Rotational isomers of N-methyl-N-arylacetamides and their derived enolates: Implications for asymmetric Hartwig oxindole cyclizations

Article abstract of DOI:10.1021/jo400385t

The rotational preferences of N-(2-bromo-4,6-dimethylphenyl)-N-methyl 2-phenylpropanamide were studied as a model of precursors for Hartwig asymmetric oxindole cyclizations. The atropisomers of this compound were separated by flash chromatography, and then the enantiomers were resolved and the interconversions of the stereocenter and the N-Ar axis were studied. Under thermal conditions, the axis is very stable. Under the basic conditions of the Hartwig cyclization, both the stereocenter and the chiral axis equilibrate via enolate formation. The N-Ar rotation barrier of a 2-phenylacetamide analogue was reduced from 31 kcal mol^-1 in the precursor to 17 kcal mol ^-1 in the enolate. Reasons for this dramatic barrier reduction and implications of both N-Ar and amide C-N rotations for Hartwig cyclizations are discussed.

Full text of DOI:10.1021/jo400385t

Article
pubs.acs.org/joc
Rotational Isomers of N‑Methyl‑N‑arylacetamides and Their Derived
Enolates: Implications for Asymmetric Hartwig Oxindole Cyclizations
Jeremie Mandel, Xiaohong Pan, E. Ben Hay, Steven J. Geib, Craig S. Wilcox,* and Dennis P. Curran*
́
́
Department of Chemistry, University of Pittsburgh, Pittsburgh, Pennsylvania 15260, United States
S
* Supporting Information
ABSTRACT: The rotational preferences of N-(2-bromo-4,6-dime-
thylphenyl)-N-methyl 2-phenylpropanamide were studied as a model
of precursors for Hartwig asymmetric oxindole cyclizations. The
atropisomers of this compound were separated by flash chromatog-
raphy, and then the enantiomers were resolved and the
interconversions of the stereocenter and the N−Ar axis were studied.
Under thermal conditions, the axis is very stable. Under the basic
conditions of the Hartwig cyclization, both the stereocenter and the
chiral axis equilibrate via enolate formation. The N−Ar rotation barrier of a 2-phenylacetamide analogue was reduced from 31
kcal mol−1 in the precursor to 17 kcal mol−1 in the enolate. Reasons for this dramatic barrier reduction and implications of both
N−Ar and amide C−N rotations for Hartwig cyclizations are discussed.
INTRODUCTION
■
Rotational isomerism of Nsp2−Csp2 single bonds spans the full
range of possibilities depending on substituents. At one end,
amide and related functional groups are planar and exhibit E/Z
isomerism.1 At the other end, functional groups with Nsp2−Ar
and related cores often prefer twisted, even orthogonal,
geometries. Such groups include N-arylamides, carbamates,
ureas, heterocycles (lactams, imides, among many others), and
so on.2,3 Both the geometry of the N−Ar bond and its rotation
barrier are crucial features in areas as diverse as enzyme/
substrate binding,4 molecular torsional balances,5 and asym-
metric synthesis.6
When reactive intermediates are involved, asymmetric
reactions of axially chiral anilides (N-arylamides) are often
ultimately dictated by the configuration of the starting
material.6,7 This is because the onward reactions of the
intermediates, radicals, anions,68 and organometallic species,8
are faster than one or more bond rotations. The result is that
only a subset of the usually considered transition states is
actually available for onward reaction in one atropisomer of a
starting material. The remaining subset is only accessible from
the other atropisomer.
Figure 1. Asymmetric Heck reactions of anilide atropisomers. Two
limiting conditions: N−Ar axis unlocked (a) or locked (b).
Consider asymmetric Heck reactions of the o-iodoanilides, as
achiral palladium catalyst under ambient temperature con-
ditions where the lock was in effect (Figure 1b).9 High chirality
transfer from the axis of 3 to the stereocenter of 4 was
observed. This shows that N−Ar bond rotation does not occur
during the Heck process of the locked substrate 3. It also
suggests that asymmetric Heck reactions like those in Figure 1a
under “unlocked” conditions might be dynamic kinetic
resolutions where the stereochemistry-determining step is
shown in Figure 1a. These were pioneered by Overman for use
in natural product synthesis.10 Cyclization of o-haloanilides like
1 with various chiral palladium catalysts produces 3-alkenyl-2-
oxindoles 2 in high yields and ee’s. Like most anilides,
precursors 1 prefer the (E)-amide rotamer (N−Ar and CO
are trans), and the plane of the N−Ar group is roughly
orthogonal to the plane of the amide.11
To address the question of which step in the Heck reaction
dictates the enantioselectivity, we prepared enantioenriched 3
that has an added o-methyl group (shown in red) to “lock” the
N−Ar bond. Then we conducted Heck cyclizations with an
Received: February 25, 2013
Published: March 28, 2013
© 2013 American Chemical Society
4083
dx.doi.org/10.1021/jo400385t | J. Org. Chem. 2013, 78, 4083−4089

The Journal of Organic Chemistry
Article
oxidative addition of a chiral Pd species into one of the rapidly
equilibrating atropisomers of 1.
Scheme 1. Synthesis and Separation of the Racemic Syn/
Anti Atropisomers of 7 Shown in R Enantiomeric Series
A second class of asymmetric reactions of o-haloanilides, Pd-
catalyzed enolate cyclizations, also produce oxindoles (Figure
2a). These transformations were pioneered by Lee and
it is present simply because the precursor (2-bromo-4,6-
dimethylaniline, 8) is readily available.
Acylation of 8 with rac-2-phenylpropanoyl chloride followed
by alkylation of the resulting sec-amide with NaH and
iodomethane provided a mixture of N-methylated atropisomers
7. These were readily separated by flash chromatography. The
less polar isomer anti-7 was isolated in 58% yield as a white
solid, while the more polar isomer syn-7 was isolated as a clear
oil in 31% yield. Here “syn” and “anti” refer to the orientations
of the bromine atom on the aryl ring relative to the phenyl
group on the stereocenter in the depicted conformation. The
combined yield of the two racemates was 89%.
Figure 2. Hartwig cyclizations of o-haloanilide enolates provide
enantioenriched oxindoles (a) and stereogenic elements of their
precursors (b).
Relative configurations of 7 were assigned by NMR
spectroscopy with the assumption that geometries like that
shown in Scheme 1 are favored. Specifically, the C−H bond of
the stereocenter is oriented approximately anti to the amide
carbonyl group. The major isomer anti-7 has a shielded singlet
Hartwig,12 and several groups have introduced new chiral
ligands.13 Kundig’s NHC ligands, in particular, give exceptional
̈
1
ees with broad substrate scope.13b Generally speaking,
cyclization of o-haloanilides like 5 with a strong base (to
make the enolate) and a chiral catalyst provides 3-phenyl-2-
oxindoles 6.14 The mechanism of this reaction may be very
different from a Heck reaction, and Hartwig provided evidence
for palladacycle intermediates.12,13e
at 1.41 ppm in its H NMR spectrum. We assign this to the o-
methyl group and attribute the upfield shift to anisotropic
shielding by the phenyl ring. The p-methyl group in anti-7 and
both aryl methyl groups of syn-7 resonate in a narrow range
2.31−2.34 ppm.
Both diastereomers were resolved into their enantiomeric
components by preparative chiral HPLC on an (S,S)-Whelk-O1
column, with the anti diastereomer providing S,M-7 and R,P-7
and the syn diastereomer providing S,P-7 and R,M-7. The
absolute configuration of enantiomer R,P-7 was determined by
solving its X-ray crystal structure (Figure 3), which also
confirmed the anti relative configuration of both the N−Ar axis
and E-geometry of the amide C−N bond.
We set out to answer two questions about this reaction. First,
the commonly drawn depictions of 5 with a planar (Z)-amide
cannot represent energy minima. Precursors like 5 have three
stereogenic elements, the stereocenter, the amide C−N bond,
and the N−Ar axis (Figure 2b). What do molecules like 5 look
like in the ground state? And second, does the configuration of
either the stereocenter15 or the chiral axis play a key role in
these cyclizations?
To address both questions, we made and studied cyclizations
of stereoisomerically pure “locked” precursors bearing addi-
tional o-methyl groups. Here we show that such precursors exist
mostly as two diastereomers, both of which have an (E)-amide
bond. To our surprise, the lock was cracked under Hartwig
cyclization conditions, apparently because the amide enolate
has a much lower N−Ar rotation barrier than its conjugate acid
precursor. We report N−Ar rotation barriers for two substrates
in both protonated and enolate forms, and we discuss the
consequences of this reaction on Hartwig oxindole cyclizations.
Figure 3. Ball and stick representation of the X-ray crystal structure of
anti-isomer R,P-7.
RESULTS AND DISCUSSION
The last link in the structure assignment, absolute
configurations of the syn diastereomers of 7, was provided by
heating the anti diastereomers to induce N−Ar bond rotation;
R,P-7 equilibrated with R,M-7 and S,P-7 with S,M-7. A
complete rotation profile in the (R)-enantiomeric series is
shown in Figure 4. The N−Ar atropisomers (anti/syn) do not
interconvert at rt, but each (E)-amide rotamer is in equilibrium
with the corresponding (Z)-amide rotamer (N−Ar and CO
cis), which must also be twisted into two atropisomers, as
■
We selected N-methyl-N-(2-bromo-4,6-dimethylphenyl)-2-phe-
nylpropanamide 7 as the “locked” precursor for resolution and
cyclization (Scheme 1). The added o-methyl group serves as
the lock by raising the N−Ar rotation barrier; however, this
methyl group is not expected to significantly alter either
ground-state geometry or isomer populations relative to
5.7a,b,9,11 The added p-methyl group in 7 should not
significantly affect either barriers or rotamer populations, and
4084
dx.doi.org/10.1021/jo400385t | J. Org. Chem. 2013, 78, 4083−4089

The Journal of Organic Chemistry
Article
We conducted a few experiments with the stereopure
precursors in the presence of palladium catalysts and NHC
ligands with similar results.16,17 In other words, the precursors 7
approached full equilibrium at a rate that was faster than
formation of product 8 (Scheme 2). This shows the
Scheme 2. Base-Catalyzed Equilibration of Both the
Stereocenter and the Axis of 7 Occurs with or without a Pd
Catalyst and Ligand
Figure 4. Interconversion of rotational isomers of 7 shown in the R
enantiomeric series. At rt, amide E/Z rotation (vertical) is fast but N−
Ar rotation (horizontal) is very slow.
stereocontrol in the Hartwig reaction is not dictated by the
original stereocenter,15 but it also shows that the “locked
substrate” approach cannot be used to test for the importance
of N−Ar atropisomer configuration in the Hartwig reaction
because the base cracks the lock.
The results suggest that the mechanism for cracking the lock
is simply formation of the enolate. Deprotonation of course
equilibrates the stereocenter (by reversible protonation), but it
also must allow for rotation of the N−Ar axis of rotamers P-9
and M-9, as shown in Figure 5. Deprotonation can form
shown. In CDCl3 at rt, the minor amide rotamer R,P-7Z in the
anti series cannot be identified in the 1H NMR spectrum of the
major atropisomer (R,P-7E/R,P-7Z > 98/2), while the E/Z
ratio in the minor syn-atropisomer series, R,M-7E/R,M-7Z, is
about 91/9.
The interconversion of antiatropisomer R,P-7 and syn-
atropisomer R,M-7 was studied by heating pure samples of each
isomer at 100 °C in tert-butylbenzene with time course
monitoring by 1H NMR spectroscopy. The equilibrium ratio is
about 75/25 in favor of R,P-7. Data were processed in the usual
way (see the Supporting Information) to give a barrier from M
to P (minor to major) of ΔG⧧ = 30.5 kcal mol−1, while the
barrier from P to M (major to minor) is ΔG⧧ = 31.2 kcal mol−1.
These barriers mean that the atropisomers of 7 have significant
stability under the thermal conditions of Hartwig reactions (t1/2
of several days at 85 °C compared to reactions times of several
hours).
Neglecting temperature dependence of equilibrium con-
stants, R-7 exists at equilibrium as about 75% of antiseries
isomer R,P-7E, 23% of syn-series isomer R,M-7E and 2% of its
amide rotamer R,M-7Z. The fourth isomer R,P-7Z (amide
rotamer of the major isomer) is present only in trace amounts.
The depiction in Figure 4 implies that the mechanism for
N−Ar bond rotation is direct interconversion of R,P-7E and
R,M-7E. This is not necessarily the case because equilibration
could occur over the Z-isomers in a Curtin−Hammett scheme
or could even be geared (coupled) with amide rotation.18
Nonetheless, it is the magnitude of the barrier, not the
mechanism of the process, that is important for the
forthcoming discussion. Here we focus on the major (E)-
amide rotamers (98% total), but these are in equilibrium with
small amounts of the corresponding (Z)-rotamers (2% total).
We next conducted control experiments by exposing R,P-7 to
NaOtBu in DME at 85 °C. These are typical conditions for
asymmetric Hartwig cyclizations,12,13 but we omitted the
palladium and ligands. Reaction progress was followed by
chiral HPLC. We expected to see one new isomer from base-
catalyzed epimerization of the stereocenter, but we quickly saw
all three of the other isomers and the system soon reached
equilibrium (3/3/1/1). Thus, the base promotes isomerization
of both the stereocenter and, unexpectedly, the N−Ar axis.
Figure 5. Enolate atropisomers based on N−Ar bond rotation and
illustrated with E-amide−Z-enolates.
isomeric amide (E/Z)-enolates. But it erases the stereocenter,
so in a simple view the diastereomers of 7 become enantiomers.
The rotation is illustrated with (Z)-enolates (Ph and OM cis),
but (E)-enolates or even twisted (chiral) enolates may also be
present.
A time course experiment at 0 °C allowed us to estimate that
the barrier for N−Ar rotation is reduced from 31 kcal mol−1 in
the precursor to about 20 kcal mol−1 in the enolate. This is a
large effect, especially considering that the amide and the
enolate both get their axial chirality from two adjacent sp2
centers.
The measured value of N−Ar rotation of the enolate 9 is
only a crude estimate for at least two reasons: (1) the kinetic
analysis is inherently problematic because two interconversions
(epimerization and N−Ar bond rotation) occur more or less
simultaneously, and (2) the analysis assumes that the
deprotonation is quantitative, which may not be the case.13f
To make a more clear-cut measurement of an enolate N−Ar
rotation barrier, we removed the α-methyl group from 7 to
4085
dx.doi.org/10.1021/jo400385t | J. Org. Chem. 2013, 78, 4083−4089

The Journal of Organic Chemistry
Article
eliminate the stereocenter and we switched to a stronger base
to ensure complete deprotonation.
rac-N-(2-Bromo-4,6-dimethylphenyl)-2-phenylacetamide 10
was synthesized and then preparatively resolved into its two
enantiomers on an (S,S)-Whelk-O1 HPLC column (Scheme
3). Standard thermal equilibration experiments as above
geometries of 11 in Figure 5 are probably good approximations
of solution structures.
To measure the N−Ar rotation barrier of the enolate, 11 was
added to a solution of LDA at −38 °C, and then aliquots were
periodically quenched and ee was assessed by chiral HPLC.
Despite possible complexities with enolate aggregates, reason-
able first-order kinetics were observed with ΔG⧧ = 17 kcal
mol−1. A similar experiment at −45 °C gave a comparable result
(see the Supporting Information). Thus formation of the
enolate reduces the N−Ar rotation barrier by almost 14 kcal
mol−1.
Scheme 3. Interconversion of Stropisomers of 10 and
Assignment of Absolute Configuration by Methylation of the
Enolate
The origin of the barrier decrease is cloaked in the structure
of the enolate. Like its amide precursor, enolate 11 has both
N−Ar and N−C(O−)C rotamers, the latter being the
analogue of E/Z amide rotamers. Thus, N−Ar rotation could
occur either on an (E) or a (Z) amide enolate rotamer or even
be coupled to amide enolate rotation.
In resonance terms, deprotonation of an amide 10 introduces
cross-conjugation in the resulting amide enolate 11 (Figure 6).
provided the barrier to interconversion of the two atropisomers,
ΔG⧧ = 30.8 kcal mol−1. Scheme 3 shows only the major (E)-
amide rotamer of 10, but the (Z)-rotamer can be observed by
1H NMR spectroscopy and the E/Z ratio at rt in CDCl3 is
about 96/4.
To assign the absolute configuration of the atropisomers, we
deprotonated the first eluting enantiomer with lithium
diisopropylamide (LDA) at −78 °C and then added iodo-
methane at the same temperature. To ensure that no N−Ar
bond rotation occurred, the reaction was quenched with NH4Cl
after 10 min, and then the products were assessed by chiral
HPLC. About 60% of the mixture was methylated products
S,M-7 and R,M-7 in a ratio of 88/12. The remaining 40% was
recovered M-10. No P-atropisomers of either the precursor 10
or the products were detected by chiral HPLC, so N−Ar
rotation of the lithium enolate is slow relative to methylation at
−78 °C. A similar experiment with P-10 gave the atropisomers
S,P-7 and R,P-7 with similar conversion and ratio.
Figure 6. Suggested effects of pyramidalization: conversion of an
anilide to an enolate or o-anilide allows pyramidalization at N and
reduces the N−Ar rotation barrier.
This suggests that carbamates and ureas might be simple
models for amide enolates. The cross-conjugation in
carbamates and ureas lowers the C(O)−N bond rotation
barrier (E/Z isomerization) relative to comparable amides.20 In
addition, the N−Ar rotation barriers of N-aryl carbamates and
ureas are also lower than those of comparable N-aryl amides
(anilides).21 Thus, the barrier-reducing effect of cross-
conjugation still plays out at the N−Ar bond even though
the two connected π-systems (carbamate/urea and aryl ring)
are orthogonal. Presumably, cross-conjugation tends to length-
en the C(O)−NAr bond and allow it to twist. In turn, the N
can pyramidalize.
The enolate methylation provides the less stable syn-
atropisomer in excess, and the diastereoselectivity (88/12) is
surprisingly high given that the two ortho substituents (Me and
Br) shielding the two faces of enolate 11 are not that different
in size. This selectivity is even more surprising in light of the
results of Simpkins and co-workers.7h,i They observed low
selectivities with related axially chiral enolates that appear to be
much more favorably biased for face selectivity (one o-
t
substituent is H and the other is Bu). They attributed the
low selectivity to alkylation of Z-amide−Z-enolates in
Recent calculations suggest that pyramidalization of the
amide nitrogen atom is an important component of N−Ar
bond rotation.22 There is evidence that various amide-type
enolates are more prone to pyramidalize than their precursors,
presumably because amide resonance is decreased on
deprotonation.23 We speculate that this ability to pyramidalize
translates to enolates like 9 and 11 and that this is an important
reason for the large decrease in the N−Ar rotation barrier.
competition with E-amide−Z-enolates.
Compared to enolate 9, the structure of enolate 11 derived
from 10 and lacking the α-methyl group is more clear-cut
(Figure 5). Amide enolates with only one substituent on the
anionic carbon atom adopt a planar, (Z)-enolate geometry to
avoid A-strain between the enolate substituent (here Ph) and
one of the N-substituents (here the Ar ring).19 Thus, the
4086
dx.doi.org/10.1021/jo400385t | J. Org. Chem. 2013, 78, 4083−4089

The Journal of Organic Chemistry
Article
To the extent that the N atom of the enolate can
pyramidalize without a large energy penalty, the atropisomer-
ism becomes a rotation of an sp3−sp2 bond rather than an sp2−
sp2 bond. Other factors being equal, the former type of rotation
is much faster.3 Compare, for example, the barriers of 12 and
13 in Figure 6.24 In substrate 13, the N pyramidalizes because it
is part of an amide acetal. This reduces the N−Ar rotation
barrier from 29 to 23 kcal mol−1.
With this detailed knowledge of the rotational preferences of
o-methyl-substituted substrates 7 and 10, we then returned to
the substrate 5 for the Hartwig cyclization. This and most other
substrates used to date lack the o-methyl group.12,13 Figure 7
shows the rotational isomers of precursor 5 inside the box.
Outside the box are intermediates in a catalytic cycle that was
proposed by Hartwig12 and has been widely adopted.13
palladium species into the precursor 5 to give an intermediate
14 with the amide carbonyl group ligated to Pd*.12 This
intermediate could form directly by insertion of Pd* into a
minor (Z)-amide rotamer, present in trace amounts.26
This is a dynamic kinetic resolution because a chiral Pd-
species can select between four isomeric (Z)-amides (M/P,
shown; R/S, only R shown). Intermediate 14 may retain an
element of axial chirality from 5, though this is now best viewed
in the context of a ring conformation. In other words, if
chirality is retained, then atropisomer P-5Z gives 14a and M-5Z
gives 14b. If the ring flip that interconverts 14a and 14b is fast
(or if the ring is planar), then the effect of the prior dynamic
kinetic resolution may be erased. But if intermediates 14a,b are
rapidly interconverting, then the next step is also a dynamic
kinetic resolution.
The next key intermediate is the palladacycle 15,12 which
must result from at least two separate steps, enolate formation
and amide C−N bond rotation. The deprotonation step can
give (E/Z)-amide enolates at the enolate CC bonds while
the rotation step involves a new element of axial chirality. (The
amide C−N bond can rotate two different ways during the
process of complexation of Pd from O, amide bond rotation,
and recompletion to C.) Reductive elimination of 15 provides
the product 6.
Notice that the path from precursor to product is less direct
than it could be The amide C−N bond rotates twice, first from
(E) to (Z) in the precursor 5, then later back to (E) in 15.
Futhermore, the formation of 14 presumably occurs from the
less populated (Z)-rotamer of 5.26 A more direct process would
be reaction of a syn or anti isomer of 5E to give 15. This
shortcut could occur in as few as two steps, deprotonation and
oxidative addition (in either order). So if the accepted
mechanism is correct, then the steps of the longer path have
the lower barriers.
Setting the possible shortcut aside and returning to the
accepted mechanism, what is the stereochemistry determining
step? Presumably the conversion of 15 to 6 occurs with
retention of configuration at the stereocenter,13e so it is not this
step but a prior one. In that case, it could well involve dynamic
kinetic resolution. All of the suggested intermediates have
multiple isomers due to the stereocenter or enolate, the N−Ar
bond or ring conformation, and of course the chiral ligand. The
dynamic kinetic resolution could occur directly at the very last
stage, formation of 15, or it could occur earlier, with the
subsequent step(s) occurring irreversibly and with chirality
transfer.
Figure 7. Steps of the Hartwig oxindole synthesis from the viewpoint
of amide rotational isomers. Precursor 5 exists as eight isomers: four in
the R-series (shown) and four in the S-series (not shown).
In substrate 5, the N−Ar rotation barrier is significantly
decreased compared to 7/10 so the atropisomers cannot be
separated.7a−c Inspection of published H NMR spectra of 5
1
and related molecules shows that there is about a 3/1 ratio of
rotational isomers. These are not (E/Z)-amide isomers, but
instead anti/syn atropisomers of the (E)-amide isomer (anti-5E
and syn-5E). Like the o-methyl group in 7 (Scheme 1), the o-
hydrogen in 5 is significantly shielded in the major antirotamer.
It is difficult to find the two minor (Z)-amide rotamers of 5
because the N−Ar atropisomers of 5 cannot be separated, and
we could not identify clear peaks for these (Z)-rotamers in
published spectra. So precursor 5 is at equilibrium at room
temperature with an anti/syn ratio of about 75/25. Both major
rotational isomers are (E)-amides, but there are probably trace
amounts of the corresponding (Z)-amide rotamers.11
CONCLUSIONS
■
We have shown that o-haloanilide precursors of Hartwig
oxindole cyclizations do not exist as planar (Z)-rotamers but
instead twisted (E)-rotamers. Precursors for such cyclization
typically have a stereocenter, so two diastereomeric atro-
pisomers are present. From the standpoint of Hartwig
cyclizations, the o-methyl group of moldel substrates 7 and
10 provides an artificially high barrier for atropisomerism (30−
31 kcal mol−1). In actual substrates like 5 lacking this Me group,
the N−Ar rotation barrier is still substantial (about 15−20 kcal
mol−1),21 but atropisomerism is nonetheless rapid compared to
onward reactions of these substrates under typical reaction
conditions (80−100 °C). This means that the first reaction of a
chiral palladium species with a precursor (amide, or its amide
enolate) in the Hartwig oxindole cyclization is a dynamic
kinetic resolution.
The basic steps of the mechanism12 are clear enough,
oxidative addition, enolate formation, conversion to an Ar−Pd-
enolate species, and reductive elimination, but the timing is less
clear. Here we focus on conformational features and represent
the chiral palladium catalysts simply as Pd*.25 Hartwig
proposed that the first step is oxidative addition of the
4087
dx.doi.org/10.1021/jo400385t | J. Org. Chem. 2013, 78, 4083−4089

The Journal of Organic Chemistry
Article
Major diastereoisomer anti-7:13b mp 108 °C; 1H NMR (CDCl3, 400
MHz) δ 7.36 (s, 1H), 7.16−7.15 (m, 3H), 6.93−6.91 (m, 2H), 6.86 (s,
1H), 3.36 (q, J = 6.8 Hz, 1H), 3.06 (s, 3H), 2.33 (s, 3H), 1.43 (d, J =
6.8 Hz, 3H), 1.41 (s, 3H); 13C NMR (CDCl3, 100 MHz) δ 174.1,
141.1, 139.6, 139.1, 138.1, 131.4, 131.1, 128.3, 127.7, 126.7, 124.1,
44.0, 34.7, 20.83, 20.77, 17.4; FTIR (neat, cm−1) 2985, 2934, 1659,
1477, 1451, 1375, 1280, 1117, 1090, 866, 813, 746; HRMS (EI) calcd
for C18H21BrNO 346.0807, found 346.0798.
The proposed mechanism for the Hartwig cyclization12
involves two amide C−N rotations (from (E) to (Z) and back)
that are formally unnecessary. This is not to say that they
cannot occur (E/Z rotations of the precursor are clearly rapid
relative to the reaction rate). However, it is also worth
considering the possibility of a more direct conversion of the
major (E)-rotamer of the precursor 5 to the product 6 (via 15)
without any intervening amide bond rotations.
Regardless of the mechanism, all the intermediates in these
cyclizations may be mixtures of multiple stereoisomers due to
presence of stereocenters in the precursor and the ligand
overlaid on several rotatable bonds. Accordingly, the step that
determines stereochemistry in asymmetric Hartwig cyclizations
could be a dynamic kinetic resolution. This might be at the
penultimate step (formation of 15), or it could occur at an
earlier step with the subsequent step(s) occurring irreversibly
and with chirality transfer.
The most general conclusion of this work is that axially chiral
anilide enolates have a much lower barrier to atropisomeriza-
tion than their amide conjugate acids. This has implications
beyond Hartwig cyclizations. For example, enantiopure anilides
are often obtained by resolution and re-equilibration of the
undesired atropisomer,7−9 but the equilibration requires forcing
conditions (high temperature, long times). Base-catalyzed
equilibrations are an attractive alternative. And enolate
reactions can be controlled either by the axial chirality of the
precursors (by conducting them at low temperatures as we did
with M-10 in Scheme 3), or by dynamic kinetic resolution (by
conducting them at higher temperatures). In essence, the
deprotonation can be considered as a simple switch, turning on
rapid rotation. Protonation turns rotation off and thereby flips
the switch back.
Minor diastereoisomer syn-7: 1H NMR (CDCl3, 500 MHz) δ 7.21−
7.15 (m, 4H), 7.09 (s, 1H), 7.03−7.01 (m, 2H), 3.30 (q, J = 7.0 Hz,
1H), 3.11 (s, 3H), 2.34 (s, 3H), 2.30 (s, 3H), 1.41 (d, J = 7.0 Hz, 3H);
13C NMR (CDCl3, 125 MHz) δ 174.0, 140.7, 139.6, 138.3, 137.6,
131.7, 130.9, 128.1, 128.0, 126.6, 124.6, 43.3, 35.7, 20.9, 20.7, 18.4;
FTIR (neat, cm−1) 2968, 2927, 1663, 1476, 1453, 1375, 1277, 1153,
1112, 814, 739; HRMS (EI) calcd for C18H21BrNO 346.0807, found
346.0798.
N-(2-Bromo-4,6-dimethylphenyl)-2-phenylacetamide. Phe-
nylacetyl chloride (160 μL, 1.2 mmol) was added to a solution of 2-
bromo-3,5-dimethylaniline (200 mg, 1 mmol) and triethylamine (420
μL, 3 mmol) in DCM (10 mL) at 0 °C under argon. The mixture was
warmed to rt and monitored by TLC. After 2 h, water was added, and
the phases were separated. The aqueous layer was extracted with
DCM. The organic phase was washed with brine, dried over MgSO4,
filtrated, and evaporated. The residue was purified by flash
chromatography (DCM) on silica gel to afford 142 mg of a white
solid (45% yield): mp 96 °C; 1H NMR (CDCl3, 500 MHz) δ 7.4−7.38
(m, 4H), 7.34−7.33 (m, 1H), 7.2 (s, 1H), 6.93 (s, 1H), 6.88 (s, 1H),
3.77 (s, 2H), 2.24 (s, 3H), 2.15 (s, 3H); 13C NMR (CDCl3, 125 MHz)
δ 169.3, 138.5, 137.6, 134.6, 131.1, 130.6, 130.5, 129.7, 129.1, 127.5,
121.7, 43.9, 20.6, 18.9; FTIR (neat, cm−1) 3226, 3026, 1654, 1604,
1565, 1519, 1496, 1453, 1350, 1282, 844, 721; HRMS (EI) calcd for
C16H17BrNO 318.0494, found 318.0468.
N-(2-Bromo-4,6-dimethylphenyl)-N-methyl-2-phenylaceta-
mide (10). To a THF slurry of NaH (186 mg, 4.7 mmol, 60% in oil)
at 0 °C was added N-(2-bromo-4,6-dimethylphenyl)-2-phenylaceta-
mide (1.14 g, 3.6 mmol) dissolved in THF (30 mL). The reaction
mixture was stirred until the solution became clear, and then
iodomethane (245 μL, 3.9 mmol) was added dropwise. The solution
was warmed to room temperature. The reaction mixture was
monitored by TLC and quenched with water after 5 h. The resulting
mixture was extracted with EtOAc. The combined organic layers were
washed with brine and dried over MgSO4, filtrated, and evaporated.
The residue was purified by flash chromatography (Hex/EtOAc 95:5)
EXPERIMENTAL SECTION
■
General Information. See the Supporting Information.
N-(2-Bromo-4,6-dimethylphenyl)-2-phenylpropanamide. 2-
Phenylpropanoyl chloride (279 mg, 1.65 mmol) was added to a
solution of 2-bromo-3,5-dimethylaniline (300 mg, 1.5 mmol) and
triethylamine (230 μL, 1.65 mmol) in DCM (15 mL) at 0 °C under
argon. The mixture was warmed to rt and monitored by TLC. After 2
h, water was added, and the phases were separated. The aqueous layer
was extracted with DCM. The organic phase was washed with brine,
dried over MgSO4, filtrated, and evaporated. The residue was purified
by flash chromatography (DCM) on silica gel to afford 439 mg of a
1
on silica gel to afford 790 mg of a colorless oil (66% yield): H NMR
(CDCl3, 600 MHz) δ 7.36 (s, 1H), 7.32−7.19 (m, 3H), 7.04−7.00 (m,
3H), 3.41 (d, J = 15 Hz, 1H), 3.23 (d, J = 15 Hz, 1H), 3.12 (s, 3H),
2.34 (s, 3H), 1.99 (s, 3H); 13C NMR (CDCl3, 150 MHz) δ 171.1,
139.9, 138.4, 138.3, 134.7, 131.7, 131.2, 129.5, 128.2, 126.6, 123.9,
40.8, 34.6, 20.8, 18.1; FTIR (neat, cm−1) 3028, 2922, 1665, 1476,
1433, 1369, 1305, 1152, 1103, 852, 813, 722; HRMS (EI) calcd for
C17H19BrNO 332.0650, found 332.0629.
1
white solid (88% yield): mp 128 °C; H NMR (CDCl3, 400 MHz) δ
7.45−7.37 (m, 4H), 7.33−7.31 (m, 1H), 7.19 (s, 1H), 6.93 (s, 1H),
6.73 (bs, 1H), 3.81 (q, J = 7.2 Hz, 1H), 2.24 (s, 3H), 2.11 (s, 3H), 1.65
(d, J = 7.2 Hz, 3H); 13C NMR (CDCl3, 100 MHz) δ 172.5, 140.8,
138.3, 137.6, 131.1, 130.6, 130.5, 129.0, 127.9, 127.6, 121.6, 47.3, 20.6,
18.9, 18.1; FTIR (neat, cm−1) 3222, 3026, 2977, 1655, 1527, 1479,
850, 718; HRMS (EI) calcd for C17H19BrNO 332.0650, found
332.0638.
N-(2-Bromo-4,6-dimethylphenyl)-N-methyl-2-phenylpropa-
namide (7). To a THF slurry of NaH (40 mg, 1 mmol, 60% in oil) at
0 °C was added N-(2-bromo-4,6-dimethylphenyl)-2-phenylpropana-
mide (0.3 g, 0.9 mmol) dissolved in THF (10 mL). The reaction
mixture was stirred until the solution became clear, and then
iodomethane (73 μL, 1.17 mmol) was added dropwise. The mixture
was warmed to room temperature. The reaction mixture was
monitored by TLC and quenched with water upon completion. The
resulting mixture was extracted with EtOAc. The combined organic
layers were washed with brine, dried over MgSO4, filtrated, and
evaporated. The residue was purified by flash chromatography (hex/
EtOAc 95:5) on silica gel to afford 175 mg of anti-7 as white solid and
95 mg of syn-7 as a colorless oil (89% yield, 3:1).
ASSOCIATED CONTENT
■
S
* Supporting Information
Details on resolution and equilibration studies, copies of typical
chromatograms, copies of NMR spectra of all new compounds,
an ORTEP diagram, and X-ray data for the crystal structure
(CIF). This material is available free of charge via the Internet
at http://pubs.acs.org/.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: daylite@pitt.edu, curran@pitt.edu.
Notes
The authors declare no competing financial interest.
4088
dx.doi.org/10.1021/jo400385t | J. Org. Chem. 2013, 78, 4083−4089

The Journal of Organic Chemistry
Article
(c) Hillgren, J. M.; Marsden, S. P. J. Org. Chem. 2008, 73, 6459.
ACKNOWLEDGMENTS
■
(d) Sole.
S. P.; Watson, E. L.; Raw, S. A. Org. Lett. 2008, 10, 2905. (f) Durbin,
M. J.; Willis, M. C. Org. Lett. 2008, 10, 1413. (g) Sole, D.; Serrano, O.
́
, D.; Serrano, O. J. Org. Chem. 2008, 73, 9372. (e) Marsden,
We thank the National Institutes of Health for funding of this
work, and we thank Prof. E. Peter Kundig, University of
Geneva, for samples of NHC ligands and helpful information.
J.M. thanks the Fulbright Foundation for a postdoctoral
fellowship award.
̈
́
J. Org. Chem. 2008, 73, 2476. (h) Satyanarayana, G.; Maier, M. E.
Tetrahedron 2008, 64, 356. (i) Watson, E. L.; Marsden, S. P.; Raw, S.
A. Tetrahedron Lett. 2009, 50, 3318. (j) Ackermann, L.; Vicente, R.;
Hofmann, N. Org. Lett. 2009, 11, 4274. (k) Pan, X.; Wilcox, C. S. J.
Org. Chem. 2010, 75, 6445−6451.
REFERENCES
■
(15) Control by the stereocenter is a “memory of chirality” process as
defined in (a) and (b), but the term has its detractors (c). (a) Zhao, H.
W.; Hsu, D. C.; Carlier, P. R. Synthesis 2005, 1−16. (b) Kawabata, T.;
Fuji, K. Top. Stereochem. 2003, 23, 175−205. (c) Cozzi, F.; Siegel, J. S.
Org. Biomol. Chem. 2005, 3, 4296−4298.
(1) Stewart, W. E.; Siddall, T. H. Chem. Rev. 1970, 70, 517−551.
(2) (a) Oki, M.; Allinger, N. L.; Eliel, E.; Wilen, S. H. Topics in
Stereochemistry; Wiley-Interscience: New York, 1983; Vol. 14, pp 1−
81. (b) Gallo, R.; Roussel, C.; Berg, U. Adv. Hetercycl. Chem 1988, 43,
174−299. (c) Eliel, E. L.; Wilen, S. Stereochemistry of Organic
Compunds; Wiley-Interscience: New York, 1994. (d) Wolf, C. Dynamic
Stereochemistry of Organic Compounds; RSC Publishing: Cambridge,
2008.
(16) Bruch, A.; Ambrosius, A.; Frohlich, R.; Studer, A.; Guthrie, D.
̈
B.; Zhang, H.; Curran, D. P. J. Am. Chem. Soc. 2010, 132, 11452−
11454.
(17) Early in this work, Kundig made a racemic sample of 7 by a
̈
(3) Examples of related cores include molecules with Nsp2−alkenyl
bonds: Clark, A. J.; Curran, D. P.; Geden, J. V.; James, N.; Wilson, P. J.
Org. Chem. 2011, 76, 4546−4551.
similar procedure and cyclized it in low ee with a chiral Pd-catalyst (ref
13b). By comparing spectra, we learned that he isolated exclusively
major atropisomer. The minor atropisomer is more polar and was
evidently removed during product purification. This is of little
importance since the atropisomers equilibrate under the cyclization
conditions.
(4) LaPlante, S. R.; D. Fader, L.; Fandrick, K. R.; Fandrick, D. R.;
Hucke, O.; Kemper, R.; Miller, S. P. F.; Edwards, P. J. J. Med. Chem.
2011, 54, 7005−7022.
(5) (a) Wilcox, C. S.; Bhayana, B. Angew. Chem. 2007, 46, 6833−
6836. (b) Wilcox, C. S.; Paliwal, S.; Kim, E. J. Am. Chem. Soc. 1998,
120, 11192−11193.
(18) (a) Pirkle, W. H.; Welch, C. J.; Zych, A. J. J. Chrom. A 1993,
648, 101−109. (b) Clayden, J.; Pink, J. H. Angew. Chem., Int. Ed. Engl.
1998, 37, 1937−1939.
(19) Mekelburger, H. B.; Wilcox, C. S. In Comprehensive Organic
Synthesis; Trost, B. M., Fleming, I., Eds.; Pergamon Press: Oxford,
1991; Vol. 2, p 99.
(6) Curran, D. P.; Qi, H.; Geib, S. J.; DeMello, N. C. J. Am. Chem.
Soc. 1994, 116, 3131−3132.
(7) (a) Guthrie, D. B.; Geib, S. J.; Curran, D. P. J. Am. Chem. Soc.
2009, 131, 15492−15500. (b) Petit, M.; Geib, S. J.; Curran, D. P.
Tetrahedron 2004, 60, 7543−7552. (c) Curran, D. P.; Chen, C. H. T.;
Geib, S. J.; Lapierre, A. J. B. Tetrahedron 2004, 60, 4413−4424.
(d) Curran, D. P.; Liu, W. D.; Chen, C. H.-T. J. Am. Chem. Soc. 1999,
121, 11012−11013. (e) Ohnishi, Y.; Sakai, M.; Nakao, S.; Kitagawa, O.
Org. Lett. 2011, 13, 2840−2843. (f) Kitagawa, O.; Yoshikawa, M.;
Tanabe, H.; Morita, T.; Takahashi, M.; Dobashi, Y.; Taguchi, T. J. Am.
Chem. Soc. 2006, 128, 12923−12931. (g) Kitagawa, O.; Takahashi, M.;
Yoshikawa, M.; Taguchi, T. J. Am. Chem. Soc. 2005, 3676−3677.
(h) Hughes, A. D.; Price, D. A.; Simpkins, N. S. J. Chem. Soc., Perkin
Trans. 1 1999, 1295−1304. (i) Hughes, A. D.; Price, D. A.; Shishkin,
O.; Simpkins, N. S. Tetrahedron Lett. 1996, 37, 7607−7110.
(8) Guthrie, D. B.; Curran, D. P. Org. Lett. 2009, 11, 249−251.
(9) Lapierre, A. J. B.; Geib, S. J.; Curran, D. P. J. Am. Chem. Soc.
2007, 129, 494−495.
(20) (a) Deetz, M. J.; Forbes, C. C.; Jonas, M.; Malerich, J. P.; Smith,
B. D.; Wiest, O. J. Org. Chem. 2002, 67, 3949−3952. (b) Smith, B. D.;
Goodenough-Lashua, D. M.; D’Souza, C. J. E.; Norton, K. J.; Schmidt,
L. M.; Tung, J. C. Tetrahedron Lett. 2004, 45, 2747−2749.
(21) (a) Julia, S.; Ginebreda, A.; Sala, P.; Sancho, M.; Annunziata, R.;
Cozzi, F. Org. Mag. Res. 1983, 21, 573−575. (b) Adler, T.; Bonjoch, J.;
Clayden, J.; Font-Bardía, M.; Pickworth, M.; Solans, X.; Sole,
́
D.;
Vallverdu, L. Org. Biomol. Chem. 2005, 3, 3173−3183. (c) For an
́
example of pKa-dependent cross-conjugation effects on rotation
barriers of heteroaryl amide C−C bonds, see: Welch, C. J.; Biba,
M.; Pye, P.; Angelaud, R.; Egbertson, M. J. Chromatogr. B 2008, 875,
118−121.
(22) Suzumura, N.; Kageyama, M.; Kamimura, D.; Inagaki, T.;
Dobashi, Y.; Hasegawa, H.; Fukaya, H.; Kitagawa, O. Tetrahedron Lett.
2012, 53, 4332−4336.
(23) Seebach, D.; Maetzke, T.; Petter, W.; Kloetzer, B.; Plattner, D.
A. J. Am. Chem. Soc. 1991, 113, 1781−1786.
(24) Ates, A.; Curran, D. P. J. Am. Chem. Soc. 2001, 123, 5130−5131.
(25) For an expanded mechanistic discussion, see ref 13f.
(26) Alternatively, the Pd* could oxidatively add to a major (E)-
amide isomer, followed by amide bond rotation and ligation of the
amide O to Pd.
(10) (a) Dounay, A. B.; Overman, L. E. Chem. Rev. 2003, 103, 2945−
2963. (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.
(c) Overman, L. E.; Poon, D. J. Angew. Chem., Int. Ed. Engl. 1997, 36,
518−521.
(11) (a) Curran, D. P.; Hale, G. R.; Geib, S. J.; Balog, A.; Cass, Q. B.;
Degani, A. L. G.; Hernandes, M. Z.; Freitas, L. C. G. Tetrahedron:
Asymmetry 1997, 8, 3955−3975. (b) Saito, S.; Toriumi, Y.; Tomioka,
N.; Itai, A. J. Org. Chem. 1995, 60, 4715−4720. (c) Itai, A.; Toriumi,
Y.; Saito, S.; Kagechika, H.; Shudo, K. J. Am. Chem. Soc. 1992, 114,
10649−10650.
(12) Lee, S.; Hartwig, J. F. J. Org. Chem. 2001, 66, 3402−3415.
(13) (a) Liu, L.; Ishida, N.; Ashida, S.; Murakami, M. Org. Lett. 2011,
13, 1666−1669. (b) Jia, Y.-X.; Katayev, D.; Bernardinelli, G.; Seidel, T.
M.; Kundig, E. P. Chem.Eur. J. 2010, 16, 6300−6309. (c) Wurtz, S.;
̈
̈
Lohre, C.; Frohlich, R.; Bergander, K.; Glorius, F. J. Am. Chem. Soc.
̈
2009, 131, 8344−8345. (d) Arao, T.; Kondo, K.; Aoyama, T.
Tetrahedron Lett. 2006, 47, 1417−1420. (e) Trillo, R. B.; Leven, M.;
Neudorfl, J. M.; Goldfuss, B. Adv. Synth. Catal. 2012, 354, 1451−1465.
̈
(f) Pan, X. Ph.D. Thesis, University of Pittsburgh, 2012, http://etd.
library.pitt.edu/ETD/available/etd-05282011-005856/.
(14) More generally, palladium-catalyzed reactions between enolates
and aryl halides provide diverse ring types and sizes. See, for example:
(a) Honda, T.; Namiki, H.; Satoh, F. Org. Lett. 2001, 3, 631.
(b) MacKay, J. A.; Bishop, R. L.; Rawal, V. H. Org. Lett. 2005, 7, 3421.
4089
dx.doi.org/10.1021/jo400385t | J. Org. Chem. 2013, 78, 4083−4089