Oxygen donor-mediated equilibration of diastereomeric alkene-palladium(II) intermediates in enantioselective desymmetrizing heck cyclizations

Article abstract of DOI:10.1021/ja072571y

This investigation examines the origin of enantioselection in the desymmetrization of an acyclic prochiral Heck cyclization precursor. High asymmetric induction (97-98% ee) is attributed to a temporary interaction of a Lewis basic oxygen donor with weakly Lewis acidic palladium(II). A series of control experiments combined with quantum-chemical model calculations provided sound evidence for a mechanism involving oxygen donor-mediated, rapid equilibration of diastereomeric alkene-palladium(II) complexes prior to the selectivity-determining ring-closing event, a Curtin-Hammett scenario. Our study also highlights the importance of the cationic pathway (triflate counter anions versus halido ligand) and alkene stereochemistry (E versus Z) in asymmetric Heck reactions.

Full text of DOI:10.1021/ja072571y

Published on Web 10/13/2007
Oxygen Donor-Mediated Equilibration of Diastereomeric
Alkene-Palladium(II) Intermediates in Enantioselective
Desymmetrizing Heck Cyclizations
Axel B. Machotta,† Bernd F. Straub,*,‡ and Martin Oestreich*,†
Contribution from the Organisch-Chemisches Institut, Westfa¨lische Wilhelms-UniVersita¨t
Mu¨nster, Corrensstrasse 40, D-48149 Mu¨nster, Germany, and Organisch-Chemisches Institut,
Ruprecht-Karls-UniVersita¨t Heidelberg, Im Neuenheimer Feld 270, D-69120 Heidelberg, Germany
Received April 12, 2007; E-mail: martin.oestreich@uni-muenster.de; straub@oci.uni-heidelberg.de
Abstract: This investigation examines the origin of enantioselection in the desymmetrization of an acyclic
prochiral Heck cyclization precursor. High asymmetric induction (97-98% ee) is attributed to a temporary
interaction of a Lewis basic oxygen donor with weakly Lewis acidic palladium(II). A series of control
experiments combined with quantum-chemical model calculations provided sound evidence for a mechanism
involving oxygen donor-mediated, rapid equilibration of diastereomeric alkene-palladium(II) complexes
prior to the selectivity-determining ring-closing event, a Curtin-Hammett scenario. Our study also highlights
the importance of the cationic pathway (triflate counter anions versus halido ligand) and alkene
stereochemistry (E versus Z) in asymmetric Heck reactions.
importance of the enantioselective intramolecular6 Heck reac-
tion,7 making it a pivotal C-C bond-forming process in complex
molecule synthesis.8 Conversely, indirect formation of a ste-
reogenic carbon is achieved in a desymmetrizing Heck reaction,
in which the site of C-C bond formation is not coinciding with
the prochiral center.
The desymmetrizing or so-called group-selective Heck reac-
tion was developed by Shibasaki almost two decades ago.9 The
diastereocontrolled cyclization of prochiral 1 (Figure 1) using
a chirally modified palladium catalyst afforded a cis-bicyclo-
[4.4.0]decane system with high enantiomeric excess (92% ee).
In this ring closure, the stereochemistry of the vicinal stereogenic
carbon atoms is set at the former prochiral center as well as at
the site of C-C bond formation since â′-hydride elimination is
clearly favored for conformational reasons. Later, Feringa
accomplished the desymmetrization of structurally related 2
(Figure 1) using a monodentate ligand in excellent enantiomeric
excess (96% ee);10 epimerization of the intermediate CR-PdII
bond through a palladium enolate enabled conventional â-hy-
dride elimination. In recent years, Lautens11 and Bra¨se12 reported
1. Introduction
The enantioselective desymmetrization of prochiral (achiral)1
as well as meso2 compounds is certainly an elegant technique
in stereoselective synthesis.3 Breaking of the symmetry plane
in these precursors is effected by differentiation of heterotopic
functional groups upon reaction with a chiral, nonracemic
reagent or catalyst. An attractive feature of this approach
originates from the fact that even processes not capable of
forming a stereogenic, tetravalent carbon might be performed
in an asymmetric sense. In principle, the Mizoroki-Heck
reaction4 is such a transformation unless â-hydride elimination,
which usually reestablishes the alkene fragment, is steered away
from the site of C-C bond formation. In order to selectively
realize alternative â′-hydride elimination, the â-carbon atom
generated in the migratory insertion step must not have any
accessible hydrogens attached to it. This is the case when a
quaternary carbon center is formed and for cyclic and, therefore,
rigid skeletons, in which a synperiplanar arrangement of the
Câ-H bond and the CR-PdII bond is conformationally inac-
cessible. The realization that asymmetrically substituted carbons
are thereby directly accessible5 largely re-evaluated the synthetic
(6) Link, J. T. In Organic Reactions; Overman, L. E., Ed.; Wiley: New York
2002; Vol. 60, pp 157-534.
(7) (a) Shibasaki, M.; Vogl, E. M.; Ohshima, T. AdV. Synth. Catal. 2004, 346,
1533-1552. (b) Guiry, P. J.; Kiely, D. Curr. Org. Chem. 2004, 8, 781-
794. (c) Donde, Y.; Overman, L. E. In Catalytic Asymmetric Synthesis;
Ojima, I., Ed.; Wiley: New York 2000; pp 675-697.
† Westfa¨lische Wilhelms-Universita¨t Mu¨nster.
‡ Ruprecht-Karls-Universita¨t Heidelberg.
(1) Willis, M. C. J. Chem. Soc., Perkin Trans. 1 1999, 1765-1784.
(2) Hoffmann, R. W. Angew. Chem., Int. Ed. 2003, 42, 1096-1109.
(3) Ward, R. S. Chem. Soc. ReV. 1990, 19, 1-19.
(8) Dounay, A. B.; Overman, L. E. Chem. ReV. 2003, 103, 2945-2964.
(9) (a) Sato, Y.; Sodeoka, M.; Shibasaki, M. J. Org. Chem. 1989, 54, 4738-
4739. (b) Ohrai, K.; Kondo, K.; Sodeoka, M.; Shibasaki, M. J. Am. Chem.
Soc. 1994, 116, 11737-11748. (c) Ohshima, T.; Shibasaki, M. Tetrahedron
1993, 49, 1773-1782. (d) Ohshima, T.; Kagechika, K.; Adachi, M.;
Sodeoka, M.; Shibasaki, M. J. Am. Chem. Soc. 1996, 118, 7108-7116.
(10) (a) Imbos, R.; Minnaard, A. J.; Feringa, B. L. J. Am. Chem. Soc. 2002,
124, 184-185. (b) Imbos, R.; Minnaard, A. J.; Feringa, B. L. J. Chem.
Soc., Dalton Trans. 2003, 2017-2023.
(4) (a) Beletskaya, I. P.; Cheprakov, A. V. Chem. ReV. 2000, 100, 3009-
3066. (b) Bra¨se, S.; de Meijere, A. In Handbook of Organopalladium
Chemistry for Organic Synthesis; Negishi, E.-i., de Meijere, A., Eds.;
Wiley: New York, 2002; Vol. 1, pp 1123-1132. (c) Larhed, M.; Hallberg,
A. In Handbook of Organopalladium Chemistry for Organic Synthesis;
Negishi, E.-i., de Meijere, A., Eds.; Wiley: New York, 2002; Vol. 1, pp
1133-1178. (d) Heck, R. F. Palladium Reagents in Organic Synthesis;
Academic Press: New York, 1985.
(11) Lautens, M.; Zunic, V. Can. J. Chem. 2004, 82, 399-407.
(12) Lormann, M. E. P.; Nieger, M.; Bra¨se, S. J. Organomet. Chem. 2006, 691,
2159-2161.
(5) Overman, L. E. Pure Appl. Chem. 1994, 66, 1423-1430.
9
10.1021/ja072571y CCC: $37.00 © 2007 American Chemical Society
J. AM. CHEM. SOC. 2007, 129, 13455-13463
13455

A R T I C L E S
Machotta et al.
Table 1. Desymmetrization of Cyclization Precursors
(E,E)-9-(E,E)-12a
T
t
ee
[%]c
yield
[%]d
entry
precursor
donor
base
[
°C]
[h]
productb
Figure 1. Structural motifs of desymmetrizing Heck cyclizations.
Scheme 1. A Novel Structural Motif for Desymmetrizationa
1
2
3
(E,E)-9
(E,E)-9
(E,E)-9
OH
OH
OH
K2CO3 80 15 (R,E,E)-13
K2CO3 60 36 (R,E,E)-13
TMP
88
94
88
2
89
92
96
93
97
97
98
74
82
36
55
87
45
31
83
84
98
40
71
86
80 15 (R,E,E)-13
4
5
6
7
8
9
10
11
12
13
(E,E)-10 OSiEt3 K2CO3 80 15 (R,E,E)-14
(E,E)-11 OMe
(E,E)-11 OMe
(E,E)-11 OMe
(E,E)-11 OMe
(E,E)-11 OMe
(E,E)-11 OMe
(E,E)-11 OMe
K2CO3 80 15 (R,E,E)-15
K2CO3 50 15 (R,E,E)-15
K2CO3 35 90 (R,E,E)-15
TMP
TMP
TMP
TMP
80 15 (R,E,E)-15
50 15 (R,E,E)-15
50 36 (R,E,E)-15
35 90 (R,E,E)-15
(E,E)-12
(E,E)-12
H
H
K2CO3 80 15 (S*,E,E)-16 18
TMP 80 15 (S*,E,E)-16
8
a All reactions were conducted with a substrate concentration of 0.1 M
in toluene with 4.0 equiv of the respective base. b The absolute configuration
was determined for (R,E,E)-15 (Sections 5 and 6 in Supporting Information);
the absolute configurations of (R,E,E)-13 and (R,E,E)-14 were assigned by
chemical correlation with (R,E,E)-15. c See Supporting Information for
details. d Yield of analytically pure product isolated after flash column
chromatography on silica gel. TMP ) 2,2,6,6-tetramethylpiperidine.
a For the detailed description of all cyclization precursor syntheses, see
Supporting Information. Do ) donor.
effect15 in a catalyst-controlled Heck cyclization. In this contri-
bution, we now disclose a full account of our investigationss
including further improved substrate (E,E)-11sdirected at
understanding the subtle factors controlling enantioselection.
Supported by quantum-chemical calculations, an unprecedented
enantioselectivity-discriminating mechanism emerges.
the desymmetrization of precursors having either a cis-bicyclo-
[3.3.0]octane (3, Figure 1) or a trans-bicyclo[4.4.0]decane (4,
Figure 1) core. Heck annulation yielded the tricyclic structures
in 99% ee and 84% ee, respectively; structural reasoning again
rationalized high diastereoselectivity and regioselective â′-
hydride elimination.
2. Results and Discussion
All these examples corroborated the impression that, in order
to allow for efficient differentiation of the enantiotopic unsatur-
ated branches, these must be incorporated into a cyclic and,
therefore, rigid framework. Nevertheless, we envisioned the
desymmetrizing Heck cyclization of open-chain precursors
having the general structure 5 (Scheme 1). Their ring closure
would produce benzannulated carbocycles 6 only with stereo-
genic information remote from the actual C-C bond-forming
site.13 Our investigation commenced with cyclization experi-
ments of a series of several oxygen-containing precursors (E,E)-
7-(E,E)-10 as well as deoxygenated (E,E)-12 (Scheme 1).14
Within this survey, we found that the level of enantioinduction
was markedly dependent on (a) the presence of the hydroxy
group, (b) its position relative to the leaving group X (n ) 1
superior to n ) 0), and (c) the leaving group in X (OTf superior
to Br) itself.14 These cooperative factors, Lewis basic oxygen
atom in ideal proximity to palladium(II) and cationic pathway
of the Heck reaction, indicated an unusual neighboring-group
2.1. Desymmetrizing Heck Cyclizations of E,E-Configured
Precursors. When subjecting prochiral diene (E,E)-9 to con-
ventional Heck conditions in toluene at 80 °C, we were pleased
to find exclusive formation of a single diastereomer (R,E,E)-
13 in good enantiomeric excess (Table 1, entry 1). The
enantiomeric excess improved significantly at 60 °C (Table 1,
entry 2), but at even lower temperatures, transtriflation from
C(sp2)-OTf to C(sp3)-OH became a competing side reaction.
Various organic and inorganic bases were also tested but had
little effect on enantioselection (e.g., Table 1, entry 3).14a It is
interesting to note that both K2CO3 and Cs2CO3 in toluene were
effective, whereas no conversion was observed when polar DMF
was used. At that time, we believed that any increase of the
level of enantioselection might only be possible if the reaction
temperature were further decreased. We hoped to achieve
exactly that by protecting the hydroxy group and thereby
preventing triflyl migration in (E,E)-9. Hence, silyl-protected
(E,E)-10 was cyclized under the initial reaction conditions. To
our surprise, this Heck reaction was sluggish, and (R,E,E)-14
(13) A 3,6-dimethoxy-functionalized substrate 5 (n ) 1) was of particular interest
to us as its cyclization would provide a catalytic asymmetric access to the
AB-ring synthon of anthracycline antibiotics: Oestreich, M.; Sempere-
Culler, F.; Machotta, A. B. Synlett 2006, 2965-2968.
(14) (a) Oestreich, M.; Sempere-Culler, F.; Machotta, A. B. Angew. Chem., Int.
Ed. 2005, 44, 149-152. (b) Coogan, M. P.; Pottenger, M. J. J. Organomet.
Chem. 2005, 690, 1409-1411.
(15) (a) Oestreich, M. Eur. J. Org. Chem. 2005, 783-792. (b) Oestreich, M.
Directed Mizoroki-Heck Reactions; Chatani, N., Ed.; Topics in Organo-
metallic Chemistry; Springer: Heidelberg, 2007; http://dx.doi.org/10.1007/
3418_2007_063.
9
13456 J. AM. CHEM. SOC. VOL. 129, NO. 44, 2007

Desymmetrizing Heck Cyclizations
A R T I C L E S
Lewis acidic cations. To test for this, we chose LiBF4 as the
source of the Lewis acid since large quantities are soluble in
THF; therefore, Heck cyclizations (E,E)-9 f (R,E,E)-13 were
conducted in homogeneous THF solution using Et3N as the base
(reference reaction: 84% ee and 70% yield at 60 °C). In a series
of experiments with variable equivalents of LiBF4 (based on
cyclization precursor (E,E)-9), a distinct effect on the level of
enantioselection was found (Figure 2). The enantiomeric excess
(∼80% ee) remained nearly unchanged in the presence of 1.0-
2.5 equiv of LiBF4; at higher concentrations of Lewis acidic
lithium cations (3.0-5.0 equiv), the enantiomeric excess (∼20%
ee) collapsed. Conversely, KBF4 (1.0-5.0 equiv) had no effect
whatsoever on the stereochemical outcome of this transformation
(∼85% ee). These results imply that, at high lithium(I)
concentrations, the Lewis basic sites at oxygen are not available
for palladium(II) coordination. Less Lewis-acidic potassium(I)
is not capable of coordinatively saturating the oxygen donor.
Figure 2. Desymmetrizing Heck cyclization of (E,E)-9 in the presence of
variable amounts of Lewis acid. Reaction conditions: (E,E)-9, Pd(OAc)2,
(R)-BINAP, LiBF4 or KBF4, Et3N, THF, 60 °C.
2.1.2. Control Experiment II. Influence of the Counter-
anion. Cationic versus Neutral Pathway. In a vast oversim-
plification, there are two mechanistic scenarios commonly
presumed to govern Heck reactions.19 These, termed cationic
and neutral pathways, refer to the formal charge at palladium-
(II) after oxidative addition/alkene coordination prior to migra-
tory insertion and are dependent on the counteranions present.
The nature of the counteranion determines the reaction path-
way:19 weakly or noncoordinating anions such as the triflate
anion will dissociate, thereby creating a vacant coordination site
at palladium(II) (cationic pathway). On the other hand, strongly
coordinating anions such as halides will remain coordinated at
palladium(II) throughout the catalytic cycle (neutral pathway).
As verified in several asymmetric Heck reactions using bidentate
ligands, this has strong implications on the enantioselectivity-
controlling alkene capture/migratory insertion process.7,20 The
latter is believed not to proceed through a pentacoordinate
alkene-palladium(II) complex.21
The desymmetrizing Heck cyclization of aryl triflate (E,E)-9
(and (E,E)-11) should proceed by the cationic pathway. Con-
sequently, oxidative addition followed by dissociation provides
a vacant Lewis acidic site at palladium(II) capable of coordinat-
ing a proximal Lewis basic oxygen donor. Starting from the
corresponding aryl bromide (E,E)-7, a neutral pathway should
be followed. On the basis of these considerations, we predicted
poor enantiocontrol for the cyclization of (E,E)-7, yet the first
experiment in this survey seemed to show the opposite (Table
2, entry 1). A closer look revealed, though, that (R,E,E)-13 was
formed with 86% ee at 5% yield which, in turn, corresponded
to a single turnover using 5.0 mol % Pd(OAc)2. At higher
temperature, conversion was raised and enantiomeric excesses
decreased to almost zero (Table 2, entries 2 and 3). From these
experiments, we concluded that stereoinduction is dependent
on bromide concentration, which gradually increases with
conversion.
was isolated in poor yield and with completely eroded enan-
tiomeric excess (Table 1, entry 4)!
This pronounced effect indicated that a suitably located
oxygen donor in the cyclization precursor might be decisive
for enantiocontrol. Aware of the precedent of a hydroxy-directed
Heck reaction,16-18 we suspected that the hydroxy group in
(E,E)-9 might interact with the palladium(II) center during the
catalytic cycle. Steric demand as well as attenuated Lewis
basicity of the oxygen donor in (E,E)-10 might account for our
experimental finding.
This hypothesis was then substantiated by an extensive survey
of the enantioselective desymmetrization of the corresponding
methyl ether (E,E)-11 (Table 1, entries 5-11). Cyclization of
(E,E)-11 under the initial reaction conditions cleanly provided
(R,E,E)-15 with an enantiomeric excess (Table 1, entry 5), which
compared well with the data obtained from the cyclization of
(E,E)-9 (Table 1, entry 1). As anticipated, the decomposition-
free ring closure of (E,E)-11 was now possible at remarkably
low temperatures; enantiomeric excesses of 92% ee at 50 °C
and 96% ee at 35 °C were high, yet chemical yields were
moderate (Table 1, entries 6-7). Importantly, substantially
higher yields and again improved enantioselectivities were
obtained by exchanging K2CO3 for TMP (Table 1, entries
8-11), a base that had been less efficient in the cyclization of
(E,E)-9 (Table 1, entry 3). Under optimized reaction conditions,
(E,E)-11 was cyclized in 98% yield and 97% ee (Table 1, entry
10).
With strong evidence for a pivotal role of oxygen in the Heck
reaction, we performed the Heck cyclization of deoxygenated
(E,E)-12 under conditions identical to those for (E,E)-9 and
(E,E)-11. In agreement with our proposal, less reactive cycliza-
tion precursor (E,E)-12 gave tetralin (S*,E,E)-16 in almost
racemic form (Table 1, entries 12 and 13).
2.1.1. Control Experiment I. Influence of Excess Lewis
Acidic Cations. We reasoned that the proposed coordination
of the oxygen donor, -OH or -OMe, to weakly Lewis acidic
palladium(II) might be severely disturbed by addition of external
(19) Cabri, W.; Candiani, I. Acc. Chem. Res. 1995, 28, 2-7.
(20) (a) Ashimori, A.; Bachand, B.; Overman, L. E.; Poon, D. J. J. Am. Chem.
Soc. 1998, 120, 6477-6487. (b) Ashimori, A.; Bachand, B.; Calter, M.
A.; Govek, S. P.; Overman, L. E.; Poon, D. J. J. Am. Chem. Soc. 1998,
120, 6488-6499.
(16) Kao, L.-C.; Stakem, F. G.; Batel, B. A.; Heck, R. F. J. Org. Chem. 1982,
47, 1267-1277.
(21) (a) Thorn, D. L.; Hoffmann, R. J. Am. Chem. Soc. 1978, 100, 2079-2090.
(b) Samsel, E. G.; Norton, J. R. J. Am. Chem. Soc. 1984, 106, 5505-
5012. (c) Goossen, L. J.; Koley, D.; Hermann, H., Thiel, W. Chem.
Commun. 2004, 2141-2143 (d) Goossen, L. J.; Koley, D.; Hermann, H.;
Thiel, W. Organometallics 2005, 24, 2398-2410. (e) Goossen, L. J.; Koley,
D.; Hermann, H.; Thiel, W. J. Am. Chem. Soc. 2005, 127, 11102-11114.
(17) Bernocchi, E.; Cacchi, S.; Ciattini, P. G.; Morera, E.; Ortar, G. Tetrahedron
Lett. 1992, 33, 3073-3076.
(18) Kang, S.-K.; Jung, K.-Y.; Park, C.-H.; Namkoong, E.-Y.; Kim, T.-H.
Tetrahedron Lett. 1995, 36, 6287-6290.
9
J. AM. CHEM. SOC. VOL. 129, NO. 44, 2007 13457

A R T I C L E S
Machotta et al.
Table 2. Desymmetrization of Cyclization Precursor (E,E)-7a
the elementary steps at the palladium center in these Heck
cyclizations.22 For this, we introduced several simplifications:
(1) The terminal phenyl substituents in the substrate were
omitted. (2) BINAP was simplified to a (Z,Z)-bis-1,4-(dimeth-
ylphosphino)buta-1,3-diene. (3) Instead of triflate, mesylate was
modeled as leaving group. Of course, we are fully aware that
this approach will not enable the reproduction of detailed steric
effects and exact enantioselectivities. Our goal was primarily
to acquire information about the coordination numbers at
palladium in catalyst intermediates, the nature of the rate-
determining step, and the nature of the stereoselectivity-
determining step.
T
t
ee
[%]b
yield
[%]c
entry
precursor
base, additive
[
°
C]
[h]
product
1
2
3
4
(E,E)-7 K2CO3
(E,E)-7 K2CO3
(E,E)-7 K2CO3
(E,E)-7 K2CO3, Ag2CO3
(1.4 equiv)
80 15 (R,E,E)-13 86
100 15 (R,E,E)-13 44
100 20 (R,E,E)-13
80 20 (S,E,E)-13 12
5
13
68
13
4
5
(E,E)-7 Ag2CO3
100 20 (S,E,E)-13
1
58
The picture for the computed elementary steps at the
palladium center is straightforward, and the qualitative conclu-
sions might be considered as reasonable also for the full,
experimental system. In the overall uncharged combination of
the 14 valence electron palladium(0) complex and the model
substrate, we arbitrarily normalized their computed free energies
to zero.
(2.0 equiv)
a All reactions were conducted with a substrate concentration of 0.1 M
in toluene with 4.0 equiv of the respective base. b See Supporting Informa-
tion for details. c Yield of analytically pure product isolated after flash
column chromatography on silica gel.
Table 3. Desymmetrization of Cyclization Precursor (E,E)-8a
BINAP-palladium(0) must be considered as the catalyst
resting state. It is unclear whether (κ2P-BINAP)Pd (M1) or η2-
alkene complex (M2) of (κ2P-BINAP)Pd is the most stable
species since the free energy differences are smaller than the
computational accuracy of our model calculations (Figure 3).
At small equilibrium concentrations, the (κ2P-BINAP)Pd frag-
ment coordinates the aryl π-system of the substrate in an η2
mode. In M3, the coordination of palladium to an unsubstituted
arene position is favored over the “active” substitution mode
in complex M4.
T
t
ee
yield
[%]c
entry
precursor
base
[
°
C]
[h]
product
[%]b
1
2
3
4
(E,E)-8
(E,E)-8
(E,E)-8
(E,E)-8
K2CO3
K2CO3
TMP
80
50
80
50
15
15
15
15
(R*,E,E)-17
(R*,E,E)-17
(R*,E,E)-17
(R*,E,E)-17
46
46
48
53
35
30
88
48
TMP
The barrier for the activation of the C-O bond of the aryl
triflate via transition state M5 is higher than expected from the
experimental observations. The mesylate model, however, is a
poorer leaving group than the triflate, and the missing solvation
in the calculations further explains the high computed barrier.
Nevertheless, the oxidative addition step is clearly the rate-
determining and rate-limiting step in the overall catalytic cycle.
The initial product of the oxidative addition is the ion pair
M6+‚OTf-. The cationic d8-ML3 palladium(II) is coordinated
in a T-shape environment by two phosphorus atoms and the
aryl ligand. The sulfonate is hydrogen-bridged to the hydroxy
group of the former substrate.
a All reactions were conducted with a substrate concentration of 0.1 M
in toluene with 4.0 equiv of the respective base. b See Supporting Informa-
tion for details. c Yield of analytically pure product isolated after flash
column chromatography on silica gel.
The cyclizations of (E,E)-7 in the presence of halide
scavengers, which provide an entry into the cationic pathway,
are still in need of an explanation. The net result from a
screening of silver salts and different bases is that yields and
enantioselectivities were poor but (S,E,E)-13 is generated with
inverted absolute configuration (Table 2, entries 4 and 5).
Similar observations have been reported by Overman in the
past.20
The sulfonate counteranion is known to be easily replaced
in the coordination sphere of palladium(II). In Figure 4, we
omitted the sulfonate anion, and the energy of the cationic
palladium(II) cation M6+ was arbitrarily normalized to zero.
Thus, absolute energy comparisons between Figure 3 on the
one side, and Figures 4 and 5 on the other side are not viable.
The palladium atom in complex M6+ is highly unsaturated.
Therefore, the coordination of the hydroxy group to the isomer
M7+ is highly exergonic. It is not surprising that the alkenes
2.1.3. Control Experiment III. Variation of the Position
of the Oxygen Donor Relative to the Palladium(II) Center.
An intramolecular interaction of the oxygen donor with the
intermediate palladium(II) center would certainly require their
ideal vicinity. Such a situation appeared to be the case for (E,E)-
9, but would it be the same for (E,E)-8, in which the position
of the hydroxy group relative to the C(sp2)-OTf bond is
changed? Selected data for the enantioselective ring closure
(E,E)-8 f (R*,E,E)-17 are summarized in Table 3 (entries 1-4).
The moderate enantioselectivities clearly demonstrate that
coordinative Pd-O interaction by a virtually planar five-
membered chelate is not favored. Chemical yields with TMP
as the base were substantially higher than with K2CO3, but the
observed enantioselectivities were invariably ∼50% ee.
2.2. Discussion of Quantum-Chemical Model Calculations.
We used quantum-chemical calculations at the B3LYP/
LACV3P**++//B3LYP/LACVP** level of theory of a simpli-
fied model system in the gas phase to obtain information on
(22) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648-5652. (b) Vosko, S. H.;
Wilk, L.; Nusair, M. Can. J. Phys. 1980, 58, 1200-1211. (c) Lee, C.; Yang,
W.; Parr, R. G. Phys. ReV. B 1988, 37, 785-789. (d) Jaguar 6.5;
Schro¨dinger, Inc.: Portland, OR, U.S.A., 2006. (e) Krishnan, R.; Binkley,
J. S.; Seeger, R.; Pople, J. A. J. Chem. Phys. 1980, 72, 650-654. (f) Hay,
P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299-310. (g) Schaftenaar,
G.; Noordik, J. H. J. Comput.-Aided Mol. Design 2000, 14, 123-134. (h)
Frisch, M. J.; Pople, J. A.; Binkley, J. S. J. Chem. Phys. 1984, 80, 3265-
3269. (i) Clark, T.; Chandrasekhar, J.; Spitznagel, G. W.; v. R. Schleyer,
P. J. Comput. Chem. 1983, 4, 294-301. (j) Becke, A. D. Phys. ReV. A
1988, 38, 3098-3100. (k) Perdew, J. P.; Zunger, A. Phys. ReV. B 1981,
23, 5048-5079. (l) Perdew, J. P. Phys. ReV. B 1986, 33, 8822-8824.
9
13458 J. AM. CHEM. SOC. VOL. 129, NO. 44, 2007

Desymmetrizing Heck Cyclizations
A R T I C L E S
Figure 3. Computed catalyst resting state and rate-determining oxidative addition step.
Figure 4. Saturation of the “hot” 14-valence electron palladium(II) cation, rapid ligand exchange, and irreversible stereoselectivity-determining alkene
insertion.
bind less tightly to the metal in complexes M9+, M10+, and
M11+. The simultaneous coordination of both the hydroxy group
and one of the two alkenes at palladium in M8+, however, is
only a ligand-exchange transition state between two 16-valence
electron species. Eighteen-valence electron palladium(II) species
are rarely local minima;23 16-valence electrons at palladium are
usually equivalent to electronic saturation.24
The somewhat lower barrier for this alcohol ligand- versus
alkene ligand-exchange rearrangement makes the latter faster
than the insertion of a coordinated alkene into the C(sp2)-Pd
bond. The hydroxy group might be considered as an intramo-
lecular mediator that facilitates ligand-exchange rearrangements,
(24) (a) Glorius, F. Angew. Chem., Int. Ed. 2004, 43, 3364-3366. (b)
Christmann, U.; Vilar, R. Angew. Chem., Int. Ed. 2005, 44, 366-374. (c)
Vicente, J.; Arcas, A. Coord. Chem. ReV. 2005, 249, 1135-1154. (d) Miura,
M. Angew. Chem., Int. Ed. 2004, 43, 2201-2203. (e) Trzeciak, A. M.;
Ziolkowski, J. J. Coord. Chem. ReV. 2005, 249, 2308-2322.
(23) Izatt, R. M.; Watt, G. D.; Eatough, D.; Christensen, J. J. J. Chem. Soc. A.
1967, 8, 1304-1308.
9
J. AM. CHEM. SOC. VOL. 129, NO. 44, 2007 13459

A R T I C L E S
Machotta et al.
Figure 5. Rearrangement into an agostic complex and â-hydride elimination.
thereby enabling the rapid equilibration toward the alkene
complex with the most facile migratory insertion. Such an
insertion is highly exergonic and thus irreversible. Despite the
formation of a C-C bond in M12+, the palladium in M13+
remains coordinated at the aryl ipso carbon.
The interaction of palladium with the aromatic π system in
M13+ is weak, and the rearrangement via the transition state
M14+ to a â-agostic complex M15+ is facile and essentially
isoenergetic (Figure 5). The coordination of the hydroxy group
at the palladium cation in M16+ is again exergonic but too weak
to prevent the eventual â-hydride elimination via transition state
M17+. Deprotonation of either the cationic palladium(II)-
alkene hydride complex M18+ or a cationic palladium(II)
hydride M19+ closes the catalytic cycle by reformation of the
palladium(0) catalyst resting state.
2.3. Mechanistic Proposal: A Curtin-Hammett-Type
Scenario.25 At the first glance, the Lewis basic oxygen donor
appears to be an enantioselectivity-directing “anchor” group.
Based on the quantum-chemical model calculations, however,
the coordination of the oxygen to the palladium(II) alkene cation
intermediate is only temporary. The oxygen ligand enhances
the rate for the equilibration of the diastereotopic alkene
coordination modes to the (R)-BINAP palladium aryl fragment.
The ring closure is facile and fast, and an efficient and even
faster equilibration is mandatory to translate the free enthalpy
barrier difference of the alkene insertion processes into a high
enantiomeric excess (Figure 6).
stantaneous after an almost barrierless rotation around a single
bond. Thus, the coordination of the alkene is fast, and almost
randomly, either one of the two diastereotopic alkene fragments
coordinates.
An associative alkene-to-alkene ligand exchange is sterically
essentially impossible in the real, nonsimplified complex, while
the insertion of the coordinated alkene into the C(sp2)-Pd bond
is fast. However, the coordination and dissociation of alkenes
in the wrong orientation such as in the oxygen-free analogues
of M10+ and M11+, which would eventually lead to a seven-
membered ring product, have to be considered as an unproduc-
tive equilibrium. Without intramolecular mediation of the alkene
ligand exchange by oxygen ligand atoms, the coordination of
the alkene fragment decides upon the diastereoselectivity of the
insertion elementary step, and thus it also decides upon the
enantioselectivity of the overall reaction.26
2.4. Diastereospecificity of the Desymmetrizing Intramo-
lecular Heck Reaction. Cyclization of Z,Z-Configured Pre-
cursors. As part of our study, we also investigated the influence
of the double bond geometry on the stereochemical course of
the intramolecular Heck reaction of 9 and 11. Of course, we
were primarily interested in enantiomeric excesses and absolute
configurations of cyclization products 13 and 15, respectively,
when using Z,Z-configured substrates (Z,Z)-9 and (Z,Z)-11.
Aside from absolute stereocontrol, the diastereospecificity (that
is, E as well as Z configuration in 9/11 is retained in 13/15)
was of interest.
Surprisingly, cyclization precursors (Z,Z)-9 and (Z,Z)-11 were
both significantly less reactive than their diastereomers (E,E)-9
Without hydroxy or methoxy groups, either one of the alkene
fragments coordinates to the initially formed 14-valence electron
palladium species. This ligand-metal bond formation is in-
(26) We also wondered whether hydrogen bonding of the hydroxy group to the
triflate might facilitate the activation of the substrate’s Ar-OTf bond by
enhancing the triflate’s leaving group capability. However, the reactions
of the methoxy derivative and the deoxygenized substrate operate at
comparable rates.
(25) (a) Curtin, D. Rec. Chem. Prog. 1954, 15, 111-128. (b) Seeman, J. I. Chem.
ReV. 1983, 83, 83-134. (c) Zefirov, N. S. Russ. J. Org. Chem. 1997, 33,
138-139. (d) http://www.iupac.org/goldbook/C01480.pdf.
9
13460 J. AM. CHEM. SOC. VOL. 129, NO. 44, 2007

Desymmetrizing Heck Cyclizations
A R T I C L E S
Figure 6. (Left) Random coordination of alkene to palladium(II) and low enantioselectivities for substrates without Lewis basic oxygen. (Right) Curtin-
Hammett-type scenario, oxygen donor-mediated equilibration and high enantioselectivities for substrates with Lewis basic oxygen.
Table 4. Desymmetrization of Cyclization Precursor (Z,Z)-9 and
and (E,E)-11 (Table 4). A reaction temperature of 100 °C finally
facilitated the cyclization of (Z,Z)-9 and (Z,Z)-11. In the presence
of K2CO3 as the base, completely isomerized (R,E,E)-13 (63%
ee) and (R,E,E)-15 (80% ee) were isolated in high yields (Table
4, entries 1 and 2). Careful analysis and purification of the crude
product allowed for the reisolation of minor amounts of (E,E)-
9, which is a sound indication of predominant pre-Heck27 rather
than post-Heck double bond isomerization by the palladium
hydride species.
(Z,Z)-11a
When using TMP as the base, we were pleased to find that
no palladium hydride-mediated E,Z isomerization to the ther-
modynamically favored E-alkenes occurred. (Z,Z)-9 and (Z,Z)-
11 cyclized in moderate to good yields and with poor enanti-
omeric excesses (Table 4, entries 3 and 4). Absolute configuration
of (R,Z,Z)-13 and (R,Z,Z)-15 was assigned by chemical cor-
relation, iodine-catalyzed isomerization, with (R,E,E)-13 and
(R,E,E)-15, respectively.
T
t
ee
yield
[%]c
entry
precursor
base
[
°
C]
[h]
product
[%]b
1
2
3
4
(Z,Z)-9
(Z,Z)-11
(Z,Z)-9
(Z,Z)-11
K2CO3
K2CO3
TMP
100
100
100
80
20
20
20
20
(R,Z,Z)-13
(R,Z,Z)-15
(R,Z,Z)-13
(R,Z,Z)-15
-
-
33
38
d
e
61f
80
TMP
a All reactions were conducted with a substrate concentration of 0.1 M
in toluene with 4.0 equiv of the respective base. b See Supporting Informa-
tion for details. c Yield of analytically pure product isolated after flash
column chromatography on silica gel. d The desired product was not
detected; instead, (R,E,E)-13 (58%, 63% ee) and (E,E)-9 (15%) were
isolated. e The desired product was not detected; instead, (R,E,E)-15 (86%,
80% ee) was isolated. f Isolated along with the 2(2E)-isomer (R,Z,E)-13
(6%) and the 4E-isomer (R,E,Z)-13 (3%).
This base-dependent difference of the diastereochemical
outcome might be rationalized by the physical state of the bases.
(27) Sonesson, C.; Larhed, M.; Nyqvist, C.; Hallberg, A. J. Org. Chem. 1996,
61, 4756-4763.
9
J. AM. CHEM. SOC. VOL. 129, NO. 44, 2007 13461

A R T I C L E S
Machotta et al.
Scheme 2. Diastereospecific Desymmetrizing Heck Cyclizations
Reductive elimination of palladium hydrides might be fast under
homogeneous conditions (TMP as the base) while being
hampered under heterogeneous conditions (K2CO3 as the base)
which, in turn, opens the undesired syn-hydropalladation reaction
channel.
The pair of diastereospecific Heck reactions ((E,E)-9 f
(R,E,E)-13 and (Z,Z)-9 f (R,Z,Z)-13) is outlined in Scheme 2.
It is a particularly nice example of stereospecific syn-carbopal-
ladation (18 f 19) as well as syn-â-hydride elimination (19 f
13). The poor catalytic turnover in the Z,Z series might originate
from an energetically unfavorable â-hydride elimination of syn-
19, forming a cis-stilbene unit (syn-19 f (R,Z,Z)-13); con-
versely, â-hydride elimination of anti-19 will liberate a ther-
modynamically more stable trans-stilbene unit (anti-19 f
(R,E,E)-13).
a Lewis basic oxygen to weakly Lewis acidic palladium(II) as
the pivotal enantioselectivity-controlling structural element,32
which mediates or, more precisely, facilitates equilibration of
diastereomeric alkene-palladium(II) complexes prior to the
irreversible, stereochemistry-determining migratory insertion. In
principle, this situation resembles an intramolecular Curtin-
Hammett scenario as the ratio of the diastereomeric alkene-
palladium(II) complexes is not reflected in the level of
enantioselection. In the absence of the Lewis basic oxygen, the
energetic barrier for interconversion of the diastereomeric
alkene-palladium(II) complexes is substantially higher than the
activation energy for alkene insertion into the C(sp2)-Pd bond.
Consequently, the ratio of the initially formed diastereomeric
alkene-palladium(II) complexes is reflected in the level of
enantioselection.
This conceptually interesting neighboring-group effect also
underscores the fundamental influence of weak donors in
palladium-catalyzed asymmetric C-C bond-forming reactions.
With a vacant coordination site at palladium(II) (cationic
pathway) and the oxygen donor in ideal proximity to it, high
chemical yield (98%) and excellent enantiomeric excess (97%
ee) were obtained under unusually mild reaction conditions
(50 °C).
3. Conclusion
A directing effect of an oxygen donor was first seen by Heck16
and, later, further verified by Cacchi and Ortar17 as well as
Kang18 in regioselective Heck arylations. Controlling regio- or
diastereoselectivity in intermolecular Heck reactions by a
covalently bound heteroatom donor has generally been an
emerging area in recent years.15,28 Several amine- and pyridine-
based systems were introduced by Hallberg,29 Carretero,30 and
Itami and Yoshida.31
The present study demonstrates for the first time the unique
influence of a weak donor on enantioselection in an asymmetric
Heck reaction. We initially assumed that a similar directing
effect governs asymmetric induction in the desymmetrizing
Heck ring closure of the acyclic prochiral cyclization precursor
(E,E)-9,14a but in contrast to the above-mentioned processes,
combined experimental and quantum-chemical examinations
revealed a refined, unprecedented mechanistic picture. A series
of control experiments substantiates temporary coordination of
The findings disclosed herein as well as a recent seminal
contribution by Curran33 clearly corroborate that exploration of
the stereochemistry-determining step in asymmetric Heck
processes is worthwhile, providing insight beyond conventional
mechanistic reasoning.
We also studied the influence of the double bond configu-
ration in the cyclization precursor on the level of enantioselec-
tivity. High enantiomeric excess was only seen for E,E
diastereomers ((E,E)-11 f (R,E,E)-15, 83%, 93% ee, TMP as
base); the corresponding Z,Z-configured substrates were less
reactive and selective ((Z,Z)-11 f (R,Z,Z)-15, 80%, 38% ee,
TMP as base). While under those conditions diastereospecific,
(28) Hoveyda, A. H.; Evans, D. A.; Fu, G. C. Chem. ReV. 1993, 93, 1307-
1370.
(29) (a) Andersson, C.-M.; Larsson, J.; Hallberg, A. J. Org. Chem. 1990, 55,
5757-5761. (b) Nilsson, P.; Larhed, M.; Hallberg, A. J. Am. Chem. Soc.
2001, 123, 8217-8225. (c) Nilsson, P.; Larhed, M.; Hallberg, A. J. Am.
Chem. Soc. 2003, 125, 3430-3431.
(32) Overman presented evidence for a unique directing effect of an alkene unit
in a diastereoselective Heck cyclization. We had also considered such an
interaction in our system (E,E)-9, but unselective ring closure of (E,E)-12
dispelled this idea (Table 1): (a) Earley, W. G.; Oh, T.; Overman, L. E.
Tetrahedron Lett. 1988, 29, 3785-3788. (b) Madin, A.; Overman, L. E.
Tetrahedron Lett. 1992, 33, 4859-4862.
(30) (a) Buezo, N. D.; Alonso, I.; Carretero, J. C. J. Am. Chem. Soc. 1998, 120,
7129-7130. (b) D´ıaz Buezo, N.; de la Rosa, J. C.; Priego, J.; Alonso, I.;
Carretero, J. C. Chem. Eur. J. 2001, 7, 3890-3900.
(31) (a) Itami, K.; Nokami, T.; Ishimura, Y.; Mitsudo, K.; Kamei, T.; Yoshida,
J.-i. J. Am. Chem. Soc. 2001, 123, 11577-11585. (b) Itami, K.; Mineno,
M.; Muraoka, N.; Yoshida, J.-i. J. Am. Chem. Soc. 2004, 126, 11778-
11779. (c) Itami, K.; Yoshida, J.-i. Synlett 2006, 157-180.
(33) Lapierre, A. J. B.; Geib, S. J.; Curran, D. P. J. Am. Chem. Soc. 2007, 129,
494-495.
(34) Sempere-Culler, F. Diploma Thesis, Albert-Ludwigs-Universita¨t, Freiburg,
Germany, 2004.
9
13462 J. AM. CHEM. SOC. VOL. 129, NO. 44, 2007

Desymmetrizing Heck Cyclizations
A R T I C L E S
it is interesting to note that, using solid K2CO3 as the base,
complete E,Z isomerization occurred ((Z,Z)-11 f (R,E,E)-15,
86%, 80% ee). Pre-Heck double bond scrambling is likely.27
the early phase of the project, Ilona Hauser for the large-scale
preparation of 3-bromo-1-trimethylsilyl-1-propyne, Gerd Fe-
hrenbach for performing the HPLC analyses, and Dr. Manfred
Keller for a X-ray crystal structure analysis (all Albert-Ludwigs-
Universita¨t Freiburg).
Acknowledgment. The research was (in part) supported by
the International Research Training Group Freiburg-Basel (GRK
1038) (predoctoral fellowship to A.B.M., 2005-2006) and the
Fonds der Chemischen Industrie. M.O. is indebted to the
Deutsche Forschungsgemeinschaft for an Emmy Noether fel-
lowship (2001-2006). B.F.S. thanks the Verband der Chemis-
chen Industrie, the Ludwigs-Maximilians-Universita¨t Mu¨nchen
as well as the Freie Universita¨t Berlin, the Dr. Otto Ro¨hm-
Geda¨chtnisstiftung, and the Deutsche Forschungsgemeinschaft
(Str 861/2-1) for financial support. We thank Dipl.-Chem.
Fernando Sempere-Culler13,14a,34 for significant contributions in
Supporting Information Available: Starting material syn-
theses, characterization data as well as 1H and 13C NMR spectra
for all new compounds, and determination of absolute config-
uration including molecular structure (X-ray); Cartesian coor-
dinates and energy data of computed model complexes. This
material is available free of charge via the Internet at
http://pubs.acs.org.
JA072571Y
9
J. AM. CHEM. SOC. VOL. 129, NO. 44, 2007 13463