Process development and large-scale synthesis of a PDE4 inhibitor

Article abstract of DOI:10.1021/op050116l

An efficient, scalable synthesis of the PDE4 inhibitor, 6-[1-methyl-1-(methylsulfonyl)ethyl]-8-(3-{(E)-2-(3-methyl-1,2,4-oxadiazol-5-yl) -2-[4-(methylsulfonyl)phenyl]vinyl}phenyl)-quinoline benzenesulfonate (10) is described. The synthesis is highly conve

Full text of DOI:10.1021/op050116l

Organic Process Research & Development 2006, 10, 36−45
Process Development and Large-Scale Synthesis of a PDE4 Inhibitor
David A. Conlon,*,† Antoinette Drahus-Paone,† Guo-Jie Ho,† Brenda Pipik,† Roy Helmy,‡ James M. McNamara,†
Yao-Jun Shi,† J. Michael Williams,† Dwight Macdonald,§ Denis Descheˆnes,§ Michel Gallant,§ Anthony Mastracchio,§
Bruno Roy,§ and John Scheigetz§
Department of Process Research and Department of Analytical Research, Merck Research Laboratories, PO Box 2000,
Rahway, New Jersey 07065-0900, U.S.A., and Medicinal Chemistry Department, Merck Frosst Centre for Therapeutic
Research, P.O. Box 1005, Pointe Claire-DorVal, Que´bec, Canada H9R
Abstract:
An efficient, scalable synthesis of the PDE4 inhibitor, 6-[1-
methyl-1-(methylsulfonyl)ethyl]-8-(3-{(E)-2-(3-methyl-1,2,4-oxa-
diazol-5-yl)-2-[4-(methylsulfonyl)phenyl]vinyl}phenyl)-
quinoline benzenesulfonate (10) is described. The synthesis is
highly convergent, generating the penultimate 9 by coupling
aldehyde 7 and oxadiazole 8 in a Knoevenagel reaction. The
process consists of a total of nine chemical steps, five of which
comprise the sequence to prepare aldehyde 7 via Skraup
reaction, bromination, sulfone formation, methylation and
The synthetic route4 used to prepare 9 by Merck medicinal
Suzuki-Miyaura cross-coupling, and a two-step sequence for
chemists was suitable for delivery of multigram quantities
the synthesis of oxadiazole 8 that includes the methylamidoxime
of the final product. However, the synthesis was not directly
and oxadiazole steps. The final two steps are Knoevenagel
amenable to large-scale production. The quinoline ring in 9
coupling and salt formation. The process produced the drug
was prepared by condensing 2-bromo-4-methylaniline (1)
candidate 10 in 46% overall yield from 2-bromo-4-methylaniline
with glycerol in a Skraup reaction. To our knowledge, the
(1) on multikilogram scale.
Skraup reaction had not been used previously on a large
scale. The Skraup reaction is usually run in refluxing mineral
acid and typically gives low yields. In addition, the Skraup
reaction is exothermic, and the potential for accumulation
of reagents and a large exotherm during the addition of
glycerol was a major concern.5 In fact, a quote from the
Organic Reactions review states that the conditions under
which the earlier Skraup syntheses were carried out often
resulted in reactions of uncontrollable Violence.6 Clearly,
Introduction
Type-4 cyclic adenosine 3′,5′-monophosphate (c-AMP)
specific phosphodiesterase (PDE-IV) is an enzyme respon-
sible for the hydrolysis of the second messenger c-AMP to
AMP in many cell types.1 High levels of c-AMP inhibit the
production of cytokines and other molecules that modulate
significant development would be required to render the
the inflammatory response. Inhibition of PDE4 suppresses
requisite Skraup reaction amenable to scale-up.
antigen-induced release of histamine in vitro, and PDE4
The second step in the reaction sequence involved
inhibitors have emerged as potential targets for intervention
benzylic bromination of the product of the Skraup reaction,
in inflammatory diseases.2 PDE4 inhibitors such as 9 are
6-methyl-8-bromoquinoline (2). The yield for the radical-
being developed for the treatment of pulmonary diseases such
as asthma and chronic obstructive pulmonary disease (COPD)
that together afflict an estimated 600 million people world-
wide.3
induced bromination was moderate (63%) due to the
incomplete conversion of starting material and the formation
of the undesired R,R-dibromo compound 3a. Furthermore,
the bromination procedure was originally performed in
* To whom correspondence should be addressed. E-mail: dave_conlon@
merck.com.
(4) (a) Descheˆnes, D.; Dube´, D.; Gallant, M.; Girard, Y.; Lacombe, P.;
Macdonald, D.; Mastracchio, A.; Perrier, H. Patent WO 01/46151, 2001.
(b) Macdonald, D.; Mastracchio, A.; Perrier, H.; Dube´, D.; Gallant, M.;
Lacombe, P.; Descheˆnes, D.; Scheigetz, J.; Roy, B.; Bateman, K.; Li, C.;
Trimble, L. A.; Day, S.; Chauret, N.; Nicoll-Griffith, D. A.; Silva, J. M.;
Huang, Z.; Laliberte´, F.; Liu, S.; Ethier, D.; Pon, D.; Muise, E.; Boulet, L.;
Chan, C. C.; Styhler, A.; Charleson, S.; Mancini, J.; Masson, P.; Claveau,
D.; Nicholson, D.; Turner, M.; Young, R. N.; Girard, Y. Bioorg. Med. Chem.
Lett. 2005, 15, 5241-5246.
† Department of Process Research, Merck Research Laboratories.
‡ Department of Analytical Research, Merck Research Laboratories.
§ Medicinal Chemistry Department, Merck Frosst Centre for Therapeutic
Research.
(1) Claveau, D.; Chen, S. L.; O’Keefe, S.; Zaller, D. M.; Styhler, A.; Liu, S.;
Huang, Z.; Nicholson, D. W.; Mancini, J. A. J. Pharmacol. Exp. Ther. 2004,
310, 752-760.
(2) (a) Sourness, J. E.; Foster, M. Drugs 1998, 1, 541-553. (b) Dyke, H.;
Montana, J. Expert Opin. InVest. Drugs 1999, 8, 1301-1325.
(3) (a) Wilhelm, R. S.; Fatheree, P. R.; Chin, R. L. Quinolines as Type IV
Phosphodiesterase I inhibitors. Patent WO 9422852, 1994. (b) Macdonald,
D.; Perrier, H.; Liu, S.; Laliberte´, F.; Rasori, R.; Robichaud, A.; Masson,
P.; Huang, Z. J. Med. Chem. 2000, 43, 3820-3823.
(5) Larsen, R. D.; Cai, D. In Category 2, Hetarenes and Related Ring Systems;
Black, D. StC., Ed.; Science of Synthesis: Houben-Weyl Methods of
Molecular Transformations, Volume 15; Georg Thieme Verlag KG:
Stuttgart, 2005; pp 389-549.
(6) Manske, R. H. F.; Kulka, M. In Organic Reactions; Adams, R., Ed.; John
Wiley & Sons: New York, 1953; Vol. VII, Chapter 2, pp 59-98.
36
•
Vol. 10, No. 1, 2006 / Organic Process Research & Development
10.1021/op050116l CCC: $33.50 © 2006 American Chemical Society
Published on Web 12/07/2005

Scheme 1. Preparation of aldehyde 7 from 2-bromo-4-methylaniline (1)
Scheme 2. Skraup reaction process
carbon tetrachloride, making this procedure unsuitable for
further scale-up.
Other process concerns were identified in the methylation
step used to introduce two methyl groups which had the
potential of forming impurities that were difficult to remove,
high levels of residual palladium resulting from the Suzuki-
Miyaura cross-coupling reaction to append the 8-aryl sub-
stituent using the homogeneous catalyst Pd(PPh3)4, control
of the double bond geometry in the Knoevenagel reaction,
and identification of a stable and pharmaceutically acceptable
form of the final product.
In this contribution, we describe the process development
and multikilogram scale preparation of compound 10, the
benzensulfonic acid salt of 9, to support programs in
preclinical and clinical development. An efficient synthesis
of this compound was accomplished by addressing several
processing issues in the original synthesis to make it more
practical prior to scale-up. Our ultimate goal was to develop
a safe and productive synthesis for preparing 10 on multi-
kilogram scale.7
Scheme 3. Quinoline impurities
To circumvent a sudden exotherm during the reaction, the
addition rate of glycerol was controlled to preclude the
accumulation of glycerol. In practice, the optimized proce-
dure is to add sodium m-nitrobenzenesulfonate and iron
sulfate heptahydrate to methane sulfonic acid, followed by
the addition of 2-bromo-4-methylaniline (1), resulting in an
exotherm that warms the batch to ∼60 °C. The reaction
mixture is then slowly heated to 120 °C, and glycerol is
added over ca. 10 h.9 It is important to allow the reaction to
proceed to a conversion of >99% since the starting material,
2-bromo-4-methylaniline (1), inhibits the following radical-
induced bromination and cannot be effectively removed
during the workup.
It is interesting to note that an excess of glycerol is
required for the Skraup reaction. In addition to the loss due
to polymerization, some of this material is consumed to form
two quinoline sulfonic acids (1c and 1d) derived from the
expected aniline byproduct 1b (Scheme 3). These byproducts
are easily removed during the extraction workup (see below).
Once the reaction is complete, the reaction temperature
is lowered to 70-80 °C, and the batch is diluted with water.
Results and Discussion
We believed that the original route to aldehyde 7 would
be viable for large-scale production if we could make the
necessary improvements. The optimized procedure for the
preparation of aldehyde 7 is shown in Scheme 1.
Skraup Reaction. The Skraup reaction is an efficient
method for the construction of the quinoline ring system.
The Skraup reaction is proposed to occur as four successive
steps (Scheme 2): (1) dehydration of glycerol to acrolein,
(2) addition of the aromatic amine to acrolein, (3) ring closure
and dehydration to form the dihydroquinoline, and (4)
oxidation of the dihydroquinoline to quinoline.8 None of the
proposed intermediates are observed by HPLC or GC even
as the reaction rate slows down near the end of the reaction.
The exothermic nature of these steps presents a significant
challenge on a large scale. Indeed, the Skraup reaction with
1 was shown to have a significant exotherm upon scale-up.
(7) Desmond, R.; Conlon, D. A.; Drahus, A.; Ho, G.-J.; Pipik, B.; LeBlond,
C.; Vailaya, A. U.S. Patent 6,835,837, 2004.
(8) (a) Li, J. J. Name Reactions: A Full Collection of Detailed Reaction
Mechanisms; Springer-Verlag: New York, 2002; pp 378-379. (b) For an
alternative mechanism, see: Eisch, J. J.; Dluzniewski, T. J. Org. Chem.
1989, 54, 1269-1274.
(9) In the pilot plant an emergency water quench was set up prior to the heatup
to 120 °C and glycerol addition.
Vol. 10, No. 1, 2006 / Organic Process Research & Development
•
37

Scheme 4. Phenolic impurities generated during the
Scheme 6. Sodium methanesulfinate coupling with 3
Skraup reaction
AIBN required a reaction temperature of ∼70 °C for
radical initiation and produced a highly toxic byproduct,
tetramethylsuccinonitrile, by homocoupling of two isobuty-
ronitrile radicals.12 While the level of tetramethylsuccino-
nitrile was expected to be low, we decided to change the
radical initiator. On the basis of the solubility of the product
3 in chlorobenzene it seemed possible that running the
reaction at a lower temperature, where the desired product
3 would crystallize directly from the reaction solution, might
be a simpler way to effectively reduce the undesired over-
bromination which forms 3a.13 Thus, Vazo-52 was selected
as the initiator, and the reaction was carried out at ∼50 °C,
giving a ratio of 7:88:5 for 2/3/3a at the end of reaction.
The use of Vazo-52 did present some challenges due to the
handling and storage issues associated with this low-
temperature initiator.14 In addition, due to the large exotherm
associated with the reaction, NBS and Vazo-52 were charged
in two portions to address safety concerns.15 Compound 3
was isolated as a mixture with succinimide, the byproduct
from NBS, by diluting the reaction mixture with cyclohexane.
Succinimide was easily removed by washing the crude solid
with aqueous methanol. From this mixture, 3 was isolated
in 75-86% yield. In addition to the increased yield, the
decomposition product from Vazo-52 is much less toxic than
that from AIBN.16
Sulfone Formation. Displacement of the benzylic bro-
mide in 3 with sodium methanesulfinate generates the sulfone
4. Initially, the sulfone formation was carried out in DMF
(1.5 mL/g substrate 3), and the reaction was quenched by
the addition of water which concomitantly crystallized 4. The
addition of water to the reaction mixture at this concentration
resulted in the formation of a thick slurry, making agitation
difficult. To simplify the isolation, the reaction was run less
concentrated (4 mL/g). The reaction mixture was allowed
to react at 15-20 °C until ∼90% conversion was achieved
and was then heated to 75-80 °C to ensure complete
conversion. Although the amount of the side product sulfinate
ester 4a (see Scheme 6) increases at a higher reaction
temperature, the elevated temperature was required for
complete conversion. The sulfinate ester 4a can be converted
to 4 under the reaction conditions13 via an intermolecular
reaction with sodium methanesulfinate.17 Furthermore, we
Scheme 5. Radical bromination of 2
Neutralization of the methanesulfonic acid is accomplished
by the addition of 10 M NaOH and 1 M NaHCO3.10 Methyl
tert-butyl ether (MTBE) extraction of bromomethylquinoline
2 followed by aqueous washes gives a solution of 2 in 90-
92% assay yield. Two phenolic impurities (1e and 1f)
(Scheme 4) are formed during the reaction at levels of 0.5-
1%, depending upon the reaction temperature and time.11 It
was found that the presence of these impurities inhibits the
radical bromination used in the subsequent step and therefore
must be controlled in the Skraup step.
These impurities were completely removed after two
dilute H3PO4 washes (0.02 M, pH ≈ 2) with ∼2-3% loss
of 2. The product 2 was then extracted into aqueous sulfuric
acid (1.2 M). This extraction did not remove 2,6-dibromo-
4-methylaniline, an impurity in the starting material which
also inhibited the radical bromination. Aqueous hydrochloric
acid could not be used for this extraction because the
hydrochloride salt of 2 crystallized from the acid solution.
Finally, bromomethylquinoline 2 was extracted into chlo-
robenzene, the solvent used in the radical bromination, after
neutralization with aqueous sodium hydroxide.
Bromination. As shown in Scheme 1, the next step in
the process is the radical bromination of 2. In our initial work
we utilized N-bromosuccinimde (NBS) and 2,2′-azobisisobu-
tyronitrile (AIBN) as the radical initiator for the benzylic
bromination of the bromomethylquinoline product from the
Skraup reaction to produce 3. The formation of 3a increased
dramatically toward the end of the reaction as the concentra-
tion of 3 increased (Scheme 5). To avoid over-bromination,
we crystallized 3 from solution by the addition of an
antisolvent (cyclohexane) after ∼70-80% conversion. Ad-
ditional NBS and AIBN were then added to push the reaction
to completion, giving a 12:75:13 ratio of 2/3/3a at the end
of reaction and 3 in 68-71% isolated yield.
(12) Tetramethylsuccinotrile is highly toxic; the oral toxicity in rats (lethal dose
for 50% of the test animals) is 60 mg/kg. DuPont Product Literature (http://
www.dupont.com/vazo/faq.html).
(13) Subsequent work demonstrated that the R,R-dibromomethyl compound could
be converted to the desired sulfone in the next step, see Xu, F.; Savary, K.;
Williams, J. M.; Grabowski, E. J. J.; Reider, P. J. Tetrahedron Lett. 2003,
44, 1283-1286.
(10) The use of 1 M NaHCO3 simplified the neutralization by removing the
need to monitor the pH.
(11) We speculated that 1e arises from a Bamberger rearrangement of the
phenylhydroxylamine intermediate formed during the partial reduction of
nitrobenzene in the Skraup reaction. Impurity 1f is formed from 1e during
the Skraup reaction. We did not determine if nitrobenzene was formed by
loss of H2SO4 from sodium m-nitrobenzenesulfonate under the acidic
reaction conditions or was present in this reagent.
(14) The maximum recommended storage temperature for Vazo-52 is 10 °C.
(15) With the split charge the maximum adiabatic temperature rise was reduced
to 65 °C.
(16) The oral toxicity in rats (lethal dose for 50% of the test animals) for the
decomposition products of Vazo-52 is 5000 mg/kg. DuPont Product
Literature (http://www.dupont.com/vazo/faq.html).
38
•
Vol. 10, No. 1, 2006 / Organic Process Research & Development

Scheme 7. Alkylation with methyl iodide
Scheme 8. Impurity formation during alkylation
Scheme 9. Carboxylic acid impurity formed from anion
oxidation
found that the product crystallized directly from the reaction
mixture upon addition of water at 50-60 °C. Due to partial
protonation, 4 is more soluble under acidic conditions;
therefore, sodium bicarbonate was added to increase the pH
and reduce the loss of 4. Sulfone 4 was isolated in 96% yield.
Methylation. In the next step of the process, it was
necessary to introduce two methyl groups on the benzylic
carbon between the aromatic ring and the sulfone group in
4 to generate 5 (Scheme 7). A significant issue for this step
was the potential to form two impurities which are difficult
to remove downstream; thus, it was critical that their
production be minimized. An undercharge of base and/or
methyl iodide resulted in the incomplete conversion of the
intermediate monomethyl derivative 5a. However, an over-
charge resulted in the formation of the dimethylethyl sulfone
5b (Scheme 8). With these factors in mind, we attempted to
carefully control the charges during our initial work with
limited success. One equivalent of base (sodium tert-
butoxide) was added followed by 1 equiv of methyl iodide.
The reaction was assayed by HPLC to monitor conversion,
and then this process was repeated. For each batch, an HPLC
assay was performed after 2 equiv of sodium tert-butoxide
and methyl iodide were added. The assay was used to
determine the amount of residual monomethylated intermedi-
ate 5a that remained (typically 6-12%). Then a third charge
of sodium tert-butoxide and methyl iodide was made on the
basis of the amount of 5a remaining.
A variety of other bases were screened for the dimethy-
lation reaction with no base showing advantage over sodium
tert-butoxide. Further improvement on the selectivity of the
reaction was realized during screening studies which explored
the effect of water concentration on this reaction. This
revealed that, at higher water concentration (4000-6000 µg/
mL, 0.9-1.3 equiv), the reaction remains a solution or very
thin slurry, and better conversion of intermediate 5a to
product 5 was obtained. In addition, at the higher water
concentrations, we did not generate the difficult-to-reject
dimethylethyl sulfone impurity 5b even with an excess
charge of reagents. At low water concentration (<1000 µg/
mL), the reaction mixture becomes a very thick slurry that
is difficult to stir, and the reaction does not go to completion.
With more water present, the sodium tert-butoxide charged
to the reaction is rapidly converted to hydroxide which is
more selective in deprotonation at the benzylic position.
Hydroxide is not strong enough to significantly deprotonate
the sulfone methyl in 5 (pKa ≈ 33)18 as needed to form 5b.
The equilibration to hydroxide as base was tested by
adding solid sodium tert-butoxide to a solution of the sulfone
and methyl iodide in dry DMF. Since there was no water
present, equilibration of base to hydroxide did not occur,
and a large amount of dimethylethyl sulfone 5b was formed.
Conversely, at the higher water concentration the benzylic
anion is generated, and less of the overalkylated ethyl
impurity 5b is formed. Attempts to replace sodium tert-
butoxide with sodium hydroxide either directly or under
phase transfer conditions did not result in an improvement.
If the acting base is sodium hydroxide, which will not
deprotonate the sulfone methyl, then an excess of base and
methyl iodide should not result in formation of large amounts
of 5b. For the large-scale synthesis, following the addition
of 2 equiv of base and methyl iodide, a standard third charge
of 15 mol % of sodium tert-butoxide and methyl iodide was
made. This procedure was successfully implemented to give
5 in 90-92% yield and 99.5 HPLC area % (LCAP).
During reaction optimization, a previously seen low-level
impurity was formed at elevated levels in several experi-
ments. This impurity was isolated and identified by LC/MS
and NMR as the carboxylic acid impurity 5c (Scheme 9).
We speculated that this impurity is generated by oxidation
of the sulfone anion and determined that degassing is a
critical parameter in the dimethylation reaction. To demon-
strate that this impurity was being formed via reaction with
oxygen, an experiment was performed where the sulfone
anion was prepared in DMF at ambient temperature without
degassing. The anion was then quenched and the resulting
reaction mixture analyzed by HPLC, revealing elevated levels
of 5c. Therefore, the reaction mixture was thoroughly
degassed using vacuum and nitrogen back fills for the large-
scale preparation, and the level of 5c at the end of reaction
was nearly undetectable, and none was observed in the
isolated product.
Suzuki-Miyaura Cross-Coupling. The final step in the
synthesis of aldehyde 7 is the Suzuki-Miyaura cross-
coupling reaction of the dimethyl sulfone 5 and 3-formylphe-
(18) Bordwell, F. G.; Bares, J. E.; Bartmess, J. E.; McCollum, G. J.; Van der
Puy, M.; Vanier, N. R.; Matthews, W. S. J. Org. Chem. 1977, 42, 321-
325.
(17) (a) Kornblum, N.; Ackermann, P.; Swiger, T. J. J. Org. Chem. 1980, 45,
5294-5298. (b) Creary, X. J. Org. Chem. 1985, 50, 5080-5084.
Vol. 10, No. 1, 2006 / Organic Process Research & Development
•
39

Scheme 10. Suzuki coupling reaction with Pd/C
Scheme 11. Homocoupled sulfone impurity
formation of the known biaryl.23 In addition, we observed
lower residual palladium levels in reactions that were
degassed. The reaction mixture is heated to 80 °C and stirred
for 1.5 h. The Pd/C and the other inorganic byproducts are
removed by filtration, and the waste cake is washed with
hot DMF to ensure a good recovery of the aldehyde. The
filtration and waste cake washing are performed at temper-
atures above 40 °C to prevent loss of the aldehyde by
crystallization.
The issue with this step was the low mass balance of the
reaction that was originally attributed to loss of product on
the waste cake during filtration. Our investigations identified
the formation of the homocoupled sulfone 7a (LC-MS/MS)
that was completely removed during the isolation. An
authentic sample of 7a was prepared using the procedure
described by Rawal et al.24 (Scheme 11).
The aldehyde intermediate 7 was isolated by the addition
of water (equal volume) to the combined filtrate and washes
at 50-60 °C. The aldehyde crystallized without seeding and
was isolated by filtration. We were pleased to find that the
residual palladium levels were very low in the isolated
material (18-80 ppm) and were reduced to <10 ppm during
further processing.
Methylamidoxime Preparation and Oxadiazole Syn-
thesis. The first step in the preparation of the oxadiazole 8
is the formation of methylamidoxime 8a from hydroxy-
lamine25 and acetonitrile.26,27 This step is complicated by the
energetic nature of the starting material, hydroxylamine, and
the product 8a. In the process, acetonitrile is heated to 70-
72 °C and a 50 wt % aqueous solution of hydroxylamine is
added slowly over at least 4 h to avoid the accumulation of
the reagents. The reaction is very exothermic, and to
minimize the potential for charging the hydroxylamine too
quickly, the charge was done in two discrete portions.
Performing the charge in this manner ensured complete
consumption of the hydroxylamine and provided an easily
controlled reaction with a negligible heat-release rate.
Following the hydroxylamine addition, the reaction mixture
was cooled to ∼30 °C, seeded, and cooled further to ca. 0
°C, and the resulting slurry was filtered. Due to the thermal
instability of the product, 8a, drying of the isolated solid
was done in a vacuum oven at ∼25 °C.
nyl boronic acid 6 (Scheme 10). Our initial screening of
several homogeneous palladium catalysts identified Pd(dppf)-
Cl2 as an excellent catalyst for this coupling.19 Unfortunately,
the isolated aldehyde 7 contained high levels of residual
palladium (3500 ppm) as well as iron (1160 ppm) from the
dppf ligand. Reduction of the residual Pd concentration and
removal of the dppf ligand to acceptable levels represented
a major challenge for the production of 7 of suitable quality.
A major objective during the design of a suitable process
was the control of the palladium level in the isolated aldehyde
7. To avoid the purification required to remove these metals,
we investigated the heterogeneous (Pd/C) catalyzed reac-
tion.20 Palladium catalysts from several different suppliers
were evaluated, and we determined that the optimum catalyst
for the reaction to form 7 had nonreduced palladium (5% or
10% Pd) distributed on the activated carbon surface (an egg-
shell dispersion). Both dry catalyst and water-wet catalyst
performed equally well in the reaction.21 The Pd/C catalyzed
Suzuki-Miyaura reaction was shown to give 7 in excellent
yield with low levels of residual palladium after a simple
filtration and precipitation. The low level of residual pal-
ladium is an important criterion for pharmaceuticals, and the
use of a heterogeneous catalyst is motivated by the ease of
separation of the palladium from the product. In the example
above, leaching of palladium from the carbon support
occurred during the reaction followed by subsequent pre-
cipitation of Pd during the course of the reaction, prompting
us to refer to this as a quasi-heterogeneous reaction.
Screening studies showed that aqueous DMF was the best
solvent system for this reaction and that powdered K2CO3
was the most effective base. Extra fine grade potassium
carbonate22 was used to minimize mass transfer limitations
ensuring complete conversion. Optimization of the Pd level
(2.6 mol %) and reaction conditions (DMF/water, 10:1, 3
equiv of K2CO3, and 80 °C) resulted in reproducible, isolated
yields of aldehyde 7 of 85-92%.
In the typical procedure all the reagents are charged, and
the reaction mixture is thoroughly degassed using vacuum
and nitrogen back fills. The degassing minimizes the loss of
3-formylphenyl boronic acid due to homocoupling and
(19) The cross-coupling reaction is inhibited by phosphine; however, chelating
phosphines such as dppf can be used. See Wallow, T. I.; Novak, B. M. J.
Org. Chem. 1994, 59, 5034-5037.
(20) Conlon, D. A.; Pipik, P.; Ferdinand, S.; LeBlond, C. R.; Sowa, J. R., Jr.;
Izzo, B.; Collins, P.; Ho, G.-J.; Williams, J. M.; Shi, Y.-J.; Sun, Y. AdV.
Synth. Catal. 2003, 345, 931-935.
(21) We found that many commercial samples of Pd/C were catalysts for this
cross-coupling. We used a dry, 5% Pd/C (C6064) from Engelhard and water
wet 5% Pd/C (A405023-5) and 10% Pd/C (A402032-10) catalysts from
Johnson Matthey in this study.
(23) Penalva, V.; Hassan, J.; Lavenot, L.; Gozzi, C.; Lemaire, M. Tetrahedron
Lett. 1998, 39, 2559-2560.
(24) Hennings, D. D.; Iwama, T.; Rawal, V. H. Org Lett. 1999, 1, 1205-1208.
(25) CAUTION: Hydroxylamine is a possible mutagen and is known to cause
severe skin, eye, and respiratory tract burns. Heating hydroxylamine
solutions may result in an exothermic decomposition, see ref 27.
(26) Eloy, F.; Lenaers, R. Chem. ReV. 1962, 62, 155-183.
(22) Extra fine K2CO3 was purchased from the Armand Products Co., Princeton,
NJ 08543.
(27) Hett, R.; Kra¨hmer, R.; Vaulont, I.; Leschinsky, K.; Snyder, J. S.; Kleine, P.
H. Org. Process Res. DeV. 2002, 6, 896-897.
40
•
Vol. 10, No. 1, 2006 / Organic Process Research & Development

Scheme 12. Formation of the oxadiazole
Scheme 13. Knoevenagel coupling to prepare 9
Scheme 14. Double bond isomers of 9
For the formation of oxadiazole 8, we initially investigated
methods with literature precedent that involved activation
of 4-methanesulfonylphenyl acetic acid (8b) with N,N′-
carbonyldiimidazole (CDI), N,N′-dicyclohexylcarbodiimide
(DCC), or formation of an acid chloride or anhydride to
facilitate coupling with methylamidoxime 8a. These proce-
dures were not suitable for large-scale processing because 8
was obtained in low-yield or required chromatography to
afford product with acceptable purity; therefore, we sought
a more efficient route to 8. The method developed involved
activation of 8b using standard peptide coupling reagents
1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochlo-
ride (EDC‚HCl) and 1-hydroxybenzotriazole (HOBt)28 fol-
lowed by displacement with methylamidoxime (Scheme 12).
It was critical to ensure that the active ester formation was
complete prior to the addition of methylamidoxime. Lower
conversion due to hydrolysis of the active ester was observed
if the water from the HOBt hydrate was not removed. In
addition, the conversion was lower when all the reagents
were charged at once. Cyclization and dehydration occurred
upon heating, and the oxadiazole was produced in 90% yield
on a 30-kg scale using this method.29
the reaction to completion. The coupling step was performed
in 2-propanol (IPA) at reflux (bp 82 °C) with piperidine as
the base. Under these conditions, the reaction remains a slurry
throughout because the starting material 7 (aldehyde, 0.15
mg/mL) and product 9 (free base, 0.03 mg/mL) have limited
solubility in IPA. Water formed during the condensation was
removed by returning the distillates to the reaction vessel
after passing through a bed of molecular sieves.
A mixture of 9/9a isomers was initially observed in the
supernatant of the reaction by HPLC (Scheme 14); however,
as the reaction proceeded, the desired product 9 crystallized
out. Toward the end of reaction we isolated the desired
product 9 by filtration, and any residual undesired 9a was
rejected in the filtrate. The reaction required 22 h to go to
completion, was cooled to 20-25 °C, and then stirred for 4
h prior to isolation of the product by filtration.
Compound 9 is light sensitive, and isomerization of the
double bond was observed upon exposure to light in both
the solid state and in solution. Care was then taken to protect
9 from exposure to light.
Benzenesulfonate Salt Formation. In an effort to identify
a stable and pharmaceutically acceptable form of the final
product, screening experiments were performed with the free
base 9, and several different crystalline salts were discovered.
One of them was the crystalline benzenesulfonate 10
(Scheme 15).
Knoevenagel Coupling. The Knoevenagel reaction30 was
an attractive approach for the formation of the penultimate
9 that contained a trisubstituted double bond (Scheme 13).
The possibility of forming a mixture of double bond
isomers was a major concern; however, the isolated product
was shown to be exclusively the desired isomer. The reaction
is reversible, and the water formed must be removed to drive
During our investigations, several different crystal forms
of 10 were observed. Development of a solvent system to
ensure reproducible production of the desired crystal form
was needed. In our investigations, it was found that solvates
were formed with methanol, ethanol, acetonitrile, and
acetone, at or below room temperature. This excluded the
use of these and similar solvents in the production of 10.
Two different solvent systems were found to reproducibly
give the desired crystal form. These systems are N,N-
dimethylformamide (DMF)/isopropyl acetate (IPAc) and
(28) CAUTION: HOBt will decompose, possibly violently, if heated above its
melting point. See Bright, R.; Dale, D. J.; Dunn, P. J.; Hussain, F.; Kang,
Y.; Mason, C.; Mitchell, J. C.; Snowden, M. J. Org. Process Res. DeV.
2004, 8, 1054-1058.
(29) Pipik, B.; Ho, G.-J.; Williams, J. M.; Conlon, D. A. Synth. Commun. 2004,
34, 1863-1870.
(30) (a) Jones, G. Org. React. 1967, 15, 204-599. (b) Tietze, L.; Beifuss, J. In
ComprehensiVe Organic Synthesis; Trost, B. M., Ed.; Pergamon: Elmsford,
NY, 1991; Vol. 2, pp 341-394.
Vol. 10, No. 1, 2006 / Organic Process Research & Development
•
41

Scheme 15. Final salt formation
resulting mixture was added 2-bromo-4-methylaniline (3.50
kg, 18.8 mol) through an addition funnel with stirring over
30 min, flushed with methanesulfonic acid (0.5 L), and
agitated until the solid dissolved (∼15 min). The addition is
exothermic, and the temperature reached ∼60 °C after the
addition. The reaction mixture was heated to 118-125 °C,
and glycerol (4.33 kg, 47.0 mol) was added through an
addition funnel over 4-8 h. After stirring at 125-133 °C
for 10-16 h, the mixture was cooled to ∼80 °C and diluted
with cold DI water (10 L) keeping the temperature <90 °C.
After cooling to ∼20 °C with an ice bath, the mixture was
neutralized with cold aqueous NaOH (10 M, 15.6 L). Solka
floc (500 g) was added, and the mixture was further
neutralized with 1 M aqueous NaHCO3 (1 M, 16 L). The
mixture was cooled to 10-20 °C, and MTBE (30 L) was
added. After stirring for 4-6 h, the mixture was filtered
through a centrifuge (10-µm bag), and the phases were
separated. The (top) MTBE layer was washed with 20 L of
brine (95% saturation). The aqueous layers were back
extracted with MTBE, and the organic phases were combined
(3.97 kg of product as a 7-9 wt % MTBE solution in 95%
yield). For characterization purposes, the hydrochloride salt
of 8-bromo-6-methylquinoline can be prepared. Mp 195-
acetic acid/isopropyl alcohol. The DMF/IPAc system was
selected for large-scale preparation because it afforded better
yields. The free base 9 and benzenesulfonic acid (1.2 equiv)
were dissolved in an 80:20 (v/v) mixture of DMF/IPAc at
40 °C. The clear, yellow solution was transferred through a
line filter to a second vessel and seeded with 10. Additional
isopropyl acetate was added to the reaction solution as an
antisolvent at 40 °C to crystallize 10. It was found that the
level of residual DMF in isolated 10 was lower if the
crystallization was performed at a temperature of 40 °C or
above.31 Once the antisolvent addition was complete, the
slurry was slowly cooled from 40 to 5 °C over 5 h and then
stirred for an additional 5 h. The product was isolated with
the desired crystal form in 91% yield and with acceptable
residual DMF levels.
1
211 °C. H NMR (400 MHz, CDCl3): δ 2.67 (s, 3H), 7.91
(s, 1H), 8.03 (dd, J ) 5.3, 8.2 Hz, 1H), 8.20 (s, 1H), 8.84
(d, J ) 8.2 Hz, 1H), 9.67 (d, J ) 4.8 Hz, 1H). 13C NMR
(100.6 MHz, CDCl3): δ 18.4, 111.5, 121.1, 126.3, 129.0,
132.5, 138.9, 140.6, 143.9, 146.1. Anal. Calcd for C10H8-
BrN: C, 54.08; H, 3.63; Br, 35.98; N, 6.31. Found: C, 53.94;
H, 3.35; Br, 36.06; N, 6.15.
Conclusion
An efficient and safe process for the preparation of the
drug candidate 10 was developed which provided many
significant advantages over the original process used by
Merck medicinal chemists. The synthesis is highly conver-
gent, coupling aldehyde 7 and oxadiazole 8 to yield the
penultimate 9. The process for preparing 10 with the
preferred crystal form was scaled up successfully and reliably
in 46% overall yield from commercially available 2-bromo-
4-methylaniline (1) on multikilogram scale in a pilot-plant
facility.
8-Bromo-6-(bromomethyl)quinoline (3). The MTBE
solution from the Skraup reaction (∼47 L, 3.95 kg, 17.8 mol)
was washed with 2 × 70 L of 0.02 M aqueous H3PO4, and
then the MTBE layer was extracted with 42 L of 1.2 M
aqueous H2SO4, followed by 14 L of 1.2 M aqueous H2SO4,
while maintaining temperature at 10-20 °C during extrac-
tion. The H2SO4 extracts were combined, and chlorobenzene
(49 L) was added. The mixture was then neutralized to pH
> 5 with 10 M aqueous NaOH (13.3 L). The phases were
separated, and the chlorobenzene layer was washed with 28
L of brine, concentrated to ∼11 L, and filtered (325 g/L,
3.46 kg, 15.6 mol). To the solution at ∼20 °C was added
NBS (1.52 kg, 8.5 mol) and Vazo-52 (155 g, 0.62 mol). The
mixture was degassed by vacuum and purged with N2 three
times. The mixture was heated to 48-52 °C over 1 h and
maintained at 48-52 °C for 8 h. The batch was cooled to
∼20 °C, and a second charge of NBS (1.52 kg, 8.5 mol)
and Vazo-52 (155 g, 0.62 mol) were made. The vessel was
purged with nitrogen and heated back to 48-52 °C for 10
h. The batch was allowed to cool to 20-25 °C, and
cyclohexane (11 L) was added over 1 h. After stirring at
∼20 °C for 2 h, the mixture was cooled to 0-5 °C and held
at this temperature for 2 h. The slurry was filtered, and the
cake was washed with 5.2 L of cold (0-5 °C) 1:2
chlorobenzene/cyclohexane, followed by 10 L of cyclohex-
ane. After air-drying for 2-4 h the cake was further dried
at 30 °C under full vacuum to remove residual solvents (<1
wt % organic solvents). The solid was slurry washed with 3
Experimental Section
General. Starting materials were obtained from com-
mercial suppliers and were used without further purification.
HPLC analyses were performed on a Hewlett-Packard Series
1100 liquid chromatograph equipped with a UV detector.
The area percent (LCAP) reported has been corrected for
detector response. NMR spectra were obtained at 400 MHz
1
for H and 100.6 MHz for 13C. All coupling constants are
reported in hertz (Hz). Microanalysis was performed by
Quantitative Technologies, Inc.
8-Bromo-6-methylquinoline (2). Sodium m-nitrobenze-
nesulfonate (2.67 kg, 11.8 mol) was added to methane-
sulfonic acid (10 L) at 20 °C followed by iron sulfate
heptahydrate (156.8 g, 0.564 mol). This addition is slightly
exothermic and raised the temperature to 27 °C. To the
(31) DMF is a Class 2 solvent, and residual levels must be controlled. See ICH
Harmonised Tripartite Guideline; Impurities: Guideline for Residual
Solvents Q3C; 17 July 1997.
42
•
Vol. 10, No. 1, 2006 / Organic Process Research & Development

× 11 L of 10 v/v% MeOH in water. The solid was dried at
30 °C under full vacuum to a constant weight (3.75 kg, 98.7
wt %, 79%). Mp 156-158 °C. 1H NMR (400 MHz,
CDCl3): δ 4.60 (s, 2H), 7.48 (dd, J ) 4.3, 8.3 Hz, 1H),
7.78 (d, J ) 1.4 Hz, 1H), 8.09-8.14 (m, 2H), 9.04 (dd, J )
1.5, 4.2 Hz, 1H). 13C NMR (100.6 MHz, CDCl3): δ 31.8,
122.4, 125.4, 127.5, 129.0, 133.9, 136.6, 136.7, 144.9, 151.7.
Anal. Calcd for C10H7Br2N: C, 39.91; H, 2.34; Br, 53.10;
N, 4.65. Found: C, 39.81; H, 2.12; Br, 53.23; N, 4.54.
8-Bromo-6-[(methylsulfonyl)methyl]quinoline (4). Pow-
dered sodium methanesulfinate (1.7 kg, 16.7 mol) was added
to dry DMF (14 L, KF ) 226 µg/mL) -3 to 0 °C, followed
by the addition of 3 (4.00 kg, 13.3 mol), and the temperature
of the slurry was maintained at 0-10 °C. The mixture was
stirred at 13-17 °C for 3 h. The mixture was then heated to
75-80 °C for 1 h. The slurry was diluted with water (42
L), while maintaining the batch temperature at 75-80 °C.
After the water addition was complete, the batch was heated
to 90 °C and then cooled to 20 °C. Sodium bicarbonate (223
g) was added, and the slurry was cooled to 0-5 °C and
stirred for 4 h. The mixture was filtered and the cake washed
sequentially with 2 × 6.7 L of cold (0-4 °C) 1:4 DMF/
water and 6.7 L of chilled water. After air-drying for 2 h,
the solid was further dried at 40-50 °C until constant weight.
Isolated yield: 3.91 kg, 12.75 mol, 96%. Typical quality:
98 wt % and 97-98 LCAP. Mp 149-152 °C. 1H NMR (400
MHz, CDCl3): δ 2.87 (s, 3H), 4.41 (s, 2H), 7.53 (dd, J )
4.3, 8.3 Hz, 1H), 7.90 (s, 1 H), 8.11 (d, J ) 1.3 Hz, 1H),
8.20 (d, J ) 8.2 Hz, 1H), 9.03, (dd, J ) 0.7, 4.0 Hz, 1H).
13C NMR (100.6 MHz, CDCl3): δ 39.8, 60.4, 122.7, 125.9,
127.2, 129.2, 130.1, 134.6, 136.8, 145.3, 152.3. Anal. Calcd
for C11H10BrNO2S: C, 44.01; H, 3.36; Br, 26.62; N, 4.67;
S, 10.68. Found: C, 43.04; H, 3.34; Br, 26.19; N, 4.46; O,
12.96; S, 10.01.
8-Bromo-6-[1-methyl-1-(methylsulfonyl)ethyl]quino-
line (5). Sulfone 4 (1648 g, 5.49 mol) was dissolved in DMF
(24 L, KF ) 5280 µg/mL), and the solution was degassed
using vacuum and nitrogen back fills; the solution was then
cooled to -10 to -5 °C under N2. Solid sodium tert-butoxide
(528 g, 5.49 mol) was added, while maintaining the tem-
perature at -10 to -5 °C. Neat methyl iodide (779 g, 5.49
mol) was added while maintaining the temperature at -10
to -5 °C. A second charge of solid sodium tert-butoxide
(528 g, 5.49 mol) was then added in one portion to the
reaction mixture at -10 °C. A second charge of the neat
methyl iodide (781 g, 5.5 mol) was added to the deep-red
reaction mixture, while maintaining the reaction temperature
between -10 and -3 °C. A third portion of sodium tert-
butoxide (79 g, 0.83 mol) was added, followed by methyl
iodide (120 g, 0.85 mol). The batch was stirred for 60 min
and for 40 min between -10 and -5 °C, and then aqueous
acetic acid solution (2 vol % acetic acid, 25 L) was slowly
added, and the temperature reached ∼10 °C. The resulting
slurry was stirred for 6 h at -5 to 0 °C and filtered. The
solid was washed sequentially with 2 × 5 L of cold (10-14
°C) 3:1 water/DMF (v/v) solution and 3 L of water and then
dried in a vacuum oven at ∼35 °C (nitrogen sweep) to a
constant weight. Isolated yield: 1647 g of 98 wt % 5, 88%
isolated yield.
1
Monomethyl methyl sulfone 5a: mp 172-176 °C. H
NMR (400 MHz, CDCl3): δ 1.90 (d, J ) 7.2 Hz, 3H), 2.75
(s, 3H), 4.35 (q, J ) 7.1 Hz, 1H), 7.54 (dd, J ) 4.1, 8.1 Hz,
1H), 7.93 (d, J ) 1.9 Hz, 1H), 8.15 (d, J ) 1.9 Hz, 1H),
8.21 (dd, J ) 1.5, 8.3 Hz, 1H), 9.10 (m, 1H). 13C NMR
(100.6 MHz, CDCl3): δ 14.2, 38.2, 64.0, 122.7, 125.8, 128.4,
129.1, 133.2, 133.4, 136.8, 145.3, 152.2. Anal. Calcd for
C12H12BrNO2S: C, 45.87; H, 3.85; Br, 25.43; N, 4.46; S,
10.21. Found: C, 45.74; H, 3.47; Br, 25.68; N, 4.65; S,
10.57.
1
Dimethyl sulfone 5: mp 205-207 °C. H NMR (400
MHz, CDCl3): δ 1.98 (s, 6H), 2.62 (s, 3H), 7.54 (dd, J )
4.1, 8.2 Hz, 1H), 8.08 (d, J ) 2.0 Hz, 1H), 8.23 (dd, J )
1.3, 8.1 Hz, 1H), 8.38 (d, J ) 2.1 Hz, 1H), 9.10 (dd, J )
1.5, 4.1 Hz, 1H). 13C NMR (100.6 MHz, CDCl3): δ 22.8,
35.0, 64.4, 122.6, 125.4, 127.8, 128.8, 132.6, 136.9, 137.3,
144.9, 152.3. Anal. Calcd for C13H14BrNO2S: C, 47.57; H,
4.30; Br, 24.34; N, 4.27; S, 9.77. Found: C, 47.47; H, 3.93;
Br, 24.43; N, 4.25; S, 9.81.
1
Dimethyl ethyl sulfone 5b: mp 210-212 °C. H NMR
(400 MHz, CDCl3): δ 1.24 (t, J ) 7.5 Hz, 3H), 1.97 (s,
6H), 2.71 (q, J ) 7.5 Hz, 1H), 7.53 (dd, J ) 4.2, 8.2 Hz,
1H), 8.07 (d, J ) 2.2, 1H), 8.22 (dd, J ) 1.6, 8.3 Hz, 1H),
8.37 (d, J ) 2.2 Hz, 1H), 9.09 (dd, J ) 1.7, 4.2 Hz). 13C
NMR (100.6 MHz, CDCl3): δ 5.2, 22.9, 41.2, 64.3, 122.5,
125.2, 127.7, 128.7, 132.6, 137.1, 137.3, 144.7, 152.2. Anal.
Calcd for C14H16BrNO2S: C, 49.13; H, 4.71; Br, 23.35; N,
4.09; S, 9.37. Found: C, 49.07; H, 4.42; Br, 23.51; N, 4.09;
S, 9.37.
3-{6-[1-Methyl-1-(methylsulfonyl)ethyl]quinolin-8-yl}-
benzaldehyde (7). Potassium carbonate (3410 g, 24.7 mol),
Pd/C (5%, 438 g), dimethyl sulfone 5 (2700 g, 96 wt %,
7.90 mol), and 3-formyl boronic acid 6 (1727 g, 11.5 mol)
were combined. The solids mixture was degassed using five
vacuum/nitrogen back-fill cycles. DMF (26.1 L) was added,
followed by water (2.6 L) under slight nitrogen purge. The
reaction mixture was degassed again using five vacuum/
nitrogen back-fill cycles and then heated at ∼80 °C for 2 h.
The waste cake and Pd/C were removed by filtering the hot
mixture (∼60 °C) through a bed of solka floc. The Pd/C
was rinsed with hot DMF (2 × 2.9 L, 60 °C). The filtrate
was heated to 50-60 °C and diluted with water (24 L). The
hazy solution was seeded, and the suspension was aged for
1 h to generate a seed bed. Additional water (8.0 L) was
slowly added. The slurry was cooled to ∼10 °C and aged
for 6-8 h. The mixture was filtered, and the solid was
washed with 1:1 DMF/water (v/v) (4 L) and water (4 L).
After drying by suction (N2 sweep) for 2-3 h, the solid was
dried in a vacuum oven (N2 sweep) until constant weight.
Isolated yield: 2446 g (85% yield corrected for wt %). Mp
182-184 °C. 1H NMR (400 MHz, CDCl3): δ 2.03 (s, 6H),
2.66 (s, 3H), 7.51 (dd, J ) 4.1, 8.2 Hz, 1H), 7.69 (t, J ) 7.6
Hz, 1H), 7.98 (m, 2H), 8.08 (d, J ) 2.1 Hz, 1H), 8.12 (d, J
) 2.2 Hz, 1H), 8.23 (m, 1H), 8.28 (m, 1H), 9.00 (dd, J )
1.8, 4.2 Hz, 1H), 10.13 (s, 1H). 13C NMR (100.6 MHz,
CDCl3): δ 22.9, 35.0, 64.7, 122.0, 128.0, 128.3, 128.9, 128.9,
Vol. 10, No. 1, 2006 / Organic Process Research & Development
•
43

129.9, 132.3, 135.8, 136.4, 136.8, 137.0, 139.7, 140.0, 145.4,
151.6, 192.4. Anal. Calcd for C20H19NO3S: C, 67.97; H,
5.42; N, 3.96; S, 9.07. Found: C, 67.86; H, 5.11; N, 3.92;
S, 8.99.
of molecular sieves for 22 h. After cooling to 20-25 °C
and stirring for 4 h, the slurry was filtered, and the solids
were washed with 2 × 5 L of 2-propanol. Isolated yield:
4.7 kg, 89%.
9: mp 201-203 °C. 1H NMR (400 MHz, CDCl3): δ 1.99
(s, 6H), 2.43 (s, 3H), 2.64 (s, 3H), 3.03 (s, 3H), 7.13 (d, J )
8.0 Hz, 1H), 7.37 (t, J ) 7.7 Hz, 1H), 7.48-7.51 (m, 2H),
7.63-7.66 (m, 3H), 7.91 (d, J ) 2.3 Hz, 1H), 7.96-7.99
(m, 2H), 8.03 (d, J ) 2.3 Hz, 1H), 8.12 (s, 1H), 8.24 (dd, J
) 1.6, 8.3 Hz, 1H), 8.96 (dd, J ) 1.7, 4.1 Hz, 1H). 13C
NMR (100.6 MHz, CDCl3): δ 11.7, 22.8, 34.9, 44.5, 64.6,
121.9, 124.2, 127.5, 128.1 (2 C), 128.5, 129.7, 129.9, 131.2,
132.4, 132.9, 133.3, 135.6, 136.9, 139.5, 139.8, 140.2, 140.4,
140.6, 145.1, 151.4, 167.8, 176.2. Anal. Calcd for
C31H29N3O5S2: C, 63.35; H, 4.97; N, 7.15; S, 10.91.
Found: C, 63.41; H, 4.81; N, 7.02; S, 10.72.
Methylamidoxime (8a). Acetonitrile (25.2 L) was heated
to 70 °C, and hydroxylamine solution (2.5 kg, 37.8 mol, 50
wt % solution) was added slowly over 4 h while the
temperature was maintained at 70 °C. (CAUTION: Heating
hydroxylamine solutions may result in an exothermic de-
composition.) The reaction is exothermic, and the temperature
is maintained by the rate of addition of hydroxylamine
solution. The reaction mixture was stirred for an additional
3 h and then cooled to 0 °C for 2 h and filtered. The isolated
solid is dried in a vacuum oven at ambient (∼25 °C)
temperature. The isolated yield was 1.51 kg (54%).
3-Methyl-5-[4-(methylsulfonyl)benzyl]-1,2,4-oxadiaz-
ole (8). Hydroxybenzotriazole hydrate (2.06 kg, 13.4 mol)
was suspended in acetonitrile (25 L), and 2.5 L was distilled
off at 1 atm under nitrogen to azeotropically remove water.
[CAUTION: Hydroxybenzotriazole will decompose, possibly
Violently, if heated aboVe its melting point (155-160 °C).]32
After cooling to 25-30 °C, 4-methylsulfonylphenylacetic
acid (2.5 kg, 11.67 mol) was added, followed by EDC
hydrochloride (2.68 kg, 14.0 mol). The resulting mixture was
stirred at 20-30 °C for 30 min. Methylamidoxime (1.14 kg,
14.0 mol) was added to the slurry over 10 min. The resulting
mixture was then heated at reflux for 12 h. The solution was
solvent switched to ethyl acetate under reduced pressure
(100-200 mBar, 40-50 °C) by continuous distillation of
∼40 L of ethyl acetate and concentrated to a final volume
of ∼27 L. After cooling to ∼20 °C, aqueous NaHCO3 (1
M, 20 L) was added slowly with vigorous stirring. The
aqueous layer was removed, and the organic layer was
washed with DI water (7.5 L). The aqueous solutions were
back extracted with 17.5 L of ethyl acetate. The combined
ethyl acetate solution was concentrated to ∼7 L (100-200
mBar, 50-70 °C) and diluted with 2-propanol (17.5 L). The
solution was further concentrated to ∼12.5 L. The solution
was allowed to cool to 10-20 °C, and the desired product
precipitated. After stirring for 2 h the mixture was filtered;
the solid was washed with 2 × 2.5 L of 2-propanol and then
air-dried for 2 h. The solid was further dried in an oven at
30-35 °C under vacuum (N2 sweep) to constant weight.
9a: 1H NMR (400 MHz, CDCl3): δ 2.02 (s, 6H), 2.46
(s, 3H), 2.63 (s, 3H), 3.08 (s, 3H), 7.18 (br d, J ) 7.8 Hz,
1H), 7.46-7.51 (o m, 2H), 7.57 (br s, 1H), 7.60-7.62 (o m,
3H), 7.73 (br d, J ) 7.8 Hz, 1H), 7.97-7.99 (m, 2H), 8.03
(d, J ) 2.3 Hz, 1H), 8.09 (d, J ) 2.3 Hz, 1H), 8.26 (dd, J
) 1.7, 8.3 Hz, 1H), 9.00 (dd, J ) 1.8, 4.2 Hz, 1H). 13C
NMR (100.6 MHz, CDCl3): δ 11.8, 22.7 (2C), 34.9, 44.6,
64.6, 121.7, 124.4, 127.6, 127.9 (2 C), 128.0, 128.2 (3C),
128.4, 129.9, 131.6, 132.3, 134.2, 135.7, 136.9, 139.4, 140.0,
140.3, 141.2, 143.5, 145.3, 151.3, 167.8, 174.9.
6-[1-Methyl-1-(methylsulfonyl)ethyl]-8-(3-{(E)-2-(3-
methyl-1,2,4-oxadiazol-5-yl)-2-[4-(methylsulfonyl)phenyl]-
vinyl}phenyl)quinoline benzenesulfonate (10). Free base
9 (996.6 g, 98.44 wt %, 1.0 equiv), 80:20 N,N-dimethylfor-
mamide/isopropyl acetate (v/v) (5.08 L) and benzenesulfonic
acid (324 g, 98 wt %, 1.2 equiv) were combined. The
resulting slurry was warmed to 40 °C for 30 min to afford
a clear yellow solution. The solution was transferred through
a line filter to a second vessel and seeded with 10 (62.1 g).
Isopropyl acetate (2.9 L) was added to the reaction solution
at 40 °C, and the slurry was stirred for 2 h to generate a
seed bed. Additional isopropyl acetate (16.2 L) was added
slowly over 3 h. Once the antisolvent addition was complete,
the reaction slurry was slowly cooled from 40 to 5 °C over
5 h and then stirred for an additional 5 h. The product was
isolated by vacuum filtration and washed with 95:5 isopropyl
acetate/DMF (1 × 5 L slurry wash and 1 × 5 L displacement
wash) and isopropyl acetate (2 × 5 L displacements washes),
followed by tray drying in a vacuum oven at 35 °C. Isolated
product (1136.5 g, 98.2 wt %, 91.3% isolated yield). Mp
1
Isolated yield: 2.6 kg, 88%. Mp 90-91 °C. H NMR (400
MHz, CDCl3): δ 7.90 (m, 2H), 7.52 (m, 2H), 4.27 (s, 2H),
3.02 (s, 3H), 2.35 (s, 3H); 13C NMR (100.6 MHz, CDCl3):
δ 176.3, 167.4, 139.9, 139.6, 129.9, 127.9, 44.4, 32.6, 11.4.
Anal. Calcd for C11H12N2O3S: C, 52.37; H, 4.79; N, 11.10.
Found: C, 52.34; H, 4.69; N, 11.0.
6-[1-Methyl-1-(methylsulfonyl)ethyl]-8-(3-{(E)-2-(3-
methyl-1,2,4-oxadiazol-5-yl)-2-[4-(methylsulfonyl)phenyl]-
vinyl}phenyl)quinoline (9). Aldehyde 7 (3.18 kg, 9.0 mol)
was suspended in 2-propanol (38 L). To the slurry was added
the oxadiazole 8 (2.4 kg, 9.5 mol) followed by piperidine
(1.05 L, 10.6 mol) at 20 °C. The mixture was heated at reflux,
and the distillates were returned to the flask through a bed
1
210 °C. H NMR (400 MHz, CDCl3): δ 1.99 (s, 6H), 2.43
(s, 3H), 2.72 (s, 3H), 3.01(s, 3H), 6.85 (s, 1H), 7.30-7.43
(m, 6H), 7.65-7.70 (m, 4H), 7.88 (d, J ) 8.2 Hz, 2H), 7.98-
8.06 (m, 3H), 8.30 (d, J ) 1.5 Hz, 1H), 8.90 (d, J ) 10.9
Hz, 1H), 9.79 (d, J ) 4.3 Hz, 1H). 13C NMR (100.6 MHz,
CDCl3): δ 176.86, 168.87, 148.79, 147.99, 147.64, 141.93,
141.27, 140.44, 139.62, 136.74, 136.45, 135.89, 135.69,
134.77, 132.99, 132.33, 132.28, 132.13, 130.67, 130.29,
130.25, 129.97, 128.92, 128.77, 126.71, 126.56, 123.78,
65.36, 44.43, 35.73, 22.91, 11.79. Anal. Calcd for
C37H35N3O8S3: C, 59.58; H, 4.73; N, 5.63, S 12.90. Found:
C, 59.51; H, 4.69; N, 5.63, S, 12.87.
(32) The acetonitrile solution of HOBt was not shock sensitive by drop weight
testing.
44
•
Vol. 10, No. 1, 2006 / Organic Process Research & Development

Acknowledgment
DiMichele for providing NMR analysis and Mr. Donald
Bachert, Mr. Thomas Vickery, and Dr. Thientu Lam for
performing the hazard evaluations testing. Finally, we thank
the reviewers for their helpful comments and suggestions.
We thank our colleagues in the Chemical Engineering
Research & Development and Analytical Research depart-
ments for their contributions to the development and
implementation of this chemistry. We thank Mr. Richard
Desmond, Mr. Bruce Foster, Dr. Raymond Cvetovich, and
Dr. Todd Nelson for their contributions to the early-stage
development of this project. We also thank Ms. Lisa
Received for review July 8, 2005.
OP050116L
Vol. 10, No. 1, 2006 / Organic Process Research & Development
•
45