Development of a suitable process for the preparation of a TNF-α converting enzyme inhibitor, WAY-281418

Article abstract of DOI:10.1021/op800090s

A suitable process for the preparation of kilogram quantities of a TNF-α converting enzyme (TACE) inhibitor (WAY-281418) was developed using isatin 13 as starting material and an efficient coupling step for the formation of sulfonamide 8 in a 15% overall yield. Process preparation of (+)-(1S,2R)-2-aminocyclopentane-1-carboxylic acid (7, (+)-cispentacin), a chiral component for WAY-281418, was successfully scaled up via an asymmetric hydroge-nation reaction. Crystallization allowed the isolation of all intermediates and the final product 9.

Full text of DOI:10.1021/op800090s

Organic Process Research & Development 2008, 12, 1253–1260
Development of a Suitable Process for the Preparation of a TNF-r Converting
Enzyme Inhibitor, WAY-281418
Youchu Wang,*,† Maria Papamichelakis, Warren Chew, John Sellstedt, Razzak Noureldin, Sam Tadayon, and
Sylvain Daigneault
Chemical DeVelopment, Wyeth Research, 1025 BouleVard Marcel-Laurin, Saint-Laurent, Quebec H4R 1J6, Canada
Rocco J. Galante and Jerry Sun
Chemical and Screening Sciences, Wyeth Research, 401 North Middletown Road, Pearl RiVer, New York 10965-1299, U.S.A.
Scheme 1. MedChem synthesis of 9
Abstract:
A suitable process for the preparation of kilogram quantities of a
TNF-r converting enzyme (TACE) inhibitor (WAY-281418) was
developed using isatin 13 as starting material and an efficient
coupling step for the formation of sulfonamide 8 in a 15% overall
yield. Process preparation of (+)-(1S,2R)-2-aminocyclopentane-1-
carboxylic acid (7, (+)-cispentacin), a chiral component for WAY-
281418, was successfully scaled up via an asymmetric hydroge-
nation reaction. Crystallization allowed the isolation of all
intermediates and the final product 9.
Introduction
TNF-R converting enzyme (TACE)1 is a membrane-bound
metalloproteinase disintegrin that processes the membrane-
associated cytokine pro-TNF-R to a biologically active soluble
form. Overproduction of the TNF-R soluble form is associated
with symptoms of a variety of infectious, autoimmune, and
inflammatory disorders including rheumatoid arthritis (RA) and
Crohn’s disease.2 Inhibition of TACE to block TNF-R over-
production is currently an important therapeutic target.3
In screening TACE inhibitors at Wyeth, WAY-281418 (9)4
was identified as a potent lead candidate with an IC50 of 0.6
nM and excellent TACE selectivity. We describe our efforts to
develop a scalable process towards the large-scale preparation
of 9 for toxicological and clinical studies.
Results and Discussion
The original MedChem route is outlined in Scheme 1. Due
to time constraints, the synthetic strategy for the preparation of
9 was not substantially changed, but the critical scale-up and
safety issues were addressed. Because the MedChem synthesis
of 7, a key chiral component for 9, was not suitable for scale-
up (Scheme 2), an asymmetric hydrogenation process was
developed instead.5
Development Synthesis of Sulfonylchloride Hydrochlo-
ride Salt 6. Starting from quinaldine (1) in the MedChem route
(Scheme 1), hydroxymethylation gave alcohol 2 in 49% yield.
This FeSO4/Fe-mediated radical reaction6 resulted in a deep-
* To whom correspondence should be addressed. E-mail: wang_youchu@
pharmatechs.com.
† Present address: Process R&D, WuXi Pharmatech Co., Ltd., 288 Fute Zhong
Rd., Waigaoqiao Free Trade Zone, Shanghai 200131, China.
(1) Black, R. A.; Rauch, C. T.; Kozlosky, C. J.; Peschon, J. J.; Slack,
J. L.; Wolfson, M. F.; Castner, B. J.; Stocking, K. L.; Reddy, P.;
Srinivasan, S.; Nelson, N.; Boiani, N.; Schooley, K. A.; Gerhart, M.;
Davis, R.; Fitzner, J. N.; Johnson, R. S.; Paxton, R. J.; March, C. J.;
Cerretti, D. P. Nature 1997, 385, 729–733.
(2) Eigler, A.; Sinha, B.; Hartmann, G.; Endres, S. Immunol.Today 1997,
18, 487–492.
(3) (a) Nelson, F. C.; Zask, A. Expert Opin. InVest. Drugs 1999, 8, 383–
392. (b) Levin, J. Curr. Top. Med. Chem. 2004, 4, 1289–1310.
(4) Levin, J.; Li, Z.; Diamantidis, G.; Lovering, F. E.; Wang, W.; Condon,
J. S.; Lin, Y.; Skotnicki, J. S.; Park, K. ꢀ-Sulfonamide Hydroxamic
Acid Inhibitors of TACE/Matrix Metalloproteinase. U.S. Patent Appl.
Pub. 0211730, 2006.
(5) (a) Tang, W.; Wu, S.; Zhang, X. J. Am. Soc. Chem. 2003, 125, 9570–
9571. (b) Zhang, X.; Tang, W. Preparation of Chiral Cyclic Amino
Acids and DeriVatiVes. U.S. Patent Appl. Pub. 0242889, 2004.
(6) Citterio, A.; Gentile, A.; Minisci, F.; Serravalle, M.; Ventura, S.
Tetrahedron 1985, 41, 617–620.
10.1021/op800090s CCC: $40.75
Published on Web 09/09/2008
2008 American Chemical Society
Vol. 12, No. 6, 2008 / Organic Process Research & Development
•
1253

Scheme 2. MedChem synthesis of 7
Scheme 3. Development synthesis of 9
reaction)8 afforded 2-methylquinoline-4-carboxylic acid (14),
which was then reduced to 2 with BH3/THF9 (generated in situ
from BF3/Et2O and NaBH4 in THF). The low yield for the BH3
reduction step was due to the solubility of 2 in water and
significant reduction of the quinoline ring under these isolation
conditions. The in situ generation of BH3/THF was found to
be more reliable and safer to control at scale relative to
purchased material. The yield from this two-step process was
less than that of the MedChem route; however, the requisite
column chromatography and tedious workup were eliminated,
the cost of goods was reduced as isatin 13 was cheaper than
quinaldine (1), and throughputs were substantially increased.
These attributes ameliorated the low yield.
Conversion of 2 to chloromethylquinoline hydrochloride salt
3 went smoothly. A suspension of 2 in toluene at room
temperature (20-25 °C) was treated with SOCl2 (2.0 equiv)
for 2 h to afford 3 after filtration of the slurry. A higher reaction
temperature (50 °C) was required if the reaction was not
complete (<1.0% of starting material). Toluene replaced the
solvent mixture of chloroform and dichloromethane used in the
MedChem route.
Figure 1. Throughput comparison between MedChem and
development syntheses.
black solution with significant tar formation, which adversely
affected the phase separation. Moreover, a tedious workup using
large volumes of CHCl3 followed by column chromatography
was required to isolate 2. Both reaction and workup throughputs7
were low (<1.0%) for which further development was war-
ranted (Figure 1). In order to provide a sufficient quantity of 2
for the downstream process development, this reaction was not
suitable.
A two-step transformation for the preparation of 2 was
designed and successfully run on a 4-kg scale in 29% yield
(two steps) (Scheme 3). No further process optimization was
done. Treating isatin 13 with KOH/water at 50 °C followed by
the addition of acetone and pH adjustment (a Pfitzinger
In the alkylation reaction of sodium 4-hydroxybenzene-
sulfonate dihydrate, 4, with 3, the MedChem group had mixed
all the reactants and reagents together in ethanol and 3.0 M
NaOH (2.2 equiv) and heated the mixture to reflux for 20 h.
After workup, 5 was obtained as its sodium salt but was
contaminated with ∼19% unreacted 4, ∼20% of NaCl, and
(8) Zemtsova, M. N.; Trakhtenberg, P. L.; Galkina, M. V. Russ. J. Org.
Chem. 2003, 39, 1803–1803.
(7) Throughput for reaction (R)/workup (W) is defined as kg of product
from 100 L of reaction/workup mixture at maximum volume expressed
as a percentage.
(9) Burkhardt, E. R.; Matos, K. Chem. ReV 2006, 106, 2617–2650, Use
of other boron-reducing reagents was not investigated, and no further
optimization studies were performed for this step.
1254
•
Vol. 12, No. 6, 2008 / Organic Process Research & Development

and 1.5 equiv of DMF. Further development work on MeCN
or THF would be needed to optimize the reaction conditions.
Due to gas (SO2) evolution during the reaction, controlled
addition of SOCl2 was required to allow the reaction to run
more safely. After SOCl2 addition, a sample of the reaction
mixture showed an exotherm from between 60 to 110 °C from
the thermal screening unit followed by an endotherm and minor
exothermic decomposition at 135 °C.
Development Synthesis of (+)-(1S,2R)-2-Aminocyclopen-
tane-1-carboxylic Acid (7). (+)-(1S,2R)-2-Aminocyclopentane-
1-carboxylic acid (cispentacin) HCl salt (7) was not readily
available in the quantities and timelines that we needed. The
MedChem group prepared 7 on gram scale via an enzyme-
catalyzed enantioselective ring-opening of unactivated alicyclic
ꢀ-lactam12 in less than 20% overall yield and 98% ee after
repeated recrystallizations (Scheme 2). However, this approach
was not scalable considering the large loading of lipolase, the
long reaction time required, and the low yield obtained. Other
alternative approaches13,14 were evaluated, but these routes
required either low temperatures (-95 °C) or column chroma-
tography purifications, which were not amenable for scale-up
to produce kilo quantities.
A route initially explored by our group was a kinetic
chemical resolution of (()-cis-2-benzyloxycarbonylaminocy-
clopentane carboxylic acid using commercially available de-
hydroabiethylamine (DAA).15 The DAA salt was obtained by
filtration in 83% yield, but resolution was not successful even
after repeated recrystallizations.
Chiral hydrogenation technology (CHT) is widely used in
the pharmaceutical industry to access chiral building blocks and
chiral molecules on industrial scale.16 Our CHT group within
Chemical Development initiated a chiral hydrogenation screen
based on Zhang’s methodology,5 and we developed a suitable
process for scale-up.
The key building block 18 for asymmetric hydrogenation
was prepared by telescoping the amination and N-acylation
reactions starting from inexpensive and commercially available
ethyl-2-oxocyclopentane carboxylate (16) (Scheme 4). The
amount of NH4OAc was reduced from 5.0 to 2.0 equiv for the
amination step, and the nonenvironmentally friendly dichlo-
romethane workup was replaced with toluene. Nonisolated 17
was directly acylated with acetic anhydride (1.5 equiv) to afford
18. The number of equivalents of Ac2O was reduced 4-fold,
and the reaction was run in the absence of base. The compound
18 was isolated by swapping toluene with water. A comparison
of the throughputs between Zhang’s method and our method
showed a significant increase in the reaction and workup
Figure 2. Solvent screen for the preparation of 6.
other unidentified minor impurities. Although NaCl can be
removed in subsequent steps, excess 4 led to undesirable
impurities.10 To drive the reaction to consume 4 and prevent
side reactions in the later steps, we modified the process such
that reactants and reagents were charged in sequence. 4 (1.1
equiv) was combined with aq NaOH (2.0 equiv) to generate
the phenoxide and was heated to 55-60 °C. To the resulting
mixture was added 3 (1.0 equiv) in one portion. The reaction
was stirred for 24 h in the presence of substoichiometric amount
of NaI (0.5 equiv) and deemed complete while 3 was <1.0%
by HPLC. The reaction was quenched with concentrated aq
HCl at 55-60 °C, adjusting the pH to ∼3. The sulfonic acid
15 was isolated by filtration free from NaCl. The slight excess
of 4 remaining was easily removed into the mother liquors. It
was necessary to quench the reaction at higher temperatures to
increase the filtration rate of the product. Acidification at room
temperature (20-25 °C) produced 15 as fine crystals that caused
a slow filtration.
In the MedChem route, 5 was treated with SOCl2 (7 equiv)
in THF at room temperature (21 °C) to afford sulfonyl chloride
hydrochloride salt 6 with 94% conversion in 86% yield.
However, due to contamination with NaCl and 4-hydroxyben-
zenesulfonyl chloride derived from 4, the actual yield for 6 was
<50%. Interestingly, when the sulfonic acid 15 was treated
under the same reaction conditions, almost no conversion was
observed by HPLC, possibly because HCl was a byproduct that
retarded the reaction. When the sodium salt 5 was used, the
byproduct NaCl precipitated out to drive the reaction. Through
a design of experiments (DoE) approach, screening of solvents
and the quantities of SOCl2 and DMF by our group, Integrated
Parallel Automation and Chemistry Technologies (InPACT),
showed the optimized conditions were 2.54 equiv of SOCl2 and
1.0 equiv of DMF in 1,2-dichloroethane at 75 °C. We were
able to conduct this transformation on a 1.67-kg scale giving
98% conversion of 15 to 6 but contaminated with ∼3.9% of
starting material as the major impurity.11 Although the reaction
in 1,2-dichloroethane performed the best in terms of conversion
(Figure 2), the hazardous nature of this solvent necessitated an
alternative. Acetonitrile and THF were chosen for further
screening in order to replace 1,2-dichloroethane. Results
revealed that MeCN could replace 1,2-dichloroethane by
increasing the amount of SOCl2 to 3.0 equiv and DMF to 1.5
equiv and that THF may be used but with 5.0 equiv. of SOCl2
(12) Forro´, E.; Fu¨lo¨p, F. Org. Lett. 2003, 5, 1209–1212.
(13) Cimarelli, C.; Palmieri, G.; Volpini, E. Synth. Commun. 2001, 31,
2943–2953.
(14) Davies, S. G.; Ichihara, O.; Lenoir, I.; Walters, I. A. S. J. Chem. Soc.,
Perkin Trans. 1 1994, 1411–1415.
(15) Konish, M.; Nishio, M.; Saitoh, K.; Miyaki, T.; Oki, T.; Kawaguchi,
H. J. Antibiot. 1989, 42, 1749–1755.
(16) Blaser, H. U.; Schmidt, E. Asymmetric Catalysis on Industrial Scale,
Challenges, Approaches and Solutions; Wiley-VCH: Weinheim, 2004.
(17) HOBt is known to be potentially explosive in the anhydrous form.
Use of the hydrated form is recommended.
(18) Transportation and storage of aqueous hydroxylamine are risky. For
details, see: am Ende, D.; Vogt, P. Org. Process Res. DeV. 2003, 7,
1029–1033.
(10) Campbell, R. W., Jr Org. Chem. 1973, 38, 1047–1047.
(11) The higher than expected level of this impurity in the product was
due to hydrolysis of the product during handling and workup.
(19) ICH Harmonised Tripartite Guideline for residual solvents Q3C(M).
Vol. 12, No. 6, 2008 / Organic Process Research & Development
•
1255

Scheme 4. Development synthesis of 7
Figure 4. Residual solvents in 18 vs reaction conversion for
19.
up to 3% (w/w, by KF) water present, a >99% conversion was
achieved. Similarly, acetic acid could be tolerated as high as
3.0% (w/w, by GC). Toluene was found to be more detrimental
to the reaction conversion. The amount of residual toluene had
to be controlled to less than 0.2% (w/w, by GC) to enable >99%
conversion. Residual solvents present in 18 were found to have
no effect on ee. Through these process development findings,
the reaction ran smoothly several times on a 600-g scale to
obtain 19 as a concentrated solution in methanol in quantitative
yield (by HPLC assay) and >99% ee. This concentrate was
telescoped into the hydrolysis step.
The methanolic solution of 19 was hydrolysed with 12.0
equiv of HCl in MeOH/water. The workup required the
concentration of the reaction mixture to <5% methanol to
ensure the efficient removal of the catalyst when the aqueous
concentrate was filtered. The aqueous filtrate was then concen-
trated to replace water and excess HCl with acetonitrile, and 7
was isolated by filtration as a white crystalline solid in 85%
yield and >99% ee. Residual ruthenium was found to be <11
ppm.
Coupling of 6 with 7. In the MedChem route, sulfonamide
8 was obtained by combining 6 with 7 in the presence of
N-methylmorpholine (NMM) in a mixture of dioxane and water
followed by extractive workup. However, the product was
isolated with up to 20% of 15 as the major impurity, which
was produced via hydrolysis of 6 during the reaction. In the
absence of water, both 6 and 7 had low solubility in dioxane
as well as other polar solvents, such as acetonitrile and DMF.
To circumvent issues associated with low solubility of 7 and 8
in a nonaqueous solvent system and to minimize the sulfonic
acid 15, we pretreated 7 with N,O-bis(trimethylsilyl)trifluoro-
acetamide (BSTFA) in the presence of NMM, which provided
a very soluble N,O-bis-silylated intermediate 20 (Scheme 5).
Figure 3. Throughput comparison between Zhang’s and
development syntheses.
throughputs (Figure 3). We were able to successfully produce
kilogram quantities with high purity using this process.
Asymmetric hydrogenation of 18 using a modified Zhang’s
procedure successfully afforded 19 in quantitative yield and
>99% ee on multigram to 100-g scales. These conditions were
nonoptimized but were sufficient for producing enough of the
chiral building block to manufacture the preclinical development
API supply. The amount of the catalyst used was a cost concern.
Significant cost reduction was achieved by reducing the amount
of expensive Ru(η2-COD)(η3-methylallyl)2 catalyst from 5 to
2 mol % and lowering the (S)-BINAP ligand from 5 to 2 mol
% and HBF4 ·Me2O from 10 to 4 mol %. Further reductions
were possible but not explored. The volume of methanol used
at small scale was reduced from 300 to 8 mL per gram of 18.
Consequently, throughputs were dramatically increased for this
step. Preparation of the “active catalyst” [adding HBF4 ·Me2O
to a mixture of Ru(η2-COD)(η3-methylallyl)2 and (S)-(-)-
BINAP] was performed in methanol, forming a heterogeneous
mixture which was used as is.
One key factor that ensured the success of the asymmetric
hydrogenation was conducting the reduction under oxygen-free
conditions, preferably using a high quality of argon (purity
>99.998%). Experiments in which nitrogen was used failed,
possibly due to a higher content of oxygen and moisture which
deteriorated the catalyst and caused the reaction to stall, yielding
a low ee of 19.
Scheme 5. Development synthesis of 8
The effect of residual solvents in 18 on reaction conversion
was investigated (Figure 4). Water had a minor impact. With
1256
•
Vol. 12, No. 6, 2008 / Organic Process Research & Development

Scheme 6. Possible formation pathway to 24
Figure 5. 1R,2R Diastereomer.
This intermediate was coupled with 6 in the presence of
NMM to afford the desired sulfonamide 8 after removal of the
TMS protecting group during aqueous workup. N,O-bis(trim-
ethylsilyl)acetamide (BSA), which was less expensive than
BSTFA, could also be used for this reaction, but TMSCl did
not work. With these process improvements, the major impurity
15 in the product was controlled to less than 2%. LC-MS
analysis further identified the 1R,2R diastereomer 21 at up to
1.5% (Figure 5). However, this impurity was effectively
removed in the next step.
Hydroxyamidation of 8 in the MedChem route using
1-hydroxybenzotriazole hydrate (HOBt,17 3.0 equiv), 1-[3-
(dimethylamino)propyl]-3-ethylcarbodiimide hydrochloride
(EDC, 3.0 equiv) and 50% aq. NH2OH (8.0 equiv) in DMF at
0∼5 °C for 18 h led to 9 in 43% yield and 85% purity. Low
yield and purity were found that was associated with excess
HOBt and EDC and this created a tedious workup. Minor
process changes (amount of reagents) made from the MedChem
route for this step increased the API yield. We did not explore
other coupling reagents. However, attempts to convert 8 to its
acid chloride by using (COCl)2/DMF/THF followed by quench-
ing the reaction mixture with aq NH2OH led to an unstirrable
mixture and an incomplete reaction (>20% of starting material
by HPLC). Reaction of 8 with the lesser amounts of HOBt (1.2
equiv) and EDC (1.2 equiv) worked well in DMF at 20-25
°C followed by quenching the reaction mixture in aq NH2OH,
which was generated in situ from solid hydroxylamine hydro-
chloride salt and Et3N in water.18 The resulting slurry was
filtered after dilution with water to provide 9 as crude API.
Minor impurities identified in the API included the 1R,2R
diastereomer 22 (Figure 6) and the pyrazolone 24 which was
only identified by LC-MS. A proposed pathway of formation
of 24 is activation of the hydroxyl group in 9 by EDC, rendering
it as a leaving group followed by nucleophilic attack by the
nitrogen of the sulfonamide group in the presence of Et3N
(Scheme 6).
(720 ppm) in the final API.19 The most stable polymorph and
the desired purity profile was obtained in acetone/water (10:1).
High residual acetone in the API was removed by a hot water
reslurry, albeit there was no impurity rejection and the purity
profile in 9 did not change. The quality of the API produced
was adequate to conduct preclinical studies.
Conclusion
In conclusion, we developed a suitable process for the
preparation of WAY-281418 (9), which consisted of seven
reaction steps and one crystallization step. The process was used
for the preparation of multikilograms of API for both early-
stage toxicological and clinical studies. The reaction and workup
throughputs for every step were significantly increased. Scale-
up issues for the synthesis of 2 were addressed using the
inexpensive isatin 13 as starting material. The coupling reaction
to prepare 8 was improved effectively. An asymmetric hydro-
genation approach for the scale-up of 7 was developed for the
successful preparation of API.
Experimental Section
General. NMR spectra were recorded at 300 MHz (1H and
13C) in DMSO-d6. HPLC analyses were determined on a Waters
Alliance 2690 HPLC instrument equipped with a PDA (model
2996) detector, and the results are described as area % (AP).
HPLC conditions for chiral purity: Column: CHIRALPAK
AD-H 250 mm × 4.6 mm, no. ADHOCE FC033. Mobile
phase: heptanes/ethanol ) 150:850 mL. Flow rate: 1.0 mL/
min. UV detector: 235 nm. Inj vol: 5 µL. Instrumentation:
Detector - WATERS PDA 996. Pump - Alliance 2690 System
G. Isocratic - time from 0 to 30 min. Residual solvent levels
were measured on a Perkin-Elmer Autosystem XL serial no.
610N0092213 GC with FID and using the following conditions:
Gamma-Dex-225, 30 m, 0.25 mm i.d. 0.25 µm at 150 °C for
40 min, inlet 200 °C, detector 220 °C. Flow rate: helium 1.0
mL/min, split 150 mL/min. Chloride content analyses were
Figure 6. 1R,2R API diastereomer 22.
Recrystallization and purification of crude API 9 was
achieved from an acetone/water solvent system after an
extensive solvent screening. Purification was also effective from
ethanol/water or THF/water combinations. However, in ethanol/
water, the undesirable ethanol solvate polymorph was formed
and using THF/water in the last step of the synthesis was not
the preferred option due to the low specification limit of THF
Vol. 12, No. 6, 2008 / Organic Process Research & Development
•
1257

determined using a column IONPAC AG17 50 mm × 4 mm
no. 001596 @ 30 °C, Dionex ED50 as detector and NaCl as
reference. Residual Ru was measured using a Perkin-Elmer
Optima 4200DV ICP. Methanol used in the preparation of 7
was bubbled (degassed) with argon prior to use.
nitrogen inlet were charged NaOH (0.471 kg, 11.78 mol),
potable water (6.56 kg), and sodium 4-hydroxybenzenesulfonate
dihydrate (4, 1.437 kg, 6.19 mol). The mixture was stirred for
10 min to obtain a solution. To the solution was then charged
ethanol (4.62 kg), NaI (76.9 g, 0.513 mol), and 4-chloromethyl-
2-methylquinoline hydrochloride salt (3, 1.30 kg, 5.70 mol).
The reaction mixture was heated to 55-60 °C over 15 min
and monitored by HPLC. After being stirred for 24 h, the
reaction mixture was quenched by adding concentrated HCl
(0.586 kg) over 30 min at 55-60 °C and adjusted pH to 3-3.3.
The resulting thick slurry was stirred for 30 min, cooled to 20
to 25 °C, and filtered, and the cake was washed with water (1
L × 3) and ethanol (1 L × 3) to give 4-(2-methylquinolin-4-
ylmethoxy)benzenesulfonic acid (15, 1.824 kg, 97% yield, 99
AP) as an off-white solid after drying in an oven under vacuum
at 70 °C for 24 h. Mp 227.2 °C. 1H NMR δ 8.44 (d, J ) 7.8
Hz, 1H), 8.17 (d, J ) 7.8 Hz, 1H), 8.07-8.12 (m, 1H), 8.03
(s, 1H), 7.90-7.94 (m, 1H), 7.58 (d, J ) 6.6 Hz, 2H), 7.14 (d,
J ) 6.6 Hz, 2H), 5.90 (s, 2H), 2.94 (s, 3H). MS: 330 [M +
1]+.
(2-Methylquinolin-4-yl)methanol (2). To a 50-L jacketed
cylindrical reactor equipped with an impeller-style agitator,
condenser, thermocouple, and nitrogen inlet were charged isatin
(13, 4.00 kg, 27.19 mol), 85% aqueous potassium hydroxide
(12.43 kg, 188.27 mol), and potable water (24.9 kg). The
reaction mixture was heated to 50 °C for 40 min. Acetone (3.16
kg, 54.37 mol) was then added over a minimum of 4 h,
maintaining the temperature at 50 °C. After being stirred for
1 h, the reaction mixture was cooled to room temperature and
acidified to pH 7 by addition of concentrated HCl (17.04 kg,
170.4 mol). The resulting KCl was filtered, and the pH of the
mother liquor was further adjusted to 3 by adding concentrated
HCl (2.52 kg) to obtain a heavy slurry which was filtered and
dried at 75 °C under vacuum to afford 2-methylquinoline-4-
carboxylic acid (14, 4.386 kg, 86% yield, 99.6 AP).
To a 30-L jacketed cylindrical reactor equipped with an
impeller-style agitator, condenser, thermocouple, and nitrogen
inlet were charged 14 (3.3 kg, 17.63 mol), THF (21.3 kg), and
boron trifluoride etherate (7.5 kg, 52.89 mol). The resulting
mixture was stirred at 20-25 °C for 15 h. To the mixture was
added NaBH4 (1.0 kg, 26.42 mol) portionwise over a minimum
of 2 h, maintaining the temperature in the range of 20 to 35
°C. After being stirred for 1 h at 32-38 °C, the reaction mixture
was added to aqueous HCl (2.5 M, 17.8 L) over a minimum of
30 min, maintaining the temperature in the range (5-30 °C).
Removal of THF by atmospheric distillation and pH adjustment
of the aqueous mixture to 10 by addition of ammonium
hydroxide (28%, v/v, 6.5 kg) over a minimum of 1 h,
maintaining the temperature in the range of 20-30 °C gave
rise to a precipitate, which was collected by filtration, washed
with toluene (3.03 kg) and water (3.5 kg), and dried at 70 °C
under vacuum for 15 h to obtain (2-methylquinolin-4-yl)metha-
4-(2-Methylquinolin-4-ylmethoxy)benzenesulfonyl Chlo-
ride Hydrochloride Salt (6). To a 3-L multinecked round-
bottom flask quipped with mechanical stirrer, condenser,
thermocouple, and nitrogen inlet were charged 15 (1.67 kg, 5.06
mol), 1,2-dichloroethane (20.7 kg), and DMF (369 g, 5.06 mol).
The mixture was heated to 70-80 °C, and thionyl chloride (1.53
kg, 27.9 mol) was added over 3 h. The reaction was stirred for
1 h, and the mixture was cooled to 20-25 °C. The suspension
was filtered and washed with 1,2-dichloroethane (3 × 2.0 kg)
to afford 4-(2-methylquinolin-4-ylmethoxy)benzenesulfonyl
chloride hydrochloride salt (6, 1.79 kg, 93% yield, 95 AP) as
a yellow solid after drying in an oven under vacuum at 50 °C
for 12 h. 1H NMR δ 8.50 (m, 2H), 8.15 (m, 1H), 8.09 (s, 1H),
7.96 (m, 1H), 7.60 (d, J ) 8.4 Hz, 2H), 7.18 (d, J ) 8.4 Hz,
2H), 5.94 (s, 3H). MS: 348 (M + 1)+. Cl- content: found
18.96% (calcd 18.49%).
Ethyl-2-acetamido-1-cyclopentene-1-carboxylate (18). To
a 2-L jacketed reactor quipped with mechanical stirrer, con-
denser, thermocouple, and nitrogen inlet were charged 2-oxo-
cylopentanecarboxylic aicd ethyl ester (16, 200 g, 1.28 mol),
ammonium acetate (198 g, 2.56 mol), and methanol (700 mL).
The mixture was heated to 65 °C and stirred for a minimum of
2 h until the reaction was complete (<0.5% of starting material
16 by HPLC). The reaction mixture was cooled and concen-
trated, and residual methanol was chased with toluene (400 mL).
The resulting toluene solution (∼800 mL) was washed with
water (500 mL × 2) and concentrated. To the concentrated
toluene solution (∼400 mL) was added acetic anhydride (196
g, 1.92 mol) and the mixture heated to 75 °C and stirred for a
minimum of 2 h until the reaction was complete (<0.5% of 17
by HPLC). Water (400 mL) was added to the reaction after
cooling, the mixture was stirred for a minimum of 1 h, and the
resulting layers were separated. The organic layer was washed
with water (400 mL), saturated sodium bicarbonate (400 mL),
and water (400 mL). The organic layer was concentrated, and
toluene was replaced with water (300 mL). The resulting slurry
was filtered and washed with water (300 mL × 3) to afford
ethyl-2-acetamido-1-cyclopentene-1-carboxylate (18, 191 g,
1
nol (2, 1.055 kg, 34% yield, 97 AP). H NMR δ 7.91-8.00
(m, 2H), 7.66-7.71 (m, 1H), 7.49-7.55 (m, 1H), 7.46 (s, 1H),
5.51 (t, J ) 5.7 Hz, OH, 1H), 5.00 (d, J ) 5.7 Hz, 2H), 3.31
(s, 3H). MS: 174 [M + 1]+.
4-Chloromethyl-2-Methylquinoline HCl Salt (3). To a
30-L jacketed cylindrical reactor (internal diameter 300 mm)
equipped with a 250 mm diameter impeller-style agitator,
condenser, thermocouple, baffle, and nitrogen inlet were charged
(2-methylquinolin-4-yl)methanol (2, 1.1 kg, 6.35 mol) and
toluene (9.52 kg). Thionyl chloride (1.51 kg, 12.7 mol) was
then slowly added over 1 h while maintaining temperature
below 25 °C. After being stirred for 2 h, the reaction was filtered
and washed with toluene (500 mL × 2) to provide 4-chlorom-
ethyl-2-methylquinoline as the HCl salt (3, 1.40 kg, 97% yield,
94 AP) after drying in oven under vacuum at 75 °C for 15 h.
1H NMR δ 8.40-8.43 (m, 2H), 8.05-8.11 (m, 1H), 8.01 (s,
1H), 7.89-7.94 (m, 1H), 5.45 (s, 2H), 2.96 (s, 3H). MS: 229
[M + 1]+. Cl- content: found 15.8% (calcd 15.6%).
4-(2-Methylquinolin-4-ylmethoxy)benzenesulfonic Acid
(15). To a 20-L jacketed cylindrical reactor equipped with an
impeller-style agitator, condenser, thermocouple, baffle, and
1258
•
Vol. 12, No. 6, 2008 / Organic Process Research & Development

76% yield, 99.7 AP) after drying at 35-40 °C, 10 mmHg, for
a minimum of 3 h until KF specification (0.5%) is met.
(+)-(1S,2R)-2-Aminocyclopentane-1-carboxylic Acid (7).
To a 5-L jacketed reactor equipped with mechanical stirrer,
condenser, thermocouple, and argon inlet were charged Ru(η2-
COD)(η3-methylallyl)2 (19.56 g, 0.0612 mol), (S)-(-)-BINAP
(38.16 g, 0.0612 mol), and degassed (argon) methanol (3.0 L)
under an inert argon atmosphere. The mixture was cooled to
0-5 °C, and to the mixture was added dropwise tetrafluoroboric
acid dimethyl ether complex (17.16 g, 15.6 mL, 0.128 mol)
over a minimum of 10 min, keeping the batch temperature
between 0 and 10 °C. The reaction mixture was stirred for 15
min at 0-10 °C and for 1 h at 20-25 °C. The resulting orange
mixture was transferred to a 2-gal hydrogenation pressure vessel
containing ethyl-2-acetamido-1-cyclopentene-1-carboxylate (18,
0.60 kg, 3.04 mol) and degassed methanol (1.8 L) under an
inert atmosphere. The mixture was hydrogenated at 75 psi, 50
°C, and at a stirring rate of 350 rpm for a minimum of 3 h. The
reaction was complete when <0.5% (GC analysis at RT: 19.5
min) of the starting material remained. The reaction mixture
was filtered through a 0.2-µm cartridge, and the cartridge was
washed with degassed methanol (2.0 L). The filtrate was
concentrated under vacuum to a volume of 2.49 L containing
19. To the concentrate containing 19 was added water (795
mL) followed by adding concentrated HCl (3.0 L) over a
minimum of 0.5 h, maintaining the batch temperature at 20-25
°C, and the mixture was heated to 80-85 °C over a minimum
of 1 h and stirred for a minimum of 24 h. The reaction was
complete when <0.5% (HPLC analysis) of starting material
19 remained. The reaction mixture was concentrated to a
volume of 3.50 L, which was filtered through a Celite pad, and
the Celite pad was washed with water (500 mL × 3). The filtrate
was concentrated under vacuum to a volume of 1.29 L. To
the concentrate was added acetonitrile (1.50 L × 3), and the
mixture was concentrated again to 1.29 L three times. To the
concentrate was added acetonitrile (1.5 L), and the resulting
slurry was stirred at 0-5 °C for 2 h. The slurry was filtered,
and the cake was washed with cold acetonitrile (0.51 L × 2)
and dried at 50 °C in a vacuum oven (10 mmHg) for a
minimum of 24 h to yield (+)-(1S,2R)-2-aminocyclopentane-
1-carboxylic acid (7, 0.423 kg, 85% yield) with 99.5% ee by
GC at RT 24.6 min. 1H NMR δ 12.8 (brs, 1H), 8.04 (brs, 3H),
3.60 (brs, 1H), 2.97 (ddd, J ) 8.4, 6.6, 6.3 Hz, 1H), 1.85-2.01
(m, 3H), 1.70-1.84 (m, 2H), 1.53-1.65 (m, 1H).
overnight (>12 h). To the reaction mixture was added water (7.16
kg) over a minimum of 30 min. The resulting slurry was stirred
for a minimum of 2 h, filtered, and washed with 1,2-dichloroethane
(2.33 kg, 1.85 L) followed by water (1.85 kg, 1.85 L) to provide
8 (1.74 kg, 85% yield, 98.2 AP). 1H NMR δ 12.1 (brs, 1H), 8.16
(m, 2H), 7.87 (m, 1H), 7.82 (d, J ) 9 Hz, 2H), 7.01 (s, 1H), 7.50
(m, 1H), 7.34 (d, J ) 9 Hz, 2H), 5.79 (s, 2H), 3.75 (m, 1H), 2.77
(s, 3H), 1.89-1.56 (m, 4H), 1.49-1.25 (m, 3H). MS: 441 [M +
1]+.
(1S,2R)-2-[4-(2-Methylquinolin-4-ylmethoxy)benzenesulfo-
nylamino]cyclopentanecarboxylic Acid Hydroxyamide: Crude
WAY-281418 (9). To a 5-L jacketed cylindrical reactor (internal
diameter 300 mm) equipped with a 250 mm diameter impeller-
style agitator, condenser, thermocouple, baffle, and nitrogen inlet
were charged 8 (0.350 kg, 0.79 mol), HOBt (0.128 kg, 0.950
mol), and DMF (0.991 kg, 1.05 L). To the mixture was added
a slurry of EDC (0.182 kg, 0.950 mol) in DMF (0.826 kg, 0.875
L) over a minimum of 30 min with a rate of addition of 15 to
20 mL/min, maintaining the temperature in the range of 20-25
°C. After being stirred for 2 h, the reaction mixture was
quenched into an aqueous solution of hydroxylamine generated
from hydroxylamine HCl (0.681 kg, 9.80 mol, 12.4 equiv) in
water (1.40 kg) upon addition of Et3N (0.992 kg, 9.80 mol)
while maintaining the temperature of the resulting mixture at
-1 to 2 °C. This was warmed to 20-25 °C and stirred for 1 h.
Water (5.25 kg) was added to the mixture over a minimum of
1 h to give a suspension. Filtration, washing with water (3 ×
0.350 kg), and drying at 60 °C, 10 mmHg, for 24 h provided
crude API (9, 0.33 kg, 91.7% yield, 97.8 AP).
Crystallization of (1S,2R)-2-[4-(2-Methylquinolin-4-yl-
methoxy)-benzenesulfonylamino]cyclopentanecarboxylic Acid
Hydroxyamide, WAY-281418 (9). To a 70-L jacketed reactor
equipped with an impeller-style agitator, condenser, thermo-
couple, baffle, and nitrogen inlet were charged the crude 9 (1.24
kg), acetone (19.6 kg) and purified water (USP, 2.48 kg). The
mixture was heated to reflux (59.1 °C) for 10 min, clarified at
50 °C through a 0.2 µm cartridge, and heated to reflux again
for 10 min. The mixture was then cooled to 50 °C to generate
crystals without seeding, held for 1 h, then further cooled to 20
°C over a minimum of 2 h, and held for 1 h again. The crystals
were collected by filtration, washed with acetone (3 × 1.0 kg),
and dried at 60 °C, 10 mmHg for 12 h to afford the purified
API 9 (0.953 kg, 76.8% yield, 99.2 AP, >99.9% ee). The
material was resuspended in purified water (USP, 9.0 kg) at 50
°C for 2 h to remove residual acetone (1.1%). Filtration and
drying at 50 °C, 10 mmHg for 16 h gave the final API 9 (0.895
kg, 99% yield, 99.2 AP, >99.9% ee). Ru < 1 ppm. Mp 195.3
°C. 1H NMR 10.45 (s, 1H), 8.80 (s, 1H), 8.11 (dd, J ) 0.9, 8.4
Hz, 1H), 7.98 (dd, J ) 0.9, 8.4 Hz, 1H), 7.80 (d, J ) 9 Hz,
2H), 7.76-7.73 (m, 1H), 7.63-7.56 (m, 2H), 7.31 (d, J ) 9
Hz, 2H), 5.71 (s, 2H), 3.50 (brs, 1H), 2.67 (s, 3H), 2.54-2.45
(m, 2H), 1.72-1.55 (m, 4H), 1.42-1.34 (m, 2H). 13C NMR δ
170.3, 161.6, 159.3, 148.1, 142.3, 134.1, 130.1, 129.6, 126.6,
124.5, 124.4, 120.8, 115.9, 67.3, 57.1, 44.9, 32.1, 28.0, 25.7,
22.4. MS: 456.1 [M + 1]+.
(1S,2R)-2-[4-(2-Methylquinolin-4-ylmethoxyl)benzene-
sulfonylamino]cyclopentanecarboxylic Acid (8). To a 20-L
jacketed cylindrical reactor equipped with an impeller-style agitator,
condenser, thermocouple, baffle, and nitrogen inlet were charged
(1S,2R)-aminocyclopentane-1-carboxylic acid HCl salt (7, 926 g,
5.59 mol), 1,2-dichloroethane (12.3 kg), and N-methylmorpholine
(570 g, 5.64 mol). The reaction mixture was stirred for 30 min.
BSTFA (2.875 kg, 11.17 mol) was then added over 10 min. The
reaction mixture was stirred at 40 °C until a solution was obtained
and held for 2 h. The solution was cooled to 5-10 °C, and
N-methylmorpholine (0.951 kg, 9.41 mol) was added over 5-10
min. 6 (1.79 kg, 4.66 mol) was added to the reaction mixture while
maintaining the temperature between 5-10 °C. After 15 min of
stirring, the reaction mixture was warmed to 21 °C and left
Vol. 12, No. 6, 2008 / Organic Process Research & Development
•
1259

Acknowledgment
Supporting Information Available
We thank the analytical team of Wyeth Canada for analysis.
HPLC conditions and retention time for intermediates 14,
2, 3, 15, 6, 8, 17, 18, 19, and API 9. This material is available
free of charge via the Internet at http://pubs.acs.org.
Note Added after ASAP Publication: The version pub-
lished on the Web September 9, 2008, has been changed. Two
authors have been added, and the acknowledgment has been
changed in the version published on the web September 27,
2008, and in the print version.
Received for review April 11, 2008.
OP800090S
1260
•
Vol. 12, No. 6, 2008 / Organic Process Research & Development