Enantioselective construction of pyrrolidines by palladium-catalyzed asymmetric [3 + 2] cycloaddition of trimethylenemethane with imines

Article abstract of DOI:10.1021/ja210981a

A protocol for the enantioselective [3 + 2] cycloaddition of trimethylenemethane (TMM) with imines has been developed. Central to this effort were the novel phosphoramidite ligands developed in our laboratories. The conditions developed to effect an asymmetric TMM reaction using 2-trimethylsilylmethyl allyl acetate were shown to be tolerant of a wide variety of imine acceptors to provide the corresponding pyrrolidine cycloadducts with excellent yields and selectivities. Use of a bis-2-naphthyl phosphoramidite allowed the successful cycloaddition of the parent TMM with N-Boc imines, and has further permitted the reaction of substituted donors with N-tosyl aldimines and ketimines in high regio-, diastereo-, and enantioselectivity. Use of a diphenylazetidine ligand allows the complementary synthesis of the exocyclic nitrile product shown, and we demonstrate control of the regioselectivity of the product based on manipulation of the reaction parameters.

Full text of DOI:10.1021/ja210981a

Article
pubs.acs.org/JACS
Enantioselective Construction of Pyrrolidines by Palladium-Catalyzed
Asymmetric [3 + 2] Cycloaddition of Trimethylenemethane with
Imines
Barry M. Trost* and Steven M. Silverman†
Department of Chemistry, Stanford University, Stanford, California 94305-5080, United States
S
* Supporting Information
ABSTRACT: A protocol for the enantioselective [3 + 2]
cycloaddition of trimethylenemethane (TMM) with imines
has been developed. Central to this effort were the novel
phosphoramidite ligands developed in our laboratories. The
conditions developed to effect an asymmetric TMM reaction
using 2-trimethylsilylmethyl allyl acetate were shown to be
tolerant of a wide variety of imine acceptors to provide the
corresponding pyrrolidine cycloadducts with excellent yields and selectivities. Use of a bis-2-naphthyl phosphoramidite allowed
the successful cycloaddition of the parent TMM with N-Boc imines, and has further permitted the reaction of substituted donors
with N-tosyl aldimines and ketimines in high regio-, diastereo-, and enantioselectivity. Use of a diphenylazetidine ligand allows
the complementary synthesis of the exocyclic nitrile product shown, and we demonstrate control of the regioselectivity of the
product based on manipulation of the reaction parameters.
of 2-((trimethylsilyl)methyl) allyl mesylate with imines was
disclosed by Kremmit and Jones.16 In 1993, the Trost group
disclosed the palladium-catalyzed reaction of 2-
((trimethylsilyl)methyl) allyl acetate (and the methyl carbo-
nate) with aromatic and aliphatic imines (Scheme 1).17 The
INTRODUCTION
■
Pyrrolidines are ubiquitous structures in organic chemistry,
appearing routinely in nature in the form of primary and
secondary metabolites as well as in other biomolecules and
non-naturally occurring pharmaceuticals.1 Reported pyrrolidine
syntheses date back to the early twentieth century with some of
the first methods accounted as part of synthetic efforts to the
Scheme 1. Achiral TMM Reaction of Donors 1 and 2 with
Tosyl Imines
tropane ring system by Richard Willstatter in 1903 and Sir
̈
Robert Robinson in 1917.2 Classical methods of pyrrolidine
synthesis include alkylations and condensations, reductions of
succinimides, cyclic enamines and pyrroles, Overman’s Aza-
Cope Mannich cascade, and numerous others.3 The stereo-
selective synthesis of functionalized pyrrolidines is a topic of
longstanding interest due to the abundance of this structural
unit in natural products and pharmaceuticals, as well as in chiral
ligands4 and organocatalysts.5 A number of enantioselective
pyrrolidine syntheses have been reported in the literature; some
recent examples include methods such as the reduction of cyclic
enamines,6 hydroamination,7 and reductive cyclization.8
reaction was conducted in toluene or THF at elevated
temperatures using triphenylphosphine or triisopropylphos-
phite as a ligand. Aromatic and aliphatic aldimines were
tolerated, though ketimine examples were scarce. Phenyl and
nitrile-substituted donors were also effective in the cyclo-
addition.
Dipolar cycloadditions also represent a well-studied approach
toward the synthesis of pyrrolidines. The most extensively
studied method involves the reaction of olefins with
azomethine ylides.9,10 Methods have been developed to control
absolute stereochemistry using chiral auxiliaries11 and through
asymmetric catalysis.12 The complementary cycloaddition of
olefins with α-amino aldehydes has also been demonstrated,13
but methods in which an all carbon 1,3-dipole is used in the
cycloaddition are rare. The Lewis acid-catalyzed ring-opening of
cyclopropanes and subsequent addition to imines has been
effected,14,15 but the most general method is the metal-
catalyzed [3 + 2] cycloaddition of trimethylenemethane with
imines. A limited account of the nickel-catalyzed cycloaddition
Due to the prevalence of this heterocycle in many molecules
of interest, its formation in an enantioselective manner is highly
desirable. The proposed process would extend the scope of our
previously developed cycloaddition with olefin substrates18 to
imines, thereby providing a new method for the generation of
chiral pyrrolidines. Prior to this work, the asymmetric reaction
had been studied in the Trost group. It was discovered that the
BPPFA ligand shown could catalyze the reaction (Scheme 2),
generating the pyrrolidine depicted in good yield and 35%
Received: January 3, 2012
Published: February 6, 2012
© 2012 American Chemical Society
4941
dx.doi.org/10.1021/ja210981a | J. Am. Chem. Soc. 2012, 134, 4941−4954

Journal of the American Chemical Society
Article
π-allylpalladium species gives the desired product 5. In general,
the cycloaddition is believed to be stepwise in nature. The
product can subsequently undergo base-promoted isomer-
ization to give an endocyclic olefin.
Scheme 2. Marrs’ Catalytic, Asymmetric Pyrrolidine
Synthesis
The catalytic cycle illustrates the inherent difficulty in the
development of an asymmetric variant of the TMM reaction.
Nucleophilic addition is almost certainly the enantiodiscrimi-
nating step. The coordinated chiral ligands on the palladium are
distal to the reaction center, and this distance likely decreases
the ability of chiral ligands to transfer stereochemical
information during the addition. Accordingly, attempts to
induce asymmetry into the TMM reaction have largely been
limited to the use of chiral auxiliaries;20 chiral catalysis has been
largely unsuccessful.21 No reports of catalytic asymmetric
cycloadditions to imines have appeared prior to our work.
enantiomeric excess.19 However, despite extensive efforts,
higher ee could not be obtained.
The catalytic cycle for this process is shown in Scheme 3.
The reaction is believed to proceed through the zwitterionic
RESULTS AND DISCUSSION
■
Palladium-Catalyzed Reaction of 3-Acetoxy-2-trime-
thylsilylmethyl-1-propene with Aldimines. Our initial
studies focused on the phosphoramidite ligands which had
enjoyed success in the olefin cycloaddition.18 We chose
benzylidene aniline (6) as a test substrate as it had been
successfully utilized in the racemic reaction, and employed
conditions which mimicked the cyclopentane synthesis (5 mol
% Pd(dba)2, 10 mol % ligand, 1.6 equivalents TMM donor,
toluene). Our choice to use excess donor was based on the fact
that it undergoes slow polymerization under the reaction
conditions. Running the reaction at 45 °C for 4 h with a variety
of ligands, we immediately observed that both conversions and
selectivities were inferior to those observed in the olefin
cycloaddition (Scheme 4). Feringa ligand L122 gave a
reasonable 71% conversion but a disappointing 3% ee. Alexakis
observed that minor alterations to the amine portion of the
phosphoramidite had beneficial effects on the reactivity and
selectivity of iridium-catalyzed asymmetric allylic alkylation
reactions.23 On the basis of these studies, we tested L2-L7.
Variants of L1 with the (R,R,R) motif all gave fairly low ee
values. Naphthyl derivative L2 gave a similar result to the
Scheme 3. Catalytic Cycle for the Palladium-Catalyzed [3 +
2] TMM Cycloaddition with Imines
Pd-TMM intermediate 3 (Scheme 3), generated in situ by
ionization of the donor followed by acetate promoted
desilylation. Addition of the nucleophilic Pd-TMM complex
to the imine followed by collapse of the zwitterionic
intermediate 4 via attack of the nitrogen nucleophile onto the
Scheme 4. Performance of Phosphoramidite Ligands in the Reaction with Benzylidene Aniline
4942
dx.doi.org/10.1021/ja210981a | J. Am. Chem. Soc. 2012, 134, 4941−4954

Journal of the American Chemical Society
Article
a
Table 1. Imine Optimization Study
entry
R
temp (°C)
% yield
% ee
b
1
2
3
4
5
6
7
Ph (6)
Ts
45
45
45
23
45
23
23
76
84
45
98
c
Bn
0
c
P(O)Ph2
Boc
0
98
87
d
Fmoc
Cbz
0
d
0
a
b
c
d
All reactions were performed at 0.2 M in toluene and used 1.6 equivalents 1. Conversion. Recovered starting material. Complex mixture.
Table 2. Palladium-Catalyzed [3 + 2] Reactions of N-Aryl
Imines
the chiral space by substituting a 2-naphthyl group for a phenyl
(L9) increased the ee of the cycloadduct to 56%, though the
conversion dropped slightly. Substitution with 4-biphenyl
groups (L10) gave still higher ee and conversion (61% and
86% respectively), and the bis-1-naphthyl ligand (L11) boosted
the ee to 73%. Bis-2-naphthyl ligand L12 gave the best results
with a 76% conversion and 84% ee. In order to examine the
effects of ring size, we turned to azetidine ligand L13. A very
facile reaction was observed with full conversion being achieved
in less than an hour, but to our disappointment, it was
discovered that the highly active catalyst generated from this
system gave a significantly lower ee.
a
Having determined that L12 provided the most suitable
catalyst system for our reaction, we turned our focus to
optimization of the imine type (Table 1). Ideally, we wanted a
group which would not only provide products in high
enantiomeric excess, but could also be easily removed. While
the use of a tosyl group (entry 2) gave a very facile reaction,
providing essentially quantitative desired product within
minutes, the material obtained was of low ee. Benzyl and
diphenylphosphinoyl imines (entries 3−4) showed no
reactivity. Use of an N-Boc imine (entry 5) gave excellent
results with the protected pyrrolidine being obtained in 98%
yield and 87% ee. Other carbamates were not successful
(entries 6−7) as neither the fluorenylmethyl or benzyl
carbonates gave significant amounts of desired product, instead
yielding only complex mixtures.
The reaction of both N-aryl and N-Boc imines was examined.
A short assessment of substituted benzylidene anilines is shown
in Table 2. The reaction worked well when the N-bound aryl
ring bore electron donating groups (entries 2−3) or with-
drawing groups (entry 4). On the C-bound ring, electron
withdrawing groups significantly enhanced reactivity as
reflected in reduced amounts of silyl acetate needed to reach
full conversion (2.5 equivalents in entry 1 versus 1.6 in entries
2−4), as well as the lower temperature required (entry 4).
Lowering the temperature to 4 °C failed to produce significant
(>10%) product from more electron rich imines (entries 1−3).
Electron rich aniline 7e showed a high level of instability, both
under the reaction conditions and during purification attempts.
Highly electron poor imines of this class, such as those derived
from nitroaniline, were unstable to the reaction conditions. Use
of a Lewis acidic additive such as In(acac)3 did not improve the
results for this substrate class.
a
All reactions were performed at 0.2 M in toluene with 5% Pd(dba)2,
10% ligand, 1.6 equiv 1, and stirred for 4 h. Yields are isolated values;
ee’s were determined by chiral HPLC or chiral GC. Reaction
performed with 2.5 equivalents 1. Conversion.
b
c
Feringa ligand with only a small increase in enantioselectivity.
Similar to what was observed in the alkene reaction, installation
of an ortho-methoxy group (L3) increased conversion levels,
and substitution at both ortho-positions (L4) increased it
further. Removal of a methyl group as in L5 provided an
inactive catalyst. Inverting a stereocenter to provide ligands of
an (R,R,S) motif L6-L7 improved the ee values, but they still
remained below 40%. We next turned to pyrrolidine-based
phosphoramidites. The catalyst formed from L8 gave the
product in 35% ee and 73% conversion. Adjusting the nature of
The use of N-Boc imines provided a broader reaction scope
(Table 3). This class proved more reactive than the substituted
benzylidene anilines. The temperature could be reduced
4943
dx.doi.org/10.1021/ja210981a | J. Am. Chem. Soc. 2012, 134, 4941−4954

Journal of the American Chemical Society
Article
a
Table 3. Palladium-Catalyzed [3 + 2] Reactions of N-Boc imines
a
All reactions were performed at 0.2 M in toluene with 5 mol % Pd(dba)2, 10 mol % ligand, 1.6 equiv 1, and stirred for 4 h. Yields are isolated values;
b
ee’s were determined by chiral HPLC or chiral GC. Reaction performed with 2.5 mol % Pd(dba)2 and 5 mol % L12.
Scheme 5. Initial Reaction of Donor 10 with N-Boc Imine 9
considerably, as low as −15 °C in some cases (entries 3, 7, 11,
15), and conversion remained high with moderate increases in
enantioselectivity observed. The reaction proved insensitive to
the nature of the aromatic substitutent with similar results
obtained regardless of substitution pattern (entries 2−4) or
electronic nature of the substitutent (entries 2−8). Steric bulk
was tolerated as well (entries 9−12). Heterocycles performed
Figure 1. ORTEP illustration of (R)-tert-butyl 2-(4-chlorophenyl)-4-
well in the reaction as can be seen by the pyridyl (entry 13),
methylenepyrrolidine-1-carboxylate (8f) with thermal ellipsoids drawn
furyl (entry 14), and thiophenyl (entry 15) substituents. The
at the 50% probability level.
pyridyl example is particularly interesting as 8m constitutes a
4944
dx.doi.org/10.1021/ja210981a | J. Am. Chem. Soc. 2012, 134, 4941−4954

Journal of the American Chemical Society
Article
Table 4. Initial Study of the Reaction of Cyano Donor 10 with Tosyl Imine 12
entry
Pd
ligand
temp (°C)
solvent
conc (M)
yield (%)
13a (ee)
14a (ee)
15a (ee)
1
CpPd(η3-C3H5)
CpPd(η3-C3H5)
Pd(dba)2
L13
L13
L13
L14
L15
L16
L12
L17
L18
L2
50
4
PhCH3
PhCH3
PhCH3
PhCH3
PhCH3
PhCH3
PhCH3
PhCH3
PhCH3
PhCH3
PhCH3
PhCH3
PhCH3
PhCH3
PhCH3
PhCH3
PhCH3
PhCH3
CH2Cl2
THF
0.2
88
0
0.7
1
2
0.2
100
100
97
0.2 (49%)
0
1 (27%)
1
3
4
0.2
0.2
0
4
CpPd(η3-C3H5)
CpPd(η3-C3H5)
CpPd(η3-C3H5)
CpPd(η3-C3H5)
CpPd(η3-C3H5)
CpPd(η3-C3H5)
CpPd(η3-C3H5)
CpPd(η3-C3H5)
CpPd(η3-C3H5)
CpPd(η3-C3H5)
CpPd(η3-C3H5)
CpPd(η3-C3H5)
CpPd(η3-C3H5)
Pd(dba)2
4
0.2
0
0.4 (66%)
1 (42%)
1 (15%)
1 (ND)
1 (32%)
1 (69%)
1 (30%)
1 (33%)
1 (10%)
1 (31%)
1 (33%)
5
4
0.2
100
43
0.7 (3.5:1, 79%, ND)
12 (1.4:1, −30%, −56%)
0.2 (92%)
0.7 (77%)
6
4
0.2
0
7
4
0.2
80
0.2 (90%)
8
4
0.2
64
9 (1.2:1, 63%, 25%)
5.2 (1.8:1, 52%, 32%)
6 (1:1, 73%, 45%)
0.5 (45%)
0
9
4
0.2
79
0
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
4
0.2
78
0
L19
L20
L21
L22
L12
L12
L12
L12
L12
L12
L12
L12
L12
L12
L12
L12
L12
L12
L12
L12
L12
L12
4
0.2
100
100
79
0
4
0.2
0.3 (39%)
0
4
0.2
0
3 (27%)
4
0.2
0
50
80
50
4
0.2
93
0
0.8 (90%)
1.1 (70%)
0.5 (82%)
0.1 (92%)
0.9 (53%)
0.3 (91%)
0.1 (91%)
0.5
1 (43%)
1 (46%)
1 (41%)
1 (33%)
1 (12%)
1 (35%)
1 (21%)
1
0.2
100
100
100
83
0
0.2
0
Pd(OAc)2
0.2
0.2 (96%)
0
CpPd(η3-C3H5)
CpPd(η3-C3H5)
CpPd(η3-C3H5)
CpPd(η3-C3H5)
CpPd(η3-C3H5)
CpPd(η3-C3H5)
CpPd(η3-C3H5)
CpPd(η3-C3H5)
CpPd(η3-C3H5)
CpPd(η3-C3H5)
CpPd(η3-C3H5)
CpPd(η3-C3H5)
CpPd(η3-C3H5)
CpPd(η3-C3H5)
40
4
0.2
0.2
100
100
complex
35
0.2 (95%)
0.2 (92%)
0
4
dioxane
Et2O
0.2
4
0.2
4
PhCH3
THF
0.01
0.01
0.02
0.02
0.02
0.02
0.02
0.02
0.02
1.25 (95%)
2.5 (95%)
1.8 (95%)
2 (90%)
0.3 (86%)
0
0
1 (37%)
1 (48%)
1 (50%)
1 (50%)
1 (50%)
1 (52%)
1
4
46
0
4
THF
100
90
1.7 (92%)
0
a
26
4
THF
b
27
4
THF
100
100
100
48
3.3 (82%)
4 (81%)
10 (81%)
0
28
29
30
31
23
50
4
THF
THF
0
PhCl
0.5 (92%)
2 (90%)
1 (33%)
1 (39%)
4
DME
100
0
0.6 (85%)
c
32
4
PhCH3
0.02
a
b
c
In(acac)3 (10 mol %) was added. DMSO (1 vol%) was added. n-Hex4NCl (10 mol %) was added.
formal total synthesis of nicotine.17 The enantiomeric excess
was reduced in the cases examined where the heterocycle was
substituted adjacent to the heteroatom (entries 16−17),
presumably because of lower steric congestion; the correspond-
ing 3-substituted isomers gave significantly higher ee’s (entries
14−15). The unsaturated system (entry 18) gave material with
low yield and ee. The catalyst loading could be lowered (entry
10) with little change in results. The absolute configuration of
the cycloadducts was shown to be as depicted via X-ray
crystallographic analysis (Figure 1).
Palladium-Catalyzed Asymmetric Cycloaddition Re-
actions of Substituted Donors with Aldimines. Having
developed a practical route to the synthesis of disubstituted N-
Boc and N-aryl pyrrolidines, we turned our attention to
substituted donors with the goal of preparing more complex
systems. Due to the success of nitrile substituted donor 10 in
our cyclopentane synthesis, we chose this as a starting point.
Using tert-butyl naphthalen-2-ylmethylenecarbamate (9,
Scheme 5), the only identifiable material was trace amounts
of the dihydropyrrole 11, which results from the facile
isomerization of the exocyclic olefin to the more stable internal
double bond isomer. Because of the sluggish nature of the
reaction, we turned to azetidine ligand L13 as we had
previously observed it to promote very rapid cycloaddition.
Gratifyingly, at 45 °C we obtained a 72% yield of the product.
Attempts to lower the temperature gave only complex mixtures.
Use of CpPd(η3-C3H5) as a precatalyst24 gave similar results.
Complex mixtures were also obtained using a phenyl-
substituted donor.
Because of the facile reaction observed with tosyl imines
when studying the parent donor, we decided to reexamine this
substrate class. We reasoned that the increased steric bulk of a
substituted donor could serve to increase enantiodiscrimination
and provide the desired product with a higher ee than what had
been observed previously. To this end, we subjected tosyl imine
12 to the reaction conditions (Table 4). We were excited to
observe a quick reaction leading to a high yield of the
anticipated, albeit isomerized, TMM cycloadduct (entry 1).
However, analysis of the mixture indicated that in addition to
14a, it contained exocyclic cyano product 15a in a 0.7:1 mixture
with 15a slightly favored. Decreasing the temperature and
reducing the reaction time provided a mixture containing a
greater proportion of 15a (entry 2). No isomerization was
observed at this temperature. The anticipated cycloadduct 13a
4945
dx.doi.org/10.1021/ja210981a | J. Am. Chem. Soc. 2012, 134, 4941−4954

Journal of the American Chemical Society
Article
Increasing the steric bulk of the azetidine component with L14
(entry 4) decreased the regioselectivity. Furthermore, this
ligand seemed more prone to promote isomerization of 13a to
14a. Less isomerization was observed with the corresponding
pyrrolidine L15 (entry 5) and the endocyclic cyano products
were obtained in good ee. We varied the ligand diastereomer
using (S,R,R)-L16 (entry 6) and found that 13a was strongly
favored, but the yield, diastereoselectivity, and enantioselectiv-
ity were low. Our previously studied naphthyl ligand L12 gave
good conversion to a mixture of cycloadducts (entry 7), with
the anticipated product 13a and its tautomer 14a being
obtained with excellent enantioselectivities. Increasing the steric
bulk of the pyrrolidine moiety provided mixtures significantly
favoring 13a (entries 8−9), but selectivity dropped. Switching
to acyclic amine-derived phosphoramidites, use of the Feringa
ligand L1 also gave a mixture favoring 13a (entry 10). No
isomerization was observed, but the enantio- and diastereose-
lectivities were poor. More bulky acyclic amine phosphor-
amidite ligands gave mixtures favoring 15a and enantioselec-
tivity was low (entries 11−12). In both of these cases, 13a was
obtained as a single diastereomer and no isomerization was
observed. We also tested known diaminophosphite ligands in
order to determine how a more electron rich phosphorus
species would affect selectivity. Ligand L21 gave a mixture
favoring the isomerized product 14a but the selectivity was low
(entry 13). Ligand L22 formed an unreactive catalyst system
(entry 14).
Figure 2. Structures of additional ligands studied in the reaction of
cyano donor 10 with tosyl imine 12.
Using L12, we observed that increasing the reaction
temperature (entries 15−16) increased the proportion of
isomerized product 14a. The enantioselectivity remained high
at 50 °C, but decreased sharply when the temperature was
increased further. Variation of the palladium source did not
drastically alter the product proportions (entries 17−18). A
brief solvent study was also conducted. While the reaction
proceeded quickly in dichloromethane at 40 °C (entry 19), the
product ratio was similar to the reaction in toluene and the ee
was reduced. Use of tetrahydrofuran and dioxane (entries 20−
21) had only a marginal effect on the relative product amounts.
The reaction in diethyl ether gave a complex mixture with no
improvement in selectivity (entry 22).
was obtained as a single diastereomer; the ee values observed
for both products were low. Lowering the temperature to −15
°C did not change the ratio. Switching to Pd(dba)2 (entry 3)
had no effect on the product distribution. We did observe that
the amount of isomerized product increased over time, and
accordingly, reactions were quenched immediately after
disappearance of starting material, generally less than two
hours.
We next studied ligand effects on the cycloaddition, and
immediately found that the enhanced reactivity of this imine
class proved differential; a variety of ligands promoted the
reaction. Additional ligands tested are depicted in Figure 2.
Scheme 6. Expanded Catalytic Cycle Depicting the Mechanism of Formation of Exocyclic Cycloadduct 15
4946
dx.doi.org/10.1021/ja210981a | J. Am. Chem. Soc. 2012, 134, 4941−4954

Journal of the American Chemical Society
Article
Table 5. Study of the Reaction of Cyano Donor 10 with Electron-Rich Tosyl Imine 19
entry
temp (°C)
solvent
time (h)
conc (M)
% yield
13b (dr, ee)
1.5 (6:1 dr)
14b (ee)
15b
1
2
3
4
50
4
THF
2
0.02
0.02
0.2
100
40
1
0
0
0
0
0
1
1
PhCH3
PhCH3
PhCH3
2
1
4
1
100
100
8 (>10:1 dr)
4
1.5
0.08
10 (>20:1 dr, 96%)
Exocyclic nitrile 15 had not been previously reported to
occur in significant quantities. Its formation can be explained by
the modified catalytic cycle shown in Scheme 6. In an
analogous fashion to the parent donor, insertion of
palladium(0) into 10 gives π-allyl species 16. The displaced
acetate serves to remove the silyl group, giving reactive species
17. In the racemic reaction and previously studied asymmetric
variants, isomerization of 17 to 18 occurs more quickly than
addition to the π-system of the acceptor. Active donor 18
proceeds in the catalytic cycle shown on the left, giving the
anticipated TMM product 13, which may isomerize to 14
depending on the reaction conditions. Donor 18 is favored on
both steric and electronic grounds. Due to the highly
nucleophilic character of the palladium-phosphoramide-donor
system and the activated nature of the tosyl imine acceptor, the
addition of 17 to the imine occurs at a competitive rate with the
isomerization (right cycle). Assuming that ring closure occurs
by attack at the unsubstituted end of the syn π-allyl, this path
would lead to product 15.
The above data can be rationalized by this explanation. Any
modification of the reaction conditions that increases the rate
of π−σ-π isomerization transforming 17 to the more stable 18
favors production of 13 and 14. Conversely, any modification
that slows the rate of the π−σ-π interconversion leads to
increased proportions of 15. Lowering the reaction temperature
decreases the unimolecular isomerization rate by a greater
amount than the bimolecular addition rate, leading to increased
15 at lower temperatures (entry 1 versus entry 2 or entry 7
versus entry 15). Use of a more coordinating solvent such as
tetrahydrofuran stabilizes palladium during isomerization,
thereby accelerating it, and products 13 and 14 are observed
in increased proportions.
The observed ligand effects can also be rationalized by this
mechanism. Use of sterically less demanding ligands such as
azetidine L13 allows for a faster rate of TMM addition to the
imine relative to the isomerization. There is a clear trend
observed when moving to very bulky ligands. Use of ligands
L17 and L18 slows the addition significantly allowing time for
the π−σ-π isomerization to occur and accordingly, larger
proportions of 13 are observed. In fact, the mismatched case of
L16 slows the addition such that the highest ratio of 13 to 15 is
observed.
The appearance of significant quantities of product 15 when
only trace quantities of this type of product were observed in
the racemic reaction may be justified in terms of the nature of
the palladium-ligand complex. In the racemic case, two
phosphorus ligands are present on palladium during the
catalytic cycle, but 31P NMR suggests that only a single
phosphoramidite is present in the asymmetric reaction. In
addition to the electronic factors that favor the isomerization of
17 to 18, steric factors also favor 18 as interactions between the
R substituent and bulky palladium-ligand complex are reduced.
With only a single ligand on the palladium however, steric
factors are mitigated.
Additionally, this mechanism provides a justification for the
high enantiomeric excess of 13 and 14, and the low
enantioselectivity observed in the formation of 15. In the
case of the parent donor, we observed low enantioselectivity
when a tosyl imine was tested (Table 1, entry 2). In the case of
the substituted donor, we felt that the increased steric bulk
would allow for a more effective enantiodiscriminating event.
This result is what is generally observed in the pathway leading
to 13 and 14. However, the nucleophilic attack leading to 15
does not have the benefit of this added steric bulk and is more
similar to the case of the parent donor and indeed, similar levels
of enantioselectivity are observed.
In order to find conditions that minimized formation of 15,
we next explored concentration effects. By decreasing the
concentration, we hoped to slow the rate of TMM addition to
the imine to allow for complete equilibration of the donor.
Decreasing the concentration to 0.01 M (entries 23−24) gave
product ratios favoring 13a with high diastereo- and
enantioselectivity, but the reactions did not go to completion.
Increasing the concentration to 0.02 M restored the yield to
quantitative levels (entries 25−32). The amount of product
arising from the left cycle in Scheme 6 remained proportionally
high, as did the enantioselectivity. Only a single diastereomer
was observed but isomerization was a persistent problem. Use
of an indium additive (entry 26) mitigated this, but also
decreased the amount of 13a formed. DMSO promoted
isomerization (entry 27). Increasing the temperature (entries
28−29) led to complete isomerization, but greatly increased the
proportion of product arising from the left catalytic cycle, giving
a 10:1 ratio at 50 °C. Switching the solvent to chlorobenzene
(entry 30) reversed the regioselectivity and use of DME (entry
31) gave results similar to THF. Addition of n-Hex4NCl, known
to facilitate π−σ-π isomerization, stopped the reaction entirely
(entry 32).
With optimized conditions in hand, we turned our attention
to expansion of the substrate scope. Using the conditions in
Table 4, entry 29 on the more electron-rich tosyl imine 19, we
observed complete conversion to a mixture of 13b and 14b.
Interestingly, no exocyclic nitrile 15b was observed (Table 5,
entry 1). As THF had been observed to facilitate isomerization,
we switched the solvent to toluene, to mitigate this and
observed exclusively 13b, albeit in low yield. Increasing the
concentration (entry 3) restored the yield and gave the product
with good regioselectivity. An intermediate concentration of
0.08 M (entry 4) provided the optimal balance of yield and
regioselectivity. Clearly the reduced rate of attack of the TMM
complex on the more electron-rich imine decreased the amount
of 15b formed.
4947
dx.doi.org/10.1021/ja210981a | J. Am. Chem. Soc. 2012, 134, 4941−4954

Journal of the American Chemical Society
Article
Table 6. Scope of the Reaction of Cyano Donor 10 with Electron-Rich Tosyl Imines
a
All reactions were performed at the stated concentration in toluene with 5% CpPd(η3-C3H5), 10% ligand, 1.5 equiv 10, and stirred until complete
consumption of starting imine was achieved as monitored by GC. Yields are combined isolated values; ee values were determined by chiral HPLC.
b
Product ratios are determined by crude NMR. Remaining material consisted of 14j.
Using the further optimized conditions, we were able to
explore the scope of the cycloaddition of N-tosyl aldimines with
our cyano donor (Table 6). The reaction scope was broad and
included oxygenated aromatic (entries 1−5), aniline-derived
(entries 6−7), electron-rich heterocyclic (entry 8), and aliphatic
imines (entries 9−10). An electron-withdrawing group was
tolerated as long as the ring remained sufficiently electron-rich
(entry 5). This product type is important as it should allow
subsequent transformation using cross-coupling reactions. In
some cases (entries 3−5, 7−8), optimal conditions were
observed when concentration was further reduced to 0.04M;
this tended to improve both the endo:exo selectivity as well as
the diastereoselectivity of the major product. In the case where
a para-diphenylaminobenzaldehyde derived imine was used
(entry 7), we observed a drop in endo:exo selectivity, even at
0.04M. This is understandable as the electron density is more
distributed in this system and therefore the imine is less
electron-rich. Aliphatic imines performed very well in this
4948
dx.doi.org/10.1021/ja210981a | J. Am. Chem. Soc. 2012, 134, 4941−4954

Journal of the American Chemical Society
Article
adjacent quaternary center performed superbly, but interest-
ingly, the sense of diastereoselectivity was reversed from what
had been previously observed and the more stable trans-
pyrrolidine was obtained. This may indicate a reversible initial
addition step. In this case, ring closure to the sterically hindered
pyrrolidine is likely slowed and the thermodynamically favored
system is ultimately the major product. The cinnamaldehyde
imine (entry 11) gave a mixture of both products in decreased
yield. Relative stereochemistry was determined by NOE
analysis. Additionally, the relative and absolute stereochemistry
was confirmed through single crystal X-ray analysis of 13d
(Figure 3)
In addition to the development of conditions that led to the
anticipated TMM cycloadduct in high chemo-, diastereo-, and
enantioselectivity, our optimization studies also elucidated
conditions that allowed for the chemoselective preparation of
exocyclic cyano-type products. This is novel reactivity, as this
product has been observed only in trace quantities previously.
Using conditions developed during these studies, we were able
to prepare a series of these cycloadducts in excellent yields and
good chemoselectivity (Table 7). An alkyne donor could even
be used in the reaction to prepare enyne 15q (entry 6).
Through developing a mechanistic understanding of the
reaction and appropriate manipulation of reaction parameters,
we were able to exhibit significant control over the identity of
the product formed in the TMM reaction of aldimines with
nitrile donor 10 (Scheme 7). Using electron-rich imines and
dilute conditions, we could obtain “normal” cycloadducts such
as 13b with high chemo-, diastereo-, and enantioselectivity.
With electron-poor aldimines, we also could selectively form
the normal, albeit isomerized, product 14a in good yield and
high enantioselectivity using THF, dilute conditions, and
elevated temperature. Careful selection of reaction parameters
also gave us the opportunity to demonstrate novel reactivity.
Using our active azetidine ligand L13 and cold, concentrated
conditions, we could form 15a in good yield and chemo-
selectivity.
Figure 3. ORTEP illustration of (2S,3R)-2-(benzo[d][1,3]dioxol-5-
yl)-4-methylene-1-tosylpyrrolidine-3-carbonitrile (13d) with thermal
ellipsoids drawn at the 50% probability level.
reaction with consistently high yields being obtained at
concentrations ranging from 0.04 M to 0.2M. Even at increased
concentrations (entry 10), no trace of products of type 15 were
observed for saturated systems. A bulky imine containing an
Table 7. Selective Preparation of Exocyclic Product
Palladium-Catalyzed Asymmetric Cycloaddition Re-
actions of a Cyano Substituted Donor with Ketimines.
We have shown that the transition metal catalyzed [3 + 2]
TMM cycloaddition is a versatile method for the regio-,
diastereo-, and enantioselective construction of pyrrolidines.
While the catalytic, enantioselective addition of carbon
nucleophiles to aldimines is well precedented, the correspond-
ing additions to ketimines are rare.25 The intrinsic lower
reactivity of ketimines due to both electronic and steric factors
undoubtedly plays a significant role. In addition to the
opportunity for novel selectivity afforded by our phosphor-
amidite ligands, we have also observed a significant increase in
reactivity of various substrate classes. This enhanced reactivity
inspired us to examine the asymmetric cycloaddition to
ketimines, despite the relatively rare and specialized nature of
reactive ketimines in the achiral TMM reaction. Such a reaction
would provide a novel route to highly substituted pyrrolidines
containing a tetrasubstituted center.
We felt that the reduced reactivity of ketimines could prove
advantageous in one regard. Because the addition of a
nucleophile to a ketimine is expected to be considerably slower
than the corresponding aldimine, the occurrence of the
exocyclic product discussed above could be avoided.Our initial
studies began with examination of the Pd-catalyzed [3 + 2]
cycloaddition of cyano-TMM donor 10 with a series of
ketimines 20a−g (Table 8). Using CpPd(η3-C3H5) and Feringa
a
Using CpPd(η3-C3H5) with corresponding alkyne donor
4949
dx.doi.org/10.1021/ja210981a | J. Am. Chem. Soc. 2012, 134, 4941−4954

Journal of the American Chemical Society
Article
Scheme 7. Control over TMM Reaction Chemoselectivity by Selection of Reaction Parameters
a
Table 8. Initial Studies of the Reaction of Donor 10 with Ketimine Substrates
b
entry
R
ligand
temp (°C)
% yield
21/22
21 % ee
22 % ee
1
4-MeOC6H4 (20a)
Bn (20b)
OCH3 (20c)
P(O)Ph2 (20d)
Ns (20e)
Ts (20f)
Ts
L1
50
50
50
50
50
50
4
0
2
L1
0
3
L1
0
4
L1
complex
77 (16)
5
L1
0.17: 1
0: 1
ND
ND
53
c
6
L1
24 (53 )
7
L1
67
79
91
1: 1.1
2.2: 1
>20: 1
86
63
8
Ts
L13
L12
L12
L12
4
95
89
9
Ts
4
>99
d
10
Ts
50
0 (87 )
e
11
MBS (20g)
50
0 (100 )
a
All reactions were performed at 0.2 M in toluene with 5% CpPd(η3-C3H5), 10% ligand, 1.5 equiv 10, and stirred for 4 h. Yields are isolated values;
b
c
d
e
ee’s were determined by chiral HPLC. Value in parentheses represents additional yield of tautomerized product of type 14. 41% ee. 96% ee. 92%
ee.
ligand L1, we were initially discouraged by the fact that PMP-
imine 20a proved completely unreactive under the conditions
utilized (entry 1), as the corresponding aldimine class had
performed well using our parent donor (Table 2). Equally
frustrating, no reactivity was observed with benzyl imine 20b or
oxime ether 20c. Diphenylphosphinoyl imine 20d provided a
complex mixture (entry 4). Analysis of the crude mixture by
NMR indicated the presence of the desired product 21d and its
tautomer containing an endocyclic olefin conjugated with the
nitrile. Unfortunately, efforts to purify the products proved
problematic. These results indicated that a strongly electron-
withdrawing group was necessary for reactivity. We tested N-
nosyl imine 20e which proved to be superior to the
phosphinoyl imine from a reactivity standpoint (entry 5).
Disappearance of the imine was observed within a short time,
but the mixture contained a complex mixture consisting of both
diastereomers as well as the endocyclic tautomer. Using N-tosyl
imine 20f we observed formation of the desired cycloadduct,
albeit as a mixture of 22f and its endocyclic olefin tautomer.
Reduction of the temperature prevented isomerization and
increased the enantioselectivity, but the diastereoselectivity
remained poor and it became clear that ligand optimization
would be required (entry 7). Azetidine ligand L13 provided a
sizable increase in enantioselectivity, although the diastereose-
lectivity remained low (entry 8). Our previously utilized bis-2-
naphthyl phosphoramidite L12 solved both the enantio- and
diastereoselectivity issues, providing the desired product in
excellent yield with a >20:1 dr and a >99% ee (entry 9). The
tautomerization event appeared to be entirely temperature
dependent with ligand L12. Using either tosyl imine 20f or 4-
methoxybenzenesulfonyl imine 20g (entries 10−11), complete
tautomerization was observed when the reaction was conducted
at 50 °C, providing the tetrasubstituted olefins in 96% and 92%
ee, respectively. Notably, while N-Boc imines derived from
aromatic aldehydes performed admirably in previous studies,
they were ineffective in this case due to their propensity to
isomerize to the enamine tautomer.
We next examined the reaction scope with respect to the
imine (Table 9). Gratifyingly, the reaction proved very general
under the optimized conditions (Table 8, entry 9), performing
well in nearly every case examined. Yields and selectivities
proved insensitive to the steric bulk of the aromatic substituent
(entry 2). The steric bulk of the aliphatic substituent could be
increased (entry 3), but conversion dropped when it became
too sterically demanding (entry 4). Increasing the temperature
of this reaction only resulted in decomposition. The reaction
was insensitive to substitution pattern of the aromatic ring
(entries 5−7) or electronic nature of the substituent.
Spirocycles could be formed in good to excellent diaster-
eoselectivity and excellent enantioselectivity (entries 8−10),
4950
dx.doi.org/10.1021/ja210981a | J. Am. Chem. Soc. 2012, 134, 4941−4954

Journal of the American Chemical Society
Article
Table 9. Initial Scope of the Reaction of Cyano Donor 10
a
with Ketimines
Figure 4. ORTEP illustration of (2S,3R)-2-(2-methoxyphenyl)-2-
methyl-4-methylene-1-tosylpyrrolidine-3-carbonitrile (23f) with ther-
mal ellipsoids drawn at the 50% probability level.
yneimine was not tolerated at 4 or 50 °C (entry 12).
Conversion remained high at low temperatures, in some
cases as low as −15 °C (entry 8). The reaction could be
successfully scaled to 1.0 mmol using 5% CpPd(η3-C3H5) and
7.5% ligand with no loss in yield or selectivity, as was shown in
the case of substrate 20f. An X-ray crystal structure analysis of
23f (entry 7) unambiguously allowed the determination of
absolute and relative stereochemistry (Figure 4).
Our proposed model for the diastereoselectivity of the
reaction is shown in Figure 5, and is based on the imine
geometry. NOE studies show that the ketimines adopt the
geometry which places the sulfonyl group syn to the methyl or
methylene group. In fact, none of the other isomer was
1
observed by H NMR in any case. We suggest that the initial
bond forming event occurs as depicted in the top right scenario,
leading to the observed, thermodynamically less favored
diastereomer26 containing the nitrile and aryl group on the
same face of the pyrrolidine ring. The other diastereomer
would arise from the top left conformer, which is highly
disfavored due to severe steric interactions between the tosyl
group and bulky π-allyl. Interestingly, the only example in Table
9 in which significant quantities of the minor diastereomer were
observed is entry 8. In this case, the methylene group is
constrained such that it is spatially closer to the aryl group; the
120° angle depicted cannot be attained. This allows the π-allyl
to rotate away from the tosyl group, mitigating the steric
interactions and understandably, this pathway becomes more
prevalent.
In order to test the validity of this model, we needed to
constrain the imine in such a manner that the sulfonyl group
would be syn to the aryl group. An imine derived from saccharin
met this criterion. Our hypothesis for this imine class is shown
in the bottom row of Figure 5. Because the sulfone is
constrained in such a manner that it is syn to the aryl group, the
most favorable conformer becomes the one shown on the
bottom left in which the opposite face of the imine approaches
the donor. This places the π-allyl away from the sulfone and
leads to the product depicted with the nitrile and the methyl
group on the same face of the pyrrolidine. The conformer
leading to the other isomer is shown in the lower right and
places the allyl close to the sulfone. This pathway is disfavored
on steric grounds. The results of the TMM reaction shown in
a
All reactions were performed at 0.2 M in toluene with 5% CpPd(η3-
C3H5), 10% ligand, 1.5 equiv 10, and stirred for 2−4 h. Yields are
combined, isolated values; ee’s were determined by chiral HPLC. 1.0
mmol scale. Reaction performed at −15 °C. Minor diastereomer.
b
c
d
though the [6.7.5] system was obtained in low yield. It
appeared that the substrate itself was unstable, as decom-
position was observed by H NMR over very short periods of
time. Heterocycles were tolerated (entry 11), albeit with
moderately decreased enantioselectivity. Unfortunately, an
1
4951
dx.doi.org/10.1021/ja210981a | J. Am. Chem. Soc. 2012, 134, 4941−4954

Journal of the American Chemical Society
Article
Figure 5. Proposed stereochemical model for diastereoselectivity of ketimine TMM reaction.
a
Scheme 8. Reaction of Cyano Donor 10 with Saccharin-
Derived Imine 24
Table 10. Scope of the Aliphatic Ketimine Cycloaddition
Figure 6. Crystal structure of saccharine imine derived cycloadduct 25.
Scheme 8 support this interpretation. The relative stereo-
chemistry was confirmed through single crystal X-ray analysis
(Figure 6). Use of this class of cyclic imine provides a method
to control the diastereoselectivity of the reaction. After removal
of the sulfone, the opposite diastereomer of that shown in
Table 9 would be obtained.
a
All reactions were performed at 0.2 M in toluene with 5% CpPd(η3-
The reaction could also be conducted with aliphatic
ketimines containing multiple enolizable sites. Equivalently
substituted aliphatic ketimines performed well (Table 10, entry
1). Nonequivalently substituted aliphatic ketimines presented a
selectivity problem when subjected to the standard conditions
(entry 2). When L12 was employed, conversion remained high
with the cyclohexyl imine, but the diastereoselectivity dropped
below acceptable levels. Ligand L1 (entry 3) restored the
diastereoselectivity, albeit at the cost of conversion and
enantioselectivity. Azetidine ligand L13 provided a compro-
mise, giving the desired product in good yield, 10:1
diastereoselectivity, and restored enantioselectivity (entry 4).
Interestingly, the diastereoselectivity in this system was
reversed relative to the less bulky substrates. As discussed
above (Table 6, entry 10), this may indicate a reversible
addition step. A more remote branch point could be tolerated
while maintaining selectivity as shown by the isobutyl and
cyclohexyl imines (entries 5−6). Interestingly, the cyclohexyl
imine also gave excellent diastereoselectivity with ligand L12
with a jump in enantioselectivity (entry 7), likely due to the
C3H5), 10% ligand, 1.5 equiv 10, and stirred for 2−4 h unless
otherwise noted. Yields are combined, isolated values; ee’s were
b
c
determined by chiral HPLC. Using L12. Using L1.
rigidity of the system. The reaction remained moderately
tolerant of sterics, as seen by these substrates. However, a
tetrasubstituted center adjacent to the imine shut down the
reaction (entry 8). All product diastereselectivities were
confirmed by NOE analysis.
It should be noted that no exocyclic nitrile was observed in
the examples above, even at concentrations of 0.2 M and low
temperatures due to the slow rate of attack of the TMM
intermediate on the ketimine relative to π−σ-π isomerization.
However, we were able to observe significant quantities of this
product through use of a nosyl ketimine. Under our standard
conditions, quantitative yield of a mixture favoring exocyclic
cycloadduct 27 was obtained, due to the faster rate of attack of
the TMM intermediate on the more electron-deficient imine
(Scheme 9).
4952
dx.doi.org/10.1021/ja210981a | J. Am. Chem. Soc. 2012, 134, 4941−4954

Journal of the American Chemical Society
Article
Scheme 9. Formation of exo Product through Cycloaddition of TMM Donor 10 with Nosyl Ketimine
Scheme 10. Functionalization and Removal of the Tosyl Protecting Group from Ketimine Cycloadduct 21f
To show the utility of the pyrrolidine cycloadducts, we
wanted to demonstrate that the sulfonyl protecting group could
be removed (Scheme 10). Initial attempts to remove the tosyl
group using a variety of conditions (SmI2, sodium naphthalide,
lithium naphthalide) resulted in significant decomposition,
likely arising from reduction of the endocyclic double bond
tautomer of the cycloadduct. Use of magnesium in methanol
effectively removed the sulfonyl group in quantitative yield, but
concomitant olefin tautomerization and reduction afforded a
mixture of pyrrolidine diastereomers. Accordingly, we felt that
prior removal of the exocyclic olefin was necessary. From
cycloadduct 21f, Rh-catalyzed hydroboration provided the
desired alcohol 28 as a single diastereomer after oxidation. The
product was silylated and subsequently treated with magnesium
in methanol to easily remove the tosyl group, providing the
anticipated pyrrolidine 29 under mild conditions. The relative
stereochemistry was confirmed by NOE analysis.
ASSOCIATED CONTENT
■
S
* Supporting Information
Detailed experimental details, compound characterization data,
and spectra. This material is available free of charge via the
Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
bmtrost@stanford.edu
Present Address
†Bristol-Myers Squibb, One Squibb Drive, New Brunswick, NJ
08903.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
Both the exocyclic double bond and nitrile provide handles
for further functionalization in complex molecule synthesis. In
this case, we show that the olefin can be hydroborated with
complete regio- and diastereoselectivity. Subsequent removal of
the tosyl group gives the free pyrrolidine which we envision will
be useful in a variety of further transformations.
We thank the NSF and NIH for their generous support of our
programs. S.M.S. thanks Eli Lilly and Roche for graduate
fellowships. We thank Johnson-Matthey for their generous gifts
of palladium salts.
REFERENCES
■
(1) (a) O’Hagan, D. Nat. Prod. Rep. 2000, 17, 435. (b) Michael, J. P.
Nat. Prod. Rep. 2005, 22, 603.
(2) (a) Willstatter, R. Annalen. 1903, 317, 204. (b) Robinson, R. J.
Chem. Soc. Trans. 1917, 111, 762. (c) Humphrey, A. J.; O’Hagan, D.
Nat. Prod. Rep. 2001, 18, 494.
(3) Some selected examples of classical pyrrolidine syntheses include
(a) Wojcik, B.; Adkins, H. J. Am. Chem. Soc. 1934, 56, 2419−2424.
(b) Karrer, P.; Portmann, P. Helv. Chim. Acta 1948, 31, 2088−2092.
(c) Dolfini, J. E.; Dolfini, D. M. Tetrahedron Lett. 1965, 2053−2058.
(d) Basha, F. Z.; DeBernardis, J. F. Tetrahedron Lett. 1984, 25, 5271−
5274. (e) Jiang, C.; Frontier, A. J. Org. Lett. 2007, 9, 4939−4942.
(f) Lorthiois, E.; Marek, I.; Normant, J. F. J. Org. Chem. 1998, 63,
2442−2450. (g) Overman, L. E.; Kakimoto, M.-A. J. Am. Chem. Soc.
1979, 101, 1310.
CONCLUSION
■
We have developed a successful catalytic, asymmetric [3 + 2]
TMM cycloaddition reaction with imine acceptors that affords
pyrrolidine products. The methodology described herein
showcases the enantioselective addition of a carbon nucleophile
to an imine. This challenging transformation is expected to be
of great synthetic utility. The ketimine additions especially
stand out, as catalytic, asymmetric additions of any nucleophile
to these systems are rare. We have accomplished this with high
levels of chemo-, diastereo- and enantioselectivity, and
generated a tetrasubstituted center in the process. The reaction
scope is broad with respect to acceptor, tolerating a variety of
aromatic and aliphatic imines. We have also demonstrated that
the reaction conditions can be tailored to favor specific reaction
pathways to afford unique products. The nitrile donor provides
access to a variety of synthetically useful intermediates after
functional group interconversion, some of which we have
demonstrated. Investigations toward further expansion of the
donor scope are ongoing and will be reported in due course.
Applications of the novel TMM reactivity showcased here in
the synthesis of complex molecules are highly anticipated.
(4) Fache, F.; Schulz, E.; Tommasino, M. L.; Lemaire, M. Chem. Rev.
2000, 100, 2159.
(5) Dalko, P. I.; Moisan, L. Angew. Chem., Int. Ed. 2004, 43, 5138.
(6) Hou, G.-H.; Xie, J.-H.; Yan, P.-C.; Zhou, Q.-L. J. Am. Chem. Soc.
2008, 131, 1366.
(7) LaLonde, R. L.; Sherry, B. D.; Kang, E. J.; Toste, F. D. J. Am.
Chem. Soc. 2007, 129, 2452.
(8) Rhee, J. U.; Krische, M. J. J. Am. Chem. Soc. 2006, 128, 10674.
(9) (a) Heine, H. W.; Peavy, R. E. Tetrahedron Lett. 1965, 6, 3123.
(b) Padwa, A.; Hamilton, L. Tetrahedron Lett. 1965, 6, 4363.
(c) Huisgen, R.; Scheer, W.; Szeimies, G.; Huber, H. Tetrahedron
Lett. 1966, 7, 397. (d) Padwa, A. In Comprehensive Organic Synthesis;
4953
dx.doi.org/10.1021/ja210981a | J. Am. Chem. Soc. 2012, 134, 4941−4954

Journal of the American Chemical Society
Article
Trost, B. M., Fleming, I., Eds.; Pergamon: Oxford, U.K., 1991; Vol. 4,
p 1069.
(10) (a) Vedejs, E.; Martinez, G. R. J. Am. Chem. Soc. 1979, 101,
6452. (b) Vedejs, E.; Martinez, G. R. J. Am. Chem. Soc. 1980, 102,
7993. (c) Vedejs, E.; West, F. G. J. Org. Chem. 1983, 48, 4773.
(d) Vedejs, E.; West, F. G. J. Org. Chem. 1985, 50, 2170.
(11) (a) Padwa, A.; Chen, Y.-Y.; Chiacchio, U.; Dent, W. Tetrahedron
1985, 41, 3529. (b) Deprez, P.; Royer, J; Husson, H.-P. Tetrahedron:
Asymmetry 1991, 2, 1189. (c) Deprez, P.; Rouden, J.; Chiaroni, A.;
Riche, C.; Royer, J.; Husson, H.-P. Tetrahedron Lett. 1991, 32, 7531.
(d) Cooper, D. M.; Grigg, R.; Hargreaves, S.; Kennewell, P.; Redpath,
J. Tetrahedron 1995, 51, 7791. (e) Grigg, R. Tetrahedron: Asymmetry
1995, 6, 2475. (f) Gothelf, K. V.; Jorgensen, K. A. Chem. Rev. 1998, 98,
863.
(12) (a) Allway, P.; Grigg, R. Tetrahedron Lett. 1991, 32, 5817.
(b) Longmire, J. M.; Wang, B.; Zhang, X. J. Am. Chem. Soc. 2002, 124,
13400. (c) Gothelf, A. S.; Gothelf, K. V.; Hazell, R. G.; Jørgensen, K.
A. Angew. Chem., Int. Ed. 2002, 41, 4236. (d) Gothelf, A. S.; Gothelf, K.
V.; Hazell, R. G.; Jørgensen, K. A. Angew. Chem., Int. Ed. 2002, 41,
4236. (e) Chen, C.; Li, X.; Schreiber, S. L. J. Am. Chem. Soc. 2003, 125,
10174. (f) Saito, S.; Tsubogo, T.; Kobayashi, S. J. Am. Chem. Soc. 2007,
129, 5364. (g) Zeng, W.; Chen, G.-Y.; Zhou, Y.-G.; Li, Y.-X. J. Am.
Chem. Soc. 2007, 129, 750. (h) Fukuzawa, S.-I.; Okim, H. Org. Lett.
2008, 10, 1747.
(13) Restorp, P.; Fischer, A.; Somfai, P. J. Am. Chem. Soc. 2006, 128,
12646.
(14) Alper, P. B.; Meyers, C.; Lerchner, A.; Siegel, D. R.; Carreira, E.
M. Angew. Chem., Int. Ed. 1999, 38, 3186.
(15) (a) Carson, C. A.; Kerr, M. A. J. Org. Chem. 2005, 70, 8242.
(b) Jackson, S. K.; Karadeolian, A.; Driega, A. B.; Kerr, M. A. J. Am.
Chem. Soc. 2008, 130, 4196.
(16) Jones, M. D.; Kemmitt, R. D. W. J. Chem. Soc., Chem. Commun.
1986, 1201.
(17) Trost, B. M.; Marrs, C. M. J. Am. Chem. Soc. 1993, 115, 6636.
(18) (a) Trost, B. M.; Stambuli, J. P.; Silverman, S. M.; Schworer, U.
̈
J. Am. Chem. Soc. 2006, 128, 13328. (b) Trost, B. M.; Cramer, N.;
Silverman, S. M. J. Am. Chem. Soc. 2007, 129, 12396. (c) Trost, B. M.;
Silverman, S. M.; Stambuli, J. P. J. Am. Chem. Soc. 2011, 133, 19483.
(19) Trost, B. M.; Marrs, C. M. Unpublished results.
(20) (a) Chaigne, F.; Gotteland, J.-P.; Malacria, M. Tetrahedron Lett.
1989, 30, 1803. (b) Trost, B. M.; Yang, B.; Miller, M. J. Am. Chem. Soc.
1989, 111, 6482.
(21) For the most successful example reported prior to our work see
Yamamoto, A.; Ito, Y.; Hayashi, T. Tetrahedron Lett. 1989, 30, 375.
(22) Feringa, B. L. Acc. Chem. Res. 2000, 33, 346.
(23) Alexakis, A.; Polet, D. Org. Lett. 2005, 7, 1621.
(24) Shintani, R.; Park, S.; Duan, W.-L.; Hayashi, T. Angew. Chem.,
Int. Ed. 2007, 46, 5901.
(25) (a) Vilaivan, T.; Bhanthumnavin, W.; Sritana-Anant, Y. Curr.
Org. Chem. 2005, 9, 873. (b) Shibasaki, M.; Motomu, K. Chem. Rev.
2008, 108, 2853.
(26) Semiempirical AM1 calculations performed using Spartan
indicate an energy difference of 1.09 kCal/mol with epi-23d being
the more stable isomer.
4954
dx.doi.org/10.1021/ja210981a | J. Am. Chem. Soc. 2012, 134, 4941−4954