Asymmetric synthesis of cis-fused bicyclic pyrrolidines and pyrrolidinones via chiral polycyclic lactams

Article abstract of DOI:10.1021/jo960185l

Condensation of racemic keto-ester 2 and 1.1 equiv of (R)-(-)-phenylglycinol in toluene with azeotropic removal of water and methanol gave rise to a tetracyclic lactam in greater than 90% yield. Examination of the crude reaction product by ¹H and ¹³C NMR, capillary GC, and HPLC revealed this product to be a single isomer, the absolute configuration of which was determined to be that illustrated by structure 3. The tetracyclic lactam 3 was stereospecifically reduced to either the cis-fused bicyclic pyrrolidine 4 or the cis-fused bicyclic pyrrolidinone 9. The chiral auxiliaries in both 4 and 9 were removed using different, novel methodologies. This highly stereoselective reaction establishes two contiguous chiral centers via the deracemization of an achiral keto-ester. This methodology has been applied to the synthesis of (1R,5R)-2-azabicyclo[3.3.0]octane 15. This important amine is now readily available from commercial materials in only three steps.

Full text of DOI:10.1021/jo960185l

J . Org. Chem. 1996, 61, 5813-5817
5813
Asym m etr ic Syn th esis of Cis-F u sed Bicyclic P yr r olid in es a n d
P yr r olid in on es via Ch ir a l P olycyclic La cta m s
Michael D. Ennis,* Robert L. Hoffman, Nabil B. Ghazal, David W. Old, and
Pamela A. Mooney
Structural, Analytical & Medicinal Chemistry, Pharmacia & Upjohn, Inc., Kalamazoo, Michigan 49001
Received J anuary 29, 1996 (Revised Manuscript Received J une 7, 1996X)
Condensation of racemic keto-ester 2 and 1.1 equiv of (R)-(-)-phenylglycinol in toluene with
azeotropic removal of water and methanol gave rise to a tetracyclic lactam in greater than 90%
1
yield. Examination of the crude reaction product by H and 13C NMR, capillary GC, and HPLC
revealed this product to be a single isomer, the absolute configuration of which was determined to
be that illustrated by structure 3. The tetracyclic lactam 3 was stereospecifically reduced to either
the cis-fused bicyclic pyrrolidine 4 or the cis-fused bicyclic pyrrolidinone 9. The chiral auxiliaries
in both 4 and 9 were removed using different, novel methodologies. This highly stereoselective
reaction establishes two contiguous chiral centers via the deracemization of an achiral keto-ester.
This methodology has been applied to the synthesis of (1R,5R)-2-azabicyclo[3.3.0]octane 15. This
important amine is now readily available from commercial materials in only three steps.
Sch em e 1
In tr od u ction
The seminal contributions of Meyers have firmly
established the synthetic usefulness of chiral bicyclic
lactams.1 Among the numerous applications of bicyclic
lactam methodology are the preparation of nonracemic,
substituted pyrrolidines and pyrrolidinones.2 Recently,
this methodology has been extended to include the
preparation of enantiomerically-pure piperidines as well.3
These nitrogen heterocycles are widespread among both
natural products and medicinally-important synthetic
compounds. In addition, nonracemic polysubstituted
pyrrolidines serve as useful auxiliaries in several impor-
tant chirality-transfer reactions. The prevalence of these
small-ring nitrogen heterocycles has encouraged the
development of numerous routes for their preparation.4
We wish to describe herein our extension of bicyclic
lactam chemistry that allows for the preparation of cis-
fused bicyclic pyrrolidines and pyrrolidinones with com-
plete stereochemical control. The development of this
new methodology emerged from the need for a practical
asymmetric synthesis of U-93385 (1), a potent serotonin-
1A agonist.5 A key feature of our synthesis is the
simultaneous establishment of two contiguous stereo-
genic centers via the deracemization of an achiral keto-
ester.
Resu lts a n d Discu ssion
Cis-F u sed P yr r olid in es: Syn th esis of U-93385.
Examination of 1 revealed an imbedded pyrrolidine, and
consideration of bicyclic lactam chemistry for its con-
struction led to an intriguing hypothesis. We felt that
condensation of racemic keto-ester 2 with (R)-(-)-phen-
ylglycinol would produce a pair of diastereomeric N-
acyliminium ion intermediates A and B (Scheme 1). We
speculated that if equilibration between A and B (per-
haps via an enamide intermediate; see ref 7) was a facile
process, then intramolecular cyclization might proceed
under thermodynamic control with selectivity for the
desired tetracycle 3. In the event, refluxing a solution
of 25 and 1.1 equiv of (R)-(-)-phenylglycinol in toluene
with azeotropic removal of water and methanol gave rise
to a tetracyclic lactam in greater than 90% yield.6
X Abstract published in Advance ACS Abstracts, J uly 15, 1996.
(1) (a) Meyers, A. I.; Romo, D. Tetrahedron 1991, 47, 9503-9569.
(b) Meyers, A. I.; Lefker, B. A. Tetrahedron 1987, 43, 5663-5676.
(2) (a) Meyers, A. I.; Burgess, L. E. J . Org. Chem. 1991, 56, 2294-
2296. (b) Burgess, L. E.; Meyers, A. I. J . Am. Chem. Soc. 1991, 113,
9858-9859. (c) Burgess, L. E.; Meyers, A. I. J . Org. Chem. 1992, 57,
1656-1662. (d) Meyers, A. I.; Snyder, L. J . Org. Chem. 1992, 57, 3814-
3819. (e) Westrum, L. J .; Meyers, A. I. Tetrahedron Lett. 1994, 35, 973-
976. (f) Tschantz, M. A.; Burgess, L. E.; Meyers, A. I. Org. Synth. 1994,
73, 221-230. (g) Bienz, S.; Busacca, C.; Meyers, A. I. J . Am. Chem.
Soc. 1989, 111, 1905-1907.
(3) Munchhof, M. J .; Meyers, A. I. J . Org. Chem. 1995, 60, 7084-
7085.
(4) See Denmark, S. E.; Marcin, L. R. J . Org. Chem. 1995, 60, 3221-
3235 for an excellent overview of the enantioselective synthesis of
substituted pyrrolidines.
(5) Lin, C.-H.; Haadsma-Svensson, S. R.; Phillips, G.; McCall, R. B.;
Piercey, M. F.; Smith, M. W.; Svensson, K.; Carlsson, A.; Chidester,
C. G.; VonVoigtlander, P. F. J . Med. Chem. 1993, 36, 2208-2218.
S0022-3263(96)00185-5 CCC: $12.00 © 1996 American Chemical Society

5814 J . Org. Chem., Vol. 61, No. 17, 1996
Ennis et al.
Ta ble 1. Su r ver y of Red u cta n ts
Sch em e 2
conditions (solvent, temp)
% yield of 4
LAH, AlCl3 (THF, -78 °C f 0 °C)
LAH (THF, -78 °C f RT)
DIBAL (THF, -78 °C f RT)
BH3·THF (THF, -78 °C f reflux)
76
62a
89
93
a
A small amount (ca. 15%) of the trans-isomer was obtained
from this reaction.
Examination of the crude reaction product by 1H and 13C
NMR, capillary GC, and HPLC revealed this product to
be a single isomer! We were subsequently able to verify
the absolute configuration of this product as that depicted
in Scheme 1 (vide infra). The stereochemical disposition
at the oxazolidine center is consistent with ample pre-
cedent from Meyers’ studies.1-3 That 3 represents the
thermodynamically favored product is supported by mo-
lecular mechanics calculations, which place the two
possible trans-ring junction isomers at substantially
higher energy relative to 3.7 This remarkably selective
reaction has been carried out without event on scales
exceeding 100 kg.8
Although the extraordinary selectivity of this conden-
sation provided the requisite stereochemical relationship
for the imbedded pyrrolidine ring of U-93385, further
elaboration was necessary to complete the synthesis.
Specifically, reduction of the lactam function in 3 along
with a stereoselective replacement of the oxazolidine
C-O bond with a hydrogen atom was needed to establish
the desired cis-fused pyrrolidine framework. Meyers has
reported that tandem reductions of this type can be
achieved using alane at -78 °C2 or DIBAL in refluxing
toluene.3 In both cases, the reductant serves as a Lewis
acid and coordinates with the oxazolidine oxygen to
promote iminium ion formation. Delivery of hydride is
directed by this departing oxygen and results in retention
of configuration. For noncoordinating reducing agents,
allylic 1,3-interactions also predict a stereoselective
reduction with retention of configuration.2b,c We too
found both alane and DIBAL effective at carrying out this
transformation, but in anticipation of production scale
processes, we briefly explored alternative reagents. As
can be seen in Table 1, we found that several reductants
could cleanly effect the stereospecific conversion of 3 to
the cis-fused pyrrolidine 4. Only when lithium aluminum
hydride was used did we witness erosion of stereochem-
ical integrity and generation of significant amounts of
the trans-ring fusion isomer (ca. 15%).9 For overall
efficiency and practicality, we found borane, another
Lewis acidic reagent, to be the reagent of choice for this
transformation. The reduced stereoselectivity witnessed
when using the noncoordinating reagent (LAH) may
reflect a lesser ability for the allylic 1,3-interactions to
direct hydride delivery relative to chelation control.
(6) All new compounds have been fully characterized by 1H NMR
(300 MHz) and 13C NMR (75.5 MHz), IR, and possess satisfactory
elemental analysis.
To intersect with the previous synthetic route to
U-93385 (and thereby confirm the absolute configura-
tion), all that remained was removal of the chiral
auxiliary from 4. Typically, removal of the phenylgly-
cinol-based auxiliary takes advantage of the benzylic
nature of the N-C bond and is accomplished using
hydrogenolytic conditions (Pd or Pd(OH)2, 3 atm H2 or
ammonium formate).2 Unfortunately, the aromatic bro-
mide present in 4 (there as a synthetic handle for later
conversion to the carboxamide in 1) precluded the use of
the normal hydrogenolytic conditions for this cleavage.
Indeed, treatment of 4 with ammonium formate in the
presence of 10 mol % Pd/C cleanly generated the over-
reduced pyrrolidine 5 in excellent yield (Scheme 2). After
considerable experimentation, we found conditions which
effectively removed the chiral auxiliary while allowing
for retention of the required aryl bromide. Following
protection of the primary alcohol of 4 as the benzyl ether
(7) These calculations were performed using MM2. A similar order-
ing of relative energies was calculated for the 1-azabicyclo[3.3.0]octane
system. A reviewer recommended experimental support for the ther-
modynamically driven equilibration we propose as the mechanism for
this condensation. Our suggestion of an equilibrium-based rationale
is founded upon a number of observations. The facile formation of
N-acyliminium ions (such as B) from bicyclic lactams has been
repeatedly described, as has the generation of enamides derived
therefrom.2g We postulate the equilibration of A and B proceeds via a
similar enamide (ii). In support of this, when a solution of 3 in toluene/
CH3OD is treated with a catalytic amount of p-TsOH (0.1 equiv) at
room temperature, one observes a gradual diminution of the benzylic
proton resonance. While it is possible that iminium ion
B is in
equilibrium with ii whereas A is not, we have no evidence to suggest
that.
(8) The scale-up of this condensation was carried out by Dr. Thomas
A. Runge, Chemical Research Preparations, Pharmacia and Upjohn,
Inc.
(9) Meyers has also observed that using LAH alone is less clean than
using alane for these reductions (see ref 2a).

Synthesis of Bicyclic Pyrrolidines and Pyrrolidinones
J . Org. Chem., Vol. 61, No. 17, 1996 5815
Sch em e 3
Sch em e 4
(91
)
(77
)
as the progenitor for the observed enamide 10.14 Simple
hydrolysis of 10 provides the lactam 11 in 83% overall
yield. Thus, the versatile intermediate 3 can serve as
an entry point into either cis-fused pyrrolidines or
pyrrolidinones.
Non -Tetr a lon e Ba sed System s. The chemistry de-
scribed above pertains to a specialized case involving
keto-tetralones and originated from a specific need for
construction of U-93385. We have shown, however, that
the remarkable control of adjacent stereocenters via
polycyclic lactams is available in other, simpler systems
as well. To illustrate, we examined the condensation of
(R)-(-)-phenylglycinol with the 2-substituted cyclo-
alkanone 12 (Scheme 4).15 We found the same high level
of stereoselectivity for 12 as we had observed for 2, again
obtaining a single isomer 13 in 83% yield. Stereoselective
reduction of 13 with borane afforded the pyrrolidine 14,
and hydrogenolytic removal of the chiral auxiliary pro-
vided (1R,5R)-2-azabicyclo[3.3.0]octane (15) in enantio-
merically pure form. Thus, our new methodology has
allowed for the synthesis of this cis-fused bicyclic amine
in only three steps! This amine, only recently prepared
in optically-active form, has been used as a chiral
auxiliary in Michael-type reactions of enamines.16
6,10 treatment with 2-chloroethyl chloroformate11 in warm
toluene and subsequent methanolysis provided the sought
after secondary amine 7 in 71% overall yield.12
The cis-fused pyrrolidine 7 is a common intermediate
with the previously-reported synthesis of 1.5 Conversion
of 7 into U-93385 (1) requires only addition of an n-propyl
group onto the secondary nitrogen followed by transfor-
mation of the aryl bromide into the carboxamide. This
overall process for the preparation of U-93385 has been
carried out as described herein on scales providing over
25 kg of final product.13
Cis-F u sed P yr r olid in on es. Meyers has demon-
strated the utility of chiral bicyclic lactams for the
preparation of pyrrolidinones based upon a selective
reduction with triethylsilane.2c,f We have also shown that
3 can be selectively reduced to the pyrrolidinone 9 using
triethylsilane and titanium tetrachloride (Scheme 3).
Again, only the cis-ring fusion was detected in this
reaction. Cleavage of the chiral auxiliary in Meyers’s
monosubstituted pyrrolidinones is typically accomplished
using dissolving metal reductions (Li, ammonia),2 condi-
tions which we again wished to avoid in order to retain
our aryl bromide. For our case, we developed a novel,
base-induced dehydration for removal of the chiral
auxiliary in 9. Treatment of 9 with lithium hydroxide
in DMSO at room temperature affords an excellent yield
of the enamide 10. We speculate the intermediacy of i
Con clu sion
We have described an extension of chiral bicyclic
lactam chemistry which allows for the preparation of cis-
fused bicyclic pyrrolidines and pyrrolidinones. This new
methodology simultaneously establishes the two contigu-
ous asymmetric centers of the bicyclic system in a single,
high-yielding reaction of a racemic starting keto-ester.
Novel transformations of the initial cyclodehydration
product provide alternatives to published procedures for
chiral auxiliary removal. These new transformations
provide access to both pyrrolidines and pyrrolidinones
containing reducible functional groups which would be
reactive under standard hydrogenolytic conditions. The
tetracyclic lactam 3 has proven rich in synthetic poten-
tial, and we have explored a number of facets of this
chemistry. Results of these studies will be the subject
of future reports from our laboratory.
(10) We chose to protect the primary hydroxyl of 4 as a benzyl for
convenience and cost reasons. We have carried out this sequence with
success using a number of other alcohol derivatives, including tert-
butyldimethylsilyl and methyl ethers as well as acetates.
(11) Olofson, R. A.; Martz, J . T.; Senet, J .-P.; Piteau, M.; Malfroot,
T. J . Org. Chem. 1984, 49, 2795-2799.
(12) Meyers provides an alternative procedure for auxiliary removal
for compounds intolerant of hydrogenolytic conditions. This involves
transformation of the primary alcohol of the auxiliary to a thiophenyl
derivative using triethylphosphine and diphenyl disulfide and then
treatment with lithio di-tert-butylbiphenyl (51% overall yield; see ref
2a).
(14) For a related reaction, see Vedejs, E.; Wilde, R. G. J . Org. Chem.
1986, 51, 117-119.
(13) The multi-kilogram preparation of U-93385 by this chemistry
was carried out by Dr. Michael F. Lipton and his associates Michael
A. Mauragis and Michael F. Veley, Chemical Research Preparations,
Pharmacia and Upjohn, Inc.
(15) During the preparation of this manuscript, an example of
condensations similar to these appeared: Ragan, J . A.; Claffey, M. C.
Heterocycles 1995, 41, 57-70. We thank Dr. Ragan for alerting us to
these results.

5816 J . Org. Chem., Vol. 61, No. 17, 1996
Ennis et al.
resulting milky aqueous suspension was extracted with dichlo-
romethane (3 × 400 mL). The combined organic phases were
washed once with brine (300 mL), dried over MgSO4, filtered,
and concentrated to give 21.17 g of a yellow syrup. This crude
product was purified by chromatography on a Waters Prep 500
using 5% ethyl acetate/hexane to give 16.99 g (83%) of the
desired benzyl ether as a pale yellow syrup: Rf 0.28 (5% ethyl
acetate/hexane); IR (mull) 2966, 2944, 2924, 2888, 2861, 2807
Exp er im en ta l Section
Proton and carbon magnetic resonance spectra were re-
corded in CDCl3 at 300 and 75.5 MHz, respectively, and are
reported in ppm on the δ scale from internal tetramethylsilane.
Infrared spectra, combustion analysis, optical rotation, and
mass spectra were determined by Physical and Analytical
Chemistry, The Upjohn Company. When necessary, solvents
and reagents were dried prior to use. Anhydrous tetrahydro-
furan refers to material that was distilled from sodium metal/
benzophenone ketyl. Dichloromethane was dried over acti-
vated 4 Å molecular sieves. Unless otherwise noted, all non-
aqueous reactions were carried out under an atmosphere of
dry nitrogen using oven-dried glassware.
Con d en sa t ion of Ket o-E st er 2 a n d (R)-P h en ylgly-
cin ol: P r ep a r a tion of Tetr a cyclic La cta m (3). A 1-L
round bottom flask was charged with keto-ester 2 (29.18 g,
0.098 mol), toluene (490 mL), and (R)-2-phenylglycinol (20.21
g, 0.147 mol) and fitted with a Dean-Stark trap. The reaction
suspension was heated to reflux, and after approximately 1 h
the reaction became homogenous. Heating was continued for
18 h, by which time the trap contained approximately 2.0 mL
of water (theory ) 1.8 mL). The reaction mixture was cooled
to room temperature and concentrated in vacuo. The resulting
crude product was purified by chromatography on a Waters
Prep 500 using 20% ethyl acetate/hexane to give 34.75 g (92%)
of 3 as a pale yellow solid, mp 123.0-125.5 °C: Rf 0.24 (20%
ethyl acetate/hexane); IR (mull) 2949, 2926, 2855, 1710, 1447,
1364, 1025, 786, 718, 702 cm-1; 1H NMR δ 7.37 (m, 6H), 7.12
(d, J ) 6.7 Hz, 1H), 7.05 (t, J ) 7.6 Hz, 1H), 5.33 (t, J ) 7.3
Hz, 1H), 4.70 (dd, J ) 8.7, 8.1 Hz, 1H), 4.23 (dd, J ) 8.8, 6.6
Hz, 1H), 3.84 (t, J ) 9.8, 1H), 3.51 (dd, J ) 17.1, 9.5 Hz, 1H),
2.84 (t, J ) 6.0 Hz, 2H), 2.57 (dd, J ) 17.1, 10.3 Hz, 1H), 2.09
(m, 1H), 1.85 (m, 1H); 13C NMR δ 176.9, 139.8, 138.7, 136.3,
131.2, 128.8, 128.1, 127.6, 125.5, 124.7, 101.2, 73.4, 57.5, 44.9,
41.1, 30.7, 27.3; [R]25D -265 (c 0.961, methanol). Anal. Calcd
for C20H18NO2Br: C, 62.51; H, 4.72; N, 3.65. Found: C, 62.55;
H, 4.76; N, 3.61.
1
1453, 1110, 734, 695 cm-1; H NMR δ 7.28 (m, 11H), 7.01 (d,
J ) 7.4 Hz, 1H), 6.92 (t, J ) 7.7 Hz, 1H), 4.50 (dd, J ) 15.7,
12.2 Hz, 2H), 4.07 (t, J ) 6.3 Hz, 1H), 3.87 (dd, J ) 9.7, 6.1
Hz, 1H), 3.77 (dd, J ) 9.7, 6.6 Hz, 1H), 3.40 (q, J ) 8.4 Hz,
2H), 3.07 (m, 1H), 2.83 (m, 2H), 2.56 (m, 3H), 1.93 (m, 1H),
1.53 (m, 1H), 1.38 (m, 1H); 13C NMR δ 140.4, 139.9, 138.7,
138.4, 130.4, 128.8, 128.3, 128.0, 127.5, 127.4, 127.2, 126.2,
125.0, 73.2, 73.1, 63.8, 58.0, 47.9, 41.7, 31.4, 27.3, 24.8; [R]25
D
-92 (c 0.9895, methanol). Anal. Calcd for C27H28NOBr: C,
70.13; H, 6.10; N, 3.03. Found: C, 69.94; H, 6.01; N, 2.87.
cis-(-)-2,3,3a ,4,5,9b -H exa h yd r o-9-b r om o-1H -b en z[e]-
in d ole (7). A solution of benzyl ether 6 (16.96 g, 36.7 mmol)
in chlorobenzene (70 mL) was treated with 1-chloroethyl
chloroformate (20.0 mL, 0.183 mol) and heated to reflux (bath
temp 150 °C). During the initial minutes of heating the
reaction darkened to a deep emerald green. After 18 h at
reflux, the now-brown reaction mixture was cooled to room
temperature and treated with a second portion of the chloro-
formate reagent (20 mL), and refluxing was continued for an
additional 4 h. The reaction was then cooled, treated with
methanol (500 mL), and reheated to reflux for 1 h. At this
point, the reaction was cooled to room temperature and
concentrated to a brown oil. This material was dissolved in
dichloromethane (300 mL) and washed with 1 M HCl (3×).
The combined aqueous washes were cooled in an ice bath and
adjusted to pH > 13 with 50% sodium hydroxide, forming a
milky solution. This basic aqueous phase was extracted with
dichloromethane (2 × 600 mL), and the combined organic
layers were dried over MgSO4, filtered, concentrated in vacuo
to give 7.91 g (86%) of 7 as a light tan oil: IR (neat) 2961,
2934, 2862, 2841, 1560, 1453, 1440, 1400, 1176, 774 cm-1; 1H
NMR δ 7.40 (d, J ) 7.7 Hz, 1H), 7.05 (d, J ) 7.3 Hz), 6.95 (t,
J ) 7.7 Hz, 1H), 3.51 (m, 2H), 3.06 (m, 1H), 2.92 (m, 1H), 2.76
(m, 1H), 2.65 (m, 2H), 2.23 (bs, 1H), 1.80 (m, 1H), 1.65 (m,
1H), 1.41 (m, 1H); 13C NMR δ 139.8, 139.5, 130.6, 127.7, 126.8,
cis-(-)-2,3,3a ,4,5,9b-Hexa h yd r o-9-br om o-3-(2(R)-1-h y-
d r oxyp h en eth -2-yl)-1H-ben z[e]in d ole (4). A solution of 3
(0.769 g, 2.00 mmol) in anhydrous THF (10 mL) was cooled to
-78 °C and treated dropwise with a 1 M solution of borane in
THF (6.0 mL, 6.00 mmol). The reaction was stirred for 2 h at
-78 °C and then for 2 h at room temperature and then finally
brought to reflux for 3 h. After stirring at room temperature
overnight, the reaction was treated dropwise with 1 M aqueous
HCl (5 mL), causing vigorous gas evolution. The reaction was
again brought to reflux for 1 h and then cooled to room
temperature and poured into brine (30 mL). The aqueous
phase was basified to pH 10 with 5 N NaOH and extracted
with dichloromethane (3 × 30 mL). The combined organic
phases were dried over MgSO4, filtered, and concentrated to
a colorless oil. The resulting crude product was purified by
flash chromatography on 60 g silica gel using 20% ethyl
acetate/hexane to give 692 mg (93%) of 4 as a colorless, tacky
solid: Rf 0.20 (15% ethyl acetate/hexane); IR (neat) 2936, 1453,
1442, 1177, 1081, 1060, 1035, 1029, 767, 703 cm-1; 1H NMR δ
7.36 (m, 4H), 7.23 (m, 2H), 7.05 (d, J ) 7.4 Hz, 1H), 6.69 (t, J
) 7.7 Hz, 1H), 4.07 (m, 2H), 3.69 (dd, J ) 9.2, 4.1 Hz, 1H),
3.42 (dd, J ) 18.4, 9.1 Hz, 1H), 3.10 (m, 2H), 2.95 (t, J ) 7.8
Hz, 1H), 2.79 (m, 1H), 2.51 (m, 2H), 2.30 (m, 1H), 2.18 (d of q,
J ) 13.7, 3.5 Hz, 1H), 1.48 (m, 1H), 1.33 (m, 1H); 13C NMR δ
140.2, 139.4, 134.7, 130.3, 129.2, 128.0, 127.7, 127.3, 126.5,
125.6, 56.2, 45.3, 42.6, 33.8, 28.4, 27.3; [R]25 -113 (c 0.6461,
D
methanol). For the hydrochloride salt: IR (mull) 2802, 2772,
2741, 2644, 2602, 2589, 2485, 2476, 1443, 771 cm-1; [R]25D -84
(c 0.8539, methanol). Anal. Calcd for C12H15NBrCl: C, 49.94;
H, 5.24; N, 4.85. Found: C, 49.81; H, 5.32; N, 4.87.
cis-(-)-2,3,3a ,4,5,9b-Hexa h yd r o-9-br om o-3-(2(R)-1-h y-
d r oxyp h en eth -2-yl)-1H-ben z[e]in d ol-2-on e (9). A solution
of 3 (15.0 g, 39.03 mmol) in CH2Cl2 (176 mL) was cooled to
-78 °C and treated with triethylsilane (13.72 mL, 85.88 mmol)
via syringe. After 15 min, TiCl4 (9.42 mL, 85.88 mmol) was
added dropwise via syringe. The reaction was stirred for
another 2 h at -78 °C and then warmed to room temperature
and quenched with saturated aqueous NH4Cl (400 mL). The
organics were separated, and the aqueous layer was extracted
with CH2Cl2 (2x). The organics were combined, washed with
brine, dried over MgSO4, filtered, and concentrated to give 22.4
g of a crude product. This material was purified on a Waters
Prep 500 using 40% ethyl acetate in hexane as the eluent to
give 12.0 g (80%) of 9 as a white foam: Rf 0.15 (40% ethyl
acetate in hexane); [R]25 -227 (c 1, ethanol); IR (mull) 3371,
D
124.5, 62.7, 61.1, 56.4, 44.6, 41.0, 31.6, 26.0, 25.9; [R]25 -127
1666, 1443, 1424, 1367, 1259, 1064, 1060, 777, 702 cm-1; H
1
D
(c 0.566, methanol). Anal. Calcd for C20H22NOBr‚0.5 H2O: C,
63.00; H, 6.08; N, 3.67. Found: C, 63.02; H, 5.83; N, 3.60.
cis-(-)-2,3,3a ,4,5,9b-Hexa h yd r o-9-br om o-3-(2(R)-1-h y-
d r oxyp h en eth -2-yl)-1H-ben z[e]in d ole (6). To a solution of
amino-alcohol 5 (16.50 g, 44.3 mmol) in anhydrous DMSO (148
mL) was added freshly-powdered potassium hydroxide (11.71
g of 85% KOH, 0.18 mol). The reaction was stirred at room
temperature for 10 min prior to the dropwise addition of benzyl
chloride (10.2 mL, 88.7 mmol), and then stirring was continued
for an additional 50 min, during which time the reaction color
gradually changed from dark-orange to yellow. At this point,
the reaction was added to ice-water (300 mL), and the
NMR δ 7.42-7.32 (m, 6H), 7.06-6.98 (m, 2H), 4.58-4.54 (m,
2H), 4.39-4.30 (m, 1H), 4.00-3.93 (m, 1H), 3.75-3.61 (m, 2H),
3.39 (q, J ) 8.8 Hz, 1H), 2.84 (t of d, J a ) 5.4 Hz, J b ) 16.8
Hz, 1H), 2.62 (d of t, J a ) 4.6 Hz, J b ) 10.5 Hz, 1H), 2.41 (q,
J ) 9.3 Hz, 1H), 1.94 (m, 1H), 1.82 (m, 1H); 13C NMR δ 175.2,
138.5, 137.3, 136.0, 131.2, 128.9, 128.1, 127.9, 127.86, 127.82,
127.1, 125.1, 64.4, 62.7, 58.7, 38.6, 37.1, 27.4, 24.2. Anal. Calcd
for C20H20N1O2Br1: C, 62.19; H, 5.22; N, 3.63. Found: C,
61.80; H, 5.17; N, 3.39. HRMS Calcd for C20H20NO2Br:
386.0756. Found: 386.0759.
cis-(-)-2,3,3a ,4,5,9b -H exa h yd r o-9-b r om o-3-(1-p h en yl-
eth ylen e)-3a -(3-p r op en yl)-1H-ben z[e]in d ol-2-on e (10). A

Synthesis of Bicyclic Pyrrolidines and Pyrrolidinones
J . Org. Chem., Vol. 61, No. 17, 1996 5817
250 mL round bottom flask equipped with a reflux condenser
was charged with 9 (5.02 g, 13.0 mmol), LiOH‚H2O (2.18 g,
52.0 mmol), and DMSO (100 mL). The reaction was heated
to 100 °C for 72 h. The brown mixture was cooled to room
temperature and poured into 1 L of H2O. The aqueous layer
was extracted with CH2Cl2 (3 × 500 mL), dried over Na2SO4,
filtered, and concentrated to a crude brown oil. The mixture
was purified on a Waters Prep 500 using 20% ethyl acetate
as the eluent to give 4.33 (90%) of the enamide 10 as a white
foam: Rf 0.44 (40% ethyl acetate in hexane); IR (mull) 1700,
resulting aqueous layer was adjusted to pH > 10 with 5 N
NaOH. The aqueous base layer was extracted with dichloro-
methane (4×) and the combined organics layers were dried
over MgSO4, filtered, and concentrated to an oil. This material
was purified by chromatography on silica gel (40% ethyl
acetate/hexane) to give the desired amino alcohol 14 (4.24 g,
91% yield) as a clear, colorless oil: IR (liquid) 2948, 2861, 2826,
1452, 1075, 1058, 1035, 1029, 767, 703 cm-1; 1H NMR δ 7.38-
7.26 (m, 3H), 7.20-7.17 (m, 2H), 3.95-3.83 (m, 2H), 3.60 (dd,
J ) 8.4, 3.6, 1H), 3.06 (br t, J ) 7.1, 2H), 2.82 (t, J ) 7.5, 1H),
1
1445, 1396, 1367, 1331, 1304, 1243, 1178, 775, 698 cm-1; H
2.36-2.28 (m, 1H), 2.04-1.96 (m, 1H), 1.91-1.24 (m, 8H); 13
C
NMR δ 7.45-7.32 (m, 6H), 7.07-6.98 (m, 2H), 5.68 (s, 1H),
5.44 (s, 1H), 4.15-4.04 (m, 1H), 3.89 (q, 1H, J ) 8.4 Hz), 3.47
(q, 1H, J ) 9.5 Hz), 2.86-2.77 (m, 1H), 2.62-2.53 (m, 1H),
2.46 (q, 1H, J ) 8.4 Hz), 1.87-1.74 (m, 2H); 13C NMR δ 173.3,
140.9, 138.7, 136.4, 135.5, 131.1, 128.6, 128.5, 127.8, 127.6,
126.2, 125.1, 113.2, 57.7, 38.4, 35.7, 26.5, 25.0; [R]25D -55 (c 1,
ethanol); HRMS calcd for C20H18NOBr: 367.0572. Found:
367.0587.
NMR δ 135.4, 129.3, 128.0, 127.6, 64.7, 64.5, 61.1, 46.8, 41.3,
33.1, 32.0, 31.6, 24.0; [R]25D -56.2 (c 1.01, EtOH); HRMS calcd
for C15H21NO: 231.1623. Found: 231.1619. Anal. Calcd for
C15H21NO: C, 77.88; H, 9.15; N, 6.05. Found: C, 77.58; H,
9.13; N, 6.08.
(1R, 5R)-2-Aza bicyclo[3.3.0]octa n e (15). A mixture of the
amino alcohol 14 (4.16 g, 18.0 mmol) and ammonium formate
(5.68 g, 90.0 mmol) in methanol (300 mL) was treated with
10% palladium on carbon (2.03 g) and stirred overnight at
room temperature. The reaction mixture was treated with 5
M HCl (40 mL) and filtered through a pad of Celite, washing
repeatedly with methanol. The filtrate and washings were
combined and concentrated in vacuo to a yellow solid, which
was dissolved in a minimum of 1 N HCl and extracted with
ether (4×). The acidic aqueous phase was cooled in an ice bath
and brought to pH > 12 with solid NaOH. This basic phase
was then extracted with minimum amounts of ether (3×, total
ether volume ) 50 mL). The organic layers were combined
and the ether, was removed by atmospheric distillation. The
remaining residue was transferred to a smaller flask and
distillation was continued. In this manner was obtained 15
(1.55 g, 77%) as a mobile, volatile oil, bp 115-117 °C: 1H NMR
δ 3.61-3.56 (m, 1H), 2.89-2.84 (m, 1H), 2.74-2.66 (m, 1H),
2.51-2.38 (m, 1H), 1.91-1.79 (m, 1H), 1.73 (br s, 1H), 1.75-
cis-(-)-2,3,3a ,4,5,9b -H exa h yd r o-9-b r om o-1H -b en z[e]-
in d ol-2-on e (11). A solution of enamide 10 (8.88 g, 24.05
mmol) in THF (200 mL) and 1 N aqueous HCl (36.08 mL) was
heated to reflux for 8 h and then cooled to room temperature
and diluted with H2O (250 mL). After concentration in vacuo,
the residue was extracted with CH2Cl2 (3×), and the organic
layers were combined, dried over MgSO4, filtered, and con-
centrated to a tan solid. The solid was triturated with hot
hexane (4×) to give 5.88 g (92%) of 11 as an off-white solid:
mp 189-190 °C; Rf 0.26 (50% acetone in hexane); IR (mull)
3159, 3083, 3055, 1700, 1647, 1437, 1328, 1303, 816, 771 cm-1
;
1H NMR δ 7.45 (d, 1H, J ) 7.5 Hz), 7.05 (quintet, 2H, J ) 7.5
Hz), 6.18 (broad s, 1H), 4.11 (d of d, 1H, J a ) 7.1 Hz, J b ) 11.6
Hz), 3.93 (d of d, 1H, J a ) 7.6 Hz, J b ) 17.3 Hz), 3.18 (q, 1H,
J ) 9.8 Hz), 2.94-2.84 (m, 1H), 2.72-2.63 (m, 1H), 2.21 (d of
d, 1H, J a ) 7.5 Hz, J b ) 17.4 Hz), 1.92-1.84 (m, 2H); 13C NMR
δ 177.5, 139.0, 136.9, 131.0, 127.7, 127.5, 125.1, 53.0, 37.7, 37.6,
1.52 (m, 3H), 1.48-1.40 (m, 2H), 1.36-1.24 (m, 2H); [R]20
)
D
27.7, 26.1; [R]25 -291 (c 1, ethanol). Anal. Calcd for C12H12
-
+ 25.3 (c 4.04, CHCl3) [lit.16a value +15.1 (c 3.51, CHCl3)].16c
Due to its volatility, this compound was characterized as the
picrate salt, mp 205-206 °C (lit.17 mp 201-203); IR (mull)
3114, 3093, 3027, 1646, 1623, 1540, 1347, 1328, 1081, 731
D
NOBr: C, 54.16; H, 4.54; N, 5.26; Br, 30.02. Found: C, 53.93;
H, 4.34; N, 5.17; Br, 29.64.
P r ep a r a tion of Tr icyclic La cta m 13. A mixture of ethyl
2-cyclopentanoneacetate (12) (7.00 g, 41.1 mmol) and (R)-
(-)-2-phenylglycinol (8.46 g, 61.7 mmol) was refluxed in
toluene (125 mL) in a flask fitted with a Dean-Stark trap.
After 16 h, the reaction was cooled and concentrated in vacuo
to an orange solid which was purified by chromatography of
silica gel (25% ethyl acetate/hexane) to give 13 (8.31 g, 83%
yield) as a light yellow solid, mp 43.5-44.0 °C: IR (mull) 1708,
1
cm-1; H NMR δ 9.13 (d, J ) 2.1, 2H), 9.09 (t, J ) 2.2, 1H),
4.15 (m, 1H), 4.41-3.22 (m, 2H), 2.90 (m, 1H), 2.33-2.21 (m,
1H), 2.12-1.51 (m, 7H); 13C NMR 168.0, 148.1, 141.6, 129.1,
119.9, 63.4, 45.6, 42.5, 32.3, 31.1, 31.2, 25.0; [R]D -2.17° (c
0.8770, EtOH). Anal. Calcd for C14H17N3O6: C, 52.01; H, 5.30;
N, 13.00. Found: C, 51.82; H, 5.28; N, 13.01.
1451, 1356, 1330, 1234, 1058, 1020, 972, 731, 702 cm-1
;
1H
Ack n ow led gm en t. We gratefully acknowledge our
process colleagues Dr. Michael F. Lipton, Michael A.
Mauragis, and Michael F. Veley for the many useful
discussions throughout the course of this project. We
would also like to thank Dr. Thomas Runge for provid-
ing large amounts of the condensation product 3.
Finally, we would like to acknowledge Dr. Chiu-Hong
Lin and the rest of the Monoamine Program Team for
the discovery of U-93385.
NMR δ 7.38-7.24 (m, 5H), 5.16 (t, J ) 7.7, 1H), 4.62 (t, J )
8.3, 1H), 3.99 (t, J ) 7.6, 1H), 2.90 (dd, J ) 17.3, 10.4, 1H),
2.77-2.68 (m, 1H), 2.49 (dd, J ) 17.3, 6.6, 1H), 2.07-1.91 (m,
2H), 1.85-1.70 (m, 3H), 1.68-1.55 (m, 1H); 13C NMR δ 180.3,
139.7, 128.7, 127.4, 125.5, 110.8, 73.4, 57.8, 41.4, 40.6, 36.6,
32.3, 24.5; [R]25D -168.5 (c 0.995, EtOH) (lit. [R]D -161, c 2.0,
CH2Cl2).15 Anal. Calcd for C15H17NO2: C, 74.05; H, 7.04; N,
5.76. Found: C, 73.66; H, 6.98; N, 5.67.
(1R,5R)-2-(2(R)-1-Hyd r oxyp h en eth -2-yl)-2-a za bicyclo-
[3.3.0]octa n e (14). A solution of 13 (4.95 g, 20.3 mmol) in
anhydrous THF (50 mL) was cooled to -78 °C and treated
dropwise with a 1 M solution of borane in THF (61.0 mL).
Stirring was continued at -78 °C for 1 h, at which point the
cooling bath was removed and the reaction was stirred for an
additional 1 h. The reaction was then heated to a gentle reflux
for 2 h, after which it was cooled to room temperature and
stirred for 48 h. The reaction was then placed in an ice-bath
and treated dropwise with 1 N HCl (61 mL; vigorous gas
evolution). When the addition was complete, the reaction was
heated to reflux for 1.5 h and then cooled and poured into brine
(200 mL). This mixture was concentrated in vacuo and the
J O960185L
(16) (a) Wallbaum, S.; Mehler, T.; Martens, J . Synth. Commun.
1994, 24, 1381-1387. (b) Stingl, K.; Martens, J . Liebigs Ann. Chem.
1994, 479-484. (c) Martens, J .; Lubben, S. Tetrahedron 1991, 47,
1205-1214. Interestingly, these authors report a rotation for the
hydrochloride salt of 15 of [R]20 ) -3.0 (c 1.2, methanol) (note the
D
change in sign). It is possible that the discrepancy between the rotation
value for 15 we obtained and that report by Wallbaum et. al. is due to
the presence of HCl in the CHCl3 used.
(17) Booth, H.; King, F. E.; Mason, K. G.; Parrick, J .; St. D.
Whitehead, R. L. J . Chem. Soc., 1959, 1050-1054.