Catalyst and pressure dependent reductive cyclizations for the diastereoselective synthesis of hexahydropyrrolo[1,2-a]quinoline-5-carboxylic esters

Article abstract of DOI:10.1002/jhet.5570430613

A diastereoselective synthesis of 1,2,3,3a,4,5-hexahydropyrrolo[1,2-a] quinoline-5-carboxylic esters has been developed using a tandem reduction-double reductive amination reaction. The nitro dicarbonyl cyclization substrates were synthesized by alkylatio

Full text of DOI:10.1002/jhet.5570430613

Nov-Dec 2006
1505
Catalyst and Pressure Dependent Reductive Cyclizations for the Diastereo-
selective Synthesis of Hexahydropyrrolo[1,2-a]quinoline-5-carboxylic Esters
Richard A. Bunce*, James E. Schammerhorn [1] and LeGrande M. Slaughter
Department of Chemistry, Oklahoma State University, Stillwater, OK 74078-3071, USA
e-mail: rab@okstate.edu
Received December 19, 2005
CO2CH3
CO2CH3
1) ozonolysis
2) catalytic hydrogenation
N
O2N
R
R
A diastereoselective synthesis of 1,2,3,3a,4,5-hexahydropyrrolo[1,2-a]quinoline-5-carboxylic
esters has been developed using a tandem reduction-double reductive amination reaction. The
nitro dicarbonyl cyclization substrates were synthesized by alkylation of methyl (2-nitro-
phenyl)acetate with 2-bromomethyl-1,5-hexadiene derivatives, followed by ozonolysis. Catalytic
hydrogenation of each substrate gave the target heterocycle, along with a deacylated product and
an adduct resulting from capture of the intermediate hydroxylamine by the side chain carbonyls.
The product ratio varied dramatically with the catalyst and the hydrogen pressure. Cyclization to
the title compounds was highly diastereoselective, producing each hexahydropyrrolo[1,2-a]-
quinoline as a single stereoisomer with the all-cis geometry. The competing processes have not
been observed in previous heterocyclization studies but can be attributed to greater strain in the
system, which slows the final ring closure.
J. Heterocyclic Chem., 43, 1505 (2006).
Introduction.
through a hexahydropyrrolo[1,2-a]quinoline intermediate,
e.g. 1 [4,5]. Since gephyrotoxin and other neurotoxins have
pharmacological value in the study of ion channel blockers,
it is important to have synthetic access to these compounds.
Two previous syntheses of hexahydropyrrolo[1,2-a]-
quinolines have been reported as part of efforts to prepare
gephyrotoxin [4,5]. These were based on a Diels-Alder cycli-
zation of an o-quinone methide N-substituted imine [4] and an
intramolecular Schmidt reaction of an azide with a benzo-
fused cyclopentyl carbocation [5]. Two additional syntheses
of the unsubstituted ring system have involved metallation
and alkylation of an N-formamidoyl-activated tetrahydro-
quinoline in the presence of pentynyl copper [6] and transition
metal-promoted intramolecular coupling of an amine with an
aryl halide [7]. All of these methods required elaborate
synthetic schemes to prepare the cyclization substrates or
special catalysts to effect ring closure. Additionally, the
unsubstituted cases provided little information on selectivity
in the ring closure or tolerance to other functionality. We
report here a simple, efficient and diastereoselective tandem
reduction-double reductive amination approach to the
synthesis of substituted 1,2,3,3a,4,5-hexahydropyrrolo[1,2-
a]quinolines using readily available precursors.
We previously reported an efficient tandem reduction-
double reductive amination route to 2,3,4,4a,5,6-hexa-
hydro-1H-benzo[c]quinolizine-6-carboxylic esters [2].
Closure of this angular 6-6-6 tricyclic system was found
to be highly diastereoselective for the production of the
all-cis isomer. Furthermore, the reaction was relatively
clean, giving only two minor products from side reactions
in closing the second ring. In the current project, we had
an interest in extending the scope of this transformation to
the preparation of the corresponding 1,2,3,3a,4,5-hexa-
hydropyrrolo[1,2-a]quinoline-5-carboxylic esters.
Of
particular interest was the comparative yield and
selectivity of the reaction for the synthesis of the more
highly strained 6-6-5 ring system [3].
The target heterocycle is relatively rare in nature. One
notable example, however, is the alkaloid gephyrotoxin (2),
a non-competitive blocker of neuromuscular transmission.
While 2 incorporates the fully reduced 6-6-5 cyclic
framework, one reported synthesis (Equation 1) proceeds
OCH3
H
15 steps
H
H
Results and Discussion.
(1)
N
N
Ref 4
H
The synthesis of the cyclization substrates is
summarized in Scheme 1. Bromodiene 4a was prepared
by alkylating the dianion of methallyl alcohol with allyl
HO
HO
H
H
1
2
\

1506
R. A. Bunce, J. E. Schammerhorn and L. M. Slaughter
Vol 43
bromide and converting the resulting alcohol to the
bromide with phosphorus tribromide [8]; bromodiene 4b
was prepared by the same method but used methallyl
bromide as the alkylating agent. Alkylation of the anion
of methyl (2-nitrophenyl)acetate (3) with 4a and 4b [9]
gave the nitro dienoic esters 5a and 5b, respectively.
Ozonolysis of these dienes followed by reductive workup
then yielded a mixture of the dicarbonyl products and
their methanol acetals [2]. Treatment of each mixture
with 3% aqueous perchloric acid in tetrahydrofuran (1:1
v/v) [10] gave the required nitro dicarbonyl substrates.
The results of our reductive cyclization studies are
shown in Scheme 2 [11]. All reactions were done in a
stainless steel reactor using a Paar shaker. In our initial
experiment with 6a, the reactor was evacuated and
flushed with hydrogen three times prior to starting the
shaker. After 3 hours under 3 atmospheres of hydrogen, a
modest 36% yield of the hexahydropyrrolo[1,2-a]-
quinoline
7
was isolated, along with 29% of
tetrahydroquinoline 8 resulting from single ring closure
and aldehyde decarbonylation [12] and 3% of a fused
cyclic hemiacetal product 9 derived from capture of the
intermediate hydroxylamine by the side chain carbonyls.
Substrate 6b followed a similar path giving 10, 8 and 11,
although the deacylation required to produce tetrahydro-
quinoline 8 was much less efficient.
Scheme 1
CO2CH3
Br
CO2CH3
Since palladium-on-carbon has often been observed
to decarbonylate aldehydes [13,14], we sought to
minimize this pathway by using platinum-based
catalysts. With platinum oxide (and also 5% platinum-
on-carbon for 6a), the decarbonylation and deacylation
processes were completely suppressed, but yields of the
desired tricyclic products were lower and much larger
proportions of the hemiacetals were observed. This
suggests that platinum catalysts are less effective in
reducing aromatic nitro groups [15], especially in the
final stage of the reaction where the N-O bond of the
hydroxylamine must be cleaved. This slower reduction
allows for trapping of the intermediate hydroxylamine
by the side chain carbonyls.
K2CO3
+
18-crown-6
CH3CN
NO2
O2N
R
84-88%
R
3
4a (R = H)
4b (R = CH3)
5a (R = H)
5b (R = CH3)
CO2CH3
1) O3, CH3OH, ꢀ78o
2) (CH3)2S, H+, ꢀ78o ꢁ 20o
O
O2N
3) 3% aqueous HClO4:THF (1:1 v/v)
90-95%
O
R
The disappointing results with platinum catalysts led
us to further investigate the use of 5% palladium-on-
carbon. Our first experiment evaluated the use of lower
6a (R = H)
6b (R = CH3)
Scheme 2
CO2CH3
CO2CH3
CO2CH3
H2
6a
+
+
CH3OH
N
N
H
N
O
catalyst
OH
7
8
9
5% Pd/C (3 atm, purged)
36%
20-25%
29%
29%
0%
3%
48-53%
3%
PtO2 or 5% Pt/C (3 atm, purged)
5% Pd/C (1 atm, purged)
37%
5%
5% Pd/C (5 atm, not purged)
66%
3%
CO2CH3
CO2CH3
CO2CH3
H2
+
+
6b
CH3OH
N
N
H
N
O
catalyst
OH
10
8
11
4%
5% Pd/C (3 atm, purged)
PtO2 (3 atm, purged)
58%
34%
48%
64%
7%
0%
36%
4%
5% Pd/C (1 atm, purged)
5% Pd/C (5 atm, not purged)
16%
4%
3%

Nov-Dec 2006
Hexahydropyrrolo[1,2-a]quinoline-5-carboxylic Esters
1507
hydrogen pressure for the reaction of 6a. Under 1
atmosphere of hydrogen, the yield of tricyclic 7 decreased
to 29% while the yield of decarbonylation product 8
increased to 37%; the proportion of hemiacetal 9
remained nearly constant at 3%. Substrate 6b gave
similar results. Thus, reduced pressures seem to favor the
decarbonylation process.
Based upon this finding, we realized that purging the
reactor prior to initiating the reaction (as was done in the
initial experiments) exposed the substrates to low
pressures of hydrogen in the presence of the catalyst and
perhaps reduced the yield of 7 and 10. The procedure
was, therefore, modified to minimize exposure of the
substrate to these conditions. After placing 6a, solvent
and catalyst in the reactor, the system was evacuated
once, shaking was started and the hydrogen pressure was
rapidly increased to 5 atmospheres. Assay of the products
from this run showed a much higher 66% yield of 7 with
only 5% of 8 and 3% of 9. A similar improvement was
also noted in the cyclization of 6b. Thus, the desired
tricyclic product appears to be favored at more elevated
hydrogen pressures with 5% palladium-on-carbon.
The mechanism of the ring closure to form the 6-6-5
tricyclic system is the same as that previously proposed
for the formation of the 6-6-6 structure (see Scheme 3).
Reduction of the aromatic nitro compound to aniline 12 is
followed by double reductive amination, first with the
closest carbonyl to yield 14 and then with the more distant
carbonyl to produce the hexahydropyrrolo[1,2-a]quinoline
derivative 7 or 10. The all-cis geometry of the product is
dictated by the carboxylic ester group in 13 and 15, which
blocks one face of the molecule forcing hydrogen to add
from the opposite side of the structure for both reductive
aminations. The shape of the molecule following the first
reductive amination also contributes to the selectivity in
the reduction of 15.
The process leading to decarbonylation of aldehyde 6a
has substantial precedent in the literature [13,14] (see
Scheme 4). Following initial cyclization to tetrahydro-
quinoline 14a [12], oxidative addition of the side chain
aldehyde to palladium on the surface of the catalyst would
form the acyl palladium complex 16. Migratory de-
insertion of carbon monoxide would then give the alkyl
palladium complex 17 and reductive elimination would
release the decarbonylated tetrahydroquinoline 8.
Normally, aldehyde decarbonylation requires heating at
150-200˚ [13,14], but in the presence of a high
concentration of hydrogen on the catalyst surface, the
process can apparently proceed at much lower
temperature.
Scheme 4
O
O
migratory
Pd
THQ
THQ
Pd
H
H
oxidative
addition
de-insertion
14a (R = H)
16
reductive
CO
THQ
THQ
Pd
H
elimination
ꢀPd-CO
H
17
8
THQ = 2-(1,2,3,4-tetrahydroquinolyl)
There is much less precedent for the deacylation of
ketone 6b. A possible mechanism, however, would
involve initial 1,2-insertion of 14b into a Pd-H bond to
generate oxygen-bound intermediate 18, followed by ꢀ-
methyl elimination to give aldehyde 14a and decarbon-
ylation as described above (see Scheme 5).
Scheme 3
CO2CH3
3 H2
6a (R = H)
6b (R = CH3)
catalyst
ꢀH2O
O
O
Scheme 5
H2N
ꢀ2 H2O
Pd
O
H
O
R
12
Pd-H
THQ
THQ
1,2-insertion
O
OCH3
O
OCH3
H2
14b (R = CH3)
ꢀ-CH3
18
decarbonylation
N
N
catalyst
ꢀH2O
O
H
O
O
H
R
8
R
H
THQ
H
H
elimination
13
14
ꢁPd-CH3
14a (R = H)
THQ = 2-(1,2,3,4-tetrahydroquinolyl)
O
OCH3
H2
7 (R = H)
10 (R = CH3)
N
catalyst
R
H
Finally, the hemiacetal product results from partial
reduction of 6 to 19 and capture of the hydroxylamine by
the side chain carbonyls (see Scheme 6). Cyclization of the
H
H
15

1508
R. A. Bunce, J. E. Schammerhorn and L. M. Slaughter
Vol 43
ground to a fine powder, dried under vacuum at 120˚ for 24
hours and stored in an oven at 120˚. Reactions were monitored
by thin layer chromatography on silica gel GF plates (Analtech
21521) with ultraviolet detection. Preparative separations were
performed by one of the following methods: (1) flash column
chromatography [21] on silica gel (grade 62, 60-200 mesh)
containing ultraviolet-active phosphor (Sorbent Technologies
UV-5) packed into quartz columns or (2) preparative thin layer
chromatography on 20-cm x 20-cm silica gel GF plates
hydroxylamine on the proximal carbonyl and loss of water
would give enamine 20 [2]. Reduction of the enamine
from the side opposite the ester would then generate the
ring closed N-hydroxytetrahydroquinoline 21 having the
side chain cis to the ester. Finally, attack of N-hydroxy
group on the second side chain carbonyl would produce the
hemiacetal. The origin of selectivity in the hemiacetal
closure is unclear. The equatorial OH in 9 was confirmed
by a single crystal X-ray structure [16] and suggests that a
steric preference rather than an anomeric effect [17] guides
the ring closure. The axial OH in 11 was deduced from the
NOESY spectrum [18], which showed no interaction of the
C-1 methyl with any other protons in the molecule; this is
only possible if the methyl occupies an equatorial position.
The observed stereochemistry in 11 would also be expected
on steric grounds [19,20], but such an argument would not
predict this to be the exclusive product.
(Analtech 02015).
Band elution for both methods was
monitored using a hand-held ultraviolet lamp. Melting points
were uncorrected. Infrared spectra were run as thin films on
sodium chloride disks and referenced to polystyrene. 1H and 13
C
Nuclear magnetic resonance spectra were measured in
deuteriochloroform at 300 MHz and 75 MHz, respectively,
using tetramethylsilane as the internal standard; coupling
constants (J) are given in Hertz. Mass spectra (electron
impact/direct probe) were obtained at 70 electron volts.
2-(Bromomethyl)-1,5-hexadiene (4a).
This compound (9.57 g, 41%) was prepared in two steps from
20 g (0.278 moles) of methallyl alcohol using the method
described by Trost and Shi [8]. The physical and spectral
properties matched those reported [8].
Scheme 6
CO2CH3
2 H2
6a (R = H)
6b (R = CH3)
2-(Bromomethyl)-5-methyl-1,5-hexadiene (4b).
catalyst
ꢀH2O
O
O
HN
HO
ꢀH2O
Methallyl alcohol was converted to its dianion on a 0.278-
mole scale as described by Trost and Shi [8] and alkylated with
28.1 g (0.208 moles) of methallyl bromide to give 17.0 g (65%)
of 2-methylene-5-methyl-5-hexen-1-ol, bp 70-72˚ at 3 mm Hg;
ir: 3386, 3072, 1647, 906 cm-1; 1H nmr: ꢀ 5.04 (m, 1H), 4.90
(m, 1H), 4.73 (m, 1H), 4.70 (m, 1H), 4.09 (s, 2H), 2.20 (m, 2H),
1.74 (d, 3H, J = 1.1), 1.61 (br s, 1H); 13C nmr: ꢀ 148.7, 145.4,
110.1, 109.5, 65.9, 35.9, 31.2, 22.4; ms: m/z 126 (M+).
Anal. Calcd. for C8H14O: C, 76.19; H, 11.11. Found: C,
76.06; H, 11.20.
R
19
O
OCH3
O
OCH3
(R = H)
(R = CH3)
9
11
N
N
O
HO
O
HO
H
H
H
R
R
20
21
Reaction of this alcohol with 18.3 g (6.36 mL, 0.068 moles)
of phosphorus tribromide in ether at 0˚ containing 1.0 g of
pyridine as outlined for 4a [8] gave 13.8 g (54%) of 4b as a light
yellow oil, bp 52˚ at 4 mm Hg; ir: 3078, 1643, 906, 889 cm-1;
1H nmr: ꢀ 5.18 (m, 1H), 4.98 (m, 1H), 4.75 (m, 1H), 4.72 (m,
1H), 3.99 (s, 2H), 2.38 (t, 2H, J = 7.6), 2.19 (t, 2H, J = 7.6), 1.75
(s, 3H); 13C nmr: ꢀ 145.1, 144.9, 115.1, 110.4, 36.8, 35.5, 31.3,
22.4; ms: m/z 188/190 (M+/M++2, ca.1:1).
Conclusions.
Our earlier tandem reduction-double reductive
amination used to prepare hexahydro-1H-benzo[c]-
quinolizines has been successfully adapted for the
preparation of the more strained hexahydropyrrolo[1,2-a]-
quinoline framework. The reaction was found to exhibit
the same high diastereoselectivity for the all-cis product,
but added strain in the smaller ring system slowed the
cyclization sufficiently to permit significant side
reactions. These competing processes were largely
circumvented by careful adjustment of catalyst and
hydrogen pressure and useful yields of the desired
Anal. Calcd. for C8H13Br: C, 50.79; H, 6.88. Found: C, 51.08;
H, 6.96.
Representative Procedure for Alkylation of Methyl (2-Nitro-
phenyl)acetate: Methyl (±)-4-Methylene-2-(2-nitrophenyl)oct-7-
enoate (5a).
The general procedure of Makosza and Tyrala was used [9].
To a stirred solution of 0.98 g (5.00 mmoles) of 3 in 25 mL of
dry acetonitrile was added 10 mg of 18-crown-6 and 5.8 g (42.0
mmoles) of anhydrous potassium carbonate. To the resulting
blue mixture was added a solution of 1.05 g (6.00 mmoles) of 4a
in 5 mL of dry acetonitrile. The reaction was stirred at 60˚ for 8
hours, then cooled to room temperature. The solids were
removed by filtration and the filtrate was concentrated under
vacuum. The remaining oil was purified by flash column
chromatography on a 25 cm x 2 cm silica gel column eluted with
increasing concentrations of ether in hexanes to give 1.27 g
tricyclic products were obtained.
underway to identify other reductive cyclizations that can
be optimized by tuning the reaction conditions.
Further work is
EXPERIMENTAL
All reactions (except hydrogenations) were run under dry
nitrogen in oven-dried glassware. Potassium carbonate was

Nov-Dec 2006
Hexahydropyrrolo[1,2-a]quinoline-5-carboxylic Esters
1509
(88%) of 5a as a light yellow oil. ir: 1738, 1643, 1528, 1351,
1
(m, 4H), 2.17 (s, 3H); 13C nmr: ꢀ 206.8, 206.0, 172.1, 148.6,
133.4, 133.2, 130.9, 128.4, 125.2, 52.4, 45.2, 42.7, 36.8, 35.9,
29.7; ms: m/z 307 (M+).
998, 909 cm-1; H nmr: ꢀ 7.87 (d, 1H, J = 8.2), 7.58 (m, 2H),
7.42 (m, 1H), 5.79 (ddt, 1H, J = 16.9, 10.4, 6.0), 5.02 (dm, 1H, J
= 16.9), 4.96 (dm, 1H, J = 10.4), 4.78 (d, 1H, J = 0.8), 4.71 (d,
1H, J = 0.8), 4.47 (dd, 1H, J = 8.2, 7.1), 3.66 (s, 3H), 2.88 (dd,
1H, J = 15.0, 8.5), 2.54 (dd, 1 H, J = 14.7, 7.1), 2.14 (m, 4H);
13C nmr: ꢀ 172.8, 149.6, 145.1, 138.0, 133.0, 132.9, 129.9,
128.1, 124.7, 114.8, 112.2, 52.3, 44.5, 39.2, 35.1, 31.8; ms: m/z
289 (M+).
Reductive Cyclization of 6a.
Caution: Though we never experienced any problems,
addition of 5% palladium-on-carbon to methanol can cause fires.
This operation should be performed under
atmosphere.
a nitrogen
Anal. Calcd. for C16H19NO4: C, 66.44; H, 6.57; N, 4.84.
Found: C, 66.61; H, 6.64; N, 4.66.
A solution of 500 mg (1.71 mmoles) of 6a in 125 mL of
methanol containing 100 mg of 5% palladium-on-carbon was
placed in a stainless steel pressure vessel in a Paar apparatus.
The vessel was evacuated once, shaking was initiated and the
apparatus was rapidly pressurized to circa 5 atmospheres (75
psi) with hydrogen gas. The reaction was continued for 3 hours
at 30˚. At the end of this time, hydrogen was purged from the
reactor and the crude reaction mixture was concentrated. The
residue was diluted with ether, and filtered through a pad of
Celite topped with a layer of anhydrous magnesium sulfate to
remove the catalyst. Removal of the ether gave a yellow oil that
was purified by preparative thin layer chromatography to give
three major bands. Band 1 contained methyl (±)-(3aR*,5S*)-
1,2,3,3a,4,5-hexahydropyrrolo[1,2-a]quinoline-5-carboxylate
(7); band 2 gave methyl (±)-(2R*,4S*)-2-ethyl-1,2,3,4-tetra-
hydroquinoline (8); band 3 yielded the cyclic hemiacetal 9
derived from methyl (±)-(2R*,4S*)-1-hydroxy-2-(3-oxopropyl)-
1,2,3,4- tetrahydroquinoline- 4-carboxylate.
Methyl (±)-7-Methyl-4-methylene-2-(2-nitrophenyl)oct-7-enoate
(5b).
This compound (1.25 g, 84%) was isolated as a light yellow
oil. ir: 1738, 1528, 1351, 892 cm-1; 1H nmr: ꢀ 7.87 (dm, 1H, J =
7.9), 7.57 (m, 2H), 7.42 (m, 1H), 4.78 (s, 1H), 4.71 (m, 2H),
4.68 (m, 1H), 4.48 (dd, 1H, J = 7.9, 6.8), 3.67 (s, 3H), 2.89 (ddd,
1H, J = 14.7, 8.2, 0.8), 2.56 (ddd, 1H, J = 14.7, 7.1, 0.8), 2.14 (s,
4H), 1.72 (s, 3H); 13C nmr: ꢀ 172.8, 149.8, 145.5, 145.3, 133.0,
132.9, 129.9, 128.1, 124.7, 112.0, 110.1, 52.3, 44.5, 39.2, 35.9,
34.0, 22.4; ms: m/z 303 (M+).
Anal. Calcd. for C17H21NO4: C, 67.33; H. 6.93; N, 4.62.
Found: C, 67.14; H, 6.99; N, 4.55.
Representative Procedure for Ozonolysis of the Dienes: Methyl
(±)-2-(2-Nitrophenyl)-4,7-dioxoheptanoate (6a).
A solution of 1.00 g (3.46 mmoles) of 5a in 125 mL of
methanol was treated with ozone at ꢁ78˚ until thin layer chroma-
tography indicated complete consumption of starting material.
Excess ozone was removed on a stream of dry nitrogen and 4.99
g (5.90 mL, 80.5 mmoles) of dimethyl sulfide and 100 mg of p-
toluenesulfonic acid were added. The mixture was gradually
warmed to room temperature and stirred for 6 hours. The
reaction was concentrated, diluted with ether, washed with
saturated sodium bicarbonate (two times) and saturated sodium
chloride (one time) and dried (magnesium sulfate). Removal of
the ether gave the keto acetal containing a small amount of the
keto aldehyde.
Methyl (±)-(3aR*,5S*)-1,2,3,3a,4,5-Hexahydropyrrolo[1,2-a]-
quinoline-5-carboxylate (7).
This compound (260 mg, 66%) was isolated as a white solid,
mp 73-75˚; ir: 1736 cm-1; 1H nmr: ꢀ 7.10 (tm, 1H, J = 8.2), 6.93
(d, 1H, J = 7.5), 6.57 (td, 1H, J = 7.5, 1.1), 6.42 (dd, 1H, J = 8.2,
1.1), 3.93 (dd, 1H, J = 12.3, 5.2), 3.77 (s, 3H), 3.41 (dddd, 1H, J
= 14.9, 10.4, 5.2, 3.0), 3.26 (m, 2H), 2.34 (ddd, 1H, J = 12.3,
5.2, 3.0), 2.10 (m, 2H), 1.91 (m, 1H), 1.82 (dt, 1H, J = 12.3,
11.5), 1.54 (m, 1H); 13C nmr: ꢀ 175.1, 144.4, 128.3, 127.0,
118.0, 115.2, 110.8, 56.3, 51.9, 46.9, 44.6, 32.9, 31.5, 23.6; ms:
m/z 231 (M+).
This mixture was dissolved in 40 mL of tetrahydrofuran and 40
mL of 3% aqueous perchloric acid was added dropwise at 0˚ [10].
The solution was stirred at 0˚ for 1 hour and at room temperature
for 4 hours then extracted with dichloromethane (two times). The
organic layer was washed with saturated sodium bicarbonate (two
times) and saturated sodium chloride (one time), dried
(magnesium sulfate) and concentrated under vacuum to give 0.91
g (90%) of 6a as a light yellow oil. This compound was used
without further purification. ir: 2840, 2725, 1737, 1715, 1528,
1351 cm-1; 1H nmr: ꢀ 9.78 (s, 1H), 7.98 (dd, 1H, J = 8.1, 1.6), 7.59
(td, 1H, J = 7.6, 1.4), 7.45 (td, 1H, J = 7.6, 1.6), 7.44 (d, 1H, J =
8.1), 4.72 (dd, 1H, J = 8.5, 4.9), 3.65 (s, 3H), 3.52 (dd, 1H, J =
17.7, 8.5), 2.92 (dd, 1H, J = 18.0, 4.9), 2.79 (m, 4H); 13C nmr: ꢀ
205.5, 200.2, 172.1, 150.1, 133.5, 133.2, 131.0, 128.5, 125.2, 52.5,
45.2, 42.9, 37.4, 34.6; ms: m/z 293 (M+).
Anal. Calcd. for C14H17NO2: C, 72.73; H, 7.36; N, 6.06.
Found: C, 72.88; H, 7.41; N, 5.93.
Methyl (±)-(2R*,4S*)-2-Ethyl-1,2,3,4-tetrahydroquinoline-4-
carboxylate (8).
This compound (19 mg, 5%) was isolated as a white solid, mp
1
84-85˚; ir: 3386, 1732 cm-1; H nmr: ꢀ 7.02 (tm, 1H, J = 7.6),
6.96 (dm, 1H, J = 7.9), 6.63 (td, 1H, J = 7.6, 1.4), 6.50 (dd, 1H, J
= 7.9, 1.1), 3.94 (dd, 1H, J = 12.0, 5.7), 3.79 (br s, 1H), 3.76 (s,
3H), 3.19 (dtd, 1H, 11.2, 6.3, 2.7), 2.20 (ddd, 1H, J = 12.8, 5.7,
2.7), 1.93 (ddd, 1H, J = 12.8, 12.0, 11.2), 1.56 (dq, 2H, J = 13.9,
7.5), 1.00 (t, 3H, J = 7.5); 13C nmr: ꢀ 175.0, 144.6, 128.1, 127.9,
118.1, 117.5, 114.7, 52.1, 52.0, 43.8, 32.0, 29.2, 9.8; ms: m/z
219 (M+).
Anal. Calcd. for C13H17NO2: C, 71.23; H, 7.76; N, 6.39.
Found: C, 71.40; H, 7.83; N, 6.26.
Methyl (±)-2-(2-Nitrophenyl)-4,7-dioxooctanoate (6b).
Methyl (±)-(2R*,4S*)-1-Hydroxy-2-(3-oxopropyl)-1,2,3,4-tetra-
hydroquinoline-4-carboxylate Cyclic Hemiacetal 9.
This compound (0.96 g, 95%) was isolated as a light yellow
oil. ir: 1739, 1715, 1528, 1351 cm-1; 1H nmr: ꢀ 7.97 (dd, 1H, J =
8.1, 1.4), 7.58 (td, 1H, J = 7.5, 1.6), 7.45 (td, 1H, J = 7.5, 1.4),
7.43 (d, 1H, J = 8.1), 4.71 (dd, 1H, J = 8.5, 4.9), 3.65 (s, 3H),
3.51 (dd, 1H, J = 17.7, 8.5), 2.92 (dd, 1H, J = 18.0, 4.9), 2.72
This compound (14 mg, 3%) was isolated as a light yellow oil
that crystallized on standing. Recrystallization from petroleum
ether gave a light yellow solid, mp 124-126˚; X-ray details [16];

1510
R. A. Bunce, J. E. Schammerhorn and L. M. Slaughter
Vol 43
1
ir: 3420, 1735 cm-1; H nmr: ꢀ 7.32 (dd, 1H, J = 8.4, 1.4), 7.18
(tm, 1H, J = 8.2), 7.03 (dt, 1H, J = 7.9, 1.1), 6.88 (td, 1H, J =
7.6, 1.4), 5.16 (ddd, 1H, J = 9.3, 6.8, 2.2), 3.99 (t, 1H, J = 8.8),
3.75 (s, 3H), 3.63 (br s, 1H), 3.05 (m, 1H), 2.21 (d, 1H, J = 8.8),
2.18 (m, 1H), 2.02 (dm, 1H, J = 12.0), 1.88 (dm, 1H, J = 10.6),
1.71-1.45 (complex, 2H); 13C nmr: ꢀ 174.3, 146.0, 128.0, 127.7,
121.8, 121.0, 117.1, 97.3, 56.4, 52.2, 43.1, 31.9, 31.8, 30.0; ms:
m/z 263 (M+).
Skinner Scholarship. Funding for the 300 and 400 MHz NMR
spectrometers of the Oklahoma Statewide Shared NMR Facility
was provided by NSF (BIR-9512269), the Oklahoma State
Regents for Higher Education, the W. M. Keck Foundation, and
Conoco, Inc. Partial funding for our mass spectrometer was
obtained from NSF and OCAST.
REFERENCES AND NOTES
Anal. Calcd. for C14H17NO4: C, 63.88; H, 6.46; N, 5.32.
Found: C, 63.92; H, 6.49; N, 5.26.
[1] Undergraduate Research Participant, 2004-2006.
[2] R. A. Bunce, D. M. Herron, L. B. Johnson and S. V. Kotturi,
J. Org. Chem., 66, 2822 (2001).
Reductive Cyclization of 6b.
This reaction was run on 500 mg (1.63 mmoles) of 6b as
described for 6a. Purification by preparative thin layer chroma-
tography gave three major bands. Band 1 contained methyl (±)-
(1R*, 3aR*, 5S*)- 1- methyl-1,2,3,3a,4,5-hexahydropyrrolo[1,2-
a]quinoline-5-carboxylate (10); band 2 gave methyl (±)-
(2R*,4S*)-2-ethyl-1,2,3,4-tetrahydroquinoline (8); band 3
yielded the cyclic hemiacetal 11 derived from methyl (±)-
(2R*,4S*)-1-hydroxy- 2- (3- oxobutyl)-1, 2, 3, 4-tetrahydroquino-
line-4-carboxylate.
[3a] The pyrrolidine ring is estimated to have 6.5 kcal/mole of
strain energy relative to piperidine, see K. B. Wiberg, D. Y. Najaki, and
K. M. Morgan, J. Am. Chem. Soc., 115, 3527 (1993); [b] J. D. Cox,
Tetrahedron, 19, 1175 (1963).
[4a] Y. Ito, E. Nakajo, M. Nakatsuka and T. Saegusa, Tetrahedron
Lett., 24, 2881 (1983); [b] R. Fujimoto, Y. Kishi and J. F. Blount, J.
Am. Chem. Soc., 102, 7154 (1980).
[5]
A
second synthesis proceeded through the same
hexahydropyrrolo[1,2-a]quinoline intermediate, see W. H. Pearson and
W. Fang, J. Org. Chem., 65, 7158 (2000).
Methyl (±)-(1R*,3aR*,5S*)-1-Methyl-1,2,3,3a,4,5-hexahydro-
pyrrolo[1,2-a]quinoline-5-carboxylate (10).
[6] A. I. Meyers and G. Milot, J. Org. Chem., 58, 6538 (1993).
[7a] For cyclization promoted by nickel(0): R. Omar-Amrani, A.
Thomas, E. Brenner, R. Schneider and Y. Fort, Org. Lett., 5, 2311
(2003); [b] For cyclization promoted by palladium(0): J. P. Wolfe, R.
A. Rennels and S. L. Buchwald, Tetrahedron, 52, 7525 (1996).
[8] B. M. Trost and Y. Shi, J. Am. Chem. Soc., 115, 9421 (1993).
[9] M. Makosza and A. Tyrala, Synth. Commun., 16, 419 (1986).
[10] T. Hudlicky, B. C. Ranu, S. M. Naqvi and A. Srnak, J. Org.
Chem., 50, 123 (1985).
[11] The modest mass balances (ca. 70%) for the reductive
cyclizations result from loss of material during chromatography.
[12] Decarbonylation could also occur before the initial ring
closure.
[13a] P. N. Rylander, Catalytic Hydrogenation Over Platinum
Metals; Academic Press: New York, 1967; pp 20-21 and 400; [b] J.
Tsuji in "Organic Synthesis via Metal Carbonyls," I. Wender and P.
Pino, Eds.; Wiley-Interscience: New York, 1977; pp 595-654; [c] M.-A.
Pearsall, A. M. Rosan, J. S. Conrad, C. A. Hendrickson, A. L. Pacchia
and D. J. Schantz, J. Chem. Educ., 72, A29 (1995).
This compound (256 mg, 64%) was isolated as a light tan oil
that darkened on standing. Prolonged storage at 0˚ gave tan
1
crystals, mp 62-64˚; ir: 1736 cm-1; H nmr: ꢀ 7.08 (tm, 1H, J =
8.2), 6.99 (dm, 1H, J = 7.6), 6.55 (td, 1H, J = 7.6, 1.1), 6.49 (d,
1H, J = 8.2), 3.98 (dd, 1H, J = 12.3, 6.0), 3.84 (quintet, 1H, J =
6.6), 3.75 (s, 3H), 3.40 (m, 1H), 2.36 (ddd, 1H, J = 12.3, 6.0,
2.2), 2.02 (m, 2H), 1.87 (dt, 1H, J = 12.3, 11.5), 1.70 (m, 2H),
1.17 (d, 3H, J = 6.3); 13C nmr: ꢀ 175.4, 143.7, 128.0 (2), 118.4,
115.1, 111.5, 57.5, 52.3, 52.0, 44.8, 33.0, 31.7, 30.3, 19.9; ms:
m/z 245 (M+).
Anal. Calcd. for C15H19NO2: C, 73.47; H, 7.76; N, 5.71.
Found: C, 73.58; H, 7.81; N, 5.61.
Methyl (±)-(2R*,4S*)-2-Ethyl-1,2,3,4-tetrahydroquinoline-4-
carboxylate (8).
[14] For a mechanistic discussion of this process, see J. Tsuji and
K. Ohno, J. Am. Chem. Soc., 90, 94 (1968).
This compound (15 mg, 4%) was isolated as a white solid.
The physical and spectral properties matched those given above.
[15] Poisoning of the catalyst surface by the tetrahydroquinoline
products may play a role in deactivating the platinum catalysts, see ref 8,
pp 16-17, 318 and 385.
Methyl (±)-(2R*,4S*)-1-Hydroxy-2-(3-oxobutyl)-1,2,3,4-tetra-
hydroquinoline-4-carboxylate Cyclic Hemiacetal 11.
[16] X-ray details for compound 9: Intensity data were measured
on a Bruker APEXII diffractometer with MoKꢁ radiation (ꢂ = 0.71073
Å) at 298(2) K. The structure was solved and refined using the Bruker
SHELXTL (Version 6.1) software package. Formula: C14H17NO4,
crystal size: 0.30 x 0.20 x 0.06 mm, formula weight: 263.29, a =
8.2069(2) Å, b = 12.2899(3) Å, c = 13.5240(4) Å, ꢃ = 103.026(2)˚, V =
1328.96(6) Å3, Z = 4, monoclinic space group P21/n, density (calculated)
= 1.316 Mg/m3, 1489 independent reflections, 1170 observed reflections
(I > 2.0(ꢄI)), R1 = 0.0335 (observed data), wR2 = 0.0957 (all data).
CCDC 290046 contains the supplementary crystallographic data for 9.
These data can be obtained free of charge from the CCDC at
www.ccdc.cam.ac.uk/data_request/cif.
This compound (14 mg, 3%) was isolated as a light yellow oil
1
that darkened on standing. ir: 3467, 1735 cm-1; H nmr: ꢀ 7.31
(dd, 1H, J = 8.2, 1.4), 7.20 (tm, 1H, J = 8.2), 7.04 (dt, 1H, J =
7.6, 1.1), 6.90 (td, 1H, J = 7.6, 1.4), 4.24 (br s, 1H), 4.03 (dd,
1H, J = 11.2, 6.8), 3.76 (s, 3H), 3.14 (m, 1H), 2.32-2.15
(complex, 2H), 1.89-1.75 (complex, 4H), 1.50 (s, 3H); 13C
NMR: ꢀ 174.1, 145.7, 128.1, 127.5, 122.2, 121.8, 117.6, 98.3,
56.5, 52.2, 43.1, 34.5, 32.4, 27.0, 26.0; ms: m/z 277 (M+).
Anal. Calcd. for C15H19NO4: C, 64.98; H, 6.86; N, 5.05.
Found: C, 65.09; H, 6.94; N, 4.93.
[17a] T. L. Gilchrist, Heterocyclic Chemistry; Longman: New
York, 1985; pp 45-48; [b] E. L. Eliel, S. L. Wilen and L. N. Mander,
Stereochemistry of Organic Compounds; Wiley: New York, 1994; pp
749-752; [c] M. B. Smith and J. March, March's Advanced Organic
Chemistry; Wiley: New York, 2001; pp 176-177; [d] F. A. Carey and R.
J. Sundberg, Advanced Organic Chemistry, Part A: Structure and
Mechanisms; Kluwer Academic/Plenum: New York, 2000; pp 151-156.
Acknowledgement.
The authors wish to thank the Oklahoma Center for the
Advancement of Science and Technology (OCAST, HR01-015)
for support of this research. J. E. S. also thanks the Oklahoma
State University College of Arts and Sciences for a Niblack
Research Scholarship and the Department of Chemistry for a

Nov-Dec 2006
Hexahydropyrrolo[1,2-a]quinoline-5-carboxylic Esters
1511
[18] D. Neuhaus and M. P Williamson, The Nuclear Overhauser
Effect in Structural and Conformational Analysis; VCH: New York,
1989; pp. 253-306.
[19] Molecular mechanics (UFF) and density functional theory (DFT)
calculations on optimized structures of the equatorial methyl and axial methyl
isomers did not indicate a clear enthalpic preference for either isomer. UFF
calculations [A. K. Rappé, C. J. Casewit, K. S. Colwell, W. A. Goddard, III
and W. M. Skiff, J. Am. Chem. Soc., 114, 10024 (1992)] favored equatorial
methyl by 1.47 kcal/mole, while DFT calculations favored axial methyl by
2.61 kcal/mole. DFT calculations were performed using the GAUSSIAN 03
package [M. J. Frisch, et al., Gaussian 03, Gaussian, Inc., Pittsburgh, PA
(2004)] at the B3LYP/6-31G** level of theory with full vibrational analysis
to derive the thermochemical results.
[20] The steric preference for equatorial CH3 is also predicted
from comparison of the conformational free energies of the CH3 and OH
groups, see ref 17d, p 140.
[21] W. C. Still, M. Kahn and A. Mitra, J. Org. Chem., 43, 2923
(1978).