(5 mmol) was added to Cp2ZrCl2 (2.5 mmol) at À78 °C.
Compound 1 (2.5 mmol) was next added, and the reaction
mixture slowly warmed to 0 °C. Subsequent ZrfZn
transmetalation using Zn(OTf)2 (2.5 mmol)7 provided
the allylzinc, which was allowed to react with an imine
(1 mmol) derived from ethanolamine (Table 1).
This method allowed the addition of allyl fragments
bearing different electron-donating moieties (Table 1, en-
tries 2, 3, and 6) but could also be applied to other groups,
such as phenyl (Table 1, entry 1), alkenyl (Table 1, entry 5),
and alkyl (Table 1, entries 4 and 7) ones. This approach
appears to be quite general, leading to the expected
compounds 2aÀg in good yield. The syn/anti selectivity is
typically moderate but could be improved with an imine
derived from a chiral aminoalcohol.2b,8
This regioselectivity reversal reflects an alternative
mechanistic pathway and may be considered as a case
of group tuning involving the O-ZrCp2Cl moiety. It was
thus thought that the formation of the linear product
could be favored by protecting the hydroxy group by a
noncovalent-Zn-bonding function. This α-reactivity has
been described in the literature with Cr9 and allylsilane10
but is rare with allylzincs.11 This aspect was thus fur-
ther studied using imines derived from 2-methoxyethy-
lamine (Table 2). The experimental procedure, Barbier
(Method A) or TaguchiÀHanzawa (Method B), was
chosen according to the structure of the initial allylzinc
derivatives.
A series of homoallylic amines were obtained in moder-
ate to good yields. In the case of allylzincs bearing an aro-
matic (Table 2, entries 1À6) or a heteroaromatic (Table 2,
entries 7À12) substituent, the nearly exclusive formation of
the linear isomer was observed, irrespective of the method
used. In the typical case of a dienylzinc reagent (Table 2,
entry 13), the linear adduct remained the major product;
however, the branched regioisomer was also observed. In
contrast, variations were observed with alkyl-substituted
allylzincs (Table 2, entries 14À16). Whereas the branched
product was the major isomer (Table 2, entry 15), or
even the unique reaction product (Table 2, entry 16)
with primaryR1 alkylgroups, itwasminor when R1 = i-Pr
(Table 2, entry 14).
Table 1. Allylation of Ethanolamine-Derived Imines
entry
R1
R2
2 (yield)
dr
1
2
3
4
5
6
7
Ph
Ph
Ph
Ph
Ph
Ph
2a (83%)
2b (84%)
2c (87%)
2d (72%)
2e (70%)a
2f (69%)
2g (64%)
4.2:1
3.6:1
4:1
3-furyl
The double bond in compounds 3 are almost all
E-configurated, except when substituted with a heteroaro-
matic fragment. In most cases, the two isomers can be
separated by simple column chromatography.
The opposite regioselectivity observed between conju-
gated and nonconjugated allylzinc derivatives might origi-
nate from the intrinsic reactivity of these two cases, which
is likely to differ (Figure 1).
N-Boc-3-indolyl
i-Pr
2.6:1
7:1
Styryl
N-Boc-3-indolyl
Ph-CH2
CH2OBn
Ph
4.9:1
2.2:1
a The regioisomer was also formed and isolated in 17% yield.
Additionally, since 2 equiv of the allylzinc reagent were
required for a complete reaction (one being consumed by
the hydroxy group), we next decided to test an imine which
was preliminarily deprotonated with a zirconocene hy-
dride. Surprisingly, the linear adduct was obtained as the
major compound in this case (Scheme 1).
However, theses differences only appear with imines
derived from an aminoether. In this case, the direct
1,2-addition at the α-position prevails leading to com-
pounds 3. This implies that the organometallic interacts
differently depending on the imine side chain (Figure 1).
The nonracemic version was initiated by using an
imine derived from phenylglycinol methyl ether. In this
case, the regioselectivity is not α-exclusive but is still in
favor of the linear product 3q (8:1), which is obtained as a
single diasteroisomer (Scheme 2).12 The formation of 3q is
consistent with a chelation-controlled cinnamyl addition,
with attack occurring at the more accessible re face
(Scheme 2).
Scheme 1
(9) Giammaruco, M.; Taddei, M.; Ulivi, P. Tetrahedron Lett. 1993,
34, 3635–3638.
(10) Wang, D.-K.; Zhou, Y.-G.; Tang, Y.; Hou, X.-L.; Dai, L.-X.
J. Org. Chem. 1999, 64, 4233–4237.
(11) (a) Bustos, F.; Gorgojo, J. M.; Suero, R.; Aurrecoechea, J. M.
(7) Several Zn(II) sources were tested, and the best results were
obtained with Zn(OTf)2.
ꢀ
Tetrahedron 2002, 58, 6837–6842. (b) Aurrecoechea, J. M.; Fernandez,
A.; Gorgojo, J. M.; Suero, R. Synth. Commun. 2003, 33, 693–702. (c)
Zhao, L.-M.; Zhang, S.-Q.; Jin, H.-S.; Wan, L.-J.; Dou, F. Org. Lett.
2012, 14, 886–889. (d) Wipf, P.; Pierce, J. G. Org. Lett. 2005, 7, 3537–
3540.
(12) The R configuration of the newly formed stereogenic center was
deduced by analogy with 7 (vide infra, Scheme 5) which was converted
into known compound 11 (vide infra, Scheme 6).
(8) For highly diastereoselective allylmetalation of an imine derived
from an aminoalcohol, see: (a) Roy, U. K.; Roy, S. Chem. Rev. 2010,
110, 2472–2535. (b) Jang, T. S.; Ku, I. W.; Jang, M. S.; Keum, G.; Kang,
S. B.; Chung, B. Y.; Kim, Y. Org. Lett. 2006, 8, 195–198. (c) Black, D. A;
Arndtsen, B. A. Org. Lett. 2006, 8, 1991–1993. (d) Babu, S. A.; Yasuda,
M.; Baba, A. Org. Lett. 2007, 9, 405–408. (e) Thirupathi, P.; Kim, S. S.
Tetrahedron 2010, 66, 8623–8628.
Org. Lett., Vol. 14, No. 12, 2012
3005