Dearomative indole (3 + 2) cycloaddition reactions

Article abstract of DOI:10.1021/ja412435b

A diastereoselective (3 + 2) dearomative annulation of 3-substituted indoles with α-haloketones has been developed. Significant regiochemical control was observed. This methodology provides easy access to highly functionalized cyclopenta- or cyclohexa-fused indoline compounds, which are common structures of many natural products. The synthetic potential of this reaction was demonstrated in the concise syntheses of the core structures of vincorine, isocorymine, and aspidophylline A. DFT studies (B3LYP-D3/6- 311++G**/MeOH) on cyclization mechanisms involving the 2-hydroxyallyl cation and its deprotonated oxyallyl cation have been performed. Under the reaction conditions, with a sparingly soluble Na₂CO ₃ base, both species may be present and both pathways are viable. Both pathways support the formation of the experimentally observed O-bound intermediate, its transformation to the final product, the regiochemical and eventual stereochemical outcome of the kinetic cyclization product, and the thermodynamic preference for formation of the final stereoisomer.

Full text of DOI:10.1021/ja412435b

Article
pubs.acs.org/JACS
Dearomative Indole (3 + 2) Cycloaddition Reactions
Hui Li, Russell P. Hughes,* and Jimmy Wu*
Department of Chemistry, Dartmouth College, Hanover, New Hampshire 03755, United States
S
* Supporting Information
ABSTRACT: A diastereoselective (3 + 2) dearomative
annulation of 3-substituted indoles with α-haloketones has
been developed. Significant regiochemical control was
observed. This methodology provides easy access to highly
functionalized cyclopenta- or cyclohexa-fused indoline com-
pounds, which are common structures of many natural
products. The synthetic potential of this reaction was
demonstrated in the concise syntheses of the core structures of vincorine, isocorymine, and aspidophylline A. DFT studies
(B3LYP-D3/6-311++G**/MeOH) on cyclization mechanisms involving the 2-hydroxyallyl cation and its deprotonated oxyallyl
cation have been performed. Under the reaction conditions, with a sparingly soluble Na2CO3 base, both species may be present
and both pathways are viable. Both pathways support the formation of the experimentally observed O-bound intermediate, its
transformation to the final product, the regiochemical and eventual stereochemical outcome of the kinetic cyclization product,
and the thermodynamic preference for formation of the final stereoisomer.
Indoles have been shown to undergo (4 + 2) cycloaddition
reactions to afford hydrocarbazoles.6c,d,8 Nevertheless, only
limited reports are available on dearomative (3 + 2)
cycloaddition of indoles.9 Kerr and Pagenkopf have reported
Lewis acid catalyzed C2/C3 cyclopentannulation of indole with
1,1-cyclopropane diesters and 2-methoxy-1-cyclopropane es-
ters, respectively.9a,b Recently, Tang and co-workers developed
a copper-catalyzed enantioselective version.9c Barluenga and co-
workers have reported an enantioselective (3 + 2) cyclo-
addition reaction of indole with alkynyl Fischer carbenes,9d
while Lian and Davies established a rhodium-catalyzed variant
with vinyldiazoacetates.9e The examples above indicate that the
use of 1,3-dipoles to dearomatize the C2C3 double bond of
indole and its derivatives can be a successful strategy to
synthesize cyclopentannulated indolines.
INTRODUCTION
■
Dearomatization of indoles has been a powerful strategy for
organic chemists to access many architecturally complex
alkaloids.1 Due to the ubiquitous nature of the indole unit in
important bioactive alkaloids, numerous chemo-, regio-, and
enantioselective methodologies have emerged. Some dearoma-
tive strategies include allylation,2 alkylation,3 arylation,4
iminium catalysis,5 and cycloaddition.6 Dearomative cyclo-
addition, which is based on the reactivity of the C2C3 double
bond of indole, is an attractive and straightforward approach to
fused indoline compounds. Moreover, indolines with a fused
five- or six-membered ring at the C2 and C3 positions are well
represented in nature (Figure 1).7
1,3-Dipoles have been of particular interest to us because, for
some time, our group has been developing annulation and
alkylation reactions of indole and related heterocycles.10,11 In
particular, the oxyallyl cation dipole usually reacts with dienes
to furnish seven-membered rings.12 However, they have also
been shown to undergo simple alkylation reactions13 as well as
(3 + 2) cycloadditions with allylsilanes,14 dienes,15 furans,16 and
enamines.17 Nonetheless, little is known about its reactivity in
dearomatization processes with substituted indoles. We
envisioned that oxyallyl or hydroxyallyl cations generated in
situ from α-haloketones may undergo formal dearomative (3 +
2) cycloaddition with indoles. Herein, we describe our progress
in this subject matter.
Received: December 6, 2013
Published: April 17, 2014
Figure 1. Natural products with fused five- and six-membered rings.
© 2014 American Chemical Society
6288
dx.doi.org/10.1021/ja412435b | J. Am. Chem. Soc. 2014, 136, 6288−6296

Journal of the American Chemical Society
Article
a
Table 1. Optimization of Reaction Conditions
b
c
d
entry
base
solvent
concn (M)
temp (°C)
time (h)
yield % (brsm)
dr
f
f
e
1
Na2CO3
Na2CO3
NaOH
MeOH/H2O
DMSO
MeOH
MeOH/H2O
MeOH
TFE
0.2
0.2
0.2
0.2
0.5
0.5
1
rt
rt
rt
rt
rt
rt
rt
rt
rt
rt
24
24
24
24
24
24
24
6.5
17
11
5 (73)
N/A
e
2
3 (82)
N/A
e
3
15 (75)
14 (73)
23 (80)
71 (95)
79 (88)
48 (48)
71 (71)
51 (51)
93 (93)
N/A
e
4
NaOH
N/A
5
Na2CO3
Na2CO3
Na2CO3
Et3N
12:1
6:1
6
7
TFE
5:1
8
TFE
1
4.6:1
6.7:1
4.6:1
6:1
9
DIPEA
DMAP
Na2CO3
TFE
1
10
11
TFE
1
TFE
1
40
10
a
d
b
c
Indole (1 equiv), 2-chlorocyclopentanone (1.4 equiv), and base (1.5 equiv). Isolated yield. Yield based on recovered starting material.
e
f
1
Determined by H NMR analysis of unpurified product. Not determined. 1:1 v/v.
Table 2. Cycloaddition Reactions of 3-Substituted Indoles
with 2-Chlorocyclopentanone
RESULTS AND DISCUSSION
■
a b
,
We initiated our study by examining the reaction between N-
benzylskatole (1) and 2-chlorocyclopentanone 2 (Table 1).
Little product was obtained after 24 h using Na2CO3 in a
mixture of MeOH/H2O or DMSO (entries 1 and 2). The use
of stronger inorganic bases such as NaOH in MeOH or
MeOH/H2O did not accelerate the reaction either and still
gave low yields after 24 h (entries 3 and 4). Inspired by the
reports of Chi, MacMillan, and Harmata,12a,13 trifluoroethanol
(TFE) was investigated as a solvent. Its use substantially
increased the reaction rate (entries 5 and 6). This is likely
because, as compared to MeOH, TFE is a stronger H−bond
donor and, hence, can more effectively promote enolate
formation/enolization as well as weaken the C−Cl bond.
Higher concentrations of the reactants also accelerated the
reaction (entries 6 and 7). We then investigated three organic
bases: Et3N, DIPEA, and DMAP (entries 8, 9, and 10), which
shortened the reaction time but furnished poorer yields. The
best yield and reactivity were obtained by conducting the
reaction in TFE (1 M) at 40 °C (entry 11).
With the optimized conditions in hand, we explored the
reaction scope between 2-chlorocyclopentanone (2) and
various 3-substituted indoles 4 (Table 2). The reaction worked
well, affording good to excellent yields and diastereoselectiv-
ities. Electron-rich indoles (Table 2, entry 10) tended to
accelerate the reaction, while indoles with a relatively strong
electron-withdrawing group (5-CO2Me) appeared unreactive.
This reaction was tolerant to different C3 substituents on
indole other than methyl, thus allowing additional function-
alities such as silylether, amino, allyl, and iodo groups to be
introduced into the products (Table 2, entries 6−10, 15). Both
the hydroxyl and amino groups needed to be protected so that
they would not act as nucleophiles and interrupt the (3 + 2)
cycloaddition. That the pendant iodine in entry 15 (Table 2)
was not destroyed is a testament to the mildness of the
reaction. C2-,C3-disubstituted indoles were not compatible
with the reaction when using 2; however, acyclic chloroketones
are amenable to cyclization with indoles possessing this
substitution pattern (vide infra). We have also examined the
reactivity of N-H and N-protected indoles. For instance, N-
a
Indole (1 equiv), 2-chlorocyclopentanone (1.4 equiv) and Na2CO3
b
c
(1.5 equiv) in TFE. Reaction concentration was 1 M. Isolated yield.
d
e
Determined by 1H NMR. Total of 2.1 equiv of 2-chlorocyclo-
pentanone used.
methyl, -Bn, -allyl, and -iPr indoles worked well, but neither N-
H nor N-Cbz protected indoles were suitable substrates. For
6289
dx.doi.org/10.1021/ja412435b | J. Am. Chem. Soc. 2014, 136, 6288−6296

Journal of the American Chemical Society
Article
the products in entries 1 and 11 (Table 2), the relative
stereochemical configuration of each diastereomer was
established by NOESY experiments (Figure 2). The relative
stereochemistries of the remaining products were assigned by
analogy.
Table 3. Cycloaddition Reactions of 3-Substituted Indoles
with Acyclic α-Haloketones
a b
,
Figure 2. Selected NOESY data for diastereomers in Table 2.
Attempts to extend the scope of cyclic α-haloketone to
include 2-chlorocyclohexanone and 2-chlorocycloheptanone
were unsuccessful. The latter was unreactive to N-benzylska-
tole; the former furnished poor yields of the desired product 7
(eq 1) while affording mostly C2 alkylated product 8,
presumably arising from 1,2-migration of a C3 alkylated
intermediate followed by aromatization. When both minor
and major isomers of 7 were purified and resubjected to the
reaction conditions, fragmentation to 8 did not occur, thereby,
precluding the possibility that 8 is derived from 7.
The scope and generality of this cycloaddition in terms of
indole substitution with acyclic α-haloketones are summarized
in Table 3. Unlike the reaction with 2-chlorocyclopentanone
(2), C2-,C3-disubstituted indoles were indeed compatible to
annulation using acyclic haloketones (Table 3, entries 1 and 2).
However, longer reaction times and, in some cases, higher
temperatures (Table 3, entries 1−4) were required. The
employment of electron-rich indole 11 (Table 3, entry 4) did
not accelerate the reaction relative to entry 3. Instead, elevated
temperatures were required to drive the reaction to completion.
This is likely due to the bulkier isopropyl group of 11. These
results also show that the (3 + 2) cycloadditions proceeded
with excellent regio- and diastereoselectivities. Single re-
gioisomers were isolated for entries 1−4. Similar results were
observed for α-haloketones 14−17 (Table 3, entries 6−9),
which also furnished their respective products as single
regioisomers. Moreover, compound 23 was formed in high
diastereoselectivities relative to the substituent on the α-
haloketone fragment (>20:1 of 23:24).
a
b
Indole (1 equiv) and Na2CO3 (1.5 equiv) in TFE. Reaction
c
d
concentration was 1.0 M. Isolated yield. As determined by 1H NMR
e
f
g
spectroscopy. 50 °C. 2.0 equiv of α-haloketones 12 used. 95% brsm.
h
i
j
k
75 °C. 1.4 equiv of α-haloketones 12 used. Not determined. α-
l
m
haloketones (2.0 equiv) Na2CO3 (2.2 equiv). 58% brsm. 59% brsm.
of 23 (minor diasteromer) and 24 (major diastereomer)
occurred upon treatment with silica gel or prolonged standing
in CH2Cl2 (eq 2).
However, in contrast to results shown in eq 2, 23 was
invariably the major diastereomer isolated in the overall
reactions for entries 6−9. In order to explain these results,
we performed stability experiments on diastereomer 24. We
discovered that a mixture of 24 and NaCl in TFE did not
epimerize to 23, while such epimerization was observed upon
It is noteworthy that, for entries 6−9, a common O-alkylated
intermediate 25 was isolated and confirmed by NMR analysis,
suggesting the possibility that α-haloketones 14−17 may
proceed via the same oxyallyl or hydroxyallyl intermediate.
We also discovered that the rearrangement of 25 to a mixture
6290
dx.doi.org/10.1021/ja412435b | J. Am. Chem. Soc. 2014, 136, 6288−6296

Journal of the American Chemical Society
Article
treatment with bicarbonate or carbonate base (eq 3). We also
discovered that simply dissolving 24 in EtOAc results in its
conversion to 23 over the course of 1 day (eq 3). These data
suggest that compound 23 is likely the thermodynamically
preferred diastereomer while 24 is the kinetically favored one.
For entries 2, 3, 5−9 (Table 3) regio- and stereochemical
assignments were established by extensive NMR analyses,
including DEPT, HMBC, HMQC, and NOESY. The regio- and
stereochemistry of the products in entries 1 and 4 (Table 3)
were assigned by analogy. Further structural confirmation of 23
was established by single crystal X-ray analysis (Figure 3).
Figure 4. (a) Kinetic profile of reaction for entry 7, Table 3. (b)
Kinetic profile of reaction for entry 9, Table 3. (All reactions were
monitored by 1H-NMR spectroscopy using hexamethyl benzene as an
internal standard).
Scheme 1. Proposed Reaction Mechanism
Figure 3. X-ray structure of compound 23.
In order to gain more detailed mechanistic information
regarding these (3 + 2) cycloaddition reactions, kinetic
experiments on the reactions for entries 7 and 9 (Table 3)
were carried out. As shown in Figure 4, they both behaved
similarly with respect to the generation of minor diastereomer
24 (purple line), which was steadily formed throughout both
reactions. The most prominent difference between the kinetic
profiles is that with the use of 17, there is a significant buildup
of the O-alkylated intermediate 25 (Figure 4b; red line) during
the first 20 h, whereas, with α-haloketone 15, the level of 25
remained low (Figure 4a). We believe this difference reflects a
change in the rate-determining step (discussion to follow). A
second significant dissimilarity is that the reaction rate with 15
was much slower as compared to that of 17. This was deduced
by comparing their consumption of N-benzylskatole (1) and/or
formation of product 23.
Scheme 1 illustrates our proposed mechanism for the (3 + 2)
cycloaddition. The reduced acidity of the α-proton of 15 (i.e.,
higher pKa) as compared to that of 17 makes 15 more resistant
to enolization. Moreover, the kinetic acidity of 15 may also be
low as a result of the stereoelectronic demands required for
enolization/tautomerization (i.e., α-proton needs to be
coplanar with the carbonyl π-system). If the activation energy
of enolization for 15 were high enough to render this step rate-
determining, this would be consistent with the lack of an
observed buildup of any intermediates along the pathway
toward product when using 15.
In contrast, we believe that enolization of 17 to give 27
(kRE2) is fast relative to kRE1. This would be followed by
6291
dx.doi.org/10.1021/ja412435b | J. Am. Chem. Soc. 2014, 136, 6288−6296

Journal of the American Chemical Society
Article
a
chloride ionization (aided by H-bonding with TFE) to generate
hydroxyallyl cation syn-28-H+. While other cycloaddition
reactions are thought to proceed through enolate oxyallyl
cations,12 DFT calculations on our system suggest that a route
via hydroxyallyl cation syn-28-H+ may also be available under
the reaction conditions (vide infra). In this scenario, N-
benzylskatole (1) then attacks syn-28-H+ to form intermediate
29 which possesses a weak C····OH interaction. Harmata,
Schreiner, and co-workers have also reported a divergence in
reactivity between oxyallyl and hydroxyallyl cations.18 Removal
of the proton with the carbonate base at this point would
furnish the observed intermediate 25. Alternatively, the
protonated form 29 can redissociate and alkylate at carbon to
generate the kinetic cycloadduct 24, which can then isomerize
to the thermodynamic product 23. We believe that when using
α-haloketone 17, the overall transformation 29 → 24 and then
to 23 is rate-determining. The observed buildup of 25 is
consistent with its formation prior to the RDS when 17 is used.
As described above, diastereomer 24 is the kinetically favored
product, but, under the reaction conditions, it epimerizes to the
more thermodynamically favored isomer 23. Our present use of
a good H-bonding solvent (TFE), in combination with a weak
non-nucleophilic base (Na2CO3), to generate oxyallyl and
hydroxyallyl cations by means of soft enolization has been
previously reported by MacMillan and Chi.13
Scheme 2. Baeyer−Villiger Oxidation
a
(a) m-CPBA, NaHCO3, 33:34 (1:3), 81% combined yield; (b)
NaOH, MeOH/H2O, rt, 1 h; (c) MeI, K2CO3, DMF, 20 min, rt, 90%
over two steps.
Scheme 3. Attempted Synthesis of Malagashanine Core
Attempts to employ compound 30 as a substrate was
unsuccessful. What we obtained was alkylation at C2 (eq 4).
Transposition of the double bond supports the intermediacy of
a hydroxallyl cation species.
Applications toward Natural Products. We have also
investigated the potential utility of using the title reaction in the
syntheses of several indoline-containing alkaloids. In this
regard, we have prepared the core structures of vincorine,7f−h
isocorymine,19 and aspidophyline A (Figure 5).7e,i
an efficient oxidant in which two lactones were isolated in a 5:1
ratio and a combined yield of 84%. However, neither of the
lactones turned out to be 37. The major product was the
undesired regioisomer 39, while the other was rearranged
lactone 40. The structures of both 39 and 40 have been
confirmed by single crystal X-ray analysis.
Although Baeyer−Villiger oxidation of compound 36 failed
to provide lactone 37, we investigated potential applications for
its regioisomer 39. Scheme 4 illustrates three short syntheses of
the core structures of vincorine, isocorymine, and aspidophyl-
line A. Debenzylation of lactone 39 by hydrogenolysis and
subsequent oxidation with Dess−Martin periodinane (DMP)
led to hemiaminal 41. Cleavage of the phthalimide group with
MeNH2 liberated the free amine, which was followed by
spontaneous ring closure to install the pyrrolidine ring of
pentacycle 42. Compound 42 maps well to vincorine and
isocorymine. By employing a similar strategy, pentacycle 46 was
obtained from 43, providing a good starting point for the
synthesis of aspidophylline A.
Figure 5. Structures of vincorine, isocorymine, and aspidophylline A
(n.b., drawn as their enantiomers).
As shown in Scheme 2, ketone 32 was subjected to Baeyer−
Villiger oxidation using m-cpba to afford lactone 34 as the
major regioisomer (3:1 mixture). Hydrolysis of 34 under
alkaline conditions, followed by esterification, furnished
cyclohexa-fused indoline 35.
We hypothesized that an analogous regioselective oxidation
of 36 could afford lactone 37, a promising precursor to
malagashanine (Scheme 3). We first examined the oxidation of
36 with m-cpba in THF at rt. The reaction turned out to be
very sluggish. Other oxidants such as H2O2 or t-BuOOH did
not perform well either. Finally, we identified peracetic acid as
Computational Studies. DFT studies were carried out
using the B3LYP-D3 functional20 and the 6-311++G** basis
set,21 as implemented in the Jaguar22 suite of programs. This
6292
dx.doi.org/10.1021/ja412435b | J. Am. Chem. Soc. 2014, 136, 6288−6296

Journal of the American Chemical Society
Article
(CO32−). Free energies of the optimized syn- and anti-isomers
of 28 and 28-H+ were calculated in a methanol solvent, and
relative values are shown in Figure 7. As expected, the syn-
Scheme 4. Synthesis of Core Structures for Vincorine,
Isocorymine, and Aspidophylline A
a
a
(a) Pd(OH)2, EtOAc, H2, rt, overnight; (b) DMP, CH2Cl2, rt, 0.5 h,
82% for two steps; (c) MeNH2, MeOH, rt, overnight, 81%; (d)
CH3CO3H, NaHCO3, THF, rt, overnight, 61%; (e) Pd/C, EtOAc, H2,
rt, overnight; (f) DMP, CH2Cl2, rt, 1 h, 90% for two steps; (g) p-TSA
(cat.), toluene, reflux, 3 h, 99%.
Figure 7. DFT calculated (B3LYP-D3/6-311++G**/CH3OH)
structures for syn and anti isomers of hydroxyallyl cation 28-H+ and
their deprotonation by solvent (MeOH) or base (CO32−) to give syn
and anti isomers of oxyallyl cation 28. Relative free energies (kcal/
mol) are in red.
dispersion corrected functional23 has proven to be essential in
modeling other cyclization reactions of indoles and especially in
assessing activation parameters for similar two-component
additions for which intermolecular dispersion interactions are
crucial components of the overall transition state energetics.24
A Poisson−Boltzmann solvation model25 as implemented in
Jaguar was applied using methanol as the solvent. The chosen
DFT method and basis perform well in predicting structures of
this type; a superimposition of the calculated and X-ray
structures of 23 is shown in Figure 6.
isomer is, in each case, significantly more stable than its anti-
analogue. Comparison of the ΔG values for deprotonation of
syn- and anti-28-H+ to give their deprotonated analogues 28
2−
were made using solvent (MeOH) or solubilized CO3 as
bases; the computationally derived absolute free energy of the
proton in methanol (−263.5 kcal/mol) was used for the former
calculation.27 Not suprisingly, loss of a proton from syn-28-H+
to the solvent is uphill by 11.8 kcal/mol, while loss of a proton
to CO32− to give HCO3− is strongly downhill by 18.2 kcal/mol.
Attempts to locate transition states for deprotonations and
protonations of these species and others shown later (Figure 8;
INT1-H+ → 25, INT3-H+ → 24) were unsuccessful since the
kinetic barrier to deprotonation should be very small.
Deprotonation of 28-H+ by solvent is unfavorable; however,
if stoichiometric CO32− were available in solution, only 28 can be
present. The solubility of Na2CO3 in anhydrous MeOH is
known to be negligible,28 and our own measurements in TFE
(shake-flask method) indicate a solubility of approximately 0.05
M. Compared to the concentration of chloroketone in the
reaction (1.4 M), nearly 4% of base may be available in solution
at a given time, raising the possibility that either 28 or 28-H+
may serve as reactive species. Because the kinetics of both the
formation of 28-H+ and its deprotonation to give 28 in the
presence of a mostly heterogeneous base is unknown, it seemed
prudent to consider the intermediacy of both syn-28-H+ and
syn-28.
Figure 6. Superimposition of X-ray structure (bronze) of compound
23 with that calculated by DFT (B3LYP-D3/6-311++G**) for its N-
methyl analogue (blue). H’s are omitted for clarity.
The key intermediate derived from either haloketone must
be the oxyallyl cation syn/anti-28 or its O-protonated form syn/
anti-28-H+. We recognize the option that 28 might also be a
singlet diradical, as discussed for the unsubstituted parent
system.26 Our own calculations (UDFT/B3LYP-D3/6-311+
+G**) confirm that the parent species does converge to a
singlet diradical structure. However, 28 invariably converges to
a closed-shell structure, which was used as the basis for further
calculations. We have also attempted to address whether 28 or
its protonated form 28-H+ is present under the reaction
conditions (vide infra). Deprotonation of 28-H+ requires a
base that could be either the solvent or the added base
First, the reaction of oxyallyl cation syn-28 with N-
methylskatole was examined using DFT. Two pathways were
found as shown in Figure 8. The first, shown in black, allows
6293
dx.doi.org/10.1021/ja412435b | J. Am. Chem. Soc. 2014, 136, 6288−6296

Journal of the American Chemical Society
Article
Figure 8. DFT calculated (B3LYP-D3/6-311++G**/CH3OH) free energy profiles for reaction of N-methylskatole with oxyallyl cation syn-28 (black
and red pathways) and hydroxyallyl cation syn-28-H+ (blue pathway), and calculated structures of transition states, intermediates, and products. Free
energy differences are in kcal/mol, and distances are in Å. Numbers in green are free energy differences between the indicated protonated species
−
plus CO32− and the corresponding deprotonated species plus HCO3 . Other than the two C−H bonds that define the stereochemistry of formation
of the second C−C bond, all other C−H hydrogen atoms are omitted for clarity.
6294
dx.doi.org/10.1021/ja412435b | J. Am. Chem. Soc. 2014, 136, 6288−6296

Journal of the American Chemical Society
Article
direct formation of 24 via a concerted but very unsymmetrical
TS1, and the second, shown in red, involves stepwise formation
of O-bound intermediate 25 via TS2, with subsequent
rearrangement to product 24 via TS3. While both pathways
involve high-energy transition states, the latter is the lower
energy path, consistent with experimental observation of O-
bound intermediate 25. Transformation of 25 to 24 occurs via
TS3, in which dissociation of the C−O bond in 25 is
accompanied by directional rotation as shown by the curved
green arrow to afford the relative stereochemical configuration
obtained during formation of the second C−C bond in 24.
Final epimerization of 24 to 23 is predicted to be downhill by
2.6 kcal/mol, consistent with the observed thermodynamics of
the overall reaction. For completeness, the analogous pathway
for regioisomeric addition of syn-28 to N-methylskatole was
located; the corresponding energy of this TS2 lies 5.7 kcal/mol
higher than that shown in Figure 8. DFT correctly predicts
formation of the O-bound intermediate and both the
regiochemistry and stereochemistry of addition, starting from
oxyallyl cation syn-28. However, earlier concerns about the
insolubility of the added base raised the possibility that the
reaction may flow not from syn-28 but rather from its
protonated analogue syn-28-H+.
Thus, while a pathway for formation of product 24 from
oxyallyl cation syn-28 is available, our DFT calculations predict
that it is a higher energy route than that emanating from the
hydroxyallyl cation syn-28-H+. As mentioned above, divergent
reactivity between oxyallyl and hydroxyallyl cations was also
previously observed by Harmata and Schreiner.18 It seems likely
that unless oxyallyl cations are generated by stoichiometric,
homogeneous, strong bases, the participation of hydroxyallyl cations
as competitive reagents cannot be automatically excluded.
CONCLUSION
■
The dearomatization of the C2/C3 double bond of 3-
substituted indoles with α-haloketones has been reported.
Both high efficiency and high diastereocontrol were observed in
the majority of cases. DFT calculations suggest that the
preferred mechanism for the formal cycloaddition may proceed
via hydroxyallyl cations rather than the corresponding oxyallyl
cations. O-Alkylated intermediates are initially formed, followed
by isomerization to the observed products. The synthetic
potential of this dearomatization process was demonstrated by
concise syntheses of the core structures of vincorine,
isocorymine, and aspidophylline A. With an eye toward
targeting malagashanine, efforts are ongoing in our laboratory
to obtain regioisomeric lactone 37.
To evaluate this possibility using DFT, reaction of the
protonated hydroxyallyl cation syn-28-H+ with N-methylska-
tole was also examined, and the resultant lowest energy reaction
profile is shown as the blue pathway in Figure 8. Initial reaction
of syn-28-H+ with N-methylskatole proceeds via TS1-H+ to
give INT1-H+ (identical to 29 in Scheme 1) in which the first
C−C bond is formed and there is a long, presumably weak,
C····OH interaction. At this stage, kinetically facile loss of a
proton (vide supra) can afford the experimentally observed
intermediate 25, so it is perhaps not surprising that, in the
presence of excess Na2CO3, a buildup of 25 is observed. It is
notable that the activation free energy for formation of
protonated 25 (INT1-H+) from the hydroxyallyl cation syn-
28-H+ and N-methylskatole is only 24.1 kcal/mol, which is
significantly lower than the 31.0 kcal/mol required for the
direct formation of 25 from oxyallyl cation syn-28 (Figure 8).
Consequently, if syn-28 and syn-28-H+ are both present in
methanol under these conditions, the reaction of syn-28-H+
with the indole is predicted to be faster. In addition, the
subsequent reaction of INT1-H+ to give INT3-H+, which is
the O-protonated form of 24, also has a considerably lower
overall activation free energy than does the unprotonated
analogue (Figure 8). Breaking the C−O bond in INT1-H+ is
facilitated by protonation, allowing dissociation via a low energy
TS2-H+ to give INT2-H+. Continued elongation of the C−O
distance, with directional rotation as shown by the curved green
arrow (Figure 8) via TS3-H+ leads to INT-3H+ with the
correct relative stereochemistry resulting from construction of
the second C−C bond. Transition states for direct formation of
either species INT2-H+ or INT3-H+ from starting materials
could not be located.
GENERAL PROCEDURE
■
To a TFE (0.18 mL) solution of 1 (40 mg, 0.18 mmol) and 2 (30 mg,
0.25 mmol) was added Na2CO3 (28 mg, 0.27 mmol). The
heterogeneous reaction mixture was stirred at 40 °C until the reaction
was judged complete as determined by thin layer chromatographic
analysis. The mixture was filtered through a short pad of Celite and
washed with 5 mL of EtOAc. The filtrate was concentrated in vacuo,
and the residue was purified via silica gel column chromatography (3%
EtOAc/hexane) to afford 3 as a colorless solid.
ASSOCIATED CONTENT
■
S
* Supporting Information
Experimental details and characterization of all new com-
pounds, CIF files, selected NOESY, HMBC, HMQC, DEPT
data, details of DFT calculations, final geometry coordinates for
all calculated structures, and larger scale graphics of Figures
7−8. This material is available free of charge via the Internet at
http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Authors
jimmy.wu@dartmouth.edu
russell.p.hughes@dartmouth.edu
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
2−
Proton loss from the reaction of INT3-H+ with CO3 is
Dedicated to our colleagues Professor David M. Lemal and
Professor Thomas A. Spencer in celebration of their 80th
birthdays. This work was supported in part by Dartmouth
College. We thank Professors David M. Lemal, Peter A. Jacobi,
and Gordon W. Gribble for useful discussions. We also thank
Dr. Richard J. Staples (Michigan State University) for
determination of X-ray structures and Wayne Casey (Dart-
mouth College) for assistance with NMR spectroscopy.
strongly downhill to give 24, so consumption of protons by
Na2CO3 appears to be essential to drive the overall reaction by
this pathway. It is also noteworthy that the blue pathway shown
in Figure 8 is proton-catalyzed and that inefficient or reversible
consumption of protons by an added base may still allow this
pathway to occur even if only small concentrations of syn-28-H
+ are present.
6295
dx.doi.org/10.1021/ja412435b | J. Am. Chem. Soc. 2014, 136, 6288−6296

Journal of the American Chemical Society
Article
(18) Harmata, M.; Huang, C.; Rooshenas, P.; Schreiner, P. R. Angew.
Chem., Int. Ed. 2008, 47, 8696−8699.
(19) (a) Bevan, C. W. L.; Patel, M. B.; Rees, A. H.; Harris, D. R.;
Marshak, M. L.; Mills, H. H. Chem. Ind. 1965, 14, 603−604.
(b) Vercauteren, J.; Massiot, G.; Le Men-Olivier, L.; Levy, J.; Delaude,
C. Bull. Soc. Chim. Fr. 1982, 9−10 (pt. 2), 291−296.
(20) (a) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785−
789. (b) Becke, A. D. J. Chem. Phys. 1993, 98, 5648−5652. (c) Becke,
A. D. J. Chem. Phys. 1993, 98, 1372−1377. (d) Stephens, P. J.; Devlin,
F. J.; Chabalowski, C. F.; Frisch, M. J. J. Phys. Chem. 1994, 98, 11623−
11627. (e) Andersson, M. P.; Uvdal, P. J. Phys. Chem. A 2005, 109,
2937−2941.
REFERENCES
■
(1) Roche, S. P.; Porco, J. A., Jr. Angew. Chem., Int. Ed. 2011, 50,
4068−4093.
(2) (a) Wu, Q. F.; Zheng, C.; You, S. L. Angew. Chem., Int. Ed. 2012,
51, 1680−1683. (b) Kimura, M.; Futamata, M.; Mukai, R.; Tamaru, Y.
J. Am. Chem. Soc. 2005, 127, 4592−4593. (c) Trost, B. M.; Quancard,
J. J. Am. Chem. Soc. 2006, 128, 6314−6315. (d) Wu, Q. F.; He, H.; Liu,
W. B.; You, S. L. J. Am. Chem. Soc. 2010, 132, 11418−11419.
(e) Kagawa, N.; Malerich, J. P.; Rawal, V. H. Org. Lett. 2008, 10,
2381−2384.
(3) Lin, A.; Yang, J.; Hashim, M. Org. Lett. 2013, 15, 1950−1953.
(4) Wu, K.; Dai, L. X.; You, S. L. Org. Lett. 2012, 14, 3772−3775.
(5) (a) Austin, J. F.; Kim, S.-G.; Sinz, C. J.; Xiao, W.-J.; MacMillan, D.
W. C. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 5482−5487. (b) Jones, S.
B.; Simmons, B.; MacMillan, D. W. C. J. Am. Chem. Soc. 2009, 131,
13606−13607. (c) Horning, B. D.; MacMillan, D. W. C. J. Am. Chem.
Soc. 2013, 135, 6442−6445.
(6) (a) Zhang, L. J. Am. Chem. Soc. 2005, 127, 16804−16805.
(b) Repka, L. M.; Ni, J.; Reisman, S. E. J. Am. Chem. Soc. 2010, 132,
14418−14420. (c) Kawano, M.; Kiuchi, T.; Negishi, S.; Tanaka, H.;
Hoshikawa, T.; Matsuo, J.; Ishibashi, H. Angew. Chem., Int. Ed. 2013,
52, 906−910. (d) Zhang, G.; Huang, X.; Li, G.; Zhang, L. J. Am. Chem.
Soc. 2008, 130, 1814−1815. (e) Martin, D.; Vanderwal, C. D. J. Am.
Chem. Soc. 2009, 131, 3472−3473. (f) Robertson, F. J.; Kenimer, B.
D.; Wu, J. Tetrahedron 2011, 4327−4332.
(7) (a) Jones, S. B.; Simmons, B.; Mastracchio, A.; MacMillan, D. W.
C. Nature 2011, 475, 183−188. (b) Magnus, P.; Gallagher, T.; Brown,
P.; Huffman, J. C. J. Am. Chem. Soc. 1984, 106, 2105−2114.
(c) Kuehne, M. E.; Seaton, P. J. J. Org. Chem. 1985, 50, 4790−4796.
(d) Wright, C. W. Nat. Prod. Rep. 2010, 27, 961−968. (e) Zu, L.; Boal,
B. W.; Garg, N. K. J. Am. Chem. Soc. 2011, 133, 8877−9. (f) Zhang,
M.; Huang, X.; Shen, L.; Qin, Y. J. Am. Chem. Soc. 2009, 131, 6013−
6020. (g) Zi, W.; Xie, W.; Ma, D. J. Am. Chem. Soc. 2012, 134, 9126−
9129. (h) Horning, B. D.; MacMillan, D. W. C. J. Am. Chem. Soc. 2013,
135, 6442−6445. (i) Li, Q.; Li, G.; Ma, S.; Feng, P.; Shi, Y. Org. Lett.
2013, 15, 2601−2603. (j) Hugel, G.; Cartier, D.; Levy, J. Tetrahedron
Lett. 1989, 30, 4513−4516. (k) Dethe, D. H.; Erande, R. D.; Ranjan, A.
J. Org. Chem. 2013, 78, 10106−10120.
(21) (a) Dunning, T. H.; Hay, P. J. In Modern Theoretical Chemistry,
Vol. 4: Applications of Electronic Structure Theory; Schaefer, H. F., III,
Ed.; Plenum: NY, 1977. (b) Hay, P. J.; Wadt, W. R. J. Chem. Phys.
1985, 82, 270−283. (c) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985,
82, 299−310. (d) Wadt, W. R.; Hay, P. J. J. Chem. Phys. 1985, 82,
284−298.
(22) Jaguar, versions 7.0−7.9, Schrodinger, L. L. C., New York, NY,
̈
2007−2012.
(23) (a) Grimme, S.; Antony, J.; Ehrlich, S.; Krieg, H. J. Chem. Phys.
2010, 132, 154104. (b) Goerigk, L.; Grimme, S. Phys. Chem. Chem.
Phys. 2011, 13, 6670−6688.
(24) Lopchuk, J. M.; Hughes, R. P.; Gribble, G. W. Org. Lett. 2013,
15, 5218−5221.
(25) (a) Tannor, D. J.; Marten, B.; Murphy, R.; Friesner, R. A.;
Sitkoff, D.; Nicholls, A.; Honig, B.; Ringnalda, M.; Goddard, W. A. J.
Am. Chem. Soc. 1994, 116, 11875−11882. (b) Marten, B.; Kim, K.;
Cortis, C.; Friesner, R. A.; Murphy, R. B.; Ringnalda, M. N.; Sitkoff,
D.; Honig, B. J. Phys. Chem. 1996, 100, 11775−11788.
(26) Bettinger, H. F. Angew. Chem., Int. Ed. 2010, 49, 670−671.
(27) Kelly, C. P.; Cramer, C. J.; Truhlar, D. G. J. Phys. Chem. B 2007,
111, 408−422.
(28) Ellingboe, J. L.; Runnels, J. H. J. Chem. Eng. Data 1966, 11,
323−324.
(8) (a) Wenkert, E.; Moeller, P. D.; Piettre, S. R. J. Am. Chem. Soc.
1988, 110, 7188−7194. (b) Cai, Q.; You, S. L. Org. Lett. 2012, 14,
3040−3043. (c) Biolatto, B.; Kneeteman, M.; Paredes, E.; Mancini, P.
M. E. J. Org. Chem. 2001, 66, 3906−3912. (e) Hsieh, M.; Rao, P. D.;
Liao, C. C. Chem. Comm 1999, 1441−1442.
(9) (a) England, D. B.; Kuss, T. D. O.; Keddy, R. G.; Kerr, M. A. J.
Org. Chem. 2001, 66, 4704−4709. (b) Bajtos, B.; Yu, M.; Zhao, H.;
Pagenkopf, B. L. J. Am. Chem. Soc. 2007, 129, 9631−9634. (c) Xiong,
H.; Xu, H.; Liao, S.; Xie, Z.; Tang, Y. J. Am. Chem. Soc. 2013, 135,
7851−7854. (d) Barluenga, J.; Tudela, E.; Ballesteros, A.; Tomas, M. J.
Am. Chem. Soc. 2009, 131, 2096−2097. (e) Lian, Y.; Davies, H. M. L. J.
Am. Chem. Soc. 2010, 132, 440−441.
(10) Han, X.; Li, H.; Hughes, R. P.; Wu, J. Angew, Chem. Int. Ed.
2012, 51, 10390−10393.
(11) Han, X.; Wu, J. Angew. Chem., Int. Ed. 2013, 52, 4637−4640.
(12) (a) Harmata, M. Chem. Commun. 2010, 46, 8886−8903.
(b) Harmata, M. Chem. Commun. 2010, 46, 8904−8922. (c) Harmata,
M. Adv. Synth. Catal. 2006, 348, 2297−2306. (d) Lohse, A. G.; Hsung,
R. P. Chem.Eur. J. 2011, 17, 3812−3822.
(13) (a) Tang, Q.; Chen, X.; Tiwari, B.; Chi, Y. Org. Lett. 2012, 14,
1922−1925. (b) Vander Wal, M. N.; Dilger, A. K.; MacMillan, D. W.
Chem. Sci. 2013, 4, 3075−3079.
(14) Giese, S.; Kastrup, L.; Stiens, D.; West, F. G. Angew. Chem., Int.
Ed. 2000, 39, 1970−1973.
(15) Marx, V. M.; Burnell, D. J. J. Am. Chem. Soc. 2010, 132, 1685−
1689.
(16) Fujita, M.; Oshima, M.; Okuno, S.; Sugimura, T.; Okuyama, T.
Org. Lett. 2006, 8, 4113−4116.
(17) (a) Hayakawa, Y.; Yokoyama, K.; Noyori, R. J. Am. Chem. Soc.
1978, 100, 1799−1806. (b) Corriu, R. J. P.; Moreau, J. J. E.; Pataud-
Sat, M. J. Org. Chem. 1990, 55, 2878−2884.
6296
dx.doi.org/10.1021/ja412435b | J. Am. Chem. Soc. 2014, 136, 6288−6296