On the mechanism of the Skraup-Doebner-Von Miller quinoline synthesis

Article abstract of DOI:10.1021/jo052410h

The mechanism of the formation of substituted quinolines from anilines and α,β-unsaturated ketones has been studied by the use of ¹³C-labeled ketones in cross-over experiments. In the reaction of doubly labeled ¹³C(2,4) mesityl oxide

Full text of DOI:10.1021/jo052410h

On the Mechanism of the Skraup-Doebner-Von Miller Quinoline
Synthesis
Scott E. Denmark* and Srikanth Venkatraman
Roger Adams Laboratory, Department of Chemistry, UniVersity of Illinois, Urbana, Illinois 61801
denmark@scs.uiuc.edu
ReceiVed NoVember 21, 2005
The mechanism of the formation of substituted quinolines from anilines and R,â-unsaturated ketones has
been studied by the use of 13C-labeled ketones in cross-over experiments. In the reaction of doubly labeled
13C(2,4) mesityl oxide, a 100% scrambling of the label in the quinoline product was observed, whereas
only a small (5-10%) amount of the starting mesityl oxide showed scrambling of the label. Similarly,
the reaction of triply labeled pulegone clearly shows that the label in the product is 100% scrambled,
whereas the label in the starting pulegone is retained. On the basis of these studies, a mechanistic pathway
for the Skraup quinoline synthesis is proposed that involves a fragmentation-recombination mechanism.
The aniline component condenses with the R,â-unsaturated ketone initially in a conjugate fashion, followed
by a fragmentation to the corresponding imine and the ketone itself. These fragments recombine to form
the quinoline product.
Introduction
Subsequently, Doebner and Von Miller replaced the glycerol
with an R,â-unsaturated ketone and conducted the reaction by
heating with an aromatic amine in the presence of an acid
catalyst or iodine (Scheme 1).5
The quinoline ring system is found in a myriad of naturally
occurring as well as medicinally active synthetic drug sub-
stances. The early success of quinolines as antimalarial drugs
has stimulated the development of various methods for their
synthesis over the last century.1 In addition, quinolines are
present as structural subunits in metabolites derived from various
flora and fauna.2 Finally, quinolines are active components in
various industrial antioxidants and dyes. Among the most
general approaches to this ring system is the classic Skraup
quinoline synthesis.3 In the early 1880s, Skraup described
heating a mixture of nitroethane, aniline, and glycerol with
concentrated sulfuric acid to form quinoline in a very low yield.4
SCHEME 1
Over the last century, various modifications have been made
to improve the yield and reproducibility of the Skraup quinoline
synthesis. Various moderators such as acetic or boric acids,
ferrous sulfate, thorium, or vanadium or iron oxides have been
used to accelerate the reaction and make it higher yielding.6
The Skraup-Doebner-Von Miller synthesis of quinolines can
also be catalyzed by a number of Lewis (SnCl4, Yb(OTf)3, Sc-
(OTf)3, ZnCl2, InCl3) and Bro¨nsted acids (TsOH, HClO4,
Amberlite) in addition to iodine.7 In addition, microwave irradia-
(1) Bradley, S. G.; Marciano-Cabral, F. Antiparasitic Drugs. In Principles
of Pharmacology; Munson, P. L., Mueller, R. A., Breese, G. R., Eds.;
Chapman & Hall: New York, 1994; Chapter 100.
(2) (a) Michael, J. P. Nat. Prod. Rep. 2005, 22, 627-646. (b) Michael,
J. P. Nat. Prod. Rep. 2004, 21, 650-668.
(3) Throughout this paper, we refer to the quinoline ring systems
generically to include both saturated and unsaturated derivatives.
(4) (a) Jones, G. Synthesis of the Quinoline Ring System. In Quinolines;
Jones, G., Ed.; Wiley: New York, 1977; Chapter 2. (b) Manske, R. H. F.;
Kulka, M. Org. React. 1953, 7, 59-98. (c) Skraup, Z. H. Monatsh. Chem.
1881, 2, 139-170. (d) Skraup, Z. H. Monatsh. Chem. 1880, 1, 316-318.
(e) Skraup, Z. H. Monatsh. Chem. 1881, 2, 587-609. (f) Skraup, Z. H.
Ber. Dtsh. Chem. Ges. 1882, 15, 897.
(5) Doebner, O.; Miller, W. V. Ber. Dtsh. Chem. Ges. 1881, 14, 2812-
2817.
10.1021/jo052410h CCC: $33.50 © 2006 American Chemical Society
Published on Web 01/24/2006
1668
J. Org. Chem. 2006, 71, 1668-1676

Skraup-Doebner-Von Miller Quinoline Synthesis
tion in combination with various Lewis acid activators has
recently been employed with good results.8 Various substituted
anilines have been used in the Skraup quinoline synthesis. For
ortho- and para-substituted anilines, the regiochemical outcome
is unambiguous. However, the structure of the quinoline pro-
ducts obtained using meta-substituted anilines is unpredictable.
The easy formation of quinolines from inexpensive starting
materials makes this an important synthetic reaction despite the
practical shortcomings. We surmised that further improvements
in the reaction yield and reproducibility would occur from a
detailed understanding of the mechanism. Thus, we embarked
on a mechanistic inquiry designed to answer several questions
about the nature of the reactive intermediates and the timing of
the bond-forming events. These studies are described in full
below.
A number of different investigators provided insights into
the mechanism by isolating various intermediates in the reaction
sequence. For example, Badger et al. reacted 8-amino-3,4
dimethylquinoline (7) with methyl vinyl ketone in glacial acetic
acid at room temperature to obtain N-(3′,4′-dimethyl-8-quinolyl)-
4-aminobutane-2-one (8), which on heating in 80% arsenic acid
forms the corresponding quinoline 9 (Scheme 3).7f,12 This
suggested that the conjugate addition is the first step in the
annulation sequence.
SCHEME 3
Background
In view of the interest in the Skraup-Doebner-Von Miller
quinoline synthesis, it is not surprising that many mechanistic
studies are already on record. Skraup himself suggested that
aldehyde anils underwent direct acid-catalyzed closure to
quinolines.4f However, this proposal was discounted by the
demonstration that 3-substituted R,â-unsaturated aldehydes
afford 2-substituted quinolines. To accommodate this fact, Ko¨nig
proposed a modification of a mechanism first suggested by
Bischler which involves the 3-anilinopropanal imine 5 as the
key intermediate (Scheme 2).9 Ko¨nig’s mechanistic proposal
was subsequently supported by deuterium-labeling experi-
ments.10 These studies showed conclusively that anils cannot
undergo direct closure but must either revert to the â-anilino
carbonyl compounds and cyclize or react via the conjugate
adducts (see below). However, in 1989, Eisch showed that under
anhydrous conditions, in the absence of free anilines, isolated
aldehyde anils undergo a rearrangement via 1,3-diazetidinium
ions to afford 2-substituted quinolines.11
Furthermore, Dauphinee and Forrest isolated imine 10 upon
treatment of aniline with acetaldehyde, which suggested that
the formation of quinoline 11 proceeds by the self-condensation
of a Schiff base followed by cyclization as illustrated in Scheme
4.13
SCHEME 4
SCHEME 2
In one of the more intriguing studies, Tung claims the
isolation of various ketone anils by the treatment of aniline with
mesityl oxide in boiling benzene.14 Reaction of the anil 12
derived from aniline and mesityl oxide (2) with excess
4-ethoxyaniline (13) resulted in the formation of a quinoline
derived only from 4-ethoxyaniline (Scheme 5). The formation
of a single quinoline product 15 led him to speculate a
SCHEME 5
(6) (a) Cohn, B. E.; Gustavson, R. G. J. Am. Chem. Soc. 1928, 50, 2709-
2711. (b) Cohn, E. W. J. Am. Chem. Soc. 1930, 52, 3685-3688. (c) Clarke,
H. T.; Davis, A. W. Org. Synth. 1922, 2, 79-83. (d) Darzens, G.; Delaby,
R.; Hiron, J. Bull Soc. Chim. 1930, 47, 227-232.
(7) (a) Arduini, A.; Bigi, F.; Casiraghi, G.; Casnati, G.; Sartori, G.
Synthesis 1981, 975-977. (b) Glinka, J. Roczn. Chem. 1965, 39, 885-
893. (c) Bowers, J. S. U.S. Patent 4514570, 1985. (d) Layer, R. W.;
Cuyahoga, F.; Son, P.-N. U.S. Patent 4069195, 1978. (e) Badger, G. M.;
Crocker, H. P.; Ennis, B. C.; Gayler, J. A.; Matthews, W. E.; Raper, W. G.
C.; Samuel, E. L.; Spotswood, T. M. Aust. J. Chem. 1963, 16, 814-827.
(8) (a) Theoclitou, M.-E.; Robinson, L. A. Tetrahedron Lett. 2002, 43,
3907-3910. (b) Ranu, B. C.; Hajra, A.; Dey, S. S.; Jana, U. Tetrahedron
59, 59, 813-819.
(9) (a) Bischler, A. Ber. Dtsh. Chem. Ges. 1892, 25, 2860-2879. (b)
Ko¨nig, W. Ber. Dtsh. Chem. Ges 1923, 56B, 1853-1855.
(10) Forrest, T. P.; Dauphinee, G. A.; Miles, W. F. Can. J. Chem. 1969,
47, 2121-2122.
(11) Eisch, J. J.; Dluzniewski, T. J. Org. Chem. 1989, 54, 1269-1274.
J. Org. Chem, Vol. 71, No. 4, 2006 1669

Denmark and Venkatraman
SCHEME 6
unidirectional ring closure of the intermediate 14 to form the
quinoline product.
oxide would result in a quinoline having a label at the C(2)
carbon from cyclization via the corresponding anil. The position
of the labels could be easily established by 13C NMR spectros-
copy.
In all of the preceding studies, the direct closure of an anil
to a quinoline was ruled out. However, in an intriguing series
of papers, Walter observed the formation of a quinoline product
from aniline and (R)-pulegone (16) whose structure could not
be explained by any existing mechanistic proposals (Scheme
6).15 To explain the formation of 22, Walter invoked a series
of electrocyclic ring closures, ring openings, and hydrogen
transfers starting from anil 17 to rationalize the formation of
the rearranged Skraup product. The mechanism proposed by
Walter was supported by the demonstration that 23 (indepen-
dently prepared) underwent a facile conversion to the expected
quinoline 22 on treatment with 3-methylcyclohexanone in 38%
yield (Scheme 7). Moreover, both Walter15 and Edwards16 have
extended this observation to develop a general synthesis of 2,2-
disubstituted dihydroquinolines starting from 2-alkenylanilines
and carbonyl compounds.
SCHEME 8
SCHEME 7
Results
1. Reaction of 4-Isopropylaniline with 13C(2)-Acetone. The
reaction of acetone with 4-isopropylaniline (1) was conducted
to establish the reaction conditions and the characteristic proton
and carbon resonances of the quinoline product. The choice of
4-isopropylaniline was based on the fact that the product
obtained was stable and easy to purify. The reaction of
4-isopropylaniline with 5 equiv of acetone containing 20 mol
% of 13C(2) acetone (100% labeled) at 150 °C for 3 h in a sealed
tube gave the quinoline 3 in 47% yield (Scheme 9). Two
resonances were clearly enriched, thus demonstrating the
incorporation of two acetone molecules to form the quinoline.
Spectroscopic analysis of the product allowed the assignment
Analysis of the mechanisms proposed by Walter and Tung
(and their variants) suggested that they could be distinguished
by the labeling experiment outlined in Scheme 8. If one
considers the Skraup reaction of aniline with 13C(2)-labeled
mesityl oxide, the 13C label should reside at C(4) of the quinoline
according to the mechanism of Tung, Ko¨nigs, and Badger. This
is in line with the suggestion that the conjugate addition of
aniline to R,â-unsaturated ketone or the corresponding imine
leads to the product formation. However, if the Walter mech-
anism is operative, an experiment with 13C(2)-labeled mesityl
SCHEME 9
(12) (a) Badger, G. M.; Crocker, H. P.; Ennis, B. C.; Gayler, J. A.;
Matthews, W. E.; Raper, W. G. C.; Samuel, E. L.; Spotswood, T. M. Aust.
J. Chem. 1963, 16, 814-827. (b) Badger, G. M.; Ennis, B. C.; Matthews,
W. E. Aust. J. Chem. 1963, 16, 828-832. (c) Badger, G. M.; Crocker, H.
P.; Ennis, B. C. Aust. J. Chem. 1963, 16, 840-844.
(13) (a) Forrest, T. P.; Dauphinee, G. A.; Deraniyagala, S. A. Can. J.
Chem. 1985, 63, 412-417. (b) Dauphinee, G. A.; Forrest, T. P. Can. J.
Chem. 1978, 56, 632-634.
(14) Tung, C. C. Tetrahedron 1963, 19, 1685-1689.
1670 J. Org. Chem., Vol. 71, No. 4, 2006

Skraup-Doebner-Von Miller Quinoline Synthesis
of the signal at 51.8 ppm to the13C(2) carbon and signal at 128.6
ppm to the olefinic carbon at 13C(4).
formation, a mass spectrometric analysis of the product would
show a high (M + 1)+ signal. For synthetic ease, we chose to
double label mesityl oxide in the C(2) and C(4) positions. The
synthesis of double-labeled mesityl oxide was achieved using
the method developed by Wayne and Adkins.18 Heating a
benzene solution of 99% 13C(2)-labeled acetone with 0.5 equiv
of freshly prepared Al(t-BuO)3 for 8 h resulted in a green
gelatinous solution. Careful chromatographic separation with
ether/pentane yielded mesityl oxide (2) in 9% and phorone (24)
in 15% yield (Scheme 11). Various bases such as CaO, NaOMe,
and Ti(OMe)4 were tried to increase the yield of the reaction,
but proved disappointing.19 Although synthetically impractical,
this method yielded double-labeled mesityl oxide to conduct
the double label cross over experiment. The isotopic distribution
of the 13C(2,4)-labeled mesityl oxide was determined using FI
mass spectrometry.
2. Reaction of 4-Isopropylaniline with 13C(2) Mesityl
Oxide. The synthesis of 13C(2)-labeled mesityl oxide was
accomplished following the method of Frangopol.17 Acetyl
chloride (99% 13C(1) labeled) was diluted with acetyl chloride
at natural abundance such that the resultant mixture contained
10 mol % of 13C(1)-labeled material. Subsequent Friedel-Crafts
reaction of the acetyl chloride with tert-butyl chloride and
anhydrous AlCl3 (neat, 25°C, 3 h) afforded 13C(2)-labeled
mesityl oxide (2) in 31% yield (Scheme 10). 13C NMR
spectroscopic analysis revealed that the label was specifically
incorporated at the C(2) position.
SCHEME 10
SCHEME 11
The key reaction of 13C(2)-labeled mesityl oxide (2) was thus
carried out with 1 (toluene, I2 (mol %), reflux, 3 h) to form the
quinoline 3 (Scheme 10). Spectroscopic analysis (13C NMR)
of 3 showed 13C incorporation in both the C(2) and C(4)
positions of the quinoline.
The surprising incorporation of the label at both the C(2) and
the C(4) position of the quinoline could be explained by one of
the following mechanistic scenarios: (1) the Skraup reaction
proceeds by a dual mechanism and both the Tung and Walter
processes are operative under the reaction conditions, (2) aniline
assists in the fragmentation of mesityl oxide into component
ketones which recombine to form product, or (3) mesityl oxide
undergoes scrambling prior to the reaction with aniline to form
the quinoline. These limiting possibilities could be distinguished
with the help of double-labeled crossover experiments.
3. Double-Labeled Mesityl Oxide Crossover Experiment.
A crossover experiment with double-labeled mesityl oxide
would assist in deciphering whether both the Walter or the Tung
mechanisms are simultaneously operative or if mesityl oxide
undergoes disproportionation before the formation of quinoline.
If both the Walter and Tung mechanisms were simultaneously
operative, a Skraup quinoline synthesis with a equimolar mixture
of double-labeled and unlabeled mesityl oxide would form the
product quinoline which has a mass distribution of M+/(M +
2)+ in the same ratio as the starting mixture. If aniline were
assisting the fragmentation of mesityl oxide into component
ketones or if mesityl oxide were scrambling prior to quinoline
The 13C(2,4)-labeled mesityl oxide was diluted with mesityl
oxide at natural abundance, and the Skraup quinoline synthesis
with 4-isopropyl aniline was conducted in refluxing toluene.
The quinoline product 3 was isolated in 47% yield, and the mass
distribution of the quinoline was analyzed by FI mass spectral
integration (Figure 1).
FIGURE 1. Mass spectral analysis of the quinoline product 3.
(15) (a) Walter, H.; Sauter, H.; Winkler, T. HelV. Chim. Acta 1992, 75,
1274-1280. (b) Walter, H.; Sauter, H.; Schneider, J. HelV. Chim. Acta 1993,
76, 1469-1475. (c) Walter, H. HelV. Chim. Acta 1994, 77, 608-614. (d)
Walter, H. Heterocycles 1995, 41, 1251-1269. (e) Walter, H. Heterocycles
1995, 41, 2427-35. (f) Walter, H. J. Prakt. Chem. 1998, 340, 309-314.
(g) Walter, H.; Sundermann, C. Heterocycles 1998, 48, 1581-1591.
(16) Edwards, J. P.; Ringgenberg, J. D.; Jones, T. K. Tetrahedron Lett.
1998. 39. 5139-5142.
The double-labeled mesityl oxide crossover experiment was
conducted using a mixture of 13C(2,4)-labeled mesityl oxide and
mesityl oxide at natural abundance in a ratio of M+/(M + 1)+/
(18) (a) Wayne, W.; Adkins, H. Org. Synth. 1941, 21, 8-10. (b) Wayne,
W.; Adkins, H. J. Am. Chem. Soc. 1940, 62, 3401-3404.
(19) (a) Knoevenagel, E. N.; Blach, L. Ber. Dtsh. Chem. Ges. 1907, 39,
3451-3457. (b) Hoffman, A. J. Am. Chem. Soc. 1909, 31, 722-724.
(17) Frangopol, M.; Genunche, A.; Negoita, N.; Frangopol, P. T.;
Balaban, A. T. Tetrahedron 1967, 23, 841-844.
J. Org. Chem, Vol. 71, No. 4, 2006 1671

Denmark and Venkatraman
(M + 2)+ 54.1%/7.5%/34.0%. The quinoline product showed
a mass distribution of M+/(M + 1)+/(M + 2)+ 33.0%/47.7%/
18.2%. For the given ratio of M+/(M + 2)+ 54.1%/34.0%, the
expected mass distribution for 100% scrambling would be
37.7%/47.4%/14.9%. The observed value for the product
corresponds to nearly 100% scrambling. Accordingly, the Tung
and Walter mechanisms are not simultaneously operating, but
a mechanism must exist which requires the separation and
recombination of the acetone units in mesityl oxide. However,
before reaching that conclusion it was necessary to prove that
mesityl oxide was not simply undergoing scrambling before
incorporation into the quinoline.
5. Synthesis of Triple-Labeled Pulegone. The experiment
conducted by Walter with pulegone as the ketone resulted in
the formation of the rearranged Skraup product 22. We reasoned
that pulegone may be a more stable ketone, and suitable labeling
of it could assist in deciphering between the Tung and Walter
mechanisms. Toward that end, pulegone was labeled at the
aliphatic and the vinylic methyl groups, the two most syntheti-
cally accessible sites (Scheme 13). Labeled pulegone (16) was
synthesized following the procedure of Corey and Chen.20 The
mixed cuprate reagent prepared from 13CH3Li21 and PhSCu was
combined with cyclohexenone to form 3-(13C)-methylcyclo-
hexanone (25) in 70% yield.22 Treatment of 3-(13C)-methylcy-
clohexanone with Li-BHT, CS2, and MeI produced the bis-
methylthio ketene acetal 26, which underwent a second mixed
cuprate methyl substitution to form the triple-labeled pulegone
(16) in 79% yield.23
4. Scrambling of Mesityl Oxide. To test the hypothesis that
the mesityl oxide underwent scrambling prior to the formation
of quinoline, a control experiment was performed using 13C-
(2)-labeled mesityl oxide (Scheme 12).
The FI mass analysis of synthetic pulegone showed an ion
distribution of M+/(M + 1)+/(M + 2)+/(M + 3)+ 0.0%/0.0%/
2.4%/87.7%. Next, the thermal stability of pulegone was
established in a control experiment. Triple-labeled pulegone (16)
was diluted with pulegone at natural abundance, and the mixture
was heated to reflux in toluene with a catalytic amount of I2
(Scheme 14). The pulegone recovered from the reaction mixture
was analyzed for mass distribution by FI mass spectrometry.
The mass distribution of pulegone used in the experiment was
identical to the pulegone isolated from the mixture. Thus, with
the knowledge that the label in the pulegone does not scramble
under the Skraup reaction conditions in the absence of the amine,
the triple-labeled pulegone crossover experiment could be
conducted.
SCHEME 12
The C(2)-labeled mesityl oxide was subjected to the condi-
tions of the Skraup quinoline synthesis. The product and the
unreacted mesityl oxide were isolated, and the extent of
scrambling was qualitatively assessed by 13C NMR spectro-
scopy. The 13C spectrum of the product showed that the label
was distributed among both the C(2) and C(4) position of the
quinoline 3. The analysis of the unreacted mesityl oxide showed
the C(4) position was also partially enriched. However, a close
inspection of the 13C NMR spectra showed that the product was
equally enriched in C(2) and C(4) positions, whereas the mesityl
oxide was only enriched to ∼10-15% at C(4). This suggested
that the mechanism of scrambling of mesityl oxide was
independent of that of the product. The scrambling of mesityl
oxide likely arises from a slower, competitive retro aldol/aldol
reaction. Even though this clearly supports the conclusion that
both the Tung and Walter mechanisms are not operating
simultaneously, a double-labeled crossover experiment with a
more stable ketone would answer this question unambiguously.
SCHEME 14
6. Triple-Labeled Pulegone Crossover Experiment. A
mixture of pulegone with a mass distribution M+/(M + 3)+
44.0%/45.2% was heated with 4-isopropylaniline in the presence
SCHEME 15
SCHEME 13
(20) Corey, E. J.; Chen, R. H. K. Tetrahedron Lett. 1973, 14, 3817-
3820.
1672 J. Org. Chem., Vol. 71, No. 4, 2006

Skraup-Doebner-Von Miller Quinoline Synthesis
of 5 mol % of I2 to form quinoline 22 in 30% yield (Scheme
15). Both of the quinoline diastereomers (1:1, 30%) and the
unreacted pulegone (66%) were isolated. The distributions of
the unreacted pulegone 16 and the product quinoline 22a are
shown in Figures 2 and 3.
The relative mass distribution of starting pulegone was M+/
(M + 1)+/(M + 2)+/(M + 3)+ 44.0%/4.2%/1.5%/45.2%. The
isotopic intensities of the (M + 1) and (M + 2) peaks were
very low, just 4.2% and 1.5%. The relative mass distribution
of the recovered pulegone from the reaction mixture was M+/
(M + 1)+/(M + 2)+/(M + 3)+ 42.1%/5.7%/1.7%/46.0% (Figure
2). The mass distribution of the recovered pulegone was within
experimental error ((5%) of the mass distribution in the starting
pulegone. This clearly established that the pulegone is stable
under the reaction conditions and no scrambling of the starting
material is observed.24
1.5%/45.2%. The distribution of the 13C label in the product
quinolines was M+/(M + 1)+/(M + 2)+/(M + 3)+ 19.7%/25.3%/
26.2%/28.8% for the nonpolar diastereomer and 21.9%/25.6%/
25.4%/27.1% for the polar diastereomer. These are similar to
the mass distribution obtained for 100% theoretical scrambling
which is M+/(M + 1)+/(M + 2)+/(M + 3)+ 24.2%/25.0%/
25.0%/25.9%. The result of complete scrambling of the label
in the quinoline product together with the unscrambled label in
the recovered pulegone demonstrates clearly that the pulegone
dissociates irreversibly into the corresponding ketone fragments
before incorporation into the quinoline nucleus.
8. Reaction of 2-Isopropenyl-4-isopropylaniline (29). The
synthesis of 2-isopropenyl-4-isopropylaniline was accomplished
as shown in Scheme 16. 4-Isopropylaniline was protected as
the acetanilide which was then brominated in acetic acid to
afford 27.25 The bromide was cross-coupled with 2-ispropenyl-
tributylstannane using PdCl2(PPh3)2 to afford anilide 28.26
Hydrolysis of the acetamide group in 28 was accomplished in
refluxing alcoholic KOH to yield 2-isopropenyl-4-isopropyl-
aniline (29).27
SCHEME 16
FIGURE 2. Mass spectral analysis of recovered pulegone.
Aniline 29 was tested for its ability to undergo the Skraup
quinoline synthesis following Walter’s precedent. Heating 29
with 20% 13C(2)-enriched acetone under the conditions de-
scribed by Walter afforded the expected quinoline 3 in 32%
yield (Scheme 17). Thus, 29 is indeed capable of undergoing
the Skraup quinoline synthesis. However, the yield of the
product obtained was lower than that obtained in the reaction
of 4-isopropylaniline with acetone, and a substantial amount of
a secondary product of the quinoline with the starting aniline
was observed.
7. Analysis of the Quinoline Products. As observed by
Walter, the quinoline products isolated were not the normal
Skraup products. The products obtained, 22a/b, were identical
to the products observed by Walter, and the two diastereomers
were easily separable by silica gel chromatography. Both of
the diastereomers were independently analyzed for isotopic
distribution.
The mass distribution of the quinoline product clearly shows
the formation of peaks from crossover products (Figure 3). If
the reaction followed either the Tung or the Walter mechanism,
the mass distribution should be the same as that of the starting
pulegone viz. M+/(M + 1)+/(M + 2)+/(M + 3)+ 44.0%/4.2%/
SCHEME 17
Discussion
The labeling experiments conducted above clearly demon-
strate that for hindered ketones such as mesityl oxide or
pulegone direct cyclization of (1) an intermediate from conjuga-
(22) (a) Adams, R.; Reifschneider, W.; Ferretti, A. Org. Synth. 1962,
42, 22-25. (b) Posner, G. H.; Whitten, C. E.; Sterling, J. J. J. Am. Chem.
Soc. 1973, 95, 7788-7800. (c) Posner, G. H.; Whitten, C. E. Org. Synth.
1976, 55, 122-127.
(23) (a) Dieter, R. K.; Silks, L. A. J. Org. Chem. 1986, 51, 4687-4701.
(b) Dieter, R. K. Tetrahedron 1986, 42, 3029-3096. (c) Dieter, R. K.; Silks,
L. A.; Fishpaugh, J. R.; Kastner, M. E. J. Am. Chem. Soc. 1985, 107, 4679-
4692.
FIGURE 3. Mass spectral analysis of quinoline product 22a (more
polar diastereomer).
(24) A control experiment in which heating triply labeled pulegone with
iodine in toluene at reflux for 8 h also showed no scrambling of the label.
(21) Oppolzer, W.; Mirza, S. HelV. Chim. Acta 1984, 67, 730-738.
J. Org. Chem, Vol. 71, No. 4, 2006 1673

Denmark and Venkatraman
SCHEME 18
SCHEME 19
tion addition, (2) the anil, or (3) a conjugate adduct of the anil
is significantly slower than the fragmentation of the enones into
their component ketones. Moreover, some step prior to the
fragmentation must be irreversible to explain the lack of
scrambling of the label in the recovered enones. This fragmenta-
tion unfortunately precludes unambiguous conclusions to be
drawn about which of the mechanisms is operative. Neverthe-
less, it is instructive to consider a number of limiting mechanistic
scenarios to narrow the possibilities and also as a framework
for designing additional experiments to discriminate among
them.
Mechanism 1: Irreversible Anil Formation. The first two
mechanistic proposals invoke the initial formation of an anil.
In the first scenario, the initial condensation of aniline 1 occurs
in an irreversible manner to form the anil 30 Scheme 18. The
anil then undergoes conjugate addition with another molecule
of aniline and reversibly fragments to the component ketone
imines 32 and 33. This provides a mechanism for the scrambling
of the intermediate anil. The anil may further form the quinoline
following the Walter mechanism to form the quinoline products
as observed. This pathway allows for the scrambling of the
product and recovery of the starting material unscrambled.
However the assumption that the “anil” is formed irreversibly
is contrary to the conventional wisdom. Moreover, the frag-
mentation of 31 into imines 32 and 33 in a reversible reaction
requires the observation of other quinoline products. In addition,
there is no obvious reason that 34 undergoes cyclization to form
a quinoline, if 31 cannot. Thus, irreversible formation of anil
30 is discarded on the basis of these two inconsistencies.28
Mechanism 2: Irreversible Fragmentation of the Anil. In
this proposal, the first step is a reversible condensation of the
aniline 1 with pulegone to form anil 30, Scheme 19. The anil
then undergoes a conjugate addition of another molecule of 1
in a reversible reaction to form 31. This intermediate is then
proposed to undergo an irreVersible fragmentation into the
corresponding imine fragments 32 and 33. Although this would
explain the lack of crossover in the recovered pulegone (and
mesityl oxide), the irreVersible fragmentation of 31 is illogical
because the recombination is required to explain the formation
of scrambled (for mesityl oxide) and rearranged (for pulegone)
quinoline products. Therefore, this mechanism can be discarded.
Mechanism 3: Irreversible Formation/Fragmentation of
a Conjugate Adduct. The third mechanism proposed in Scheme
20 is based on the isolation of similar intermediates by Badger
et al.12 The first step is an reversible conjugate addition of aniline
to the R,â-unsaturated ketone (pulegone). Although conjugate
addition of amines can be irreversible29 (see below), the
formation of a tertiary center makes this less favorable.30 This
amine intermediate 36 undergoes an irreversible fragmentation
to form the acetone imine 32 and 3-methylcyclohexanone.
Recombination of these two components explains the results
of the labeling experiments and observation of a rearranged
Skraup product.31 The higher reactivity of the acetone imine as
(28) A modification of this sequence wherein the formation of the anil
30 is reversible and the conjugate addition to form 31 is irreversible will
lead to the same labeling results and will have the same logical inconsisten-
cies.
(29) Jung, M. E. In ComprehensiVe Organic Synthesis; Trost, B. M.,
Fleming, I., Eds.; Pergamon Press: New York, 1991; Vol. 4; Chapter 1.1.
(30) A recent study of the conjugate addition of nitrogen nucleophiles
to unsaturated ketones catalyzed by Bronsted acids shows that the addition
is reversible at room temperature. Wabnitz, T. C.; Spencer, J. B. Org. Lett.
2003, 5, 2141-2144.
(25) Sterling, E. C.; Bogert, M. T. J. Org. Chem. 1939, 4, 20-28.
(26) (a) Farina, V.; Krishnamurthy, V.; Scott, W. J. Org. React. 1997,
50, 1-652. (b) Banwell, M. G.; Cameron, J. M.; Collis, M. P.; Crisp, G.
T.; Gable, R. W.; Hamel, E.; Lambert, J. N.; Mackay, M. F.; Reum, M. E.;
Scoble, J. A. Aust. J. Chem. 1991, 44, 705-728.
(31) This mechanism is also allowed in the case of mesityl oxide because
of a similar scrambling of labels in the product 3.
(27) Hock, H.; Kropf, H. Chem. Ber. 1956, 89, 2436-2438.
1674 J. Org. Chem., Vol. 71, No. 4, 2006

Skraup-Doebner-Von Miller Quinoline Synthesis
SCHEME 20
a nucleophile in the condensation with 3-methylcyclohexanone
explains the formation of a single product (albeit in low yield)
and is in line with the reactivity of various enamines with
aldehydes and ketones.32 Finally, the new anil 37 suffers a
conjugate addition with another molecule of aniline 1 and
condenses to via 34 form quinoline 22. A modification of this
sequence wherein the first step is irreversible and the second
reversible will lead to the same labeling results.33
pathway. Our studies using labeled ketones shows that the R,â-
unsaturated ketones undergo fragmentation into their corre-
sponding ketone components. The aniline adds to the R,â-
unsaturated ketone initially in a conjugate fashion, followed by
a fragmentation to the corresponding imine and one of the
component ketones. These fragments recombine to form an anil
which leads to the quinoline product by conjugate addition of
a second molecule of aniline followed by cyclization. We are
presently considering studies to further distinguishing the
mechanisms proposed above. We are also exploring the methods
of making quinolines derived from unsymmetrical ketones in
accordance with our proposal.
The foregoing mechanism rationalizes the observation of
scrambled (and rearranged) quinoline products by starting from
a direct conjugate addition. Mechanisms involving initial
formation of anils can also be invoked but must still account
for the lack of scrambling of the label in the recovered ketones.
From the foregoing analysis, it would appear that neither the
Tung nor the Walter mechanisms are operative. However,
Tung’s demonstration that the preformed anil of mesityl oxide
forms the quinoline product in the presence of additional aniline
does show that intermediates such as 33 or 35 are competent.
The problem lies in explaining the unique formation of 22 from
Walter’s experiments (in low yield). To resolve these incon-
sistencies, crossover experiments should be conducted on the
preformed anil 34, thus precluding the intermediacy of 30 in
mechanism 1. If no crossover is observed, then the Walter
mechanism may be operative. However, this is of questionable
relevance to the Skraup-Doebner-Von Miller reaction if the
mechanism changes, depending upon the starting material used.
Our current view favors mechanism 3 shown in Scheme 20.
This mechanism accounts for all the labeling results, can explain
the unique formation of a rearranged Skraup product from
pulegone, and is precedented in the studies by Badger et al.
This mechanism can also explain the small amount of scram-
bling observed in the recovered mesityl oxide but not in the
recovered pulegone. This leakage must arise from self-
condensation of acetone formed from the fragmentation of the
conjugate adduct (related to 36, Scheme 20), whereas the
3-methylcyclohexanone must revert back to pulegone by addi-
tions to imines. Self-condensation leads to a different product.
Experimental Section
General Experimental Procedures. See the Supporting Infor-
mation.
1,2-Dihydro-6-isopropyl-2,2,4-trimethylquinoline. A solution
of 4-isopropylaniline (500 mg, 3.7 mmol), mesityl oxide (725 mg,
7.3 mmol, 2.0 equiv), and iodine (10 mg, 0.037 mmol 1 mol %) in
2 mL of dry toluene was heated at reflux under N2 for 5 h. The
dark brown reaction mixture was concentrated in vacuo at 50 °C,
and the crude residue was purified by silica gel chromatography
(hexanes/EtOAc, 24/1) to yield the quinoline product as a clear,
pale yellow liquid (303 mg, 38%).
Alternate Procedure. A solution of 4-isopropylaniline (2.0 g,
14.8 mmol) and iodine (45 mg, 0.18 mmol, 1.25 mol %) in a three-
necked, round-bottom flask was attached with a Vigreux column
and a distillation setup. Acetone (10 mL, 7.91 g, 136.2 mmol, 9.2
equiv) was introduced dropwise over 2 h through a second neck
with a dropping funnel while the internal temperature was
maintained at 150 °C. The unreacted acetone distilled out. Once
the addition of acetone was complete, the reaction mixture was
concentrated in vacuo and the reaction mixture was purified by
silica gel chromatography (hexanes/EtOAc, 24/1) to afford the
quinoline product as a colorless liquid. The chromatographed liquid
was further purified by Kugelrohr distillation (1.8 g, 60%). A
substantial portion of the quinoline product polymerized on
distillation: bp 180 °C (0.01 mmHg, ABT); Rf 0.68 (hexane/EtOAc,
1
7/2); H NMR (400 MHz) 6.93 (d, J ) 1.9, 1 H, Ar-HC(7)), 6.87
(dd, J ) 2.0, 7.9, 1 H, Ar HC(9)), 6.40 (d, J ) 8.0, 1 H, ArHC-
(10′)), 5.31 (d, J ) 0.81, HC(3)), 3.6 (br s, 1 H, HN(1)), 2.80 (s,
J ) 6, 1 H, HC(14)), 2.01 (d, 3 H, H3C(13)), 1.28 (s, 6 H, 2 ×
H3C(12, 13)), 1.22 (d, J ) 6.8, 6 H, 2 × H3CC(15,16)); 13C NMR
(126 MHz) 141.24 (C(6)), 137.5 (C(8)), 128.6 (C(3)), 128.4 (C(4)),
126.1 (C(9)), 121.7 (C(7)), 121.3 (C(5)), 112.8 (C(10)), 51.8 (C(2)),
33.5 (C(14)), 31.1 (C(11,12)), 24.3 (C(15, 16)), 18.7 (C(13)); IR
(neat) 3450 (s), 2958 (s), 2932 (s), 2872 (s), 1709 (s), 1492 (m),
1465 (m), 1406 (m), 1380 (m), 1328 (m), 1127 (m), 1097 (m),
1075 (m), 1011 (s), 750 (s), 700 (s); MS (EI, 70 eV) 200 ([M -
CH3]+, 11), 168 (15), 158 (5), 141 (20), 128.1 (15), 115 (25), 91
(23), 77 (100), 65 (19); FI MS (150 °C) (M+, 100). Anal. (C15H21N
Conclusions
The condensation of aniline derivatives with R,â-unsaturated
ketones to form quinoline follows a complex mechanistic
(32) This also explains the lack of self-condensation of the ketone product
which would give rise to scrambling of the label in mesityl oxide. (a)
Cervinka, O. In The Chemistry of Enamines; Rappoport, Z., Ed.; Wiley:
New York, 1994; Part 1; Chapter 9. (b) Alt, G. H.; Cook, A. G. In Enamines,
2nd ed.; Cook, A. G., Ed.; Dekker: New York, 1988; Chapter 4.
(33) We prefer the former proposal wherein the fragmentation is
irreversible because that also helps explain why 36 does not form imine 31
and cyclize to a normal Skraup product.
J. Org. Chem, Vol. 71, No. 4, 2006 1675

Denmark and Venkatraman
(215.15)) Calcd: C, 83.73; H, 9.76; N, 6.50. Found: C, 83.73; H,
9.79; N, 6.39.
Data for 22b: Rf ) 0.47 (hexane/ether, 6/1); [R]D +32 (c )
1.52, CHCl3); 1H NMR (500 MHz, methanol-d4) 6.86 (d, 1 H, J )
1.8 Hz), 6.83 (dd, 1 H, J ) 2.0, 8.2 Hz), 6.50 (d, 1 H, J ) 8.0 Hz),
5.72 (d, 1 H, J ) 1.4 Hz), 2.73 (h, 1 H, J ) 6.9 Hz), 1.97 (s, 3 H),
1.99-1.95 (m, 2 H), 1.65-1.64 (m, 2 H), 1.57-1.54 (m, 2 H),1.19
(d, 6 H, J ) 7.1 Hz), 1.20-1.13 (m, 1 H), 0.84 (d, 3 H, J ) 6.4
Hz), 0.87-0.82 (m, 2 H); 13C NMR (126 MHz, methanol-d4) 143.3,
139.2, 130.8, 127.1, 126.7, 124.0, 122.4, 114.9, 54.6, 39.7, 35.8,
34.8, 30.1, 29.8, 23.1, 22.7, 19.2; IR (KBr, cm-1) 3362 (m), 3019
(m), 2953 (s), 2949 (s), 2939 (s), 2923 (s), 2919 (s), 2888 (s), 2865
(s), 2843 (s) 1645 (s), 1610 (s), 1581 (s), 1499 (s), 1461 (s), 1455
(s), 1379 (s), 1361 (s), 1347 (s), 1331 (s), 1290 (s), 1252 (m), 1244-
(m), 1219 (m), 1203 (m), 1190 (m) 1178 (m), 1151 (m), 1106 (s)
1069 (s), 1055 (s), 954(m), 884 (m), 810 (m), 646 (w), 573 (m);
MS (FI) 269 (M+, 100). Anal. (C19H27N, 269.42) Calcd: C, 84.70;
H, 10.10; N, 5.20. Found: C, 84.82; H, 9.94; N, 5.30.
Synthesis of Quinolines 22a and 22b: (3R)-6′-Isopropyl-3,4′-
dimethylspiro[cyclohexane-1,2′(1′H)-quinoline]. A solution of
4-isopropylaniline (0.50 g, 3.70 mmol), (R)-pulegone (0.70 g, 4.625
mmol, 1.25 equiv, technical grade), and iodine (19 mg, 0.08 mmol
2 mol %) was heated to reflux in toluene (5.0 mL) for 13 h. The
dark brown solution was concentrated in vacuo, and the residue
was purified by silica gel chromatography (hexanes/ether 6/1) to
provide 206 mg (21%) of 22a and 200 mg (20%) of 22b as pale-
yellow, viscous oils. Data for 22a: Rf ) 0.69 (hexane/ether, 6/1);
[R]D -51.3 (c ) 1.05, CHCl3); 1H NMR (500 MHz, methanol-d4)
6.87 (d, 1 H, J ) 2.0 Hz), 6.82 (dd, 1 H, J ) 2.0, 8.0 Hz), 6.59 (d,
1 H, J ) 8.0 Hz), 5.18 (d, 1 H, J ) 1.5 Hz), 2.74 (h, 1 H, J ) 7.0
Hz), 1.95 (d, 3 H, J ) 1.5 Hz), 1.91-1.85 (m, 2 H), 1.69-1.52
(m, 4 H), 1.17 (d, 6 H, J ) 6.9 Hz), 1.14-1.11 (dq, 1 H, J ) 4.4,
13.7 Hz), 0.85 (d, 3 H, J ) 6.4 Hz), 0.87-0.80 (m, 2 H); 13C NMR
(126 MHz, methanol-d4) 141.3, 137.7, 129.9, 128.9, 125.5, 122.6,
120.8, 113.6, 53.1, 45.7, 36.1, 34.0, 33.4, 26.3, 23.4, 21.8, 20.0,
17.4; IR (KBr, cm-1) 3409 (m), 3018 (m), 2951 (s), 2940 (s), 2921
(s), 2918 (s), 2885 (s), 2866 (s), 2844 (s), 1650 (m), 1610 (s), 1582
(s), 1500 (s), 1455 (s), 1418 (m), 1378 (m), 1361 (m), 1331 (m),
1285 (s), 1264 (m), 1203 (m), 1178 (m), 1151 (m), 1053 (m), 1046
(m), 958 (s), 885 (m), 809 (m), 645 (w), 533 (w); MS (FI) 269
(M+, 100). Anal. (C19H27N, 269.42) Calcd: C, 84.70; H, 10.10; N,
5.20. Found: C, 84.56; H, 9.89; N, 5.08.
Acknowledgment. We are grateful to LIGAND Pharma-
ceuticals for financial support. We also thank Stephen Mullen
for recording the mass spectra.
Supporting Information Available: Preparation and charac-
terization of labeled 2 and 16 along with mass spectral analysis of
all cross over and control experiments. This material is available
free of charge via the Internet at http://pubs.acs.org.
JO052410H
1676 J. Org. Chem., Vol. 71, No. 4, 2006