Mechanistic investigation of enediyne-connected amino ester rearrangement. Theoretical rationale for the exclusive preference for 1,6- or 1,5-hydrogen atom transfer depending on the substrate. A potential route to chiral naphthoazepines

Article abstract of DOI:10.1021/jo202580y

Memory of chirality (MOC) and deuterium-labeling studies were used to demonstrate that the cascade rearrangement of enediyne-connected amino esters 1a and 1b evolved through exclusive 1,5- or 1,6-hydrogen atom transfer, subsequent to 1,3-proton shift and Saito-Myers cyclization, depending on the structure of the starting material. These results were independently confirmed by DFT theoretical calculations performed on model monoradicals. These calculations clearly demonstrate that in the alanine series, 1,5-hydrogen shift is kinetically favored over 1,6-hydrogen shift because of its greater exergonicity. In the valine series, the bulk of the substituent at the nitrogen atom has a major influence on the fate of the reaction. N-Tosylation increases the barrier to 1,5-hydrogen shift to the benefit of 1,6-hydrogen shift. The ready availability of 1,6-hydrogen atom transfer was explored as a potential route for the enantioselective synthesis of naphthoazepines.

Full text of DOI:10.1021/jo202580y

Article
pubs.acs.org/joc
Mechanistic Investigation of Enediyne-Connected Amino Ester
Rearrangement. Theoretical Rationale for the Exclusive Preference
for 1,6- or 1,5-Hydrogen Atom Transfer Depending on the Substrate.
A Potential Route to Chiral Naphthoazepines
Damien Campolo,† Anouk Gaudel-Siri,*,† Shovan Mondal,† Didier Siri,† Eric Besson,†
Nicolas Vanthuyne,‡ Malek Nechab,*,† and Michele P. Bertrand*,†
̀
†Institut de Chimie Radicalaire, CNRS UMR 7273, Equipes CMO and CTM, and ‡ISM2, CNRS UMR 7313, Equipe Chirosciences,
́
Aix-Marseille Universite, 13397 Cedex 20, Marseille, France
S
* Supporting Information
ABSTRACT: Memory of chirality (MOC) and deuterium-
labeling studies were used to demonstrate that the cascade
rearrangement of enediyne-connected amino esters 1a and 1b
evolved through exclusive 1,5- or 1,6-hydrogen atom transfer,
subsequent to 1,3-proton shift and Saito−Myers cyclization,
depending on the structure of the starting material. These results
were independently confirmed by DFT theoretical calculations
performed on model monoradicals. These calculations clearly
demonstrate that in the alanine series, 1,5-hydrogen shift is
kinetically favored over 1,6-hydrogen shift because of its greater
exergonicity. In the valine series, the bulk of the substituent at the nitrogen atom has a major influence on the fate of the reaction.
N-Tosylation increases the barrier to 1,5-hydrogen shift to the benefit of 1,6-hydrogen shift. The ready availability of 1,6-
hydrogen atom transfer was explored as a potential route for the enantioselective synthesis of naphthoazepines.
that was nearly totally controlled by the MOC phenomenon.
We discuss herein the comparative behavior of structurally
related enediynes derived from valine and alanine, respectively.
INTRODUCTION
■
We have recently reported the enantioselective cascade
rearrangement of enediyne 1a which evolved through the
mechanism depicted in Scheme 1.1,2 The reaction proceeded
with memory of chirality3 and retention of the configuration of
the starting material owing to the generation of intermediate C
in a chiral conformation. The racemization half-life time of the
captodative center in diradical C is long compared to the time
scale of the recombination step (d). By opposition to
enantiopure enediynes in which the stereogenic center was
included in a 5-membered ring,1,2 substrate 1a led to an
undesired olefinic product (3a) that resulted from a
disproportionation step (e),4 which competed with the
biradical recombination leading to tetrahydrobenzoisoquino-
lines 2a. These reactions that were initially performed using
alumina as the base were shown to proceed very readily with a
recoverable nanocatalyst, i.e., mesoporous silica grafted with a
tertiary amino group (GA-SBA-15).5 Overall yields were
significantly improved, and the reaction could be carried out
indifferently in benzene or in acetonitrile. The main drawback
in this methodology was the lack of diastereoselectivity which
resulted in the formation of two enantioenriched diastereomers
with two contiguous stereocenters. This drawback was very
recently overcome by designing a one-pot strategy.6 Tandem
RESULTS AND DISCUSSION
■
Synthetic Results. When enediyne (S)-1b was submitted
to similar experimental conditions (in the presence of 0.2 equiv
of GA-SBA-15), (S)-4b was isolated as the only product at the
expense of any tricyclic product issued from a recombination
process (eq 1). Intrigued by the formation of olefin 4b, while
3b was expected on the ground of the common prevalence of
1,5-hydrogen shifts, we have checked the influence of other
experimental conditions on the fate of (rac)-1b. In the presence
of an excess of base, 3b could also be formed, but indirectly. It
resulted from 4b through the slow migration of the double
bond (eq 2).
It became obvious that 4b was the kinetic product and,
therefore, that it resulted from 1,6-hydrogen atom transfer
rather than from 1,5-transfer of the hydrogen atom in
captodative position.
The kinetic preference for 1,6-hydrogen shift was unambig-
uously demonstrated by measuring the optical purity of 4b
generated as the unique reaction product from optically pure
1b (ee >99%) (eq 1); the reaction proceeded without loss of
́
Crabbe homologation (leading to monosubstituted allenes)/
Myers−Saito cyclization-induced rearrangement was shown to
Received: December 21, 2011
Published: February 16, 2012
lead to heterocyclic compounds bearing only one stereocenter
© 2012 American Chemical Society
2773
dx.doi.org/10.1021/jo202580y | J. Org. Chem. 2012, 77, 2773−2783

The Journal of Organic Chemistry
Article
Scheme 1. General Mechanism
originating from the integration of aromatic signals are
probably responsible for the gap between the effective and
the theoretical rate of deuteration.7 Conversely, as shown in eq
4, the unique product resulting from (rac)-d1-1b, i.e., (rac)-d1-
4b, was exclusively deuterated at the captodative position,
which confirmed that in this case only 1,6-hydrogen shift had
occurred.
The rearrangement of (rac)-d3-1a, prepared from the
corresponding commercially available (rac)-d3-alanine (99.5%
D), is reported in eq 5. It corroborated the results reported in
eq 3. Deuterium was recovered exclusively in the CD3 group in
the tricyclic diastereomeric products ((rac)-d3-2a). In addition,
in d3-3a deuterium was scrambled between the terminal
olefinic carbon and the methylene group of the benzylic
sulfone. In all likelihood, disproportionation is an intra-
molecular pathway.
chirality. The enantiopurity of the starting material was entirely
recovered in the product.
The formation of 3a through 1,5-hydrogen shift to 1-
naphthyl radical B (Scheme 1) was therefore questioned. The
pathways leading to 3a and 3b were confirmed by deuterium
labeling of substrates 1a and 1b as reported in eqs 3−5,
respectively.
Racemic 1a and 1b, monodeuterated at the captodative
position, were prepared from monodeuterated racemic N-
tosylalanine and N-tosylvaline methyl esters, respectively.
Deuteration of the precursors was accomplished upon treat-
ment with K2CO3 in MeOD (99% D) at 70 °C (see the
Supporting Information).
As shown in eq 3, both monodeuterated products 2a and 3a
resulting from the rearrangement of (rac)-d1-1a were labeled
exclusively at the aromatic ring, as expected from exclusive 1,5-
hydrogen shift. The rate of deuteration with respect to the
starting material was determined from the relative area of the
residual singlet signal of the unlabeled aromatic proton. Errors
An interesting feature was noted. The relative ratio of
isolated products (3S,4S)-2a:(3S,4R)-2a:3a that was initially
42:33:21 (Scheme 1) or 40:35:21 (eq 3)5 became 45:38:10 in
the experiment performed on (rac)-d3-1a (eq 5). A primary
kinetic isotope effect slowed down disproportionation and
2774
dx.doi.org/10.1021/jo202580y | J. Org. Chem. 2012, 77, 2773−2783

The Journal of Organic Chemistry
Article
Chart 1
increased the global yield of tricyclic products from 75 to 83%
at the expense of d3-3a.8 It is to be noted that apparently
disproportionation affected similarly the pathways leading to
both diastereomers (the ratio of which remained constant
within the limits of experimental errors).
Examples of 1,5-hydrogen abstraction are widespread in the
literature. Applications of Barton and Hoffmann−Loeffler
reactions, based on O- and N-centered radicals, respectively,
are well documented.9 Similarly, many synthetic strategies are
based on radical translocation from vinyl or aryl radicals which
have become very popular and powerful synthetic tools.9c,d,10
Comparatively, examples of 1,6-hydrogen shift are less
numerous.11
The only data found in the literature concerning 1-naphthyl
radicals have been reported by De Mesmaeker.14,19 The latter
also mentioned competitive 1,5-, 1,6-, and 1,7-H transfers, but
even these results cannot be correlated to ours, since in these
reactions, 1-naphthyl radicals abstracted hydrogen atoms from
an aliphatic chain located in position 8 and not in position 2.
The above disclosed rearrangement of 1b was rather
surprising since 1,6-hydrogen shift is kinetically favored to the
exclusion of any 1,5-hydrogen shift, even though the latter
should be more, or at least, as exothermic as the previous one.
DFT calculations were undertaken on model monoradicals to
confirm or infirm these observations.
Computational Section. First the translocation reaction
was investigated starting from simple models, i.e., radicals 5−8
bearing a methyl group as substituent at nitrogen, for the sake
of reducing calculation time (Chart 1, eqs 6−13). Then, based
on these data, new explorations of the potential energy surface
were achieved to calculate the transition state geometries for
the rearrangements of N-tosylated monoradicals 9−12 (Chart
1, eqs 14−21). The latter are closely related to the intermediate
biradicals involved in the enediynes cascade transformation.
Computational Details. DFT calculations were performed
to determine the profiles of the competitive reactions using the
Gaussian 09 program.20 The geometries were fully optimized at
the UB3LYP/6-31G(d) level of theory. Vibrational frequencies
were calculated at the UB3LYP/6-31G(d) level to determine
the nature of the located stationary points. Zero-point energies
and thermodynamic data were calculated using the specified
scaling factor (0.9603).21 All transition-state geometries were
confirmed by intrinsic reaction coordinates (IRC) calculations.
In order to enhance the accuracy of the calculations, single-
point energies were calculated at the UB3LYP/6-311++G-
(3df,3pd) level of theory. The spin contamination was low for
all radical species (<S2> values ranged between 0.754 and
0.760).
Peculiar attention was given to literature data where 1,5- and
1,6-hydrogen shifts could be competitive pathways.12−14 As a
rule of thumb, 1,5-hydrogen shift is generally favored over 1,6-
hydrogen shift. According to theoretical calculations based on
butoxy and pentoxy radical, the preference for 1,5-hydrogen
shift (six-membered ring transition state) over 1−6 hydrogen
shift (seven-membered ring transition sate), results from the
more favorable activation entropy of the former. Activation
enthalpy of the rearrangement of pentoxy radical was found
only 0.7 kcal per mole lower than that of butoxy radical.15
Exceptions to the rule occur when 1,6-hydrogen shift is more
exothermic than 1,5-migration, however, in most cases,
mixtures of products resulting from the competition between
these two processes were formed.
The unique example of recent theoretical calculations using
B3LYP DFT functional, and concerning independent 1,5- and
1,6-hydrogen abstractions by aryl radical, deals with the
methodology developed by Curran13 from o-iodoanilide
derivatives.16,17
According to Schiesser data, the rate constant for 1,5-
hydrogen transfer from a captodative position in such
structures ranges between 1.0 and 4.9 × 107 s−1.18 In the
very same paper, the authors have observed that 1,6- or 1,7-
radical translocations competed with 1,5- hydrogen shift
depending on the structure. However, these structures are
completely different from ours, particularly in regard of the
planarity of the anilide moiety and extrapolations to the fate of
radicals of type B (Scheme 1) might be hazardous.
Some model reactions (Chart 1, eqs 6−9) were also
calculated at the G3(MP2)-RAD level of theory to validate
the selection of UB3LYP/6-311++G(3df,3pd)//UB3LYP/6-
31G(d) method. B3LYP methods are known to underestimate
energies values, however the energy gaps are consistent with
those obtained from G3(MP2)-RAD calculations which are
2775
dx.doi.org/10.1021/jo202580y | J. Org. Chem. 2012, 77, 2773−2783

The Journal of Organic Chemistry
Article
known to be highly reliable for radical species (Figure 1, Table
1).22
predisposed to 1,6-H shift, in the case of radicals 9, 10 and 12.
No energy difference is observed in the case of radical 11 (see
ΔGconf in Table 1).
The predicted ratio of products Xa and Xb (X = 5−12),
issued from the competitive 1,5- and 1,6-H shift reactions,
respectively, are given in Table 1 (they were determined from
the calculated ΔΔG⧧ values). From the calculations performed
on the series of N-methyl model radicals derived from alanine
(radicals 5 and 7), 1,5-H transfer is obviously much faster than
1,6-H transfer (see ΔG⧧ in Table 1). It must be noted that the
least stable conformer (ready to undergo 1,6-H shift from the
methyl group) of any of these starting radicals should account
for less than 1% at 298 K according to Boltzmann partition.
According to theoretical calculations, 1,5-H transfer remains
favored over 1,6-H transfer, even in the valine series (radicals 6,
8), but the barrier for 1,6-hydrogen shift is decreased as
compared to the alanine series (the incidence of the reaction
exergonicity is discussed in the following). Both aryl and 1-
naphthyl radicals behave similarly.
In this first series of model radicals, i.e., N-methyl derivatives,
in all cases, 1,6-H shift is enthalpically disfavored and the
activation enthalpic governs the fate of the radical, although the
activation entropy is more favorable for the seven-membered
ring transition state (ΔS⧧ values show that the six-membered
ring transition state geometry for 1,5-H transfer is more
constrained in most cases but radical 8).
Figure 1. Linear correlation between free energy values calculated at
the UB3LYP/6-311++G(3df,3pd)//UB3LYP/6-31G(d) and the G3-
(MP2)-RAD levels.
Computational Data. The activation free energy (ΔG⧧),
the activation enthalpy (ΔH⧧), and entropy (ΔS⧧), together
with the reaction exergonicity (ΔG0) are reported in Table 1
for each reaction (Chart 1, eqs 6−21). It must be noticed that
the intrinsic reaction coordinate (IRC) calculations led, in most
cases, to slightly different conformations of the reactive starting
radical depending on the nature of the transition state (six- or
seven-membered ring). The free energy differences ΔGconf
between these conformers of the starting radical are given in
Table 1. The conformations obtained from the seven-
membered transition states are the least stable ones for N-
methylated model radicals and the situation is reversed for N-
tosylated radicals. IRC calculations show that the conformer
predisposed to 1,5-H shift is less stable than the conformer
It must be noted that, not surprisingly, the homolysis of the
captodative C−H bond corresponds to the most exergonic
reactions (see ΔΔG0 in Table 1). Due to the difference
between the bond dissociation energy (BDE) of the C−H bond
in the methyl group and that of the tertiary C−H bond in the
isopropyl group the enthalpic factors should contribute to a
lesser extend to the activation barrier in the case of valine
derivatives.
The captodative character of radicals 9−12 may be
questioned,23 and the simple comparison of the calculated
free energy of reactions (ΔG0) involving 1,5-H migration, i.e.,
Table 1. ΔGconf, ΔG⧧, ΔΔG⧧, ΔH⧧, ΔG0, ΔΔG0, and ΔG⧧ in kcal/mol and ΔS⧧ in cal/mol/K, for 1,5- and 1,6- Translocations
i
of Radicals 5−12, Calculated at 298 K at the UB3LYP/6-311++G(3df,3pd)//UB3LYP/6-31G(d) Level of Theory
b
a c
ΔG⧧
d
a c
−
e
reaction
ΔGconf
−
ΔΔG⧧ (Xa/Xb)
ΔH⧧
ΔS⧧
ΔG0
ΔΔG0
ΔG⧧
i
5→5a
0.0
4.4 (6.4)
10.9 (12.6)
4.4 (6.0)
6.7 (8.0)
3.6
−6.5 (5.8 × 104)
2.8
10.0
2.3
5.0
1.9
9.6
1.4
4.6
3.7
6.2
8.2
5.5
2.7
6.2
8.4
6.1
−5.5
−2.9
−6.9
−5.6
−5.8
−4.5
−9.4
−9.7
−5.3
−5.7
−2.6
+2.4
−3.6
−6.4
−7.9
−4.6
−34.0 (−29.4)
−7.1 (−7.6)
−33.0 (−27.5)
−16.5 (−12.5)
−34.4
−7.6
−31.6
−16.9
−28.1
−10.3
−22.2
−21.7
−26.9
−11.1
−20.9
−20.7
−26.9
−16.5
−26.8
−14.7
−17.8
−0.5
17.2
12.3
16.8
11.1
16.2
12.8
16.2
12.6
15.5
12.5
15.7
13.5
14.1
13.1
16.0
16.2
5→5b
3.2 (3.9)
0.0
6→6a
−2.3 (49)
6→6b
3.3 (3.8)
0.0
7→7a
−7.3 (2.2 × 105)
−3.3 (2.6 × 102)
−2.6 (81)
7→7b
2.7
10.9
8→8a
0.0
4.2
8→8b
2.0
7.5
9→9a
0.9
5.3
9→9b
0.0
7.9
10→10a
10→10b
11→11a
11→11b
12→12a
12→12b
3.5
9.0
4.2 (8.3 × 10−4
−4.3 (1.4 × 103)
3.3 (3.8 × 10−3
)
0.0
4.8
0.0
3.8
−15.8
−0.2
0.0
8.1
5.2
10.8
)
0.0
7.5
a
b
G3(MP2)-RAD values are given in parentheses. Free energy difference between the two most stable conformers of radicals 5−12 resulting from
c
d
IRC calculations. Free energies of the competitive 1,5- and 1,6-hydrogen atom transfers are given relative to the most stable conformer. ΔΔG⧧ =
⧧
e
ΔG⧧
− ΔG⧧(1,6‑H), Xa/Xb = e− ΔΔG /RT is the product ratio resulting from the competitive 1,5- and 1,6-hydrogen atom transfers. ΔΔG0 =
ΔG0(1,5‑H)
(1,5‑H) − ΔG0(1,6‑H)
.
2776
dx.doi.org/10.1021/jo202580y | J. Org. Chem. 2012, 77, 2773−2783

The Journal of Organic Chemistry
Article
5→5a/9→9a, 6→6a/10→10a, 7→7a/11→11a, 8→8a/12→
12a, shows that the stabilization energy of radicals bearing a N-
Ts group at nitrogen is lower than that of radicals bearing a N-
Me group by 5.9 to 10.9 kcal/mol. However, it is difficult to
estimate the relative contribution of electronic and steric
effects.24
Table 2. Calculated C−H Distances (in Å) in the Transition
States for 1,5- And 1,6-translocations of radicals 5−12 at the
UB3LYP/6-311++G(3df,3pd)//(UB3LYP/6-31G(d) Level
at 298 K
d(C−H)
d(CAr−H)
reaction d(C···H)
(substrate)
d(CAr···H)
product
L
In agreement with Hammond’s postulate,25 the activation
free energy is influenced by the reaction exergonicity. In
Marcus formalism,26 the activation free energy (ΔG⧧) is the
5→5a
5→5b
6→6a
6→6b
7→7a
7→7b
8→8a
8→8b
9→9a
9→9b
1.234
1.286
1.228
1.249
1.232
1.286
1.224
1.254
1.231
1.291
1.238
1.097
1.104
1.097
1.097
1.096
1.094
1.097
1.097
1.098
1.094
1.099
1.505
1.399
1.515
1.468
1.509
1.401
1.520
1.467
1.503
1.395
1.482
1.086
1.087
1.087
1.085
1.087
1.087
1.089
1.086
1.088
1.087
1.087
0.327
0.583
0.306
0.397
0.322
0.611
0.295
0.412
0.320
0.640
0.352
sum of an intrinsic activation free energy (ΔG⧧ ) and a
i
thermodynamic contribution including the reaction free
enthalpy (ΔG0) (Figure 2). The intrinsic barrier represents
10→
10a
10→
1.255
1.227
1.289
1.24
1.097
1.098
1.094
1.098
1.098
1.451
1.501
1.399
1.482
1.460
1.088
1.089
1.088
1.088
1.088
0.435
0.313
0.627
0.360
0.404
10b
11→
11a
11→
11b
12→
12a
12→
12b
1.247
Figure 2. Relationships between ΔG0, ΔG⧧, and ΔG⧧ in Marcus
i
In these cases, 1,6-H shift is disfavored not only enthalpically
but also entropically. The difference in activation entropy
between six- and seven-membered ring transition states varies
between 0.4−2.8 cal/mol/K, which corresponds to 0.12−0.83
kcal/mol at 298 K.
The situation is completely reversed in the valine series
(radicals 10 and 12), since the exergonicities of the two
processes become similar. According to calculations, for these
radicals 1,6-H shift should be favored and moreover be the only
pathway, which is attested by the experimental results obtained
from 1b. By opposition to radicals 9 and 11, in the cases of
radicals 10 and 12, 1,6-H shift is favored both enthalpically and
entropically.
The reaction profiles and geometries of transition states for
the competitive evolutions of radical 11 and 12 are given in
Figures 3 and 4, respectively. Figure 4 gives clear evidence that
the bulky substituent attached to nitrogen controls the
preferred conformation of radical 12 and contributes to
increase the activation free energy of 1,5-hydrogen shift.
The main differences in torsional interactions around C−N
and C−S single bonds are illustrated in Figure 5. The simple
examination of the main torsional angles accounts for the
higher energy of the conformer predisposed to 1,5-H transfer.
Experimental Evidence for the Influence of N-Tosyl
Group. In order to test the validity of the theoretical
prediction, enediyne ( )-13 was prepared from N-methylvaline
methyl ester. The difficulties in preparing strictly the N-methyl
analogue of 1b led us to use the strategy recently devised using
formalism.
pure transition states effects for thermoneutral process. The
thermodynamic contribution lowers the activation free energy
for exergonic processes, or conversely increases the activation
free energy for endergonic reactions.
The values of the intrinsic barriers (ΔG⧧ ) clearly indicate
i
that 1,6-hydrogen shift should be largely favored over 1,5-
hydrogen shift in nearly all cases but radicals 11 and 12, for
which a competition should be expected between the two
reactions. The exergonicity driving force is responsible for the
lowering of the activation free energy and for the exclusive
preference for 1,5-H shift predicted in the case of radicals 5−9.
Due to the higher exergonicity of the process, the transition
state of 1,5-H migration is earlier than that of 1,6-H shift, which
should alleviate steric crowding. The C−H distances in the
corresponding transition structures are given in Table 2, where
L (eq 22) is a measure of the precocity of the transition state.27
As expected in the alanine series, the transition state for 1,5-H
shift is much earlier than the transition state for 1,6-H shift; the
difference in L parameters is less pronounced in the valine
series.
⧧
d(C···H) − d(C···H)
rad
L =
⧧
d(C ···H) − d(C ···H)
Ar
Ar
product
(22)
Calculations suggest that the nature of the substituent at
nitrogen has a determining influence on the kinetics of the two
competitive reactions. In the case of N-tosyl derivatives the
situation becomes strongly dependent on the nature of the
amino ester moiety. In the alanine series (radicals 9 and 11),
whatever the structure of the radical (aryl or 1-naphthyl), and
́
one-pot Crabbe homologation/tandem enynyne−allene
( )-14 rearrangement.6 The experimental results are reported
in Scheme 2.
The one-pot multistep reaction led to a mixture of products
( )-15 (28%) and ( )-16 (25%), resulting from exclusive 1,5-
H shift but from both the captodative position and the methyl
group. Despite the reaction enthalpy is in favor of hydrogen
even though 1,6-H shift remains intrinsically favored (ΔG⧧ in
i
Table 1), 1,5-H shift is much faster owing to the high
exergonicity of this process (ΔG⧧ and ΔG0 in Table 1).
2777
dx.doi.org/10.1021/jo202580y | J. Org. Chem. 2012, 77, 2773−2783

The Journal of Organic Chemistry
Article
Figure 3. Competitive reaction pathways for the translocation of
radical 11 (activation free energies and exergonicities in kcal/mol).
Figure 5. Compared torsional interactions around the two C−N and
the C−S bonds in the two reactive conformers of radical 12 (on the
left, views of the conformer predisposed for 1,5-H shift; on the right,
views of the conformer predisposed to 1,6-H shift. Hydrogen atoms
were removed for the sake of clarity).
Scheme 2. Tandem Crabbe Homologation/Rearrangement
of ( )-13
Figure 4. Competitive reaction pathways for the translocation of
radical 12 (activation free energies and exergonicities in kcal/mol).
abstraction from the captodative position, hydrogen abstraction
from the methyl group competes. It is enhanced by the statistic
factor (3H/1H) and probably by steric effects. Side products
resulting from oxidative degradation were detected in the crude
mixture that could not be quantified nor identified. These data
confirmed the dramatic incidence of the methyl group on the
fate of diradical D the isolated products resulted from 1,5-H
shift in agreement with theoretical predictions made for radical
8 (Chart 1, eqs 12 and 13, and Table 1). The additional
incidence of the absence of substituent at the benzylic radical
center, contributes to an important slowing down of the
disproportionation pathway. No product resulting from
disproportionation was isolated or detected in the crude
mixture.6
Cascade Rearrangements Leading to Naphthoaze-
pines. The availability of tricyclic products resulting from 1,6-
hydrogen shift should be reinforced by favorable enthalpic
factors. Thus, in the absence of any competitive 1,5-H shift, the
enantioselective synthesis of seven-membered ring cyclic amino
esters bearing a quaternary stereogenic center was an obvious
application of the methodology.
Homopropargylic aminoester (S)-17 (85% ee) was prepared
from (S)-N-tosylalanine methyl ester. Submitted to the action
2778
dx.doi.org/10.1021/jo202580y | J. Org. Chem. 2012, 77, 2773−2783

The Journal of Organic Chemistry
Article
Scheme 3. Cyclization of (S)-17
Scheme 4. Rearrangement of 4
of GA-SBA-15, (S)-17 led to (1S,2S)-18 in 38% yield and 56%
ee (Scheme 3).28 The major product 19 resulted again from
disproportionation.
migration from the position α to oxygen in the oxazolidinone
ring to the benzylic radical center leading to 22 in 40 or 49%
yield. The reaction is enantioselective, oxazepin 21 was
obtained as a single diastereomer in 24% yield and 84% ee
from (R)-20 and its enantiomer was isolated in 27% yield and
77% ee from (S)-20.29
The high ee provides an indirect proof that the rotations
around the non benzylic C−C and C−N single bonds β and γ
are still hindered as compared to rotation around bond α in
diradical H. However, this investigation showed that the
increased flexibility in the side chain enables disproportionation
pathways that were not observed in experiments performed on
the previous set of homologous propargylic substrates,
undergoing the closure of six-membered rings in the last
elementary step of the cascade rearrangement.1 This is a severe
limitation to the yield in enantio-enriched naphtoazepines that
could be prepared according to this pathway.
In spite of the increased number of rotational degrees of
freedom in the side chain bearing the reactive captodative
center in biradical G and as this radical center (issued this time
from 1,6-hydrogen transfer) was formed in a chiral con-
formation, memory of chirality was still observed. This means
that recombination via rotation around the two single bonds
adjacent to the aromatic system, i.e., bonds α and δ, remained
the fastest pathway. Naphtho-azepine 18 was isolated as a single
isomer in 56% ee.29
In order to suppress any competitive disproportionation,
attempts to synthesize optically pure enediyne derived from
phenyl glycine failed, and only the racemic starting material
could be obtained (this is due to the acidity of the proton in the
benzylic captodative position).
In order to overcome the increased acidity of the proton in
the synthesis of the enediyne substrate, and at the same time to
improve the level of memory of chirality by reducing the
conformational flexibility, enediyne (S)- and (R)-20 were
prepared from the oxazolidinones derived from (S)- and (R)-
phenyl glycinols, respectively. The results are shown in Scheme
4. Previous studies had shown that the tetracyclic sulfones
derived from these oxazolidinones were easily epimerized in
acetonitrile. Therefore, the rearrangement of 20 was achieved
under different experimental conditions, and benzene was first
selected as the most suitable solvent to minimize the possible
epimerization of the cis and trans isomers of product 21.30
Much to our disappointment, the increased flexibility of the
side chain promoted a disproportionation, that is, hydrogen
CONCLUSION
■
In summary, the cascade rearrangement of enediynes of type 1,
starting with 1,3-proton shift immediately followed by Saito−
Myers cyclization, was shown to evolve via exclusive 1,5- or 1,6-
hydrogen shift depending on the structure of the starting
material. Computational data confirmed these kinetic prefer-
ences. The dramatic influence of the substituent attached to the
nitrogen atom of the amino ester moiety was emphasized. In
the alanine derivative 1a, exclusive 1,5-hydrogen abstraction of
the hydrogen atom in captodative position is well accounted for
by the incidence of the greater exergonicity of this process as
compared to 1,6-hydrogen shift from the methyl group.
2779
dx.doi.org/10.1021/jo202580y | J. Org. Chem. 2012, 77, 2773−2783

The Journal of Organic Chemistry
Article
(3S*, 4S*)-d1-2a (63 mg, 35%, 81% deuterium labeling), and d1-3a
(38 mg, 21%, 75% deuterium labeling).
Conversely, the “exclusive” kinetic preference for 1,6-hydrogen
abstraction over 1,5-migration of the hydrogen atom in
captodative position, observed in the rearrangement of the
valine derivative 1b, is quite unexpected. It is the bulky tosyl
group attached to nitrogen which contributes to increase the
activation free energy of 1,5-hydrogen shift. This is rationalized,
according to theoretical calculations, by the higher energy of
the conformation that must be reached for model radical 12 to
be ready to transfer the hydrogen atom from the captodative
position. The validity of calculations was confirmed by the
rearrangement of rac-enediyne 13.
The ready occurrence of 1,6-hydrogen transfer was
tentatively exploited, with compounds where no competitive
1,5-hydrogen shift is available, to achieve the enantioselective
synthesis of naphthoazepines based on the phenomenon of
memory of chirality. The main drawback is that, whenever
possible without introducing torsional constraints, dispropor-
tionation becomes the major pathway with respect to
recombination in the last elementary step.
Methyl (3S*,4R*)-3-Methyl-2,4-bis[(4-methylbenzene)-
sulfonyl](10-2H)-1H,2H,3H,4H-benzo[g]isoquinoline-3-carbox-
ylate (d1-2a). HRMS (ESI): m/z calcd for [M + NH4]+
C30H32DN2O6S2 582.1837, found 582.1833. 1H NMR (400 MHz,
CDCl3) δ: 7.85 (2H, d, J = 8.3), 7.73 (1H, br d, J = 8.8), 7.59 (1H, br
d, J = 7.8), 7.50−7.40 (2.25H, dt (2H) plus superimposed residual s
(0.25H)), 7.33−7.26 (5H, m (s superimposed to two d)), 7.08 (2H, d,
J = 7.9), 4.89 (1H, d, J = 15.3, A part of an AB pattern), 4.73 (1H, s),
4.63 (1H, d, J = 15.1, B part of an AB pattern), 3.15 (3H, s), 2.42 (3H,
2
s), 2.39 (3H, s), 2.34 (3H, s). H NMR (61.4 MHz, acetone) δ: 7.73
(br s).
Methyl (3S*,4S*)-3-Methyl-2,4-bis[(4-methylbenzene)-
sulfonyl](10-2H)-1H,2H,3H,4H-benzo[g]isoquinoline-3-carbox-
ylate (d1-2a). HRMS (ESI): m/z calcd for [M + NH4]+
C30H32DN2O6S2 582.1837, found 582.1835. 1H NMR (400 MHz,
CDCl3) δ: 7.99 (2H, d, J = 8.3), 7.77 (1H, d, J = 7.9), 7.57 (0.19H, s,
residual CHar), 7.54−7.46 (1H, m), 7.43 (2H, m), 7.33 (1H, d, J =
8.1), 7.36−7.29 (1H, m), 7.25 (2H, d, J = 8.1), 7.00 (2H, d, J = 8.1),
6.86 (1H, s), 4.60 (2H, br s), 4.50 (1H, s), 4.01 (3H, s), 2.44 (3H, s),
2.33 (3H, s), 1.76 (3H, s). 2H NMR (61.4 MHz, acetone) δ: 7.85 (br
s).
Methyl 2-{N-[(3-{[(4-Methylbenzene)sulfonyl](2H1)methyl}-
naphthalen-2-yl)methyl](4-methylbenzene)sulfonamido}(2H2)-
prop-2-enoate (d1-3a). HRMS (ESI): m/z calcd for [M + NH4]+
C30H32DN2O6S2 582.1837, found 582.1840. 1H NMR (400 MHz,
CDCl3) δ: 7.88 (2H, d, J = 8.3), 7.77 (1H, s), 7.74 (2H, d, J = 8.3),
7.79−7.64 (2H, superimposed m), 7.50−7.44 (2H, m), 7.43 (0.25H,
residual s), 7.37−7.32 (4H, 2 × d, J = 8.1), 6.14 (1H, s), 5.44 (1H, s),
4.84 (2H, s), 4.77 (2H, s), 3.53 (3H, s), 2.48 (3H, s), 2.47 (3H, s). 2H
NMR (61.4 MHz, acetone) δ: 7.75 (br s).
Rearrangement of (rac)-d1-1b. Monodeuterated (rac)-d1-4b
was prepared according to the procedure already described for the
synthesis of 4b, using (rac)-d1-1b (101 mg, 0.17 mmol) and GA-SBA-
15 (24 mg, 0.03 mmol) in CH3CN (10 mL). The reaction mixture was
stirred for 18 h at 80 °C. After workup, purification by liquid
chromatography on silica gel (pentane/Et2O, 70/30 to 60/40) led to
(rac)-d1-4b (66 mg, 65% yield, and 83% deuterium labeling) as a
yellow oil.
EXPERIMENTAL SECTION
■
Rearrangement of (S)-1b. A solution of (S)-1b (95 mg, 0.16
mmol) and GA-SBA-15 (23 mg, 0.03 mmol) in CH3CN (10 mL) was
stirred for 18 h at 80 °C. Then the solvent was removed in vacuo, and
the residue was purified by flash chromatography on silica gel
(pentane/Et2O, 90/10 to 50/50). This led to 4b (63 mg, 66%) as a
yellow oil.
Methyl (2S)-3-Methyl-2-{N-[(3-{[(4-methylbenzene)sulfonyl]-
methyl}naphthalen-2-yl)methyl](4-methylbenzene)-
sulfonamido}but-3-enoate (4b). ee > 99% (Chiralpak IA, hexane/
EtOH 70/30, 1 mL/min, tR(R) = 11.44 min, tR(S) = 12.69 min, k(R) =
2.81, k(S) = 3.23, α = 1.15, RS = 1.74). [α]25 +29 (c 1.7, CHCl3).
D
HRMS (ESI): m/z calcd for [M + NH4]+ C32H37N2O6S2 609.2088,
found 609.2091. 1H NMR (400 MHz, CDCl3) δ: 7.92 (1H, s), 7.71−
7.69 (1H, br d, J = 8.2), 7.64−7.62 (1H, br d, J = 7.5), 7.54 (2H, d, J =
8.0), 7.51 (2H, d, J = 8.0), 7.47−7.40 (2H, m), 7.39 (1H, s), 7.23 (2H,
d, J = 8.0), 7.16 (2H, d, J = 8.0), 5.07 (1H, s), 4.81 (1H, s), 4.74 (1H,
s), 4.66 (2H, s), 4.62 (1H, d (A part of AB pattern), J = 14.3), 4.55
(2H, d (B part of AB pattern), J = 14.3), 3.43 (3H, s), 2.43 (3H, s,
CH3), 2.36 (3H, s), 1.58 (3H). 13C NMR (100 MHz, CDCl3) δ:
170.0, 144.9, 143.7, 138.8, 136.7, 135.4, 134.2, 133.1, 132.2, 132.1,
129.9, 129.7 (2 × C), 129.4 (2 × C), 128.9 (2 × C), 127.8, 127.5,
127.4 (2 × C), 126.8, 126.4, 124.6, 117.4, 64.4, 60.0, 52.0, 46.3, 21.8,
21.6, 21.5.
(rac)-Methyl 3-methyl-2-{N-[(3-{[(4-methylbenzene)-
sulfonyl]methyl}naphthalen-2-yl)methyl](4-methylbenzene)-
sulfonamido}(2-2H)but-3-enoate (d1-4b). HRMS (ESI): m/z
calcd for [M + NH4]+ C32H36DN2O6S2 610.2150, found 610.2136.
1H NMR (400 MHz, CDCl3) δ: 7.92 (1H, s), 7.71−7.69 (1H, m),
7.64−7.62 (1H, m), 7.53 (2H, d, J = 8.3), 7.51 (2H, d, J = 8.3), 7.47−
7.40 (2H, m), 7.39 (1H, s), 7.23 (2H, d, J = 8.0), 7.16 (2H, d, J = 8.0),
5.07 (0.16H, residual s), 4.81 (1H, s), 4.74 (1H, s), 4.66 (2H, s), 4.62
(1H, d (A part of AB pattern), J = 14.3), 4.55 (2H, d (B part of AB
pattern), J = 14.3), 3.43 (3H, s), 2.43 (3H, s), 2.36 (3H, s), 1.58 (3H,
Methyl 3-Methyl-2-{N-[(3-{[(4-methylbenzene)sulfonyl]-
methyl}naphthalen-2-yl)methyl](4-methylbenzene)-
sulfonamido}but-2-enoate (3b). A solution of (S)-4b (67 mg, 0.11
mmol) and GA-SBA-15 (160 mg, 0.23 mmol) in CH3CN (7 mL) was
stirred for 7 days at 80 °C. Then the mixture was filtered and washed
with dichloromethane. The solvent was removed in vacuo, and the
residue was purified by flash chromatography on silica gel (pentane/
Et2O, 80/20 to 40/60). This led to 3b (47 mg, 70%) as a yellow oil.
HRMS (ESI): m/z calcd for [M + NH4]+ C32H37N2O6S2 609.2088,
found 609.2087. 1H NMR (400 MHz, CDCl3) δ: 7.89 (2H, d, J = 8.3),
7.79−7.69 (5H, m), 7.53−7.44 (2H, m), 7.39 (1H, s), 7.35 (4H, d, J =
8.0), 5.54 (1H, d, J = 13.8), 5.38 (1H, d, J = 13.8), 4.38 (1H, d, J =
13.5), 4.31 (1H, d, J = 13.8), 4.31 (3H, s), 2.47 (3H, s), 2.46 (3H, s),
1.87 (3H, s), 1.31 (3H, s). 13C NMR (100 MHz, CDCl3) δ: 164.9,
160.04, 144.9, 143.53, 136.6, 136.5, 133.3, 133.06, 133.00, 132.1,
132,0, 129,9 (2 × C), 129,5 (2 × C), 128.8 (2 × C), 128.0 (2 × C),
127.9, 127,6, 127.1, 126.9, 125.4, 121.4, 59.1, 51.1, 50.9, 23.8, 22.0,
21.8, 21.7.
2
s). H NMR (61.4 MHz, acetone) δ: 5.08 (br s).
Rearrangement of (rac)-d3-1a. A solution of (rac)-d3-1a (200
mg, 0.35 mmol) and GA-SBA-15 (50 mg, 0.07 mmol) in CH3CN (20
mL) was stirred for 15 h at 80 °C. Then the solvent was removed in
vacuo, and the residue was purified by flash chromatography on silica
gel (pentane/dichloromethane, 50/50 to 80/20). This led to three
products, (3S*,4R*)-d3-2a (90 mg, 45%, 100% deuterium labeling),
(3S*,4S*)-d3-2a (77 mg, 38%, 100% deuterium labeling), and d3-3a
(20 mg, 10%, 100% deuterium labeling).
Methyl (3S,4R)-3-(2H3)Methyl-2,4-bis[(4-methylbenzene)-
sulfonyl]-1H,2H,3H,4H-benzo[g]isoquinoline-3-carboxylate
(d3-2a). HRMS (ESI): m/z calcd for [M + NH4]+ C30H30D3N2O6S2
584.1963, found 584.1958. 1H NMR (400 MHz, CDCl3) δ: 7.87 (2H,
d, J = 8.3), 7.75 (1H, br d, J = 7.9), 7.61 (1H, br d, J = 7.6), 7.50 (1H,
s), 7.50−7.40 (2H, m), 7.33−7.26 (5H, m (s superimposed to two d)),
7.08 (2H, d, J = 7.9), 4.91 (1H, d, J = 15.3, A part of an AB pattern),
4.74 (1H, s), 4.63 (1H, d, J = 15.1, B part of an AB pattern), 3.17 (3H,
Rearrangement of (rac)-d1-1a. A solution of (rac)-d1-1a (183
mg, 0.32 mmol) and GA-SBA-15 (46 mg, 0.06 mmol) in CH3CN (18
mL) was stirred for 15 h at 80 °C. Then the solvent was removed in
vacuo, and the residue was purified by flash chromatography on silica
gel (dichloromethane/Et2O, 100/00 to 99/1). This led to three
products, (3S*, 4R*)-d1-2a (73 mg, 40%, 75% deuterium labeling),
2
s), 2.40 (3H, s), 2.37 (3H, s). H NMR (61.4 MHz, acetone) δ 2.34
(br s).
Methyl (3S,4S)-3-(2H3)Methyl-2,4-bis[(4-methylbenzene)-
sulfonyl]-1H,2H,3H,4H-benzo[g]isoquinoline-3-carboxylate
2780
dx.doi.org/10.1021/jo202580y | J. Org. Chem. 2012, 77, 2773−2783

The Journal of Organic Chemistry
Article
(d3-2a). HRMS (ESI): m/z calcd for [M + NH4]+ C30H30D3N2O6S2
584.1963, found 584.1958. 1H NMR (400 MHz, CDCl3) δ: 7.99 (2H,
d, J = 8.3), 7.77 (1H, d, J = 7.9), 7.57 (1H, s), 7.54−7.46 (1H, m),
7.45−7.40 (2H, m), 7.33 (2H, d, J = 8.1), 7.25 (2H, d, J = 8.3), 7.00
(2H, d, J = 8.1), 6.87 (1H, s), 4.60 (2H, br s), 4.50 (1H, s), 4.01 (3H,
the signal of the proton at 5. 95 ppm and the signal of the methyl
protons at 2.55 ppm.
Methyl 2-{N-[2-(3-{[(4-Methylbenzene)sulfonyl]methyl}-
naphthalen-2-yl)ethyl](4-methylbenzene)sulfonamido}prop-2-
enoate (19). HRMS (ESI): m/z calcd for [M + NH4]+ C31H35N2O6S2
595.1931, found 595.1928. 1H NMR (400 MHz, CDCl3) δ: 7.73 (1H,
d, J = 8.0), 7.69−7.66 (1H, m), 7.68 (1H, superimposed s), 7.66 (1H,
superimposed s, CHar), 7.62 (2H, d, J = 8.5), 7.59 (2H, d, J = 8.5),
7.50−7.41 (2H, m), 7.26 (2H, d, J = 8.0), 7.23 (2H, d, J = 8.0), 6.44
(1H, s), 5.81 (1H, s), 4.49 (2H, br s), 3.65 (3H, s), 3.64−3.60 (2H,
m), 2.97−2.93 (2H, m), 2.43 (3H, s), 2.43 (3H, s). 13C NMR (100
MHz, CDCl3) δ: 164.3, 145.0, 143.8, 136.0, 135.7, 135.2, 133.5, 132.6,
132.1, 129.8 (2 × C), 129.6 (2 × C), 129.2, 128.8, 128.7 (2 × C),
127.8 (2 × C), 127.7, 127.2, 127.0, 126.2, 124.8, 59.6, 52.6, 50.3, 32.5,
21.8, 21.7. One carbon is missing probably due to overlapping.
Rearrangement of (R)-20. (R)-20 (100 mg, 0.21 mmol) and GA-
SBA-15 (29 mg, 0.04 mmol) were dissolved in benzene (10 mL). The
mixture was heated at 80 °C for 16.5 h. The solvent was removed in
vacuo, and the residue was purified by flash chromatography on silica
gel (pentane/AcOEt, 80/20 to 50/50). This led to (2S,3R)-21 (24 mg,
24%) as a yellow amorphous solid and 22 (40 mg, 40%) as a yellow
amorphous solid.
2
s), 2.44 (3H, s), 2.33 (3H, s). H NMR (61.4 MHz, acetone) δ 1.67
(br s).
Methyl 2-{N-[(3-{[(4-Methylbenzene)sulfonyl](2H1)methyl}-
naphthalen-2-yl)methyl](4-methylbenzene)sulfonamido}(2H2)-
prop-2-enoate (d3-3a). HRMS (ESI): m/z calcd for [M + NH4]+
C30H30D3N2O6S2 584.1963, found 584.1961. 1H NMR (400 MHz,
CDCl3) δ: 7.88 (2H, d, J = 8.3), 7.77 (1H, s), 7.77−7.72 (3H,
superimposed m), 7.69−7.64 (1H, m), 7.52−7.45 (2H, m), 7.43 (1H,
s), 7.37−7.32 (4H, 2 superimposed d, J = 8.1), 4.84 (1H, s), 4.77 (2H,
2
s), 3.53 (3H, s), 2.48 (3H, s), 2.47 (3H, s). H NMR (61.4 MHz,
acetone) δ: 6.15 (br s), 5.63 (br s), 5.00 (br s).
Rearrangement of ( )-13. (CH2O)n (26 mg, 0.88 mmol), CuI
(34 mg, 0.18 mmol), 1,4-dioxane (1 mL), enediyne ( )-13 (100 mg,
0.35 mmol), and dicylohexylamine (115 mg, 0.63 mmol) were added
sequentially into a flask equipped with a reflux condenser under an
argon atmosphere. The resulting mixture was stirred at reflux for 2 h.
Solvent was evaporated under vacuo, the residue was dissolved in
DCM and filtered over a short pad of Celite, and the filtrate was
concentrated and purified by column chromatography on silica gel
using pentane/Et2O (95:5 to 90:10) as eluent to afford ( )-15 (29
mg, 28%) and ( )-16 (26 mg, 25%).
Methyl 2-Methyl-3-(propan-2-yl)-1H,2H,3H,4H-benzo[g]-
isoquinoline-3-carboxylate ( )-15. HRMS (ESI): m/z calcd for
[M + H]+ C19H24NO2 298.1802, found 298.1801. 1H NMR (400
MHz, CDCl3) δ: 7.75−7.69 (2H, m), 7.62 (1H, s), 7.46 (1H, s), 7.40−
7.33 (2H, m), 4.10 (1H, d (A part of AB pattern), J= 16.3), 3.77 (1H,
d (B part of AB pattern), J = 16.3), 3.60 (3H, s), 3.31 (1H, d (A part
of AB pattern), J = 16.3), 3.02 (1H, d (B part of AB pattern), J = 16.3),
2.58 (3H, s), 2.45 (1H, sept, J = 6.8), 1.06 (3H, d, J = 6.8), 1.00 (3H,
d, J = 6.8). 13C NMR (100 MHz, CDCl3) δ: 173.8, 133.0, 132.6, 132.2,
132.0, 127.3, 127.2, 126.5, 125.2, 125.1, 124.0, 68.6, 55.3, 51.1, 38.2,
31.3, 29.8, 18.0, 16.6.
Methyl 2-{1H,2H,3H,4H-Benzo[g]isoquinolin-2-yl}-3-methyl-
butanoate ( )-16. HRMS (ESI): m/z calcd for [M + H]+
C19H24NO2 298.1802, found 298.1803. 1H NMR (400 MHz,
CDCl3) δ: 7.74−7.69 (2H, m), 7.57 (1H, s), 7.50 (1H, s), 7.40−
7.35 (2H, m), 4.00−3.90. (2H, AB pattern, JAB = 16.3), 3.73 (3H, s),
3.10−3.02 (3H, m), 3.01 (1H, d, J = 10.8), 2.79 (1H, m), 2.27−2.16
(1H, m), 1.03 (3H, d, J = 6.8), 0.95 (3H, d, J = 6.8). 13C NMR (100
MHz, CDCl3) δ: 172.3, 134.2, 133.6, 132.4, 132.0, 127.3, 127.2, 126.8,
125.3, 125.3, 124.6, 74.4, 53.3, 50.8, 47.2, 30.4, 27.3, 20.0, 19.5.
Rearrangement of (S)-17. Compound (S)-17 (100 mg, 0.17
mmol) and GA-SBA-15 (24 mg, 0.03 mmol) were dissolved in
CH3CN (10 mL). The mixture was heated at 80 °C for 22 h. Then the
solvent was removed in vacuo, and the residue was purified by flash
chromatography on silica gel (pentane/dichloromethane, 50/50 to 0/
100). This led to (1S, 2S)-18 (38 mg, 38%) as a yellow oil and 19 (52
mg, 52%) as a yellow oil.
(2S,3R)-2-[(4-Methylbenzene)sulfonyl]-3-phenyl-5-oxa-7-
azatetracyclo[8.8.0.03,7.012,17]octadeca-1(18),10,12(17),13,15-
pentaen-6-one (21). ee = 84% (Chiralpak IB, hexane/EtOH/
chloroform 60/30/10, 1 mL/min, tR(2S, 3R) = 6.56 min, tR(2R, 3S) =
8.21 min, k(2S,3R) = 1.19, k(2R,3S) = 1.68, α = 1.41 and RS = 2.70).
[α]25 −103 (c 0.48, CHCl3). HRMS (ESI): m/z calcd for [M +
D
NH4]+ C29H29N2O4S 501.1843, found 501.1842. 1H NMR (400 MHz,
CDCl3) δ: 7.59 (1H, d, J = 8.0), 7.43−7.23 (10H, m), 7.15 (1H, tt, J =
8.3 and 1.5), 6.98 (1H, s), 6.91 (2H, d, J = 8.0), 5.72 (1H, d, J = 9.0),
5.19 (1H, s), 4.40 (1H, ddd, J = 14.3, 7.3 and 2.8), 4.25 (1H, br t, J =
14.0), 4.07 (1H, d, J = 9.0), 3.26 (1H, pseudo t, J = 13.0), 2.80 (1H, br
dd, J = 15.3 and 4.0), 2.12 (3H, s). 13C NMR (100 MHz, CDCl3) δ:
158.2, 144.9, 140.0, 138.1, 136.4, 135.8, 132.8, 131.4, 130.0, 129.5 (2 ×
C), 129.2 (2 × C), 128.8, 128.5 (2 × C), 128.3, 127.3, 127.0, 126.9,
126.1, 125.9 (2 × C), 77.8, 74.0, 67.7, 40.9, 35.8, 21.4. In full
agreement with the Chem-3D model showing a π-stacking interaction,
the stereochemistry was assigned from 2D NMR experiments. The
NOESY spectrum shows a clear cross-peaks between the signal of the
proton in position α to the tosyl group at 5.19 ppm and the s signal of
one aromatic proton of the naphthyl group at 6.98 ppm and the d of
one proton of the CH2O group of the carbamate ring at 5.72 ppm.
3-[2-(3-{[(4-Methylbenzene)sulfonyl]methyl}naphthalen-2-
yl)ethyl]-4-phenyl-2,3-dihydro-1,3-oxazol-2-one (22). HRMS
(ESI): m/z calcd for [M + NH4]+ C29H29N2O4S 501.1843, found
1
501.1840. H NMR (400 MHz, CDCl3) δ: 7.69−7.65 (2H, m), 7.54
(2H, d, J = 8.3), 7.51 (1H, s), 7.49−7.37 (6H, m), 7.23 (2H, d, J =
8.3), 7.13−7.09 (2H, m), 6.74 (1H, s), 4.28 (2H, s), 3.82−3.77 (2H,
m), 2.96−2.91 (2H, m), 2.42 (3H, s). 13C NMR (100 MHz, CDCl3)
δ: 156.1, 145.0, 135.5, 134.5, 133.5, 132.8, 132.2, 129.9, 129.8 (2 × C),
129.7, 129.4, 129.3 (2 × C), 128.8 (2 × C), 128.6 (2 × C), 127.8,
127.2, 127.1, 126.3, 126.3, 124.7, 123.9, 59.3, 43.0, 31.8, 21.8.
Rearrangement of (S)-20. (S)-20 (200 mg, 0.41 mmol) and GA-
SBA-15 (58 mg, 0.08 mmol) were dissolved in benzene (20 mL). The
mixture was heated at 80 °C for 16.5 h. The solvent was removed in
vacuo, and the residue was purified by flash chromatography on silica
gel (pentane/AcOEt, 80/20 to 50/50). This led to (2R,3S)-21 (54 mg,
27%) as a yellow amorphous solid and 22 (98 mg, 49%) as a yellow
amorphous solid.
Methyl (1S,2S)-2-Methyl-1,3-bis[(4-methylbenzene)-
sulfonyl]-1H,2H,3H,4H,5H-naphtho[2,3-d]azepine-2-carboxy-
late (18). ee = 56% (Chiralpak IC, hexane/EtOH 50/50, 1 mL/min,
tR(1S,2S) = 15.80 min, tR(1S,2S) = 28.43 min, k(1R,2R) = 4.27,
k(1R,2S) = 8.45, α = 1.98, RS = 7.05). [α]25 +23 (c 0.6, CHCl3).
D
HRMS (ESI): m/z calcd for [M + NH4]+ C31H35N2O6S2 595.1931,
found 595.1928. 1H NMR (400 MHz, CDCl3) δ: 7.73 (1H, d, J = 8.0),
7.62 (1H, s), 7.47 (2H, d, J = 8.3), 7.41 (1H, dt, J= 1.0 and 6.8), 7.33−
7.29 (3H, m), 7.22 (1H, d, J = 8.0), 7.11 (2H, d, J = 8.0), 7.05 (2H, d,
J = 8.0), 6.68 (1H, s), 5.95 (1H, s), 4.11 (1H, ddd, J = 6.8, 11.8 and
14.8), 3.92−3.80 (2H, m), 3.62 (3H, s), 3.00 (1H, ddd, J = 1.8, 5.3 and
14.8), 2.55 (3H, s), 2.35 (3H, s), 2.32 (3H, s). 13C NMR (100 MHz,
CDCl3) δ 172.5, 145.7, 142.8, 139.3, 136.2, 135.9, 135.7, 133.7, 131.7,
129.3, 129.3 (2 × C), 129.22 (2 × C), 129.21 (2 × C), 129.0, 127.7,
127.1, 126.8, 126.4 (2 × C), 125.7, 78.4, 66.6, 53.4, 45.4, 33.0, 23.3,
21.6, 21.5. The cis stereochemistry was assigned from 2D NMR
experiments. The NOESY spectrum shows a clear cross-peak between
(2R,3S)-2-[(4-Methylbenzene)sulfonyl]-3-phenyl-5-oxa-7-
azatetracyclo[8.8.0.03,7.012,17]octadeca-1(18),10,12(17),13,15-
pentaen-6-one (21). ee = 77% (Chiralpak IB, hexane/EtOH/
chloroform 60/30/10, 1 mL/min, tR(2S,3R) = 6.56 min, tR(2R,3S) =
8.21 min, k(2S,3R)= 1.19, k(2R,3S) = 1.68, α = 1.41 and RS = 2.70).
[α]25 +118 (c 0.45, CHCl3).
D
2781
dx.doi.org/10.1021/jo202580y | J. Org. Chem. 2012, 77, 2773−2783

The Journal of Organic Chemistry
Article
N. Chimia 2008, 62, 510−513. (c) Robertson, J.; Pillai, J.; Lush, R. K.
Chem. Soc. Rev. 2001, 30, 94−103.
ASSOCIATED CONTENT
* Supporting Information
Synthesis of starting materials, HPLC analyses, NMR spectra
for all new compounds, and computational details. This
material is available free of charge via the Internet at http://
pubs.acs.org
■
S
(11) For examples of exclusive 1,6-hydrogen shift without possible
competition of 1,5-hydrogen shift, see: (a) Gramain, J.-C.; Remuson,
́
R.; Vallee, D. J. Org. Chem. 1985, 50, 710−712. (b) Furata, K.; Nagata,
T.; Yamamoto, H. Tetrahedron Lett. 1988, 29, 2215−2218. (c) Kraus,
G. A.; Chen, L. J. Am. Chem. Soc. 1990, 112, 3464−3466. (d) Bosch,
E.; Bachi, M. D. J. Org. Chem. 1993, 58, 5581−5582. (e) Bogen, S.;
Fensterbank, L.; Malacria, M. J. Am. Chem. Soc. 1997, 119, 5037−5038.
(f) Manabe, T.; Yanagi, S.-I.; Ohe, K.; Uemura, S. Organometallics
1998, 17, 2942−2944. (g) Lin, H.; Schall, A.; Reiser, O. Synlett 2005,
2603−2606. (h) Sakaguchi, N.; Hirano, S.; Matsuda, A.; Shuto, S. Org.
Lett. 2006, 8, 3291−3294. (i) Yoshimura, Y.; Yamazaki, Y.; Wachi, K.;
Satoh, S.; Takahata, H. Synlett 2007, 111−114.
AUTHOR INFORMATION
Corresponding Authors
*(A.G-S.) E-mail: anouk.siri@univ-cezanne.fr.
*(M.N.) E-mail: malek.nechab@univ-provence.fr.
*(M.P.B.) E-mail: michele.bertrand@univ-cezanne.fr.
■
(12) For competitive processes, see: (a) Ayer, W. A.; Law, D. A.;
Notes
Piers, K. Tetrahedron Lett. 1964, 5, 2959−2963. (b) Jose,
́
I. C.;
The authors declare no competing financial interest.
Francisco, C. G.; Hernandez, R.; Salazar, J. A.; Suarez, E. Tetrahedron
́
́
Lett. 1984, 25, 1953−1956. (c) McDonald, C. E.; Beebe, T.; Beard,
M.; McMillen, D.; Selski, D. Tetrahedron Lett. 1989, 30, 4791−4794.
(d) Bogen, S.; Fensterbank, L.; Malacria, M. J. Org. Chem. 1999, 64,
819−825. (e) Wille, U. Chem.Eur. J. 2002, 8, 340−347. (f) Wille, U.
J. Am. Chem. Soc. 2002, 124, 14−15. (g) Wille, U.; Heuger, G.;
Jargstorff, C. J. Org. Chem. 2008, 73, 1413−1421 and previous
references cited therein. (h) Harrowen, D. C.; Stenning, K. J.; Whiting,
S.; Thompson, T.; Walton, R. Org. Biomol. Chem. 2011, 9, 4882−4885.
ACKNOWLEDGMENTS
We thank the ANR (JCJC MOCER2) for financial support and
■
the Universite
́
de Provence for the postdoctoral grant of Dr S.
Mondal. This work was also supported by the computing
facilities of the CRCMM.
REFERENCES
■
́ ́
(i) Francisco, C. G.; Freire, R.; Herrere, A. J.; Perez-Martín, I.; Suarez,
(1) Nechab, M.; Campolo, D.; Maury, J.; Perfetti, P.; Vanthuyne, N.;
Siri, D.; Bertrand, M. P. J. Am. Chem. Soc. 2010, 132, 14742−14744.
(2) For selected examples of other cascade rearrangements of
enediynes, following on from Bergman or Saito−Myers cyclizations,
see: (a) Grissom, J. W.; Klingberg, D.; Huang, D.; Slattery, B. J. J. Org.
Chem. 1997, 62, 603−626 and references therein. (b) Li, H.; Petersen,
J. L.; Wang, K. K. J. Org. Chem. 2003, 68, 5512−5518. (c) Tarli, A.;
Wang, K. K. J. Org. Chem. 1997, 62, 8841−8847. (d) Wang, K. K.;
Wang, Z.; Tarli, A.; Gannett, P. J. Am. Chem. Soc. 1996, 118, 10783−
10791. (e) Wang, K. K.; Zhang, H.-R.; Petersen, J. L. J. Org. Chem.
1999, 64, 1650−1656.
(3) (a) Fuji, K.; Kawabata, T. Chem.Eur. J. 1998, 4, 373−376.
(b) Zhao, H.; Hsu, D. C.; Carlier, P. R. Synthesis 2005, 1−16.
(c) Kawabata, T.; Fuji, K. In Topics in Stereochemistry; Denmark, S. E.,
Ed.; Wiley and Sons, Inc.; New York, 2003; Vol. 23, pp 175−205.
(4) For examples of disproportionation of biradicals, see: (a) Bucher,
G.; Mahajan, A. A.; Schmittel, M. J. Org. Chem. 2009, 74, 5850−5860.
(b) Choe, T.; Khan, S. I.; Garcia-Garibay, M. A. Photochem. Photobiol.
Sci. 2006, 5, 449−451. (c) Wagner, P. J.; Laidig, G. Tetrahedron Lett.
1991, 32, 895−898. (d) Wagner, P. J.; Chiu, C. J. Am. Chem. Soc.
1979, 101, 7134−7135.
(5) Nechab, M.; Besson, E.; Campolo, D.; Perfetti, P.; Vanthuyne, N.;
Bloch, E.; Denoyel, R.; Bertrand, M. P. Chem. Commun. 2011, 47,
5286−5288.
(6) Mondal, S.; Nechab, M.; Vanthuyne, N.; Bertrand, M. P. Chem.
Commun. 2012, 48, 2549−2551.
(7) For (3S*,4S*)-d1-2a the deuterated position gives rise to a
residual s at 7.57 ppm (0.19H)/CHTs s at 4.50 ppm (1H). For
(3S*,4R*)-d1-2a the deuterated position gives rise to a residual s
superimposed to a m between 7.51 and 7.59 ppm (2.25H)/CHTs s at
4.72 ppm (1H). For d1-3a the deuterated position gives rise to a
residual s at 7.43 ppm (0.25H)/CH at 6.14 (1H) and 5.41 (1H),
respectively.
(8) For the competitive migration of H/D in Schmittel rearrange-
ment, see: Bucher, G.; Mahajan, A. A.; Schmittel, M. J. Org. Chem.
2009, 74, 5950−5860.
E. Tetrahedron 2007, 63, 8910−8920.
(13) (a) Curran, D. P.; Kim, D.; Liu, H. T.; Shen, W. J. Am. Chem.
Soc. 1988, 110, 5900−5902. (b) Snieckus, V.; Cuevas, J.-C.; Sloan, C.
P.; Liu, H.; Curran, D. P. J. Am. Chem. Soc. 1990, 112, 896−898.
(c) Curran, D. P.; Somayajula, K. V.; Yu, H. Tetrahedron Lett. 1992,
33, 2295−2298. (d) Curran, D. P.; Shen, W. J. Am. Chem. Soc. 1993,
115, 6051−6059. (e) Yamasaki, N.; Eichenberger, E.; Curran, D. P.
Tetrahedron Lett. 1994, 35, 6623−6226. (f) Curran, D. P.; Xu, J. Y. J.
Am. Chem. Soc. 1996, 118, 3142−3147. (g) Curran, D. P.; Liu, W.;
Chen, C. H.-T. J. Am. Chem. Soc. 1999, 121, 11012−11013.
(14) (a) Denenmark, D.; Hoffmann, P.; Winkler, T.; Waldner, A.; De
Mesmaeker, A. Synlett 1991, 621−624. (b) Denenmark, D.; Winkler,
T.; Waldner, A.; De Mesmaeker, A. Tetrahedron Lett. 1992, 33, 621−
624.
(15) (a) Dorigo, A. E.; Houk, K. N. J. Am. Chem. Soc. 1987, 109,
2195−2197. Conversely, in Norrish II photochemical reactions, 1,5-
hydrogen transfer would be favored over 1−6 hydrogen migration by
both enthalpic and entropic factors, see: (b) Dorigo, A. E.; McCarrick,
M. A.; Loncharic, R. J.; Houk, K. N. J. Am. Chem. Soc. 1990, 112,
7508−7514.
(16) Rey, V.; Pierini, A.; Penenory, A. B. J. Org. Chem. 2009, 74,
́
̃ ̃
1223−1230.
(17) For calculations of 1,5-H migration following on from a
Bergman rearrangement, see: Baroudi, A.; Mauldin, J.; Alabugin, I. V. J.
Am. Chem. Soc. 2010, 132, 967−979.
(18) Zeng, L.; Kaoudi, T.; Schiesser, C. H. Tetrahedron Lett. 2006,
47, 7911−7914.
(19) It can be noted that the presence of a nitrogen atom protected
as an acetamide or a trifluoroacetamide in the chain makes the
enthalpically favored 1,7-H migration of a benzylic hydrogen atom
compete with 1,5-hydrogen shift.
(20) Gaussian 09, Revision A.02: Frisch, M. J.; Trucks, G. W.;
Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.;
Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji,
H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.;
Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.;
Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao,
O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.; Peralta, J. E.;
Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.;
Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.;
Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega,
N.; Millam, N. J.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.;
Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.;
(9) For reviews, see: (a) Cekovic, Z. J. Serb. Chem. Soc. 2005, 70,
287−318. (b) Cekovic, Z. Tetrahedron 2003, 59, 8073. (c) Feray, L.;
Kuznetsov, N.; Renaud, P. In Radicals in Organic Synthesis; Renaud, P.,
Sibi, M. P., Eds.; Wiley-VCH: Weinheim, 1998; Vol. 2, pp 246−278.
(d) Togo, H. In Avanced Free Radical Reactions for Organic Synthesis;
Elsevier: Oxford, 2004; pp 171−186.
(10) (a) Den
(b) Renaud, P.; Beaufils, F; Den
́
es
̀
, F.; Beaufils, F.; Renaud, P. Synlett 2008, 2389−2399.
es, Feray, L.; Imboden, C.; Kuznetsov,
́
̀
2782
dx.doi.org/10.1021/jo202580y | J. Org. Chem. 2012, 77, 2773−2783

The Journal of Organic Chemistry
Article
Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.;
Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.;
̈
Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.;
Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian, Inc.,
Wallingford, CT, 2009.
(21) Scott, A. P.; Radom, L. J. Phys. Chem. 1996, 100, 16502−16513.
(22) (a) Henry, D.; Sullivan, M.; Radom, L. J. Chem. Phys. 2003, 118,
4849−4860.
(23) Viehe, H. G.; Janousek, Z.; Merenyi, R.; Stella, L. Acc. Chem. Res.
1985, 18, 148−154.
(24) For recent discussions on radical stabilization energies, see:
(a) Coote, M. L.; Lin, C. Y.; Beckwith, A. L. J.; Zavitsas, A. A. Phys.
Chem. Chem. Phys. 2010, 12, 9597−9610. (b) Menon, A. S.; Wood, G.
P. F; Moran, D.; Radom, L. J. Phys. Chem. A 2007, 111, 13638−13644.
(c) Poutsma, M. L. J. Org. Chem. 2011, 76, 270−276. (d) Wood, G. P.
F.; Moran, D.; Jacob, R.; Radom. J. Phys. Chem. A 2005, 109, 6318−
6325.
(25) Hammond, G. S. J. Am. Chem. Soc. 1955, 77, 334−338.
(26) (a) Marcus, R. A. Annu. Rev. Phys. Chem. 1964, 15, 155−196.
(b) Marcus, R. A. J. Phys. Chem. 1968, 72, 891−899. (c) For
applications to reactions involving intermolecular hydrogen atom
transfers, see ref 15b and Newcomb, M.; Manek, M. B.; Glenn, A. G. J.
Am. Chem. Soc. 1991, 113, 949−958. For an application to radical
cyclizations, see: (d) Alabugin, I. V.; Manoharan, M. J. Am. Chem. Soc.
2005, 127, 12583−12594. (e) Alabugin, I. V.; Timokhin, V. I.; Abrams,
J. N.; Manoharan, M.; Abrams, R.; Ghiviriga, I. J. Am. Chem. Soc. 2008,
130, 10984−10995.
(27) For calculations referring to hydrogen abstraction by imide-N-
oxyl radicals, see: Arnaud, R.; Milet, A.; Adamo, C.; Einhorn, C.;
Einhorn, J. J. Chem. Soc., Perkin Trans. 2 2002, 1967−1972 and
references cited therein.
(28) The absolute configurations, resulting from retention of the
configuration of the starting stereocenter, were assigned by analogy
with our previous results; see ref 1.
(29) Stereochemistry was assigned from a NOESY experiment which
showed cross-peaks between the proton α to the tosyl group at 5.95
ppm and the protons of the methyl group at 2.55 ppm.
(30) Effectively, the enantiomeric excess of naphthoazepine (2R,3S)-
33 formed in acetonitrile was only 65% instead of 84% in benzene.
2783
dx.doi.org/10.1021/jo202580y | J. Org. Chem. 2012, 77, 2773−2783