chlorides such as 5, and the transition metal-catalyzed
preparation of ArMgCl was developed instead.4 The proce-
dure is rather complicated, and the influence of the added
catalyst on our copper-catalyzed reaction is unpredictable.
Consequently, we turned our attention to aryl bromides
(ArBr), which are, in general, easily convertible into ArMgBr
(8) by routine operation.
Table 1. Effect of LiCl and MgCl2 on the CuCN-Catalyzed
Reaction of 1 with PhMgBr (8a)a
entry
additive
2a:3a
yield of 2a, %b,c
1
2
3
4
5
6
7
8
-
70:30
86:14
90:10
92:8
93:7
87:13
93:7
63
80
75
84
94 (94)
ndd
90 (72)
83 (73)
LiCl (1 equiv)
LiCl (2 equiv)
LiCl (3 equiv)
LiCl (4 equiv)
MgCl2 (1 equiv)
MgCl2 (3 equiv)
MgCl2 (4 equiv)
93:7
a Reactions were carried out with 8a (3 equiv) and CuCN (0.3 equiv) in
the presence of the additive (LiCl or MgCl2) in THF at 0 °C for 1 h. b Yields
were determined by 1H NMR analysis. c Isolated yields are shown in
parentheses. d Not determined.
was recorded with ratios of LiCl/8a (entries 4 and 5) and of
MgCl2/8a (entries 7 and 8) equal to or slightly greater than
one.7 These selectivities as well as the yields of 2a in these
entries were almost equal to those obtained with PhMgCl
(93:7, 90% NMR yield).3
The problem associated with the magnesium bromide 8,
however, is its low regioselectivity. For example, application
of the protocol shown in eq 1 to PhMgBr (8a) afforded a
70:30 mixture of 2a and 3a. We postulated that an inorganic
chloride would convert ArMgBr to ArMgCl by halide
exchange, and thus contribute to higher regioselectivity. As
presented herein, this idea was proved to be the case.
Moreover, the new protocol was successfully utilized as the
key reaction in the synthesis of AH-13205 (9),5 which is an
analogue of PGA1 (10) with EP2-receptor agonist activity.6
Furthermore, we demonstrated that the protocol is applicable
to alkenyl-magnesium bromides, thus showing a wider
applicability of the present strategy using monoacetate 1 in
organic synthesis.
High regioselectivity was also observed with other ArMg-
Br by using the protocol established above (Table 2). Thus,
Table 2. CuCN-Catalyzed Reaction with ArMgBr 8a-fa
2, 3, 8
entry suffix
yield
of 2, %b,c
Ar
additive
2:3
1
2
3
4
5
6
7
8
9
10
11
12
13
b
b
b
c
c
c
d
d
e
e
f
2-MeC6H4
2-MeC6H4
2-MeC6H4
4-MeC6H4
4-MeC6H4
4-MeC6H4
2-MeOC6H4
2-MeOC6H4
4-MeOC6H4
4-MeOC6H4
3,4-OCH2O-C6H3
3,4-OCH2O-C6H3
3,4-OCH2O-C6H3
-
70:30
91:9
93:7
75:25
91:9
94:6
91:9
94:6
90:10
91:9
75:25
93:7
93:7
77
82 (80)
87
80
94 (83)
86
68
77 (71)
78
91
75
LiCl
MgCl2
-
LiCl
MgCl2
LiCl
MgCl2
LiCl
MgCl2
-
f
f
LiCl
MgCl2
99 (92)
98 (85)
To find effective inorganic chloride(s), LiCl, NaCl, KCl,
MgCl2, CaCl2, and AlCl3 were each added in the CuCN-
catalyzed reaction of 1 with PhMgBr (8a) (3 equiv) in THF
at 0 °C. Among them, LiCl and MgCl2 did improve the native
regioselectivity of 8a (Table 1, entry 1). The best selectivity
a Reactions were carried out with ArMgBr (3 equiv) and CuCN (0.3
equiv) in the presence of LiCl or MgCl2 (4 equiv) in THF at 0 °C for 1 h.
b Yields were determined by 1H NMR analysis. c Isolated yields are shown
in parentheses.
(4) Bogdanovic´, B.; Schwickardi, M. Angew. Chem., Int. Ed. 2000, 39,
4610-4612.
the native selectivities of 2- and 4-Tol-MgBr (8b and 8c,
respectively) (entries 1 and 4) were improved by addition
of LiCl or MgCl2 (entries 2, 3 and 5, 6) to the levels (>90%
selectivity) previously observed with 2- and 4-Tol-MgCl.
Similarly, 2- and 4-(MeO)C6H4 groups were attached to 1
regioselectively in good yields (entries 7-10). Grignard
reagent 8f, prepared easily from 5 (X ) Br) and Mg, also
furnished high regioselectivity (entries 12 and 13). Note that
the corresponding magnesium chloride is hardly prepared
from 3,4-OCH2OC6H3Cl by the routine procedure as re-
ported.4
(5) A review: Nials, A. T.; Vardey, C. J.; Denyer, L. H.; Thomas, M.;
Sparrow, S. J.; Shepherd, G. D.; Coleman, R. A. CardioVasc. Drug ReV.
1993, 11, 165-179.
(6) (a) Spada, C. S.; Nieves, A. L.; Woodward, D. F. Exp. Eye Res. 2002,
75, 155-163. (b) Nials, A. T.; Coleman, R. A.; Hartley, D.; Sheldrick, R.
L. G. Br. J. Pharmacol. 1991, 102, 24P. (c) Coleman, R. A.; Kennedy, I.;
Sheldrick, R. L. G. Br. J. Pharmacol. 1987, 91, 323P.
(7) Since the 1-1.3 ratios of the chlorides (LiCl, MgCl2) over PhMgBr
seem to be insufficient for complete conversion of PhMgBr to PhMgCl,
we presume that among two PhMgCl- and PhMgBr-CuCN complexes
being in equilibrium with each other, the former complex of higher
regioselectivity reacts faster than the latter. The following facts support
this hypothesis: reaction with PhMgCl was completed at -18 °C for 1 h,
while reaction with PhMgBr at -18 °C for 1 h was incomplete, affording
a mixture of the starting monoacetate 1 and the products 2a and 3a.
184
Org. Lett., Vol. 7, No. 2, 2005