Synthesis of a naphthyridone p38 MAP kinase inhibitor

Article abstract of DOI:10.1021/jo061618f

Compound 1 is a p38 MAP kinase inhibitor potentially useful for the treatment of rheumatoid arthritis and psoriasis. A novel six-step synthesis suitable for large-scale preparation was developed in support of a drug development program at Merck Research Laboratories. The key steps include a tandem Heck-lactamization, N-oxidation, and a highly chemoselective Grignard addition of 4-(N-tert-butylpiperidinyl)-magnesium chloride to a naphthyridone N-oxide. The N-oxide exerted complete chemoselectivity via chelation in directing the Grignard addition to the α position as opposed to 1,4-addition on the enelactam. The dihydropyridyl adduct was in situ aromatized with isobutylchloroformate followed by heating in pyridine. Syntheses of Grignard precursor, N-tert-butyl-4-chloro-piperidine, were accomplished via transamination with a quaternary ammonium piperidone or via addition of methylmagnesium chloride to an iminium ion. Utilizing this chemistry, multi-kilogram preparation of compound 1 was successfully demonstrated. American Chemical Society.

Full text of DOI:10.1021/jo061618f

Synthesis of a Naphthyridone p38 MAP Kinase Inhibitor
John Y. L. Chung,* Raymond J. Cvetovich,* Mark McLaughlin, Joseph Amato,
Fuh-Rong Tsay, Mark Jensen, Steve Weissman, and Daniel Zewge
Process Research, Merck Research Laboratories, P.O. Box 2000, Rahway, New Jersey 07965
john_chung@merck.com
ReceiVed August 3, 2006
Compound 1 is a p38 MAP kinase inhibitor potentially useful for the treatment of rheumatoid arthritis
and psoriasis. A novel six-step synthesis suitable for large-scale preparation was developed in support of
a drug development program at Merck Research Laboratories. The key steps include a tandem Heck-
lactamization, N-oxidation, and a highly chemoselective Grignard addition of 4-(N-tert-butylpiperidinyl)-
magnesium chloride to a naphthyridone N-oxide. The N-oxide exerted complete chemoselectivity via
chelation in directing the Grignard addition to the R position as opposed to 1,4-addition on the ene-
lactam. The dihydropyridyl adduct was in situ aromatized with isobutylchloroformate followed by heating
in pyridine. Syntheses of Grignard precursor, N-tert-butyl-4-chloro-piperidine, were accomplished via
transamination with a quaternary ammonium piperidone or via addition of methylmagnesium chloride to
an iminium ion. Utilizing this chemistry, multi-kilogram preparation of compound 1 was successfully
demonstrated.
Introduction
elimination half-lives, so their effects cannot be rapidly with-
drawn in the event of an infection. A low molecular weight,
orally active cytokine inhibitor would not have these liabilities.
In addition, p38R kinase activity is necessary for the upregu-
lation of COX-2 in response to stress.3 Thus, a p38R inhibitor
would be expected to have analgesic effects similar to those of
the Coxibs. Several p38 inhibitors are under clinical develop-
ment at various companies, and the ability of these compounds
to inhibit TNFR release in man has been demonstrated.4
Research at Merck has identified compound 1 as a potent
p38 MAP kinase inhibitor (Figure 1), possessing novel entities
outside the scope of other currently known inhibitors with
p38 mitogen-activated protein (MAP) kinases are intracellular
soluble serine/threonine kinases that positively regulate the
production and action of several pro-inflammatory mediators.1
In particular, the release of tumor necrosis factor-R (TNF-R)
and interleukin-1 (IL-1) in response to stress is dependent on
p38 kinase activity, and therapeutic strategies to target these
cytokines have been clinically validated. Biological agents that
sequester TNF-R have shown impressive clinical efficacy in the
treatment of rheumatoid arthritis (RA), Crohn’s Disease, and
psoriasis.2 Despite the success of these biological agents, several
important unmet medical needs remain. All of these biological
agents must be administered parenterally. They have very long
(3) Lee, J. C.; Badger, A. M.; Griswold, D. E.; Dunnington, D.; Truneh,
A.; Votta, B.; White, J. R.; Young, P. R.; Bender, P. E. Ann. N.Y. Acad.
Sci. 1993, 696, 149-170.
(4) (a) Faas, M. M.; Moes, H.; Fijen, J. W.; Muller Kobold, A. C.;
Tulleken, J. E.; Zijlstra, J. G. Clin. Exp. Immunol. 2002, 127, 337-343.
(b) Branger, J.; van den Blink, B.; Weijer, S.; Madwed, J.; Bos, C. L.;
Gupta, A.; Yong, C. L.; Polmar, S. H.; Olszyna, D. P.; Hack, C. E.; van
Deventer, S. J.; Peppelenbosch, M. P.; van der Poll, T. J. Immunol. 2002,
168, 4070-4077.
(1) Chen, Z.; Gibson, T. B.; Robinson, F.; Silvestro, L.; Pearson, G.;
Xu, B.; Wright, A.; Vanderbilt, C.; Cobb, M. H. Chem. ReV. 2001, 101,
2449-2476.
(2) There are currently three TNFR sequestrants marketed in the U.S.
(Enbrel, Remicade, and Humira) with additional agents under development.
Anakinra, an IL-1 receptor antagonist, has been approved for the treatment
of RA.
10.1021/jo061618f CCC: $33.50 © 2006 American Chemical Society
Published on Web 10/04/2006
8602
J. Org. Chem. 2006, 71, 8602-8609

Synthesis of a Naphthyridone p38 MAP Kinase Inhibitor
SCHEME 2. Early Synthesis of Stannane 6
FIGURE 1. p38 MAP kinase inhibitor 1.
SCHEME 1. Early Synthesis of 1
Specifically, the 2,4-difluorophenyl ring occupies a hydrophobic
pocket near Thr-106 that is much smaller in most other kinases,
and the amide carbonyl forms a unique double hydrogen bond
with the protein backbone. This particular sequence of amino
acids, which is necessary for binding 1 with high affinity, is
unique to p38R and p38â kinases. To further the studies of this
compound, we needed an efficient synthesis that would be
amenable to large-scale synthesis.
An early synthesis5 of 1 involved an 18-step synthetic
procedure in its longest linear sequence with an overall yield
of approximately 2% (Schemes 1 and 2). In this synthesis,
naphthyridone 5 was prepared in 15 steps from 2-amino-3-
methylpyridine via lactamization of pyridine 4 followed by
olefin introduction. The N-tert-butyl piperidine moiety was
installed via a Stille coupling using unsaturated piperidinyl
stannane 6, followed by a double bond hydrogenation. This
hydrogenation suffered low yield due to competing lactam olefin
reduction and dechlorination, which produced a mixture of
products and required chromatographic purification. Stannane
6 was prepared in four steps from N-ethyl-4-piperidone employ-
ing hexamethyldistannane in the final step. The synthesis and
the use of stannane 6 for the installation of N-tert-butyl-4-
piperidine constituted major drawbacks in this synthesis. It is
clear that to have an efficient large-scale synthesis of 1, we
needed a streamlined synthesis of the naphthyridone core and
a “greener” means of preparing and installing the N-tert-butyl-
4-piperidine in such a way that it avoids toxic tin reagents, the
subsequent hydrogenation, and the chromatographic purification.
As previously noted, the sp2-sp2 Stille coupling route
suffered from the necessary reduction steps. Thus, alternate
protocols for introducing the piperidine moiety via a sp3-sp2
coupling were preferred. Because alkyl Grignard reagents have
been known to add to pyridine,8 pyridine N-oxide,9,10 and
2-halopyridine11 with or without activation or transition metal
increased potency in whole blood and much improved physi-
cochemical properties.5 The compound contains a naphthyridone
core, which retained the desired selectivity.6 The lipophilic tert-
butyl piperidine moiety attached to the naphthyridone core
greatly improved whole blood and in vivo activity. The tert-
butyl group proved to be optimal for moderating rates of
metabolism and improving pharmacokinetic properties.7 Com-
pound 1 also has excellent selectivity against other kinases.
(8) (a) Gilman, H.; Eisch, J.; Soddy, T. S. J. Am. Chem. Soc. 1959, 81,
4000-4003. (b) Comins, D. L.; Abdullah, A. H. J. Org. Chem. 1982, 47,
4315-4319. (c) Comins, D. L.; Stroud, E. D.; Herrick, J. J. Heterocycles
1984, 22, 151-157. (d) Comins, D. L.; Myoung, Y. C. J. Org. Chem. 1990,
55, 292-298. (e) Nishiwaki, N.; Minakata, S.; Komatsu, M.; Ohshiro, Y.
Chem. Lett. 1989, 773-776. (f) Webb, T. R. Tetrahedron Lett. 1985, 26,
3191-3194. (g) Al-Arnaout, A.; Courtois, G.; Miginiac, L. J. Organomet.
Chem. 1987, 333, 139-153. (h) Kim, Y.-C.; Jacobson, K. A. Synthesis
2000, 119-122. (i) Fakhfakh, M. A.; Fanck, X.; Fournet, A.; Hocquemiller,
R.; Figade`re, B. Tetrahedron Lett. 2001, 42, 3847-3850.
(9) (a) Kato, T.; Yamanaka, H. J. Org. Chem. 1965, 30, 910-913. (b)
van Bergen, T. J.; Kellogg, R. M. J. Org. Chem. 1971, 36, 1705-1708. (c)
Schiess, P.; Ringele, R. Tetrahedron Lett. 1972, 13, 311-312. (d) Schiess,
P.; Monnier, C.; Ringele, P.; Sendi, E. HelV. Chim. Acta 1974, 57, 1676-
1691. (e) Webb, T. R. Tetrahedron Lett. 1985, 26, 3191-3194.
(5) Doherty, J. B.; Stelmach, J. E.; Chen, M.-H.; Liu, L.; Hunt, J. A.;
Ruzek, R. D.; Goulet, J. L.; Wisnoski, D. D.; Natarajan, S. R.; Rupprecht,
K. M.; Bao, J.; Miao, S.; Hong, X. WO patent 2002058695, 2002.
(6) Hunt, J. A.; Kallashi, F.; Ruzek, R. D.; Sinclair, P. J.; Ita, I.;
McCormick, S. X.; Pivinchny, J. V.; Hop, C. E. C. A.; Kumar, S.; Wang,
Z.; O’Keefe, S. J.; O’Neill, E. A.; Porter, G.; Thompson, J. E.; Woods, A.;
Zaller, D. M.; Doherty, J. B. Bioorg. Med. Chem. Lett. 2003, 13, 467.
(7) Bao, J.; Hunt, J. A.; Miao, S.; Rupprecht, K. M.; Stelmach, J. E.;
Liu, L.; Ruzek, R. D.; Sinclair, P. J.; Pivnichny, J. V.; Hop, C. E. C. A.;
Kumar, S.; Zaller, D. M.; Shoop, W. L.; O’Neill, E. A.; O’Keefe, S. J.;
Thompson, C. M.; Cubbon, R. M.; Wang, R.; Zhang, W. X.; Thompson, J.
E.; Doherty, J. B. Bioorg. Med. Chem. Lett. 2006, 16, 64-68.
J. Org. Chem, Vol. 71, No. 22, 2006 8603

Chung et al.
SCHEME 4. Heck-Lactamization of Acrylanilide 18 with
SCHEME 3. Retrosynthetic Analysis of 1
3-Chloro-4-iodo-pyridine
SCHEME 5. Synthesis of Naphthyridone 17
from efforts to develop this synthetic strategy into a practical
chemical synthesis that could be utilized for multi-kilogram
production.
Results and Discussion
Tandem Heck-Lactamization. In route development ex-
ploratory experiments, the Heck coupling of 2,6-dichlorophen-
ylacrylamide (18)15 with 4-chloro-3-iodopyridine (21) was
examined to discover any difficulties in the Heck coupling with
N-arylacrylamides, because no known Heck coupling reaction
using N-arylacrylamides was known to us (Scheme 4). Reaction
in dimethylacetamide (DMAC) at 130 °C in the presence of
Pd(OAc)2 was slow and produced many products, including one
of very particular interest, naphthyridone 22, which was isolated
in 5% yield. This product was not anticipated because Heck
product 23 was not expected to cyclize due to the trans
orientation of the olefin.
Wishing to examine the potential of this reaction, and desiring
to prepare a synthetically useful intermediate for the synthesis
of 1, the coupling of acrylanilide 18 with tri-halosubstituted
pyridine 19 was explored (Scheme 5). The preparation of
2,6-dichlorophenylacrylamide (18) from acryloyl chloride and
2,6-dichloroaniline, initially performed in methylene chloride
in the presence of TEA, gave only a 40% yield of product. This
catalysis, we envisioned that Grignard reagent 1212 derived from
4-chloro-N-tert-butyl-piperidine 13 would be ideal for such
direct coupling with either naphthyridone 14, N-oxide 15, or
chloronaphthyridone 16, preferably without heavy metal ca-
talysis (Scheme 3). Compounds 14-16 in turn may be accessible
from 17 via a Suzuki-Miyaura arylation.13 Naphthyridone 17
bearing the appropriate substituents could be constructed by our
recently discovered Heck-lactamization14 process from acryla-
mide 18 and trihalopyridine 19. Described herein are the results
(10) (a) Cervinka, O.; Fabryova, A.; Matouchova, L. Collect. Czech.
Chem. Commun. 1963, 28, 535-538. (b) Fakhfakh, M. A.; Franck, X.;
Fournet, A.; Hocquemiller, R.; Figadere, B. Tetrahedron Lett. 2001, 42,
3847-3850. (c) Takeuchi, I.; Ozawa, I.; Shigemura, K.; Hamada, Y.; Ito,
T.; Ohyama, A. Yakugaku Zasshi 1979, 99, 451-457. (d) Eberson, L.;
Cardellini, L.; Greci, L.; Poloni, M. Gazz. Chim. Ital. 1988, 118, 35-39.
(e) Tagawa, Y.; Nomura, M.; Yamashita, H.; Goto, Y.; Hamana, M.
Heterocycles 1999, 51, 2385-2397.
(11) (a) Thorsett, E. D.; Stermitz, F. R. J. Heterocycl. Chem. 1973, 10,
243-244. (b) Fu¨erstner, A.; Leitner, A.; Mendez, M.; Krause, H. J. Am.
Chem. Soc. 2002, 124, 13856-13863. (c) Fu¨rstner, A.; Leitner, A. Angew.
Chem., Int. Ed. 2002, 41, 609-612. (d) Wallace, D. J.; Chen, C. Tetrahedron
Lett. 2002, 43, 6987-6990. (e) Thomson, C. G.; Beer, M. S.; Curtis, N.
R.; Diggle, H. J.; Handford, E.; Kulagowski, J. J. Bioorg. Med. Chem. Lett.
2004, 14, 677-680. (f) Corley, E. G.; Conrad, K.; Murry, J. A.; Savarin,
C.; Holko, J.; Boice, G. J. Org. Chem. 2004, 69, 5120-5123. (g) Ohmiya,
H.; Yorimitsu, H.; Oshima, K. Chem. Lett. 2004, 33, 1240-1241.
(12) (a) Remy, D. C.; Rittle, K. E.; Hunt, C. A.; Anderson, P. S.; Arison,
B. H.; Engelhardt, E. L.; Hirschmann, R.; Clineschmidt, B. V.; Lotti, V. J.;
Bunting, P. R.; Ballentine, R. J.; Papp, N. L.; Flataker, L.; Witoslawski, J.
J.; Stone, C. A. J. Med. Chem. 1977, 20, 1013-1019. (b) Sindelar, K.;
Metysova, J.; Holubek, J.; Svatek, E.; Protiva, J.; Protiva, M. Collect. Czech.
Chem. Commun. 1982, 47, 3077-3093.
(13) (a) Miyaura, N.; Suzuki, A. Chem. ReV. 1995, 95, 2457. (b) Suzuki,
A. J. Organomet. Chem. 1999, 576, 147. (c) Suzuki, A. In Metal-Catalyzed
Cross-Coupling Reactions; Diederich, F., Stang, P. J., Eds.; Wiley-VCH:
Weinheim, 1998. (d) Chemler, S. R.; Trauner, D.; Danishefsky, S. J. Angew.
Chem., Int. Ed. 2001, 40, 4544. (e) Suzuki, A.; Brown, H. C. Organic
Syntheses Via Boranes; Aldrich Chemical Co., Inc.: Milwaukee, WI, 2003;
Vol. 3.
(14) Cvetovich, R. J.; Reamer, R. A.; DiMichele, L.; Chung, J. Y. L.;
Chilenski, J. J. Org. Chem. 2006, 71, 8610-8613.
(15) Cvetovich, R. J.; DiMichele, L. Org. Process Res. DeV. 2006, 10,
944-946.
8604 J. Org. Chem., Vol. 71, No. 22, 2006

Synthesis of a Naphthyridone p38 MAP Kinase Inhibitor
TABLE 1. N-Oxidation Solvent Screen
SCHEME 6. Suzuki-Miyaura Coupling of 17 and 26
concentration
(mL/g)
MCPBA
(equiv)
rxn
time (h)
N-oxide
(mol %)
entry
solvent
was due mainly to the insolubility of the dichloroaniline
hydrochloride salt that formed. So, DMF was used as the
solvent. In the absence of added bases, such as TEA, the reaction
achieved 50% conversion at 20 °C and 100% conversion by
warming to 60 °C. However, acrylanilide 18 was prone to
polymerization at these temperatures. In exploring other sol-
vents, the reaction was performed in dimethylacetamide (DMAC).
In this solvent, the reaction was accomplished rapidly at 15-
20 °C. No polymerization was observed, even when heated to
80 °C. When acryloyl chloride was identified as a disadvanta-
geous cost factor ($300/kg), acrylic acid ($6/kg) was dissolved
in DMAC to which thionyl chloride ($3/kg) and then 2,6-dichlo-
roaniline were added at -5 °C, warmed to 20 °C, and then
product was precipitated by the addition of water to produce
acrylanilide 18 in 92% isolated yield.
1
2
3
4
5
6
CH2Cl2
CH2Cl2
acetone
acetone
toluene
THF
6
6
10
10
6
2
16
16
26
168
24
24
97.6
97.0
85.2
12.8
96.6
92.2
1.5
1.75
1.5a
1.45
1.5
6
a Oxone as oxidant.
SCHEME 7. Optimized N-Oxidation Conditions
Pyridine 19 was prepared from 4-amino-2-chloro-pyridine
(24) in two steps. Specifically, Sandmeyer bromination of 24
followed by lithiation-iodination of dihalopyridine 25 afforded
trihalide 19 in 80% overall yield. Initial coupling reactions of
halopyridine 19 and acrylanilide 18 in DMAC at 130 °C were
slow, but did produce naphthyridone 17 in 10-15% yield. The
major product was bis-Heck coupling at the iodo- and bromo-
positions on the pyridine, and the presence of mono-Heck
product was detected in very small amount. Improvements in
the formation of the desired product resulted with the switch to
ethylene carbonate as solvent along with the addition of
potassium bromide and small amounts of water. Treatment of
19 with 2,6-dichlorophenylacrylamide (18) in ethylene carbonate
at 90-100 °C in the presence of Pd(OAc)2, NaOAc, and KBr
afforded the tandem Heck-lactamization and produced key
naphthyridone 17 in 80% yield on lab scales (1-10 mmol) and
65% on kilogram scales.14 The formation of naphthyridone 17
did not appear to be dependent on the formation of mono-Heck
product, and indeed its independent synthesis and subsequent
exposure to the coupling conditions failed to produce naphthy-
ridone. Instead, it produced bis-Heck product in the presence
of acrylanilide 18, and only decomposition in its absence. The
use of phosphine-based ligands resulted in polymerization of
acrylanilide 18 and produced little naphthyridone or Heck
products.
Suzuki Coupling. The Suzuki-Miyaura coupling between
chloro-naphthyridone 17 and difluorophenyl boronic acid (26)
was initially carried out in refluxing toluene using Pd(dppf)Cl2
as the coupling catalyst. This reaction required the use of 2-2.5
equiv of boronic acid 26 to achieve complete consumption of
chloronaphthyridone 17. It was subsequently found that under
biphasic conditions using 1 mol % Pd(OAc)2, 3 mol % Ph3P,
10 mol % KCl, 10 mol % tetrabutylammonium chloride, toluene,
and aqueous Na2CO3 at 90 °C for 3 h (Scheme 6), the reaction
could be completed with the use of 1.5 equiv of boronic acid
26, affording 14 in 96% reaction yield and 91% isolated yield
after crystallization on 3 kg scale. Although the goal was to
reduce the use of expensive commercial boronic acid, this was
partially offset with the use of tetrabutylammonium chloride.
The preparation of boronic acid 26 from the corresponding
bromide and isopropylborate produced boronic acid that resulted
in faster reactions requiring less of the reagent. The rate
acceleration was traced to the presence of isopropyl alcohol in
the reaction mixture. Subsequently, simpler conditions were
developed that afforded the same yields accomplished above
using 1 mol % Pd(OAc)2, 2 mol % Ph3P, Na2CO3 in 2-propanol,
using only 1.1 equiv of boronic acid 26.
N-Oxidation Reaction. With 14 in hand, the N-oxidation
reaction was screened with respect to oxidants. Initial attempts
to oxidize 14 using MCPBA in acetonitrile at 55-60 °C resulted
in a maximum of 88% conversion. Subsequent control studies
revealed that N-oxide 15 was unstable at elevated temperatures
in that it tended to revert back to starting material.16 Therefore,
the N-oxidation was best performed at 22 °C. A solvent screen
identified several leads. The best conversion (97.6%) was
attained using CH2Cl2 and 2 equiv of MCPBA at 22 °C
overnight (Table 1, entry 1). Because CH2Cl2 is generally
avoided in large-scale reactions, toluene was used on large scale,
achieving a 96.6% conversion (see Table 1, entry 5) and a
reaction yield of 96%. A highly streamlined isolation was also
developed to take advantage of the low solubility of 15 (1 mg/
mL) and high solubility of MCBA (110 mg/mL) in MTBE.
Thus, four volumes of MTBE were added as anti-solvent to
the reaction mixture (Scheme 7), and by stirring at -10 °C,
crystalline 15 was obtained as an MTBE solvate, free of MCBA,
in high yield and excellent purity (98.5 LCAP; 93% recovery).17
The overall isolated yield for the step was 88% on 2 kg scale.
(16) (a) Kurbatova, A. S.; Kurbatov, Yu. V.; Avezov, M. R.; Otrosh-
chenko, O. S.; Sadykov, A. S. Nauchn. Tr. Samark. Gos. UniV. 1969, 167,
38-42. (b) Barnes, J. H.; Hartley, F. R.; Jones, C. E. L. Tetrahedron 1982,
38, 3277-3280.
J. Org. Chem, Vol. 71, No. 22, 2006 8605

Chung et al.
SCHEME 8. Preparation of Chloride 13 via
Transamination (Method 1)
SCHEME 9. Preparation of Chloride 13 via Iminium Ion
(Method 2)
SCHEME 10. Grignard Formation and NMR Chemical
Shifts of 12 and 13 in THF-d8
Treatment of N-oxide 15 with POCl3 then produced chloronaph-
thyridone 16 in 85% yield.18
Synthesis of 4-Chloro-1-tert-butylpiperidine (13). 4-Chloro-
N-tert-butylpiperidine was prepared from N-methyl-4-piperidone
(method 1) or from 4-chloropiperidine (method 2).19 In the first
method, the tert-butyl group was installed via transamination
of quaternary salt 27 with tert-butylamine to afford N-tert-butyl-
4-piperidone (28) (Scheme 8). For this reaction, we found
efficient and rapid removal of dimethylamine was crucial for
high yield due to product instability. Removal of dimethylamine
via physical means such as molecular sieve 13X, K2CO3, or
distillation worked well on small to mid-scales. Yet on multi-
kilogram scale, for example, molecular sieve 13X method gave
lower yields (52% distilled isolated yield on 35 mol scale vs
77% on 1 mol scale) due to decomposition from the extended
reaction time required. This is presumably due to a smaller
surface-to-volume ratio in a conventional scale-up set up. After
the scale-up, we found sodium acrylate to be an excellent,
selective dimethylamine Michael acceptor over tert-butylamine.
Under an optimized condition with water as the solvent, this
reaction was demonstrated on 100 g scale in 87% isolated yield.
Product 28 obtained from the above reactions was then
subjected to a Raney Ni reduction to afford hydroxy piperidine
29 in 98% yield. Conversion of 29 to chloride 13 using thionyl
chloride was initially met with difficulties with as much as 50%
yield of olefin 30 being generated. Subsequently, 13 was
prepared in 85% yield in the presence of 0.5 equiv of
tetrabutylammonium chloride (TBAC), which suppressed the
formation of elimination product 30 (2%), presumably via
trapping the intermediate carbonium ion. This step was suc-
cessfully demonstrated on multi-kilogram scale, and the crude
product could be used for Grignard generation without further
purification (Scheme 10). The overall yield for the four-step
process was 65-70%, and the only potential issue for this route
was the high cost of TBAC used in the chlorination step.
As an alternate route to 13, we explored the generation of a
tert-butyl group by the reaction of a dimethyliminium salt
prepared from 4-chloropiperidine with methyl Grignard (Scheme
9). The advantage of this iminium route is that the starting
material already contains the halogen, obviating the need for
the carbonyl reduction/alcohol chlorination sequence used in
the piperidone-based route. Initial work determined that hydro-
chloride salt 20b itself would not undergo iminium formation
due to insufficient solubility. Examination of alternative acid
salts revealed that the triflic acid and hydrogen bromide salts
both gave efficient iminium formation. The latter was chosen
for further development because of its useful physical properties.
Iminium bromide 31 can be isolated by simple filtration of the
reaction mixture and is relatively stable in the absence of water.
A through-process was developed in which hydrochloride salt
20b is subjected to counterion exchange with sodium bromide
and then converted to iminium salt 31 in the same pot. Treatment
of 31 with methylmagnesium chloride in THF at -15 °C was
found to afford good yields of 13 along with 5-10% of
4-chloropiperidine, which can be separated during extraction
by treatment with mild acid. The secondary amine side-product
results from deprotonation of iminium salt 31 to give an enamine
(observed by NMR) that is hydrolyzed upon aqueous workup.
This iminium route was demonstrated in multiple runs of 50-
100 g. Piperidine 13 was isolated in 71% overall yield by
aqueous workup and was obtained in high purity (99% by GC).
Grignard Formation. The Grignard formation of 4-chloro-
N-tert-butylpiperidine (13) using magnesium metal was best
initiated by addition of 5-10 mol % of isopropylmagnesium
chloride20 and heating the mixture in refluxing THF for 2-5 h
to consume all starting material. Typically, starting from a 90:
10 mixture of 13 and olefin 30, the yield of Grignard 12 was
about 90% as an 84:8:8 mixture of 12, 30, and 32 based on
NMR and titration analyses. The NMR chemical shifts of the
carbon and proton next to magnesium chloride in 12 were
significantly upfield shifted relative to the chloro 13, consistent
with reported values (Scheme 10).
(17) Naphthyridone N-oxide 15 MTBE solvate exhibited an exotherm
of 111.6 cal/g with an initiation temperature of 163 °C. The differential
thermal analyzer also detected an exotherm starting at ∼160 °C. The
compound is not shock sensitive based on drop weight measurements.
(18) Compound 16 has also been prepared by an alternate route: Chung,
J. Y. L.; Cai, C.; McWilliams, J. C.; Reamer, R. A.; Dormer, P. G.;
Cvetovich, R. J. J. Org. Chem. 2005, 70, 10342-10347.
(19) (a) Amato, J. S.; Chung, J. Y. L.; Cvetovich, R. J.; Reamer, R. A.;
Zhao, D.; Zhou, G.; Gong, X. Org. Process Res. DeV. 2004, 8, 939-941.
(b) Amato, J. S.; Chung, J. Y. L.; Cvetovich, R. J.; Gong, X.; McLaughlin,
M.; Reamer, R. A. J. Org. Chem. 2005, 70, 1930-1933.
(20) Knochel, P.; Dohle, W.; Gommermann, N.; Kneisel, F. F.; Kopp,
F.; Korn, T.; Sapountzis, I.; Vu, V. A. Angew. Chem., Int. Ed. 2003, 42,
4302-4320.
8606 J. Org. Chem., Vol. 71, No. 22, 2006

Synthesis of a Naphthyridone p38 MAP Kinase Inhibitor
SCHEME 11. Attempted Cross Coupling
SCHEME 13. Reactions of Compounds 14 and 16 with
Grignard 12
SCHEME 12. Reaction of 14 and Grignards with TMSOTf
or Phenyl Chloroformate
environment around the pyridine nitrogen, both favoring selec-
tive activation of the ene-lactam system.
Next, additions of Grignard 12 to these substrates without
activation were investigated (Scheme 13). In the reactions with
14 and 16, the major products were dimeric adducts 39 and 40,
respectively, based on LC/MS analysis. It appeared that 1,4-
addition is still the preferred pathway, and the resulting enolate
intermediates, which previously were quenched by trimethylsilyl
triflate or the chloroformate, now underwent a second 1,4-
addition to produce the proposed dimeric structures in Scheme
13.
Significantly, when N-oxide 15 was treated with 1.1 equiv
of Grignard 12, the reaction proceeded in a titration-like manner
to give a single product identified as dihydropyridine 41a by
LC/MS analysis (Scheme 14). Attempts to isolate 41b from the
deep burgundy solution led to significant decomposition, which
is not surprising because dihydropyridines are air-unstable.
Various in-situ dehydrating conditions (refluxing THF, solid
NaOH, aqueous HCl) and reagents (MsCl, Me2NCOCl, Ac2O,
TMSCl) were examined to aromatize the dihydropyridine ring.
Isobutyl chloroformate was found to be the best activator, which
when added to 41a rapidly made a ∼1:1 mixture of 1 and
carbonate 42 (identified by LC/MS, but not isolated). Reversing
the order of Grignard and isobutyl chloroformate addition led
to poorer yields. Addition of KOtBu or NaOEt to this 1:1
mixture did not improve the yield of 1. Heating at 110 °C in
nBuOH consumed 42 but did not yield additional 1; instead,
alcohol 43 was formed in significant amounts (which was
crystallized and confirmed by NMR and LC/MS). Heating in
N-methylpyrrolidone (NMP) at 110 °C gave an improved 75%
reaction yield along with several impurities. Ultimately, switch-
ing the solvent to pyridine and heating at 110 °C for 9 h afforded
1 in 92% reaction yield!
These observations could be rationalized by the supposition
that 42 exists as a 1:1 mixture of E and Z isomers with respect
to the carbonate and the piperidyl groups (Scheme 15). The
Z-carbonate can undergo a base-induced anti-elimination at low
temperatures, whereas anti-elimination is not possible for the
E-carbonate and requires a syn-elimination via thermolysis.
An efficient isolation and purification of 1 from the Grignard-
thermolysis reaction mixture was also developed. The crude 1
HCl salt was directly crystallized from the reaction mixture in
99% recovery by adding aqueous NH4Cl solution. The organic
phase efficiently removed many impurities and color. The free
base was prepared with aqueous Na2CO3 and extraction with
IPAC, whereupon 1 crystallized in 92% recovery and 80%
overall isolated yield from N-oxide 15 with a purity of >98.5%
by HPLC. The structure of 1 was confirmed via NMR and
The “all-in” nature of this Grignard preparation necessitated
a safety assessment before this step could be run in large scale.
The results from the calorimetry study indicated that the
Grignard formation generated a steady output of heat over a 3
h period and was therefore unlikely to suffer a thermal runaway.
Consequently, a large run (11.4 mol scale) was conducted
following the same “all-in” protocol. Unlike small-scale lab runs
of this reaction, no isopropylmagnesium chloride was required
to initiate Grignard formation, and after several hours at reflux
a dark brown solution of desired Grignard 12 was furnished.
NMR analysis showed that olefin 30 had increased from 2% to
around 10% during Grignard formation. Titration assay of the
Grignard solution gave a concentration of 0.52 M (versus an
expected 0.6 M) for an 87% yield.
Installation of N-tert-Butylpiperidine. With both piperidine
Grignard 12 and naphthyridone 14-16 in hand, we proceeded
to study the coupling of these two pieces. Initial cross-coupling
studies of chloro-naphthyridone 16 with isopropylmagnesium
chloride or Grignard 12 under Fe or Pd catalysis did not give
the desired products. Instead, the 1,4-addition product, lactam
33, was the major product (Scheme 11).
We then turned our attention to substrate 14. Activation of
pyridine nitrogen of 14 using phenyl chloroformate or tert-
butyldimethylsilyl triflate followed by the addition of Grignard
reagents 12 was studied (Scheme 12). In all cases, the desired
products (e.g., 35) were not obtained. Again, the 1,4-addition
adducts were the main products: saturated amide 37 when the
silicon activator was used, and compound 38 in the case of
chloroformate. Possibly, the strongly electron-withdrawing
difluorophenyl group rendered the pyridine nitrogen less reactive
than the lactam oxygen in these examples. Additionally, the
reaction utilizing the bulky silicon activator cleanly produced
37, a fact probably attributable to the combined effects of the
oxophilic character of silicon and the sterically encumbered
J. Org. Chem, Vol. 71, No. 22, 2006 8607

Chung et al.
SCHEME 14. Conversion of N-Oxide 15 to Compound 1
SCHEME 15. Elimination Pathways of Carbonate 42
The progress of the reaction was followed by HPLC condition
A (see Supporting Information), and the reaction was stopped when
the chloronaphthyridone 17 was < 0.1%. When the reaction was
complete (about 2-2.5 h), the solution was cooled to 32 °C, and
CH2Cl2 (11 L) was added. The solution was filtered through 0.86
kg of Solka Floc to remove palladium. The reactor was rinsed with
toluene (12 L), which was then passed through the Solka Floc. The
filtrate was transferred to a 100 L extractor and the phases separated.
The organic layer was washed with 10% aqueous NaCl (16 L),
then water (10 L). The organic layer was assayed via HPLC (96%
yield), and the solution was concentrated to dryness.
EtOAc (16.5 L) was charged to the crude product (4.3 mL/g),
and the mixture was heated to 70 °C to dissolve all solids. After
20 min at 70 °C, the solution was cooled to 0 °C, and crystallization
ensued, aged for 2 h, and filtered. The wet cake was washed with
cold EtOAc (3 × 1.7 L) and vacuum-dried at 50 °C overnight under
a stream of N2 to afford 3.39 kg (91%) of compound 14: 1H NMR
(CDCl3) δ 8.54 (d, J ) 5.8, 1H), 7.73 (dd, J ) 3.5, 10.0, 1H),
7.65-7.55 (m, 3H), 7.49 (dd, J ) 7.6, 8.7, 1H), 7.10 (dt, J ) 2.0,
8.1, 1H), 7.02 (dt, J ) 2.0, 9.8, 1H), 6.80 (d, J ) 10.0, 1H), 6.45
(d, J ) 5.8, 1H); 13C NMR (CDCl3) δ 164.0 (dd, JCF ) 12.0, 251.7),
160.2, 160.0 (dd, JCF )12.0, 251.7), 154.0, 149.8, 145.1, 138.1 (d,
JCF ) 3.1), 135.1, 133.3 (dd, JCF ) 4.5, 9.8), 132.4, 131.5, 129.6,
123.5, 122.4 (dd, JCF ) 3.9, 15.3), 115.3, 112.4 (dd, JCF ) 3.8,
21.5), 108.6, 104.6 (t, JCF ) 25.6); 19F NMR (CDCl3) δ -108.0
(d, JFF ) 8.6), -110.6 (d, JFF ) 8.6); HRMS m/z [M + H]+ calcd
for C20H11Cl2F2N2O 403.0216, found 403.0215.
Suzuki-Miyaura Coupling (Method 2). A 250 mL, round-
bottom flask equipped with a magnetic stirrer, thermocouple, and
N2/vacuum inlet adapter was charged sequentially with chloronaph-
thyridone 17 (3.91 g, 12.0 mmol), 2,4-difluorophenylboronic acid
(2.27 g, 14.4 mmol), Pd(OAc)2 (26.9 mg, 0.12 mmol), Ph3P (94.3
mg, 0.36 mmol), Na2CO3 (2.54 g, 24.0 mmol), and 2-propanol (60
mL). The suspension was degassed by cycling vacuum and N2 three
times. The resulting white suspension was aged at 80 °C for 12 h.
When the chloronaphthyridone 17 was all consumed (assay by
HPLC condition A), the suspension was cooled to 50 °C and acetone
(60 mL) was added. The suspension was cooled to 20 °C, and the
solids were removed by filtration to provide a brown solution of
Suzuki adduct 14 (136 mL containing 4.84 g, 100%). The Pd can
be removed by treatment of this solution with silacycle thiol-3, MP-
TMT, or Ecosorb C-933. Addition of one volume of water and
cooling to 0 °C provided the Suzuki adduct in >95% isolated yield.
single-crystal X-ray analysis. This process was successfully
demonstrated on multi-kilogram scale.
In summary, we have developed an efficient synthesis of 1
that is amenable for large-scale synthesis. The key steps included
a tandem Heck-lactamization and a highly efficient tert-
butylpiperidinyl Grignard addition to naphthyridone N-oxide.
The N-oxide exerted complete chemoselectivity via chelation
in directing the Grignard addition to the R position as opposed
to a 1,4-addition on the ene-lactam. The optimized conditions
to aromatize the dihydropyridine intermediate were achieved
with isobutylchloroformate followed by thermolysis in pyridine.
The Grignard precursor, N-tert-butyl-4-chloro-piperidine, was
efficiently prepared by a transamination method or via a methyl
Grignard addition to the iminium ion. Utilizing this chemistry,
multi-kilogram preparation of the crystalline compound 1 was
successfully demonstrated.
Experimental Section
Suzuki-Miyaura Coupling. Preparation of 1-(2,6-Dichlo-
rophenyl)-5-(2,4-difluorophenyl)-1,6-naphthyridin-2(1H)-one (14)
(Method 1). A 72 L, round-bottom flask equipped with a mechan-
ical stirrer, thermocouple, condenser, and N2 inlet was charged
sequentially with toluene (15.6 L), chloronaphthyridone 17 (3.00
kg, 9.23 mol), tetrabutylammonium chloride (264 g, 0.95 mol), KCl
(70.8 g, 0.95 mol), 2,4-difluorophenyl boronic acid (1.81 kg, 11.46
mol), Pd(OAc)2 (20.7 g, 0.092 mol), PPh3 (72.5 g, 0.277 mol), and
additional toluene (4 L). The mixture was degassed three times by
cycling vaccum/N2, and stirred for 15 min at rt under N2. The slurry
was cooled to 10 °C with an ice-water bath. Na2CO3 (2.97 kg,
28.02 mol) and water (15 L) were added (an exotherm to 20 °C
was observed), and the slurry was heated to reflux (∼90 °C).
Preparation of 1-(2,6-Dichlorophenyl)-5-(2,4-difluorophenyl)-
1,6-naphthyridin-2(1H)-one 6-Oxide (15). A 100 L cylindrical
extractor equipped with a mechanical stirrer and thermocouple was
charged sequentially with toluene (10.2 L), difluorophenyl-naph-
thyridone 14 (1.70 kg, 4.21 mol), and 70 wt % MCPBA (1.063 kg,
8608 J. Org. Chem., Vol. 71, No. 22, 2006

Synthesis of a Naphthyridone p38 MAP Kinase Inhibitor
6.16 mol). A solution results after 15 min at 25 °C. The progress
of the reaction was followed by HPLC (condition B), and the
reaction was stopped when the difluorophenyl-naphthyridone 14
was less than 4 mol % left. When the reaction was complete (about
18 h), MTBE (40.8 L) was charged to the cloudy mixture. The
solution was stirred for 30 min at rt, then cooled to -10 °C and
aged for 18 h. When the product concentration in the supernatant
was under 2.1 mg/mL, the mixture was filtered and the wet cake
was washed with cold MTBE (3 × 5 L). The wet cake was vacuum-
dried at rt under a stream of air for 5 h to afford 1882 g of N-oxide
as a yellow solid. The purity of the solid was 81.5 wt % (88%
yield) containing 13.3 wt % MTBE and 3.4 wt % toluene as
determined by NMR and HPLC analysis. This material was used
as is in the next step. A solvent-free analytical sample was obtained
by prolonged vacuum-drying: 1H NMR (CDCl3) δ 8.30 (d, J )
7.5, 1H), 7.63 (m, 1H), 7.61 (m, 1H), 7.45-7.57 (m, 2H), 7.44
(dd, J ) 1.0, 10.0, 1H), 7.14 (dt, J ) 2.3, 8.2, 1H), 7.09 (dt, J )
2.3, 9.2,1H), 6.85 (d, J ) 10.0, 1H), 6.49 (d, J ) 7.5, 1H); 13C
consumption of the Grignard adduct intermediate. This THF
solution of carbonate 42/compound 1 could be stored at 20 °C under
nitrogen overnight without problem.
THF was distilled (70-95 °C/ambient pressure) and replaced
with anhydrous pyridine (17.5 L). The resulting slurry was heated
at 110 °C for 10 h. The solution was cooled to 20 °C, LC assayed
to confirm the 1 to carbonate 42 ratio was g98% (at 210 nm), and
concentrated to e1/2 volumes (∼9.5 L) at 30-35 °C/20-25 Torr.
Saturated aqueous NH4Cl (20 L) and water (10 L) were added,
and the resulting mixture (pH 8) was adjusted to pH 5.8 with 3 N
HCl (∼7.4 L). IPAC (5.8 L) and MTBE (5.8 L) were added, and
the mixture was stirred at 20 °C for 1.5 h, and then filtered and
1
washed with /2 saturated NaCl solution (7.8 L) and MTBE (7.8
L). The wet cake was vacuum-dried to give ∼4.5 kg of crude 1
HCl salt as a beige solid.
Compound 1 HCl salt was charged into a mixture of IPAC
(39 L) and H2O (23 L), and 3.75 equiv of solid Na2CO3
monohydrate (2.26 kg, 18.2 mol) was added (to pH ∼10). After
the mixture was stirred at 20 °C for 1 h, the layers were separated,
and the organic was washed with H2O (2 × 23 L) and assayed by
HPLC for yield (92%). The organic was filtered, concentrated (10-
25 °C /28-30 in. Hg), and then seeded with 1 to induce
crystallization if required. The mixture was concentrated and flushed
with IPAC until H2O e 100 ppm. The volume was adjusted such
that 1 concentration was ∼0.33 g/L (final vol ∼7.5 L). Heptane
(25 L) was slowly added over 4 h to give a target ratio of ∼18%
IPAC/heptane and a supernatant concentration of e9 g/L. The
mixture was stirred at 20 °C overnight, filtered, and washed with
10% IPAC/heptane (10 L). The wet cake was vacuum-dried at 20
°C under N2 to afford 2.1 kg of 1 free base (80% yield): purity
NMR (CDCl3) δ 164.6 (dd, JCF ) 12.3, 253.6), 160.8 (dd, JCF
12.3, 253.6), 158.9, 142.3, 141.1, 136.3, 135.8, 135.0 (d, JCF
14.4), 133.2 (dd, JCF ) 4.7, 9.6), 131.8, 131.6, 129.7 (d, JCF
8.0), 125.8, 118.0, 113.7 (dd, JCF ) 4.6, 16.0), 112.6 (d, JCF
)
)
)
)
3.0), 112.5 (dd, JCF ) 3.7, 21.6), 111.5, 105.3 (t, JCF ) 25.0);
HRMS m/z [M + H]+ calcd for C20H11Cl2F2N2O2 419.0160, found
419.0165.
Grignard Formation: Preparation of (1-tert-Butylpiperidin-
4-yl)(chloro)magnesium (12). To a flask containing magnesium
metal (291 g, 12.0 mol) was added THF (11.8 L, H2O < 50 µg/
mL). A solution (H2O < 150 µg/mL) of chloride 13 (30.2 wt % in
THF, 6.62 kg, 7.2 L, 11.4 mol) was then added over 5 min. The
brown heterogeneous mixture was then heated in 10 °C stages
toward 65 °C. At 35 °C, the reaction was initiated and self-heated
to 65 °C and began to reflux steadily. The steam bath was then set
to maintain reflux for 5 h to ensure complete conversion to
Grignard. The resultant dark brown solution of Grignard reagent
was then titrated to determine the molarity. Molarity was determined
as 0.52 M versus a theoretical 0.6 M (87% yield). Use tests
confirmed the reagent was of sufficient quality for use in subsequent
steps. See Scheme 9 for partial NMR data in THF-d8.
Preparation of 7-(1-tert-Butylpiperidin-4-yl)-1-(2,6-dichloro-
phenyl)-5-(2,4-difluorophenyl)-1,6-naphthyridin-2(1H)-one (1).
To a yellow suspension of naphthyridone N-oxide 15 (2500 g,
81.5%, 4.86 mol) in THF (24.3 L), cooled to 2 °C, was added 1.1
equiv of Grignard 12 solution (10.3 L, 0.52 M in THF, 5.35 mol)
over 1.5 h at <3 °C to give a deep burgundy solution. After being
stirred for 1 h at 2-5 °C, the starting material was completely
consumed as judged by HPLC (<0.35%; condition C), and 1.5
equiv of isobutyl chloroformate (0.965 L, 7.29 mol) was slowly
added over 1.5 h at <8 °C. The mixture was warmed to 20 °C
over 1-2 h and aged at this temperature for 1 h, and the resulting
orange-red solution was assayed by HPLC to confirm complete
1
99.6%; H NMR (CD3CN) δ 7.74 (ddd, J ) 9.8, 3.4, 0.8, 1H),
7.71 (d, J ) 7.9, 2H), 7.63-7.59 (om, 2H), 7.20-7.14 (om, 2H),
6.65 (d, J ) 9.8, 1H), 6.34 (s, 1H), 3.09-3.06 (m, 2H), 2.61 (tt, J
) 12.1, 4.2, 1H), 2.10 (td, J ) 11.7, 2.3, 2H), 1.77-1.73 (m, 2H),
1.66-1.59 (m, 2H), 1.01 (s, 9H); 13C NMR (CD3CN) δ 167.9, 164.7
(dd, JCF ) 249.0, 12.2), 161.1, 160.9 (dd, JCF ) 248.4, 12.2), 154.1,
146.4, 139.6 (d, JCF ) 3.1), 135.5, 134.4 (dd, JCF ) 9.8, 4.9), 133.3,
133.0, 130.7, 123.8 (dd, JCF ) 15.3, 4.3), 122.7, 114.2, 113.1 (dd,
JCF ) 21.4, 3.7), 106.4, 105.2 (t, JCF ) 26.2), 54.3, 47.0, 45.6,
33.7, 26.5; HRMS m/z [M + H]+ calcd for C29H28Cl2F2N3O
542.1577, found 542.1576.
Acknowledgment. We thank Mr. Robert Reamer for as-
sistance with NMR assignments, and Ms. Jennifer Chilenski
for obtaining the X-ray structure.
Supporting Information Available: General experimental
methods, NMR spectra for compounds 14, 15, and 16, and X-ray
crystal structure data in CIF format for compound 1. This material
is available free of charge via the Internet at http://pubs.acs.org.
JO061618F
J. Org. Chem, Vol. 71, No. 22, 2006 8609