Development of synthetic routes, via a tropinone intermediate, to a long-acting muscarinic antagonist for the treatment of respiratory disease

Article abstract of DOI:10.1021/op3002933

This contribution describes the development of two synthetic routes to an investigational muscarinic antagonist for the treatment of chronic obstructive pulmonary disease. The first route used a starting material which was in plentiful supply within the GSK network and was used to make material for early clinical trials and safety assessment studies. Further investigations identified a second, potential long-term manufacturing route from commercially available building blocks, using substrate control to install the two stereocentres with excellent selectivity. A key step was a substrate-directed epoxide reduction which also gave rise to a minor byproduct through a skeletal rearrangement of the tropane ring. A deuterium-labeling experiment was carried out, which shed light on the origin of the byproduct, and also guided the improvement of reaction conditions.

Full text of DOI:10.1021/op3002933

Article
pubs.acs.org/OPRD
Development of Synthetic Routes, via a Tropinone Intermediate, to a
Long-Acting Muscarinic Antagonist for the Treatment of Respiratory
Disease
Robert N. Bream,*,† Doug Hayes, David G. Hulcoop, and Alexandra J. Whiteman
GlaxoSmithKline Medicines Research Centre, Gunnels Wood Road, Stevenage, Hertfordshire SG1 2NY, United Kingdom
S
* Supporting Information
ABSTRACT: This contribution describes the development of two synthetic routes to an investigational muscarinic antagonist
for the treatment of chronic obstructive pulmonary disease. The first route used a starting material which was in plentiful supply
within the GSK network and was used to make material for early clinical trials and safety assessment studies. Further
investigations identified a second, potential long-term manufacturing route from commercially available building blocks, using
substrate control to install the two stereocentres with excellent selectivity. A key step was a substrate-directed epoxide reduction
which also gave rise to a minor byproduct through a skeletal rearrangement of the tropane ring. A deuterium-labeling experiment
was carried out, which shed light on the origin of the byproduct, and also guided the improvement of reaction conditions.
hronic obstructive pulmonary disease (COPD) is a fatal
treatment with trimethylsilyl (TMS)-cyanide and a Lewis acid.5
Clung disease characterized by progressive and irreversible A screen of reaction conditions revealed tin tetrachloride in
airflow limitation with systemic consequences.1 It is the fifth
DCM as optimal, affording a 9:1 ratio of substitution to
elimination products. The desired nitrile was purified as its HCl
salt 4 by selective crystallization from the crude reaction
mixture. On laboratory scale, this process reliably gave 70%
yield. On the single batch carried out in pilot-plant equipment,
poor mixing resulted in poor recovery and a lower 49% yield.
Alkylation of the nitrile free-base with benzylbromoethyl ether
afforded a 5:2 ratio of products, favoring undesired isomer 5.
Tropane 4 was therefore demethylated by treatment of its free
base with 1-chloroethylchloroformate, and the resultant
secondary amine 6 was realkylated with benzylbromoethyl
ether to afford tertiary amine 7.6 Initially, demethylation was
carried out using diisopropylamine as base. Under these
leading cause of death worldwide, killing 250 people every
hour, and is the only disease in the World Health Organization
(WHO) top five with an increasing death rate.2 While there is
currently no cure, treatment focuses on bronchodilatation,
targeting adrenergic and muscarinic receptors in the lung to
cause smooth muscle relaxation.3 Inhaled corticosteroidal anti-
inflammatories are also used.4 As part of GlaxoSmithKline’s
long-standing commitment to the development of novel
medicines for the treatment of pulmonary disease, tropane
(1) was designed as a long-acting, selective, and reversible
muscarinic antagonist which might provide a once-daily inhaled
treatment for COPD (Figure 1).
conditions, some dealkylation of Hunig’s base was observed,
̈
resulting in contamination of the isolated product with
ethylisopropylamine and diisopropylamine hydrochloride.
These impurities were readily purged during downstream
steps. However, on scale up to 8.5 kg in pilot-plant equipment,
these impurities sublimed during solvent distillations resulting
in caking of the condenser system with white powder which
was difficult to remove. Therefore, a second process, using
NaHCO3 as base in toluene was developed, affording reliably
higher yields without amine impurities on laboratory scale.
Alkylation with iodomethane afforded 1 as the major product in
7:1 solution ratio. The isolated ratio was improved to 99:1 by
ripening the resultant slurry with added methanol before
filtration. Subsequent recrystallization from aqueous methanol
afforded product in greater than 99.5:0.5 dr (Scheme 1).
Route A provided material for early toxicological assessment
and the first human clinical trials. However, the dealkylation/
realkylation sequence adds two steps to the sequence, and
tropane ester 2 is expensive. As a long-term manufacturing
Figure 1. Long-acting muscarinic antagonist for COPD.
Of key concern in designing the synthetic route was the
diastereoselective alkylation of the tropane nitrogen to afford
the requisite quaternary ammonium salt. Early studies revealed
that alkylation to afford a quaternary ammonium salt occurred
preferentially from the less hindered side of the nitrogen,
adjacent to the methylene bridge. The initial synthetic route
(route A) used tropane ester 2 as the starting material as it
already had the correct stereochemistry at C3 and was available
in bulk quantities from other active research programmes
within GlaxoSmithKline. Treatment with phenyllithium
provided the tertiary alcohol 3 in good yield as a crystalline
solid. Substitution of the hydroxyl group was accomplished by
Received: October 19, 2012
Published: March 7, 2013
© 2013 American Chemical Society
641
dx.doi.org/10.1021/op3002933 | Org. Process Res. Dev. 2013, 17, 641−650

Organic Process Research & Development
Article
Scheme 1. Route A to troponium salt diastereomers
Scheme 2. Initial attempts to install C3 stereochemistry with benzyloxyethyl group present
strategy, more efficient access to 1 was sought from cheap and
abundant starting materials. Given the required order of amine
alkylation, we sought to install the benzyloxyethyl side chain at
an early stage. The first strategy involved reduction of an olefin
to generate the required stereocentre at C3. Both endo- and exo-
olefins, 9 and 10, were synthesized in short order from
tropinone 11. However, reduction of olefins 9 by either
hydrogenation or conjugate reduction7 failed to give syntheti-
cally useful diastereoselectivity at the C3. Diastereomeric ratios
from 1:1 to 2:1 suggested that the olefin was too far from the
steric bulk of the molecule for significant distinction between its
two faces. Olefin 10 proved unreactive to a range of
hydrogenation catalysts, even at elevated temperatures and
pressures, likely due to the hindered steric environment
provided by the tropane ring and the bulky diphenylacetonitrile
group. (Scheme 2).
as the limiting reagent, provided access to tropinone 11 in
greater than 90% purity as a solution in 2-MeTHF (Scheme 3).
A seven-factor, ten-experiment, DoE factor-screening study was
used to explore the reaction.10 Key findings were that 1.05
equiv of 2,5-dimethoxytetrahydrofuran proved insufficient for
complete consumption of amine 12, while high levels (1.55
equiv) resulted in the formation of increased levels of the main
pyrrole impurity, formed by condensation of 2,5-dimethoxyfur-
an and amine 12. A small reduction in the charge of sodium
acetate (from 4 equiv to 3.5 equiv) was found to be beneficial
to minimize formation of this impurity. The highest levels of
the desired product were obtained in solution using the
conditions reported in the Experimental Section.
Corey−Chaykovsky epoxidation provided 13 as a single
diastereomer with the ylid attacking from the less hindered
upper face.11 It was found that using tBuOH rather than DMSO
as solvent increased the reaction rate 10-fold. While stereo-
chemistry of 13 could not be assigned directly, the major
byproduct 14, made by addition of a second equivalent of ylid
to the distal terminus of the epoxide, was isolated and
characterized, and the configuration of 13 was assigned by
Alternative methods, using key tropinone 11 as the starting
material, were therefore investigated. This was readily accessed
by functionalization of amine 12 in a three-component
Robinson tropinone reaction.8,9 Optimization of reaction
conditions using a DoE approach, setting the valuable amine
642
dx.doi.org/10.1021/op3002933 | Org. Process Res. Dev. 2013, 17, 641−650

Organic Process Research & Development
Article
Scheme 3. Revised route to muscarinic antagonist 1
Table 1. Optimization of reaction conditions for isomerization of epoxide 13
solvent
THF
base
ratio 15:16
time (min)
Et2NLi
iPr2NLi
Et2NLi
Et2NLi
Et2NLi
14:1
1:1
15
45
THF
2-MeTHF
PhMe
6:1
15
1:7.5
1:1
5
DME
>500
analogy. Isomerization of the epoxide to allylic alcohol 15 was
attempted under a range of Lewis acidic and basic conditions.12
Rearrangement under acidic conditions gave only small
amounts of the desired allylic alcohol, along with degradation
products. Treatment with lithium amide bases gave the desired
product along with amine addition products 16.
A screen of solvents and bases (Table 1) revealed that less
coordinating solvents such as PhMe gave a faster reaction but
favored the amine addition product. DME gave a much slower
and unselective reaction. Interestingly, using a more hindered
base (LDA) also favored the addition product, which therefore
seems unlikely to be formed by a simple SN2 addition. One
possibility is that the epoxide is deprotonated at the terminal
carbon and then opens by rearrangement via a carbene,
followed by protonation of the lithiated intermediate.13 The
combination of lithium diethylamide in THF was optimal,
giving the desired allylic alcohol in 14:1 ratio which could be
further purified by crystallization as its citric acid salt 17.
Simpler acid salts were not crystalline.
the tertiary amine. The minimization of byproducts due to
alcohol debenzylation and reduction of potential π-allyl species
was of key concern. After a series of screening experiments,
looking at catalyst (metal and support), solvent, temperature
and pressure, we settled on a 5 wt % platinum on graphite
catalyst as optimum: a diastereomeric ratio in excess of 99:1
was achieved, and other byproducts were kept below 1%.
Unfortunately, 30 wt % of metal catalyst was required to drive
this reaction to completion, likely due to sequestering of the
catalyst by impurities such as diamine 16. Using allylic alcohol
15, purified by silica gel chromatography, catalyst loading could
be reduced to 5 wt %. However, chromatographic purification
required a high loading of silica which would be impractical on
manufacturing scale. Furthermore, the beneficial effect of
purification could not be reproduced by multiple recrystalliza-
tions of the citrate salt 17. Unbowed, we proceeded to
investigate subsequent synthetic transformations. Alcohol 18
was converted to the mesylate 19 which, unpurified, reacted
with the anion of diphenylacetonitrile, affording 7 and
intercepting route A (cf. Scheme 1). At this point, alternative
solvent mixtures for the methylation reaction were investigated
with a view to finding a system which would be compatible with
the aqueous workup required after diphenylacetonitrile
The selective reduction of allylic alcohol 15 was next
investigated. It was hoped that the desired diastereomer 18
would be favored not only by sterics, but by delivery of the
catalyst to the upper face of the olefin by the coordination with
643
dx.doi.org/10.1021/op3002933 | Org. Process Res. Dev. 2013, 17, 641−650

Organic Process Research & Development
Article
Scheme 4. Directed epoxide reduction
a
Table 2. Optimization of reaction conditions for selective reduction of epoxide 13
entry
reagent
solvent
temperature (oC)
20 (%)
21 (%)
15 (%)
22 (%)
1
NaB(CN)H3/BF3·OEt2
Et3SiH
THF
−78 to 20
−20 to 20
−20 to 20
−20 to 20
−20
−20
−20
−20
−20
−20
−78
−20
−20
12
0
54
0
11
0
0
0
0
0
0
8
0
0
0
0
0
23
0
2
PhMe
PhMe
PhMe
PhMe
PhMe
PhMe
THF
3
BH3·Et3N
BH3·pyridine
Red-Al
0
0
0
4
0
0
0
5
0
100
100
100
18
0
0
6
LiAlH4
0
0
7
LiBH4
0
0
8
LiBH4/BH3·THF
Dibal-H
8
66
98
16
3
9
THF
2
10
11
12
13
Dibal-H
PhMe
PhMe
DCM
TBME
81
89
79
81
3
Dibal-H
8
Dibal-H
7
14
13
Dibal-H
6
a
Conditions: A 0.37 M solution of epoxide 13 in the given solvent was treated with reductant (and Lewis acid) at the given temperature and stirred
until all epoxide was consumed. Ratio calculated by absolute HPLC area.
Scheme 5. Mechanistic insight into the formation of pyrollidine 22
alkylation. The diastereomeric ratio (dr) is solvent-dependent,
and the best selectivity was achieved using a 9:1 mixture of
MiBK and MeOH, resulting in an absolute dr of 9:1 and, after
ripening, an isolated dr of 22:1. Under these optimized
conditions, and after a single recrystallization, the API produced
was of comparable quality to that produced via route A.
The large platinum catalyst loading required for hydro-
genation of the non-chromatographed allylic alcohol 15 made
this route prohibitively expensive. To circumvent this problem,
we speculated that a direct reduction of epoxide 13 might be
possible by delivery of a reducing agent to the upper face of the
oxirane ring (Scheme 4).14 The effect of a range of reductants,
with and without Lewis acids, on the epoxide was next
investigated. A mixture of products was observed: the desired
internal reduction product 20, tertiary alcohol 21, derived from
reduction at the unsubstituted end of the epoxide, the allylic
alcohol 15, formed by Lewis acid-mediated isomerization, and
the skeletal rearrangement product 22 (Scheme 4).
Strong reducing agents (LiAlH4, LiBH4, Red-Al, entries 5−7)
attacked the epoxide from the least substituted end, affording
mainly 21. Milder reducing agents (borane, triethylsilane,
entries 2−4) gave no reaction at all. Lewis acidic additives
increased the amounts of products 15 and 22 (entries 1 and 8),
presumably via a cationic intermediate. Dibal-H in non-
coordinating solvent offered the best combination of Lewis
acidity and reducing power, giving the desired product in 80%
644
dx.doi.org/10.1021/op3002933 | Org. Process Res. Dev. 2013, 17, 641−650

Organic Process Research & Development
Article
Scheme 6. Route B to drug substance 1
yield at −20 °C. Toluene was selected as solvent due to the
ease of solvent swap from 2-MeTHF used in the previous
telescoped step. The proportion of alcohol 20 could be further
increased to 89% by performing the reaction at −78 °C.
Interestingly, switching from toluene to THF completely
changed the reaction course, affording 22 as the major product
(Table 2).
EXPERIMENTAL SECTION
■
General Methods. All reactions were carried out in water,
methanol, and acetone-washed glassware under a nitrogen
atmosphere. Solvents and reagents were used without any
1
purification or drying. H and 13C NMR spectra were acquired
on a Bruker spectrometer at frequencies of 400 and 100 MHz,
respectively. High-resolution mass spectra were recorded on a
linear ion trap combined with a Fourier transform ion cyclotron
resonance mass spectrometer using an electrospray ionization
source operated in positive ion mode. IR spectra were recorded
as solids or neat oils. HPLC chromatograms were recorded on a
C18(2) column at 40 °C; eluent gradient: 100% 0.05% v/v
TFA in water to 95% 0.05% v/v TFA in MeCN over 8 min.
Melting points were recorded using an automated melting
point apparatus.
To investigate the mechanism of formation of 22, we made
deuterium-labeled epoxide 13 using CD3I and D6-DMSO as the
ylid source.15 On treatment with Dibal-H in toluene at −78 °C,
only deuterium-labeled products analogous to 20 and 21 were
formed, in a 1:1 ratio (Scheme 5). Neither of the products
analogous to 15 or 22 was detected. The change in product
ratio from that of the unlabeled experiment implies a secondary
kinetic isotope effect. This provides indirect evidence that a
cationic sp2 intermediate, 23, is involved in the generation of
products 20, 15, and 22 since carbon−deuterium bonds offer
inferior hyperconjugative stabilization.16 Tertiary alcohol 21 is
formed through an SN2 mechanism, and no reduction in its rate
of formation was observed: a secondary kinetic isotope effect
would not be expected for this pathway as no cationic sp2
intermediate is formed at the reacting carbon. A mechanism for
the formation of 22 is proposed, invoking the Lewis acid-
mediated formation of a tertiary carbocation and subsequent
ring fragmentation, stabilized by the nitrogen lone pair. A
similar fragmentation pathway was observed during synthetic
work into epibatidine analogues.17
The epoxide reducation reaction has been incorporated into
the new route, Route B, thereby eliminating a step. The new
six-step route is shown in Scheme 6, split into two distinct
stages with only two isolated intermediates. The process was
demonstrated on 0.2 mol scale, and recrystallization was
performed as described in the original process to deliver drug
substance with the same crystalline form and purity as those of
the substance generated via route A. Further development and
scale up of this route to check performance in a pilot-plant
setting were thwarted by unrelated project issues.
2-(8-Methyl-8-azabicyclo[3.2.1]octan-3-yl)-1,1-diphenyle-
thanol 3. Methyl 2-(8-methyl-8-azabicyclo[3.2.1]octan-3-yl)-
acetate 2 (9.0 kg, 42.6 mol) was dissolved in dibutyl ether
(DBE, 19.9 L) and added to a solution of phenyllithium (20 wt
% in dibutyl ether, 40.7 kg, 96.2 mol), maintaining internal
temperature below 5 °C. The solution was then warmed to 20
°C and stirred until all starting material and intermediates had
been consumed as judged by HPLC. The reaction solution was
chilled to 12 °C and quenched with water (45 L), maintaining
the internal temperature below 30 °C. The resultant slurry was
then aged at 12 °C for 1 h with stirring. The product was
collected by vacuum filtration and washed with water/acetone
(10 °C, 95:5 v/v, 2 × 45 L), followed by acetone (10 °C, 22.7
L) before drying in vacuo at 50 °C to afford 2-(8-methyl-8-
azabicyclo[3.2.1]octan-3-yl)-1,1-diphenylethanol 3, 12.2 kg,
38.0 mol, 89%. δH ((CD3)2SO, 400 MHz) 1.16 (2H, d, J =
12.9 Hz), 1.61−1.96 (7H, m), 2.03 (3H, s), 2.46 (2H, d, J = 5.1
Hz), 2.88 (2H, s), 5.36 (1H, br s), 7.14 (2H, tt, J = 7.6, 1.3 Hz),
7.26 (4H, t, J = 7.6 Hz), 7.42 (4H, dd, J = 7.6, 1.3 Hz); δC
((CD3)2SO, 100 MHz) 22.8, 26.2, 37.3, 39.6, 49.0, 59.7, 77.4,
125.9, 125.9, 127.6, 148.7; νmax cm−1 (solid) 3056.7, 2912.5,
1492.3, 1446.7, 1319.0, 1265.6, 1211.5, 1171.8, 1129.0, 1062.0,
1029.8, 1011.5, 976.1, 953.5, 935.9, 909.0, 895.7, 837.0, 818.2,
779.8, 753.3, 739.5, 696.7; HRMS m/z calc’d for C22H28NO
322.2165, found 322.2159; mp 145−148 °C.
In conclusion, we have developed two synthetic routes to an
investigational muscarinic antagonist drug 1. Route A took
advantage of the ready availability of tropane ester 2 within our
network and was used to make 5 kg of active substance in our
pilot plant. The inefficient dealkylation/alkylation sequence led
to the development of a second route, which employs readily
available starting materials and provides access to 1 in 6
synthetic steps with just two isolated intermediates.
Laboratory-Scale Reaction. 3-(8-Methyl-8-azabicyclo-
[3.2.1]octan-3-yl)-2,2-diphenylpropanenitrile Hydrochloride
4. Trimethylsilylcyanide (36.0 mL, 273 mmol) was added to
a suspension of 2-(8-methyl-8-azabicyclo[3.2.1]octan-3-yl)-1,1-
diphenylethanol 3 (25.0 g, 77.8 mmol) in DCM (250 mL) at
645
dx.doi.org/10.1021/op3002933 | Org. Process Res. Dev. 2013, 17, 641−650

Organic Process Research & Development
Article
−5 °C. Tin tetrachloride (32.0 mL, 273 mmol) was added
dropwise to the suspension at such a rate to maintain the
internal temperature at −5 °C. The suspension was then
warmed to 25 °C over 30 min and stirred for 2 h until the
reactions was considered complete by HPLC. The slurry was
added slowly to aqueous NaOH solution (5 M, 500 mL) at 0
°C over 5−10 min. The DCM layer was separated, and the
aqueous layer washed with DCM (2 × 50 mL). The organic
layers were combined and washed with water (2 × 50 mL), and
the organic solvent was evaporated by vacuum distillation and
MIBK (300 mL) added. The solution was then concentrated to
200 mL by vacuum distillation and passed through a Cuno
Zetacarbon filter, eluting with an additional 200 mL of MIBK.
The solution was then concentrated to 300 mL, and a solution
of HCl in 2-propanol (5 M, 17.2 mL, 86 mmol) was added at
45 °C over 45 min. The resulting slurry was stirred for 2 h at 80
°C, then cooled to 20 °C over 1 h, and aged for 1 h. The
product was collected by vacuum filtration and washed with
MIBK (2 × 50 mL) before drying in vacuo at 50 °C to afford 3-
(8-methyl-8-azabicyclo[3.2.1]octan-3-yl)-2,2-diphenylpropane-
nitrile hydrochloride 4 (20.0 g, 54.4 mmol, 70%) as a white
solid. δH ((CD3)2SO, 400 MHz) 1.66 (2H, d, J = 14.4 Hz),
1.72−1.83 (1H, m), 2.10−2.24 (4H, m), 2.30−2.43 (2H, m),
2.55 (3H, s), 2.93 (2H, d, J = 5.3 Hz), 3.61−3.80 (2H, m), 7.33
(2H, tt, J = 7.3, 1.5 Hz), 7.42 (4H, t, J = 7.3 Hz), 7.48 (4H, dd,
J = 7.3, 1.5 Hz), 10.64 (1H, br s); δC ((CD3)2SO, 100 MHz)
22.9, 23.9, 34.1, 37.8, 44.8, 51.9, 61.8, 122.3, 126.8, 128.0,
129.0, 139.7; νmax cm−1 (solid) 2944.7, 2483.1, 1478.7, 1449.2,
1393.8, 1335.5, 1088.3, 1052.1, 968.6, 945.6, 881.7, 831.7,
760.6, 747.5, 694.5, 663.3; HRMS m/z calc’d for C23H27N2
331.2169, found 331.2168; mp 245−247 °C.
Single Pilot-Plant Scale Batch. 3-(8-Methyl-8-
azabicyclo[3.2.1]octan-3-yl)-2,2-diphenylpropanenitrile Hy-
drochloride 4. Trimethylsilylcyanide (9.35 kg, 94.2 mol) was
added to a suspension of 2-(8-methyl-8-azabicyclo[3.2.1]octan-
3-yl)-1,1-diphenylethanol 3 (8.65 kg, 26.9 mol) in DCM (86.5
L) at −5 °C over 7 min, followed by a line-wash of DCM (0.5
L). Tin tetrachloride (24.5 kg, 94.0 mol) was added over 75
min to the suspension at such a rate to maintain the internal
temperature at −5 °C, followed by a line-wash of DCM (0.5 L).
The suspension was then warmed to 25 °C over 30 min and
stirred for 2 h until considered complete by HPLC. The slurry
was added slowly to aqueous NaOH solution (5 M, 86.5 L) at 0
°C over 90 min, following by a line-wash of DCM (8.65 L).
The DCM layer was separated, and the aqueous layer was
washed with DCM (2 × 17.3 L). The organic layers were
combined and washed with water (2 × 17.3 L), and the organic
solvent was evaporated by vacuum distillation and MIBK (34.6
L) added. The solution was then concentrated to 34.6 L by
vacuum distillation and passed through a Cuno Zetacarbon
filter, eluting with an additional 17.3 L of MIBK. The solution
was then concentrated to 51.9 L, and a solution of HCl in 2-
propanol (5 M, 5.54 kg, 29.6 mol) was added at 45 °C over 45
min. The resulting slurry was stirred for 1 h at 80 °C, then
cooled to 20 °C over 1 h, and aged for at least 1 h. The product
was collected by vacuum filtration and washed with MIBK
(52.0 L, 17.3 L) before drying in vacuo at 50 °C to afford 3-(8-
methyl-8-azabicyclo[3.2.1]octan-3-yl)-2,2-diphenylpropaneni-
trile hydrochloride 4 (3.15 kg, 8.59 mol, 49.6%) as a white
solid.
chloride 5 (15.0 g, 7.55 mol) was suspended in toluene (150
mL), and aqueous sodium hydroxide (1.33 M, 45 mL) was
added before stirring for 30 min. The aqueous layer was
removed, and the organics were distilled down to 75 mL under
atmospheric pressure. The solution was cooled to 50 °C, and
sodium bicarbonate (7.5 g, 88.2 mmol) was added followed by
1-chloroethyl chloroformate (11.1 mL, 102 mmol) at such a
rate that the temperature remained below 55 °C. The reaction
was stirred for 2 h until considered complete by HPLC, at
which point, water (45 mL) was added and stirring continued
for a further 1 h at 50 °C. A solution of aqueous potassium
carbonate (18 g in 45 mL) was then added carefully to the
reaction at 80 °C and stirred for 30 min, and the layers were
allowed to settle, and the aqueous layer was removed. The
organic layer was washed successively at 80 °C with water (45
mL) and brine (11.25 g NaCl in 30 mL water). A solution of
HCl in 2-propanol (5 M, 9.0 mL, 45 mmol) was added slowly
to the organic fraction at 80 °C and cooled to 60 °C. The
mixture was stirred for 30 min at 60 °C and then cooled to 20
°C. The slurry was stirred for at least 60 min and product
collected by filtration. The filter cake was washed with TBME
(2 × 75 mL) and the product dried in vacuo at 50 °C to afford
3-(8-azabicyclo[3.2.1]octan-3-yl)-2,2-diphenylpropanenitrile
hydrochloride 6 (12.3 g, 34.9 mmol, 85%) as a white solid. δH
(CDCl3, 400 MHz) 1.63 (2H, d, J = 15.4 Hz), 2.00−2.09 (2H,
m), 2.11−2.19 (1H, m), 2.27−2.38 (2H, m), 2.45 (2H, ddd, J =
15.4, 7.6, 3.5 Hz), 2.71 (2H, d, J = 5.8 Hz), 3.92−4.00 (2H, m),
7.28−7.41 (10H, m) 9.47 (2H, br s); δC (CDCl3, 100 MHz)
24.6, 25.8, 33.8, 46.9, 51.4, 54.0, 122.1, 126.9, 128.2, 129.1,
139.3; νmax cm−1 (solid) 3365.1, 2883.1, 2787.9, 1636.4, 1490.2,
1447.3, 1410.5, 1382.4, 1207.7, 1035.4, 951.1, 770.9, 756.3,
747.4, 711.3, 693.4, 656.2; HRMS m/z calc’d for C22H25N2
317.2012, found 317.2006; mp 221−223 °C.
Pilot-Plant Scale Reaction. 3-(8-Azabicyclo[3.2.1]octan-
3-yl)-2,2-diphenylpropanenitrile Hydrochloride 6. 3-(8-Meth-
yl-8-azabicyclo[3.2.1]octan-3-yl)-2,2-diphenylpropanenitrile hy-
drochloride 5 (2.77 kg, 7.55 mol) was suspended in toluene
(5.4 L) and acetone (8.1 L); diisopropylethylamine (2.38 kg,
18.4 mol) was added over 10 min and the resulting solution
heated to 50 °C. 1-Chloroethyl chloroformate (2.65 kg, 18.5
mmol) was then added over 1 h at such a rate that the
temperature remained below 55 °C. The solution was stirred
for at least 2 h until the reaction was considered complete by
HPLC. The reaction was then distilled under atmospheric
conditions to ∼17.6 L, diluted with MIBK (32.4 L), and heated
to 80 °C. A solution of potassium carbonate (3.24 kg, 23.4
mmol, in 8.1 L water) was then added to the reaction at 80 °C
over 1 h, the biphasic mixture stirred for 30 min, the layers
allowed to settle, and the aqueous layer removed. The organic
layer was then washed with water at 80 °C (8.1 L) and brine
(2.03 kg NaCl, in 5.4 L water). A solution of HCl in 2-propanol
(5 M, 1.49 kg, 8.3 mol) was added over 30 min to the organic
fraction at 80 °C and then the reaction cooled to 60 °C. The
resulting slurry was stirred for 30 min at 60 °C and then cooled
to 20 °C. The slurry was stirred for at least 60 min and product
collected by filtration. The filter cake was washed with MIBK (2
× 13.5 L) and the product dried in vacuo at 50 °C to afford 3-
(8-azabicyclo[3.2.1]octan-3-yl)-2,2-diphenylpropanenitrile hy-
drochloride 6 (2.54 kg, 7.36 mol, 89.2%) as a white solid.
The isolated product was contaminated with ethylisopropyl-
amine and diisopropylamine hydrochloride.
Laboratory Scale. 3-(8-Azabicyclo[3.2.1]octan-3-yl)-2,2-
diphenylpropanenitrile Hydrochloride 6. 3-(8-Methyl-8-
azabicyclo[3.2.1]octan-3-yl)-2,2-diphenylpropanenitrile hydro-
8-(2-(Benzyloxy)ethyl)-3-(2-cyano-2,2-diphenylethyl)-8′-
methyl-8-azabicyclo[3.2.1]octan-8-ium Iodide 1. To a stirred
646
dx.doi.org/10.1021/op3002933 | Org. Process Res. Dev. 2013, 17, 641−650

Organic Process Research & Development
Article
suspension of 3-(8-azabicyclo[3.2.1]octan-3-yl)-2,2-diphenyl-
propanenitrile hydrochloride 6 (2.54 kg, 7.20 mol) in acetone
(15.1 L) were added potassium carbonate (2.99 kg, 21.6 mol)
followed by benzylbromoethyl ether (1.86 kg, 8.67 mol) and
the resultant suspension heated to 60 °C for at least 18 h. The
suspension was cooled to 20 °C and filtered to remove
inorganic material. The filter cake was washed with acetone (2
× 9.6 L), and the combined organics were distilled down to
17.9 L. The solution was cooled to 0 °C, and methyl iodide
(1.23 kg, 8.67 mol) was added, maintaining temperature below
0 °C. The resulting slurry was allowed to warm to 20 °C and
stirred for 18 h. MeOH (3.81 L) was added and the slurry
heated at 55 °C for at least 18 h. After the slurry cooled to 20
°C and aged for 1 h, the precipitated product was collected by
filtration. The filter cake was washed with acetone (3 × 10.2 L)
and the product dried in vacuo at 50 °C to afford intermediate
grade 8-(2-(benzyloxy)ethyl)-3-(2-cyano-2,2-diphenylethyl)-8-
methyl-8-azabicyclo[3.2.1]octan-8-ium iodide 1 (3.06 kg, 5.16
mol, 72%) as a white solid. A portion of the resultant material
(700 g, 1.18 mol) was recrystallized from 17.5 L of 5% aqueous
MeOH to afford 8-(2-(benzyloxy)ethyl)-3-(2-cyano-2,2-diphe-
nylethyl)-8′-methyl-8-azabicyclo[3.2.1]octan-8-ium iodide 1
(593 g, 1.00 mol, 85%) as a crystalline white solid. δH
(CD3OD, 400 MHz) 1.74 (2H, d, J = 16.4 Hz), 2.10−2.21
(1H, ddd, J = 14.9, 9.3, 5.9 Hz), 2.26−2.48 (4H, m), 2.56 (2H,
ddd, J = 16.4, 9.3, 3.8 Hz), 2.97 (2H, d, J = 5.9 Hz), 3.02 (3H,
s), 3.68 (2H, t, J = 5.1 Hz), 3.85−3.92 (4H, m), 4.56 (2H, s),
7.27−7.37 (7H, m), 7.41 (4H, t, J = 7.6 Hz), 7.47 (4H, dd, J =
7.6, 1.7 Hz); δC (CD3OD, 100 MHz) 22.6, 25.4, 33.5, 47.0,
50.3, 53.6, 56.3, 65.1, 70.3, 74.6, 123.7, 128.4, 129.1, 129.3,
129.4, 129.8, 130.4, 138.9, 141.4; νmax cm−1 (solid) 2973.3,
2946.7, 2872.8, 2231.5, 1600.6, 1582.9, 1495.9, 1457.5, 1448.3,
1412.1, 1356.0, 1124.1, 1048.4, 1025.8, 915.2, 900.6, 779.4,
755.0, 741.0, 714.3, 695.4; HRMS m/z calc’d for C32H37ON2
465.2900, found 465.2887; ion chromatography: I−, expected
21.4%, found 21.9%; mp 218−220 °C.
O-Benzylethanolamine 12:9 Ethanolamine (40.5 g, 0.663
mol) was added dropwise to a suspension of sodium hydride
(60% dispersion in mineral oil, 26.5 g, 0.663 mol) in THF (640
mL) at 0 °C. The temperature was then raised to reflux and the
solution aged for 30 min, at which point, benzyl chloride (75.5
g, 0.596 mol) was added dropwise. The solution was aged at
reflux for 2 h, then cooled to 20 °C and quenched with water
(100 mL). The THF was removed by distillation and the pH
adjusted with 1 M aqueous HCl (200 mL). The aqueous layer
was extracted with DCM (3 × 150 mL) and the remaining
aqueous fraction basified with 5% aqueous NaOH (70 mL).
This was extracted with DCM (3 × 150 mL), and the
combined organic fractions were dried over Na2SO4, filtered,
and condensed in vacuo. The resultant oil was distilled (1.5
mbar, 100−102 °C) to afford O-benzylethanolamine 12 (69.0
g, 0.456 mol, 77%) as a colourless oil. δH (CDCl3, 400 MHz)
1.28 (2H, br s), 2.91 (2H, t, J = 5.1 Hz), 3.53 (2H, t, J = 5.1
Hz), 4.55 (2H, s), 7.28−7.39 (5H, m).
(8-(2-(Benzyloxy)ethyl)-8-azabicyclo[3.2.1]octan-3-yl)-
methanol Hydrochloride 18. 2,5-Dimethoxyfuran (35.2 g,
0.266 mol) was heated to 80 °C for 30 min in water (100 mL)
with concentrated HCl (37%, 2.80 mL, 0.034 mol). The
solution was then cooled to 0 °C. In a separate vessel, acetone-
1,3-dicarboxylic acid (43.6 g, 0.298 mol) was added to a
solution of NaOAc (61.2 g, 0.746 mol), O-benzylethanolamine
12 (32.2 g, 0.213 mol), and concentrated HCl (37%, 2.98 mL,
0.036 mol) in water (200 mL) at 5 °C. The furan (dialdehyde)
solution was then added and washed in with an additional 100
mL water. The solution was aged at 5 °C for 10 min, heated to
45 °C over 2 h, and aged at 45 °C until considered complete by
HPLC. NaCl (150 g) was added followed by a solution of
NaOH (25 g) in water (30 mL). The biphasic solution was
then extracted with 2-MeTHF (3 × 200 mL). The combined
organics were then distilled down to a volume of 100 mL at
atmospheric pressure to afford a solution of 8-(2-(benzyloxy)-
ethyl)-8-azabicyclo[3.2.1]octan-3-one 11. In a separate vessel, a
t
8′-(2-(Benzyloxy)ethyl)-3-(2-cyano-2,2-diphenylethyl)-8-
methyl-8-azabicyclo[3.2.1]octan-8-ium Iodide 5. A sample of
mother liquors after crystallization of 8-(2-(benzyloxy)ethyl)-3-
(2-cyano-2,2-diphenylethyl)-8′-methyl-8-azabicyclo[3.2.1]-
octan-8-ium iodide 1 was evaporated to dryness to afford a
yellow oily solid, 19.4 g. This was triturated with acetone (40
mL) for 15 min and filtered, washing the cake with acetone (2
× 10 mL). The liquors were again evaporated to afford a yellow
solid and further triturated with acetone (40 mL) for 15 min.
The slurry was filtered, washing the cake with acetone (2 × 10
mL). Finally, the liquors were taken up in acetone (50 mL) at
55 °C. The mixture was allowed to cool to ambient
temperature and aged overnight. The resultant slurry was
filtered and washed with acetone (2 × 10 mL) and dried in
vacuo to afford 8′-(2-(benzyloxy)ethyl)-3-(2-cyano-2,2-diphe-
nylethyl)-8-methyl-8-azabicyclo[3.2.1]octan-8-ium iodide 5
(1.5 g) as a white solid. δH (MeOD, 400 MHz) 1.78 (2H, d,
J = 16.4 Hz), 2.15 (1H, ddd, J = 14.7, 9.1, 5.6 Hz), 2.25−2.35
(2H, m), 2.40−2.53 (4H, m), 2.97 (2H, d, J = 5.6 Hz), 3.09
(3H, s), 3.54 (2H, t, J = 4.2 Hz), 3.89−3.98 (4H, m), 4.56 (2H,
s), 7.26−7.37 (7H, m), 7.40 (4H, t, J = 7.6 Hz), 7.47 (4H, d, J
= 8.3 Hz); δC (MeOD, 100 MHz) 23.1, 25.7, 33.6, 42.2, 50.1,
53.5, 61.9, 65.0, 69.2, 74.5, 123.7, 128.4, 129.2, 129.2, 129.4,
129.7, 130.3, 138.8, 141.4; νmax cm−1 (solid) 2943.9, 1492.8,
1449.3, 1358.4, 1219.0, 1105.6, 1084.4, 1019.9, 931.7, 899.5,
750.0, 696.7; HRMS m/z calc’d for C32H37N2O 465.2900,
found 465.2890; mp 161−165 °C.
solution of KOtBu in BuOH (1 M, 266 mL, 0.266 mol) was
added to a stirred suspension of trimethylsulfoxonium iodide
(58.7 g, 0.266 mol) in 2-MeTHF (200 mL) at 20 °C and the
mixture aged for 1 h. The solution of 8-(2-(benzyloxy)ethyl)-8-
azabicyclo[3.2.1]octan-3-one 11 was then added over 20 min
and washed in with a further 50 mL 2-MeTHF. The solution
was then heated to 75 °C for at least 16 h until considered
complete by HPLC. Toluene (200 mL) was added and the
reaction quenched at 50 °C by addition of water (300 mL).
The aqueous fraction was decanted and the organics washed
with water (200 mL) and distilled down to 200 mL. Toluene
(200 mL) was added and the solution again distilled down to
200 mL to afford a solution of 8-(2-(benzyloxy)ethyl)-8-
azaspiro[bicyclo[3.2.1]octane-3,2′-oxirane] 13 which was
cooled to −25 °C. A solution of Dibal-H in toluene (1.5 M,
213 mL, 0.320 mol) was charged to a clean dry vessel and
cooled to −78 °C. The solution of 8-(2-(benzyloxy)ethyl)-8-
azaspiro[bicyclo[3.2.1]octane-3,2′-oxirane] 13 was added over
at least 1.5 h, maintaining temperature below −60 °C, and
washed in with a further 2 × 50 mL toluene. The solution was
aged at −60 °C for 30 min, until complete epoxide
consumption as judged by HPLC, and quenched by slow
addition of ethyl acetate (100 mL). The solution was then
added slowly to a saturated solution of Rochelle’s salt (500 mL)
at 0 °C and aged with vigorous stirring for 4 h. The aqueous
layer was decanted, and the organics were washed with water
(200 mL). The solution was concentrated to 300 mL, and
647
dx.doi.org/10.1021/op3002933 | Org. Process Res. Dev. 2013, 17, 641−650

Organic Process Research & Development
Article
toluene (300 mL) was added. The solution was again distilled
down to 300 mL and cooled to 65 °C, where 2-propanol (11.4
mL) was added. A solution of HCl in 2-propanol (5 M, 42.6
mL, 0.213 mol) was added dropwise, and the mixture was
seeded with 100 mg of (8-(2-(benzyloxy)ethyl)-8-
azabicyclo[3.2.1]octan-3-yl)methanol hydrochloride 18 at 60
°C and aged for 30 min. The suspension was cooled to 10 °C
over 2 h, aged for 2 h, filtered, washed with toluene (3 × 100
mL) and dried in vacuo to afford (8-(2-(benzyloxy)ethyl)-8-
azabicyclo[3.2.1]octan-3-yl)methanol hydrochloride 18 (30.3 g,
97.2 mmol, 46%) as an off-white solid: δH (CDCl3, 400 MHz)
1.83−2.06 (5H, m), 2.20−2.36 (4H, m), 3.19−3.36 (2H, m),
3.59 (2H, d, J = 8.6 Hz), 3.83 (2H, t, J = 5.0 Hz), 3.92−4.04
(2H, m), 7.28−7.41 (5H, m); δC (CDCl3, 100 MHz) 25.6,
31.2, 31.4, 52.7, 63.7, 65.9, 66.6, 74.5, 129.2, 129.3, 129.7,
139.0; νmax cm−1 (solid) 3280.7, 2984.7, 2862.8, 2648.4, 2591.2,
2549.7, 1477.3, 1451.7, 1369.6, 1324.1, 1253.4, 1114.8, 1085.8,
1057.8, 1031.3, 980.9, 940.2, 917.8, 829.0, 735.1, 695.1, 655.0;
HRMS m/z calc’d for C17H26NO2 276.1958, found 276.1956;
mp 144−146 °C.
with water (100 mL) and most of the THF removed by
distillation. The bi-phasic mixture was then extracted with
MiBK (3 × 100 mL) and distilled down to 120 mL to afford a
solution of 3-(8-(2-(benzyloxy)ethyl)-8-azabicyclo[3.2.1]octan-
3-yl)-2,2-diphenylpropanenitrile 7. MeOH (11 mL) was added
and the solution cooled to 0 °C. Iodomethane (4.65 mL, 74.6
mmol) was added and after aging for 1 h the mixture allowed to
warm to 20 °C and aged for a further 16 h. The slurry was
heated to 60 °C for 6 h, then cooled to 20 °C over 1 h and aged
for a further 1 h. The slurry was filtered and the resulting solid
washed with MiBK (2 × 100 mL) and dried in vacuo to afford
8-(2-(benzyloxy)ethyl)-3-(2-cyano-2,2-diphenylethyl)-8′-meth-
yl-8-azabicyclo[3.2.1]octan-8-ium iodide 1 (28.0 g, 47.3 mmol,
82%) as an off-white solid. The resultant material (10.0 g, 16.9
mmol) was recrystallized from 250 mL 5% aqueous MeOH to
afford 8-(2-(benzyloxy)ethyl)-3-(2-cyano-2,2-diphenylethyl)-
8′-methyl-8-azabicyclo[3.2.1]octan-8-ium iodide 1 (8.2 g, 13.8
mmol, 82%) as a crystalline white solid. Analytical data match
those previously reported.
Intermediates 19 and 7 were isolated and subsequently
characterized by evaporation of in-process solutions and
purification.
Intermediates 11 and 13 were isolated and subsequently
characterized by evaporation of in process solutions and
purification.
(8-(2-(Benzyloxy)ethyl)-8-azabicyclo[3.2.1]octan-3-yl)-
methyl methanesulfonate 19 was purified by silica gel
chromatography, eluting with 0−10% MeOH in TBME: δH
(CDCl3, 400 MHz) 1.41 (2H, d, J = 13.9 Hz), 1.49−1.57 (2H,
m), 1.97−2.08 (2H, m), 2.09−2.22 (3H, m), 2.62 (2H, t, J =
6.1 Hz), 3.00 (3H, s), 3.24−3.33 (2H, m), 3.60 (2H, t, J = 6.1
Hz), 4.18 (2H, d, J = 7.3 Hz), 4.54 (2H, s), 7.25−7.38 (5H, m);
δC (CDCl3, 100 MHz) 26.9, 27.9, 31.1, 37.4, 51.6, 58.6, 69.7,
73.1, 74.0, 127.5, 127.5, 128.3, 138.3; νmax cm−1 (oil) 2922.3,
2851.3, 1454.1, 1352.9, 1264.6, 1172.4, 1086.2, 1027.9, 948.4,
887.9, 846.5, 811.9, 731.5, 698.1; HRMS m/z calc’d for
C18H28NO4S 354.1734, found 354.1734.
3-(8-(2-(Benzyloxy)ethyl)-8-azabicyclo[3.2.1]octan-3-yl)-
2,2-diphenylpropanenitrile 7 was purified by silica gel
chromatography, eluting with 0−5% MeOH in TBME: δH
(CDCl3, 400 MHz) 1.25 (d, J = 13.9 Hz), 1.63−1.71 (2H, m),
1.89−2.02 (3H, m), 2.08 (2H, ddd, J = 13.4, 8.3, 5.2 Hz), 2.54
(2H, t, J = 6.3 Hz), 2.62 (2H, d, J = 5.8 Hz), 3.13−3.21 (2H,
m), 3.53 (2H, t, J = 6.3 Hz), 4.50 (2H, s), 7.23−7.41 (15H, m);
δC (CDCl3, 100 MHz) 25.2, 26.8, 36.5, 48.4, 51.4, 51.6, 58.7,
69.9, 73.1, 122.7, 127.0, 127.5, 127.5, 127.8, 128.3, 128.8, 138.5,
140.4; νmax cm−1 (oil) 2928.4, 2236.0, 1598.8, 1494.0, 1449.3,
1320.2, 1205.6, 1073.1, 1029.3, 1001.9, 909.3, 818.6, 731.1,
694.8; HRMS m/z calc’d for C31H35N2O 451.2744, found
451.2738.
(8-(2-(Benzyloxy)ethyl)-8-azabicyclo[3.2.1]oct-3-en-3-yl)-
methanol Citrate 17. Diethylamine (48.0 mL, 0.464 mol) was
added dropwise to a solution of nBuLi in THF (2.5 M, 185 mL,
0.464 mol) maintaining temperature below 0 °C. After aging
for 30 min, the solution was cooled to −20 °C and a solution of
8-(2-(benzyloxy)ethyl)-8-azaspiro[bicyclo[3.2.1]octane-3,2′-ox-
irane] 13 (50.7 g, 0.185 mol) in THF (70 mL) added over 30
min, maintaining temperature below −10 °C, and washed in
with an additional 30 mL THF. After 1 h water (500 mL) was
added maintaining temperature below 20 °C, and the THF
removed by distillation. The residue was extracted with 2-
MeTHF (300 mL) and the solution dried over Na2SO4 and
filtered. The resultant solution was passed through a Cuno
Zetacarbon filter, eluting with 2-MeTHF (200 mL) and
evaporated to dryness. The resultant oil was taken up in IPA
(500 mL) and 2-MeTHF (200 mL) and heated to 75 °C. A
8-(2-(Benzyloxy)ethyl)-8-azabicyclo[3.2.1]octan-3-one 11,
purified by silica gel chromatography, eluting with 20−80%
EtOAc in heptane: δH (CDCl3, 400 MHz) 1.55−1.65 (2H, m),
1.99−2.10 (2H, m), 2.18 (2H, dd, J = 16.1, 1.5 Hz), 2.69 (2H,
dd, J = 16.1, 4.4 Hz), 2.86 (2H, t, J = 5.8 Hz), 3.57−3.64 (2H,
m), 3.69 (2H, t, J = 5.8 Hz), 4.58 (2H, s), 7.28−7.40 (5H, m);
δC (CDCl3, 100 MHz) 27.9, 47.2, 49.9, 59.3, 69.8, 73.2, 127.5,
127.6, 128.4, 138.2, 210.0; νmax cm−1 (oil) 2951.3, 1710.1,
1496.3, 1453.5, 1411.0, 1347.7, 1318.5, 1277.8, 1196.4, 1098.3,
1071.7, 1027.9, 1007.8, 941.7, 907.4, 840.4, 735.5, 697.2;
HRMS m/z calc’d for C16H22NO2 260.1645, found 260.1643.
8-(2-(Benzyloxy)ethyl)-8-azaspiro[bicyclo[3.2.1]octane-3,2′-
oxirane] 13 purified by silica gel chromatography, eluting with
30−80% EtOAc in heptane: δH (CDCl3, 400 MHz) 1.12 (2H,
d, J = 14.2 Hz), 1.92−2.06 (4H, m), 2.38 (2H, dd, J = 14.2, 3.2
Hz), 2.39 (2H, s), 2.69 (2H,, J = 6.3 Hz), 3.31−3.39 (2H, m),
3.62 (2H, t, J = 6.3 Hz), 4.56 (2H, s), 7.25−7.39 (5H, m); δC
(CDCl3, 100 MHz) 26.1, 38.7, 48.4, 51.5, 54.7, 59.5, 69.9, 73.1,
127.5, 127.6, 128.3, 138.5; νmax cm−1 (oil) 3030.4, 2942.2,
1496.1, 1453.8, 1346.9, 1325.4, 1218.4, 1101.5, 1072.1, 1055.0,
1028.3, 1010.8, 940.2, 903.8, 843.5, 785.6, 734.1, 696.8; HRMS
m/z calc’d for C17H23NO2 274.1802, found 274.1798.
8-(2-(Benzyloxy)ethyl)-3-(2-cyano-2,2-diphenylethyl)-8′-
methyl-8-azabicyclo[3.2.1]octan-8-ium iodide 1. Triethylamine
(16.0 mL, 115 mmol) and methanesulfonyl chloride (5.33 g,
68.9 mmol) were added sequentially to a stirred suspension of
(8-(2-(benzyloxy)ethyl)-8-azabicyclo[3.2.1]octan-3-yl)-
methanol hydrochloride 18 (17.9 g, 57.4 mmol) in THF (150
mL) at 0 °C. After 1 h the resultant slurry was filtered and the
cake washed with a further 100 mL THF. The combined
organics were distilled down to 100 mL to afford a solution of
(8-(2-(benzyloxy)ethyl)-8-azabicyclo[3.2.1]octan-3-yl)methyl
methanesulfonate 19. In a separate vessel, a solution of KOtBu
in tBuOH (1 M, 115 mL, 115 mmol) was added to a solution of
diphenylacetonitrile (22.2 g, 115 mmol) in THF (80 mL) at 20
°C. After 1 h the solution of (8-(2-(benzyloxy)ethyl)-8-
azabicyclo[3.2.1]octan-3-yl)methyl methanesulfonate 19 was
added and washed in with a further 20 mL THF; the solution
was heated to reflux for at least 1 h until complete consumption
of mesylate as indicated by HPLC. The reaction was quenched
648
dx.doi.org/10.1021/op3002933 | Org. Process Res. Dev. 2013, 17, 641−650

Organic Process Research & Development
Article
solution of citric acid (39.2 g, 0.204 mol) in IPA (300 mL) at
75 °C was added and the mixture cooled to 65 °C. The
solution was seeded with 100 mg (8-(2-(benzyloxy)ethyl)-8-
azabicyclo[3.2.1]oct-3-en-3-yl)methanol citrate 17 and aged at
65 °C for 1 h. The slurry was then cooled to 20 °C over 3 h and
aged at 20 °C for 3 h. The resultant slurry was filtered, and the
cake was washed with 2-MeTHF (2 × 200 mL) and dried in
vacuo to afford (8-(2-(benzyloxy)ethyl)-8-azabicyclo[3.2.1]oct-
3-en-3-yl)methanol citrate 17 (56.9 g, 0.122 mol, 66%) as a
white solid: δH ((CD3)2SO, 400 MHz) 1.62−1.74 (1H, m),
1.95 (2H, m), 2.04−2.26 (2H, m), 2.48 (2H, d, J = 15.2 Hz),
2.55 (2H, d, J = 15.2 Hz), 2.55−2.64 (1H, m), 3.15 (2H, t, J =
5.1 Hz), 3.69 (1H, dd, J = 11.2, 5.1 Hz), 3.73 (1H, dd, J = 11.2,
5.1 Hz), 3.79 (2H, s), 3.94 (1H, dd, J = 6.1, 5.1 Hz), 4.03 (1H,
t, J = 5.4 Hz), 4.49 (2H, s), 5.74 (1H, d, J = 5.4 Hz), 7.23−7.38
(5H, m); δC ((CD3)2SO, 100 MHz) 27.1, 31.8, 31.9, 58.5, 58.6,
62.8, 65.7, 71.5, 72.2, 119.4, 127.6, 127.6, 128.4, 136.8, 137.9,
171.5, 176.9; νmax cm−1 (solid) 3378.6, 2860.9, 1726.1, 1575.8,
1433.0, 1399.5 1349.4, 1316.8, 1251.1, 1206.4, 1171.4, 1108.3,
1075.0, 1038.9, 1012.7, 979.6, 913.0, 892.8, 846.8, 791.9, 748.5,
699.1, 664.9; HRMS m/z calc’d for C17H23NO2 274.1802,
found 274.1805; mp 120−122 °C.
and stirred with saturated aqueous Rochelle’s salt (20 mL) for 2
h. The biphasic solution was extracted with EtOAc (2 × 20
mL), dried over Na2SO4, filtered, and condensed in vacuo. The
resultant oil was purified by silica gel chromatography, eluting
with 0−10% 2 M NH3/MeOH in DCM to afford 3-(1-(2-
(benzyloxy)ethyl)-4-methylenepyrrolidin-2-yl)propan-1-ol 22
(77.6 mg, 0.282 mmol, 77%) as an orange oil: δH (CDCl3,
400 MHz) 1.66−1.83 (4H, m), 2.26−2.37 (2H, m), 2.41−2.53
(2H, m), 2.61−2.70 (1H, m), 3.08−3.23 (2H, m), 3.60−3.70
(2H, m), 3.94 (1H, d, J = 12.0 Hz), 3.98 (1H, d, J = 12.0 Hz),
4.53 (2H, s), 4.87 (1H, s), 5.10 (1H, s), 7.25−7.38 (5H, m); δC
(CDCl3, 100 MHz) 22.7, 27.6, 38.3, 54.1, 54.5, 63.8, 66.8, 68.9,
73.2, 116.5, 127.6, 127.7, 128.3, 138.2, 145.7; νmax cm−1 (oil)
3371.8, 3066.5, 2860.2, 1647.3, 1496.3, 1453.9, 1359.4, 1305.8,
1250.1, 1205.5, 1094.1, 1028.5, 960.9, 900.8, 848.6, 734.9,
969.8; HRMS m/z calc’d for C17H26O2N 276.1958, found
276.1960.
8-(2-(Benzyloxy)ethyl)-3-methyl-8-azabicyclo[3.2.1]octan-
3-ol 21. A solution of RedAl in PhMe (3.5 M, 0.157 mL, 0.549
mmol) was added dropwise to a solution of 8-(2-(benzyloxy)-
ethyl)-8-azaspiro[bicyclo[3.2.1]octane-3,2′-oxirane] 13 (100
mg, 0.366 mmol) in toluene (1 mL) at −30 °C. After 30
min, the reaction was quenched with EtOAc (1 mL) and stirred
with saturated, aqueous Rochelle’s salt (20 mL) for 2 h. The
reaction was extracted with EtOAc (2 × 20 mL), dried over
Na2SO4, filtered, and condensed in vacuo. The resultant oil was
purified by silica gel chromatography, eluting with 0−5% 2 M
NH3/MeOH in DCM to afford 8-(2-(benzyloxy)ethyl)-3-
methyl-8-azabicyclo[3.2.1]octan-3-ol 21 (82.6 mg, 0.300
mmol, 82%) as colourless oil: δH (CDCl3, 400 MHz) 1.12
(1H, br s), 1.16 (3H, s), 1.57 (2H, d, J = 14.2 Hz), 1.83−1.97
(4H, m), 2.02−2.10 (2H, m), 2.62 (2H, t, J = 6.4 Hz), 3.18−
3.26 (2H, m), 3.59 (2H, t, J = 6.4 Hz), 4.54 (2H, s), 7.23−7.38
(5H, m); δC (CDCl3, 100 MHz) 25.8, 34.1, 44.1, 51.4, 58.9,
69.0, 69.9, 73.0, 127.4, 127.5, 128.3, 138.4.
Various impurities were isolated from reaction mixtures
during route investigation work; their isolation is documented
below:
4-(2-(Benzyloxy)ethyl)octahydro-1H-3,7-methanoindolizin-
4-ium-7-olate 14 was isolated as a hygroscopic solid by filtration
of the white precipitate formed when excess sulfoxonium ylid is
added to 8-(2-(benzyloxy)ethyl)-8-azabicyclo[3.2.1]octan-3-
one 13 during the epoxidation reaction: δH (MeOD, 400
MHz) 1.74 (2H, d, J = 12.5 Hz), 1.90−1.97 (2H, m), 2.00−
2.08 (2H, m), 2.38 (2H, t, J = 10.8 Hz), 2.47−2.55 (2H, m),
3.40−3.45 (2H, m), 3.86−3.93 (2H, m), 3.94−4.00 (2H, m),
4.17−4.25 (2H, m), 4.60 (2H, s), 7.28−7.41 (5H, m); δC
(MeOD, 100 MHz) 31.1, 34.5, 43.1, 53.4, 60.4, 63.3, 64.9, 68.3,
74.5, 129.2, 129.3, 129.7, 138.9; νmax cm−1 (solid) 3292.0,
2960.8, 2881.8, 1591.0, 1453.0, 1435.9, 1363.9, 1352.7, 1327.6,
1263.7, 1225.5, 1207.2, 1152.6, 1139.7, 1111.0, 1074.1, 1032.9,
1020.4, 990.5, 956.4, 929.1, 886.0, 835.1, 818.0, 759.5, 707.8;
HRMS m/z calc’d for C18H26O2N 288.1958, found 288.1955.
8-(2-(Benzyloxy)ethyl)-3-((diethylamino)methyl)-8-
azabicyclo[3.2.1]octan-3-ol 16 was isolated by purification of
the free-based liquors from isolation of (8-(2-(benzyloxy)-
ethyl)-8-azabicyclo[3.2.1]oct-3-en-3-yl)methanol citrate 17 by
silica gel chromatography on a SNAP KP-NH Biotage cartridge,
eluting with 0−25% TBME in cyclohexane: δH (CDCl3, 400
MHz) 0.99 (6H, t, J = 7.1 Hz), 1.54 (2H, d, J = 13.4 Hz), 1.76
(2H, dd, J = 13.6, 3.0 Hz), 1.83−1.92 (2H, m), 2.17 (2H, dd, J
= 13.6, 6.1 Hz), 2.23 (2H, s), 2.57 (4H, q, J = 7.1 Hz), 2.62
(2H, t, J = 6.3 Hz), 3.20−3.29 (2H, m), 3.60 (2H, t, J = 6.3
Hz), 4.54 (2H, s), 7.24−7.42 (5H, m); δC (CDCl3, 100 MHz)
12.2, 25.7, 43.1, 48.8, 51.6, 59.3, 66.5, 67.9, 69.9, 73.1, 127.5,
127.6, 128.3, 138.5; νmax cm−1 (oil) 3400.7, 2925.3, 2852.9,
1496.4, 1454.4, 1384.5, 1373.0, 1349.01294.5, 1258.1, 1227.8,
1202.5, 1079.4, 1059.8, 1028.6, 976.0, 953.5, 930.2, 887.7,
875.3, 832.2, 787.1, 732.3, 696.4, 661.6; HRMS m/z calc’d for
C21H35 N2O2 347.2693, found 347.2693.
ASSOCIATED CONTENT
* Supporting Information
NMR data for key compounds. This material is available free of
charge via the Internet at http://pubs.acs.org.
■
S
AUTHOR INFORMATION
Corresponding Author
*E-mail: Robert.N.Bream@gmail.com
■
Present Address
†European Medicines Agency, 7 Westferry Circus, Canary
Wharf, London E14 4HB, United Kingdom.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We acknowledge Michael R. Webb for valuable advice in
preparing this manuscript and for supervision of A.J.W. We
thank Andrew Richards for help with the DoE to investigate
factors for the tropinone synthesis.
3-(1-(2-(Benzyloxy)ethyl)-4-methylenepyrrolidin-2-yl)-
propan-1-ol 22. A solution of Dibal-H in THF (1 M, 0.549
mL, 0.549 mmol) was added dropwise to a solution of 8-(2-
(benzyloxy)ethyl)-8-azaspiro[bicyclo[3.2.1]octane-3,2′-oxir-
ane] 13 (100 mg, 0.366 mmol) in THF (1 mL) at −30 °C.
After 30 min, the reaction was quenched with EtOAc (1 mL)
REFERENCES
■
(1) (a) World Health Report; World Health Organization: Geneva,
Switzerland; 2000. See http://www.who.int/whr/2000/en/index.html.
(b) Global Strategy for the Diagnosis, Management, and Prevention of
Chronic Obstructive Pulmonary Disease. From the Global Strategy for
the Diagnosis, Management and Prevention of COPD, Global
649
dx.doi.org/10.1021/op3002933 | Org. Process Res. Dev. 2013, 17, 641−650

Organic Process Research & Development
Article
Initiative for Chronic Obstructive Lung Disease (GOLD), 2013.
Available from: http://www.goldcopd.org.
(2) Lopez, A. D.; Shibuya, K.; Rao, C.; Mathers, C. D.; Hansell, A. L.;
Held, L. S.; Schmid, V.; Buist, S. Eur. Respir. J. 2006, 27, 397.
(3) (a) Liesker, J. J.; Wijkstra, P. J.; Ten Hacken, N. H.; Koeter, G.
̈
H.; Postma, D. S.; Kerstjens, H. A. Chest 2002, 121, 597. (b) Salpeter,
S. R.; Buckley, N. S.; Salpeter, E. E. J. Gen. Intern. Med. 2006, 21, 1011.
(c) Calverley, P. M.; Anderson, J. A.; Celli, B.; Ferguson, G. T.;
Jenkins, C.; Jones, P. W.; Yates, J. C.; Vestbo, J. N. Engl. J. Med. 2007,
356, 775.
(4) Gartlehner, G.; Hansen, R. A.; Carson, S. S.; Lohr, K. N. Ann.
Fam. Med. 2006, 4, 253.
(5) (a) Davis, R.; Untch, K. G. J. Org. Chem. 1981, 46, 2985.
(b) Kanai, T.; Kanagawa, Y.; Ishii, Y. J. Org. Chem. 1990, 55, 3274.
(c) Mizuno, A.; Hamada, Y.; Shioiri, T. Synthesis 1980, 1007. (d) Gang
Chen, G.; Wang, Z.; Wu, J.; Ding, K. Org. Lett. 2008, 10, 4573.
(6) (a) Koreeda, M.; Luengo, J. I. J. Org. Chem. 1984, 49, 2081.
(b) Thomas, J. B.; Brieaddy, L. E.; Boldt, J. G.; Perretta, C.; Carroll, F.
I. Synthesis 2007, 1481.
(7) (a) Lipshutz, B. H.; Keith, J.; Papa, P.; Vivian, R. Tetrahedron Lett.
1998, 39, 4627. (b) Mahoney, W. S; Brestensky, D. M.; Stryker, J. M.
J. Am. Chem. Soc. 1988, 110, 291. (c) Jurkauskas, V.; Sadighi, J. P.;
Buchwald, S. L. Org. Lett. 2003, 5, 2417. (d) Appella, D. H.; Moritani,
Y.; Shintani, R.; Ferreira, E. M.; Buchwald, S. L. J. Am. Chem. Soc.
1999, 121, 9473.
(8) Burks, J. E., Jr.; Espinosa, L.; LaBell, E. S.; McGill, J. M.; Ritter, A.
R.; Speakman, J. L.; Williams, M. Org. Process Res. Dev. 1997, 1, 198.
(9) Synthesis of amine 12: Hu, X. E.; Cassady, J. M. Synth. Commun.
1995, 25, 907.
(10) Factors examined were diacid stoichiometry, 2,5-dimethoxyfur-
an stoichiometry, HCl stoichiometry, NaOAc stoichiometry, water
volume, dialdehyde (from furan hydrolysis) addition temperature, and
heating rate.
(11) (a) Ng, J. S. Synth. Commun. 1990, 20, 1193. (b) Corey, E. J.;
Chaykovsky, M. J. Am. Chem. Soc. 1962, 84, 867. (c) Corey, E. J.;
Chaykovsky, M. J. Am. Chem. Soc. 1965, 87, 1353.
(12) (a) Crandall, J. K.; Chang, L.-H. J. Org. Chem. 1967, 32, 435.
(b) Crandall, J. K.; Chang, L.-H. J. Org. Chem. 1968, 33, 2375.
(c) Mordini, A.; Ben Rayana, E.; Margot, C.; Schlosser, M. Tetrahedron
1990, 46, 2401. (d) Varie, D. L.; Beck, C.; Borders, S. K.; Brady, M.
D.; Cronin, J. S.; Ditsworth, T. K.; Hay, D. A.; Hoard, D. W.; Hoying,
R. C.; Linder, R. J.; Miller, R. D.; Moher, E. D.; Remacle, J. R.; Rieck, J.
A., III; Anderson, D. D.; Dodson, P. N.; Forst, M. B.; Pierson, D. A.;
Turpin, J. A. Org. Process Res. Dev. 2007, 11, 546. (e) Danishefsky, S. J.;
Masters, J. J.; Young, W. B.; Link, J. T.; Snyder, L. B.; Magee, T. V.;
Jung, D. K.; Isaacs, R. C. A.; Bornmann, W. G.; Alaimo, C. A.; Coburn,
C. A.; Di Grandi, M. J. J. Am. Chem. Soc. 1996, 118, 2843. (f) Paquette,
L. A.; Lo, H. Y. J. Org. Chem. 2003, 68, 2282.
(13) (a) Hodgson, D. M.; Bray, C. D.; Kindon, N. D. J. Am. Chem.
Soc. 2004, 126, 6870. (b) Hodgson, D. M.; Fleming, M. J.; Stanway, S.
J. J. Am. Chem. Soc. 2004, 126, 12250. (c) Hodgson, D. M.; Fleming,
M. J.; Stanway, S. J. J. Org. Chem. 2007, 72, 4763.
(14) (a) Brown, H. C.; Yoon, Y. M. J. Am. Chem. Soc. 1968, 90, 2686.
(b) Hutchins, R. O.; Taffer, I. M.; Burgoyne, W. J. Org. Chem. 1981,
46, 5214. (c) Finan, J. M.; Kishi, Y. Tetrahedron Lett. 1982, 23, 2719.
(d) Taber, D. F.; House, J. B. J. Org. Chem. 1994, 59, 4004. (e) Yoon,
N. M.; Kim, K. E. J. Org. Chem. 1987, 52, 5564. (f) Hanessian, S.;
Yang, Y.; Giroux, S.; Mascitti, V.; Ma, J.; Raeppel, F. J. Am. Chem. Soc.
2003, 125, 13784. (g) Dai, L.-X.; Lou, B.-L.; Zhang, Y.-Z.; Guo, G.-Z.
Tetrahedron Lett. 1986, 27, 4343. Chochrek, P.; Kurek-Tyrlik, A.;
Michalak, K.; Wicha, J. Tetrahedron Lett. 2006, 47, 6017.
(15) Doria, G.; Romeo, C.; Forgione, A.; Sberze, P.; Tibolla, M.;
Corno, M. L.; Cadelli, G. Eur. J. Med. Chem. 1981, 16, 367.
(16) Anslyn, E. V.; Dougherty, D. A. Modern Physical Organic
Chemistry; University Science Books: Sausolito, CA. 2006; p 428.
(17) Olivo, H. F.; Colby, D. A.; Hemenway, M. S. J. Org. Chem. 1999,
64, 4966.
650
dx.doi.org/10.1021/op3002933 | Org. Process Res. Dev. 2013, 17, 641−650