.
Angewandte
Communications
Iron-catalyzed cross-coupling reactions have been sug-
gested to proceed by an oxidative addition of a low-valent
affecting the yield of 2 (Table 1, entry 6). However, the
replacement of [Co(acac)3] by [Co(acac)2] decreased the yield
of 2 (54%) (Table 1, entry 7). On the basis of these results, we
selected [Co(acac)3] as the cobalt source, as it is less
hygroscopic than CoCl2, and tmeda, which is cheaper than
N,N’-tetramethyl-trans-cyclohexanediamine or dppe, as the
additive.
À
iron intermediate into the C1 Br bond, a process which
involves the formation of a carbon-centered radical.[8,11] As
a carbon-centered anomeric radical intermediate can be
hypothesized to account for the observed a selectivity
(Scheme 2),[4b,6,12] and as carbon-centered radical intermedi-
ates were also suggested to occur in cobalt-catalyzed cross-
coupling reactions between alkyl halides and Grignard
reagents,[13,14] we turned our attention to the cobalt-catalyzed
coupling of O-acetyl-a-bromo-d-mannose (1) with PhMgBr
(Table 1). Moreover, as cobalt-catalyzed cross-coupling reac-
tions between aryl Grignard reagents and alkyl halides do not
suffer from b-H elimination,[13] and tolerate a-oxygenated
substituents on alkyl halides,[14a,c–g,i–l] the formation of glucal 3
should be decreased.
The optimized conditions were applied to evaluate the
scope and limitations of the cobalt-catalyzed cross-coupling
of aryl Grignard reagents with 1-bromo glycosides 1, 4, and 6;
the results are reported in Table 2. With 1-bromo mannose
derivative 1 (Table 2, entries 1–8), the cross-coupling was
diastereoselective, as only the a isomer was detected. Elec-
tron-rich or electron-poor Grignard reagents led to similar
results (70–83% yield of isolated product; Table 2, entries 1–
5); however, bulkier Grignard reagents induced lower yields
(Table 2, entries 6 and 7). The cross-coupling was also
possible with an N-methylindole Grignard reagent, which
afforded the corresponding coupling product in 53% yield
(Table 2, entry 8). On the other hand, with o-CF3-PhMgCl, 2-
thienylmagnesium bromide, and an N-methylpyrazole
derived Grignard reagent, no cross-coupling products were
observed, and 1 was recovered (Table 2, entries 9–11). With
1-bromo glycoside 4, which is derived from glucose, the
coupling with aryl Grignard reagents was still diastereo-
selective in favor of the a isomers, which were obtained in
good yields, but with a lower a/b ratio (1.3:1–3:1; Table 2,
entries 12–16). When O-acetyl-a-bromo-d-galactose (6) was
used as the electrophile, the cross-coupling was diastereo-
selective (only the a isomer was detected) and the C-aryl
glycosides were obtained in moderate to good yields (Table 2,
entries 17–21).
Table 1: Optimization of the cobalt-catalyzed cross-coupling of PhMgBr
and 1-bromo glycoside 1.[a]
Entry
Cobalt source
(mol%)
Ligand
(mol%)
Yield of 2
3[c]
[%][b]
1
2
[Co(acac)3] (5)
CoCl2 (5)
–
–
10%
13%
trace
–
3
CoCl2 (5)
74%
–
4
5
CoCl2 (5)
dppe (5)
73%
76%
75%
54%
–
–
–
–
[Co(acac)3] (5)
[Co(acac)3] (1)
[Co(acac)2] (5)
tmeda (5)
tmeda (1)
tmeda (5)
6
As an extension of this study, and owing to the synthetic
value of C-vinyl glycosides, the reactivity of bromides 1, 4, and
6 with alkenyl Grignard reagents was evaluated (Table 3). We
were delighted to isolate the expected C-vinyl glycosides with
full a selectivity for bromides 1 and 6 (Table 3, entries 1, 2, 5,
and 6). In the case of 1-bromo glycoside 4, a moderate a/b
selectivity (1.5:1) was observed (Table 3, entry 4). Unexpect-
edly, vinylMgBr failed to react, and the starting bromide was
recovered (Table 3, entry 3).
7[d]
[a] Conditions: 1 (0.1m in THF), PhMgBr (1m in THF; 1.5 equiv) was
added at a rate of 2 mLminÀ1; once the addition was complete, the
reaction medium was warmed to RT. [b] Yield of isolated product.
[c] Detected by GC-MS. [d] 2.2 equiv of PhMgBr were used. acac=
acetylacetonate, dppe=1,2-bis(diphenylphosphino)ethane, tmeda =
N,N’-tetramethylethylenediamine.
The cobalt-centered catalytic system in the cross-coupling
of O-acetyl-a-bromo-d-mannose (1) and PhMgBr (1.5 equiv)
in THF was examined, and the results are reported in Table 1.
The use of [Co(acac)3] (5 mol%) in the absence of any ligand
led to product 2 in only 10% yield with traces of glucal 3
(Table 1, entry 1), and the use of CoCl2 (5 mol%) led to 2 in
only 13% yield (Table 1, entry 2). With the combination of
CoCl2 and N,N’-tetramethyl-trans-cyclohexanediamine as
a ligand, (5 and 6 mol%, respectively),[14f] the expected
product 2 was obtained in 74% yield as a single a isomer
(Table 1, entry 3). Replacing N,N’-tetramethyl-trans-cyclo-
From a mechanistic point of view, the stereochemical
outcome of this cobalt-catalyzed cross-coupling seems to
support a radical pathway, as suggested in the literature.[13,14]
In our case, the formation of an anomeric radical intermediate
at C1 during the oxidative addition of low-valent cobalt
species would induce an a-selective cross-coupling.[4b,6,12] To
verify if a radical pathway is operational in the cobalt-
catalyzed cross-coupling of 1-bromo glycosides with aryl
Grignard reagents, the reactivity of d-olefinic 1-bromo glyco-
side 7 was examined (Scheme 3). Treatment of 7 with
PhMgBr under our reaction conditions [PhMgBr
(1.5 equiv), [Co(acac)3] (5 mol%), tmeda (5 mol%)] pro-
duced an epimeric mixture of bicyclic compound 8 (d.r. = 1:1),
which was isolated in 88% yield. This product results from the
formation of an anomeric radical that leads to a 5-exo-trig
cyclization followed by cross-coupling with PhMgBr, which
suggests that the cyclization is faster than the direct cross-
coupling.[16]
hexanediamine
with
1,2-bis(diphenylphosphino)ethane
(dppe; 5 mol%)[14l] led to a similar result (73% yield of 2;
Table 1, entry 4), and the use of [Co(acac)3] (5 mol%) and
tmeda (N,N’-tetramethylethylenediamine)[14a] (5 mol%)
afforded 2 in 75% yield as a single a isomer (Table 1,
entry 5).[15] The catalytic loading could be decreased to
1 mol% of [Co(acac)3] and tmeda without significantly
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 11101 –11104