Practical synthesis of a cathepsin S Inhibitor: Route identification, purification strategies, and serendipitous discovery of a crystalline salt form

Article abstract of DOI:10.1021/jo902650b

Chemical Equation Presented A redox economical strategy resulted in a concise, modular synthesis of compound 1, a potent Cathepsin S inhibitor. Starting from three building blocks, crude drug substance was prepared, in a two-step sequence in high yield. Efficient purification of the crude drug substance was accomplished via the formation of an unusual monoethyl oxalate salt.

Full text of DOI:10.1021/jo902650b

pubs.acs.org/joc
Practical Synthesis of a Cathepsin S Inhibitor: Route Identification,
Purification Strategies, and Serendipitous Discovery of a Crystalline
Salt Form
Xiaohu Deng,* Jimmy T. Liang, Matthew Peterson,† Raymond Rynberg, Eugene Cheung,‡
and Neelakandha S. Mani
Johnson & Johnson Pharmaceutical Research & Development LLC, 3210 Merryfield Row, San Diego,
California 92121. †Current address: Amgen Inc., 360 Binney St, Cambridge, MA 02139.
‡TransForm Pharmaceuticals, 29 Hartwell Avenue, Lexington, MA 02421.
xdeng@its.jnj.com
Received December 16, 2009
A “redox economical” strategy resulted in a concise, modular synthesis of compound 1, a potent
Cathepsin S inhibitor. Starting from three building blocks, crude drug substance was prepared in a
two-step sequence in high yield. Efficient purification of the crude drug substance was accomplished
via the formation of an unusual monoethyl oxalate salt.
Introduction
earlier were peptidic compounds that covalently bind to
the cysteine at the active site. In 2004, our laboratories
disclosed the first highly potent nonpeptidic, noncovalent
cathepsin S inhibitor.3 Subsequent medicinal chemistry ef-
forts4 identified compound 1 as a potent inhibitor of cathe-
psin S. For the purpose of further pharmacologically
profiling this compound, a practical, large-scale synthesis
was required.
Cathepsin S is a cysteine protease of the papain super-
family. Unlike the ubiquitous housekeeping enzymes cathe-
psins B, D, and L, cathepsin S is mainly expressed in antigen
presenting cells such as dendritic cells, B cells, and macro-
phages. It is believed that cathepsin S is involved in antigen-
presentation to the cell surface for recognition by the CD4þ
T cells, which in turn triggers an immune response.1 There-
fore, cathepsin S inhibitors have been proposed to be poten-
tial therapeutics for a range of immunological disorders and
have been intensively pursued by many major pharmaceu-
tical companies.2 Most cathepsin S inhibitors discovered
(3) (a) Thurmond, R. L.; Beavers, M. P.; Cai, H.; Meduna, S. P.; Gustin,
D. J.; Sun, S.; Almond, H. J.; Karlsson, L.; Edwards, J. P. J. Med. Chem.
2004, 47, 4799–4801. (b) Thurmond, R. L.; Sun, S.; Sehon, C. A.; Baker,
S. M.; Cai, H.; Gu, Y.; Jiang, W.; Riley, J. P.; Williams, K. N.; Edwards, J. P.;
Karlsson, L. J. Pharm. Exp. Ther. 2004, 308, 268–276.
(4) (a) Gustin, D. J.; Sehon, C. A.; Wei, J.; Cai, H.; Meduna, S. P.;
Khatuya, H.; Sun, S.; Gu, Y.; Jiang, W.; Thurmond, R. L.; Karlsson, L.;
Edwards, J. P. Bioorg. Med. Chem. Lett. 2005, 15, 1687–1691. (b) Grice,
C. A.; Tays, K.; Khatuya, H.; Gustin, D. J.; Butler, C. R.; Wei, J.; Sehon,
C. A.; Sun, S.; Gu, Y.; Jiang, W.; Thurmond, R. L.; Karlsson, L.; Edwards,
J. P. Bioorg. Med. Chem. Lett. 2006, 16, 2209–2212. (c) Wei, J.; Pio, B. A.;
Cai, H.; Meduna, S. P.; Sun, S.; Gu, Y.; Jiang, W.; Thurmond, R. L.;
Karlsson, L.; Edwards, J. P. Bioorg. Med. Chem. Lett. 2007, 17, 5525–5528.
(d) Ameriks, M. K.; Axe, F. U.; Bembenek, S. D.; Edwards, J. P.; Gu, Y.;
Karlsson, L.; Randal, M.; Sun, S.; Thurmond, R. L.; Zhu, J. Bioorg. Med.
Chem. Lett. 2009, 19, 6131–6134. (e) Ameriks, M. K.; Cai, H.; Edwards, J. P.;
Gebauer, D.; Gleason, E.; Gu, Y.; Karlsson, L.; Nguyen, S.; Sun, S.;
Thurmond, R. L.; Zhu, J. Bioorg. Med. Chem. Lett. 2009, 19, 6135–6139.
(1) (a) Villandangos, J. A.; Bryant, R. A. R.; Deussing, J.; Driessen, C.;
Lennon-Dumenil, A-M; Riese, R. J.; Roth, W.; Saftig, P.; Shi, G-P;
Chapman, H. A.; Peters, C.; Ploegh, H. L. Immunol. Rev. 1999, 172, 109–120.
(b) Nakagawa, T. Y.; Rudensky, A. Y. Immunol. Rev. 1999, 172, 121–129.
(c) Hsing, L. C.; Rudensky, A. Y. Immunol. Rev. 2005, 207, 229–241.
(2) (a) Kim, W.; Kang, K. Expert Opin. Ther. Pat. 2002, 12, 419–432.
(b) Leroy, V.; Thruairatnam, S. Expert Opin. Ther. Pat. 2004, 14, 301–311.
(c) Thrumond, R. L.; Sun, S.; Karlsson, L.; Edwards, J. P. Curr. Opin. Invest.
Drugs 2005, 6, 473–482. (d) Link, J. O.; Zipfel, S. Curr. Opin. Drug Discovery
Dev. 2006, 9, 471–482. (e) Gupta, S.; Kumar, S. R.; Dastidar, S.; Ray, A.
Expert Opin. Ther. Targets 2008, 12, 291–299.
1940 J. Org. Chem. 2010, 75, 1940–1947
Published on Web 02/15/2010
DOI: 10.1021/jo902650b
r
2010 American Chemical Society

Deng et al.
JOCArticle
SCHEME 1. Original Synthesis of Compound 1
SCHEME 2. Modular “Redox Economical” Approach
The original synthesis of compound 1 involved a 7-step
sequence (Scheme 1) from compound 2, which had been
prepared previously in a three-step, one-pot process (see
Experimental Section for details). Because this synthesis of 1
was originally tailored to facilitate analogue production, it did
not capitalize on opportunities for convergence. We hoped to
address this issue in our revised route. In addition, from
economical and environmental points of view, we planned
to avoid expensive, environment-unfriendly reagents such as
the Dess-Martin reagent and HATU. Finally, we deemed it
critical to minimize unit operations and eliminate column
chromatographic purification; because of its high molecular
weight (699 Da) and polarity, purification of the final drug
substance 1 was particularly challenging. Herein, we report
our endeavor toward the efficient synthesis of drug substance
1, which features a modular, two-step route from three readily
available building blocks. The purification of the final drug
substance was accomplished through the formation of an
unusual monoethyl oxalate salt. Overall, this synthesis is
operationally simple, high-yielding, chromatography-free,
and amenable for large-scale production.
no intermediary refunctionalizations ...”.5 To approach that,
the practitioners of industrial process research must strive to
minimize oxidation state manipulations. This concept has
been further delineated recently by Baran and Hoffmann,
who coined the term “redox economy”.6 In the original
synthesis of 1, two chemical steps were devoted to counter-
productive oxidation state adjustments with expensive
NaBH(OAc)3 and the Dess-Martin reagent. In addition,
the sequence was performed in a linear fashion starting from
the rather expensive advanced intermediate 2. Strategically,
a modular approach from optimal building blocks would be
more efficient and economical. To that end, we devised a new
retro-synthetic analysis (Scheme 2).
To access the fully elaborated left-hand building block 3,
we chose the readily available and inexpensive 1-bromo-3-
chloropropane as the starting material with the hope that the
reactivity difference of bromide and chloride would provide
the requisite chemoselectivity (Scheme 3). The oxidation
state and functionality of this building block were keys to
the potential for the realization of redox economy. Unfortu-
nately, a number of literature procedures7 afforded a mixture
(5) Hendrickson, J. B. J. Am. Chem. Soc. 1975, 97, 5784–5800.
(6) Burns, N. Z.; Baran, P. S.; Hoffmann, R. W. Angew. Chem., Int. Ed.
2009, 48, 2854–2867.
(7) (a) Murty, M. S. R.; Jyothirmai, B.; Krishna, P. R.; Yadav, J. S. Synth.
Commun. 2003, 33, 2483–2486. (b) Wang, J-Q; Gao, M.; Miller, K. D.;
Sledge, G. W.; Zheng, Q-H. Bioorg. Med. Chem. Lett. 2006, 16, 4102–4106.
Results and Discussion
Hendrickson has formalized the concept of an “ideal
synthesis” as one that “creates a complex skeleton ... in a
sequence only of successive construction reactions involving
J. Org. Chem. Vol. 75, No. 6, 2010 1941

Deng et al.
JOCArticle
SCHEME 3. Synthesis of Compound 3
SCHEME 5. Synthesis of Compound 9
SCHEME 4. Alkylation of Compound 2 with 3
SCHEME 6. Synthesis of Compound 10
of 3 and dialkylation product 5 even in the presence of a large
excess of 1-bromo-3-chloropropane. The low selectivity was
probably due to the high reaction temperature required for
the alkylation of sterically hindered 2-methylmorpholine. To
address this problem, we treated 2-methylmorpholine with
NaH to effect complete deprotonation. Subsequent addition
of equimolar 1-bromo-2-chloropropane and warming to
65 °C gave rise to the desired product 3 with effective
suppression of the dialkylation pathway. A 100-g scale
reaction was performed to provide 3 in 74% yield, and no
column chromatographic purification was needed.
The alkylation reaction of compound 2 with 3 was then
investigated. It is well-known in the literature,8 and from our
own experience, that N1:N2 regioselectivity is always an issue
upon alkylation of 1-H-pyrazoles. Indeed, on a 0.5-g scale
test run, the alkylation reaction provided a mixture of 6 and 7
in a 9:1 ratio and 75% yield (Scheme 4). A similar regioiso-
meric ratio was observed in the reaction between compound
2 and 3-bromo-propanol in the original synthesis. Our past
experience with the regioisomeric pyrazole products9 has
taught us that recrystallization often provides an effective
purification strategy. Unfortunately, in this case, efforts to
generate crystalline solids of either regioisomer 6 or 7 were
unsuccessful. More problematically, on a 5-g scale, the
reaction stalled at ∼50% conversion. Addition of excess
reagents (3 and base) and increase of reaction temperature
failed to drive the reaction to completion. The cause of this
scale-up issue is not fully understood.
also allow selective removal of the undesired N2-alkylated
byproduct. Therefore, we decided to prepare 9, which was
easily accomplished in two steps (Scheme 5). The deprotec-
tion of the Boc group of compound 2 proceeded uneventfully
to afford 8 in quantitative yield. The isolation of 8 involved
the simple addition of water followed by neutralization of
TFA with NaOH. In the oxalamide formation step, an
inexpensive coupling reagent, CDI, was used to replace the
expensive HATU/HOAT used in the original synthesis. A
water-miscible solvent, DMF, was purposely chosen, and
compound 9 was directly precipitated in 95% yield by simply
adding water to the reaction solution.
With 9 in hand, the alkylation reaction with 3 was in-
vestigated. Two regioisomers were again observed in a ∼9:1
ratio favoring desired product 10 (Scheme 6). Gratifyingly,
the desired regioisomer 10 was preferentially precipitated by
addition of water to the DMF reaction solution. Subsequent
trituration from hot EtOH provided pure, highly crystalline
material in 70% yield. Thus, the difficult regioselectivity
problem was solved with a simple purification strategy. It is
noteworthy that no scale-up issues were observed in this case
(up to 250 g scale).
To circumvent these problems, we turned our attention to
an alternative intermediate 9. It was noticed that molecules
with the oxalamide group tended to be highly crystalline
solids. Besides the obvious operational advantages of solid
intermediates in scale-up campaigns,10 crystallization might
Before the crucial Sonogashira cross-coupling reaction
could be performed, a suitable large-scale synthesis of build-
ing block 4 was needed. Under various direct reductive
amination conditions on 4-ethynyl-benzaldehyde with NaBH-
(OAc)3, the desired product 4 was invariably contaminated
with dialkylation product 11 (20-40%) (Scheme 7). The
alternative stepwise reductive amination conditions (MeOH/
NaBH4), which usually result in enhanced selectivity,11 also
gave rise to significant quantities of 11. Interestingly,
upon TMS protection of the starting material, the two-step
reductive amination procedure provided 4 exclusively, probably
(8) (a) Meanwell, N. A.; Rosenfeld, M. J.; Wright, J. J. K.; Brassard,
C. L.; Buchanan, J. O.; Federici, M. E.; Fleming, J. S.; Seiler, S. M. J. Med.
Chem. 1992, 35, 389–397. (b) Wang, X.; Tan, J.; Zhang, L. Org. Lett. 2000, 2,
3107–3109.
(9) (a) Deng, X.; Mani, N. S. Org. Lett. 2006, 8, 3505–3508. (b) Liang,
J. T.; Mani, N. S.; Jones, T. K. J. Org. Chem. 2007, 72, 8243–8350. (c) Deng,
X.; Mani, N. S. J. Org. Chem. 2008, 73, 2412–2415. (d) Deng, X.; Mani, N. S.
Org. Lett. 2008, 10, 1307–1310.
(10) Anderson N. G. Practical Process Research & Development; Aca-
demic Press: New York, 2000.
(11) Abdel-Magid, A. F.; Mehrman, S. J. Org. Process. Res. Dev. 2006,
10, 971–1031.
1942 J. Org. Chem. Vol. 75, No. 6, 2010

Deng et al.
JOCArticle
SCHEME 7. Synthesis of Compound 4
SCHEME 8. Synthesis of Compound 1
SCHEME 9. Production of Pure, Solvent-Free 1, L-Tartrate
Salt
due to the increase in the rate of the imine reduction step.12 It is
worthwhile to note that removal of the TMS protecting group
was achieved in situ upon the addition of NaBH4, which made
the sequence a highly efficient one-pot operation. Compound 4
was isolated in 90% yield, and no column chromatographic
purification was needed.
With building blocks 4 and 10 in hand, the optimization of
the Sonogashira cross-coupling reaction was studied. In the
active pharmaceutical ingredient (API) production process,
heavy metal contamination of the final drug substance is
always a major concern. We quickly identified that the
unusually high loading of Pd/Cu catalysts required in the
original synthesis was the result of poor solubility of com-
pound 10 in THF. Switching the solvent to DMF, the
catalyst loading was successfully lowered to 0.25% Pd /
0.5% CuI (Scheme 8). At a slightly elevated temperature
(50 °C), the Sonogashira reaction was complete in 16 h.
Again, because of the use of water-miscible solvent DMF,
crude 1 was precipitated by the addition of water in 85%
yield and ∼89% HPLC assay purity.
under high vacuum (0.1 Torr) at 70 °C for several days.
Trituration in other solvents such as EtOH resulted in the
trapping of the incoming solvent. Finally, azeotropic dis-
tillation of an aqueous solution of the L-tartrate salt success-
fully lowered the organic solvent residue level to <0.5 wt %.14
Subsequent lyophilization provided the final API as an
amorphous white powder. This protocol was used to pro-
duce over 100 g of material for various animal studies
(Scheme 9).
Unfortunately, this approach is not suitable for larger-
scale production. For purification of API on large scale,
recrystallization of free base or a salt remains the most
practical option. Attempts at recrystallization or trituration
of free base 1 with numerous combinations of common
solvents (MeOH, EtOH, IPA, t-amylOH, EtOAc, CH3CN,
THF, 2-MeTHF, CH2Cl2, CHCl3, hexanes, heptane toluene,
acetone, MTBE, DMF, iPrOAc, 2-butanone, nitromethane,
DME) failed to provide any crystalline material. Further-
more, improvements to the purity of 1 were minimal. This
result did not come as a total surprise, because understand-
ably, the two flexible side chains present in compound 1
would impede the formation of a rigid crystalline lattice. In
the pharmaceutical industry, a common strategy to deal with
amorphous drug substances is to form a salt, which may
possess higher crystallinity, a desirable characteristic for
drug development. Therefore, significant efforts were under-
taken toward salt screening with both automated equipment
and manual optimization. Over 30 pharmaceutically accep-
table, structurally diverse acids15 were investigated in vari-
ous common solvents (Figure 1). In addition to the pre-
viously identified tartaric acid, only three other acids
To finish the production of the final API, we required a
purification of crude 1. This seemingly trivial task proved to
be particularly challenging. Because of time constraints, we
resorted to column chromatography to produce an early
batch of the API (Scheme 9). Poor product recovery (∼60%)
was observed probably due to the high polarity and instabi-
lity of the API on normal phase silica gel. As a result of its
high water solubility, the L-tartrate salt of 113 was identified
as the potential salt form early. However, even under opti-
mized salt formation conditions in a mixed EtOAc/MeOH
solvent at 60 °C, the L-tartrate salt obtained was amorphous
and hygroscopic and contained about 8 wt % solvent
residues, which were extremely difficult to remove even
(12) Abdel-Magid, A. F.; Harris, B. D.; Maryanoff, C. A. Synlett 1994,
81–83.
(14) Direct lyophilization was not effective removing the organic solvent
residues.
(13) There was no different to use L- or D-tartaric acid. Naturally
occurring L-tartaric acid was chosen for the cost reason.
(15) Pharmaceutical Salts: Properties, Selection, and Use; Stahl, P. H.,
Wermuth, C. G., Eds.; Wiley-VCH: New York, 2008.
J. Org. Chem. Vol. 75, No. 6, 2010 1943

Deng et al.
JOCArticle
FIGURE 1. Common pharmaceutically accepted acids.
SCHEME 10. Formation of a Crystalline Solid
(H3PO4, oxalic acid, and 5-oxo-pyrrolidine-2-carboxylic
acid) provided solid salts. Unfortunately, all four salts were
amorphous and failed to reject the impurities.
Finally, a serendipitous breakthrough occurred during the
study of the solubility of 1 in diethyl oxalate. This unortho-
dox “solvent” was purposefully chosen due to its structural
similarity to the oxalamide side chain of 1. When compound
1 was dissolved in diethyl oxalate, a white solid 12 precipi-
tated almost immediately (Scheme 10). Optical microscope
study of 12 (Figure 2) suggested a highly crystalline material;
this crystallinity was confirmed by powder XRD analysis
(Figure 3). More significantly, upon formation of 12, the
purity of the drug substance increased from 89% to 96%
based on HPLC analysis.
FIGURE 2. Microscope photo of crystal 12.
Careful NMR studies revealed that crystalline material 12
was actually the monoethyl oxalate salt of 1, which was
unambiguously confirmed by single crystal X-ray crystal-
lography (Figure 4). Apparently, diethyl oxalate was rapidly
hydrolyzed under catalysis by 1 to generate monoethyl
oxalate,16 which was acidic enough to complex with com-
pound 1 to form the salt 12. Control experiments showed
that compound 1 in >99.5% purity also afforded the same
salt 12, which ruled out the possibility of the impurities
serving as the catalysts for hydrolysis. Interestingly, the same
hydrolysis and salt formation was observed with
(PhCH2)2NH but NOT with intermediate 9, suggesting that
the secondary amine portion of 1 was responsible for the
catalytic hydrolysis process. Extensive literature search re-
vealed a single report of this unusual amine mediated
hydrolysis of oxalate with tert-butylamine.17 To confirm this
theory, compound 1 was treated with preformed monoethyl
oxalate, and the identical crystalline salt 12 was obtained.
Careful examination of the single crystal X-ray structural
data suggested that both carbonyl groups and the length of
FIGURE 3. Powder X-ray diffraction pattern of crystal 12.
its alkyl side chain were essential for the construction of the
crystal lattice. To test the hypothesis, a number of 2-oxo-
carboxylic acids (Figure 5) were investigated for the salt
formation process. Indeed, other than monoethyl oxalate,
only 2-oxo-pentanoic acid fits the structural requirements to
provide a crystalline salt.
Encouraged by these results, the salt formation process
was carefully optimized. It was found that in CH3CN at
50 °C, addition of monoethyl oxalate to the crude 1 (89%
purity) followed by slow cooling to room temperature
produced the crystalline salt 12 in >98.5% purity. The salt
(16) Presumably with water present in the solvent or in the air.
(17) Petyunin, G. P. Farm. Zh. (Kiev) 1980, 3, 40–42.
1944 J. Org. Chem. Vol. 75, No. 6, 2010

Deng et al.
JOCArticle
FIGURE 4. Single crystal X-ray structure of 12.
SCHEME 11. Synthesis of Compound 2
FIGURE 5. 2-Oxo-carboxylic acids.
12 could be directly used as the final API or easily converted
to other suitable forms for future development.
Conclusion
In summary, we devised a modular, redox economical
synthesis of 1 from three readily available building blocks.
This two-step synthesis is high-yielding, operationally sim-
ple, and amenable for large-scale manufacture. An unusual
amine-mediated diethyloxalate hydrolysis reaction was ob-
served. On the basis of these findings, salt formation with
monoethyl oxalate provided an efficient purification method
for the final drug substance.
The enamine prepared above was dissolved in CH2Cl2 (1.6 L),
and then Et3N (209 mL, 1.5 mol, 1.5 equiv) was added. At 0 °C
under N2, a solution of 4-chloro-3-iodo-benzoyl chloride (290 g,
1.0 mol, 1.0 equiv) in CH2Cl2 (400 mL) was added over 30 min.
The ice bath was then removed, and the reaction solution was
stirred at room temperature for 3 h. All volatile solvents were
removed under vacuum, and the residue was redissolved in
EtOH (1.5 L). At 0 °C, anhydrous NH2NH2 (47 mL, 1.5 mol,
1.5 equiv) was added over 30 min. (exothermic reaction). The
reaction solution was stirred at room temperature for 16 h. The
precipitated white solid was collected by filtration and washed
with cold EtOH to afford the pure compound 2 as a white solid
(333 g, 0.73 mol, >95% purity, 73%). The mother liquor was
concentrated and was partitioned between CH2Cl2/H2O. The
organic layer was washed with water, dried over Na2SO4, and
concentrated. The crude product was recrystallized from hot
CH3CN to give another crop of the pure material (74 g, 0.16 mmol,
16%). The combined yield was 89% over three steps: mp 173
-175 °C. 1H NMR (600 MHz, DMSO) δ 8.16 (d, J = 1.8 Hz,
1H), 7.65 (d, J = 8.3 Hz, 1H), 7.58 (s, 1H), 4.56 (s, 2H), 3.63 (t,
J = 5.8 Hz, 2H), 2.70 (t, J = 5.8 Hz, 2H), 1.42 (s, 9H). 13C NMR
(151 MHz, DMSO) δ 154.0, 136.6, 135.9, 129.6, 126.9, 110.1,
99.4, 79.1, 63.1, 42.5, 40.5, 27.9. (Due to the tautomeric nature of
Experimental Section
3-(4-Chloro-3-iodo-phenyl)-1,4,6,7-tetrahydro-pyrazolo[4,3-c]-
pyridine-5-carboxylic Acid tert-Butyl Ester (2). Compound 2 was
synthesized via a 3-step sequence in a one-pot fashion (Scheme 11).
In a 3-L round-bottom flask equipped with a Dean-Stark
trap, a reflux condenser, and an internal thermocouple were
added sequentially N-Boc-piperidone (200 g, 1.0 mol, 1.0 equiv),
toluene (2 L), morpholine (92 mL, 1.05 mol, 1.05 equiv), and
p-toluenesulfonic acid (1.0 g, 0.005 mmol, 0.5% equiv). The
reaction solution was refluxed under N2 for 16 h (about 18 mL of
water was collected). The solvent was evaporated to afford 4-
morpholin-4-yl-3,6-dihydro-2H-pyridine-1-carboxylic acid tert-
butyl ester, which was used directly in the next reaction
(colorless oil, ∼270 g, 100%).
J. Org. Chem. Vol. 75, No. 6, 2010 1945

Deng et al.
JOCArticle
this compound, a few carbon signals were undetectable under
then extracted with EtOAc (3 ꢀ 15 mL). The organic layers were
combined and extracted with 1 mol/L HCl (40 mL).18 The
aqueous layer that contained compound 3 was cooled in an ice
bath and basified with slow addition of NaOH pellets to pH
9-10, which was then extracted with EtOAc (3 ꢀ 15 mL). The
combined organic layer was dried with Na2SO4 and concen-
trated to afford the title compound as a colorless oil (1.3 g,
7.4 mmol, 74%): 1H NMR (500 MHz, CDCl3) δ 3.78 (dtd, J =
11.3, 3.1, 0.9 Hz, 1H), 3.70-3.62 (m, 2H), 3.60 (t, J = 6.4 Hz,
2H), 3.23 (dd, J = 11.1, 9.0 Hz, 1H), 2.90 (dt, J = 13.0, 7.6 Hz,
1H), 2.72 (dt, J = 11.7, 2.8 Hz, 1H), 2.51-2.37 (m, 1H), 2.36
standard 13C NMR conditions even at
a 50 mg/mL
concentration.) IR (neat, cm-1): 3239 (m), 2978 (w), 1650 (s),
1428 (s). HRMS-ESI (m/z): [M þ H]þ calcd for C17H19 N3O2ClI
460.0283, found 460.0265.
3-(4-Chloro-3-iodo-phenyl)-4,5,6,7-tetrahydro-1H-pyrazolo-
[4,3-c]pyridine (8). In a 3-L round-bottom flask equipped with
mechanical stirring and an internal thermocouple, compound 2
(260 g, 0.57 mol, 1.0 equiv) was suspended in CH2Cl2 (750 mL).
TFA (250 mL) was added at room temperature over 20 min to
form a clear solution. The reaction mixture was stirred at room
temperature for 16 h. A lot of the TFA salt of compound 8
precipitated as a white solid. Water (2 L) was added, and then
saturated NaOH aqueous solution was added to adjust to pH >
12. The mixture was stirred for 3 h. The white solid precipitated
was collected by filtration, washed with water and dried in a
vacuum oven to provide 8 as a white solid (205 g, 0.57 mol,
100%). This material was used in the next reaction without
further purification. The compound was characterized as the
2.26 (m, 2H), 1.99-1.83 (m, 2H), 0.98 (d, J = 6.3 Hz, 3H); 13
C
NMR (126 MHz, CDCl3) δ 73.0, 67.4, 55.2, 51.3, 50.5, 43.1,
29.3, 14.1. HRMS-ESI (m/z): [M þ H]þ calcd for C8H17ClNO
178.0993, found 178.0992.
2-{3-(4-Chloro-3-iodo-phenyl)-1-[3-(3-methyl-morpholin-4-yl)-
propyl]-1,4,6,7-tetrahydro-pyrazolo[4,3-c]pyridin-5-yl}-2-oxo-
acetamide (10). In a 5-L flask equipped with mechanical stirring
and an internal thermocouple, compound 9 (190 g, 0.44 mol, 1.0
equiv) was dissolved in DMF (1.8 L). Cs2CO3 (180 g, 0.55 mol,
1.25 equiv) and compound 3 (86 g, 0.48 mol, 1.1 equiv) were
added sequentially. The reaction mixture was then stirred at
50 °C under N2 for 4 h and cooled to room temperature. Water
(3.5 L) was added slowly over 1 h, and the suspension was stirred
at room temperature overnight. The white solid precipitated was
collected by filtration, washed with water, and dried in a vacuum
oven. The crude compound was triturated from hot EtOH (∼1.5 L)
to provide pure 10 as a white solid (176 g, 0.31 mol, 70%):19 mp
189 -190 °C. (Two sets of NMR peaks were observed due to the
tautomeric nature of this compound.) 1H NMR (600 MHz,
DMSO) δ 8.19 (d, J = 2.0 Hz, 0.7H), 8.14 (s, 0.7H), 8.13 (d, J =
2.1 Hz, 0.3H), 8.08 (s, 0.3H), 7.77 (s, 0.7H), 7.72 (s, 0.3H), 7.65
(s, 0.3H), 7.64 (s, 0.7H), 7.59 (dd, J = 8.4, 2.0 Hz, 0.7H), 7.48
(dd, J = 8.3, 2.1 Hz, 0.3H), 4.71 (s, 0.6H), 4.68 (s, 1.4H),
4.12-3.97 (m, 2H), 3.89-3.77 (m, 0.7H), 3.78-3.70 (m, 1.3H),
3.78-3.62 (m, 1H), 3.56 (dt, J = 10.9, 3.0 Hz, 1H), 3.47 (dt, J =
14.9, 2.8 Hz, 1H), 3.14-3.05 (m, 1H), 2.96-2.74 (m, 2H),
2.74-2.63 (m, 2H), 2.34-2.23 (m, 1H), 2.12 (ddd, J = 9.4,
5.7, 3.1 Hz, 2H), 1.92 (dq, J = 13.9, 7.1 Hz, 2H), 0.84 (d, J =
6.3 Hz, 3H); 13C NMR (151 MHz, DMSO) δ 165.5, 165.1, 164.3,
163.7, 142.0, 141.7, 138.3, 137.9, 136.7, 136.6, 135.9, 135.8,
133.7, 133.6, 129.6, 126.9, 126.8, 110.2, 109.7, 99.35, 99.27,
72.0, 72.0, 66.5, 54.6, 50.2, 49.6, 42.6, 42.3, 38.0, 37.7, 25.8,
25.7, 22.1, 20.9, 13.3. IR (neat, cm-1): 3328 (w), 3061 (w), 1694
(s), 1669 (s), 1637 (s), 1433 (s). HRMS-ESI (m/z): [M þ H]þ calcd
for C22H27N5O3ClI 572.0920, found 572.0918.
1
TFA salt: mp 180 -182 °C. H NMR (600 MHz, DMSO) δ
12.77 (s, 1H), 9.32 (s, 2H), 8.16 (d, J = 2.1 Hz, 1H), 7.66 (d, J =
8.3 Hz, 1H), 7.59 (dd, J = 8.3, 2.1 Hz, 1H), 4.40 (s, 2H), 3.43 (d,
J = 5.3 Hz, 2H), 2.97 (t, J = 6.1 Hz, 2H). 13C NMR (151 MHz,
DMSO) δ 158.5 (q, JC-F = 35.4 Hz, 1C), 140.9, 138.7, 136.9,
136.5, 132.0, 129.6, 127.3, 115.9 (q, JC-F = 293.0 Hz, 1C), 106.1,
99.5, 40.5, 40.0, 19.0. IR (neat, cm-1): 3231 (w), 3067 (w), 1754
(w), 1659 (s), 1437 (s). HRMS-ESI (m/z): [M þ H]þ calcd for
C12H11N3ClI 359.9759, found 359.9751.
2-[3-(4-Chloro-3-iodo-phenyl)-1,4,6,7-tetrahydro-pyrazolo-
[4,3-c]pyridin-5-yl]-2-oxo-acetamide (9). In a 5-L flask equipped
with mechanical stirring, CDI (110.5 g, 0.68 mol, 1.2 equiv) was
suspended in DMF (1.5 L) and then oxalamic acid (60.7 g,
0.68 mol, 1.2 equiv) was added. After stirring at room tempera-
ture for 3 h, a nearly clear solution was formed. Compound 8
(205 g, 0.57 mol, 1.0 equiv) was added as a solid over 10 min. The
reaction was complete in 30 min. Water (2.5 L) was added slowly
over 2 h and stirred at room temperature overnight. The
precipitated white solid was collected by filtration, washed with
water and dried in a vacuum oven to provide 9 (231 g, 0.53 mol,
95%). This material was used in the next reaction without
further purification: mp 235-237 °C. Two sets of NMR peaks
were observed due to the tautomeric nature of this compound:
1H NMR (600 MHz, DMSO) δ 13.02 (s, 1H), 8.20 (d, J = 2.0
Hz, 0.7H), 8.15 (s, 0.7H), 8.13 (dd, J = 2.0 Hz, 0.3H), 8.11 (s,
0.3H), 7.77 (s, 0.7H), 7.72 (s, 0.3H), 7.67 (s, 0.3H), 7.66 (s, 0.7H),
7.61 (dd, J = 8.4, 2.0 Hz, 0.7H), 7.50 (dd, J = 8.3, 2.0 Hz, 0.3H),
4.72 (s, 0.6H), 4.70 (s, 1.4H), 3.81 (t, J = 5.9 Hz, 0.6H), 3.73 (t,
J = 5.7 Hz, 1.4H), 2.82 (t, J = 5.6 Hz, 1H), 2.75 (dd, J = 10.5,
4.6 Hz, 1H); 13C NMR (151 MHz, DMSO) δ 165.6, 165.3, 164.5,
163.8, 136.7, 136.6, 136.1, 136.0, 129.6, 127.0, 126.9, 109.6,
109.2, 99.45, 99.37, 42.69, 42.62, 22.5, 21.4. (Due to the tauto-
meric nature of this compound, a few carbon signals were
undetectable under standard 13C NMR conditions even at a
50 mg/mL concentration.) IR (neat, cm-1): 3333 (w), 3177 (w),
1671 (s), 1651 (s). HRMS-ESI (m/z): [M þ H]þ calcd for
C14H12N4O2ClI 430.9766, found 430.9752.
(4-Chloro-benzyl)-(4-ethynyl-benzyl)-amine (4). 4-Trimethyl-
silanylethynyl-benzaldehyde (10 g, 49 mmol, 1.0 equiv) was
dissolved in MeOH (200 mL), and 4-chlorobenzylamine (7.4 g,
52 mmol, 1.05 equiv) was added. After stirring at room tem-
perature overnight, NaBH4 (1.9 g, 49 mmol, 1.0 equiv) was
added slowly. After stirring at room temperature for 1 h, 1 mol/
L HCl aqueous solution was slowly added to quench the
unreacted NaBH4. MeOH was evaporated, and the residue
was partitioned between CH2Cl2 and saturated NaHCO3 aqu-
eous solution. The organic layer was separated, dried over
MgSO4 and concentrated. The crude compound was passed
through a short pad of silica gel to remove the baseline impu-
rities with EtOAc/hexanes as eluent to provide 4 as a light yellow
4-(3-Chloro-propyl)-3-methyl-morpholine (3). In a 25-mL
round-bottom flask with a magnetic stir bar, 3-methyl-morpho-
line (1 g, 9.9 mmol, 1.0 equiv) and 1-bromo-3-chloro-propane
(3.1 g, 19.8 mmol, 2.0 equiv) were dissolved in THF (5 mL). NaH
(0.79 g, 60 wt % in mineral oil, 2.0 equiv) was added in 2
portions. Mild bubbling occurred during the addition. The
reaction slurry was heated to 65 °C for 18 h. GC/MS analysis
indicated the completion of the reaction. The reaction solution
was carefully quenched with addition of ice-water (40 mL) and
1
oil (11 g, 43 mmol, 90%): Rf = 0.54 (1:1 EtOAc/hexanes). H
NMR (600 MHz, CDCl3) δ 7.47-7.44 (m, 2H), 7.32-7.26 (m,
6H), 4.70 (s, 0H), 3.78 (s, 2H), 3.75 (s, 2H), 3.06 (s, 1H); 13C
NMR (151 MHz, CDCl3) δ 141.1, 138.6, 132.7, 132.2, 129.5,
128.5, 128.0, 120.7, 83.6, 77.0, 52.7, 52.3. IR (neat, cm-1): 3294
(m), 2911 (w), 2827 (w), 2106 (w), 1487 (s). HRMS-ESI (m/z): [M þ
H]þ calcd for C16H14NCl 256.0888, found 256.0871.
(18) The small amount of unreacted 3-methyl-morpholine remained in
the organic layer.
(19) The undesired regioisomer was completely rejected in this process.
1946 J. Org. Chem. Vol. 75, No. 6, 2010

Deng et al.
JOCArticle
2-{3-(4-Chloro-3-{4-[(4-chloro-benzylamino)-methyl]-phenyl-
ethynyl}-phenyl)-1-[3-(3-methyl-morpholin-4-yl)-propyl]-1,4,6,7-
tetrahydro-pyrazolo[4,3-c]pyridin-5-yl}-2-oxo-acetamide (1, free
base). In a 5-L flask equipped with mechanical stirring and an
internal thermocouple were added compound 10 (120 g,
0.21 mol, 1.0 equiv), compound 4 (56.3 g, 0.22 mol, 1.0 equiv),
DMF (1.5 L), and Et3N (120 mL, 0.84 mol, 4.0 equiv) sequen-
tially. A stream of N2 was bubbled into the solution for 15 min.
A mixture of Pd(PPh3)2Cl2 (0.37 g, 0.5 mmol, 0.0025 equiv) and
CuI (0.2 g, 1.0 mmol, 0.005 equiv) was added as a solid under N2.
The solution was degassed with N2 for another 10 min. The
reaction solution was stirred at 50 °C overnight. The reaction
solution was cooled to room temperature, and water (2 L) was
added upon stirring. The liquid layer was decanted, and the oily
precipitate was partitioned between EtOAc (2 L) and water (1 L)
and saturated NaHCO3 (500 mL). The organic layer was
separated, dried over MgSO4, and concentrated to provide the
crude material as a foamy yellow solid (145 g, ∼85% purity
based on HPLC analysis). The crude material was purified on
silica gel with 2 mol/L NH3 in MeOH/CH2Cl2 as eluent to
provide the drug substance in >98% purity:20 mp 61-64 °C.
Two sets of NMR peaks due to the tautomeric nature of this
compound: 1H NMR (600 MHz, DMSO) δ 8.17 (s, 0.7H), 8.12
(s, 0.3H), 7.90 (d, J = 1.9 Hz, 0.7H), 7.84 (d, J = 2.1 Hz, 0.3H),
7.81 (s, 0.7H), 7.75 (s, 0.3H), 7.67 (s, 0.3H), 7.65 (s, 0.7H), 7.63
(dd, J = 8.5, 2.0 Hz, 0.7H), 7.58 -7.54 (m, 2H), 7.52 (dd, J =
8.5, 2.1 Hz, 0.3H), 7.43 (d, J = 8.1 Hz, 2H), 7.40 -7.32 (s, 4H),
4.78 (s, 0.6H), 4.74 (s, 1.4H), 4.12-3.99 (m, 2H), 3.91-3.75 (m,
2H), 3.71 (s, 2H), 3.68-3.62 (m, 3H), 3.60-3.53 (m, 1H),
3.52-3.44 (m, 1H), 3.11 (dd, J = 10.8, 8.8 Hz, 1H), 2.87 (ddd,
J = 33.1, 18.1, 10.4 Hz, 3H), 2.74-2.64 (m, 2H), 2.29 (dd, J =
10.0, 4.3 Hz, 1H), 2.19-2.06 (m, 2H), 1.94 (dt, J = 13.6, 6.8 Hz,
2H), 0.84 (d, J = 6.3 Hz, 3H); 13C NMR (151 MHz, DMSO) δ
165.5, 165.2, 164.3, 163.7, 142.7, 142.43, 142.31, 142.28, 139.7,
138.2, 137.9, 133.17, 133.09, 132.63, 132.60, 131.3, 131.0, 129.82,
129.78, 129.73, 129.6, 128.2, 128.0, 127.15, 127.05, 122.50,
122.46, 119.78, 119.75, 110.2, 109.7, 94.66, 94.64, 85.44, 85.39,
72.07, 72.04, 66.6, 54.6, 51.7, 51.3, 50.2, 49.7, 46.3, 42.8, 42.4,
38.2, 37.8, 25.84, 25.79, 22.2, 21.0, 13.3. IR (neat, cm-1): 2956
(w), 2843 (w), 2219 (w), 1694 (s), 1640 (s), 1452 (m). HRMS-ESI
(m/z): [M þ H]þ calcd for C38H40N6O3Cl2 699.2612, found
699.2624.
the pure 1, L-tartrate salt as an amorphous white powder (97 g,
80%): mp 80-83 °C. Two sets of NMR peaks due to the
tautomeric nature of this compound: 1H NMR (600 MHz,
DMSO) δ 8.16 (s, 0.7H), 8.11 (s, 0.3H), 7.89 (d, J = 1.9 Hz,
0.7H), 7.84 (d, J = 2.0 Hz, 0.3H), 7.79 (s, 0.7H), 7.73 (s, 0.3H),
7.71-7.59 (m, 3.7H), 7.56 -7.51 (m, 2.3H), 7.50-7.40 (m, 4H),
4.77 (s, 0.6H), 4.73 (s, 1.4H), 4.20 (s, 3H), 4.12-4.02 (m, 2H),
3.98 (s, 2H), 3.95 (s, 2H), 3.90-3.73 (m, 2H), 3.74-3.66 (m, 1H),
3.63-3.56 (m, 1H), 3.55 3.46 (m, 1H), 3.18 -3.10 (m, 1H),
2.99-2.68 (m, 4H), 2.45-2.35 (m, 1H), 2.26-2.16 (m, 2H),
2.04-1.90 (m, 2H), 0.88 (d, J = 6.3 Hz, 3H); 13C NMR (151
MHz, DMSO) δ 173.7, 165.5, 165.1, 164.3, 163.7, 142.7, 142.4,
138.3, 138.0, 137.45, 137.38, 135.04, 135.00, 133.22, 133.14,
132.61, 132.58, 132.4, 131.5, 130.9, 129.87, 129.82, 129.76,
129.4, 128.2, 127.3, 127.2, 122.30, 122.25, 120.98, 120.93,
110.2, 109.7, 94.2, 85.99, 85.94, 71.9, 71.6, 66.2, 54.7, 50.4,
50.1, 49.6, 46.2, 42.7, 42.3, 38.2, 37.8, 25.5, 25.5, 22.1, 20.9,
13.0. IR (neat, cm-1): 3009 (w), 2985 (w), 1691 (s), 1634 (s), 1592
(s), 1455 (m).
Compound 1, Monoethyl Oxalate (12). Two procedures were
used to prepare the monoethyl oxalate salt. (A) Compound 1
(free base, 100 mg) was dissolved in 4 mL of CH3CN at 50 °C.
Diethyl oxalate was added dropwise until the solution became
cloudy. The solution was left at 50 °C without stirring under N2
for 16 h. The precipitated crystalline solid was collected by
filtration and washed with CH3CN to afford the title salt. (B)
Compound 1 (free base, 1.5 g, 1.0 equiv) was dissolved in
CH3CN at 50 °C and then monoethyl oxalate (0.25 g, 1.0 equiv)
was added. The solution was stirred at 50 °C for 16 h and cooled
to room temperature. The precipitated crystalline solid was
collected by filtration and washed with CH3CN (1.27 g, 85%).
Identical salts were obtained with both methods: mp 152.9 °C.
Two sets of NMR peaks due to the tautomeric nature of this
compound: 1H NMR (600 MHz, DMSO) δ 8.16 (s, 0.7H), 8.12
(s, 0.3H), 7.90 (s, 0.7H), 7.84 (s, 0.3H), 7.80 (s, 0.7H), 7.73
(s, 0.3H), 7.71-7.52 (m, 6H), 7.53 -7.42 (m, 4H), 4.77 (s, 0.6H),
4.73 (s, 1.4H), 4.13 -4.04 (m, 6H), 4.02 (dd, J = 14.2, 7.1 Hz,
2H), 3.94-3.34 (m, 7H), 3.25-3.15 (m, 1H), 3.03-2.72 (m, 4H),
2.34 (br. s, 2H), 2.00 (br. s, 2H), 1.19 (t, J = 7.1 Hz, 3H), 0.93
(d, J = 6.3 Hz, 3H); 13C NMR (151 MHz, DMSO) δ 165.9,
165.5, 165.2, 164.4, 163.8, 163.1, 142.8, 142.5, 138.4, 138.0,
133.3, 133.2, 133.0, 132.7, 132.6, 131.6, 131.4, 129.94, 129.88,
128.4, 127.5, 127.4, 122.3, 122.2, 121.53, 121.49, 110.3, 109.8,
94.1, 86.3, 71.3, 65.9, 62.6, 59.0, 55.0, 50.0, 49.64, 49.61,
46.2, 42.7, 42.4, 38.2, 37.8, 25.3, 22.2, 21.0, 14.1, 12.8. IR
(neat, cm-1): 3457 (w), 3009 (w), 2986 (w), 1724 (m), 1694 (s),
1644 (s), 1454 (m).
Compound 1, L-Tartrate. The free base 1 (100 g, 0.14 mol, 1.0
equiv, >98% purity) was dissolved in EtOAc (1.5 L). L-Tartaric
acid (21.5, 0.14 mol, 1.0 equiv) dissolved in MeOH (150 mL) was
added over 1 h. After stirring at room temperature for an
additional 1 h, the precipitated solid was collected by filtration
and washed with EtOAc under N2.21 This amorphous material
contained ∼8 wt % residue solvents (EtOAc and MeOH) based
Acknowledgment. We wish to thank Prof. Scott E. Denmark,
University of Illinois, Urbana-Champaign, Prof. David
W. C. MacMillan, Princeton University, Dr. John Wiener,
Dr. Michael Ameriks and Dr. James P. Edwards for helpful
discussions; Dr. Jiejun Wu, Ms. Heather McAllister, and
Ms. Bita Naderi for analytical support; and Dr. Daniel J.
Pippel for proof-reading this manuscript.
1
on H NMR integration.22 Azeotropic distillation of the aqu-
eous solution (∼500 mL water) at 50 °C under vacuum success-
fully lowered the organic solvent residue to <0.5 wt %.
Subsequent lyophilization23 of the aqueous solution afforded
(20) The recovery was ∼60%. We also noticed slow decomposition of this
material on silica gel during the chromatography process.
(21) The material is highly hygroscopic. Without N2 flow, it gummed up
upon filtration.
(22) Removal of the residue solvents was extremely difficult. High
vacuum at 50 °C for a few days failed to reduce the residue solvent level.
Trituration in various solvents (EtOH, THF, 2-methyl-THF, PrOAc, etc.)
resulted in the trapping of the coming solvents.
(23) Direct lyophilization was not effective on removing the residue
organic solvents.
Supporting Information Available: General experimental
methods, 1H and 13C NMR spectra for compounds 1-4, 8-10,
and 12 and chiral HPLC chromatograms of compound 1. This
material is available free of charge via the Internet at http://
pubs.acs.org.
i
J. Org. Chem. Vol. 75, No. 6, 2010 1947