.
Angewandte
Communications
[a]
ꢀ
Table 1: Optimization studies of the C N bond formation.
Entry
Additive
Result[b]
1
2
3
4
5
6
none
4 recovered
4 recovered
SnCl2
CuCl2
CuCl
B(OMe)3
ZnCl2
5 (trace) + 4 (major)
5 (trace) + 4 (major)
5 (9%) + 4 (16%) + unidentified by-products
5 (54%) + 4 (8%)
[a] Reaction conditions: nBuLi (2.2 equiv), THF, 0 8C, 2 h, then additive
(1.1 equiv) 1 h, then I2 (1.3 equiv), 0 8C!23 8C, 2 h. All reactions were
conducted at a concentration of 0.3m of 4 in THF. [b] Yield determined
by 1H NMR analysis using hexamethylenetetramine (HMTA) as an
internal standard. THF=tetrahydrofuran.
Scheme 2. The prepared isoindolines. Reaction conditions: Substrate
in THF (0.075m), nBuLi (2.2 equiv), 08C, 2 h, then ZnCl2 (1.1 equiv),
1 h, then, I2 (1.3 equiv); warm to 238C over 2 h. Because of the
inherent instability of the isoindoline products, the yields were
determined by H NMR analysis using HMTA as an internal standard.
[a] The amine precursor to isoindoline 8 was found to be volatile,
which may account for the reduced yield of 8. THF=tetrahydrofuran.
generated from 1)[14] and furthermore, could facilitate depro-
tonation of the benzylic methyl group by breaking down
organolithium aggregates.[15] To that end, we investigated
several additives. Stannous chloride, cupric chloride, and
cuprous chloride were ineffective and led mostly to the
recovery of the starting material (entries 2–4). However,
trimethyl borate (see entry 5), which we investigated as an
1
ꢀ
additive on the basis of the successful oxidative C N bond-
forming reaction reported by Shenvi and co-workers,[16] led
mostly to decomposition but gave an encouraging 9% yield of
the product, along with the starting amine 4 (16% recovery).
Pleasingly, ZnCl2 proved to be effective as an additive and led
to the formation of isoindoline 5 in 54% yield (as determined
by 1H NMR analysis)[17] along with some recovered 4 (8%). A
slight improvement in the yield of 5 (59% yield; 68% yield
based on remaining starting material) could be obtained by
conducting the reaction at a lower concentration of 0.075m in
THF. Although the balance of the material was mostly
accounted for by nonspecific decomposition, trace amounts of
imine and iodinated by-products were formed as well.
Generally, the yield of isolated 5 was low because of the
instability of the isoindoline moiety.
groups relative to methyl groups by using alkyllithium
bases.[18–20] N-Aryl substrates yield only small amounts of
the corresponding isoindolines (see 16), thus indicating
a deleterious effect of N-aryl substitution on the reaction.
Overall, the formation of isoindolines by the C,N-dianion
oxidation protocol proceeds in modest to good yield, which is
partly attributable to the stated inherent instability of
3
alkylated isoindolines.[21] To our knowledge, these C(sp ) N
ꢀ
couplings cannot be effected using any other established
methods.
3
ꢀ
The one-pot oxidative C(sp ) N coupling illustrated in
Scheme 2 spurred an investigation of the dianion formation/
oxidation in a more challenging system involving amide
substrates that would generate less nucleophilic lithium
amides.[22] With our established reaction conditions, N-
acylated isoindolines 17 and 18 (Scheme 3) were formed in
good yields. Unlike the alkylated isoindoline compounds
(Scheme 2), these acylated variants could be purified without
substantial loss of material. Furthermore, isoindolinones (19–
23a) were also formed in modest to good yield. Notably,
products such as 21 were formed without competing ortho
lithiation of the dimethoxybenzene moiety and a methylene
As illustrated in Scheme 2 (only products are shown), the
optimal conditions identified in Table 1, entry 6, with ZnCl2 as
the additive, are applicable to a range of tolyl substrates and
give the corresponding isoindoline products. In the case of 6,
a tertiary amine group, which could lead to competing
lithiation elsewhere, is tolerated. In addition, lithiation of
the pendant phenyl group present in 7 does not appear to
significantly compete with the desired reaction. Steric bulk
proximal to the secondary amine group does not adversely
affect the efficiency of the reaction (compare yields of 5–7 and
8–11). Additionally, substituents on the benzenoid portion of
the substrates are tolerated, including an alkyl group (see 12),
which could undergo lateral deprotonation, and a methoxy
group (see 13), which could promote ortho lithiation. Under
ꢀ
C H bond could be functionalized with good fidelity (see
isoindolinone 23b).
Using our protocol, even substrates that form six- and
seven-membered rings (Scheme 4), the formation of which is
not possible using the HLF or Suꢀrez reactions, participate in
3
ꢀ
the oxidative C(sp ) N coupling. For example, N,N-dialkyl
the current conditions,
amine 24a yields seven-membered azacycle 25a in 51% yield
under our standard reaction conditions, whereas amide 24b
affords 25b in 46% yield (as determined by NMR spectros-
copy).[23] Notably, an attempted HLF reaction with the N-
3
ꢀ
C(sp ) N coupling involving a methylene group proceeds in
reduced yield (see 15). This result likely points to the
associated kinetic challenge in deprotonating methylene
2
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
These are not the final page numbers!