Organic Letters
Letter
Table 1. Reaction Optimization
a
a
a
entry
Ph[B]
solvent
4a yield (%) (E/Z)
4a/5a (L/B)
6a yield (%)
1
2
3
4
5
6
7
8
PhBpin
PhBpin
PhBpin
PhB(OH)2
PhB(OH)2
PhB(OH)2
PhB(OH)2
PhB(OH)2
THF/H2O (5/1)
THF
MeOH
MeOH
DMF
toluene
dioxane
THF
52 (11/1)
0
13/1
0
0
2
2
0
13
4
0
54 (10/1)
76 (10/1)
33 (5/1)
76 (>20/1)
85 (>20/1)
13/1
15/1
16/1
10/1
14/1
>20/1
b
89 (86) (>20/1)
a
b
1
Determined by H NMR analysis. Isolated yield.
could afford the desired trans-1,4-adduct. However, the high
levels of chemo-, regio-, and stereocontrol would be nontrivial
because of the competitive generation of the 1,2-adduct and
cis-1,4-adduct as well as the direct addition of carbon
nucleophiles to carbonyls. In 1999, Tamaru and co-workers
reported a seminal work on a nickel-catalyzed three-
component coupling of aldehydes, dienes, and organometallic
reagents (basically Me2Zn and Ph2Zn) for the synthesis of
homoallylic alcohols.12 The narrow scope of organometallics,
modest levels of selectivity, and functional group compatibility,
however, significantly limit the synthetic utility.
reaction conditions at ambient temperature was able to afford
1,4-disubstituted homoallylic alcohol [trans-1,4-adduct (4a)]
in 52% yield with high levels of regio- and stereocontrol over
the 1,2-adduct (13/1 L/B) and cis-1,4-adduct (11/1 E/Z)
(Table 1, entry 1). The same conditions but in the absence of
H2O led to no conversions (entry 2). When MeOH was used
as the solvent, results similar to those of the aqueous condition
were obtained (entry 3). Using PhB(OH)2 instead of PhBpin,
the yield was improved to 76%. We found protic solvents were
not necessary, and solvents would dramatically affect the yield
and selectivity when PhB(OH)2 was used as a coupling
partner. While polar solvent DMF leads to low stereoselectivity
(5/1 E/Z) and yield, arene and ether-type solvents (toluene
and dioxane) gave 4a in excellent stereocontrol (>20/1 E/Z)
and high yields, accompanied by a substantial amount of
byproduct via direct carbonyl addition (6a) (entries 4−8).
Fortunately, THF was identified as the optimal solvent to
furnish 4a in 86% isolated yield with excellent regio- and
stereoselectivity (>20/1 E/Z, >20/1 L/B), and with no
observation of 6a (entry 8). It is noteworthy that the mild
and practical reaction condition is used at room temperature
with no exogenous base, which might permit good functional
group tolerance.
Under the optimized reaction conditions, we first examined
the scope of this reaction using various aldehydes. As shown in
Table 2, we found a wide variety of homoallylic alcohols were
obtained in good to high yields with excellent regio- and
stereoselectivity (>20/1 E/Z, >20/1 L/B) with very few
exceptions. The reaction was not sensitive to the steric effect of
the aldehyde substituent, and bulky 2,4,6-trimethyl benzalde-
hyde afforded the coupling product in high yield (4c).
Aromatic aldehydes with electron-donating and -withdrawing
groups were both suitable substrates (4a−4z). In particular,
strong electron-deficient substrates (4j−4l) that might readily
undergo direct carbonyl arylation were found to be compatible.
Many functional groups such as ethers (4d and 4e), a
trifluoromethyl group (4h), a trifluoromethoxy group (4i), a
cyano group (4k), an ester group (4l), and an alkynyl group
(4m) were tolerated under this mild condition. Interestingly,
aldehyde with a Bpin moiety was applicable (4p), suggesting
the reaction could effectively differentiate arylboronic acid and
esters. It is noteworthy that substrates with Cl and Br
substituents were viable, indicating the method is orthogonal
to traditional Suzuki−Miyaura reaction, and these handles
provide opportunities for further manipulation (4n−4p).
In contrast, the relative stability and easy availability of
organoboronic acids and esters have imparted good functional
group tolerance and great operational simplicity to the
Suzuki−Miyaura couplings, making it one of the most
commonly used reactions in organic chemistry.13 We felt
that the use of arylboronic acids and esters instead of difficult-
to-handle organometallic reagents would greatly facilitate the
synthesis of homoallylic alcohols. Nevertheless, due to the low
intrinsic nucleophilicity of organoboronic acid, strong bases are
generally required to convert “transmetalation-inactive” metal-
(II) halide into “transmetalation-active” metal(II) alkoxide.14
However, the addition of a base would promote competitive
protodeboronation of organoboronic acid and be incompatible
with base-sensitive functional groups, thus limiting the
substrate scope.15 We speculated the oxa−nickelacycle
intermediate mentioned above that contains a Ni(II) alkoxide
moiety would probably allow an efficient transmetalation of
organoboronic acid without an exogenous base.16 As part of
our ongoing research on nickel catalysis,17 we here report a
base-free highly regio- and trans-selective Ni-catalyzed three-
component coupling of butadiene, organoboronic acids, and
carbonyl compounds. This protocol provides a general,
efficient, and modular synthetic approach to 1,4-disubstituted
and 1,1,4-trisubstituted homoallylic alcohols with an excep-
tionally broad substrate scope and an extraordinary tolerance
to even highly base-sensitive functional groups (Scheme 1c).
We commenced our studies with the model reaction of
butadiene, benzaldehyde, and phenylboronic acid pinacol ester
(PhBpin) in the presence of a nickel catalyst [Ni(cod)2]. We
first tested various phosphine, and NHC ligands chelated
nickel catalysts with or without the addition of a base, but none
led to observation of the desired product 4a (see the
catalyst in a THF/H2O mixture under base- and ligand-free
2572
Org. Lett. 2021, 23, 2571−2577
Li, Yu-Qing
