Chemical development of the casein kinase i - Epsilon inhibitor: 3-(3-fluorophenyl)sulfanyl-1 H -pyrrolo[3,2- b pyridine-2-carboxylic acid amide

Article abstract of DOI:10.1021/op200171a

The development of a scalable process for 3-arylsulfanyl-1H-pyrrolo[3,2- bpyridine-2-carboxylic acid amides (1), potent casein kinase I inhibitors, is described. The rapid identification of suitable reaction conditions expedited the lab scale synthesis of drug substances for early toxicological evaluations. Further improvements were made to achieve a safe and cost-effective process to meet increasing demands for drug substances to support clinical studies. This paper describes the synthesis at multikilogram scale.

Full text of DOI:10.1021/op200171a

ARTICLE
pubs.acs.org/OPRD
Chemical Development of the Casein Kinase I - Epsilon Inhibitor:
3-(3-Fluorophenyl)sulfanyl-1H-pyrrolo[3,2-b]pyridine-2-carboxylic
Acid Amide
Bao-Guo Huang,* Gregory Kubiak, John J. Shay, Clive Pemberton, James Peers, Reda G. Hanna,
Matthew R. Powers, Juan A. Gamboa, Ann M. Gelormini, Hyacinthe Yarabe, and Duane E. Rudisill
Molecular Innovative Therapeutics, Sanofi-Aventis US, 1041 Route 202-206, Bridgewater, New Jersey 08807, United States
ABSTRACT: The development of a scalable process for 3-arylsulfanyl-1H-pyrrolo[3,2-b]pyridine-2-carboxylic acid amides (1),
potent casein kinase I inhibitors, is described. The rapid identification of suitable reaction conditions expedited the lab scale
synthesis of drug substances for early toxicological evaluations. Further improvements were made to achieve a safe and cost-effective
process to meet increasing demands for drug substances to support clinical studies. This paper describes the synthesis at
multikilogram scale.
’ INTRODUCTION
Our attention focused on an indirect approach to introduce
the methyl group. Reaction of 2-chloro-3-nitropyridine with
diethyl sodiomalonate followed by decarboxylation was reported
to provide 2-methyl-3-nitropyridine in 63% yield after vacuum
distillation.9 The procedure used to prepare diethyl sodiomalo-
nate required heating sodium metal in diethyl malonate at 120 °C
and was not feasible for scale-up. A process with a better safety
profile was developed using diethyl malonate and a base (Scheme 2).
After a base screening, including potassium tert-butoxide, sod-
ium tert-butoxide, sodium hydride, sodium ethoxide, and sodium
bis(trimethylsilyl)amide (NaHMDS), sodium tert-butoxide was
selected. N-Methylpyrrolidinone (NMP) was determined to be
the superior solvent as it enhanced the solubility of the base and
accelerated the reaction. Excess diethyl malonate (2.0 equiv) was
required to suppress formation of the bis-arylation byproduct 7.
Upon completion of the arylation in NMP at 50 °C after 2 h, the
resulting 6 was decarboxylated by treatment with aqueous
sulfuric acid at 100 °C for 12 h to complete a one-pot process.
The product was extracted with toluene, and the toluene solution
was evaporated in vacuo to provide 2 as an oil. The process was
carried out at ca. 500 g scale to produce 2-methyl-3-nitropyridine
in 80% yield.
At pilot plant scale, the product was prepared as a concen-
trated toluene solution (vide infra). The toluene was partially
distilled off at a pot temperature up to 90 °C and a vacuum of 100
Torr. Compound 2 was thermally stable (decomposition tem-
perature: 263 °C by DSC) under the distillation conditions. The
process was performed in the pilot plant at 8.75 kg scale,
affording 2 as a toluene solution (∼60 wt %) in 90% yield.
Development of Oxalate Condensation (Step 2). A solvent
screen determined that the reaction in THF was much faster than
in ethanol (2 h vs 3 d). The reaction outcome was highly
dependent on the order of addition. Addition of diethyl oxalate
to a mixture of sodium ethoxide and 2, or addition of sodium
ethoxide to a mixture of diethyl oxalate and 2, resulted in
Casein kinase Iε, one of the isoforms of protein kinases, acts in
a molecular pathway that regulates the circadian rhythm.1
Following the discovery of the casein kinase I inhibitors, 3-ar-
ylsulfanyl-1H-pyrrolo[3,2-b]pyridine-2-carboxylic acid amides
1a,b (Scheme 1), and their potential use for the treatment of
depression,2 multikilo quantities were required for pharmacolo-
gical profiling and clinical trials. This led to the search for a safe,
high-yielding, and reliable process.
’ DISCUSSION
The Discovery chemistry route (Scheme 1) involved a Suzuki
coupling of the commercially available 2-chloro-3-nitropyridine
with methylboronic acid3 followed by alkylation of the carbanion
of 2 with diethyl oxalate to provide the ketoester 3. Reductive
cyclization of the ketoester via catalytic hydrogenation led to the
backbone of 4-azaindole. The sequential amidation and thioalk-
ylation of the azaindole ester 4 gave the drug substance 1a or 1b.
The Suzuki coupling suffered from a modest yield, low
reproducibility, and a long reaction time as well as costly
reagents. An undesired chromatographic purification was also
used for the product purification. Step 2 was very sluggish taking
up to 3 days. The amidation required a pressure reactor and
suffered a slow conversion. The final step required a long reaction
time at 100 °C and subsequent chromatographic purification.
The first aim was to provide workable conditions that could be
used on a large lab scale (e.g., multihundred grams) with relative
ease of processing, in reasonable yield, and in a short time frame.
The second target was to ensure that further scale-up in the pilot
plant was achieved at multikilogram scale.
Development of Ring Methylation Process (Step 1). Sev-
eral alternative cross-coupling conditions were briefly examined
using Me3Al,4 MeMgCl,5 and trimethylboroxine,6 but none
provided a yield greater than 40%. Direct coupling of 2-chloro-
3-nitropyridine with ethyl pyruvate gave 3 in a modest yield
(<50%),7,8 and the need to use an expensive ligand, 2-dicyclo-
hexylphosphino-20-(N,N-dimethylamino)biphenyl, made the
method less attractive.
Received: June 27, 2011
Published: August 03, 2011
r
2011 American Chemical Society
1040
dx.doi.org/10.1021/op200171a Org. Process Res. Dev. 2011, 15, 1040–1045
|

Organic Process Research & Development
ARTICLE
Scheme 1. Discovery route to 4-azaindole amides 1a,ba
a Reagents and conditions: Step 1. MeB(OH)2, K2CO3, Pd(PPh3)4 (10 mol %), dioxane, 100 °C, 2ꢀ3 d, 57% after chromatography; Step 2. diethyl
oxalate, NaOEt, EtOH, rt, 3 d, 64%; Step 3. Pd/C, H2 (53 psi), EtOH, rt, 4 h, 89%; Step 4. NH3 in MeOH, rt, 3 d, 86%; Step 5. for 1a: diphenyl disulfide,
Cs2CO3, DMF, 100 °C, 2 d, 60% after chromatography; for 1b: bis(3-fluorophenyl) disulfide, Cs2CO3, DMF, 100 °C, 20 h, 68% after chromatography.
Scale: 0.1ꢀ10 g. Overall yield: 16.8%.
Scheme 2. Preparation of 2-methyl-3-nitropyridine via mal-
onate intermediate 6
Table 1. Effect of concentration of 2 in toluene on ketoester
formation
Wt % of 2
Isolated
Purity
Batch Size
in toluene
Yields (%)
(Area% by HPLC)
31 g
24 g
24 g
34%
52%
70%
82%a
94%
91%
99.2%
99.6%
98.6%
a 18.6% remaining starting material.
of 2 in the toluene solution affected the rate of the reaction as well as
the isolated yield (Table 1). While a dilute solution (34 wt %) gave a
relatively lower conversion, a concentrated solution (70 wt %) af-
forded a product with a slightly lower purity. A moderate concen-
tration (ca. 50 wt %) was deemed to provide the optimal results.
Oxalic acid, coprecipitated out with 3, was shown to slow
down filtration and inhibited initiation of subsequent transfer
hydrogenation. This was overcome by ensuring that a sufficient
amount of water and isopropyl alcohol was used to wash the filter
cake. The process was performed in the plant at 9 kg scale in 84%
yield and a purity of 96.5 wt %.
Scheme 3. Preparation of ketoester 3
Development of Pyrrole Ring Cyclization (Step 3). The
original catalytic hydrogenation conditions were used for the lab
preparation. A strong exotherm with an adiabatic temperature
rise of 32 °C was observed during the initial hydrogen uptake at
57 g scale. After the exotherm subsided, the mixture was held to
40 °C to progress to completion. The crude product, containing
ca. 2% of a six-membered ring byproduct 8,10 was reslurried in a
mixture of ethyl acetate and heptane at reflux, giving 4 in >90%
recovery (<0.2% impurity 8). The process was used in the lab to
prepare 640 g of 4.
formation of a significant amount (>10%) of undesired dimeric
product. A large exotherm was generated when mixing diethyl
oxalate with sodium ethoxide, which could be the cause for the
dimer formation. For optimal results, it was essential to dose 2 to
a premixed solution of diethyl oxalate (3.0 equiv) and sodium
ethoxide (2.0 equiv) in THF at 4 °C. The resulting mixture was
then warmed to room temperature and held for 2 h. Upon
treatment with aqueous ammonium chloride, the ketoester 3 was
isolated by vacuum filtration in 88% yield. The oxalic acid
byproduct, resulting from excess diethyl oxalate, was removed
with a water slurry. The procedure was used in the lab to prepare
over 600 g of the ketoester 3 (Scheme 3).
On a multikilogram scale run, 2 was prepared as a toluene
solution to avoid vacuum distillation to dryness. The concentration
The catalytic hydrogenation was used for the quick lab batch
delivery, but the process was not amenable for further scale-up
1041
dx.doi.org/10.1021/op200171a |Org. Process Res. Dev. 2011, 15, 1040–1045

Organic Process Research & Development
ARTICLE
Scheme 4. Intermediates to the azaindole ester 4
Scheme 5. Intermediates and byproducts observed in the amidation reaction
Scheme 6. Thioalkylation of azaindole amide 5
Scheme 7. Proposed mechanism for thiophenyl addition
due to pilot plant limitations at that time. Effort was then focused
on the development of a transfer hydrogenation to achieve the
desired transformation. The reaction was believed to proceed
through intermediate 9 and cyclized hydroxylamine 10 as
observed by following the reaction closely with HPLC and
LCMS (Scheme 4).
lower than that for reactions that successfully initiated. In light of
the performance variability observed for various lots of the
ketoester, an additional 10% of the ammonium formate solution
was charged to ensure a smooth initiation of the reaction. The
catalytic decomposition of ammonium formate resulted in a pH
increase that might help solubilize the oxalic acid.
Two lab batches (25 and 65 g) were carried out using the final
process, providing 4 in 80ꢀ87% yield.
Ammonium formate, formic acid, and ammonium formate
with acetic acid were examined as hydrogen donors. The reaction
with formic acid alone was relatively slow (<10% after 2 h).
Attention was next turned to the ammonium formateꢀacetic
acid system. The rate of reaction was shown to be comparable to
ammonium formate by itself, and the acetic acid minimized the
amount of sublimate. An attempt was made to perform the
reaction with Pd/C and ammonium formate in the absence of
acetic acid, with the hope that at a lower temperature (35 °C) the
ammonium salt byproduct would not sublime. Also, eliminating
acetic acid might simplify the workup by avoiding the need for
neutralization. Upon completion of the reaction (pH = 8ꢀ9), a
significant amount of ammonium salt sublimate had still formed
on the reactor walls and at the bottom of the condenser. This led
to a decision to use the ammonium formateꢀacetic acid process.
While working on the development of this step, it became
apparent that the conditions required for the initiation of the
transfer hydrogenation were dependent on the quality of the
ketoester 3 used. The reaction using one particular batch of
ketoester (containing 3.2 wt % of oxalic acid) did not initiate at all
under the standard conditions, where the pH of the mixture was
The process was transferred to the pilot plant, and four batches
at 8 kg scale were completed. The reaction and workup proceeded
satisfactorily, in good agreement with the observations in the lab. A
total of 22.1 kg of 4 were produced in 86.9ꢀ89.9% yield.
Development of Amidation (Step 4). The amidation was
initially performed with a solution of ammonia in methanol
taking 3 d to complete at room temperature. To improve the rate
of conversion, the reaction was carried out at 50 °C with less
MeOH (15 parts) and was complete in 36 h. The reaction
mixture started as a suspension that dissolved on heating, and the
product precipitated upon cooling. The mixture was filtered and
washed with methyl tert-butyl ether to provide 5 in 90ꢀ95%
yield. The reaction in methanol proceeded through a trans-
esterification to the methyl ester 11a; thus the impurities observed
1042
dx.doi.org/10.1021/op200171a |Org. Process Res. Dev. 2011, 15, 1040–1045

Organic Process Research & Development
ARTICLE
Scheme 8. Scale-up Synthesis of Drug Substance 1a,ba
a 1st Generation Process Chemistry: Step 1. (a) diethyl malonate, NaOtBu, NMP, 50 °C, 2 h; (b) 6 M aq. H2SO4, 100 °C, 7 h, 80%; Step 2. diethyl oxalate,
NaOEt, THF, rt, 2 h, 88%; Step 3. Pd/C, EtOH, H2 (50 psi), 40 °C, 4 h, 85%; Step 4. NH3 in MeOH, 50 °C, 36 h, 90%; Step 5. (a) diphenyl disulfide,
Cs2CO3, NMP, 120 °C, 24 h, 84%; (b) bis(3-fluorophenyl) disulfide, Cs2CO3, NMP, 120 °C, 2 h, 95%. Scale: 10ꢀ330 g, Overall yield: 45.2%. 2nd
Generation Process Chemistry: Step 1. (a) diethyl malonate, NaOtBu, NMP, 50 °C, 2 h; (b) 6 M aq. H2SO4, 100 °C, 12 h, 90.1%; Step 2. diethyl oxalate,
NaOEt, THF, rt, 2 h, 84% after reslurry; Step 3. Pd/C, HCOONH4, CH3COOH, EtOH, 35 °C, 2 h, 89.2%; Step 4. NH3 in ethylene glycol, 65 °C, 6 h,
95%; Step 5. bis(3-fluorophenyl) disulfide, Cs2CO3, NMP, 120 °C, 2 h, 97.6%. Scale: 5ꢀ30 kg, Overall yield: 62.6%.
in this reaction included the ethyl and the methyl esters (Scheme 5).
Also observed was the corresponding carboxylic acid 11b, which
was washed away in the filtrate.
Alternative conditions were explored to shorten the reaction
times. Literature indicated that similar reactions have been run
in aqueous ammonium hydroxide solution.11 Efforts toward
this end showed that the starting material was consumed in just
5 h at 60 °C at ∼20 psi. The best isolated yield from the small
scale reactions was 64%. A significant amount of 11b was
formed.
An alternative procedure utilizing ethylene glycol reduced the
pressure needed to ∼1 psi and decreased the reaction time to 6 h.
Ammonia is reported to have a higher solubility in ethylene
glycol than in other solvents such as chlorobenzene, acetone, and
butanol.12 Utilizing this solvent, the amidation was complete at
65 °C in 6 h. This reaction did not generate significant pressure
(<1 psi). The yield range was 90ꢀ97% with a target purity of
greater than 98%. The main impurity was the glycol ester 11c. It
was shown that approximately 2% of glycol ester impurity does
not impact the next step in the synthesis.
The reaction was believed to proceed as depicted in Scheme 7.
The azaindole amide 5 was deprotonated, and the resulting anion
attacked the SꢀS bond of the aryl disulfide to form the
thiophenyl azaindole. The subsequent hydrogen transfer from
C3 to N1 was readily achieved with formation of a stable
aromatic system. The anion was hydrolyzed to produce product
1a or 1b.
It was found that 2.0 equiv of base were essential to drive the
reaction to completion. It was also determined that the reaction
proceeded faster in NMP than in DMF. The reaction progressed
to completion at 120 °C using 2.5 equiv of cesium carbonate and
1.2 equiv of the disulfide. The crude product was precipitated
from the reaction mixture by adding water. The excess aryl
disulfide was removed by slurrying the crude product in a mixture
of ethyl acetate and heptane. The crude drug substance was then
recrystallized from MeOH in >85% recovery. In the case of the
fluorophenyl disulfide, impurity 12 was observed in the isolated
product, suggesting an additional thiophenyl moiety reacting
with 1b.
The plant operators preferred to add the azaindole ester 4
followed by the ethylene glycol. Subsurface addition of gaseous
ammonia to the resulting slurry was carried out at room tem-
perature in a closed system, and no heat generation was observed.
The reaction was heated to 65 °C for 2 h and then cooled to
20 °C. After adding water, the amide 5 was isolated by filtration.
The cake was washed three times with water to remove residual
ethylene glycol. The campaign in the plant utilizing the ethylene
glycol process provided 17 kg of 5 in four batches in high yield
(90ꢀ94%) and purity (98.4ꢀ98.7%).
The stoichiometry of bis(3-fluorophenyl) disulfide has a
strong effect on the formation of 12. About 0.67% 12 formed
when 1.25 equiv of the reagent was used. Use of 1.1 equiv of the
reagent worked well, sufficiently driving the reaction to comple-
tion while 12 was not detected. The ethylene glycol carried over
from step 4 was removed during workup. Under normal condi-
tions, up to 1ꢀ2 wt % ethylene glycol was well tolerated.
Development of Thiophenyl Addition (Step 5). In the
original Discovery chemistry route, the amide 5 was treated with
the aryl disulfide (1.5 equiv) and cesium carbonate (1.2 equiv)
in DMF at 100 °C for 2 days, providing 60% of 1a (Scheme 6).
The low yield was due to an incomplete reaction. Removal of
the remaining 5 from the desired product 1a required flash
chromatography.
1043
dx.doi.org/10.1021/op200171a |Org. Process Res. Dev. 2011, 15, 1040–1045

Organic Process Research & Development
ARTICLE
Both 1a and 1b were prepared on lab scale to support early
profiling. Only 1b was selected as a drug candidate and then
prepared at 5 kg scale in the pilot plant.
The reactor was heated to 50 ( 3 °C and held for 2 h. A sample
was taken and analyzed by HPLC, and the reaction was con-
sidered complete when e0.5 A% of 2-chloro-3-nitropyridine
remained (actual amount = 0.14 A%). A solution of 6 M sulfuric
acid (57 kg) was charged over 25 min at 50ꢀ60 °C (note: the
sulfuric acid addition is exothermic, and the batch thickens
during the first 20% of the charge. As the addition proceeds, a
complete solution is obtained with the color change from dark
brown to orange). The reaction was heated to 100 ( 3 °C and
held for 10 h. A sample was taken and analyzed by HPLC. The
reaction was considered complete when e1 A% of the monoe-
ster 6 remained (actual amount =0.44 A% A% = Area%). The
batch was cooled to 15 ( 5 °C and transferred to a glass lined
reactor. Toluene (76 kg) was charged to the reactor. A solution of
45 wt % potassium hydroxide (105 kg) was charged over 45 min
at 20 ( 10 °C. The batch was filtered through Celite (4.0 kg).
The filter cake was washed with water (16.4 kg), and the toluene
layer was separated (agitated for 15 min, then settled for 15 min).
The aqueous layer was extracted with toluene (76 kg). The
combined toluene layers (146.8 kg) were distilled at 100 Torr,
and reactor temperature was gradually adjusted up to 90 °C until
the batch was at the minimum stirred volume (∼20 L). The
batch was cooled to 20 ( 10 °C to yield a solution of 2 in toluene
(total weight: 11.8 kg, 58.5 wt %, 90.1% yield). An aliquot was
’ CONCLUSIONS
This paper describes the development of lab synthesis and a
subsequent scale up process for casein kinase I inhibitors (1a and
1b). Introduction of the 2-methyl group was readily achieved via
a two-step one-pot process in 90% yield using the cost-effective
malonate chemistry. Transfer hydrogenation was implemented
in the pilot plant to achieve the desired reductive cyclization.
Amidation in ethylene glycol went to completion within a few
hours under atmospheric pressure, eliminating the need for a
pressure reactor. The overall yield was improved from 17% to
63%. A total of 28 kg of 1b was prepared and provided for clinical
trials. The improved process is summarized in Scheme 8.
’ EXPERIMENTAL SECTION
General. 2-Chloro-3-nitropyridine was purchased from JVR
Fine Chemie Pvt., India, and bis(3-fluorophenyl) disulfide was
sourced from Zhejiang Shou & Fu Chemical Co., China. Mass
spectra were obtained on a Finnigan MAT TSQ 700 mass
spectrometer using electron impact (EI) at 70 eV and chemical
ionization (CI) with the relative peak height in percent and the
molecular ion designated as M given in parentheses.
1
distilled to provide a sample for analysis: H NMR (300 MHz,
CDCl3) δ: 8.70 (d, 1H), 8.25 (d, 1H), 7.35ꢀ7.40 (m, 1H), 2.85
(s, 3H).
HPLC was performed on Luna 5 μ phenyl-hexyl 4.6 mm ꢁ
150 mm. Flow rate: 1.0 mL/min; column temperature: 25 °C,
equilibrate column for 5 min between runs. Typical total run
time: 35 min.
Method for step 1 and 2: Mobile phase: (A) 0.1% TFA in
water; (B) 0.1% TFA in acetonitrile. Injection volume: 10 μL, λ =
240 nm; diluting solvent: 50/50 solvent A/solvent B. Gradient
from 30% to 40% solvent B over 10 min, then from 40% to 70%
over 16 min. Retention times: 2-chloro-3-nitropyridine, 6.5 min;
2-methyl-3-nitropyridine 2, 4.6 min; aryl malonate 6, 6.3 min;
bis-aryl malonate 7, 14.8 min; ketoester 3, 13.1 min; diethyl
oxalate: 6.1 min; oxalic acid, 1.5 min.
Method for step 3: Mobile phase: (A) 0.1% TFA in water; (B)
0.1% TFA in acetonitrile. Injection volume: 5 μL, λ = 230 and
290 nm; diluting solvent: methanol. Gradient from 5% to 40%
solvent B over 15 min, then from 40% to 95% solvent B over
10 min. Retention times: hydroxyamine 9, 4.6 min; azaindole
ethyl ester 4, 7.4; N-hydroxy azaindole 10, 8.0 min.
Method for steps 4 and 5: Mobile phase: (A) Dissolve 3.86 g of
ammonium acetate in a 1000 mL of HPLC grade water (pH 6.9).
Filter the buffer using a 0.45 m nylon filter. (B) Acetonitrile.
Injection volume: 5 μL, λ = 230 and 290 nm; diluting solvent:
50/50 water/acetonitrile. Gradient from 5% to 40% solvent B
over 15 min, then from 40% to 95% solvent B over 10 min.
Retention times: azaindole amide 5, 6.8 min; glycol ester 11c, 8.7
min; azaindole ester 4, 12.9 min; 1b, 15.6 min; byproduct 12,
20.8 min; bis(3-fluorophenyl) disulfide, 25.2 min.
2-Hydroxy-3-(3-nitropyridin-2-yl)acrylic Acid Ethyl Ester
(3). A glass lined reactor was charged with sodium ethoxide
(8.8 kg, 129.3 mol) and THF (65 kg). The reaction was adjusted
to 10 ( 3 °C. Diethyl oxalate (29 kg, 198.4 mol) was charged
over 30 min at 10 ( 5 °C. The batch was held at this temperature
for 30 min. The toluene solution of 2 (8.5 kg, 48.9 wt % and
8.35 kg, 58.1 wt %, corresponding to 9.0 kg of 2, 65.2 mol) was
charged over 20 min at 10 ( 5 °C. The reactor was warmed to
20 ( 3 °C and held for 2 h. A sample was taken and analyzed by
HPLC. The reaction was considered complete when <1.0 A% of
2 remained (actual amount = 0.2 A%). Ammonium chloride
solution (87 kg, preprepared from 17.4 kg of ammonium chloride
and 69.6 kg of water) was charged over 1 h at 20 ( 5 °C. Water
(16 kg) and isopropyl alcohol (32 kg) were added, and the
mixture was stirred at 20 ( 5 °C for 1 h. The batch was filtered,
and the cake was washed with water (2 ꢁ 54 kg). Once the rate of
filtrate collection slowed, the cake was transferred to a second
glass lined reactor. Water (210 kg) was charged, and the sus-
pension was stirred (100 rpm) at 20 ( 5 °C for 1 h. The batch
was filtered, and the cake was washed with water (2 ꢁ 27 kg).
The filter cake was dried at 130 Torr and 45 ( 5 °C to reach
LOD e 2.0 wt % (actual: 0.41 wt % in 2 days) to provide 3 (13.1
1
kg, 84.1% yield, 96.46 wt %). H NMR (300 MHz, CDCl3) δ:
14.60 (br, 1H), 8.70 (d, 1H), 8.40 (m, 1H), 7.35ꢀ7.40 (m, 2H),
13
3
4.20 (q, 2H, OCH2CH ), 1.40 (t, 3H, OCH2CH3); C NMR:
(75 MHz, CDCl3) δ: 163.2, 156.2, 151.5, 150.0, 142.0, 134.7,
121.2, 98.2, 62.4, 14.4.
2-Methyl-3-nitropyridine (2). A glass lined reactor was
charged with sodium tert-butoxide (11.7 kg, 121.7 mol) and
NMP (36 kg) at 20 ( 3 °C. Diethyl malonate (17.7 kg, 110.5 mol)
was charged over 45 min at 20 ( 10 °C. The batch was agitated at
20 ( 3 °C for 15 min. A preprepared solution of 2-chloro-3-
nitropyridine (8.75 kg, 55.2 mol) in NMP (23 kg) was charged
over 45 min at 20 ( 5 °C (note: higher temperatures must
be avoided as it can lead to formation of diaryl byproduct 7).
1H-Pyrrolo[3,2-b]pyridine-2-carboxylic Acid Ethyl Ester
(4). A glass lined reactor was charged with ketoester 3 (8 kg,
33.6 mol) followed by a suspension of 10% Pd/C (1.9 kg) in
water (8.8 kg). Ethanol (3C, 66 kg) and glacial acetic acid (10 kg,
5.0 equiv) were added, and the batch was heated to 35 ( 3 °C. A
preprepared solution of ammonium formate (0.74 kg) in water
(0.96 kg) was charged over 15 min at 35 ( 3 °C (note: an
exotherm and off-gas should be observed as an indication of
1044
dx.doi.org/10.1021/op200171a |Org. Process Res. Dev. 2011, 15, 1040–1045

Organic Process Research & Development
ARTICLE
initiation of the reaction). The second portion of the ammonium
formate solution (6.66 kg of ammonium formate in 8.64 kg of
water) was charged over 2 h at 35 (3°C. The batch was held for 1 h.
A sample was taken and analyzed by HPLC. The reaction was
considered complete when <0.1 A% of 3 and <0.5 A% intermediate
hydroxylamine remained (actual amount: 0.009 A% 3 and 0.21 A%
hydroxylamine). Celite (1.1 kg) was charged, and the batch was
agitated for 15 min. The filter was precoated with Celite (1.7 kg). The
batch was filtered through Celite, and the filter cake was washed with
ethanol (3C, 13 kg) and water (16 kg), sequentially (note: keep the
Pd residue water wet as the dry catalyst is highly flammable). The
combined filtrate and washes were transferred to a glass lined reactor,
and water (96 kg) was charged. The batch was distilled at 60 °C and
150 Torr until the batch was about half volume. The batch was then
cooled to 5 ( 5 °C over 1 h, and the pH was adjusted to 8.05 (target
range: 8.0 ( 0.5) using aqueous sodium hydroxide solution (2.5 M,
24 kg). The batch was stirred at 5 ( 5 °C for 1 h and filtered. The
filter cake was washed with water (2 ꢁ 16 kg). The cake was dried at
60 ( 5 °C and 130 Torr to provide azaindole ester 4 (5.7 kg, 89.2%
yield, 98.7 wt % pure). 1H NMR (300 MHz, CDCl3) δ:8.40(s, 1H),
3-(3-Fluorophenyl)sulfanyl-1H-pyrrolo[3,2-b]pyridine-2-
carboxylic Acid Amide (1b). A glass lined reactor was charged
with 5 (5 kg, 31.0 mol), cesium carbonate (26 kg, 79.8 mol, 2.5
equiv), and NMP (52 kg). The mixture was stirred at 25 ( 5 °C.
Bis(3-fluorophenyl) disulfide (8.7 kg, 34.2 mol, 1.1 equiv) was
charged. The batch was heated to 120 ( 5 °C and held for 1 h.
The batch was cooled to 75 ( 5 °C for sampling. The reaction
was considered complete when e0.5 A% of 5 present (actual
amount: 0.14 A%). Water (50 kg) was charged over 45 min at
15ꢀ30 °C. The batch was stirred at 20 ( 5 °C for 1 h and then
filtered, and the cake was washed with water (2 ꢁ 25 kg). The
cake (15.6 kg) was transferred into a Hastelloy reactor. Heptane
(50 kg) was charged followed by ethyl acetate (17 kg). The
mixture was stirred at 20 ( 5 °C for 1 h and filtered. The cake was
washed with heptane (2 ꢁ 8 kg) and dried at 40 ( 5 °C and 130
1
Torr to give 1b (8.7 kg, 97.6% yield). H NMR (300 MHz,
DMSO-d6) δ: 12.60 (br, 1H), 8.55 (s, 1H), 8.20 (d, 1H), 7.95
(m, 2H), 7.30 (m, 2H), 6.80ꢀ7.00 (m, 3H); 13C NMR: (75
MHz, DMSO-d6) δ: 161.3, 145.4, 145.3, 137.0, 130.9, 130.8,
129.9, 122.0, 122.0, 120.8, 119.7, 112.8, 112.5, 100.1.
3
7.90 (d, 1H), 7.20ꢀ7.40 (m, 2H), 4.45 (q, 2H, OCH2CH ), 1.40 (t,
13
3H, OCH2CH3); C NMR: (75 MHz, CDCl3) δ: 161.6, 144.0,
144.0, 131.1, 120.9, 119.5, 106.5, 61.2, 23.5.
’ AUTHOR INFORMATION
Corresponding Author
*E-mail: baoguo.huang@sanofi-aventis.com.
1H-Pyrrolo[3,2-b]pyridine-2-carboxylic Acid Amide (5). A
glass lined reactor was charged with azaindole ester 4 (5.5 kg, 28.9
mol) and ethylene glycol (56.1 kg). The mixture was stirred at 23 (
3 °C. Ammonia (3.6 kg) was charged through a dip pipe to the
closed reactor at 20ꢀ35 °C. The mixture was stirred for 15 min at
25 ( 3 °C. The reactor was heated to 65 ( 3 °C over 40 min while
venting to the acid scrubber. The batch was held at 65 (3 °Cfor6h
in the closed reactor. A sample was taken and analyzed by HPLC.
The reaction was considered complete when g98 A% of 5 was
present (actual amount: 98.0 A%). Water (23 kg) was charged, and
the batch was cooled to 5 ( 2 °C over 30 min and held at that
temperature for 1 h. The mixture was filtered, and the filter cake was
washed with water (2 ꢁ 33 kg) and dried at 65 °C until water
content was <0.5% by Karl Fischer (KF: 0.15% in 24 h) to give 5
(4.45 kg, 95% yield, 97.3 wt %). 1HNMR(300MHz, DMSO-d6) δ:
11.60 (br, 1H), 8.40 (s, 1H), 8.20 (br, 1H), 7.80 (d, 1H), 7.60 (br,
1H), 7.15ꢀ7.25 (m, 2H); 13C NMR: (75 MHz, DMSO-d6) δ:
162.5, 145.2, 143.8, 134.3, 129.5, 119.6, 118.8, 102.2.
’ ACKNOWLEDGMENT
The authorsare grateful tomany colleaguesatSanofi-Aventis for
their support on technical, process safety, and sourcing activities.
We would like to thank Drs. Andrew Bridge, Franz Weiberth, and
David Lythgoe for their helpful discussions.
’ REFERENCES
(1) Eide, E. J.; Woolf, M. F.; Kang, H.; Woolf, P.; Hurst, W.;
Camacho, F.; Vielhaber, E. L.; Giovanni, A.; Virshup, D. M. Mol. Cell.
Biol. 2005, 25 (7), 2795–807.
(2) Metz, W. A.; Halley, F.; Dutruc-Rosset, G.; Choi-Sledeski, Y. M.;
Bernard, P. G.; Fink, D. M.; Doerflinger, G.; Huang, B.-G.; Gelormini,
A. M.; Gamboa, J. A.; Giovanni, A.; Roehr, J. E.; Tsay, J. T.; Camacho, F.;
Hurst, W. J.; Harnish, S. W.; Chiang, Y. U.S. Pat. Appl. Publ. 2005, 30
pp. CODEN: US 20050131012 A1 20050616 CAN 143:43870 AN
2005:527395.
(3) Niu, C.; Li, J.; Doyle, T. W.; Chen, S.-H. Tetrahedron 1998,
54, 6311.
(4) Hirota, K.; Kitade, Y.; Kanbe, Y.; Isobe, Y.; Maki, Y. Synthesis
1993, 2, 213.
(5) Huang, B.-G.; Tang, D. U.S. Patent 5,874,606, Feb. 23, 1999,
CAN 130:110059 AN 1999: 130424.
(6) Gray, M.; Andrews, I. P.; Hook, D. F.; Kitteringham, J.; Voyle, M.
Tetrahedron Lett. 2000, 41, 6237.
(7) Rutherford, J. L.; Rainka, M. P.; Buchwald, S. L. J. Am. Chem. Soc.
2002, 124, 15168.
3-Phenylsulfanyl-1H-pyrrolo[3,2-b]pyridine-2-carboxylic
Acid Amide (1a). A mixture of diphenyl disulfide (177.8 g, 1.5
equiv), Cs2CO3 (351.9 g, 2.0 equiv), and amide 5 (87.5 g, 0.54 mol)
in NMP (1.20 L) was heated to 120 °C for 21 h. Additional Cs2CO3
(87.97 g, 0.5 equiv) was added, and the reaction was held at 120 °C
for another 4 h. The reaction was judged to be complete (94% of 1a,
1.4% of unreacted 5, 4.5% of PhSSPh). The reaction mixture was
cooled to rt and combined with the crude mixture from another batch
(starting from 78.1 g of amide 5) for workup. The combined mixture
was poured into iceꢀwater (6 L) and stirred for 1 h. The brown
colored solid was filtered-off, washed with water (2 ꢁ 300 mL), and
air-dried for 6 h. The solids were twice slurried with 20% EtOAc/
heptane (2.0 L) at rt to remove PhSSPh. The crude product was
divided into two batches (150 and 154.8 g), and each treated with
charcoal in THF (6.0 L) at reflux for 1 h. The mixture was cooled to
room temperature and filtered through Celite, and the filtrate was
concentrated to give light brown solids (122 and 137 g), respectively.
(8) Jorgensen, M.; Lee, S.; Liu, X.; Wolkowski, J. P.; Hartwig, J. F.
J. Am. Chem. Soc. 2002, 124, 12557.
(9) Liu, M.-C.; Lin, T.-S.; Sartorelli, A. C. Synth. Commun. 1990, 20
(19), 2965.
(10) Suzuki, H.; Gyoutoku, H.; Yokoo, H.; Shinba, M.; Sato, Y.;
Yamada, H.; Murakami, Y. Synlett 2000, 8, 1196.
(11) Silvestri, R.; De Martino, G.; La Regina, G.; Artico, M.; Massa,
S.; Vargiu, L.; Mura, M.; Loi, A. G.; Marceddu, T.; La Colla, P. J. Med.
Chem. 2003, 46, 2482.
1
The isolated yield was 94%. H NMR (300 MHz, DMSO-d6) δ:
12.60 (br, 1H), 8.50 (s, 1H), 8.20 (br, 1H), 7.95 (d, 2H), 7.20 (m,
4H); 13C NMR: (75 MHz, DMSO-d6) δ: 161.1, 145.3, 144.9, 136.8,
136.4, 129.9, 128.9, 125.9, 125.4, 120.5, 119.5, 101.0.
(12) Short, I.; Sahgal, A.; Hayduk, W. J. Chem. Eng. Data 1983,
28, 63–66.
1045
dx.doi.org/10.1021/op200171a |Org. Process Res. Dev. 2011, 15, 1040–1045