The synthesis of a dopamine D2 partial agonist for the treatment of schizophrenia

Article abstract of DOI:10.1021/op800307k

The synthesis of the phosphoric acid salt of dopamine D2 partial agonist 2-{4-[4-(7-fluoro-naphthalen-1-yl)-piperazin-1-yl]- butoxy}-5,6,7,9-tetrahydro- 1,7,9-triaza-benzocyclohepten-8- one (1) is reported. The most prominent feature of the molecule is a seven-membered ring urea functionality that has been prepared via an efficient one-pot, three-step transformation. The original synthesis from the Medicinal Chemistry group provided precursor 13 in 10 steps and 2% overall yield, required four chromatographies and employed unsafe reagents such as 2-iodoxybenzoic acid (IBX) and HClO₄. The optimized synthetic route for the preparation of phosphate salt 1 consists of 12 linear steps with a 10% overall yield. Safer and more robust reaction conditions have been developed with only one required chromatography. Another key step in the synthesis is the coupling of iodide 25 with naphthalenopiperazine 12 to provide 13. Due to the difficulty to purify this intermediate, a protocol had to be developed to obtain crude material with the required purity, suitable for use in the subsequent salt formation step. Finally, considerable work was carried out to determine the most stable polymorph of the API. As a result, a robust set of conditions has been developed for the formation of phosphoric acid salt 1, providing the desired polymorph in excellent yield and purity.

Full text of DOI:10.1021/op800307k

Organic Process Research & Development 2009, 13, 555–566
The Synthesis of a Dopamine D2 Partial Agonist for the Treatment of Schizophrenia
Javier Magano,*,† Alison Acciacca,† Anne Akin,† Benjamin M. Collman,† Brian Conway,† Michael Waldo,† Michael H. Chen,‡
and Kenneth E. Mennen§
Pfizer Global Research and DeVelopment, Eastern Point Road, Groton, Connecticut 06340, U.S.A.
Abstract:
The synthesis of the phosphoric acid salt of dopamine D2 partial
agonist 2-{4-[4-(7-fluoro-naphthalen-1-yl)-piperazin-1-yl]-
butoxy}-5,6,7,9-tetrahydro-1,7,9-triaza-benzocyclohepten-8-
one (1) is reported. The most prominent feature of the
molecule is a seven-membered ring urea functionality that
has been prepared Wia an efficient one-pot, three-step
transformation. The original synthesis from the Medicinal
Chemistry group provided precursor 13 in 10 steps and 2%
overall yield, required four chromatographies and employed
unsafe reagents such as 2-iodoxybenzoic acid (IBX) and
HClO4. The optimized synthetic route for the preparation
of phosphate salt 1 consists of 12 linear steps with a 10%
overall yield. Safer and more robust reaction conditions
have been developed with only one required chromatogra-
phy. Another key step in the synthesis is the coupling of
iodide 25 with naphthalenopiperazine 12 to provide 13. Due
to the difficulty to purify this intermediate, a protocol had
to be developed to obtain crude material with the required
purity, suitable for use in the subsequent salt formation step.
Finally, considerable work was carried out to determine the
most stable polymorph of the API. As a result, a robust set
of conditions has been developed for the formation of
phosphoric acid salt 1, providing the desired polymorph in
excellent yield and purity.
Figure 1. Structure of dopamine D2 partial agonist 1.
treatment with current agents3 can lead to tardive dyskinesia,
blood dyscrasias, weight gain, diabetes mellitus, akathisia,
dyslipidemia, QT interval prolongation and Parkinsonism.1,4
Therefore, there is a clear need for new and better options to
treat this serious disorder.
Recently, the phosphoric acid salt of 2-{4-[4-(7-fluoro-
naphthalen-1-yl)-piperazin-1-yl]-butoxy}-5,6,7,9-tetrahydro-
1,7,9-triaza-benzocyclohepten-8-one (1) (Figure 1) has been
identified at Pfizer as a dopamine D2 partial agonist candidate
for the treatment of schizophrenia. This contribution describes
the route development work for the preparation of this com-
pound in kilogram quantities for clinical studies.
Medicinal Chemistry Route. Prior to the identification of
phosphoric acid salt 1 as the desired target, the efforts of the
Medicinal Chemistry group were directed toward the preparation
of free base 13 (Scheme 1).5 The synthesis started with
commercially available 2,6-dichloropyridine (2), which was
treated with concentrated ammonia at high pressure and
temperature for 2 days to give 2-amino-6-chloropyridine (3) in
fair yield.6 The amino group was converted to the corresponding
pivalamide 4 and subsequent low-temperature formylation of
the pyridine ring with DMF afforded aldehyde 5, which was
purified by chromatography.7a Chloride displacement with
4-benzyloxy-1-butanol in DMF8 followed by Wittig olefination
under cryogenic conditions with (methoxymethyl)triphenylphos-
phonium chloride and PhLi as base generated methyl vinyl ether
Introduction
Dopamine D2 receptor antagonists and partial agonists have
been used clinically for treating central nervous system (CNS)
disorders, such as schizophrenia and bipolar disorder. In
particular, schizophrenia is a chronic, highly debilitating mental
disorder that is equally common in men and women and affects
1% of the population.1 There are two million people affected
in the United States alone, which results in 330,000 hospitaliza-
tions each year.2 Thirty percent of the patients do not respond
adequately to current antipsychotic treatments or have continued
needs for improvement in core symptoms such as negative
symptoms, cognitive decline, and/or overall schizophrenic
symptomatology. Safety remains a major concern, since chronic
(3) Gray, J. A.; Roth, B. L. Mol. Psychiatry 2007, 12, 904.
(4) Vohora, D. Curr. Opin. InVest. Drugs 2007, 8, 531.
(5) (a) Clark, J. D.; Davis, J. M.; Favor, D.; Fay, L. K.; Franklin, L.;
Henegar, K. E.; Johnson, D. S.; Nichelson, B. J.; Ou, L.; Repine, J. T.;
Walters, M. A.; White, A. D.; Zhu, Z. U.S. Pat. Appl. Publ.
20050043309 A1, 2005. CAN 142:261526. (b) Favor, D. A.; Johnson,
D. S.; Repine, J. T.; White, A. D. PCT Int. Appl. WO 2006103559
A1, 2006. CAN 145:397556. (c) Chen, M. H. G.; Magano, J.; Mennen,
K. E. PCT Int. Appl. WO 2007116265 A1, 2007. CAN 147:448819.
(6) (a) Kaminski, J. J.; Perkins, D. G.; Frantz, J. D.; Solomon, D. M.;
Elliott, A. J.; Chiu, P. J. S.; Long, J. F. J. Med. Chem. 1987, 30, 2047.
(b) Ikemoto, T.; Kawamoto, T.; Wada, H.; Ishida, T.; Ito, T.; Isogami,
Y.; Miyano, Y.; Mizuno, Y.; Tomimatsu, K.; Hamamura, K.; Takatani,
M.; Wakimasu, M. Tetrahedron 2002, 58, 489.
* Author for correspondence. E-mail: Javier.Magano@Pfizer.com.
† Pfizer Global Research and Development, Eastern Point Road, Groton, CT
06340, U.S.A.
‡ Current address: Beijing PharmaSciences, Yizhuang, Beijing, China 100176.
§ Current address: Array BioPharma, 3200 Walnut Street, Boulder, CO 80301,
U.S.A.
(7) (a) Turner, J. A. J. Org. Chem. 1990, 55, 4744. (b) Fenlon, E. E.;
Murray, T. J.; Baloga, M. H.; Zimmerman, S. C. J. Org. Chem. 1993,
58, 6625.
(1) Schultz, S. H.; North, S. W.; Shields, C. G. Am. Fam. Physician 2007,
75, 1821.
(8) Repine, J. T.; Johnson, D. S.; Stuk, T.; White, A. D.; Stier, M. A.; Li,
T.; Yang, Z.; Maiti, S. N. Tetrahedron Lett. 2007, 48, 8189.
(2) Jain, K. K. Exp. Opin. InVest. Drugs 2000, 9, 2935.
10.1021/op800307k CCC: $40.75  2009 American Chemical Society
Published on Web 03/23/2009
Vol. 13, No. 3, 2009 / Organic Process Research & Development
•
555

Scheme 1. Medicinal Chemistry Route to Free Base 13a
a Reagents and conditions: (a) cc NH4OH, 180 °C, 48 h, steel bomb, 64%. (b) Pivaloyl chloride, TEA, CH2Cl2, 0 to 20 °C, 94%. (c) (i) n-BuLi (1.6 M in hexanes),
N,N,N′,N′-tetramethylethylenediamine, THF, -70 °C; (ii) DMF, -70 to 20 °C, 30%. (d) 4-Benzyloxy-1-butanol, NaH, DMF, 0 °C, 67%. (e) PH3P(CH2OCH3)Cl, PhLi,
Et2O, -60 to 20 °C, 80%. (f) 2 N NaOH, EtOH, reflux, 18 h, 74%. (g) (i) Cl3CCONCO, CH2Cl2, 20 °C; (ii) HClO4, Et2O, 20 °C; (iii) NaOH, H2O/MeOH, 90%. (h)
H2, 10% Pd/C, MeOH, 35 °C, 100%. (i) IBX, MeCN, reflux. (j) 12, NaBH(OAc)3, TEA, DCE/THF, 20 °C, 30% (2 steps).
7 as a mixture of E/Z isomers.9 Because both 6 and 7 are oils,
they were purified by chromatography on silica. Pivalamide 7
was treated with dilute NaOH in ethanol at reflux10 to provide
the unstable intermediate aminopyridine 8, which was carried
forward without purification. 8 was then subjected to a one pot,
three-step sequence that involved: (a) the reaction with trichlo-
roacetyl isocyanate11 to give an acyclic trichloroacetylated urea
intermediate; (b) the unmasking of the latent aldehyde func-
tionality with perchloric acid12 followed by intramolecular attack
of the amino group of the urea on the aldehyde to give a
trichloroacetylated seven-membered ring urea intermediate; (c)
the addition of NaOH in methanol to remove the trichloroacetyl
group on the urea and give intermediate 9 in excellent yield (a
more in-depth discussion on this step is presented in Scheme 6
in the Kilo Laboratory Route section below). Hydrogenation
of the double bond and debenzylation to deprotect the hydroxy
group could be done simultaneously to afford alcohol 10 in
quantitative yield. The oxidation of 10 with IBX13 to generate
aldehyde 11 followed by reductive amination14 in the presence
of naphthalenopiperazine 1215 gave the target compound albeit
in low yield after chromatographic purification. As an alternative
to chromatography, 13 could also be purified by recrystallization
from either CH2Cl2 or CHCl3, but the very low recovery and
the toxicity of these two solvents made this protocol impractical.
This synthetic approach consisted of 10 linear steps with a 2%
overall yield.
Our group employed this route to explore the chemistry as
well as to prepare small quantities of API, and several
modifications were quickly implemented to turn it into a safer
process. Thus, 1-methyl-2-pyrrolidinone (NMP) was substituted
for DMF for the preparation of aldehyde 6 to avoid the
potentially dangerous NaH/DMF combination.16 Also, 4 M HCl
in dioxane was employed as a replacement for HClO4 in diethyl
ether. Finally, a different oxidation protocol for the preparation
of aldehyde 11 was developed through the use of SO3 ·pyridine
complex17 or the traditional Swern conditions with oxalyl
chloride18 to avoid the handling of shock-sensitive IBX. Even
after these changes, several limitations were still envisioned that
(9) Maehr, H.; Smallheer, J. M. J. Org. Chem. 1981, 46, 1752.
(10) Warner, P.; Barker, A. J.; Jackman, A. L.; Burrows, K. D.; Roberts,
N.; Bishop, J. A. M.; O’Connor, B. M.; Hughes, L. R. J. Med. Chem.
1992, 35, 2761.
(15) Magano, J.; Akin, A.; Chen, M. H.; Giza, K.; Moon, J.; Saenz, J.
Synth. Commun. 2008, 38, 3631.
(16) For an example on the risks of thermal runaway reactions using the
NaH/DMF combination see: DeWall, G. Chem. Eng. News 1982, 60,
5.
(11) (a) Speziale, A. J.; Smith, L. R. J. Org. Chem. 1963, 28, 1805. (b)
Ramirez, F.; Telefus, C. D. J. Org. Chem. 1969, 34, 376.
(12) Sørensen, U. S.; Bleisch, T. J.; Kingston, A. E.; Wright, R. A.; Johnson,
B. G.; Schoepp, D. D.; Ornstein, P. L. Bioorg. Med. Chem. 2003, 11,
197.
(17) (a) Roush, W. R.; Newcom, J. S. Org. Lett. 2002, 4, 4739. (b) Smith,
A. B.; Adams, C. M.; Lodise Barbosa, S. A.; Degnan, A. P. J. Am.
Chem. Soc. 2003, 125, 350.
(18) (a) Fang, X.; Bandarage, U. K.; Wang, T.; Schroeder, J. D.; Garvey,
D. S. J. Org. Chem. 2001, 66, 4019. (b) Semmelhack, M. F.; Jiang,
Y.; Ho, D. Org. Lett. 2001, 3, 2403.
(13) Frigerio, M.; Santagostino, M. Tetrahedron Lett. 1994, 35, 8019.
(14) Denmark, S. E.; Fu, J. Org. Lett. 2002, 4, 1951.
556
•
Vol. 13, No. 3, 2009 / Organic Process Research & Development

Scheme 2. Convergent Approach for the Synthesis of 13a
a Reagents and conditions: (a) TEA, NMP, 20 °C. (b) K2CO3, MeOH/H2O, 20 °C. (c) 5, KOt-Bu, THF, 0 to 20 °C, 70% (3 steps). (d) Ph3P(CH2OCH3)Cl, PhLi,
THF, -60 to 20 °C, 90%. (e) 50% NaOH, EtOH, reflux. (f) (i) Cl3CC(O)NCO, CH2Cl2, 20 °C; (ii) HClO4, Et2O, 20 °C; (iii) MeOH, 20 °C, 50% (2 steps). (g) H2, Pd/C,
MeOH, 40 °C or NH4HCO2, MeOH, reflux.
would not make this route amenable to scale, such as a high
pressure, time-consuming step in specialized equipment, two
cryogenic steps, four chromatographies, the use of pyrophoric
phenyllithium in diethyl ether, a reductive amination step that
gave lower yields and purities as scale increased and, most
important, the very low overall yield. Thus, when our group
was requested to prepare larger amounts of 13 to support clinical
studies, we decided to look for other approaches to assemble
this molecule.
Alternative Approaches to 13. Since the original process
consisted of a long sequence of linear steps, we first looked at
the possibility of a more convergent route. As a result, the
alternative approach shown in Scheme 2 was implemented.
The synthesis started with the reaction between piperazine
hydrochloride 12 and 4-bromobutyl acetate (14). After a base
and solvent screen, it was found that the best conditions for
this coupling were triethylamine in 1-methyl-2-pyrrolidinone
(NMP) at 20 °C, which gave crude ester 15 as an oily solid.
The hydrolysis of the ester group in 15 was carried out with
potassium carbonate in aqueous methanol to produce alcohol
16 as a dark-brown oil which, without further purification, was
allowed to react with aldehyde 5 to give ether 17 in 70% yield
after chromatography for the three steps combined. Aldehyde
17 was then subjected to the same sequence of steps as in
Scheme 1 to provide unsaturated cyclic urea 20. The final
double-bond reduction step to generate 13 proved to be more
difficult than expected. Several conditions (catalytic hydrogena-
tion, phase transfer hydrogenation) were tried to force the
reaction to completion, but in all cases, the reactions stalled at
90-95% conversion. When higher temperatures or longer
reaction times were employed, more impurity formation was
observed. In addition, the purging of unreacted 20 through
recrystallization could only be accomplished at the expense of
a low recovery. Due to the lack of crystallinity of some of the
intermediates, which prevented the development of efficient
methods of purification, the need for chromatography after the
Wittig reaction to prepare 18, and the inefficient reduction step,
this route was not pursued further.
Other approaches that were tested for the preparation of
the seven-membered ring urea moiety are shown in Scheme 3.
We tried to introduce the two-carbon chain at the 3-position of
the pyridine ring through the use of chloro- or bromoacetalde-
hyde diethyl acetal19 or bromoacetonitrile,20 but only starting
material 4 was recovered, and no desired products 21 or 22
were detected. Also, a Henry reaction21 between aldehyde 5
and nitromethane was attempted to generate nitroalkene 23, but
substrate decomposition was observed.
Kilo Laboratory Route to Phosphoric Acid Salt 1. Based
on these unfavorable results and the pressing need to deliver
large amounts of material for clinical studies, we decided to
optimize the original Medicinal Chemistry route to turn it into
a more robust process. At the same time, the phosphoric acid
salt 1 had been identified as the preferred form of the API, and
a protocol for the reproducible preparation of the targeted solid
form had to be developed as well. After optimization of the
(19) van Dijk, J. T. M.; Hartwijk, A.; Bleeker, A. C.; Lugtenburg, J.;
Cornelisse, J. J. Org. Chem. 1996, 61, 1136.
(20) Frejd, T.; Klingstedt, T. Synthesis 1987, 40.
(21) (a) Hoffman, J. M.; Wai, J. S.; Thomas, C. M.; Levin, R. B.; O’Brien,
J. A.; Goldman, M. E. J. Med. Chem. 1992, 35, 3784. (b) Mogilaiah,
K.; Rao, R. B. Synth. Commun. 2002, 32, 747.
Vol. 13, No. 3, 2009 / Organic Process Research & Development
•
557

3-hydroxyglutaronitrile23 which, in the presence of dry hydrogen
chloride in ether,24 provides the HCl salt of 3. Due to the risks
associated with the generation of HCN and the low overall yield
(28% for the two steps combined), this route was ruled out.
Intermediate 3 was then treated with pivaloyl chloride to
provide pivalamide 4. Screening of inorganic bases (Na2CO3,
K2CO3, KHCO3, NaHCO3) identified NaHCO3 as the base of
choice, and a simple filtration to remove the inorganic salts and
precipitation of the product from heptane afforded material in
excellent yield and purity. A drop in yield was observed when
this process was run on scale compared to laboratory experi-
ments (86% vs 90-94%), where there was essentially no
difference in performance between NaHCO3 and TEA. Even
though the reason for this lower yield is still unclear, a possibility
is a slower crystallization rate on large scale. On the other hand,
the use of NaHCO3 had the advantage that this intermediate
could be isolated without the need for an aqueous workup,
which both simplified the number of operations and reduced
the amount of waste considerably. Another finding during the
kilo laboratory campaign was that, unlike in laboratory experi-
ments, a strong gas evolution was observed during the addition
of the pivaloyl chloride after an induction period. This led to
some foaming and solvent reflux that was easily controlled by
the reflux condenser of the reactor.
The optimization of the subsequent formylation step focused
on both improving the yield and attempting to run the reaction
at higher temperature. Better yield and purity were obtained
when 4-formylmorpholine7b was employed in place of DMF.
During laboratory experiments it was noticed that the addition
of a THF solution of 4-formylmorpholine to the dianion of 4
at -70 °C caused the mixture to thicken considerably, which
prevented a good mixing even with an overhead stirrer. Even
though this same event was observed on scale, an efficient
agitation during the morpholine addition ensured a good mixing
of the reagents, and the integrities of the shaft and blade were
never compromised. Another key factor for the successful
outcome of this reaction was the predrying of the 4-formyl-
morpholine over 4 Å molecular sieves (water content <0.01%
by Karl Fischer analysis), since the presence of water resulted
in incomplete consumption of the starting material and lower
purity. Unfortunately, when the reaction was run at higher
temperature (-40 to -20 °C instead of -70 °C), lower yield
and purity were obtained as well. As alternatives to the
pivalamide group,25 the formylation step was also examined
with the corresponding carbamate or carbonate, but decomposi-
tion was observed in both cases.
Scheme 3. Alternative Approaches for the Preparation of
the Seven-membered Ring Urea
reaction conditions, a modified kilo laboratory route was
implemented, which is shown in Scheme 4.
The optimized synthesis started with 2,6-dichloropyridine
(2), which was converted into 2-amino-6-chloropyridine (3) in
the presence of concentrated NH4OH in a sealed, stainless steel
reactor by heating to 180 °C. Intermediate 3 crystallized from
the solution upon cooling and was easily isolated by filtration.
Up to 10% of 2,6-diaminopyridine was formed as a side
product, but it was easily purged in the filtrates due to its higher
solubility in water. With the goal of accelerating the manufacture
of this intermediate, the progress of the reaction was monitored
at different intervals, and it was found that it was complete after
only 12 h compared to the 48 h that was reported by our
Medicinal Chemistry group, which resulted in an increased
yield. One of the possible reasons for this increase in yield may
be the production of less 2,6-diaminopyridine when the reaction
time is shortened. At the same time, it was found that 180 °C
was the optimal temperature since the reaction proceeded at a
considerably slower rate below this value and hardly proceeded
at all below 120 °C. Another finding was that vigorous stirring
is necessary to push the reaction to completion since in one
instance and, due to a malfunction of the overhead stirrer, the
process stalled at partial conversion. Even though this trans-
formation required the use of specialized equipment, we decided
to proceed with it since it worked well and to outsource this
intermediate for our kilo laboratory campaign. Unfortunately,
due to the limited availability from commercial sources (only
7 out of the 9 kg that was required could be supplied by the
manufacturer at the start of the campaign), we were forced to
run several 200-g batches due to equipment size limitations in
our own laboratories to quickly prepare 2 kg of this material to
meet our needs. The reaction is reproducible up to this scale
under this condition and regardless of the source of 2,6-
diaminopyridine.22
The chloride displacement with 4-benzyloxy-1-butanol was
modified to avoid the use of NaH in combination with solvents
(23) Bergeron, S.; Chaplin, D. A.; Edwards, J. H.; Ellis, B. S. W.; Hill,
C. L.; Holt-Tiffin, K.; Knight, J. R.; Mahoney, T.; Osborne, A. P.;
Ruecroft, G. Org. Process Res. DeV. 2006, 10, 661.
(24) Ferrarini, P. L.; Biagi, G.; Livi, O.; Primofiore, G.; Carpene`, M.
J. Heterocycl. Chem. 1981, 18, 1007.
(25) (a) Fuhrer, W.; Gschwend, H. W. J. Org. Chem. 1979, 44, 1133. (b)
Adelstein, G. W.; Yen, C. H.; Haack, R. A.; Yu, S.; Gullikson, G.;
Price, D. V.; Anglin, C.; Decktor, D. L.; Tsai, H.; Keith, R. H. J. Med.
Chem. 1988, 31, 1215. (c) Conley, R. A.; Barton, D. L.; Stefanick,
S. M.; Lam, M. M.; Lindabery, G. C.; Kasulanis, C. F.; Cesco-Cancian,
S.; Currey, S.; Fabian, A. C.; Levine, S. D. J. Heterocycl. Chem. 1995,
A different approach found in the literature for the prepara-
tion of 3 involves the reaction between epichlorohydrin and a
mixture of NaCN and hydrochloric acid (35%) to generate
´
(22) 2,6-Diaminopyridine was purchased from Sigma-Aldrich Co. and
Lancaster.
32, 761. (d) Ubeda, J. I.; Villacampa, M.; Avendan˜o, C. Synthesis
1998, 1176.
558
•
Vol. 13, No. 3, 2009 / Organic Process Research & Development

Scheme 4. Kilo Lab Route for the Preparation of 1a
a Reagents and conditions: (a) cc NH4OH, 180 °C, 12 h, steel bomb, 77%. (b) Pivaloyl chloride, NaHCO3, CH2Cl2, -5 to 20 °C, 86%. (c) (i) n-BuLi (1.6 M in
hexanes), N,N,N′,N′-tetramethylethylenediamine, THF, -70 to -20 °C; (ii) 4-formylmorpholine, -70 to 20 °C, 49%. (d) 4-Benzyloxy-1-butanol, NaOt-Bu, DMF, 0
to 20 °C. (e) Ph3P(CH2OCH3)Cl, KOt-Bu, THF, 0 °C, 61% (2 steps). (f) 50% NaOH, EtOH, reflux, 1 h. (g) (i) Cl3CCONCO, THF, 20 °C; (ii) 4 M HCl, EtOAc, 5
to 20 °C; (iii) MeOH, 20 °C, 81% (2 steps). (h) H2, 10% Pd/C, MeOH, 40 °C, 95%. (i) MsCl, TEA, THF, 30-35 °C, 84%. (j) NaI, acetone, reflux, 71%. (k) 12, K2CO3,
MeCN, 70 °C, 95%. (l) (i) 1 M H3PO4, MeOH/EtOAc, 50 °C; (ii) cooling to 20 °C over 5 h, 87%.
Scheme 5. Failed Approach for the Preparation of
Since Wittig product 7 is an oil and the use of the crude material
proved inadequate to afford good-quality intermediates in the
following steps, the chromatographic purification of this inter-
mediate was essential. This operation was carried out on silica
and with 4:1 heptane/ethyl acetate as mobile phase.27 Upon
concentrating the fractions that contained desired product, the
residual heptane and ethyl acetate were displaced with toluene,
and olefin 7 was used in the next step as a toluene solution.
The combined yield for the chloride displacement and Wittig
olefination steps was 61%, in comparison with the 53% that
was obtained when intermediate 6 was also purified via
chromatography. We speculate that the difference in yield may
be due to the instability of 6 in the presence of silica. The
instability of this intermediate was further confirmed when
material purified by chromatography darkened and showed
degradation by HPLC analysis after a few weeks on storage.
Wittig product 7 also showed signs of degradation (darkening
and decreased purity by HPLC) upon storage over prolonged
periods of time.
Intermediate 9
such as NMP or DMF. It was observed that sodium tert-
butoxide in DMF performed well in this transformation to afford
ether 6 as a dark-brown oil. In addition, the presence of the
aldehyde group, due to its electron-withdrawing nature, was
fundamental for the success of this transformation, since when
this same reaction was attempted on chloride 26 (Scheme 5),
no detectable amount of 9 was produced even upon heating to
50 °C. In the Medicinal Chemistry route, intermediate 6 was
purified by chromatography since it is an oil. We discovered
that the crude material telescoped as a THF solution was a good
substrate for the subsequent Wittig olefination, and as a result,
the chromatographic purification could be eliminated.
Without isolation, aldehyde 6 underwent a Wittig reaction
with (methoxymethyl)triphenylphosphonium chloride to provide
masked aldehyde 7 as a 3:2 mixture of E/Z isomers. Potassium
tert-butoxide26 proved a good substitute for phenyllithium, and
the reaction could be carried out in THF at 0 °C instead of at
cryogenic temperatures as reported in the original synthesis.
(27) The chromatographic purification conditions were as follows: Column:
Hipersep 600 LP with an internal diameter of 60 cm. Stationary phase:
Merck flash silica (63-200 µm, 80 kg). Mobile phase: 4:1 heptane/
ethyl acetate, with a flow rate of 3 L/min (nitrogen pressure). UV
detection wavelength: 335 nm. Polar impurities were removed in the
first injection, and the fractions that contained early eluting impurities
and desired product were concentrated. A solvent exchange into toluene
was performed to give, after concentration, a final volume of 20 L.
The column was back washed with 3 volumes of acetone and, after
re-equilibration with 4:1 heptane/ethyl acetate, the 20 L from the first
injection was divided into two portions and passed through the column
with an acetone back wash with re-equilibration in between.
(26) (a) Wu¨nsch, B. Arch. Pharm. 1990, 323, 493. (b) Harvey, R. G.; Dai,
Q.; Ran, C.; Penning, T. M. J. Org. Chem. 2004, 69, 2024. (c) Garcı´ia-
Ruano, J. L.; Alema´n, J.; Aranda, M. T.; Ferna´ndez-Iba´n˜ez, M. A.;
Rodrı´guez-Ferna´ndez, M. M.; Maestro, M. C. Tetrahedron 2004, 60,
10067.
Vol. 13, No. 3, 2009 / Organic Process Research & Development
•
559

Scheme 6. Step-wise Sequence for the Preparation of Cyclic Urea 9
The synthesis continued with the hydrolysis of the pivala-
mide group in 7 with 50% sodium hydroxide in ethanol at
reflux. By increasing the base concentration, the reaction time
was reduced from 18 to only 1 h without affecting the purity
or yield of aminopyridine 8. Intermediate 8 is a waxy solid that
also has a strong tendency to darken upon standing, and for
this reason, it was telescoped into the next step as a toluene
solution and used immediately.
With crude 8 in hand, the stage was set for the three-step,
one-pot process for the formation of cyclic urea 9 (Scheme 6).
The synthesis of seven-membered ring ureas has been reported
in the literature. They can be prepared from bromourethanes
and KNH2,28 semicarbazides Via treatment with sodium nitrite
followed by spontaneous decomposition and cyclization of the
resulting nitrene,29 the reaction of 1,4-diamines with triphos-
gene30 or 1,1′-carbonyldiimidazole31 and oxidative rearrange-
ment of bicyclic six-membered ring ureas in the presence of
N-bromosuccinimide.32 In our case, we opted for a novel
approach that involved the treatment of aminopyridine 8 with
trichloroacetyl isocyanate33 in THF at 0 °C to give trichloro-
acetylated urea 27, which could be detected by LC-MS.
Without isolation, to 27 was added a 4 M HCl solution in ethyl
acetate, freshly prepared from acetyl chloride and methanol,34
to afford aldehyde 28.35 Commercially available 4 M HCl
solution in dioxane can also be used to carry out this depro-
tection step, but the use of this solvent was avoided due to its
carcinogenicity. Aldehyde 28 is a short-lived species that
cyclized in situ to give a mixture of cyclic trichloroacetylated
urea 29, also detected by LC-MS, and 9. Since the conversion
of 29 to 9 was a slow process under the reaction conditions,
we investigated ways to promote this transformation. It was
found that a simple addition of methanol to the reaction mixture
at 20 °C was sufficient to complete the removal of the
trichloroacetyl group and give 9 in very good yield. Unlike the
original protocol, the need for a strong base such as NaOH was
circumvented. To the best of our knowledge, this is an
unprecedented, one-pot approach for the synthesis of seven-
membered ring ureas.
The synthesis continued with the simultaneous hydrodeben-
zylation and double bond hydrogenation of 9 over 10% Pd/C
in methanol to give alcohol 10 in almost quantitative yield. This
intermediate has very low solubility in most solvents, and after
a solvent exchange, it could be easily isolated in high yield
and purity from an MTBE slurry. Due to reactor volume
limitations at the time of the campaign, this same protocol had
to be repeated to convert all available substrate.
To avoid the difficult-to-scale reductive amination, it was
decided to replace aldehyde 11 (Scheme 1) with iodide 25
(Scheme 4) in the coupling with piperazine 12. Thus, crude
alcohol 10 was converted to mesylate 24 under standard
conditions.36 The mesylation of 10 only went to completion
when the internal temperature was maintained in the 30-35
°C range during the MsCl addition, whereas at lower temper-
atures the process stalled at approximately 90-95% conversion.
The addition of more MsCl had no appreciable effect on the
conversion of unreacted alcohol. At this point we do not have
a satisfactory explanation for this observation. One possible
reason is the presence of some residual water or methanol from
the previous hydrogenation step, even though the Loss on
Drying test for alcohol 10 showed only a 0.03% weight loss.
(28) Spangler, R. J.; Boop, D. C.; Kim, J. H. J. Org. Chem. 1974, 39,
1368.
(29) Hiremathm, S. P.; Badami, P. S.; Purohit, M. G. Indian J. Chem.,
Sect. B: Org. Chem. Incl. Med. Chem. 1985, 24B, 1115.
(30) Bauvois, B.; Puiffe, M.-L.; Bongui, J.-B.; Paillat, S.; Monneret, C.;
Dauzonne, D. J. Med. Chem. 2003, 46, 3900.
(31) Mayer, P.; Brunel, P.; Chaplain, C.; Piedecoq, C.; Calmel, F.;
Schambel, P.; Chopin, P.; Wurch, T.; Pauwels, P. J.; Marien, M.;
Vidaluc, J.-L.; Imbert, T. J. Med. Chem. 2000, 43, 3653.
(32) Ganguly, A. K.; Wang, C. H.; Biswas, W. D.; Misiaszek, J.; Micula,
A. Tetrahedron Lett. 2006, 47, 5539.
(33) For some recent examples on the use of trichloroacetyl isocyanate for
the synthesis of ureas, see: (a) Pietsch, M.; Gu¨tschow, M. J. Med.
Chem. 2005, 48, 8270. (b) Nishiyama, T.; Isobe, M.; Ichikawa, Y.
Angew. Chem., Int. Ed. 2005, 44, 4372. (c) Hanessian, S.; Marcotte,
S.; Machaalani, R.; Huang, G.; Pierron, J.; Loiseleur, O. Tetrahedron
2006, 62, 5201.
(35) At this point it is unknown whether this intermediate is present in
solution as the aldehyde species or as a cyclic aminal. Additional
studies with techniques such as ReactIR would be necessary to clarify
this point.
(36) Hayakawa, M.; Kaizawa, H.; Kawaguchi, K.; Ishikawa, N.; Koizumi,
T.; Ohishi, T.; Yamano, M.; Okada, M.; Ohta, M.; Tsukamoto, S.;
Raynaud, F. I.; Waterfield, M. D.; Parker, P.; Workman, P. Bioorg.
Med. Chem. 2007, 15, 403.
(34) Nudelman, A.; Bechor, Y.; Falb, E.; Fischer, B.; Wexler, B. A.;
Nudelman, A. Synth. Commun. 1998, 28, 471.
560
•
Vol. 13, No. 3, 2009 / Organic Process Research & Development

Scheme 7. Preparation of Piperazine Hydrochloride 12a
a Reagents and conditions: (a) NaOt-Bu, 1,1′-bis(di-tert-butylphosphinoferrocene) palladium dichloride, toluene, reflux. (b) (i) HCl gas, EtOH, 0 °C to rt; (ii)
Si-thiol silica gel (Pd scavenger), 35 °C, THF/water, 81% (3 steps).
We did not investigate other alternatives such as the tosylate
since mesylates are usually more reactive and MsCl, being a
liquid, is easier to handle on scale than TsCl.
been employed as an efficient means for identifying the
thermodynamically most stable polymorph.38 Stable form
screening experiments exploit the fact that a suspension that is
saturated with respect to a metastable polymorph is supersatu-
rated with respect to the stable polymorph. This supersaturation
is most efficiently removed in solvents in which the compound
of interest is the most soluble, especially those in which the
solubility is greater than 8 mmol.39 Eventually, the least soluble
form will crystallize at the expense of the more soluble
metastable form(s) such that the equilibrium suspension consists
only of the thermodynamically stable form.
Forms A (monohydrate), B (anhydrate), and C (anhydrate)
of phosphate salt 1 had been encountered previously during
small-scale crystallization and salt formation experiments. Form
A was initially recovered from salt formation experiments using
acetonitrile and water mixtures. Scale-up of this procedure from
200 mg to 10-14 g resulted in incomplete salt formation. Form
B of phosphate salt 1 was recovered from small-scale suspen-
sions of Form A in neat ethyl acetate and tetrahydrofuran.
Attempts to reproduce Form B on a scale larger than 100-200
mg were unsuccessful and yielded anhydrous Form C. Forms
A and B were never recovered again as pure phases and were
not fully characterized. It was noted that Forms A and B were
observed to be somewhat high in phosphorous content.
Stable form screening experiments were employed to identify
the thermodynamically stable polymorph as well as establish
preliminary relative stability of the other crystalline forms.
Compound 1 was suspended in 15 different solvents (Table 1)
for two weeks, and the resulting solids were collected and
analyzed using PXRD. Suspensions were prepared using
material that was a mixture of Forms A and C. Form C was
recovered from 13 of these solutions. A dihydrate (Form D)
and a methanol solvate (Form E) were recovered from water
and methanol suspensions, respectively.
Iodide 25 was prepared by treating mesylate 24 with sodium
iodide in acetone at reflux.37 Once the reaction was complete,
25 was precipitated from solution through the addition of water.
Based on previous laboratory work, we were concerned about
the impurity profile of the iodide and the possibility of not being
able to purge some of the byproducts in the next step. As a
result, we decided to upgrade its purity Via recrystallization from
acetonitrile, which gave material of excellent purity (97.5% a/a)
in 71% yield. Even though on laboratory scale the purity of
this intermediate was always g99% by HPLC (a/a), the slightly
lower purity of this batch did not compromise the quality of
the free base 13 produced in the next step.
The second coupling partner to generate 13, piperazine
hydrochloride 12, was synthesized on scale through a procedure
developed in our laboratories that involved a Pd-catalyzed cross-
coupling reaction between 1-bromo-7-fluoronaphthalene (30)
and 1-Boc-piperazine (31) and subsequent deprotection under
acidic conditions (Scheme 7).15
With 12 and 25 in hand, an investigation of the reaction
conditions for their coupling was conducted. Because of the
difficulty to purify free base 13 by recrystallization, the goal
was to develop conditions that provided good quality material
that could be used in the final salt formation step without any
further purification. Thus, several bases (triethylamine, diiso-
propylethylamine, Li2CO3, Na2CO3, K2CO3 and Cs2CO3) and
solvents (MeCN, CHCl3, acetone, DMF, NMP, 1,2-dichloro-
ethane) were screened for this. The best condition was the use
of 1.05 equiv of piperazine 12 in the presence of K2CO3 and
MeCN at 70 °C for 4 h, which minimized impurity build-up.
Once all the iodide had been consumed, the addition of water
precipitated 13 from solution and provided this material in
excellent yield and purity.
Thermodynamically Stable Polymorph Determination.
Prior to the formation of phosphate salt 1, it was desired to
identify the thermodynamically stable polymorph and subse-
quently develop a protocol for its reproducible preparation. The
thermodynamically stable polymorph under ambient conditions
is generally targeted for pharmaceutical development if it is
known. Although this solid form is the least soluble at ambient
conditions, it is also the least likely form to convert during
processing, formulation, scale-up or storage. The accelerated
kinetics of solution-mediated phase transformation have recently
Form C of compound 1 was the only anhydrous/nonsolvated
form recovered from the stable form screen; however, the
solubility in these systems was poor. Therefore, compound 1
was suspended in ethanol at 50 °C for 1 week (solubility )
(38) (a) Cardew, P. T.; Davey, R. J. Proc. R. Soc. London, Ser. A 1985,
398, 415. (b) Rodríguez-Hornedo, N.; Murphy, D. J. Pharm. Sci. 1999,
88, 651. (c) Gu, C.; Young, V.; Grant, D. J. W. J. Pharm. Sci. 2001,
90, 1878. (d) Miller, J. M.; Collman, B. M.; Greene, L. R.; Grant,
D. J. W.; Blackburn, A. Pharm. DeV. Technol. 2005, 10, 291. (e) Gong,
Y.; Collman, B. M.; Mehrens, S. M.; Lu, E.; Miller, J. M.; Blackburn,
A.; Grant, D. J. W. J. Pharm. Sci. 2008, 97, 2130.
(39) The term “mmol” is used in a broad sense in the literature to represent
the minimum solubility needed to ensure confidence in stable form
screening, and it refers to millimol/L. It assures that a certain number
of molecules go into solution rather than a mass.
(37) Dugave, C.; Kessler, P. Tetrahedron Lett. 1994, 35, 9557.
Vol. 13, No. 3, 2009 / Organic Process Research & Development
•
561

Table 1. Stable Polymorph Screen Summary of Compound
1
Table 2. Thermodynamic Stability Relationships Between
Forms C and D at Various aw values
solubility
water
solvent system
activity (aw) initial form recovered form
solvent system
mg/mL
mmol
form
ethanol-water
0.2
0.2
0.2
0.2
0.3
0.3
0.3
0.4
0.5
0.83
0.92
C
C
water
3.5
0.25
1.5
<0.1
<0.1
2
<0.1
1.25
<0.1
<0.1
<0.1
1
<0.1
>11.8
4.64
6.23
0.45
2.67
-
D
C
C
C
C
C
C
C
C
C
C
C
C
E
C
isopropanol-water
ethanol-water
C
C
acetone
C+D
C+D
C+D
D
C+D
C+D
C+D
D
acetonitrile
dichloromethane
ethyl acetate
ethanol
isopropanol-water
ethanol-water
-
isopropanol-water
isopropanol-water
isopropanol-water
methanol-water
isopropanol-water
92% RH (2 weeks)
3.56
-
D
D
hexane
C
D
isopropanol
methyl ethyl ketone
glyme
2.23
-
C
D
C
D
-
-
C
C
methyl tert-butyl ether
tetrahydrofuran
toluene
methanol
ethanol (55 °C)
1.78
-
Form C was stored in a 92% relative humidity (RH) chamber
for 2 weeks (aw ) 0.92). Form conversion to dihydrate was
observed in solutions where aw was 0.4 and higher. No form
changes were observed in solutions below aw of 0.4. Therefore
caw was likely somewhere between 0.2 and 0.4. However, Form
C did not convert to dihydrate in the solid state after being stored
at 92% RH for two weeks.
-
8.27
8.3 mmol) and, from this suspension, Form C was obtained,
further supporting the hypothesis that it was the stable anhydrous
form of compound 1.
Form D (dihydrate) and Form E (methanol solvate) of
compound 1 were the thermodynamically stable forms in a pure
water and methanol environments, respectively. Form E was
relevant from a synthesis standpoint, and its preparation from
Form C was investigated. Thus, Form C was suspended in 6.3,
10 and 20% methanol solutions in ethyl acetate. Only Form C
was recovered, indicating that a substantial amount of methanol
was required to convert it to Form E. A methanol solvate was
then avoidable and would not be considered for dosing in
humans.
The dihydrate (Form D) of compound 1 was potentially more
problematic. Solid-state transition to Form D dihydrate may
occur since water is ubiquitous in the environment. Furthermore,
stable hydrates are less soluble in water relative to anhydrous
forms of the same material. Although the magnitude of the
aqueous solubility difference between Forms C and D was
unknown, there was a strong desire to avoid the potential
decrease in bioavailability of the stable hydrate.
The hydration state of an anhydrous/hydrate system depends
on the water activity (abbreviated aw) of the surrounding system
in relation to the critical water activity (abbreviated caw) of
hydrate formation. The caw of a hydrate/anhydrous pair at a
given temperature is a state thermodynamic variable. If the caw
for compound 1 was hypothetically 0.45 at 25 °C, then exquisite
thermodynamic control could be used to recover and store the
desired form. For example, crystallization conditions with aw
< 0.45 would favor anhydrous Form C. Furthermore, storage
of Form C in relative humidity <45% would guarantee no form
change to dihydrate in this scenario.
Project timelines and compound availability in early
development necessitated that a caw range was identified
in lieu of definition of the finite thermodynamic point.
The hypothetically low caw range presented possible
challenges for storage of desired Form C of compound 1.
However, solid-state transformation to dihydrate was not
seen even though the compound was stored well above
the hypothetical caw. Furthermore, the relatively fast
kinetics of transformation in solution failed to induce
transformation to dihydrate when aw was less than 0.4.
In summary, Form C was robustly prepared in anhydrous
solvents, Forms A and B easily converted to Form C, and Form
D was thermodynamically stable in water activities relevant to
environmental conditions (see Figure 2 for a summary of the
stability relationship between Forms A, B, C and D). However,
the kinetics of transformation were slow, especially in the solid
state. Routine desiccated storage conditions could then be
employed to maintain humidity such that the desired anhydrous
form of compound 1 was preserved. Therefore, Form C was
identified as the target solid form for early development.
With this information in hand, the preparation of Form C
of the phosphoric acid salt of 1 was carried out by dissolving
free base 13 in ethyl acetate at 50-55 °C followed by the
addition of a 1 M solution of phosphoric acid in methanol. Due
to its poor solubility in this medium, the salt crystallized
immediately from the solution. and after a slow, controlled
Water activities in binary aqueous and organic solutions have
been described in the literature as an effective method to
determine caw of a hydrate/anhydrous system.40 Forms C and
D were suspended in solutions with aw ranging from 0.2 to
0.83 for approximately 24 h (Table 2). Additionally, anhydrous
(40) (a) Grant, D. J. W.; Higuchi, T. Solubility BehaVior of Organic
Compounds; Wiley: New York, 1990. (b) Ghosh, S.; Grant, D. J. W.
Int. J. Pharm. 1995, 114, 185. (c) Zhu, H.; Yuen, C.; Grant, D. J. W.
Int. J. Pharm. 1996, 135, 151. (d) Zhu, H.; Grant, D. J. W. Int.
J. Pharm. 1996, 139, 33.
Figure 2. Thermodynamic stability relationship between an-
hydrous and hydrated forms of compound 1.
562
•
Vol. 13, No. 3, 2009 / Organic Process Research & Development

cooling to 20 °C over 5 h, the final API was obtained in
was added, and the distillation was resumed until the volume
in the reactor was 22 L. This protocol was repeated once more
to a final volume of 22 L, and the resulting suspension was
stirred at 20 °C for 1 h. The solid was filtered, washed with
heptane (12 L) and dried under vacuum (15 Torr) at 40 °C to
give 12.80 kg (86%) of pivalamide 4 as a pale-yellow solid.
Mp: 85-87 °C. HPLC retention time: 15.25 min. HPLC purity:
excellent yield and purity.
Conclusions
In conclusion, an optimized protocol for the preparation
of D2 partial agonist 1 has been developed that overcomes
many of the synthetic limitations present in the original
route, such as the low overall yield, the use of unsafe
reagents and conditions and the need for multiple chro-
matographic purifications. However, further work would
have to be carried out to turn it into a truly scalable
process, which includes alternatives for the preparation
of 2-amino-6-chloropyridine (3) under milder conditions
and without resorting to the use of specialized equipment,
the elimination of the remaining chromatography and the
generation of aldehyde 5 under noncryogenic temperatures.
In addition, a more convergent route would be desirable.
But, at the same time, the improvements that were
implemented during the kilo-laboratory campaign in-
creased the overall yield from 2 to 10% and allowed for
the quick preparation of multikilogram quantities of API
to make possible the timely advance of the associated drug
development program. Due to the discontinuation of this
program, no further work will be carried out in the near
future to address these issues.
1
99.8% a/a. H NMR (400 MHz, CDCl3) δ 1.26 (s, 9H),
6.96-7.01 (m, 1H), 7.59 (d, J ) 7.87 Hz, 1H), 7.96 (br s, 1H),
8.12 (d, J ) 8.24 Hz, 1H). 13C NMR (100 MHz, CDCl3) δ
27.5, 40.0, 112.1, 119.7, 141.1, 148.9, 151.6, 177.2. HRMS
(EI) exact mass calcd for C10H14ClN2O (M + H) 213.0795,
found 213.0788.
Synthesis of Aldehyde 5. To a solution of pivalamide 4
(12.75 kg, 60.2 mol) and N,N,N′,N′-tetramethylethylenediamine
(7.0 kg, 60.2 mol) in THF (154 L) at -70 °C under nitrogen
was added n-BuLi (1.6 M in hexanes, 94.0 L, 150.5 mol) while
the internal temperature was held below -60 °C. The mixture
was warmed to -20 °C over 2 h and stirred at this temperature
for an additional 2 h. The mixture was then cooled to -70 °C,
and a solution of 4-formylmorpholine (20.8 kg, 180 mol) in
THF (13 L) was slowly added while the internal temperature
was maintained below -60 °C. The resulting thick suspension
was warmed to 20 °C and stirred for 18 h. The dark brown
mixture was cooled to 0 °C, and saturated ammonium chloride
(80 L) and 1 M HCl (12 L) were added to bring the pH down
to 7 ( 0.5 while the internal temperature was held below 10
°C. The two layers were allowed to separate, and the aqueous
layer was extracted with toluene (64 L). The aqueous layer was
extracted with additional toluene (32 L), and the combined
organic extracts were concentrated at reduced pressure until the
volume reached 45 L. Toluene (29 L) was added to the mixture,
and the distillation was resumed until the volume reached 45
L. Hexanes (64 L) was added over 30 min, and the resulting
suspension was stirred at 20 °C for 30 min. The solid was
filtered, washed with a 2:1 MTBE/hexanes mixture (51 L) and
dried under vacuum (15 Torr) at 20 °C to give 7.15 kg (49%)
of aldehyde 5 as a tan solid. Mp: 137-138 °C. HPLC retention
time: 10.11 min. HPLC purity: 92.7% a/a. 1H NMR (400 MHz,
CDCl3) δ 1.31 (s, 9H), 7.14 (d, 1H), 8.01 (d, J ) 8.06 Hz,
1H), 9.87 (s, 1H), 10.87 (br s, 1H). 13C NMR (100 MHz,
CDCl3) δ 27.5, 40.8, 116.6, 119.3, 145.3, 151.9, 155.9, 176.7,
192.2. HRMS (EI) exact mass calcd for C11H14ClN2O2 (M +
H) 241.0744, found 241.0738.
Synthesis of Aldehyde 6. To a solution of 4-benzyloxy-1-
butanol (4.09 kg, 22.7 mol) in DMF (19 L) at 0 °C under
nitrogen was added a freshly prepared solution of NaOt-Bu
(6.55 kg, 68.1 mol) in DMF (16 L) while the internal
temperature was held below 10 °C. The resulting mixture was
stirred at 0-5 °C for 30 min, and a solution of aldehyde 5 (7.10
kg, 29.5 mol) in DMF (14.3 L) was slowly added while the
internal temperature was maintained below 10 °C. The reaction
was warmed to 20 °C and stirred for 1 h. The mixture was
cooled to 5 °C, and water (103 L) and MTBE (82 L) were
added. The layers were separated, and the aqueous layer was
extracted with MTBE (36 L). The combined organic extracts
were washed with water (38 L) and saturated brine (44 L), and
the solvent was removed at reduced pressure until the batch
Experimental section
Reaction completion and product purity were evaluated by
HPLC using the following conditions: column: YMC PackPro
C18, 150 mm × 4.6 mm, 3 µm; wavelength: 215 nm;
temperature: 25 °C; injection volume: 10 µL after diluting the
sample with 50:50 water/acetonitrile to a concentration of
approximately 0.35 mg/mL of product; eluent (A) water (0.1%
TFA) (B) ACN (0.1% TFA), gradient: (0 min) (A) 70%, (B)
30%; (30 min) (A) 20% (B) 80%; (35 min) (A) 20% (B) 80%;
(35.1 min) (A) 70% (B) 30%; (45 min) (A) 70% (B) 30%.
Synthesis of 2-Amino-6-chloropyridine (3). A 4-L stainless
steel reactor was charged with 2,5-dichloropyridine (2) (200.0
g, 1.35 mol) and aqueous, concentrated ammonium hydroxide
(1 L). The reactor was sealed, and the mixture was heated at
180 °C for 12 h with good stirring. The contents of the reactor
were cooled to 20 °C, and the resulting solid was filtered,
washed with water (3 × 100 mL) and dried under vacuum (15
Torr) at 45 °C to give 134.1 g (77%) of 3 as a light green solid.
Mp: 72-75 °C. HPLC retention time: 2.67 min. HPLC purity:
99.6% a/a. 1H NMR (300 MHz, CDCl3) δ 5.12 (br s, 2H), 6.35
(d, J ) 7.91 Hz, 1H), 6.58 (d, J ) 7.58 Hz, 1H), 7.31 (t, J )
7.75 Hz, 1H). 13C NMR (75 MHz, CDCl3) δ 106.8, 112.9,
140.2, 149.4, 159.1. MS (ES+): 129.4 (M + H)+.
Synthesis of Pivalamide 4. To a mixture of 2-amino-6-
chloropyridine (3) (9.00 kg, 70 mol) and sodium bicarbonate
(11.76 kg, 140 mol) in CH2Cl2 (87 L) at -5 °C under nitrogen
was added pivaloyl chloride (12.67 kg, 105 mol) while the
internal temperature was held below 25 °C. The resulting
mixture was then stirred at 20 °C for 5 h. The suspension was
filtered to remove the inorganic solids, and these solids were
washed with CH2Cl2 (16 L). The filtrates were concentrated at
reduced pressure to a final volume of 27 L. Heptane (18 L)
Vol. 13, No. 3, 2009 / Organic Process Research & Development
•
563

volume reached 24 L. THF (30 L) was added, and the
distillation was resumed until the batch volume reached 28 L
to give a THF solution of aldehyde 6 that was used without
further purification in the next step. HPLC retention time: 27.55
min. HPLC purity: 60.5% a/a. An analytical sample was purified
by flash chromatography on silica (heptane/ethyl acetate 9/1 as
mobile phase) to fully characterize this intermediate. 1H NMR
(400 MHz, CDCl3) δ 1.35 (s, 9H), 1.73-1.93 (m, 4H), 3.52
(d, J ) 6.41 Hz, 2H), 4.52 (s, 4H), 6.46 (d, 1H), 7.21-7.35
(m, 5H) 7.78 (d, J ) 8.24 Hz, 1H) 9.69 (s, 1H) 11.49 (br s,
1H). 13C NMR (100 MHz, CDCl3) δ 25.9, 26.6, 27.6, 27.7,
41.1, 67.5, 70.1, 73.0, 105.5, 111.0, 127.7, 127.8, 128.5, 138.7,
145.7, 153.2, 167.0, 176.9, 191.8. HRMS (EI) exact mass calcd
for C22H28N2O4 (M + H) 385.2127, found 385.2125.
with water (3 × 18 L) and saturated brine (20 L) and dried
over MgSO4 (5 kg). The solution was filtered, and the solvent
was removed at reduced pressure until the batch volume reached
10 L. Toluene (30 L) was added, and the solution was
concentrated at reduced pressure until the batch volume reached
10 L. Karl Fischer analysis showed 0.05% water content. This
toluene solution of aminopyridine 8 (3/2 E/Z mixture of isomers)
was used in the next step without further purification. HPLC
retention times: 11.09 and 11.26 min. HPLC purity: 90.8% a/a.
An analytical sample was purified by flash chromatography on
silica (CH2Cl2/ethyl acetate 9/1) to fully characterize this
1
intermediate. H NMR (400 MHz, CDCl3) δ 1.72-1.91 (m,
4H), 3.51-3.58 (m, 2H), 3.66 (s, 1.2H) 3.73 (s, 1.8H),
4.14-4.22 (m, 2H), 4.30-4.85 (m, 4H), 5.02 (d, J ) 6.96 Hz,
0.6H), 5.55 (d, J ) 12.64 Hz, 0.4H) 6.04-6.12 (m, 1H), 6.73
(d, J ) 12.64 Hz, 0.4H), 7.22-7.39 (m, 5H), 7.65 (d, J ) 8.24
Hz, 0.6H). 13C NMR (100 MHz, CDCl3) δ 26.2, 26.6, 27.4,
56.8, 60.6, 65.9, 66.0, 70.2, 73.1, 98.8, 99.0, 99.4, 100.4, 107.4,
108.3, 127.7, 127.8, 128.6, 138.2, 138.8, 140.9, 146.3, 149.7,
154.1, 154.4, 162.0, 162.3. HRMS (EI) exact mass calcd for
C19H25N2O3 (M + H) 329.1865, found 329.1859.
Synthesis of Urea 9. Reactor A: to a mixture of methanol
(4 L) and ethyl acetate (12 L) cooled at 10 °C under nitrogen
was slowly added acetyl chloride (5.92 L, 83.3 mol) while
maintaining the internal temperature below 25 °C. The resulting
mixture was stirred at 20 °C for 1 h to give a 4 M HCl solution
in ethyl acetate.
Reactor B: to the toluene solution of 8 that was obtained in
the previous step was added THF (61 L), and the mixture was
cooled to 0 °C. Trichloroacetyl isocyanate (3.77 kg, 20.0 mol)
was slowly added while maintaining the internal temperature
below 6 °C. The resulting mixture was warmed to 20 °C for
1 h. HPLC analysis showed complete consumption of amine 8
to give acyclic urea 27 (HPLC retention times: 28.92 and 29.06
min). The mixture was cooled to 5 °C, and the 4 M HCl solution
in ethyl acetate previously prepared in Reactor A was transferred
to Reactor B while the internal temperature was held below 10
°C. The mixture was warmed to 20 °C and stirred for 24 h.
HPLC analysis showed full consumption of urea 27 and a 3/1
mixture of 9 and acetylated urea 29 (HPLC retention time: 24.76
min). Methanol (28 L) was added, and the reaction was stirred
for an additional 8 h at 20 °C to fully convert intermediate 29
to 9. One molar NaOH (80 L) was added to adjust the pH to 7,
followed by the addition of ethyl acetate (65 L). The layers
were separated, and the aqueous layer was extracted with ethyl
acetate (15 L). The combined organic extracts were washed
with saturated brine (25 L), and the solvent was removed at
reduced pressure until the batch volume reached 10 L. Methanol
(71 L) was added followed by water (35.4 L) over 30 min.
The suspension was stirred at 20 °C for 30 min, and the solid
was filtered, washed with a 1:1 methanol/water mixture (22 L)
and dried under vacuum (15 Torr) at 40 °C to give 4.66 kg
(77%, two steps) of cyclic urea 9 as a beige solid. Mp: 79-80
°C. HPLC retention time: 22.63 min. HPLC purity: 98.5% a/a.
1H NMR (400 MHz, CDCl3) δ 1.79-1.90 (m, 2H), 1.90-2.01
(m, 2H), 3.57 (t, J ) 6.23 Hz, 2H), 4.30 (d, J ) 6.23 Hz, 2H),
4.54 (s, 2H), 6.47 (d, J ) 3.85 Hz, 1H), 6.56 (br s, 1H), 6.67
(d, J ) 8.61 Hz, 1H), 7.24-7.39 (m, 5H), 7.74-7.82 (m, 2H),
Synthesis of Methyl Vinyl Ether 7. To a suspension of
(methoxymethyl)triphenylphosphonium chloride (11.34 kg,
33.0 mol) in dry THF (13.8 L) at 0 °C under nitrogen
was added a 1 M solution of KOt-Bu in THF (57.3 L,
57.3 mol) while the internal temperature was maintained
below 5 °C. The resulting deep-red mixture was stirred
at 0-5 °C for 30 min, and the THF solution of aldehyde
6 obtained in the previous step was slowly added while
maintaining the internal temperature below 5 °C. After
15 min, aqueous, saturated ammonium chloride (36 L),
water (31 L) and MTBE (42 L) were added, and the layers
were separated. The organic phase was washed with
saturated brine (25 L), and the solvent was removed under
reduced pressure until the batch volume reached 25 L.
This crude material was chromatographed on silica
(heptane/ethyl acetate 4/1 as mobile phase),27 and the
fractions that contained product were concentrated at
reduced pressure. Residual heptane and ethyl acetate were
displaced with toluene (2 × 35 L) to give 7.38 kg (61%,
two steps; yield calculated based on Loss on Drying test
of a small sample) of 7 as a bright-yellow toluene solution
(27.45 kg). Olefin 7 was obtained as a 3/2 E/Z mixture of
isomers. HPLC retention times: 23.18 and 23.53 min.
1
HPLC purity: 95.1% a/a. H NMR (400 MHz, CDCl3) δ
1.28-1.35 (m, 9H), 1.71-1.90 (m, 4H), 3.53 (t, J ) 6.23
Hz, 2H), 3.61 (s, 1H), 3.72 (s, 2H), 4.18-4.30 (m, 2H),
4.50 (s, 2H), 5.07 (d, J ) 6.96 Hz, 0.6H), 5.57 (d, J )
13.00 Hz, 0.4H), 6.11 (d, J ) 7.14 Hz, 0.6H), 6.50-6.56
(m, 1H), 6.83 (d, J ) 13.00 Hz, 0.4H), 7.22-7.38 (m,
4.8H), 7.50-7.57 (m, 1H), 7.77 (br s, 0.6H), 8.03 (d, J )
8.42 Hz, 0.6H). 13C NMR (100 MHz, CDCl3) δ 26.1, 26.6,
27.8, 39.7, 56.3, 60.7, 66.1, 66.2, 70.2, 73.0, 100.4, 100.8,
107.8, 108.6, 117.9, 119.5, 127.7, 127.8, 128.5, 137.4,
138.8, 141.3, 144.5, 144.7, 147.1, 149.1, 161.3, 161.6,
176.6, 176.7. HRMS (EI) exact mass calcd for C24H33N2O4
(M + H) 413.2440, found 413.2433.
Synthesis of Aminopyridine 8. To the toluene solution of
7 obtained in the previous step was added ethanol (72 L)
followed by 50% NaOH solution in water (13.2 kg) while the
internal temperature was held below 70 °C. The hazy mixture
was heated to reflux for 1 h, and after the mixture cooled to 40
°C, toluene (38 L) and water (70 L) were added. The layers
were separated, and the aqueous layer was extracted with
toluene (38 L). The combined organic extracts were washed
564
•
Vol. 13, No. 3, 2009 / Organic Process Research & Development

9.13 (br s, 1H). 13C NMR (100 MHz, CDCl3) δ 26.0, 26.7,
66.4, 70.1, 73.1, 104.0, 106.5, 116.9, 122.9, 127.8, 128.6, 133.0,
138.7, 144.0, 152.7, 160.6. Anal. Calcd for C19H21N3O3: C:
67.24; N: 12.38; H: 6.24. Found: C: 67.03; N: 12.21; H: 6.17.
Synthesis of Alcohol 10. A Parr shaker reactor was charged
with a solution of urea 9 (150 g, 442 mmol) in methanol (1.6
L). Pd/C, 10% (Johnson Matthey, 1940 carbon (shell, unre-
duced, water wet), 15 g) was added, and the contents of the
reactor were pressurized and shaken with hydrogen gas to 50
psig and heated to 35 °C. When the hydrogen uptake ceased
(usually after 3 h), the mixture was filtered, and the catalyst
cake was washed with methanol (250 mL). Due to reactor
volume limitations, this same protocol was repeated to convert
all available substrate. All the filtrates were combined together
to give a crude solution of alcohol 10 in methanol. The solvent
was removed at reduced pressure until the batch volume reached
4 L. MTBE (13.5 L) was added over 10 min and the resulting
suspension was cooled to 10 °C and stirred for 1 h. The solid
was filtered, washed with MTBE (4 L) and dried under vacuum
(15 Torr) at 40 °C to give 2.96 kg (90%) of alcohol 10 as a
white solid. Mp: 116-118 °C. HPLC retention time: 3.05 min.
was added over 1 h, and the resulting suspension was stirred at
20 °C for 30 min. The solid was filtered, washed with water
(20 L) and dried under vacuum (15 Torr) at 40 °C for 12 h to
give crude iodide 25. To crude 25 was added acetonitrile (14
L), and the suspension was heated to 60 °C to give a clear
solution. The solution was cooled to 10 °C over 1 h, and the
solid was filtered, washed with acetonitrile (3 L) and dried under
vacuum (15 Torr) at 40 °C to give 1.78 kg (71%) of iodide 25
as an off-white solid. Mp: 105-107 °C. HPLC retention time:
1
15.68 min. HPLC purity: 97.5% a/a. H NMR (400 MHz,
CDCl3) δ 1.73-1.83 (m, 2H), 1.86-1.97 (m, 2H), 2.89 (t, J )
8.79 Hz, 2H), 3.16 (t, J ) 6.87 Hz, 2H), 3.95-4.05 (m, 2H),
4.09 (t, J ) 6.13 Hz, 2H), 6.13 (d, J ) 8.06 Hz, 1H), 6.48 (br
s, 1H), 7.25 (d, J ) 7.87 Hz, 1H), 8.31 (br s, 1H). 13C NMR
(100 MHz, CDCl3) δ 6.8, 23.7, 30.0, 30.4, 46.0, 65.3, 101.7,
116.1, 136.1, 154.5, 155.8, 162.1. Anal. Calcd for C12H16IN3O2:
C: 39.90; N: 11.63; H: 4.47. Found: C: 39.89; N: 11.39; H:
4.43.
Synthesis of Free Base 13. A mixture of iodide 25 (1.77
kg, 4.9 mol), piperazine 12 (1.37 kg, 5.14 mol) and potassium
carbonate (2.03 kg, 14.7 mol) in acetonitrile (21.6 L) was heated
to 70 °C under nitrogen. After 4 h, the reaction was cooled to
40 °C, and water (21.8 L) was added over 30 min. The resulting
slurry was stirred at 20 °C for 30 min, and the solid was filtered,
washed with a 1:1 water/acetonitrile mixture (5.4 L) and dried
under vacuum (15 Torr) at 40 °C to give 2.17 kg (95%) of 13
as a white solid. Mp: 120-121 °C. HPLC retention time: 8.86
min. HPLC purity: 98.7% a/a. 1H NMR (400 MHz, CDCl3) δ
1.68-1.93 (m, 4H), 2.50-2.60 (m, 2H), 2.78 (br s, 4H), 3.00
(t, J ) 8.92 Hz, 2H), 3.17 (br s, 4H), 4.07-4.17 (m, 2H), 4.23
(t, J ) 6.22 Hz, 2H), 5.19 (br s, 1H), 6.26 (d, J ) 7.88 Hz,
1H), 7.13 (d, J ) 7.47 Hz, 1H), 7.19-7.29 (m, 1H), 7.32-7.41
(m, 2H), 7.55 (d, J ) 8.30 Hz, 1H), 7.75-7.87 (m, 2H), 8.53
(br s, 1H). 13C NMR (100 MHz, CDCl3) δ 23.8, 27.2, 46.0,
53.0, 53.9, 58.5, 66.4, 101.9, 107.3, 107.5, 116.0, 116.1, 116.4,
123.6, 125.3, 130.9, 131.8, 136.2, 149.4, 154.6, 155.4, 159.6,
162.0, 162.4. Anal. Calcd for C26H30FN5O2: C: 67.37; N: 15.11;
H: 6.52; F, 4.10. Found: C: 67.45; N: 14.46; H: 6.56; F, 3.94.
HRMS (EI) exact mass calcd for C26H31FN5O2 (M + H)
464.2462, found 464.2457.
Synthesis of Phosphoric Acid Salt 1. Preparation of the
phosphoric acid solution in methanol: Phosphoric acid
(0.45 kg, 4.59 mol) was added to methanol (4.6 L) at 20
°C under nitrogen. The mixture was stirred until all the
solids dissolved to give a 1 M solution of phosphoric acid
in methanol.
To a solution of 13 (2.15 kg, 4.63 mol) in ethyl acetate (64.5
L) at 50-55 °C under nitrogen was added the 1 M solution of
phosphoric acid in methanol previously prepared while the
internal temperature was held in the 50-55 °C range. The
resulting slurry was stirred for 20 min at 50-55 °C and then
cooled to 20 °C over 5 h. The solid was filtered, washed with
ethyl acetate (2 × 8 L) and dried under vacuum (15 Torr) at
50 °C to give 2.28 kg (87%) of salt 1 as a white solid. Mp
(DSC): 192.7 °C. HPLC purity: 98.8% a/a. 1H NMR (400 MHz,
DMSO-d6) δ 1.63-1.84 (m, 4H), 2.64-2.77 (m, 2H), 2.81-3.02
(m, 5H), 3.12 (br s, 4H), 3.91 (t, J ) 8.71 Hz, 2H), 4.18 (t, J
) 6.01 Hz, 2H), 6.27 (d, J ) 7.88 Hz, 1H), 7.04 (br s, 1H),
1
HPLC purity: 99.5% a/a. H NMR (400 MHz, DMSO-d6) δ
1.45-1.58 (m, 2H), 1.65-1.77 (m, 2H), 2.90 (t, J ) 8.79 Hz,
2H), 3.39-3.47 (m, 2H), 3.90 (t, J ) 8.79 Hz, 2H), 4.12 (t, J
) 6.59 Hz, 2H), 4.46 (t, J ) 5.13 Hz, 1H), 6.22 (d, J ) 7.87
Hz, 1H), 6.98 (br s, 1H), 7.45 (d, J ) 8.06 Hz, 1H), 7.98 (br
s, 1H). 13C NMR (100 MHz, DMSO-d6) δ 23.5, 25.9, 29.6,
46.2, 61.0, 66.5, 101.1, 116.7, 137.0, 154.7, 154.9, 162.2. Anal.
Calcd for C12H17N3O3: C: 57.36; N: 16.72; H: 6.82. Found: C:
57.25; N: 16.40; H: 6.74.
Synthesis of Mesylate 24. To a solution of alcohol 10
(2.14 kg, 8.50 mol) and triethylamine (3.55 L, 25.5 mol)
in THF (86 L) at 25-30 °C under nitrogen was added
methanesulfonyl chloride (1.46 kg, 12.7 mol) while
maintaining the internal temperature in the 30-35 °C
range. After 30 min, water (9.8 L) was added, and the
suspension was concentrated at reduced pressure until the
batch volume reached 40 L. Water (58 L) was added over
1 h, and the suspension was stirred at 20 °C for 30 min.
The solid was filtered, washed with water (20 L) and dried
under vacuum (15 Torr) at 40 °C to give 2.33 kg (84%)
of mesylate 24 as an off-white solid. HPLC retention time:
7.05 min. HPLC purity: 89.8% a/a. An analytical sample
was purified by flash chromatography on silica (CH2Cl2/
ethyl acetate 95/5) to fully characterize this intermediate.
Mp: 158-160 °C. 1H NMR (300 MHz, DMSO-d6) δ
1.71-1.91 (m, 4H), 2.94 (t, J ) 8.57 Hz, 2H), 3.20 (s,
3H), 3.93 (t, J ) 8.74 Hz, 2H), 4.09-4.35 (m, 4H), 6.28
(d, J ) 7.91 Hz, 1H), 7.00 (br s, 1H), 7.51 (d, J ) 7.91
Hz, 1H), 7.98 (br s, 1H). 13C NMR (75 MHz, DMSO-d6)
δ 23.5, 25.3, 26.1, 37.2, 46.3, 66.1, 70.8, 101.1, 117.0,
137.1, 154.7, 162.1. HRMS (EI) exact mass calcd for
C13H20N3O5 (M + H) 330.1124, found 330.1115.
Synthesis of Iodide 25. A mixture of mesylate 24 (2.32 kg,
7.0 mol) and sodium iodide (3.17 kg, 21.1 mol) in acetone (45.5
L) under nitrogen was heated at reflux for 3 h. The suspension
was cooled to 30 °C, and the solvent was removed at reduced
pressure until the batch volume reached 25 L. Water (53.5 L)
Vol. 13, No. 3, 2009 / Organic Process Research & Development
•
565

7.20 (d, J ) 7.47 Hz, 1H), 7.37-7.46 (m, 2H), 7.49 (d, J )
7.88 Hz, 1H), 7.66 (d, J ) 7.88 Hz, 1H), 7.73 (dd, J ) 11.20,
2.49 Hz, 1H), 7.88-8.09 (m, 2H), 9.25 (br s, 3H). 13C NMR
(100 MHz, DMSO-d6) δ 22.4, 23.5, 26.8, 46.3, 51.9, 53.0, 57.2,
66.3, 101.2, 107.2, 107.4, 116.5, 116.9, 124.2, 126.1, 129.8,
129.9, 132.0, 132.1, 137.1, 154.7, 155.0, 159.5, 162.1. Anal.
Calcd for C26H33FN5O6P: C: 55.61; N: 12.47; H: 5.92; F, 3.38;
P, 5.52. Found: C: 55.21; N: 12.32; H: 5.89; F, 3.35; P, 5.24.
for the purification of intermediate 7, the Ann Arbor kilo
laboratory staff for their work during the preparation of the first
clinical batch and Dr. Kevin Henegar for reviewing this
manuscript.
Supporting Information Available
Copies of 1H and 13C NMR spectra for compounds 1, 3, 4,
5, 6, 7, 8, 9, 10, 13, 24 and 25 and PXRD data of compound
1. This material is available free of charge via the Internet at
http://pubs.acs.org.
Acknowledgment
We thank Norm Colbry for the preparation of 2-amino-6-
chloropyridine, Brian Moon for obtaining the safety data,
Ronald Morris for obtaining the HRMS data, Ming Zeng and
Larry Gibson from the Separation Sciences group in Groton
Received for review December 12, 2008.
OP800307K
566
•
Vol. 13, No. 3, 2009 / Organic Process Research & Development