Efficient synthesis of a highly selective NPY-5 receptor antagonist: A drug candidate for the treatment of obesity

Article abstract of DOI:10.1021/op0600963

A concise and practical synthesis of highly selective NPY-5 receptor antagonist 1 is described. The animopyrazine intermediate 3 was synthesized via either monobromination of aminopyrazine or palladium-catalyzed regioselective debromination of dibromopyrazine followed by an efficient Suzuki-Miyaura coupling. For the preparation of the spirolactone piperidine 2, significantly improved yield was achieved by using a combination of n-BuMgCl and n-BuLi. This protocol also dramatically increased the thermal stability of the aryllithium intermediate and eliminated the requirement for costly cryogenic conditions. The union of the spirolactone piperidine 2 and aminopyrazine 3 via a carbonyl group was accomplished using phenyl chloroformate delivering the target molecule in high yield.

Full text of DOI:10.1021/op0600963

Organic Process Research & Development 2006, 10, 822−828
Efficient Synthesis of a Highly Selective NPY-5 Receptor Antagonist: A Drug
Candidate for the Treatment of Obesity
Takahiro Itoh,* Shinji Kato, Nobuaki Nonoyama, Toshihiro Wada, Kenji Maeda, and Toshiaki Mase
Process R & D, Banyu Pharmaceutical Co. Ltd, Kamimutsuna 3-Chome-9-1, Okazaki, Aichi 444-0858, Japan
Matthew M. Zhao,* Jake Z. Song, David M. Tschaen, and James M. McNamara
Process Research, Merck Research Laboratories, Merck & Co. Inc., Rahway, New Jersey 07065, U.S.A.
Abstract:
oligonucleotides or NPY antibodies acutely decreases feed-
ing.
3
A concise and practical synthesis of highly selective NPY-5
receptor antagonist 1 is described. The animopyrazine inter-
mediate 3 was synthesized via either monobromination of
aminopyrazine or palladium-catalyzed regioselective debromi-
nation of dibromopyrazine followed by an efficient Suzuki-
Miyaura coupling. For the preparation of the spirolactone
piperidine 2, significantly improved yield was achieved by using
a combination of n-BuMgCl and n-BuLi. This protocol also
dramatically increased the thermal stability of the aryllithium
intermediate and eliminated the requirement for costly cryo-
genic conditions. The union of the spirolactone piperidine 2 and
aminopyrazine 3 via a carbonyl group was accomplished using
phenyl chloroformate delivering the target molecule in high
yield.
Compound 1 is a highly selective NPY receptor antagonist
with >4000-fold selectivity for the human Y5 receptors over
other types of NPY receptors. It is being investigated for
the prevention and treatment of diabetes, obesity, and obesity-
4
related disorders.
Structurally, the target molecule 1 consists of a spirolac-
tone moiety 2 and a phenylpyrazine subunit 3 connected via
a urea linkage (Scheme 1). The most efficient approach to
the spirolactone fragment 2 appeared to be via aryllithium
addition to protected piperidone 5 followed by in situ
lactonization. Potential complications include quenching of
the aryllithium by the carboxylic acid and enolization of the
piperidone. The phenylpyrazine 3 could be synthesized by
bromination of aminopyrazine 6 followed by a Pd-catalyzed
Suzuki-Miyaura coupling with phenylboronic acid.
Introduction
Results and Discussion
Obesity is a growing epidemic and is currently affecting
approximately 30% of the adult population in the Western
world. It is caused by a chronic imbalance between energy
intake and energy expenditure. A body-mass-index (BMI)
Selective monobromination of aminopyrazine is a long
standing problem. In the literature, it has been reported to
5
be a low yielding reaction. In our hands, adding solid NBS
to a solution of aminopyrazine 6 afforded the bromopyrazine
7 in 50-65% yields on a small scale. However, on a larger
scale, whereas NBS was added slowly or in portions to
control the exothermic reaction, the yield dropped precipi-
tously (30%). Large amounts of a dark gummy solid
appeared in the reaction mixture which further complicated
the workup and isolation of the product.
2
above 30 kg/m significantly increases the risk of numerous
diseases particularly diabetes, hypertension, and dyslipidemia
and reduces life expectancy. Obesity is imposing an increas-
ingly severe burden to the healthcare systems. Neuropeptide
Y (NPY) is a highly conserved peptide that is involved in
1
regulating feeding behavior. Thus far six types of NPY
receptors have been identified with the Y5 type dominant.
It has been shown that chronic administration of NPY leads
to obesity in rodents, while injection of NPY antisense
With N-bromoacetamide, the reaction initially appeared
to be much slower based on the rate of increase of the
2
(
3) Schaffhauser, A. O.; Stricker-Krongrad, A.; Brunner, L.; Cumin, F.; Gerald,
C.; Whitebread, S.; Criscione, L.; Hofbauer, K. G. Diabetes 1997, 46, 1792-
1798.
*
To whom correspondence should be addressed. E-mail: takahiro_itoh
merck.com; matthew_zhao@merck.com.
1) (a) Parker, E.; Van Heek, M.; Stamford, A. Eur. J. Pharmacol. 2002, 440,
@
(
(4) (a) Erondu, N. E.; Fong, T. M.; MacNeil, D. J.; Van Der Ploeg, L. H. T.
PCT Int. Appl. WO 2005000217 A2 20050106, 2005. (b) Erondu, N. E.;
Fong, T. M.; MacNeil, D. J.; Van Der Ploeg, L. H. T.; Kanatani, A. PCT
Int. Appl. WO 2004110375 A2 20041223, 2004. (c) Fong, T. M.; Erondu,
N. E.; MacNeil, D. J.; Mcintype, J. H.; Van Der Ploeg, L. H. T. PCT Int.
Appl. WO 2004110368 A2 20041223, 2004. (d) MacNeil, D. J.; Mcintyre,
J. H.; Van Der Ploeg, L. H. T.; Ishihara, A. PCT Int. Appl. WO 2004009015
A2 20040129, 2004. (e) Fukami, T.; Kanitani, A.; Ishihara, A.; Ishii, Y.;
Takahashi, T.; Haga, Y.; Sakamoto, T.; Itoh, T. U.S. Patent US 2002188124
A1 20021212, 2002. (f) Fukami, T.; Kanatani, A.; Ishihara, A.; Ishii, Y.;
Takahashi, T.; Haga, Y.; Sakamoto, T.; Itoh, T. PCT Int. Appl. WO
2001014376 A1 20010301, 2001.
1
73-187. (b) Gehlert, D. R.; Heiman, M. Annu. Rep. Med. Chem. 1997,
3
2, 21-30. (c) Gerald, C.; Walker, M. W.; Criscione, L.; Gustafson, E. L.;
Batzl-Hartmann, C.; Smith, K. E.; Vaysse, P.; Durkin, M. M.; Laz, T. M.;
Linemeyer, D. L.; Schaffhauser, A. O.; Whitebread, S.; Hofbauer, K. G.;
Taber, R. I.; Branchek, T. A.; Weinshank, R. L. Nature 1996, 382, 168-
1
71. (d) Marsh, D. J.; Hollopeter, G.; Kafer, K. E.; Palmiter, R. D. Nat.
Med. 1998, 4, 718-721. (e) Gehlert, D. R. Neuropeptides 1999, 33, 329-
38.
3
(
2) (a) Paposinho, P. D.; Pierroz, D. D.; Broqua, P.; White, R. B.; Pedrazzini,
T.; Aubert, M. L. Mol. Cell. Endocrinol. 2001, 185, 195-204. (b) Stanley,
B. G.; Kyrkouli, S. E.; Lampert, S.; Leibowitz, S. F. Peptides 1986, 7, 1189-
(5) (a) Sato, N.; Takeuchi, R. Synthesis 1990, 659-660. (b) Sato, N. J.
Heterocycl. Chem. 1982, 19, 673-674. (c) De Bie, D. A.; Ostrowicz, A.;
Geurtsen, G.; Van der Plas, H. C. Tetrahedron 1988, 44, 2977-2983.
1
192. (c) Beck, B.; Stricker-Krongrad, A.; Nicolas, J. P.; Burlet, C. Int. J.
Obes. 1992, 16, 295-302.
8
22
•
Vol. 10, No. 4, 2006 / Organic Process Research & Development
10.1021/op0600963 CCC: $33.50 © 2006 American Chemical Society
Published on Web 06/14/2006

Scheme 1. Retrosynthetic analysis
Scheme 2. Bromination side reactions
agent. We found that both NBS and DBH were unstable in
THF, DMSO, DME, dioxane, NMP, and various aqueous/
organic systems. The most promising solvents were DMF
and MeCN. DBH was selected due to its higher solubility
and more robust solution stability. It should be pointed out
that both bromine atoms were utilized. N-Bromoacetamide
was much more expensive and not readily available on scale.
Thus, by simultaneously adding solutions of DBH and
aminopyrazine 7 in MeCN/DMF to a reaction flask at 0-5
reaction temperature and was thus added in one portion.
Surprisingly, the temperature of the reaction mixture started
to increase rapidly by 15 °C after a few minutes of induction
period. To our surprise, the assay yield was 78%, the highest
observed thus far! However, on scale-up, whereas the
N-bromoacetamide was added slowly, the yield dropped back
to 35%. Other brominating agents such as 1,3-dibromo-5,5-
dimethylhydantoin (DBH) behaved similarly. Unfortunately,
rapid addition would create a serious safety hazard on scale-
up for the highly exothermic reaction. The other alternative
for controlling the exothermic reaction was to slowly add
the substrate to the brominating agent. However, this mode
of addition produced large amounts of dibromopyrazine 8.
The barrier for the second bromination was surprisingly low.
In fact, dibromopyrazine 8 could be obtained cleanly by
simply adding the substrate to 2 equiv of brominating agent.
°
C, the bromination became reproducible and the yield was
significantly improved to 70%. While this procedure was
convenient and scalable, it was still not optimal. As the
reaction progressed, the accumulated product led to its
increased participation in the above-mentioned side reactions.
A flow-cell reactor, where the substrate and DBH solutions
are rapidly mixed and then pushed out of the reactor, should
eliminate this problem. Thus, the reaction was conducted in
a simple flow-cell reactor prepared from a syringe and a coil
of polypropylene tube. Solutions of the substrate and the
brominating agent were added simultaneously to the syringe
with stirring and then pushed through the coil. With this
setup, the yield was further improved to 80%. This protocol
was successfully demonstrated on a kilogram scale using a
small round-bottom flask as the flow-cell reactor. Careful
calibration of the pumps to maintain a proper reagent/
substrate ratio was critical for optimal results.
Alternatively, selective reduction of dibromopyrazine 8,
obtained by simply adding aminopyrazine to 2.0 equiv of
brominating agents (Vide supra), would provide the desired
bromopyrazine 7. Literature precedence on palladium-
catalyzed regioselective Stille and Sonogashira couplings of
We hypothesized that the low yield on slow addition was
due to competitive N-bromination followed by dimerization
with the starting material to give 10 (Scheme 2). The
presence of excess starting material favored this side reaction.
Further N-bromination of the dimer followed by eliminating
8
at the 3-position provided further impetus for investigating
7
this approach.
Screenings of various hydride donors, palladium sources,
ligands, bases, and solvents for the regioselective debromi-
nation were carried out (Table 1). Initial results indicated
HBr gave an azo-dye type compound, 2,2′-azobispyrazine
6
1
2, observed by LC-MS analysis. Compound 6 could also
participate in analogous side reactions, further aggravating
the problem. This hypothesis was supported by the fact that
decreased yields were accompanied by increased acidity of
the reaction mixture as well as the appearance of large
amounts of dark precipitate.
that formic acid was the best hydride donor and Pd(OAc)
was a good palladium source for the reaction. Other
palladium catalysts such as Pd(PPh and PdCl (PPh also
gave acceptable results. Polar aprotic solvents such DMAc
and DMSO were the preferred solvents, and the best ligand
2
)
3 4
2
3 2
)
We reasoned that simultaneous addition of the substrate
and the brominating agent should substantially reduce this
N-N bond forming side reactions by minimizing the
presence of excess starting material. This mode of addition
required a reasonable stable solution of the brominating
appeared to be P(2-furyl)
expensive (entry 3). The much more economical Ph
performed almost as well (91% yield). Triethylamine and
i-Pr NEt afforded essentially the same results (entries 6, 7).
3
, which was unfortunately rather
3
P
2
(
7) (a) Nakamura, H.; Takeuchi, D.; Murai, A. Synlett 1995, 1227-1228. (b)
Nakamura, H.; Aizawa, M.; Takeuchi, D.; Murai, A.; Shimoura, O.
Tetrahedron Lett. 2000, 41, 2185-2188. (c) Wu, C.; Nakamura, H.; Murai,
A.; Shimomura, O. Tetrahedron Lett. 2001, 42, 2997-3000.
(
6) (a) 2,2′-Azobispyrazine: Carlucci, L.; Ciani, G.; Proserpio, D. M.; Rizzato,
S. New J. Chem. 2003, 27, 483-489. (b) Brown, E. V.; Granneman, G. R.
J. Am. Chem. Soc. 1975, 97, 621-627.
Vol. 10, No. 4, 2006 / Organic Process Research & Development
•
823

Table 1. Effect of ligand, base, and solvent for regioselective
reduction of 8
failed. Therefore, a search for reaction conditions compatible
with both steps was carried out in DME/EtOH/H O (5/2/1),
2
entry
palladium/ligand
base
solvent
7
6
the preferred solvent system for the Suzuki-Miyaura
coupling (Table 2).
1
2
3
4
5
6
7
8
9
Pd(OAc)
Pd(OAc)
Pd(OAc)
Pd(OAc)
Pd(OAc)
Pd(OAc)
Pd(OAc)
2
/P(o-tol)
/P(t-Bu)
/P(2-furyl)
/PPh
/PPh
/PPh
/PPh
3
Et
Et
3
3
Et
N
N
DMAc 33
DMAc 39 10
DMAc 97
DMAc 88
DMSO 90
MeCN 61
5
It was ound that formic acid, ammonium formate, sodium
2
3
formate, or (EtO)
base (entries 1-3). A reasonable yield of the desired product
6 was obtained using NaBH as the hydride source (entry
). Triethylsilane appeared to be the best hydride donor for
the debromination affording 95% yield of bromopyrazine 7
entry 5). Expensive commercial Pd(PPh or a more
economical in situ prepared catalyst from Pd(OAc) and PPh
were very effective for the reactions. On the other hand,
PdCl /PPh was much less so (entry 7). Simple Pd/C was
almost completely ineffective (entry 8). We also observed
that 2.0 equiv of Et SiH were required for the reaction in
the absence of base. By adding 0.5 equiv of Na CO to
SiH charge could be
3
SiH gave poor results in the absence of a
2
3
3
N
2
3
1
0.3
2
2
3
Et
3
Et
N
2
3
3
N
4
2
3
Et
3
N
4
2
3
2
i-Pr NEt DMSO 91
Pd(OAc)
Pd(OAc)
2
/PPh
/PPh
3
DBU
n-Bu
DMSO 87 10
DMSO 68 1
(
3 4
)
2
3
3
N
2
3
Scheme 3. Mechanism of the regioselective debromination
2
3
3
2
3
neutralize the generated HBr, the Et
reduced to 1.2 equiv.
3
Thus, dibromopyrazine 8 was treated with 1.2 equiv of
SiH in the presence of 4 mol % of Pd(OAc) , 8 mol %
PPh , and 0.5 equiv of Na CO in DME/EtOH/H O (5/2/1)
at 80 °C for 2 h to give bromopyrazine 7. The reaction
mixture was then treated with 1.1 equiv of PhB(OH) and
.5 of equiv Na CO at 80 °C for 2 h affording 2-amino-5-
Et
3
2
3
2
3
2
2
1
2
3
phenylpyrazine (3). The yield for the reduction step was as
high as 96%, and the overall isolated yield of 3 was 87%
(eq 3). Thus we have developed an efficient and high yielding
DBU gave a somewhat lower yield and more overreduction
product 6 (entry 8). Longer chain tertiary amine n-Bu N gave
3
a substantially lower yield (entry 9).
Interestingly, bidentate ligands such as DPPF and BINAP
were much less effective than monodentate ligands. This
observation supports our hypothesis that participation of the
neighboring amino groups as a ligand was the key for the
observed high regioselectivity (Scheme 3).
8
With the bromopyrazine 7 in hand, its coupling with
phenylboronic acid was investigated. It was found that
excellent results were obtained with 0.5 mol % PdCl ‚dppf
2
as the catalyst. The reaction went to completion in 2-8 h to
give the coupling product 3 in 93% assay yield (eq 2). The
one-pot procedure for the synthesis of 2-amino-5-phe-
nylpyrazine 3 from dibromopyrazine employing sequential
palladium-catalyzed regioselective reduction and a Suzuki-
Miyaura coupling reaction.
Subsequently, we also developed two alternative synthetic
routes to the phenylpyrazine 3 via condensation of isoni-
trosoacetophenone with aminomalononitrile and aminoac-
etonitriles, respectively. The results have been published
presence of water appeared to promote the reaction as the
reaction was very sluggish in the absence of water. The
product was crystallized from toluene/heptane in 88% yield.
Since palladium was used for the regioselective debro-
mination step, it would be very advantageous to perform the
Suzuki-Miyaura coupling in a one-pot fashion using the
residual palladium catalyst. However, directly carrying
forward the debromination reaction mixture to the next step
9
elsewhere.
For the synthesis of spirolatone fragment 2, aryllithium
addition to N-protected piperidone appeared to be the most
efficient approach (eq 4). Treatment of 2-bromobenzoic acid
with 2.2 equiv of n-BuLi at -78 °C followed by N-benzyl
piperidone and acidification afforded the desired spirolactone
7
albeit in rather low yield. The best yield (58% by HPLC
(
8) (a) Suzuki, A. J. Organomet. Chem. 1999, 576, 147-168. (b) Thompson,
(9) (a) Albaneze-Walker, J.; Zhao, M.; Baker, M.; Dormer, P.; McNamara, J.
Tetrahedron Lett. 2002, 43, 6747-6750. (b) Itoh, T.; Maeda, K.; Wada,
T.; Tomimoto, K.; Mase, T. Tetrahedron Lett. 2002, 43, 9287-9290.
W. J.; Jones, J. H.; Lyle, P. A.; Thies, J. E. J. Org. Chem. 1988, 53, 2052-
2
055. (c) Jones, K.; Keenan, M.; Hibbert, F. Synlett. 1996, 509-510.
824
•
Vol. 10, No. 4, 2006 / Organic Process Research & Development

Table 2. Screening of reducing agents and palladium catalysts for reduction
entry
reductant (equiv)
palladium/ligand
base (equiv)
7
6
1
2
3
4
5
6
7
8
9
HCO
HCO
2
H (2.0)
NH (2.0)
SiH (2.0)
(2.0)
SiH (2.0)
Pd(PPh
Pd(PPh
Pd(PPh
Pd(PPh
Pd(PPh
Pd(OAc)
3
3
3
3
)
)
)
)
4
4
4
4
none
19%
16%
8%
trace
trace
ND
1%
2%
7%
7%
1%
2%
28%
3%
2
4
none
none
none
none
none
none
none
(EtO)
NaBH
Et
3
4
70%
95%
86%
59%
13%
96%
66%
38%
3
3
)
4
Et
Et
Et
Et
Et
Et
3
SiH (2.0)
SiH (2.0)
SiH (2.0)
2
/PPh
3
3
3
PdCl
Pd/C
2
/PPh
3
3
SiH (1.2)
SiH (1.2)
SiH (1.2)
Pd(OAc)
Pd(OAc)
PdCl /PPh
2
/PPh
3
Na
Et
Na
2
CO
N (1.0)
CO aq. (0.5)
3
aq. (0.5)
10
1
3
2
/PPh
3
3
1
3
2
3
2
3
assay) was obtained by slow addition of n-BuLi.10 As
expected, quench of the aryllithium by the carboxylic acid
was a major side reaction. Another side reaction was the
enolization of piperidone. Various amounts of compound 8
via direct n-BuLi addition to the piperidone were also
observed. Furthermore, the reaction required cryogenic
conditions which would drive up the cost significantly for
large scale production.
Table 3. Effect of alkylmagnesium reagents
yield (%) by HPLC
entry reagent (equiv) n-BuLi (equiv) 13 PhCO
2
H
14
1a
none
n-Bu Mg (0.5)
MeMgBr (1.0)
i-PrMgCl (1.1)
n-BuMgCl (0.8)
n-BuMgCl (0.9)
n-BuMgCl (1.0)
2.2
1.05
1.5
1.5
1.3
1.3
1.3
58
67
63
71
73
76
76
19
28
20
17
10
13
11
2
3
4
5
6
7
2
16
7
13
a
Reaction was carried out at -78 °C.
Scheme 4. Final product
To avoid the premature halogen-metal exchange and the
resulting internal quench, the feasibility of performing the
deprotonation step with a less active alkylmagnesium reagent
followed by subsequent metal-halogen exchange with
n-BuLi was investigated. We also envisioned that, in the
presence of Mg , the aryllithium should be rapidly converted
to the thermally more stable arylmagnesium and thus could
potentially obviate the costly cryogenic conditions. This
approach proved to be quite successful. Sequential treatment
in excellent yield (91%). Thus we have developed a very
concise and economical synthesis of the spirolactone 2. All
raw materials were very inexpensive, and the requirement
for cryogenic conditions was eliminated.
2+
With both the spirolactone 2 and the phenylpyrazine 3 in
hand, their coupling to the target molecule 1 was investigated.
A number of methods for the synthesis of unsymmetrical
of 2-bromobenzoic acid with 0.52 equiv of n-Bu
2
Mg and
1.1 equiv of n-BuLi followed by N-benzyl piperidone
1
2
afforded the spirolactone 13 in 67% yield after in situ
lactonization. The reaction could be successfully carried out
at -20 to 0 °C without any instability problems. This
procedure proved to be of general utility, and the detailed
results were disclosed recently.11
Subsequently, various other alkylmagnesium reagents
were examined, and the results are summarized in Table 3.
We found that MeMgBr gave a somewhat lower yield than
ureas have been reported. We decided to use the phenyl
12g
chloroformate protocol for ease of operation, low cost, and
mildness of reaction conditions. The aminopyrazine 3 was
chosen as the first coupling partner instead of the spirolactone
amine 2, as the latter should be more active and thus should
proceed under milder conditions for the formation of a urea
linkage. This consideration was important as the urea moiety
in the final product was somewhat labile to hydrolysis under
strong basic conditions. For the preparation of the activated
carbamate 15, aqueous workup and isolation should be
avoided if possible as it is prone to hydrolysis. Ideally, the
n-Bu
2
Mg (entry 3). i-PrMgCl afforded quite good results
(71% yield, entry 4). The best yield (76%) was obtained with
0.9-1.0 equiv of n-BuMgCl (entries 6 and 7). The reaction
has been successfully scaled up to a multikilogram scale
affording the spirolactone in 74% isolated yield.
Removal of the benzyl protective group was accomplished
by a simple hydrogenolysis of 13 in MeOH at 20-30 °C
with Pd/C as the catalyst. The desired product 2 was isolated
(
12) (a) Kitazaki, T.; Ichikawa, T.; Tasaka, A.; Hosono, H.; Matsushita, Y.;
Hayashi, R.; Okonogi, K.; Itoh, K. Chem. Pharm. Bull. 2000, 48, 1935-
1946. (b) Ichikawa, T.; Yamada, M.; Yamaguchi, M.; Kitazaki, T.;
Matsushita, Y.; Higashikawa, K.; Itoh, K. Chem. Pharm. Bull. 2001, 49,
1110-1119. (c) Katritzky, A. R.; Pleynet, D. P. M.; Yang, B. J. Org. Chem.
1997, 62, 4155-4158. (d) Adamiak, R. W.; Stawi n˜ ski, J. Tetrahedron Lett.
1977, 18, 1935-1936. (e) Matsumura, Y.; Satoh, Y.; Onomura, O.; Maki,
(
10) Beak, P.; Musick, T. J.; Chen, C. W. J. Am. Chem. Soc. 1988, 110, 3538-
542.
11) Kato, S.; Nonoyama, N.; Tomimoto, K.; Mase, T. Tetrahedron Lett. 2002,
3, 7315-7317
T. J. Org. Chem. 2000, 65, 1549-1551. (f) Batey, R. A.; Santhakumar, V.;
Yoshina-Ishii, C.; Taylor, S. D. Tetrahedron Lett. 1998, 39, 6267-6270.
(g) Kitteringham, J.; Shipton, M. R.; Voyle, M. Synth. Commun. 2000, 30,
1937-1943.
3
(
4
Vol. 10, No. 4, 2006 / Organic Process Research & Development
•
825

Figure 1. XRPD chart of each crystal forms of 1.
product would crystallize from the reaction media while the
hydrochloride salt of the organic base remained in the mother
liquor. Thus, proper selection of base and solvent was crucial.
When pyridine was used as the base and THF was used as
the solvent, a thick gel was obtained. After switching to
acetonitrile as the solvent, the product 15 precipitated as a
very slow-filtering fine suspension. Fortunately, excellent
results were obtained when a THF/MeCN mixed solvent
system was employed. The product crystallized nicely and
was directly filtered in 92-95% yield while pyridine
hydrochloride salt remained in the mother liquor.
The final coupling of the carbamate 15 with the spiro-
lactone 2 is quite facile in solvents such as MeCN, THF,
DMF, DMSO, etc. DMF was selected owing to the higher
solubility of the substrate and product in this solvent. By
2
adding i-Pr NEt as the base, the hydrochloride salt of the
spirolactone could be used directly. The order of addition
Figure 2. Correlation of polymorphism for thermodynamic
stability.
form B under anhydrous slurry state conditions. The bio-
availability of the anhydrous form D was nearly equal to
that of form B. However, it was necessary to mill form D
finely to acquire the same bioavailability as form B. As
mentioned above, form D was not a good candidate for
clinical study due to absorption and desorption of water
reversibly depending on the relative humidity and technical
issues of milling. But form B was a metastable form under
aqueous conditions, so that the chemical stability tests of
form B were investigated. As a result, form B was not
converted to form D under high humidity, and the conversion
from B to D was kinetically very slow under aqueous slurry
state conditions. With these data in hand, it was clear that
form B was the more preferable form than form D for clinical
study. After the coupling reaction, 1 was obtained as form
was quite important as the carbamate 15 was unstable to
base (i-Pr
the reaction was carried out by mixing the HCl salt of
spirolactone 2 and i-Pr NEt for 0.5 h followed by addition
of the carbamate 15. The mixture was then heated to 40-
2
NEt) in the absence of the coupling partner. Thus,
2
4
9
5 °C and stirred for 2-4 h affording the final product 1 in
2% isolated yield (Scheme 4).
Compound 1 has four crystal forms; form A, B, C, and
2
D by crystallization from a DMF/H O system. Then, form
D was easily converted to form B by the slurry state
transformation in MeCN.
D (Figure 1). The most stable polymorph is usually preferable
for pharmaceutical purposes because the crystalline form
might affect the mechanical stability of tablets and the
bioavailability of the active compound. It was necessary to
minimize the possibility of subsequent polymorphic changes
for the development of the drug. As deduced from many
polymorph studies, we found that forms A and C were less
thermodynamically stable than forms B and D (Figure 2).
Form D is the most thermodynamically stable form and exists
in hydrous and anhydrous forms. The results of the hygro-
scopic analysis showed that the vacuum-dried form D was
converted to the hydrous form under high humidity, and the
hydrous form D was converted to the hemihydrate form D
under low humidity. Also form D was easily converted to
In summary, we have developed a very concise and
practical synthesis of the highly selective NPY-5 receptor
antagonist 1. A much improved monobromination of ami-
nopyrazine by using a flow-cell reactor was developed.
Alternatively, sequential palladium-catalyzed regioselective
debromination of dibromopyrazine and Suzuki-Miyaura
coupling gave the phenylpyrazine fragment 3. A practical
and economical synthesis of the spirolactone fragment 2 and
a high-yielding coupling with the aminopyrazine 3 complete
the synthesis of the target molecule 1.
Experimental Section
General. All reagents were purchased and used without
any further purification. Organic solvents were dried over 4
826
•
Vol. 10, No. 4, 2006 / Organic Process Research & Development

Å molecular sieves. H and 13C NMR spectra were taken in
1
overnight to give 1.82 kg (88% yield) of 3 as a yellow solid.
(b) One-pot debromination/Suzuki-Miyaura coupling pro-
cedure: A mixture of the dibromopyrazine 8 (100 g, 395
DMSO-d at 400 and 100 MHz, respectively. Hydrogen
multiplicity (C, CH, CH
6
2 3
, CH ) information was obtained
from DEPT spectra.
mmol), Et
mL, 0.5 equiv), and DME/EtOH/H
degassed by vacuum/N -fill cycles. Pd(OAc)
PPh (8 mol %) were added, and the mixture was degassed
3
SiH (75.7 mL, 1.2 equiv), 1.5 M aq. Na
O (5/2/1, 1.5 L) was
(4 mol %) and
2 3
CO (132
2-Amino-5-bromopyrazine (7). (a) Simultaneous addi-
2
tion mode: A solution of 2-aminopyrazine 6 (3.5 kg, 36.8
mol) in DMF/MeCN (7 L/22.8 L) and a solution of DBH
2
2
3
(
5.5 kg, 19.3 mol) in DMF/MeCN (3.5 L/33.9 L) were added
twice more. The mixture was heated to 80 °C and aged for
2 h. After confirming the completion of the reaction by
HPLC, the reaction mixture was cooled to 30-40 °C. PhB-
simultaneously to a flask containing 3.5 L of MeCN at 0 °C
over 1 h. The reaction mixture was stirred for 0.5 h, quenched
with aq. Na
vacuo. Solka-Floc (filter aid, 1.1 kg) and aq. Na
wt %, 33.0 L) were added, and the mixture was stirred for
h and then filtered. The filtrate was saturated with solid
2
S
2
O
3
‚5H
2
O (10%, 7.0 L), and concentrated in
(OH)
mmol) were added, and the mixture was degassed by
vacuum/N -fill cycles. It was heated to refluxing temperature
2 2 3
(53.0 g, 435 mmol) and solid Na CO (62.8 g, 593
2
CO (2.0
3
2
1
and aged for 2 h. After confirming the completion of the
reaction by HPLC, it was cooled to ambient temperature and
THF (500 mL) was added to dissolve the product. The
organic layer was separated, washed with brine (500 mL ×
NaCl (∼10.0 kg) and then extracted with 3/2 EtOAc/n-
heptane (35.0 L, 2 × 16.0 L). The combined organic extract
was washed with brine and then treated with Darco-KB
overnight. The product was crystallized by a solvent switch
to heptane furnishing 4.48 kg of 7 as a yellow solid (65%
yield corrected for 93% purity). (b) Flow-cell reaction: A
small flow-cell reactor was made by connecting a 20 mL
syringe with polypropylene tubing coiled to a ∼70 mL
internal volume. A magnetic stir bar is placed in the syringe
reactor, and the latter was capped with a rubber septum. Both
the syringe and the coil were immersed in a cooling bath
2 4
2), treated with Darco-KB (5.0 g) and Na SO (10.0 g)
overnight, and then filtered and washed with toluene/THF
(1/1, 300 mL). The filtrate was concentrated under reduced
pressure to 400 mL, and then heptane (350 mL) was added
over 1 h. After aging for 2 h, the product was filtered, washed
with toluene/n-heptane (1/1, 200 mL), and dried to give 58.9
g of 3 as a yellow solid in 87% yield corrected for purity.
1
Mp 147.5-148.0 °C; H NMR (DMSO-d
6
) δ 8.51 (d, 1H, J
(
-20 °C). The outlet of the coil led to a 2 L flask maintained
) 1.5 Hz), 8.00 (d, 1H, J ) 1.5 Hz), 7.91 (d, 2H, J ) 7.5
Hz), 7.40 (d, 2H, J ) 7.7, 0.5 Hz), 7.29 (d, 1H, J ) 7.7
at 0 °C. A solution of 2-aminopyrazine 6 (35.0 g, 368 mmol)
in DMF/MeCN (70/228 mL) and a solution of DBH (55.2
g, 193 mmol) in DMF/MeCN (35/339 mL) were added
simultaneously via a syringe pump to the flow-cell reactor
at a rate of 3.0 and 3.7 mL/min, respectively, with vigorous
stirring. After completing the addition, the system was purged
by pumping through acetonitrile. Following the workup and
isolation, the procedure above gave 51.0 g of 7 as a yellow
1
3
Hz), 6.56 (brd s, 2H); C NMR (DMSO-d
6
) δ 155.4 (C),
139.6 (C), 139.3 (CH), 137.6 (C), 131.9 (CH), 129.1 (CH),
-
1
127.8 (CH), 125.2 (CH); IR (KBr, cm ) 3343, 3177, 1651,
1588, 1538, 1479, 1447, 1389, 1216, 1025, 1010, 883, 783,
9 3
751, 694. Anal. Calcd for C10H N : C, 70.16; H, 5.30; N,
24.54. Found: C, 69.90; H, 4.97; N, 24.53.
1′-Benzylspiro[isobenzofuran-1(3H),4′-piperidin]-3-
one Hydrochloride (13). n-BuMgCl (2.0 M in THF, 4.5 L,
9.0 mol) was added slowly to a solution of 2-bromobenzoic
acid 4 (2.0 kg, 9.95 mol) in THF (20 L) followed by n-BuLi
(1.56 M in n-hexane, 8.3 L, 12.9 mol) at -15 to -5 °C.
The mixture was stirred for 0.5 h, and then a solution of
1-benzyl-4-piperidone (2.02 L, 10.9 mol) in n-heptane (6.0
L) was added over 1 h. After aging for 1 h, methyl tert-
butyl ether (MTBE, 10.0 L) and AcOH (3.14 L, 54.8 mol)
in water (20 L) were added sequentially. The mixture was
heated to 35 to 40 °C and stirred for 3 h to affect the
lactonization. It was then cooled to 25 °C, and the layers
separated. The aqueous layer was extracted with MTBE (10.0
L) and then combined with the organic layer. The combined
organic layer was then washed sequentially with NaOH (2.0
solid (75% yield corrected for 94% purity). Mp 114.5-115.5
1
°
C; H NMR (DMSO-d
6
) δ 8.00 (d, J ) 1.5 Hz, 1H), 7.68
(
d, J ) 1.5 Hz, 1H), 6.60 (brd, 2H); 13C NMR (DMSO-d
δ 155.7 (C), 144.0 (CH), 132.6 (CH), 124.1 (C); IR (KBr,
6
)
-
1
cm ) 3309, 3167, 1631, 1567, 1531, 1460, 1380, 1206,
1
2
4
4 4 3
105, 1009, 863. Anal. Calcd for C H N Br: C, 27.61; H,
.32; Br, 45.92; N, 24.15. Found: C, 27.52; H, 2.19; Br,
5.58; N, 24.03.
2
-Amino-5-phenylpyrazine (3). (a) Synthesis from 6: A
mixture of toluene (13.4 L), water (9.5 L), DMF (2.38 L),
CO (2.47 kg, 17.9 mol), 2-amino-5-bromopyrazine 6 (2.2
kg, 94 wt %, 11.9 mol), PhB(OH) (1.6 kg, 13.1 mol), and
PdCl ‚dppf‚CH Cl (87.1 g, 0.12 mol) was degassed by
vacuum/nitrogen-fill cycles and then heated to reflux (∼87
C) until the reaction was complete (5-8 h). It was cooled
K
2
3
2
2
2
2
°
N, 10.0 L), K
2 3 2
CO (1.0 kg, 7.24 mol in 5.0 L of H O), and
to <40 °C, and then THF (11.8 L) was added to dissolve
the product. The organic layer was washed with brine (11.3
L, twice). It was then treated with Darco-KB (414 g) and
Na SO (2.07 kg) for overnight. The mixture was filtered,
2 4
and the filter cake was washed with 1/1 toluene/THF (6.0
L). The filtrate was concentrated under reduced pressure to
H O (10.0 L) and then concentrated and flushed with MeOH
2
(24.0 L) until residual water was <0.2 wt %. A solution of
HCl in EtOAc (4.0 N, 2.32 L, 9.28 mol) was slowly added
followed by methyl tert-butyl ether (24.4 L) to crystallize
the spirolactone 13 as its HCl salt. The solid was collected
by filtration, washed with MeOH/MTBE (1/5), and then dried
1
8.5 L, and then n-heptane (7.5 L) was added over 1 h. After
to give 2.4 kg of 13 HCl salt (74% yield). Mp 275 °C; H
aging for 2 h, the product was collected by filtration and
NMR (DMSO-d
6
) δ 11.5-11.8 (br s, 1H), 7.75-7.9 (2H,
washed with 1/1 toluene/n-heptane (3.8 L) and then dried
m), 7.6-7.75 (m, 3H), 7.4-7.55 (m, 4H), 4.76 (d, J ) 5
Vol. 10, No. 4, 2006 / Organic Process Research & Development
•
827

-
1
Hz, 0.2H), 4.45 (d, J ) 5.0 Hz, 1.8H), 3.1-3.5 (m, 4H),
.83 (t, J ) 13.0 Hz, 2H), 1.94 (d, J ) 14 Hz, 2H). Free
(CH), 122.4 (CH); IR (cm ) 3196, 2978, 1754, 1560, 1444,
1365, 1298, 1238, 1213, 1164, 1075, 1059, 1012, 928, 910.
2
1
base: mp 106.5-107.5 °C; H NMR (DMSO-d
J ) 7.6 Hz, 1H), 7.78-7.73 (m, 2H), 7.61-7.55 (m, 1H),
.36-7.31 (m, 4H), 7.29-7.22 (m, 1H), 3.58 (s, 2H), 2.84
brd d, J ) 11 Hz, 2H), 2.35 (td, J ) 12.5, 2.0 Hz, 2H),
6
) δ 7.80 (d,
13 3 2
Anal. Calcd for C17H N O : C, 70.09; H, 4.50; N, 14.42.
Found: C, 69.92; H, 4.43; N, 14.34.
1′-{[(5-Phenyl-2-pyrazinyl)amino]carbonyl}-spiro-
[isobenzofuran-1(3H),4′-piperidin]-3-one (1). A mixture of
7
(
2
.23 (td, J ) 13.3 Hz, 4.3 Hz, 2H), 1.60 (d, J ) 11.8 Hz,
2
2‚HCl monohydrate (1.68 kg, 7.5 mol), i-Pr NEt (990 mL,
13
2H); C NMR (DMSO-d
6
) δ 168.7 (CdO), 153.6 (C), 138.2
5.67 mol), and DMF (13.9 L) was stirred for 0.5 h, and then
carbamate 15 (1.98 kg, 6.8 mol) was added. (Carbamate 15
should be added last. Otherwise, it would decompose upon
(
(
(
C), 134.4 (CH), 129.4 (CH), 128.9 (CH), 128.1 (CH), 126.9
CH), 125.1 (CH), 124.8 (C), 121.9 (CH), 84.5 (C), 62.1
CH
-
1
2
), 49.2 (CH
32, 731, 694. Anal. Calcd for C19
2
), 35.4 (CH
2
); IR (cm ) 1763 (CdO), 1071,
: C, 77.79; H,
2
mixing with i-Pr NEt.) The mixture was heated to 40-45
9
6
H29NO
2
°C for 2 h, cooled to ambient temperature, and then quenched
with AcOH (119 mL, 2.08 mol). Water (4.2 L) was added,
and the mixture was stirred for 1 h to initiate the crystal-
lization. More water (5.1 L) was added over 2-4 h, and the
mixture was stirred for 2 h. The product was filtered and
washed with DMF/H O (1:1) and water to give the 2.76 kg
2
(97% yield) of 1 (monohydrate, form D). It was converted
into the anhydrous form by suspending in MeCN (7.9 L)
.53; N, 4.77. Found: C, 77.66; H, 6.22; N, 4.70.
Spiro[isobenzofuran-1(3H),4′-piperidin]-3-one hydro-
chloride Monohydrate (2). A solution of 13 HCl salt (2.0
kg, 6.06 mol) in 10/1 methanol/water (35.2 L) was hydro-
genated under 25 psi of hydrogen in the presence of 10%
Pd/C (50 wt % wet, 400 g) at 30 °C for 6 h. The catalyst
was filtered off and rinsed with MeOH (8.0 L). The filtrate
was concentrated and flushed with methanol at 45-50 °C
until a water content of 6-9% and a final volume of 6.4 L
were achieved. The mixture was cooled to 30 °C, and then
MTBE (4.8 L) was added over 2 h to crystallize the product.
The product was filtered, washed with 1/5 MeOH/MTBE
overnight to give 2.5 kg of 1 (92% yield). Mp 207-207.6
1
°C; H NMR (DMSO-d
6
) δ 9.77 (s, 1H), 9.19 (d, J ) 1.5
Hz, 1H), 8.90 (d, J ) 1.5 Hz, 1H), 8.07 (td, J ) 7.2, 1.5 Hz,
1H), 8.07 (d, J ) 7.2 Hz, 1H), 7.84 (d, J ) 7.6 Hz, 1H),
7.78 (dd, J ) 7.4, 0.8 Hz, 1H), 7.76 (dt, J ) 7.2, 0.8 Hz,
1H), 7.59 (dt, J ) 7.3, 0.8 Hz, 1H), 7.48 (tt, J ) 7.4, 1 Hz,
2H), 7.41 (tt, J ) 7.3, 1.5 Hz, 1H), 4.37 (brd d, J ) 13.7
Hz, 2H), 3.23 (t, J ) 12.2 Hz, 2H), 2.27 (dt, J ) 13.7, 4.5
(3.2 L, twice), and dried at 25 °C to give 1.46 kg of 2 (93%
yield corrected for purity, 6.4% water). Mp 268 °C (dec);
1
H NMR (DMSO-d
6
) δ 9.59 (brd s, 2H), 7.85 (d, J ) 7.6
1
3
Hz, 1H), 7.83 (dt, J ) 7.5, 1.0 Hz, 1H), 7.64 (dd, J ) 7.5,
6
Hz, 2H), 1.68 (d, J ) 13.4 Hz, 2H); C NMR (DMSO-d )
1
3
2
1
1
.0 Hz, 1H), 7.59 (d, J ) 7.6 Hz, 1H), 3.44-3.39 (m, 4H),
δ 169.1 (C), 154.4 (C), 153.6 (C), 149.9 (C), 145.1 (C), 139.1
(CH), 136.6 (C), 136.3 (CH), 135.0 (CH), 130.0 (CH), 129.3
(CH), 129.2 (CH), 126.3 (CH), 125.7 (CH), 125.2 (C), 122.5
.12 (dt, J ) 13.0, 2.8 Hz, 2H), 2.60 (dt, J ) 14.2, 4.6 Hz,
H), 1.85 (d, J ) 14.3 Hz, 2H); 13C NMR (DMSO-d
) δ
68.7 (C), 152.6 (C), 135.5 (CH), 130.5 (CH), 126.0 (CH),
25.0 (C), 121.9 (CH), 82.5 (C), 40.6 (CH ), 32.05 (CH );
6
-
1
(CH), 85.1 (C), 41.4 (CH
(brd), 3058, 2951, 2871, 1765, 1668, 1542, 1502, 1446, 1358,
1230, 1064, 930, 790, 694. Anal. Calcd for C23 : C,
2 2
), 35.5 (CH ); IR (cm ) 3291
2
2
-
1
IR (cm ) 3600-2500 (brd), 1769, 1637, 1581, 1465, 1445,
319, 1263, 1079, 1018, 957, 932, 764, 693. Anal. Calcd
: C, 60.13; H, 5.89; Cl, 14.79; N, 5.84.
Found: C, 59.90; H, 5.88; Cl, 15.06; N, 5.78.
5-Phenyl-pyrazin-2-yl)-carbamic Acid Phenyl Ester
15). Phenyl chloroformate (1.36 kg, 1.09 L, 8.69 mol) was
20 4 3
H N O
1
68.99; H, 5.03; N, 13.99; O, 11.99. Found: C, 68.78; H,
4.88; N, 13.93.
for C12H14ClNO
2
(
Acknowledgment
(
We acknowledge Dr. N. Yasuda, Merck & Co., Inc., for
added to a mixture of 3 (1.45 g, 98 wt %, 8.28 mol) and
pyridine (786 g, 804 mL, 1.2 equiv) in THF (6.1 L) and
MeCN (8.2 L) at 20-30 °C over 3 h. The mixture was stirred
for 1 h, filtered, washed with 2/1 MeCN/THF (4.5 L), and
his critical reading of this manuscript.
Supporting Information Available
One-pot debromination/Suzuki-Miyaura coupling pro-
1
13
dried to give 2.22 kg (92% yield) of the carbamate 15 as a
cedure and H and C NMR data for 1, 2, 3, 13, and 15.
This material is available free of charge via the Internet at
http://pubs.acs.org.
1
white solid. Mp 199.5-200.5 °C; H NMR (DMSO-d
6
) δ
11.17 (s, 1H), 9.15 (d, J ) 1.5 Hz, 1 H), 9.00 (d, J ) 1.5
Hz, 1H), 8.09 (dd, J ) 7.1, 1.5 Hz, 2H), 7.53-7.43 (m, 5H),
13
7
(
1
.31-7.25 (m, 3H); C NMR (DMSO-d
C), 147.8 (C), 146.6 (C), 139.8 (CH), 136.2 (C), 135.0 (CH),
30.0 (CH), 129.6 (CH), 129.4 (CH), 126.5 (CH), 126.3
6
) δ 152.3 (C), 150.7
Received for review May 8, 2006.
OP0600963
828
•
Vol. 10, No. 4, 2006 / Organic Process Research & Development