K. Nakata et al. / Tetrahedron Letters 55 (2014) 5774–5777
5775
Table 1
Preliminary investigation for reaction between 1a–d and PhMgX (2a, 9)
Conversiona (%)
Yield of 3aa (%)
Yield of (Ph)2 (%)
a
Entry
Substrate
PhMgX (1.5 equiv)
Cu cat. (30 mol %)
Additive (2 equiv)
1
2
1a (X = Br)
1a
PhMgBr (2a)
2a
—
CuI
—
—
0
32
0
28
<5
9
3
1a
1a
1a
1a
1a
1a
1a
1a
PhMgCl (9)
2a
9
CuI
CuI
CuI
CuI
CuI
CuI
—
60
59
10
<5
9
6
5
<5
12
8
<5
11
7
<5
5
<5
<5
4b
5
LiCl
LiCl
LiF
LiBr
LiI
LiCl
LiCl
LiCl
LiCl
LiCl
LiCl
LiBr
LiCl
LiCl
100
100
35
100
87
87
100
95
53
80
0
15
Quant (83)c
Quant
30
Qunat
79
76
Quant
93
48
6
7
8
9
10
11
12
13
14
15
16
17
2a
2a
2a
2a
2a
2a
2a
2a
2a
2a
2a
2a
CuCl
CuBr
CuCN
Cu(OAc)2
(CuOTf)2ÁC6H6
CuI
CuI
CuI
CuI
1a
1a
1a
75
0
0
6 (X = Cl)
6 (X = Cl)
7 (X = I)
8 (X = OTs)
100
100
Quant (87)c
Quant (83)c
a
b
c
Determined by 1H NMR spectroscopy using C6H5OMe as an internal standard.
Reaction in Et2O under otherwise the same conditions resulted in 10% conversion.
Isolated yield.
Since the incomplete conversion mentioned above was incon-
Table 2
Reaction of 1a with various ArMgBr
sistent with the hypothesis that the coupling of n-BuBr (derived
from PhBr and n-BuLi) with PhLi/MgCl2 proceeded more rapidly
than the allylic coupling of 4 with PhLi/MgCl2 (Scheme 1), we then
postulated the acceleration of the coupling reaction by LiCl, which
was produced as a byproduct in the preparation of PhMgCl from
PhLi and MgCl2. In practice, the addition of LiCl (2 equiv) indeed
accelerated the CuI-assisted reaction of bromide 1a with PhMgBr
(2a) (entry 4). The use of 1.0 and 0.5 equiv of LiCl was, however,
insufficient and resulted in recovery of ca. 10% of 1a as determined
by 1H NMR spectroscopy of the crude reaction mixture (see
Supplementary material). Reaction with PhMgCl (9) proceeded as
well, although the amount of Ph2 was slightly increased (entry
5). No acceleration of the reaction was observed in Et2O (data
not shown).13,14
The effect of other LiX on the reaction was examined next. As
shown in entries 6–8 in Table 1, LiBr was found to be as effective
as LiCl. In addition, the magnesium salts were poorly active (data
not shown). Of the several copper salts investigated (entries
9–13), CuBr exhibited similar catalytic behavior as that of CuI.
The substrate scope was then explored by subjecting the substrates
6–8 to the coupling reaction. Both iodide 7 and tosylate 8 were
found to be as reactive as bromide 1a (entries 16 and 17), whereas
chloride 6 only marginally afforded the desired product (entries 14
and 15).
Entry
1
Ar
ArMgBr
Product
Yielda (%)
91
2b
3b
3c
3d
3e
2
3
4
2c
2d
2e
91
87
87
5
6
2f
3f
85
2g
3g
89b
a
Isolated yield.
Compound 3g:1a = 94:6 after 24 h. Cf. 84:16 after 2 h.
b
To determine the scope of the method optimized for the cou-
pling of bromide 1a with 2a (Table 1, entry 4), different ArMgBr
were reacted with 1a at room temperature for 2 h. As summarized
in Table 2, p-Me- and p-MeOC6H4MgBr (2b,d) afforded the desired
coupling products 3b and 3d in 91% and 87% yields, respectively
(entries 1 and 3). Notably, the steric hindrance at the o-position
was not a significant matter (entries 2 and 4), and the reactivity
of o-substituted Ar reagents was estimated to be similar to that
observed for the methods developed previously.5f–l,6c,6f–i,7c,7f–i,8a,10
The sterically more congested reagent, 2,6-Me2C6H3MgBr (2f), also
participated in the coupling reaction to afford 3f in 85% yield (entry
5). In several previous publications, the coupling of 2,6-Me2C6H3
reagents has been examined to afford the desired products5l and
in others to be little reactive.5f,g,k,8a Compared to these previous
results, the compatibility with such a sterically hindered reagent
is an advantage of the present method. Furthermore, the electron
withdrawing reagent, p-FC6H4MgBr (2g), furnished 3g in good
yield (entry 6).
Next, iodides 1b and 1c with the CN and CO2Et groups, respec-
tively, were subjected to the present reaction with PhMgBr (2a)
(Scheme 2). Both of the reactions proceeded smoothly to afford
3h and 3i in good yields, respectively.
We then targeted olivetol dimethyl ether (3j), which is a key
intermediate for the syntheses of cannabinoids.15 Presently, 3,5-
(MeO)2C6H3X (X = Cl, Br) are available at a reasonable price,16
and thus the coupling of C5H11X with the Grignard reagent 2h or
2i would be a convenient approach to 3j (Scheme 3). However,