of 1a at room temperature could be achieved again using
catalytic amounts of both NaH (0.1 equiv) and 15-C-5 (0.2
equiv) as long as the reaction time was increased (entry 17).
As far as the mechanistic rationalization of the results is
concerned, we reasoned that the remarkable efficacy of these
new reaction conditions could be directly related to the
particular nature of the enolate counterion. Indeed, upon NaH
deprotonation, the malonamide substrates are expected to
give rise to planar, resonance-stabilized, Na-chelated eno-
lates, which are at the same time conjugated with the
intrinsically planar amide function. As a consequence, in the
ground-state conformation of the reactive intermediate, all
twelve9 atoms making up the Na-chelated malonamide
enolate will roughly lie in the same R plane (Scheme 2, A).
Scheme 3. Chelated vs Counterion-Free Pathway in the
Cyclization of 1 to 2
cated the different possible metal-free TSs at the B3LYP/
LACVP* level of theory in Jaguar v 4.2.13 The solvent was
simulated as a continuum using the PB-SCRF model14 in
Jaguar with parameters appropriate for dichloromethane.15
The relative energies for the eight possible TSs are shown
in Table 2.
Scheme 2
Table 2. Relative Energies and Length of the Forming Bond
for the Eight Possible TSs
On the other hand, in an ideal intermolecular enolate/η3-
allyl-PdX interaction, optimal orbital overlap requires a
parallel disposition between the plane containing the enolate
(R′) and that containing the allyl fragment (â). We can thus
infer that the transition state (TS) of the cyclization process
(Scheme 2, B) has to feature a significant loss of conjugation
with respect to the Na-chelated ground-state conformation.
As a consequence, we believe that the remarkable reactiv-
ity of these new cyclization conditions is likely to be due to
the generation of a highly reactive zwitterionic10 enolate/η3-
allyl intermediate, that can be achieved under either biphasic
PTC conditions or Na-sequestered homogeneous conditions
(Scheme 3).11
Erel
C-C bond length
entry syn/anti allyl endo/exo E/Z (kJ/mol)
(Å)
1
2
3
4
5
6
7
8
syn
syn
syn
syn
anti
anti
anti
anti
exo
exo
endo
endo
exo
exo
endo
endo
E
Z
E
Z
E
Z
E
Z
0.0
12.5
81.9
100.0
5.1
20.9
53.6
65.9
2.62
2.69
2.79
2.78
2.63
2.71
2.58
2.60
One immediately notices the large energy difference
between endo and exo cyclization. In all cases, the exo
cyclization is favored by more than 30 kJ/mol compared to
the most facile of the endo cyclization pathways. In Figure
1 are shown the two most accessible TSs, corresponding to
entries 1 and 5 in Table 2.
Compared to our earlier model calculations, which in-
cluded the Na+ counterion, the transitions states are earlier
and more reactant-like. The forming bond has a length of
over 2.6 Å in the TS, as compared to ca. 2.4 Å when the
Na+ counterion was included (Figure 1).
To support the above speculation, and in line with our
earlier contribution in this area,12 we have performed a
computational model study. Accordingly, we have used a
small model system for the Pd-allyl enolate intermediate,
(H3P)2Pd[CH2CHCH]-CH2-NMe-CO-CHCHO, and lo-
(9) Methyl groups in methyl esters are known to favor a planar s-cis
conformation. See for example: Stereoelectronic Effects in Organic
Chemistry; Deslongchamps, P., Ed.; Pergamon Press: Oxford, 1983.
(10) At present we cannot rule out the alternative ionization by anionic
palladate species [L2Pd(OAc)]Na, which would result in a completely naked
enolate instead of the zwitterionic form, in line with the observations of
Amatore, Jutand, and co-workers: Amatore, C.; Jutand, A.; M’Barki, M.
A.; Meyer, G.; Mottier, L. Eur. J. Inorg. Chem. 2001, 873-880. However,
the reaction does not seem to be sensitive to the presence of halides; cf.
entries 5 and 6, Table 1.
This indicates that the free anion is inherently more
reactive than the chelated complex, as expected. Looking
more closely at the geometries of the TSs, we see that
(11) Deprotonation of the pronucleophile under biphasic PTC conditions
is expected to take place in the interfacial region. (a) Makosza, M.; Krylowa,
I. Tetrahedron 1999, 55, 6395-6402. (b) Mason, D. Magdassi, S.; Sasson,
Y. J. Org. Chem. 1990, 55, 2714-2717. (c) Gobbi, A.; Landini, D.; Maia,
A.; Petricci, S. J. Org. Chem. 1998, 63, 5356-5361. (d) Lygo, B.; Andrews,
B. Acc. Chem. Res. 2004, 37, 518-525.
(13) Jaguar 4.2; Schro¨dinger, Inc.: Portland, OR, 2000. See: http://
(14) Marten, B.; Kim, K.; Cortis, C.; Friesner, R. A.; Murphy, R. B.;
Ringnalda, M. N.; Sitkoff, D.; Honig, B. J. Phys. Chem. 1996, 100, 11775-
11788.
(15) Solvent model is essential for the success of the calculations; without
it, the ring closure occurs without a barrier: Hagelin, H.; A° kermark, B.;
Norrby, P.-O. Chem. Eur. J. 1999, 5, 902-909.
(12) Norrby, P.-O.; Mader, M. M.; Vitale, M.; Prestat, G.; Poli, G.
Organometallics 2003, 22, 1849-1855.
Org. Lett., Vol. 7, No. 6, 2005
997
Madec, David
