The methodology was extended to the functionalization
of cyclic enamides 7aꢀd (Scheme 3). In contrast with the
three-component reaction involving acyclic precursors, a
high level of diastereocontrol was observed (1H NMR)
with the trans-isomer generally predominating, as shown
by the X-ray structure determination of compound 8e
obtained from xanthate 1b and amide 7b. In line with
previous observations, yields dropped with the introduc-
tion of a chain R to the nitrogen center (e.g., 8d).
Reaction with amides 7aꢀd allowed various substituents
on the nitrogen. Reaction of amide 7c afforded the addi-
tion product 8c, albeit in modest yield but high diaste-
reocontrol, without a trace of the 5-exo-trig cyclization
product, indicating that the intermolecular trapping of
the radical intermediate IIb (Scheme 1) by sulfone 3 is
faster than the cyclization onto the butenyl fragment. The
matched polarity between radical IIb and vinylsulfone 3 is
likely at the origin of this reactivity (vide infra).
overbalance the natural tendency of the C-centered radical
to react at the less hindered site.
Scheme 4. Carbo-alkenylation of Cyclic Ene-carbamates 9aꢀe
Carbo-alkenylation of cyclic ene-carbamates led to si-
milar observations (Scheme 4). For instance, the presence
of an alkyl chain R to nitrogen in 9a,b led to modest yields
of the three-component adducts 10a,b, having the trans-
relative configuration (1H NMR). Varying the size of the
ene-carbamate ring led to contrasting results.
As mentioned above, enamides and ene-carbamates are
electron-rich olefins that favor the addition of electron-
deficient radical species such as those derived from
xanthates 1a,b. These additions generate a reactive radical
such as IIb,12 which possesses an allylic character13 as
shown by the canonical forms below (Scheme 5). Ab initio
calculations at the DFT level14 indicates that spin density
in IIb is mainly concentrated R to the nitrogen as in IIb-i.
Scheme 3. Carbo-alkenylation of Cyclic Enamides 7aꢀd
·
The free rotation around the C ꢀN bond and thus the
absence of allylic strain (A1,3), as in resonance form IIb-ii,
would thus explain the low level of stererocontrol observed
in acyclic series (Table 1 and Scheme 2). This contrasts
with the high stereocontrol observed in radical addition
to related enamine systems,11 where A1,3 was invoked.
Computational studies have also been performed on
the reaction between radical intermediate IV (arising from
the addition of 1b onto olefin 9d) and vinylsulfone 3
(Scheme 5). The lowest energy conformation, which ex-
·
hibits a nearly planar C NCdO system, clearly shows that
While excellent yield was obtained with 6-membered
ring carbamate 9d, modest and low yields were observed
respectively for the 7- and 5-membered ring systems 9e and
9c. The formation of 10f indicates that the presence of a
substituent (Me) at C5 is not sufficient to control the
diastereofacial selectivity during the attack of the xanthate
precursor, while the trans-configuration between C2 and
C3 is again observed. Finally, the formation of 10g, albeit
in low yield, is noteworthy and is the result of the addition
of a radical species at the most sterically hindered site of a
trisubstituted olefin, indicating that a strong driving force
the top face is hindered by the ketone chain (pseudoaxial),
forcing the sulfone to approach through the bottom face.
This transition state proved to be in excellent agreement
with experiment, as 10h arising from the three-component
reaction between 9d, 1b, and 3 was formed with complete
diastereocontrol (Scheme 6).
(12) Electron-rich radicals such as IIb react very rapidly with vinyl-
sulfone 3, leading to the expected adduct, without a trace of the xanthate
transfer product (e.g., quenching the reaction between 1a, 3 and 9d after
10 min led to 30% of the expected 10d and no xanthate transfer product;
see the Supporting Information). Such transfer products are occasion-
ally found7c and are likely intermediates in these reactions when using
less reactive olefins.
(13) Quiclet-Sire, B.; Zard, S. Z. Org. Lett. 2008, 10, 3279–3282.
(14) DFT calculations were performed with the Gaussian09 software
package (Supporting Information). The M06-2X exchange correlation
functional was used with a standard double-ζ 6-31þG(d,p) basis set.
(10) Olefin 2g may have also lost its enamide character (overlap
between the olefin and the nitrogen lone pair) as a result of A1,3 strain
(we thank a reviewer for this comment).
(11) (a) Renaud, P.; Schubert, S. Angew. Chem., Int. Ed. Engl. 1990,
29, 433–435. (b) Schubert, S.; Renaud, P.; Carrupt, P.-A.; Schenk, K.
Helv. Chim. Acta 1993, 76, 2473–2489.
Org. Lett., Vol. XX, No. XX, XXXX
C