γ- and δ-lactams through palladium-catalyzed intramolecular allylic alkylation: Enantioselective synthesis, NMR investigation, and DFT rationalization

Article abstract of DOI:10.1002/chem.201001300

The Pd-catalyzed intramolecular allylic alkylation of unsaturated amides to give γ- and δ-lactams has been studied in the presence of chiral ligands. Ligand (R)-3,5-tBu-MeOBIPHEP (MeOBIPHEP=6,6'-dimethoxybiphenyl-2,2- diyl)bis(diphenylphosphine)) afforded the best results and allowed the cyclization reactions to take place in up to 94:6 enantiomeric ratio. A model Pd-allyl complex has been prepared and studied through NMR spectroscopic analysis, which provided insight into the processes responsible for the observed enantiomeric ratios. DFT studies were used to characterize the diastereomeric reaction pathways. The calculated energy differences were in good agreement with the experimentally observed enantiomeric ratios. A transient existence: The Pd-catalyzed intramolecular allylic alkylation of unsaturated amides to give γ- and δ-lactams in the presence of (R)-3,5-tBu-MeOBIPHEP takes place in up to 94:6 enantiomeric ratio (e.r.; see scheme). The energies of the diastereomeric transition states are in good agreement with the experimentally observed enantiomeric ratios.

Full text of DOI:10.1002/chem.201001300

FULL PAPER
DOI: 10.1002/chem.201001300
g- and d-Lactams through Palladium-Catalyzed Intramolecular Allylic
ACHTUNGTRENNUNG lkylation: Enantioselective Synthesis, NMR Investigation, and DFT
AHCTUNGTREGRNNUN ationalization
Xavier Bantreil,[a] Guillaume Prestat,*[a] Aitor Moreno,[b] David Madec,[a]
Peter Fristrup,[c] Per-Ola Norrby,[d] Paul S. Pregosin,[b] and Giovanni Poli*[a]
Dedicated to Professor Alberto Brandi on the occasion of his 60th birthday
Abstract: The Pd-catalyzed intramolec-
ular allylic alkylation of unsaturated
amides to give g- and d-lactams has
been studied in the presence of chiral
ligands. Ligand (R)-3,5-tBu-MeOBI-
take place in up to 94:6 enantiomeric
ratio. A model Pd–allyl complex has
been prepared and studied through
NMR spectroscopic analysis, which
provided insight into the processes re-
sponsible for the observed enantiomer-
ic ratios. DFT studies were used to
characterize the diastereomeric reac-
tion pathways. The calculated energy
differences were in good agreement
with the experimentally observed enan-
tiomeric ratios.
PHEP
(MeOBIPHEP=6,6ꢀ-dimeth-
Keywords: cyclization
·
density
oxybiphenyl-2,2-diyl)bis(diphenylphos-
phine)) afforded the best results and
allowed the cyclization reactions to
functional calculations · enantiose-
lectivity · lactams · palladium
Introduction
In 1998, we reported an intramolecular palladium-catalyzed
route to 3,4-substituted g-lactams.[1] The cyclization process,
which is based on the reaction between a resonance-stabi-
lized carbanion and an allylic acetate linked by an amide
function, is completely regio- and stereoselective, thus lead-
ing exclusively to the trans product through a 5-exo process
(Scheme 1). This strategy turned out to be very general,
which we later applied to the synthesis of various racemic
natural or nonnatural compounds of biological interest.[2]
Scheme 1. Intramolecular palladium-catalyzed route to 3,4-substituted g-
lactams. EWG=electron-withdrawing group.
The development of an asymmetric version of this cycliza-
tion process to access g-lactams in enantio-enriched form
was, therefore, considered as the next challenge. We report
herein details of an investigation devoted to this subject.[3]
Since the first example of asymmetric allylic alkylation
(AAA) reaction reported by Trost and Strege[4] in 1977, this
class of reaction has been extensively studied.[5] The success
of this methodology relies on the ease with which palladium
tolerates various functional groups and the ability to use a
[a] Dr. X. Bantreil, Dr. G. Prestat, Dr. D. Madec, Prof. G. Poli
UPMC Univ Paris 06
Institut Parisien de Chimie Molꢁculaire (UMR CNRS 7201)
FR2769, 4, Place Jussieu, 75252, case 183, 75005, Paris (France)
Fax : (+33)144-277-567
E-mail: guillaume.prestat@upmc.fr
giovanni.poli@upmc.fr
[b] Dr. A. Moreno, Prof. P. S. Pregosin
Laboratory of Inorganic Chemistry
ꢀ
ꢀ
ꢀ
ꢀ
variety of nucleophiles to form new C C, C O, C N, C S,
ETHZ HCI Hçnggerberg 8093 Zꢂrich
Wolfgang Pauli Strasse 10, 8093 Zꢂrich (Switzerland)
ꢀ
and C H bonds. With stabilized carbanion nucleophiles, it is
possible to control the stereochemistry both in the allyl
group and in a prochiral nucleophile.[6] However, the intra-
molecular version of the AAA reaction is much rarer.[7–11]
The lack of successful examples in this field may stem
from the mechanism of the reaction itself (illustrated with
our substrate in Scheme 2). Oxidative addition of (Z)-1a
with a Pd0 complex gives rise to the h3-allylic anti-A com-
plex, which might cyclize or equilibrate through a p-s-p
mechanism to the h3-syn-B complex (Scheme 2). Alterna-
[c] Dr. P. Fristrup
Department of Chemistry
Technical University of Denmark
Kemitorvet, Building 201, 2800 Lyngby (Denmark)
[d] Prof. P.-O. Norrby
Department of Chemistry
University of Gothenburg
Kemigꢃrden 4, 41296 Gçteborg (Sweden)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201001300.
Chem. Eur. J. 2011, 17, 2885 – 2896
ꢄ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2885

Scheme 2. Reagents and reaction conditions. i) BSA (1.2 equiv), KOAc
(10 mol%), [Pd2A(dba)3] (5 mol%), PPh3 (0.5 equiv), THF, reflux (69%);
ii) BSA (1.2 equiv), KOAc (10 mol%), [{Pd
(h3-C3H5)Cl}2] (2.5 mol%),
CHTUNGTRENNUNG
AHCTUNGTRENNUNG
dppe (7.5 mol%), CH2Cl2, RT, 2 h, (85%). BSA=N,O-bis(trimethylsilyl)-
acetamide, dba=dibenzylideneacetone, dppe=1,2-bis(diphenylphosphi-
no)ethane.
Figure 1. Models of syn- and anti-cyclizing transition states.
tively, the latter complex may be directly generated from
(E)-1a. It is noteworthy that these intermediates are diaste-
reomeric isomers and, in the presence of a chiral ligand,
might lead to lactam 2 with different enantioselectivities.
Results and Discussion
In the course of a preceding study, we applied computation-
al methods to this reaction and located the transition states
(TS) that arise from models of A and B.[12,13] In particular,
the syn-TS of type B was lower in energy than the analo-
gous anti isomer A by 13 kJmolꢀ1, as calculated by DFT
methods (Figure 1). Although we lacked information about
the activation barriers associated with the generation of
each isomeric h3-allyl complex and that related to anti–syn
isomerization, we speculated that (E)-1a might cyclize
under milder conditions than (Z)-1a.
Indeed, after a short optimization, we found that when
(E)-1a was treated with [{PdACHTNUGRTNEUNG
(h3-C3H5)Cl}2] (2.5 mol%),
Figure 2. Chiral ligands tested for the AAA reaction.
dppe (7.5 mol%) as the ligand, and the couple BSA/KOAc
(1.2 equiv and 10 mol%, respectively) as the base, the reac-
tion could be carried out in CH2Cl2 at room temperature to
yield the desired pyrrolidone in 85% yield as a single trans
diastereomer (Scheme 2, step ii). These conditions are defi-
nitely more favorable than those required to cyclize (Z)-1a
(i.e., heating to reflux in THF; Scheme 2, step i).
Such a different behavior between isomers (Z)-1a and
(E)-1a strongly suggests that under the above reaction con-
ditions (i.e., CH2Cl2 at room temperature) no anti-to-syn
equilibration takes place because only the syn h3-allyl com-
plex can cyclize. These new conditions set the stage for
screening chiral ligands.
The set of enantiopure ligands tested is depicted in
Figure 2. DIOP (L1),[14] tBu-PHOX (L2),[15] the Trost ligand
L3,[16] and a member of the Walphos family of ligands L4[17]
required thermal activation to reach reasonable conversion
of (E)-1a into 2 and none led to a significant enantiomeric
ratio (Table 1, entries 1–4). A Josiphos-derived ligand, that
is, (R)-(S)-Xyliphos (L5),[18] and BINAP (L6)[19] allowed sat-
isfactory conversions at room temperature, with the latter
affording a promising enantiomeric ratio of 79:21 (Table 1,
entry 6).
2886
www.chemeurj.org
ꢄ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 2885 – 2896

g- and d-Lactams through the Allylic Alkylation of Unsaturated Amides
Table 1. AAA reaction of (E)-1a: screening of chiral ligands.[a]
FULL PAPER
Table 3. AAA reaction of (E)-1a: screening of atropoisomeric ligands.[a]
Entry
Ligand
T
Yield [%]
e.r.
Configuration[d]
Entry
Ligand
T
Yield [%]
e.r.[b,c]
1
2
3
4
5
6
L1
L2
L3
L4
L5
(R)-L6
reflux
reflux
reflux
reflux
RT
60
36
79
7
72
42
50:50[b]
57:43[b]
60:40[c]
50:50[c]
53:47[c]
79:21[c]
3R,4R
3R,4R
3S,4S
3R,4R
3R,4R
3R,4R
1
2
3
4
5
6
7
8
(R)-L6
(R)-L7
(R)-L7
(R)-L7
(R)-L10
(R)-L8
(R)-L11
(R)-L11
RT
RT
08C
reflux
08C
RT
42
65
44
50
41
50
73
67
79:21
86:14
86:14
84:16
86:14
83:17
83:17
83:17
RT
RT
reflux
[a] Reagents and reaction conditions: (E)-1a (1 equiv), BSA (1.2 equiv),
KOAc (10 mol%), [{Pd
(h3-C3H5)Cl}2] (2.5 mol%), ligand (7.5 mol%),
CH2Cl2, reflux, 2 h. [b] Enantiomeric ratios were determined by H NMR
spectroscopic analysis in the presence of [Eu(hfc)3] (hfc=3-(heptafluoro-
ACHTUNGTRENUNG
1
[a] Reagents and reaction conditions: (E)-1a (1 equiv), BSA (1.2 equiv),
KOAc (10 mol%), [{Pd
(h3-C3H5)Cl}2] (2.5 mol%), ligand (7.5 mol%),
N
ACHTUNGTRENNUNG
propyhydroxymethylene)-d-camphorate). [c] Enantiomeric ratios were
determined by chiral HPLC. [d] The absolute configuration was estab-
lished by correlation with (R)-N-benzoyl-3-ethylpyrrolidine[20] and com-
parison of its optical rotation (see the Supporting Information).
CH2Cl2, reflux, 2 h. [b] Enantiomeric ratios were determined by chiral
HPLC. [c] The absolute configuration of the major enantiomer was
3R,4R.
Solvent effects were tested using L6 as the ligand. Tolu-
ene, Et2O, and THF led, even at reflux, to poor conversion
and the enantiomeric ratios were lower than those obtained
in CH2Cl2 (Table 2, entry 1 versus entries 2–4). Replacement
of BSA/KOAc by Cs2CO3 led to an increase in yield, un-
fortunately accompanied by a decrease in enantiomeric
ratio (Table 2, entry 1 versus entry 5).
Table 2. AAA reaction of (E)-1a with (R)-BINAP: solvent effects.[a]
Entry
Solvent
T
Yield [%]
e.r.[b,c]
Figure 3. Array of C2-symmetric atropoisomeric ligands examined.
1
2
3
CH2Cl2
Et2O
THF
toluene
CH2Cl2
RT
42
14
25
17
83
79:21
–
57:43
64:36
68:32
reflux
reflux
reflux
RT
those reported earlier. Moreover, we have shown that under
these conditions (E)-1a could cyclize, and the ring closure
from the syn intermediate was favored by 5 kJmolꢀ1 over
that from the alternative anti intermediate (Figure 4). These
results indicated that, under the limit cases, the new condi-
tions feature either an enhanced reactivity of the h3-allylpal-
4
5[d]
[a] Reagents and reaction conditions: (E)-1a (1 equiv), BSA (1.2 equiv),
KOAc (10 mol%), [{Pd
(h3-C3H5)Cl}2] (2.5 mol%), (R)-BINAP (L6;
7.5 mol%), solvent. [b] Enantiomeric ratios were determined by
1H NMR with [Eu
(hfc)3]. [c] The absolute configuration of the major
AHCTUNGTRENNUNG
ACHTUNGTRENNUNG
isomer was 3R,4R. [d] Cs2CO3 (1.2 equiv) was used as the base.
Next, a selection of chiral C2-symmetric atropoisomeric li-
gands was tested (see Table 3 and Figure 3). All of these li-
gands afforded higher enantiomeric ratios (e.r.) than those
obtained with L6. The best result was obtained with (R)-
MeO-BIPHEP (L7)[21] at room temperature, which gave
86:14 e.r. in 65% yield (Table 3, entry 2). Raising the tem-
perature to allow reflux or cooling to 08C did not lead to
major changes (Table 3, entries 3 and 4). The use of DI-
FLUORPHOS (L10)[22] at 08C led to similar results, where-
as (R)-3,5-tBu-MeOBIPHEP (L8)[23] and SYNPHOS
(L11)[24] were less efficient (Table 3, entries 6–8).
We recently demonstrated that the palladium-catalyzed
allylic alkylation reaction of (Z)-1a could efficiently be run
in a biphasic medium.[25,2c] These conditions, which involve
counterion-free enolate intermediates, were milder than
Figure 4. Qualitative energy profiles under the biphasic conditions.
Chem. Eur. J. 2011, 17, 2885 – 2896
ꢄ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2887

G. Poli et al.
Table 4. AAA reaction of (E)-1 and (Z)-1 under biphasic conditions.[a]
ladium complexes with respect to the previous conditions
(Figure 4, path a), or, more likely, an anti–syn isomerization
barrier lower than that required for cyclization (Figure 4,
path b). This latter Curtin–Hammet-type case implies that
Entry
Substrate
Ligand
T
Yield [%]
e.r.[b]
1
2
3
4
5
6
7
8
9
(E)-1a
(E)-1a
(E)-1a
(E)-1a
(E)-1a
(Z)-1a
(Z)-1a
(Z)-1a
(Z)-1a
(R)-L6
(R)-L11
(R)-L7
(R)-L8
(R)-L8
(R)-L8
(R)-L8
(S)-L9
(S)-L9
RT
RT
RT
RT
08C
RT
08C
RT
08C
88
–
86:14[c]
–
ꢀ
the C C bond-formation occurs exclusively through the
most stable (syn) h3-allylpalladium complex independently
of the E or Z geometry of the starting substrate. Thus, we
decided to test some chiral ligands under these new condi-
tions.
80
77
47
62
56
66
43
86:14[c]
90:10[c]
92:8[c]
88:12[c]
92:8[c]
17:83[d]
15:85[d]
The AAA reaction of (E)-1a under biphasic conditions
was first tested with L6 (7.5 mol%) in the presence of [{Pd-
ACHTUNGTRENNUNG
(h3-C3H5)Cl}2] (2.5 mol%) as the palladium source, KOH
[a] (E)-1 or (Z)-1 (1 equiv), [{PdACHTUNGTRNEUNG
(h3-C3H5)Cl}2] (2.5 mol%), ligand
(7.5 mol%), nBu4NBr (10 mol%), KOHaq (2 equiv), CH2Cl2/H2O, RT,
2 h. [b] Enantiomeric ratio was determined by chiral HPLC. [c] The abso-
lute configuration of the major isomer was 3R,4R. [d] The configuration
of the major enantiomer was 3S,4S.
(2.0 equiv) as the base, and nBu4NBr (10 mol%) as the
phase-transfer reagent in a biphasic system CH2Cl2/H2O
(1:1, v/v). These conditions gave the expected g-lactam in
88% yield with a satisfactory 86:14 e.r. (Table 4, entry 1).
Surprisingly, SYNPHOS (L11) could not promote the reac-
tion, thus leaving the starting material unchanged (Table 4,
entry 2). The limited stability of this ligand under basic bi-
phasic conditions might be responsible for this result.
Ligand L7 led to the same result as the BINAP ligand in
terms of e.r. value but in slightly lower yield (Table 4,
entry 3). Finally, 90:10 e.r. was reached by using L8[23,26] with
77% yield (Table 4, entry 4). The e.r. value increased to 92:8
by carrying out the reaction at 08C, albeit with a decrease in
yield to 47% (Table 4, entry 5).
modifying the departing ability of the leaving group may
affect the energy of the transition states, thus leading to the
diastereomeric h3-allyl intermediates and the level of enan-
tioselectivity of the process. Furthermore, modification of
the loading of the palladium catalyst may affect the kinetics
of the exchange between these h3-allyl intermediates
(through p-s-p or a direct inversion mechanism). On the
Interestingly, (Z)-1a as the
starting substrate led to almost
similar results as (E)-1a in
terms of yield and enantioselec-
tivity (Table 4, entries 6 and 7
versus entries 4 and 5).[27] The
use of the more sterically de-
manding
(S)-3,5-tBu-4-MeO-
MeOBIPHEP (L9)[23] did not
improve the e.r. value (Table 4,
entries 8 and 9).
If we consider a three-dimen-
sional simplified qualitative
energy profile for the reaction
under study involving only the
syn-allyl complexes (wherein
the intermediate h3-allyl com-
plexes are higher in energy
than the starting substrate), two
limit cases may be distinguished
(Figure 5, cases 1 and 2).
ꢀ
If the C C bond formation is
faster that the ionization step,
the former step becomes irre-
versible and exchange between
the diastereomeric allyl com-
plexes becomes negligible; as a
result, the ionization step is
enantiodiscriminating (EDS) as
well as rate determining (RDS;
Figure 5, case 1). In this case, Figure 5. Qualitative energy profiles for the enantioselective cyclization reaction under study.
2888
www.chemeurj.org
ꢄ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 2885 – 2896

g- and d-Lactams through the Allylic Alkylation of Unsaturated Amides
FULL PAPER
ꢀ
Table 6. Influence of the EWG on the AAA reaction.[a]
other hand, if ionization is faster than the C C bond forma-
tion, rapid equilibration of the diastereomeric allyl com-
plexes through a retro reaction, p-s-p, or a direct inversion
mechanism[28] becomes possible and C C bond formation
ꢀ
becomes the EDS as well as the RDS (Figure 5, case 2). In
this case, the degree of enantioselectivity is not expected to
vary upon modification of the leaving-group ability or cata-
lyst loading.[9,29]
The influence of the nature of the leaving group was
tested by using ligand L8 (Table 5). Replacement of the ace-
tate leaving group in (E)-1a by a para-methoxybenzoate or
Entry
Substrate
EWG
Yield [%]
e.r.[b,c]
1
2
3
(E)-1a
(E)-1d
(E)-1e
CO2Me
COMe
SO2Ph
88
50
58
86:14
70:30
53:47
[a] Reagents and conditions: (E)-1 (1 equiv), [{PdACTHNUTRGNEUNG
(h3-C3H5)Cl}2]
(2.5 mol%), (R)-BINAP (7.5 mol%), nBu4NBr (10 mol%), aqueous
KOH (50%wt, 2 equiv), CH2Cl2/H2O = 1:1, RT, 2 h. [b] Enantiomeric
ratio was determined by chiral HPLC. [c] The absolute configuration of
the major enantiomer was 3R,4R.
Table 5. Study of the influence of the leaving group (LG) and the cata-
lyst loading on the AAA reaction.[a]
Entry
Substrate
LG
[Pd] [mol%]
Yield [%]
e.r.[b,c]
1
2
3
4
5
(E)-1a
(E)-1b
(E)-1c
(E)-1a
(E)-1a
OAc
PMB
OCO2Me
OAc
OAc
2.5
2.5
2.5
0.5
0.25
77
36
58
41
3
90:10
88:12
89:11
89:11
90:10
[a] Reagents and conditions: (E)-1a–c (1 equiv), [{PdACTHNUTRGNEUNG
(h3-C3H5)Cl}2]/(S)-
L8=1:3, nBu4NBr (10 mol%), aqueous KOH (50%wt, 2 equiv), CH2Cl2/
H2O, RT, 2 h. [b] Enantiomeric ratios were determined by chiral HPLC.
[c] The absolute configuration of the major isomer was 3S,4S. PMB:
para-methoxybenzoate.
Scheme 3. Synthesis of the homologated cyclization precursor 5.
a methyl carbonate (worse and better leaving groups with
respect to acetate, respectively) induced a decrease in the
yield without affecting the e.r. value (Table 5, entries 1–3).
Similarly, decreasing the loading of the palladium catalyst to
0.5 and 0.25 mol% did not affect the e.r. value (Table 5, en-
tries 1, 4, and 5). These results suggest a reversible ioniza-
tion and a rate-determining step as well as an enantiodiscri-
mine gave N-but-3-enylbenzylamine, which was acylated
with methyl malonyl chloride to yield precursor 4. Finally, a
cross-metathesis step involving allylic diacetate in the pres-
ence of 3 mol% of the Grubbs second-generation catalyst
allowed the formation of 5 in a 90:10 E/Z ratio.
With precursor 5 in hand, various conditions for the palla-
dium-catalyzed allylic alkylation reaction were tested in the
presence of the achiral, bidentate ligand dppe (Table 7). By
ꢀ
minating C C bond-formation step.
Finally, by using L6 as the
ligand and switching the elec-
tron-withdrawing group from
Table 7. AAA reaction of precursor 5.
methoxycarbonyl to phenylsul-
fone and acetyl brought about a
constant decrease in the enan-
tiomeric ratio of the cyclized
product (Table 6).
Our attention next focused
Entry
Conditions[a]
Pd source
Base
Solvent
CH2Cl2
CH2Cl2
THF
THF
CH2Cl2/H2O
Yield [%][b]
on the cyclization of precursor
5, homologated by one carbon
atom on the allylic moiety with
respect to 1. This new cycliza-
tion precursor was easily ob-
tained in three steps from com-
mercial products (Scheme 3).
1
2
3
4
5
A
A
A
A
B
[{Pd
[{Pd
[{Pd
U
BSA/KOAc
BSA/KOAc
NaH
NaH
KOH
56
67
41
52
87
[c]
Pd
A
[{Pd
A
[a] Conditions A: 5 (1 equiv), [Pd] (5 mol%), dppe (7.5 mol%), base (1.2 equiv), KOAc (0.1 equiv when
needed). Conditions B: 5 (1 equiv), [{Pd
(h3-allyl)Cl}2] (2.5 mol%), dppe (7.5 mol%), nBu4NBr (10 mol%),
aqueous KOH (50%wt, 2 equiv), CH2Cl2/H2O=1:1. [b] Compound 6 was obtained in a trans/cis ratio of 90:10.
bromo-1-butene and benzyla- [c] The reaction was carried out at reflux temperature of the solvent.
AHCTUNGTRENNUNG
Thus,
a reaction between 4-
Chem. Eur. J. 2011, 17, 2885 – 2896
ꢄ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2889

G. Poli et al.
using 5 and under the reaction conditions optimized for (E)-
1a, piperidinone 6 was obtained in 56% yield (Table 7,
entry 1). Increasing the temperature to allow reflux slightly
improved the yield to 67% (Table 7, entry 2), but changing
the solvent, base, and source of palladium did not allow a
higher conversion (Table 7, entries 3 and 4). Again, biphasic
conditions were very efficient, thus leading to the desired pi-
peridinone 6 in 87% yield. We note that 6 was always ob-
tained in a diastereomeric trans/cis ratio of 90:10. Such a
ratio appears to reflect a thermodynamic equilibrium, as
crude mixtures that consist of 50:50 trans/cis isomers isomer-
ized on standing to afford a 90:10 ratio of products.
Scheme 4. Synthesis of intermediate palladium complexes.
The chiral ligands from Figures 2 and 3 were evaluated in
the AAA reaction of 5 under the biphasic conditions. Once
again L1–L3 and L5 did not lead to satisfactory enantiomer-
ic ratios (Table 8, entries 1–3), whereas L6 gave a promising
Scheme 5. Synthesis of the cationic h3-allylphosphine–palladium salts.
Table 8. AAA reaction of precursor 5.[a]
The solution structure and dynamics of the chloro-bridged
1
complex 8 were investigated in CD2Cl2. The H NMR spec-
trum shows the presence of two sets of signals in a ratio of
approximately 2:1, which is ascribed to the presence of two
different geometric isomers 8A and 8B associated with the
Entry
Ligand
T
Yield [%][b]
e.r.[c]
1
2
3
4
5
6
7
8
L2
L3
L5
RT
RT
RT
RT
RT
08C
RT
08C
55
16
84
83
83
83
78
65
31:69
43:57
48:52
86:14
12:88
11:89
8:92
ꢀ
restricted rotation about the N C(=O) bond (Scheme 6).
(R)-L6
(S)-L7
(S)-L7
(S)-L8
(S)-L8
6:94
[a] Reagents and conditions: 5 (1 equiv), [{PdACTHNUTRGNEUNG
(h3-allyl)Cl}2] (2.5 mol%),
ligand (7.5 mol%), nBu4NBr (10 mol%), aqueous KOH (50%wt,
2 equiv), CH2Cl2/H2O=1:1. [b] Compound 6 was obtained in a trans/cis
ratio of 90:10. [c] Enantiomeric ratios were measured by chiral HPLC of
the major trans diastereomer (laevorotatory/dextrorotatory isomers); the
absolute configuration of 6 was not determined.
Scheme 6. Conformational population of dimer 8.
The proton spin–spin coupling patterns for the central
allyl protons b and b’ confirm that in both isomers the sub-
stituent on the terminal allyl carbon is in the syn position
(relative to the central allyl proton). The 1H,1H NOESY
spectrum of 8 shows a strong NOE interaction between the
two nonequivalent anti protons at the “a” and “c” positions
in keeping with this syn geometry. Figure 5 shows a section
of the spectrum of 8 recorded at 273 K. We note that in the
major conformer A the benzylic protons e show NOE inter-
actions with the methylene protons g, whereas in the minor
isomer B these are absent. The minor isomer B reveals a b’/
g’ interaction between the central allyl and methylene pro-
tons, respectively, which is absent in isomer A. These obser-
vations are in agreement with the proposed presence of the
two amide isomers A and B. Note that no exchange peaks
are present between these two species at 273 K (Figure 6).
Figure 7 shows a section of the phase-sensitive 1H,1H
NOESY spectrum of 8 recorded at ambient temperature.
The observation of relatively intense cross peaks concerned
e.r. value of 86:14 (Table 8, entry 4). The remaining C2-sym-
metric atropoisomeric ligands were tested. Ligand L8 gave
the best results again. Indeed, piperidinone 6 was obtained
in 78% yield with 8:92 e.r. in the presence of L8 (Table 8,
entry 7), which could be increased to 6:94 e.r. by carrying
out the reaction at 08C (Table 8, entry 8).
NMR spectroscopic studies: To obtain further insight with
respect to the factors responsible for the observed enantio-
selectivity, several model Pd salts were prepared. Specifical-
ly, acetate (E)-1a was converted into its chlorinated ana-
logue 7 and treated with [Pd2ACTHNUTRGNEUNG
(dba)3][30] to give the chloro-
bridged h3-allylpalladium complex 8 (Scheme 4).
In a subsequent step, the dinuclear species 8 was treated
with a variety of mono- and bis(phosphine) ligands in the
presence of AgBF4 to afford the corresponding cationic h3-
allylphosphine–palladium salts 9a–e (Scheme 5).
2890
www.chemeurj.org
ꢄ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 2885 – 2896

g- and d-Lactams through the Allylic Alkylation of Unsaturated Amides
FULL PAPER
Figure 7. Section of the 1H,1H NOESY spectrum of 8 in CD2Cl2 at ambi-
ent temperature. The exchange peaks (open circles) between the central
allyl protons b and b’ and benzylic protons e and e’, respectively, show
that the two isomers exchange.
1
Figure 6. Section of the H,1H NOESY spectrum of 8 in CD2Cl2 at 273 K.
with the two central allyl protons b and b’ proves that these
two isomers are exchanging. In addition, the diastereomeric
benzylic protons e1 and e2 in rotamer A show exchange
cross peaks to benzylic protons e1’ and e2’ in rotamer B. Fur-
thermore, the diastereomeric benzylic protons in each
isomer show intramolecular contacts (e1/e2 and e1’/e2’). This
study was helpful in understanding the solution behavior of
the phosphine salts 9a–e.
nance of which appears at approximately d=5.87 ppm,
there are two overlapping triplets. In a minor isomer, at ap-
proximately d=5.55 ppm two well-resolved triplets arise
from only one relatively large trans coupling constant, which
implies an anti-allyl geometry for this isomer. Note that the
presence of an anti isomer suggests an h3-h1-h3 mecha-
nism.[32]
The 31P NMR spectra in
CD2Cl2 for these bisphosphine
salts 9a–e share many similar
features. There is
a major
isomer composed of an AB
spin system in which 2J-
ACHTUNGTRENNUNG
(31P, 31P)=53–55 Hz for 9b and
9c and 2J
(31P, 31P)ꢁ43 Hz for
the PPh3 analogue 9a, which is
consistent with a cis arrange-
ment of the phosphine ligands.
The minor species (20% or
less) also feature AB spin sys-
tems with coupling constants
2J
ACHTUNGTRENNUNG
for a representative 31P NMR
spectrum).[31]
In the 1H NMR spectra for
9a–e, the splitting patterns of
the central allyl protons in the
major isomers (assuming that
3Jtrans >3Jcis) are indicative of a
syn-allyl
arrangement
Figure 8. 31P NMR spectrum of 9b in CD2Cl2 that shows the different AB (or AX) spin systems present in so-
(Figure 9). Specifically, in the
major isomer of 9b, the reso-
lution.
Chem. Eur. J. 2011, 17, 2885 – 2896
ꢄ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2891

G. Poli et al.
In summary, the treatment
of the allyl chloride 7 with
[Pd2ACTHNUGRTNEUNG(dba)3] in dichloromethane
leads to the corresponding di-
nuclear palladium complex,
present as two amide conform-
ers that exchange at ambient
temperature. In both of these
isomers, the substituent on the
allyl unit is in the syn position.
The corresponding phosphine
allyl salts 9a–e are relatively
stable and reveal a number of
species in solution. The major
isomers in CD2Cl2 have a syn-
allyl arrangement and the
amide carbonyl function is in a
pseudo-trans position relative
to the benzyl substituent. The
minor isomers that are present
are in mutual exchange; fur-
1
Figure 9. Section of the H NMR spectrum of 9e in CD2Cl2 that shows the splitting patterns of the central allyl
protons in the major and minor isomers.
1
As expected, the H,1H NOESY spectra of the major spe-
cies in 9a–e show NOE interactions between the two none-
quivalent anti protons of the allyl ligand. In addition, all of
the major isomers show NOE interactions between the
methylene protons g and the benzylic protons e, thus indi-
cating the presence of rotamers of type A, that is, the amide
carbonyl function in a pseudo-trans position relative to the
benzyl substituent. In summary, on the basis of the NMR
spectroscopic data, the major species present in solution 9a–
e in dichloromethane are in agreement with structures of
type A (Scheme 6).
Due to the low concentration of the minor isomers and
1
the strong overlap of the resonances in the H NMR spectra
of 9a–e, even at high B0 fields, it was not possible to com-
pletely characterize these species. However, inspection of
1
the H NMR spectra in the region of the central allyl proton
resonances reveals that for all of the salts there are 2–3 ad-
1
ditional minor syn and anti isomers.[33] The H,1H NOESY
spectra of 9a–e at ambient temperature show that all the
isomers and specifically the syn and anti isomers are in
mutual exchange (Figure 10).
Finally, 9b was tested in the cyclization chemistry. Treat-
Figure 10. Section of the 1H,1H exchange spectroscopy (EXSY) map of
9e that shows the molecular exchange between the isomers.
ment of this complex with BSA (1.0 equiv) and KOAc
(1.0 equiv) in CH2Cl2 for 2 h at room temperature gave the
expected pyrrolidone 2 in 38% yield with 70:30 e.r. (3R,4R).
Although the enantioselectivity is poorer with respect to
that obtained in the catalytic experiment (compare with
Table 1, entry 6; 79:21 e.r.), this result suggests that the h3-
allylpalladium complexes 9b–e mimic the behavior of the
species present during catalysis. As a consequence, the infor-
mation deducible from the analysis of the stoichiometric
complexes, most notably that concerning the stereochemical
outcome, can be reasonably transposed to the catalytic ver-
sion (Scheme 7).
Scheme 7. Cyclization of the h3-allylpalladium complex 9b
2892
www.chemeurj.org
ꢄ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 2885 – 2896

g- and d-Lactams through the Allylic Alkylation of Unsaturated Amides
FULL PAPER
thermore, at least one of these
has an anti arrangement of the
allyl substituent.
Computational investigation:
To add further support to the
proposed reaction mechanism,
a theoretical investigation of
ꢀ
the crucial C C bond forma-
tion was carried out. In line
with earlier work on allylpalla-
dium complexes,[34] we em-
ployed DFT in combination
with the B3LYP functional[35]
as incorporated in Jaguar.[36]
The LACVP* basis set was
employed, which entails a 6-
31G* basis set for light atoms
and an effective core potential
(ECP) and basis set for palla-
dium.[37] In the palladium-cata-
lyzed allylic alkylation reac-
tion, solvation effects can have
a large influence on the struc-
ture of the transition states;
therefore, we employed a po-
larized Boltzmann self-consis-
tent reaction field (PB-
SCRF[38]) with parameters suit-
able for dichloromethane. Ini-
tially, the reaction was investi-
gated using PH3 model ligands
on palladium (similar to our
earlier work).[25a] However, for
the final evaluation of the pos-
ꢀ
sible reaction paths the full
BINAP ligand was included to
evaluate the energy difference
between diastereomeric reac-
Figure 11. Overview of the four different diastereomeric ring-closing conformations optimized with a fixed C
C bond distance of 2.5 ꢅ resulting in the formation of 2.
tion paths. Accordingly, both the enantiomeric forms of the
transition states obtained using model phosphine ligands
In all cases, the reaction was exothermic by approximately
150 kJmolꢀ1 when decreasing the C C distance from 3.5 to
ꢀ
were combined with an optimized {(R)-Pd
N
1.5 ꢅ. During these investigations, it became clear the
B3LYP functional did not adequately describe the nonbond-
ed interactions responsible for discriminating between the
different possible reaction pathways. This deficiency of the
B3LYP functional is well known and the problem has been
addressed by either appending a classic dispersion term[40] or
by using a functional that incorporates kinetic energy-densi-
ty terms.[41] Among the most successful of the latter ap-
proaches are the M0x family of functionals reported by
Zhao and Truhlar.[42] Herein, we chose the M06 functional,
which was optimized with particular focus on organometallic
systems.[42]
ment. The resulting complexes were optimized while keep-
ꢀ
ing only the forming C C bond fixed. The presence of the
chiral ligand did not change the relative energies of the
ring-closing TS significantly, which allowed the investigation
to be limited to only the four most stable exo-TSs, that is,
those involving interaction between an E-configured enolate
and the syn- or anti-configured allyl moieties. The initially
located TSs were further optimized by using a transition-
state force-field, which allowed a conformational search of
the full substrates to be performed. The best conformers
that lead to either the 3R,4R or 3S,4S enantiomer of 2 sub-
mitted to a series of DFT minimizations while keeping the
For systems of this size and complexity, the location of
the transition states with implicit solvation models is nontri-
vial. In our experience, more reliable energy differences can
be obtained in the allylic alkylation by using a minimization
ꢀ
developing C C bond fixed to 1.5, 2.0, 2.5, 3.0, and 3.5 ꢅ.
ꢀ
The resulting optimized structures with a C C distance of
2.5 ꢅ are shown in Figure 11.[39]
Chem. Eur. J. 2011, 17, 2885 – 2896
ꢄ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2893

G. Poli et al.
ꢀ
with a fixed C C distance. In
ꢀ
these systems, a C C distance
of 2.5 ꢅ was close to the TS
and the energy difference be-
tween the two diastereomeric
transition states that arise
from the syn-allyl complex was
7.2 kJmolꢀ1, found by using
DFT/M06. This result is in
good agreement with the sense
and the extent of induction ob-
served
(84:16 e.r.
experimentally
corresponds to
4.1 kJmolꢀ1 at room tempera-
ture). The anti-allyl transition
states are significantly less fa-
vored. The most favorable
anti-allyl TS is approximately
14 kJmolꢀ1 higher in energy
than the best syn-allyl TS.[43,44]
In a Curtin–Hammett situa-
tion, the energies of the inter-
mediates become irrelevant,
and the selectivity is deter-
mined solely by the relative
energy of the competing tran-
sition states. In the current
case, reaction through the anti
manifold should, therefore, ac-
count for less than 1% of the
final product.
The investigation of the re-
action that leads to the forma-
tion of the six-membered ring
piperidone 6 was carried out in
a similar fashion as described
for 2. Also here the overall re-
action was exothermic by ap-
ꢀ
Figure 12. Overview of the four different diastereomeric ring-closing conformations optimized with a fixed C
C bond length of 2.5 ꢅ that results in the formation of 6.
proximately 150 kJmolꢀ1, and the energy differences be-
tween the four diastereomeric transition states were small
(Figure 12). Also in the formation of the six-membered ring,
the most favorable reaction pathway arises from a syn-con-
figured h3-allylpalladium complex. The most-favored syn
diastereomer is favored by 7.4 kJmolꢀ1 with respect to the
syn pathway that leads to the minor enantiomer. The simi-
larity to the syn-allyl complexes determined for the forma-
tion of the g-lactam 2 is striking; thus, we find it plausible
that the absolute configuration of 6 is 3R,4R in analogy to 2.
However, for the TS that arises from the anti-allyl complex,
there is a clear difference between the g- and d- lactams.
These complexes were energetically very unfavorable in the
formation of 2, whereas they are only 4–6 kJmolꢀ1 higher in
energy in the formation of 6 than the TSs that arise from
the syn-allyl complex. Fortunately, also for this pair of dia-
stereomers, the formation of (3R,4R)-6 is favored, which can
explain that a high enantioselectivity is also observed for
this substrate.
Conclusion
In summary, we have reported an AAA reaction that allows
the formation of g- and d- lactams in good yields with good
enantioselectivities. The use of the C2 symmetric atropoiso-
meric ligand 3,5-tBu-MeOBIPHEP under biphasic condi-
tions developed in our laboratory furnished pyrrolidinones
and piperidinones in 92:8 and 94:6 enantiomeric ratios.
A designed synthesis allowed the preparation, isolation,
and cyclization of the h3-allylpalladium complex expected to
form transiently in the catalytic experiments.
The observed enantioselectivity can be rationalized by
using DFT/M06 calculations on the full BINAP ligand and
complete substrates. The overall good agreement between
the calculated diastereo- and enantioselectivities and the ex-
perimentally observed values gives hope that computer-gen-
erated catalyst optimization is possible for these systems.
2894
www.chemeurj.org
ꢄ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 2885 – 2896

g- and d-Lactams through the Allylic Alkylation of Unsaturated Amides
FULL PAPER
inov, E. M. Grenzer, J. Am. Chem. Soc. 1997, 119, 7879–7880; d) R.
Kuwano, Y. Ito, J. Am. Chem. Soc. 1999, 121, 3236–3237; e) B. M.
Trost, G. M. Schroeder, J. Am. Chem. Soc. 1999, 121, 6759–6760;
f) S.-L. You, X.-L. Hou, L.-X. Dai, X.-Z. Zhu, Org. Lett. 2001, 3,
149–151; g) B. M. Trost, G. M. Schroeder, J. Kristensen, Angew.
Chem. 2002, 114, 3642–3645; Angew. Chem. Int. Ed. 2002, 41, 3492–
3495; h) R. Kuwano, K. Uchida, Y. Ito, Org. Lett. 2003, 5, 2177–
2179; i) B. M. Trost, W. Tang, J. Am. Chem. Soc. 2003, 125, 8744–
8745; j) B. M. Trost, C. Pissot-Soldermann, I. Chen, G. M. Schroed-
er, J. Am. Chem. Soc. 2004, 126, 4480–4481; k) D. C. Behenna,
B. M. Stoltz, J. Am. Chem. Soc. 2004, 126, 15044–15045; l) J. T.
Mohr, B. M. Stoltz, Chem. Asian J. 2007, 2, 1476–1491.
[7] K. Yamamoto, J. Tsuji, Tetrahedron Lett. 1982, 23, 3089–3092.
[8] a) J. P. GenÞt, S. Grisoni, Tetrahedron Lett. 1988, 29, 4543–4546;
b) N. Kardos, J. P. GenÞt, Tetrahedron: Asymmetry 1994, 5, 1525–
1533.
[9] G. Koch, A. Pfaltz, Tetrahedron: Asymmetry 1996, 7, 2213–2216.
[10] a) B. M. Trost, K. L. Sacchi, G. M. Schroeder, N. Asakawa, Org.
Lett. 2002, 4, 3427–3430; b) C. Butts, E. Filali, G. Lloyd-Jones, P-O.
Norrby, D. Sale, Y. Schramm, J. Am. Chem. Soc. 2009, 131, 9945–
9957; c) B. M. Trost, D. A. Thaisrivongs, J. Am. Chem. Soc. 2009,
131, 12056–12057 and ref. quoted.
[11] J. Bian, M. Van Wingerden, J. M. Ready, J. Am. Chem. Soc. 2006,
128, 7428–7429.
[12] P.-O. Norrby, M. M. Mader, M. Vitale, G. Prestat, G. Poli, Organo-
metallics 2003, 22, 1849–1855.
Experimental Section
General procedures for AAA reactions:
Allylic alkylation under homogenous conditions: The bidentate ligand
(11.7 mmol, 7.5 mol%) was added to a solution of the allylpalladium chlo-
ride dimer (1.4 mg, 3.9 mmol, 2.5 mol%) in dichloromethane (500 mL).
The solution was stirred at room temperature for 5 min. A solution of
acyclic substrate
1 or 5 (0.157 mmol, 1 equiv) in dichloromethane
(1.5 mL), BSA (46 mL, 0.188 mmol, 1.2 equiv), and potassium acetate
(1.6 mg, 15.7 mmol, 0.1 equiv) were successively added. The resulting so-
lution was stirred at the appropriate temperature until full completion
(monitoring by TLC). A saturated aqueous NH4Cl solution was added
and the aqueous phase was extracted with dichloromethane (3ꢆ). The
collected organic phases were washed with brine, dried over MgSO4, and
the solvent was removed in vacuo. The crude product was purified by
flash chromatography.
Allylic alkylation under biphasic conditions: The allylpalladium chloride
dimer (1.4 mg, 3.9 mmol, 2.5 mol%) and the bidentate ligand (11.7 mmol,
7.5 mol%) were added to a solution of tetrabutylammonium bromide
(0.1 equiv) in dichloromethane (500 mL). The solution was stirred at
room temperature for 5 min. A solution of acyclic substrate 1 or 5
(0.157 mmol, 1 equiv) in dichloromethane (1.5 mL), water (2 mL), and
50% KOH aqueous solution (63 mL, 0.314 mmol, 2 equiv) were succes-
sively added. The resulting biphasic system was stirred vigorously at
room temperature for 2 h. A saturated aqueous NH4Cl solution was
added and the aqueous phase was extracted with dichloromethane (3ꢆ).
The collected organic phases were washed with brine, dried over MgSO4,
and the solvent was removed in vacuo. The crude product was purified
by flash chromatography.
[13] Erratum: P.-O. Norrby, M. M. Mader, M. Vitale, G. Prestat, G. Poli,
Organometallics 2006, 25, 4234.
[14] H. B. Kagan, T.-P. Dang, J. Am. Chem. Soc. 1972, 94, 6429–6433.
[15] a) P. von Matt, A. Pfaltz, Angew. Chem. 1993, 105, 614–615; Angew.
Chem. Int. Ed. Engl. 1993, 32, 566–569; b) J. Sprinz, G. Helmchen,
Tetrahedron Lett. 1993, 34, 1769–1773; c) G. J. Dawson, C. G. Frost,
J. M. J. Williams, S. J. Coote, Tetrahedron Lett. 1993, 34, 3149–3150.
[16] B. M. Trost, D. L. van Vranken, C. Bingel, J. Am. Chem. Soc. 1992,
114, 9327–9343.
Acknowledgements
[17] T. Sturm, W. Weissensteiner, F. Spindler, Adv. Synth. Catal. 2003,
345, 160–164.
[18] H.-U. Blaser, W. Brieden, B. Pugin, F. Spindler, M. Studer, A. Togni,
Top. Catal. 2002, 19, 3–16.
[19] A. Miyashita, A. Yasuda, H. Takaya, K. Toriumi, T. Ito, T. Souchi,
R. Noyori, J. Am. Chem. Soc. 1980, 102, 7932–7934.
[20] C. Andrꢁs, J. P. Duque-Soladana, R. Pedrosa, J. Org. Chem. 1999,
64, 4273–4281.
CNRS and UPMC are acknowledged for financial support. The sponsor-
ship of COST Action D40 “Innovative Catalysis: New Processes and Se-
lectivities” is also kindly acknowledged. We thank Dr. Hans-Ulrich
Blaser and Dr. Benoit Pugin (Solvias AG) for a gift of the ligands and
Prof. J.-P. GenÞt and Dr. V. Ratovelomanana-Vidal for a sample of
ligand L3. P.F. thanks the Carlsberg Foundation, Lundbeck Foundation,
the Danish Council for Independent ResearchjTechnology, and Produc-
tion Sciences for financial support.
[21] R. Schmid, J. Foricher, M. Cereghetti, P. Schꢇnholzer, Helv. Chim.
Acta 1991, 74, 370–389.
[22] a) S. Jeulin, S. Duprat de Paule, V. Ratovelomanana-Vidal, J.-P.
GenÞt, N. Champion, Angew. Chem. 2004, 116, 324–329; Angew.
Chem. Int. Ed. 2004, 43, 320–325; b) S. Jeulin, S. Duprat de Paule,
V. Ratovelomanana-Vidal, J.-P. GenÞt, N. Champion, Proc. Natl.
Acad. Sci. USA 2004, 101, 5799–5804.
[23] R. Schmid, E. A. Broger, M. Cereghetti, Y. Crameri, J. Foricher, M.
Lalonde, R. K. Mꢂller, M. Scalone, G. Schoettel, U. Zutter, Pure
Appl. Chem. 1996, 68, 131–138.
[24] a) S. Duprat de Paule, N. Champion, V. Ratovelomanana-Vidal, J.-P.
Genet, P. Dellis, French patent 0112499, PCT FR02/03146,
W003029259; b) S. Duprat de Paule, S. Jeulin, V. Ratovelomanana-
Vidal, J.-P. Genet, N. Champion, P. Dellis, Tetrahedron Lett. 2003,
44, 823–826.
[25] a) D. Madec, G. Prestat, E. Martini, P. Fristrup, G. Poli, P.-O.
Norrby, Org. Lett. 2005, 7, 995–998; b) G. Maitro, G. Prestat, D.
Madec, G. Poli, Synlett 2006, 1055–1058; See also: c) G. Maitro, G.
Prestat, D. Madec, G. Poli, J. Org. Chem. 2006, 71, 7449–7454; d) S.
Vogel, X. Bantreil, G. Maitro, G. Prestat, D. Madec, G. Poli, Tetrahe-
dron Lett. 2010, 51, 1459–1461.
[1] G. Giambastiani, B. Pacini, M. Porcelloni, G. Poli, J. Org. Chem.
1998, 63, 804–807.
[2] a) G. Poli, G. Giambastiani, J. Org. Chem. 2002, 67, 9456–9459;
b) S. Lemaire, G. Giambastiani, G. Prestat, G. Poli, Eur. J. Org.
Chem. 2004, 2840–2847; c) M. B. T. Thuong, S. Sottocornola, G. Pre-
stat, G. Broggini, D. Madec, G. Poli, Synlett 2007, 1521–1524; d) See
also: D. Craig, C. J. T. Hyland, S. E. Ward, Synlett 2006, 2142–2144.
[3] For a preliminary report, see: X. Bantreil, G. Prestat, D. Madec, P.
Fristrup, G. Poli, Synlett 2009, 1441–1444; therein, the cyclized
structure 2, incorrectly indicated as the 3S,4R isomer instead of
3S,4S, depicts the major enantiomer when using the S-configurated
atropoisomeric ligands.
[4] B. M. Trost, P. E. Strege, J. Am. Chem. Soc. 1977, 99, 1649–1651.
[5] a) B. M. Trost, D. L. van Vranken, Chem. Rev. 1996, 96, 395–422;
b) A. Pfaltz, M. Lautens, Allylic Substitution Reactions in Compre-
hensive Asymmetric Catalysis II (Eds.: E. N. Jacobsen, A. Pfaltz, H.
Yamamoto), Springer, Berlin, 1999, pp. 834–884; c) B. M. Trost,
M. L. Crawley, Chem. Rev. 2003, 103, 2921–2943; d) B. M. Trost, J.
Org. Chem. 2004, 69, 5813–5837; e) Z. Lu, S. Ma, Angew. Chem.
2007, 120, 264–303; Angew. Chem. Int. Ed. 2007, 47, 258–297; B. M.
Trost, Chem. Sci. 2010, 1, 427–440.
[26] G. Trabesinger, A. Albinati, N. Feiken, R. W. Kunz, P. S. Pregosin,
M. Tschoerner, J. Am. Chem. Soc. 1997, 119, 6315–6323.
[27] This behavior suggests that a rapid anti-to-syn equilibration of the
h3-allylpalladium intermediates takes place under these biphasic
conditions; however, it should be kept in mind that under these
[6] a) T. Hayashi, K. Kanehira, T. Hagihara, M. Kumada, J. Org. Chem.
1988, 53, 113–120; b) M. Sawamura, H. Nagata, H. Sakamoto, Y.
Ito, J. Am. Chem. Soc. 1992, 114, 2586–2592; c) B. M. Trost, R. Rad-
Chem. Eur. J. 2011, 17, 2885 – 2896
ꢄ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2895

G. Poli et al.
same conditions the cyclization of substrates that incorporate fixed
nonisomerisable anti-allylpalladium complexes has been known (see
reference [25a]).
Sitkoff, A. Nicholls, M. Ringnalda, W. A. Goddard, III, B. Honig, J.
Am. Chem. Soc. 1994, 116, 11875–11882.
[39] The computed transition states clearly indicate that the trans-3,4-dis-
ubstituted g-lactam 2 is kinetically preferred over the cis isomer;
however, it should be kept in mind that the reaction conditions
imply that control of the stereochemistry at the active methylene
position is likely due to a base-promoted thermodynamic equilibra-
tion.
[40] a) M. Elstner, P. Hobza, T. Frauenheim, S. Suhai, E. Kaxiras, J.
Chem. Phys. 2001, 114, 5149–5155; b) S. Grimme, J. Comput. Chem.
2004, 25, 1463- 1473.
[41] Y. Zhao, N. E. Schultz, D. G. Truhlar, J. Chem. Theory Comput.
2006, 2, 364–382.
[42] a) Y. Zhao, D. G. Truhlar, Theor. Chem. Acc. 2008, 120, 215–241;
b) Y. Zhao, D. G. Truhlar, Acc. Chem. Res. 2008, 41, 157–167;
c) B. B. Averkiev, Y. Zha, D. G. Truhlar, J. Mol. Cat. A, 2010, 324,
80–88.
[28] It should be noted that generation of the h3-allylpalladium complex
from an allyl acetate precursor is a reversible and thermodynamical-
ly disfavored process; see: T. Yamamoto, O. Saito, A. Yamamoto, J.
Am. Chem. Soc. 1981, 103, 5600–5602; as a consequence, easy equi-
libration between the two diastereomeric h3-allylpalladium com-
plexes in case 2 of Figure 5 should be possible through the retro re-
action independently of the height of the direct interconversion bar-
rier (through p-s-p or “direct inversion”).
[29] a) J. C. Fiaud, J. Y. Legros, J. Org. Chem. 1990, 55, 4840–4846; b) C.
Amatore, A. Jutand, L. Mensah, G. Meyer, J.-C. Fiaud, J.-Y. Legros,
Eur. J. Org. Chem. 2006, 1185–1192.
[30] B. ꢅkermark, S. Hansson, A. Vitagliano, Organometallics 1991, 10,
2592–2599.
[31] The AB spin system centered around d=15.4 ppm differs in that it
[43] The XYZ coordinates and solution-phase energies for the olefin
complex and allyl complex are available in the Supporting Informa-
ACTHNUTRGNEUNG
shows a significantly smaller 2J(31P, 31P) coupling constant of J=
17 Hz in the 31P NMR spectrum of 9b; however, such a small cou-
pling constant was only observed in the spectra of 9b and 9c.
[32] P. S. Pregosin, R. Salzmann, Coord. Chem. Rev. 1996, 155, 35–68.
[33] The allyl face in 9a–e is prostereogenic; as a consequence, two dif-
ferent diastereomers will be obtained in the case of 9b–e depending
on which face the allyl will preferentially react, thus increasing the
number of possible isomers by a factor of two and complicating the
analysis.
ꢀ
tion; restrained optimizations with fixed C C bond distances of 1.5,
2.0, 2.5, 3.0, and 3.5 ꢅ for each of the four diastereomeric reaction
pathways.
[44] In a related study, Trost et al. studied the AAA reaction of allylic
carbonates to give enantioenriched chromanes with the standard
Trost ligand: B. M. Trost, H. C. Shen, L. Dong, J. P. Surivet, C. Syl-
vain, J. Am. Chem. Soc. 2004, 126, 11966–11983; this study suggest-
ed a mechanism involving the initial generation of the kinetic h3-al-
lylpalladium intermediate diastereoisomer followed by a p-s-p equi-
libration into the more stable and more reactive diastereomeric
complex (Curtin–Hammet conditions), which eventually cyclizes; al-
though we did not address a study to verify if such a mechanism is
operational in our case too, we stress that under ideal Curtin–
Hammet conditions the degree of enantioselection is independent of
the rate of formation and the energies of the intermediate diastereo-
meric h3-allylpalladium complexes as it depends exclusively on the
energy difference between the competing cyclizing transition states.
[34] a) P. Fristrup, M. Ahlquist, D. Tanner P.-O. Norrby, J. Phys. Chem. A
2008, 112, 12862–12867; b) N. Svensen, P. Fristrup, D. Tanner, P.-O.
Norrby, Adv. Synth. Catal. 2007, 349, 2631–2640; c) P. Fristrup, T.
Jensen, J. Hoppe, P.-O. Norrby, Chem. Eur. J. 2006, 12, 5352–5360.
[35] a) C. Lee, W. Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785–789;
b) A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652; c) P. J. Ste-
phens, F. J. Devlin, C. F. Chabalowski, M. J. Frisch, J. Phys. Chem.
1994, 98, 11623–11627.
[36] Jaguar, version 7.6, release 211, Schrodinger, LLC, New York, 2008.
[37] P. J. Hay, W. R. Wadt, J. Chem. Phys. 1985, 82, 299–310.
[38] a) B. Marten, K. Kim, C. Cortis, R. A. Friesner, R. B. Murphy, M. N.
Ringnalda, D. Sitkoff, B. Honig, J. Phys. Chem. 1996, 100, 11775–
11788; b) D. J. Tannor, B. Marten, R. Murphy, R. A. Friesner, D.
Received: May 13, 2010
Revised: November 26, 2010
Published online: February 3, 2011
2896
www.chemeurj.org
ꢄ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 2885 – 2896