5
L1
L1
L1
L1
L1
hexane
toluene
THF
DMF
MeCN
40
40
40
40
40
40
40
40
40
25 (96 h)
40
40
25
25
25
25
25
42
48
32
15
19
0
71
92
92
37
43
32
15
18
0
70
92
92
89
0
95
99
98
98
98
97
48
24
98
87
67:33
66:34
64:38
68:32
72:28
N.A.[d]
82:18
83:17
82:18
85:15
N.A.[d]
94:6
Optimization: For the initial set of experiments, the
prototype hydroxy amino phosphane chiral ligand
L1 (10 mol%) was used in combination with
CuOtBu (10 mol%) for the reaction between
6
L1-OMe toluene
7
8
9
L1
L1
L1
L1
EtOH
iPrOH
tBuOH
iPrOH
ꢀ
CyCHO (1a; 0.2 mmol) and HC CSiiPr3 (2a;
10
11
12
13
14
15
16[f]
17
18
19
20[f]
21[f]
91
0
2 equiv) to give propargylic alcohol 3a in various
media, including nonpolar and polar aprotic sol-
vents and protic solvents with a constant reaction
time of 48 hours (the results are summarized in
Table 1). To our delight, the reaction proceeded at
408C in aprotic solvents such as hexane, toluene,
THF, DMF, and MeCN in low or moderate yields
and enantioselectivities (Table 1, entries 1–5). In
accord with the hydrogen-bonding-based catalyst
design, change of the hydroxy group of L1 into a
methoxy group (L1-OMe) completely inhibited the
reaction (toluene, 658C), thus indicating a critical
role for the alcoholic site (Table 1, entry 6).
L1-OMe tBuOH
L2
L2
L2
L2
L2
ent-L2
L2
L2
iPrOH
iPrOH
tBuOH
100
100
100
100
100
100
54
95:5
96:4
97:3
97:3
tBuOH+H2O[e]
tBuOH+H2O[e]
tBuOH+H2O[e]
toluene
97:3
40
40
78:22
67:23
95:5
THF
29
100
87
L2
L2
tBuOH/toluene (3:1) 25
tBuOH/toluene (1:1) 25
93:7
[a] Determined by 1H NMR spectroscopic analysis of the crude mixture. [b] Yield of
3a isolated by column chromatography on silica gel. [c] Determined by HPLC analy-
sis. [d] Not applicable. [e] Five equivalents of H2O to 1a were used. [f] 1.2 equivalents
of 2a were used.
&
2
&
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Cu-Catalyzed Enantioselective Alkynylation of Aldehydes with Terminal Alkynes
FULL PAPER
Next, protic solvents were examined despite the expected
dissociation of the ligand–substrate hydrogen bonds. Alco-
holic solvents, such as EtOH, iPrOH, and tBuOH, gave
much higher conversions and yields, and, interestingly, the
enantiocontrol was markedly improved (the R/S enantio-
meric ratio (e.r.) was as high as 83:17 with iPrOH; Table 1,
entries 7–9). The reaction at 258C resulted in even higher
enantiocontrol (85:15 e.r.) at the expense of the conversion
rate (96 h, 91% conv.; Table 1, entry 10). Importantly, no re-
action occurred with L1-OMe, even in the alcoholic solvent
(tBuOH) (Table 1, entry 11). These results strongly suggest
that a directional ligand–substrate hydrogen bond occurs in
alcoholic solvents. Furthermore, alcoholic solvent molecules
must participate in the enantiocontrol through hydrogen-
bonding interactions.
In our effort to optimize the chiral ligand, it appeared
that both the reaction rate and enantiocontrol could be im-
proved by introducing sterically demanding hydrocarbon
substituents at the a position of the hydroxy group. Conse-
quently, ligand L2 with a neopentyl group was optimal, thus
affording 3a in 95% yield of the isolated product with
94:6 e.r. (Table 1, entry 12). The
Figure 2. A gram-scale reaction of 1a and 2a with a bench-stable catalyst
precursor.
tems with the single components (Table 1, entries 20 and
21).
Upon scaling up the reaction (i.e., 1a: 10 mmol, 2a:
12 mmol) with a lower catalyst loading (Cu: 5 mol%), we
confirmed that the air-sensitive copper source CuOtBu
could be replaced with a bench-stable CuCl/K2CO3 system
without affecting the yield and enantioselectivity (Figure 2).
The addition of H2O showed no effect under these condi-
tions.
Substrates: The range of aldehydes is shown in Table 2 (en-
tries 1–12). The reaction of an aliphatic aldehyde 1b with a
branch at the b position proceeded with high enantioselec-
enantioselectivity with L2 was
95:5 e.r. when the reaction tem-
Table 2. Scope of aldehydes 1 and alkynes 2.[a]
perature was decreased to 258C
and was further increased to
96:4 e.r. when tBuOH was used
as the solvent (Table 1, en-
tries 13 and 14). Interestingly,
the addition of a small amount
of H2O (5 equiv to 1a) in-
creased the enantioselectivity to
give (R)-3a with 94% ee (i.e.,
97:3 e.r.; Table 1, entry 15). The
role of H2O is unclear. The
amount of alkyne 2a could be
decreased to 1.2 equivalents
while leaving the quantitative
Entry
R1CHO 1
Alkyne 2
Ligand
Product
Yield [%]
e.r.[b]
ꢀ
1
2
CyCH2CHO (1b)
iPrCHO (1c)
HC CSiiPr3 (2a)
L2
L2
(R)-3b
(R)-3c
75
87
95:5
93:7
ꢀ
HC CSiiPr3 (2a)
ꢀ
3
HC CSiiPr3 (2a)
L2
(R)-3d
99
97:3
ꢀ
4
HC CSiiPr3 (2a)
L2
(R)-3e
98
97:3
yield
unchanged
(Table 1,
entry 16). To obtain the S isom-
er of 3a, the anitipode of L2
(ent-L2), which was synthesized
from a d-proline-based starting
material (see the Supporting In-
formation), was successfully ap-
plied to this protocol (Table 1,
entry 17).
Aprotic solvents, such as tol-
uene and THF, were significant-
ly inferior to the alcoholic sol-
vents for optimizing both the
conversion rate and enantiocon-
trol (Table 1, entries 18 and 19).
Mixed solvent systems com-
posed of tBuOH and toluene
(3:1 or 1:1) gave e.r. values in-
ꢀ
5
6
7
8
9
tBuCHO (1 f)
PhCHO (1g)
PhCHO (1g)
3-MeO2C-C6H4CHO (1h)
4-F-C6H4CHO (1i)
4-Cl-C6H4CHO (1j)
3-MeO-C6H4CHO (1k)
3-thienyl-CHO (1l)
CyCHO (1a)
CyCHO (1a)
CyCHO (1a)
CyCHO (1a)
CyCHO (1a)
CyCHO (1a)
CyCHO (1a)
PhCHO (1g)
PhCHO (1g)
HC CSiiPr3 (2a)
L2
L2
L3
L3
L3
L3
L3
L3
L2
L2
L2
L2
L2
L2
L2
L3
L3
L3
(R)-3 f
(R)-3g
(R)-3g
(R)-3h
(R)-3i
(R)-3j
(R)-3k
(S)-3l
(R)-4a
(R)-5a
(R)-6a
(R)-7a
(R)-8a
(R)-9a
(R)-10a
(R)-8g
(R)-9g
(R)-10g
42
71
80
93
91
95
95
73
92
97
91
96
94
81
98
80
84
86
86:14
89:11
95:5
93:7
94:6
95:5
96:4
95:5
93:7
93:7
92:8
94:6
87:13
91:9
86:14
94:6
92:8
91:9
ꢀ
HC CSiiPr3 (2a)
ꢀ
HC CSiiPr3 (2a)
ꢀ
HC CSiiPr3 (2a)
ꢀ
HC CSiiPr3 (2a)
ꢀ
10
11
12
13
14
15
16
17
18
19
20
21
22
HC CSiiPr3 (2a)
ꢀ
HC CSiiPr3 (2a)
ꢀ
HC CSiiPr3 (2a)
ꢀ
HC CiBu (2b)
ꢀ
HC CtBu (2c)
ꢀ
HC CCH2N
(CH2Ph)2 (2d)
ꢀ
HC CCMe2OSiMe2tBu (2e)
ꢀ
HC CPh (2 f)
ꢀ
HC CC6H4-4-OMe (2g)
ꢀ
HC CC6H4-4-CF3 (2h)
ꢀ
HC CPh (2 f)
ꢀ
HC CC6H4-4-OMe (2g)
ꢀ
PhCHO (1g)
HC CC6H4-4-CF3 (2h)
[a] Reaction conditions: 48 h (entries 3, 4, 8–11, 13–22), 60 h (entry 12), and 72 h (entries 1, 2, 5–7). H2O
termediate between those sys- (5 equiv to 1) was added in entries 1–5. [b] Determined by HPLC analysis.
Chem. Eur. J. 2013, 00, 0 – 0
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
3
&
ÞÞ
These are not the final page numbers!

S. Mori, M. Sawamura et al.
tivity, but the yield was moderate due to self-condensation
of the aldehyde (Table 2, entry 1). On the other hand, a-
branched aldehydes (1c–e) are generally suitable substrates
in terms of both yield and enantioselectivity (Table 2, en-
tries 2–4). The tolerance toward N-based functional groups,
such as tertiary amino and carbamate groups, is an attractive
feature of this protocol. Nevertheless, tBuCHO (1 f) reacted
with moderate enantioselectivity (Table 2, entry 5). For the
reaction of PhCHO (1g), the Cy2P-based ligand L3 gave
higher enantioselectivity (95:5 e.r.) than the Ph2P-based
ligand L2 (Table 2, entries 6 vs. 7). The superiority of L3 ap-
plied over a range of substituted aromatic aldehydes (1h–k)
and a heteroaromatic aldehyde (1l) (Table 2, entries 8–12).
Various enantioenriched propargylic alcohols 4a–10a and
8b–10b with different substituents at the alkyne terminus
were obtained through copper-catalyzed protocols with L2
or L3 (Table 2, entries 13–22). Aliphatic alkynes 2b–2e with
different degrees and patterns of branching reacted with 1a
with reasonably high enantioselectivities (Table 2, en-
tries 13–16). Functionalized alkynes such as N,N-dibenzyl-
propargylamine (2d) and an O-protected propargylic alco-
hol 2e reacted with high yields and enantioselectivities (6a,
7a; Table 2, entries 15 and 16). Although the aromatic al-
kynes 2 f–h reacted with 1a in somewhat lower enantioselec-
tivities than the aliphatic alkynes (Table 2, entries 17–19),
these alkynes were suitable substrates for the reaction with
the aromatic aldehyde 1g catalyzed by using the Cu/L3
system (Table 2, entries 20–22).
observed between L2 and the product 3a, thus supporting
the monomeric nature of the catalytic species (see Figure S1
in the Supporting Information). Because the reaction was
conducted in an alcoholic solvent, participation of an exoge-
nous alcohol molecule through hydrogen bonding was con-
sidered throughout the catalytic cycle.
First, a CuI alkoxide complex A with an h2-coordinated
alkyne ligand is produced through the reaction between a
P,N,OH-ligand L and CuOtBu or CuCl/K2CO3 in the pres-
ence of an alkyne (i.e., 2). The alkoxide oxygen atom and
the terminal hydrogen atom of the alkyne likely form a hy-
drogen-bond bridge with an alcohol molecule, which is
either from the alkoxide complex or solvent. The p-complex
A should be in equilibrium with h1-acetylide complex B
through intramolecular proton transfer. In the acetylide
complex B, the OH group of L stays bound to the Cu atom
at the oxygen atom and provides an acidic hydrogen-bond
donor site, while the acetylide carbon atom offers a basic
(nucleophilic) site. Replacement of the alcohol molecule of
B with an aldehyde (i.e., 1) forms a hydrogen-bonded com-
plex C, which is a precursor for a nucleophilic addition step.
The aldehyde group in C should have an interaction with an
alcoholic solvent. The proton-assisted, carbon–carbon bond-
forming, nucleophilic addition proceeds through a six-cen-
tered transition state D(TS) to yield the alkoxocopper(I)/
propargylic alcohol complex E. Finally, the alkyne-exchange
equlibrium between E and substrate 2 releases the propar-
gylic alcohol 3 to complete the catalytic cycle. The proton-
assisted nucleophilic addition should be an enantiodiscrimi-
nating step.
Proposed reaction pathway: A proposed reaction pathway,
which accounts for most of the mechanistically important
experimental results, is illustrated in Figure 3. This cycle is
based on the hydrogen-bond-based, cooperative, six-cen-
tered transition-state model shown in Figure 1b. It is to be
noted that a nonlinear effect in enantiomeric excess was not
Quantum-mechanical studies: To understand how the chiral
catalysts impart high enantioselectivity, we carried out DFT
calculations for the transition states (TSs) of the enantiodis-
À
criminating C C bond-formation step (cf. D(TS) in
Figure 3), not only for the modeled reactions between
ꢀ
MeCHO and HC CSiMe3 (models 1a and 2a: chiral ligand
= L1 and L2, respectively), but also for the reactions be-
ꢀ
tween 1a and HC CSiiPr3 (2a) [model 1b and 2b: ligand =
L1 (Table 1, entry 10) and L2 (Table 1, entries 14–16 and
Figure 2), respectively]. Geometry optimizations were per-
formed at the B3LYP[21,22] level with the TZVP basis set[23]
for Cu and with the 6-31G(d) basis set[24] for the other
atoms (termed BI basis sets). Single-point energies of the
optimized geometries were calculated at the B3LYP-D-
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
(termed BII basis sets); energy was adjusted to introduce
the effects of the solvent polarity of tBuOH (e=12.0) and
empirical dispersion corrections.
First, the TS geometries were optimized for a simplified
ꢀ
model (model 1a) with L1, MeCHO, and a C CSiMe3
group to investigate the range of conformers (40 for each of
the major and minor enantiomers and 80 in total) of the pyr-
rolidine[28] and P,N-chelate rings with the copper atom at the
B3LYP level with a smaller basis set (termed B0 basis sets;
see the Supporting Information for details). This process re-
Figure 3. A proposed reaction pathway.
&
4
&
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Cu-Catalyzed Enantioselective Alkynylation of Aldehydes with Terminal Alkynes
FULL PAPER
2
À
sulted in convergence to eight TS structures for each isomer
(see Table S2 in the Supporting Information). The two most
meaningful conformers out of the eight were further opti-
mized at the B3LYP level with a larger basis set BI (see Fig-
ure S5 in the Supporting Information). Figure 4 shows the
most stable structures of TSs that lead to the major (M) and
conformers without a C H···O hydrogen bond are no less
than 16.0 kJmolÀ1 higher in Gibbs free energy (B3LYP-D-
AHCTUNGTREG(NNNU PCM)/BI//B3LYP/B0; see Table S3 in the Supporting Infor-
mation). Consequently, the directional two-point hydrogen-
bonding orients the carbonyl group. Nevertheless, the M
and m structures are almost equal in stability, and calcula-
tions of the Boltzmann distribution at 258C from the Gibbs
free energies of all the optimized structures for this modeled
system give an M/m abundance ratio (e.r.) as low as
57.4:42.6 (see Table S4 in the Supporting Information).
We then performed calculations for a more advanced
ꢀ
model system (model 1b) with L1, 1a, and a C CSiiPr3
group (2a-H), which corresponds to the experiment that
showed the moderate enantiomeric ratio of 85:15 (Table 1,
entry 10). To decrease the cost of the calculations, only a
single conformation of the CyCHO molecule and the most
stable three conformations of the SiiPr3 moiety were consid-
ered; the latter were obtained by conformational analysis of
2a at the HF/6-31G* level with the Spartan 08 program.[38]
The calculations, albeit with considerable ambiguity due to
the conformational constraints in the CyCHO molecule,
gave a reasonable M/m abundance ratio of 74.9:25.1 e.r. (see
Figure S10 and Table S6 in the Supporting Information).
After preliminary examination of the other simplified
model system (model 2a) with the neopentyl-substituted
ligand L2 (see Figure S11 and Table S7 in the Supporting In-
formation), which gave a M/m ratio of 64.0:36.0 e.r., we per-
formed calculations for a system (i.e., model 2b) with L2,
Figure 4. The most stable structures of the major (M) and minor (m)
À ꢀ
transition states for model 1a (L1, MeCHO, C CSiMe3). Phenyl groups
are shown as light-blue balls for clarity.
ꢀ
1a, and a C CSiiPr3 group (2a-H). These conditions corre-
minor (m) enantiomeric products, in which the pyrrolidine
ring adopts an envelope-type conformation with an out-of-
plane N atom and an equatorial hydroxymethylene side
chain. A reasonable value of 72.9 kJmolÀ1 was obtained for
the Gibbs activation energy of alkynylation through the
most stable TS(M) (see Figure S7 in the Supporting Infor-
mation). In contrast, four-centered TSs, in which the carbon-
yl oxygen atom is directly coordinated to the Cu atom with
the OH group of the ligand free from coordination, were
greater than 70 kJmolÀ1 higher in total electronic energy rel-
ative to the six-centered TSs (see Table S5 and Figures S8
and S9 in the Supporting Information). Accordingly, for fur-
ther studies, we decided to consider only six-centered TSs
that adopted the pyrrolidine and P,N-chelate-ring conforma-
tions as in the most stable TSs of model 1a (Figure 4).
spond to the optimal reaction conditions in the experiments
(Table 1, entries 13–15, and Figure 2). Geometry optimiza-
tions were conducted for 27 conformers for each of TS(M)
and TS(m), with different conformations of the neopentyl
group (ꢂ3), the CyCHO molecule (ꢂ3), and the SiiPr3
group (ꢂ3; see Table S8 and Figure S12 in the Supporting
Information). The most stable structures of M and m are
shown in Figure 5a. The M/m abundance ratio based on the
Boltzmann distribution at 258C from the Gibbs free ener-
gies of all the optimized structures was 96.9:3.1 e.r. (see
Table S8 in the Supporting Information) in accord with the
efficient enantiocontrol (up to 97:3 e.r. in tBuOH at 258C)
with the Cu/L2 system in the experiments.
The neopentyl group of L2 is relatively distant from both
the alkyne and aldehyde substrates, thus overhanging the
Cu-bound hydroxy group with its tBu hammerhead: the hy-
drogen atom on the carbon atom at the a position to the
OH group and the OH oxygen atom have van der Waals
contacts with the nearest tBu hydrogen atoms (Figure 5a).
The P-phenyl groups (omitted in Figure 5a) are also located
in regions where no direct interaction with the substrates
occurs. Despite a lack of chiral ligand–substrate steric inter-
actions (which is unusual in enantioselective catalysis), the
directional two-point hydrogen bond arranges the aldehyde
carbonyl group asymmetrically in a well-defined manner. As
a result, the difference in the steric environment around the
aldehyde between TS(M) and TS(m) is evident, as shown in
the views from the plane of the aldehyde along the develop-
A surprising feature of the TSs of model 1a (Figure 4) is
that, in addition to the hydrogen bond between the OH
group and aldehyde oxygen atom with H···O atomic distan-
ces of 1.52 and 1.51 ꢁ, the pyrrolidine ring has a direct inter-
action with the carbonyl oxygen atom through a nonclassical
2
À
C
H···O hydrogen bond with an H···O atomic distance of
À
2.24 ꢁ for both TS(M) and TS(m). Although the C H bond
lengths are normal, the H···O atomic distances are consider-
ably shorter than the sum of van der Waals radii of the H
3
À
and O atoms (ca. 2.6 ꢁ). Similar sp -C H···O interactions
are found in organic crystal structures[29] and biomole-
cules,[30] although they have rarely been referred to in the
studies on artificial catalyst systems.[31–37] Other optimized
Chem. Eur. J. 2013, 00, 0 – 0
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
5
&
ÞÞ
These are not the final page numbers!

S. Mori, M. Sawamura et al.
2
À
À
C
H···O and O H···O hydro-
gen-bonding distances by ap-
proximately 5% (see Fig-
ure S13 in the Supporting Infor-
mation). Although the possibili-
ty of a certain degree of overes-
3
À
timation of the sp -C H···O
hydrogen bond is not ruled out,
it is difficult to explain the effi-
cient enantiocontrol without
considering this secondary in-
teraction (see Figure S4 in the
Supporting Information for var-
ious higher energy conformers
without
a C-H···O hydrogen
bond in model 1a).
Conclusion
A copper-catalyzed enantiose-
lective alkynylation of alde-
hydes with terminal alkynes
with prolinol-based hydroxy
amino phosphane chiral ligands
has been developed. This reac-
tion presents a case in which
ligand–substrate
hydrogen-
bonding interactions cooperate
with a metal center in protic
solvents. Quantum-mechanical
calculations show the occur-
3
À
rence of a nonclassical sp -C
H···O hydrogen bond as a sec-
ondary interaction between the
ligand and the carbonyl sub-
strate, which results in highly
directional
catalyst–substrate
Figure 5. The most stable structures of the M and m transition states for model 2b (L2, 1a, and 2a-H). a) Side
two-point hydrogen bonding.
The enantioselective catalysis is
applicable for both aliphatic
and aromatic aldehydes in com-
views (the phenyl groups are shown as light-blue balls for clarity). b) Views (from the front) from the plane of
À
the aldehyde along the developing C C bond.
À
ing C C bond (Figure 5b). The Cy substituent in TS(M) is
arranged perpendicularly to the axis of the acetylide, where-
as that it is eclipsed in the TS(m). Consequently, TS(m) en-
counters larger steric repulsion between the Cy substituent
of the aldehyde and SiiPr3 substituent of the alkyne. The ne-
opentyl group in L2, which is distal to the substrates, might
play the role of an anchor to enhance the directionality of
the hydrogen bonds, thus producing a well-defined chiral re-
bination with various alkynes with different terminal sub-
stituents, thus providing a useful method to prepare enan-
tioenriched propargylic alcohols, thus eliminating the prob-
lems of existing systems, such as a limited substrate range or
the use of precious metals and organic bases in large quanti-
ties.[12–16] Catalyst–substrate hydrogen-bonding interactions
3
À
in protic solvents and sp -C H···O hydrogen bonds would
be useful new concepts to understand the mechanisms of co-
operative asymmetric catalysis and for future catalyst
design. On the other hand, the question of why aliphatic al-
dehydes and aromatic aldehydes favor different P substitu-
ents (PPh2 vs. PCy2) and of how the alcoholic solvents par-
ticipate in the catalysis remain to be elucidated.
action environment.
To confirm that the sp -C H···O hydrogen bond exists
3
À
even in an alcoholic solvent, we attached a single MeOH
molecule to the aldehyde C=O oxygen atom of the most-
stable conformer of TS(M) of model 2b through a hydrogen
bond (cf. D(TS) in Figure 3). This step caused only a small
structural change in the TS except for the elongation of the
&
6
&
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Cu-Catalyzed Enantioselective Alkynylation of Aldehydes with Terminal Alkynes
FULL PAPER
[16] J. Ito, R. Asai, H. Nishiyama, Org. Lett. 2010, 12, 3860–3862.
[17] J. W. Bode, E. M. Carreira, J. Am. Chem. Soc. 2001, 123, 3611–3612.
[18] J. A. Marshall, M. P. Bourbeau, Org. Lett. 2003, 5, 3197–3199.
[19] For a review on the enantioselective direct alkynylation of imines
with terminal alkynes, see: P. de Armas, D. Tejedor, F. Garcꢃa-Tella-
do, Angew. Chem. 2010, 122, 1029–1032; Angew. Chem. Int. Ed.
2010, 49, 1013–1016.
[20] ”TRAP” is an abbreviation for “Trans-chelating chiral phosphane”.
See: a) M. Sawamura, H. Hamashima, Y. Ito, Tetrahedron: Asymme-
try 1991, 2, 593–596; b) M. Sawamura, H. Hamashima, M. Suga-
wara, R. Kuwano, Y. Ito, Organometallics 1995, 14, 4549–4558.
[21] A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652.
Experimental Section
Procedure for a gram-scale reaction of 1a and 2a with a bench-stable
catalyst precursor (Figure 2): CuCl (49.5 mg, 0.5 mmol), K2CO3 (138 mg,
1.0 mmol), and L2 (312 mg, 0.7 mmol) were placed in a two-necked
round-bottom flask containing a magnetic stirring bar. The flask was
sealed with a rubber septum, evacuated, and filled with argon. tBuOH
(20 mL) was added to the flask, and the mixture was stirred at 258C for
5 min. Alkyne 2a (2.7 mL, 12 mmol) and aldehyde 1a (1.2 mL, 10 mmol)
were added to the flask. After stirring for 72 h at 258C, the reaction mix-
ture was concentrated, and the residue was subjected to silica gel chro-
matography (hexane/EtOAc 95:5) to give 3a (2.86 g, 9.7 mmol) in 97%
yield.
[22] C. Lee, W. Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785–789.
[23] A. Schꢄfer, C. Huber, R. Ahlrichs, J. Chem. Phys. 1994, 100, 5829–
5835.
[24] W. J. Hehre, L. Radom, P. v. R. Schleyer, J. A. Pople, Ab Initio Mo-
lecular Orbital Theory, Wiley, New York, 1986.
[25] S. Grimme, J. Comput. Chem. 2006, 27, 1787–1799.
Acknowledgements
ˇ
[26] S. Miertus, E. Scrocco, J. Tomasi, Chem. Phys. 1981, 55, 117–129.
[27] J. Tomasi, B. Mennucci, R. Cammi, Chem. Rev. 2005, 105, 2999–
3093.
[28] L. Carballeira, I. Pꢅrez-Juste, C. VanAlsenoy, J. Phys. Chem. A
2002, 106, 3873–3884.
[29] G. R. Desiraju, Acc. Chem. Res. 1996, 29, 441–449.
[30] P. Chakrabarti, S. Chakrabarti, J. Mol. Biol. 1998, 284, 867–873.
This work was supported by CREST (JST) to M.S and by ACT-C (JST)
to M.S. and S.M. We thank MEXT for the financial support through the
Global COE grant (Project No. B01: Catalysis as the Basis for Innova-
tion in Materials Science) to M.S. and the Grant-in-Aid for Scientific Re-
search on Innovative Areas “Molecular Activation Directed toward
Straightforward Synthesis” to S.M. The generous allotment of computa-
tion time from the Research Center for Computational Science (RCCS;
National Institutes of Natural Sciences, Japan) is also gratefully acknowl-
edged.
À
[31] For formyl C H···O hydrogen bonds in enantioselective Lewis acid-
catalyzed reactions of aldehydes, see: E. J. Corey, T. W. Lee, Chem.
Commun. 2001, 1321–1329.
[32] TS models involving hydrogen bonds between an N-adjacent sp C
3
À
H bond of N-acylimidazolinium salts and a carboxylate oxygen
atom have been proposed for the enantioselective acyl transfer cat-
alysis; see: I. Shiina, K. Nakata, K. Ono, Y. Onoda, M. Itagaki, J.
[1] G. A. Jeffrey, An Introduction to Hydrogen Bonding, Oxford Uni-
versity Press, Oxford, 1997.
[2] M. Sawamura, Y. Ito, Chem. Rev. 1992, 92, 857–871.
[3] T. Ikariya, K. Murata, R. Noyori, Org. Biomol. Chem. 2006, 4, 393–
406.
[4] A. G. Doyle, E. N. Jacobsen, Chem. Rev. 2007, 107, 5713–5743.
[5] T. Akiyama, Chem. Rev. 2007, 107, 5744–5758.
[6] T. Hayashi, K. Kanehira, T. Hagihara, M. Kumada, J. Org. Chem.
1988, 53, 113–120.
[7] Y. Ito, M. Sawamura, T. Hayashi, J. Am. Chem. Soc. 1986, 108,
6405–6406.
[8] S. Morikawa, K. Michigami, H. Amii, Org. Lett. 2010, 12, 2520–
2523.
[9] A. P. Brogan, T. J. Dickerson, K. D. Janda, Angew. Chem. 2006, 118,
8278–8280; Angew. Chem. Int. Ed. 2006, 45, 8100–8102.
[10] Y. Hayashi, Angew. Chem. 2006, 118, 8281–8282; Angew. Chem. Int.
Ed. 2006, 45, 8103–8104.
[11] The Noyori–Ikariya asymmetric transfer hydrogenation of ketones
catalyzed by transition-metal amides and related catalysis are impor-
tant exceptions; see ref. [3] and C. Wang, X. Wu, J. Xiao, Chem.
Asian J. 2008, 3, 1750–1770.
[12] A review on the enantioselective alkynylation of carbonyl com-
pounds and imines: B. M. Trost, A. H. Weiss, Adv. Synth. Catal.
2009, 351, 963–983.
Am. Chem. Soc. 2010, 132, 11629–11641.
[33] Hydrogen-bonding interactions between an N-adjacent sp -C H
2
À
bond and a carboxylate oxygen atom have been proposed for acyl
transfer catalysis with 4-(N,N-dimethylamino)pyridine (DMAP);
see: S. Xu, I. Held, B. Kempf, H. Mayr, W. Steglich, H. Zipse,
Chem. Eur. J. 2005, 11, 4751–4757.
À
[34] C H···O interactions in cinchona alkaloid-catalyzed enantioselective
methanolysis of meso-cyclic anhydrides: H. Yang, M. W. Wong, J.
Am. Chem. Soc. 2013, 135, 5808—5818.
2
À
[35] For sp -C H···O interactions in enantioselective thiourea organoca-
talysis, see: K. M. Lippert, K. Hof, D. Gerbig, D. Ley, H. Hausmann,
S. Guenther, P. R. Schreiner, Eur. J. Org. Chem. 2012, 5919—5927.
3
À
[36] List proposed the occurrence of sp -C H···O=P interactions in the
transition state of the Brønsted acid-catalyzed enantioselective SN2-
À
type O-alkylation reactions; the C H bond is located at the pentag-
ˇ
´
onal carbon center with a partial positive charge; see: I. Coric, J. H.
Kim, T. Vlaar, M. Patil, W. Thiel, B. List, Angew. Chem. 2013, 125,
3574–3577; Angew. Chem. Int. Ed. 2013, 52, 3490–3493.
À
À
[37] For X-ray crystal structures involving (2-pyridyl)3C H···Cl interac-
tions in PdIV complexes coordinated with a tridentate pyridine
ligand, see: A. Maleckis, J. W. Kampf, M. S. Sanford, J. Am. Chem.
Soc. 2013, 135, 6618–6625.
[38] Spartan 08, Wavefunction, Inc., Irvine, CA, 2008.
[13] N. K. Anand, E. M. Carreira, J. Am. Chem. Soc. 2001, 123, 9687–
9688.
[14] R. Takita, K. Yakura, T. Ohshima, M. Shibasaki, J. Am. Chem. Soc.
2005, 127, 13760–13761.
[15] a) Y. Asano, K. Hara, H. Ito, M. Sawamura, Org. Lett. 2007, 9,
3901–3904; b) Y. Asano, H. Ito, K. Hara, M. Sawamura, Organome-
tallics 2008, 27, 5984–5996.
Received: April 5, 2013
Revised: June 27, 2013
Published online: && &&, 0000
Chem. Eur. J. 2013, 00, 0 – 0
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
7
&
ÞÞ
These are not the final page numbers!

S. Mori, M. Sawamura et al.
Efficient enantioselective direct car-
bonyl addition of terminal alkynes is
achieved through ligand–substrate
Asymmetric Synthesis
T. Ishii, R. Watanabe, T. Moriya,
H. Ohmiya, S. Mori,*
M. Sawamura* ................ &&&& — &&&&
two-point hydrogen bonds consisting
3
À
À
of O H···O and sp -C H···O interac-
tions that cooperate with copper in
alcoholic solvents (see picture).
Cooperative Catalysis of Metal and
3
À
À
O H···O/sp -C H···O Two-Point
Hydrogen Bonds in Alcoholic
Solvents: Cu-Catalyzed Enantioselec-
tive Direct Alkynylation of Aldehydes
with Terminal Alkynes
&
8
&
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!