Radical and heck cyclizations of Diastereomeric o -haloanilide atropisomers

Article abstract of DOI:10.1021/ja108795x

The outcomes of radical cyclizations and Heck reactions of N-(cyclohex-2-enyl)-N-(2-iodophenyl)acetamides depend critically on the configurations of the chiral axis and the stereocenter. In substrates without an ortho-methyl group, the diastereomeric prec

Full text of DOI:10.1021/ja108795x

Published on Web 12/10/2010
Radical and Heck Cyclizations of Diastereomeric o-Haloanilide
Atropisomers
David B. Guthrie, Steven J. Geib, and Dennis P. Curran*
Department of Chemistry, UniVersity of Pittsburgh, Pittsburgh, PennsylVania 15261, United States
Received September 29, 2010; E-mail: curran@pitt.edu
Abstract: The outcomes of radical cyclizations and Heck reactions of N-(cyclohex-2-enyl)-N-(2-iodophe-
nyl)acetamides depend critically on the configurations of the chiral axis and the stereocenter. In substrates
without an ortho-methyl group, the diastereomeric precursors interconvert slowly at ambient temperatures.
Cyclization of enriched mixtures of diastereomers provided similar yields of acetyl tetrahydrocarbazoles or
dihydrocarbazoles, suggesting that interconversion of the radical or organometallic intermediates also occurs.
Diastereomers of N-(cyclohex-2-enyl)-N-(2-iodo-4,6-dimethylphenyl)acetamides with an additional ortho-
methyl group did not interconvert at ambient temperatures and were readily resolved. In radical cyclizations,
syn diastereomers were prone to cyclize, while anti isomers were not. Strikingly, Heck reactions gave the
opposite result; anti isomers were prone to cyclization and syn isomers were not. Heck reactions of allylic
acetates occur with ꢀ-hydride elimination when acetate is trans to palladium and with ꢀ-acetoxy elimination
when acetate is cis. This is surprising because prior studies have suggested that a trans relationship of
palladium and acetoxy is essential for acetate elimination. Analyses of the results provide insights into
mechanisms for radical cyclization and for insertion and elimination in the Heck reaction.
Introduction
Radical cyclizations of axially chiral o-haloanilides typically
occur with high chirality transfer from the axis to the new
stereocenter. In one major class of reactions, cyclizations of
acryloyl anilides provide 2-indolones with a new stereocenter
in high selectivity.1 This chirality transfer is not perturbed by
nearby stereocenters. For example, valine-derived (M,S)-1 and
(P,S)-1 can be separated and cyclized to give predominantly
(R,S)-2 and (S,S)-2, respectively (Figure 1).2 The chiral axis,
not the nearby stereocenter, controls the outcome, so either
configuration of the new stereocenter in 2 can be accessed from
a single enantiomer of valine.
Likewise, radical cyclizations of N-allyl-o-haloanilides also
occur with high chirality transfer (Figure 2).3 For example,
cyclization of (M)-3 provides (R)-4 with high chirality transfer.
This result shows that no rotation about the N-Ar bond occurs
in the radical intermediate and that there is a facially selective
cyclization. The transition-state model for such cyclizations is
shown in 5. Related anionic cyclizations and Heck reactions
are similarly stereoselective.4,5
Figure 1. Radical cyclizations of diastereomeric valine-derived o-iodoa-
nilides.
a substituent, resulting in diastereomers as shown in 6a,b. Unlike
1 in Figure 1, whose stereocenter is extraannular (outside of
the forming ring), the new stereocenters of 6a,b are intraannular
(part of the forming ring). Is the outcome of these cyclizations
controlled by the chiral axis, the stereocenter, or both? There
are two examples of such reactions in the literature;6,7 however,
it is not currently clear what controls selectivity.
We were interested in perturbing such cyclizations by
replacing one of the hydrogen atoms on the N-CH2 group with
(1) (a) Curran, D. P.; Liu, W. D.; Chen, C. H.-T. J. Am. Chem. Soc. 1999,
121, 11012–11013. (b) Ates, A.; Curran, D. P. J. Am. Chem. Soc.
2001, 123, 5130–5131. (c) Petit, M.; Geib, S. J.; Curran, D. P.
Tetrahedron 2004, 60, 7543–7552.
Here we describe the results of a detailed study of radical
cyclizations of N-cyclohexenyl-o-iodoanilides. It turns out that
(2) (a) Jones, K.; McCarthy, C. Tetrahedron Lett. 1989, 30, 2657–2660.
(b) Petit, M.; Lapierre, A. J. B.; Curran, D. P. J. Am. Chem. Soc.
2005, 127, 14994–14995.
(6) (a) Alcaide, B.; Almendros, P.; Rodr´ıguez-Vicente, A.; Ruiz, M. P.
Tetrahedron 2005, 61, 2767–2778. (b) Alcaide, B.; Moreno, A. M.;
Rodr´ıguez-Vicente, A.; Sierra, M. A. Tetrahedron: Asymmetry 1996,
7, 2203–2206.
(3) Curran, D. P.; Chen, C. H. T.; Geib, S. J.; Lapierre, A. J. B.
Tetrahedron 2004, 60, 4413–4424.
(4) Guthrie, D. B.; Curran, D. P. Org. Lett. 2009, 11, 249–251.
(5) Lapierre, A. J. B.; Geib, S. J.; Curran, D. P. J. Am. Chem. Soc. 2007,
129, 494–495.
(7) Lampard, C.; Murphy, J. A.; Rasheed, F. Tetrahedron Lett. 1994, 35,
8675–8678.
9
10.1021/ja108795x  2011 American Chemical Society
J. AM. CHEM. SOC. 2011, 133, 115–122 115

A R T I C L E S
Guthrie et al.
Scheme 1. Synthesis of Diastereomeric o-Iodoanilides
Figure 2. Example of chirality transfer in radical cyclizations of N-allyl-
o-haloanilides, transition-state model 5 showing diastereotopic hydrogens
targeted for replacement, and diastereomeric precursors 6a,b of cyclizations
with potential for intraannular chirality transfer.
was then acylated with acetyl chloride to produce a mixture of
atropisomers 9a. The two diastereomers were partially separable
by column chromatography; the first eluting compound, anti-
9a, was isolated in 58% yield in a 95/5 dr (anti-to-syn ratio).
The second eluting compound, syn-9a, was collected in 22%
yield in an 18/82 dr. The inability to isolate atropisomerically
pure samples of 9a is due to competitive N-Ar bond rotation
during chromatography and isolation (see below).
The pair of o,p-dimethylated substrates 9b was made in the
same fashion. Alkylation of 7b with 3-bromocyclohexene
afforded 8b in 61% yield. Acylation with acetyl chloride
produced two diastereomers of 9b, which were separated by
flash chromatography. Less polar anti-9b (45% yield) was
followed by syn-9b (40% yield). There is no interconversion
of these atropisomers during isolation, so the samples are
isomerically pure.
Single crystals of both syn- and anti-9b were grown by slow
evaporation from a 3:1 hexane:CH2Cl2 solvent mixture, and the
X-ray crystal structures were solved. Isomer anti-9b crystallized
with a pair of enantiomers in the unit cell; the (P,R) structure
is shown in Figure 3. In this structure, the planes of the amide
group and the aromatic ring are nearly perpendicular, with a
dihedral angle of 84.1°. The N-cyclohexenyl bond is rotated
such that the hydrogen R to the amide nitrogen is pointing
toward the o-iodine. The dihedral angle between this C-H bond
and the N-carbonyl amide bond is 46.6°. The cyclohexene ring
exists in a half-chair conformation with the amide group in a
pseudoaxial position. In this structure, the iodine atom is held
away from the alkene, implying that the radical derived from
anti-9b is not predisposed for cyclization.
both the chiral axis and the stereocenter are important in
controlling the outcome of the reaction. This results in cycliza-
tions of diastereomeric precursors that give structurally different
products (rather than diastereomeric products) with very high
selectivity.8 We have previously reported one example of such
a reaction in the context of a new radical phosphanylation with
memory of chirality.9 Here we report rotation barriers, crystal
structures, and tin hydride experiments that allow us to
understand this result.
We also extend the study to Heck cyclizations10 of the same
precursors with surprising results. Again, diastereomeric precur-
sors react to give different (rather than stereoisomeric) products,
but the pathways of the Heck reactions are not anticipated by
the radical cyclizations. The results provide insights into both
the insertion and ꢀ-hydride elimination steps of the Heck
reaction.11
Results and Discussion
Synthesis and Configuration of Atropisomeric N-Cyclohe-
xenyl-o-iodoanilides. Introduction of a stereocenter on the allyl
substituent of an axially chiral anilide provides products with
two stereocenters and therefore introduces a new element of
diastereoselectivity in subsequent cyclizations. To simplify this
aspect, we chose an N-cyclohexenyl substituent, which was
expected to provide only products with a cis ring fusion.12 Two
pairs of diastereomeric anilidesswith and without ortho-methyl
substituentsswere prepared as shown in Scheme 1.
2-Iodoaniline 7a was alkylated with racemic 3-bromocyclo-
hexene to produce N-cyclohexenyl aniline 8a in 87% yield. This
Slow evaporation of a solution containing racemic syn-9b
deposited several crystals. The one chosen for analysis consisted
(8) Kumar, R. R.; Kagan, H. B. AdV. Synth. Catal. 2010, 352, 231–242.
(9) Bruch, A.; Ambrosius, A.; Fro¨hlich, R.; Studer, A.; Guthrie, D. B.;
Zhang, H.; Curran, D. P. J. Am. Chem. Soc. 2010, 132, 11452–11454.
(10) (a) The Mizoroki-Heck Reaction; Oestreich, M., Ed.; John Wiley &
Sons: Chichester, U.K., 2009. Gooflen, L.; Baumann, K. In Handbook
of C-H Transformations; Dyker, G., Ed.; Wiley-VCH: Weinheim,
2005; pp 277-286. (b) Alonso, D. A.; Na´jera, C. In Compounds with
All-Carbon Functions. Alkenes; de Meijere, A., Ed.; Georg Thieme
Verlag: Stuttgart, 2009; Vol. 47a, pp 439-479.
(11) (a) Jutand, A. In The Mizoroki-Heck Reaction; Oestreich, M., Ed.;
John Wiley & Sons: Chichester, U.K., 2009; pp 1-50. (b) Knowles,
J. P.; Whiting, A. Org. Biomol. Chem. 2007, 5, 31–44.
(12) Curran, D. P.; Porter, N. A.; Giese, B. Stereochemistry of Radical
Reactions: Concepts, Guidelines, and Synthetic Applications; VCH:
Weinheim, 1996; pp 51-71.
Figure 3. ORTEP and traditional representations of anti-9b.
9
116 J. AM. CHEM. SOC. VOL. 133, NO. 1, 2011

Cyclizations of Haloanilide Atropisomers
A R T I C L E S
Figure 5. Equilibration plot of syn-9a and anti-9a over time.
Figure 4. ORTEP and traditional representations of the two crystal forms
of syn-9b.
of only the (M,R) enantiomer, containing two unique structures
in the unit cell (Figure 4). In both structures the N-aryl plane
was again nearly perpendicular to the plane of the amide, with
dihedral angles of 84.8° and 85.6°. However, the two structures
differ in the orientation of the N-cyclohexenyl group. In the
left structure of Figure 4, the allylic hydrogen R to nitrogen is
pointed toward the ortho-methyl group, with a dihedral angle
C-HR/N-carbonyl of 48.5°. In the right structure, the R-hy-
drogen is oriented toward the o-iodine, and the dihedral angle
between the C-HR bond and carbonyl-N bond is -41.5°. This
molecule showed only one set of resonances in 1H NMR spectra,
so it is likely that these conformations rapidly interconvert in
solution. Importantly, the o-iodine radical precursor is near the
alkene, so the radical derived from syn-9b seems predisposed
to cyclize.
Figure 6. Barriers to rotation for N-cyclohexenyl atropisomers of 9a,b.
comparable to the experimentally determined barrier to rotation
for a related N-cyclohexyl o-iodoanilide.13
The barriers to rotation for 9b were determined in the same
manner. However, 9b is much more stable, and heating to 153
°C was necessary to effect rotation on a practical time scale
(see Supporting Information). The rotation of syn-9b to anti-
9b has a barrier of ∆G‡rot ) 36.0 kcal/mol, and the reverse
process has ∆G‡rot ) 36.6 kcal/mol.
Radical Reactions of N-Cyclohexenyl Atropisomers 9a,b. We
next subjected both atropisomers of 9a and 9b to typical tin
hydride reductions (Scheme 2). An 18/82 anti/syn mixture of
9a was dissolved in a minimal amount of benzene with Bu3SnH,
and Et3B initiator was added rapidly to the stirred solution
(Scheme 2). Despite the high initial concentration of tin hydride
([Bu3SnH]i ) 0.8 M), the cyclization product 11a was isolated
in 81% yield. The direct reduction product 10a was isolated in
only 6% yield. When the 95/5 anti/syn mixture of 9a was reacted
under the same conditions, similar results were obtained.
Cyclized 11a was isolated in 76% yield, and directly reduced
10a was collected in 10% yield.
The chromatographic and spectroscopic behavior of syn- and
anti-9b allowed us to characterize atropisomeric compounds 9a
as anti or syn by analogy. During column chromatography on
silica gel, anti-9b was less polar than syn-9b. Furthermore, 1H
NMR spectroscopy showed that the alkenyl and allylic (R to
N) protons of anti-9b were shifted downfield relative to those
of syn-9b. After separation of the atropisomers of 9a, the first
eluting atropisomer indeed showed a downfield shift for the three
diagnostic signals, so it is anti-9a. This pattern held true for all
N-cyclohexenyl-o-iodoanilides prepared.
Rotation Barriers of the N-Aryl Bonds of 9. Because anti-9a
and syn-9a are diastereomers, their interconversion by rotation
of the N-aryl bond occurs with two different barriers to rotation
if Keq * 1. To determine the barriers, a sample of syn-9a (28/
72 dr) in a 9:1 hexanes/EtOAc solution was heated at 36 °C,
and aliquots were periodically analyzed by HPLC. The final
composition of the equilibrium mixture of 9a was 65/35 anti/
syn. The first-order isomerization of syn-9a to anti-9a was
plotted against time (Figure 5).
The slope of the graph in Figure 5 corresponds to the rate
constant of rotation from syn-9a to anti-9a (Figure 6) at 36 °C;
k1 ) 5.2 × 10-4 s-1. From the equilibrium values, the rate
constant for the reverse rotation is k-1 ) 2.8 × 10-4 s-1. The
barriers to rotation for these two processes at 36 °C are ∆G‡rot
) 22.8 and 23.1 kcal/mol, respectively. These values are
The Bu3SnH-mediated radical reactions of the pure atropi-
somers of 9b gave strikingly different results. Reduction of syn-
9b with Bu3SnH (0.10 M) and Et3B gave cyclized 11b in 92%
yield along with only 4% of reduced 10b. This was an
inseparable mixture of atropisomers in a 2:1 ratio, although the
configurations of the atropisomers were not assigned. To confirm
the identity of 10b, catalytic hydrogenation over Pd/C produced
(13) Adler, T.; Bonjoch, J.; Clayden, J.; Font-Bard´ıa, M.; Pickworth, M.;
Solans, X.; Sole´, D.; Vallverdu´, L. Org. Biomol. Chem. 2005, 3, 3173–
3183.
9
J. AM. CHEM. SOC. VOL. 133, NO. 1, 2011 117

A R T I C L E S
Guthrie et al.
Scheme 2. Tin Hydride Reductions of syn- and anti-9a,b
Figure 8. Regioselectivity in cyclizations of 13a,b depends on the
o-substituent.
TS model 5 (Figure 2). Rapid cyclization and subsequent
reaction with tin hydride gives 11b. The failure of anti-12b to
cyclize in competition with reduction (even at low tin hydride
concentration) is surprising. Its N-Ar bond cannot rotate, so it
cannot access the transition state of syn-12b. Transition states
for cyclization are apparently inaccessible by rotation of the
N-cyclohexenyl bond. These transition states do not resemble
5 in geometry and must be rather high in energy.
By analogy to the results with 12b (R ) Me), radical syn-
12a, which lacks the ortho-methyl group (R ) H), is prone to
cyclize, while anti-12a is not. Because anti-9a and syn-9a give
similar results at high tin hydride concentration, the intercon-
version of anti/syn-12a must be rapid relative to the onward
reaction. Assuming that anti-12a does not cyclize directly, at
least 75% of anti-12a must have undergone rotation to syn-12a
to account for the total yield of 11a. This implies that the rotation
process interconverting anti-12a and syn-12a is at least three
times faster than the competing reduction by Bu3SnH.
an N-cyclohexylanilide with a simpler spectrum (see Supporting
Information). In contrast, tin hydride reduction of anti-9b gave
10b in 71% yield, again as a 2:1 mixture of atropisomers.
Decreasing the Bu3SnH concentration 100-fold gave a crude
product that was not as clean. Analysis of the 1H NMR spectrum
of this product showed significant amounts of 10b, but no
cyclized 11b. A reaction in refluxing benzene, initiated by
AIBN, gave similar results to this room-temperature experiment.
In short, we could not detect formation of 11b from anti-9b at
any concentration or temperature studied.
Figure 7 shows an analysis of the results of cyclizations of
9a,b. ortho-Methyl substrates syn/anti-9b give different prod-
ucts, so their derived radicals syn/anti-12b do not interconvert
by N-Ar bond rotation. When radical syn-12b is formed (Figure
7), the preferred conformation of the cyclohexenyl group holds
the alkene close to the radical in a geometry similar to that of
The rate constant kH for reduction of the phenyl radical is
5.2 × 108 M-1 s-1 at room temperature.14 In the experiments
shown in Scheme 2, the mean tributyltin hydride concentration15
[Bu3SnH]c ) 0.55 M, so the rate of reduction of anti-12a is
about 2.9 × 108 s-1. We can therefore identify a lower limit
for the rotational rate constant of anti-12a to syn-12a of kr )
8.7 × 108 s-1
.
Radical Reactions of Diastereomeric Atropisomers with
Two Acceptors. Independent observations by Storey and Jones16
and our group5,17 have illuminated an unusual regioselectivity
effect in the cyclizations of compounds with two radical
acceptors such as 13a and 13b (Figure 8). When no second
ortho substituent is present, as in 13a (R ) H), cyclization
occurs exclusively onto the acryloyl group to give 14a with no
trace of 15a. When an ortho-methyl group is present, however,
13b (R ) Me) cyclizes preferentially onto the N-allyl group to
give 15b with no 15a. This difference was attributed to steric
hindrance between the second o-substituent and the N-allyl
group, which slows the cyclization onto the crotonyl group by
disfavoring twisting of the N-Ar bond.5
To address the question of how a stereocenter next to nitrogen
influences this interesting regiodivergent behavior, we prepared
substrates 16a and 16b with an additional alkene acceptor at
the amide carbonyl (Scheme 3). While both atropisomers of
16a (made from 8a) were visible by TLC analysis, N-aryl bond
rotation was too rapid for separation by flash chromatography,
and 16a was isolated as a 64/36 anti/syn mixture. Atropisomers
(14) Johnston, L. J.; Lusztyk, J.; Wayner, D. D. M.; Abeywickreyma, A. N.;
Beckwith, A. L. J.; Scaiano, J. C.; Ingold, C. K. J. Am. Chem. Soc.
1985, 107, 4594–4596.
(15) Newcomb, M. Tetrahedron 1993, 49, 1151–1176.
(16) Jones, K.; Storey, J. M. D. J. Chem. Soc., Chem. Commun. 1992, 1766–
1767.
Figure 7. Competitive reactions of atropisomeric radicals 12a,b.
(17) Lapierre, A. J. B. Ph.D. Dissertation, University of Pittsburgh, 2005.
9
118 J. AM. CHEM. SOC. VOL. 133, NO. 1, 2011

Cyclizations of Haloanilide Atropisomers
A R T I C L E S
Scheme 3. Radical Reactions of Substrates 16a,b with Two
Acceptors
Figure 9. Pathways for reactions of radicals derived from 16a.
of ortho-methyl anilide 16b were separable by column chro-
matography; anti-16b was isolated in 38% yield, and syn-16b
was isolated in 29% yield.
These substrates were then reacted under radical conditions
as shown in Scheme 3. When the 64/36 anti/syn mixture of
16a was treated with Bu3SnH at room temperature, a mixture
of products was formed. The first eluting N-cyclohexenylin-
dolone, 17a, was obtained in 46% yield as an inseparable 5:1
mixture of diastereomers. Tricycle 18a was also collected in
53% yield.
Substrate syn-16b was reacted with Bu3SnH (0.01 M) and
Et3B at room temperature in benzene to give 18b in 81% yield
as the sole product. There was no evidence for direct reduction
(product not shown) or for formation of 17b. The reaction of
anti-16b under similar conditions gave a complex mixture of
products in poor mass balance.
The results in the 16b (R ) Me) series are consistent with
the observations of 13b in Figure 8. Specifically, syn-16b has
both the ortho-methyl group (favors cyclization to the N-allyl
side) and the configuration at the stereocenter that is prone to
cyclization. Because of these matched effects, a high yield of
18b is obtained. But anti-16b has mismatched effects. The
ortho-methyl group favors cyclization of the N-cyclohexenyl
side, while the configuration of the stereocenter disfavors it.
Indeed, neither cyclized product was observed in the crude
reaction mixture.
Pathways for formation of 17a and 18a from 16a are shown
in Figure 9. Radical anti-19a, formed directly from anti-16a,
should cyclize preferentially onto the crotonyl group to give
cyclized radical 20aR, precursor of 17aR. This regioselectivity
follows from the result with 13a (Figure 8) and from the
configuration (anti), which is mismatched for N-cyclohexenyl
cyclization. The stereoselectivity follows from the usual model1,2
(and has not been confirmed). Radical syn-19a, produced by
iodine atom abstraction from syn-16a, has two cyclization
pathways available. Cyclization onto the crotonyl group (as with
13a) gives radical 20aꢀ, leading to product 17aꢀ. Alternatively,
the syn-configuration is matched for cyclization onto the
Figure 10. Room-temperature Heck reaction of an axially chiral substrate.
cyclohexenyl group to produce radical 21a, which leads to 18a.
Aryl radicals anti-19a and syn-19a may also interconvert by
an N-aryl bond rotation, as shown by the prior results (Figure
7).
If the N-aryl rotation processes of anti-19a and syn-19a (kr,
kr′) are significantly slower than cyclization processes (kc1, kc2,
kc3), then the ratio of products 17aR/(17aꢀ + 18a) should reflect
the starting anti/syn ratio of 64/36. However, this is not the
case. For example, the yield of 18a (53%) significantly exceeds
the percentage of its precursor syn-16a (36%). Therefore, anti-
19a and syn-19a must equilibrate to some extent at room
temperature before cyclizing.
Heck Reactions of Diastereomeric Atropisomers. We have
previously investigated chirality transfer in room-temperature
Heck reactions of axially chiral o-iodoanilides (Figure 10).5 In
these reactions, the axis of chirality of the substrates dictated
the stereochemistry of the product. For example, cyclization of
(M)-22 gives (R)-23 with about 86% chirality transfer.
Because substrates 9b have restricted rotation about both the
N-aryl and N-cyclohexenyl bonds, we surmised on the basis of
the radical results that one atropisomer would cyclize under
Heck conditions, while the other would not. In the event, this
prediction was realized, but in the reverse! We subjected both
diastereomers of 9b to the room-temperature Heck conditions
of Hartwig (Scheme 4).18 When syn-9b was treated with
catalytic Pd2(dba)3 and P(t-Bu)3 ligand with Et3N base, no
cyclization to 24b occurred. Conducting a similar reaction of
syn-9b in the presence of excess tert-butyl acrylate (to trap aryl-
Pd intermediates) provided bimolecular Heck adduct syn-25b
(18) Stambuli, J. P.; Stauffer, S. R.; Shaughnessy, K. H.; Hartwig, J. F.
J. Am. Chem. Soc. 2001, 123, 2677–2678.
9
J. AM. CHEM. SOC. VOL. 133, NO. 1, 2011 119

A R T I C L E S
Guthrie et al.
Scheme 4. Room-Temperature Heck Reactions of 9ba
Figure 11. Proposed pathway for Heck cyclization of anti-9b.
Radical and Heck Reactions of N-Cyclohexenyl Allylic
Acetates. In the final step of the Heck reaction, the ꢀ-elimination
of a heteroatomic leaving group (for example, -Cl, -OAc,
-OCH3) typically proceeds more quickly than competing
ꢀ-hydride elimination when both pathways are available.21
Recently, Lautens and co-workers developed a procedure for
the efficient Heck coupling of aryl iodides and allylic acetates.22
The final step of this reaction presumably involves a chemose-
lective ꢀ-acetoxy elimination. Very little isomerization of the
double bond was seen in most products. We wondered whether
judicious placement of an acetoxy group in such substrates could
avoid a palladium hydride intermediate, thereby circumventing
isomerization.
a Conditions: 10 mol % Pd2(dba)3 · CHCl3 40 mol % P(t-Bu)3 · HBF4,
Et3N, DMF, rt, with or without trap.
in 69% yield. Notice that the axial chirality is retained in this
bimolecular process.9 The room-temperature Heck reaction of
anti-9b effected intramolecular cyclization, producing 74% of
24b as an inseparable mixture of alkene isomers in a 55:45 ratio.
Performing the Heck reaction of anti-9b in the presence of tert-
butyl acrylate formed trapped adduct anti-25b in 31% yield
along with 24b in 68% yield (25/75 mixture of alkene isomers).
Again, the bimolecular product 25b was found with retention
of axial chirality.
In contrast to the radical reactions of 9b, in which only the
syn atropisomer is prone to cyclize, only anti-9b is prone to
cyclize under room-temperature Heck conditions. We suggest
that this dramatic difference originates from fundamental
differences in the radical and Heck reaction transition-state
geometries. In the radical reaction, only two atoms need to
approach: the radical center and the alkene carbon being
attacked. The usual trajectory for approach (∼109° attack
angle)19 is readily accommodated from the syn radical.
However, in the insertion step of the Heck reaction, four
atoms come together: both atoms of the CAr-Pd bond approach
both atoms of the alkene.11 This approach is not easily
accommodated from the syn precursor because the Pd atom is
placed under the ring rather than near the alkene.20
A proposed pathway for the Heck cyclization of anti-9b is
shown in Figure 11. Oxidative addition of the palladium catalyst
into the carbon-iodine bond of anti-9b produces arylpalladium
intermediate 26. N-aryl rotation, concomitant with N-cyclohe-
xenyl bond rotation, allows for coordination between the
palladium and the alkene on the concave face of the cyclohexene
ring, as in 27. All four key participating atoms are well
positioned, and migratory insertion from this face results in cis-
fused intermediate 28. Only one hydrogen is available for syn
ꢀ-hydride elimination, so 24b ∆3,4 is presumably the primary
product. Subsequent palladium hydride reinsertion and elimina-
tion can migrate the double bond to the ∆4,5 position.
With this in mind, a series of disubstituted cyclohexenes
containing a cis- or a trans-acetoxy group relative to the anilide
group were prepared as described in the Supporting Information.
The precursor structures and the products from radical and Heck
cyclizations are summarized in Table 1. In the ortho-methyl
series (27c,t), separation was possible and the atropisomers were
stable. The syn and anti isomers of 26c and 26t (ortho-hydrogen
substituent) could not be separated. However, upon standing at
room temperature, a sample of neat oil 26c (60/40, anti/syn)
spontaneously formed crystals possessing the anti configuration
(Figure 12). The selective crystallization of one diastereomeric
atropisomer, a dynamic thermodynamic resolution, has been
performed previously to produce diastereoenriched anilides from
an equilibrium mixture.1b The torsion angle between the amide
and aromatic planes of anti-26c is 83.8°, essentially perpen-
dicular, as expected. The cyclohexene ring exists in a half-chair
conformation with the bulkier amide group pseudoequatorial
1
and the acetate pseudoaxial. H NMR analysis by dissolution
of a crystal at low temperature (see Supporting Information)
confirmed that anti-26c is the major atropisomer in CDCl3
solution after equilibrium is reached.
Under tin hydride conditions (Table 1), all compounds
containing an allylic acetate group behaved similarly to unsub-
stituted analogue 9a or 9b. trans-Disubstituted 26t (60/40 anti/
syn) provided tricycle 28t in 89% yield. Similarly, cis isomer
(21) Zhu, G.; Lu, X. Organometallics 1995, 14, 4899–4904.
(22) (a) Mariampillai, B.; Herse, C.; Lautens, M. Org. Lett. 2005, 7, 4745–
4747. (b) Lautens, M.; Tayama, E.; Herse, C. J. Am. Chem. Soc. 2005,
127, 72–73. (c) Similar selectivities in ꢀ-eliminations have been
observed by Dr Christelle Herse of the University of Toronto as cited
in the Thesis of Dr. Brian Mariampillai, University of Toronto, pp
158-159, 2009. We thank Prof. Mark Lautens for sharing this
information.
(19) Beckwith, A. J. J.; Schiesser, C. H. Tetrahedron 1985, 41, 3925–
3941.
(20) (a) Mori, M.; Nakanishi, M.; Kajishima, D.; Sato, Y. Org. Lett. 2001,
3, 1913–1916. (b) Mori, M.; Nakanishi, M.; Kajishima, D.; Sato, Y.
J. Am. Chem. Soc. 2003, 125, 9801–9807.
9
120 J. AM. CHEM. SOC. VOL. 133, NO. 1, 2011

Cyclizations of Haloanilide Atropisomers
A R T I C L E S
Table 1. Radical and Heck Cyclizations of Allylic Acetates
Figure 12. X-ray crystal structure of anti-26c.
Figure 13. Suggested mechanism of Heck reaction with chemoselective
ꢀ-elimination.
reaction of anti-27c afforded directly reduced 30c (96% yield)
as a 67/33 mixture of atropisomers. In short, the substrates
lacking the o-Me group (26c,t) gave cyclized products in similar
yields. When the o-Me group was present (27c,t), syn-isomers
cyclized and anti-isomers did not.
Four of the substrates were then subjected to Heck conditions
according to the procedure developed by Lautens.22 In a typical
experiment, 26t (60/40 anti/syn) was combined in wet aceto-
nitrile with BuNMe2 and catalytic amounts of Pd2(dba)3 ·CHCl3
and P(o-tol)3. The mixture was placed in a sealed tube and
heated at 80 °C until Pd-black precipitated. After purification,
tricyclic vinyl acetate 31 was isolated in 47% yield. Under the
same reaction conditions, anti-27t produced 32 in 73% yield,
again as the vinyl acetate. The behavior of cis-disubstituted
cyclohexene systems under these conditions was studied as well.
Under Heck conditions, 26c (60/40 anti/syn) gave alkene 33 in
82% yield as the sole product. Similarly, anti-27c produced
alkene 24b in 73% yield. On the basis of the results with syn-
9b, the syn atropisomers of 27t and 27c were expected to be
unreactive toward Heck cyclizations, so these were not examined.
A mechanistic rationale for these results is shown in Figure
13 with substrates anti-27c and anti-27t as examples. Oxidative
addition of the Pd0 catalyst occurs to give aryl-PdII intermediate
34c or 34t. Concomitant N-aryl and N-cyclohexenyl rotation
followed by insertion gives alkylpalladium intermediate 35c or
35t. These undergo ꢀ-elimination with the syn substituent R2.
In the reaction of anti-27c, R2 in palladium intermediate 35c is
an acetate group, so IPdIIOAc is eliminated to give alkene 24b.
When anti-27t is the substrate, R2 in 35t is H, and ꢀ-hydride
elimination occurs to give vinyl acetate 32.
a 1.5 equiv of Bu3SnH, 1.0 equiv of Et3B, [Bu3SnH] ) 0.01 M,
i
PhH, rt. b 5 mol % Pd2 (dba)3, CHCl3, 22 mol % P(o-tol)3 BuNMe2,
10:1 MeCN/H2O sealed tube, 80 °C. c [Bu3SnH] ) 0.2 M. d Reaction
i
not attempted.
26c (60/40 anti/syn) cyclized to 28c in 92% yield. ortho-Methyl
substrates syn-27t and syn-27c also underwent cyclization,
producing 29t (77% yield) and 29c (71% yield), respectively.
When anti-27t was treated with Bu3SnH at an initial hydride
concentration of 0.01 M, multiple products were formed.
However, increasing the [Bu3SnH]i to 0.20 M produced 30t in
96% yield in a 67/33 ratio of atropisomers. Similarly, the radical
The ability of 35t to undergo ꢀ-acetoxy elimination when
R2 ) OAc is unusual. Previous studies have shown that
9
J. AM. CHEM. SOC. VOL. 133, NO. 1, 2011 121

A R T I C L E S
Guthrie et al.
ꢀ-elimination of heteroatom substituents during the Heck
reaction proceeds with the leaving group anti-periplanar to the
palladium.18,19,23 However, those studies were performed under
acidic conditions, with Pd(OAc)2 catalyst in AcOH solution.
Both anti- and syn-eliminations of aryloxy groups have been
observed during Heck reactions in the presence of trialkylamine
base.22c,24 The propensity for this system to undergo syn-acetoxy
elimination is indicative of a change in preferred geometry for
ꢀ-elimination under basic conditions.
temperatures and were readily resolved. In radical cyclizations,
the syn-diastereomers were prone to cyclize while the anti
isomers were not (bimolecular reduction occurred instead).
Strikingly, Heck reactions gave the opposite result; anti isomers
were prone to cyclize and syn isomers preferred bimolecular
reactions. Heck reactions of allylic acetates occur with ꢀ-hydride
elimination when acetate is trans to palladium and with
ꢀ-acetoxy elimination when acetate is cis. This is surprising
because prior studies have suggested that a trans relationship
of palladium and acetoxy was favored for acetate elimination.
Conclusions
The results of radical cyclizations and Heck reactions of N-(3-
cyclohexenyl)-ortho-iodoacetanilides depend critically on the
configurations of the chiral axis and the stereocenter. In
substrates 9a, 13a, 16a, 26t, and 26c without an additional ortho-
methyl group, the precursor diastereomers interconvert slowly
at ambient temperatures. Radical and Heck cyclizations of
enriched mixtures of diastereomers occurred in high yields and
provided similar results, suggesting that interconversion of the
radical or organometallic intermediates also occurs.
These results show the dramatic effects that axial chirality
can have on the outcome of radical cyclizations and Heck
reactions. The effects can be either reinforced or negated by
the additional stereocenter present in the substrates, thereby
providing insight into geometries of intermediates and transition
states.
Acknowledgment. We thank the U.S. National Science Foun-
dations (NSF) for funding.
Diastereomers of substrates 9b, 13b, 16b, 27t, and 27c with
an additional methyl group did not interconvert at ambient
Supporting Information Available: Experimental procedures
and compound characterization data. This material is available
free of charge via the Internet at http://pubs.acs.org.
(23) Cheng, J. C.-Y.; Daves, J.; Doyle, G. Organometallics 1986, 5, 1753–
1755.
(24) Sinou, D.; Bedjeguelal, K. Eur. J. Org. Chem. 2000, 4071–4077.
JA108795X
9
122 J. AM. CHEM. SOC. VOL. 133, NO. 1, 2011