Unexpected 5,6,7,8,9,10-hexahydro-6,6-pentamethylenephenanthridines and 2,3,4,5-tetrahydro-4,4-tetramethylene-1 H -cyclopenta[ c ]quinolines from skraup-doebner-von miller quinoline synthesis and their implications for the mechanism of that reaction

Article abstract of DOI:10.1021/jo202681r

The real mechanism of the Skraup-Doebner-Von Miller quinoline synthesis remains controversial and not well understood despite several mechanistic studies reported on the matter. A series of unexpected and unusual 5,6,7,8,9,10-hexahydro-6,6-pentamethylenephenanthridines and 2,3,4,5-tetrahydro-4,4-tetramethylene-1H-cyclopenta[c]quinolines have been obtained through the Skraup-Doebner-Von Miller quinoline synthesis. On the basis of these unexpected results and in agreement with some of the previously reported quinoline syntheses, an alternative mechanistic pathway is proposed for this variant of the reaction. It involves the formation of a Schiff base through a reaction between the ketone and the aniline derivative in the first step, followed by a cycloalkenylation at the ortho-position to the amine functional group of the aniline derivative, and an annulation in the final step to close the quinoline ring, leading to a dihydroquinoline derivative. To the best of our knowledge, this is the first report of such a mechanistic pathway being proposed for any variant of the Skraup-Doebner-Von Miller quinoline synthesis.

Full text of DOI:10.1021/jo202681r

Article
pubs.acs.org/joc
Unexpected 5,6,7,8,9,10-Hexahydro-6,6-
pentamethylenephenanthridines and 2,3,4,5-Tetrahydro-4,4-
tetramethylene-1H-cyclopenta[c]quinolines from Skraup−Doebner−
Von Miller Quinoline Synthesis and Their Implications for the
Mechanism of That Reaction
Jean Fotie,*,† Hilaire V. Kemami Wangun,‡ Frank R. Fronczek,§ Nancy Massawe,† Bijay T. Bhattarai,†
Jessica L. Rhodus,† Thomas A. Singleton,⊥ and D. Scott Bohle⊥
†Department of Chemistry and Physics, Southeastern Louisiana University, SLU 10878, Hammond, Louisiana 70402-0878, United
States
‡Harbor Branch Oceanographic Institute, Centre of Marine Biomedical and Biotechnology Research, Florida Atlantic University, 5600
US 1 North, Fort Pierce, Florida 34946, United States
§Department of Chemistry, Louisiana State University, Baton Rouge, Louisiana 70803-1804, United States
⊥Department of Chemistry, McGill University, 801 Sherbrooke Street West, Montreal, Quebec H3A 0B8, Canada
S
* Supporting Information
ABSTRACT: The real mechanism of the Skraup−Doebner−
Von Miller quinoline synthesis remains controversial and not
well understood despite several mechanistic studies reported
on the matter. A series of unexpected and unusual 5,6,7,8,9,10-
hexahydro-6,6-pentamethylenephenanthridines and 2,3,4,5-
tetrahydro-4,4-tetramethylene-1H-cyclopenta[c]quinolines
have been obtained through the Skraup−Doebner−Von Miller
quinoline synthesis. On the basis of these unexpected results
and in agreement with some of the previously reported
quinoline syntheses, an alternative mechanistic pathway is
proposed for this variant of the reaction. It involves the formation of a Schiff base through a reaction between the ketone and the
aniline derivative in the first step, followed by a cycloalkenylation at the ortho-position to the amine functional group of the
aniline derivative, and an annulation in the final step to close the quinoline ring, leading to a dihydroquinoline derivative. To the
best of our knowledge, this is the first report of such a mechanistic pathway being proposed for any variant of the Skraup−
Doebner−Von Miller quinoline synthesis.
nitroethane, aniline, and glycerol in the presence of
concentrated sulfuric acid resulted in the formation of quinoline
in a very low yield. Doebner and Von Miller later reported that
the use of aniline and an α,β-unsaturated ketone (mesityl
oxide) or 2 equiv of acetone in the presence of a catalytic
amount of iodine or acid resulted in 1,2-dihydro-2,2,4-
trimethylquinoline.9 Since then, several modifications and
optimizations of the Skraup reaction using varieties of catalysts
have been reported.10 Despite the intensive research in the area,
the detailed mechanism of this family of reactions remains
controversial. Skraup first proposed that the reaction goes
through an aldehyde anil intermediate, which undergoes acid-
catalyzed annualation to quinoline.8a It was later shown that
anils cannot undergo direct closure,11 and the mechanism was
dismissed. Today’s most accepted mechanisms involve a series
INTRODUCTION
■
Quinolines and dihydroquinolines are found as important
components of naturally occurring and synthetic compounds
clinically used for the treatment of infectious diseases and in
synthetic dyes. In fact, quinolines such as quinine and
chloroquines are well-known for their antimalarial activity,1
while dihydroquinolines are known to display, among other
biological activities, antioxidant,2,3 anti-inflammatory,4 and
hormone receptor modulating5 properties. Furthermore,
ethoxyquin (6-ethoxy-1,2-dihydro-2,2,4-trimethylquinoline) is
an FDA-approved antioxidant commonly used as a preservative
in the food processing industry.6 We have recently discovered
that some dihydroquinolines also display outstanding anti-
trypanosomal activities.7
In view of their numerous applications, several methods of
synthesis enabling an easy access to this family of compounds
have been developed over the years. Around 1880, Skraup
reported in a series of papers8 that heating a mixture of
Received: December 29, 2011
Published: February 15, 2012
© 2012 American Chemical Society
2784
dx.doi.org/10.1021/jo202681r | J. Org. Chem. 2012, 77, 2784−2790

The Journal of Organic Chemistry
Article
of fragmentations and recombinations,12 in which aldehyde
anils must first undergo a rearrangement through 1,3-
diazetidinium ions. However, Denmark and Venkatraman,13
when investigating these mechanistic pathways using 13C-
labeled ketones in crossover experiments, found that none of
these mechanisms were complete on their own. They
concluded that this family of reactions follows a complex
mechanistic pathway.13 More importantly, they suggested that
the mechanism of these reactions may actually change
depending upon the starting material used, and that substituted
ketones such as pulegone may not actually follow the
“scrambling” mechanism.13
polymerize very quickly, generating oligomers (di-, tri-, and
tetramers) as observed when monitoring the reaction by GC−
MS for two hours. As a consequence, 2-cyclohexenylcyclohex-
anone (3; Figure 2) was always present in appreciable amount
Figure 2. Structure of 2-cyclohexenylcyclohexanone (3) and 2-
cyclopentylidenecyclopentanone (4) obtained from iodine-catalyzed
polymerization of cyclohexanone and cyclopentanone, respectively.
As part of our continuing effort to optimize the
antitrypanosomal activity of dihydroquinolines,7 a series of
closely related 5,6,7,8,9,10-hexahydro-6,6-pentamethylenephe-
nanthridines (1; Figure 1) were designed, and in an attempt to
each time cyclohexanone was used as starting material, while 2-
cyclopentylidenecyclopentanone (4; Figure 2) was produced
each time cyclopentanone was used as a starting material
throughout this entire investigation. The trimers and tetramers
were present only in a very limited amount (about 5−10%), as
observed in the GC−MS spectra. It is important to notice that
cyclopentanone dimerizes to yield predominantly an α,β-
unsaturated ketone, while cycloclohexanone dimerizes to
produce mainly a β,γ-unsaturated ketone. The difference in
the geometry of these two starting materials may be the key
factor behind the observed difference in reactivity.
Considering our unsuccessful attempt, and with the fact that
cyclohexanone and cyclopentanone have proven to polymerize
very quickly when heated in the presence of iodine, and since in
our previous studies 0.05 equiv of iodine as catalyst was found
to be the optimal conditions for the preparation of
dihydroquinolines through the Skraup−Doebner−Von Miller
reaction,7 we decided to modify Mammen et al.’s14 reaction
conditions. In addition to using just a catalytic amount of iodine
instead of 12 equiv, an excess (5 equiv) of ketones was also
used in order to make up for the self-polymerization. In these
conditions, 4-chloroaniline, 4-ethoxyaniline, or aniline in the
presence of 5 equiv of cyclohexanone or cyclopentanone at 165
°C for 72 h did not produce the expected products. Instead,
surprising cycloakenylated derivatives of the expected com-
pounds were systematically obtained in a good yield (Scheme
1).
In fact, 4-chloroaniline in the presence of cyclohexanone in
these conditions yielded 2-chloro-4-cyclohexenyl-5,6,7,8,9,10-
hexahydro-6,6-pentamethylenephenanthridine (5a, 59%), while
in the presence of cyclopentanone, it yielded 8-chloro-6-
cyclopentenyl-2,3,4,5-tetrahydro-4,4-tetramethylene-1H-
cyclopenta[c]quinoline (6a, 58%). In the same conditions, 4-
ethoxyaniline in the presence of cyclohexanone yielded 4-
cyclohexenyl-2-ethoxy-5,6,7,8,9,10-hexahydro-6,6-pentamethy-
lenephenanthridine (5b, 53%), while aniline in the presence of
cyclopentanone, yielded 6-cyclopentenyl-2,3,4,5-tetrahydro-4,4-
tetramethylene-1H-cyclopenta[c]quinoline (6b, 46%). More
interestingly, aniline in the presence of cyclohexanone yielded
2,4-dicyclohexenyl-5,6,7,8,9,10-hexahydro-6,6-pentamethylene-
phenanthridine (7, 63%), a derivative with cyclohexenyl groups
at the ortho and para positions to the nitrogen atom of the
dihydroquinoline ring system (Scheme 2). In all these cases, an
increase in the amount of ketones did not have any impact on
the yield of the reaction. Nevertheless, the increase in the
amount of iodine appeared to exacerbate the polymerization of
these cyclic ketones as observed in the GC−MS spectra.
Figure 1. Basic structure of 5,6,7,8,9,10-hexahydro-6,6-pentamethyle-
nephenanthridines (1) and 2,3,4,5-tetrahydro-4,4-tetramethylene-1H-
cyclopenta[c]quinolines (2), with the numbering used in the naming
of prepared derivatives.
prepare these compounds through the Skraup−Doebner−Von
Miller reaction, some rather unexpected cycloalkenylated
derivatives of the target molecules were obtained. These
compounds appeared to be systematically cycloalkenylated at
either the 2 or the 4 positions or at both positions. These
unexpected results prompted us to extend the investigation to
2,3,4,5-tetrahydro-4,4-tetramethylene-1H-cyclopenta[c]-
quinolines (2; Figure 1). In light of these new and unexpected
results, we proposed an alternative and straightforward
mechanistic pathway for the Skraup−Doebner−Von Miller
quinoline synthesis, at least as far as cyclic ketones are
concerned.
RESULTS AND DISCUSSION
■
In 1985, Mammen et al.14 reported that they were able to
prepare a series of 5,6,7,8,9,10-hexahydro-6,6-pentamethylene-
phenanthridines by heating a mixture of aniline derivatives and
2 equiv of cyclohexanone in the presence of 12 equiv of iodine.
These compounds can be considered as close derivatives of 1,2-
dihydroquinolines with the only difference that instead of using
an open-chain ketone in their Skraup−Doebner−Von Miller
quinoline synthesis, they used cyclohexanone. Thus, as part of
our investigation on the optimization of the antitrypanosomal
activity of 1,2-dihydroquinolines,7 we designed some of these
derivatives. When following the same reaction conditions as
Mammen et al.14 using 4-chloroaniline as the starting material,
we obtained mainly a clear sticky oil that was characterized as 2-
cyclohexenylcyclohexanone (3), the starting material (4-
chloroaniline), and a trace of an unknown material for which
GC−MS data did not match the expected quinoline derivative.
Curious about the origin of 2-cyclohexenylcyclohexanone, we
decided to heat up a mixture of cyclohexanone or cyclo-
pentanone in the presence of a catalytic amount of iodine (5%
mol). Under these conditions, these two compounds proved to
2785
dx.doi.org/10.1021/jo202681r | J. Org. Chem. 2012, 77, 2784−2790

The Journal of Organic Chemistry
Article
Scheme 1. Reactions between 4-Chloroaniline, 4-Ethoxyaniline, or Aniline and Cyclohexanone or Cyclopentanone
Scheme 2. Reaction between Aniline and Cyclohexanone
Scheme 4. Retrosynthesis of Compounds 8 and 9
While investigating whether iodine can actually catalyze a
cycloalkenylation on an aromatic ring system, ethoxyquin (6-
ethoxy-1,2-dihydro-2,2,4-trimethylquinoline) was submitted to
the same reaction conditions. In fact, ethoxyquin reacts with
cyclohexanone and cyclopentanone in the presence of a
catalytic amount of iodine to yield 8-cyclohexenyl-6-ethoxy-
1,2-dihydro-2,2,4-trimethylquinoline (10, 54%) and 8-cyclo-
pentenyl-6-ethoxy-1,2-dihydro-2,2,4-trimethylquinoline (11,
52%), respectively, as shown in Scheme 5.
Furthermore, 2-chloroaniline in which one of the ortho-
positions to the nitrogen atom is already substituted, in the
same conditions, yielded mainly the noncycloalkenylated
derivative (Scheme 3), although some traces of the para-
Scheme 3. Reaction between 2-Chloroaniline and
Cyclohexanone or Cyclopentanone
Scheme 5. Reaction between Ethoxyquin and
Cyclohexanone or Cyclopentanone
cycloalkenylated (para to the nitrogen atom) derivatives were
detected in the GC−MS spectra.15
In fact, 2-chloroaniline in the presence of cyclohexanone
yielded mainly 4-chloro-5,6,7,8,9,10-hexahydro-6,6-pentam-
ethylene-phenanthridine (8, 38%), while in the presence of
cyclopentanone, it yielded mainly 6-chloro-2,3,4,5-tetrahydro-
4,4-tetramethylene-1H-cyclopenta[c]quinolines (9, 35%). All
these compounds were characterized based on their NMR (1H,
13C, COSY, gHSQC, and HMBC) and HRESI-MS spectra. The
structures of compounds 5a, 5b, and 7 were unambiguously
determined by single crystal X-ray diffraction, and the obtained
molecular structures of these compounds are shown in the
Supporting Information.
At this stage of the investigation, it was obvious that in
addition to the usual Skraup−Doebner−Von Miller quinoline
synthesis, a regioselective (ortho/para to the amine of the
dihydroquinoline ring) cycloalkanylation was taking place.
More importantly, this latter reaction sequence seems to be
playing an important role in the Skraup−Doebner−Von Miller
reaction since, by simple retrosynthesis (Scheme 4), it could be
observed that at least one of the ortho-positions to the amino
function of the aniline derivative used as starting material
appeared to be cycloalkenylated at some point during the
course of the reaction, even in compounds 8 and 9, which did
not display an apparent cycloalkenyl substitution.
Furthermore, during the preparation of 5a, 1-(5-chloro-2-
(cyclohexylideneamino)phenyl)cyclohexanol (12, 0.1%) was
also isolated as a key reaction intermediate (Scheme 6). This is
the only case in which the isolation of such an intermediate was
successfully (mainly luckily) achieved. Several other attempts to
isolate such an intermediate during other reactions were proven
unsuccessful.
On the basis of the data collected throughout this study, we
hypothesized that the Skraup−Doebner−Von Miller reaction
Scheme 6. 1-(5-Chloro-2-
(cyclohexylideneamino)phenyl)cyclohexanol (12) Isolated
as a Side Product during the Synthesis of 5a
2786
dx.doi.org/10.1021/jo202681r | J. Org. Chem. 2012, 77, 2784−2790

The Journal of Organic Chemistry
Article
Scheme 7. Proposed Mechanistic Pathway for Skraup−Doebner−Von Miller Quinoline Synthesis Involving Cyclic Ketones
involved in the synthesis of the above-listed cycloalkenylated
5,6,7,8,9,10-hexahydro-6,6-pentamethylene-phenanthridines
and 2,3,4,5-tetrahydro-4,4-tetramethylene-1H-cyclopenta[c]-
quinolines may be proceeding through three basic sequences:
(1) the formation of a Schiff base through a reaction between
the ketone and the aniline derivative in the first step, followed
by (2) a cycloalkenylation at the ortho-position to the amine
functional group of the aniline derivative, and (3) an annulation
in the final step to close the quinoline ring, leading to a
dihydroquinoline derivative as described in Scheme 7.
The formation of a Schiff base as the initial step in the
mechanism of the Skraup−Doebner−Von Miller quinoline
synthesis is well accepted.12,13 Furthermore, Edwards et al.16
have previously reported that o-aminostyrenes can easily react
with ketones/cyclic ketones in the presence of a catalytic
amount of iodine to yield the corresponding spiro-dihydroqui-
noline derivatives as shown in Scheme 8. These findings were
recently confirmed by Denmark and Venkatraman.13
al.16 spiro-dihydroquinoline synthesis on the other, are clear
indications that the mechanism proposed in this study is
definitely the preferred pathway for this variant of the Skraup−
Doebner−Von Miller quinoline synthesis. However, since the
iodine-catalyzed ortho-vinylation of arylamines through a direct
reaction with open-chain ketones produced only traces of the
expected products during this study, this mechanism might not
apply to open-chain ketones. This is an additional proof that
the Skraup−Doebner−Von Miller quinoline synthesis follows
complex mechanistic pathways that might be dependent upon
the starting material, as previously suggested by Denmark and
Venkatraman.13
CONCLUSION
■
A series of unexpected and unusual 5,6,7,8,9,10-hexahydro-6,6-
pentamethylenephenanthridines and 2,3,4,5-tetrahydro-4,4-tet-
ramethylene-1H-cyclopenta[c]quinolines have been obtained
through the Skraup−Doebner−Von Miller quinoline synthesis.
These data have led to the development of an alternative and
straightforward mechanistic pathway for the Skraup−Doeb-
ner−Von Miller quinoline synthesis, at least as far as cyclic
ketones are concerned. More importantly, this report provides
an easy method for direct access to regioselective cyclo-
alkenylated dihydroquinoline derivatives through a single-pot
Skraup−Doebner−Von Miller quinoline synthesis. This study
also suggests that iodine can catalyze a one-pot regioselective
cycloalkenylation on aromatic rings (at least as far as arylamines
are concerned). Additional studies to determine the full scope
of such a one-pot regioselective cycloalkenylation reaction on
aromatic rings are already underway in our lab, and the
obtained results will be part of future reports.
Scheme 8. Reaction between o-Aminostyrenes and Cyclic
Ketones According to Edwards et al.16
Since this study shows that iodine can effectively catalyze an
ortho-cycloalkanylation on arylamines, in addition to the fact
that Edwards et al.16 have already proven that o-aminostyrenes
can easily react with ketones/cyclic ketones in the very same
conditions to yield the corresponding spiro-dihydroquinoline
derivatives, the mechanistic pathway proposed in this report is
viable and fully supported by strong data. More importantly,
the final step of this mechanism consists in a typical 6π-
electrocyclization of 2-azahexa-1,3,5-triene (electrocyclic clo-
sure),17 followed by a sigmatropic rearrangement ([1,5]-H
shift) resulting in a dihydroquinoline ring.
The fact that cyclohexanone and cyclopentanone were found
to quickly polymerize through self-condensation (cf. com-
pounds 2 and 3) is an indication that the pathway involving the
fragmentation and recombination (cf. ref 12) cannot be
completely excluded. Nevertheless, the high yield of the
cycloalkenylation reaction on one hand, and the Edwards et
EXPERIMENTAL SECTION
■
General Methods. NMR data were collected on two different
spectrometers, one operating at 600 MHz for 1H and 150 MHz for 13
C
1
and on another at 300 MHz for H and 75 MHz for 13C. Chemical
shifts were referenced to the solvent (CDCl3 δH observed at 7.25 ppm
and δC observed at 77.2 ppm). These data were recorded at 27 °C,
and all two-dimensional spectra were recorded with 2048 data points
for x-domain and 256 for y-domain. The edited g-HSQC and g-
HMBC spectra were optimized for 8 and 140 Hz, respectively. It is
important to mention that because of the complexity of the structure
of these molecules, and the succession and overlap of CH2 signals in
the 1H NMR spectra, it was difficult to accurately integrate these
protons in some cases. Thus, a special sequence known as DEPT-
HMQC was used for accurate integration of protons attached on each
carbon in these cases. Crystal structure analyses of 5a, 5b, and 7 were
2787
dx.doi.org/10.1021/jo202681r | J. Org. Chem. 2012, 77, 2784−2790

The Journal of Organic Chemistry
Article
based on data collected at low temperature on either a Nonius
KappaCCD diffractometer equipped with MoKα radiation or a Bruker
Kappa Apex-II CCD diffractometer with CuKα radiation. Melting
points were measured on a digital capillary melting point apparatus
calibrated with benzoic acid (≥99.5%) [mp 122.38 °C (lit), obtained
122.4−122.6 °C]. Reaction mixtures were monitored using a 200-MS
GC−MS ion trap mass spectrometer or by TLC silica gel 60 F254
plates. Gravity and flash column chromatography were performed
using type 60A silica gel (60−230 mesh). All the compounds were
further purified using silica gel GF preparative 1000 or 1500 μm
UV254 plates or recrystallized from a mixture of hexanes and ethyl
acetate (for compounds 5a, 5b, and 7).
Materials. All chemicals and solvents were purchased from major
chemical suppliers and were used without further purification unless
stated otherwise.
Preparation and Characterization of Compounds. All
reactions were performed at 160−165 °C for 72 h in the presence
of 0.05 equiv of iodine and 5 equiv of ketones.
2-Cyclohexenylcyclohexanone (3). This clear sticky oil was
present throughout the entire investigation each time cyclohexanone
was used as a starting material, but it was isolated only once for
characterization through a purification on a silica gel column using
hexanes−ethyl acetate (95:5) as eluent: 1H NMR (CDCl3, 600 MHz)
δ 1.51−1.69 (6H, m), 1.75−2.01 (8H, m), 2.15−2.39 (2H, m), 2.85−
2.87 (1H, m), 5.40 (1H, m); it is important to mention that the fact
that two protons from the same carbon appear at very different
chemical shift makes it difficult to accurately integrate these protons;
13C NMR (CDCl3, 150 MHz) δ 22.5, 22.9, 24.9, 25.3, 27.3, 27.7, 31.9,
42.2, 58.8, 123.7, 135.9, 211.8; HRESI-MS [M + H]+ m/z 179.1426
(calcd. 179.143042 for C12H19O).
by reacting 4-ethoxyaniline, also known as p-phenitidine (5 g, 36
mmol), and cyclohexanone (17.9 g, 182 mmol) in the presence of a
catalytic amount of iodine (463 mg, 1.82 mmol). After purification on
a silica gel column using hexanes−ethyl acetate (9:1), compound 5b
was obtained as a yellowish needles (7.3 g, 53%): mp 106.4−107.3 °C;
1H NMR (CDCl3, 600 MHz) δ 1.09−1.13 (2H, m), 1.26−1.31 (2H,
m), 1.35 (3H, t, J = 6.8 Hz), 1.40−1.42 (2H, m), 1.56−1.58 (2H, m),
1.65−1.72 (8H, m), 1.75−1.79 (2H, m), 2.10−2.13 (2H, m), 2.21−
2.25 (4H, m), 2.36−2.38 (2H, m), 3.96 (2H, q, J = 6.8 Hz), 4.84 (1H,
brs, NH), 5.83 (1H, m), 6.46 (1H, d, J = 2.8 Hz), 6.62 (1H, d, J = 2.7
Hz); 13C NMR (CDCl3, 150 MHz) δ 15.2, 21.6, 22.4, 22.5, 23.1, 23.5,
24.9, 25.5, 25.6, 25.8, 30.3, 32.1, 54.7, 64.0, 108.1, 112.6, 125.7, 125.8,
127.2, 129.9, 132.4, 135.9, 137.2, 151.1. The spiro-quaternary carbon
(C-6) appears at 54.7 ppm, while the g-HSQC and g-HMBC
sequences enable us to find the key correlation between the protons at
δ 5.83 ppm (the proton on the cyclohexenyl double bond carried by
the carbon at 125.7 ppm) and the quaternary aromatic carbon (C-4) at
δ 125.8 ppm. HRESI-MS: [M + H]+ m/z 378.2806 (calcd. 378.279142
for C26H36NO). The structure of this compound was unambiguously
determined by single crystal X-ray diffraction. Crystal data: C26H35NO,
Mr = 377.55, monoclinic space group P21/c, a = 18.545(3), b =
10.021(2), c = 11.5921(15) Å, β = 92.327(10)°, V = 2152.5(6) Å3, T =
95 K, Z = 4, Dx = 1.165 Mg m−3, θmax = 27.9°(MoKα), R = 0.050 for
5106 data and 287 refined parameters. This structure has some
disorder involving alternate conformations of six-membered rings. The
crystal structure data are deposited in the Cambridge database (CCDC
845945).
8-Chloro-6-cyclopentenyl-2,3,4,5-tetrahydro-4,4-tetra-
methylene-1H-cyclopenta[c]quinoline (6a). This compound was
obtained by reacting 4-chloroaniline (5 g, 39 mmol) and cyclo-
pentanone (16.5 g, 196 mmol) in the presence of a catalytic amount of
iodine (498 mg, 1.96 mmol). After purification on a silica gel column
using hexanes−ethyl acetate (9:1), compound 6a was obtained as
yellowish oil (7.4 g, 58%). This compound, like some other derivatives
obtained from cyclopentanone, showed some signs of auto-oxidation,
as it turned dark-greenish upon standing, a phenomenon previously
observed with dihydroquinolines in general7 and more importantly
with cylopentyl-spiro-derivatives of dihydroquinoline as reported by
Edwards et al.16 Nevertheless, this oxidation was not as dramatic as
previously reported,18 as indicated by the 13C NMR and HRESI-MS
2-Cyclopentylidenecyclopentanone (4). This clear sticky oil
was present throughout the entire investigation each time cyclo-
pentanone was used as a starting material. It was also isolated only
once for characterization through a purification on a silica gel column
1
using hexanes−ethyl acetate (95:5) as eluent: H NMR (CDCl3, 600
MHz) δ 1.61−1.65 (4H, m), 1.83−1.85 (2H, m), 2.21−2.24 (4H, m),
2.46−2.47 (2H, m), 2.70−2.71 (2H, m); 13C NMR (CDCl3, 150
MHz) δ 20.1, 25.3, 27.0, 29.6, 32.6, 34.3, 39.8, 127.9, 158.6, 207.4;
HRESI-MS [M + H]+ m/z 151.1118 (calcd. 151.111742 for
C10H15O).
1
spectra: H NMR (CDCl3, 600 MHz) δ 1.67−1.70 (6H, m), 1.83−
2-Chloro-4-cyclohexenyl-5,6,7,8,9,10-hexahydro-6,6-pen-
tamethylenephenanthridine (5a). This compound was obtained
by reacting 4-chloroaniline (5 g, 39 mmol) and cyclohexanone (19.3 g,
196 mmol) in the presence of a catalytic amount of iodine (498 mg,
1.96 mmol). After purification on a silica gel column using hexanes−
ethyl acetate (9:1), compound 5a was obtained as a yellowish needles
1.85 (2H, m), 1.95−2.03 (4H, m), 2.49−2.52 (2H, m), 2.54−2.56
(2H, m), 2.60−2.65 (4H, m), 4.64 (1H, brs, NH), 5.90 (1H, m), 6.73
(1H, d, J = 2.0 Hz), 6.84 (1H, d J = 2.0 Hz); 13C NMR (CDCl3, 150
MHz) δ 22.4, 23.3, 23.8, 31.4, 32.2, 33.9, 36.7, 39.6, 65.1, 121.1, 121.7,
122.0, 122.8, 125.9, 128.4, 132.0, 138.3, 140.1, 140.7. In this case, the
spiro-quaternary carbon (C-4) appears at 65.1 ppm, while the g-HSQC
and g-HMBC sequences enable us to find the key correlation between
the protons at δ 5.90 ppm (the proton on the cyclopentenyl double
bond carried by the carbon at 125.9 ppm) and the quaternary aromatic
carbon (C-6) at δ 122.8 ppm. HRESI-MS: [M + H]+ m/z 326.1669
(calcd. 326.167004 for C21H25NCl).
1
(8.4 g, 59%): mp 127.9−128.7 °C; H NMR (CDCl3, 600 MHz) δ
1.11−1.13 (2H, m), 1.27−1.29 (2H, m), 1.42−1.46 (2H, m), 1.58−
1.59 (2H, m), 1.66−1.71 (8H, m), 1.77−1.78 (2H, m), 2.10- 2.11
(2H, m), 2.20−2.22 (4H, m), 2.31−2.33 (2H, m), 5.00 (1H, brs, NH),
5.83 (1H, m), 6.80 (1H, d, J = 2.0 Hz), 6.93 (1H, d, J = 2.4 Hz); 13C
NMR (CDCl3, 150 MHz) δ 21.5, 22.3, 23.0, 23.4, 24.9, 25.4, 25.6,
25.7, 30.1, 32.5, 54.8, 120.6, 122.0, 125.0, 125.9, 126.0, 128.0, 130.5,
135.0, 137.1, 137.4. The spiro-quaternary carbon (C-6) appears at 54.8
ppm, while the g-HSQC and g-HMBC sequences enable us to find the
key correlation between the protons at δ 5.83 ppm (the proton on the
cyclohexenyl double bond carried by the carbon at 125.9 ppm) and the
quaternary aromatic carbon (C-4) at δ 126.0 ppm. HRESI-MS: [M +
H]+ m/z 368.2140 (calcd. 368.213954 for C24H31NCl). The structure
of this compound was unambiguously determined by single crystal X-
ray diffraction. Crystal data: C24H30ClN, Mr = 367.94, triclinic space
6-Cyclopentenyl-2,3,4,5-tetrahydro-4,4-tetramethylene-1H-
cyclopenta[c]quinoline (6b). This compound was obtained by
reacting aniline (5 g, 53.7 mmol) and cyclopentanone (22.6 g, 268
mmol) in the presence of a catalytic amount of iodine (682 mg, 2.68
mmol). After purification on a silica gel column using hexanes−ethyl
acetate (95:5), compound 6b was obtained as yellowish oil (7.2 g,
46%). This compound displayed better stability than its counterpart 6a
1
1
as indicated by the H and 13C NMR spectra: H NMR (CDCl3, 600
MHz) δ 1.70−1.72 (6H, m), 1.85−1.90 (2H, m), 2.00−2.03 (4H, m),
2.50−2.52 (2H, m), 2.54−2.60 (2H, m), 2.66−2.68 (4H, m), 4.64
(1H, brs, NH), 5.91 (1H, m), 6.59 (1H, t, J = 7.6 Hz), 6.80 (1H, dd, J
= 7.7 and 1.4 Hz), 6.90 (1H, dd, J = 7.6 and 1.4 Hz); 13C NMR
(CDCl3, 150 MHz) δ 22.5, 23.4, 23.8, 31.5, 32.2, 33.9, 36.8, 39.6, 65.0,
116.3, 120.5, 121.5, 122.2, 126.6, 127.3, 132.8, 138.5, 139.7, 141.8. In
this case, the spiro-quaternary carbon (C-4) appears at 65.0 ppm, while
the g-HSQC and g-HMBC sequences enable us to find the key
correlation between the proton at δ 5.91 ppm (the proton on the
cyclopentenyl double bond carried by the carbon at 126.6 ppm) and
group P1, a = 8.1800(10), b = 9.9875(13), c = 13.2073(15) Å, α =
̅
97.344(6), β = 98.122(8), γ = 111.627(8)°, V = 973.8(2) Å3, T = 95 K,
Z = 2, Dx = 1.255 Mg m−3, θmax = 33.1°(MoKα), R = 0.044 for 7378
data and 257 refined parameters. This structure has some disorder
involving alternate conformations of six-membered rings. The crystal
structure data are deposited in the Cambridge database (CCDC
845943).
4-Cyclohexenyl-2-ethoxy-5,6,7,8,9,10-hexahydro-6,6-pen-
tamethylenephenanthridine (5b). This compound was obtained
2788
dx.doi.org/10.1021/jo202681r | J. Org. Chem. 2012, 77, 2784−2790

The Journal of Organic Chemistry
Article
the quaternary aromatic carbon (C-6) at δ 121.5 ppm. HRESI-MS: [M
+ H]+ m/z 292.2060 (calcd. 292.205976 for C21H26N).
(2H, q, J = 6.9 Hz), 5.37 (1H, s), 5.71 (1H, m), 6.46 (1H, d, J = 2.8
Hz), 6.62 (1H, d, J = 2.8 Hz); 13C NMR (CDCl3, 150 MHz) δ 15.2,
18.9, 22.2, 23.3, 25.5, 29.8, 30.0, 50.8, 51.5, 64.2, 109.5, 114.1, 123.4,
127.2, 129.1, 129.9, 134.0, 136.0, 150.7. The g-HSQC and g-HMBC
sequences enable us to find the key correlation between the proton at
δ 5.71 ppm (the proton on the cyclohexenyl double bond carried by
the carbon at 127.2 ppm) and the quaternary aromatic carbon (C-6) at
δ 129.1 ppm. HRESI-MS: [M + H]+ m/z 298.2158 (calcd. 298.216541
for C20H28NO).
2,4-Dicyclohexenyl-5,6,7,8,9,10-hexahydro-6,6-pentame-
thylenephenanthridine (7). This compound was obtained by
reacting aniline (5 g, 53.7 mmol) and cyclohexanone (26.3 g, 268
mmol) in the presence of a catalytic amount of iodine (682 mg, 2.68
mmol). After purification on a silica gel column using hexanes−ethyl
acetate (95:5), compound 7 was obtained as a transparent needles
1
(14.0 g, 63%): mp 146.5 −147.9 °C; H NMR (CDCl3, 600 MHz) δ
1.11−1.14 (2H, m), 1.28−1.32 (2H, m), 1.42−1.48 (2H, m), 1.56−
1.59 (2H, m), 1.63−1.79 (14H, m), 2.10−2.13 (2H, m), 2.17−2.18
(2H, m), 2.19−2.23 (2H, m), 2.25−2.28 (2H, m), 2.38−2.40 (2H, m),
2.41−2.43 (2H, m), 5.03 (1H, brs, NH), 5.83 (1H, m), 6.00 (1H, m),
6.91 (1H, d, J = 2.1 Hz), 7.05 (1H, d, J = 2.1 Hz); 13C NMR (CDCl3,
150 MHz) δ 21.5, 22.5, 22.6, 22.7, 23.4, 23.5, 24.9, 25.5, 26.0, 27.6,
30.4, 32.4, 54.7, 117.3, 121.4, 123.1, 124.0, 125.9, 127.2, 129.2, 131.3,
135.9, 136.2, 136.7, 137.2. The spiro-quaternary carbon (C-6) appears
at 54.7 ppm, while the g-HSQC and g-HMBC sequences enable us to
find key correlations between the protons at δ 5.83 ppm (the proton
on the ortho-cyclohexenyl double bond carried by the carbon at 127.2
ppm) and the quaternary aromatic carbon (C-4) which also appears at
δ 129.2 ppm, and between the proton at 6.00 ppm (the proton on the
para-cyclohexenyl double bond carried by the carbon at 121.4 ppm)
and the quaternary aromatic carbon (C-2) at δ 131.3 ppm. HRESI-
MS: [M + H]+ m/z 414.3157 (calcd. 414.315527 for C30H40N). The
structure of this compound was unambiguously determined by single
crystal X-ray diffraction. Crystal data: C30H39N, Mr = 413.62,
orthorhombic space group Pbca, a = 9.7385(4), b = 11.6917(6), c =
41.5907(19) Å, V = 4735.5(4) Å3, T = 90 K, Z = 8, Dx = 1.160 Mg
m−3, θmax = 59.3°(CuKα), R = 0.053 for 3376 data and 293 refined
parameters. This structure has some disorder involving alternate
conformations of six-membered rings. The crystal structure data are
deposited in the Cambridge database (CCDC 845944).
4-Chloro-5,6,7,8,9,10-hexahydro-6,6-pentamethylenephe-
nanthridine (8). This compound was obtained by reacting 2-
chloroaniline (5 g, 39.2 mmol) and cyclohexanone (19.2 g, 196 mmol)
in the presence of a catalytic amount of iodine (498 mg, 1.96 mmol).
After purification on a silica gel column using hexanes−ethyl acetate
(97.5:2.5), compound 8 was obtained as yellowish oil (4.3 g, 38%).
This compound also showed some signs of decomposition as observed
on the 1H NMR spectrum: 1H NMR (CDCl3, 600 MHz) δ 1.50−1.53
(2H, m), 1.58−1.61 (2H, m), 1.65−1.73 (8H, m), 2.13−2.14 (2H, m),
2.35−2.36 (2H, m), 5.04 (1H, brs, NH), 6.59 (1H, dd, J = 7.8, and 8.2
Hz), 6.97 (1H, d,d J = 8.2 and 2.4 Hz), 7.05 (1H, dd, J = 7.8 and 2.4
Hz); 13C NMR (CDCl3, 150 MHz) δ 21.2, 22.5, 23.0, 25.0, 25.3, 25.6,
32.8, 55.1, 117.5, 118.2, 120.6, 125.0, 125.6, 127.0, 136.9, 138.5; the
spiro-quaternary carbon (C-6) appears at 55.1 ppm; HRESI-MS [M +
H]+ m/z 288.1507 (calcd. 288.151354 for C18H23NCl).
8-Cyclopentenyl-6-ethoxy-1,2-dihydro-2,2,4-trimethylqui-
noline (11). This compound was obtained by reacting ethoxyquin
(2.02 g, 9.3 mmol) and cyclopentanone (3.9 g, 46.5 mmol) in the
presence of a catalytic amount of iodine (118 mg, 0.46 mmol). After
purification on a silica gel column using hexanes−ethyl acetate (9:1),
compound 11 was obtained as yellowish oil (1.37 g, 52%). This
compound, as for some other derivatives obtained from cyclo-
pentanone, showed some signs of auto-oxidation, as it turned dark-
greenish upon standing. However, this oxidation was not very dramatic
as indicated by the 13C NMR and HRESI-MS spectra:18 1H NMR
(CDCl3, 600 MHz) δ 1.25 (6H, s), 1.38 (3H, t, J = 6.9 Hz), 1.99−2.00
(2H, m), 2.00 (3H, s), 2.57−2.58 (2H, m), 2.66−2.67 (2H, m), 3.96
(2H, q, J = 6.9 Hz), 4.20 (1H, brs, NH), 5.40 (1H, s), 5.90 (1H, m),
6.58 (1H, d, J = 2.8 Hz), 6.64 (1H, d, J = 2.8 Hz) ; 13C NMR (CDCl3,
150 MHz) δ 15.2, 19.1, 23.3, 30.3, 33.9, 36.8, 51.6, 64.2, 110.0, 113.6,
123.2, 128.4, 129.1, 129.8, 134.9, 141.2, 150.4. The g-HSQC and g-
HMBC sequences enable us to find the key correlation between the
proton at δ 5.90 ppm (the proton on the cyclopentenyl double bond
carried by the carbon at 129.1 ppm) and the quaternary aromatic
carbon (C-6) at δ 123.2 ppm. HRESI-MS: [M + H]+ m/z 284.2009
(calcd. 284.200891 for C19H26NO).
1-(5-Chloro-2-(cyclohexylideneamino)phenyl)cyclohexanol
(12). This compound was obtained as a side product during the
preparation of 5a, in which 4-chloroaniline (5 g, 39 mmol) and
cyclohexanone (19.3 g, 196 mmol) were allowed to react in the
presence of a catalytic amount of iodine (498 mg, 1.96 mmol). After
purification on a silica gel column using hexanes−ethyl acetate
(85:15), compound 12 was obtained as white solid (12 mg, 0.1%): mp
1
103.4 −103.9 °C; H NMR (CDCl3, 300 MHz) 1.65−1.71 (3H, m),
1.90−1.99 (3H, m), 2.27−2.37 (3H, m), 3.01−3.13 (3H, m), 3.16−
3.20 (3H, m), 3.23−3.28 (3H, m), 4.15−4.19 (2H, m), 4.16 (1H, brs,
OH) (this OH proton is hidden by the 2H around 4.15−4.19, and that
explains the broad base observed for those protons and the fact that
the integral shows 3H instead of 2H), 7.55 (1H, dd, J = 9.3 and 1.8
Hz), 7.73 (1H, d, J = 1.8 Hz), 7.88 (1H, d, J = 9.3 Hz); 13C NMR
(CDCl3, 75 MHz) 23.8, 24.2, 25.3, 31.2, 31.3, 34.9, 78.3, 123.0, 125.6,
129.3, 129.4, 131.4, 144.5, 152.3, 158.9; HRESI-MS [M + H]+ m/z
306.1429 (C18H25ClNO), with the base peak [M − H2O + H]+ m/z
288.1518 (calcd. 288.151354 for C18H23NCl).
6-Chloro-2,3,4,5-tetrahydro-4,4-tetramethylene-1H-
cyclopenta[c]quinolines (9). This compound was obtained by
reacting 2-chloroaniline (5 g, 39.2 mmol) and cyclopentanone (16.5 g,
196 mmol) in the presence of a catalytic amount of iodine (498 mg,
1.96 mmol). After purification on a silica gel column using hexanes−
ethyl acetate (95:5), compound 9 was obtained as yellowish oil (3.54
ASSOCIATED CONTENT
■
S
* Supporting Information
1H NMR and 13C NMR of 3, 4, 5a, 5b, 6a, 6b, 7, 8, 9, 10, 11,
and 12; the molecular structures obtained from single crystal X-
ray diffraction as well as the single crystal X-ray diffraction data
(CIF files) for compounds 5a, 5b, and 7; the HRESI-MS
spectra of 6a, 8, and 11 that showed some signs of
decomposition in their 1H NMR spectra. This material is
available free of charge via the Internet at http://pubs.acs.org.
1
g, 35%): H NMR (CDCl3, 600 MHz) δ 1.69−1.72 (2H, m), 1.78−
1.81 (4H, m), 1.89−1.91 (2H, m), 2.01−2.04 (2H, m), 2.51−2.52
(2H, m), 2.63−2.65 (2H, m), 4.53 (1H, brs, NH), 6.51 (1H, dd, J =
8.3 and 7.6 Hz), 6.77 (1H, dd, J = 7.6 and 2.3 Hz), 7.01 (1H, dd, J =
8.3 and 2.3 Hz); 13C NMR (CDCl3, 150 MHz) δ 22.6, 23.9, 31.3, 32.3,
40.4, 65.3, 116.5, 116.8, 121.6, 121.7, 127.4, 131.9, 138.9, 139.7; in this
case, the spiro-quaternary carbon (C-4) appears at 65.3 ppm; HRESI-
MS [M + H]+ m/z 260.1206 (calcd. 260.120054 for C16H19NCl).
8-Cyclohexenyl-6-ethoxy-1,2-dihydro-2,2,4-trimethylquino-
line (10). This compound was obtained by reacting ethoxyquin (1.5 g,
6.9 mmol) and cyclohexanone (3.4 g, 34.5 mmol) in the presence of a
catalytic amount of iodine (88 mg, 0.35 mmol). After purification on a
silica gel column using hexanes−ethyl acetate (9:1), compound 10 was
obtained as yellowish oil (1.11 g, 54%): 1H NMR (CDCl3, 600 MHz)
δ 1.22 (6H, s), 1.36 (3H, t, J = 6.9 Hz), 1.68−1.69 (2H, m), 1.75−1.76
(2H, m), 1.98 (3H, s), 2.16−2.18 (4H, m), 3.42 (1H, brs, NH), 3.96
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: jean.fotie@selu.edu.
Notes
The authors declare no competing financial interest.
2789
dx.doi.org/10.1021/jo202681r | J. Org. Chem. 2012, 77, 2784−2790

The Journal of Organic Chemistry
Article
(18) They turned into dark gummy compounds regardless of the
ACKNOWLEDGMENTS
■
1
method of purification, and the H NMR of these compounds show
J.F. acknowledges the financial support of Southeastern
Louisiana University through a Faculty Enhancement Grant.
D.S.B. gratefully acknowledges support from NSERC and the
CRC.
some signs of decomposition. However, the 13C NMR and the HRESI-
MS spectra were very decent, even after the sample was kept on the
bench for 24 h, suggesting that the decomposition of this type of
compound is not as dramatic as previously reported by Edwards et
al.16
REFERENCES
■
(1) Fotie, J. Key Natural Products in Malaria Chemotherapy: From
Quinine to Artemisinin and beyond. In Bioactive Natural Products:
Opportunities and Challenges in Medicinal Chemistry; Brahmachari, G.,
Ed.; World Scientific Publishing: Singapore, 2011; Chapter 3, pp 223−
271.
(2) Thorisson, S.; Gunstone, F.; Hardy, R. J. Am. Oil Chem. Soc.
1992, 69, 806−809.
(3) Lockhart, B.; Bonhomme, N.; Roger, A.; Dorey, G.; Casara, P.;
Lestage, P. Eur. J. Pharmacol. 2001, 416, 59−68.
(4) Dillard, R.; Pavey, D.; Benslay, D. J. Med. Chem. 1973, 16, 251−
253.
(5) Takahashi, H.; Bekkali, Y.; Capolino, A.; Gilmore, T.; Goldrick,
S.; Nelson, R.; Terenzio, D.; Wang, J.; Zuvela-Jelaska, L.; Proudfoot, J.;
Nabozny, G.; Thomson, D. Bioorg. Med. Chem. Lett. 2006, 16, 1549−
1552.
(6) Shah, A.; Pierson, J.; Pavlostathisa, S. Water Res. 2005, 39, 4251−
4263.
(7) (a) Fotie, J.; Kaiser, M.; Delfín, D. A.; Manley, J.; Reid, C. S.;
Paris, J. M.; Wenzler, T.; Maes, L.; Mahasenan, K. V.; Li, C.;
Werbovetz, K. A. J. Med. Chem. 2010, 53, 966−982. (b) Fotie, J.;
Fronczek, C. F.; Burns, K. A.; Fronczek, F. R.; Bain, C.; Bohle, D. S.;
Poudeu, F. P. Acta Crystallogr. 2010, E66, 1660. (c) Allen, K. M.; Fotie,
J. Synthesis of unusual dihydroquinoline derivatives through a
modified Skraup reaction. ACS Joint 66th Southwest and 62nd
Southeastern Regional Meeting, New Orleans, LA, Dec 1−4, 2010.
(d) Reid, C. S.; Patrick, D. A.; He, S.; Fotie, J.; Premalatha, K.; Tidwell,
R. R.; Wang, M. Z.; Liu, Q.; Gershkovich, P.; Wasan, K. M.; Wenzler,
T.; Brun, R.; Werbovetz, K. A. Bioorg. Med. Chem. 2011, 19, 513−523.
(8) (a) Skraup, Z. H. Monatsh. Chem. 1880, 1, 316−318. (b) Skraup,
Z. H. Monatsh. Chem. 1881, 2, 139−170. (c) Skraup, Z. H. Monatsh.
Chem. 1881, 2, 587−609. (d) Skraup, Z. H. Ber. Dtsch. Chem. Ges.
1882, 15, 897.
(9) Doebner, O.; Miller, W. V. Ber. Dtsch. Chem. Ges. 1881, 14,
2812−2817.
(10) (a) Reddelien, G.; Thurm, A. Chem. Ber. 1932, 65, 1511−1521.
(b) Craig, D. J. Am. Chem. Soc. 1938, 60, 1458−1465. (c) Theoclitou,
M.-E.; Robinson, L. A. Tetrahedron Lett. 2002, 43, 3907−3910.
(d) Kamakshi, R.; Reddy, B. S. R. Catal. Commun. 2007, 8, 825−828.
(e) Hegedus, A.; Hell, Z.; Vargadi, T.; Potor, A.; Gresits, I. Catal. Lett.
̈
2007, 117, 99−101.
(11) Konig, W. Ber. Dtsch. Chem. Ges. 1923, 56B, 1853−1855.
̈
(12) (a) Eisch, J. J.; Dluzniewski, T. J. Org. Chem. 1989, 54, 1269−
1274. (b) Walter, H.; Sauter, H.; Winkler, T. Helv. Chim. Acta 1992,
75, 1274−1280. (c) Walter, H.; Schneider, J. Heterocycles 1995, 41,
1251−1269. (d) Walter, H.; Sundermann, C. Heterocycles 1998, 48,
1581−1591.
(13) Denmark, S. E.; Venkatraman, S. J. Org. Chem. 2006, 71, 1668−
1676.
(14) Mammen, H.; Dietzel, S.; Martin, H. D. Z. Chem. 1985, 25,
435−436.
(15) Monitoring these reactions by GC−MS showed that a very
small amount of the para-alkenylated derivative was present, but we
never managed to isolate them.
(16) Edwards, J. P.; Ringgenberg, J. D.; Jones, T. K. Tetrahedron Lett.
1998, 39, 5139−5142.
(17) For 6π-electrocyclization of azahexa-1,3,5-trienes, see: (a) Chen,
Z.-B.; Hong, D.; Wang, Y.-G. J. Org. Chem. 2009, 74, 903−905.
(b) Molina, P.; Pastor, A.; Vilaplana, M. J. Tetrahedron 1993, 49,
́
7769−7778. (c) Palacios, F.; Herran, E.; Alonso, C.; Rubiales, G.;
Lecea, B.; Ayerbe, M.; Cossío, F. P. J. Org. Chem. 2006, 71, 6020−
6030.
2790
dx.doi.org/10.1021/jo202681r | J. Org. Chem. 2012, 77, 2784−2790